Roles of the Two ATP Binding Sites of ClpB from Thermus thermophilus *

As a member of molecular chaperone Hsp100/Clp family, T ClpB from Thermus thermophilus has two nucleotide binding domains, NBD1 and NBD2, in a single polypeptide, each containing WalkerA and WalkerB con-sensus motifs. To probe their roles, mutations were introduced into the WalkerA or WalkerB motifs of each or both of the NBDs. The results are as follows. 1) For each of the NBDs, the ability of nucleotide binding is lost by mutations in the WalkerA motif but is retained by mutations in the WalkerB motif. 2) Each NBD has a casein-stimulatable small basic ATPase activity that is lost when the WalkerB motif is mutated. 3) T ClpB assembles into a uniform 580-kDa oligomer when ATP is present at 55 °C, and only the mutants in the WalkerA motif in NBD1 fail to assemble, indicating that ATP binding to NBD1 but not hydrolysis is necessary and sufficient for the assembly. 4) Chaperone function of T ClpB was lost when the WalkerA motif in each of the NBDs was mutated. Mutants in the WalkerB motifs of each NBD retained some chaperone activity. Hsp104, thermotolerance, cell heat-induced The functional chaperone cooperation between stress (2, 3)

As a member of molecular chaperone Hsp100/Clp family, TClpB from Thermus thermophilus has two nucleotide binding domains, NBD1 and NBD2, in a single polypeptide, each containing WalkerA and WalkerB consensus motifs. To probe their roles, mutations were introduced into the WalkerA or WalkerB motifs of each or both of the NBDs. The results are as follows. 1) For each of the NBDs, the ability of nucleotide binding is lost by mutations in the WalkerA motif but is retained by mutations in the WalkerB motif. 2) Each NBD has a caseinstimulatable small basic ATPase activity that is lost when the WalkerB motif is mutated. 3) TClpB assembles into a uniform 580-kDa oligomer when ATP is present at 55°C, and only the mutants in the WalkerA motif in NBD1 fail to assemble, indicating that ATP binding to NBD1 but not hydrolysis is necessary and sufficient for the assembly. 4) Chaperone function of TClpB was lost when the WalkerA motif in each of the NBDs was mutated. Mutants in the WalkerB motifs of each NBD retained some chaperone activity.
Hsp104, 1 originally identified as a factor for thermotolerance, is a molecular chaperone involved in the recovery of the cell from heat-induced damage (1). The functional chaperone cooperation between Hsp104 and Hsp70 enables yeast to recover from heat stress in vivo (2,3) and reactivates proteins that had been chemically denatured and allowed to aggregate in vitro (4). The bacterial homolog of Hsp104 is ClpB, a ubiquitous protein found in various strains of bacteria (5). Escherichia coli ClpB is a heat shock factor ( 32 )-dependent heat shock protein (6,7), and E. coli mutants lacking clpB exhibit growth defects at 44°C and undergo a more rapid death at 50°C than the wild type (7). Overproduction of other chaperones cannot compensate the deleterious effect of the clpB null mutants at 50°C (8). We have demonstrated that heat-inactivated proteins were rescued by the cooperation between DnaK (Hsp70)-DnaJ-GrpE and ClpB from a thermophilic eubacteria, Thermus thermophilus HB8 (9,10). Such cooperation was also reported for E. coli chaperones and yeast mitochondrial ones (11)(12)(13)(14).
Hsp104/ClpB is a member of the Hsp100/Clp family that includes proteins of varied functions (15). As known so far, members of this family can self-assemble into ring-shaped hexamers or heptamers in the presence of ATP and possess ATPase activity that is stimulated by casein (16 -22). Recently, it has been recognized that ATP-dependent remodeling and unfolding of proteins underlie the functions of the members of this family. ClpA, a major member of this family, is involved in ATP-dependent protein degradation through association with the unrelated protease ClpP. Without ClpP, ClpA can disassemble inactive dimers of the plasmid P1 initiator protein, RepA, into active monomers (23) and unfold SsrA (a signal peptide that is responsible for recruiting truncated proteins to ClpAP for degradation)-tagged protein (24). ATP-dependent proteolysis by the ClpAP complex is based on these activities of ClpA. ClpX and ClpY (HslU) are also known to be regulatory subunits of the ATP-dependent protease and are complexed with the unrelated proteases ClpP and ClpQ (HslV), respectively. Similar to ClpA, ClpX shows protein remodeling and unfolding activity (25,26). The Hsp100/Clp family is divided into two major classes. Members of the first class (Class1) have two nucleotide binding domains, NBD1 (N-terminal side) and NBD2 (C-terminal side) in a single polypeptide. Each of the two NBDs contains consensus sequence motifs, WalkerA (GXXGXGKT, where X is the variable) and WalkerB (hhhhDE, where h is the hydrophobic residue) (27). Although each of these two domains is well conserved throughout all members of the Class1 Hsp100/Clp family, there is only a limited sequence similarity between NBD1 and NBD2. Members of another class (Class2) have only one NBD homologous to NBD2 of the Class1 Clps. ClpB, Hsp104, and ClpA belong to the Class1, and ClpX and ClpY belong to Class2 (15).
The role of the NBDs in the Hsp100/Clp family has been studied using site-directed mutagenesis in vivo and in vitro (3, 16, 22, 28 -35). As expected, the replacement of the Lys residue in the NBD of ClpY by Thr resulted in a loss of adenine nucleotide binding, hexamer formation, and ATPase function (31). The distinct roles of the two NBDs have been probed for several Class1 members. For yeast Hsp104, mutations in the NBD2 affected oligomer assembly more severely than the same mutations in NBD1 (16,32). NBD2 mutants of Hsp104 had some ATPase activity, but NBD1 mutants did not (32,35). On the contrary, two independent studies on E. coli ClpA (29,30) showed that mutation of the Lys residue in the WalkerA motif of NBD1 resulted in a loss of ability to form hexamers, whereas the equivalent mutant of NBD2 was defective in ATP hydrolysis and was unable to activate protease activity of ClpAP complex. For E. coli ClpB, mutants of the WalkerA motif in either NBD1 or NBD2 possess significant ATPase activity, but only the former lost the ability of ATP-dependent assembly of a uniform 580-kDa oligomer in the presence of salt (22). Recently, similar experiments were carried out for ClpB from T. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Supported by a research fellowship from the Japan Society for the Promotion of Science for Young Scientists.
§ To whom correspondence should be addressed. Fax: 81-45-924-5277; E-mail: myoshida@res.titech.ac.jp. 1 The abbreviations used are: Hsp, heat shock protein; NBD, nucleotide binding domain; G6PDH, glucose-6-phosphate dehydrogenase; TDnaKJ complex, the trigonal complex of DnaK 3   (TClpB) by Schlee et al. (34), but the results differed from those of E. coli ClpB. They observed that wildtype TClpB did not assemble into a uniform oligomer with a distinct molecular mass in the presence of ATP and salt, but that rather it existed as a dynamic mixture of various oligomeric states with dimers and tetramers being the dominant species. Mutants of TClpB in WalkerA motif of either NBD1 or NBD2 were shown to have only little effect on these oligomeric states. All of their measurements were carried out at 25°C.
Here, using TClpB, we report that at 55°C, a physiological temperature for T. thermophilus, TClpB exists as a stable 580-kDa oligomer in the presence of ATP and salt, and that the mutation of the WalkerA motif in NBD1 but not NBD2 prevents the formation the 580-kDa oligomer. Thus, the previous discrepancy between E. coli ClpB and TClpB on the role of WalkerA motif in NBD1 was resolved. In addition, we clarified that WalkerB motif of each NBD is essential for ATP hydrolysis activity of its own NBD.

EXPERIMENTAL PROCEDURES
Proteins-The recombinant plasmid pMCB1 (9) containing a gene for TClpB was used for mutagenesis. Site-directed mutagenesis was performed using polymerase chain reaction with the QuikChange TM sitedirected mutagenesis kit from Stratagene. The mutations were confirmed by DNA sequence analysis. The proteins were expressed in E. coli (BL21 (DE3)) and purified as described previously (9). TDnaKJ complex and TGrpE were also expressed in E. coli (BL21 (DE3)) carrying pMKJ8 (TDnaKJ complex) and pMGE3 (TGrpE), respectively, and purified as described previously (36,37). Throughout this paper, concentrations of substrate proteins are expressed as monomers and those of T. thermophilus chaperones are noted as the trigonal complex for TDnaKJ complex (38), dimer for TGrpE. TClpB and its mutants are expressed as a monomer or hexamer as indicated.
Fluorescence Measurements-The fluorescent nucleotide analogous 2Ј(3Ј)-O-NЈ-methylaniloyl-aminoadenosine-5Ј-diphosphate (Mant-ADP) was prepared as described previously (39). Fluorescence measurements were performed with a Hitachi F4500 fluorometer. The excitation wavelength was 360 nm, and emission spectra from 400 -500 nm were collected. Purified wild-type or mutant TClpB (1 M as hexamer) in 50 mM MOPS-NaOH, pH 7.5, 150 mM KCl, 5 mM MgCl 2 (1.2 ml) were preincubated for more than 10 min in a sealed stirring cuvette at 55°C. Mant-ADP was added, and the fluorescence spectra were measured when equilibration was reached (1 min). As a control, the fluorescence spectra of the buffer containing the same components, other than TClpB, with appropriate concentrations of Mant-ADP were measured, and the differences of the emission intensities at 440 nm were plotted against concentrations of Mant-ADP. Apparent dissociation constants of Mant-ADP for TClpB (monomer) were calculated by fitting the titration data to the model of simple 1:1 binding taking into consideration the decrease of concentrations of free Mant-ADP as shown in Equation 1 where ⌬Fl is the change of the fluorescence intensity at 440 nm, C is the constant, E is the TClpB monomer, F is the Mant-ADP, and K d is the apparent dissociation constant of Mant-ADP for TClpB.
The displacement of Mant-ADP by ADP or ATP was detected by monitoring the decrease in fluorescence at 440 nm by adding Mg-ADP or Mg-ATP. Dissociation constants of ADP and ATP for TClpB (monomer) were calculated by fitting the titration data to the model of competitive binding as shown in Equation 2 Fl where Fl is the fluorescence intensity at 440 nm, C 1 and C 2 are the constants, E is the TClpB monomer, F is the Mant-ADP, K d1 is the apparent dissociation constant of Mant-ADP for TClpB, S is the ADP or the ATP, and K d2 is the apparent dissociation constant of ADP or ATP for TClpB. Data were analyzed with KaleidaGraph 3.0 (Synergy Software).
Gel-filtration Analysis-Purified wild-type or mutant TClpB (1 mg/ ml) in 50 mM MOPS-NaOH, pH 7.5, 5 mM MgCl 2 , 1 mM dithiothreitol, the indicated concentrations of KCl, and the indicated nucleotide was preincubated for more than 1 min at the loading temperature. The mixture was then centrifuged for 1 min (no obvious precipitation). An aliquot (100 l) of the solution was loaded on a HPLC gel-filtration column TSK G-3000SW XL (Tosoh). The column was preequilibrated, and the protein was eluted at a flow rate 0.5 ml/min with the same buffer as described above but without dithiothreitol at the indicated temperature. The elution of proteins was monitored by absorbance at 290 nm. Molecular size standards used were thyroglobulin (669 kDa), ferritin (440 kDa), TDnaKJ complex from T. thermophilus (330 kDa), catalase (232 kDa), G6PDH from Bacillus stearothermophilus (212 kDa), and aldorase (158 kDa). Elution profiles of the heat stable standard proteins (TDnaKJ complex, G6PDH) at 55°C were the same as those at 20°C, ensuring that the column properties were not changed at high temperature.
Protein Cross-linking-Purified TClpB (10 g/ml) in 50 mM MOPS-NaOH, pH 7.5, 150 mM KCl, 5 mM MgCl 2 , and 0.5 mM dithiothreitol was incubated with glutaraldehyde (final concentration 0.01%) in the absence or the presence of the indicated nucleotide at the indicated temperature for 30 min. The reaction was terminated by adding 150 l of 1 M glycine, pH 6.25, to 300 l of the reaction mixture, and proteins were precipitated by trichloroacetic acid (final concentration 10%) on ice for 10 min. The precipitates were resuspended in the sample buffer for electrophoresis and adjusted to neutral pH by 1 M Tris-HCl, pH 7.5. Proteins were analyzed with 3.5% polyacrylamide gel electrophoresis in the presence of sodium dodecylsulfate and were visualized with silver staining.
ATPase Assay-ATPase activities were assayed in a reaction mixture (200 l) in 50 mM MOPS-NaOH, pH 7.5, 150 mM KCl, 10 mM MgCl 2 , indicated concentrations of ATP, and 20 g of casein when indicated. The reaction was initiated by the addition of ATP. After incubation of the reaction mixtures at 55°C for an appropriate period, an aliquot was transferred into 2% perchloric acid, and the released P i was measured by malachite green method (40,41).

WalkerA Motif Mutations of Each NBD Abolished Nucleotide
Binding in the NBD-We introduced three types of mutation into each of NBDs (Fig. 1). One type was mutations in the WalkerA motifs, a replacement of Lys-Thr by Ala-Ala (KT/AA). The other two were mutations in WalkerB motifs, a replacement of Asp by Asn (D/N) and a replacement of Glu by Gln (E/Q). In addition to these mutations in a single NBD, we made five types of "double" mutants of TClpB, combining the mutations in NBD1 and in NBD2 as mentioned above.
First, we investigated the nucleotide binding ability of the wild-type and mutant TClpB by using the fluorescence nucleotide analog Mant-ADP at 55°C. When Mant-ADP was mixed with the wild-type TClpB, fluorescence increased severalfold, and the emission peak was blue-shifted. The extent of the fluorescence increases at 440 nm induced by the wild-type, and some of the mutant TClpBs were plotted against the concentrations of Mant-ADP (Fig. 2, A and B), and the apparent K d values were calculated by fitting the data to a simple 1:1 binding scheme (Table I) Fig. 2A, inset). A single mutant carrying mutations in WalkerA-NBD1 (1KT/AA) showed almost the same fluorescence increase as the wild type, and the K d value was slightly smaller than that of the wild type. A single mutant carrying mutations in WalkerA-NBD2 (2KT/AA) showed only small fluorescence increase. However, the titration curve obeyed a saturation curve, and the K d value (14 M) was calculated from these data ( Fig. 2A, inset). We also measured the decreases of Mant-ADP fluorescence by adding Mg-ADP or Mg-ATP in the presence of wild-type or mutant TClpB (Fig. 3). The apparent K d values of ADP and ATP were calculated by fitting the data to the model of the competitive binding (Table I). The competitive titration of Mg-ADP against Mant-ADP suggested that the affinities for ADP of the 1KT/AA and 2KT/AA mutations were almost the same each other (Table  I). Thus, NBD2 has a stronger affinity for Mant-ADP than NBD1 but has almost the same affinity for ADP as NBD1. These results are consistent with those obtained from measurements at 25°C reported in a previous paper on TClpB (34), although the affinities were slightly lower. In the case of ATP, as TClpB has intrinsic ATPase activity, the apparent K d values calculated from displacement titration might be influenced by the binding of ADP generated by ATP hydrolysis. As the values of apparent K d of ATP for all mutants and wild-type TClpB were higher than the K d values of ADP for each TClpB, the values represent the lower limit. Among them, 1KT/AA, 2KT/AA mutants, and combined mutants have very low or no ATPase activities (see below), thus the influences of generated ADP on the K d value of ATP for these mutants would be relatively low. The apparent K d values of ATP for NBD2 (1KT/ AA) and NBD1 (2KT/AA) were essentially the same. The effect of mutations of WalkerB motifs of both NBDs was also examined. Mutations in WalkerB-NBD1 (1D/N and 1E/Q) did not affect the ADP and ATP binding properties of TClpB. Double mutants of WalkerB-NBD1 and WalkerA-NBD2 (1D/N-2KT/AA and 1E/Q-2KT/AA) did not change significantly the binding affinities of a single WalkerA-NBD2 mutant (2KT/AA), although the K d of ATP for 1D/N-2KT/AA was slightly higher. The mutations of WalkerB-NBD2 (2D/N and 2E/Q), either alone or in combination with WalkerA-NBD1 mutation, resulted in a somewhat stronger binding affinity for ADP. Thus, the mutations of WalkerB of both NBDs did not impair or even strengthen the binding of ADP and ATP.
WalkerB Mutations Inactivated the ATPase in the Mutated NBD-We measured the ATPase activities of mutants of TClpB with or without 0.1 mg/ml casein at 55°C at 5 mM ATP (Fig. 4, A and B). ATPase activities of all the single mutants with the exception of 1E/Q were lower than that of the wild  Two ATP Binding Sites of ClpB type but were stimulated by casein, although the degrees of stimulation were different from each other. Unlike other mutations, 1E/Q mutation had an enormously increased ATPase activity, 10-fold that of the wild type, and this activity was not stimulated but rather was slightly inhibited by casein. The real reason of this high casein-insensitive ATPase of the 1E/Q mutant is not known, but one speculative possibility is that residue Glu-271 in the wild-type TClpB may be committed to the suppression of ATP hydrolysis, which can be released by unfolded protein. Both WalkerA-NBD1 (1KT/AA) and WalkerA-NBD2 (2KT/AA) mutants retained some small ATPase activities, which were stimulated when casein was present. Because the 1KT/AA mutant is unable to bind nucleotide at NBD1, the residual ATPase activity of this mutant is solely attributed to the ATPase activity of intact NBD2. Similarly, ATPase activity of 2KT/AA is solely attributed to intact NBD1. The sum of the ATPase activities of the two mutants is far smaller than the activity of the wild type. Thus, each of NBD1 and NBD2 has a basic ATPase activity represented by 2KT/AA and 1KT/AA, respectively, and cooperative stimulation is assumed between two NBDs in the wild-type TClpB (34). To understand the roles of the WalkerB motifs of each of NBDs in the basic ATPase activity, we measured ATPase activities of the double mutants, WalkerA-NBD1/WalkerB-NBD2 (1KT/AA-2D/N and 1KT/AA-2E/Q) and WalkerA-NBD2/WalkerB-NBD1 (2KT/AA-1D/N and 2KT/AA-1E/Q) (Fig. 4B). No significant ATPase activity was detected for any of the mutants with or without casein. These results indicate that the two acidic residues, Asp-Glu in WalkerB motifs, play a critical role in the ATP hydrolysis reaction occurring in each of NBDs.
Temperature-and ATP-dependent Assembly of TClpB into a 580-kDa Oligomer-It has been reported that the yeast Hsp104 and E. coli ClpB oligomerize to a uniform-defined oligomer in an ATP-dependent manner (16,19,22). This defined oligomer of ClpB has been proposed to be a heptamer rather than a hexamer (22), but we refer to them as a 580-kDa oligomer in this study. At first, oligomer formation of wild-type TClpB was assessed by gel-filtration HPLC (Fig. 5A, upper trace). At 20°C in the absence of nucleotide, the TClpB peak appeared with an apparent size of 400 -450 kDa (4 -5-mer of a 96-kDa monomer). When 2 mM ADP was present in the elution buffer for chromatography, the retention time of TClpB was significantly delayed, and the estimated molecular size of the dominant species was ϳ250 kDa (2-3-mer). In the presence of 2 mM ATP, TClpB eluted at the position of ϳ5-mer. However, under any conditions tested at 20°C, TClpB was eluted as a broad band with a long tail. This finding was not because of interaction with the column matrix as the analysis using another type of gel (Superose 6) gave very similar elution profiles as those described above (data not shown). These results agree with those reported by Schlee et al. (34), who measured the oligomeric state of TClpB at 25°C. Gel filtration chromatography at 55°C at which temperature TClpB is active as a chaperone (9, 10) produced different elution profiles (Fig. 5A, middle trace). With 2 mM ATP present in the elution buffer, TClpB was eluted as a nearly symmetric peak at 12.5 min, which corresponded to an apparent molecular size of 580 kDa (6-mer). Without nucleotide, TClpB eluted as a long-tailed low peak at the position of ϳ580 kDa, suggesting the decay of the oligomers during elution. With 2 mM ADP, the peak of TClpB appeared at 13.0 min corresponding to 500 kDa (ϳ5-mer). In the presence of ADP, E. coli ClpB tends to exist as dimer (19), and Hsp104 exists as hexamer (16). The present case is different from both cases. The buffers used in the above experiments contained 150 mM KCl. It was reported that the oligomer of TClpB is destabilized when KCl concentration is increased (34). In fact, at a higher KCl concentration (300 mM), the elution of TClpB without nucleotide was significantly delayed (Fig. 5A, lower trace). However, the effect of elevated KCl concentration was minimal on the elution of TClpB in the presence of ATP. Thus, the oligomeric states of TClpB vary depending on the temperature, salt concentration, and the nucleotide present, and it exists as a stable oligomer with M r ϭ 580 kDa at only 55°C and in the presence of ATP, the same conditions where TClpB is active as a chaperone. This finding was further supported by the crosslinking experiments. TClpB was covalently cross-linked by incubation with glutaraldehyde at 55°C. In the absence of nucleotide, the product was comprised of molecular species of various sizes as shown by a ladder pattern on SDS gel electrophoresis (Fig. 5B, lane 2). In the presence of 3 mM ATP, the cross-linked product was electrophoresed as a single band corresponding to the sixth band of the ladder pattern of lane 2 (Fig. 5B, lane 3). Although this observation suggests that TClpB exist as hexamer in the presence of ATP, the heptameric ring structure of E. coli ClpB was observed with electron microscopy previously (22), and the number of monomers in the ring has been remained unsettled. With other conditions such as 3 mM ADP at 55°C or 3 mM ATP at 20°C, the cross-linked products showed a ladder pattern on SDS gel electrophoresis similar to that of lane 2 (data not shown).
ATP Binding to NBD1 Is Crucial for Assembly of TClpB into the 580-kDa Oligomer-The oligomer states of TClpB mutants were analyzed by gel-filtration chromatography at 55°C in the presence of 150 mM KCl (Fig. 6). Without nucleotide, TClpB mutants were eluted as a broad peak with a long tail (Fig. 6, left panel) as seen for the wild type. When the elution buffer contained 2 mM ADP, the elution of the WalkerA-NBD1 mutant was much delayed, whereas all other single mutants eluted at 13.0 min (ϳ5-mer) as did the wild type (Fig. 6, center panel). When the elution buffer contained 2 mM ATP, the WalkerA-NBD1 mutant eluted at 13.0 min with a considerable tailing (Fig. 6, right panel). When the KCl concentration was elevated to 300 mM in the presence of 2 mM ATP, elution was further delayed. On the other hand, elution patterns of other mutants in the presence of 2 mM ATP, either at 150 or 300 mM KCl, were very similar to those of the wild type, a nearly symmetric peak at 12.5 min, the position of the 580 kDa oligomer (Fig. 6, right  panel). Because the mutants of WalkerB-NBD1 cannot catalyze ATP hydrolysis at NBD1 as described earlier but still can form a stable 580-kDa oligomer, ATP hydrolysis at NBD1 is not essential for the assembly to form a 580-kDa oligomer. The results of mutations of NBD2 indicate that ATP binding and hydrolysis at NBD2 are not prerequisite for the assembly. To summarize, the ATP binding to NBD1 but not hydrolysis, is necessary and sufficient for the assembly of a stable 580-kDa oligomer of TClpB.
Chaperone Activities of the Mutant TClpB-We previously demonstrated chaperone activity of TClpB in the reactivation of the heat-inactivated proteins. Cooperation with the TDnaKJ-GrpE set was essential for this activity (9, 10). Using ␣-glucosidase and G6PDH as substrate proteins, chaperone activities of TClpB mutants were measured. Substrate proteins were inactivated at high temperatures (73°C for ␣-glucosidase and 72°C for G6PDH) in the presence of TDnaKJ complex ϩTGrpE ϩ 3 mM ATP, and then the temperature was shifted down to 55°C with concomitant addition of TClpB. When the The temperature was shifted down to 55°C, and TClpB or its mutants (0.05 M hexamer) were immediately added to the solutions. After a 120-min incubation at 55°C, recovered enzyme activities were measured. The recoveries gained after a 120-min incubation at 55°C are shown as percent of the activities before heat inactivation. The column marked None shows the recovered activities when any TClpB was not added. Experimental details are described under "Experimental Procedures." wild-type TClpB was added, ϳ70% activities were recovered after a 2-h incubation for both ␣-glucosidase and G6PDH. When the WalkerB mutants of either of the NBDs were added instead of the wild type, the substrate proteins were reactivated, although the yield was somewhat diminished (Fig. 7, A  and B). Conversely, when the WalkerA mutants of either of the NBDs were added, no reactivation was observed. We examined several other experimental conditions such as the addition of TClpB prior to incubation at 72°C, a 5-fold excess of TClpB, or 300 mM KCl, but the results were essentially similar, although the reactivation yield at 300 mM KCl was slightly lower (data not shown). These results indicate that ATP binding to both NBDs is required for chaperone activity. ATP hydrolysis at only one of the NBDs can support some small chaperone activity, but efficient activity requires communication between two intact NBDs. DISCUSSION The functional role of each of the two NBDs of ClpB/Hsp104 family has been contradictory in the previous papers. WalkerA motif in NBD2 (yeast Hsp104) or the WalkerA motif in NBD1 (E. coli ClpB) or none of them (TClpB) has been implicated in playing an essential role in oligomerization (16,22,34). A part of this discrepancy is now resolved by this work, as TClpB behaves similarly to E. coli ClpB when experiments are carried out at 55°C rather than at 25°C (34) where the oligomers are unstable. Thus, ATP binding to NBD1 is necessary for ClpB to form a stable oligomer. In addition, the result of WalkerB mutants in NBD1 of TClpB provides good evidence that ATP binding alone but not ATP hydrolysis at NBD1 enables TClpB to form stable 580-kDa oligomers. This report has also clarified the role of the WalkerB motifs of TClpB for the first time in the Hsp100/Clp family. The WalkerB motifs of this family have always hhhhDE sequence (where h is the hydrophobic residue). Substitution of the carboxylate group of Asp or Glu with carboxyamide group (Asn or Gln) abolishes the basic ATPase activity of each of the NBDs. Therefore, both Asp and Glu in the WalkerB motif are necessary for ATP hydrolysis. Considering the crystal structure of HslU, a member of Hsp100/Clp family (21), one can assume that Asp is ligated to Mg 2ϩ , and Glu is acting as a water-attacking base. Some different effects of mutations between D/N and E/Q substitutions of the WalkerB motif of the same NBD, such as ATPase activities of 1D/N and 1E/Q mutants, may be caused by such a difference.