Role of BrnQ1 and BrnQ2 in Branched-Chain Amino Acid Transport and Virulence in Staphylococcus aureus

Abstract

The branched-chain amino acids (BCAAs; Ile, Leu, and Val) not only are important nutrients for the growth of Staphylococcus aureus but also are corepressors for CodY, which regulates virulence gene expression, implicating BCAAs as an important link between the metabolic state of the cell and virulence. BCAAs are either synthesized intracellularly or acquired from the environment. S. aureus encodes three putative BCAA transporters, designated BrnQ1, BrnQ2, and BrnQ3; their functions have not yet been formally tested. In this study, we mutated all three brnQ paralogs so as to characterize their substrate specificities and their roles in growth in vitro and in vivo. We demonstrated that in the community-associated, methicillin-resistant S. aureus (CA-MRSA) strain USA300, BrnQ1 is involved in uptake of all three BCAAs, BrnQ2 transports Ile, and BrnQ3 does not have a significant role in BCAA transport under the conditions tested. Of the three, only BrnQ1 is essential for USA300 to grow in a chemically defined medium that is limited for Leu or Val. Interestingly, we observed that a brnQ2 mutant grew better than USA300 in media limited for Leu and Val, owing to the fact that this mutation leads to overexpression of brnQ1. In a murine infection model, the brnQ1 mutant was attenuated, but in contrast, brnQ2 mutants had significantly increased virulence compared to that of USA300, a phenotype we suggest is at least partially linked to enhanced in vivo scavenging of Leu and Val through BrnQ1. These data uncover a hitherto-undiscovered connection between nutrient acquisition and virulence in CA-MRSA.

INTRODUCTION

Staphylococcus aureus is a highly successful human pathogen that succeeds at infecting virtually every body site, causing skin, soft tissue, respiratory, bone, joint, and endovascular infections (1). Maintenance of metabolic homeostasis is important for its infection process, as the majority of genes necessary for infection identified in large-scale signature-tagged mutagenesis (STM) screens fall into the categories of metabolism, transport, and biosynthesis (2, 3). Acquisition of host-derived nutrients, specifically amino acids, appears to be an important mechanism of meeting nutritional needs, as a large proportion of attenuated STM strains contain mutations in amino acid transporters (3).

Amino acid transporters are ubiquitous in bacteria and are typically selective for transporting either one amino acid or several amino acids with related structures (4). The branched-chain amino acids (BCAAs; Ile, Leu, and Val) are hydrophobic amino acids typically found in the core in globular proteins or in the transmembrane domain in cell surface proteins (5). The transport mechanisms described for acquisition of the BCAAs in bacteria include secondary transporters belonging to the Leu, Ile, Val:cation symporter (LIVCS) family (e.g., BrnQ, BraB, BraZ, and BcaP) (4, 6–9) and the LIV-I ABC transporter (10–13). With the exception of an ABC transporter in Streptococcus pneumoniae (14), secondary transport, which couples the movement of an ion, usually Na⁺ or H⁺, down its concentration gradient with the movement of another molecule against its concentration gradient, is the primary means of BCAA acquisition in Gram-positive bacteria.

S. aureus needs BCAA transport for growth, as it exhibits an auxotrophic phenotype for Leu and Val, despite possessing the genes necessary for their biosynthesis (15–17). The mechanisms of BCAA acquisition, however, have not been described for this species. Strategies to maintain intracellular levels of BCAAs in S. aureus are of additional interest since the BCAAs also act as corepressors of the global transcriptional regulator CodY (18–20). CodY, a highly conserved regulatory protein in low-G+C Gram-positive bacteria, uses binding of BCAAs and GTP to sense the metabolic status of the cell, the downstream effect of this being the adaptation of the cell to nutrient limitation (21, 22). CodY is activated through direct interaction with the BCAAs as well as GTP (18–20, 23) and, in its active state, represses transcription of upwards of 100 genes (17, 24, 25). Depletion of BCAAs and GTP results in derepression of CodY target genes, the products of which are involved in a range of cellular processes depending on the species, including sporulation (23), biofilm formation (26, 27), protein degradation and utilization (22), and amino acid metabolism and transport (17, 24). More recently, CodY has emerged as a regulator of virulence in S. aureus (28–30), implicating CodY as an important link between the nutrient status of the cell and virulence.

Recent characterization of the CodY regulon in S. aureus revealed amino acid metabolism and transport genes as a major class of CodY targets, including the brnQ1 and brnQ2 BCAA transporter genes (17, 24). We therefore hypothesized that BCAA transport is involved in adaptation to nutrient limitation and might also serve as an important mechanism to maintain intracellular pools of BCAAs that affect CodY activity and, in turn, regulate virulence of S. aureus. Despite the potential importance of BCAA transport in regulating virulence in S. aureus, no BCAA transporter has, as yet, been functionally characterized for a role in BCAA uptake. In this study, we sought to characterize the transport function of the brnQ genes in S. aureus and investigate their role in nutrient acquisition and virulence in a murine model of S. aureus infection. Our data indicate that the brnQ genes have evolved distinct BCAA transport specificities and, consequently, differentially affect the fitness of S. aureus in vivo.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Bacterial strains used in this study are described in Table 1. Methicillin-resistant S. aureus (MRSA) isolate pulsed-field gel electrophoresis type USA300 LAC that has been cured of the erythromycin resistance plasmid was used in all experiments as the wild-type (WT) strain. S. aureus strains were grown either in tryptic soy broth (TSB) or in a chemically defined medium (CDM), described previously (31). Complete CDM was composed of the following (final concentrations, μM): alanine (672), arginine (287), aspartic acid (684), cysteine (166), glutamic acid (680), glycine (670), histidine (129), isoleucine (228), leucine (684), lysine (342), methionine (20), phenylalanine (240), proline (690), serine (285), threonine (260), tryptophan (50), tyrosine (275), valine (684), thiamine (56), nicotinic acid (10), biotin (0.04), pantothenic acid (2.3), MgCl₂ (1,000), CaCl₂ (100), monopotassium phosphate (40,000), dipotassium phosphate (14,700), sodium citrate dehydrate (1,400), magnesium sulfate (400), ammonium sulfate (7,600), and glucose (27,753). The concentrations of individual amino acids were modified in some experiments as indicated. Where required, chloramphenicol (10 μg ml⁻¹), erythromycin (3 μg ml⁻¹), ampicillin (100 μg ml⁻¹), and tetracycline (4 μg ml⁻¹) were added to growth media. All growth curves were performed at 37°C with shaking using a flask/volume ratio of at least 7:1. For growth curves performed in TSB, S. aureus strains were pregrown for 8 h in TSB and subcultured into fresh TSB to a starting optical density at 600 nm (OD₆₀₀) equivalent of 0.0025. For growth curves performed in CDM, S. aureus strains were pregrown for 8 h in complete CDM and subcultured into fresh CDM to a starting OD₆₀₀ equivalent of 0.0025; as indicated in Results, either complete CDM or CDM with altered concentrations of amino acids was used. The optical densities of the cultures were measured until stationary phase was reached or until desired time points were reached.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Source or reference
Strains
S. aureus
USA300	USA300 LAC cured of antibiotic resistance plasmid	This study
RN4220	rK− mK+; capable of accepting foreign DNA	46
RN6390	Prophage-cured laboratory strain	47
Newman	WT clinical isolate	48
H2324	Newman brnQ3::Tet; Tetr	This study
H2568	USA300 ΔbrnQ1 (SAUSA300_0188)	This study
H2563	USA300 ΔbrnQ2 (SAUSA300_0306)	This study
H2578	USA300 brnQ3::Tet; Tetr (SAUSA300_1300)	This study
H2589	USA300 ΔbrnQ1 ΔbrnQ2 brnQ3::Tet; Tetr	This study
H2994	USA300 brnQ2::ΦNΣ; Eryr	41
H3001	USA300 codY::ΦNΣ; Eryr	41
E. coli
DH5α	F− ϕ80dlacZΔM15 recA1 endA1 gyrA96 thi-1 hsdR17(rK− mK−) supE44 relA1 deoR Δ(lacZYA-argF)U169 phoA	Promega
Plasmids
pKOR1	E. coli/Staphylococcus shuttle vector allowing allelic replacement in S. aureus	32
pDG1513	Antibiotic resistance cassette	49
pMAD	Temperature-sensitive E. coli/S. aureus shuttle vector	50
pRMC2	Anhydrotetracycline-inducible expression vector; Apr in E. coli; Cmr in S. aureus	34
pSO1	pRMC2 containing brnQ1; Cmr	This study
pSO2	pRMC2 containing brnQ2; Cmr	This study
pSO3	pRMC2 containing brnQ3; Cmr	This study

^aAbbreviations: Tet^r, Cm^r, Ery^r, and Ap^r, resistances to tetracycline, chloramphenicol, erythromycin, and ampicillin, respectively.

Mutagenesis of brnQ genes and complementation of mutants.

Deletions of the brnQ1 and brnQ2 genes were constructed using the pKOR-1 plasmid as described previously (32). Primer sequences were based on the published USA300 FPR3757 genome and are displayed in Table S1 in the supplemental material (33). Briefly, sequences flanking the brnQ locus were PCR amplified and a PCR amplicon of the joined DNA fragments was recombined into the temperature-sensitive vector pKOR1 using the BP Clonase reaction (Invitrogen), resulting in the pKOR-1ΔbrnQ1 and pKOR-1ΔbrnQ2 plasmids. These plasmids were first passaged through S. aureus RN4220 before electroporation into strain USA300. Precise in-frame allelic replacement of the brnQ genes was established by a two-step process of temperature shifting and antisense counterselection and confirmed by PCR and DNA sequencing. The brnQ3::Tet knockout allele consisted of a tetracycline resistance cassette, excised from plasmid pDG1513 with restriction enzyme PstI, flanked by DNA sequences homologous to regions upstream and downstream of brnQ3, leaving the first 69 bp of the start of the gene and the last 34 bp of the gene. The knockout allele was cloned into the temperature-sensitive Escherichia coli/S. aureus shuttle vector pMAD and then passaged through S. aureus RN4220 at permissive temperatures prior to being transduced using phage 80α into S. aureus RN6390. Recombinant RN6390 was cultured at 30°C to mid-log phase before the incubation temperature was shifted to 42°C. The bacteria were further incubated for 16 h before being plated onto tryptic soy agar (TSA) containing tetracycline. Colonies were screened for sensitivity to erythromycin, which indicates a loss of the pMAD backbone DNA after integration of the knockout allele into the S. aureus chromosome via double homologous recombination. The brnQ3::Tet mutation was confirmed by PCR and mobilized to S. aureus USA300 by transduction using phage 80α. Transposon insertions in the brnQ2 and codY genes were identified in the Nebraska Transposon Mutant Library with strain identifications (IDs) of NE605 and NE1555, respectively. The transposons were transduced into our laboratory strain of USA300 using phage 80α, and insertion was confirmed by PCR.

For complementation of the brnQ mutations, each of the brnQ genes, including their native promoters, was cloned into the shuttle vector pRMC2 (34). The resultant plasmids are listed in Table 1.

Radioactive transport assays.

The protocol for transport was adapted from a previously described protocol (35). Cultures were grown overnight at 37°C in complete CDM and subcultured into complete CDM at a starting OD₆₀₀ of 0.1. Bacteria were grown to mid-exponential phase (OD₆₀₀ of 1.0), harvested by filtration on 0.45-μm-pore-size membrane filters, and washed with phosphate-buffered saline (PBS) before being resuspended in CDM lacking amino acids. Cells were heated at 37°C for 10 min prior to the assay. The ¹⁴C-labeled amino acid of interest (PerkinElmer, MA) was added to cells at a final concentration of 1 μM. An aliquot of cells was removed at 20, 40, and 60 s and rapidly filtered through 0.45-μm membrane filters. The filters were immediately washed with 10 ml of 0.1 M LiCl₂ at room temperature. Filters were dried and placed in scintillation vials containing 4 ml of Cytoscint scintillation cocktail (Fisher Scientific). ¹⁴C radioactivity was measured using an LS 6500 scintillation system (Beckman).

To determine transport kinetics, cells prepared as described above were incubated with a 500 nM, 1 μM, 2 μM, or 4 μM concentration of a ¹⁴C-labeled amino acid. An aliquot of cells was filtered as described above after 20 s. The initial velocity of uptake for each substrate concentration was plotted to determine transport kinetics, and the K_m and V_max values were extrapolated.

qPCR.

Quantitative PCR (qPCR) was performed as described previously (31). In brief, RNA was prepared from 5-ml cultures of S. aureus strains grown in TSB to an OD₆₀₀ of 0.6 or in CDM to an OD₆₀₀ of 1.0. Cells were collected and RNA was extracted using an Aurum total RNA minikit as described by the manufacturer (Bio-Rad, Hercules, CA). RNA (500 ng) was reverse transcribed using SuperScript II (Invitrogen, Carlsbad, CA) and PCR amplified using SensiFast SYBR (Bioline, Taunton, MA) and the primers described in Table S1 in the supplemental material. Data were normalized relative to expression of the reference gene rpoB.

Murine model of systemic S. aureus infection.

All protocols were reviewed and approved by the University of Western Ontario's Animal Use Subcommittee, a subcommittee of the University Council on Animal Care. Seven-week-old BALB/c mice were purchased from Charles River Laboratories (Wilmington, MA) and housed in microisolator cages. Bacteria were grown to exponential phase (OD₆₀₀ of approximately 3.0) in TSB, pelleted by centrifugation, washed twice in PBS, and resuspended in PBS. Bacterial suspensions in 100-μl volumes were administered via tail vein injection (6 × 10⁶ to 7 × 10⁶ CFU). Ninety-six hours postchallenge, kidneys and livers were aseptically excised and placed in 4-ml volumes of PBS containing 0.1% (vol/vol) Triton X-100 in sterile 15-ml conical tubes. Organs were homogenized for 10 s, and bacterial loads were calculated following serial dilution on TSB agar plates. Data are presented as log₁₀ CFU recovered per organ, or organ pair in the case of kidneys. For experiments with the hypervirulent strain, mice were monitored every 3 h starting at 12 h postchallenge, weighed, and assessed for behavioral responses to provocation. Mouse gait in response to stimulation was used to assign a behavioral response: normal, mild, moderate, or severe. Severe behavioral responses were considered criteria for premature euthanasia. Data were analyzed using the Student unpaired t test, and P values of <0.05 were considered to indicate statistical significance.

Tissue sectioning and histology.

Kidneys were removed from infected mice and fixed for 24 h in 10% neutral formalin buffer. Five-micrometer sections of embedded tissue were used for hematoxylin and eosin (H&E) staining and Gram staining. A veterinary pathologist inspected 4 sections/kidney pair from 3 mice in the challenge group for the presence of staphylococcal abscesses and tissue abnormalities. Data shown are representative images.

RESULTS

Identification of brnQ genes in USA300 and generation of isogenic mutants.

The brnQ genes identified in the studies by Pohl et al. and Majerczyk et al. (17, 28) are allelic with brnQ1 (SAUSA300_188) and brnQ2 (SAUSA300_306) in USA300 FPR375. Bioinformatic analyses revealed a third potential BrnQ paralog encoded by SAUSA300_1300 with 35% identity to BrnQ1 and 38% identity to BrnQ2; we designated the gene brnQ3. BrnQ3 has 100% amino acid sequence identity to the encoded product of a brnQ gene previously cloned from S. aureus RN450; however, its function as a BCAA transporter was never formally tested (36). The three paralogs share significant sequence similarity with one another: 54% for BrnQ1 and BrnQ2, 54% for BrnQ1 and BrnQ3, and 57% for BrnQ2 and BrnQ3. The proteins have similar lengths: 451 amino acids (aa) for BrnQ1, 435 aa for BrnQ2, and 447 aa for BrnQ3. TMPred predicts that they all have 12 transmembrane domains. Searches of the databases using HHPred revealed the top hits to be transmembrane transporters for a variety of substrates. The genes are located at independent loci in the USA300 genome, and a CodY box (i.e., CodY-binding motif) is positioned 117 bp upstream of the brnQ1 start codon and 85 bp upstream of the brnQ2 start codon (Fig. 1). Using qPCR, we confirmed statistically significant CodY regulation of brnQ1 and a trend (albeit not significant) toward CodY regulation of brnQ2 in USA300 cells grown to early exponential growth phase (OD = 0.6) in TSB (data not shown). To ascertain a function for each of the brnQ paralogs, in-frame, markerless deletions of the brnQ1 and brnQ2 coding regions were constructed, and a tetracycline resistance cassette was used to replace the majority of the brnQ3 coding region. A triple mutant, herein referred to as the brnQ1-2-3 mutant, was also constructed.

FIG 1.

Genomic context for the three brnQ paralogs in USA300. The gene ID in USA300 FPR3757 is indicated. Abbreviations: HYP, hypothetical protein; RocD, ornithine-oxo-acid aminotransferase.

BrnQ1 transports Ile, Leu, and Val.

To assess the specific role of each of the BrnQ transporters, we performed uptake assays with radiolabeled BCAAs. Cells were grown in a chemically defined medium (CDM) to late exponential phase (OD = 1), a growth phase that we ensured, using qPCR, did not demonstrate CodY-dependent repression of the brnQ genes (data not shown). We measured uptake of BCAAs in the isogenic single mutants but also in the brnQ1-2-3 mutant complemented with each gene provided in trans. We observed that compared to that of the WT, uptake of each of Leu, Ile, and Val was decreased in the brnQ1 mutant (Fig. 2). While the brnQ1 mutant was similar to the brnQ1-2-3 mutant in terms of Leu and Val transport, the brnQ1-2-3 mutant had decreased Ile uptake compared to that of the brnQ1 mutant, suggesting the involvement of one of the other two transporters in Ile uptake. Importantly, expression of brnQ1 in trans (on pSO1) in the brnQ1-2-3 mutant resulted in increased uptake of each of Ile, Leu, and Val to levels at or above those seen in the WT (Fig. 2). Together, these data indicate that BrnQ1 is capable of transporting all three BCAAs.

FIG 2.

BrnQ1 transports isoleucine, leucine, and valine. Cells were harvested from cultures grown to late exponential phase in complete CDM. Transport of ¹⁴C-labeled BCAAs, as indicated, was measured at 20, 40, and 60 s. Data points are the means from three biological replicates ± SDs.

BrnQ2 is an Ile transporter.

Using the approach described in the preceding section, we next determined the substrate specificity of BrnQ2. Relative to that in the WT, transport of Ile, Leu, and Val was not decreased in the brnQ2 mutant but was, in fact, increased (Fig. 3). Given the just-described data detailing the involvement of BrnQ1 in transport of the three BCAAs, we attribute the transport seen in the brnQ2 mutant strain to increased expression of brnQ1 (see below). When we examined uptake in the brnQ1-2-3 mutant overexpressing brnQ2 in trans (on pSO2), we found that Leu and Val transport was not restored, while uptake of Ile was restored (Fig. 3). These data implicate BrnQ2 as a transporter selective for Ile, with little to no role in transport of Leu and Val, at least under the conditions used in this study.

FIG 3.

BrnQ2 transports isoleucine. Cells were harvested from cultures grown to late exponential phase in complete CDM. Transport of ¹⁴C-labeled BCAAs, as indicated, was measured at 20, 40, and 60 s. Data points are the means from three biological replicates ± SDs.

BrnQ3 is not a functional BCAA transporter.

Similarly to the assays described in the preceding sections, we assessed BCAA transport in the brnQ3 mutant to validate its predicted function (36). Our results demonstrate that transport of Ile, Leu, and Val was not affected in this strain, relative to that in the WT, and that expression of brnQ3 in trans (on pSO3) in the brnQ1-2-3 mutant background did not appreciably restore transport of any of the three BCAAs (Fig. 4). Since this was a negative result, we used qPCR to confirm that brnQ3 was expressed from pSO3 in the brnQ1-2-3 mutant to levels comparable to those of brnQ1 expressed from pSO1 (data not shown). Combined, these data led us to conclude that BrnQ3 does not function as a transporter of BCAAs, at least under the conditions used in this study.

FIG 4.

BrnQ3 in not a BCAA transporter. Cells were harvested from cultures grown to late exponential phase in complete CDM. Transport of ¹⁴C-labeled BCAAs, as indicated, was measured at 20, 40, and 60 s. Data points are the means from three biological replicates ± SDs.

Mutation of the brnQ3 gene was previously identified as causing salt sensitivity in S. aureus strain RN450 (37). To examine this phenomenon further, we streaked USA300, along with its isogenic brnQ1, brnQ2, and brnQ3 mutants, on TSB agar and TSB agar containing 2 M NaCl. Although all strains grew relatively slower on agar plates containing 2 M salt, mutation of the brnQ genes had no effect on the ability of strain USA300 to grow in the presence of high concentrations of NaCl, including the brnQ1-2-3 mutant (see Fig. S1 in the supplemental material). It is likely that the absence of the salt-sensitive phenotype in our study is due to strain differences between RN450 and USA300, since the assay was performed identically, using 2 M NaCl, as previously described (37).

Transport kinetics.

To gain further insight into the roles of the three BrnQ paralogs in BCAA transport, we determined the kinetic properties of BCAA transport in strains expressing each of BrnQ1, BrnQ2, and BrnQ3 by measuring the initial velocity of uptake in response to various substrate concentration. The kinetic plots are shown in Fig. 5, and apparent K_m and V_max data are provided in Table 2. The data indicate that BrnQ1 is a major determinant of uptake of all three BCAAs, while BrnQ2 appears specific to Ile, albeit with a lower affinity for Ile than that of BrnQ1.

FIG 5.

Kinetics of BCAA transport in S. aureus. Values are the means from 3 independent cultures ± SDs.

TABLE 2.

K_m and V_max values for the BrnQ transporters^a

Transporter	Leucine		Isoleucine		Valine
	Km	Vmax	Km	Vmax	Km	Vmax
BrnQ1	1.62	6.85	0.93	10.22	1.79	12.81
BrnQ2			8.94	18.13

^aK_m values are listed in μM, while V_max values are listed in pmol/μg/min.

BrnQ1-mediated Leu and Val acquisition is required for optimal growth of USA300 in CDM.

We next sought to investigate the physiological role of the functional BCAA transporters in meeting the nutritional requirements of S. aureus growth. To do so, we selected CDM, described previously (31), enabling the limitation or omission of amino acids of interest. In complete CDM, we consistently observed a slight growth delay for the brnQ1 mutant compared to the WT strain (Fig. 6A), indicating that BrnQ1 is required for growth in this medium. In comparison, all strains grew equally well in TSB (see Fig. S2 in the supplemental material). We next investigated the amino acid specificity of the growth phenotype in CDM. S. aureus has been reported to be auxotrophic for Leu and Val despite the presence of BCAA biosynthetic genes (15–17), indicating a critical role for BCAA uptake to support growth. First, to confirm the auxotrophic phenotype of USA300, the WT strain was grown in complete CDM as well as CDM lacking Ile (CDM^−Ile), Leu (CDM^−Leu), or Val (CDM^−Val). As expected based on the literature, WT USA300 was not growth impaired in CDM^−Ile compared to growth in complete CDM, indicating that USA300 is capable of de novo synthesis of Ile that is sufficient to support immediate growth. In contrast, we observed that growth in CDM^−Leu was significantly delayed through the first 12 h of incubation, but by 24 h the cells had attained a biomass equivalent to that obtained in complete CDM (Fig. 6B). For growth in CDM^−Val, we observed an even greater defect. A significant growth delay was seen through the initial 24 h of incubation, followed by growth that resulted in a biomass equivalent to that of the WT after 48 h (Fig. 6B). This suggests that Leu and Val synthesis is delayed, possibly owing to the repression of the biosynthetic genes by CodY, Gcp, and/or unrecognized factors (17, 28, 38). We observed that the growth delay in CDM^−Leu was consistent upon serial subculturing, excluding the accumulation of suppressor mutations.

FIG 6.

BrnQ1 is required for optimal growth in CDM limited for leucine or valine. (A) Strains were pregrown in complete CDM and subcultured into complete CDM. Growth was monitored until stationary phase was reached. (B) WT USA300 was grown in complete CDM and subcultured into complete CDM or CDM lacking Ile, Leu, or Val. Growth yield was monitored at 12, 24, and 48 h. (C and D) Strains were pregrown in complete CDM and subcultured into CDM in which all amino acids were present at the same concentration as in CDM, except for alterations to the concentration of Leu (C) or Val (D). Growth yield was measured at 12 h. For panels E and F, growth curves were performed on the strains as indicated. Cultures were pregrown in complete CDM and subcultured into CDM in which all amino acids were present at the same concentrations as in CDM, except that Leu was at 1% of the concentration in complete CDM (E) and Val was at 1% of the concentration in complete CDM (F). Data shown are the means from three biological replicates ± SDs and were analyzed using the Student t test (compared to WT). **, P < 0.01; ***, P < 0.001.

We next assessed the role of BrnQ1 in acquiring Leu and Val to promote early and rapid growth, since the BCAA uptake data shown in Fig. 2 through 4 implicated BrnQ1 as the dominant Leu and Val transporter. Growth levels of WT USA300 and the brnQ1 mutant were compared in complete CDM and CDM with various amounts of Leu and Val limitation. The brnQ1 mutant was impaired for growth compared to WT USA300 upon limitation of Leu or Val to concentrations ≤5% of that in complete CDM (complete CDM contained 684 μM Leu and 684 μM Val) (Fig. 6C and D). Growth of the brnQ1 mutant was fully restored to WT levels upon complementation with brnQ1 in trans (Fig. 6C and D).

To investigate the possible involvement of BrnQ2 and BrnQ3 in the acquisition of Leu or Val for growth, we limited these two amino acids individually to 1% of the concentration in complete CDM and compared the growth, over 24 h, of the WT to that of the brnQ1-2-3 mutant carrying empty vector or the brnQ genes. We observed that the brnQ1-2-3 mutant, as expected based on the above-described data, was impaired for growth (Fig. 6E and F), and neither vector expressing brnQ2 or brnQ3 restored this defect. In contrast, expression of brnQ1 allowed rapid growth such that the strain grew quicker than the WT in media limited for Leu (Fig. 6E). Overexpression of brnQ1 had an even more pronounced effect on the growth enhancement of S. aureus in media limited for Val (Fig. 6F).

Combined, our data indicate that BrnQ1-mediated Leu and Val acquisition is required for optimal growth of S. aureus when these amino acids are limited.

Mutation of brnQ2 results in overexpression of brnQ1.

The data shown in Fig. 3 demonstrated that a brnQ2 mutant transported Ile, Leu, and Val to levels higher than those of the WT strain. This was an interesting observation that led us to investigate the possible regulatory effect of a brnQ2 mutation on brnQ1 expression, given that our data have shown that BrnQ1 is a major Ile, Leu, and Val transporter. As shown in Fig. 7A, qPCR demonstrated that brnQ1 expression was over 40-fold higher in the brnQ2 mutant than in the WT. We confirmed that this phenomenon was specifically due to mutation of brnQ2 and not some secondary mutation in the brnQ2 mutant since complementation of the brnQ2 mutant with brnQ2 in trans reversed expression of brnQ1 to levels that were even lower than in the WT.

FIG 7.

Mutation of brnQ2 results in overexpression of brnQ1. (A) The expression level of brnQ1 was evaluated in S. aureus strains, as indicated. Data were normalized relative to the expression of rpoB, and expression of brnQ1 in WT cells was set to 1 as the comparator. (B and C) Cultures were pregrown in complete CDM and subcultured into CDM in which all amino acids were present at the same concentrations as in CDM, except that Leu was at 1% of the concentration in complete CDM (B) and Val was at 1% of the concentration in complete CDM (C). Data shown are means from three biological replicates ± SDs. **, P < 0.01.

Given that a mutation of brnQ2 enhances expression of brnQ1, we hypothesized that a brnQ2 would grow quicker than the WT in media limited for either Leu or Val. As we did for experiments shown in Fig. 6E and F, we compared the growth of selected strains in media that contained 1% of the concentration of either Leu or Val found in complete CDM. In support of our hypothesis, we observed that a brnQ2 mutant demonstrated early and rapid growth, compared to that of the WT, in medium limited for Leu, and more so in medium limited for Val. The growth kinetics of this mutant were virtually identical to those of a brnQ1 mutant overexpressing brnQ1 from a plasmid (Fig. 7B and C).

BrnQ1 is required for USA300 fitness in vivo.

We next investigated the contribution of BCAA acquisition to the fitness of USA300 in vivo using an established bacteremia murine model of infection. Groups of 10 female BALB/c mice were infected via tail vein injection with 5 × 10⁶ to 7 × 10⁶ CFU of WT USA300 or the isogenic mutant. Mice infected with either the WT or the brnQ1 mutant were sacrificed on day 4 postinfection, and the bacterial burdens in livers and kidneys were assessed. Mice infected with the brnQ1 mutant had significantly lower CFU in the kidneys and livers than mice infected with WT USA300 (Fig. 8), suggesting that USA300 encounters BCAA limitation in vivo and that BrnQ1 is necessary for their acquisition in order to meet growth requirements.

FIG 8.

BrnQ1 is required for full S. aureus virulence in a mouse model of systemic infection. BALB/c mice were infected with 5 × 10⁶ to 7 × 10⁶ CFU of WT USA300 (n = 9) or the ΔbrnQ1 mutant (n = 10) via tail vein injection. Mice were sacrificed on day 4 of infection, kidneys (A) and livers (B) were harvested and homogenized, and dilutions were plated for CFU determination. Data were analyzed using the Student unpaired t test. *, P < 0.05; ***, P < 0.001.

Deletion of brnQ2 results in increased virulence, including severe kidney pathology.

Unexpectedly, mice challenged with the brnQ2 mutant, in the same manner as described above, lost a significant percentage of body weight compared to WT-infected mice by 24 h postinfection and exhibited severe signs of infection, requiring premature sacrifice (Fig. 9A and B). Although exhibiting no signs of disease that would warrant euthanasia at the 24-h time point, WT-infected mice were also sacrificed at this time to compare bacterial burdens in organs. Kidneys and livers from mice infected with the brnQ2 mutant had significantly higher CFU than kidneys and livers from WT-infected mice (Fig. 9C and D). Upon visual inspection, and in contrast to those of mice infected with the WT, the kidneys from all mice infected with the brnQ2 mutant were found to have numerous small white abscesses on the renal cortex (Fig. 9E). Infected kidneys were sectioned for histology and stained to visualize tissue pathology. H&E staining of kidneys from WT-infected mice displayed no signs of inflammation, and few abscesses were found in any of the sections analyzed (Fig. 9Fi). Overall, few bacteria were present in the tissue based on Gram staining (Fig. 9Fii). Upon examination of several sections, we were able to observe clusters of Gram-positive bacteria, but this was a rare case (Fig. 9Fiii). H&E and Gram staining of kidneys from brnQ2-infected mice revealed multiple foci of bacteria surrounded by neutrophils throughout the kidney (Fig. 9G and H). Abscesses appeared to be surrounded by fibrin deposits, typical of S. aureus abscesses (39).

FIG 9.

Deletion of brnQ2 results in increased virulence and severe renal abscess formation. BALB/c mice were infected with 5 × 10⁶ to 7 × 10⁶ CFU of WT USA300 (n = 9) or the ΔbrnQ2 mutant (n = 10) via tail vein injection. Mice were monitored every 3 h and assessed for provoked behavior responses. (A and B) Mice infected with the ΔbrnQ2 mutant exhibited severe signs of infection and lost significantly more weight than WT-infected mice within the first 24 h of infection, requiring euthanasia. (C and D) Kidneys and livers were harvested from both groups of mice at 24 h postinfection and plated for CFU determination. (E) Representative kidneys harvested from mice infected with either WT USA300 or the ΔbrnQ2 mutant, demonstrating gross pathology. (F) H&E and Gram staining of representative kidneys harvested from WT-infected mice. (G and H) H&E and Gram staining of representative kidneys from brnQ2 mutant-infected mice. **, P < 0.01; ***, P < 0.001.

To provide additional support for the conclusion that the hypervirulence phenotype of our brnQ2 deletion mutant was indeed solely attributable to mutation of brnQ2, we repeated the infection experiments with an independently derived brnQ2 mutation. For this, we transduced the brnQ2::ΦNΣ transposon insertion, obtained from the Nebraska Transposon Mutant Library (40, 41), into our laboratory USA300 strain. Mice infected with USA300 brnQ2::ΦNΣ showed significant morbidity and a lack of a provoked response within the 24- to 48-h time period and required early euthanasia (data not shown). Similar to mice infected with the brnQ2 deletion mutant, mice infected with USA300 brnQ2::ΦNΣ exhibited hyperabscessing and severe parenchymal damage throughout the kidney (data not shown). It is important to note that the challenge dose used in all experiments is within the “normal” range used by us and other laboratories that use this model to study S. aureus pathogenesis. The mice do not normally require euthanasia before at least 96 h, and they frequently survive much longer if the infection is allowed to proceed.

In an attempt to understand the basis for the hypervirulence phenotype, we investigated whether the brnQ2 deletion mutant exhibited an altered secreted protein profile compared to that of WT USA300 when grown in CDM; however, we did not observe differences in protein secretion that might explain the hypervirulence phenotype (data not shown). Additional growth conditions will likely need to be explored in order to more closely recapitulate the environment that USA300 encounters in vivo.

DISCUSSION

S. aureus necessitates expression of transporters for Leu and Val acquisition that would allow for growth, since it has been reported to be auxotrophic for Leu and Val, despite carrying the genes for biosynthesis (15, 16). Here, we demonstrate that BrnQ1 plays a dominant role in meeting Leu and Val nutritional requirements in vitro, especially in a medium containing either of these amino acids in the low micromolar range, and postulate that this is also the case in vivo, since a brnQ1 mutant is attenuated in vivo. In stark contrast, BrnQ2 is selective for Ile and acquisition of Ile appears to play an important role in regulating USA300 virulence. BrnQ3 does not function as a BCAA transporter, and although it was previously identified to contribute to the osmotolerance of S. aureus, we were unable to replicate this finding. Our ongoing studies will investigate the ligand specificity of BrnQ3, and the importance of this is demonstrated by the finding that a brnQ3 mutant is attenuated in vivo (42).

Our observation that USA300 is able to grow in the absence of exogenous Leu or Val contrasts previous reports that S. aureus is auxotrophic for these amino acids. This is likely explained by the extension of the growth period monitored in this work, since appreciable growth of USA300 in media lacking Leu and Val is not observed until 24 and 48 h, respectively (Fig. 6B). We presume that the biosynthesis of Leu and Val is derepressed at the later stages of growth, as we excluded the possibility of development of suppressor mutations in CDM^−Leu. At present, the regulatory mechanisms that govern the regulation of synthesis of the isoleucine, leucine, and valine group (ILV) under these conditions is not known, but plausible mediators of repression include CodY and Gcp, both known repressors of the ILV biosynthetic genes in S. aureus (17, 28, 38). Although we did not investigate this in further detail, of note, we observed that growth in CDM^−Leu and CDM^−Val (i.e., growth dependent on Leu and Val synthesis) is iron dependent (data not shown). Indeed, several enzymes involved in BCAA biosynthesis contain Fe-S clusters, including IlvD, LeuC, and LeuD. LeuCD are repressed in Bacillus subtilis upon iron starvation (43, 44), and a similar link between iron availability and BCAA biosynthesis might exist in S. aureus. This phenotype is relevant when considering the levels of BCAA biosynthesis in vivo, as iron is limited during S. aureus infection (45) and consequently might impart a greater importance to BCAA acquisition in vivo.

BrnQ1 plays a predominant role in BCAA transport, as demonstrated by both radioactive transport assays and the growth assays in which Leu and Val are present in low micromolar concentrations (Fig. 2 and 6). However, two lines of evidence suggest the presence of at least one additional BCAA transporter in S. aureus. First, a brnQ1 mutant (Fig. 6) and a brnQ1-2-3 mutant (data not shown) are not defective for growth, compared to the WT, even at early time points, in CDM containing concentrations of Leu or Val of 100 μM or higher, and our transport assays indicate that the brnQ1-2-3 mutant still transports some Ile, Leu, and Val (Fig. 2 to 4). We are currently investigating additional genes that contribute to BCAA transport, such as a gene homologous to the CodY-regulated transporter BcaP described for Lactococcus lactis (9). BcaP functions as the high-affinity BCAA transporter in L. lactis, whereas BrnQ is the low-affinity transporter (9). Given that BrnQ1 in USA300 is necessary for growth when Leu or Val is in low micromolar concentrations, and not when they are present at higher concentrations in the growth medium, we predict that BrnQ1 is the high-affinity transporter in this species. In line with this, the brnQ1 mutant is attenuated in vivo, suggesting that BrnQ1 functions as a primary means of BCAA transport during infection.

We hypothesize that the disparate outcomes of infection with the brnQ1 versus the brnQ2 mutants may be at least be partially, if not wholly, dependent on the expression level of brnQ1 and, thus, the ability of S. aureus to scavenge Leu and Val when they are present at low concentrations. Indeed, our in vitro data demonstrate that the loss of brnQ1 adversely affects growth in low-Leu and low-Val media, while the loss of brnQ2 positively affects growth in the same media, as a result of high-level expression of brnQ1. Although we do not yet understand the molecular basis of this phenomenon, it is possible that the hypervirulence of brnQ2 mutants is due to the overexpression of brnQ1. Another possible explanation is the overexpression of toxins and secreted enzymes in a brnQ2 mutant. Although we were unable to detect this occurring during growth of the brnQ2 mutant in CDM, this does not rule out increased toxin expression during in vivo growth through signals present in vivo that are absent in CDM.

Finally, it is interesting to consider the possible effects of mutating brnQ1 and brnQ2 on CodY activity, both in vitro and in vivo, and this is the subject of our ongoing studies. At least part of their involvement in virulence may be linked to their ability to manipulate intracellular levels of CodY effector molecules, with implications for the expression of CodY-regulated genes.