Hfq Is a Global Regulator That Controls the Pathogenicity of Staphylococcus aureus

The Hfq protein is reported to be an RNA chaperone, which is involved in the stress response and the virulence of several pathogens. In E. coli, Hfq can mediate the interaction between some sRNAs and their target mRNAs. But it is controversial whether Hfq plays an important role in S. aureus. In this study, we found that the deletion of hfq gene in S. aureus 8325-4 can increase the surface carotenoid pigments. The hfq mutant was more resistant to oxidative stress but the pathogenicity of the mutant was reduced. We reveal that the Hfq protein can be detected only in some S. aureus strains. Using microarray and qRT-PCR, we identified 116 genes in the hfq mutant which had differential expression from the wild type, most of which are related to the phenotype and virulence of S. aureus. Among the 116 genes, 49 mRNAs can specifically bind Hfq protein, which indicates that Hfq also acts as an RNA binding protein in S. aureus. Our data suggest that Hfq protein of S. aureus is a multifunctional regulator involved in stress and virulence.


Introduction
Hfq protein was originally identified in E. coli as a host factor for the replication of the Qb phage RNA [1] and is considered to be a multifunctional regulator of a variety of targets in bacteria [2]. Hfq regulates post-transcriptional gene expression in E. coli, specifically mRNA translation, stability or polyadenylation [3,4].
Recently most studies have focused on the RNA chaperone function of Hfq in mediating the interaction between sRNA(small RNA) and mRNA [5]. In E. coli, some known sRNAs can interact with Hfq [6]. In Salmonella, Hfq associates with almost half of the co-immunoprecipitated sRNAs [7]. It is reported that Hfq can influence the stability of several sRNAs [8][9][10]. Besides acting as an RNA chaperone, E. coli Hfq has ATPase activity and can mildly affect transcription and translation in vitro [11,12].
The inactivation of Hfq in E. coli can affect pleiotropic phenotypes, such as decreased growth rates, negative supercoiling of plasmids in stationary phase, increased carbon source oxidation, and increased sensitivity to ultraviolet light [13]. Hfq is also involved in stress response or virulence of a variety of pathogenic bacteria, such as Neisseria meningitides, Salmonella, Vibrio cholerae and Pseudomonas aeruginosa [7,[14][15][16][17][18][19].
Staphylococcus aureus, a major Gram-positive pathogen, can cause a broad range of diseases, from minor skin infections to lifethreatening diseases, such as toxic shock syndrome and septicemia [20]. Many reports identify that RNAIII can act as antisense RNA and regulate the expression of several virulence genes, such as spa (Staphylococcal protein A, an important virulence factor of S. aureus) and hla (alpha-toxin gene, a major virulence factor of S. aureus) [21]. There is also an Hfq-like protein (8.9kD) in S. aureus, which can form homo-hexamer [22]. The structure of Hfqbinding RNA has been identified [3,22]. Under in vitro condition, it is found that Hfq can specifically bind RNAIII and spa mRNA, which suggests that S. aureus Hfq can also act as an RNA chaperone [23]. However, Bohn et al. showed that Hfq did not play a crucial role in stress response, spa mRNA expression, or exoprotein expression in S. aureus RN6390, COL and Newman. They also tested ,2000 phenotypes in the RN6390 hfq mutant by the Phenotype Microarray (PM) Technology [24], and found Hfq was not involved in the stress response, resistance to several chemical agents, or metabolic pathways [24]. It was also reported that S. aureus Hfq could not affect SA1000 expression, which is a fibrinogen-binding protein and identified as a target of RNAIII [25]. Geisinger et al. suggested that the weak transcription of hfq was possibly due to the loss of the gene's major promoter [26].
In this study, we found the carotenoid pigment increased in the hfq mutant of S. aureus 8325-4 (Dhfq-8325) and the mutant was less toxic during infection of MDBK cell and a mouse model of peritonitis than its parental strains. The results of the microarray studies showed that the expression of the 116 genes was altered in Dhfq-8325. We also found that 48 mRNAs of the 116 genes can specifically bind Hfq, which suggests that Hfq may be an RNA binding protein. We believe Hfq is a multifunctional regulator of gene expression and plays a major role in the infection of S. aureus. Our conclusion is contrary to the finding by Bohn et al. The difference of these findings may be due to differing expression patterns of the Hfq protein in different S. aureus strains.

Deletion of Hfq in S. aureus 8325-4 increased the carotenoid pigments
We constructed the hfq gene deletion mutant (Dhfq-8325) derived from S. aureus 8325-4. The result showed that Dhfq-8325 displayed high intensities of yellow pigmentation compared to the wild type strain. After the restoration of Hfq activity in Dhfq-8325, the color of the restored strain (rs-hfq-8325) was recovered ( Figure 1A, 1B).The intensity of the yellow carotenoid pigment of S. aureus is one of the classical criteria for identification of this species. The genes of crtM and crtN, which encode dehydrosqualene synthase and dehydrosqualene desaturase respectively, are essential for the pigment's synthesis of S. aureus. Both genes are located in the same operon [27]. We quantified the mRNA level of crtM in 8325-4, Dhfq-8325 and rs-hfq-8325 by qRT-PCR, and showed that the mRNA level of crtM increased in Dhfq-8325 ( Figure 1C). This may explain why the carotenoid pigment of Dhfq-8325 was increased.
It was reported that the pigment acted as an antioxidant for S. aureus' resistance to neutrophil killing [28]. We tested the susceptibility of Dhfq-8325 to oxidant, and showed that Dhfq-8325 was indeed more resistant to oxidant than the wild type (Figure 2A, 2B). In addition, we found that the number of the mutant bacterium survived in neutrophils was significantly larger than that of the wild type ( Figure 2C). We also found that Dhfq-8325 didn't display other phenotypic variations including biofilm formation, cell growth, surface charge and hydrophobicity (data not shown).
The bacterial strains used in the work of Bohn et al. were RN6390, COL and Newman, which were different from the ones in our work [24]. Although the mRNA of hfq could be detected in such strains by RT-PCR, we wondered whether this mRNA could be translated into the protein. So we extracted the same amount of total cellular proteins from 8325-4, RN6390 and other 11 strains, and detected the expression of Hfq by Western blot with the specific anti-Hfq antibodies prepared in our lab. The results revealed that Hfq protein could be found in 8325-4 and other 7 strains, but not in COL and RN6390 ( Figure 3). The lack of Hfq may explain the non-detectable effect of Hfq in S. aureus COL and RN6390 in the work of Bohn et al. [24]. The sequences of all the genes encoding Hfq protein including their promoter regions in these strains are confirmed to be the same (data not shown). We also compared the transcription of hfq in 8325-4 and RN6390 by qRT-PCR using 16S rRNA as the control. No significant difference was observed (data not shown). This result suggested that there might be some post-transcription regulation of Hfq protein. Recently Becher et al. reported that mRNA of hfq (SACOL1324) could be detected in COL by DNA microarray, but the Hfq protein was undetectable [29]. This result is consistent with our observation. The different expression profiles of Hfq in S. aureus strains needs further investigation.
Dhfq-8325 was susceptive to the whole blood killing in spite of its increased golden pigments The work of Liu et al. showed that the number of DcrtM-8325 survived in the whole blood decreased compared with WT because of reduced pigment synthesis [28]. We performed a similar experiment to verify whether Dhfq-8325 was more resistant to the whole blood than wild type. It was surprising that the number of surviving Dhfq-8325 cells was significantly less than the wild type ( Figure 4). These results suggest that some other regulators in Dhfq-8325 might be more important than the pigments for S. aureus survival in the whole blood. The deletion of hfq decreased the toxicity of S. aureus It is well known that the secreted toxins are important for infection caused by S. aureus. We compared the cytotoxicity of the hfq mutant with that of the wild type. After incubation with the supernatants from different strains, the survival cell numbers were quantified by CCK8 assay. We found that the toxicity of the supernatant of Dhfq-8325 significantly decreased compared to its parental stain ( Figure 5). However, the cytotoxicity of the supernatant of RN6390 and COL did not alter after the hfq gene deletion ( Figure 5). In order to investigate if Hfq protein could influence the toxicity in other strains besides 8325-4, we constructed an hfq-deletion mutant from ATCC25923, which was an Hfq positive strain ( Figure 3). Similarly, the survival percentage of cells after treatment of the supernatant of ATCC25923 was lower than that of the hfq mutant ( Figure 5). The apoptosis and necrosis of MDBK cells induced by the supernatant was tested using flow cytometry. We found that the percentage of apoptosis and necrosis induced by the supernatant of Dhfq-8325 was significantly lower than that of the wild type ( Figure 6). In our further study, the pathogenicity of Dhfq-8325 was assessed in a murine peritonitis model, with an equal number of bacterium from the wild type and Dhfq-8325 injected  intraperitoneally into the mice. The result showed that at the different time points (6 h, 10 h, 20 h, 24 h and 30 h), the survival number of the Dhfq-8325 group was larger than that of the wild type group( Figure 7A), which was consistent with the results of the whole blood survival assay and the cytotoxicity assay. The pathogenicity of Dhfq-25923 was also assessed in the same murine peritonitis model, and the result showed that the survival number of the Dhfq-25923 group was larger than that of ATCC25923 group ( Figure 7B).
Identification of the changed genes in ?hfq-8325 using microarray assay The phenotype and pathogenicity of Dhfq-8325 were different from that of the wild type, so we analyzed the global change of gene expression in Dhfq-8325 by microarray. The results revealed that the expression levels of 116 genes were altered in Dhfq-8325 (fold change$1.5), most of which were involved in the stress response and pathogenicity of S. aureus. Among the 116 genes, 33 had declined expression level (Table S1), while 83 had elevated expression level, compared to the wild type (Table S2). It was found that the level of crtM in Dhfq-8325 was increased, which was consistent with the result of qRT-PCR. The expression of some virulence genes (such as serine protease, cysteine protease and staphylococcal nuclease) and global regulators (such as sarA) was changed after hfq deletion, which could affect the pathogenicity of Dhfq-8325. In order to validate the microarray data, the level of four genes (two up-regulated genes and two down-regulated in Dhfq-8325) was tested. The results of qRT-PCR were consistent with the microarray data ( Figure 8).
Although it remained controversial whether Hfq plays an important role in S. aureus, it has been confirmed that Hfq protein could specifically bind RNAIII [23]. We obtained the RNAs that could potentially bind Hfq protein by immunoprecipitation in vitro and these RNAs were further analyzed by microarray. We identified about 300 mRNAs which were enriched in the pool of Hfq IP, including RNAIII (Table S3). Then we compared these RNAs with the 116 genes identified in Dhfq-8325 (Table S1, S2), and we found that 49 of the 116 genes in Dhfq-8325 were enriched in the pool of RNAs that bind Hfq protein. It is likely these genes may be regulated by Hfq.

Discussion
Hfq was discovered as an E. coli host factor required for replication of RNA phage Qb [1]. Hfq protein belongs to the large family of Sm and Sm like proteins. The sequence alignment shows that E. coli Hfq protein and S. aureus Hfq have the conserved Sm1 sequence motif, and the resolved structures show both of them can form homo-hexameric stuctures [3]. More recent studies show that Hfq protein can act as an RNA chaperone to mediate RNA-RNA interaction in E. coli [10]. But it is controversial whether Hfq plays an important role in S. aureus. Many reports reveal that S. aureus RNAIII acts as an antisense RNA to regulate its target genes expression [23,25,26], and it is a potential target of S. aureus Hfq in vivo [23]. However, Boisset et al. found that Hfq protein exerted no  major effect in vivo on SA1000, SA2353, rot mRNA levels or RNAIII stability in S. aureus RN6390 [25]. Furthermore Bohn et al.
reported that the deletion of Hfq had no noticeable effect on virulence gene expression nor was this protein highly expressed in S. aureus RN6390 [24].
In our work, we found that the deletion of Hfq can increase the pigments synthesis in 8325-4, while the color of the bacteria was recovered after the hfq gene was restored in Dhfq-8325. These results suggest that Hfq can influence the phenotype of S. aureus. The reason why our conclusion is different from that of Bohn et al. may be due to the different strains used in the experiments. Although S. aureus 8325-4 and RN6390 both originate from S. aureus 8325, the RN6390 has been generated by three rounds of transduction from 8325-4, and they have been validated to be different strains [30][31][32]. We also found that the Hfq protein can be detected in 8325-4 and other strains, but not in RN6390. The lack of Hfq expression may explain that there were no detectable changes induced by inactivation of Hfq in some strains. The reason why Hfq can not be detected in some strains is not clear and is now being investigated in our lab.
Although the pigments increased in Dhfq-8325, we found the survival number of Dhfq-8325 in the whole blood decreased significantly and the Dhfq-8325 was less toxic to mammalian cells and mice. These results suggest that the pigments are just one factor of S. aureus pathogenicity and there may be some other more crucial factors regulated by Hfq. We then compared the gene expression profile of Dhfq-8325 with that of wild type. The transcription analysis showed that the expression of many genes was altered. Several serine protease and cysteine proteases (such as sspA, sspB and splC) were down-regulated in Dhfq-8325 and identified as the important factors for S. aureus infection. A transposon mutant of the ssp operon had an attenuated virulence in three separate animal models (mouse abscess, bacteremia and wound infection) [33]. This may explain why the pathogenesis of Dhfq-8325 was decreased in the animal model. Some recent researches have suggested that the studies defining the regulatory roles of sarA and agr using RN6390 are not always representative of the events that occur in clinical isolates of S. aureus [34,35]. So we constructed an hfq deletion mutant from ATCC25923, which is a clinical isolate, to investigate the  biological function of Hfq. Our results showed that the cytotoxiticy and pathogenesis of the Dhfq-25923 was significantly decreased as compared to its parental strain which suggests that Hfq can influence the toxicity of lab strains and clinical isolates.
The RNA chaperone activity of S. aureus Hfq has not been confirmed [23][24][25]. Although Hfq protein of E. coli has a longer C terminal-tail than that of S. aureus, the tail appears to be dispensable in the RNA binding function of Hfq [3]. Here we obtained the RNAs that bind Hfq in S. aureus 8325-4 by immunoprecipitation. The binding RNAs were then identified using microarray technology. We found that about 40% of the mRNAs of the changed genes in Dhfq-8325 could specifically bind Hfq protein. The results suggest that the Hfq protein in S. aureus can also act as an RNA binding protein such as in E. coli. We are investigating the mechanism how these genes can be regulated by Hfq directly or indirectly.

Ethics Statement
All animal experimental protocols of the study are in accordance with the national guidelines for the use of animals in scientific research. It's also approved by Animal Care and Use Committee of Beijing Institute of Basic Medical Sciences, with the approval number BMS-081210.

Bacterial strains and growth conditions
The strains used in this study are listed in Table 1. Strains were grown in 5 ml of brain heart infusion (BHI, BD) or Luria-Bertani (LB) medium at 37uC for 12 h with shaking at 200 rpm in a 25-ml test tube. Cells from 1 ml of pre-culture were transferred to 100 ml of BHI or LB medium in a 500-ml flask and incubated at 37uC on a rotary shaker at 200 rpm. S. aureus strains were routinely grown in BHI and E. coli strains were grown in LB medium either with no antibiotics, or with 20 mg/ml erythromycin, 100 mg/ml ampicillin and 50 mg/ml kanamycin.

Expression and purification of Hfq protein
The hfq gene was amplified from the genome of S. aureus 8325-4 with the primers (hfq NcoI and hfq XhoI) in Table 2. The PCR product was cloned into the expression vector pET28(a) as an NcoI-XhoI fragment, generating pET28(a)-hfq, which was subsequently transformed into E. coli strain BL21 (DE3). Expression of the recombinant Hfq protein was induced by adding 0.1 mM isopropyl b-D-thiogalactopyranoside (IPTG) to a growing culture at OD 600 = 0.4. Purification of recombinant protein on Q Sepharose Fast Flow and SP Sepharose Fast Flow columns (Amersham Biosciences) was carried out according to the manufacturer's instructions.

Production of polyclonal antibodies against Hfq
Female New Zealand White rabbits were first immunized by subcutaneous injection with 1 mg of Hfq protein in Phosphate Buffered Saline (PBS) mixed with complete Freund's adjuvant in a total volume of 1 ml. And the subsequent booster injections, i.e., 1 mg of Hfq in PBS emulsified in incomplete Freund's adjuvant, were administered at 3 and 6 weeks after the primary immunization. At the eighth week, the sera were collected and the antibody titers were determined by ELISA.

Construction of hfq deletion mutant
The mutant was constructed using the method described previously [36] with some modifications. In order to create a deletion mutant of hfq in the chromosome of 8325-4, two regions of DNA flanking the hfq gene were amplified by PCR using the primers (Up-hfq F-EcoRI and Up-hfq R-KpnI; Down-hfq F-KpnI and Down-hfq R-SalI) with restriction sites listed in Table 2. The upstream fragment (659 bp) was digested with EcoRI and KpnI, and the downstream fragment (680 bp) was digested with KpnI and SalI. The two fragments were cloned together into pMD19T digested with EcoRI and SalI. The resulting construct was digested with KpnI, and then a 1.6-kb kanamycin cassette which was amplified from the plasmid of pTZ-TRAP::kan provided by Dr. Balaban N was inserted. The resulting plasmid was digested with EcoRI and SalI, and a fragment harboring kanamycin resistance between the upstream and downstream fragments was ligated into pAUL-A digested with EcoRI and SalI to create plasmid pAUL-A-Dhfq. pAUL-A has a temperature-sensitive origin of replication that is active in S. aureus at 30uC but not at 42uC. The recombinant plasmid initially isolated from E. coli, was introduced into S. aureus RN4220 by electroporation and colonies resistant to kanamycin and erythromycin were selected after growth at 30uC. The resistant clones were subjected to a temperature shift to 42uC to select the plasmid integration into the chromosome. Bacteria resistant to kanamycin but sensitive to erythromycin were selected. The mutation was confirmed by PCR, and followed by transduction into strains 8325-4, RN6390, COL and ATCC25923 with phageW11 to create the mutant strains from which the coding region of hfq (234 bp) was deleted (Table 1).

Restoring Hfq Activity
Primers rs-hfqF and rs-hfqR (listed in Table 2) were designed to PCR-amplify a 1406 bp fragment encompassing the gene encoding for Hfq, the promoter region and termination site with SalI/EcoRI sites. S. aureus 8325-4 chromosomal DNA was used as a template. The PCR product (''whole hfq'') was digested and cloned into the SalI/EcoRI digested pMAD, which could replicate in E. coli at 37uC and in S. aureus at 30uC and could be used as a suicide vector when grown at 42uC in S. aureus. The resulting plasmid (pMAD-rs-hfq) was used to transform into E. coli DH5a. Cells were selected on LB plates containing 100 mg/ml ampicillin. Plasmid was isolated from positive clones and used to transform S. aureus RN4220. The transformants were selected on tryptic soy agar plates containing 10 mg/ml erythromycin at 30uC. Then the plasmid was isolated from the positive clone (RN4220 containing pMAD-rs-hfq) and transformed to Dhfq-8325 by electroporation. The transformants were selected on tryptic soy agar plates containing 10 mg/ml erythromycin at 30uC for 48 h, and then transferred to 42uC for 12 h. The restoring colonies (rs-hfq-8325) were confirmed by PCR and RT-PCR analysis.

Oxidant susceptibility assay
The assay to test susceptibility to oxidants was performed as previously described [37] with some modifications. For hydrogen peroxide susceptibility assay, S. aureus cells were grown to early exponential phase and harvested by centrifugation. Hydrogen peroxide (H 2 O 2 ) was added at a 1.5% final concentration to the bacteria (2610 9 /ml) and incubated at 37uC for 1 h. Then, 1,000 U/ml of catalase (Sigma-Aldrich) was added to quench residual H 2 O 2 . Dilutions were plated on tryptic soy agar plates for enumeration of surviving cells.
The assay to test susceptibility to singlet oxygen was performed as previously described [37] with some modifications as well. S. aureus strains were grown to early exponential phase and harvested by centrifugation. The bacteria (10 8 /ml) were incubated at 37uC in a 24-well culture plate. 200 ml of 0.1 mg/ml methylene blue was added to the wells and the plate was situated at exactly 10 cm from a 100-W light source. The viable cells were assessed after 1.5 h by plating dilutions on tryptic soy agar plates.

Neutrophil intracellular survival assay
Neutrophils for intracellular survival assays were purified from healthy human volunteers as previously described [37]. Bacteria were washed twice in PBS, diluted to a concentration of 4.5610 6 CFU (colony forming unit) in 100 ml RPMI1640 with 10% fetal calf serum and mixed with 3610 5 neutrophils in the same media. The mixture was centrifuged at 700 g for 5 min, and then incubated at 37uC in a 5% CO 2 incubator. After 10 min gentamicin was added to a final concentration of 400 mg/ml for killing the extracellular bacteria. Then, the contents of sample wells were withdrawn and centrifuged to pellet the neutrophils. Neutrophils were washed twice, and then lysed in 0.02% Triton X-100. CFU was calculated by serial dilution plated on Todd-Hewitt agar (THA; 1.5% agar in Todd-Hewitt broth) plate.
Whole-blood killing assay (WBKA) S. aureus strains were grown for 12 h and harvested by centrifugation (10,000 g, 1 min). Bacteria were washed twice in PBS, and then 1610 4 cells were diluted to the volume of 25 ml in PBS, and mixed with 75 ml freshly drawn mouse blood in heparinized tubes. The tubes were incubated at 37uC for 4 h with agitation, and then surviving cells were enumerated by plating the dilutions on THA plates.

Cytotoxicity assay
The supernatant of S. aureus was collected as described [38,39] with some modifications. Briefly, S. aureus cells were grown to stationary phase at 37uC. Growth culture was centrifuged at 6,000 g for 10 min at 4uC. The supernatant was collected and filtered through a 0.22 mm filter to remove residual cells.
Madin-Darby Bovine Kidney (MDBK, ATCC-CCL22) cells were resuspended at a concentration of 1610 5 cells/ml and added to a 96-well plate (100 ml/well) at 37uC in a 5% CO 2 incubator for 6 h. The supernatant of S. aureus strains (5 ml/well) was added, with BHI broth as the control. The CCK8 assay was used to test the number of survival cells as per the protocol (Dojindo) after 12 h incubation. The cell viability percentage was calculated as: Viability percentage (%) = (Absorption value of supernatant of treatment group)/(Absorption value of supernatant of control group) 6100%.
For flow cytometric analysis, the MDBK cells were resuspended at a concentration of 1610 6 cells/ml and added to a 12-well plate (1 ml/well) at 37uC in a 5% CO 2 incubator for 6 h. The cultivated cells were incubated with BHI or with supernatant (100 ml/well) of S. aureus strains for 1 h. Prior to harvesting, the cells were washed twice with PBS, trypsinized, and pelleted. Then cells were resuspended at a concentration of 1610 6 cells/ml in Binding Buffer (0.01 M HEPES/NaOH, pH 7.4, 14 mM NaCl, 0.25 mM CaCl 2 ). 500 ml aliquots of cells were added into FACS tubes and mixed with 25 ng/ml fluorescein isothiocyanate-labeled annexin V and 10 mg/ml propidium iodide (PI) to incubation for 15 min at room temperature in the dark. Then the cells were analyzed immediately by flow cytometry.

RNAs binding to Hfq were obtained by immunoprecipitation
The immunoprecipitation assays were prepared as described previously [23] with some modifications. The purified Hfq protein (100 mg) was incubated with 5 ml rabbit immune serum containing S. aureus Hfq-specific antibodies or rabbit normal serum at 37uC for 30 min. The captured antibodies were then bound to 1 ml of protein A sepharose CL-4B (Sigma-Aldrich) at 37uC for 30 min. After four washes with 2 ml washing buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.5% Triton X-100), the total RNA of S. aureus was added and incubated at 37uC for 30 min. After another four washes with 2 ml washing buffer, the binding complexes were eluted from protein A sepharose with 1 ml of 0.1 M Glycine-HCl, pH 2.5, and neutralized with 100 ul of 1 M Tris-HCl, pH 8.0. Then the mixture was extracted by phenol, followed by RNA precipitation and the RNAs which were bound to Hfq were analyzed by microarray. Calculation of enrichment factor was performed as follows: enrichment factor = signal intensities of Hfq IP/signal intensities of control IP.

Transcriptional microarray analysis
Gene expression profiles of the Dhfq-8325 and wild type strains were analyzed by using Affymetrix S. aureus arrays as described previously [40] with some modifications as follows. The S. aureus array contains probe sets to over 3,300 S. aureus open reading frames (ORFs) and 4,800 intergenic region sequences based on the updated S. aureus genomic sequences of N315, Mu50 and COL [41]. Total bacterial RNA was extracted from S. aureus which were grown for 6 h with shaking at 37uC using Trizol (Invitrogen) as previously described [41]. DNase digestion of 80 ml of total RNA was performed with 10 U of RNase-free DNase I (Promega) and 10 ml of the 106 reaction buffers in a total reaction volume of 100 ml for 30 min at 37uC. The RNA (10 mg) was reversely transcribed to cDNA by using M-MLV reverse transcriptase (Promega) and random primers (Promega) for 1 h at 42uC. The fragmented cDNAs were then directly labeled with biotin by using a biotin-ddUTP kit (Affymetrix). Biotinylated cDNA was hybridized to the GeneChips. Microarrays were scanned using an Affymetrix GeneChip Scanner 3000 and image data were extracted using GeneChip Operating Software (GCOS) Version 1.4. Two independent labeling reactions and hybridizations were carried out for each RNA sample. This experiment was completed by Bioassay Laboratory of CapitalBio Corporation. Normalization and expression analysis were performed with DNA-chip analyzer (dChip). Invariant set normalization was used to normalize arrays at the probe level. Genes with variations (n-fold$1.5) were collected. All data is MIAME compliant, and the raw data has been deposited in the database ArrayExpress (No. E-TABM-1005).

Preparation of S. aureus total proteins and Western blot
The cell extract was prepared as described [41] with some modifications as follows. Cells were grown for 6 h with shaking at 37uC. Equal numbers of cells (1610 10 CFU) were collected and then resuspended in 1 ml PBS buffer containing lysostaphin (100 mg/ml, Sigma-Aldrich). After incubation for 30 min at 37uC, the mixture was centrifuged at 13,000 g for 10 min and the precipitate was resuspended in 1 ml PBS. The mixture was sonicated on ice three times for 30 s each and centrifuged at 13,000 g for 10 min. The supernatant was precipitated by adjusting filtered supernatants to 10% trichloroacetic acid (TCA) and incubated at 4uC for 4 h. After centrifugations at 12,000 g for 10 min, precipitated proteins were washed twice in ice-cold 96% ethanol, and air dried. The proteins were resolved in an appropriate volume of a solution containing 7 M urea and 2 M thiourea. The samples were then subjected to 15% SDS-PAGE and the proteins were blotted onto Hybond-ECL nitrocellulose membrane (Amersham Biosciences). The membrane was blocked in 5% non-fat dry milk at 37uC for 2 h, probed with 1:500 diluted polyclonal rabbit anti-Hfq antibodies (prepared by ourselves) for 1 h at room temperature, and washed twice in PBS with 0.5% Tween 20 (PBST). Then the membrane was incubated in a 1:5,000 solution of HRP-conjugated goat anti-rabbit secondary antibody (Jackson) at room temperature for 1 h. After further washing with PBST, the membrane was assayed by the enhanced chemiluminescence (ECL) Western blotting detection system (Pierce).

Quantitative reverse transcription PCR (qRT-PCR)
For cDNA synthesis, 1 mg of total RNA was mixed with 500 ng of random hexamer (Promega). Samples were incubated at 65uC for 10 min with 5 ml of 56 first-strand buffer, 2 ml of 5 mM dNTP, 2 0U of RNasin (Takara), 1 ml of M-MLV reverse transcriptase (Promega) and distilled water to a total volume of 25 ml. The qPCR reaction mixture contained 12.5 ml of 26SYBR green PCR mix (GenePharma), 0.3 mM of gene-specific forward and reverse primers, and 1 ml of template, made up to a final volume of 25 ml with distilled water. The primers are shown in Table 2. Cycling parameters were set as follows: initial activation step at 95uC for 10 min, denaturation at 94uC for 30 s, annealing at 58uC for 30 s, and extension at 72uC for 40 s. Melting curve analysis was performed at from 58uC to 95uC with stepwise fluorescence acquisition at every 1uC s 21 . Melting curves observed for each gene were confirmed to correspond to the correct amplicon size by agarose gel electrophoresis of the PCR products. The levels of gene expression were calculated by relative quantification using 16S rRNA as the endogenous reference gene. All samples were amplified in triplicate and the data analysis was carried out using the MxPro qPCR system software (Stratagene).

Statistical analysis
All quantitative data were analyzed using Student t-tests. P,0.05 was considered to be statistically significant.