Global Regulatory Impact of ClpP Protease of Staphylococcus aureus on Regulons Involved in Virulence, Oxidative Stress Response, Autolysis, and DNA Repair

Abstract

Staphylococcus aureus is an important pathogen, causing a wide range of infections including sepsis, wound infections, pneumonia, and catheter-related infections. In several pathogens ClpP proteases were identified by in vivo expression technologies to be important for virulence. Clp proteolytic complexes are responsible for adaptation to multiple stresses by degrading accumulated and misfolded proteins. In this report clpP, encoding the proteolytic subunit of the ATP-dependent Clp protease, was deleted, and gene expression of ΔclpP was determined by global transcriptional analysis using DNA-microarray technology. The transcriptional profile reveals a strong regulatory impact of ClpP on the expression of genes encoding proteins that are involved in the pathogenicity of S. aureus and adaptation of the pathogen to several stresses. Expression of the agr system and agr-dependent extracellular virulence factors was diminished. Moreover, the loss of clpP leads to a complete transcriptional derepression of genes of the CtsR- and HrcA-controlled heat shock regulon and a partial derepression of genes involved in oxidative stress response, metal homeostasis, and SOS DNA repair controlled by PerR, Fur, MntR, and LexA. The levels of transcription of genes encoding proteins involved in adaptation to anaerobic conditions potentially regulated by an Fnr-like regulator were decreased. Furthermore, the expression of genes whose products are involved in autolysis was deregulated, leading to enhanced autolysis in the mutant. Our results indicate a strong impact of ClpP proteolytic activity on virulence, stress response, and physiology in S. aureus.

The Clp proteases were first identified in Escherichia coli and consist of an ATPase specificity factor (ClpA or ClpX in E. coli; ClpX, ClpC, or ClpE in Bacillus subtilis) and a proteolytic domain (ClpP) that contains a consensus serine protease active site (33). In E. coli, ClpP-mediated proteolysis is regulated by heat shock and removes abnormal proteins that accumulate during stress conditions, recycles amino acids from nonessential proteins during starvation, and contributes to the clearance of truncated peptides from stalled ribosomes by the SsrA-tagging system (34, 65). Moreover, it has been shown that Clp proteases play a significant role in certain processes regulating cellular functions via proteolysis (33, 36, 45). For example, in E. coli ClpXP is involved in degradation of regulator proteins including the alternative sigma factor SigS, the UmuD SOS protein, and different phage proteins (23, 31, 34). Regulatory proteolysis is presumably determined by certain amino acid sequences which serve as a degradation signal. Flynn et al. have identified more than 50 proteins in E. coli as potential substrates for proteolysis by ClpXP (19). Further substrate proteins with regulatory functions were identified in other bacteria such as CtrA in Caulobacter crescentus, sigma s and FlhC/FlhD in Salmonella enterica serovar Typhimurium, PopR in Streptomyces lividans, and HdiR in Lactococcus lactis (44, 69, 73, 74). In B. subtilis Clp-specific target proteins were recognized which are involved in peptidoglycan synthesis, competence, sporulation, and heat shock regulation (30, 48, 50).

In addition, several studies indicate that ClpP proteolytic activity is critical for virulence of pathogenic bacteria, including S. enterica serovar Typhimurium, Streptococcus pneumoniae, Listeria monocytogenes, and Staphylococcus aureus (26, 29, 37, 58, 63, 73, 75). Interestingly, ClpP seems to be essential for survival of L. monocytogenes in murine macrophages (28, 29). Furthermore, clpX and clpP mutants of S. aureus, respectively, were attenuated in a murine abscess model (26), and ClpC plays an important role for long-term survival and for intracellular replication of this pathogen (12, 25). More recently, Frees et al. found that clpP deletion (ΔclpP) in S. aureus 8325-4 resulted in impaired virulence properties (26). In this study, the global regulatory agr locus was repressed in the ΔclpP strain, giving rise to a reduced α-toxin and extracellular protease activity. Moreover, the ΔclpP strain was more sensitive against hydrogen peroxide and not able to replicate intracellularly. The authors suggested that the reduced virulence of the ΔclpP strain is most likely due to reduced agr-regulated virulence gene expression rather than to decreased stress resistance (26). In addition, there are indications that Rot (repressor of toxin) in complex with RNAIII is a substrate of Clp-dependent degradation regulating serine protease sspA expression (27). All these reports suggest that ClpP proteolytic activity is important not only for cell physiology but also for regulation of virulence properties of pathogenic bacteria.

In order to get a more comprehensive picture of the role of ClpP protease on global transcription in S. aureus and how it relates to physiology and virulence, a ΔclpP strain was constructed in strain S. aureus 8325, and global gene expression was studied by comparative DNA microarray hybridization. We report here that clpP deletion affects the expression of genes belonging to specific regulons which are involved in adaptation to changes in the physiological state of the cell as well as in virulence.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. E. coli strain DH5α and S. aureus strains were grown in Luria-Bertani broth (LB) unless otherwise indicated. The recombinant E. coli and S. aureus clones were cultivated under selective pressure with either ampicillin (100 μg ml⁻¹), chloramphenicol (10 μg ml⁻¹), or erythromycin (10 μg ml⁻¹), respectively. For growth curves and RNA isolation, overnight cultures of S. aureus were diluted to an optical density at 600 nm (OD₆₀₀) of 0.01 in LB and were incubated at different temperatures (20°C, 30°C, 37°C, 42°C, and 45°C) under aerobic conditions with orbital shaking (180 rpm). Samples were collected in intervals during the first 10 h. Samples for RNA isolation were collected in the exponential growth phase (OD₆₀₀ of 1.0).

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant genotype or phenotype	Source or reference
Strains
E. coli DH5α	λ− φ80dlacΔM15 Δ(lacZYA-argF)U169 recA1 endA1 hsdR17 (rK− mK−) supE44 thi-1 gyrA relA1	68
S. aureus
RN4220	NCTC 8325-4-r (restriction mutant, with 11-bp deletion in rsbU)	49
8325	NCTC 8325 (wild-type, with 11-bp deletion in rsbU)	Laboratory stock
8325ΔclpP	clpP deletion strain of 8325	This study
8325ΔclpP+	8325ΔclpP, containing pHPS9clpP	This study
8325ΔagrA	agrA deletion strain of 8325	This study
8325ΔagrC	agrC deletion strain of 8325	This study
Plasmids
pBT2	Shuttle vector; Apr in E. coli; Cmr in S. aureus	9
pEC1	Apr EmrermB fragment in pUC18	9
pBT2ΔclpP	Deletion vector for clpP; ermB fragment flanked by fragments upstream and downstream of clpP in pBT2; Apr in E. coli; Emr and Cmr in S. aureus	This study
pHPS9	Shuttle vector; Cmr in E. coli; Emr in S. aureus	35
pHPS9KclpP	pHPS9, containing clpP fragment for ΔclpP complementation	This study

Construction of the S. aureus ΔclpP mutant.

For construction of a ΔclpP mutant in S. aureus, two PCR fragments of 1,046 bp and 943 bp, encompassing the up- and downstream regions of the clpP gene in strain 8325, corresponding to SA0724 of strain N315, were amplified using primer with added HindIII and BamHI restriction sites for the upstream fragment and PstI and EcoRI for the downstream fragment (see Table S2 in the supplemental material). The plasmid pEC1 containing the ermB resistance cassette was digested with BamHI and PstI. The up- and downstream fragments and the ermB cassette were ligated in one step into the temperature-sensitive shuttle vector pBT2, which was digested with EcoRI and HindIII (9). Construction of this deletion vector was carried out in E. coli DH5α and subsequently introduced into S. aureus strain 8325 by transduction with phage φ85 (72). In this strain gene inactivation was carried out as described by Brückner (9). Successful homologous recombination and loss of the plasmid were proven by Southern blot hybridization and PCR.

Complementation of the ΔclpP strain.

For complementation of the ΔclpP strain, an 824-bp PCR fragment containing the entire SA0723 locus and the ribosome-binding site was amplified by PCR using primer with added BamHI and EcoRI restriction sites and ligated into the shuttle vector pHPS9 (35). The plasmid was transformed into RN4220 by electroporation and transduced into the ΔclpP strain by using phage φ85. Clones were selected using erythromycin and chloramphenicol.

RNA preparation and Northern blot analysis.

Total RNA was isolated from S. aureus cultures in the exponential growth phase (OD₆₀₀ of 1.0). Bacteria were harvested with the addition of RNA Protect (QIAGEN, Hilden, Germany) according to the manufacturer's instruction. Prior to RNA isolation bacteria were lysed using glass beads in a Fast Prep shaker (Qbiogene, Heidelberg, Germany) for 45 s at a speed of 6.5 units. RNA was isolated using a QIAGEN RNeasy kit according to the standard QIAGEN RNeasy protocol. Ten to twenty micrograms of total RNA was used to perform a denaturing agarose gel electrophoresis and Northern blot hybridization as described previously (2). The probes were generated by PCR by using the primer sets listed in Table S2 in the supplemental material and were labeled by use of an ECL kit (Amersham Biosciences, Freiburg, Germany). Hybridization was performed as described in the manufacturer's instructions. The signals were quantified by densitometric scanning.

Semiquantitative reverse transcription-PCR (RT-PCR).

Reverse transcription was performed using 2 μg of DNase I-treated RNA samples, a random hexamer primer mix and Superscript III TM reverse transcriptase (Invitrogen, Karlsruhe, Germany) at 50°C for 1 h. The cDNA was adjusted to 40 μl with double-distilled water and amplified in different PCRs (including negative controls) with primers specific for the corresponding genes (for primers, see Table S2 in the supplemental material).

Microarray analysis.

S. aureus N315 full genome microarrays containing PCR products of 2,334 genes were used for microarray analysis (Scienion, Berlin, Germany). Each slide contained 6,336 features corresponding to duplicate copies of each open reading frame (ORF) and several controls. Total RNA for DNA microarray analysis was isolated from cultures in the exponential growth phase at an OD₆₀₀ of 1.0 at 37°C. Reverse transcription and fluorescent labeling reactions were performed using 10 μg of total RNA using random primers and Superscript III reverse transcriptase (Invitrogen), and cDNA was concomitantly labeled using the dyes Cy3 and Cy5 according to the manufacturer's instructions (Scienion). RNA obtained from four different biological experiments was utilized, and a reverse labeling (dye switch) experiment was performed to minimize bias due to differential dye bleaching or incorporation of the Cy3 and Cy5 dyes during the RT reaction. Microarray hybridization (16 h at 50°C) and washing of the slides were performed according to the manufacturer's instructions. Hybridized slides were scanned using a Genepix 4000B laser scanner (Axon Instruments Inc., Union City, CA). Bioinformatic analyses on the slide hybridization results of each single experiment were performed using Genepix Pro3.0 (Axon Instruments Inc.). Data of each image were normalized to the mean ratio of means of all features. Different experiments were normalized to each other using Expressionist software, version 3.1 (Genedata, Martinsried, Germany). Mean values and standard deviations of gene expression ratios based on two spot replicates on each microarray and four different hybridization experiments were calculated in Microsoft Excel XP.

Triton X-100-induced autolysis assays.

Autolysis assays were performed as described by Mani et al. (56). Bacteria were grown in tryptic soy broth (TSB) containing 1 M NaCl to an OD₆₀₀ of ∼0.7 at 37°C with shaking at 250 rpm. After one wash with phosphate-buffered saline (PBS), cells were resuspended in the same volume of 0.05 M Tris-HCl buffer (pH 7.5) containing 0.1% Triton X-100 and were incubated at 30°C with shaking. The optical density was measured in intervals. Results were normalized to an OD₆₀₀ at time zero, and percent lysis was calculated.

Physiological analysis.

For analysis of physiological changes in the mutant, an API-Staph test system was used according to the manufacturer's instructions (bioMérieux, Nürtingen, Germany).

Fibronectin binding assays.

The capability of the S. aureus strains to bind fibronectin was measured by using a radiometric assay described by Hussain et al. (42).

Infection experiments and gentamicin-lysostaphin protection assays.

Gentamicin-lysostaphin protection assays were performed as described by Agerer et al. (1). Briefly, overnight cultures of S. aureus were diluted 1:100 in TSB and were cultured to an OD₆₀₀ of 1.0. Bacteria were harvested and washed twice with PBS. For gentamicin-lysostaphin protection assays, 293T cells (2 × 10⁵ cells/well) were infected with bacteria at a multiplicity of infection of 20. After 2 h of coincubation at 37°C, the culture medium was replaced by Dulbecco's modified Eagle's medium-10% calf serum containing 50 μg/ml gentamicin and 20 μg/ml lysostaphin, and cells were further incubated for 45 min. Cells were washed with PBS, and intracellular bacteria was released by incubation in 1% saponin in PBS for 20 min at 37°C. Samples were diluted in PBS and plated on TSB agar plates for determination of the recovered CFU.

Scanning electron microscopy.

For scanning electron microscopy, staphylococci were grown overnight in TSB medium on polystyrene chamber slides at 37°C. After the medium was decanted, the slides were washed three times with 1× PBS, mounted on aluminum stubs, and shadowed with gold. For visualization, a scanning electron microscope (Zeiss DSM962) was used at 15 kV.

RESULTS AND DISCUSSION

Clp proteases are responsible for degradation of misfolded proteins under certain stress conditions (31). Recently, it has been reported that S. aureus ClpP is required for growth at reduced and elevated temperatures on solid medium (26). To obtain a more detailed view of the ability of a ΔclpP strain to cope with nonpermissive temperatures, growth experiments in liquid culture were performed by comparing growth at different temperatures of the ΔclpP strain to the parent strain 8325 and the complemented mutant ΔclpP⁺ strain. The ΔclpP strain showed a growth defect at all temperatures tested (37°C, 30°C, 20°C, 42°C, and 45°C) (Fig. 1A). The temperature sensitivity of the mutant was especially observed at reduced temperatures. At 20°C the mutant grew for 6 h with a similar growth rate as the wild type; however, subsequently the cells ceased growth (Fig. 1A). To investigate transcriptional activation of clpP at reduced temperatures, Northern blot analysis was performed after growth of the wild-type strain at 37°, 30°, and 20°C, respectively. A threefold induction of clpP could be observed at 30°C, and transcription of the clpP gene was increased fourfold at 20°C (Fig. 1B). These results suggest an increase of ClpP protease activity at lower temperatures which might be an essential response of S. aureus to survive under these conditions. Low temperatures, similar to heat shock, cause extensive protein denaturation and subsequent aggregation (24). Likewise, in the cyanobacterium Synechococcus sp., ClpP1 is essential for adaptation and growth at 25°C (64). The growth kinetics suggests that ClpP is even more important for growth at low temperatures than at higher temperatures. Importantly, during the first hours of growth, there was no significant difference in the doubling time of mutant and wild-type cells, but growth stops in the logarithmic growth phase (Fig. 1A). This is probably due to an accumulation of misfolded and aggregated proteins that prevents further expression of functional proteins. Notably, the morphology of colonies of the ΔclpP strain showed a reduction in size (∼0.8-fold) compared to the wild-type strain 8325 (Fig. 2). After prolonged incubation at 37°C, the wild type became slightly yellowish while the mutant remained white (data not shown). A slightly different cell surface of the ΔclpP strain was observed by scanning electron microscopy, and it appeared to be more rough and irregular (Fig. 2). All effects in the deletion mutant could be restored by complementation (data not shown).

FIG. 1.

(A) Growth kinetics of S. aureus 8325 wild-type (♦), 8325ΔclpP (▪), and 8325ΔclpP⁺ (▴) strains grown at 37°C, 30°C, 20°C, 42°C, and 45°C. The results are representative of three independent experiments. (B) Northern blot analysis of clpP transcription in S. aureus 8325 at various temperatures (left) and at various time points during the growth phase (at indicated OD₆₀₀ values) at 37°C.

FIG. 2.

Scanning electron microscopy of strain 8325 wild-type (A and B) and the isogenic ΔclpP mutant (C and D). Cells of the ΔclpP strain show a rougher and more irregular surface and decreased cell size than the wild-type strain. Preparation of samples was performed as described in Materials and Methods.

Global transcriptional profile of ΔclpP mutant of strain 8325.

There is increasing evidence that the Clp protease complex is involved in not only the degradation of misfolded proteins under stress conditions but also the regulation of protein expression and secretion (32, 36). In several bacterial pathogens, including S. aureus, virulence is strongly influenced by the activity of ClpP (26). To learn more about the regulatory role of ClpP in S. aureus, transcriptome analysis was performed by comparing exponentially growing (OD₆₀₀ of 1.0) ΔclpP mutant and parental strain 8325 using an S. aureus full genome chip. We decided to analyze gene expression at this time point because clpP transcription was maximal in the logarithmic growth phase (Fig. 1B). Moreover, it has been demonstrated by DNA microarray analysis that ∼97% of all genes are expressed at the end of the exponential growth phase (70). The experiments presented here revealed a reduced transcription of 227 ORFs in the ΔclpP strain, whereas transcription of 197 ORFs was increased. The expression of genes belonging to several regulons which play a role in virulence, oxidative stress, redox state, SOS response, metal homeostasis, and anaerobic growth were affected by the deletion of clpP. The expression data of the different categories are described and discussed in the following sections.

Virulence factor expression.

Expression of 46 virulence-associated genes was differentially regulated in the ΔclpP strain (Table 2). Genes that encode adhesins, including those encoding fibrinogen-binding proteins (clfA and clfB), the fibronectin binding proteins (fnbA and fnbB), and the elastin-binding protein (epbS), were induced in the mutant strain. Many exoenzymes were down-regulated, including alpha-toxin encoded by hla, V8 serine protease encoded by sppA, the serine proteases encoded by the spl operon, the metalloproteinase encoded by aur, a lipase precursor (encoded by lip), the cysteine proteinase staphopain (encoded by SA1725), a staphylococcal nuclease (encoded by nuc), and glycerol ester hydrolase (encoded by geh). Other virulence factors, such as those encoded by the cap operon (including 16 genes, capA-P) and an immunoglobulin G-binding protein (sbi), were down-regulated, whereas fmtB, isaB, and SA2447 (encoding a hypothetical protein, similar to streptococcal hemagglutinin protein) were up-regulated (Table 2). In addition, transcription of the ica operon (icaADBC), encoding products responsible for synthesis of the polysaccharide intercellular antigen, which is involved in biofilm formation of staphylococci, was strongly down-regulated in the mutant (Table 2).

TABLE 2.

Virulence-associated factors of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted function	Expression ratio of WT/ΔclpPa
Up-regulated factors
Adhesins
SA0742	clfA	Fibrinogen-binding protein A, clumping factor (LPXTG)	0.3
SA1268	ebhB	Hypothetical protein, similar to streptococcal adhesin emb	0.5
SA1312	ebpS	Elastin binding protein	0.4
SA2161		Hypothetical protein, attachment to host cells and virulence	0.4
SA2423	clfB	Clumping factor B (LPXTG)	0.4
SA2290	fnbB	Fibronectin-binding protein homolog (LPXTG)	Up
SA2291	fnbA	Fibronectin-binding protein homolog (LPXTG)	Up
Toxin, SA1811	hlb	Truncated beta-hemolysin	0.4
Other
SA0891		Hypothetical protein; similar to ferrichrome ABC transporter	0.5
SA1964	fmtB	FmtB protein (LPXTG)	0.4
SA1979		Hypothetical protein, similar to ferrichrome ABC transporter	0.5
SA2337	feoB	Ferrous iron transport protein B homolog	0.3
SA2356	isaA	Immunodominant antigen A	0.5
SA2431	isaB	Immunodominant antigen B	0.4
SA2447	hsa	Hypothetical protein, similar to streptococcal hemagglutinin protein (LPXTG)	0.3
Down-regulated factors
Adhesins
SA0587	mntC	Lipoprotein; streptococcal adhesin PsaA homologue	2.5
SA2459	icaA	Intercellular adhesion protein A	4.3
SA2460	icaD	Intercellular adhesion protein D	9.0
SA2461	icaB	Intercellular adhesion protein B	2.0
SA2462	icaC	Intercellular adhesion protein C	2.9
Toxins
SA1007	hla	Alpha-hemolysin precursor	3.8
SA1813		Hypothetical protein; similar to leukocidin chain lukM precursor	4.3
Exoenzymes
SA0022		Hypothetical protein; similar to 5′ nucleotidase (LPXTG)	3.2
SA0309	geh	Glycerol ester hydrolase	3.9
SA0746	nuc	Staphylococcal nuclease	5.0
SA0901	sspA	Serine protease; V8 protease; glutamyl endopeptidase	4.1
SA1628	splD	Serine protease SplD	3.6
SA1629	splC	Serine protease SplC	7.7
SA1630	splB	Serine protease SplB	3.6
SA1725		Staphopain, cysteine proteinase	3.4
SA2430	aur	Zinc metalloproteinase aureolysin	18.8
SA2463	lip	Triacylglycerol lipase precursor	3.7
Other
SA0144	capA	Capsular polysaccharide synthesis enzyme Cap5A	5.6
SA0145	capB	Capsular polysaccharide synthesis enzyme Cap5B	4.6
SA0146	capC	Capsular polysaccharide synthesis enzyme Cap8C	4.1
SA0147	capD	Capsular polysaccharide synthesis enzyme Cap5D	4.4
SA0148	capE	Capsular polysaccharide synthesis enzyme Cap8E	3.3
SA0149	capF	Capsular polysaccharide synthesis enzyme Cap5F	3.2
SA0150	capG	Capsular polysaccharide synthesis enzyme Cap5G	2.6
SA0151	capH	Capsular polysaccharide synthesis enzyme Cap5H	2.6
SA0152	capI	Capsular polysaccharide synthesis enzyme Cap5I	2.5
SA0153	capJ	Capsular polysaccharide synthesis enzyme Cap5J	2.1
SA0154	capK	Capsular polysaccharide synthesis enzyme Cap5K	2.6
SA0155	capL	Capsular polysaccharide synthesis enzyme Cap5L	2.5
SA0156	capM	Capsular polysaccharide synthesis enzyme Cap5M	2.7
SA0157	capN	Capsular polysaccharide synthesis enzyme Cap5N	2.1
SA0158	capO	Capsular polysaccharide synthesis enzyme Cap8O	3.0
SA0159	capP	Capsular polysaccharide synthesis enzyme Cap5P	2.1
SA0252	lrgA	Holin-like protein LrgA	11.2
SA0253	lrgB	Holin-like protein LrgB	12.7
SA0566		Hypothetical protein; similar to iron-binding protein	2.5
SA0841		Hypothetical protein; similar to cell surface protein Map-w	2.3
SA1709		Hypothetical protein; similar to ferritin	3.1
SA2206	sbi	Immunoglobulin G-binding protein SBI	4.3

^aRatio of gene expression of wild-type (WT) versus the ΔclpP mutant strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type. Up, increased transcription in the ΔclpP strain confirmed by RT-PCR.

Since many of the deregulated virulence factors are regulated by the global regulatory agr system (61), we investigated the expression of RNAIII, the effector molecule of the agr system, by Northern blot analysis. As shown in Fig. 3A, the RNAIII transcript levels were about threefold decreased in the ΔclpP strain, confirming the results by Frees et al. (26, 27). Thus, the observed changes in the expression of agr-regulated genes could be the direct result of down-regulation of RNAIII effector molecule levels. In addition, transcription of sarA, a global repressor of protease expression (13, 47), was up-regulated 2.5-fold in the ΔclpP strain. The strong repression of the metalloprotease aureolysin gene aur (18.8-fold) in particular might be due to the overexpression of sarA, as it has been shown that aur is most sensitive to repression by SarA (47). The mechanism of how ClpP regulates expression of agr and sarA remains unknown, especially if AgrA, AgrC, or SarA is a substrate of proteolytical cleavage by ClpP. Recently, Frees et al. (27) suggested that agr and ClpXP act epistatically on extracellular gene expression and that possibly Rot, a repressor of toxin expression, links the agr regulatory system with the ClpXP machinery, where Rot is targeted by ClpXP in the presence of accumulating RNAIII (27). However, further work has to be done to unravel the role of ClpP in the network of virulence factor regulation in S. aureus. Possibly, different regulators are substrates of ClpP-dependent proteolysis, as has been suggested for Rot (27). The microarray expression data of extracellular proteases could be corroborated by reduced proteolysis on milk-agar plates (data not shown). Moreover, we confirmed the down-regulation of hla expression which has been shown previously (26) by RT-PCR (Fig. 3B). Down-regulation of hla is most probably the result of low RNAIII expression. In addition, sarT, a repressor of hla, was up-regulated in the mutant, which might contribute to the decreased hla expression levels.

FIG. 3.

Transcriptional analysis of selected genes in 8325 wild-type (lane 1), ΔclpP (lane 2), and ΔclpP⁺ (lane 3) strains. RNA was isolated from exponentially growing cells (OD₆₀₀ of 1.0). (A) Northern blot analysis of RNAIII expression by hybridization with an RNAIII-specific probe. (B) Semiquantitative RT-PCR for transcriptional analysis of hla, fnbA, fnbB, and clfA. As a control, expression of 16S rRNA and clpP was determined.

Regulation of virulence factor expression in S. aureus is extremely complex, involving at least four two-component systems (agrAC, arlRS, saeRS, and srrAB), several transcription factors (encoded by sarA, sarS, sarT, sarR, and rot), and an alternative sigma factor (σ^B) (reviewed in reference 61). Here, we show that deletion of clpP exerted a strong impact on transcription of virulence-associated genes, many of which are under the control of global regulatory systems. The transcription of the agr system, arlRS, and sigB was down-regulated, whereas sarA and sarT were up-regulated in the ΔclpP strain. However, not all data of our study fit into the current concept of regulatory events leading to expression of a distinct virulence gene. For example, the arlRS system acts divergently to agr in the regulation of virulence determinants including hla, hlb, lip, and sspA, whose transcription is increased in an arlS mutant, as well as RNAIII transcription (22). In our study, arlRS expression was decreased by a factor of 3, and agrAC and RNAIII expression was also decreased by a factor of 2 to 4. This suggests that the impact of ArlRS on agr-regulated gene expression was superseded by other regulatory processes or that the level of ArlRS expression was still sufficient to depress RNAIII production. Alternatively, the reported regulatory impact of ArlRS on agr may reflect the fact that it was mainly investigated in strain 8325-4; however, Fournier and Klier (21) stated that in strain 8325, the strain used in this study, regulation might be different than in strain 8325-4 (21). Recently, Liang et al. investigated the Arl regulon by DNA microarray analysis (53). It was shown that ArlRS up-regulates the transcription of agrBDCA as well as hld located within the regulatory RNAIII in strain WCUH29. These results are in contrast to previous reports showing a repressive effect of ArlRS on agr RNAII and RNAIII expression (22). Further work has to be done to clarify the exact role of ArlRS on gene regulation in different genetic backgrounds. The strong impact of the clpP deletion on certain regulatory pathways of virulence factor expression clearly indicates a link between ClpP protease activity and regulation of virulence traits.

Internalization of the ΔclpP strain in 293T cells.

As several adhesins including fibronectin binding proteins A and B (encoded by fnbA and fnbB) were up-regulated in the ΔclpP strain (Fig. 3), we tested the ability of the ΔclpP strain to invade human epithelial cells. Interestingly, the rate of internalization by 293T cells increased about ∼10-fold compared to the parent strain (Fig. 4). The isogenic strains 8325ΔagrA and 8325ΔagrC, which were taken as controls, showed no significant differences in internalization rate, indicating an agr-independent mechanism responsible for increased internalization of the ΔclpP strain. The fibronectin-binding proteins FnbA and FnbB serve as the main surface proteins of S. aureus that mediate adherence to host cells by binding of fibronectin, which interacts with β₁α₅-integrins on the surface of host cells. In turn, β₁α₅-integrin clustering triggers the uptake of S. aureus by a zipper-like mechanism (1, 71). Gene expression data of both fnbA and fnbB were excluded from microarray analysis due to differences in homology between N315 and 8325 DNA sequences. Thus, the expression of these genes was analyzed by RT-PCR, revealing an induction of expression of fnbA and fnbB by threefold compared to the wild type (Fig. 3). Furthermore, fibronectin binding capacity was analyzed. The ΔclpP mutant showed a 2.3-fold increased capability to bind fibronectin in comparison to the wild type (data not shown). These results suggest that at least one reason for the increased internalization rate could be the overexpression of FnbA and FnbB. Recently, Frees et al. (24) investigated the intracellular replication of a clpP mutant of S. aureus strain 8325-4 in MAC-T cells, a bovine mammary epithelial cell line. ΔclpP cells were not able to replicate intracellularly, as indicated by bioluminescence (25). In contrast to that study, where the internalization rate was not affected by the clpP deletion, we clearly observed a significant increase in the internalization rate of ΔclpP cells compared to the wild type. Since we used the human kidney cell line 293T and our strain background was 8325, it has to be clarified whether the observed differences are due to the S. aureus strain background or the host cell line.

FIG. 4.

Internalization of ΔclpP mutant cells was increased in 293T cells. Relative internalization of different isogenic mutants of strain 8325 (ΔagrA, ΔagrC, ΔclpP, and complemented ΔclpP⁺ strains) is compared to internalization of 8325 wild type (Wt). Means ± standard deviations of four experiments are given.

Autolysis.

Expression of regulators of murein hydrolases (encoded by lytSR, lrgAB, arlSR, and rat) was mostly decreased in the ΔclpP strain, while transcription of lytM was increased (Tables 2 to 4). To determine the effect of clpP deletion on autolysis, an assay was performed treating cells with Triton X-100. The ΔclpP strain showed a strong induction of autolysis starting after 30 min of growth compared to wild-type and ΔclpP⁺ strains, confirming the microarray data (Fig. 5). The two-component system lytSR is involved in regulation of peptidoglycan hydrolases. In S. aureus a lytS mutant showed increased autolysis, altered levels of hydrolase activity, and a rough cell surface (10). lrgA and lrgB are positively regulated by lytSR, and their products show similarities to a bacteriophage murein hydrolase transporter family of proteins known as holins, which negatively affect peptidoglycan hydrolases (10). Interestingly, as mentioned above, expression of arlSR, encoding a two-component system (TCS) involved in autolysis, was also reduced in the clpP mutant (20, 53). Recently, DNA microarray analysis revealed a down-regulation of lytSR and lrgAB by ArlRS (53). Thus, it is tempting to speculate that decreased arlRS expression in the mutant contributes to the enhanced autolysis in the ΔclpP strain. In addition, the transcriptional regulator rat is described to be a repressor of autolysis and belongs to the MarR and SarA protein families (43). This type of repressor was also down-regulated in the clpP mutant. Altogether, the transcriptional profile of genes involved in autolysis may reflect the strong influence of ClpP protease activity on the regulation of autolysis in S. aureus.

TABLE 4.

Putative YycFG-controlled genes of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted functiona	Expression ratio of WT/ΔclpPb	Position (orientation)c	Putative YycF-binding sequenced
Up-regulated
SA0265e	lytM	Peptidoglycan hydrolase	0.2	−142 (+)	TGTAATGACAATGTAAT
SA0674e		Putative anion-binding protein	0.5	−16 (+)	TGTAATCAAATTGTAAT
SA1221e		Thioredoxin reductase	0.3	−113 (−)	TGTTAAGAAAATGTAAA
SA1305e	hu	DNA-binding protein II	0.4	−58 (+)	TGTAATGCTTGTGTTAA
SA1312e	ebpS	Elastin binding protein	0.4	−22 (+)	TGTAAAATCATTGTAAT
SA1898e		HP; similar to SceD precursor	0.5	−113 (+)	TGTAATCACTGTGTAAA
SA2093e	ssaA	Secretory antigen precursor SsaA homolog	0.2	−266 (−)	TGTTACAAATTTGTAAT
				−138 (−)	TGTTAACGTTTTGTAAT
SA2097e		HP; similar to secretory antigen precursor SsaA	0.4	−123 (−)	TGTTATTGATTTGTAAA
SA2285	aap	HP; similar to accumulation-associated protein	0.3	−34 (+)	TGTAAATTCACTGTAAG
SA2290	fnbB	Fibronectin-binding protein homolog	Up	−121 (−)	TGTTAACTTTATGTATA
SA2353e		HP; similar to secretory antigen precursor SsaA	0.4	−166 (−)	TGTTATCATAATGTAAT
SA2356e	isaA	Immunodominant antigen A	0.5	−140 (+)	TGTAAAGAAAGTGTAAT
SA2447-SA2440f	hsa	HP; similar to streptococcal hemagglutinin protein	0.3-0.5	−388 (−)	TGTAATATATGTGTAAT
SA2481e		Predicted sulfur transferase	0.3	−24 (+)	TGTTATAAGCATGTTAA
Down-regulated
SA0129e	sasD	HP	4.5	−16 (+)	TGTAATCAAATTGTAAT
SA0616-SA0617f	vraF	ABC transporter ATP-binding protein	2.1-2.4	−65 (+)	TGTTAGTCATATGTTAA
SA0682e-SA0681f		Putative di-tripeptide ABC transporter	2.4-3.1	−246 (−)	TGTTATTTTAATGTAAC
SA0913e-SA0910f	qoxA	Putative quinol oxidase polypeptide II QoxA	2.2-2.6	−53 (+)	TGTAAATATTGTGTAAT
SA1945e-SA1944f		Mannose-6 phosphate isomerase Pmi homolog	3.0-3.3	−179 (+)	TGTTAAAGTACTGTAAA
YycF consensus sequencee					TGTWAHNNNNNTGTWAH

^aHP, hypothetical protein.

^bRatio of gene expression of wild-type (WT) versus the ΔclpP strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type. Up, increased transcription in the ΔclpP mutant confirmed by RT-PCR.

^cPosition of the putative YycF-binding sequence relative to the translational start site in base pairs. Orientation (+/−) is given relative to the transcription of the respective gene.

^dBoldface, 100% conserved residues; italics, nucleotides conserved in more than half of all sequences.

^ePutative YycF-binding sequence (16). lytM, ssaA, and isaA were experimentally confirmed.

^fFirst and last ORF of putative transcription unit.

FIG. 5.

Autolysis of whole cells of S. aureus 8325 wild-type (♦), 8325ΔclpP (▪), and complemented mutant ΔclpP⁺ (▴) strains by Triton X-100. The results are expressed as lysis percentages as described in Materials and Methods. The average of two independent experiments is shown.

Heat shock regulation.

The loss of ClpP leads to accumulation of misfolded proteins similar to stress conditions, resulting in an increased demand for chaperones and proteases which are typically induced under heat shock conditions. The transcription of CtsR and HrcA, the main regulators of the heat shock response, is completely derepressed, as are the genes of the corresponding heat shock regulon (Table 3) as described for clpP mutants of S. aureus and S. pneumoniae (25, 66). In S. aureus the HrcA regulon (hrcA-grpE-dnaK-dnaJ and groESL) is embedded within the CtsR regulon (ctsR-mcsA-mcsB-clpC, clpB, and the HrcA regulon) (11). Hence, its derepression could be the result of inactivation of the CtsR repressor. In B. subtilis HrcA requires GroE to adopt its active conformation. Decreased levels of free GroE by association with misfolded proteins under heat shock conditions lead to inactivation of HrcA and a derepression of transcription of the HrcA regulon (59). The activity of the repressor CtsR is modulated by McsA and McsB and results in targeted degradation of CtsR by ClpCP in response to several stresses (14, 50). McsA contains a CXXC motif which might serve as a sensor of oxidative conditions. In B. subtilis elevated temperatures and oxidative stress conditions (H₂O₂, paraquat, NO, and diamide) give rise to an inactivation of CtsR and a derepression of transcription of corresponding genes (3, 52, 60). In S. aureus CtsR accumulates in cells lacking ClpP due to limited degradation by the Clp proteolytic machinery (25). As transcription of heat shock genes controlled by CtsR was induced in the clpP mutant, this would imply that CtsR accumulates in an inactive conformation in the ΔclpP strain and is not able to bind to the promoter region of those genes.

TABLE 3.

Genes encoding putative regulators of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted functiona	Putative transcription unitb (5′→3′)	Expression ratio of WT/ΔclpPc
Up-regulated
SA0017	yycF (vicR)	Two-component response regulator	yycF (0.5)-yycG (0.4)	0.5
SA0298		HP; similar to regulatory protein PfoR		0.5
SA0480	ctsR	Repressor of class III stress genes	ctsR (0.3)-SA0481 (0.1)-SA0482 (0.1)-clpC (0.1)	0.3
SA0573	sarA	Staphylococcal accessory regulator A		0.4
SA1041	pyrR	Pyrimidine operon repressor chain A	pyrR (0.5)-pyrP (0.4)-pyrB (0.4)-pyrC (0.4)-pyrAA (0.4)-pyrAB (0.5)-pyrF (0.5)-pyrE (0.4)	0.5
SA1098	codY	Transcription pleiotropic repressor CodY		0.5
SA1139	glpP	Glycerol uptake operon antiterminator		0.5
SA1174	lexA	SOS regulatory LexA protein		0.4
SA1411	hrcA	Heat-inducible transcriptional repressor	hrcA (0.4)-grpE (0.4)-dnaK (0.3)-dnaJ (0.3)	0.4
SA1897		HP; similar to transcriptional activator TenA	SA1897 (0.3)-thiD (0.4)-thiM (0.5)-thiE (0.4)	0.3
SA2286	sarT	SarA homologue		0.5
SA2320		HP; similar to regulatory protein PfoR	SA2320 (0.3)-SA2319 (0.4)-SA2318 (0.2)	0.3
SA2418		HP; similar to two-component RR	SA2418 (0.5)-SA2417 (0.4)	0.4
Down-regulated
SA0250	lytS	Two-component sensor HK	lytS (2.1)-lytR (2.1)	2.1
SA0322		HP; similar to transcription regulator, MarA family	SA0322 (3.5)-svrA (2.8)	3.5
SA0454	purR	pur operon repressor homologue		2.1
SA0641	rat	HP; similar to transcriptional regulator		3.8
SA1248	arlR	Two-component RR	arlR (3.6)-arlS (2.1)	3.6
SA1509		COG1327: predicted transcriptional regulator		2.4
SA1748		HP; similar to transcription regulator, GntR family	SA1748 (2.6)-SA1747 (ND*)-SA1746 (2.1)-SA1745 (2.3)-SA1744 (2.2)	2.6
SA1801		Antirepressor		4.3
SA1843	agrC	Accessory gene regulator C	agrB (ND*)-agrD (ND*)-agrC (2.4)-agrA (1.9)	2.4
SA1869	sigB	Sigma factor B	rsbU (4.2)-rsbV (ND*)-rsbW (4.5)-sigB (4.1)	4.1
SA2089	sarR	SarA homologue		2.3
SA2108		HP; similar to transcription regulator, RpiR family		2.0
SA2180	nreB	HP; similar to two-component HK	nreA (3.1)-nreB (3.2)-nreC (2.4)	3.2

^aHP, hypothetical protein.

^bValues in parentheses indicate relative expression levels of genes organized in one putative operon. ND*, ORF not represented on microarray used.

^cRatio of gene expression of wild-type (WT) versus the ΔclpP strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type.

Transcription of regulatory proteins was strongly affected by clpP deletion.

The genes of five TCSs were differentially expressed in the clpP mutant compared to the wild type: four were down-regulated, including lytSR, arlRS, agrAC, and a TCS with homology in sequence and orientation with nreBC of S. carnosus (Table 3). In contrast, the essential YycG/YycF TCS was up-regulated. Furthermore, the expression of 10 putative regulators was reduced, including those encoded by rat/mgr and sarR; an antirepressor encoded by SA1801; and two putative transcriptional regulators, encoded by SA0322 (MarA family) and SA1748 (GntR family). In addition, transcription of 10 transcriptional regulators was increased, including those encoded by ctsR, hrcA, sarA, sarT, codY, and lexA; a putative transcriptional regulator similar to TenA, encoded by SA1897; and a hypothetical protein similar to the regulator protein PfoR, encoded by SA2320. The genes of the sigB operon and the sigB-dependent asp23 were down-regulated. Notably, although strain 8325 is regarded as a functional sigB mutant due to an 11-bp deletion in rsbU, sigB transcription could be detected, suggesting a residual SigB activity in strain 8325. Likewise, Palma and Cheung (62) detected a reduced (by up to 50%) but still present expression of SigB-dependent genes in an rsbU mutant of the wild-type strain FDA486 (62). The observed strong influence of clpP deletion on transcription of regulators suggests that ClpP proteolytic activity may serve as an important mechanism to control gene expression in S. aureus. Therefore, a genome-wide in silico sequence analysis was performed using known consensus sequences of regulatory proteins including Fur (ferric uptake regulator), PerR, MntR, LexA, Fnr/ArcR, and YycFG to assess the impact of ClpP on expression of genes belonging to several regulons. This analysis revealed a strong impact of clpP deletion on the expression of genes that may be part of these regulons. However, it has to be stressed that the in silico recognition sequence search was solely based on known or putative consensus sequences and that for most members of specific regulons no experimentally confirmed data are available. For those genes for which regulator binding has been experimentally confirmed, this information was included in the analysis. The conclusion that ClpP might be involved in the regulation of the transcription of members of the Fur, PerR, MntR, LexA, Fnr/ArcR, and YycFG regulons was based on the observation that a significantly higher portion of ORFs with a putative recognition sequence of one of these regulators upstream of the translational start was deregulated (between 33 and 63%) than the overall percentage of deregulated genes (approximately 19% of all ORFs).

Impact of ClpP on expression of genes of the essential YycFG regulon.

The highly conserved YycF/YycG (VicR/VicK) TCS has been demonstrated to be essential in several gram-positive bacteria by regulation of cell wall biosynthesis and cell division (16, 41, 46, 57). In S. aureus a mutation in yycF results in a lethal phenotype at nonpermissive temperatures, and its essentiality has been proven by regulated expression of yycF using a conditional mutant system (16, 51, 57). In our experiments, deletion of clpP increased transcription of the yycFG locus. Regulation of yycFG transcription is presently unknown. Autoregulation can be ruled out as no YycF-specific recognition sequence can be found in the upstream region. Interestingly, we could identify a putative Crp/Fnr-like consensus sequence 62 bp upstream of the translation start of yycF (see below). The YycF-specific DNA-binding sequence consists of two repetitive hexamers: [TGT(A/T)A(A/T/C)-5N-TGT(A/T)A(A/T/C)] identified in B. subtilis and S. aureus (16, 41). In S. aureus N315 the consensus sequence could be found upstream of 31 ORFs (16). For three genes (lytM, ssa, and isaA) binding of the response regulator to the consensus was demonstrated recently (16). In the ΔclpP strain 16 putative members of the described yycFG regulon were deregulated, including the three experimentally confirmed genes lytM, ssa, and isaA (Table 4). In addition, we identified four additional putative yycFG-regulated genes: aap, hsa, fnbB, and vraF. Fourteen out of the 20 genes were up-regulated in the mutant including genes involved in cell wall synthesis (lytM, ssa, and two ssa homologous genes, SA2097 and SA2353) and virulence (isaA, ebpS, SA1898, fnbB, and hsa). Overall, 57% of all putative YycFG-regulated genes were deregulated in the ΔclpP strain.

ClpP controls metal ion homeostasis and oxidative stress proteins.

In B. subtilis and S. aureus, genes involved in iron and manganese homeostasis are regulated by three Fur homologous repressors, Fur, PerR, and Zur, and in addition by the DtxR homolog MntR. Fur is a transcriptional repressor controlling genes involved in iron uptake. Fur-regulated genes possess a so-called Fur box located upstream of the start codon. To find putative Fur-regulated genes, we used the postulated Fur box GATAATGATWATCATTATC for a consensus sequence search (40). We found that the expression of 6 out of 12 genes with a putative Fur box in the N315 genome was differentially regulated in the ΔclpP strain (Table 5). Interestingly, all of these genes including the Fur-dependent iron transporters feoB and feoB2 and a gene coding for a thioredoxin-homologous protein (SA2162) were up-regulated in the ΔclpP strain, indicating a lower Fur repressor activity in the clpP mutant.

TABLE 5.

Putative Fur-controlled genes of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted function	Expression ratio of WT/ΔclpPa	Position (orientation)b	Putative Fur boxc
SA0162	aldA	Aldehyde dehydrogenase homolog	0.3	−234 (−)	CTTGAGAATAATTCTCATTAAA
SA1982d-SA1980e	feoB2	Putative transporter	0.3-0.4	−24 (+)	AATGATAATGATTCTTATTATC
SA1979		Putative ferrichrome ABC transporter	0.4	−41 (−)	ATTGATAACAATTATCATTGTC
SA2001		Putative oxidoreductase, aldo/keto reductase family	0.5	−135 (+)	ATTGATAATTATGATAATCATA
SA2162d		Putative thioredoxin reductase	0.4	−92 (+)	ATTGATAATTATTATCATTTAA
SA2337d	feoB	Ferrous iron transport protein	0.3	−20 (+)	AGTGATAATGATTATTATTTCT
Fur consensus sequence (40)					NNNGATAATGATTATCATTATC

^aRatio of gene expression of wild-type versus the ΔclpP strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type.

^bPosition of the putative Fur box relative to the translational start site in base pairs. Orientation (+/−) is given relative to the transcription of the respective gene.

^cBoldface, 100% conserved residues; italics, nucleotides conserved in more than half of all sequences.

^dPredicted Fur box (40).

^eFirst and last ORF of putative transcription unit.

Furthermore, we analyzed the transcription of putative PerR-regulated genes. PerR controls as a Mn-dependent repressor a peroxide defense regulon. Members of this regulon, like catalase and peroxidases, detoxify reactive oxygen species (ROS); others, like ferritin or MrgA, store iron. Using an adapted consensus sequence postulated by Horsburgh et al. (39), we found 36 putative PerR-regulated genes in the genome of strain N315; 12 out of these 36 genes were deregulated in the ΔclpP strain (Table 6). Ten out of the 12 genes were up-regulated in the mutant, including the known PerR-controlled genes ahpCF, nfrA, and trxB. However, transcription of ftnA (ferritin) was decreased, suggesting incomplete derepression of the PerR regulon in the ΔclpP strain or other yet unknown regulatory mechanisms. In S. aureus and other bacteria, peroxide defense mechanisms and iron homeostasis are linked with manganese (Mn) transport that is controlled by MntR. MntR regulates as a Mn-dependent repressor the expression of two transport systems, mntABC and mntH (38). It has been proposed that mntABC represents the major Mn transport system in S. aureus that is regulated by several metal-depending repressors including PerR. Expression of mntABC has been shown to be induced at high Mn concentrations, while expression of mntH is repressed (38). Manganese acts in a dual way as an antioxidant and as a cofactor of enzymes like catalases, superoxide dismutases, and peroxidases. It is assumed that manganese protects S. aureus against ROS as a scavenger of either superoxide (O₂⁻) or hydrogen peroxide (H₂O₂) (38). Therefore, these bacteria possess a basal protection against ROS and are not required to activate the energy-dependent PerR defense regulon. Consistently, it was shown that Mn(II) acts as a repressor of PerR (38). However, under high oxidative stress conditions, the Mn-based defense mechanism becomes inadequate, giving rise to induction of the H₂O₂-sensitive PerR regulon. The deletion of clpP has a drastic effect on the expression of genes of the PerR, Fur, and MntR regulons, which strongly suggests that ClpP proteolytic activity is a key element in the defense against ROS under aerobic growth conditions. As mntABC (SA0587 to SA0589) expression was decreased in the clpP mutant, this would suggest that manganese transport is affected by the clpP deletion. Consequently, a decreased intracellular Mn level could contribute to oxidative stress conditions and derepression of the PerR regulon. The exact role of ClpP in coping with oxidative stress remains to be defined; however, the observed deregulation of oxidative stress-related regulons underlines the importance of functional ClpP activity for oxidative stress response. Importantly, Frees et al. (25) reported that the ClpP mutant in strain 8325-4 was more sensitive to hydrogen peroxide than the wild type (25).

TABLE 6.

Putative PerR-controlled genes of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted function	Putative transcription unit (5′→3′)	Expression ratio of WT/ΔclpPa	Position (orientation)b	Putative PerR boxc,e
SA0229		Conserved hypothetical protein	SA0230d-SA0229	0.4	−238 (−)	AATTAAATTATTATTTT
SA0298		Putative regulatory protein PfoR		0.5	−118 (−)	ATAATAATTATTATTAA
SA0366e	ahpC	Alkyl hydroperoxide reductase subunit F	ahpC-ahpF	0.3-0.4	−59 (+)	ATTAGAATTATTATAAT
SA0367e	nfrA	Putative nitro/flavin reductase		0.3	−93 (+)	AGTTCAATTATTAACTT
SA0719e	trxB	Thioredoxine reductase		0.4	−634 (+)	CATATAATTATTATTAT
SA0891		Putative ferrichrome ABC transporter		0.5	−390 (+)	AGATTAATTATTAAATA
SA0914	chiB	Putative chitinase B		2.8	−137 (−)	GAAATAATTATTATTTT
SA1268	ebhB	Similar to streptococcal adhesin	ebhB-ebhA	0.5	−252 (+)	TTTATAATTATTATAAA
SA1407		Conserved hypothetical protein		0.3	15 (+)	CTTTCAATTATTATTAA
SA1617		Similar to latent nuclear antigen	SA1617-SA1621 (SA1620d)	0.1-0.4	−191 (+)	TTTACAATTATTAAATT
SA1709	ftnA	Putative ferritin		3.1	−77 (+)	ATTATAATTATTATTAT
SA1897		Putative transcriptional activator TenA	SA1897-thiD-thiM-thiE	0.3-0.5	−261 (+)	TATAGAATTATTATTTA
PerR consensus sequence (39)						ATTATAATTATTATAAT

^aRatio of gene expression of wild-type (WT) versus the ΔclpP strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type.

^bPosition of the putative PerR recognition sequence relative to the translational start site in base pairs. Orientation (+/−) is given relative to the transcription of the respective gene.

^cBoldface, 100% conserved residues; italics, nucleotides conserved in more than half of all sequences.

^dORF not represented on the microarray.

^ePutative PerR box (39, 72).

Identification of putative LexA-regulated genes.

LexA regulates genes involved in repair of DNA damage. In E. coli LexA and the LexA homologous repressor HdiR have been recognized as substrates of ClpP-derived proteolysis (19, 69). Recently, a LexA-dependent regulation of fibronectin-binding protein B has been reported in S. aureus (6). In order to assess the impact of clpP deletion on the expression of putative LexA-regulated genes, a consensus sequence search using the B. subtilis recognition sequence (CGAACRNRYGTTCG) was performed (76). Without variation of the recognition motif, no putative LexA-regulated gene within the N315 genome could be identified. However, if we used a sequence adapted to GAAC-N₄-GTTC, we recognized 12 out of 20 putative LexA-dependent genes which were differentially regulated in the mutant (Table 7). Importantly, all of these 12 genes were up-regulated in the mutant, including both known LexA-regulated genes, recA and fnbB. Moreover, putative LexA-regulated genes like umuC, uvrA, and lexA itself were up-regulated in the ΔclpP strain. In addition, we found the LexA recognition motif upstream of two genes belonging to the PerR regulon (ahpC and the ferric ABC transporter SA0891 gene). The expression of the fibronectin-binding protein fnbB was determined by RT-PCR as the DNA microarray experiments did not allow a clear prediction (Fig. 3B). The expression data indicate a derepression of the LexA-regulated SOS-DNA repair regulon, which might be the consequence of increased DNA damage due to the reduced capability of the ΔclpP strain to cope with oxidative stress and to remove unfolded proteins.

TABLE 7.

Putative LexA-controlled genes of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted function	Expression ratio of WT/ΔclpPa	Position (orientation)b	Putative LexA-binding sequencec
SA0366	ahpC	Alkyl hydroperoxide reductase subunit F	0.3	−308 (+)	CGAACAAATATTCT
SA0714d	uvrA	Exinuclease ABC subunit A	0.4	−65 (+)	CGAAAGATTTAGAT
SA0891		Putative to ferrichrome ABC transporter	0.5	−354 (+)	TGAACAATTGTTGT
SA0993	uvrC	Excinuclease ABC subunit C	0.5	−79 (+)	CGAAGATGTTGATT
SA1128d	recA	RecA	0.4	−86 (+)	CGAACAAATATTCG
				−129 (−)	CGAACAAACGTGCT
SA1174d	lexA	SOS regulatory LexA protein	0.4	−58 (+)	CGAACAAATGTTTG
SA1180		Similar to exonuclease SbcD	0.5	−15 (+)	CGAACAAATGTTCT
SA1196d	umuC	Similar to DNA-damage repair protein	0.5	−35 (−)	CGAACACGTGTTCT
SA2090d	fnbB	Fibronectin-binding protein homolog	Up	−58 (+)	CGAACAATATAGAA
				−86 (−)	TGAAAAAAAGCGAG
SA2091d	fnb	Fibronectin-binding protein homolog	Up	−59 (+)	CGAACAATATAGAC
SA2375		Similar to dihydroorotate dehydrogenase	0.4	−223 (−)	TGAACAATGGTTAG
SA2473		Hypothetical protein	0.4	−205 (−)	TGAACGTTGGTTAC
LexA consensus sequenced					GAAC-N4-GTTC

^aRatio of gene expression of wild-type versus the ΔclpP strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type. Up, increased transcription in the ΔclpP mutant confirmed by RT-PCR.

^bPosition of the putative LexA recognition sequence relative to the translational start site in base pairs. Orientation (+/−) is given relative to the transcription of the respective gene.

^cBoldface, 100% conserved residues; italics, nucleotides conserved in more than half of all sequences.

^dPredicted LexA binding sequence (6, 76).

Anaerobic growth.

The physiological examination of the ΔclpP strain by using the API Staph test system showed that arginine deiminase activity was reduced (Fig. 6A). Arginine deiminase is encoded by arcA, which is located in an operon (arcABCDR) whose transcription is induced under anaerobic conditions and which is controlled by catabolite repression (15, 78). The arginine deiminase system is used by many prokaryotes to produce ATP under anaerobic conditions by catalyzing the conversion of arginine to ornithine, ammonia, and CO₂. Expression of the arc operon is controlled by regulatory proteins of the Crp/Fnr family (54). S. aureus and other gram-positive bacteria carry a gene coding for a Crp/Fnr homologous protein (arcR), located downstream of arcA in the N315 genome. A consensus sequence search using the ArcR recognition sequence of Bacillus licheniformis within all deregulated genes of the ΔclpP strain resulted in the identification of 19 genes that carry an arcR consensus sequence upstream of the transcriptional start site resembling the B. licheniformis ArcR binding site TGTGA-N₆-TCACG (55) (Table 8). Among these, 15 genes were down-regulated, and 10 of them are preferentially expressed under anaerobic conditions, including those encoding arginine deiminase (arcA), formate acetyltransferase (pflB), lactate dehydrogenase (lctE), nitrite extrusion protein (narK), and alcohol-acetaldehyde dehydrogenase (adhE). Overall, expression of almost two-thirds (19 out of 30) of all genes with a putative ArcR consensus sequence in front of the translational start were influenced by clpP deletion, suggesting a significant impact of ClpP on regulation of ArcR-dependent gene expression. Moreover, the transcription of other genes that are involved in anaerobic growth was affected in the ΔclpP strain. The TCS NreBC regulates anaerobic respiration in Staphylococcus carnosus by controlling transcription of the nitrate reductase operon (narGHIJ) and nitrite reductase (nir) (18). In S. aureus N315 we could identify a putative TCS with high homology to nreABC of S. carnosus (SA2181 to SA2179). In the ΔclpP strain nreABC as well as the nar and nir operons were down-regulated (see Table S1 in the supplemental material). Consequently, the ΔclpP strain showed a growth defect under anaerobic conditions on solid medium (data not shown). All these data indicate that ClpP is essential for growth and survival of S. aureus under anaerobic conditions, probably due to regulating the activity of the arginine deiminase pathway and, furthermore, nitrate and nitrite respiration.

FIG. 6.

Arginine deiminase (A) and urease (B) activity of 8325 wild-type (Wt), ΔclpP, and complemented mutant ΔclpP⁺ strains after 4 h of incubation (urease) or after 16 h of incubation (arginine deiminase). API Staph test was performed according to the manufacturer's instructions (BioMérieux). +++, ++, and − indicate very high, high, and no enzymatic activity, respectively.

TABLE 8.

Putative ArcR-controlled genes of S. aureus differentially expressed in the ΔclpP strain

N315 ORF	Gene name	Description or predicted functiona	Expression ratio in WT/ΔclpPb	Position (orientation)c	Putative ArcR-binding sequenced
Down-regulated
SA0143	adhE	Alcohol-acetaldehyde dehydrogenase	6.0	−22 (+)	TTGTGAAATAATTCACAA
SA0218-SA0219e	pflB	Formate acetyltransferase	7.8-15.7	−79 (+)	ATGTGAAAAAAATCACAA
SA0232	lctE	l-Lactate dehydrogenase	12.1	−210 (+)	ATGTGAAATAAATCACAA
SA0293	nirC	HP; similar to formate transporter NirC	4.4	−55 (−)	TTGTGAATAATTTCACAA
SA0295		HP; outer membrane protein precursor	5.7	−141 (+)	ATGTGATAGGTCTCCCAT
SA0562	adh1	Alcohol-dehydrogenase I	3.5	−289 (+)	TTGTGAATTAATTCACAT
SA0641	rat	Transcriptional regulator	3.8	−69 (+)	TTGTGAATTAATAAACAA
SA1272	ald	Alanine dehydrogenase	3.8	−21 (−)	TTGTGAATAATTTCACAA
SA1813	lukM	HP; similar to leukocidin chain lukM precursor	4.3	−49 (−)	ATGTGAATAATATCACAA
SA2156	lctP	l-Lactate permease lctP homolog	5.5	−107 (+)	TTGTGAAAAAAATCACAT
SA2176	narK	Nitrite extrusion protein	3.6	−151 (−)	TTGTGAAAAAGTGAACAT
SA2189-SA2188e	nirR	HP; similar to NirR	7.7-10.1	−47 (−)	TTGTGAAAAGAATCACAT
SA2268		HP	17.8	−112 (+)	TTGTGAAATACATCACAA
SA2428	arcA	Arginine deiminase	3.4	−62 (+)	ATGTGAATATAATCACAT
SA2430	aur	Zinc metalloproteinase aureolysin	18.8	−215 (−)	TTGTGAAAATATTAACAA
Up-regulated
SA0017-SA0018e	yycFG	Response regulator/histidine kinase	0.4-0.5	−62 (−)	TTGTGTAAAAAATCACAG
SA0175		Conserved HP	0.4	−42 (−)	TTGTGAAAATAATCACAA
SA2311		HP; similar to NAD(P)H-flavin oxidoreductase	0.5	−168 (+)	TTGTGAAAAATATCACAA
SA2373-SA2371e		HP	0.5	−95 (+)	TTTTGAATATAATCACAA
ArcR consensus sequencef					TGTGAA-N5-TCACA

^aHP, hypothetical protein.

^bRatio of gene expression of wild-type (WT) versus the ΔclpP strain. Values of ≥2 indicate decreased expression, and values of ≤0.5 indicate increased expression in the ΔclpP strain compared to the wild type.

^c Position of the putative ArcR-binding sequence relative to the translational start site in base pairs. Orientation (+/−) is given relative to the transcription of the respective gene.

^dBoldface, 100% conserved residues; italics, nucleotides conserved in more than half of all sequences.

^e First and last ORF of putative transcription unit.

^fIn B. licheniformis (TGTGA-N₆-TCACG) (55).

Urease activity.

It was striking that the urease activity test revealed a strong induction in the ΔclpP strain after 4 h of incubation, whereas the parental strain and the ΔclpP⁺ strain did not show any activity at this time point (Fig. 6B). This observation is clearly consistent with the microarray data showing an induction of the complete ure operon (SA2081; ureABCEFGD) in the ΔclpP strain by 5- to 10-fold at an OD₆₀₀ of 1.0 (see Table S1 in the supplemental material). These results were also confirmed by RT-PCR (data not shown). Recently, induction of the ure operon was described in S. aureus biofilms and in a rot mutant (5, 67). In the ΔclpP strain no alteration of transcription of rot could be detected. As urease catalyzes the hydrolysis of urea to form ammonia and CO₂, it has been suggested that a high urease activity may indicate attempts of bacteria to neutralize acidic environments. For example, urease activity of Helicobacter pylori is essential to colonize the acidic environment present in the stomach (17). Alternatively, the ure operon is induced in response to nitrogen starvation, e.g., in B. subtilis and Corynebacterium glutamicum (4, 7). Interestingly, carbamoyl phosphate synthetase transcription (pyrAA [SA1045] and pyrAB [SA1046]) was induced twofold in the mutant, but other enzymes of the urea cycle were not. Thus, increased levels of carbamoyl phosphate may be sufficient to generate a higher concentration of urea that is toxic for the cell and, consequently, has to be inactivated by urease. Further work has to be done to clarify the exact role of high urease activity for pH balance and/or nitrogen metabolism in ClpP-deficient cells.

Concluding remarks.

Global DNA expression analysis using DNA microarray technology revealed a broad impact of the S. aureus ClpP protease on several regulons involved in virulence, heat shock response, oxidative stress response, DNA repair, autolysis, and anaerobic growth. Targets of proteolytic ClpP activity in S. aureus are presently not known; however, the clustering of deregulated genes suggests that the expression of genes within specific regulons is controlled by ClpP-dependent proteolysis. In E. coli many proteins cleaved by ClpXP are involved in the oxidative stress response and a shift between aerobic and anaerobic growth. It has been suggested that ClpXP degrades proteins whose Fe-S clusters have been damaged by oxidation (19). Many results presented in this study are consistent with this idea. Possibly, ClpP plays a major role in the maintenance of reducing conditions within the cell by degradation of oxidized proteins. In consequence, oxidation-susceptible proteins like Spx may accumulate in the ΔclpP strain (77). An important challenge for the future will be to identify substrates of ClpP proteolytic activity in S. aureus and to clarify the role of functional ClpP for the infection process. In addition, ClpP may serve as an attractive new target for anti-infective agents. Interestingly, acyldepsipeptides, a new class of antibiotics that targets ClpP protease, has been recently identified (8). Surprisingly, the antimicrobial activity of the compound was not due to an inhibition of the target ClpP, but bacterial cells were killed by uncontrolled ClpP-dependent proteolysis. These observations impressively stress the importance of controlled ClpP-mediated proteolysis for protein homeostasis in bacterial cells.