Proteomics opens doors to the mechanisms of developmentally regulated secretion

. The program of multicellular development in Dictyostelium discoideum culminates with the assembly of a rugged, environmentally resistant spore coat around each spore cell. After synthesis, the proteins that will constitute the coat are stored in prespore vesicles (PSVs) until an unknown developmental signal triggers the PSVs to move to the cell surface where they fuse with the plasma membrane and secrete their cargo by exocytosis. These events occur synchronously in 80% of the cells in each developing multicellular aggregate, and thus the system offers a unique opportunity to study the developmental regulation of protein secretion in situ . Proteomic analysis of purified PSVs identified many of the constituent proteins, which in turn has lead to novel hypotheses and new experimental avenues regarding the molecular mechanisms regulating secretion from the PSVs.

Dictyostelium discoideum cells synchronously secrete proteins from unique vesicles during development.
The eucaryotic cellular slime mold D. discoideum provides a unique opportunity to study the mechanisms regulating a specific secretory event that is triggered by the developmental program (5). D. discoideum cells divide mitotically as long as there is a source of nutrients. The cells resemble mammalian cells, having a simple plasma membrane and typical organelles.
When the cells are depleted of nutrients, starvation initiates the onset of a complex and now well described developmental program leading to the generation of environmentally resistant spores The Dictyostelium discoideum spore coat is the end product of regulated secretion.
Terminally differentiated spores are surrounded by a rugged polarized tri-laminar spore coat which is composed of approximately equal parts of cellulose and glycoproteins (10). The outer layer is electron dense and comprised of loosely associated proteins that can be removed by cold sodium dodecyl sulfate. The middle layer is cellulose, and the inner layer is made up of proteins that are covalently coupled and require heat and denaturing agents for extraction (11,12). A galactose/N-acetyl-galactosamine-containing polysaccharide (GPS) is also part of the spore coat, residing proximal to the plasma membrane of the cell. Biochemical studies of isolated spore coats have identified about 10 abundant proteins and several minor proteins (13,14). (A complete list of spore coat proteins, corresponding antibodies and references can be found at http://www.biosci.missouri.edu/alexander/). Four of the most abundant spore coat glycoproteins -SP90, PsB/SP85, SP70 and SP60exist as a preassembled multiprotein complex (PsB complex) that is held together by a combination of covalent and noncovalent bonds (15,16). These spore coat proteins are synthesized during aggregation. They are localized exclusively to the prespore cells during the slug stage. During terminal differentiation and spore formation, these proteins, along with other less well-characterized proteins, are deposited into the spore coat by a process of regulated secretion. A series of spore coat protein deletion mutants was used to define the order of assembly of these proteins into the complex (17). Immunostaining has shown that SP96 resides only in the outer layer of the coat, while PsB/SP85 is located only in the inner layer. Thus, the PsB complex is incorporated into the coat with a specific polarity and spans the central cellulose layer. Although it is not known how this polarity is achieved, it is important to note that the PsB by guest on May 6, 2020 complex has an endogenous cellulose binding activity, which is necessary for proper spore coat assembly (17)(18)(19)(20). The spores represent an important evolutionary advantage for this organism allowing the cells to remain viable under extreme environmental conditions including drought, heat and radiation.
The prespore vesicles and regulated secretion of the spore coat proteins.
The prespore cells are defined by the presence of unique secretory vesicles -the prespore vesicles (PSVs) -which appear de novo at the time of aggregation (21). They were originally characterized as carrying a cargo of GPS, which was known to become part of the spore coat. It was later shown that some of the major spore coat proteins, including the PsB multiprotein complex, were also cargo within the PSVs. The origin of the PSVs is a mystery. They appear concomitantly with the appearance of GPS and the spore coat proteins, which are coordinately expressed (15,16,22,23). The proteins are packaged into the PSVs soon after their synthesis ( Figure 2). At terminal differentiation, an undefined developmental signal induces the PSVs to migrate to the cell periphery, where they fuse with the plasma membrane and secrete their cargo to the extracellular space where it assembles into the spore coat ( Figure 2). When aggregates of cells containing PSVs are made to enter an alternative developmental sequence, where the aggregates become elongated motile "slugs", they do not complete spore differentiation and morphogenesis (24). The PSVs in these cells never secrete unless the aggregates are induced to re-enter the normal developmental sequence. When this happens, secretion from the PSVs is rapidly and coordinately initiated, coinciding with spore formation. Thus, the induction of the secretory process is a true signal-mediated event. Essentially nothing is known about the by guest on May 6, 2020 molecular mechanisms that underlie this precise spatial and temporal fusion of the PSVs with the plasma membrane, and the coordination of secretion with terminal differentiation.
The questions and a proteomic strategy. Several significant questions regarding secretion from the PSVs needed answering: 1) What is the genesis of the PSVs?; 2) What is the routing signal(s) used by PSV cargo proteins?; 3) What is the developmental signal that activates the secretion process?; 4) What is the mechanism of fusion between the PSVs and the plasma membrane? and 5) What are the biochemical mechanisms of spore coat protein assembly? Answering these questions required a biochemical approach because there was no reasonably convenient way to screen or select for mutations that affect secretion late in development. An axiom of biochemistry has been that it was necessary to define the components of a reaction or system in order to begin to understand how it functioned. Proteomics is merely a larger version of this strategy. Thus, we set out to purify the PSVs in order to identify and study their constituent proteins.
Purified PSVs begin to provide the answers.
We developed a purification scheme aimed at providing purified PSVs for biochemical analysis. (25,26). We used enzymatic and immunological assays to show that the PSVs were highly purified and devoid of markers for other organelles. Immuno-electron microscopic analysis showed that purified vesicles had the same morphology and contents of PSVs observed in cells. Moreover, examination of the PSVs purified from cells at different times throughout development showed that they increased in size, but not number, while continuing to take up more of the spore coat protein cargo. by guest on May 6, 2020 One dimensional SDS-PAGE separation of the proteins in the purified PSVs showed that the PSVs contained many more proteins than just the contents of the future spore coat. The purified PSV preparations were immediately useful for testing some ideas. Based on work on regulated secretion in yeast, we hypothesized that the PSVs would contain small GTP-binding proteins that regulate fusion. γ 32 GTP-binding assays showed that PSVs indeed contained a PSVspecific developmentally regulated GTP-binding protein, although its identity was still unknown.
In addition, we asked if PSVs contained clatharin, which is often associated with endo-and exocytotoic vesicles. We showed that the PSVs did not contain clatharin although the cells contained it in abundance. This "one protein at a time" approach was severely limited by our preconceived ideas regarding mechanism and by the availability of assays and antibody reagents (25). Clearly, we needed a more global approach to identify all the proteins in the PSVs so that we could begin to understand the process of developmentally regulated secretion in this system.
The ability to obtain biochemical quantities of pure PSVs allowed us to use a proteomic approach.
The proteomic solution.
The purified PSV preparations were solubilized and separated by two-dimensional polyacrylamide gel electrophoresis (2D PAGE). Approximately 100 proteins were separated.
Multiple PSV preparations produced reproducible separations, further confirming the specificity of the purification scheme. Eighty of the most abundant protein spots were excised from the 2D gel, digested with trypsin and the peptides were analyzed by MALDI-TOF spectrometry. These analyses provided identification of more than 50% of the proteins in the PSVs (27; includes a full by guest on May 6, 2020 list and accession numbers). The D. discoideum genome is not yet completely sequenced, and we anticipate the identification of the rest of the PSV proteins upon completion of the sequencing project. Except for the spore coat proteins none were known previously to be associated with the PSVs. The identified proteins fell into several functional groups including: 1. Spore coat proteins -SP96, SP87, SP60. These are the known cargo proteins of the prespore vesicle which ultimately assemble into the spore coat following secretion. Although these proteins are heavily O-glycosylated (28), they were identifiable.
2. Pumps/ATPases -F1 ATPase, P-type ATPase. Proton pumps have been shown to be associated with other secretory vesicles. The acidification of the lumen of these vesicles results in the concentration of their cargo. A similar role is expected for the PSV associated pumps.
3. Calcium binding proteins -Calreticulin and calfumirin. Calcium has previously been associated with the regulation of vesicle traffic protein secretion (29), but assigning functions to specific Ca 2+ binding proteins has been difficult because of the large number of such proteins and because they often have multiple functions. Specifically associating these proteins with the PSVs allows targeted studies regarding their role in this organelle. Interestingly, calreticulin has been previously associated with the endoplasmic reticulum and recently shown by proteomics to function in the phagosome as well (30,31). 4. Protein disulfide isomerases -Thioredoxin 2 and Thioredoxin 3. It had been hypothesized that disulfide isomerases were involved in cross-linking the proteins during spore coat maturation and that they might be delivered to the extracellular space via the PSVs (10,32). The role of these enzymes in cross-linking spore coat proteins can now be examined in strains lacking these enzymes. The genes encoding these proteins are also expressed early in development, and it is not currently known whether the thioredoxins in PSVs are synthesized de by guest on May 6, 2020 novo later in development. In yeast, it was shown that thioredoxin associates with SNARE complexes forming the LMA1 protein complex (33,34). Thus, the thioredoxins found in the PSVs may play a role in vesicle docking at the plasma membrane as well. 5. cAMP regulated prespore specific proteins -PB74, DG17, and D7. These proteins are expressed exclusively in prespore cells and their cognate genes are regulated by cAMP (35)(36)(37).
They have frequently been used as markers for cell differentiation, but there has not been any previous suggestion as to their function in spore cell differentiation.
6. Cytoskeletal elements -actin, coactosin, profilin and BiP. The identification of these proteins suggests an actin based mechanism for transporting the PSVs to the plasma membrane for fusion.
7. Esterases -crystal protein and D2. Both proteins have been shown to be associated in a complex with esterosomes, and eventually localize to the spore coat (38,39). Esterase activity has also been shown to be involved in the lysis of the spore coat during germination (40). We predict that these enzymes are assembled into the spore coat after secretion. The crystal protein has also been shown to associate with the actin cytoskeleton (41), and may therefore function in anchoring the PSVs to the cytoskeleton for transport to the cell surface.
8. Regulatory proteins -PI3 kinase, NDP kinase, and Rab7. Rab GTPases are known to regulate the rate and timing of vesicle traffic (42). The absolute level of GTP-bound Rab associated with secretory vesicles, rather than the rate of GTP hydrolysis per se, dictates fusion competence (43,44). The Rab7 GTPase was previously shown to function in endosomal traffic in mitotically dividing D. discoideum cells (45). NDP kinase is the GTP-binding protein that was shown earlier to be associated with the purified PSVs (25). It uses ATP to generate other nucleoside triphosphates including GTP, is localized at sites of increased GTP concentration and it thought by guest on May 6, 2020 to be the major source of GTP in the cell (46). Interestingly, NDP kinase is also expressed early in development, and a mechanism must exist for importing it to the PSVs. The localization and activity of NDP kinase may control the GTP-bound state of Rab7, which in turn may regulate the fusion of the PSVs with the plasma membrane.
9. Other proteins. An important outcome of the PSV proteome analysis was the identification of proteins for which there are corresponding ESTs, but no known function. Although we do not know the specific biochemical functions of these proteins, we do know that they are an integral part of the PSVs and therefore function in some step of regulated secretion. The homology of these gene products with those of other organisms suggests a similar function in those species as well. It will be especially exciting to elucidate the function of these novel proteins.
Lessons learned.
The proteomic analysis of the PSVs resulted in a number of important observations that bear on any study of an isolated organelle. 1) The PSV proteome was specific. A limited and reproducible set of proteins was identified in the analysis, indicating that these proteins are intimately associated with this organelle. 2) Interestingly, several of the PSV proteins previously had been implicated in other cellular/developmental functions, such as those described for Rab7 and NDP kinase. This suggests that proteins can have multiple roles in cells at different times in development, and that function can be dictated by specific subcellular localization (although how this changes during development remains to be elucidated). This observation also has been made in other systems. Analysis of phagosomes revealed the presence of endoplasmic reticulum proteins and lead to the novel discovery that the ER is a source of phagosome membrane (30,31). 3) This study has suggested functional roles for cell-type by guest on May 6, 2020 specific proteins of previously unknown function. Many genes in D. discoideum are activated only in the developing prespore cell population, and have long been used as markers for differentiation. Although much is known about the regulation of these genes, their function in prespore cells has remained unknown. The proteomics study shows that some of the cognate proteins have a direct role in PSV structure and/or function. 4) One of the PSV proteins was homologous to a protein of unknown function in Arabidopsis. Thus we suggest that the Arabidopsis protein will have a function in protein secretion. This demonstrates that a proteomic analysis in one organism can help understand a protein's function in another organism. 5) Perhaps most importantly, the identification of specific PSV associated proteins readily suggests hypotheses that can be tested to determine the precise mechanism of developmental control of PSV biogenesis, signal perception, vesicle movement and fusion with the plasma membrane.
The next step -back to molecular genetics.
The most obvious initial approach is to use homologous recombination to make gene disruptions in each of the identified PSV proteins and monitor the effect of these mutation on spore differentiation. Presumably, at least some of these would block spore maturation but it would still be unclear what mechanistic event was affected.
A novel strategy is to make a general tester strain in which to systematically interrogate the effects of various null mutations. This strain would express one of the spore coat genes (e.g., the gene encoding SP60) fused with the gene encoding green fluorescent protein (GFP) under the control of its own promoter. (The loss of SP60 has no discernable effect on the growth and development of D. discoideum.) The GFP-SP60 would localize to -and thus define -the PSVs. by guest on May 6, 2020 Single and double null mutations of the genes of interest can then be produced in this genetic background, and fluorescent localization of the GFP-SP60 can be used to assess in living cells if the biogenesis, maturation and fusion of the PSVs are disrupted in the mutants. This will allow the association of each of the cognate proteins with a specific step of the pathway of developmentally regulated secretion from the PSVs (Figure 3). The molecular genetic tools available for use with D. discoideum make this approach eminently feasible.
An interesting example of going from protein identification to function is the case of NDP kinase and Rab7. NDP kinase is the major source for GTP in the cell. Its biochemistry has been extensively studied in D. discoideum as well as other systems (47)(48)(49)(50)(51). Since it is associated with the PSVs in a developmentally regulated manner, its continuous expression and localization may be required to generate GTP to drive vesicle fusion. The role of NDP-kinase can be studied by replacing the wild-type copy of the gene in the GFP-SP60 tester strain with a tetracycline (tet R ) inducible copy (52). This allows the expression of NDP kinase to be turned on and off at different times during differentiation, and to follow its role in controlling PSV secretion by examining GFP-SP60 localization.
The interaction of NDP-kinase with Rab7 can be probed further using crosslinking and co-immunoprecipitation approaches that have been used successfully in other proteomic applications as well (53,54). We propose that NDP kinase may be the proximal source of GTP for the Rab7, which suggests that it plays a major role in regulating the onset of the secretory event. As such we believe that the two proteins may be physically associated.

Downloaded from
Lastly, we can test directly whether NDP kinase supplies GTP to Rab7, by incubating prespore cells with γ 32 P-ATP, and sampling throughout late development. The GTP generated by NDP kinase will be labeled with the γ 32 P from ATP. The association of this γ 32 P-GTP with Rab7 will be detectable by immunoprecipitating Rab7. No γ 32 P-GTP should be associated with Rab7 if the NDP kinase is switched off. This will provide direct evidence that NDP kinase generates GTP for Rab7 activity.
The foregoing discussion shows that the global discovery-based nature of proteomics is having a major impact on the progress of elucidating fundamental mechanisms in cell biology     Null mutants (knockouts) will be produced by homologous recombination in a standard tester strain expressing a GFP-SP60 fusion. This will allow assessment of the secretory process by monitoring the fate of the fluorescent PSVs throughout development. The predicted classes of by guest on May 6, 2020 phenotypes include A) PSV genesis is blocked; B) PSVs are made but do not move to plasma membrane during development; C) PSVs move to plasma membrane but do not fuse; and D) Fusion and exocytosis occur but spore coat proteins cannot assemble into a mature spore coat. by guest on May 6, 2020