Refolding kinetics of staphylococcal nuclease and its mutants in the presence of the chaperonin GroEL1

J. Mol. Biol. (1998) 277, 733±745

Refolding Kinetics of Staphylococcal Nuclease and Its Mutants in the Presence of the Chaperonin GroEL Galina P. Tsurupa, Teikichi Ikura, Tadashi Makio and Kunihiro Kuwajima* Department of Physics School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku Tokyo 113, Japan

We have analyzed the effect of the chaperonin GroEL on the refolding kinetics of staphylococcal nuclease and its three mutants by stopped-¯ow ¯uorescence measurements. It was found that a transient folding intermediate of staphylococcal nuclease was tightly bound to GroEL and refolded in the GroEL-bound state without releasing the non-native protein in solution, and the refolding rate in the GroEL-bound state was 0.01 sÿ1. The GroEL-affected refolding of the nuclease appears to be in decided contrast to that of apo-a-lactalbumin reported in our previous study, wherein a-lactalbumin was shown to be more weakly bound by GroEL and to refold in the free state in solution. In spite of the apparent difference between the proteins, the GroEL-affected refolding reactions of both the proteins can be represented by a common uni®ed reaction scheme. On the basis of this scheme, the binding constant between the nuclease intermediate and GroEL was estimated to be larger than 109 Mÿ1. The stoichiometry of binding of the nuclease and its mutants to GroEL was found to be two (nuclease/GroEL 14-mer). The increase in ionic strength resulted in a weakening of the interaction between the nuclease and GroEL, which was attributed to a weakening of the electrostatic attraction between the two proteins as a result of electrostatic screening by ions. Although ATP was found to accelerate the GroELaffected refolding of the nuclease, the refolding rate was still far from the rate of the free refolding. The free refolding behavior of the nuclease and its mutants was restored in the presence of the cochaperonin GroES and ATP. # 1998 Academic Press Limited

*Corresponding author

Keywords: molecular chaperone; chaperonin; GroEL; staphylococcal nuclease; protein folding

Introduction It is now well established that protein folding in vivo is not a spontaneous process but requires a series of helper proteins, including molecular chaperones (Gething & Sambrook, 1992; Buchner, 1996; Hartl, 1996). Molecular chaperones facilitate protein folding by recognizing immature nonnative molecules of a protein formed at an early stage of its folding and preventing their unproducPresent address: G. P. Tsurupa, Laboratory of Chemical Prospecting, National Institute of Sericultural and Entomological Science, 1-2 Ohwashi, Tsukuba 305, Japan. Abbreviations used: apo-a-LA, apo-a-lactalbumin; SNase, staphylococcal nuclease; MDH, malate dehydrogenase; AAT, aspartate aminotransferase. 0022±2836/98/130733±13 $25.00/0/mb981630

tive aggregation. Among a variety of molecular chaperones, the Escherichia coli chaperonin GroE system has been best characterized physicochemically. This system comprises GroEL, an oligomeric protein of 14 identical 57 kDa subunits, and GroES, which forms a ring of seven identical subunits of 10 kDa (Viitanen et al., 1990; Schmidt et al., 1994a; Fenton & Horwich, 1997). It has been shown by electron microscopy and X-ray crystallography that GroEL consists of two heptameric rings stacked to form a cylinder with a central cavity (Ishii et al., 1992; Langer et al., 1992; Saibil et al., 1993; Braig et al., 1994, 1995; Chen et al., 1994). Each GroEL subunit consists of an equatorial domain, which is responsible for the assembly of the complex, an apical domain, which contains the binding site for a target protein and co-chaperonin GroES, and a small intermediate domain, which is # 1998 Academic Press Limited

734 mobile and connects the equatorial and apical domains (Chen et al., 1994; Fenton et al., 1994; Roseman et al., 1996; Fenton & Horwich, 1997). At present, however, there is no uni®ed scheme to explain all known experimental data concerning GroEL-mediated protein folding. It has been reported that some proteins refold in the free state in solution through repeated cycles of binding and release of the proteins in the non-native state (Todd et al., 1994; Burston et al., 1996; Katsumata et al., 1996a; Sparrer et al., 1996), while other proteins are known to refold in the GroEL-bound state (Corrales & Fersht, 1995; Itzhaki et al., 1995; Clark & Frieden, 1997). The retardation effect of GroEL on refolding kinetics is relatively small for some proteins but much more remarkable for others. What are the main factors of target proteins that give rise to such different effects of GroEL on the protein folding? It has been shown that GroEL interacts with a number of different conformational states formed in the process of protein folding: the unfolded state (Badcoe et al., 1991; Okazaki et al., 1994, 1997; Zahn et al., 1994, 1996a,b; Walter et al., 1996), the molten globule state (Goloubinoff et al., 1989; Hayer-Hartl et al., 1994; Robinson et al., 1994; Katsumata et al., 1996a,b), the more-structured folding intermediates (Gervasoni et al., 1996; Goldberg et al., 1997), and even the quaternarystructured folding intermediates (Lilie & Buchner, 1995). Therefore, the question of what conformational protein state is recognized by GroEL may not be relevant, but different proteins do appear to form transient or stable complexes with GroEL in rather different states. It is known that the binding of a polypeptide to GroEL is mediated favorably by hydrophobic interactions (Mendoza et al., 1991; Fenton et al., 1994; Itzhaki et al., 1995; Lin et al., 1995; Katsumata et al., 1996b). However, electrostatic interactions are also important for the recognition of a target protein by GroEL (Okazaki et al., 1994; Katsumata et al., 1996a,b). In previous studies, we have investigated the refolding kinetics of apo-a-lactalbumin (apo-a-LA) in the presence of GroEL (Katsumata et al., 1996a,b) and have found that GroEL binds apoa-LA with a binding constant of about 106 Mÿ1 and that there is a rapid equilibrium between binding and dissociation of apo-a-LA and GroEL. Therefore, the GroEL-affected refolding of apo-aLA occurs in the free state in solution. It has also been found that the increase of ionic strength results in an increase of the binding strength between apo-a-LA and GroEL. We attributed this result to non-speci®c screening by ions of the electrostatic repulsion between apo-a-LA and GroEL, both of which are negatively charged at neutral pH. Here, we have chosen positively charged proteins, staphylococcal nuclease (SNase) and its three mutants, as target proteins of GroEL and investigated their refolding kinetics in the presence of GroEL. The results show that the refolding kinetics of SNase are much more retarded by GroEL than

GroEL-affected Refolding of Staphylococcal Nuclease

are those of apo-a-LA. The transient folding intermediates of SNase and its three mutants bind to GroEL very tightly, so that the further refolding occurs in the GroEL-bound state. The effect of ionic strength is found to be opposite that for apo-a-LA. In spite of these differences between SNase and apo-a-LA, we show that the GroELaffected refolding reactions of both these proteins can be represented by a common uni®ed scheme, which allows us to estimate the binding constant (>109 Mÿ1) and other molecular parameters for the GroEL-affected refolding of SNase.

Results Unfolding equilibria and refolding kinetics of SNase and its mutants The equilibrium unfolding of SNase and its three mutants (P47 T/P117G, A90 S, and A69 T) induced by acidi®cation was studied by tryptophan ¯uorescence at 332 nm (excited at 295 nm) and 20 C. Figure 1 shows the unfolding transition curves given by the pH-dependence of the ¯uorescence intensity of the proteins. All the proteins are fully in the native state at pH 7.5 and fully in the acid unfolded state at pH 2.5, but they have different stabilities against the pH change. The double mutant P47 T/P117G is most stable, having a transition midpoint at pH 3.4, while the A69 T mutant is least stable with a transition midpoint at pH 4.3. The stability order of the proteins against the acid unfolding shown here is apparently the same as the stability order against the ureainduced unfolding previously reported (Kalnin & Kuwajima, 1995; Ikura et al., 1997). The kinetics of the free refolding of SNase and its mutants were studied by stopped-¯ow ¯uorescence measurements. The refolding was initiated by rapid mixing of the acid-unfolded protein solution with a ninefold volume of the refolding buf-

Figure 1. The pH dependence of the tryptophan ¯uorescence intensity at 335 nm (excited at 295 nm) for SNase and its mutants: wild-type (*), P47 T/P117G (~), A90 S (^), and A69 T (!) mutants at 20 C. The concentration of proteins was 0.2 mg/ml. All proteins are in the acid-unfolded state at pH 2.5 and in the native state at pH 7.5.

735

GroEL-affected Refolding of Staphylococcal Nuclease

Figure 2. The refolding kinetics of the SNase P47 T/P117G mutant in the absence (black) and presence of 0.59 mM (red) and 4.7 mM (green) of GroEL at pH 7.5 and 20 C. The refolding was initiated by the pHjump from pH 2.5, where the SNase was in the acid-unfolded state, to pH 7.5, where the protein was in the native state. The kinetics were followed by stopped-¯ow ¯uorescence measurements. The inset shows the data of a manual-mixing experiment of refolding kinetics of SNase P47 T/P117G mutant in the presence of 4.7 mM of GroEL. The concentration of SNase was 1 mM.

fer to give a ®nal pH of 7.5. The refolding process was followed by monitoring time-dependent changes in the tryptophan ¯uorescence. A typical kinetic trace for the P47 T/P117G mutant is shown in Figure 2. The observed kinetics were ®tted by the non-linear least-squares method with the equation X Ai eÿki t …1† A…t† ˆ A…1† ‡ i

where A(t) and A (1) were the observed ¯uorescence intensities at time t and in®nite time, respectively, and Ai and ki were the amplitude and the apparent ®rst-order rate constant, respectively, of the ith kinetic phase. The kinetic parameters of refolding for all the proteins obtained by the nonlinear least-squares ®tting are listed in Table 1. Wild-type SNase and the P47 T/P117G mutant show very similar kinetics and at least four phases each, although the kinetics of the wild-type protein can also be ®tted with the three-exponential equation because the amplitude of the third phase is suf®ciently small (4% of the total ¯uorescence change). The two single mutants also show multiphasic kinetics (at least three phases for A90 S and at least two for A69 T) but with the slower phases being more predominant. The previous studies of SNase and its proline mutants have shown that the multi-phasic refolding kinetics are due to (i) the presence of multiple parallel folding pathways for different proline isomers and (ii) slow isomerizations of proline peptide bonds in the unfolded states (Ikura et al., 1997; Walkenhorst et al., 1997).

The kinetic parameters presented here for the refolding from the acid-unfolded state for wildtype and mutant SNase are in good agreement with those previously reported for the refolding from the urea-induced unfolded state (Kalnin & Kuwajima, 1995; Ikura et al., 1997). Effect of GroEL on the kinetic refolding of SNase and its mutants To investigate the effect of GroEL on the kinetic refolding of SNase, we measured the refolding kinetics in the presence of different GroEL concentrations, both smaller and larger than the concentration equimolar to SNase. A preliminary study has shown that wild-type SNase and the P47 T/P117G mutant exhibit essentially the same behavior in the GroEL-affected kinetic refolding, so we have investigated the details of the kinetics only for the P47 T/P117G mutant. Figure 2 shows kinetic refolding curves measured by the timedependent ¯uorescence changes for the P47 T/ P117G mutant in the presence of two different concentrations of GroEL, as well as the refolding curve in the absence of GroEL. The refolding curves in the presence of GroEL in Figure 2 are those after the correction for GroEL background ¯uorescence changes (Katsumata et al., 1996a). Although GroEL itself has no tryptophan residues, it was made ¯uorescent by contamination with tryptophan-¯uorescent compounds. When the molar concentration of GroEL was a half of the concentration of P47 T/P117G (red

Table 1. Kinetic parameters of the refolding of SNase and its mutants Protein Wild-type P47T/P117G A90S A69T

k1 (sÿ1)

A1

k2 (sÿ1)

A2

k3 (sÿ1)

A3

k4 (sÿ1)

A4

14.1  0.2 16.9  0.4 ± ±

ÿ577  7 ÿ730  9 ± ±

1.77  0.12 1.78  0.13 3.20  0.54 ±

ÿ366  9 ÿ228  8 ÿ149  19 ±

0.27  0.09 0.17  0.08 0.51  0.02 0.27  0.01

ÿ44  21 ÿ76  32 ÿ708  16 ÿ569  15

0.020  0.043 0.020  0.058 0.040  0.005 0.057  0.003

ÿ122  160 ÿ76  140 ÿ279  8 ÿ458  12

736

GroEL-affected Refolding of Staphylococcal Nuclease

curve), we observed four phases in the refolding with the rate constants essentially the same as those in the absence of GroEL (black curve), but the amplitude of the fastest phase was decreased with concomitant increases in the amplitude in the slower (the third and the fourth) phases. The refolding kinetics in the presence of 4.7 mM GroEL (14-mer), which was 4.7 times the concentration of P47 T/P117G, was measured by both stopped-¯ow and manual-mixing techniques (Figure 2). In this condition, the refolding kinetics became a singlephase process, and it was retarded by three orders of magnitude as compared with the fastest phase of the free refolding in the absence of GroEL (green curve and an inset of Figure 2). The kinetic parameters of the P47 T/P117G refolding at various GroEL concentrations are summarized in Table 2. As can be seen, none of the rate constants of the four phases of refolding are affected signi®cantly by GroEL until the concentration of GroEL reaches four times that of the target protein. Only the amplitude of the fastest phase is decreased with increase in GroEL concentration, and at the same time the amplitude of the slowest phase increases. There are also changes in the amplitudes of the second and third phases. Although the detailed molecular mechanisms of the SNase folding have not yet been fully established, the major phase of refolding is known to correspond to the refolding of the major proline isomer of the protein from the partially folded state, which may be identical to the acid unfolded state, to the native state (Ikura et al., 1997). Therefore, we will only focus on the major phase (the fastest phase, in this case) of refolding in the following analysis. Table 2 shows that a third amount of GroEL already shows a signi®cant effect on the fastestphase amplitude but that the rate constant of this phase is not changed by GroEL. This indicates that the binding between GroEL and a folding intermediate of SNase is so tight as to be almost irreversible under this condition (see Discussion). The addition of the saturating concentration of GroEL (>four times that of SNase) results in a single phase with a rate constant similar to that of the fourth phase of the free refolding. This means that the intermediate trapped by GroEL refolds very slowly with a half time of about 70 seconds. If the

folding occurs only after SNase is released from GroEL, an increase in GroEL concentration must result in a decrease in the refolding rate constant. According to the law of mass action, the increase of GroEL concentration decreases the population of free SNase, leading to the decrease of the apparent refolding rate (see Gray & Fersht, 1993). Nevertheless, we have observed the same rate constant at different concentrations of excess GroEL for wildtype and P47 T/P117G SNase and for the slowfolding mutants, A90 S and A69 T, for which the GroEL binding must occur much faster than the free refolding (see below; Tables 2 to 4). This indicates that SNase refolds in the GroEL-bound state without being released in the non-native state in solution. Addition of GroEL in a concentration equimolar to that of SNase was not suf®cient to fully arrest the fastest phase of refolding of the protein. A GroEL concentration of at least four times that of SNase was required to fully arrest the reaction. There are two possible explanations for this observation. One is that one molecule of SNase binds to more than one GroEL 14-mer. The other is that the refolding of SNase (wild-type and P47 T/P117G) occurs so quickly that it competes with the association reaction between GroEL and the intermediate of SNase, and hence the equimolar concentration of GroEL cannot fully arrest the fastest refolding of the protein. In order to test these explanations, we investigated the effect of GroEL on the refolding kinetics of SNase mutants, A90 S and A69 T, that show slower refolding (Table 1). The kinetic parameters of the refolding reactions of A90 S and A69 T at various concentrations of GroEL are shown in Tables 3 and 4, respectively. From these Tables, 75 and 85% of the amplitudes of the major phase (the second phase for A90 S and the ®rst phase for A69 T) have been lost in the kinetics of A90 S and A69 T, respectively, at 0.3 mM GroEL, and in both the cases, the addition of equimolarconcentration of GroEL almost fully arrested the major phase of refolding. The results thus demonstrate that the requirement of a fourfold molar excess of GroEL to fully arrest the fastest-phase refolding of P47 T/P117G, as shown above, is in fact a result of the competition between the refolding and the target binding by GroEL.

Table 2. Kinetic parameters of the P47T/P117G mutant refolding in the presence of different GroEL concentrations [GroEL]/ [P47T/P117G] 0 0.30 0.59 0.89 1.2 1.8 2.4 3.6 4.7 a

k1 (sÿ1)

A1

k2 (sÿ1)

A2

k3 (sÿ1)

A3

k4 (sÿ1)

A4

16.9  0.4 21.8  2.1 18.9  1.0 18.3  1.9 26.3  4.3 18.2  3.0 20.3a 20.3a

ÿ730  9 ÿ622  32 ÿ563  15 ÿ500  31 ÿ436  35 ÿ398  42 ÿ258  34 ÿ154  24

1.78  0.13 2.94  0.62 1.41  0.14 3.06  0.57 2.26  0.55 1.16  0.24 2.10a 2.10a

ÿ228  8 ÿ249  24 ÿ255  13 ÿ212  28 ÿ230  26 ÿ243  32 ÿ140  21 ÿ50  15

0.17  0.08 0.59  0.10 0.15  0.05 0.18  0.03 0.28  0.10 0.16  0.08 0.26a 0.26a

ÿ76  32 ÿ225  29 ÿ111  20 ÿ210  20 ÿ166  27 ÿ158  33 ÿ100  22 ÿ3  13

0.020  0.058 0.030  0.006 0.010  0.008 0.020  0.005 0.020  0.005 0.010  0.005 0.020  0.004 0.010  0.003 0.010  0.004

ÿ76  140 ÿ105  7 ÿ163  27 ÿ238  19 ÿ356  22 ÿ433  93 ÿ746  31 ÿ658  306 ÿ657  190

The rate constants were ®xed during non-linear least-squares ®tting of the kinetic parameters

737

GroEL-affected Refolding of Staphylococcal Nuclease Table 3. Kinetic parameters of the A90S mutant refolding in the presence of different GroEL concentrations [GroEL]/[A90S] 0 0.13 0.30 0.40 0.59 0.89 1.2 2.4 a

k1 (sÿ1)

A1

k2 (sÿ1)

A2

k3 (sÿ1)

A3

3.20  0.54 5.59  0.24 4.50a 4.50a 4.50a 4.50a 4.50a

ÿ149  19 ÿ123  6 ÿ59  6 ÿ43  7 ÿ37  8 ÿ29  14 ÿ21  9

0.51  0.02 0.43  0.01 0.23  0.01 0.18  0.02 0.15  0.03 0.31  0.09 0.35a

ÿ708  16 ÿ458  5 ÿ386  10 ÿ238  22 ÿ173  27 ÿ101  11 ÿ44  4

0.040  0.005 0.029  0.001 0.026  0.001 0.028  0.002 0.020  0.002 0.018  0.002 0.011  0.001 0.010  0.008

ÿ279  8 ÿ517  4 ÿ510  8 ÿ586  18 ÿ525  16 ÿ487  16 ÿ589  33 ÿ466  31

The rate constants were ®xed during non-linear least-squares ®tting of the kinetic parameters.

The stoichiometry of binding of a target protein to GroEL can be calculated from the loss of the major phase amplitude in the refolding kinetics in the presence of GroEL. If the amplitude of the refolding phase of 1 mM SNase in the absence of GroEL is
F0 and that in the presence of a mM GroEL is
Fx, then the number of moles (n) of SNase bound per 14-mer of GroEL is given by the following equation: n ˆ (
F0 ÿ
Fx)/ a(
F0 ÿ
F1), where
F1 is the limiting amplitude at saturating GroEL and expected close to zero (Corrales & Fersht, 1995). The calculations for the two single mutants from the data of Tables 3 and 4 show that the stoichiometry tends toward two molecules of the refolding intermediate per 14-mer of GroEL; n is calculated at 1.6 to 1.8 and 1.1 to 2.1 for A90 S and A69 T, respectively, at 0.3 to 0.4 mM GroEL (Tables 3 and 4). From the results shown above, the refolding of SNase in the presence of GroEL can be represented by the following scheme

Scheme 1

where A denotes the folding intermediate of SNase, which is realized at the beginning of the refolding from the acid-unfolded state and refolds to the native state (N) with a rate constant of kf in the absence of GroEL. The binding of GroEL (G) to A occurs irreversibly with a bimolecular rate constant kb, and produces GA and GA2, in which one and two molecules, respectively, of the SNase intermediate are bound by one molecule of the GroEL 14-mer. Because the apparent refolding rate of SNase as measured by the change in ¯uorescence intensity in the presence of excess GroEL does not depend on GroEL concentration (see above), the SNase intermediate bound to GroEL refolds without dissociation, and the rate constant of this refolding is given by kg in Scheme 1. The results of Tables 2 to 4 show that the kg is about 0.01 sÿ1. It is likely that the refolding rate, kg, is different in the different complexes, GA and GA2, but this difference cannot be distinguished in the present data. On the basis of Scheme 1, it is possible to roughly estimate kb. At 0.59 mM of GroEL, the amplitude of the fastest-phase refolding of P47 T/P117G is 77% of the amplitude of the free refolding (Table 2). Therefore, if we assume that the two target-binding sites on GroEL are independent of each other, kb is given by kb  (0.23/ 0.77) kf/[g], where kf  20 sÿ1 and [g] is the concentration of the target binding site (i.e. [g] ˆ 2 [GroEL]  10ÿ6 M). The value of kb is thus estimated in the order of 107 sÿ1 Mÿ1.

Table 4. Kinetic parameters of the A69T mutant refolding in the presence of different GroEL concentrations [GroEL]/[A69T] 0 0.13 0.30 0.40 0.59 0.89 1.2 2.4 a

k1 (sÿ1)

A1

k2 (sÿ1)

A2

0.27  0.01 0.27  0.01 0.24  0.01 0.32  0.04 0.29  0.06 0.30a 0.30a

ÿ569  15 ÿ418  6 ÿ399  9 ÿ129  7 ÿ83  8 ÿ39  9 ÿ36  9

0.057  0.003 0.016  0.001 0.019  0.001 0.021  0.001 0.017  0.001 0.014  0.003 0.010  0.002 0.010  0.001

ÿ458  12 ÿ391  7 ÿ405  7 ÿ658  6 ÿ581  11 ÿ420  29 ÿ370  74 ÿ354  56

The rate constants were ®xed during non-linear least-squares ®tting of the kinetic parameters.

738

GroEL-affected Refolding of Staphylococcal Nuclease

Table 5. Kinetic parameters of GroEL-affected refolding of the P47T/P117G mutant in the presence of Mg-ATP [GroEL]/ [P47T/P117G] 0 0.59 1.2 2.4 4.7

k1 (sÿ1)

A1

k2 (sÿ1)

A2

k3 (sÿ1)

A3

k4 (sÿ1)

A4

12.6  0.5 26.1  4.3 19.0  6.7

ÿ1002  27 ÿ378  35 ÿ195  42

2.30  0.18 1.10  0.18 2.76  0.46

ÿ488  24 ÿ370  48 ÿ351  35

0.19  0.07 0.19  0.02 0.17  0.01 0.10  0.01 0.11  0.01

ÿ131  29 ÿ837  31 ÿ992  29 ÿ1134  116 ÿ1057  89

0.020  0.050 0.010  0.007 0.020  0.004 0.010  0.010 0.011  0.025

ÿ161  55 ÿ330  68 ÿ441  25 ÿ583  29 ÿ187  120

Activity recovery of SNase refolded in the GroEL-bound state In Scheme 1, we have assumed that SNase refolds in the GroEL-bound state, then the refolded native SNase is released from GroEL. In order to con®rm the validity of this assumption, we performed the following activity recovery experiments. First, the SNase (P47 T/P117G) refolding was started in the presence of a twice molar excess of GroEL. At 10, 30, and 240 seconds after the refolding was started, an aliquot of the refolding reaction mixture was added to the activity assay solution that contained GroEL in the same concentration as in the refolding reaction mixture. Because the GroEL concentration in solution did not change, there was no dissociation due to dilution of the refolding reaction mixture into the activity assay solution. The results of the experiments have shown that all the activity curves obtained after the different times of refolding were nearly parallel with the control curve that had been obtained by the addition of native SNase to the activity assay solution. The activity curves indicated the presence of a lag phase, which lasted for about 100 seconds, in the time course of the activity assay, and the refolding of SNase suf®ciently proceeded during this lag phase because the lag phase was longer than the half time (70 seconds) of the SNase refolding in the GroELbound state (data not shown). Essentially identical results were obtained for wild-type and A69 T mutant SNase. These activity recovery results together with the kinetic refolding data clearly indicate that SNase refolds in the GroEL-bound state and that the refolded SNase is fully active in the DNA hydrolysis assay. Effect of ATP and cochaperonin GroES ATP alone is known to be suf®cient to induce productive release of some target proteins (Laminet et al., 1990; Viitanen et al., 1991; Schmidt & Buchner, 1992; Kubo et al., 1993), although the

full chaperonin action of GroEL requires both ATP and GroES (Viitanen et al., 1990; Mendoza et al., 1991; Brandsch et al., 1992; Zheng et al., 1993; Corrales & Fersht, 1996). To investigate the effect of ATP on the GroEL-affected refolding of SNase, we measured the refolding kinetics of P47 T/ P117G in the presence of ATP at various concentrations of GroEL, and obtained the results in Table 5. It appears that ATP does not lead to essential change in the scheme of the refolding of SNase in the presence of GroEL. The amplitude of the fastest phase decreases with an increase of GroEL concentration, but the refolding rate constants do not change. However, in the presence of ATP, the increase in the amplitude by GroEL occurs mainly in the third phase, and at a saturating concentration of GroEL (two to ®ve times the concentration of SNase) the major phase has a rate constant of about 0.1 sÿ1 (Table 5). This means that the SNase refolding in the GroEL-bound state, at a saturation concentration of GroEL, is accelerated about ten times from 0.01 to 0.1 sÿ1 by ATP. Because the free refolding of SNase was not signi®cantly affected by ATP (data not shown), the acceleration of the refolding was caused by binding and/or hydrolysis of ATP on GroEL (see Discussion). The addition of all components of the GroE system (GroEL, GroES, and Mg-ATP) restored the characteristics of the free refolding of SNase and its mutants (data not shown). Effect of ionic strength Our previous studies of the GroEL-affected refolding kinetics of a-lactalbumin have shown that electrostatic screening is an important factor to determine the binding strength between GroEL and a target protein (Katsumata et al., 1996b). To investigate the effect of the electrostatic interactions on the binding between GroEL and SNase, we studied the refolding kinetics of SNase in the GroEL-bound state at various concentrations of NaCl (the concentrations of SNase and GroEL

Table 6. In¯uence of ionic strength on the kinetic parameters of GroEL-affected refolding of the P47T/P117G mutant Ionic strength (mM) 10 50 150

k1 (sÿ1)

1.67  0.27

A1

k2 (sÿ1)

A2

ÿ153  12

0.05  0.01 0.08  0.01 0.11  0.01

ÿ343  24 ÿ433  11 ÿ361  9

739

GroEL-affected Refolding of Staphylococcal Nuclease

were 1 and 2.6 mM, respectively; Table 6). Similar results were obtained for the two single mutants (data not shown). It was found that the refolding is accelerated by about two orders of magnitude when the ionic strength is increased from 10 mM to 150 mM and that this effect is independent of the ion species used, because different salts (NaCl, KCl, and K2SO4) were found to be equally effective when the ionic strength was kept constant (data not shown). The free refolding kinetics of wildtype and mutant SNase in the absence of GroEL were not affected by salts, so that the acceleration of the refolding was caused by the effect of electrostatic interactions on the GroEL-SNase binding. This effect was the opposite of that previously observed in the a-lactalbumin refolding (Katsumata et al., 1996b), a difference possibly attributable to the difference in net charge between SNase and a-lactalbumin (see Discussion).

at 0.010 sÿ1, indicating that the refolding occurs in the GroEL-bound state and that the refolding rate in the bound state is identical for wild-type SNase and the three mutants studied. The above scheme of the SNase refolding in the presence of GroEL is apparently in contrast to that of the GroEL-affected refolding kinetics of apo-a-LA, as reported in our previous studies (Katsumata et al., 1996a). In apo-a-LA, the increase of GroEL concentration results in a decrease of the refolding rate without any change in the amplitude of the observed kinetics. This means that the binding of apo-a-LA to GroEL is reversible and that apo-a-LA refolds in the free state after dissociation from GroEL. The refolding of apo-a-LA in the presence of GroEL can thus be represented by Scheme 2.

Discussion We have studied the effect of GroEL on the refolding kinetics of SNase and its three mutants and have found that GroEL interacts strongly with a transient folding intermediate and that the refolding occurs in the GroEL-bound state. The refolding rate of the GroEL-bound protein is retarded more than 1000-fold in the case of wildtype SNase and the P47 T/P117G mutant, and 30 to 50-fold in the case of the A90 S and A69 T mutants. These results appear to be the reverse of our previous results on the GroEL-affected kinetic refolding of apo-a-LA, in which the refolding occurred in the free state with reversible association-dissociation between apo-a-LA and GroEL (Katsumata et al., 1996a). Thus, we will ®rst examine the differences between the GroEL-affected kinetic refoldings of SNase and apo-a-LA and discuss schematic models that represent the refolding reactions of these proteins in the presence of GroEL. Then, we will discuss details of the parameters required for describing the GroEL-affected refolding of SNase, i.e. binding parameters between GroEL and SNase, effect of ionic strength on the refolding, and effect of ATP and GroES on the refolding. The GroEL-affected kinetic refolding: schematic models and computer simulations The present results are interpreted in terms of Scheme 1. The data of the kinetic experiments have shown that the rate constant of the major refolding phase in SNase and its mutants is not changed by GroEL, while the amplitude of the major phase is decreased with an increase in GroEL concentration. This decrease in amplitude has been observed at a GroEL concentration much smaller than the concentration equimolar to SNase, indicating that the binding of the SNase intermediate to GroEL is very tight and almost irreversible. In the presence of a large excess of GroEL, the refolding rate levels off

Scheme 2

Here, we have also assumed that a maximum of two molecules of apo-a-LA are bound to GroEL, following the stoichiometry of binding of SNase to GroEL found in this study, although a one to one ([GroEL] : [apo-a-LA]) binding was assumed in the previous study. Because there is a rapid equilibrium between the binding and dissociation of apo-a-LA and GroEL, the binding constant Kb of the refolding intermediate (A) to GroEL (G) can be obtained from the GroEL-affected refolding kinetics of apo-a-LA, and has been estimated at 106 Mÿ1 (Katsumata et al., 1996a). In order to interpret the refolding behavior of both SNase and apo-a-LA in the presence of GroEL, we propose the following uni®ed scheme (Scheme 3).

Scheme 3

740

GroEL-affected Refolding of Staphylococcal Nuclease

Figure 3. The simulations of the effect of GroEL on the apparent refolding kinetics of a target protein. Scheme 3 was used in this model. (a and b) The case of rapid equilibrium between the binding and dissociation of GroEL and the target protein, and with the refolding occurring in the free state in solution. One mM concentration of the target protein was used, along with the following rate constants: kf ˆ 0.06 sÿ1, kb ˆ 106 sÿ1 Mÿ1, kd ˆ 1 sÿ1, kg ˆ 10ÿ2 sÿ1. (c and d) The case of the practically irreversible binding of the target protein to GroEL. One mM concentration of the target protein was used, along with the following rate constants: kf ˆ 20 sÿ1, kb ˆ 107 sÿ1 Mÿ1, kd ˆ 10ÿ3 sÿ1, kg ˆ 10ÿ2 sÿ1.

Here, kd is the dissociation constant of the GroEL-bound protein, and, as in Scheme 1, we assume that kb, as well as kd and kg, are the same for GA and GA2 (we cannot distinguish the differences by the present experiments). When, in Scheme 3, kd4kg and kd4kf, the effect of GroEL on the refolding kinetics of a target protein becomes equivalent to that seen in apo-a-LA refolding (Scheme 2). When kd5 kg, the refolding behavior in the presence of GroEL is identical to that found in SNase refolding (Scheme 1). The refolding behavior of a target protein in the presence of GroEL depends on the relations among the kinetic parameters kf , kb , kd, and kg. To test this model, we next carried out computer simulations of the GroEL-affected refolding kinetics of model proteins in accordance with Scheme 3 (see Materials and Methods). The results are illustrated in Figure 3, in which the effects of GroEL concentration on the apparent rate constants and on the amplitudes of refolding are shown for the following two cases: (1) refolding

occurring in the free state with rapid equilibrium of reversible binding and dissociation between the target protein and GroEL ((kb[g] ‡ kd)4(kf ‡ kg)); and (2) practically irreversible binding of the target protein to GroEL (kb[g]4kd) and the refolding in the GroEL-bound state (kg4kd). In the ®rst case (Figure 3 (a) and (b)), there is apparently only one kinetic phase, and the apparent rate constant kapp of this phase decreases with an increase of GroEL concentration and comes close to the value of kg at a high concentration of GroEL. In the second case (Figure 3(c) and (d)), there are practically two observable kinetic phases that have rate constants of kf and kg, because we have assumed that the binding of the folding intermediate to GroEL is not accompanied by the ¯uorescence change and occurs much faster than the refolding (kb[g]4kf). We have also assumed that dissociation of native SNase is not included in the observed kinetics because the native protein is expected to dissociate quickly. The amplitude of the fast phase that has the rate constant kf decreases and the amplitude of

741

GroEL-affected Refolding of Staphylococcal Nuclease

the slow phase that has the rate constant kg increases with increasing GroEL concentration. Interestingly, when the rate of binding of the intermediate to GroEL is comparable to the free refolding rate (kb[g]  kf), the apparent rate of the fast phase increases with an increase of GroEL concentration as if GroEL were accelerating the refolding. This apparent rate increase is not caused by the acceleration but by a rapid consumption of the free refolding species at the rate constant kb[g]. For wild-type SNase and P47 T/P117G mutant, kb[g] is comparable to kf, but this apparent rate increase is not clear in the experimentally observed rate constants (Table 2), since the amplitude of the fast phase becomes too small when the apparent rate is signi®cantly increased. As a result, the experimentally observed refolding kinetics of apo-a-LA and SNase in the presence of GroEL are in good accordance with the simulation results, and the kinetics of apo-a-LA and SNase correspond to the ®rst and second cases, respectively, in the simulations. Stoichiometry of the target binding by GroEL From the kinetic data for two single mutants, A90 S and A69 T, we have calculated the stoichiometry of binding of the folding intermediates to GroEL as two ([SNase]/[GroEL]) at a GroEL concentration of 0.3 to 0.4 mM ([SNase] ˆ 1.0 mM; Tables 3 and 4). The stoichiometry of binding to GroEL has been reported as one for some proteins (Corrales & Fersht, 1995; Itzhaki et al., 1995; Lin et al., 1995) and two for others. The presence of two binding sites on GroEL has been shown for maltose-binding protein (Sparrer et al., 1996) and eukaryotic DHFR (Clark et al., 1996; Clark & Frieden, 1997). A recent study has suggested that the football-shaped GroEL-GroES complex having GroES on both sides of GroEL can accommodate and handle two target proteins simultaneously (Sparrer et al., 1997; Llorca et al., 1997). Thus, the binding by GroEL of two molecules of the target protein is probably more common than previously thought. Parameters for the GroEL-affected refolding Scheme 3 allows us to estimate the equilibrium binding constant Kb (ˆ kb/kd) between SNase and GroEL, as well as the other parameters for the GroEL-affected refolding. We have found that the refolding is in competition with the binding by GroEL for wild-type SNase and the P47 T/P117G mutant, and on the basis of Scheme 1, we have estimated kb in the order of 107 Mÿ1 sÿ1. From the kinetic data for the presence of a large excess of GroEL, the kg is estimated at 10ÿ2 sÿ1, so that the kd must be smaller than 10ÿ2 sÿ1. Taken together, these parameter values indicate that the binding constant Kb between the SNase folding intermediate and GroEL must be larger than 109 Mÿ1. This value is about three orders of magnitude larger

than that for apo-a-LA (Kb ˆ 106 Mÿ1). On the other hand, the kb value for apo-a-LA was previously estimated at 106 Mÿ1 sÿ1, so that the kd must be about 1 sÿ1. These parameter values were used in the computer simulations shown above. It has been reported that barnase refolds in the GroEL-bound state (Gray & Fersht, 1993). For this protein, kb and kg were estimated at 3.5  107 Mÿ1 sÿ1 and 10ÿ2 sÿ1, respectively (Corrales & Fersht, 1995). From the reported kinetic data, it is clear that kd must be smaller than 10ÿ2 sÿ1 for barnase, so that Kb is estimated in the order of 3.5  109 Mÿ1. These values (Kb, kb, kd, and kg) are thus very similar to those found for SNase here. The kinetic parameters of the GroEL-affected refolding were also reported for maltose-binding protein (Sparrer et al., 1996). This protein has a Kb value of about 1011 Mÿ1 for its fast phase and of about 1010 Mÿ1 for its slow phase of refolding, a kd of 10ÿ3 ÿ 10ÿ2 sÿ1, and a kb of about 107 ÿ 108 Mÿ1 sÿ1. These values thus indicate that maltose-binding protein is more strongly bound to GroEL than SNase. However, maltose-binding protein refolds in the free state with repeated cycles of the binding and release of the folding intermediate, because kd is larger than kg. It is also noted that the kg values for wild-type SNase and the three different mutants are identical even though their intrinsic refolding rates in the free state differ by two orders of magnitude. This means that the molecular mechanism controlling the refolding of the target proteins in the GroELbound state is different from that of the free refolding but may relate to a structural dynamic phenomenon of the GroEL molecules. Effect of ionic strength The present results show that the kb for SNase is one order of magnitude larger and the kd is two orders of magnitude smaller than the corresponding values for apo-a-LA. What causes these differences between SNase and apo-a-LA? We have assumed that the difference in net charge results in the different kb and kd values between the proteins because the electrostatic interactions are expected to play important roles in the target recognition by GroEL. At neutral pH, apo-a-LA is negatively charged with a net charge of ÿ7, SNase is positively charged with a net charge of ‡12, and GroEL is highly negatively charged with a net charge of ÿ18 per monomer. Therefore, there must exist an electrostatic repulsion between apo-a-LA and GroEL and an electrostatic attraction between SNase and GroEL. Because of the electrostatic attraction, SNase must be more strongly recognized by GroEL than is apo-a-LA, and the effect of ionic strength on the GroEL-affected refolding is expected to be the opposite of that previously observed for apo-a-LA. In order to test the above assumption, we have investigated the effects of ionic strength on the GroEL-affected refolding of SNase, and compared

742 the results with those for apo-a-LA. Our previous studies on apo-a-LA have shown that an increase in ionic strength results in an increase in binding strength between the apo-a-LA folding intermediate and GroEL, which ®nally leads to more effective retardation of the refolding by GroEL (Katsumata et al., 1996b). We interpreted this effect in terms of the electrostatic screening by ions, which decreases the repulsion between the two highly negatively charged protein molecules, apoa-LA and GroEL (Katsumata et al., 1996a). Here, we have observed that the increase in ionic strength results in the opposite effect for SNase and accelerates the rate of the refolding in the presence of GroEL (Table 6). These results for SNase and apo-a-LA thus provide clear evidence that the electrostatic interactions are important for the target recognition by GroEL. It is known that GroEL-target complex derives its stability from non-speci®c hydrophobic interactions (Martin et al., 1991; Mendoza et al., 1991; Landry et al., 1992; Fenton et al., 1994; Itzhaki et al., 1995; Lin et al., 1995). Thus, the stronger binding of target proteins to GroEL previously observed under conditions of increased salt concentrations has often been attributed to the intensi®cation of hydrophobic interactions by salt, and it has also been proposed that salt increases the accessible hydrophobic surface of the GroEL oligomer (Horowitz et al., 1995). Although hydrophobic interactions are dominant in the target recognition by GroEL, some of the data previously explained by the salt-induced intensi®cation of hydrophobic interactions may simply be due to the electrostatic screening effect. Studies on the refolding of chymotrypsin inhibitor 2 and its mutants in the presence of GroEL have also shown that hydrophobic and positively charged side-chains tend to interact favorably with GroEL, whereas negatively charged side-chains tend to repel GroEL (Itzhaki et al., 1995). Maltose-binding protein is negatively charged (a net charge of ÿ7 at neutral pH), and the increase in ionic strength results in stronger binding between this protein and GroEL (Sparrer et al., 1996). It has been reported that isozymes of malate dehydrogenase (MDH; Staniforth et al., 1994) and aspartate aminotransferase (AAT; Mattingly et al., 1995; Widmann & Christen, 1995) have different af®nities for GroEL. It appears that this fact can also be interpreted in terms of differences in net charge. Mitochondrial MDH has a net charge of ‡3 and binds more strongly to GroEL than its cytosolic homolog, which has a net charge of ÿ26. In the case of AAT isozymes, mitochondrial AAT, having a net charge of ‡8, binds more strongly than the cytosolic homolog (net charge of ‡2) or AAT from E. coli (net charge of ÿ8). Effect of ATP and the cochaperonin GroES Here, we have found that addition of ATP increases the refolding rate of SNase in the GroELbound state to as high as 0.1 sÿ1, although this rate

GroEL-affected Refolding of Staphylococcal Nuclease

is still much less than the rate of the free refolding in solution. It is interesting to note that the observed rate constant in the presence of 5 mM ATP is the same as the turnover rate constant of ATP hydrolysis of GroEL (Yifrach & Horovitz, 1995). It is thus very likely that the ATP hydrolysis is a rate-limiting step in the refolding of the target protein in the GroEL-bound state. Our data also show that ATP alone is not able to restore the free refolding rate of SNase. The free refolding rate was restored in the presence of the cochaperonin GroES and ATP. It has been reported that ATP binding alone is suf®cient to trigger a conformational switch that leads to a low-af®nity state of GroEL for the target protein (Yifrach & Horovitz, 1995; Sparrer et al., 1996; Aharoni & Horovitz, 1997) and is sometimes suf®cient to release the target protein in solution (Mizobata et al., 1992; Schmidt et al., 1994a; Fenton & Horwich, 1997). In the case of SNase and its mutants, the low-af®nity state may maintain the complex. Usually, there is a correlation between the overall complex stability and the conditions for ef®cient release and refolding of the GroEL-bound proteins, and the release of tightly bound ligands requires the presence of cochaperonin GroES (Schmidt et al., 1994b; Fenton & Horwich, 1997). Our data on the effects of ATP and cochaperonin GroES on the refolding of SNase and its mutants con®rm this conclusion.

Materials and Methods Expression and purification of proteins The expression plasmid containing the mutation P47 T/P117G was constructed using the MUTA-GENE phagemid in vitro mutagenesis kit (BIO-RAD; Kunkel, 1985) from the expression plasmid pMT7-SN-P117G containing the T7 promoter, the M13 intergenic region, and a gene of the P117G mutant of SNase (Ikura et al., 1997). The nucleotide sequence of the mutant plasmid was identi®ed by an autosequencer ALF express (Pharmacia). The A90 S and A69 T mutants of SNase were expressed and puri®ed from the E. coli AR120 strain carrying recombinant plasmids (pL9) that contained the mutant genes (Kalnin & Kuwajima, 1995); the E. coli cells carrying the plasmids were a gift from D. Shortle (Shortle & Meeker, 1989). Wild-type and P47 T/P117G mutant of SNase were expressed and puri®ed from E. coli cells BL21(DE3)/pLysS that contained expression plasmids pMT7-SN and pMT7-SN-P47 T/P117G (Ikura et al., 1997). GroEL was prepared from E. coli TG1 cells that contained the expression plasmid pKY206, which was a gift from K. Ito (Ito & Akiyama, 1991). The puri®cation was done as described previously (Katsumata et al., 1996a). The concentration of SNase was determined spectrophotometrically at 280 nm with an extinction coef®cient of E1% 1 cm ˆ 9.3. The concentration of GroEL, which refers to the 14-mer, was determined spectrophotometrically using an extinction coef®cient of E1% 1 cm ˆ 2.4 at 280 nm. The extinction coef®cient of GroEL was the same as that previously used (Katsumata et al., 1996a,b) and signi®cantly larger than the extinction coef®cient (E1% 1 cm ˆ 1.8) expected from the amino acid composition (seven tyro-

743

GroEL-affected Refolding of Staphylococcal Nuclease sine residues and no tryptophan per monomer) of GroEL (Pace et al., 1995). The difference of the extinction coef®cient from the expected value may be due to contamination with tryptophan containing compounds. More recent preparations of our GroEL by repeated cycles of ATP treatments and ion-exchange chromatography have an extinction coef®cient of E1% 1 cm ˆ 2.1 at 280 nm, and preliminary studies have shown that the effects of GroEL on the refolding kinetics of target proteins are identical between the different preparations of GroEL (T. M. et al., unpublished data). Equilibrium experiments Equilibrium ¯uorescence intensity was measured in a Jasco FP-777 spectro¯uorometer at 20 C. The concentration of SNase and its mutants was 0.2 mg/ml in solution that contained 50 mM Na-cacodylate, 50 mM NaCl, 2 mM EDTA, and various amount of HCl. The ¯uorescence intensity excited at 295 nm was measured at 332 nm with a 1-cm cell, and the spectral bandwidth was 1.5 nm for both excitation and emission. Stopped-flow fluorescence measurement All the refolding kinetics were measured at 20 C with a Unisoku stopped-¯ow ¯uorometer (specially designed by Unisoku company). The refolding reaction was followed by monitoring time-dependent changes in the tryptophan ¯uorescence intensity. The excitation wavelength was 295 nm and the emission light was ®ltered through a SC32 ®lter to cut off the excitation light below 320 nm. In the refolding of SNase and its mutants, the protein (10 mM) was ®rst acid denatured at pH 2.5 in solution A that contained 20 mM MgCl2, 50 mM KCl, and an appropriate concentration of HCl. The refolding was initiated by rapid mixing of SNase in solution A with ninefold volume of the refolding buffer (10 mM Na-cacodylate, pH 8.0) to give a ®nal pH of 7.5. To investigate the effect of GroEL on the refolding kinetics of SNase, the refolding buffer contained an appropriate amount of GroEL, and the ®nal pH was 7.5. For measurement of the effects of ATP and cochaperonin GroES, the refolding buffer also contained 5 mM ATP and/or GroES at a concentration twice that of GroEL. We measured the refolding kinetics of SNase in the presence of GroEL at various ionic strengths. Various concentrations of salt species (NaCl, KCl and K2SO4) were added into the refolding buffer in these measurements. SNase activity measurement Enzyme activity measurements of SNase and its mutants were made by the method of Cuatrecasas et al. (1967), which was based on increase of the absorbance at 260 nm due to hydrolysis of nucleic acids by SNase. For the measurements during the refolding of SNase in the presence of GroEL, ®rst, the refolding reaction was initiated by mixing the solution A, which contained 10 mM of SNase, with ninefold volume of the refolding buffer that contained 2.2 mM of GroEL. After 10, 30, and 240 seconds, 25 ml of the GroEL-affected SNase refolding reaction mixture was added to 2.5 ml of the assay solution that contained 50 mg/ml of boiled DNA from salmon testes (Sigma Co.), 10 mM CaCl2, and 2.2 mM of GroEL in the refolding buffer. The assay was done at 20 C

Computer simulations for the GroEL-affected refolding For the computer simulations of the refolding of a target protein in the presence of GroEL, Scheme 3 was rewritten with the differential equations such that d‰NŠ ˆ kf ‰AŠ ‡ kg ‰gAŠ dt d‰AŠ ˆ ÿkf ‰AŠ ÿ kb ‰gŠ‰AŠ ‡ kd ‰gAŠ dt d‰gŠ ˆ ÿkb ‰gŠ‰AŠ ‡ kd ‰gAŠ ‡ kg ‰gAŠ dt d‰gAŠ ˆ kb ‰gŠ‰AŠ ÿ kd ‰gAŠ ÿ kg ‰gAŠ; dt where [g] is the concentration of stoichiometric binding sites of GroEL, which is twice the molar concentration of GroEL in Scheme 3, and [gA] is the concentration of the binding site occupied by the folding intermediate (A) of the target protein. We assume for convenience that the binding reactions of the target protein to the two stoichiometric binding sites of GroEL occur independently. These differential equations were solved numerically by the Runge-Kutta method (Press et al., 1988) using kinetic parameters shown in the corresponding Figure (Figure 3). The apparent rate constants of refolding of the target protein were obtained by non-linear least-squares ®tting to equation (1) of the curves obtained by the simulations.

Acknowledgments The authors thank Professor David Shortle and Dr Alan Meeker for the gift of strains of overproducing the A69 T and A90 S mutants of SNase, and Professor Koreaki Ito for the gift of the plasmid pKY206. They also thank Mr K. Maki and Professor H. Kataoka for their HPLC and amino acid sequence analyses. This work was supported by Grants-in-Aid for Scienti®c Research from the Ministry of Education, Culture and Science of Japan and by a grant (RG-331/93 M) from the International Human Frontier Science Program (HFSP) Organization. G.P.T. was a Postdoctoral Fellow of the Japan Society for the Promotion of Science.

References Aharoni, A. & Horovitz, A. (1997). Detection of changes in pairwise interactions during allosteric transitions: coupling between local and global conformational changes in GroEL. Proc. Natl Acad. Sci. USA, 94, 1698± 1702. Badcoe, I. G., Smith, C. J., Wood, S., Halsall, D. J., Holbrook, J. J., Lund, P. & Clarke, A. R. (1991). Binding of a chaperonin to the folding intermediates of lactate dehydrogenase. Biochemistry, 30, 9195± 9200. Braig, K., Otwinowski, Z., Hegde, R., Boisvert, D. C., Joachimiak, A., Horwich, A. L. & Sigler, P. B. (1994). The crystal structure of the bacterial chaperÊ . Nature, 371, 578±586. onin GroEL at 2.8 A Braig, K., Adams, P. D. & BruÈnger, A. T. (1995). Conformational variability in the re®ned structure Ê resolution. Nature of the chaperonin GroEL at 2. 8 A Struct. Biol. 2, 1083± 1094.

744 Brandsch, R., Bichler, V., Schmidt, M. & Buchner, J. (1992). GroE dependence of refolding and holoenzyme formation of 6 ± hydroxy-D-nicotine oxidase. J. Biol. Chem. 267, 20844± 20849. Buchner, J. (1996). Supervising the fold: functional principles of molecular chaperones. FASEB J. 10, 10± 19. Burston, S. G., Weissman, J. S., Farr, G. W., Fenton, W. A. & Horwich, A. L. (1996). Release of both native and non-native proteins from a cis-only GroEL ternary complex. Nature, 282, 96 ±99. Chen, S., Roseman, A. M., Hunter, A. S., Wood, S. P., Burston, S. G., Ranson, N. A., Clarke, A. R. & Salbil, H. R. (1994). Location of a folding protein and shape changes in GroEL-GroES complexes imaged by cryo-electron microscopy. Nature, 371, 261 ± 264. Clark, A. C. & Frieden, C. (1997). GroEL-mediated folding of structurally homologous dihydrofolate reductases. J. Mol. Biol. 268, 512± 525. Clark, A. C., Hugo, E. & Frieden, C. (1996). Determination of regions in the dihydrofolate reductase structure that interact with the molecular chaperonin GroEL. Biochemistry, 35, 5893± 5901. Corrales, F. J. & Fersht, A. R. (1995). The folding of GroEL-bound barnase as a model for chaperoninmediated protein folding. Proc. Natl Acad. Sci. USA, 92, 5326± 5330. Corrales, F. J. & Fersht, A. R. (1996). Toward a mechanism for GroEL. GroES chaperone activity: an ATPase-gated and -pulsed folding and annealing cage. Proc. Natl Acad. Sci. USA, 93, 4509± 4512. Cuatrecasas, P., Fuchs, S. & An®nsen, C. B. (1967). Catalytic properties and speci®city of the extracellular nuclease of Staphylococcus aureus. J. Biol. Chem. 242, 1541± 1547. Fenton, W. A. & Horwich, A. L. (1997). GroEL-mediated protein folding. Protein Sci. 6, 743± 760. Fenton, W. A., Kashi, Y., Furtak, K. & Horwich, A. L. (1994). Residues in chaperonin GroEL required for polypeptide binding and release. Nature, 371, 614± 619. Gervasoni, P., Staudenmann, W., James, P., Gehrig, P. & Plueckthun, A. (1996). b-Lactamase binds to GroEL in a conformation highly protected against hydrogen/deuterium exchange. Proc. Natl Acad. Sci. USA, 93, 12189± 12194. Gething, M. J. & Sambrook, J. (1992). Protein folding in the cell. Nature, 355, 33±45. Goldberg, M. S., Zhang, J., Sondek, S., Matthews, C. R., Fox, R. O. & Horwich, A. L. (1997). Native-like structure of a protein-folding intermediate bound to the chaperonin GroEL. Proc. Natl Acad. Sci. USA, 94, 1080± 1085. Goloubinoff, P., Christeller, J. T., Gatenby, A. A. & Lorimer, G. H. (1989). Reconstitution of active dimeric ribulose bisphosphate carboxylase from an unfolded state depends on two chaperonin proteins and Mg-ATP. Nature, 342, 884± 889. Gray, T. E. & Fersht, A. R. (1993). Refolding of barnase in the presence of GroE. J. Mol. Biol. 232, 1197± 1207. Hartl, F. U. (1996). Molecular chaperones in cellular protein folding. Nature, 381, 571± 580. Hayer-Hartl, M. K., Ewbank, J. J., Creighton, T. E. & Hartl, F. U. (1994). Conformational speci®city of the chaperonin GroEL for the compact folding intermediates of a-lactalbumin. EMBO J. 13, 3192± 3202. Horowitz, P. M., Hua, S. & Gibbons, D. L. (1995). Hydrophobic surfaces that are hidden in chaperonin

GroEL-affected Refolding of Staphylococcal Nuclease Cpn60 can be exposed by formation of assemblycompetent monomers or by ionic perturbation of the oligomer. J. Biol. Chem. 270, 1535± 1542. Ikura, T., Tsurupa, G. P. & Kuwajima, K. (1997). Kinetic folding and cis/trans prolyl isomerization of staphylococcal nuclease. A study by stopped-¯ow absorption, stopped-¯ow circular dichroism and molecular dynamics simulations. Biochemistry, 36, 6529± 6538(Corrections. Biochemistry, 36, 11050).. Ishii, N., Taguchi, H., Sumi, M. & Yoshida, M. (1992). Structure of holo-chaperonin studied with electron microscopy: oligomeric cpn10 on top of two layers of cpn60 rings with two stripes each. FEBS Letters, 299, 169± 174. Ito, K. & Akiyama, Y. (1991). In vivo analysis of integration of membrane proteins in Escherichia coli. Mol. Microbiol. 5, 2243± 2253. Itzhaki, L. S., Otzen, D. E. & Fersht, A. R. (1995). Nature and consequences of GroEL-protein interactions. Biochemistry, 34, 14581± 14587. Kalnin, N. N. & Kuwajima, K. (1995). Kinetic folding and unfolding of staphylococcal nuclease and its six mutants studied by stopped-¯ow circular dichroism. Proteins: Struct. Funct. Genet. 23, 163 ± 176. Katsumata, K., Okazaki, A. & Kuwajima, K. (1996a). Effect of GroEL on the refolding kinetics of alactalbumin. J. Mol. Biol. 258, 827 ± 838. Katsumata, K., Okazaki, A., Tsurupa, G. P. & Kuwajima, K. (1996b). Dominant forces in the recognition of a transient folding intermediate of a-lactalbumin by GroEL. J. Mol. Biol. 264, 643± 649. Kubo, T., Mizobata, T. & Kawata, Y. (1993). Refolding of yeast enolase in the presence of the chaperonin GroE. The nucleotide speci®city of GroE and the role of GroES. J. Biol. Chem. 268, 19346± 19351. Kunkel, T. A. (1985). Rapid and ef®cient site-speci®c mutagenesis without phenotypic selection. Proc. Natl Acad. Sci. USA, 82, 488 ± 492. Laminet, A. A., Ziegelhoffer, T., Georgopoulos, C. & PluÈckthun, A. (1990). The Escherichia coli heat shock proteins GroEL and GroES modulate the folding of the b-lactamase precursor. EMBO J. 9, 2315± 2319. Landry, S. J., Jordan, R., McMacken, R. & Gierasch, L. M. (1992). Different conformations for the same polypeptide bound to chaperones DnaK and GroEL. Nature, 355, 455 ± 457. Langer, T., Pfeifer, G., Martin, J., Baumeister, W. & Hartl, F. U. (1992). Chaperonin-mediated protein folding: GroES binds to one end of the GroEL cylinder, which accommodates the protein substrate within its central cavity. EMBO J. 11, 4757 ±4765. Lilie, H. & Buchner, J. (1995). Interaction of GroEL with a highly structured folding intermediate: iterative binding cycles do not involve unfolding. Proc. Natl Acad. Sci. USA, 92, 8100± 8104. Lin, Z., Schwarz, F. P. & Eisenstein, E. (1995). The hydrophobic nature of GroEL-substrate binding. J. Biol. Chem. 270, 1011±1014. Llorca, O., Marco, S., Carrascosa, J. L. & Valpuesta, J. M. (1997). Symmetric GroEL-GroES complexes contain substrate simultaneously in both GroEL rings. FEBS Letters, 405, 195±199. Martin, J., Langer, T., Boteva, R., Schramel, A., Horwich, A. L. & Hartl, F. U. (1991). Chaperonin-mediated protein folding at the surface of groEL through a ``molten globule''-like intermediate. Nature, 352, 36±42.

GroEL-affected Refolding of Staphylococcal Nuclease Mattingly, J. R., Jr, Iriarte, A. & Martinez-Carrion, M. (1995). Homologous proteins with different af®nities for groEL. The refolding of the aspartate aminotransferase isozymes at varying temperatures. J. Biol. Chem. 270, 1138± 1148. Mendoza, J. A., Rogers, E., Lorimer, G. H. & Horowitz, P. M. (1991). Chaperonins facilitate the in vitro folding of monomeric mitochondrial rhodanese. J. Biol. Chem. 266, 13044± 13049. Mizobata, T., Akiyama, Y., Ito, K., Yumoto, N. & Kawata, Y. (1992). Effects of the chaperonin GroE on the refolding of tryptophanase from Escherichia coli. Refolding is enhanced in the presence of ADP. J. Biol. Chem. 267, 17773. Okazaki, A., Ikura, T., Nikaido, K. & Kuwajima, K. (1994). The chaperonin GroEL does not recognize apo-a-lactalbumin in the molten globule state. Nature Struct. Biol. 1, 439± 446. Okazaki, A., Katsumata, K. & Kuwajima, K. (1997). Hydrogen-exchange kinetics of reduced a-lactalbumin bound to the chaperonin GroEL. J. Biochem. (Tokyo), 121, 534± 541. Pace, C. N., Vajdos, F., Fee, L., Grimsley, G. & Gray, T. (1995). How to measure and predict the molar absorption coef®cient of a protein. Protein Sci. 4, 2411± 2423. Press, W. H., Flannery, B. P., Teukolsky, S. A. & Vetterling, W. T. (1988). Numerical Recipes in C, Cambridge University Press. Robinson, C. V., Gross, M., Eyles, S. J., Ewbank, J. J., Mayhew, M., Hartl, F. U., Dobson, C. M. & Radford, S. E. (1994). Conformation of GroELbound a-lactalbumin probed by mass spectrometry. Nature, 372, 646 ± 651. Roseman, A. M., Chen, S. X., White, H., Braig, K. & Saibil, H. R. (1996). The chaperonin ATPase cycle: mechanism of allosteric switching and movements of substrate-binding domains in GroEL. Cell, 87, 241 ± 251. Saibil, H. R., Zheng, D., Roseman, A. M., Hunter, A. S., Watson, G. M. F., Chen, S., Auf der Mauer, A., O'Hara, B. P., Wood, S. P., Mann, N. H., Barnett, L. K. & Ellis, R. J. (1993). ATP induces large quaternary rearrangements in a cage-like chaperonin structure. Curr. Biol. 3, 265± 273. Schmidt, M. & Buchner, J. (1992). Interaction of GroE with all-b-protein. J. Biol. Chem. 267, 16829± 16833. Schmidt, M., Buchner, J., Todd, M. J., Lorimer, G. H. & Viitanen, P. V. (1994a). On the role of groES in the chaperonin-assisted folding reaction. Three case studies. J. Biol. Chem. 269, 10304± 10311. Schmidt, M., BuÈcheler, U., Kaluza, B. & Buchner, J. (1994b). Correlation between the stability of the GroEL-protein ligand complex and the release mechanism. J. Biol. Chem. 269, 27964 ±27972. Shortle, D. & Meeker, A. K. (1989). Residual structure in large fragments of staphylococcal nuclease: effects

745 of amino acid substitutions. Biochemistry, 28, 936 ± 944. Sparrer, H., Lilie, H. & Buchner, J. (1996). Dynamics of the GroEL protein complex: Effects of nucleotides and folding mutants. J. Mol. Biol. 258, 74 ± 87. Sparrer, H., Rutkat, K. & Buchner, J. (1997). Catalysis of protein folding by symmetric chaperone complexes. Proc. Natl Acad. Sci. USA, 94, 1096± 1100. Staniforth, R. A., Cortes, A., Burston, S. G., Atkinson, T., Holbrook, J. J. & Clarke, A. R. (1994). The stability and hydrophobicity of cytosolic and mitochondrial malate dehydrogenases and their relation to chaperonin-assisted folding. FEBS Letters, 344, 129 ± 135. Todd, M. J., Viitanen, P. V. & Lorimer, G. H. (1994). Dynamics of the chaperonin ATPase cycle: implications for facilitated protein folding. Science, 265, 659 ± 666. Viitanen, P. V., Lubben, T. H., Reed, J., Goloubinoff, P., O'Keefe, D. P. & Lorimer, G. H. (1990). Chaperonin-facilitated refolding of ribulosebisphosphate carboxylase and ATP hydrolysis by chaperonin 60 (groEL) are K‡ dependent. Biochemistry, 29, 5665± 5671. Viitanen, P. V., Donaldson, G. K., Lorimer, G. H., Lubben, T. H. & Gatenby, A. A. (1991). Complex interactions between the chaperonin 60 molecular chaperone and dihydrofolate reductase. Biochemistry, 30, 9716± 9723. Walkenhorst, W. F., Green, S. M. & Roder, H. (1997). Kinetic evidence for folding and unfolding intermediates in stphylococcal nuclease. Biochemistry, 36, 5795± 5805. Walter, S., Lorimer, G. H. & Schmid, F. X. (1996). A thermodynamic coupling mechanism for GroELmediated unfolding. Proc. Natl Acad. Sci. USA, 93, 9425± 9430. Widmann, M. & Christen, P. (1995). Differential effects of molecular chaperones on refolding of homologous proteins. FEBS Letters, 377, 481± 484. Yifrach, O. & Horovitz, A. (1995). Nested cooperativity in the ATPase activity of the oligomeric chaperonin GroEL. Biochemistry, 34, 5303± 5308. Zahn, R., Spitzfaden, C., Ottiger, M., WuÈthrich, K. & PluÈckthun, A. (1994). Destabilization of the complete protein secondary structure on binding to the chaperone GroEL. Nature, 368, 261± 265. Zahn, R., Perrett, S. & Fersht, A. R. (1996a). Conformational states bound by the molecular chaperones GroEL and SecB: A hidden unfolding (annealing) activity. J. Mol. Biol. 261, 43± 61. Zahn, R., Perrett, S., Stenberg, G. & Fersht, A. R. (1996b). Catalysis of amide proton exchange by the molecular chaperones GroEL and SecB. Science, 271, 642± 645. Zheng, X., Rosenberg, L. E., Kalousek, F. & Fenton, W. A. (1993). GroEL, GroES, and ATP-dependent folding and spontaneous assembly of ornithine transcarbamylase. J. Biol. Chem. 268, 7489± 7493.

Edited by P. E. Wright (Received 22 July 1997; received in revised form 8 January 1998; accepted 12 January 1998)