pta and lipL are genetically linked in Staphylococci

S. aureus preferentially uses glycolysis to metabolize glucose (Fig 1A) [4,19,43]. The lipoic acid-dependent PDH complex serves as a major exit point from the glycolytic pathway, where the decarboxylation of pyruvate leads to the generation of acetyl-CoA. During periods of carbon overflow in vitro, acetyl-CoA is predominantly shuttled into the Pta-AckA pathway to generate ATP and acetate via substrate-level phosphorylation (Fig 1A), in part because carbon catabolite repression suppresses TCA cycle gene expression [44,45]. Following glucose consumption, acetate is reassimilated via acetyl-CoA synthase (Acs), for use in the TCA cycle (Fig 1A). Central to glycolysis and TCA cycle function is LipL-dependent lipoylation of the E2 subunits of PDH and OGDH (Fig 1A). In a prior study, we determined that the lipL (SAUSA300_0571) open reading frame begins two nucleotides downstream from the stop codon of pta (SAUSA300_0570), implying a gene pairing that links lipoyl transfer activity with the Pta-AckA pathway [26]. Considering pta and ackA are genetically linked in classically studied microorganisms such as Escherichia coli, we wondered if this unusual pta-lipL linkage observed in S. aureus was widespread. We conducted a comparative synteny analysis of the pta gene environment. We surveyed representative Gram-negative or Gram-positive bacterial pathogens and Mycobacterium tuberculosis, as well as commensals Lactococcus lactis subs. cremoris, Bacillus subtilis, and E. coli, and found that most Gram-negative pathogens surveyed were likely to have ackA and pta positioned in an operon, like E. coli (Fig 1B). By contrast, analysis of M. tuberculosis and Gram-positive pathogens showed that pta and lipL were in close genetic proximity for some Gram-positives (Bacillus anthracis and S. aureus), whereas others were positioned at greater than 70kb nucleotide distances (~70kb for Streptococcus pyogenes, ~ 2500kb for Listeria monocytogenes, and ~300kb for Enterococcus faecalis) (Fig 1C). S. aureus was the only species with pta and lipL genetically linked in a putative operon, potentially implying a unique genetic arrangement for Staphylococci (Fig 1C). Indeed, coagulase-positive and negative Staphylococci all have the same pta-lipL organization, indicating it is a conserved feature of the genus (Fig 1D).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. pta and lipL are genetically linked in Staphylococci.

(A) Model of acetate fermentation in S. aureus. Glucose is converted to pyruvate, the primary end-product of glycolysis. Pyruvate has several fates, one of which is its oxidative decarboxylation by pyruvate dehydrogenase (PDH) to generate Acetyl-CoA. Pyruvate can also be directly decarboxylated (CidC) to produce acetate. Acetyl-CoA can enter the tricarboxylic acid (TCA) cycle or is used by the phosphotransacetylase (Pta) to generate Acetyl-phosphate (Ac-P), the substrate for acetate kinase (AckA). The activity of AckA generates Acetate and ATP. PDH activity is regulated by lipoic acid (lipoyl) attachment, which is mediated by delivery of lipoic acid from the H subunit of the glycine cleavage system (GcvH) by the amidotransferase, LipL. (B-D) Synteny analysis of pta and lipL genes. (B) Phylogenetic tree of Gram-negatives, including pta synteny analysis adjacent to the relevant microorganism. (C) Phylogenetic tree of Gram-positives and M. tuberculosis, including pta synteny analysis adjacent to the respective microorganism. (D) Phylogenetic tree of Staphylococci including pta synteny analysis. (E) Normalized and mapped transcript reads across the pta-lipL region in WT and a ΔlipL mutant. (F) qRT-PCR analysis of RNA extracted from WT, Δpta, and ΔlipL strains at 6 hours of growth. ****, p < 0.0001 by one-way ANOVA with Tukey’s post hoc test. qRT-PCR experiments were conducted with two independent biological replicates in technical triplicate. Boxplots indicate the median and quartiles. Whiskers indicate the range.

https://doi.org/10.1371/journal.ppat.1014183.g001

To establish that S. aureus co-transcribes pta and lipL, we conducted RNA sequencing of WT S. aureus and a ΔlipL mutant grown to late exponential phase. The read analysis indicated that pta and lipL transcripts consist of a single mRNA encoding both genes (Fig 1E). The deletion of lipL led to the generation of a truncated transcript but did not negatively impact pta expression (Fig 1E). To validate the RNAseq results, we measured transcript levels of pta, lipL and the junction between the pta and lipL genes via qRT-PCR. Transcript levels were similar between the pta, lipL, and the junction in WT S. aureus (Fig 1F). Junctional amplicons were not detected from the Δpta and ΔlipL mutants (Fig 1F). Altogether, these data indicate that the genetic link between pta and lipL is conserved in Staphylococci and the genes are co-transcribed.

Loss of either pta or lipL compromises acetate production

Given the evidence for co-transcription of pta and lipL in S. aureus and the role of Pta in overflow metabolism, we sought to determine how these genes contribute to acetate production via Pta-AckA and metabolic flux in general. We generated in-frame single and double deletions of pta and lipL in the methicillin-resistant Staphylococcus aureus (MRSA) strain LAC [46,47]. We complemented pta and lipL in single copy in the chromosome with constitutive expression [48], and subsequently determined if the loss of pta or lipL resulted in growth defects or shifts in metabolite production. The Δpta mutant had a marginal delay in exit from logarithmic phase growth in tryptic soy broth (TSB) compared to WT (Fig 2A). A ΔlipL mutant, on the other hand, had compromised growth kinetics, including dramatically reduced growth in TSB, consistent with previous work from our group (Fig 2A) [1]. A significant component of the ΔlipL mutant growth defect in TSB is due to loss of LipL-dependent lipoylation of the BCODH complex, which is essential for branched-chain fatty acid synthesis [5]. Indeed, growth of ΔlipL mutant was substantially restored in TSB supplemented with branched-chain carboxylic acids (BCCAs) (isobutyric acid, 2-methylbutyric acid, and isovaleric acid) that bypass the requirement for BCODH or with use of the lipL complement strain [1,26] (Fig 2A). To test for defects in lipoylation of E2 subunits of enzyme complexes, we conducted immunoblots on whole cell lysates of the Δpta and ΔlipL mutants and their complement strains grown in TSB + BCCAs. Consistent with prior results, the ΔlipL mutant had defective lipoylation compared to WT [26] (Fig 2B), whereas the Δpta mutant had evidence for increased lipoylation on E2 subunits (Fig 2B). Thus, while lipoylation itself is not a direct effect of Pta, the degree of lipoylation on E2 subunits is increased in its absence (see below). These data indicate that the effects of pta and lipL on growth kinetics and lipoylation in broth are separable and largely driven by LipL-dependent effects on BCODH activity.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Loss of either pta or lipL compromises acetate production.

(A) Growth of WT, Δpta, ΔlipL, Δpta + pta, and ΔlipL + lipL strains in TSB and TSB + BCCAs [10 mM isobutyric acid (IB), 9 mM 2-methylbutyric acid (2MB), 9 mM isovaleric acid (IV) + 10 mM sodium acetate (NaAc)] containing 14 mM glucose. (B) α-lipoic acid immunoblots of whole cell lysates from the indicated strains after 9-hours of growth in TSB medium. The presented blot and Coomassie-stained gel are representative of at least three independent experiments. (C) Quantification of glucose, acetate, lactate and formate from WT, Δpta, Δpta + pta culture supernatants over time in TSB with 14 mM glucose. (D) Quantification of glucose, acetate, lactate and formate from WT, ΔlipL, ΔlipL + lipL culture supernatants over time in TSB with 14 mM glucose and BCCA supplementation. Each metabolite quantification assay is representative of three independent experiments, with each timepoint measured in technical triplicate. Errors bars indicate standard deviation from the mean. Some lines on growth curve graphs (Δpta + pta, and ΔlipL + lipL) are obscured because they overlap.

https://doi.org/10.1371/journal.ppat.1014183.g002

We next assessed glycolytic flux over time by measuring glucose and acetate levels in media containing 14 mM glucose under aerobic conditions to promote carbon overflow metabolism [28,49]. In addition to acetate, we measured lactate and formate as alternative pyruvate fermentation products in S. aureus [31,32,50]. As previously reported, glucose was rapidly consumed within 6 hours of growth along with a commensurate increase in acetate that peaked at 6 hours and was reassimilated in stationary phase (Fig 2C) [27,28]. The Δpta mutant consumed all the glucose in the media by 6 hours with no differences compared to the WT strain (Fig 2C). Peak acetate levels from the Δpta mutant were ~13 mM, a ~ 57% reduction compared to WT (Fig 2C). There were no differences in lactate or formate levels produced by the Δpta mutant compared to WT (Fig 2C). Conversely, the ΔlipL mutant had a delay in the glucose consumption rate relative to WT (Fig 2D). Peak acetate production from the ΔlipL mutant was around 8 mM, a ~ 70% reduction compared to WT (Fig 2D). There was also a commensurate increase in lactate production by the ΔlipL mutant, which peaked at ~8 mM at 6 hours (Fig 2D). Growth and metabolite quantification assays using glucose-free media or low glucose supplementation (3.5 mM) established that acetate production requires LipL and that levels depend on the amount of glucose added. Compensatory increases in lactate production by a ΔlipL mutant are only observed when glucose is provided as a carbon source (14 mM and 3.5 mM) (S1 Fig and Fig 2D). Together, these data suggest that both Pta and LipL are required for optimal acetate fermentation during overflow metabolism and under conditions where glucose is limiting. While deletion of either pta or lipL leads to compensatory acetate production, presumably by other acetogenic pathways, lactate fermentation is only induced upon deletion of lipL in the presence of glucose, potentially to offset a pyruvate bottleneck during overflow that is associated with inactivation of PDH.

LipL and Pta coordinate PDH activity and acetate production

pta and lipL are co-transcribed (Fig 1). The deletion of either gene alone shifts acetate production and metabolic flux either downstream (Δpta) or upstream (ΔlipL) of PDH (Figs 1A and 2). We hypothesized that the tandem expression of pta-lipL might promote coordination of pyruvate decarboxylation with the delivery of acetyl-CoA to Pta-AckA to improve metabolic efficiency. To test this, we assessed the effect of loss of the entire pta-lipL locus on growth and metabolite production. The Δpta ΔlipL mutant had severely compromised growth in TSB, which was largely complemented when TSB was supplemented with BCCAs or when using a Δpta ΔlipL + lipL complement strain (Fig 3A). The Δpta ΔlipL + pta strain did not restore growth in TSB without BCCAs (Fig 3A). Immunoblots on whole cell lysates of the Δpta ΔlipL double mutant and complement strains revealed that only those strains lacking lipL had defective lipoylation profiles compared to WT (Fig 2B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Pta and LipL coordinate PDH activity and acetate production.

(A) Growth of WT, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL strains in TSB and TSB + BCCAs [10 mM isobutyric acid (IB), 9 mM 2-methylbutyric acid (2MB), 9 mM isovaleric acid (IV) + 10 mM sodium acetate (NaAc)] containing 14 mM glucose. (B) Quantification of glucose, acetate, lactate, and formate from WT, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL culture supernatants over time in TSB with 14 mM glucose and BCCA supplementation. Each metabolite quantification assay is representative of three independent experiments, with each timepoint measured in technical triplicate. Error bars indicate standard deviation from the mean.

https://doi.org/10.1371/journal.ppat.1014183.g003

The WT and Δpta ΔlipL + lipL strains consumed all glucose by 6 hours, whereas the Δpta ΔlipL double mutant and Δpta ΔlipL + pta had delays in glucose consumption (Fig 3B). Notably, the Δpta ΔlipL double mutant did not produce additional acetate (Fig 3B). The Δpta ΔlipL + pta strain mirrored this phenotype and acetate levels did not accumulate beyond ~5 mM, a ~ 83% reduction compared to peak acetate production in the WT and Δpta ΔlipL + lipL strains (Fig 3B). Intriguingly, the Δpta ΔlipL + lipL strain had delayed acetate reassimilation. Furthermore, compensatory lactate production by the Δpta ΔlipL double mutant and Δpta ΔlipL + pta strains was consistent with a redirection of pyruvate toward lactate fermentation (Fig 3B). Taken together, these data confirm that defects in growth associated with the Δpta ΔlipL mutant are due to activities of LipL that are independent of Pta-AckA and flux through PDH. Furthermore, loss of both pta and lipL halts acetate production entirely, redirecting pyruvate towards alternative fates, such as lactate fermentation. Constitutive expression of lipL favors alternative modes of acetate production that do not depend on Pta-AckA.

Pta and LipL proteins directly interact

Prior functional and computational studies suggest that metabolic flux can be enhanced by protein-protein interactions and protein complex formation [13–15]. Considering our demonstration of the coordinated expression of pta and lipL and the necessity of both for optimal acetate production, we hypothesized there might be a direct interaction between Pta and LipL that could facilitate acetate fermentation. We first used AlphaFold3 to predict the likelihood of a direct interaction between Pta and LipL. The AlphaFold3 model predicted a potential interaction between Pta and LipL, with a template modeling of 0.76 and an interface predicted template modeling (ipTM) score of 0.61 (Fig 4A). Considering the ipTM score of the predicted interaction falls at the likelihood threshold for a potential interaction, we opted to validate this prediction by expressing and purifying Pta (Pta-6xHis) and LipL (LipL-6xHis) and conducting a biophysical assessment of a direct interaction using microscale thermophoresis (MST). Recombinant Pta was labeled with the amine-reactive dye, N-hydroxysuccinimide (NHS) and titrated against unlabeled LipL (Fig 4B). MST determined that Pta interacts with LipL in solution with moderate affinity (Kd of 490 + /- 105 nM) (Fig 4B). To determine if the proteins interact in vivo, we cloned the coding sequences of both pta and lipL into the adenylate cyclase-based Bacterial two-hybrid (BACTH) system [51,52]. Pta and LipL interacted strongly when expressed in E. coli, achieving similar levels of β-galactosidase activity as the leucine zipper positive control (Fig 4C). In considering the possibility that an interaction between Pta and LipL might improve efficiency of metabolic flux through PDH to Pta-AckA, we reasoned that loss of Pta would enhance the demand for LipL-mediated delivery of lipoic acid to E2 proteins of PDH and OGDH as a compensatory measure to drive carbon flux. Indeed, a Δpta mutant had increased lipoylation on E2-PDH and E2-OGDH at 3 and 6 hours compared with WT and the Δpta + pta strain (Fig 4D and 4E). This increase in lipoylation was not associated with an increase in pdhC (E2-PDH) or sucB (E2-OGDH) transcripts at 3 or 6 hours (S2A Fig), suggesting increased delivery of lipoic acid to E2-PDH and E2-OGDH by LipL. Collectively, these results indicate Pta and LipL directly interact in vitro and in vivo. Furthermore, the loss of Pta results in increased delivery of lipoic acid to E2-PDH and E2-OGDH as a compensatory measure.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Pta and LipL proteins directly interact.

(A) AlphaFold3 prediction of the interaction between Pta (blue) and LipL (pink). pTM (0.76) and an ipTM (0.61) are displayed graphically below the model. (B) Microscale thermogram of Pta-NHS titrated against the recombinant LipL. The top panel shows the thermophoretic time-traces of four independent experiments, and the blue and pink areas represent time spans used to obtain the fluorescence cold (F_c) and hot (F_h) regions, respectively. The middle panel shows the binding curve with the line of best fit using the 1:1 binding model with an error surface projection confidence of 95%. The residuals between the data and fit are shown in the bottom panel. Kd is presented as the mean ± standard deviation of four independent replicates. (C) Bacterial adenylate cyclase two-hybrid (BACTH) assay was used to assay an interaction between Pta and LipL. Graph displays β-galactosidase activity (Miller Units) from E. coli BTH101 co-transformed with pUT18C-pta and pKT25-lipL, pUT18C-zip and pKT25-zip (positive control – PC) or pUT18C and pKT25 vectors (negative control – NC). ***, p < 0.001 by one-way ANOVA with Tukey’s post hoc test. Cultures of E. coli BTH101strains co-transformed with the aforementioned plasmid combinations were also spotted on LB agar plates with X-Gal as an indicator. Data are representative of at least three independent experiments. (D-E) α-lipoic acid immunoblots of whole cell lysates from WT, Δpta, and Δpta + pta strains after 3 and 6-hours of growth in TSB medium. Densitometric quantification of E2-PDH and E2-OGDH bands in Δpta and Δpta + pta strains relative to WT bands. *, p < 0.05; **, p < 0.01 by Kruskal-Wallis test with Dunn’s post hoc analysis. The presented blots and Coomassie-stained gels are representative of four independent experiments. The mean and standard error of the mean are shown.

https://doi.org/10.1371/journal.ppat.1014183.g004

The coordinated activity of Pta and LipL is important during infection of the skin, whereas additional functions of LipL promote systemic infection

Previous work from our group found that LipL is required for optimal S. aureus infection in both a murine sepsis model and a murine skin and soft tissue infection model [26]. Follow up studies determined that S. aureus is mostly refractory to lipoic acid deficiency in murine skin but not kidneys on account of its ability to bypass LipL-dependent branched-chain fatty acid synthesis by incorporating host unsaturated fatty acids [5]. These results do not preclude the possibility that the coordinated activity of Pta and LipL that drives acetate fermentation might contribute to infection in one or both infection models. To test this, we infected mice intravenously or intradermally with WT, Δpta, Δpta + pta, ΔlipL, ΔlipL + lipL, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL strains. At 96 hours post intravenous infection, kidneys from Δpta mutant-infected animals had ∼100-fold fewer CFUs compared to WT (Fig 5A). By contrast, the kidneys from ΔlipL mutant-infected mice had ~ 10000-fold fewer CFUs (Fig 5A). The kidneys from Δpta ΔlipL, and Δpta ΔlipL + pta infected mice phenocopied ΔlipL mutant-infected mice with ~10000-fold fewer CFUs (Fig 5A). Systemic infection with Δpta + pta, ΔlipL + lipL, and Δpta ΔlipL + lipL complement strains restored CFUs in the kidney to nearly WT levels (Fig 5A). At 72 hours post intradermal infection, abscesses from mice infected either with Δpta, ΔlipL, and Δpta ΔlipL mutants phenocopied one another with ∼10-fold fewer CFUs compared to WT (Fig 5B). Infection with the Δpta + pta and ΔlipL + lipL complement strains restored abscess CFUs to nearly WT levels (Fig 5B). Infection with either Δpta ΔlipL + pta or Δpta ΔlipL + lipL strains did not restore CFU to WT levels (Fig 5B). Considering the identical attenuation for Δpta, ΔlipL, and Δpta ΔlipL, Δpta ΔlipL + pta or Δpta ΔlipL + lipL strains during infection of the skin, we conclude that the coordinated activity between Pta and LipL is required for S. aureus survival at this site. In contrast, during systemic infection, the contributions of Pta and LipL to virulence are separable and dominated by independent roles for LipL in promoting branched fatty acid synthesis, as we have previously published [5].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. The coordinated activity of Pta and LipL is important during infection of the skin, whereas additional functions of LipL promote systemic infection.

(A) Bacterial burden in the kidneys of mice 96 hours after bloodstream infection with 1.0 x 10⁷ CFU of WT, Δpta, Δpta + pta, ΔlipL, ΔlipL + lipL, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL strains. Animal numbers displayed are as follows: WT, N = 36; Δpta, N = 29; Δpta + pta, N = 28; ΔlipL, N = 20; ΔlipL + lipL, N = 19; Δpta ΔlipL, N = 20; Δpta ΔlipL + pta, N = 8; and Δpta ΔlipL + lipL, N = 8. ns, not significant; *, p < 0.05; ***, p < 0.001; ****, p < 0.0001 by Kruskal-Wallis test with Dunn’s post hoc analysis. (B) Bacterial burden in the skin of mice 72 hours after intradermal infection with 1.0 x 10⁷ CFU of WT, Δpta, Δpta + pta, ΔlipL, ΔlipL + lipL, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL. Animal numbers displayed are as follows: WT, N = 16; Δpta, N = 16; Δpta + pta, N = 16; ΔlipL, N = 16; ΔlipL + lipL, N = 16; Δpta ΔlipL, N = 16; Δpta ΔlipL + pta, N = 16; and Δpta ΔlipL + lipL, N = 16. ns, not significant; **, p < 0.01; ****, p < 0.0001 by Kruskal-Wallis test with Dunn’s post hoc analysis. log₁₀ CFU per organ or abscess is displayed for each infected mouse along with the median as a measure of central tendency. Graphs represent combined data from at least two independent experiments.

https://doi.org/10.1371/journal.ppat.1014183.g005

Loss of lipL induces significant transcriptional changes in S. aureus

Our data suggest that a block to exit from glycolytic metabolism at PDH redirects metabolic flux on account of pyruvate accumulation in the cell. To test this, we quantified intracellular pyruvate levels from WT, Δpta, Δpta + pta, ΔlipL, ΔlipL + lipL, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL strains at 3 hours. We found that all strains harboring a ΔlipL mutation had approximately twice the amount of intracellular pyruvate compared to WT (~745 pmol for a ΔlipL mutant versus ~363 pmol for WT) (Fig 6A). The redirection of pyruvate metabolism is further substantiated by increases in lactate and Pta-AckA - independent generation of acetate (Figs 2D and 3B). Given the importance of LipL in activating the PDH, OGDH, and BCODH complexes [1,26], we hypothesized that, in addition to the shifts in glucose metabolism at the PDH node, LipL might also drive a metabolic landscape that is shifted in a ΔlipL mutant on account of disruptions to TCA cycle activity and branched fatty acid synthesis. To test this possibility, we grew WT and ΔlipL mutant strains to late-exponential phase and performed bulk RNA-sequencing. Our results confirmed several major changes in gene expression patterns between the WT and ΔlipL mutant (Fig 6B and 6C). The ΔlipL mutant significantly upregulated genes involved in virulence, histidine catabolism and oxidative metabolism and significantly downregulated genes involved in purine nucleotide biosynthesis and branched-chain amino acid biosynthesis (Fig 6D). Specifically, the loss of lipL increased expression of several cytoxin genes (lukS-PV, hla, psmβ1, and psmβ2), gudB (glutamate dehydrogenase), the hut operon (histidine uptake) and the qox operon (electron transport chain), while decreasing expression of the pur operon (purine biosynthesis), as well as the leu and the ilv operons (branched-chain amino acid biosynthesis) (Fig 6E). Furthermore, we saw substantially increased levels of the pyruvate oxidase cidC (SAUSA300_2477) (Fig 6C, 6E and 6F). CidC generates acetate directly from pyruvate and does not require PDH activity [37]. We also saw increased levels of alsD transcripts (Fig 6E). AlsSD diverts carbon flow towards neutral byproducts to prevent CidC-mediated acidification of the cell [36]. To validate our RNAseq results, we assessed cidC transcript levels in WT, Δpta, ΔlipL, Δpta + pta, and ΔlipL + lipL strains via qRT-PCR. Indeed, we found a ~ 8-fold increase of cidC transcripts in both the Δpta and ΔlipL mutants (Fig 6F). Overall, RNAseq analysis confirmed that the loss of lipL significantly alters the S. aureus transcriptome, including changes to pathways involved in virulence, oxidative metabolism, amino acid metabolism, and nucleotide biosynthesis. Furthermore, the RNAseq dataset revealed increased levels of cidC transcripts, which were confirmed by qRT-PCR in both the Δpta and ΔlipL mutants.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Loss of lipL induces significant transcriptional changes in S. aureus.

(A) Quantification of intracellular pyruvate from WT, Δpta, ΔlipL, Δpta + pta, ΔlipL + lipL, Δpta ΔlipL, Δpta ΔlipL + pta, and Δpta ΔlipL + lipL strains after 3 hours of growth in TSB. ns, not significant; **, p < 0.01 by Kruskal-Wallis test with Dunn’s post hoc analysis. Graphs represent combined data from three independent experiments. (B) Principal component analysis (PCA) of RNA sequencing datasets for the WT and ΔlipL mutant triplicate samples. (C) Volcano plot of transcriptional changes observed in a ΔlipL mutant compared to WT S. aureus during late-exponential phase growth. Each dot represents one gene. Vertical dotted lines represent log₂ fold change = -2 and 2 cut-offs. Horizontal dotted line represents adj. p value = 0.05. (D) Analysis of upregulated and downregulated pathways in a ΔlipL mutant compared to WT. Black bars represent log(p-value); vertical dotted line is p-value < 0.05 threshold of significance as determined by NIH DAVID Bioinformatics pipeline. Blue bars represent the number of downregulated genes associated with the respective pathway, and pink bars represent the number of upregulated genes associated with the respective pathway. (E) Fold change of specific genes in the ΔlipL mutant compared to WT. (F) qRT-PCR analysis of RNA extracted from WT, Δpta, ΔlipL, Δpta + pta, and ΔlipL + lipL strains after 3 hours of growth in TSB. ****, p < 0.0001 by one-way ANOVA with Tukey’s post hoc test. qRT-PCR experiments were conducted with two independent biological replicates in technical triplicate. Boxplots indicate the median and quartiles. Whiskers indicate the range.

https://doi.org/10.1371/journal.ppat.1014183.g006

The deletion of both pta and cidC significantly reduces acetate production by S. aureus

Prior studies showed that inactivation of the Pta-AckA pathway led to redirection of carbon into the TCA cycle [28,53]. Furthermore, loss of LipL or Pta leads to transcriptional changes that are suggestive of premature entry into the TCA cycle, increased oxidative metabolism, and increased acetate production by CidC (Fig 6F) [28]. Increased cidC levels in both Δpta and ΔlipL strains signify demand for acetate production in S. aureus regardless of whether the block occurs before or after PDH. To confirm prior studies and test if Pta and CidC are the primary sources of acetate, we generated an in-frame deletion of cidC in both WT and a Δpta mutant of S. aureus. The ΔcidC mutant had no effect on growth in TSB, whereas a Δpta ΔcidC double mutant had a marginal delay in transition to stationary phase growth (Fig 7A), like the Δpta single mutant (Fig 2A). The Δpta ΔcidC and Δpta ΔcidC + cidC strains, but not the ΔcidC or the Δpta ΔcidC + pta strains had evidence of increased lipoylation on E2 subunits consistent with more lipoylation occurring in the absence of Pta, but not CidC (Fig 7B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. The deletion of both pta and cidC significantly reduces acetate production by S. aureus.

(A) Growth of WT, ΔcidC, Δpta ΔcidC, ΔcidC + cidC, Δpta ΔcidC + pta, and Δpta ΔcidC + cidC strains in TSB and TSB + BCCAs [10 mM isobutyric acid (IB), 9 mM 2-methylbutyric acid (2MB), 9 mM isovaleric acid (IV) + 10 mM sodium acetate (NaAc)] containing 14 mM glucose. (B) α-lipoic acid immunoblots of whole cell lysates from the indicated strains after 9-hour subculture growth in TSB medium. The presented blot and Coomassie-stained gel are representative of three independent experiments. (C) Quantification of glucose, acetate, lactate, and formate from WT, ΔcidC, and ΔcidC + cidC culture supernatants over time in TSB with 14 mM glucose. (D) Quantification of glucose, acetate, lactate, and formate from WT, Δpta ΔcidC, Δpta ΔcidC + pta, and Δpta ΔcidC + cidC culture supernatants over time in TSB with 14 mM glucose. Each metabolite quantification assay is representative of three independent experiments, with each timepoint measured in technical triplicate. Error bars indicate standard deviation from the mean.

https://doi.org/10.1371/journal.ppat.1014183.g007

When measuring glucose, acetate, lactate, and formate levels, we noted the ΔcidC mutant consumed all glucose in the media by 6 hours with no difference compared to the WT (Fig 7C). Peak acetate production was ~ 30 mM for all strains (Fig 7C). Furthermore, there was negligible production of lactate and formate (Fig 7C). The Δpta ΔcidC double mutant also consumed all glucose by 6 hours and had no delay in the glucose consumption compared to the WT (Fig 7D). Consistent with previous reports, there was negligible production of acetate from the Δpta ΔcidC double mutant (~5 mM, a ~ 83% reduction compared to WT). (Fig 7D) [28]. Peak acetate production from the Δpta ΔcidC + cidC was ~ 13 mM (Fig 7D), whereas acetate production from Δpta ΔcidC + pta was the same as WT (Fig 7D). There was no lactate or formate produced, consistent with our earlier observations that compensatory lactate fermentation only occurs when PDH is inactivated (Fig 7D). These data corroborate prior work and indicate that cidC is dispensable for S. aureus growth in broth yet is activated to compensate for the loss of Pta to promote acetate production (compare Figs 2C and 7D) [28,36].

Acetate production is required for S. aureus infection

In prior studies, CidC was implicated in the pathogenesis of biofilm-associated infection in the heart, suggesting CidC has a role in virulence [36]. Furthermore, the increase of cidC transcripts in the Δpta and ΔlipL mutants (Fig 6F) led us to hypothesize that acetate generation via CidC could be an important compensatory strategy to promote S. aureus survival in vivo, especially if acetate generation in and of itself were important for infection. To begin to test this possibility, we determined if cidC was required for S. aureus bloodstream and skin infection using planktonic cultures. We infected mice intravenously or intradermally with WT, ΔcidC, and ΔcidC + cidC strains. At 96 hours post intravenous infection, we isolated and homogenized kidneys from ΔcidC mutant-infected animals and found no difference in CFUs compared to WT (Fig 8A). At 72 hours post intradermal infection, there was also no difference in CFUs between strains (Fig 8A). We then infected mice intravenously or intradermally with WT, Δpta ΔcidC, Δpta ΔcidC + pta, and Δpta ΔcidC + cidC strains. After systemic infection, the kidneys from Δpta ΔcidC double mutant-infected animals had ∼10000-fold fewer CFUs compared to WT (Fig 8B). This result was in stark contrast to a Δpta mutant, which had a ~ 100-fold reduction in CFUs (Fig 5A). Complementation of the double mutant with pta restored kidney CFUs to nearly WT levels (Fig 8B), whereas complementation of the double mutant with cidC partially restored virulence to levels previously observed for a Δpta mutant (~100-fold fewer CFUs compared to WT) (Fig 8B). At 72 hours post intradermal infection, we found ∼100-fold fewer CFUs were recovered from the abscesses of Δpta ΔcidC double mutant compared to WT (Fig 8B). Like the kidneys, complementation with pta restored CFU levels to WT levels (Fig 8B), whereas complementation with cidC in the skin was not evident (Fig 8B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Acetate production is required for S. aureus infection.

(A) Upper panel - Bacterial burden in the skin of mice 72 hours after intradermal infection with 1.0 x 10⁷ CFU of WT and ΔcidC strains. Animal numbers displayed are as follows: WT, N = 18; ΔcidC, N = 18. Lower panel - Bacterial burden in the kidneys of mice 96 hours after bloodstream infection with 1.0 x 10⁷ CFU of WT, ΔcidC strains. Animal numbers displayed are as follows: WT, N = 34; ΔcidC, N = 33. NS, not significant. Statistics were calculated by Mann-Whitney test. (B) Upper panel - Bacterial burden in the skin of mice 72 hours after intradermal infection with 1.0 x 10⁷ CFU of WT, Δpta ΔcidC, Δpta ΔcidC + pta, and Δpta ΔcidC + cidC strains. Animal numbers displayed are as follows: WT, N = 29; Δpta ΔcidC, N = 29; Δpta ΔcidC + pta, N = 28; Δpta ΔcidC + cidC, N = 28. Lower panel - Bacterial burden in the kidneys of mice 96 hours after bloodstream infection with 1.0 x 10⁷ CFU of WT, Δpta ΔcidC, Δpta ΔcidC + pta, and Δpta ΔcidC + cidC strains. Animal numbers displayed are as follows: WT, N = 16; Δpta ΔcidC, N = 16; Δpta ΔcidC + pta, N = 16; Δpta ΔcidC + cidC, N = 16. ***, p < 0.001; ****, p < 0.0001 by Kruskal-Wallis test with Dunn’s post hoc analysis. (C) Upper panel - Bacterial burden in the skin of mice 72 hours after intradermal infection with 1.0 x 10⁷ CFU of WT, Δagr::tet, Δpta ΔcidC, and Δpta ΔcidC Δagr::tet strains. Animal numbers displayed are as follows: WT, N = 16; Δagr::tet, N = 16; Δpta ΔcidC, N = 16; Δpta ΔcidC Δagr::tet, N = 16. Lower panel - Bacterial burden in the kidneys of mice 96 hours after bloodstream infection with 1.0 x 10⁷ CFU of WT, Δagr::tet, Δpta ΔcidC, and Δpta ΔcidC Δagr::tet strains. Animal numbers displayed are as follows: WT, N = 12; Δagr::tet, N = 12; Δpta ΔcidC, N = 12; Δpta ΔcidC Δagr::tet, N = 12. *, p < 0.05; **, p < 0.01; ****, p < 0.0001 by Kruskal-Wallis test with Dunn’s post hoc analysis. log₁₀CFU per organ or abscess is displayed for each infected mouse along with the median as a measure of central tendency. All graphs represent combined data from at least two independent experiments.

https://doi.org/10.1371/journal.ppat.1014183.g008

Among the pathways significantly upregulated in RNAseq analysis of a ΔlipL mutant (Fig 6E) were genes involved in virulence. Toxin genes, including hla, psmβ1, psmβ2, and lukS-PV, were amongst the most significantly upregulated (Fig 6F). These toxin genes are all regulated by the Accessory Gene Regulatory (Agr) two-component quorum sensing system [54,55]. Despite the increase in Agr-dependent virulence gene expression, a ΔlipL mutant remained attenuated in vivo (Fig 5) arguing that virulence factor expression is not contributing to in vivo phenotypes in this strain. Nevertheless, this outcome prompted us test if in vivo phenotypes linked to acetate production were potentially due to changes in Agr activity. To test this possibility, we infected mice intravenously or intradermally with WT, Δagr::tet, Δpta ΔcidC, and Δpta ΔcidC Δagr::tet strains. At 96 hours post-intravenous infection, kidneys from Δagr::tet mutant-infected animals had similar CFUs to WT (Fig 8C), consistent with previous reports [56,57]. Conversely, kidneys from Δpta ΔcidC and Δpta ΔcidC Δagr::tet mutant-infected animals had identical reduction in CFUs compared to WT (∼10000-fold) (Fig 8C). At 72 hours post intradermal infection, abscesses from mice infected with Δagr::tet mutant had ∼10-fold fewer CFUs compared to WT, which is consistent with previous literature [58] (Fig 8C). Abscesses from mice infected with Δpta ΔcidC double mutant had ∼100-fold fewer CFUs compared to WT (Fig 8C). Mice infected with Δpta ΔcidC Δagr::tet triple mutant had ∼1000-fold fewer CFUs, demonstrating an additive effect (Fig 8C). Taken together, these data suggest that acetate production mediated by Pta and CidC is required for infection and is Agr independent.

Ethics statement

All animal experiments were conducted in ABSL2 facilities under Institutional Animal Care and Use Committee (IACUC)-approved protocols from the University of Illinois Chicago (Protocol #25–042) and Loyola University Chicago Health Sciences Division (Protocol #2020025), in accordance with United States Department of Agriculture (USDA), Public Health Service (PHS), and Office of Laboratory Animal Welfare (OLAW) guidelines. The University of Illinois Chicago holds a current PHS Animal Assurance (#A3460-01, through 04/30/2027), is Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) accredited (#000186, 04/01/2024), and USDA licensed (#33-R-0018). Loyola University Chicago holds a current PHS Animal Assurance (#A3460-01, through 02/28/2026), is AAALAC-accredited (#000180, 11/18/2022), and USDA-licensed (#33-R-0024). All procedures complied with institutional ethical standards.

Phylogenetic tree construction and synteny analysis

The pta coding sequence was used in independent searches of the Seed analysis viewer (https://pubseed.theseed.org/FIG/seedviewer.cgi?page=Minimal) [102,103]. The pta gene in Staphylococcus aureus subsp. aureus USA300 used for search was taxon ID 367830; feature ID fig|367830.3.peg.650) and is annotated as phosphate acetyltransferase (EC 2.3.1.8). All genomes were compared by a standard search of the Seed database. The genomes of M. tuberculosis and selected Gram-positive and Gram-negative bacteria representing the most frequent genetic organizations of pta, lipL, and ackA were then chosen for subsequent analysis. Taxonomic assignments for these genomes were confirmed in NCBI (TaxIDs: 242231, 604162, 99287, 575584, 1051003, 272620, 83333, 196600, 187410, 451515, 224308, 1027396, 160490, 1773, 226185, 416870, 373153, 191218, 1385523, 1242971, 1155131, 904338, 1189311, 435837, 1134914, 342451). Phylogenetic trees were constructed using the selected genomes in PhyloT_v2 (https://phylot.biobyte.de) and visualized using the interactive Tree of Life tool, iTOL_v7 (https://itol.embl.de) [104].

Bacterial strains and growth conditions

The strains used in this study are described in S1 Table. Escherichia coli strains were grown in Lysogeny Broth (LB) (BD Biosciences), and S. aureus strains were grown in Tryptic Soy Broth (TSB) (BD Biosciences). For experiments with no or low glucose (3.5 mM), S. aureus strains were growin in TSB without dextrose (BD Biosciences) with or without D-glucose supplementation (Amresco). Unless stipulated, strains were grown at 37°C, shaking at 200 RPM. For strains harboring a ΔlipL mutation, branched-chain carboxylic acids (BCCAs) (10 mM isobutyric acid [Sigma-Aldrich], 9 mM 2-methylbutyric acid [Alfa Aesar], 9 mM isovaleric acid [Sigma-Aldrich]) and 10 mM sodium acetate (NaAc; Sigma-Aldrich) were added to the media to promote growth [23,61]. When needed, media were supplemented with the following agents of selection: ampicillin (100 µg/ml), chloramphenicol (10 µg/ml), kanamycin (50 µg/ml), cadmium chloride (0.1 mM), anhydrous tetracycline (1 μg/ml) (AnTet), and sodium citrate (10 mM).

Generation of in-frame deletion mutants

Oligonucleotide pairs (S2 Table) were designed to amplify fragments of ~500–1000 nucleotides upstream and downstream of the targeted genes for mutagenesis using S. aureus LAC genomic DNA as a template. Purified amplicons were then used as templates in a splicing by overlap extension (SOE) PCR to yield the final 1,000–2,000 bp amplicon containing a fusion of the upstream and downstream gene fragments. Each amplicon was digested with KpnI and SacI restriction endonucleases and ligated into the multiple cloning site (MCS) of pIMAY using T4 DNA ligase [105]. Ligation reactions were transformed into E. coli IM08B [46], and transformants containing the ligated amplicon were verified by PCR. Plasmids were purified from IM08B and subsequently transformed into S. aureus competent cells by electroporation (1800 V, 10 μF, 600 Ω, 2 mm cuvette). Mutagenesis was performed as previously described [47]. Briefly, S. aureus cells containing pIMAY plasmids were grown at 30°C to allow for plasmid replication. Plasmid recombination into the chromosome was induced at 37°C with selection on chloramphenicol. To facilitate plasmid excision, cells were returned to the permissive temperature for plasmid replication (30°C) and passaged without antibiotic to promote plasmid loss. Bacteria were then plated on TSA plates supplemented with AnTet and incubated at 30°C to counter-select for cells that still contain the pIMAY plasmid. Chloramphenicol sensitive and AnTet-resistant colonies were screened for the desired mutation by PCR.

Generation of complement strains and phage transduction

Complementation strains were generated using the pJC1111 plasmid [48]. Genes for complementation were amplified from S. aureus genomic DNA using the primer pairs listed in S2 Table, while the constitutive P_HELP promoter was amplified from the pIMAY plasmid as template [47]. Purified amplicons were fused by SOE PCR and subsequently cloned into the MCS of pJC1111 using Sac1 and Sal1 restriction endonucleases for digestion and T4 DNA ligase for ligation. The recombinant plasmids were propagated in E. coli DH5α and transformed into SaPI-1 integrase-expressing S. aureus (RN9011) to allow single-copy integration in the chromosome at the SaPI-1 site [48]. Bacteriophage Φ11 was used to package the integrated complementation plasmids from RN9011, and the phage was used for transduction of genetic material into the recipient strain of interest, as previously described [106]. Briefly, a single colony of the RN9011 donor strain was grown overnight at 37°C with shaking. The overnight culture was diluted to an OD₆₀₀ of 0.2 in TSB. 5 mL of diluted culture was added to 5 mL of TMG buffer (10 mM Tris pH 7.5, 5 mM MgCl₂, 0.01% gelatin (v/v), 5 mM CaCl₂) in a 15 mL conical tube containing 100 μL of Φ11 phage and incubated at room temperature overnight until lysis was complete. The following day, 100 μL of the filter sterilized phage lysate was added to an overnight culture of the recipient strain and incubated at 37°C with shaking for 20 minutes. The infected recipient was washed three times with 1 mL of 40 mM sodium citrate and then plated on TSA supplemented with 0.1 mM cadmium chloride and 40 mM sodium citrate to select for transductants. All transductants were confirmed by PCR.

Generation of Δagr Δpta ΔcidC strain

A marked Δagr::tet mutation was transduced into a Δpta ΔcidC mutant strain, as described above, to generate the Δagr::tet Δpta ΔcidC mutant strain. The Δagr::tet mutation spans the agr operon [58].

Generation of 6x-Histidine tagged Pta and LipL expression plasmids

The pta and lipL genes were amplified from S. aureus LAC genomic DNA using the primer pairs in S2 Table. The resulting amplicons were cloned into the pET21a (Novagen) expression vector using Nde1 and Sal1 restriction endonucleases for digestion and T4 DNA ligase for ligation. The resulting plasmids were transformed into chemically competent E. coli DH5α. Plasmids were purified from DH5α and transformed into chemically competent E. coli lysY/I^q (NEB). All recombinant expression plasmids were verified by whole plasmid sequencing (Plasmidsaurus).

Purification of 6x-Histidine-tagged proteins

To purify Pta-6xHis and LipL-6xHis, a single colony of E. coli lysY/I^q containing pET21a-pta-6xHis or pET21a-lipL-6xHis was used to start an overnight culture in 20 mL of LB at 37°C with shaking at 200 RPM. The bacteria were diluted (1:100) in 1 L of fresh LB and incubated at 37°C with shaking at 200 RPM until reaching an OD₆₀₀ of ~0.3. Protein expression was induced by supplementing the culture with IPTG (0.1 mM) and incubating at 37°C for 3 hours with shaking at 200 RPM. Following induction, bacteria were collected by centrifugation at 10000 RPM for 20 min, and pellets were stored at -80°C until use. Frozen bacterial pellets were thawed and resuspended in 60 mL lysis buffer (25 mM imidazole, 50 mM Tris-HCl, 300 mM NaCl, 1 mM phenylmethylsulfonyl fluoride (PMSF), pH 8). The bacterial suspensions were sonicated on an ice bath for 10 minutes with intervals of 10 s ON and 50 s OFF at an amplitude of 340 W using a Branson 550 Sonicator. Cell debris was removed by centrifugation at 15000 RPM for 30 min, and the lysate was collected and passed through a 0.22 µm syringe filter. The clarified lysates were then mixed with 1 ml of nickel-NTA resin (Qiagen), preequilibrated with lysis buffer, and incubated on a rotisserie overnight a 4°C. After incubation, the mixture was poured into a glass gravity flow chromatography column (Bio-Rad), and the resin was washed with 50 mL of lysis buffer, followed by elution of the bound protein with 6 ml elution buffer (500 mM imidazole, 50 mM Tris-HCl, 300 mM NaCl, pH 8). Purified proteins were dialyzed at 4°C in 10-kDa molecular weight cutoff (MWCO) dialysis cassettes (Thermo Scientific). Dialysis conditions were as follows: 1 L buffer 1 (100 mM imidazole, 50 mM Tris-HCl, 300 mM NaCl, pH 8) for 3 hours, 1 L buffer 2 (25 mM imidazole, 50 mM Tris-HCl, 300 mM NaCl, pH 8) for 3 hours, 1 L buffer 3 (50 mM Tris-HCl, 300 mM NaCl, pH 8) overnight, and 1 L buffer 4 (50 mM Tris-HCl, 300 mM NaCl, pH 8) for 3 hours. Protein concentration was measured using a bicinchoninic acid (BCA) kit (Thermo Scientific). Protein purity was confirmed by resolving the purified proteins using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by staining gels with Coomassie blue. Purified proteins were stored in aliquots at -80°C until use.

RNA purification

S. aureus strains were struck onto fresh TSA plates and incubated overnight at 37°C. The next day, a single colony from the plate was inoculated into 3 mL of TSB in a 15 mL conical tube and incubated overnight at 37°C with shaking at 200 RPM. The following day, 250 μL of overnight culture was added to 25 mL of TSB + BCCAs in a 250 mL flask. Flasks were placed at 37°C with shaking at 200 RPM. For RNA sequencing, cultures were grown to late exponential phase (OD₆₀₀ = 0.6) and for qPCR, strains were grown for 6 hours. 10 mL of each culture was removed and subsequently normalized to the OD₆₀₀ of the strain with the lowest turbidity to ensure equivalent biomass. Cultures were centrifuged at 11000 RPM and washed twice with 15 mL PBS. Cell pellets were fixed in 15 mL of a 1:1 ethanol: acetone solution and stored at -80°C until RNA extraction. RNA was subsequently purified using a RNeasy Micro kit (Qiagen) with an additional DNase digestion using Turbo DNase (Thermo). Briefly, bacterial pellets were thawed on ice and centrifuged for 15 minutes at 4000 RPM at 4°C to remove the ethanol:acetone solution. The pellet was air-dried and was washed twice with 500 μL of 1x PBS, followed by suspension of the washed pellet in 500 μL RLT buffer containing 10 μL/mL of β-mercaptoethanol. Suspensions were transferred to RNase-free Lysing Matrix B tubes (MP Biomedicals). Tubes were placed in a Fast Prep-24 5G (MP Biomedicals) bead disruption system, followed by two sequential cell lysis steps at 5.0 speed for 20 s and 4.5 speed for 20 s. Samples were kept on ice for 5 minutes between the two lysis steps. After lysis, samples were centrifuged at 10000 RPM for 20 min, and the supernatant was removed. 500 μL of RLT buffer was added to supernatants and the samples were vortexed, followed by addition of 500 μL of 100% ethanol to precipitate the RNA, followed by addition to a Qiagen RNeasy spin column. All remaining steps were performed according to the manufacturer’s protocol. The RNA was eluted in 40 μL of DEPC-treated water. A second DNase digestion was conducted using 1 U Turbo DNase in solution (Invitrogen). The RNA was precipitated with 100% ethanol, added to an RNeasy spin column, washed 3 times in 450 μL wash buffer, and eluted in 50 μL of DEPC-treated water. The purified RNA was quantified via Nanodrop and RNA quality was assessed either by resolving the RNA on a 1.2% agarose/formaldehyde gel or by tape station analysis. RNA was stored at -80°C until cDNA synthesis or RNA sequencing.

qRT-PCR and RNA sequencing

For qRT-PCR, cDNA was first synthesized from 5 ng of purified RNA using the GoScript reverse transcription system following the manufacturer’s protocol (Promega). Gene expression was assessed using SYBR Green (Bio-Rad) with the primer pairs listed in S2 Table on a QuantStudio 5 real-time PCR system. Fold change relative to housekeeping genes (gyrB or sigA) was calculated using the threshold cycle (2 − ΔΔCT) method. RNA sequencing was completed by SeqCoast Genomics LLC. Briefly, samples were prepared for RNA sequencing using an Illumina Stranded Total RNA Prep Ligation with Ribo-Zero Plus Microbiome and IDT For Illumina Unique Dual Indexes. Sequencing was performed on the Illumina NextSeq2000 platform using a 300-cycle flow cell kit to produce 2 x 150 bp paired reads. Read demultiplexing, read trimming, and run analytics were performed using DRAGEN v4.2.7, an on-board analysis software on the NextSeq2000. The quality of the sequencing data was first assessed using FastQC v0.12.0. Next, all reads were trimmed and quality filtered using BBDuk v39.33 employing Illumina’s adapters and the program’s default settings. To remove any potential contaminating reads, the data were subjected to another round of quality processing. This was performed by splitting the reads and removing the contaminating reads, using BBSplit v39.33 and the reference genome of S. aureus subsp. aureus LAC. Due to the lower support of S. aureus subsp. aureus LAC reference genome (NC_007793) in the downstream analysis tools and databases, the processed reads were instead aligned to the reference genome of S. aureus subsp. aureus TCH1516 using Bowtie 2 v2.5.4. SAMtools v1.22.1 were then employed to convert the output files of Bowtie 2, followed by the counting of the mapped reads to the genomic features using featureCounts (from the Subread package, v2.1.1). Finally, differential expression analysis was conducted employing DESeq2 (version 1.40.2). Specifically, count matrices and sample metadata were imported with the DESeqDataSetFromMatrix function and normalized using DESeq2’s median-of-ratios method. Afterwards, gene-wise dispersions were estimated, and a negative-binomial generalized linear model was fit to the counts using a design formula that included the condition factor. Statistical significance of coefficients was assessed using Wald test, and resulting p-values were adjusted for multiple testing by the Benjamini–Hochberg procedure to control the false discovery rate. Log2-fold changes were reported after effect-size shrinkage using the apeglm method, which provides a more stable and interpretable log-fold change estimate for ranking and visualization. Genes with adjusted p-value ≤ 0.05 were considered differentially expressed and consequently further analyzed. The data obtained from the differential expression analysis were further processed employing Pathview v1.48.0 for mapping and rendering the differentially expressed genes on relevant pathway graphs, clusterProfiler v4.16.0 for gene set enrichment analysis, and the individual genes manually parsed using Microsoft Excel (v2505). Normalized and mapped reads were visualized in Integrative Genomics Viewer (IGV).

Pathway analysis

Functional pathway enrichment was carried out using approaches similar to those described in a prior study [107]. In summary, the set of differentially expressed genes was assigned to biological pathways using the free Database for Annotation, Visualization and Integrated Discovery (DAVID) provided by NIAID/NIH [108,109]. Gene identifiers were converted to NCTC8325 locus tags in the SAOUHSC ortholog format with the help of AureoWiki, an online resource for S. aureus. The DAVID analysis returned unbiased functional annotations, suggested pathway groupings, and the number of genes associated with each pathway.

Growth curves

S. aureus strains were struck onto fresh TSA plates and incubated overnight at 37°C. The next day, a single colony from the plate was inoculated into 3 mL of TSB in a 15 mL conical tube and incubated overnight at 37°C with shaking at 200 RPM. Strains containing a ΔlipL mutation were grown with complete BCCAs overnight. The next day, strains were washed three times with 5 mL 1 x PBS and subcultured 1:100 (2 μL into 198 μL fresh TSB) in a flat-bottom 96-well plate (Costar) in technical triplicate. Strains were grown in TSB + BCCAs without the addition of sodium acetate. The plate was incubated in a Tecan Spark plate reader at 37°C with orbital shaking at 3 mm with a frequency of 180 RPM. Bacterial growth was monitored by measuring OD₆₀₀ every hour for 24 hours.

Metabolite quantification

S. aureus strains were struck onto fresh TSA plates and incubated overnight at 37°C. The next day, a single colony from the plate was inoculated into 3 mL of TSB in a 15 mL conical tube and incubated overnight at 37°C with shaking at 200 RPM. The following day, 250 μL of overnight culture was added to 25 mL of TSB + BCCAs without sodium acetate in a 250 mL flask. Flasks were placed at 37°C with shaking at 200 RPM. For glucose, lactate, and formate measurements, 120 μL of culture was removed every hour for six hours. For acetate measurements, 60 μL of culture was collected every three hours for 12 hours, with a final collection at 24 hours. Samples were collected in 1.5 mL Eppendorf tubes at each time point and were centrifuged at 13000 RPM for 5 minutes. The cellfree supernatant was stored at -20°C until assaying metabolite levels. Acetate, glucose, formate, and D- and L-lactate concentrations were determined using kits purchased from R-Biopharm, according to the manufacturer’s protocol with minor changes. Briefly, for all metabolites, 10 μL of each sample was incubated with 200 μL of reagent 1 for 3 minutes at room temperature in a flat-bottom 96-well plate (Costar) in technical triplicate. Following the primary incubation, 50 μL of reagent 2 was added to the reaction mixture and incubated for 15 minutes. A Tecan Spark plate reader was used to read the absorbance at 340 nm for each metabolite. Final calculations of metabolite concentrations were determined using a standard curve.

Intracellular pyruvate quantification

S. aureus strains were struck onto fresh TSA plates and incubated overnight at 37°C. The next day, a single colony from the plate was inoculated into 3 mL of TSB in a 15 mL conical tube and incubated overnight at 37°C with shaking at 200 RPM. The following day, 250 μL of overnight culture was added to 25 mL of TSB + BCCAs without sodium acetate in a 250 mL flask and placed at 37°C with shaking at 200 RPM. At 3 hours, cultures were normalized to the lowest optical density and 1 mL of normalized culture was added to 1.5 mL Eppendorf tubes, centrifuged at 10,000 RPM for two minutes, and washed twice with 1 mL 1x PBS. The bacterial suspension was mixed with pyruvate assay buffer from a commercial Pyruvate Assay Kit (Abcam) and transferred to screw cap microcentrifuge tubes (Fisher Scientific), which were preloaded with 250 µL of 0.1 mm glass beads (Electron Microscopy Sciences). Tubes were placed in a Fast Prep-24 5G (MP Biomedicals) bead disruption system, followed by two sequential cell lysis steps at 5.0 speed for 20 s and 4.5 speed for 20 s. Samples were kept on ice for 5 minutes between the two lysis steps. After lysis, samples were centrifuged at 10000 RPM for 20 min, and 150 µL of the supernatant was collected and assayed according to the manufacturer’s protocol.

Determination of protein lipoylation and densitometry

S. aureus strains were struck onto fresh TSA plates and incubated overnight at 37°C. The next day, a single colony from the plate was inoculated into 3 mL of TSB in a 15 mL conical tube and incubated overnight at 37°C with shaking at 200 RPM. Overnight cultures were diluted 1:100 in 10 mL fresh TSB in 50 mL conical tubes and incubated for 9 hours at 37°C with shaking at 200 RPM. Bacteria were centrifuged at 8000 RPM for 10 minutes, supernatants were removed, and cell pellets were resuspended in 1 mL 1x PBS. Bacterial suspensions were transferred to screw cap microcentrifuge tubes (Fisher Scientific), which were preloaded with 250 µL of 0.1 mm glass beads (Electron Microscopy Sciences). Tubes were placed in a Fast Prep-24 5G (MP Biomedicals) bead disruption system, followed by two sequential cell lysis steps at 5.0 speed for 20 s and at 4.5 speed for 20 s. Samples were kept on ice for 5 minutes between the two lysis steps. After lysis, samples were centrifuged at 11,000 RPM for 20 minutes, and 150 µL of the supernatant was collected in 1.5 mL centrifuge tubes containing 50 µL of 4X SDS sample buffer (0.2 M Tri-HCl, pH 6.8, 8% SDS, 5.5 M glycerol, 0.02 M EDTA, 0.6 M β-mercaptoethanol, and 6 mM Bromophenol blue). Samples were boiled for 10 minutes and stored at -20°C. Thawed cell lysates were loaded onto 15% polyacrylamide gels after normalizing to the original OD₆₀₀ of the culture, followed by SDS-PAGE at 100 V for 2 hours. Resolved proteins were either stained with Coomassie blue or were transferred to 0.2 μm Immobilon polyvinylidene difluoride (PVDF) membranes (Millipore Sigma) at 20 V for 90 min. After transfer, membranes were incubated for 1 hour in Tris-buffered saline + 0.1% TWEEN 20 (Amresco) (TBST) supplemented with 5% bovine serum albumin (BSA) (GoldBio) and human IgG (Sigma) (1:2,000). Rabbit polyclonal α-lipoic acid antibody (Calbiochem) was added to the membranes at a 1:7,500 dilution followed by incubation overnight at 4°C. The following day, the membrane was washed three times for 10 minutes each in ∼20 mL of TBST. Alkaline phosphatase (AP)-conjugated goat anti-rabbit IgG (H + L) (Invitrogen) was then added at a 1:5000 dilution in 5% BSA in TBST for 1 hour followed by three 10 minutes washes in ∼20 mL of TBST. Membranes were developed with 5-bromo-4-chloro-3-indoyl-phosphate/nitro blue tetrazolium color development substrate (GoldBio). Densitometry analysis was performed using the ImageJ software on four independent experiments to determine the lipoylation of each band relative to the control band set at a value of 1.

Bacterial Adenylate Cyclase Two-Hybrid (BACTH) system

S. aureus pta and lipL genes were amplified (S2 Table) and cloned into plasmids pKT25 and pUT18c to generate translational fusions with the T25 or T18 domains of the Bordetella pertussis adenylate cyclase [51]. To clone the lipL gene into BACTH vectors, Sma1 and Pst1 restriction endonucleases were used for digestion and T4 DNA ligase for ligation. To clone the pta gene into BACTH vectors, Sma1 and Xba1 restriction endonucleases were used for digestion and T4 DNA ligase for ligation. The resulting plasmids were transformed into E. coli DH5α and plated on LB agar containing kanamycin (50 μg/mL) or ampicillin (100 μg/mL). Transformants containing the ligated amplicon were verified by PCR. The recombinant plasmids were purified from DH5α E. coli and then co-transformed into E. coli strain BTH101 and plated on LB agar containing both kanamycin (50 μg/mL) and ampicillin (100 μg/mL). To assay for protein-protein interactions, single colonies were inoculated into 3 mL LB in 15 mL conical tubes containing both antibiotics overnight. The following day, strains were diluted 1:100 and grown for 6 hours at 30°C with shaking at 200 RPM. Upon reaching late-exponential phase (~6 hours), 10 μL culture was spotted on LB agar containing IPTG (0.5 mM), ampicillin (100 μg/mL), kanamycin (50 μg/mL), and X-Gal (100 μg/mL). Plates were incubated for 24 hours at 30°C to allow for bacterial growth. An interaction was defined by the presence of blue color relative to positive and negative controls. The negative control, empty BACTH plasmids, yield colorless colonies. The positive control was BTH101 co-transformed with pKT25-zip and pUT18C-zip, plasmids expressing leucine zipper - adenylate cyclase fusions that dimerize to yield blue colonies.

β-galactosidase activity was quantified based on previous methods [110]. Briefly, single colonies were inoculated in 3 mL LB in 15 mL conical tubes containing both antibiotics overnight. The next day, the cultures were diluted 1:100 in 3 mL LB and grown for 6 hours with IPTG (0.5 mM), ampicillin (100 μg/mL) and kanamycin (50 μg/mL) with shaking at 200 RPM at 30°C until reaching an OD₆₀₀ ~ 1.0. Strains were OD normalized and 100 μL of each sample was added to a 96-well plate (Costar) in triplicate. The cultures were centrifuged at 4000 RPM, and the bacterial pellets were resuspended in 80 μL of Z-buffer (60 mM Na₂HPO₄·7H₂O, 40 mM NaH₂PO₄·H₂O, 10 mM KCl, 1 mM MgSO₄, 50 mM β-mercaptoethanol, pH 7). The resuspended cells were then permeabilized with 10 μL of 0.1% SDS and 10 μL of chloroform. Plates were centrifuged at 4000 RPM, the supernatant was transferred to a new 96-well plate, mixed with 20 μL ONPG (4 mg/mL), and the reaction was allowed to proceed for 15–30 min. Reactions were stopped by adding 30 μL 1 M Na₂CO₃ and absorbance was measured at 420 nm. The negative control, BTH101 carrying the empty pKT25 and pUT18c plasmids, and the positive control, BTH101 containing pKT25-zip and pUT18c-zip plasmids, are described above. β-galactosidase activity is reflected relative to the negative and positive controls, with the negative control assigned an activity value of 0 in arbitrary units and positive control assigned an activity of 1000 arbitrary units.

Microscale Thermophoresis (MST)

Experiments were performed as described previously with minor modifications [111]. Briefly, pure recombinant Pta was labeled with a NanoTemper 2nd Generation Red N-hydroxysuccinimide (NHS) Dye (Nanotemper Technologies). For protein labeling, Pta was first buffer exchanged into buffer M (20 mM MES, 100 mM NaCl, 10% glycerol, pH 7.5) to remove incompatible buffer components. 10 µM Pta was mixed with 3X excess of the Red NHS dye (300 µM) and incubated in the dark for 30 minutes for the dye to react with the protein. The protein-dye mixture was passed through a gel filtration column (Column B, Nanotemper Technologies) to remove unreacted and excess dye. For a 1:1 protein-dye conjugate, the degree of label (DOL) of the protein was determined using the formula: A650/195,000/M/cm × concentration of labeled protein. A650; absorbance at 650 nm, and 195,000/M/cm is the molar absorbance of the Red NHS dye. A 1:1 protein dye ratio occurred at a DOL value between 0.6 and 1. After labeling, protein samples were aliquoted into 10µL volumes, flash frozen using liquid nitrogen, and stored at -80^oC until used. For microscale thermophoresis experiments, the unlabeled protein partner was serially diluted in low-binding tubes with buffer M supplemented with 1 mM BME and 0.05% Tween 20, pH 7.5, and titrated against 20 nM of the labeled proteins. The protein mixtures were loaded into standard Monolith NT.115 capillary tubes, and thermophoresis was determined using the following parameters: 20% excitation power and high MST Power at 25°C.

Thermophoresis results were analyzed using the PALMIST [112] and GUSSI [112] analysis pipeline. Briefly, data from the MST software (Mo. Control v1.6.1) were imported into the PALMIST software and a preset T-jump (TJ) was applied to the data using a 1:1 binding model with a 95% confidence interval. Datasets were examined for kinetic effects to ensure that thermophoresis achieved equilibrium. After data analysis, GUSSI was used for figure rendering. The top panel in Fig 4B indicates the normalized fluorescence data points, the middle panel represents the binding curve, and the lower panel represents the residuals (a plot between the fitted line and the data). Error bars indicate the standard deviation of four independent replicates.

Murine infection models

S. aureus strains were struck onto fresh TSA plates and incubated overnight at 37°C with shaking at 200 RPM. The next day, a single colony from the plate was inoculated into 3 mL of TSB in a 15 mL conical tube and incubated overnight at 37°C with shaking at 200 RPM. The overnight cultures were diluted 1:100 in 3 mL fresh TSB and grown for 3 h at 37°C with shaking at 200 RPM. Bacteria were pelleted by centrifugation at 3900 RPM at 4°C for 5 minutes. The cells were washed three times with 5 mL of 1X PBS. Bacterial cell suspensions were normalized to an OD₆₀₀ of ~0.32 (~1 × 10⁸ CFU/mL), and serial dilutions were plated onto TSA plates to enumerate CFU and ensure accuracy of inocula. Six-week-old female Swiss Webster mice (Envigo) were anesthetized with ketamine/xylazine (100/10 mg/kg) via intraperitoneal injection. For systemic infections, 100 μL of OD-normalized S. aureus in PBS (~1.0 × 10⁷ CFU) was injected directly into the bloodstream via the retro-orbital sinus. Mice were monitored daily. At 96 h postinfection, mice were euthanized, and the kidneys were recovered, homogenized, serially diluted, dilutions were plated onto tryptic soy agar followed by incubation at 37°C overnight to enumerate CFU. For skin infections, bacterial suspensions (~1.0 × 10⁷ CFU) were injected intradermally into anesthetized mice on each side of a shaved flank region. Mice were monitored daily. At 72 h postinfection, mice were euthanized, and abscesses were excised, homogenized, and plated to enumerate CFU.

Statistical analysis

For in vitro growth assays, western blots, and metabolite quantifications, data shown are representative of at least three independent experiments conducted in triplicate. qRT-PCR experiments included two independent biological replicates conducted in technical triplicate with all data shown. All animal studies include combined data from at least two independent experiments. Statistical analyses were conducted using GraphPad Prism version 10. Comparisons among three or more groups were conducted using either one-way ANOVA followed by Tukey’s post hoc test or Kruskal–Wallis test followed by Dunn’s post-test, whereas differences between two groups were analyzed with the Mann–Whitney test. A p-value below 0.05 was considered statistically significant. Statistical significance is indicated as follows: *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.