Conserved residues in the N-domain of the AAA+ chaperone ClpA regulate substrate recognition and unfolding Annette H. Erbse 1, *, Judith N. Wagner 1, *, Kaye N. Truscott 2 , Sukhdeep K. Spall 2 , Janine Kirstein 1,3 , Kornelius Zeth 4 ,Ku ¨ rsad Turgay 1,3 , Axel Mogk 1 , Bernd Bukau 1 and David A. Dougan 1,2 1 Zentrum fu ¨ r Molekulare Biologie Heidelberg, Universita ¨ t Heidelberg, Heidelberg, Germany 2 Department of Biochemistry, La Trobe University, Melbourne, Australia 3 Institut fu ¨ r Biologie, Freie Universita ¨ t Berlin, Berlin, Germany 4 MPI fu ¨ r Entwicklungsbiologie, Tubingen, Germany The AAA+ superfamily [1] is an extensive group of proteins involved in a broad range of biological func- tions. Its members are present in all kingdoms of life and often play a crucial role in cell maintenance. In bacteria, several AAA+ proteins (e.g. ClpA, ClpB, ClpX, HslU and Lon) are central to the protein qual- ity-control network [2]. They employ a common mech- anism, involving the binding and hydrolysis of ATP, to mediate the unfolding ⁄ disassembly of a variety of proteins, including large macromolecular complexes [3]. Although several of these proteins share consider- able sequence similarity, they demonstrate distinct substrate specificity. For example, in Escherichia coli, ClpA is responsible, either directly or indirectly via the adaptor protein ClpS, for recognition of substrates such as SsrA-tagged proteins or N-end rule substrates Keywords AAA+; binding; ClpA; SsrA; unfolding Correspondence D. A. Dougan, Department of Biochemistry, La Trobe University, Melbourne 3086, Australia Fax: +61 3 9479 2467 Tel: +61 3 9479 3276 E-mail: d.dougan@latrobe.edu.au B. Bukau, Zentrum fu ¨ r Molekulare Biologie Heidelberg, Universita ¨ t Heidelberg, INF 282, Heidelberg D-69120, Germany Fax: +49 6221 54 5894 Tel: +49 6221 54 6795 E-mail: bukau@zmbh.uni-heidelberg.de *These authors contributed equally to this work (Received 22 November 2007, revised 10 January 2008, accepted 14 January 2008) doi:10.1111/j.1742-4658.2008.06304.x Protein degradation in the cytosol of Escherichia coli is carried out by a variety of different proteolytic machines, including ClpAP. The ClpA com- ponent is a hexameric AAA+ (ATPase associated with various cellular activities) chaperone that utilizes the energy of ATP to control substrate recognition and unfolding. The precise role of the N-domains of ClpA in this process, however, remains elusive. Here, we have analysed the role of five highly conserved basic residues in the N-domain of ClpA by monitor- ing the binding, unfolding and degradation of several different substrates, including short unstructured peptides, tagged and untagged proteins. Inter- estingly, mutation of three of these basic residues within the N-domain of ClpA (H94, R86 and R100) did not alter substrate degradation. In contrast mutation of two conserved arginine residues (R90 and R131), flanking a putative peptide-binding groove within the N-domain of ClpA, specifically compromised the ability of ClpA to unfold and degrade selected substrates but did not prevent substrate recognition, ClpS-mediated substrate delivery or ClpP binding. In contrast, a highly conserved tyrosine residue lining the central pore of the ClpA hexamer was essential for the degradation of all substrate types analysed, including both folded and unstructured proteins. Taken together, these data suggest that ClpA utilizes two structural ele- ments, one in the N-domain and the other in the pore of the hexamer, both of which are required for efficient unfolding of some protein substrates. Abbreviations AAA+, ATPase associated with various cellular activities; FITC, fluorescein isothiocyanate; GFP, green fluorescent protein; kR, lambda repressor. 1400 FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS [4,5]. Once recognized, these substrates are unfolded by the AAA+ protein, in an ATP-dependent manner, and translocated through the central pore of the oligo- mer into the associated ClpP peptidase, where they are degraded into short peptides. AAA+ proteins usually contain an N-terminal domain (N-domain) that serves as a docking site for various adaptor proteins [6–10]. ClpA consists of three domains: an N-domain and two ATP-binding domains referred to as the D1 and D2 domains. Interestingly, deletion of the N-domain from ClpA not only abol- ishes binding of the adaptor protein, ClpS, but addi- tionally modulates ClpA substrate specificity [8,11–13]. This change in substrate specificity is poorly under- stood, and the mechanism by which the N-domains might regulate ClpA function is controversial, although it has been proposed that the N-domain controls bind- ing of ClpA to ClpP [14]. Interestingly, there is also considerable debate regarding the role of the ClpB N-domain (which shares a common fold with the N-domain of ClpA) in substrate selection [15–17]. One difficulty in understanding the role of the N-domain of ClpA stems from the variety of activities exhibited by various DNClpA constructs tested, each containing different lengths of ‘linker’ residues that connect the N-domain to the D1 domain. In order to avoid the potential problems associated with ‘ragged’ ends of DNClpA, we chose to create several single and double point mutations within the N-domain to probe N-domain function. Here, using mutational analysis, we report the iden- tification of a structural element composed of con- served basic amino acids (R90 and R131), located within the N-domain of ClpA, that dramatically alters the ability of ClpA to degrade selected substrates. This element, although dispensable for the recognition of the SsrA tag per se, modulates the binding, unfolding and subsequent degradation of SsrA-tagged protein substrates. We propose that this element plays an important role in the binding and subsequent release of substrates, by triggering ‘local’ unfolding of the sub- strate. We speculate that the ATP-dependent global unfolding of some protein substrates is initiated through productive binding to the substrate via two elements in ClpA, one in the N-domain and the other in the pore of the ClpA hexamer. In the case of short unstructured peptides or unfolded proteins such as casein, binding to the tyrosine residues in the hexamer- ic pore of ClpA is sufficient for substrate translocation to occur; however, in other cases such as SsrA-tagged protein substrates, binding at both sites is required for translocation-mediated global unfolding to proceed efficiently. Results Two conserved arginine residues (R90 and R131) within the N-domain are required for full ClpA function We were interested to understand how substrates are recognized and subsequently unfolded by ClpA. As mutation of the tyrosine residue located in the pore has been demonstrated to inhibit degradation of all substrates tested [18], we postulated that substrate dis- crimination must arise from an alternative region within ClpA. Based on previous findings showing that deletion of the N-domain of ClpA dramatically reduced the rate of degradation of GFP–ssrA and to a lesser extent casein [8,11], we speculated that the N-domain facilitates an early binding step, contribut- ing to specific recognition of substrates such as SsrA- tagged proteins. In order to further study the role of the N-domains in substrate recognition, we compared the amino acid sequences of this region in several AAA+ proteins (Fig. 1). From this analysis, we noted a high occurrence of conserved basic residues distrib- uted throughout the domain, several of which (R86, R90, H94, R100 and R131) flanked a hydrophobic groove (Fig. 2A). To test the role of these basic resi- dues, we constructed a number of single (R86A, R90A and R131A) and double (H94A ⁄ R100A and R90A ⁄ R131A) point mutations in the N-domain of ClpA (Fig. 2A). First, we compared the degradation of SsrA-tagged GFP by wild-type and mutant ClpAP complexes (Fig. 2B). The ClpP-dependent degradation of GFP– ssrA mediated by either the single mutant R86A (Fig. 2B, open inverted triangles) or the double mutant H94A ⁄ R100A (Fig. 2B, filled diamonds) was unaf- fected. In contrast the rate of ClpP-mediated degrada- tion by the single mutants R90A (Fig. 2B, open diamonds) and R131A (Fig. 2B, open triangles) was reduced approximately threefold when compared to wild-type ClpA (Fig. 2B, open circles). Interestingly, when we combined these two single point mutants to create the double mutant R90A ⁄ R131A (herein referred to as RR ⁄ AA), the degradation of GFP–ssrA was reduced dramatically (Fig. 2B, filled circles). Although these mutant proteins exhibited different abilities with regard to mediation of GFP–ssrA degradation (Fig. 2B), the basal ATPase activity was not affected (Fig. 3E, compare lanes 1 and 4). Given that the ATPase activity of ClpA is dependent on its oligomeri- zation [19], as the nucleotide is bound between two adjacent subunits, this result suggests that the overall hexameric structure of RR ⁄ AA was maintained. A. H. Erbse et al. Substrate recognition and unfolding by ClpA FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS 1401 To determine whether this dramatic change in ClpP- mediated degradation of GFP–ssrA by RR ⁄ AA was simply due to an inability to bind ClpP, we performed co-immunoprecipitation experiments using a-ClpP anti- serum (Fig. 2C). The co-immunoprecipitation of ClpA with the a-ClpP antiserum was specific, as recovery of ClpA required the addition of both ATPcS (a non- hydrolysable analogue of ATP) and ClpP (Fig. 2C, lane 3). Importantly, RR ⁄ AA (Fig. 2C, lane 4) did not show any change in ClpP interaction when compared to wild-type ClpA (Fig. 2C, lane 3), as determined by quantification of ClpA amounts after co-immunopre- cipitation (Fig. 2C, lower panel), suggesting that the overall structure of the RR ⁄ AA mutant is not compro- mised. Likewise, the other N-domain mutants tested (i.e. R86A and H94A ⁄ R100A) also exhibited wild-type ClpA behaviour (data not shown). An alternative explanation for the lack of GFP–ssrA degradation exhibited by RR ⁄ AA could be that the N-domain was structurally compromised as a result of mutations in this region. To confirm that neither the N-domain structure nor the overall structure of these mutant proteins were adversely affected, we tested the degradation of a model N-end rule substrate, FR-lin- ker–GFP [20]. The ClpP-mediated degradation of this substrate class requires specific interaction between ClpS and the N-domain of ClpA. Consequently, dra- matic changes to the structure of the N-domain of ClpA would inhibit ClpS binding and thereby ClpS- dependent degradation of this substrate. As expected, the ClpP-mediated degradation of FR-linker–GFP by wild-type ClpA required the addition of ClpS (Fig. 2D). Importantly, like wild-type ClpA, RR ⁄ AA was also able to support the ClpS-dependent degrad- ation of FR-linker–GFP (Fig. 2D), demonstrating a functional interaction between ClpS and the N-domain of RR ⁄ AA, and this result suggests that neither the local nor the overall structure of the RR ⁄ AA mutant was compromised. Mutation of the conserved arginine residues has only a moderate effect on degradation of short unstructured peptides and the model unfolded protein, casein To determine whether RR ⁄ AA also demonstrated an inability to degrade other known ClpAP substrates, we examined the ClpP-dependent degradation of several model ClpA substrates, including the N-terminal domain of the k repressor fused to the SsrA tag (kR– ssrA) [21], two short peptides, and the model unfolded protein a-casein [22]. As for GFP–ssrA, the rate of fluorescein-labelled kR–ssrA degradation mediated by RR ⁄ AA was dramatically reduced when compared to wild-type ClpA (Fig. 3A, filled circles and open circles). Interestingly, the rate of RR ⁄ AA-mediated degradation was not significantly altered for an SsrA- tagged peptide (Fig. 3B), indicating that recognition of the SsrA tag is not affected by RR ⁄ AA. Moreover, two other unfolded substrates, a-casein (Fig. 3C) and a 21-amino-acid polypeptide derived from r 32 (a loosely folded protein) [23] (Fig. 3D), were also degraded by RR ⁄ AA with similar kinetics to wild-type ClpA, in a ClpP-dependent manner. In contrast to the Fig. 1. Multiple sequence alignment of the N-domain of bacterial ClpA homologues and E. coli ClpB. Amino acid sequences of the N-domain of ClpA from E. coli (P0ABH9), V. cholera (Q9KSW2), P. aeruginosa (Q9I0L8), X. fastidosa (Q87DL7), B. japonicum (Q89JW6), C. crescentus (Q9A5H9), N. meningitidis (Q9JZZ6), D. radiodurans (Q9RWS7), C. acetobutylicum (Q97I30) and H. pylori (O24875) were aligned together with the amino acid sequence of the N-domain of E. coli ClpB (P63284). Conserved hydrophobic residues are highlighted in grey, conserved basic residues are highlighted in blue, and conserved acidic residues are highlighted in red. Amino acid numbering corresponds to the ClpA sequence from E. coli. Residues chosen for mutation are indicated by asterisks. Substrate recognition and unfolding by ClpA A. H. Erbse et al. 1402 FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS substrate-dependent degradation exhibited by RR ⁄ AA, the ClpA pore mutant (Y259A), which is unable to translocate GFP–ssrA [18], prevented the degradation of both kR–ssrA (Fig. 3A, open diamonds) and fluo- rescein isothiocyanate-labelled casein (FITC-casein; data not shown). Consistent with these results, casein stimulated the ATPase activity of both wild-type ClpA and RR ⁄ AA (Fig. 3E), while, in contrast, SsrA-tagged GFP only stimulated the ATPase activity of wild-type ClpA (Fig. 3E). Together, these data suggest that RR ⁄ AA has a reduced ability to initiate unfolding of more tightly folded proteins, but retains full ability to translocate short unstructured peptides and model unfolded proteins into the ClpP chamber for degrada- tion. RR ⁄ AA delays the release and subsequent unfolding of certain protein substrates Before testing the unfolding activity of RR ⁄ AA, we wished to compare the ability of the RR ⁄ AA mutant to bind to the various substrates tested. To do this, we constructed a ClpA variant in which the glutamic acid residue within the Walker B motif of each AAA domain (E286, E565) was changed to alanine. This double Walker B mutant (herein referred to as dWB) Fig. 2. Two conserved arginine residues flanking a hydrophobic groove are essential for N-domain function. (A) Structure of the ClpA N- domain. ClpA is shown as a ribbon diagram (dark grey), and the side chains of R86, R90, H94, R100 and R131 are represented as a ball and stick (blue) flanking the putative peptide-binding groove (orange). The surface of the N-domain is shaded light grey, and R86, R90, H94, R100 and R131 are highlighted in blue. (B) The ClpP-mediated degradation of GFP–ssrA was monitored by fluorescence in the presence of wild-type ClpA (open circles), R86A (inverted open triangles), H94A ⁄ R100A (filled diamonds), R90A (open diamonds), R131A (open triangles) and RR ⁄ AA (filled circles). (C) The interaction between wild-type ClpA (lane 3) or RR ⁄ AA (lane 4) with ClpP, assessed by co-immunoprecipita- tion using a-ClpP antiserum, was visualized by staining of the protein bands using Coomassie brilliant blue following separation by SDS– PAGE. In the absence of added ATPcS (lane 1) or ClpP (lane 2), ClpA was not co-precipitated. The relative amount of ClpA binding to ClpP was determined from quantification of three independent experiments. Error bars represent the standard error of the mean. A non-specific protein band is indicated by an asterisk. (D) The functional interaction between ClpS and ClpA (wild-type and RR ⁄ AA) was observed by moni- toring the ClpS-dependent degradation of FR-linker–GFP (in the presence of ClpP). A. H. Erbse et al. Substrate recognition and unfolding by ClpA FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS 1403 and the corresponding mutant in RR⁄ AA (referred to as RR ⁄ dWB) were used to monitor substrate binding as determined by co-elution of substrate–ClpA complexes during gel filtration. Initially, we tested the ability of dWB and RR ⁄ dWB to interact with FITC- casein. As a control, in the absence of ClpA, FITC-casein eluted in a single peak at 21.5 mL (Fig. 4A, open circles). However, upon addition of ATP and dWB (Fig. 4A, open triangles) or RR ⁄ dWB (Fig. 4A, filled diamonds), the FITC-casein peak shifted and formed two new peaks, the largest of which co-eluted with the ClpA hexamer (Fig. 4A, grey block). Quantification of this peak indicated that approximately 30 and 40 pmol of FITC-casein were bound to the hexamers of dWB and RR ⁄ dWB respec- tively. Next we compared the ability of kR–ssrA (Fig. 4B) and GFP–ssrA (Fig. 4C) to bind to dWB or RR ⁄ dWB. As controls, each substrate (in the absence of ClpA) was also separated by gel filtration and the amount of substrate was quantified in the hexamer region of the gel filtration profile (Fig. 4B,C, lane 1). Similarly, as a further control, each substrate in the presence of dWB (Fig. 4B,C, lane 2) or RR ⁄ dWB (Fig. 4B,C, lane 4) was also quantified after separation by gel filtration in the absence of ATP. These controls demonstrated a strict requirement for ATP in the interaction between dWB ClpA and each substrate tested. Interestingly, under the same conditions, although very little change in the binding of FITC- casein was observed, approximately threefold more Fig. 3. RR ⁄ AA exhibits different abilities with regard to degradation of various ClpA substrates. (A) ClpP-mediated degradation of fluores- cein-labelled kR–ssrA by ClpA (open circles), RR ⁄ AA (closed circles) and Y259A (open diamonds) was monitored by an increase in fluores- cence (excitation at 490 nm and emission at 520 nm). (B) ClpP-mediated degradation of a SsrA tagged peptide (50 l M) was monitored in the absence of ClpA (ClpP) and the presence of wild-type (ClpA) or mutant (RR ⁄ AA) proteins. (C) Time course of a-casein degradation by ClpA or RR ⁄ AA in the presence of ClpP. (D) ClpP-mediated degradation of a short unstructured peptide derived from r 32 (QRKLFFNLEKTKQRLGWFNQC) by RR ⁄ AA is not compromised. ClpP-mediated degradation of the peptide (50 lM) was monitored over time in the presence of wild-type ClpA (open circles) or RR ⁄ AA (filled circles). The amount of peptide remaining was determined by quantification of the Coomassie-stained band following separation of the proteins by Tris ⁄ Tricine SDS–PAGE. (E) The ATPase activity of wild-type ClpA (lanes 1–3), RR ⁄ AA (lanes 4–6) and Y259A (lanes 7–9) was determined either in the absence of substrate (white bars; lanes 1, 4 and 7, respectively) or in the presence of GFP–ssrA (grey bars; lanes 2, 5 and 8, respectively) or a-casein (black bars; lanes 3, 6 and 9, respec- tively). The ATPase activity (relative to ClpA in the absence of substrate) was determined from three independent experiments. Error bars represent the standard error of the mean. Substrate recognition and unfolding by ClpA A. H. Erbse et al. 1404 FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS SsrA-tagged substrate co-eluted with RR ⁄ dWB when compared to dWB ClpA (Fig. 4B,C). These data are consistent with the notion that RR ⁄ AA is able to bind each substrate but exhibits a change in the release of some substrates (e.g. kR–ssrA). This lack of release is expected to hinder unfolding and ultimately reduce degradation of the substrate. To further test the possibility that RR ⁄ AA has a compromised unfolding activity, we compared the abil- ity of wild-type ClpA and RR ⁄ AA to unfold SsrA- tagged GFP in the presence of the GroEL trap [24]. As expected wild-type ClpA, in the absence of ClpP, was able to unfold GFP–ssrA (Fig. 5A, open circles) but the unfolding ability of RR ⁄ AA (Fig. 5A, filled circles) was strongly compromised. Surprisingly, the kinetics of unfolding by RR ⁄ AA measured using the GroEL trap were slower than expected. As this method does not directly measure the change in sub- strate conformation and may be affected by rapid refolding of the substrate, we chose to validate this finding using a more sensitive and direct approach. Thus, hydrogen–deuterium exchange was used to mea- sure the unfolding of GFP–ssrA in the presence and absence of either wild-type ClpA or RR⁄ AA. Follow- ing incubation of GFP–ssrA (28 954 Da) in deuterated buffer, the mass of GFP–ssrA rapidly increased to 29 034 Da within the first 5 min of the experiment. This change in mass occurred in the absence (data not shown) and the presence of wild-type or mutant ClpA (indicated by the hash symbol, #, in Fig. 5B,C), and resulted from the rapid exchange of 80 accessible amide protons. In the absence of ClpA, the remaining amide protons within the protected core did not exchange over a period of 2 h (data not shown). In the Fig. 4. Mutations in the N-domain do not prevent substrate interaction. (A) FITC-casein (500 pmol) was separated by gel filtration in the presence of 2 m M ATP (open circles), 160 pmol dWB ClpA 6 plus 2 mM ATP (open triangles) or 160 pmol RR ⁄ dWB ClpA 6 plus 2 mM ATP (filled diamonds) as described in Experimental procedures. The molecular mass standards thyroglobulin (669 kDa), ferritin (440 kDa), aldolase (232 kDa) and ovalbumin (43 kDa) eluted as indicated by the arrows labelled 669, 440, 232 and 43 respectively. The position at which ClpA 6 eluted is indicated with an arrow labelled ClpA 6 . The amount of casein bound was calculated from the peak elution (boxed in grey) that co- eluted with ClpA 6 . (B) Fluorescein-labelled kR–ssrA (450 pmol) was separated by gel filtration without the addition of ATP (white bar, lane 1), in the presence of 160 pmol dWB ClpA 6 without (lane 2) or with addition of 2 mM ATP (lane 3), or in the presence of 160 pmol RR ⁄ dWB ClpA 6 without (lane 4) or with addition of 2 mM ATP (lane 5) as described in (A). (C) GFP–ssrA (990 pmol) was separated by gel filtration without the addition of ATP (lane 1), in the presence of 160 pmol dWB ClpA 6 without (lane 2) or with addition of 2 mM ATP (lane 3), or in the presence of 160 pmol RR ⁄ dWB ClpA 6 without (lane 4) or with addition of 2 mM ATP (lane 5) as described in (A). A. H. Erbse et al. Substrate recognition and unfolding by ClpA FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS 1405 presence of ClpA and ATP, we observed a further peak (Fig. 5B, asterisk), which arises from incorpora- tion of deuterium into the core region of GFP–ssrA as a result of its unfolding. With time, the relative amount of this heavier species (29 130 Da) increased, reflecting complete unfolding of all GFP–ssrA by ClpA (Fig. 5D, open circles). In contrast, the rate of RR ⁄ AA-mediated unfolding (Fig. 5D, filled circles) was significantly slower than that of wild-type ClpA, with more than half of the GFP–ssrA still folded after 30 min (Fig. 5C, asterisk). Taken together, these data suggest that the change in degradation of GFP–ssrA mediated by RR⁄ AA stems from a delayed release of substrate, which results in reduced unfolding of the substrate. Discussion As for most AAA+ proteases, ClpA utilizes the hydrolysis of ATP to drive substrate unfolding and translocation into the associated peptidase (ClpP). To date, however, the role of the N-domains in this pro- cess has not been well defined as several conflicting roles have been proposed. Despite this, one aspect of the N-domain function is unambiguous – it is essential for ClpS binding and hence the delivery of N-end rule substrates to ClpAP. Currently, much of our mecha- nistic understanding of the ClpAP machine is based largely on the use of model proteins such as casein and GFP–ssrA. Previous studies have demonstrated that a ring of tyrosine residues located in the pore of the ClpA hexamer is essential for the translocation and degradation of all substrates [18]. In contrast, various N-domain deletions of ClpA have exhibited differing affects on substrate degradation [8,11,12], which may simply result from reduced ClpP interaction [14]. In order to better understand N-domain function, we analysed in detail both the sequence and three-dimen- sional structure of the ClpA N-domain [25]. In this study, we have identified an element within the N-domain of ClpA (composed of two conserved basic residues, R90 and R131) that flanks a hydropho- bic groove. This element, via an unknown mechanism, contributes to the dynamic nature of substrate inter- action with ClpA. In contrast to mutation of the hexameric pore tyrosine residue (which abolishes degradation of all substrates examined), the RR ⁄ AA mutant alters the unfolding of certain substrate types. For example, SsrA-tagged proteins are bound by RR ⁄ AA but release of the substrate is inhibited (Fig. 4). This slow substrate release appears to be spe- cific for SsrA-tagged proteins and was not observed for the model unfolded protein casein or short peptide substrates (including an SsrA-tagged peptide) as deter- mined by rapid degradation of these peptides (Fig. 3B,D). RR ⁄ AA also exhibited a reduced rate of GFP–ssrA unfolding as measured by hydrogen–deute- rium exchange or in the presence of the GroEL trap (Fig. 5). Collectively, these data confirm that the SsrA tag does not bind to the N-domain of ClpA, and sug- gest that these basic residues influence substrate release from the N-domain, which in turn allows substrate unfolding to proceed. Importantly, in contrast to previous studies on the N-domain of ClpA (which examined the effect of removing the entire domain and resulted in dramatic affects on ATPase activity or ClpP binding [14,26]), our site-directed mutagenesis approach has allowed us Fig. 5. Mutations in the N-domain reduce substrate unfolding. Unfolding of GFP–ssrA was monitored (A) in the presence of the GroEL trap D87K upon the addition of ClpA (open circles) or RR ⁄ AA (closed circles), (B) by hydrogen–deuterium exchange in D 2 O buffer in the presence of ClpA and ATP, or (C) by hydrogen–deuterium exchange in D 2 O buffer in the presence of RR ⁄ AA and ATP. (D) The relative amount of ‘unfolded’ GFP–ssrA (29 130 Da) was determined in the presence of wild-type ClpA (open circles) or RR ⁄ AA (filled circles). Substrate recognition and unfolding by ClpA A. H. Erbse et al. 1406 FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS to specifically probe N-domain function. The RR ⁄ AA mutant not only displays normal basal ATPase activity but also retains the ability to interact with both ClpS and ClpP and most importantly is able to translocate short peptides and unfolded protein substrates into the ClpP chamber for degradation. Despite these normal activities, RR ⁄ AA shows a dramatic decrease in the ability to degrade SsrA-tagged proteins (Figs 2 and 3). Nevertheless, this defect does not result from a simple inability to interact with the SsrA tag (as demonstrated by efficient degradation of the SsrA peptide by RR ⁄ AA and stable binding of two SsrA-tagged pro- teins to RR⁄ dWB). In contrast, this region appears to be involved in a more general binding task, which is consistent with previous findings [18,27], and moreover may modulate the ability of ClpA to bind, unfold and ultimately degrade substrates such as GFP–ssrA. Interestingly, ClpS was observed to interact directly with R131 in a ClpS–ClpA complex [28]. However, this interaction is not required for ClpS-mediated action [25], and hence most likely mimics a substrate interaction, suggesting that R131 may interact directly with some substrates. Nevertheless, using an in vitro crosslinking approach [29], which permits the detection of dynamic interactions, we did not observe an inter- action between various N-domain residues (His94, Leu109 and Val134) located in close proximity to R90 and R131 and a substrate (data not shown). Impor- tantly, these variants were able to crosslink to ClpS and mediated the degradation of GFP–ssrA and FR- linker–GFP by ClpP (data not shown). Thus it remains unclear whether these arginine residues are directly involved in substrate binding. Of note, although the N-domains of ClpA and p97 are not structurally related, mutations in several basic residues (R95G, R155C, R155H) within the N-domain of p97 have been implicated in the inclusion body myopathy asso- ciated with Paget’s disease of bone and fronto-tempo- ral dementia [30]. Interestingly, in the crystal structure of p97, these basic residues are not surface-exposed but face the AAA domain, in close proximity to the Walker A motif. Hence, it is appealing to speculate that the arginine residues in ClpA do not regulate sub- strate unfolding directly through interaction with the substrate, but instead coordinate substrate bind- ing ⁄ release via an interaction elsewhere in ClpA. Inter- estingly, although RR ⁄ AA has a dramatic inhibitory effect on the degradation of SsrA-tagged GFP, it does not affect binding or delivery of the model N-end rule substrate (FR-linker–GFP) by ClpS (Fig. 2), which suggests one of two possibilities. Firstly, that the defect in RR ⁄ AA-mediated unfolding is not dependent on the global thermodynamic stability of the substrate, but rather correlates with local unfolding of the sub- strate (i.e. unfolding of the N- or C-terminus). This can be understood by examining the N- and C-termi- nal structures of GFP. The first 11 N-terminal residues of GFP form an a-helix, which leads into a parallel b-sheet. In contrast, the last 12 amino acids of GFP form a b-strand leading into an anti-parallel b-sheet [31]. In the case of GFP–ssrA, release of this substrate from ClpA may be more effective than from RR ⁄ AA, allowing ClpA to perform the ATP-dependent unfold- ing step more efficiently. Interestingly, it has been pro- posed [32,33] that an a-helix is more easily unfolded than a b-sheet. Therefore, these data support the idea that RR ⁄ AA has reduced ability to trigger local unfolding of a substrate (at the N- or C-terminus) and are consistent with the idea that local unfolding by the N-domain may be required before global unfolding of the substrate can proceed, as has been suggested for the AAA+ protein PAN [34]. Alternatively, different ClpA substrates may utilize various recogni- tion ⁄ unfolding pathways – some that require these arginine residues in the N-domain (e.g. GFP–ssrA) and other that do not (e.g. N-end rule substrates, delivered by ClpS). Therefore, ClpS-delivered sub- strates may bypass the need for these residues in the N-domain. In this case, it is appealing to speculate that ClpS itself may act as the second binding site required for unfolding, thereby replacing a need for the N-domain. Experimental procedures Proteins ClpA, ClpP and GFP–ssrA were over-expressed from an isopropyl thio-b-d-galactoside-inducible plasmid and puri- fied from the clarified lysates as described previously [8]. All ClpA mutant proteins were purified as for wild-type ClpA. kR–ssrA was labelled with fluorescein as described previously [21]. Purification of FR-linker–GFP was per- formed as previously described [20]. FITC-casein was obtained from Sigma (St Louis, MO, USA). All proteins were > 95% pure as determined by Coomassie-stained SDS–PAGE. Protein concentrations were determined using a Bradford assay system (Bio-Rad, Munich, Germany) using BSA purchased from Pierce (Rockford, IL, USA) as a standard, and refer to the protomer. Unfolding and protein degradation assays GFP–ssrA degradation was monitored by changes in fluorescence (excitation at 400 nm and emission at 510 nm). Degradation of fluorescein-labelled kR–ssrA and A. H. Erbse et al. Substrate recognition and unfolding by ClpA FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS 1407 FITC-labelled a-casein was monitored by changes in fluo- rescence (excitation at 490 nm and emission at 520 nm) using a Perkin-Elmer fluorescence spectrometer LS50B (Waltham, MA, USA). The reactions were carried out as described previously [24]. Unfolding assays for GFP–ssrA were performed in the presence of GroEL trap D87K as previously described [24]. Non-fluorescent degradation assays (GFP–ssrA, a-casein, FR-linker–GFP and peptides) were preformed as previously described [8,20]. Unless other- wise stated, 1 lm ClpA (wild-type and all mutant ClpA proteins) and 1 lm ClpP were used. Samples were removed from the reactions at the indicated time points and degra- dation was stopped by the addition of sample buffer. Pro- tein substrates were separated by 15% SDS–PAGE and peptide substrates by 16.5% Tris ⁄ Tricine SDS–PAGE. Proteins were visualized by Coomassie brilliant blue stain- ing. When required, protein bands were quantified using geleval1.21 (FrogDance Software, Dundee, UK). Analysis of ClpA–ClpP complexes by co-immunoprecipitation To assess ClpA–ClpP complex formation, wild-type or mutant ClpA (1 lm) and ATPcS(2mm ) were preincubated in co-immunoprecipitation buffer (50 mm Tris ⁄ HCl pH 7.5, 300 mm NaCl, 40 mm Mg-acetate, 10% glycerol) at room temperature for 2 min prior to the addition of ClpP (1 lm). After a further 3 min incubation at room temperature, the protein samples were mixed by end-over-end rotation for 1 h at 4 °C with Protein A–Sepharose obtained from Sigma (St Louis, MO, USA) containing pre-bound antibodies against E. coli ClpP. Following removal of the unbound fraction, the Protein A–Sepharose beads were washed three times with ice-cold co-immunoprecipitation buffer contain- ing 10 mm ATP, then bound proteins were eluted with 50 mm glycine, pH 2.5. Proteins were separated by 10% SDS–PAGE and detected by Coomassie brilliant blue staining. Gel filtration and substrate binding Gel filtration was carried out at 4 °C using a Superose 6 column (GE Healthcare, Uppsala, Sweden) in buffer con- taining 20 mm Tris ⁄ HCl pH 7.4, 100 mm KCl, 40 mm NaCl, 10 mm MgCl 2 ,5mm dithiothreitol, 0.1 mm EDTA, 5% glycerol with or without 2 mm ATP. Fractions of 250 lL were collected in a 96-well plate and samples analy- sed by fluorescence and 15% SDS–PAGE. ATPase assay The ATPase activity of wild-type and mutant ClpA (0.5 lm) was measured at 660 nm, in degradation buffer (25 mm Tris ⁄ HCl pH 7.5, 100 mm NaCl, 100 mm KCl, 20 mm MgCl 2 , 0.05% Triton X-100, 10% glycerol) in the absence or presence of unlabelled substrate (5 lm). The reaction was started by the addition of 2 mm ATP and stopped by the addition of 800 lL malachite green solution (0.034% malachite green, 0.1% Triton X-100 and 10.5 gÆ L )1 ammonium molybdate in 1 N HCl) and 100 lL of 34% citrate. Hydrogen–deuterium exchange and mass spectrometry GFP–ssrA (2 lm) was diluted 75· into deuterated buffer (50 mm Tris ⁄ HCl pH 7.5, 300 mm NaCl, 10% glycerol, 0.5 mm dithiothreitol, 10 mm ATP). Where appropriate, ClpA or RR ⁄ AA (2 lm) was added at the start of the exchange reaction (t = 0 min); for the control experiment, equal volumes of non-deuterated buffer were added. At indicated time points, samples were removed from the reac- tion. The hydrogen–deuterium exchange was stopped by rapidly lowering the pH to 2.4 at 4 °C. All subsequent steps were carried out on ice to minimize back exchange. The pro- teins were separated on a micro-C4RP column connected to an ESI-QTOF mass spectrometer (Applied Biosystems, Foster City, CA, USA) using an acetonitrile gradient. Acknowledgements We thank E. Weber-Ban (Eidgeno ¨ ssiche Technische Hochschule Zurich) for providing fluorescently labelled kR–ssrA. This research was supported by the Austra- lian Research Council Discovery Project scheme (DP0450051), the Deutsche Forschungsgemeinschaft priority program ‘Proteolysis in Prokaryotes: Protein Quality Control and Regulatory Principles’ and Aus- tralian Research Council QEII Fellowships to D.A.D and K.N.T. References 1 Neuwald AF, Aravind L, Spouge JL & Koonin EV (1999) AAA+: a class of chaperone-like ATPases asso- ciated with the assembly, operation, and disassembly of protein complexes. Genome Res 9, 27–43. 2 Dougan DA, Mogk A & Bukau B (2002) Protein fold- ing and degradation in bacteria: to degrade or not to degrade? That is the question. Cell Mol Life Sci 59 , 1607–1616. 3 Ogura T & Wilkinson AJ (2001) AAA+ superfamily ATPases: common structure – diverse function. Genes Cells 6, 575–597. 4 Mogk A, Schmidt R & Bukau B (2007) The N-end rule pathway for regulated proteolysis: prokaryotic and eukaryotic strategies. Trends Cell Biol 17, 165–172. Substrate recognition and unfolding by ClpA A. H. Erbse et al. 1408 FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS 5 Varshavsky A (1996) The N-end rule: functions, myster- ies, uses. Proc Natl Acad Sci USA 93, 12142–12149. 6 Meyer HH, Shorter JG, Seemann J, Pappin D & War- ren G (2000) A complex of mammalian ufd1 and npl4 links the AAA-ATPase, p97, to ubiquitin and nuclear transport pathways. EMBO J 19, 2181–2192. 7 Rouiller I, Butel VM, Latterich M, Milligan RA & Wilson-Kubalek EM (2000) A major conformational change in p97 AAA ATPase upon ATP binding. Mol Cell 6, 1485–1490. 8 Dougan DA, Reid BG, Horwich AL & Bukau B (2002) ClpS, a substrate modulator of the ClpAP machine. Mol Cell 9, 673–683. 9 Dougan DA, Mogk A, Zeth K, Turgay K & Bukau B (2002) AAA+ proteins and substrate recognition, it all depends on their partner in crime. FEBS Lett 529,6– 10. 10 Dougan DA, Weber-Ban E & Bukau B (2003) Targeted delivery of an ssrA-tagged substrate by the adaptor pro- tein SspB to its cognate AAA+ protein ClpX. Mol Cell 12, 373–380. 11 Lo JH, Baker TA & Sauer RT (2001) Characterization of the N-terminal repeat domain of Escherichia coli ClpA-A class I Clp ⁄ HSP100 ATPase. Protein Sci 10, 551–559. 12 Singh SK, Rozycki J, Ortega J, Ishikawa T, Lo J, Ste- ven AC & Maurizi MR (2001) Functional domains of the ClpA and ClpX molecular chaperones identified by limited proteolysis and deletion analysis. J Biol Chem 276, 29420–29429. 13 Ishikawa T, Maurizi MR & Steven AC (2004) The N-terminal substrate-binding domain of ClpA unfoldase is highly mobile and extends axially from the distal surface of ClpAP protease. J Struct Biol 146, 180–188. 14 Hinnerwisch J, Reid BG, Fenton WA & Horwich AL (2005) Roles of the N-domains of the ClpA unfoldase in binding substrate proteins and in stable complex for- mation with the ClpP protease. J Biol Chem 280, 40838–40844. 15 Barnett ME, Zolkiewska A & Zolkiewski M (2000) Structure and activity of ClpB from Escherichia coli. Role of the amino- and carboxyl-terminal domains. J Biol Chem 275, 37565–37571. 16 Beinker P, Schlee S, Groemping Y, Seidel R & Rein- stein J (2002) The N terminus of ClpB from Thermus thermophilus is not essential for the chaperone activity. J Biol Chem 277, 47160–47166. 17 Mogk A, Schlieker C, Strub C, Rist W, Weibezahn J & Bukau B (2003) Roles of individual domains and con- served motifs of the AAA+ chaperone ClpB in oligo- merization, ATP hydrolysis, and chaperone activity. J Biol Chem 278, 17615–17624. 18 Hinnerwisch J, Fenton WA, Furtak KJ, Farr GW & Horwich AL (2005) Loops in the central channel of ClpA chaperone mediate protein binding, unfolding, and translocation. Cell 121, 1029–1041. 19 Singh SK & Maurizi MR (1994) Mutational analysis demonstrates different functional roles for the two ATP-binding sites in ClpAP protease from Escherichia coli. J Biol Chem 269, 29537–29545. 20 Erbse A, Schmidt R, Bornemann T, Schneider-Mergen- er J, Mogk A, Zahn R, Dougan DA & Bukau B (2006) ClpS is an essential component of the N-end rule path- way in Escherichia coli. Nature 439, 753–756. 21 Reid BG, Fenton WA, Horwich AL & Weber-Ban EU (2001) ClpA mediates directional translocation of sub- strate proteins into the ClpP protease. Proc Natl Acad Sci USA 98, 3768–3772. 22 Herskovits TT (1966) On the conformation of caseins. Optical rotatory properties. Biochemistry 5, 1018–1026. 23 Rist W, Jorgensen TJ, Roepstorff P, Bukau B & Mayer MP (2003) Mapping temperature-induced conforma- tional changes in the Escherichia coli heat shock tran- scription factor sigma 32 by amide hydrogen exchange. J Biol Chem 278, 51415–51421. 24 Weber-Ban EU, Reid BG, Miranker AD & Horwich AL (1999) Global unfolding of a substrate protein by the Hsp100 chaperone ClpA. Nature 401, 90–93. 25 Zeth K, Ravelli RB, Paal K, Cusack S, Bukau B & Dougan DA (2002) Structural analysis of the adaptor protein ClpS in complex with the N-terminal domain of ClpA. Nat Struct Biol 9, 906–911. 26 Seol JH, Yoo SJ, Kim KI, Kang MS, Ha DB & Chung CH (1994) The 65-kDa protein derived from the inter- nal translational initiation site of the clpA gene inhibits the ATP-dependent protease Ti in Escherichia coli. J Biol Chem 269, 29468–29473. 27 Piszczek G, Rozycki J, Singh SK, Ginsburg A & Maurizi MR (2005) The molecular chaperone, ClpA has a single high affinity peptide binding site per hexamer. J Biol Chem 277, 12221–12230. 28 Guo F, Esser L, Singh SK, Maurizi MR & Xia D (2002) Crystal structure of the heterodimeric complex of the adaptor, ClpS, with the N-domain of the AAA+ chaperone, ClpA. J Biol Chem 277, 46753–46762. 29 Schlieker C, Weibezahn J, Patzelt H, Tessarz P, Strub C, Zeth K, Erbse A, Schneider-Mergener J, Chin JW, Schultz PG et al. (2004) Substrate recognition by the AAA+ chaperone ClpB. Nat Struct Mol Biol 11, 607–615. 30 Watts GD, Wymer J, Kovach MJ, Mehta SG, Mumm S, Darvish D, Pestronk A, Whyte MP & Kimonis VE (2004) Inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia is caused by mutant valosin-containing protein. Nat Genet 36, 377–381. 31 Ormo M, Cubitt AB, Kallio K, Gross LA, Tsien RY & Remington SJ (1996) Crystal structure of the Aequorea A. H. Erbse et al. Substrate recognition and unfolding by ClpA FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS 1409 [...].. .Substrate recognition and unfolding by ClpA A H Erbse et al victoria green fluorescent protein Science 273, 1392– 1395 32 Kenniston JA, Baker TA, Fernandez JM & Sauer RT (2003) Linkage between ATP consumption and mechanical unfolding during the protein processing reactions of an AAA+ degradation machine Cell 114, 511–520 1410 33 Lee C, Schwartz MP, Prakash... Iwakura M & Matouschek A (2001) ATP-dependent proteases degrade their substrates by processively unraveling them from the degradation signal Mol Cell 7, 627–637 34 Navon A & Goldberg AL (2001) Proteins are unfolded on the surface of the ATPase ring before transport into the proteasome Mol Cell 8, 1339–1349 FEBS Journal 275 (2008) 1400–1410 ª 2008 The Authors Journal compilation ª 2008 FEBS . productive binding to the substrate via two elements in ClpA, one in the N-domain and the other in the pore of the ClpA hexamer. In the case of short unstructured. with the N-domain of ClpA) in substrate selection [15–17]. One difficulty in understanding the role of the N-domain of ClpA stems from the variety of activities