The Chaperone ClpX Stimulates Expression of Staphylococcus aureus Protein A by Rot Dependent and Independent Pathways

The Clp ATPases (Hsp100) constitute a family of closely related proteins that have protein reactivating and remodelling activities typical of molecular chaperones. In Staphylococcus aureus the ClpX chaperone is essential for virulence and for transcription of spa encoding Protein A. The present study was undertaken to elucidate the mechanism by which ClpX stimulates expression of Protein A. For this purpose, we prepared antibodies directed against Rot, an activator of spa transcription, and demonstrated that cells devoid of ClpX contain three-fold less Rot than wild-type cells. By varying Rot expression from an inducible promoter we showed that expression of Protein A requires a threshold level of Rot. In the absence of ClpX the Rot content is reduced below this threshold level, hence, explaining the substantially reduced Protein A expression in the clpX mutant. Experiments addressed at pinpointing the role of ClpX in Rot synthesis revealed that ClpX is required for translation of Rot. Interestingly, translation of the spa mRNA was, like the rot mRNA, enhanced by ClpX. These data demonstrate that ClpX performs dual roles in regulating Protein A expression, as ClpX stimulates transcription of spa by enhancing translation of Rot, and that ClpX additionally is required for full translation of the spa mRNA. The current findings emphasize that ClpX has a central role in fine-tuning virulence regulation in S. aureus.


Introduction
Staphylococcus aureus is an opportunistic pathogen capable of causing a variety of diseases in humans, ranging from localized infections of skin and soft tissue to life-threatening systemic infections [1]. The pathogenicity of S. aureus relies on a wide array of surface-bound and secreted virulence factors that provide the bacterium with the ability of tissue binding, tissue destruction, and immune evasion [2]. These virulence factors are coordinately produced in a growth phase dependent manner. The cell-surface associated factors are primarily expressed during exponential growth phase, whereas expression of the secreted factors is induced upon transition to stationary phase. Central for this regulation is the quorum sensing agr locus [3].
Protein A is a major surface bound virulence factor found in all examined strains of S. aureus [4]. It is well-known for its ability to bind the Fc-region of IgG from several mammalian species [5]. Additionally, Protein A can bind von Willebrand factor, and is capable of inducing inflammatory responses in the host [5,6]. Accordingly, the importance of Protein A in infections has been demonstrated in several animal models [7][8][9].
Expression of Protein A is regulated by growth phase and is controlled by complex regulatory networks acting at both the transcriptional, translational and post-translational levels [10][11][12][13]. The complex regulation of Protein A expression has been schematically depicted in Fig. 1. At the pinnacle of this regulatory network is the Agr quorum sensing system reviewed in [14]. The effector molecule of the agr quorum sensing system is RNAIII, a small regulatory RNA that is strongly induced in the postexponential growth phase [14]. RNAIII is 514 nucleotides long and folds into a complex secondary structure comprising 14 distinct stem-loops [15]. Recent evidence supports that RNAIII fulfils its role as a global virulence regulator primarily by controlling translation of target genes [12,[16][17][18]. In regard to Protein A expression, RNAIII acts both directly and indirectly to reduce synthesis of the protein. Directly, RNAIII down regulates expression of Protein A by binding to the spa mRNA [12]. This binding reduces Protein A synthesis at two levels, as it both inhibits translation and promotes degradation of the spa mRNA [12]. Indirectly, RNAIII reduces transcription of spa by inhibiting translation of Rot [17]. As illustrated in Fig. 1, Rot activates spa transcription directly by binding to the spa promoter, and additionally, by enhancing transcription of sarS, encoded directly upstream of spa [11,13]. Similar to Rot, SarS activates spa transcription through direct interactions with the spa promoter region [10,11,13,19]. The activator activities of Rot and SarS are counteracted by the repressor SarA [10,13]. Present models suggest that SarS and Rot act synergistically to promote spa transcription, whereas the repressor SarA and the activator SarS compete for overlapping binding sites in the spa promoter [10,11,13].
We recently revealed an additional layer of regulation by showing that the levels of spa transcript and Protein A were severely reduced in mutants lacking ClpX [20,21]. Notably, we additionally showed that the clpX gene is essential for virulence of S. aureus [21]. The ClpX protein belongs to the family of closely related Clp ATPases that function as molecular chaperones [22,23]. Additionally, ClpX can interact with the unrelated ClpP peptidase forming a proteolytic complex [24][25][26]. In the ClpXP protease, ClpX serves to specifically recognize and subsequently unfold and translocate the substrate into the ClpP proteolytic chamber for degradation [24,[26][27][28]. Interestingly, we found that Protein A expression was un-affected by the absence of ClpP [20,21]. We therefore hypothesized that it is the ClpP independent chaperone activity of ClpX that is required for synthesis of Protein A. Further genetic analyses indicated that ClpX stimulates spa transcription by controlling either activity or synthesis of the transcriptional regulator Rot (Fig. 1) [20,21]. The present study was undertaken to elucidate the molecular mechanism by which ClpX regulates Protein A expression. Here, we report that expression of Protein A requires a threshold level of Rot. In the absence of ClpX, the Rot content is below this threshold level resulting in lack of Protein A expression. An important finding is that ClpX enhances expression of Rot at the level of translation, and that ClpX additionally enhances translation of the spa mRNA. Hence, ClpX performs dual roles in regulating Protein A expression, as ClpX stimulates transcription of spa indirectly by enhancing translation of Rot and additionally enhances translation of the spa mRNA.

Bacterial strains and growth conditions
Bacterial strains and plasmids are listed in Table 1. S. aureus 8325-4 and SA564 were used as wild-type strains. As primary recipient for plasmids the restriction deficient strain S. aureus RN4220 [29,30] was used. S. aureus strains were grown in Tryptic Soya Broth media (TSB; Oxoid) under vigorous agitation (200 rpm) at 37uC. Usually 20 ml of medium was inoculated in 200 ml flasks to allow efficient aeration of the medium. For solid medium, 1.5% agar was added to give TSA plates. Erythromycin (5 mg ml 21 ), chloramphenicol (5 mg ml 21 ), or tetracycline (5 mg ml 21 ) was added as required. Upon receipt of the low-passage isolate SA564 (from G. Somerville) the strain was cultured once and stored frozen at 280uC. In all experiments we used SA564 streaked freshly from this stock. Escherichia coli strains Top10, Top10F' or StrataClone Solopack Competent Cells were used for DNA cloning and were grown in Luria-Bertani (LB; Oxoid) broth or on LB agar plates at 37uC. Tetracycline (10 mg ml 21 ), kanamycin (50 mg ml 21 ), chloramphenicol (35 mg ml 21 ) or ampicillin (50 mg ml 21 ) was added as required.

Plasmid and strain construction
SA564DclpX. For construction of SA564DclpX, pSaDclpX from RN4220 [21] was electroporated into competent SA564 cells at 30uC as described in [31]. A strain containing a 655 bp in frame deletion in the clpX gene was obtained by allelic replacement as described in [21], except that the final plasmid-loss was performed at 37uC. Verification of the chromosomal clpX deletion was done as described in [21].
Restoration of the chromosomal clpX locus in 8325-4DclpX and SA564 DclpX As we were unable to clone an error-less clpX gene in E. coli an alternative strategy was employed to complement the clpX deletion in 8325-4 and SA564: In the course of attempting to create a clpX deletion in the strain Newman following the same procedure as described in the present manuscript and in [21], we obtained the desired chromosomal deletion of clpX (D. Frees, unpublished data). However, following replacement recombination these colonies  [10,11,13,41]. Translation of spa and rot mRNAs is inhibited by RNAIII [12]. We hypothesized that ClpX stimulates Protein A expression by stimulating either synthesis or activity of Rot (indicated by the dotted arrow still maintained a plasmid-borne copy of the intact clpX gene (and its promoter). This plasmid was transformed into SA564 and 8325-4, and used to restore the chromosomal clpX locus by doing the same double cross-over procedure as initially used to create the clpX deletion [21]. Colonies containing an intact chromo-somal copy of the clpX gene were identified by colony PCR using the primers SaclpX385f (59-GACGATGCAGAACAACGTG) and SaclpX2447r (59-CCATACCCTGGAACATCCAC), and subsequently plasmid loss was achieved by growing at the nonpermissive temperature. Construction of pASrot expressing anti-sense rot from an inducible promoter To vary the Rot content in SA564 we used the expression vector pRMC2 [32]. This plasmid harbors an improved copy of the tetO promoter in pALC2073 that ensures complete repression when non-induced and high expression when induced by anhydrotetracycline. The rot gene was PCR-amplified from 8325-4 chromosomal DNA using the primers KpnI-rotF (59-TTAC-CATGGTACC ATTGGGAGATGTTTAGCATG) and SacI-rotR (59-ACATCAGAGCTC CCAACGTAATCATGCTCCAT) which, except for the added restriction sites (underlined), are the same primers that J. Oscarsson used to clone rot in pWA163 [11]. The PCR-fragment was TOPO-cloned (Invitrogen), and after confirming that the resulting plasmid contained an error-less rotgene, it was cloned into pRMC2 in the anti-sense orientation using SacI and EcoRI. The plasmid was first introduced into RN4220, purified from this strain using the Qiagen Plasmid Midi prep kit and then introduced into SA564 by electroporation as described by [31].
pLOJ133. A fragment containing the rpoB-promoter region, the rpoB-leader, and the Shine-Dalgarno and startcodon was PCR amplified using the primers Prpofwd and PrpoRev2 (59-GC-GGATCCTGCCAAACAGATTCAC-39), digested with EcoRI and BamHI and ligated into pLUG520 where the PrpoB-spafragment had been removed using the same enzymes. This resulted in pLOJ133; PrpoB translationally fused to lacZ in a pLUG520 derivative.
The plasmids were introduced into S. aureus strain RN4220 by electroporation and purified using the miniprep kit from Omega. The plasmid isolation procedure was modified by incubating the cell suspension in solution I containing 50 mg ml 21 lysostaphin (Sigma) for 1 h at 37uC. The purified plasmids were subsequently introduced into other S. aureus strains by electroporation. Transductions were performed as described by [34] using Q11.

Overexpression and purification of Rot and ClpX Histagged fusion proteins and generation of anti-Rot-and anti-ClpX antibodies
A fragment containing the full-length rot gene was generated by PCR using primers o-loj-rot-7 (59-GGATTTCCATATGAA-AAAAGTAAATAACGACACTG-39) and o-loj-rot-10 (59-GGG-AATTCTTACACAGCAATAATTCGGTTTAAACTATTTT-GC-39), digested with NdeI and EcoRI and cloned in-frame with six N-terminal histidine codons downstream of the IPTG-inducible T7 promoter of pET28a+ (Novagen) resulting in plasmid pLOJ104. DNA sequence analysis was performed to verify the construction and the absence of mutations. However, we were unable to recover error-less clones and for the purpose of antibody production we chose to continue with a clone containing a mutation leading to an Ala129RArg substitution.
A fragment containing the full-length clpX gene was generated by PCR using primers ClpX100fwd (59-caccATGTTTAAATT-CAATGAAGATGAAGAAAATTTG-39) and ClpX100rev (59-TTA AGCTGATGTTTTACTATTATTAATTAAATTGCC-39), TOPO-cloned into pET100D/TOPO (Invitrogen) in-frame with six N-terminal histidine codons downstream of the IPTGinducible T7 promoter resulting in plasmid pET100clpX. DNA sequence analysis was performed to verify the construction and the absence of mutations. However, we were unable to recover errorless clones and for the purpose of antibody production we chose to continue with a clone containing a single amino substitution.
To induce synthesis of the His-tagged Rot and ClpX proteins in E. coli, ER2566/pLOJ104 and ER2566/pET100clpX were grown in 2xLB to an OD 600 of 1.0 and then induced by adding IPTG to a final concentration of 1.0 mM for 2 h. The overexpressed proteins were purified on a Nickel NTA resin as described by the manufacturer (Qiagen) and 200 mg of pure protein was used to immunize a rabbit (antibodies were produced by CovalAB Company) using standard procedures (Sambrook et al., 1989).

Overexpression and purification of Hfq protein and generation of anti-Hfq-antibodies
For purification of Hfq, the intein system (Impact-CN; New England Biolabs) was used. Protein purification was performed as described previously [35]. Anti-Hfq-antibody was produced by Charles River Laboratories by immunizing rabbits with purified Hfq protein. The specificity of the antibody was tested by Western blot analysis on purified Hfq protein and protein extracts.

Western blotting
S. aureus strains were grown in TSB as specified above. Exponential cells were harvested at OD 600 = 1.0 (+/20.1), and post-exponential samples at OD 600 = 2.0 (+/20.1), or OD 600 = 3.0 (+/20.1), and used for isolation of intracellular proteins. One ml aliquots were harvested and kept at 280uC until all samples were collected. The cell pellets were thawed on ice, resuspended in 50 mM TrisHCl, Ph = 7.4 to a calculated OD 600 of 10.0. PMSF, Dnase, Rnase and lysostaphin were added to the samples, and they were incubated at room-temperature for 15 min. Cellular debris was removed by centrifugation and the protein concentration of each sample was measured using the Bradford dye-binding procedure from Bio-Rad. For immunoblotting a total of 5 mg (or 10 mg for Hfq-immunoblotting) of each sample was loaded on NuPAGEH 4-12% Bis-Tris gels (Invitrogen) using MOPS-Buffer (Invitrogen). The proteins were blotted onto a polyvinylidene difluoride membrane using the XCell II blotting module (Invitrogen). The membranes were pre-blocked with IgG. The Rot and ClpX proteins were probed using Rabbit-anti-Rotantibody or Rabbit-anti-ClpX-antibody at a 1:2000 dilution. Protein A was probed using monoclonal Mouse-anti-Protein-Aantibodies (Sigma) at a 1:10000 dilution. The Hfq protein was probed using anti-Hfq antibodies at a 1:2500 dilution. Bound antibody was detected with the WesternBreeze Chemiluminescent Anti-Rabbit kit or Anti-mouse kit (Invitrogen). All Western blots were repeated at least three times with similar results.

RNA extraction and Northern blot analysis
Cells were grown in TSB at 37uC with shaking to the indicated OD 600 . 1 ml aliquots were harvested and immediately frozen and stored at -80uC. Cells were lysed mechanically using the FastPrep system (Bio101; Q-biogene), and RNA was isolated using the RNeasy mini kit (QIAGEN, Valencia, Calif.) according to the manufacturer's instructions. Total RNA was quantified by spectrophotometric analysis (l = 260 nm), and 5 mg of RNA of each preparation was loaded onto a 1% agarose gel and separated in 10 mM sodium phosphate buffer as described previously [20]. RNA was transferred to a positively charged nylon membrane (Boehringer Mannheim) by capillary blotting as described by Sambrook et al. Hybridization was performed according to [20] using gene-specific probes that had been labeled with [32P]dCTP using the Ready-to-Go DNA-labeling beads from Amersham Biosciences. Internal fragments of the rot gene (amplified with the primers o-loj-rot-7 and o-loj-rot-10), spa (amplified with the primers described in [20]) or the rnalll gene (amplified with the primers 5'-GATCACAGAGATGTTATGG + 5'-CATAG-CACTGAGTCCAAGG) were used as template in the labeling reactions. All steps were repeated in three independent experiments giving similar results.

Measurements of b-galactosidase activity
Cells were grown in TSB as described above to the indicated OD 600 . 1 ml aliquots were harvested and immediately frozen and stored at 220uC. Cell pellets were resuspended in 300 ml 0.9% NaCl containing 5 mg/ml lysostaphin and incubated at 37uC for 15 min until the samples became clear indicating lysis was complete. Cellular debris was removed by centrifugation and the samples were transferred to fresh tubes. The protein concentration of each sample was measured using the Bradford dye-binding procedure from Bio-Rad. Expression of b-galactosidase was quantified by adding 400 ml Z-buffer contaning 1 mg/ml ONPG to 100 ml of each sample as described previously [36]. Samples were incubated at 37uC until the solution had become yellow or not longer than 180 min. The reaction was stopped by adding 500 ml 1 M NaCO 3 and b-galactosidase activity was measured at OD 420 and the specific activity (1 unit = 1 nmole o-nitrophenol produced per min per mg protein) was calculated. All steps were repeated in three independent experiments giving similar results.
Statistical analysis. All significant values were determined using the 2 sample t-test with unequal variances. Error bars represent standard deviations.

Cells lacking ClpX contain reduced levels of the Rot transcriptional regulator
Similar to the transcriptional regulator Rot, ClpX is required for transcription of the spa gene, encoding Protein A [20,21]. Based on this observation we proposed that the chaperone activity of ClpX is required either to activate Rot, or, to control expression of Rot [20]. To estimate the level of Rot protein in the clpX mutant cells, we prepared polyclonal antibodies against Rot. The Rot antibodies were used to determine the relative amount of Rot protein in wild-type cells and in cells carrying deletions in clpX, clpP, or agr. As translation of the rot mRNA is inhibited by RNAIII [17], the Rot level was assayed both in exponential growth phase (prior to induction of RNAIII) and in post-exponential growth phase (RNAIII induced) (Fig. 2). The anti-Rot antibodies recognized a protein of the expected 16 kDa that was absent in protein extracts derived from the rot mutant ( Fig. 2A). In wild-type cells the Rot protein was most abundant during exponential growth phase ( Fig. 2A), which is consistent with the inhibitory role of RNAIII on rot translation [17,18]. In post-exponential cells the Rot level was gradually reduced, reaching 75% and 50% of the exponential level at OD 600 = 2.0 and 3.0, respectively. In accordance with the proposed model, the down regulation of Rot was clearly dependent on agr. Interestingly, cells lacking ClpX contained 2-3 fold less Rot protein than wild type-cells at all assayed time-points ( Fig. 2A). As restoration of the chromosomal clpX locus, resulted in wild-type expression of both Rot and Protein A we conclude that ClpX indeed has a stimulatory effect on Rot synthesis ( Fig. 2B and data not shown). The post-exponential down regulation of Rot was still observed in the clpX mutant, indicating that the effect of ClpX and RNAIII on Rot expression is mediated by different pathways. Moreover, cells lacking clpP contained wild-type levels of Rot protein at all assayed time-points (Fig. 2C), signifying that ClpX enhances Rot expression independently of ClpP.

Expression of spa is activated above a threshold level of Rot
After confirming that ClpX is required for full expression of Rot, we examined whether the 2-3 fold reduction of the Rot level can explain the dramatically reduced Protein A level of the clpX mutant [20]. To control the Rot-level, we employed a rot mutant that harbours a plasmid carrying the rot gene transcribed from a xylose inducible promoter [11]. Protein samples were derived from exponential cultures grown with varying amounts of xylose, and were used for Western blot detection of Rot and Protein A. As seen in Fig. 3, upper panel, Rot expression increased with increasing concentrations of xylose, and with 4% added xylose the Rot level approached the wild-type level. The experiment clearly demonstrated that Protein A synthesis was dependent on the Rot levels ( Fig. 3, central panel). Specifically, Protein A expression remained low until the level of Rot protein constituted approximately 75% of the wild-type level. As Rot controls Protein A synthesis at the level of transcription, this finding indicates that a threshold concentration of Rot must be exceeded to obtain induction of spa transcription. Importantly, in the clpX mutant, the Rot level was below this threshold level (the level was comparable to the Rot level in cells grown in the presence of 0.5-1% xylose). Hence, this experiment confirms that the reduced level of Rot in the clpX mutant can explain the severely reduced spa transcription in this strain [20]. Additionally, our results suggest that Protein A expression is regulated by a threshold mechanism.

The essential roles of ClpX and Rot in Protein A expression are conserved in a clinical strain
The increasing focus on virulence gene regulation in clinical strains has emphasized that regulatory models established in laboratory strains may not be representative of the regulatory mechanisms observed in clinical isolates [37][38][39]. To assess if the role of ClpX in Rot and Protein A synthesis is conserved in clinical isolates, we made the 651 bp in-fame deletion of the clpX gene in the low-passage isolate SA564 derived from a patient with toxic shock syndrome [37]. Expression of Protein A and Rot in the SA564DclpX strain and SA564 wild type strain was compared using Western blot analysis (Fig. 4A). Strikingly, the stimulatory role of ClpX on Rot/ProteinA synthesis appears even more pronounced in the SA564 background, as deletion of clpX in SA564 substantially reduced (approximately 6 fold) the cellular level of Rot, and concomitantly decreased spa transcription and Protein A synthesis to less than 5% of the wild-type level (Fig. 4A). Upon complementation of the clpX locus, Protein A and the Rot level were returned to wild-type levels (Fig. 4B). We conclude, that the positive roles of ClpX on Rot and Protein A synthesis are conserved in a low-passage clinical strain, suggesting that the underlying mechanism is conserved among S. aureus strains.
The SA564 strain expresses considerably more Protein A than 8325-4, and we therefore examined if Protein A expression is dependent on a threshold-level of Rot, as was observed for 8325-4 ( Fig. 3). To reduce the cellular Rot level in SA564, the rot gene was cloned in the anti-sense orientation under control of the tetO promoter in the expression vector pRMC2 (see Materials and Methods for details; 32). As can be seen in Fig. 5, we could achieve a minor reduction (2 fold at the highest concentration of anhydrotetracycline) in the cellular Rot concentration by inducing transcription of the rot anti-sense mRNA. Interestingly, this modest reduction of Rot reduced expression of Protein A as much as 10 fold. This finding supports that below a threshold level of Rot, Protein A synthesis is greatly reduced. rot transcription is slightly reduced in clpX and clpP mutant cells We next attempted to pin-point the level at which ClpX mediates its affect on Rot synthesis. We first examined if ClpX controls Rot synthesis at the level of transcription. To this end, Northern blot analysis was performed to determine the level of rot transcript in exponential and post-exponential wild-type, clpP, and clpX mutant cells (Fig. 6). The Northern blot revealed that both clpX and clpP mutant cells contained rot-transcript levels that were slightly reduced compared to wild-type levels. As described above, the Rot protein level in the clpP mutant strain was comparable to the wild-type level, suggesting that the reduced rot transcription in clpX and clpP mutant strains is not reflected at the protein level. The Northern blot also revealed that in all strains transcription of rot increased in the post-exponential growth phase, as also reported by others [18]. The post-exponential induction of rot transcription occurs concomitantly with the down-regulation of Rot at the protein level emphasizing the significance of posttranscriptional regulation in Rot expression. From this experiment, we conclude that ClpX controls Rot synthesis mainly at the post-transcriptional level.

ClpX does not control Rot expression by stimulating transcription of RNAIII
Next, we addressed if ClpX controls Rot expression by enhancing expression of RNAIII that is an inhibitor of Rot translation [17]. To this notion, the level of RNAIII in wild-type, agr, and clpX mutant cells in exponential and post-exponential growth phase was determined by Northern blot analysis (Fig. 7). The experiment demonstrated that the RNAIII level in the clpX mutant does not deviate from the wild-type levels. Hence, we ruled out that the reduced Rot level in the clpX mutant is accomplished through an increase in RNAIII synthesis.

ClpX stimulates translation of rot mRNA
To investigate directly whether ClpX regulates Rot translation, a translational fusion between the 59-rot-leader and lacZ was constructed as depicted in Fig. 8A. This construct encompasses the RNAIII interacting region, the ribosome binding site, and the first three amino acid codons of the rot leader translationally fused to lacZ [17,33]. Transcription of the fusion is driven by the constitutive promoter P rpoB . The plasmid harbouring the rottranslational fusion was introduced into wild-type, and agr, clpX, clpP, and agr;clpX mutant strains, and b-galactosidase activity was assayed in samples derived from post-exponential cultures (Fig. 8B). The assay revealed that b-galactosidase activity was almost five fold higher in the agr mutant than in the wild-type (Fig. 8B). This suppressive effect of RNAIII on rot-translation is in the same range as observed by others and, thus, validates the constructs [17]. While samples derived from clpP mutant cells expressed wild-type levels of b-galactosidase (data not shown), b-galactosidase activity was reduced approximately two fold compared to wild-type in clpX mutant cells. This two-fold difference was statistically significant (wild-type different from clpX; P-value = 0.008), indicating that ClpX indeed has a positive effect on translation of rot. Interestingly, b-galactosidase activity in the agr, clpX double mutant was substantially reduced compared to the activity measured in the agr single mutant strain. This suggests that even though the inhibitory effect of RNAIII is relieved by a deletion of the agr locus, translation of Rot still depends on ClpX. Taken  To reduce the level of Rot in SA564, anti-sense rot mRNA was expressed from the tetO promoter in pRMC2 in the presence of increasing concentrations of the inducer anhydrotetracyclin. Anhydrotetracyclin was added at OD 600 = 0.4 and proteins were extracted from samples derived at OD 600 = 2.0. Western blotting was performed with Rot antibody (upper panel) and protein A mono-clonal antibody (second panel). To visualize equal loading the signals detected from non-specific binding of the Rot antibody to an unknown cellular protein is shown. All protein levels were quantified relatively to the corresponding signal obtained in the un-induced culture (100%) using Gene Snap from Perkin Elmer. doi:10.1371/journal.pone.0012752.g005 together; these results specify that ClpX stimulates rot translation, and that the role of ClpX in rot translation is not linked to the RNAIII mediated inhibition of rot translation.
ClpX stimulates translation of spa mRNA but not of rpoB mRNA Huntzinger et al. showed that translation of spa (Protein A), similar to translation of Rot, is inhibited by RNAIII [12]. To examine if ClpX has a role in translation of Protein A, the plasmid, pLUG520, carrying the P rpoB -spa-lacZ translational fusion used by Huntzinger et al. [12] was introduced into wild-type and mutant strains. This construct encompasses the constitutive rpoB promoter and the first 63 nucleotides of the spa mRNA (including the sites for RNAIII interaction) fused in frame with lacZ deleted of its ribosomal binding site, and thus specifically measures the translational activity of the spa-leader. In samples derived from post-exponential cultures, b-galactosidase activity was approximately 1.5 fold higher in the agr background than in the wild-type background, in accordance with previous results obtained by Huntzinger et al. [12] (Fig. 8C). Importantly, the absence of ClpX reduced b-galactosidase activity expressed from the spa-lacZ translational fusion approximately 5 fold, both in wild-type and in cells devoid of agr/RNAIII. We conclude that translation of spa mRNA, similar to translation of rot mRNA, is positively affected by ClpX and that the positive effect of ClpX on spa and rot translation occurs independently of RNAIII.
To investigate if ClpX also contributes to translation of mRNAs not interacting with RNAIII, we examined expression of the house-keeping gene rpoB in the clpX mutant by using a translational fusion between the rpoB promoter and lacZ (pLOJ133). With this plasmid, b-galactosidase activity was similar in samples derived from all strains demonstrating that neither RNAIII nor ClpX regulates expression of RpoB (Fig. 8D).

Discussion
Protein A is an abundant surface protein that by its ability to bind the Fc fragment of immunoglobulins from mammalian species is believed to be important for immune evasion of S. aureus [5]. Previously, we have shown that ClpX (but not ClpP), similar to the transcriptional regulator Rot, is essential for transcription of the spa gene encoding Protein A, and therefore we hypothesized that the chaperone activity of ClpX is required for either synthesis or activity of Rot [20]. In the present study we confirm that ClpX is indeed required for full expression of Rot, as cells devoid of ClpX contain approximately 2-3 fold less Rot protein than wildtype cells. This finding raised the question whether the relatively modest 2-3 fold reduction in the cellular Rot level is sufficient to abolish spa transcription. We confirmed this by varying Rot expression from an inducible promoter and demonstrated that a 2 fold reduction in Rot expression causes a dramatic 90% reduction in Protein A expression in both 8325-4 and the low-passage clinical isolate SA564 [40]. This finding emphasizes that major changes in Protein A expression may be achieved by small reductions in the Rot level and suggests that a threshold level of Rot is needed for synthesis of protein A.
Rot stimulates Protein A transcription both directly, by binding to the spa promoter, and in-directly by enhancing transcription of  Cells of the wild-type and the agr, and clpX mutants were grown in TSB and at the indicated optical densities total RNA was extracted and separated on an agarose gel. RNA was blotted onto a nitrocellulose membrane and probed with an RNAIII specific probe. Below the blot is a picture of the ethidium stained agarose gel. The blot presented is representative of three independent experiments. doi:10.1371/journal.pone.0012752.g007 Figure 8. Translation of Rot and Protein A is decreased in the absence of ClpX. (A) Schematic drawing of the lacZ translational fusion points in pLOJ132 and pLOJ133. pLOJ132 is a P rpoB -rot-lacZ translational fusion. Transcription is driven from the constitutive P rpoB promoter. The rotleader encompassing the RNAIIII interacting region from +162 to +250 (numbered with respect to the transcriptional start site) is translationally fused to the lacZ gene. pLOJ133 is a P rpoB -lacZ translational fusion. (B) b-Galactosidase activity measured in post-exponential (OD 600 = 2.0) wild-type-and, clpX, agr, agr;clpX mutant strains harbouring pLOJ133. The specific activities are presented as a percentage of the wild-type activity (100%). P-values were calculated using the two-sample t-test with unequal variances: wild-type different than clpX: P-value = 0.008; agr different than agr;clpX: Pvalue = 0.0004. The statistical calculations were based on five independent experiments. (C) Protein A translation was measured as b-Galactosidase activity expressed from pLUG520 in wild-type-and clpX, agr, and agr;clpX mutant cells in the post-exponential growth phase (OD 600 = 2.0). Specific sarS, encoding another positive activator of spa transcription [11,13,41]. Interestingly, transcription from the spa promoter was suggested to be regulated by negative (SarA) and positive regulators (SarS and Rot) competing for overlapping operator sites [13]. Therefore we propose that the threshold level of Rot (and SarS) is required to relieve the repression imposed by SarA, and that the ClpX dependent expression of Rot is required to exceed this threshold level. It is tempting to speculate that additional genes of the Rot regulon are subject to a similar regulatory mechanism.
As deletion of clpX severely decreased the Rot level (6 fold compared to the max 3-fold in 8325-4), and accordingly abolished expression of Protein A in the low-passage isolate SA564, we propose that the mechanism by which ClpX controls the Rot and Protein A level is conserved among clinical S. aureus isolates.
Attempts to pin-point the level at which ClpX mediates its effect on Rot synthesis demonstrated that ClpX stimulates expression of Rot both at the level of transcription and translation ( Fig. 6 and 8). Based on the finding that transcription of rot was decreased similarly by the absence of either ClpX or ClpP, we propose that ClpX controls rot transcription as part of the ClpXP proteolytic complex. Notably, the clpP mutant strain contained wild-type levels of Rot suggesting that the reduction in rot transcription observed in clpX and clpP mutant strains is not reflected in the cellular Rot level, and hence that the ClpP independent effect of ClpX in controlling Rot synthesis is achieved mainly at the posttranscriptional level. Accordingly, we were able to show that ClpX does indeed stimulate translation of the rot mRNA, and that ClpX additionally enhances translation of the spa mRNA. Initially this observation prompted us to suspect that ClpX modulates the interaction between RNAIII and its target mRNA's. This interaction involves base pairings between RNAIII and secondary hair-pin structures in the target mRNA by which dual loop-loop complexes are formed. These interactions block ribosomal access to the Shine-Dalgarno of the respective messengers, thereby inhibiting translation [17,18]. However, as we observed a substantial ClpX dependent stimulation of spa and rot translation in cells lacking RNAIII, we conclude that ClpX stimulates translation independently of RNAIII.
Secondary structure determinations of the spaand rot-mRNAs reveal that the translational signals are partially obscured in the stem-loop structures, also when not bound to RNAIII [12,17]. Consequently, it is possible that the structures of the spa and rot transcripts required for interactions with RNAIII could hamper interactions with the ribosomes even in the absence of RNAIII. Hence, one way ClpX may promote translation of the rot and spa mRNAs is by, directly or indirectly, unfolding the stem-loop structures to facilitate interaction of these mRNAs with the translational apparatus.
In E. coli, the abundant RNA chaperone Hfq has a central role in translational stimulation of specific mRNAs [42]. Hfq is the only RNA chaperone yet identified in S. aureus. Accordingly, Hfq could be a candidate through which ClpX mediates its function in translational stimulation. However, recently the biological significance of Hfq in S. aureus was questioned by the lack of phenotypes of an hfq mutant and by the lack of detection of the hfq mRNA [18,40]. Consistent with this finding we were not able to detect Hfq in our strains when performing Western blot analysis using anti-Hfq polyclonal antibodies (raised against Hfq purified from S. aureus, see Materials and Methods, Supplemental Fig. S1). Furthermore, as both expression of Protein A and Rot did not change in strains lacking Hfq [40], it seems unlikely that the stimulatory effect of ClpX on translation of the rotand spa transcripts is mediated through Hfq. Rather, S. aureus may harbour a yet unidentified RNA chaperone that is subject to regulation by ClpX.
An intriguing alternative possibility is that ClpX could affect translation through direct binding to specific mRNAs. The ClpX family of proteins contains a C4 zinc-binding domain in the Nterminal region and this type of domain has been implicated in nucleic acid binding in both eukaryotic and prokaryotic organisms [43,44]. In addition, a eukaryotic relative of ClpX, the HSP101 heat shock protein of plants specifically binds to and stimulates translation of the A-rich 59 un-translated leader of the tobacco mosaic virus [45]. Inspection of the three-dimensional structure of ClpX, furthermore, reveals a striking resemblance to the RNA binding protein Hfq as both proteins form homo-hexameric ringstructures with similar sizes of the central pores. Further experiments are underway to determine the mechanisms of ClpX mediated translational stimulation.
In conclusion, we have shown that ClpX has a dual role in stimulating Protein A synthesis, as ClpX enhances both transcription (through enhancing synthesis of Rot), and translation of spa. Our findings emphasize that ClpX through its role in fine-tuning the level of the central regulator Rot holds an important role in controlling global expression of virulence factors in S. aureus. Figure S1 Hfq could not be detected in Western blot analysis. Cells were harvested in early stationary phase or late stationary phase. Cells were lysed as described in materials and methods and total protein extract were separated on an SDS-gel. 10 mg protein was loaded (0.5 mg purified S. aureus Hfq protein).and samples were heated to 95uC for 20 min prior to loading. For comparison of protein sizes, the PageRulerTM Plus Prestained Protein Ladder (Fermentas) was included (not shown). The Hfq antibody specifically recognizes the monomeric and multimeric species of the purified Hfq protein, whereas no bands at corresponding sizes were observed in any of the cellular protein samples tested. Found at: doi:10.1371/journal.pone.0012752.s001 (2.49 MB TIF)