Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sertoli cell microtubules: their polarity and binding to spermatid associated ectoplasmic specializations Redenbach, Darlene Marie 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_redenbach_darlene_ma.pdf [ 8.21MB ]
JSON: 831-1.0086715.json
JSON-LD: 831-1.0086715-ld.json
RDF/XML (Pretty): 831-1.0086715-rdf.xml
RDF/JSON: 831-1.0086715-rdf.json
Turtle: 831-1.0086715-turtle.txt
N-Triples: 831-1.0086715-rdf-ntriples.txt
Original Record: 831-1.0086715-source.json
Full Text

Full Text

SERTOLI CELL MItROTUBULES:THEIR POLARITY AND BINDINGTO SPERMATID-ASSOCIATEDECTOPLASMIC SPECIALIZATIONSbyDARLENE MARIE REDENBACHB.S.R., The University of British Columbia, 1982M.Sc., The University of British Columbia, 1986A THESIS SUBMITTED IN PARTIAL FULFULLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Anatomy)We accept this thesis as conformingto the required s)aF1-a1rdTHE UNIVERSITY OF BRITISH COLUMBIAApril, 1992© Darlene Marie Redenbach, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of________________The University of British ColumbiaVancouver, CanadaDate- .2-DE-6 (2/88)ABSTRACTDuring spermatogenesis, spermatogenic cells are moved through the blood testisbarrier from the basal to the apical compartment, where they become oriented parallelto the long axis of the Sertoli cell, and situated within Sertoli cell crypts. They aremoved again toward the base of the epithelium, before being translocated across theepithelium for release into the tubule lumen. Crypts are lined with unique actincontaining submembrane structures called ectoplasmic specializations (ESs) that formpart of the Sertoli cell-spermatid junction. ESs consist of the Sertoli cell membrane, afenestrated cistern of smooth endoplasmic reticulum, and a highly ordered interveningarray of actin filaments. ESs are thought to participate in establishing junctionaldomains at Sertoli cell-spermatid adhesion junctions, which serve to anchor thedeveloping spermatids within the Sertoli cell crypts. Sertoli cell microtubules occuradjacent to the endoplasmic reticulum of the ES (ESER), oriented parallel to the longaxis of the cell, and to the direction of spermatid translocation. Other investigators havedescribed linkages between the ESER and adjacent microtubules. Sertoli cellmicrotubules have been suggested to aid in orientation and positioning of spermatogeniccells, within the seminiferous epithelium. It is proposed that this may be achieved by amicrotubule-based transport mechanism known to be involved in establishing andmaintaining organelle positioning in other cells. As part of a study to test the hypothesisthat spermatid translocation is a microtubule-based event, the polarity of Sertoli cellmicrotubules was determined. The potential for binding between spermatid-ESs andmicrotubules was assayed, and the binding characterized, using a selection of conditionsknown to alter organelle-microtubule interaction in other systems. The results of thisstudy indicate that Sertoli cell microtubules are orientated with their minus-enddirected toward the apical surface of the cell and that microtubules bind to spermatid-EScomplexes, are releasable in the presence of nucleotides, and share binding propertieswith known mechanoenzymes. These results are consistent with the hypothesis thatspermatids are moved through the seminiferous epithelium by a microtubule-basedtransport mechanism.IIITABLE OF CONTENTSABSTRACT.i ITABLE OF CONTENTS iiiUST OF TABLES xLISTOF FIGURES xiABBREVIATIONS x iiiACKNOWLEDGEMENTS x ivCHAPTER 1: INTRODUCTION 1OVERVIEW 2BACKGROUND 6TESTIS 6Cellular organization of the testis 6Function of the testis 6Terminology of spermatogenesis 1 4Positional changes during spermatogenesis 1 4SPERMATIDS 1 6Overview of sperm iogenesis 1 6Morphological changes during spermiogenesis 1 6Positional changes of spermatids during spermiogenesis 2 0Regulation of spermatogenic cell development 23SERTOU CELLS 24General morphology 24Membrane bounded organelles 25Mitochondria 28Exocytotic and Endocytotic compartments 2 8Golgi apparatus 29Endoplasmic reticulum 2 9Cytoskeleton 32Microfilaments 32Intermediate filaments 33Microtubules 34Ectoplasmic specializations 34Organization 3 7Relationship with cytoskeleton 3 8Linkages 39ESER 39ivStage specific changes .4 0Functions 41Regulation 4 2Changing events in Sertoli cells during spermatogenesis 43SERTOLI-GERM CELL RELATIONS 44Processes 44Lateral processes 44Apical processes 45Penetrating processes 45Tubulobulbar processes 46Junctions 47Blood testis barrier 4 7Sertoli cell-Spermatid junction 4 8MICROTUBULES 49Introductory statement 49Microtubule structure and kinetics 5 0Microtubule organizing centers (MTOC5) 5 1Microtubule structures 53Tubulin isoforms 53Post-translational modification of tubulin 5 3Microtubule associated proteins (MAPs) 54Microtubule polarity 56Microtubule based intracellular transport 5 7Mechanoenzymes 6 1Kinesin 62molecular structure 6 3biochemical properties 63sensitivity to inhibitors 64rate and direction of transport 64other properties 65Cytoplasmic dynein 65molecular structure 66biochemical properties 66sensitivity to inhibitors 6 6rate and direction of transport 6 7other properties 6 7Other putative microtubule based motors for organelle transport 6 7visiken 67VlOS sea urchin egg motor .68Reticulomyxa motor 6 8C. elegans motor 6 8Kar 3 and ncd non claret disjunctional gene product 6 9Organelle-microtubule binding proteins 69170 kD protein from Hela cells 69The mechanochemical cycles of kinesin and cytoplasmic dynein 70Mechanoenzyme inhibitors and how they work 73EHNA 74NEM 74Vanadate 74AMPPNP 77Binding Assays 77The organelle-microtubule binding complex 79Where is the motor located9 80How are motors kept available for bidirectional transport9 8 1Are the motors alone sufficient for transport9 8 2How are motors regulated’ 83Microtubule-dependent organelle positioning 84Cellular organization (microtubules in cells in general) 8 7MICROTUBULES IN SERTOLI CELLS 90Distribution 90Sertoli cell MTOC 9 3Tubulin post translational modifications in Sertoli cells 9 4Stage dependent changes in Sertoli cell microtubule distribution 94Sertoli cell MAPs 96Relationship of microtubules to other organelles in Sertoli cells 97Function of Sertoli cell microtubules 9 7Sertoli cell shape 9 8Organelle positioning and translocation during spermatogenesis 1 00Influence of Sertoli cell microtubules on the head shape of 101spermatogenic cellsMicrotubule perturbation 1 02DEVELOPMENT OF THE HYPOTHESIS 104HYPOTHESIS STATEMENT 1 08PREDICTIONS 1 08EXPERIMENTAL OUTUNE 1 09Microtubule polarity study 1 09viMicrotubule-spermatid-ES binding study .1 09CHAPTER 2: MICROTUBULE POLARITY 111INTRODUCTiON 112DETERMINATION OF MICROTUBULE POLARITY: THE APPROACH 11 2DETERMINATION OF MICROTUBULE POLARITY: THE STRATEGY 11 3MATERIALS AND METHODS 115MATERIALS 11 5Animals 11 5Chemicals and supplies 11 5Buffers 115METHODS 116Preparation of purified tubulin 11 6Hook decoration 11 7Electron microscopy 1 22Negative stain electron microscopy 1 25Video enhanced differential interference contrast microscopy 1 25RESULTS 127MICROTUBULE POLARITY 1 27Effects of lysis decoration buffer 1 27Sampling for montage 1 36Scoring criteria 1 36Microtubule polarity: observations 1 41Microtubule polarity: counts 1 4 1DISCUSSION 151MICROTUBULE POLARITY IN SERTOLI CELLS 1 51SUMMARY 153CHAPTER 3: BINDING ASSAY 158INTRODUCTION 158BINDING ASSAY CRITERIA: COMPONENTS 158Spermatid-ES 158Labelled microtubules 1 63BINDING ASSAY: ESTABLISHING THE CRITERIA 1 63Development of the binding assay 1 63Establishing that counts represented microtubules 1 64CHARACTERISTICS OF BINDING 1 64Reversal of binding 1 64viiExperiments to characterize 3MI and spermatid-ESs binding 1 64LOCALIZATION OF LABEL 1 65Morphological support for results of binding assay 1 65CALCIUM AS A PROPOSED REGULATOR 1 67MATERIALS AND METHODS 168MATERIALS 1 68Animals 1 68Chemicals and reagents 1 68General 168Buffers 168Primary antibodies 169Secondary antibodies 1 69Binding assay treatment reagents 1 69METHODS 170General 170Protein determinations 1 70SDS-PAGE gels 1 70Isolation of 3MTx- spermatid-ES binding assay components 170Tubulin purification for binding assay 1 70Preparation of 3HGTP labelled, taxol stabilized microtubules (3MTx)1 72Length measurement of3HGTP labelled, taxol stabilized 1 72microtubules (3MT)Spermatid-ES isolation 1 73Binding assay: preparation 1 76Sample preparation for binding assay 1 76Gradient preparation for binding assay 1 76Binding assay: methods 1 79Running the binding assays 1 79Binding assays: experimental design 1 79Design of binding assay experiments (general) 1 79Experiments for which total sample added to gradients was not 150 tl180Direct effect of treatments on microtubules 180Design of matched experiments 1 8 1Design of [MT] experiment 1 8 1Topical binding assay 1 85Preparation of treatment materials used in binding assays 1 85Isolation of rat testis crude supernatant for cytoplasmic 185dynein enriched MAP preparationVIIIIsolation of cytoplasmic dynein enriched MAP preparation from 186testis crude supernatantSertoli Cell enriched isolation from 21 day old rats 1 87Preparation of cytosol: Sertoli cell enriched preparation from 1 8821 day rat testisData collection for binding assays 1 88Data analysis 1 89Immunohistochemistry 1 89Immunofluoresence: sample preparation and microscopy 1 89(5A6: antitubulin)Rhodamine phalloidin staining 1 90Preparation of normal mouse lgG: control for 5A6 1 90anti-tubulin antibodyAutoradiography of binding assay 1 9 1Autoradiography: methods 1 91Electron Microscopy 1 92Routine electron microscopy 1 92Methods for calcium study 193RESULTS 196COMPONENTS OF THE BINDING ASSAY 1 96Spermatid-ES5 196Ectoplasmic specialization endoplasmic reticulum: ESE R 202Endogenous microtubules remain attached during warm isolation 202Microtubule isolation and measurement 21 3Labelling exogenous microtubules 21 3Incorporation of label and stability of microtubules over time 21 6Microtubule stability: effect of agents for characterization 2 1 6of microtubule bindingBINDING ASSAY: ESTABLISHING THE CRITERIA 21 6Buffer conditions 21 7Spermatid-ES enrichment by gradients 218Controls: “same spin controls” in every spin 224Sampling 224BINDING ASSAY: RESULTS 231Microtubules bind to spermatid-ESs 231Label entering the gradient with spermatid-ESs is 231primarily from microtubulesThere is a time course to 3MTx - spermatid-ES binding 232ixNucleotides reverse 3MTx - spermatid-ES binding 233The effect of ATP depletion on3M.I-spermatid-ES binding 245Microtubule concentration has a nonlinear effect on binding 245- spermatid-ES binding is dynamic 249Characterization of 3MTx - spermatid-Es binding 254LOCALIZATION OF LABEL TO SPERMATID HEADS 258Labelled microtubules are localized to the spermatid-ES 258Actin and tubulin dual staining of amorphous clusters 266Pursuit of a proposed regulation of events around the spermatid-ES 271DISCUSSION 280COMPONENTS OF THE BINDING ASSAY 280Spermatid-ESs 280THE BINDING ASSAY MEASURES MICROTUBULE- SPERMATID-ES BINDING 282Effects of the sucrose gradient 282Counts represent microtubule binding 282Binding time course 283Effects of added nucleotide 284Competition by unlabelled microtubules 285CHARACTERIZATION OF BINDING 285LOCALIZATION OF3MT-SPERMATID-ES BINDING 286Label associates with spermatid heads 286Localization to amorphous clusters 287ANTIMONY PRECIPITATION EVIDENCE OF CALCIUM IN ESER IS INCONCLUSIVE 287CHAPTER 4: DISCUSSION 289INTRODUCTORY REMARKS 290MICROTUBULE POLARITY IN SERTOLI CELLS 2902MIxSPERMATID-ES BINDING 297BIBUOGRAPHY 308xLIST OF TABLESCHAPTER 2Table Il-I Classification of microtubules according to hook decoration 1 42cold/warm incubationTable Il-lI Percent of microtubules with clockwise, counterclockwise and 1 43ambiguous hook decoration: cold/warm incubation.Table Il-Ill Classification of microtubules according to hook decoration 1 48exogenous tubulin/warm incubationTable II-IV Percent of microtubules with clockwise, counterclockwise and 1 48ambiguous hook decoration: exogenous tubulin/warmincubationCHAPTER 3Table Ill-I Sample data from binding assay 229Table Ill-Il Non-microtubule component of binding: summary 235Table Ill-Ill Effect of the addition of excess cold label: 5X cold taxol stabilized 252microtubules on3MT-spermatid-ES bindingTable lll-IV Effect of the non-microtubule component of 5X cold taxol stabilized ..252microtubule sample on3MT-spermatid-ES bindingTable Ill-V Characterization of3MT-spermatid-ES binding 257xiUST OF FIGURESCHAPTER 1Fig. 1-1: Organization of the testis 8Fig. 1-2: Stages of the seminiferous epithelium 11Fig. 1-3 Sertoli cell Regulation 13Fig. 1-4: Spermiogenesis 1 8Fig. 1-5 Spermatid translocation in the seminiferous epithelium 22Fig. 1-6: Sertoli cell fine features 27Fig. 1-7: Ectoplasmic specializations 3 7Fig. 1 -8: Mechanoenzyme cycles 72Fig. 1-9: Mechanoenzyme inhibitors 76Fig. 1-10: Continuous organelle circulatory system: radial model 89Fig. 1-11: Microtubule distribution in Sertoli cells 92Fig. 1-12: Microtubule-based spermatid translocation model 1 07CHAPTER 2Fig. 2-1: Evidence of purity and polymerization capability of bovine brain 11 9tubulinFig. 2-2: Microtubule decoration: sampling methods for EM sections 1 24Fig. 2-3: Verification of spermatid tail axoneme orientation 129Fig. 2-4: Rat seminiferous epithelium: highly extracted following incubation in ..1 3 1lysis/decoration bufferFig. 2-5: Low power view of an area of seminiferous epithelium that has been 1 33incubated in lysis decoration buffer.Fig. 2-6: Rat seminiferous epithelium moderately extracted in 1 35lysis/decoration bufferFig. 2-7: Rat seminiferous epithelium only slightly extracted in 138lysis/decoration bufferFig. 2-8: Montage of rat seminiferous epithelium incubated in lysis decoration ....1 40bufferFig. 2-9: Montage of rat seminiferous epithelium incubated in warm 1 47lysis/decoration bufferFig. 2-10: Hook decoration of seminiferous epithelium, incubated 150immediately at 35°C.Fig. 2-11: Summary diagram of microtubule polarity results 1 56CHAPTER 3Fig. 3-1: Isolation of spermatids from seminiferous epithelium with intact ESs. .1 60Fig. 3-2: Transillumination methods of seminiferous cycle staging 1 62Fig. 3-3: Spermatid-ES isolation 1 75Fig. 3-4: Binding assay: methods 1 78Fig. 3-5: Microtubule concentration experiment: strategy 1 83Fig. 3-6: Appearance of spermaid-ES isolate 1 98Fig. 3-7: Actin staining associated with the heads of isolated spermatids 201Spermatid ESs.Fig. 3-8: ESER remains attached to spermatid head following spermatid-ES 205isolation.Fig. 3-9: Tubulin staining pattern on spermatid-ESs under warm or cold 207conditions.Fig. 3-10: Control figure for 5A6 antibody staining for tubulin (1) 209Fig. 3-1 1: Control figure for 5A6 antibody staining for tubulin (2) 2 1 2Fig. 3-12: Length and fine structure of taxol stabilized microtubules used for 215XIIMT-spermatid- ESsFig. 3-13: Testing for microtubule migration using sucrose gradients in 220MT-spermatid-ESs binding assays.Fig. 3-14: Establishing spin criteria for binding assay: SDS page gels 222Fig. 3-15: Establishing centrifugation conditions for binding 223assay using3MTX counts.Fig. 3-16: Centrifugation of crude spermatid-ES isolate on a discontinuous 22630-45-60% sucrose gradient results in enrichmentof spermatid-ESs.Fig. 3-17: Binding assay method: short sample experiment 228Fig. 3-18: Results sample binding assay 230Fig. 3-19: Non-microtubule component of binding 234Fig. 3-20: Time course for3MT-spermatid-ES binding: overlay plots 237Fig. 3-21: Time course for MT-spermatid-ES binding: collapsed data 239Fig. 3-22: Effect of 10 mM MgATP on3MT-spermatid-ES binding 241Fig. 3-23: Reduction in tubulin staining on MT-spermatid-ES binding after 243treatment with 10 mM MgATP.Fig. 3-24: Comparison of effects of MgATP and GTP on3MT-spermatid-ES 244bindingFig. 3-25: ATP depletion effect on3MT-spermatid-ES binding 246Fig. 3-26: Effect of varying spermatid-ES concentration on3MT-spermatid-ES .248binding.Fig. 3-27: Effect of varying microtubule concentration on MT-spermatid-ES ....250binding.Fig. 3-28: 3MT-Spermatid-ES binding expressed as a function of microtubule ....251concentration.Fig. 3-29: Characterization of3MT-spermatid-ES binding: matched experiments.256Fig. 3-30: Localization of label in MT-spermatid-ES binding from pre-gradient26lautoradiography samples.Fig. 3-31: DIC and bright field micrograph pairs to show localization of label in ...2633MT-spermatid-ES binding from post-gradientautoradiography samples.Fig. 3-32: DIC and bright field micrograph pairs to show localization of label in ....2653MTX- spermatid-ES binding in the presence of 10mM MgATP frompost-gradient autoradiography samplesFig. 3-33: Actin staining of amorphous clusters, like those seen in the 268autoradiography slides, show that spermatid heads are presentin the clusters.Fig. 3-34: MT-spermatid-ESs stain for both actin and tubulin 270Fig. 3-35: Localization of calcium in seminiferous epithelium (Method 1) 275Fig. 3-36: Localization of calcium in seminiferous epithelium (Method 2) 277Fig. 3-37: Localization of calcium in seminiferous epithelium (Method 4) 279CHAPTER 4Fig. 4-1: Summary diagram of microtubule-based spermatid translocation 305modelFig. 4-2: Summary diagram of microtubule-spermatid-ES binding 307XIIIABBREVIATIONS5A6 monoclonal antibody to alpha tubulinADP adenosine diphosphateAMPPNP adenylimidodiphosphate: nonhydrolyzable analogue of ATPATP adenosin triphosphateDIG differential interference contrast microscopyEHNA erythro-9-[3(2hydroxynonyl)] adenine analogue of adenosineEM electron microscopyER endoplasmic reticulumectoplasmic specializationESER endoplasmic reticulum of ectoplasmic specializationFSH follicle stimulating hormoneGTP guanosine triphosphateLH leutenizing hormoneLM light microscopyMAPs m icrotubu le associated proteins3MTx microtubules labelled with tritium GTP (3), and stabilized withtaxol3MTxspermatid-ES spermatid-ESs complexes with bound microtubulesMT microtubuleNEM N-ethylmaleimide, sulfhydryl alkylating agentinorganic phosphatespermatid-ES spermatids, isolated with intact ectoplasmic specializationsTuJ1 monoclonal antibody to beta tubulinxivACKNOWLEDGEMENTSFirst, I would like to thank my husband, Jak and my children, Karoline,Katharine and Jason, for their patience and understanding during the pursuit of this goal.Without their love, and encouragement this would not have been possible.A dept of gratitude is owed to the members of my supervisory committee: Drs.Wayne yogi, Nelly Auersperg, Bruce Crawford, Harold Kasinsky, and Kim Boekelheide,for their guidance, friendship, and support through this study. In addition, the manyhours spent in sharing their experience and their equipment, and in reading this thesisare gratefully acknowledged.I would especially like to thank my mentor, Dr. A. Wayne Vogl, for his support,and to acknowledge his enthusiasm and dedication to the pursuit of science, that hasinstilled in me an appreciation for the process of science. His pursuit of excellence inscience and teaching provides an inspiration to all his students and colleagues.I would like to thank Dr. C. Slonecker, the faculty, staff and students in theAnatomy department at UBC, in particular David Pfeiffer, for their friendship, fruitfuldiscussions and many good times.I am very grateful to Dr. Kim Boekeiheide and the members of his lab, in thedepartment of Pathology and Laboratory Medicine at Brown University, that were sogenerous with their friendship, and expertise during the completion of this study.Appreciation is due, Franke Crymble and Jak Redenbach for excellent art work.z -Ic•) I0-UC- m-I 5-zIntroduction 2OVERVIEWSertoli cells have been described as nurse or nurturing cells for developing malegerm cells. They make up the columnar cells of the seminiferous epithelium acrosswhich developing germ cells must be translocated before they can be released into thetubule lumen. The dependence of germ cells on Sertoli cells, implied by the intimaterelationship between these two cell types, has long been assumed and continuinginvestigation supports that conclusion. A great deal of work has been done to describe theevents of spermatogenesis and elucidate the underlying mechanisms, yet precisely whatrole the Sertoli cell plays and how it may serve to nurture the developing germ cell isnot entirely understood.Sertoli cell-germ cell relations are highly complex and constantly undergoingelaborate changes, including the formation of a variety of Sertoli-Sertoli and Sertoligerm cell junctions during spermatogenesis (Nicander, 1967; Gilula et al., 1976;Gravis, 1979; Dym and Fawcett, 1970; Russell et al., 1983b; Russell and Peterson,1985; Russell, 1977a, 1980; Grove and VogI, 1989). Unique actin containingstructures called ectoplasmic specializations (ESs) form within Sertoli cells, adjacentto regions of attachment to spermatids, lining the Sertoli cell crypts that hold developingspermatids (Brokelmann, 1963; Flickinger and Fawcett, 1967; Nicander, 1967;Russell, 1977a; yogI, 1989; yogI et al., 1991a,b). ESs are thought to assist inestablishing membrane domains for the adhesion junction and have been proposed to playa role in the orientating and positioning of spermatids (yogI, 1989; Grove and VogI,1989; Grove et at., 1990; VogI et at., 1991a).ESs consist of an hexagonal array of actin filaments sandwiched between theSertoli cell plasma membrane and a fenestrated cistern of endoplasmic reticulum (forreviews see VogI, 1989; VogI, et al., 1991a). Bundles of Sertoli cell microtubulesoccur adjacent to the endoplasmic reticulum of the ES (ESER) (Fawcett, 1975; Russell,1977c; VogI, 1988, 1989; Amlani and VogI, 1988; VogI, et al., 1991a). Themicrotubules surround elongating spermatids and are oriented parallel to the long axisintroduction 3of the Sertoli cell and to the direction of spermatid translocation (Fawcett, 1975; VogI,1988, 1989; Amlani and VogI, 1988, Neely and Boekelheide, 1988; Hermo et al.,1991; Redenbach and yogI, 1991). Significantly, ESs occur during the orientation andtranslocation of spermatids and are dismantled just prior to the release of latespermatids into the tubule lumen (Russell, 1977b,c, 1984). The microtubules haveboth a temporal and structural relationship to ESs, suggesting a coordinated function.The close association of microtubules with ESER is reminiscent of other systemsin which membrane bound organelles, observed adjacent to microtubules, can bedemonstrated to translocate along microtubule tracks. Much attention has been focusedon microtubule based transport in axons (Smith, 1972; Lasek and Brady, 1985; seeOkabe and Hirokawa, 1989; Sheetz et al., 1989) and, more recently, in a number ofother cell types (Allen et al., 1981b; Koonce and Schliwa, 1985; Hayden et al., 1983;McNiven et al., 1984; Steuer et al., 1990; see Kelly, 1990b; Kreis, 1990; Schroer andSheetz, 1991a). It is becoming increasingly apparent that microtubule-based transportis fundamental to a surprising number of housekeeping functions in cells generally(Terasaki, et al., 1984; van der Sluijs et al., 1990; Duden et al., 1990; see Kelly,1990b; Kreis, 1990; Schroer and Sheetz, 1991a), in addition to being the underlyingmechanism of many highly specialized events (Steuer et al., 1990; Troutt and Burnside,1988a; Vallee et al., 1989b). Microtubule associated proteins, and in particularmechanoenzymes or ‘motors’, are thought to mediate microtubule dependent events, andare currently under intense investigation (Vallee et al., 1989a,b; Schroer et al., 1989;Balch; Vale, 1990; Schroer and Sheetz, 1991a; Schroer and Sheetz, 1989; Sloboda andGilbert, 1989; Kelly, 1990a;. There is recent evidence that other microtubule bindingproteins may also function to position organelles (Rickard and Kreis, 1990, 1991;Scheel and Kreis, 1991a,b).A system for translocating spermatogenic cells, in general, through theseminiferous epithelium has the following requirements: 1) developing spermatogeniccells must be translocated from the base to the apex of the seminiferous epithelium,Introduction 4without breaking the seal that maintains the exclusive basal and apical environments; 2)provision must be made for the temporal and spatial economy to simultaneouslyaccommodate more than one stage of maturation of spermatogenic cells within the apicalenvironment; 3) spermatids must complete their differentiation and yet be released onlywhen they are precisely ready; 4) a mechanism is required for orientation andmovement of spermatids situated within Sertoli cell crypts.A number of occurrences in Sertoli cells address these requirements.Compartmentalization is established by the blood testis barrier (Dym and Fawcett,1970), a tight junction formed between lateral processes of Sertoli cells that serves toseparate the apical and basal environments. The translocating spermatogenic cells areshuttled toward the tubule lumen by the formation of Sertoli cell lateral processes thatsegregate the germ cells from subsequent generations of translocating cells (Russell,1977b; Weber et al., 1983). Spermatids are retained in apical crypts of Sertoli cellsby an adhesive junction, the ectoplasmic specialization, which is progressivelydismantled prior to spermatid release. As spermatids undergo differentiation, they areprecisely oriented and positioned, being returned within their crypts toward the base ofthe Sertoli cell, then translocated to the tubule lumen for release. One factor thatremains to be explained is the mechanism by which spermatids are oriented andpositioned in the seminiferous epithelium.The means of providing spermatid orientation and positioning may lie in a novelapplication of an ubiquitous mechanism: microtubule-based transport. A potentialinteraction between the ER of ESs (ESER) and adjacent bundles of microtubules, orientedin the direction of germ cell translocation, is the basis of the microtubule-basedspermatid translocation model proposed in this study. The model is consistent with whatis known of Sertoli cell morphology and of microtubule-based transport. Amicrotubule-based motility model for orientation, positioning, and translocation ofspermatids, is feasible whether these events are viewed as passive on the part of thespermatid, or reciprocally regulated between spermatids and Sertoli cells. BecauseIntroduction 5spermatids do not appear to move independently, the model satisfies the conditions of ahighly active roll for Sertoli cells while not excluding a feedback role for spermatids.Most importantly, it accommodates what is known of the morphology of Sertoli cells; inparticular, at their interface with developing germ cells. It is supported by analogy toknown systems of intracellular transport; however, it introduces what may be a uniqueexample of intracellular transport, the regulation of an extracellular event.This thesis is presented in four parts. Chapter 1 provides the backgroundinformation for the study as a whole, the development of the hypothesis, and theapproach used to test the hypothesis. Chapter 2 describes the Sertoli cell microtubulepolarity portion of the study and Chapter 3 the assay developed to examine the potentialbinding of microtubules to spermatid-ESs, each with introduction, methods, results, anddiscussion sections pertaining to that portion of the study. Chapter 4 provides adiscussion of the significance of the findings from chapters 2 and 3 and their relevanceto the proposed hypothesis.The introduction, in this chapter, begins with a description of the events ofspermatogenesis and the morphological features of spermatids and Sertoli cells that willbe encountered in this study. These introductory sections are followed by a detaileddescription of microtubules and their function in microtubule-based transport,generally. What is currently known about microtubules in Sertoli cells is thenreviewed and a model developed for their proposed function in microtubule-basedspermatid translocation. Finally, the rationale and approach to testing this model ispresented.Introduction 6BACKGROUNDTESTISCellular organization of the testisThe mammalian testis consists of closely packed seminiferous tubules andintervening interstitial tissue enclosed in a connective tissue capsule. Seminiferoustubules are comprised of fluid filled tubes lined by seminiferous epithelium andensheathed by a squamous sheet of myoid cells. Fig. 1-1 shows the cellular organizationof the testis and the seminiferous epithelium, in the rat. Unless otherwise indicated,descriptions apply to the animal model used in this study, rat testis. The seminiferousepithelium is a modified columnar epithelium consisting of two cell types, a sessile,non-proliferating population of ‘nurse cells’, or Sertoli cells, and a highly proliferativepopulation of spermatogenic cells, undergoing complex morphological and positionalchanges (Sertoli, 1865; Leblond and Clermont, 1952a,b; Perey et al., 1961; Russell,1977c; yogI et al., 1983a; Vogl et al., 1991a,b). The interstitial compartment iscomprised of clusters of steroidogenic Leydig cells in addition to vascular and lymphatictissue. The structural compartmentalization of the testis, into seminiferous tubules andinterstitial tissue, reflects its functional duality.Function of the testisThe mature mammalian testis participates in both exocrine and endocrineactivities. Its primary exocrine function, spermatogenesis, involves the production andrelease of spermatozoa under the nurturing influence of Sertoli cells. Its endocrinefunction, steroidogenesis, is carried out primarily in the interstitial compartment byLeydig cells.Spermatogenesis has been described in detail in a number of species(Brökelmann, 1963; Leblond and Clermont, 1952a,b; Clermont, 1972; Perey et al.,1961; Russell et al., 1990). In rat, full maturation takes approximately 52 daysIntroduction 7Figure 1-1: Organization of the testis: This light micrograph shows a cross-section of rat testis with a number of seminiferous tubules and the interstitial tissue.The seminiferous tubules are composed of a modified columnar epithelium (seminiferousepithelium) which includes Sertoli cells and intervening spermatogenic cells. Sertolicells extend from the base to the tubule lumen identified by highly infolded nuclei withprominent nucleoli, situated at the base of the epithelium. Spermatogenic cells aresituated between the Sertoli cells and include spermatagonia that lie along the base of theepithelium, spermatocytes, identified by large round nuclei found midway between thebase and the lumen of the epithelium, and spermatids, recognized by their darklystaining nuclei. Elongate spermatids are found deep within the epithelium from stages IIto VI, being deepest at stage V, and extend into the tubule lumen by stage VIII. They canbe identified by their elongate nuclei during these stages. Released spermatozoa are seenin the tubule lumen. Surrounding the seminiferous tubule are myoid cells. Theinterstitial compartment is made up of vascular elements, and clusters of round Leydigcells. Seminiferous tubules at different stages of development are included in this field.(Micrograph courtesy of W. yogI.):--‘‘:-.:•-•—-•.•..•,•.•__.,•1i...‘1..‘I‘•/•••.1•’I•.0Introduction 9(Clermont, 1972). Before one generation of spermatogenic cells have completedspermatogenesis, the subsequent generation begins, resulting in a highly coordinatedstacking of spermatogenic cells from four to five generations. The time between theinitiation of one generation of spermatogenesis to the initiation of the next is one ‘cycle’of the seminiferous epithelium, requires about 13 days (Clermont, 1972) and producesspecific ‘cell associations’ (Leblond and Clermont 1 952a). These ‘cell associations’arise from precisely coordinated morphological and positional changes in spermatogeniccells as they progress together through the seminiferous epithelium. Their invariablecoexistence has been used to distinguish the 14 ‘stages’ of spermatogenesis,conventionally used to identify events in spermatogenesis. Fig. 1-2 shows the cycle ofthe seminiferous epithelium the the cell associations that occur at each stage (adaptedfrom Perey et al., 1961).Sertoli cells undergo morphological and physiological, stage dependent, changesin order to provide the changing milieu required to nurture spermatogenic cells atdifferent stages of development. Presumably, all spermatogenic cells in a given cellassociation thrive in a similar environment.The endocrine function of the testis is steroidogenesis, mainly the province ofLeydig cells (see Means et al., 1980; Purvis and Hansson, 1981). Fig. 1-3 shows thefundamental feedback mechanisms by which steroidogenic functions of the testis areregulated (adapted from Purvis and Hansson, 1981). The pituitary releases two trophichormones that influence testicular function: follicle stimulating hormone (FSH) andluteinizing hormone (LH). Stimulated by LH, Leydig cells synthesize and secretetestosterone which concentrates to very high levels in the seminiferous epithelium.FSH, through the second messenger system and LH, directly influence protein synthesis,secretion, and enzyme activity in Sertoli cells which in turn regulate spermatogenesis.Sertoli cells secrete tubule fluid and a number of proteins including: androgen bindingprotein, that binds testosterone to keep local levels high, inhibin, to regulate FSH levels,plasminogen activator, which is involved in the cycle of tissue resorption, and a numberintroduction 1 0Figure 1-2: Stages of the seminiferous epithelium: This chart shows 14 cellassociations, in vertical columns (roman numerals) that invariably coexist in theseminiferous epithelium of the rat. The 14 cell associations occur consecutively in theorder shown here, making up the 14 stages of the cycle of the seminiferous epitheliumin the rat, as described by Leblond and Clermont (1 952a). Four and one half cycles arerequired to complete spermatogenesis; however, after one cycle, the next generation ofspermatogenic cells are initiated resulting in the accumulation of five generations ofgerm cells in the epithelium by the time the mature germ cells are released. ‘A’indicates type A spermatogonia, ‘In’, intermediate spermatogonia, and ‘B’, type Bspermatogonia. Primary spermatocytes are indicated as ‘P1’ (preleptotene), ‘L’(leptotene), ‘Z’ (zygotene), ‘P’ (pachytene), ‘Di’ (diplotene) and secondaryspermtocytes ‘II’. The 19 steps of spermiogenesis occur from stage I through stage XIVof the 4th cycle and are completed by stage VIII of the fifth cycle. (This chart is modifiedfrom Perey et al., 1961. Details shown in the developing spermatids are adapted fromClermont and Rambourg 1978; Russell et al.,1990).itXlixIA1%AlIllII“IxIXXllIlAHAIntroduction 1 2Figure 1-3: Sertoli cell regulation: This diagram gives an overview of thehormonal regulation of the testis. The stimulatory effects of the steroid hormone:testosterone (from Leydig cells) and the protein hormone: follicle stimulating hormone(from the pituitary gland) activate protein synthesis, protein secretion, and enzymeactivity. The effects of follicle stimulating hormone are activated through a secondmessenger system and are proposed to be modulated, in part, by calcium sequestrationand release by the ER (adapted from Purvis and Hansson, 1981).‘3nhHnsec.r€4ory rocessepro4&n e.515enyrne. cv4ypkosphDrIae. kiru’scEo ke1eAonrcne)orgons- E2.e— b&ro.vioraeffe.cA.sone.Introduction 1 4of other proteins. Of particular interest here is the second messenger cascade, initiatedby FSH, that triggers the many functions under calcium/calmodulin modulation,including proposed cytoskeleton regulation. The response of Sertoli and Leydig cells totrophic hormone involves inter-organ, intercellular and even intracellular feedbackmechanisms that modulate the response (Purvis and Hansson, 1981). The completepicture is infinitely complex and includes paracrine regulatory events by whichspermatogenic cells are thought to influence Sertoli cell activities, both of which areonly partially understood (Saez et al., 1985).Terminology of spermatogenesisA complex terminology has evolved to distinguish the spermatogenic cell typesand their maturation. Spermatogenesis can be subdivided into three stages,spermatocytogenesis (proliferative phase), meiosis, and spermatid differentiation(spermiogenesis). ‘Spermatogonia’ include stem cells and early mitotic spermatogeniccells. Cells undergoing the first and second meiotic divisions are ‘primary and secondaryspermatocytes’ respectively. At the completion of meiosis, they become ‘spermatids’that undergo morphological differentiation and are released as ‘spermatozoa’ from theepithelium. ‘Spermatogenesis’ is the complete process of transformation ofspermatogonia into spermatozoa. ‘Spermatocytogenesis’ is the process by whichspermatogonia become spermatocytes and includes a number of mitotic divisions.“Meiosis” is the formation of haploid spermatids from spermatocytes. ‘Spermiogenesis’is the differentiation of spermatids which culminates at ‘spermiation’, the release ofmature germ cells into the tubule lumen, at which point they are referred to as‘spermatozoa”(see Leblond and Clermont, 1 952a).Positional changes during spermatogenesisDuring spermatocytogenesis, spermatogonia reside on the basal lamina of theseminiferous epithelium between Sertoli cells. They undergo mitosis, after which theyIntroduction 1 5assume one of two fates, to remain as stem cells, or become committed to developmentand begin their migration toward the tubule lumen. The committed spermatogonia carryout a number of incomplete mitotic divisions, thereafter remaining linked with theirclones by intercellular bridges, first described by Von Ebner (see Perey et al., 1961;Gondos, 1984). They then begin the next phase of spermatogenesis, meiosis.The area between Sertoli cells, in which spermatogenic cells migrate, is dividedinto apical (or adluminal) and basal compartments by a very tight occluding junction,the blood-testis barrier, (Dym and Fawcett, 1970). The barrier is formed betweenneighbouring Sertoli cells, separating the environment of the the basal compartment,accessible to blood borne substances, from that of the adluminal compartment, accessibleonly to Sertoli cell lumenal secretions.During the second phase, meiosis, primary spermatocytes synchronously beginthe first of two meiotic divisions as they pass from the basal compartment through theblood testis barrier, to the adlumenal compartment. This is achieved, partly, by theextension of Sertoli cell lateral processes between the basal lamina and the migratingcell, to re-establish the tight junction of the blood testis barrier, thereby separatingmigrating spermatocytes from the basal compartment. Once the seal is complete, thebarrier above the spermatocyte is disassembled and the spermatocyte occupies theadluminal compartment (Clermont, 1972). During the pachytene stage of the firstmeiotic division, chromosome crossover occurs; spermatocytes are renderedimmunologically different. Having been isolated from their vascular supply, they aredependent on the Sertoli cell for nourishment and protection from immunological attack.After pre-leptotene spermatocytes arrive in the adluminal compartment, they completethe first meiotic division and then, as secondary spermatocytes, undergo a rapid secondmeiotic division to become early, or round haploid spermatids, ending the meiotic stage.During the third stage, spermiogenesis, spermatids undergo differentiation,become oriented to the long axis of Sertoli cells, and are moved from their lateralposition between Sertoli cells to become embedded in apical orypts within individualIntroduction 1 6Sertoli cells. During their encasement in the crypts, spermatids are brought deeperwithin the Sertoli cell, then returned to the apical region of the Sertoli cell, and releasedinto the tubule lumen. The reason for this return to a basal position is not known. Thereis no evidence to suggest germ cells play an active role in their positioning ortrans location.After spermiation, spermatozoa become motile and continue to undergo subtlechanges as they complete their journey through the male reproductive tract.SPERM ATI DSThe steps of spermioaenesisBased on morphological changes in differentiating spermatids, spermiogenesishas been divided into 19 ‘steps’ (not to be confused with stages) (Leblond and Clermont,1952a,b; Clermont, 1972; for detailed account of staging criteria, see Russell et al.,1990). The occurrence of these 19 steps is shown in Fig. 1-2. The morphologicalchanges that occur during the 19 spermatid steps (Arabic numerals) are illustrated inFig. 1-4.During spermiogenesis, two types of changes occur in spermatogenic cells,morphological and positional.Morphological changes in spermatids during spermiogenesisThe 19 steps of spermiogenesis, are grouped into four phases: Golgi phase, capphase, acrosome phase, and maturation phase (Leblond and Clermont, 1952; for detailssee Russell, et al. 1990) (Figure 1-4). The Golgi phase (stages I to Ill) begins thedifferentiation of round spermatids and is characterized by initiation of the the futureaxoneme from the distal centriole and the formation of an acrosome granule from theGolgi. During the cap phase (stages IV to VII), a head cap grows out of the acrosomalgranule and spreads over the surface of the nucleus describing a 140° arc by the end ofIntroduction 1 7Figure 1-4: Spermiogenesis: This diagram shows the 19 recognizable steps ofspermiogenesis that occur in four phases (LeBlond and Clermont, 1952a): Golgi phase,cap phase, acrosomal phase, and maturation phase. See text for details of morphologicalfeatures of the 19 stages. ( G) Golgi, (AV) acrosomal vesical, (N) nucleus, (T) tail, (C)cap, (M) manchette, (A) acrosome, (VF) ventral fin, (DF) dorsal fin, (MIT)mitochondria, (RB) residual body. (The diagram is adapted from Leblond and Clermont,1952a; Clermont and Rambourg, 1978; Russell et al., 1990).I819\. iiN‘7GoI& ha.sec.c&f psecLc.rosornc.t phc*.se.rc4.ion‘9Introduction 1 9stage VII (Russell et al., 1990); meanwhile, the paired centrioles move closer to thenucleus, with the proximal one becoming attached to the nucleus and the distal oneassociated with the developing axoneme, the future tail. The next phase sees thebeginning of spermatid elongation. At the onset of the acrosome phase (stages VIII toXIV), the acrosome becomes oriented toward the base of the epithelium and the cytoplasmtoward the tubule lumen, with the acrosomal cap diametrically opposed to the developingtail. The caudal migration of spermatid cytoplasm results in an anterior displacement ofthe acrosome capped nucleus until it abuts against its rostral plasma membraneimmediately adjacent to the neighbouring Sertoli cell. A sheath of microtubules, themanchette, begins to assemble from a point midway along the nucleus extending caudallyover the developing tail. After the beginning of manchette development in step 8 (Cole,1988; Russell, 1990), the nucleus begins condensation and elongation. By the end of theacrosome phase, the acrosome extends well over the nucleus with a space between theacrosome and nucleus ventrally, becoming pointed at the apical end. The nucleus reachesits maximum length at step 14, shortening slightly thereafter. By stage XIV, a ridgeforms on the dorsal aspect of the acrosome. The caudal displacement of spermatidcytoplasm continues to be maneuvered into an isolated lobe of residual cytoplasm,approaching the manchette which is beginning to dismantle. At the end of stage XIV, anew generation of spermatids begin spermiogenesis. The maturation phase begins at stepXV and sees the brief appearance of a dorsal fin on the acrosome. This marks thebeginning of the assembly of the middle piece. At step 17, the acrosome moves forwardand cytoplasm accumulates in the concave side of the head. During step 18, thecytoplasm becomes segregated into a residual body which is retained by the Sertoli cellat spermiation. The nucleus and acrosome become increasingly curved during steps 18and 19. Stages Ito XIV coincide with steps 1-14. Step 15 coincides with stage I, step16 with stages II and Ill, step 17 with stages IV and V, step 18 with stage VI, and step 19with stage VII, followed by release of the free spermatozoa after stage VIII.Introduction 20Spermatogenic cells form elaborate, specialized, microtubule structures, onlysome of whose functions are readily apparent. Microtubules form mitotic and meioticspindles that act as scaffolds for chromosome movement during cell division. In addition,an elaborate and extensively cross linked sheath of microtubules, the manchette, extendsposteriorly over the caudal portion of elongating spermatids, surrounding the developingaxoneme (Rattner and Brinkley, 1972; Cole et al., 1988). The function of manchettesremains controversial. Flagellar axonemes form the basis of the structurally complexspermatid tail that transduces microtubule-microtubule interaction into spermmotility. Arms, that extend from the microtubule doublets that form the outercytoskeleton of the axoneme, are the basis of its kinetic properties. The arms are anaxonemal form of the mechanoenzyme, dynein. Studies on the mechanochemicalproperties of axonemal dynein, and their role in sperm motility have provided a basisfor the characterization of cytoplasmic mechanoenzymes (Tash, 1989). Isolated anddemembranated axonemes have been used to study microtubule based motility (Paschaland Vallee, 1987) and have been used as a source of stable microtubules in studies ofmicrotubule dynamics (Borisy and Bergen, 1982).Positional chanaes of spermatids during spermiogenesisDifferentiating spermatids (step 8) become oriented perpendicular to the base ofthe epithelium in stage VIII. At this time, they are moved from the lateral compartmentand become situated within a crypt of an individual Sertoli cell (Russell, et al., 1983a).Within Sertoli cell crypts, spermatids become oriented to the long axis of the Sertolicell (stage VIII step 8) (Figs 1-2 and 1-5). The crypt deepens, carrying the spermatidonce more toward the base of the Sertoli cell, passing spermatogenic cells of subsequentgenerations that are being moved toward the lumen. The basal migration occurs betweenstages II and VII with elongating spermatids reaching their most basal position duringstages IV - V (see Clermont, 1972). These events are illustrated in Fig 1-5 (adaptedfrom Clermont, 1972; Clermont and Rambourg, 1978 and Leblond andIntroduction 2 1Figure 1-5: Spermatid translocation in the seminiferous epithelium: Thisdiagram depicts the stages of the seminiferous epithelium during which spermatids aremoved from an apical, to a basal position, deep within Sertoli cell crypts and returned tothe lumen for release. In each figure a Sertoli cell can be recognized by its infoldednucleus (S). Numbers indicate the steps of spermiogenesis for spermatids during thesestages. Other cell types include spermatogonia (A)type A, (In) intermediate; (B) typeB, (mM) intermediate in mitosis, (BM) type B in mitosis, and spermatocytes (P1)preleptotene and (P) pachytene. Arabic numbers indicate the steps of spermiogenesis ofspermatids during these stages. The spermatids reach their deepest position during stageV. Stages of spermatogenic cells are as indicated in Figure 1-2. (Adapted from:Clermont, 1972; Clermont and Ramborg, 1978; and Russell et al., 1990).112S4ae. IVS4ce. VVIIIIntroduction 23Clermont, 1952a), showing the relative positions of spermatogenic cells during thebasal and subsequent apical spermatid migration.A consequence of the positional changes of spermatogenic cells, that establishspecific cell associations, is the segregation of spermatid residual cytoplasm (Russell,1984). As the spermatid is moved toward the base of the epithelium, the residualcytoplasm becomes segregated from the elongating spermatid and anchored by processesextending from the Sertoli cell. As the spermatid returns to the tubule lumen, thespermatid nucleus is attached to the residual cytoplasm by a slender stalk. Duringspermiation, the residual cytoplasm is retained by the Sertoli cell and undergoes partialautolysis and then phagocytosis by the Sertoli cell (Russell, 1984).Very little is known about the mechanisms that bring about these positionalchanges; however, in the absence of any evidence of inherent motility in thespermatogenic cells, they are thought to be mediated by Sertoli cells. During the initialmigration toward the lumen the combined formation of Sertoli cell processes and theaddition of subsequent generations of spermatogenic cells may explain the lumenalmigration of spermatocytes. It has been suggested that unique actin containingstructures, ectoplasmic specializations, lining Sertoli cell crypts during the orientationand translocation of spermatids, may participate in these positional changes (Fawcett,1975; VogI et al., 1983a; Russell, 1977c; VogI, 1989).Regulation of spermatogenic cell developmentLittle is known about how the development of spermatogenic cells is programmedor regulated. There are bridges between spermatids of the same stage, making possible asyncytial development of clones (Gondos, 1964). What was initially thought to be awave of development passing along the tubule, suggesting a time line of morphogenesis,is no longer considered to result from the transmission of a wave like signal (Perey etal., 1961). Their stage-dependent changes in synthesis, secretion, and morphologyIntroduction 24imply that Sertoli cells play a very important role in the nourishment, and protection,as well as translocation and positioning of spermatogenic cells.SERTOLI CELLSOne of the first features to be described in Sertoli cells, was their intimaterelationship with spermatogenic cells (Sertoli, 1865), prompting the suggestion thatthey may play a role in the formation of spermatozoa. The responsibility of Sertoli cellsfor the nurturing of spermatogenic cells begins very early in development. EmbryonicSertoli cells arise from an undifferentiated blastema, not from the branching of a preexisting epithelium as is frequently thought, are present before Leydig cells, and play acentral role in the differentiation of the testis (Magre and Jost, 1991). In fetal testis,Sertoli cells carry out an endocrine function, secreting a Mullerian inhibitingsubstance, preventing spermatogenesis. Later, with the establishment of a basal tightjunctions and cell polarity, they assume an exocrine role, the production of spermatozoa.Mature Sertoli cells are non-proliferative, but exhibit stage specificphysiological and morphological variation. Not only are germ cell associations constantwith each other, but each cell association corresponds to specific Sertoli cell events.Although the causal relationships of most of these events is not known, it is thecoordination of events in Sertoli cells with stages of spermatogenesis, that has led to theassertion that Sertoli cells largely regulate the events of spermatogenesis. However, theinfluence is not all one way, with paracrine effects of germ cells on Sertoli cells alsoknown to occur (Saez, 1985; Kerr, 1988).General morphologySertoli cells are essentially columnar cells that are continually being modified incorrelation with changes in neighbouring spermatogenic cells. The fine structure ofSertoli cells has been studied extensively (Fawcett, 1975; Ritzen et al., 1981; Russellet al., 1983a; PlOen and Ritzen; 1984; Tindall et al, 1985; Kerr, 1988) and reflectsIntroduction 25their morphological and physiological support of spermatogenic cells. Fig. 1-6 is adiagramatic representation of the fine structural features found in Sertoli cells.Sertoli cells extend from the base to the lumen of the seminiferous tubule. Forpurposes of discussion, the cell can be divided into three regions: the infranuclear orbasal region, the nuclear region, and the supranuclear or apical region. The basal regionis a narrow band between the nucleus and the base of the cell, with the nuclear regionalso situated deep within the cell. The apical region includes the extensive area above thenucleus that forms the body or stalk of the cell and a variety of highly pleomorphicprocesses. The most elaborate changes in volume, shape, and placement of organelles,occur in the apical region of the cell, corresponding to changes occuring inspermatogenic cells (Kerr, 1988); yet it is the most highly organized region, with mostorganelles oriented to the long axis of the cell. The cell volume is greatest from stageXIII to XIV and lowest at stage VIII, with most changes occurring in the apical portion ofthe cell (Kerr, 1988).Sertoli cells can be easily discerned from adjacent spermatogonia by their highlyinfolded euchromatic nucleus, containing two perikaryosomes and a nucleolus, suggestiveof synthetic activity. Centrioles have been described in a supranuclear position(Nagano, 1966).Membrane bounded organellesThe presence of a large Golgi apparatus (Rambourg et al., 1979), lipid droplets,and elaborate accumulations of SER, argues for steroid synthesis in Sertoli cells,particularly in such species as ground squirrel (yogI, 1983a). However, the role ofSertoli cells in steroidogenesis is still not clear.Introduction 26Figure. 1-6: Sertoli cell fine features: This diagram illustrates the finefeatures of Sertoli cells. The figure is not intended to represent any specific stage ofspermatogenesis. Shown are spermatogenic cells surrounded by Sertoli cell processes.Note that in the apical portion of the cell, microtubules and membranous organelles areoriented parallel to the long axis of the cell; whereas, in the basal portion they are morerandomly distributed. Bundles of microtubules pass adjacent to the ESER of apical ESs.Spermatids are situated in Sertoli cell crypts that are lined by apical ESs. The bloodtestis barrier is closely associated with the basal ES. (SER) smooth endoplasmicreticulum, (IFs) intermediate filaments, (RER) rough endoplasmic reticulum, (mit)mitochondria, (ECM) extra cellular matrix, and (ES) ectoplasmic specialization.e. CS5J ° rour4 :\T3i1;, e spernc*Uc1GC WI l4eri,processoj11 1...:full,1,,IllIltuul Ittliflill I Ittltttttulttlttjljtttfttl,,ttE. C. W27pe.n&cra.-l jrprocessmicro-ubu.1e-—ItVIntroduction 28MitochondriaBasally, mitochondria are small and haphazardly organized, while apically theyare numerous, elongated, and oriented parallel to the long axis of the cell, closely packedbetween the abundant microtubules (Fawcett, 1975).Exocytotic and Endocytotic compartmentsSertoli cells synthesize and secrete a wide range of proteins (Ritzen et al., 1981Kerr, 1988) including inhibin, androgen binding protein and plasminogen activator.Fawcett (1975) describes a paucity of vacuoles or membrane bound granules that wouldsupport the extensive export of a secretory product.Morales and coworkers, (1986) describe an apical fluid phase endocytosis at allstages of spermatogenesis in which endocytotic vesicles fuse to become endosomes andtransform into acid phosphatase-positive secondary lysosomes. Exclusively duringstages VII through IX the endosomes fuse with phagocytosed residual bodies. Lysosomesincrease in number from stages X though XIV, decreasing from stages I to II, andremaining low from IV to VIII. This period also reflects a rapid turnover rate. Inaddition to elimination of their own end products of metabolism, Sertoli cells play a rolein the reduction and phagocytosis of spermatid residual cytoplasm. However, much ofthe residual cytoplasm is thought to undergo autolysis before it is phagocytosed by theSertoli cell. The increase in lysosomes coincides with the elimination of spermatidresidual cytoplasm. Ueno and Mon (1990) describe an earlier wave of primary thensecondary lysosomes, in a basal location, between stages IV to VI prior to the increasereported by Morales et al (1986). The explanation for this increase remainsspeculative.Introduction 29Golgi apparatusThe Golgi apparatus in Sertoli cells is well developed and found primarily in thebasal portion of the cell (Fawcett, 1975; Rambourg et al., 1979). It consists ofsaccular regions, showing classical cis and trans regions, interconnected into onecontinuous organelle (Rambourg et al, 1979), with an unusual relationship to ERelements situated at trans but not cis regions.The Golgi shows a biphasic, stage dependent change in volume, maximum at stageVIII and XIII to XIV and minimum at stage VII and lX-Xl (Ueno and Mori, 1990). Thischange in volume is accompanied by a repositioning from the usual basal location of theGolgi to a more apical location, as described by Rambourg and coworkers in a 3D model ofGolgi in stage V-Vlll Sertoli cells (Rambourg et al., 1979). The authors suggest thatthis may be in response to an increased need for glycoproteins with changes in lysosomeand plasma membrane in these regions apically, and that it is dependent on the presenceof late spermatids (Ueno et al, 1991). In that positioning of the Golgi is microtubuledependent (see Kreis, 1990), with the Golgi being maintained at the microtubuleminus-end in rat and mouse pituitary and monkey kidney immortalized cells (Skoufiaset al., 1990), microtubules could be expected to play a role in the stage dependentrelocation of the Golgi in Sertoli cells.Endoplasmic reticulumIn Sertoli cells, rough endoplasmic reticulum and free ribosomes, located mainlyat the base of the cell, are not prominent features (Nagano, 1966; Fawcett, 1975). Thevolume of RER is greatest at the time of spermatid release and lowest in stage XIII (Uenoand Mori, 1990; Kerr, 1988).Sertoli cell SER demonstrates species and stage dependent variation, both involume and distribution (Nagano, 1966; Fawcett, 1975; yogI, et al., 1983a; Pudney,1986; Clermont et al., 1980). Its greatest volume (three times the minimum volume)Introduction 3 0occuring at stages Xllf-XIV and minimum at stages VII to VIII are reciprocal to thechanges in RER (Kerr, 1988).In the rat, abundant SER predominates in the apical region, particularly inassociation with round spermatids (Fawcett, 1975; Rambourg et al., 1979). Duringstage VI, while the spermatid is still deeply recessed in the epithelium, flattenedcisterns of SER surround the head of the elongate spermatid while other SER profilesextend throughout the apical cytoplasm, oriented with the long axis of the cell. Early instage VII, as the spermatid is being moved toward the lumen, bundles of microtubules areinterspersed between the layers of SER. During stage VII, the residual cytoplasm movesbasally, the nucleus becomes more curved, and tubulobulbar processes extend from theconcave surface of the head. Becoming more fenestrated, the cisternae of SER give way totubular forms, except at ES sites. The tubular forms of the ER enter the apical processfrom the apical stalk, extend along the dorsal surface as well as span the ventralcurvature of the spermatid, anastomosing with the SER that cups the bulbous porLions ofthe tubulobulbar processes, as they cross from the tip to the base of the nucleus(Clermont et al., 1980). The flattened cisternae of the ESER are disassembled alongwith the actin filaments, followed by disintegration of the tubular complexes and its SERsystem, leaving only the microtubules as spermiation proceeds.In the ground squirrel, a large volume of SER is transported through the body ofthe Sertoli cell and into the apical stalk where it accumulates in cytoplasm surroundingthe late spermatid. (yogI et al., 1983a). The function of this dense layer of SER is notknown. In this species, colchicine perturbation of microtubules results in the failure ofthe SER to be transported to the apical position, suggesting a role for microtubules in theSER positioning (VogI, et al., 1983b).Much of what has been learned about the functions of SER is derived fromspecialized systems such as steroidogenesis in the SER of endocrine cells, calciumsequestration in sarcoplasmic reticulum of striated muscle and metabolism of toxicIntroduction 3 1substances in the SER in liver. Steroidogenesis, a common function of SER, is consideredlimited in Sertoli cells.Intracellular calcium plays a role in a wide range of cell events and as would beexpected, is unevenly distributed in the cytoplasm. In a 1984 review of calcium incells, Somlyo pointed out, “Unfortunately for biologists, living systems often evolve touse different structures and mechanisms to solve the same problems (photoreceptorsand contractile regulation are but two examples), and grand unifying schemes withfundamental principles applicable to ‘all cells’ are rare. In the regulation of cellcalcium by the ER in all nucleated cells, we may have a rare example.” (Somlyo, 1984).Currently, with such ubiquitous processes as microtubule-based transport andmembrane targeting by coat proteins being described, a unifying scheme such as calciumsequestration by the SER, as described by Somlyo (1984), seems even more plausible.Calcium probes such as equorin used in Chironomus salivary glands (Rose andLowenstein, 1975), indicate that selective calcium sequestration by SER is able tomaintain heterogeneous microenvironments of calcium concentration within single cells.This permits local control of cell functions in which calcium acts as the secondmessenger. Calcium is a known regulator of the cytoskeleton and its chelation has beenshown to mimic FSH induced shape changes in cultured Sertoli cells (see Dedman et al.,1979). In addition, the binding of microtubule associated proteins to tubulin isregulated by phosphorylation of microtubule associated proteins modulated by calciumcalmodulin and cAMP dependent kinases (Bershadsky and Vasileiv, 1988c). Franchiniand Camatini (1 985b) report the presence of calcium in the ESER in guinea pig Sertolicells, using a pyroantimonate precipitation technique. The data shown are not convincingand are inconsistent with the results of Kierszenbaum and coworkers (1971) in mouseSertoli cells. The possibility that local calcium levels could be maintained by the ESERto regulate the cytoskeletal associated events around ESs bears further study.The constant and elaborate, stage dependent, changes in membranous organellesthat occur in Sertoli cells, most of which are microtubule dependent in cells generallyIntroduction 32(See Schroer and Sheetz, 1991a), indicate a crucial role for microtubules in organellepositioning in Sertoli cells. The parallel orientation of many organelles andmicrotubules may reflect this association.CytoskeletonIn keeping with their role of providing morphological support for spermatogeniccells, in addition to their many intracellular functions, Sertoli cells possess anelaborate cytoskeleton of actin filaments, intermediate filaments, and microtubules(Amlani and yogI, 1988; yogI, 1988; 1989; yogI et al., 1992). Except where they arepart of stabilized structures, actin filaments and microtubules are highly dynamicpolymers, undergoing constant assembly/disassembly in equilibrium with theirrespective soluble cytosolic pools. They both participate in structural and motilityfunctions. In contrast, intermediate filaments are thought to be mainly architectural infunction and less dynamic. The stability of each of these cytoskeletal elements as well astheir association with each other, with membrane bounded organelles, and the Sertolicell membrane is modified by the presence of a vast array of cytoskeleton associatedproteins.microfilamentsMicrofilaments (filamentous actin) are assembled from monomers of globularactin (g-actin) polarized with a slow growing and a fast growing end, identified withmyosin decoration as the barbed and pointed end respectively. They occur throughout theSertoli cell, particularly in regions of the plasma membrane, as well as in areas ofspecialized function such as at ES sites, junctional sites, and tubulobulbar complexes(Franke, et al., 1978; yogI et al., 1986; see for review yogI, 1989; yogI et al., 1992).Microfilaments, identified by their 6 nm diameter size and ability to bindphallotoxins, are particularly prominent at ES sites where they form a highly orderedand stable hexagonal array between a cistern of endoplasmic reticulum and the SertoliIntroduction 3 3cell membrane (for review, yogI, 1989). Fine strands between actin filaments suggestthat they may be highly crosslinked to each other and to their adjacent membranes.Actin filaments at ES sites are very stable, being resistant to mechanical perturbationand the effects of cytochalasin, an actin disrupting agent (Russell et al., 1988).intermediate filamentsIn immature Sertoli cells, intermediate filaments are of vimentin, and to a lesserextent cytokeratin, types. Only the vimentin intermediate filaments are present inhealthy Sertoli cells of the mature rat (Franke, et at., 1979; Amlani and VogI, 1988;Hail et al., 1991). Intermediate filaments, identified ultrastructuraily as l2nmfilaments, are particularly prominent at the base of the cell and among infoldings of, andsurrounding, the nucleus (Fawcett, 1975; Franke et at., 1979; Ploen and Ritzen, 1984;BrOkelmann, 1963) giving the impression of a halo (Fawcett, 1975) that excludesother organelles from the nuclear region. From the nucleus they extend to desmosomelike junctions formed with translocatirig germ cells (Amlani and yogI, 1988).Intermediate filaments appear to form a stabilizing scaffold between the nucleus andintercellular junctional sites, possibly adding stability to the epithelium and serving toposition the nucleus at the base of the cell (Franke, et al., 1979). The stabilizingscaffold extends to intermediate filaments that form a network at the base of the cell inclose association with hemidesmosome-like junctions between Sertoli cells and theextracellular matrix. Small bundles of 8-12 intermediate filaments reach from thenucleus into the apical portion of the cell, extending along the dorsal surface ofelongating spermatids through a defect in the ESER to approach the cell membrane(Amlani and yogI, 1988; VogI, 1989; yogI et at., 1991). The distribution ofintermediate filaments changes with spermatogenesis, particularly in those associatedwith ESs. They are most prominent when spermatids are at their deepest location (StageV), diminishing thereafter (Amlani and VogI, 1988). It is proposed that these bundlesmay serve an anchoring function at Sertoli cell crypts.Introduction 34microtubulesIt was not long after the first discovery of microtubules that they were describedin the cells of the seminiferous epithelium (Christensen, 1965; Fawcett, 1975). Theseminiferous epithelium is a showcase of microtubule architecture, with many highlyspecialized microtubule structures in developing germ cells (Wolosewick and DeMay,1982; Cherry and Hsu, 1984; Hermo et al., 1991) and cytoplasmic microtubules inSertoli cells (Christensen, 1965; Fawcett, 1975; Amlani and yogI, 1988; yogI et al.,1983a; VogI, 1988, 1989; Hermo et al., 1991). In contrast to the complexity ofmicrotubule structures found in germ cells, the conspicuous feature of cytoplasmicmicrotubules in Sertoli cells is their abundance. They are most prominent in thesupranuclear portion of the cell in which they are oriented parallel to one another and tothe long axis of the Sertoli cell. As they extend into the apical processes, microtubulessurround ESs that line Sertoli cell crypts and conform to stage dependent morphologicaland positional changes that occur in developing spermatids.In that Sertoli cell microtubules are a prime focus of this study, they will bediscussed later in greater detail following a description of the molecular, kinetic, andfunctional properties of microtubules in cells generally.Ectoplasmic specializationsUnique actin containing structures called ‘ectoplasmic specializations’ (ESs)(Russell, 1977c) form within Sertoli cells in two locations: apically, they occuradjacent to regions of spermatid attachment to Sertoli cells; basally, they are present atthe blood testis barrier (Brokelmann, 1963; Flickinger and Fawcett, 1967; Nicander,1967; Dym and Fawcett, 1970; Russell, 1977c; for review see yogI, 1989; 1991a,b).ESs are found on both sides of Sertoli-Sertoli cell junctions, but only on the Sertoli cellside of Sertoli-spermatid junctions (Romrell and Ross, 1979). Ribosomes are aconstant feature of the outer face of the ESER at basal junctions but rarely occur on theIntroduction 3 5apical ESER. Close association with intermediate filaments and microtubules are only afeature at apical ESs. The location of basal and apical ESs is indicated diagrammaticallyin Fig. 1-6 and their fine structure in Fig. 1-7.OrganizationESs form adjacent to early spermatids becoming aligned with the rostral surfaceof the spermatid as the acrosome and nucleus migrate to the plasma membrane duringstep 8. At this point, orientation of the spermatid occurs. Failure of acrosomeformation following injection with anti-cancer drug procarbazine indicates thatacrosome-ES association is required for early spermatids to become aligned within theepithelium and enter Sertoli cell crypts (Russell, et al., 1983a).As elongation ensues, ESs line the Sertoli cell crypts in areas adjacent to theacrosome. Although it is generally considered that ESs occur only adjacent to spermatidheads, Weber and coworkers, in a morphometric study of stage V rat Sertoli cell,describe ESs that extend as far as the midpiece of step 17 spermatids (Weber, et al.,1983). ESs are more concentrated over the dorsal than the concave ventral surface instep 17 spermatids (Weber, et aL, 1983).For purposes of correlating structure and function, VogI and coworkers,(1991a) divide ESs into four major domains: 1) extracellular, 2) integral membrane(plasma membrane), 3) ectoplasmic, and 4) endoplasmic reticulum domains.The extracellular domain spans the intercellular space and would contain theextracellular regions of adhesion molecules that would be present in the second domain,the integral membrane domain. In an attempt to identify adhesion molecules that mayprovide intercellular adhesion at ES sites, Pfeiffer and coworkers (Pfeiffer et al.,1991) were able to demonstrate a Beta 1 integrin subunit at apical and basal sites whichcould potentially occupy these first two domains. They were unable to detect E cadherin,N Cam or A Cam at these same sites (Pfeiffer et al., 1991). The third domain, theectoplasmic domain, can be subdivided into three zones, encompassing the middle zone ofIntroduction 3 6Figure 1-7: Ectoplasmic specializations: These micrographs show theultrastructural features of apical ectoplasmic specializations in the rat (a,b) andsquirrel (b,c) testis. Microtubules (arrowheads) can be seen in close association withthe endoplasmic reticulum of ectoplasmic specializations or ESER (long fine arrows) ineach case. Actin filaments (asterisks) are situated between the ESER and the Sertoli cellmembrane (thick short arrow). Note the fine linkages that can be seen between theESER and microtubules (short thin arrow). Bars: a-c = 250 tim, d= 100 tim.37.-;;S.,. .Introduction 3 8actin filaments and flanking zones of actin membrane attachment. The zones ofattachment of actin to the plasmalemma is a likely location of vinculin, shown to bepresent at ESs sites (Grove and VogI, 1989; Grove et al., 1989; Pfeiffer and VogI,1991). The actin zone contains an orderly array of actin filaments with an inter-filament spacing approximately 10 nm apart which appear to be crosslinked by finestrands, among which the actin binding proteins fimbrin (Grove and VogI, 1989) andalpha actinin (Russell and Goh, 1988), present at ES sites, may occur. Very littleinformation is available about the fourth domain, the endoplasmic reticulum domain.There is some evidence that the ESER may be continuous with the ER of the rest of thecell (Clermont et al., 1980), but whether its function is that of general ER, or if theESER has specialized functions requires further study. Calcium has been reported in theESER (Franchi and Camatini, 1985b) but due to the insensitivity of calcium localizationtechnique used corroborative data are needed. In view of the increasing number ofenzymes and coating protein markers available for ER and Golgi, this information may beavailable soon.A provisional fifth domain, the cytoplasmic domain, is described. While not yetdetermined to be a part of the ES, microtubules occur in close association with thecytoplasmic face of ESs. At apical ESs, microtubules come in contact with the ESER,reminiscent of other systems in which microtubules are associated with membranebound organelles and involved in microtubule based transport. It remains to be shownwhether the fifth zone is a functional component of ESs.Relationship with cytoskeletonIntermediate filaments extend from the nucleus to ESs, reach along the dorsalsurface of the spermatid, and pass through a gap in the ESER to reach the junctionalmembrane. Bundles of microtubules pass adjacent to ESs and are oriented parallel to thepath of spermatid translocation. The close association of ESs with cytoskeletal elementssuggests a role in positioning of spermatids at apical sites.Introduction 3 9LinkagesFine cross-linking strands have been described between actin filaments and fromthem to the adjacent membranes (Franke et al., 1978; Russell, 1977c). At ESs, actinfilaments are firmly attached to the Sertoli cell membrane and to the cytoplasmic face ofthe ER component (Russell, 1977c) such that during very vigorous isolation of Sertolicell organelles, ribbons of ESs are seen with both the Sertoli cell and ER membranesstill attached (unpublished observations). The microfilaments and ER are tightly boundto the junctional site and remain attached when the cell is perturbed mechanically(Franke et al. 1978; Masri et al., 1987; yogI and Soucy, 1985 VogI et al., 1985a;1986.) or with collagenase. However, they can be dissociated from the adjacentmembrane by treatment with trypsin (Romrell and Ross, 1979; Masri et al., 1987;yogI et al, 1986), resulting in the loss of the intense actin positive staining with NBDphallacidin, typical of ESs (Masri et al., 1987). Filament-filament or filament-membrane proteins alpha actinin (Franke, 1978) and fimbrin (Grove and yogi, 1986b)are present but myosin is not (yogI and Soucy, 1985). The cross-linking of actinfilaments and their adjacent membranes contributes to the structural continuity to thespermatid to ESER linkage.Strands, occasionally seen extending between the ESER and microtubules (Fig. 1-7), were first described by Russell (1977c). The nature of these linkages is not knownnor is it known whether they are discrete structures.ESERThe ESER is thought to arise from the ER of the Sertoli cell (Clermont et al.,1980) which extends into the apical cytoplasm and surrounds developing spermatids.The stage dependent changes in the ER that associates with the ESER have been describedearlier. The ESER is seen as a flattened fenestrated cistern of ES that lies adjacent to theactin component of ESs from their first appearance adjacent to early spermatids. Thefunction of the ESER is not known. Franchi and Camatini (1985b) have reported theIntroduction 40presence of calcium in the ESER, using calcium precipitating agents, but the evidence isweak and requires confirmation. It is not known whether the ESER forms a functionalunit with any of the membranous organelles that make up the secretory, Golgi, orendosomal compartments.Stage specific changesAlthough it is not known what specifies their formation, the current view holdsthat ES complexes arise from precursors in a cytosolic pool with basal ESs formingindependently from apical ESs, rather than being cycled between apical and basal sites asan intact structure as had been proposed earlier by Russell (1984). Basal ESs form atapproximately 16 to 19 days in the rat, coinciding with with the establishment of theblood testis barrier (Gilula et al., 1976) and the beginning of spermatogenesis. Theyare dismantled and reassembled with the shuttling of spermatocytes into the adluminalcompartment, with little change in their overall pattern.At apical sites, ES5 show stage specific changes in their organization (yogI andSoucy, 1985, yogI et at, 1985). The appearance of apical ESs has been reported asearly as the mid pachytene spermatocytes in the mouse (Ross and Dobler, 1975; Ross,1976), although Weber and coworkers (Weber, 1983) indicate that few ESs areassociated with spermatocytes, in the rat. They are present by the time spermatidsbegin differentiation and persist until step 18, just before spermiation (Russell,1977c, 1984). Stage dependent reorganization of the actin filament pattern, of ESs,during spermatogenesis, has been described in squirrel (yogI and Soucy, 1985; Pfeifferand yogI, 1991; yogI et al., 1991a) and rat (yogI et al., 1985b) and corresponds withspecies specific shaping of the spermatid head. Timing of ES disassembly shows speciesvariation, being before (opossum and bandicoot), slightly before or during (guinea pig,chinchilla, stallion, bull, sheep, pig, dog, cat, rabbit, and rat) or after (mouse)disengagement of spermatozoa (see Russell, 1984). In rat, as ESs are gradually beingIntroduction 4 1dismantled, during stage VII, as a consecutive series of tubulobulbar complexes form andresorb, immediately prior to spermiation at stage VIII.FunctionsThe numerous roles that have been suggested for the ES include: stabilization ofthe junctional membrane (Suarez-Quain and Dym 1984), intercellular attachment(Brökelmann, 1963; Nicander, 1967; Ross, 1976), cell communication (Flickingerand Fawcett, 1967), anchoring of the spermatid (Fawcett, 1975), regulation of surfacecontour and positioning of the crypt (Russell, 1977c), enzymatic digestion (Tindall etal, 1985), regulation of spermatogenesis (Saez et al., 1985), or secretion of steroids(Pudney, 1986), among which two functions seem most likely (yogI, 1989):participation in establishment of adhesion domains at sites of the intercellularattachment and positioning of germ cells within the seminiferous epithelium (forreviews see yogI, 1989; VogI, et al., 1991a,b). Rather than being mutually exclusive,these roles are complementary.The adhesion hypothesis is supported by a number of findings. The intimateassociation of ESs with sites of spermatid-Sertoli cell and Sertoli-Sertoli cellattachment suggests that the ESs play a role in their regulation. The naturally occuringbreakdown in adhesion accompanies the resorption of ESs during the remodelling of theblood testis barrier and at spermiation (yogi et al., 1983a). The pharmacologicaldisruption of actin filaments with cytochalasin D (Russell, et al., 1988) leads to loss ofadhesion at basal sites and release of spermatids. Evidence that ES actin filaments are ofuniform polarity, demonstrated by Si decoration in ground squirrel (VogI et aI., 1986)and rat (Masri et al., 1987), that myosin is not present in mammalian ESs, and thatfilaments do not contract defines a non-contractile role for ESs (VogI and Soucy, 1985;Vogi et al., 1983 a). There is biochemical and immunological evidence that vinculin,typically present at adhesion junctions, is present at ES sites (Grove and yogI, 1989;Grove et al., 1990; Pfeiffer and yogI, 1991). Finally, when spermatids are42mechanically removed from the seminiferous epithelium, the linkage betweenspermatids, the Sertoli cell membrane, and ESs remains intact (Romrell and Ross,1979; yogI et al., 1985; Masari et al., 1987; Grove and VogI, 1989), indicating anadhesive continuity between ES and spermatid-Sertoli cell junctions. Taken together,these data suggest that ESs are part of actin-associated intercellular adhesion junctions(yogI et al., 1986; Grove and yogI, 1989; VogI, 1989) which remain presentthroughout spermiogenesis and serve to anchor spermatids within Sertoli cell crypts.Based on their presence at sites of spermatid-Sertoli cell attachment and closeassociation with elements of the Sertoli cell cytoskeleton, they have also been suggestedto participate in orienting and positioning spermatids within the epithelium.RegulationVery little is known about the regulation of ESs. There is some evidence thatthere may be an indirect calcium effect on ESs by a calcium-calmodulin pathway.Treatment of seminiferous tubules with trifluoroperazine, a calmodulin inhibitor,results in disorganization of actin filaments and displacement of ER (Franchi andCamatini, 1985a). Using a pyroantimonate method to localize calcium in guinea pigtestis, these authors report the presence of calcium at ES sites (Franchi and Camatini,1985b). Intratesticular injections of high doses of dibutyryl cyclic AMP results inretention of spermatozoa past their normal release at stage VIII through stages VIII, IXand X, accompanied by retained fluid in the subacrosomal space and surrounding themiddle piece of the unreleased spermatozoa, and a failure of the dismantling of ESs andformation of tubulobulbar process (Gravis, 1980). The author concludes that the effectof dibutyryl cyclic AMP is to delay the dismantling of ESs and the consequentialformation of tubulobulbar processes, producing persistent adhesion of the spermatidsand retention of cytoplasmic fluid respectively. The mechanism behind these events isnot known.Introduction 43Changina physiological events in Sertoli cells during spermatogenesisMorphological and positional changes that occur in spermatogenic cells have beendescribed above. These events occur in a changing milieu established primarily by theSertoli cell. The stage dependent changes in Sertoli cell organelles have been describedand indicate that the physiological activities of Sertoli cells and the environment theyestablish, in the adluminal compartment and tubule lumen, change withspermatogenesis.Using a ‘transillumination’ technique, similar to that introduced by Perey andcoworkers (1961), Parvenin and Ruokonen (1982) have developed an approach tostudying the physiological events of the spermatogenic cycle, devised to relate theexternal morphology of isolated seminiferous tubules with the conventional stagingaccording to Leblond and Clermont (1952). Lengths of seminiferous tubules aredissected free from the interstitial tissue, and staged along their length, after whichsegments of the tubule are used to identity stage specific activities of the seminiferousepithelium. This allows for stage specific sampling of the seminiferous tubule forstudies in which material from selected stages is required. The markers for determiningstages of the seminiferous epithelium in dissected individual tubules are described inChapter 3 Methods.Stimulated by FSH and LH, Sertoli cells secrete a wide range of testis specificproteins, many of which are stage dependent (for review, Griswold, 1988).Kangasneimi and coworkers (1990), using transillumination techniques, (1990) haveshown that FSH stimulated cAMP production is elevated between stages II and VI andsuggested that this may correlate with the transport of spermatids deep within theepithelium. Mali et al (1985) found the LH stimulation of testosterone is highest at theonset of meiosis, a time when the rate of RNA transcription is greatest (stages VII-X1)and spermiation has just occurred. Levels of plasminogen activator, a protease thoughtto be involved with tissue reconstruction, are highest just before spermiation (stagesVII and VIII), coinciding with reorganization of the blood testis barrier.Introduction 44SERTOLI-GERM CELL RELATIONSProcessesOf particular importance to the events of spermatogenesis are the elaborate cellprocesses and complex cytoskeleton of the Sertoli cell. Because of the translocation ofgerm cells through the epithelium, intimate Sertoli-Sertoli and Sertoli cell-germ cellrelationships exist (Weber, et al., 1983). Specialized structures are formed in bothcell types that serve to modify the Sertoli cell-germ cell relationship (Russell, 1980).The Sertoli cells interface with each other and with developing germ cells by extending anumber of cytoplasmic processes. These structures include the lateral, apical, andpenetrating processes of Sertoli cells and tubulobulbar processes of spermatids(Russell, 1980).lateral processesThe Sertoli cell extends lateral processes which interlock with each other and enshroudthe developing germ cells. The seminiferous epithelium is divided into basal and apicalcompartments by tight junctions, circumscribing the base of the Sertoli cells (Dym andFawcett, 1980). This provides for separate apical and basal environments as well as animmunological barrier for germ cells having completed meiosis. Sertoli cells areinvolved in the ushering of spermatocytes from the basal to the apical compartment byboth surrounding, and therefore isolating, translocating germ cells. As the germ cell ismoved upward (a seemingly passive process) new lateral processes form behind thetranslocating germ cell then only after the new barrier is formed is the barrier abovethe germ cell dismantled, ensuring that there is not any communication between the twocompartments (Russell, 1977a).45Apical processesApical processes extend from a microtubule rich stalk and support developingspermatids (Russell et al., 1983b; Weber et al., 1983, VogI et al, 1983a). Afterspermatids are returned from their basal migration in Sertoli cell crypts, they becomeextended into the lumen on the apical stalk, held by apical processes. This isparticularly evident in the ‘simple’ seminiferous epithelium of the squirrel wherespermatids are extended into the tubule lumen on apical processes, extending from thinapical stalks (yogI et al., 1983a,b). The elaborately convoluted surface of the apicalportion of Sertoli cells contrasts sharply with other epithelial cells which typicallypossess microvilli and apically placed tight junctions.Penetrating processesPenetrating processes, from the Sertoli cell, invaginate into germ cell cytoplasmduring its segregation from the rest of the spermatid (Nicander, 1967; Morales andClermont, 1982; Vogl et al, 1985b). The presence of coated pits, in both the Sertolicell penetrating processes and adjacent germ cells, suggests there is an exchangebetween the two cells (yogI et. al., 1985b). Organelle content is variable, frequentlyincluding coated vesicles and pits. Penetrating processes have been suggested to aid inrelocation and possible anchoring of residual cytoplasm, to be retained at spermiation,and to participate in intercellular exchange.yogI et al., (1983a, 1985a) have shown that, in squirrel testis, penetratingprocesses initially do not contain microtubules, but as the residual cytoplasm isextended into the tubule lumen, microtubules become abundant in these processes,suggesting that microtubules may participate in the maintenance of the shape of theseprocesses and participate in the movement or removal of residual cytoplasm from thelumen to within the epithelium, following spermatid elongation.Introduction 46Tubulobulbar processesTubulobulbar complexes form apically as cytoplasmic evaginations from theventral surface of elongate spermatids, or basally from evaginations of Sertoli cells,coupled to invaginations of the adjacent Sertoli cell cytoplasm (Russell and Clermont,1976; Russell, 1979a,b, 1984). They include a long slender penetrating process, thereciprocal Sertoli cell membrane, and its associated underlying structures.Tubulobulbar processes terminate in a bulbous swelling which may or may not have afurther short tubular process. Surrounding the tubular portion, is a network of actinfilaments (yogI et al., 1985b; VogI, 1989); capping the bulbar portion, is a cistern ofendoplasmic reticulum (Russell and Clermont, 1976; Russell, 1979a). Theendoplasmic reticulum is continuous with general ER (Russell, 1984). Thetubulobulbar complex terminates in a coated pit. Basally, they occur in discontinuitiesof ESs and apically, occur at sites where ESs are being dismantled (Russell, 1979a;Clermont et al., 1980).Apically, a series of these process form and disintegrate just before spermiation(mid stage VII to stage VIII) (Russell, 1979a). Their extent is species specific, beingprominent in rat spermatids (Russell and Clermont, 1976; Russell, 1979a,b), less soin opossum, vole, guinea-pig, mouse, hamster, rabbit, dog, monkey and human (Russelland Malone, 1981), but seldom seen in the squirrel (yogI 1989). Their function isunknown, but a correlation of their appearance apically at stage VII (Russell, 1979a)with a loss of cellular volume from spermatids suggests a function of fluid regulation(Russell, 1979c). As apical tubulobulbar processes persist until just prior to the lossof Sertoli-germ cell contact (stage VIII), an anchoring function is also possible (Russelland Clermont, 1976). Basally, they appear between stages II to V (Russell, 1979b andare resorbed during the apical movement of spermatocytes through the blood-testisbarrier late in stage VII (Russell, 1977a, 1979b). It is interesting that they occur attimes when ESs are being modified and may be involved in the elimination of junctionalmembrane domains (see yogI, 1989 for discussion).Introduction 47JunctionsJunctional sites associated with the specialized intercellular relations of theseminiferous epithelium change during spermatogenesis. There are a variety ofjunctions which occur at the Sertoli-Sertoli cell and Sertoli-germ cell interface(Brokelmann, 1963). A tight junctional complex forms between lateral processes atthe base of Sertoli cells, forming the blood testis barrier and is closely associated with,but distinct from, the adhesive junctions that occur at basal ES sites. At early germcell-Sertoli cell contacts, desmosome-like junctions and gap junctions are encountered(Nagano, 1966; Russell, 1977a,c, Russell and Peterson, 1985). These junctionsdecrease in number until the second meiotic division of the germ cell is complete.Finally unique adhesive junctions (apical ESs) form between spermatids and Sertolicells.Blood testis barrierUnlike many other epithelia, the tight junction of the seminiferous epitheliumforms at the base of adjacent epithelial cells. Staining for filamentous actin at basal ESsites provides a close approximation of the pattern of tight junctions of the blood-testisbarrier with which basal ESs are closely associated. The actin staining of basal ESsindicates an interconnecting honeycomb shaped junctional network, when epithelialsheets are viewed enface (yogI et al, 1985a; yogI, 1989), marking the ring like seal ofthe blood-testis barrier. The tight junction separates the basal and adluminalcompartments (Dym and Fawcett, 1970) establishing the unusually tight blood testisbarrier, evidenced by its impermeability to lanthanum (Dym and Fawcett, 1970) andthe large number of parallel sealing strands seen on freeze fracture replicas (Gilula, 1976). The formation of the junction, indicated by the appearance of actin filamentsbetween days 16 and 22 in the rat (Russell et al., 1989) and the exclusion oflanthanum, coincides with the beginning of spermatogenesis and is required for theIntroduction 48directional secretion of tubule fluid. An adhesion junction occurs in close associationswith the tight junction, resulting in a close association of ESs with this junctionalcomplex.The blood testis barrier is thought to provide two functions: 1) the creation of aseparate adluminal environment suitable for the nourishment of developing germ cells,and 2) immunological protection for post pachytene spermatocytes (Dym and Fawcett,1970)Sertoli cell-Spermatid junctionSertoli cell-spermatid junctions are unique adhesion junctions, formed betweenspermatids and Sertoli cells, and possessing subsurface specializations, ESs, on theSertoli cell side of the junction. Details of this junction have been included in previousdiscussions but a number of points should be summarized. These junctions occur inareas of contact between spermatids and Sertoli cells. That they are adhesion junctionsis shown by the retention of the Sertoli cell membrane and ESs, at these sites, duringmechanical removal of spermatids from the epithelium. They do not prevent the passageof such molecules as lanthanum and are therefore not tight junctions. It is thought thattheir adhesive function serves to anchor spermatids in Sertoli cell crypts in which theyare moved to the base of Sertoli cells, before being translocated to the tubule lumen. Itis through these junctions and their subsurface specializations that microtubules mayexert their influence on positioning and translocation of spermatids. Loss ofadhesiveness at these junctions, which occurs at spermiation, coincides with thedismantling of ESs.Introduction 49M ICROT U B U LESIntroductory statementAlthough the activities of microtubule based transport were observed as long agoas 1879, when microscopist, Joseph Leidy, described movement of small granules oninvisible microtubule tracks in the freshwater foraminer, Gromia, it is not untilrecently that the fundamental and ubiquitous involvement of microtubules in cellactivities has begun to be appreciated. The microtubule story has been closely tied toadvances in microscopy. Leidy was unable to see the tracks that supported particlemovement because microtubules, just 25nm in diameter, are well below the limit ofresolution of 200 nm (lnoué 1986) in the light microscope.Taking advantage of the improved resolution of electron microscopy, tubularfibrils were first identified by Grassier in 1965, Faure, et al., in 1958, and Roth in1958, in axonemes of cilia and flagella, and Porter in 1954 in plants (see Gibbons1988). These structures, were christened “microtubules” by Slautterback (1963),and first reviewed extensively by Porter in 1966. While electron microscopy soonrevealed the ubiquitous occurrence of microtubules, it was the work of Allen andcoworkers (Allen et al., 1981c,1982, Hayden et al. 1983) who, using computerassisted, video enhanced differential interference contrast microscopy (Allen et al.1981a; lnoue, 1981), and extruded axoplasm of the squid giant axon (Brady et al.,1982, 1985; Vale et al., 1985c), brought to life the dynamic nature of microtubules.Immunocytochemical staining, viewed with fluorescence microscopy, proved to be aninvaluable tool for studying microtubule associated events (Horio and Hotani, 1986;Sammak and Borisy, 1988). These techniques, coupled with the purification andcharacterization of the major microtubule protein, tubulin (see Mandelkow andMandelkow, 1989b), and its copurifying proteins, microtubule associated proteins orMAPs (see Olmsted, 1986), made possible in vivo (Allen et al., 1982; Gilbert andSloboda, 1984; Hayden et al., 1983) and in vitro (Vale et al., 1985b,d; Paschal et al.,Introduction 501987; see Vale and Toyoshima, 1989) assays of microtubule-based transport. Thesestudies have stimulated the isolation and characterization of a number of microtubuleassociated ATPases, mainly falling into two families; kinesin (Vale et al.,1985a) andcytoplasmic dynein (Pratt, 1980; Paschal and Vallee, 1987), candidates formicrotubule-based transport motors. More recently, other potential candidates formicrotubule-organelle binding proteins (Rickard and Kreis, 1991; Van der Sluijs etal., 1990; Scheel and Kreis, 1991a,b) have been described. The extensive role ofmicrotubules in organelle transport and positioning is rapidly being elucidated (seeSchroer and Sheetz, 1991 a).Microtubule structure and kineticsMicrotubules can assemble from purified tubulin, but native microtubules aremade of ‘microtubule protein’: that is, tubulin and its associated copurifying proteins,referred to as “microtubule associated proteins or MAPs” (Borisy et al., 1974;Williams and Detrich, 1979; Timasheff and Grisham, 1980; Vallee; 1982; McKeithanand Rosenbaum, 1989; Mandelkow and Mandelkow, 1989a,b).Tubulin exists as a dimer comprised of alpha and beta tubulin monomers, totalmolecular weight 100 kD, sharing up to 60% homology. The tubulin dimer has oneexchangeable and one nonexchangeable GTP binding site. Above a critical concentration,tubulin dimers assemble into microtubules, catalyzed by GTP binding at theexchangeable site (Timasheff and Grisham, 1980; Shelanski et al., 1973 for reviewssee McKeitham and Rosenbaum, 1989; Mandelkow and Mandelkow, 1989a). Tubulindimers assemble, head to tail, into protofilaments which in turn form lateralassociations at a 100 offset, resulting in a tubule format with a spiral orientation ofneighbouring dimers (see Mandlekow and Mandlekow, 1989 a,b). Native microtubulesapproximately 25 nm in diameter are generally comprised of 13 protofilaments, buthigher or lower numbers occur. The heterogeneity of dimers confers a structural andkinetic polarity on microtubules, giving them a slow growing minus-end and a fastIntroduction 5 1growing plus-end (Allen and Borisy, 1974). Tubulin dimer exchange occursexclusively at the ends of the polymer (Margolis, 1982; Rothwell et al., 1985). Invivo, cytoplasmic microtubules are in dynamic equilibrium with a cytosolic pool, whichin turn regulates tubulin mRNA levels (Bershadsky, 1988).Microtubule protein can be isolated by exposure of crude cell extracts to a seriesof temperature-dependent microtubule assembly/disassembly cycles (Borisy et al.,1974; Williams and Lee, 1982; Lee, 1982). Typically, microtubules are assembled at37°C, pelleted by centrifugation and depolymerized at 0°C. Tubulin is then separatedfrom the copurifying proteins (MAPS) by ion exchange chromatography (see Timasheffand Grisham 1980; Vallee, 1986b). Alternatively, capitalizing on the stabilization ofmicrotubules at all temperatures by taxol, the microtubule protein is polymerized inthe presence of taxol (Vallee, 1982, 1 986a) and MAPs removed by high salt buffer,after which soluble MAPs and MAP free microtubules are separated by centrifugation.Taxol promotes tubulin polymerization by lowering the critical concentration oftubulin required for assembly and strengthening inter-tubulin bonding (Wilson et at1985), producing short stable microtubules (Schiff et al., 1979). Taxol binding, at asite distinct from the exchangeable GTP catalytic site (Horowitz, 1986), can proceed inthe absence of GTP (Shelanski, 1970) and is able to stabilize microtubules againstdepolymerization conditions (Schiff et al., 1979; Horowitz et al., 1981,86; Wilson etal, 1985). These conditions include cold temperatures, Ca2’ buffer dilution, anddepolymerizing agents such as nocodozole, vinbiastine, and coichicine (Timasheff andGrisham, 1980 Shelanski, et al., 1979, Na and Timasheff, 1982, Dedmann et al.1971).Microtubule Organizing Centers (MTOC)In many cells, microtubules radiate from a centrally located centrosome thatconsists of electron dense pericentrosomal material, and variably, a pair of centrioles.Centrosomes are the preferred site for microtubule assembly in vivo (Soltys andIntroduction 52Borisy, 1985) and nucleate microtubules in vitro (Mitchison and Kirschner, 1984;Kuriyama, 1984, Bergen and Borisy, 1980), determining their length, number,direction of growth, polarity, and protofilament number (Kuriyama, 1984; Bornens etal., 1987; Brinkley et al., 1981; Mitchison and Kirschner, 1984; Soltys and Borisy,1985; Evans et at., 1985). These centrally located structures were initially consideredthe organizers of the cytoskeleton (Tilney ,1971), but a more dynamic scheme is nowrecognized in which MTOCs are able to assume different forms and cellular positions(Brinkley, 1985), the obvious example being cell division.With the polarization of cells during mouse blastocyst formation, centrioleschange position and microtubules ultimately become oriented with the long axis of thecell (Houliston et al., 1987; Fleming and Johnson, 1988). Similarly, theestablishment of polarity in MDCK cells as they reach confluence, is accompanied byseparation of centrioles, dispersion of the pericentrosomal material, realignment ofmicrotubules, and increased microtubule stability (Bacallao et al., 1989; BrO et at.,1990), as evidenced by reduced tubulin incorporation and increased microtubule half-life (Pepperkok et al.,, 1990). These changes do not occur in fibroblasts, grown toconfluence (Pepperkok et al., 1990; Bacallao et al., 1989).Structures known to act as MTOCs include basal bodies, kinectochores, andcentrosomes, of which only kinetochores nucleate microtubules at their plus-end(Brinkley et at., 1985). Not all microtubules are nucleated at centrosomes oridentifiable MTOCs (Fleming and Johnson, 1 988). Particularly graphic examples ofthis are the reorganization of microtubule polarity and direction of pigment transportthat occurs in amputated segments of melanophore arms (McNiven et al, 1984;McNiven, and Porter, 1986) and chicken sensory neurites (Baas et at., 1987) and thereorganization of microtubule polarity that accompanies dendrite differentiation (Baaset al., 1988, 1989), indicating a potential for local regulation of microtubule polarity.Recently, it has been shown that gamma tubulin, a component of pericentrosomalmaterial, nucleates microtubule assembly in vitro. Moreover, gamma tubulinIntroduction 53antibodies block microtubule assembly in vivo (Zheng et al., 1991; Stearns et al.,1991; Joshi et al., 1991), and may be a candidate for nucleation at other sites.Microtubule structuresWe do not have to look any farther than the testis to find a variety of microtubulestructures such as mitotic and meiotic spindles, axonemes, manchettes, and cytoplasmicmicrotubules, all functionally distinct. The question rises: how are these structural andfunctional microtubule differences determined? There are at least three levels at whichstructural and functional diversity of microtubules may be specified; genetic expressionof tubulin isoforms, post translational modification, or regulation by associatedproteins.Tubulin isoformsWhile there are multiple gene families encoding both alpha and beta tubulin,more than one gene will encode a single isoform or a single gene may encode multipleisoforms (Vilasante et al., 1986; McKeith and Rosenbaum, 1989). Furthermore,multiple isoforms can coexist in a single microtubule, or single isoforms occur in abroad range of microtubule structures in the same cell (Greer and Rosenbaum, 1989).Isoforms can be substituted for one another (Hoyle and Raff, 1990) and in vitro, diverseisoforms can assemble indiscriminately. In a recent review summarizing the availableevidence for genetic determination of tubulin function, Bulinski and Gundersen (1991)conclude that, while the possibility of isoform specific microtubule functions cannot beexcluded, the existence of multigene families of tubulin is not likely to be a significantdeterminant of microtubule structure or function (Greer and Rosenbaum, 1989).Post-translational modification of tubulinPost-translational tubulin modification, is a rapid and reversible means ofproducing covalently diverse tubulins and has been examined in a variety of tubulinIntroduction 54containing structures (for reviews see McKietham and Rosenbaum, 1984; Bulinski andGundersen, 1991: Greer and Rosenbaum, 1989). Known modifications include:phosphorylation, acetylation, and detyrosination. In comparative studies of tubulinmodification and microtubule stability, using structural models of retinoblastoma cells(Schuize et al., 1987) and testis (Hermo et al., 1991) and morphogenetic models offibroblast orientation, neurite outgrowth, and myogenesis (Bulinski and Gundersen,1991), support has been found for a correlation between microtubule stability and thepresence of detyrosinated and acetylated forms of tubulin. The simple assumption is thatdetyrosination and acetylation confer stability on tubulin. Accumulating evidencesuggests that the correlation may be circumstantial; that detyrosination and acetylationare modifications that may occur on “older” microtubules, explaining the correlationwith stability. Furthermore, the correlation between the stability/instability ofmicrotubules and known categories of post-translation modification is not complete(Schultz et al., 1987; McBeath and Fujiwara, 1989). Idriss and coworkers (1991)have shown that kinetic indicators of dynamic instability are not changed by microtubuledetyrosination. The current consensus is that these modifications are coincidental: notbeing a consequence, nor a cause of stability Schulze et al., 1987; Greer and Rosenbaum,1989; Bulinski and Gundersen, 1991; Idriss et al., 1991). Functional diversity, onlycircumstantially related to post-translational modifications, may be primarilymodulated by the presence of MAPs and other regulatory proteins.Microtubule associated proteins (MAPs)A number of proteins copurify with tubulin, for which Sloboda coined the term“microtubule associated proteins” or MAPs, classified generally according to their sizedetermined on SDSPAGE gels (Vallee and Bloom, 1984; Paschal et al, 1987; for reviewssee Olmsted, 1986; Sloboda and Gilbert, 1989; Wiche, 1989; Vallee, 1990b). ManyMAPs are cAMP or calcium/calmodulin protein kinase stimulated phosphoproteins that,in turn, stimulate microtubule assembly (Wiche, 1989). MAPs have been dividedIntroduction 55into: two high molecular weight classes, MAP 1 and MAP 2; a second heterogenous group(..200kDa) that figures predominantly in non neuronal sources, including MAP 3 andMAP4; and a group of smaller MAPS, that includes Kinesin (11O-l3OkDa) and tau(52-55 kDa) (Wiche, 1989).Further characterization has now identified at least five subgroups of the highmolecular weight (HMW) MAPs: MAPS 1A,1B,1C, and MAPS 2A and 2B. In neuronaltissue MAP 2, like tau, is heat stable distinguishing it from MAP 1. MAP properties, thebasis of their terminology, vary considerably depending on the source and conditions ofisolation, leading to some confusion. For example, a number of different MAPs have beendesignated MAP 1C (Wiche, 1989), including a low affinity mechanoenzyme whose yieldis potentiated by isolation on microtubules that have been assembled with taxol in theabsence of any triphosphate nucleotides (Paschal et al., 1987). This MAP has beenidentified as cytoplasmic dynein (Paschal et al., 1987), an ATPase and potentialretrograde motor for organelle transport (Paschal and Vallee, 1987). Although notstrictly ‘MAPs’, additional proteins are found to copurify with tubulin under specialconditions, including the mechanoenzyme, kinesin. In a broad sense, ‘microtubulebinding proteins’ would include all proteins that can bind to microtubules including:tubulin copurifying MAPs; transiently associating mechanoenzymes; and more stablybinding microtubule binding proteins.MAPs share a number of functions in addition to promotion of microtubuleassembly. They mediate interactions between microtubules and with organelles,including other cytoskeletal elements (Gyoeva and Gelfand, 1991). These associationsrange from the dynamic interactions of mechanoenzymes to stable interactions producedby crosslinking or spacing mechanisms (Hems, et al., 1991) and can be translated intoa wide variety of cellular functions.Introduction 56Microtubute polarityThe asymmetry of tubulin dimers confers a physiological (Allen and Borisy,1974; Summer and Kirschner, 1979; Bergen and Borisy, 1980; Mandelkow andMandelkow, 1989a) and structural (Allen and Borisy, 1974; Mitchison and Kirschner,1984; Horio and Hotani, 1986) polarity on microtubules, resulting in differentassembly rates at the fast growing (plus) and slow growing (minus) ends (Borisy,1978). Two morphological techniques have been devised to determine microtubulepolarity in vivo : microtubule decoration with purified dynein (Haimo, 1982), or hookformations with purified tubulin (Heidemarin and McIntosh, 1980; Heidemann andEuteneuer, 1982; McIntosh and Euteneuer, 1984; Heidemann, 1991). Incubation ofmicrotubules in tubulin containing hook promoting buffer induces curved protofilamentsheets to assemble along microtubules with a consistent handedness that gives theappearance of hooks, when viewed in cross section, indicating the pole from which thedecorated microtubules are being viewed (Heidemann and Mcintosh, 1980). This methodcorrectly identifies over 90% of microtubules of known polarity, the remaining lessthan 10% being of the ‘wrong’ polarity. The occasional decoration of a singlemicrotubule with hooks of both clockwise and counterclockwise handedness supports theassumption that those of opposite polarity are “wrong” in 5-10% of cases. When viewedfrom the plus-end, hooks form in a clockwise direction. Using the tubulin hooktechnique, cytoplasmic microtubules have been found to be unipolar with their plus-ends directed toward the periphery of the cell in axons (Burton and Paige, 1981;Heidemann et al, 1981; Baas et al., 1988), melanophores (McNiven et at., 1984;Euteneuer and McIntosh, 1981; Heidemann et al., 1981) chromatophores, (McNiven andPorter, 1986), neurites (Baas et al., 1987), photoreceptors (Troutt and Burnside,1988a), Reticulomyxa (Euteneuer et al., 1989) and ciliary axonemes in a number ofcells (Heidemann and McIntosh, 1980; Euteneuer and McIntosh, 1981). The singleexample of nonuniform polarity occurs when unipolar microtubutes in developingneuronal cells become bipolar, coincidental with the onset of dendrite specificIntroduction 57morphology (Baas et al., 1988, 1989; Black and Baas, 1989; Burton, 1988).Recently, examples have been found in which the minus-ends of microtubules aredirected toward the periphery of the cell, including teleost retinal pigment epithelialcells (Troutt and Burnside, 1988b) insect ovarioles (Stebbings and Hunt, 1983) andDrosophila wing epidermal cells (Mogensen et al., 1989).Cytoplasmic microtubules are highly dynamic and subject to changes in polarityunder physiological conditions such as establishment of cell polarity (Houliston et al.,1987; Bacallao et at., 1989; Bré et at., 1990; Pepperkok et al., 1990) and dendriteformation (Baas et at., 1988, 1989), and experimental conditions, such as amputationof cell segments (McNiven et at, 1984; McNiven and Porter, 1986; Baas et al., 1987).Microtubule polarity has important functional implications in cells, mostnotably in those which depend on important asymmetrical microtubule-based transportactivities, such as the orthograde transport of neurosecretory products and retrogradetransport of degradation products that occurs in axons. The more subtle activities,involved in organelle transport and positioning, also depend on vectorial microtubuletransport (see Schroer and Sheetz, 1991a).Microtubule based intracellular transportAlthough intracellular motility was first described as long ago as 1879, almost acentury passed before scientists began to elucidate the mechanisms of intracellularmotility. White morphologists, reported microtubules in an increasing number ofstructures and cell types, physiologists studied the transport of radiolabelledsubstances, as well as vesicles, along axons (Smith, 1972). However, not untildifferential interference contrast microscopy was combined with video and computerimage enhancement (Allen, at. al., 1981a,1982; lnoué, 1986, Schnapp, 1986), was itpossible to visualize the participating organelles of intracellular motility, many assmall as 25 nm, in real time, bringing the two fields together (Allen et at., 1981c,1985; Brady et al., 1982; Lasek and Brady, 1985; Schnapp et al., 1985).Introduction 58Much interest initially centered around two model systems of intracellularmotility, pigment granule dispersal and aggregation, and axonal transport (for reviewssee Vale, 1987; Sloboda and Gilbert, 1989). The aggregation and dispersion of pigmentgranules were proposed to occur on a dynamic scaffold, the microtrabecular lattice(Ellisman and Porter, 1980). Current evidence, such as the sensitivity ofchromatophore distribution to vanadate (Ogawa, 1987), supports a mechanoenzymedriven microtubule-based transport mechanism.Axonal transport, has been studied extensively since the first radiolabellingevidence of its existence. By ligating individual crab axons, Smith (1972) showed thatorganelles accumulate at both sides of the ligature, indicating that axonal transport isbidirectional. The organelles undergoing anterograde and retrograde transport werefound to differ. With the development of Allen’s video enhanced contrast differentialinterference contrast (AVECDIC) microscopy, Allen and coworkers (1981a,c; 1982)began following the events of axonal transport in the squid giant axon (Allen et al.,1981c, 1982). Brady and coworkers (1982) found that axoplasm extruded from thesquid giant axon retains its transport activity, arguing against any involvement of theaxonal membrane in axonal transport. In diluted extruded axoplasm, vesicles movebidirectionally along filaments (Brady, et. al., 1982; Allen et. al., 1985; Vale et al,1985c); stirring of the axoplasm causes disordered transport. Vesicles were observedto transport bidirectionally on single transport filaments (Allen et al., 1983; Schnappet al., 1985; Koonce and Schliwa, 1985). The diameter of the transport filaments wasdistorted because of the light diffracting effect of differential interference contrastmicroscopy which causes structures smaller that 200nm to appear uniformly as 200nm. For this reason, the size of the transport filament, could not be measuredaccurately from the AVECDIC images (lnoué,1986). Immunofluoresence and videoenhanced microscopy followed by slam freezing and rotary shadow electron microscopyon the same area of axoplasm, verified that the transport filaments are singlemicrotubules (Schnapp et. al., 1985; Koonce and Schliwa, 1985). Allen and coworkersintroduction 59(1985) used taxol to stabilize axonal microtubules, showing that treadmilling is not themechanism of microtubule-based transport in axons.To study microtubule based transport, Vale and coworkers (Vale et. al., 1985a-d) devised an elegantly simple in vitro assay in which isolated axoplasmic vesicles, inthe presence of a high speed supernatant (S2) from axoplasm were added to MAP-freemicrotubules that had been adhered to glass slides. They found that: 1) in the presence ofS2, axoplasmic vesicles transported along microtubules, 2) polystyrene beads, in thepresence of S2 (or preincubated in S2), transport along microtubules; and 3) glassslides that had been coated with S2, (even if the slide had been rinsed with a buffer thatdoes not support transport), supported the movement of microtubules. In the presenceof S2, vesicle transport was shown to be sensitive to heat, trypsin, AMPPNP (nonhydrolyzable analogue to ATP), 1OOpM vanadate, but not 20j.tM vanadate (phosphateanalogue), or NEM (sulphydryl alkylating agent). Unlike their behavior in vivo, allaxoplasmic vesicles were found to transport at the same rate (1 .63 (Vale etal., 1985b,d), independent of vesicle size. Furthermore, in all these studies thetransported vesicle (or surface) moved toward the plus-end of the microtubule. Thesefindings indicate that: a ‘motor’ is present in the S2 high speed supernatant, the motor isable to bind microtubules, and the motor is able to bind vesicles (or form or anonspecific binding to beads or glass) at its non microtubule end. Although the nonmicrotubule end of the ‘motor’ binds nonspecifically to the glass or polystyrene beads,its binding to vesicles is highly specific.In similar experiments using a crude axoplasm supernatant (Si), they foundtransport to be bidirectional and vesicles passed one another on the same microtubulewithout interference (Vale, et. al., 1985d). The Si supernatant supported both minusend and plus- end directed transport and was inhibited by NEM and by a lowerconcentration of vanadate (2OjiM) than was S2 supported transport (Vale et al, 1985d).This indicated the presence of a minus-end directed motor in the crude Si supernatant,60that had not been present in the S2 preparation. Taking advantage of the finding thatvesicle transport in axons can be immobilized by the addition of the nonhydrolyzableanalogue of ATP, AMPPNP (Lasek and Brady, 1985; Brady et al., 1985), Vale andcoworkers isolated a plus-end directed motor, kinesin, from squid axoplasm (Vale et al.,1985a,b). Subsequently, the retrograde transport factor has been found to be acytoplasmic form of the mechanoenzyme dynein that powers intermicrotubule transportin axonemes (Pratt, 1980; Paschal et a!., 1987; Paschal and Vallee, 1987), now called‘cytoplasmic dynein’.Gilbert and Sloboda (1984) injected rhodamine labelled vesicles into freshlyextruded axoplasm, confirming that, in vivo, vesicles are transported in both directions,but do not reverse during their journey. Vesicles are able to move on purifiedmicrotubules, indicating that the motor resides on the vesicle. Moreover, bypredigestion of vesicle surface proteins with triton, they showed that vesicle surfaceproteins are required for transport.In vivo, organelles transport along axonal microtubules at a variety of ratesvarying over a 100 fold range (Pratt, 1989a) depending on load, rate limiting step, andphosphorylation state of the enzyme, and the presence of other governors (Pratt,1989a). Vale, (Vale at. al., 1985ac) showed that transport rates varied in intactaxoplasm but in extruded, diluted axoplasm and in vitro assays, vesicles travelled at thesame rate of 2.2p.m/second (Vale et al., 1985c), irrespective of size (Vale , 1987;Okabe and Hirokawa, 1989). It may be that mechanoenzymes have an inherent rate oftransport which, in vivo, is differentially impeded by the viscous drag imposed by thecytoplasm.In axons, vesicles transport along microtubules in both directions, transferringfrom one microtubule to another, passing one another without interference, and seldomchanging direction. As they transfer from one microtubule to another they temporarilytravel on two filaments at once. These observations have led to the conclusion thatvesicles may have multiple motor binding sites (Vale et al, 1985c).Introduction 6 1Axoplasm preparations remain viable for hours, even after two-fold dilutionwith buffer, (Lasek and Brady, 1985; Allen et al, 1985; Cohn et al., 1987) but, withtime, vesicle transport slows. Addition of MgATP restores transport, indicatingtransport is MgATP dependent (Allen et al., 1985).The terms ‘anterograde’ and ‘retrograde’ transport were initially used todistinguish the direction of microtubule-based transport in axons but with new modelsbeing studied, the terms are misleading; they imply that all microtubules are orientedwith their plus end directed toward the cell periphery and they don’t apply to cell freesystems. The terms minus-end and plus-end directed motors indicate the direction oftransport on microtubules and are more accurate.An earlier hypothesis proposed that microtubule polarity was the soledeterminant of the direction of microtubule based transport, like a one way street. Thishas been challenged by the fact that, with the notable exception of dendrites in neurons(Burton, 1988, Baas et al., 1988), cytoplasmic microtubules are unipolar tracks(Burton and Paige, 1981; Heidemann et al, 1981; Stebbings and Hunt, 1983; Euteneueret al., 1989; Troutt and Burnside, 1988a,b; Baas et al, 1989; Bacallao et al., 1989;Black and Baas, 1998; Mogensen et al., 1998; Redenbach and yogI, 1991). Usingflagellar microtubules as tracks, Gilbert and coworkers (1985) showed that axoplasmicvesicle transport occurs bidirectionally on the unipolar microtubules. The in vitroassays of Vale and coworkers (Vale et al, 1985a-d) clearly showed that the polarity ofmicrotubules does dictate the direction of transport, for given motors. This leftunanswered: 1) how does bidirectional transport occur on single microtubules; 2) howis the direction of transport determined for a given vesicle and; 3) at what level is thisdecision regulated?MechanoenzymesThe issue of direction of transport was partially addressed with the discovery ofthe two families of microtubule associated mechanoenzymes, kinesin (Vale et al,Introduction 621985a,b,d; Brady, 1985) and cytoplasmic dynein (Paschal and Vallee, 1987). A greatdeal of evidence supports the proposal that kinesin and cytoplasmic dynein are the plus-end and minus-end directed motors respectively (Vale et al., 1985a,d; Paschal et al.,1987; Vale, 1987, 1990). However, the elegant simplicity of this story may be shortlived, in light of a number of exceptions reported recently (Malik and Vale, 1990; Vale,1990; Walker et al., 1990; Hirokawa et al, 1990; McDonald et al., 1990; Schliwa etal., 1991). The characterization of the mechanoenzyme families, by their molecularstructure, biochemical properties, sensitivity to inhibitors, and rate and direction oftransport provides criteria for their family members and a basis for comparison ofthose that may belong to other groups. Many of the properties of these putative motorfamilies have been derived from microtubule binding, ATPase activity and motilityassays which, while good predictors of one another, are not the same (Vallee, et al.,1 989b). The mechanoenzyme activity of these motors is stimulated by microtubules.The effect of organelle binding is not known. The properties of mechanoenzymes varyconsiderably, depending on their source (Hollenbeck, 1989a,b), experimentalconditions, and purity (Wagner et al., 1989). Purified systems can also be expected todiffer substantially from in vivo systems. These differences in method ofcharacterization have important implications for this study. Bearing this in mind, anattempt is made here to generalize the properties representative of each family,acquired under a variety of test conditions.Kinesin:Kinesin, was the first microtubule-based mechanoenzyme to be identified (Valeet al., 1985a,d; Brady, 1985; for reviews see Pratt 1989a; McIntosh and Porter,1989; Vale, 1987; Bloom et al., 1989). It has been isolated from a number of speciesincluding squid (Vale et al., 1985a,b), chicken (Brady, 1985), cow (Kutznetsov andGelfand, 1986), sea urchin egg (Scholey et al., 1988), and pig (Amos, 1987). Thebasis of kinesin isolation is its tight binding to microtubules in the presence of63AMPPNP, or ATP depletion (Cohn et at., 1989; Amos, 1987), and exclusive GTPdependent release from microtubules (Paschal et at., 1987a). In addition, generalmembers of the kinesin family are distinguished from cytoplasmic dyneins by their:support of plus-end directed organelle transport, response to GTP, and relativeinsensitivity to the pharmacological agents vanadate, NEM, and EHNA. The kinesinfamily of mechanoenzymes share the following characteristics.molecular structure:Combined biochemical and ultrastructural data indicate that kinesin is a tetramerwith a molecular weight of 350kD (Vale et al., 1985b) composed of two globular heavychains (90-135 kD (Vale et at, 1985d) joined by a bent stalk, that terminates in afeathered end comprised of 2 light chains (55-80 kD) (Mcintosh and Porter, 1989;Vale et at., 1985d). It has a sedimentation coefficient of 9.5 S. The ATPase activityresides in the heavy chains (Yang et al., 1989 Hirokawa et al., 1989), which is thoughtto be the microtubule binding site (Amos, 1987; Gilbert and Sloboda, 1986), while thefeathered end, is the organelle binding site (Yang et al, 1989; see Sheetz, 1989).Although usually seen in a bent formation (Amos, 1987; Hirokawa et at., 1989), themolecule extends to 80 nm in length (Hisanga et al., 1989; Hirokawa et at., 1989) bentto 32nm (Hirokawa et al., 1989).biochemical properties:Kinesin has a low ATPase activity (Brady, 1985; Kutznetsov and Gelfand,1986), that is stimulated by microtubules or Ca2 (Cohn et al., 1987). It forms rigorcomplexes with microtubules in the presence of the nonhydrolyzable analogues of ATP,which are releasable with ATP and GTP (Lasek and Brady, 1985; Vale et at, 1985a;Porter et at., 1987).Introduction 64sensitivity to inhibitors:A number of inhibitors modify both the ATPase activity and transport capabilityof kinesin, although not always to the same extent (for reviews see Penningroth, 1989;Vale, 1990; Vallee et al., 1989; Wagner et al., 1989). Kinesin differs fromcytoplasmic dynein in its generally low sensitivity to NEM (Vale et al, 1985a; Lye,1989) and insensitivity to vanadate dependent photocleavage (Porter et al., 1987;Scholey et al., 1988). Although kinesin is relatively insensitive to vanadate atconcentrations lower than 25 jiM, both ATPase activity and motility are reduced by 100iiM vanadate (Lasek and Brady, 1985; Vale et al., 1985d; Porter et al, 1987).However, inhibitions at lower concentrations have been reported (Wagner et al., 1989).Porter and coworkers (1987) found that kinesin based motility declined in the presenceof 3-5 mM NEM, but, unlike with other inhibitors, the effect is caused by failure tobind, rather than formation of rigor complexes. In contrast, kinesin ATPase activity ismarkedly reduced and transport activity suspended by quantities as low as 5jiM AMPPNP(Okaba and Hirokawa, 1989), or by the ATP depleting agents apyrase or hexokinase andglucoserate and direction of transport:Kinesin’s plus-end directed organelle transport activity (Vale et al, 1985a,cd;Brady, 1985; Porter et al., 1987ab, Scholey et al., 1988) is dependent, in a dosedependent manner, on MgATP (as low as 10 j.tM)(Porter et aI., 1987) and GTP (above 1mM) (Vale,et aL, 1985c) at an in vitro rate of 0.6jim/sec (Vale et al, 1985c; Porter etal, 1987a). Low concentrations of ATP slow the rate of transport and the number oforganelles participating. Kinesin’s transport properties diminish with increased purity(Schroer et al., 1989; Schroer and Sheetz, 1991b), supporting the growing view thataccessory proteins are required for transport activity.Introduction 65other properties:Although kinesin can be isolated in both organelle and microtubule fractions, 60-80% is found in the cytosolic fractions (Hollenbeck, 1989a; Vale et at., 1985; Sheetz,1987, 1989). However, even without the use of AMPPNP or ATP depletion, lowquantities of kinesin can be released from microtubules (Porter et at., 1987).Furthermore, it has been identified immunocytochemically with organelles, in vivo, andin isolated vesicle fractions (Brady and Pfister, 1991; Pfister et al., 1989b).Cytoplasmic dyneinThe isolation of cytoplasmic dynein was much more elusive, although its presencewas anticipated, by analogy to axonemal dynein, even before the discovery of kinesin(Pratt, 1980). Its presence was suspected as the minus-end motor in the crudeaxoplasm supernatant Si (Vale et at., 1985d). It has been identified as the highmolecular weight MAP, MAP1C (Paschal et at, 1987; Paschal and Vallee, 1987;Gibbons, 1988), but the term cytoplasmic dynein is in current use, based on is closeresemblance to axonemal dynein (Pratt, 1989a; Gilbert and Sloboda, 1986 Paschal andVallee, 1987,1 988 Paschal et at., 1987). It has been isolated in a wide variety of cellsincluding, bovine brain (Paschal et at., 1987), amoeba (Euteneuer et al., i989a),nematode (Lye et al., 1987), rat liver (Collins and Vallee, 1989), rat testis (Collinsand Vallee, 1989; Neely and Boekelheide, 1988), squid optic lobe (Schnapp and Reese,1989), and chick brain (Hollenbeck, i989a). Like kinesin, cytoplasmic dyneinsupports organelle transport and can bind nonspecifically to glass at the nonmicrotubule end. The basis of its isolation is its binding to microtubules in the absenceof ATP, selective retention during a GTP release step (to remove any bound kinesin), andthen release by ATP. Although there is considerable yield of cytoplasmic dynein usingthis method of isolation, the majority of cytoplasmic dynein is thought to be associatedwith membrane bound organelles.Introduction 66molecular structure:Cytoplasmic dynein is a 140 kD mechanoenzyme consisting of two identicalglobular heads, (Stebbings, 1988; Neely et al., 1989), that comprise the heavy chainsand one or more intermediate and light chains (Vallee, et al., 1989a) sedimenting at20S. The ATPase activity resides on the heavy chains (Vallee et al., 1989a; see Pratt,1 989a), considered to be the microtubule binding end, with variable intermediate (68-72 kD) and light chains (50-55kD) (Neely and Boekelheide, 1988; Collins and Vallee,1989), thought to make up the stalk and small globular organelle binding end (forreview see Schroer and Sheetz, 1991a).biochemical properties:Cytoplasmic dynein has a weak ATPase activity (Paschal et al., 1987) which isstimulated markedly by microtubules in a dose dependent manner (Shpetner et at.,1988). Although dynein hydrolyzes other nucleotides, the rigor complexes it formswith microtubules are releasable only by MgATP (Paschal et al., 1987; Paschal andVallee, 1987) but some forms of cytoplasmic dynein (sea urchin egg dynein) are lessdiscriminating in their nucleotide dependence than others (Pratt, 1986b). Unlikekinesin, but like myosin, it does not form rigor complexes with nonhydrolyzableanalogues to ATP. Its increased binding to microtubules with ATP depletion varies withthe source of cytoplasmic dynein (Lye, 1989).sensitivity to inhibitors:Cytoplasmic dynein is distinguishable from kinesin by its sensitivity to NEM, atconcentrations greater than 1 mM (Shpetner et al., 1988)(being reported as low aslOOJiM) (Okawa and Hirokawa, 1989; Lye, 1989; Paschal et al., 1988; Paschal andVallee, 1987), vanadate, at concentration as low as 10 jiM increasing to lOOjiM(Shpetner et al., 1988), EHNA, at 1-5 mM, although mainly at high ionic strength(Okaba and Hirokawa, 1989; Pratt, 1989a; Shpetner et al., 1988), and vanadateIntroduction 67dependent photocleavage (Lye, 1989; Neely and Boekelheide, 1988). ATP depletiondecreases transport (Vale et al, 1985b).rate and direction of transport:Cytoplasmic dynein supports minus-end directed microtubule based transport(Paschal and Vallee, 1987; Schroer et al., 1988a,b) at a rate of 2.0 tim/sec (Porter etal., 1987).other properties:Based on immunocytochemical staining and negative staining, dynein is thought. toreside mainly on organelles but is also present in solution (Pratt, 1989a) and can beisolated on microtubules.Other putative microtubule-based motors for organelle transportInterpreting the characteristics of kinesin and cytoplasmic dynein broadly, mostisolated motors or proposed motors fit into the two families. However, some decidedly donot. In some cases, it is not known whether these represent a unique family of motors,or if the experimental conditions have produced the contradictions. The failure of highlypurified kinesin to support transport is but one example. They are described here toillustrate the potential for other motor families, as well as the differences that may leadto a better understanding of the underlying mechanisms of organelle transport.visikenGilbert and Sloboda (1986, 1989) have isolated a microtubule associatedprotein: ‘vesiken’, present in both vesicle and microtubule fractions, that cross reactswith MAP2 antibodies. Although transport has not been demonstrated directly, antivesiken antibodies label vesicles and can block vesicle transport. Unlike dynein, vesikendoes not photocleave in the presence of vanadate.Introduction 6810 S sea urchin egg motorCollins and Vallee (1986) have isolated a 10 S sea urchin egg motor with thefollowing properties: microtubule stimulated ATPase activity, hydrolyzes ATP and GTPand is unaffected by vanadate. However, it remains bound in the presence of ATPReticulomyxa motorKoonce and coworkers (Koonce et al., 1985, 1986; Koonce and Schliwa 1985a;Euteneuer et al., 1 988b, 1 989a,c) report a 20-23S putative motor from reticulomyxathat shares many properties with cytoplasmic dynein. It is insensitive to AMPPNP. ATPdepletion increases binding which is released with ATP but not GTP. The motor issubject to vanadate dependent photocleavage and sensitive to 100 pM vanadate, EHNA andmillimolar concentrations of NEM. Intriguingly, it supports bidirectional transport,which reverses readily and can be reversed by phosphorylation.The pharmacological agents inhibit both directions of transport in the same manner,suggesting that the bidirectional transport is not an artifact of isolation (Porter andJohnson, 1989).C. elegans motorLye and colleagues (1987) have isolated a 400 kD putative motor from C.elegans, that resembles both kinesin and cytoplasmic dynein. It was initially reported tohave ATPase activity that is sensitive to vanadate (5 jiM), 1 mM NEM, 5 mM AMPPNP,subject to vanadate dependent photocleavage, and does not release from microtubuleswith GTP. However, it was found to support plus-end directed microtubule-basedtransport. Subsequent studies showed it to be minus-end directed, illustrating thecharacterization problems that occur with different experimental conditions.69A similar danger lies in extrapolating untested parameters from thegeneralization of a group of tested effects (Malik and Vale, 1 990), as the next exampleillustrates.Kar 3 and ncd non claret disjunctional gene productUsing a genetic approach, a putative motor has been encoded from Drosophila nonclaret disjunctional gene from Drosophila. It has properties similar to kinesin withAMPPNP binding, and has a 40-45% homology with kinesin heavy chain (McDonald etal., 1990; McDonald and Goldstein, 1990; Walker et al, 1990; see Vale and Goldstein,1990), suggesting that it belongs to the kinesin family, and it was presumed it would bea plus-end directed motor. However, it is a minus-end directed motor. Its homologywith kinesin is only in the heavy chain (motor) component. The ATPase site is on theopposite terminus from that of kinesin.Organelle-microtubule binding proteinsIt is safe to say that all microtubule-based mechanoenzymes are microtubulebinding proteins. Not all support organelle-microtubule binding; some, for exampleaxonemal dynein, support inter-microtubule transport. Of those proteins that do linkorganelles to microtubules, not all are mechanoenzymes. The term organellemicrotubule binding proteins may best describe these proteins.170 kD protein from Hela cellsThere is some evidence that not all the proteins for which microtubule binding isaffected by nucleotides are motors, but may represent another class of proteins thatfunction to anchor, rather than transport, organelles. Why then do they require ATP?The nucleotide dependence may reflect regulation of microtubule binding byphosphorylation as that described in the 1 70 kD protein reported by Rickard and Kreis(1990, 1991; Rickard and Kreis, 1990), rather than the use of nucleotides forintroduction 70transport. The 170 kD protein is released from microtubules with ATP,in vivo but notin vitro. The presence of ATP prevents binding, due to a change in the phosphorylationstate of 170 kD protein. These authors propose that the ATP dependence is in an enzymemediating binding between endocytic vesicles and microtubules (Scheel and Kreis,1991b), resulting in release of 170 kD from microtubules with ATP and GTP, in thepresence of cytosolic factors. 170 kD protein is unaffected by AMPPNP,immunolocalizes with microtubule pathways in vivo becoming dispersed withmicrotubule disruption. This protein is reminiscent of the 173 kD protein in testis(Neely and Boekelheide, 1988), that is releasable with ATP and GTP. The exactnucleotide interaction remains to be clarified. Proteins such as this one, that arereleasable with both ATP and GTP may have missed detection in other systems becausethey would not be recovered during kinesin isolation, because they are not AMPPNPresponsive, and would have been lost in a dynein preparation because they would notsurvive the early GTP step to release kinesin.The mechanochemical cycles of kinesin and cytoplasmic dyneinIn order to explain some of the effects of pharmacological properties on theATPase activity and motor capabilities of these motors, it is useful to understand theirmechanochemical cycles. Although these cycles are not fully elucidated some informationis available.The mechanochemical cycles of microtubule-kinesin (MK) and microtubuledynein (MD) cycle can be compared to the well studied actin-myosin cycle (AM). Thecycles of these three mechanoenzymes share three properties: 1) the formation ofenzyme-cytoskeletal complexes in the absence of ATP, 2) dissociation of the complexwith the addition of ATP (complexed to Mg), 3) activation of enzyme ATPase activity inthe presence of the cytoskeletal cofactor (Scholey et aL, 1988).In the actomyosin cycle (see Fig.1-8), ATP has two functions. It provides energyfor the power stroke and reduces affinity of the linking partners. The well studiedIntroduction 7 1Figure 1-8: Mechanoenzyme cycles: This figure is an abbreviated version of theevents of the mechanochemical cycles of the interaction of myosin with actin, dyneinwith microtubules (upper figure), and kinesin with microtubules (lower figure). Inthe upper figure, using the well studied actomyosin cycle as the example, the cyclebegins with myosin bound to actin (AM), ATP then binds to AM (AMATP) reducing theaffinity of myosin for actin and resulting in release of myosin-ATP from actin (MATP).ATP is then hydrolyzed to ADP and inorganic phosphate (ADP P1) and Pi is released toyield myosin ADP (MADP). The ADP is released (AM). The kinesin-microtubulemechanochemical cycle is similar except that the microtubule is not released until afterthe release of inorganic phosphate. In the actin-myosin cycle the rate limiting step isphosphate release. In the kinesin-microtubule cycle, the rate limiting step is ADPrelease. The difference between the two cycles is reflected in the fact that actin andmyosin (dynein and microtubules) are released before hydrolysis, but kinesin andmicrotubule release follows hydrolysis. This is the basis for their different response tonon hydrolyzable analogues of ATP.72-Actin tlyo5in and Hicrottibule dynein pathwaysAM AM•ATPAM.ADP— M-ADP M.ADP.Pitiicrotubule Kinesin pathwayMK MKATP MKATPK ADP— MK- ADP—. IlK ADPIntroduction 73actomyosin system provides a model. ATP binds to the rigor complex, actomyosin,producing dissociation to actin and myosin, the ATP is hydrolyzed producing an activatedmyosinADPPi complex with increased affinity for actin. As a result of the higheraffinity, actin binds to myosin and the phosphate (Pi) is released, providing energy forthe power stroke. This cycle would appear to be the same for cytoplasmic dynein as well.The following evidence indicates that the situation is different for kinesin. As anonhydrolyzable analogue of ATP, APMPNP stabilizes the transient intermediates thatoccur just before hydrolysis (Lasek and Brady, 1985) at a point which, in the myosinor dynein cycles dissociation has already occurred, yet in the kinesin cycle dissociationhas not occurred. This indicates that, unlike the case of myosin and dynein, dissociationfollows not precedes hydrolysis, for kinesin (Lasek and Brady, 1985; Vale, 1987). Therate limiting steps are different for the two enzymes with that of kinesin being ADPrelease, and dynein being phosphate release (Pi). For kinesin, it is the fast release of Piand the very slow release of ADP that results in dissociation occuring after hydrolysis(Chilcote and Johnson, 1989; Hackney, 1988, Hackney et at., 1989). The result is thatAMPPNP, by blocking hydrolysis, blocks release of the rigor complexes, in kinesin only.Myosin and dynein have a low affinity for AMPPNP, and in these two motors, AMPPNP isless effective in dissociating rigor complexes than ATP (Chilcote and Johnson, 1989),although slow dissociation occurs. If AMPPNP is added to rigor complexes formed bydynein (or myosin), the binding may be reduced by the slow activity of AMPPNP, actingas a slow substitute for ATP. The effects of AMPPNP for both kinesin and dynein aresmall compared with ATP, and high concentrations of ATP negate these effects. Now itmay be useful to look more closely at the inhibitors of these events.Mechanoenzyme inhibitors and how they workPenningroth (1989) points out that characterization of mechanoenzymeproperties in the presence of selected inhibitors is most reliable in a well definedIntroduction 74system, such as axonemal dynein based transport, where the only ATPase activity isexpressed as inter-microtubule sliding. The complexities of organelle transport, evenin fairly well defined systems, presents some problems of interpretation ofmechanoenzyme behavior. The composition of the motor complex is unknown, as is thepresence of activators that may be present in the cytoplasm. Bearing that in mind, theinhibitors EHNA, Vanadate, NEM, and AMPPNP are frequently used to characterizeputative motors and their properties (See Fig. 1-9 for the molecular structures ofEHNA, vanadate, and AMPPNP).EHNA: (erythro-9(3-(2-Hydroxynonyl) adenine is a structural analogue ofadenosine. It binds at a site that is separate from the catalytic site, inhibits dyneinATPase activity in a dose dependent manner, is reversible with ATP, is effective below0.2 mM ATP, gives a mixed response to dynein, and generally behaves similarly toreduction in ATP. It is more inhibitory on motility mediated by dynein than kinesin(Vale et al., 1985b)NEM: (N-ethylmalemide) is a sulphydryl alkylating agent interfering withSH bonds, and is blocked by ATP, indicating its effect is near the catalytic site. As aninhibitor, it is not very specific in that it potentially affects any SH groups. It doeshowever illustrate the interpretive difficulties that occur comparing ATPase activity,motility, and binding assays. Unlike ATP depletion, which would increase binding butreduce transport, NEM would reduce both, because it interferes with cytoplasmic dyneinby preventing binding rather than by blocking hydrolysis and forming rigor complexes(Porter et al., 1987; Dabora and Sheetz, 1988a).Vanadate: V03 is an analogue of phosphate. It forms by attaching to ADP, withhydrolysis of ATP needed for its formation. It binds near but not in the catalytic site; itdoes not compete directly with ATP; and its effects depend on ionic strength. It acts as ageneral phosphate poison. Therefore, if there is no turnover of binding, no change wouldbe expected. Its effect may be dependent on the relationship between the timing of therate limiting step and product release in the mechanochemical cycle.Introduction 75Figure 1-9: Mechanoenzyme inhibitors: This figure indicates the molecularstructure of analogue of ATP or its components used as inhibitors to characterizemicrotubule-spermatid-ES binding. The inhibitors are shown on the left and theircounterparts on the right. EHNA is a structural analogue for adenosine; vanadate is ananalogue for phosphate; and AMPPNP is a nonhydolyzable analogue for ATP.7(DNHxxç>H HJX>HOCK;N.eryhro-[z-(-W3ckaxynonjl))Ac1er (ii) tde.rosne.0/90nodaA e.9-P-OHPhosphAe.0I I0 0NHHJ%N No 0 0II II ii0—P—o— p—I I I0 0 0 HAMP-PI4P AlPIntroduction 77Alternatively it may be influencing the phosphorylation state of the enzyme or accessoryproteins.AMPPNP: (adenyl imidodiphosphate) is a non hydrolyzable analogue of ATP. Itbinds at the catalytic site inhibiting ATPase activity by blocking the hydrolysis step.This points to a potential difference in the mechanochemical pathway of dynein andmyosin compared with kinesin. The formation of rigor complexes by AMPPNP impliesthat product release follows ATP hydrolysis (kinesin), whereas the absence of rigorcomplexes implies that product release is before hydrolysis (dynein and myosin),although it could conceivably reduce future binding potential in a cycling system as thehydrolysis step increases affinity for binding, at least in the case of actin and myosin.Binding AssaysOne of the criteria for classification of a protein as a microtubule dependentmechanoenzyme is evidence of participation in microtubule-based transport. This isassayed indirectly, in the form of microtubule-stimulated ATPase activity, in ATPaseactivity assays, or directly, in the observation microtubule-based transport, inmotility assays. The formation of rigor complexes can be detected in motility assays,because, if binding occurs followed by the formation of rigor complexes, transport stops.However, transport also stops if binding does not occur. Therefore, lack of motility inmotility assays could be manifested by both binding and non binding, in binding assays.For example when NEM prevents binding, it produces non-transport in motility assaysand non binding, in binding assays. In contrast, ATP depletion produces rigor, alsoproducing non-transport in motility assays but binding in binding assays. This exampleillustrates that binding assays are not synonymous with motility assays.Binding assays have been used for a number of reasons. In a complex system,they provide a means to test a relationship between a given organelle and microtubulesand they allow a known relationship to be characterized. One aspect of microtubulebased organelle transport that has had very little study is the events that occur at theIntroduction 78organelle-motor site. Recently, there has been interest in possible activators formotors that are thought to operate at this site (Schroer et at., 1989; Schroer andSheetz, 1991a,b; Brady and Pfister, 1991).One of the firstbinding assays was used to test the binding of porcine pituitarysecretory granules to microtubules, based on cosedimentation on a sucrose gradient(Sherline et al., 1977). The granules bind microtubules, but not tubulin, in theabsence of ATP and are releasable with both ATP and GTP. Binding is mildly inhibited byAMPPNP. Using a morphological approach, Suprenant and Dentler (1982) demonstratedbinding of anglefish pancreas secretory granules to microtubules. This interaction isreleasable with ATP. Pratt (1986b) assembled vesicle-microtubule complexes in thepresence of supernatant from sonicated axoplasm, producing rigor complexes that werereleasable with ATP or AMPPNP. Polypeptides isolated from the vesicle-microtubulecomplexes showed ATPase activity and remained with the vesicles upon microtubuledepolymerization. Only one of the polypepties extracted from the vesicle-microtubulecomplex required ATP to rebind. Rothwell and coworkers (1989) combined humanneutrophil granules and taxol stabilized microtubules to observe binding that wasincreased with vanadate and decreased with NEM, AMPPNP, and ATP. Van der Sluijs andcoworkers (1990) examined exocytotic vesicle-microtubule binding in the absence ofATP; nucleotide release did not occur at 1mM ATP or 1mM GTP. Exocytotic vesiclemicrotubule binding was found to require cytosol. The binding was sensitive to heat,trypsin, or NEM treatment of the cytosol but not of the vesicles, suggesting that theactivator was soluble. Scheel and Kreis (1991a,) used a new approach to test thebinding of exocytic vesicle to microtubules. Microtubules were complexed to magneticbeads by immunoadsorption to provide an affinity matrix. Trypsin treatment of thevesicles blocked binding, as did heat, trypsin and NEM treatment of the cytosol. Theywere releasable with ATP and GTP but not affected by AMPPNP, ATPyS, GTPPNP, andATPyS (all nonhydrolyzable nucleotide analogues).Introduction 79In each case, described above, the binding assay established a relationshipbetween the organelle and microtubules but further approaches were required todetermine if the binding was mediated by a mechanoenzyme. For example, in an attemptto identify the protein binding exocytic vesicles to microtubules, Scheel and Kreis(1991b) immunodepleted the cytosol of 170 kD protein, kinesin, or cytoplasmicdynein, of which only the removal of 170 kD affected binding, as did salt treatment of theorganelles. This indicates that, in this case, binding is not mediated by dynein or kinesinbut requires a 170 kD protein. They found a time course to the binding, requiring 45minutes to reach maximum binding, It will be remembered from a description ofmicrotubule binding proteins, that a 170 kD protein binds to microtubules, and isreleasable from microtubules when it is phosphorylated by a cytosolic factor. It is notthought to be a motor but whether it has a tethering role, or is involved in regulation ofother motors is not known (Scheel and Kreis, 1991a,b).This group of binding experiments illustrates the diverse properties that havebeen identified in binding assays between membranous organelles and microtubules.These binding assays serve to verify the relationship between microtubules and theseorganelles and characterize their interaction. It is clear that the microtubule-organellebinding properties determined in binding assays will not be the same as the ATPaseactivity or motility assays. Binding only tests part of the mechanochemical cycle andtherefore does not distinguish between mechanoenzymes and non motor microtubulebinding proteins. They may also be testing a different aspect of motor proteinregulation. In most cases, cytosolic factors have been required for binding. Thissupports the growing consensus that accessory factors, either in the cytosol or on theprotein complex, are required for microtubule-organelle interaction.The organelle-microtubule binding complexOne issue, that has not been resolved in microtubule-based organelle transport,is: What constitutes the essential components for transport? It encompasses a numberIntroduction 80of problems. Firstly, what elements constitute the organelle-motor binding, that is,does the motor reside on the organelle and if so is there direct binding (Lacey and Haimo,1992) or are integral membrane receptors required? Secondly, in the case ofbidirectional transport, how does the motor for minus-end directed transport remainavailable during plus-end directed transport, as for example in axons? Thirdly, how isthe direction of transport regulated, that is, how is the appropriate motor activated?Each question has implications for the other.Where is the motor located?Central to this issue is the question: does the motor reside on the vesicle, on themicrotubule, or in the cytosol? The answers have been contradictory, which may resultfrom the questionable assumption that immunolocalization in vivo and colocalization onisolated cell homogenate fractions are measuring the same thing (Brady and Pfister,1991). Hollenbeck (1989) used an antiserum to kinesin, to determine the distributionof kinesin following sequential extraction with saponin, 1% triton and finally SDS, incell fractions that contained cytoplasm, membrane bounded organelles and cytoskeleton.He reported that, in fibroblasts, 68% of kinesin is in soluble form, 32% in membraneor organelle associated form and none is present in cytoskeletal fraction. However,Pfister et. al. (1 989a) used monoclonal antibodies to the 124 kD heavy chain and 64 kDlight chain, showing that kinesin, in cultured cells, associates with membrane boundedorganelles, but not Golgi or nuclear membrane and associates with microtubules, only inthe presence of AMP-PNP. Cytoplasmic dynein has been shown to reside withmembranous organelles (Pfarr et al., 1990; Steuer et al, 1990). Brady and Pfister(1991) address this contradiction of where the motor resides, by suggesting thatorganelle-microtubule binding is weak enough to cause release during isolation, oralternatively, that their concentration in the cytosolic pool is not sufficient to bedetected by immunocytochemical means. A third possibility is that motors may changetheir state such that they are less visible’ to antibodies in the soluble state.Introduction 8 1Presumably much of the variability in the need for cytosolic factors in binding andtransport assays may arise from conditions that mediate the organelle-motor or motormicrotubule interactions and or their activators.In motility assays, three states are observed: vesicles transport alongmicrotubules (motility); vesicles remain bound without transport (binding); orvesicles ‘fall off’ microtubules (dissociated). In keeping with the ATPase activity ofmechanoenzymes being in the microtubule binding domains, it is assumed that much ofthe pharmacological inhibitors reflect activity at the motor-microtubule site, makingthis the most likely site for regulation. It is not known what causes vesicles to abandontheir cycling and dissociate from microtubules, as opposed to forming rigor complexesand remaining bound, but immobile. Nor is it known how high their affinity is tomicrotubules during transport. The question is: where do they reside when transportstops?How are motors kept available for bidirectional transport?This problem is most clearly illustrated in axons, in which organelles travellong distances to the axon terminus and, in modified form, return to the cell body.Protein synthesis occurs only in the cell body; therefore, any motor travellingretrograde would have to be first carried to the terminus. Probably similar but lessdramatic events occur in less polarized cells. If kinesin is in the soluble fraction inaxons, how could it get to the terminus to be used for the return trip? In axons,anterorade and retrograde travelling organelles are different in size and can bedistinguished from one another. Axons were ligated and colocalization of cytoplasmicdynein on retrograde and anterograde vesicles was examined with immunocytochemicalprobes. Cytoplasmic dynein was present on both classes of organelle (Hirokawa et al.,1990), while kinesin was much more abundant on the anterograde than the retrogradevesicles (Hirokawa et al., 1991). However, antibodies to kiriesin inhibit bothanterograde and retrograde transport (Brady et al., 1990). These data suggest that theIntroduction 82retrograde motor cytoplasmic dynein is carried to the axon terminus in some inactivestate ferried by kinesin, then transfers to the retrograde vesicle and returns with alimited amount of inactive kinesin to the cell body. The reason for kinesin’s interferencewith bidirectional transport is not clear.Are the motors alone sufficient for transport?Very little is known about vesicle-motor interactions and how they are affectedduring isolation. Lacey and Haimo have shown that dynein binds directly to extractedsynaptosomes (Lacey and Haimo, 1992). This binding is not ATP releasable, supportingthe impression that ATP dependent regulation is probably not occuring at the organellemotor site. In contrast, many motility assays fail if vesicles are salt extracted (Schroeret al., 1988a,b; Pratt, 1986), or treated with trypsin Gilbert and Sloboda, 1986).Covering the middle ground, Vale et al. (1985c; Schroer et al, 1988) foundmembranous organelles moved along purified microtubules in the absence of solublefactors, but were further stimulated by addition of high speed cytosol. Thecontradictions may lie in the purity of the components or differences between systems.Kinesin and cytoplasmic dynein are both known to support microtubule basedtransport; however, Schroer and coworkers have shown that neither of these motorssupport transport when they are highly purified (Schroer et al., 1988a,b; Sheetz andSchroer 1989; Schnapp and Reese, 1989). Illustrative of this problem, is a study inwhich kinesin was removed from the crude axoplasmic supernatant Si, whichpreviously had been shown to support bidirectional transport (Schroer et al., 1988a).The kinesin depleted supernatant was then added to vesicles that had been washed in highsalt to reduce contamination by supernatant proteins. The result was a reduction inminus-end and plus-end organelle transport. When the removed kinesin was furtherpurified on an affinity column and returned to the Si, anterograde transport was notfully restored, suggesting that something had been removed with the kinesin and had beenlost during the affinity purification. This was verified in that, the elutant from theIntroduction 83kinesin purification column was able to support some bidirectional transport. Thesefindings could be explained by the existence of accessory proteins to mediate organellebinding having been removed with kinesin, or by the occurrence of active and inactivestates of the enzyme and necessary regulators. Sheetz and coworkers (1989) haveproposed that there is an ‘organelle translocation complex’ which may include: themotor(s), accessory protein(s), and an organelle (with a motor binding site). Therecent finding of two potential activators (Schroer and Sheetz, 1991b) supports theproposal that highly purified motors are not sufficient for transport.How are motors regulated?Sheetz and coworkers (1989) have proposed two models to explain theavailability of motors for directional transport. They reconcile the existence of a factorthat is progressively removed during the purification of cytoplasmic dynein (Sheetz etal, 1989; Schroer et al., 1989) which they have termed an accessory factor, withactive and inactive states of the motor. In the first model, the organelle has separatebinding sites for the minus-end and plus-end motors. Organelle receptors are proposedto be active or inactive, resulting in an active complex dictating transport direction andand inactive complex carrying the inactive motor. The second model involves a singlemotor complex containing both motors, in which the motor complex has twoconformations to activate either the minus-end or plus-end component of the motorcomplex. Regulation could occur by a number of methods including: phosphorylation,activation by ‘activator’ proteins (Schroer and Sheetz, 1991b), inactivation byselective proteolysis (Sheetz et al, 1989), or cooperative action between the twomotors.Schroer and Sheetz (1991b) point out that in the presence of highly purifiedmotors, microtubules bind to and transport along glass slides, but do not transportvesicles. This indicates that the transport potential of mechanoenzymes can be inhibitedby their binding to vesicles. Furthermore these workers isolated two activators that areIntroduction 84capable of converting the tethered binding of vesicles to microtubules mediated by highlypurified motors, to vesicle transport. They propose that these activators have beenpresent in the mechanoenzyme preparations that have supported vesicle transport. Thisfinding has important implications for both binding assays and motility assays. Thedifference between binding and non-binding in binding assays, or transport and nontransport in motility assays, may lie not with the motor itself, but whether potentialactivators have been retained in the system used.In that the direction of transport is not the only issue for mechanoenzymes, butorganelle specificity as well, current interest in membrane sorting by capping proteinsmay reveal other mechanoenzyme regulation mechanisms. There may prove to be arelationship between activating proteins and capping proteins involved in organellesorting.Microtubule-dependent orpanelle positioningTwo basic cell functions have been linked with microtubule-based transport:organization and distribution of organelles (Schroer et al, 1988, 1989; Schnapp andReese, 1989; Pfister et al, 1989b), which includes the formation of endoplasmicreticulum networks (Lee et al, 1989; Teraskai et al, 1984; Vale and Hotani, 1988);and events of mitosis and meiosis (Mitchison, 1986).Studies of microtubule-based transport have been extended to a multitude of celltypes, leading to the realization that it is the underlying mechanism for a vast number offundamental cell functions. It will soon be easier to list the cell events that are notmicrotubule dependent than those that are (see Schroer and Sheetz, 1991 a for review).In order to integrate the many cellular functions in which they participate,organelles are transported along, or positioned in, specific locations in the cell.Positioning involves two states transport to and maintenance at the destination, that is,motility or binding.Introduction 8 5Many of the processes in the cell, that involve membranous organelles, aremicrotubule dependent and require the movement of organelles along pathways bothmorphological and physiological (for review Bershadsky and Vasiliev, 1 988).Essentially, they are being either transported or anchored. These two processes mayeven be combined, as in the shaping and elongating of membrane profiles in ER,lysosomes and possibly even elongation of mitochondria involve anchoring in some areaswhile moving in another. They may involve the same mechanisms, although differentproteins, or the same proteins with different instructions.A number of methods have been used to establish the microtubule dependency oforganelle positioning: the effects of microtubule perturbation, colocalization of theorganelle with microtubules, or by direct observation with video enhanced microscopy.Fig. 1-10 illustrates the many organelle pathways that are microtubule dependent. Ingeneral, the pathways that require organeiles either cover relatively long distances, orthere is a short life span of the components, requiring efficiency in both cases (forreviews see Duden et al., 1990; Kreis, 1990; Kelly, 1990a,b; Schroer and Sheetz,1991a; Bomsel et al., 1990). Many pathways spanning short distances such as, betweenGolgi stacks, membrane to early endosome, late secretory granules to membrane, orfrom ER to cis Golgi, do not require microtubules.In the secretory pathway, formation and maintenance of ER networks ismicrotubule dependent (Terasaki et al., 1984; Lee and Chen, 1988); it collapses withmicrotubule disruption (Terasaki et al., 1984) and can be formed in vitro by a kinesindependent process (Vale and Hotani, 1988; Dabora and Sheetz, 1988b). Although thepathway from ER to Golgi is not microtubuIe dependent, the salvage compartment, inwhich vesicies are returned from trans Golgi to ER to retain ER specific proteins,requires microtubules (Achier et al, 1989; van Zeijl and Matlan, 1990). In polarizedcells, the targeting of proteins to the basolateral surface is by bulk flow, but to apicalsurfaces is microtubule dependent. This was illustrated in a particularly interestingstudy in which microtubule disruption caused the failure of targeting of apical proteins86in the brush border of intestinal epithelial cells (Achier et al, 1989). Profiles ofbrush border structures, complete with microvilli and actin components, were seen tomistarget to basolateral surfaces, or assemble within the cytoplasm. Glycosylation of ofmistargeted apical membrane proteins indicates that the mistargeting occurs after theypass through the Golgi (van Zeijl and Matlin, 1990). Exocytotic vesicies have beenshown to bind to microtubules by a protein that is thought to be involved in anchoringrather than transport (Scheel and Kreis, 1991a,b).Placement of the Golgi is microtubule dependent, in that disruption ofmicrotubuies causes dispersal of the Golgi, although its internal structure remainsintact and the Golgi is still able to function (Bacallao et al., 1989; Ho et al., 1989; seeKreis 1990 for review). Removal of the disrupting agent causes microtubule dependenttransport of Golgi to the microtubule minus-end in some cell types (see Duden et al.,1990). In the endocytotic pathway, the movement of early endosomes is microtubuleindependent, but their transport to and fusion with late endosomes or lysosomes requiresmicrotubules (Bomsel et al., 1990), as does lysosome elongation (see Duden et al.,1990).It would seem, with the positioning of most organelles and their transport duringcell activities, that there must be an overall coordination in what Heuser (1989) hascalled the “continuous intracellular circulatory system’. Alan and Vale (1991), havedescribed a cell cycle dependent regulatory factor that serves to activate microtubulebased organelle positioning. Using extracts from metaphase and interphase cells, theyshowed that organelles from both cell types could participate in microtubule-basedtransport and membrane fusion in the presence of interphase cell extracts, but notmetaphase cell extracts. This indicates that these events are activated in interphasecells by a regulatory mechanism that is absent in metaphase cells.Certainly, organelle positioning on microtubule tracks results in cell specificorganelle patterns with some common underlying themes.Introduction 87Cellular organization (microtubules in cells in general)Until very recently, it was assumed that microtubules radiate from aperinuclear position, oriented with their minus-ends directed toward the cell surface.In this plan, the organelles would be placed by appropriate minus-end or plus-enddirected motors to be positioned or transported, as required, along microtubule tracks(for reviews see Gibbons, 1988; Duden et al., 1990, Kelly, 1990a,b). Both kinesinand cytoplasmic dynein have been implicated in these events (for review, see Schroerand Sheetz, 1991a). However, Bacallao and coworkers (1989) have shown that duringthe polarization of MDCK cells, there is reorganization of cellular organelles andredistribution of microtubules. This results in separation of the paired centrioles,migration of centrosomal material to the apical region of the cell, from whichmicrotubules emanate, with their minus-ends toward the cell surface and their plus-ends toward the nuclear region of the cell. Microtubules in the basal area are lessorganized. The junctional complex is reported to move upward toward the apical region.The Golgi is repositioned in a ribbon-like complex extending apically from the nucleus.The authors noted a striking similarity of the Golgi pattern of the polarized MDCK to theGolgi pattern of Sertoli cells, as described by Rambourg and coworkers(1979). Thisextensive reorganization, indicates that polarized epithelial cells may have a differentcontinuous organelle circulatory system than the model shown in Fig. 1-10. Althoughthere is little direct evidence, it is reasonable to assume that Sertoli cell microtubulesparticipate in most of these housekeeping activities.Schroer and Sheetz (1 991 a) have proposed three models of cellular organizationto accommodate these observations. The first, the radial model accounts for the planshown in Fig. 1-10), typical of what has been observed in many cell lines. The second, alinear model accounts for the overlapping microtubule plan that has been identified inaxons. The third, addresses the findings of the MDCK cell organization (Bacallao et al.,1989) and is supported by findings of minus-end out orientation of the Drosophila wingIntroduction 88Figure 1-10: Continuous organelle circulatory system: radial model:This diagram indicates the classical radial model of microtubule and membranousorganelle distribution and pathways that have been shown to be microtubule dependent(for review see Schroer and Sheetz, 1991a). The microtubule dependent pathwaysinclude: formation of ER networks; membrane recycling from the transitioncompartment; positioning of the Golgi apparatus; some secretory vesicle transport;movement of endosomes from the early to late endosome compartment; cycling to thetrans Golgi network; transport of endosomes to lysosomes; and the shaping of lysosomes.Those pathways that are thought to be microtubule independent are movement of vesiclesbetween Golgi stacks, movement of ER to Golgi, and the early events of endocytosis andlate events of exocytosis. (Modified from Schroer and Sheetz, 1991a; Duden et al.,1990; Kreis, 1991; Kelly, 1991 a,b)89Introduction 9 0epidermal cell (Mogensen et al., 1989) and retinal pigment epithelial cell (Troutt andBurnside, 1988b), in which microtubule minus-ends are directed toward the cellperiphery. The prospect of microtubule polarity being different in polarized epithelialcells, has prompted a numbers of reviewers to emphasize the importance of determiningmicrotubule polarity in cells to the understanding of the events of organelle positioning,and the motors that carry it out (Boekeiheide et al, 1989; Kelly, 1990a,b; Duden et al.,1990; Kreis, 1990).MICROTUBULES IN SERTOLI CELLSDistributionThe distribution of microtubules in Sertoli cells was first described byChristensen (1965) and later in greater detail by others (Fawcett, 1975; yogI et al.,1983a,b; VogI, 1988; Amlani and VogI, 1988; Hermo et al., 1991). Sertoli cellmicrotubules have been reported to be abundant in the Sertoli cells of all mammalianspecies studied. Although mention is made of them in most accounts of Sertoli cellmorphology, only a few studies have addressed their distribution in Sertoli cells (yogI etal., 1983a,b; yogI, 1988; Amlani and VogI, 1988), tending instead to focus observationson germ cell microtubules (Wolosewick and deMay, 1982; Cherry and Hsu, 1984).In Sertoli cells of both the squirrel and rat, microtubules are prominent in thebody of the cell, extending from a supranuclear position into the apical stalk and apicalprocesses that support late spermatids (Amlani and yogI, 1988; yogI, 1988). In Fig. 1-11, immunofluoresent staining with a polyclonal antibody to tubulin, shows the parallelorientation of microtubules in apical portions of rat seminiferous epithelium. Inimmunofluoresence studies, few microtubules have been described at the base of the cell(yogI et al., 1983a,b, Amlani and VogI, 1988 and VogI, 1988) but they have beendetected basally with antibodies to tyrosinated alpha tubulin (Hermo et al., 1991).There is ultrastructural evidence for a limited number of somewhat randomly organizedmicrotubules in this location (Fawcett, 1975).Introduction 9 1Figure 1-11: Microtubule distribution in Sertoli cells: These are pairedphase contrast and immunofluoresence light micrographs of apical fragments of Sertolicells, reacted with a polyclonal antibody to tubulin (gift of Kegi Fugiwara). Earlyspermatogenic cells (asterisk in phase image), are seen as dark areas in (b).Microtubules do not occupy the lateral processes that surround these earlyspermatogenic cells. Late spermatids can be identified by their dark nuclei and tails inthe phase image in (a). The late spermatids in (a) correspond to the negative images in(b) indicating the deep placement of spermatids in Sertoli cell crypts. Positive stainingSertoli cell microtubules are seen in (b) oriented parallel to the long axis of the cell andsurrounding the late spermatids. Tubulin containing axonemes in germ cells do not stainin this preparation. bar = 10 jim. (Micrograph courtesy of W. yogI from VogI et al.,1991b and 1992)474)Introduction 9 3Microtubules do not extend into lateral processes that encircle meioticspermatocytes. However, they are present in lateral processes surrounding late roundspermatids, in squirrel (Amlani and yogI, 1988), and early elongating spermatids, inrat (yogI, 1988). As spermatids become oriented perpendicular to the base of theepithelium and become situated within Sertoli cell crypts, they become richlyensheathed by dense bundles of microtubules, continuous with those of the Sertoli cellstalk.Thick bundles of microtubules extend into apical processes that support Sertolicell crypts running parallel to the long axis, as well as the path of transport, ofmaturing spermatids. Subsequently, they become concentrated, adjacent to theacrosome, around translocating spermatids, and remain there until spermiation. Theappearance of microtubules in Sertoli cell processes with their regular spacing andadjacent organelles has been compared to that in axons (Fawcett, 1975; yogI et al.,1983a; Neely and Boekelheide, 1988). Microtubules course past the cistern of ESER aswell as other membrane bounded organelles. Elongate mitochondria are prominent inareas rich in microtubules. Considering this distribution, Sertoli cell microtubuleshave opportunity, through ESs, to serve as tracks on which to translocate germ cells.Disruption of Sertoli cell microtubules following colchicine injections, resulted in basalaccumulation of smooth ER and failure of elongate spermatids to be moved apically to thetubule lumen. This is consistent with apical translocation of ER and spermatids onmicrotubule tracks.Sertoli cell MTOCThe site of microtubule organization, in Sertoli cells, has received virtually noattention. Paired centrioles have been observed in a supranuclear position (Nagano,1966). However, Sertoli cell microtubules do not appear to radiate from this area,instead running parallel to the body of the cell (Amlani and yogI, 1988; VogI, 1988),typical of epithelial cells generally (Kelly et al., 1990a,b; Schroer and Sheetz, 1991a).Introduction 94In MDCK cells, grown to confluence, microtubule organizing centers have been describedto relocate, and microtubules reorganize, during the establishment of cell polarity(Bacallao et al., 1989; Bré et al, 1989). This may occur in Sertoli cells.Tubulin post translational modifications in Sertoli cellslmmunocytochemical staining with antibodies to tyrosinated tubulin indicates aprogressive increase in density of staining in the apical region, from the time thatspermatids are deeply situated in Sertoli cell crypts to the time they reach their mostapical position: stages II to VIII (Hermo et al., 1991), concentrated in apical processessurrounding step 15 to 19 spermatids. This finding is supported by observations of yogIand coworkers (yogI et al. 1983a,b; Amlani and VogI, 1988; yogI, 1988) which placethe maximum density of microtubules in apical processes surrounding elongatespermatids beginning at stage V. Specific changes that occur in relation todifferentiating spermatids are described below.Staae dependent changes in Sertoli cell microtubule distributionInformation about stage dependent changes in Sertoli cell microtubules isprimarily morphological. A number of changes in the distribution of Sertoli cellmicrotubules during spermatogenesis have been described, including their distributionin lateral, penetrating, and apical processes. Microtubules do not extend into the lateralprocesses that envelop spermatocytes and early spermatids (yogI et al., 1 983a; VogI etal., 1988; Amlani and yogI, 1988), being first observed in these processes as theysurround elongating spermatids that are being reoriented to the base of the Sertoli cell.Microtubules are most abundant surrounding elongate spermatids as they becomesituated within Sertoli cell crypts (stage II to IV), descend deep within the epithelium(stage VI- stage V), and are subsequently moved toward the tubule lumen, supported bySertoli cell apical processes (stage VII stage)(VogI et al., 1983a; yogI, 1988; Amlaniand yogI, 1988; Hermo et al., 1991). Microtubules pass adjacent to ESs and around theIntroduction 95elongating spermatid, reaching into the most apical confines of the cell (Fig. 1-1 1). Asmicrotubules extend into apical processes, the modifications they undergo to conform tothe changing shape of germ cells are different from a simple accommodation to an objectplaced in a stream of microtubules. There is a species difference in this conformationimmediately adjacent to spermatid heads (yogI, 1988, Amlani and VogI, 1988).In both squirrel and rat, the microtubules remain faithful to the shape of thespermatid head, but take on a species specific distribution when apical processes extendlate spermatids toward the epithelial lumen (Amlani and yogI, 1988; yogI, 1988). Insquirrel, microtubules surround the saucer shaped spermatid head forming a C shape,lining the inner aspect of the rim of the lip of the acrosome. In immunofluoresenceimages of fixed and fragmented sections of apical Sertoli cells, in squirrel, thesemicrotubules appear to have lost their continuity with the microtubules of the stalk(VogI et al., 1983a; Amlani and VogI, 1988). An accumulation of the elaborate networkof SER adjacent to the concave surface of the squirrel spermatid head corresponds to theC shaped microtubule confirmation, suggesting a relationship between the two events.The reorganization of microtubules adjacent to spermatid heads in the rat issimilar, although less striking, forming a canoe shape that cradles the late spermatid,and concentrating particularly at the dorsal and ventral aspects of the head (yogi,1988). In rat, the microtubules surrounding spermatid heads appear to retain theircontinuity with those in the apical processes. These specific differences suggest morethan a simple bypassing of spermatid heads but an intimate involvement in thisrelationship, possibly involving placement and orientation of the spermatid and theassembly of nearby ER (Clermont et al., 1980).Although stage and species specific changes occur in the distribution of Sertolicell microtubules during spermatogenesis; not a lot is known about the factors thatinduce this redistribution, or for that matter, the more subtle alterations that occur inmicrotubules themselves. The profile of post-translationaly modified tubulins inSertoli cell microtubules, during different stages of spermatogenesis, indicates thatIntroduction 9 6stage dependent changes do not arise from the tubulin modifications presently known(Hermo et aL, 1991) The functional capacity of microtubules is also modified byinteractions with other molecules including ions, nucleotides, regulatory proteins, andMAPs.Sertoli cell MAPsCytoplasmic dynein has been isolated in high quantities in testis (Neely andBoekelheide, 1988; Collins and Vallee, 1989) and in Sertoli cell enriched preparations(Neely and Boekelheide, 1988) In a profile of testis MAPs, Neely and Boekelheidedescribe HMW MAPs similar in electrophoretic properties to brain MAP, MAP1 andMAP 2, although they do not share immunoreactivity with brain MAPs MAP1A and MAP2. They differ from brain MAPs in a reduced ability to promote microtubulepolymerization and lack of heat stability. The MAP content changes with developmentwith a prominent MAP, ‘HMW-.2’, present at all time points. Furthermore, this MAP ispresent in Sertoli cell enriched and germ cell depleted (cryptorchid and 2,5-hexanedione treated) testis preparations, indicating a likely localization in Sertoli cells.HMW-2 (Neely and Boekelheide, 1988) is released from microtubules by MgATP andpossesses microtubule stimulated ATPase activity, which, in turn, is inhibited byvanadate (75% at 5j.tM, 58% at 5OiiM), vanadate dependent photocleavage and EHNA(4mM but not 0.4mM). In addition, Neely and Boekelheide report the presence of a 173kD protein, which is partially released with 5 mM GTP. In a MAP profile from 5, 10,15, 24 day old and adult rat testis, a MAP of a slightly slower electrophoretic mobilitythan HMW2, ‘HMW-3)’ is first identified as a minor component at 24 days and is amajor component in adult testis. It has not yet been established whether this MAP onlyoccurs while late spermatids are present, or whether it is a variant of another MAPpresent earlier.Introduction 9 7Relationship of microtubules to other organelles in Sertoli cellsThe close packing and parallel alignment of Sertoli cell microtubules is reflectedin the shape and orientation of nearby organelles. Numerous elongate mitochondria andlong profiles of smooth endoplasmic reticulum conform, in alignment and shape, tomicrotubules in Sertoli cell apical processes. In contrast to the shape and orientation ofthese organelles in the apical portion of Sertoli cells, Fawcett (1975) describesmitochondria to be less elongate and randomly distributed in the base of Sertoli cells, asare microtubules. Microtubules occur in the vicinity of elaborate changes in ERdistribution that occur in Sertoli cells of the rat (Clermont et al., 1980) and squirrel(yogI et al., 1983a,b). While it is not known whether microtubules play a role inshaping these organelles in Sertoli cells, a role has been established for shaping of ER invitro (Vale and Hotani, 1988; Dabora and Sheetz, 1988) and in vivo (Lee et al., 1989).Mitochondria have been shown to form links with microtubules (Hirokawa, 1982;Linden et al., 1989) and have been observed to travel along microtubule tracks in axons(Forman et al., 1983). It is not known whether Sertoli cell mitochondria and otherorganelles, observed in close relation to Sertoli cell microtubules, are undergoingtransport or being maintained in that location.Function of Sertoli cell microtubulesMost of the suggested functions of microtubules in Sertoli cells arise fromanalogy with microtubule functions that have been demonstrated in other systems (Valeand Hotani, 1988; Allen at al., 1981; Duden et al., 1990; Sheetz et al, 1986; Schnappet al., 1986; Schliwa, 1984; van ZeijI and Matlin, 1990; van der Sluijs et al., 1990;Balch, 1989; Ho et al., 1989; Heuser, 1989; AchIer et al, 1989; Burnside, 1989;Bomsel et al, 1990; for reviews see Schroer and Sheetz, 1991 a; Duden et al, 1990; andKelly, 1 990a,b). Collectively, these and other studies have provided an extensive list ofongoing functions in interphase cells, which are generally agreed to be microtubuledependent, including: maintenance of cell shape, transport for late events in endocytosisIntroduction 98and early events in exocytosis, recycling of membrane from cis Golgi to ER, targeting ofapical membrane proteins, formation of ER network; Golgi assembly and placement; andlysosome shaping and distribution, all described earlier. Although most of these eventsapply to all cells, few have been demonstrated in Sertoli cells. Apart from microtubuledisruption studies and their inherent problems, most evidence of microtubule functionin Sertoli cells is morphological and somewhat speculative. The complexity of theseminiferous epithelium presents technical constraints for direct testing of Sertoli cellmicrotubule function. Functions that have been suggested for Sertoli cell microtubulesinclude: maintenance of cell shape, shaping of the spermatid head, positioning andtranslocation of organeltes, and translocation of residual cytoplasm and spermatids. Theabundance of microtubules and MAPs in Sertoli cells argues for an importantmicrotubule function, possibly one that is unique to Sertoli cells. Sertoli cells havebeen relegated the special function of nurturing and maintaining germ cells duringspermatogenesis (Dym, 1977; Ritzen et al., 1981; Griswold, 1988). A special functionfor Sertoli cell microtubules may be directed toward that end.Sertoli cell shapeOne of the first functions attributed to microtubules was maintenance of cellshape (Porter, 1966). Extensive changes occur in Sertoli cell shape (Clermont et al.,1980; Dym,. 1977; Fawcett, 1975; Gravis, 1979; Morales and Clermont, 1982; Ross,1976; Russell, 1980, 1984; Russell and Malone, 1980; yogI, 1988; yogI et al.,1983a,b, 1985a, Amlani and yogI, 1988). They are directly related tospermatogenesis, and result from the accommodation of Sertoli cell processes to thechanging shape and position of developing spermatids, and the selective retention ofresidual cytoplasm.Sertoli cells lack the apical tight junctional complexes typical of many polarizedepithelia. Unlike most other epithelia, Sertoli cells are highly convoluted, however theyare tightly bound to apically placed germ cells. The extensive germ cell-Sertoti cell99attachment and the interconnection between germs cells by intercellular bridges mayserve to substantially stabilize much of the apical region of the epithelium.Reinforcement by microtubules may add to this stability. Microtubule bundles inSertoli cells run parallel to one another maintaining an even spacing, forming whatlooks on electron micrographs to be a zone of exclusion around each microtubule. Withfine linkages frequently seen between microtubules, the easy assumption is that a cross-linked scaffold exists. MAPs have been proposed as the source of linkages seen extendingfrom microtubules and have been proposed to provide the zone of exclusion. How couldSertoli cell microtubules be cross-linked to reinforce cell shape and at the same timeallow passage of organelles along microtubule tracks? The question of microtubulecross-linking has been addressed in a study testing the effect of a potential cross-linkingMAP, MAP-2 in the presence of the mechanoenzyme, kinesin (Massow et al., 1989;Hems et al., 1991). MAP-2 is a larger molecule than kinesin. It was found that MAP-2, included in the incubation solution of kinesin on a glass substrate, acted as a spacerdisabling kinesin’s ability to transport microtubules. However, if MAP-2 waspreincubated with the microtubules and not the substrate, glass bound kinesin was ableto interact with microtubules, and induce transport, in spite of microtubule boundMAP-2 (Hems et al., 1991). It was deduced that MAP-2 may act as a spacer protein notby cross-linking microtubules but by repelling adjacent microtubules. Such amechanism for the creation of a network of evenly spaced microtubules would addressthe problem of moving organelles through a cross-linked scaffold at the same time asproviding the function of maintaining cell shape. Although there is no evidence for MAP2 in Sertoli cell microtubules, a similar mechanism would provide for the movement oforganelles including spermatid bound ESs through a field of microtubules without energyconsumptive binding and release and without diminishing the role of microtubules in themaintenance of cell shape. Support for a role for Sertoli cell microtubules in themaintenance of cell shape comes from evidence that when Sertoli cell microtubules aredisrupted with coichicine, Sertoli cells loose their columnar shape, or become unstableIntroduction 1 00with the apical portion bulging into the lumen, in some cases sloughing a liningconsisting of apical Sertoli cells and associated spermatids (Russell, et al., 1981; yogIet al., 1983b).Organelle positioning and translocation during spermatogenesisThe unvarying association of specific stages of developing germ cells with oneanother in the seminiferous cycle (BrOkelmann, 1963; Perey, et al, 1961; Clermont,1972) and stage dependence of cycling biochemical and morphological events in Sertolicells indicate that the Sertoli cell milieu changes in a constant and regulated manner.Dynamic changes that occur in the positioning of such organelles as Golgi andendoplasmic reticulum are stage dependent (Clermont et al, 1980; Rambourg et at.,1979; Ueno et al, 1991) and therefore likely to be important to the overall program.In the face of constant change in cell shape, critical positioning of organelles is requiredto place components of the synthetic and secretory pathways in positions of optimumefficiency for the synthesis of proteins, targeting of membrane proteins, and delivery ofsecretory products to their destination. Considering the distances involved, leaving suchpositioning to diffusion would be inefficient.Evidence for a role of Sertoli cell microtubules in the shaping and distribution oforganelles is indirect. In the apical region of the cell, organelles are observed in closeassociation with microtubules; they are oriented parallel to microtubules; and theyassume elongate shapes consistent with elongation of membrane bounded organelles bythe adjacent microtubules. Sertoli cell microtubules have been reported in areas ofaccumulated ER (Clermont et al, 1980; VogI, et al., 1983a). The close temporal andspatial association of ER with microtubules next to the acrosome lip of late spermatid insquirrel, as well as the failure of that network of ER to accumulate apically followingmicrotubule disruption (VogI, 1983b), is consistent with a role for transporting andaccumulating ER in that region (yogi, 1983a; Amlani and yogI, 1988). A role forSertoli cell microtubuies in the orientation and shaping of organelles can only beIntroduction 1 0 1predicted by analogy to other systems, where positioning and shaping of organelles hasbeen shown to be microtubule dependent (Lee et al., 1989; Vale and Hotani, 1988).Influence of Sertoli cell microtubules on the head shape of spermatogeniccellsThe role of microtubules, of either spermatid or Sertoli cell origin, in theshaping of spermatid heads has been the subject of a long-standing debate (Fawcett et al.,1971). Sertoli cell microtubules do not lie adjacent to spermatids, but are separated byESs. It seems unlikely that microtubules could exert a force through these structures tomodify the shape of the spermatid head. In addition, the orientation of Sertoli cellmicrotubules is not consistent with the head shape changes that occur. Evidence ofSertoli cell microtubules as contributors to shaping of spermatid head is indirect andcomes mainly from perturbation of microtubules in hamsters (Rattner, 1970), mouse(Handel, 1979; Wolosewick and Bryan, 1977), rat (Aoki, 1980; Russell et al 1989),and squirrel (VogI et al., 1 983b). Of those that reported on the effects of Sertoli cellmicrotubule perturbation on head shape, yogI et al. (1983b) and Handel (1979) founddistortion in head shape while Russell et al. (1981) did not. However, the latter groupquestioned whether the apparently normal spermatid heads, observed in their study, hadin fact been exposed to the agent before being sloughed into the tubule lumen.Considering the limitations of these studies, the evidence is scant. Aside from theperturbation studies, the only other observation addressing the question of microtubuleinvolvement in the determination of head shape has been that of dense bands of Sertolicell microtubules which conform intimately to the inner lip of the acrosome duringshaping of spermatid head in the squirrel (Amlani and Vogl, 1988), and are notseparated by ESs. These authors concede the possibility that this band of Sertoli cellmicrotubules may have some influence on formation of that portion of the acrosome. Thesuggestion that the spermatid nuclear shape is due to chromosome condensation (Fawcett,introduction 1 021971) while attractive because of the wide range of species specific spermatid shapesthat occur (Fawcett et al. 1971), has no experimental support.Microtubule perturbationA number of pharmacological agents including vinblastine, nocodozole,colchicine, and colcemid, disrupt microtubules, while taxol stabilizes microtubules.Investigators have enlisted these agents to disrupt or stabilize microtubules in attemptsto explain the role of microtubules in cells. Studies of this type have always beenviewed with caution because: with systemic treatment they produce profound morbidityin the animal; it is difficult to determine if the observed effects are primary effects dueto loss of the microtubule scaffold, or secondary to the disruption of another microtubuledependent process; and these agents do not bind exclusively to microtubules. Thesecautions are increasingly germane with the blossoming list of cell events in whichmicrotubules are implicated (Schroer and Sheetz, 1991a).Viewed in that light, a number of investigators have used microtubule inhibitingagents to understand the function of microtubules in hamsters (Rattner, 1970), mouse(Handel, 1979; Wolosewick and Bryan, 1977), rat (Aoki, 1980; Russell et al 81,89),and squirrel (yogI et al., 1983b) Sertoli cells. The agents used were microtubuledepolymerizing drugs; colcemid (Wolosewick and Bryan, 1977; Rattner, 1970),colchicine (Handel, 1979; Wolosewick and Bryan; 1977; Russell et al.1981; Aoki,1980; yogI et al., 1983b); nocodozole (Aoki, 1980) and vinblastine (Wolosewick andBryan, 1977; Russell et al., 1981). In most reports, differences betweendepolymerizing agents (or their doses) were essentially a matter of degree.A common finding of microtubule disruption studies is the sloughing of apicalportions of Sertoli cells and associated spermatids. This may result from a loss ofreinforcement by microtubules in the Sertoli cell stalk. The remaining Sertoli celltakes on a cuboidal shape. Probably one of the most dramatic findings of microtubuledisruption is that of the changes in squirrel Sertoli cell morphology duringintroduction 1 03spermatogenesis. The extensive sloughing is not seen in the squirrel seminiferousepithelium treated with colchicine. In untreated squirrel testis, large caches ofendoplasmic reticulum are produced at the base of Sertoli cells and delivered to thevicinity of the spermatid head as spermatids elongate, become aligned with the long axisof Sertoli cells and move toward the lumen (yogI, 1983a). During this time, theresidual cytoplasm that has been segregated from the rest of the spermatid, is extendedinto the tubule lumen. It is then drawn basally, possibly with the help of Sertoli cellpenetrating processes, to be phagocytosed by Sertoli cells. Concurrent withspermiation, the mass of SER that is situated on the concave side of the spermatid headreturns to its basal position. When squirrel Sertoli cell microtubules are disruptedwith colchicine, the SER that has not ascended to the apex of the cell fails to do so (yogI etal., 1983b). Equally, the SER that has ascended to that position does not return. Theresidual cytoplasm is not extended into the tubule lumen but remains at the apicalsurface of the Sertoli cell. These events suggest that intracellular transport of ER isunable to occur in the absence of microtubules. It also suggests that cell shape is notmaintained, with residual cytoplasm no longer suspended into the tubule lumen.Probably one of the most interesting changes that occurs with the disruption ofSertoli cell microtubules, is the failure of spermatids to be moved toward the lumen. Insquirrel testis, treated with colchicine, elongate spermatids remain, in their crypts, inthe basal portion of the Sertoli cell, failing to be translocated to the tubule lumen (yogiet al., 1983b). Similarly, spermatids fail to be transiocated to the tubule lumen instage VI following intratesticular injections of taxol, in rat (Russell et al., 1989b).These findings support a proposed model in which spermatids are translocated across theepithelium by a mechanism of microtubule-based transport (yogi, 1989; yogI et al.,1983a, 1991a,b, 1992; Redenbach and yogI, 1991; Redenbach et al, 1991).Introduction 1 04DEVELOPMENT OF THE HYPOTHESISSertoli cell microtubules are suspected to be involved with numerous motilityrelated events, during spermatogenesis, such as redistribution of Sertoli cell organellesand translocation of spermatids in the seminiferous epithelium (Amlani and yogI, 1988;Russell, 1977a,b; Fawcett, 1975, yogI, 1989). The abundance of microtubules andMAPs in Sertoli cells lends credence to the notion that there may be a special role formicrotubules in Sertoli cells, beyond the housekeeping functions of a secretory cell andthe maintenance of cell shape. It is reasonable to assume that the abundant, highlyorganized and stage specific distribution of microtubules in Sertoli cells may signifythat, in addition to supporting the many microtubule-dependent functions common toother cells, Sertoli cell microtubules participate in functions unique to Sertoli cells.In rats, spermatids are translocated toward the apex of Sertoli cells, interruptedat least once by a basal excursion, before finally achieving the apical position fromwhich they are released. ESs are junctional complexes that are part of an adhesionjunction, assemble prior to the positioning of spermatids, and persist until immediatelyprior to spermatid release. ESs line Sertoli cell crypts in which spermatids aretranslocated (see yogI, 1989). The ER of the ES (ESER) is linked to a highly orderedlayer of actin filaments which is linked to the Sertoli cell plasma membrane. Thismembrane is in turn adherent to the spermatid. The ES-spermatid unit remains intactwhen spermatids are mechanically separated from Sertoli cells (Franke et al., 1978;Grove and yogI, 1989; Romrell and Ross, 1979). Through this linkage, forces appliedto ESs would be transmitted to spermatids. Parallel arrays of microtubules occuradjacent to the cytoplasmic face of the ESER. The structural components are present fora microtubule-based transport system to position and transport spermatogenic cells inthe Sertoli cell cytoplasm. Spermatid positioning has long been attributed to Sertolicells.An hypothesis, proposed by Christensen (1965) and later by Fawcett (1975),held that the morphology of Sertoli cell microtubules implied an involvement in theIntroduction 1 05movement of spermatids within the seminiferous epithelium. This notion was furtherextended by Russell (1977a), who suggested that spermatid translocation may be amicrotubule-based transport event. Reasoning that if microtubules can bind tomembrane bounded organelles, it follows that they may interact in the same way with theESER to transport spermatids along microtubule tracks. The hypothesis has been furthermodified to include the movement of spermatids on microtubule tracks bymechanoenzymes (yogI, 1989; VogI et al, 1989; Redenbach and VogI, 1991).Fig. 1-12 is a proposed model of microtubule-based spermatid translocationmodel to explain movement of spermatids within the seminiferous epithelium. Aspermatid is shown having been moved deep within a Sertoli cell crypt. Just deep to theSertoli cell membrane, the crypt is lined by an ectoplasmic specializations consisting ofactin filaments flanked on their cytoplasmic side by a cistern of endoplasmic reticulum(ESER). Microtubules occur adjacent to the ESER. Mechanoenzymes are illustrated as abridge between the microtubules and ESER. In this model, it is proposed thatmechanoenzymes may act as motors to move the spermatid-ES complexes alongmicrotubule tracks toward the base of the epithelium and then back to the tubule lumen.This model is supported by a number of pieces of evidence. First, microtubulesare oriented parallel to the direction of spermatid translocation (Christensen, 1965;Fawcett, 1987; yogI et al., 1983; Amlani and yogI, 1988; yogI, 1988). Second,microtubules occur adjacent to the ESER (for reviews see yogI, 1 988; VogI et al.,1991a,b; 1992). Third, linkages have been reported between microtubules and theESER (Russell, 1977b). Fourth, cytoplasmic dynein is present in high concentration intestis (Neely and Boekelheide, 1988; Collins and Vallee, 1989) and has been isolatedfrom Sertoli cell enriched preparations (Neely and Boekelheide, 1988), suggesting anactive role for microtubule based motility in Sertoli cells. Fifth, ESs remain attached tospermatids when the seminiferous epithelium is mechanically disrupted, indicating thatIntroduction 1 06Figure. 1-12: Microtubule-based translocation model. This is a model toexplain the movements of spermatids within the seminiferous epithelium. A spermatidis depicted as having been moved toward the base of the epithelium, deep within a Sertolicell crypt (left), and then returned toward the lumen of the epithelium (right) forrelease. The crypt is lined by an ectoplasmic specialization (ES). The ER of the ES(ESER) is linked through the actin network to the Sertoli cell membrane, and across theadhesion junction to the spermatid. This linkage remains intact during mechanicaldisruption of the seminiferous epithelium Forces applied to the ESER could betransmitted through this linkage to the spermatid. Microtubules occur adjacent to theESER. It is suggested that the ESER acts as a vehicle that is moved along microtubuletracks, resulting in movement of spermatids through the seminiferous epithelium.Mechanoenzymes are illustrated as a bridge between the microtubules and ESER. It isproposed that mechanoenzymes may act as motors to move the spermatid-ES complexalong microtubule tracks, toward the lumen of the seminiferous epithelium. Arequirement of this model is that microtubules are able to bind to the ESs.Sertoli Cell-MicrotubulesESERMicrofilaments/Mechanoenzymes 1Spermatid%07Introduction 1 08there is a tight linkage from the ESER, transmitted through the actin network of the ESto the Sertoli-spermatid junction, providing a means of directly linking events at themicrotubule ESER interface with spermatids. The direction of this proposed transportwould necessarily be a function of available mechanoenzymes and the polarity of Sertolicell microtubules. Taken together, this evidence supports the notion that spermatidtranslocation is a microtubule based-event.HYPOTHESIS STATEMENTHYPOTHESIS:Elongate spermatids are oriented and positioned in theseminiferous epithelium by means of a microtubule-basedtransport mechanism, through their attachment toectoplasmic specializations, which serve as vehicles, beingmoved along Sertoli cell microtubule tracks, powered bySertoli cell mechanoenzymes.PREDICTIONS1) If spermatids are oriented and positioned on microtubules, they should, throughtheir linkage to ESs, be able to bind to microtubules.2) If spermatid-ES binding is mediated by a mechanoenzyme or other dynamicallyregulated linking protein, the binding will share some characteristics withmicrotubule-organelle binding observed in other systems.Sertoli cell microtubules share microtubule distribution with other polarizedepithelial cells, in that they are oriented parallel with the long axis of the cell. As is thecase in other cells, they are likely to be unipolar in orientation. Furthermore, they mayIntroduction 1 09share the minus-end out polarity that has been demonstrated in polarized MDCK cells.This would be consistent with the presence of an abundance of cytoplasmic dynein, aknown minus-end directed mechanoenzyme. As part of the study to test the microtubulebased spermatid translocation hypothesis, the polarity of Sertoli cell microtubules wasdetermined. This is the focus of Chapter 2: Microtubule Polarity.Spermatid-ES-microtubule binding is required for microtubule-basedspermatid translocation. The development of a binding assay to test for spermatid-ESmicrotubule binding and its characterization are the focus of Chapter 3: BindingAssay.EXPERIMENTAL OUTLINEMicrotubule polarity studyIn this study, a morphological technique was used to determine the polarity ofSertoli cell microtubules, in which blocks of rat testis were treated with lysisdecoration buffer, with and without exogenous tubulin, adapted from the method ofHeidemann and Mcintosh (1980; Heidemann, 1991). Polarity was assessed bydetermining the direction of hook formation on Sertoli cell microtubules, using montagesconstructed from photographs of thin sections through the seminiferous epithelium,parallel to the base of the epithelium. Microtubules are decorated with clockwise hookswhen observed from their plus-ends to their minus-ends. Section orientation wasverified by the direction of axonemal dynein arms in spermatid axonemes, beingclockwise when viewed from the tip to the base of the axoneme.Microtubule-spermatid-ES binding studyAn in vitro assay was developed to test the potential for spermatids, throughtheir attachment to Sertoli cell ESs, to bind to microtubules. Microtubules wereassembled in the presence of [3H] labelled GTP and taxol to provide radiolabelled, stableIntroduction 11 0microtubules. Spermatid-ESs complexes were isolated, by a squash technique, from ratseminiferous tubules, and incubated with the labelled microtubules.Binding was characterized using pharmacological agents and nucleotides that havebeen used to characterize microtubule binding in other systems.The results of this study have implications for the Sertoli cell and its role inspermatogenesis and, in a broader sense, for cellular microtubules in general and theirrole in microtubule dependent organelle positioning. Furthermore, they imply that anintracellular activity, microtubule-based transport, can extend its influence to anextracellular event.Microtubule polarity 111CHAPTER 2MICROTUBULE POLARITYMicrotubule polarity 11 2INTRODUCTIONDETERMINATION OF MICROTUBULE POLARITY: THE APPROACHTo determine the polarity of microtubules in Sertoli cells, a morphologicaltechnique, introduced by Heidemann and Mcintosh (1980), was adapted for use in testis.This technique involves the use of a lysis/decoration buffer to lyse cells and decorateendogenous microtubules, usually with the addition of exogenous tubulin. The buffercomponents, essential to induce the assembly of curved protofilament sheets on existingmicrotubules, are a high salt buffer, preferably 0.5 M PIPES, and DMSO. The additionof purified tubulin contributes to the soluble tubulin pool that provides tubulinsubstrate for hook assembly. A number of modifications of the original technique havebeen reported in the literature, usually invoked because of buffer penetrationconstraints, specific to the tissue being studied. The technique has been usedsuccessfully without exogenous tubulin in tissue with abundant microtubules (Trouttand Burnside, 1988a,b).The technique is a valid measure of microtubule polarity (Heidemann andMcIntosh, 1980). This was demonstrated by determining the direction of tubulin hookson microtubules grown from basal bodies for which the polarity was known (Heidemannand McIntosh). The handedness of the hook formation identified microtubule polarityaccurately in at least 90% of microtubules examined. This accuracy has been consistentin subsequent use of this method. The assumption was made that the remainingmicrotubules, decorated in an opposite direction to their known polarity, were decorated‘erroneously’. One alternate interpretation is that exogenous microtubules may haveformed from exogenous tubulin, added with the decoration buffer. Because themicrotubules of ‘erroneous’ polarity are generally oriented parallel to the others, theformer explanation is preferred. In the literature, microtubule systems with apolarity of 90 % have been interpreted as “unipolar” (Troutt and Burnside, 1988a,b; Mogensen et al., 1989; Burton, 1988; Baas et al, 1987, 1988; McIntosh and113Euteneuer; McNiven et al., 1984; Euteneuer and Mcintosh, 1981; Burton and Paige,1981).DETERMINATION OF MICROTUBULE POLARITY: THE STRATEGYThe degree of hook formation can be varied to some degree, by altering the bufferand incubation conditions, or by changing the concentration of tubulin added to the lysisdecoration treatment (Heidemann and Euteneuer, 1982; Heidemann, 1990. However, inmany tissues, the rate and depth of penetration of the treating buffer will vary, within,as well as between, samples. Preliminary experiments showed this to be true for testis.For this reason, a number of different buffer conditions, as well as tissue from twospecies, were used in this study to ensure that buffer conditions did not play a role inthese observations. In view of the fact that the microtubules are abundant in Sertolicells, the use of hook decoration buffer, without exogenous tubulin, was also employed.The majority of Sertoli cell microtubules arise from a supranuclear location andcourse parallel to the long axis of the cell, extending into the apical stalk, Of particularinterest in this study are microtubules found in the supranuclear portion of the Sertolicell as well as in the vicinity of differentiating spermatids with associated ESs. Sectionswere taken from a supranuclear position, in areas which included spermatids. In thislocation, spermatid tails were frequently included in the section. Microtubules wereparallel to each other in these areas.Purified, polymerization competent, bovine brain tubulin was the source ofexogenous tubulin for this study. There was a concern that microtubules wouldpolymerize from the exogenous tubulin before gaining access to the cytoplasmicmicrotubules. For this reason, the concentration of exogenous tubulin was kept low;however, the assembly promoting nature of the buffer and the presence of endogenoustubulin could alter that threshold, particularly at higher temperatures. For this reasona range of temperature and incubation time combinations were tried, both with andwithout exogenous tubulin.Microtubule polarity 11 4In these experiments, care was taken to follow the ‘sidedness’ of the sections,ensuring that the assumed direction from which the tissue was viewed was correct.Dynein arms on axoneme microtubules are always oriented in a clockwise direction,when viewed from their base to their tip (Bershadsky, 1988, chapter 2). The presenceof spermatid tail axonemes was used to provide verification of the section orientation, inmany sections.Taken together, these experiments consistently give the same answer to thequestion: What is the polarity of Sertoli cell microtubules?Microtubule polarity 11 5MATERIALS AND METHODSMATERIALSAnimalsTestes from six adult, Sprague Dawley rats (200 to 275 gms) and onereproductively active ground squirrel (Citellus. lateralis) (244 gms) were used in themicrotubule polarity study. The animals were maintained in the animal care facilities atthe University of B.C. Beef brains were acquired from Intercontinental Packers,Vancouver, B. C.Chemicals and suppliesUnless otherwise mentioned all chemicals used for the microtubule polaritystudies were from Sigma Chemical Co. (Springfield Missouri). Taxol was the generousgift of Dr. Matthew Suffness, National Cancer Institute USA. Chemicals for electronmicroscopy were from JB EM and embedding media from Polysciences (WarringtonPA.). SS34 rotor is from Sorvall. The other rotors are from Beckman.BuffersPEM: 100 mM PIPES, 1 mM MgCI2, 1 mM EGTA, pH 6.9. PEM4M or PEM8M:PEM with 4M or 8M glycerol.Lysis decoration buffer (0.5 M Pipes, 2.5% DMSO, 1 mM EGTA, 1 mM MgCI2,1% triton X100, 0.02% SDS (optional), 0.5% deoxycholate, 1 mM GTPwith or without exogenous tubulin). For detailed description of makinglysis decoration buffer, (see Heidemann, et al., 1980). In experimentswhere tubulin was dialyzed in 2X PEM and then mixed 1:1 with theremaining components of lysis/decoration buffer, these components were2X(0.4 M Pipes, 2.5% DMSO, 1% triton X100, 0.02% SDS (optional),0.5% deoxycholate, 1 mM GTP).116METHODSPreparation of purified tubulin:Microtubule protein, for the microtubule decoration experiments, was preparedfrom two beef brains, using a temperature dependent assembly-disassembly protocolmodified from that of Williams and Lee (1982). Two steer heads were obtained fromIntercontinental Packers, packed in ice and transported to the lab. Brains were removedand dissected free of vascular and connective tissue (yield 700g brain). The tissue wasplaced in PEM4M buffer (70% of original volume), homogenized in an Oster blenderfor 50 seconds, and centrifuged at 23,500g for 5 minutes at 4°C in an SS34 rotor(14,000 rpm). The pellet was discarded. The supernatant was centrifuged at 39,000gfor 30 minutes at 4°C in SS34 rotor. The supernatant was brought to 1 mg/mI GTP,incubated at 37°C for 35 minutes then centrifuged at 35°C, at 39,000g for 30 minutesin the SS34 rotor (18,000 rpm). The supernatant was discarded. The pellet wasresuspended in 15% of original homogenate volume in PEM4M, solubilized on ice for30-40 minutes then centrifuged at 39,000g for 30 minutes at 4°C. The supernatantwas again made 1 mg/mI GTP, incubated at 37°C for 30 mm, and then centrifuged at39,000g for 30 minutes at 35°C in an SS34 rotor. The pellet was resuspended on ice in10% original volume in PEM without glycerol, solubilized for 30 minutes withhomogenization, and cleared at 39,000g for 30 minutes at 4°C in SS34 rotor. Thesupernatant, containing 19 mg/mI protein (Biorad), was frozen and stored in liquidnitrogen (protein concentration 1 9 mg/mI).Microtubule protein was thawed, as needed, cleared with a 39000g spin at 4°Cfor 20 minutes, cycled one more time through one warm, and then one cold step asdescribed above, and purified by a phosphocellulose chromatography. A chromatographycolumn was packed with activated P-il (Whatman), equilibrated with PEM buffer (noglycerol) pH 6.9 and loaded with 2.5mIs of microtubule protein. Purified tubulin wasrecovered from column fractions, monitored for tubulin content by A280, and dialyzed117into PEM4M glycerol. It was stored in liquid nitrogen (final concentration 0.6 -0.7mg/mi). Tubulin purity was verified by SDS PAGE gel electrophoresis (Fig 2-1: d)using a 7% gel, based on the method of Laemmli (1970).Purified tubulin from the column did not appear to polymerize in a post columnwarm cycle; therefore, the polymerization competency of post column purified tubulinwas checked (see below). A pellet of polymerized tubulin was not visible in warm cycledpost column tubulin, at this concentration (0.6 to 0.7 mg/mI); therefore, verificationof polymerization competency was sought using three methods: 1) negative staining oftubulin incubated on sea urchin axonemes isolated and demembranated according to themethod of Stephens (1 986a) (Fig. 2-1 :b), 2) negative stain electron microscopy (Fig2-1 :c), and 3) video enhanced differential interference contrast microscopy (Fig 2-1:d) of purified tubulin in the presence of 20i.tM taxol at 37°C. Although selfpolymerization did not occur, taxol stabilization or sea urchin axoneme nucleationresulted in polymerization.For microtubule decoration experiments, aliquots of purified tubulin werethawed immediately prior to use, concentrated to one half original volume in anultrafree-PF filtration unit (Millipore), then dialyzed against 1000 volumes of 2XPEM buffer (100 mM PIPES, 1 mM EGTA and 1mM MgCl) buffer at 4°C for 2 hrs, andfinally brought to predialysis volume. Dialyzed tubulin in 2X PEM was diluted 1:1 with2X concentration of remaining components of lysis/decoration buffer, to give a finallysis/decoration buffer the same as that used for endogenous tubulin preparationsdescribed below, with final a tubulin concentration of 0.6 to 0.7 mg/mI. Proteinconcentration was determined after the method of Bradford (1976) using Bioradreagents.Hook decorationThe method for microtubule decoration was modified after the procedure ofHeidemann and Mcintosh (Heidemann and Mcintosh, 1980, Heidemann et al., 1980).Microtubule polarity 11 8Figure 2-1. Evidence of purity and polymerization capability of purifiedbovine brain tubulin.(a, b, c) Micrographs from the same image of taxol polymerized microtubules from thepurified bovine brain tubulin used in microtubule polarity experiments. This is theappearance of the microtubules (a) viewed with DIG video-enhanced microscopy , (b)further contrast enhanced by computer, using expansion of assigned grey levels, and (c)‘cleaned’ by computer subtraction of out of focus background image.(d) SDS PAGE gel (7%) loaded with 3X cycled bovine brain microtubule protein (lane1) before purification on phosphocellulose column, (lane 2) purified tubulin afterMAPs were removed by chromatography, and (lane 3) molecular weight markers(ovalbumin 45kD, bovine plasma albumin 66kD, rabbit muscle phosphorylase b 97.4kD, beta-galactosidase 116 kD, and rabbit muscle myosin 205 kD). (Direction ofmigration on gel is from top to bottom). Prominent bands of alpha and beta tubulin (>)are seen (approximately 55kD) (lanes 1 and 2) with minor bands of microtubuleassociated proteins in microtubule protein (lane 1) which were subsequently removedby chromatography (lane 2).(e) Negative stain EM image of purified tubulin polymerized in the presence of 20jiMtaxol.(f) Video enhanced, DIG image of purified tubulin polymerized in the presence of 20 iMtaxol, from the same sample as in (e).(g) Isolated and demembranated sea urchin sperm axonemes incubated with and without(inset: center) exogenous bovine brain tubulin to illustrate polymerization capability ofpurified tubulin is shown using negative stain EM. A high power view shows themicrotubule protofilament detail of axoneme nucleated microtubule growth in areamarked with an astrisk (inset upper right). This also marks the end of the axoneme thatsupported longer microtubule growth (plus-end).Its©5K21 30C4Microtubule polarity 1 2 1Animals were anesthetized with halothane and the testes removed, decapsulated, placed inlysis/decoration buffer (with or without exogenous tubulin), and quickly cut into 2 mmcubes. Samples were incubated in lysis/decoration buffer:1) with exogenous tubulin, for 30 minutes on ice (4°C),10 minutes at 22°C and15 minutes at 37°C, (rat testis): control la - without exogenous tubulin, for 30minutes on ice, 10 minutes at 22°C andl5 minutes at 37°C. or control lb - withoutexogenous tubulin (and omitting the cold step) incubated for 40 minutes at 22°C and 15minutes at 37°C (rat),2) without exogenous tubulin for 10 minutes on ice, followed by 20 minutes at33°C (rat and squirrel testis). control 2 - (omitting the cold step) - 30 minutes at33°C (rat and squirrel testis). Number 2 gave the best results.This technique has been used widely to examine microtubule polarity (Heidemannand McIntosh, 1980) with, and without, the addition of exogenous tubulin using anumber of different incubation conditions, including warm and cold preincubationtemperatures. The rationale here for using an initial cold step, was to permit theexogenous tubulin to penetrate the tissue in the depolymerized form, favoring itscontribution to hook formation rather than self assembly. However, using theseconditions, microtubules may depolymerize, resulting in the examination of anexclusively cold stable population, or depolymerize and repolymerize in a mannerdifferent to what was present before the incubation. To address this concern, regardingthe effect of the cold step on endogenous microtubules, additional experiments werecarried out in which purified tubulin was added without the initial cold incubation. Inthese experiments, tissue was excised from the animal and immediately incubated inlysis decoration buffer with exogenous tubulin for 15 minutes at 22°C followed by 15minutes at 33°C. If self assembly occurred, it was anticipated that it would occur withrandom polarity. In the earlier experiments, controls had included the use of 33°C butin the absence of exogenous tubulin; therefore, they showed limited decoration, albeitconsistent with experimental findings. In order to determine the effect of incubationMicrotubule polarity 1 22entirely at 35°C with added tubulin, a further experiment was done in which the tissuewas immediately incubated in lysis buffer at 35°C for 30 minutes with exogenoustubulin. To reduce self polymerization of exogenous tubulin before tubulin gained accessto the cell, the tubulin in 2X PEM buffer was quickly warmed to 35°C in 2X PEM bufferand added to the remaining components of lysis decoration buffer at 35°C, immediatelybefore tissue incubation.Electron microscopyAll EM fixation and processing steps were carried out at room temperature.Tissue blocks were fixed for electron microscopy in 3% glutaraldehyde in PEM for 45minutes, with or without 1.0% tannic acid, rinsed 3 times, and postfixed in bufferedosmium tetroxide for 30 to 45 minutes. (Adequate fixation was also obtained using 2%glutaraldehyde for 30 minutes). Samples were rinsed 3 times in PEM and then stained“en bloc”, in aqueous 1% uranyl acetate for 1 hr, dehydrated and embedded in Polybed812 (Polysciences, Inc., Warrington, PA). Sections were cut and collected on 300 meshcopper grids. Sectioning strategy is illustrated in Fig. 2-2. Sections were stained withuranyl acetate and lead citrate for 8 minutes each, and viewed on a Phillips 300transmission EM at 80kV.In order to be certain of Sertoli cell orientation, the sampling strategy was asfollows: Blocks were positioned so that the seminiferous tubule to be sectioned wasoriented parallel to, and just deep to, the block face. (Fig. 2-2). They were trimmedwith enantomeric shaped mezas to eliminate errors in section orientation from the blockto final photographs. Seminiferous tubules were sampled serially from the base to theapex of the epithelium. A series of sections were cut, consisting of semi-thin sectionsfor light microscopy, interspersed between thin sections for electron microscopy,proceeding from the base to the apex of the epithelium. Semi-thin sections were stainedwith toludine blue to establish the orientation and approximate tissue depth of thecorresponding intervening thin EM sections. Areas for observation were selected fromMicrotubule polarity 1 23Figure 2-2. Microtubule decoration: sampling method for EM sections.Strategy used to orientate the seminiferous tubule such that Sertoli cells could besectioned perpendicular to their long axes. The morphology of the seminiferous tubulehas been simplified to show a simple columnar epithelial tube of Sertoli cells eachsupporting a single spermatid, which in turn projects into the tubule lumen. Assectioning progresses toward the apical surface of Sertoli cells, the spermatid isincluded in the section. By selecting cells from a central strip of the section (shadedarea), structures, which are orientated with their long axes parallel to that of theSertoli cell, such as spermatid tails or microtubules, are viewed in cross section(depicted as circles in the section).Rough TrimmedBlockFinal TrimmedBlock12.4Ion Grid,74oShadin9lndatesMicrotubule polarity 1 25the central strip of the seminiferous tubule sections where the knife passed at rightangles to the long axis of Sertoli cells.To provide further verification of the tissue orientation, advantage was taken ofthe inherent polarity of sperm tail axonemes, which extend their distal ends into thelumen of the seminiferous tubule. When a spermatid axoneme is viewed from the base tothe tip, its microtubules are being viewed looking toward their plus end (Heidemann andMcIntosh, 1980). In this orientation, dynein arms extend from the A toward the Bsubfiber, and are thereby pointing in a clockwise direction. It was verified that, whenobserved from the base to the apex of the seminiferous epithelium, dynein arms of allspermatid axonemes extended in a clockwise direction providing an internal check on theorientation of the tissue, in most sections.Negative stain electron microscopyDroplets of axoneme nucleated or taxol stabilized microtubules from purifiedtubulin, were placed on copper grids, that had been coated with parlodian, coated withcarbon, next dipped in acetone to remove parlodian, and finally glow discharged. Tissuewas allowed to settle; then blotted gently. A few drops of aqueous 1% uranyl acetatewere applied, rinsed carefully with 2 drops of distilled water, and allowed to dry. Theywere then visualized on a Phillips 300 at 60-80 kV.Video enhanced differential interference microscopyTubulin polymerization capacity was checked with video enhanced differentialinterference contrast microscopy. Microtubules, polymerized from the purified tubulinwith 20j.tM taxol, were placed in a tissue chamber as described by Schnapp (1980).Briefly, this consisted of placing sample within a partial circle of vacuum grease appliedto a polylysine coated acid cleaned slide (or coverslip fixed to a specially tooled adaptor).#0 spacers were placed on either side of the grease, and all covered with #0 coverslip.A drop of sample was applied with its meniscus rising just slightly above the level ofthe grease. The sides were sealed with Varlap (1:1:1 vaseline/lanolin/paraffin) leavingMicrotubule polarity 1 26access at the front and back to exchange solutions. These techniques reduced the tendencyof the sample to stream. Samples were observed, under oil, using a ZeissPhotomicroscope I, fitted with a 200/4 mercury light source, heat filters, and DICoptics. The image was captured by Dage video camera with manual gain and backgroundgrey level control. The camera was fitted with on Zeiss zoom attachment fitted with ai OX eyepiece, and the image displayed on a Sony monitor screen. Images were recordedon a Sony 3/4 inch video cassette recorder and photographed with a Pentax cameramounted on a tripod to allow greater than 1 second exposures to fuse the video image.(video microscopy: Allen et al., 1981a; lnoué, 1981,1986)Microtubule polarity 1 27RESULTSMICROTUBULE POLARITYEffects of lysis decoration bufferSertoli cells were examined at a level where microtubules are associated withESs surrounding spermatids in order to determine the polarity of microtubules inlocations where they may be involved in orientation and translocation of germ cells. Inmost cases, this included the presence of sperm tails of more deeply placed spermatids.Figure 2-3 shows an area of seminiferous epithelium that has been treated withlysis/decoration buffer and sectioned as shown in Fig. 2-2. A number of sperm tailaxonemes are included in the section. This section is oriented such that axonemes areviewed looking toward the tubule lumen. Dynein arms, projecting from A subfibers, areoriented in a clockwise direction.Figures 2-4 to 2-8 are electron micrographs from four different samples,incubated in cold then warm lysis decoration buffer, illustrating the consistency of theresults in spite of the variability of tissue appearance encountered. The degree ofmembrane extraction, nuclear decondensation, and germ cell collapse that varied withdepth and location of the seminiferous tubule within the tissue block can be seen.In the seminiferous epithelium there are complex relationships between Sertolicells and germ cells. In more highly extracted tissue, there was little to mark theseelaborate cell boundaries (Fig. 2-4). It was found that, incubation in lysis/decorationbuffer produced some decondensation of chromatin, as has been reported previously(Euteneuer and McIntosh, 1980), further reducing tissue boundary indicators(compare Figs. 2-4, and 2-6). However, in many areas, selective shrinkage of germcells was observed (Figs. 2-7 to, 2-9) permitting observations to be limitedexclusively to Sertoli cell microtubules, in those samples. In other samples, manchettesMicrotubule polarity 1 28Figure 2-3. Verification of spermatid tail axoneme orientation.(a) An area of seminiferous epithelium, which has been incubated in lysis/decorationbuffer, in which the axonemes of a number of spermatid tails can be seen in crosssection. This section is viewed as though looking toward the distal end of the axoneme.The orientation of dynein arms extending from A subfibers is in a clockwise direction.Included in the section are a number of microtubules decorated with hooks oriented(thick arrow) in a clockwise direction.(b) Micrograph showing greater detail of a spermatid axoneme indicated by an asteriskin (a). Note dynein arms (arrows) which project from A subfibers (arrowheads) areoriented in a clockwise direction.bars: a = him, b = polarity 1 30Figure 2-4. Rat seminiferous epithelium, highly extracted followingincubation in lysis/decoration buffer.(a) Section of seminiferous epithelium, oriented as though looking from the base towardthe lumen of the epithelium. The spermatid nucleus has undergone decondensation andmembranes are fragmented. Actin filaments of an ES are seen adjacent to the spermatid(arrows). Decorated microtubules can be seen with hooks oriented in a clockwisedirection. Tangentially sectioned microtubules are seen in this section (arrowhead).(b) Enlargement of decorated microtubule cluster and remnant of adjacent ES indicatedin (a).(c) Axoneme indicated in (a). Note A subfibers (arrowheads) and dynein arms(arrows) oriented in a clockwise direction.(d) Hookdecorated microtubules indicated in (a).bars: a = 1 jim, b = 0.1 jim.© ©Isb. N\,132/Figure 2-5. Low power view of an area of seminiferous epithelium thathas been incubated in lysis decoration buffer. Groups of decorated microtubulesare seen between elongating spermatids (arrows). This is a typical view in which theorientation of microtubules (arrows) can be compared to the associated sperm tails(arrowhead). The area outlined with a dotted line, in this figure, is shown at highermagnification in following = lj.tm4134 NFigure 2-6. Rat seminiferous epithelium moderately extracted followingincubation in lysis decoration buffer.(b) Micrograph of Sertoli cell microtubule decoration. Membranes are disrupted, butmore intact than those seen in Fig. 2-3. A spermatid nucleus and adjacent acrosome areindicated. An ES (arrows) and nearby decorated microtubules (arrowheads) can be seen.Manchette microtubules do not show hook formation. Sertoli cell microtubule hooks areoriented in a clockwise direction.(a) Axoneme from the same section as shown above, showing clockwise orientation ofdynein arms (arrows) in same field as clockwise decoration of Sertoli cell microtubules.bars: a = 1 jim, b = 0.25 jim.‘4P/I.-’..Microtubule polarity 1 36were easily distinguished from cytoplasmic microtubules, providing an additionalboundary marker (Fig 2-5).Decorated microtubules must be cut in precise cross section for hook formationto be clear. In all fields, some microtubules were found that were slightly tangential tothe plane of the knife, appearing as short striated bundles (Fig 2-6). Although thetangential microtubules were decorated, as indicated by their striated appearance, theycould not be scored for polarity. Because there was no pattern to the distribution oftangential microtubules, it is presumed that those that could be scored wererepresentative of all those present.Sampling for montaaesAreas showing different degrees of extraction were chosen for sampling. To becertain that the results were representative of all microtubules in the area, large fieldsof epithelium were sampled. These were recorded by sampling either the entire area of a300 mesh grid square or strips across an entire grid, photographing them at 18,800Xon Agfa plate film. Montages 3, 4 and 5 are derived from this method of sampling. The50 to 80 negatives for each field were printed at final magnification of approximately50,000X. Resulting micrographs were then spliced together into working montages andscored with the aid of a hand held magnifying glass. Figure 2-8 is a reduction of montage3. In addition, small montages were photographed and assembled from areas moreenriched with microtubules (montages 1 and 2).Scoring criteriaTo determine the polarity of microtubules from each montage, microtubules wereclassified according to the following criteria. Microtubules were grouped into“clockwise”, “counterclockwise” or “ambiguous” categories. These were furthersubdivided into “hook” or “cluster” groups. In all microtubules scored in the “hook”category, hooks arose from a single microtubule (first order hooks). Hooks on hookswere not counted, nor were closed hooks. In those few cases where first order hooks ofMicrotubule polarity 1 37Figure 2-7. Rat seminiferous epithelium only slightly extractedfollowing incubation in lysis/decoration buffer.(a) Micrograph of seminiferous epithelium, orientated as though looking toward theepithelial lumen. In this sample, decondensation of the spermatid nucleus is minimal.Microtubule decoration of Sertoli cell microtubules is mostly with hooks orientated in aclockwise direction (small arrowheads). A microtubule with counter-clockwiseorientation is indicated (arrow). ES is indicated with large arrowhead.(b) An axoneme from the same section shows clockwise orientation of dynein arms.bars: a= 1 jim, b = O.25jim0I a,—Microtubule polarity 1 39Fig. 2-8. Montage of rat seminiferous epithelium incubated inlysis/decoration buffer. Section of rat seminiferous epithelium, incubated inlysis/decoration buffer, viewed as though looking toward the tubule lumen.(a) Reduced photograph of a working montage assembled from micrographs of an entiresquare of 300-mesh EM grid. Selective shrinkage of germ cells has occurred.(b) An enlargement of the area indicated, showing microtubules oriented in a clockwisedirection.(c) Axoneme from the area indicated on the montage, also showing clockwise directeddynein arms.(d) Axoneme from an adjacent grid square. The clockwise orientation of dynein arms canbe seen (arrows).bars: a = lOjim, b = 0.25 tim, c = 0.liim./Microtubule polarity 1 4 1both orientations were present, microtubules were scored as “ambiguous”. In order notto bias the scoring by excluding the heavily decorated microtubules, those with onlysecond order hooks were assigned to a separate category of “clusters”. In no case weremore than 10% of the secondary hooks on clusters of a reverse orientation and thereforenone were considered ambiguous.Microtubule polarity: observationsIn samples that included cold incubation (on ice), virtually all cytoplasmicmicrotubules were decorated. In contrast, the easily distinguishable manchettemicrotubules were rarely decorated and axonemes were never observed to be decoratedwith tubulin.In samples with added tubulin, initially incubated on ice, microtubule decorationoften consisted of elaborate clusters. In those without added tubulin, initially incubatedon ice, decoration was less complex; however, the total number of microtubules was low.In samples not incubated on ice, without added tubulin, the total number of microtubuleswas higher, but the number that were decorated was low. In this group, although thosethat were decorated seldom possessed more than one hook, their orientation was the sameas those with cold incubation. Under all incubation conditions, regardless of themicrotubule density or complexity of hook formation (Fig. 4), the majority of decoratedmicrotubules possessed hooks curving in the same direction, indicating uniformpolarity. Similarly, variability in the extent and pattern of hook decoration has beenreported in other studies (Heidemann et al., 1981).Microtubule polarity: countsTable Il-I shows the scoring of microtubules in each montage assembled from ratseminiferous epithelium (montages 1-4). Ground squirrel Sertoli cells were alsoexamined and decorated microtubule counts from a representative montage ofseminiferous epithelium, from this animal, are included in Table II- I (montage 5).Microtubule polarity 1 42Tab. Il-I. Classification of microtubules according to hook decoration:cold/warm incubation.Clockwise Counterclockwise AmbiguousMont- Hooks Clusters Hooks Clusters Hookse1 2 3 4 5 1 2 31 3 0 1 1 0 9 0 0 0 0 02 1 5 2 5 0 5 0 1 0 0 13 8 3 3 0 1 21 0 0 0 0 04 71 34 14 3 5 1 16 6 2 0 35 72 164 1 0 0 9 1 0 0 0Orientation of decorated cytoplasmic microtubules in montages of seminiferousepithelium which was incubated in cold followed by warm lysis decoration buffer.Microtubules were viewed as though looking toward the apex of the Sertoli cell, scoredfor polarity, and classified according to the orientation and number of hooks formed.Values given indicate the number of microtubules decorated with clockwise,counterclockwise or ambiguous hook decoration. The maximum number of clockwise andcounterclockwise open hooks that were observed to originate directly from a singlemicrotubule, was 5 and 3 respectively, in these samples. Some microtubules wereobserved, in which no hooks arose directly from a single microtubule but formed anelaborate structure in which multiple open hooks of the same orientation arose from amicrotubule bearing only closed hooks. These were scored as “clusters”. Montages 1 to4 are from rat, and montage 5 is from ground squirrel seminiferous epithelium.Microtubule polarity 1 43Tab. Il-lI. Percent of microtubules with clockwise, counterclockwiseand ambiguous hook decoration: cold/warm incubation.Clockwise Counter clockwise Ambiguous M icrotubu lesMontage Total % Total % Total % Total1 14 1 00.0 0 0.0 0 0.0 142 18 90.0 1 5.0 1 5.0 203 36 100.0 0 0.0 0 0.0 364 128 83.0 24 15.0 3 2.0 155Average 93 . 3 5.0 1 .75 93 90.3 10 9.7 0 0.0 103The total and percent of microtubules decorated with hooks of clockwise,counterclockwise or ambiguous orientation for each montage scored in Table li-I. Theaverage number of microtubules with hooks oriented in a clockwise direction formontages 1 to 4 is 93.3%. Counts from the ground squirrel epithelium are indicated inmontage 5 for comparison.1 44Table Il-Il indicates the percentage of microtubules of each orientation, in eachmontage. These data indicate that between 83 and 100% of microtubules were orientedin the same direction, the average being 93.3%; that is, when viewed from the base toapex of the epithelium, an average of 93.3% of microtubules were oriented in aclockwise direction.In addition, axonemes from the same field as these microtubules, show clockwiseorientation of dynein arms. In order to pursue this point, axonemes were sought thatwere in clear cross section in, and surrounding, the area sampled for each montage andin many other samples observed. The orientation of dynein arms was found to be thesame as hooks on surrounding cytoplasmic microtubules, in each case (see Figs 2-3 to -2-8).In the experiments in which exogenous tubulin was added without an initialincubation on ice, a similar variation in tissue appearance was noted. Using the samecriteria as was used for in the cold incubation group (Tables li-Il and Il-Il), fourmontages from the experiments using warm lysis/decoration buffer with added tubulinwere scored for polarity (Tables Il-Ill and ll-IV). Microtubules were again found to beof uniform polarity and oriented with the minus-end directed toward the apex of the cell(Fig. 2-9).Using this microtubule decoration technique, quantitative differences in thedensity of microtubules and the complexity of their decoration were seen. Thesedifferences depend on the area of sampling, incubation temperatures, and on the presenceof exogenous compared with endogenous tubulin; however, in all cases, the qualitativefindings do not vary. That is, most microtubule hooks are of the same orientation,indicating a uniform polarity. When cytoplasmic microtubules are observed from thebase to the apical surface of the epithelium, hook decoration is in the clockwisedirection, indicating that the minus ends of the cytoplasmic microtubules are directedtoward the apex of the cell (Redenbach and yogI, 1991). Additional experiments inwhich tissue was incubated directly in 35°C lysis decoration buffer, with exogenousMicrotubule polarity 1 45tubulin, were found to have decorated microtubules in the extracellular space (Fig. 2-10), indicating that polymerization had occurred before tubulin was able to enterSertoli cells. However, the microtubules were found to be oriented in a random manner,both outside and in some areas within the cell membrane (Fig 2-10). They could bedistinguished from the highly ordered microtubules oriented parallel to one another andof uniform polarity. Furthermore, where the majority of microtubules were unipolarand oriented with the long axis of the cell, they were oriented, as had been found in allthe other samples, with their minus-ends directed toward the apex of the cell. In viewof the possible artifact of self polymerizing microtubules, no counts were made fromthis group.Microtubule polarity 1 46Fig. 2-9. Montage of rat seminiferous epithelium incubated in warmlysis/decoration buffer.(a) Micrograph of a portion of montage 9 (Tabs. Il-Ill, -IV); rat seminiferousepithelium that was incubated in warm lysis/decoration buffer in the presence ofexogenous tubulin. The field includes part of a lightly staining Sertoli cell which isflanked by darker staining germ cells (broad arrows). The tissue is viewed as thoughlooking toward the apical surface of the cell. The clockwise orientation of hooks oncytoplasmic microtubules (long arrows) indicates that, in Sertoli cells, microtubulesare oriented with their minus ends directed toward the apical surface of the cell.(b) Higher magnification of decorated microtubules from tissue shown above.bars: a = 0.5 lIm, b = 0.1 jim.:J4A••‘,-II.“‘.1Microtubule polarity 1 48Tab. Il-Ill. Classification of microtubules according to hook decoration:exogenous tubulin/warm incubation.Clockwise Counterclockwise AmbiguousMont- Hooks Clusters Hooks Clusters Hooksages1 2 3 4 5 1 2 36 25 13100 1 0 0 2 0 0 07 9 7 10 1 0 0 0 0 0 08 11 15 5 3 0 0 0 0 0 0 09 17 9 92 0 1 2 0 0 0 0Hook orientation of decorated cytoplasmic microtubules in montages of seminiferousepithelium which was incubated in warm lysis decoration buffer containing exogenoustubulin. As in Table Il-I, microtubules are viewed as though looking toward the apex ofthe Sertoli cell, scored for polarity, and classified according to the orientation andnumber of hooks formed. These data are from montages from rat seminiferousepithelium.Tab. Il-tV. Percent of microtubules with clockwise, counterclockwiseand ambiguous hook decoration: exogenous tubulin/warm incubation.Clockwise Counter clockwise Ambiguous M icrotubu lesMontage Total % Total % Total % Total6 49 95.0 2 5.0 0 0 517 18 100.0 0 0 0 0 188 34 100.0 0 0 0 0 349 38 95.0 2 5.0 0 0 40Average 97.5 2.5 0.0The total and percent of microtubules decorated with hooks of clockwise,counterclockwise or ambiguous orientation for each montage scored in Table Il-Ill.Microtubule polarity 1 49Figure 2-10: Microtubule decoration in seminiferous epitheliumincubated immediately in lysis decoration buffer with exogenous tubulinat 35°C. Micrographs of seminiferous epithelium incubated at 35°C in lysisdecoration buffer to which purified bovine brain tubulin was added immediately prior toincubation. There was evidence of polymerization and decoration of microtubules beforeentry to cells and of haphazard organization of microtubules within Sertoli cells, bothconsidered to be due to self polymerization of exogenous tubulin.(a)Base of Sertoli cell (broad arrow), extracellular matrix including collagen fibers(arrow heads), and myoid cells (asterisk). Decorated microtubules are seen in theextracellular matrix (short arrow), an artifactual location for microtubules.(b) Close up view of decorated microtubules shown in (a). Collagen fibers of theextracellular matrix are seen. The same decorated microtubules are identified as in (a).Collagen fiber is identified with an arrowhead.(c,d) Examples of haphazard orientation of microtubules found in Sertoli cells. Thesewere not seen in Sertoli cells incubated initially on ice or at 22°C.(e) Area of stalk of Sertoli cell, cut perpendicular to cell axis, observed from base toapex of the cell. In areas such as this, haphazardly organized microtubules are stillpresent, but the majority of microtubules are caught in cross section and oriented in aclockwise direction, as shown here.(f) Axoneme from same field as shown in (e).bars: a =, b= O.lp.m, c=, d= O.5iim , e= O.2im , f= 0‘1’Microtubule polarity 1 5 1DISCUSSIONThe specific findings of the microtubule polarity portion of this study aresummarized in this section and conclusions drawn from the data are discussed. Theimplications of these findings and their relevance to the issue of microtubule-basedspermatid translocation will be discussed more fully in chapter 4.MICROTUBULE POLARITY IN SERTOLI CELLSThe results of this study lead to the conclusion that Sertoli cell microtubules areunipolar. Combining the counts from all montages from rat testis, an average of 95.4 %percent of Sertoli cell microtubules were oriented in one direction. This is well withinthe accepted range of reliability of microtubule polarity indicated by hook formation onmicrotubules of known polarity, a standard used to assert unipolar orientation ofm Ic rotu bu le s.Sertoli cell microtubules are oriented with their minus-end directed toward theapical surface of the cell. Alternate L.M. and E.M. sections were used to monitor level ofsampling within the apical portion of Sertoli cells and precautions were taken to ensurethat section orientation was not lost during processing. The presence of sperm tailaxonemes was used to verify the orientation of tissue sections, in most cases. Whenviewed from the base toward the tip of the spermatid tail, axonemal dynein arms areoriented in a clockwise direction. From that same perspective, viewing from the base tothe apex of the epithelium, hooks on Sertoli cell microtubules form in a clockwisedirection. The coexistence of clockwise oriented dynein arms, on sperm tail axonemes,and clockwise hook formation, on cytoplasmic microtubules, in the same sections,confirms the finding that the polarity of microtubules are oriented with their minusends directed toward the apical surface of the cell.The use of a number of modifications of the lysis/decoration parameters indicatesthat methods do not play a role in the quality of results reported here. TheMicro tubule polarity 1 52lysis/decoration buffer was variably disruptive of the tissue, both within and betweensamples. Nuclear decondensation occurred in some samples, as has been observed byothers (Euteneuer and McIntosh, 1980). Because tissue preservation, as well as thecomplexity of hook decoration, varied with the depth of block sampled, it was clear thatbuffer penetration varied. In many samples, osmotic damage occurred in the form ofselective condensation of germ cell cytoplasm. However, the latter feature facilitatedcounting of Sertoli cell microtubules. Regardless of the extent of tissue extraction ormicrotubule density, microtubule orientation was the same. The microtubuleorientation in rat and squirrel testis were the same.Hook formation was minimal on manchette microtubules and not seen on any ofthe sperm tail axonemes. The factors that are responsible for the remarkable stabilityof manchette and sperm tail axoneme microtubules may alter their potential to becomedecorated with the curved protofilament sheets. Manchette and sperm tail axonememicrotubules have been reported to contain detyrosinated and acetylated tubulins, whileSertoli cell cytoplasmic microtubules have been shown to be comprised exclusively oftyrosinated tubulins (Hermo et al., 1991). Detyrosination and acetylation are posttranslational modifications that occur in tubulins comprising more stable structures,but appear to be coincidental with microtubule longevity, rather than a cause of it. Itmay also be that hook formation is hindered by manchette MAPs and axonemal MAPs andassociated structures.In selecting the parameters of time and temperature, there was a trade offbetween the time required for tissue penetration and polymerization of exogenoustubulin. To confront the possible interpretation that cold incubation conditions resultedin preferential selection of a cold resistant population of microtubules, or alteration inpolarity due to depolymerization and repolymerization, a number of preincubation andincubation temperature combinations were used. Initial preincubation conditions on iceor at 22°C were used to avoid self polymerization of exogenous tubulin, before it gainedaccess to the Sertoli cell. To get a more complete picture, some of the experiments were1 53carried out in which exogenous tubulin was omitted from the lysis/decoration buffer.This seemed reasonable because Sertoli cells have a very high concentration ofmicrotubules, in most areas, and quite likely an adequate soluble tubulin pooi, both ofwhich may provide a source of tubulin for hook formation. The density of decoratedmicrotubules was reduced following either with the use of cold preincubation or theomission of exogenous tubulin and were higher when preincubation was carried out at22°C. Regardless of the incubation conditions used, the polarity indicated by hookdecoration was the same.To more closely approximate in vivo temperatures, experiments were alsocarried out using immediate incubation at 35°C. It was considered that if selfpolymerization of exogenous tubulin occurred, it would be with random polarity.Microtubules of haphazard orientation were seen in a supranuclear location of Sertolicells when tissue was immediately incubated at 35°C, even though the tubulin was notadded until immediately before incubation. Randomly oriented microtubules are notobserved in the apical regions of Sertoli cells in control tissue. In addition, followingthe immediate warm incubation in lysis/decoration buffer with exogenous tubulin,decorated microtubules were observed in the interstitial space, decidedly an artifact.Counts were not made from tissue from experiments incubated immediately at 35°C.However, even in this group, microtubules from areas in which microtubules werelargely parallel showed microtubule hook formation, consistent with the other data.SUMMARYIn reviewing data from all of the experiments, the results consistently showedthat Sertoli cell microtubules are oriented with the minus end directed toward the apicalsurface of the cell. Figure 2-1 1 illustrates these findings showing a Sertoli cell, withspermatids embedded in Sertoli cell crypts, sectioned in the apical portion of the cell,viewed from the base to the apex of the cell. The coexistence of clockwise orientedaxoneme dynein arms and clockwise oriented cytoplasmic microtubule hooks isMicrotubule polarity 1 54illustrated. The resulting model provides for cytoplasmic microtubules oriented withtheir minus-end directed toward the apex of the cell.The implications of these findings are discussed in chapter 4.Microtubule polarity 1 55Figure 2-11: Summary diagram of Sertoli cell microtubule polarityobservations. This is a schematic summary of the results of hook formation followingmicrotubule decoration of Sertoli cell microtubules. No attempt is made to represent aspecific stage of the seminiferous epithelial cycle. 1. A Sertoli cell as depicted in themodel diagram sectioned to reveal a cross section of spermatids and surroundingmicrotubules. The resulting section (2.) shows the handedness of hook formation ondecorated microtubules and the direction of axonemal dynein arms on axonemes. Whenthe dynein arms are oriented in a clockwise direction, as shown, the section is beingviewed from the base to the apical surface of the Sertoli cell, and microtubules areoriented with their minus-ends directed toward the apical surface of the cell.Is’Sertoji Cell1.9. c.w zC) Iz G)mCl) -<011 58INTRODUCTIONThe purpose of this section is first, to describe some of the methods used for thebinding assay and second, to provide an outline for the approach used to answer thequestion: do microtubules bind to spermatid-ESs?.BINDING ASSAY CRITERIA: COMPONENTSSpermatid-ESsIt was necessary to develop a method for the isolation of spermatids, with their ESsconsistently intact, in sufficient numbers for a binding assay. It has been shown thatwhen spermatids are mechanically removed from the seminiferous epithelium, ESsremain intact and attached to isolated spermatids (Fig. 3:1 in Masri et al., 1987;Romrell and Ross, 1979; Fig 2 in yogi et al., 1986; Grove and yogI, 1989); that is, thespermatid with its membrane, the Sertoli cell membrane, the underlying actin network,and the ESER are stably linked and remain intact. This relationship is illustrated in Fig.3-1. ESER has been shown to remain attached although it is subject to osmotic damagewith some buffers. Isolation of Sertoli cell sheets in a PBS buffer (yogI and Soucy,1985), followed by pipetting to remove spermatids (Masri et al., 1987; yogI et al.,1986; Grove and yogI, 1989), used to investigate the molecular components of ESs,proved too vigorous for the ESER. The buffer did not protect the exposed ESER fromosmotic effects. A more gentle isolation method was developed to ensure retention of ESsand buffer components adjusted to protect the ESER.Following a method adapted from an earlier method of Perey and coworkers (1961)by Parverien and Ruokonen (1982), individual seminiferous tubules were isolated andstages identified. Figure 3-2 illustrates this method of stage identification. Squasheswere made from sections of the tubule to verify that spermatids could be expressed fromall stages of the tubules using this method. The method of individual tubule isolation wasthen used throughout the binding assay, but sampling was not stage specific.Binding assay 1 59Figure 3-1: Isolation of spermatids-ESs from seminiferous epitheliumwith intact ES5: This diagram shows a spermatid situated in a Sertoli cell crypt thatis lined with an ES. During spermatid isolation with a gentle squash technique, thespermatid is removed from the Sertoli cell carrying with it the ES: the Sertoli cellmembrane, the intact actin filament network, and the ESER.160Binding assay 1 6 1Figure 3-2: Transillumination methods of seminiferous cycle staging:This diagram shows the morphological criteria used to determine the stages in isolatedsingle seminiferous tubules (adapted from Parvinen and Ruokonen, 1982). This methodof identifying seminiferous epithelium stages in intact tubules allows for stage specificinvestigation of physiological and morphological events during spermatogenesis, and wasused in this study to verify that spermatids could be recovered from all stages ofspermatogenesis.I62I II— HI 1\b/—\/ I)(— XI XII XIII XI’I—I: .. • tll ill’ • • • • •_______• :•:..:- STRONG SPOT -- PALE - \WEA< SPOTDARK 2O4E1 63Labelled microtubulesIn preliminary studies, a ‘topical’ binding assay was developed. Isolated spermatidswere allowed to adhere to a glass slide and then subjected to a number of bindingconditions. The binding of exogenous microtubules and variety of MAP preparations tospermatids that had been adhered directly to glass slides was tested. Although somegeneral trends could be seen, results were difficult to interpret and findings could not bequantified. It became apparent that to properly characterize spermatid-ES-microtubulebinding, it was necessary to distinguish between the fate of added microtubules andendogenous microtubules that may have been retained either during isolation, from thetreatment condition, or by nonspecific binding to the slide. A marker was needed toquantify binding, in order to demonstrate consistently when binding occurred and tocharacterize binding, in the presence of pharmacological agents that may potentiallyinhibit or enhance microtubule binding. To quantify spermatid-ES-microtubule binding,the measurement of labelled exogenous microtubules provided a means of ‘ignoring’endogenous microtubules retained during isolation. Radiolabelled exogenousmicrotubules stabilized with taxol (3MT) were used to provide a quantitativecomparison of3MTx-spermatid-ES binding under selected binding conditions.BINDING ASSAY: ESTABLISHING THE CRITERIADevelopment of the binding assayAn assay was developed in which spermatid-ESs were combined with 3MTxs andallowed to bind. In order to confine the measurement to bound microtubules, unboundmicrotubules had to be removed prior to counting. In order to measure microtubulebinding to spermatid-ESs, after binding had occurred, it was necessary to separatespermatid-ESs from other cellular material that accompanied spermatid-ESs as theywere isolated from the seminiferous tubule during the squash technique. Both of theserequirements were addressed using centrifugation of the combined 3MT and spermatid1 64ESs, over a sucrose gradient, determining conditions that would enable the spermatidES, with their bound 3MTs(3MT - spermatid-ESs) to enter the gradient while 3MTsalone did not.Establishing that counts represented microtubulesIt was necessary to show that label was incorporated, stably, into the exogenousmicrotubules, and that the label that become bound to spermatid-ESs was microtubule inorigin. In this study, incorporation was measured by comparing counts before and afterthe removal of microtubules by centrifugation. It has been shown that microtubules,labelled by polymerization of purified tubulin in the presence of [3H] GTP and taxol,remain stable (Wilson et al., 1985). Stoichiometric quantities of GTP are not requiredif microtubule assembly occurs in the presence of taxol. Stability was measured bycomparing microtubule counts, pelleted from aliquots taken from the same stock, overtime. The strategy, used to determine the proportion of counts that could be attributed tomicrotubule binding, was to remove microtubules from an aliquot of 3MTx stock andcompare binding with and without microtubules.CHARACTERISTICS OF BINDINGReversal of binding.Mechanoenzymes have ATPase activity and are removed from microtubules withnucleotides, ATP and GTP breaking kinesin-microtubule rigor bonds and ATP breakingcytoplasmic dynein-microtubule rigor bonds. Some of the putative motors arereleasable with both GTP and ATP. Concentrations of ATP and GTP routinely used forrelease in MAP isolation, were used in this study.Experiments to characterize 3MIx and spermatid-ESs bindingIf binding could be demonstrated between 3MTx and spermatid-ESs, manipulation ofbinding, using conditions known to influence MAP-microtubule binding, could be used to1 65characterize that binding. While there is a fair amount of information available aboutthe ATPase activities and motility properties of the motor-microtubule site, less isknown about its binding properties and even less is known about the binding at the motororganelle site (Brady and Pfister, 1991). In the absence of specific agents to test thenature of motor-microtubule binding, aside from nucleotide specificity,pharmacological agents that affect microtubule based motility, and mechanoenzymeATPase activity were selected for these studies. Direct comparisons of binding, ATPaseactivity and motility are not valid. However, use of these agents provided someinformation as to the nature of the binding and a basis for comparison with otherorganelle-microtubule binding studies.The concentrations of pharmacological agents were selected from a consensus of thelevels that might distinguish between the two known families of motors. They included 2mM NEM, 1mM EHNA, 10 and 100 p,M vanadate, as well as 5 mM GTP and 10 mMMgATP. The effect of added Sertoli cell cytosol and rat testis MAPs were also tested.LOCALIZATION OF LABELMorphological suDport for results of binding assayIn order for counts, in gradient fractions, to be considered representative of boundmicrotubules entering the gradient, it was necessary to determine if the bindingoccurred at ES sites. In that microtubules would be labelled with[3H]-GTP,autoradiography of 3MT combined with spermatid-ES5, was used to localize 3MT -spermatid-ESs binding.To examine the potential binding of microtubules at ESs, in this portion of the study:1) A method was developed to isolate spermatids, with ESs attached, with a minimumof associated endogenous microtubules.2) Microtubules were assembled from purified bovine brain tubulin, in thepresence of[3HIGTP and stabilized with taxol to be used as labelled exogenous1 66microtubules. The incorporation of label, stability, and length of the microtubuleswere checked.3) Buffer conditions were defined to protect the ESER from osmotic effects andaddress the anticipated requirements for binding.4) Differential centrifugation conditions were determined that met the criteria formeasurement of exogenous microtubule binding to spermatid-ESs: enrichment forspermatid-ESs and separation of microtubules bound to spermatid-ESs from thosethat were not bound.5) The validity of considering counts entering the gradient as microtubule bindingwas checked by determining the percent of counts that could be attributed to themicrotubule component of 3MTx6) 3MTx - spermatid-ES binding was quantified by counting the radiolabelledexogenous microtubules present with spermatid-ESs in the gradient fraction:binding assays.7) The effect of nucleotides on 3MTX - spermatid-ES binding was examined for itspotential as a tool to reverse binding.8) 3MT - spermatid-ES binding was characterized by imposing selected conditionsincluding: the effect of varying substrate concentration; the possibility of turnoverof binding; the effect of adding Sertoli cell cytosoi or a cytoplasmic dynein enrichedMAP preparation; and the effect of a selection of known mechanoenzyme inhibitors(10 mM MgATP, 5 mM GTP ± ATP depletion, 2 mM NEM, 1 mM EHNA, 1 mMAMPPNP, 10 pM vanadate, and 100 jiM vanadate, and ATP depletion with 10units/mi hexokinase with D glucose).9) Label was localized by autoradiography, with and without the release of bindingusing 10 mM MgATP.1 67CALCIUM AS A PROPOSED REGULATORIt has been proposed that the ESER may provide regulation of local events byselective uptake and release of calcium (Franchi and Camatini, 1985). Results ofcalcium localization studies in mouse (Kierszenbaum et al., 1971) and guinea pig testis(Franchi and Camatini, 1985) were contradictory. Mechanoenzymes are highlysensitive to calcium dependent proteases, released during tissue disruption, andtherefore calcium is customarily removed from in vitro assays of cell motility by EGTA.Little is known about the effect of calcium on microtubule based transport. Microtubulebased transport occurs in the presence of EGTA. To test the initial assumption, thatcalcium is sequestered by the ESER, the methods used for localization of calcium inmouse (Kierszenbaum et al., 1971) and guinea pig testis (Franchi and Camatini, 1985)were repeated in rat testis. In addition a number of other methods were used.1 68MATERIALS AND METHODSMATERIALSAnimalsRat testes used for the binding assay studies were from male Sprague Dawleyrats, housed in animal care facilities at University of British Columbia or BrownUniversity (Charles River CD). To ensure that all stages of spermatid development werewell represented in tissue used in the binding assays, only mature rats of greater than250 gms were used (Ekwall et aL, 1984) for these experiments. Animals used forisolation of Sertoli cell cytosol and testis cytoplasmic dynein enriched MAP preparationswere 21 days of age. Animals used in the binding assays were anesthetized withpentobarbital or killed with CO2.Chemicals and reacentsGeneralUnless otherwise indicated, all chemicals and reagents were obtained from Sigmachemical Co. (St. Louis Mo.). 3H-GTP (NET-305 Guanosine 5’-triphosphate,tetrasodium salt, [8-5’-3H] was from NEN Research Products Wilmington DE.BuffersA4M: 0.1 M MES (2[N-morpholino]ethane sulfonic acid), 1.0 mM EGTA (bis(13-aminoethyl ether)-N,N1- etracetic acid), 0.5 mM MgCI2, 4.0 M glycerol, pH6.75. (plus protease inhibitors: 2 jig/mI leupeptin (4.2 jiM), 0.1 mg/mIsoybean trypsin inhibitor, and 0.2mM PMSF (phenyl-methylsulfonyl fluoride).1 69.E.EM: 0.1 M PIPES (piperazine-N, N-bis (2-ethane sulphonic acid), 1 mM MgCI2, 1mM EGTA, pH 6.9 (pIus protease inhibitors: 10 jig/mI leupeptin, 0.5 gig/mIpepstatin,10 jig/mi soybean trypsin inhibitor and 0.5mM PMSF)PEM 250: PEM with 250 mM sucrose (plus protease inhibitors leupeptin,0.5 jig/mI pepstatin, 10 jig/mI soybean trypsin inhibitor and 0.5 mM PMSF.Ea: phosphate buffered saline (0.15 M NaCI, 4.0 mM Na/K phosphate titrated to pH7.3), 5 mM KCI).ME: (0.1 M MES, 1 mM EGTA, 0.5 mM MgCI2, 0.1 mM EDTA(ethylenediaminetetracetic acid), 1 mM MgATP and 20 jiM taxol, pH 6.75)PIPES buffer -K:100 mM PIPES (titrated with KOH),1 mM EGTA, 1.0 M sucrose(or 250 mM if used in binding assay) plus protease inhibitors: 0.1 mg/mIsoybean trypsin inhibitor, 2 jig/mI leupeptin, 2 jig/mI pepstatinPrimary antibodies:5A6 monoclonal antibody to alpha tubulin and TuJ1 monoclonal antibody to betatubulin were the generous gift of Dr. David L. Brown (Aitchison and Brown, 1986).Secondary antibodies:FITC - conjugated goat anti-mouse affinity purified IgG (Organon TeknikaCappel, Malvern P.A.) was used as a secondary antibody to 5A6, and Tujl. Mouse IgG,used in control, was purified from normal mouse serum (Sigma S3509) using anImunoPure (G) IgG purification kit (PIERCE, Rockford III.).Binding assay treatment reagentsAM PPN P: non hydrolyzable analogue of ATP (5’-adenylylimidodiphosphate)Vanadate: sodium orthovanadateEHNA: erythro-9-[3-(2-hydroxynonyl)] adenineNEM: N-ethylmaleimideBinding assay 1 70Protease inhibitors were added to buffers from stocks, kept at -80°C. ConcentratedPMSF in DMSO was added to buffer immediately before interacting with tissue byvigorous stiring.METHODSGeneral protocolsProtein determinationsProtein concentrations were done by the methods of Lowry et al. (1951) orBradford (1976) as indicated.SDS PAGE gelsSDS PAGE gels were done according to Laemmli (1970), using 7% SDS PAGEminigels stained with Coomassie blue stain or 7.5% PhastGels (Phast Gel systemPharmacia LKB, Piscataway, N.J.), stained with silver stain (Phast gel method 3).Isolation of 3MTx - spermatid-ES binding assay componentsTubulin purification for binding assayFor this part of the study, tubulin was purified using a temperature dependentassembly, disassembly method using DEAE-Sephacel (Pharmacia) after the firsttemperature dependent cycle. This method provides equal purity but an improved yieldover that used for microtubule decoration experiments described in chapter 2. Brains,from freshly killed steer, were placed immediately on ice and transported to the lab(approximately 20 minutes). Tissue was homogenized in a cold Waring blender in A4Mbuffer (75 mi/i 00 g of tissue) 2X for 15 seconds each at low speed and 30 seconds athigh speed. Homogenate was centrifuged at 36,000 rpm (100,000 g), for 45 minutes,171at 4°C, in 42.1 rotor. The crude supernatant was brought to 0.1 mM GTP and added toDEAE-Sephacel (1/2 original supernatant volume), which had been equilibrated with1.0 M sodium glutamate, (pH 6.6) and mixed with DEAE-Sephacel by gentle rocking at4°C for 45 minutes. It was then centrifuged at 2,800 rpm, for 2 minutes, in a lEG CRU5000 centrifuge. The supernatant was discarded. The tubulin-bound Sephacel waswashed by being resuspended in 1.0 M sodium glutamate with 0.1 mM GTP (pH 6.6),allowed to sit for 2 minutes, and respun at 2,800 rpm, for 2 minutes, all at 4°C. Thewash step was repeated two more times, discarding the wash each time. Tubulin was theneluted from the Sephacel by incubating in 1/4 X original supernatant volume of 0.85 MNaCI with 1.0 M sodium glutamate and 0.1mM GTP (pH 6.6) for 20 minutes at 4°C. Itwas then spun at 2800 rpm, for 2 minutes in an IEC CRU 5000 centrifuge, at 4°C andcollected eluate saved. The eluting proceedure was repeated 2 more times by gentlyresuspending the tubulin bound sephacel, incubating it for 10 minutes and spinning at2800 rpm in lEG CRU 5000 centrifuge, for 2 minutes, each time saving the eluate.Excess Sephacel was centrifuged from the eluate. The eluate was brought to 1.0 mM GTPand the tubulin allowed to polymerize for 30 minutes at 37°C. The microtubules werecentrifuged at 35,000 rpm (96,000g) at 37°C, for 30 minutes using a 42.1 rotor. Thesupernatant was discarded and the microtubule pellet resuspended and homogenized in1/6 of original crude supernatant volume of 1 .0 M sodium glutamate. The microtubuleswere depolymerized on ice for 30 minutes, centrifuged at 36,000 rpm (100,000 g), at4°C, for 30 minutes, in 42.1 rotor. This “iX cycled” pure tubulin (MAP free) waseither frozen in liquid nitrogen and stored at -80°C at this point (1X cycled tubulin),before the final recycle step, or recycled one more time before freezing (2X cycledtubulin). To recycle theiX tubulin, it was placed in 1M sodium glutamate with 1mMGTP, polymerized for 30 minutes at 37°C, centrifuged at 36,000 rpm (100,000 g), at37°C, for 30 minutes, in 5OTi rotor. The pellet was resuspended and homogenized insmall volume of PEM 250, depolymerized for 30 minutes on ice, and cleared at at39,500 rpm (100,000g), at 4°C, for 30 minutes, using a 50 rotor. This purified “2X1 72cycled” tubulin (MAP free) was aliquoted and frozen in liquid nitrogen, then stored at -80°C. Protein concentration was determined according to the method of Bradford.Preparation of 3HGTP labelled, taxol stabilized, microtubules (3MTx):For the binding assays, 2X cycled tubulin was polymerized in the presence of3HGTP and taxol to produce labelled, stabilized microtubules as described by Wilson andco-workers (1985). All gradient solutions contained 20tM taxol to maintainmicrotubule stability in bound microtubules entering the gradient. Taking advantage ofthe fact that in the presence of taxol stoichiometric amounts of GTP are not required forpolymerization (Vallee, 1982), a substoichiometeric amount of GTP (labelled plusunlabelled) was used to ensure complete incorporation of competent GTP intomicrotubules and therefore minimize the residual label left in solution. To labelmicrotubules, 3HGTP (1 .0 iCi/50 jil of final microtubule solution) was included in thepolymerization step(3MT). MAP free microtubules were assembled by polymerizationof purified bovine brain tubulin ([1X] = 1.6 mg/mI) at 37°C for 30 minutes in PEM250 buffer, with 5 iiM GTP, and 20 jiM taxol (MT).Length measurement of 3HGTP labelled, taxol stabilized, microtubules(3MTx):Microtubules, polymerized at lx concentration (1.6mg/mI) in the presence of2OjiM taxol and 5j.tM GTP, (as prepared for binding assays but without 3HGTP label)were examined using negative stain electron microscopy techniques and microtubulelengths determined. Microtubules were placed on carbon coated EM grids, negativelystained with 1% aqueous uranyl acetate and observed using Phillips 300 EM at 60 kv.They were photographed, using Agfa plate film, and lengths were measured directly fromthe negatives. Using the negative, montages were formed, where necessary, to measurethe longer microtubules. All microtubules, for which both ends were clearly visible,were measured in 14 fields, a total of 103 microtubules.1 73Spermatid-ES isolationFigure 3-3 illustrates a ‘poke and squash method’ developed for isolation ofspermatid-ESs. Testes were removed, decapsulated and placed in PEM 250 on ice.Seminiferous tubules were carefully teased apart to increase access of cold buffer.Under a dissecting microscope, approximately 15 mm lengths of individual seminiferoustubules were removed from the testis and placed on an ice cold glass slide. Care wastaken to avoid overlap of tubules. The slide was replaced on ice frequently during theisolation to avoid having the spermatids warmed by the microscope lamp. Tubules werepoked at approximately 3 mm intervals and covered with a glass coverslip. Spermatidswere expressed from the tubules by very gently pressing the coverslip onto the slide(under direct observation of the dissecting scope). The coverslip was removed and,using fine forceps, all seminiferous tubules were removed from the slide and coverslipand discarded. Using a minimum of PEM 250 buffer, the remaining isolate was gentlyrinsed, from both the slide and coverslip, into a cold collecting container and kept colduntil just before use. For stage specific isolation, the testis was incubated in PEM 250with 2mg/mi collagenase for 3 minutes, teased gently apart, and rinsed in PEM 250,without collagenase, before isolating individual seminiferous tubules. Staging ofseminiferous tubule seqments was done according to Parvinen and coworkers (Parvinenand Tapani-Perttula, 1972; Parvinen and Ruokonen,1982), as elaborated byKangasneimi and coworkers (1990). A complete wave of seminiferous epithelium, fromone dark zone-pale zone change-point to the next, was identified. The wave was cut intofour segments, based on the transillumination pattern of pale zone (stages IX to Xli),weak zone (XIII and I), strong spot zone (Ii to VI), and dark zone (VII and VIII). Thesegments were placed on a microscope slide, gently squashed, and immediately visualizedby phase microscopy, to determine if spermatids could be recovered from all stages.1 74Figure 3-3. Spermatid-ESs isolation. Steps for spermatid-ESs isolation areindicated in the figure. The essential features of this isolation technique are 1) Tissue iskept cold at all steps, including being returned onto ice repeatedly while being handled onmicroscope, to encourage depolymerization of endogenous microtubules. 2) Theseminiferous tubules are arranged on the slide such that they do not overlap, to ensurethat pressure is applied evenly to the tubules. 3) Holes are poked in the seminiferoustubules to allow spermatids to exit all along the tubule, reducing the pressure requiredto get a good yield. This can be observed under the microscope. 4) Tissue is recoveredfrom both the slide and the coverslip. 5) A minimum amount of buffer is used to rinsetissue from slide and coverslip, to discourage dilution of cytosolic factors. Spermatidsare allowed to settle and the cold buffer is taken from the top for each rinse, to reducethe total amount of buffer added.L75Spermatid - ES IsolationPlace on Ice inPEM 250 butter withprotease inhibitorsPlace isolatedS rmatidSsoniceLJTease apart 1 cm lengths ofseminiferous tubules (canselect specific stage)-Keep cold--Add cover slip-Press down gentlyon coversllpRemove coverslipV 1Discard tubulewalls from slideand coverslip--Remove testisPEM with proteaseInhibitorsslidePoke holes Inseminiferous tubules Remove any clurrps ofcells adhering to tubulesifIfRinse spermatidstrom slide andcoverslip1 76Binding assay: preparationSample preparation for binding assayFigure 3-4 illustrates the method used for the binding assays. Labelled andstabilized microtubules were prepared using 1 1iCi/50 ii of tubulin (3MT). Isolatedspermatid-ESs were prepared under cold conditions and kept on ice. Spermatid-ESswere quickly brought to room temperature and combined with 3MTx, at a ratio of 1:2(3Mt:spermatidESs) (in most cases 50 iil:100 uI) and either used immediately(time=0) or allowed to incubate at room temperature for the duration indicated by theexperiment. Samples made up in this way are referred to as ‘no treatmentcontrols’, serving as a basis for comparison with treated groups. To avoid dilution ofsample or gradient solutions, all treatment solutions were added at 100X the finalconcentration, to both sample and gradient solutions.Gradient preparation for binding assayGradients were prepared from stock solutions immediately before each spin.Stocks of 60, 45 or 30% sucrose in PEM buffer, with freshly added protease inhibitorsand 20 jiM taxol, were used for the gradients. Treatments, such as 10 mM MgATP thatwere added to samples, were also included in all levels of the gradient on which thetreated sample was to be loaded, to prevent reversal of binding by dilution of the agentbeing tested. (Treatments with: added excess cold microtubules; dynein enriched MAPs;Sertoli cell cytosol, and different concentrations of MTs or supernatants, were notincluded in the gradients.) Each gradient consisted of 150 jil of PEM6O%, 200 jilPEM45%, 200 jil PEM3O%, layered from bottom to top. Gradients were loaded with150 jil of sample: 50 jil 3MTX +100 jil spermatid-ESs (exceptions are noted).Binding assay 1 77Figure 3-4. Binding assay: methods: This figure illustrates the method used forrunning binding assays. Labelled microtubules (3MTxs) are assembled by addingtritium labelled GTP(3HGTP) at 1 ici/5O .tl tubulin stock, to tubulin in PEM25O plustaxol and GTP buffer, and polymerized for 30 minutes at 37°C. Spermatid-ESs, arequickly warmed to room temperature and immediately combined with 3MTxs at 50 jil3MTx for each 100 jil of spermatid-ES preparation. Gradients are assembledimmediately prior to the spin for which they are to be used. If the tissue is to betreated, treatment agent is added (from 100X concentration stock) to each sucrose stockfor those gradients. Gradient levels are marked with an indelible pen (that will notbleed color into the scintillation fluid) after the application of each level of sucrose, toprovide markers for accurate cutting of the gradient after freezing. The spermatid-ES +3MTX mixture is loaded onto the sucrose gradient (150 uI! gradient). The gradientsare centrifuged in a Beckman SW 65 rotor (3 bucket rotor), for 5 minutes, at roomtemperature (the spin conditions must be verified for the centrifuge used). Gradientsare removed from the rotor, inspected, and immediately frozen at -70°C. The frozengradients are cut as shown in the figure, midway between sample/30% and 30/45%,and between 30/45 and 45/60% interfaces. 4.5 mIs of scintillation fluid (plus 500u1dH2O/sample) are added to each sample and counts/minute determined. Combined countsfrom the 30/45%, and 30/45 and 45/60% gradient fractions are expressed as apercent of the total count.Binding Assay78Centrifuge Beckman SW 65rotor 5000 rpm 5 minutesroom temperatureExamine andfreeze to -70 CCut-withrazorbladeAdd scmtillat Ionfluid3MTXSpermatid-ESsI Load ontogradientSample (1501.tI)30% (200.iJ)45% (2O0il)60% (150d)(PEM%Sucrose)TopCount % total countCombine counts B & CExpress as % of total countAdd tubulin In PEM25Owith taxol and GTP3HGTPEvaportate ETOHlpCi/50 ii tubulin3MTX1:2spermatid- ES5iii N Count30/45interface45/60interface% total countBinding assay 1 79Binding assay: methodsRunning the binding assaysThe time required for isolation of fresh spermatid-ESs was a limiting factor onthe volume of sample that could be used in the binding assays. The SW 65 rotor(Beckman), with #356860 adaptors, provided the ideal volume (800 jil capacity) forthe gradients. Because preliminary experiments indicated that it was imperative thatcentrifuge conditions were constant from spin to spin, the same ultracentrifuge(Beckman L7-55) and rotor (Beckman SW 65) were used for all binding assays. Adefined proceedure for ultracentrifugation was followed, duplicating the rate ofacceleration, timing by use of “hold” option, precise duration of spin (verified byrevolutions on meter), and rate of deceleration.Use of the SW 65 rotor (with 3 buckets) dictated that only three experimentalconditions could be included in any spin, one being the ‘no treatment control’ that servedas a basis of comparison for other conditions in any given spin. Gradients were preparedimmediately before each spin. 150 jil of sample were loaded onto the sucrose gradients,spun at 5000 rpm, for 5 minutes, at 220C. The gradients were inspected andimmediately frozen to -80°C. A number of techniques that potentially could introduceerror such as: methods of sampling from gradients, effects of freezing, means used tostandardize the spin conditions, and effect of gradient solutions on scintillation fluid,were checked.Binding assays: experimental designDesign of binding assay experiments (general)Experiments were designed around the constraints of 3 gradients/spin.Consideration was given to an apparent time course of binding and different effects ofBinding assay 1 80adding treatments before or after binding was established. In general, three conditions,one being the ‘no treatment control’ and the other two being treatment conditions, wererepeated throughout the entire experiment (see Figs. 3-17 and 3-18 for an example ofa short experiment employing the general design). A ‘no treatment control’ was includedin every spin, unless the effect of a second treatment was imposed simultaneously on ‘notreatment’ and the other two treated samples in the same spin (for example, 10 mM ATPeffect on ‘no treatment control, 10 j.tM vanadate and 100 p.M vanadate). In that case, the‘no treatment control’ value was extrapolated from the ‘no treatment control’ values ofthe previous and subsequent spins. This experimental design was used to test 10 mMMgATP, 2 mM NEM, 1 mM EHNA, 1 mM AMPPNP, 5 mM GTP (with and without ATPdepletion), 10 and 100 p.M vanadate, ATP depletion with l8units/ml hexokinase and Dglucose, cold 5X [MTxI competition (or microtubule depleted supernatant), 1X [3MTx]supernatant, Sertoli cell cytosol and cytoplasmic dynein enriched testis MAPpreparation. Spins to test the effect of 10 mM MgATP on all these conditions wereincluded.Experiments for which total sample added to gradients was not 150 p.1In experiments in which treatment material could not be added at bOXconcentration (for example, competition with the addition of 5X cold microtubules) totalvolume of sample was greater than 150 p.1. The effect of dilution, by the treatmentmaterial, was controlled for by the addition of equal amounts of buffer to the ‘notreatment control’.Direct effect of treatments on microtubulesTo rule out the possibility that treatments influenced binding by a direct affect onmicrotubules, each treatment was added directly to aliquots of polymerizedmicrotubules. Treated microtubules were pelleted in a Beckman airfuge and theBinding assay 1 81microtubule mass in the pellet compared with ‘no treatment’ microtubules to look forany direct effects of treatments on microtubules.Design of matched experimentsA series of experiments were designed to look at the effects of a number ofbinding inhibitors on early and established binding, matching as closely as possible fortime and spermatid-ES and microtubule concentration. They were carried out with afirst series of spins using samples to which treatments had been added before bindingwas begun (early binding) followed by a second series using samples treated at least 2hours after binding had begun (established binding). This was achieved by dividing eachof the 3MTx stock and the isolated spermatid-ES stock in half. For “early binding”,treatments were added to the aliquot of spermatid-ESs and then combined with the 3MTx.This was time = 0. For the second series, “established binding”, the 3MTx andspermatid-ES samples were combined at time = 0, but treatments were not added untilafter the first series of spins were complete, that is at the end of the “early binding”runs. For both the early binding and established binding series, the first gradient wasprepared immediately before treatments were added so that data from the first spinscould capture the earliest possible effects of treatment. The effect of 10 mM MgATP wastested in each group. For this series of matched experiments, the microtubuleconcentration was O.5X.Design of [MT] experimentFigure 3-5 shows the design of the microtubule concentration experiment. Totest the MT concentration effect on 3MT- spermatid-ES binding, a stock of 3MTx wasassembled at lox tubulin concentration (1X[3MT= 1.6mg/mI, ie 0.53 mg/mI in finalassay sample). Stocks of 0.1 X, O.25X, 0.5X, 1 .OX, 3.OX, and i ox [3MTx] were dilutedfrom the lOX[3MTx] stock. Polymerization was done before dilution as polymerizationproperties depend on tubulin concentration.Binding assay 1 82Figure 3-5. Microtubule concentration experiment: strategy: This figureshows the design of the microtubule concentration experiment. Three things weredetermined: 1) the effect of microtubule concentration [T] on3MT-spermatid-ESbinding (binding assay); 2) the reliability of the assumed tubulin concentrations(concentration assay): and 3) any loss of counts from pelleted microtubules betweenspin 1 and 6 (cycling of label)(stability of label). Tubulin was polymerized at 10 Xconcentration and all dilutions made from that stock (hence data expressed in absolutecounts). [T]= tubulin concentration in microtubule samples. The binding assay wascarried out as described earlier, using the tubulin from each of tubulin concentrationstocks with a 1 X sample included in every spin to act as the control. For theconcentration assay, counts and tubulin concentrations were determined directly fromtubulin concentration stocks, verifying that assumed dilutions were reliable. To test thestability of label, samples from each tubulin stock was tested, timed to coincide withspins 1 and 6. At each time period, three samples from each stock were spun over 15%PEM sucrose cushions, in a Beckman airfuge, at maximum speed, for 30 minutes. Fromthese airfuge spins, counts of pelleted microtubules were determined as shown. Pelletedmicrotubule counts showed that counts pelleted from the samples did not change duringthe experiment; label does not cycle from the taxol polymerized microtubules.L83Figure 3—5: Microtubule concentr6tlon experimentU Ispermetid—Ess I [Lioxi MTx_________________________________I4 — [X I 3> I IX O.5X O.25X o.ioxjbinding sg COUfl concentration asseg-countI counts nd tubulin concentrationI (T]from each concentration stockI (Tistabilitg of labelU 00 airfuge spins plus for each concentration000 2 3 1000 CT 9at the beginning of spin Smicrotubules are pefleted in airfuge0 00 count (o) count count supernatant (TIsupernatant count000 pellet countU U U ii airfuge th. beginning of spin 6microtubules are pefleted in airfug.0 00 1 2 supernatant CT]U 00CT] C.. (T]9 supernatantcount(c) count count pellet count000000Binding assay 1 84Samples from each concentration stock were used in two ways: 1)binding assayswere done to test the effect of MT concentration on3MT-spermatid-ES binding, and 2)tubulin concentration and MT incorporation assays were run, concurrently with thebinding assays, to verify tubulin concentration and label stability (Fig. 3-5).1) Binding assays: Samples from each of the 0.1, 0.25, 0.5, 3X and lOX 3MTconcentrations were added to the spermatid ESs. The resulting 3MT - spermatid-ES5were allowed to bind for 1 .5 hrs (established binding) and tested for binding on sucrosegradients. Samples from each [3MT] were loaded onto the gradients four times in variedorder, except the 1X [3MT] - spermatid-ES sample which served as the ‘no treatmentcontrol’ for each of the ten spins.2) Protein concentration and label incorporation: For this part of theexperiment, MT protein concentrations were determined (Bradford) and scintillationcounts made from each MTx stock. This allowed the comparison between the number ofcounts in samples loaded in each of the binding assays with the number of counts insamples directly aliquoted from the stock, to verify the tubulin concentration in eachstock. To test for possible cycling of microtubules over the time period of theexperiment, and therefore loss of label, three samples were taken to coincide with thebeginning of the binding assay spins and three more to coincide with the midpoint of thebinding assay spins. In each case, 50 il of sample was pelleted over a 125 ill cushion of15% sucrose in PEM in a Beckman airfuge, at full speed (130,000 to 150,000 rpm),for 20 minutes. Percent incorporation of tubulin into microtubules was determined foreach sample and total counts were done for each of the six samples at each concentration.The counts and protein checks verified that the measured protein concentrations of thesamples were consistent with the assumed concentration for the binding assays,(0.1X -10 X), that the percent incorporation into pellets was essentially the same at all 3MTxconcentrations, and that there was no turnover in the microtubules, ie, the label did notcycle out of the pellet during the experiment.Binding assay 1 85Topical binding assayIn preliminary experiments, an attempt was made to carry out the binding assaydirectly on acid cleaned, polylsysine coated glass slides, evaluating binding by thepresence of antitubulin staining on spermatid heads. Spermatid-ESs were isolated at theselected temperature and then spun through a cushion of 30% sucrose in a Beckmanairfuge at maximum speed for 3 minutes. The isolated spermatid-ESs were collectedfrom the bottom of the tube, placed on glass slides, and allowed to settle for fifteenminutes. Spermatid-ESs were then reacted with microtubules directly on the slide,using a variety of conditions and temperatures. While some trends were observed withthese assays, they were replaced with the radiolabelled technique for three reasons. 1)To show that binding occurs, it was deemed necessary to demonstrate that exogenousmicrotubules (labelled in the gradient assays) could be bound. The topical assays did notdistinguish between exogenous and endogenous microtubules. 2) Because microtubulesattached to slides, their presence could not unambiguously be interpreted as binding. 3)The radiolabelled gradient assays could be quantified.Preparation of treatment materials used in binding assaysIsolation of rat testis crude supernatant for cytoplasmic dynein enrichedMAP preparationRat testis crude supernatant and cytoplasmic dynein enriched MAPs wereprepared after the method of Neely and Boekelheide (1988). Testes were removed, from25 CO2 killed rats (21 days of age), decapsulated, and put immediately into iced A4Mbuffer with protease inhibitors (see buffers). Testes were homogenized using Douncehomogenizer at 1300 rpm (3-5 times up and down) with 1:1 buffer:testis by wt (yieldSN volume about 40 mIs). The resulting homogenate was spun at 36,000 rpm(100,000g) at 4°C, for 30 mm, in a 42.1 rotor. The supernatant was frozen in liquidBinding assay 1 86nitrogen and stored at -80°C until required for cytoplasmic dynein enriched MAPpreparation.Isolation of cytoplasmic dynein enriched MAP preparation from crudesupernatantCrude supernatant from rat testes (42 mIs) was thawed and cleared bycentrifugation at 40,000 rpm (130,000g), at 4°C, for 45 minutes, in a 42.1 rotor.The supernatant was brought to 20 jiM taxol and underlayed with 25% sucrose in MESbuffer (see buffers) using an 18 g needle. The supernatant was incubated for 10 minutesat 37°C to initiate the tubulin polymerization, then returned to ice for a further 20minutes. (Microtubule polymerization is stablilzed from this point on by 20 jiM taxol).Microtubules were pelleted at 36,000 rpm (100,000g) at 4°C, for 30 minutes in42.1 rotor. The pellet was resuspend in MES buffer plus 201iM taxol (about 1/10 volof testes homogenate), gently homogenized and placed on ice for 5 minutes. Microtubuleswere again pelleted at 36,000 rpm, at 4°C, for 30 minutes in a 50 rotor (75,000 g).The pellet was resuspended (in same vol about 1/20 original testis vol) in MES bufferwith 2OjiM taxol, and 5 mM GTP and 5 mM MgCI2, homogenized gently, and allowed to sitat room temperature for 10 minutes. After a further 5 minute incubation at 37°C, itwas spun at 36000 rpm, (90,000 g) at 29°C , for 30 minutes with type 50 Ti rotor.The pellet was resuspended in 1/20 volume of MES buffer with 20 jiM taxol, 1mM GTP,1 mM MgCI2 and 1 mM ATP, gently homogenized, and allowed to sit for 10 minutes atroom temperature and then for 5 minutes at 37°C. It was again spun for 36,000 rpm,(90,000 g) at 29°C , for 30 minutes with type 50 Ti rotor. MAPS were eluted from themicrotubules by resuspending the pellet in MES buffer (1/20 of original supernatantvolume) with 20 jiM taxol, 1 mM GTP and 10 mM MgCI2 and 10 mM ATP. Followinggentle homogenization, microtubules were incubated for 15 minutes at roomtemperature, thenlO minutes at 37°C, and spun for 30 minutes at 29°C at 36000 rpm,in a 50 Ti rotor. This supernatant contained ATP eluted cytoplasmic dynein (HMW2)Binding assay 1 87enriched rat testis MAPs. The cytoplasmic dynein enriched MAP fraction was salt elutedon exocellulose GF5 40-100 iiM desalting gel (Pierce) at 4°C into PEM 250 and usedfor binding assays immediately after being removed from the column (concentration =210 tg/ml). For binding assays, cytoplasmic dynein (HMW2) enriched rat testis MAPswere loaded at 1/100 dilution final concentration = sample. This HMW2 MAPenriched preparation (considered to be MAP 1 C or cytoplasmic dynein) has been shownto be present, independent of the development of mature germ cells, and is considered tobe of Sertoli cell origin, as it was present in Sertoli cell enriched preparations from tworat models of germ cell depleted testes (Neely and Boekelheide, 1988).Sertoli Cell enriched isolation from 21 day old ratsTestes were removed from 21 day old rats, immediately placed into CMF Hanks(Hanks balanced salt solution: calcium and magnesium free), and washed 3X with sterileCMF Hanks. Testes were then decapsulated and moved into fresh CMF Hanks, and minced.Testes pieces were transfered to a solution of: 1 .9 ml 40X trypsin, 0.5 mg DNase I, and48.1 ml CMF Hanks and incubated for 30 mm at 32.5°C on a rotating table at 150 rpm,with occasional added swirling. Tissue was poured into a 50 ml centrifuge tube andallowed to sediment for 5 mm. The supernatant was discarded. 20 ml HBSS were addedto the tubules, which were then pipetted 10-15 times with a 10 cc pipette and allowedto sediment for 5 minutes. The supernatant was removed. The pelleted tubule fragmentswere added to 35 ml of CMF Hanks with; 45mg collagenase, 0.6mg DNase I, and 30 mghyaluronidase. The fragments were incubated for 60 mm at 32.5°C, rotating at 150 rpmwith occasional swirling, then poured into a 50 ml centrifuge tube and spun at 1000rpm in a clinical centrifuge for 3 minutes. The supernatant was removed and 0.1%soybean trypsin inhibitor added to the pellet 0.35m1 in 35m1 DMEM/F12 (pen/strep,gent., L- glutamine). The cells were pipetted 10-15 times with a 10cc pipette to breakup the pellet, and spun down at 1000 rpm for 3 minutes. The supernatant was discardedBinding assay 1 88and the cells resuspend in 20m1 F12/DMEM (pen/strep, gent., L- glutamine). Cellswere pipetted 10 times to break up any cell clumps.Preparation of cytosol: Sertoli cell enriched preparation from 21 day rattestisSertoli cells, isolated from 21 day old rats as described above, were washed oncein 20 mM HEPES-KOH, 1mM EDTA, 1 mM MgCI2 pH 7.4 and resedimented bycentrifuging at 1000 g for 5 mm. The supernatant was removed and 1 volume of 100mM PIPES-K, 250 mM sucrose with protease inhibitors pH 6.9 was added. The cellswere homogenized using a Dounce homogenizer and centrifuged for 10 mm at 43,000g at4°C. The supernatant was collected and recentrifuged, at 50,000 rpm (150,000g), at4°C, for 90 minutes in a 50 Ti rotor, to obtain the Sertoli cell enriched cytosolicextract. The extract was frozen in liquid nitrogen and stored at -80°C. lOjil PMSF(fromloo mM stock) per ml buffer was added just prior to use.Data collection for binding assaysFrozen gradients were cut into three fractions with a razor blade as shown in Fig.3-4. Each fraction was place in a scintillation vial, mixed with 500 tl of deionizedwater and 4.5 mIs of scintillation fluid (Optifluor: Packard) mixed well until clear,allowed to settle, and counted. Counts were always very low in the 45/60 fractioncompared with the 30/45 fraction. Any changes in that trend could be traced back tofaults in gradient assembly. Because this pattern was consistent and no differences incell content were noted between the fractions, data from the 30/45 and 45/60 fractionswere combined for analysis. Data from any gradients in which the appearance ofinspected gradients, spin duration, control values, or counts loaded, cast doubt on thespin were not included.Total counts loaded onto gradients were consistent throughout any givenexperiment, indicating that sample did not change over time. Variability of counts in1 89gradient fractions for ‘no treatment controls’ was low within any experiment. However,there were differences in ‘no treatment counts’ between experiments. Counts areexpressed as a percent of the same spin ‘no treatment control’, thereby reducing theeffects of between experiment variability.Sucrose concentration was determined to have no effect on the scintillation fluid.Repeat counts in a second counter or in the same counter after a period of 12 hrs,indicated that there were no spurious effects of chemiluminesence or counter efficiency.Data analysisFor the determination of significant main effects of experimental treatments,data was analyzed by analysis of variance (ANOVA); significance at P < 0.05. WhereANOVA showed significant main effects, Tukey’s paired tests were used for between groupcomparisons.Immu nohistochemistryImmunofluoresence: sample preparation and microscopy (5A6: anti-tub uli n).Spermatid isolate was adhered to acid-cleaned polylysine coated glass slides. Itwas then fixed in 3.7% paraformaldehyde in PEM 250 for 10 mm, rinsed 3 times, 5minutes each, in PEM 250 and plunged into -20°C acetone for 5 minutes. Slides wereleft to airdry briefly after which tissue was rehydrated with 0.1% BSA/PBS for 5minutes. Non-specific sites were blocked with 5% normal goat serum in 0.1%BSA/PBS for 30 mm. Tissue was incubated with primary antibody (diluted in 0.1%BSA/PBS withl % normal goat serum) for 60 minutes at 37°C in moisture chambers,then rinsed 3 times with 0.1% BSA/PBS at 5 minutes/wash. Tissue was incubated withFITC congugated goat anti mouse secondary antibody (diluted in 0.1% BSA/PBS withl%normal goat serum) for 45 minutes, in moisture chambers, at room temperature,Binding assay 1 90washed 3 times with 0.1% BSA/PBS for 5 minutes! wash, mounted in 50% glycerolwith p-phenylenediamine (Johnson and Nogueira Araujo, 1981), and sealed with clearnail polish. Slides were kept at 4°C in the dark and photographed within 24hrs ofstaining.Antibody final dilutions were: 5A6 anti alpha tubulin (1:4000), and TuJ1 antibeta tubulin (1:10,000) for primary anti-tubulin antibodies and FITC - congugatedanti mouse lgG (1:200) for the secondary antibody. Controls used were: normal mouseserum, lgG purified from normal goat serum, 0.1% BSA/PBS withl% normal goatserum substituted for primary antibodies or both (no 10, no 20, no 10 and no 20), andtubulin block for 5A6.Rhodamine phalloidin stainingFor actin staining, rhodamine phalloidin (1:200) was used (and tissue processedas described above, eliminating the primary antibody steps. For dual staining rhodaminephalloidin was added at the same time as the secondary antibody.Preparation of normal mouse IgG: control for 5A6 anti-tubulin antibodyMouse lgG was isolated from normal mouse serum (Sigma) using an ImunoPure(G) lgG purification Kit (PIERCE, Rockford III). Normal mouse serum (Sigma S3509)was clearified by centrifugation in clinical centrifuge at maximum speed for 15minutes. It was then diluted 1:1 with ImmunoPure Immobilized Protein G column andwas equilibrated with binding buffer. Serum was applied to the column, which had beenwashed with 10 ml binding buffer. Normal mouse lgG was eluted from the column with 6mIs of elution buffer, monitoring 1 ml fractions with A280 (Gilforld spectrometer) tolocate IgG fraction. lgG fraction was desalted on desalting column using equilibrationbuffer (kit) and eluted with PBS. Resulting IgG was concentrated back to originalvolume using centricon microconcentrator (AMICON).Binding assay 1 9 1Autoradiography of binding assayAutoradiography: methodsTissue was prepared as described for binding assay for no treatment control and10mM MgATP treated groups. Samples which had not been enriched on gradients andsamples recovered from the combined 30/45 and 45/60 interfaces were processed forautoradiography. Microtubules were labelled using 10 j.tCi/50 Ltl of tubulin atlXconcentration (1 OX that used in binding assays). To avoid possible morphological damageby freezing, tissue for autoradiography was drawn slowly from the 30/45 and 45/60interfaces using a Hamilton syringe pushed through a small hole made in the wall of thecentrifuge tube. To provide a count check on the material used for autoradiography,binding assays were done using parallel samples from the ‘no treatment controls’ and 10mM MgATP treated groups. 10 mM MgATP reduced binding to 20% of ‘no treatmentcontrols’.Samples for autoradiography were placed directly on acid washed glass slides andafter being allowed to adhere, were fixed with 3.7% paraformaldehyde in PEM 250,incubated for 5 minutes in -20°C acetone, rinsed in PEM 250 and allowed to air dryovernight. Using deflected red light, slides were dipped in melted emulsion at 40°C(Kodak), allowed to dry for 30 minutes and suspended in light proof boxes containingsilica gel dessicant pouches. Boxes were wrapped in 3 layers of foil and stored at 4°C.For development, slides were brought to room temperature, removed from the boxesunder deflected red light, developed in Kodak Dl 9 (diluted to 50% with dH2O) for 3minutes, placed in stop bath of 8m1 Kodak stop indicator/50 ml dH2O for 30 seconds,fixed in 25% sodium thiosulfate for 5 minutes, and rinsed for 30 minutes. Slides weredehydrated in ascending series of ethanol (70, 95, 95, 100, and 100%), placed in 2changes of xylene, and the coverslip mounted with permount. Before photographing,emulsion was scraped from the back surface of the glass slides with a razor blade.1 92Periodic development determined that the optimum exposure time for silver graindevelopment was 3 weeks.A number of observations cast doubt on the reliability of carrying out quantativemeasures for the autoradiography experiments. Firstly, during processing forimmunocytochemistry, it was observed that spermatid loss occurred at each rinse step.In addition, spots were seen associated with small amorphous clumps in autoradiographyslides. Together these observations suggested that the 3MT- spermatid-ESs, that hadadhered to the glass slide, may be disturbed by processing. Secondly, when 3MTx-spermatid-ES5 were exposed to either paraformaldehyde or acetone, total counts ingradient fractions were reduced over controls. This was suprising in that under anyother treatment conditions, the total counts loaded were consistent. (see results).The following experiment was done to check the effects of processing (rinse stepsand acetone and paraformaldehyde treatment) on 3MT- spermatid-ES binding. MTSpermatid-ES (cold MT) sample was allowed to bind. Three conditions, 10 mM MgATP,ATP depleted with hexokinase and D-glucose, and ‘no-treatment control’ were eachsampled in the following ways 1) before being put onto the gradient (pre-gradientprocessed), 2) after being removed from the gradient but unfixed (post gradientunprocessed) and 3) after being removed from the gradient and processed for either 5A6anti tubulin, without primary antibody or with 5A6 and rhodamine phalloidin for duallocalization of tubulin and actin (post gradient processed). Fixative proceedure was thesame as that used for immunhistochemistry described earlier. Slides were examined andphotographed immediately following the experiment on a Zeiss axiovert fitted withfilters for FITC and rhodamine and with DIC optics.Electron MicroscopyRoutine electron microscopyFor experiments described in this chapter, tissue processed for electronmicroscopy was immersion fixed using 1 .5% glutaraldehyde and 1 .5%1 93paraformaldehyde. Following initial fixation, tissue was processed for electronmicroscopy as described in Chapter 2. Any temperature or buffer variations aredescribed with the results.Methods for calcium studyFor these studies, potassium antimony (EM Sciences) was boiled in dH2O (pH8.6), cooled to room temperature and centrifuged at 10,000 rpm (16,000g) at 20°Cfor 30 minutes in a GSA rotor (Sorval) to remove insoluble precipitate. All EMfixation in this series involved tissue preparation, and fixation, followed by routinedehydration and embedding (see Chapter 2 for routine dehydration, embedding, andsectioning). For all experiments, tissue was observed with and without stain (uranylacetate and lead citrate for 6 minutes each) to distinguish antimonate from metal stain,both visually and with elemental analysis. Calcium localization micrographs werephotographed on a Phillips 300 electron microscope, from thin sections that had beenstained with uranyl acetate and lead citrate to show ESER membranes.The literature is repleat with potassium antimony precipitation methods forcalcium localization. Three other calcium precipitation methods were tried. Onlypotassium antimony consistently gave positive results.Method 1: (Tandler et al., 1970; Kierszenbaum et al., 1971) Fresh decapsulatedrat and squirrel testes and hearts were cut into aproximately 1 mm cubes, and incubatedat room temperature in 2% potassium antimony, in deionized, distilled water, for 6 hrs.(Potassium antimonate acts as a primary fixative). The blocks were then hardened in5% paraformaldehyde in potassium antimony, for 16 hours, at room temperature, andwashed 2X 15 minutes in dH2O. Samples were post fixed in cold 2% osmium tetroxide(OSO4) in dH2O for 1 hour (parallel samples were processed with 0S04 excluded.Samples were rat testis, squirrel testis and rat heart muscle.Binding assay 1 94Method 2: (Klein et al., 1972; as used by Franchi and Camatini, 1985). Rattestis was cut into blocks and incubated in 2% glutaraldyhyde with 2% potassiumantimony in 0.01 M acetic acid for 4 hours, on ice. Tissue was washed 3X with cold .01 Mactic acid with 2% potassium antimony, postfixed in 1% 0S04 with potassium antimony(pH 7) for 1 hour, washed with cold .01 M acetic acid with 2% potassium antimony, 3X10 minutes each and processed for EM.Method 3: (Borgers et al. 1981). The proceedure is the same as method 1 except:potassium antimony incubation was for 12 hours and was preceeded by a 30 minuteHEPES-NaCI buffer (137 mM NaCI and 10 mM Hepes) preincubation step with either;buffer only, 5mM EGTA, or 4 mM calcium chloride at room temperature. Thepreincubated tissue was rinsed briefly in one change of potassium antimony to removeexcess preincubation solutions and then incubated in fresh potassium antimony.Method 4: (Spicer et al. 1969) Rat and squirrel testes and hearts were cut intocubes, and incubated for 1 hour at 4°C in 2.5% potassium antimony with 1% 0S04 (pH7.4 with acetic acid), dehydrated and embedded. A variation on this method (Cramer andGaIl, 1979) using 2% potassium antimony incubation for 1 hour on ice, was also tried.Three other methods were tried which have been used with limited success in theliterature.Method 5: In this method, oxalate is used, to bind calcium, and antimony, toproduce an electron dense precipitate (Borgers et al., 1977). Tissue was immersed in3% glutaraldehyde with 90mM oxalate at 4°C (pH 7.4), rinsed for 24 hours with 7.5%sucrose with 90mM oxalate at 4 °C, post fixed for 2 hours with 1% 0s04 and 2%potassium antimony in 0.1N acetic acid at 4°C.Method 6: NHA (N,N-Naphthaloylhydroxylamine) is used to precipitate calcium(Zechmeister, 1 979).Binding assay 1 95Method 7: Tetrabutylamonium fluoride method (Poenine and Epel, 1987). Blockof rat testis and rat striated muscle were incubated inlOOmM tetrabutylamoniumfluoride with 1% glutaraldehyde in 25mM Hepes buffer (pH 7.2) at room temperature,for 15 minutes, rinsed in the same buffer with glutaraldehyde removed, post fixed on icein same buffer with 1% 0S04 and processed for EM.Binding assay 1 96RESULTSThis portion of the work consists of a combined biochemical and morphologicalapproach to study the binding potential between microtubules and isolated spermatid-EScomplexes. The purpose of these experiments was to: 1) define the components of thebinding assay, 2) establish and test the criteria set for the binding assay, 3)characterize binding under a number of different conditions and 4) determine thelocation of label in the assay.COMPONENTS OF THE BINDING ASSAYSpermatid-ESsIsolation of spermatids, by gentle poking and squashing of seminiferous tubulesegments, resulted in a crude isolate comprised of spermatids from different stages ofspermiogenesis. In order to determine that spermatids could be recovered from a rangeof epithelial stages, including when they are deeply placed in Spermatid crypts,spermatids were isolated from segments of seminiferous tubules, using thetransillumination method (see methods). The squash method yielded spermatids fromeach of the segments. Figure 3-6: a-e shows the phase contrast microscopic appearanceof spermatids immediately after they had been squashed from the tubule segments whilethey were still clustered together and surrounded by cytoplasm. To better visualize thespermatid head, they are also shown after spermatids from these isolates had beenseparated from one another. Figure 3-6:f is the DIC image of a general stage isolateafter it had been rinsed from the slide. While spermatids remain still closely associatedwith one another, the phase contrast image highlights mainly the nuclei of spermatidheads. When spermatids are separated from the surrounding cell material, their phasecontrast image changes. The outline of the acrosome can then more readily be seen,giving the impression of a larger spermatid head, with a slightly greater curvature, ItBinding assay 1 97Figure 3-6. Appearance of spermatid-ESs isolate. Spermatid-ES wereisolated from defined stages of seminiferous epithelium to verify that all stages ofelongate spermatids can be recovered using the seminiferous tubule squash method. Tofacillitate separation, clumps of seminiferous tubule were incubated for 3 minutes inPEM 250 with 2mg/mI collagenase, and then rinsed with PEM 250 (withoutcollagenase). Complete waves of seminiferous tubules were teased from the clump andtransillumination pattern zones from one dark-zone to pale-zone change point, to thenext, were identified (see introduction for explanation of zones). To examine spermatidssquashed from different stages, waves were cut into four zones (see figure 3-2): ‘palezone’ (stages IX to XII), ‘weak spot zone’ (XIII and I), ‘strong spot zone’ (II to VI), and‘dark zones’ (VII and VIII) and isolated from each of the four segments by very gentlesquashing. (a-e) phase contrast (f) DIC micrographs; tissue unfixed.(a) Isolate from weak spot zone: stages XIII to I, spermatid steps XIII to XV,(b) Separation of spermatids from isolate preparation in (a) to visualize individualspermatids. The acrosome identified only as a pale corona in (a) is more visible in (b)(arrow). During these stages (spermatid steps XIII to XV), phase image of flagellabecome thicker.(c) Tissue from stages II to VI (strong spot zone): late spermatid steps XVI to XVIII.During these stages elongate spermatids achieve their deepest position within Sertolicell crypts.(d) Spermatid separated from the isolate in (C) showing long strands (arrowhead)typically seen extending from heads of spermatid-ESs isolated under warm conditions.The acrosome extends beyond the anterior tip of the nucleus. Together, these featuresgive the isolated spermatid in (d) a more curved appearance than those that are stillassociated with surrounding cytoplasm. The dorsal surface of the nucleus (small whitearrow) can be distinguished from a strand extending along the dorsal surface (arrowhead).(e) Individual spermatids in looser association with the isolate than surrounded them in(c) providing better visualization of the more completely developed flagellum(f) DIG image of multistage isolate, after being rinsed onto slide. Note the spermatidsare surrounded with other cellular material in addition to the spermatids.bars: a-e = 10 jim, f = 20 jim..ilBinding assay 1 99is not known whether being removed from the confines of the cytoplasm results in actualincrease in curvature. It was found that spermatids could be removed from all foursegments of epithelium using the squash method, that their appearance is slightlydifferent when viewed in isolation, and that a wide range of stages are represented in thegeneral isolate. Although it is to be expected that binding characteristics may varybetween stages, determining which stages bind microtubules was beyond the scope of thisstudy.In addition to elongate spermatids, the crude isolate contained round spermatids,nuclei, organelles, and other cell components (Fig 3-6: f). Differential centrifugationof the crude isolate, on a 30/45/60% sucrose gradient at high speed: 100,000g for 45minutes (Beckman Ti50 rotor, 36,000 rpm) or (Beckman SW4O rotor, 27,000 rpm)and very low speed (5000 rpm for 5 minutes), all yielded an elongate spermatid-ESenriched isolate that accumulated mainly at the 30/45 interface. Figure 3-16 showsspermatid enrichment from a crude spermatid isolate before (a) and after (b,c,d) usingspin conditions ultimately chosen for the binding assays. Much of the cell debris hasbeen removed.When spermatids are mechanically removed from the seminiferous epithelium,the Sertoli cell ESs remain attached (Masri et al., 1987; Romrell and Ross, 1979;Grove and yogI, 1989). For the binding assay, it was necessary to verify that, during thesquash method of spermatid-ES isolation, ESs remain attached and intact.In ESs, actin consists of a highly stable and ordered array of filaments thatencompass all but the most posterior portion of the spermatid head. It has been wellestablished spermatid heads stain intensely with rhodamine phalloidin, a stain forfilamentous actin, only when ESs are present (yogI et al., 1986). Isolated spermatidsdemonstrated the bright actin staining surrounding spermatid heads, indicating thepresence of ESs (Fig. 3-7 ). The ES actin staining pattern did not coincide with that ofthe spermatid nucleus but provided a wider sillhoutte, accommodating the acrosome, thatdid not extend the full length of the nucleus posteriorly, (compare the outline of theBinding assay 200Figure 3-7. Actin staining associated with the heads of isolatedspermatids: Spermatid-ESs. Spermatids, isolated by the squash method, stain foractin with rhodamine phalloidin in a pattern consistent with the location of ESs.Spermatids, shown here, were isolated by the squash method, (under warm conditions)and centrifuged on a sucrose gradient to give an enriched spermatid-ES preparation.Phase (a,c,f) and fluorescence (b,d,e,g,h) images of spermatids stained with rhodaminephalloidin to identify filamentous actin (b,d,g). Actin staining (b) surrounds thespermatid head but does not extend the full length of the nucleus posteriorly. A dottedline superimposed on the phase image in (a) indicates the location of the ES identified byactin staining in (b). (Note the difference in position between the caudal end of thespermatid nucleus and the actin staining of the ES). These data indicate that ESs remainattached to spermatids when they are isolated by this method.(C) Phase contrast image of an isolated spermatid-ES, dual stained with rhodaminephalloidin for filamentous actin (d) and 5A6 antibody for tubulin (e) showing thatendogenous tubulin is present but differs in distribution from that of actin. The tubulinpositive strand extending from the spermatid head is a common finding in warm isolatedspermatid-ESs (double arrows) and is not seen in the actin staining. Spermatid-ESs atthis stage, a time when tubulobulbar processes are forming but actin staining givesevidence of presence of ESs, are at the apex of the epithelium, and may retain elements ofthe apical process during isolation.(f) Spermatozoa or very late stage spermatid, identified in the phase image by its highlycondensed and acutely curved nucleus, negative for both (g) actin and (h) tubulinstaining. This is consistent with the fact that ESs are no longer present on late stagespermatids at the time of spermiation and that the actin and tubulin staining are notstaining structures on the spermatid itself.bars: a-h = 10 jim.10Binding assay 202spermatid nucleus, indicated by black dots, with the actin staining pattern (Fig.3-7a,b).As expected, spermatids at a stage immediately prior to spermiation, identified by theirhighly curved heads, do not stain intensely for actin because the ESs have degraded by thetime spermatids are released (Fig. 3-7 f-h).Ectoplasmic specialization endoplasmic reticulum: ESERThe outer face of the ER is not readily removed from the actin component. In anattempt to establish conditions which would ensure that the inner, cytoplasmic face ofthe ESER was not lost, a sucrose containing buffer was used. On phase images the headsappeared to have some material around the heads. DIOC6, a stain initially thought to beER specific, stained all membranes and organelles including mitochondriaindiscriminately, proving a poor tool to identify ESER. For EM of these spermatids, thetissue was pelleted in an Eppindorf centrifuge at every step of preparation. Loosespermatids were lost during wash steps and those in the pellet condensed with adjacentmaterial around the head, making it more difficult to demonstrate the full shape of theER. Spermatid-ESs, isolated in 100 mM sucrose PEM buffer showed some dilation ofthe ER when processed for EM (Fig. 3-8 a,b). As a result of the prolongedcentrifugation required when preparing the 250 mM sucrose, material becamecollapsed, and the ER was flattened, leaving only thin profiles of ESER visible (Fig. 3-8c).Endogenous microtubules remain attached during warm isolationIsolation temperature used during spermatid-ES isolation is important to thepresence of endogenous microtubules. “Warm isolation” was done at room temperature,and kept somewhat warm by the microscope lamp. Ice was not used in any step. Phaseimages of spermatids isolated under warm conditons revealed long strands extending,from the anterior and posterior aspects of the spermatid head, parallel to the long axis ofthe spermatid. Figure 3-9 shows the phase appearance and anti-tubulin antibody 5A6Binding assay 203staining of isolated spermatids using the warm (a-h) and cold isolation (i-I) methods.(Controls for this antibody are shown in Figs. 3-10 and 3-11). Long strands, seenassociated with the heads of the warm isolated spermatids, were positive for tubulinstaining, suggesting that endogenous microtubules remain attached to the heads with thisisolation.For “cold spermatid-ES isolation”, every attempt was made to keep the isolatecold throughout the proceedure. The testis was placed on ice immediately after beingremoved from the animal, and tubule segments were teased apart in ice cold buffer. Theslide used on the microscope stage to poke and squash the tubule segments was cold andwas replaced on ice as necessary, during the procedure. The isolated spermatids wereimmediately rinsed with ice cold buffer and kept cold in a collecting dish until used(methods Fig. 3:3). Following cold isolation, the long strands were generally absent andthe staining with tubulin was greatly reduced or absent. Occasionally, some endogenousmicrotubules remained. For the radiolabelled binding assays in this study, allspermatid-ES isolations were done using the cold isolation method to reduce boundendogenous microtubules. These data suggest that endogenous microtubules, under warmisolation conditions, remain associated with the spermatid head and support thehypothesis that microtubules bind to spermatids.To explore the relationship of the tubulin staining strands with actin stainingESs, isolated spermatids (warm isolation) were dual stained with rhodamine phalloidinand 5A6 (anti-tubulin antibody with FITC 2° mouse lgG probe). Staining indicated thatboth tubulin and actin were present but formed different patterns. Figure 3-7: c,d,eshow an isolated spermatid-ES stained for both actin and tubulin. Tubulin stain extends,as do the strands seen in the phase image, beyond the confines of the ES, delineated byactin staining. Late stage spermatids and spermatozoa lack both actin and tubulinstaining, consistent with the loss of ESs just prior to spermiation (Fig. 3-7 f,g,h). Theacutely curved head and lack of tubulobulbar processes, of this image, suggests that thespermatid/spermatozoa has been isolated immediately before, or after, spermiation.Binding assay 204Figure 3-8. ESER remains attached to spermatid head followingspermatid-ES isolation. Electron micrographs of the fine structure of spermatidESs that have been isolated in ice cold PEM containing sucrose. The ER remains attachedto spermatids although it remains somewhat dilated when PEM with 100 mM sucroseconcentration is used.(a, b) ESER of spermatids isolated in PEM with 100 mM sucrose (PEM 100). ESER(arrows) remains attached to the actin component of the ectoplasmic specialization(arrowheads).(c) Cistern of ER surrounding spermatid heads (arrows) following isolation in PEMwith 250 mM sucrose (PEM 250).bars: a= 1 jim , b = 0.5 jim , c = 0.5 jim.Lua4w:..,:.Binding assay 206Figure 3-9. Tubulin staining pattern on spermatid-ESs isolated underwarm or cold conditions. Spermatid-ESs, when isolated under warm conditions,retain endogenous tubulin in strands associated with the spermatid head.(a,b; c,d; e,f; g,h) Phase contrast images and tubulin staining pattern of spermatid ESsisolated under warm conditions.(i,j; k,l) Phase contrast images and tubulin staining pattern following isolation ofspermatid-ESs under cold conditions. When spermatid-ES isolation is carried out onice, the tubulin staining pattern is altered or lost.bars: a-I = 10 jim.207IBinding assay 208Figure 3-10. Control figure for 5A6 antibody staining for tubulin (1).Spermatid-ESs in these micrographs isolated under warm conditions (endogenoustubulin only). These micrographs show phase contrast and fluorescence images ofsamples spermatid-ESs stained for tubulin using 10 monoclonal antibody to alphatubulin 5A6 and FITC conjugated goat- anti mouse 2° antibody (lgG). All 10 and 2°antibody buffers contained 1% normal goat serum and 0.1% bovine serum albumin inphosphate buffered saline. Purified mouse lgG was reacted with testis sections processedin the same way as the spermatid-ESs shown in this figure and there was no evidence ofnon specific staining seen. The 5A6 tubulin antibody consistently stained both Sertolicell and exogenous bovine brain microtubules.(a,b) Immunocytochemical staining with 1° monoclonal antibody 5A6 for alpha tubulinand FITC congugated goat anti mouse 2° antibody.(c, d; e, t) 1° antibody is omitted from buffer.(g,h) 1° and 2° antibodies are omitted from buffer(i,j) 10 antibody is replaced with normal mouse serum. Remnant of tissue is included toshow that non-specific staining of tissue does not occurbars: a-j = 10 urn‘M 4209a\0Binding assay 2 11Figure 3-11. Control figures for 5A6 antibody staining for tubulin (2).(a,b; i,j) Phase and fluorescence images of spermatid-ESs isolated under warmconditions (endogenous tubulin only) stained with 5A6 monoclonal antibody to alphatubulin using the same spermatid-ES isolation conditions as shown in Fig. 3-10.(c,d; e,f) Tublin block of 5A6 antibody. No staining was seen when excess tubulin wasadded to the 5A6 antibody before reacting with spermatids.(g,h) Tuji monocloncal antibody to beta tubulin (gift from David Brown). Stainingpattern for Tuji antibody to beta tubulin was the same as for 5A6 antibody to alphatubulin.bars: a-j = 10 jim.®•,rIIIIbIIpb_L_/2.12.Binding assay 21 3Microtubule isolation and measurementsTubulin that is polymerized in the presence of taxol yields short stablemicrotubules (Schiff et al., 1979; Horowitz et al., 1981; Vallee, 1982). Figure 3-12shows taxol polymerized bovine brain microtubules, an unlabelled equivalent of thoseused in the binding assays. Negative stain EM highlights the protofilament structure ofmicrotubules (Fig 3-12 e). Of the 103 microtubules measured, over 14 fields, themean length was 2.8 ±1.2 jim. The length of taxol stabilized microtubules measuredhere would be approximately 15 to 20% of the length of the spermatid head. Tubulinstaining associated with spermatids removed from the sucrose gradient generallyconformed to the shape of the spermatid head. They were never seen to form the longstrands seen in spermatids isolated under warm conditions. In a “topical binding”experiment in which an excess of additional cytoplasmic dynein enriched MAPs, and taxolpolymerized microtubules were added to isolated spermatid-ESs directly on the slide,the anti-tubulin stained microtubules associated with a spermatid can be seen as shortstrands (Fig. 3-12: b,c).Labelling exogenous microtubulesTaxol reduces the critical concentration for microtubule assembly and stabilizesthe polymer to depolymerizing conditions of cold or microtubule disrupting agents(Schiff et al, 1979, Horowitz et al., 1981). In the presence of taxol, microtubulepolymerization is able to proceed without added nucleotides (Vallee 1982; Paschal et aI.,1991). Microtubules, for binding assays, were assembled from purified brain tubulinin the presence of a substochiometric amount of GTP (5 jiM) plus 3HGTP (1 jiCi/5Ojiltubulin) and 20 jiM taxol. This ensured maximum incorporation of competent3HGTPinto exogenous microtubules before they were added to the spermatid-ES preparation.Binding assay 21 4Figure 3-12. Length and structure of taxol stabilized microtubules usedfor MT - Spermatid-ES. MAP-free microtubules assembled from purified bovinebrain tubulin (1.6 mg/mI) for 30 minutes at 37°C in the presence of 20 jiM taxol. Theresulting polymerized and taxol stabilized microtubules were stained with 1% aqueousuranyl acetate on parlodian and carbon coated copper EM grids, and visualized on aPhillips 300 at 80 kv. Fields of positively stained microtubules were photographed,(including consecutive fields to encompass the longer microtubules) and microtubulelengths counted from the negatives.(a, d) EM micrographs of positively stained taxol polymerized microtubule images suchas those from which microtubule lengths were determined.(b, c) Phase and 5A6 anti tubulin fluorescence micrographs of a cold isolatedspermatid-ES, allowed to adhere to a polylysine coated slide and then incubated withtaxol stabilized microtubules in the presence of a cytoplasmic dynein enriched MAPpreparation from rat testis. (See ‘topical binding assay’ in methods). The shortexogenous microtubules which extend from the spermatid head are seen to particularadvantage in this figure (arrow). The short taxol assembled microtubules, in thisfigure, contrast with the long tubulin positive strands of endogenous microtubulesassociated with the head of the spermatids isolated under warm conditions.(e) EM image of negatively stained taxol polymerized microtubules showing the tinestrands of tubulin protofilaments (small arrows). bars: a and d = 1 jim, b and c =lOjim, e = 0.ljim.2.15*4Binding assay 2 1 6Incorporation of label and stability of microtubules over timeThe percent of total label that pelleted with microtubules (sample pelleted over a15% sucrose cushion in Beckman airfuge at maximum speed for 30 minutes), was takento be the percent incorporation of label. In untreated, taxol stabilized microtubules,prepared for binding assays, this was 43 ± 9.2%. Taxol stabilization prevents cyclingof the label from the exogenous microtubules (Wilson and coworkers, 1985). Thequantity of label present in microtubules, aliquoted from the same stock, pelleted fromsamples collected beginning at time = 0 hrs; or time = 2 hrs remained constant.Furthermore, an HapparentH decline in binding over the duration of the experimentwould have been expected if the label was being cycled from bound exogenousmicrotubules; “apparent” because it would be label, not microtubules that would bedeclining. No such effect was observed.Microtubule stability: effect of aaents for characterization ofmicrotubule bindingTo ensure that changes in microtubule- spermatid-ES binding were not due todirect microtubule injury by the respective agents, incorporation of label was measuredin microtubules exposed to agents used to characterize binding. The incorporation oflabel in all of the agents was within 98% ± 8% of the incorporation of the untreatedmicrotubules in same spin controls. This was well within the variability of samplingfor pelleted microtubules.BINDING ASSAY: ESTABLISHING THE CRITERIAIn a series of “topical binding” experiments, immunohistochemistry was used toexamine binding of exogenous microtubules to isolated spermatids that had been adheredto glass slides. Using this method, the results were difficult to interpret because someendogenous microtubules remained and observations could not be quantitated. TheseBinding assay 21 7experiments indicated that, to characterize spermatid-microtubule binding, anunambiguous method using labelled exogenous microtubules was required to characterizespermatid-ES-microtubule binding.Buffer conditionsAs described above, a number of buffer conditions were examined to protect theESER from hypo-osmolar etfects. PEM with 250 mM sucrose was deemed to be ofsufficient osmolarity to isolate spermatid-ESs with the ESER intact. Protease inhibitorswere added to protect protease sensitive MAPs that might be present and required forbinding (Wiche, 1989).When crude spermatid-ES isolate was placed on a discontinuous sucrose gradientincluding 30, 45 and 60% sucrose, and spun at 92000g for 45 minutes (4°C or 29°C),spermatids-ESs were primarily found at the 30/45 interface, with a minor amount atthe 45/60 interface. There was no discernable difference in the stages of spermatidspresent at the 30/45 and the 45/60 interface; however, the occasional cluster ofspermatid heads and a rare fragment of tubule wall were found at the lower level.Inspection of material from the bottom of the gradients indicated that material did notmigrate through the 60% sucrose. Using SDS PAGE gel electrophoresis, it wasdetermined that during high speed spins taxol stabilized microtubules were present atall levels of the gradient, with the greatest concentration at the 30/45 interface.The need for spin parameters in which spermatids would consistently enter thegradient, while unbound microtubules would not, was satisfied by a short duration, lowspeed spin. Figure 3-13 is a SDS PAGE gel of fraction samples from gradients loadedwith spermatid-ESs, spermatid-ESs incubated with 3MT, and 3MT alone, spun at5,000 rpm for 5 minutes. It was determined that any increase in speed or time (over5,000 rpm for 5 minutes) allowed taxol stabilized microtubules to enter the gradient(Fig. 3-14 and 3-15). Spins of 5,000 rpm for 5 minutes did not. Figure 3-14(bottom) shows an SDS PAGE gel of spermatid-ESs + 3MT samples using threeBinding assay 21 8different spin conditions. The 5,000 rpm for 5 minutes provided a maximum amount oftissue at the 30/45 interface.Using the ultracentrifuge with Beckman SW65 rotor, spermatid-ESs migratedinto the gradient, collecting mainly at the 30/45 interface with only a minor amountcollecting at the 45/60 interface (Figs. 3-13 and 3-14). Although the force used wasavailable from other centrifuges, precise control of the acceleration and deccelerationtime, duration and total number of cycles available in the ultracentrifuge were criticalto consistency between spins and maintenance of the criteria: 3MTx - spermatid-ESsenter the gradient and 3MT alone do not. To that end, the same centrifuge was used forall binding assay experiments. Using the SW65 Beckman rotor with adaptors for asucrose gradient capacity of 550ii1 and a sample load of 150 jil accommodated theproblem of small yield from spermatid-ES isolation. Although results from the 30/45and 45/60 interface were eventually pooled, the two levels of sucrose were countedseparately as an internal check. The count in the 45/60 fraction was always very smallcompared with that in the 30/45 fraction. Use of the three levels of sucrose meant thatthe occasional error in the construction of the gradient could be caught on visualinspection and the data deleted. Such errors were also obvious in the fraction counts.Spermatid-ES enrichment by gradientsThe primary purpose for putting the sample for the binding assay on a sucrosegradient was to ‘wash out’ unbound microtubules. It simultaneously had the effect ofenriching the fraction for spermatid-ESs, leaving much of the other cellular materialbehind. Figure 3-16 shows the composition of a spermatid isolate before being applied(Fig. 3-16a) and after being recovered (Fig. 3-16 b,c,d) from the gradient.Binding assay 21 9Figure 3-13. Testing for microtubule migration using sucrose gradientsin MTx - spermatid-ES binding assays.Spermatid-ESs, MT- spermatid-ESs, and MT alone, were centrifuged ondiscontinuous 30, 45, and 60 % sucrose gradients at 5,000 rpm for 5 minutes(Beckman Ultracentrifuge model L7-55). Samples were prepared from tissuerecovered from the 45/60, 30/45, and sample/30% sucrose interfaces and loaded onto7.5% SDS polyacrylamide gels. Gels were stained with comassie blue stain.(Lane 1) Molecular weight markers (carbonic anhydrase 29 kD, egg albumin 45 kD,bovine albumin 66 kD, phosphorylase B 97.4 kD, beta galactosidase 116 kD, and Myosin205 kD.)(Electrophoresis is from top to bottom of gel).(Lanes 2, 3, and 4) Isolated spermatid-ES samples, recovered from 45/60, 30/45 andsample/30 % sucrose gradient interfaces respectively.(Lanes 5, 6, and 7) Isolated spermatids, to which taxol polymerized bovine brainmicrotubules had been added; from 45/60, 30/45 and sample/30 % sucrose gradientinterfaces respectively. Arrow indicates tubulin in lane 7. (Lanes 8, 9, and 10) Taxolpolymerized microtubules only, recovered from 45/60, 30/45 and sample/30 %sucrose gradient interfaces respectively.Proteins from spermatid-ESs collect mainly at the 30/45 interfaces (lanes 3 and 6).Tubulin is also present in these fractions, possibly darker in lane 6. Tubulin is presentin the 30%/45% fraction in the spermatid-ES samples when MTs are added (lane 7)but not in 30%/45% fractions when the spermatid-ES samples do not have added MTs(lane 4). There was no tubulin found to enter the gradient, that is, in the 30/45 or45/60 fractions, when spermatid-ESs were not present. Tubulin loaded onto thegradient without spermatids (lanes 8, 9 and 10) remains in sample/30 fraction. Thesedata indicated that MTs do not enter the gradient, except in the presence of spermatidESs, under these spin conditions.IISi30 45’60 30/451 2 3 4 5 6 7 8 9 101‘Z2O.— tub ulinS /30IM 45/60 30/45•-iiIi*iSf30 45/60 30/45iiIm icrotu bules-IIsp erma tids spe rm at ids+microt u buleSBinding assay 221Figure 3-14. Establishing spin criteria for binding assay: SOS page gels:Gel data from a variety of the spin parameters tested to determine optimum parametersfor maximum migration of MTx - spermatid-ES5 but without migration of MT alone.The format and lane assignments for the four gels in the upper part of this plate areexactly the same as those shown in the previous figure.(upper left and upper right) Gels used to test the contents of gradient fractions underspin conditions ultimately used in the binding assays (in which MT- spermatid-ESssamples were spun at 5000 rpm for 5 minutes). Spin criteria of requiring thatunbound microtubules do not migrate into gradient fractions, are met with theseparameters.(middle left) Gel testing spin parameters, 15000 rpm, 5 minutes.(middle right) Gel testing spin parameters, 5000 rpm for 30 minutes.At 5000 rpm for 5 minutes, tubulin was present only in the sample/30% fraction (lane10). At both 15000 rpm for 5 minutes and 5000 rpm for 30 minutes tubulin waspresent in both the 30%/45% (middle left, lane 9 and middle right lane 9) andsample/30% fractions (middle left, lane 10 and middle right lane 10). These dataindicated that the spin conditions of 5,000 rpm for 5 minutes fit the criteria thatspermatid-ES enter the gradient while microtubules alone do not.(lower gel) SDS PAGE gel loaded with gradient fractions from three different spermatidES + MTx binding assays.To compare the amounts of MT-spermatid-ESs that enter the gradient using these spinconditions, the 45/60, 30/45 and sample/30 interfaces were collected from each assay.(lanes 2,3,4) 45/60, 30/45 and sample/30 interfaces from assay spun at 5,000 rpmfor 30 minutes. (lanes 5,6,7) 45/60, 30/45 and sample/30 interfaces from assayspun at 15,000 rpm for 5 minutes. (lanes 7,8,9) 45/60, 30/45 and sample/30interfaces from assay spun at 5,000 rpm for 5 minutes. The majority of MTspermatid-ESs were found in the 30/45 interface, when using the parameters: 5000rpm for 5 minutes. The 5000 rpm for 5 minutes satisfied two criteria: adequate entryof MT bound spermatid-ESs and no entry of MT alone. These data were consistent withradiolabelled data.5,000 rpm, 5 mm1 3 4 5 6 7 8 9 104>15,000 rpm, 5 mm3 4 6 7 8 9101.4.5,000 rpm, 30 mm>1 2 3 4 5 6 7 8 9 10‘lQ.LS2 3 4 5 6 7 8 9 1011 I>I II III II II IILii_i i5K30’ 15K5’ 5K5’Figure 3-15. Test of differential centrifugation conditions for bindingassay using 3MT counts. The previous figure provides SDS gel evidence thatmicrotubule protein is not detected in the gradient in the absence of spermatid-ESs.This data gives evidence that counts, associated with microtubules are not detected in thegradient when loaded without spermatid-ESs. A number of different centrifugationconditions were tested which could provide for migration of labelled microtubules boundto spermatid-ESs and at the same time not result in the migration of unboundmicrotubules. Data are give for three of the conditions tested and expressed as a percentof the total label that entered the gradient using 15,000 (15K5’), 10,000 (10K5’) and5,000 rpm (5K5’) for 5 minutes for each spin. One sample each of: labelledmicrotubules + buffer (3MTx); 3MTx + isolated spermatids (no treatment control);3MTx-spermatid-ESs + 10mM MgATP (+10mM ATP), was included in each spin.Centrifugation at 5,000 rpm for 5 minutes met the required criteria. (* indicates 0.0%counts entering gradient, which do not register on the graph.) This illustrates theimportance of using spin criteria that eliminate the counts from microtubules enteringthe gradient.Binding assay4-C004-04-•1-04-C00.3020100223• 3MTx + bufferno treatment+ 1OmMATPexp 1 exp 2a exp 2b15K 5’ 10K 5’ 5K 5224Controls: “same spin controls” in every spinThe Beckman SW65 rotor, has only three buckets, limiting the number ofconditions that can be spun together to three. An untreated control (‘same spin control’)was included in each spin. Sample size and therefore ‘total counts loaded’ wereequivalent, in all gradients for a given experiment (exceptions are noted). Figure 3-17, 3-18 and Table Ill-I show a short experiment that is illustrative of the generalexperimental design, results, and method of expressing data. Measures of 3MTx-spermatid-ESs entering the gradient are combined counts from fractions 30/45 and45/60, expressed as the percent of total counts from all three fractions: ‘percent oftotal count’. To examine the treatment effect of an agent, it was necessary to factor out,between-experiment variability and variability caused by time. These effects would bereflected in ‘no treatment control’ spins. For that data, each count from a treated groupwas expressed as a fraction of the count from the ‘no treatment control’ from the samespin: ‘fraction of same spin control’.Initially, one gradient with 3MTX and buffer only, was included in each run tocheck for migration of microtubules alone. Consistently, no label was found in thesegradients. This was then checked only periodically. None of the treatment agents used tocharacterize 3MTx - spermatid-ES binding caused migration of 3MTX and buffer alone.Epididymal sperm, with or without added microtubules, did not migrate in thesame way, on the sucrose gradient, as spermatid-ESs, making them inappropriate as acontrol.SamplingHigh variablity was encountered when spermatid-ESs were sampled using thetop, side or bottom access to the gradients. Tissue was also damaged passing through theneedle (see discussion of autoradiography experiments). Contamination of labeloccurred and not all the material was recovered. Gradients were fast frozen and cut intothree fractions (mid way between S/30 and 30/45 and between 30/45 and 45/60Binding assay 225Figure 3-16. Centrifugation of crude spermatid-ES isolate on adiscontinuous 30-45-60% sucrose gradient results in enrichment ofspermatid-ESs. DIC appearance of spermatid-ESs isolate (without added tubulin)before and after being enriched on a sucrose gradient, using the same conditions as thebinding assay.(a) DIG appearance of spermatid-ES isolate prior to being enriched by a sucrosegradient.(b, c and d) DIC appearance of the spermatid-ES isolate shown in (a) after beingenriched on the sucrose gradient. Samples shown here were collected by syringe fromthe 30/45 % sucrose interface and are unfixed. Figures (b and c) show typicalappearance of isolated material. Figure (d) shows clusters of cellular material that isseen in some fields. In (d) spermatid heads can be seen in the cluster.bars: a-d = 20pm.‘9—Binding assay 227Figure 3-17. Binding assay: method: short sample experiment. Design of ashort experiment that typifies the binding assay experiments used throughout this study.3MTx were combined with isolated spermatid-ESs (1:2) to give 3MT - spermatid-ESs,or with buffer (1:2) to give 3MTx+ buffer. The 3MT - spermatid-ESs were furtherdivided into ‘no treatment control’ and 3MT - spermatid-ES with added 10 mM MgATPgroups. Gradients were prepared immediately before each spin. Three conditions weretested in each spin. A ‘no treatment control’ was used in every spin. Initially a 3MTxwith buffer was included in each spin, however, with the repeated result that 3MTx didnot enter the gradient, this condition was replaced by other treatments in laterexperiments. A typical ‘treatment’, 10 mM MgATP, was used in this experiment.(Early vs late binding) For ‘early’ binding experiments, the 3MT - spermatid-ESswere loaded immediately onto the gradient for the first spin. For established binding,3MTX - spermatid-ESs binding was allowed to proceed for at least 1 hour before beingassayed. Binding in this experiment is early binding.(Early vs late treatmeni For early treatment, the 10 mM MgATP was added at the sametime as the 3MT and the first spin begun as quickly as possible. When treatment wasimposed on established binding, treatment was added to 3MT - spermatid-ESs the afterthe established binding period. Treatment in this experiment is on early binding.Spermatid-ESsExperimental Design22S3MT÷ bufferNo treatmentcontrol3MTX- spermatid-ESs(no treatment control)3MTX- spermatid-ESs+ 10mM MgATPBinding assay 229Tab. Ill-I. Sample data from binding assayspin no treatment 10 mM MgATP 3MTx + buffer % same spincontrol1 7.0 1.6 0.0 232 7.2 1.6 0.0 223 8.1 1.4 0.0 174 8.2 1.5 0.0 18This short binding assay consisted of four spins having one gradient of each condition:3MT - spermatid-ESs only (untreated), 3MTx - spermatid-ESs with 10 mM MgATP,and 3MT with buffer alone, in each spin. Samples were incubated for 30 minutesbefore first spin to allow time for 3MT - spermatid-ES binding. Gradients were spunat 5,000 rpm for 5 minutes. Data represents counts that were collectively found in30/45 and 45/60 gradient fractions, expressed as a percent of total counts loaded ontogradient. Each value from the 10 mM MgATP group is expressed as a fraction of thesame spin control (column % same spin control).Binding assay 2303MTx- spermatid-ES binding10•.5spinFigure 3-18. Results of sample binding assay. 3MT with buffer only (opencircle), 3MT with spermatid-ESs treated with 10 mM Mg ATP (open square) and samespin controls (solid square) are shown as a function of spin number and expressed as apercent of same spin control (See Table Ill-I). 3MT with buffer only do not enter thegradient. 10 mM MgATP significantly reduces binding.231interfaces) with a razor blade to circumvent these sampling inconsistencies. Becauserepeated freezing and thawing of gradients is an accepted method to produce continuousgradients, gradients of different colors were processed to see if the freezing caused anymixing of the gradient levels. No mixing occured. Problems frequently encountered inscintillation counting, chemiluminesence and interaction of gradient components withscintillation fluid, also were checked.BINDING ASSAY: RESULTSMicrotubules bind to spermatid-ESsThe first evidence of spermatid-ES-3MT binding, in the assays, was thepresence of counts in gradient fractions using conditions in which 3MT alone could notenter the gradient. Figures 3-17, 3-18 and Table Ill-I show the data from a shortexperiment in which the effects of: 3MT+ buffer; spermatid-ES + 3MT (notreatment); and spermatid-ES + 3MTX +10 mM MgATP are compared. In thisexperiment, 3MT and spermatid-ESs were incubated for 30 minutes before the spinswere started. 3MT did not enter the gradient alone as shown by the 0% counts in thegradient. Spermatid-ESs entered the gradient and with them, radioactive counts,evidence of labelled microtubules. Assuming that the label represents microtubules, andthat the binding is to spermatid-ES5 (the major component of the gradient fraction), twofindings central to this study are supported by these data: 1) Binding occurs betweenmicrotubules and spermatid-ESs and 2) this binding can be reversed by treatment with10mM MgATP. The assumption that the label represents microtubules is examined first.Label entering the gradient with sperimatid-ESs is primarily frommicrotubulesIn order to equate counts entering the gradient with microtubules bound tospermatid-ESs, it was necessary to determine what percent of the counts present in theBinding assay 232gradient could be attributed to microtubule and non microtubule associated label. Thereasoning was that, if all microtubules were removed from the 3MT sample bypelleting, the remaining label would constitute non microtubule associated label in thesample. 1200 tl of 3MTx sample were spun over a PEM15% sucrose cushion at50,000 rpm for 30 minutes and the top 900 uI were recovered as ‘3MT supernatant’(3MTxSN). The percent of count remaining in the supernatant after pelleting averaged52.5%, suggesting that at least 47.5 % of the label was incorporated into microtubulesof sufficient size to be pelleted (tubulin concentration was reduced by 90% bypelleting). This was consistent with microtubule incorporation values generally. The3MTXSN was then substituted for 3MTx. Controls were 3MT+ buffer and 3MT-spermatid-ESs. Removing labelled microtubules from 3MTx sample has the effect oflowering the total counts loaded onto the gradients. For these experiments, binding wastherefore expressed as ‘absolute counts entering the gradient’ (rather than % of totalcounts). Table Ill-Il summarizes the results of these experiments. These data indicatethat the majority of counts entering the gradient are associated with labelledm icrotubu le s.Figure 3-19 shows the microtubule and non microtubule components of bindingover time. There is an increase in counts entering the gradient over the first two timepoints of sequential spins. Calculations show that in both these points, nonmicrotubulecomponent is 25% of microtubule component. This raised the question of a time coursein binding.There is a time course to 3MTx - spermatid-ES bindingWhen the data from ‘no treatment controls’ from each experiment were expressedover time, (plotted on a composite graph), a pattern of binding emerged that wastypically lower at early time points before achieving a plateau. In order to define thistime course, count data from ‘no treatment controls’ of seven experiments, matched inBinding assay 233their time courses, were plotted over time (Fig 3-20). In these experiments, 3MTx andspermatid-ESs were mixed (time = 0 minutes), immediately placed on the sucrosegradient, and the first spin begun (early binding). Data from the spins were found to bedifferent, by analysis of variance, when grouped into either 15 or 30 minute intervals(P < 0.05) (Fig 3-21). Pairwise comparisons using Tukey’s test showed that the first15 minute interval was different from all but the second, and the second was differentfrom most but not all of the others. When they were grouped into 30 minute intervals,the first interval was different from the others and none of the others were differentfrom each other. This indicates that there is a time course to the binding and with lowerbinding in the first 30 minutes and a plateau occuring by approximately 60 minutes.This time course pattern was found to be consistent with a composite scatter plot of theother experiments in the binding study that had included measures during early binding.Nucleotides reverse bindingSome MAPs can be released from microtubules with nucleotides, a fact that isutilized routinely for the isolation of MAPs (Wiche, 1989), in particularmechanoenzymes.The effect of 10mM MgATP on 3MTx - spermatid-ES binding was tested in 20experiments. In some experiments 10 mM MgATP was a prime focus of the experiment(one of the two treatment conditions being examined). In the other experiments it wastested episodically to measure ATP release on other agents under study. The meanreduction of binding by 1 0mM MgATP was determined for each experiment. The methodof expressing the binding is shown in table Ill-I. Counts (combined 30/45 and 45/60fractions) are expressed as percent of total count. In order to compare the effect oftreatments, such as ATP, in a number of different experiments, the percent of totalcount is then expressed as a ratio of “no treatment, same spin control” values. Figure3-18 and Table Ill-I show that 3MTx - spermatid-ES binding is reduced by the additionof 10 mM MgATP. Over 20 experiments, 10 mM Mg ATP reduced binding to 36 ± 13%04-04-C00)C04-C00Figure 3-19. Non-microtubule component of binding. Spermatid-ESs reactedwith 3MT (solid squares) or microtubule depleted 3MTx(3MTSN) (open squares). Aninitial increase in binding is seen in both3MTx-spermatid-ES and3MTSN-spermatid-ES binding. The average percent of binding attributed to microtubule component ofbinding was 75%.Binding assay100000234Non microtubule component of binding800006000040000200000• 1•1•1 I0 20 40 60 80 100 120 140timeBinding assay 235Table Ill-Il. Non-microtubule component of binding: summaryexperiment 1 experiment 2final tubulin concentration 0.53 mg/mI 0.27mg/miin control sample (1 x) (0.5X)percent of binding 25 ± 2 % 8 ± 2%accounted for bynonmicrotubule componentof sample incubated withspermatid- ESs*calculated percent of 75 ± 2 % 92 ± 2 %binding accounted for bymicrotubule component ofsample incubated withspermatid-ESspercent of label entering 0.0 % 0.0%the gradient accounted forby non microtubulecomponent in the absence ofspermatid-ESs+percent of label 48% 47%incorporated intomicrotubule component‘Non-microtubule component’ is the supernatant from the same labelled microtubulesample as that used for the control but first having been depleted of microtubules byhigh speed centrifugation substituted for the 3MT sample in spins with spermatidESs* or with buffer only+. Controls were 3MT + spermatid-ESs (‘no treatmentcontrol’) and 3MT + buffer. The percent of label accounted for by non-microtubulecomponent is the percent of binding in the microtubule depleted supernatant divided bythe percent of binding in the control sample. Percent incorporation of label intomicrotubules was calculated as counts removed by centritugation/original counts inmicrotubule sample before centrifugation. Results indicate that at least 75-92% ofcounts entering gradient can be attributed to the microtubule component of the label.Microtubule-depleted sample with buffer only (replacing spermatid-ESs) does notenter the gradient (no counts in gradient fractions).Binding assay 236Figure 3-20. Time course of 3MTx - spermatid-ES binding: overlay plotsof time course from matched experiments. The source of these data is the 3MT -spermatid-ES controls from seven experiments, matched with respect to microtubuleconcentration and time course of binding. 3MTx and isolated spermatid-ESs werecombined at time = 0 minutes. For each time point, the percent of counts entering thegradient is expressed as a ratio of the experiment mean. Each experiment has beenassigned a different symbol.237Binding assaya,a,Ea,Ea,‘aa,.0.Da,0a,0Time course of 3MTx - spermatid-ES binding1.501.251.00•0.75O.500.25 -0.000I— I60 120 180time (mm)240Binding assay 238Figure 3-21. Time course of 3MTx - spermatid-ES binding: collapseddata. Values from Fig. 3-20 shown above have been collapsed into 15 minute intervals.Plot lines have been removed. Solid line indicates mean time line for the sevenexperiments. The time course for these seven matched experiments is consistent with asimilar composite graph for the time course for all of the binding assays done in thisstudy (not shown). There is a significant increase in binding during the first hour.1.50 -c 1.25SSE4- 1.00•Ca,EO.750xa,O.50025a,0S0. 0.00. -0Binding assay 239Time course of 3MTx - spermatid-ES binding (15mm intervals)xS.— .Sx •.30 60 90 120 150 180 210 240time (mm)240of same spin controls. Release appeared to be immediate (within the time constraints ofthis assay) and occured whether the ATP was added before or after binding has becomeestablished.Support for the binding assay evidence of release by 10 mM MgATP is shown inFigures 3-22 and 3-23. Parallel samples using 3MT or MT were incubated withspermatid-ESs, with or without 10 mM MgATP, run on gradients and processed forscintillation counting (labelled), or recovered with a needle and processed for antitubulin immunofluoresence (unlabelled). The anti-tublin staining pattern was presentin controls but essentially disappeared from MTx - spermatid-ES + 10 mM MgATPgroup. Unlike the antibody staining controls seen in Figures 3-10 and 3-11, there aresmall remnants of staining around the spermatid head, consistent with the findings of thebinding assays. The binding assay run parallel to the immunofluoresence showed areduction to 31% of controls in 10 mM MgATP treated group.5mM GTP is used to selectively release the mechanoenzyme, kinesin, frommicrotubules (Vallee, 1982). In five experiments, 5mM GTP reduced 3MT -spermatid-ES binding to 43 ± 21% of controls. This compares with 10mM MgATPrelease values of 36 ± 14%. When 10mM MgATPATP and 5mM GTP were compareddirectly in the same spins, the two effects were not significantly different.To address the possibility of the reconstitution of ATP from GTP by nucleosidediphosphokinase activity in which the phosphate from GTP is bound to ADP, GTP in thepresence of 18 units/mI hexokinase and 10 mg/mID glucose was tested. The binding wasnot different between the two groups.As described with the 10 mM MgATP group data, a number of factors may becontributing to the variability in nucleotide binding release encountered. However,when 10 mM MgATP and 5 mM GTP were combined, the binding was always reduced toless than half of the binding that remained after treatment with either nucleotide alone.In some way, the effect was additive. Figure 3-24 illustrates the comparison of 10 mMMgATP, 5 mMGTP and their combined effects.241Effect of 10mM M9ATP on 3MTx-spermatid-ES binding10.0-0U::0 100 200 300 400time (minutes)Figure 3-22. Effect of 10 mM MgATP on 3MTx - spermatid-ES binding.Reduction of binding following treatment with 10 mM MgATP (open squares). Treatmentwith 10 mM MgATP and the initiation of binding were both at time = 0 minutes.Microtubule concentration in final samples was 0.53 mg/mI. The typical time course ofbinding of the untreated group 3MTx - spermatid-ESs, resulting in an increase inbinding during the first hour, is seen in the control group (closed squares). In thisexperiment, reduced binding by ATP treatment is evident by the first spin (at 18minutes) and maintained. The mean ratio of 10 mM Mg ATP/same spin control valuesfor this experiment is 27.4 ± 10.5 that is, only 27% of the binding remains aftertreatment with 10 mM MgATP (first spin omitted in calculation to avoid time effect).Binding assay 242Figure 3-23. Reduction in tubulin staining on MT - Spermatid-ESsbinding after treatment with 10 mM MgATP. Paired DIG and fluorescencemicrographs of isolated spermatids incubated with exogenous microtubules (MTx -spermatid-ESs) that have been stained with 5A6 anti tubulin monoclonal antibody.(a,b) MT - spermatid-ESs untreated control(c,d; e,f; g,h) MTx - spermatid-ESs treated with 10 mM MgATP. Unlike the completeabsence of staining in controls for 5A6, these spermatids, treated with 10 mM MgATP,retain remnants of staining for tubulin in some cases (arrow), but staining is greatlyreduced compared with the untreated controls. This is consistent with the count datafrom binding assays.bars: a-h = 10 jim.in1NIBinding assay000•1-04-C0000.121086420Nucleotide effects on 3MTx - spermatid-ES bindingexperiment244Figure 3-24. Comparison of nucleotide effects on 3MTx - spermatid-ESbinding. Effects of 10 mM Mg ATP (solid bar), 5 mM GTP (diagonal hatched bar) and10 mM MgATP+5 mM GTP combined (fine hatched bar) in the same spins. These dataindicate that, under the conditions used in these binding assays, the two nucleotidescombined reduce 3MTx - spermatid-ES binding to a greater degree than either alone.1 2 3245The effect of ATP depletion on3L[-spermatid-ES bindingAnother aspect of the nucleotide question to be considered was that if ATP reducesbinding, will the complete depletion of ATP increase binding? Exogenous ATP was notadded to the ‘no treatment control groups’ to avoid a possible ATP induced release.However, low levels of ATP of the capacity to regenerate ATP might retained in thespermatid-ES isolate. Because the degree to which ATP was present and its change overtime could not monitored, an ATP depleting system in which hexokinase (18u/ml) and Dglucose (10mM) were used to exhaust endogenous ATP was added to the spermatid-ESisolate (and to the gradient solutions).Figure 3-25 shows 6 experiments in which the effect of ATP depletion wasexamined. When these 6 experiments are compared by analysis of variance, onlyexperiments 1 and 6 are significantly different from the controls. Although experiment5 appears to be reduced, due to the variability of measures, there is no statisticaldifference registered by analysis of variance. These results are consistent with thenotion that in some cases the residual nucleotide is low and in others, there is sufficientATP regenerating capability or residual ATP that there is a change in binding valueswhen an ATP depletion system is introduced. Again, the ATP question appears to becomplex. What is clear is that high concentrations of MgATP produce a release of bindingand complete depletion of ATP contributes to an increase in binding in somepreparations.Microtubule concentration has a nonlinear effect on bindingThe initial assumption of the binding was that it would be analogous to actinmyosin rigor and therefore not dynamic (not cycling), a condition one would expect in apurified motor-microtubule system in vitro. This was predicated on the finding thatATP is required for microtubule based, organelle transport in vivo and in vitro (Vale,1987; Cohn et al., 1987; Shimizu et al., 1991). Evidence presented so far suggests thatBinding assay 246Ca,0L.S 2000L.C00C0a,0.a,0.IFigure 3-25. ATP depletion effect on 3MTx - spermatid-ES binding.Effect of ATP depletion on 3MTx - spermatid-ES binding in six experiments. Hexokinaseand D glucose were added to buffers used in six 3MTx - spermatid-ES binding assayexperiments, to deplete any residual ATP. Analysis of variance between ATP depletedgroup and controls for the six experiments indicates that binding was significantlyincreased by ATP depletion in experiments 1* and 6*, but not significantly differentfrom controls in experiments 2-5. Line at 100% indicates control values. Althoughexperiment 5 would appear to be significantly reduced from the control value, analysisof variance of all the groups compared with the variance of no spin controls, indicatesthat it is not.250150100experiment2473MT-spermatid-ES binding is stable, with a time course that establishs a plateauofbinding and an increase in binding with the removal of ATP. If that is so, then it isreasonable to expect that the effect of varying the substrate concentration would resultin a linear increase in binding to a point of saturation, which would then plateau in spiteof further increases in substrate concentration. 3MTx- spermatid-ES binding does notfulfill that prediction.Some variability is to be expected in the spermatid composition and betweenisolations. When comparing binding in the presence of a fixed concentration of tubulinwhile varying the spermatid-ES concentration from the same isolate, there is a linearrelationship between spermatid-ES concentration and binding. This is illustrated inFigure 3-26. This leads to the assumption that, at these concentrations, an increase inavailable binding sites results in a proportional increase in binding.Based on the linear relationship of spermatid-ES concentration to binding at afixed microtubule concentration, if the number of binding sites was held constant(sampling from the same isolate) then a rigor-like binding would predict a linearrelationship between binding and microtubule concentration, until saturation ofavailable binding sites occured. Figure 3-5 shows the design of a study done to check therelationship between tubulin concentration and binding, using spermatid-ESs sampledfrom the same isolate. A number of precautions were taken. To ensure thatincorporation of label into microtubules was not influenced by microtubuleconcentration, label was incorporated at the lox concentration and all dilutions madefrom the same iox stock. Bindinwas allowed to become established before the firstspins were begun. The microtubule concentration experiment was designed to checkthree technical considerations: 1) do the microtubules retain their label over theduration of the experiment? 2) is the sampling of the tubulin concentrations reliable?3) do the counts reflect the tubulin concentration? The first question was addressed bycomparing the counts and their respective tubulin concentrations early and late in theexperiment. There was no loss of counts or microtubule concentration, verifying thatBinding assay64- 5,0o4-0.4- 3,•1-04-00L.00.0[spermat id-ES]248Figure 3-26. Effect of varying spermatid-ES concentration on 3MTx -spermatid-ES binding. Spermatid-ESs from the same isolation were diluted to 1/3and 2/3 of controls. To avoid an effect of time, binding was allowed to proceed for twohours before [spermatid-ES] effects were tested. Samples from each of the threeconcentrations were included in each spin.1/3 2/3 3/3249label does not cycle from the microtubules. This was consistent with previous findings(Wilson et al., 1985) and the known stability of taxol stabilized microtubules (Schiff etal., 1979; Horowitz et al., 1981; Vallee, 1982). Secondly, microtubule concentrationin stocks did not change with repeated sampling, although there was, understandably,more variability in sampling with the lowest concentrations. Thirdly, the counts fromboth the assays and the microtubule concentration determination portion of the studyvaried in direct proporation to the assumed tubulin concentration from the stocksolutions. Because dilutions were made after the label had been incorporated such thatthe counts loaded reflected the tubulin concentration being tested, the ‘total counts’ loadedon each gradient were (unlike the other assays) not equivalent. For this reason, bindingdata is expressed in absolute ‘counts’ rather than ‘percent of counts loaded’.The results of the binding assay using different microtubule concentrations areshown in Figs. 3-27 and 3-28. Binding increases with increasing tubulin concentrationin a non linear manner. The slope of binding vs microtubule concentration changes at lxconcentration (Fig.. 3-28). One of a number of possible explanations is that there maybe at least one low and one high affinity binding site being assayed. The binding does notshow evidence of saturation even at the very high microtubule concentrations used. Anattempt was made to explore the kinetics of this binding, plotting the bound/free versusbound and drawing a Scatchard plot to calculate the Km. It became evident that thissystem would need to be defined more precisely before the kinetics could be studied; notenough is known about the binding to make the assumptions required for kinetic analysis.These data suggest that the binding is not rigor-like; that turnover of binding may beoccuring. If turnover was occuring, the introduction of excess amounts of unlabelledmicrotubules would replace the labelled microtubules reducing apparent binding.3M.I.x - spermatid-ES binding is dynamicThe shape of the plot of substrate concentration vs binding suggested that 3MTx-spermatid-ES binding in this system, may not be static, but may be in dynamicBinding assay 25040000 -______________________________________________—s— lox30000 -—-——lx20000-_______oax10000lx [MT] = 0.53mg/mi0•. i • I • I • I • I0 2 4 6 8 10 12SPINFigure 3-27. Effect of varying microtubule concentration on 3MT -spermatid-ESs binding. Counts from combined 30/45 and 45/60 gradientfractions of MT..spermatid-ESs binding using six different microtubuleconcentrations. Radiolabelled microtubules were polymerized and stabilized with taxolat lOX concentration (10 X 1.6 mg/mI = 16.0 mg/mI), then diluted into 0.1 to loxsamples. (Samples loaded on the gradient were 3MT : 1OOil spermatid-ESs).Final concentrations in binding assay, ie. after adding spermatid-ESs, were 0.1 X to 10X (0.53 mg/mI). Microtubule concentration were assigned in varying order for twotubes, the third one being a 1X[3MT] sample that was included in each spin, to act as acontrol. Due to randomization, two samples of 0.5X were included in spin 3: the valueswere the same and they have been placed side by side. Data is expressed as counts thatwere recovered from the gradient. Counts loaded were proportional to the [3MT]because the [3MT] samples were diluted after the microtubules had been polymerizedand stabilized. Counts entering the gradient increased with increasing[3MTx].Binding assayU)IzD0C.)400003000020000100000[MT]251Figure 3-28. MT..spermatid-ESs binding expressed as a function ofmicrotubule concentration. With the concentration of spermatid-ESs heldconstant, 3MT - spermatid-ESs binding increases with the increase in [3MT]. Therate of increase is nonlinear and does not show saturation at microtubule concentrationsused.0 2 4 6 8 10 12Binding assay 252Table Ill-Ill. Effect of the addition of excess cold label: 5X cold taxolstabilized microtubules on 3MTX - spermatid-ES bindingcompetition by3MT concentration competition by dilution by displacement bycold 5X MTx cold 5X MTx1.OX 48±10% 46±3%0.5X 18± 5 % 24±2%To test the effect of competition of unlabelled microtubules (cold) with labelledmicrotubule-spermatid-ES binding, excess unlabelled microtubules (5Oiil/ gradientof 5X concentration of unlabelled microtubules MTx) were added either at the onset(dilution) or after established binding (displacement) of 5Ojil of 3MTx and 100 uIspermatid-ESs. 50 u’1 of buffer only were added to the controls to compensate for theaddition of of 3MT. 1X tubulin concentration = 1.6 mg/mI (before being addedto the spermatid-ESs). Competition of cold label at the onset of binding would have thesame effect as diluting the labelled microtubules at a time when binding sites areequally available to labelled and non labelled microtubules. Competition after theestablishment of binding (here over 1 hour) indicates the degree to which theunlabelled microtubules are able to displace already bound 3MT. These experimentsshow that microtubules bound to spermatid-ESs can be displaced, suggesting aturnover of binding.Table lll-IV. Effect of the non-microtubule component of 5X cold taxolstabilized microtubule sample on 3MTx - spermatid-ES bindingcompetition by3MT concentration competition by dilution by displacement bysupernatant from supernatant fromcold 5X MT cold 5X MT1.OX 83±8% 94±6%0.5X 68± 8 % 69±11%The purpose of this part of the study was to determine the effect of the microtubule andnon microtubule components of cold competition. Aliquots of non labelled 5Xmicrotubules were pelleted and the supernatant used as a microtubule-depletedcontrol for the experiments described above. These data indicate that it is primarilythe microtubule component of the excess cold label (5X cold MTx) that is responsiblefor the effect of competition to 3MT - spermatid-ESs binding.In both tables, figures quoted are percent of same spin control253equilibrium. To test this, binding (no treatment control) was compared with apparentbinding (competition by the addition of an excess of unlabelled microtubules). ATP wasnot added to the system, nor was it depleted: the isolate was the same as for ‘no treatmentcontrols’ of all the binding assays. If the excess of unlabelled microtubules are added atthe onset of binding, they simply represent a dilution of label by competing directly forthe available sites. The reasoning was: if unlabelled microtubules are added afterbinding has become established, the unlabelled microtubules will need to replace alreadybound labelled microtubules. A change in apparent binding would indicate turn over ofbinding. Eventually the number of labelled microtubules may approach that in thediluted group, but only if the system is dynamic. To avoid the effects of dilution ofspermatid-ES5 by the addition of microtubules, buffer, equal to the volume of addedunlabelled microtubules, was added to the controls. This has the effect of increasing themicrotubule concentration in the test sample while decreasing it in the control.However, the result of the minor changes in tubulin concentration, according to the datapresented above, would be to increase binding, the opposite effect expected if unlabelledmicrotubules compete with labelled microtubules. Table Ill-Ill shows the effects of coldcompetition of 5 fold concentration of microtubules added to compete with labelledmicrotubules. The binding was reduced following competition with added excessunlabelled microtubules. This would suggest that labelled microtubules had beenremoved, and replaced by unlabelled microtubules.If microtubules are responsible for the reduced binding by competition, thenremoving them from the competing sample should eliminate the competition effect. Ashad been done to measure the microtubule component in the labelled binding, theunlabelled sample was centrifuged to remove pelletable microtubules leaving a5XMTsupernatant. Table lll-IV shows the dilution and competition effects of the nonmicrotubule component of 5XMTsample on binding. Comparison of tables Ill-Ill andlIl-lV indicate that the majority of the competition effect can be attributed tom icrotu bu le s.254Characterization of 3MTx - spermatid-Es bindingIf3MTX - spermatid-ES binding, like that of cytoplasmic dynein and kinesin, canbe reversed by added nucleotides, it most likely posesses ATPase activity. However,because binding can be competed off by unlabelled microtubules, it may be undergoing aslow turnover. In order to further characterize3MT-spermatid-ES binding, it wasmeasured in the presence of agents known to affect MT-mechanoenzyme interaction. Asdescribed in Chapter 1, the effects of a variety of agents that have been used tocharacterize mechanoenzyme activity and binding to microtubules depends on theconditions of the binding, the source of the mechanoenzyme and the nature of the activitybeing tested (for example: motility vs ATPase activity). Very few studies have examinedthe effect of these inhibiting agents on binding directly. The plan here was to select anumber of agents that are known to alter MAP-microtubule interactions and use them tocharacterize3MT-spermatid-ES binding with a view to comparing these propertieswith other organelle-MT binding, bearing in mind that these agents may be affectingmore than one site and/or protein.A number of agents (see table lll-V) were used to test their affect onspermatid-ES binding. A series of experiments were then carried out using these agentsin a design matched in time course, including early and established binding, and asnearly as possible in substrate concentration, to characterize 3MTx - spermatid-ESbinding. Based on the assumption that a high affinity binding may be occuring at thelower tubulin concentrations, the matched series were carried out with O.5Xmicrotubule concentration. Results from these experiments showed the effects weresimilar at 1X and O.5X microtubule concentrations.The agents used to characterize binding were: 10 and 100 1iM vanadate, 2 mMNEM, 1 mM EHNA, 1 mM AMPPNP, added cytoplasmic dynein enriched MAP prep, Sertolicell enriched cytosol, unlabelled excess microtubules, 5 mM GTP (with and withoutdepletion), and ATP depletion with hexokinase and D glucose. The effect of 10 mM MgATPBinding assay 255Figure 3-29. Characterization of3MT - spermatid-ESs binding:matched experiments. Data from a series of experiments designed to re-examine anumber of treatments using conditions matched as closely as possible for substrateconcentration and time course (see methods). Each experiment included an early bindingseries in which inhibitors were added to spermatid-ESs immediately before combining3MTx and spermatid-ESs, and an established binding series in which binding was allowto become established before adding the inhibitor. It provides for the characterization ofbinding with respect to a number of properties of 3MT spermatid-ESs binding. Inall of the graphs in this figure, solid squares (--) are ‘no treatmentcontrols’ (3MT - sperma tidESs).(top left) “cold 5X MT competition” shows the ‘apparent’ decrease in binding followingcold competition by the addition of an excess of unlabelled microtubules (-12’) and thesupematant from the excess cold microtubules binding (-*-).(top right) Binding with added Sertoli cell cytosol from 21 day old rat testis (--. ),and cytoplasmic dynein enriched testis MAP preparation (-rj--). ANOVA shows thatoverall, the effects of adding either the Sertoli cell cytosol or the cytoplasmic dyneinenriched MAP preparation are not significant. While there did appear to be an increasein binding in the presence of early addition of cytoplasmic dynein enriched MAPpreparation, at each time period, it was not significant by ANOVA at the concentrationused.(middle left) 3MTx - spermatid-ES binding is significantly decreased with lOOi.Lmvanadate (*-) but not bum vanadate (-EJ-).(middle right) 3MTx - spermatid-ES binding is significantly decreased with 2 mM NEM(—fl-’ ) but not 1mM EHNA ( -*- ). The late treatment group is joined by dotted line.The spin at 190 minutes is from the early treatment group.(bottom left) 3MT - spermatid-ES binding is significantly reduced by 5 mM GTP (—o-—)with or without ATP depletion. The addition of ATP depletion enzymes hexokinase andsubstrate D-glucose to GTP (to remove any contamination of ATP in the GTP) did notchange the result ( - ).(bottom right) 3MT - spermatid-ES binding was not significantly altered by thenonhydrolyzable ATP analogue AMPPNP ( -12- ).C00a0•1-0unlabelled 5X MTx competition20151050 I • I • I0 60 120 180 2402015C0010a0•1-0 5.time10 tM and 100 tM vanadate0-0C00a0•1-0time (mm)180C00a0Sertoli cell cytoso) andcytoplasmic dynein enriched MAPs20151050200010•0.4-04-C00a4-04-.4-05.0time (mm)2402mM NEM and 1mM EHNA— I I—0 60 120 180 240time (mm)0 60 120 180I...I • • • I60 120q305mM GTP+/- ATP depletion 1 mM AMPPNP10201510500 • I • I • 10 60 120 180 240 0 60 120 180 240time (mm) time (mm)Binding assay 257Table Ill-V. Effects of selected agents on 3MTx- spermatid-ES bindingCHARACTERIZATION OF 3MTx - SPERMATlDES BINDINGCondition tested Early binding Established binding1OmMMgATP5mMGTP .1.5 mM GTP + ATP depletion1mMAMPPNPl0iiMvanadate100 jiM vanadate1mMEHNA2mMNEMSertoli cell enrichedcytosolCytoplasmic dyneinenrichedMAPsexcess unlabelled ‘1m icrotu bu le sThis chart summarizes the effects of a numbers of selected treatments on 3MTx-spermatid-ES binding. Changes in binding, which were found to be significant byANOVA and pairwise comparisons using Tukey’s test (at p= .01) are indicated asdecreased, t increased or unchanged.258was tested on all these treatments. The effects of these agents are shown in Table Ill-V.Those indicated as changed were significantly different from same spins controls byanalysis of variance of all experiments involving that agent. To illustrate the nature ofthese differences, sample data from the matched series of experiments are shown inFigure 3-29.All agents were tested at the onset of binding and after binding had becomeestablished. Only the dynein enriched MAP preparation showed any difference betweenearly and late binding; however, its effect on binding was not statistically significant atthis dynein concentration. The effect of treatment is not affected by timing. It may bethat the dynamic nature of the binding eliminates any effect of early or late use of theagent as a factor in binding.LOCALIZATION OF LABEL TO SPERMATID HEADSLabelled microtubules are localized to the spermatid-ESThere are a number of pieces of evidence that support the contention thatmicrotubules bind to spermatid-ESs. 1) Under the assay conditions, microtubules donot enter the gradient alone, 2) The major source of label found in gradient fractions isfrom microtubules, 3) The presence of label in the gradient fraction is reduced with theaddition of 10 mM MgATP, 4) Label can be competed away with excess coldmicrotubules. 5) Binding is dependent on substrate concentration, and 6) Binding isinhibited by some of the same inhibitors that affect other MAPs. Morphological datashows that spermatid ESs are highly enriched in gradient fractions. It follows that themicrotubules are attaching to the spermatid-ESs and with an enzyme profile that isinfluenced by agents that similarly affect known mechanoenzymes. However, none ofthis data directly addresses the question: Are microtubules binding to the head of thespermatid, that is to the zone covered by ESs? It has been shown here, that isolatedspermatid-ESs stain for actin and, following cold isolation, stain weakly and variably259for tubulin, but, when isolated under warm conditions, retain long strands of tubulinpositive structures.Autoradiography experiments were carried out to localize the label, taken to be acorrelate for exogenous microtubules. Tissue was prepared for the autoradiographyexperiments in the same way as for the binding assay, except to avoid deleterious effectsof freezing, gradients were sampled from the side of the centrifuge tube with a syringe.Pre-gradient samples (stock as loaded onto gradient) and samples that had been treatedwith 10 mM MgATP were included in the study. Parallel binding assays were run.In the pre-gradient samples, label was associated with a number of structures,including spermatids, round structures and clumps of debris. After being enriched and“washed” by the gradient the amount of label (due to unbound microtubules andnonspecific label) was reduced, and primarily observed in association with twostructures, spermatid heads and amorphous clusters of unknown origin. Bearing inmind that microtubules do not enter the gradient on their own, any ‘free’ label is likelyto be primarily from two sources: 1) background picked up by the emulsion, or bycontamination (preliminary experiments showed some contamination occurs whensampling from the side) and 2) label released after entering the gradient, includinglabelled microtubules that have come free from attached spermatids. A series ofexperiments were done to examine the content of amorphous clusters to see if theycontained ESs, as defined by having a coexistence of actin and tubulin (see below).Micrographs in Fig. 3-30 are paired DIC and bright field images to show thedistribution of label in pre-gradient samples. DIC images allow for identification of themorphology of the structures and brightfield enhances the identification of label. Figure3-31 shows paired images from gradient material from ‘no treatment controls’. Labelwas seen associated with spermatid heads as would be predicted for 3MTx - spermatidES binding. Very little label was seen associated with the (membrane covered)spermatid tail. Figure 3-32 shows paired images of gradient material from 10 mMBinding assay 260Figure 3-30. Localization of label in 3MT - spermatid-ES binding frompre-gradient autoradiography samples. Localization of label in 3MT -spermatid-ES binding was followed, by autoradiography.(a, b; c, d) Localization of radioactive label, from 3MT, on spermatid-ES isolate, afterbeing allowed to develop in the dark for three weeks. Paired DIC and bright field imageshave been included to show the structure and highlight the spots produced bydevelopment of emulsion by beta particles emitted from[3H]GTP label incorporated intoexogenous microtubules. These micrographs illustrate the location of the label in 3MT -spermatid-ESs samples before being loaded onto gradients. Label is associated withspermatid head (arrows) as well as other cellular material.bars: a-d = 10 lIm.4.4.•-‘‘C’4ya.4..•1$4.•,*.:%;: ISS 0.S.4t.t.4..U,©•*481%*e48P,.“‘4.I,‘‘‘4445’ S£Binding assay 262Figure 3-31. DIG and brightfield micrograph pairs to show localizationof label in 3MT - spermatid-ES binding from gradient autoradiographysamples.(a,b; c,d; e,f) Paired DIC and bright field images of label in 3MTx - spermatid-ESbinding, sampled from the 30/45 gradient fraction, in autoradiography experimentssampled at three weeks. DIG images show the spots produced by 3MT label (white dots)and the structures with which they are associated. While DIG images allow for the bettervisualization of structure, bright field images enhance the visualization of theautoradiography spots (black spots). Label was mainly associated with two types ofstructures, the heads of spermatids (arrows) and amorphous clumps as shown in e and f(arrow heads).bars: a-f = 10 tm.rfl.3....*...I(414:---::-I****..S--:.•.....%q.-I•-.--.•t;:Ia4004-.Binding assay 264Figure 3-32. DIC and brightfield micrographs pairs to show localizationof label in 3MTx - spermatid-ES binding, in the presence of 10 mMMgATP, in gradient autoradiography samples. These paired DIG and brightfieldimages (a,b; c,d; e,f) show the limited association of labelled 3MT with spermatidheads, following treatment with 10 mM MgATP. Label is seen as white spots in DICimages (a,c,e) and black spots in bright field images (b,d,f). The three samples showthe effect of ATP on a range of spermatid stages. Comparison of these results with thoseof the previous figure shows that label that is associated with spermatid heads in 3MT -spermatid-ES binding is markedly reduced with the addition of 10 mM MgATP.bars: a and b = 10 jim.IA‘94.4*44-I266MgATP treated samples. Label is greatly reduced and is not associated with spermatidheads to the same degree as in untreated samples.These results are consistent with those described earlier, (Fig. 3-23) in whichthe 5A6 anti tubulin antibody staining was observed on 10mM MgATP treated materialand no treatment controls from gradient fractions. The tubulin staining was greatlyreduced following treatment with 10mM MgATP. Unlike controls without primaryantibody, some remnants of tubulin staining remained.Actin and tubulin dual staining of amorphous clustersIn some areas, label was associated with amorphous clusters of unknownmaterial. In some DIC images, varying the focus revealed what appeared to be heads,embedded in some of the larger clusters. Similarly processed tissue (without label) wasstained with rhodamine phalloidin to identify actin, to see if ESs were present inamorphous clusters. Figure 3-33 shows head shaped actin staining, in some clusters,indicating that some of the larger clusters may contain ESs. Figure 3-33: c,d showclumping of material that occasionally occured following processing of gradient material.Figure 3-33: g,h shows the round cells that were sometimes present, but did not stainfor actin.It was reasoned that if microtubules were brought into the gradient on ESs,tubulin would only be present on actin staining structures. On the other hand, not allactin staining structures would necessarily have tubulin. Figure 3-34 shows dualstaining of samples from ‘no treatment control’ gradient fractions, stained to identifyboth tubulin and actin. Most spermatids heads stained for both actin and tubulin. Thosestructures that stained for tubulin, also stained for actin. However, those that stainedfor actin did not always have associated tubulin (see Fig. 3-33: d,e,f, for a head that haslost its tubulin staining, and Fig. 3-33: g.h.i for cellular material that is positive foractin but not tubulin). These may have been ESs that had been separated from thespermatid heads during processing.Binding assay 267Figure 3-33. Actin staining of amorphous clusters, like those seen in theautoradiography slides, show that ESs are frequently present in theclusters. Amorphous clusters, found in gradient fractions, stained with rhodaminephalloidin for filamentous actin. Clusters are similar to those seen in gradient fractionsin autoradiography experiments. These experiments were carried out to answer thequestion: Are there ESs in the clusters seen in autoradiographs? Shown is materialfrom samples that had been recovered, by needle, from gradients and fixed and treatedwith cold acetone, as was done for autoradiography samples.(a,b; , c,d; e,f; g,h) are DIC and fluorescence micrograph pairs that show the structureand filamentous actin staining of structures similar to clusters seen in theautoradiographs. Many of the clusters stained for actin and, in addition, the shape ofspermatid heads can be seen in many of the DIG images (arrows). Not all clustersstained for actin see e,f and g,h pairs (arrowheads). Many of these clusters appear tocontain spermatid heads or remnants of ESs.bars: a-h = 20‘9Ca\0IBinding assay 269Figure 3-34. 3MT - spermatid-ESs stain for both actin and tubulin.Dual staining of 3MT - spermatid-ES5 showing the presence of both actin and tubulinin 3MT - spermatid-ES and in other structures that may be remnants of ESs that haveadhered to the polylysine coated slides. (a, d, g) phase contrast images; (b, e, h)staining for tubulin; (C, f, i) staining for actin. The dual staining shows that in allstructures where tubulin staining is positive, actin staining is also positive. Thereverse is not the case: not all actin staining is accompanied by positive tubulin staining(arrows) see spermatid head that stains for actin but not tubulin in figs (d-f). Note theclusters that stain for actin but not tubulin in figs (g-i). This would suggest that thesestructures may contain ESs with associated microtubules, possibly disturbed byprocessing. Although ESs may be present without microtubules, microtubules are notpresent without the ES (actin).bars: a-i= 25 urn.H-‘-271Pursuit of a proposed regulation of events around the spermatid-ESElongating spermatids are transported in two directions during spermatogenesis,toward the base of the epithelium between stages IV and V and returned to the lumenduring stage VI. The function of this basal migration is not known; nor is its regulation.Equally, the means of regulating microtubule based transport in systems generally, isnot known. In the axon, minus-end and plus-end directed transport occurs in thepresence of kinesin and cytoplasmic dynein, each exclusive in their direction oftransport. It is not known how their direction is specified. There is very littleinformation as to whether motors coexist on the same organelles, how the appropriatemotor is activated, or how its direction is assigned. Based on the evidence thatendoplasmic reticulum plays a calcium sequestering role in cells generally (Somlyo,1984), it has been proposed that ER serves to produce local microenvironments for theregulation of the multitude of calcium-dependent events in cells which may includemicrotubule-based transport (Burton et al., 1985). The ESER is situated between twocytoskeletal elements, actin and microtubules differentially sensitive to calcium. Usingprecipitation of calcium by antimonate, Franchi and Camatini (1985) reported thatcalcium is sequestered in the ESER, in the guinea pig; however, work by Kierszenbaunand coworkers (1971) showed localization of calcium in early spermatids, and a doubleline of extracellular precipitate, but no Sertoli cell deposits, surrounding latespermatids, in the mouse. To explore the role of the ESER, an attempt was made toverify the findings of a calcium sequestering activity of the ESER. A variety of protocolswere used in an attempt to demonstrate the reported calcium uptake in the ESER.The pyroantimonate methods gave mixed results (methods 1-4) which dependedon the technique and did not support, nor refute a conclusion that calcium is sequesteredin ESER (Fig. 3-35,36 and 37). Method one (Kierszenbaum et al., 1971; Tandler etal., 1970), used by Kierszenbaum and coworkers to show an intercellular distributionof potassium antimonate adjacent to late spermatids in mouse testis, demonstrated thesame intercellular distribution in rat testis. The double lines of precipitate described,Binding assay 272in mouse testis, were seen only occasionally in rat tissue and were within the Sertolicell- late spermatid intercellular space. If the method, using prolonged fixation withantimonate and hardening with paraformaldehyde (Kierszenbaum et al., 1971; Tandleret al., 1970), was completed with incubation in 0S04 (Fig. 3-35), no precipitate wasseen in the ESER. Without the final 0S04, there was a loss of membranes and moredispersion of precipitate (not shown). If the tissue was preincubated with calciumchloride or EGTA, it was not materially changed (method: 3). Antimony was included inthe prefixation rinse after the EGTA incubation, and being able to successfully competewith EGTA for calcium, may have stabilized the calcium location. The EGTApreincubation may not have performed its initial intent, to act as a control in whichcalcium ions were removed.The second method, incorporating fixation in combined 2% glutaraldehyde and2% K-pyrantimonate (Klein et al., 1972) as used by Franchi and Camatini, to showintercellular and ER location of antimonate in quinea pig testis, showed very limitedprecipitate in the ESER and a wide spread distribution of precipitate throughoutintercellular space (Fig. 3-36). Although there was only limited vesiculation of ESERin these sections, it contrasts with the data presented by Franchi and Camatini in whichthe ESER is highly vesiculated. In addition they show widening of the intercellularspace, a finding not seen in this study.The fourth method, most frequently used in calcium localization studies,consisted of incubation in combined antimony and 0s04. With this method there wasconsiderable vesiculation of ESER, which often contained precipitate. However,precipitate was also spread throughout the cytoplasmic space and was not present in theintercellular space (Fig. 3-37). The high degree of vesiculation of the ESER, and thepresence of precipitate throughout cell, casts doubt on whether the precipitate isidentifying calcium in the ER or has actually been captured after precipitate formed.The three other methods tried gave poor results. No precipitate was seen intestes or cardiac muscle with method 5: oxalate antiomny method of Borgers et al.,Binding assay 2731977). Using the NHA method, described by Zechmeister (1979), no precipitate wasseen in rat testis or cardiac muscle. Using the tetrabutylamonium fluoride method(Poenine and Epel, 1987), deposits of precipitate coalesced into large deposits that boreno relationship to any specific structure. No data is shown from these methods.These results, using the precipitation of calcium antimony to localize calcium intestis, in part, replicate previous findings and extend the information on localization ofcalcium in testis. However, taken together, they are contradictory and indicate that thismethod does not provide satisfactory evidence of celcium sequestration by the ESER.Binding assay 274Figure 3-35: Localization of calcium in seminiferous epithelium (Method1). Micrograph of rat testis, treated with potassium antimony for calcium localization(method 1).(a) Using method 1, precipitate is almost exclusively in an intercellular (arrow heads)location, with occasional evidence of entry into the cell (long arrow), but not localizedwithin ESER (short arrows).(b) A higher magnification of the area marker with a large arrow in (a) shows theprecipitate to be between the germ cell membrane (on the left of the precipitate) andSertoli cell membrane (on its right). Microfilaments and the endoplasmic reticulum ofthe ES are seen just deep to the Sertoli cell membrane.(c) Rat cardiac muscle (control for method 1) shows antimony is present in thesarcoplasmic reticulum and in intracellular location (small arrows). Antimony has hadto cross the cardiac myofiber membrane to precipitate calcium within the sarcoplasmicreticulum or in the myofibrillar space, as shown here.bars:a=1tm,b=O.25pm,c=1im.4‘“Vrii_ -“$$“ 44—IC1:•g’’2..4275rt :44..Binding assay 276Figure 3-36: Localization of calcium in the seminiferous epithelium(Method 2). Micrograph of rat testis treated with potassium antimony using method 2.(a) Extensive intercellular precipitate (short arrows), occasional ESER precipitate(arrowhead) and large deposits in intracellular vesicles are seen surrounding latespermatids.(b) Antimony depostis mainly in intracellular vesicles (long arrows) and intercellularlocation.(c) ESER does not contain antimony, but intercellular deposts are seen (short arrows).bars:a=1tm,b=1tm,c=O.5tm,d=O.5tm.p1‘1d.:i’_-S,(:IC46Binding assay 278Figure 3-37: Localization of calcium in seminiferous epithelium (Method4), Localization of antimony precipitate using a method mostly commonly used incalcium localization experiments. Here precipitate is in ESER in some areas (smallarrowhead) but not others (large arrowheads). It is also located in vesicles in thecytoplasm (short arrows) as well as being ‘free’ in the cytoplasm (long arrows). Thelight background of the cytoplasm suggests that the cytoplasm has been extracted fromthis cell and the ‘free’ calcium may be attached to remaining cytoskeleton. Extensivevesiculation of the ESER occurs with this method and with longer incubation,membranes are lost.bars: a,b = 0.5 im.:.-dj——. .“:‘c2.19:-Igc-Sk,i’• E-.’fI —_Binding assay 280DISCUSSIONThe findings of this part of the study, designed to test binding betweenSpermatid-ESs and microtubules, are summarized here. The results and specificconclusions that can be drawn from these findings are summarized in this section. Theimplications of these findings and their relevance to the proposal that spermatids aretransported through the seminiferous epithelium by a microtubule-based transportmechanism are discussed in Chapter 4.COMPONENTS OF THE BINDING ASSAYSpermatid-ESsThe components used in the binding assays, isolated spermatid-ESs and labelledtaxol assembled bovine brain microtubules, were acceptable representatives of their invivo counterparts: spermatids with ESs and Sertoli cell microtubules. The spermatidES isolation method developed for the binding assays, involving poking and then gentlesquashing of the seminiferous epithelium, yielded spermatids from all stages. Isolationof spermatids from segments of seminiferous tubules that had been staged bytransillumination, showed that all stages of elongate spermatids could be recovered bythis spermatid isolation technique. This was further supported by the presence of allspermatid stages in the isolated tissue as seen in the micrographs from this study. Thepresence of ESs on isolated spermatids was indicated by bright actin staining withrhodamine phalloidin, in a shape characteristic of ES distribution, and by microfilamentnetworks, in electron micrographs. These observations are consistent with the findingsof others in which ESs remained intact during mechanical isolation of spermatids(Romrell and Ross, 1979; Franke et al., 1978; Grove and VogI, 1989; Masari et al.,1987; yogI, et al., 1986), with profiles of ESER still attached. When removed from theconfines of Sertoli cells, in buffers without sucrose, the ESER tends to dilate, and inBinding assay 281some cases break into smaller vesicles, still attached to the actin network. Some dilationof the ESER was still present in EMs of spermatids isolated in PEM buffer with 100 mMsucrose. Therefore, to protect the ESER further from osmotic effects, 250 mM sucrosewas used in all buffers, as employed by Pratt to protect organelles from osmotic damagein the isolation of vesicle-microtubule complexes (Pratt, 1986).Endogenous microtubules, were retained during isolation of spermatid-ESs underwarm conditions, and seen as long strands that extended beyond the ends of spermatidheads, with anti-tubulin immunofluoresence. These were largely removed duringisolation under cold conditions. In contrast, exogenous microtubules stabilized withtaxol, used in the binding assay were seen closely associated with the spermatid head,using the same tubulin probe. This is consistent with the exogenous microtubules,assembled with taxol for this study, being an average length of of 2.8 jim (about 15 to20% of the length of the spermatid heads). Endogenous microtubules, depolymerizedduring the cold isolation, may have reassembled; however, they would be unlabeled andtherefore ignored’ by the binding assay. Immunocytochemical staining of spermatidESs, that had been incubated with exogenous microtubules and recovered from gradients,indicated that tubulin was present only on spermatids that stained positively forfilamentous actin. Not all actin containing structures stained for tubulin. This suggeststhat the presence of ESs was necessary for the presence of microtubules, butmicrotubules were not present without ESs.Labelled, taxol polymerized microtubules remained stable throughout the assay.Exogenous microtubules were labelled with[3HJGTP. Assembled from bovine braintubulin, in the presence of taxol and substoichiometric amounts of[3H]GTP, theyincorporated an average of 43 ± 9.2% of the available label. The number of counts leftin the supernatant by pelleted, labelled microtubules, divided by the number of totalcounts in the 3MT sample, (X100) was taken as the percent of label incorporated intomicrotubules, bearing in mind that some microtubules may be of insufficient length topellet. Some breakdown of GTP during storage and preparation steps could beBinding assay 282anticipated. Label did not cycle from microtubules during the time of the experiment,indicating that the microtubules were stable. Similar microtubule stability wasreported by Wilson and coworkers (Wilson et al., 1985) and is further supported bythe observation that binding did not appear to decline over the course of the binding assayexperiments. If microtubules were cycling and thereby losing their label, there wouldhave been an ‘apparent’ reduction in binding (detected by counts), but not real binding oflabelled and unlabeled microtubules. Microtubule stability was unchanged in thepresence of all agents used to characterize microtubule binding and therefore did notcontribute to the changes in binding that occurred with these agents.THE BINDING ASSAY MEASURES MICROTUBULE- SPERMATID BINDINGEffects of the sucrose gradientSDS gel electrophoresis and scintillation counts from gradient fractions,indicated that microtubules did not enter the gradient alone, with the differentialcentrifugation parameters used in these assays. In addition to ‘washing out’ of unboundmicrotubules, the sucrose gradient served to provide a fraction enriched for spermatidESs, having separated them from the other cellular components of the isolate.Therefore, the counts in the gradient were primarily associated with spermatid-ESs.Counts represent microtubule bindingThe possibility remained that even though microtubules became labelled, othersources of label such as breakdown products of[3H]GTP may have taken part in nonspecific binding with spermatid-ESs and thereby entered the gradient. To address thispossibility, it was necessary to determine what proportion of the label that entered thegradient, in these assays, could be attributed to microtubules. Microtubules wereremoved from 3MT stock, by pelleting, giving a ‘microtubule free’ 3MTX supernatant.Measurement of binding with 3MT or with ‘microtubule free’ 3MT supernatant, in theBinding assay 283same spins, provided a measure of the microtubule and non-microtubule components ofthe label that entered the gradient sample and therefore the microtubule and nonmicrotubule components of binding. Between 75-92% of counts found in the gradientfractions could be attributed to labelled microtubules. Taken together, these datasupport the interpretation that: 1) the majority of counts present in the gradient areassociated with spermatid-ESs can be attributed to microtubules, and 2) that themajority of counts represent specific3MT-spermatid-ES binding.MAP isolation procedures, particularly for mechanoenzymes, take advantage ofthe fact that high concentrations of nucleotides can be used to release microtubulebinding proteins from microtubules. In this study, the reduced number of countspresent in the gradient fractions, following the addition of 10 mM MgATP, supports theinterpretation that it was microtubules that were bound to spermatid-ESs.Furthermore, binding is directly influenced by changes in either spermatid-ES ormicrotubule concentration, supporting the interpretation that spermatid-ESs andmicrotubules are both involved in the binding.Binding time courseAfter an initial increase during the first hour, binding remained relativelystable over the course of the experiment, in untreated controls. Scheel and Kreis(1991), in an assay of endocytic carrier vesicle-microtubule binding report a similartime course, with an increase in binding over the first 45 minutes. This may representthe time required for the system to stabilize. There was an expected variability betweenexperiments. This was dealt with, in part, by expressing binding as a percent of samespin controls. Similarly, the effects of treatment agents showed some variabilitybetween experiments: therefore, the significance of treatment effects was tested byanalysis of variance.Binding assay 284Effects of added nucleotideThe amount of ATP, present or being replenished in the spermatid-ES isolate wasnot known. Hexokinase uses ATP as its exclusive substrate for the breakdown of Dglucose and is widely employed as a tool to deplete ATP from the cell contents. In thisstudy, ATP depletion from spermatid-ES isolate resulted in increased microtubulebinding in some, but not all, binding assays to which it had been added. A possibleinterpretation is that binding may have increased as a result of ATP depletion, in somecases and in the others, that either the ATP was not depleted sufficiently to affectbinding, or that ATP, (or the capacity for its renewal) was already very low, in thosesamples.The concentrations of nucleotides, 10 mM MgATP and 5 mM GTP, chosen to testfor binding release in the binding assays were guided by experiments in whichnucleotides were used to release mechanoenzymes from microtubules duringmechanoenzyme isolation (Paschal et al., 1987). The effect of 10 mM MgATP was testedin virtually all binding assays in this study, both on ‘no treatment controls’ as well as onthe treated groups. Binding was reduced in every case. The addition of 10 mM MgATPreduced binding to an average of 36 ± 13% of ‘no treatment controls’. Although thevalues from these data tended to cluster into two groups, the differences between thesetwo groups could not be determined from the available data. 5 mM GTP reduced bindingto 43 ± 21% of controls. In preparations containing cytosolic enzymes, nucleosidasediphosphokinase activity removes the terminal phosphate from available nucleotides toreconstitute ATP. This can result in the presence of ATP in GTP containing solutions(Paschal et al., 1989; VaIlee et al, 1989). In assays in which 5 mM GTP was added, inthe presence of hexokinase and glucose to deplete ATP, binding was not different from theaddition of GTP alone. Increasing the Mg concentration did not change the GTP effect(1mM MgCI2 was present in all buffers). The effect of 10 mM MgATP and 5 mM GTPwere not statistically different when they were compared in the same spins. However,when they were combined, binding was reduced to less than one half of that of eitherBinding assay 285agent alone. One of a number of explanations for the apparent additive effect of ATP andGTP is that the effects of the two nucleotides may be mediated at different binding sites.What is evident is that high levels of both ATP and GTP reduce3MT-spermatid-ESbinding. AMPPNP, a non hydrolyzable analogue of ATP did not significantly changebinding. It is of interest to note that, in experiments testing AMPPNP effects, valueswere somewhat lower than in controls; however, taking into account the betweenexperiment variability, the difference was not significant.Competition by unlabeled microtubulesUnder the conditions of these assays, competition by excess unlabeledmicrotubule competition may indicate that there was slow turnover of 3MTx-spermatid-ES binding. This is based on the evidence that label is not cycling frommicrotubules and labelled microtubules can be competed away by the addition of excessunlabeled microtubules. The turnover was unlike the rapid turnover that would occurduring the dynamic binding and release in microtubule based motility, in which caseinsufficient rigor complexes would have formed to detect binding. However, it suggeststhat these events are ATP dependent and potentially dynamic. Competition by unlabeledmicrotubules supports the contention that the label in gradient fractions is microtubulein origin.CHARACTERIZATION OF BINDING3MTxspermatidES binding was reduced by 2 mM NEM and high concentrationsof vanadate, but unchanged by EHNA, low concentrations of vanadate, and the addition ofSertoli cell enriched cytosol and MAP enriched preparations. Binding was reversiblewith the addition of MgATP and GTP and unchanged with AMPPNP. Depletion of ATP byhexokinase and D-glucose increased binding in some preparations. The relevance ofthese observations are discussed in Chapter 4.Binding assay 286LOCALIZATION OF3MTX-SPERMATID-ES BIN DINGLabel associates with spermatid headsData from the binding assays supported the contention that microtubules wereable to bind to isolated spermatid-ESs, but did not localize the binding. Theautoradiography data indicated that the label, shown to be primarily microtubule originin gradients, was associated with two structures in autoradiographs.In pre-gradient samples, label was more intense and widely distributed than ingradient samples. This is in keeping with the presence of both bound and unboundmicrotubules. In the pre-gradient samples, label was associated with spermatid headsbut not exclusively. In gradient samples, where unbound microtubules were notpresent, label was associated mainly with the heads, but not the tail of spermatids.Immunocytochemical staining of material, treated in the same manner as theautoradiography samples, showed that the microtubules are localized to the heads ofMT-spermatid-ES complexes. Taken together, these data support the conclusion thatthe label is indicative of microtubules and that they bind to spermatid headsOn autoradiographs, the label on3MT-spermatid-ESs treated with 10 mMMgATP is reduced in number and is no longer associated with microtubule heads. Thisprovides further support for the specificity of binding to the spermatid head. There area number of possible explanations for the retention of some label in3MT-spermatid-ESs treated with 10 mM MgATP. Because 10 mM MgATP is included in the gradientsolutions, further microtubule release may occur in the gradient. Also, it was observedin preliminary experiments that sampling gradient fractions from the side of the tube,as was done for the autoradiography experiments, results in some contamination bylabel. Finally, as explained below, rinse steps tended to wash spermatids from slides,possibly separating ESs from the spermatid heads. For these reasons, the observationswere not quantified.Binding assay 287Localization to amorphous clustersLabel associated with small amorphous clusters of unknown composition. Todetermine the source of these clusters and their composition, two approaches were used.The first was to see if the clusters were present in the gradient samples or if they hadformed as a result of processing. Samples of gradient fractions were observed with DICmicroscopy before any processing was carried out. Only occasional clusters wereobserved in those samples. However, after processing for immunofluoresence, clusterswere observed more frequently, often containing DIG images of spermatid heads. Thesewere thought to be effects of rinsing of the slides during preparation. The concentrationof spermatids was reduced by processing, indicating that spermatids were beingdisplaced from the slide. The second approach was to determine if the clusters containedESs, either having been separated from the spermatid heads, or clumped together duringprocessing. The presence of actin, detected with rhodamine phalloidin, was considered tobe evidence of the presence of ESs. Samples, treated in the same manner as theautoradiography samples, were double stained to identify actin and tubulin. Actin andtubulin were frequently associated with amorphous clusters, suggesting that they maycontain ESs that were associated with material that accumulated during processing, orseparated from spermatids. This is consistent with an observed loss of spermatids, fromglass slides, during processing. Moreover, tubulin staining did not occur without actin,suggesting that the presence of microtubules required ESs.ANTIMONY PRECIPITATION EVIDENCE OF CALCIUM IN ESER ISINCONCLUSIVEThe contradictions between reports of calcium localization in guinea pig (Franchiand Camatini, 1985) and mouse (Tandler et al., 1970) testis may be related to thetechniques used, rather than species differences. The methods of antimony precipitationto localize calcium, used in these studies were repeated in rat testis with similarlycontradictory results in the same species. The use of other techniques, all of which haveBinding assay 288been used to localize calcium in a variety of tissues, increased the range of possibleinterpretations. Method 3, was developed to address the possibility that calcium ionsmoved from their in vivo location as a result of treatment with fixatives. In this casethe antimony acts as the primary fixative, with aldehyde fixatives added after antimonyhas precipitated calcium in its in vivo location. That would indicate that calcium, insignificant amounts to be identified, is only present in the intercellular space. Thetissue preservation is better in these sections. The fact that antimony acts as theprimary fixative and that it identifies intracellular calcium in cardiac muscle, suggests,but does not prove, that it is not prevented from entering the cell. It may be thatcalcium enters the cell in tissue which loses membrane before it is fixed andvesiculation surrounds calcium that would otherwise be in the intercellular space. Theenormous range of variations in antimony precipitation methods used to localize calciummay be an index to the difficulty interpreting this technique. The only conclusionavailable from this work is that corroborative methods are needed before it can beaccepted that calcium is present in the ESER.0 0C.)CoIC) C-UCl)-ICl)m5 zCDDiscussion 290INTRODUCTORY REMARKSThe purpose of this study has been threefold: firstly, to determine the polarity ofSertoli cell microtubules; secondly, to examine the potential for binding between ESs andmicrotubules; and lastly, to characterize that binding. It has been shown in this study,that Sertoli cell microtubules are oriented with their minus-ends directed toward theapical surface of the cell. This observation has important implications for the Sertolicell and for cells generally. Data from the MT-spermatid-ES binding assays, reportedhere, indicate that ESs are able to bind microtubules. The binding can be released byATP and GTP, and shares properties with organelle-microtubule binding in othersystems. These findings are consistent with a role for microtubules in positioningspermatids within the seminiferous epithelium and the hypothesis of a microtubulebased transport mechanism for spermatid translocation.MICROTIJBULE POLARITY IN SERTOLI CELLSIn a number of recent reviews of the mechanisms involved in the organization ofintracellular organelles (Kreis, 1990; Kelly, 1990a,b; Schroer and Sheetz, 1991),the authors have emphasized the importance of microtubule orientation to organelledistribution, in particular, the need to account for the observation that microtubules inpolarized epithelial cells may be organized differently than those in many cell types onwhich the classical model has been based. In view of the microtubule dependence of manycellular events, many of which are mediated by mechanoenzymes with a preferreddirection of transport, the orientation of cytoplasmic microtubules has importantconsequences for these events, particularly in epithelial cells.The hook decoration method has been used to determine microtubule polarity in anumber of cells and cell free systems. Unipolar microtubule arrays have beendemonstrated in cat post-ganglionic sympathetic fibers (Heidemann et al., 1981),chicken sensory neurites (Bacallo et al., 1989), rat hippocampal neurons and frogolfactory neurons (Baas et al., 1988; Burton and Paige, 1981), angelfish and teleost291melanophores (Euteneuer and McIntosh, 1981; (Phillips and Satir, 1988), teieostphotoreceptors (Troutt and Burnside, 1988a) and heliozoan axopodia (Euteneuer andMcIntosh, 1981), sunfish retinal pigmented epithelial cells (Troutt and Burnside,1988b), Madin Darby canine kidney (MDCK) cells (Bacallao et al., 1989), andDrosophila wing epidermal cells (Mogensen et al., 1989). The only exception to theunipolar distribution of cytoplasmic microtubuies, reported thus far, are the bipolarmicrotubule arrays that have been described in dendrites ((Baas et al., 1988; Burton,1988). interestingly, in developing nerve processes all neuronal microtubules areunipoiar and have their positive end directed toward the cell periphery until the firstsign of dendritic specific morphology, at which time dendrite microtubules becomebipolar (Baas et al., 1989).in the majority of cell types studied, cytoplasmic microtubules all oriented withtheir minus-ends radiating from a supranuclear microtubule organizing center andtheir plus-end directed toward the cell periphery. The cell types in which thismicrotubule orientation has been demonstrated include: cat post-ganglionic sympatheticfibers Heidemann et al., 1981), chicken sensory neurites (Bacallao et al., 1989), rathippocampal neurons and frog olfactory neurons (Baas et al., 1988; Burton, 1981),angelfish and teleost melanophores (Euteneuer and Mcintosh, 1981; McNiven andPorter, 1986), teleost photoreceptors (Troutt and Burnside, 1988a) and heliozoanaxopodia (Euteneuer and McIntosh, 1981). Recently, a few exceptions to this plus-endout microtubule pattern have been reported, describing cells in which microtubules areoriented with their minus-ends directed peripherally. These include sunfish retinalpigmented epithelial cells (Troutt and Burnside, 1988b), Madin Darby canine kidney(MDCK) cells (Bacallao et al., 1989), and Drosophila wing epidermal cells (Mogensenet al., 1989). Sertoli cells share their ‘unusual’ polarity with these cells.In the classic cell model (Kreis, 1990; Kerr, 1991a,b) of cellular organization,microtubules are thought to be nucleated from centrosomal MTOCs (typically locatedcentrally in a perinuclear position), stabilized at their proximal minus-ends, andDiscussion 292growing at their peripherally-directed plus-ends (Kirschner, 1980). There areexceptions to this pattern. Microtubules of developing dendrites switch from a unipolardistribution, with their plus-end directed peripherally, to a bipolar distribution in theabsence of a demonstrable MTOC. Similarly, amputated neurite segments are capable ofreorienting their microtubules in the absence of any known MTOC (Baas et al. 1987).In those cells in which microtubules have been shown to be oriented with their minus-ends out, microtubule-MTOC relationships are atypical. In retinal pigmented epithelialcells and MDCK cells, both with their MT minus-end directly peripherally, cytoplasmicmicrotubules do not appear to associate with perinuclear centrosomes (Trout andBurnside, 1988b). Nucleating centers may relocate to an apical position, as has beendemonstrated in Drosophila wing cells (Mogensen et al., 1989), an epithelial cell inwhich microtubules are oriented with their minus-ends directed toward the apicalsurface of the cell Mogensen et al., 1989). While it is clear MTOCs, with their pairedcentrioles and centrosomal material, are capable of nucleating microtubule, othermicrotubule organizing mechanisms probably exist. A recent report of the nucleatingcapacity of gamma tubulin in the absence of a MTOC (Joshi et al., 1991) is potentiallysuch a mechanism.Sertoli cell microtubules occur parallel to the long axis of the cell and extendfrom the body into apical processes (Amlani and yogI, 1988, yogI, 1988). Althoughpaired centrioles have been reported in a supranuclear position (Nagano, 1 966) inSertoli cells, microtubules do not appear to arise from a central point. The distributionof centrosomal material has not been described in Sertoli cells. Further studies areneeded to determine the source of microtubule nucleation in Sertoli cells.Other than in the cultured MDCK cell, the results reported here are the onlyexample of mammalian epithelial tissue for which microtubule polarity has beendetermined. Although the cells that have been reported to share their minus-end outmicrotubule polarity with Sertoli cells are from a diverse group of organisms, theyshare at least one feature, they are all epithelial cells. The finding that MDCK cellsDiscussion 293reorganize their microtubule polarity as they establish cell polarity (Pepperkok et al.,1990; Bré et al., 1990; Bacallo et al., 1989) has given rise to a new model, theparallel model for microtubule organization in epithelial cells (Schroer and Sheetz,1991). Based on the limited examples that have been reported, Schroer and Sheetz(1991) have suggested that the minus-end out polarity, reported here for Sertoli cells,may not be unusual. They have proposed three models of intracellular organization basedon microtubule organization: 1) the radial model: in which microtubules radiate from anucleating center with their plus ends directed peripherally, typical of many culturedcells (Kreis, 1991; Kelly, 1990a,b); 2) the linear model: in which microtubules arearranged in series with slight overlap at their ends, unipolar in their orientation, asfound in axons (Burton and Paige, 1981), and 3) parallel: microtubules arranged in aparallel configuration with their minus ends directed toward the apical surface of thecell in polarized epithelial cells(Mogensen et al., 1989; Troutt and Burnside, 1988b;Bacallao et al, 1989; Redenbach and yogI., 1991).There are a number of possible consequences of the polarity of cytoplasmicmicrotubules. As described earlier, the transport and distribution of most membranebounded organelles depend on a microtubule-based transport system, as does the entire“continuous intracellular circulatory system” (Heuser, 1989) of the cell.Furthermore there is evidence that microtubule polarity plays a vital role. Amputationof embryonic chicken sensory neurites (Baas et al., 1987) or teleost melanophore arms(McNiven et al., 1984) results in a reversal of the microtubule polarity at the cut endwithin the amputated segments. In severed melanophore processes, a similar reversal ofmicrotubule polarity coincides precisely with the reversal of pigment aggregation in thedistal cut end (McNiven and Porter, 1986). It is not known what factors institute orregulate these changes in microtubule polarity. However, there is some signal toreestablish a functional pattern. This relationship between polarity and functionimplies that once the mechanoenzymes and their regulating signals are in place, thepolarity of the microtubules is critical.Discussion 294It is increasingly apparent that the distribution of cellular organelles isprimarily based on a preferential direction of transport of specific organelles (Blackand Baas, 1989; Kelly, 1990a,b; Kreis, 1990) presumably mediated bymechanoenzymes that have a preferred direction of transport. Therefore the polarity ofmicrotubules will influence the distribution of organelles in the cell. It will beinteresting to see whether organelles in polarized cells that exhibit a minus-end outpolarity, have a unique distribution, or whether they employ different motors for theirpositioning than their counterparts in non-polarized cells.Reorientation of microtubules to a minus-end out pattern, in epithelial cells,appears to be linked to the establishment of cell polarity. The formation of cell-celljunctions and the development of polarity, in cultured MDCK (epithelial) cells,immediately precedes the reorganization of microtubules (Bacallao et al., 1989).Therefore the new microtubule pattern is probably a response to rather than a cause ofpolarization. Following their reorganization, microtubules have greater stability (Bréet al., 1990; Pepperkok et al., 1990). The apical migration of electron dense materialis proposed to be centrosomal material that assumes a microtubule nucleating role, inthe apical region of the cell, following cell polarization of MDCK cells. It is not knownwhat mechanisms are responsible for the change in organelle position during theestablishment of cell polarity, or with the change to bidirectional polarity of dendritemicrotubules, at the first sign of dendrite specific morphology (Baas et al., 1988,1989). However, the coincidence of the alteration of microtubule polarity with changesin tissue specific morphology suggests a causal relationship.In secretory cells, the assignment of secretory products, as well as apically orbasally targeted membrane proteins to their appropriate surface, may depend on thepolarity of microtubules and the mechanoenzymes and the regulatory proteins that areavailable. It will be interesting to determine the microtubule orientation for cells thathave more than one active secretory surface. With the basal targeting of proteins beingDiscussion 295a bulk flow phenomenon, apical targeting may be a departure from the constitutivepattern, imposed by special constraints of polarization.Sertoli cell microtubules are likely to participate in the same microtubuledependent functions as other cells, with their microtubule polarity influencing theseevents. The minus end-out orientation may have a causal relationship with theestablishment of cell polarity, as Sertoli cells assume their epithelial morphology. Asdescribed earlier, during spermatogenesis, Sertoli cell organelles undergo marked stagedependent positional changes of cellular organelles including repositioning of the Golgiand elements of ER. Sertoli cells are secretory cells (Griswold, 1988), assuming theresponsibility for establishing an environment to maintain viable spermatogenic cellsthrough release of secretory products into the seminiferous tubule lumen, which iscontinuous with the adluminal compartment that houses spermatids. Much of thesecretory pathway is microtubule dependent in cells generally, and can also be expectedto be microtubule dependent in Sertoli cells. Many of the Sertoli cell secretory productsare moved to the seminiferous tubule lumen, toward the microtubule minus-end. Theelaborate shape changes that occur during spermatogenesis would require activemembrane recycling. This pathway is, in part, microtubule dependent (See Schroer andSheetz, 1991; Kreis, 1990) and would involve an important role for Sertoli cellmicrotubules. In addition, the lysosomal pathway, also partially microtubule dependent,is active and stage dependent in Sertoli cells as they not only dispose of their ownunwanted cellular end products, but are responsible for the elimination of nonviablespermatogenic cells and residual bodies. In view of the range and extent of microtubuledependent activities occurring in Sertoli cells, incurred by both the added responsibilityimposed by continuous stage dependent changes and the responsibility assumed for manynurturing functions on behalf of the spermatogenic cells, it is not surprising to find thatSertoli cells possess an abundance of microtubules. It follows that they would alsopossess an abundance of the required microtubule associated proteins to mediate thesefunctions. Furthermore, with Sertoli cell microtubules directed with their minus-endDiscussion 296out, an abundance of minus end-directed motors could be expected. Consistent with thisis the finding that a high concentration of MAPs is present in testis (Neely andBoekelheide, 1988; Collins and Vallee, 1989) and Sertoli cell enriched fractions,including the minus-end directed motor, cytoplasmic dynein (Neely and Boekelheide;1988).Discussion 297MIx-SPERMATID-ES BINDINGSertoli cells are thought to mediate the translocation of spermatids within theseminiferous epithelium. The events of spermatogenesis are highly complex and theyare not readily amenable to manipulation in vivo. For this reason, an in vitro systemhas been developed to examine the potential for spermatid-ES-microtubule interaction.This approach has been rewarding for the study of microtubule-based transportgenerally (Vale et al., 1985a-d).There are three methods in common use to test organelle-microtubuleinteractions in vitro. The first is the use of binding assays (Pratt, 1986; Suprenant andDentler, 1982; Van der Sluij et al., 1990; Scheel and Kreis, 1991a,b; Rothwell et al.,1989; Sherline et al., 1977). These involve assaying the interaction of organelles withmicrotubules under selected conditions, usually by direct observation or differentialsedimentation. The advantage of the binding assay is that it can be used in less definedsystems and is often the place to start in systems with many unknown components or incases where the questions being asked defy simplification of the components. The secondmeans of testing microtubule-organelle interaction is the motility assay (Vale andToyoshima, 1989). These assays are particularly useful when the elements involvedare small and amenable to observation with DIG video enhanced microscopy. Manyorganelle transport systems are either simple, or well defined, allowing for studies tobe carried out in vitro. Motility assays are most useful in systems that can be readilymanipulated, and have the advantage of providing information about the wholemechanochemical cycle, allowing for the study of dynamic parameters. The third methodis a biochemical progression from the first two and involves characterization of theATPase activity of the enzyme. This generally requires that the motor be isolated(Paschal et at., 1987; Pratt, 1986a, 1991).The advantage of the binding assay is that it can be used to test the interaction oforganelles with microtubules before the participating partners have been defined, inorder to find out what characteristics can be used to further test the interaction. It298therefore tests the potential for microtubule binding and can be used to partiallycharacterize that binding. The Si crude supernatant that Vale and coworkers (1985d)isolated from axoplasm contained factors that transport vesicles in both directions,participants from two separate events. Not unlike the early in vitro studies of squidaxoplasm, binding assays in less defined systems are testing all the motors present inthe system, collectively. Similarly, the spermatid-ES isolation technique samples fromall stages of spermatogenesis. Because the transport of spermatids toward the base of theepithelium and toward the apical surface involve a different directions of transport anddifferent stages of spermatogenesis, it is reasonable to predict that there may be morethan one motor to mediate these events. The method of spermatid isolation used in thisstudy recovers spermatid-ESs from all stages. It is important to remember that theresults of the binding assays in this study represent3MT-spermatid-ES bindingproperties of all stages in combination.The wealth of information that has been gained from the ATPase activity assaysand motility assays to study mechanoenzyme behavior has contributed to the study ofmore complex systems. Bindings assays are being employed to look at specificorganelle-microtubule interactions in an attempt to test whether those organelles maybe part of microtubule dependent pathways (see Schroer and Sheetz, 199ia). There area number of studies, using binding assays, that closely parallel the method used in thisstudy, and provide some comparison for the data presented here (Pratt, 1986;Suprenant and Dentler, 1982; Van der Sluij et al., 1990; Scheel and Kreis, 1991a,b;Rothwell et al., 1989; Sherline et al., 1977). Details of these studies have beendescribed in an earlier section: binding assays, chapter 1.Mechanoenzymes share three properties by which they can be identified (Scholeyet al., 1988): 1) formation of enzyme cytoskeleton interactions in absence of ATP. 2)dissociation of that interaction with nucleotide, usually ATP, and 3) cytoskeletonactivated ATPase activity. The binding assays used in this study provide evidence of thefirst two properties in microtubule-spermatid-ES interactions: that binding occurs,299and that it is releasable with nucleotides. Without the isolation of the linking protein(s)ATPase activity could not be measured in this system. Here, binding in the absence ofadded ATP and dissociation with nucleotides, taken together with an apparent turnover ofbinding, support, but do not prove that the linking protein is a mechanoenzyme.Spermatid-ES complexes share their potential to stably bind microtubules withall of the organelle-microtubule binding assays described earlier. This bound state isanalogous to the arrest of motility and depression of ATPase activity that occurs in theabsence of ATP and transport and increased ATPase activity in the presence of ATP, thatare measured in motility assays and ATPase activitiy assays respectively. The kinesinmicrotubule interaction is releasable by ATP and GTP but the dynein-microtubuleinteraction is releasable only with ATP. 3MT spermatid complexes were releasablewith both 5 mM GTP and 10 mM ATP. An unexpected finding was the greater release thatoccurred with a combination of ATP and GTP, compared with either alone. Although thiscould be a dose dependent effect, most mechanoenzyme isolation techniques employ thelevels of nucleotide used in this study, on the assumption that they are in excess of whatis required for maximum release (Paschal et al., 1991; Neely and Boekelheide, 1988).One possible interpretation is that ATP and GTP are acting at different sites andtherefore have an additive effect. Of six binding assays cited earlier, five formed rigorcomplexes that were releasable with ATP. Of the three in which GTP and ATP releasewere both tested, two showed release with both nucleotides The common interpretationis that in the absence of ATP, rigor complexes form and that with the addition of ATP(and in some cases GTP), there is a decrease in motor-microtubule affinity and rigor isreleased (Lasek and Brady, 1985; Chiocotin and Johnson, 1989; Hackney, 1988).Rickard and Kreis (1991) offer an alternate explanation that arose from aninconsistency in their data. They have identified a 1 7OkD microtubule binding protein“ppl 70” that mediates binding between exocytic vesicles and microtubules. (Rickardand Kreis, 1991; Scheel and Kreis 1991a.b) Initially ppl7O was found to be releasedfrom microtubules by ATP and GTP in crude preparations, but not with affinity purifiedDiscussion 3 00protein. They have extended their findings to show that nucleotide dependent releaserequires cytosolic extracts. Furthermore, they report that 170 kD-microtubulebinding is regulated by phosphorylation of 170 kD (also referred to as ppl7O), bycytosolic kinases in the present of excess ATP. Removal of ATP restricts kinase activity.Phosphatases in the cytosol dephosphorylate the ppl7O protein, resulting in thedephosphorylated form and its release from microtubules. Scheel and Kreis (1991 a,b)have used a novel binding assay, employing an affinity matrix of microtubules complexedto magnetic beads and exposed to endocytic vesicles, to show that pp 170 is likely theprotein that mediates endocytic carrier vesicle-microtubule binding. They interpretthe cytosol requirement as a need for a source of kinases and phosphatases to regulate theon/off cycling. The endocytic carrier vesicle-microtubule binding is releasable withATP and GTP, presumably, not as a stimulus for the mechanochemical cycle, but toprovide for phosphorylation of ppl 70- to achieve the off state.Depletion of ATP frequently increases the formation of rigor complexes and isused to increase mechanoenzyme binding in both organelle and microtubule-basedmechanoenzyme isolation. In this study ATP depletion produced mixed results,increasing binding in two of six experiments, leaving it unchanged in four. Because thebasal level of ATP was not monitored in these assays the basal level is not known. It maybe that the residual ATP or the possible regeneration of ATP, was already sufficientlylow to maximize binding in four of the six experiments. There appears to be someturnover of binding occurring, evidenced by the competition with excess unlabeledmicrotubules, but it is slow enough to provide enough rigor complexes to detect binding.Microtubule-spermatid-ESs were unaffected by AMPPNP.In most experiments AMPPNP reduced binding slightly but accounting forvariance, the reduction was not significant. In the presence of AMPPNP, kinesin-likemotors would be predicted to increase the formation of rigor complexes, while a dyneinlike motor would slightly decrease binding. If kinesin were present, an increase inDiscussion 30 1binding may be expected because of the slow release of ADP from kinesin (the ratelimiting step), which occurs after ATP hydrolysis and before kinesin-microtubuleseparation. If dynein were present, a decrease in binding would be expected because ofthe relative ineffectiveness of AMPPNP as a substitute for ATP in releasing rigorcomplexes with dynein or myosin (Penningroth, 1989). Although dynein-microtubulerigor complexes would be released with ATP, they are less effectively released withAMPPNP.In the binding assays cited earlier, three groups tested the effect of NEM onbinding: two found reduced binding at 1-5 mM levels. Kinesin is less sensitive to NEMthan dynein but the results of experiments vary. NEM is a sulfhydral alkylating agentthat interferes with transport by altering SH groups before the motor interacts with thecytoskeleton. Its reduction in binding is therefore approximately analogous to areduction in motility. In this study, 2mM NEM reduced microtubule spermatid-ESbinding. Although NEM is used as a probe for dynein-like motors, it can potentiallyinterfere with any SH group and therefore its specificity is poor (Penningroth, 1989).EHNA is a structural analogue of adenosine and is a more effective inhibitor ofdynein than kinesin. It has been shown to reduce ATPase activity in cytoplasmic dyneinin testis at 4 mM but not at 0.4mM (Neely and Boekelheide, 1988). It was ineffective inchanging binding in this binding assay. It was not used in the binding assays describedabove.The addition of cytosol from Sertoli cells or MAP enriched fractions from testisdid not alter binding, in this study. It is possible that the linking protein(s) werealready present in sufficient amounts for maximum binding or that the concentration ofproteins in the added fractions used in this assay were too low.The reduced binding with 100 jiM vanadate, in this study, is more puzzling. Theeffects of vanadate on motility and ATPase activity occur at lower levels for cytoplasmicdynein than for kinesin, although they are quite variable between systems. In motilityassays, vanadate blocks ATP by forming ADP-vanadate complexes from ATP and thenDiscussion 302occupying an adjacent site effectively limits access of ATP, reducing transport butincreasing rigor formation. In this study,3MT-spermatid-ES binding was reduced by100 jiM vanadate, but not significantly reduced by 10 jiM vanadate. Porter and Johnson(1989) describe the potential for vanadate and ADP to partially mimic the effect of ATPin dissociating rigor complexes in dynein microtubule binding, If dynein was involved inthis binding assay, this effect may explain the reduced binding by high concentrations ofvanadate (Porter and Johnson, 1989)Binding increased with increased microtubule concentration. Even in thepresence of very high concentrations of microtubules saturation did not occur. Thechange in slope of binding with increased microtubule concentration may reflect theexistence of two binding sites one of high and one of low affinity, or more than oneenzyme.The results of the binding assay suggest that spermatid ESs bind to microtubulesand are released by ATP and GTP. As has been the case with other organelle-microtubulebinding assays, the characteristics of3MT-spermatid-ES binding do not align with asingle known motor. They share properties with both dynein and kinesin. The ATPeffects, on3MT-spermatid-ES binding, are similar to those that occur with kinesin ordynein type motors, (for reviews see Vale, 1987,1990; McIntosh and Porter, 1989),but they are also similar to the the binding properties of ppl7O that have, upon furtherinvestigation been shown to be regulated by phosphorylation/dephosphorylation, andmay not have microtubule-based transport capabilities (Rickard and Kreis, 1991;Scheel and Kreis, 1991a,b). As has been done by Kreis and coworkers, additionalapproaches are required to determine if the linking protein in3MTx-spermatid-ESbinding is a motor.Considering the complexity of the microtubule-spermatid-ES system, as isolatedfor this study, it is possible that more than one motor (and/or accessory or microtubulebinding protein) may contribute simultaneously to the binding properties describedhere. What is clear is that3MT-spermatid-ES binding shares two properties with303known motors: it binds to microtubules and it can be released with nucleotides. Thethird property, whether it participates in microtubule-based transport, requiresfurther study. The findings presented here do not prove, but are consistent with thespermatid-ES microtubule binding being mediated by a mechanoenzyme.The localization of tubulin with immunofluoresence or radiolabelled microtubuleby autoradiography, indicate that microtubules associate with spermatid-ESs. Thisevidence is further supported by the microtubule release by ATP using theseparameters.SUPPORT FOR THE MODELOrganelle-microtubule binding in cells generally provides for the positioning oforganelles and translocation along functional pathways. Microtubule spermatid-ESbinding has been tested in this study because of a suspected function in the positioningand translocation of spermatids within the seminiferous epithelium. The evidencepresented here provides two additional pieces of information for the proposed model ofmicrotubule based-spermatid translocation. First, that Sertoli cell microtubules areoriented with their minus-ends directed toward the apical surface of the cell, andsecondly, that binding occurs between ESs and microtubules. Figure 4-1 is adiagramatic representation of these findings and Figure 4-2 depicts the ES-microtubulelinkage as a motor. Two possible motor complexes are illustrated in which the motor-ER binding may be either direct, or mediated through an accessory membrane protein.Further evidence is needed to determine whether the binding is indeed a mechanoenzymecapable of spermatid-Es transport along microtubule tracks.VogI et al (1 991 a) have described four domains that make up ESs. A fifth domainis provisionally described at the ER microtubule site. The data presented here supportthe possibility that the cytoplasmic domain with its microtubules and links to the ESERare in fact a functional part of ectoplasmic specializations.Discussion 304Figure 4-1: Summary diagram of microtubule-based spermatidtranslocation model. This is the microtubule-based spermatid translocation modelpresented in chapter 1. The results of the microtubule polarity study, that Sertoli cellmicrotubules are oriented with their minus-ends directed toward the apical surface ofthe Sertoli cell, have been added to the model.0C’)1’Discussion 3 06Figure 4-2: Summary diagram of microtubule-spermatid-ES binding:This diagram illustrates the structural details of apical ESs in Sertoli cells. Thestructural link from the ESER, across the actin network, to the Sertoli cell andspermatid membranes are shown. A microtubule situated on the cytoplasmic face of theESER is shown bound to the ESER, consistent with the results of the MT-spermatid-ESbinding assays. Two possible formats for the linkage are shown, one in which the linkingprotein provides a direct linkage, and one in which binding is mediated by an integralmembrane protein receptor. The details of this linkage are speculative. Thatmicrotubule-spermatid-ES binding is releasable with nucleotides and shares someproperties with known mechanoenzymes suggests that the linking protein may be amotor. These findings suggest that the cytoplasmic domain is a functional component ofESs.0ci)0(I)0O5COoctSEEEEo_0oC—Ewt-DoC/)CO<Cl):3-Qw-2C/)0wBibliography 308BIBLIOGRAPHYAchier, C.,D. Filmer, C. Merte and D. Drenckhahn. 1989. Roie of microtubules inpolarized delivery of apical membrane proteins to the brush border of theintestinal epithelium. J. Cell Biol. 109: 179-189.Allan, V.J. and R.D. Vale. 1991. Cell cycle control of microtubule-based membranetransport and tubule formation in vitro. J. Cell Biol. 113 (2): 347-359.Allen, C., and G.G. Borisy. 1974. Structural polarity and directional growth ofmicrotubules of Chlamydomonas flagella. J. Cell Biol. 90: 381 -402.Allen, R.D., N.S. Allen, and J.L. Travis. 1981a. Video-enhanced contrast differentialinterference contrast (AVEC-DIC) microscopy: a new method capable of analyzingmicrotubule related motility in the reticulopodial network of Allogromialaticollaris. Cell Moth. 1: 291-302.Allen, R.D., J. Metuzals, I. Tasaki, S.T. Brady and S.P. Gilbert. 1982. Fast axonaltransport in squid giant axon. Science 218: 1127-1128.Allen, R.D., J.L. Travis, J.H. Hayden, N.S. Allen, A.C. Breuer and L.J. Lewis. 1981b.Cytoplasmic transport: moving ultrastructural elements common to many celltypes revealed by video-enhanced microscopy. Cold Spring Harbour Symposiumon Quantitative Biology. 46:85-89.Allen R.D., D.G. Weiss, J.H. Hayden, D.T. Brown, H. Fujiwake, and M. Simpson. 1985.Gliding movement of, and bidirectional transport along, single nativemicrotubules from squid axoplasm: evidence for an active role of microtubules incytoplasmic transport. J. Cell Biol. 100: 1736-1752.Amlani, S., and A.W. Vogl. 1988. Changes in the distribution of microtubules andintermediate filaments in mammalian Sertoli cells during spermatogenesis. Anat.Rec. 220: 143-160.Amos, L.A. 1987. Kinesin from pig brain studied by electron microscopy. J. Cell. Sci.87: 105-111.Bibliography 309Baas, P.W, M.M. Black, and G.A. Banker. 1989. Changes in microtubules polarityorientation during the development of hippocampal neurons in culture. J. CellBiol. 109 (6): 3085-3094.Baas, P.W, J.S. Deitch, M.M. Black, and G.A. Banker. 1988. Polarity orientation ofmicrotubules in hippocampal neurons: uniformity in the axon and nonuniformityin the dendrite. Proc. Nati. Acad. Sd. 85: 8335-8339.Baas, P.W., L.A. White, and S.R. Heidemann. 1987. Microtubule polarity reversalaccompanies regrowth of amputated neurites. Proc. Nati. Acad. Sd. 84: 5272-5276.Bacallao, R., C. Antony, C. Dotti, E. Karsenti, E.H.K. Steizer, and K. Simmons. 1989. Thesubcetlular organization of Madin-Darby canine kidney cells during theformation of a polarized epithelium. J. Cell Biol. 109: 2817-2832.Batch, W.E. 1989. Biochemistry of interorganelle transport. J. Biol. Chem. 264 (29):16965-16968.Bergen, L.C., and G.G. Borisy. 1980. Head-to-tail polymerization of microtubules invitro. etectron microscope analysis of seeded assembly. J. Cell Biol. 84(1):141-50.Bershadsky, A.D., and J.M. Vasiliev. 1 988a. Cytoskeleton and internal organization of thecell. In Cytoskeleton. Plenum Press, N.Y. 19: 167-201.Bershadsky, A.D., and J.M. Vasiliev. 1988b. Reorganization of Cytoskeleton. InCytoskelton. Plenum Press, N.Y.: 21 7-250.Bershadsky, A.D., and J.M. Vasiliev. 1 988c. Systems of microtubules. In Cytoskeleton.Plenum Press, New York. 73-131.Black, M.M., and P.W. Baas. 1989. The basis of polarity in neurons. Trends in Neurosci.12(6): 211-214.Bloom, G.S., M.C. Wagner, K.K. Pfister, P.L. Leopold and S.T. Brady. 1989. Involvementof microtubles and kinesin in the fast axonal transport of membrane-boundedorganelles. In Cell Movement: Kinesin, Dynein, and Microtubule Dynamics. VolBibliography 31 02. F.D. Warner and J.R. Mcintosh, editors. Alan R. Liss Inc., New York. 321-333.Boekeiheide K., M.D. Neeiy, and T.M. Sioussat. 1989. The Sertoli cell cytoskeleton: atarget for toxicant induced germ cell loss. Toxicol App!. Pharmicol. 101: 373-389.Bomsel, M., R. Parton, S.A. Kuznetsov, T.A. Schroer and J. Gruenberg. 1990.Microtubule- and motor- dependent fusion in vitro between apical andbasoiateral endocytic vesicles from MDCK cells. Cell 62:719-731.Borgers, M., M. DeBrabander, J. van Reempts, F. Awouters, and W.A. Jacob. 1977.Intranuclear microtubules in lung mast cells of guinea pigs in anaphylacticshock. Lab. Invest. 37 (1): 1-9.Borgers, M., F. Thoné, and J.M. van Nueten, 1981. The subcellular distribution ofcalcium and the effects of calcium-antagonists as evaluated with a combinedoxalate-pyroantimonate technique. Acta Histochem. Supp!. XXIV, S. 327-332.Borisy, G.G., 1978. Polarity of microtubles of the mitotic spindle. J. Mol. Biol. 124:565-70.Borisy, G.G., and L.C. Bergen. 1982. A direct method for analyzing the polymerizationkinetics at the two ends of a microtuble. Meth. Cell Biol. 24: 171-187.Borisy, G.G., J.M. Marcum, J.B. Olmstead, D.B. Murphy, and K.A. Johnson. 1974.Purificatrion of tubulin and associated high molecular weight proteins fromporcine brain and characterization of microtubule assembly in vitro. Ann. N. Y.Acad. Sci. 253: 107-132.Bornens, M., M. Paintrand, J. Bergs, M.C. Marty, and E. Karsenti. 1987. Structural andchemical characterization of isolated centrosomes. Cell Motil. Cytoskeleton. 8:238-24 9.Bradford, M.M. 1976. A rapid method for the quantitation of microgram quantities ofprotein utilizing the principle of protein-dye binding. Anal. Biochem. 72: 248-254.Bibliography 3 11Brady, S.T. 1985. A novel brain ATPase with properties expected for the fast axonaltransport motor. Nature. 317: 73-75.Brady, S.T., R.J. Lasek and R.D. Allen. 1982. Fast axonai transport in extruded axoplasmfrom squid giant axon. Science. 218: 1129-1131.Brady, S.T., and K.K. Pfister. 1991. Kinesin interactions with membrane boundedorganelles in vivo and in vitro. J. Cell Sci. Supp. 14: 103-108.Brady, S.T., K.K. Pfister, G.S. Bloom. 1990. A monoclonal antibody against kinesininhibits both anterograde and retrograde fast axonal transport in squid axoplasm.Proc. NatI. Acad. Sci. 87 (3): 1061-1065.Bré, M.H., R. Pepperkok, A.M. Hill, N. Levilliers, W. Ansorge, E.H.K. Stelzer and E.Karsenti. 1990. Regulation of microtubule dynamics and nucleation duringpolarization in MDCK ii cells. J. Cell Biol. 111 (6):3013-3021.Brinkley, B. R. 1985. Microtubule organizing centers. Annu. Rev. Cell Biol. 1:145-172.Brinkley, B.R., S.M. Cox, D.A. Pepper, L. Wible, S.L. Brenner, and R.L. Pardue. 1981.Tubulin assembley sites and the organization of cytoplasmic microtubles incultured mammalian cells. J. Cell Biol. 90: 554-562.Brökelmann, J. 1963. Fine structure of germ cells-and Sertoli cells during the cycle ofthe seminiferous epithelium in the rat. Zeitschrift für Zellforschung. 59: 820-850.Bulinski, J.C. and G.G. Gundersen. 1991. Stabilization and post-translationalmodification of microtubules during cellular morphogenesis. BioEssays.13 (6):285-293.Burnside B. 1989. Microtubule sliding and the generation of force for cell shape change.In Cell Movement: Kinesin, Dynein and Microtubule Dynamics. Vol 2. F. D.Warner and J. R. Mcintosh, editors. New York: Alan R. Liss Inc. 169-1 89.Burton, P.R. 1988. Dendrites of mitral cell neurons contain microtubules of oppositepolarity. Brain Res. 473: 107-115.Bibliography 31 2Burton, P.R., and L.A. Laveri. 1985. The distribution, relationships to other organelles,and calcium-sequestering ability of smooth endoplasmic reticulum in frogolfactory axons. J. Neurosci. 5 (11): 3047-3060.Burton, P.R., and J.L. Paige. 1981. Polarity of axoplasmic microtubules in the olfactorynerve of the frog. Proc. Nati. Acad. ScL 78 (5): 3269-3273.Cherry, L.M. and T.C. Hsu. 1984. Antitubulin immunofluoresence studies ofspermatogenesis in the mouse. Chromosoma. 90: 265-274.Chilcote, T.J., and K.A. Johnson. 1989. Microtubule-dynein cross-bridge cycle and thekinetics of 5’-Adenylyl Imidodiphosphate (AMPPNP) binding. In Cell Movement:The Dynein ATPases. Vol 1. F.D. Warner, P. Sater, l.R. Gibbons, editors. Alan R.Liss Inc., N. Y.: 235-250.Christensen, A.K. 1965. Microtubules in Sertoli cells of the guinea pig testis. Anat. Rec.151: 335a.Clermont, Y. 1972. Kinetics of spermatogenesis in mammals: seminiferous epitheliumcycle and spermatogonial renewal. Physiol. Rev. 52 (1): 198-236.Clermont, Y., J. McCoshen and L. Hermo. 1980. Evolution of the endoplasmic reticulumin the Sertoli cell cytoplasm encapsulating the heads of late spermatids in the rat.Anat. Rec. 196: 83-99.Clermont, Y., and A. Rambourg 1978. Evolution of the endoplasmic reticulum during ratspermiogenesis. Am. J. Anat. 151: 191-212.Cohn, S.A., A.L. Ingold, and J.M. Scholey. 1987. Correlation between the ATPase andmicrotubule translocating activities of sea urchin egg kinesin. Nature. 328:1 60-1 63.Cohn, S.A., A.L. Ingold, and J.M. Scholey. 1989. Quantative analysis of sea urchin eggkinesin-driven microtubule motility. J. Biol. Chem. 264(8): 4290-4297.Cole, A., Mi. Meistrich, L.M. Cherry, and P.K. Trostle-Weige. 1988. Nuclear andmanchette development in spermatids of normal and azh/azh mutant mice. Biol.of Reprod. 38: 385-401.Bibliography 31 3Collins, C.A., and R.B. Vallee. 1989. Preparation of microtubules from rat liver andtestis: cytoplasmic dynein is a major microtubule associated protein. Cell Moth.Cytoskeleton. 14: 491-500.Cramer E. B., and J. I. Gallin. 1979. Localization of submembranous cations to theleading end of human neutrophils during chemotaxis. J. Cell Biol. 82: 369-379.Dabora, S.L., and M.P. Sheetz. 1988. The microtubule-dependent formation of atubulovesicular network with characteristics of the ER from cultured cellextracts. Cell. 54: 27-35.Dedman, J.R., B.R. Brinkley, and A.R. Means. 1979. Regulation of microfilaments andmicrotubules by calcium and cyclic AMP In Advances in Cyclic NucleotideResearch. Vol. 11. P. Greengard, G.A. Robinson, editors. Raven Press, N. Y.: 131-174.Duden, R., W.C. Ho, V.J. Allan and T.E. Kreis. 1990. What’s new in cytoskeletonorganelle interactions? Relationship between microtubules and the Golgiapparatus. Path. Res. Pract. 186 (4): 535-541.Dym, M. 1977. The role of the Sertoli cell in spermatogenesis. In Male ReproductiveSystem. R.D. Yates and M. Gordon, editors. Masson Pub. Co., St. Paul. 155-1 69.Dym, M. and Y. Clermont. 1970. Role of spermatogonia in the repair of the seminiferousepithelium following irradiation of the rat testis. Am. J. Anat. 128: 265-282.Dym, M., and D.W. Fawcett. 1970. The blood-testis barrier in the rat and thephysiological compartmentation of the seminiferous epithelium. Biol. Reprod. 3:308-326.Ekwall, H., A. Jansson, P. Sjoberg, and L. PlOen. 1984. Differentiation of the rat testisbetween 20 and 120 days of age. Arch of Androh 13: 27-36.Ellisman, M.A. and K.R. Porter. 1980. Microtrabecular structure of the axoplasmicmatrix: visualization of crosslinking structures and their distribution. J. CellBiol. 87:464-479.Bibliography 3 1 4Euteneuer, U., L.T. Haimo, and M. Schliwa. 1 989a Microtubule bundles of Reticulomyxanetworks are of uniform polarity. Eur. J. Cell Biol. 49 (2): 373-376.Euteneuer, U., K.B. Johnson and M. Schliwa. 1989b. Photolytic clevage of cytoplasmicdynein inhibits organelle transport in Reticulomyxa.. Eur. J. Cell Biol. 50: 34-40.Euteneuer U., and J.R. Mcintosh. 1981a. Polarity of some motility-relatedmicrotubules. Proc. NatI. Acad. Sd. 78(1): 372-376.Euteneuer, U., and J.R. Mcintosh. 1981b. Structural polarity of kinetochoremicrotubules in PtK1 Cells. J. Cell Biol. 89: 338-345.Evans, L., T.J. Mitchison, and M.W. Kirschner. 1985. influence of the centrosome on thestructure of nucleated microtubules. J. Cell Biol. 100: 1185-1191.Fawcett, D.W. 1975. Ultrastructure and function of the Sertoli cell. In Handbook ofPhysiology, Section 7, Endocrinology, Vol. 5: Male Reproductive System. R.O.Greep, editor. Williams and Wilkins, Baltimore, pp. 21-55.Fawcett, D.W., W.A. Anderson, and D.M. Phillips. 1971. Morphogenic factors influencingthe shape of the sperm head. Dev. Biol. 26(2): 220-251.Fleming, T.P., and M.H. Johnson. 1988. From egg tp epithelium. Ann. Rev. Cell Biol. 4:459-85.Flickinger, C., and D.W. Fawcett. 1967. The junctional specializations of the Sertolicells in the seminiferous epithelium. Anat. Record. 158: 207-222.Forman, D.S., K.J. Brown, and D.R. Livengood. 1983. Fast axonal transport inpermiablized lobster giant axon is inhibited by vanadate. J. Neurosci. 3(6):1279-1288.Franchi, E., and M. Camatini. 1985. Morphological evidence for calcium stores atSertoli-Sertoli and Sertoli-spermatid interrelations. Cell Biol. mt. Rep. 9 (5):441 -446.Franke, W.W., C. Grund, A. Fink, K. Weber, B.M. Jockusch, H. Zentgraf, and M. Osborn.1978a. Location of actin in the microfilament bundles associated with theBibliography 3 1 5junctional specializations between Sertoli cells and spermatids. Biol. Cell. 31:7-14.Franke, W.W., C. Grund, and E. Schmid. 1979. Intermediate-sized filaments present inSertoli cells are of vimentin type. Eur. J. Cell Biol. 19: 269-275.Franke, W.W., M. Hergt, and C. Grund. 1987b. Rearrangement of the vimentincytoskeleton during adipose conversion: formation of an intermediate filamentcage around lipid globules. Cell. 49: 131-141.Gelfand, V. I. 1989. Cytoplasmic microtubular motors. Curr. Opin. Cell Biol. 1: 63-66.Gibbons, I.R. 1988. Dynein ATPases as microtubule motors. J. Cell Chem. 263( 31):15837-15840.Gilbert, S.P., R.D. Allen, and R.D. Sloboda. 1985. Translocation of vesicles from squidaxoplasm on flagellular microtubules. Nature. 315: 245-248.Gilbert, S.P., and R.D. Sloboda. 1989. A squid dynein isoform promotes axoplasmicvesicle translocation. J. Cell Biol. 109: 2379-2394.Gilbert, S.P., and R.D. Sloboda. 1984. Bidirectional transport of fluorescently labeledvesicles introduced into extruded axoplasm of squid Loligo pealeL J. Cell Biol. 99:445-452.Gilbert, S.P., and R.D. Sloboda. 1986. Identification of a MAP2-like ATP-bindingprotein associated with axoplasmic vesicles that translocate on isolatedmicrotubules. J. Cell Biol. 103: 947-956.Gilula, N.B., D.W. Fawcett, and A. Aoki. 1976. The Sertoli cell occluding junctions andgap junctions in mature and developing mammalian testis. Dev. Biol. 50: 142-168.Gondos, B. 1984. Germ cell differentiation and intercellular bridges. In Ultrastructureof Reproduction. J. VanBlerkom, P.M. Motta, editors. Martinus Nyhoff Pub.Boston.Gravis, C.J. 1979. Interrelationships between Sertoli cells and germ cells in the Syrianhamster. Z. mikrosk-anat. Forsch. Leipzig. 93 (2): S 321 -342.Bibliography 3 1 6Gravis, C.J. 1980. Ultrastructural observations on spermatozoa retained within theseminiferous epithelium after treatment with dibutyrl cyclic AMP. Tiss. CelL 12(2): 309-322.Greer, K. and J.L. Rosenbaum. 1989. Post-translational modifications of tubulin. In CellMovement: Kinesin, Dynein and Microtubule Dynamics, Vol 2. F. D. Warner andJ. R. McIntosh, editors. New York, Alan Liss Inc. 47-66.Griswold, M.D. 1988. Protein secretions of Sertoli cells. mt. Rev, of Cyt. 110: 133-156.Grove, B.D., D.C. Pfeitfer, S. Allen, and A.W. yogI. 1990. Immunofluorescencelocalization of vinculin in (“junctional“) ectoplasmic specializations of ratSertoli cells. Am. J. Anat. 188: 44-56.Grove, B. D., and A. W. yogI. 1989. Sertoli cell ectoplasmic specializations: a type ofactin-associated adhesion junction? J. Cell Sd. 93: 309-323.Hackney, D.D. 1988. Kinesin ATPase: rate-limiting ADP release. Biochem. 85: 6314-6318.Hackney, D.D., A.S. Malik, and K.W. Wright. 1989. Nucleotide-free kinesin hydrolyzesATP with burst kinetics. J. Biol. Chem. 264(27): 15943-15948.Haimo, L.T. 1982. Dynein decoration of microtubules-determination of polarity. Meth.Cell Biol. 24: 189-206.Hall, E.S., J. Eveleth and K. Boekelheide. 1991. 2, 5-hexanedione exposure alters therat Sertoli cell cytoskeleton in intermediate filaments and actin. Toxicol. Appi.Pharacol.:111(in press).Handel, M.A. 1979. Effects of coichicine on spermiogenesis in the mouse. J. Embryol.Exp. Morph. 51: 73-83.Hayden, J.H., R.D. Allen, and R.D. Goldman. 1983. Cytoplasmic transport in keratocytes:direct visualization of particle translocation along microtubules. Cell Motil. 3(1): 1—19.Bibliography 31 7Heidemann, S.R. 1991. Microtubule polarity determination based on formation ofprotofilament hooks. Meth. in Enzymol. 196: 469-477.Heidemann, S.R., and U. Euteneuer. 1982. Microtubule polarity determination based onconditions for tubulin assembly in vitro. Meth. Cell Biol. 24 : 207-216.Heidemann, S.R., J.M. Landers, and M.A. Hamborg. 1981. Polarity orientation of axonalmicrotubules. J. Cell Biol. 91: 661-665.Heidemann, S.R., and J.R. McIntosh. 1980. Visualization of the structural polarity ofmicrotubules. Nature. 286: 517-519.Heidemann, S. R., G.W. Zieve, and J.R. McIntosh. 1980. Evidence for microtubulesubunit addition to the distal end of mitotic structures in vitro. J. Cell Biol. 87:152-159.Hems, S., Y.H. Song, H. Willie, E. Mandelkow, and E.M. Mandelkow. 1991. Effect ofMAP2, Map2C, and tau on kinesin-dependent microtubule motility. J. Cell Sci.Supp. 14: 121-124.Hermo L., R. Oko and N. B. Hecht. 1991. Differential post-translational modifications ofmicrotubules in cells of the seminiferous epithelium of the rat: a light andelectron microscope immunocytochemical study. Anat Rec. 229: 31-50.Heuser, J. E. 1989. Mechanisms behind the organization of membranous organelles incells. Curr. Opin. in Cell Biol. 1: 98-102.Hirokawa, N. 1982. Cross-linker system between neurofilaments, microtubules andmembranous organelles in frog axons revealed by the quick-freeze, deep-etchingmethod. J. Cell Biol. 94: 129-142.Hirokawa, N., K.K. Pfister, H. Yorifuji, M.C. Wagner, S.T. Brady, and G.S. Bloom. 1989.Submolecular domains of bovine brain kinesin identified by electron microscopyand monoclonal-antibody decoration. Cell. 56: 867-878.Hirokawa, N., R. Sato-Yoshitake, N. Kobayashi, K.K.Pfister, G.S. Bloom, and S.T. Brady.1991. Kinesin associates with anterogradely transported membranous organellesin vivo. J. Cell Biol. 114 (2): 295-3 02.Bibliography 31 8Hirokawa, N., R. Sato-Yoshitake, T. Yoshida, and T. Kawashima. 1990. Brain dynein(MAPIC) localizes on both antergradely and retrogradely transportedmembranous organelles in vivo. J Cell Biol. 111: 1027-1037.Hisanaga, S., H. Murofushi, K. Okuhara, R. Sato, Y. Masuda, H. Sakai, and N. Hirokawa.1989. The molecular structure of adrenal medulla kinesin. Cell Motil.Cytoskeleton. 12: 264-272.Ho, W.C., V.J. Allan, G. van Meer, E.G. Berger and T.E. Kreis. 1989. Reclustering ofscattered Golgi elements occurs along microtubules. Eur. J. Cell Biol. 48: 250-263.Hollenbeck, P.J. 1989a The distribution, abundance, and subcellular localization ofkinesin. J. Cell Biol. 108 (6): 2335-2342.Hollenbeck, P.J. 1 989b. The transport and assembly of the axonal cytoskeleton. J. CellBiol. 1 08: 223-227.Horio,T., and H. Hotani. 1986. Visualization of the dynamic instability of individualmicrotubules by dark-field microscopy. Nature. 321: 605-607.Horowitz, S.B., L. Lothstein, J.J. Manfredi, W. Mellado, J. Parness, S.N. Roy, P.B.Schiff, L. Sorbara, and R. Zeheb. 1986. Taxol: Mechanisms of action andresistance. Ann. N. Y. Acad. ScL 733-744.Horwitz, S.B., J. Parness, P.B. Schiff, and J.J. Manfredi. 1981. Taxol: a new probe forstudying the structure and function of microtubules. Cold Springs HarborSymposium on Quantitative Biology. 46 (1): 209-226.Houliston, E., S.J. Pickering, and B. Maro. 1987. Redistribution of microtubules andpericentriolar material during the development of polarity in mouseblastomeres. J. Cell Biol. 104: 1299-1308.Hoyle, H.D., and E.C. Raff. 1990. Two Drosophila beta tubulin isoforms are notfunctionally equivalent. J. Cell Biol. 111: 1009-1026.Bibliography 31 9Idriss, H., D.K. Stammers, C.K. Ross and R.G. Burns. 1991. The dynamic instability ofmicrotubules is not modulated by alpha Tubulin Tryosinylation. Cell Moth.Cytoskele ton. 20: 30-37.lnoué, S. 1981. Video image processing greatly enhances contrast quality, and speed inpolarization-based microscopy. J. Cell Biol. 89: 346-356.lnoué, S. 1986. Video Microscopy. New York, Plenum Press.Joshi, H.C., M.J. Palacio, and D.W. Cleveland. 1991. Gamma tubulin is required for invivo microtubule nucleation. J. Cell Biol. 115(3): 170a abst. 991.Kangasniemi, M., A Kaipia, P. Mali, J. Toppari, I. Huhtaniemi, and M. Parvinen. 1990.Modulation of basal and FSH-dependent cyclic AMP production in ratseminiferous tubules staged by an improved transillumination technique. Anat.Rec. 227: 62-76.Kelly, R.B. 1990a. Association between microtubules and intracellular organelles.Curr. Opin. in Ce/I Biol. 2: 105-108.Kelly, R.B. 1990b. Microtubules, membrane traffic, and cell organization. Cell 61: 1-3.Kerr, J.B. 1988. An ultrastructural and morphometric analysis of the Sertoli cellduring the spermatogenic cycle of the rat. Anat. Embryol. 179: 191-203.Kierszenbaum, A.L., C.M. Libanati, and C.J. Tandler. 1971. The distribution of inorganiccations in mouse testis. J. Cell Biol. 48: 314-323.Klein, R.L., S.S. Yen, and A. Thureson-Klein 1972. Critique on the K-pyroantimonatemethod for semiquantitative estimation of cations in conjunction with electronmicroscopy. J. Histochem. Cytochem. 20 (1) 65-78.Koonce, M.P., U. Euteneuer, and M. Schliwa. 1986. Reticulomyxa: a new model system ofintracellular transport. J. Cell Sd. Suppl. 5: 145-159.Koonce, M.P., and M. Schliwa. 1985. Bidirectional organelle transport can occur in cellprocesses that contain single microtubules. J. Cell Biol. 100: 322-326.Bibliography 320Kreis, T. E. 1990. Role of microtubules in the organization of the Golgi apparatus. CellMot. Cytoskeleton. 15: 67-70.Kuriyama, R. 1984. Activity and stability of centrosomes in Chinese hamster ovarycells in nucleation of microtubules in vitro. J. Cell Sd. 66: 277-295.Kuznetsov, S.A.M, and V.1. Gelfand. 1986. Bovine brain kinesin is a microtubuleactivated ATPase. Proc. Nati. Acad. ScL 83: 8530-8534.Lacey, M.L. and L.T. Haimo. 1992. Cytoplasmic dynein is a vesicle protein. J. Biol. Chem.267: (in press- March).Laemmli. U. K. 1970. Clevage of structural proteins during the assembly of the head ofbacteriophage T4. Nature 227: 680-685.Lasek, R.J., and S.T. Brady. 1985. Attachment of transported vesicles to microtubules inaxoplasm is facilitated by AMP PNP. Nature. 316: 645-647.Leblond, C.P. and Y. Clermont. 1952a. Definition of the stages of the cycle of theseminiferous epithelium in the rat. Ann. N. Y. Acad. Sd. 55: 548-573.Leblond, C.P., and Y. Clermont. 1952b. Spermiogenesis of rat, mouse, hamster andguinea pig as revealed by the periodic acid-fuchsin sulfurous acid technique. Am.J. Anat. 90:167-215.Lee, J.C. 1982. Purification and chemical properties of brain tubulin. Meth Cell Biol.24: 9-30.Lee, C., M. Ferguson and L.B. Chen. 1989. Construction of the endoplasmic reticulum. J.Cell Biol. 109: 2045-2055.Linden, M., B.D. Nelson, D. Loncar and J.F. Leterrier. 1989. Studies on the interactionbetween mitochondria and the cytoskeleton. J. Bloen. and Biomem. 21 (4): 507-518.Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall. 1951. Protein measurementwith the Folin phenol reagent. J. Biol. Chem. 193: 265-275.Lye, R.J., C.M. Pfarr, and M.E. Porter. 1989. Cytoplasmic dynein and microtubuletranslocators. In Cell Movement: Kinesin, Dynein, and Microtubule Dynamics. VolBibliography 3212. F.D. Warner and J.R. McIntosh, editors. Alan R. LiSS Inc., New York. 141-154.Lye, R.J., M.E. Porter, J.M. Scholey, and J.R. McIntosh. 1987. Identification of amicrotubule-based cytoplasmic motor in the nematode C. elegans. Cell. 51(2):309-18.Magre, S. and A. Jost. 1991. Sertoli cells and testicular differentiation in the rat fetus.J. Elect. Microsc. Tech. 19: 172-188.Mali, P., M.J. Weish, J. Toppari, K.K. Vihko, and M. Parvinen. 1985. Cell interactionsin the rat seminiferous epithelium with special reference to the cellulardistribution of calmodulin. Med. Biol. 63: 237-244.Malik, F. and R. D. Vale 1990. A new direction for kinesin. Nature 347: 713-714.Mandelkow, E., and E.M. Mandelkow. 1989a. Microtubule structure and tubulinpolymerization in Curr. Opin. Cell Biol. 1: 5-9.Mandelkow, E. and E.-M. Mandelkow. 1989b. Tubulin, microtubules, and oligomers:molecular structure and implications for assembly. In Cell Movement: Kinesin,Dynein, and Microtubule Dynamics. Vol 2. F.D. Warner and J.R. Mcintosh,editors. Alan R. Liss Inc., New York. 23-45.Margolis, R.L. 1982. Measurement of steady-state tubulin flux. Meth. Cell Biol. 24:145-158.Masri, B.A., L.D.Russell, and A.W. VogI. 1987. Distribution of actin in spermatids andadjacent Sertoli cell regions of the rat. Anat. Rec. 218: 20-26.McBeath, E. and K. Fujiwara. 1989. Coalignment of microtubules, cytokeratinintermediate filaments, and collagen fibrils in a collagen-secreting cell system.Eur. J. Cell Biol. 50: 510-521.McDonald, H.B., and L.S.B. Goldstein. 1990. Identification and characterization of a geneencoding a kiriesin-like protein in Drosophila.. Cell. 61: 991-1000.McDonald, H.B., R.J. Stewart, and L.S.B. Goldstein. 1990. The kinesin-like ncd proteinof Drosophila is a minus-end directed microtubule motor. Cell: 1159-1165.Bibliography 322Mcintosh, J. R. and M. E. Porter. 1989. Enzymes for microtubule-dependent motility. J.Biol. Chem. 264 (11): 6001-6004McIntosh, J.R. and U. Euteneuer. 1984. Tubulin hooks as probes for microtubulepolarity: an analysis of the method and an evaluation of data on microtubuleopolarity in the mitotic spindle. J. Cell Biol. 98: 525-533.McKeithan, T. W. and U. L. Rosenbaum. 1984. The biochemistry of microtubules. CellMuscle Motil. 5: 255-288.McNiven, M.A., M. Wang, and K.R. Porter. 1984. Microtubule polarity and the directionof pigment transport reverse simulataneously in surgically severed melanoporearms. Cell. 37: 753-765.McNiven, M.A., and K.R. Porter. 1986. Microtubule polarity confers direction topigment transport in chromatophores. J. Cell Biol. 103: 1547-1555.Means, A.R., J.R. Dedman, J.R. Tash, D. J. Tindall, M. Van Sickle, and M.J. Welsh. 1980.Regulation of the testis Sertoli cell by follicle stimulating hormone. Annu. Rev.Physiol. 42: 59-70.Mitchison, T.J. and M.W. Kirschner. 1984a. Dynamic instability of microtubule growth.Nature. 312: 237-242.Mitchison, T.J. and M.W. Kirschner. 1 984b. Microtubule assembly nucleated by isolatedcentrosomes. Nature. 312: 232-237.Mogensen, M. M., J. B. Tucker, and H. Stebbings. 1989. Microtubule polarities indicatethat nucleation and capture of microtubules occurs at cell surfaces in Drosophila.J. Cell Biol. 108: 1445-1552.Moralis, C., Y. Clermont, and N.J. Nadler. 1986. Cyclic endocytotic activity and kineticsof lysosomes in Sertoli cells of the rat: a morphometric analysis. Biol. of Reprod.34: 207-218.Moralis, C., and V. Clermont. 1982. Evolution of Sertoli cell process invading thecytoplasm of rat spermatids. Anat. Rec. 203: 233-244.Bibliography 323Na, G.C., and S.N. Timasheff. 1982. Physical properties of purified calf brain tubulin.Meth. Enzymol. 85: 393-416.Nagano, T. 1966. Some observations on the fine structure of the Sertoli cell in thehuman testis. Zeitschrift für Zellforschung. 73: 89-106.Neeley, M.D., H.P. Erickson and K. Boekelheide. 1990. HMW-2 the Sertoli cellcytoplasmic dynein from rat testis, is a dimer composed of nearly identicalsubunits. J. Biol. Chem. 265 (15) 8691-8698.Neely, M. D., and K. Boekelheide. 1988. Sertoli cell processes have axoplasmic features:an ordered microtubule distribution and an abundant high molecular weightmicrotubule-associated protein (cytoplasmic dynein). J. Cell Biol. 107: 1767-1776.Neighbors, B.W., R.C. Williams, and J.R. McIntosh. 1988. Localization of kinesin incultured cells. J. Cell Biol. 106: 1193-1204.Nicander, L. 1963. Some ultrastructural features of mammalian Sertoli cells. J.Ultrast. Res. 8: 190-191a.Nicander, L. 1967. An electron microscopical study of cell contacts in the seminiferoustubules of some mammals. Zeitschrift für Zellforschung. 83: 375-397.Ogawa, K., H. Hosoya, E. Yokota, T. Kobayashi, Y. Wakamatsu, K. Ozato, S. Negishi, and M.Obika. 1987. Melanoma dynein: evidence that dynein is a general motor formicrotubule-associated cell motilities. Eur. J. Cell Biol. 43: 3-9.Okabe, S., and N. Hirokawa 1989. Axonal transport. Curr. Opin. Cell Biol. 1: 91-97.Olmstead, J.B. 1986. Microtubule-associated proteins. Annu. Rev. Cell B1oL 2: 421-457.Parvinen, M., and A. Ruokonen. 1982. Endogenesis steroids in the rat seminiferoustubules. Comparison of the stages of the epithelial cycle isolated bytransillumination-assisted microdissection. J. Androl. 3: 211-220.Parvinen, M., and T. Vanha-Perttula. 1972. Identification and enzyme quantitation of thestage of seminiferous epithelial wave in the rat. Anat. Rec. 174: 435-450.Bibliography 324Paschai, B.M., H.S. Shpetner and R.B. Vailee. 1987. MAP 1C is a microtubule-activatedATPase which translocates microtubules in vitro and has dynein-like properties.J. Cell Biol. 105: 1273-1282.Paschal, B.M., H.S. Shpetner, and R.D. VaIlee. 1991. Purification of brain cytoplasmicdynein and characterization of its in vitro properties. Meth. in Enzymol. 196:181-191.Paschal, B.M., and R.B. Vallee. 1987. Retrograde transport by the microtubuleassociated protein MAP1C. Nature 330: 181-183.Penningroth, S.M. 1989. Chemical inhibitors of the dynein adenosine triphosphatases. InCell Movement: The dynein ATPases, Vol I. F.D. Warren, P. Satu, l.R. Gibbons,editors. Alan R. Liss, Inc., New York: 167-179.Pepperkok, R., M.H. Bré, J. Davoust. and T.E. Kreis. 1990. Microtubules are stabilizedin confluent epithelial cells but not in fibroblasts. J. Cell Biol. 111 (6): 3003-3012.Perey, B., Y. Clermont, and C.P. Leblond. 1961. The wave of the seminiferous epitheliumin the rat. Am. J. Anat. 108: 47-77.Pfarr, C.M, M. Cove, P.M. Grissom, T.S. Hays, M.E. Porter, and J.R. McintoSh. 1990.Cytoplasmic dynein is localized to kinetochores during mitosis. Nature 345:263-265.Pfeiffer, D.C., S. Delhar, S.W. Byers, and A.W. VogI. 1991. Evidence that an integrinmay be present at specialized sites (‘ectopiasmic specializations’) ofintercellular adhesion in the seminiferous epithelium. J. Cell Biol. 115(3): p482a abst. 2803.Pfeiffer, D. S. and A. W. yogI., 1991. Evidence that vinculin is co-distributed with actinbundles in ectoplasmic (‘junctional’) specializations of mammalian Sertoli cells.Anat. Rec. 231: 89-100.Pfister, K. K., M. C. Wagner, D. L. Stenoien, S. T. Brady, and G. S. Bloom. 1989b.Monoclonal antibodies to kinesin heavy and light chains stain vesicle-likeBibliography 325structures, but not microtubuies, in cultured cells. J. Cell Biol. 108 (4):1453-1463.Plôen, L, and E.M. Ritzén. 1984. Fine structural features of Sertoli cells. InUltrastructure of reproduction. J. Van Blerkom, P.M. Motta, editors. MartinusNijhoff Pub., Boston. 67-74.Poenie, M. and D. Epel. 1987. Ultrastructural localization of intracellular calciumstores by a new cytochemical method. J. Histochem. Cytochem. 35 (9): 939-956.Porter, K.R. 1966. Cytoplasmic microtubules and their functions. In Ciba FoundationSymposium on Principles of Biomoiecular Organizations. G.E.W. Wolstenholmeand M. O’Connor, editors. London: J & A Churchill Ltd. 308-345.Porter, M. E. and K. A. Johnson. 1989. Dynein structure and function. Annu. Rev. CellBiol. 5:119-151.Porter, M.E., J.M. Scholey, D.L. Stemple. G.P.A. Vigers, R.D. Vale, M.P. Sheetz, and J.R.Mcintosh. 1987. Characterization of the microtubule movement produced by seaurchin egg kinesin. J. Biol. Chem. 262 (6): 2794-2802.Pratt, M. M. 1980. The identification of a dynein ATPase in unfertilized sea urchin eggs.Dev. Biol. 74: 364-378.Pratt, M.M. 1986a. Purification of cytoplasmic dynein from Strongylocentrotus seaurchin eggs. Meth. Enzymol. 134: 325-337.Pratt, M.M. 1986b. Stable complexes of axopiasmic vesicles and microtubuies: proteincompositon and ATPase activity. J. Cell Biol. 103: 957-968.Pratt, M.M. 1989. Cytoplasmic dynein and related adenosine triphosphatases.In CellMovement: Kinesin, Dynein, and Microtubule Dynamics. Vol 2. F.D. Warner andJ.R. Mcintosh, editors. Alan R. Liss Inc., New York. 115-124.Pudney, J. 1986. Fine structural changes in Sertoii and Leydig cells during thereproductive cycle of the ground squirrel Citellus lateralis. J. Reprod. Fert. 77:37-49.Bibliography 326Purvis, K., and V. Hansson. 1981. Hormonal Regulation of Spermatogenesis. mt. J.Andrology Supp.3: 81-143.Rambourg, A., Y. Clermont, and L. Hermo. 1979. Three-dimensional architecture of theGolgi apparatus in Sertoli cells of the rat. Am. J. Anat. 154: 454-476.Rattner, J.B., and B.R. Brinkley. 1972. Ultrastructure of mammalian spermiogenesis.Ill the organization and morphogenesis of the manchette during rodentspermiogenesis. J. Ultrastru. Res. 41: 209-218.Redenbach, D.M. and A.W. VogI. 1991. Microtubule polarity in Sertoli cells: a model formicrotubule-based spermatid transport. Eur. J. Cell Biol. 54:277-290.Redenbach, D.M., K. Boekelheide, and A.W. VogI. 1991. Dynamic binding of spermatid-EScomplexes to microtubules: a model of microtubule based spermatid translocationin mammalian Sertoli cells. J. Cell Biol. 115: 40 abst. 229.Rickard, J.E. and T.E. Kreis. 1991. Binding of ppl7O to microtubules is regulated byphosphorylation. J. Biol. Chem. 266 (26) 17597-17605.Rickard, J.E. and T.E. Kreis. 1990. Identification of a novel nucleotide-sensitivemicrotubule-binding protein in HeLa cells. J. Cell Biol. 110 (5): 1623-1633.Ritzen, E.M., C. Boitani, and M. Parvinen. 1981a. Cyclic secretion of proteins by ratseminiferous tubules depending on the stage of spermatogenesis. mt. J. ofAndrology. 3: 51-58.Ritzen, E.M., V. Hansson, and F.S. French. 1981b. The Sertoli cell. In The Testis. H.Burger and D. de Kretsen, editors. Raven Press. New York. 175-194.Romrell, L.J. and M.H. Ross. 1979. Characterization of Sertoli cell-germ cell junctionalspecializations in dissociated testicular cells. Anat. Rec. 193 (1):23-42.Ross, M.H. 1976a. Sertoli-Sertoli junctions and Sertoli-spermatid junctions afterefferent ductule ligation and lanthanum treatment. Am. J. Anat. 148: 49-56.Ross, M.H. 1976b. The Sertoli cell junction specialization during spermiogenesis and atspermiation. Anat. Rec. 186: 79-104.Bibliography 327Ross, M.H. and J. Dobber. 1975. The Sertoli cell junctional specializations and theirrelationship to the germinal epithelium as observed after efferent ductuleligation. Anat. Rec. 183: 267-292.Rothwell, S.W., W.A. Grasser, and D.B. Murphy. 1985. Direct observations ofmicrotubule treadmilling by electron microscopy. J. Cell B10L 101: 1637-42.Rothwell, S.W., J. Nath, and D.G. Wright. 1989. Interactions of cytoplasmic granuleswith microtubles in human neutrophils. J. Cell Biol. 108: 2313-2326.Russell, L.D. 1977a. Desmosome-like junctions between Sertoli and germ cells in therat testis. Am. J. Anat. 148: 301-312.Russell, L.D. 1979a. Further observations on tubulobulbar complexes formed by latespermatids and Sertoli cells in the rat testis. Anat. Rec. 194 (2): 213-232.Russell, L.D. 1977b. Movement of spermatocytes from the basal to the adluminalcompartment of the rat testis. Am. J. Anat. 148: 313-328.Russell L.D. 1977c. Observations on rat Sertoli ectoplasmic (“junctional”)specializations in the association with germ cells of the rat testis. Tissue Cell 9:475-498.Russell, L.D. 1979b. Observations on the inter-relationships of Sertoli cells at the levelof the blood-testis barrier: evidence for formation and resorption of SertoliSertoli tubulobulbar complexes during the spermatogenic cycle of the rat. Am. J.Anat. 155: 259-280.Russell, L. D. 1980. Sertoli-germ cell interactions: a review. Gamete Res. 3:179-202.Russell, L.D. 1979c. Spermatid-Sertoli tubulobulbar complexes as devices forelimination of cytoplasm from the head region of late spermatids of the rat. Anat.Rec. 194 (2): 233-246.Russell, L.D. 1984. Spermiation-the sperm release process: Ultrastructuralobservations and unresolved problems. In Electromicroscopy in Biology andMedicine. J. van Berkhon and P. Motta editors. Martinus Nijhoff, Boston. 46-66.Bibliography 328Russell, L.D., A. Bartke and J.C. Goh. 1989a. Postnatal development of Sertoli barrier,tubular lumen and cytoskeleton of Sertoli and myoid cells in the rat, and theirrelationship to tubular fluid secretion and flow. Am. J. Anat. 184: 179-189.Russell, L.D., and Y. Clermont. 1976. Anchoring device between Sertoli cells and latespermatids in rat seminiferous tubules. Anat. Rec.185 (3): 259-278.Russell, L.D., R.A. Ettlin, A.P. Sink a Hikim, and E.D. Clegg. 1990. Histological andHistopathological Evaluation of the Testis. Cache River Press, Clearwater,Florida.Russell, L.D., and J.C. Goh. 1988. Localization of aipha-actinin in the rat testis:preliminary observations. In: Development and function of one reproductiveorgans. Ares-Serno Symposium. M. Parvinen, I. Huhtanieme and L. Pelliniemi,editors. Rome Ans Serono Symp. Rev no. 14. Vol. II: 237-244.Russell, L.D., J.C. Goh, R.M.A. Rashed, and A.W. VogI. 1988. The consequences of actindisruption at Sertoli ectoplasmic specialization sites facing spermatids after invivo exposure of rat testis to cytochalasin D. Biol. of Reproduction. 39: 105-118.Russell, L.D., I.P. Lee, R. Ettlin, and R.N. Peterson. 1983a. Development of the acrosomeand alignment, elongation and entrenchment of spermatids in procarbazinetreated rats. Tiss. Cell. 15 (4): 615-626.Russell, L.D. and J.P. Malone. 1981. A study of Sertoli-spermatid tubulobulbarcomplexes in selected mammals. Tiss. Cell. 12(2): 263-285.Russell L.D., J.P. Malone and D.S. MacCurdy. 1981. Effect of the microtubule disruptingagents, colchicine and vinblastine, on seminiferous tubule structure in the rat.Tiss. Cell.13 (2): 349-367.Russell, L.D., P. Myers, J. Ostenburg, and J. Malone. 1980. Sertoli ectoplasmicspecializations during spermatogenesis. In Testis Development Structure andFunction. C. Steinberger and E. Steinberger, editors. Raven Press N. Y. 55-63.Bibliography 329Russell, L.D., and R.N. Peterson. 1985. Sertoli Cell Junctions: Morphological andfunctional correlates. mt. Rev. Cytol. 94: 177-211.Russell, L.D., N.K. Saxena, and T.T. Turner. 1989b. Cytoskeletal involvement inspermiation and sperm transport. Tiss. Cell. 21(3): 361-79.Russell, L.D., M. Tallon-Doran, J.E. Weber, V. Wong, and R.N. Peterson. 1983b. Three-dimensional reconstruction of a rat stage V Sertoli cell: Ill a study of specificcellular relationships. Am. J. Anat. 167: 181-192.Saez, J.M., E. Tabone, M.H. Perrard-Sapori, and M.A. Rivarola. 1985. Paracrine role ofSertoli cells. Med. 8101. 63: 225-236.Sammak, P.J., and G.G. Borisy. 1988. Direct observations of microtubule dynamics inliving cells. Nature. 332: 724-726.Scheel, J., and T.E. Kreis. 1991a. Motor protein independent binding of endocyticcarrier vesicles to microtubules. J. Biol. Chem. 266 (27): 18141-18148.Scheel, J., and T.E. Kreis. 1991b. Motor protein independent binding of endocyticcarrier vesicles to microtubules. J. Cell Biol. 115 (3): 478a (abst. 2779).Schelanski, M.L., F. Gaskin, and C.R. Cantor. 1973. Microtubule assemby in the absenceof added nucleotides. Proc. NatI. Acad. Sd. 70: 765-768.Schiff, P.B., J. Fant, and S. B. Horwitz. 1979. Promotion of microtubule assembly invitro by taxol. Nature 277: 665-667.Schliwa, M. 1984. Mechanism of intracellular organelle transport. In Cell, and MuscleMotility, vol 5. J. W. Shaw, editor. New York, Plenum Press.1-81.Schliwa, M., T. Shimizu, R.D. Vale, and U. Euteneuer. 1991. Nucleotide specificities ofanterograde and retrograde organelle transport in Reticulomyxa areindistinguishable. J. Cell Biol. 112: 1199-1203.Schnapp, B.J. 1986. Viewing single microtubules by video light microscopy. Meth.Enzymol. 134: 561-573.Schnapp, B.J. and T.S. Reese. 1989. Dynein is the motor for retrograde axonal transportor organelles. Proc. NatI. Acad. Sci. 86 (5): 1548-1552.Bibliography 330Schnapp, B.J., R.D. Vale, M.P. Sheetz, and T.S. Reese. 1986. Microtubules and themechanism of directed organelle movement. Ann. N.Y. Acad. Sd. 466: 909-91 8.Schnapp, B.J., R.D. Vale, M.P. Sheetz and T.S. Reese. 1985. Single microtubules fromsquid axoplasm support bidirectional movement of organelles. Cell 40: 455-462.Scholey, J.M., S.A. Cohn, and A.L. Ingold. 1989. Biochemical and mobile properties of seaurchin egg kinesin: Relationship to the pathway of the kinesin-microtubuleATPase. In Cell Movement: Kinesin, Dynein, and Microtubule Dynamics. Vol 2.F.D. Warner and J.R. McIntosh, editors. Alan R. Liss Inc., New York.Schroer, T.A., B.J. Schnapp, T.S. Reese and M.P. Sheetz. 1988. The role of kinesin andother soluble factors in organelle movement along microtubules. J. Cell Biol.107: 1785-1792.Schroer, T. A. and M. P. Sheetz. 1991a. Functions of microtubule-based motors. Annu.Rev. Physiol. 53: 629-652.Schroer, T.A., and M.P. Sheetz. 1989. Role of Kinesin and Kinesin-associated proteins inorganelle transport. In Cell Movement: Kinesin, Dynein, and MicrotubuleDynamics. Vol 2. F.D. Warner and J.R. McIntosh, editors. Alan R. Liss Inc., NewYork. 295-306.Schroer, T.A., and M.P. Sheetz. 1991b. Two activators of microtubules-based vesicletransport. J. Cell Biol. 115 (5): 1309-1318.Schroer, T. A., E. R. Steuer and M. P. Sheetz. 1989. Cytoplasmic dynein is a minus end-directed motor for membranous organelles. Cell 56:937-946.Schulze, E., D.J. Asai, J.C. Bulinski and M. Kirschner. Posttranslational modification andmicrotubule stability. J. Cell Biol. 1 05:2167-2177, 1 987.Sertoli, E. 1865. DelI’ esistenza di particolari cellule ramificate nei canalicoliseminiferi del testicolo umano. B.P. Setchell, translator. Morgagni. 7: 31-39.Sheetz, M.P., R.D. Vale, B. Schnapp, T. Schroer and T. Reese. 1986. Vesicle movementsand microtubule-based motors. J. Cell Sc Suppl. 5: 181-1 88.Bibliography 331Sheetz, M.P. 1989. Kinesin structure and function In Cell Movement: Kinesin, Dynein,and Microtubule Dynamics. Vol 2. F.D. Warner and J.R. Mcintosh, editors. Alan R.Liss Inc., New York. 277-285.Sheetz, M.P., E.R. Steuer, and T.A. Schroer. 1989. The mechanism and regulation of fastaxonal transport. Trends in Neurosci. 12 (11): 474-8.Sherline, P., Y.C. Lee, and L.S. Jacobs. 1977. Binding of microtubules to pituitarysecretory granules and secretory granule membranes. J. Cell Biol. 72: 380-389.Shpetner, H.S., B.M. Paschal, and R.B. VaIlee. 1988. Characterization of themicrotubule-activated ATPase of brain cytoplasm dynein (MAPIC). J. Cell Biol.107 (3): 1001-1009.Skoufias, D. A., T.L. Burgess, and L. Wilson. 1990. Spacial and temporal colocalization ofthe Golgi Appatatus and microtubules rich in detyrosinated tubulin. J. Cell Biol.111: 1929-1937.Slautterback, D.B. 1963. Cytoplasmic microtubules. I. Hydra. J. Cell Biol. 18: 367-388.Sloboda, R.D., and S.P. Gilbert. 1989. Microtubule-associated proteins and intracellularparticle motility. In Cell Movement: Kinesin, Dynein, and MicrotubuleDynamics. Vol 2. F.D. Warner and J.R. Mcintosh, editors. Alan R. Liss Inc., NewYork. 223-232.Smith, R.S. 1972. Detection of organelles in myelinated nerve fibres by dark-fieldmicroscopy. Can. J. Physiol. & Pharm. 50: 467-469.Smith, R. S. 1980. The short-term accumulation of axonally transported organelles inthe region of localized lesions of single myelinated axons. J. Neurocytol. 9: 39-65.Soltys, B.J., and G.G. Borisy. 1985. Polymerization of tubulin in vivo: direct evidencefor assembly onto microtubule ends and from centrosomes. J. Cell Biol. 1682-1689.Bibliography 332Somlyo, A.P. 1984. Cellular site of calcium regulation. Nature 309: 516-517Spicer, S. S., J. H. Hardin, and W. B. Greene. 1969. Nuclear precipitates inpyroantimonate-osmium tetroxide-fixed tissues. J. Cell Biol. 39: 216-221.Stebbings, H. 1988. Microtubule motors: cytoplasmic dynein graduates. Nature. 336:14-15.Stebbings, H. and C. Hunt. 1983. Microtubule polarity in the nutritive tubes of insec