Acid-activatable cysteine proteinases in the cellular slime mold Dictyostelium discoideum.

Studies of the cysteine proteinases of the cellular slime mold Dictyostelium discoideum have been aided by a simple acid treatment step that was incorporated into the standard one-dimensional gelatin-sodium dodecyl sulfate-polyacrylamide gel electrophoresis assay procedure. The step involved immersing the separating gel in 10% (v/v) glacial acetic acid for 30-60 s immediately after electrophoresis. This modified approach revealed the presence of acid-activatable forms of some enzymes with noticeable increases in their ability to hydrolyze gelatin, a substrate present in the sodium dodecyl sulfate-polyacrylamide gels, and peptidyl amidomethylcoumarins. The activation has been analyzed using extracts of dormant spores from which cysteine proteinase activity had previously appeared low or virtually absent. The major acid-activatable proteinase had an apparent molecular mass of 48 kDa. Its activation was not due to autocatalysis as it was not prevented by mercuric chloride, an inhibitor of the enzyme, and was not accompanied by a significant change in electrophoretic mobility. It was most likely due to a conformational change and/or the removal of a low molecular weight inhibitor. The acid treatment has also revealed the presence of acid-activatable cysteine proteinases in vegetative cells, in which cysteine proteinase activity is present at high levels, as well as among enzymes from the developmental cells which have much lower cysteine proteinase activity. Indeed novel developmental forms were detected at some stages. These results provide additional insight concerning cysteine proteinase expression at various stages during development in the slime molds. A developmental model is presented which suggests that the crypticity of the cysteine proteinases in dormant spores may be governed by proton pumps and endogenous lysosomotropic agents.

Studies of the cysteine proteinases of the cellular slime mold Dictyostelium discoideum have been aided by a simple acid treatment step that was incorporated into the standard one-dimensional gelatin-sodium dodecyl sulfate-polyacrylamide gel electrophoresis assay procedure. The step involved immersing the separating gel in 10% (v/v) glacial acetic acid for 30 -60 s immediately after electrophoresis. This modified approach revealed the presence of acid-activatable forms of some enzymes with noticeable increases in their ability to hydrolyze gelatin, a substrate present in the sodium dodecyl sulfate-polyacrylamide gels, and peptidyl amidomethylcoumarins. The activation has been analyzed using extracts of dormant spores from which cysteine proteinase activity had previously appeared low or virtually absent. The major acid-activatable proteinase had an apparent molecular mass of 48 kDa. Its activation was not due to autocatalysis as it was not prevented by mercuric chloride, an inhibitor of the enzyme, and was not accompanied by a significant change in electrophoretic mobility. It was most likely due to a conformational change and/or the removal of a low molecular weight inhibitor. The acid treatment has also revealed the presence of acidactivatable cysteine proteinases in vegetative cells, in which cysteine proteinase activity is present at high levels, as well as among enzymes from the developmental cells which have much lower cysteine proteinase activity. Indeed novel developmental forms were detected at some stages. These results provide additional insight concerning cysteine proteinase expression at various stages during development in the slime molds. A developmental model is presented which suggests that the crypticity of the cysteine proteinases in dormant spores may be governed by proton pumps and endogenous lysosomotropic agents.
Dictyostelium discoideum possesses multiple forms of cysteine proteinase (ddCP) (1) which have been shown to differ from one another during growth (2) and development (3). Net cysteine proteinase activity is high in vegetative cells but decreases during the developmental phase (3), partly as a result of secretion (3,4). Consequently, both the stalk cells and the dormant spores that are formed during development possess very low levels of detectable activity; however, some cysteine proteinase activity is found in the extracellular matrix surrounding spores in fruiting bodies (5). Net intracellular cysteine proteinase activity returns to appreciable levels during spore germination (5). In addition to the cysteine proteinases, D. discoideum also produces an aspartic proteinase (ddAP58) whose level remains relatively constant throughout the life cycle.
There are three developmentally regulated genes present in D. discoideum whose sequences indicate that they encode cysteine proteinases. Predicted products of the cprA (CP1) and cprB (CP2) genes have considerable similarity to members of the papain superfamily (6,7), whereas the product of the third gene (CP3) may be a truncated cysteine proteinase (8). Intriguingly, all three genes are transcribed at times when net cysteine proteinase activity is decreasing, and hence their transcription appears inconsistent with activity levels of the cysteine proteinases detected during various stages of development (1). To date, attempts to show a meaningful relationship between the transcription of the developmentally regulated cysteine proteinase genes and the activity levels of the cysteine proteinases recorded during development have been unsuccessful. However, the possibility that gene products may have been present as inactive forms was not examined in detail.
Studies of many lysosomal enzymes, including cysteine proteinases, have indicated that the mechanism involved in the production of a mature and active conformation requires the cleavage of a pro-sequence. This cleavage, with mammalian cathepsin L (9 -12) and cathepsin B (13) for example, may be brought about in vitro at acid pH, suggesting that the lower pH of the lysosome may be responsible for triggering activation in vivo. In D. discoideum, comparisons of preprocessed and mature forms of various lysosomal enzymes have indicated molecular weight differences that could be clearly resolved using SDS-PAGE 1 techniques (14). It has also been reported that "increasing the lysosomal/endosomal pH from 5.4 to 6.4 with ammonium chloride does not prevent proper sorting of lysosomal hydrolases, but does prevent the cleavage of the lysosomally localized intermediate forms of the enzymes" (15). Thus, it appears plausible that in vivo, the optimal activity of the lysosomal cysteine proteinases might require an acidic pH.
We will describe the conditions required for an in vitro acid activation of the cysteine proteinases in D. discoideum. The results of this work provide information that may relate to the transcriptional and translational events governing the developmentally regulated cysteine proteinase genes in the cellular slime molds and support a proposed model whereby the crypticity of the cysteine proteinases in dormant spores is controlled by proton pumps and endogenous lysosomotropic agents (16).

Methods
Organisms Used and Sample Preparation-Experiments were performed using D. discoideum strains SG1 (ATCC 44840) and, for axenically grown myxamoebae, AX2 (ATCC 24397). Strain SG1 was grown in the presence of Escherichia coli for spore production (5) and in association with Klebsiella aerogenes when myxamoebae and developmental cells were required. Axenic cells of strain AX2 were grown on HL5 medium (2). Extracts from dormant spores as well as from the late stages of development were prepared by vortexing with glass beads (5). Spore casings were obtained from dormant spores by centrifuging at 4,500 ϫ g for 2 min to pellet the glass beads and any spores that remained intact, then centrifuging the supernatant for 2 min at 15,000 ϫ g in an Eppendorf microcentrifuge and resuspending the pelleted spore casings in ice-cold H 2 O. Cells from all other stages were harvested and lysed with 0.1% (v/v) Triton X-100 (4) and then either stored frozen at Ϫ20°C or used immediately in experiments. Freezing had no apparent effect on the levels of proteinase activity observed.
Electrophoresis and Proteinase Detection-Proteinases were detected following their electrophoretic separation using the gelatin-SDS-PAGE assay system, which consisted primarily of 10% (w/v) acrylamide minigels (0.75 mm) containing 0.2% (w/v) gelatin, Tris-HCl buffer, pH 8.8, and 0.1% (w/v) SDS (3,17). For acid activation, the standard treatment involved immersing the separating gel in 10% (v/v) acetic acid for 30 -60 s immediately after the electrophoretic separation was complete and prior to its subsequent washing with 2.5% (v/v) Triton X-100 for 30 min. Proteinase bands were then developed by incubating the gel at 25°C for 16 h in acetic acid/sodium acetate buffer, pH 4.0, containing 1 mM dithiothreitol. Alternatively, proteinase bands were detected with fluorogenic peptidyl amidomethylcoumarins using a modification of the method described previously (17). Unless otherwise indicated, gelatin was not included in these gels. Following electrophoresis, gels were transferred to 0.1 M sodium phosphate buffer, pH 6.0, containing 20 mM cysteine, either immediately or after a 30-s incubation in 10% (v/v) acetic acid. Gels were incubated for 10 min with one change of buffer and were then transferred to the same buffer containing 0.05 mM substrate. Bands of proteinase activity were recorded as described (17) during a 60-min incubation at room temperature.
When a second electrophoretic separation was required, samples were first run under standard conditions using narrow lanes (15-well comb in the Bio-Rad Mini Protean system). One method involved loading two identical samples in adjacent lanes in the center of the gel and, after electrophoresis, the gels were cut between these two lanes. One section of the gel was treated with 10% (v/v) acetic acid as above. The sections were then washed for 1 min five times in resolving gel buffer (0.375 M Tris HCl, pH 8.8) containing 0.1% (w/v) SDS and then placed between glass plates for the Mini Protean system so that the original sample lanes ran along the top edge of the gel sections from left (top of the original lane) to right (bottom). Gaps were then filled with fresh separating gel mix and electrophoresis continued until the marker dye, added to fresh electrophoresis running buffer, had reached the bottom of the gel. Activity was then detected using the fluorogenic substrate Boc-Val-Leu-Lys-NMec as described above. After activity had been detected the gel was stained with Coomassie Blue R-250. An alternative method, used with either gelatin or fluorogenic substrates, involved the following. Samples were run in narrow lanes, as above. After electrophoresis, strips the width of a single lane were cut from the gel, and these were treated, if required, with 10% (v/v) acetic acid for 30 s. Gelatin gels were then incubated with 2.5% (v/v) Triton X-100 for 30 min, and all gels were finally immersed in electrophoresis running buffer (0.25 M Tris, 0.192 M glycine, 0.1% (w/v) SDS). Two gel strips were laid horizontally on the top of a new gel so that they covered the gel, and electrophoresis carried out until marker dye had reached the bottom of the gel. Activity was detected as described above.

RESULTS
Acid Activation of Spore Proteinases-The standard one-dimensional gelatin-SDS-PAGE technique used for analyzing the proteinases present in D. discoideum (3) involves electrophoretic separation of the enzymes in a sample containing 1% (w/v) SDS and 2.5% (v/v) mercaptoethanol, followed by a wash with 2.5% (v/v) Triton X-100, and then incubation of the gel in an appropriate buffer system, usually acetate buffer, pH 4.0, containing a reducing agent such as 1 mM dithiothreitol or 20 mM cysteine. During an analysis of proteinases present in extracts from dormant spores, we found that if this procedure was interrupted by treating the gels briefly in 10% (v/v) acetic acid before washing with Triton X-100, greater activity levels of some proteinase bands were observed. Without this acid treatment step, the major detectable band of activity corresponded to the aspartic proteinase ddAP58 (5); minor bands apparently corresponding to cysteine proteinases ddCP48 and ddCP43 were occasionally observed using a 16-h incubation time, depending on the amount of protein loaded. After acid treatment the ddAP58 band was unchanged, but the activity of the cysteine proteinases, especially that corresponding to ddCP48, was greatly enhanced (Fig. 1). Without acid activation, high levels of ddCP48 activity only appeared when heat-activated spores were germinated (5) and rarely if germination occurred by autoactivation. All of the activatable proteinases were present at higher levels in extracts from dormant spores than in samples of spore casings (Fig. 1). Some extracellular matrix proteinases (5), with apparent molecular masses in the 20 -30-kDa range, were also activated by the acid treatment. However, the major enzyme detected in the absence of acid treatment (ddCP18 with an apparent molecular mass of 18 kDa) had reduced activity after acid treatment (Fig. 1).
The activation was due to an in vitro effect on the enzyme and not on the gelatin substrate. When gelatin-containing gels were pretreated with acetic acid, the subsequent band pattern was identical to that obtained with nontreated gelatin gels. In addition, pretreatment of gelatin stock solutions with 10% (v/v) acetic acid, whether in the presence or absence of SDS, did not affect the observed results (data not shown).
Confirmation that the effect of the acid was on the enzyme rather than substrate was obtained by using peptidyl amidomethylcoumarins. These were not added until after the acid and subsequent washing steps were completed. Detection of proteinases using these fluorogenic substrates did not necessarily require treatment of the gels with Triton X-100, and gels could simply be washed in phosphate buffer, pH 6.0, before addition of substrate. After acid treatment the same 48-kDa proteinase detected with gelatin as substrate could be seen using Boc-Val-Leu-Lys-NMec (Fig. 2). The presence of gelatin inhibited the activity toward the fluorogenic substrate, an effect that has also been noted with cysteine proteinases from other stages of D. discoideum. 2 For this reason gelatin was routinely omitted from the gels whenever fluorogenic substrates were used. For the 48-kDa proteinase, Boc-Val-Leu-Lys-NMec was the best fluorogenic substrate tested, and the enzyme also hydrolyzed Z-Phe-Arg-NMec well. The enzyme had lower activity toward H-Leu-Val-Tyr-NMec, Bz-Phe-Val-Arg-NMec, and Suc-Leu-Tyr-NMec. No activity was detectable with Z-Arg-Arg-NMec. In its preference for substrates with bulky amino acids at the Pro 3 and Pro 2 positions and inability to hydrolyze Z-Arg-Arg-NMec the enzyme resembled mammalian cathepsins L and S rather than cathepsin B (18,19).
Boc-Val-Leu-Lys-NMec was adopted as the preferred low molecular weight substrate for the detection of acid-activatable spore proteinases. The activity of the spore enzymes toward this substrate was not totally dependent on a reducing agent (not shown), but since a reducing agent was required for the detection of cysteine proteinases at other stages of the D. discoideum life cycle and cysteine was found to be most appropriate, the latter was included routinely at a concentration of 20 mM.
Conditions Required for Activation of the Spore Cysteine Proteinase-The activation of the 48-kDa proteinase required a weak acid with an incubation in 10% (v/v) (1.7 M) glacial acetic acid for 30 -60 s being optimal. Neither 1.7 M HCl nor 10% (v/v) formic acid, which are both stronger acids than acetic acid, was effective, but 10% (v/v) propionic acid, also a weak acid, did activate the enzymes. Sodium acetate (1.7 M) had no effect. Activation could also be achieved using 0.1 M glycine buffer, pH 2.0, although a longer incubation step was required (2 min). The standard length of the acetic acid treatment was 30 -60 s, but even a 15-s incubation period proved sufficient for activation. However, if the treatment was prolonged for more than 60 s less activity was observed, presumably because of subsequent denaturation of the activated enzyme.
The activation of cysteine proteinases at low pH is not unique to enzymes of D. discoideum, although with one exception (20) it has not been demonstrated in the manner described here. In most other cases activation has been shown to result from limited proteolysis of inactive pro-enzymes, and an intramolecular autocatalytic conversion has been implicated in studies of recombinant enzymes (13,21). It has been shown that mercuric ions, which inhibit cysteine proteinase activity, block both the activation of a latent cathepsin L (12) and the processing of the recombinant papain precursor (21). To test whether an autocatalytic mechanism might be involved in the activation process described here, gels were incubated with HgCl 2 immediately before treatment with acetic acid. HgCl 2 was confirmed to be an inhibitor of the activated 48-kDa spore enzyme, an effect reversed by the inclusion of 20 mM cysteine in the gel incubation buffer (not shown). If gels were treated with HgCl 2 before transfer to acid, activation still occurred (Fig. 3); samples pretreated with HgCl 2 (for example, lanes 3 and 7) showed the same degree of activation as the equivalent untreated samples (lanes 4 and 8). Identical results were obtained whether or not electrophoresis was carried out in the presence of a reducing agent, which might have affected the ability of HgCl 2 to inhibit the enzyme. Thus, in spite of the presence of an inhibitor of the enzyme at the time of acid treatment, the enzyme was still activated. This strongly suggested that an autocatalytic processing step was not involved.
Attempts to activate proteinases prior to electrophoresis by either preparing or washing the samples with a low pH buffer have not proven successful. Incubation in 0.1 M glycine HCl buffer, pH 2.0, alone failed to activate the major spore proteinase (data not shown). Because of a likely requirement for a detergent in the activation process, SDS was added to the buffer, but no activity at all was detected after this treatment. It is possible that under these conditions other proteolytic enzymes inactivated the cysteine proteinases.
Acid Activation Did Not Affect Electrophoretic Mobility-To investigate further the possible mechanism of activation, the 2 M. J. North and K. Nicol, unpublished data.

FIG. 2. Demonstration of acid-activatable spore cysteine proteinase with Boc-Val-Leu-Lys-NMec as substrate.
Samples of spore extract (containing 10 g of protein) were run on two gels, one containing no protein substrate (lanes 1 and 2) and one that contained 0.2% (w/v) gelatin (lanes 3 and 4). Control lanes (lanes 1 and 3) were cut from the gels and transferred immediately to 0.1 M phosphate buffer,   3. Effect of mercuric chloride on acid activation. Samples of spore extract (containing 200 g of protein) were layered onto two gels. For panel A, 2-mercaptoethanol was omitted from the sample buffer. After electrophoresis the gels were cut into strips and treated as follows. Strips 1-4 were incubated in buffer lacking cysteine; strips 5-8 were incubated in buffer ϩ 20 mM cysteine. Strips 3, 4, 7, and 8 were treated with 10% (v/v) acetic acid for 30 s before incubation in buffer. Strips 1, 3, 5, and 7 were incubated in 1 mM HgCl 2 for 5 min before acid treatment or direct transfer to buffer. All strips were eventually incubated with Boc-Val-Leu-Lys-NMec, with or without cysteine, as indicated above, for 24 min before the bands were photographed. electrophoretic mobility of the activated proteinase was examined. Since activation has only been achieved with proteinases in gels, two-dimensional electrophoresis was used. Various procedures were tested to minimize a loss of activity of the acidtreated enzyme during the second run, which posed a major problem. However, whenever activity was detectable after the second run, regardless of the procedure used (see "Methods"), the mobility of the major acid-activatable proteinase was not apparently affected by the acid treatment. For example, Fig. 4 shows the result of using large blocks of the first gel for the second run and Boc-Val-Leu-Lys-NMec as the substrate. After the first electrophoresis run one gel block was acid treated, the other not, but the mobility of the proteinase during the second run was the same in both. The lack of effect of acid treatment was confirmed by the location of the activity band on the diagonal line of Coomassie Blue-stained proteins. Any protein whose mobility had changed between the two runs would lie above or below this line.
Acid Activation of Proteinases Present at Other Stages-Vegetative myxamoebae have much higher cysteine proteinase activity than spores (1,3,5). Acid treatment did have some effect on the myxamoebal proteinase pattern, although only certain proteinases were affected. With Boc-Val-Leu-Lys-NMec as substrate two proteinases, ddCP48 in bacterially grown cells and ddCP51 in axenic cells were detectable only after acid treatment. The other proteinases did not require acid treatment (Fig. 5A). With gelatin-SDS-PAGE, however, it was not always possible to distinguish any effect of acid treatment (Fig.  5B). The difference between the effects observed with fluorogenic substrates and gelatin is likely to be due to the time required for detecting activity (less than 1 h and 16 h, respectively) and may reflect the fact that the vegetative proteinases, unlike the spore proteinase, are activated slowly during incubation at pH 4, the pH of the buffer used to detect bands in gelatin gels.
As development proceeds, so cysteine proteinase activity declines. This has been shown using a number of different detection methods, including the standard gelatin-SDS-PAGE technique (1, 3). During the aggregation stage acid treatment enhanced the activity of some of the proteinases, but no new proteinases were detected. Following aggregation, acid treatment revealed the presence of a number of activatable enzymes. These were not always readily detectable using gelatin-SDS-PAGE but could be seen clearly with Boc-Val-Leu-Lys-NMec as substrate (Fig. 6). A proteinase with an apparent molecular mass of 60 kDa was detected only after acid treatment, whereas the activity of another, with an apparent molecular mass of 24 kDa, was enhanced significantly by acid treatment. The level of these two enzymes, together with ddCP18, increased after aggregation (from 12 h onward). All were confirmed as cysteine proteinases by their sensitivity to E-64. Some of the proteinases detected at this stage were probably associated with the extracellular matrix and slime sheath. For example, the 24-kDa enzyme appears to be one of those present in the spore matrix (see Fig. 1). However, preliminary results indicate that this is not true of the 60-kDa enzyme, which was was not removed from slugs by washing and appeared to be intracellular. DISCUSSION During the final stages of sporulation in the life cycle of eukaryotic microorganisms, it seems reasonable to assume that the activity of lysosomal enzymes such as proteinases should become minimal, consistent with the dormant state of the organisms. In addition to the normal reductions of transcriptional and translational events during sporulation (22), there FIG. 4. Electrophoretic mobility of the major acid-activated proteinase in spore samples detected with Boc-Val-Leu-Lys-NMec. Samples of spore extract (10 g of protein) were run in two central lanes of a gel. After electrophoresis the gel was cut into two blocks that were arranged for a second electrophoresis run at 90°C to the first as described under "Experimental Procedures." Block A was treated with 10% (v/v) acetic acid for 30 s after the first run; block B was treated with acid after the second run. Activity was detected with Boc-Val-Leu-Lys-NMec and the photograph taken after 30 min. Above blocks A and B is a lane in which a sample of spore extract was run at the edge of the first gel to show the position of the acid-activated band after the first run (photograph taken after 10-min incubation with Boc-Val-Leu-Lys-NMec). Blocks C and D are blocks A and B, respectively, stained with Coomassie Blue. The position of the activity seen in blocks A and B is outlined. are several other strategies that an organism might utilize to minimize lysosomal enzyme activity, including the secretion of enzymes, the synthesis of inhibitors, or alterations in the conformational state of the enzymes. Conversely, during the germination of dormant structures, increases in the activity of the lysosomal enzymes could result from de novo synthesis, the loss of inhibitors from preexisting enzymes, or conformational changes in preexisting enzymes.
The results described here show that to detect some of the Dictyostelium cysteine proteinases using substrate-SDS-PAGE it was necessary to include a brief acid treatment step. The changes in cysteine proteinases during the different stages of the life cycle detected previously by gelatin-SDS-PAGE, i.e. without acid treatment, were consistent with the changes in net activity detected in cell extracts using cysteine proteinasespecific substrates such as peptide nitroanilides (5, 23). Thus it seems likely that the activatable enzymes were initially in an inactive form in the samples analyzed. We conclude that the lack of activity of these enzymes was not an artifact resulting simply from the procedures involved in the electrophoretic analysis and is therefore of physiological significance. The cysteine proteinases in D. discoideum may thus be synthesized and packaged in lysosomes as relatively inactive forms. It is not sufficient to free the enzymes from the spore lysosomes and incubate them at their optimal pH with protein or peptide derivatives to demonstrate enzymatic activity.
Mammalian lysosomal cysteine proteinases are synthesized as pro-enzymes containing an amino-terminal pro-region consisting of 62 amino acids, in the case of cathepsin B, or approximately 100 amino acids in the case of other cysteine proteinases such as cathepsins H, L, and S. The enzymes are activated by removal of the pro-region shortly after leaving the Golgi apparatus and entering the lysosome. Although there is some evidence suggesting the involvement of other proteolytic enzymes, for example metalloproteinases and aspartic proteinases (24,25), it is more likely that an autocatalytic process is involved (see Ref. 13). Typically, a 35-40-kDa inactive proenzyme is converted to a 22-33-kDa active enzyme that may then be further proteolytically processed. The cysteine proteinases of D. discoideum for which sequence information is available (6 -8, 26) are, like the aforementioned cathepsins, members of the papain superfamily (peptidase family C1 (27)), and it seems likely that all of them are synthesized as precursors containing a typical pro-region. A possible explanation for activation, therefore, might have been the acid-induced removal of the pro-peptide, as has been observed with cysteine protein-ases from mammalian cells (9 -12), tick eggs (28), and Leishmania mexicana mexicana (20). However, this would have resulted in a detectable shift in electrophoretic mobility, and this was not apparent for the 48-kDa spore enzyme or any of the other bands detectable on gels (Fig. 4). A second possibility was that the pro-enzyme was activated without the removal of the pro-region; the apparent molecular mass of the activated 48-kDa enzyme was sufficiently large to have included the proregion. However, sequence data for two of the active enzymes produced by vegetative cells of D. discoideum suggest that this is unlikely. ddCP38 from bacterially grown cells and ddCP42 from axenic cells (respective molecular masses of approximately 38 and 40 kDa as determined by SDS-PAGE), are both larger than cysteine proteinases found in other organisms but have amino-terminal sequences that align closely with those of other mature cysteine proteinases. 3 Thus D. discoideum cysteine proteinases can be large, even without a pro-region at the amino terminus, and this is most likely due to inserts within the mature sequence and glycosylation (26). It is therefore probable that the activatable enzymes had lost the pro-peptide prior to being treated with acid.
The resolution achieved with the two-dimensional gel systems may not have allowed the detection of any change in mobility resulting from a slight reduction in molecular weight, and so the possibility that a small peptide was removed cannot be entirely discounted. However, in view of the lack of effect on activation of the cysteine proteinase inhibitor HgCl 2 , it seems more likely that the activation was due to a conformational change or the removal of a small, tightly bound inhibitor. The only cysteine proteinase inhibitor reported in D. discoideum is too large (14 kDa) (29) to have had any involvement in acid activation as its removal from any proteinase would have affected electrophoretic mobility.
There are precedents for proteinase activation involving conformational changes; indeed it has been suggested that a change in conformation occurs before bond cleavage which removes the pro-region of the cathepsin L precursor (9). Another group of proteinases, the matrix metalloproteinases, are synthesized in a latent form that is inactive because of the formation of an intramolecular complex between the single cysteine residue in its pro-peptide domain and the essential zinc atom in the catalytic domain. A variety of activators including conformational perturbants, reversible sulfhydryl group modifiers, FIG. 6. Appearance of acid-activatable proteinase during the developmental phase. SG1 cells were harvested, allowed to develop on nitrocellulose filters, samples collected, and extracts prepared as described under "Experimental Procedures." Cells were not washed. Non-gelatin gels were used, and each lane was loaded with 10 g of protein. Proteinase activity was recorded after a 13-min incubation with Boc-Val-Leu-Lys-NMec. Panel A, the gel was not treated before transfer to Triton X-100. Panel B, the gel was treated with 10% (v/v) acetic acid before transfer to Triton X-100. The apparent molecular masses (in kDa) of the major proteinases, indicated on the left, are those determined previously (Ref. 3) and were confirmed by reference to standard markers. The positions of two new acid-activated proteinases are indicated by arrows. and irreversible sulfhydryl group modifiers trigger the socalled "cysteine switch" (30). Activation is achieved through dissociation of the cysteine from the complex. It is possible that the activation described here might be achieved through a similar mechanism but in this case involving the dissociation of an ionic bond.
This in vitro activation technique worked not only for spore cysteine proteinases but was able to enhance the activity from cells at other developmental stages, although to different degrees. Thus, it is possible that many of the cysteine proteinases in D. discoideum can exist in at least two activity states, active and inactive, which can be altered by an acid shock. The extent to which acid treatment is needed to detect activity in vitro may reflect, in part at least, the degree of exposure to low pH already experienced by individual proteinases within the cells.
It has become clear that the mature lysosomal enzymes of D. discoideum become acidified with the aid of specific and discrete proton pump vacuoles (31)(32)(33). Our hypothesis for the production of dormant spore cryptic cysteine proteinases involves the synthesis of cysteine proteinases and their packaging into lysosomal vesicles that do not make immediate contact with the proton pump vacuoles. Although this would prevent the activation of new cysteine proteinases, it might not be sufficient to deactivate the existing enzymes completely. It is possible that the synthesis of a lysosomotropic agent such as ammonia, which is produced in large quantities during development (34), or glutamine, which rises from 1.5 mM in amoebae to 72.8 mM in spores (35), shocks the cysteine proteinases into the inactive form by serving as a sink for any protons previously present in the lysosome (16). Indeed it has now been found that treatment of the acid-activated proteinases with ammonia deactivates the enzymes, which can subsequently be reactivated with a second acid treatment. 4 We also suggest that the dormant spore lysosome should have a relatively high pH, unlike the low pH of the functioning lysosomes in vegetative cells. During the spore germination process the fusion of the organelle with a functioning proton pump vacuole would provide a sufficient drop in pH to convert the inactive forms of the cysteine proteinases into their active conformations.
Further research is required to determine the validity of our hypothesis and to ascertain whether acid-activatable proteinases are inactive within the cell and indeed if their activation has physiological significance. Interestingly, it was the developmental proteinases, including the spore enzymes, which were much less active without the incorporation of the acid treatment step in the analysis. Additional and more complex scenarios for the process of reducing lysosomal enzyme activities in vivo may involve: 1) the appearance of a developmentspecific inhibitor; 2) a difference in proteinase location, which affects conformation or inhibitor attachment; or 3) the synthesis of a new set of developmental cysteine proteinases that bind more tightly to an inhibitor present throughout the life cycle or which require special conditions to gain an active conformation. Additional research is also required to determine whether the acid activation phenomenon occurs with other enzymes and developmental systems. Such work is essential to understanding the specific mechanism(s) involved and their importance in regulating development. Thus far, preliminary results from this work involving acid-activatable forms of the cysteine proteinases in D. discoideum suggest that at least some of the proteinases revealed by this treatment during development correspond to the developmentally regulated cprA and cprB gene products. It is not yet known whether the 48-kDa acidactivated spore proteinase is one such or if this is identical to the vegetative enzyme ddCP48. Purification of the activatable proteinases is now in progress to establish their relationship to the predicted products of these genes, to investigate the activation mechanism in more detail, and to determine the extent to which it relates to the control of proteolysis during the development and germination of the dormant structures in the cellular slime molds.