The YjbH Protein of Bacillus subtilis Enhances ClpXP-Catalyzed Proteolysis of Spx

Abstract

The global transcriptional regulator Spx of Bacillus subtilis is controlled at several levels of the gene expression process. It is maintained at low concentrations during unperturbed growth by the ATP-dependent protease ClpXP. Under disulfide stress, Spx concentration increases due in part to a reduction in ClpXP-catalyzed proteolysis. Recent studies of Larsson and coworkers (Mol. Microbiol. 66:669-684, 2007) implicated the product of the yjbH gene as being necessary for the proteolytic control of Spx. In the present study, yeast two-hybrid analysis and protein-protein cross-linking showed that Spx interacts with YjbH. YjbH protein was shown to enhance the proteolysis of Spx in reaction mixtures containing ClpXP protease but not ClpCP protease. An N-terminal truncated form of YjbH with a deletion of residues 1 to 24 (YjbH^Δ1-24) showed no proteolysis enhancement activity. YjbH is specific for Spx as it did not accelerate proteolysis of the ClpXP substrate green fluorescent protein (GFP)-SsrA, a GFP derivative with a C-terminal SsrA tag that is recognized by ClpXP. Using inductively coupled plasma atomic emission spectroscopy and 4-(2-pyridylazo) resorcinol release experiments, YjbH was found to contain zinc atoms. Zinc analysis of YjbH^Δ1-24 revealed that the N-terminal histidine-rich region is indispensable for the coordination of at least one Zn atom. A Zn atom coordinated by the N-terminal region was rapidly released from the protein upon treatment with a strong oxidant. In conclusion, YjbH is proposed to be an adaptor for ClpXP-catalyzed Spx degradation, and a model of YjbH redox control involving Zn dissociation is presented.

In the spore-forming bacterium Bacillus subtilis, the protein Spx is a global transcriptional regulator that exerts both positive and negative control of multiple genes during thiol-specific oxidative stress (34, 57). Cells undergoing disulfide stress exhibit an elevated Spx concentration and heightened Spx activity. This leads to induction of the Spx regulon, which includes trxA (encoding thioredoxin), trxB (thioredoxin reductase), and other genes that function in the oxidative stress response and in cysteine biosynthesis (10, 34). Spx-dependent activation of trxA and trxB requires interaction of the protein with RNA polymerase (RNAP) holoenzyme (33, 37). This interaction also results in repression of operons that require an activator for transcription initiation (35, 54).

The spx gene was first identified as a suppressor locus of clpP and clpX mutations (32). ClpX proteins form hexameric, ring-shaped complexes and belong to the AAA+ (for ATPases associated with a variety of cellular activities) Clp/Hsp100 family of proteins (2). ClpX is the ATPase and substrate-binding component of the ClpXP protease. ClpX also functions as an unfoldase and translocase that can denature and deliver substrate proteins to the proteolytic chamber, which is composed of two heptameric rings of the ClpP subunits (45). In wild-type cells under normal growth conditions, Spx is present at nearly undetectable levels because of ClpXP-catalyzed proteolysis (35, 36). The clpP and clpX mutants are severely defective in genetic competence, sporulation, and growth in minimal medium. Nearly all of these properties of the clpX and clpP mutants are overcome by a null mutation in spx (32). Thus, the proteolytic turnover of Spx by ClpXP is an important function in growing B. subtilis cells.

ClpX and ClpP orthologs are found in most bacteria, mitochondria, and chloroplasts. ClpXP plays an important role in determining the quality and quantity of the proteome in cellular processes associated with developmental programs and during times of stress. Through proteomic studies of Escherichia coli ClpXP, five distinct degradation signals were identified, including three sequence motifs at the N terminus of natural substrates and two sequence motifs found at the C terminus (14, 15, 27). In B. subtilis the C-terminal residues of Spx are required for its degradation by ClpXP (34). This sequence shows similarity to the SsrA-tagged AANDENYALAA peptide, which is a known recognition sequence for ClpXP in Escherichia coli and B. subtilis (16, 50). Aside from direct sequence recognition by ClpX, an adaptor protein is sometimes required for tethering a substrate to the protease via contact with ClpX. For example, the response regulator RssB in E. coli recognizes the stationary-phase sigma factor σ^S and delivers it to ClpXP for degradation (4). In exponentially growing cells, RssB is kept in an active form to quickly facilitate the turnover of σ^S (56). ClpXP also utilizes the adaptor SspB, a ribosome-associated protein that can specifically recognize SsrA-tagged proteins for degradation (28). In B. subtilis the ATP-dependent protease ClpCP utilizes the adaptors MecA, YpbH, and McsB (22, 36, 39, 47). MecA targets ComK, the competence transcriptional regulator, in exponential-phase cells to facilitate its rapid turnover by ClpCP. In response to changes in the nutritional environment and high cell density, MecA is inhibited by the small anti-adaptor peptide ComS, resulting in release of ComK from the proteolytic complex (47, 48). This results in heightened ComK expression and activation of the late competence genes (6, 49). Thus, Clp protease can recognize target proteins through a broad range of target signals and with the aid of substrate-specific adaptors (45).

Besides SsrA-tagged proteins, only a few protein targets of ClpXP have been identified in B. subtilis. Recently, the 52-residue-long Sda peptide, which blocks sporulation in response to defects in replication initiation in B. subtilis (7), was found to be a substrate for ClpXP (7, 43). ClpXP also functions in activation of the SigW regulon in B. subtilis by degrading the anti-sigma protein RsiW (53) as part of the cell envelope stress response. Thus far, there have been no adaptor proteins found that function in ClpXP substrate recognition in B. subtilis.

In a recent study, Larsson and coworkers discovered that a null mutation in the yjbH gene causes a dramatic increase in Spx concentration while having no effect on spx transcription (25). They have proposed that the product of yjbH might control the ClpXP-catalyzed proteolysis of Spx by altering the conformation of the substrate and thereby uncovering a recognition element that is targeted by the protease. The yjbH gene is linked to spx and resides in an operon with yjbI that encodes a truncated hemoglobin (YjbI) with peroxidase activity (11).

In this report, we show that Spx proteolysis in vitro by ClpXP is enhanced by YjbH protein while another ATP-dependent and adaptor-dependent protease, ClpCP, shows little effect from YjbH. Data are presented indicating that YjbH binds zinc and that metal binding requires the YjbH N-terminal region. YjbH is proposed to affect the ClpXP-dependent proteolytic control of Spx by serving as an adaptor for ClpXP.

MATERIALS AND METHODS

Bacterial strains, plasmids, and chemicals.

B. subtilis JH642 was used as the parental strain for all genetic manipulations. E. coli Top10 (Invitrogen) was used for general cloning procedures. Medium components were from Difco. All restriction/modifying enzymes were from New England Biolabs. Oligonucleotide primers were from Invitrogen; plasmid pET-23a and E. coli BL21(DE3)pLysS were from Novagen. The Ni²⁺-nitrilotriacetic acid (NTA) resin and PCR/plasmid purification kits were from Qiagen (Germany). High-Q anion-exchange prepack columns were procured from Bio-Rad. All analytical-grade chemicals were from Sigma unless otherwise stated.

Construction of insertion mutant of yjbH.

According to previous work (42), the yjbH gene contains an alternative start codon (TTG) preceded by a putative ribosome-binding site (GGAGG) and encodes an additional 24 amino acids compared with the predicted yjbH gene coding sequence (23) (Subtilist; http://genolist.pasteur.fr/SubtiList/). Primers oZY07-42 (5′-C AGC GTC GAC ATG TTT GTA GAC CC-3′) and oZY07-43 (5′-CAT AAG TCG ACA GCC TCA AGC ATA TGC CC-3′) were used to amplify the yjbH gene from B. subtilis strain JH642 chromosomal DNA. The PCR fragment (from +62 to +936) was digested with SalI and then ligated with pUC19 that had been digested with the same enzyme to generate pZY36. Plasmid pZY36 was cleaved with BglII before treatment with T4 DNA polymerase to create blunt ends and then further ligated with a tetracycline resistance cassette from pDG1515 (17) (cleaved with BamHI and EcoRI and then treated with T4 DNA polymerase to create blunt ends) to generate pZY38 [yjbH::tet (forward orientation)] and pZY30 [yjbH::tet (reverse orientation)]. The plasmids were introduced by transformation into B. subtilis strain JH642 with selection for tetracycline resistance to obtain integrants bearing the tetracycline resistance cassette at position +486 of yjbH to generate ORB6952 [trpC2 pheA1 yjbH::tet (forward)] and ORB6953 [trpC2 pheA1 yjbH::tet (reverse)].

Yeast two-hybrid procedure.

Plasmid pSN11 (35) was constructed by subcloning spx from pMMN470 into pGBKT7 (Clontech). The yjbH gene (900 bp) was amplified by PCR using primers 5′-ATA TGG ATC CTA ACA AAC TAT CAG CAT GAG C-3′ (forward) and 5′-ATA TCT CGA GCT ATT TTT CAC ATG ATT GAT ATT C-3′ (reverse) (restriction sites are underlined) from genomic DNA of B. subtilis JH642. After restriction digestion with BamHI and XhoI, the PCR fragment was cloned into vector pGADT7 (Clontech) to generate pYJBHAD. Plasmid pTYJBHAD was generated by amplifying an N-terminally truncated yjbH gene (828 bp) with primers 5′-ATA TGG ATC CTA ATG TTT GTA GAC CCT TTA TGT CC-3′ (forward) and 5′-ATA TCT CGA GCT ATT TTT CAC ATG ATT GAT ATT C-3′ (reverse), followed by insertion into the BamHI-XhoI site in plasmid pGADT7. Plasmid pSN11, a pGBKT7 derivative carrying an spx-GAL4 DNA-binding domain (BD) fusion, and pYJBHAD or pTYJBHAD, carrying, respectively, a full-length and truncated yjbH-GAL4 activation domain (AD), were used to transform Saccharomyces cerevisiae PJ69-4A (His⁻ Ade⁻ Trp⁻ Leu⁻) (21), which carries a GAL2-ADE2 and GAL2-HIS3 fusion controlled by GAL4. The yeast two-hybrid reporter strain (PJ69-4A) expressing different BD and AD fusion proteins was assayed for growth on an adenine dropout plate or a histidine dropout plate. The interactions between Spx-αCTD (RpoA α C-terminal domain) and Spx-αCTD (rpoA^cxs-1, Y263C) were used as positive and negative controls, respectively.

Expression and purification of C-terminally His₆-tagged YjbH and YjbH^Δ1-24.

The full-length yjbH gene (900 bp) was PCR amplified from genomic DNA of B. subtilis JH642 with primers 5′-ATA TGC TAG CAT GAC AAA CTA TCA GCA TGA GC-3′ (forward) and 5′-ATA TCT CGA GTT TTT CAC ATG ATT GAT ATT CAT C-3′ (reverse) (restriction sites are underlined). The PCR product was cloned into vector pET-23a at the NheI and XhoI sites to produce plasmid pYPET. To generate pTYPET carrying the coding sequence of the yjbH gene encoding a deletion of residues 1 to 24 (yjbH^Δ1-24), a fragment (828 bp) was amplified using a different forward primer, 5′-ATA TGC TAG CAT GTT TGT AGA CCC TTT ATG TCC-3′, and cloned into the NheI-XhoI-digested pET-23a vector. The pYPET- and pTYPET-transformed E. coli BL21(DE3) pLysS cells were grown at 37°C in LB broth containing 50 μg ml⁻¹ ampicillin and 5 μg ml⁻¹ chloramphenicol to an A₆₀₀ of 0.5 to 0.6, and expression was induced with 0.3 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 3 h at 30°C. Cells from a 1-liter culture were resuspended in 10 ml of buffer A (50 mM NaH₂PO₄, 300 mM NaCl, pH 8.0). One tablet of EDTA-free protease inhibitor cocktail (Roche Applied Science, Germany) was added per 10 ml of resuspended cells. The cells were subjected to freeze-thaw cycles and disrupted by a French press. The inclusion bodies were separated by centrifugation at 17,000 × g for 30 min at 4°C. YjbH and YjbH^Δ1-24 were purified from inclusion bodies by dissolving in buffer B (100 mM NaH₂PO₄, 10 mM Tris-Cl, and 6 N guanidine HCl, pH 8.0) for 16 h at 25°C. The clear supernatant was obtained by centrifugation at 17,000 × g for 30 min. The supernatant was applied to a preequilibrated Ni²⁺-NTA column. The YjbH and YjbH^Δ1-24 proteins were purified under denaturing conditions as described by the manufacturer (Qiagen, Germany). For the in-column refolding, protein bound to the Ni²⁺-NTA column was washed with buffer B by gradually decreasing the guanidine HCl concentration from 6 N to 1 N with an interval of 1 N. At the end of the gradient, the column was washed with 30 column volumes of buffer C (50 mM NaH₂PO₄, 100 mM KCl, 30 mM imidazole, and 5% glycerol). The protein was eluted in buffer D (50 mM NaH₂PO₄, 100 mM KCl, 200 mM imidazole, and 5% glycerol). The fractions containing YjbH and YjbH^Δ1-24 were examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. To further improve the purity, proteins were applied to a High-Q (anion-exchange prepack columns; Bio-Rad) cartridge. The YjbH and YjbH^Δ1-24 proteins were eluted with 100 mM to 500 mM KCl gradients and analyzed by 12% SDS-PAGE. The fractions containing pure YjbH and YjbH^Δ1-24 were concentrated by an Amicon filtration device and dialyzed against buffer E (50 mM HEPES-KOH, pH 7.6, 100 mM KCl, 5% glycerol and 5 mM dithiothreitol [DTT]). The protein concentration was measured using protein concentration dye reagent (Bio-Rad), and bovine serum albumin was used to make the standard curve.

Preparation of Zn-saturated YjbH and YjbH^Δ1-24.

Zinc-saturated YjbH or YjbH^Δ1-24 was prepared by incubating freshly purified proteins (100 μM) with 0.1 mM zinc acetate or ZnCl₂ in the presence of 1 mM DTT in a 1-ml reaction mixture at 30°C for 30 min. Unbound zinc was removed from the proteins by dialyzing against buffer E.

Glutaraldehyde cross-linking of YjbH and Spx.

Cross-linking reactions were carried out in 50 mM HEPES-KOH, pH 7.6, 100 mM KCl, 5% glycerol, and 5 mM DTT. Spx (10 μM) was mixed with different concentrations of YjbH (0, 2.5, 5, and 10 μM) in a 10-μl reaction volume. After incubation at 25°C for 15 min, 150 μM glutaraldehyde was added to the reaction mixture, and incubation was continued for 90 min at 25°C. The cross-linking reaction was stopped by addition of 3 μl of SDS-PAGE loading buffer. The cross-linking products were resolved by 12% SDS-PAGE. The gel-resolved protein was transferred to a nitrocellulose membrane in a Bio-Rad transfer apparatus by standard procedures, and immunoblot analysis was carried out with anti-Spx antibodies (36) to visualize the Spx-YjbH cross-linked complex.

In vitro proteolysis reactions.

In vitro proteolysis reaction mixtures were assembled under conditions described previously (55), with some modifications. The reactions were carried out in 50 mM HEPES-KOH (pH 7.6), 50 mM KCl, 10 mM Mg acetate, 5 mM DTT, 5 mM ATP, 5 mM creatine phosphate, 0.05 U/ml creatine kinase (Sigma), and Spx or green fluorescent protein (GFP)-SsrA (concentrations are indicated in figure legends). One-hundred-microliter reaction mixtures were incubated at 30°C in the presence of ClpP, ClpX, or ClpC and MecA (concentrations are indicated in figure legends). At indicated time intervals (figure legends), a 12-μl sample from each reaction mixture was collected and treated with 3 μl of stop buffer (SDS loading dye in 0.1 M DTT). The proteins were then resolved on a 12% SDS-PAGE gel, followed by staining with Coomassie blue. Levels of Spx were defined as ratios of Spx band intensity to ClpP band intensity since ClpP concentrations in each set of reactions were equal. The Spx/ClpP ratio in a reaction mixture containing no ClpX was given the value 100%.

Complementation of yjbH with His-tagged yjbH and yjbH^Δ1-24.

For yjbH and yjbH^Δ1-24, DNA was PCR amplified using the forward primers 5′-ATA TAA GCT TAA AAG GTG GTG AAC TAC TAT GAC AAA CTA TCA GCA TGA G-3′ and 5′-ATA TAA GCT TAA AAG GTG GTG AAC TAC ATG TTT GTA GAC CCT TTA TG-3′, respectively, with the reverse primer 5′-ATA TGC TAG CTC AGT GGT GGT GGT GGT G-3′ and pYPET DNA as a template. PCR products were digested with HindIII and NheI and cloned at identical sites in plasmid pDR111 (gift from D. Rudner and A. D. Grossman) to generate plasmids pYHISCOMP and pY^Δ1-24HISCOMP. Resulting plasmids were transformed to the strain ORB6952 [JH642 yjbH::tet (forward)], and transformants were selected on Difco sporulation medium agar containing spectinomycin and tetracycline. To check the integration of yjbH-his and yjbH^Δ1-24-his at the amyE locus, clones were plated on an LB-starch plate and grown for 16 h at 37°C. The clones having integration of yjbH-his at the amyE locus did not show a starch-degrading phenotype when stained with iodine solution. The resulting strains (ORB7364 and ORB7534) were examined to determine if yjbH-his and yjbH^Δ1-24-his complements the yjbH null mutation. To determine the relative concentration of Spx in the wild-type (JH642), yjbH mutant (ORB6952), yjbH-his complemented (ORB7364), and yjbH^Δ1-24-his complemented (ORB7534) strains, cells were collected from cultures at mid-log phase, and whole-cell lysates were prepared by protoplast isolation. Fifteen micrograms of protein was resolved by 12% SDS-PAGE. The protein was transferred on a nitrocellulose membrane, and immunoblotting was carried out with anti-Spx antibodies.

ICP-AES.

Purified YjbH (20 μM) and YjbH^Δ1-24 (14.5 μM) were dialyzed against buffer E and subsequently diluted 80-fold in metal-free MilliQ water to achieve a final volume of 2 ml. Samples were then subjected to inductively coupled plasma atomic emission spectroscopy (ICP-AES) using an Optima 2000 DV optical emission spectrometer (Perkin Elmer). To accurately estimate the amount of bound metal obtained, values were corrected for the dilution factor.

Estimation of Zn by PAR.

The compound 4-(2-pyridylazo)resorcinol (PAR) is a metal chelator, which absorbs light at 500 nm when complexed with a free zinc ion, with a ɛ₅₀₀ of 66,000 M⁻¹ cm⁻¹ (18). To release the Zn atom(s) from the proteins, 3 μM YjbH or YjbH^Δ1-24 was treated with proteinase K (20 μg/ml) for 30 min at 52°C; then 100 μM PAR was added to the samples, and absorbance was recorded at 500 nm. The molar concentration of zinc was measured by correlating the absorbance with absorbance of standard curves of known concentrations of zinc acetate. In order to measure the time course of zinc release after oxidation, YjbH and YjbH^Δ1-24 (∼100 μM) were initially saturated with zinc as described above. Proteins were then diluted to 2.5 μM in buffer E containing 1 mM DTT. To start the reaction, 100 μM PAR was added in the samples. Nonspecifically bound zinc was trapped by PAR in the absence of oxidation, as monitored by a slight increase in the A₅₀₀. When absorbance did not increase further, 10 mM H₂O₂ was added, and the A₅₀₀ was monitored every 30 s for 10 min at 25°C. The absorbance change was converted to the amount of released zinc using a standard curve of known amounts of zinc acetate.

RESULTS

Spx interacts with YjbH.

The phenotype of the yjbH null mutant suggested that the yjbH product might participate in the proteolytic control of spx by serving as an adaptor to bind and deliver Spx protein to the ClpXP protease for degradation. To examine whether Spx interacts with YjbH, the yeast two-hybrid system was used. Full-length yjbH and yjbH^Δ1-24 coding sequences were inserted into pGADT7 to produce YjbH-GAL4 (AD) and YjbH^Δ1-24-GAL4 (AD) fusions, respectively. Each of the fusions was introduced by transformation into a yeast strain containing pSN11 (Spx-GAL4 BD) (35). The physical interaction of Spx and YjbH proteins was monitored by the activation of ADE2 and HIS3 as indicated by growth in the absence of either adenine or histidine. As shown in Fig. 1A, full-length YjbH interacts with Spx; however, deletion of 24 amino acids at the N terminus eliminates Spx binding. A previously established interaction between Spx and αCTD of RNAP served as a positive control, whereas Spx and αCTD^cxs-1 (rpoA^cxs-1, a Y263C codon substitution) was used as a negative control for interaction. Yeast two-hybrid results suggest that Spx interacts with YjbH, and for successful interaction, the histidine-rich N terminus of YjbH is essential. To further confirm the results of in vivo interaction, in vitro protein-protein interaction studies of Spx and YjbH were undertaken. Thus, we attempted to purify YjbH and YjbH^Δ1-24 from the heterologous host, E. coli. The full-length yjbH (900 bp, with alternate start codon, TTG) and yjbH^Δ1-24 (828 bp, with ATG start) were PCR amplified, cloned into expression vector pET-23a, and overexpressed in E. coli BL21(DE3)pLYS. The truncated version of YjbH corresponds to the predicted product of the yjbH coding sequence reported in the BSORF (http://bacillus.genome.ad.jp/) and Subtilist (http://genolist.pasteur.fr/SubtiList/) websites. The full-length coding sequence was uncovered and reported by Rogstam and coworkers (42). The overexpression of the YjbH and YjbH^Δ1-24 proteins at 37°C resulted in the formation of inclusion bodies. Reducing the temperature of incubation during induction or coexpression of cold shock molecular chaperones did not assist in producing a soluble form of YjbH or YjbH^Δ1-24 protein (data not shown). Therefore, we undertook purification of YjbH and YjbH^Δ1-24 by Ni²⁺-NTA affinity chromatography under denaturing conditions. The inclusion bodies were resuspended in buffer containing 6 N guanidine HCl. The denatured YjbH or YjbH^Δ1-24 was immobilized onto the Ni²⁺-NTA resin, and after in-column renaturation, soluble protein was obtained in the eluate. To further improve the protein purity, anion-exchange chromatography was carried out, and protein purity was analyzed by SDS-PAGE (Fig. 1B). On SDS-PAGE, recombinant YjbH and YjbH^Δ1-24 correspond to a molecular mass of ∼37 kDa and 34 kDa, respectively, which is in agreement with predicted theoretical molecular masses of YjbH and YjbH^Δ1-24 with a His₆ tag.

FIG. 1.

Spx interacts with YjbH. (A) Spx interacts with YjbH in the yeast two-hybrid assay. Gal4p AD-YjbH specifically interacts with Gal4p DNA BD-Spx but not with Gal4p BD alone. The N-terminal deletion of 24 amino acids from YjbH (AD-YjbH^Δ1-24) also resulted in loss of the interaction with Spx. The yeast two-hybrid reporter strain (PJ69-4a) expressing different BD and AD fusion proteins was assayed for growth on an adenine dropout plate or a histidine dropout plate. The interaction between Spx-αCTD and Spx-αCTD (rpoA^cxs-1) was used as a positive and negative control, respectively. (B) Purification of YjbH and YjbH^Δ1-24 of B. subtilis from E. coli. YjbH and YjbH^Δ1-24 were overexpressed in E. coli BL21(DE3)pLysS, followed by purification on an Ni²⁺-NTA affinity column under denaturing conditions. A Coomassie brilliant blue-stained 12% reducing SDS-PAGE gel shows the purified YjbH-His₆ (4 μg) and YjbH^Δ1-24-His₆ (4 μg). Molecular mass markers are shown in the lane on the left. (C) Glutaraldehyde cross-linking of Spx and YjbH. A concentration-dependent cross-linking of Spx with 0 to 10 μM YjbH/YjbH^Δ1-24 was carried out as described in Materials and Methods. The Spx-YjbH complex was resolved by 12% SDS-PAGE. Immunoblotting (IB) was carried out with anti-Spx polyclonal serum, followed by incubation with alkaline phosphatase-conjugated anti-rabbit immunoglobulin G secondary antibodies. The blot was developed using 5-bromo-4-chloro-3-indolylphosphate/nitroblue tetrazolium as a substrate. (D) Spx-YjbH complex formation is oxidation sensitive. The glutaraldehyde cross-linking was carried out in the presence of the thiol oxidant diamide, and complex formation was examined as described above.

In order to examine the interaction between Spx and YjbH protein, cross-linking studies were conducted. The cross-linking of Spx with YjbH and YjbH^Δ1-24 confirmed that Spx forms a complex exclusively with YjbH in a concentration-dependent manner (Fig. 1C). These data are in agreement with the yeast two-hybrid results that the N-terminal histidine-rich region of YjbH is essential for interaction with Spx as no cross-linked product was formed in the reaction of YjbH^Δ1-24 with Spx (Fig. 1C). In order to examine the influence of disulfide stress on Spx and YjbH interaction, cross-linking was carried out in the presence of the thiol oxidant diamide. Spx forms a complex with YjbH under reducing conditions (Fig. 1D), but diamide inhibited complex formation, indicative of a possible redox switch governing Spx-YjbH interaction.

YjbH enhances the ClpXP-mediated proteolysis of Spx.

The results above and previous findings (25) strongly suggest that YjbH can function as an adaptor protein that targets Spx for ClpXP-catalyzed proteolysis. To test this speculation, in vitro proteolysis of Spx in reaction mixtures containing ClpXP and ClpCP was carried out in the presence of YjbH. As shown in Fig. 2A and B, YjbH enhanced the degradation of Spx by ClpXP. In the presence of full-length YjbH, ∼98% of Spx was degraded in 15 min of proteolysis (Fig. 2A and C and 3). On the other hand, in a reaction mixture with no YjbH protein, only ∼37% of Spx was degraded in the same period of proteolysis (Fig. 2B and C and 3). As demonstrated previously, YjbH^Δ1-24 failed to interact with Spx, and it also failed to enhance the degradation of Spx by ClpXP (Fig. 3). Indeed, in the presence of YjbH^Δ1-24, only ∼38% of Spx was degraded in 15 min of proteolysis, which is comparable to what occurred in the reaction mixture containing no YjbH (Fig. 3B).

FIG. 2.

YjbH enhances ClpXP proteolysis of Spx. Effect of YjbH on ClpXP-catalyzed proteolysis of Spx in vitro. Spx (11 μM), ClpX (3 μM), and ClpP (11 μM) with an ATP-generating system (creatine kinase) were incubated at 30°C for the times indicated in the absence (A) or presence (B) of YjbH (10 μM). Reactions were performed in proteolysis reaction buffer containing 50 mM HEPES-KOH (pH 7.6), 50 mM KCl, and 10 mM Mg acetate as described in Materials and Methods. The reactions in 12-μl samples were stopped by adding 3 μl of SDS-PAGE loading buffer containing 100 mM DTT. Samples were analyzed on a 12% SDS-PAGE gel, followed by staining with Coomassie blue. Creatine phosphate kinase (CrK; 0.05 U/μl), 5 mM ATP, and 5 mM creatine phosphate were used as an ATP-regenerating system. (C) The plot of Spx band intensities against time of reaction was derived from two independent proteolysis experiments. The intensities of ClpP protein in each lane were used as internal controls (55). The Spx/ClpP ratio in the 0-min reaction was referred to as 100%.

FIG. 3.

The YjbH^Δ1-24 mutant does not enhance Spx proteolysis. (A) Spx (6 μM), ClpX (6 μM), and ClpP (12 μM) were incubated at 30°C in the presence of ATP, an ATP-generating system (creatine kinase), and 10 mM DTT with or without YjbH or YjbH^Δ1-24 (6 μM). The top panel shows an SDS-PAGE gel profile of reactions without YjbH. The second panel shows a gel profile of reaction mixtures containing YjbH, and the bottom panel is a gel profile of reaction mixtures containing YjbH^Δ1-24. (B) A plot of Spx band intensities with the ClpP band as an internal control is shown. The inset contains an immunoblot of extracts from cells lacking yjbH and complemented with His-tagged yjbH and yjbH^Δ1-24 to examine accumulation of Spx. Cells were harvested at mid-log phase from the cultures, and cell lysates were prepared by protoplast isolation. The same amount of protein (15 μg) was loaded in each lane, and immunoblotting was carried out with anti-Spx antiserum. Lane 1, JH642; lane 2, ORB6952 (yjbH::tet); lane 3, ORB7364 [yjbH::tet (reverse) P_hyperspank-yjbH-his] without IPTG; lane 4, ORB7364 with IPTG; lane 5, ORB7534 [yjbH::tet (reverse), P_hyperspank-yjbH^Δ1-24-his] without IPTG; lane 6, ORB7534 with IPTG.

To determine if the His₆-tagged version of YjbH was active in vivo, it was produced in a B. subtilis strain bearing a yjbH::tet null mutation. As reported previously (26), the yjbH null mutant showed poor growth and elevated Spx protein concentration. Both growth and low Spx concentration were restored to the complemented strain (Fig. 3B, inset). Production of the His₆-tagged YjbH^Δ1-24 protein did not result in complementation of the yjbH null mutation (Fig. 3B, inset).

In order to investigate whether YjbH is a specific adaptor for Spx, an in vitro proteolysis reaction mixture was assembled that contained His₆-GFP-tagged SsrA and ClpXP. Either in the presence or absence of YjbH, SsrA-His₆-GFP was degraded at similar rates (Fig. 4), suggesting that YjbH is a specific adaptor protein for Spx. ClpCP protease complex has also been implicated in the proteolysis of Spx in the presence of another adaptor protein, MecA (36). We determined whether YjbH could also function as an adaptor for ClpCP in an in vitro proteolysis reaction mixture containing Spx. There was almost no degradation of Spx with or without YjbH (Fig. 5). However, in the presence of MecA, ClpCP efficiently catalyzed the proteolysis of Spx. Hence, the above results provide strong evidence for the role of YjbH as an adaptor protein to enhance ClpXP-catalyzed degradation of Spx.

FIG. 4.

Effect of YjbH on ClpXP-catalyzed proteolysis of GFP-SsrA in vitro. (A) GFP-SsrA (6 μM), ClpX (6 μM), and ClpP (12 μM) were incubated at 30°C in the presence of ATP, an ATP-generating system (creatine kinase), and 10 mM DTT with or without YjbH (6 μM) in a proteolysis reaction as described in the legend of Fig. 3. Twelve-microliter samples were taken at 0-, 5-, 10-, 20-, 30-, 45-, and 60-min time points. The top panel is a gel profile of reaction mixtures that did not contain YjbH, and the bottom panel shows gel profiles of reaction mixtures containing YjbH. The plot of GFP-SsrA/ClpP band intensity ratios against time of reaction was derived from two independent proteolysis experiments (B). The intensities of ClpP protein in each lane were used as internal controls. The GFP-SsrA/ClpP ratio in the 0-min reaction was referred to as 100%.

FIG. 5.

Effect of YjbH on ClpCP proteolysis of Spx in vitro. (A) Spx (6 μM), ClpC (3 μM), and ClpP (12 μM) were incubated at 30°C in the presence of ATP, an ATP-generating system (creatine kinase), and 10 mM DTT with or without YjbH (6 μM)/MecA (2.5 μM) in a proteolysis reaction as described in Materials and Methods. Twelve-microliter samples were taken at 0-, 2-, 4-, 6-, 8-, 10-, and 15-min time points. The top panel shows a gel profile of reaction mixtures containing no adaptor protein. The second panel shows gel profiles of reaction mixtures containing YjbH, and the lower panel is a gel profile of reaction mixtures containing MecA. The plot of Spx/ClpP band intensity ratios against time of reaction was derived from two independent proteolysis experiments (B). The intensities of ClpP protein in each lane were used as internal controls. The Spx/ClpP ratio in the 0-min reaction was referred to as 100%.

YjbH is a zinc-binding protein.

The fact that Spx protein levels increase upon oxidative stress in B. subtilis suggests possible redox control of proteolysis as part of the network that controls Spx concentration. Indeed, we had previously reported the potential role of the zinc-binding domain of ClpX, which is required for the ClpXP-dependent degradation of Spx, as a possible site of redox control (55). During normal growth, ClpX coordinates a Zn atom that not only helps in substrate and adaptor recognition but also contributes to the multimerization of the protein (3, 38, 51). Under oxidative stress, the Zn atom disassembles from ClpX, resulting in loss of ClpX's ability to interact and degrade Spx (55). However, to perform its function as a transcriptional activator during oxidative stress, Spx might also have to disassociate from the adaptor protein YjbH (25). Metal coordination and removal are common mechanisms to alter the conformation or biochemical function of a protein in response to changes in the redox state of the cell (19). The amino acid sequence of YjbH indicates the presence of a cluster of His residues at the N terminus upstream from a CXXC motif (Fig. 6A). Though the arrangement of His and Cys residues is not always indicative of a metal binding site, we sought to determine if YjbH binds a metal and, if so, whether the N terminus His cluster is required for metal binding. Zn-saturated YjbH was prepared and analyzed for the presence of Zn using ICP-AES and a colorimetric method that measures released Zn upon proteinase K treatment with PAR (Materials and Methods). Emission spectroscopy as well as colorimetric analysis showed the presence of Zn(II) atoms in YjbH (Fig. 6B and C). The exact stoichiometry of the zinc binding to the YjbH was difficult to pinpoint since both methods accounted for different Zn contents. However, we believe that ICP-AES is a more reliable method because Zn estimation by PAR is associated with nonspecific chelation of other divalent cations. Zinc estimation of the YjbH^Δ1-24 protein, having the N-terminal deletion of the histidine-rich cluster (residues 1 to 24), by ICP-AES and PAR showed the presence of half of the Zn content in comparison to that with full-length YjbH. Therefore, we conclude that full-length YjbH possibly binds to two Zn atoms and that the N-terminal His-rich region plays a pivotal role in coordination of at least one Zn atom.

FIG. 6.

YjbH is a Zn binding protein. (A) Diagrammatic representation and positions of histidine and cysteine residues in YjbH and YjbH^Δ1-24. (B) ICP-AES of YjbH and YjbH^Δ1-24 to show presence of Zn. The AES was carried out as described in Materials and Methods. Twenty micromolar YjbH and 14.5 μM YjbH^Δ1-24 were used in the measurement. (C) Estimation of zinc in YjbH and YjbH^Δ1-24 by PAR assay. The presence of zinc in the proteins was measured after proteinase K treatment, as described in Materials and Methods, followed by Zn estimation with 0.1 mM PAR. Light absorption was monitored at 500 nm. The amount of Zn was calculated from the standard curve generated with known concentrations of ZnCl₂.

Oxidation causes Zn release from YjbH.

In order to investigate the possibility that Zn-bound YjbH can sense the change in the redox state of the environment, oxidation of YjbH was carried out with hydrogen peroxide. Zn-saturated YjbH was treated with H₂O₂, and Zn release was monitored by PAR assay. As demonstrated in Fig. 7, out of possibly two zinc atoms, one was released rapidly from full-length YjbH upon oxidation with a half-life (t_1/2) of ∼1.5 min. About 3.5 mol of zinc was released per mol of YjbH. In contrast, the absorbance change was marginal in YjbH in the presence of the reducing agent DTT, implying that zinc is tightly bound to the reduced YjbH to resist chelation by PAR. Interestingly, the YjbH^Δ1-24 protein showed release of a single Zn atom upon oxidation, with a t_1/2 of ∼10 min, which was comparable to the rate of release of the second Zn atom from full-length YjbH (t_1/2 of ∼10 min). Both proteins contain the C-terminal His₆ tag, which could account for the binding of the second Zn atom. The above results suggest that the histidine-rich N-terminal region of YjbH coordinates a possible redox-sensing Zn atom, which upon oxidation rapidly dissociates from the protein to induce a conformational change, resulting in release of Spx from the YjbH-Spx complex. Difficulties were encountered in trying to obtain YjbH free of Zn due to the reduced solubility of apo-YjbH. The only soluble, active preparation of YjbH that was obtained was after renaturation of denaturant-treated protein in the presence of Zn ion. Hence, proteolysis reactions with YjbH lacking Zn were not performed.

FIG. 7.

Release of Zn from YjbH upon oxidation estimated by PAR assay. Zn-saturated YjbH and YjbH^Δ1-24 were prepared as described in Materials and Methods. The protein was diluted to 2.5 μM in 20 mM HEPES-KOH (pH 7.6) with 1 mM DTT and 0.1 mM PAR. The samples were divided into two tubes, into one of which 10 mM H₂O₂ was added. Zinc release was monitored by recording light absorption for 20 min at 500 nm every 30 s after the addition of H₂O₂. Absorption from untreated and reduced YjbH and YjbH^Δ1-24 was also recorded. The molar amount of released zinc per mol of YjbH was plotted.

DISCUSSION

The involvement and influence of Spx in critical cellular processes demand complex mechanisms to control its production at multiple levels of the gene expression process. Previous studies have shown that expression of spx is controlled at the transcriptional level by repressors, YodB and PerR (26), and by alternative forms of RNAP (13, 40, 41, 46). Initial studies of posttranslational regulation of spx revealed the role of ATP-dependent proteases, ClpXP and ClpCP, in the degradation of Spx protein (35, 36). The clpP or clpX mutant cells showed high levels of Spx. In vitro proteolysis of Spx by ClpCP in the presence of an adaptor (either MecA or YpbH) was shown to be more efficient than the degradation catalyzed by ClpXP (55). However, a clpC null mutant did not show any significant accumulation of Spx in comparison to the wild-type cells (S. Nakano and P. Zuber, unpublished data). Therefore, under normal growth conditions, ClpXP-mediated proteolysis of Spx seems to be the principle mechanism of posttranslational regulation of Spx concentration. Genetic studies revealed that yjbH mutant cells showed reduced sensitivity to the thiol oxidant diamide. Transcriptome analyses of the yjbH mutant uncovered changes in the synthesis of many transcripts belonging to the Spx regulon due to the accumulation of Spx protein (25). A possible biochemical function of YjbH was proposed in which YjbH serves as an adaptor protein that interacts with Spx to stimulate conformational change. According to the reported model, the change in Spx conformation involves the exposure of the C terminus of Spx, which enhances the ability of the ClpXP protease complex to recognize Spx as a substrate (25).

Regulators that are controlled by proteolysis often are subject to other mechanisms of regulation that target synthesis and activity. Such regulators also have profound effects on essential processes, and their persistent presence in the cell results in impaired growth and development. The general stress response sigma factor RpoS of E. coli is controlled at the transcriptional and translational levels (24). RpoS is also subject to regulated proteolysis by ClpXP and the substrate-specific adaptor protein RssB (4, 5, 31). The CtrA protein that governs stalk-to-swarmer cell conversion in Caulobacter crescentus is a member of the response regulator family of two-component regulatory proteins and is controlled by phosphorylation (12, 52). The concentration of CtrA is regulated by proteolysis that is catalyzed by the ClpXP protease (44). Recent studies have uncovered the RcdA protein that functions in CtrA turnover (30). As with results for yjbH, the phenotype of rcdA suggested that the gene might encode an adaptor that directed the ClpXP-catalyzed proteolysis of CtrA. However, in vitro studies showed that RcdA did not affect the rate of CtrA proteolysis by ClpXP (9) and suggest a different role for RcdA in CtrA control. In contrast, the purified YjbH protein fulfills two of the functions of an adaptor in that it binds to the substrate (Spx) and enhances ClpXP-catalyzed proteolysis of the substrate. It is not yet known if YjbH interacts with ClpX, as has been shown with the ClpXP adaptor of E. coli, SspB (28). YjbH could act as a tether for the substrate by also contacting the ClpX subunit of ClpXP, or it could simply render Spx sensitive to proteolysis by altering the conformation of the substrate, as suggested by Larsson et al. (25).

In the reactions described herein, YjbH is not consumed during Spx proteolysis, which is similar to E. coli SspB (28). This is unlike the SspB ortholog SspBα (8) and unlike MecA (47, 48), which is an adaptor required for degradation of the competence transcription factor, ComK, by ClpCP. This raises the question of how YjbH concentration and activity might be controlled. The yjbH gene resides within an operon with yjbI, and both genes are members of the Spx regulon (25, 34). Induction of the negative control factor for Spx (YjbH) would impart a negative feedback effect that might allow for a rapid return to the low Spx levels that characterize unperturbed cells after stress is relieved. The presence of a CXXC motif at the N terminus of YjbH suggests another level of control that involves a redox mechanism governing the adaptor activity of the protein.

In the present study we showed that YjbH could coordinate with Zn atoms. The arrangement of cysteine and histidine residues in the protein does not show a characteristic Zn-coordinating site found in other Zn-binding proteins (Fig. 4A). However, the possibility of a new Zn-coordinating motif cannot be ruled out. Emission spectroscopy as well as colorimetric analysis clearly showed the presence of Zn atoms in YjbH. Here, it is noteworthy that YjbH preparation was carried through denaturing conditions and that protein was allowed to fold artificially on a metal affinity column. Therefore, it was extremely difficult to establish the exact number of properly folded YjbH molecules in a given population. Thus, we believe it is difficult to exactly pinpoint the stoichiometry of Zn binding with YjbH. Moreover, there was evident a difference in the Zn estimations from ICP-AES and PAR analysis. However, we believe ICP-AES is more reliable, as discussed in Results. Hence, it was concluded that YjbH possibly coordinates two Zn atoms. Analysis of the YjbH^Δ1-24 protein uncovered a single Zn atom, suggesting a role for the N-terminal histidine and cysteine residues in coordination of at least one Zn atom. The Zn binding to the mutant YjbH^Δ1-24 and the more oxidant-resistant Zn interaction with the wild-type YjbH could be mediated by the C-terminal His₆ tag. The Zn coordinated by the N-terminal histidine and cysteine residues is rapidly released from the protein upon treatment with hydrogen peroxide (Fig. 7), suggesting that the N terminus bound to Zn might function in sensing the redox state of the cellular environment. The role of metal atoms in redox sensing and in the function of metalloproteins has been extensively studied. The Zn-binding anti-sigma factors have a redox-sensing domain that bears a Zn atom, which controls the anti-sigma activity depending upon the redox state of the cell (1, 29). In Hsp33 during oxidative stress, Zn atom is released from the protein, allowing the formation of two intramolecular disulfide bonds in the protein. The disulfide formation results in a global structural rearrangement and the induction of chaperone activity (20). Assuming a similar situation for YjbH, coupled with the finding that it acts as an adaptor protein for proteolysis of a transcriptional regulator required for the induction of the thioredoxin and thioredoxin reductase genes under disulfide stress, the results imply that YjbH should also sense the change in thiol-disulfide homoeostasis in B. subtilis. That treatment of YjbH with a strong oxidant results in the rapid expulsion of bound Zn from the protein indicates a possible role for the Zn atom in maintaining the global structure of the protein under reducing conditions. Possibly under disulfide stress, the Zn atom of YjbH disassembles from the protein, allowing free thiols to form an intramolecular disulfide(s) in the N-terminal metal-binding domain. The formation of intramolecular disulfide(s) may lead to the change in the structural determinants in YjbH that governs its interaction with Spx. Subsequently, the ability of YjbH to enhance the degradation of Spx via ClpXP might be lost. Thus, during disulfide stress the YjbH-Spx complex is destabilized, resulting in accumulation of free, active Spx molecules in the cell. Identifying the ligands of YjbH that coordinate Zn and the possible disulfide bond formation and its role in conformation change and interaction with Spx will be attractive goals for future studies.