(Received for publication, June 4, 1994, and in revised form, October 25, 1996)
From the Department of Biochemistry, University of Texas Health Science Center, San Antonio, Texas 78284
The interaction of GroEL with urea-unfolded dihydrofolate reductase (DHFR) has been studied in the presence of DHFR substrates by investigating the ability of GroES to release enzyme under conditions where a stable GroES-GroEL-DHFR ternary complex can be formed. In these circumstances, GroES could only partially discharge the DHFR if ADP was present in the solution and approximately half of the DHFR remained bound on the chaperonin. This bound DHFR could be rescued by addition of ATP and KCl into the refolding mixture. The stable ternary complex did not show any significant protection of bound DHFR against proteolysis by Proteinase K. These results are in contrast to those observed with the GroEL-DHFR complex formed by thermal inactivation of DHFR at 45 °C in which GroES addition leads to partial protection of bound DHFR. Thus, the method of presentation influences the properties of the bound intermediates. It is suggested that the ability of GroES to bind on the same side of the GroEL double toroid as the target protein and displace it into the central cavity depends on the way the protein-substrate is presented to the GroEL molecule. Therefore, the compact folding intermediate formed by thermal unfolding can be protected against proteolysis after GroES binds to form a ternary complex. In addition, structural changes within GroEL induced by the experimental conditions may contribute to differences in the properties of the complexes. The more open urea-unfolded DHFR binds on the surface of chaperonin and can be displaced into solution by the tighter binding GroES molecule. It is suggested that the state of the unfolded protein when it is presented to GroEL determines the detailed mechanism of its assisted refolding. It follows that individual proteins, having characteristic folding intermediates, can have different detailed mechanisms of chaperonin-assisted folding.
The chaperonin GroEL,1 together with the co-chaperonin GroES, has been shown to facilitate protein folding (1). The initial step of the chaperonin-assisted folding cycle is interaction of the unfolded polypeptide with GroEL (2). The mechanism of the subsequent release of the bound protein from the complex has been studied intensively (3, 4, 5, 6). GroEL is an oligomer in which its 14 identical subunits are arranged as two, stacked, 7-membered rings to form a cylindrical structure with a central cavity. It has been shown that during the chaperonin cycle two distinct types (cis and trans) of the GroEL-GroES-protein substrate complex can be formed (4). In case of the trans complex, unfolded polypeptide is bound to the GroEL ring opposite to the ring occupied by GroES. Release and rebinding of GroES during the chaperonin cycle results in formation of a cis complex in which the GroES is bound on the same GroEL ring occupied by the protein-substrate (4). When this cis complex is formed, it has been suggested that the unfolded polypeptide is displaced into the central cavity which is capped by the bound GroES, thus placing the previously bound protein in a protected space. It has been suggested that final release of the protein-substrate occurs from the cis complex only (4, 7). These studies have used unfolded proteins that have high affinity for GroEL and require for successful folding the complete chaperonin system including co-chaperonin GroES, ATP, and KCl. There are several proteins that are able to at least partially dissociate from their complexes with GroEL without ATP hydrolysis (8, 9, 10). These proteins generally display low affinity for GroEL. It has been suggested that both GroES and the target proteins use linked binding sites on the surface of GroEL (11). Assuming that the strength of GroEL-GroES interaction does not depend on the nature of the protein-substrate, one can suggest that the diverse proteins that bind to GroEL represent a spectrum of binding affinities for GroEL relative to the binding affinity for GroES. Further, a single protein may fold (or misfold) through multiple states that have different affinities for GroEL. In the present study, possible differences in the mechanisms of chaperonin interactions for weak and tight binding proteins have been explored by comparing interactions of urea- and heat-unfolded bovine dihydrofolate reductase (DHFR).
GroEL was expressed in E. coli and purified as described previously (12). Bovine DHFR, Proteinase K, dihydrofolate (H2F), NADPH, ATP, and ADP were purchased from Sigma.
Preparation of Urea-unfolded DHFRDHFR was unfolded by
dialyzing against 8 M urea (50 mM Tris, 1 mM EDTA, 10 mM ME) for at least 2 h at
room temperature. Protein concentration of DHFR was determined
spectrophotometrically based on the extinction coefficient at 280 nm
reported as 27,280 cm
1 M
1
(13).
Urea-unfolded DHFR was rapidly diluted
into buffer A (50 mM Tris, 10 mM
MgCl2, 10 mM ME, 0.1 mM
H2F, and 0.075 mM NADPH, pH 7.8) with or
without GroEL and nucleotide as described in the individual figure
legends. Final concentrations of the enzyme varied from 0.24 to 0.3 µM. The concentration of the GroEL was varied from 0 to
1.2 µM, and the final urea concentration did not exceed
0.5 M. The solution was mixed, and, after a 1-min
incubation at room temperature, enzyme activity was analyzed by the
decrease in the NADPH concentration detected spectrophotometrically at 340 nm. In some experiments, GroES (final concentration varied from 0 to 1.2 µM) was added to the refolding solution after
30 s of incubation. The enzyme activity was detected as above
after an additional 30-s incubation. Resulting data were analyzed in terms of the initial rate of the reaction.
Urea-unfolded DHFR was rapidly
diluted to buffer A (50 mM Tris, 50 mM KCl, 10 mM MgCl2, 10 mM ME, pH 7.8) with
or without GroEL. Final concentrations of the proteins were: 0.24 µM DHFR and 0.277 µM GroEL. The final
concentration of urea did not exceed 0.6 M.
After a 5-min incubation at room temperature, the mixture was applied
on a Superose 12 gel filtration column (1.5 × 30 cm, flow rate 1 ml/min) equilibrated with buffer A without ME. Fractions containing
GroEL were collected, pooled, and concentrated using Ultrafree-MC
concentration units with 30,000 cutoff (Millipore). SDS Laemmli gel
electrophoresis was used to confirm the presence of DHFR in these
fractions. GroES was added to some solutions to a final concentration
of 0.7 µM. ADP was added to all solutions to give a final
concentration of 2 mM. Proteinase K was then added to 0.2%
of the total protein (w/w), and the solution was incubated for 10 min
at room temperature, after which the reaction was stopped by adding
phenylmethylsulfonyl fluoride to a final concentration of 3 mM. Solutions were incubated for 10 min at room
temperature, boiled, and analyzed on the 12% SDS Laemmli gel system.
Protein was visualized by silver staining (14). The density of the
protein bands was analyzed using NIH Image software.
Bovine DHFR was unfolded in 8 M urea. Dilution
of the unfolded material to the buffer, containing no denaturant,
results in fast unassisted refolding, detected as an appearance of
enzymatic activity of the native protein. The kinetics of NADPH
oxidation by dihydrofolate, catalyzed by native DHFR, have been used in this study to analyze the amount of successfully folded enzyme in
solution. As shown in Fig. 1, introduction of GroEL to
the refolding solution results in a decrease of the enzyme activity. A
similar result has been reported for the mouse DHFR whose refolding can
be completely arrested upon binding to GroEL (8). Fig. 1 demonstrates
that almost 80% of DHFR can be captured at 1:1 ratio of GroEL to the
unfolded DHFR whether or not ADP is present in solution.
The GroEL-DHFR complex slowly dissociates (data not shown and Ref. 8). Under conditions when enzyme substrates (NADPH and dihydrofolate) are present in the solution, considerable dissociation was detected after 1 h of incubation (data not shown). The mouse DHFR-GroEL complex has been reported to completely dissociate after 3 h of incubation under similar conditions (8). These data suggest that the enzyme exists in dynamic equilibrium between free and bound states.
As with several other proteins, the addition of ATP alone can lead to the release of fully active bovine DHFR from its complex with GroEL. These weakly bound proteins contrast with the class of tightly bound proteins, like rhodanese (2), 6-hydroxy-D-nicotine oxidase (15), and ornithine transcarbamylase (16). These latter proteins require the presence of the complete chaperonin system, including the co-chaperonin GroES, in order to be successfully released. The presence of GroES, although not required, does facilitate dissociation of DHFR from the GroEL-DHFR complex (8).
GroES Can Partially Discharge DHFR from the Complex with GroEL in the Presence of ADPFig. 2 shows that GroEL can
arrest approximately 72% of the spontaneous folding of DHFR
(open squares, Fig. 2). Addition of GroES to the preformed
GroEL-DHFR complex results in partial discharge of the active enzyme
(closed squares, Fig. 2). The amount of the released protein
increases up to about 1:1 the ratio of GroES to GroEL-DHFR complex.
Subsequent addition of GroES does not result in any considerable change
of the free enzyme concentration. The maximum level of DHFR released
(65%) represents about 50% of the total protein that had been
captured by GroEL (Fig. 2). The remaining half of the captured enzyme
could be released upon addition of ATP and KCl, producing a 95% yield
of active protein. As a control, a similar titration of GroES to the
GroEL-DHFR complex with no nucleotide present in the solution was
carried out. Since the GroEL-GroES interaction is not possible under
this condition (data not shown and Ref. 4), no significant discharge of
the DHFR over a control sample was observed within the time of this
experiment.
Digestion Analysis of the GroEL-DHFR-GroES Complex
The
presence of a nucleotide is not required for the binding of an unfolded
protein-substrate to GroEL (Fig. 1; Ref. 2). Therefore, a random
distribution of the bound nucleotide (ADP) relative to the site of the
unfolded polypeptide binding on two GroEL toroids can result in two
distinct types of GroES-GroEL interactions (4). Thus, GroES binding can
occur on the same toroid (cis) where protein-substrate is
bound or on the opposite site (trans) of the GroEL oligomer.
These two complexes can be distinguished by using protease digestion
analysis as was reported earlier (4). In the present study, two
different methods were used to form the GroEL-DHFR complex. In the
first case, DHFR was unfolded in 8 M urea for 15 h
before dilution in the presence of GroEL. In the second case, the
complex was formed by heating together GroEL and native DHFR at
45 °C for 10 min as reported earlier (7). Complexes were purified
using size exclusion chromatography, and fractions containing GroEL
were collected and concentrated. Solutions were supplemented with ADP
and GroES (when needed). Digestion analysis was performed using
Proteinase K (final concentration 0.2% w/w). Two different results
were observed. While unfolded DHFR can be partially (42 ± 5%)
protected by GroEL-GroES interaction if the GroEL-DHFR complex was
preformed by heat denaturation (Fig. 3, lane
7), no protection could be detected if the complex was formed by
addition of urea-unfolded protein (Fig. 3, lane 3). The
positive controls, containing GroEL-DHFR complexes, formed by using
urea or heat-denatured enzyme and consequently treated with ADP and
GroES, but not Proteinase K, are shown in lanes 1 and
6, respectively. These results suggest that, due to the
considerable differences in the protein conformation between chemically
and thermally unfolded protein, the resulting GroEL-protein complexes have different properties, leading to the different mechanism of the
GroES-facilitated protein release.
Several different mechanisms have been suggested for the function of the GroEL-GroES system (4, 6). The release of the bound polypeptide from its complex with GroEL is one of the central steps that such a mechanism must address. Several aspects of these mechanisms are based on the ability to form relatively stable complexes involving GroEL in the presence of ADP. It was demonstrated (7) that ternary complexes containing DHFR, GroEL, and GroES could be formed by thermal denaturation and that a significant fraction of the bound DHFR was protected against Proteinase K digestion. This was taken to imply that GroES could bind in one of two ways to the complex. First, GroES could bind over the central cavity of GroEL on the same toroid as DHFR (cis), thus capping the complex and ejecting the DHFR into the chamber formed from the central cavity in the GroEL toroid and the overlying GroES dome; or second, GroES could bind on the toroid opposite to the bound DHFR (trans) leaving the DHFR on the surface. In this model, the DHFR in the cis complex is protected from Proteinase K, while that in the trans complex can be proteolyzed.
The present results are consistent with the observations for the thermally denatured protein, but the results are quite different for the urea-unfolded DHFR. While nearly 50% of DHFR can be protected by GroEL-GroES interaction in the former case, no protection was detected in the case of urea-unfolded protein-substrate (Fig. 3). Thus, it is clear that the complexes formed in these different ways have different characteristics. In contrast to the thermally produced complex, the urea complex can be partially discharged by GroES, and the undischarged part of the bound DHFR is sensitive to Proteinase K digestion. The considerable conformational differences between heat- and chemically unfolded protein can explain this phenomenon. DHFR in 8 M urea is likely to be fully unfolded, and the collapsed states that form on dilution could bind extensively to the binding sites on one of the two GroEL toroids. These extensive surfaces and the rapid collapse of the DHFR on dilution would favor binding close to the surface of the chaperonin. This surface location would be favored by the fact that the protein is introduced to a relatively unperturbed GroEL molecule. Since protein binding and GroES binding appear to involve the same or similar sites, peripheral binding of DHFR would lead to a weak complex in which cis GroES binding could release DHFR into the bulk solution by direct displacement. The trans ternary complex would be relatively stable in the presence of ADP. Thus, urea-unfolded DHFR would form a complex with GroEL, and GroES would release 50% by direct displacement and leave 50% as a trans complex that is susceptible to Proteinase K.
Thermally produced DHFR-GroEL complexes are different from complexes formed using urea-denatured DHFR. Thermal experiments that demonstrate the complex are performed by heating DHFR together with GroEL at 45 °C, so that thermal effects on both DHFR and GroEL must be considered. Bovine DHFR has a thermal transition with a midpoint at ~50 °C (17), so the 45 °C incubation does not lead to fully heat-denatured protein. In addition, many proteins, even after complete thermal denaturation, still contain significant ordered structures that can be unfolded further (18). In addition, GroEL itself has been reported to undergo two thermal transitions between 25 and 35 °C (19), so that temperature-induced formation of complexes influences the structure of GroEL. This suggestion is supported by the observation that there are proteolytic bands that must come from GroEL after proteolytic treatment of the thermally induced complexes (Fig. 3). It is possible, then, that heating DHFR in the presence of GroEL results in small, partially unfolded structures that can bind inside of the perturbed GroEL cavity. Then, the subsequent addition of GroES can result in forming a capped cis complex as described previously. The heating together of DHFR and GroEL is important, since the control experiments that have been carried out in the present study showed that no complexes between GroEL and DHFR can be formed when GroEL is preheated at 45 °C for 10 min and then added to native DHFR at 25 °C (data not shown). When DHFR was heated in the absence of GroEL, a precipitate formed, and the protein that remained in the solution could not be captured by GroEL at 25 °C.
The present data support the concept that the ability of the unfolded polypeptide to form a stable complex with GroEL at different steps of its folding pathway can influence the detailed mechanism of its assisted refolding. This would be in keeping with suggestions that chaperonin-facilitated folding can be modulated by the properties of both the protein-substrate and the GroEL (20, 21, 22).