A Comparison of the GroE Chaperonin Requirements for Sequentially and Structurally Homologous Malate Dehydrogenases: The importance of folding kinetics and solution environment .

: Escherichia coli malate dehydrogenase (EcMDH) and its eukaryotic counterpart, porcine mitochondrial malate dehydrogenase (PmMDH) are highly homologous proteins with significant sequence identity (60%) and virtually identical, native structural folds. Despite this homology, EcMDH folds rapidly and efficiently in vitro and does not seem to interact with GroE chaperonins at physiological temperatures (37 o C), while PmMDH folds much slower than EcMDH and requires these chaperonins to fold to the native state at 37 o C. Double-jump experiments indicate that the slow folding behavior of PmMDH is not limited by proline isomerization. Although the folding enhancer glycerol (<5M) does not alter the renaturation kinetics of EcMDH, it dramatically accelerates the spontaneous renaturation of PmMDH at all temperatures tested. Kinetic analysis of PmMDH renaturation with increasing glycerol concentrations suggests that this osmolyte increases the on-pathway kinetics of the monomer folding to assembly competent forms. Other osmolytes such as TMAO, sucrose, and betaine also reactivate PmMDH at non-permissive temperatures (37 o C). Glycerol-jump experiments with preformed GroEL-PmMDH complexes indicate that the shift between stringent (requires ATP and GroES) and relaxed (only requires ATP) complex conformations are rapid (< 3-5 sec). The similarity in irreversible misfolding kinetics of PmMDH measured with glycerol or the activated chaperonin complex (GroEL-GroES-ATP) suggests that these folding aids may influence the same step in the PmMDH folding reaction. Moreover, the interactions between glycerol-induced PmMDH folding intermediates and GroEL-GroES-ATP are diminished. Our results support the notion that the protein folding kinetics of sequentially and structurally homologous proteins, rather than the structural fold, dictate the GroE chaperonin requirement. folding intermediates populations and chaperonins (9,10). The folding intermediate structures of homologous proteins that require chaperonins to fold are predicted to have longer lifetimes and/or slower on-pathway folding rates. In addition, our studies indicate that some osmolytes will profoundly influence the chaperonin requirements and that osmolytes certainly may influence the folding kinetics and stabilization of on-pathway protein folding intermediates.

Introduction mitochondrial isozymes does not hold, because rhodanese, a highly stringent mitochondrial chaperonin substrate, has a slightly acid pI. In addition, the potential chaperonin substrate proteins identified from high affinity chaperonin-protein complexes isolated from E. coli cell extracts have a wide range of global pI values (7). Furthermore, the role that specific folding motifs play in dictating interactions between mammalian DHFR and the chaperonin has been recently questioned (8).
Rather than identifying correlations between chaperonin stringency and primary sequences or native tertiary and quaternary structures of homologous proteins, it appears that the folding kinetics and the lifetime of the folding intermediates (9,10) may be more reasonable properties that ultimately define the chaperonin requirements. In more recent work, Clark and Frieden provide strong evidence that kinetic factors may indeed explain the differences in chaperonin requirements for the homologous E. coli and mammalian DHFR isozymes (11,12). Specifically, an intermediate population of mammalian DHFR, existing as a partially folded species, can interact with the chaperonin before refolding to the native conformation (11,12). Unfortunately, these studies have only examined the interaction of the homologous DHFR substrates with the nucleotide-free highaffinity form of GroEL, a species that may have very transient lifetimes in vivo. In addition, the recent observation that the activated GroE species (i.e. GroEL, GroES and ATP) may actively unfold bound substrates suggests that the high affinity form of GroEL represents an oversimplification of chaperonin-substrate interaction (13). Thus, if we are to understand the dynamic nature of the interactions between chaperonins and their protein substrates, we must examine systems where E. coli substrates and their homologous mitochondria counterparts interact with more physiologically relevant, activated chaperonin species.
In this current work, we have compared the folding of a highly homologous pair of malate dehydrogenase enzymes, Escherichia coli (EcMDH) and porcine mitochondria (PmMDH), in the absence and presence of the GroE chaperonins. The correlative studies of Hartl and coworkers indicate that intrinsic GroEL polypeptide substrates contain predominantly αβ tertiary folds containing two or more domains. PmMDH has this same type of fold and has been suggested to be a model in vitro substrate of the bacterial GroE chaperonins (7). The homologous PmMDH and EcMDH proteins in our study fold within cellular environments that contain group I chaperonins (GroEL/GroES in E. coli and Hsp60/Hsp10 in mitochondria). In addition, these two isozymes are more highly related to each other than any other homologous proteins used to study chaperonin interactions.
EcMDH are highly similar with respect to sequence and structure (~60% sequence identity, ~80% sequence similarity), we find no evidence that EcMDH binds to or even interacts with the activated chaperonin complex (GroEL-GroES-ATP). Since the renaturation rate and probably the folding speed of EcMDH are more rapid, an increase in the folding of PmMDH should also diminish the chaperonin requirement.
Numerous investigators have found that including polyols in the refolding buffer can lead to increased protein refolding rates (15,16). We have found that the inclusion of the osmolyte, glycerol, in the refolding buffer leads to an enormous increase in both yield and reactivation rate of PmMDH at all temperatures tested.
Under these conditions, the initial interactions between the refolding PmMDH subunits and the activated GroE chaperonin complex are no longer detectable. These studies suggest that the folding kinetics of the protein, rather than its particular structural fold, dictate the observed variability in chaperonin interactions with structurally homologous isozymes.

Methods and Materials The GroE Molecular Chaperonins, GroEL and GroES
The Escherichia coli chaperonins, GroEL and GroES, were isolated from the lysate of cells containing the appropriate overexpression plasmid (gifts from Dr. Edward Eisenstein and Dr. George Lorimer, respectively). GroEL and GroES were purified as described by Voziyan and Fisher (17) and Eisenstein et al. (18). Since GroEL and GroES do not contain tryptophan residues, the removal of tryptophan containing contaminants, as assayed by second derivative analysis of the absorption spectra and tryptophan fluorescence were used as a criterion for purity of the chaperonin preparations as well as by silver stained SDSpolyacrylamide gel electrophoresis (19).

Malate Dehydrogenase (MDH)
Porcine mitochondrial malate dehydrogenase (PmMDH) and Escherichia coli malate dehydrogenase (EcMDH) were purchased from Sigma Chemical Company. The purity of each protein was examined by SDSpolyacrylamide gel electrophoresis followed by a silver staining procedure to resolve the protein bands. No foreign protein bands for PmMDH or EcMDH were observed. Under these conditions, both isozymes were completely unfolded as accessed by the complete loss of activity.

Denaturation and Renaturation of MDH
Complete unfolding of secondary and tertiary structures were confirmed by both tyrosine fluorescence and near UV circular dichroism spectroscopy (data not shown). Renaturation of a small aliquot of concentrated MDH was initiated by a rapid 100-fold dilution of standard buffer alone or with various combinations of 1 µM GroEL, 2 µM GroES, 5 mM ATP, and different concentrations of glycerol. The final MDH subunit concentration was 0.1 µM, unless otherwise stated. ATP was added to the refolding solution containing the GroE chaperonins 30 -60 seconds before the initiation of protein folding.

MDH Enzymatic Activity Assay
The enzymatic activity of MDH was determined at 37 o C using a substrate analog, ketomalonic acid (1 mM) and 0.2 mM NADH under standard buffer conditions, and following the rate of oxidation of NADH at 340nm on a SLM Aminco 3000 spectrophotometer (21,22). The use of ketomalonic acid as a competitive substrate for MDH eliminates the use of the natural substrate oxaloacetate, which undergoes significant decarboxylation to pyruvate at room temperature (22). The final MDH subunit concentration in the assay mixture was between 0.05 and 0.09 µM. At these MDH concentrations, for both EcMDH and PmMDH, the absorbance decline of NADH was linear within the time range of the data acquisition (1-3 min).

90° Light Scattering
A SLM 8000S fluorescence spectrophotometer with both the excitation and the emission wavelengths set at 360 nm was used to measure the formation of protein aggregates during the spontaneous renaturation of PmMDH with or without 35% glycerol in standard buffer conditions at 37°C. The bandpass for the emission monochromator was set at 5 nm. The final PmMDH subunit concentration in these experiments was 1 µM.

Proline Isomerization "Double Jump" Experiment
The "double jump" experiment was designed to detect the effect of proline isomerization as it pertains to the renaturation kinetics of proteins (23,24). The term, "double jump", refers to the change of environment of a protein during the transition (i) from native to denaturing solution conditions, and (ii) from denaturing back to native solution conditions. Native PmMDH was denatured at 20°C for various times from 30 seconds to 6 hours before a rapid dilution with standard buffer at 20°C was performed to initiate spontaneous folding. In these experiments, PmMDH was completely unfolded after 30 sec as determined by tyrosine fluorescence. The enzymatic activity of PmMDH and EcMDH was used to monitor protein renaturation.

Commitment to the Native State
The commitment experiments were originally designed to examine the efficiency of the release of protein substrates in either a native state or committed to fold to the native state from chaperonin-substrate complexes (25). This methodology was used to determine whether PmMDH folding intermediates could interact with various components of the GroE chaperonin-mediated protein folding mechanism (i.e. GroEL, GroES, and ATP) in the presence and absence of 35% glycerol. In the presence and absence of 35% glycerol, PmMDH renaturation was initiated either spontaneously, with GroEL alone, with GroEL and ATP, or with GroEL, GroES, and ATP at 37°C in standard buffer conditions. Polyclonal anti-GroEL antibodies were added to aliquots of renaturing protein at 15 seconds or nearly 2 hours after the initiation of protein refolding. The final concentration of solution components was as follows: PmMDH monomers, 0.1 µM; GroEL, 1 µM; GroES, 2 µM; ATP, 5 mM; anti-GroEL antibodies, 50 µM. After the addition of anti-GroEL, the immunoprecipitants were rapidly removed by spinning the mixture in a microcentrifuge for 20-30 seconds. The supernatant was extracted, incubated at 37°C, and analyzed for enzymatic activity 2 hours after the initiation of refolding.

Irreversible Misfolding Kinetics
To measure irreversible misfolding rates, either the nucleotide-activated chaperonin complex (GroEL-GroES-ATP) or 35% glycerol were added to samples of refolding PmMDH at different times ranging from 0 to 60 seconds after spontaneous folding was initiated in standard buffer at 37°C according to the protocol outlined in (33). After the folding reaction took place for 4 -6 hours to reach maximum refolding yields, the enzymatic activity of each sample was determined. The final concentration of the solution components was as follows: initiated, the glycerol concentration was rapidly changed to either 35% or 3.5%, respectively, with and without ATP. The final concentration of the solution components was as follows: PmMDH monomers, 0.1 µM; GroEL, 1 µM; ATP, 5 mM. The renaturation of dimeric PmMDH was monitored by the recovery of enzymatic activity.

Curve Fitting
The renaturation data for PmMDH and EcMDH were fit to two different models. PmMDH renaturation was modeled from the folding reaction scheme of PmMDH as determined by Ranson

Renaturation of PmMDH
The rate constants associated with M → U, N → M, and Agg → I agg were assumed to be negligible.
For EcMDH, the renaturation data showed better fits to a single exponential rise (rather than a double exponential rise) using a non-linear least squares procedure. However, the renaturation profiles of EcMDH observed when GroEL alone was present were best fit to a double exponential rise. The best fits were determined based on reduced χ 2 values and residuals.

Results
Comparison of the sequence and structure of EcMDH and PmMDH. A comparison of the amino acid sequences between E. coli and porcine mitochondrial malate dehydrogenase reveals a large degree of sequence identity and similarity, 58.2% and 80.3%, respectively ( Figure 1A). Functionally significant residues show even larger degrees of conservation. For example, the subunit interfaces of the dimeric structures are 74% identical (26). Hydropathy index (27)  Although both of these proteins can potentially interact with a group I chaperonin family (i.e. GroE or Hsp60) in their respective host organisms, it is clear that the EcMDH, under the defined solution conditions, does not need the GroE chaperonin system during refolding (Figure 2A). The renaturation kinetics are virtually identical for spontaneous refolding (refolding without chaperonins) or folding with GroEL when either ATP or ATP and GroES are present. All three kinetic profiles showed optimal fits to a single exponential rise function (0.35 ± 0.01 sec -1 ). When ATP is absent, the high-affinity nucleotide-free GroEL oligomer slows down the renaturation of EcMDH and the kinetic profile now fits a double exponential rise (k 1 = 0.0074 sec -1 and k 2 = 0.00037 sec -1 ). In all cases, the activity yields of refolded protein are high (80-100% regain of original activity).
In contrast, under exactly the same solution conditions, PmMDH absolutely requires the complete chaperonin system, GroEL-GroES-ATP to refold and reactivate ( Figure 2B). No reactivation is observed when refolding is initiated without the chaperonins present or when GroEL or GroEL and ATP are present. The chaperonin requirements of PmMDH agree with previous studies (4,29). In addition to the differences in the chaperonin requirements, the refolding yields of PmMDH are substantially lower (~60%) when compared to the recoveries observed with EcMDH (~90-100%). It is clear that the folding mechanisms for these two proteins under these solution conditions are different.

Proline isomerization and folding of PmMDH. Comparative folding studies between homologous
prokaryotic and eukaryotic aspartate transaminase performed by Widmann and Christen (2000) indicate that the slower folding rate of the mitochondrial enzyme is due, in part, to differences in proline isomerization (14). A sequence comparison between EcMDH and PmMDH reveals that there are more proline residues present within the mitochondrial primary sequence than in its bacterial homologue (21 vs. 13 prolines, respectively). To test the possibility that these additional proline residues are responsible for the slower folding rate of PmMDH, we At glycerol concentrations of up to 35% (4M), the refolding and reassembly rates of EcMDH at 20 o C are virtually unaffected ( Figure 4A). In stark contrast, the reassembly and reactivation of PmMDH is dramatically increased under the same conditions ( Figure 4B). The renaturation profile is influenced by the concentration of glycerol present in the renaturation mixture ( Figure 5A, Table 1). The optimal concentration range for glycerol-enhanced renaturation rates and yields are from 20 to 35% glycerol. At higher glycerol concentrations (> 50%), the reactivation rates are slower, but the overall yields are higher than those observed in the absence of glycerol ( Figure 5A). Furthermore, including glycerol in the renaturation mixture at "nonpermissive" physiological temperatures of 37 o C also accelerates the refolding and assembly reaction with the refolding yields as high as ~80% ( Figure 5B, Table 1). Under these temperature conditions, which are nonpermissive in the absence of glycerol, the optimal renaturation rates are also observed around glycerol concentrations of about 35% ( Figure 5B).

Change in the chaperonin requirements for PmMDH folding in the presence of glycerol. When glycerol
is added to the renaturation solution, the requirement for GroES is relaxed. Denatured PmMDH can now be substantially reactivated in 35% glycerol with GroEL and ATP instead of just GroEL, GroES, and ATP. ( Figure   6). It is important to note that the high affinity chaperonin (GroEL alone) can still completely arrest the renaturation of PmMDH ( Figure 6). This observation indicates that, in the presence of glycerol, the folding intermediates of PmMDH still have a high affinity for GroEL and can form tight, arrested complexes with GroEL. But more importantly, the spontaneous folding and the GroEL-GroES-ATP reactivation profiles are virtually identical in the presence of glycerol ( Figure 6, Table 1). Curiously, the folding and assembly of PmMDH in the presence of the GroE chaperonins is faster with glycerol than without it (Table 1).
Given this observation, we wanted to test if the PmMDH folding intermediates, generated in the presence of glycerol, will even interact with the activated chaperonin complex (GroEL,GroES, and ATP). In order to observe this initial interaction, a rapid immunoprecipitation technique with anti-GroEL antibodies is shown that glycerol can also prevent large-scale rhodanese aggregation (32).

Glycerol jump experiments with GroEL-PmMDH complexes. Since the chaperonin requirements for
PmMDH become less stringent when glycerol is present, this raised the following question. Will a GroEL-PmMDH complex formed in the presence of glycerol retain a non-stringent chaperonin requirement if this system is rapidly switched from a glycerol containing solution to one without glycerol? Starting with an arrested PmMDH-GroEL complex, a jump from low (0%) to high glycerol (35%) solution concentrations shows that PmMDH is released from GroEL as soon as ATP is added and it refolds to an active state ( Figure 9). In fact, if the jump solution contains ATP, the reactivation kinetics is identical to the situation where ATP is added to a GroEL-PmMDH complex that is formed under high glycerol conditions. However, if the chaperonin-PmMDH complex is formed in the presence of 35% glycerol and the solution conditions are rapidly jumped to lower glycerol concentrations (3.5%) in the presence of ATP, reactivation of PmMDH is no longer observed ( Figure 9). Only when GroES is included in the refolding mixture at low glycerol concentrations can PmMDH once again reactivate (data not shown). These experiments indicate that the glycerol dependent shift in chaperonin stringency is less than 3-5 sec (typical manual mixing times).  isozymes were originally studied by Jaenicke and coworkers (34). These investigators found that the reactivation of cytoplasmic MDH follows first order kinetics and are independent of concentration. In contrast, the mitochondrial isozyme (PmMDH) reactivation is found to be concentration dependent. One model used to fit the data assumes that the monomer rapidly assembles into an inactive dimer that undergoes a unimolecular reaction before an active dimer is formed, a bi-unimolecular mechanism (34). Since there is an enhanced renaturation of PmMDH in glycerol, we set up experiments to determine if the molecular mechanism of PmMDH assembly changes in the presence of this osmolyte to be independent of protein concentration.
The easiest way to determine if glycerol affects the mechanism of the kinetic processes is to carry out the renaturation of PmMDH in 35% glycerol at two different concentrations ( Figure 11). We found that there is a noticeable concentration dependent increase in the renaturation kinetic profiles of PmMDH in the presence of glycerol. These results suggest that the mechanism of this reaction still depends on a second order process.   Table 2).

Discussion
It has been repeatedly demonstrated that homologs in protein folds do not translate into similar protein interactions with the GroE chaperonins (4-6). Data collected by Frieden and Clark strongly suggest that the differences in folding kinetics of dihydrofolate reductase isozymes (DHFR) determine of chaperonin interactions (35,36). Based on structural differences these investigators also suggested that additional unstructured loop regions in the protein determined chaperonin interactions (6). However, Horwich and coworkers suggest that these unstructured regions may be irrelevant to the interaction between GroEL and DHFR, and it is the folding core that is responsible for the kinetic control of chaperonin-substrate interaction (8). Regardless of the exact mechanism of interaction, these experiments with DHFR homologs indicate that folding speed may be a crucial parameter governing chaperonin-protein substrate interactions.
It is important to note that DHFR falls into the Class II substrates category for chaperonin interaction (37). These substrates do not require the complete contingent of the chaperonins (i.e. GroEL and GroES) to fold (38). The homologous malate dehydrogenase isozymes, EcMDH and PmMDH, used in our work naturally fold in the presence of group I chaperonin (Cpn60/Cpn10) systems in vivo. Since PmMDH absolutely requires the entire set of GroE chaperonins and ATP to mediate its folding under non-permissive in vitro conditions, this in vitro substrate is classified as a potential Class III substrate (37,38). Based on the identities of a broad set of chaperonin substrates isolated by co-immunoprecipitation experiments from E. coli cell extracts, Hartl and coworkers suggested that the αβ fold, with a particular reference to the αβ fold of porcine mitochondrial malate dehydrogenase, correlate with potential chaperonin substrate identification (7). Our experiments have shown that, unlike its structurally homologous porcine mitochondrial counterpart, E. coli MDH folds very rapidly without the assistance of the GroE chaperonins and does not appear to interact with any of the physiologically relevant activated forms of the GroE chaperonins. Our results strongly suggest that structural folds cannot be used alone to identify potential class III chaperonin substrates. Based on our results however, we do predict that a substantial portion of the potential Escherichia coli chaperonin substrates identified by Houry et al., (7) should have slower folding/assembly kinetics than the other E. coli proteins that do not require chaperonin assistance to fold.
Widmann and Christen have found that the folding rates of three different oligomeric E. coli proteins are much faster than their mitochrondial homologues. These investigators have found that the mitochondrial aspartate aminotransferase folds slower than its E. coli counterpart because of cis-trans proline isomerization.
Furthermore, these investigators also note that the EcMDH folds faster than PmMDH at 25 o C. Unlike mitochondrial asparate aminotransferase, our results suggest that proline isomerization is not the rate limiting step for renaturation of PmMDH. In addition, if proline isomerization was the rate limiting step, one would also predict that isomerization will still retard the folding rate even when folding enhancers such as glycerol are added to the refolding mixture. If so, the renaturation rates for the short pulsed denatured PmMDH (i.e. short incubation times in the denaturant) with glycerol could be comparable to those rates observed with EcMDH renaturation. However, contrary to this expectation, variable denaturation times of PmMDH did not result in disparate renaturation kinetics in the presence of glycerol (Data not shown).
Our results indicate that the renaturation of porcine mitochondrial malate dehydrogenase, PmMDH, is enhanced by glycerol addition at all temperatures tested and that the chaperonin requirements diminish in its presence (39). 90 o light scattering experiments have confirmed that the addition of glycerol inhibits the largescale off-pathway aggregation reaction. We initially thought that the aggregation prevention may be sufficient to facilitate the rapid renaturation of PmMDH. It has been shown in silico that a small addition of urea at nondenaturing concentrations may prevent early aggregation of some proteins (40 (21) truly reflects the molecular events, the fits to our data indicates that the collapse or accumulation of folded monomers to assembly competent forms is accelerated or enhanced when glycerol is present (Table 1). Apparently, this glycerol dependent collapse may be more substantial than the increase in assembly competent monomers observed when chaperonins alone are used to facilitating folding (Table 1). However, we must be cognizant of the fact that we are only examining the renaturation profiles and that the molecular events preceding the association and renaturation (i.e. the folding reaction toward an assembly competent PmMDH monomer) cannot be directly determined by the analysis presented here. The PmMDH reactivation reflects the final step in the formation of the native active dimer (21,34 We have observed that various osmolytes also accelerate PmMDH reactivation with or without chaperonins. In addition, with glycerol present, the interaction between activated chaperonin and refolding PmMDH subunits cannot be detected by rapid immunoprecipitation. Given the observed influence of glycerol on spontaneous PmMDH folding, we tend to favor the explanation that osmolytes, particularly glycerol, primarily influences the folding of the PmMDH subunits. Indeed, other investigators have provided evidence suggesting that osmolytes may enhance folding because these compounds may decrease the activation energy barrier for folding, leading to a more rapid polypeptide collapse (15,42).
Even though the molecular basis behind the glycerol-induced enhancement of the PmMDH renaturation rates cannot be determined by the data presented here, it is clear that this enhancement leads to a reduction in the chaperonin requirements. The accumulated literature data supports the notion that some polyols such as sucrose and glycerol can stabilize proteins by shifting the equilibrium toward more compact or collapsed states, perhaps explaining the observed increase in rates of the apparent folding reactions (15,16). Likewise, a reduction in the population of transient expanded states may also diminish the available species that lead to offpathway aggregation (43). Consequently, this decline in the population of transient expanded states may explain why the activated chaperonin no longer interacts very well with the refolding PmMDH monomers.
Although the exact mechanism governing this accelerated folding are not clear, polyols may decrease the volume of the transient state intermediate (15). By measuring specific transfer free energies of amino acid and peptide backbone between aqueous and osmolyte containing solutions, Bolen and coworkers have uncovered a possible molecular explanation for osmolyte-induced stabilization (44,45). Experimental evidence suggests that the denatured and native states of proteins are both destabilized by osmolytes. However, Bolen's thermodynamic data indicates that the destabilization of the denatured state far exceeds the destabilization free energy of the native state because the peptide backbone solubility decreases in osmolyte solutions. This disparity in destabilization results in a net increase in the free energy of stabilization (44,45). If we extend this property to the energy levels of folding intermediates and transition states, an acceleration in folding can result primarily from a destabilization of the intermediate state, a stabilization (or smaller destabilization) of the more native and compact transition states or osmolytes may affect both states. The bottom line is that any subsequent decrease in the activation energy barriers (the folding landscapes) will result in faster folding rates (46,47). With glycerol present, faster folding of the PmMDH intermediate towards an assembly competent, native-like monomer may result from any decrease in the activation energy differences between folded intermediates and transition states. Using the kinetic model presented by Clarke and coworkers (21,22), our analysis suggests that there may be an increase in the microscopic rate constant for the collapse toward an assembly competent monomer by at least 1 to 4 orders of magnitude (Table 1). Unfortunately, the analysis performed herein does not provide any information about the energy levels of the transition states. However, any resulting increase in the concentration of assembly competent monomers will naturally lead to an increase in apparent rate of the bimolecular assembly reaction and may bypass the need for any chaperonin assistance.
The effect of glycerol on the entire reaction mechanism also shows that the off-pathway kinetics have changed. Clarke and coworkers have suggested that the chaperonins influence the partitioning reactions of offpathway intermediates (21,22). The fit of our data to the model predicts that the aggregation reaction will also decrease by an order of magnitude (see k 5 , Table 1). The slower aggregation reaction coupled with the rapid collapse to assembly competent monomers correlates with the observed loss in large aggregates formation during spontaneous refolding ( Figure 8). The effect of glycerol on renaturation of PmMDH may not be specific because we also observe an accelerated and/or enhanced folding (37 o C) of PmMDH by other osmolytes such as betaine, sucrose, and trimethyl amine N-oxide (TMAO). Curiously, betaine and TMAO do not change the chaperonin requirements when refolding PmMDH binds to the high affinity chaperonin (Table 2). Here again is another situation where the GroES requirement is not dictated by "permissivity" of the folding conditions (17,39). Since various osmolytes show differential interactions with proteins and change the chaperonin requirements (48-50), our future studies will examine how betaine and TMAO effect PmMDH folding kinetics.
In summary, our experimental results support the notion that the folding rates, rather than the structural identity or folding class of chaperonin substrates, are the key determinants that dictate interactions between transient folding intermediates populations and chaperonins (9,10). The folding intermediate structures of homologous proteins that require chaperonins to fold are predicted to have longer lifetimes and/or slower onpathway folding rates. In addition, our studies indicate that some osmolytes will profoundly influence the chaperonin requirements and that osmolytes certainly may influence the folding kinetics and stabilization of onpathway protein folding intermediates.      Table 1.  glycerol ( )). The kinetic parameters resulting in the best fit to the renaturation profiles of PmMDH are listed in Table 1. The kinetic constants are listed in Table 1.