Characterization of the N-terminal repeat domain of Escherichia coli ClpA—A class I Clp/HSP100 ATPase

Abstract

The ClpA, ClpB, and ClpC subfamilies of the Clp/HSP100 ATPases contain a conserved N-terminal region of ∼150 residues that consists of two approximate sequence repeats. This sequence from the Escherichia coli ClpA enzyme is shown to encode an independent structural domain (the R domain) that is monomeric and ∼40% α-helical. A ClpA fragment lacking the R domain showed ATP-dependent oligomerization, protein-stimulated ATPase activity, and the ability to complex with the ClpP peptidase and mediate degradation of peptide and protein substrates, including casein and ssrA-tagged proteins. Compared with the activities of the wild-type ClpA, however, those of the ClpA fragment missing the R domain were reduced. These results indicate that the R domain is not required for the basic recognition, unfolding, and translocation functions that allow ClpA-ClpP to degrade some protein substrates, but they suggest that it may play a role in modulating these activities.

The Clp/HSP100 family of proteins is composed of hexameric ATPases that function as protease regulatory subunits, protein unfolding machines, and disassembly chaperones in bacteria, some archaea, yeast, and metazoans, including a wide variety of plants, insects, and mammals (Squires and Squires 1992; Schirmer et al. 1996). Family members range in size from ∼350 to 1000 amino acids and have been divided into two classes, each containing multiple subfamilies (Fig. 1A ▶). Class and subfamily distinctions depend on the number and type of domains. For example, class I family members contain two ATPase domains, whereas class II proteins have a single ATPase domain. ATP binding stabilizes hexamer formation, and ATP hydrolysis is required for the catalysis of protein denaturation (Hoskins et al. 1998; Maurizi et al. 1998). In their capacity as protease regulatory subunits, the Clp ATPases mediate substrate binding as well as unfolding and translocation of the unfolded protein to an associated multimeric protease (Hoskins et al. 1998). A sensor and substrate discrimination (SSD) domain, which is present in all Clp/HSP100 family members, has been implicated in recognition of some protein substrates (Smith et al. 1999). The crystal structure of the class II enzyme HslU (ClpY) (Bochtler et al. 2000) provides a three-dimensional view of an SSD domain, a class II ATPase domain, and an intermediate domain found only in the HslU subfamily. Relatively little is known about other Clp/HSP100 domains, however, and domain designations usually have been based on sequence comparisons rather than experimental studies.

Fig. 1.

(A) Schematic representation of the domain structures of different classes of Clp ATPases. Eukaryotic ClpC enzymes have additional N-terminal sequences that are not shown. (B) Multiple sequence alignment of R-domain sequences from bacterial ClpA, ClpB, and ClpC enzymes.

The class I ClpA, ClpB, and ClpC subfamilies contain N-terminal sequences of ∼150 amino acids, which in turn consist of two repeats of ∼75 residues (Fig. 1 ▶). We refer to this 150-residue sequence as the repeat or R region and, in this paper, characterize the structural and function properties of this region from the Escherichia coli ClpA protein. Several biochemical activities have been demonstrated for E. coli ClpA, including protein-stimulated ATP hydrolysis (Hwang et al. 1988), ATP-dependent disassembly of RepA dimers (Wickner et al. 1994), and ATP-dependent denaturation of GFP-ssrA, that is, green fluorescent protein containing a C-terminal ssrA degradation tag (Weber-Ban et al. 1999). Hexameric rings of ClpA also combine with the double-ring ClpP peptidase to form the ClpA-ClpP protease (Wang et al. 1997; Grimaud et al. 1998). ClpA-ClpP was initially shown to catalyze ATP-dependent degradation of casein (Katayama et al. 1987), which is denatured and serves as an excellent nonspecific protease substrate. ClpA-ClpP also degrades specific substrates possessing stably folded native structures, including RepA and proteins marked for destruction by addition of the ssrA degradation tag (Wickner et al. 1994; Gottesman et al. 1998).

In E. coli, alternate translation start sites within the mRNAs encoding ClpA and ClpB remove the R regions of these proteins, which results in truncated variants designated ClpA65 and ClpB79. The ATPase activity of purified ClpB79, unlike that of wild-type ClpB, was not stimulated by casein or other denatured proteins; thus, it was suggested that the R region might be involved in binding these substrates (Park et al. 1993). Purified ClpA65 (residues 169–758) had markedly reduced ATPase activity and did not support ClpP-dependent degradation of casein (Seol et al. 1994). This result suggested that the R region of ClpA may be required for the protease activity of ClpA-ClpP and may play a role in interactions with ClpP.

Here, we characterize the properties of a protein fragment corresponding to residues 1–161 of E. coli ClpA; we also characterize a truncated ClpA variant (residues 162–758; ClpAΔR) that is missing the R region. We find that the isolated N-terminal fragment is an independent, stably folded domain. The ClpAΔR fragment is also stably folded, activates the peptidase activity of ClpP, and mediates ClpP-dependent degradation of casein and ssrA-tagged protein substrates. Hence, although the R domain is highly conserved, it is not required for ClpP interactions, for recognition of some ClpA substrates, or for the unfolding and translocation functions of ClpA-ClpP that are required for protein degradation. ClpAΔR does have reduced ATPase activity and is less active than wild-type ClpA in mediating ClpP-dependent protein degradation. The implications of these results and the potential roles of the R domain are discussed.

Results

Sequence comparisons

Figure 1B ▶ shows a multiple alignment of the R-region sequences from a set of bacterial and eukaryotic ClpA, ClpB, and ClpC family members. Although the overall number of sequence identities is relatively low and no position is absolutely conserved, the alignments generally have E-values of <10⁻²⁰ in PSIBLAST searches (Altschul et al. 1997) and thus are highly significant. To allow comparison, we have arranged Figure 1B ▶ so that the first R-region sequence repeat is on the top and the second repeat is directly below. Each repeat has ∼70–75 residues and shows a reasonably conserved pattern of positions in which hydrophobic side chains are present. Again, PSIBLAST searches confirm that the sequence relationships between the two repeats are statistically significant, with many alignments between the first and second repeats having E-values of 10⁻¹⁰ or lower.

Biophysical characterization of the R domain and ClpAΔR

An R-region fragment with an N-terminal His6 tag followed by residues 1–161 of E. coli ClpA was cloned, overexpressed, and purified. This fragment includes both sequence repeats of the R region. The C-terminal junction was chosen to fall within a sequence rich in Pro, Gly, Ser, and Glu that we suspected would be unstructured or form a loop in intact ClpA. The far-UV circular dichroism (CD) spectrum of the R-region fragment had minima at 208 and 222 nm, which is characteristic of an α-helical structure (Fig. 2 ▶). The mean residue ellipticity at 222 nm was that expected for a protein with ∼40% α-helical content. In GuHCl unfolding experiments, the R-region fragment showed the cooperative melting expected for a stably folded native protein (Fig. 3A ▶), and denaturation was fully reversible. These experiments show that residues 1–161 of ClpA encode an independent structural domain.

Fig. 2.

CD spectra of the R domain, ClpAΔR, and ClpA (0.5 μM in monomer equivalents in 25 mM KPi, at pH 7.5, 50 mM KCl, 10% glycerol) at 25°C. The solid line represents the sum of the R domain and ClpAΔR spectra.

Fig. 3.

Denaturation of the R domain (A), ClpAΔR (B), and ClpA (C) by GuHCl. Experiments were performed at 25°C by mixing protein in 25 mM KPi (pH 7.5), 50 mM KCl, and 10% glycerol with protein at the same concentration in buffer containing 7 M GuHCl. Ellipticity at 230 nm was allowed to equilibrate and was averaged over a period of 2 min.

Because the R-domain sequence consists of two approximate sequence repeats, it seemed possible that the sequence might contain two independent folding units. Several observations, however, indicated that this was not the case. First, both GuHCl denaturation experiments (Fig. 3A ▶) and thermal unfolding studies (data not shown) showed a single transition with no sign of biphasic unfolding. Second, we were unable to isolate stable subfragments by proteolytic digestion of the purified 1–161 R-domain fragment by using trypsin, chymotrypsin, proteinase K, or V8 protease. Third, we subcloned ClpA fragments containing repeat 1 (residues 1–75) and repeat 2 (residues 76–161) under the control of a strong T7 promoter, but these fragments were not stably expressed in E. coli, presumably because they were degraded rapidly by cellular proteases.

To address the role of the R domain in ClpA structure and function, we cloned, overexpressed, and purified a fragment (designated ClpAΔR) containing an N-terminal His6 tag followed by residues 162–758 of ClpA. ClpAΔR had a far-UV CD spectrum (Fig. 2 ▶) and GuHCl melting curve (Fig. 3B ▶) indicative of a stably folded structure. Hence, the R domain is not required for stable folding of the remaining portions of ClpA. As shown in Figure 2 ▶, addition of the individual CD spectra of ClpAΔR and the R-domain fragment gives a spectrum that is indistinguishable from that of intact ClpA. This indicates that separation of the R domain from the rest of ClpA does not significantly change the secondary structure of either fragment. Moreover, the GuHCl denaturation profile of intact ClpA shows two transitions. The first broad transition, with a midpoint at ∼1.5 M GuHCl, corresponds to denaturation of ClpAΔR. The second transition, with a midpoint at ∼3 M GuHCl, corresponds to denaturation of the R domain. Thus, the R domain remains folded while the remaining portions of ClpA denature and does not appear to appreciably stabilize these other regions.

The purified R-domain fragment behaved as a monomer in analytical equilibrium ultracentrifugation experiments performed at protein concentrations of 10, 50, and 250 μM (Fig. 4 ▶). In gel filtration experiments, ClpAΔR appeared to chromatograph as a mixture of monomers and dimers in the absence of nucleotide and near the position expected for a hexamer in the presence of ATPγS (data not shown). In accord with previous studies (Maurizi et al. 1998), we found that ClpA seemed to be largely dimeric without ATPγS and hexameric with nucleotide. These results indicate that the R domain neither forms oligomers by itself nor is required for ClpA oligomerization.

Fig. 4.

Equilibrium analytical ultracentrifugation of the R domain (250 μM in monomer equivalents in 50 mM KPi, at pH 7.5, 300 mM KCl, 0.1 mM EDTA) at 20°C and 25,000 rpm. The lines represent the expected absorbance distribution for a monomer (solid line; Mr 20246) and a dimer (dashed line; Mr 40492).

ClpAΔR exhibits reduced but protein-stimulated ATPase activity

ClpAΔR was approximately half as active as ClpA in a coupled ATP-hydrolysis assay (Fig. 5 ▶). Clearly, ClpAΔR retains the ability to hydrolyze ATP but is a somewhat poorer ATPase than is intact ClpA. ATP hydrolysis was stimulated ∼1.5- to twofold by λ N-cI-ssrA protein for both ClpAΔR and ClpA (Fig. 5 ▶). As a result, we conclude that the R domain is required for neither ATPase activity nor for substrate stimulation of this activity, although it may play some role in modulating the rate of ATPase hydrolysis.

Fig. 5.

ATPase activities of ClpA and ClpAΔR. Reactions (50 μL) contained 0.1 μM ClpA or ClpAΔR (monomer equivalents) with or without 1 μM λ N-cI-ssrA in 50 mM HEPES (pH 7.5), 300 mM KCl, 0.1% NP-40, 10% glycerol and were performed at 37°C. ATP hydrolysis in a coupled reaction was assayed by the disappearance of NADH as monitored by a reduction in absorbance at 340 nm.

ClpAΔR activates degradation of peptides and proteins by ClpP

The binding of ClpA to ClpP stimulates the intrinsic peptidase activity of ClpP and is also required for ClpA-ClpP–dependent degradation of protein substrates (Thompson et al. 1994). Figure 6 ▶ shows hydrolysis of succinyl-Leu-Tyr 7-amido-4-methylcoumrin (AMC) by ClpP alone, by ClpA-ClpP, and by ClpAΔR-ClpP. Both ClpA and ClpAΔR stimulated the peptidase activity of ClpP by a factor of ∼1.6-fold. This result indicates that ClpAΔR, like ClpA, is capable of forming an active complex with ClpP.

Fig. 6.

Effects of ClpAΔR and ClpA on ClpP-mediated hydrolysis of succinyl-Leu-Tyr-AMC monitored by absorbance at 414 nm. Reactions (100 μL) contained 3.7 pmoles ClpP₁₄ 1 mM ATP, 0.125 mM succinyl-Leu-Tyr-AMC, and an ATP regeneration system. ClpAΔR or ClpA (7.4 pmoles in hexamer equivalents) was added to the appropriate reactions.

ClpAΔR, like intact ClpA, promoted degradation of casein and of protein substrates containing the ssrA degradation tag (Fig. 7 ▶). ClpAΔR-ClpP degraded casein and λ N-cI-ssrA at ∼20% to 30% of the rate observed with ClpA-ClpP (Fig. 7A,B ▶). ClpAΔR-ClpP also catalyzed degradation of GFP-ssrA but only at ∼2.5% the rate of ClpA-ClpP, based on estimates of initial velocity (Fig. 7C ▶). In all cases, degradation was dependent on ClpP and ATP (data not shown). As was also seen with ClpA-ClpP (Gottesman et al. 1998), ClpAΔR-ClpP degradation of λ N-cI-ssrA was reduced 5- to 10-fold when the C terminus of the ssrA degradation tag was changed from Ala-Ala to Asp-Asp (data not shown). Hence, ClpAΔR-ClpP engaged in proteolytic activity against casein and substrates, which are recognized through their C-terminal residues. In the degradation assays shown in Figure 7 ▶, ClpA or ClpAΔR, ClpP, and ATP were preincubated for 15 min, and reactions were initiated by adding substrate. For ClpAΔR, order-of-addition experiments revealed a kinetic lag of ∼10–20 min, when the substrate was included in the preincubation mix and the reactions were instead started by the addition of one of the other reaction components (Fig. 8 ▶). No lag was seen in similar order-of-addition experiments performed with wild-type ClpA, indicating that deletion of the R domain affects the rate of assembly of active ClpA-ClpP complexes.

Fig. 7.

Degradation of different substrates by ClpAΔR-ClpP and ClpA-ClpP. (A) Casein degradation. (B) Degradation of ³⁵S-labeled λ N-cI-ssrA at 37°C monitored by release of trichloroacetic acid–soluble radioactivity. Reactions (20 μL) contained 10 pmoles λ N-cI-ssrA, 0.2 pmoles ClpA₆ or ClpAΔR6, and 0.1 pmole ClpP₁₄ in 25 mM HEPES (pH 7.5), 25 mM potassium acetate, 5 mM MgCl₂, 10% glycerol, 0.02% Nonidet P-40. (C) Degradation of GFP-ssrA at 37°C monitored by loss of fluorescence at 411 nm. Reactions (100 μL) contained 9.4 pmoles GFP-ssrA, 0.5 pmoles ClpP₁₄, and 1.0 pmole ClpA₆ or ClpAΔR₆ in 50 mM HEPES (pH 7.5), 300 mM NaCl, 20 mM MgCl₂, 10% glycerol. All reactions contained 2 mM ATP and an ATP regeneration system containing 0.1 mg/mL creatine kinase and 5 mM creatine phosphate.

Fig. 8.

Order-of-addition experiments. Proteolysis reactions lacking one component were pre-incubated at 37°C for 15 min. The missing component was then added, and aliquots were removed at the times indicated. Each curve is labeled with the component added last. Reactions (20 μL) contained 10 pmoles λ N-cI-ssrA, 0.2 pmoles ClpAΔR₆, 2 mM ATP with regeneration system, and 0.1 pmoles ClpP₁₄ in 25 mM HEPES (pH 7.5), 25 mM potassium acetate, 5 mM MgCl₂, 10% glycerol, 0.02% Nonidet P-40.

Discussion

We have established that a fragment consisting of the N-terminal 161 residues of the E. coli ClpA protein corresponds to an independent structural domain that we refer to as the R domain. The R domain, which is highly conserved in several Clp/HSP100 subfamilies, is monomeric and partially α-helical. Moreover, although this domain contains two approximate sequence repeats, it behaves as a single cooperatively folded unit. CD and denaturation studies indicate that the R domain of ClpA does not interact strongly with the remaining portions of the protein. Indeed, a ClpA fragment, lacking the R domain, is also stably folded and displays approximately the same denaturation behavior as the corresponding region in intact ClpA. The ClpAΔR fragment forms oligomers in an ATP-dependent fashion, displays protein-stimulated ATPase activity, and stimulates the peptidase activity of ClpP. ClpAΔR is also able to combine with ClpP to catalyze ATP-dependent degradation of several known ClpA substrates, including casein, λ N-cI-ssrA, and GFP-ssrA. These results eliminate a number of potential roles for the R domain. Specifically, the R domain cannot be required for the overall folding or oligomerization of ClpA, nor is it essential for ATPase activity or needed for ClpP interactions, nor is it necessary for the recognition, unfolding, or translocation activities that allow ClpA-ClpP to degrade some protein substrates. ClpA65 (residues169–758) is a natural ClpA variant, synthesized from an alternate translation start, which lacks the R domain (Seol et al. 1994). In contrast to the results reported here for ClpAΔR (residues 162–758), ClpA65 was shown to have a much larger reduction in ATPase activity (20-fold versus 2-fold), and ClpA65-ClpP was found not to degrade casein (Seol et al. 1994). Whether these discrepancies reflect the minor sequence variations between ClpA65 and ClpAΔR or procedural differences in protein purification or assay design is not known. However, N-ethylmaleimide (NEM) modification of Cys47 in the R domain has also been shown to block ClpP-mediated proteolysis (Seol et al. 1997). Therefore, although the R domain is not required for protein degradation, it appears to be positioned in a manner that allows NEM modification to prevent the normal degradation cycle. Thus, it is attractive to consider the possibility that binding of accessory factors to the R domain might also negatively regulate the proteolytic activity of ClpA. Deletion of the R domain of ClpA slows assembly of complexes with ClpP and reduces the ATPase and protein degradation activities. Interestingly, the proteolytic defect of ClpAΔR-ClpP relative to ClpA-ClpP was significantly greater for the GFP-ssrA substrate (40-fold) than it was for the λ N-cI-ssrA or casein substrates (three- to fivefold). Indeed, degradation of GFP-ssrA by ClpAΔR-ClpP was probably detected only because the GFP fluorescence assay was sufficiently sensitive to detect low levels of degradation. Native GFP is more thermodynamically stable than is the native N-terminal domain of λ repressor, and casein is denatured. Because ClpAΔR-ClpP degrades the latter proteins at similar rates, it is clear that there is no simple correlation between protein stability and degradation rates. Nevertheless, the R domain may serve, in some fashion, to allow ClpA to degrade or remodel proteins like GFP that are hyperstable.

We assume that the R domain of ClpA plays an important biological role because of its strong evolutionary conservation and because the E. coli clpA gene expresses a natural isoform that is missing in this domain (Seol et al. 1994). Hence, it seems unlikely that this domain simply acts to increase the activity of the remaining parts of ClpA modestly. What else might the R domain do? One possibility is that it functions in recognition of classes of ClpA substrates distinct from casein and ssrA-tagged substrates. In preliminary experiments, for example, we did not detect ClpAΔR-ClpP degradation of RepA, a ClpA substrate recognized via peptide sequences located near the N terminus (Hoskins et al. 2000). The R domain might also act as a docking site for accessory factors in the cell or function to mediate proper subcellular localization. Additional experiments are required to test these models. We anticipate that the R-domain fragment and the ClpAΔR fragment will provide valuable tools for these studies.

Materials and methods

Construction, expression, and protein purification

The clpA gene used for these studies contained the Met169→Thr mutation, which prevents the translational start that gives rise to the ClpA65 fragment (Seol et al. 1994); the gene was provided by John Flanagan in the overexpression vector pET9a(ClpAT). This mutant is indistinguishable from wild-type ClpA in its ATPase activity and ability to support ClpP-mediated protein degradation (Seol et al. 1995). Genes encoding the R-domain fragment (ClpA residues 1–161) and the ClpAΔR fragment (residues 162–758) were amplified by PCR, using the clpA (Met169→Thr) gene as template and with primers encoding NdeI and BamHI sites. The amplified DNA was cleaved with both restriction enzymes and subcloned between the NdeI and BamHI sites of pET15b, resulting in addition of a His6 tag to the N terminus of each protein fragment, to generate plasmids pJL101 (encoding the R-domain fragment) and pJL102 (encoding the ClpAΔR fragment). ClpA (Met169→Thr) protein was expressed in E. coli strain BL21/DE3/plysS. Cells were freshly transformed with plasmid pET9a(ClpAT) and grown at 37°C in Luria-Bertani broth to an OD₆₀₀ of 0.6. Isopropyl β-D-thiogalactoside was then added to a final concentration of 1 mM; after growth for an additional 3 h, cells were harvested by centrifugation. They were then resuspended in 100 mM Tris (pH 7.8) and 10% sucrose (4.3 mL/g cells) and lysed by sonication; debris was removed by centrifugation at 30,000g. Ammonium sulfate was added to the supernatant to 40% saturation. After 20 min at 4°C, the pellet fraction was collected by centrifugation and resuspended (6.25 mL/g cells) in buffer A (50 mM Tris-Cl, at pH 7.5, 100 mM KCl, 1 mM DTT, 10% glycerol). This material was dialyzed against two 6-L changes of the same buffer and loaded onto a 0.5 × 5-cm Mono S column (Pharmacia) equilibrated in buffer A. The column was developed using a linear gradient from 100 mM to 1 M KCl (20 mL total) in buffer A. Column fractions were assayed by SDS–PAGE, and those containing purified ClpA were pooled, flash frozen in liquid nitrogen, and stored at −80°C. The R-domain and ClpAΔR fragments were purified from E. coli strain BB101 transformed with pJL101 or pJL102, respectively. Cells were grown, induced, harvested, and lysed as described above. For the R-domain purification, the cleared lysate was mixed with an equal volume of buffer B (50 mM HEPES, at pH 7.5, 12 mM imidazole, 250 mM NaCl, 10% glycerol) and loaded onto a 3-mL Ni²⁺-NTA column (Qiagen) equilibrated in buffer B. The column was washed with 100 mL of buffer B, and protein was eluted with 50 mM HEPES (pH 7.5), 250 mM NaCl, 250 mM imidazole, 10% glycerol. For the ClpAΔR purification, the cleared lysate was precipitated with 40% ammonium sulfate, resuspended in buffer B, and purified as described on a 3-mL Ni²⁺-NTA column. Fractions containing ClpAΔR protein were pooled, dialyzed twice against buffer C (50 mM Tris, at pH 8.5, 100 mM KCl, 1 mM DTT, 10% glycerol), and loaded onto a 0.5 × 5-cm Mono Q column (Pharmacia) equilibrated in buffer C. The Mono Q column was developed with a linear gradient from 100 mM to 1 M KCl in buffer C. For both the R-domain and ClpAΔR, fractions containing protein of >95% purity were pooled, flash frozen, and stored at −80°C. Purified E. coli ClpP protein and GFP-ssrA (Andersen et al. 1998) were gifts of Yong-In Kim. His₆-tagged λ N-cI-ssrA protein was 35S-labeled and affinity purified on a Ni²⁺-NTA column as described (Gottesman et al. 1998). Fluorescein-conjugated casein was purchased from Molecular Probes. RepA protein was a gift of Sue Wickner.

Protein characterization

CD spectra (25°C; 25 mM KPi at pH 7.5, 50 mM KCl, 10% glycerol) were taken using an AVIV model 60 spectrapolarimeter in a 1-cm pathlength cuvette at 1-nm intervals from 200 to 250 nm with an averaging time of 30 sec at each wavelength. Denaturation experiments (25°C) were performed by mixing protein dissolved in 25 mM KPi (pH 7.5), 50 mM KCl, 10% glycerol with protein at the same concentration dissolved in 7 M GuHCl. After equilibration to a constant CD signal, the ellipticity at 230 nm was averaged for a period of 2 min. Denaturation was reversible as judged by the recovery of native ellipticity after dilution from 7 M GuHCl. Analytical ultracentrifugation experiments were performed using a Beckman Optima XL-A centrifuge. The R domain (10, 25, or 250 μM in 50 mM Kpi, at pH 7.5, 300 mM KCl, 0.1 mM EDTA) was centrifuged at 15,000 or 25,000 rpm using a Beckman 60 Ti rotor. Absorbance readings, measured at 214 nm, 220 nm, or 280 nm, were performed at 3-h intervals to determine when samples reached equilibrium (generally over 21 h). Ten scans were averaged and then analyzed according to Laue et al. (1992) to determine apparent molecular weights. In all cases, the weight average molecular weight was within 10% of that expected for a monomer.

Samples of ClpA or ClpAΔR (50 μL of 10 or 40 μM in monomer equivalents) were chromatographed on a 2.4 mL Superose 6 gel-filtration column in 50 mM Tris [pH 7.5 or 8.4], 300 mM KCl, 25 mM MgCl₂, 10% glycerol. Apparent molecular weights were determined by comparisons of elution volumes with those of Bio-Rad molecular weight standards.

Activity assays

A stock of succinyl-Leu-Tyr-AMC (100 mM; Sigma) in dimethylsulfoxide was prepared as a substrate for ClpP peptidase assays. Reactions (100 μL) were performed at 37°C in 50 mM HEPES (pH 7.5), 300 mM KCl, 0.1% NP-40, 10% glycerol and contained 3.7 pmoles ClpP₁₄, 7.4 pmoles ClpA₆ (Met169→Thr) or ClpAΔR₆, 1 mM ATP, 0.125 mM succinyl-Leu-Tyr-AMC, and an ATP regeneration system consisting of 0.1 mg/mL creatine kinase and 5 mM creatine phosphate. AMC release was monitored by fluorescence emission intensity at 414 nm (excitation at 357 nm).

GFP-ssrA degradation assays (37°C) were monitored by loss of fluorescence at 511 nm (excitation at 467 nm) and contained 50 mM HEPES (pH 7.5), 300 mM NaCl, 20 mM MgCl₂, 10% glycerol, 9.4 pmoles GFP-ssrA, 0.5 pM ClpP₁₄, and 1.0 pM ClpA₆ or ClpAΔR₆ in a total volume of 100 μL. Degradation of ³⁵S-labeled λ N-cI-ssrA by ClpA-ClpP or ClpAΔR-ClpP (37°C) was assayed by the release of ³⁵S counts soluble in 10% cold trichloroacetic acid, using the same buffer, enzyme, and substrate concentrations (Gottesman et al. 1998). Fluorescein-conjugated casein was dissolved in 50 mM Tris (pH 7.5), 300 mM KCl, 10% glycerol, and its degradation was monitored by increases in fluorescence at 419 nm (excitation at 394 nm). The substrate and enzyme concentrations for these assays were the same as those used in the GFP-ssrA assays. All proteolysis reactions contained 2 mM ATP, 0.1 mg/mL creatine kinase, and 5 mM creatine phosphate as an ATP regeneration system.

ATPase activity was assayed via the loss of reduced nicotinamide adenine dinucleotide (NADH) (monitored by changes in absorbance at 340 nm) in a reaction in which NADH oxidation was coupled to ATP hydrolysis (Karon et al. 1995). Reactions (50 μL) were performed at 37°C and contained 2.5 mM ATP, 1 mM NADH, 18.75 U/mL pyruvate kinase, 21.45 U/mL lactate dehydrogenase, 7.5 mM phosphoenolpyruvate in 50 mM HEPES (pH 7.5), 300 mM KCl, 0.1% NP-40, 10% glycerol, and 0.1 μM ClpA or ClpAΔR monomer. Identical reactions with the addition of 1 μM λ N-cI-ssrA were used to measure substrate stimulation of ATPase activity.