FEBS Letters 584 (2010) 929–934

journal homepage: www.FEBSLetters .org

Nucleotide utilization requirements that render ClpB active as a chaperone

Urko del Castillo, José Ángel Fernández-Higuero, Sergio Pérez-Acebrón 1, Fernando Moro, Arturo Muga *

Unidad de Biofísica (CSIC-UPV/EHU), and Departamento de Bioquímica y BiologíaMolecular (UPV/EHU), Facultad de Ciencia y Tecnología, Universidad del País Vasco,P.O. Box 644, Bilbao, Spain

a r t i c l e i n f o

Article history:Received 6 December 2009Revised 28 December 2009Accepted 12 January 2010Available online 19 January 2010

Edited by Felix Wieland

Keywords:ClpBChaperonesProtein oligomerAggregate reactivation

0014-5793/$36.00 � 2010 Federation of European Biodoi:10.1016/j.febslet.2010.01.029

Abbreviations: T1, ClpB(E279A); T2, ClpB(E678A);ClpB(K611A); T1N2, ClpB(E279A/K611A); N1T2, ClpB(K21

Thermus thermophilus* Corresponding author. Fax: +34 944648500.

E-mail address: arturo.muga@ehu.es (A. Muga).1 Present address: German Cancer Research Center (

ology), Im Neuenheimer Feld 581, Heidelberg 69120, G

a b s t r a c t

ClpB is a member of the AAA+ superfamily that forms a ring-shaped homohexamer. Each protomercontains two nucleotide binding domains arranged in two rings that hydrolyze ATP. We extend hereprevious studies on ClpB nucleotide utilization requirements by using an experimental approachthat maximizes random incorporation of different subunits into the protein hexamer. Incorporationof one subunit unable to bind or hydrolyze ATP knocks down the chaperone activity, while the wthexamer can accommodate two mutant subunits that hydrolyze ATP in only one protein ring. Foursubunits seem to build the functional cooperative unit, provided that one of the protein rings con-tains active nucleotide binding sites.� 2010 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

Bacterial ClpB is an AAA+ molecular machine that uses the en-ergy of ATP binding and hydrolysis to extract unfolded proteinmonomers from aggregates, and to remodel and translocate themthrough its axial channel for productive reactivation [1–3]. Disag-gregation and reactivation of protein aggregates in the cell requiresthe combined action of ClpB and the DnaK system, that alone areunable to reactivate large insoluble aggregates [4–7]. As a memberof class I of the Hsp100 chaperone family, ClpB contains two nucle-otide binding sites (NBD) per protomer [8]. Therefore, in the func-tional hexameric assembly there are 12 nucleotide binding sitesarranged in two rings, each containing one of the two nucleotidebinding sites of the protomers that most likely are allostericallycommunicated to thread the substrate protein through the centralpore during aggregate reactivation. Ring-1 is located close to the N-terminus of the protein and ring-2 comprises the C-terminal ATPbinding sites. So far, there is neither allosteric model to explainthe functionality of this protein nor a clear view of the intra- andinter-subunit communications. Besides the two nucleotide bindingdomains (NBDs), the N-terminal domain is assumed to play a role

chemical Societies. Published by E

T1T2, ClpB(E279A/E678A); N2,2A/E678A); TClpB, ClpB from

Division of Molecular Embry-ermany.

in substrate recognition, but appears to be dispensable for disag-gregating small aggregates in vitro [9,10] and the C-terminal do-main is involved in subunit oligomerization [11]. The M domain,which is specific for ClpB and its homolog Hsp104 and strictly re-quired for protein disaggregation [12–15], is inserted into NBD1and forms a large coiled-coil structure built by four helices.

An interesting property of this protein is that subunits can eas-ily exchange between hexamers, indicating that the oligomericstructure is highly dynamic [16,17]. A recent study with TClpBhas used this property to show that insertion of an inactive subunitwithin the hexamer impairs its chaperone activity [16]. Using thesame mixing approach, it has been recently shown that ClpB canmodify the mechanism of ATP utilization depending on the sub-strate and the presence of the DnaK system [18]. While in the ab-sence of DnaK the protein can remodel the substrate withapproximately three active protomers per hexamer, in the pres-ence of DnaK the characteristics of the active hexamer have notbeen defined so accurately. We have used an alternative experi-mental approach to generate hybrid hexamers containing wt ClpBand two different types of mutant subunits, which are defective inATP binding or hydrolysis either in one or both nucleotide-bindingrings of the hexamer. This approach involves denaturation of theprotein variants whose subunits will form the hybrid hexamers,and refolding of the heterooligomer after removing the denaturantby dialysis. Our data show that incorporation of one single rigidsubunit unable to bind or hydrolyze ATP abolishes the chaperoneactivity of the hexamer. However, the hexamer can incorporatetwo mutant subunits able to hydrolyze ATP at either protein ring,

lsevier B.V. All rights reserved.

930 U. del Castillo et al. / FEBS Letters 584 (2010) 929–934

suggesting that the active, conformationally alternating ring allowsthe rearrangement of the four wt subunits, so that their nucleotidebinding sites can communicate to productively thread and processthe substrate.

2. Materials and methods

2.1. Protein expression and purification

Wt ClpB and ClpB mutants were expressed in BB4561 Esche-richia coli strain and purified as previously described [19]. ClpB sin-gle point mutants were cloned using Stratagene Quickchange II kitand mutations were confirmed by DNA sequencing. DnaK was ex-pressed, purified, and extensively dialyzed to obtain nucleotide-free samples [20]. DnaJ, GrpE were obtained as previously reported[21,22]. Protein concentration was determined by the colorimetricBradford assay (Bio-Rad), except for DnaK, that was determinedspectrophotometrically using e280 = 15.8 � 103 M�1 cm�1 [23]. Ma-late dehydrogenase (MDH) was purchased from Sigma and lactatedehydrogenase (LDH) and pyruvate kinase (PK) from Roche.

2.2. Hybrid ClpB hexamers

Hybrid chaperones were prepared using the method describedfor GroEL [26,27], slightly modified. Briefly, pure homooligomersor mixtures of ClpB variants (final protein concentration 10 mg/ml) at different subunit molar ratios were denatured in buffer50 mM Tris–HCl, 5 mM b-mercaptoethanol, 8 M urea, pH 7.5(25 �C, 2 h). Unfolded proteins or protein mixtures were renaturedby diluting them 15 times into buffer 50 mM Tris–HCl,20 mM MgCl2, 5 mM ATP and 5 mM b-mercaptoethanol, pH 7.5.To remove residual urea, they were dialyzed overnight againstthe same buffer at 4 �C. Finally, the samples were centrifuged dur-ing 10 min at 13 000 rpm to eliminate any insoluble material, andthe renatured hexamers were concentrated with centricon YM-50filters (Millipore).

2.3. Size-exclusion chromatography (SEC)

Wild type ClpB or ClpB mutants (50 lM monomer) were loadedon a Superose 6 PC 3.2/30 (GE Healthcare) equilibrated in 50 mMTris–HCl, 50 or 150 mM KCl, 2 mM ATP, pH 7.5. Experiments wereperformed at 25 �C.

2.4. ATPase activity

Steady state ATPase activity measurements were carried out in50 mM Tris–HCl (pH 7.5) buffer containing 5 mM MgCl2 and thedesired KCl concentration (50 or 150 mM) at 25 �C, as described[20]. Protein and ATP concentration were 2 lM (monomer) and1 mM, respectively, in the presence of an ATP-regenerating system[24]. Substrate-induced stimulation of the ATPase activity wasdetermined in the presence of 10 lM a-caseine (Sigma). Reactionswere followed measuring the absorbance decay at 340 nm in aCary spectrophotometer (Varian).

Fig. 1. Biochemical properties of renatured homo-oligomers of wt ClpB and itsmutants. (A) ATPase activity in the absence (black) or in the presence (grey) of a-casein. (B) Reactivation of aggregated MDH by native and renatured wt ClpB and T1.(C) Protection against aggregation of thermally denatured MDH by native andrefolded T1T2. The absorbance at 550 nm of a MDH (6 lM) sample was followed inthe presence of 6 lM (monomer) T1T2. R refers to the protein variant that has beenunfolded in 8 M urea and renatured after removing the denaturant by dialysis. In(A) and (B) data are means ± S.E.M. (n = 3), and in (C), representative experiments oftwo replicates are shown.

2.5. Reactivation of MDH aggregates

MDH (2 lM monomer) was denatured and aggregated by incu-bating the protein for 30 min at 47 �C in 50 mM Tris–HCl, 150 or50 mM KCl, 20 mM MgCl2, 2 mM DTT, pH 7.5. The sample was di-luted to 1 lM in the same buffer containing 1.5 lM wt ClpB eitheralone or in hybrid hexamers containing the indicated molar ratio ofmutant subunits, 1 lM DnaK, 0.3 lM DnaJ, and 0.2 lM GrpE, andreactivation was started by adding 2 mM ATP to the sample includ-

ing an ATP regenerating system (6 mM phosphoenolpyruvate and20 ng/ml pyruvate kinase) at 30 �C. MDH activity was recordedafter 3 h reactivation as previously described [25]. The concentra-tion of wt ClpB in the hybrid hexamers was kept constant(1.5 lM monomer), since below 0.7 lM the chaperone activity issensitive to protein dilution. Therefore, changes in chaperoneactivity are due to the formation of hybrid hexamers.

2.6. MDH aggregation

The ability of T1T2 or hybrid hexamers, containing differentmolar ratios of wt ClpB and T1T2 subunits, to avoid aggregationof thermally denatured MDH was followed by measuring the tur-bidity at 550 nm of MDH samples (6 lM) at 47 �C. Experimentswere performed in buffer 50 mM Tris–HCl, 150 mM KCl, 20 mMMgCl2, 2 mM DTT, 10 mM ATP, pH 7.5. T1T2 concentration wasmaintained constant in the hybrid hexamers (6 lM), since thisconcentration of pure mutant was found to suppress MDH aggre-gation under the same experimental conditions. To estimate thenumber of T1T2 subunits involved in MDH binding, similar exper-

U. del Castillo et al. / FEBS Letters 584 (2010) 929–934 931

iments were carried out at increasing ATP concentrations. Datawere fitted to the Hill equation y ¼ ymax

xn

knþxn, where y is the per-centage of protection against MDH aggregation, x is the ATP con-centration, n is the Hill coefficient and K the dissociationconstant. Kinetic measurements were performed in a Cary spectro-photometer (Varian).

3. Results and discussion

We have used a new experimental method to generate ClpBhexamers containing different proportions of randomly distributedwt and different mutant subunits with substitutions in one or bothnucleotide binding sites. Mutation of the highly conserved lysineresidues at the Walker A motif (K212 and K611 in NBD1 andNBD2, respectively) by alanine impairs nucleotide binding at themutated NBD (N – null-substitution). Replacement of the con-served glutamic acid residues of the Walker B motif (E274 andE678 at the NBD1 and NBD2, respectively) by alanine results inNBDs that bind but not hydrolyze nucleotide (T – trap-substitu-tion) [12]. Hybrid oligomers were obtained by denaturating theproteins in 8 M urea, followed by removal of the chaotropic agentby dialysis and refolding of the protein particle (Fig. S1). The alter-native method that has been used to generate heterohexamers ofClpB is mixing homohexamers of different ClpB variants that as aconsequence of their dynamic nature exchange subunits in solu-tion [16,17]. However, the oligomeric intermediates involved insubunit shuffling are as yet unknown, and therefore the unfold-

Fig. 2. Biochemical properties of hybrid ClpB hexamers containing wt and mutant subuniT1N2 (B) and N1T2 (C) protomers in the wt ClpB hexamer on MDH reactivation. Experdescribed in [16], that assumes that a hybrid hexamer displays chaperone activity when imore subunits (red), four or more subunits (green), five or more subunits (blue), and sconditions. ATPase activity of the same heterohexamers formed by different subunit mATPase activity values (circles) are compared to those estimated (triangles) considering tThe ATPase activity was assayed in the absence (filled symbols) and presence (empty s(n = 3).

ing/refolding strategy employed herein maximizes random incor-poration of different subunits into the hybrid hexamer, andtherefore allows a reliable modeling of the experimental data.The heterohexamers generated by this method will, in addition,be at equilibrium, and protomer exchange would not vary theirsubunit composition provided that different subunits have similartendency to associate/dissociate (see below).

Modeling the distribution of different ClpB subunits within theprotein hexamer has been made using the binomial distributionfunction previously described for TClpB, that calculates the proba-bility of a protein specie with a defined proportion of mutant andwt subunits to occur at a given experimental molar ratio, but doesnot consider the possible distinct configurations that a specific spe-cie can adopt, i.e., the relative position of different subunits withinthe hexamer (Fig. S1) [16]. We also compare data presented hereinwith recently published results obtained with mixing experiments,in which different amounts of homohexamers of wt ClpB and someof the protein mutants used in this study are mixed before theexperiment [16,18].

To test if the refolding method modifies the biochemical prop-erties of the proteins, the ATPase and chaperone activities of re-folded homohexamers were compared with those of the nativeproteins. Data shown in Fig. 1 demonstrate that the ATPase activityin the absence and presence of substrate is not significantly mod-ified by the denaturation/refolding process (Fig. 1A). The ATPaseactivity of mutant T2, both in the absence and presence of sub-strate, is around four times higher than that of wt ClpB in spite

ts defective in ATP binding or hydrolysis. Effect of increasing the amount of T1T2 (A),imental values (filled circles) are compared with those estimated with the model

t contains one or more subunits (black), two or more subunits (dark yellow), three orix (pink) wt subunits. Wt ClpB reactivates 64 ± 4% MDH under these experimentalolar ratios of wt ClpB and T1T2 (D), T1N2 (E) and N1T2 (F) subunits. Experimentalhe activity of the corresponding homooligomers and the experimental molar ratios.ymbols) of a-casein. Inset, ATPase activity per wt subunit. Data are means ± S.E.M.

Fig. 3. Protection against aggregation of thermally denatured MDH by wtClpB:T1T2 hexamers. (A) Comparison of the percentage of protection for wtClpB:T1T2 hexamers containing different subunit molar ratios (filled circles) withthe theoretical values estimated [16] assuming that a hybrid hexamer can stablybind the substrate when incorporates one or more subunits (black), two or moresubunits (dark yellow), three or more subunits (red), four or more subunits (green),five or more subunits (blue), and six (pink) mutant subunits. Experiments wereperformed in buffer 50 mM Tris–HCl, 150 mM KCl, 20 mM MgCl2, 2 mM DTT,10 mM ATP, pH 7.5. (B) Protection against aggregation of 6 lM MDH by 6 lM T1T2monomer, as a function of ATP concentration. Analysis of the data with the Hillequation gives Kd = 59 ± 3 lM and n = 3.9 (solid line). Data are means ± S.E.M.(n = 2–3).

932 U. del Castillo et al. / FEBS Letters 584 (2010) 929–934

of having one NBD inactivated. This suggests, as previously re-ported [12], that communication between both NBDs regulatestheir ATPase activity. The T1 mutant shows a residual chaperoneactivity with MDH as substrate that recovers upon reconstitution,similarly to wt ClpB (Fig. 1B), as it does the ability of T1T2 to formstable complexes with substrates in the presence of ATP (Fig. 1C).Size exclusion chromatography also indicates that all mutants, ex-cept those unable to bind nucleotides at NBD1, form stable hexa-mers at 150 mM KCl and approximately 5 lM protein monomer(not shown). N1T2 was able to oligomerize at lower ionic strength(i.e., 50 mM KCl), and therefore hybrids containing this mutantwere prepared and characterized at this salt concentration. Thus,under the experimental conditions used in these experiments wtClpB and the different protein variants have a similar tendencyto form hexamers, and their refolded conformations retain na-tive-like functional properties.

We first analyzed the effect of incorporating mutant subunits inwhich both NBDs were mutated, either with a trap or null substi-tution, into the protein hexamer. An inhibition of the chaperoneactivity is observed at the lowest mutant:wt molar ratio forT1T2, T1N2 at 150 mM KCl, and for N1T2 at 50 mM KCl (Fig. 2).A model assuming that hexamers are inactive when one mutantsubunit is intercalated into the ClpB hexamer nicely fits the exper-imental data, indicating that a single subunit of these mutantsabolishes the chaperone activity of the protein (Fig. 3). Since ClpBvariants with a trap (T) mutations in both rings are able to stablybind substrate proteins [28], this inhibition could be due to forma-tion of a stable complex with the substrate that prevents the fold-ing process. Alternatively, mutant subunits could impair thethreading mechanism and therefore the chaperone-mediatedaggregate reactivation. To decipher which alternative was takingplace, the ability to prevent aggregation of thermally denaturedMDH of wt:T1T2 heterooligomers was studied. At a wt:mutantsubunit molar ratio of 4:2, when the chaperone activity of hybridoligomers was almost completely lost (Fig. 2A), the heterooligo-mers were unable to prevent aggregation of denatured MDH, aswt ClpB (Fig. 3A). These data strongly suggest that the inabilityof these hybrids to reactivate protein aggregates is not due to for-mation of stable complexes with protein substrates but rather toan impairment of the conformational change that triggers sub-strate translocation. The similar refolding inhibitory effect exertedby hybrids containing N1T2 subunits, unable to form stable com-plexes with substrates (not shown), supports this hypothesis. Theestimated protection percentages assuming that hybrid hexamerscan stably bind substrates when a certain number of T1T2 mutantsubunits are incorporated into the hexamer are shown in Fig. 3A.Comparison of these values with the experimental data revealsthat incorporation of at least four mutant protomers within thehexamer is sufficient to interact in a stable manner with the sub-strate (Fig. 3A). Furthermore, formation of T1T2–substrate com-plex, measured as protection against aggregation of MDH atconstant mutant subunit concentration (6 lM), was followed as afunction of ATP concentration (Fig. 3B). Data analysis with the Hillequation gives Kd = 59 ± 3 lM and n = 4, suggesting again that atleast four chaperone subunits are involved in substrate binding.

We next studied the ATPase activity of these heterohexamers(Fig. 2). Incorporation of increasing amounts of T1T2 subunits re-sults in a bell-shaped activation of the wt subunits regardless ofthe presence of substrate protein, that reaches its maximum at3:3 molar ratio (Fig. 2B). However, intercalation of T1N2 (Fig. 2D)or N1T2 (Fig. 2F) protomers does not significantly modify theiractivity, suggesting that the activation effect is elicited only bythe ATP-bound T1T2 symmetric conformation. The comparison be-tween the chaperone and ATPase activities of these hybrids indi-cates that their disaggregating activity does not depend on howthe active protomers within the hexamers hydrolyze ATP, but

rather on intersubunit communication. This communication is bro-ken when subunits defective in ATP binding or hydrolysis in bothNBDs are incorporated into the hexamers.

A different picture is observed for the chaperone activity of het-erohexamers composed of wt ClpB and variants in which only oneNBD is mutated (T1, T2 and N2) (Fig. 4). Intercalation of mutantprotomers able to hydrolyze ATP in one binding site within thehexamer reduces reactivation of MDH aggregates to a lower extentthan ATPase inactive mutants (Fig. 4), as recently observed [18].The experimental ATPase activity values of these hybrids agreewith those obtained by the arithmetic combination of the activityof the corresponding homohexamers, indicating that communica-tion that controls ATPase activity of the active nucleotide bindingsites is maintained in the heterooligomers (not shown). We nextcompare the experimental chaperone activity of the hybrid oligo-mers with the values theoretically estimated assuming that the ac-tive proteins require the presence of a given number of wt subunitswithin the hexamer (Fig. 4A–C). This comparison indicates that hy-brid oligomers containing at least four active subunits can disag-gregate and reactivate MDH, provided that the mutant subunitshydrolyze ATP in one NBD. The coincidence of the number of sub-units involved in substrate binding and threading is expected sinceto process the substrate, the hexamer must first interact with it.The similarity of the chaperone activity values presented hereinand those obtained in a recent study by mixing homooligomers[18], demonstrates that subunit exchange is a fast [16,17] processthat efficiently shuffles protomers of different oligomers.

It is interesting to note that the efficiency of mutant subunits topoison the chaperone activity of the wt hexamers depends on hownucleotide is hydrolyzed in each protein ring. Intercalation of one

U. del Castillo et al. / FEBS Letters 584 (2010) 929–934 933

protomer of mutants T1T2, T1N2 and N1T2 abolishes the activity,while the wt hexamer can accommodate two subunits of ClpB vari-ants that hydrolyze ATP at either ring of the protein without loos-ing its reactivation activity. ClpB needs to couple hydrolysis in bothnucleotide binding rings to translocate the unfolded substratesince the chaperone activity of mutants that hydrolyze ATP in onlyone ring of the protein is either abolished or severely impaired[12]. Our data suggest that the rigid, ‘‘frozen” mutant conforma-tions, either symmetric (T1T2) or asymmetric (T1N2 and N1T2),would hamper the conformational change that triggers inter-ringcommunication and thus couples ATP hydrolysis in both NBDs.The structural plasticity required to generate the inter-ring signalwould be achieved with only one fully active in ATP hydrolysis,‘‘conformationally alternating” ring and four active nucleotidebinding sites in the other ring. Thus, four subunits might buildthe functional cooperative unit that binds and process the sub-strate, if one of the protein rings contains active nucleotide bindingsites. A similar number of active subunits has been reported for an-other two hexameric unfoldases, ClpX [29] and HslU [30]. Althoughthey both could bind six ATPs in solution, they appear to bind amaximum of four nucleotide molecules. The recent crystal struc-tures of nucleotide-free and nucleotide-bound ClpX hexamers re-veals a striking asymmetry due to differences in rotationbetween the large and small NBDs of individual subunits, which re-sult in two subunits that cannot bind nucleotide [31]. In contrast,these two subunits in the ClpB hexamer must be able to hydrolyze,

Fig. 4. Chaperone activity of hybrid ClpB hexamers containing wt ClpB and mutantsubunits able to hydrolyze ATP at either of the nucleotide binding rings of theprotomer. Effect of increasing the amount of T1 (A), T2 (B) and N2 (C) subunits inthe wt ClpB hexamer on substrate reactivation. Experimental values (filled circles)are compared with the theoretical ones estimated assuming that a hybrid hexamercan reactivate MDH aggregates when it contains one or more subunits (black),two or more subunits (dark yellow), three or more subunits (red), four or moresubunits (green), five or more subunits (blue), and six (pink) wt subunits [16]. WtClpB reactivates 64 ± 4% MDH under these experimental conditions. Data aremeans ± S.E.M. (n = 3).

not only to bind, ATP in one NBD. This difference is functionallyimportant, since it allows the rearrangement of the ClpB subunitsthat couples the power stroke at both nucleotide binding rings toprovide vectorial work on polypeptide substrate. The fact thatthe fully active ring can be any of the two protein rings also sup-ports that inter-ring communication can be established in bothdirections, and therefore that it could be used by the protein toconcert ATP hydrolysis at both NBDs.

Acknowledgements

We thank J.M. Vilar for generating mathematical models, A.Mogk for the ClpB plasmid and E. Bilbao for technical assistance.Financial support from the Ministerio de Educación y Ciencia(Grant BFU2007-64452), the Universidad del País Vasco and Gobi-erno Vasco (Grant IT-358-07) and Diputación Foral de Bizkaia(Grant DIPE08/18) is gratefully acknowledged. S.P.A. and U. del C.were recipients of fellowships from Gobierno Vasco and Ministeriode Educación y Ciencia, respectively. F.M. is supported by a Ramóny Cajal contract.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.febslet.2010.01.029.

References

[1] Maurizi, M.R. and Xia, D. (2004) Protein binding and disruption by Clp/Hsp100chaperones. Structure 12, 175–183.

[2] Sauer, R.T. et al. (2004) Sculpting the proteome with AAA(+) proteases anddisassembly machines. Cell 119, 9–18.

[3] Mogk, A., Haslberger, T., Tessarz, P. and Bukau, B. (2008) Common and specificmechanisms of AAA+ proteins involved in protein quality control. Biochem.Soc. Trans. 36, 120–125.

[4] Glover, J.R. and Lindquist, S. (1998) Hsp104, Hsp70, and Hsp40: a novelchaperone system that rescues previously aggregated proteins. Cell 94, 73–82.

[5] Goloubinoff, P., Mogk, A., Zvi, A.P., Tomoyasu, T. and Bukau, B. (1999)Sequential mechanism of solubilization and refolding of stable proteinaggregates by a bichaperone network. Proc. Natl. Acad. Sci. USA 96, 13732–13737.

[6] Motohashi, K., Watanabe, Y., Yohda, M. and Yoshida, M. (1999) Heat-inactivated proteins are rescued by the DnaK. J-GrpE set and ClpBchaperones. Proc. Natl. Acad. Sci. USA 96, 7184–7189.

[7] Zolkiewski, M. (1999) ClpB cooperates with DnaK, DnaJ, and GrpE insuppressing protein aggregation. A novel multi-chaperone system fromEscherichia coli. J. Biol. Chem. 274, 28083–28086.

[8] Schirmer, E.C., Glover, J.R., Singer, M.A. and Lindquist, S. (1996) HSP100/Clpproteins: a common mechanism explains diverse functions. Trends Biochem.Sci. 21, 289–296.

[9] Beinker, P., Schlee, S., Groemping, Y., Seidel, R. and Reinstein, J. (2002) The Nterminus of ClpB from Thermus thermophilus is not essential for the chaperoneactivity. J. Biol. Chem. 277, 47160–47166.

[10] Barnett, M.E., Nagy, M., Kedzierska, S. and Zolkiewski, M. (2005) The amino-terminal domain of ClpB supports binding to strongly aggregated proteins. J.Biol. Chem. 280, 34940–34945.

[11] Barnett, M.E., Zolkiewska, A. and Zolkiewski, M. (2000) Structure and activityof ClpB from Escherichia coli. Role of the amino-and -carboxyl-terminaldomains. J. Biol. Chem. 275, 37565–37571.

[12] Mogk, A., Schlieker, C., Strub, C., Rist, W., Weibezahn, J. and Bukau, B. (2003)Roles of individual domains and conserved motifs of the AAA+ chaperone ClpBin oligomerization, ATP hydrolysis, and chaperone activity. J. Biol. Chem. 278,17615–17624.

[13] Kedzierska, S., Akoev, V., Barnett, M.E. and Zolkiewski, M. (2003) Structure andfunction of the middle domain of ClpB from Escherichia coli. Biochemistry 42,14242–14248.

[14] Schirmer, E.C., Homann, O.R., Kowal, A.S. and Lindquist, S. (2004) Dominantgain-of-function mutations in Hsp104p reveal crucial roles for the middleregion. Mol. Biol. Cell 15, 2061–2072.

[15] Lee, U., Wie, C., Escobar, M., Williams, B., Hong, S.W. and Vierling, E. (2005)Genetic analysis reveals domain interactions of Arabidopsis Hsp100/ClpB andcooperation with the small heat shock protein chaperone system. Plant Cell17, 559–571.

[16] Werbeck, N.D., Schlee, S. and Reinstein, J. (2008) Coupling and dynamics ofsubunits in the hexameric AAA+ chaperone ClpB. J. Mol. Biol. 378, 178–190.

[17] Haslberger, T., Zdanowicz, A., Brand, I., Kirstein, J., Turgay, K., Mogk, A. andBukau, B. (2008) Protein disaggregation by the AAA+ chaperone ClpB involves

934 U. del Castillo et al. / FEBS Letters 584 (2010) 929–934

partial threading of looped polypeptide segments. Nat. Struct. Mol. Biol. 15,641–650.

[18] Hoskins, J.R., Doyle, S.M. and Wickner, S. (2009) Coupling ATP utilization toprotein remodeling by ClpB, a hexameric AAA+ protein. Proc. Natl. Acad. Sci.USA. pnas.org/cgi/doi/10.1073/pnas.0911937106.

[19] Woo, K.M., Kim, K.I., Goldberg, A.L., Ha, D.B. and Chung, C.H. (1992) The heat-shock protein ClpB in Escherichia coli is a protein-activated ATPase. J. Biol.Chem. 267, 20429–20434.

[20] Moro, F., Fernandez, V. and Muga, A. (2003) Interdomain interaction throughhelices A and B of DnaK peptide binding domain. FEBS Lett. 533, 119–123.

[21] Zylicz, M., Yamamoto, T., McKittrick, N., Sell, S. and Georgopoulos, C. (1985)Purification and properties of the dnaJ replication protein of Escherichia coli. J.Biol. Chem. 260, 7591–7598.

[22] Mehl, A.F., Heskett, L.D. and Neal, K.M. (2001) A GrpE mutant containing theNH(2)-terminal ‘‘tail” region is able to displace bound polypeptide substratefrom DnaK. Biochem. Biophys. Res. Commun. 282, 562–569.

[23] Jordan, R. and McMacken, R. (1995) Modulation of the ATPase activity of themolecular chaperone DnaK by peptides and the DnaJ and GrpE heat shockproteins. J. Biol. Chem. 270, 4563–4569.

[24] Norby, J.G. (1988) Coupled assay of Na+, K+-ATPase activity. MethodsEnzymol. 156, 116–119.

[25] Schlieker, C., Tews, I., Bukau, B. and Mogk, A. (2004) Solubilization ofaggregated proteins by ClpB/DnaK relies on the continuous extraction ofunfolded polypeptides. FEBS Lett. 578, 351–356.

[26] Shiseki, K., Murai, N., Motojima, F., Hisabori, T., Yoshida, M. and Taguchi, H.(2001) Synchronized domain-opening motion of GroEL is essential forcommunication between the two rings. J. Biol. Chem. 276, 11335–11338.

[27] Sot, B., Galan, A., Valpuesta, J.M., Bertrand, S. and Muga, A. (2002) Salt bridgesat the inter-ring interface regulate the thermostat of GroEL. J. Biol. Chem. 277,34024–34029.

[28] Weibezahn, J., Schlieker, C., Bukau, B. and Mogk, A. (2003) Characterization ofa trap mutant of the AAA+ chaperone ClpB. J. Biol. Chem. 278, 32608–32617.

[29] Hersch, G.L., Burton, R.E., Bolon, D.N., Baker, T.A. and Sauer, R.T. (2005)Asymmetric interactions of ATP with the AAA+ ClpX6 unfoldase: allostericcontrol of a protein machine. Cell 121, 1017–1027.

[30] Yakamavich, J.A., Baker, T.A. and Sauer, R.T. (2008) Asymmetric nucleotidetransactions of the HslUV protease. J. Mol. Biol. 380, 946–957.

[31] Glynn, S.E., Martin, A., Nager, A.R., Baker, T.A. and Sauer, R.T. (2009) Structuresof asymmetric ClpX hexamers reveal nucleotide-dependent motions in a AAA+protein-unfolding machine. Cell 139, 744–756.