The Small RNA Teg41 Regulates Expression of the Alpha Phenol-Soluble Modulins and Is Required for Virulence in Staphylococcus aureus

The alpha phenol-soluble modulins (αPSMs) are among the most potent toxins produced by Staphylococcus aureus. Their biological role during infection has been studied in detail; however, the way they are produced by the bacterial cell is not well understood. In this work, we identify a small RNA molecule called Teg41 that plays an important role in αPSM production by S. aureus. Teg41 positively influences αPSM production. The importance of Teg41 is highlighted by the fact that a strain containing a deletion in the 3′ end of Teg41 produces significantly less αPSMs and is attenuated for virulence in a mouse abscess model of infection. As the search for new therapeutic strategies to combat S. aureus infection proceeds, Teg41 may represent a novel target.

cytolytic peptides called phenol-soluble modulins (PSMs). PSMs are amphipathic, alpha helical peptides that vary in size depending on their classification. The ␣ type PSMs are typically 20 to 25 amino acids in size, while the ␤ type PSMs are around 40 to 45 amino acids. The ␣ type PSMs have been the focus of intense study in recent years, as they have been implicated in contributing to the high virulence potential of communityacquired methicillin-resistant S. aureus (CA-MRSA) strains (2,3). While the role of PSMs in S. aureus infection has been extensively investigated (for a comprehensive review of PSMs and S. aureus virulence, see reference 4), their production within the S. aureus cell is less well studied.
There are five ␣ type PSMs in S. aureus. Four of them (named ␣PSM1 to ␣PSM4) are encoded on a polycistronic transcript (the ␣PSM transcript [ Fig. 1]), while the fifth, the ␦-toxin, is encoded within the RNAIII transcript. Interestingly, both of these transcripts are regulated by the agr system by direct binding of AgrA to the promoter region (5). Therefore, while the production of ␣ type PSMs is linked to population density and virulence (through the agr system), little else is known about the mechanisms that control PSM production within the bacterial cell. PSMs do not contain a secretion signal sequence and are secreted from the cell via a dedicated ABC transport system (the Pmt system) (6).
Small RNAs (sRNAs) represent an understudied class of regulatory molecules in S. aureus (7)(8)(9)(10). They are often overlooked because the majority of sRNA genes (including that of RNAIII) are not included in GenBank genome annotation files (due to the small size of the ␣PSM ORFs, the ␣PSM locus/transcript is typically absent in S. aureus annotated genome files too). To address this, we recently performed a comprehensive mapping study which annotated 303 individual sRNAs (and the ␣PSM transcript) on the genome of S. aureus CA-MRSA strain USA300 (11). These annotated files serve as a valuable resource for analysis of sRNAs in S. aureus. In addition to providing a reference genome that can be used for transcriptomic studies (of sRNAs and ␣PSMs), these files clearly outline the location on the chromosome of each sRNA gene relative to proteincoding genes.
During the annotation process, we observed that the gene encoding a previously unstudied sRNA (called Teg41) was located immediately downstream of, and divergently transcribed from, the ␣PSM transcript (Fig. 1). The close proximity of Teg41 to the ␣PSMs led us to hypothesize that Teg41 may play a role in regulating ␣PSM production. In this study, we demonstrate that Teg41 positively regulates expression of the ␣PSMs. Overproduction of Teg41 leads to increased erythrocyte hemolysis and increased production of ␣PSMs. We identify the region of Teg41 responsible for regulating the ␣PSMs and demonstrate that deletion of this region (i) results in a decrease in ␣PSM production and (ii) attenuates S. aureus virulence in a murine abscess model of infection. Together, these data demonstrate for the first time regulation of the ␣PSMs by an sRNA and further demonstrate the important role played by sRNAs in regulating virulence in S. aureus.

RESULTS
The sRNA Teg41 contributes to hemolytic activity in S. aureus. Recently, we performed a comprehensive reannotation of the S. aureus genome in which all previously identified sRNAs were mapped to their reported chromosomal locations (11). During the annotation process we observed that the gene encoding a previously unstudied sRNA (called Teg41) was located immediately downstream of, and divergently transcribed from, the alpha phenol-soluble modulin (␣PSM) transcript (Fig. 1). The close proximity of Teg41 to the ␣PSMs led us to hypothesize that Teg41 may play a role in regulating ␣PSM production. To test this hypothesis, we attempted to create a Teg41 deletion strain by allelic exchange. Despite repeated attempts, we were unable to delete the Teg41 gene. In the absence of a Teg41 mutant strain, we elected to examine the effect of Teg41 overproduction. The Teg41 gene, under the control of its native promoter, was expressed on a multicopy plasmid (pMK4) in S. aureus strain USA300 TCH1516. The ␣PSM peptides are potent hemolysins; therefore, to test the potential contribution of Teg41 to ␣PSM production, we examined the hemolytic activity of S. aureus cells ( Fig. 2A) and cell-free culture supernatants (Fig. 2B). When we compared wild-type S. aureus (containing the empty vector) to the Teg41-overproducing strain, a significant increase in hemolytic activity was observed in both S. aureus cells and cell-free culture supernatants taken from the Teg41-overproducing strain (Fig. 2). This result shows that Teg41 contributes to S. aureus hemolytic activity, potentially by regulating ␣PSM activity.
To investigate whether Teg41 overexpression leads to increased hemolysis in multiple S. aureus backgrounds, the Teg41 overexpression plasmid and empty vector control were transduced into five wild-type S. aureus strains: USA300 JE2, USA300 Overproduction of Teg41 results in an increase in hemolytic activity in cell-free 15-h culture supernatants in various wild-type S. aureus backgrounds. Hemolytic activity in the wild-type strain containing the vector control (WT pMK4) is set at 100%, and the relative hemolytic activity of the Teg41-overproducing strain (WT pMK4_Teg41) is indicated as a percentage. All statistical analyses were performed using Student's t test. Statistical significance is indicated by bars and asterisks as follows: *, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.005.
Overproduction of Teg41 could result in increased hemolytic activity due to a regulatory function, or alternatively Teg41 could itself encode a hemolytic peptide. When we examined the Teg41 sequence, we identified one potential open reading frame (ORF) capable of encoding a peptide of 24 amino acids ( Fig. 3A and B). To investigate whether Teg41 elaborates this peptide (and it possesses hemolytic activity), we transduced the Teg41-overproducing plasmid into Staphylococcus epidermidis and performed hemolysis assays using cell-free culture supernatants. No hemolytic activity was observed in S. epidermidis, strongly suggesting that Teg41 does not encode a hemolytic peptide (Fig. 3C). Taken together, the data above indicate that overexpression of Teg41 leads to an increase in hemolytic activity in staphylococcal strains carrying genes that encode the ␣PSM peptides.
Teg41 is highly conserved in S. aureus. The Teg41 transcript was first identified by Beaume et al. in 2010 (12). Prior to that, work by Geissman et al. had predicted the existence of an sRNA at the corresponding location (which they termed RsaX05), due to the presence of orphan promoter and terminator sequences (13). Teg41 transcript has subsequently been identified in a number of published RNAseq experiments and has been given a variety of different names and designations (including sRNA095 [14], srn_1080 [15], and SAUSA300s087 [11]). Here we refer to it as Teg41, the name used upon the first experimental identification of the sRNA (12). While Teg41 has been observed in a number of studies, there are no reports investigating the role of Teg41 in S. aureus.
Teg41 was reported by Geissman et al. (13) and Beaume et al. (12) as a transcript located between the genes for the ␣PSMs and ndhF. ndhF encodes an NADH dehydrogenase involved in the transfer of electrons from NAD to quinones in the cellular membrane (16). Due to its close proximity and similar orientation to ndhF, it was speculated that Teg41 may function as a cis-acting riboswitch or 5= UTR for this gene ( Fig. 1) (12, 13). The results presented herein ( Fig. 2A and below) suggest that this is not the case, and instead, they imply that Teg41 is a stand-alone transcript that influences hemolytic activity of S. aureus. The sequence of Teg41 is highly conserved in S. aureus. A BLAST search of the Teg41 sequence from S. aureus strain USA300 FPR3757 identified 161 Teg41 homologues that displayed 100% sequence identity. The minimum sequence identity of S. aureus Teg41 homologues in the database is 94%. Outside of S. aureus, Teg41 is found in only one other Staphylococcus species, the closely related Staphylococcus argenteus. Interestingly, this pattern of distribution among staphylococcal strains (i.e., present only in S. aureus and S. argenteus) mirrors that of the ␣PSM operon. In contrast to this, ndhF, the gene downstream of Teg41, is found in all staphylococcal species. Thus, the ␣PSM operon and Teg41 are genetically linked, while a number of staphylococcal spp. contain a copy of the ndhF gene that is not preceded by Teg41.
Overproduction of Teg41 in a psm␣ mutant does not result in increased hemolytic activity. Due to the increase in hemolytic activity observed upon Teg41 overproduction and the close proximity of Teg41 to the psm␣ gene, we hypothesized that Teg41 is regulating ␣PSM production. To test this hypothesis, we overexpressed Teg41 in a psm␣ mutant strain (i.e., one in which the entire psm␣ locus has been deleted). If the hemolytic phenotype observed upon Teg41 overproduction is independent of the ␣PSMs, then an increase in hemolytic activity would be expected upon Teg41 overproduction in a psm␣ mutant strain. Results demonstrated that this was not the case. Overproduction of Teg41 in the JE2 wild-type S. aureus strain resulted in an increase in hemolysis, similar to that observed previously (Fig. 2); however, overproduction of Teg41 in the psm␣ mutant strain did not result in a significant increase in hemolytic activity (Fig. 4A). This result shows that the increase in hemolysis observed following overproduction of Teg41 is dependent on the ␣PSMs. Furthermore, it confirms that the increase in hemolysis is not due to a Teg41 encoded peptide, as no increase in hemolytic activity was observed in the psm␣ mutant strain overproducing Teg41.
To confirm that Teg41 overproduction leads to increased ␣PSM-dependent hemolysis, we overproduced Teg41 in a hla mutant strain. The hla gene encodes the S. aureus ␣-hemolysin (Hla), another potent hemolytic toxin. Overproduction of Teg41 in the hla mutant led to an increase in hemolytic activity, similar to that observed in the wild-type strain (Fig. 4B), clearly demonstrating that the increase in hemolysis observed is not due to the activity of Hla. It should be noted that hla mutant strains are still hemolytic (Fig. 4B). This is consistent with studies that demonstrate that Hla is highly hemolytic toward rabbit and sheep erythrocytes but relatively inactive against human erythrocytes (17).
Overproduction of Teg41 results in increased ␣PSM production. The results above strongly suggest that increased Teg41 production leads to increased ␣PSM levels. To test this hypothesis and determine whether ␣PSM peptide levels are increased upon Teg41 overproduction, we performed a butanol extraction to isolate and purify the PSM peptides from S. aureus culture supernatants (18). The extracted peptides were used in hemolysis assays and analyzed by SDS-PAGE (Fig. 5). The relative degree of hemolytic activity of extracted peptides (Fig. 5C) was similar to that observed in cell-free culture supernatants before butanol extraction (Fig. 5A). These results indicate that the Teg41dependent hemolysin was isolated during the butanol extraction procedure.
SDS-PAGE analysis of butanol extractions shows one band corresponding to the PSMs ( Fig. 5B; see also Fig. S1 in the supplemental material). An increase in PSM abundance was observed when Teg41 was overproduced in the wild-type background (Fig. 5B). Densitometry analysis on serially diluted samples indicates that PSM levels are increased approximately1.8-fold when Teg41 was overproduced (Fig. S2). An overall decrease in PSM levels was observed in the psm␣ mutant strain; however, a band was still detected (presumably representing the ␦-toxin). Overproduction of Teg41 in the psm␣ mutant strain did not result in increased PSM production.
To investigate further the nature and composition of the peptides visualized by SDS-PAGE, butanol-extracted peptides from the wild-type and psm␣ mutant strains with and without Teg41 overproduction were subjected to mass spectrometry analysis. Fragment ion spectra corresponding to the delta-toxin (Hld) peptide were identified in all samples, confirming that the band visible in the psm␣ mutant strains is Hld. In addition to Hld, peptides corresponding to ␣PSM1 to -4 were detected in extracts from wild-type bacteria, but not from the psm␣ mutant. This confirms that Teg41 overproduction leads to an increase in ␣PSM production, which in turn accounts for the increase observed in hemolysis.
The 3= region of Teg41 is required for hemolytic activity. To investigate the molecular mechanism through which Teg41 regulates the ␣PSMs, we began by analyzing the sequences of Teg41 and the ␣PSM transcript. We used the program intaRNA (19)(20)(21) to search for potential interactions between these two RNA molecules. The results generated predict an interaction between the 3= region of Teg41 (nucleotides 183 to 194) and the ␣PSM transcript between nucleotides 376 and 388 ( Fig. 6A and B). Interestingly, this region of the ␣PSM transcript is located immediately downstream of the translation start codon for PSM␣4 (Fig. 6B). On the basis of the results presented so far and the predicted interaction between Teg41 and the ␣PSM transcript, we hypothesized that a direct interaction between the 3= region of Teg41 and the ␣PSM transcript increases ␣PSM peptide production. To test this hypothesis, we constructed a mutant strain in which 24 nucleotides in the 3= region of Teg41 were deleted (i.e., the region predicted to interact with the ␣PSM transcript). As previously mentioned, our attempts to create a complete Teg41 deletion strain were unsuccessful; however, our attempt to delete the 3= region of Teg41 (Teg41Δ3) was successful. To test the hypothesis that deletion of the ␣PSM interaction site of Teg41 would lead to a decrease in hemolysis, we examined the hemolytic activity of culture supernatants from the Teg41Δ3= strain. Results show a 10-fold reduction in hemolytic activity in the Teg41Δ3= strain compared to the wild type (Fig. 6C). Providing full-length Teg41 in trans, on a plasmid, completely restored hemolytic activity in the Teg41Δ3= strain. To confirm that the results of the hemolysis assay in Fig. 6C were due to alterations in PSM production, culture supernatants were subjected to butanol extraction. SDS-PAGE analysis of the extracted peptides shows a reduction in PSM levels in the Teg41Δ3= strain (Fig. 6D). Providing Teg41 in trans (Teg41Δ3= pMK4_Teg41) increased PSM levels although not completely to the level in the wild-type strain. Together, these results clearly demonstrate that Teg41 is required for S. aureus ␣PSM production and hemolytic activity. Moreover, the results are consistent with a model whereby the 3= end of Teg41 is responsible for modulating ␣PSM levels.
Teg41 is required for virulence in an abscess model of infection. The results presented above ( Fig. 6C and D) demonstrate a reduction in PSM production and hemolytic Bands of similar intensities were observed in the wild-type strain with and without the empty vector. A band with increased intensity was observed in the wild-type strain overproducing Teg41. An overall decrease in PSM levels was observed in the ␣PSM mutant, which did not vary upon overproduction of Teg41. (C) Hemolytic activity of butanol-extracted peptides. Butanol-extracted samples were dissolved in water and used in human blood hemolysis assays. A similar result was obtained to that shown in panel A (using culture supernatants), indicating that the hemolysin responsible was purified during the butanol extraction process. Hemolytic activity in the wild-type strain containing the vector control (WT pMK4) is set at 100%, and the relative hemolytic activity of all other strains is indicated as a percentage. Statistical analyses were performed using Student's t test. Statistical significance is indicated as follows: *, P Ͻ 0.05; ns, not significant. The strain background was JE2.
Teg41 Regulates ␣PSM Production in S. aureus ® activity when Teg41 is disrupted. To compare the virulence of wild-type S. aureus and the Teg41Δ3= strain, we employed a murine abscess model of infection. Groups of 12 mice were subcutaneously injected with S. aureus, and following a 7-day infection, abscesses were measured and the number of bacteria present was determined. Abscess area was significantly reduced in mice infected with the Teg41Δ3=strain compared to the wild type (Fig. 7A). Additionally, a 302-fold reduction in bacterial numbers was observed in the abscesses of mice infected with the Teg41Δ3= strain compared to those infected with the wild type (Fig. 7B). Histopathology of excised abscesses shows widespread necrosis and muscle destruction in mice infected with the wild-type strain, as well as a high concentration of inflammation-associated leukocytes (Fig. 7C). Conversely, mice infected with the Teg41Δ3=strain display intact muscle tissue, no necrosis, and a lower concentration of leukocytes (Fig. 7D). Taken together, these results show that the reduction in ␣PSM production in the Teg41Δ3= strain results in attenuation and clearly demonstrates that Teg41 is required for virulence in S. aureus. Growth curve aureus culture supernatants is significantly reduced in the Teg41Δ3= strain. A 10-fold reduction in hemolytic activity was observed in culture supernatants from the Teg41Δ3= strain. Hemolysis was restored to wild-type levels by introducing full-length Teg41 on a plasmid (Teg41Δ3= pMK4_Teg41). Hemolytic activity in the wild-type strain is set at 100%, and the relative hemolytic activity of the other strains is indicated as a percentage. Statistical analyses were performed using Student's t test (**, P Ͻ 0.01; ns, not significant). (D) PSM levels are significantly reduced in the Teg41Δ3= strain. Culture supernatants from panel C were butanol extracted and analyzed by SDS-PAGE. A reduction in PSM peptide levels was observed in the Teg41Δ3= strain. Expressing full-length Teg41 from a plasmid in the Teg41Δ3= strain (Teg41Δ3= pMK4_Teg41) resulted in an increase in PSM production. The strain background was AH1263.

Zapf et al.
® analysis demonstrated that that there is no difference in growth rate between the wild-type and Teg41Δ3= strains (Fig. S3). Virulence of S. aureus in the murine abscess model is strongly linked to the activity of the alpha-toxin, Hla. Hla production has been shown to be dependent on ␣PSM levels, with delayed induction of hla expression observed in an ␣PSM mutant strain (22). Therefore, we hypothesized that the attenuation in virulence observed in the Teg41Δ3= strain may be due in part to decreased Hla production. To test this hypothesis, we performed hemolysis assays using S. aureus culture supernatants and rabbit blood (as previously mentioned, rabbit erythrocytes are exquisitely sensitive to Hlamediated lysis). Results show a reduction in activity from Teg41Δ3= culture supernatants compared to the wild-type culture supernatant, strongly suggesting reduced Hla production in this strain (Fig. 8).
To confirm that the virulence phenotypes observed in the Teg41Δ3= strain were due to the 24-nt deletion in Teg41, we subjected the Teg41Δ3= strain and the parental wild-type strain to whole-genome sequencing. Analysis of the genome sequencing data confirmed the deletion in the 3= region of Teg41. No additional polymorphisms, insertions, or deletions were observed in any S. aureus virulence genes and/or regulatory proteins in the Teg41Δ3= strain.
The 3= end of Teg41 is necessary and sufficient for S. aureus hemolysis. The data presented above (Fig. 6) show that the 3= end of Teg41 is necessary for ␣PSMdependent hemolytic activity in S. aureus. Expression of full-length Teg41 from a Teg41 Regulates ␣PSM Production in S. aureus ® plasmid restored hemolysis and PSM production in the Teg41Δ3= strain (Fig. 6). To investigate further the role of the 3= end of Teg41, we constructed a plasmid expressing a truncated form of Teg41 missing 47 nt at the 3= end (Teg41_5=). When the truncated form of Teg41 (Teg41_5=) was expressed in wild-type S. aureus or in the Teg41Δ3= strain, there was no increase in hemolytic activity (Fig. 9A). This result specifically shows that the 3= end of Teg41 is necessary to restore hemolysis when ectopically expressed from a plasmid. We next wanted to investigate whether expression of the 3= end of Teg41 alone is sufficient for hemolytic activity. To do so, we examined the ability of the isolated 3= end of Teg41 (Teg41_3=) to induce hemolysis in the wild-type and Teg41Δ3= strains. Full-length Teg41 and Teg41_3= were cloned into the inducible plasmid pCN51 (generating plasmids pCN51_Teg41 and pCN51_Teg41_3=), and each plasmid, along with the vector control, was transformed into wild-type S. aureus and the Teg41Δ3= strain. Expression from each plasmid was induced by the addition of CdCl 2 , and hemolytic activity of culture supernatants was determined using human blood. As previously demonstrated, overexpression of full-length Teg41 resulted in an increase in hemolytic activity in wild-type S. aureus and restored hemolytic activity in the Teg41Δ3= strain (Fig. 9B). Interestingly, expression of Teg41_3= also resulted in an in increase in hemolytic activity and restoration of hemolytic activity in the wild-type and Teg41Δ3= strains, respectively. This result clearly demonstrates that the 3= end of Teg41 is both necessary and sufficient for ␣PSM-mediated hemolytic activity in S. aureus. In both the wild-type and Teg41Δ3= strains, the increase in hemolytic activity observed from expression of Teg41_3= was not as great as that observed when expressing full-length Teg41. This suggests that while the hemolysis effect is mediated by the 3= end of Teg41, the 5= end may still play a role in Teg41 activity.
Analysis of Teg41 and ␣PSM transcripts. Teg41 was initially reported, and is annotated, as a 205-nt transcript (12); however, previous studies have reported different sizes, with estimates ranging from 146 nt (14) to 300 nt (13). To confirm the size of Teg41 and investigate the effect of the 24-nt chromosomal deletion in the Teg41Δ3= strain, we performed Northern blotting using a Teg41-specific probe and total RNA extracted from wild-type S. aureus and the Teg41Δ3= strain at 6 h growth. Results show a band approximately 200 nt in size in the wild-type bacteria (Fig. 10A), confirming the predicted size of Teg41 (12) and demonstrating that the Teg41 transcript is independent of ndhF. In contrast, Teg41 was barely detected in the Teg41Δ3= strain. A very faint band was detected, running at a slightly lower apparent molecular weight, which we hypothesize is the truncated form of Teg41 (i.e., Teg41_3=). The large reduction in Teg41 abundance in the Teg41Δ3= strain suggests that truncated Teg41 is probably unstable and is rapidly degraded in the cell. To confirm that the band detected corresponds to truncated Teg41, we performed reverse transcriptase quantitative PCR (RT-qPCR) using primers that anneal to the 5= end of Teg41 (which is present in the Teg41Δ3= strain). Using RNA from 6-h growth as the template, results show that truncated Teg41 was detected but that the transcript is approximately threefold less abundant (Fig. 10B). These results show that truncated Teg41 is being expressed in the Teg41Δ3= strain but is present at a lower level than full-length Teg41 in the wild-type strain.
Our data clearly show that Teg41 positively influences ␣PSM production. Positive gene regulation by sRNAs is less common than negative regulation and whenever reported often occurs through the stabilization of transcripts, which leads to increased translation. To determine whether the presence of Teg41 leads to increased abundance of the ␣PSM transcript, we performed Northern blotting to examine ␣PSM transcript levels in the wild-type and Teg41Δ3= strain. Using RNA extracted at 3 h and 6 h of growth, we observed decreased ␣PSM transcript levels in the Teg41Δ3= strain compared to the wild type (Fig. 10C). The reduction in ␣PSM transcript levels in the Teg41Δ3= strain was confirmed by RT-qPCR (Fig. 10D).
Together, these results show that deleting the 3= end of Teg41 leads to decreased Teg41 and ␣PSM transcript levels in the cell. While the decrease in ␣PSM transcript Overexpression of full-length Teg41 (pCN51_Teg41) or the Teg41 3= end (pCN51_Teg41_3') results in an increase in hemolytic activity in both wild-type S. aureus and the Teg41Δ3= strain. In panels A and B, hemolytic activity in the wild-type strain without plasmid was set at 100%, and the relative hemolytic activity of all other strains was indicated as a percentage. Statistical analyses were performed using Student's t test. Statistical significance is indicated as follows: **, P Ͻ 0.01; ns ϭ not significant. The strain background was AH1263.
Teg41 Regulates ␣PSM Production in S. aureus ® levels could be due to decreased promoter activity, we hypothesize that in the absence of Teg41, the ␣PSM transcript is rapidly degraded, leading to decreased ␣PSM peptide production and attenuation of virulence.

DISCUSSION
Over the last decade, the biological role and contribution of the ␣PSMs to S. aureus virulence have become clear, but there are still significant gaps in our understanding of how they are produced in the bacterial cell. Although PSM␣1 to PSM␣4 are encoded within the same polycistronic transcript (Fig. 1), recent in vitro studies have shown that the levels of the four peptides vary. A study of MRSA isolates demonstrated that PSM␣4 is commonly the most abundant ␣PSM, followed by PSM␣1 (23). PSM␣2 and PSM␣3 are typically the least abundant. A similar pattern was reported in MSSA strains (24). The mechanism responsible for this variation in ␣PSM levels is unknown; however, this study and the discovery that the ␣PSM transcript is subject to sRNA-mediated regulation may begin to shed some light on it. For the first time (that we are aware of), we have demonstrated sRNA-mediated regulation of the ␣PSM peptides. Teg41 positively influences ␣PSM production with ␣PSM levels increased upon Teg41 overproduction and decreased in the Teg41Δ3= strain (i.e., when the 3= region of Teg41 is deleted). Our Northern blot analysis shows decreased abundance of the ␣PSM transcript in the Teg41Δ3= strain, suggesting that Teg41 regulates the ␣PSMs at the level of transcription or transcript stability. sRNA-mediated, positive gene regulation frequently occurs at the level of transcript stability, leading us to propose a model whereby binding of Teg41 to the ␣PSM transcript stabilizes the transcript, facilitating increased translation of the ␣PSM peptides. This proposed model does not preclude the possibility that Teg41 also functions at the translational level. Binding of Teg41 could result in a conformational change to the ␣PSM transcript that increases or decreases ribosomal access to one or more ␣PSM ribosome binding sites. In support of this idea, the predicted structure of the ␣PSM transcript shows all four ␣PSM ribosome-binding sites (RBS) located within base-paired regions (Fig. 11). Thus, while a transcript stability mechanism seems likely, further experimental investigation is necessary to fully understand the mechanism(s) underlying Teg41-mediated regulation of the ␣PSMs. Interestingly, a recent bioinformatic analysis of published RNAseq data, performed by Subramanian et al. (25), examined the expression profile of sRNAs in S. aureus and predicted an important role for Teg41 (called sRNA095 in their study) in S. aureus pathogenesis.
Here we provide experimental evidence to confirm this prediction, validating the approach used in this study. The sequences of sRNAs typically give little valuable information regarding their nature or function. Initial reports suggested that Teg41 was likely a 5= UTR or riboswitch for the downstream ndhF gene (12,13). Our analysis also revealed the possibility that a short peptide was encoded within the Teg41 RNA. The data presented in this study clearly demonstrate that neither of these is the case. Rather, Teg41 is a stand-alone, trans-acting regulatory RNA that regulates production of the ␣PSMs. While it is impossible to rule out the existence of a hemolytic Teg41-derived peptide, three important results demonstrate that this potential peptide is not contributing to hemolysis. First, overproduction of Teg41 in Staphylococcus epidermidis did not result in an increase in hemolytic activity (Fig. 3). Second, overproduction of Teg41 in an ␣PSM mutant did not result in an increase in hemolytic activity (Fig. 4). If Teg41 were to encode a hemolytic peptide, it is highly likely that an increase in hemolysis would have been observed in these experiments. Finally, expressing the 3= end of Teg41 from a plasmid (pCN51_Teg41_3') caused an increase in hemolytic activity in the wild-type and Teg41Δ3= strain. The Teg41_3= segment expressed from this plasmid does not contain the potential ORF, thereby conclusively demonstrating that a hemolytic peptide is not responsible for the Teg41-mediated increase in hemolysis observed. The ␣PSM operon is a rare example of a locus in S. aureus that is regulated by the agr system independently of RNAIII (5). The response regulator AgrA binds directly to the ␣PSM promoter and activates transcription. This ties ␣PSM expression to cell density. A second virulence-associated global regulator, MgrA, was recently shown to bind to the ␣PSM promoter and regulate expression of the ␣PSM operon (26). In contrast to AgrA, MgrA represses expression of the ␣PSM transcript. The Teg41 and ␣PSM promoters are found in close proximity to each other; therefore, it is tempting to speculate that AgrA and/or MgrA (which both bind in this region) may also have direct roles in activating the Teg41 promoter. We are currently investigating this possibility, mapping the Teg41 promoter and examining the expression of Teg41 under a variety of conditions. Previous work by Queck et al. (5) has shown that additional regulatory site(s) are located upstream of the AgrA binding site, within the Teg41 gene. We are also investigating the roles of these elements in the regulation of Teg41 in an effort to better understand how Teg41 controls ␣PSM production.
The data presented in this study clearly demonstrate regulation of the ␣PSMs by Teg41; however, additional studies will be required to fully understand the role of Teg41 in the S. aureus cell and to determine whether the regulation of Teg41 is transcriptional or posttranscriptional. We consider it likely that our inability to create a Teg41 deletion strain is due to the removal of critical promoter elements for the downstream ndhF gene while attempting to delete Teg41. Although not considered essential, it is likely that a strain deficient in ndhF expression would exhibit severe growth defects. Deleting 24 nucleotides at the 3= end of Teg41 was possible, and the resulting strain (Teg41Δ3') did not exhibit any growth defect compared to the wild type (see Fig. S3 in the supplemental material). This result strongly suggests that the ndhF promoter is intact in this strain and ndhF expression is unaffected. Even if ndhF expression is altered in the Teg41Δ3= strain, our data demonstrate that it is not responsible for the hemolysis phenotype observed. Complementation of the Teg41Δ3= hemolysis defect was possible with plasmid expression of Teg41 or Teg41_3=, demonstrating that the hemolysis phenotype is Teg41 dependent.
The work presented herein exemplifies the growing appreciation of the importance of sRNAs in S. aureus and further expands the list of sRNAs known to play a role in controlling virulence (10). It is the first time an sRNA has been shown to regulate expression of the ␣PSMs. Furthermore, it highlights the value of including annotations for sRNAs directly on reference genomes (11). Work is ongoing in our lab to fully explore the molecular mechanism through which Teg41 controls ␣PSM production. We acknowledge that a direct interaction between Teg41 and the ␣PSM transcript has yet to be established. We consider it likely that such an interaction exists; however, we do not discount the possibility that the regulation could be indirect. If this is the case, then the indirect regulation would appear to be mediated by the 3= end of Teg41, as removal of this region resulted in a significant reduction in ␣PSM production, and expressing this region on a plasmid restores hemolysis in the Teg41Δ3= strain. Regulation of ␣PSM production by Teg41 could potentially represent a novel target for therapeutic intervention, as disrupting Teg41 activity is likely to dramatically reduce the virulence potential of S. aureus.

MATERIALS AND METHODS
Strains, plasmids, and primers. All bacterial strains and plasmids used in this study are listed in Table 1. Oligonucleotides are listed in Table 2. S. aureus cultures were routinely grown at 37°C with shaking in tryptic soy broth (TSB), and Escherichia coli cultures were grown at 37°C with shaking in lysogeny broth (LB). Staphylococcus epidermidis cultures were grown at 37°C with shaking in B2 broth (27). Where appropriate, the following antibiotics were used at the concentrations indicated: chloramphenicol (10 g ml Ϫ1 ), erythromycin (5 g ml Ϫ1 ), lincomycin (25 g ml Ϫ1 ), and ampicillin (100 g ml Ϫ1 ). To induce promoter activity in strains containing plasmid pCN51, CdCl 2 was used at a concentration of 2 M. For comparative analysis of supernatants, S. aureus cultures were synchronized as previously described (28). Briefly, overnight 5-ml cultures of each strain were diluted 1:100 in 10 ml of fresh, prewarmed TSB, and grown for 3 h to mid-exponential phase. The 3-h, mid-exponential-phase cultures were subsequently diluted into 25 ml of fresh TSB at a starting OD 600 of 0.05. The cultures were then grown for the time indicated, typically 15 h. amplified using primer pair #0232/#0488 and digested with the restriction enzymes EcoRI and BamHI. The resulting fragment was cloned into plasmid pMK4. The Teg41-overproducing plasmid pCN51_Teg41 (pRKC0473) was constructed as follows. A 205-nt region of the S. aureus chromosome, containing Teg41, was amplified using primer pair #0301/#0302 and digested with the restriction enzymes EcoRI and SalI. The resulting fragment was cloned into plasmid pCN51. The Teg41 3= plasmid pCN51_Teg41_3= (pRKC0628) was constructed as follows. A 62-nt region of the S. aureus chromosome, containing the 3= end of Teg41, was amplified using primer pair #0580/#0581 and digested with the restriction enzymes EcoRI and SalI. The resulting fragment was cloned into plasmid pCN51.
After plasmid construction, clones were selected by transforming ligation mixtures into E. coli DH5␣ and selecting on ampicillin agar plates. Plasmid sequence was confirmed by restriction digestion and DNA sequencing. Once confirmed, plasmids were introduced into E. coli RN4220 by electroporation and subsequently transferred to additional strains by phage transduction (29). To introduce plasmids into S. epidermidis strains, total DNA (which contains both genomic DNA and plasmid DNA) was prepared from S. aureus strains RKC0072 and RKC0474 (30). Aliquots of each DNA isolation (which contain the desired plasmids) were then electroporated into S. epidermidis strain 1457 as previously described (31).
To construct the Teg41Δ3= strain (JLB162), a chromosomal fragment upstream of Teg41 was amplified using PCR from the AH1263 chromosome with primers JBKU89 and JBKU90. The product was digested with EcoRI and KpnI and then ligated into the same site of pJB38 to generate pJB1037. Next, the downstream fragment was amplified using primers JBKU92 and JBKU93, digested with KpnI and SalI, and ligated into the same sites of pJB1037, yielding pJB1039. pJB1039 was introduced into strain RN4220 by electroporation and subsequently transferred into strain AH1263 by phage transduction as previously described (29). Allelic exchange was performed (32), and the mutation was confirmed by digestion with KpnI of a PCR product using primers flanking the mutation. Additionally, total DNA was extracted (30), and PCR was performed with primers JBKU94 and JBKU95 to amplify a region outside the sequence used for pJB1039. This region was completely sequenced to ensure that no unintended changes were made during the mutant construction process.
Cell-free blood hemolysis assay. Synchronized, cell-free, S. aureus culture supernatants were diluted 1:2 in reaction buffer (40 mM CaCl 2 , 1.7% NaCl) and incubated at 37°C in a tube revolver with 25 l of whole blood from humans or rabbits. Following a 10-min incubation, the samples were centrifuged at 5,500 ϫ g, and 100 l of the supernatant was transferred to a 96-well plate. The degree of erythrocyte lysis was determined by reading the absorbance of the samples at OD 543 .
Whole-blood hemolysis assay. Bacterial strains were grown to mid-exponential phase, pelleted, washed with PBS, resuspended in PBS, and used to inoculate 2-ml aliquots of human blood at an equivalent of OD 600 of 0.05. Inoculated samples were incubated with agitation at 37°C. At the time points indicated, 200-l samples were withdrawn from each sample, and the intact erythrocytes were pelleted by centrifugation at 10,000 rpm. The degree of erythrocyte lysis was determined by OD 543 measurement of the resulting supernatant. At each time point, additional samples were withdrawn, and the number of bacteria was determined by serially diluting and plating on TSA.
Butanol extraction of PSMs. PSMs were isolated from S. aureus culture supernatants using the previously published butanol extraction method (18). Five milliliters of synchronized cell-free culture supernatants was combined with 2 ml of n-butanol. The samples were incubated with shaking at 37°C for 1 h. After this incubation, the top organic layer was removed and dried down via vacuum centrifugation for 6 h at 5,000 rpm. To visualize PSMs, extracts were dissolved in water, mixed with 6ϫ loading buffer, separated on 12% SDS-PAGE gels, and stained with Coomassie brilliant blue.