INTRODUCTION

Developing cells interact with, and are influenced by, neighboring cells and the surrounding environment. These interactions contribute to the determination of cell fate and patterning in developing organisms. Cellular interactions are mediated by both diffusible small molecules and extracellular macromolecules, which can be either components of the cell surface or the extracellular matrix (ECM)¹ (Edelman, 1986; Ekblom et al., 1986). These extracellular macromolecules play structural and instructive roles in morphogenesis including cell-cell adhesion, cell migration, and induction of cell differentiation (Prochniantz and Theodore, 1995; Ruoslahti and Yamaguchi, 1991). In plants, the cell wall is the main ECM which functions to control cell shape and cell expansion, both of which are essential to plant morphogenesis (Cassab and Varner, 1988; Varner, 1989). Various models of cell wall structure and assembly have been proposed, but elucidating the basic biochemical activities required for cell wall formation and modification remains a major area of investigation (Carpita and Gibeaut, 1993). The ECMs of both animal and plant cells are composed of a variety of molecules including glycoproteins, proteoglycans, and glycolipids. In addition to these components, plant cell walls are cellulose based and contain substantial amounts of heteropolysaccharides.

At least four different general forms of ECM are produced during Dictyostelium development: 1) an ECM is secreted by aggregating cells which facilitates cell movement (Williams and Joss, 1993); 2) during slug migration, a well defined sheath is secreted and surrounds the multicellular slug (but not individual cells). The sheath ECM is composed of glycoproteins and cellulose (Freeze and Loomis, 1977a, 1977b; Grant and Williams, 1983). 3) At culmination, the anterior prestalk cells produce an ECM that surrounds the multicellular stalk. This ECM is also composed of cellulose and proteins (Blanton, 1993; Freeze and Loomis, 1978), two of which are known to be high molecular weight glycoproteins encoded by the ecmA and B genes (McRobbie et al., 1988). The stalk and slug sheaths have similar general biochemical characteristics (Freeze and Loomis, 1978). 4) Also during culmination, the posterior prespore cells secrete a variety of glycoproteins from the prespore vesicles (PSVs). These secreted glycoproteins combine with cellulose (which is not in the PSVs) to form the environmentally resistant spore coat. Clearly these ECMs play central roles throughout morphogenesis. However, there is only a rudimentary understanding of their composition, biochemical function, and the mechanisms controlling their developmentally regulated secretion and assembly.

Our work has focused on the last of these ECMs, the spore coat. We have characterized a family of developmentally regulated glycoproteins that share a common O-linked oligosaccharide modification (Alexander et al., 1988). Addition of this modification to proteins is dependent on the wild-type modB allele, and the oligosaccharide epitope is recognized by the MUD50 monoclonal antibody (mAb) (Alexander et al., 1988; Murray et al., 1983; West and Loomis, 1985). The MUD50 reactive oligosaccharide modification is found on proteins that are expressed at several different stages of morphogenesis but is particularly prominent during the migrating slug stage. The majority of the MUD50 reactive proteins at this developmental stage are expressed in prespore cells. Some of these proteins are membrane bound while others are soluble. Several of these MUD50 reactive proteins have now been shown to localize to ECMs. The sheathin proteins are a component of the sheath that surrounds the migrating slug (Zhou-Chou et al., 1995), and the polymorphic PsB (prespore protein B) glycoprotein becomes part of the extracellular spore coat (Smith et al., 1989; Watson et al., 1994; West et al., 1986).

The PsB glycoprotein is recognized by the mAb MUD102 and is part of a larger specific multiprotein complex² (Watson et al., 1993). Of these proteins, only PsB carries the modB oligosaccharide. The p112/SP96 and p78 proteins in the complex are glycosylated with a fucose containing oligosaccharide. Both of these oligosaccharides require the modC and E genes for biosynthesis, but only the fucose containing epitope on p112/SP96 is recognized by the MUD62 mAb (Champion et al., 1991, 1995; Gonzalez-Yanes et al., 1989; Watson, 1994). p112/SP96 is highly phosphorylated although the significance of this phosphorylation is unclear (Watson et al., 1993). The PsB multiprotein complex is preassembled in PSVs where it is stored awaiting the signal for secretion. During the terminal stage of spore maturation, these PSVs fuse in a developmentally regulated manner with the cell plasma membrane and the PsB complex is secreted. After secretion it becomes part of the spore coat ECM (Watson et al., 1994). Additional proteins, including the fucosylated and phosphorylated SP75, are incorporated at this time (Watson, 1994; Watson et al., 1994). Within the spore coat, the PsB complex is positioned asymmetrically so that p112/SP96 is detected by immunofluorescence on the outside of the spore coat while the PsB protein is detected only on the inside (Richardson and Loomis, 1992; Watson et al., 1994). Thus, the complex may span the middle layer of the coat which is known to be composed primarily of cellulose (Hemmes et al., 1972; West and Erdos, 1990). Nothing is known about the developmental signals that control the initiation of secretion or the molecular mechanisms underlying the fusion event.

Although our studies demonstrated that the PsB complex is part of the spore coat, a specific biochemical activity that would account for its localization had not been defined. In addition, nothing was known about the specific assembly of the individual protein subunits in the multiprotein complex. Both questions are fundamental to our understanding of the organization and function of ECMs. In this study, we demonstrate that the PsB complex has a specific binding activity for cellulose, which is consistent with its localization in the spore coat. We also demonstrate the order of assembly and relative topology of the PsB complex proteins and show that subunit assembly is required for the cellulose binding activity.

MATERIALS AND METHODS

Dictyostelium discoideum cells were grown in association with Klebsiella aerogenes on SM plates to late log phase (2-3 × 10⁸ cells/100-mm plate). Development was initiated by scraping the cells from the plates and washing away the remaining bacteria by differential centrifugation in lower pad solution buffer (20 mM KCl, 2.5 mM MgCl₂, 40 mM potassium phosphate buffer (pH 6.5) containing 0.5 mg/ml streptomycin sulfate) (Sussman, 1987). The washed cells were plated for development on sterile lower pad solution saturated black paper filters (Thomas Scientific 4740C20) at 10⁸ cells per 40-mm diameter filter (Soll, 1987; Sussman, 1987). Strain WS380B was used throughout this study. Its development is highly synchronous and is shown schematically in Fig. 2. Note that WS380B is a wild isolate and its development is accelerated compared to that of the standard lab strains NC4 and its axenic derivative Ax3. Mutants lacking specific spore coat proteins (cot mutants) were described by Fosnaugh et al. (1994). Table I describes the strains used in this work.

Table I. Strains used in this study Strain Genetic background Modification References WS380B WT None Erdos et al. (1973) HL328 AX3 pyr5-6 point mutation Kuspa and Loomis (1994) TL55 (cotA) HL328 Disruption in gene for SP96 Fosnaugh et al. (1994) TL26 (cotB) HL328 Disruption in gene for SP70 Fosnaugh et al. (1994) TL52 (cotC) HL328 Disruption in gene for SP60 Fosnaugh et al. (1994) TL56 (cotABC) HL328 Disruptions in genes for SP96, SP70, and SP60 Fosnaugh et al. (1994)

For radiolabeling, filters containing developing cells were transferred to a fresh 60-mm Petri dish containing a 100-µl drop of lower pad solution containing 200 µCi of [³⁵S]methionine. The developmental stage for labeling is noted in the text for each experiment. Cells were harvested by transferring the filters to a 15-ml disposable plastic tube containing 5 ml of cold H₂O. Vortexing released the cells, or aggregates, from the filters, and they were collected by centrifugation at 500 × g in a clinical centrifuge. The cells were washed once in cold H₂O and frozen as pellets at 80 °C. The cells were lysed by thawing the frozen pellets and adding 0.5 ml of 10 mM Tris-HCl, pH 8.0, containing 1% Nonidet P-40 and 0.2 mM AEBSF (protease inhibitor). The same procedure was used to lyse both labeled and unlabeled cells. These lysates are used for immunoprecipitation, affinity chromatography on cellulose (see below), or Western analysis (Alexander et al., 1991; Watson et al., 1994, 1993). The antibodies used in this study, and their reactivities, are described in Table II.

Table II. Antibodies used in this study and their cognate proteins Name Protein(s) and epitope recognized Antibody type Reference MUD50 PsA, PsB, sheathins and others; modB oligosaccharide Mouse IgG Alexander et al. (1988), Grant and Williams (1983), and Smith et al. (1989) MUD51 Sheathins; peptide backbone Mouse IgG Grant and Williams (1983) MUD62 p112/SP96 and SP75; fucose oligosaccharide Mouse IgG Champion et al. (1991), Champion et al. (1995), and Grant and Williams (1983) MUD102 PsB; peptide backbone Mouse IgM Smith et al. (1989) and Watson et al. (1993) Anti-discoidin Discoidin I and II proteins Rabbit Ig Alexander et al. (1983) Jab2 ST310 and ST430, stalk proteins Mouse IgG Wallace et al. (1984)

Affinity chromatography was used to assay for cellulose binding activity. Fibrous cellulose from cotton (Sigma C-6288) was used as the affinity matrix. The columns were washed with 20 column volumes of wash buffer (10 mM Tris-HCl, 150 mM NaCl, pH 7.4) immediately before use. For experiments using radioactively labeled material, 0.4 g of cellulose were used for each column. 8 × 10⁷ cells were lysed as described above. The extracts were loaded onto the columns at 4 °C and allowed to interact with the cellulose for 1 h. The column was then washed with 25 column volumes of wash buffer to remove unbound proteins, followed by washing with 10 column volumes of wash buffer containing 0.5 M NaCl to remove loosely bound proteins. Ten column volumes of wash buffer were then run through the column to remove any remaining NaCl solution. Bound proteins were eluted with 10 column volumes of 6 M urea in wash buffer, followed by 10 column volumes of wash buffer, which was all collected as one fraction. The column was then washed with 10 column volumes of 1% SDS in wash buffer. The fractions were concentrated using acetone precipitation or lyophilization and analyzed by SDS-PAGE.

Strain WS380B cells were grown, allowed to develop, and harvested at the slug stage of development as described above. These aggregates were then lysed in Nonidet P-40 lysis buffer containing either 0.15 M cellobiose, trehalose, glucose, galactose, or in the absence of any sugars (control). Lysates were then loaded on cellulose columns pre-washed with buffers containing the respective sugars or with buffer alone (control). The columns were then washed with the same buffers. Elutions were carried out with urea as described in the previous section. The proteins in the load and urea elutions from each column were separated by SDS-PAGE and blotted to nitrocellulose. Identical blots were reacted with MUD102 or MUD62 to test for the presence of PsB or SP96, respectively (see Table II).

RESULTS

Based on our earlier demonstration that the PsB complex localizes to the cellulose containing spore coat (Watson et al., 1994), we hypothesized that the complex may have endogenous cellulose binding activity and that this activity may play a role in spore maturation. We directly tested this idea by assaying for cellulose binding proteins in extracts of migrating slug-stage cells. We rationalized that the cellulose binding activity of proteins might be masked following secretion since the proteins would be already bound to cellulose. To avoid this difficulty, we used extracts from slug-stage cells when the proteins of the complex are sequestered in the prespore vesicles which do not contain cellulose (Erdos and West, 1989). To assay for cellulose binding proteins, an extract of slug-stage cells was prepared and chromatographed on a column of cellulose. After loading the extract, the column was washed with 0.5 M NaCl, to remove any loosely bound proteins, followed by further elutions with 6 M urea and 1% SDS. Each elution was concentrated by lyophilization so that protein samples representing equal cell equivalents could be compared by SDS-PAGE. The results of these experiments are shown in Fig. 1.

Coomassie Brilliant Blue staining of the fractions showed very few proteins which were retained on the cellulose. Four protein bands at 218, 70, 40, and 30 kDa were bound and eluted with urea. Further elution with 1% SDS released more of each of these proteins, although in later experiments complete elution could be accomplished with urea alone. Silver staining revealed additional minor bands. Interestingly the 218-kDa band visualized with Coomassie Brilliant Blue is only faintly visible with the silver stain.

Together these results confirmed that proteins do bind to cellulose, but they did not identify any of the proteins or answer the primary question of whether the PsB multiprotein complex has cellulose binding activity. To address these two issues, parallel lanes were blotted onto nitrocellulose for staining with antibodies. Staining with the MUD102 mAb demonstrated that PsB is bound to cellulose and is eluted with urea. Apparently, the concentration of the PsB protein is below that required for visualization with Coomassie Brilliant Blue (see above). Previous data indicate that all the PsB protein is in complex form (Watson et al., 1993), suggesting that the entire complex binds to cellulose. Staining with MUD62 supports this contention. The MUD62 mAb recognizes both p112/SP96 and SP75 via a common fucose containing epitope. p112/SP96 is part of the PsB complex while it resides in the PSVs but SP75 is only recruited to the complex following secretion (Watson, 1994). The reactivity with MUD62 in Fig. 1 showed that p112/SP96 is bound to cellulose while SP75 did not bind. These data indicate that the entire PsB complex is specifically bound to cellulose, as SP75 is not bound (see below for further confirmation that the entire PsB complex is bound).

PsB contains the MUD50 reactive O-linked oligosaccharide. We wished to determine if other proteins with this carbohydrate modification are also bound to cellulose. Studies on plants indicate that there are interactions between carbohydrates and cellulose (Varner, 1989). Therefore, we probed another Western blot with MUD50 and observed that only PsB and two other proteins of molecular masses 34 kDa and 40 kDa are bound to cellulose. These two lower molecular weight proteins are the sheathin proteins which are part of the ECM sheath that surrounds migrating slugs and have been shown to form a complex which colocalizes with cellulose (Zhou-Chou et al., 1995). We have shown that these bands are reactive with the anti-sheathin antibody MUD51 (data not shown). Importantly, all the other MUD50 reactive glycoproteins are not retained on the cellulose, again emphasizing the specificity of the binding.

Antibody JAB2 reacts with two related stalk-specific ECM components ST310 and ST430 (McRobbie et al., 1988; Wallace et al., 1984). Based on their localization to the cellulose-rich stalk, we thought that these proteins might also have in vitro cellulose binding activity. However, as is evident in Fig. 1, neither ST310 nor ST430 was bound to cellulose.

The discoidin proteins are a family of cytoplasmic proteins with carbohydrate-binding (lectin) activity that are expressed during aggregation but are extremely stable and persist throughout the remainder of development (Alexander et al., 1990, 1992). The discoidin proteins preferentially bind to galactose residues, but also bind to a lesser extent to other hexoses including glucose (Rosen et al., 1973). As expected, the discoidin proteins bind to cellulose in our assay. The discoidin proteins are most likely the 30-kDa band seen by Coomassie Brilliant Blue and silver staining as they constitute 1-2% of the cellular protein at this stage of development.

The Western analyses of PsB and SP96 described above suggested that the entire PsB complex binds to cellulose. We have directly demonstrated this by pulse labeling developing WS380B cells at hourly intervals throughout morphogenesis with [³⁵S]methionine. The labeled cells were harvested, lysed, and divided into three fractions for analysis. Part of each sample was fractionated directly by SDS-PAGE (Fig. 2, panel A) to show the entire spectrum of [³⁵S]methionine-labeled proteins at each developmental stage. As in previous work (Watson et al., 1993), there is a major increase in protein synthesis when the cells enter aggregates as reflected by [³⁵S]methionine incorporation. A second aliquot of each sample was immunoprecipitated with MUD102 and panel B shows that the entire PsB complex was being synthesized at the expected developmental times. A third aliquot of each sample was affinity chromatographed on cellulose. It is clear from panel C that the entire PsB complex (i.e. all the protein subunits) is bound to cellulose. In addition, recovery of the radiolabeled proteins is essentially complete after elution with urea (compare to the corresponding MUD102 immunoprecipitations in panel B). The data show the specificity of the binding and indicate that the complex has the cellulose binding activity while residing in the PSVs prior to secretion.

Having established that the PsB multiprotein complex did have cellulose binding activity, we investigated the nature of this binding. We tested the possibility that the interaction was lectin-like and could be abrogated with an exogenously added sugar hapten (Sharon, 1993). Fig. 3 shows that this is not the case. None of the mono- and disaccharides including cellobiose (the disaccharide prepared from cellulose) were able to block the binding of the PsB complex to cellulose. As shown by staining with MUD102 and MUD62, neither the binding of PsB nor p112/SP96 to cellulose was affected by the added sugars. Again, little SP75 is bound to the cellulose demonstrating the binding specificity of the PsB complex. In contrast, the binding of the discoidin lectin proteins was inhibited by its specific hapten galactose as well as cellobiose (data not shown). Consistent with these data, we never observed PsB complex binding to other polysaccharide matrices including Sepharose and Sephadex (Watson et al., 1993).

Our previous localization studies showed that the PsB protein was localized to the inner layer of the spore coat and that an associated protein, p112, was identical to an outer spore-coat protein called SP96 (Watson et al., 1994). The genes for SP96 as well as two other spore-coat proteins, SP70 and SP60, have been cloned and used to make a series of single, double and triple gene disruption mutants (Fosnaugh et al., 1994). The cotA, B-, and -C genes encode SP96, -70, and -60, respectively (Table I). We wished to determine if the cotB and -C genes encoded additional subunits of the PsB complex.

Each of the cot mutants and the wild-type controls (both WS380B and HL328, the parent of the cot mutants) were labeled for 1 h with [³⁵S]methionine beginning at the slug stage of development (morphology equivalent to 9 h in Fig. 2). Labeled cell lysates were prepared, and an aliquot was taken for SDS-PAGE to demonstrate that the level and extent of labeling in each strain was comparable (Fig. 4, Lanes 1-6). Two additional aliquots of each lysate were immunoprecipitated with either the MUD102 or MUD62 mAb which react with PsB and p112/SP96, respectively. The proteins in these immunoprecipitates were separated by SDS-PAGE. These results, shown in Fig. 4 (Lanes 7-12 and 13-18, respectively), reveal which of the PsB complex proteins remain associated with these two proteins in each of the mutant strains. A final aliquot of lysate from each strain was tested for cellulose binding activity in our in vitro assay and the bound proteins were analyzed by SDS-PAGE (Fig. 4, Lanes 19-24). The data from these experiments have been summarized in the model for the structure of the PsB complex presented in Fig. 5 which is helpful when reviewing the data in Fig. 4.

Immunoprecipitation of extracts of WS380B or HL328 with either MUD102 or MUD62 demonstrated the presence of the complete PsB multiprotein complex. Note that in strain HL328, PsB is a larger polymorphic form (Smith et al., 1989; Watson et al., 1993) and should be used for comparison to the cot mutants. The MUD102 immunoprecipitates showed the 4 major proteins of the complex (Fig. 4, Lanes 7 and 8). Immunoprecipitation with MUD62 showed the presence of an additional band (Lanes 13 and 14). This is SP75 which we have shown is highly phosphorylated and fucosylated. It is not part of the complex while it is stored in the PSVs, but becomes incorporated into the complex after secretion (Watson, 1994). This experiment also demonstrates that SP75 is polymorphic and that a smaller form exists in HL328. We have shown previously that SP96 exists in excess in the cells, i.e. it is the only subunit protein that is not completely incorporated into the PsB complex, and this explains why the SP96 and SP75 bands in the MUD62 immunoprecipitates appear more prominent than the other proteins in the complex.³ A protein we have called p200 appears to be associated with the complex in these experiments (note that this is not the p218 seen by Coomassie Brilliant Blue staining in the cellulose binding studies in Fig. 1). We are not sure at this time whether this is a bona fide subunit of the PsB complex as it is not present in all immunoprecipitations (Watson et al., 1993, 1994; see under ``Discussion''). The PsB complex from both control strains specifically binds to cellulose (Lanes 19 and 20).

Immunoprecipitation with MUD102 shows that a partial PsB complex lacking SP96 is made in this mutant strain (Fig. 4, Lane 9). This confirms our immunological data showing that SP96 is identical to p112 (Watson et al., 1994). The partial complex is not immunoprecipitated by MUD62 because of the lack of the SP96 subunit. MUD62 only immunoprecipitated SP75 by virtue of its fucose epitope and confirmed that this protein is not recruited to the complex prior to secretion (Lane 15). The partial complex from the cotA mutant does not have cellulose binding activity, implicating SP96 in this role (Lane 21).

Immunoprecipitation of extracts of the cotB mutant with MUD102 showed that the complex subunit p78 is identical to SP70 (Fig. 4, Lane 10). Moreover, the partial complex assembled in this mutant has only two proteins, PsB and the smallest subunit p58. This demonstrates that SP96 is attached directly to SP70 and that SP70 is attached in turn to either PsB or p58. Interestingly, immunoprecipitation with MUD62 showed that SP96 is synthesized and stable in this mutant strain, although it is not incorporated into the complex because of the lack of SP70 (Lane 16). The partial complex lacking both SP96 and SP70 is not capable of binding to cellulose (Lane 22). It is very significant that the free SP96 does not bind to cellulose, indicating that its assembly into the complex is required for this activity.

Immunoprecipitation with MUD102 or MUD62 of cotC mutant extracts demonstrates a partial complex containing SP96, PsB, and SP70 but lacking the smallest subunit (Fig. 4, Lanes 11 and 17). This identifies p58 as SP60. Moreover, this result shows that SP70 is directly bound to PsB. The partial complex lacking SP60 possesses full cellulose binding activity, indicating that SP60 is not part of the binding site (Lane 23).

Little PsB is present in the MUD102 immunoprecipitate of the lysate of the triple mutant, indicating that free PsB may not be stable in the absence of all the other subunits of the complex (Fig.4, Lane 12).

DISCUSSION

We have previously characterized a family of modB dependent O-glycosylated proteins which are reactive with the MUD50 monoclonal antibody. One of these, the PsB glycoprotein, is part of a preassembled multiprotein complex stored in the PSVs. The PsB complex is synchronously secreted at the end of development and becomes a central part of the spore coat (Watson et al., 1993, 1994). In this report we have addressed two fundamental issues, viz. the biochemical function of the complex and the assembly of its protein subunits.

The data show that the PsB complex binds specifically to cellulose. Moreover, the immunoprecipitation analysis of spore coat protein mutants showed that assembly of SP96 into the PsB complex is required for activity. Neither free SP96 nor a partial complex lacking SP96 show binding activity. Although the assembly of SP96 into the complex is clearly required for cellulose binding activity the precise binding site is not known. SP96 may acquire the correct conformation for binding activity once it is assembled into the complex. Alternatively, SP96 assembly into the complex may induce a conformational change in one or more of the other subunits which is required for the cellulose binding activity.

Our hypothesis that the PsB complex would have this biochemical activity was based on our earlier observations that it localized to the cellulose containing spore coats (Watson et al., 1994). We also predicted that we would be able to demonstrate binding because the PsB complex in slug cells resides in PSVs where it is not already bound to cellulose (Watson et al., 1994; West and Erdos, 1990). This may help explain why no specific in vitro cellulose binding proteins have been demonstrated for plants (Varner, 1989). If cellulose binding proteins do exist in plants, they may be rapidly and irreversibly incorporated into the cellulose containing cell walls and would not be available for in vitro binding assays. We also have been able to demonstrate that the MUD50 reactive sheathin proteins can bind to cellulose. Earlier microscopic studies showed that these proteins co-localized with cellulose in the slug sheath as footprints where the migrating slug made contact with the substratum (Zhou-Chou et al., 1995). This localization suggested that these proteins may be important in slug migration and is consistent with our earlier observation that slugs of modB mutant strains cannot migrate (Alexander et al., 1988). In addition, at least four other Coomassie Blue staining proteins bind to cellulose. One of these bands is the discoidin proteins, but we have not yet identified any of the other three or pursued further characterization. It is interesting to note that only the discoidin proteins were observed in the radiolabeling studies in Fig. 2. The other proteins may be synthesized slowly and therefore are not labeled strongly with [³⁵S]methionine.

The binding of the PsB complex to cellulose is specific. Unlike lectins, the binding is not inhibited by sugar haptens. Presumably, the complex recognizes a higher order structure of the cellulose and binds by hydrogen bonding (Watson et al., 1987). Moreover, binding is not affected by either 0.5 M NaCl or 0.01 M EDTA (data not shown). Protein denaturants including urea and SDS both release the complex by disrupting the tertiary structure of the proteins (Creighton, 1983).

Most previously demonstrated cases of cellulose binding proteins are limited to cellulases of bacteria and fungi. These cellulases are divided into three domains: a catalytic domain, a proline and threonine-rich spacer region (O-glycosylated in fungi), and a cellulose binding domain (CBD) (Gilkes et al., 1992; Stahlberg et al., 1988). Dictyostelium produces a cellulase with three domains similar to the fungal and bacterial enzymes (Ramalingam et al., 1992). Generally, proteins with CBDs have been thought to bind to cellulose as monomers, but recent experiments indicate that several CBD containing proteins may bind as dimers in Cellulomonas fimi (Guang-Yi et al., 1995) or as multiprotein complexes in Clostridium cellulolyticum and Clostridium cellulovorans (Gehin and Petitdemange, 1995; Shoseyov and Doi, 1990). In one of these multiprotein complexes, the protein with cellulose binding activity does not appear to have the catalytic function (Shoseyov and Doi, 1990; Shoseyov et al., 1992). Amino acid deletion and substitution analyses have identified regions and individual amino acids in CBDs that are important for the binding activity (Goldstein and Doi, 1994; Guang-Yi et al., 1995; Linder et al., 1995). The nature of the PsB complex CBD remains to be identified.

All celluloses are chemically identical in primary composition. They are synthesized as long 1,4-glucan chains which differ in their degree of polymerization, crystallinity, and other physical properties depending on the biological source (Saxena et al., 1994). The individual glucan chains associate to form microfibrils which are stabilized by hydrogen bonds and these polymers vary in structure depending on the organism in which they are synthesized (Okano et al., 1989). In Dictyostelium, the degree of crystallinity of cellulose varies depending on the developmental stage of the aggregates with crystallinity increasing as development proceeds (Freeze and Loomis, 1978). In addition, Dictyostelium cellulose may be modified with mannose residues (Freeze and Loomis, 1977a, 1978). These subtle modifications in cellulose structure could explain why the binding of the PsB complex to commercially available cellulose is not always complete.

Many proteins function as part of higher order multiprotein complexes, but it is rare that much is known about the assembly and structure of the complex (Hurtley and Helenius, 1989). Our analysis of the spore coat mutants directly addressed this issue and resulted in three major findings.

1) The relative order of the assembly of the proteins was defined and is illustrated in the model for the wild-type PsB complex in Fig. 5. Interestingly, in mutants lacking a protein subunit the remaining subunits are stable and assemble into partial complexes. It is not yet known whether the partial complexes are transported to the PSVs and secreted.

2) Several components of the PsB multiprotein complex have now been identified as spore coat proteins. Specifically, our p78 is the cotB gene product SP70 and our p58 is the cotC gene product SP60. Using immunological criteria we had previously identified the highly phosphorylated protein p112 as SP96 and the analysis of the cotA mutant confirms this (Watson et al., 1994). The work reported here, together with earlier studies (Devine et al., 1982; Watson et al., 1994), clearly shows that the SP70 subunit (cotB gene product) is distinct from the highly phosphorylated protein SP75 that is reactive with the MUD62 antibody (Champion et al., 1991; Watson et al., 1994). We have shown that SP75 is not part of the preassembled complex in the PSVs, but that it is added to the complex during or after secretion (Watson, 1994). Although SP70 is not reactive with the MUD62 antibody, we have demonstrated that it is fucosylated because its molecular size is altered in the modC and E mutants which are affected in a step in the addition of fucose oligosaccharides (Champion et al., 1991, 1995; Gonzalez-Yanes et al., 1989; Watson, 1994). This is in agreement with an earlier report that showed incorporation of [¹⁴C]fucose into SP70 (Devine et al., 1982). The available data do not indicate that the cotC gene product SP60 is either phosphorylated or glycosylated (Watson 1994; Watson et al., 1993). Attempts to identify the spiA protein, which is expressed late in spore differentiation (Richardson and Loomis, 1992), as a member of the PsB complex have been unsuccessful.

3) In the course of this and earlier work, we have sometimes observed a 200-kDa protein (p200 in Fig. 4) in immunoprecipitates of the PsB complex. Its appearance is inconsistent (e.g. it is present in Fig. 4 but not in Fig. 2) and may depend on the duration of the labeling period and/or genetic background. Work from another group suggested that a 200-kDa band extracted from spores was a nonspecific aggregate of other spore coat proteins (Gonzalez-Yanes et al., 1989). However, our evidence indicates that this is a unique protein: (a) p200 is recognized specifically by the monoclonal antibody MUD143 (data not shown); (b) none of the other antibodies to PsB, SP96, or SP75 react with p200 in Western analyses (Watson et al., 1993, 1994; Fig. 1); (c) the p200 band is not phosphorylated indicating that neither SP96 nor SP75 are components of it (Watson et al., 1994); and (d) attempts to disrupt p200 have been unsuccessful (data not shown). In addition, p200 does not appear to carry the modB or modC, D, E-dependent O-linked oligosaccharides as it is not reactive either with the MUD50 or MUD62 antibodies, respectively. We have not yet determined whether it is N-glycosylated.

These data significantly advance our understanding of the assembly and specific biochemical function of the PsB multiprotein complex during Dictyostelium development. However, several important developmental and cell biological questions remain to be answered. Nothing is known about the routing and localization of the PsB complex to the PSVs where they are stored awaiting the developmental signal to be secreted. Generally, protein oligomerization occurs in the endoplasmic reticulum (Hurtley and Helenius, 1989), suggesting that one of the PsB complex subunits acts as an address tag for the entire complex. Indeed, it is possible that the oligosaccharide modifications of the PsB complex proteins are involved since glycosylation has been shown to be involved in intracellular targeting in Dictyostelium as well as other systems (Ebert et al., 1989; Fiedler and Simons, 1995; Freeze et al., 1989; Matter and Mellman, 1994). In addition, nothing is known about the biochemical make up of the PSVs and especially about the PSV membrane which must be involved in the import of the protein cargo, the receipt of the signal for secretion and the eventual fusion of the PSVs with the plasma membrane during secretion. Our observation that the SP75 protein is not in the preassembled complex suggests that there may be more than one type of PSV and that components of the spore coat are sequestered from each other until secretion.

The nature of the molecular signal which induces the developmentally regulated secretion remains a central mystery. We have shown that the PsB multiprotein complex accumulates during the slug stage of development. We predict that secretion is coupled to the transition from slug migration to fruit construction which is under the control of several environmental factors (Newell et al., 1969; Sussman and Brackenbury, 1976). These environmental cues, in turn, appear to control the level of the morphogen cAMP which is known to regulate cell-type gene expression and cytodifferentiation (Loomis, 1988; Sussman and Schindler, 1978). The ability of cAMP to induce secretion is directly testable and may be extended to an in vitro system of isolated PSVs and plasma membranes. In addition, the wealth of mutants in cAMP binding proteins and G-proteins (Firtel, 1995) should allow us to dissect out the precise signal transduction pathway involved in the induction of secretion as we and others have done for various aspects of early gene expression (Blusch et al., 1995; Wu et al., 1995).

Overall, this work has demonstrated a specific biochemical activity for the PsB complex which correlates with its localization in the spore coat. In addition, the work has provided a unique view of the assembly of this multiprotein complex. The study opens up a number of challenging lines of investigation which further address issues central to both the genesis of the Dictyostelium spore coat as well as the developmental control of the assembly and organization of ECMs in general.