Uncoating of Coated Vesicles by Yeast hsp70 Proteins*

The ability of hsp7O isoenzymes from wild-type and mutant yeast strains to uncoat bovine brain coated vesicles was analyzed and compared with that of the brain uncoating ATPase. Results show that, among the four major cytoplasmic isoenzymes produced in wild-type yeast, almost all of the activity is associated with the SSAl and SSA2 isoenzymes. The SSBl and SSB2 isoenzymes have almost no uncoating activity and are not found in the clathrin-hsp70 complexes formed during the uncoating reaction. Using hsp70 mutant yeast strains we find a marked difference in uncoating activity between the SSAl and SSA2 isoenzymes, although there is only a 3% difference between their amino acid sequences. The SSA4 isoenzyme, which is produced only under stress conditions, has an uncoating activity intermediate between SSAl and SSA2. These results suggest that the ability of hsp70 isoenzymes to uncoat clathrin-coated vesicles is restricted to certain mem- bers of the hsp70 family and can be affected by subtle changes in amino acid sequence. We also investigated the uncoating activity of mixtures of isoenzymes and find that the isoenzyme with lower uncoating activity reduces the activity of the isoenzyme with higher un- coating activity possibly by occupying binding sites on coated vesicles.

It has recently become clear that the bovine brain uncoating ATPase, which strips clathrin off of bovine brain clathrincoated vesicles is a member of the hsp70 class of heat-shock proteins (Chappell et al., 1986;Ungewickell, 1985). This large class of proteins includes both proteins which are present constitutively in the cell and proteins which are induced during heat shock (for reviews, see Craig, 1985 andLindquist, 1986). All of these proteins bind ATP very tightly, and one of their distinguishing properties is their ability to be isolated in an almost pure state with the use of an ATP-agarose affinity column.
Thus far, several rather diverse functions have been found for the hsp7O proteins, in addition to their ability to dissociate clathrin from coated vesicles. The most primitive of the hsp7O proteins, the dnaK protein of Escherichia coli, appears to be involved in numerous cellular functions as well as in bacteriophage replication (Phillips and Silhavy, 1990;Tilly and Yarmolinsky, 1989;Zylicz et al., 1983). Yeast has at least eight different hsp7O proteins (Lindquist, 1986), two of which have been shown to be involved in translocation of proteins across the membranes of endoplasmic reticulum (Chirico et al., 1988) and mitochondria (Deshaies et al., 1988), respectively. In * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. mammalian cells, the hsp70 proteins appear to be involved in the uncoating process mentioned above, and, in addition, they appear to play a role in antigen presentation (Vanbuskirk et al., 1989) and nascent protein folding (Beckmann et al., 1990). Furthermore, BiP or GRP78, which is a member of this family located in the endoplasmic reticulum (Munro and Pelham, 1987), appears to be involved in complexing with aggregated proteins (Kozutsnmi et al., 1988) and also in assisting multichain proteins to fold into correct quaternary structures (Bole et al., 1986;Flynn et al., 1989). All of these functions have been described under the umbrella term molecular chaperone (Ellis, 1987;Hemmingsen et al., 1988) suggesting that these proteins, in general, either fold, unfold, or otherwise alter the conformation of proteins to which they bind. For example, the role of the hsp70 proteins in translocation is thought to involve unfolding proteins so that they assume the correct conformation when they are presented to the machinery which is involved in transporting them across the membrane.
Recently, Flynn et al. (1989) found that both BiP and the uncoating ATPase are able to interact with a wide variety of peptides. Although this fits with the molecular chaperone idea, it does raise a question as to whether the hsp70 proteins show significant specificity in their interactions with various substrates. In the present study we investigated the ability of the cytoplasmic hsp70 proteins from yeast to dissociate clathrin from bovine brain coated vesicles. Using yeast mutants to produce specific hsp7O proteins (Craig and Jacobsen, 1984;Craig and Jacobsen, 1985), we find that these proteins have very different abilities to dissociate clathrin from coated vesicles. The SSBl and SSB2 isoenzymes apparently have almost no uncoating activity, while a mixture of SSAl and SSAB isoenzymes shows considerable uncoating activity, but less than that exhibited by the bovine brain uncoating ATPase. Furthermore, even the SSAl and SSAB isoenzymes, which differ only in a few amino acids, show a marked difference in their uncoating abilities, with the SSA2 isoenzyme approaching the activity of the bovine brain uncoating ATPase. Our results also suggest that the isoenzyme with lower uncoating activity may inhibit the activity of the isoenzyme with higher uncoating activity in a mixture of different isoenzymes possibly by occupying the binding sites on clathrin-coated vesicles.

MATERIALS AND METHODS
Purification of Bovine Brain Coated Vesicles and Uncoating ATPme-Coated vesicles were prepared from calf brain according to the precedure of Nandi et al. (1982). Generally, coated vesicles containing 60 mg of clathrin were obtained from about 2.5 kg of brain tissue (10 calf brains). The coated vesicles were stored at 4 "C at a clathrin triskelion concentration of about 15 PM. Bovine brain uncoating ATPase was purified by the method of Schlossman et al. (1984) with slight modifications as described previously (Greene and Eisenberg, 1990).
Purification of Yeast h p 7 0 Proteins-The yeast strains used were developed in the laboratory of Dr. E. Craig at the University of Wisconsin. All steps in the yeast hsp7O protein preparation were carried out at 0-4 "C unless otherwise indicated. Cells were grown in YPD medium (1% yeast extract, 2% peptone, and 2% glucose) until the Am of the culture reached 6 at which point they were harvested by Millipore filtration. The cell paste was then dropped into liquid nitrogen and stored at -70 "C until use. In a typical preparation, 60 g of the cell paste were suspended in 180 ml of homogenization buffer (buffer A 40 mM imidazole, 75 mM KCl, 5 mM magnesium acetate, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 10 p~ leupeptin, 10 p~ pepstatin A, pH 7.0) and mixed with 240 g of glass beads (0.45 mm in diameter, from B. Braun Melsungen AG). The cells were broken using a Biospec bead-beater (model 909-1) with four to five cycles of 2-min beating and 10-min sitting. A Zeiss light microscope was used to determine that over 90% of the cells were broken. The cell homogenate was then centrifuged at 5,000 rpm for 10 min in a Sorvall SS34 rotor. The supernatant was centrifuged again at 34,000 rpm for 150 min in a Beckman Ti-50.2 rotor. The resulting supernatant (normally about 130 ml, with a protein concentration of 5 mg/ml) was then loaded onto a 5-ml ATP-agarose (Sigma, Cat. A2767) column equilibrated with elution buffer (buffer C 20 mM imidazole, 25 mM KC1, 10 mM ammonium sulfate, 2 mM magnesium acetate, 1 mM dithiothreitol, pH 7.0) at a flow rate of 1 ml/min. The column was then washed with 30 ml of buffer C, 40 ml of buffer C containing 1 M NaCl, and 30 ml of buffer C again, and finally, eluted with buffer C containing 1 mM ATP. The yeast hsp7O protein was precipitated by addition of ammonium sulfate to 70% saturation to the ATP-eluted fraction, and the precipitate was collected by a 20min centrifugation at 18,000 rpm in a Sorvall SS34 rotor. The protein was dialyzed against buffer C overnight before use. This preparation method yielded hsp70protein which was at least 90% pure determined by densitometric scanning of the SDS' gels. The hsp7O protein concentration was determined using its extinction coeffient = 4.5, molecular mass = 70 kDa) which had previously been determined by amino acid composition analysis.
Assay for Coated Vesicle Uncoating Activity-The uncoating activity of the yeast hsp70 protein was measured as described by Greene and Eisenberg (1990). Briefly, the coated vesicles were mixed with hsp7O protein in the presence of MgATP and incubated at 25 "C. The reaction mixture was then centrifuged in a Beckman TL-100 centrifuge to pellet the remaining coated vesicles. The supernatant, containing soluble hsp70 protein and clathrin released from coated vesicles, was analyzed by SDS gel. The Coomassie Blue-stained hsp70 protein and clathrin bands were quantified using an LKB Ultrascan XL laser densitometer.
FPLC Column Chromatography-FPLC was performed using a Pharmacia LKB Biotechnology Inc. Superose-6 column (HR 16/50) and a Pharmacia FPLC system with Gp-250 programmer, P-500 pumps, Single-Path UV-1 monitor, and Frac-100 fraction collecter. The column was equilibrated in buffer C, pH 7.0. For separating the upper and lower bands, 2 ml of purified hsp70 protein at a concentration of 15 PM were loaded onto the column. For separating clathrin and clathrin-hsp70 complexes from free enzyme in the uncoating Supernatant, 2 ml of the supernatant of a uncoating reaction mixure with 10 p~ hsp7O protein and coated veslcles containing 1 p~ clathrin triskelion were loaded onto the column. The column was run at a flow rate of 1.5 ml/min. Fractions of 1 ml were collected and analyzed by SDS-polyacrylamide gel electrophoresis.

RESULTS
Purification of Yeast hsp70 Protein-A high degree of purification of hsp7O proteins from yeast can be obtained simply by chromatography of crude lysates on an ATP-agarose column ( Fig. 1). After elution from the ATP-agarose column, we obtained a doublet at approximately 70 kDa with an upper to lower band ratio of 3 to 2 in stain intensity (lune 4 ) . Since we were purifying the hsp7O proteins from wild-type yeast, which contains a number of hsps, this doublet may be due to the presence of a mixture of isoenzymes in our final preparation. Comparison of the uncoating activity of the yeast hsp70 protein and brain uncoating ATPase. The uncoating of the brain coated vesicles by yeast hsp70 protein and brain uncoating ATPase was carried out in buffer C (see "Materials and Methods") with an ATP regenerating system (30 units/ml creatine phosphokinase and 15 mM phosphocreatine from Sigma) at 25 "C. The enzyme concentration for both yeast hsp70 protein and brain uncoating ATPase is 0.8 p~. The total coated vesicles used contained 0.5 p~ clathrin triskelions. The amount of clathrin in the uncoating supernatant was determined by densotometric scanning of SDS gels. 0, uncoating by yeast hsp70 protein; e, uncoating by bovine brain uncoating ATPase. The initial burst of uncoating was estimated by extrapolation of the linear steady-state rate (dashed lines).
For comparison the bovine brain uncoating ATPase is shown in lune 5. Densitometric scanning of the SDS gel showed that the yeast hsp70 protein is more than 90% pure. Uncoating Activity of the Yeast hsp70 Protein-We next tested the ability of the purified wild-type yeast hsp7O protein preparation to release clathrin from bovine brain coated vesicles. Using brain uncoating ATPase, we had previously found that there was an initial burst of uncoating followed by slow steady-state uncoating activity (Greene and Eisenberg, 1990). Fig. 2 shows a comparison of the time course of uncoating activity by the yeast hsp70 proteins prepared from wild-type yeast and the brain uncoating ATPase. Like the brain uncoating ATPase, the yeast protein showed an initial burst of uncoating followed by slow steady-state uncoating activity. However, the magnitude of the initial burst of uncoating and the steady-state rate of uncoating were only about 40 and 25%, respectively, of the values observed with the brain enzyme. Fig. 3 shows the magnitude of the initial burst of uncoating as a function of enzyme concentration with both wild-type yeast hsp7O protein and brain uncoating ATPase. As we found previously with brain uncoating ATPase, the initial burst of yeast hsp70 protein and brain uncoating ATPase. The initial burst of clathrin release determined by extrapolation as in Fig. 2 is plotted as a function of yeast hsp70 protein (0) and brain uncoating ATPase (0). Total coated vesicles used contained 0.2 p~ clathrin triskelions.
uncoating was stoichiometric with three enzyme molecules dissociating one clathrin molecule. However, with yeast hsp70 protein, the initial burst of uncoating was markedly decreased with considerably more enzyme required to dissociate the clathrin from the coated vesicles in the initial burst. Furthermore, even at a ratio of hsp7O enzyme uncoating ATPase to clathrin much higher than the stoichiometric binding ratio, the yeast hsp70 protein did not release clathrin from coated vesicles as extensively as did stoichiometric amounts of brain enzyme.

Separation of the Upper and Lower h p 7 0 Bands by FPLC-
One reason that the wild-type yeast hsp7O preparation was less effective than the brain enzyme may be that it is a mixture of various isoenzymes of yeast which have different levels of uncoating activity. We, therefore, attempted to determine the activity of the individual isoenzymes present in the wild-type yeast preparation. We began by attempting to separate the wild-type hsp7O proteins on a Superose-6 column using FPLC. Fig. 4 shows that the upper and lower bands observed on a one-dimensional SDS gel are partially separated by chromatography on FPLC. The upper band elutes first followed by the lower band.
We next tested the different fractions from the FPLC column for uncoating activity (Fig. 5, upper panel). Qualitatively, it is clear that much more clathrin was released by the earlier column fractions, which contained mostly the upper band, than by the later fractions. The amount of dissociated clathrin as a function of the protein in the upper and lower bands was quantified by densitometry (Fig. 5, lower panel). It is clear that almost all of the uncoating activity was associated with the upper band on the one-dimensional gel, but the lower band showed little or no uncoating activity.
To determine which yeast isoenzymes were associated with the upper and lower bands, we compared a two-dimensional gel of the wild-type enzyme with two-dimensional gels of the upper and lower bands separated by the FPLC column. Fig.  6a shows that the hsp7O that we originally isolated from wildtype yeast was mainly a mixture of SSAl and SSA2 isoenzymes and SSBl and SSB2 isoenzymes. These isoenzymes were originally identified by analysis of strains containing single mutations (Chappell et al., 1986;Craig and Jacobsen, 1984;Craig and Jacobsen, 1985). Fig. 6b shows that the upper band from the SDS gel was composed mainly of the SSAl and SSA2 isoenzymes, while the lower band (Fig. 6c) was composed mainly of the SSBl and SSB2 isoenzymes. Therefore, most of the uncoating activity observed with the wildtype yeast appeared to be associated with the SSAl and SSA2 and ; about 50-60% of the free clathrin is complexed with enzyme following separation of free enzyme from uncoated clathrin and its associated bound enzyme on an FPLC column (data not shown). If indeed the SSBl and SSBP isoenzymes have very little activity, it might be expected that, following the uncoating reaction with wild-type enzyme, considerably less of these isoenzymes would bind to clathrin than the SSAl and SSA2 isoenzymes. To test this, we used FPLC to separate the free enzyme from the uncoated clathrin and its associated bound enzyme. With the wild-type yeast enzyme the clathrin was found to be about 35% saturated with enzyme; somewhat less complex than was observed with the brain uncoating ATPase. Fig. 7a shows an FPLC chromatogram of the uncoating supernatant of wildtype hsp70 and Fig. 7b shows two-dimensional gels of wildtype hsp70 proteins and isoenzymes found associated with a. FPLC Chromatogram of Uncoating Supernatant Fraction Number

Two-Dimensional Gels of hsp70
Wild-Type hsp7O Bound hsp70  FIG. 7. Binding of yeast hsp70 isoenzymes to clathrin during the uncoating reaction. a, the supernatant from the uncoating reaction was chromatographed on an FPLC Superose-6 column to separate the clathrin-hsp70 complexes from free hsp7O protein. Three fractions at the clathrin peak (fraction numbers 38-40), containing released free clathrin triskelions and clathrin-hsp70 complexes, were pooled for two-dimensional gel electrophoresis. b, two-dimensional gels of the wild-type yeast hsp7O protein used in the uncoating reaction (left); and the hsp70 isoenzymes associated with the clathrin peak (right). clathrin during uncoating. As can be seen, the SSAl and SSA2 isoenzymes, but almost no SSBl and SSBP isoenzymes were complexed with clathrin. Interestingly, the two-dimensional gels in Fig. 7b suggest that a little more SSAP than SSAl isoenzyme was complexed with clathrin compared to the ratio found in the purified wild-type hsp70 protein. This suggests that the SSA2 isoenzyme may have a slightly higher activity than the SSAl isoenzyme.
Since it was possible that the SSB isoenzymes bound to the coated vesicles in a dead-end complex even though they were unable to uncoat clathrin from coated vesicles, we also analyzed the uncoating pellet which contains the coated vesicles which sediment following the uncoating reaction. We found that the clathrin was 30% saturated with SSAl and SSA2, but no SSBl and SSB2 isoenzymes bound to the pelleted coated vesicles (Fig. 8). These results are consistent with almost all of the uncoating activity observed with wild-type enzyme most likely being carried out by the SSAl and SSAP isoenzymes. Since the SSBl and SSB2 isoenzymes neither bind to the dissociated clathrin or to the coated vesicles ( Fig.  7b and 8), these results strongly suggest that the presence of SSBl and SSB2 isoenzymes has no effect on the uncoating reaction carried out by SSAl and SSA2.
Uncoating Activity of Different hp70 Isoenzymes-Since the SSBl and SSBP isoenzymes make up about 40% of the mixture of wild-type enzymes, the fact that they show almost no activity partially explains our observation that the wildtype enzyme was not as active as the brain uncoating ATPase. However, even taking this into consideration, the yeast enzyme still showed considerably less activity than the brain uncoating ATPase. Therefore, we wondered whether there was also a difference in the activity of the SSAl and SSAB isoenzymes. It might be expected that these two isoenzymes would have about the same activity since there is only a 3% difference in their amino acid sequences (Slater and Craig, 1989a). However, their regulatory sequences are quite different (Slater and Craig, 1989a) suggesting that they may have different functions in the yeast cell. Therefore, using yeast mutants which lack either the SSAl or SSAB isoenzyme we tested the activity of these individual isoenzymes. Fig. 9 shows the uncoating activity of the isoenzymes purified from ssal mutant (ssal-) or ssa2 mutant (ssa2-). For comparison we also show the activity of the wild-type enzyme and the brain uncoating ATPase. Note that the preparations of purified SSAl and SSAZ isoenzymes did contain the SSB isoenzymes, but based on the results presented above, we would expect that they have no effect on the observed uncoating activity. Therefore, the activity of the yeast enzymes are plotted as a function of the amount of upper band on onedimensional SDS gel, i.e. on the amount of SSAl and/or SSA2 isoenzyme present. Surprisingly, despite their similar amino acid sequences there was a marked difference in the activity of the SSAl and SSA2 isoenzymes. The activity of the SSAB isoenzyme approached that of the brain uncoating ATPase while the SSAl isoenzyme had less activity than the mixture of isoenzymes isolated from wild-type yeast. Furthermore, even at very high concentrations of enzyme it appeared that the SSAl isoenzyme did not dissociate as much clathrin from coated vesicles as the SSA2 isoenzyme.
Interestingly, these data suggest that the activity of the wild-type enzyme is not just the sum of the activities of the SSAl and SSAB isoenzymes. If this were the case, when the wild-type enzyme was twice the concentration of the SSAB isoenzyme, the activity of the wild-type enzyme should be equal to or greater than the activity of the SSA2 isoenzyme. This is because nearly equal amounts of SSAl and SSAB isoenzymes are present in preparations from wild-type yeast (Figs. 6a and 76). However, at high enzyme concentration where the activity of the all of the yeast enzymes levels off, the activity of the wild-type enzyme (solid circles, Fig. 9) was, in fact, significantly less than the activity of SSA2 isoenzyme (open triangles, Fig. 9) at half of the concentration. An identical result was obtained if a mixture of purified isoenzymes  ssa2-, A), and brain uncoating ATPase (W) was measured as in Fig. 3. The initial burst of uncoating by the mixture of equal amount of isoenzymes from ssaland ssa2-strains are indicated by 0. from ssal-and ssa2-strains was used (open circles, Fig. 9) instead of wild-type enzyme. This suggests that in addition to the SSAl isoenzyme having less activity than the SSA2 isoenzyme, it may actually inhibit the activity of the SSA2 isoenzyme. T o further investigate this phenomenon, we studied the effect of the yeast isoenzymes on the activity of the bovine brain uncoating ATPase. As can be seen in Table I, both with SSAl and SSAB, mixtures of the yeast isoenzymes with the uncoating ATPase show considerably less than additive uncoating activity. It is possible that this effect is due to a competition between the yeast isoenzymes and the uncoating ATPase for binding sites on the coated vesicles. The binding of the yeast enzymes may compete with the binding of the more active uncoating ATPase thus reducing the uncoating activity which occurs. A similar effect may also explain why the SSAl and SSA2 isoenzymes do not exhibit additive uncoating activity.
To assess more fully the differences in activity among the different hsp70 isoenzymes, we isolated hsp7O isoenzymes from different mutant strains. The two-dimensional gels in Fig. 10 show the hsp70 isoenzymes present in different yeast mutant strains, and Table   I1 shows the activity of these different preparations. In each case we have presented the activity per amount of upper band of the isoenzyme. The ssal-ssa2 double mutant (DS16) produces SSA4 isoenzyme (Craig and Jacobsen, 1984), which is not expressed in wild-type yeast

TABLE I
Effect of yeast isoenzymes on the initial burst of uncoating by brain uncoating ATPase Values presented are the amounts (in p~) of clathrin released during initial burst of uncoating (see Fig. 2). The enzyme concentrations refer to the concentrations of upper bands including brain uncoating ATPase, SSAZ in ssal-, and SSAl in ssa2-.   except under stress conditions. SSA4 isoenzyme, which is probably produced to compensate for the loss of the SSAl and SSAZ isoenzymes, migrates as an upper band on onedimensional gels like the SSAl and SSA2 isoenzymes. The uncoating activity of the SSA4 isoenzyme is intermediate between the SSAl and SSAZ isoenzymes. Interestingly, the mutant which is lacking the SSBl and SSBZ isoenzymes (T151) had significantly lower activity than the wild-type enzyme. The two-dimensional gel of the hsp7O proteins isolated from this yeast strain indicates that it had considerably more SSAl than SSAZ isoenzyme (Fig. lo), in contrast to the wild-type yeast which had approximately equal amounts of the two isoenzymes. This may account for the lower activity of the enzyme isolated from this strain of yeast compared to the wild-type enzyme.
In the present study we tested the ability of five of the six cytoplasmic proteins to uncoat clathrin-coated vesicles. Our results show that, while the SSA1, SSA2, and SSA4 isoenzymes are able to uncoat bovine brain coated vesicles, the SSBl and SSBZ isoenzymes, which are 65-75% similar in amino acid sequence to the SSA isoenzymes (Slater and Craig, 1989b), are unable to carry out significant uncoating. Furthermore, they do not appear to bind to clathrin dissociated from coated vesicles or to the coated vesicles themselves. Therefore, although our preparations of SSA1, SSA2, and SSA4 isoenzymes are contaminanted with the SSB isoenzymes, we consider it very unlikely that they are affecting the uncoating activity which we observe. Furthermore, since our two-dimensional gels suggest that SSCl and BiP are only present in very small amounts in our preparations and SSA3 is not expressed, we think it likely that the preparations of SSA1, SSA2, and SSA4 isoenzymes which we prepared from the mutant yeast strains are acting as relatively pure preparations of a single isoenzyme.
Not only is there a major difference in the uncoating activity of the SSA and SSB isoenzymes, but, in addition, the SSAl isoenzyme shows considerably less activity than the SSA2 isoenzyme although they differ in only a few amino acids. Comparing the amino acid sequences of SSAl and SSAS isoenzymes, Slater and Craig have shown that SSAl has four amino acid substitutions at the N-terminal end, which is close to the putative nucleotide binding site, and 11 amino acid substitutions in the rest of the molecule with a three-amino acid insertion at the C-terminal end, which is presumably where clathrin binds (Slater and Craig, 1989a). Furthermore, six of the 15 amino acid substitutions are conservative in nature.
It is remarkable that this small number of changes is apparently sufficient to alter the way SSAB binds clathrin and/or nucleotide and therefore lower the uncoating activity of SSAl compared to SSA2. Flynn et al. (1989) recently demonstrated that a large number of rather nonspecific peptides bind to the constitutive mammalian hsp70 protein which raises the question of how specifically the hsp70 proteins interact with their various substrates. The data presented in this paper suggest that, at least in regard to the uncoating reaction, the interaction appears to be quite specific. Not only did the various isoforms of the hsp70 proteins differ markedly in their ability to uncoat clathrin-coated vesicles but in addition subtle differences in amino acid sequence had a significant effect on uncoating activity. This is of particular interest, because up to the present time the most clearly defined activity of the constitutive mammalian hsp7O protein is its ability to strip clathrin off of coated vesicles. We conclude from these data that the ability to uncoat bovine brain coated vesicles is specific to certain members of the hsp7O family and depends on subtle differences in amino acid sequence. In addition, these data suggest that it is the SSA isoenzymes from yeast which are most similar to the bovine brain uncoating ATPase with the SSAB isoenzyme approaching the bovine brain uncoating ATPase in activity.
One question which arises from this work is whether SSAS, or any of the other yeast 70,000 heat-shock proteins, are actually involved in uncoating clathrin-coated vesicles in yeast. One reason that the SSAZ isoenzyme may not be as active as the brain enzyme is that we are not using yeast clathrin-coated vesicles as substrate. In fact, there is evidence that clathrin-coated vesicles occur in yeast and, if they occur, presumably at some point they must be uncoated. In this regard, it is of interest that SSAB isoenzyme is the only SSA protein which is entirely constitutive; presumably the uncoating of clathrin-coated vesicles is a constitutive process in yeast. However, at the present time, there is no evidence as to whether SSAS isoenzyme or the other yeast hsp70 proteins are involved in this process. The only defined function for the yeast hsp7O proteins is their role in the translocation of proteins into the endoplasmic reticulum (Chirico et al., 1988) and mitochondria (Deshaies et al., 1988). Given the relationship between the mammalian uncoating ATPase and the SSA proteins in yeast it will be of interest to determine if the mammalian protein can replace the yeast protein in this function in vivo.
One other interesting point noted in these studies is that, when mixtures of isoenzymes are used, the isoenzyme with lower activity can actually inhibit the activity of the isoenzyme with higher activity. This is true both for mixtures of SSAl with SSA2 and for mixtures of either SSAl or SSA2 with bovine brain uncoating ATPase. It may be related to the observation that, with the yeast enzymes, complete uncoating of the coated vesicles is not observed even at very high concentrations of enzyme. Although it is not clear why complete uncoating does not occur, it may be due to the fact that some of the coated vesicles are so stable that they cannot be uncoated by the less active enzyme during the initial burst of uncoating. However, if the less active enzyme binds to these stable coated vesicles and prevents the more active enzyme from uncoating them, it could explain why the presence of less active enzyme inhibits the activity of more active enzyme. We are currently studying the binding of the uncoating ATPase to clathrin under various conditions, and it is possible that these studies will shed light on whether a dead-end complex can, in fact, form between the yeast isoenzyme and the clathrin-coated vesicles.