Sea urchin morphogenesis and cell–hyalin adhesion are perturbed by a monoclonal antibody specific for hyalin

The hyaline layer (HL) is the external extracellular matrix of the sea urchin embryo which is secreted at fertilization. The HL can be dissolved in Ca²⁺-free medium (Herbst, 1900; Faust et al. 1959; Vacquier, 1969; Kane, 1970) and hyalin, the predominant protein, can then be purified by alternating Ca²⁺-induced precipitation with Ca²⁺ removal and solubilization (Kane, 1973). Hyalin is composed of approximately 25 % acidic residues, 3·5 % basic residues (Stephens & Kane, 1970; Citkowitz, 1971) and possibly 2–3% carbohydrate (Citkowitz, 1971). Sedimentation rate experiments performed by Stephens & Kane (1970) yield molecular weight values ranging from 100 to 350 ×10³ for 5. purpuratus and 370 to 760×10³ for C. atratus, whereas similar experiments reported by Gray et al. (1986) yield a molecular weight of 900×10³ for S. purpuratus hyalin. In gel electrophoresis experiments, hyalin migrates at a very large apparent molecular weight too great for quantification (Citkowitz, 1972). In addition, data presented by Citkowitz (1972) and Gray et al. (1986) provide support for the idea that pure hyalin preparations may be composed of at least two very high molecular weight species. Studies with immunofluorescent staining using polyclonal antisera, and incorporation of [³H]leucine have detailed the release and spread of hyalin from cortical granules and new synthesis of hyalin beginning between mesenchyme blastula and gastrula (McClay & Fink, 1982). In blastula-stage embryos, the basal components of the HL evolve into a distinct layer, the apical lamina (AL), located between the apical surface of the blastomeres and the HL (Wolpert & Mercer, T963). The HL-AL contains 145 and 175×10³M_r glycoproteins in addition to hyalin (Hall & Vacquier, 1982). Herbst (1900) found that sea urchin embryos incubated in sea water lacking divalent cations dissociated into blastomeres. Subsequently, this same treatment was used to dissolve the HL and extract hyalin components (Faust et al. 1959; Vacquier, 1969; Kane, 1970). Hyalin was then conveniently purified by repeated cycles of Ca²⁺-induced polymerization, followed by Ca²⁺ removal and resolubilization (Kane, 1973).

Herbst’s (1900) experiments and later observations indicating that microvilli fink the apical surfaces of the blastomeres with the HL (Dan, 1960), led Gustafson & Wolpert (1967) to postulate that the HL participates in morphogenesis as an anchoring substrate for epithelial sheets undergoing changes in shape. Consistent with this hypothesis, recent data establish that-ectoderm and endoderm cells bind to hyalin throughout embryogenesis, whereas mesenchyme cells do not (McClay & Fink, 1982; Fink & McClay, 1985).

In this work, we have discovered that a monoclonal antibody (McA Tg-HYL) specific for hyalin protein blocks cell-hyalin adhesion and inhibits gastrulation and arm rudiment formation in sea urchin embryos. This antibody recognizes hyalin from six species believed to have diverged between 100 and 200 ×10⁶ years ago (Durham, 1966). Our data indicate that McA Tg-HYL recognizes a highly conserved epitope of the hyalin protein involved in cell binding to the HL. It appears likely that by binding to this epitope McA Tg-HYL inhibits the normal association of embryonic cells with the HL, interfering with the epithelial folding required for gastrulation and arm rudiment formation.

T. gratilla and C. atratus embryos were raised at 22–24 °C in buffered sea water (Kane, 1973). S. purpuratus and L. pictus embryos were reared at 15°C and at pH8·0.

Crude egg hyalin protein was the generous gift of Dr Robert Kane or was purified using 1 M-glycine, 2 mM-EDTA (ethylene diamine tetra-acetic acid), 10 mM-iodoacetamide, aprotinin 100 Kallikrein Units ml⁻¹, lmw-PMSF (phenyl methyl sulphonyl fluoride) at pH 8· 3 (Kane, 1973) from freshly fertilized T. gratilla eggs, or blastula-stage embryos where indicated.

Monoclonal antibodies were generated by the University of Hawaii monoclonal antibody facility using sonicated, dis-tilled-water-lysed, pluteus-stage whole T. gratilla embryos as immunogen. Hybridomas were screened by indirect immunofluorescence as described below. McAs from ascites fluid were purified by the method of Goding (1983) and concentrated in Amicon Centricon-30 miniconcentrators to approximately 10 mg ml⁻¹. These purified Me As were used in all experiments. McA concentrations were determined by measuring OD₂₈₀ and using an extinction coefficient of 1–4 OD units mg⁻¹ ml⁻¹. Fab’ fragments were prepared by the method of Goding (1983).

Previously developed fixation and fluorescent antibody staining procedures (Dr Richard Allen, personal communication; Wang et al. 1982) were modified for sea urchin embryo whole mounts. Embryos were harvested by centrifugation and washed once in cold acid sea water (ASW: 0·2M-acetic acid in MFSW adjusted to pH 4·5 with NaOH) and twice with ice-cold MFSW. Embryo pellets were resuspended with 5 vols of ice-cold MFSW. An equal volume of freshly prepared, ice-cold 20% formalin in MFSW, pH 8-5, was added to the embryos, which were then placed on ice for 30 min with occasional stirring. Fixed embryos were washed three times in 20 vols of ice-cold MFSW followed by extraction with 10 vols of acetone at −20°C. After extraction for 20min, the acetone concentration was diluted to 50 % with ice-cold Tris-buffered sea water pH 8·5 (TSW 8 Previously developed fixation and 5: MFSW, 10 mM-Tris-HCl pH 8·5), and the embryos gently pelleted. Embryos were then washed three times with 20 vols ice-cold TSW 8·5, and kept as approx. 50% (v/v) suspensions in TSW 8·5, 0·02% NaN₃ or in 0·15M-sodium borate buffer pH8·3 at 4°C.

2–3 drops of fixed, extracted, embryos were added to 100 μl spent hybridoma culture medium or dilute ascites fluid and gently agitated. Negative controls lacked primary antibody or used a non-murine primary antibody, the positive control was a rabbit antibody to T. gratilla yolk protein. Fixed embryos were reacted with Me As at room temperature for 20min. Embryos were diluted in PBS pH 7·2 and pelleted at low speed for a total of three washes to remove unbound antibody. Washed embryos were reacted for 20 min with FITC goat anti-mouse IgG (Miles-Yeda Laboratories Cat. no. 65-171) antibody at a final concentration of 1:32. Embryos were washed twice and resuspended in 70% glycerol, 30% PBS pH 7·2, 0·02% NaN₃. For paraffin sections, embryos were fixed, embedded, sectioned and immunofluorescence performed as previously described (Wessel et al. 1984). Sections treated with periodic acid prior to reaction with primary antibody were incubated in 20mM-sodium periodate, 50mM-sodium acetate pH4·5 in the dark for an hour at room temperature. The slides were washed once in ddH₂O and reacted for 10min with 50mM-sodium borohydride in PBS, washed in PBS prior to antibody binding. For immunoperoxidase localization, paraffin sections were probed with a Vectastain ABC kit (Vector labs, cat. no. PK 4002) according to the manufacturers’ instructions.

T. gratilla embryos were harvested by settling or gentle centrifugation about half an hour prior to addition of tracer and resuspended to 10 % (v/v) with TSW 8·5. Typically, the suspensions were also quantified by counting the embryos in 50μl drops. [³⁵S]methionine (Amersham or NEN, >1000 Ci mmol⁻¹ in tricine buffer) was added to the embryos to a final specific activity of 100 pCi ml⁻¹. Embryos were incubated for 3h (0–3, 9–12, 21–24, 33–36, 45–48h postfertilization) at room temperature (22-24°C). Labelled embryos were washed once with 20 vols of ASW, resuspended in cold TSW 8·5, aliquotted, pelleted for 10s in a Brinkmann Microfuge, flash frozen in liquid nitrogen and stored at −80°C until needed.

Immunoprecipitations were performed by the method of Rohrschneider et al. (1979) with the following modifications. Frozen radioactive embryos were sonicated in 5 vols of RIPA buffer (radio immunoprecipitation assay buffer; 10 mM-Tris-HCl pH7·2, 0·15M-NaCl, 1% sodium deoxycholate, 1 % Triton X-100, 01 % SDS, 100 Kallikrein Units of aprotininml⁻¹,25 mM-EDTA, 0·01 % pepstatin A, 10 mM-iodoacetamide, 1 mM-phenylmethylsulphonylfluor-ide). Lysates were precleared by incubation with 0·01 vols 10 mg ml⁻¹ normal rabbit IgG or 50 μl of rabbit anti-mouse serum (Dako Z-259, Accurate Chemicals, New York) for 30min on ice, followed by addition of 100 mg packed formalin-fixed Staphylococcus aureus per 50 μl normal rabbit IgG and another 20min on ice. Aliquots of precleared lysates containing 10⁶-10⁷ cts min⁻¹ were reacted with 25 μg of monoclonal antibody for 3h at 4 °C. Antigen-antibody complexes were isolated by addition of 10 mg S. aureus which had been precoated with rabbit anti-mouse serum. Immunoprecipitated proteins were resolved on 4% polyacrylamide gels and detected by fluorography with Enhance (NEN).

SDS-polyacrylamide slab gel electrophoresis was performed according to the method of Laemmli (1970) with T:C ratios taken from Citkowitz (1971). Gels were silver stained by the method of Merrill et al. (1984), or using Coomassie R-250. Soluble hyalin protein, quantified by the method of Lowry et al. (1951), was denatured in 4 vols l·25×Laemmli sample buffer containing 25 mM-EDTA, and heated to 70°C for 5 min. Samples run under reducing conditions included 1·5% DTT in the sample buffer.

Electrophoretic transfers of proteins from slab gels to nitrocellulose (NC) were performed according to Towbin et al. (1979) with the following modifications made to the apparatus. Scotchbrite pads were replaced with 1cm thick sheets of sponge scavenged from packing materials. Platinum wire electrodes were replaced by stainless steel 11×13 cm sheets (Vaessen et al. 1981). Transfers were performed at 8–10 V cm⁻¹ for 4h, in a Tupperware container surrounded by an ice bath. Blots were stored under transfer buffer in the refrigerator for at least 24 h to maximize the binding of the transferred proteins.

Blots were briefly rinsed in TBS pH7·6 (TBS: Trisbuffered saline, 150mM-NaCl, 10mM-Tris-HCl) containing 5 mM-MgC1₂ and transferred to heat-sealable plastic (Seal-a-Meal) bags where all reactions and washes were carried out. Transfers were quenched with 5 % BSA in TBS/MgCl₂ for an hour, reacted with 5 μg ml⁻¹ primary antibody in quenching buffer for 1 h and washed twice for 15 min with TBS/MgC1₂. Secondary antibody reagent (goat-affinity-isolated antibodies to mouse IgG and IgM, alkaline phosphatase conjugated, Boehringer Mannheim biochemicals cat. no. 60527) was diluted 1:2000 in quenching buffer, reacted to the blot for an hour and washed twice. Bound antibody complexes were visualized using the chromogenic substrate of Leary et al. (1983).

To test for morphogenetic effects of the antibodies on whole embryos, varying amounts of purified concentrated monoclonal antibodies were added to 1 ml samples of a 0·1% embryo culture (1000 embryos ml⁻¹). The embryos were allowed to develop at room temperature in the case of T. gratilla, C. atratus, L. variegatus and A. punctulata and at 15 °C for S. purpuraras and L. pictus. Development was monitored by microscopy without coverslips. Specimens were photographed on a Zeiss Universal microscope, equipped with a camera and bright-field strobe illumination or differential interference contrast optics, using Agfa Superpan 200 black and white film or Ektachrome 160.

Centrifugal adhesion assays were performed as described by McClay et al. (1981) and Lotz et al. (1988) with the following modifications. Cell adhesion to substrate rather than to cell monolayers was measured. Quenching of excess protein-binding capacity was done in 10 % BLOTTO (Johnson et al. 1984). 20μgml⁻¹A. punctulata hyalin (purified by the method of Kane, 1973) was bound to microtitre plate wells for 2 h at room temperature and wells were rinsed twice between hyalin binding and BLOTTO and between BLOTTO and 30 μg ml⁻¹ McA Tg-HYL binding. Antibody incubations were performed for 1 h at room temperature. Control Fab’ fragments were made from mouse nonimmune serum by the method of Goding (1983). A. punctulata embryos were labelled for 6h in Iμ Ci ml⁻¹ [³H]leucine prior to dissociation in hyalin extraction medium (Fink & McClay, 1985). Dissociated cells were resuspended in TSW 8·5 and 5μl (9·0×10⁵ cells) of cell suspension were added to each well. Microtitre wells were then topped up with 50μg ml⁻¹ McA Tg-HYL in TSW 8-5 and remained at room temperature for 1h prior to centrifugation. Cells were spun onto substrate-coated wells for 8 min at 20 g, the plates inverted and spun off for 8 min at 25g.

The cell-cell reaggregation assay used is a modification of the method of Noll et al. (1979). Blastula-stage T. gratilla embryos were dissociated to single cells by the method of Kane (1973), resuspended in a 1:1 mixture of 1M-glycine 2 mM-EDTA pH8-3/MFSW, pelleted and washed once in MFSW. Cell pellets were resuspended in MFSW to a concentration of 2×10⁷ cells ml⁻¹. 10³ cells were added to microtitre plate wells (Flow Labs, FB-48-CLEAR) containing 0, 1-0,10 or 100 μg McA Tg-HYL or a control McA in a final volume of 95 pl (adjusted with TSW 8·5), and allowed to reaggregate at room temperature. Reaggregates were photographed on a Nikon inverted microscope, with Agfa Superpan 200 film.

An in vitro assay for the gelation of hyalin in the presence of Ca²⁺ ions was adapted from the sponge aggregation factor gelation assay developed by Rice (1983). 12·5 μl 0·04% Alcian blue 8GX (Sigma A-5268) was added to 25 μg of soluble T. gratilla hyalin in 0·5 ml TBS pH 7·6 in a glass scintillation vial. 5μl of 2M-CaCl₂ was added to the vial which was then immediately swirled at 60 revs min⁻¹ at room temperature, yielding blue clumps of gelled hyalin within minutes. To test the effect of a monoclonal antibody on hyalin gelation, 1, 10, 20, 50 or 100pg of monoclonal antibody or Fab’ fragments were included prior to the addition of Ca²⁺. Ca²⁺-induced gelation of hyalin was assessed 5 min after the addition of CaCl₂.

300 hybridomas obtained from a single fusion of spleen cells from a mouse immunized with whole T. gratilla pluteus-stage embryos were tested for indirect immunofluorescence staining of specific embryonic structures. Two hybridoma supernatants (McA Tg-HYL and McA 92), which were observed to react to the HL of T. gratilla (Fig. 1), were recloned twice to ensure monoclonal provenance of the antibodies. The specificities of the antibodies were monitored at all steps by indirect immunofluorescence. Cross reactivity of the antibodies by immunofluorescence was tested on all of the sea urchin species used. One of the antibodies, the subject of this report, exhibited immunofluorescence staining patterns with five other species (C. atratus, S. purpuratus, A. punctulata, L. variegatus and L. pictus) similar to the pattern observed with T. gratilla (Fig. 1). All McAs referred to in this work are mouse IgG1 subtype.

The HL-AL components that are recognized by McA Tg-HYL were examined using immunoprecipitation of total radioactive proteins from T. gratilla embryos incubated at various stages with [³⁵S]methionine. Although hyalin is present in large amounts throughout development, hyalin synthesis was not observed before gastrula stage (Fig. 2D) in agreement with the observation of McClay & Fink (1982). Fig. 2 shows a fluorogram of the immunoprecipitated proteins separated on an SDS-polyacrylamide gel in the presence of reducing agent. As shown in lanes D-F, hyalin migrates under these conditions as a smeared doublet with an apparent relative molecular mass significantly in excess of 300×10³. Hyalin is the most prominent material immunoprecipitated by McA Tg-HYL as evidenced by the broad bands at the top of lanes D, E and F which are respectively from gastada-, prism- and pluteus-stage embryos. The higher molecular weight hyalin band (which is also the predominant species as determined by its Coomassie and silver staining on gels) is synthesized earlier than the lower molecular weight portion of the doublet. However, at least three other polypeptides with apparent of 190, 140 and 105 ×10 ³ were also immunoprecipitated from prism and pluteus stages. Control lanes of immunoprecipitates with a monoclonal antibody of the same subclass but with an unrelated specificity (chick bursa cell extracellular matrix proteoglycans) show few radioactive bands, indicating that McA Tg-HYL specifically is involved in the immunoprecipitation of hyalin as well as the 190, 140 and 105 ×10³ proteins. These proteins, which are immunoprecipitated along with hyalin, will be shown below to be components of the HL-AL and may be analogous to AL proteins previously characterized in S. purpuratus (Hall & Vacquier, 1982).

Ca²⁺ precipitation and immunoblot analysis of HL-AL components removed from blastula embryos by the standard 1 M-glycine, 2 mM-EDTA extraction procedure provided strong evidence concerning the specificity of McA Tg-HYL. When the antibody is reacted to a protein gel blot of T. gratilla hyalin purified via three cycles of Ca²⁺ precipitation (Fig. 3A) only the major band gives a signal (Fig. 3B). The same band reacts in immunoblots of crude glycine extract from blastula-stage embryos (Fig. 3C). A control immunoblot using nonimmune mouse serum instead of McA Tg-HYL gives no signal (Fig. 3D). The crude extract clearly contains numerous proteins that are not recognized by McA Tg-HYL, including the 190, 140 and 105×10³ bands as seen in autoradiographs of Ca²⁺-soluble (Fig. 3E) and -precipitable (Fig. 3F) material from the glycine extract of [³⁵S]methionine-labelled blastulae. Although there is high molecular weight material seen in the Ca²⁺-soluble material (Fig. 3E) it does not comigrate with hyalin. These data demonstrate that McA Tg-HYL is specific for hyalin. This result is confirmed by the fact that McA Tg-HYL, reacted to gel blots of McA Tg-HYL immunoprecipitates, does not bind to anything other than the immunoprecipitated hyalin band (D. Adelson, unpublished data).

It is interesting to note that the 190, 140 and 105×10³ bands are enriched in the Ca²⁺-precipitable fraction, consistent with the idea that they are HL-AL proteins normally associated with hyalin. These hyalin-associated proteins are coordinately synthesized independently of hyalin since they are evident at 0-3h (Fig. 2B), blastula (Figs 2C, 3E,F), at prism (Fig. 2E), at pluteus (Fig. 2F) but not at gastrula (Fig. 2D). The Ca²⁺-soluble material contains high molecular weight bands which migrate below hyalin, but due to their size may include the large subunit of laminin.

Evidence for the chemical nature of the McA Tg-HYL epitope conies from Fig. 4 (C,D) where 20 mw-periodic acid oxidation does not destroy immunoreactivity compared to a control carbohydrate epitope which is no longer bound by McA Ic10 (Nelson & McClay, 1986) after 20 niM-periodic acid treatment (Fig. 4A,B). In addition, protein gel blots of limited proteolytic digests of hyalin are not recognized by McA Tg-HYL (D. Adelson, unpublished data). This indirect evidence is consistent with a peptide rather than carbohydrate epitope.

The immunofluorescence cross-reactivity seen in all sea urchin species tested indicated to us that the McA Tg-HYL epitope might be a significant functional domain of the hyalin protein. This idea was tested by incubating developing embryos in McA Tg-HYL. Irrespective of the stage of embryos, significant alteration of further morphogenesis occurred in 5–70% of the treated embryos when 5 μg ml⁻¹ of the antibody was added to 1000 embryos ml⁻¹. 100% of the treated embryos were affected when 10 μg ml⁻¹ or more of antibody was added per 1000 embryos ml⁻¹. Blastocoel expansion, gastrulation and arm rudiment formation were all inhibited. In addition, the epithelial cells of stunted embryos were seen to round up and the epithelium thicken compared to controls. Similar results were obtained with six species, T. gratilla, L. pictus, L. variegatus, S. purpuratus, A. punctulata and C. atratus. Embryos of T. gratilla, L. pictus and L. variegatus evidenced a loosening of the HL from the underlying cells (Fig. 5B, Fig. 6A) which often progressed to extensive delamination of the HL (Fig. 6A). Fab’ fragments made from McA Tg-HYL were tested at concentrations ranging from 6pgml⁻¹ to 30pgml⁻¹ on L. variegatus embryos and produced stunting as well (data not shown).

Fig. 5 shows the results obtained with T. gratilla when McA Tg-HYL at 10 μg ml⁻¹ was added to (1) late-blastula-stage embryos, when the vegetal plate thickens prior to primary invagination of the archen-teron, (2) late-gastrula-stage embryos when the gut is complete and (3) prism-stage embryos when arm growth is beginning. Addition of the antibody before gastrula stage inhibits invagination of the archenteron (Fig. 5B) and prevents completion of invagination if added during gastrulation. In T. gratilla, the cells of the presumptive endoderm pile up in the blastocoel of the stunted embryo (Fig. 5B indicated by PE), similar effects are seen in L. variegatus (Fig. 6A,B). When the antibody is added to spherical gastrulae (Fig. 5D), the embryos remain spherical (Fig. 5E) rather than becoming prisms (Fig. 5E). If the antibody is added to prisms (Fig. 5G), the embryos remain as prisms (Fig. 5H) rather than developing into plutei (Fig. 51). Differentiation of tissues in these embryos appears to progress in spite of the failure to undergo normal morphogenetic changes. Pigment cells appear on schedule and in appropriate locations. Although triradiate spicules are formed in treated embryos, they are not able to extend to normal dimensions apparently because of the failure of arm formation. The reversibility of the antibody is addressed by the following observation; at concentrations of 10μgml⁻¹ or less of McA Tg-HYL per 1000 embryos ml⁻¹ and 24–36 h after stunting, T. gratilla embryos often recovered and continued with normal morphogenesis. This spontaneous recovery was also seen in L. variegatus but with slightly different timing (Fig. 6C), presumably due to turnover of the hyalin-antibody complex or to eventual overwhelming of the antibody by newly synthesized hyalin. Specific experiments to assess the reversibility of the stunting were deemed impractical since antibodies, unlike low molecular weight drugs, cannot be washed off the embryos. Controls in all experiments performed included growing embryos in the presence of McA 92 which also stains the hyalin layer by indirect immunofluorescence and immunoprecipitates the same four bands as McA Tg-HYL, but whose epitope is not retained after gel transfer to nitrocellulose (Adelson, 1985). McA 92 at 100 ng ml⁻¹ had no effect on embryos of any species studied. Frames C, F and I in Fig. 5 show examples of such controls for T. gratilla.

A possible mode of action of McA Tg-HYL on the HL is the weakening or depolymerization of the HLby binding of the antibody to the protein domain responsible for the protein-protein interactions stabilizing the hyalin gel. An assay for Ca²⁺-induced gel formation of microgram quantities of hyalin was developed based on the visualization of small gels by Alcian blue staining (Rice, 1983). Fab’ fragments and intact McA Tg-HYL had no effect on hyalin gelation, implying that destabilization of the hyalin gel was not involved in the inhibition of morphogenesis caused by McA Tg-HYL. This experiment does not rule out a mechanism where McA Tg-HYL dissociates the HL from the AL.

Cell-cell interactions are often important in morphogenesis and hyalin has been implicated in sea urchin embryo cell adhesion (Spiegel & Spiegel, 1975; Timourian & Watchmaker, 1975). It seemed possible that McA Tg-HYL inhibited morphogenesis by affecting cell-cell association. This idea was examined by testing the effects of the antibody on the reaggregation of dissociated blastula cells. These cells reaggregate normally in the presence of McA Tg-HYL, as well as with a control McA, at concentrations of 10 μg to 1m g ml⁻¹ (see Fig. 7, panels B,C,D,F,G,H).

Reaggregation was inhibited by a rabbit polyclonal antiserum at 500 μ gml⁻¹, previously shown to inhibit aggregation (Noll et al. 1979; Fig. 7, panel A). These results indicate that cell reaggregation and cell-cell association is not perturbed by McA Tg-HYL. There is a slight qualitative difference in the morphology of the reaggregates in panel D, but this was at antibody concentrations 100 times greater than those used in stunting experiments. Since significant cellular metabolism is required for reaggregation of these cells (Giudice, 1965), these results also indicate that McA Tg-HYL is not exerting its morphogenetic effects by some general cytotoxic reaction.

We hypothesized that McA Tg-HYL might be exerting its stunting effect by covering a portion of the hyalin molecule involved in cell-hyalin adhesion. This hypothesis was tested by using a previously described adhesion assay (McClay et al. 1981) as modified by Lotz et al. (1988). Gastrula-stage embryos grown in [³H]leucine were completely dissociated and the cells added to microtitre wells previously coated with hyalin. Some of the wells were treated with anti-hyalin Fab’ fragments and controls were treated with equal amounts of nonspecific mouse Fab’ fragments. The cells were centrifuged onto the substrates to allow them to bind, followed by inversion of the microtitre plates and centrifugation in the opposite direction to shear unbound or loosely bound cells off the substrate. The unbound cells and cells bound to the substrate were separately quantified via scintillation counting. Background binding was determined by adding cells to wells that had not been coated with hyalin. The results in Table 1 show 20μgml⁻¹ Fab’ fragments made from McA Tg-HYL reduced the number of A. punctulata cells bound to A. punctulata-hyaVm-coateà microtitre plate wells by between 23 % and 50 %. The differences in values seen between the two experiments reported in Table 1 can be attributed to two factors; higher background binding during experiment 2 and differing amounts of hyalin remaining on the cell surfaces after dissociation. The latter difference could affect results two ways; by tying up Fab’ fragments and by contributing to adhesion via cellular hyalin-substrate hyalin polymerization. In spite of this day-to-day variability, these results support the hypothesis that the antibody is capable of specifically perturbing cell-hyalin interactions.

The specificity of McA Tg-HYL for the major component of hyalin is demonstrated by the combined results of Figs 2 & 3. Previous work by Citkowitz (1972) and Gray et al. (1986) shows that purified hyalin is not homogeneous, but contains a major and a minor component. It is not known whether these two components are present as precursor and product or whether they are related isoforms. The metabolic labelling data presented in Fig. 2 do not support a precursor/product relationship unless the processing step is itself developmentally regulated. The protein immunoblot data of Fig. 3 show that the antibody reacts specifically with the predominant species seen when gel lanes of purified hyalin are silver stained. The 190,140 and 105 × 10³M_T polypeptides seen in the immunoprecipitations (Fig. 2) do not react with the antibody, although their presence can be ascertained from the autoradiographs (Fig. 3E,F) of the Cabsoluble and -precipitable material from the blastula glycine extract. Blastula-stage embryos were chosen for this experiment since they do not synthesize hyalin but do synthesize the 190, 140 and 105 ×10³ proteins. This allowed us to determine that none of the metabolically labelled material reacted specifically with McA Tg-HYL. We could thus attribute the presence of the 190, 140 and 105xlO³ bands in immunoprecipitates to other factors, most likely to interactions with hyalin itself.

The in vivo effects of McA Tg-HYL are quite specific; epithelial cells in treated embryos round up and cause the epithelium to thicken, followed by the inhibition of archenteron invagination and arm rudiment formation. In addition, three closely related species exhibited gross delamination of the epithelium from the HL. We ruled out the antibody’s effects being a function of general toxicity for two reasons; (1) mesodermal differentiation proceeds without delay as evidenced by spicule formation and the appearance of pigment cells (Fig. 5H), (2) reaggregation experiments performed in the presence of the antibody reveal no delay or abnormal reaggregate morphology (Fig. 7). Furthermore, the antibody did not affect hyalin protein gelation in vitro or cell-cell association (Fig. 7) which has been postulated to involve the hyalin protein (Spiegel & Spiegel, 1975; Timourian & Watchmaker, 1975). Having ruled out these explanations, we decided to test whether the antibody had a direct effect on cell-hyalin adhesion. McA Tg-HYL Fab’ fragments at a concentration sufficient for in vivo stunting were found to inhibit cell hyalin adhesion by 23–50% (Table 1). This observation, along with the fact that the antibody causes the separation of the HL from the ectoderm, provides the clearest evidence concerning the function of this epitope. Taken together with its periodate resistance (Fig. 4B,C) and its susceptibility to proteolytic digestion (Adelson, unpublished data) the abovementioned data suggest that the epitope is part of a cell binding protein domain of the hyalin molecule recognized by epithelial cells (Fink & McClay, 1985).

The immunological cross-reactivity of McA Tg-HYL to embryos from divergent groups of sea urchins indicates that the recognized epitope is strongly conserved. Of the five species of urchins in addition to T. gratilla that bind the antibody, L. pictus and L. variegatus are the most closely related. This agrees with our observation that the antibody only causes gross separation and delamination of the HL in these three species. The affinity of the antibody for the epitope may be reduced or some portion of the epitope may be qualitatively different in the remaining species which diverged from T. gratilla 200 million years ago, such that full inhibition of cell attachment to the HL does not occur. However, the epitope still must be recognized to a significant extent since the antibody stains the HL by immunofluorescence and prevents further morphogenesis. Isolation of this epitope and analysis of its sequence or structure from several species should provide interesting insights into the nature of hyalin function.

Protein domains involved in cell-substratum attachment and migration have been identified on fibronectin and discoidin I (Pierschbacher & Ruoslahti, 1984; Yamada & Kennedy, 1984; Springer et al. 1984). These are related tetrapeptide sequences which have also been identified in such evolutionarily divergent proteins as the fibrinogen alpha chain, the E. coli lambda receptor, sindbis virus coat protein, alpha lytic protease and testis-specific basic protein (Pierschbacher & Ruoslahti, 1984). A decapeptide including the cell binding tetrapeptide domain from fibronectin has been shown to inhibit gastrulation in amphibians and neural crest cell migration in birds (Boucaut et al. 1984). In Dictyostelium discoideum, the tetrapeptide present in the discoidin I sequence is required for cellular streaming during morphogenesis (Springer et al. 1984).

A distinct cell binding domain (Graf et al. 1987) is also present on laminin, an extracellular matrix protein which has recently been demonstrated in the HL and basal lamina of sea urchin embryos (McCarthy et al. 1987). McCarthy & Burger (1987) report that intact monoclonal anti-laminin injected into the blastocoel is followed by cellular elongation, rounding up of cells and subsequent loss of epithelial character. They do not report subsequent morphogenetic effects, whether the effects of microinjected Fab’ fragments are similar or if direct incubation of embryos in their antibody has any effect on epithelial structure. There is no reason to believe that McA Tg-HYL is similar to this anti-laminin monoclonal either in specificity or in function. Our antibody is unambiguously directed against hyalin, which is only present on the apical surface of the embryo. In addition, although our antibody initially has similar effects on the morphology of the epithelium, it has been demonstrated that this is almost certainly due to its inhibitory effect on cell-hyalin adhesion and not a function of antibody-mediated crosslinking of hyalin. We also report subsequent significant morphological impairment of the embryo as a result of our antibody treatment.

The inhibition of sea urchin embryo morphogenesis by a monoclonal antibody that specifically recognizes hyalin provides new evidence that the HL plays an important role in morphogenesis. Ideas of this nature are persistent in the literature of sea urchin development (Chambers, 1940; Dan, 1960; Gustafson & Wolpert, 1967). Dan (1960) postulated that the HL provides an anchorage for epithelial cells while they are undergoing the stress and changes in shape taking place during invagination. Gustafson & Wolpert (1967) speculated that differences between the adhesiveness of cells to cells and cells to the HL could account for the primary invagination of the archenteron. Other epithelial folding such as outpocketing of the arm rudiments could be produced by reverse inequalities in adhesiveness. Citkowitz (1971) put forward experimental evidence based on nonspecific proteolytic digestion of embryo surfaces that the HL was important for gastrulation. Our data support and extend this conclusion. Horstadius (1939) noted that the initial stages of arm formation occurred in the absence of spicular development, but that the full extent of arm formation required spiculogenesis. In our experiments, spicules are formed (Fig. 5) but fail to grow when arm rudiment formation is inhibited in the presence of McA Tg-HYL. This provides support for the idea that full arm extension involves a feedback interaction between the spicule-forming mesodermal cells and the HL-binding ectodermal cells, essentially as suggested by Hörstadius (1939). The mesenchymal syncytium must recognize and enter the outpocketing of the arm epithelium to produce further spicule growth. This syncytial extension may depend on mesenchymal cell binding to specific features of the basal lamina produced by the overlying ectodermal cells (Katow & Solursh, 1981; Wessel et al. 1984).

In conclusion, we propose that McA Tg-HYL inhibits epithelial folding by blocking cell-hyalin adhesion. Whether or not this adhesive interaction is mechanically sufficient to anchor the epithelial sheet during morphogenesis is unknown. We would not be surprised if epithelial integrity were also reduced due to disruption of the epithelial cells’ cytoskeleton, which would occur subsequent to loss of cell-hyalin contact.

D.L.A. is indebted to Carol Burdsal for performing adhesion assays, Drs Albert Benedict, Karen Yamaga, Leslie Tam and John Spencer for their immunological expertise, Drs David R. McClay, Robert Kane, Greg Dolecki, Steve Black, Mark Alliegro and Charles Etten-sohn for helpful discussions, Margaret Lotz-Bousvarous for the periodate protocol and Susan Nelson for the use of Ic10.