COMMUNICATION:
ATP Hydrolysis Is Critical for Induction of Conformational Changes in GroEL That Expose Hydrophobic Surfaces*

(Received for publication, December 5, 1996, and in revised form, January 2, 1997)

Boris M. Gorovits , Jesse Ybarra and Paul M. Horowitz Dagger

From the Department of Biochemistry, University of Texas Health Science Center, San Antonio, Texas 78284

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

The degree of hydrophobic exposure in the molecular chaperone GroEL during its cycle of ATP hydrolysis was analyzed using 1,1'-bis(4-anilino)naphthalene-5,5'disulfonic acid (bisANS), a hydrophobic probe, whose fluorescence is highly sensitive to the environment. In the presence of 10 mM MgCl2 and 10 mM KCl the addition of ATP, but not ADP or AMP-PNP, resulted in a time-dependent, linear increase in the bisANS fluorescence. The rate of the increase in the bisANS fluorescence depended on the concentrations of both GroEL and the probe. The effect could be substantially inhibited by addition of excess ADP or by converting ATP to ADP using hexokinase, showing that the increase in the bisANS fluorescence was correlated with ATP hydrolysis. The rate of ATP hydrolysis catalyzed by GroEL was uncompetitively inhibited in the presence of bisANS. This uncompetitive inhibition suggests that the probe can interact with the GroEL-ATP complex. The inability of the nonhydrolyzable ATP analog, AMP-PNP, to cause a similar effect is explained by the interaction of bisANS with a transient conformational state of GroEL formed consequent to ATP hydrolysis. It is suggested that this short lived hydrophobic exposure reflects a conformational shift in GroEL that results from electrostatic repulsion between the bound products of ATP hydrolysis, and it plays an important role in the mechanism of the chaperonin cycle.


INTRODUCTION

The molecular chaperones have been shown to facilitate protein folding and transport of nascent polypeptides in vivo (1). The chaperonin system, involving Escherichia coli proteins GroEL and GroES, has been demonstrated to promote in vitro refolding of a wide variety of proteins (4), including rubisco1 (2) and rhodanese (3). This refolding activity is associated with ATP binding and hydrolysis.

Recently, a thermodynamic partitioning model for hydrophobic binding of polypeptides by GroEL has been suggested (5, 6). This model implies that there is a thermodynamic partitioning between folding of the protein to internalize its hydrophobic residues and the hydrophobic interaction of the protein with hydrophobic surfaces on GroEL. This proposal emphasizes the importance of investigating the role of hydrophobic interactions in the mechanism of GroEL-assisted protein folding. In addition, all models for interactions involving GroEL during the folding cycle recognize the importance of nucleotide binding and hydrolysis in driving the chaperonin cycle that alternately binds interactive intermediates and releases species that are capable of folding. No models to date have identified explicitly how the act of hydrolysis is coupled to the alteration of hydrophobic interactions. Does the hydrolysis serve simply to exchange ATP and ADP or does the hydrolytic event provide the motive force for the requisite conformational changes? If it is the latter, how could the hydrolysis be coupled to the structural events?

In the present study, we have investigated the level of hydrophobic exposure in the molecular chaperone GroEL during its cycle of ATP hydrolysis. The hydrophobic probe, 1,1'-bis(4-anilino)naphthalene-5,5'-disulfonic acid (bisANS), has been used. The influence of different nucleotides has been studied. In the presence of MgCl2 and KCl, addition of ATP, but not ADP or AMP-PNP, resulted in a linear time-dependent increase in the bisANS fluorescence. It is shown that the increase in the bisANS fluorescence is correlated with ATP hydrolysis itself, rather than with the binding of nucleotides. The importance of the hydrophobic exposure, occurring during the ATP hydrolysis, is discussed.


MATERIALS AND METHODS

Reagents and Proteins

1,1'-bis(4-Anilino)naphthalene-5,5'-disulfonic acid was purchased from Molecular Probes (Eugene, OR). ATP, ADP, and AMP-PNP were purchased from Sigma. All reagents were analytical grade.

Protein Purification

The chaperonin GroEL was expressed in E. coli and purified as described previously (7). After purification, GroEL was dialyzed against 50 mM Tris-HCl, pH 7.6, containing 1 mM dithiothreitol, then made 10% (v/v) in glycerol, rapidly frozen, and stored at -70 °C. The GroEL protomer concentration was measured using the bicinchoninic acid protein assay (Pierce) according to the procedure recommended by the manufacturer and assuming a molecular mass of 60 kDa.

Standard Buffer

The following solution was used as a standard buffer: 50 mM Tris-HCl, pH 7.8, 10 mM MgCl2, 10 mM KCl.

Fluorescence Detection of the bisANS Interaction with GroEL

Fluorescence measurements were made on a SLM 500c spectrofluorometer (SLM Instruments, Urbana, IL). Samples contained 0.1-0.4 µM GroEL and 0.2-40 µM bisANS in the corresponding Standard Buffer. Samples were excited at 399 nm, and fluorescence was monitored at 500 nm.

Catalysis of the ATP Hydrolysis by GroEL

The ATPase activity of GroEL was measured by the method of Viitanen et al. (8) with the modifications of Horovitz et al. (9). The concentration of GroEL oligomer was 10 µM in Standard Buffer, containing 0-40 µM [gamma -32P]ATP (72-235 Ci/mol) and 0-40 µM bisANS. The reaction was carried out for 0-16 min.


RESULTS

The Binding of the Hydrophobic Probe bisANS Changes in the Presence of Nucleotides

One of the methods for monitoring changes in the exposure of hydrophobic surfaces in a protein involves use of noncovalent fluorescent probes. One such probe, bisANS, has been widely used to study conformational changes in proteins (10, 11). bisANS fluorescence is increased in the presence of GroEL, showing that some of the hydrophobic sites of the protein are solvent-exposed under normal conditions (12). The signal reaches its maximum value within the time of the manual mixing and remains stable (Fig. 1). The presence of nucleotide in the solution results in the change of the binding behavior that depends on the nature of the nucleotide added. While the presence of ADP or the nonhydrolyzable ATP analog, AMP-PNP, causes only a slow increase in the bisANS fluorescence (Fig. 1), addition of ATP induces a faster increase in the bisANS fluorescence (Fig. 1). The fluorescence intensity reaches a plateau after prolonged incubation (data not shown). For example, under the conditions shown in Fig. 1 (0.4 µM GroEL, 1 mM ATP, 10 µM bisANS) the fluorescence plateaus at a value of approx 10 units after a 10-min incubation.


Fig. 1. Fluorescence of hydrophobic probe, bisANS, induced by binding to GroEL, linearly increases in the presence of ATP. bisANS (10 µM) was added as indicated to the solution of 0.4 µM GroEL, 10 mM MgCl2, 10 mM KCl. Nucleotides were added to the final concentration of 1 mM. Samples were excited at 399 nm, and fluorescence was monitored at 500 nm.
[View Larger Version of this Image (17K GIF file)]


The rate of the increase of the bisANS fluorescence at a given initial ATP concentration depends on the concentrations of both the protein and the bisANS. For example, at 5 µM bisANS, increasing GroEL concentrations result in higher rates of bisANS fluorescence increase, which reach an apparent maximum rate at [GroEL] = 0.3-0.4 µM (data not shown). It should be noted that under these conditions there are 5.6 µM GroEL monomers so there is no large excess of bisANS. The dependence of the rates on the bisANS concentration is somewhat more complex. The rates of the fluorescence change increase up to about a 5 µM concentration of the probe, and then the rates decrease as the bisANS concentration is raised further. The latter effect can be explained by the inhibition of the ATPase activity of GroEL at high concentrations of the probe, >5 µM (see below). For example, the rate of GroEL-catalyzed ATP hydrolysis is inhibited 46.6% at [bisANS] = 40 µM.

To elucidate the type of inhibition of the GroEL-catalyzed ATP hydrolysis by bisANS, the reaction was performed at different concentrations of bisANS (Fig. 2A). The initial ATPase velocities were fit to the Hill equation. The dependences of the maximal initial ATPase velocities (Vmax), apparent dissociation constants (K), and Hill coefficients (n) were analyzed as a function of [bisANS]. Both Vmax and K were reduced at high concentrations of the probe (Fig. 2B), indicating that bisANS is an uncompetitive inhibitor of the ATPase activity of GroEL. The value of the Hill coefficient obtained in the absence of bisANS (n = 2.41) corresponds to the previously reported value (9). It is interesting that n does not change considerably up to 40 µM bisANS, although the rate of the reaction is reduced to 46.6% of the initial value (Fig. 2B). ATPase activities were constant after preincubation of GroEL for at least 10 min at the concentrations of bisANS used. This indicates that bisANS does not cause any progressive changes in GroEL under the present experimental conditions.


Fig. 2. A, ATPase activity of GroEL measured at different ATP and bisANS concentrations. The concentration of GroEL oligomer was 10 nM. The lines represent data fit by using the Hill equation. The bisANS concentrations were 0 (squares), 10 (circles), 40 (triangles), and 100 (diamonds) µM. The ATPase assays were performed as described under "Materials and Methods." B, the dependence of K and Vmax as a function of bisANS concentration. The concentration of GroEL oligomer was 10 nM. ATPase activity assays were performed as described under "Materials and Methods." The squares and circles correspond to K and Vmax values, respectively. The values of K and Vmax were normalized assuming those at 0 M bisANS equal 1. The lines are shown to guide the eye.
[View Larger Version of this Image (13K GIF file)]


bisANS was observed to dissociate from GroEL when samples that had been incubated with ATP and bisANS were diluted 1:10 into buffer without the probe to give a final bisANS concentration of 1 µM. The decrease of the fluorescence signal could be fit with two exponentials having approximately equal amplitudes and rate constants of 47.5 × 10-3 s-1 and 5.3 × 10-3 s-1 without ATP in the dilution buffer and 82.2 × 10-3 s-1 and 12.1 × 10-3 s-1 in the presence of 0.2 mM ATP. These results demonstrate reversibility of the bisANS-GroEL interaction, and they show that ATP can accelerate both binding and dissociation of the probe.

The Increase in the bisANS Fluorescence Is Related to the GroEL-catalyzed ATP Hydrolysis

Several possibilities were examined as contributing to the increasing bisANS fluorescence in the presence of GroEL and ATP. No covalent phosphorylation of GroEL was detected up to 4 h under conditions employed in the present study (data not shown). Additional experiments demonstrated that ATP hydrolysis did not cause a permanent modification of GroEL. ATP and GroEL were premixed, and ATP hydrolysis was allowed to proceed in the absence of bisANS. When bisANS was then added to replicate samples after various intervals, the initial fluorescence increment and the subsequent rates of increase of the bisANS signal were close to those observed if the bisANS had been added at t = 0. Finally, under conditions where GroEL ATPase activity was largely suppressed (low concentration of K+, 1 mM (8)), no significant increase in the bisANS fluorescence was observed (data not shown).

Addition of increasing amounts of ADP to the mixture containing 0.1 µM GroEL, 0.2 mM ATP, and 5 µM bisANS, successfully inhibited the fluorescence increase (Fig. 3). The effect was completely quenched when the ADP concentration reached 1 mM. Excessive amounts of ADP (>2 mM) caused a slight decrease in the fluorescence signal (data not shown).


Fig. 3. The time-dependent increase of the bisANS fluorescence in the presence of ATP and GroEL can be inhibited by excess ADP. ADP was added to the solution containing 0.1 µM GroEL oligomer, 5 µM bisANS, 0.2 mM ATP, 10 mM MgCl2, 10 mM KCl in 50 mM Tris buffer, pH 7.8. Samples were excited at 399 nm, and fluorescence was monitored at 500 nm. The line is shown to guide the eye.
[View Larger Version of this Image (13K GIF file)]


The increase in the bisANS fluorescence in the presence of GroEL and ATP could also be quenched by converting ATP to ADP using hexokinase and glucose (Fig. 4). In this experiment hexokinase (5 units) was added to the solution containing 0.4 µM GroEL, 0.2 mM ATP, 10 µM bisANS, and 0.1 M glucose. ATP is quickly hydrolyzed to ADP in the enzymatic reaction catalyzed by hexokinase. As a result, the rate of increase in the bisANS fluorescence was largely inhibited (Fig. 4, lower trace).


Fig. 4. Conversion of ATP to ADP in the presence of hexokinase can inhibit the time-dependent increase of the bisANS fluorescence in the presence of ATP and GroEL. Hexokinase (5 units) was added (lower curve) to the solution containing 0.1 µM GroEL oligomer, 5 µM bisANS, 0.2 mM ATP, 0.1 M glucose, 10 mM MgCl2, 10 mM KCl in 50 mM Tris buffer, pH 7.8. The upper curve shows a control experiment in which no hexokinase was added. Samples were excited at 399 nm, and fluorescence was monitored at 500 nm.
[View Larger Version of this Image (17K GIF file)]



DISCUSSION

Hydrophobic interactions have been suggested to be the main driving force for the association of protein-substrates with GroEL (13). Therefore, the oligomer of the chaperonin should possess, or be able to express under certain conditions, hydrophobic surfaces, that are necessary for this interaction. By using the hydrophobic probe bisANS, only limited accessible hydrophobic regions are found in the intact chaperonin GroEL (3). Due to the high degree of the intramolecular flexibility, GroEL is able to expose hydrophobic sites under certain conditions (14, 15). For example, a noticeable increase in the number of bisANS binding sites in the unperturbed GroEL 14-mer can be induced upon addition of salt (12) or under conditions where GroEL is dissociated into folded monomers (e.g. >= 2.0 M urea) (16).

Although interaction of the GroEL oligomer with nucleotide under conditions where no hydrolysis is possible results in a conformational change (17), the present data show that there is no considerable change in the level of accessible hydrophobic surfaces (Fig. 1).

In the presence of K+, GroEL possesses weak ATPase activity (18). Under such conditions, addition of the hydrophobic probe results in a time-dependent increase in the bisANS fluorescence (Fig. 1), demonstrating the ability of the probe to bind to the hydrophobic surfaces that can be exposed on GroEL during the catalytic cycle of ATP hydrolysis. Since excess ADP or complete conversion of ATP to ADP by using the hexokinase-catalyzed reaction can completely inhibit the effect, the increase in the bisANS fluorescence can be directly correlated to the GroEL-catalyzed ATP hydrolysis. It should also be noted that the addition of ATP does not induce quickly a large fluorescence increment (Fig. 1, nucleotide added). This supports further the idea that the increase in bisANS fluorescence follows the hydrolysis and not the binding of ATP.

The results are consistent with the model presented in Scheme 1. In this model, the binding of ATP to GroEL is followed by the hydrolysis step where an intermediate state (GroEL-ADP-Pi), I, is produced. This GroEL conformation (shown in brackets) is not stable, and it transiently contains both bound ADP and Pi. In fact, because of the positive cooperativity among the ATPase sites within one ring of the GroEL oligomer, the products from the hydrolysis of 7 mol of ATP would be bound to that ring. The electrostatic repulsion between the bound ADP and Pi presumably contributes to this instability. The transient state, I, would have increased accessibility of hydrophobic surfaces, and, therefore, it would be able to bind bisANS with a consequent increase of the fluorescence signal. It has been shown that substrate and ligand interactions can modulate hydrophobic exposure in GroEL, so it cannot be excluded that the interactions of bisANS can contribute to the appearance of hydrophobic sites. Thus, the metastable state, I, or one that is accessible from it forms only after ATP hydrolysis and is not observed with any equilibrium states that can be produced by incubation with nucleotides. This potential malleability of GroEL follows ATP hydrolysis and does not depend on ATP binding directly. Since the probe is a low molecular weight molecule and, unlike an unfolded protein, does not have tendency to sequester its hydrophobic sites by folding, bisANS remains bound to GroEL. In addition, the short lifetime of the transient GroEL state formed during hydrolysis might not permit bisANS to equilibrate with all available sites on the GroEL molecule during one cycle of ATP hydrolysis. This non-equilibration would explain why the rate of ATP hydrolysis is greater than the rate of bisANS binding. Therefore, the fluorescence increases in a time-dependent manner. State I is readily reversible, since no exposed hydrophobic surfaces can be detected in the forms binding ADP or AMP-PNP nor in the form to which ATP initially binds.


[View Larger Version of this Image (7K GIF file)]

Scheme 1.

bisANS is able to uncompetitively inhibit the ATPase activity of GroEL. This suggests that the probe is able to interact with species related to the enzyme-substrate complex. The inhibitor constant (Ki) is estimated as 35 µM, and the complex is able to only slowly hydrolyze ATP, since the maximum velocity of ATP hydrolysis for the bisANS-bound complex is estimated to be 9% of the corresponding value in the absence of the inhibitor.

The fact that the rate of bisANS fluorescence increase depends on both bisANS and GroEL concentrations demonstrates that at low concentrations of the probe, the GroEL-bisANS interaction is controlled by the mass action law. This shows that the state of the protein that is able to interact with bisANS (transition state of ATP hydrolysis) is metastable, and its concentration controls the rate of the chaperonin-probe interaction.

A large conformational change in the GroEL molecule has been detected by using electron microscopy (19). Interestingly, similar changes have been found when GroEL is incubated with ADP and Pi, but not with ADP alone (20). These changes were observed by electron microscopy of samples incubated in 30 mM ADP and Pi before fixation. In the present study, no large increases in hydrophobic exposure were observed at 3 mM ADP and Pi (data not shown). This presumably reflects the apparently large energy from concerted ATP hydrolysis. The conformational shift, occurring during ATP hydrolysis by GroEL, has been suggested to result in the formation of an open conformation in which the apical domain of each GroEL monomer moves upward and outward. This has the effect of exposing previously hidden areas of the GroEL molecule. A GroEL mutant has recently been prepared which was constrained by inserting a disulfide bridge in such a manner that movements of the apical domain of each GroEL monomer would be hindered. This mutant is locked in a closed conformation, and the hydrolysis of ATP is inhibited although ATP is bound (21). Those results are consistent with the picture presented here showing that ATP hydrolysis and structural changes are coupled, and hydrolysis per se and not only nucleotide binding have important conformational consequences. The same method was used to suggest that the release of substrate protein from GroEL is directly coupled with the ATP hydrolysis step (21).

The results presented here are consistent with the suggestion that GroEL can undergo a large conformational shift accompanied by large hydrophobic exposure during ATP hydrolysis. Since the unstable, transient GroEL conformation generated during ATP hydrolysis contains both ADP and Pi bound to it, the high energy of the electrostatic repulsion between those two species could produce the force necessary for GroEL to undergo the conformational shift. This molecular movement enables GroEL to unfold partially folded protein bound on the chaperonin. The subsequent release of the substrate protein, due to its tendency to fold, will allow the polypeptide to form a native-like conformation or correct a misfolded state. Further, it follows that the major exposure of hydrophobic surfaces is transient, thus explaining why no extensive hydrophobic surfaces are commonly detected on GroEL in complexes at equilibrium.


FOOTNOTES

*   The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    To whom correspondence should be addressed: Dept. of Biochemistry, University of Texas Health Science Center, 7703 Floyd Curl Dr., San Antonio, TX 78284. Tel.: 210-567-3737; Fax: 210-567-6595.
1   The abbreviations used are: rubisco, ribulose-bisphosphate carboxylase/oxygenase; AMP-PNP, 5'-adenylyl beta ,gamma -imidodiphosphate; bisANS, 1,1'-bi(4-anilino)naphthalene-5,5'-disulfonic acid.

REFERENCES

  1. Ellis, R. J., and Van der Vies, S. M. (1991) Annu. Rev. Biochem. 60, 321-347 [CrossRef][Medline] [Order article via Infotrieve]
  2. Goloubinoff, P., Christeller, J. T., Gatenby, A. A., and Lorimer, G. H. (1989) Nature 342, 884-889 [Medline] [Order article via Infotrieve]
  3. Mendoza, J. A., Rogers, E., Lorimer, G. H., and Horowitz, P. H. (1991) J. Biol. Chem. 266, 13044-13049 [Abstract/Free Full Text]
  4. Viitanen, P. V., Gatenby, A. A., and Lorimer, G. H. (1992) Protein Sci. 1, 363-369 [Abstract/Free Full Text]
  5. Zahn, R., Axmann, S. E., Riicknagel, K.-F., Jaeger, E., Laminet, A. A., and Pluckthun, A. (1994) J. Mol. Biol. 242, 150-164 [CrossRef][Medline] [Order article via Infotrieve]
  6. Zahn, R., and Pluckthun, A. (1994) J. Mol. Biol. 242, 165-174 [CrossRef][Medline] [Order article via Infotrieve]
  7. Clark, A. C., Hugo, E., and Frieden, C. (1996) Biochemistry 35, 5893-5901 [CrossRef][Medline] [Order article via Infotrieve]
  8. Viitanen, P. V., Lubben, T. H., Reed, J., Goloubinoff, P., O'Keefe, D. P., and Lorimer, G. H. (1990) Biochemistry 26, 5665-5671
  9. Horovitz, A., Bochkareva, E. S., Kovalenko, O., and Girshovich, A. S. (1993) J. Mol. Biol. 231, 58-64 [CrossRef][Medline] [Order article via Infotrieve]
  10. Rosen, C.-G., and Weber, G. (1969) Biochemistry 8, 3915-3920 [Medline] [Order article via Infotrieve]
  11. Horowitz, P. M., and Criscimagna, N. L. (1990) J. Biol. Chem. 265, 2576-2583 [Abstract/Free Full Text]
  12. Horowitz, P. M., Hua, S., and Gibbons, D. L. (1995) J. Biol. Chem. 270, 1535-1542 [Abstract/Free Full Text]
  13. Lin, Z., Schwarz, F. P., and Elsenstein, E. (1995) J. Biol. Chem. 270, 1011-1014 [Abstract/Free Full Text]
  14. Gorovits, B. M., and Horowitz, P. M. (1995) J. Biol. Chem. 270, 13057-13062 [Abstract/Free Full Text]
  15. Gorovits, B. M., Raman, C. S., and Horowitz, P. M. (1995) J. Biol. Chem. 270, 2061-2066 [Abstract/Free Full Text]
  16. Mendoza, J. A., Demeler, B., and Horowitz, P. H. (1994) J. Biol. Chem. 269, 2447-2451 [Abstract/Free Full Text]
  17. Gorovits, B. M., and Horowitz, P. M. (1995) J. Biol. Chem. 270, 28551-28556 [Abstract/Free Full Text]
  18. Hendrix, R. W. (1979) J. Mol. Biol. 129, 375-392 [Medline] [Order article via Infotrieve]
  19. Roseman, A., Chen, S., White, H., Braig, K., and Saibil, H. (1996) Cell 87, 241-251 [Medline] [Order article via Infotrieve]
  20. Llorca, O., Marco, S., Carrascosa, J. L., and Valpuesta, J. M. (1997) J. Struct. Biol., in press
  21. Murai, N., Makino, Y., and Yoshida, M. (1996) J. Biol. Chem. 271, 28229-28234 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.