Mutations in a Membrane Permease or hpt Lead to 6-Thioguanine Resistance in Staphylococcus aureus

ABSTRACT

We recently discovered that 6-thioguanine (6-TG) is an antivirulence compound that is produced by a number of coagulase-negative staphylococci. In Staphylococcus aureus, it inhibits de novo purine biosynthesis and ribosomal protein expression, thus inhibiting growth and abrogating toxin production. Mechanisms by which S. aureus may develop resistance to this compound are currently unknown. Here, we show that 6-TG-resistant S. aureus mutants emerge spontaneously when the bacteria are subjected to high concentrations of 6-TG in vitro. Whole-genome sequencing of these mutants revealed frameshift and missense mutations in a xanthine-uracil permease family protein (stgP [six thioguanine permease]) and single nucleotide polymorphisms in hypoxanthine phosphoribosyltransferase (hpt). These mutations engender S. aureus the ability to resist both the growth inhibitory and toxin downregulation effects of 6-TG. While prophylactic administration of 6-TG ameliorates necrotic lesions in subcutaneous infection of mice with methicillin-resistant S. aureus (MRSA) strain USA300 LAC, the drug did not reduce lesion size formed by the 6-TG-resistant strains. These findings identify mechanisms of 6-TG resistance, and this information can be leveraged to inform strategies to slow the evolution of resistance.

INTRODUCTION

Staphylococcus aureus is an important human pathogen that causes mild to severe disease, including skin and soft tissue infections, pneumonia, bacteremia, toxic shock syndrome, osteomyelitis, and endocarditis (1, 2). Treatment for S. aureus infections has been complicated due to the emergence of multidrug-resistant S. aureus strains such as methicillin-resistant S. aureus (MRSA) (3). Currently, vancomycin and daptomycin are used to treat drug-resistant S. aureus infections, but resistance against these antibiotics has already been documented (4–6). To combat drug-resistant S. aureus infections, it is critically important to discover new antibiotics, chemically modify existing ones to improve utility, repurpose known drugs traditionally not used for antibacterial activity, or use drugs in combination to forestall evolution of resistance. In addition to antibiotics, antivirulence compounds could also be used to combat bacterial infections. These compounds reduce the virulence potential of a pathogen, which allows the host immune system and traditional antibiotics to eliminate the pathogen more effectively. Since these compounds do not directly impair the growth of pathogens, they exert less selective pressure for resistance. For example, a compound was recently discovered to reverse β-lactam resistance in S. aureus and synergize with traditional antibiotics (7). However, resistance against antivirulence drugs can still develop, and this needs to be characterized in detail to inform alternative treatment options and to mitigate the evolution of drug resistance (8).

To discover anti-S. aureus drugs, several groups screened coagulase-negative staphylococcal species (CoNS), because these bacteria can often be found together with S. aureus and compete for nutrient-poor niches such as human skin. To successfully compete against S. aureus, CoNS have developed several strategies to antagonize S. aureus, one of which is to secrete antimicrobials that antagonize S. aureus growth and virulence. One of these classes of antimicrobials is bacteriocins; these are ribosomally synthesized antimicrobial peptides produced by Gram-positive bacteria that have a narrow spectrum of activity and target bacterial species that are closely related to the producing strain (9). Therefore, bacteriocins have few off-target effects and could be promising for targeting specific pathogens such as S. aureus. Indeed, recent research revealed that the CoNS Staphylococcus hominis and Staphylococcus epidermidis produce bacteriocins that specifically kill S. aureus on human skin (10). These bacteria have also successfully been used in a phase 1 randomized clinical trial to treat infection in atopic dermatitis, demonstrating the protective effects that CoNS provide the host against S. aureus (11).

In addition to bacteriocins, CoNS can produce other bactericidal compounds that kill S. aureus. For instance, the human commensal Staphylococcus lugdunensis has been shown to produce a novel non-ribosomally synthesized thiazolidine-containing cyclic peptide termed lugdunin that is potent against S. aureus (12). In addition to its bactericidal activity, lugdunin has also been shown to stimulate production of antimicrobial peptides such as LL-37 and recruit host immune cells, which further eradicates S. aureus in vivo (13). Remarkably, resistance against lugdunin has not been detected despite passaging S. aureus in subinhibitory concentrations of lugdunin for 30 days under laboratory conditions (12). Since resistance against lugdunin has not emerged and the antibiotic has been shown to be potent against S. aureus, researchers have begun to chemically modify the compound to create lugdunin variants that show similar activity to the parent compound (14).

Besides bactericidal compounds, several CoNS species have been shown to antagonize S. aureus by targeting agr quorum sensing in S. aureus. The agr system is extremely well characterized in S. aureus and is known to be critically important for its pathogenesis, as quorum sensing regulates the transcription of many toxins required for infection (15–20). Importantly, CoNS, including Staphylococcus caprae, Staphylococcus simulans, and Staphylococcus schleiferi, secrete unique autoinducing peptides that block the agr quorum sensing system in S. aureus. While the agr system is not essential for S. aureus growth, this strategy allows these CoNS to indirectly gain a competitive advantage by preventing S. aureus from effectively colonizing the host (21–23). These microbial interactions between different staphylococcal species inadvertently result in protection of the host from infection by the more virulent S. aureus. Moreover, autoinducing peptides from these species could be used as a drug to counter S. aureus skin infections (21, 22). Altogether, these studies demonstrate that CoNS have evolved several strategies to outcompete S. aureus to gain a competitive advantage. Clearly, CoNS are a rich reservoir of anti-S. aureus agents that can be leveraged to treat S. aureus infections.

We have recently discovered that a number of Staphylococcus chromogenes, S. epidermidis, S. pseudintermedius, and S. capitis strains produce the purine analog 6-thioguanine (6-TG), a purine biosynthesis inhibitor that impairs S. aureus growth in purine-limited growth media (24). Interestingly, these species are all known to colonize similar niches as S. aureus, and they may have evolved to produce 6-TG to successfully compete against S. aureus in these habitats (25–28). 6-TG is also antagonistic in vivo, as it can be used to prevent lesion formation during MRSA subcutaneous infection (24). We showed that 6-TG reduces toxin production in S. aureus by downregulating expression of genes encoding the global virulence regulator (agr) and ribosome biogenesis (24). Here, we demonstrate that S. aureus can evolve resistance against 6-TG, and this occurs through loss-of-function mutations in a gene encoding a xanthine-uracil family permease or through missense mutations in the hypoxanthine phosphoribosyltransferase gene. This study thus demonstrates that resistance to 6-TG can occur through mutations that alter the activation of the 6-TG prodrug or through mutations that block entry of the prodrug into the S. aureus cell.

RESULTS

Selection of 6-TG-resistant S. aureus.

We previously showed that growth of the community-acquired MRSA strain USA300 LAC (here termed S. aureus) was slightly sensitive to 6-TG at 10 μg/ml on nutrient rich tryptic soy agar (TSA) (24). However, S. aureus was much more sensitive to 6-TG on purine-deplete RPMI medium at the same concentration, since the drug is a known inhibitor of purine biosynthesis (29). Therefore, to exert strong selective pressure for 6-TG-resistant S. aureus clones, we used a high concentration of 6-TG (100 μg) in a disc diffusion assay on RPMI medium. As expected, we observed a large zone of inhibition around the disc after 24 h of incubation. Following 48 h of incubation at 37°C, we detected small colonies that appeared within the zone of inhibition, an observation that was replicated across several independent experiments (Fig. 1A). Furthermore, we reduced 6-TG concentrations in this assay and noticed this phenomenon occurred with concentrations as low as 3 μg, although the zone of inhibition was much smaller and fewer small colonies appeared in the zone (data not shown).

FIG 1.

Selection for 6-TG-resistant S. aureus clones. (A) Disc diffusion assay using 100 μg 6-TG against S. aureus on RPMI medium incubated for 48 h at 37°C. Arrows point to spontaneously arising S. aureus colonies that appear within the zone of inhibition. (B) Quantification of zone of inhibition generated using 100 μg 6-TG on RPMI medium against wild-type S. aureus and spontaneous colonies from panel A. Results are from 3 biological replicates. Data are shown as means ± standard deviations (SDs) (n = 3). *, P ≤ 0.05 by one-way analysis of variance (ANOVA) with Dunn’s multiple-comparison test; n.s., not significant.

To determine whether the colonies arising spontaneously were indeed resistant to 6-TG-mediated growth inhibition, these colonies were picked and regrown in tryptic soy broth (TSB) in the absence of 6-TG. These cultures were then streaked on RPMI medium and examined for sensitivity to 6-TG using the disc diffusion assay. While the growth of wild-type S. aureus was inhibited extensively, several, but not all, of the clones that grew out from the initial selection maintained resistance to 6-TG (Fig. 1B). Moreover, we determined that the MIC of 6-TG in chemically defined medium (CDM) is 12.5 μg/ml for wild-type S. aureus. The growth of the resistant clones, however, although slightly lessened, was not completely inhibited at concentrations >12.5 μg/ml (as it was for the wild type [WT]) and up to at least 400 μg/ml (Table 1). Since we had removed selective pressure during growth of the resistant clones in TSB, these data suggest that not only could we select for S. aureus resistant to 6-TG but also that this resistance was heritable and therefore genetically encoded.

TABLE 1.

6-TG MICs against S. aureus and 6-TG-resistant mutants^a

Strain	MIC (μg/ml)
S. aureus USA300	12.5
R1	>400
R3	>400
R5	>400
R6	>400
R7	>400
R8	>400
R9	>400
R10	>400
R11	>400

^aMICs of 6-TG were determined using minimal chemically defined media. Concentrations up to 400 μg/ml were tested against all strains.

Identification of mutations that result in 6-TG resistance in S. aureus.

To determine the genetic changes that occurred in the stably resistant S. aureus clones, we sequenced the genomes of two independently grown cultures of our WT USA300 LAC stock along with 9 of the 6-TG-resistant isogenic clones that were selected from 9 independent experiments performed as described above. We then mapped sequence reads to the parent S. aureus FPR_3757 genome. Remarkably, for each of the sequenced 6-TG-resistant clones, we detected mutations in only one of two genes despite the fact that 6-TG is a potent mutagen (29–31).

We found that for resistant clones R1, R5, R6, and R8, mutations occurred in SAUSA300_2207. This gene is annotated in some genomes as pbuG encoding a putative guanine permease but, in other genomes, is annotated as encoding a putative xanthine/uracil permease family protein; here, we term the gene stgP, for six thioguanine permease. For stgP (resistant mutants R1, R5, R6, and R8), all mutations were indels (i.e., insertions and deletions) and, for three of the four mutants, resulted in frameshifts and the introduction of stop codons which prematurely truncated the protein (Table 2). Mutant R1 had a three-base insertion which introduced an Ile residue after amino acid 388.

TABLE 2.

Whole-genome sequencing reveals mutations that confer resistance against 6-TG^a

Mutant	Gene no.	Gene name	Nucleotide mutation	Nucleotide position	Amino acid mutation	Amino acid position
R1	SAUSA300_2207	6-Thioguanine permease	ATC insertion	1017	Introduction of Ile	338_339
R3	SAUSA300_0488	Hypoxanthine phosphoribosyltransferase	G→T	226	Gly→Val	76
R5	SAUSA300_2207	6-Thioguanine permease	A insertion	102	Introduction of stop codon	34_35
R6	SAUSA300_2207	6-Thioguanine permease	T insertion	102	Introduction of stop codon	34_35
R7	SAUSA300_0488	Hypoxanthine phosphoribosyltransferase	GG→AT	241–242	Gly→Ile	81
R8	SAUSA300_2207	6-Thioguanine permease	A deletion	467–468	Introduction of stop codon	166
R9	SAUSA300_0488	Hypoxanthine phosphoribosyltransferase	C→T	320	Thr→Ile	107
R10	SAUSA300_0488	Hypoxanthine phosphoribosyltransferase	A→T	214	Ser→Cys	72
R11	SAUSA300_0488	Hypoxanthine phosphoribosyltransferase	A→C	181	Thr→Pro	61

^aResistant S. aureus clones were isolated from a 6-TG disc diffusion assay on RPMI medium, and genomes were prepared and sequenced. The bioinformatics software Geneious was used to identify mutations in these resistant clones.

For resistant clones R3, R7, R9, R10, and R11, mutations occurred in SAUSA300_0488, encoding hypoxanthine phosphoribosyltransferase (hpt). In contrast to mutations in stgP, mutations in hpt were single nucleotide polymorphisms (SNPs) causing single amino acid substitutions (Fig. 2A, Table 2). That no indels or nonsense mutations were found in hpt is not surprising given that, in several bacterial species, hpt is essential (32, 33) and is an indication that, in S. aureus, it is also an essential gene. In an attempt to provide a rational explanation for the effect of the hpt missense mutations on Hpt function, we mapped the S. aureus apoHpt structure (PDB 4RQB_B) to the holoHpt structure from Bacillus anthracis (PDB 6D9R); the latter structure was of HprT bound to phosphoribosyl pyrophosphate (PRPP) and 9-deazaguanine, an analog inhibitor (Fig. 2B and C). This showed us that all but one of the missense mutations map to the putative substrate binding region of S. aureus HprT, leading us to believe that the mutant proteins are not functional in converting the prodrug 6-TG to its active form. The one mutation occurring outside the substrate binding region of the protein was a Thr→Pro mutation which could presumably have consequences on the substrate binding site as well.

FIG 2.

6-TG-resistant S. aureus clones harbor mutations in stgP and hpt. (A) Genetic map of mutations found in 6-TG-resistant S. aureus clones in stgP and hpt from 9 independent experiments. (B) Structural model of the Hpt enzyme from Bacillus anthracis binding to phosphoribosyl pyrophosphate and the noncatalyzable guanine analog 9-deazaguanine. (C) Structural model of the Hpt enzyme from S. aureus. Mutations in hpt that were identified in 6-TG-resistant S. aureus clones are mapped in this model. (D) Wild-type S. aureus, R1 pEmpty, R1 pstgP R10 pEmpty, and R10 phpt were serially diluted and plated onto RPMI medium containing 10 μg/ml 6-TG. (E) Wild-type S. aureus, R1, and R10 were grown in TSB with 6-TG (10 μg/ml) or a vehicle control, and OD₆₀₀ was measured every hour for the first 8 h. The OD₆₀₀ was also measured at the indicated times. Results are pooled from 4 biological replicates from two independent experiments. Data are shown as means ± SDs. (F) Wild-type S. aureus, R1, and R10 were grown in RPMI medium with 6-TG (10 μg/ml) or a vehicle control, and OD₆₀₀ was measured at the indicated times. Results are pooled from 4 biological replicates from two independent experiments. Data are shown as means ± SDs.

To confirm that mutations in the hpt or stgP gene confer 6-TG resistance, we complemented the mutants with their respective wild-type genes and plated the bacteria on RPMI medium containing 6-TG. As expected, R1 and R10 carrying an empty vector (here termed R1 pEmpty and R10 pEmpty, respectively) are resistant to 6-TG. However, when R1 and R10 were complemented with the wild-type versions of stgP and hpt (here termed R1 pstgP and R10 phpt, respectively), they became resensitized to 6-TG and grew similarly to wild-type S. aureus (Fig. 2D). Moreover, liquid growth experiments revealed that R1 and R10 are resistant to 6-TG in nutrient-rich TSB and in purine-deplete RPMI medium, whereas wild-type S. aureus was sensitive in both types of media (Fig. 2E and F). As expected, wild-type S. aureus was more sensitive to 6-TG when grown in RPMI medium (as it is purine free) than when grown in TSB. Taken together, these data confirm that the spontaneously arising mutations we identified in stgP and hpt cause S. aureus to become resistant to 6-TG-mediated growth inhibition.

To complement our efforts to identify resistant mutants, described above, we also screened the entire Nebraska transposon insertion library by growing the mutants on RPMI medium supplemented with 6-TG (34). In this screen, we only identified the stgP transposon mutant as resistant (data not shown), supporting the notion that the only knockout mutation in a nonessential gene in S. aureus conferring resistance to 6-TG is in this permease-encoding gene. Since hpt is an essential gene, this mutant was not present in the Nebraska transposon insertion library. Our combined results support the contention that mutations in stgP and hpt result in 6-TG resistance in S. aureus.

6-TG does not affect toxin production in S. aureus clones that have mutations in stgP and hpt.

In prior work, we demonstrated that in addition to growth inhibition via inhibition of de novo purine biosynthesis, 6-TG also significantly reduced toxin production in S. aureus by downregulating the expression of genes that code for ribosome biogenesis and by downregulating expression of the accessory gene regulator (agr) system (24). Thus, we next questioned whether R1 (as a representative mutant for stgP) and R10 (as a representative mutant for hpt) were also resistant to the effects 6-TG imparts on secreted proteins. To determine this, we cultured the strains in TSB supplemented with 6-TG (10 μg/ml) or a vehicle control overnight, collected supernatants, and analyzed the secreted protein profile using SDS-PAGE. We observed that the secreted protein profile of wild-type S. aureus, but not R1 pEmpty and R10 pEmpty, had substantially less secreted protein when treated with 6-TG than with the vehicle control (Fig. 3A). Furthermore, both R1 pstgP and R10 phpt were resensitized to 6-TG, as the secreted protein profiles of these strains were similar to that of the 6-TG-treated S. aureus culture (Fig. 3A). We performed Western blot analysis of Hla from cell culture supernatants from cultures treated with a vehicle control or 6-TG, as this toxin is known to be important for S. aureus pathogenesis (35–38). Here, we confirmed that S. aureus produces less Hla when cultures were treated with 6-TG than when treated with a vehicle control, while Hla levels were unchanged in R1 pEmpty and R10 pEmpty cultures under either condition (Fig. 3B). However, only R1 pstgP and not R10 phpt Hla levels were reduced in response to 6-TG relative to that in R1 pEmpty and R10 pEmpty, respectively, suggesting that the mutation in hpt in R10 may have a dominant negative effect and antagonize or override the effects of wild-type hpt expressed from the plasmid (Fig. 3B). Taken together, these data demonstrate that mutations in stgP and hpt in S. aureus result in resistance against the effects that 6-TG has on toxin production in addition to its effect on growth.

FIG 3.

Secreted protein profiles of stgP and hpt S. aureus mutants are unaffected by 10 μg/ml 6-TG. (A) S. aureus was grown in the presence of 6-TG (10 μg/ml) or a vehicle control to an OD₆₀₀ of 7.0, and secreted proteins in the supernatant were TCA precipitated and separated by SDS-PAGE. (B) Protein samples in panel A were immunoblotted and probed with anti-alpha toxin antibody in a Western blot. Numbers indicate molecular mass markers.

Mutations in hpt but not stgP confer resistance against high 6-TG concentrations in vitro.

Next, we questioned whether higher 6-TG concentrations could curtail growth and/or toxin production in R1 and R10. Similarly to that for previous experiments, we cultured wild-type S. aureus, R1, and R10 in TSB supplemented with different concentrations of 6-TG (1.3 to 160 μg/ml) or a vehicle control overnight, collected supernatants, and analyzed the effects of these 6-TG concentrations on secreted Hla. We observed that wild-type S. aureus was exquisitely sensitive to 6-TG, as growth was impaired with every concentration used (Fig. 4A). As expected, both R1 and R10 were resistant at concentrations of 6-TG up to 10 μg/ml, concentrations of 6-TG to which wild-type S. aureus was sensitive. However, growth of R1 was impaired at concentrations of 20 μg/ml and greater (Fig. 4B). In contrast to R1, the growth of R10 was unchanged when treated with all concentrations of 6-TG tested (i.e., as high as 160 μg/ml) (Fig. 4C). Next, we tested whether these same 6-TG concentrations could affect Hla production in these strains. We observed that Hla production by wild-type S. aureus was reduced by 6-TG with concentrations as low as 2.5 μg/ml, while Hla production by R10 was unchanged in response to all 6-TG concentrations used. Interestingly, Hla levels produced by R1 were diminished at concentrations of 80 μg/ml and higher, demonstrating that high 6-TG concentrations can still affect R1 (Fig. 5). These data demonstrate that while R10 (hpt missense mutants) is completely resistant to the effects 6-TG imparts on growth and toxin production, R1 (stgP permease mutants) can be sensitized to 6-TG when high concentrations of the drug are used.

FIG 4.

The growth of an stgP mutant but not an hpt mutant is affected by high 6-TG concentrations. Wild-type S. aureus (A), R1 (stgP mutant) (B), and R10 (hpt mutant) (C) were grown in 6-TG (1.3 to 160 μg/ml) in TSB. The OD₆₀₀ was measured after 24 h. Results are from 3 independent experiments. Data are shown as means ± SDs (n = 3). **, P ≤ 0.01; ****, P ≤ 0.0001; n.s., not significant by one-way ANOVA with Dunnett’s multiple-comparison test.

FIG 5.

Hla production by an stgP mutant but not an hpt mutant is affected by high 6-TG concentrations. Wild-type S. aureus, R1 (stgP mutant), and R10 (hpt mutant) were grown in 6-TG (1.3 to 160 μg/ml) in TSB. Secreted proteins in the supernatant from the cultures were TCA precipitated and separated by SDS-PAGE. Protein samples were immunoblotted and probed with anti-alpha toxin antibody in a Western blot.

Prophylactic 6-TG does not inhibit tissue necrosis in subcutaneous infections caused by 6-TG-resistant S. aureus mutants.

We previously demonstrated that prophylactic subcutaneous treatment of mice with 6-TG protects against the necrotic lesion-forming ability of USA300 LAC, presumably via downregulation toxins in vivo (24). Given our results presented above that demonstrate that R1 and R10 are resistant to both the growth inhibiting and toxin-downregulating effects of 6-TG in vitro, we reasoned this should hold true in vivo as well. Therefore, we prophylactically administered 6-TG or a vehicle control before subcutaneously infecting mice with either S. aureus, R1, or R10. As expected, 6-TG significantly reduced the size of necrotic lesions caused by the wild-type USA300 LAC relative to those with the vehicle control. Remarkably, the sizes of the lesions caused by R1 and R10 were unaffected by 6-TG (Fig. 6A and B). Taken together, these data show that while 6-TG has profound negative effects on S. aureus-induced lesion size, it does not affect lesion size caused by R1 and R10. We attribute this to our observation that the prodrug does not perturb Hla production in these mutants at the concentration used in this experiment.

FIG 6.

S. aureusstgP and hpt mutants are resistant to 6-TG in vivo. (A) Representative images showing necrotic skin lesions 4 days postinfection with S. aureus, R1, and R10 with/without prophylactic 6-TG treatment (20 μg). (B) Quantitation of necrotic skin lesion size 4 days postinfection with S. aureus, R1, or R10 with/without prophylactic 6-TG treatment (20 μg). Results are from one experiment with 4 animals per group. Data are shown as means ± SDs. ***, P ≤ 0.001; **, P ≤ 0.01; n.s., not signficant by two-sided unpaired t test.

Coinfections with 6-TG-producing bacteria do not abrogate lesion formation caused by the hpt mutant in a subcutaneous infection.

Next, we questioned whether coinfections with 6-TG-producing bacteria would prevent lesions caused by R1 and R10 in a subcutaneous infection. To test this, we coinfected mice using wild-type S. aureus, R1, or R10 with avirulent S. aureus RN4220 pEmpty (non-6-TG producer) or S. aureus RN4220 ptgs (6-TG producer). These strains were previously used in coinfections with S. aureus, where RN4220 ptgs inhibits S. aureus USA300 lesion formation and RN4220 pEmpty does not (24). Our previous work demonstrated that coinfections with wild-type S. aureus and RN4220 pEmpty resulted in lesion formation, while coinfections with RN4220 ptgs significantly minimized skin damage (Fig. 7A and B) (24). Interestingly, lesions caused by only R10, and not R1, were unaffected by coinfections with RN4220 ptgs (Fig. 7A and B). Altogether, these results suggest that mutations in hpt confer greater resistance to 6-TG than mutations in stgP in vivo.

FIG 7.

S. aureushpt mutants, but not stgP mutants, are resistant to 6-TG-producing bacteria during coinfection. (A) Representative images showing necrotic skin lesions 4 days postinfection with S. aureus, R1, or R10 plus RN4220 pEmpty or plus RN4220 ptgs in a 1:1 ratio. (B) Quantitation of necrotic skin lesion size 4 days postinfection with S. aureus, R1, or R10 plus RN4220 pEmpty or plus RN4220 ptgs in a 1:1 ratio. Results are from one experiment with 4 animals per group. Coinfection data using wild-type S. aureus + RN4220 pEmpty and RN4220 ptgs are duplicated from our previous work (24), and this infection was done in conjunction with R1 and R10 + RN4220 pEmpty or + RN4220 ptgs. Data are shown as means ± SDs. ***, P ≤ 0.001; **, P ≤ 0.01; n.s., not signficant by two-sided unpaired t test.

DISCUSSION

S. aureus is an important human pathogen and a significant cause of morbidity and mortality. Due to the emergence of multidrug-resistant S. aureus strains, treatment options for S. aureus infections have become limited. To discover new antimicrobials to treat these infections, researchers have begun investigating CoNS, since these bacteria cocolonize with S. aureus and compete for survival. One strategy that CoNS use to outcompete S. aureus is the production of various antimicrobial compounds, and these products have been shown to have the potential to treat S. aureus infections by eliminating the bacteria or reducing virulence (10–13, 21, 22).

However, to prolong the usefulness of these antimicrobials, it is important to characterize resistance mechanisms which will provide insight for strategies that can be used to reduce antimicrobial resistance. Indeed, the determinants that cause S. aureus strains to be susceptible or resistant to some CoNS-derived compounds have been characterized to some extent. For instance, autoinducing peptides (AIPs) from CoNS reduce S. aureus virulence by targeting the agr quorum sensing system, but the level of inhibition is dependent on agr classes; certain AIPs better inhibit quorum sensing from certain agr classes than from other classes (21, 22). Resistance against bacteriocins in S. aureus is dependent on levels of expression of mprF, an enzyme that increases the cationic charge of the bacterial cell surface by adding l-lysine to negatively charged phosphatidylglycerol in the cell membrane. This modification results in repulsion of positively charged antimicrobial peptides, including bacteriocins (11). Interestingly, resistance against the antibiotic lugdunin has not been demonstrated (12). By understanding how resistance against antimicrobials occurs, researchers can develop strategies to reduce resistance by using appropriate doses of new antimicrobials and by combination antibiotic therapy.

Our recent study showed that 6-TG, a CoNS-derived compound, is a potent antivirulence compound and can be used as an antimicrobial for drug-resistant S. aureus infections (24). However, there have been no studies done on 6-TG resistance in S. aureus, and characterization of resistance mechanisms needs to be performed to understand how this pathogen copes with this purine biosynthesis inhibitor.

In this study, we show that S. aureus can become spontaneously resistant against 6-TG in vitro under purine-deplete conditions. Using whole-genome sequencing and bioinformatic analyses, we identified that some resistant S. aureus clones harbor frameshift mutations in stgP. This gene is annotated as a putative xanthine/uracil permease family protein and is likely a transporter that imports 6-TG into the cell. Perhaps it is unsurprising that a defect in a putative nucleobase transporter results in 6-TG resistance, since other nucleobase transporters have been shown to import 6-TG in other bacterial species (39, 40). Interestingly, stgP mutants are still susceptible to high 6-TG concentrations (see Fig. 4 and 5), suggesting that there is/are one or more lower-affinity transporters that facilitate 6-TG uptake in S. aureus. One such transporter could be PbuX, a putative xanthine permease that is encoded in an operon involved in guanine metabolism (41), and this idea requires further investigation. Nonetheless, our results suggest that stgP is the major transporter that imports 6-TG into S. aureus. These in vitro observations are notable when one considers the coinfection data (Fig. 7) demonstrating that expression of the tgs operon and, by extension, production of 6-TG by a heterologous bacterium, while not affecting pathogenic potential of the hpt mutant, diminishes the virulence of the stgP mutant. We speculate that the continued production of 6-TG in this confined subcutaneous environment likely results in localized 6-TG concentrations that surpass the threshold concentration required for the function of a secondary 6-TG transporter to become relevant and thus inhibit virulence.

A number of independent mutations that were selected based on resistance to 6-TG by S. aureus were missense mutations in hpt, an enzyme important in the purine salvage pathway (42, 43). In humans, Hpt is the enzyme that converts 6-TG from a prodrug into the active 6-thioguanine monophosphate (TGMP) and is used to treat various cancers and inflammatory bowel disease (29, 31, 44–46). Moreover, 6-TG has been shown to interact with Hpt in Mycoplasma pneumoniae (33). Based on this knowledge, we hypothesize that 6-TG is converted into TGMP by Hpt and that TGMP is the active metabolite that blocks de novo purine biosynthesis in S. aureus. Hence, mutations in S. aureus Hpt may prevent 6-TG from being converted into its active antimicrobial form, resulting in 6-TG resistance. Interestingly, all mutations identified in hpt were nucleotide substitutions, suggesting that the enzyme is impaired in some manner but still functional. Indeed, Hpt has pleiotropic functions in bacteria and has been shown to play significant roles in virulence and survival in several Gram-positive bacteria, including Listeria monocytogenes and Bacillus subtilis (32, 47). However, investigations on S. aureus Hpt are limited, and the gene product has only been shown, thus far, to be important for survival under amino acid-limiting conditions and high salt concentrations in vitro (48). Therefore, it will be interesting to test whether Hpt is important for other aspects of S. aureus biology, including its role in infection and virulence.

The effectiveness of medical intervention for bacterial infections is heavily dependent on the number of antimicrobials to which bacteria are susceptible. Due to the emergence of multidrug-resistant S. aureus strains, many antibiotics that were once useful have become obsolete. To overcome this problem, researchers have focused their endeavors on discovering and characterizing new antimicrobial agents to treat infections caused by drug-resistant S. aureus strains. However, efforts must also be focused on identifying antimicrobial resistance mechanisms, as this knowledge could help to prolong the usefulness of antibacterial drugs. Our study reveals two mechanisms by which S. aureus can become resistant to the antivirulence compound 6-TG. By understanding how 6-TG resistance occurs, it will be possible to develop strategies to mitigate resistance against this compound.

MATERIALS AND METHODS

Ethics statement.

All experiments involving animals were reviewed and approved by the Animal Use Subcommittee of the University of Western Ontario and were performed according to the Canadian Council on Animal Care guidelines.

Bacterial strains and plasmids.

Bacterial strains and plasmids are listed in Table 3. Unless otherwise noted, strains were cultured at 37°C with shaking at 200 rpm. Escherichia coli was grown in Luria broth (LB) or on LB agar containing 100 μg/ml ampicillin to maintain plasmids when necessary. All S. aureus strains were grown in tryptic soy broth (TSB) or on TSB agar plates (TSA). Staphylococcal plasmids were maintained with growth in 10 μg/ml chloramphenicol when necessary. RPMI 1640 medium was used as purine-deplete medium.

TABLE 3.

Bacterial strains used in this study

Strain	Description	Source or reference
S. aureus USA300	USA300 LAC, cured of resistance plasmids	Lab stock
S. aureus RN4220	rK− mK+; capable of accepting foreign DNA	Lab stock
R1	6-TG-resistant S. aureus mutant	This study
R3	6-TG-resistant S. aureus mutant	This study
R5	6-TG-resistant S. aureus mutant	This study
R6	6-TG-resistant S. aureus mutant	This study
R7	6-TG-resistant S. aureus mutant	This study
R8	6-TG-resistant S. aureus mutant	This study
R9	6-TG-resistant S. aureus mutant	This study
R10	6-TG-resistant S. aureus mutant	This study
R11	6-TG-resistant S. aureus mutant	This study
R1 pEmpty	R1 carrying pALC2073	This study
R1 pstgP	R1 carrying the wild-type version of stgP from S. aureus	This study
R10 pEmpty	R10 carrying pALC2073	This study
R10 phpt	R10 carrying the wild-type version of hpt from S. aureus	This study
RN4220 pEmpty	RN4220 carrying pALC2073	24
RN4220 ptgs	RN4220 carrying the 6-TG biosynthetic operon from S. chromogenes ATCC 43764 in pALC2073	24

Disc diffusion assays.

The optical density at 600 nm (OD₆₀₀) values of bacteria were normalized to 1.0. Sterile cotton swabs were used to spread plate a suspension of the culture onto solid media. Sterile paper discs were placed on top of the bacteria, and 10 μl of 10 mg/ml (100 μg) 6-thioguanine dissolved in 0.1 M NaOH was dropped onto the disc. Plates were incubated at 37°C for 24 h unless specified otherwise.

Determining the MIC of 6-TG.

Wild-type S. aureus and 6-TG resistant S. aureus mutants were streaked on TSA, and single colonies were inoculated in TSB to grow overnight at 37°C. The OD₆₀₀ values of bacteria were normalized to 1.0, and 10 μl of the suspension (OD₆₀₀ of 0.01) was inoculated into 1 ml of CDM that contained 6-TG (3.12 to 400 μg/ml). The components of CDM were described previously (49). Briefly, CDM was composed of the following (final concentration): alanine (672 μM), arginine (287 μM), aspartic acid (684 μM), cysteine (166 μM), glutamic acid (680 μM), glycine (670 μM), histidine (129 μM), isoleucine (228 μM), leucine (684 μM), lysine (342 μM), methionine (20 μM), phenylalanine (240 μM), proline (690 μM), serine (285 μM), threonine (260 μM), tryptophan (50) μM, tyrosine (275 μM), valine (684 μM), thiamine (56 μM), nicotinic acid (10 μM), biotin (0.04 μM), pantothenic acid (2.3 μM), MgCl₂ (1,000) μM, CaCl2 (100 μM), monopotassium phosphate (40,000 μM), dipotassium phosphate (14,700 μM), sodium citrate dehydrate (1,400 μM), magnesium sulfate (400 μM), ammonium sulfate (7,600 μM), and glucose (27,753 μM). Cultures were grown for 24 h at 37°C, and OD₆₀₀ was measured. The concentration of 6-TG that completely inhibited growth was determined to be the MIC.

Whole-genome sequencing.

S. aureus genomic DNA was isolated by phenol-chloroform extraction from single colonies inoculated into overnight liquid cultures. The extracted genomic DNA for wild-type S. aureus and resistant clones was tagmented using the Nextera tagmentation kit (Illumina) using a modified protocol (50). The tagmented products were sequenced on an Illumina NextSeq 550 with paired-end sequencing (2 by 150 bp) at the Microbial Genome Sequencing Center in Pittsburgh, PA. DNA sequence reads were trimmed using Trimmomatic version 0.36 (51) with the following parameters: ILLUMINACLIP, NexteraPE:2:30:10:8:true; LEADING, 20; TRAILING, 20; SLIDINGWINDOW, 4:20; MINLEN, 36. The reads were assembled using the SPAdes version 3.13 assembler (52) in careful mode. The assembled genome of wild-type S. aureus was aligned to the published S. aureus USA300_FPR3757 genome (CP000255) and annotated using Prokka version 1.12 (53). For SNP/indel analyses, the resistant S. aureus clones’ genomes were then aligned to the wild-type S. aureus genome in Geneious Prime 2020.1.2 (Biomatters, Ltd., Aukland, New Zealand) using the Geneious aligner and the default settings. Coverage statistics for the whole-genome sequencing can be found in Table 4.

TABLE 4.

Genome sequencing coverage statistics

S. aureus clone	Mean depth of coverage (×)	Median Q30 score (%)
Wild type USA300-LAC	198	≥89.8
R1	174	≥90.5
R3	193	≥90.8
R5	191	≥90.6
R6	207	≥90.8
R7	182	≥89.7
R8	157	≥90.4
R9	185	≥90.5
R10	164	≥89.7
R11	250	≥90.6

Cloning of stgP and hpt genes.

Genomic DNA was isolated from S. aureus by phenol-chloroform extraction. Primers containing a KpnI or SacI restriction site were used to amplify stgP and hpt (Table 5). The amplified product was ligated into pALC2073 that was similarly digested and was transformed into E. coli. Plasmids isolated from E. coli were introduced into S. aureus RN4220.

TABLE 5.

Primers used in this study

Name	Sequence	Description
USA300_stgP_F_KpnI	TTTTTTGGTACCGTTCGCTTTTGGTATTGTTT	Forward primer to amplify stgP from S. aureus
USA300_stgP_R_SacI	TTTTTTGAGCTCCGTTTTTTGTGTCTGGATTTA	Reverse primer to amplify stgP from S. aureus
USA300_hpt_F_KpnI	TTTTTTGGTACCAAAAATGGAGATGAATAGCG	Forward primer to amplify hpt from S. aureus
USA300_hpt_R_SacI	TTTTTTGAGCTCATTTCTCAAACGCATAGTAA	Reverse primer to amplify hpt from S. aureus

Assessing the effect of 6-TG on S. aureus growth kinetics.

Single colonies of S. aureus wild type, R1, and R10 were inoculated into 5 ml TSB and grown overnight. Thirty milliliters of TSB with 10 μg/ml of 6-TG or a vehicle control was inoculated with bacteria from S. aureus, R1, or R10 overnight cultures such that the starting OD₆₀₀ was 0.05. The OD₆₀₀ values of the cultures were measured every hour for 8 h. The OD₆₀₀ was also measured at 12 and 24 h.

Nebraska transposon insertion library screening.

Frozen stocks were inoculated directly into TSB and grown overnight in 96-well plates. Ten microliters of the culture was inoculated onto RPMI medium supplemented with 3 μg/ml 6-TG, and growth of each mutant was compared to that of wild-type S. aureus. Mutants that appeared more sensitive or resistant to 6-TG than wild-type S. aureus were screened again. In this second screen, wild-type S. aureus and the mutants were grown in TSB overnight, and OD₆₀₀ was normalized to 1.0. Cultures were serially diluted and plated onto RPMI medium and RPMI medium supplemented with 3 μg/ml 6-TG, and relative 6-TG-mediated growth inhibition of each mutant was compared to that of the wild type. Mutants that grew poorly on RPMI medium alone were excluded from further analysis.

TCA precipitation and gel electrophoresis.

Single colonies of S. aureus grown on TSA were inoculated in TSB with 1.3 to 160 μg/ml 6-TG or a vehicle control and grown to an OD₆₀₀ of ∼6.0 to 7.0. Proteins in the supernatant were harvested through trichloroacetic acid (TCA) precipitation. Briefly, supernatant derived from 6.0 OD₆₀₀ units of bacterial culture were mixed with equal volumes of ice-cold 20% (wt/vol) TCA for 3 h, washed twice with ice-cold 70% ethanol, and air dried. The proteins were dissolved in 1× Laemmli SDS-PAGE reducing buffer and boiled for 10 min. Proteins were separated on 12% SDS-PAGE gels and stained with InstantBlue ultrafast protein stain overnight for visualization of protein bands.

Western blotting.

Western blot analysis on alpha toxin was performed by transferring the proteins in the SDS-PAGE gel to a nitrocellulose membrane. The membrane was then blocked with phosphate-buffered saline (PBS) supplemented with 0.1% (vol/vol) Tween 20 and 5% (wt/vol) milk. After blocking, the membrane was incubated with rabbit anti-Hla (1:500) polyclonal antibody (Sigma-Aldrich) for 2 h and washed with PBS-Tween 20 (3 times). Donkey anti-rabbit IRDye 800 (1:20,000) (Rockland) was used as a secondary antibody for 45 min, and the membrane was washed with PBS-Tween 20 (3 times) prior to imaging. Membranes were scanned using an Odyssey Clx imaging system (Li-Cor Biosciences).

Subcutaneous skin infections.

Eight- to ten-week-old female BALB/c mice were shaved and subcutaneously injected with 25 μl of 800 μg/ml 6-TG (equivalent to 20 μg 6-TG or 1 mg/kg) or a vehicle control 1 day before infection. This dosage was chosen based on the therapeutic concentration of 6-TG used in both mice and in humans for inflammatory bowel disease (54, 55). Overnight cultures (grown in TSB) of S. aureus USA300, R1, and R10 were subcultured to an OD₆₀₀ of 0.05 in TSB and grown to an OD₆₀₀ of 2.0 to 2.2. Bacterial cells were pelleted and washed in PBS twice. Bacterial cells were then normalized to an OD₆₀₀ of 3.7, and 25 μl of the bacterial suspension (∼5.0 × 10⁷ CFU) was mixed with 25 μl of 800 μg/ml 6-TG and subcutaneously injected into the animals. Infected mice were monitored daily for 4 days and sacrificed at 96 h postinfection. Lesion sizes were analyzed in ImageJ.

Subcutaneous skin coinfections.

Eight- to ten-week-old female BALB/c mice were shaved 1 day before infection. Overnight cultures (grown in TSB) of S. aureus USA300, R1, and R10, RN4220 pEmpty, and RN4220 ptgs were subcultured to an OD₆₀₀ of 0.05 in TSB and grown to an OD₆₀₀ of 2.0 to 2.2. RN4220 pEmpty and RN4220 ptgs were grown in the presence of antibiotic. Bacterial cells were pelleted and washed in PBS twice. S. aureus USA300 (wild type, R1, and R10) was normalized to an OD₆₀₀ of 3.7, while RN4220 pEmpty was normalized to an OD₆₀₀ of 5.4 and RN4220 ptgs was normalized to an OD₆₀₀ of 10.8. Twenty-five microliters (∼5.0 × 10⁷ CFU) of S. aureus USA300 (wild type, R1, or R10) was mixed with 25 μl (∼5.0 × 10⁷ CFU) of RN4220 pEmpty or RN4220 ptgs and subcutaneously injected into the animals. Infected mice were monitored daily for 4 days and sacrificed at 96 h postinfection. Lesion sizes were analyzed in ImageJ.

Data availability.

The raw sequence reads for the 6-TG-resistant S. aureus LAC mutants (R1, R3, R5, R6, R7, R8, R9, R10, and R11) sequenced in this study can be found in the NCBI Sequence Read Archive under BioProject accession PRJNA721392.