Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Actin-related intercellular adhesion junctions in vertebrate Sertoli cells Pfeiffer, David Carl 1997

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

Item Metadata

Download

Media
831-ubc_1998-272257.pdf [ 42.52MB ]
Metadata
JSON: 831-1.0088784.json
JSON-LD: 831-1.0088784-ld.json
RDF/XML (Pretty): 831-1.0088784-rdf.xml
RDF/JSON: 831-1.0088784-rdf.json
Turtle: 831-1.0088784-turtle.txt
N-Triples: 831-1.0088784-rdf-ntriples.txt
Original Record: 831-1.0088784-source.json
Full Text
831-1.0088784-fulltext.txt
Citation
831-1.0088784.ris

Full Text

ACTIN-RELATED INTERCELLULAR ADHESION JUNCTIONS IN VERTEBRATE SERTOLI CELLS by DAVID CARL PFEIFFER BA, The University of North Carolina at Charlotte, 1987 M . S c , The University of British Columbia, 1990 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES (Department of Anatomy) We accept this thesis as conforming to t h ^ required standard^-THE UNIVERSITY OF BRITISHCOLUMBIA August, 1997 © David Carl Pfeiffer, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date AlovewvUr 1*5. I W DE-6 (2/88) 11 A B S T R A C T In the eutherian mammal seminiferous epithelium, the cytoplasm of Sertoli cells adjacent to certain regions of intercellular attachment exhibits unusual structural complexes. At these sites, a layer of non-contractile, hexagonally packed actin filaments is positioned between the Sertoli cell plasma membrane and an underlying cistern of endoplasmic reticulum. The filament layer along with the endoplasmic reticulum and the plasma membrane involved with attachment are referred to as an "ectoplasmic specialization". Ectoplasmic specializations are found apically at sites of attachment to spermatids and basally at sites of attachment to neighboring Sertoli cells. Similarly structured complexes are not found in other cell types nor in Sertoli cells of other vertebrate species. Ectoplasmic specializations of eutherian mammals have been hypothesized to be a highly specialized form of actin-related adhesion junction. If this is true, then non-mammalian homologues of this junction should be present and, in general, should more closely resemble typical actin-related adhesion junctions found i n other cell types than do mammalian ectoplasmic specializations. I have tested this prediction by conducting a systematic comparative survey of the morphology and composition of actin-related adhesion junctions in Sertoli cells of non-mammalian vertebrates. My results indicate that vertebrate Sertoli cells use actin-related adhesion junctions as a common mechanism for adhesion to elongating spermatids and to neighboring Sertoli cells. In non-mammalian vertebrate classes, the junctions appear to possess contactile properties and, in general, more closely resemble the actin-related adhesion junctions of other cell types then do eutherian ectoplasmic specializations. Structural differences in the junctions between classes indicate that different strategies for intercellular adhesion have evolved at these sites within various vertebrate classes. These data suggest that eutherian mammal ectoplasmic specializations represent a modified form of actin-related adhesion junction. I speculate on what the precursors of iii >' the present-day eutherian junctions may have been like as well as on how and why structural adaptations may have occurred at these sites. iv TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF FIGURES viii ACKNOWLEDGEMENTS xiii CHAPTER 1. GENERAL INTRODUCTION 1 Actin-Related Adhesion Junctions 2 Vertebrate Classification 7 The Vertebrate Germinal Epithelium 8 a) Anamniotes 9 b) Amniotes 13 Adhesion Junctions within the Germinal Epithelium 16 Mammalian Ectoplasmic Specializations 17 Development of the Hypothesis 23 Statement of Hypothesis... 26 Predictions 26 Experimental Approach 26 I, CHAPTER 2. ACTIN-RELATED INTERCELLULAR ADHESION JUNCTIONS IN THE GERMINAL COMPARTMENT OF THE TESTIS IN THE HAGFISH (Eptatretus stouti) AND LAMPREY (Lampetra tridentatus) 28 V Introduction 29 Materials and Methods 32 Results 35 Discussion 39 CHAPTER 3. ACTIN-RELATED INTERCELLULAR ADHESION JUNCTIONS IN THE GERMINAL COMPARTMENT OF THE TESTIS IN THE DOGFISH {Squalus acanthias) AND RATFISH (Hydrolagus colliei) 55 Introduction 56 Materials and Methods 59 Results 62 Discussion 68 CHAPTER 4. ACTIN-RELATED INTERCELLULAR ADHESION JUNCTIONS IN THE GERMINAL COMPARTMENT OF THE TESTIS IN THE BOWFIN {Amia calva) AND GUPPY (Poecilia reticulata) 90 Introduction 91 Materials and Methods 94 Results 96 Discussion 104 vi CHAPTER 5. ACTIN-RELATED INTERCELLULAR ADHESION JUNCTIONS IN SERTOLI CELLS OF THE ROOSTER (Gallus domesticus), TURTLE {Psuedemys scripta) AND ALLIGATOR (Alligator mississippiensis) 135 Introduction 136 Materials and Methods 139 Results 147 Discussion 154 CHAPTER 6. "ECTOPLASMIC SPECIALIZATIONS" IN SERTOLI CELLS OF A MARSUPIAL, THE VIRGINIA OPOSSUM (Didelphus virginiana) 187 Introduction 188 Materials and Methods 191 Results 194 Discussion 198 CHAPTER 7. ACTIN FILAMENTS ASSOCIATED WITH THE BASAL SERTOLI CELL SURFACE IN THE ALLIGATOR (Alligator mississippiensis) AND TURTLE (Psuedemys scripta) 209 Introduction 210 Materials and Methods 212 Results 214 Discussion 217 vii CHAPTER 8. SUMMARY AND CONCLUSIONS 231 BIBLIOGRAPHY 254 LIST OF FIGURES CHAPTER 1 Fig. 1.1. Spermatocyst formation in anamniotes 11 Fig. 1.2. Position of intercellular junctions within a typical epithelial cell and within a eutherian mammal Sertoli cell 19 Fig. 1.3. Electron micrograph of a eutherian mammal ectoplasmic specialization 22 CHAPTER 2 Fig. 2.1. Light and electron micrographs of hagfish germinal epithelium 46 Fig. 2.2. Electron micrographs of hagfish Sertoli cells 48 Fig. 2.3. Actin labelling at sites of intercellular contact between Sertoli cells of the hagfish 50 Fig. 2.4. Light and electron micrographs of lamprey germinal epithelium 52 Fig. 2.5. Junctional sites between lamprey Sertoli cells 54 CHAPTER 3 Fig. 3.1. Sertoli cell/germ cell actin-related junctions in the dogfish 75 Fig. 3.2. Inter-Sertoli cell actin-related junctions in the dogfish.. 77 I X Fig. 3.3. Actin labelling at sites of contact between Sertoli cells and germ cells in the dogfish 79 Fig. 3.4. Actin labelling at sites of contact between Sertoli cell in the dogfish 81 Fig. 3.5. Myosin labelling at sites of contact between Sertoli cells and germ cells in the dogfish 83 Fig. 3.6. Sertoli cell/germ cell actin-related junctions in the ratfish 85 Fig. 3.7. Inter-Sertoli cell actin-related junctions in the ratfish 87 Fig. 3.8. Actin labelling at sites of intercellular contact within the germinal epithelium of the ratfish 89 CHAPTER 4 Fig. 4.1. Light micrographs of toluidene blue stained sections of spermatogenically active bowfin testis 110 Fig. 4.2. Light micrographs of toluidene blue stained sections of post-spermiation bowfin testis 112 Fig. 4.3. Electron micrographs of attachment sites between Sertoli cells and spermatocytes in the bowfin 114 Fig. 4.4. Junctions between apical Sertoli cell processes in the bowfin 116 Fig. 4.5. Light and electron micrographs of post-spermiation bowfin testis 118 Fig. 4.6. Junctional complexes between adjacent Sertoli cells in the bowfin 120 Fig. 4.7. High power electron micrograph of a Sertoli cell junctional complex in the bowfin 122 X Fig. 4.8. Actin and myosin labelling in post-spermiation bowfin testis 124 Fig. 4.9. Light and electron micrographs of guppy spermatocysts 126 Fig. 4.10. Sertoli cell/germ cell contact sites in the guppy 128 Fig. 4.11. Inter-Sertoli cell junctional complexes in the guppy 130 Fig. 4.12. Actin and myosin labelling in guppy spermatocysts 132 Fig. 4.13. Actin and myosin labelling in tangentially sectioned guppy spermatocysts 134 CHAPTER 5 Fig. 5.1. Sertoli cell actin-related junction formed around a spermatid head in the rooster 160 Fig. 5.2. Sertoli cell actin-related junctions formed spermatid heads in the alligator 162 Fig. 5.3. Sertoli cell actin-related junction formed around a spermatid head in the turtle 164 Fig. 5.4. Actin, myosin and vinculin labelling with the seminiferous epithelium of the rooster 166 Fig. 5.5. Controls for actin labelling within the seminiferous epithelium of the rooster 168 Fig. 5.6. Controls for myosin labelling within the seminiferous epithelium of the rooster 170 Fig. 5.7. Controls for vinculin labelling within the seminiferous epithelium of the rooster 172 xi Fig. 5.8. Actin and myosin labelling within the seminiferous epithelium of the alligator 174 Fig. 5.9. Actin and myosin labelling within the seminiferous epithelium of the turtle 176 Fig. 5.10. Controls for myosin labelling within the seminiferous epithelium of the turtle 178 Fig. 5.11. Actin and myosin labelling in Sertoli cell regions surrounding spermatids that have been mechanically dissociated from turtle testis 180 Fig. 5.12. Actin labelling in mechanically fragmented turtle seminiferous epithelium 182 Fig. 5.13. Evidence that Sertoli cell actin-related junctions in the turtle can be induced to contract 184 Fig. 5.14. Quantitative analysis of contraction experiments involving Sertoli cell actin-related adhesion junctions in the turtle 186 CHAPTER 6 Fig. 6.1. Electron micrographs of early-staged Sertoli cell/spermatid actin-related adhesion junctions in the opossum 204 Fig. 6.2. Sertoli cell actin-related adhesion junctions in the opossum 206 Fig. 6.3. Actin labelling in within the seminiferous epithelium of the opossum..., 208 CHAPTER 7 Fig. 7.1. Basal actin filament concentrations in an alligator Sertoli cell 222 Fig. 7.2. Bundles of actin filaments near the base of an alligator Sertoli cell....224 Fig. 7.3. Actin filament bundles near in the basal cytoplasm of alligator Sertoli cells 226 Fig. 7.4. Basal actin filament concentrations in turtle Sertoli cells 228 Fig. 7.5. Actin labelling in the basal cytoplasm of alligator and turtle Sertoli cells 230 CHAPTER 8 Fig. 8.1. Summary diagram of actin-related adhesion junctions within vertebrate Sertoli cells 238 XLLL ACKNOWLEDGEMENTS I would like to express my sincere thanks to the members of my Supervisory Committee, Dr. Virginia Diewert, Dr. Harold Kasinsky and Dr. Tim O'Connor, for critically reading my thesis during its preparation and for offering their helpful comments and suggestions. I would also like to thank Dr. Joanne Weinberg for her help with the statistical analysis of my data. Most of all, I would like to thank my Supervisor, Dr. A. Wayne Vogl, for his guidance and encouragement and, especially, for his friendship. This work was supported by a grant from the Medical Research Council of Canada awarded to Dr. A.W. Vogl. 1 CHAPTER 1 General Introduction 2 A C T I N - R E L A T E D ADHESION JUNCTIONS: Cell adhesion is essential for the ordering of animal cells into tissues and for the subsequent maintenence of tissue architecture. Cells must be able to adhere to one another and/or to their substrate. Moreover, the mechanisms of cell adhesion must be controllable so that cells may be organized into the distinctive, complex three-dimensional patterns of animal tissues. A variety of different mechanisms have evolved that meet this demand. One of the methods of adhesion most frequently seen involves the formation of specialized cortical regions in which adhesion molecules are concentrated in plasma membrane domains that are linked to underlying elements of the cytoskeleton. Cytoskeletal-related adhesion sites are employed by cells in a diverse array of situations, ranging from contiguously organized cells of epithelia to isolated cultured cells in contact with surrounding components of the extracellular matrix. Based on the type of cytoskeletal filament associated with adhesion sites, two general categories of filament-related adhesion sites can be distinguished; intermediate filament-related adhesion sites and actin-related adhesion sites. Each of these categories can be further subdivided into cell/cell and cell/substratum adhesion sites. Intermediate filament-related adhesion sites consist mainly of desmosomes (cell/cell) and hemidesmosomes (cell/substratum) (Schwarz et al., 1990; Garrod, 1993; Gumbiner, 1996). At both sites, intermediate filaments converge on a distinct plaque at the plasma membrane. The plaque is, in part, made up of a series of molecules that link the intermediate filaments to elements in the membrane responsible for intercellular adhesion (Franke et al., 1981; Mueller and Franke, 1983; Borradori and Sonnenberg, 1996). At desmosomes, there are at least two adhesion elements (desmoglein and desmocollin) present and these are members of the cadherin superfamily (Garrod, 1993; Koch and Franke, 1994). Hemidesmosomes appear to morphologically resemble half-desmosomes; however, their molecular composition is 3 very different. The main adhesive element is the oc6(B4 integrin receptor (Sorinenberg et al., 1991). The plaque proteins that link to intermediate filaments to the adhesive elements also differ from those present at desmosomes (Borradori and Sonnenberg, 1996). These linking and adhesion molecules of intermediate filament-related sites differ from those present at actin-related adhesion sites (Geiger et al., 1983). Actin-related adhesion sites, the focus of this thesis, display considerable variety in cytoarchitecture. In polarized epithelial cells, the adhesion sites typically form zonular junctions, termed zonula adherens, that encircle cell apices (Farquhar and Palade, 1963). By contrast, actin-related adhesion junctions between cardiac muscle cells are sheet-like in nature (fascia adherens) (Tokuyasu et al., 1981). Still in other cells, such as ocular lens cells, the adhesion junctions are punctate (Lo, 1988). While the overall appearance of these junctions may differ, the manner through which adhesion is accomplished follows a consistent plan. Actin filaments in the cortical cytoplasm are linked indirectly through a membrane-bound plaque of proteins to transmembrane adhesion proteins. The membrane adhesion proteins then function to physically bind adjacent plasma membranes together or link the plasma membrane to molecules of the extracellular matrix (Geiger et al., 1985a,b; Burridge etal., 1988; Takeichi, 1988; Geiger, 1989; Geiger et. al, 1995; Gumbiner, 1996). Although the precise molecular organization of actin-related adhesion sites has not fully been determined, a number of the proteins that make up these junctions have been identified. Certain proteins are present at both cell/cell and cell/substratum actin-related adhesion junctions while others are restricted to only one of these junction types. At cell/cell actin-related adhesion junctions or adherens junctions, the transmembrane adhesion molecules are classical cadherins. These molecules are directly responsible for the binding of one cell to its neighbor and function in a calcium-dependent, homophilic manner, that is, the cadherin on one cell binds directly to an identical molecule on an adjacent cell if appropriate calcium levels are present 4 (Takeichi, 1991). At the junction site, cadherins are linked indirectly through a series of proteins to the actin cytoskeleton. Specifically, the intracellular domain of the cadherin molecule binds directly to B-catenin (Jou etal., 1995). B-catenin interacts with oc-catenin, which in turn binds to actin filaments (Rimm et al., 1 995) . Interestingly, the molecule plakoglobin can occur in place of B-catenin in the cadherin/catenin complex (Sacco et al., 1995). When either B-catenin or plakoglobin are present in the cadherin/catenin complex, the other molecule is absent (Nathke et al., 1 994) . A variety of other proteins have also been identified within the undercoat of adherens junctions. Some of the better characterized examples include: a-Actinin (a 90-kD actin-binding protein which has been found to interact with the cadherin/catenin complex via the a-catenin molecule - Knudsen et al, 1995), vinculin (a 116 -kD protein that binds directly to actin as well as to the COOH-terminus of oc-actinin -Menkel et al., 1994; Jockusch et al., 1995), zyxin (an 82-kD protein that binds to the NH2-terminus of oc-actinin - Crawford etal., 1992), and radixin (a 82-kD protein that caps the barbed or fast growing ends of actin filaments and which has been postulated to link these ends to the undercoat of the junctions - Tsukita et al., 1 989) . Other proteins known to occur within the adherens junction undercoat include a large number of signal transduction molecules (Yamada and Geiger, 1997). The specific interactions most of these kinases and associated proteins have with other elements of the adherens junction are not yet clear. Adding further complexity to these adhesion sites is the fact that some proteins appear to assume both structural and regulatory roles. The protein B-catenin, for example, is not only involved with physically linking cadherins to the underlying actin cytoskeleton, but is also thought to be a regulatory component of the cadherin/catenin complex and thus of cadherin-mediated adhesion (Balsamo et al . , 1996) . 5 At actin-related cell/substratum adhesion sites or focal adhesions, the number of proteins present is great and the various interactions between them also appear to be complex (Burridge etal, 1988; Jockusch et al., 1995). The transmembrane adhesion molecules at these sites are members of the well-studied integrin family of receptors. These molecules are heterodimers, each comprised of an a and B subunit (Hynes, 1992). The extracellular domains of these receptors bind to elements of the extracellular matrix. While several different integrins have been localized at focal adhesion sites, most studies on integrin-cytoskeletal interactions have focussed on the oc5B1 integrin (the fibronectin receptor). At focal adhesion sites, it is the B subunit of this integrin that appears to interact directly with underlying cytoskeletal-associated proteins. The a subunit appears to assume mainly a regulatory role (Hynes, 1 992) . Directly associating with the cytoplasmic domain of the B1 subunit are the proteins a -actinin (Otey et al., 1990) and talin (a 270 kD protein which is not found at cel l /cel l contacts - Drenckhahn et al., 1988). These two proteins may, in turn, bind directly to actin but they also appear to bind to a number of other proteins present at these sites. Some of the many proteins identified at focal adhesions include: paxillin (a 68 kD protein which likely assumes a regulatory role and which is not found at cel l /cel l contacts - Turner et al., 1990), vinculin (in addition to actin and oc-actinin, appears to bind to talin and paxillin - see Jockusch et al., 1995), zyxin (Beckerle, 1986), and tensin (a large multifunctional protein with structural and regulatory properties - see Jockusch et al., 1995). As is the case at adherens junctions, a large number of signal transduction molecules are also present at focal adhesions. The most abundant protein found at actin-related adhesion junctions is actin. At cell/cell junctions, loose bundles or networks of actin filaments relate to the undersurface of the attachment sites. The filaments making up the bundles or networks are cross-linked to one another by actin-binding proteins and filament polarity within these structures is randomly organized (Hirokawa et al., 1983). At cell/substratum 6 junctions, actin filaments are commonly organized into stress fibers. The actin filaments of stress fibers are cross-linked into parallel bundles; however, the filaments are still loosely bundled and exhibit random polarity. One common feature of both locations is that the actin filaments associate with the myosin II protein and possess contractile properties (Isenberg et al., 1976; Owaribe et al., 1981; Burgess et al . , 1982; Hirokawa et al., 1983; Mooseker, 1985) Many of the proteins that occur at actin-related adhesion sites also occur at other locations within the cell. a-Actinin, for example, which is found at both cell/cell and cell/substratum adhesion sites, is also found in stress fibers where it is thought to cross-link actin filaments (Lazarides and Burridge, 1975). Certain proteins, however, are found exclusively at actin filament-related adhesion sites. Vinculin is one such protein and hence it is considered a marker protein for this form of junction (Geiger, 1982; Geiger et al., 1983). What has become apparent is that actin-related adhesion junctions represent more than an incorporation of the cytoskeleton into the adhesion process. They represent supramolecular complexes in which an elaborate system of molecular interactions occurs. With a variety of signal transduction molecules concentrated at these sites, actin-related adhesion junctions not only mediate adhesion but also appear to play important roles in intracellular signalling (Yamada and Geiger, 1997). The fact that a broad spectrum of cell types exhibit actin-related adhesion junctions indicates that these sites assume fundamental roles in cells in general. Despite the widespread distribution of actin-related junctions amongst cells and that they appear to be essential to tissue organization, these junctions are not well characterized in the germinal epithelium of the vertebrate testis. Most of the information that is available is from studies that have focussed on the mammalian system. Very little is known about the general cell adhesion mechanisms or role act in-7 related adhesion junctions play within the germinal epithelium of non-mammalian vertebrates. VERTEBRATE CLASSIFICATION: The phylum Chordata is subdivided into three subphyla: Urochordata, Cephalochordata, and Vertebrata (some include Hemichordata as a fourth subpylum of Chordata) (Romer, 1970). Common to all members of this phylum are the presence of a notochord and, at least at some point during development, gill slits. The focus of this study is on the largest subdivision of this phylum, Vertebrata. The vertebrates include those chordates that possess a backbone or vertebral column. The members of this subphylum have evolved in diverse directions. Based on distinguishing features, present-day vertebrates are divided into a series of seven classes. These include: Agnatha (hagfish and lampreys), Chondrichthyes (sharks, skates, rays and ratfish or chimaeras), Osteichthyes (bony fish), Amphibia (frogs, toads, newts, salamanders and caecilia), Reptilia (turtles, lizards, snakes, alligators and crocodiles), Aves (birds) and Mammalia (monotremes, marsupials and placental mammals). A further method of grouping vertebrate classes is based on the type of embryonic development present, specifically on whether or not an amniotic membrane is formed around the developing embryo. In amniotes (Reptilia, Aves and Mammalia), this embryonic membrane is present and encloses the embryo in a fluid-filled space. Anamniotes (Agnatha, Chondrichthyes, Osteichthyes and Amphibia) lack an amniotic membrane during development. Anamniotes frequently have a simpler mode of reproduction in which eggs are laid in water and young develop there. This latter method of vertebrate classification proves particularly useful in comparative studies on the male reproductive tract. As will be seen, the basic 8 organization of the germinal epithelium differs considerably between anamniotes and amniotes. Directly related to these organizational differences are differences in the process of spermatogenesis between these two vertebrate groups. THE V E R T E B R A T E GERMINAL EPITHELIUM: The vertebrate testis can be viewed as being made up of two distinct compartments: a germinal compartment and an interstitial compartment (Grier, 1 993; Pudney, 1993). The germinal compartment is lined by an epithelium termed the germinal epithelium or, in more commonly in mammals, the seminiferous epithelium. It is the site of spermatogenesis. The interstitial compartment lies outside of the germinal epithelium, separated from it by a basement membrane, and is the site of androgen-producing Leydig cells as well as connective tissue elements, blood vessels and lymphatic vessels. The germinal epithelium, as defined by Grier (1993), is comprised of three distinct components: Sertoli cells, spermatogenic cells and an acellular basement membrane. All three components must be present for a germinal epithelium to exist. This basic arrangement appears to be an ancient and conserved feature. It is found within all present-day vertebrates and, indeed, within all members of the phylym Chordata (Grier, 1993). Apart from the conservation of this tripartite structure, however, the organization of the epithelium as a whole and the spacial relationships between and amongst its cells are seen to differ between certain vertebrate classes. Two general patterns of organization become apparent. One is shared by anamniote classes and the other is specific to the amniotes. 9 a) Anamniotes: In all anamniotes (Agnatha, Chondrichthyes, Osteichthyes and Amphibia), the basic structural unit in which spermatogenesis occurs is a cyst-like structure termed the spermatocyst (Grier, 1993). The germinal epithelium lines the spermatocyst and, depending on the species, may be formed by a single Sertoli cell and an associated cohort of spermatogenic cells or multiple Sertoli cells and their associated cohorts of spermatogenic cells. In all cases, the Sertoli cells associate with a single isogenetic clone of spermatogenic cells. Development of spermatocysts involves an initial association between a mesenchymal-like Sertoli cell and a single undifferentiated primary spermatogonium (Pudney, 1993, 1995) (Fig. 1.1). In some species, this initial association involves several Sertoli cells and a single spermatogonium. Once a basement membrane has been laid down, the timing of which varies between anamniote classes, the Sertoli cell and spermatogenic cell are separated from the interstitial compartment and are thus termed a spermatocyst (Grier, 1993). Spermatogenesis begins when the primary spermatogonium, which has become completely surrounded by a Sertoli cell, undergoes a series of repeated mitotic divisions to produce a mass of secondary spermatogonia. These spermatogenic cells are housed within the lumen of the spermatocyst, the walls of which are formed by the Sertoli cell. Secondary spermatogonia divide meiotically to produce spermatocytes. With a further meiotic division, these cells yield a new cell stage termed spermatids. Early spermatids enter the process of spermiogenesis. This process involves a complex series of events that dramatically change the morphology of the spermatogenic cells. During spermiogenesis, the spermatid nucleus condenses and, in most species, elongates. Furthermore, a flagellum forms, a reduction in and redistribution of cytoplasm occurs, and, in most classes, an acrosome forms. By the completion of spermiogenesis, maturing spermatids have attained the species specific morphology of spermatozoa. The final 1 0 FIGURE 1.1. Diagram illustrating the general features of spermatocyst formation in anamniotes. A: Initial association of a Sertoli cell (SC) with a primary spermatogonium (PG). B: The early spermatocyst is formed when a Sertoli cell completely surrounds the germ cell. C: The primary spermatogonium then undergoes a number of mitotic divisions to produce secondary spermatogonia (SG). Spermatogenic cells are completely enclosed within the spermatocyst, the walls of which are formed by thin cytoplasmic extensions of the Sertoli cell. D: Following meiotic divisions, spermiogenesis begins with the formation of spermatids (ST). In many anamniote species, the heads of spermatids become positioned within apical recesses of the Sertoli cell. E: A clone of mature spermatozoa is eventually released from the spermatocyst as the walls of the cyst breakdown. (Modified from Pudney, 1993). II 1 2 release of mature spermatozoa from the spermatocyst involves the rupture of the Sertoli cell processes that form the spermatocyst walls. (Pudney, 1993, 1995). With the onset of spermatogenesis, concomitant changes in Sertoli cells are seen as well. The initially mesenchymal-like Sertoli cells begin to differentiate. Variations exist between anamniotes in some of the specifics of Sertoli cell differentiation but, i n general, it appears to involve the development of organelles and frequently a hypertrophy of the Sertoli cell near the latter stages of spermiogenesis (Pudney, 1993). Tight junctions are established between adjacent Sertoli cells and these appear to form a permiability barrier within the epithelium of anamniotes after the meiosis stage of spermatogenesis. Its barrier functions have been demonstrated in species from Osteichthyes and Amphibia (Abraham et al., 1980; Marcaillou and Szollosi, 1 980; Franchi et al., 1982; Bergmann et al., 1983; Parmentier et al., 1985). This barrier is thought to function in a fashion similar to the "blood-testis" barrier of mammals (see below). By later stages of spermatogenesis, Sertoli cells and maturing spermatids develop a close association in many anamniote species (Pudney, 1993). This involves a positioning of each spermatid head (the nucleus and acrosome region) within an apical recess or crypt in the Sertoli cell. The organization of spermatocysts within anamniote classes varies. In the more primitive arrangement (Agnatha and Chondrichthyes), spermatocysts occur as spherical structures. These may be present as isolated units interconnected by mesenteric tissue (hagfish) or they may be grouped together into a distinct, albeit primative, testis (lamprey and chondrichthyan species) (Grier, 1993). A more advanced arrangement is seen in Osteichthyes and Amphibia. In species of these classes, spermatocysts are organized into elongated lobules or form the walls of tubules. Here, the basement membrane does not surround each spermatocyst completely. Rather, it is restricted to the perimeter of the entire lobule or tubule (Grier, 1993). 1 3 b) A m n i o t e s : With the evolution of amniotes, a new anatomical arrangement of the germinal epithelium emerged. The cystic pattern of anamniotes is replaced by a new tubular organization in which a permanent population of Sertoli cells is present. This restructuring has included profound changes in the cellular relationships that exist during spermatogenesis. Unlike the anamniote organization, where each Sertoli cell or group of Sertoli cells is associated with a single clone of developmentally identical spermatogenic cells, each amniote Sertoli cell is associated with several, developmentally different cohorts of spermatogenic cells. Cohorts of spermatogenic cells are not enveloped by Sertoli cells, as occurs in anamniotes. Rather, spermatogenic cells are located laterally between adjacent Sertoli cells with the most immature stages near the base of the epithelium. This produces a structurally complex epithelium consisting of Sertoli cells and multiple stages of developing spermatogenic cells. This tubular organization is present in all modern-day amniote species (Pudney, 1993). Accompanying these changes in the organization of the germinal epithelium have been changes in the morphology of Sertoli cells. In general, Sertoli cells of amniote species are columnar in shape. In the adult, they form a sessile and nondividing population of cells that constitutes the major structural element of the germinal epithelium. They are situated on a basement membrane that separates the epithelium from an underlying lamina propria containing a contractile layer(s) of myoid cells. From this foundation, Sertoli cells the extend toward the tubule lumen. Cytoplasmic processes extend from the lateral and apical surfaces of the cell bodies creating irregular contours. Each Sertoli cell varies in shape from its neighbors. Interposed between and attached to the Sertoli cells are the smaller and more numerous germ cells. Because there is a continual production and upward migration of germ cells through the epithelium, each amniote Sertoli cell, at any given point in time, 1 4 is in contact with many germ cells that are at different stages of differentiation (Fawcett, 1975; de Kretser and Kerr, 1988). During spermatogenesis, amniote germ cells proliferate and pass through the same series of events seen in anamniote classes. The most immature cells (spermatogonia) are located basally within the epithelium. As they begin the process of differentiation, they gradually become more apically positioned. During this upward migration, the cells complete meiosis and enter spermiogenesis. Incomplete cytokinesis during spermatogenesis results in germ cells remaining attached to one another forming isogenetic clones that move through the epithelium as units. During spermiogenesis, spermatids become positioned in apical recesses (crypts) of Sertoli cells. It is within these crypts that the germ cells develop the morphological features characteristic of mature spermatozoa. Eventually, during the process known as spermiation, mature spermatids are released from the epithelium as spermatozoa (reviewed by Russell, 1993a). One significant structural feature that sets amniote Sertoli cells apart from anamniote Sertoli cells and from other epithelial cell types in general, is the position of the inter-Sertoli cell tight junctions. In contrast to other cell types, amniote Sertoli cells form tight junctions with one another close to the base of the epithelium. Consequently, this partitions the germinal epithelium into two compartments (Dym and Fawcett, 1970). A small basal compartment, in which the most immature germ cells are housed, is formed below the junctions while a larger adluminal compartment, in which the more differentiated germ cells are situated, is created above them. These tight junctions form the major structural elements of the "blood-testis barrier" (Dym and Fawcett, 1970; Fawcett et al., 1970; Dym, 1973; Gilula etal . , 1976; Byers e ta l . , 1993). They form a continuous zone of intercellular contact throughout the germinal epithelium, creating a permeability barrier that effectively "seals off" the adluminal compartment from the basal one. Access of immune cells, or antibodies, to the adluminal 1 5 compartment is thus prevented. These permeability junctions also allow the Sertoli cell to produce an adluminal microenvironment essential for normal germ cell maturation (Fawcett, 1975; de Kretser and Kerr, 1988). This basally located permeability barrier has been demonstrated in all amniote classes (Reptilia - Baccetti et al., 1983; Bergmann et al., 1984; Cavicchia and Miranda, 1988;. Aves - Osman et al., 1980; Bergmann et al., 1984; Mammalia - Dym and Fawcett, 1970). A further difference between the blood-testis barrier of amniotes and anamniotes relates to the time point at which a functional permiability barrier is established by Sertoli cells. While in anamniote species the barrier appears to form after the meiosis stage of spermatogenesis is complete, in amniotes a functional barrier is established earlier, at the onset of meiosis (Bergmann et al., 1984; Dym and Fawcett, 1970). During the course of spermatogenesis, generations of spermatogenic cells pass through a fixed series of changes as they differentiate and mature. Directly tied in to these events are changes in the cellular associations amniote Sertoli cells form with the developing germ cells. These changes have been best documented within eutherian mammals (placental mammals) (Fawcett, 1975; de Kretser and Kerr, 1988; Morales and Clermont, 1993). As spermatogenic cells transform in shape and undergo a general apical migration through the epithelium, the contours of Sertoli cells change. An example of this is seen as clones of early spermatocytes (preleptotene spermatocytes) located within the basal compartment of the epithelium traverse the inter-Sertoli cell tight junctions forming the blood-testis barrier to reach the adluminal compartment. At this stage, Sertoli cells form tight junctions both above and below the spermatocytes, creating a transitory "intermediate compartment" (Russell, 1977a). Breakdown of tight junctions above the spermatocytes then permit the germ cells to enter the apical compartment (Russell, 1977c; Byers et al., 1993). A further and more dramatic change in the cellular associations that takes place in the germinal epithelium occurs during spermiogenesis in certain eutherian mammals. 16 In these species, the apical Sertoli cell crypts containing elongating spermatids invaginate deep into the epithelium at one point, then later become positioned in Sertoli cell stalks that extend into the tubule lumen. Spermiation then takes place from this lumenal site (see reviews by Russell, 1980; 1993a). Given the dynamic intercellular relationships that exist within the germinal epithelium of the vertebate testis, the process of maintaining adhesion between the cells of the epithelium is very important. Within the germinal epithelium, this process is governed for a large part by Sertoli cells. Sertoli cells establish and maintain adhesion junctions with each other, with developing spermatogenic cells and with the underlying substratum. In the following section, I review some of the adhesion mechanisms used by cells of the germinal epithelium. ADHESION JUNCTIONS WITHIN THE GERMINAL EPITHELIUM: In epithelial cells in general, cytoskeletal-related adhesion junctions are formed between adjacent cells and between cells and the substratum. Cell/cell adhesion junctions, including both actin-related junctions and desmosomes, are typically clustered together near the apical surface of cells. Here they occur in close proximity to a zone of tight junctions formed between neighboring cells. The inter-epithelial cell tight junctions act to form an apical permeability barrier across the epithelium. The fact that these three junction types characteristically occur together led Farquhar and Palade (1963) to propose the concept of the epithelial cell "junctional complex". A fourth type of intercellular junction, the gap junction, may also occur in the vicinity of the junctional complex, but its presence and its positioning are variable. Gap junctions function mainly in cell/cell communication. Apically located junctional complexes consisting of tight junctions, actin-related adhesion junctions and desmosomes are a common feature of many epithelial cell types. 1 7 Studies on the adhesion mechanisms within the germinal epithelium have focussed mainly on the mammalian system. Typical appearing junctional complexes are not seen within the germinal epithelium of mammals. Instead, a modified form of junctional complex is present basally between neighboring Sertoli cells (Fig. 1.2). At this site, Sertoli cells form extensive tight junctions with one another, thereby establishing the blood-testis barrier. Also occurring at these basal sites, interspersed with the tight junctions, are gap junctions (Nicander, 1967; Dym and Fawcett, 1970 Fawcett et al., 1970; Gilula etal . , 1976; Enders, 1993). Desmosome-like junctions have been reported to occur occasionally between the basal aspects of Sertoli cells as well as between Sertoli cells and certain germ cell stages (Russell, 1977b; Russell and Peterson, 1985; Russell, 1993b). Noticeably absent from Sertoli cell junctional complexes are typical actin-related adhesion junctions such as the zonula adherens that characterize junctional complexes in most other epithelia. However, broad zones of close membrane apposition do occur within Sertoli cell junctional complexes. In mammals, these regions are associated with unique accumulations of cytoplasmic filaments and cisternae of the endoplasmic reticulum and have been termed "ectoplasmic specializations" (Russell, 1977a). Mammal ian E c t o p l a s m i c Spec ia l i za t ions : The plasma membrane of each Sertoli cell lies in close apposition to several different neighboring cells and follows a highly irregular course. Zones of intercellular adhesion with adjacent cells are present at sites along its contour. Some of these adhesion sites correspond to intermediate filament associated desmosome-like junctions (Russell, 1977b, Russell and Peterson, 1985; Russell, 1993b). At two notable locations, however, each Sertoli cell forms broad zones of strong intercellular adhesion from which desmosome-like junctions are generally absent. These zones are found basally at the level of Sertoli cell junctional complexes between adjoined Sertoli cells, and apically 1 8 FIGURE 1.2. Diagram illustrating the position of intercellular junctions within a typical epithelial cell and a eutherian mammal Sertoli cell. In epithelial cells in general, tight junctions, zonulae adherens and desmosomes occur together as an apical junctional complex formed between neighboring cells. The actin filaments of the zonulae adherens are loosely arranged in bundles or as a network. In Sertoli cells of eutherian mammals, tight junctions are formed basally between adjacent cells. Also occurring in the region of these basal tight junctions are desmosome-like junctions and an actin filament complex termed the ectoplasmic specialization. The actin filaments of ectoplasmic specializations are arranged in highly ordered paracrystalline arrays and are linked to underlying cisternae of endoplasmic reticulum. Ectoplasmic specializations are also formed by Sertoli cells around elongating spermatids housed within apical Sertoli cell recesses. Typical Epithelial Cel l Eutherian Mammal Sertoli Cel l 2 0 along the crypts housing elongating spermatids. The intercellular space at these sites is uniform and is reduced from 150-200 A to less than 100 A (Flickinger and Fawcett, 1967; Dym and Fawcett, 1970). Consistently found in association with these sites are arrays of actin filaments situated between the Sertoli cell plasma membrane and an underlying cistern of endoplasmic reticulum (Fig. 1.3). These complexes, consisting of the filament layer, the endoplasmic reticulum and regions of the plasma membrane involved with adhesion, are referred to as ectoplasmic specializations (Russell, 1977a). The actin filaments of ectoplasmic specializations are highly ordered and they collectively assume a unipolar orientation (Toyama, 1976; Vogl et al., 1983, 1986) . The filaments are grouped into bundles (Flickinger and Fawcett, 1967; Dym and Fawcett, 1970; Vogl and Soucy, 1985) in which the filaments are hexagonally packed (Dym and Fawcett, 1970; Franke et al., 1978), with an average interfilament spacing of 10-11 nm (Franke et al., 1978; Grove and Vogl, 1989). Sertoli cell ectoplasmic specializations are dynamic complexes. The apical migration and dramatic shape changes of the developing germ cells during spermatogenesis are marked by the assembly and disassembly of ectoplasmic specializations. As germ cells leave the basal compartment en route for the adluminal compartment, they pass through the blood-testis barrier. This is accompanied by the dissociation of junctions and ectoplasmic specializations above the germ cells and the simultaneous formation of new ones below (Russell, 1977c). Likewise, ectoplasmic specializations located at the apical Sertoli cell/germ cell attachment sites undergo an assembly-disassembly process. While there appears to be interspecies variations within the class Mammalia with respect to the precise germ cell stage that ectoplasmic specializations first associate with (see de Kretser and Kerr, 1988), well-developed ectoplasmic specializations are invariably found around elongating spermatid heads. This relationship is maintained throughout spermatid maturation during which considerable rearrangement of ectoplasmic specialization filaments occurs (Vogl and Soucy, 1 985; 2 1 FIGURE 1.3. Transmission electron micrograph of part of an apical Sertoli cell crypt containing an attached spermatid. Ectoplasmic specializations in the Sertoli cell cytoplasm immediately adjacent to the spermatid are indicated by the black brackets. Ectoplasmic specializations consist of a layer of actin filaments together with a cistern of endoplasmic reticulum on one side of the filament layer and regions of the plasma membrane involved with intercellular attachment on the other. Actin filaments within the layer are often grouped into distinct bundles, one of which is indicated here by the arrowhead. X78.000. Bar = 0.1 um. (Micrograph from Pfeiffer, 1990). 7* 23 Vogl et al., 1991). An eventual dismantling of these complexes takes place during spermiation at which time there is a corresponding loss of adhesiveness at the attachment sites (Ross, 1976; Russell, 1977a; Russell etal., 1980; Suarez-Quian and Dym, 1980; Vogl et al., 1983; Russell, 1993a). DEVELOPMENT OF THE HYPOTHESIS The precise functional relationship that ectoplasmic specializations assume with apical and basal Sertoli cell adhesion sites is unknown. Similar structures are not found in any other cell type. Since their original description by Brokelmann (1 9 6 3 ) , numerous hypotheses have been generated. That the filament type present at ectoplasmic specializations is actin has suggested to some authors that ectoplasmic specializations may possess contractile properties since contractility is a common feature of many actin filament networks found in other cell types. Various hypotheses reflected this view including proposals that, through contraction, ectoplasmic specializations may function to retain elongating spermatids within Sertoli cell crypts (Toyama et al., 1976; Russell, 1977a), or that ectoplasmic specialization contraction may play an active role in spermiation (Gravis, 1978), perhaps by pulling the plasma membrane away from the germ cells: However, more recent findings indicate that, in mammalian species, ectoplasmic specializations are not capable of contraction. Myosin II is absent from the structures and glycerinated models do not contract in the presence of appropriate buffers (Vogl and Soucy, 1985). Other hypotheses of ectoplasmic specialization function have emphasized a more structural role at adhesion sites. Among these are proposals that ectoplasmic specializations are involved in strengthening and supporting Sertoli cell adhesion sites (Russell, 1977a; Vogl and Soucy, 1985), or that ectoplasmic specializations may serve as a scaffolding through which other cytoskeletal elements can act on adhesion sites to 24 influence the positioning and movement of maturing germ cells through the epithelium (Russell, 1977a; Ross, 1976). Common to the majority of hypotheses for ectoplasmic specialization function is the underlying theme of ectoplasmic specialization participation, either direct or indirect, in the process of intercellular attachment. Interestingly, in returning to the original descriptions of ectoplasmic specializations (Brokelmann, 1963; Flickinger and Fawcett, 1967; Nicander, 1967), the suggestion was put forward that functionally ectoplasmic specializations may not differ substantially from other filament-related junctions, where actin or intermediate filaments are involved in the intercellular adhesion process. These early descriptions provided little evidence for any interaction between ectoplasmic specializations and the apical and basal intercellular attachment domains of the plasma membrane. Considerable evidence, however, now indicates that ectoplasmic specializations are functionally associated with these membrane zones, and likely do play a direct role in intercellular adhesion. Observations supporting this conclusion include: (1) Ectoplasmic specializations occur only at sites of intercellular attachment, these being the basal Sertoli/Sertoli cell junctions and the apical Sertoli/spermatid adhesion sites; (2) When spermatids are mechanically separated from Sertoli cells, ectoplasmic specializations remain attached to the spermatids (Russell, 1977a; Franke et al., 1978; Romrell and Ross, 1979; Vogl et al., 1985; Vogl and Soucy, 1985). For ectoplasmic specializations to be removed, addition of trypsin is required, indicating that there is a proteinaceous linkage between ectoplasmic specializations and the spermatids (Romrell and Ross, 1979). (3) Vinculin, a molecular marker for actin-related adhesion junctions, has been demonstrated to be present at ectoplasmic specializations through both immunoblotting and immunolocalization studies (Grove and Vogl, 1989; Grove et al., 1990; Pfeiffer and Vogl, 1991). (4) The dismantling of ectoplasmic specializations during spermatogenesis, as germ cells move through the blood-testis barrier or prior to 25 spermatid release from the epithelium, corresponds to a loss of intercellular adhesiveness. Experimental data further strengthens this point: disruption of ectoplasmic specialization actin filaments with the drug cytochalasin D results in a decrease or loss of adhesion between cells at apical sites and a change in permeability at basal Sertoli cell adhesion sites (Russell et al., 1988; Webber et al., 1988). (5 ) Studies with immunological probes indicate that a member of the integrin family of receptors is present at ectoplasmic specializations (Pfeiffer et al., 1991; Palombi et al., 1992; Salanova et al., 1995). Integrins are a known class of adhesion molecules of which some members have been found at cell/cell contacts in other systems (Carter et al., 1990; Larjava et al., 1990a; Lampugnani et al., 1991). An additional line of evidence that mammalian ectoplasmic specializations function in intercellular adhesion comes from comparative studies. Surprisingly little is known about how Sertoli cells are physically related to both neighboring cells and to the underlying extracellular matrix in non-mammalian vertebrates. In most studies that have been done, particularly in anamniotes, junctions are either not described or can not be clearly described as being actin or intermediate associated; however, there are some notable exceptions. In a study of the ratfish testis, Stanely and Lambert (1985) clearly demonstrated the presence of actin-related adhesion junctions between Sertoli cells and spermatids. Moreover, they demonstrated that the protein myosin II is present at these sites and that the filaments are loosely arranged. In a comparative ultrastructural study of some select non-mammalian vertebrate species, Sprando and Russell (1987a) described microfilament-related junctions between Sertoli cells and spermatids. Actin bundles have also been demonstrated in Sertoli cell regions adjacent to spermatids in lizards (Baccetti et al., 1983). Results of these preliminary studies suggest that structures homologous to ectoplasmic specializations are found in non-mammalian vertebrates, but that, morphologically, the non-mammalian junctions differ substantially from their mammalian counterparts. Significantly, these preliminary 26 studies indicate that the non-mammalian junctions appear to more closely resemble the typical actin-related adhesion junctions of other cell types than do mammalian ectoplasmic specializations. S T A T E M E N T OF HYPOTHESIS Based on cumulative observations of ectoplasmic specialization morphology and the interaction of these structures with Sertoli cell attachment sites, the following hypothesis has been formulated: Ectoplasmic specializations of eutherian mammalian Sertoli cells are a highly specialized form of actin-related intercellular adhesion junction related to the adherens junctions of many other epithelial cell types. The unusual structurual features of these sites in eutherian mammals reflect their specialized nature and that they represent a unique strategy for intercellular adhesion that evolved within this group of vertebrates. PREDICTIONS If eutherian mammalian ectoplasmic specializations are a specialized form of actin-related adhesion junction, then non-mammalian vertebrate homologues of this junction should be present and, in general, they should more closely resemble typical actin-related adhesion junctions of other cell types than do mammalian ectoplasmic specializations. EXPERIMENTAL A P P R O A C H The general approach I have taken to verify these predictions has involved a comparative study of the morphology and composition of actin-related adhesion junctions 27 in Sertoli cells of non-mammalian vertebrates. In representative species of the classes Agnatha, Chondrichthyes, Osteichthyes, Reptilia, Aves and the Mammalian subclass Metatheria, I have determined 1) whether or not actin-related adhesion junctions are present, 2) where these junctions are located in the cell (Sertoli cell/Sertoli cell; Sertoli cell/spermatogenic cell; Sertoli cell/extracellular matrix) and 3) the basic design and composition of these junctions. I have ordered the findings of my study into six chapters (chapters 2 - 7). The first three chapters focus on the actin-related adhesion junctions formed by Sertoli cells of three anamniote classes (Agnatha, Chondrichthyes and Osteichthyes). Each of these chapters focusses on a separate class. In chapters 5 and 6, I examine these Sertoli cell junctions within amniote classes - chapter 5 focussing on the classes Reptilia and Aves and chapter 6 focussing on metatherian (marsupial) mammals. Finally, in chapter 7, I examine the relationships actin filaments have with the basal surface of reptilian Sertoli cells. The results of this comparative study of adhesion junctions of vertebrate Sertoli cells not only provide information relevant to the adhesion hypothesis proposed for eutherian mammal ectoplasmic specializations, but also 1) Provide information on how vertebrates use different intracellular strategies to accomodate variations between species in the organization and kinetics of sperm production, and 2) Provide one of the first descriptions of the comparative cell biology of a junction site. 28 CHAPTER 2 Actin-Related Intercellular Adhesion Junctions in the Germinal Compartment of the Testis in the Hagfish (Eptatretus stouti) and Lamprey (Lampetra tridentatus) 29 I n t r o d u c t i o n The germinal compartment of the vertebrate testis is lined by two distinct populations of cells, Sertoli cells and spermatogenic cells (Fawcett, 1975; Grier, 1993; Pudney, 1993). Together these two cell types combine to form an epithelium termed the germinal epithelium or, in mammals, the seminiferous epithelium. The presence of two different cell types within the epithelium, one of which is permanent (Sertoli cells) and one of which is actively dividing and differentiating (spermatogenic cells), leads to dynamic and often complex cell associations. As a consequence, mechanisms by which intercellular adhesion are established and maintained within the epithelium are of prime importance. To date, the best studied model of intercellular adhesion in the germinal epithelium is the mammalian system. Like other epithelial cell types in general, mammalian Sertoli cells and developing spermatogenic cells express various adhesion molecules on their surfaces. While the full complement and distribution of adhesion molecules expressed in the mammalian seminiferous epithelium is unknown, the emerging pattern appears to be complicated. For example, some adhesion molecules appear to be expressed only at certain stages of testicular development and others appear to be expressed only by specific cell stages during spermatogenesis (Byers et al., 1 993; Munro and Blaschuk, 1996). A second and perhaps more important means by which intercellular attachment is accomplished within the germinal epithelium is through well-defined adhesion junctions. At such sites, adhesive elements in the plasma membrane are concentrated into distinct zones that are related to underyling cytoskeletal filaments. Two forms of these junctions are constructed by mammalian Sertoli cells: intermediate filament junctions of the desmosome type and, at certain locations, a structurally unique actin filament junction thought to be involved with adhesion termed the ectoplasmic 30 specialization (Russell, 1977a). Each of these junction types occurs as part of the junctional complex formed basally between adjacent Sertoli cells. Sertoli cells also form these junctions with germ cells at specific stages during germ cell development. For example, Sertoli cells form mainly desmosome-like junctions with germ cells in the early stages of development (spermatocytes), but form mainly ectoplasmic specializations with germ cells in later stages of differentiation (spermatids) (Russell, 1993b). Although considerable interest and attention has been directed towards the mechanisms of intercellular attachment in the seminiferous epithelium in mammals, relatively little is known about such processes in the non-mammalian vertebrate testis. Initial studies have revealed that there are some fundamental differences in the structure of Sertoli cell intercellular junctions, such as the ectoplasmic specialization, between mammals and non-mammalians vertebrates. In mammals these specialized junctions exhibit unique highly ordered arrays of actin and subsurface cisternae of endoplasmic reticulum (Vogl et al., 1993), whereas in non-mammalian vertebrates they appear to more closely resemble typical actin-related junctions seen in other cell types (Baccetti et al., 1983; Stanely and Lambert, 1985; Sprando and Russell, 1987a). As part of a larger study on the mechanisms of intercellular attachment within the vertebrate testis, I have examined the intercellular junctions found within the germinal epithelium of the hagfish (Eptatretus stouti) and lamprey (Lampetra tridentatus). Hagfish and lamprey are the sole extant members of the class Agnatha, the most primitive group of vertebrates. As such, they represent a potentially valuable source of information on what the early relationships between cells may have been like within the germinal epithelium. Currently very little is known regarding adhesion mechanisms in this class. In this chapter I have focussed on the relationships found within the maturing germinal epithelium of the hagfish and lamprey, i.e. the germinal epithelium related to 3 1 the spermatid stage of germ cell development. In mammals and in many other non-mammalian vertebrates, it is at this latter stage of germ cell differentiation that Sertoli cells not only form junctions between themselves but also form ectoplasmic specializations around spermatids (Russell, 1993b; Vogl et al., 1993). 32 Materials and Methods A total of fifteen large (43-62 cm in length) Pacific hagfish (Eptatretus stouti) were obtained from Pacific Bio-Marine Laboratories, Inc. (Venice, CA, USA). Of these, nine were found to be mature males and were used in this study. Twelve Pacific lamprey (Lampetra tridentatus) (22-26 cm in length) were caught near Port Alberni, B.C., Canada, and then maintained for several months in the Department of Zoology at the University of British Columbia. Eight of these were mature males and were used in this study. Electron Microscopy: Hagfish and lamprey were anesthetized with MS222 (tricaine: 1g/l) buffered with sodium bicarbonate (2g/l), decapitated and the testes dissected out and placed in fixative containing 1.5% paraformaldehyde (Fisher Scientific, Pittsburgh, PA), 1.5% glutaraldehyde, 0.1 M sodium cacodylate buffer (J.B. EM Services, Inc.) (pH 7.3) . Testes were then cut into small blocks (mm3) in fixative and allowed to immersion fix for 2 hrs. Tissue blocks were washed with 0.1 M sodium cacodylate (pH 7.3) and then postfixed in cold (ice) 1% osmium tetroxide (J.B. EM Services, Inc.) in 0.1 M sodium cacodylate for 1 hr. Tissue blocks were next washed with distilled H2O, stained en block with 1.0% uranyl acetate (aqueous) for 1 hr, and again washed with distilled H2O. Following dehydration through a graded series of ethyl alcohols, blocks were embedded i n (Polybed 812 Polysciences, Inc.).. Thick sections were stained with 1% toluidene blue and photographed on a Zeiss Axiophot microscope. Thin sections were stained with uranyl acetate and lead citrate and then photographed on a Philips EM 300 operated at 80 kV. Chemicals and other compounds used for this set of experiments and all others in this study were obtained from Sigma Chemical Co. (St. Louis, MO), unless otherwise noted. 33 Fluorescence Microscopy: (a) Tissue Preparation: Only testes from hagfish were used for localization of actin and myosin II by immunofluorescence. Testes to be used for fluorescence microscopy were dissected out of anesthetized, decapitated hagfish and placed in fixative (3.0% paraformaldehyde in phosphate buffered saline (PBS) - (150 mM NaCI, 5 mM KCI, 3.2 mM Na2HP04, 0.8 mM KH2PO4, pH 7.3). Testes were then cut into blocks of approximately 1 c m 3 . After immersion fixation for 30 min, the tissue was washed three times for 10 min each in PBS buffer, embedded in Tissue-Tek O.C.T. compound (Miles Inc.), and then frozen in liquid nitrogen. Frozen sections of 8um were cut with a cryostat and attached to polylysine-coated slides. The slides were immediately placed into cold ( -20°C) acetone for 5 min and then air dried for approximately 30 min. (b) Localization of Actin: Localization of filamentous actin by fluorescence was accomplished using rhodamine phalloidin (Molecular Probes, Inc.). Tissue sections were rehydrated, for 1 0 min, with PBS containing 0.1% bovine serum albumin (BSA) and then incubated, for 2 0 min, in 1.65 x 1 0 " 6 M rhodamine phalloidin (test for filamentous actin). Controls sections for rhodamine phalloidin staining were incubated, for 20 min, in one of the following: (1) PBS (containing 0.1% BSA) + 1.65 x 1 0 - 6 M rhodamine phalloidin + 7.76 x 10" 4 M phallacidin (Molecular Probes, lnc.)(competitive specificity control); (2) PBS (containing 0.1% BSA) + 7.76 x 1 0 - 4 M phallacidin (control for phallacidin in reagent 1); (3) PBS (containing 0.1% BSA) (control for autofluorescence). All slides were washed with PBS, mounted in 1:1 glycerohPBS containing 0.02% sodium 34 azide, and then examined on a Zeiss Axiophot fitted with a filter set for detecting rhodamine. (c) Localization of Myosin II: The localization of myosin II by immunofluorescence was accomplished using and a rabbit antiserum raised against human platelet myosin II. Tissue sections were rehydrated in TPBS (0.05% Tween-20 in PBS) containing 0.1% BSA and 5% normal goat serum (NGS). The slides were drained then incubated, at 37°C for 1 hr, with primary myosin antiserum at a dilution of 1:50 in TPBS containing 0.1% BSA and 1.0% NGS. After washing three times with TPBS containing 0.1% BSA, the sections were incubated, at 3 7 ° C for 1 hr, with a goat anti-rabbit IgG conjugated to fluorescein (Sigma Co.) diluted 1:32 with TPBS/0.1% BSA. The slides were washed three times with TPBS/0.1% BSA, then mounted with the same buffer diluted 1:1 with glycerol and containing 0.02% sodium azide. Slides were examined on a Zeiss Axiophot fitted with a filter set for detecting fluorescein isothiocyanate. Controls for myosin II staining included: (1) substitution of the immune serum with preimmune serum; (2) substitution of the primary antiserum with buffer; (3 ) substitution both of the primary and of the secondary antibodies with buffer. The primary myosin II antiserum and the matched preimmune serum were kindly provided by Dr. Kegi Fujiwara (National Cardiovascular Center Research Institute, Osaka, Japan). 35 R e s u l t s Hagf ish (Eptatretus stouti): The hagfish testis is made up of clusters of spermatocysts. Each spermatocyst consists of a peripherally located layer of Sertoli cells and a population of spermatogenic cells undergoing synchronous development (Figs. 2.1a,b). As seen in Fig. 2.1a, the stage of germ cell development differs from one hagfish spermatocyst to another. In this chapter I have focussed on the intercellular relationships found within spermatocysts containing germ cells in the latter stages of their differentiation, that is, germ cells undergoing spermiogenesis. Maturing spermatocysts are easily recognizable by the fact that the spermatids they contain possess an elongate head shape (Fig. 2.1b). Ultrastructurally, Sertoli cells of maturing spermatocysts are seen to form a simple squamous to cuboidal epithelium that rests on a basal lamina (Fig. 2.1c). While Sertoli cells extend cytoplasmic processes between early germ cell stages (data not shown), similar processes are not evident between spermatids in maturing spermatocysts. Rather, spermatids appear to lie free within the lumen of the spermatocyst. The lateral borders of Sertoli cells, the regions along which adjacent Sertoli cells contact each other, follow highly convoluted courses. This results in a considerable overlapping and redundancy of the lateral plasma membrane of each cell (Fig. 2.1c). At the apical end of these borders are sites of intercellular contact that are related to an underlying layer of microfilaments (Fig. 2.1 c-e). In high power electron micrographs, the filaments are arranged in a loose network that lacks any well-defined organization (Fig. 2.1d,e). Other forms of cytoskeletally-related junctions, such as desmosomes, do not occur in association with these apically located microfilament junctions. Sertoli cell desmosome-like junctions are seen, although infrequently, in earlier staged 36 spermatocysts. When present, these small "spot" junctions are formed along the basolateral Sertoli cell membrane and show no association with the apically located microfilament junctions (Fig. 2.2a). In this study, I was unable to determine whether Sertoli cells form these desmosome-like junctions with adjacent Sertoli cells or whether they form them with early staged spermatogenic cells. During the course of spermiogenesis, elongating spermatids occasionally lie i n close association with Sertoli cells (Fig. 2.2b). Entrenchment of elongating spermatids within Sertoli cells, however, is not seen. Furthermore, there is no evidence of any form of junctional contact between Sertoli cells and elongating spermatids. In the hagfish, a layer of cytoskeletal filaments is frequently seen close to the basal surface of the Sertoli cell. This layer is restricted to the region immediately adjacent to the basal plasma membrane and appears to be made up of actin filaments (Fig. 2.2c). The filaments often occur in loose bundles in which the filaments are aligned parallel to one another and parallel to the base of the cell. While the filaments course along or near the plasma membrane, distinct attachment plaques or focal adhesion junctions are not obvious. Moreover, no close association of the filaments with organelles such as the endoplasmic reticulum is seen. In fact, organelles appear to be excluded from the narrow zone occupied by the filaments. Interestingly, such basal accumulations of filaments are not as obvious in Sertoli cells lining the walls of earlier spermatocysts (data not shown). When fixed frozen sections are stained with the filamentous actin probe rhodamine phalloidin, fluorescence occurs as linear bands at the margin between neighboring Sertoli cells (Fig. 2.3a). I interpret these sites as corresponding to the microfilament related junctions apparent at the ultrastructural level. In cross sections of spermatocysts, the linear bands often appear as isolated, short, vertical tracts near the luminal surface of the epithelium, while in more grazing sections the bands appear to be continuous with one another. Labelling with rhodamine phalloidin also occurs in the 37 region of the spermatocyst wall (Fig. 2.3a). Presumably the majority of this fluorescence represents the boundary cells located outside of the epithelium. Some of this fluorescence, however, may be due to labelling of actin bundles near the base of Sertoli cells. Labelling within the germinal epithelium with the myosin II antiserum was inconclusive. Specific staining was not observed in any of the controls for rhodamine phalloidin staining (data not shown). Lamprey (Lampetra tridentatus): As in the hagfish, the lamprey testis is made up of spermatocysts, each consisting of a layer of Sertoli cells surrounding a population of spermatogenic cells. During spermiogenesis, spermatids acquire an elongate head shape and are seen to lie free within the spermatocyst lumen (Fig. 2.4a). No association with Sertoli cells is evident. At the ultrastructural level, the Sertoli cell layer in maturing lamprey spermatocysts is seen to be a simple squamous epithelium sitting on a basal lamina. As in the hagfish, the lateral borders of Sertoli cells follow very turtuous courses (Fig. 2.4b). Concentrated at the luminal end of these borders are sites of intercellular contact that are related to underlying elements of the cytoskeleton (Figs. 2.4b,c - 2.5). Ultrastructurally, two distinct forms of filament associated junctions are seen. The first form is more luminal in position than the second and is characterized, in low power micrographs, by an electron dense band that follows the apical zone of contact between two neighboring cells (Fig. 2.4b). This junction type appears to be associated with microfilaments which, in high power micrographs, are seen to form a loose carpet adjacent to the plasma membrane (Figs. 2.4c, 2.5). By appearance, these junctions very closely resemble the microfilament related junctions seen in similar locations in the hagfish. At these sites, adjacent plasma membranes are closely apposed. The second and 3 8 most conspicuous form of filament associated junction is a classically appearing desmosome. (Figs. 2.4c, 2.5). At these sites, adjacent plasma membranes are separated by an intermediate dense line and submembrane plaques are associated with bundles of intermediate filaments (Figs. 2.4, 2.5). By size, the intermediate filaments associated with the desmosomes are clearly larger in diameter than the filaments related to the more apically positioned microfilament junctions. In the lamprey, occasional filaments are seen near the basal plasma membrane of the Sertoli cell but they show no organization and represent at most a minor feature of the cell. Well-developed basal concentrations of actin filaments, as seen in the hagfish, are not obvious. 39 D i s c u s s i o n In this study I examined intercellular junctions within the germinal epithelium of hagfish (Eptatretus stouti) and lamprey (Lampetra tridentatus), representatives of the most primitive group of vertebrates, class Agnatha. While there have been a number of studies on the attachment mechanisms utilized by cells of the mammalian seminiferous epithelium, relatively few such studies have been carried out within the germinal epithelium of non-mammalian vertebrates. Virtually nothing is known about the intercellular adhesion mechanisms in the class Agnatha. In fact, even general ultrastructural information on the agnathan testis is limited. Structurally the agnathan testis has been defined as polyspermatocystic (Grier , 1992). It is comprised of clusters of spermatocysts, each consisting of a single layer of Sertoli cells and an isogenetic population of germ cells (Callard, 1991). The most primitive polyspermatocystic testis within the vertebrate line occurs within the hagfish (Grier, 1993). A discrete organ is not present, rather groups of spermatocysts are linked together by mesentery-like tissue. Germ cell development within individual spermatocysts is synchronous but may differ from one spermatocyst to another (Dodd and Sumpter, 1984). Unlike in hagfish, lamprey spermatocysts are organized together into a distinct organ in which all spermatocysts develop synchronously (Grier, 1 992) . At the completion of spermatogenesis in both hagfish and lamprey, spermatozoa are released directly into the abdominal cavity (Grier, 1993; Pudney, 1993). Sperm ducts are lacking in this class. My results indicate that one of the mechanisms used by agnathan Sertoli cells for intercellular adhesion involves cytoskeletal-related junctions. In other epithelial systems in general, neighboring cells are linked together by an apically located junctional complex consisting of an actin filament-related zone, the zonula adherens, and an underlying layer of intermediate filament-related desmosomes (Farquhar and Palade, 4 0 1963). The actin-related zone is contractile (Owaribe etal . , 1981; Burgess, 1982; Hirokawa et al., 1983; reviewed by Mooseker, 1985) and contains both tight (Madara, 1987; Drenckhahn and Dermietzl, 1988; reviewed by Madara, 1992) and adhesion junctions (Hirokawa etal., 1983; Boiler eta. , 1985; Hirano etal., 1987; Gumbiner, 1992). In the hagfish and lamprey, adjacent, Sertoli cells are also linked together apically by a zone of microfilament-related junctions. The microfilaments appear to be loosely organized as is commonly seen at the contractile zonulae adherens of other cell types. In the hagfish, these areas label with rhodamine phalloidin, indicating that the filament type associated with these junctions is actin. A second form of junction, an intermediate filament-related desmosome, is situated immediately below this zone in the lamprey. In this species, the microfilament-related junctions and the underlying desmosomes together form what closely resembles a "typical" junctional complex as described in other epithelia (Farquhar and Palade, 1963). That agnathan Sertoli cells should employ cytoskeletal-related junctions to help maintain intercellular adhesion is perhaps not that surprising given that it is a common mechanism among epithelial cells in general, including Sertoli cells of higher vertebrates. What is of interest, however, is the structure these junctions and their position within the agnathan germinal epithelium. In mammals, a "modified" junctional complex is present between neighboring Sertoli cells. Unlike those of typical epithelial cells, the mammalian Sertoli cell junctional complex is formed basally within the epithelium. It exhibits several other unusual features including atypical intermediate filament junctions referred to as "desmosome-like" junctions (Russell, 1977b), and a unique non-contractile actin filament junction termed the "ectoplasmic specialization" (Russell, 1977a). In mammals, ectoplasmic specializations consist of highly ordered actin filaments and underlying cisternae of endoplasmic reticulum. The agnathan Sertoli cell junctions, by contrast, appear indistinguishable from those seen in typical junctional complexes of other epithelial cell types. Positionally 4 1 they are located apically within the epithelium. In both species the actin-related zone closely resembles the adherens junctions of other epithelial cells in terms of the filament organization and absence of endoplasmic reticulum. The loose filament organization suggests that, like adherens junctions in general, the actin-related zone in agnathan Sertoli cells may be contractile. I was unable to conclusively demonstrate labelling at these sites with an antibody raised against the myosin II protein; however this may have been due to a species recognition problem between the antibody and antigen. The antibody I used was raised against the mammalian form of non-muscle myosin II. When intermediate filament-related junctions are present below the actin zone, as occurs in the lamprey, they are seen to closely resemble classically described desmosomes of other cell types. Given the structural similarities between the agnathan Sertoli cell junctional complexes and those other epithelial cell types, it would appear there has been little modification by agnathan Sertoli cells on the basic epithelial junctional complex design. While hagfish lack apical desmosomes as part of this complex, a loosely organized actin-related zone, a basic element of the epithelial junctional complex, is present. Although I did not identify tight junctions in this study, one might predict they would also occur apically at the level of these actin filament junctions as is common in other epithelial cell types. In lamprey a typical epithelial junctional complex is seen. This basic junctional complex design also appears to have been retained within the bony fish (Abraham et al., 1980; see also Chapter 4) but has been altered in several other vertebrate classes. For example, initial studies indicate that within the class Chondrichthyes, a form of junctional complex occurs basally between Sertoli cells (Pudney, 1993). It is the site of the Sertoli cell tight junctions, has associated cytoskeletal filaments, and may or may not have a layer of endoplasmic reticulum related to it, depending on the species (see Chapter 3). Among reptiles (Baccetti et al. , 1983; Bergmann et al., 1984; Cavicchia and Miranda, 1988) and birds (Osman et al . , 42 1980; Bergmann et al., 1984), Sertoli cell tight junctions are also situated basally within the epithelium but they appear to lack well defined junctional complexes associated with them. The most extensive modifications on the basic junctional complex design appear to have occurred in seminiferous epithelium of mammals. A particularly interesting feature of the germinal epithelium of both hagfish and lamprey is absence of a close physical relationship between Sertoli cells and elongating spermatids. This is very different from the condition seen in other vertebrate classes. Typically the heads of elongating spermatids become positioned deep within apical recesses of Sertoli cells. Around these spermatid heads Sertoli cells form an actin filament junction, the ectoplasmic specialization. Agnathan Sertoli cells appear not to form any close relationship with elongating spermatids and consequently lack any w e l l -developed attachment junctions with spermatids. The very fact that spermatids of both hagfish and lamprey develop an elongate head shape is of interest as well, since they are known to be (lampreys) or assummed to be (hagfish) external fertilizers (Jespersen, 1975). Generally it is the spermatids of internal fertilizers, not external fertilizers, that acquire an elongate head shape and maintain a close association with Sertoli cells throughout their development. Sperm cells of external fertilizers tend to retain a round head shape and often develop within the lumena of the spermatocysts, at a considerable distance from Sertoli cells. Why hagfish and lampreys display certain features normally associated with internal fertilization is open for speculation. Afzelius (in Nicander, 1969:55) has proposed that the ancestors of lampreys were, in fact, internal fertilizers and that certain ancestral features have been retained. This explanation could well account for features in sperm morphology seen in present day lampreys that are normally associated with internal fertilization. Perhaps a similar switch from internal to external fertilization occurred during hagfish evolution as well. 4 3 In addition to the cytoskeletal elements related to the apical junctional complexes, I observed well-developed filament concentrations in a second location i n hagfish Sertoli cells. In this species, a loose mat of filament bundles occurs close to the basal plasma membrane. Based on their appearance, I interpret these filaments to be actin. Similar bundles of actin filaments have also been reported at the base of Sertoli cells in several species of reptiles (Unsicker, 1973; Baccetti et al., 1983; Chapter 7 ) , a teleost (Arenas et al., 1995), and are a common feature in a variety of other cell types (Wong et al., 1983; Sugimoto et al., 1989; Byers and Fujiwara, 1982; Hay, 1983) . Unlike the basal actin filaments reported in some reptile Sertoli cells (Chapter 7 ) , those seen in the hagfish do not appear to have elements of the endoplasmic reticulum associated with them. When present at the base of the cell, subsurface actin filaments are generally thought to play an important role in the attachment of the cell to the basal lamina (Geiger et al., 1985; Burridge etal., 1985; Geiger, 1989) and, perhaps, some form of contractile role as well. Given the similar appearance of the actin filament bundles at the base of hagfish Sertoli cells with those seen in other cell types, it is not unreasonable to speculate that in the hagfish they may be involved in cell-substratum adhesion as well. If so, then this differs considerably from the mammalian condition where adhesion of the cell to the basal lamina is thought to be achieved in part by intermediate filament-related hemidesmosome-like junctions (Russell, 1977c). In mammals actin filaments are relatively sparse along the basal plasma membrane of the Sertoli cell. Intermediate filaments rather than actin form a loose carpet along the basal surface of mammalian Sertoli cell (Vogl et al., 1983; Vogl 1989; Vogl et al., 1993). Unlike in the hagfish, I did not observe significant concentrations of either actin or intermediate filaments near the base of the cell in the lamprey. In this respect lamprey Sertoli cells appear similar to those of many other species of non-mammalian vertebrates where a well-developed basal filament population is not obvious. 44 In summary, my results indicate that Sertoli cells in the germinal epithelium of both the hagfish and the lamprey are linked together by apically located junctional complexes. These junctional complexes closely resemble the classically described junctional complexes of other epithelial cell types in appearance and location. The actin component of agnathan Sertoli cell junctional complexes, like the adherens component of typical junctional complexes, is loosely organized. In the hagfish a well-developed layer of actin filaments is present at the base of the Sertoli cell as well, suggesting a role in cell-substratum adhesion. Interestingly, Sertoli cells of these species do not form a close physical contact with elongating spermatids and ectoplasmic specializations or other well defined junctions are not observed between Sertoli cells and developing spermatids, despite the fact that the spermatids in each of these species acquire elongate head shapes. Given the ancient lineage of agnathans, it is tempting to speculate that the intercellular attachment mechanisms used by Sertoli cells of present day hagfish and lamprey may resemble those used by early vertebrate Sertoli cells. If so, then my findings suggest that the adhesion mechanisms of early vertebrate Sertoli cells may have closely resembled those used currently by other epithelial cell types, i.e. intercellular adhesion was maintained in part through apically located junctional complexes consisting of actin-related adherens junctions and intermediate filament-related desmosomes. Modifications on this original basic design could have led to the range of Sertoli cell junctional systems seen in present-day vertebrate classes. 45 FIGURE 2.1. Shown in panels (a) and (b) are light micrographs of toluidene blue stained plastic sections of hagfish testis. Panel (a) shows, at low magnification, portions of three spermatocysts. Spermatogenic cells fill the lumena of the spermatocysts. Note that the stage of germ cell development differs between spermatocysts. Younger spermatocysts are filled with spherically-shaped germ cells whereas maturing spermatocysts house elongating spermatids. A maturing spermatocyst is visible at the top of the panel (asterisk). Panel (b) shows, at higher power, a maturing spermatocyst. Elongating spermatids (large arrowheads) are clearly evident within the lumen. The wall of the spermatocyst is formed by a simple squamous epithelium of Sertoli cells (small arrowheads). Densely staining boundary cells occur outside of the epithelium (asterisk). At the ultrastructural level, hagfish Sertoli cells of maturing spermatocysts are seen to be squamous-shaped cells with highly convoluted lateral plasma membranes (arrowheads) (panel c). The Sertoli cell nucleus is indicated (N). Near the lumenal border of the cells, an intercellular junction (curved arrow in panel c) is seen. Shown in panel (d) is a higher power view of two of these apical Sertoli cell junctions. The junctions are related to an underlying layer of microfilaments (arrowheads). As seen at higher power in panel (e), the microfilaments occur as a loosely organized network (arrowheads). Panel (a), X215. Bar = 50 um. Panel (b), X830. Bar = 20 um. Panel (c), X11,100. Bar = 1 um. Panel (d), X21.100. Bar = 1 um. Panel (e), X24.600. Bar = 0.5. 47 FIGURE 2.2. Electron micrographs of hagfish Sertoli cells. Shown in panel (a) is a desmosome-like junction formed between a Sertoli cell (large asterisk) and a second (small asterisk) unidentified cell type (either a Sertoli cell or early spermatogenic cell). Sertoli cell intermediate filaments (small arrowheads) are clearly seen streaming down to a desmosomal attachment plaque (large arrowhead). When present, Sertoli cell desmosome-like junctions are present basally within the epithelium. Junctions between Sertoli cells and elongating spermatids are not formed in the hagfish. As seen in panel (b), elongating spermatids lie free within the spermatocyst lumen. Entrenchment of spermatids within Sertoli cell apical recesses does not occur. The asterisk indicates a spermatid head. Shown in panel (c) are concentrations of loosely organized microfilaments (arrowheads) near the basal plasma membrane of a Sertoli cell. Panel (a), X58.750. Bar = 0.25 um. Panel (b), X15.500. Bar = 1 um. Panel (c ) , X33.780. Bar = 0.5 um. 4 9 FIGURE 2.3. Paired fluorescence (panel a) and phase (panel a') micrographs of a fixed frozen section of a hagfish spermatocyst that has been labelled with the actin probe rhodamine phalloidin. An asterisk in each micrograph indicates the position of the boundary cells which lie outside of the epithelium. In this slightly grazing section, tracts of fluorescence (arrowheads) are clearly seen near the lumenal surface of the epithelium. These tracts appear to form a fairly continuous band near the apex of the epithelium. Positionally, this fluorescence corresponds to the level at which the Sertoli cell junctions are observed ultrastructurally. Strong labelling is also seen in the region of the spermatocyst wall (asterisk). Presumably this fluorescence is emitted by the boundary cells; however, it may also be due in part to the filament bundles seen basally within Sertoli cells. X530. Bar = 20 um. 50 5 1 FIGURE 2.4. Panel (a) is a light micrograph of a toluidene blue stained plastic section of Lamprey testis in which several spermatocysts are visible. The lumen of one spermatocyst is indicated by an asterisk. Note that, at this late stage of spermatogenesis, elongate spermatids fill the lumena of the spermatocysts. A single layer of flattened Sertoli cells (arrowheads) is apparent around the wall of each spermatocyst. At the ultrastructural level, as seen in panel (b), junctional complexes (curved arrow) between adjacent Sertoli cells are visible near the lumenal border of the epithelium. The lateral plasma membranes of Sertoli cells (arrowheads) follow highly convoluted courses. The lumen of the spermatocyst is indicated with an asterisk. At higher magnification (panel c), the inter-Sertoli cell junctional complex can be seen to be made up of two distinct filament-related junctions. Most apical in position is a zone of microfilament-related junctions (small arrowheads). Below this is a desmosome (large arrowhead). Panel (a) X690. Bar = 20 urn. Panel (b) X26,480. Bar = 1 um. Panel (c) X60.680. Bar = 0.25 um. 5 3 FIGURE 2.5. Junctional complexes between adjacent Sertoli cells in the lamprey. Microfilaments of the more apical junctions (bracket) appear to be loosely organized. Intermediate filaments (small arrowheads) can be seen streaming towards the underlying desmosomes (large arrowheads). A highly vaculated residual body (star) is present within one of the Sertoli'cells. Similar appearing residual bodies, cast off by maturing spermatids, are common within spermatocyst lumena at this late stage of spermatogenesis. An asterisk indicates the position of the spermatocyst lumen. X52.250. Bar = 0.5 urn. 5 5 CHAPTER 3 Actin-Related Intercellular Adhesion Junctions in the Germinal Compartment of the Testis in the Dogfish (Squalus acanthias) and Ratfish (Hydrolagus colliei) 56 I n t r o d u c t i o n In the vertebrate testis, the germinal epithelium is formed by two cell types: a somatic population of Sertoli cells and a male gamete population of spermatogenic cells (Fawcett, 1975; Grier, 1993; Pudney, 1993). Intercellular adhesion between cells of this epithelium is maintained for a large part by Sertoli cell adhesion junctions. In mammals, two different and very distinct forms of Sertoli cell adhesion junctions are present, both of which are linked to underlying elements of the cytoskeleton (Russell, 1993b; Vogl et al., 1993). The first is an intermediate filament junction of the desmosome type. The second is a structurally unique form of adhesion junction related to actin filaments which has been given the term "ectoplasmic specialization" (Russell, 1977a). Mammalian Sertoli cells form ectoplasmic specializations at two locations. They are found basally between adjacent Sertoli cells at the level of the Sertoli cell junctional complex. They are also found in a more apical position around the Sertoli cell crypts which house elongating spermatids (Nicander, 1967; Dym and Fawcett, 1970; Flickinger and Fawcett, 1970; Russell, 1977a,b). When compared with actin-related adhesion junctions of other cell types, the mammalian ectoplasmic specialization is unusual in a number of respects. For example, the actin filaments of mammalian ectoplasmic specializations are highly ordered and lack the contractile function which is commonly found at actin-related adhesion junctions in other cell types (Vogl and Soucy, 1985). As an adhesion junction, the mammalian ectoplasmic specialization is also unusual in that it has incorporated into it underlying components of the endoplasmic reticulum. To date, most studies on intercellular adhesion junctions within the germinal epithelium have focussed on the mammalian condition. Preliminary work within the class Chondrichthyes, however, suggests that a modified form of ectoplasmic specialization is present in at least one species of this non-mammalian vertebrate class as well. In the ratfish, Stanley and Lambert (1985) described bundles of Sertoli cell filaments concentrated around the heads of elongating spermatids. It was demonstrated through labelling experiments with NBD-phallacidin and anti-actin antibodies that the filaments were actin. Significantly, the contraction mediating protein myosin II was also detected at these sites. This indicates that, while these sites in the ratfish appear to be analogous to the apical ectoplasmic specializations of mammalian Sertoli cells, fundamental differences exist between the two, such as the potential for contraction. In this chapter I further examine the intercellular junctions formed between Sertoli cells and elongating spermatids and between adjacent Sertoli cells in the ratfish (Hydrolagus colliei). I also examine these junctions in the spiny dogfish (Squalus acanthias), another representative of this class. The class Chondrichthyes is a very old vertebrate group consisting of fish with cartilaginous skeletons. At the evolutionary level, this group represents the emergence of the modern jawed fish. Internal fertilization is a characteristic of all present-day chondrichthyan species (Callard et al., 1988). The class is divided into two subclasses, the Elasmobranchii (sharks, skates, and rays) and the Holocephali (ratfish or chimaera). Although detailed reproductive biology studies have been carried out on relatively few species of this class, results of these investigations indicate that testicular organization and the process of spermatogenesis have been highly conserved in this class, that is, the process appears to be very similar amongst the various members of this class. Spermatogenesis occurs within spermatocysts which are lined by a layer of Sertoli cells. In chondrichthyan species, each Sertoli cell encompasses and isolates an isogenic clone of spermatogenic cells. An individual Sertoli cell and its clone of spermatogenic cells is termed a spermatoblast. The spermatocyst is formed by a mass of spermatoblasts which eventually become peripherally organized around a central lumen (Pudney, 1993, 1995). During the process of spermiogenesis, spermatid heads 58 elongate and become situated in apical recesses of Sertoli cells. It is at this point that Stanley and Lambert (1985) noted ratfish Sertoli cells form ectoplasmic specialization-like junctions around the heads of elongating spermatids. In this chapter I present evidence that chondrichthyan Sertoli cells possess actin-related adhesion junctions and that these junction are present where ectoplasmic specializations are known to occur in mammalian Sertoli cells. These chondrichthyan junctions appear to be similar to actin-associated junctions of other cell types in that they may be contractile. They also share characteristics in common with mammalian ectoplasmic specializations. My results corroborate those of Stanely and Lambert (1985) on the Sertoli cell-spermatid junctions of the ratfish. Moreover, my results extend the observations of this earlier study to include information on the inter-Sertoli cell junctions of the ratfish as well as on the Sertoli cell-spermatid and inter-Sertoli cell junctions of the dogfish. Taken together, my results suggest that the actin-related adhesion junctions at these sites may be a common feature of Sertoli cells in species of this class. 59 Materials and Methods A total of five mature spiny dogfish (Squalus acanthias) and three mature ratfish (Hydrolagus colliei) were used in this study. Two of the dogfish were caught near the Department of Fisheries and Oceans Biological Sciences Branch West Vancouver Laboratory and then transported to the University of British Columbia, where they were used the same day. All other fish in this study were caught near Bamfield, British Columbia by personnel of the Bamfield Marine Station. Of these, the dogfish were maintained for several days in a large holding tank at the Bamfield Marine Station until the time of use, while the ratfish were maintained overnight at the same facility until the time of use. Electron Microscopy: Fish were anesthetized with MS222 (tricaine: 1g/l) buffered with sodium bicarbonate (2g/l). Transection of the cervical spinal cord was then performed while fish were under deep anesthesia. Testicular tissue was dissected out and immediately placed in a fixative containing 1.5% paraformaldehyde (Fisher Scientific), 1.5% glutaraldehyde, and 0.1 M sodium cacodylate (J.B. EM Services, Inc.) (pH 7.3). Tissue was cut into small blocks (mm 2) and allowed to fix by immersion for 2.5 hrs. Following this, some tissue blocks (from one dogfish caught off West Vancouver) were washed in 0.1 M sodium cacodylate and then immediately processed for electron microscopy as previously described (see Chapter 2). Other tissue blocks (from dogfish and ratfish caught near Bamfield) were first stored for approximately 36 hrs in 0.1 M sodium cacodylate before being further processed using standard techniques for electron microscopy. Thick sections were stained with 1% toluidene blue and photographed on a Zeiss Axiophot microscope. Thin sections were stained with uranyl acetate and lead 60 citrate and then photographed on a Philips 300 electron microscope operated at 80 kV. Chemicals and other compounds used for this set of experiments and all others in this study were obtained from Sigma Chemical Co. (St. Louis, MO) unless otherwise noted. Fluorescence Microscopy: (a) Tissue Preparation Testicular tissue to be used for fluorescence microscopy was removed from anesthetized dogfish and ratfish and immediately placed in fixative (3.0% paraformaldehyde in phosphate buffered saline (PBS) - (150 mM NaCI, 5 mM KCI, 3.2 mM Na2HP04, 0.8 KH2P04, pH 7.3). Tissue was cut into blocks ( c m 3 ) and allowed to fix by immersion for 30 min before being further washed and frozen for fluorescence experiments as previously described (see Chapter 2). (b) Localization of Actin and Myosin II Localization of filamentous actin and myosin II in cryosections by immunofluorescence was accomplished using rhodamine phalloidin and a rabbit antiserum raised against human platelet myosin II respectively. The labelling protocols followed for both these probes have been previously described (see Chapter 2). Controls for rhodamine phalloidin included (1) PBS (containing 0.1% BSA) + 1.65 x 10-6 M rhodamine phalloidin + 7.76 x 10-4 M phallacidin (competitive specificity control); (2) PBS (containing 0.1% BSA) + 7.76 x 10-4 M phallacidin (control for phallacidin in reagent 1); (3) PBS (containing 0.1% BSA) (control for autofluorescence). Controls for myosin II staining included: (1) substitution of the immune serum with preimmune serum; (2) substitution of the primary antiserum with buffer; (3) substitution both of 61 the primary and of the secondary antibodies with buffer. The primary myosin 11 antiserum and the matched preimmune serum were kindly provided by Dr. Kegi Fujiwara (National Cardiovascular Center Research Institute, Osaka, Japan). All sections processed for fluorescence were examined on a Zeiss Axiophot photomicroscope fitted with filter sets for detecting rhodamine and fluorescein isothiocyanate. 62 R e s u l t s Dogfish (Squalus acanthias): Spatial Relationships Within the Germinal Epithelium: Within the spermatogenic testis, each Sertoli cell surrounds and isolates a clone of germ cells thereby forming a spermatoblast. During spermiogenesis, spermatid heads begin to elongate as chromatin condenses and cytoplasm is displaced posteriorly. With further elongation each spermatid head develops a distinct helical shape. Concomitant with this is a repositioning of spermatids within the spermatoblast. As the spermatid heads elongate, they become situated in recesses or crypts within the apical Sertoli cell cytoplasm. This begins as a loose bundle of spermatids in the apex of the Sertoli cell in which the spermatids are oriented such that their acrosomes point toward the base of the Sertoli cell. As spermiogenesis proceeds, spermatid heads gather closer together and eventually form a compact bundle close to the basal surface of the Sertoli cell. (a) Junctions Between Sertoli Cells and Spermatogenic Cells: At the ultrastructural level, elongating spermatids can easily be seen within the apical Sertoli cell recesses (3.1a). Also readily apparent is a dense layer of microfilaments in the Sertoli cell cytoplasm immediately adjacent to each spermatid head (Figs. 3.1a,b). These filaments closely follow the contour of the spermatid head and, although not highly organized, tend to be oriented parallel to the long axis of the spermatid head. The filaments extend beyond the spermatid head down toward the base of the Sertoli cell where, when viewed in cross-section, they are seen to form a hollow filamentous sheath (Fig. 3.1c). Elements of the endoplasmic reticulum are frequently 63 seen near these filaments but are not a consistent feature. No direct linkage or association between these Sertoli cell microfilaments and the endoplasmic reticulum is apparent (Fig. 3.2a). At higher magnifications, a site of apparent contact between the Sertoli cell and each elongating spermatid is visible (Fig. 3.1 d). In these regions, the Sertoli cell plasma membrane is closely and uniformly applied to the adjacent spermatid. In this same location, there are detectable linkages between adjacent plasma membranes and very distinct linkages between the acrosome and plasma membrane of the spermatid. The Sertoli cell microfilaments appear to be closely applied to this site. (b) Junctions Between Adjacent Sertoli Cells: Each spermatocyst is formed by numerous Sertoli cells. Adjacent Sertoli cells contact each other along their lateral plasma membranes. At the utrastructural level, sites of intercellular contact can be seen basally between neighboring Sertoli cells (Fig. 3.2b). These sites consist of a layer of indistinct filamentous material next to the plasma membrane (Fig. 3.2c). Deep to the filamentous layer, and apparently in contact with it, are cisternae of endoplasmic reticulum. Focal sites of membrane fusion between adjacent Sertoli cells, presumably tight junctions, are occationally seen at the level of the filament associated contact sites (Fig. 3.2d). Of interest is the observation that, in material in which the Sertoli cell cytoplasm has been largely extracted as an artefact of tissue preparation, the filamentous layer and underlying endoplasmic reticulum remain associated with the Sertoli cell plasma membrane (Fig. 3.2e). This indicates that the filamentous layer and underlying endoplasmic reticulum are firmly bound to each other as well as to the plasma membrane at these contact sites. 64 (c) Fluoresecence Localization of Actin and Myosin II: When stained with the filamentous actin probe rhodamine phalloidin, Sertoli cell regions related to spermatid bundles label intensely (Fig. 3.3a). Specifically, the staining is detected around the spermatid heads and corresponds to those regions in which dense concentrations of Sertoli cell microfilaments are seen at the ultrastructural level (see above). Weak labelling with rhodamine phalloidin is also detected around the margins of Sertoli cells, particularly when the plane of section has grazed through the wall of a spermatocyst (Fig. 3.4). The site of this labelling corresponds to the filamentous junctions observed between adjacent Sertoli cells at the ultrastructural level. Staining with rhodamine phalloidin was greatly reduced when blocked with phallacidin (Fig. 3.3b), indicating that the rhodamine phalloidin I used was specifically binding to filamentous actin. No fluorescence was detected in any of the other control slides for rhodamine phalloidin (Figs. 3.3c,d). Spermatocysts treated with an immunological probe raised against the myosin 11 molecule (Fig. 3.5a) yield a staining pattern around the spermatid heads which is very similar to that seen with rhodamine phalloidin. This specific and intense fluorescence corresponds to those regions containing the Sertoli cell microfilaments. Unlike with rhodamine phalloidin, I did not detect any significant myosin II staining around the margins of Sertoli cells. No staining as described above was observed in any of the control slides for myosin II; however, some diffuse nonspecific staining occurred in the basal cytoplasm of Sertoli cells when sections were incubated with preimmune serum in place of the antiserum (Fig. 3.5b). No fluorescence was detected in any of the other control slides for the myosin II antiserum ( Figs. 3.5c,d). 65 No difference in staining with rhodamine phalloidin or with the myosin 11 antiserum was noted between the tissue from the West Vancouver dogfish (tissue processed the same day) and the tissue from the Bamfield dogfish (tissue stored in buffer after fixation). Ratf ish (Hydrolagus colliei) Spatial Relationships Within the Germinal Epithelium: The organization of Sertoli cells and spermatogenic cells within the active spermatocyst of the ratfish very closely resembles that seen in the dogfish. As in the dogfish, elongating spermatids become positioned within apical recesses of Sertoli cells. (a) Junctions Between Sertoli Cells and Spermatogenic Cells: At the ultrastructural level, the heads of elongating spermatids can be seen embedded within the apical cytoplasm of Sertoli cells. The anterior end of each spermatid head is surrounded by a dense layer of Sertoli cell microfilaments which appear to be loosely organized into a sheath. As in the dogfish, the microfilaments extend deep into the Sertoli cell cytoplasm from the spermatid head (Figs. 3.6a,b) In general, the filaments are oriented parallel to the long axis of the spermatid head. In cross section (Fig. 3.6c) , the filaments are seen to be particularly concentrated on one side of the spermatid head. At later stages of spermiogenesis, elongating spermatid heads are gathered tightly together and move deep within the Sertoli cell cytoplasm. The Sertoli cell cytoplasm in which the spermatid bundle is situated is largely devoid of organelles (Fig. 3.6d). Elements of the endoplasmic reticulum are not seen around individual microfilament 66 sheaths; however, cisternae of endoplasmic reticulum are seen around the perimeter of the whole bundle of spermatid heads (Fig. 3.6d). (b) Junctions Between Adjacent Sertoli Cells: As in the dogfish, adjacent Sertoli cells in the ratfish are in contact with each other along their lateral plasma membranes. At the ultrastructural level, microfilament associated contact sites are seen basally between Sertoli cells. The microfilaments appear to be loosely organized and are more heavily concentrated along these basal contact sites than in the dogfish (Fig. 3.7a,b). Cisternae of endoplasmic reticulum are occasionally seen in the vicinity of the microfilaments but do not appear to associate with them as is seen in the dogfish. Interspersed along these contact sites are tight junctions (Fig. 3.7b and inset). (c) Fluoresecence Localization of Actin and Myosin II: In fixed frozen sections of ratfish spermatocysts, the overall pattern of fluorescence obtained with rhodamine phalloidin (Figs. 3.8a,b) is very similar to that obtained in the dogfish with this actin probe; that is, specific fluorescence occurs in (1 ) regions associated with the heads of elongating spermatids, and (2) along the lateral borders of Sertoli cells. Staining associated with the heads of elongating spermatids is intense and more extensive than that seen in the dogfish. This is quite evident when the actin staining in the ratfish (Fig. 3.8a) is compared with the actin staining in the dogfish (Fig. 3.3a) (both micrographs are printed at the same magnification). Staining associated with the lateral borders of ratfish Sertoli cells also appears more intense than that seen in the dogfish. As in the dogfish, this staining is seen most easily in 67 grazing sections of spermatocysts (Fig. 3.8b) where it is apparent as a continuous band a fluorescence around the margins of Sertoli cells. No specific fluorescence was detected in any of the control slides for rhodamine phalloidin staining (data not shown). I was unable to detect any fluorescence in ratfish spermatocysts stained with the antiserum raised against myosin II. 68 Discussion In this chapter I provide evidence that Sertoli cells in the ratfish (Hydrolagus colliei) and in the spiny dogfish (Squalus acanthias), two different species from the class Chondrichthyes, form actin filament-related junctions with neighboring Sertoli cells and with elongating spermatids. The junctions are found where ectoplasmic specializations are known to occur in mammalian Sertoli cells, but differ structurally from their mammalian counterparts. At the most basic level, the organization of the chondrichthyan testis differs considerably from that of the mammalian condition. The bulk of the mammalian testis consists of seminiferous tubules. Spermatogenesis occurs within the seminiferous epithelium that lines the tubules. By contrast, the chondichthyan testis is made up of a mass of spermatocysts and is thus defined as polyspermatocystic (Grier, 1 992) . Spermatocysts are lined by many Sertoli cells, each of which surrounds and houses a clone of spermatogenic cells. The chondrichthyan Sertoli cell/spermatogenic cell unit is identified as a spermatoblast. Spermatogenesis occurs synchronously in all spermatoblasts of a single spermatocyst (Pudney, 1993, 1995). Despite the significant differences in testicular organization between these two classes, parallels do exist in terms of the physical relationships that develop between Sertoli cells and spermatogenic cells at certain stages of spermatogenesis. In both classes, elongating spermatid heads become positioned deeply within apical recesses or crypts of Sertoli cells during spermiogenesis. In chondrichthyan species this occurs within the spermatoblast while in mammals this takes place within the seminiferous epithelium. Adjacent Sertoli cells in both classes also share the common feature of being linked together basally by tight junctions. In both classes this creates a permeability barrier "low down" within the germinal epithelium. Here I describe a further structural similarity between Sertoli cells of these two classes: the presence of 69 microfilament-related junctions around the apical Sertoli cell crypts and around the basally located tight junctions. In mammalian Sertoli cells, a specialized form of actin filament-associated junction, the ectoplasmic specialization, is consistently seen around the apical crypts which contain elongating spermatids. These junctions also occur basally at the level of the inter-Sertoli cell junctional complexes (Nicander, 1967; Dym and Fawcett, 1 970; Flickinger and Fawcett, 1970; Russell, 1977a). Mammalian ectoplasmic specializations are unusual amongst actin-related junctions in that they are non-contractile (Vogl and Soucy, 1985). They possess highly ordered actin filaments which are tightly cross-linked into non-contractile bundles. These bundles are further linked to the plasma membrane and also to underlying cisternae of endoplasmic reticulum. Functionally they are thought to be primarily involved with intercellular adhesion (Grove and Vogl, 1989; Vogl, 1989; Vogl et al., 1991a,b, 1993). My results, along with those of a previous study in the ratfish (Stanley and Lambert, 1985), indicate that a modified form of ectoplasmic specialization is also present in Sertoli cells of chondrichthyan species. In the dogfish and ratfish these junctions occur at the similar apical and basal locations as ectoplasmic specializations do in mammals. As with mammalian ectoplasmic specializations, these chondrichthyan junctions consist of a layer of microfilaments. Labelling with rhodamine phalloidin at both the apical and basal sites confirms that the filament type present is actin. These chondrichthyan junctions, however, exhibit several structural differences from mammalian ectoplasmic specializations. In a number of ways they more closely resemble typical actin-related adhesion junctions seen in other cell types. For example, the actin filaments at both the apical and basal sites appear to be loosely organized, i.e. they lack the highly ordered appearance of filaments seen at mammalian ectoplasmic specializations. Furthermore, positive labelling of these sites apically in the dogfish (this chapter) and in the ratfish (Stanley and Lambert, 1985) with probes for myosin 7 0 II suggest that these sites are contractile. Contractility is a common feature of actin associated adhesion junctions in other cell types in general but one that is absent at mammalian ectoplasmic specializations (Vogl and Soucy, 1985). Although I was unable to detect myosin staining at either the apical or basal sites in the ratfish, this may have been due to a lack of inter-species reactivity between the antibody I used (raised against mammalian myosin II) and the antigen (ratfish myosin II). This interpretation is supported by the fact that Stanley and Lambert (1985) were able to detect myosin II at these sites apically in the ratfish with a different antibody. I was also unable to detect myosin labelling basally in the dogfish but this may have been a result of it being present in amounts below the level of detection by the methods I used. Consistent with this interpretation is the fact that the basal actin, with which the myosin would presumably associate, is only barely detectable by fluorescence microscopy in this species. In mammals one of the notable features of the ectoplasmic specialization is that i t has cisternae of endoplasmic reticulum firmly bound to the filament layer. This is true of both the apical and basal sites. My results indicate that in chondrichthyan species an association of the endoplasmic reticulum with the junction sites is variable. In neither the dogfish nor ratfish does the endoplasmic reticulum appear to be directly linked to the apical microfilaments as is seen in mammals; however, in both species elements of the endoplasmic reticulum are often found near the apical microfilament junctions. In the dogfish endoplasmic reticulum occurs near many of the filament sheaths around the spermatid heads. In the ratfish the endoplasmic reticulum occurs around the perimeter of the entire group of filament sheaths related to a bundle of spermatid heads. Basally, a consistent association between the endoplasmic reticulum and the filament layer at the junction sites is seen in the dogfish. Moreover, in this species the endoplasmic reticulum appears to be directly linked to the filament layer at these basal sites. This forms a structural unit that closely resembles the mammalian ectoplasmic 7 1 specialization. In the ratfish, a consistent association of endoplasmic reticulum with the junction sites is not seen. My observations along with those of Stanley and Lambert (1985) indicate that these chondrichthyan junctions are the structural homologues of the mammalian Sertoli cell/spermatid ectoplasmic specializations. They are present where ectoplasmic specializations occur in mammals and, like the latter, consist of filamentous actin. As with mammalian ectoplasmic specializations, it is likely that these chondrichthyan junctions are also involved with intercellular attachment. Evidence supporting this conclusion includes the results of experiments utilizing squash preparations of ratfish testis (Stanley and Lambert, 1985). Gentle squash preparations causes the fragmentation of spermatocysts. Elongating spermatids break free from Sertoli cells during such treatment, taking with them the Sertoli cell junctions and the underlying actin filaments. Such observations indicate firm attachment between Sertoli cells and spermatids at these junctional sites. In high power electron micrographs (this chapter; Stanley and Lambert, 1985), intercellular densities can be seen spanning the space between the Sertoli cell and spermatid plasma membranes at these junction sites. These have been interpreted as part of the adhesive elements at the apical junction sites (Stanley and Lambert, 1985). Evidence that the basal sites are involved with intercellular adhesion includes the observation that extraction of the Sertoli cell cytoplasm as an artefact of tissue preparation does not disrupt the association the microfilaments and underlying endoplasmic reticulum have with the plasma membrane. While this observation does not directly demonstrate that the basal sites are involved with intercellular adhesion, it does indicate that these two elements are firmly bound to one another and to the plasma membrane as one would predict in an actin-related adhesion junction. In both the dogfish and the ratfish I observed several sites of focal membrane contact between the plasma membranes of adjacent Sertoli cells. Similar sites have been reported by others in the dogfish (see Pudney, 1993) and presumably represent tight junctions. As with the tight junctions of mammalian Sertoli cells, these sites are found low down within the epithelium close to the basal lamina. What is of interest is that these sites occur within the zone in which the basal microfilament junctions are formed between neighboring dogfish and ratfish Sertoli cells. This suggests the existence of a form of junctional complex basally within this class, with an adhesion component and a permeability barrier. Unlike at the classical junctional complexes of other epithelial cell types, I did not note desmosomes to be part of this complex. While these chondrichthyan junctions share several features in common with the typical actin-related adhesion junctions found in other cell types as well as with ectoplasmic specializations of mammalian Sertoli cells, they also exhibit some unusual features of their own. At actin-related intercellular adhesion junctions in general, actin filaments tend to be restricted to the area immediately surrounding the contact site. This arrangement holds true at mammalian ectoplasmic specializations as well. In both dogfish and ratfish Sertoli cells, however, the actin filament zone is greatly expanded from each of the apical Sertoli cell crypts. In these species the filaments are oriented parallel to the long axis of each spermatid head and they extend away from the crypts and down toward the basal surface of the Sertoli cell. In the ratfish they have been described as being anchored to the base of the Sertoli cell (Stanley and Lambert, 1985). Although I was unable to determine if the filaments are in fact attached in some manner to the base of the cell in either species, my observations confirm that the filaments do course down to the vicinity the basal plasma membrane of the Sertoli cell. The contractile nature of these filaments, coupled with an association with the base of the Sertoli cell, suggest that these filaments may be involved with functions in addition to intercellular attachment. Stanley and Lambert (1985) have proposed that, through contraction, these filaments may play a role in positioning elongating spermatids within the spermatoblast. During spermiogenesis in this class, spermatids situated within Sertoli cell crypts become 7 3 oriented toward the base of the Sertoli cell and packed into a tight bundle. My observations in the dogfish and the ratfish are consistent with those of Stanley and Lambert (1985) in the ratfish and suggest that these junctional filaments are in a position to assume a role in spermatid orientation. If so, this differs dramatically from the mammalian condition, where the actin filaments of ectoplasmic specializations are non-contractile and are thought to be primarily involved with stablilizing the adhesion site. A direct role of these junctional filaments in spermatid orientation may, in fact, be peculiar to this class. In no other class has a link between the actin filaments around the apical crypts and the base of the cell been proposed. In summary, my results indicate that actin-related adhesion junctions are found in chondrichthyan Sertoli cells at sites where ectoplasmic specializations are known to occur in mammalian Sertoli cells. In certain key respects, such as the filament organization and an apparent capacity for contraction, the chondrichthyan junctions more closely resemble the typical actin-related adhesion junctions of other cell types than they do ectoplasmic specializations of mammals. As seen in this study, the presence of these junctions in two different chondrichthyan species, one representing the Elasmobrachii subclass (the dogfish) and the other representing the Holocephali subclass (the ratfish), suggest that these Sertoli cell junctions may be a common feature shared by species of this class. 74 FIGURE 3.1. Electron micrographs of dogfish spermatocysts showing filament-related attachment sites between Sertoli cells and elongating spermatids. At low magnification, as seen in panel (a), the heads of elongating spermatids (asterisks) embedded within the apical cytoplasm of a Sertoli cell can easily be seen. Note that dense concentrations of Sertoli cell microfilaments (arrowheads) are closely related to each spermatid head and that these filaments extend away from the heads toward the basal surface of the Sertoli cell. Panel (b) shows a higher magnification of these microfilaments around two spermatid heads, one of which is indicated by an asterisk. As is seen around the spermatid head on the right, the microfilaments tend to be oriented parallel to the long axis of the spermatid head (arrows) and extend for a considerable distance away from the spermatid head (arrowheads). In cross section, panel (c), the microfilaments are seen to form a sheath (arrowheads) around each spermatid head (asterisks). In this micrograph the plane of section has passed below the spermatid heads on the left such that only the filamentous sheaths are apparent. Panel (d) is a similar cross sectional view seen at higher power. In this micrograph a single spermatid head is surrounded by the dense concentration of Sertoli cell microfilaments. Of particular interest is the attachment site between the Sertoli cell and spermatid plasma membranes. A series of well-developed and highly ordered linkages, some of which are indicated by arrows, spans the intracellular space between the acrosome and plasma membrane of the spermatid at these sites. These linkages are limited to one side of the spermatid head and the underlying Sertoli cell microfilaments appear to be closely applied to this attachment site. Panel (a), X16,250. Bar = 1 um. Panel (b), X21.215. Bar = 0.5 |xm. Panel (c), X28.780. Bar = 0.5 um. Panel (d), X62.530. Bar = 0.25 urn. 75 76 FIGURE 3.2. Electron micrographs of dogfish spermatocysts. Panel (a) shows a cross section Sertoli cell microfilaments (arrowheads) around an elongating spermatid head. Note that elements of the endoplasmic reticulum (asterisks) course nearby the microfilaments but that no direct association between the two is visible. Cisternae of endoplasmic reticulum also occur at the level of the intercellular junctions formed between neighboring Sertoli cells (panel b). At these sites the endoplasmic reticulum (arrowheads) does appear to be closely linked to the junctions. These junctions are formed basally between Sertoli cells (panel b) and, as seen in panel (c), consist of a filamentous layer (curved arrows) interposed between the plasma membrane and an underlying cistern of endoplasmic reticulum (asterisks). At high magnification (panel d), focal sites of membrane contact (arrowheads) are occationally seen in the regions of these filament-related Sertoli cell junctions. These sites likely represent tight junctions. Elements of the endoplasmic reticulum are indicated by asterisks. Note in panel (e) that when the Sertoli cell cytoplasm has been largely extracted as an artefact of tissue preparation, the filamentous layer (curved arrows) and underlying endoplasmic reticulum (asterisks) remain associated with the Sertoli cell plasma membrane at these basal junction sites. This suggests that the two are firmly bound to the membrane at these sites. Panel (a), X51.590. Bar = 0.25 um. Panel (b), X30.030. Bar = 0.5 um. Panel (c), X90.600. Bar = 0.2 um. Panel (d), X218,950. Bar = 0.1 um. Panel (e), X45.740. Bar = 0.25 urn. 77 78 FIGURE 3.3. Paired fluorescence and phase micrographs of fixed frozen sections of dogfish spermatocysts showing the distribution of filamentous actin (panel a,a') along with the controls (panels b,b'-d,d') which were run for actin staining. Strong staining with rhodamine phalloidin (panel a,a') is seen around the tightly packed anterior ends of spermatid heads (arrows), sites where Sertoli cell microfilaments are seen at the ultrastructural level. When similar sections are incubated with rhodamine phalloidin in the presence of phallacidin (panel b,b'), labelling of these sites is greatly reduced. No staining is seen when sections are incubated with phallacidin alone (panel c,c') of with buffer alone (panel d,d'). X628. Bar = 20 um. 77 80 FIGURE 3.4. Paired fluorescence and phase micrographs of a fixed frozen section of a dogfish spermatocyst stained with rhodamine phalloidin. In this grazing section, weak labelling for filamentous actin is detected basally around the perimeter of Sertoli cells (arrowheads)(panel a). Several clusters of spermatids embedded within the Sertoli cells are visible by. phase (panel a'). The plane of section has fortuitously passed through the anterior ends of some of these spermatid clusters, around which strong labelling for actin is seen (arrows). X628. Bar = 20 um. SI 82 FIGURE 3.5. Paired fluorescence and phase micrographs of fixed frozen sections of dogfish spermatocysts showing the distribution of myosin II (panel a,a') and the controls (panels b,b'-d,d') which were run for myosin II labelling. When stained with a myosin II antiserum, intense specific labelling occurs around the anterior ends of spermatid heads (arrows)(panel a,a'). The location and pattern of this labelling matches that seen with rhodamine phalloidin in similar sections and corresponds to the location where Sertoli cell microfilaments are seen at the ultrastructural level. A diffuse staining with the antiserum is also present throughout the basal Sertoli cell cytoplasm; however, this is likely non-specific as a similar pattern is also seen when sections are incubated with preimmune serum (panel b,b'). No specific staining is present in sections incubated with preimmune serum (panel b,b'), nor in sections in which primary (panel c,c') or both primary and secondary antibodies have been replaced by buffer alone (panel d,d'). X628. Bar = 20 um. 84 FIGURE 3.6. Electron micrographs of ratfish spermatocysts showing filament-related attachment sites between Sertoli cells and elongating spermatids. Panel (a) shows elongating spermatids (asterisks) embedded within the apical cytoplasm of a Sertoli cell. Note the Sertoli cell microfilaments (arrowheads) closely related to each spermatid head and extending deeper into the apical Sertoli cell cytoplasm. Panel (b) is a similar view of these Sertoli cell microfilaments (arrowheads) shown at higher magnification. The filaments are aligned parallel to the long axis of the spermatid head. In panel (c) a cross section through the anterior ends of several spermatid heads (asterisks) is seen. Each is surrounded by a sheath of Sertoli cell microfilaments (arrowheads) and this sheath appears to be concentrated to one side of the spermatid head. In late stages of spermiogenesis, elongating spermatids (asterisks in panel d) are tightly bundled together with their heads near the basal surface of the Sertoli cell. The Sertoli cell cytoplasm into which the spermatid bundle projects is largely devoid of organelles (panel d). Note cisternae of endoplasmic reticulum occur near the. perimeter of the spermatid bundle (arrowheads). Panel (a), X23,060. Bar = 1.0 um. Panel (b ) , X42,340. Bar = 0.25 um. Panel (c), X39,750. Bar = 0.25. Panel (d), X15,530. Bar = 1.0 um. 86 FIGURE 3.7. Electron micrographs of basal microfilament-associated junctions formed between adjacent Sertoli cells in the ratfish. In panel (a) a zone of loosely organized microfilaments (arrowheads) closely follows the basolateral plasma membranes of Sertoli cells where the cells are in contact with each other. In panel (b) similar clumps of microfilaments (asterisks) are seen along the basolateral Sertoli cell plasma membranes. Note tight junctions (arrowheads) between neighboring Sertoli cells are occasionally seen in the vicinity of the microfilaments. The inset shows the tight junctions (arrowheads) of panel (b) at higher magnification. Panel (a), X131,000. Bar = 0.1 um. Panel (b), X94.550. Bar = 0.25 urn. Inset, X168.890. Bar = 0.1 um. 8 8 FIGURE 3.8. Paired fluorescence and phase micrographs of fixed frozen sections of ratfish spermatocysts showing the distribution of filamentous actin. Intense labelling with rhodamine phalloidin occurs around the anterior ends of elongating spermatid heads (arrows) which occur in tightly clumped groups. This labelling corresponds to location where Sertoli cell microfilaments are seen at the ultrastructural level. Also visible is a more diffuse staining basally within Sertoli cells (arrowheads) (panel a). This likely represents the actin related to the intercellular junctions formed basally between neighboring Sertoli cells. In more grazing sections through the spermatocyst wall (panel b,b'), the full extent of this basal actin is appreciated. Labelling is apparent around the margins Sertoli cells (arrowheads). Labelling of Sertoli cell actin filaments related to elongating spermatids (arrows) is also visible. X628. Bar = 20 um. 9 0 CHAPTER 4 Actin-Related Intercellular Adhesion Junctions in the Germinal Compartment of Testis in the Bowfin (Amia calva) and Guppy (Poecilia reticulata) 9 1 I n t r o d u c t i o n The germinal compartment of the vertebrate testis consists of spermatogenic cells and Sertoli cells (Grier 1993). Often, there occur intercellular junctions between and amongst these two cell populations. Although these junctions are well characterized in mammals (Byers et al. 1993; Vogl et al. 1993), little is known about the mechanisms of attachment between cells in the germinal compartment in non-mammalian vertebrates. In this chapter I examine the role intercellular junctions play in the germinal compartment in the class Osteichthyes, the bony fish. In the class Osteichthyes, as in other anamniote classes, the structural unit in which the germinal compartment is found is termed the spermatocyst (Grier, 1 993; Pudney, 1993). Sertoli cells form the walls of spermatocysts and surround the more centrally located population of spermatogenic cells. Adjacent Sertoli cells abut one another and are reported as squamous to columnar in shape, depending on the species and stage of spermatogenesis (see Pudney, 1993). Previous studies have demonstrated the presence of Sertoli cell tight junctions in several species of bony fish (Abraham et al. , 1980; Marcaillou and Szollosi, 1980; Bergmann et al., 1984; Parmentier et al. , 1985). Occasional small desmosomes have also been noted between adjacent Sertoli cells in some species (Abraham et al., 1984; Bergman et al., 1985; Arenas et al., 1995), as have complex membrane interdigitations along the lateral borders of Sertoli cells (see Pudney, 1993). Intercellular adhesion functions have been attributed to all of these features (Pudney, 1993). Well-developed attachment plaques or organized junctional complexes between Sertoli cells, however, have not yet been described. Moreover, we l l -defined junctions, such as ectoplasmic specializations, have not been reported between Sertoli cells and spermatids in any bony fish species. To further investigate the extent of intercellular adhesion junctions and the role they assume in the bony fish germinal epithelium, I here examine junctions between neighboring Sertoli cells and between 92 Sertoli cells and spermatids in two species of bony fish, the bowfin Amia calva and the guppy Poecelia reticulata. Selecting representative bony fish species for a comparative study requires some care. With more than 20,000 different species currently separated into 35 recognized orders, the class Osteichthyes is an immense and diverse vertebrate group. Many present-day bony fish species in fact make poor representatives of their class i n comparative studies in that they are highly specialized, that is, over time they have evolved considerably away from the ancestral or "primitive" form of bony fish and now, in certain characteristics, no longer represent the "prototypical" bony fish. The bowfin is thought to have appeared early on the phylogenetic scale but evolved little since that time (Patterson, 1982). It therefore serves as a good "primitive" representative of this class. For this study on the germinal epithelium, another particularly important feature to consider when selecting representative species is the type of structural relationship that forms between Sertoli cells and the developing germ cells. Within this class, two different patterns of Sertoli cell/germ cell association are seen. In most species of bony fish, germ cells retain a round head shape as they develop and form at most only a minor structural relationship with Sertoli cells. In general these species tend to be external fertilizers (Jamieson, 1991). The bowfin serves as an example of this group of bony fish. Contrasting this group is a second, relatively small number of bony fish species belonging to families in the order Cyprinodontiformes. In these species, a close association between the developing germ cells and Sertoli cells is established at certain stages of spermatogenesis, that is, during the process of spermiogenesis. The germ cells of these species tend to acquire a more elongate head shape as they mature. Interestingly, these species also tend to be internal fertilizers. Previous studies of intercellular junctions between Sertoli cells and germ cells in bony 93 fish (Sprando and Russell, 1987a,b) have not focussed on species from this second group. The guppy is a representative of this group. 94 Materials and Methods A total of fourteen mature bowfin {Amia calva) and five mature, wild type guppies (Poecilia reticulata) were used in this study. Bowfin were caught in Lake Ontario, then shipped to the University of British Columbia where they were kept in large holding tanks in the Department of Zoology until the time of use. Guppies were kindly provided by Dr. N. R. Lily from a colony maintained in the Department of Zoology at the University of British Columbia. Individuals composing the parent stock of this colony were originally obtained from Trinidad. Electron Microscopy: For electron microscopy, 12 bowfin and three guppies were anesthetized with MS222 (tricaine: 1g/l) buffered with sodium bicarbonate (2g/l), and the testes dissected out in fixative (1.5% paraformaldehyde, 1.5% glutaraldehyde, 0.1 M sodium cacodylate buffer, pH 7.3). Testes were immersion fixed for 2 hrs and then processed for electron microscopy as previously described (see Chapter 2). Thick sections were stained with 1% toluidene blue and photographed on a Zeiss Axiophot microscope. Thin sections were stained with uranyl acetate and lead citrate and then photographed on a Philips EM 300 operated at 80 KV. Fluorescence Microscopy: (a) Tissue Preparation: Testes to be used for fluorescence microscopy were removed in fixative (3.0% paraformaldehyde in phosphate buffered saline (PBS) - (150 mM NaCI, 5 mM KCI, 3,2 95 mM Na2HP04, 0.8 mM KH2PO4, pH 7.3) from 10 bowfin and two guppies that had been anesthetized. After immersion fixation for 30 min, the testes were washed in buffer, embedded directly in OCT compound (bowfin) or placed in a pocket of fixed rat liver and then embedded in OCT compound (guppy), and then frozen in liquid nitrogen. Frozen sections of 8um were cut with a cryostat and attached to polylysine coated slides. The slides were immediately placed into cold ( -20°C) acetone for 5 min and then air dried for approximately 30 min. (b) Localization of Actin and Myosin II: Localization of filamentous actin and myosin II by immunofluorescence was accomplished using rhodamine phalloidin and a rabbit antiserum raised against human platelet myosin II respectively. The labelling protocols followed for both these probes have been previously described (see Chapter 2). Controls for rhodamine phalloidin included (1) PBS (containing 0.1% BSA) + 1.65 x 10-6 M rhodamine phalloidin + 7.76 x 10-4 M phallacidin (competitive specificity control); (2) PBS (containing 0.1% BSA) + 7.76 x 10-4 M phallacidin (control for phallacidin in reagent 1); ( 3 ) PBS (containing 0.1% BSA) (control for autofluorescence). Controls for myosin II staining included: (1) substitution of the immune serum with preimmune serum; (2 ) substitution of the primary antiserum with buffer; (3) substitution both of the primary and of the secondary antibodies with buffer. The primary myosin II antiserum and the matched preimmune serum were kindly provided by Dr. Kegi Fujiwara (National Cardiovascular Center Research Institute, Osaka, Japan). All sections processed for fluorescence were examined on a Zeiss Axiophot photomicroscope fitted with filter sets for detecting rhodamine and fluorescein isothiocyanate. 96 R e s u l t s Bowfin (Amia calva) Spatial Relationships Within the Germinal Epithelium: In the bowfin, the spermatogenic testis consists of a series of anastomosing tubules (Grier, 1993) which are lined by spermatocysts. Each spermatocyst is formed by a single layer of Sertoli cells which extend long, thin cytoplasmic processes out into the lumen of the tubule. These processes, in turn, join up with those of neighboring Sertoli cells to complete the walls of cyst (Fig. 4.1a). Contained within each cyst is a cohort of developing spermatogenic cells (Fig. 4.1a,b). In this species, spermatogenic cells possess a round head shape throughout development and retain this head shape even as mature spermatozoa. Spermiation is accomplished by the breakdown of the spermatocyst wall and the release of spermatozoa into the lumen of the tubule (germinal compartment) (Fig. 4.1b). Following this, Sertoli cells are incorporated into the walls of the tubule (Fig. 4.2a). Spermatocysts are no longer present at this stage. Eventually, as the testis enters regression, all spermatozoa are shed from the tubules and the epithelium made up of Sertoli cells degenerates (Fig. 4.2b). The Spermatogenically Active Testis: The spatial relationships between cells of the active spermatocyst make i t possible for intercellular attachment junctions to occur at up to three separate locations: 1) between Sertoli cells and spermatogenic cells; 2) along the lateral borders of adjacent Sertoli cells; and 3) between the long, thin, cytoplasmic processes that extend apically from one Sertoli cell to another. 97 (a) Junctions Between Sertoli Cells and Spermatogenic Cells: Within the spermatogenic testis, cohorts of developing spermatogenic cells f i l l each spermatocyst. Despite the close proximity many of the spermatogenic cells have to the Sertoli cell layer, intimate contact between the two cell types, i.e. entrenchment of the spermatid stage within Sertoli cells, is not seen. Contact between Sertoli cells and spermatogenic cells is limited to very small, focal attachment sites. These sites are formed between Sertoli cells and spermatogenic cells which are in the early stages of differentiation, i.e. spermatocyte stage. Such sites are evident only at the ultrastructural level (Fig. 4.3a,b) and appear as localized densities along the plasma membrane at sites of contact between the two cell types. An indistinct filamentous layer is related to these attachment sites. The indistinct composition of the filamentous layer suggests that the junctions may be associated with microfilaments rather than of filaments belonging to the intermediate class. The attachment sites appear to be few i n number and, in this study, I was unable to determine if Sertoli cells construct these sites with all spermatocytes within a single cohort, or with only those located at periphery of each cyst. No junctional contacts of any kind are formed between Sertoli cells and the spermatid stage of development. (b) Junctions Between the Lateral Borders of Adjacent Sertoli Cells: Within the active spermatocyst, Sertoli cells are squamous to cuboidal shaped cells that extend thin, apical cytoplasmic processes out to meet those of neighboring Sertoli cells. The lateral borders of Sertoli cells, the regions along which adjacent Sertoli cells contact each other, follow convoluted courses. Concentrated at the luminal end of these borders are sites of intercellular contact that are related to underlying elements of the cytoskeleton (Figs. 4.4). 98 Ultrastructurally, two distinct forms of filament-associated junctions are seen. The first form is more luminal in position than the second and is characterized by an electron dense band that follows the apical zone of contact between two neighboring cells (Fig. 4.4). The second and most conspicuous form of filament-associated junction is a desmosome. The most luminal of the two junction types appears to be associated with microfilaments (Fig. 4.4). In high power electron micrographs, the filaments related to this junction form a carpet adjacent to the plasma membrane and are clearly of smaller diameter than the intermediate filaments associated with nearby desmosomes. At these sites, adjacent plasma membranes are closely apposed. Desmosomes in the bowfin have a classical appearance (Fig. 4.4). Adjacent plasma membranes are separated by an intermediate dense line and submembrane plaques are associated with bundles of intermediate filaments. Often these bundles appear to interconnect adjacent desmosomes. In the bowfin, these two Sertoli cell junction types are arranged together in what resembles a classical junctional complex. Although I could not conclusively identify gap or tight junctions in the material I examined, I did occasionally note focal sites of contact between Sertoli cells. These sites tended to be located at the apex of the junctional complex and may be tight junctions. In addition to the apically located junctional complex, other Sertoli cell microfilament-related junctions are occasionally seen deeper within the epithelium and unassociated with the apical junction complex. These junctions are apparent as small (spot) attachment sites along the lateral Sertoli cell plasma membrane. 99 (c) Junctions Between Sertoli Cell Apical Processes: Within the spermatogenic testis of the bowfin, spermatocyst walls are formed by long, thin cytoplasmic processes that extend apically from neighboring Sertoli cells. The joining of these processes leads to the completion of each cyst (Fig. 4.4). At the sites of union between these cytoplasmic processes, small filament-related junctions are seen (Fig. 4.4). These junctions are visible only at the ultrastructural level and, by appearance, closely resemble the microfilament-related junctions that are present at the level of the Sertoli cell junctional complex (see above). The similarity with the latter junction type holds true in terms of the indistinct composition and organization of the filaments at these junction sites. The absence of an intermediate dense line and submembrane plaques which characterize desmosomes further suggest that these small junctions are not intermediate filament-related but rather are microfilament-related. The Post-Spermiation Testis: Following the release of spermatozoa from the cysts, Sertoli cells become incorporated into the tubule walls. The testis now appears as a series of epithelial lined chambers (Fig. 4.5a,b). Individual tubules are lined by a single layer of Sertoli cells (Fig. 4.5b). At the ultrastructural level, the Sertoli cell layer is seen to be a simple squamous epithelium separated from connective tissue by a basal lamina (Fig. 4.5c). The thin, apical cytoplasmic processes seen in Sertoli cells of the active spermatocyst are now absent. Along the luminal surface, the occasional microvillus is present. The lateral borders of Sertoli cells now take particularly convoluted courses. This results in a considerable overlapping and redundancy of the lateral plasma membrane of each cell (Fig. 4.5c). The material separating these epithelial lined chambers consists of a matrix rich tissue containing sparse cells and small blood vessels. Spherical shaped 100 spermatozoa are densely concentrated in the luminal spaces of the ducts. No attachment or physical association of any kind is seen between Sertoli cells and spermatozoa. Junctional complexes are still formed between adjacent Sertoli cells. As in Sertoli cells of the active spermatocyst, these complexes are positioned apically between neighboring Sertoli cells and consist of two different forms of filament-associated junctions: 1) a more apical layer of microfilament related junctions and 2) an underlying layer of desmosomes (Figs. 4.6a,b, 4.7). Ultrastructurally, the appearance of these junctions individually and their arrangement together as a junctional complex is indistinguishable from those formed between adjacent Sertoli cells of active spermatocysts. Often, a microvillus is located very near these junctional complexes. Small focal or "spot" junctions related to microfilaments are still occasionally seen deeper within the epithelium between Sertoli cells (Fig. 4.6a.). In high power electron micrographs, focal sites of contact are seen between Sertoli cells near the apex of the junctional complex (Fig. 4.7). These may represent tight junctions. Fluorescence Localization of Actin: Due to difficulties in obtaining bowfin testes with active spermatocysts, the localization of actin was conducted only in the post-spermiation testis, i.e. tissue in which the inter-Sertoli cell junctional complexes are still present but in which the apical Sertoli cell processes are no longer seen. In fixed frozen sections of such tissue stained with rhodamine phalloidin, fluorescence occurs in two major locations: ( 1 ) Around the periphery of the tubules, presumably in cells that are the equivalent of the actin-rich peritubular cells characteristically seen in higher vertebrates; (2) Within the epithelium. Within the epithelium, the fluorescence occurs as linear bands, which I interpret as being at the margin between neighboring Sertoli cells. The fluorescence 101 pattern of these bands is distinct from the pattern emitted by the more peripheral wall region. In cross sections of tubules, the linear bands often appear as isolated, short, vertical tracts near the luminal surface of the epithelium (Fig. 4.8a), while in more grazing sections the bands appear to be continuous with one another and form a honeycomb pattern (Fig. 4.8b). No fluorescence was observed in any of the control slides for the rhodamine phalloidin staining (data not shown). Fluorescence Localization of Myosin II: As with the detection of actin, localization of myosin II was conducted only in post-spermiation tissue. In fixed frozen sections labelled with a probe for myosin II, the overall pattern of fluorescence seen is similar to that obtained with the probe for actin. Specific and intense fluorescence is present around the periphery of each tubule. Less intense fluorescence, appearing as short bands, is also present within the epithelium (Fig. 4.8c,d). The location and pattern of these bands are similar to that of the linear staining seen with the actin probe within the epithelium. No fluorescence was observed when preimmune serum was substituted for the primary antiserum, when the primary antiserum was omitted, nor when both the primary and secondary antibodies were omitted (data not shown). 102 G u p p y {Poecilia reticulata) Spatial Relationships Within the Germinal Epithelium: In the guppy testis, defined as lobular (Grier, 1993), active spermatocysts are lined by a simple squamous to cuboidal epithelium made up of Sertoli cells. Spermatogenic cells occur centrally within each cyst lumen. As spermiogenesis proceeds, the head shape of spermatids changes from round to an elongate form. With further development is the appearance of a specialized intercellular association that forms between Sertoli cells and the elongating spermatids. Elongating spermatids become aligned with their heads embedded or entrenched within the apical cytoplasm of the Sertoli cells (Fig. 4.9a). Spermiation involves the release of non-encapsulated clusters of spermatozoa, known as spermatozeugmata, into efferent ducts. (a) Junctions Between Sertoli Cells and Spermatogenic Cells: At a specific stage of spermiogenesis, elongating spermatids become embedded within apical crypts of Sertoli cells (Fig. 4.9a). At this stage, an apical zone of weak and homogeneous staining is visible in regions of Sertoli cells that surround spermatid heads (Fig. 4.9a). When viewed with the electron microscope, this zone is composed of clumps of filamentous material, each of which is related to a recess containing a spermatid head (Fig. 4.9b). At higher magnification, linkages (Fig. 4.10a,b) are evident between the plasma membrane of the Sertoli cell and that of the spermatid head. Interestingly, microfilaments are not restricted to crypt regions, but extend into long "microvil l i" that insinuate themselves between proximal regions of spermatid tails (Fig. 4.10c). 103 (b) Junctions Between the Lateral Borders of Adjacent Sertoli Cells: Junctional complexes occur between adjacent Sertoli cells. These complexes occur at the apex of the epithelium and consist both of microfilarnent and of intermediate filament related components (Fig. 4.11a-c). Microfilament-related components (Fig. 4.11a,b) occur apically while junctions related to intermediate filaments (Fig. 4.11 a,c) occur in a more basal position. Intermediate filament-related components are t obviously identifiable as desmosomes (Fig. 4.11a,c). The nature of the microfilament-related junctions is unclear; however, they occur in positions where tight, adhesion (intermediate) and gap junctions occur in other epithelia (Farquhar and Palade, 1963). Fluorescence Localization of Actin and Myosin II: When stained with the filamentous actin probe rhodamine phalloidin (Fig. 4.12a,a'), Sertoli cell regions associated with spermatids stain intensely. This staining occurs in regions around spermatid heads and also in elongate (thread-like) structures related to the proximal parts of spermatid tails (Fig. 4.12a,a'). Labelling with the immunological probe for myosin II (Fig. 4.12b,b') is concentrated in areas associated with apical recesses. Regions of the epithelium corresponding to junctions between neighboring Sertoli cells also label with probes for filamentous actin and myosin II. This is particularly obvious in tangential sections through the cyst wall (Fig. 4.13a,a'-b,b'). In sections such as these, labelled apical junctional complexes of numerous cells are viewed en face and together form a hexagonal or honeycomb pattern (Fig. 4.13a,a'-b,b'). Specific staining was not observed in any of the controls for rhodamine phalloidin staining nor for anti-myosin II labelling (data not shown). 104 D i s c u s s i o n In the bowfin and the guppy, as in other bony fish (see Grier 1975, 1 993; Pudney, 1993), spermatogenesis occurs in spermatocysts which are lined by a simple squamous to cuboidal epithelium formed of Sertoli cells. Spermatogenic cells differentiate within the lumena of the cysts and are ultimately released from the cysts as spermatozoa. Structurally the testes of the bowfin and guppy represent the two different patterns of testicular organization that are present within the class Osteichthyes. In the unrestricted spermatogonia! type, which is by far the most common and which is present in the bowfin, spermatogonia are found along the entire length of the germinal compartment (tubules or lobules) (Grier, 1993; Pudney, 1993). Contrasting this is the second pattern of testicular organization: the restricted spermatogonia type. In this pattern, typified by the guppy, spermatogonia occur in cysts only at the distal ends of lobules (Grier 1993; Pudney 1993). As spermatocysts mature, they gradually translocate to the proximal ends of the lobules where clusters of spermatozoa are released into efferent ducts. This type of testicular organization is seen only in one order of bony fish, Cyprinodontiformes. Spatial relationships between Sertoli cells and spermatogenic cells in the bowfin and in the guppy are similar during the early stages of spermatogenesis. In both species, most spermatogenic cells appear not to be directly associated with the spermatocyst epithelium. Only the spermatogenic cells located at the periphery of spermatocysts lie near or perhaps in contact with Sertoli cells (Billard 1970; Grier 1975; Billard 1984). As spermatogenesis progresses, however, two very different patterns emerge. While in the bowfin Sertoli cell/germ cell relationships remain distant, in the guppy an association develops between all spermatogenic cells and the spermatocyst epithelium, leading eventually to a stage where elongating spermatids become situated in recesses 105 within the apices of Sertoli cells (Billard 1970; Grier 1975; see Billard 1986, Fig. 5). Apical "finger-like" extensions of Sertoli cells are thought to anchor the spermatids to the epithelium (Grier 1975; Pudney 1993). The presence of these "extensions", which are composed of an electron lucent material, produces a relatively organelle-free zone of apical Sertoli cell cytoplasm around the spermatid heads in the guppy (Grier 1975). Well-developed cytoskeletal-associated junctions between Sertoli cells and spermatids have not been previously described at these sites in the guppy (see Grier 1975, and Pudney 1993) or in any other bony fish species (Sprando and Russell, 1987a,b). In this study I present evidence that adhesion junctions occur between Sertoli cells and spermatids in the guppy, but not in the bowfin, and that these junctions are related to actin. In the guppy fibrous linkages can be seen between the plasma membranes of Sertoli cells and the entrenched, elongating spermatids heads at the attachment sites. Moreover, these attachment regions are clearly related to underlying concentrations of loosely arranged microfilaments in Sertoli cells. These filaments occupy much of the organelle free zone in the apical cytoplasm. This zone stains intensely with rhodamine phalloidin, indicating that the filaments are actin, and with probes for myosin II. Cisternae of endoplasmic reticulum are not associated with the actin. That previous studies designed to characterize attachment junctions between Sertoli cells and spermatids in bony fish (Sprando and Russell 1987a,b) did not detect well-developed junctions is most likely due to the fact that the species examined were ones in which a close physical relationship between Sertoli cells and spermatids is not formed. Species examined in these studies are similar to the bowfin in this respect. In these species, spermatogenic cells retain a round head shape throughout spermiogenesis. While this pattern may be the most common in this class, this study demonstrates that in species such as the guppy, where a close physical association between Sertoli cells and 106 spermatids does form, well-developed Sertoli cell/spermatid junctions are constructed. Interestingly, the species in which this close Sertoli cell/spermatid association is formed are also those in which spermatids acquire an elongate head shape. In the guppy, actin filament concentrations occur in a second location as well: in elongate processes (microvilli) that extend into the lumen amongst proximal regions of spermatid tails. These microvilli, however, do not appear to be attached to the spermatids and are therefore unlikely to play a major role in intercellular adhesion. Intercellular junctions also occur between lateral plasma membranes of adjacent Sertoli cells in the bowfin and guppy but, unlike the Sertoli cell/spermatid junctions, these Sertoli/Sertoli cell junctions appear to be very similar between the two species. Previous studies have reported membrane interdigitations, desmosomes and tight junctions between adjacent Sertoli cells in several different bony fish species, including those in the guppy (Billard 1970; Grier 1975; Bergmann et al. 1984). My results indicate that, in the bowfin and the guppy, junctions between neighboring Sertoli cells occur at the apex of the epithelium and are arranged in the form of a complex not unlike classically described junctional networks in other epithelia (Farquhar and Palade 1963). Although I could not confirm the presence of tight junctions in either species using standard transmission electron microscopy, others have demonstrated the presence of these junctions in the guppy indirectly with tracers (Marcaillou and Szollosi 1980; Bergmann et al. 1984). What I did observe in this study is a zone of microfilament-related junctions near the lumenal margin of the epithelium. That these loosely arranged filaments are actin is indicated by their size, relative to intermediate filaments located elsewhere in the cell, and the labelling of these junctional regions with rhodamine phalloidin. As in other epithelial systems in general (Hirokawa et al. 1983; Philp and Nachmias 1985; reviewed by Mooseker 1985), these regions also label with probes for myosin II. In typical junctional complexes, the actin-related zone contains tight (Madara 1987; Drenckhahn and Dermietzel 1988; reviewed by Madara 1992) and 107 adhesion junctions (Hirokawa etal . 1983; Boiler etal . 1985; Hirano et al. 1 987; Gumbinar 1992). Unlike Sertoli/Sertoli cell junctions in many other vertebrates, elements of the endoplasmic reticulum are not found as part of the junctional complex i n the bowfin nor in the guppy. Immediately basal to the actin-related junctions is a zone of intermediate filament-related desmosomes. A second site of inter-Sertoli cell junctions is seen in the bowfin: between the thin apical cytoplasmic extensions that join one Sertoli cell to another to complete each spermatocyst. Based on the similar appearance these small junctions have to the microfilament-related junctions present at the level of the Sertoli cell junctional complexes, I interpret these apical spot junctions as being microfilament-related as well. Given the location of these junctions in the thin apical walls of the spermatocyst, i t is possible that the dissolution of these junctions at the time of spermiation could be the event that leads to the opening of the spermatocyst and the release of spermatozoa. I conclude from my observations that well-developed actin-related adhesion junctions form between Sertoli cells and spermatids in the guppy but not in the bowfin. The actin associated with these junctions is loosely arranged, probably has contractile properties (contains myosin II), and is not related to an underlying cistern of endoplasmic reticulum. My observations also indicate that junctional complexes occur between neighboring Sertoli cells in the bowfin and the guppy. Between Sertoli cells, these complexes are situated at the apical margin of the epithelium. A ring of actin filaments, which is postulated to have contractile properties because it contains myosin II and is not related to elements of the endoplasmic reticulum, is associated with the junctions and is most likely related to adhesion and occluding components of the complexes. These results suggest that, within the class Osteichthyes, junctional complexes between Sertoli cells may be a widespread phenomonon but that w e l l -developed junctions between Sertoli cells and spermatogenic cells only develop in species where there is intimate contact between Sertoli cells and spermatids. These species 108 appear to be the ones in which spermatids acquire an elongate head shape. On the basis of location and cytoskeletal composition, I propose that the actin-related junction components I identify in the bowfin and the guppy are the structural and functional correlates of ectoplasmic specializations described in other vertebrates (Russell, 1977a; Sprando and Russell 1987a; see also Chapters 5 and 6). 109 FIGURE 4.1. Light micrographs of toluidene blue stained plastic sections of spermatocysts of the bowfin testis. Shown in panel (a) is a cross-section of a seminiferous tubule with a patent lumen (large asterisk) filled with mature spermatozoa. Several spermatocysts line the walls of the tubule. Each spermatocyst is formed by Sertoli cells which surround a cohort of developing and differentiating spermatogenic cells. Clusters of developing spermatids within two different spermatocysts are indicated (small asterisks). Spermatocyst walls are completed by thin cytoplasmic processes (small arrowheads) of adjacent Sertoli cells. Sertoli cell nuclei (large arrowheads) can be seen along the walls of the tubule. Shown in panel (b) are three spermatocysts at a later stage of spermatogenesis. Housed within are maturing spermatogenic cells about to be released as spermatozoa. The thin spermatocyst walls formed by Sertoli cell processes are clearly evident (small arrowheads). As is seen in the spermatocyst furthest on the left, spermiation occurs when the walls of the spermatocyst open (curved arrow) to release mature spermatozoa into the lumen of the tubule. Sertoli cell nuclei (large arrowheads) are visible. X628. Bar = 20 urn. 110 111 FIGURE 4.2. Light micrographs of toluidene blue stained plastic sections of the bowfin testis. Shown in panel (a) is a cross-section of a post-spermiation seminiferous tubule. Spermatozoa fill the lumen (asterisk) of the tubule. Sertoli cell nuclei are indicated (arrowheads). Note that spermatocysts are no longer present and that Sertoli cells are restricted to the tubule wall as a simple squamous epithelium. As the season progresses, all spermatozoa are eventually shed from the tubules and the testis regresses, as seen i n panel (b). At this point, a dramatic proliferation of connective tissue cells begins within the matrix regions that separate the tubules. Concomitant with this is a collapse of the tubules and a degeneration of the Sertoli cell epithelium that lines the tubules. The lumen of one tubule is indicated (asterisk). Note that the lumena are now devoid of all spermatozoa. Two candidates for Sertoli cells (arrowheads) are seen along the periphery of one tubule. X628. Bar = 20 um. 113 FIGURE 4.3. Electron micrographs of a bowfin spermatocyst showing attachment sites formed between Sertoli cells and spermatocytes. At low magnification, as seen in panel (a), a tightly packed cohort of spermatocytes is visible within a spermatocyst. Nuclei of three spermatocytes (SC) and of one Sertoli cell (SE) are labelled. Also indicated are the lumen of the seminiferous tubule (asterisk) and the junctional complex (curved arrow) formed between adjacent Sertoli cells (see Fig. 4.4). Very small attachment sites are visible between Sertoli cells and spermatocytes (arrowheads). Panel (b) shows a higher magnification of these attachment sites (arrowheads). Such sites appear as localized densities along the plasma membrane at points of contact between the two cell types. Panel (a) X8.100. Bar = 2 urn. Panel (b) X14,750. Bar = 1 um. 115 FIGURE 4.4. Junctions between apical Sertoli cell processes in' the bowfin. Shown here are portions of two squamous shaped Sertoli cells joined together by an apically positioned junctional complex. Microfilament related junctions (small arrows) can be seen above desmosomes (large arrow). Completion of spermatocyst walls in the bowfin is accomplished by thin apical processes that extend from one Sertoli cell to another. Small filament associated junctions occur at the sites of union of these processes (arrowheads). By appearance these filament related junctions closely resemble the microfilament junctions seen at the level of the Sertoli cell junctional complex. The inset shows, at higher magnification, the junctions between the apical Sertoli cell processes. The lumen of the spermatocyst is indicated (small asterisk) as is the lumen of the tubule (large asterisk). X14,030. Bar = 1 urn. Inset X28.210. Bar = 0.5 urn. 116 117 FIGURE 4.5. Panels (a) and (b) are light micrographs of toluidene blue stained plastic sections of the post-spermiation bowfin testis. As seen at low power in panel (a), the lumena of tubules are densely packed with spermatozoa (asterisks) following spermiation. At higher magnification, as shown in panel (b), a single layer of Sertoli cells (small arrowheads) is apparent around the wall of each tubule. Spermatocysts are no longer evident. Sertoli cell nuclei are occasionally seen around the tubule wall (large arrowheads). At the ultrastructural level, panel (c), the squamous shape of post-spermiation Sertoli cells is clearly evident. Note that the thin apical Sertoli cell processes which form the walls of the active spermatocyst are now absent. The lateral plasma membranes of Sertoli cells (arrowheads) now follow highly tortuous courses. Inter-Sertoli cell junctional complexes (curved arrow), situated at the apical margin of the epithelium, are still present. Panel (a) X341. Bar = 50 u.m. Panel (b) X640. Bar = 50 urn. Panel (c) X16,860. Bar = 1 urn. 119 FIGURE 4.6. Junctional complexes between two adjacent Sertoli cells in the post-spermiation testis of the bowfin. As shown in panel (a), the apically positioned Sertoli cell junctions appear very similar to those formed by Sertoli cells of the active spermatocyst (see Fig. 3.4). Located most apically is a zone of microfilament associated junctions (arrowheads). Immediately basal to this is a desmosome (curved arrow). Occasional small spot junctions (arrows) that appear to be microfilament based can be seen more basally along the highly convoluted Sertoli cell plasma membranes. These junctions occur independently of the apical junctional complex. The head of a sperm cell (asterisk) can be seen in the tubule lumen. A more grazing section through the apical junctional complex is shown in panel (b). The extent of the microfilament related junctions (arrows) and underlying desmosomes (large arrowheads) is easily seen. Bundles of intermediate filaments (small arrowheads) appear to interconnect adjacent desmosomes. Panel (a) X28,460. Bar = 0.5 urn. Panel (b) X17,940. Bar = 1 um. lao 121 FIGURE 4.7. High power electron micrograph of a junctional complex formed between two Sertoli cells in a bowfin spermatocyst. These complexes occur near the apex of the epithelium and consist of both microfilament and intermediate filament associated junctions. Most luminal in position are areas of contact related to microfilaments. A zone of classical appearing intermediate filament-related desmosomes is located immediately basal to these. The microfilaments (small arrows) are seen to be smaller in diameter than the desmosomal intermediate filaments (larger arrows). Sites of focal membrane contact (arrowheads), possibly tight junctions, are occasionally visible at the apex of the junctional complex. X94.290. Bar = 0.25 urn. 123 FIGURE 4.8. Paired fluorescence and phase micrographs of fixed frozen sections of post-spermiation bowfin testis that have been labelled with rhodamine phalloidin (a,a'-b,b') and an immunological probe for myosin II (c,c'-d,d'). In each micrograph, the asterisks indicate the position of the tubule walls. When stained with rhodamine phalloidin, intense fluorescence is seen in the tubule walls (a,b), presumably from peritubular cells. In cross-sections of the tubules (a,a'), isolated, vertical tracks of fluorescence (arrowheads) are also detectable near the lumenal surface of the epithelium, most likely representing the borders of adjacent Sertoli cells. In more grazing sections of the tubules (b,b'), the tracks appear to be continuous with one another and form a honeycomb pattern (arrowheads). When stained for myosin II, a very similar pattern of labelling is observed. Short vertical bands of fluorescence (arrowheads) are seen within the epithelium in cross sections of tubules (c,c'), while in more grazing sections (d,d') these bands appear to link up (arrowheads). Intense labelling with the myosin probe is observed in the tubule walls (c,d). X717. Bar = 20 um. 125 FIGURE 4.9. Shown in panel (a) is a light micrograph of toluidene blue stained plastic section of a spermatocyst of the guppy testis. At this stage of spermatogenesis, spermatids (arrows) are entrenched in the apical surface of Sertoli cells (arrowheads) which form a simple cuboidal epithelium which lines the spermatocyst. Notice the zones of homogeneous staining in Sertoli cell regions that surround the entrenched spermatid heads (asterisks). Shown in panel (b) is a low magnification electron micrograph of a Sertoli cell containing entrenched spermatids (three of which are indicated by white asterisks) in its apical surface. Notice that concentrations of microfilaments (black asterisks) are related to each of the recesses and extend deeper into the apical cytoplasm of the Sertoli cell. Microvilli (arrowheads) that extend into regions of the lumen occupied by spermatid tails are visible at the top of the micrograph. The Sertoli cell (N) nucleus is indicated. Panel (a) X1.365. Bar = 10 jam. Panel (b) X19,000. Bar = 1 um. 127 FIGURE 4.10. Regions of contact between Sertoli cells and entrenched spermatids in the guppy spermatocyst are shown in panels (a) and (b). Fibrous linkages (arrows) are obvious between the plasma membrane of Sertoli cells and that of spermatids. Also evident in panel (a) is a loose arrangement of microfilaments around and extending basally away from the recess. In panel (c) a microvillus (arrowheads) extends from the apical border of a guppy Sertoli cell into regions of the lumen occupied by spermatid tails. Note that microfilaments in the microvillus appear to extend out of the apical network of filaments associated with the recesses containing spermatids. Panel (a) X74,100. Bar = 0.25 um. Panel (b) X101.900. Bar = 0.2 um. Panel (c) X19,760. Bar = 1 um. 129 FIGURE 4.11. Apical junctional network between neighboring Sertoli cells of the guppy. Most apical in position are areas of contact related to loosely arranged microfilaments. These areas are indicated in panels (a) and (b) by brackets; small asterisks indicate clusters of microfilaments. Immediately basal to these areas is a well developed zone of intermediate filament-related desmosomes indicated by the large arrowheads in panels (a) and (c). Panel (a) X65.625. Bar = 0.25 um. Panel (b) X76.120. Bar = 0.25 um. Panel (c) X76.140. Bar = 0.25 um. 131 FIGURE 4.12. Paired fluorescence and phase micrographs of fixed frozen sections of guppy spermatocysts in which spermatids (black arrowhead in panel a') are entrenched in the apical margins of Sertoli cells. When stained with rhodamine phalloidin (a,a'), intense labelling occurs in the apical cytoplasm of Sertoli cells associated with recesses containing the elongate spermatids (large white arrowhead in panel a and inset). Elongate threads of staining (small arrow in panel a) extend into regions of the lumen containing spermatid tails. Also notice that there is some staining at the base of the epithelium (small arrowhead in panel a and inset), most likely in cells of the spermatocyst wall. Shown in panels (b,b') is a section similar to that shown in panel (a,a'), but which has been treated with an immunological probe for myosin II. Note that the most intense staining occurs in the apical zone of Sertoli cell cytoplasm (arrowhead in panel a and inset) related to proximal ends of spermatid heads. Although the staining pattern is similar to that obtained with rhodamine phalloidin, notice that labelling does not appear to extend into the lumen amongst spermatid tails as does the staining with the actin probe. Panels (a,a'-b,b') X690. Bars = 20 um. Insets X1.080. Bars = 10 um. 133 FIGURE 4.13. Paired fluorescence and phase micrographs of fixed frozen sections of guppy spermatocysts stained with rhodamine phalloidin (a,a') and an immunological probe for myosin II (b,b'). In these grazing sections, regions of intercellular junction between neighboring Sertoli cells are cut en face. A distinct honeycomb pattern of labelling (large arrowheads in panel a) is clearly visible in material stained for actin and to a lesser extent in material labelled for myosin II (large arrowheads in panel b). Notice that a linear pattern of staining for actin (small arrowheads in panel a) and for myosin II (small arrowheads in panel b) can be detected in cells of the spermatocyst wall. Panels (a,a'-b,b') X800. Bars = 2 um. 135 CHAPTER 5 Actin-Related Adhesion Junctions in Sertoli Cells of the Rooster, Turtle and Alligator 136 I n t r o d u c t i o n Ectoplasmic specializations in vertebrate Sertoli cells consist of a layer of actin filaments together with adjacent regions of the plasma membrane involved with intercellular attachment. In eutherian mammals, ectoplasmic specializations occur at sites of attachment to spermatids and adjacent to basally situated junctions (blood-testis barrier) between Sertoli cells (Nicander, 1967; Dym and Fawcett, 1970; Flickinger and Fawcett, 1970; Russell, 1977a,b). A cistern of endoplasmic reticulum is linked to the cytoplasmic surface of the actin layer. In non-mammalian vertebrates, ectoplasmic specialization-like junctions are reported mainly in areas of attachment to spermatids (Baccetti etal., 1983; Stanley and Lambert, 1985; Sprando and Russell, 1987a). A cistern of endoplasmic reticulum is not always evident (Cooksey and Rothwell, 1 973; Osman et al., 1980; Sprando and Russell, 1987a). The function of ectoplasmic specializations in general is not entirely clear; however, there is a growing body of evidence to support the hypothesis that they are primarily a form of actin-associated adhesion junction related to the zonulae adherens of other epithelia (reviewed by Vogl, 1989; Grove et al., 1990; and Vogl et al., 1991a,b). One component of this evidence is that ectoplasmic specializations react positively with immunological probes for vinculin - a molecular marker for actin associated adhesion junctions (Grove and Vogl, 1989; Grove et al., 1990; Pfeiffer and Vogl, 1991) . Although the full complement of integral membrane molecules responsible for intercellular adhesion at these sites have not been identified conclusively, there is evidence that, in mammals, one of the elements is an integrin (Pfeiffer et al., 1991; Palombi et al., 1992; Salanova et al., 1995). One property that characterizes the actin complexes associated both with the zonulae adherens and with actin associated adhesion junctions in general, is that of contractility. In fact, the filaments associated with the zonula adherens form a 137 "contractile ring" that circumscribes the apices of the cells in regions of the adhesion junction (Burgess, 1982; Hirokawa et al., 1983; reviewed by Mooseker, 1985). In contrast, ectoplasmic specializations of eutherian mammals are not contractile. Actin filaments within the structures are hexagonally packed (Dym and Fawcett, 1970; Russell, 1977a,b; Franke et al., 1978) and are arranged in a unipolar fashion (Toyama, 1976; Vogl etal., 1986). Moreover, the structures do not contain myosin II (Suarez-Quian and Dym, 1984; Vogl and Soucy, 1985) nor can they be induced to contract in vitro in the presence of ATP with or without calcium (Vogl and Soucy, 1985). The situation in non-mammalian vertebrates may be different. In these animals, ultrastructural studies (Osman et al., 1980; Baccetti et al., 1983; Stanley and Lambert, 1985; Sprando and Russell, 1987) indicate that the arrangement of actin filaments at sites where ectoplasmic specializations occur in mammals closely resembles that of filaments in actin bundles associated with intermediate (adherens) junctions in general; that is, the filaments are not hexagonally packed, but appear to form "loosely" packed bundles. Based on this observation, one might predict that ectoplasmic specialization-like junctions in the testes of non-mammalian vertebrates may possess contractile properties. Data from a previous study of a non-mammalian vertebrate, the ratfish, is consistent with this prediction (Stanley and Lambert, 1 985) . The microfilament-related junctions at apical sites in Sertoli cells of this animal stain positively with an immunological probe for myosin II, indicating that the basic components necessary to generate contraction are present in this system. In this thesis I have provided further evidence that ectoplasmic specialization-like junctions of non-mammalian vertebrates may be contractile. Myosin II is detectable at the apically located junctions in the dogfish and the guppy and, in both species, the filaments at these sites are arranged in loosely packed bundles (see Chapters 2 and 3). Our laboratory suspects that ectoplasmic specializations in mammals are unique in many respects and that similar structures in non-mammalian vertebrates may more 138 closely resemble the basic form of actin-associated adhesion junction commonly found i n other cell types. In this chapter I report the results of an experiment designed to verify the prediction that ectoplasmic specializations in non-mammalian vertebrates, like actin-associated adhesion junctions in general, have contractile properties. I demonstrate that these structures in two additional classes of non-mammalian vertebrates (Aves and Reptilia) react positively with probes both for actin and for myosin II. Moreover, in the turtle, ectoplasmic specializations can be induced to contract in vitro when exposed to appropriate conditions. These results, together with the observation that these structures in the rooster react with an immunological probe for vinculin, are consistent with the general hypothesis that ectoplasmic specializations are a form of actin-associated adhesion junction. 139 Materials and Methods Animals: A total of 4 roosters (Gallus domesticus), 8 turtles (Pseudemys scripta) and 2 alligators (Alligator mississippiensis) were used in this study. The roosters were obtained from the animal care center at the University of British Columbia and the turtles were obtained from Nasco (Lemberger, Wl). Animals were anesthetized either halothane administered via the respiratory system (roosters) or with sodium pentobarbitone administered intraperitoneal^ (turtles). The alligator tissue was obtained from two animals killed by Louisiana Department of Wildlife and Fisheries personnel at the Rockefeller Wildlife Refuge (Grand Chenier, Louisiana). Electron Microscopy: Testes from 2 roosters, 2 turtles and 2 alligators were processed for electron microscopy. Rooster and turtle samples were processed as follows: Animals were anesthetized and their hearts exposed. An 18G syringe needle, attached to a perfusion apparatus, was inserted into the left ventricle and the inferior vena cava was cut. The animals were perfused with phosphate-buffered saline (150 mM NaCI, 5 mM KCI, 3.2 mM [\la2HPO4, 0.8 mM KH2PO4, pH 7.3) for approximately two minutes followed immediately by perfusion with an EM fixative containing 1.5% paraformaldehyde (Fisher Scientific, Pittsburgh, PA), 1.5% glutaraldehyde and 0.1M sodium cacodylate (J.B. EM Services, Inc.) (pH 7.3). After 20-30 min, the perfusion was stopped and the testes were removed. They were cut into small blocks, fixed an additional 2 hr in the same fixative, washed with buffer, and then post-fixed in cold (ice) 1% osmium tetroxide (J.B. EM Services, Inc.) in 0.1 M sodium cacodylate. The blocks were washed 140 with dH20, treated with 1% aqueous uranyl acetate for 1 hr, washed again with dH20, and then dehydrated through alcohol and embedded in Polybed 812 (Polysciences, Inc). Alligator tissue samples were processed as follows: Testes were removed from freshly killed animals and cut into small blocks (Imm^) that were immersion-fixed in EM fixative (see above) for 3 hrs. Following this, tissue blocks were stored in 0.1 M sodium cacodylate for 4 days , then washed twice in this buffer and post-fixed in cold (ice) 1% osmium tetroxide (J.B. EM Services, Inc.) in 0.1 M sodium cacodylate. Alligator tissue samples were then further processed in the same manner as the rooster and turtle tissue described above. Thick (1 pirn) sections were stained with toluidene blue. Thin sections were stained with both uranyl acetate and lead citrate and were examined on a Philips 30 0 electron microscope operated at 80 kV. Chemicals and other compounds used for this set of experiments and all others in this study were obtained from Sigma Chemical Co. (St. Louis, MO), unless otherwise noted. Tissue Preparation for Localization of Actin. Myosin II and Vinculin: To obtain testicular material for fluorescence localization of actin and myosin 11, testes from 2 roosters, three turtles and 2 alligators were used. Anesthetized roosters and turtles were perfused, via the left ventricle, with 3% paraformaldehyde in phosphate buffered saline (PBS). After 10 min, the animals were then perfused with PBS for 20-30 min. The testes were removed and cut into blocks of approximately 1 c m 3 . The blocks were frozen at liquid nitrogen temperatures in Tissue-Tek O.C.T. Compound (Miles Inc.) and stored in liquid nitrogen until they were sectioned for fluorescence localization of these two proteins. In two of the three turtles, the testes did not perfuse well so they were cut into blocks and fixed by immersion for an additional 10 min and then washed with PBS prior to being frozen. Alligator testes were excised 141 from freshly killed animals and cut into blocks of approximately 1 c m 3 in 0.1 M sodium cacodylate containing 3% paraformaldehyde. After immersion-fixing is this fixative for 1 hr, the blocks were washed once with 0.1 M sodium cacodylate and stored for 4 days in this buffer. Blocks were then washed twice with sodium cacodylate and frozen at liquid nitrogen temperatures in Tissue-Tek O.C.T. Compound (Miles Inc.) until they were sectioned for the fluorescence localization of actin and myosin II. Only rooster testes were used for vinculin staining. This tissue was processed in the same manner as the other rooster tissue described above. In order to verify the location of actin and myosin II staining associated with spermatids of the turtle, testes from two additional animals were excised, placed in 3% paraformaldehyde in PBS, and minced with scalpels. After 10 min the material was collected, placed in a centrifuge tube and washed three times with PBS. Following this, the material was aspirated repeatedly with a pipette to mechanically fragment the tissue. The resulting suspension was placed on polylysine coated slides, treated with cold acetone ( -20°C) and then air dried. Staining for actin and for myosin was done as indicated below for tissue sections. Spermatids and attached junctions, which together had been separated from the seminiferous epithelium and could be identified with phase microscopy, were evaluated for reactivity with the fluorescence probes. Fluorescence Localization of Actin: Frozen sections (7-10 urn thick) of fixed testis were collected on polylysine coated slides and immediately plunged into cold ( - 2 0 ° C ) acetone. After 5 min the slides were removed and air dried for approximately 30 min. The sections were rehydrated, for 10 min, with PBS and then incubated, for 20 min, in one of the following: (1) PBS + 1.65 x 1 0 " 6 M NBD-phallacidin (nitrobenzoxadiazole-phallacidin, Molecular Probes, Inc.) (test for filamentous actin); (2) PBS + 1.65 x 1 0 - 6 M NBD-phallacidin + 1.04 142 x 1 0 " 4 M phalloidin (competitive specificity control); (3) PBS + 1.04 x 1 0 " 4 M phalloidin (control for phalloidin in reagent 2); (4) PBS (control for autofluorescence). Mechanically fragmented tissue of the turtle was stained for filamentous actin in a similar fashion to sections except that 1.65 x 1 0 " 6 rhodamine phalloidin (Molecular Probes, Inc.) was used instead of NBD-phallacidin in solutions (1) and (2), and 7.76 x 10" 4 M phallacidin was used instead of phalloidin in solutions (2) and (4). All slides were washed with PBS, mounted in 1:1 glycerol:PBS containing 0.02% sodium azide, and then examined either on a Zeiss Photomicroscope III or a Zeiss Axiophot each fitted with filter sets for detecting fluorescein isothiocyanate or rhodamine. Immunofluorescence Localization of Myosin II and Vinculin: Frozen sections of testis (7-10 um thick) were collected on polylysine coated slides and immediately plunged into cold ( - 2 0 ° C ) acetone (myosin) or cold ( - 2 0 ° C ) methanol (vinculin). After 5 min the slides were removed and air dried for approximately 30 min. Tissue was rehydrated in TPBS (0.05% Tween-20 in PBS) containing 0.1% bovine serum albumin (BSA) and 5% normal goat serum (NGS). The slides were drained then incubated, at 37°C for 1 hr, with primary antiserum at a dilution of 1:50 (myosin) or a concentration of approximately 50 ug/ml (vinculin) in TPBS containing 0.1% BSA and 1.0% NGS. After washing three times with TPBS containing 0.1% BSA, the sections were incubated, at 37°C for 1 hr, with a goat ant i -rabbit IgG conjugated to fluorescein (Sigma Co.) diluted 1:32 with TPBS/0.1% BSA. The slides were washed three times with TPBS/0.1% BSA then mounted with the same buffer diluted 1:1 with glycerol and containing 0.02% sodium azide. The primary rabbit antiserum raised against human platelet myosin II and the matched preimmune serum were kindly provided by Dr. Kegi Fujiwara (National 143 Cardiovascular Center Research Institute, Osaka, Japan). The primary antiserum to vinculin was produced in rabbits against vinculin isolated from human platelets (Grove and Vogl, 1989; Grove et al., 1990). The antiserum was affinity purified as described elsewhere (Grove and Vogl, 1989). Controls for myosin II staining included the following: (1) substitution of the immune serum with preimmune serum; (2) substitution of the primary antiserum with buffer; (3) substitution both of the primary and of the secondary antibodies with buffer. Controls for vinculin staining included: (1) substitution of the affinity purified antibody with a 1:25 dilution of the non-adsorbed serum from the affinity column; (2 ) preincubation of specific (affinity purified) antibody with antigen (1 mg/ml); ( 3 ) substitution of primary antibody with buffer; (4) substitution of primary antibody and of secondary antibody with buffer. Contraction Experiments: (a) Rationale: If ectoplasmic specialization-like junctions in non-mammalian vertebrates are contractile, then they should contract when exposed to appropriate conditions. The approach I chose to verify this prediction was similar to that used by investigators working with other epithelia (Hirokawa etal., 1983) and previously used by Vogl and Soucy (1985) on the mammalian seminiferous epithelium; that is, I prepared glycerinated samples of the epithelium and exposed them to buffers containing ATP with or without calcium. These epithelial preparations were then fixed and the junctions labelled with NBD-phallacidin. Diameters of junctions exposed to treatment buffers were compared to those of junctions exposed to control buffers. A reduction in diameter of the structures exposed to treatment buffers was interpreted as evidence of contraction. 144 I chose to use turtles for these experiments because ectoplasmic specialization-like junctions in these animals are larger than those in the rooster. I felt that, using the fluorescence assay, any induced contraction would be easily detected in the turtle system. A preliminary experiment, in which the results were evaluated qualitatively, was followed by two additional experiments in which data were analyzed statistically. Each experiment was done with epithelium derived from a different animal. (b) Preparation of Epithelial Fragments: The protocol for obtaining epithelial fragments from turtle testes was generally similar to that used to obtain similar material from mammalian testes (Vogl and Soucy, 1985). Testes were removed from anesthetized animals and placed in cold (ice) PBS containing 20 mM EDTA. The testes were decapsulated and the material minced, for no more than ten minutes, with two scalpels used in a scissor-like fashion. The material, containing epithelial fragments, tubule walls, intact tubules and interstitial tissue, was collected and placed into centrifuge tubes. A small aliquot from each animal was fixed with 3% paraformaldehyde in PBS for 10 min, washed with PBS, and labelled with NBD-phallacidin as described in section (d) below. (c) Preparation of Glycerinated Epithelial Samples: Glycerinated samples were prepared generally as described by Vogl and Soucy (1985) for mammalian seminiferous epithelium. The material containing epithelial fragments was centrifuged for three minutes at setting 7 in an IEC clinical centrifuge. The supernatant was discarded and replaced with cold (ice) extraction buffer containing 50% (wt/vol) glycerol in 100 mM KCI, 3.2 mM Na2HPC>4, 0.8 mM KH2PO4, 5.0 mM EDTA, 0.1 mM PMSF and 1.0 mg/ml soybean trypsin inhibitor (pH 7.0). The material was gently resuspended and left overnight at 4°C. 145 (d) Contraction Protocol: Contraction experiments were modelled after those described by Hirokawa et al. (1983) for intestinal epithelia. The testicular tissue in extraction buffer was centrifuged for 3 min. at setting 7 in an IEC clinical centrifuge and the supernatant replaced with cold (ice) control buffer (25.0 mM piperazine-N,N'-bis(2-ethane sulfonic acid) (PIPES), 50.0 mM KCI, 5.0 MgCl2, 0.2 mM DTT, 1.0 mM EGTA). The tissue was resuspended and left on ice for no more than 10 min. This step was repeated and the tissue was then resuspended in one of the following: (1) control buffer; ( 2 ) control buffer containing 0.9 mM CaCl2; (3) control buffer containing 2.0 mM ATP; (4) control buffer containing 2.0 mM ATP and 0.9 mM CaCl2- The samples were incubated for 10 min at 37°C , centrifuged, and the resulting pellets were gently resuspended in treatment buffers containing 3% paraformaldehyde. The material was fixed for 10 min, centrifuged, washed three times with PBS, and then processed for localization of actin by fluorescence. (e) Localization of Actin: Fixed and washed tissue suspensions were centrifuged and the tops of the pellets, that contained mainly epithelial fragments, were resuspended in a small amount of PBS. The fragments were placed on polylysine coated slides and all excess fluid removed with pipettes. The slides were placed in cold ( - 2 0 ° C ) acetone for 5 min, air dried, then stained with NBD-phallacidin as described above. (f) Quantitative Analysis Clusters of junctions were photographed on a Zeiss Axiophot photomicroscope. Photographic negatives of ten clusters per treatment were printed at identical magnification. The diameters of ten individual junctions within each cluster were measured using a Kontron Image Analysis System and the mean diameter for each cluster 146, was calculated. This resulted in ten diameter values (n=10) for each of the four treatments (buffer, buffer + calcium, buffer + ATP, and buffer + ATP + calcium). The data were analyzed using an ANOVA to test for general effects of treatment and aTuky's post-hoc test to check for differences among treatments. All quantitative data were left in "pixel" units, as determined by the image analysis system, and were not converted to true metric values. The final magnification of the photographic prints used to obtain measurement data for the second experiment was slightly greater than that of prints used to obtain measurement data for the first experiment. 147 R e s u l t s (A) Ultrastructure Ectoplasmic specialization-like junctions associated with elongate spermatids of the rooster, turtle and alligator consist of a layer of filaments underlying the Sertoli cell plasma membrane (Figs. 5.1-5.3). As reported by others (Sprando and Russell, 1987a), the filaments do not appear hexagonally packed. A cistern of endoplasmic reticulum is associated with the junctions in the turtle and alligator, but is absent from those of the rooster. No junctions similar to those between Sertoli cells and spermatids were observed, in either species, between neighboring Sertoli cells at the base of the epithelium. Rather, attachment junctions in these regions appeared to be either desmosomes, or small junctions not obviously associated with intermediate filaments. These junctions were also observed between Sertoli cells and early spermatogenic cells. (B) Localization of Actin. Myosin II and Vinculin in Fixed Frozen Sections Rooster ( Gallus domesticus): Probes for actin, myosin II and vinculin reacted positively in the seminiferous epithelium and tubule walls of the rooster testis (Fig. 5.4a,b,c). Significantly, all three probes were specifically reactive with regions associated with the heads of elongate spermatids. These regions correspond to those sites known from ultrastructural studies to contain apically situated ectoplasmic specialization-like junctions (Osman et al. , 1980; Sprando and Russell, 1987a; this chapter). Actin: In the seminiferous epithelium, specific fluorescence due to NBD-phallacidin binding occurred in three major locations: (1) In regions of the epithelium immediately adjacent to the tubule lumen, (2) In regions associated with the heads of apically 148 situated spermatid heads, and (3) in focal regions in the basal half of the epithelium (Fig. 5.4a). The fluorescence immediately adjacent to the lumen appeared punctate in pattern and was somewhat more diffuse than that associated with the spermatid heads situated deeper in the epithelium. The fluorescence signal emitted from regions associated with groups of elongate spermatids was intense relative to that elsewhere i n the epithelium and occurred as clusters of short linear bands. Each band appeared to be positioned around the anterior region of a spermatid head. No linear staining, characteristic of basally situated ectoplasmic specializations in mammals, was observed in basal regions of the seminiferous epithelium in the rooster. Rather, fluorescence i n these regions was generally punctate or focal (Fig. 5.4a). Occasionally, tracts of punctate fluorescence appeared to outline the margins of early spermatogenic cells (Fig. 5.5a). The most intense fluorescence in seminiferous tubules of the rooster was emitted by the tubule wall, presumably due to labelled filamentous actin in peritubular cells (Fig. 5.4a). Controls for staining with the NBD-phallacidin are shown in Figure 5.5. No staining as described above was observed in any of the control slides; however, some nonspecific autofluorescence was present in the seminiferous epithelium. This fluorescence was generally diffuse or granular in nature and occurred in the middle to outer third of the epithelium. Although often associated with clusters of spermatids, the staining was distinct from the linear bands of specific fluorescence associated with the anterior ends of spermatid heads. Myosin II: The overall pattern of fluorescence obtained with the probe for myosin 11 (Fig. 5.4b) was similar to that obtained with the probe for actin; that is, specific fluorescence occurred in (1) regions immediately adjacent to the lumen, (2) regions associated with the anterior ends of elongate spermatids, (3) areas near the base of the epithelium, and (4) cells within the tubule wall. In regions immediately adjacent the 149 tubule lumen, staining was intense (Fig. 5.4b) and was similar in pattern to that described above for the actin probe NBD-phallacidin. Staining associated with the anterior ends of elongate spermatids was somewhat more focal or punctate than that obtained with the actin probe; however, linear bands of fluoresence, similar to those obtained with NBD-phallacidin, were observed in association with some of the spermatids (Fig. 5.4b). Although a positive signal appeared to occur in the basal half of the epithelium (Figs. 5.4b), the fluoresence was much more diffuse than that obtained with the actin probe. Like NBD-phallacidin, the myosin II antiserum was strongly reactive with cells in the tubule wall (Figs. 5.4b). Controls for staining with the myosin II antiserum are shown in Figure 5.6. No fluorescence, as described above, was observed when preimmune serum was substituted for the antiserum (Fig. 5.6b), when the primary antiserum was omitted (Fig. 5.6c), nor when both the primary and secondary antibodies were omitted (Fig. 5.6d). Vinculin: In seminiferous tubules of the rooster, the vinculin probe reacted specifically and strongly at two major locations: (1) Regions associated with the anterior ends of elongate spermatids; (2) Cells in the tubule wall. Interestingly, fluorescence associated with each spermatid was patterned in the shape of a "U" that appeared to cap the anterior end of the spermatid head (Figs. 5.4c). As with NBD-phallacidin and the myosin probe, the vinculin antibody reacted strongly with cells in the tubule wall. No convincing signal was detected in basal regions of the epithelium nor in regions immediately adjacent to the lumen. Controls for the vinculin staining are shown in Figure 5.7. Preincubation of antibody with antigen and incubation of tissue sections with antibody in the presence of antigen competitively reduced fluorescence in the tubule wall and completely eliminated fluoresence associated with elongate spermatids (Fig. 5.7b). No specific fluorescence was observed when the primary antibody was replaced with the non-adsorbed serum 150 from the affinity column (Fig. 5.7c), when the primary antibody was omitted (Fig. 5.7d), nor when both primary and secondary antibodies were omitted (Fig. 5.7e). Alligator (All igator mississipiensis): As with the rooster, I treated fixed frozen sections of alligator testis with probes for actin and myosin II. These probes reacted specifically with regions associated with elongate spermatids and with cells in the tubule wall (Fig. 5.8a,b). Actin: In the seminiferous epithelium, intense labelling occurred in regions of elongating spermatids. This staining was apparent as linear bands (Fig. 5.8a) and corresponded to the position of elongating spermatid heads seen by phase contrast (Fig 5.8a') A punctate pattern of fluorescence was also seen in basal regions of the epithelium, and often this fluorescence appeared to outline the margins of early spermatogenic cells. Intense staining of regions associated with the tubule wall was apparent (Fig. 5.8a). Presumably, much of this staining was due to labelling of peritubular cells; however, ultrastructural data reported in Chapter 7 indicate that some of this fluorescence may have resulted from labelling of microfilament bundles found adjacent to the basal plasma membrane of Sertoli cells. No specific fluorescence as described above was observed in any of the control slides for NBD-phallacidin staining (data not shown). Autofluorescence similar to that observed in the rooster was not present in the alligator. Myosin II: In the seminiferous epithelium, specific staining with the myosin II probe occurred in regions around elongating spermatid heads (Fig. 5.8b). The fluorescence was not as intense or as well defined as that seen in similar areas with the actin probe in this species. Rather, the fluorescence appeared as a diffuse labelling around the anterior ends 151 of groups of elongating spermatid heads. Strong labelling also occurred in regions associated with the tubule wall. This most likely represented peritubular cells and possibly basal portions of Sertoli cells (see Chapter 7). No specific fluorescence, as described above, was seen in any of the control slides run for the myosin II probe in the alligator (data not shown). Turtle ( Pseudemys scripts): In fixed frozen sections of turtle testis, probes for actin and myosin II labelled tissue in a similar pattern to that seen in the alligator. These probes reacted specifically with regions associated with elongate spermatids and with cells in the tubule wall (Fig. 5.9a,b). Actin: In the seminiferous epithelium, staining with NBD-phallacidin occurred mainly in association with elongate spermatids (Fig. 5.9a). The probe appeared to label regions around the anterior end of each head. As in the rooster and alligator, punctate staining was present in basal regions of the epithelium. Staining of regions associated with the tubule wall was intense. Much of this staining was presumably due to labelling of peritubular cells; however, some of this fluorescence may have represented basal Sertoli cell microfilament bundles (see Chapter 7). No specific staining, as described above, was seen in any of the control slides (data not shown). Autofluorescence similar to that observed in the rooster was not observed. Myosin II: Like NBD-phallacidin, the probe for myosin II stained cells in the tubule wall, presumably peritubular cells and possibly basal regions of Sertoli cells, and reacted specifically with regions associated with the anterior ends of elongate spermatids (Fig. 5.9b,c). In the latter location, staining generally appeared less extensive in 152 distribution than that observed with the actin probe. Also, fluorescence associated with individual spermatids at one stage of spermatogenesis often occurred in a "V" shaped configuration that capped the anterior end of each spermatid head (Fig. 5.9c). This pattern corresponded with the actin pattern present at the same stage of spermatogenesis (compare the myosin pattern in Fig. 5.9c with the pattern indicated by the small arrowheads in Fig. 5.12a). No specific fluorescence was observed when primary antiserum was replaced with preimmune serum (Fig. 5.10b), when primary antiserum was omitted (Fig. 5.10c), nor when primary and secondary antibodies were omitted (Fig. 5.1 Od). Mechanically Isolated Spermatids of the Turtle: To verify that myosin II staining was associated with ectoplasmic specialization-like junctions in the turtle, spermatids were mechanically isolated from immersion fixed seminiferous epithelium and stained for either actin or myosin. With both probes, elongate profiles of specific staining were observed along the anterior ends of some spermatid heads (Fig. 5.11). The actin pattern was indicative of the position of junctions in this species and suggested that, like in mammals, ectoplasmic specialization-like junctions remained attached to spermatids mechanically dissociated from the epithelium. The pattern of myosin II staining was similar to that of actin, thereby indicating that the observed positive staining for myosin II was most likely in the junctions. 153 (C) Contraction Experiments Actin Distribution in Isolated Seminiferous Epithelium of the Turtle: In isolated segments of turtle epithelium prior to glycerination, NBD-phallacidin labelling occurred mainly in regions associated with clusters of elongate spermatids (Fig. 5.12a,a'). Within each cluster of spermatids, cylindrical columns of fluorescence were present around, and appeared to extend from, the anterior ends of the needle-shaped spermatid heads. This was particularly evident when spermatids together with attached.regions of Sertoli cells were mechanically detached from the epithelium (Fig. 5.12b,b'). Hexagonal patterns of intense actin staining similar to those observed at the "blood-testis barrier" in isolated mammalian epithelium were not observed in epithelial fragments of the turtle (Fig. 5.12a,a'). Rather, staining in regions corresponding to basal Sertoli-Sertoli cell junctions in mammals was at best weak and diffuse. Contraction of Glycerinated Epithelial Samples: In glycerinated samples of turtle epithelium, groups of actin-related junctions had often separated from their associated clusters of spermatids and formed "ball-like" networks. Examples of these networks in samples treated with buffer or with buffer containing ATP are shown in Fig. 5.13a and Fig. 5.13b respectively. The diameters of individual junctions within these networks were statistically smaller in samples treated with ATP or with ATP in the presence of calcium than in samples incubated in buffer alone or in buffer containing calcium (see Fig. 5.14). Controls for NBD-phallacidin staining in these experiments were all negative. 154 D i s c u s s i o n In this chapter I present evidence that, unlike ectoplasmic specializations in Sertoli cells of mammals, the actin-related junctions in similar cells of the turtle, alligator and rooster have contractile properties. In eutherian mammals, ectoplasmic specializations consist of a layer of actin filaments together with a cistern of endoplasmic reticulum on one side of the filament layer and regions of the plasma membrane involved with intercellular attachment on the other (Brokelmann, 1963; Flickinger and Fawcett, 1967; Nicander, 1967; Dym and Fawcett, 1970; Russell, 1977a). The actin filaments are organized into distinct bundles and are cross-linked to each other and to adjacent membranes (Russell, 1977a,b; Franke et al., 1978; Romrell and Ross, 1979; Vogl et al., 1986; Grove and Vogl, 1989; Yazama et al., 1991). Ectoplasmic specializations are mainly found in two locations in mammalian Sertoli cells: next to sites of attachment to spermatids and next to basal sites of junction between adjacent Sertoli cells. At sites of attachment to spermatids, ectoplasmic specializations surround Sertoli cell invaginations (apical crypts) in which the spermatids are embedded (Fawcett, 1975; Russell, 1977a; Russell and Peterson, 1985). At basal locations, the structures circumscribe Sertoli cells in regions of the "blood-testis" barrier (Dym and Fawcett, 1970; Russell 1977b; Weber et al., 1 983; Vogl and Soucy, 1985) . Although the function of ectoplasmic specializations is not entirely clear, it has previously argued (Grove and Vogl, 1989; Vogl et al., 1991a; 1993) that they are most likely a form of actin-associated adhesion junction. Evidence consistent with this hypothesis includes the following observations: 155 (1) Ectoplasmic specializations occur both at apical and at basal sites of intercellular attachment and the form of junction common to both sites appears to be of the adhesion type. (2) Ectoplasmic specializations remain tightly adherent to spermatids mechanically dissociated from the epithelium (Frank et al., 1978; Romrell and Ross, 1979; Vogl and Soucy, 1985; Vogl et al., 1985;1986; Masri et al., 1987; Grove and Vogl, 1989). (3) The natural or pharmacological disruption of actin filaments at apical sites results in a loss of intercellular adhesion (Russell 1977a,b; Russell et al., 1988; Weber et al. , 1988) . (4) Immunological probes for vinculin, a molecular marker for actin-associated adhesion sites, reacts with ectoplasmic specializations at both the light and ultrastructural level (Grove and Vogl, 1989; Grove et al., 1990; Pfeiffer and Vogl, 1991) and reacts with a band of the appropriate molecular weight for vinculin on immunoblots of testicular fractions enriched for ectoplasmic specializations (Grove and Vogl, 1989). (5) Immunological probes for integrins, a known class of adhesion molecules of which some members have been found at cell/cell contacts in other systems (Carter et al. , 1990; Larjava etal., 1990; Lampugnani et al., 1991), react with epithelial regions containing ectoplasmic specializations (Pfeiffer et al., 1991; Palombi et al., 1 992; Salanova et al., 1995). Mammalian ectoplasmic specializations are not contractile. Constituent actin filaments are hexagonally packed (Dym and Fawcett, 1970; Russell, 1977a,b; Franke et al., 1978) and have a uniform polarity (Toyama, 1976; Vogl et al., 1986). In addition, immunological probes for myosin II do not react with the structures (Suarez-Quian and Dym, 1984; Vogl and Soucy, 1985) nor do glycerinated samples of ectoplasmic specializations contract when exposed to conditions that generate contraction in other systems (Vogl and Soucy, 1985). These features of the actin bundles in mammalian 156 ectoplasmic specializations are different than those of actin bundles associated with junctions in most other epithelia where filament bundles are contractile and in which the filaments are "loosely" cross-linked (Burgess, 1982; Hirokawa et al., 1983; reviewed by Mooseker, 1985). Ectoplasmic specialization-like junctions have been described by others in non-mammalian vertebrates; however, unlike those in mammals, the structures have been reported to be largely restricted in location to sites of contact between Sertoli cells and spermatids. Only in the cartilaginous fish is a structural homologue of the ectoplasmic specialization found at sites of contact between neighboring Sertoli cells (see Chapter 3). The finding in this study that ectoplasmic specialization-like junctions of the rooster react positively with an immunological probe for vinculin support the general hypothesis that these sites are a form of actin-associated adhesion junction. Interestingly, the actin filaments in ectoplasmic specialization-like junctions in non-mammalian vertebrates are not hexagonally packed. Rather, they are loosely arranged into bundles (Osman et al., 1980; Baccetti et al., 1983; Stanley and Lambert,, 1985; Sprando and Russell, 1987a; Chapters 3 and 4). Moreover, these junctions in several non-mammalian vertebrates (ratfish - Stanley and Lambert, 1985; dogfish -Chapter 3; guppy - Chapter 4) react with immunological probes for myosin II. Taken together, these observations indicate that ectoplasmic specialization-like junctions in non-mammalian vertebrates may possess contractile properties. A number of my observations in this chapter are consistent with the prediction that ectoplasmic specialization-like junctions in non-mammalian vertebrates may be contractile. First, the ultrastructure of the actin bundles in the turtle, alligator and rooster is typical of non-muscle contractile bundles; that is, the filaments are loosely arranged and are not hexagonally packed (see also Osman et al., 1980; Sprando and Russell, 1987a). Second, probes for myosin II react positively with regions in fixed frozen sections known to contain the sites. That this reactivity is in the ectoplasmic 157 specialization-like junctions and not within the germ cells themselves is indicated by the high intensity of the staining, the position of the label at the anterior end of the spermatid head, and the general co-distribution of the staining with that of the actin probe. Third, glycerinated samples of turtle ectoplasmic specialization-like junctions contract when exposed to standard contraction buffers. The observed decrease in diameter of these structures after incubation in contraction buffers is consistent with the presence of actin bundles organized in a spiral arrangement in regions adjacent to and along the long axis of the needle-shaped spermatid heads. Ultrastructural studies in the turtle indicate that the filaments are, in fact, arranged in this fashion around the spermatid heads (Sprando and Russell, 1987a; this study). Contraction of such a system would acount for the observed reduced diameter of each junctional network. My findings in the turtle, alligator and rooster system, together with data from the ratfish (Stanley and Lambert, 1985), dogfish (Chapter 3) and guppy (Chapter 4), indicate that the ability to contract may be a general property of ectoplasmic specialization-like junctions in non-mammalian vertebrates. Perhaps one of the more interesting features of the seminiferous epithelium i n non-mammalian vertebrates is the general lack of distinct adhesion plaques adjacent to Sertoli-Sertoli cell junctions. This is true even though (a) in many of these vertebrates a blood-testis barrier (tight junction network) occurs at least at some time during spermatogenesis and (b) ectoplasmic specialization-like junctions are present at sites of attachment to spermatids. These observations are consistent with the hypothesis that ectoplasmic specializations are not directly related to permeability junctions, but are primarily associated with junctions of the adhesion type. Of possible significance is my observation of small focal sources of fluorescence in basal regions of the seminiferous epithelium in tissues stained with NBD-phallacidin. It is possible that these sites are small focal adhesion junctions not unlike those described by Drenckhaln and Franz (1986) between mammalian enterocytes. Moreover, the structures I have identified 158 with the actin probe may be one or more of the focal junction types, described by others, in ultrastructural studies of the turtle and rooster seminiferous epithelium (Cooksey and Rothwell, 1973; Osman et al., 1980). In this chapter I present evidence that ectoplasmic specialization-like junctions in the rooster, turtle and alligator have contractile properties. This is unlike the situation in mammals where the same structures are not contractile, but is similar to the situation in most other epithelia where actin bundles associated with intercellular adhesion junctions (particularly the zonulae adherens) generally possess contractile properties. It is tempting to speculate that ectoplasmic specializations in mammals evolved from an actin-associated adhesion junction that initially had contractile properties and from which myosin II was eventually lost and in which the filaments became hexagonally packed. 159 FIGURE 5.1. Electron micrograph of a spermatid head and the associated Sertoli cell regions of a rooster. Actin filaments of the ectoplasmic specialization-like junction are indicated by the arrowheads. The spermatid head is indicated by the large white asterisk. Unlike in the turtle and alligator, these actin-related junctions of the rooster do not contain a cistern of endoplasmic reticulum. In this micrograph, the residual lobe of the spermatid overlaps the thin layer of Sertoli cell cytoplasm in which the ectoplasmic specialization-like junction is found. Two thin extensions of this lobe are indicated by the arrows. X50.122. Bar = 0.5 urn. 160 161 FIGURE 5.2. Electron micrographs of ectoplasmic specialization-like junctions associated with spermatids of the alligator. In panel (a) the actin filaments are indicated by arrowheads and the acrosome of the spermatid head is indicated by an asterisk. Shown in panel (b) is a cross-section through two junctions. On the right, the anterior tip of one spermatid head (asterisk) is seen surrounded by the actin filaments (arrow) of an ectoplasmic specialization-like junction. In the junction on the left, the plane of section has passed below the level of the spermatid head. Consequently, only the actin filaments (arrow) are visible. Note the cisternae of endoplasmic reticulum (arrowheads) closely related to the actin filaments of each ectoplasmic specialization-like junction. Panel (a) X27.900. Bar = 0.5 um. Panel (b) X30,215. Bar = 0.5 um. 165 163 FIGURE 5.3. Electron micrograph of an ectoplasmic specialization-like junction associated with a spermatid of the turtle. In this grazing section through the junction site, the spermatid head (asterisk) can be seen at the top of the micrograph. Actin filaments are indicated by the arrowheads and the related cistern of endoplasmic reticulum is indicated by the arrows. X63,100. Bar = 0.5 um. 165 FIGURE 5.4. Paired fluorescence and phase micrographs in which the distributions of actin (panels a,a'), myosin II (panels b,b') and vinculin (panels c,c') in fixed frozen sections of rooster seminiferous epithelium are indicated. In each micrograph, the asterisks indicate the position of the tubule wall. Large arrowheads in the phase micrographs indicate the anterior ends of spermatid heads. Ectoplasmic specialization-like junctions associated with the anterior ends of elongate spermatids are clearly labelled (medium sized arrowheads) in tissue treated with the actin probe (panel a). Also labelled are focal sites near the base (small arrowheads) and at the apex (arrows) of the epithelium. The myosin II probe labels ectoplasmic specialization-like junctions (arrowheads) and sites at the apex of the epithelium (arrows). Basal staining is more diffuse than that seen with the actin probe. Regions associated with the anterior ends of spermatids, known to be sites at which ectoplasmic specialization-like junctions occur, react with the affinity purified vinculin antibody (arrowhead in panel c). Cells in the tubule wall, presumably peritubular cells, react strongly with all three probes (asterisks). X1,063. Bar = 10 um. 166 167 FIGURE 5.5. Controls for NBD-phallacidin staining of fixed frozen sections of rooster seminiferous tubules. Positive staining with the probe for actin filaments is shown in panels (a,a'). Staining occurs in ectoplasmic specialization-like junctions attached to the anterior ends of spermatid heads (large arrowheads), at focal sites in basal regions of the epithelium (small arrowheads), and in linear streaks at the apex of the epithelium (arrows). Cells in the tubule wall also react with the actin probe (asterisks). Similar staining does not occur when sections are incubated with NBD-phallacidin in the presence of phalloidin (panels b,b'), with phalloidin alone (panels c,c'), or with buffer alone (panels d,d'); however, there does occur some diffuse or "granular" autofluorescence (arrowheads). This autofluorescence is often present in the region of spermatid clusters but it is distinct from the linear bands of specific fluorescence seen at the anterior ends of spermatid heads after staining with the actin probe. X965. Bar = 10 um. 169 FIGURE 5.6. Controls for myosin II localization, by immunofluorescence, in fixed frozen sections of rooster seminiferous tubules. Positive staining with the myosin II probe is shown in panels (a,a'). Notice that regions adjacent to the anterior ends of spermatid heads, known to be sites at which ectoplasmic specialization-like junctions occur, react strongly with the antiserum (arrowhead). There also appears to be some diffuse staining in basal regions of the epithelium. Cells in the tubule wall also are strongly reactive with the probe (asterisk). No specific staining is present in sections incubated with preimmune serum (panels b,b'), nor in sections from treatment regimens in which the primary (panels c,c') or both the primary and secondary antibodies have been replaced by buffer alone (panels d,d'). X800. Bar = 10 urn. 171 FIGURE 5.7. Controls for vinculin localization, by immunofluorescence, in fixed frozen sections of rooster seminiferous tubules. Positive staining with affinity purified antibodies againist human platelet vinculin is shown in panels (a,a'). Regions adjacent to the anterior ends of elongate spermatids are reactive with the probe (arrowheads), as are cells in the tubule wall (asterisk). Preincubation of antibodies with antigen and treatment of tissue sections with antibodies in the presence of antigen results in the elimination of staining in the epithelium and dramatically reduces staining in the tubule wall (asterisk) (panels b,b'). No specific staining occurs in sections in which the antibody is replaced by the non-adsorbed fraction from the affinity column (panels c,c'), the primary antibody is deleted from the protocol (panels d,d'), or both the primary and the secondary antibodies are omitted from the protocol (panels e,e'). The large arrowheads in the phase micrographs indicate the positions of clusters of elongate spermatids. X782. Bar = 10 um. 173 FIGURE 5.8. Paired fluorescence and phase micrographs of actin (panels a,a') and myosin II (panels b,b') distribution in fixed frozen sections of alligator seminiferous epithelium. NBD-phallacidin strongly labels ectoplasmic specialization-like junctions at sites of attachment to apically situated spermatids (arrowhead in panels a,a'). The myosin II probe (panels b,b') stains similar sites around spermatid heads but the fluorescence pattern is more diffuse (arrowheads). In each micrograph, the asterisks indicate the location of the tubule wall. Both the actin probe and the myosin II probe strongly label elements of the tubule wall, presumably peritubular cells and possibly basal regions of Sertoli cells. X620. Bar = 10 urn. 175 FIGURE 5.9. Paired fluorescence and phase micrographs of actin (panels a,a') and myosin II (panels b,b' and c,c') distribution in fixed frozen sections of turtle seminiferous epithelium. NBD-phallacidin labels ectoplasmic specialization-like junctions (arrowheads in panel a) at sites of attachment to apically situated spermatids. Similar sites stain with the myosin II probe (panels b,b'). The spermatids shown in panel (b) are at an earlier stage of spermiogenesis than those shown in a. The probe for myosin II reacts in "V-shaped" patterns adjacent to the ends of the spermatid heads (arrowheads). This is particularly evident when images similar to that shown in panel b are magnified (panel c) (arrowheads indicate anti-myosin II staining). The V-shaped patterns obtained with the myosin II probe and shown here are similar to those obtained with the actin probe and indicated by small arrowheads in Fig. 5.12a. Asterisks indicate the tubule wall. Cells in the tubule wall react positively both with NBD-pallacidin and with the myosin II probe. Panels (a,a') X695; Panels (b,b') X933; Panels (c,c') X1,825. Bars = 10 um. 176 © . © I 0 177 FIGURE 5.10. Controls for myosin II localization, by immunofluorescence, in fixed frozen sections of turtle seminiferous tubules. Shown in panels (a,a') is specific staining with antisera raised against human platelet myosin II. Regions adjacent to spermatid heads and known to contain ectoplasmic specialization-like junctions react positively with the myosin II probe (arrowhead). Also reactive are cells in the tubule wall (asterisk). No speicific staining is apparent when the preimmune serum is substituted for immune serum (panels b,b'), when the primary antiserum is omitted from the protocol (panels c,c'), or when both the primary and secondary antibodies are omitted from the protocol (panels d,d'). X1,192. Bar = 10 um. 179 FIGURE 5.11. Shown here are paired phase and fluorescence micrographs of spermatids and attached Sertoli cell regions that together have been mechanically dissociated from fixed turtle testis. The cells in panels (a,a') have been labelled with rhodamine phalloidin. Actin, presumably in an ectoplasmic specialization-like junction, is clearly visible (arrowhead) at the anterior end of the spermatid in the lower third of the micrographs. A small amount of staining is present in anterior regions of the middle spermatid while there is no staining associated with the upper cell. The cells in panels (b,b') have been labelled with a probe for non-muscle myosin II. Specific staining has the same pattern of distribution as cells labelled with rhodamine phalloidin. Positive staining, presumably in an ectoplasmic specialization-like junction, is indicated by the arrowhead. Panels (a,a') X2,144. Bar = 10 um. Panels (b,b') X2.185. Bar = 10 um. ISO 181 FIGURE 5.12. Fragments of turtle seminiferous epithelium that have been fixed and labelled with NBD-phallacidin. The fragments of epithelium were obtained by mincing decapsulated testes in PBS containing EDTA. In reasonably intact segments of the epithelium, such as that shown in panels (a,a'), ectoplasmic specialization-like junctions associated with clusters of elongate spermatids (large arrowheads) stain intensely and generally appear as linear bands of fluorescence that lie adjacent to and extend from the anterior ends of the spermatid heads. Fluorescence associated with early elongate spermatids (small arrowheads) and presumably emitted by labelled actin in forming ectoplasmic specialization-like junctions is not as intense as at later stages and often occurs in "V-shaped" patterns that cap the ends of the spermatid heads. Staining amongst cells at the base of the epithelium is diffuse in the image shown in panel (a) (small arrows). This staining pattern is much different than that which occurs in mammals where an intense signal is emitted from actin in ectoplasmic specializations associated with junctions between adjacent Sertoli cells (see Vogl and Soucy, 1 9 8 5 ) . These basally situated ectoplasmic specializations do not occur in turtles. Shown in panels (b,b') is an elongate spermatid that has been mechanically separated from the epithelium. The spermatid head is outlined by the white dots. Actin in an ectoplasmic specialization-like junction that has remained attached to the anterior end of the head is clearly labelled by the fluorescent probe (arrowhead). Panels (a,a') X1.030. Panels (b,b') X1.920. Bars = 10 um. 183 FIGURE 5.13. Evidence that turtle ectoplasmic specialization-like junctions can be induced to contract. In the experiments summarized here, clusters of glycerinated ectoplasmic specialization-like junctions were incubated in control buffer (panel a) or in control buffer containing ATP (panel b), ATP + C a + + (data not shown), or C a + + (data not shown). In material treated either with ATP alone or with ATP in the presence of C a + + , the diameters of individual junctions (arrowheads) were smaller than in tissue incubated in control buffer or in control buffer containing C a + + (see Fig. 5.14). The overall size of clusters also appeared reduced in samples treated with ATP or ATP + C a + + when compared to controls; however, we did not analyze this apparent difference statistically. X1,920. Bar = 10 um. 185 FIGURE 5.14. Results of two independent contraction experiments in which glycerinated ectoplasmic specialization-like junctions of turtles were treated with buffer, buffer + C a + + , buffer + ATP, and buffer + ATP + C a + + . The V-axis indicates diameter values in pixel units (measured on a Kontron Image Analysis System). The magnification of prints used to measure diameters was somewhat less in experiment (a) than in experiment (b). The mean value and standard deviation are indicated for each treatment. In both experiments, ANOVA demonstrated a significant effect of treatments (experiment a, F(3,36)=16.43, P<0.001; experiment b, F(3,36)=41.45, P<0.001). Tukey's post-hoc analysis indicated that mean diameters of buffer treated junctions were not significantly different from those of similar structures incubated with buffer containing calcium (experiment a, P=0.677; experiment b, P=0.999), but were significantly different from the mean diameters of junctions treated with buffer containing ATP (experiment a, P<0.001; experiment b, P<0.001) and buffer containing ATP + C a + + (experiment a, P<0.001; experiment b, P<0.001). Similarly, mean diameters of ectoplasmic specialization-like junctions treated with buffer containing calcium were significantly different from those of the structures treated with buffer containing ATP (experiment a, P=0.001; experiment b, P<0.001) and buffer containing ATP + C a + + (experiment a, P=0.001; experiment b, P<0.001). Mean diameters of ectoplasmic specialization-like junctions treated in buffer containing ATP were not significantly different from those incubated in buffer containing ATP + C a + + (experiment a, P=0.994; experiment b, P=0.975). 136 M Buffer m C a + + m ATP E2 A T P + C a + + 40.0-37.5-35.0-187 CHAPTER 6 Actin-Related Adhesion Junctions in Sertoli Cells of a Marsupial, the Virginia Opossum (Didelphus virginiana) 188 I n t r o d u c t i o n Ectoplasmic specializations of mammalian Sertoli cells are unusual actin filament-containing complexes found at certain sites of intercellular attachment (Brokelmann, 1963; Flickinger and Fawcett, 1967; Nicander, 1967; Dym and Fawcett, 1970; Russell, 1977a,b). They are made up of the Sertoli cell plasma membrane at the attachment site, a layer of actin filaments and a deeper saccule of endoplasmic reticulum (reviewed by Vogl et al., 1991 and Vogl et al., 1993). In eutherian mammals (placental mammals), the actin filaments are highly ordered and are arranged into hexagonally packed (Dym and Fawcett, 1970; Russell, 1977a,b; Franke et al., 1978), unipolar (Toyama, 1976; Vogl et al., 1983), non-contractile (Vogl and Soucy, 1985) bundles. Ectoplasmic specializations are found at sites of attachment to spermatid heads and at the level of the inter-Sertoli cell tight junctions. Functionally, ectoplasmic specializations are suspected to be involved with intercellular adhesion (Grove and Vogl, 1989; Grove et al., 1990; reviewed by Vogl et al., 1991a,b and Vogl et al., 1993). As shown previously in this thesis (Chapters 3-5), homologues of eutherian mammal ectoplasmic specializations do exist in non-mammalian vertebrates but, when present, they differ considerably from their eutherian counterparts (see also: Osman et al., 1980; Baccetti et al., 1983; Stanely and Lambert, 1985; Sprando and Russell, 1987a). For example, certain key features of non-mammalian vertebrate ectoplasmic specialization-like junctions more closely resemble the actin-associated adhesion junctions of other cell types than do eutherian ectoplasmic specializations. One such feature is related to the potential for contraction. Unlike ectoplasmic specializations of eutherian mammals which are not contractile (Vogl and Soucy, 1985), the junctions present in non-mammalian vertebrate Sertoli cells do appear to have contractile properties. The actin filaments of the sites are arranged into loose bundles or networks, 189 the protein myosin II is present, and the sites can be induced to contract in vitro when exposed to the appropriate conditions (see Chapter 5). Given that significant differences exist between the Sertoli cell ectoplasmic specializations of eutherian mammals and the structural equivalents of non-mammalian vertebrate Sertoli cells, it is of considerable interest to determine the condition of these junctions in non-placental mammals. Non-placental mammals such as the opossum, a representative of the marsupials (Metatheria), are in many ways regarded as "primitive" mammals (Romer, 1970). Data on the Sertoli cell junctions of such non-placental mammals may not only provide information on the strategies for intercellular adhesion which have evolved within this group but may also help shed further light on what the precursor of the unusual eutherian mammal ectoplasmic specialization may have been like. Marsupials (Metatheria) represent an early group of mammals which is thought to have diverged from the main line that lead to the more progressive group of placental mammals (Eutheria) (Romer, 1970). Currently, little is known about the Sertoli cell junctions of marsupial species. Preliminary ultrastructural observations on select species of marsupials indicate that ectoplasmic specialization-like junctions are formed by apically by Sertoli cells around spermatid heads (bandicoot - Sapsford et al., 1 969; opossums- Phillips, 1970; Rattner, 1972; Russell and Malone, 1980). Structurally these junctions appear to resemble those found in non-mammalian vertebrates. No descriptions of inter-Sertoli cell ectoplasmic specializations in marsupials are available in the literature. In this chapter, I examine the actin-related junctions formed by Sertoli cells of an American species of opossum, Didelphus virginiana, a species considered characteristic of the marsupial group (Romer, 1970). I present evidence that Sertoli cells of the opossum form ectoplasmic specialization-like junctions both apically with spermatids and basally with adjacent Sertoli cells. Moreover, I show that ectoplasmic 190 specialization-like junctions in this species share features in common with those found in eutherian mammals as well as with those found in non-mammalian vertebrate species. The fully assembled Sertoli cell/spermatid junctions in this species closely resemble the actin-related junctions found in non-mammalian vertebrates whereas the basally located inter-Sertoli cell junctions are more similar to the ectoplasmic specializations of eutherian mammals. Materials and Methods 191 Animals: One adult male Virginia opossum {Didelphis virginiana) was live-trapped near Shreveport, Louisiana, and transported to the Department of Cell Biology and Anatomy at the Louisiana State University Medical Center where tissue was processed the same day. Three additional adult animals were obtained from Northeastern Wildlife, South Plymouth, New York, and maintained for up to ten days in the Department of Anatomy at the University of British Columbia. These latter animals were exposed to a 12hr:12hr light:dark cycle and given free access to food and water. All animals were anesthetized with a ketamine-HCI (30 mg/kg)/rompum (6 mg/kg) mixture administered intramuscularly. Electron Microscopy: Testes from two opossums were processed for electron microscopy. Tissue from one animal was first perfused with fixative and then further fixed by immersion whereas tissue from the second animal was simply fixed by immersion. For the perfusion, the animal was anesthetized and its heart exposed. The left ventricle was cut and a canula inserted and tied into place via a ligature placed through the transverse sinus. The inferior vena cava was next cut and the perfusion begun. The animal was perfused briefly (5-10 sec) with PBS (see Chapter 2) containing heparin and procain-HCI before being perfused with EM fixative (see Chapter 2) for 10 min. Following this, the testes were removed, cut into small blocks (1 m m 3 ) in fixative and then allowed to fix by immersion for an additional 2 hrs. Testes from the second anesthetized animal were simply removed and placed immediately in EM fixative in which they were cut into 192 small blocks (1 mm 3 ) . These blocks were allowed to immersion fix for 2.5 hrs. Further processing of tissue from both animals for electron microscopy was carried out as previously described (see Chaper 2). Thick sections were stained with 1% toluidene blue and photographed on a Zeiss Axiophot microscope. Thin sections were stained with uranyl acetate and lead citrate and the photographed on a Philips EM 300 operated at 8 0 kV. Chemicals and other compounds used for this set of experiments and all others in this study were obtained from Sigma Chemical Co. (St. Louis, MO), unless otherwise noted. Fluorescence Microscopy: (a) Tissue Preparation: Testes from three opossums were processed for fluorescence microscopy. Tissue from two of the animals was perfused with fixative. The perfusion was conducted in the same manner described above with the only difference being that a fluorescence fixative (see Chapter 2) was used in place of the EM fixative. Following a 10 min perfusion with fixative, the tissue was washed with PBS by perfusion for 15 min and then cut into blocks (1 c m 3 ) . Testes from the third anesthetized animal were simply removed and placed immediately in fluorescence fixative. They were then cut into blocks (1 c m 3 ) , allowed to fix by immersion for 30 min., and finally washed in PBS three times for 1 0 minutes each. Washed tissue from all three animals was then embedded in Tissue Tek OCT compound (Miles, Inc.) and frozen and sectioned as described in Chapter 2. (b) Localization of Actin and Myosin II: Localization of filamentous actin in cryosections by immunofluorescence was accomplished using rhodamine phalloidin. The labelling protocol followed for this probe 193 has been previously described (see Chapter 2). Controls for rhodamine phalloidin included (1) PBS (containing 0.1% BSA) + 1.65 x 10-6 M rhodamine phalloidin + 7.76 x 10-4 M phallacidin (competitive specificity control); (2) PBS (containing 0.1% BSA) + 7.76 x 10-4 M phallacidin (control for phallacidin in reagent 1); ( 3 ) PBS (containing 0.1% BSA) (control for autofluorescence). For the localization of myosin II in cryosections of opossum testis, I used two different myosin antibodies. The first was a rabbit antiserum raised against human platelet myosin II. It was kindly provided by Dr. Kegi Fujiwara (National Cardiovascular Center Research Institute, Osaka, Japan). The labelling protocol and controls run for this probe were described in Chapter 2. The second myosin antibody was a commercially available (Sigma Chemical Co.) rabbit antiserum raised against bovine uterus myosin 11 (whole molecule). It was used at a dilution of 1:10. Controls for this antibody included: (1) substitution of the primary antiserum with normal rabbit serum; (2) substitution of the primary antiserum with buffer; (3) substitution both of the primary and of the secondary antibodies with buffer. All sections processed for fluorescence were examined on a Zeiss Axiophot photomicroscope fitted with filter sets for detecting rhodamine and fluorescein isothiocyanate. 194 R e s u l t s Ultrastructure: Sertoli cell ectoplasmic specialization-like junctions, consisting of the Sertoli cell plasma membrane, a layer of microfilaments and underlying saccules of endoplasmic reticulum, occur at two different locations within the seminiferous epithelium of the opossum: 1) Apically apposed to the head regions of elongating spermatids, and 2 ) Basally between neighboring Sertoli cells. a) Junctions Formed Apically with Spermatids: During the course of spermiogenesis in opossum species, the round nuclei of early spermatids undergo an unusual shape transformation to produce the complex "U" form present in late spermatids and spermatozoa (Von Holstein, 1967; Orsi and Ferreira, 1978). Part of this nuclear shaping process involves a progressive flattening of the nucleus in an anterior-posterior direction, i.e. along the perpendicular axis of the spermatid. As a consequence of this flattening, the portion of the spermatid head which lies in direct contact with the Sertoli cell appears very blunt when viewed at the ultrastructural level (Fig. 6.1a). By the mid-stages of spermiogenesis, a region of filamentous material representing the developing junction is visible on the Sertoli cell side of this contact site. The filamentous region is restricted to the area adjacent to the Sertoli cell plasma membrane. Also present are cisternae of endoplasmic reticulum. These are related to the lateral edges of the filamentous area but extend for a considerable distance away from the filaments down into the apical Sertoli cell cytoplasm (Fig. 6.1a). These cisternae appear to outline an organelle-free zone below the spermatid head. When viewed at higher 195 magnification, the filamentous region is seen be made up of two different populations of Sertoli cell microfilaments: 1) A zone of filaments that is related to the contact site with the spermatid head, and 2) A laterally positioned, narrow band of filaments that flanks the inner surface of the endoplasmic reticulum. The second population of filaments is not directly associated with the contact site (Fig. 6.1b). Significantly, the two filament populations differ from one another in their organization (Fig. 6.1c). The filament zone applied to the attachment site with the spermatid head is made up of a loosely organized network of microfilaments. This network makes up the bulk of the filaments present. The narrow band of filaments, by contrast, consists of more ordered microfilaments which run parallel to one another and parallel to the flattened surface of the spermatid head (Fig. 6.1c,d). When the plane of section passes longitudinally through the attachment site, these filaments are seen in cross-section (Fig. 6.1c,d) as a band of ordered filaments on either side of the loosely organized network of filaments. In more grazing sections, the filaments of each band appear to stream towards each other (data not shown). This suggests that they form a filamentous ring around the periphery of the loosely organized network of filaments. These ordered filaments follow the inner surface of the endoplasmic reticulum for a short distance away from the spermatid head but cb not appear to line the entire inner surface of the endoplasmic reticulum in this area. By later stages of spermiogenesis, the appearance of the Sertoli cell/spermatid attachment site has changed dramatically. The microfilament zone has expanded considerably and is now present as a cone of microfilaments projecting from the contact site with the spermatid head down toward the base of the Sertoli cell (Fig. 6.2a). It is comprised entirely of a loosely organized network of microfilaments (Fig. 6.2a,b). The narrow band of highly ordered filaments seen in association with earlier stages is no longer present. By this later stage, the entire network of microfilaments is outlined by a layer of endoplasmic reticulum. Interestingly, the region now occupied by this filament network appears to correspond to the organelle-free zone which the endoplasmic 196 reticulum outlined earlier (compare the microfilament zone of the junction in Fig. 6.2a with the area outlined by endoplasmic reticulum in Fig. 6.1a). b) Junctions Formed Basally between Sertoli Cells: The lateral borders of Sertoli cells follow irregular courses. Well-developed microfilament junctions, which by all accounts appear to be ectoplasmic specializations, are formed basally along these borders between adjacent Sertoli cells. These junctions begin close to the basal lamina and consist of a layer of microfilaments closely apposed to the plasma membrane. A deeper layer of endoplasmic reticulum is consistently seen in conjunction with the filament layer (Fig. 6.2c). The microfilaments show some degree of ordering in that they are generally oriented parallel to one another and parallel to the Sertoli cell plasma membrane (Fig. 6.2d). Localization of Actin and Myosin II: Actin: In fixed frozen sections of opossum testis, the filamentous actin probe rhodamine phalloidin reacted positively at three locations within the seminiferous epithelium: 1 ) In regions associated with the apically positioned spermatid heads, 2) In regions of the epithelium immediately next to the tubule lumen, and 3) In regions near the base of the epithelium. Specific and strong labelling with the probe also occurred outside of the epithelium in the tubule wall. The fluorescence pattern emitted from regions near spermatid heads differed depending on the stage of spermatid present. Around mid-stage spermatids, identified in phase-contrast by the fact that the spermatid heads were beginning to flatten, a weak signal was detected over the head region. This was 197 surrounded by a thin peripheral ring of stronger fluorescence (Fig. 6.3a,a') which may correspond with the location where the narrow band of highly ordered filaments was seen ultrastructurally. Around later staged spermatids, identified in phase-contrast by their distinct "U" shaped nuclei, intense fluorescence occurred immediately next to the flattened anterior surface of the spermatid heads (Fig. 6.3b,b'). The pattern of this signal was in the form of a cone or block of fluorescence next to each spermatid head and appeared to correspond to the location where the elaborate microfilament network was observed at the ultrastructural level. The fluorescence near the tubule lumen and near the base of the epithelium did not appear to vary at different stages of spermatogenesis. The fluorescence near the lumen formed a punctate pattern (Fig. 6.3b,b'). It was not possible to determine if it was emitted by Sertoli cells or spermatogenic cells. The labelling near the base of the epithelium was intense and present as linear bands (Figs. 6.3a,a',b,b') which presumably represent the basal ectoplasmic specialization-like junctions seen at the ultrastructural level. This basal pattern of staining was similar to that which characterizes labelled ectoplasmic specializations in mammals (Vogl and Soucy, 1985). In more grazing sections through the basal regions of the epithelium (Fig. 6.3c,c'), these bands of fluorescence were seen to be continuous with each other around the margins of Sertoli cells and formed a honey-comb pattern. No specific fluorescence, as described above, was detected in any of the control slides for rhodamine phalloidin staining (data not shown). Myosin II: In this set of experiments I incubated fixed frozen sections of opossum testis with two different antibodies raised against the myosin II protein. Staining with each probe within the epithelium was inconclusive. All control slides for the myosin probes were negative (data not shown). 198 D i s c u s s i o n In this chapter I present evidence that Sertoli cells of a marsupial, the Virginia opossum (Didelphus virginiana), form ectoplasmic specialization-like junctions apically with spermatids as well as basally with neighboring Sertoli cells. The junctions share features in common with those found in Sertoli cells of eutherian mammals (placental mammals) and non-mammalian vertebrate species. In eutherian mammals, ectoplasmic specializations represent a unique Sertoli cell actin filament complex related to intercellular attachment sites. They consist of a layer of hexagonally packed actin filaments along with the adjacent plasma membrane at the attachment site. A layer of endoplasmic reticulum is linked to the cytoplasmic face of the actin filament layer. They are formed at two specific locations within the eutherian Sertoli cell: apically at sites of attachment to spermatids and basally at the level of the inter-Sertoli cell tight junctions ("blood-testis barrier") (Brokelmann, 1 963; Flickinger and Fawcett, 1967; Nicander, 1967; Dym and Fawcett, 1970; Russell, 1977a,b). In eutherian mammals, ectoplasmic specializations are not contractile (Vogl and Soucy, 1985). Previous studies in marsupial species (Sapsford etal., 1969; Phillips, 1 970; Russell and Malone, 1980) have noted the presence of Sertoli cell ectoplasmic specialization-like junctions formed apically with maturing spermatids. My results in the opossum confirm that these apical actin-related junctions, when fully assembled, consist of an prominent extension of microfilaments from the attachment site down into the apical Sertoli cell cytoplasm. This filamentous zone is bounded by a layer of endoplasmic reticulum. My results confirm that the main filament type present is actin, like at ectoplasmic specializations of eutherian mammals (Toyama,1976; Franke et al. , 1978; Vogl et al., 1983, 1985, and 1986; Vogl and Soucy, 1985; Suarez-Quian and Dym, 1988). Although marsupials are closely related to eutherian mammals, the 199 overall structure of these apical ectoplasmic specialization-like junctions much more closely resembles those present in non-mammalian vertebrates. Typically in non-mammalian vertebrates, the actin filaments of the apical ectoplasmic specialization-like junctions extend down and away from the attachment site as a loose network or bundle of contractile filaments (see Chapters 3-5). In this study, I noted that the filament organization at the fully assembed apical junction is also primarily organized as a loose network of microfilaments. Inter-Sertoli cell actin-related junctions have not previously been reported in a marsupial species. Here I found well-developed ectoplasmic specialization-like junctions are present between adjacent Sertoli cells. These inter-Sertoli cell junctions are formed basally between neighboring Sertoli cells, as are those of eutherian mammals. Interestingly, these basal junctions exhibit a more ordered actin filament pattern than is seen at the apical Sertoli cell/spermatid sites. The filaments occur in a layer in which the filaments are oriented parallel to one another. By appearance the filament organization resembles that seen at the zonulae adherens of other epithelial cell types (Hirokawa and Tilney, 1982; Hirokawa et al., 1983). A hexagonal packing of the filaments, as occurs in the unipolar, non-contractile filament arrays of eutherian mammal ectoplasmic specializations, is not obvious. Nevertheless, that the layer of actin filaments shows some ordering and is positioned basally between the Sertoli cell plasma membrane and an underlying cistern of endoplasmic reticulum is reminiscent of the eutherian condition. When visualized with rhodamine phalloidin, particularly in tangential sections, these inter-Sertoli cell ectoplasmic specialization-like junctions appear to form continuous belts around the perimeters of Sertoli cells. This, too, closely resembles the eutherian condition (Vogl and Soucy, 1985; Vogl et al., 1985). In no other amniote class has a distinct and continuous "ectoplasmic specialization" been identified around the perimeters of Sertoli cells. 2 0 0 Given the filament organizations of the apical and basal actin-related junctions in the opossum, an organization that suggests a contractile nature, one might predict that the protein myosin II would be present at one or both sites. In this study, however, labelling experiments for myosin II were inconclusive at both sites. One of the particularly interesting findings in this study is related to the assembly of the apical Sertoli cell/spermatid junctions in this species. The fully developed junction, consisting of an elongated network of actin filaments circumscribed by endoplasmic reticulum, is preceded by an earlier form in which two distinct populations of actin filaments are present. One of these early filament groups is organized as a loose network and is directly associated with the Sertoli cell attachment site to the spermatid head. Its filament organization closely resembles that of the later, much larger filament network of the fully developed junction. It is probable that a simple expansion of this early filament network is what gives rise to the elongated filament network seen later. Functionally, this loose network of actin filaments is likely involved with intercellular adhesion as occurs at actin-related adhesion junctions in other systems. It is the second early filament group which is more noteworthy. It is organized very differently from the first. It consists of a narrow band of highly ordered actin filaments that appear to be arranged as a ring around the outside of the early filament network. This filament band lines the undersurface of a portion of endoplasmic reticulum that extends away from the contact site with the spermatid head. Significantly, the filament ordering and their relationship to elements of the endoplasmic reticulum very closely resemble that seen at the ectoplasmic specializations of eutherian mammals. When the filaments are viewed in cross-section, they are seen to be oriented parallel to each other in a manner similar to the tightly cross-linked, non-contractile filament arrays of eutherian ectoplasmic specializations. Although this filament band makes up a minor component of the overall filaments present, it is noteworthy in that i t 201 represents the first point along the vertebrate line where this "eutherian" filament organization is found at Sertoli cell/spermatid ectoplasmic specializations. In all non-mammalian vertebrate classes the actin filaments associated with ectoplasmic specialization-like junctions appear to be organized entirely into contractile networks or bundles (see Chapters 3-5). Regarding the functional significance of these highly ordered filaments in the opossum, it would seem they may play a different role at the attachment site than do the actin filaments of eutherian mammal ectoplasmic specializations. In eutherians, the highly ordered filaments of the ectoplasmic specialization follow the contours of the attachment site closely and are thought to assume a role in intercellular adhesion (Grove and Vogl, 1989; Grove et al., 1990; reviewed by Vogl et al., 1991a,b and Vogl et al. , 1993). In the opossum, however, it would seem unlikely that these ordered filaments are directly involved with intercellular adhesion for the following reasons: First, they are spacially unrelated to the contact site. The narrow band in which they occur extends directly away from the contact site and down into the Sertoli cell cytoplasm. Secondly, this band of highly ordered filaments appears to be a transient feature of the apical junctions in this species. It is lost by the point the expanded filament network of the site is fully established. These facts point to a different primary role for these ordered filaments. It is possible that their presence may in some way be tied in with the progressive downward expansion of the loose filament network from the contact site. This is suggested by the observations that they extend partially down into the area where network will later expand and, as mentioned, they are lost once the loose network of filaments is fully established. Alternatively, these ordered filaments may offer further structural support to the attachment site and thereby play an indirect role in intercellular adhesion. Closely related to the outer surface of the ordered filament band are cisternae of endoplasmic reticulum. Of possible significance is the fact that this endoplasmic 2 0 2 reticulum extends as a line well beyond the filaments and outlines a cone-shaped, organelle-free zone below the spermatid head. Interestingly, this outlined zone appears to correspond to the area into which the loose filament network will later expand. This suggests that the endoplasmic reticulum may be laid down first and that this is followed by an expansion of the filament network out to it. Although there is no direct evidence to support this, these observations suggest that the band of highly ordered filaments and/or the endoplasmic reticulum may in some way be involved with establishing the future elongated filament network, perhaps by defining where it will lie or by playing a more direct role in its expansion. In summary, I present evidence that well-developed ectoplasmic specialization-like junctions are formed by Sertoli cells in the opossum. At the apically located Sertoli cell/spermatid sites, the overall structure and filament arrangement is similar to those found in non-mammalian vertebrates. Significantly, a small transient group of ordered filaments is seen at these apical sites. The filaments at the basally located inter-Sertoli cell junctions are more ordered and occur in loose bundles similar to the zonulae adherens of other cell types. The overall structure of these basal sites more closely resembles that of the ectoplasmic specializations of eutherian mammals. Taken together, these observations indicate that Sertoli ectoplasmic specializations of the opossum share features in common with those found both in eutherian mammals and non-mammalian vertebrates and, in many ways, represent an "intermediate" between the two different organizations. 203 FIGURE 6.1. Electron micrographs of developing Sertoli cell ectoplasmic specialization-like junctions associated with spermatids at the mid-stage of spermiogenesis in the opossum. At low magnification, as shown in panel (a), the flattened shape of the spermatid head (asterisk) is clearly evident. The developing junction is apparent as a region of filamentous material (arrows) on the Sertoli cell side of the contact site. Note that a line of endoplasmic reticulum (arrowheads) extends down from the filamentous region and appears to outline an organelle-free zone in the apical Sertoli cell cytoplasm. In panel (b) the Sertoli cell filamentous region is seen to be made up of two different groups of microfilaments. One group (asterisk) is closely related to the contact site with the spermatid head while the other group (curved arrows) extends away from the contact site as a narrow band along the inner surface of the endoplasmic reticulum (arrowheads). Panel (c) shows a similar image to panel (b), but at higher power. Note that the organization of the two microfilament groups differs. The filaments related to the attachment site to the spermatid head are organized as a loose filamentous network (asterisk). The filaments that extend away from the contact site (curved arrows) on the inner surface of the endoplasmic reticulum (arrowheads) are highly ordered. The microfilaments of this latter group are oriented parallel to each other and parallel to the flattened surface to the spermatid head (panel d). Panel (a) X9.100. Bar = 2 urn. Panel (b) X21,980. Bar = 1 um. Panel (c) X80.215. Bar = 0.25 um. Panel (d) X11 6,380. Bar = 0.1 um. 205 FIGURE 6.2. Electron micrographs of Sertoli cell ectoplasmic specialization-like junctions in the opossum. Each ectoplasmic specialization-like junction associated with a later staged spermatid consists of a greatly expanded microfilament zone (arrowheads) in the apical Sertoli cell cytoplasm (panel a). As shown in panel (b), the microfilaments form a loosely organized network (asterisk) which is circumscribed by endoplasmic reticulum (arrowheads). The band of highly ordered microfilaments present at earlier stages is now absent. Panel (c) shows the structure of the basally located junction formed between neighboring Sertoli cells. It consist of a layer of microfilaments (curved arrows) positioned between the plasma membrane and underlying cisternae of endoplasmic reticulum (asterisks). As seen in panel (d), the microfilaments (arrowheads) related to these basal junctions appear to course parallel to each other and parallel to the plasma membrane. Panel (a) X19,600. Bar = 1 um. Panel (b) X40.590. Bar = 0.5 um. Panel (c) X71.400. Bar = 0.1 um. Panel (d) X71.430. Bar = 0.1 um. %0Q> 207 FIGURE 6.3. Matched fluorescence and phase micrographs in which the distribution of actin in fixed frozen sections of opossum seminiferous epithelium are shown. In each micrograph, the large asterisks indicate the position of the tubule wall. In tissue labelled with rhodamine phalloidin, positive staining occurs in regions associated with spermatid heads. Around mid-stage spermatids (panel a,a'), diffuse fluorescence (small asterisks) is visible over the region of the spermatid head. This is surrounded by a more intense peripheral band of staining (small arrows). Near later stage spermatids (panel b,b'), intense blocks of fluorescence (arrows) are seen associated with the flattened anterior surface of the spermatid heads. Also labelled at both stages are focal sites near the tubule lumen (small arrowheads in panel b) as well as intense linear bands near the base of the epithelium (large arrowheads in panels a and b). The latter presumably represent the basally located inter-Sertoli cell ectoplasmic specialization-like junctions. In more grazing sections through the epithelium (panel c,c'), the basal linear bands (arrowheads) appear to link up with one another and form a honey-comb pattern. Cells in the tubule walls label intensely with the actin probe at all stages. X700. Bar = 20 um. 209 CHAPTER 7 Actin Filaments Associated with the Basal Sertoli Cell Surface in the Alligator and Turtle 210 I n t r o d u c t i o n Sertoli cells occur in seminiferous tubules of the testis where, along with a population of germ cells, they form the seminiferous epithelium. As with most other types of epithelial cells, Sertoli cells are situated on and anchored to a basal lamina. In many cells, attachment to the substratum is accomplished by matrix receptors. The cytoplasmic domains of these receptors are thought to be indirectly linked to elements of the cytoskeleton, most often to actin filaments or to intermediate filaments. Their extracellular domains bind the cell directly to specific components of the extracellular matrix. These transmembrane receptors commonly belong to the integrin family of molecules (Hynes, 1987; Buck and Horwitz, 1987; Albelda and Buck, 1990). The types of integrins present at cell/substratum junctions appear to vary depending on the filament type associated with the contact site (Carter etal., 1990a,b; Sonnenberg et al., 1990; Stepp et al., 1990). Often these receptors are clustered into morphologically distinct regions of the plasma membrane that are associated with subsurface bundles of filaments. Hemidesmosomes are examples of substrate adhesion junctions that are related to intermediate filaments (Staehelin, 1975; Cowin et al. , 1985; Schwarz et al., 1990) while focal contacts are examples of junctions related to actin filaments (Geiger et al., 1985; Burridge et al., 1988; Geiger, 1989). The mechanisms by which Sertoli cells establish and maintain attachment to the basal lamina are not well understood. In mammals, intermediate filaments of the vimentin type form a loose carpet near the Sertoli cell basal plasma membrane (Vogl et al., 1983; Amlani and Vogl, 1988; Vogl, 1989) and appear to interact with hemidesmosome-like structures (Connell, 1974; Russell, 1977). Actin filaments are present at the base of the cell, but form only a minor component of the cytoskeleton in 211 this region (Vogl et al., 1983; Vogl, 1989). The integral membrane elements involved with cell/substratum adhesion in Sertoli cells have not been identified. In some non-mammalian vertebrates, intermediate filaments are not concentrated at the base of Sertoli cells; rather, actin filaments predominate. This is true in the hagfish (see Chapter 2). Two studies on species from the class Reptilia (tortoise - Unsicker, 1974; lizards from the genus Lacerta - Baccetti et al., 1 9 83) also report that actin filaments are abundant at the base of the Sertoli cell. Furthermore, the basal actin filaments in these reptilian species appear to be organized into discrete bundles. I have studied two additional representative species of the class Reptilia, the American alligator (Alligator mississippiensis) and the red-ear turtle (Psuedemys scripta). In this chapter I present evidence consistent with the conclusion that Sertoli cells of these species possess bundles of actin filaments adjacent to the basal plasma membrane. This suggests that basal actin accumulations in Sertoli cells may be a reasonably widespread feature in the class Reptilia. This is in contrast to the mammalian situation where the majority of filaments present at the base of similar cells are of the intermediate type. I propose that the actin bundles in reptiles may participate in anchoring Sertoli cells to the substratum. 212 Materials and Methods Animals: Two alligators (Alligator mississippiensis) and 4 turtles (Pseudemys scripta) were used in this study. The alligator tissue was obtained from two animals destroyed by Louisiana Department of Wildlife and Fisheries personnel at the Rockefeller Wildlife Refuge (Grand Chenier, Louisiana). Turtles were obtained from Nasco (Lemberger, Wl ) . The turtles were anesthetized with sodium pentobarbitone administered intraperitoneal^. Electron Microscopy: Samples of testes from 2 alligators and 2 turtles were processed for electron microscopy as follows. Testes were removed and cut into small blocks ( m m 3 ) that were immersion fixed for 3 hrs in a fixative containing 1.5% paraformaldehyde (Fisher Scientific, Pittsburgh, PA), 1.5% glutaraldehyde (J.B EM Services, Inc.) and 0.1 M sodium cacodylate (J.B. EM Services, Inc.) (pH 7.3). Following this, blocks of alligator tissue were stored in 0.1 M sodium cacodylate for 4 days, then washed twice in this buffer and post-fixed in cold (ice) 1% osmium tetroxide (J.B. EM Services, Inc.) in 0.1 M sodium cacodylate. After fixation, turtle tissue was imediately washed three times with 0.1 M sodium cacodylate. Tissue from both species was then processed for electron microscopy as previously described (see Chapter 2). Thick (1u.m) sections were stained with toluidene blue. Thin sections were stained with both uranyl acetate and lead citrate and were examined on a Philips 300 electron microscope operated at 80 kV. 213 Tissue Preparation for Localization of Actin: Testes from alligators were excised and cut into blocks of approximately 1 c m 3 in 0.1 M sodium cacodylate containing 3% paraformaldehyde. After immersion fixing in this fixative for 1 hr, the blocks were washed once with 0.1 M sodium cacodylate and stored for 4 days in this buffer. Blocks were then washed twice with 0.1 M sodium cacodylate and frozen at liquid nitrogen temperatures in Tissue-Tek O.C.T. Compound (Miles Inc.). Turtle testes were cut into similar sized blocks and immersion fixed in phosphate buffered saline (PBS) containing 3% paraformaldehyde for 1 hr. Blocks were then washed three times with PBS and frozen in the same manner as were the alligator blocks. All frozen blocks were stored in liquid nitrogen until they were sectioned for fluorescence localization of actin. Fluorescence Localization of Actin: Frozen sections (7-10 um thick) of fixed alligator and turtle testes were collected on polylysine coated slides and immediately plunged into cold ( -20°C) acetone. After 5 min the slides were removed and air dried for approximately 30 min. The subsequent labelling for filamentous actin with the probe rhodamine phalloidin was carried out as previously described (see Chapter 2). Control slides run for rhodamine phalloidin staining included those previously described (see Chapter 2). Following actin labelling, all slides were examined on a Zeiss Axiophot photomicroscope fitted with filter sets for detecting rhodamine. 214 R e s u l t s (a) Ultrastructure: Alligator (Alligator mississippiensis): In the alligator, well-developed filament concentrations occur on or near the Sertoli cell plasma membrane adjacent to the basal lamina. The filaments occur in distinct bundles in which individual elements are uniformly aligned and are loosely clumped together into what resemble small "haystacks" when the bundles are cut in certain planes (Figs. 7.1 & 7.2). These filament bundles often appear to be in contact with the plasma membrane; however, distinct attachment plaques are not obvious. Immediately apical to basal regions containing the filament bundles are perinuclear regions in which intermediate filaments predominate (Fig. 7.2). The perinuclear intermediate filaments are clearly larger in diameter than filaments of the basal bundles (inset of Fig. 7.2). The basal filament bundles appear to form a branching network over the basal surface of the cell as is clearly evident in grazing sections of this region (Fig. 7.3a). Individual bundles are oriented parallel to the basal surface of the cell (Fig. 7.3b). Frequently a cistern of endoplasmic reticulum is associated with the bundles (Figs. 7.3a,b). Ribosomes occur along the cytoplasmic face of the reticulum. Turtle (Pseudemys scripta): Basal concentrations of filaments also occur in turtle Sertoli cells. As in the alligator, the filaments are aligned parallel to the basal plasma membrane (Figs. 7.4a,b). The filaments are of a similar diameter to those present in peritubular cells 215 and are oriented parallel to one another in a loose arrangement. Along certain portions of the basal surface of some cells, the microfilaments appear to be segregated into distinct bundles which are often found within ridge-like processes of the Sertoli cell (Fig. 7.4a). In other regions, this pattern is not observed and the filaments form a continuous layer (Fig. 7.4b). In general, the basal filament population in the turtle is restricted to the area immediately adjacent to the basal plasma membrane. Unlike in the alligator, the bundles, when cut in cross-section, do not appear to extend apically into the cytoplasm. In addition, the bundles are not obviously associated with cisternae of endoplasmic reticulum. (b) Fluorescence Localization of Actin: Alligator: (Alligator mississippiensis): In fixed frozen sections, rhodamine phalloidin strongly labels cells of the tubule wall (Fig. 7.5a). Presumably these are peritubular cells which are known to contain a large amount of actin. In addition, linear tracts of fluorescence occur in regions somewhat removed from the tubule wall in what I am interpreting as basal regions of the seminiferous epithelium. These fluorescent tracts appear as either vertical or horizontal bands depending on plane of section. These tracts occur in a location similar to that of filament aggregations observed in electron micrographs described above. In some regions, where the epithelium has apparently separated away from the tubule wall during tissue preparation, dense bands of fluorescence are clearly associated with basal regions of the epithelium (Fig. 7.5a). I suspect that this fluorescence may be eminating from actin bundles in the Sertoli cell. In control sections (data not shown), the concentration of phallacidin that I used competitively reduced but did not eliminate the fluorescence described above. I did not 216 observe any fluorescence in sections treated with buffer containing phallacidin or with buffer alone. Turtle: (Pseudemys scripts): Fixed frozen sections of turtle seminiferous tubules stained with rhodamine phalloidin (7.5b) display a pattern of labelling similar to that seen in the alligator. Cells of the tubule wall label intensely with the fluorescent phalotoxin. Also present are linear tracts of fluorescence somewhat removed from the tubule wall in regions that I am interpreting as the base of the seminiferous epithelium. 217 > D i s c u s s i o n In this chapter I present evidence that large bundles of actin filaments are present adjacent to the basal plasma membrane of Sertoli cells in the American alligator and the red-ear turtle. This situation is unlike that reported for basal regions of mammalian Sertoli cells where intermediate filaments predominate (Vogl etal. , 1 983; Vogl, 1989; Vogl et al., 1993). Both in alligator and in turtle Sertoli cells, distinct bundles of filaments are related to the plasma membrane in regions where it associates with the basal lamina. The bundles are oriented parallel to the plasma membrane and, at least in the alligator, appear to form a network. These bundles in the alligator are large in girth and, when viewed in cross-section, appear as columnar stacks of filaments. In this species, cisternae of endoplasmic reticulum are related to the bundles. Filament bundles in the turtle are narrower than those of the alligator and are often found in folds of the plasma membrane. Filament bundles in the alligator and turtle are morphologically similar to those reported in Sertoli cells of some other reptilian species. In the tortoise, Testudo graeca, filament bundles are organized in the same fashion as they are in the turtle: that is , filament bundles occur in folds or "foot processes" of the basal plasma membrane (Unsicker, 1974). In two species of lizards, Lacerta muralis and Lacerta sicula, the filaments are described as forming "a thick layer close to the plasma membrane, forming short, dense bundles close to the basal lamina" (Baccetti et al., 1983). These observations, together with mine, indicate that basal filament concentrations may be a general feature of reptilian Sertoli cells. Reptilia may be unique among the non-mammalian vertebrate classes in this sense. To date, well-developed basal filaments have been described in Sertoli cells of only one other class, Agnatha, and even here basal 218 filaments are not a consistent feature between species, i.e. hagfish exhibit them but lamprey do not (Chapter 2). Evidence consistent with the conclusion that the filaments in basal bundles of reptilian Sertoli cells are actin includes the following observations: (a) The diameter (70A) of filaments in the tortoises (Unsicker, 1974) and in lizards (Baccetti et al. , 1983) is within the range of that for actin filaments; (b) The filaments are smaller in diameter than perinuclear intermediate filaments in the alligator; (c) In lizards, the filaments have a 5 nm periodicity, in rotary shadow deep-etch preprarations, that is characteristic for filamentous actin (Baccetti et al., 1983); (d) In the turtle, the filaments are of a similar diameter to those present in peritubular cells which stain intensely with fluorescent phallotoxins; (e) In the alligator and turtle (this chapter), and in lizards (Baccetti et al., 1983), regions associated with the base of the seminiferous epithelium, which correspond to the location of filament bundles seen at the ultrastructural level, label with rhodamine phalloidin. The presence of actin filaments in a layer or in bundles immediately adjacent to the plasma membrane in regions associated with extracellular matrix is not unique to reptilian Sertoli cells. Endothelial cells possess actin filaments in basal regions (Wong etal., 1983; Dreckhahn and Wagner, 1986), as do mesothelial cells (Sugimoto et al. , 1989), fish scleroblasts (Byers and Fujiwara, 1982), and embryonic avian corneal epithelial cells (Sugrue and Hay, 1981; 1982; Hay, 1983). In addition, many mesenchymal cells possess an actin cortex (Hay, 1983; Singer etal., 1984). In other cells, such as in skeletal muscle cells (Shear and Bloch, 1985), in smooth muscle cells (Small and Sobieszek, 1980; Geiger et al., 1981, 1985), and in cells grown in culture (Singer, 1979; Geiger et al., 1984), distinct bundles of actin are related to the plasma membrane in specific regions where the cell is firmly anchored to extracellular matrix. In all of these cells, actin is thought to be linked to matrix receptors, such as integrins, in the plasma membrane and thereby participate in the process by which the cells are 219 firmly attached to the substratum (Geiger et al., 1985; Burridge et al., 1988; Geiger, 1989) . The function of actin filament bundles in the Sertoli cells of reptiles is unknown. Based on analogy with the systems just described, one of several possibilities is that they may play a role in cell/substrate adhesion. If this is the case one would predict that basal actin filament bundles are linked, in some way, to elements in the membrane involved with attachment. Morphologically, the filaments are closely related to the plasma membrane, although submembrane dense plaques are not generally apparent. Any elements that may link the actin filaments to the plasma membrane remain to be identified as do integral membrane adhesion molecules themselves. Alternative, or perhaps additional, functions relate to potential contractile properties of the filament bundles. Unsicker (1974) concluded, based solely on ultrastructural observations, that the filament bundles may be contractile. He further implied that contraction of the bundles might contribute to the overall contractility of seminiferous tubules, which would presumably facilitate the transport of spermatozoa. The observation that the filaments are organized into loose bundles is consistent with the filaments having a contractile potential. The presence of contractile properties might also provide some degree of tensile support to the basal region of Sertoli cells. If, in fact, cell/substratum adhesion in reptilian Sertoli cells is actin related, as the morphology tends to indicate, the mechanism of attachment to the basal lamina used by Sertoli cells in reptiles is different from the strategy used by the same cells in mammals. In adult mammals, intermediate filaments are the most abundant filament type found in basal regions of Sertoli cells. The filaments form a layer adjacent to the plasma membrane and are related to hemidesmosome-like structures. Actin filaments are generally scarce at the base of the cell, except for reports of actin filament bundles in the pig (Toyama, 1975; Toyama et al., 1979; van Vorstenbosh et al., 1984). It would appear that reptilian Sertoli cell/substrate attachment is mainly actin associated while 220 in mammals it is mainly intermediate filament associated. This would indicate that the relative proportions of actin filament versus intermediate filament related matrix receptors may be different in the two vertebrate classes. In reptiles, those receptors and linking elements that relate to actin filaments may be more prevalent than those that relate to intermediate filaments, while in mammals the proportions are reversed. Interestingly, actin filaments have been reported adjacent to the basal plasma membrane in the developing rat (Ramos and Dym, 1979; Magre and Jost, 1983), but are superceeded by intermediate filaments as the animal matures. This could indicate that there is a shift away from actin based substrate adhesion towards a more intermediate filament based adhesion with further development. In summary, I have presented evidence that bundles of actin filaments are present in the basal cytoplasm of alligator and turtle Sertoli cells. Together with reports of similar bundles in other reptilian species, my data indicate that these structures may be a widespread phenomenon within this vertebrate class. I speculate that these filaments may play a role in cell/substrate adhesion. 221 FIGURE 7.1. Basal regions of alligator Sertoli cells. The plane of section is oblique to the tubule wall. Spermatogenic cells are indicated by the asterisks. Clearly evident are bundles of microfilaments (arrowheads) located immediately adjacent to the basal plasma membrane of the Sertoli cell. X8.970. 223 FIGURE 7.2. Shown here is the basal region of an alligator Sertoli cell. The plane of section is perpendicular to the tubule wall. Bundles of microfilaments (arrowheads) occur adjacent to the plasma membrane. The bundles are cut in cross-section or at an oblique angle to their longitudinal axis, and appear as "haystack-like" structures that project apically towards the perinuclear region. The Sertoli cell nucleus is indicated by the asterisk. The inset is a magnified view of the region indicated by the small rectangle. In the inset, microfilaments of one of the basal bundles are indicated by the small arrowheads while intermediate filaments that occur in the perinuclear cytoplasm are indicated by the larger arrowheads. The arrow indicates a microtubule. The microfilaments are smaller in diameter than the intermediate filaments. X22.917. Inset, X37.962. 225 FIGURE 7.3. Microfilament bundles in the basal cytoplasm of alligator Sertoli cells. As shown in panel (a), microfilament bundles (indicated by the asterisks) appear to form a "lattice-like" network adjacent to the plasma membrane. This particular section is cut at an acutely oblique angle to the tubule wall. Note that cisternae of endoplasmic reticulum (arrowheads) surround the filament bundles. Panel (b) shows a longitudinal section of a basal microfilament bundle (asterisk) in an alligator Sertoli cell. Observe that the filaments are uniformly aligned and that a cistern of endoplasmic reticulum (arrowheads) is associated with the bundle. Panel (a) X14,655. Panel (b) X28,830. 2 2 6 227 FIGURE 7.4. Basal regions of turtle Sertoli cells and adjacent structures in the tubule wall. Peritubular cells are indicated by the asterisks. In panel (a), distinct microfilament bundles (arrowheads) occur in folds of the basal plasma membrane of the Sertoli cell. Notice that the diameters of filaments in the peritubular cells are the same as those of filaments in basal regions of the Sertoli cell. Shown in panel (b) is a section, similar to that shown in panel (a), through the basal cytoplasm of a turtle Sertoli cell. Note that the plasma membrane of this cell is not corrugated and that a continuous layer of microfilaments (arrowheads) lies adjacent to it. Panel (a) X56,400. Panel (b) X72,308. 229 FIGURE 7.5. Fluorescence micrographs of a fixed frozen sections of alligator (panel a) and turtle (panel b) seminiferous tubules stained with rhodamine phalloidin. The white asterisk in each panel indicates the position of the seminiferous epithelium. The small black asterisks indicate the position of the tubule wall. In the alligator (panel a) , intense labelling occurs in cells of the tubule wall. Also labelled are linear tracts, apparently within the base of the epithelium, that are somewhat separated from the tubule wall and are oriented either parallel or perpendicular to it (arrowheads). In regions where the epithelium has separated from the wall, similar bands of fluorescence remain associated with the epithelium (arrows). In the turtle (panel b), as in the alligator, cells of the tubule wall are intensely labelled. The arrowheads indicate small bands of fluorescence that are apparently located in basal regions of the epithelium. Panel (a) X807. Panel (b) X944. a 3 0 CHAPTER 8 Summary and Concl 232 Summary and Conclusions Ectoplasmic specializations of eutherian mammal Sertoli cells consist of a layer of hexagonally packed actin filaments together with a cistern of endoplasmic reticulum on one side of the filament layer and regions of the plasma membrane involved with intercellular attachment on the other. The actin filaments are arranged in a unipolar fashion, are cross-linked to each other and to adjacent membranes, are organized into bundles, and are not part of a contractile system. Ectoplasmic specializations are found apically at sites of attachment to spermatids, and basally at sites of attachment to adjacent Sertoli cells. Similarly structured complexes are not found in any other cell type. Moreover, the unusual features of the ectoplasmic specialization appear to be restricted to mammals, specifically to placental mammals (Eutheria). Morphologically similar complexes have not been reported in Sertoli cells of any other vertebrate species. The function(s) of ectoplasmic specializations and the significance of their unusual structure in eutherian mammals are not entirely clear. One of the current hypotheses is that ectoplasmic specializations function primarily as a form of act in-related adhesion junction not unlike the zonulae adherens of other types of epithelial cells. Several lines of evidence, summarized elsewhere in this thesis (General Introduction, Chapter 5), support this hypothesis. I have further tested the hypothesis that ectoplasmic specializations of eutherian Sertoli cells are a specialized form of actin-associated adhesion junction. I have predicted that homologues of these junctions will occur in Sertoli cells of non-mammalian vertebrates but that these homologues will more closely resemble, in morphology and composition, actin-associated adhesion junctions found generally in cells than do ectoplasmic specializations in mammals. The approach I have used to verify this prediction has involved a systematic survey of the morphology and composition of 233 actin-associated adhesion junctions in Sertoli cells of non-mammalian vertebrates. What follows is a summary of the key findings of my study, along with the relavent findings of others and the general conclusions I have drawn. ACTIN-RELATED ADHESION JUNCTIONS ARE PRESENT IN SERTOLI CELLS OF NON-MAMMALIAN VERTEBRATES Ultrastructurally, Sertoli cell microfilament-associated junctions are detectable in representative species of all non-mammalian vertebrate classes. My own results include observations from all vertebrate classes except Amphibia. Based on the published micrographs of others (Fig. 15 in Burgos and Vitale-Calpe, 1967; Sprando and Russell, 1987a; Fig. 33 in Pudney, 1993), these junctions are clearly present in Sertoli cells of this class as well. In all classes, these Sertoli cell junctions are located in regions of attachment to spermatids and/or in regions of attachment to adjacent Sertoli cells. That the filament type present at these intercellular junctions is actin is indicated by their size relative to other cytoskeletal elements and by their ability to label with specific probes for filamentous actin. Evidence that these non-mammalian Sertoli cell attachment sites function in intercellular adhesion include the following observations: 1) Immunological probes for the protein vinculin react specifically with these sites in the rooster (Chapter 5). Vinculin is considered to be a molecular marker for identifying actin-associated adhesion sites in general. 2) Intercellular linkages are detectable between adjacent plasma membranes at the attachment sites in certain non-mammalian species (dogfish - Chapter 3; guppy -Chapter 4). Visible intercellular linkages are a common feature of adhesion junctions. 3) The attachment sites remain tightly adherent to spermatids that have been mechanically fragmented from the epithelium (ratfish - Stanely and Lambert, 1985; turtle - Chapter 5). 234 ACTIN-RELATED ADHESION JUNCTIONS OF NON-MAMMALIAN SERTOLI CELLS OCCUR WHERE ECTOPLASMIC SPECIALIZATIONS ARE FOUND IN MAMMALIAN SERTOLI CELLS In mammalian Sertoli cells, ectoplasmic specializations are formed mainly at two locations: 1) apically around the heads of elongating spermatids, and 2) basally at the level of the inter-Sertoli cell tight junctions that make up the"blood-testis" barrier. My results indicate that, in general, actin-related junctions are also found at these sites in non-mammalian vertebrate Sertoli cells (Fig. 8.1). Supporting my observations are those of others who have noted the presence of these junctions around spermatid heads (ratfish - Stanely and Lambert, 1985; bullfrog, rooster, turtle - Sprando and Russell, 1987a; rooster - Osman et al., 1980; lizards - Baccetti et al., 1983) and between adjacent Sertoli cells (fish - Arenas et al., 1995) in non-mammalian vertebrate species. Interestingly, while the apical Sertoli cell junctions around spermatids are seen fairly consistently throughout non-mammalian vertebrate classes, there are some exceptions. In the class Agnatha, Sertoli cell junctions with spermatids appear to be absent (Chapter 2). Similarly, in the class Osteichthyes, these Sertoli cell junctions are absent in the bowfin (Chapter 4) and in the bluegill (Sprando and Russell, 1987a), but are present in the guppy (Chapter 4). It appears that the important factor determining whether or not the Sertoli cell/spermatid junctions will form is the type of association that exists between Sertoli cells and spermatids. In those species in which a close physical association between Sertoli cells and spermatids develops, that is, one in which the spermatids are situated in apical Sertoli cell crypts, Sertoli cells form actin-related junctions around the spermatid heads. This is the case in mammals (eutherian and metatherian) and in most non-mammalian vertebrate classes. Species in which this occurs are internal fertilizers and tend to have spermatids that acquire an elongate head shape. In a second, a close association between Sertoli cells and spermatids does not exist. In these species, spermatids lie free within the spermatocyst lumen and entrenchment of 235 spermatids within Sertoli cell crypts is not seen. This is the case in the class Agnatha and in most species of the class Osteichthyes. Sertoli cells of these species do not form actin-related junctions with spermatids. These species are external fertilizers and, at least for the Osteichthyes species, tend to have spermatids that retain a round head shape. Inter-Sertoli cell actin-related junctions of non-mammalian vertebrates, like ectoplasmic specializations of eutherian mammals, occur in regions where tight junctions are found. This appears to hold true regardless of the position of the Sertoli cell tight junctions. In most anamniote classes, the zone of tight junctions is positioned apically between adjacent Sertoli cells (presumably in Agnatha - Chapter 2; Osteichthyes - Abraham et al., 1980; Marcaillou and Szollosi, 1980; Parmentier et al. , 1985; Amphibia - Franchi et al., 1982; Bergmann etal., 1983). Chondrichthyes is an exception in that the tight junctions in this anamniote class are positioned basally between Sertoli cells (Chapter 3). In amniotes, the inter-Sertoli cell tight junctions are also positioned basally between cells (Reptilia - Baccetti et al., 1983; Cavicchia and Miranda, 1988; Aves - Osman et al., 1980; Bergmann etal., 1983; Mammalia - Dym and Fawcett, 1970). In all cases, when inter-Sertoli cell actin-related junctions are present, they occur in close association with the tight junction zone. In the two classes where well-developed actin-related junctions are not seen between Sertoli cells (Reptilia and Aves), I did detect small focal sources of fluorescence near the base of the epithelium in tissue stained for actin (Chapter 5). These may represent focal act in-related junctions. The basal position of this fluorescence does correspond to the general vicinity of the Sertoli cell tight junctions in these species. Combined, these observations indicate that the inter-Sertoli cell actin-related junctions of non-mammalian vertebrates, like ectoplasmic specializations of eutherian mammals, occur in close association with the zone of Sertoli cell tight junctions. 236 FIGURE 8.1. Schematic representation of intercellular junctions formed between adjacent Sertoli cells and between Sertoli cells and spermatogenic cells within different vertebrate classes. The intent of this diagram is not to imply that one form necessarily gave rise to the next. Rather, it is to illustrate the different strategies for intercellular adhesion that have evolved within Sertoli cells of different vertebrate classes. In a "typical" epithelial cell, as seen at the bottom of the diagram, intermediate filament-related desmosomes, actin-related adhesion junctions and tight junctions occur together as part of an apical junctional complex. The junctional-associated actin filaments are arranged in loose networks or loose bundles. Agnatha: Sertoli cells display an arrangement of junctions similar to that seen in other epithelial cell types. No junctions are formed between Sertoli cells and elongating spermatids in this class. ChondrichthyesJVlodifications on the basic epithelial cell junctional complex design have occurred within Sertoli cells of chondrichthyan species. Junctions between adjacent Sertoli cells are located basally and consist of tight junctions and actin-related junctions. The actin filaments are loosely arranged and are occasionally associated with underlying cisternae of endoplasmic reticulum. Junctions with elongating spermatids consist of loose bundles of actin filaments that extend to the base of the Sertoli cell and are associated with endoplasmic reticulum. Osteichthyes and A m p h i b i a : In these classes, inter-Sertoli cell junctions occur together as part of an apically positioned junctional complex similar to those seen in other epithelial cell types. Sertoli cell junctions with spermatids are seen only in those species in which spermatids acquire an elongate head shape. In these species, Sertoli cells form well-developed actin-related junctions around the elongating heads of spermatids. Actin filaments of the junctions are loosely organized as networks or bundles. Elements of the endoplasmic reticulum do not occur in association with the junctions. Repti l ia and A v e s : Inter-Sertoli cell tight junctions are positioned basally between adjacent cells, a characteristic of all amniote Sertoli cells. A well-developed junctional complex at the level of the tight junctions is 2 3 7 not seen in these two classes. Small, focal filament-related junctions are seen along the lateral borders of neighboring Sertoli cells. Sertoli cell junctions formed with elongating spermatids consist of actin-related junctions in which the filaments are loosely arranged. Cisternae of endoplasmic reticulum are associated with the Sertoli cell/spermatid junctions in reptilian species but not in avian species. An additional general feature of the class Reptilia is the presence of basal concentrations of actin filaments in Sertoli cells. These are organized as loose networks or loose bundles and, depending on the species, may or may not be associated with elements of the endoplasmic reticulum. Metatheria: In Sertoli cells of marsupial mammals, well-developed act in-related junctions occur basally in the region of the tight junctions. The actin filaments of these junctions are arranged as loose bundles and are associated with endoplasmic reticulum. Sertoli cell/spermatid junctions consist of loosely arranged actin networks with underlying cisternae of endoplasmic reticulum. A peripheral zone of paracrystalline actin is associated with these apical junctions as they are assembled but is lost by the time the junctions are fully assembled. Euther ia: Sertoli cells of placental mammals exhibit inter-Sertoli cell junctional complexes basally between neighboring Sertoli cells. These complexes consist of tight junctions, desmosome-like junctions and well-developed actin-related junctions termed ectoplasmic specializations. The actin filaments occur in highly ordered paracrystalline arrays and are linked to underlying cisternae of endoplasmic reticulum. Similarly structured actin-related junctions (ectoplasmic specializations) are formed by Sertoli cells around elongating spermatids. 239 ACTIN-RELATED ADHESION JUNCTIONS OF NON-MAMMALIAN VERTEBRATE SERTOLI CELLS ARE CONTRACTILE One of the unusual features of ectoplasmic specializations in eutherian mammals is that they are not contractile. My results indicate that the actin-related adhesion junctions of non-mammalian vertebrate Sertoli cells, like actin-related adhesion junctions of other epithelia in general, are contractile. This is suggested by several observations: 1) The actin filaments at the attachment sites are loosely organized into bundles or networks. A loose packing arrangement of actin filaments is one characteristic of contractile systems. My results indicate that a loose filament organization is a general characteristic of these Sertoli cell junctions throughout all classes of non-mammalian vertebrates. Observations by others (Osman et al., 1980; Baccetti et al., 1983; Stanely and Lambert, 1985; Sprando and Russell, 1987a) in select non-mammalian vertebrate species, including in representatives from the class Amphibia (Sprando and Russell, 1987a), support this conclusion. 2) The protein myosin II is detectable at Sertoli cell actin-related adhesion junctions in species from a number of different non-mammalian vertebrate classes (dogfish -Chapter 3; ratfish - Stanely and Lambert, 1985; bowfin and guppy - Chapter 4; turtle, alligator and rooster - Chapter 5). 3) Sertoli cell actin-related junctions isolated from the turtle can be induced to contract in vitro in the presence of ATP (Chapter 5). 2 4 0 ACTIN-RELATED ADHESION JUNCTIONS OF NON-MAMMALIAN VERTEBRATE SERTOLI CELLS MORE CLOSELY RESEMBLE THE JUNCTIONS OF OTHER CELL TYPES THAN DO ECTOPLASMIC SPECIALIZATIONS OF EUTHERIAN MAMMALS Structurally, ectoplasmic specializations of eutherian mammals differ considerably from the junctions of other cell types. Actin-related adhesion junctions of non-mammalian vertebrate Sertoli cells, however, more closely resemble the actin-related adhesion junctions of other cell types than do ectoplasmic specializations of eutherian mammals. For example: 1) The actin filaments of the non-mammalian and marsupial Sertoli cell junctions are organized loosely into bundles or networks as occurs at junctions in other cell types in general. At ectoplasmic specializations of eutherian mammals, actin filaments are highly ordered into unipolar, hexagonal arrays. 2) The non-mammalian Sertoli cell junctions possess contractile properties, a characteristic of actin-related adhesion junctions in general. Ectoplasmic specializations of eutherian mammals are not contractile. 3) In most anamniote classes, inter-Sertoli cell actin-related junctions occur along with desmosomes and presumably tight junctions as part of a "typical" appearing apical junctional complex. In many cases (Agnatha - Chapter 2; Osteichthyes - Chapter 4; certain Amphibians - Fig. 33 in Pudney, 1993), the inter-Sertoli cell actin-related junctions are structurally indistinguishable from the zonulae adherens of other epithelial cell types. Apically located junctional complexes consisting of these three junction types are a characteristic of epithelial cells in general. 4) Elements of the endoplasmic reticulum are not a consistent feature of the non-mammalian Sertoli cell junctions, that is, the endoplasmic reticulum is associated with the sites in certain classes but not in others. In general, the actin-related adhesion junctions of other cell types are not related to the endoplasmic reticulum. A cistern of endoplasmic reticulum is a characteristic of eutherian ectoplasmic specializations. 241 ACTIN-RELATED ADHESION JUNCTIONS OF NON-MAMMALIAN VERTEBRATE SERTOLI CELLS FORMED WITH SPERMATIDS AND WITH ADJACENT SERTOLI CELLS ARE THE STRUCTURAL AND FUNCTIONAL HOMOLOGUES OF MAMMALIAN ECTOPLASMIC SPECIALIZATIONS Several lines of evidence, as described above, indicate that the Sertoli cell act in-related adhesion junctions formed with spermatids and with adjacent Sertoli cells in non-mammalian vertebrates are the structural and functional correlates of eutherian mammal ectoplasmic specializations. Observations consistent with this include: 1) The main filament type that comprises the non-mammalian Sertoli cell junctions is actin, as is the case at eutherian ectoplasmic specializations. 2) The non-mammalian Sertoli cell junctions occur at the same specific locations as eutherian ectoplasmic specializations do. 3) The non-mammalian Sertoli cell junctions appear to function in intercellular adhesion, as is suspected of eutherian ectoplasmic specializations. SERTOLI CELL ECTOPLASMIC SPECIALIZATIONS OF EUTHERIAN MAMMALS ARE A FORM OF ACTIN-RELATED ADHESION JUNCTION My results are consistent with the hypothesis that ectoplasmic specializations in eutherian mammal Sertoli cells are a form of actin-related adhesion junction. Evidence from my comparative studies verify the prediction that structural and functional homologues of eutherian ectoplasmic specializations are present in non-mammalian vertebrate Sertoli cells. My results further verify the prediction that these non-mammalian junctions more closely resemble the actin-related adhesion junctions found generally in cells than do eutherian ectoplasmic specializations. These comparative data add to the existing evidence that the unusally structured eutherian ectoplasmic specializations are a form of actin-related adhesion junction. 242 WHY DO SERTOLI CELLS FORM ACTIN-RELATED ADHESION JUNCTIONS? Why vertebrate Sertoli cells consistently form actin-related adhesion junctions at the same two specific sites is unknown. It is perhaps not that surprising to find Sertoli cells form actin-related junctions at the level of the inter-Sertoli cell tight junctions. One can see a direct homology between the inter-Sertoli cell actin-related junctions and the zonulae adherens of other epithelial cells. What is less obvious is why Sertoli cells consistently form actin-related adhesion junctions with spermatids but not usually with earlier-staged spermatogenic cells. With earlier-staged cells, Sertoli cells generally form much smaller focal attachment sites of the desmosome variety. This suggests that spermatids present a different adhesion requirement to Sertoli cells and that Sertoli cells respond by forming well-developed actin-related adhesion junctions. What this different requirement is remains unclear but it may very well be related to S the shape transformations the spermatid heads go through. While earlier-staged spermatogenic cells remain essentially round in shape, spermatid morphology changes dramatically as these cells differentiate and mature. During the process of spermiogenesis, the spermatid nucleus condenses and elongates, an acrosome forms (i n most classes) and the cytoplasmic volume around the head is reduced considerably. Well-developed actin-related adhesion junctions may respresent the Sertoli cell's strategy for maintaining adhesion to the very dynamic contours of these cells. EVOLUTIONARY SPECULATIONS This study confirms that the unique structure of eutherian ectoplasmic specializations, with their hexagonally packed, non-contractile actin filaments, is peculiar to this group of mammals. Eutherian ectoplasmic specializations differ considerably from their marsupial and non-mammalian counterparts and similar structured junctions are not found in any other cell type. The significance of the unusual structure of eutherian ectoplasmic specializations remains unclear; however, the observations of this study do allow for speculation on what the evolutionary precursors of this junction may have been like and, perhaps, what some of the selective pressures acting on these sites may have been. a) Intercellular Adhesion Strategies within the Germinal Epithelium In epithelial cells in general, actin-related adhesion junctions are a common mechanism for establishing and maintaining . intercellular adhesion as well as cell/substatum adhesion. What becomes apparent upon examining the actin-related adhesion junctions of the vertebrate germinal epithelium is that different strategies for intercellular adhesion appear to have evolved within different classes. These various strategies are reflected in the structure of the Sertoli cell actin-related junctions. In most anamniote classes, for example, Sertoli cells form actin-related adhesion junctions that closely resemble those found in other epithelial cell types. This is particularly true of the inter-Sertoli cell junctions which generally occur apically along with desmosomes and tight junctions as part of a "typical" epithelial cell junctional complex. This suggests the actin-based adhesion mechanism used by Sertoli cells of these classes is not that different from those used by other epithelial cell types. Modifications on this basic junction design could have led to the range of different Sertoli cell junctions now seen in various classes. One such modification has occurred in the class Chondrichthyes, where the actin filaments of the apically located Sertoli cell-spermatid junctions extend down to and are presumably linked to the base of the cell (Stanely and Lambert, 1 985) . These "modified" junctions are still a contractile form of actin-related adhesion junction, but through the link to the base of the cell they have now likely acquired an additional role(s), such as the positioning and bundling of spermatids (see Chapter 3 discussion). Another and more dramatic remodelling of the basic junction design appears 244 to have occured in eutherian mammals, where myosin II has been lost from the sites and the actin filaments have become cross-linked. Different stategies for Sertoli cell/substatum adhesion appear to have evolved within different vertebrate classes as well. Adhesion to the substrate within eutherian mammals appears to be mediated primarily through intermedite filament-related junctions of the hemidesmosome variety (Russell, 1977c). This is contrasted by the class Reptilia where well-developed basal concentrations of actin filaments suggest that Sertoil cells of this class utilize an actin-based cell/substatum adhesion mechanism (see Chapter 7). These contractile basal actin filaments may have also evolved additional functions in this class, such as a role in the overall contraction of the seminiferous tubules (Unsicker, 1974). b) Precursor Junctions and Their Structural Remodelling From the observations of this study, it is reasonable to speculate that the evolutionary precursor of the modern-day eutherian ectoplasmic specializations was likely a contractile actin-related adhesion junction. Early in vertebrate evolution these Sertoli cell junctions likely resembled the adherens junctions of other epithelia. From these early junctions, one can envision some of the changes which may have occurred at these sites. Elements of the endoplasmic reticulum became associated with these contractile junctions in several classes, perhaps through the process of convergent evolution. In early eutherian mammals, the modifications were more extensive and included the loss of myosin II and a re-ordering of the actin filaments into hexagonal arrays. The filament re-ordering was likely a result of the expression of new actin binding proteins not previously found at these sites. For example, a likely candidate of one of these proteins is fimbrin, a protein which has been tentatively identified at eutherian ectoplasmic specializations (Grove and Vogl, 1989). Fimbrin has been shown to bundle actin filaments into hexagonal arrays in vitro (Bretscher 1981). Although 245 localization experiments for this protein at the actin-related junctions of non-mammalian vertebrate Sertoli cells have not been conducted, one would predict that i t would be found only in the eutherian form of these junctions based on its filament bundling characteristics. Among other modifications which may have occurred at these Sertoli cell junctions during the course of evolution are ones that involved the type(s) of cell adhesion molecules present. E-cadherin, the cell adhesion molecule which characterizes the adherens junctions of other epithelia, appears to be absent at eutherian ectoplasmic specializations (Byers et al., 1994). This likely represents another example of how the protein composition of these Sertoli cell junctions in eutherian mammals has diverged from that of the adherens junctions of other epithelia. It is possible that E-cadherin may be present at the actin-related Sertoli cell junctions of non-mammalian vertebrates, but this has not been demonstrated. The type(s) of adhesion molecules which function i n place of E-cadherin at eutherian ectoplasmic specializations is/are unknown; however, one molecule that has now been localized to these junctions is the a6B1 integrin. As with E-cadherin, the distribution of integrins with non-mammalian vertebrate Sertoli cells is unknown. Given the unique structure of eutherian ectoplasmic specializations, it is possible that the presence of an integrin at these sites is a eutherian adaptation. c) Reasons for Remodelling Sertoli Cell Junctions Why the Sertoli cell actin-related adhesion junctions of many vertebrate classes have evolved structural adaptations is unknown. One of the common structural adaptations of these sites includes the incorporation of elements of the endoplasmic reticulum into the junctions. This is seen in several non-mammalian vertebrate classes as well as in eutherian mammals. The association of this organelle with the contractile junctions of non-mammalian vertebrates may have evolved as part of a regulatory mechanism of these sites. The endoplasmic reticulum is known to be a storage sink for 246 calcium, a cation involved with the activation of non-muscle myosin II. Calcium is also a known second messenger in various signal transduction pathways. Controlled release of calcium from the junction-associated endoplasmic reticulum could therefore play one of several different regulatory roles at these sites, such as in the contraction of the actin bundles or in the assembly/disassembly of the junctions. At the non-contractile ectoplasmic specializations of eutherian mammals, one can envision the association of the endoplasmic reticulum with the junction site evolving as part of a regulatory system as well. The endoplasmic reticulum at eutherian ectoplasmic specializations may also have evolved additional roles not found at the non-mammalian vertebrate junctions. For example, there is now considerable evidence indicating that the observed movement of spermatids within the eutherian seminiferous epithelium during spermatogenesis is generated via a microtubule-based vesicle transport system. The proposed mechanism by which this occurs is one in which apical ectoplasmic specializations and attached spermatids are moved as units along microtubule tracts i n the Sertoli cell (Redenbach and Vogl, 1991; Vogl et al., 1993; Vogl, 1996). In this model, the endoplasmic reticulum at the junction site is a crucial link between the attachment plaque to the spermatids and the underlying microtubule tracts in the Sertoli cell. Without the endoplasmic reticulum, microtubule-based transport of the junctions and attached spermatids would not be possible. How a Sertoli cell junction could evolve a structural and functional link to a microtubule-based transport system is intriguing. Observations from this study suggest that the evolutionary precursors of the modern-day eutherian junctions were likely contractile actin-related adhesion junctions not unlike the Sertoli cell junctions currently seen in other vertebrates. If so, then elements of the endoplasmic reticulum were likely already associated with the junctions as local calcium stores. Another requirement of the proposed microtubule-based transport system is that high concentrations of microtubules occur parallel to the Sertoli cell crypts housing 247 spermatids. This has been demonstrated in eutherian Sertoli cells (Vogl et al., 1 983; Amlani and Vogl, 1988; Vogl 1988). Significantly, precedents for this microtubule arrangement exist in non-mammalian Sertoli cells. There are several reports of an abundance of microtubules coursing parallel to the long axis of the Sertoli cell and surrounding the embedded spermatids in amphibians (Burgos and Vitale-Calpe, 1967; Reed and Stanely; 1972; Pudney, 1993), in reptiles (Pudney, 1993; Pfeiffer and Vogl - personal observations) and in birds (see Pudney, 1993). These observations indicate that, while a microtubule-based transport system involving these Sertoli cell junctions and attached spermatids may be unique to eutherian mammals, the necessary elements for such a system were likely already present in the evolutionary precursors of eutherian ectoplasmic specializations. If so, then the early association of the endoplasmic reticulum with the junctions (which fulfilled one function) may have pre-adapted the sites for a future, unrelated function (a role in a microtubule-based transport system). Similarly, an abundance of microtubules near these early junctions may have p r e -adapted the system for a later role in the transport of junctions. This process of p r e -adaptation may have played an important role in the evolution of the eutherian ectoplasmic specialization. A role of eutherian ectoplasmic specializations in spermatid transport might explain the some of the further structural adaptations present at these junctions. One requisite of this microtubule-based transport model is that elements of the junction must be stablilized to the extent that movement of the microtubule-based motors wil l result in movement of the junction and attached spermatid as a unit. A re-ordering of the actin filaments at the sites into tightly cross-linked, non-contractile filament bundles may represent one way of structurally reinforcing and further stabilizing the adhesion sites. As a microtubule-based transport system evolved an association with the adhesion junctions, factors which provided further stability and reinforcement to the sites may have been positively selected for. As mentioned above, some of these factors may have 248 included the expression of new actin binding proteins at these sites. At this point it is important to emphasize that, although apical eutherian ectoplasmic specializations may have assumed a role in spermatid transport, intercellular adhesion still remains the primary function of these junctions. The proposed microtubule-based transport system for spermatid movement utilizes modified actin-related adhesion junctions Sertoli cells form with spermatids. Turning from the eutherian condition to the different adhesion strategies of other vertebrate classes, it is interesting to note that a role in spermatid movement also appears to have evolved independently in a separate class of vertebrates, Chondrichthyes. During spermiogenesis in this class, elongating spermatids are gathered together within the apex of each Sertoli cell and then pulled deep down as a tight bundle towards the base of the Sertoli cell. This movement appears to be generated by the Sertoli cell actin-related adhesion junctions; however, the underlying mechanism is thought to be through a direct contraction of the junctions (Stanely and Lambert, 1 985; see Chapter 3 discussion) and not through a link to a microtubule-based transport system. In this class, selective pressures acting on the junctions have shaped the sites in a very different direction from that which has occurred at eutherian junctions. WHEN IS A SERTOLI CELL ACTIN-RELATED ADHESION JUNCTION AN "ECTOPLASMIC SPECIALIZATION"? As shown in this study, structural homologues of eutherian ectoplasmic specializations do occur in non-mammalian vertebrate Sertoli cells; however, these homologues resemble in several ways the more "typical" actin-related adhesion junctions of other epithelial cell types. In fact, in several anamniote classes the inter-Sertoli cell actin-related junctions very closely resemble the zonulae adherens of other epithelia. This raises the interesting question of terminology. Should the non-mammalian vertebrate homologues of eutherian ectoplasmic specializations be referred 2 4 9 to as "ectoplasmic specializations" even though, in certain classes, they structurally and functionally appear to be "typical" actin-related adhesion junctions? Or, should the term "ectoplasmic specialization" be reserved for only the unique eutherian mammal form of this junction? Russell (1977a) originally suggested the term "ectoplasmic specialization" be used to refer to the eutherian form of these structures since they are morphologically unique and since, at that time, there was little evidence to indicate that these structures function as an intercellular junction. Subsequently, the term "ectoplasmic specialization" has also been used to refer to the non-mammalian vertebrate homologues of this junction (Baccetti et al., 1983; Sprando and Russell, 1987a; Arenas et al. , 1995). Considerable evidence now indicates that eutherian ectoplasmic specializations and the Sertoli cell homologues in non-mammalian vertebrates are a form of intercellular junction, specifically one related to adhesion. Given this, it could be argued that a term that recognizes this fact would now be more appropriate than the term "ectoplasmic specialization". The terms "actin-related adhesion junctions" or simply "adherens junctions" could be used accurately to describe these Sertoli cell junctions in certain vertebrate classes (Agnatha, Osteichthyes) and would indicate their homology to the adherens junctions of other epithelial cell types. However, such terms do not recognize the fact that in many non-mammalian classes these Sertoli cell junctions display features not found at "typical" adherens junctions. For example, in several classes elements of the endoplasmic reticulum are associated with the sites. Moreover, in several classes the actin filaments associated with the apical Sertoli cell-spermatid junctions extend down and away from the junction site for a considerable distance. The most extreme case of this occurs in the class Chondrichthyes where the actin filaments extend down from the contact site to the base of the Sertoli cell (Chapter 3). 250 In returning to the early studies of these sites, the term "junctional specialization" was used by Flickinger and Fawcett (1967) to describe these sites in eutherian mammals. This term is perhaps more desirable than "ectoplasmic specialization". It specifies that these structures are a form of intercellular junction yet also indicates that they differ from the adherens junctions of other cell types. I propose that the term "junctional specialization", rather than "ectoplasmic specialization", be used to denote these structures in non-mammalian vertebrate classes as well as in metatherian mammals. Futhermore, I propose that the term "ectoplasmic specialization" be reserved for the eutherian form of this junction only. This latter term is now deeply rooted in the literature regarding the eutherian condition and to change it at this point would add confusion. Its continued use, in fact, does help emphasize that the eutherian form of this junction is structurally unique and that it differs considerably from the Sertoli cell junctions in other vertebrate classes. To indicate the homology between these eutherian junctions and their non-mammalian counterparts, it may be useful to modify this term to "ectoplasmic ('junctional') specialization", the term used by Russell (1977a) in the title of his initial study on these structures. Finally, if the term "junctional specialization" is to be used to describe the non-eutherian homologues of ectoplasmic specializations, it is important to define the criteria that must be met before a Sertoli cell junction can be termed a "junctional specialization". First, the junction must occur at one of the two locations where ectoplasmic specializations are found in eutherian mammals. Secondly, the junction must consist of actin filaments. Regarding this second point, the term "junctional specialization" therefore refers only to the actin-related junction at the level of the inter-Sertoli cell junctional complex and does not include the often nearby tight junctions, desmosomes or desmosome-like junctions. 251 SUMMARY In this study, I provide information on the comparative cell biology of a Sertoli cell intercellular junction. My data show that vertebrate Sertoli cells use actin-related adhesion junctions as a common mechanism for adhesion to adjacent Sertoli cells and to spermatids. In non-mammalian vertebrate classes the junctions appear to have contractile properties and more closely resemble the actin-related adhesion junctions of other cell types than do eutherian ectoplasmic specializations. These data suggest that ectoplasmic specializations of eutherian Sertoli cells represent a modified form of actin-related adhesion junction. I speculate on what the evolutionary precursors of the modern-day eutherian junctions may have been like as well as on how and why structural adaptations may have occurred at these sites. My observations indicate that different strategies for intercellular adhesion have evolved within the germinal epithelium in various vertebrate classes and that these differences are reflected in the structure of the Sertoli cell actin-related adhesion junctions. 252 BIBLIOGRAPHY Abraham, M., E. Rahamim, H. Tibika, and E. Golenser. 1980. The blood-testis barrier in Aphanius dispar (Teleostei). Cell Tissue Res., 211:207-214. Albelda, S.M., and C A . Buck. 1990. Integrins and other cell adhesion molecules. FASEB J . , 4:2868-2880. Amlani, S., and A.W. Vogl. 1988. Changes in the distribution of microtubules and intermediate filaments in mammalian Sertoli cells during spermatogenesis. Anat. R e c , 220:143-160 . Arenas, M.I., B. Fraile, M.P. De Miguel, and R. Paniagua. 1995. Cytoskeleton in Sertoli cells of the mosquito fish (Gambusia affinis holbrooki). Anat. R e c , 24:225-234. Baccetti, B., Bigliardi, E., Talluri, M. V., and A.G. Burrini. 1983. The Sertoli cell in lizards. J . Ultrastruct. Res., 85:11-23. Balsamo, J . , T.C. Leung, H. Ernst, M.K.B. Zanin, S. Hoffman, and J . Lilien. 1996. Regulated binding of a PTP1B-like phosphatase to N-cadherin control of cadherin-mediated adhesion by dephosphorylation of B-catenin. J . Cell Biol., 134:801-813. Beckerle, M.C. 1986. Identification of a new protein localized at sites of cell-substrate adhesion. J . Cell Biol., 103:1679-1687. Bergmann, M., H. Greven, and J . Schindelmeiser. 1983. Observations on the blood-testis barrier in a frog and a salamander. Cell Tissue Res., 232:189:200. Bergmann, M., J . Schindelmeiser, H. Greven. 1984. The blood-testis barrier in vertebrates having different testicular organization. Cell Tissue Res., 238:145-150. Billard, R. 1970. La spermatogenese de Poecilia reticulata III. - Ultrastructure des cellules de Sertoli. Ann. Biol. anim. Bioch. Biophys., 10:37-50. Billard, R. 1984. Ultrastructural changes in the spermatogonia and spermatocytes of Poecilia reticulata during spermatogenesis. Cell Tissue Res., 237:219-226. Billard, R. 1986. Spermatogenesis and spermatology of some teleost fish species. Reprod. Nutr. Develop., 26:877-920. Boiler, K., D. Vestweber, and R. Kemler. 1985. Cell-adhesion molecule uvomorulin is localized in the intermediate junctions of adult intestinal epithelial cells. J . Cell Biol. , 100:327-332. Borradori, L. and A. Sonnenberg. 1996. Hemidesmosomes: roles in adhesion, signaling and human diseases. Curr. Opin. Cell Biol., 8:647-656. Brokelmann, J . 1963. Fine structure of germ cells and Sertoli cells during the cycle of the seminiferous epithelium in the rat. Z. Zellforsh. Microsk. Anat., 59:820-850. Buck, C.A., and A.F. Horwitz. 1987. Cell surface receptors for extracellular matrix molecules. Annu. Rev. Cell Biol., 3:179-205. 253 Burgess, D.R. 1982. Reactivation of intestinal epithelial cell brush border motility: ATP-dependent contraction via aterminal web contractile ring. J . Cell Biol., 9 5 : 8 5 3 -863. Burridge, K., K. Fath, T. Kelly, G. Nuckolls, and C. Turner. 1988. Focal adhesions: Transmembrane junctions between the extracellular matrix and the cytoskeleton. Annu. Rev. Cell Biol., 4:487-525. Byers, H.R., and K. Fujiwara. 1982. Stress fibers in cells in situ: immunofluorescence visualization with anti-actin, anti-myosin, and anti-alpha-actinin. J . Cell Biol. , 93:804-81 1 . Byers, S., R-M. Pelletier, C. Suarez-Quian. 1993. Sertoli cell junctions and the seminiferous epithelium barrier. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp 431-446. Byers, S., B. Jegou, C. MacCalman, and O. Blaschuk. 1993. Sertoli cell adhesion molecules and the collective organization of the testis. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp. 461-476. Callard, G.V. 1991. Spermatogenesis. In: Pang P., Schreibman, M. (eds) Vertebrate Endocrinology: Fundamentals and Biochemical Implications, vol. 4, part A. Academic Press, New York, pp. 303-341. Carter, W.G., E.A. Wayner, T.S. Bouchard, and P. Kaur. 1990a. The role of integrins a2B1 and a3B1 in cell-cell and cell-substrate adhesion of human epidermal cells. J . Cell Biol., 110:1387-1404. Carter, W.G., P. Kaur, S.G.Gil, P.J. Gahr, and E.A. Wayner. 1990b. Distinct functions for integrins a3B1 in focal adhesions and a6B4/bullous pemphigoid antigen in a new stable anchoring contact (SAC) of keratinocytes: Relation to hemidesmosomes. J . Cell Biol., 111:3141-3154. Cavicchia, J.C. and J.C. Miranda. 1988. The blood-testis barrier in lizards during the annual spermatogenic cycle. Microsc. Electron Biol. Celular., 12:73-87. Connell, C.J. 1974. The Sertoli cell of the sexually mature dog. Anat. Rec , 178:333A. Cooksey, E.J., and B. Rothwell. 1973. The ultrastructure of the Sertoli cell and its differentiation in the domestic fowl (Gallus domesticus). J . Anat., 114:329-345. Crawford, A.W., J.W. Michelsen, and M.C. Beckerle. 1992. An interaction between zyxin and alpha-actinin. J . Cell Biol., 116:1381-1393. Dodd, J.M., and J.P. Sumpter. 1984. Fishes. In: Lamming G.E. (ed.), Marshall's Physiology of Reproduction, vol. 1, Churchill Livingstone, New York., pp. 1-79. Drenckhahn, D., and H. Franz. 1986. Identification of actin, alpha-actinin-, and vinculin-containing plaques at the lateral membrane of epithelial cells. J . Cell Biol. , 102:1843-1852. Drenckhahn, D., and J . Wagner. 1986. Stress fibers in the splenic sinus endothelium in situ: Molecular structure, relationship to the extracellular matrix, and contractility. J . Cell Biol., 102:1738-1747. 254 Drenckhahn, D., and R. Dermietzel. 1988. Organization of the actin filament cytoskeleton in the intestinal brush border: a quantitative and qualitative immunoelectron microscope study. J . Cell Biol., 107:1037-1048. Dym, M., and D.W. Fawcett. 1970. The blood-testis barrier in the rat and the physiological compartmentation of the seminiferous epitheium. Biol. Reprod., 3 :308-326. Enders, G.C. 1993. Sertoli-Sertoli and Sertoli-germ cell communications. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp. 4 4 7 - 4 6 0 . Farquhar, M.G., and G.E. Palade. 1963. Junctional complexes in various epithelia. J . Cell Biol., 17:375-409. Fawcett, D.W. 1975. Ultrastructure and function of the Sertoli cell. In: Handbook of Physiology. R.O. Greep, ed. Williams and Wilkins, Baltimore, Maryland, Sect. 7, Vol. 5, pp. 21-55. Flickinger, C , and D.W. Fawcett. 1967. The junctional specializations of Sertoli cells in the seminiferous epithelium. Anat. R e c , 158:207-222. Franchi, E., M. Camatini, and I. deCurtis. 1982. Morphological evidence of a permeability barrier in urodele testis. J . Ultrastruct. Res., 80:253-263. Franke, W.W., C. Grund, A. Fink, K. Weber, B.M. Jockusch, H. Zentgraf, and M. Osborn. 1978. Localization of actin in the microfilament bundles associated with the junctional specializations between Sertoli cells and spermatids. Biol. Cell, 31:7-14. Garrod, D.R. 1993. Desmosomes and hemidesmosomes. Curr. Opin. Cell Biol., 6 :682-687. Geiger, B. 1989. Cytoskeleton-associated cell contacts. Curr. Opin. Cell Biol., 1:103-109. Geiger, B., A.N. Dutton, K.T. Tokuyasu, and S.R. Singer. 1981. Immunoelectron microscope studies of membrane-microfilament interactions: Distributions of alpha-actinin, tropomyosin, and vinculin in intestinal epithelial brush border and chicken gizzard smooth muscle cells. J . Cell Biol., 91:614-628. Geiger, B., Z. Avnur, G. Rinnerthaler, H. Hinssen, and V.J. Small. 1984. Microfilament-organizing centers in areas of cell contact: Cytoskeletal interactions during cell attachment and locomotion. J . Cell Biol., 99:83s-91s. Geiger, B., Z. Avnur, T. Volberg, and T. Volk. 1985. Molecular domains of adherens junctions. In The Cell in Contact, (eds. G. M. Edelman and J . P. Thiery), pp. 461 -489 . Neuroscience Institute Publication, John Wiley and Sons, New York. Geiger, B., S. Yehuda-Levenberg, and A.D. Bershadsky. 1995. Molecular interactions in the submembrane plaque of cell-cell and cell-matrix adhesions. Acta. Anat., 154:46-62. 255 Grier, H.J. 1975. Aspects of germinal cyst and sperm development in Poecilia latipinna (Teleostei: Poeciliidae). J . Morph., 146:229-250. Grier, H.J. 1992. Chordate testis: The extracellular matrix hypothesis. J . Exp. Zool., 261 :1 51 -1 60. Grier, H.J. 1993. Comparative organization of Sertoli cells including the Sertoli cell barrier. In: The Sertoli cell (eds. L.D. Russell and M.D. Griswald). Cache River Press, Clearwater, Florida, pp. 703-739. Grove, B.D., and A.W. Vogl. 1989. Sertoli cell ectoplasmic specializations: A type of actin-associated adhesion junction? J . Cell Sci., 93: 309-323. Grove, B.D., D.C. Pfeiffer, S. Allen, and A.W. Vogl. 1990. Immunofluorescence localization of vinculin in ectoplasmic ("junctional") specializations of rat Sertoli cells. Am. J . Anat., 188:44-56. Gumbiner, B.M. 1992. Cell adhesion molecules in epithelia. In: Tight junctions (ed. M. Cereijido). C R C Press Inc., Boca Raton, Florida, pp. 91-104. Gumbiner, B.M. 1996. Cell adhesion: The molecular basis of tissue architecture and morphogenesis. Cell, 84:345-357. Hay, E.D. 1983. Cell and extracellular matrix: Their organization and mutual dependence. Modern Cell Biol., 2:509-548. Hirano, S., A. Nose, K. Hatta, A. Kawakami, and M. Takeichi. 1987. Calcium-dependent cell-cell adhesion molecules (cadherins): subclass specificities and possible involvement of actin bundles. J . Cell Biol. 105:2501-2510. Hirokawa, N., T.C.S. Keller, III, R. Chasan, and M.S. Mooseker. 1983. Mechanism of brush border contractility studied by the quick-freeze, deep-etch method. J . Cell Biol. , 96:1 325-1 336. Hynes, R.O. 1987. Integrins: A family of cell surface receptors. Cell, 48:549-554. Hynes, R.O. 1992. Integrins: versatility, modulation, and signaling in cell adhesion. Cell, 69:11-25. Isenberg, G., P.C. Rathke, N. Hulsmann, W.W. Franke, and K.E. Wohlfarth-Botterman. 1976. Cytoplasmic actomyosin fibrils in tissue culture cells. Direct proof of contractility by visualization of ATP-induced contraction in fibrils isolated by laser microbeam dissection. Cell Tissue Res., 166:427-443. Jamieson, B.G.M. 1991. Fish evolution and systematics: evidence from spermatozoa. Cambridge University Press, Cambridge, England. Jespersen, A. 1975. Fine structure of spermiogenesis in Eastern Pacific species of hagfish (Myxinidae). Acta. Zool., 56:189:198. Jockusch, B.M., P. Bubeck, K. Giehl, M. Kroemker, J . Moschner, M. Rothkegel, M. Rudiger, K. Schlutter, G. Stanke, and J . Winkler. 1995. The molecular architecture of focal adhesions. Annu. Rev. Cell Dev. Biol., 11:376-416. 256 Jou, T., D.B. Stewart, J . Stappert, W.J. Nelson, and J.A. Marrs. 1995. Genetic and biochemical dissection of protein linkages in the cadherin-catenin complex. Proc. Natl. Acad. Sci. USA, 92:5067-5071. Knudsen, K.A., A.P. Soler, K.R. Johnson, and M.J. Wheelock. 1995. Interaction of oc-actinin with the cadherin/catenin cell-cell adhesion complex via a-catenin. J . Cell Biol., 130:67-77. Koch, P.J. and W.W. Franke. 1994. Desmosomal cadherins: another growing multigene family of adhesion molecules. Curr. Opin. Cell Biol., 6:682-687. Lampugnani, M.G., M. Resnati, E. Dejana, and P.C. Marchisio. 1991. The role of integrins in the maintenance of endothelial monolayer integrity. J . Cell. Biol. , 1 12:479-490. Larjava, H., J . Peltonen, S.K. Akiyama, S.S. Yamada, H.R. Gralnick, J . Uitto, and K.M. Yamada. 1990. Novel function for Bi integrins in keratinocyte cell-cell interactions. J . Cell Biol., 110:803-815. Lazarides, E., and K. Burridge. 1975. Alpha-actinin: immunofluorescent localization of a muscle structural protein in nonmuscle cells. Cell, 6:289-298. Madara, J.L. 1987. Intestinal absorptive cell tight junctions are linked to the cytoskeleton. Am. J . Physiol., 253: C171-175. Madara, J.L. 1992. Relationships between the tight junction and the cytoskeleton. In: Tight junctions (ed. M. Cereijido). CRC Press Inc., Boca Raton, Florida, pp. 105-119. Magre, S., and A. Jost. 1983. Early stages of the differentiation of the rat testis: relations between Sertoli and germ cells. In The Seventh Symposium of the British Society for Developmental Biology, Current problems in germ cell differentiation, (eds. A. McLaren and C. C. Wylie), pp. 201-214. Cambridge University Press, Cambridge. Marcaillou, C , and A. Szollosi. 1980. The "blood-testis" barrier in a nematode and a fish: a generalizable concept. J . Utrastruct. Res. 70:128-136. Masri, B.A., L.D. Russell, and A.W. Vogl. 1987. Distribution of actin in spermatids and adjacent Sertoli cell regions of the rat. Anat. R e c , 218:20-26. Menkel, A.R., M. Kroemker, P. Bubeck, M. Ronsiek, G. Nikolai, and B.M. Jockusch. 1994. Characterization of an F-actin-binding domain in the cytoskeletal protein vinculin. J . Cell Biol., 126:1231-1240. Mooseker, M.S. 1985. Organization, chemistry, and assembly of the cytoskeletal apparatus of the intestinal brush border. Ann. Rev. Cell Biol. 1:209-241. Morales, C , and Y. Clermont. 1993. Structural changes of the Sertoli cell during the cycle of the seminiferous epithelium. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp. 305-329. Munro, S.B., and O.W. Blaschuk. 1996. A comprehensive survey of the cadherins expressed in the testes of fetal, immature, and adult mice utilizing the polymerase chain reaction. Biol. Reprod., 55:822-827. 257 Nanthke, I.S., L. Hinck, J.R. Swedlow, J . Papkoff, and J.W. Nelson. 1994. Defining interactions and distributions of cadherin and catenin complexes in polarized epithelial cells. J . Cell Biol., 125:1341-1352. Nicander, L. 1967. An electron microscopical study of cell contacts in the seminiferous tubules of some mammals. Z. Zellforsh. Microsk. Anat. 83:375-397. Nicander, L. 1969. Comparative studies on the fine structure of vertebrate spermatazoa. In: Proc. Intern. Symp. Comparative Spermatology, (Baccetti, B., ed.), Rome. Osman, D.I., H. Ekwall, and L. Ploen. 1980. Specialized cell contacts and the blood-testis barrier in the seminiferous tubules of the domestic fowl (gallus domesticus). Int. J . Androl., 3:553-562. Otey, C.A., F.M. Pavalko, and K. Burridge. 1990. An interaction between alpha-actinin and the beta 1 integrin subunit in vitro. J . Cell Biol., 111:721-729. Owaribe, K., R. Kodama, and G. Eguchi. 1981. Demonstration of contractility of circumferential actin bundles and its morphogenetic significance in pigmented epithelium in vitro and in vivo. J . Cell Biol., 90:507-514. Palombi, F., M. Salanova, G. Tarone, D. Farini, M. Stefanini. 1992. Distribution of B1 integrin subunit in rat seminiferous epithelium. Biol. Reprod., 47:1173-1182. Parmentier, H.K., J.G.M. van den Boogaart, and P.M. Timmermans. Physiological compartmentation in gonadal tissue of the common carp (Cyprinis carpio L ) . A study with horseradish peroxidase and monoclonal antibodies. Cell Tissue Res., 242:75-81. Pfeiffer, D.C. 1990. Master's Thesis. University of British Columbia, Vancouver, B.C., Canada. Pfeiffer, D . C , and A.W. Vogl. 1991a. Evidence that vinculin is co-distributed with actin bundles in ectoplasmic ("junctional") specializations of mammalian Sertoli cells. Anat. R e c , 231:89-100. Pfeiffer, D.C, S. Dedhar, S.W. Byers, and A.W. Vogl. 1991. Evidence that an integrin may be present at specialized sites ('ectoplasmic specializations') of intercellular adhesion in the seminiferous epithelium. J . Cell Biol., 115:482a. Philp, N.J., and V.T. Nachmias. 1985. Components of the cytoskeleton in the retinal pigmented epithelium of the chick. J . Cell Biol., 101:358-362. Pudney, J . 1993. Comparative cytology of the non-mammalian vertebrate Sertoli cell. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp 611-657. Pudney, J . 1995. Spermatogenesis in nonmammalian vertebrates. Microscopy Res. Tech. , 32:459-497. Ramos, A.S. JR., and M. Dym, M. 1979. Ultrastructural differentiation of rat Sertoli cells. Biol, of Reprod., 21:909-922. 258 Rimm, D.L., E.R. Koslov, P. Kebriaei, C D . Cianci, and J.S. Morrow. 1995. oc1(E)-catenin is an actin-binding and actin-bundling protein mediating the attachment of F -actin to the membrane adhesion complex. Proc. Natl. Acad. Sci. USA, 92:8813-8817. Romer, A.S. 1970. Who's who among the vertebrates. In: Romer, The vertebrate body. W.B. Saunders Co., Philadelphia, pp. 30-76. Romrell, L.J., and M.H. Ross. 1979. Characterization of Sertoli cell-germ cell junctional specializations in dissociated testicular cells. Anat. R e c , 193:23-42. Russell, L. 1977a. Observations on rat Sertoli ectoplasmic ('junctional') specializations in their association with germ cells of the rat testis. Tissue & Cell, 9:475-498. Russell, L. 1977b. Desmosome-like junctions between Sertoli cells and germ cells in the rat testis. Am. J . Anat., 301-312. Russell, L. 1977c. Movement of spermatocytes from the basal to the adluminal compartment of the rat testis. Am. J . Anat., 148:313-328. Russell, L. 1980. Sertoli-germ cell interrelations: a review. Gamete Res., 3:179-202. Russell, L , and R.N. Peterson. 1985. Sertoli cell junctions: morphological and functional correlates. Int. Rev. Cytol., 94:177-211. Russell, L , J . C Goh, R.M.A. Rashed, and A.W. Vogl. 1988. The consequences of actin disruption at Sertoli ectoplasmic specialization sites facing spermatids after in vivo exposure of rat testis to cytochalasin D. Biol. Reprod., 39:105-118. Russell, L. 1993a. Role in spermiation. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp. 269-303. Russell, L. 1993b. Morphological and functional evidence for Sertoli-germ cell relationships. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp. 365-390. Sacco, P.A., P.M. McGrananhan, M.J. Wheelock, and K.R. Johnson. 1995. Identification of plakoglobin domains required for association with N-cadherin and oc-catenin. J . Biol. Chem., 270:20201-20206. Salanova, M., M. Stefanini, I. De Curtis, and F. Palombi. 1995. Integrin receptor a6B1 is located at specific sites of cell-to-cell contact in rat seminiferous epithelium. Biol. Reprod., 52:79-87. Schwarz, M.A., K. Owaribe, J . Kartenbeck, and W.W. Franke. 1990. Desmosomes and hemidesmosomes: Constitutive molecular components. Annu. Rev. Cell Biol., 6:461-491. Shear, C.R., and R. Bloch. 1985. Vinculin in subacrolemmal densities in chicken skeletal muscle: localization and relationship to intracellular and extracellular structures. J . Cell Biol., 101:240-256. Singer, I.I. 1979. The fibronexus: A transmembrane association of fibronectin-containing fibers and bundles of 5 nm microfilaments in cultured human fibroblasts. Cell , 24:481-492. 2 5 9 Singer, 1.1., D.W. Kawka, D.M. Kazazis, and R.A.F. Clark. 1984. In vivo co-distribution of fibronectin and actin fibers in granulation tissue: Immunofluorescence and electron microscope studies of the fibronexus at the myofibroblast surface. J . Cell Biol. , 98 :2091-2106 . Small, J.V., and A. Sobieszek. 1980. The contractile apparatus of smooth muscle. Int. Rev. Cytol., 64:241-306. Sonnenberg, A., J . Calafat, H. Janssen, H. Daams, L.M.H. van der Raaij-Hemler, R. Falcioni, S.J. Kennel, J.D. Aplin, J . Baker, M. Loizidou, and D. Garrod. 1991. Integrin a6/B4 complex is located in hemidesmosomes, suggesting a major role in epidermal cell-basement membrane adhesion. J . Cell Biol., 113:907-917. Sprando, R.L., and L.D. Russell. 1987a. A comparative study of Sertoli cell ectoplasmic specializations in selected non-mammalian vertebrates. Tissue Cell, 19:479-493. Sprando, R.L., and L.D. Russell. 1987b. Germ cell-somatic cell relationships: A comparative study of desmosome-like and gap junctions during spermatogenesis i n selected non-mammalian vertebrates. Scanning Microsc, 1:1249-1255. Staehelin, L.A. 1974. Structure and function of intercellular junctions. Int. Rev. Cytol., 39 :191-283 . Stanley, H.P., and C.C. Lambert. 1985. The role of a Sertoli cell actin-myosin system in sperm bundle formation in the ratfish, Hydrolagus colliei (Chondrichthyes, Holocephali). J . Morph., 186:223-236. Stepp, M.A., S. Spurr-Michaud, A. Tisdale, J . Elwell, and I.K. Gipson. 1990. a6B4 integrin heterodimer is a component of hemidesmosomes. Proc. Natl. Acad. Sci. USA, 87:8970-8974. Suarez-Quian, C.A., and M. Dym. 1984. Further observations on the microfilament bundles of Sertoli cell junctional complexes. Ann. N. Y. Acad. Sci. 438:476-480. Sugimoto, K., S. Fujii, Y. Ichikawa, and I. Nakamura. 1989. Expression of stress fibers in bullfrog mesothelial cells in situ. Cell Tissue Res., 258:373-380. Sugrue, S.P., and E.D. Hay. 1981. Response of basal epithelial cell surface and cytoskeleton to solubilized extracellular matrix molecules. J . Cell Biol., 91:45-54. Sugrue, S.P., and E.D. Hay. 1982. Interaction of embryonic corneal epithelium with exogenous collagen, laminin, and fibronectin. Role of endogenous protein synthesis. Dev. Biol., 92:97-106. Takeichi, M. 1991. Cadherin cell adhesion receptors as a morphogenetic regulator. Science (Wash. DC), 251:1451-1455. Toyama, Y. 1975. Ultastructure study of crystalloids in Sertoli cells of the normal, intersex and experimental cryptorchid swine. Cell Tiss. Res., 158:205-213. Toyama, Y., T. Obinata, and H. Holtzer. 1979. Crystalloids of actin-like filaments in the Sertoli cell of the swine testis. Anat. R e c , 195:47-62. 260 Toyama, Y. 1986. Actin-like filaments in the Sertoli cell junctional specializations in the swine and mouse testis. Anat. Rec. 186:477-492. Tsukita, S., Y. Hieda, and S. Tsukita. 1989. A new 82-kD barbed end-capping protein (radixin) localized in the cell-to-cell adherens junction: purification and characterization. J . Cell Biol., 108:2369-2382. Turner, C.E., J . Glenney Jr., K. Burridge. 1990. Paxillin: a new vinculin-binding protein present in focal adhesions. J . Cell Biol., 111:1059-1068. Unsicker, K. 1973. Contractile filamentous structures in Sertoli cells of the greek tortoise (Testudo graeca)? Experientia, 30:272-273. van Vorstenbosch, C.J.A.H.V., E. Spek, B. Colenbrander, and C.J.G. Wensing. 1984. Sertoli cell development of pig testis in the fetal and neonatal period. Biol. Reprod., 31 , pp. 565-577. Vogl, A.W. 1989. Distribution and function of organized concentrations of actin filaments in mammalian spermatogenic cells and Sertoli cells. Int. Rev. Cytol., 119:1-56. Vogl, A.W. 1996. Spacially dynamic intercellular adhesion junction is coupled to a microtubule-based motility system: evidence from an in vitro binding assay. Cell Motil. Cytoskeleton, 34:1-12. Vogl, A.W., and L.J. Soucy. 1985. Arrangement and possible function of actin filament bundles in ectoplasmic specializations of ground squirrel Sertoli cells. J . Cell Biol. , 100:814-825. Vogl, A.W., Y.C. Lin, M. Dym, and D.W. Fawcett. 1983. Sertoi cells of the golden-mantled ground squirrel (spermophilus lateralis): A model system for the study of shape change. Am. J . Anat., 168:83-98. Vogl, A.W., L.J. Soucy, and G.J. Lew. 1985. Distribution of actin in isolated seminiferous epithelia and denuded tubule walls of the rat. Anat. Rec , 213:63-71. Vogl, A.W., B.D. Grove, and G.J. Lew. 1986. Distribution of actin in Sertoli cell ectoplasmic specializations and associated spermatids in the ground squirrel testis. Anat. R e c , 215: 331-341. Vogl, A.W., D.C. Pfeiffer, and D.M. Redenbach. 1991a. Sertoli cell cytoskeleton: influence on spermatogenic cells. In: Comparative Spermatology 20 Years After. B. Baccetti, ed. Serona Symposia, Italy, Raven Press. 75:709-716. Vogl, A.W., D.C. Pfeiffer, and D.M. Redenbach. 1991b. Ectoplasmic ("junctional") specializations in mammalian Sertoli cells: influence on spermatogenic cells. Ann. N.Y. Acad. Sci. 637:175-202. Vogl A.W., D.C. Pfeiffer, D.M. Redenbach, and B.D. Grove. 1993. The Sertoli cell cytoskeleton. In: Russell LD, Griswald MD (eds) The Sertoli cell. Cache River Press, Clearwater, Florida, pp. 39-86. Weber, J .E . , L.D. Russell, V. Wong, and R.N. Peterson. 1983. Three-dimensional reconstruction of a rat stage V Sertoli cell: II. Morphometry of Sertoli-Sertoli and Sertoli-germ cell relationships. Am. J . Anat. 167:163-179. 261 Weber, J .E. , T.T Turner, K.S.K. Tung, and L.D. Russell. 1988. Effects of cytochalasin D on the integrity of the Sertoli cell (blood-testis) barrier. Am. J . Anat., 182:130-147. Wong, A.J. , T.P. Pollard, and I. Herman. 1983. Actin filament stress fibers in vascular endothelial cells in vivo. Science (Wash. D. C ) , 219:867-869. Yamada, K.M., and B. Geiger. 1997. Molecular interactions in cell adhesion complexes. Curr. Opin. Cell Biol., 9:76-85. Yazama, F., H. Sawada, K. Hirosawa, Y. Hayashi, and T. Nishida. 1991. Deep-etch visualization of Sertoli cell (blood-testis) barrier in the boar. Tissue and Cell , 23 :235-246 . 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0088784/manifest

Comment

Related Items