Investigations into σ^B-Modulated Regulatory Pathways Governing Extracellular Virulence Determinant Production in Staphylococcus aureus

Abstract

The commonly used Staphylococcus aureus laboratory strain 8325-4 bears a naturally occurring 11-bp deletion in the σ^B-regulating phosphatase rsbU. We have previously published a report (M. J. Horsburgh, J. L. Aish, I. J. White, L. Shaw, J. K. Lithgow, and S. J. Foster, J. Bacteriol. 184:5457-5467, 2002) on restoring the rsbU deletion, producing a σ^B-functional 8325-4 derivative, SH1000. SH1000 is pleiotropically altered in phenotype from 8325-4, displaying enhanced pigmentation, increased growth yields, and a marked decrease in secreted exoproteins. This reduction in exoprotein secretion appears to result from a sixfold reduction in agr expression. In this study we have undertaken transposon mutagenesis of SH1000 to identify components involved in the modulation of extracellular proteases and α-hemolysin compared to 8325-4. In total, 13 genes were identified displaying increased α-hemolysin transcription and extracellular proteolysis. Phenotypic analysis revealed that each mutant also had decreased pigmentation and a general increase in protein secretion. Interestingly this phenotype was not identical in each case but was variable from mutant to mutant. None of the genes identified encoded classic regulatory proteins but were predominantly metabolic enzymes involved in amino acid biosynthesis and transport. Further analysis revealed that all of these mutations were clustered in a 35-kb region of the chromosome. By complementation and genetic manipulation we were able to demonstrate the validity of these mutations. Interestingly transcriptional analysis revealed that rather than being regulated by σ^B, these genes appeared to have a role in the regulation of σ^B activity. Thus, we propose that the loss of individual genes in this chromosomal hot spot region results in a destabilization of cellular harmony and disruption of the σ^B regulatory cascade.

The pathogenic success of Staphylococcus aureus is largely due to its bewildering array of secreted and surface-associated virulence determinants (37, 41). The result of such a repertoire of infection components is that this organism is able to cause a range of infections from benign or minor wound infections to far more serious and life-threatening conditions such as septicemia and endocarditis. The damaging virulence proteins produced by S. aureus are subject to multilevel and multifactorial regulation both temporally and spatially in response to the environments encountered during pathogenesis (42). This responsive and adaptive nature is thought to be central to the disease-causing ability of the organism and is largely the result of the multiple control mechanisms that it employs in gene regulation.

Two very well characterized global regulatory loci exist in S. aureus, agr and sarA, which through individual and combined approaches mediate the expression of virulence determinant production (1, 5, 11, 12, 13, 14, 15, 16, 22, 25, 42, 43, 46). In addition to these there are also a number of other regulatory loci, including 16 two-component regulators (3, 13, 41) and an ever-growing family of SarA homologues (13). To further complicate matters, S. aureus also possesses an alternative sigma factor, σ^B, which has recently been shown to be a potent and effective regulator of a wide range of genes (6, 62, 63). Numerous virulence-associated loci have been shown to be regulated by this factor, including coa, sarA, fnbA, clfA, sspA, hla, and seb (4, 18, 19, 23, 50, 55).

Investigations into σ^B have revealed that the lab strain 8325-4, used by many researchers working on the biology of S. aureus, contains an 11-bp deletion in the σ^B-controlling phosphatase rsbU (20, 21, 34). This deletion results in severe reduction of σ^B function. We created an unmarked variant of 8325-4, called SH1000, carrying a repaired and functional rsbU locus, which demonstrated that the restoration of σ^B function resulted in pleiotropic differences between the two strains (23). Compared to 8325-4, SH1000 displayed a number of phenotypic variations including increased pigmentation, decreased growth lag phase, increased final growth yield, and a marked decrease in secreted exoproteins, including α-hemolysin and the major proteases (23, 52).

Analysis of SarA production in SH1000 revealed that levels remained constant throughout growth at both the transcriptional and the protein levels, with no variation observed when a sigB mutation was introduced into the strain (23). Further characterization of SH1000 revealed that the levels of agr transcription were markedly decreased compared to 8325-4 (23). Using an hld (RNAIII)-lacZ reporter gene fusion a sixfold reduction in maximal transcription was observed compared to 8325-4 or an SH1000 sigB mutant. Thus, the apparent reduction in extracellular virulence determinant production in SH1000 would appear to be the result of the down-regulation of the agr locus. Yet this cannot be a direct σ^B effect, as sigma factors can serve only to upregulate gene transcription. Therefore, either σ^B is upregulating a potent repressor of agr activity or it is indirectly down-regulating a potent activator of agr transcription.

Thus, we undertook transposon mutagenesis in order to identify components involved in this σ^B-mediated reduction in virulence determinant production. To achieve this, we developed two functional transposon screens, one designed to isolate components that affect hla transcription through the use of a hla::lacZ fusion and the other serving to identify components controlling extracellular proteolysis via the use of protease activity agar. A number of components were identified, and their role in the σ^B regulation cascade is investigated and discussed.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

S. aureus and Escherichia coli strains and plasmids are listed in Table 1. E. coli was grown in Luria-Bertani (LB) medium at 37°C. S. aureus was grown in 100 ml brain heart infusion (BHI; 1:2.5 flask/volume ratio) at 37°C with shaking at 250 rpm (Oxoid) (9), unless indicated otherwise. When required, antibiotics were added at the indicated concentrations: ampicillin, 100 mg liter⁻¹ (E. coli); CdCl₂, 250 mg liter⁻¹ (S. aureus); tetracycline, 5 mg liter⁻¹ (S. aureus); erythromycin (Ery), 5 mg liter⁻¹ (S. aureus); and lincomycin, 25 mg liter⁻¹ (S. aureus).

TABLE 1.

Strains, plasmids, and oligonucleotides used in this study

Strain, plasmid, or primer	Genotype or descriptiona	Reference(s) or source
Strains
E. coli DH5α	φ80 Δ(lacZ)M15 Δ(argF-lac)U169 endA1 recA1 hsdR17 (rK− mK+) deoR thi-1 supE44 gyrA96 relA1	48
S. aureus
8325-4	Wild-type strain cured of prophages	Lab stocks
RN4220	Restriction-deficient transformation recipient	Lab stocks
SH1000	Functional rsbU derivative of 8325-4 rsbU+	23
LES07	8325-4 pAZ106::sspA (sspA+)	23
LES08	SH1000 pAZ106::sspA (sspA+)	23
LES60	SH1000 pRN3208	This study
LES65	SH1000 telA::Tn551 pLES16	This study
LES68	SH1000 telA::Tn551 pLES16 pAISH::hla (hlA+)	This study
LES71	SH1000 telA::Tn551 pAISH::asp23 (asp23+)	This study
LES72	SH1000 lysA::Tn551 pAISH::asp23 (asp23+)	This study
LES73	SH1000 brnQ::Tn551 pAISH::asp23 (asp23+)	This study
LES74	8325-4 pAISH::asp23 (asp23+)	This study
LES75	SH1000 pAZ106::sigB (sigB+)	This study
LES76	SH1000 pAZ106::sigB (sigB+) telA::kan	This study
KS1	SH1000 pAISH::asp23 (asp23+)	This study
JLA512	8325-4 pAISH::hla (hlA+)	This study
JLA513	SH1000 pAISH::hla (hlA+)	This study
JLA541	SH1000 pRN3208 pAISH::hla (hlA+)	This study
JLA608	SH1000 pAZ106::lysC (lysC+)	This study
JLA609	MJH502 pAZ106::lysC (lysC+)	This study
JLA610	SH1000 pAZ106::lysA (lysA+)	This study
JLA611	MJH502 pAZ106::lysA (lysA+)	This study
J68	SH1000 pAZ106::telA (telA+)	This study
J69	SH1000 telA::kan	This study
JLA612	MJH502 pAZ106::telA (telA+)	This study
MC100	8325-4 pAZ106::sigB (sigB+)	Lab stocks
MJH502	SH1000 sigB::tet	23
Plasmids
pAISH1	Tetr derivative of pMUTIN4	2
pAZ106	Promoterless lacZ erm insertion vector	31
pOB	pGEM3Zf(+)-based cloning vector	23
pJA5	pAISH1 containing a 2.2-kb hla fragment	This study
pRN3208	TS shuttle vector harboring Tn551	32
pMK4	cm shuttle vector	54
pLES15	pMK4 containing a 253-bp telA promoter fragment	This study
pLES16	pLES15 containing a 1,273-bp telA complementation fragment	This study
pSIM02	pAISH containing a 630-bp asp23 fragment	This study
pJIM1	pOB containing a 4.3-kb telA fragment	This study
pJIM9	pAZ106 containing a 2.0-kb telA fragment	This study
pJIM68	pJIM1 containing a kanamycin resistance cassette in telA	This study
Primers
OLTN	ATAGAGAGATGTCACCGTCAAG
ASP23F	ACAAAGCTTAAGAATGTAGAATATTTCACA
ASP23R	ACAGGATCCATTAGTGAATGTATCTAA
OL330	ATGGAATTCGGTGCGATGCCGAAGCAA
OL331	CGCCCCGGGTTGTTAGGAAGGGCAGAG
OL332	CGCCCCGGGCGCTACCCGATACGTGCA
OL333	ATGGGATCCGAGTATATGTCCAAGCATG
OL334	ATGGAATTCCTGGCCAGTCACTGAATAACC
OL335	CGCCCCGGGGAACGCGTCCGAATACTTG
OL336	CGCCCCGGGGCGTCTCAACGAAGATGA
OL337	ATGGGATCCACTTAATGTCCCAAGCTC
OL338	ATGGGATCCGTTTCTGCTATGTACAAC
OL339	ATGGAATTCCCTAGGAAAGCGAAGCCATTC
JKL1	CAACGCATGCAGCTTGTGGACCGCGG
JKL3	GGCGGTACCATGTCTAACTACAGCATGCGG

^aRestriction sites are underlined.

Construction of hla and asp23-lacZ fusions using pAISH1.

pAISH1 is a derivative of pMUTIN4 (55) that has had the erythromycin (erm) resistance cassette exchanged for tetracycline (tet) (2). To generate the hla-lacZ fusion, a 2.2-kb EcoRI/NotI-flanked fragment (containing the 5′ region of the hla gene and upstream regulatory sequences) from the plasmid pPF4 (9) was cloned into pAISH1, producing pJA5. To generate the asp23-lacZ fusion, a PCR fragment using primer pair ASP23F-ASP23R (630 bp) was cloned into pAISH1 to produce pSIM02. Plasmids were then transformed into electrocompetent RN4220 (33, 49) and selected on BHI agar plates containing tetracycline. One clone from each was confirmed by Southern blotting before being used to generate a phage lysate using φ11. This was then used to transduce S. aureus SH1000, with clones being identified and confirmed. This created strains JLA513 (hla-lacZ hla⁺) and KS1 (asp23-lacZ asp23⁺).

β-Galactosidase assays.

Levels of β-galactosidase activity were measured as described previously (24). Fluorescence was measured using a Victor plate reader (Wallac) with an 0.1-s count time and calibrated with standard concentrations of 4-methylumbelliferone (MU). One unit of β-galactosidase activity was defined as the amount of enzyme that catalyzed the production of 1 pmol MU min⁻¹ unit of optical density at 600 nm (OD₆₀₀)⁻¹. Assays were performed on duplicate samples and the values averaged. The results presented here were representative of three independent experiments that showed less than 10% variability.

Transposon mutagenesis.

A single colony of S. aureus strain LES60 (SH1000 pRN3208) for protease screens or JLA541 (SH1000 pRN3208 pAISH::hla) for α-hemolysin screens was inoculated into 10 ml BHI containing CdCl₂ and erythromycin and grown overnight at 37°C with shaking (250 rpm). The OD₆₀₀ was determined for each, and this was used to inoculate 50 ml of prewarmed BHI (CdCl₂ and Ery) to a starting OD₆₀₀ of 0.02. These cultures were then grown at 37°C with shaking until OD₆₀₀ reached ∼1.0. Three-milliliter aliquots were removed, and cells were harvested by centrifugation before being resuspended in 100 ml BHI containing only erythromycin, prewarmed to 42°C. These cultures were then grown at this temperature, which is nonpermissive for replication of the plasmid, with shaking (250 rpm) until OD₆₀₀ reached 0.3 to 0.4. Three milliliters of each culture was then transferred to a further 100 ml BHI (Ery), again prewarmed to 42°C. Cultures were again grown at this elevated temperature and allowed to grow to stationary phase (18 h). After this time cells were harvested and resuspended in 4 ml BHI containing 10% (vol/vol) glycerol, giving an approximate CFU/ml of 5 × 10⁹. Cells were snap frozen in liquid N₂ and stored at −20°C. The efficiency of transposition was determined by serial dilution of cells in phosphate-buffered saline and plating onto BHI agar containing Ery alone or Ery and CdCl₂. The insertion frequency was calculated by comparing the number of colonies that grew on plates containing Ery with the number of colonies that grew on plates containing Ery and CdCl₂. A minimum of 95% insertion was found in each case.

DNA sequence analysis of transposon mutants.

In order to determine the exact site of transposon insertion, short single-read DNA sequencing reactions were carried out. Genomic DNA was isolated from clones using the QIAGEN DNeasy kit, and sequencing PCRs were carried out using Pwo polymerase (Roche) and oligonucleotide OLTN (Table 1). This primer hybridizes 68 bp inside the transposon Tn551, and the sequencing reaction proceeds into the flanking chromosomal DNA. Using this method 300 to 400 bp of useful chromosomal DNA sequence was obtained for all transposon mutants. This was then used to conduct BLAST analysis of the TIGR CMR database for S. aureus strain COL, allowing the exact determination of transposon insertion sites.

Protein analysis.

Extracellular and total protein sample preparation and analysis were conducted using 12% (wt/vol) sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described previously (9). Zymography was conducted as detailed by McAleese et al. (38). All protein samples are equivalent to 1.0 OD₆₀₀ units of original culture.

Western blotting.

Proteins were blotted onto a polyvinylidene difluoride membrane (Bio-Rad) and detected using antisera raised against either SspA or Hla. Antibodies were diluted 1:2,500 and used according to standard techniques (48). Alkaline phosphatase-conjugated goat anti-rabbit secondary antibody (diluted 1:30,000) was used to detect SspA or Hla colorimetrically.

Complementation of the telA transposon mutant.

telA (encoding a protein with homology to the tellurite resistance protein of Rhodobacter sphaeroides) is the third gene of a polycistronic operon that is almost certainly driven by a single promoter at the beginning of the locus. Thus, the technique of promoter fusion complementation was used as described previously (53) in order to generate a functional complementation construct. Firstly a promoter PCR product was generated using primer pair OL334-OL335 (253 bp) and cloned into pMK4 (54), creating pLES15. The functional telA gene was then PCR generated using primer pair OL336-OL337 (1,273 bp) and cloned into pLES15, creating complementation construct pLES16. This construct was then electroporated into RN4220, with correct clones being confirmed by restriction digestion. A clone was then used to generate a φ11 lysate which was used as a donor for transduction of the SH1000 telA transposon mutant, generating complemented strain LES65. This strain was also transduced with the hla-lacZ construct from JLA512, creating LES66.

Construction of lysC, lysA, and telA-lacZ fusion strains.

The lysC-lacZ (lysC encodes an aspartokinase) fusion construct was transduced (using φ11) from MDW41 (59) into SH1000 and MJH502 (SH1000 sigB) (23) to create strains JLA608 (lysC-lacZ hla⁺) and JLA609 (lysC-lacZ hla⁺ sigB), respectively. The lysA-lacZ (lysA encodes a diaminopimelate decarboxylase) construct was transduced from SPW2 (58) into SH1000 and MJH502 (SH1000 sigB) to create strains JLA610 (lysA-lacZ hla⁺) and JLA611 (lysA-lacZ hla⁺ sigB), respectively. A telA-lacZ fusion strain was created in this study. Primer pair JKL1-JKL3 was used to generate a 4.3-kb fragment extending from the cspB gene upstream of telA to the brnQ gene downstream of telA. This was cloned into pOB (23), creating pJIM1. A 2.0-kb telA fragment was excised from pJIM1 using SphI and BglII and cloned into similarly digested pAZ106 (31), creating pJIM9. This was then transformed into electrocompetent RN4220 (33, 49) and selected on BHI agar plates containing erythromycin. One clone was confirmed by Southern blotting before being used to generate a phage lysate using φ11. This was then used to transduce S. aureus SH1000 and MJH502 (SH1000 sigB), creating strains J68 (telA-lacZ hla⁺) and JLA612 (telA-lacZ hla⁺ sigB), respectively. Again clones were confirmed by Southern blot analysis.

Construction of a telA::kan cassette mutant.

A kanamycin resistance cassette was ligated into a naturally occurring BglII site (in the center of the telA gene) in pJIM1 (construction described above) to create pJIM68. This plasmid was electroporated into S. aureus RN4220 selecting for single-crossover events, with clones resistant to both kanamycin and erythromycin. The mutation was then transduced into SH1000, with clones selected for kanamycin resistance and erythromycin sensitivity, creating strain J69, which was confirmed by Southern blotting.

RESULTS

The modulation of environmental growth conditions cannot relieve the σ^B-mediated repression of hla and sspA.

Using lacZ reporter gene fusions for both hla (α-hemolysin) and sspA (V8 protease), we undertook to investigate the role of environmental conditions in the expression of these two important virulence determinants. These studies were conducted in both 8325-4 and SH1000 in an attempt to understand whether environmental conditions could act as a signal to relieve the potential transcriptional repression of sspA and hla mediated by σ^B in SH1000. The conditions tested were increased aeration (1:10 versus 1:2.5) or temperature (43°C) and growth in 1 M NaCl, 20 mM sucrose, or 5% (vol/vol) ethanol. Standard condition transcriptional analysis for these two operons (37°C, 1:2.5 aeration) was also conducted in parallel for comparative purposes. Under none of the tested conditions was a relative increase in transcription of hla or sspA observed in SH1000 compared to 8325-4 (data not shown). Interestingly the salt-mediated increase in transcription of sspA observed by Lindsay and Foster (36) is not apparent in SH1000, while in 8325-4 a minimum of a twofold upregulation is observed.

The development of functional transposon screens to identify genetic components involved in the σ^B-mediated repression of α-hemolysin and protease synthesis.

Environmental signaling does not appear to explain the profound differences in virulence determinant expression between 8325-4 and SH1000. As this effect is observed at the level of transcription, we undertook transposon mutagenesis in SH1000 in order to identify specific effector components involved in this regulatory process. Functional screens were thus developed for both α-hemolysin expression and protease activity and were used in conjunction with Tn551 mutagenesis. S. aureus SH1000 produces no detectable proteolytic activity compared to 8325-4 (23, 52). The use of growth medium containing 1% (wt/vol) dried skimmed milk has been documented previously (28) as a measure of proteolytic activity of individual bacterial colonies. Analysis of the relative proteolytic activity of 8325-4 and SH1000 using BHI protease agar revealed no detectable proteolysis associated with SH1000 but significant amounts (7- to 10-mm zones of clearing) for strain 8325-4 (data not shown). Therefore, SH1000 Tn551 libraries (∼200 colonies per plate) were grown overnight on protease agar plates containing 1% (wt/vol) dried skimmed milk with clones being analyzed for the presence of proteolytic zones of clearing.

To identify components involved in the reduction of hla expression in SH1000, an hla-lacZ fusion was constructed using the tetracycline-marked suicide vector pAISH (see Materials and Methods). Analysis of the expression of hla in SH1000 using this strain revealed negligible levels of hemolysin expression when using X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) as a substrate (data not shown). Thus, transposon insertions within genes that result in upregulation of hla transcription would lead to a blue coloration of colonies, akin to that observed in 8325-4. Thus, SH1000 Tn551 libraries were generated containing this hla-lacZ fusion, incubated on BHI agar plates containing X-Gal, and analyzed for the presence of blue colonies.

Identification of genes involved in the putative σ^B-mediated repression of α-hemolysin and protease synthesis.

In total 15,000 colonies were analyzed in each transposon screen. Thirty-seven colonies were identified that demonstrated extracellular proteolytic activity, and 13 colonies were identified demonstrating an increase in α-hemolysin expression. Interestingly those mutants that demonstrated an increase in hla transcription also demonstrated an increase in extracellular proteolysis (Fig. 1A). Moreover, those clones that were identified on the basis of an increase in protease production also demonstrated an increase in hla transcription when transduced into JLA513 (SH1000 hla::lacZ) (Fig. 1B). Each of the transposon insertion mutants was then lysed using phage φ11 and transduced back to SH1000 in order to confirm absolute linkage of the observed phenotype to the transposon insertion. All mutations identified in the hla::lacZ screen were then transduced back into JLA513 (hla::lacZ). One hundred percent cotransduction of proteolysis or hla expression alteration with the insertion was observed with the 50 mutants. Single-read DNA sequencing was undertaken, and the derived sequence was used to identify the specific location of transposon insertion for each of the mutants (Table 2).

FIG. 1.

Extracellular proteolysis and hla transcription of transposon mutants. (A) Strains were patched onto BHI agar containing 1% (wt/vol) dried skimmed milk and incubated overnight to measure extracellular proteolysis. Strain order: 1, SH1000; 2, MJH502 (SH1000 sigB); 3, 8325-4; 4, SH1000 Tn551::lysC; 5, SH1000 Tn551::asd; 6, SH1000 Tn551::cspA; 7, SH1000 Tn551::ykuQ; 8, SH1000 Tn551::hipO; 9, SH1000 Tn551::lysA; 10, SH1000 Tn551::aclP; 11, SH1000 Tn551::telA; 12, SH1000 Tn551::brnQ; 13, SH1000 Tn551::norQ. (B) Strains were patched onto BHI agar containing X-Gal and incubated overnight to measure transcription of hla. Strain order: 1, JLA513 (SH1000 hla::lacZ); 2, JLA373 (SH1000 hla::lacZ sigB); 3, PC322 (8325-4 hla::lacZ); 4, SH1000 Tn551::lysA hla::lacZ; 5, SH1000 Tn551::ykuQ hla::lacZ; 6, SH1000 Tn551::hipO hla::lacZ; 7, SH1000 Tn551::lysA hla::lacZ; 8, SH1000 Tn551::aclP hla::lacZ; 9, SH1000 Tn551::telA hla::lacZ; 10, SH1000 Tn551::brnQ hla::lacZ; 11, SH1000 Tn551::norQ hla::lacZ.

TABLE 2.

Identification of transposon insertion sites

Inactivated genea	Function	Insertion pointb	No. identifiedc
SACOL0482	Conserved hypothetical protein	99 bp upstream	1
SACOL1006	Conserved hypothetical protein	262/804	4 (2)
SACOL1428 (lysC)	Aspartokinase	327/1428	1
SACOL1432 (ykuQ)	Tetrahydrodipicolinate acetyltransferase	256/717	2 (2)
SACOL1433 (hipO)	Hippurate hydrolase	221/1433	2
SACOL1435 (lysA)	Diaminopimelate decarboxylase	Multiple internal	17 (3)
SACOL1438 (cspA)	Cold shock protein A	13 bp upstream	7
SACOL1439 (aclP)	Acylphosphatase	29 bp upstream	1
SACOL1441 (telA)	Tellurite resistance	882/1441	4 (1)
SACOL1443 (brnQ)	Branched-chain amino acid transporter	1221/1341	2 (1)
SACOL1445 (norQ)	Unknown function	19/789	1
SACOL1947	Conserved hypothetical protein	619/984	1
SACOL2057 (rsbU)	σB-regulating phosphatase	278/999	7 (4)

^aGene number and annotation (in parentheses) are from S. aureus COL genome at TIGR (http://www.tigr.org/cmr).

^bLocation of transposon insertion in the respective open reading frames in relation to the length of the entire gene.

^cNumber of independent mutations for each gene. The number of clones from the hla-lacZ screen is shown in parentheses.

From this list of 50 mutants 41 of them had insertions within open reading frames, with a further nine insertions intergenically between coding regions. Regarding the 41 open reading frame insertions it was apparent that these represented only 10 different genes, while the nine intergenic insertions accounted for only three different loci. Importantly insertions within the rsbU loci were identified within both screens, adding credibility to the efficacy of the study. Those loci identified on multiple occasions were not always the result of insertion at exactly the same site. For example multiple insertion sites were observed for SACOL1006, SACOL1435 (lysA), SACOL1438 (cspA), and SACOL2057 (rsbU). However, the same insertion site was found for each of the SACOL1432 (ykuQ), SACOL1433 (hipO), and SACOL1441 (telA) clones, suggesting that these were likely to be siblings. Those insertions that were intergenic were all within a very short distance of the initiation codon of the respective locus. Thus, the insertions within aclP (29 bp upstream) and cspA (13 bp upstream) appear to directly interrupt putative transcriptional elements that precede these genes (data not shown).

Localization of transposon insertions reveals a Tn551 hot spot.

Analysis of the location of transposon insertions led to the observation that many of the genes are clustered together within a particular region of the S. aureus genome (i.e., between gene numbers SACOL1428 and SACOL1446). Indeed, the genes lysC, ykuQ, hipO, and lysA form part of the dap operon (59), while telA and aclP insertions are also contained within the same transcriptional unit. This is in turn contiguous to the cspA insertions, which reside close to norQ and brnQ, which are separated by only one gene. Interestingly the dap operon is clustered near the tel operon, which in turn is located closely to the brnQ/norQ operons. Each operon is separated from the next by a region of approximately 1 kb in each case. Indeed, with the exception of rsbU and the conserved hypothetical proteins identified all of the transposon insertions identified in this screen are located within a 35-kb region of the S. aureus chromosome (Fig. 2). Interestingly, transposon insertions within genes SACOL1423 and SACOL1452 (lab stocks), which flank this region of the chromosome, do not demonstrate the same proteolytic and hemolytic phenotype as the mutants identified in this screen (data not shown). Thus, these genes could represent boundaries of this hot spot region.

FIG. 2.

Genetic organization of the hot spot region for Tn551 insertion yielding increased exoprotein production in SH1000. Genes identified in this screen are shaded dark gray; those uninvestigated are shaded light gray. Those shaded black contain transposon insertions (from lab stocks) which do not demonstrate the increased proteolytic and hla expression phenotype in the SH1000 background, suggesting a boundary of the hot spot region.

It is curious that transposon insertions within genes of apparently unrelated function in a clustered region of the chromosome appeared to have almost identical mutant phenotypes. Therefore, it was important to ensure that the phenotypes observed for mutants within this hot spot region of the chromosome were not a result of altered DNA topology resulting from the insertion of exogenous DNA. This was achieved by patching the SH1000 lysC transposon mutant onto protease agar alongside JLA608 (SH1000 lysC::lacZ lysC⁺). Whereas lysC has been insertionally inactivated in the transposon mutant, JLA608 possesses a functional copy of lysC, as well as a promoterless β-galactosidase gene fused to an additional, recombinationally duplicated lysC promoter. JLA608 (SH1000 lysC::lacZ lysC⁺) did not demonstrate the same increase in protease activity as did the transposon insertion mutant, suggesting that it is indeed the physical disruption of gene function that has caused the phenotype (data not shown).

Transposon insertions result in SH1000 returning to an 8325-4-like phenotype.

Having determined that the transposon mutants identified demonstrated a genuine phenotype of increased proteolysis and hla expression, selected representative clones were then subjected to further analyses. Exoprotein analysis was conducted on all mutants in conjunction with Western blotting using anti-SspA and anti-Hla antibodies (Fig. 3). In each case a distinct and clear increase in the amount of exoproteins, and the amount of SspA and Hla secreted, was observed compared to the parental strain, SH1000. It should be noted, however, that this increase in exoprotein production was not uniform across all transposon mutants. It is apparent that some mutations produce a more profound phenotype than others do. For example rsbU mutation results in an 8325-4-like exoprotein profile while lysA mutation results in only a minor alteration from the SH1000 profile. A further variable characteristic was the presence and the amount of pro-SspA as detected by antibodies. In low-protease-producing strains such as SH1000 it is common for SspA to exist in equal amounts in its processed and unprocessed forms (52); however, high-protease-producing strains tend to show a massive increase in the amount of active protease and a very small amount of proprotease. This is seen in the Western blot with anti-SspA antibodies, where SH1000 and lysA, for example, display a clear doublet that reacts with the anti-SspA antibody. In contrast the rsbU transposon mutant shows a great deal more protein and considerably more active SspA.

FIG. 3.

Analysis of the secreted proteins of selected transposon mutants. (A) Exoprotein analysis. (B) Western blot analysis using anti-SspA antibodies. (C) Western blot analysis using anti-Hla antibodies. Strain order: 1, SH1000 Tn551::lysA; 2, SH1000 Tn551::hipO; 3, SH1000 Tn551::lysC; 4, SH1000 Tn551::telA; 5, SH1000 Tn551::aclP; 6, SH1000 Tn551::cspA; 7, SH1000 Tn551::norQ; 8, SH1000 Tn551::brnQ; 9, SH1000 Tn551::rsbU; 10, SH1000 Tn551::ykuQ; 11, SH1000 Tn551::SACOL1947; 12, SH1000. All protein samples are equivalent to 1.0 OD₆₀₀ units of original culture.

Zymogram analysis of the transposon mutants further demonstrated that this increase in proteolytic activity, importantly, was not confined to the SspA protease alone but was true for all extracellular proteases (Fig. 4). Once again, however, this phenotype was not identical among mutants, with strains such as the rsbU and telA strains demonstrating highly proteolytic profiles, while other such as the lysA and hipO strains demonstrated less activity. This variable phenotype was further qualified by analysis of the coloration of strains. The restoration of full σ^B activity in SH1000 results in an increase in pigmentation of the strain compared to its parent 8325-4 (23). Thus, each of the mutants was patched onto an agar plate along with SH1000 and documented for pigmentation (Fig. 5). As can be seen, the intensity and degree of coloration among strains vary considerably, with strains such as the ykuQ and brnQ strains displaying a more orange and SH1000-like phenotype, while strains such as the telA and norQ strains are less pigmented.

FIG. 4.

Extracellular proteolytic activity of the identified transposon mutants. Arrows indicate the activity of each protease, as determined previously (51). Strain order: 1, SH1000 Tn551::lysA; 2, SH1000 Tn551::hipO; 3, SH1000 Tn551::lysC; 4, SH1000 Tn551::telA; 5, SH1000 Tn551::aclP; 6, SH1000 Tn551::cspA; 7, SH1000 Tn551::norQ; 8, SH1000 Tn551::brnQ; 9, SH1000 Tn551::rsbU; 10, SH1000 Tn551::ykuQ; 11, SH1000 Tn551::SACOL1947; 12, SH1000.

FIG. 5.

Effect of transposon insertions on pigmentation. Strains were patched onto BHI agar and incubated overnight in order to visualize the colony pigmentation. Strain order: 1, SH1000; 2, MJH502 (SH1000 sigB); 3, SH1000 Tn551::lysC; 4, SH1000 Tn551::asd; 5, SH1000 Tn551::cspA; 6, SH1000 Tn551::ykuQ; 7, SH1000 Tn551::hipO; 8, SH1000 Tn551::lysA; 9, SH1000 Tn551::aclP; 10, SH1000 Tn551::telA; 11, SH1000 Tn551::brnQ; 12, SH1000 Tn551::norQ.

Complementation analysis of a representative transposon mutant.

In order to confirm that the phenotypes observed here were a direct result of the loss of function of the genes in question, a telA mutant was selected for complementation analysis. This mutation was selected for further analysis based on the phenotypic analysis, which revealed it to have the most pronounced and 8325-4-like phenotype of the SH1000 transposon mutants characterized in this study. The telA gene was cloned in the S. aureus shuttle vector pMK4 and reintroduced in trans into the transposon mutant. The complemented strain revealed, when patched onto protease agar plates, that the increased proteolysis observed in the transposon screen had indeed disappeared and that the strain had returned to an SH1000-like phenotype (data not shown). Moreover complementation of the telA mutant with in trans telA resulted in the return of full, SH1000-like pigmentation to the strain. In order to obtain quantitative evidence that complementation caused phenotype reversion, the telA-complemented transposon mutant was combined with the hla-lacZ reporter fusion and found to be decreased in hla expression when analyzed for LacZ activity during growth (Fig. 6). Indeed, complementation of telA led to a reduction in expression very closely resembling the levels seen for the parental strain, SH1000.

FIG. 6.

Complementation analysis of the telA transposon mutant of SH1000. Assays of hla-lacZ fusion activity were conducted during growth of SH1000 (▪), 8325-4 (⧫), SH1000 Tn551::telA (▴), and SH1000 Tn551::telA pMK4::telA (•). Cultures were grown at 37°C with shaking, and samples were taken at the times indicated. Specific β-galactosidase activity was determined. Results are representative of at least three separate experiments.

Relief of virulence determinant repression may result from altered σ^B activity in the SH1000 transposon mutants.

In order to determine the specific interplay between the identified genes and σ^B, transcriptional analysis of representative genes was undertaken in the 8325-4 and SH1000 backgrounds. Using lacZ reporter gene fusions for lysA, lysC, and telA, we were able to demonstrate that there is no alteration in the levels of expression of each of these genes in the presence or absence of a functional σ^B (data not shown). Thus, if σ^B does not affect the transcription of these genes, then it is possible that they in turn affect σ^B, putatively at the level of its activity. This was determined via the measurement of expression of asp23, a gene whose transcription is known to be solely controlled by σ^B (21). An asp23::lacZ fusion was created using pAISH1 and introduced into the lysA, brnQ, and telA representative mutant strains. These strains were then assayed during growth for asp23 expression (Fig. 7A).

FIG. 7.

Effect of mutations on asp23 expression. (A) Assay of transcription from asp23-lacZ fusions during growth. Strains SH1000 (▪), SH1000 Tn551::lysA (•), SH1000 Tn551::brnQ (▴), SH1000 Tn551::telA (⧫), and 8325-4 (○) were grown at 37°C with shaking, and samples were taken at the times indicated for β-galactosidase activity. The results are representative of at least three separate experiments. MUG, methylumbelliferyl-β-glucuronide. (B) Whole-cell proteins were prepared and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The identity of Asp23 (arrow) was determined by N-terminal sequence analysis (VDNNXAXQAYDQX). Strains: 1, SH1000; 2, MJH502 (SH1000 sigB); 3, SH1000 Tn551::lysC; 4, SH1000 Tn551::telA; 5, SH1000 Tn551::brnQ.

Transcription of asp23 in the lysA, brnQ, or telA mutants is altered compared to both SH1000 and 8325-4. In the lysA mutant strain the level of asp23 transcription is close to that of SH1000, indicating that the activity of σ^B is relatively high in this strain. However, the amount of asp23 transcription in the brnQ and telA strains is reduced between three- and fourfold from SH1000 (2,032 [SH1000], 746 [brnQ], and 433 [telA] maximal methylumbelliferyl-β-glucuronide units at T = 3 h), indicating that the activity of σ^B is relatively low in these strains. This was further corroborated by analysis of cytoplasmic proteins and N-terminal sequencing which revealed the presence of Asp23 in an extract of SH1000 (Fig. 7B). The relative amounts of Asp23 protein in the transposon mutants were reduced compared to SH1000 (Fig. 7). To further support this hypothesis, a telA mutant was generated so that we might pair it with our existing sigB-lacZ fusion, in order to specifically determine that the effects seen with asp23 did indeed relate to altered sigB transcription. In an SH1000 telA::kan sigB::lacZ strain, transcription of sigB was reduced by approximately fourfold (15,429 [SH1000] and 4,169 [telA] maximal methylumbelliferyl-β-glucuronide units; data not shown). Thus, the variable phenotype and σ^B activity levels within each of these mutants suggest that the genes identified modulate the activity of σ^B and thus as a result the levels of extracellular protein synthesis.

DISCUSSION

Virulence determinant production in S. aureus occurs in a highly coordinated fashion in response to environmental conditions, requiring the interplay of multiple genetic regulators. It has been observed that the 8325-4 strain lineage used by many labs has a deletion in the rsbU gene (20, 34) and that repair of this deletion results in a dramatic decrease in exoprotein production (7, 23). RsbU is part of the regulatory network that controls the activity of the alternative sigma factor, σ^B (21). σ^B has been shown to be a pleiotropic regulator of virulence determinant production and to be required for pathogenesis in a septic arthritis model (23, 26). We have previously created an unmarked derivative of 8325-4 with a repaired rsbU gene called SH1000. Characterization of SH1000 revealed that the level of agr transcription was markedly decreased compared to 8325-4, apparently independently of SarA (23). In other work in our laboratory we have also shown that this effect is not mediated through two other major regulators, Rot and SarHI (unpublished observations). Thus, the specific mechanism or effector of this σ^B-mediated repression of agr, and consequently the agr regulon (including hla and sspA), remains unidentified.

Bacillus subtilis possesses a number of genes involved in upregulating σ^B activity (rsbP, rsbQ, rsbR, rsbS, rsbT, and rsbX) that are absent from the S. aureus genome (27, 60, 61). In B. subtilis, the Rsb proteins regulate σ^B activity in response to signals detected by two separate pathways; RsbP and RsbQ constitute the energy stress pathway, while RsbS, RsbT, RsbR, RsbX, and RsbU form part of the environmental stress pathway (reviewed in reference 45). In this latter example RsbSTRX, in response to environmental signals, modulate the phosphorylation status of RsbU, which in turn serves to modulate σ^B activity. As S. aureus lacks the RsbSTRX signaling pathway, it is reasonable to assume that another, as yet unidentified mechanism may exist to modulate the phosphorylation status of RsbU. Thus, we undertook analysis of the transcription of hla and sspA in cultures grown under differing environmental conditions in order to determine whether it was possible that environmental signals could modulate RsbU and thus in turn σ^B. In all conditions tested we found no stimulation of hla or sspA transcription in SH1000 compared to 8325-4. Therefore, we undertook two independent, functional transposon screens in order to identify components in SH1000 involved in the σ^B regulatory cascade. One screen involved an hla::lacZ fusion and was focused at the level of transcription, while an extracellular proteolysis screen was directed at the level of protein activity. A number of components were identified in both screens, mutations in which led to both an increase in transcription of hla and extracellular proteolytic activity. This commonality is no surprise, as exoprotein repression by σ^B is mediated through reduced agr activity, and both hla and the extracellular proteases are regulated in a similar, agr-dependent manner (8, 9, 10, 36, 47, 52).

Mapping of the transposon insertions revealed that rather than identifying known or putative global regulatory loci, we had in fact found a series of unrelated, metabolic genes, all of which are involved in the repression of virulence determinant production in SH1000. ykuQ, lysC and lysA, and hipO are all located within the dap operon (59). The dap operon consists of eight genes, six of which are involved in the biosynthesis of the amino acids lysine, methionine, threonine, and isoleucine. aclP encodes an acylphosphatase, an enzyme responsible for the cleavage of carboxyl-phosphate bonds in acylphosphates (40), which is thought to have a role in the glycolytic pathway and also in pyrimidine biosynthesis (44). Although TelA is homologous (33% identity, 106/317 amino acids) to the tellurite resistance protein of Rhodobacter sphaeroides, no such similar role for it has been identified in S. aureus (J. K. Lithgow and S. J. Foster, unpublished data). telA was originally identified in a starvation survival recovery transposon screen (58), and the mutant was shown to be incapable of recovery after prolonged starvation of carbon, amino acids, and glucose as well as possessing an increased sensitivity to acid stress. The other genes of discernible function identified were norQ, which displays homology to nitrate reductases from a variety of organisms; brnQ, which encodes a branched-chain-amino-acid transporter (56); and cspA, which specifies one of the three S. aureus cold shock proteins (29). Interestingly this last gene has been characterized in a recent study where it was found that mutations of cspA in the COL background result in decreased expression of the S. aureus pigmentation gene crtN and, more importantly, in sigB itself (30). The remaining genes either are conserved hypothetical proteins with unknown functions (SACOL0482, SACOL1006, and SACOL1947) or are known to be involved in regulation (rsbU).

Thus, we have a series of seemingly unrelated genes, mutations in which in SH1000 result in relief of repression of hla and extracellular proteases. Yet through the tests employed within this study we have confirmed that this collection of genes does indeed play a very real role in the regulatory process. One of the most interesting observations was that the phenotypes of all of these mutants, while similar, were not identical. It became apparent that even though they all produced a similar final result (i.e., increased hla transcription and extracellular proteolysis), the degree and severity of these phenotypes appeared variable between mutants. Thus, if each of these were acting by exactly the same mechanism along a defined pathway, it would be expected that they would produce an identical unwavering phenotype in each example. Investigations into the regulation of representative examples of these transposon mutant genes revealed that σ^B apparently plays no role in their transcription. Interestingly we have shown that these genes, or rather the lack of them in the transposon mutants, result in reduced σ^B activity. Thus, it would appear that rather than serving as σ^B-modulated components, they in fact impinge upon the regulatory cascade upstream of σ^B. As none of the identified proteins possesses any DNA binding motifs, any repression of σ^B activity must occur in an indirect manner.

As alluded to earlier, S. aureus lacks all of the σ^B environmental signaling components found in B. subtilis. Thus, it is possible that in the absence of this extended environmental sensing network the genes identified in this screen may replace the functions of the missing rsb genes. While this is an interesting proposal, in the absence of specific proof we must assume that all σ^B activity signaling in S. aureus is routed through the phosphatase RsbU. Yet the specific phosphorylation state of RsbU still requires mediation in order for σ^B to be discretely regulated. How this occurs, and in response to what, is unknown, but it is clear that there are as yet unidentified factors at work. Therefore, we propose that it is here, at the specific level of the control of the RsbU phosphorylation state, that the effect of these transposon mutants is exerted (Fig. 8). It is unlikely that they act directly to influence this; however, it is possible that they destabilize the precise balance of the cell, preventing full signaling through RsbU. Indeed, they are all metabolic genes and must each contribute to the general fitness and survival of the cell. Any decrease in the activity of RsbU, and concomitantly RsbV, results in the formation of the RsbW-SigB complex and blocks σ^B function (45). Importantly the mutations do not fully block the action of RsbU, or in exactly the same way, as seen by the variable phenotype displayed.

FIG. 8.

Model of the σ^B regulatory network and its impact on virulence determinant expression. (Adapted with permission from the work of Senn et al. [51].) “X” refers to the proposed missing mediator of the agr repression effect in SH1000.

Our screens have identified several genes within a specific chromosomal region with roles in the control of σ^B activity. Many of these genes have been identified in previous transposon screens using Tn551 and Tn917 (17, 35, 39, 57, 58, 59). In fact the region has been shown to form a hot spot for insertion of these two transposons (4). We have excluded the possibility that the observed phenotypes are artifactual or are the result of distortion of DNA topology. Other screens have shown roles for members of the hot spot region in pathogenesis, virulence determinant regulation, or starvation survival (17, 35, 39, 57, 58, 59). Consequently, it is possible that this region of the genome may represent a cluster of genes of fundamental importance to the virulence and survival of S. aureus.