Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Characterization of starfish yolk and cortical granule proteins, and of a novel extracellular proteoglycan… Reimer, Corinne L. 1994

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

Item Metadata


831-ubc_1994-894490.pdf [ 6.55MB ]
JSON: 831-1.0099230.json
JSON-LD: 831-1.0099230-ld.json
RDF/XML (Pretty): 831-1.0099230-rdf.xml
RDF/JSON: 831-1.0099230-rdf.json
Turtle: 831-1.0099230-turtle.txt
N-Triples: 831-1.0099230-rdf-ntriples.txt
Original Record: 831-1.0099230-source.json
Full Text

Full Text

CHARACTERIZATION OF STARFISH YOLK AND CORTICAL GRANULE PROTEINS,AND OF A NOVEL EXTRACELLULAR PROTEOGLYCAN IMPLICATED INDIGESTIVE TRACT MORPHOGENESISByCORINNE L. REIMERB. Sc., The University of British Columbia, 1986M. Sc., The University of British Columbia, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Anatomy)We accept this thesis as conforming to the required standard..THE UNIVERSITY OF BRITISH COLUMBIAJanuary 1994© Corinne L. ReimerIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. t is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)__________________________Department of_________________The University of British ColumbiaVancouver, CanadaDateDE-6 (2/88)ABSTRACTExtracellular matrix (ECM) is thought to play a major role in morphogenesis by influencing processessuch as cell migration and differentiation, although the specific mechanisms involved are poorlyunderstood. This study examined the ECM and egg storage granules of starfish (Pisaster ochraceus)embryos, and attempted to identity components important for digestive tract morphogenesis. Threemonoclonal antibodies were developed with specificities for the ECM, yolk and cortical granules inPisaster eggs and embryos. These antibodies were then used to localize, isolate and characterize theantigens through early development, using immunohistochemistry, immunocytochemistry,immunochemical and biochemical techniques. The first antibody, PM1, binds to a large extracellularproteoglycan, which appears in the blastocoel matrix at mid-gastrulation, and is synthesized only byendodermally-derived tissues. The use of PM1 antibody as a function blocking agent in live embryocultures suggested that it plays an important role in digestive tract morphogenesis. A second antibodyrecognizes a protein localized in cortical granules of unfertilized eggs. The majority of these granules arelocated in the peripheral egg cytoplasm and are released at fertilization. However, a secondmorphologically identical population of granules remain dispersed throughout the egg cytoplasm, andappear to contribute to ECMs of the developing embryo, including the blastocoel ECM, basementmembranes, and the hyaline layer. The function of this protein is currently unknown; however, it has adifferent storage and secretion profile from the PMI proteoglycan, suggesting its role in the blastocoelmatrix may be different. A third antibody recognizes proteins stored in yolk granules located throughoutthe egg and cells of the developing embryo, which do not appear to contribute to ECMs duringembryogenesis. Partial biochemical characterizations using the anti-yolk antibody revealed that there areseveral molecular species of yolk proteins present in the oocyte, and that their molecular compositionchanges during embryogenesis. Depletion of the yolk proteins is not significant until the larval stage,suggesting that they do not play a major role until later in development.IllTABLE OF CONTENTSABSTRACT iiTABLE OF CONTENTS iiiLIST OF TABLES viLIST OF FIGURES viiLIST OF APPENDICES bcLIST OF ABBREVIATIONS xACKNOWLEDGEMENTS XIII. INTRODUCTION 1General overview 1PART A: The PM1 proteoglycan In starfish gut morphogenesis 31. Overview 32. Proteoglycans 4(a) Structure of proteoglycans 4(b) Functions of proteoglycans 7(C) Proteoglycans in development 93. Starfish embryos: a model system for studying cell-ECM interactions during morphogenesis 114. Digestive tract morphogenesis in Pisaster ochraceus 125. Statement of the problem (Part A) 15Part B: Yolk and cortical granules In starfish eggs 1 61. Overview 162. Yolk proteins 173. Vitellogenesis 184. Yolk proteins of echirioids 205. Fate of the yolk proteins in sea urchins 206. Functions of the major yolk proteins in the sea urchin 21(a) Nutrient store 21(b) Extracellularfunctions 227. Cortical granules 24(a) General 24(b) Cortical granules of echinoderms 248. Statement of the problem (Part B) 27II. MATERIALS AND METHODS 291. Embryo preparation 292. Fixation and embedding of oocytes and embryos 303. Monoclonal antibody production 30(a) Preparation of the immunogen 33i) Detergent homogenization of embryos 33i Protease inhibitors 33Con A-affinity chromatography 34iv) Immunogen preparation and immunization 38(b) Hybridoma production 38(c) Ascites production 43(d) Purification of PM1 antibody 43(e) 1gM fragmentation 44iv4. immunohistochemistry .47(a) lmmunofluorescence staining 47(b) PM1 and P21 2/PH3C8 double immunofluorescence staining 47(C) Double immunofluorescence with PC3H2 and PY4F8 48(d) Lectin histochemistry 48(e) Immunogold electron microscopy 495. SDS PAGE and Western blotting 49(a) Gels and blotting membranes 49(b) Developmental Western blot analysis 50Sample preparation and SDS-PAGE 50ii) Western blots 51(C) Digestive tract and coelomic fluid analysis 516. Isolation and characterization of antigens 52(a) Immunoaffinity purification of the PM1 antigen 52(b) Immunoprecipitations with PY4F8 53(c) Biochemical Stains 53i) Coomassie Blue 54ii) Alcian Blue 54iii) Periodic acid/Schiff (PAS) 54(d) Sodium periodate treatment 54(e) Lectin labelling of Western blots 557. Analysis of antigens by enzymatic digestion 56(a) Sugar-degrading enzymes 56(b) Endoglycosidase F digestion 58(C) Trypsin 588. In vivo perturbation studies 58(a) PM1 antibody perturbations 58(b) B-D-Xyloside 59(c) Tunicamycin 60ill. RESULTS 6 1PART A: The PM1 proteoglycan in starfish gut morphogenesis 6 11. Immunolocalization of the PM1 antigen 61(a) Immunofluorescence 61(b) Immunogold 642. Developmental distribution of PM1 immunoreactivity 693. lmmunoblot analysis of the PM1 antigen in development 794. Perturbation of development with PM1 antibody 725. Biochemical characterization of the PM1 antigen 77(a) Biochemical stains 77(b) Subunit analysis 80(c) Enzymatic digestions 806. Epitope determination 85(a) Periodate treatment 85(b) Endo F and trypsin treatments 87(C) Analysis with sugar-degrading enzymes 887. Perturbation of development with tunicamycin and 13-xyloside 88(a) Morphological effects 88(b) Effects on PM1 synthesis 91PART B: Yolk and Cortical Granule Proteins 941. Specificity of antibodies against starfish egg granules 942. Ultrastructural analysis of PY4F8 immunoreactivity 94V3. Ultrastructural localization of the PC3H2 antigen in the egg 994. Localization of the PC3H2 antigen in early development 995. Distribution of PY4F8 immunoreactivity during development 104(a) lmmunofluorescence 104(b) lmmunogold 1076. Immunoblot analysis of the PY4F8 and PC3H2 antigens 1127. lmmunoblot analysis of PY4F8-yolk antigens during development 1158. Ontogeny of the major yolk proteins 1159. Lectin characterization of the yolk proteins 12010. Epitope characterization of the PY4F8 yolk antigens 123IV. DISCUSSION 126PART A: The PM1 proteoglycan in starfish gut morphogenesis 1 26(I) Notes on methodology 1 261. Monoclonal antibody production 1262. Antibody perturbation studies 1273. Affinity purification of the PM I antigen 128(II) Biochemical Characterizations of the PM1 antIgen 1 291. Evidence that the PM1 antigen is a proteoglycan 129(a) Localization in the Golgi suggests the PM1 contains sugar residues 129(b) Carbohydrate groups not typical of glycoproteins 130(c) Carbohydrate groups not typical of known vertebrate proteoglycans 131(d) 13-xyloside blocks PM1 proteoglycan synthesis 133(Ill) Role of PM1 proteoglycan In gut morphogenesis 1331. Localization and secretion of the PM 1 proteoglycan in the blastocoel matrix 1332. In vivo perturbation studies with the PM1 proteoglycan 137(IV) Conclusions and further work 1 38PART B: Starfish yolk and cortical granule proteins 1 39(I) Yolk proteins of P. ochraceus 1 391. Specificity of the Pisaster yolk 4F8 antibody 1402. PY4F8 immunoreactivity during development 1413. Ontogeny of the Pisaster yolk proteins 142(a) Identification of a starfish vitellogenin 1 42(b) Effect of 6-mercaptoethanol on PY4F8 immunoreactivity 143(C) Vitellogenin processing and transport 1434. Characterization of the starfish yolk proteins 1445. Utilization of yolk proteins 1466. Conclusions 147(ii) A cortical granule antigen of P. ochraceus 1 481. The PC3H2-antigen is a cortical granule protein 1482. The PC3H2 antigen is released during the cortical reaction 1493. The PC3H2 antigen is present in the early embryo 1503. Conclusions 151Final Summary and Conclusions 1 52REFERENCES 154APPENDICES 172viLIST OF TABLESTable 1. Proteolytic inhibitors 35Table 2. Enzymes and chemicals used in the structural analysis of glycoconjugates 57vIUST OF FIGURESFig. 1. Chemical structures of glycosaminoglycans 5Fig. 2. The early development of Pisaster ochraceus 13Fig. 3. Formation of the fertilization membrane as viewed with Nomarski phase contrastmicroscopy 25Fig. 4. Protocol for monoclonal antibody production against starfish ECM and yolk proteins 31Fig. 5. Con A-FITC staining of the late gastrula embryo (P. ochraceus) 36Fig. 6. Electrophoretic analysis of the Con A-affinity fraction from embryo homogenates 39Fig. 7. lmmunohistochemical analysis of mouse tail bleed 41Fig. 8. Electrophoretic analysis of PM1 pepsin fragmentation 45Fig. 9. Immunofluorescence localization of PM1 antibody in the early larvae 62Fig. 10. Immunogold localization of the PM 1 antigen in the ECM and Golgi of late gastrulaembryos 65Fig. 11. Immunogold localization of the PM1 antigen in mesenchyme and endoderm cells 67Fig. 12. lmmunofluorescence localization of the PM1 antigen in early development 70Fig. 13. Western blot analysis of the PM1 antigen in early development 73Fig. 14. Effect of PM1 antibody on developing embryos in vivo 75Fig. 15. Characterization of the PM1 antigen with Coomassie blue, PAS and Alcian blue 78Fig. 16. Electrophoretic analysis of the PM1 antigen under reducing and non-reducingconditions 81Fig. 17. Electrophoretic analysis of the PM1 antigen after treatment with GAG-degradingenzymes 83Fig. 18. Analysis of PM1 immunoreactivity after treatment with periodic acid, endoglycosidase Fand trypsin 86Fig. 19. Western blot analysis of the PM1 antigen after treatment with GAG-degrading enzymesand fucosidase 89Fig. 20. Effects of tunicamycin and B-xyloside on P. ochraceus development and on PM1synthesis 92Fig. 21. Immunofluorescence localization of yolk and cortical granule antigens 95Fig. 22. Immunogold localization of yolk antigens in the unfertilized oocyte 97Fig. 23. Immunogold localization of the PC3H2 antigen in the unfertilized oocyte 100Fig. 24. Immunofluorescence localization of the PC3H2 antigen in early development 102Fig. 25. Immunogold localization of the PC3H2 antigen in the early embryo 105Fig. 26. Immunofluorescence localization of yolk proteins in the early embryo and larva 108Fig. 27. Immunogold localization of yolk proteins in the gastrula 110vi”Fig. 28. Western blot analysis of yolk and cortical granule antigens in the oocyte 113Fig. 29. Western blot analysis of yolk proteins in early development 116Fig. 30. Western blot analysis of yolk proteins in the adult intestine and coelomic fluid 118Fig. 31. Con A and WGA binding in the unfertilized oocyte 121Fig. 32. Characterization of PY4F8-immunoprecipitate with lectins and endoglycosidase Fdigestion 124LIST OF APPENDICESAppendix 1: Freeze substitution of starfish oocytes and embryos 172Appendix 2: Con A-Sepharose Affinity Chromatography 174Appendix 3: Protein Quantitation 176Appendix 4: Immunization of mice with Con A-specific embryo fraction 178Appendix 5: Hybridoma production for monoclonal antibodies 179Appendix 6: Recloning hybridomas 181Appendix 7: Expansion and freezing of Positive Clones 182Appendix 8: Isotyping monoclonals with the Serotec isotyping kit 183Appendix 9: Photography 184Appendix 10: Preparation of Colloidal Gold 185Appendix 11: Microtitration assay for determination of the correct protein concentration for goldsol stabilization 186Appendix 12: Gold conjugation to Rabbit anti-mouse lgG/M 187Appendix 13: Gradient Gels 188Appendix 14: PM 1 Immunoaffinity Column Preparation and Chromatography 190Appendix 15: Immunoprecipitations 192Appendix 16: Buffers 193xLIST OF ABBREVIATIONSB-xyloside B-D-xylopyranosideB L basal laminaBM basement membraneCAPS 3-(Cyclohexylamino)- 1 -propanesulfonic acidCon A Concanavalin ADMEM Dulbeccos modified Eagles mediumDMSO dimethyl sulfoxideECM extracellular matrixEDTA ethylenediamine tetraacetic acidEGTA [ethyleneglycol-bis-(13-amino ethyl ether) N,N,N,N’-tetraacetic acidEndo F endoglycosidase FFITC fluorescene isothiocyanateGAGs glycosaminoglycansg gravityHAT hypoxanthine aminopterin thymidineHL hyaline layerlgG immunoglobulin G1gM immunoglobulin MMr relative mobilityN-linked asparagine-linked0-linked serine/threonine-linkedPAGE polyacrylamide gel electrophoresisPAS periodic acid SchiffPBS phosphate buffered salinePC3H2 Pisaster cortical 3H2PEG polyethylene glycolPC3H2 Pisastercortical 3H2xiPH3C8 P!sasterhyahne 3C8PM 1 Pisaster matrix-iPY4F8 Pisaster yolk 4F8PMSF phenylmethyldisulfonyl fluoridePVDF polyvinylidene difluoridePY4F8 Pisaster yolk 4F8SDS sodium dodecyl sulfateTBST Tris buffered saline, tweenTEM transmission electron microscopyTBS Tris buffered salineVBS veronal buffered salineVtg vitellogeninWGA wheat germ agglutininxl’ACKOWLEDGEMENTSI would like to gratefully acknowledge the continued support of my supervisor, Dr. Bruce Crawford. Hisguidance, encouragement and demand for excellence in the lab and at the desk has made my graduateeducation second to none. I feel very privileged to have had him as a mentor and friend over the years. Iwould also like to thank the members of my thesis supervisory committee, Drs. Wayne yogi, NellyAuersperg and Ravi Shah, for their commitment to help me through my Ph. D., offering guidance, criticalevaluation of my work, and encouragement. The department of Anatomy at UBC has been a home to mefor several years, and I would like to thank all its members, past and present, for making it a truly remarkableplace to be. I am especially thankful for the leadership of Dr. Charles Slonecker, who was departmenthead for most of my years there. His support of the graduate student program and enthusiasm in teachingwill not be forgotten. Special thanks also go to people close to me without whose support as friends andcolleagues, I most certainly would not have survived the Ph. D. experience, particularly to ZeidMohamedali, Dave Woods, and Steve Cumming. Finally, I am indebted to my parents, for their love andencouragement over the years, and for believing in the path I have chosen to follow. I could not havedone it without them.I. INTRODUCTIONGeneral OverviewMorphogenesis is the process whereby cells and tissues are organized into highly ordered threedimensional structures that make up an organism. Although it is easy to observe the structural changesthat occur as development proceeds, the underlying mechanisms that are responsible for these changesare not readily apparent. During morphogenesis, cells participate in interactions as they organizethemselves into the complex patterns characteristic of organs. These interactions which simply putinvolve cell-cell contact or cell-extracellular matrix (ECM) contact, are necessary for the development andmaintenance of proper tissue architecture. Together, these interactions are responsible for directing cellbehavior during morphogenesis, which includes adhesion, migration, growth and differentiation,ultimately resulting in the emergence of structure in the embryo. The ECM is a major component inembryonic systems, occupying the vast spaces and cavities that are formed from the blastula stageonwards. ECM provides an environment through which cells can migrate, and a substratum for theiradhesion and guidance. In addition, it can affect growth and cell differentiation (reviewed by Hay, 1991;Adams and Watt, 1993; Lin and Bissell, 1993). When investigating the role of ECM in morphogenesis,the developmentally regulated appearance of specific components in the ECM is often taken as anindication that these components have a specific role in the morphogenetic events that are occurring atthat time. The object of many studies has therefore been to study components of ECM duringdevelopment by investigating when they are synthesized, where they are secreted, and whatmorphogenetic events are occurring during this time. In some cases, components appearing inembryonic ECMs are not synthesized de novo, but are derived from maternal stores of the oocyte.Questions then arise as to when, where, and how these components are stored, and how theirdeployment to the ECM regulated. In many cases, it is unclear whether ECM components that are storedin yolk and other granules of the oocyte play a different role than ECM components that are synthesizedduring early development (Alliegro etah, 1992).2Investigations into the details of the mechanisms by which ECM influences cells to form complex tissuesare no simple task. In most cases, these processes occur inside opaque embryos so that it is very difficultto observe the events directly. Most research in this field has therefore focused on describing the stagesof organogenesis using fixed tissue, or alternatively by showing how the behaviour of particular cells andthe molecules they make change under different conditions in tissue culture. While tremendousadvances have been made in reconstituting the natural cell-ECM environments using tissue culture,these model systems fall short of the in vivo situation, as the complete repertoire of ECM proteins and thefactors they bind to are still not known (Passaniti, 1992). In this respect, there has been more and moreattention devoted to studying the mechanisms of cell-ECM interactions in invertebrate organisms, sincethey contain many of the ECM elements found in vertebrates (reviewed in Har-El and Tanzer, 1993), andsince these elements are sometimes easier to study in morphologically simpler organisms, which are oftentranslucent and thus permit the visualization of morphogenesis in vivo. In the present study, efforts weremade to investigate cell-ECM interactions involved in early morphogenesis of the starfish Pisasterochraceus. As little is known about the characterization of the ECM components in this organism, Ichoose to approach these studies by developing monoclonal antibodies against components of theembryonic ECM, which in some cases were also localized in egg storage granules. In this way, thesecomponents could be identified, characterized and their tissue distributions mapped throughembryogenesis. There are two main areas of work which resulted from the investigations, and they arepresented in distinct sections throughout the thesis. Part A documents the work resulting from amonoclonal antibody against an extracellular proteoglycan of the embryo and early larvae, while Part Brepresents the work resulting from 2 antibodies generated against components of the embryo that werealso localized in storage granules of the oocyte.3PART A: The PM1 proteoglycan In starfish gut morphogenesis1. OverviewDuring embryonic development, epithelial rearrangements result in the formation of cavities, which allowspace for the migrations of mesenchyme cells, as well as for movements and foldings of epithelial sheets.Extensive work on many different species has shown that these cavities are filled with an extracellularmatrix (ECM), composed mainly of collagens (types I and II fibrillar, and type IV basement membrane), noncollagenous glycoproteins (fibronectin, laminin, and tenascin) and proteoglycans (heparin sulfate,chondroitin sulfate). Additional proteins include nidogen, and thrombospondin (reviewed by Hay, 1991).These components have binding sites for each other and are capable of forming an interlacing network orweb of molecules (reviewed by Hardingham and Fosang, 1992; Adams and Watt, 1993). That the ECMplays a key role in interactions which occur during morphogenesis, such as providing a mechanicalsubstratum upon which epithelial and mesenchymal cells can adhere and migrate, has been suggested byseveral investigators (Grobstein, 1954; Pierce, 1966; Hay, 1981;Thiery etaL, 1983; Ekblom etah, 1986).However, the mechanisms involved in ceII-ECM interactions have remained largely a mystery untilrecently, when cell-binding sites within individual ECM components and specific cell surface receptorshave been identified (reviewed by Hynes, 1992). Of all the ECM components, proteoglycans are amongthe most poorly understood in their capacity to affect cell behavior during morphogenesis. Because oftheir structural complexity, the functional domains of individual proteoglycans have been fairly difficult tostudy and characterize, and therefore much more has been learned of the role of ECM glycoproteins suchas fibronectin and laminin in morphogenesis. However, it is clear that proteoglycans play an important rolein morphogenesis, and with the recent advent of molecular biological techniques and domain-specificantibodies, this field is rapidly progressing. In this section, proteoglycans will be introduced primarily asmacromolecules of the extracellular environment, and I will discuss what is known of their role indevelopmental processes. In addition, the starfish model system that was used to study proteoglycanfunction during development will be introduced.42. Proteoglycans(a) Structure ofproteoglycansProteoglycans are present in phylogenetically diverse species, being abundant in sponges, the mostancestral of the known metazoan animals (Misevic and Burger, 1990), as well as in sea cucumbers (Vieiraand Mourao, 1988; Kariya etah, 1990), fruit flies (Brower etah, 1987) and cockroaches (Carbonetto etah,1983), in addition to vertebrates. They are structurally very complex, in that each contains a core proteinwith one or more covalently bound glycosaminoglycan (GAG) chains. GAGs are linear polymers ofrepeating disaccharides that contain one hexosamine and a carboxylate and/or a sulfate ester. There arefour classes of GAGs, and their chemical structures are summarized in Fig. 1 (Wight et a!., 1991). Theseinclude hyaluronic acid, chondroitin sulfate and its epimerized homologue, dermatan sulfate, keratansulfate, and heparan sulfate /heparin. Except for hyaluronic acid, all GAGs are synthesized covalentlybound to a core protein. Their attachment to the core protein occurs through a linkage region in asequence which consists of a xylose, galactose, galactose and uronic acid residue, followed by therepeating disaccharide units that make up the GAG chain proper.The core proteins of about 20 proteoglycans have been sequenced and they have been given nameswhich often reflect their biological activities (e.g. aggrecan, which facilitates aggregation of cartilage matrixcomponents) (reviewed by Kjellén and Lindahl, 1991). These proteoglycans are distributed at variouslocations in tissue, some occurring in intracellular locations, and others in cell surface or extracellularmatrices. Those secreted and deposited into the ECM include aggrecan, versican, decorin, biglycan,fibromodulin, and in general are rich in chondroitin and/or dermatan sulfate residues. The basementmembrane proteoglycans, also a secreted ECM, are characteristically rich in heparan sulfate. A secondclass include cell surface proteoglycans, which may be anchored to the plasma membrane through ahydrophobic peptide domain (fibroglycan, syndecan, betaglycan, thrombomodulin, and CD44), by a lipidanchor (glypican) or by association with other membrane proteins (invariant chain). These cell surfaceproteoglycans are also rich in heparan sulfate residues. A third class resides intracellularly in secretory5Fig. 1. Chemical structures of glycosaminoglycansIn this figure, the repeating disaccharide backbone structure of the four classes ofGAGs are shown (from Wight eta!., 1991).(A) Hyaluronic acid (HA), also known as hyaluronan or hyaluronate, has the simplestGAG structure, consisting of an alternating polymer of N-acetylglucosamine andglucuronic acid. A single molecule can have a molecular weight of up to 10 million,which corresponds to approximately 25,000 repeat disaccharides. HA is the onlyGAG which is not synthesized covalently bound to a protein core, nor is it sulfated..(B) Chondroitin sulfate (CS)/Dermatan Sulfate (DS). CS has the same basicbackbone structure as HA but with N-acetylgalactosamine replacing the Nacetyiglucosamine. lndMdual chains are seldom more than 100 kDa, correspondingto approximately 250 repeat disaccharides. The most common sites of sulfation arethe 4 and 6 positions of the N-acetylgalactosamine residue (Fig. 1 B, dotted and solidarrows). Dermatan sulfate, the epimerized form of CS, in which the D-glucuronic acidis converted to L-iduronic acid, can be sulfated at position 2 (Fig. 1 B, asterisk).(C) Keratan sulfate (KS) consists of a backbone structure with alternating Nacetylgiucosamine and galactose residues. Individual chains are seldom more than40 kDa, which corresponds to approximately 80 disaccharide repeats.(D) Heparan sulfate (HS) and Heparin have the same backbone structure, that ofrepeating N-acetylglucosamine and glucuronic acid, which differ from the other GAGsin their xl A linkage. Individual chains are usually below 50 kDa. Less than 50% ofthe N-acetyl groups are converted to N-sulfates in HS whereas usually 70% or moreare converted in heparin.6A. Hyaluronic Acid-1 ,4-gTcUA-f3-1 ,3-gIcNAc-3-B. Chondrojtin/Dermatan Sulfate-1 ,4-gIcUA-J3,3-gaINAc-f3--1 ,4-idoUA-a-C. Keratan Sulfate-1 ,3-gal-f3-1 ,4-gIcNAc-3-0. Heparan Sulfate/Heparin-1 ,4-gIcUA-,4-gIcNAc-a--1 ,4idoUA-a-0AcAcHOAc• OHHOSAc7granules (serglycin, chromogranin A proteoglycan) (Kjellén and Lindahl, 1991; Gallagher, 1989; Esko,1991). Although many other proteoglycans exist, cDNA or genomic clones of the core proteins have notyet been reported, and they are therefore referred to on the basis of their GAG content.(b) Functions of proteoglycansThe structural complexity of proteoglycans arises from the fact that these macromolecules havedifferent core proteins and different numbers and lengths of individual GAG chains. Some also containmore than one type of GAG chain; for example, syndecan, aggrecan and serglycin. Yet other containasparagine (N)-linked oligosaccharides typical of glycoproteins, and serine/threonine (0)-linked glycansfound in mucins, in addition to their extensive GAG chains (reviewed by Jackson eta!., 1991). Thesestructural complexities enable proteoglycans to take part in a wide variety of biological functions, only a fewof which are understood. Some of these include:(1) Providing resilience to tissues. GAG chains, with their highly charged sulfate and carboxylate groupsdominate the physical properties of the protein to which they are attached. Proteoglycans in the ECMthus function physically as creators of a water-filled compartment; their high fixed negative charge attractscounter ions, and the osmotic imbalance caused by a local high concentration of ions draws water from thesurrounding areas (Hardingham and Bayliss, 1990).(2) Storage sites for growth factors. Several proteoglycans have been shown to bind growth factors,including those with chondroitin sulphate chains, which bind to Platelet factor 4 (Penn et aL, 1988), as wellas betaglycan and decorin, which bind TGF-B (Andres et aL, 1989; Yamaguchi et aL, 1990). This suggeststhat proteoglycans are indirectly able to modulate the activity of growth factors, and may provide localtissue-bound reservoirs of growth factors. For example, in Chinese hamster ovary cells, the binding ofTGF-f3 by decorin directly neutralizes the activity of the growth factor (Ruoslahti and Yamaguchi, 1991).8(3) Matrix organizers. It is likely that proteoglycans play a role in the assembly of other ECM components,and there are several indications that proteoglycan-collagen interactions are important in the regulation ofcollagen fibrillogenesis and matrix assembly (Hascall and Hascall, 1981; Hardingham and Fosang, 1992).Studies have shown that when corneal dermatan sulfate proteoglycan is biochemically altered, theorganization of corneal stroma was disrupted, including focal alterations in collagen I ibril packing and adisruption of lamellar organization (Hahn and Birk, 1992). Proteoglycans also link other matrix proteinstogether, and are often found as a component of ECM networks. The large cartilage proteoglycanaggrecan exists as huge multimolecular aggregates comprising numerous proteoglycan monomers whichare non-covalently bound to hyaluronan and are stabilized by a glycoprotein (link protein) (Hardinghamand Bayliss, 1990).(4) Substratum for migrating cells. Several studies indicate that extracellular and cell surfaceproteoglycans are important for cell migration. This is supported by the findings that migrating endothelialcells exhibit increased chondroitin sulfate and dermatan sulfate proteoglycan synthesis as compared tosessile cells in vitro (Kinsella and Wight, 1986). Another proteoglycan Perlecan, which is associated withlaminin in the basement membrane, has been shown to promote neurite outgrowth (Hantaz-Ambroise eta!., 1987). In addition, the use of mutant Chinese hamster ovary cells that are defective in GAG synthesishave provided insight into the mechanisms of cell surface GAG chains in adhesion. Studies have shownthat cell surface heparin sulfate proteoglycans can mediate attachment to type V collagen,thrombospondin and fibronectin (reviewed by Esko, 1991). Several cell surface vertebrate proteoglycanshave also been implicated with cell migration, including those of the syndecan family, thrombomodulin,and CD44 (Hardingham and Fossang, 1992). In addition, primary mesenchyme cell migration in sea urchinembryos is blocked when proteoglycan synthesis is disrupted (Akasaka et a!., 1980; Solursh eta!., 1986).This could be due to a disruption of a chondroitin sulfate/dermatan sulfate proteoglycan present on seaurchin primary mesenchyme cells, which has been implicated with cell migration (Lane and Solursh, 1991).9(5) Growth factor activity. Epidermal growth factor-like motifs in some proteoglycans (i.e. aggrecan) mayfunction in modulating the proliferative and metabolic activities of chondrocytes and tibroblast(Hardingham and Fosang, 1992), as similar functions have been attributed to glycoproteins such aslaminin and tenascin bearing these motifs (Engel, 1991).(c) Proteoglycans in developmentProteoglycans as a class of macromolecules have been identified and studied in a variety ofdevelopmental systems, although clear evidence for the functions of specific proteoglycans indevelopmental processes remains sparse. Several experimental approaches have been used todetermine the roles that proteoglycans play in development. The most common of these involves theexamination of temporal and spatial changes in proteoglycan composition and/or distribution duringdevelopment. Several different systems have been investigated, including human bone formation, inwhich the distribution of biglycan and decorin were studied using antibodies and cDNA probes (Bianco eta!., 1990); chick neural crest development, in which antibodies were used to study the distributions ofcytotactin and its chondroitin sulfate proteoglycan ligand (Tan et a!., 1987; Hoffman et a!, 1988; Perris eta!., 1991); and rat skin formation, in which biochemical quantifications of chondroitin sulfate proteoglycanrevealed a striking decrease in concentrations during the transition from the fetus to the newborn(Habuchi et aL, 1986).Other studies aimed at understanding the role of proteoglycans in development have involvedphysically changing their properties and then examining the developmental consequences of thesealterations. The most common of these have involved the use of the chemical (3-D xyloside, which inhibitsnormal synthesis of those proteoglycans with 0-linked xylose-mediated GAG chains (all but keratin sulfateproteoglycans). The 13-xyloside competes with core protein xylosides for GAG chain attachment, therebycreating a free xyloside-GAG molecule (Okayama etaL, 1973; Galligani et aL, 1975). Numerous studieshave examined the effects of 13-xyloside on developmental processes, including chick feather formation10(Goetinck and Carlone, 1988), murine renal development (Platt et al., 1987; Lelongt et a!., 1988), seaurchin mesenchymal migration (Solursh eta!., 1986; Lane and Solursh, 1991) and avian corneal stroma(Hahn and Birk, 1992). Although there is no typical morphogenetic response to this treatment,development and/or matrix structure is affected in all cases, which suggests that proteoglycans play animportant role in development.In echinoids, proteoglycan function has also been studied by culturing the embryos in sulfate-free seawater. Since all GAGs with the exception of hyaluronic acid contain sulfate groups, this treatment isthought to alter the normal synthesis of proteoglycans. Sulfate deprivation results in the reduction in a 15-30 nm diameter granular component present in the blastocoel and basal lamina of sea urchin embryos(Katow and Solursh, 1979). Scanning electron microscopical investigations of sulfate-deprived embryoshave also revealed an inhibition of pseudopodia formation by mesenchyme cells and abnormalarrangements of cells in the blastocoel (Akasaka eta!., 1980), suggesting that some sulfated glycoproteinor proteoglycan conjugates are responsible for normal gastrulation and mesenchyme cell migration in seaurchin embryos (Katow and Solursh, 1981; Akasaka and Terayama, 1983; Venkatasubramanian andSolursh, 1984). Although these experiments provide information about sulfated glycoconjugates duringdevelopment, they do not provide information about specific proteoglycans. One way in which singleproteoglycans can be targeted is by using monospecific antibodies, which can be used to block thefunction of proteoglycans in vivo. This approach has been used often to study the functions ofglycoproteins in development. For example, in early amphibian embryos, fibronectiri appears as acomponent of loose fibnllar matrix which underlies the blastocoel roof during early gastrulation. This matrixnetwork has been shown to be the substratum for migrating involuting mesodermal cells, and if the cellsare deprived of this contact with antibodies directed against fibronectin, the embryos respond with acollapse and wrinkling of the blastocoel roof, which results from an inhibition of mesoderm migration (Leeeta!., 1984; Boucaut et aL, 1984). Studies of this nature have also been carried out in sea urchin embryosto examine the function of the major protein of the external ECM, hyalin. After incubation in sea water11containing a monoclonal antibody against this protein, embryos fail to gastrulate, and are inhibited in armrudiment formation (Adelson and Hymphreys, 1988).A third approach that has been used to study proteoglycans in development processes has made useof mutations that have occurred in organisms. One of the more extensively studied proteoglycans in thisregard is aggrecan, for which several mutants have been described at the level of the core protein, itsglycosylation, and sulfation in the chicken, mouse and turkey. These mutants characteristically havecartilaginous rudiments that are reduced in size, with the more severe affects being present in mutationsoccurring at the level of the core protein or its glycosylation (reviewed by Goetinck, 1985). It thus appearsthat aggrecan plays an important structural role in the development of cartilage. While studies focused atthe level of the gene are quite obviously very powerful, such studies are far less accessible than the moretraditional methods. Additionally, the ground work of identifying and characterizing ECM components hasyet to be accomplished in several organisms that are commonly used for studies in development. One ofthese is the starfish embryo.3. Starfish embryos: a model system for studying cell-ECM interactions duringmorphogenesisStarfish embryos and early larvae offer advantages over many other organisms for the study of cell-ECMinteractions during morphogenesis. During most of their early development they exhibit a relatively simplemorphology consisting of an ectoderm and endoderm, which are separated by a blastocoel containingmesenchyme cells and an extensive gel-like ECM (Strathmann, 1989; Crawford, 1990). This ECM is richin alcianophilic fibers (Crawford, 1989, 1990), similar to those described in vertebrate and sea urchinembryos (Endo and Noda, 1977; Katow and Solursh, 1979; Kawabe et aL, 1981). When combined withtheir optical transparency and the fact that, unlike sea urchin embryos, they do not form spicules, theevents of morphogenesis are easily visualized in this organism. In addition, they can be raised in large12synchronous cultures to obtain the amount of material required for many biochemical studies. Despite themany advantages of starfish for developmental studies we know relatively little about their ECMcomponents. Lectin labelling studies at the light and transmission electron microscope (TEM) level haveshown that this material is rich in carbohydrate moieties, specifically, Con A (Concanavalin A) and Wheatgerm agglutinin binding sites (Reimer and Crawford, 1991; Reimer eta!., 1992). In addition, observationsby Crawford and Crawford (1992) have demonstrated that starfish ECM contains very large (SepharoseCL-2B-excluded) sulfated glycoproteins which do not contain known GAGs typical of vertebrate (reviewedby Hardingham and Fosang, 1992) or other invertebrate (Lane and Solursh, 1991; Kariya et aL, 1990)proteoglycans. Although such elements have been identified, neither isolation of individual componentsnor a functional analysis of the effect of these components on the different morphogenetic events in earlystarfish development has yet been accomplished.4 Digestive tract morphogenesis in Pisaster ochraceusDigestive tract morphogenesis in embryos of the starfish, P. ochraceus, has been describedpreviously (Crawford and Abed, 1983; Abed and Crawford, 1986), and is summarized in Fig. 2.Briefly, it begins with the onset of gastrulation when ectodermal cells at the vegetal region of theembryo begin to ingress into the blastocoel, forming the archenteron or primitive gut endoderm. Asdevelopment proceeds, the archenteron elongates and gives rise to the mesenchyme cells, whichundergo an epithelial-mesenchymal transition and migrate off the expanded archenteron tip. Thesemesenchyme cells migrate through the ECM-rich blastocoel and are believed to re-organize as well assynthesize some elements of the matrix (Crawford and Reimer, unpublished observations). As thearchenteron continues to elongate, a blister of basal lamina (BL) is formed at its tip, which thenextends over to and fuses with the BL of the presumptive stomodeal ectoderm. The newly formedtube of BL acts as a conduit along which endoderm and stomodeal cells appear to migrate, leading to13FIg. 2. The early development of Pisaster ochraceus(A) Embryos form a hollow blastula that is enveloped in a fertilization membrane at 24hours post-fertilization, when grown at 12° C.(B) Gastrulation begins in the hatched blastula, with an invagination of cells at the vegetalregion that form the archenteron (a) or primitive gut endoderm.(C) At 3 days post-fertilization, mesenchyme cells (mc) form off the expanded tip of thearchenteron and migrate through the blastocoel (b).(D) The archenteron continues to elongate, and a blister of basement membrane (bm)forms at its tip; although the basement membrane is not visible, its position can beinferred from the presence of scattered cells on its endodermal surface (D, E-arrows).Concurrently, the archenteron begins to form coelomic pouches (c) at its ends from whicha second population of mesenchyme cells emerge.(E) At a slightly later stage in this side view, while the bm cannot be seen, its position ismarked by the flattened surfaces of cells located within it, and extends over to thepresumptive stomodeal ectoderm.(F) The embryo in “E” turned 90° shows both coelomic pouches (C) which are bulging offthe side of the archenteron, as well as the well developed blister of bm involved in mouthformation..(G) Further development involves segmentation of the gut into the esophageal (e) andstomach (5) regions. Some mesenchyme cells (arrows) have settled on the esophaguswhere they will develop into smooth muscle cells with processes wrapping around theesophagus.(H) A similar stage as (G), but turned 90°. Note the stomach (s), esophagus (e), andcoeloms (C), as well as the mesenchyme cells which are migrating through the blastocoelwith processes extended (arrows).(I) This 8 day-old feeding bipinnaria larvae has a well-developed segmented digestivetract, including a mouth (m), esophagus (e), stomach (s) and intestine (i). Bar = 50 jim.15the formation of a mouth. At the same time, the archenteron forms two pouches at its end, the coeloms,and soon after a second population of mesenchyme cells emerge from this region. These mesenchymecells appear to represent a distinct population, which migrate through the blastocoel and settle on theesophageal region of the developing gut tube, where they attach to the BL and differentiate into smoothmuscle (Crawford, 1990). Further development involves segmentation of the gut into the esophageal,stomach and intestinal regions, and with the formation of the mouth, the feeding bipinnaria larvae isformed. The coelomic pouches detach from the gut and remain suspended in the matrix adjacent to themidline.5. Statement of the Problem (Part A)Although digestive tract morphogenesis of starfish development presents an interesting system inwhich to study cell-ECM interactions for the reasons mentioned above, ECM components of this organismremain uncharacterized. The organization of the matrix does, however, appear morphologically similar toother well characterized matrices, and since several other invertebrates contain ECM componentscommon to many species, it is likely that these components are also present in the starfish embryo. Instudying cell-ECM interactions that occur during starfish development, information pertaining to ECMmechanisms in general development can thus be obtained. The primary objective of this work wastherefore to isolate and characterized a component or components of the starfish ECM that werespecifically involved with digestive tract morphogenesis to attempt to understand the mechanisms bywhich ECM influences cell behavior during development. As previous attempts at identifying starfish ECMin situ with antibodies directed against vertebrate ECM components failed, it was necessary to use adifferent approach in identifying and studying starfish ECM. This involved generating starfish-specificmonoclonal antibodies against a pool of proteins extracted from embryo homogenates, and then usingimmunohistochemistry to screen for a specific component of the ECM in the blastocoel. The antibody wasthen used to describe the synthesis and secretion patterns of the antigen during early starfish embryo16morphogenesis. A further objective was to use the antibody in various biochemical techniques to bothisolate and characterize the antigen. And finally, to address the question of whether the ECM componentwas morphogenetically active during this period in development, the antibody was introduced into livingcultures of starfish embryos as a function-blocking perturbation agent.Part B: Yolk and cortical granules In starfish eggs1. OvervIewThe unfertilized oocyte contains various granules that serve as storage sites for proteins, the majority ofwhich are yolk granules or platelets. These granules contain the yolk proteins, which are derivatives of thelarge precursor, vitellogenin. Vitellogenins are closely related, both functionally and structurally in widelydiveient species, including worms, sea urchins, fish, frogs and chickens (reviewed by Byrne etat 1989).Although yolk proteins of many different species have been studied and characterized, and theirontogeny, transport and uptake mechanisms into the egg, including receptor characterization is known,the details of how yolk is processed and utilized in embryos during development remain obscure. Thefunction traditionally ascribed to yolk is that it serves as a store of nutrients for embryogenesis by providingraw materials (amino acids, carbohydrates, and lipids) for use by the developing embryo (reviewed byWilliams, 1967). However, there is little direct evidence for this proposed function (Byrne et a!., 1989).Another population of storage granules in the egg are the cortical granules which, as their name implies,are located in the cortical egg cytoplasm. In many species, these granules undergo exocytosis atfertilization, and the contents are thought to provide a block to polyspermy (reviewed by Cran and Esper,1990). Other components are poorly characterized and therefore many functions of the cortical granulesremain unknown. While proteins are stored in distinct granules of the unfertilized oocyte, in many cases itis unclear what the significance of this discrete packaging is, and several types of storage granules haveyet to be investigated thoroughly and compared through phylogeny. There have been reports that17protein stored in yolk and cortical granules contribute to ECMs of the early embryo (Outenreath et aL,1988; Gratwohl etah, 1991), although literature is sparse and in some cases contradictory. In the presentstudy, an attempt was made to identify yolk and cortical granules components in the starfish and to studytheir course in early development in order to try to come to a better understanding of their functions.2. Yolk proteinsYolk proteins are abundant in the unfertilized oocyte, and are stored in yolk platelets or granules whichoften comprise the bulk volume of an egg. During embryogenesis in oviparous species, it is thought thatthese proteins eventually undergo catabolism to provide raw materials, such as amino acids,carbohydrates and lipids, for use by the developing embryo, although very little is known about theutilization of yolk material (Schechtman, 1956; Deuchar, 1963; Mancuso, 1964). The majority of proteinsstored in the yolk are phosphoglycoproteins, and are derived from the large precursor vitellogenin (Pan etaL, 1969; Wahli, 1988). In oviparous animals, these proteins are thought to function much like the milknutrient proteins in mammals or the storage proteins of plant seeds (Anderson, 1974). The amount ofyolk present in eggs varies depending on the development strategies or organisms. For example, inmany invertebrates and lower chordates (tunicates) the eggs are smaller and contain a moderateproportion of yolk, (approximately 30% of the total egg volume) (Harvey, 1956). These animals rely onother sources of nutrition in the early stages of development, and are observed to form a larval stagewhich can feed itself fairly rapidly. Development then continues from this mobile self-feeding form(Postlehwait and Giorgi, 1985; Strathmann, 1987). Other organisms such as mammals form a placentawhich supplies food and oxygen for the embryo during its long gestation, and therefore their yolkrequirement is small. Amphibian eggs contain much more yolk than do eggs of Amphioxus or theechinoderms, however these embryos still develop a feeding larval stage which is efficient in obtainingfood (Karasaki, 1963). On the other extreme are the eggs of fishes, reptiles, birds, and most smallerstarfish, in which the majority of the egg volume is yolk. As these animals develop either without a feeding18larval stage or placental attachment, they require a sufficient amount of yolk for nourishment throughoutdevelopment.3. VltellogeneslsVitellogenesis refers to the synthesis and accumulation of yolk proteins in the growing oocytes (Wahli eta!., 1981). As mentioned earlier, vitellogenins are closely related, both functionally and structurally inwidely divergent species, including worms, sea urchins, fish, frogs and chickens (reviewed by Byrne eta!.1989). In all species, vitellogenin synthesis occurs in specific tissues, although this varies among species.Vitellogenin is expressed in the intestine of the nematodes (Kimble and Sharrock, 1983), the intestineand gonads of echinoderms (Shyu et at, 1986)1, the fat-body of female insects (Bownes, 1986)2, andthe liver of vertebrates (Bergink and Wallace 1974; Wang and Williams, 1982; Wahli et at, 1981). Commonto all these major sites of synthesis are their endodermal origin.Vitellogenin expression occurs as a specific response to hormones, specifically estrogen invertebrates, juvenile hormone in most insects, or ecdysone in some dipterans (Wang and Williams, 1982;Wahli and Ryffel, 1985). Often, a complex cascade of hormonal interactions is involved in the pathway; forexample, in the amphibian during mating season, the hypothalamus secretes gonadotrophic releasinghormone, which acts on the pituitary to secrete gonadotrophic hormones. These stimulate follicle cells tosecrete estrogen, which then act on the liver to activate the synthesis and secretion of vitellogenin. Theestrogen induces vitellogenin at both the transcriptional and translational levels, and stabilizes thevitellogenin mRNA, increasing its half-life from 16 hours to 3 weeks (reviewed by MaIler, 1985).1Echinoderms appear to be unique in that at least in the sea urchin, Vtg is synthesized by bothsexes and not only in the intestine but also in the gonads.2Drosophlla yolk proteins are synthesized in the ovary in addition to the fat body.19Following synthesis, the vitellogenins are secreted into the body fluid, blood, hemolymph or coelomicfluid, depending on the organism, and are then transported to the ovary where they are taken up by thedeveloping oocyte. This involves the clustering of proteins to specific receptors on the egg plasmamembrane to form coated pits, which are then endocytosed to become coated vesicles (Goldstein et aL,1979). In the bird, amphibian, fish and insect, it has been shown that the coated vesicles deliver thevitellogenins to yolk granules where they are stored for utilization later on in embryonic development(reviewed by Byrne eta!., 1989).After synthesis in the rough endoplasmic reticulum, vitellogenins generally undergo various post-translational modifications (e.g. glycosylation, phosphorylation, proteolytic cleavage) prior to theirsecretion, during their transit to the oocyte, and/or following their selective uptake (Wang et a!., 1983;Sharrock 1984). Once in the oocyte they are cleaved, usually via acid hydrolases such as B-cathepsins, toform the mature yolk proteins of the egg (Medina et aL, 1988). The proteolytic products vary among somespecies in both number and size. In vertebrates, vitellogenin is split into two smaller proteins: the heavilyphosphorylated phosvitin (30-35 kDa) and the lipoprotein lipovitellin (120 kDa) (Wiley and Wallace, 1981).The former has a highly unusual composition, in that more than half of its residues are serine residues,most of which are phosphorylated. These two proteins are packaged together into membrane-boundedyolk platelets (reviewed by Byrne eta!., 1989). Invertebrate mature yolk proteins generally do not containthese phosvitins (Tirumalai and Subramoniam, 1992; Martinez and Wheeler, 1991; Scott and Lennarz,1989), and thus there is an interesting evolutionary pattern which exists and because of it, vitellogeninhas been used extensively for the study of the molecular processes of evolution204. Yolk proteins of echinoldsInformation on the protein composition of yolk or of its utilization is non-existent in the starfish.However, yolk proteins of the sea urchin, a closely related species have been studied rather extensivelyover the past 10 years. In the sea urchin Strongylocentrotus purpuratus, a vitellogenin of 155 kDa issynthesized and post-translationally modified to a 195 kDa form in cells of the adult intestine (Shyu at a!.,1986) and coelomocytes (Harrington and Ozaki, 1986); it is then secreted into the coelomic fluid, andtaken up by the ovary and the oocytes. In the oocyte, the vitellogenin is further modified to a mature 160-180 kDa yolk glycoprotein, where it has been localized to yolk platelets using immunogold electronmicroscopy (Harrington and Easton, 1982; Shyu et a!., 1986; Scott and Lennarz 1989). Scott andLennarz (1989) used a monoclonal antibody raised against the 90 kDa protein from S. Purpuratus to testfor cross-reactivity to other species, including those from the Asteroids, Ophiuroids, and Holothuroids aswell as widely divergent species such as chicken, Xenopus !aevis, and Drosophila. While the antibodycross-reacted with all sea urchin species studied, in no case did the polyclonal antibody recognize anyglycoproteins of the other species. This indicates that although yolk proteins are found in many differentspecies throughout the animal kingdom, they contain epitopes which are highly specific among membersof a phylogenetic class.5. Fate of the yolk proteins In sea urchinsIn many different sea urchins examined, the general pattern which has been observed is adisappearance of the major 160-180 kDa yolk protein of the egg during embryogenesis with theconcomitant appearance of glycoproteins having lower molecular masses. In S. purpuratus, theseglycoproteins have molecular masses of 115, 108, 90, 83 and 68 kDa. In addition, a homologous set ofyolk glycoproteins with similar molecular masses in the embryos of the sea urchins Arbacia punctulata, L.pictus, Hemicentrotus puicherrimus, Anthocidaris crassispina, and the sand dollar Dendraster excentricus21were identified (Ozaki, 1980; Yokota and Kato 1988; Scott and Lennarz, 1989). Studies have found thatthese lower molecular mass proteins are in fact derived from the major vitellogenin by a process involvinglimited proteolysis (Kari and Rottmann, 1985). The initiation of this step-wise proteolysis occurs atdifferent stages of development in the sea urchin, for example proteolysis is initiated at the early blastulastage (12 hours) in S. purpuratus., whereas in L Pictus, it is initiated at the gastrula stage (48 hours)(Scott and Lennarz, 1989).Several observations have indicated that the step-wise proteolysis of sea urchin vitellogenin occurs viathe actions of a cathepsin B-like enzyme, which is activated by acidic conditions (Yokota and Kato, 1988;Mallya et a!., 1992). There is a transient acidification of the yolk platelets, which is sufficient to activate thecathepsin-B enzyme in both S. purpuratus and L. pictus (Mallya et a!., 92). In addition, inhibitors ofcathepsin B activity have been shown to block the degradation of the major yolk protein in the sea urchinS. purpuratus (Mallya et a!., 1992). And finally, a partial purification of a cathepsin B-like enzyme from seaurchin eggs has been reported (Okada and Yokota, 1990).6. Functions of the major yolk proteins In the sea urchin(a) Nutrient storeAs with other organisms, the precise function of sea urchin yolk proteins in embryos during earlydevelopment is unknown, although they are believed to provide an energy source for use during earlydevelopment. A puzzling observation is that despite changes in molecular masses of yolk proteins, nodecrease in the amount of yolk protein in the yolk granules is detected over the course of development.Furthermore, the chemical composition of yolk platelets in terms of phospholipid, triglyceride, hexose,sialic acid, and RNA, remains unchanged from the zygote to the pluteus larva (Armant et a!., 1986). Forexample, the embryo makes the 84 and 65 kDa proteins from the 180 kDa protein without using them fornutrition during its non-feeding (pre-larval) stages (Armant et a!., 1986). Thus, while the constituents of22the yolk granules are proteolytically processed but not subsequently degraded to free amino acids toreplenish the cellular pool, their possible function remains unclear. A similar situation has been observedin frogs, where the yolk platelets remain in much the same condition until their disappearance at the larvalstage (Karasaki, 1963). Additionally, in some insects, yolk proteins are not depleted until just beforehatching (McGregor and Loughton, 1974).To explain these observations, it has been proposed (Armant et a!., 1986) that in the sea urchin, theseproteins may be used as emergency stores of nutrients for the feeding pluteus larva. The proposition,however, was challenged by (Scott et aL, 1990), who showed that starved larvae exhibited only a slightreduction in the amount of yolk glycoproteins and were stunted as compared with larvae given an externalfood source, which developed normally and showed a significant decrease in the level of yolkglycoproteins. Thus, it appears that the yolk may not be serving as a reserve store for larvae, but rather itsutilization appears to be dependent on conditions of growth.(b) Extracellular functionsThere is a growing body of evidence to suggest that yolk proteins become incorporated intomembranes and other extracellular structures of early embryos. One such protein which has beenidentified recently is the sea urchin glycoprotein toposome (Noll et aL, 1985; Mantranga eta!., 1986). It isone of the major components of yolk granules in the sea urchin, and is present as a 22S glycoproteincomplex, consisting of 6-160 kDa subunits which are proteolytically processed during development intofragments ranging from 70-155 kDa. In the egg, toposome is stored in yolk granules and in the centralelectron-dense compartment of the cortical granules, and is also present in the plasma membranes. Atfertilization, the protein stored in the cortical granules is exocytosed and becomes part of a double layerenveloping the hatched blastula on the outside of the hyaline layer, i.e. the apical lamina. The toposomesin the yolk granules are processed by partial proteolysis and secreted in this form to all the externalsurfaces of newly formed cells (Gratwohl et al., 1991). In the blastula, the unprocessed 160 kDa form is23found on the outside of the hyaline layer, outlining the two borders of the apical lamina as well associatedwith the microvilli. All plasma membranes of the developing embryo are covered with toposomesoriginating from yolk granules. By contrast, hyalin, another ECM protein that contributes to the hyalinelayer, is stored exclusively in the homogeneous part of the cortical granules in the oocyte. In the hatchedblastula nearly all hyalin is seen on the inside of the apical lamina, but it does not appear associated withplasma membranes. Cervello and Matranga (1989) have shown that toposome displays the characteristicsof a cell adhesion molecule. They propose that vitellogenin is the primordial form of a cell-adhesionmolecule, and that the biologically active form, toposome, is a cleavage product of it.Reports of yolk proteins appearing in extracellular sites have also been observed in other organisms.Sanders et aL (1990) has used antibodies to identify endogenous lectins in chick embryos and has shownthat these proteins, which are stored in the yolk platelets of unfertilized oocytes, are secreted into theECM at the epiblast stage. Hamazaki et a!. (1989) has shown in the Japanese rice patty fish (Oryziaslatipes) that a glycoprotein present only in yolk granules of spawning female (SF substance) becomeslocalized to the inner layer of the ovarian egg envelope of the growing oocyte. Furthermore, this proteinshares many characteristics of vitellogenin, in that they are both synthesized by the liver and not theooctye, and are transported there through the circulatory system. And finally, Outenreath et aL (1988)have found that an endogenous lectin, XL-43 (from X Iaevis) is present in cortical granules and in vesiclesin the unfertilized oocyte, and becomes localized to the ECM during development, where it may functionas a substrate for cell migration.The results of these studies seem to indicate that the nutritional role traditionally attributed to what hasbeen loosely termed “yolk” must be questioned. The new concept emerging is that much of what hasbeen regarded as yolk may be material stored for the assembly of membranes during periods of rapidembryonic and larval growth, or for extracellular matrices. However, several reports have failed to showthat proteins stored in yolk granules are secreted in the above manner, and in the case of sea urchins, themajor yolk glycoproteins remain within these organelles even after proteolytic processing, and are not24translocated to other organelles such as lysosomes or the cell surface as suggested by Gratwohl et aL,(1991). It appears that the material stored in yolk platelets is perhaps more heterogeneous thanpreviously thought. What is also evident, is that there still remain many unanswered questions on thesubject of yolk protein function and utilization during embryogenesis.7. CortIcal granules(a) GeneralAnother type of granule found in many oocytes including those of mammals, amphibians andinvertebrates are the cortical granules, which, as their name implies, underlie the plasma membrane in thecortical cytoplasm of unfertilized eggs. These are not as numerous as yolk, and generally are very short-lived once development begins. In many organisms, such as sea urchins, frogs and hamsters, theirfunction appears to be restricted to the events occurring at fertilization, where they undergo a triggeredexocytosis and help to create a block to polyspermy. In other species, such as some polychaetes,molluscs, mussels and clams, they are not released upon egg activation, and so perform a differentfunction (reviewed by Anderson, 1974; Gulyas, 1980). Several aspects of their function still remainunknown, largely because the composition of these granules is poorly characterized. They have beenshown to contain fucosyl and sialyl-rich glycoconjugates in the mouse (Lee et aL, 1988) as well as certainproteases and ovoperoxide in the mouse and sea urchin (reviewed by Cran and Esper, 1990). However,the physiological role of these components has yet to be defined. Perhaps the best studied organism inthis area has been the sea urchin, in which the elevation of the fertilization membrane, a directconsequence of cortical granule release, is very pronounced. This phenomenon, which also occurs instarfish (Fig. 3) is unlike the response of the mammalian oocyte, in which case the zona pellucida becomesrefractory to further sperm penetration but does not result in any overt structural change.25Fig. 3. Formation of the fertilization membrane as viewed with Nomarsklphase contrast microscopy.This figure shows the dramatic appearance of the fertilization membrane or envelop afterfertilization in the starfish, P. ochraceus.(A) Immature oocyte, showing an intact germinal vesicle (arrows) which surrounds thenucleus (n).(B) A mature oocyte 5 minutes after fertilization. Note that the germinal vesicle hasbroken down, and that a fertilization membrane is becoming apparent (arrows).(C) One hour after fertilization, the fertilization membrane (fm) is well raised off the surfaceof the egg.(D) Eight hours after fertilization at the first cleavage stage, the appearance of 2 polarbodies (pb) is noted, as well as a prominent fertilization membrane (fm).(E, F) At the 4 cell stage through to the blastula, the embryo continues to develop withinthe confines of the fertilization membrane. Shortly after, the blastula “hatches” out of thismembrane and begins to swim freely. b = blastocoel. Bar = 50 jim.1727(b) Cortical granules of echinodermsThe contents of sea urchin cortical granules are among the best characterized. Some of thecomponents become part of the fertilization membrane (Baginski et aL, 1982), while others become a partof the hyaline layer, the ECM associated with the apical ectoderm of the embryos. One of these proteinsis hyalin, which is stored both in the cortical granules as well as in low density granules dispersedthroughout the egg (Hylander and Summers, 1982; Gratwohl et al., 1991), and which is the majorcomponent of the embryonic hyaline layer. Another protein stored in both the cortical and yolk granules istoposome, which is deployed to the hyaline layer shortly after fertilization, and which is also associatedwith plasma membranes of the egg and early embryo (Gratwohl et aL, 1991). Starfish cortical granuleshave been identified morphologically, and are similar to those of sea urchins, in that they have a veryelectron dense component and a varied morphology. Several histochemical studies have been carriedout on cortical granules in starfish (Crawford and Abed, 1986, Sousa and Azevedo, 1989) which haveshown that they contain acid mucopolysaccharides and acid phosphatase components that areexocytosed upon fertilization. However, information about their protein constituents and how this relatesto other species is sparse.8. Statement of the problem (Part B)Yolk proteins are highly conserved through phylogeny, which suggests that they have an essentialfunction to play during development and morphogenesis, yet the mechanisms of yolk utilization stillremain a mystery. Some reports indicate that yolk is progressively degraded through embryonicdevelopment, while others indicate that depletion occurs later in development when the organism hasreached a feeding larval stage. Yet there are other reports that some of the proteins stored in yolk are notcatabolized for nutrition, but are secreted into ECMs of the developing embryo. The primary objectives ofthis part of the study were: (1) to investigate the yolk proteins of the starfish P. ochraceus, and determinewhether they are catabolized for nutrition during early development; and (2) to investigate whether there28is any evidence that ECM proteins of the early embryo are derived from maternal stores in the oocyte, andif so, determine whether they are stored in the yolk granules and/or other storage granules. The approachused was similar to that in Part A. This involved generating antibodies against proteins of the gastrulastage embryo, and then using immunohistochemistry to localize components of yolk granules and otherstorage granules of the embryo and oocyte. The antibodies were also used to carefully trace the patternof distribution of these antigens throughout development using both immunofluorescence andimmunogold TEM. A final objective involved the use of the antibodies to characterize the antigens, sothat they could be compared with known yolk and other components of oocyte storage granules.29II. MATERIALS AND METHODS1. Embryo preparationRipe adult starfish (Pisaster ochraceus) collected from the intertidal zone near Sidney, B.C. weremaintained in sea water tanks (12°C) in the Department of Zoology, University of British Columbia. Theadults were kept under conditions of constant light in order to prevent spawning. Embryo cultures wereprepared and maintained as previously described (Crawford and Abed, 1983). Sea water used forculturing was collected from the Department of Fisheries and Oceans in West Vancouver, and was filteredthrough a Whatman #1 filter and aerated prior to use. Ovaries removed from the adult female by excising 1arm were placed in 0.1 mg/mI 1-methyl adenine in sea water at 12°C so as to induce the oocytes tocomplete meiosis so they would be ready for fertilization. After approximately 70 minutes, the breakdownof germinal vesicles was complete, indicating that the oocytes had reached full maturity, and they werewashed in aerated filtered sea water. Testes isolated from adult male starfish, also by excising 1 arm, wereplaced in a plastic petri dish and kept dry until the eggs were ready for fertilization. A few drops of spermwere then place in 20 ml sea water to make a cloudy suspension, and alter a few minutes, the sperm werechecked for motility by microscopic observation. Sufficient eggs to cover 1/2-3/4 of the bottom of a 1 literplastic beaker were fertilized with 0.5 ml dilute sperm solution in 400 ml sea water. Following hatching atapproximately 24 hours, the swimming embryos were poured into new beakers to separate them from thedebris on the bottom of the beakers. Embryos at various developmental stages were harvested by gentlecentrifugation (125 x g for 3 minutes), the majority of sea water was removed, and the embryos were eitherstored at -20°C for further use in biochemical studies, or fixed by freeze-substitution forimmunohistochemical studies.302. Fixation and embedding of oocytes and embryosIt was crucial for the success of this study to combine methods that maintained maximum tissuestructure and immunoreactivity. For this reason, material was fixed by freeze-substitution according to themethod of Campbell et aL (1991), which is detailed in Appendix 1. Embryos of various developmentalstages were cryoprotected for 30 minutes in 15% 2,3-butanediol (Sigma) in sea water, and thenapproximately 1 of thick suspension was placed on 50 mesh nickel TEM grids. In the case of immatureand mature oocytes, the cryoprotectant was omitted as it caused excessive tissue collapse. Residualliquid was removed from the embryos with filter paper, and the grid together with the embryos was quicklyplunged into liquid propane that had been pre-cooled to -196°C with liquid nitrogen. The frozen embryoswere then placed in either absolute ethanol or into Alcian blue (Marivac)-saturated ethanol which had beenpre-cooled to -90°C with a dry ice-acetone bath. They were maintained at low temperature (below -85°C)for a period of 5 days with daily additions of liquid nitrogen, during which time the ethanol substituted forthe water. Subsequently, they were slowly warmed to room temperature and embedded into a plasticresin, either JB4 (Polysciences) or LR White (JBS), as per manufacturers instructions. Material embeddedin JB4 was sectioned at 1-1.5 jim and stained for immunofluorescence as described below. Material in LRWhite was sectioned for TEM at 50-60 nm and stained using the colloidal gold technique, also asdescribed below.3. Monoclonal antibody productionIn the following section, the steps required for immunogen preparation and monoclonal antibodyproduction and purification are detailed (summarized in Fig. 4). This involved: 1) obtaining a detergentextract of proteins from starfish embryos; 2) selecting for an ECM-rich fraction of the extract which wasused to immunize mice for monoclonal antibody production; 3) screening the hybridoma supernatantusing thin section immunohistochemistry.31Fig. 4. Protocol for monoclonal antibody production against starfish ECMand yolk proteinsEmbryos at the late gastrula stage were homogenized in a detergent buffer, and theextracted proteins were passed over a Con A-Sepharose affinity column to obtain aglycocorijugate-enriched fraction from the embryo homogenate. The Con A-bindingfraction was released from the column using the specific competitor x-methyl mannoside.The fraction was then used to immunize mice for monoclonal antibody production. Tailbleeds and hybridoma supernatants were tested for antibodies directed to the ECM andegg proteins using immunofluorescence on thin plastic sections of fixed embryos.323. Immunization4. Hybridoma Production0005. Screening with inimunofluorescenceFITC1. Detergent Extraction of Embryos2. Con A-Sepharose FractionationPeak I: Non-binding fractionPeak 2: Con A-binding fraction1FractionsApply a methyl rnannoside33(a) Preparation of the immunogen1) Detergent homogenization of embryosAll steps of the embryo homogenization and extraction were performed on ice to retard the activity ofprotease action. Starfish gastrulae which had been reared, concentrated and stored at -20°C, asdescribed above, were thawed quickly and solubilized with an equal volume of the extraction bufferconsisting of 20 mM Tris, 0.5 M NaCI, pH 7.4 with 1% Brij 56 [polyoxyethylene 10 cetylether (Sigma)]. Theembryos were homogenized first using a Dounce tissue homogenizer (10 plunges), followed by a briefsonication with a Fisher probe sonic dismembrator (15 seconds at 45%). The homogenate was thenextracted for 30 minutes on ice, centrifuged at 35,000 rpm (100,000 g) for 1 hour at 4°C, and thesupernatant was collected and either stored at -70°C or used immediately.Brij 56, a non-ionic detergent, was chosen to aid membrane breakup during the homogenization.Although other detergents such as Triton X, Nonidet P-40 and the Tween series, are often used as mildreagents in homogenization buffers, all of these detergents have a high absorbance at 280 nm due to thepresence of phenol rings. These detergents are therefore incompatible with chromatographicprocedures in which protein monitoring off the column uses UV absorbance at a 280. Brij 56, although amild non-ionic detergent, does not contain phenol rings and therefore does not interfere with proteinmonitoring if this method of detection is used.ii) Protease inhibitorsSince the release of intracellular proteases is unavoidable after ultrasonic treatment, protease inhibitorswere an essential addition to all buffers which were used for embryo homogenization and extraction. Theprotease inhibitor cocktail used in most of the studies included PMSF (phenylmethylsulfonylfluoride),active against serine proteases at concentrations of 1 mM, pepstatin A, active against acid proteases atconcentrations of 1 .tg/ml, iodoacetamide, active against covalent thiol proteases at concentrations of 10mM, and the potent metalloproteinase inhibitors EDTA (ethylenediamine tetraacetic acid) and EGTA[ethyleneglycol-bis-(f3-amino ethyl ether) N,N,N’,N’-tetraacetic acid] active at concentrations of 1-5 mM34(see Table 1 for substrate specificities of protease inhibitors). However, homogenization, extraction, andrunning buffers that were used for Con A-affinity chromatography did not include the metalloproteinaseinhibitors EDTA or EGTA, as Con A requires calcium ions for binding activity.iii) Con A-affinity chromatographyA large number of ECM components contain complex carbohydrate chains which have a high content ofmannose residues at their terminal end. Examples include glycoproteins such as laminin, fibronectin,several collagens, as well as proteoglycans, many of which contain N and 0-linked oligosaccharide chainsin addition to GAG chains. As Concanavalin A (Con A) has a very strong affinity for mannose residues, it isoften used as a general marker of such carbohydrate groups. To determine whether starfish embryonicECM contained Con A-binding sites, sections of late gastrula-stage embryos were stained with FITC-ConA (Fig. 5). Positive staining was observed in large granules of all cell types, as well as throughout the ECMrich blastocoel and over basement membranes. For this reason, it was chosen to obtain an ECM-ennchedfraction from the embryo extract using solid phase affinity chromatography. The details of this procedureare listed in Appendix 2. A 10 ml Con A-Sepharose 4B (Sigma) column was pre-equilibrated with 10column volumes of washing buffer (20 mM Tris-HCI, 0.5 M NaCI, pH 7.4), and pre-washed with 3 columnvolumes of the elution buffer (0.5 M cz-methylmannoside in the washing buffer). For chromatography, theembryo extract prepared as described above [section 3(a)i] was first dialyzed against 20 mM Tris, 0.5 MNaCI, pH 7.4, and then 1 ml was applied to the lectin column. The sample was recycled through thecolumn 2 times, and then washed through with 10 column volumes of washing buffer. The bound fractionwas released from the Con A-Sepharose with the addition of 3 ml elution buffer, and the peak wasmonitored with an a 280 ultraviolet detector (Pharmacia). The fractions under the peak were pooled andconcentrated in dialysis tubing (MW cutoff 12-14,000; Spectrum Medical) that was placed overpolyethylene glycol (PEG) beads (16,000-20,000 MW; Sigma). After a 20-fold reduction in volume wasachieved, the approximate protein concentration of the solution was determined using UV absorption at280 nm (see Appendix 3 for details), and the sample was analyzed with SDS-PAGE (Fig. 6).Table1:ProteolyticInhibitorsinhibitorProteasetargetEffectiveStockCommentsConcentrationEDTAIEGTAMetalloproteinases1-5mM0.5MinH20,pH8.0(divalentcation-dependentproteases)PepstatinAAcidproteases1jImI1mg/mIDMSOC*)01PMSFSerineproteases1mM5mg/mIAcetoneAddfreshateachstep(somethiolproteasesandcarboxypeptidases)(hydrolyzesrapidly)lodoacetamideTriosephosphatedehydrogenase10mM10mg/mIPBSAddfreshateachstep(covalentthiolproteases)36Fig. 5. Con A-FITC staining of the late gastrula embryo (P. ochraceus).A section of a gastrula stage embryo (P. ochraceus) that has been fixed by freezesubstitution in ethanol, embedded in JB4 plastic resin, and sectioned at 1.5 jim. Stainingwas performed using biotin-tagged Con A followed with streptavidin-FITC.(A) Labelling of the large intracellular granules in cells of the ectoderm (ec), endoderm(en) and mesenchyme cells (mc) is evident. Con A also labels the matrix of the blastocoel(b). Bar=36jim.(B) A higher magnification of an embryo prepared as in A, showing a region of the lowergut endoderm and the ectoderm which are separated by matrix in the blastocoel.Labelling of the basement membranes (bm) underlying the epithelium is observed. inaddition, the hyaline layer (hi), which surrounds the embryo on the apical surface of theepithelium, shows binding by Con A. Bar = 10 Immunogen preparation and immunizationFor immunogen preparation, the Con A-bound fraction was dialyzed against PBS (no azide) and theprotein concentration of the solution was adjusted to 4 mg/mI PBS. Equal parts were then mixed withFreund’s incomplete adjuvant (Sigma) to form an emulsion. This was achieved using two, 1 ml glasssyringes that were fitted with a 20 gauge copper connecter. In one syringe, 0.5 ml of the protein solutionwere loaded, while 0.5 ml of the adjuvant were loaded in the other syringe, and the mixture was displacedback and forth through the connecter until a white stiff emulsion was produced. The immunogen wasthen introduced subcutaneously into the backs of 4 week old BALB/c mice (100 jil immunogen permouse, corresponding to approximately 200 jig protein per mouse). One month following the initialimmunization, the mice were boosted by intraperitoneal injection with 100 .tg of concentrated proteinsolution containing no adjuvant. Five days following the boost, a test bleed was taken from the tail. Theblood was suspended in 2% blotto at a concentration of 1:10 and 1:100, and tested on thin sections ofJB4-embedded freeze substituted embryos (section 4). Mice were boosted by intraperitoneal injection ofpure protein solution at 1 month intervals until a positive tail bleed was obtained (Fig. 7). The spleen wasthen harvested for hybridoma production 5 days following the final boost. Details of the immunizationprotocol are summarized in Appendix 4.b) Hybridoma productionMonoclonal antibodies were prepared according to the protocol summarized in Kannangara et a!.(1989), and are detailed in Appendix 5. Confluent cultures (5-6 petri dishes, 9 mm) of mouse myelomacells were prepared and washed in DMEM. The spleen from one mouse was harvested, the cells wereisolated, and mixed together with the myeloma cells. The fusogen was added and the cell mixture waswashed and resuspended in HAT media with 20% FCS. Thymocytes, to be used as a feeder cellpopulation, were harvested and mixed with the fused cell mixture, which was then plated into 24-welltissue culture plates (Linbro). After 3-4 days, 100 jtl HT media were added to each of the wells, and 7 days39Fig. 6. Electrophoretic analysis of the Con A-affinity fraction fromembryo homogenates.The figure shows proteins separated by electrophoresis on 3-10% gradient gels. Lane 1shows the crude detergent extract from starfish gastrulae, while lane 2 shows the Con Aaffinity-purified fraction from the crude extract. In all figure of this thesis, whereappropriate, molecular mass standards indicated on the left are in kilodaltons (kDa).Mnb116—97_L%%66—45-%2420—205—23—41Fig. 7. Immunohlstochemical analysis of mouse tail bleed(A) A mid-saggital section of a plastic-embedded gastrula fixed by freeze-substitution andstained with serum isolated from a mouse that had been immunized with a Con A-enriched fraction of embryo homogenates. Antibody binding to several ECM structures isapparent, including the diffuse matrix of the blastocoel (b) and the hyaline layer (hI), aspecialized ECM found at the apex of the epithelium surrounding the embryo and liningthe digestive tract lumen. Bar = 40 jim.(B) A higher magnification of an embryo prepared as in A, showing part of the ectodermand gut endoderm. Staining of the basement membranes (arrowheads), the hyaline layer(hi) as well as light staining of large granules (9) in the epithelial cells are apparent.Bar=l5jim.LL43post-fusion, the cells were replenished with fresh media. Screening of the hybridoma supernatant wasstarted about 10 days post-fusion, and was done on embryo sections as described in section 4. Desirableclones were selected based on immunolocalization studies in gastrula stage embryos. These were thensubcloned and expanded for further production of hybridoma supernatant (Appendices 6, 7). The cloneswere assayed for their isotype using the Serotec mouse morioclonal isotyping kit. Details of this assay areoutlined in Appendix 8. The following monoclonals were used for this study: PM 1 (Pisaster matrix 1) ofthe 1gM isotype, PY4F8 (Pisaster yolk 4F8), PC3H2 (Pisaster cortical 3H2), and Pisaster 212, all of theIgGi isotype, and PH3C8 (Pisasterhyaline 3C8) of the lgG2a isotype.(C) Ascites productionFor large scale antibody production, the PM1 hybridomas were grown in culture to densities of roughly1 x 106 cells per ml and the cells were collected and washed in PBS (no azide) for ascites production.Ascites tumors were induced in pristane-primed BALB/c mice by the intraperitoneal injection of 5-10 x 106cloned hybridoma cells. Approximately 7 - 10 days after injection, ascites fluid was drained from theperitoneal cavities of the infected mice using 20 gauge sterile needles and gentle belly massage. Themice were put under mild sedation with Halothane for this procedure. The ascites fluid was clarified bycentrifugation (10 minutes, 10,000 x g), and defatted as follows: To 5 ml ascitic fluid were added 5 ml VBS(veronal-buffered saline; 4 mM barbitone, 0.15 M NaCI, 0.8 mM Mg2,0.3 mM Ca 2+), pH 7.2, and 150 mgsilicon dioxide powder. The mixture was incubated at room temperature for 30 minutes with occasionalshaking, and then centrifuged 2000 x g for 20 minutes (Neoh et aL, 1986).(d) Purification of PMI antibodyAfter the PM1 isotype was determined to be 1gM, further purification of the ascites was carried out usingSephacryl S-300 gel filtration chromatography. Initially, conventional size separation chromatography wasperformed, in which tris-buffered saline was the equilibrating and washing buffer. This, however, gave44inadequate peak separation from a-2 macroglobulin, which also eluted in the first peak with 1gM. Inaddition, there was some overlap with peak 2, lgG. Subsequently, the low ionic strength procedure ofBouvet et aL (1984) was used with superior results. In this procedure, a 2.5 x 100 cm column (Pharmacia)was prepared at room temperature and equilibrated with a low ionic strength buffer (5 mM sodiumphosphate, pH 7.5; no azide). Five ml of purified ascites were loaded onto the column, followedimmediately by 50 ml high ionic strength buffer (50 mM sodium phosphate, 2.0 M NaCI, 20 mM sodiumazide, pH 7.5). The proteins were then eluted with the equilibrating buffer (low ionic strength), andfractions of 5.0 ml were collected into tubes containing 250 uI high ionic strength buffer. Because of theinsolubility of 1gM in the veiy low ionic strength buffer, it was retained in the column, and therefore it elutedin the final peak, and was well separated from the other proteins in the ascites fluid. The 1gM was thenconcentrated in dialysis bags placed over PEG beads, or alternatively via ultrafiltration (Centricon-50;Amicon).(e) 1gM fragmentationPM1 antibody was digested into smaller lgG-like fragments with an estimated Mr of 180 kDa, using thelow temperature pepsin proteolysis method of Pascula and Clem (1992). Under these conditions,reasonably homogeneous proteolytic fragments can be generated with a single step digestion. Toachieve this, 1 ml of a 10 mg/mI purified 1gM solution was dialyzed against 20 mM sodium acetate, 0.15 MNaCI, pH 4. Crystalline pepsin dissolved in the same buffer was added at a ratio of 1:15 w/w enzyme toantibody, and the 1gM was digested at 4°C for 24 hours. The digestion was stopped by increasing the pHof the solution to inactivate the pepsin with the addition 0.1 ml of 2 M Tris-HCI, pH 8.6, and the fragmentswere separated using gel filtration on a Sephacryl S300 column as above. Sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS PAGE) was carried out under non-reducing conditions to confirmthe digestion of 1gM (900 kDa) to smaller fragments with Mr of 130 kDa (see Fig. 8).45FIg. 8. Electrophoretic analysis of PM1 pepsin fragmentation.The figure shows PM1 antibody, an 1gM class monoclonal with a molecular mass of 900kDa, which has been purified from ascites fluid using gel chromatography, and analyzedby SDS polyacrylamide electrophoresis under non-reducing conditions on 3-10%gradient gels before and after proteolysis with pepsin. Lane 1, undigested PM1, showinga strong band at 900 kDa. Lane 2 shows the separation of PM1 fragments after 24 hourdigestion with pepsin (4°C). In addition to some undigested material at the top of the gel,the predominant band is seen at 130 kDa. Lane 3 shows a 130 kDa enriched traction(arrow) of the digest isolated by gel filtration chromatography; only a small amount ofundigested 1gM is present in this traction.I(Y)cijH,.474. Immunohistochemistry(a) Immunofluorescence stainingThin sections (1-1.5 jim) of JB4-embedded freeze substituted material were stained with themonoclonal antibodies PM1, P212 or PY4F8. The sections were first pre-incubated with 0.05% rabbit lgG(Sigma) in PBS containing 0.2% Carnation non-fat milk powder (PBSIblotto) for 30 minutes. This stepensured that sites prone to non-specific binding were blocked, a technique which reduced thebackground staining substantially. Next, undiluted hybridoma supernatant was applied to the sections for1 hour. After 2 washes of 15 minutes each in PBS/blotto, the sections were incubated for 1 hour in FITCconjugated secondary antibodies that were prepared in rabbits against either mouse 1gM or mouse lgG,depending on the isotype of the primary antibody. The sections were washed as above, mounted with16% gelvatol, 0.4% DABCO (1 .4-diazabicyclo [2.2.2]octane; Aldrich) 30% glycerol in PBS, pH 7.2 (Taylorand Heimer, 1974; Johnson et aL, 1982), and photographed with Fugicolor 1600 film or Kodak TMAX3200 on a Zeiss Axiophot Photomicroscope equipped with epifluorescence optics (see Appendix 9 fordetails). Control sections were stained as above with normal pooled mouse 1gM or lgG replacing theprimary antibody.(b) PM1 and P212/PH3C8 double Immunofluorescence stainingMultiple immunofluorescence labelling was carried out in some cases to compare the localization of twodifferent monoclonals on the same tissue section. This involved the PM1 antibody together with one oftwo different monoclonal antibodies that were developed in this lab: P3C8 which binds to the hyalinelayer, and P212 which binds to a plasma membrane-associated antigen. The sections were first labelledwith the hybridoma supernatants P3C8 or P212 (both lgG class antibodies) for 1 hour and washed asabove, after which biotin-labelled goat anti-mouse secondary antibodies were applied to the sections for 1hour. The sections were washed as above, and incubated with a streptavidin-texas red fluorochrome(Molecular Probes) for 1 hour. Next, the sections were washed and PM1 immunostaining was done as48described above. Sections were washed, mounted and photographed by double exposure on a ZeissAxiophot Photomicroscope as described above. Controls were stained as above with mouse 1gM and lgGreplacing the primary antibodies.(c) Double immunofluorescence with PC3H2 and PY4F8For double label immunofluorescence using 2 monoclonals of the lgG class, the sections were labelledwith PY4F8 hybridoma supernatant for 1 hour and washed as above, after which a biotin-labelled goatanti-mouse secondary antibody was applied to the section for 1 hour. Sections were washed and probedwith a streptavidin-texas red fluorochrome (Molecular Probes) for 1 hour. Next, the sections were washedand stained with PC3H2 hybridoma supernatant as described above, washed, mounted andphotographed by double exposure on a Zeiss Axiophot Photomicroscope. Controls were stained asabove with mouse lgG replacing the primary antibodies.(d) Lectin histochemistrySections were preincubated with 2% blotto for 30 minutes and then were stained for 1 hour with FITCconjugates of the following lectins (Sigma): Con A (Concanavalin A), WGA (wheat germ agglutinin),RCA12O (ricinus communis agglutinin-120), UEA-l (ulex europeus agglutinin) and SBA (soybeanagglutinin). The lectins were diluted in 2% blotto to a final concentration of 200 jig/mI. Sections werethen washed in 2% blotto for 15 minutes, and prepared for photography as described above. Forcontrols, the lectins were first incubated for 1 hour with 1 M concentrations of their target sugars: Thesewere a-methyl mannoside for Con A, N-acetyl glucosamine for WGA, galactose for RCA1 20, L-Fucose forUEA-I, and N-acetyl galactosamine for SBA. The lectins were then were applied to tissue sections asdescribed above.49(e) Immunogold electron microscopyUltrathin sections (50-60 nm) of LR White-embedded material were picked up on parlodion/carboncoated 100 mesh nickel grids and stained as follows. The grids were floated on drops of the preincubation buffer consisting 0110% normal serum of the secondary host (either goat or rabbit serum;Pierce) in PBS/blotto for 1 hour, followed by a 90 minute incubation in undiluted hybridoma supernatant.The grids were then washed 2 times over 30 minutes with PBS/blotto, and floated on a goat anti-mouselgG/M-colloidal gold conjugate that was prepared in the lab (see Appendix 10 for details). This involvedfirst preparing gold particles with a mean particle diameter of 25 nm (Au25) after the method of Frens,(1973), in which chloroauric acid is reduced with sodium citrate to form a heterodisperse gold sol. Thegold particles were then coupled to goat anti-mouse IgGIM after the methods of Slot and Geuze (1985),and is detailed in Appendices 11 and 12. The grids were washed as above, rinsed in distilled water, andstained with saturated aqueous uranyl acetate (10 minutes) and lead citrate (5 minutes) (Reynolds, 1963).Electron microscopy was performed on a Philips 301 TEM. Control grids were stained as above, withnormal mouse lgG or 1gM replacing the primary antibody.5. SDS PAGE and Western blotting(a) Gels and blotting membranesGradient acrylamide gels (3-10%) were used for the majority of the experiments, because they facilitatedthe separation of large antigens better than the low percentage conventional acrylamide gels (detailed inAppendix 13). These gels also offered several advantages for Western blotting. While low percentageacrylamide gels were often prone to adhere to transfer membranes during electro-elution, the gradientgels were less likely to adhere to the membranes. There were still some problems with nitrocellulose,because it adhered to the low percentage regions of the gradient gels. However, for most of theexperiments, the Biorad PVDF (polyvinylidene difluoride) membrane was used and proved superior to50nitrocellulose, offering higher binding affinity of large proteins, higher retention of proteins during elution,and lower background for immunostaining.b) Developmental Western blot analysisi) Sample preparation and SDS-PAGEOocytes and embryos at the following stages were used for the developmental Western blot analysis:Unfertilized immature oocytes, blastulae (2 days post-fertilization), gastrulae (4 days post-fertilization), andbipinnaria larvae (9 days post-fertilization). One ml of packed eggs or embryos were suspended in anequal volume of ice-cold extraction buffer. For PM1 Western analysis, the extraction buffer used was ahigh ionic strength buffer designed to solubilize extracellular proteoglycans, and contained 4 M guanidinehydrochloride, 50 mM sodium acetate, 10 mM EDTA, 1 mM PMSF, 1 mM iodoacetamide, and 1 ig/mlpepstatin A. For the PY4F8 Western analysis, the extraction buffer was designed to solubilize intracellularproteins and those contained within storage granules; it consisted of 10 mM Tris, 50 mM NaCI, 1% TritonX-100 (Sigma), 0.5% sodium deoxycholate (Sigma), 0.1% SDS (BDH), pH 7.3, with the protease inhibitorsmentioned above. Tissue homogenization was performed on ice, first using a Dounce tissuehomogenizer (10 plunges), followed by a brief sonication with a Fisher probe sonic dismembrator (15seconds at 45%). The homogenate was then extracted overnight at 4°C with constant rotation andcentrifuged at 35,000 rpm (100,000 g) for 1 hour in at 4°C in a Sorvall SS65 rotor. The supernatant wascollected, and prepared for electrophoresis by boiling in SDS-PAGE sample reducing buffer at a ratio of1:10. Analysis of the antigens was performed on 3-10% gradient gels in a Biorad vertical mini-slab systemutilizing the buffer system described by Laemmli (1970). For total protein visualization and for PM1Western blot analysis, lanes were loaded with approximately 10 jig/mI protein/lane as determined by theBiorad DC Protein Assay. For PY4F8 immunoblots, lanes were loaded with approximately 5 jig/lane ofreduced sample, or with 2 jig/lane non-reduced sample.51ii) Western blotsGels were equilibrated for 30 minutes in transfer buffer [10 mM CAPS (3-[Cyclohexylamino]-1-propanesulfonic acid; Sigma), pH 11.0], and proteins were electroeluted onto PVDF membrane (Biorad)for 1 hour at 100 v. Membranes were rinsed 15 minutes in distilled water, and then either stained for totalprotein with 0.025% Coomassie Brilliant Blue R-250 (Sigma) in 40% methanol, or processed forimmunoblotting. lmmunoreactive bands were visualized as follows: Membranes were blocked for 1 hourat 37°C in TBST (10 mM Tris, 0.15 M NaCl, 0.2% Tween 20; BDH) with 5% milk powder and 0.5% normalgoat serum. Following this, the membranes were rinsed in washing buffer (TBS with 0.1% milk powder),and were incubated overnight at room temperature with a 1:1 dilution of primary hybridoma supernatant inTBST, with 1% milk powder and 1% BSA. The membranes were then washed 3 times for 15 minutes inthe TBS washing buffer, and incubated for 90 minutes in goat anti-mouse lgG/M-biotin (Pierce) diluted inTBST with 1% milk powder. The membranes were washed as above, and incubated in streptavidin-HRP(Pierce) diluted in TBST with 1% milk powder for a further 90 minutes. After a final wash, immunoreactivebands were visualized by incubating the strips with 4-chloronaphthol (Sigma; 0.03% 4-chloronaphthol in50 mM Tris, pH 7.6 with 10 jl of 30% hydrogen peroxide) at room temperature with agitation for 10-30minutes. Apparent molecular masses (Mr) for SDS-PAGE and Western blots were estimated bycomparison with Sigma high molecular weight standards and mouse Engelbreth-HoIm-Swarm (EHS)laminin (Sigma), the latter of which has an estimated Mr of 900 kDa (non-reducing) and 420/220 kDa(reducing).(C) Digestive tract arid coelomic fluid analysisDigestive tract material was isolated from adult female starfish and suspended in an equal volume ofextraction buffer. Extraction was performed with a tissue homogenizer and with mild ultrasonic treatmentas described in section 3(a), and the supernatant was collected, and prepared for electrophoresis byboiling in SDS-PAGE sample buffer at a ratio of 1:10. Coelomic fluid isolated from adult female starfishafter the removal of 1 arm was concentrated 20-fold via ultrafiltration (Centricon-50) prior to suspension in52SDS-PAGE sample buffer. Analysis of the antigens was performed on 3-10% gradient gels in a BioRadvertical mini-slab system utilizing the buffer system described by Laemmli (1970).6. Isolation and characterization of antigens(a) Immunoaffinitypurification of the PM1 antigenFor further purification of the PM1 antigen, a PM1 immunoaffinity column was prepared following aprotocol described in Appendix 14. Purified PM1 antibody from 2 ml ascites fluid was coupled to 1 ml AffiGel 10 (BioRad) according to manufacturers directions. The antibody was first dialyzed against couplingbuffer (0.25 M sodium bicarbonate, pH 8.7) at 4°C for 24 hours. Affi-gel 10 was then activated by washingwith 10 volumes of ice-cold distilled water; the dialyzed antibody was added quickly to 1 ml of the activatedAffi-gel 10, and the mixture was rotated end over end for 4 hours at 4°C. Following coupling, 0.1 ml oil Methanolamine HCL (pH 8) were added to block any remaining active esters, and the gel was rotated afurther 1 hour at room temperature. The coupled gel was transferred to a 3 ml column made from an emptysyringe barrel, and equilibrated with TBS, pH 7.5. Affinity chromatography was carried out at 4°C asfollows: 1 ml of guanidine hydrochloride embryo extract was dialyzed against TBS overnight, and thenapplied to the affinity column at a flow rate of 5 mI/hour. After extensive washing with TBS, the PM1antigen was eluted off the column using 2 ml 0.1 M triethylamine, pH 11.0. The peak was detected usingUV (a 280) , and was collected in 1 ml fractions into tubes containing 100 tl 1 M Tris, pH 6.0; fractions werepooled, concentrated via ultrafiltration (Centricon-50), and stored at -70°C. The protease inhibitors PMSFand pepstatin A were routinely used throughout the procedure, and PMSF was supplemented every 18hours.53(b) Immunoprecipitations with PY4F8Embryo extracts were immunoprecipitated with PY4F8 (or P212 as a control) hybridoma supernatantusing Protein-A Sepharose 4B (Sigma) and goat anti-mouse lgG (Pierce) as a linker (Appendix 15). Allsteps were performed at 4°C. Briefly, 100 iJ settled Protein A-beads were washed twice for 10 minuteseach with embryo extraction buffer (10 mM Tris, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS,pH 7.3) (pre-wash) and then washed with the immunoprecipitation buffer, consisting of 20 mM Tris, 150mM NaCI, 0.1% Triton X-100, 25 mM EDTA, 1 mM iodoacetamide, 1 mM PMSF, 1 ig/ml pepstatin A, and 1mM EGTA, pH 7.8. Following this, 50 p1 of rabbit anti-mouse lgG linker were added to the beads at aconcentration of 50 jig/mI, and the mixture was rotated end over end for 1 hour. The beads were thenwashed for 30 minutes with 3 changes of buffer, after which 100 p1 hybridoma supernatant were added tothem. They were rotated for 1 hour, and then washed for 30 minutes as above. Qocyte or embryodetergent extract (100 p1) was added along with a fresh dose of PMSF (10 p1 of a 1 mg/mI stock of PMSF inacetone), and the beads were rotated for 1 hour. The final wash was extensive to ensure all unboundproteins were removed from the bead suspension (4 changes of buffer over 30 minutes). The antigenwas separated from the beads by resuspending the bead suspension in 200 p1 of reducing SDS-PAGEsample buffer, and heating them to 85°C for 15 minutes. To collect the sample, the beads werecentrifuged for 1 minute at 125 x g, the supernatant was collected, and stored at -20°C until analysis bySDS-PAGE was required.(c) Biochemical stainsAliquots of affinity-purified PM1 antigen were subject to SDS-PAGE on 3-10 % gradient gels, and werestained with various dyes (see below) for biochemical characterization. All gels were photographed on alight box using Kodak high contrast copy film, and film was developed with D19 full strength for 6 minutes.54i) Coomassie BlueFor total protein detection, gels were placed in Coomassie blue (0.25% in 50% methanol, 10% aceticacid) for 30 minutes, and then destained with several changes of 10% methanol, 10% acetic acid.ii) Alcian BlueFor glycosaminoglycan (GAG) detection, Alcian blue staining using the critical electrolyte method ofWall and Gyi (1988) was used as follows: after SDS-PAGE, the gel was washed exhaustively to remove alltraces of SDS, first with 50% methanol, 7% acetic acid for 1 hour, then with distilled water for 1 hour.These steps were repeated once before the gel was finally transferred to a solution of 0.2% Alcian blue8GX (Marivac), 50 mM MgCI2 in 3 % acetic acid overnight. The gel was then destained with severalchanges of 50 mM MgCl, in 3 % acetic acid. To increase the contrast of the light blue bands, a yellowfilter was used during photography.iii) Periodic acid/Schiff (PAS)PAS stain (Lillie, 1951) was used to detect glycoproteins as follows: After SDS-PAGE, the gel wasfixed overnight in 40% methanol, 7% acetic acid, rinsed in distilled water, and placed in 1% periodic acid in3% acetic acid for 50 minutes. The gel was washed in several changes of distilled water, and placed inSchiff’s reagent in the dark for 50 minutes. Following this, the gel was washed with 3 changes of 0.5%sulfite wash (0.5 g potassium metabisulphite, 5 % N HCI), and then rinsed in distilled water. To increasethe contrast of the pink bands, a green filter was used for photography.(d) Sodium periodate treatmentStrips of PVDF containing affinity-purified PM1 antigen were pre-incubated in TBS with 5% milk powderand 0.5% normal goat serum for 1 hour at 37°C. For the conventional periodate treatment, the method ofWoodward eta!. (1985) was used, in which blots were rinsed in 50 mM sodium acetate, pH 4.5, and thenincubated in 50 mM periodic acid (Sigma) in 50 mM sodium acetate, pH 4.5, for 1 hour at room temperature55in the dark. Control strips were incubated in buffer alone as above. Both experimental and control blotswere then rinsed with acetate buffer for 3 changes of 10 minutes each, and exposed to 50 mM sodiumborohydride (Sigma) in PBS for 30 minutes at room temperature. Alternatively, some strips were treatedwith periodate in a high ionic strength buffer after the method of Scott and 1-larbinson (1968) as follows:the strips were blocked as above, then rinsed in 50 mM sodium acetate, 0.2 M sodium perchlorate (BDH),pH 3.0, and incubated in 50 mM periodic acid in 50 mM sodium acetate, 0.2 M sodium perchlorate, pH 3.0for 24 hours at 37°C in the dark. Control blots were incubated in buffer alone under the same conditionsof time and temperature. Both the experimental and control blots were then rinsed in the high ionicstrength buffer for 3 changes of 10 minutes each, and placed in 50 mM sodium borohydride in PBS for 30minutes at room temperature. Following this, blots from both treatments were rinsed with TBS for 3changes of 10 minutes, and were processed altogether for immunoblotting as above.(e) Lectin labelling of Western blotsStrips of PVDF containing PY4F8 immunoprecipitations were pre-incubated for 1 hour in a blockingsolution to reduce non-specific binding of the lectins. For Con A blots, the block was TBS with 5% milkpowder, however milk powder was not compatible with WGA staining, and therefore 3% BSA was used toblock the blots in this case. After blocking, the blots were rinsed briefly in buffer without blocking agent,and incubated overnight at room temperature in a 1:500 dilution of Con A-biotin (Pierce) in TBST with 1%milk powder, or WGA-biotin (Sigma) in (PBST) PBS with 0.2% Tween 20. Following rinsing for 3 x 10minutes in TBS or PBS respectively, the blots were treated for 1 hour with a 1:500 dilution of streptavidinHRP in TBST with 1% milk powder for Con A-probed blots, or in PBST for WGA-probed blots. After a finalwashing step, the blots were processed for detection using 4 chloronaphthol as described in section 5above.567. Analysis of antigens by enzymatic digestionPrior to all digestions, aliquots of pure PM1 antigen were dialyzed in semi-micro dialysis tubing (MWcutoff 12-14,000; Spectrum Medical) against the appropriate digestion buffer, to ensure that the correctbuffer and pH requirements of the enzymes were satisfied. Protease inhibitors (1 mM PMSF, and 1 jj.g/mimM pepstatin A) were included in each digestion buffer.a) Sugar-degrading enzymesPure PM1 antigen was digested with the following enzymes (for a list of enzymes, their sources andtheir substrate specificities, see Table 2): chondroitinase AC and ABC (0.2 U/mi in 0.1 M Tris, 30 mMsodium acetate, pH 7.3 with 10 mM EDTA) , heparitinase (0.15 U/mi in 0.1 M sodium acetate, 1 mM calciumacetate, pH 7.0), testicular hyaluronidase (20 mg/mi in 0.1 M Tris, 30 mM sodium acetate, pH 5.6), andbovine kidney a-L fucosidase (0.2 U/mi in 0.1 M citrate-phosphate, pH 5.5) (Appendix 16). Followingfucosidase digestion, some samples were dialyzed in 0.1 M Tris, 30 mM sodium acetate, pH 7.0 with 10mM EDTA and further digested with chondroitinase ABC or AC as above. All digestions were carried outat 37°C for 24 hours, except for the heparitinase digestion, which was carried out at 43°C for the same timeperiod. The reactions were stopped by boiling in reducing or non-reducing SDS-PAGE sample buffer for5 minutes.The following substrates were used to check the activity of the GAG-degrading enzymes (Sigma):choridroitin sulfates A and B, heparitin sulfate (bovine intestinal mucosa), and hyaluronic acid (humanumbilical cord). Solutions of 0.1 mg/mi were digested with the enzymes listed above using the sameenzyme concentrations, pH, time and temperature. Following the digestions, 10 il aliquots were spottedonto Whatman #1 filter paper, and allowed to dry for 30 minutes. The filter paper was then stained for 15minutes with 1% Alcian blue in 3% acetic acid, and destained with 3% acetic acid. Control digestsreceived no enzyme treatment but were incubated as above.Table2:EnzymesandchemicalsusedinthestructuralanalysisofglycoconjugatesEnzymeSource1SubstratesReferencesTesticularhyaluronidaseBovinetestesHyaluronate,Meyer,1970chondroitin4and6sulfatesdermatansulfate(inpart)ChondroitinaseABCProteusvulgarisChondroitin4and6sulfatesYamagataetaL,1968(chondroitinABClyase)Dermatansulfateandhyaluronate(slowly)ChondroitinaseACArthrobacterChondroitin4and6sulfatesSuzuki,1972(chondroitinAClyase)aurescenshyaluronate(slowly)HeparinaseIllFlavobacteriumHeparin,heparansulfateLinkerandHovingh,1972(heparitinaseI)heparinum-cc-L-FucosidaseBovinekidneyp-nitrophenylx-L-fucosidelijimaetaL,1971EndoglycosidaseFFlavobacteriumN-linkedhighmannoseandcomplexoligosElderandAlexander,1982meningosepticum8-DnitropyranoxylosideInhibits0-linkedoligoassemblyonproteoglycans2Roden,1980(B-xyloside)TunicamycinInhibitsassemblyofN-linkedoligosaccharidesElbein,19871AllenzymesandchemicalslistedobtainedfromSigma2Anexceptiontothisiscornealkeratansulfate,whichisN-linked.58(b) Endoglycosidase F digestionFor endoglycosidase F (endo F) digestion, aliquots of pure antigen or extracts prepared from gastrulastage embryos were first dialyzed overnight in 0.1 M sodium phosphate, 50 mM EDTA, 0.5% Nonidet P40, 0.1 % SDS, pH 6.1 with protease inhibitors. To 5 jil of the antigen or homogenate was then added 20U/mI endo F, and the digestion was carried out for 18 hours at 37°C (Elder and Alexander, 1982).Samples, including controls which were incubated in digestion buffer without enzymes, were subject toSDS-PAGE on 3-10% gradient gels under reducing conditions, and were transferred to PVDF membraneand probed with PMI or PY4F8 antibodies as described in Section.5.(c) TrypsinPM1 antigen was also digested with trypsin 1:250 (Difco) at a concentration of 1 mg trypsin/pureantigen. The digestion was allowed to proceed for 3 hours at 37°C. Control digests included PM1 alonewith protease inhibitors under the same conditions of time and temperature.8. In viva perturbatIon studies(a) PM1 antibody perturbationsThe in vivo effects of PM 1 antibody on the development of blastulae, early gastrulae and postmesenchyme-stage gastrulae were examined. The embryos were placed into 24-well tissue culture plates(Linbro) containing freshly airated Millipore-filtered (0.22 jim) sea water supplemented with 0.06 mg/mIgentamycin sulfate (Sigma), and containing purified PM1 antibody (whole molecule or fragmented) atconcentrations ranging from 0 jig/mI to 100 jig/mI. Control embryos were grown in sea water containingpooled mouse 1gM or lgG at the same concentrations as was used for the experimentals. The embryoswere grown at 12°C, and the development of both experimentals and controls was monitored withNomarski differential interference microscopy. Some of the embryos were removed from the treated sea59water 4 hours after the addition of antibody, fixed by freeze-substitution in 100% ethanol, and embeddedin JB4. These embryos were sectioned (1-1.5 pm) and were immunolabelled with a secondary antibody(rabbit anti-mouse FITC) to determine if the PMI antibody had entered the blastocoel. The embryos werephotographed at the end of the experimental period using DIC optics as follows: immobilization of theswimming embryos was achieved with a brief fixation in 0.5% glutaraldehyde in 80% sea water, after whichthey were were placed on glass slides with covergiass held off the slide at a distance of 0.5 mm withplasticine. Photographs were taken on Fugicolor 100 film with a 16 x objective, using a Zeissphotomicroscope II fitted with a DIC slider and color prism. At the end of the experimental period, asdetermined when the controls reached mouth formation stage (5 1/2 days), experimental and controlembryos were fixed by freeze-substitution and embedded into JB4 or LR White for morphologicalinvestigations at both the light and electron microscopic level. LM sections were stained withRichardson’s cationic dye (0.5% methylene blue, 0.5% Azure Il, 0.5% borax).(b) 8-D-XylosideNitrophenyl 13-D-xylopyrannoside (B-xyloside; Sigma) was solubilized in 95% ethanol and mixed with2 volumes of PBS, pH 7.4, for a final stock concentration of 20 mg/mI. Embryos at the pre-mesenchymeearly gastrulae stage (3 days) were transferred into sea water containing 5 mM and 0.5 mM concentrationsof 8-xyloside, and development was monitored. Controls consisted of embryos incubated in seawatercontaining 5 mM and 0.5 mM Nitrophenyl (3-D-galactopyrannoside (Sigma) prepared in the samemanner. The embryos were collected daily and transferred into sea water with freshly supplemented 8-xyloside. At 5 1/2 days of development, embryos from the various treatments were photographed withDIC optics as above, and fixed by freeze substitution, embedded into JB4, and sectioned for PM1immunostaining.60(C) TunicamycinThree day embryos were transferred into sea water containing 0.2 and 2 jig/mI tunicamycin (Sigma),prepared from a 1 mg/mI stock solution in DMSO (controls included DMSO alone). The embryos werecollected and transferred into fresh incubation media daily until the controls reached mouth formationstage (5 1/2 days). At this time, embryos were photographed using DIG optics as described above, orfixed by freeze-substitution, embedded into JB4 and sectioned for PM1 immunofluorescence staining,also as described above.61III. RESULTSPART A: The PM1 proteoglycan In starfish gut morphogenesis1. lmmunoiocaIization of the PM1 antigen(a) ImmunofluorescenceImmunofluorescence staining with the PM1 antibody revealed that the PM1 antigen was presentthroughout the blastocoel ECM of early bipinnaria stage larvae (Figs 9A, 9B; green fiuorochrome). Thematrix was very strongly labelled at this late stage in development, and had a granular appearance. ThePM1 label was also present in the digestive tract (endoderm) as well as in the digestive tract lumen wherelabelling was strongest in the esophageal region (Fig. 9A). Coelomic pouches (not shown), which are alsoderived from the endoderm, showed a similar pattern of labelling. The PM1 antibody did not bind to thehyaline layer, the other major extracellular component of the embryo which is located on the apical side ofthe ectodermal epithelium surrounding the entire embryo (Fig. 9A-red), although label was present overregions of the ECM lining the gut, which is continuous with the hyaline layer and which has beendescribed as morphologically similar to the hyaline layer (Crawford and Abed, 1986). This area ofspecialized ECM (hereafter referred to as the gut ECM) displays at least 2 epitopes that are not present inthe hyaline layer, including the PM1 antigen (Fig. 9A) and the PG5F9 antigen (data not shown), indicatingthat although it appears similar morphologically, it is not identical to the hyaline layer. Labelling ofbasement membranes (BM) was not obvious, because it was difficult to distinguish BM labelling fromlabelling of the adjacent ECM at the light microscopic level. In addition to extracellular localizations in theblastocoel and digestive tract lumen, the PM1 antigen was also detected intracellularly in the endodermalcells of the digestive tract, suggesting that synthesis of the antigen might be occurring in these cells (Fig.9B). PM1 labelling in these cells was characterized by brightly stained granules, which were often moretowards the apex of the cells. Labelled granules were also sometimes aligned at the cell-cell interface,suggesting that the material was between the epithelial cells. Mesenchyme cells actively migratingthrough the blastocoel displayed filopodial extensions and were closely associated with the PM 1 antigen-62Fig. 9. Immunotluorescence localization Ct PM1 antibody in the earlylarvae.(A) A 1 1/2 jim median sagittal section through an early bipinnaria larva, which has beenfixed by freeze-substitution into ethanol and embedded in plastic (JB4). The double-labelled section shows all elements of the digestive tract, including the mouth (m),esophagus (e), stomach (s) and intestine (i). PM1 antibody (green) binds to the matrixthroughout the blastocoel (b) and is also found in the lumen of the digestive tract, with anespecially high concentration of label in the esophageal region. No staining of the hyalinelayer (hi) is observed with the PM1 antibody, which in this section has been stained withan anti-hyaline layer antibody (red). Bar = 40 jim.(B) A higher magnification view showing the gut endoderm of a late gastrula embryoprepared as above, and stained with PM1 antibody. The endodermal cells (en) arestrongly labelled and show numerous bright granules in the cell apices. Other PM1positive granules are aligned in rows (arrowheads). Simultaneous observations withphase contrast microscopy (not shown) suggests that they are following the lateral cellularborders. Labelling is also observed over the gut ECM, which is associated with the apicalsurtaces of endodermal cells, and is especially pronounced in the esophageal (e) region.Bar= l5jim.(C) A section through a late gastrula embryo prepared as above, which has been stainedwith both PM1 antibody (green), and with the P212 antibody (red). The latter antibodybinds to an antigen associated with membranes of both mesenchyme and epithelial cells.Mesenchyme cells are actively migrating through the PM1-rich matrix, and also containintracellular PM 1-positive granules (arrow). Bar = 10 jim.64rich matrix, as were the basal surfaces of the epithelial cells (Fig. 9C). Some mesenchyme cells alsoshowed labelling of intracellular granules, indicating that these cells may be synthesizing the antigen aswell (Fig. 9C).(b) Immunogo!dImmunogold staining with the PM1 antibody revealed several aspects of localization, synthesis andtrafficking routes that were not obvious with immunofluorescence. Gold particles were typically not foundover the major strands of blastocoel matrix, but rather on amorphous material that was often associatedwith them. Only a few gold particles were associated with the basal lamina (BL) suggesting that little of theantigen was localized here (Fig. bA, B). Golgi complexes of endodermal cells were typically located in theapex of cells close to the gut lumen. Gold particles were found almost exclusively over the trans-Golgiregion, as well as in Golgi-associated vesicles (Fig. bC, D), indicating that the PM1 antibody mightrecognize a carbohydrate epitope, as carbohydrates are known to be added and modified here (Kornfeldand Kornfeld, 1985). The PM1 antigen was also detected in the gut ECM, which borders the apicalsurface of endodermal cells (Fig. 1 OD). However it was not present in the morphologically similar hyalinelayer, which borders the apical surfaces of ectodermal cells. Gold particles were also present intercellularlybetween the lateral borders of adjacent cells (Fig. hA), suggesting that the antigen may reach theblastocoel by exocytosis at the lateral cell borders below the junctional complexes, followed by migrationbetween the cells and through the BL. The Golgi complexes of mesenchymal cells were also decoratedwith immunogold (Fig. 11 B), indicating that synthesis may also have been occurring in these cells. Controlsections stained as above with mouse 1gM replacing PM1 antibody showed no gold labelling (data notshown).65Fig. 10. Immunogold localization of the PM1 antigen in the ECM andGolgi of late gastrula embryos.In this figure and the one following (Fig. 11), embryos were fixed by freeze-substitution inethanol and embedded in LR White resin. As osmium has been eliminated from thefixation protocol to maintain maximum antigenicity, the visualization of membranes is poor.(A) A TEM showing the basal region of 2 endoderm cells, their accompanying basallamina (bI), and the matrix-filled blastocoel cavity. Although the ECM is heavily labelledwith gold particles, very few gold particles are observed over the bl. Bar = 0.6 jim.(B) A higher magnification of the ECM shown in Figure A, demonstrating that the labellingwith PM 1 is over amorphous regions which are associated with the fibrous meshwork ofmatrix, rather than directly on the major fibers. Bar = 0.2 jim.(C) This figure shows the Golgi apparatus located in the apical region of an endodermalcell. Note that PM I labelling in the Golgi appears to be restricted to the trans-Golgi region(arrows), as well as to Golgi-associated vesicles (arrowheads), suggesting that theantibody may be recognizing a carbohydrate epitope. Bar = 0.3 jim.(D) A section through the apical region of an esophageal endoderm cell of an earlybipinnaria larva, showing the gut ECM which borders the apical surfaces of the endodermcells, extending into the lumen of the digestive tract. Gold particles are observed in theGolgi complex (gc) as well as in associated vesicles, indicating that the PM1 antigen issecreted in vesicles. Some of the PM1 antigen is directed towards the lumen of thedigestive tract, where it is found over the gut ECM (straight arrows). In addition, isolatedaggregates of gold labelled material are also located between the Golgi apparatus and thebase of the cell (curved arrows) suggesting that this material may be moving towards theblastocoel. Bar = 1.0 jim¶rg.•4s’)“.7.•’‘•j:1-$.t4.iJ..F1‘I.:...ts.——.%*)•..••:....I.,..0.4.•.—•...•I;,••.;:SI%:,-i;L‘*‘-;.4.’U¶0,S%•tt:•ç3s.17••..••t.q.F.‘a.V••::‘,.•p•-••.•%?p,•t•tfltOb’Aj.t,.4••t•4•.—w••F%-a.a.a_!11L$z••0r•‘•.,.4—‘Vb4t--J!c!‘::;It..1 rIdt•.‘IC1••••1.Si14Jtlb••Sap:_4i_.R,-...s___rt-••)_I-,.-‘-:“‘-,•44:.)%E:.IJç-:t:k.d[I67FIg. 11. immunogold localization of the PM1 antigen In mesenchyme andendoderm cells.(A) A TEM of the basal region of 2 gut epithelial cells of a gastrula stage embryo,demonstrating that material labelled with PM1 antibody is found in the space between the2 cells. This suggests that transport of the antigen from the Golgi complex to theblastocoel may occur via an intercellular route, with material that is secreted at the lateralcell borders basal to the cell junctions passing between the cells and through the basallamina (bI) to enter the blastocoel. Bar = 1.0 ELm.(B) A TEM of a mesenchyme cell from a gastrula showing gold particles that are alignedalong the Golgi (arrows), suggesting that the PM1 antigen may also be synthesized bythese cells. Bar = 2.0 pm.0•,r4)....4,..tAt’1•‘S.T•,;r?!?‘..‘tf‘tgI:.1.:...•1’t‘t•r.4,‘4,?.692. Developmental dIstribution of PM1 ImmunoreactivltyPM1 immunoreactivity was first detected in the vegetal plate of the hatched blastula, where it waslimited to a few brightly labelled granules in some of the vegetal plate cells, and some diffuse light stainingthroughout the blastocoel (Fig. 12A). Shortly after gastrulation, intracellular staining of the endodermalcells was more obvious, and was primarily found in the apices of these cells (Fig. 12B). As in the blastulastage, very little if any PM1 was detected in the blastocoel cavity of the early gastrula. By mid gastrula, asdevelopment of the mouth and digestive tract progressed, both PM1 staining and newly formedmesenchyme cells were observed in the blastocoel (Fig. 12C). PM1 staining of the blastocoel during thelater stages of gastrulation was dramatically increased; strongly positive material was observed throughoutthe blastocoel, while some staining of the gut ECM was also apparent (Fig. 12D). As stated before, PM1immunoreactivity was found throughout the blastocoel and gut of the early larvae. Throughoutdevelopment, intracellular staining appeared to be restricted to endodermally-derived cells, including thedigestive tract, coelomic pouches and mesenchyme cells. PM 1-positive material was never observed inthe ectoderm or stomodeal cells of the developing embryo/larva. No non-specific staining was observedin any of the normal mouse serum controls (data not shown).3. Immunoblot analysis of the PM1 antigen In developmentAnalysis of the appearance of PM1 in development was also undertaken using Western blots. Whenexamining material extracted from eggs through larval stages of development, a dramatic change in theapparent molecular masses of extractable material was seen over time, probably due to the depletion ofthe large yolk proteins, which are abundant in the early embryo (Fig. 13A) (see Part B). When the same 4developmental stages, unfertilized oocyte, blastula, gastrula and early larva, were probed with PM1antibody, staining was not present in the unfertilized oocyte. A faint band of immunoreactivity at about600 kDawas70Fig. 12. Immunofluorescence localization of the PM1 antigen in earlydevelopment.The figure shows the localization of PM1 immunoreactivity during early developmentusing 1 1/2 im thick sagittal sections through embryos of different stages. Bar, 30 pm.(A) A blastula-stage embryo in which labelling with PM1 is restricted to a few brightlystained granules in cells found in the vegetal plate region (arrowheads), and to some verylightly stained material in the blastocoel (b).(B) An early gastrula-stage embryo, stained as above, in which endodermal cells havebrightly labelled granules located primarily in the cell apices (arrowheads). There is stillalmost no labelled material in the blastocoel.(C) A mid-gastrula embryo in which mesenchyme cell (mc) formation has begun. PM1staining of ECM in the blastula has increased significantly at this stage. Note thatintracellular staining is still restricted to the endoderm (arrowheads).(D) A late gastrula embryo in which mouth formation is almost complete, and archenteronelongation and bending has occurred. The blastocoel is heavily labelled with PM1antibody, and the endodermal cells have many labelled granules, suggesting that this is avery active period in PM1 synthesis,72detected at the blastula stage (Fig. 13B). By the gastrula stage, this band was stained strongly, and at theearly larval stage, the intensity and size of the band increased further, indicating that there was a gradualincrease in the amount of PM1 antigen present during gastrulation and early larval development. (Fig.13B). The same amount of protein was loaded in each lane (10 jig) in order that the concentration of PM1antigen in the egg and embryos could be compared through development. Because the amount of PM1antigen in the embryo steadily increased during development, the material from gastrula and larval stagescontained too much of the antigen for it to be separated optimally on the gel. This resulted in broadlystained, over-developed bands in lanes L and G (Fig. 13B), which none-the-less clearly indicated that thePM1 antigen steadily accumulated through development.4. PerturbatIon of development with PM1 antIbodyThe function blocking experiments in which PM1 antibody was added directly to cultures of embryosinitially involved the use of whole unfragmented antibody. Because of the large size of this antibody (1gMclass, 900 kDa), the experiments were repeated with 130 kDa antibody fragments (see section 3e above);by using the smaller fragments, it was thought that the possibility of non-specific steric-related interactionsoccurring at the site of antibody binding would be diminished. The results of experiments in which wholeand fragmented antibody were used were essentially the same, and the data reported in this section (Fig.14) represents experiments in which fragmented PM1 antibody was used. Since the PM1 antibody wasadded directly to sea water in which early gastrula embryos were swimming, rather than administereddirectly into the blastocoel by injection, it was necessary to determine whether it had successfully enteredthe blastocoel. To achieve this, some embryos were removed from the incubation media 4 hours after theaddition of antibody and stained with a secondary anti-PM 1 antibody. These embryos showed positivestaining in both the area of the blastocoel as well as in the endodermal cells (Fig. 14B), indicating thatindeed the PM1 antibody had entered the embryo. No staining was apparent in embryos treated withnormal mouse 1gM in the incubation medium (Fig. 14C).73Fig. 13. Western blot analysis of the PM1 antigen In early development.Crude embryonic extracts containing 10 jig/mI of protein/lane from either unfertilizedoocytes (0), blastulae (BI), gastrulae (G) and early larvae (L) were separated byelectrophoresis on 3-10% gradient gels. The material was then transferred onto PVDFmembranes and stained either for total protein with Coomassie blue (A), or for PM1immunoreactivity (B). PM1 does not appear to be present in the oocyte (0). It is firstdetected in the blastula stage where a faint band of staining with an apparent Mr greaterthan 600 kDa appears (BI, arrowhead). Following this, PM1 levels increase dramaticallyduring gastrulation (G) and continue to increase during larval development (L).Overloading of the lanes contributes to the smeared bands in lanes G and L. The arrowmarks the origin of the running gel.14A BMrOBIG L OBIGL42O:i-205—.___tzI116—_66—29—- j75Fig. 14. Effect of PM1 antibody on developing embryos in vivo.This figure demonstrates the effect of PM1 antibody on the development of early gastrulaembryos. The antibody was added directly to sea water containing swimming earlygastrula embryos.(A) A 1.5 jim mid sagittal section through a 3 day embryo fixed just prior to the addition ofPM 1 antibody.(B) The embryo has been fixed by freeze substitution 4 hours after exposure to 10 jig/miPM1 antibody, and probed with a FITC labelled anti-mouse 1gM. The presence of label inthe blastocoel demonstrates that the 1gM penetrates into this region. The labelling seenin the endoderm is probably due to PM1 antibody which is bound to material located onthe surface or between the endodermal cells.(C) An embryo grown for 6 hours in the presence of non-specific mouse 1gM (control) forcomparison with Fig. B. The archenteron is more elongated than that of the experimentalembryos. In addition, mesenchyme cells, which are not yet present in the experimentals(B), are beginning to be formed in the blastocoel of the controls.(D) A median sagittal section through a control embryo (treated with mouse 1gM) at 5 1/2days of development. Mesenchyme cells are present in the blastocoel, the coeloms (C)are present, a mouth (m) has formed, the gut has become segmented, and an oral hood(oh) is forming.(E) A median sagittal section through an embryo of the same age as that seen in Fig. D,following treatment with PM1 for 2 1/2 days. Although some mesenchyme cells haveformed, the archenteron has elongated little beyond that seen in Fig. B (4 hours posttreatment) and shows no evidence of segmentation; the mouth and coeloms have alsofailed to form.(F) This embryo is similar to that seen in Fig. E but has been photographed using DIGphase contrast optics; the mouth and coeloms have failed to differentiate and appear tobe represented by a ball of disorganized cells at the tip of the archenteron (arrow). Bar =40 jim.ILrn77Embryos from 3 different developmental stages were treated with PM1 antibody. These included theblastula, early gastrula and mesenchyme-cell gastrula stages. When blastula-stage embryos were grownin the presence of PM1 antibody, they continued to develop normally for 1 day, however shortly afterarchenteron formation, the embryos were affected by this treatment. These embryos resembled earlygastrulae that were grown under the same conditions (Fig. 14). By 5 1/2 days of development, thecontrol embryo had formed mouth and coeloms similar to normal development (Fig. 14D). However, themorphology of embryos treated with 10 pg/mI antibody was very different from the controls. Theseembryos had a shortened unsegmented gut; in some cases the tip of the archenteron remained rounded(Fig. 14E), and in others it appeared to consist of a knot of irregular cells (Fig. 14F). No coelom or mouthformation was evident, and the blister of BM usually seen to extend off the tip of the archenteron waslacking. Although mesenchyme cells did form in these embryos, their arrangement in the blastocoelappeared random, and there was no evidence of mesenchyme cell differentiation. Gastrulae at a slightlylater stage of development, (mesenchyme cell-stage) were incubated with the PM1 antibody. Theseembryos were able to form a digestive tract, however, segmentation and size was altered somewhat,idicating that the effect of the antibody on development at this later stage was not as severe (data notshown).5. BiochemIcal characterization of the PM1 antigen(a) Biochemical stainsAliquots of affinity-purified PM1 antigen were subject to SDS-PAGE and then stained with dyes specificfor proteins and carbohydrates to determine the biochemical nature of this antigen. Although staining ofthe total embryo homogenate was evident in all cases, the pure antigen was stained only with Alcian blue;no Coomassie blue- or PAS-positive material was observed in lanes containing the PM1 antigen (Fig. 15).In addition, the PM1 antigen was not detected with silver staining (data not shown). The Alcian blue692. Developmental distribution of PM1 immunoreactivityPM1 immunoreactivity was first detected in the vegetal plate of the hatched blastula, where it waslimited to a few brightly labelled granules in some of the vegetal plate cells, and some diffuse light stainingthroughout the blastocoel (Fig. 12A). Shortly after gastrulation, intracellular staining of the endodermalcells was more obvious, and was primarily found in the apices of these cells (Fig. 1 2B). As in the blastulastage, very little if any PM1 was detected in the blastocoel cavity of the early gastrula. By mid gastrula, asdevelopment of the mouth and digestive tract progressed, both PM1 staining and newly formedmesenchyme cells were observed in the blastocoel (Fig. 12C). PM1 staining of the blastocoel during thelater stages of gastrulation was dramatically increased; strongly positive material was observed throughoutthe blastocoel, while some staining of the gut ECM was also apparent (Fig. 12D). As stated before, PM1immunoreactivity was found throughout the blastocoel and gut of the early larvae. Throughoutdevelopment, intracellular staining appeared to be restricted to endodermally-derived cells, including thedigestive tract, coelomic pouches and mesenchyme cells. PM 1 -positive material was never observed inthe ectoderm or stomocleal cells of the developing embryo/larva. No non-specific staining was observedin any of the normal mouse serum controls (data not shown).3. Immunoblot analysis of the PM1 antigen in developmentAnalysis of the appearance of PM1 in development was also undertaken using Western blots. Whenexamining material extracted from eggs through larval stages of development, a dramatic change in theapparent molecular masses of extractable material was seen over time, probably due to the depletion ofthe large yolk proteins, which are abundant in the early embryo (Fig. 13A) (see Part B). When the same 4developmental stages, unfertilized oocyte, blastula, gastrula and early larva, were probed with PM1antibody, staining was not present in the unfertilized oocyte. A faint band of immunoreactivity at about600 kDawas70Fig. 12. Immunofluorescence localization of the PM1 antigen in earlydevelopment.The figure shows the localization of PM1 immunoreactivity during early developmentusing 1 1/2 im thick sagittal sections through embryos of different stages. Bar, 30 pPm.(A) A blastula-stage embryo in which labelling with PM1 is restricted to a few brightlystained granules in cells found in the vegetal plate region (arrowheads), and to some verylightly stained material in the blastocoel (b).(B) An early gastrula-stage embryo, stained as above, in which endodermal cells havebrightly labelled granules located primarily in the cell apices (arrowheads). There is stillalmost no labelled material in the biastocoel.(C) A mid-gastrula embryo in which mesenchyme cell (mc) formation has begun. PM1staining of ECM in the blastula has increased significantly at this stage. Note thatintracellular staining is still restricted to the endoderm (arrowheads).(D) A late gastrula embryo in which mouth formation is almost complete, and archenteronelongation and bending has occurred. The blastocoel is heavily labelled with PM1antibody, and the endodermal cells have many labelled granules, suggesting that this is avery active period in PM1 synthesis.IL72detected at the blastula stage (Fig. 13B). By the gastrula stage, this band was stained strongly, and at theearly larval stage, the intensity and size of the band increased further, indicating that there was a gradualincrease in the amount of PM1 antigen present during gastrulation and early larval development. (Fig.138). The same amount of protein was loaded in each lane (10 tg) in order that the concentration of PM1antigen in the egg and embryos could be compared through development. Because the amount of PM1antigen in the embryo steadily increased during development, the material from gastrula and larval stagescontained too much of the antigen for it to be separated optimally on the gel. This resulted in broadlystained, over-developed bands in lanes L and G (Fig. 13B), which none-the-less clearly indicated that thePM1 antigen steadily accumulated through development.4. Perturbation of development with PM1 antibodyThe function blocking experiments in which PM1 antibody was added directly to cultures of embryosinitially involved the use of whole unfragmented antibody. Because of the large size of this antibody (1gMclass, 900 kDa), the experiments were repeated with 130 kDa antibody fragments (see section 3e above);by using the smaller fragments, it was thought that the possibility of non-specific steric-related interactionsoccurring at the site of antibody binding would be diminished. The results of experiments in which wholeand fragmented antibody were used were essentially the same, and the data reported in this section (Fig.14) represents experiments in which fragmented PM1 antibody was used. Since the PM1 antibody wasadded directly to sea water in which early gastrula embryos were swimming, rather than administereddirectly into the blastocoel by injection, it was necessary to determine whether it had successfully enteredthe blastocoel. To achieve this, some embryos were removed from the incubation media 4 hours after theaddition of antibody and stained with a secondary anti-PM 1 antibody. These embryos showed positivestaining in both the area of the blastocoel as well as in the endodermal cells (Fig. 14B), indicating thatindeed the PM1 antibody had entered the embryo. No staining was apparent in embryos treated withnormal mouse 1gM in the incubation medium (Fig. 1 4C).73Fig. 13. Western blot analysis of the PM1 antigen in early development.Crude embryonic extracts containing 10 p.g/ml of protein/lane from either unfertilizedoocytes (0), blastulae (BI), gastrulae (G) and early larvae (L) were separated byelectrophoresis on 3-10% gradient gels. The material was then transferred onto PVDFmembranes and stained either for total protein with Coomassie blue (A), or for PM1immunoreactivity (B). PM1 does not appear to be present in the oocyte (0). it is firstdetected in the blastula stage where a faint band of staining with an apparent Mr greaterthan 600 kDa appears (BI, arrowhead). Following this, PM1 levels increase dramaticallyduring gastrulation (G) and continue to increase during larval development (L).Overloading of the lanes contributes to the smeared bands in lanes G and L. The arrowmarks the origin of the running gel.1442O-dj205— bb_pr29— -_•AMrOBIGLBOBIG L116—66—75FIg. 14. Effect of PM1 antibody on developing embryos In vivo.This figure demonstrates the effect of PM1 antibody on the development of early gastrulaembryos. The antibody was added directly to sea water containing swimming earlygastrula embryos.(A) A 1.5 im mid sagittal section through a 3 day embryo fixed just prior to the addition ofPM 1 antibody.(B) The embryo has been fixed by freeze substitution 4 hours after exposure to 10 jig/mIPM1 antibody, and probed with a FITC labelled anti-mouse 1gM. The presence of label inthe blastocoel demonstrates that the 1gM penetrates into this region. The labelling seenin the endoderm is probably due to PM1 antibody which is bound to material located onthe surface or between the endodermal cells.(C) An embryo grown for 6 hours in the presence of non-specific mouse 1gM (control) forcomparison with Fig. B. The archenteron is more elongated than that of the experimentalembryos. In addition, mesenchyme cells, which are not yet present in the experimentals(B), are beginning to be formed in the blastocoel of the controls.(D) A median sagittal section through a control embryo (treated with mouse 1gM) at 5 1/2days of development. Mesenchyme cells are present in the blastocoel, the coeloms (c)are present, a mouth (m) has formed, the gut has become segmented, and an oral hood(oh) is forming.(E) A median sagittal section through an embryo of the same age as that seen in Fig. D,following treatment with PM1 for 2 1/2 days. Although some mesenchyme cells haveformed, the archenteron has elongated little beyond that seen in Fig. B (4 hours post-treatment) and shows no evidence of segmentation; the mouth and coeloms have alsofailed to form.(F) This embryo is similar to that seen in Fig. E but has been photographed using DIGphase contrast optics; the mouth and coeloms have failed to differentiate and appear tobe represented by a ball of disorganized cells at the tip of the archenteron (arrow). Bar =40 jim.Lm77Embryos from 3 different developmental stages were treated with PMI antibody. These included theblastula, early gastrula and mesenchyme-cell gastrula stages. When blastula-stage embryos were grownin the presence of PM1 antibody, they continued to develop normally for 1 day, however shortly afterarchenteron formation, the embryos were affected by this treatment. These embryos resembled earlygastrulae that were grown under the same conditions (Fig. 14). By 5 1/2 days of development, thecontrol embryo had formed mouth and coeloms similar to normal development (Fig. 1 4D). However, themorphology of embryos treated with 10 jig/mI antibody was very different from the controls. Theseembryos had a shortened unsegmented gut; in some cases the tip of the archenteron remained rounded(Fig. 14E), and in others it appeared to consist of a knot of irregular cells (Fig. 14F). No coelom or mouthformation was evident, and the blister of BM usually seen to extend off the tip of the archenteron waslacking. Although mesenchyme cells did form in these embryos, their arrangement in the blastocoelappeared random, and there was no evidence of mesenchyme cell differentiation. Gastwlae at a slightlylater stage of development, (mesenchyme cell-stage) were incubated with the PM1 antibody. Theseembryos were able to form a digestive tract, however, segmentation and size was altered somewhat,idicating that the effect of the antibody on development at this later stage was not as severe (data notshown).5. BiochemIcal characterization of the PM1 antigen(a) Biochemical stainsAliquots of affinity-purified PM1 antigen were subject to SDS-PAGE and then stained with dyes specificfor proteins and carbohydrates to determine the biochemical nature of this antigen. Although staining ofthe total embryo homogenate was evident in all cases, the pure antigen was stained only with Alcian blue;no Coomassie blue- or PAS-positive material was observed in lanes containing the PM1 antigen (Fig. 15).In addition, the PM1 antigen was not detected with silver staining (data not shown). The Alcian blue78Fig. 15. CharacterIzation of the PM1 antigen with Coomassie blue, PASand Alcian blue.Guanidine extracts of bipinnaria larvae (E) and affinity-purified PM1 antigen (PM1) havebeen separated on SDS-PAGE and stained with Coomassie blue (A), PAS (B) and Alcianblue/magnesium chloride (C). In the case of PAS, EHS laminin (420/210 kDa) was addedas a control (LN). Note that PM1 fails to stain with Coomassie blue and PAS (arrows),whereas both PM1 (arrow) and an equivalent band in the larval extract (E) stain with Alcianblue, suggesting that this molecule is a proteoglycan.a,U)U)cE00C)IwciIICl)C>%0)0a,00U)Ca,a.00wIF.3+LuU)Ca)1; #3’II I I I I I00 U (0 (0100)ON 0 ‘ (0 C1N80stained a single band migrating at 600 kDa, suggesting that the PM1 antigen contains many negativelycharged alcianophilic groups characteristic of proteoglycans and GAGs (Fig. 15C).(b) Subunit analysisTo determine the subunit nature of the PM1 antigen, it was subject to SDS-PAGE under both reducingand non-reducing conditions and then visualized by staining with Alcian blue, and by Western blotanalysis. Fig. 16A shows that the apparent Mr of the PM1 antigen was the same under reducing (R) andnon-reducing (NR) conditions when subject to SDS-PAGE. This would suggest that the PM1 antigen mayconsist of 1 large molecule rather than an aggregate of subunits. In addition, when the pure antigen wastransferred to PVDF membrane and immunostained with PM1 antibody, the major bands present wereagain identical in their electrophoretic mobility (16B), indicating that the antibody is indeed binding to theaffinity-pure antigen. The additional minor band migrating with at 900 kDa present under non-reducingconditions was also present in the control lane, and therefore represented some PM1 antibody (1gMsubclass, 900 kDa) that was eluted off of the affinity column along with the PM1 antigen duringchromatography.(C) Enzymatic digestionsTo further assay for the presence of GAG chains characteristic of proteoglycans, the PM1 antigen wassubject to enzymatic degradation by hyaluronidase, chondroitinase ABC, chondroitinase AC andheparitinase, and was then analyzed for GAG content with SDS-PAGE with Alcian blue staining. Fig. 17shows that in no cases was the staining intensity affected significantly by enzymatic treatment atconcentrations which degrade known standards.81Fig. 16. Eiectrophoretic analysis of the PM1 antigen under reducing andnon-reducing conditions.Pure PM1 antigen was subject to SDS-PAGE under reducing and non-reducingconditions and then stained with Aician blue (A) or transferred to PVDF and stained withPM1 antibody (B).(A) The PM1 antigen has the same apparent Mr under reducing (R) and non-reducing(NR) conditions, indicting that it contain covalently-bound subunits.(B) A Western blot showing a strong band of immunoreactivity at 600 kDa in both thereducing (R) and non-reducing (NR) lanes. An additional minor band apparent in the nonreducing lane (arrow) also present in the controi lane (C) represents PM1 antibody (1gMsubclass, 900 kDa), that has been eluted off the affinity column with the antigen. This blotdemonstrates that the PM1 antibody does indeed recognize the affinity-purified antigen,and confirms that electrophoretic mobility is identical under both reducing and nonreducing conditions.82A BR NR RNRCrI420—83Fig. 17. Electrophoretlc analysis of the PM1 antigen after treatmentwith GAG-degrading enzymes.Aiiquots of affinity-pure PM1 antigen were first digested with various GAG-degradingenzymes (37°C, 24 hours) and then analyzed by SDS-PAGE under reducing conditions.The gels were stained with Aician blue in order to detect GAGs. Lane 1 shows the controldigest in which the antigen was incubated without enzymes. Lane 2, chondroitinase ABC(ch. ABC)-digested materiai; lane 3, chondroitinase AC (ch. AC)-digested antigen;. lane4, antigen digested with heparitinase (hep.); and lane 5, hyaiuronidase (hyai.)-digestedantigen. The same amount of starting antigen was used in au the incubations. The bandin lane 2 appears to be siightiy weaker in intensity when compared to the control lane (1),indicating that there may be dermatan sulfate groups present on the PM1 antigen.L____t.C)C•)mo 0)0 —.C)WC)856. Epltope determination(a) Periodate treatmentAntibody binding to the trans Golgi region of endodermal and mesenchyme cells suggested that theepitope was dependent on carbohydrate residues. To further investigate the nature of this epitope,sodium periodate treatment and enzymatic digestions of the antigen were performed (Fig. 18). Whenstrips of PVDF membrane containing PM 1 antigen were subject to sodium periodate oxidation underconventional conditions (20-50 mM periodate in 50 mM sodium acetate for 1 hour at room temperature)(Fig. 18A, lane 2), PM1 immunoreactivity did not appear to be different from the control (Fig. 18A, lane 1).When periodate oxidation was carried out in a high ionic strength buffer (0.2 M sodium perchiorate) theimmunoreactivity was reduced significantly (Fig. 18, lane 4) when compared with the control blot (lane 3),suggesting that the PM1 antibody is directed towards a carbohydrate-dependent epitope that containshighly negatively charged groups.(b) Endo F and trypsin treatmentsFurther investigations included enzymatic digestions of aliquots of the PM1 antigen, followed byWestern blot analysis. After digestion with endo F, an enzyme which removes N-linked carbohydratemoieties, no difference in band profiles was observed when compared with the control digestion (Fig.18B, lanes 1 and 2). This indicated that the PM1 antibody is not directed towards an N-linkedcarbohydrate epitope. In addition, the PM 1 antigen did not appear to be sensitive to enzymaticdegradation by this enzyme at concentrations that degraded control glycoproteins, since no shift in theapparent Mr of the band was evident after digestion. Trypsin treatment of the purified antigen resulted inthe total loss of PM1 immunoreactivity, suggesting that the antigen may contain polypeptide components(Fig. 18B, lanes 3 and 4).86Fig. 18. Analysis of PM1 Immunoreactlvlty after treatment with periodicacid, endoglycosidase F and trypsin.The figure shows the effect of periodate oxidation (A) endogiycosidase F and trypticdigestions (B) on the PM1 epitope.(A) Pure antigen was subject to SDS-PAGE under reducing conditions and transferredto PVDF. Strips of the membrane were then subject to periodate oxidation underconventional conditions (iow ionic strength; lane 2) or with 0.2 M sodium perchiorate inthe buffer (lane 4). Following periodate treatment, PM1 strips were reduced with sodiumborohydrate and then immunostained with PM1 antibody. Controls strips were incubatedin buffer (lane 1-low ionic strength, iane 3-high ionic strength) without periodate and thentreated similarly to the experimentais. Only under conditions of high ionic strength waspenodate effective in reducing PM1 immunoreactivity (iane 4, arrow), suggesting that theepitope contains highiy charged anionic groups which are characteristic of GAGs.(B) Aliquots of pure antigen were subject to enzymatic digestion (24 hours, 37° C) withendoglycosidase F (lane 2) or with trypsin (iane 4) prior to Western blot anaiysis with thePM1 antibody. After endo F digestion (iane 2), PM1 immunoreactivity is identical to that ofthe controi digestion (lane 1). However, after digestion with trypsin, PM1immunoreactivity is almost completeiy abolished (lane 4) when compared with the controldigestion (lane 3), indicating that the PM1 antigen has a polypeptide component.91•A BMr 1 2 3 4 1 2 3 4.1.. -9OO__.::420— rP!205—Periodate Endo F Trypsin88(C) Analysis with sugar-degrading enzymesWhen aliquots of PM1 antigen were subject to the same GAG-degrading enzymes as used above andthen transferred to PVDF, there was no noticeable difference in the intensities of PM1 immunoreactivebands or in the electrophoretic mobility of the antigen (Fig. 19, lanes 2-5), indicating as above that thePMI antigen does not contain GAGs typical of those associated with vertebrate proteoglycans.Furthermore, when PM I antigen was subject to a-fucosidase digestion prior to chondroitinase ABC or ACdigestion (Fig. 19, lane 7, 8), there was no difference in PM1 immunoreactive band profiles whencompared with the control (Fig. 19, lane 6), indicating that the PM 1 epitope does not appear to be blockedfrom chondroitinase digestion by fucose residues as is the case in some sea cucumbers and sponges(Kariya, et aL, 1990; Vieira and Mourao, 1988; Misevic and Burger, 1993).7. PerturbatIon of development with tunlcamycln and 6-xyloslde(a) Moiphological effectsTreatment with both tunicamycin and 13-xyloside disrupted mouth and digestive tract morphogenesis inP. ochraceus embryos. Tunicamycin, which inhibits glycoprotein synthesis by preventing the assembly ofN-linked oligosaccharide chains to proteins, was effective at a very small dose. At a time when controlembryos had undergone mouth and coelom formation, embryos continually exposed to 0.2 jig/mItunicamycin from early in gastrulation were characterized by a smaller rounded blastocoel with a veryshortened archenteron displaying no evidence of segmentation. It was, however, bent over to one sideof the ectoderm, as if attempting to establish contact with the presumptive stomodeal region for mouthformation (Fig. 20A). When embryos were grown for the same time period in the presence of 5 mM Bxyloside, which negatively competes for 0-linked xylose-mediated GAG chain attachment to the proteincore of proteoglycans, they were smaller than normal and the archenteron extended almost the entirelength of the blastocoel. There was little or no segmentation or differentiation of the archenteron, and noapparent bending (Fig. 20C). In both tunicamycin- and B-xyloside-treated embryos, mesenchyme cell89FIg. 19. Western blot analysis of the PM1 antigen after treatment withGAG-degrading enzymes and fucosidase.Aliquots of PM1 antigen were subject to enzymatic digestion with various GAG-degradingenzymes prior to Western blot analysis with the PM1 antibody. In addition to digestionwith chondroitinase ABC (ch. ABC), chondroitinase AC (ch. AC), heparitinase (hep.) andhyaluronidase (hyal.) alone (lanes 2-5), aliquots of the antigen were also digested with afucosidase (fu) prior to digestion with ch. ABC and ch. AC (lanes 7, 8) to determine iffucose residues were blocking the activities of the chondroitiriases. In all cases,immunoreactivity remains the same when compared with the controls (lanes 1, 6),indicating that the PM1 epitope is not aftered by the above enzymes. Arrow indicates thestart of the running gel.gC.)C.)__‘I’C0<< aQ€0 II0 .C .C42O—1234567891formation occurred despite the fact that blastocoel expansion was drastically reduced. The (3-xyloside-treated embryos strongly resembled those that were grown in the presence of the PM1 antibody.(b) Effects on PM1 synthesisEmbryos were treated with tunicamycin as described above, fixed at 5 1 /2 days growth, and processedfor thin section immunofluorescence. These embryos showed strong PM1 labelling of material in theblastocoel (Fig. 20D) that was of a comparable staining intensity with that of the controls (Fig. 20E). Thisindicated that PM1 synthesis and/or secretion was occurring at normal levels, even though the embryoshad experienced a developmental perturbation as a result of the tunicamycin treatment. However,embryos that were treated similarly with B-xyloside and examined at 5 1/2 days growth showed a significantreduction in the amount of PM1 staining in the blastocoel (Fig. 20F) when compared with the controls.This suggested that the synthesis and/or post translational modification of the PM1 antigen was disruptedby treatment with B-xyloside, and that the PM1 antigen shares characteristics of 0-linked proteoglycans.92Fig. 20. Effects of tunicamycin and 13-xyloside on P. ochraceusdevelopment and on PM1 synthesis.This figure shows the effects of tunicamycin and 13-xyloside on embryonic development(A-C) and on PM1 synthesis (D-F). Bar = 50 m for A-F inclusive. (A-C) Embryos at theearly gastrula stage were exposed to 0.2 tunicamycin, 5 mM l3-xyloside or controlmedia, which was changed daily for 3 days. At 5 1/2 days, the tunicamycin-treatedembryos (A) are smaller than controls (B), and have a rounded blastocoel with a shortenedarchenteron (a) which bends over to one side. The 13-xyloside-treated embryos (C) arealso smaller than the controls, and have an elongated essentially non-partitionedarchenteron (a) that extends almost the length of the blastocoel (C). All controls aresimilar, having a normal tripartite differentiated digestive tract (B).(D-F) Embryos corresponding to those in Figures A-C were fixed by freeze-substitution inethanol and embedded in JB4. Mid-saggital sections were then immunostained with thePM1 antibody. Labelling of the matrix in the tunicamycin-treated embryo (D) wascomparable in intensity to that of the control embryo (E), indicating that althoughtunicamycin disrupts normal development, it does not significantly alter the synthesis ofthe PM1 antigen. However PM1 labelling of B-xyloside embryos was significantly reduced(F) when compared to the control, indicating that this inhibitor of 0-linked GAG assemblyhas a direct inhibitory effect on PM1 antigen synthesis and/or assembly (tun =tunicamycin; xyl = B-xyloside; cont = control). Bar = 50 j.m.€1,IlAX--TH‘.2I••94PART B: Yolk and Cortical Granule Proteins1. SpecifIcity of antibodies against starfish egg granulesWhen immature starfish (P. ochraceus) oocytes (those with germinal vesicies intact) were stained withthe monoclonal antibodies PY4F8 and PC3H2, two distinct populations of granules were observed (Fig.21). The granules labelled by the PY4F8 antibody clearly formed a major constituent of the oocytevolume; they were the larger of the 2 types, and were dispersed throughout the cytoplasm of the egg(Fig. 21 A). This pattern of distribution is typical of yolk storage granules in other species, and suggeststhat the labelled granules seen in this starfish oocyte also contain yolk. The second granule type waslabelled by the PC3H2 antibody; these granules were concentrated in the peripheral egg cytoplasm, andappeared to represent cortical granules. A few were also dispersed in a random fashion throughout thecytoplasm of the immature oocyte, although in lesser numbers (Fig. 21B). Double labelimmunofluorescence revealed that these 2 types of vesicles were distinct (Fig. 21 C); the PY4F8-labelledgranules reached diameters of 4 I.Lm, and were absent form the cortical region of the egg. The PC3H2-labelled granules, roughly 1.5 p.m in diameter, were primarily in the cortical cytoplasm but were alsoscattered throughout the cytoplasm interspersed between PY4F8-Iabelled granules, indicating that theywere a distinct population.2. Ultrastructural analysis of PY4F8 immunoreactivityUltrastructural observations of mature oocytes fixed by freeze-substitution in Alcian blue-saturatedethanol indicated that a large part of the egg cytoplasm was occupied by a homogeneous population ofgranules of intermediate staining density. These same granules were decorated with immunogold labelafter staining with PY4F8 antibody (Fig. 22A). Also present in the mature oocyte were the highly electrondense cortical granules located in the peripheral cytoplasm; these granules did not have PY4F8-go)d label95Fig. 21. immunofluorescence localization of yolk and cortical granuleantigens.Immature oocytes fixed by freeze substitution in ethanol and embedded in JB4 resinhave been sectioned (1.5 jim) and immunolabelled with 2 antibodies against cytoplasmicgranules.(A) The PY4F8 monoclonal antibody labels ubiquitous granules dispersed throughoutthe egg cytoplasm that are characteristic of yolk granules or platelets. No staining isobserved over the nucleus (n). Bar (A and B) = 40 jim.(B) A serial section of the oocyte shown in Figure A but stained with PC3H2 monoclonalantibody shows labelling of a population of vesicles which are located primarily in theperipheral egg cytoplasm, but which are dispersed throughout the cytoplasm as well,although in lesser numbers. Staining of the nucleus (n) is again absent.(C) A section through an immature oocyte prepared as above, which has been stainedwith both PY4F8 (red) and PC3H2 (green) identifies 2 different types of granules: the redPY4F8-labelled granules are larger, reaching diameters of 4 jim, and are absent from thecortical region of the egg, while the granules labelled with PC3H2 (green), are roughly 1jim in diameter, and are found predominantly at the cortical cytoplasm just underlying theplasma membrane (pm). There are also some PC3H2-positive granules interspersedbetween the larger yolk granules throughout the cytoplasm (arrowheads). Bar = 15 jim.q697Fig. 22. Immunogold localization of yolk antigens in the unfertilizedoocyte.TEMs of an unfertilized (mature) oocyte that has been prepared by freeze-substitution inAlcian blue-saturated ethanol, embedded in LR White resin, and stained with PY4F8immunogold. Osmium has not been used in the fixative to maintain maximum antigenicity,and therefore the visualization of membranes is poor.(A) At this low magnification, the oocyte cytoplasm shows many yolk vesicles (y) ofintermediate staining density which are decorated with PY4F8-gold particles (see Fig.22B), and the highly electron dense cortical granules (cg), which are much smaller and arelocated at the cell periphery. Also present are electron translucent vacuoles (v), somewhich contain amorphous material, but which have no gold label associated with them.Surrounding the oocyte adjacent to the egg pm is the vitelline membrane (vi), and next tothat, the extensive jelly coat (jc). Bar = 3.0 pm.(B) At high magnification, it can be observed that the yolk vesicles (y) are evenlydecorated with PY4F8-gold label while the cortical granules (cg) have no label associatedwith them. In addition, another granule of intermediate staining density (g’) is apparent inthe cortex, but this granule has no gold label associated with it, suggesting that it mayrepresents a distinct population. Bar = 1.0 pm.:•,•.t‘;:,V...,.••i.....••‘4;i*4.f‘.4•.V..V..a••j--,ri-..%%i,...,,t—•Vq.;VbP•-,aV-,V.V•.V••—‘a4•,,,‘,,V.‘r?..:’‘--aV.*‘•a,••..•:•.•,V•••.•.•..Ø•..,..‘•:‘:‘-:-[pV••••:.d’•?%I.--‘..•c,“:••-..i••‘•k1•‘AVV-•,bVVVV.V•V‘:f‘h••.•:.•.-aV•“.‘-:14:.-..,J_r•\•V••••‘•••V••V‘•*t‘‘é’V..;rtD’.•,•4,.99associated with them (F. 22B). Other constituents of the egg cytoplasm included clear vacuoles, whichwere scattered throughout the egg cytoplasm, as well as a second population of granules, which wereintermediate in density but fewer in number when compared with the yolk granules. These granules werenot labelled with the immunogold PY4F8, indicating that there are at least 2 distinct populations ofgranules having intermediate staining density.3. Ultrastructural localization of the PC3H2 antigen In the eggColloidal gold labelling of mature oocytes with PC3H2 (Fig. 23) confirmed that this antibody was directedto material in the structures previously identified as cortical granules (Crawford and Abed, 1986). Thesemembrane bound granules were characterized by a highly electron dense periphery and a low densitycore, although patches of intermediate density were sometimes apparent. Labelling in this case waspredominantly associated with the electron dense regions in the periphery (Fig. 23, inset). In addition, afew gold particles were found associated with the plasma membrane. No gold decoration of the yolkgranules was observed.4. Localization of the PC3H2 antigen In early developmentAs is the case in other echinoderms, cortical granules in starfish exhibit a cortical reaction which involvesthe release of their contents and elevation of the fertilization membrane (Holland, 1980). To observe thefate of the PC3H2 antigen during oocyte maturation and early development, immunofluorescencestaining of sections of oocytes and early embryos were compared. Whereas in the unfertilized immatureoocyte, granules were dispersed throughout the central and cortical cytoplasm, (Fig. 24A), the matureoocyte, observed just after fertilization, showed fewer granules in the central cytoplasm, and a higherconcentration in the cortical cytoplasm (Fig. 24B). These activated eggs showed staining outside of the100Fig. 23. Immunogold localization of the PC3H2 antigen In theunfertilized oocyte.A TEM of the cortical region of a mature oocyte stained with PC3H2 immunogold. Severalgold decorated cortical granules (cg) are evident, and are located in the peripheralcytoplasm just under the egg plasma membrane. These membrane bound granules arediverse in their shapes, but typically have an electron dense component in the periphery,and an electron translucent region in the core. Patches of intermediate density are alsoapparent and are associated with the outer core. Gold particles are predominantlyassociated with the electron dense regions in the periphery (inset, bar = 0.3 iim). Someparticles are also associated with the plasma membrane (arrows), however, no particles areassociated with the yolk granules (y). vi = vitelline membrane. Bar = 1.5 lIm./0/4___e-. •4 • . . .. .4a-‘Ti,.c L,S :- .1 1w 1: ,• S”.è V• ‘. - .; •‘. . :- h .1 .1...- a• ) s’-’’—. - ‘I : — - •.- %•. —r.. • .y . .—- •• ;..ef -•.w ‘ ‘ r -Ia+1-_p- p. -•. .,; ._______-p102Fig. 24. immunofluorescence localization of the PC3H2 antigen in earlydevelopment.This figure compares the localization of PC3H2 immunoreactMty in 1.5 jim sections of theimmature and just fertilized oocytes as well as in the 2 cell-stage and blastula stageembryo, using 1.5 jim thick sections. Bar, A-C = 10 jim; D = 40 p.m.(A) in this immature oocyte, labelling is present in both the cortical granules which areconcentrated in the peripheral cytoplasm just beneath the plasma membrane (pm), and ingranules scattered throughout the oocyte cytoplasm. Some light staining of the plasmamembrane is also evident.(B) A section through an oocyte which has been matured with the hormone 1-methyladenine, fertilized, and fixed shortly after, showing that the majority of brightly stainedgranules are now concentrated in the cell periphery, with only a few scattered throughoutthe cytoplasm. A fertilization membrane (fm) has formed and is raised from the surface ofthe cell plasma membrane (pm). In addition, staining is evident on the outer side of theplasma membrane, indicating that some cortical granules have been activated and arereleasing their contents into the space limited by the fertilization membrane.(C) A section through a fertilized embryo after the first cleavage. The brightly stainedcortical granules located at the periphery of the egg are no longer present at this stage.However, label is now concentrated on the outer side of the oocyte plasma membrane inthe space limited by the fertilization membrane (fm), the perivitelline space (pvs). Severalbrightly stained granules still persist in the cell cytoplasm.(D) A lower magnification view of an early blastula stage embryo, showing an ECM-filledblastocoel surrounded by blastomeres, some of which are still folded into the center ofthe blastocoel. The label is absent over the fertilization membrane (fm); however, brightlystained granules are still present in the blastomeres (arrowheads). The staining of thematrix in the blastocoel (b) is also evident.C”0104cell plasma membrane in the space limited by the fertilization membrane (perivitelline space), suggestingthat the contents of these granules was released into this space. By the second cleavage stage, therewas a dramatic change in the PC3H2-staining pattern; brightly labelled granules were no longer present inthe cortical cytoplasm, and a diffuse pattern of staining was present throughout the perMtelline space(Fig. 24C). Some cytoplasmic granules were labelled, however they were not located in the apical regionof the cells. At 24 hours post-fertilization, the embryo had developed into a blastula, and PC3H2 stainingwas no longer present in the apical region, either in the perivitelline space nor associated with thefertilization membrane. However, some labelled granules were still present in the blastomeres, and inaddition, labelling of the ECM of the blastocoel was evident (Fig. 24D).Ultrastructural observations of blastula and gastrula stage embryos with PC3H2 immunogold revealedthat these labelled granules had a morphology resembling the granules of the oocyte cortical cytoplasm,including an electron dense component over which gold particles were localized (Fig. 25A-C). In thegastrula, gold label was localized over ECM fibers of the blastocoel and the basal lamina (Fig. 25B ), as wellas over one element of the hyaline layer (Fig. 25C). Further observations at this stage revealed thatregions of the Golgi complexes were also labelled with a few gold particles, indicating that some newsynthesis or post-translational modifications of the antigen was occurring. The cortical granules wereoften closely situated to the Golgi complex, suggesting that they may be providing a storage site for thenewly synthesized antigen (Fig. 25C).5. DIstributIon of PY4F8 immunoreactlvlty durIng development(a) ImmunofluorescenceFurther analysis was undertaken to study the fate of the PY4F8 immunoreactive yolk granules and theircontents during development to determine whether they undergo depletion in a manner similar to knownyolk proteins. lmmunofluorescence on 1.5 j.m sections of different staged embryos revealed that the105Fig. 25. Immunogold localization of the PC3H2 antigen in the earlyembryoThe figure shows TEMs of the unhatched biastula stage (A) and gastrula stage (B, C)embryos which have been fixed by freeze substitution in ethanol and processed forPC3H2 immunogoid labeiling.(A) The unhatched blastula shows the presence of an electron dense PC3H2-positivegranule, which resembles the cortical granules of the unfertilized egg. Gold particles arenot associated with the fertilization membrane (fm). However some gold particles arefound in the region just below it over elements of the developing hyaline layer (arrows).Bar=0.6jim.(B) A section through an early gastrula stage embryo showing positive staining with thePC3H2 antibody of the basal lamina (bI) of endodermal cells, as well as over elements ofthe matrix in the biastocoel (*) Note that the electron dense granule (cg) that ispositioned near the base of the cell also has gold label associated with it. Bar = 0.6 j.m.(C) This figure shows the apical region of 2 ectoderm cells of the same embryo as (B).Gold label is found over a cortical granule (cg) as well as over regions of 2 Golgi complexes(gc). As particles are found just outside the cg membrane, material is either beingsecreted or is being taken up by the granule for storage. in addition, one region of thehyaline layer (hi) is also decorated with PC3H2 immunogold (arrows). Bar = 0.6 pm.4ik?: .,%‘.‘3tr• ‘•-.‘. ‘‘3dr 4‘:‘• ‘.4%- “:?‘14.- - ---...fl •‘ ‘•%‘4 e‘4 ‘•d, •S1cg• “-- . , a-, .• 1a -i”:; ‘:: ‘:---,:p-•t•-• !‘ -—•-. • •. -—-— *4‘%. I..— j‘• •,•.;•:‘-‘•-‘--ee.• •• •• 4• -. _•.-a• ----• --C -IS‘II0BVA4/fm4Is-f,/41*t4Sr,. ‘,,--%,,#I, •t•rI4107granules observed in the oocytes were also present in early development, including the blastula stage (2day old) and gastrula stage (4 day old) embryos (Fig. 26 A-D). The granules were stained with intensitiesthat were comparable to those of the egg, suggesting that the amount of protein in the granules did notchange substantially over the this time period (Fig. 26A-D). These putative yolk granules were seen in allthree germ layers, including the ectoderm, endoderm and mesenchyme cells of the gastrula (Fig. 26D).Labelling was confined to the intracellular granules, with no label apparent in extracellular spaces or otherareas of the cells. Stained granules were still present in all three germ layers of the early larva (Fig. 26E),although the intensity of the stain was reduced in comparison with that in the earlier stages ofdevelopment (Fig. 26A-D). It was only at the late bipinnaria larval stage (12 days post fertilization) when thecomplete depletion of this material was observed (Fig. 26F). Labelling with PY4F8 was never seenoutside of these granules, either within the cells or in extracellular locations.(b) ImmunogoldTEM observations of the early larva revealed granules that were intermediate in staining density with amorphology that resembled the yolk granules found in the oocyte. These granules also labelled withPY4F8-immunogold (Fig. 27), confirming the results with immunofluorescence. In addition to thesegranules, several other membrane-bound structures, including Golgi associated vesicles and electrontranslucent vacuoles were observed. No evidence of gold label was apparent outside of the granuleseither in the cytoplasm or in the extracellular matrices, suggesting that the material was not being releasedinto the cell cytoplasm or extracellular spaces.108Fig. 26. Immunofluorescence localization of yolk proteins in the earlyembryo and larva.In this series, eggs and embryos of various stages of development have been fixed byfreeze-substitution into ethanol, and thin sections (1.5 pm) of JB4-embedded materialhave been stained with the PY4F8 antibody. Bar = 50 pm.(A) The immature oocyte showing bright staining throughout the cytoplasm in roundgranules, but which is excluded from the nucleus (n).(B) A 2-cell stage embryo, showing cells with abundant stained granules throughout thecytoplasm.(C) A blastula stage embryo, showing intense labelling of granules in the blastomeres.The unstained regions in the cell apices represent nuclei. b = blastocoel.(D) A mid-sagittal section through a gastrula, in which differentiation of the archenteroninto the different regions of the digestive tract has begun. Labelled granules are still seenthroughout the organism in all 3 germ layers, including the ectoderm (ec), the endoderm(en) and mesenchyme cells (mc).(E) An early larval stage of development, showing a well differentiated digestive tract, witha mouth (m), esophagus (e), stomach (5) and intestine (i). Note that although manygranules are still present throughout the organism, the staining intensity of thesegranules has decreased dramatically over this period of development.(F) A late bipinnaria larva, showing some elements of the digestive tract, including theesophagus (e), stomach (s), intestine (I) as well as a coelomic pouch (cp). At this laterstage in development, no labelled granules are present in the organism, indicating thatthe material has been completely depleted, and that no new synthesis of this material isoccurring. This figure has been over-exposed during photography compared to thosepreceding it in order to increase the contrast.00110Fig. 27. Immunogold localization of yolk proteins In the gastrula.A TEM showing ectodermal cells of an early bipinnaria larva, prepared by freeze-substitution into Alcian blue-saturated ethanol, and labelled with PY4F8-gold. Goldparticles are located over a single putative yolk granule (y), which has an intermediatestaining density. Label is not located over smaller granules of similar staining density (g’),nor is staining observed in any other areas of the cell. Other structures in the cell includemitochondria (m), a nucleus (n), a cilium rootlet (c), lipid vacuoles (v) and on the apical cellsurface, the hyaline layer (hi). Bar = 0.8 rim.Ifr:•:4.4•••f‘41126. lmmunoblot analysis of the PY4F8 and PC3H2 antigensDetergent extracted material from immature oocytes was subject to Western blot analysis to identity theproteins constituent of the 2 granules types. Material probed with the PC3H2 antibody (cortical granule-specific) identified a single band of 125 kDa (Fig. 28A). When the same extract was probed with thePY4F8 antibody (yolk granule-specific), 3 bands were stained: a major band of 164 kDa, a minor band of100 kDa and a doublet at 70 kDa. (Fig. 28A). To investigate the subunit nature of the yolk antigens, non-reduced samples (prepared without B-mercaptoethanol) were transferred to PVDF and probed with thePY4F8 antibody for comparison with reduced extracts (Fig. 28B). Under non-reducing condition, thesame 3 major bands were present as those found in the reduced lane, indicating that these 3 yolk proteinsare not present as a disulfide- linked multimer in the immature oocyte. In addition, a new band of 45 kDawas present under these conditions. As this band was not present under reducing conditions, it ispossible that the 13-mercaptoethanol, while not breaking inter-chain disulfide bonds, was alteringimmunogenicity by destroying intra-chain disulfide bonds of the 45 kDa protein. Furthermore, there was asignificant difference in the antigenicity of the yolk proteins prepared under reducing and non-reducingconditions. Blots from non-reducing gels were several fold more antigenic, i.e., 0.2 ig total oocyte extractunder non-reducing conditions had a comparable band signal as 1 protein in the reducing gels whenprobed with PY4F8 antibody. This further suggests that the i3-mercaptoethanol altered the antigenicity ofthe yolk proteins, and indicates that PY4F8 may be directed towards a protein rather than a carbohydrateepitope.113Fig. 28. Western blot analysis of yolk and cortical granule antigens Inthe oocyte.The figure shows detergent-extracted material from immature oocytes which has beensubject to SDS-PAGE on 3-10% acrylamide gels under reducing (A) and non-reducing(B) conditions, and then transferred to PVDF for immunoblotting.(A) Lane 1 was stained for total protein with Coomassie blue. Lane 2 represents 5 tgtotal protein which was immunostained with PC3H2, revealing a single band at 125 kDa.Lane 3 represents 1 jig of total protein which was probed with PY4F8 and shows 4 majorbands at 164 kDa, 100 kDa and a doublet at 74 and 64 kDa.(B) In this figure, material was prepared for electrophoresis without (3-mercaptoethanol:Lane 1 was stained for total protein with Coomassie blue; lane 2 represents 0.2 jig ofoocyte extract probed with PY4F8, showing a similar band profile as that observed underreducing conditions (A, lane 3). One additional band is stained at 48 kDa under nonreducing conditions.I ‘•tA420—205—_B116_I66—45—0I.123‘I121157. Immunoblot analysis of PY4F8-yolk antigens during developmentDetergent extracts from 4 different developmental stages, immature oocytes, blastulae, gastrulae andbipinnaria larvae were separated by SDS-PAGE gels under reducing conditions, and the proteins weretransferred to PVDF and probed with PY4F8. Coomassie blue staining of the extractable proteinsrevealed a general decrease in proteins larger than 120 kDa during embryonic and larval development(Fig. 29A). PY4F8 immunoreactive bands were present in all the embryonic stages although no bandswere observed in the larval stage extracts (Fig. 29B). In the unfertilized oocyte (0), 3 bands were presentat 164 kDa, 100 kDa and a doublet at 70 kDa; at the blastulae stage (B), the 100 kDa band was no longerpresent, however the other 2 bands were identical to those of the oocyte. At the gastrula stage (G), only abroad band between 64-74 kDa remained, although a few minor bands were evident at 116 and 100 kDa.At the advanced larval stage (L), there was no evidence of immunoreactive bands. These results areconsistent with the immunohistochemical findings, and suggest that the PY4F8 antigens areprogressively depleted and/or broken down over embryonic and larval development.8. Ontogeny of the major yolk proteinsTo determine the possible source of the yolk antigens, both the adult intestine and coelomic fluids wereexamined for PY4F8-reactive proteins. The 3 major yolk proteins of the oocyte were present in both theintestine and coelomic fluids, suggesting that synthesis of the yolk proteins may be occurring in theselocations (Fig. 30). In addition, under non-reducing conditions, a region of immunoreactivity at 400 kDa,not seen in the oocyte, was present in both intestinal extract and coelomic fluid (Fig. 30, arrows). Underreducing conditions, however, this 400 kDa band was not present, suggesting that this high molecularmass protein may be a disulf ide-linked complex.116FIg. 29. Western blot analysis of yolk proteins In early development.Western blot analysis with PY4F8, in which detergent extracted proteins from eitherunfertilized oocytes (0), blastulae (BI), gastrulae (G) and bipinnaria larva (L) wereseparated on SDS-PAGE (3-10% gradient gels) under reducing conditions, andtransferred to PVDF membranes.(A) Coomassie blue-stained membrane, in which 10 jig protein was loaded/lane showedthat the higher molecular mass proteins are gradually lost during development.(B) Western blot in which 1 jig total protein/lane was probed with the PY4F8 antibody. Inthe oocyte (0), 4 major bands are present at 164 kDa, 100 kDa and a doublet at 64 and 74kDa. At the blastula stage (BI), the decrease of the 100 kDa band was apparent, and atthe gastrula stage (G), only 1 major broad band remained between 64-74 kDa. However afew faint bands between 100 and 160 kDa were still present. Extract from the larval stage(L) showed no immunoreactive bands.I Il420—.205—.116—97—66—45—Mr OBIG L OBIGLL..Coomassie Western118FIg. 30. Western blot analysis of yolk proteIns in the adult intestine andcoelomic fluid.The figure shows PY4F8 immunoreactivity to components of the adult intestine (A) andcoelomic fluid (B). Intestinal extract and coelomic fluid was separated by electrophoresison 3-10% gradient gels under both reducing and non-reducing conditions, and thentransferred to PVDF and either stained with Coomassie blue (lanes 1 and 3) or probedwith PY4F8 (lanes 2 and 4).(A) intestinal extract prepared without B-mercaptoethanol and stained with PY4F8showed several immunoreactive bands at 400, 164, 120, 100 and 74 kDa (lane 2). Underreducing conditions, only 2 major bands at 164 and 74 kDa were apparent (lane 4),suggesting that the intestine contains a large (400 kDa) multimeric disultide linked yolkprotein.(B) A band profile similar to that observed in the non-reduced intestinal extract (A, lane 2)was observed when coelomic fluid is probed with PY4F8 under non-reducing conditions(lane 2). Again, a region of immunoreactivity was present around 400 kDa (arrow),although it was a broader band than in the intestine; in addition, the 4 bands with lowermolecular masses were also present. Under reducing conditions (lane 4),immunoreactivity was restricted to a single band migrating as a doublet at 164 kDa.I IiCoelomicAIntestine/ —420—205—-.1--I--116—66—45—Ba——12I ;1tdiI3412 341209. Lectin characterization of the yolk proteinsOf the different FITC-conjugated iectins used in this study, only WGA and Con A stained granulesresembling the PY4F8-specific yolk granules (Fig. 31). The staining pattern was essentially the same forboth of these lectins, characterized by strong labelling of the ubiquitous yolk granules of the oocytes.Positive labelling with Con A suggested that the granules contained glycoproteins of the high mannose orcomplex type, while positive labelling with WGA suggested that glycoproteins with terminal neuraminicacid or N-acetyl glucosamine residues were also contained within these granules.To determine if the PY4F8 yolk antigens were in fact the yolk constituents being recognized by Con Aand WGA, the antigens were isolated by immunoprecipitation from homogenized oocytes, and were thentransferred to PVDF where they were probed with Con A and WGA. Coomassie blue staining of theimmunoprecipitated PY4F8 antigens from oocyte extracts revealed 2 major bands at 164 kDa and 116kDa, and 3 minor bands at 400 kDa, 100 kDa and 74 kDa (Fig. 32A, lane 1). In addition,immurloprecipitating IgG fragments were present at 55 kDa in both the PY4F8 immunoprecipitation andcontrol lanes (Fig. 32A, lane 2). When the same samples were probed with PY4F8 antibody, similar bandswere detected; however, no staining of the 116 kDa band was observed, indicating that this major proteinmay have co-precipitated together with the yolk antigens (Fig. 32A, lane 3). Alternatively, this protein mayhave represented a subunit lacking the PY4F8 epitope but derived from one of the larger yolk antigens.When the immunoprecipitations were probed with Con A, all of the PY4F8 yolk antigens were stained,including bands at 400, 164, 100, 74 and 64 (Fig. 32A, lane 5) as well as the 116 kDa band. With WGA,the 2 larger yolk antigens (400 and 164 kDa) stained positive, but no labelling of the 100, 74 or 64 kDaproteins was observed (Fig. 32A, lane 7). Control lectin blots revealed no binding under the sameconditions. These results suggested that: (1) the yolk antigens were indeed glycoproteins; (2) all of theantigens contained N-linked high mannose or complex type (Con A-reactive) oligosaccharide chains; and(3) the larger yolk antigens also contained WGA-reactive sites, indicating the presence of neuraminic acid121Fig. 31. Con A and WGA binding in the unfertilized oocyte.The figure shows 1.5 m sections through immature oocytes which have been preparedby freeze-substitution and stained with the FITC-labelled lectins Con A (A) and WGA (B).Both lectins appear to stain yolk granules found throughout the egg. Staining isexcluded from the nuclei (n) and from the jelly coat which surrounds the oocyte. Bar = 50jtm.cir4123or N-acetyl giucosamine terminal residues, characteristic of N-linked complex oligosaccharides and 0-linked oligosaccharides (Beeley, 1985).10. Epltope characterization of the PY4F8 yolk antigensTo further investigate the extent of glycosylation, as well as to determine the nature of the PY4F8epitope, yolk proteins were treated with Endo F, an enzyme which removes all N-linked oligosaccharidegroups from proteins. After digestion, all of the the yolk antigens appeared to have a greaterelectrophoretic mobility than the controls (Fig. 32B, lane 2), indicating that they were sensitive to Endo Ftreatment and contained a considerable amount of N-linked carbohydrate. Despite this, immunoreactivityafter endo F treatment was retained, showing that the epitope was not dependent on the theseoligosaccharide moieties, but rather was directed to the protein core and/or to 0-linked oligosaccharides.124Fig. 32. CharacterizatIon of PY4F8-immunopreclpltate with lectlns andendoglycosidase F digestion.(A) This figure demonstrates the binding patterns of Con A and WGA on immunoprecipitated yolk(PY4F8) antigens from immature oocytes. Immunoprecipitations and their controls were run on 3-12 % gradient gels, transferred to PVDF, and stained as follows: Lanes I and 2 show Coomassieblue staining of PY4F8 immunoprecipitated antigens (IP) and the control immunoprecipitation (IPC). in addition to the immunoprecipitated lgG fragments (50 kDa; arrow), which were present inboth lanes 1 and 2, several other bands were apparent in lane 1, including strong bands at 164and 116 kDa, and weaker bands at 400, 100, and 74 kDa. in lane 3, the immunoprecipitatedantigens were probed with PY4F8 hybridoma supernatant, which detected bands at 164, 100, 74and 64 kDa, in addition to the fragmented lgG (arrow). Lane 4 represents the same PY4F8-immunoprecipitation as in lane 3 which was probed with a control monoclonal antibody, and whichshowed specificity only for the fragmented lgG bands. The 116 kDa Coomassie stained bandpresent in lane 1 (arrow) was not present in lane 3, indicating that this protein co-precipitated withthe PY4F8 antigens, but was not a PY4F8-specific yolk protein. Lanes 5 and 6 show Con Astaining of the immunoprecipitated antigens (Con A) and the control immunoprecipitation (Con A-C). Several Con A-positive bands were present, including all of the bands detected in lanes 1 and3, indicating that they were all glycosylated and contain terminal mannose residues. No bandswere apparent in the control lane 6. Lanes 7 and 8 represent blots identical to those in lanes 5and 6, but stained with WGA. In this case, only the 2 larger yolk antigens and the 116 kDa coprecipitated band were stained, indicating that these glycoproteins also contain terminalneuraminic acid and br N-acetylglucosamine residues.(B) This figure shows the effects of Endoglycosidase F on PY4F8 yolk antigens. Detergentextracted proteins from blastulae were digested for 18 hours at 37°C, separated by SDS-PAGEunder reducing conditions, and transferred to PVDF for immunodetection with PY4F8supernatant. Controls were incubated without the enzyme and treated as above. When probedwith PY4F8, 3 major bands were present in the control digest (lane 1), including a band at 164kDa, a doublet at 100 kDa, and a broad band at 70 kDa, which was representative of the typicalband profile observed in early embryos. After digestion with Endo F, several immunoreactivebands were present (lane 2), indicating that the PY4F8 epitope was not dependent on the N-linked oligosaccharides. In addition, the Endo F-digested proteins migrated with a greaterelectrophoretic mobility, and there were no major bands which corresponded to those of thecontrol, indicating that most if not all of the yolk proteins recognized by PY4F8 were sensitive toEndo F digestion and contained N-linked oligosaccharides.‘.30C)’0)CoIb3,0)-‘0(0(110)0)(11IIIIIIE.1II‘p‘p-cPY4F8PY4F8-CConAConA-CWGAWGA-C0),.0IbJUI0)0)UiOIIIII_IITIEIfHII14-1wContEndoFS.CI126IV. DISCUSSIONPART A: The PM1 proteogiycan in starfish gut morphogenesis(I) Notes on methodology1. Monoclonal antibody productionIn this study, monoclonal antibodies were generated to isolate, characterize and study amorphogenetically active blastocoel-specific component in starfish gut morphogenesis usingimmunochemical and in vivo perturbation techniques. Monoclonal antibodies were chosen as a researchtool, because with this technique a relatively impure antigen can be used to generate a host of antibodies,and then the particular one(s) of interest can then be selected for. Previously, monoclonal antibodieshave been generated to blastocoel matrix components of P. ochraceus embryos; these antibodies,however, also shared epitopes with elements of the hyaline layer (Crawford and Crawford, 1992). In usingantibodies such as these for function blocking studies, it was difficult to separate effects on the hyaiinelayer from those on biastocoel matrix components. For this reason, it was necessary to generate newmonoclonais with biastocoei-specific epitopes. Attempts at reducing the number of antibodies generatedagainst the hyaline layer by immunizing mice with a hyalirie layer-deficient embryo homogenate were notsuccessful, as many antibodies with specificities for the hyaiine layer were still generated. Hence a newapproach was used to prepare the immunogen in the present study. This involved obtaining aglycoconjugate-enriched fraction from embryo homogenates using a Con A-Sepharose affinity column.While many positive clones were generated, it was not possible to use conventional techniques such asthe ELISA (enzyme-linked immunosorbent assay) in screening the various hybridoma supernatants,because this procedure did not permit the level of resolution required to select for blastocoel-specificimmunoreactivity. Instead, supernatant from all positive wells (wells containing successful fusions) wastested on thin (1 .5iim) sections of gastrula-stage embryos using indirect immunofluorescence; in this way,antibodies specific for the blastocoel ECM could be distinguished from those which had dual specificity for127both the hyaline layer and the ECM, or other specificities. The resolution afforded by the plastic-embedded sections was very advantageous; at the light microscopic level, labelling of the blastocoelmatrix could be distinguished readily from that of basement membranes, the hyaline layer, and in factdifferent regions of the hyaline layer. In addition, labelling of intracellular granules was easy to detect, andtherefore the synthesis and storage of matrix components could be investigated. When this techniquewas used in combination with tissue that was fixed by freeze-substitution in ethanol, antigenicity of manycomponents did not appear to be compromised, as monoclonals having both protein and carbohydratespecificities labelled the tissue strongly. This was, however, a very labor intensive procedure, as thehyaline layer was particularly antigenic; in some cases, the entire 96 well plate of hybridomas was positive,and while almost all produced antibodies directed against the hyaline layer, 1 clone was generated thatproduced the blastocoel-specific antibody, Pisaster matrix 1 (PM1).2. Antibody perturbation studiesStudies involving the use of antibodies to block the functions of ECM components have beenperformed extensively in both cell culture and in vivo systems. When performed in the whole organism,perturbation studies of this type usually involve micro-injecting the antibodies directly into tissue to ensurethat they are delivered to a specific location. Alternately, if the targeted structure is on the outer aspect ofcells, as is the hyaline layer of echinoderms, perturbation experiments can be performed by simplyincubating the embryos in sea water containing antibodies. A similar approach was used in the presentstudy, that of transferring embryos into sea water containing PM1 antibody or antibody fragments.Although it was not assumed that the antibody would gain access to the blastocoel, embryos treated inthis way were severely affected in their development, while controls incubated in normal mouse 1gM werenot. This suggested that the PM1 antibody was having a specific effect on a developmental process,resulting in the disruption of morphogenesis. Subsequent experiments were therefore designed toshow that this effect on development was the result of the PM1 antibody binding to and interfering withthe PM1 antigen in the blastocoel. This involved probing the tissues with an FITC-labelled anti-lgM128antibody shortly after the embryos were transferred into sea water containing the PM1 antibody. Theresults showed specific labelling of material in the blastocoel and endodermal cells but not in other areasof the embryo or hyaline layer, which demonstrated that the PM1 antibody did indeed gain access to theblastocoel when embryos were simply transferred into sea water containing it. This suggests that theeffect on development was a result of a specific interaction between the PM1 antigen and antibody. It isunclear whether the intact or fragmented 1gM was actively taken up by the epithelial cells, or whether theantibody simply diffused into the blastocoel under these conditions. Since the epithelial cells are oftenloosely attached in the region of the endoderm, it is possible that during archenteron expansion, theirjunctional complexes are transient, permitting antibody to leak through into the blastocoel.3. Affinity purification of the PM1 antigenSeveral factors owing to the biochemical properties of the PM1 antibody (1gM subclass; 900 kDa) andthe large size of the PM1 antigen (>600 kDa) made the affinity purification of this molecule somewhatchallenging. Initially, the antibody was coupled to CNBr-activated Sepharose 4B. However forundetermined reasons, the coupling efficiency with this support resin was extremely low. Affi-gel 10provided a reasonable alternative, although intact 1gM was still difficult to couple efficiently. Attempts weremade to first fragment the 1gM with pepsin, and this improved the coupling efficiency somewhat. Asecond difficulty was the size of the antigen (>600 kDa). With a molecule of this size, the binding of asingle epitope is not strong enough to withstand the vigorous column washings that are required in theaffinity purification procedure, to ensure that non-specific proteins are eluted from the column. Thereforeit is likely that a significant amount of the PM1 antigen was lost during the washing procedures. A furtherfactor was the buffer conditions required for affinity chromatography. Even though 4 M guanidinechloride was used to extract the PM1 proteoglycan from the embryos, this buffer was not compatible withaffinity chromatography, as it would have denatured the column, and instead Tris buffer was used. It isprobably that a considerable amount of material was lost due to insolubility in the Tris buffer. For these129reasons, up to 30 column runs were required to obtain enough material to carry out the experiments in thepresent study. The benefits of obtaining a pure antigen did, however, outweigh the above negativefactors.(II) Biochemical Characterizations of the PM1 antIgen1. Evidence that the PM1 antigen Is a proteoglycan.Several experimental findings led to the suggestion that the PM1 antigen is a carbohydrate-containingmolecule and that it shares structural features that are commonly associated with those of proteoglycans.These findings are discussed in the following section.(a) Locallzation in the Golgi suggests the PM1 antigen contains sugar residuesThe PM1 antigen is first detected in the trans Golgi network and in Golgi associated vesicles. Studiesshow that carbohydrates are usually modified and/or added to protein or lipid cores in this region of theGolgi complex (for review see Kornfeld and Kornfeld, 1985). These include: (1) oligosaccharides linked toasparagine residues (N-linked), that are characteristic of glycoproteins; (2) oligosaccharides linked toserine, threonine or hydroxylysine residues (0-linked) that are characteristic of mucins; (3) GAG chainstypical of proteoglycans. The localization of PMI immunogold particles to the trans Golgi region and postGoigi secretory vesicles and not to the cis Goigi or ribosomes suggests that the antibody is directedagainst a region of the antigen which is undergoing processing or modifications of its carbohydrate region.The nature of the sugar groups cannot, however, be determined from these observations.130(b) Carbohydrate groups not typical of glycoproteinsSeveral experimental findings indicated that the carbohydrate groups associated with the PM1 antigenare not those typically associated with glycoproteins. First, glycoproteins generally contain N-linkedoligosaccharides that are of the high mannose or complex type. Endoglycosidase F, an enzyme whichspecifically cleaves such N-linked oligosaccharides (Elder and Alexander, 1982), did not affect the PM1antigen, as both PM1 immunoreactivity as well as the apparent Mr of affinity pure PM1 were unchangedafter treatment, while control glycoproteins showed a definite mobility shift after treatment. Secondly, thevicinal hydroxyl groups of such N-linked oligosaccharides can be efficiently oxidized with periodic acid togenerate aldehyde groups, which can then be reduced with sodium borohydride (Bobbitt, 1956). Thistreatment changes the conformation of the carbohydrate groups, and often will affect antigenicity of anantibody that is directed towards a carbohydrate epitope (Woodward et al., 1985). Periodate oxidation ofacidic polysaccharides (GAGs), however, proceeds very slowly because of the strong repulsion betweenperiodate ion and the polyanions. This effect can be overcome by increasing the ionic strength of theoxidizing buffer, which decreases the degree of ionic repulsion between periodate ions and highlycharged GAG groups (Scott and Harbinson, 1968). In the present study, when PM1 was oxidized underconditions of low ionic strength, little or no effect was observed, suggesting that it does not containoligosaccharides characteristic of glycoproteins. However,when the ionic strength was increased with 0.2M sodium perchlorate, immunoreactivity was reduced significantly. This strongly suggests that theepitope is dependent upon highly charged acidic groups such as GAGs.Experiments in which embryos were treated with tunicamycin provided yet further evidence to supportthe argument that PM1 is not a glycoprotein. Tunicamycin is a drug which inhibits the first step in theformation of N-linked oligosaccharide synthesis, and is often used to inhibit glycoprotein synthesis(Duskin and Mahoney, 1982). Embryos that were grown in the presence of very low concentrations oftunicamycin (0.2 ig/ml), although severely affected by the drug, were still capable of synthesizing and131secreting the PM1 antigen. This indicated that the antigen was not significantly affected by this inhibitor ofglycoprotein synthesis and that it is therefore probably not a glycoprotein.Finally, the PAS reaction, which involves first the oxidation of vicinol diols to generate dialdehydes, andthen the visualization of these dialdehydes with the Schiff reagent (McManus, 1946), is used to identifyoligosaccharides typical of glycoproteins, as N-linked oligosaccharides are rich in vicinol diols. Staining ofthe affinity-purified PM1 antigen with PAS was negative, again suggesting that the PM1 antigen is not aglycoprotein. Staining was observed, however, with Alcian blue, a cationic dye which binds strongly topolyanionic groups such as GAGs (Scott, 1972). The reaction can be made more specific by adding 0.5 Mmagnesium chloride to the Alcian blue solution. When this is done, all commonly known GAGs arestained, but nonspecific stain uptake by negatively charged polyion species found in other proteins doesnot occur (Wall and Gyi, 1988). The specific staining of the PM1 antigen therefore suggests that itcontains highly acidic groups typical of those found in GAGs. Furthermore, despite the fact that it doesnot stain with Coomassie blue, the antigen is sensitive to tryptic digestion which indicates that thesehighly acidic groups are indeed linked to a polypeptide core, as they are in proteoglycans.(C) Carbohydrate groups not typical of known vertebrate proteoglycansProteoglycans are common elements of ECMs, and have been described in a variety of vertebrate andnon-vertebrate systems (reviewed by Hardingham and Fosang, 1992; Goetinck and Winterbottom, 1991,Har-El and Tanzer, 1993). While many proteoglycans contain some N and 0-linked oligosaccharides(aggrecan from a rat chondrosarcoma, KSPG from monkey cornea Nilsson et aL, 1982; 1983), all containGAG chains (Roden, 1980). There are different classes of vertebrate GAGs, which include the chondroitinsulfates, dermatan sulfate, keratan sulfate, and heparin sulfates, and they are high in glucuronic residues,as well as N-acetylgalactosamine, N-acetyl glucosamine, and galactose. These GAGs can beexperimentally removed from their protein core with enzymes that are available for each GAG. Thisprovides a convenient diagnostic tool for their identification. In addition, as mentioned above, the132negative polyanionic structure of the GAG enables them to bind strongly to cationic dyes such as Alcianblue, and this binding is often used to identify proteoglycans and hyaluronic acid in tissues.When the PM1 antigen was treated with the vertebrate GAG-degrading enzymes chondroitinase ABC,chondroitinase AC, heparitinase and ovine testicular hyaluronidase, and then separated by SDS PAGEand stained with Alcian blue, no significant effect was observed despite the fact that under similarconditions, known standards were degraded. There did appear to be a slight reduction in stainingintensity after digestion with chondroitinase ABC, indicating that there may be some dermatan sulfategroups present on the PM1 proteoglycan. However, no change in the electrophoretic mobility of theband was apparent, indicating that a very small amount if any had been digested away. Western blots ofthe PMI antigen digested with each of the enzymes listed above showed that immunoreactivity andelectrophoretic mobility was also essentially unchanged when compared with the controls. This isconsistent with earlier observations in the same species (P. ochraceus) in which no proteoglycanssensitive to the above enzymes could be identified (Crawford and Crawford, 1992).Although it is known that GAGs from echinoderms contain glucuronic acid, N-acetylgalactosamine andsulfate (Motohiro, 1960; Vieira and Mouräo, 1988; Kariya eta!., 1990), suggesting that they are similar toknown vertebrate GAGs such as the chondroitin sulfates, the observations in this study suggest that PM1is quite distinct from known vertebrate proteoglycans. In other studies of invertebrate proteoglycans, theanalysis of several chondroitin sulfate proteoglycans isolated from the sea cucumber and sponge haveindicated that they have fucose-containing branched GAGs which are insensitive to digestions withchondroitinases (Kariya, eta!., 1990; Vieira and Mourao, 1988; Misevic and Burger, 1990). This does notappear to be the case for PM1, as digestion of the antigen with a-fucosidase prior to chondroitinasedigestion failed to have any effect.133d) !3-xyloside blocks PM1 antigen synthesisAlthough the PM1 antigen contains highly negatively charged groups characteristic of GAGs, which arelinked to a polypeptide core, the GAGs do not resemble those commonly found in vertebrate systems. Inspite of this, the B-xyloside inhibition experiment provided strong evidence that the PM1 antigen sharesstructural characteristics of known vertebrate GAGs. As previously mentioned, (3-xyloside competes withcore protein xylosides for GAG chain attachment, thereby creating a free xyloside-GAG molecule(Okayama et aL, 1973; Galligani et aL, 1975) and inhibiting the assembly of proteoglycans. When embryoswere grown in the presence of this chemical, the level of PM1 proteoglycan synthesis and/or secretionwas reduced significantly. This indicated that although these starfish GAGs appear different from otherknown GAGs, their attachment to the polypeptide core involves a xylose-mediated linkage. This type oflinkage is characteristic of known proteoglycans, such as those containing chondroitin sulfate, dermatansulfate and heparan sulfate residues.(Ill) Role of PM1 proteoglycan in gut morphogenesis1. LocalIzation and secretion of the PM1 proteoglycan in the blastocoel matrixThe first indication that the PM1 proteoglycan may play a specific role in gut morphogenesis was theobservation of its secretion into the matrix of the blastocoel during the time at which digestive tractdevelopment and differentiation occurs. lmmunotluorescence labelling showed the PM proteoglycan hada granular appearance and was distributed throughout the blastocoel, which suggested that it waslocalized in the network of matrix fibers present in the blastocoel ECM. Ultrastructural immunogoldlabelling studies were pursued in order to define the exact localization of this proteoglycan in the matrixnetwork. Previous studies in P. ochraceus have demonstrated that matrix fixed with Alcian blue andexamined by TEM appears as a loose meshwork containing fibers of intermediate electron density that areencrusted with short dense fibers. These shorter dense fibers change to granules after fixation with134ruthenium red (Crawford, 1989; Strathman, 1989), indicating that they contain highly anionic structuressuch as GAGs (Thyberg et a!., 1973). Although these components have not been previouslycharacterized or localized in the starfish, it has been suggested that the ECM fibers may contains collagentype I and Il, since they have a similar appearance to those described and characterized in other embryonicsystems, such as the developing chick cornea and tibia (Hendrix et aL, 1981; Hay, 1978), and that thealcian blue/ruthenium red-stained material contains proteoglycans. In the present study afterimmunogold PM1 staining, gold particles were present primarily on the short filamentous aggregatesthroughout the matrix, but not on the thicker fibrils, suggesting that these aggregates are indeedrepresentative of proteoglycans.Stronger evidence for the role of the PM1 proteoglycan as a morphogenetically important molecule forgut development comes from the observations of its developmentally regulated pattern of secretioninto the blastocoel matrix during development. It can not be determined from the present studies whetherthe synthesis of the proteoglycan is also regulated, since the means of detection was with the PM1antibody, which is directed to a carbohydrate epitope, and therefore detects the proteoglycan when it isglycosylated to its mature form prior to secretion. While early in development there are only trace amountsof the proteoglycan in the blastocoel, the amount of detectable antigen increases markedly during midgastrulation, so that over a 24 hour time period, there is a rapid accumulation of material in both theendodermal cells and the blastocoel. At the same time, development of the digestive tract begins. Thisinvolves the formation of mesenchyme cells, which are active in migration and possible remodelling of thematrix, and are involved with esophageal muscle formation. In addition, the elongation of the primitive guttube (archenteron), and its differentiation, including epithelial segmentation and bending, leading tomouth and coelom formation occurs. The time at which the PM1 proteoglycan appears in the blastocoelsuggests that it may be a “gastrulation specific” matrix component and that its functions may be centeredaround events occurring during gut formation. It is important to note that the secretion of the PM1proteoglycan into the blastocoel is not simply an indication of the time at which all blastocoel matrixcomponents are secreted. Several other matrix components, including the PC3H2 antigen (discussed135later), are abundant in the blastocoel of the blastula-stage embryo. Therefore, the observation of PM1secretion at the gastrula stage indicates that it is developmentally regulated, and the fact that it isproduced in endoderm and mesenchyme cells but not by ectodermal cells may also indicate that it hasfunctions relating to the specific activities of these cell populations.That the PM1 proteoglycan is secreted into the blastocoel at the time during which gut morphogenesisis active suggests that it may participate in cell-ECM interactions during this time. Double labelexperiments with the PM1 antibody and the P212 antibody (directed against a cell-surface antigen) havesuggested that the mesenchyme and epithelial cells have a close physical interaction with the PM1component of the blastocoel matrix during this time. These observations have important ramifications, asit is know from other systems that cellular behavior, including migration and differentiation duringmorphogenesis, is controlled not only by its developmental lineage, but also by other cells and by theECM in the milieu (Stoker et aL, 1990). In vitro, the dramatic effects of ECM on the differentiation state ofnumerous cell populations have been observed, including hepatocytes, mammary cells, keratinocytes,Sertoli cells, granulosa cells, myoblasts, myocytes and endothelial cell, to name only a few. In every case,the addition of specific ECM molecules dramatically alters both the morphology and the function of thesecells (reviewed by Damsky and Werb, 1992). For example, fibronectin has been shown to be a regulatorof keratinocyte differentiation, in that contact with fibronectin prevents the cells from undergoing terminaldifferentiation, and from expressing a marker of terminal differentiation, involucrin (Green, 1977; Watt,1984; Nicholson and Watt, 1991). Other studies have shown that mammary epithelial cells are regulatedin their capacity to produce B-casein by the ECM (Streuli and Bissell,1990). And yet other studies haveshown that laminin can regulate hepatocyte differentiation as measured by albumin transcription (Caron,1990).Evidence for the role of ECM in promoting cell migration has been shown in several in viva systems.Studies using function-blocking antibodies have shown that fibronectin is important in avian neural crestcell migration (Sheppard et aL, 1991) and amphibian mesenchyrne cell migration (Boucaut and Darribére,1361983; Boucaut et a!., 1984). In addition, thrombospondin and tenascin are important in stimulatingcerebellar granule cell migration in the brain (O’Shea eta!., 1990; Erickson and Bourdon, 1989), to nameonly a few. Although the detailed mechanisms responsible for ECM influences on cell migration anddifferentiation are not yet known in full, the general hypothesis for the ECM-cell regulatory mechanism isas follows: Extracellular signals are transmitted across the cell membrane via transmembrane receptors,such as integrins, which recognize ECM molecules (Hynes, 1987). The changes in these receptors,triggered by the ligand binding, in turn, cause rearrangements of the cytoskeletal network. This affectsthe intracellular cascade of signal transduction leading to changes in gene expression (Bissel andAggeler, 1987; Adams and Watt, 1993), and therefore in the growth and differentiation state of the cells.These studies provide evidence that cells are regulated by glycoprotein constituents of their ECMenvironment, and it is possible that the PM1 proteoglycan, which appears in the matrix just at the timewhen gut morphogenesis occurs, may play a similar role in starfish gut development. However,proteoglycans are structurally very different from glycoproteins, and it can not be assumed that theirpresence in extracellular matrices qualifies them to be key players in controlling cell migration anddifferentiation. There is, however, some evidence from in vitro systems to indicate that proteoglycans areimportant for cell migration. One study has shown that migrating endothelial cells exhibit increasedchondroitin sulfate and dermatan sulfate proteoglycan synthesis as compared to sessile cells in vitro(Kinsella and Wight, 1986). Another study has shown that the removal of these proteoglycans from earlyrat embryos inhibits emigration of cranial neural crest cells in vivo (Morris-Kay and Tuckett, 1989).However, somewhat contradictory are the findings that a mixture of the proteoglycans decorin andbiglycan inhibits the attachment of fibroblasts to fibronectin. Similarly, a chondroitin sulfate proteoglycanhas been shown to inhibit neural crest cell migration on several ECMs, and this inhibition is mediated byinteractions with cell-surface associated hyaluronic acid (Perris and Johansson, 1990). In thesesituations, the proteoglycans may inhibit migration by binding to the cell surface and hindering other celladhesion receptors or by binding matrix proteins and thereby masking cell attachment sites (Ruoslahti,1989), presumably by binding to fibronectin and in doing so, sterically hindering the fibronectin-integrin137interactions. (Lewandowska et aL, 1987). These studies show that proteoglycans are certainly legitimatecandidates as ECM regulatory proteins in processes of cell migration, and therefore a case can be madefor the PM 1 proteoglycan as a key participant in gut rnorphogenesis.2. In vivo perturbation studies with the PM1 proteoglycanThe localization of the PM1 proteoglycan in the blastocoel matrix at the time at which gutmorphogenesis is occurring, although a “pre-requisite” it it is to be considered as a morphogeneticallyimportant molecule, is not an evidence in itself. However, function blocking studies with the PM1antibody did provide strong evidence for its putative role as an important matrix component indevelopment of the gut. Embryos that were subject to treatment with the PM1 antibody at the earlygastrula stage are affected in several ways. Their overall size represents only 60% of that of the controls,indicating that blastocoel expansion has been limited. The most noticeable feature of these embryos is,however, the lack of differentiation of the digestive tract. Unlike the control embryos, which havedeveloped segments in the endodermal gut tube to yield distinct regions representing the esophagus,stomach and intestine, the PM1 perturbed embryos have a very shortened non-partitioned endodermwhich extends approximately 50% of the length of the embryo. They fail to form the blister of BL from thetip of the archenteron that in normal development may guide the archenteron to the presumptive mouthregion of ectoderm. In spite of this, mesenchyme cell formation from the tip of the archenteron does notappear to be interrupted, and several migrating cells can be observed throughout the blastocoel.Although it is not known how the PM1 antibody interferes with development, it does so presumably by aspecific interacting with the PM1 proteoglycan, and thereby blocking the normal functions of this matrixcomponent. This could be in the form of a direct interference whereby a cell matrix interaction is physicallyobstructed. Alternatively, the PM1 antibody could be preventing interactions with other matrixcomponents that are required for the general organization and/or assembly of the matrix, and which wouldthen affect cell migration and differentiation adversely.138Although few other reports exist that demonstrate the effects of function-blocking antibodies againstproteoglycans in vivo, some studies have shown that proteoglycans are important for morphogenesis byusing competition and enzymatic digestion procedures. The injection of heparin, which interferes withheparan sulphate proteoglycan-mediated interactions, retards gastrulation, and affects neuraldevelopment in Xenopus (Mitani, 1989), while injection of heparitinase into Xenopus blastulaerandomizes the development of right/left asymmetry (Yost, 1992), although it is not known whether cellsurface associated or extracellular heparan sulfate is being interfered with in these studies. Further work isrequired to clarify this area of research, but these studies do support the idea that proteoglycans can playkey roles in controlling the events of cell migration during morphogenesis. It therefore seems possiblethat similarly, during starfish development, the mesenchyme and epithelial cells involved with gutformation could be receiving important cues required for their migration and differentiation from theelements of the matrix located in the blastocoel, such as the PM1 proteoglycan.3. Conclusions and further workThis study has shown that a large proteoglycan, PM1, is secreted into the blastocoel matrix of earlygastrulae during the time at which development of the digestive tract begins, and that it is involved withthe morphogenetic events that occur during this time. Since PM1 proteoglycan-binding proteins orreceptors have not been identified, its specific role in development remains unknown. Possible functionsbased on what is known about proteoglycans could include providing resilience for blastocoel expansion,playing a role in the organization of the blastocoel matrix, binding other ECM components and growthfactors, or providing a substrate for migrating cells. The present observations indicate that the PM1antibody is directed to a carbohydrate region on the proteoglycan, and furthermore that in binding to thisregion, digestive tract morphogenesis can be inhibited in vivo. This suggests that the sugar chains on thePM1 proteoglycan may act to facilitate cell migration, differentiation and/or matrix assembly during mouth139and digestive tract formation. Further understanding of this process clearly depends on a betterknowledge of the structure of the PM1 proteoglycan, as well how it interacts with other elements of theECM or with cell surface proteins. Although the enzymatic removal of sugar groups was unsuccessful,further attempts could employ the use of chemical deglycosylation techniques, such as hydrolysis withhydrogen fluoride or trifluoromethane sulphonic acid. In this way, a thorough analysis of the sugarconstituents of the PM1 proteoglycan could be carried out. Furthermore, sequencing of the PM1proteoglycan core protein would make possible its comparison with other known proteoglycans; in thisway, common structural motifs could be identified, such as such as the arginine-glycine-asparagine cellbinding domain, or regions having growth factor activity. As the PM1 antibody is directed towards a sugarepitope, this antibody could not be used to screen a cDNA expression library. However, a polyclonalantibody could be developed against the affinity-purified proteoglycan, and then used to screen for thePM1 clone. Other studies could investigate PM1 proteoglycan binding proteins and receptors with theuse of PM 1 proteoglycan affinity columns, or with co-immunoprecipitation techniques. Knowing what thebinding sites of the PM1 proteoglycan are and how they are related to other known proteoglycans wouldaid in establishing the mechanisms by which it affects digestive tract morphogenesis.PART B: Starfish yolk and cortIcal granule proteins(I) Yolk proteins of P. ochraceusYolk proteins of oviparous species are conserved through a wide range of phylogenetically diversespecies, which suggests that they perform similar functions in many organisms. These are thought toeventually undergo catabolism to provide raw materials (amino acids, carbohydrates, and lipids) for use bythe developing embryo; however, little direct evidence exists for this. A major objective of this part of thestudy was to use the monoclonal antibody Pisaster yolk 4F8 to investigate yolk proteins of the starfish P.ochraceus. The immunogen used to generate the Pisaster yolk monoclonal was derived from a digest of140gastrula-stage embryos. The predominant yolk species in the starfish at this stage are the 64-74 kDaproteins, and it is therefore probable that the antibody was generated against these low molecular masscomponents. In a fashion similar to the PM1 antibody studies, the anti-yolk antibody was then used tobiochemically characterize and localize the yolk proteins in tissue during early embryonic and larvaldevelopment as well as in the adult starfish, to understand more clearly the role that yolk serves in thedeveloping organism. This study represents a starting point for the characterization of starfish yolkproteins.1. Specificity of the Pisaster yolk 4F8 antibodyThe PY4F8 antibody stains the granules that take up the majority of the egg volume, indicating that theantigens are stored here. These granules are distributed throughout the egg with a decreased incidencein the cortical cytoplasm, where the predominant organelles are the cortical granules. When oocyles wereviewed with the electron microscope after immunogold labelling with PY4F8 antibody, gold particles wereassociated with granules of an intermediate staining density which approach 4 jim in diameter. Whenembryos at various stages of development were stained with the PY4F8 antibody, positive-stainedgranules were present throughout the embryonic life span of the starfish. During the early larval stage,however, the staining intensity of the granules was significantly reduced, and at the same time, it did notappear as though this material was undergoing secretion, as no immunolabelling was present in any otherregions of the embryo. Western blots of egg homogenates demonstrated that the PY4F8 antibodyrecognized several proteins having apparent molecular masses of 164, 100, 74 and 64 kDa. Theseproteins, although different in size, shared a common structural feature in that they were all recognized bythe same monoclonal antibody. This type of pattern, that of a family of proteins in the egg which share acommon epitope, is characteristic of yolk proteins in other species, including nematodes (Sharrock,1984), goldfish (de Vlaming et a!., 1980) and chicken (Wallace and Morgan, 1986). In these species, theproteins are stored in yolk platelets or granules, which range from 1-4 jim in diameter and are the141predominant organelle in the egg. Typically, such yolk platelets and proteins are progressively degradedthrough development (Anderson, 1974). The similar pattern of molecular composition, localization anddepletion during development that is observed in P. ochraceus eggs suggests that the PY4F8 antibodyrecognizes starfish yolk proteins which are stored in yolk granules.2. PY4F8 Immunoreactlvlty durIng development.The PY4F8-stained granules are present through the early stages of development. The granulesapparent at the gastrula stage were comparable in intensity to those present in the unfertilized egg,indicating that a major loss of proteins and granules is not occurring at this stage. Although the stainingintensity of the yolk granules remained strong through early development, the starfish PY4F8-antigenschanged in molecular composition during this time. While in the oocyte, 4 major proteins were present,with the development to the blastula, there was a loss of the 100 kDa species. Further development tothe gastrula resulted in the loss of the 164 kDa species, leaving a considerable amount of lower molecularmass material which migrated as a broad band between 64-74 kDa. Although densitometry studies werenot performed, the immunoblot profiles indicated that as the larger proteins were lost during earlydevelopment, the amount of staining around the 64-74 kDa region steadily increased before finallydecreasing at the larval stage. Direct observations of the yolk proteins in embryonic tissue withimmunogold labelling showed that no labelling was present outside of the yolk granules, suggesting thatthe 164 and 100 kDa proteins were not being secreted into other subcellular organelles or intoextracellular spaces to become part of matrices.This pattern of change in molecular masses has been observed in the yolk proteins of the sea urchin S.purpuratus, where along with the disappearance of the 160 kDa yolk protein of the egg, glycoproteinswith a lower molecular masses appear (115, 108, 90, 83 and 68 kDa) and are gradually depleted duringdevelopment (Scott and Lennarz, 1989). Homologous sets of proteins with similar molecular mass have142been observed in the embryos of several other sea urchins such as L. pictus and A. punctulata (Kari andRottmann, 1985; Scott and Lennarz, 1989; Lee et a!., 1989), Hemicentrotus puicherrimus andAnthocidaris crassispina (Yokota and Kato, 1988), as well as in the sand dollar Dendraster excentricus(Scott and Lennarz, 1989). The biochemical composition and depletion profile during development addadditional evidence that the PY4F8 antigens of P. ochraceus are yolk proteins.It is unclear how the change in the molecular compositions of yolk proteins through developmentoccurs in this organism. In other organisms, such as the sea urchin S. purpuratus, it has been shown thatyolk proteins undergo a stepwise proteolysis, such that a single limited proteolytic cleavage of a 180 kDaegg yolk protein gives rise to 2 intermediate molecular weight glycoproteins during embryogenesis (Leeet aL, 1989). In P. ochraceus, the amount of the 164 kDa protein decreases while the 64-74 kDa proteinsappear to steadily accumulate during embryogenesis, which suggests that these lower Mr proteins may bederived from the breakdown of the larger one. This assumption can, however, not be made based on thepresent data, and further biochemical analysis of the individual proteins is necessary to determine whetherthey are related and if so, what the possible cleavage patterns are.3. Ontogeny of the Pisaster yolk proteins(a) Identification ala starfish viteiogeninYolk proteins of other species are known to be synthesized as high molecular mass precursors(vitellogenins) in organs distant from the ovary; in particular, an echinoderm yolk protein precursorvitellogenin has been located in coelomic fluid (Harrington and Ozaki, 1986) and intestines (Shyu et aL,1986) of adult sea urchin. Attempts were made to locate and identify a precursor vitellogenin of thestarfish mature yolk proteins, and to do this, coelomic fluid and intestinal tissue was examined for PY4F8-reactive proteins. These investigations revealed a 400 kDa protein was present in both intestine andcoelomic fluid isolated from adult females. Rather than a single defined band, there were several bands143giving rise to a broad area of immunoreactivity around the 400 kDa region on the blot, possibly as a resultof varying degrees of glycosylation on the same protein. In addition, several smaller proteins weredetected, which correspond to the mature yolk proteins of the egg. The shared immunoreactMty impliedthat these high molecular mass proteins share structural features with the lower molecular mass yolkproteins of the egg. This suggests that the smaller molecular mass proteins of the egg are derivatives ofthis large protein of the intestines and coelomic fluid, and supported identification of the 400 kDa speciesas a starfish vitellogenin.(b) Effect of 13-mercaptoethanol on PY4F8 immunoreactivityCoelomic and intestinal extracts which were treated with 13-mercaptoethanol prior to immunoblot analysisdid not exhibit the 400 kDa PY4F8 immunoreactive bands seen in untreated material. This suggested thatthe starfish vitellogenin contained disulfide-linked subunits which were broken up into monomers underreducing conditions. However, in the PY4F8 immunoprecipitate derived from oocyte-extracted proteins,a 400 kDa band was present even after treatment in reducing buffer. This protein was not immunoreactiveto PY4F8, but stained with Coomassie blue and with the lectins Con A and WGA. This implies thattreatment with (3-mercaptoethanol destroys the epitope, perhaps by cleaving intra-chain disulfide bonds,rather than causing the breakdown of the 400 kDa complexes. Further to this argument is the observationthat although a similar Western blot band profile was observed in oocyte extract under both reducing andnon-reducing conditions, the antigenicity of the non-reduced sample was at least 5-fold greater, indicatingthat the treatment with 13-mercaptoethanol alters the immunoreactivity of the monomeric yolk proteins insome way.(C) Vitellogenin processing and transportVitellogenins of other species vary in molecular mass from 170 kDa in the nematode (Sharroch, 1984),and 380 kDa in the goldfish (DeViaming etal., 1980), to 460 kDa in the toad (Wiley and Wallace, 1981).144After synthesis in an organ distant from the ovary such as the liver, intestine or by coelomocytes, they aresecreted into a body fluid such as blood, hemolymph or coelomic fluid, and then transported to the ovarywhere they are taken up by receptor-mediated endocytosis. The time and location of processing ofvitellogenins varies among species, and can occur prior to their secretion, during their transit to theoocyte, and/or following their selective uptake (Wang eta!., 1983; Sharrock, 1984), giving rise to matureglycoproteins. The fact that very little of the 400 kDa species was observed in the oocyte indicates thatthe amount of vitellogenin taken up by the oocytes was probably small in comparison with the lowermolecular mass derivatives, or if large quantities are taken up, they are rapidly processed to the smallermolecular mass forms. This suggests that in P. ochraceus, vitellogenin is processed into the mature yolkproteins primarily in the intestine or coelomic fluid, and that these mature proteins are then transported tothe ovary where they are taken up by the oocytes.4. Characterization of the starfish yolk proteinsVitellogenins and mature yolk proteins are typically glycoproteins rich in lipids and phosphate groups.The results of this study suggest that the extent of glycosylation of P. ochraceus yolk proteins is verysubstantial. Following treatment with Endo F, an enzyme which enzymatically cleaves asparagine (N)-linked oligosaccharides chains (typical of glycoproteins) from the core protein, major bands were presentat 100 kDa, a doublet at 55 kDa, a doublet at 45 kDa and a single band at 32 kDa, whereas in the control,bands were present at 164, 100 (doublet) and 64/74 kDa. Although direct correlations betweenglycosylated and unglycosylated bands can not be made, the fact that the largest protein in the enzymetreated lane migrates at 100 kDa suggests that the 164 kDa major yolk protein has undergone a mobilityshift of at least 60 kDa, which represents a substantial loss in apparent molecular mass. This is evengreater than that observed in the sea urchin by Armant et aL (1986), who observed a shift of 35 kDa afterthe enzymatic deglycosylation of the major yolk protein in A. punctulata.145Further evidence that the yolk antigens are glycoproteins comes from lectin binding studies. Lectins aresugar-binding proteins of non-immune origin that bind to or precipitate glycoconjugates (Goldstein et aL,1973). The ability of lectins to bind specifically and reversibly to carbohydrates has been exploited in awide variety of techniques used to study glycoproteins and oligosaccharides. There are 2 general classesof N-glycosidically linked oligosaccharide chains, one containing a polymannose chain and a secondcomposed of a complex chain that includes mannose residues but which terminates in sialic acid units.Those oligosaccharide chains linked by serine/threonine residues (0-linked) generally do not containmannose residues (Lennarz, 1983). Con A binds to a-mannose and a glucose residues (So andGoldstein, 1969), and as N-linked oligosaccharides of both the polymannose and complex types are richin mannose residues, Con A is often used to identify glycoproteins with these sugar chains (Beeley,1985). Con A binding of the Pisaster yolk proteins revealed that both the mature yolk proteins as well asthe vitellogenin are N-linked polymannose containing glycoproteins, as are yolk proteins from otherspecies, including nematodes (Winter, 1992) the leech (Baert et aL, 1991), and sea urchins (Scott andLennarz, 1989).Another lectin, WGA, has a different sugar-binding specificity, in that it preferentially binds to sialic acidand N-acetyl glucosamine residues (Allen et aL, 1973; Greenaway and LeVine, 1973). As mentionedabove, terminal sialic acid residues are characteristic of complex type N-linked oligosaccharides and of 0-linked oligosaccharides, but not of polymannose-type chains. When the Pisaster yolk proteins werestained with WGA, only the 400 and 164 kDa forms stain positively, suggesting that the lower molecularmass proteins are of the polymannose type (i.e. do not contain terminal sialic acid residues). This indicatesthat although the family of Pisaster yolk proteins share structural similarities, in that they all contain thePY4F8 epitope, there appear to be variations within their oligosaccharide chains.1465. Utilization of yolk proteinsAlthough yolk granules in certain other embryonic systems are known to provide a store of nutrients andenergy (Williams, 1967), the exact mode of utilization of the major yolk proteins by the growing embryo hasnot been clearly established. There have been several reports that material stored in yolk granules issecreted to the ECM. For example, the protein toposome identified in sea urchin eggs (No)) et al., 1985)is stored in yolk granules, and is released to become part of the hyaline layer during embryogenesis(Gratwohl eta!., 1991). Another study (Outenreath eta!., 1988) identified a lectin in yolk granules ofXenopus that was secreted into the ECM of early embryos. In the present study, proteins recognized bythe PY4F8 antigen were found exclusively in the yolk granules of oocytes, and during embryogenesis,localization of the yolk proteins remained in these granules. No evidence of the yolk protein was evidentin other subcellular compartments or in extracellular matrices, suggesting that the yolk proteinsrecognized by PY4F8 are not secreted into these areas, or if they are, the secreted antigens are alteredsuch that the epitope is no longer present. This does not rule out the possibility that other proteins arestored in the yolk granules that may be secreted in a similar way to those described above. While Pisastereggs do contain at least one protein that is secreted into extracellular matrices during development, muchlike the above-mentioned proteins (the PC3H2 antigen; see below), this protein is not stored in yolkgranules but is a component of the cortical granules.While the traditional notion of yolk is that it is serves as a nutrient store during early development, thereseems to be little change in the amount of yolk present in the early stages of Pisaster. The proteins andthe yolk granules appear to undergo depletion only once the organism has reached the feeding larvalstage. This suggests that the principle function of yolk may be as a nutrient store for the larva, rather thanfor the early embryo. A similar observation has been made of sea urchin yolk proteins, where it has beenfurther observed that larvae given a supply of food deplete their yolk much more efficiently than thosewhich are starved (Scott eta!., 1990). This could indicate that when no food is available for the feedinglarvae, the yolk material may be used sparingly, to ensure survival until which time the larvae obtains a147suitable external food source. However, if the larvae have a readily available supply of food, the yolkmaterial may in such circumstances, be used towards rapid growth and development, as there would beno requirement for an internal nutrient store. It would be of interest to study the yolk of non-feedinglarvae, such as Crossasterpaposus or Hipasteria spinosa, and compare the structures and depletionprofiles of their yolk proteins with those of feeding larvae. These species give rise to large eggs (about 1mm in diameter) which are very yolky compared with eggs of feeding larvae. The fact that these eggs areabout 5 times as large, and contain much more yolk than those which undergo a feeding larval stage(Strathmann, 1987), would seem to indicate that the yolk in this case is required for nourishment duringthe period of larval development, since they are unable to feed. Further characterization of starfishvitellogenin and its derivatives, as well as examining different conditions under which yolk proteins of thestarfish are depleted and the rates at which this occurs may provide further information on how theseproteins are utilized for nutrition.6. ConclusionsAlthough yolk proteins from many species have been studied extensively, and as a result, theirontogeny, transport and uptake mechanisms including receptor characterization is known, a key questionstill remains in the field of yolk biology, and that centers around the biochemistry of yolk protein utilization.The depletion profile of the yolk in P. ochraceus is suggestive that, like several other species, yolk isprobably not important during early stages of development, the time during which it is generally assumedto be utilized. The study has also shown that a family of yolk proteins recognized by the PY4F8 antibodyare not secreted into the ECM or other subcellular organelles, suggesting that they are degraded fornutritive stores. Since the depletion is not obvious until the feeding larval stage is reached, the yolk mayrepresent a safety store of food in the case that the larva does not encounter external food sources at thisstage in development. Further studies are necessary to understand the biology of yolk utilization instarfish and in other species. In this respect, the isolation of yolk granule enzymes, which may be148responsible for the limited proteolysis of the yolk proteins, may provide clues to their depletion and how itis regulated.(II) A cortical granule antigen of P. ochraceusThis final part of the study involved the identification and immunohistochemical analysis of a corticalgranule protein in the unfertilized egg and early embryo. This was of interest for 3 major reasons. First,immunohistochemistry was used to investigate the storage, synthesis and secretion patterns of anotherblastocoel ECM component, and to compare this with that of the PM1 proteoglycan. Secondly, it was ofinterest to compare the storage and secretion of another type of granule in the egg with that of theubiquitous yolk granules. And finally, the investigation of a component of starfish cortical granules wasundertaken to study possible functions based on its localization during egg activation and earlydevelopment, since cortical granule proteins have not been identified in the starfish until now.1. The PC3H2-antigen is a cortical granule protein.A general characteristic of cortical granules is their diverse morphology, including shapes resemblingspindles, tubules, dumb-bells or pears (reviewed by Anderson, 1972). There are 2 common featuresamong cortical granules of many different species. The first is that they are positioned in the corticalcytoplasm of mature eggs, and the second is that they have a very electron-dense component to themthat is unmatched by any other organelle in the ooctye (Anderson, 1968; Bal, 1970; Schuel, 1978). In thepresent study, the PC3H2 antibody-labelled granules were located primarily in the peripheral cytoplasm ofPisaster eggs, and had a distribution similar to those of cortical granules. When eggs were viewed with theelectron microscope, gold-labelled antibody identified granules of about 1.5 im in diameter having avaried morphology consisting of an electron dense outer component, and an electron translucent149component towards the core with patches of intermediate stain density interspersed throughout thegranules. These are morphologically similar to cortical granules that have previously been described instarfish eggs, including Patina miniata (Holland, 1980), Pisaster ochraceus (Crawford and Abed, 1986),and Marthasterias glacialls (Sousa and Azevedo 1989), and somewhat similar to cortical granules in otherspecies such as hamsters (Cherr et aL, 1988) and sheep (Cran et aL, 1988). These observationstherefore indicate that the PC3H2 antibody is directed towards a component of starfish cortical granules.2. The PC3H2 antigen Is released during the cortical reactionSince the observations of Harvey (1911) that the cortical granules of sea urchins are no longer visiblesubsequent to egg activation, investigators have agreed that the cortical granules of many organisms arereleased during fertilization and play a role in blocking polyspermy. However, only a few components ofcortical granules from a handful of species have been characterized, and many of their functions duringthe cortical reaction remain unknown. Furthermore, several organisms have cortical granules which do notparticipate in the cortical reaction at all, such as the the annelid Chaetopterus, the amphineuran molluscMoalia and the mussel Mytilus (reviewed by Anderson, 1972). Some of the components of corticalgranules are involved with the formation of the fertilization membrane. For example, the cortical granulelectin of Xenopus eggs is a metalloglycoprotein that combines with jelly coat components to form thefertilization membrane (Nishihara et a!., 1986). Other cortical granule constituents identified includefucosyl and sialyl-rich glycoconjugates (Lee eta!., 1988), and ovoperoxidase (Gulyas and Schmell, 1980)in the mouse; sulfated acid mucopolysaccharides in sea urchins and starfish (Schuel et aL, 1974; Sousaand Azevedo, 1989); and proteases (Alliegro and Schuel, 1988), 8-1 ,3-glucanase (Wessel et a!.,1987),and and hyalin (Hylander and Summers, 1982), in sea urchins. In the present study, the PC3H2 antibody,which recognized a 125 kDa protein located within the cortical granules, was used to determined ifPisaster cortical granules undergo exocytosis during the cortical reaction, and if so, what the fate of thiscomponent was during early development. Shortly after fertilization, the PC3H2 antigen was released150from the granules located in the peripheral cytoplasm, and labelling was detected in the perivitellinespace, the space intervening between the egg plasma membrane and the fertilization membrane. Theantigen was not detected in the fertilization membrane, and by the blastula stage, no label in theperivitelline space was evident, indicating that it may have a transient role during fertilization.3. The PC3H2 antigen is present in the early embryoit was first reported by Anderson (1972) that when cortical granules are present, depending on theorganism, they may or may not participate in the cortical reaction, and it was suggested that there were 2types of cortical granules, one of fertilization, and one not yet classified. In P. octiraceus, while during eggmaturation, many of the cortical granules not already at the periphery move towards it, there are severalPC3H2-stained granules which remain in the subcortical and central cytoplasm, and appear to be randomlyscattered, interspersed with the yolk granules. Investigations at the TEM level have shown that they havea similar morphology to those in the cortical region, indicating that they are probably members of the samegranule population. This is confirmed by the fact that they are also stained with PC3H2 immunogold label.In addition, observations from several different batches of oocytes have revealed a consistent pattern,indicating it is not random. These granules that are left behind do not undergo exocytosis at fertilization,but are still present later in development in the blastula and gastrula stages, and when viewed with theelectron microscope, have a similar morphology to the peripherally located granules. Furthermore,PC3H2- immunogold label was found over the same electron-dense regions, suggesting that they are thesame type of granules.However, while these cortical granules that have been “left behind” do not secrete material during eggactivation, they do appear to secrete the PC3H2 antigen into the blastocoel, the basment membranes andone region of the hyaline layerduring embryogenesis. In addition, immunogold electron microscopyshows gold label associated with Golgi complexes. The label is concentrated in the trans Golgi near the151lateral edges, and in most cases, the Golgi are in close proximity to at least one cortical granule. Goldparticles were observed just on the outer sides of the cortical granule membranes as well, which suggeststhat these granules may be serving as storage sites for newly synthesized PC3H2 antigen. The presentobservations indicated that the starfish cortical granules release the PC3H2 antigen at 2 different times,the first occurring at egg activation, and the second during embryogenesis, where it becomes a majorcomponent of all 3 extracellular matrices in the embryo, the hyaline layer, the blastocoel ECM, and thebasement membranes. A similar observation has been made in the sea urchin with the protein hyalin.This protein, although located in cortical granules which are exocytosed at fertilization, is also present inrandomly distributed cortical vesicles, which do not move to the periphery during egg maturation, nor arethey exocytosed during the cortical reaction (Hylander and Summers, 1982). These granules are thoughtto function in the early embryo as a hyalin reservoir, in the case that renewal of the hyaline layer is required.4. ConclusIonsThe present observations indicate that one of the components of egg cortical granules is a 120 kDaprotein that is released upon egg activation into the perivitelline space, but which is not incorporated intothe fertilization envelope, as are several other components of cortical granules in other species. Thisindicates that while the PC3H2 antigen may be involved in blocking polyspermy, it does not contribute to apermanent structural barrier. While all cortical granules located in the peripheral egg cytoplasm appear toundergo exocytosis, some cortical granules resist movement to the cell periphery and release theircontents later in development. One of these components, the PC3H2 antigen, is secreted to become apart of the ECM of the blastocoel, basement membranes, as well as one region of the hyaline layer. It isunknown whether this protein has a similar function at fertilization and during embryonic development.However, during the cortical reaction, its localization in the perMtelline space is transient, whereas duringdevelopment, it clearly remains as a structural component of the ECM.152FINAL SUMMARY AND CONCLUSIONS (Parts A & B)This study has investigated several components of starfish embiyonic ECM and egg storage granules,with the intentions of studying how ECM affects cell behavior during morphogenesis. A major part of thework presented investigated the role of one particular matrix component in starfish gut morphogenesis.By generating monoclonal antibodies to starfish gastrula extract, an antibody specific for matrix in theblastocoel was isolated, the PM1 antibody. Using this antibody as a research tool, a large extracellularcomponent of the blastocoel ECM (estimated Mr > 600 kDa) was identified. This component wasdetermined to be a proteoglycan, based on the fact that it was stained with Alcian blue under conditions ofhigh ionic strength. In addition, the antigen was sensitive to trypsin, indicating that it contained apolypeptide component. And finally, its assembly was disrupted by the inhibitor of proteoglycansynthesis, B-D-xyloside. Further investigations at the histochemical level revealed that the PM1proteoglycan is synthesized and/or undergoes maturation first during early gastrulation. Secretion intothe blastocoel does not occur until mid-gastrulation, suggesting that it may have a specific role in themorphogenesis of the digestive tract which begins here at that time. Function blocking studies using thePM1 antibody in embryo cultures resulted in severely malformed embryos when compared to the controls,with a marked effect on digestive tract morphogenesis. This suggests that the PM1 proteoglycan plays aspecific role in the development, perhaps by providing resilience for blastocoel expansion, playing a rolein the organization of the blastocoel matrix, binding other ECM components and growth factors, orproviding a substrate for migrating cells.Another component of the blastocoel ECM, the PM3H2 antigen, has a distribution in early developmentthat is very different from that of the PM1 proteoglycan. This antigen is present in unfertilized oocytes,where is it stored in the cortical granules. Although a significant amount of the antigen is exocytosedduring the cortical reaction which occurs shortly after fertiliation, it also appears to be secreted into variousECMs during the course of embryonic development. Whereas the PM1 antigen is secreted into the ECM153of the gastrula-stage embryo, the PC3H2 antigen first appears in the blastocoel at the blastula-stage. Thissecretion does not appear to be dependent upon new transcription, as the maternally-derived antigenstored in cortical granules is released into the ECM. This is in contrast to the PM1 antigen, which is notsynthesized or matured until the gastrula stage. Although the PC3H2 antigen localized to severaldifferent ECMs, it was found in the same network of blastocoel matrix as the PM1 proteoglycan. The 2different storage and secretion profiles of these ECM components suggest that they may functiondifferently during morphogenesis. Further characterization of the PC3H2 antigen will bring us to a closerunderstanding of its function during fertilization and development.Although there are reports of protein stored in yolk granules that contribute to ECMs during earlydevelopment, this does not appear to be the case for the family of yolk glycoproteins recognized by thePY4F8 antibody. The results indicate that these Pisaster yolk proteins have a similar localization anddepletion profile as those “classical” yolk proteins documented in many other species. The PY4F8recognizes several major yolk proteins of the egg that are derived from a 400 kDa vitellogenin precursor,and like yolk proteins in other species, the vitellogenin appears to be synthesized and/or stored in theintestines and coelomic fluid. As the smaller yolk proteins are also detected in the intestine and coelomicfluid, there is the possibility that the vitellogenins are broken down at these remote sites and transportedto the ovary in this form. Biochemical studies have shown that the starfish yolk proteins are considerablyrich in glycoproteins, as endoglycosidase F digestion causes a significant shift in electrophoretic mobilitiesof the proteins. In addition, Con A and WGA binding to the proteins suggests that they contain N andmaybe 0-linked oligosaccharide chains. Although the PY4F8 antibody recognizes the classical yolkproteins, which are thought to provide a store of nutrients for the developing antibody by undergoingcatabolism, they are not depleted to a significant degree until larval development begins, which indicatesthat these proteins may be represent a safety store of food for the feeding larva. This study by no meansrules out the possibility that other proteins are stored in the yolk granules, which may contribute to ECMsin a similar way as has been reported in several other organisms.154REFERENCESAbed, M. and Crawford, B. J. (1986). Ultrastructural aspects of mouth formation in the starfish Pisasterochraceus. J. Morphol. 188, 239-250.Adams, J. C. and Watt, F. M. (1993). Regulation of development and differentiation by the extracellularmatrix. Development 117, 1183-1198.Adelson, D. L. and Humphreys, T. (1988). Sea urchin morphogenesis and cell-hyalin adhesion areperturbed by a monoclonal antibody specific for hyalin. Development 104, 391-402.Akasaka, K., Amemiya, S., Terayama, H. (1980). Scanning electron microscopical study of the inside ofsea urchin embryos (Pseudocentrotus depressus). Effects of aryl 13-xyloside, tunicamycin, anddeprivation of sulfate ions. Exp. Cell Res. 129, 1-13.Akasaka, K. and Terayama, H. (1983). Sulfated glycan present in the EDTA extract of Hemicentrotusembryos (mid-gastrula). Exp. Cell Res. 146, 177-185.Allen, A. K., Neuberger, A., Sharon, N. (1973). The purification, composition and specificity of wheat-germ agglutinin. Biochem. J. 131, 155-162.Alliegro, M. C., Black, S. D., McClay, D. R. (1992). Deployment of extracellular matrix proteins in sea urchinembryogenesis. Microsc. Res. Tech. 22, 2-10.Alliegro, M. C. and Shuel, H. (1988). lmmunochemical localization of the 35-kDa sea urchin egg trypsin-likeprotease and its effects upon the egg surface. Dev. Biol. 125, 168.Anderson, E. (1968). Oocyte differentiation in the sea urchin, Arbacia punctulata, with particular referenceto the origin of cortical granules and their participation in the cortical reaction. J. Cell BioL 37, 514-539.Anderson, E. (1972). Oogenesis. (J. D. Biggers, A. W. Schuetz, Eds), pp. 87-117, Univ. Park Press,Baltimore, Maryland.Anderson, E. (1974). Comparative aspects of the ultrastructure of the female gamete. mt. Rev. Cytol.Supp!. 4, 51-56.155Armant, D. R., Carson, 0. 0., Decker, G.L., Welply, J. K., Lennarz, W. J. (1986). Characterization of yolkplatelets isolated from developing embryos of Arbacia punctulata. 0ev. BioL 113, 342-355.Baert, J.-L., Britel,M., Slomianny, M.-C., Delbart, C., Fournet, B., Sautiere, P. and Malecha, J. (1991). Yolkproteins in leech. Identification, purification and characterization of vitellin and vitellogenin. Eur. J.Biochem.. 201, 191-198.Baginski, R. M., McBlaine, P. J., Carroll, E. J. Jr. (1982). Novel procedures for collection of sea urchin eggcortical granule exudate: partial characterization and evidence for post-secretion processing. Gamete Res.6, 39-45.Baginski, R. M., McBlaine, P. J., Carroll, E. J. Jr. (1982). Novel procedures for collection of sea urchin eggcortical granule exudate; partial characterization and evidence for post-secretion processing. Gamete Res.6, 39-46.Bal, A. K. (1970). Selective staining of the ultrastructural components of cortical granules and Golgicisternae of sea urchin eggs. Z Zellforsch 104, 471-477Beeley, J. G. (1985). Glycoprotein and proteoglycan techniques. in “Laboratory techniques inbiochemistry and molecular biology” (R. H. Burdon and P. H. vanKnippenberg, Ed), Vol. 16, pp. 5-28,Elsevier, Amsterdam.Bergink, E. W. and Wallace, R. A. (1974). Precursor-product relationship between amphibian vitellogeninand the yolk proteins, lipovitellin and phosvitin. J. Biol. Chem. 249, 2897-2903.Bianco, P., Fisher, L. W., Young, M. F., Termine, J. D. and Gehron Robey, P. (1990). Expression andlocalization of the two small proteoglycans biglycan and decorin in developing human skeletal and non-skeletal tissues. J. Histochem. Cytochem. 38, 1549-1563.Bissell, M. J. and Aggeler, J. (1987). Dynamic reciprocity: how do extracellular matrix and hormones directgene expression. “Mechanisms of Signal Transduction by Hormones and Growth Factors”. (M. Cabot,Ed), pp. 251-262. Liss, New York.Bobbitt, J. M. (1956). Periodate oxidation of carbohydrates. Adv. Carbohyd. Chem. 11, 1-41.156Boucaut, J.-C. and Darribére, T. (1983). Fibronectin in early amphibian embryos: Migrating mesodermalcells are in contact with a fibronectin-rich fibrillar matrix established prior to gastrulation. Cell Tissue Res.234, 135-145.Boucaut, J. C., Darribere, T., Boulekbache, H., Thiery, J. P. (1984). Prevention of gastrulation but notneurulation by antibodies to fibronectin in amphibian embryos. Nature (London) 307, 364-366.Bouvet, J.-P., Pires, R., Pillot, J. (1984). A modified gel filtration technique producing an unusualexclusion volume of 1gM: A simple way of preparing monoclonal 1gM. J. Immunoh Meth. 66, 299-305.Bownes, M. (1986). Expression of genes coding for vitellogenin (yolk protein). Annu. Rev. Entomol. 31,507-531.Brower, D. L., Piovant, M., Salatino, R., Brailey, J., Hendrix, M. J. C. (1987). Identification of a specializedextracellular matrix component in Drosophila imaginal disc. Dev. Biol. 119, 373-381.Byrne, B. M., Gruber, M. and G. AB. (1989). The evolution of egg yolk proteins. Prog. Biophys. Molec.Biol. 53, 33-69.Campbell, S. S., Crawford, B. J. and Reimer, C. L. (1991). A simple ethanol-based freeze-substitutiontechnique for marine invertebrate embryos which allows retention of antigenicity. J. Microscopy 164, 197-215.Carbonetto, S., Gruver, M. M. and Turner, D. C. (1983). Nerve fiber growth in culture on fibronectin,collagen and glycosaminoglycan substrates. J. NeuroscL 3, 2324-2335.Caron, J. M. (1990). Induction of albumin gene transcription in hepatocytes by extracellular matrix. Mo!.CelL BioL 10, 1239-1243.Cervello, M. and Matranga, V. (1989). Evidence of a precursor-product relationship between vitellogeninand toposome, a glycoprotein complex mediating cell adhesion. Cell Different. Develop. 26, 67-76.Cherr, G. N., Drobnis, E. Z., Katz, D. F. (1988). Localization of cortical granule constituents before andafter exocytosis in the hamster egg. J. Exp. ZooL 246, 81-93.157Cran, D. G., Moor, R.M., Irvine, R. F. (1988). Initiation of cortical reaction in hamster and sheep oocytes inresponse to inositoltrisphosphate. J. Cell ScL 91, 139-144.Cran, D. G. and Esper, C. R. (1990). Cortical granules and the cortical reaction in mammals. J. Reprod.Fert., SuppL42, 177-188.Crawford, B. and Abed, M. (1986). Ultrastructural aspects of the surface coatings of eggs and larvae of thestarfish, Pisaster ochraceus, revealed by Alcian blue. J. Morphol. 187, 23-37.Crawford, B. J. (1972). The ultrastructure and differentiation of chick pigmented retinal cells in clonalculture. PhD Thesis, University of Washington.Crawford, B. J. (1989). Ultrastructure of the basal lamina and its relationship to extracellular matrix ofembryos of the starfish Pisasterochraceus as revealed by anionic dyes. J. Morphoh 199, 349-361.Crawford, B. J. (1990). Changes in the arrangement of the extracellular matrix, larval shape, andmesenchyme cell migration during asteroid larval development. J. Morphol. 206, 147-161.Crawford, B. J. and Abed, M. (1983). The role of the basal lamina in mouth formation in the embryo of thestarfish Pisaster ochraceus. J. MorphoL 176, 235-246.Crawford, T. J. and Crawford, B. J. (1992). Characterization and localization of large sulfated glycoproteinsin the extracellular matrix of the developing asteroid Pisaster ochraceus. Biochem. Cell BioL 70, 91-98.Damsky, D. H. and Werb, Z. (1992). Signal transduction by integrin receptors for extracellular matrix:cooperative processing of extracellular information. Curr. Opin. Cell BioL 4, 772-781.Deuchar, E. M. (1963). Tracing amino acids from yolk protein into tissue protein. Acta Embryol. MorphohExp.6, 311-323.DeVlaming, V. L., Wiley, H. S., Delahunty, G., Wallace, R. A. (1980). Goldfish (Carassius auratus)vitellogenin: induction, isolation, properties and relationship to yolk proteins. Comp. Biochem. Physiol.87B, 613-623.Duskin, D. and Mahoney, W. C. (1982). Relationship of the structure and biological activity of the naturalhomologues of tunicamycin. J. B10L Chem. 257, 3105-3109.158Ekblom, P., Vestweber, D. and Kemler, R. (1986). Cell-matrix interactions and cell adhesion duringdevelopment. Ann. Rev. Cell Blot 2, 27-47.Elbein, A. D. (1987). Inhibitors of the biosynthesis and processing of N-linked oligosaccharide chains.Ann. Rev. Biochem. 56, 497-534.Elder, J. H. and Alexander, S. (1982). endo-B-N-acetylglucosaminidase F: Endoglycosidase fromFlavobacterium meningosepticum that cleaves both high-mannose and complex glycoproteins. Proc.NatL Acad. Sc!. USA 79, 4540-4544.Endo, Y. and Noda, Y. D. (1977). Ultrastructure of blastocoel of sea urchin embryos. Zoot Mag. 86, 309.Engel, J. (1991). Common structural motifs in proteins of the extracellular matrix. Cur. Opin. Cell Blot 3,779-785.Erickson, H. P. and Bourdon, M. A. (1989). Tenascin: a extracellular matrix protein prominent inspecialized embryonic tissues and tumors. Annu. Rev. Cell Blot 5, 71-92.Esko, J. D. (1991). Genetic analysis of proteoglycan structure, function and metabolism. Curr. Opin. CellBiol. 3, 805-81 6.Frens, G. (1973). Controlled nucleation for the regulation of the particle size in monodisperse goldsuspensions. Nat. Phys. Sc, 241, 20-22.Gallagher, J. T. (1989). The extended family of proteoglycans: social residents of the pericellular zone.Curr. Opin. Cell Blot 1, 1201-1218.Galligani, L., Hopwood, J., Schwartz, N. B., Dorfman, A. (1975). Stimulation of synthesis of freechondroitin sulfate chains by B-D xylosides in cultured cells. J. Biol. Chem. 250, 5400-5406.Goetinck, P. F. (1985). Developmental Mechanisms: Normal and Abnormal. (L. W. Lash, Ed), pp. 137-1 47Liss, New York.Goetinck, P. F. and Carlone, D. L. (1988). Altered proteoglycan synthesis disrupts feather patternformation in chick embryonic skin. Dev. Blot 127, 179-186.159Goetinck, P. F. and Winterbottom, N. (1991). “Biochemistry and Physiology of the Skin”. (L. Goldstein,Ed), Oxford Univ. Press, New York.Goldstein, I. J., Reichert, A. M., Misaki, A., Gorin, P. A. J. (1973). An “extension” of the carbohydratebinding specificities of concanavalin A. Biochim. Biophys. Acta 317, 500-504.Goldstein, J. L., Anderson, R. G. W., Brown, M. S. (1979). Coated pits, coated vesicles, and receptor-mediated endocytosis. Nature (London) 279, 679-685.Gomori (1955). Preparation of buffers for use in enzyme studies. Meth. Enzymoh 1, 141.Gratwohl, E. K.-M., Kellenberger, E. Lorand, L., Noll, H. (1991). Storage, ultrastructural targeting andfunction of toposomes and hyalin in sea urchin embryogenesis. Mech. Devel. 33, 127-138.Green, H. (1977). Terminal differentiation of cultured epidermis cells. Cell 11, 405-41 6.Greenaway, P. J. and LeVine, D. (1973). Binding of N-acetyl-neuraminic acid by wheat-germ agglutinin.Nature New Biol. 241, 191-192.Grobstein, C. (1954). Tissue interaction in the morphogenesis of mouse embryonic rudiments in vitro, in“Aspects of Synthesis and Order of Growth”. (G. Rudnick, Ed), pp. 233-256. Princeton University Press,Princeton.Gulyas, B. J. (1980). Cortical granules of mammalian eggs. mt. Rev. Cytoh 63, 357-392.Gulyas, B. J. and Schmell, E. D. (1980). Ovoperoxidase activity in ionophore treated mouse eggs. I.Electron Microscopic Locallzation 3, 267-278.Habuchi, H., Kumata, K., Suzuki, S. (1986). Changes in proteoglycan composition during development ofrat skin: The occurrence in fetal skin of a chondroitin sulfate proteoglycan with high turnover rate. J. BioLChem. 261, 1031-1040,Hahn, R. A. and Birk, D. E. (1992). B-D xyloside alters dermatan sulfate proteoglycan synthesis and theorganization of the developing avian corneal stroma. Development 115, 383-393.160Hamazaki, T. S., Nagahama, Y., luchi, I., Yamagami, K. (1989). A glycoprotein from the liver constitutes theinner layer of the egg envelop (zona pellucida interna) of the fish, Oryzias latipes. Dev. Biol. 133, 101-110.Hantaz-Ambroise, D., Vigny, M., Koenig, J. (1987). Heparan sulfate proteoglycan and laminin mediate twodifferent types of neurite outgrowth. J. Neurosci. 7, 2293-2304.Har-El, R. and Tanzer, M. L. (1993). Evolution of the extracellular matrix in invertebrates. FASEB J. 7,1115-1123.Hardingham, T. E. and Bayliss, M. T. (1990). Proteoglycans of articular cartilage changes in aging and injoint disease. Semin. Arth. Rheum., Supph 1, 12-33.Hardingham, T. E. and Fosang, A. J. (1992). Proteoglycans: many forms and many functions. FASEB J. 6,861 -870.Harrington, F. E. and Easton, D. P. (1982). A putative precursor to the major yolk protein of the sea urchin.Dev. 8101. 94, 505-508.Harrington, F. E. and Ozaki, H. (1986). The major yolk glycoprotein precursor in echinoids is secreted bycoelomocytes into the coelomic plasma. Cell Differ. 19, 51-57.Harvey, E. B. (1956). The American Arbacia and other sea urchins. Princeton University Press, Princeton.Harvey, E. N. (1911). Studies on the permeability of cells. J. Exp. Zool. 10, 507-.556.Hascall, V. C. and Hascall, G. K. (1981). Proteoglycans. “Cell Biology of Extracellular Matrix”. (E. D. Hay,Ed), pp. 39-63. Plenum Press, New York.Hay, E. D. (1978). Fine structure of embryonic matrices and their relation to the cell surface in rutheniumred-fixed tissues. Growth 42, 399-423.Hay, E. D. (1981). Collagen and Embryonic Development. “Cell Biology of Extracellular Matrix”. (E. D. Hay,Ed), pp. 379-405. Plenum Press, New York.161Hay, E. D. (1991). Collagen and other matrix glycoproteins in embryogeriesis. “Cell Biology of ExtracellularMatrix, Second Ed.”. (E. D. Hay, Ed), pp. 419-462. Plenum Press, New York.Hendrix, M. J., Hay, E. D., Von der Mark, K., Linsenmayer, T. F. (1981). Immunohistochemical localizationof collagen types I and II in the developing chick cornea and tibia by electron microscopy. Invest.Ophthalmoh Vis. Sc!. 22, 359-375.Hoffman, S., Crossin, K. L., Edelman, G. M. (1988). Molecular forms, binding functions, anddevelopmental expression patterns of cytotactin and cytotactin-binding proteoglycan, an interactive pairof extracellular matrix molecules. J. Cell Biol. 106, 5 19-532.Holland, N. 0. (1980). Electron microscopic study of the cortical reaction in eggs of the starfish (Patinaminiata). Cell Tissue Res. 205, 67-76.Hylander, B. L., Summers, R. G. (1982). An ultrastructural immunocytochemical localization of hyalin in thesea urchin egg. 0ev. Sb!. 93, 368-380.Hynes, R. 0. (1987). Integrins: A family of cell surface receptors. Ce1148, 549-554.Hynes, R. 0. (1992). lntegrins: Versatility, modulation and signalling in cell adhesion. Cell 69, 11-25.lijima, Y., Muramatsu, T., Egami, F. (1971). Purification of x-L-fucosidase from the liver of a marinegastropod, Turbo cornutus, and its action on blood group substances. Arch. Biochem. Biophys. 145, 50-54.Jackson, R. L., Busch, S. J., Cardin, A. 0. (1991). Glycosaminoglycans: Molecular properties, proteininteractions, and role in physiological processes. Physiol. Rev. 71, 481-539.Johnson, G. D., Davidson, R. S., McNamee, K. C., Russell, G., Goodwin, D., Holborow, E. J. (1982).Fading of immunofluorescence during microscopy: a study of the phenomenon and its remedy. J.Immuno!. Methods 55, 231-242.Kannangara, T., Wieczorek, A. and Lavender, D. P. (1989). lmmunoaffinity columns for isolation ofabscisic acid in conifer seedlings. Physiologica, Plantarum. 75, 369-373.162Karasaki, S. (1963). Studies on amphibian yolk. 5. Electron microscopic observations on the utilization ofyolk platelets during embryogenesis. J. Ultrastruct. Res. 9, 225-247.Kari, B. E. and Rottmann, W. L. (1985). Analysis of changes in a yolk glycoprotein complex in thedeveloping sea urchin embryo. Dev. Biol. 108, 18-25.Kariya, Y., Watabe, S., Hashimoto, K. and Yoshida, K. (1990). Occurrence of chondroitin sulfate E inglycosaminoglycan isolated from the body wall of sea cucumber Stichopus japonicus. J. 8101. Chem. 265,5081-5085.Katow, H. and Solursh, M. (1979). Ultrastructure of blastocoel material in blastulae and gastrulae of the seaurchin Lytechinus pictus. J. Exp. Zool. 210, 561 -567.Katow, H. and Solursh, M. (1981). Ultrastructural and time-lapse studies of primary mesenchyme cellbehavior in normal and sulfate-deprived sea urchin embryos. Exp. Cell Res. 136, 233-245.Katow, H. and Solursh, S. (1981). Ultrastructural and time-lapse studies of primary mesenchyme cellbehavior in normal and sulfate-deprived sea urchin embryos. Exp. Cell Res. 136, 233-245.Kawabe, T. T., Armstrong, P. B. and Polluck, E. G. (1981). An extracellularfibrillar matrix in gastrulating seaurchin embryos. Dev. 8101. 85, 509-5 15.Kimble, J. and Sharrock, W. J. (1983). Tissue specific synthesis of yolk proteins in Caenorhabditiselegans. Dev. 810!. 96, 189-196.Kinsella, M. G. and Wight, T. N. (1986). Modulation of sulfated proteoglycan synthesis by bovine aorticendothelial cells during migration. J. Biol. Chem 102, 679-687.Kjellén, L. and Lindahl, U. (1991). Proteoglyans: Structure and interactions. Annu. Rev. Biochem. 60,443-475.Kornfeld, F. and Kornfeld, S. (1985). Assembly of asparagine-linked oligosaccharides. Annu. Rev.Biochem. 54, 631-664.Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophageT4. Nature (London) 227, 680-685.163Lane, M. C. and Solursh, M. (1991). Primary mesenchyme cell migration requires a chondroitinsulfate/dermatan sulfate proteoglycan. Dev. Biol. 143, 389-397.Lee, G., Hynes, H., Kirschner, M. (1984). Temporal and spatial regulation of fibronectin in early Xenopusdevelopment. Cell 36, 729-740.Lee, G. F., Fanning, E. W., Small, M. P., Hille, M. B. (1989). Developmentally regulated proteolyticprocessing of a yolk glycoprotein complex in the embryos of the sea urchin Strongylocentrotuspurpuratus. Cell Differ. Dev. 265-18,Lee, G. F., Fanning, E. W., Small, M. P., Hille, M. B. (1989). Developmentally regulated proteolyticprocessing of a yolk glycoprotein complex in the embryos of the sea urchin Strongylocentrotuspurpuratus. Cell Differ. 0ev. 265-18,Lee, S. H., Ahuja, K.K., Gilburt, D. J. and Whittingham, D. G. (1988). The appearance of glycoconjugatesassociated with cortical granule release during mouse fertilization. Development 102, 595-604.Lelongt, B., Makino, H., Dalecki, T. M., Kanwar, Y. S. (1988). Role of proteoglycans in renal development.0ev. Biol. 128, 256-276.Lennarz, W. J. (1983). Glycoprotein synthesis and embryonic development. CRC Critical Review inBiochemistry 14, 257-272,Lewandowska, K., Choi, H. U., Rosenberg, L. C., Zardi, L., Gulp, L. A. (1987). Fibronectin-mediatedadhesion of fibroblasts: Inhibition by dermatan sulfate proteoglycan and evidence for a crypticglycosaminoglycan-binding domain. J. Cell Biol. 105, 1443-1454.Lillie, R. D. (1951). Simplification of the manufacture of Schiff reagent for use in histochemical procedures.Stain Technol. 26, 163-165.Lin, C. Q. and Bissel, M. J. (1993). Multi-faceted regulation of cell differentiation by extracellular matrix.FASEB J. 7, 737-743.Linker, A. and Hovingh, P. (1972). Heparinase and heparitinase from flavobacteria. Methods Enzymol.28, 902-911.164MaIler, J. L. (1985). Qocyte maturation in amphibians. “Development Biology. A ComprehensiveSynthesis”. (L. W. Browder, Ed), Vol.1. pp. 289-311 . Plenum Press, New York.MalIya, S. K., Partin, J. S., Valdizan, M. C. and Lennarz, W. J. (1992). Proteolysis of the major yolkglycoproteins is regulated by acidification of the yolk platelets in sea urchin embryos. J. Cell 8101. 117,1211-1221.Mancuso, V. (1964). Ultrastructural changes in the cytoplasm of Ciona intestinalls oocytes. Acta. Embryo!.Morphoh Exp. 7, 269-295.Martinez, T. and Wheeler, D. (1991). Identification of vitellogenin in the ant, Camponotus festinatus:Changes in hemolymph proteins and fat body development in workers. Arch. Insect Biochem. Phys. 17,143-1 55.Matranga, V., Kuwasakik B. and Noll, H. (1986). Functional characterization of toposomes from sea urchinblastula embryos by a morphogenetic cell aggregation assay. EMBO J. 5, 3125-3132.McManus, J. R. A. (1946). Histological demonstration of mucin after periodic acid. Nature (London) 158,202.Medina, M. P., Leon, P., Vallejo, C. G. (1988). Drosophila cathepsin B-like proteinase: a suggested rolein yolk degradation. Arch. Biochem. Biophys. 263, 355-363.Meyer, K. (1970). Hyaluronidases. “The Enzymes”. (P. D. Boyer, Ed), Vol.5. pp. 307-320. AcademicPress, New York/London.Misevic, G. N. and Burger, M. M. (1990). Involvement of a highly polyvalent glycan in the cell-binding of theaggregation factor from the marine sponge Microciona prolifera. J. Cell. Biochem. 43, 307-314.Mitani, 5. (1989). Retarded gastrulation and altered subsequent development of neural tissues inheparin-injected Xenopus embryos. Development 107, 423-435.Morris-Kay, G. M. and Tuckett, F. (1989). Immunohistochemical localization of chondroitin sulfateproteoglycans and the effects of chondroitinase ABC in 9-11 day rat embryos. Development 106, 787-798.165Motohiro, T. (1960). Studies on mucoprotein in marine products-Ill. Crystalization of itin sulfatefrom the mucoprotein in Cucumaria japonica. Nippon Suisan Gakkaishi 26, 1171-1174.Neoh, S. H., Gordon, C., Potter, A., Zola, H. (1986). The purification of mouse monoclonal antibodies fromascitic fluid. J. Immunol. Methods 9, 231-235.Nicholson, L. J. and Watt, F. M. (1991). Decreased expression of fibronectin and the cz5Bl integrin duringterminal differentiation of human keratinocytes. J. Cell Sd. 98, 225-232.Nilsson, B., De Luca, S., Lohmander, L. S. and Hascall, V. C. (1982). Structures of N-linked and 0-linkedoligosaccharides on proteoglycan monomer isolated from the Swarm rat chondrosarcoma. J. BioL Chem.257, 10920-10927.Nilsson, B., Nakazawa, K., Hassell, J. R., Newsome, D. A., Hascall, V. C. (1983). Structure ofoligosaccharides and the linkage region between keratan sulfate and the core protein on proteoglycansfrom monkey cornea. J. Biol. Chem. 258, 6056-6063.Nishihara, R., Wyrick, R. E., Working, P. D., Chen, Y.-H. and Hedrick, J. L. (1986). Isolation andcharacterization of a lectin from the cortical granules of Xenopus laevis eggs. Biochemistry 25, 6013-6020.Noll, H., Matranga, V., Cervello, M., Humphreys, T., Kuwasaki, B. and Adelson, D. (1985). Characterizationof toposomes from sea urchin blastula cells: A cell organelle mediating cell adhesion and expressingpositional information. Proc. NatL Acad. Sci. USA 82, 8062-8066.O’Shea, K. S., Rheinheimer, J. S. T., Dixit, V. W. (1990). Deposition and role of thrombospondin in thehistogenesis of the cerebellar cortex. J. Cell Biol. 110, 1275-1284.Okada, Y. and Yokota, Y. (1990). Purification and properties of cathepsin B from sea urchin eggs. Comp.Biochem. Physiol. 96B, 381-386.Okayama, M., Kimata, K., Suzuki, S. (1973). The influence of p-nitrophenyl B-D xyloside on the synthesison the synthesis of proteochondroitin sulfate by slices of embryonic chick cartilage. J. Biochem. 74,1069-1073.166Outenreath, R. L., Roberson, M. M., Barondes, S. H. (1988). Endogenous lectin secretion into theextracellular matrix of early embryos of Xenopus laevis. Dev. BioL 125, 187-194.Ozaki, H. (1980). Yolk Proteins of the sand dollar Dendraster excentricus. Dev. Growth Differ. 22, 365-372.Pan, M. L., Bell, W. J., Telfer, W. H. (1969). Vitellogenic blood protein synthesis by insect fat body.Science 165, 393-394.Pascual, D. W. and Clem, L. W. (1992). Low temperature pepsin proteolysis. An effective procedure formouse 1gM F(ab’)2 fragment production. J. Immunoh Methods 146, 249-255.Passaniti, A. (1992). Extracellular matrix-cell interactions: Matrigel and complex cellular pattern formation(Correspondence). Lab. Invest. 67, 804-808.Penn, J.-P., Bonnet, F., Mailet, P., JoIIès (1989). Characterization and N-terminal sequence of humanplatelet proteoglycan. Biochem. J. 255, 1007-1013.Perris, R., Krotoski, D., Lallier, T., Domingo, C., Sorrell, J. M., Bronner-Fraser, M. (1991). Spatial andtemporal changes in the distribution of proteoglycans during avian neural crest development.Development 111, 583-599.Perris, R. and Johansson, S. (1990). Inhibition of neural crest cell migration by aggregating chondroitinsulfate proteoglycans is mediated by their hyaluronin-binding region. Dev. BioL 137, 1-12.Pierce, B. M. (1966). The development of basement membranes of the mouse embryo. Dev. Biol. 13,231 -249.Platt, J. L., Brown, D. M., Granlund, K., Oegema, T. R., Klein, D. J. (1987). Proteoglycan metabolismassociated with mouse metanephric development: morphologic and biochemical effects of B-D-xyloside.Dev. Sb!. 123, 293-306.Postlethwait, J. H. and Giorgi, F. (1985). Vitellogenesis in Insects. “Developmental Biology. AComprehensive Synthesis”. (L. W. Browder, Ed), Vol.1. pp. 85-126. Plenum Press, New York.167Reimer, C. L., Crawford, B. J., Crawford, T. J. (1992). Basement membrane lectin binding sites aredecreased in the esophageal endoderm during the arrival of presumptive muscle mesenchyme in thedeveloping asteroid Pisaster ochraceus. J. Morphol. 212, 291 -303.Reimer, C. L. and Crawford, B. J. (1990). Lectin histochemistry of the hyaline layer in the asteroid, Pisasterochraceus. J. Morphol. 203, 361-375.Reynolds, E. S. (1963). The use of lead citrate at high pH as an electron opaque stain in electronmicroscopy. J. Cell BioL 17, 208-212.Roden, L. (1980). Structure and metabolism of connective tissue proteoglycans. “The Biochemistry ofGlycoproteins and Proteöglycans”. (W. J. Lennarz, Ed), pp. 267-372. Plenum Press, New York.Ruoslahti, E. (1989). Proteoglycans in cell regulation. J. Biol. Chem. 264, 13369-13372.Ruoslahti, E. and Yamaguchi, Y. (1991). Proteoglycans as modulators of growth factor activities. Cell 64,867-869.Sanders, E. J., Zalik, S. E., Schneider, W. J., Ledsham, I. M. (1990). The endogenous lectins of the chickblastoderm are present in association with an apolipoprotein in distinct organelles and in the extracellularmatrix. Roux’s Arch. Dev. BioL 199, 295-306.Schechtman, A. M. (1956). Uptake and transfer of macromolecules by cells with special reference togrowth and development. mt Rev. Cytoh 5, 303-322.Schuel, H., Kelly, J. W., Berger, E. R., Wilson, W. L. (1974). Sulfated acid mucopolysaccharides in thecortical granules of eggs. Effects of quaternary ammonium salts on fertilization. Exp. Cell Res. 88, 24.Schuel, H. (1978). Secretory functions of egg cortical granules in fertilization and development. GameteRes. 1, 299-382.Scott, J. E. (1972). Histochemistry of Alcian blue Ill. The molecular biological basis of staining by Alcianblue 8GX and analogous phthalocyanins. Histochemie 32, 191-212.Scott, J. E. and Harbinson, R. J. (1968). Periodate oxidation of acid polysaccharides. Inhibition by theelectrostatic field of the substrate. Histochemie 14, 215-220.168Scott, L. B., Leahy, P. S., Decker, G. L., Lennarz, W. J. (1990). Loss of yolk platelets and yolkglycoproteins during larval development of the sea urchin embryo. Dev. BioL 137, 368-377.Scott, L. B. and Lennarz, W. J. (1989). Structure of a major yolk glycoprotein and its processing pathwayby limited proteolysis are conserved in echinoids. Dev. B1oL 132, 91-102.Sharrock, W. J. (1984). Cleavage of two yolk proteins from a precursor in Caenorhabditis elegans. J.Molec. Biol. 174, 419-431.Sheppard, A. M., Hamilton, S. K., Pearlman, A. L. (1991). Changes in the distribution of extracellular matrixcomponents accompany early morphogenetic events of mammalian cortical development. J. Neurosci.11, 3928-3942.Shyu, A.-B., Raff, R. A. and Blumenthal, T. (1986). Expression of the vitellogenin gene in female and malesea urchin. Proc. NatL Acad. Sd. USA 83, 3865-3869.Slot, J. M. and Geuze, H. J. (1985). A new method of preparing gold probes for multiple labellingcytochemistry. Eur. J. CelL Biol. 38, 87-93.So, L.-L. and Goldstein, I. J. (1969). Protein-carbohydrate interactions. Part XXI. The interaction ofConcanavalin A with D-fructans. Carbohydr. Res. 10, 231-244.Solursh, M., Mitchell, S. L. and Katow, H. (1986). Inhibition of cell migration in sea urchin embryos by B-Dxyloside. Dev. B1oL 118, 325-332.Sousa, M. and Azevedo, C. (1989). Starfish sperm-oocyte jelly binding triggers functional changes incortical granules. A study using acid phosphatase and ruthenium red ultrastructural histochemistry.Histochemistry 90, 353-357.Stoker, A. W., Streuli, C.H., Martins-Green, M., Bissell, M. J. (1990). Designer microenvironment for theanalysis of cell and tissue function. Cuff. Opin. Cell Biol. 2, 864-874.Strathmann, M. F. (1987). Reproduction and Development of Marine Invertebrates of the Northern PacificCoast. University of Washington Press, Seattle.169Strathmann, R. R. (1989). Existence and functions of a gel filled primary body cavity in development ofEchinoderms and Hemichordates. Biol. BulL 176, 25-31.Streuli, C. H. and Bissel, M. J. (1990). Expression of extracellular matrix components is regulated bysubstratum. J. Cell BioL 110, 1405-1415.Suzuki, S. (1972). Chondroitinases from Proteus vulgaris and Flavobacterium heparinum. MethodsEnzymol. 28, 911-917.Tan, 5.-S., Crossin, K. L., Hoffman, S. and Edelman, G. M. (1987). Asymmetric expression in somites ofcytotactin and its proteoglycan ligand is correlated with neural crest cell distribution. Proc. NatL Acad. Sd.USA 84, 7977-7981.Taylor, C. E. D. and Heimer, G. V. (1974). Measuring immunofluorescence emission in terms of standardinternational units. J. BioL Stands. 2, 1-8.Thiery, J. P., Rovasio, R. A., Duband, J. L. and Delouvee, A. (1983). The extracellular matrix in earlyembryogenesis. “Hormones and Cell Regulation, Volume 7”. (J. E. Dumont Nunez, J. and R.M. Denton,Ed), pp. 319-334. Elsevier Biomedical Press.Thyberg, J., Lohmander, L. S., Friberg, U. (1973). Electron microscopic demonstration of proteoglycansin guinea pig epiphyseal cartilage. J. Ultrastruct. Res. 45, 407-427.Tirumalai, R. and Subramoniam, T. (1992). Purification and characterization of vitellogenin and lipovitellinsof the sand crab Emerita asiatica: Molecular aspects of crab yolk proteins. M0L Reprod. Deveh 33, 16-26.Venkatasubramanian, K. and Solursh, M. (1984). Adhesive and migratory behavior of normal and sulfate-deficient sea urchin cells in vitro. Exp. Cell Res. 154, 421-431.Vieira, R. P. and Mourao, P. A. (1988). Occurrence of a unique fucose-branched chondroitin sulfate in thebody wall of a sea cucumber. J. BioL Chem. 263, 181 76-18183.Wahli, W., Dawid, I. B., Ryffel, G. V., Weber, R. (1981). Vitellogenesis and the vitellogenin gene family.Science 212, 298-304.Wahli, W. (1988). Evolution and expression of vitellogenin genes. Trends Genetics 4, 227-232.170Wahli, W. and Ryffel, G. U. (1985). XenopusVitellogenin Genes. “Oxford Surveys on Eukaryotic Genes”.(N. McLean, Ed), Vol.2. PP. 96-120 . Oxford University Press.Wall, R. S. and Gyi, T. J. (1988). Alcian blue staining of proteoglycans in polyacrylamide gels using the“critical electrolyte concentration” approach. Anal. Biochem. 175, 298-299.Wallace, R. A. and Morgan, J. P. (1986). Chromatographic resolution of chicken phosvitin. Multiplemacromolecular species in a classic vitellogenin-derived phosphoprotein. Biochem. J. 240, 871-878.Wang, S., Smith, D. E., and Williams, D.L. (1983). Purification of avian vitellogenin Ill: comparison withvitellogenin I and II. Biochemistry 22, 6206-621 2.Wang, S. and Williams, D. L. (1982). Biosynthesis of the vitellogenins. Identification and characterizationof nonphosphorylated precursors to avian vitellogenin I and vitellogenin II. J. Biol. Chem. 257, 3837-3846.Watt, R. M. (1984). Selective migration of terminally differentiating cells from the basal layer of culturedhuman epidermis. J. Cell Biol. 98, 16-21.Wessel, G. M. and McClay, D. R. (1987). Gastrulation in the sea urchin embryo requires deposition of crosslinked collagen within the extracellular matrix. Dev. Biol. 121, 149-1 65.Wight, T. N., Dick, K., Heinegárd, Hascall, V. C. (1991). Proteoglycans. Structure and Function. “CellBiology of Extracellular Matrix”. (E. D. Hay, Ed), pp. 45-78. Plenum, New York.Wiley, H. S. and Wallace, R. A. (1981). The structure of vitellogenin. Multiple vitellogenins in Xenopuslaevis give rise to multiple forms of the yolk proteins. J. Biol. Chem. 256, 8626-8634.Wiley, H. S. and Wallace, R. A. (1981). The structure of vitellogenin. Multiple vitellogenins in Xenopuslaevis give rise to multiple forms of the yolk proteins. J. B1oL Chem. 256, 8626-8634.Williams, J. (1967). Yolk Utilization. “The Biochemistry of Animal Development”. (R. Weber, Ed), Vol.11. pp.343-378 . Academic Press, New York/London.171Winter, C. E. (1992). The yolk polypeptides of a free-living rhabditid nematode. Comp. B!ochem. PhysioL103B,189-196.Woodward, M. P., Young, W. W., Bloodgood, R. A. (1985). Detection of monoclonal antibodies specificfor carbohydrate epitopes using periodate oxidation. J. Immunol. Methods 78, 143-153.Yamagata, T., Saito, H., Habuchi, 0. and Suzuki, S. (1968). Purification and properties of bacterialchondroitinases and chondroitin sulfatases. J. Biol. Chem. 243, 1523-1535.Yamaguchi, Y., Mann, D., Ruoslahti, E. (1990). Negative regulation of transforming growth factor-B by theproteoglycan decorin. Nature London) 346, 281-284.Yokota, Y. and Kato, K. H. (1988). Degradation of yolk proteins in sea urchin eggs and embryos. CellDiffer. 23, 191-200.Yost, H. J. (1992). Regulation of vertebrate left-right asymmetries by extracellular matrix. Nature 357, 158-161.172APPENDICESAppendix 1: Freeze substitution of starfish oocytes and embryos(After Campbell eta!., 1991)1. Harvest oocytes and embryos on ice and transfer to 15 ml conical glass centrifuge tubes.2. Resuspend embryos but not oocytes in 10 ml of cryoprotectant solution, consisting of 15% 2,3-butanediol prepared in sea water, and allow to equilibrate for 30 minutes.3. Prepare freezing apparatus in a fume hood, by filling a Dewar flask with liquid nitrogen, and placingan insert cup into the Dewar flask. Fill the cup with cooking grade propane (Mastercraft), andadjust an eddy current motor with an attached stirring rod so that the rod goes at least half way intothe insert cup. LIQUID PROPANE IS VERY EXPLOSIVE-USE EXTREME CAUTION, ANDENSURE NO SPARKS COME NEAR THE CRYOGENIC APPARATUS.4. After the embryos have equilibrated, remove as much solution as possible, and place 1 ill of thethick suspension on a 50 mesh EM grid. Use filter paper to draw off as much liquid as possible,and to ensure an even monolayer is present. Using a pair of cross-closing forceps, quickly plungethe grid with embryos deep into the propane for 10-15 seconds and then into the surroundingliquid nitrogen for 10 seconds. The grid can now be transferred into the pre-cooled substitutingmedia (see below), or alternatively, the grids can be placed into freezer vials containing liquidnitrogen and stored in a liquid nitrogen freezer for long term storage. Eggs can be frozen in thesame manner, but they should not be incubated in cryoprotectant prior to freezing.5. To substitute the eggs or embryos the ethanol must first be pre-cooled to 90 C. To do this,place approximately 150 g of dry ice into a wide mouth thermos, and slowly add acetone until thethermos is half full. Then pour in liquid nitrogen and break up the acetone as it freezes until thebath reaches between -90 and -95° C.6. Place 10 ml absolute ethanol or alcian blue (Marivac)-saturated ethanol into glass vials and precool in the dry ice-acetone bath for at least 10 minutes.7. Quickly transfer the grids with embryos into the pre-cooled substituting medium, place the vialsquickly into the thermos, and store in a -70° freezer. Let the embryos substitute for 5 days, andmaintain the temperature below -80° C with daily additions of liquid nitrogen.1738. After the substitution is complete, gradually bring the vials up to room temperature, and wash withabsolute ethanol twice. The embryos are now ready to be embedded into JB4 or LR White resins.9. Infiltrate the embryos with several changes of uncatalyzed resin for at least 24 hours. For JB4,polymerize the tissue in 9 mm aluminum dishes at room temperature, ensuring that at least 5 ml ofsolution is used per dish, and that a second dish has been placed over top in order to exclude airfrom the surface. Alternatively, polymerize the LR Whe in beam capsules at 550 c.174Appendix 2: Con A-Sepharose affinity chromatographyColumn Volume: 3 mlWashing buffer: 20 mM Tris-HCI, 0.5M NaCI, 0.1% Brij 56, pH 7.4Eluent: 0.5 M methyl a-D-mannopyranoside in washing bufferTemperature: 20°CElution Rate: 5 ml /hourProtease inhibitors included in all buffers: 1 mM PMSF, 1 jig/mI pepstatin A, 10 mM EDTA1. Place 3 ml of Con A-Sepharose in column, made from the empty 5 mm syringe barrel.2. Wash column with 10 column volumes (30 ml) TBS (20 mM Tris, 0.15 M NaCI, pH 7.4) to removethe thimerosal, which is used as a preservative.2. Wash with 1 column volume of 1 M NaCI in 20 mM M Tris-HCI and 0.1% Brij 56, pH 7.4, to reducethe free release of Con A into solution.3. Wash with 1 column volume 0.1 M methyl-c-D-mannopyranoside in 20 mM Tris-HCI, 0.5 M NaCI,0.1% brij 56.4. Wash with 5 column volumes of washing buffer.5. Apply 1.0 ml detergent solubilized sample (embryo extract) to the lectin column, and recyclethrough column 2 times. Wash with 10 column volumes (30 ml) of washing buffer, or until a flat lineis achieved with the UV recorder (a280), indicating that all unbound protein has been eluted fromthe column.6. Apply 3 ml 0.2 M methyl-D-a-mannopyranoside eluting buffer to the column, and stop the columnflow for 60 minutes after all 3 ml has entered the resin. This allows dissociation to take place priorto elution, and ensures a good recovery.7, Resume column flow with washing buffer, and collect the peak in 1 ml fractions.8. Concentrate the protein containing fractions in dialysis tubing placed over beads of PEG (MW20,000) and store at -70°C.175Regeneration of column:1. Wash with 10 column volumes of 0.1 M Tris-HCI, 0.5 M NaCI, pH 8.5.2. Wash with 10 column volumes 0.1 M sodium acetate buffer, 0.5 M NaCI, pH 4.5 (with 1 mM calciumchloride, 1 mM magnesium chloride and 1 mM manganese chloride to preserve the bindingactivity of Con A).3. Re-equilibrate with TBS, or store in above buffer, pH 6.0 with 0.01% thimerosol.176Appendix 3: Protein quantificationUV Detection at cx 280:This is a quick method of determinating the concentration of a known pji protein solution, for which theextinction coefficient is known. This method has the advantage that none of the sample is destroyedduring the reading. The absorbance maximum at 280 nm is due primarily to the presence of tyrosine andtryptophan residues.1%Protein E280 (i.e. the absorbance of a 10 mg/mI solution at 280 nm)13.61gM 11.81. Read the absorbance versus a buffer control at 280 nm.2. Calculate the approximate concentration using the following equation:Concentration of sample = absorbance at 280 nm x 10 mg/mIextinction coefficient at 280 nmFor protein solutions that are contaminated with nucleic acids, an approximate concentration can bedetermined using the following equation:Protein concentration (mg/mI) = (1.55 xA280) - (0.76 xA260).Biorad DC protein assay:This assay is based on the reaction of proteins with an alkaline copper tartrate solution and Folin reagent.i.e. the reduction of Folin reagent by copper-treated proteins. Color development occurs primarily bytryptophan and tyrosine residues.The protocol was followed according to manufacturers directions:1. Serial dilutions of the standard (BSA) was prepared, ranging from 1.5 mg/mI to 0.1 mg/mI.2. One hundred uI of the standard solution (or sample solution) were placed in 10 ml test tubes.1773. Five hundred p.1 of reagent A (an alkaline copper tartrate solution) were added to each test tube,and the mixture was vortexed briefly.4. Four ml reagent B (a dilute Folin regent) were added to each test tube, and the mixture wasvortexed briefly.5. Solutions were allowed to stand 15 minutes while the color reaction developed, and absorbancewas read at 750 nm within the hour. The standard absorbances were plotted, and sampleabsorbances were determined based on comparison with the standard curve.178Appendix 4: immunization of mice with Con A-specific embryo fraction1. Prepare immunogen in Freund’s incomplete adjuvant, using equai voiumes of protein soiutionand adjuvant. immunize 4 week-oid BALB/c mice by subcutaneus injections of 100 jiiimmunogen/mouse. Each mouse shouid receive no more than 200 protein peradministration.2. Boost the mice at 1 month intervais with intraperitonealiy injections of 100 ti pure antigen (noFreund’s adjuvant).3. Five days after each boost, take a test bieed from the tail as foliows: use a sharp steriie razor biadeto remove a small part of the tail end(2 mm), and collect 4 drops of blood by massaging the tailproximal to distal. Suspend the blood in 500 .d 2% biotto, and spin in a clinical centrifuge toremove the red blood cells. Use serum at dilutions of 1:10 and 1:100 to test for antibody titer onJB4-embedded embryo sections using indirect immunofiuorescence microscopy.4. Continue to monitor antigen response with test bleeds, always taken 5 days after the boost.When the immune response is satisfactory, harvest the spleen for hybridoma production 3-5 daysfollowing the boost.179Appendix 5: Hybridoma production for monocional antibodies:(After Kannangara et aL,1989)Growth of Myeloma cell line1. 5-7 days before the fusion, remove 1 vial of myeloma cells from the liquid nitrogen freezer. Thawat 45° C, and wash with 10 ml Dulbecco’s Modified Eagle Medium (DMEM) with no fetal calf serum(FCS).2. Centrifuge at 800 x g for 10 minutes.3. Seed cells into 20% Fetal calf serum (FCS) in DMEM in sterile petri or T.C. dish.4. Subculture into 10% FCS in DMEM when cells cover 50-70% of the surface area of dish.Fusion Protocol1. Harvest myeloma cells: 5-6 petn dishes (50-70% covered) are required per spleen.Place all of the cells in one 50 ml centrifuge tube and centrifuge 800 x g for 10 minutes. Decantoff liquid.2. Wash cells with 10 ml DMEM twice with gentle resuspension and centrifugation as above.3. Prepare tusogen:a) Add 5 g polyethylene glycol (PEG) 4000 MW and 4 ml DMEM to a beaker.b) Heat gently with stirring at 45° C until dissolved.C) Cool to room temperature.d) Add 1 ml dimethyl sulfoxide (DMSO).e) Filter through a 0.45 im Millipore filter into a sterile tube.f) Place in a 37° C water bath until required.4. Harvest spleen from mouse as follows: sacrifice the mouse in a chamber filled with C02, spray themouse with 70% ethanol and placed on a paper towel. Using sterile techniques, make a smallincision in the abdominal skin and pull the skin in an anterior and posterior direction away from the180abdomen, ensuring that the skin is pulled well over the head of the mouse to maintain a sterileenvironment in the abdominal cavity. Dissect out the spleen from the mesentary and placed in asterile petri dish containing 10 ml DMEM. Cover spleen with a 1 inch square of sterile gauze, andsqueeze through the gauze with the sterile end of a syringe plunger. Centrifuge as in step 1.5. Mix spleen and myeloma cells together in a round bottom 12 ml tube and centrifuge as above.6. Remove all liquid from the pellet of mixed spleen and myeloma cells.7. Precise timing and volumes are required for the following steps:a) Add 1 ml of fusogen to cells over a time period of 1 minute while stirring the pellet constantlywith the pipette tip.b) Add 1 ml DMEM over a time period of 1 minute to the cells with stirring, to slowly dilute out thePEG solution.c) Add 2 additional ml DMEM over the next minute with stirring.d) Add 6 additional ml DMEM over the next 3-5 minutes with stirring.8. Spin down cells and resuspend in 5 ml (HAT) media (media supplemented with hypoxanthine,aminopterin and thymidine), and with 20% FCS included.9. To isolate a feeder cell population, sacrifice 2, 4-6 week BALB/c mice and remove their thymuses.Follow the same sterile techniques as for spleen removal, but in this case, after the removal ofskin, use a pair of sharp scissors to open the rib cage and access the area deep to the manubriumwhere the thymus lies. Be careful not to rupture the heart or major vessels. Mince the thymusesthrough gauze using the same technique that was used on the spleen, wash in DMEM, andresuspend in 10 ml HAT media.10. Mix thymocytes with the fused cell mixture, and dilute to 50-100 ml with HAT. Plate out at 100 jilper well into sterile 96 well plates, and incubate at 37° C with 5% C02.11. After 3-4 days, add 100 jil HT media (HAT without aminopterin).12. After 7 days, remove 180 jil media and add 180 fresh HT media.13. After 10 days, test clones for antibody production by removing 30 p.1 of hybridoma supernatantand performing immunofluorescence microscopy on thin sections of JB4-embedded embryos.181Appendix 6: Recloning hybridomasOnce the selected hybridomas have grown to occupy a large part of the well, recloning must be carried outin order to obtain wells containing only one clone; this should be done immediately, as prolonged growthcan result in selection against the particular clone of interest.1. Sacrifice 2 BALB/c mice and remove thymuses as described in appendix 4. Suspend the cells in50 ml DMEM supplemented with 20% FCS.2. Add 100 tl of thymocyte suspension to each well of a 96 well plate; add 200 jii to the first well ineach row (across), and 0 jil to the last well in each row.3. Add 15 of cells from the original well of clones to the first well in each row (A-H). Use 4 rows/clone to ensure a positive single clone.4. Make serial dilutions of each row by removing 100 pifrom well #1 and adding to #2 and so on untilwell #12 now has 100 pi. When diluting, mix each well vigorously by pipetting up and downseveral times before removing the 100 pi.5. Incubate plates at 370 C, and visually inspect each well daily for single colony formation. After 8-10days, retest the supernatant from wells containing single clones.182Appendix 7: Expansion and freezing of positive clones1. Once positive clones have been selected and have been recloned to ensure that all hybridomasare derived from a single clone, the cells can be expanded for further use. This is done bygradually increasing the surface area available to the growing cells, starting with 24 well tissueculture wells, and then sterile 60 mm, 90 mm, and finally 150 mm petri or tissue culture dishes. Inevery expansion, cells should be left undisturbed until the bottom of the well or plate is crowded.The cells should have smooth puffy plasma membranes, which indicates that they are still in logphase growth, as opposed to wrinkled, collapsed and irregular ones, which indicates they aredead or dying.2. Harvest 2, 90 mm plates of cells in log phase growth; centrifuge at 800 x g for 10 minutes andresuspend the cells in 5 ml of freezing media (1 ml FCS, 0.5 ml DMSO and 3.5 ml DMEM).3. Aliquot exactly 1 ml into 1.5 ml sterile freezer vials, and place the vials into a foam-insulatedcardboard box, and then quickly into a -70° C freezer overnight to allow the cells to freeze slowly.4. Remove vials from freezer and store them in liquid nitrogen indefinitely.5. To resurrect the hybridomas, remove 1 vial of cells from the liquid nitrogen freezer, and gentlythaw them in a 37° C water bath. Wash with 10 ml DMEM, and resuspend in 10 ml DMEMsupplemented with 20% FCS. Plate all the cells in one 90 mm dish, and leave undisturbed untilexpansion is required.183Appendix 8: isotyping monoclonals with the Serotec Isotyplng kit.This diagnostic test is based on red cell agglutination. Sheep red blood cells are coupled to antibodiesdirected to the 6 different mouse monoclonal isotypes. The blood cells are then incubated withhybridoma supernatant in microtiter plates, and positive agglutination is easily visible on the bottom of theplates, indicating the isotype of the supernatant.Protocol:1. Make 1/10 and 1/50 dilutions of hybridoma supernatants in PBS; ensure that highly clonedsupernatants are used.2. Pipette 30 l of diluted supernatant into each of eight wells across a “U” bottomed microtiterplate.3. Add 30 jil of each isotyping reagent (antibody-coupled sheep RBCs) to the wells containing thesupernatant.4. Cover the plate and place on a shaker for 6 seconds.5. Leave plate on a cool flat surface with no vibration for 1 hour.6. Read results as follows: partial or full carpet of cells over bottom of well indicates that agglutinationhas occurred, and is therefore a positive result; a small red circle of cells at the bottom of the wellindicate that no agglutination has occurred, and therefore is a negative result.Appendix9:PhotographyMicroscopyExposuretime(Sec.)FilmDeveloper/conditionsLightandDICmicroscopy:-color-BANAutomaticAutomaticFugicolor100KodakT-MAX100Automatic,-1densityKodakD-76,7mm. @22°CFluorescencelightmicroscopy:-color-8/W3-7seconds5-10secondsFugicolor1600KodakTMAXp3200Automatic,+3densityKodakTMAX,lomin.@22°CGelsandX-rays(Reprovitwithlightbox)-color-BANF8,1/15,1/30,1/60Fil,1/8,1/15,1/30Fugicolor100with80AfilterKodakhighcontrastcopyAutomatic,-2densityKodakD-19,6min.@22°CBlots(Reprovitwithsideillumination)-color-B/WTEM(BAN)F8,1/2,1Fli,1/2,1/4,1/82secondsFugicolor100with80AfilterKodakhighcontrastcopyKodakfinegrain(5302)Automatic,-2densityKodakD-19,6mm. @22°CKodakD-19,5mm. @22°CReprints(Reprovit,sideillumination):-BAN-colorF8,1/15,1/30F8,1/2,1,smallprintsF8,1,2,largeprintsF8,1/4,1/2,immunofluorescenceKodakT-MAX100Fugicolor100with80AfilterKodakD-76,7mm. @22°CAutomaticAutomaticAutomatic,+2densityPaper:-ToprintB/WprintsfromB/Wnegatives:IlfordmultigradeIllRC,90s,KodakDektol1:2-ToprintBANprintsfromcolornegatives:KodakpanalureselectRC(FH),KodakDektol1:2185Appendix 10: Preparation of colloidal gold(After Slot and Geuze, 1985)To make 100 ml gold sol, prepare the following solutions:Solution I: chloroaurlc acid1%HAuCI4 1rrdistil led water 79 mlSolution II: citric acid (reducIng soin)1% tn-sodium citrate (aqueous) 4 ml1% tannic acid (aqueous) 0-0.5 mldistilled water up to 20 ml1% Tannlc Acid 1% sodium citrate Particle size (nm)0.5ml 4m1 Gnm0.125m1 4mi 10-l5nmI 4 ml 20-25 nm/ 2 ml 30-35 nm/ 1 ml 60-70 nmProcedure:1. Heat solutions I and II separately in a 60° C water bath.2. Add reducing solution II to solution I quickly with stirring; ensure temperature remains constant at60° C to avoid the formation of a heterodisperse sol.3. Keep the solution stirring until color change to red or purple is complete; the reaction time willincrease proportionally to particle diameter size.4. Storeat4°CTo size gold particles:1. Place 1 jii gold sol on a coated 100 mesh grid, and air dry.2. Photograph gold particles with a magnification of at least x 16,000.3. Measure the diameter size on the negative, and use the following formula:Negative measurement of diameter (mm x 106 = diameter (nm)magnification186Appendix 11: Microtltratlori assay for determination of the correct proteinconcentration for gold 501 stabilization.Protocol:1. Add 100 jil distilled water to each of 10 1.5 ml Eppendorf tubes.2. Prepare a 1 mg/mI solution of protein to be assayed in distilled H20. Dialyze if necessary against 2mM sodium borate, pH adjusted just basic to the p1 of the protein (9.0 for lgG).3. To the first tube, add 100 jil of the protein solution. Serially dilute by removing 100 il from firsttube, adding it to the 2nd tube and pipettirig up and down 3 times; then continue subsequentdilutions to the 9th tube, and leave 10th tube protein free.4. Adjust the pH of 5.0 ml gold sol to just basic of the p1 of the protein being assayed using 0.2MK2C03.5. Add 500 jil gold sol to each well and pipette stir. Stand 15 minutes.6. Assess the resistance of mixture to saft-induced flocculation by adding 100 jil of 10% NaCI. Stand5 minutes.7. The last well to maintain a red color represents the end point for protein stabilized gold.8. Calculate protein stabilizing concentration per ml gold sol and double for experimental proteinstabilizing concentration187Appendix 12: Gold conjugation to rabbit anti-mouse IgG!M1. Dilute 100 l of stock protein solution (2 mg/mI) up to 500 j.ti in 2 mM sodium borate-HCI, pH 9.0,and dialyze against borate buffer overnight.2. Adjust the pH of 10 ml gold sol to 9.0 with 0.2 M K2C03.3. Add 0.5 ml dialyzed protein all at once to stirring gold sol, and stir for 30 minutes.4. Add 0.5 mi 10% BSA to stirring gold soiution and continue to stir a further 5 minutes.5. Place in 1.5 ml Eppendort tubes and centrifuge 15,000 x g for 45 minutes.6. Resuspend in 10 mM Tris with 50 mM NaCI, 0.1% BSA (pH 8.2), and gradually increase the saitconcentration to 150 mM NaCI by diaiysis over 24 hours.7. Add sodium azide to a final concentration of 0.02%, and store at 4° C for 1 year.188Appendix 13: Gradient gels(Use distilled water and electrophoresis grade reagents throughout)Reagents:Lower Tris (1.5 M) Dissolve 36.34 g Tris base and 0.8 g SDS in 150 ml H20. Titrate pH to8.8 with 6N HCI. Add H20 to final volume of 200 ml, and Millipore filter(0.45j.t).Upper Tris (0.5M) Dissolve 12.11 g Tris base and 0.8 g SDS in 150 ml H20. Titrate pH to6.8 with 6N HCI. Add H20 to final volume of 200 ml, and Millipore filter(0.45ji).30% Acrylamide Dissolve 7.5 g acrylamide and 0.2 g bis-acrylamide in H20 to a(2.5% cross-linker) final volume of 25 ml, and Millipore filter (0.45ji).Electrophoresis Buffer: Dissolve 3.03 g Tris base, 14.41 g glycine and 1 g SDS in 1 L H20.10% Ammonium. Persulfate Dissolve 100 mg in 1 ml H20. Store the solution at 40 C and discard afterone day.To Pour 7 MinI Slab Gradient Gels:Stock Solutions 3 % 1 2%30% acrylamide 2.0 ml 8.0 mlLower Tris (1.5 M) 5.0 ml 5.0 mlH20 13.0 ml 7.0 mlDegas for 5 minutes.TEMED lOj.tI lOp110% ammonium persulfate 75 uI 75 p1Protocol:1. Prepare acrylamide solutions and degas for 10 minutes.2. To set up gradient gel apparatus, place the gradient maker on magnetic stirrer, place a small stir barin the column with outflow tubing attached, and ensure that connecting port is closed off.3. Add TEMED and APS to acrylamide solutions, swirl gently, and pour into gradient maker, ensuringthat the 3% acrylamide solution is poured into the side with outflow tubing.4. Attach outflow tubing to multi-gel apparatus, open connecting port on gradient maker, turn on themagnetic stirrer, and lastly open stopcock on outflow tubing. Allow the acrylamide to pour inslowly, so that the all the solution has entered within 2-3 minutes.1895. Disconnect outflow tubing, and rinse out gradient apparatus immediately. Then place 200 jilwater-saturated isobutanol on each gel to ensure a flat even surface during polymerization.6. Rinse the isobutanol off the gels after 1 hour, and continue to polymerise for a further hour; storegels in plastic bags at 40 C.7. Prior to PAGE run, allow gels to equilibrate to room temperature and pour a stacking gel over top.Stacking gels: (makes 2)Stock Solutions 2.5%30% acrylamide 0.5 mlUpper Iris (0.5 M) 1.5 mlH20 4.0 mlTEMED 6il10% ammonium persulfate 45jilStacking gels are poured according to manufacturers instructions (BioRad) by first placing a comb (eitherfor multilane or preparative runs) at a 450 angIe on top of the running gel, and then slowly pouring theaciylamide in to avoid bubble formation.190AppendIx 14: PM1-immunoaffinity column preparation and chromatographyCoupling Protocol:Affi-Gel 10 (BioRad) is a N-hydroxysuccinimide ester of a derivatized cross-linked agarose gel beadsupport, which allows spontaneous coupling of ligands in a quick 1-step procedure. The couplingprotocol of PM1 to Affi-Gel 10 was based on the manufacturers specifications, however because PM1 isan 1gM class monoclonal, some alterations in time, pH and buffer strength were necessary. All procedureswere performed at 40 C.1. Purified PM1 antibody from 2 ml ascftes fluid was dialyzed against the coupling buffer (0.25 Msodium bicarbonate, pH 8.7) overnight, and 100 jil were removed and saved for later use todetermine the efficiency of protein coupling.2. Aft-Gel 10 was removed from the freezer, and 3 ml of the slurry suspension in isopropyl alcoholwere washed with 20 ml ice-cold distilled water using a Buchner filter with Whatman #54 hardenedfilter paper. One ml of the washed and now activated gel was quickly transferred to a 10 ml snapcap round bottom vial and stored on ice.3. The Jd buffered 1gM protein solution was diluted to 3 ml in coupling buffer (see above) and thenadded to the gel; the vial was rotated end over end for 1 hour at 4° C, and then for a further 4hours at room temperature. At this time, 100 jii of supernatant were removed and checked forcoupling efficiency, using UV absorption at 280 nm. For this, the supernatant was diluted in 0.01N HCI to lower the pH to prevent coupling by-products (hydroxysuccinimide) from interfering withthe absorption reading.4. After coupling efficiency was established, 0.1 ml 1 M ethanolamine HCI (pH 8.0) were added tothe slurry to block any remaining active ester groups, and rotation was continued for 1 hour.191Chromatography protocol:1. The coupled gel was transferred to a 3 ml column made from an empty syringe barrel which wasconnected to a Pharmacia ultraviolet detector (a 280). The column was washed with couplingbuffer until a flat line on the UV recorder was obtained.2. Prior to sample application, the column was pre-equilibrated with 20 column volumes of TBS, pH7.4. Guanidine HCI embryo extract (0.5 ml), which had been dialyzed against TBS overnight andfreshly supplemented with 1 mM PMSF, was then applied to the column at a flow rate of 5 mI/hourThe extract was recycled through the column 2 times, and the column was then washed with TBSuntil a flat line was achieved, indicating that all unbound protein was washed from the column.3. The PM1 antigen was then eluted off the column using 2 ml of 0.1 M triethylamine, pH 11.0. Thepeak was collected in 1.0 ml fractions into tubes containing 100 jil 1 M Tris, pH 6.0. Fractions wereimmediately pooled and concentrated 10 fold via ultrafiltration using Centricon-50 filters (Amicon),which have a molecular cutoff of 50 kD. The affinity-purified PM1 antigen was stored in smallaliquots at -70° C.4. The PM1 affinity-column was washed extensively with TBS and re-used successfully for 5additional chromatography runs, with only a minor decrease in peak size. For storage of thecolumn, 0.2% sodium azide was added to the TBS to prevent bacterial growth.192Appendix 15: ImmunoprecipltatlonsReagents:Embryo extraction buffer: 10 mM Tris, 1% triton X-100, 0.5% sodium deoxycholate, 0.1% SDS(sodium dodecyi sulphate), pH 7.3.immunoprecipitatior buffer: 20 mM Tris, 150mM NaCI, 0.1%triton X-100, pH 7.8.Protease inhibitors: 25 mM EDTA, 1 mM iodoacetamide, 1 mM PMSF, 1 in all buffers) pepstatin A, and 1 mM EGTAPerform all procedures at 40 C in 1.5 ml Eppendorf tubes.1. Pre-wash 50 J.Ll settled Protein A-Sepharose 4B fast flow (Sigma) with embryo extraction buffer byrotation end over end for a total of 3, 10 minute washes. After each wash, settle the Protein A-beadswith a brief (10 seconds) gentle spin in a clinical centrifuge.2. Incubate the washed Protein A-Sepharose with the linker antibody (goat anti-mouse lgG) for 1 hour(add 20 iI of a 1 in 40 dilution of whole serum).3. Wash with immunoprecipitation buffer as above three times over 30 minutes.4. Incubate with 10-100 jii hybridoma supernatant for 1 hour with constant rotation.5. Wash as in step 3, and save 25 jil of the bead suspension to be used later as a controlimmunoprecipitation (i.e. without embryo extract).6. incubate with 200 Iii embryo extract 1-2 hours with rotation. Be sure to add a fresh dose of PMSF(10 jii of a 1 mg/mI stock of PMSF in acetone) at this time.7. Wash with immunoprecipitation buffer thoroughly (3-5 times over 30 minutes).8. Resuspend beads in 50 il reducing sample SDS-PAGE buffer, and heat to 85° C for 15 minutes.Centrifuge 1 minute at 125 x g, and carefully remove supematant from bead pellet. Use immediatelyfor SOS-PAGE, or store at -20° C.193Appendix 16: BuffersPhosphate Buffered Saline (PBS)(Crawford, 1972)Use 50 ml Stock V and 50 ml Stock Vi and bring up to 1 L with distilled H20; adjust pH to 7.4Stock V Stock VINaC1 160.Og(0.14M) MgSO47H0 15.4g (0.003M)KCL 8.Og (0.005M) CaC122H20 1 .6g (0.0005M)KH2PO4 3.Og (0.OO1M) orNa2HPO47H20 5.8g (0.OO1M) CaC126H20 2.4g (0.0005M)oranhydrous 3.lg (0.OO1M)Dissolve in distilled H20 and make up to 1 L. Dissolve Mg first, add calcium, then bring upto 1 L with distilled H20.2% Blotto:Dissolve 2 g of Carnation non-tat milk powder in 1 liter PBS with 0.1% sodium azide; adjust pH to 7.4.TrIs Buffered Saline (TBS)(Sigma)TBS ranges commonly from 20-50 mM with 0.15 M NaCi. The pH varies with the temperature, therefore tomake a 1 L solution of 50 mM Tris, follow the table below:pH@ 5°C pH@25°C pH@37°C Trizma HCI Trizma Basegrams/Liter7.76 7.20 6.91 7.02 0.677.97 7.40 7.12 6.61 0.978.18 7.60 7.30 6.06 1.398.37 7.80 7.52 5.32 1.978,58 8.00 7.71 4.44 2.65194Citrate-Phosphate Buffer, pH 5.6:For 100 ml of buffer, mix 42 ml 0.1 M citric acid stock with 58 ml 0.2 M Na2HPO4 stock.Sodium Phosphate Buffer:(Gomori, after Sorensen, 1955)pH @ 25° C x ml 0.2 M Na2HPO4 y ml 0.2 M NaH2PO45.8 4.0 46.06.0 6.15 43.856.6 18.75 31.257.2 36.0 14.07.6 43.5 6.58.0 47.35 2.65


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items