Conserved and context-dependent changes in pathogen and host cell gene expression during Staphylococcus aureus intracellular infection

Abstract

Background

The opportunistic intracellular lifestyle of Staphylococcus aureus contributes to its ability to maintain chronic infective states. However, knowledge of pathogen-encoded factors that enable S. aureus to invade and persist within human cells, host cell responses to infection, and how such features vary across different bacterial strains and host cell types remains incomplete. Here, we used dual RNA-seq to examine alterations in S. aureus and host cell transcription following bacterial invasion, comparing results from two different S. aureus strains (JE2 and ATCC29213) and human cell lines (THP-1-derived macrophages and CCL-185 lung epithelial cells) over time.

Results

The fates of bacteria and host cells, and their changes in gene expression were dependent on the combination of strain and cell type examined. Transcriptional changes largely occurred early in infection (by 3 h), with subtler changes over longer timescales (24 h). Some bacterial alterations, including upregulation of metal ion homeostasis and pyrimidine biosynthesis and downregulation of anaerobic respiration and certain virulence factors, were consistent across infections, indicating general importance to intracellular pathogenesis. Other changes occurred only in particular S. aureus strains or during infection of specific host cell lines. Host cell analysis revealed marked differences in transcriptional response between infected cell types, and specific alterations dependent on the infecting S. aureus strain. Phenotypic contributions of bacterial genes with the most significant expression changes during infection were evaluated using CRISPRi knockdown experiments, most of which confirmed significant alterations in host cell invasion phenotypes under the conditions tested (26 of 32, 81%).

Conclusions

This work identifies a limited number of conserved transcriptional changes occurring during S. aureus intracellular pathogenesis that were common to all host cell-pathogen combinations examined, suggesting these are key to the infection process. However, most transcriptional changes were contingent and specific to the strain of bacteria, type of human cell, or combination of the two, indicating significant variation in context-dependent responses.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05297-w.

Introduction

Staphylococcus aureus is a prevalent opportunistic pathogen responsible for a high burden of morbidity and mortality worldwide [1–5]. Although most frequently associated with skin and soft tissue infections [6], S. aureus can cause a spectrum of clinical diseases ranging from relatively benign to life-threatening and can impact a variety of organ systems [3, 5, 7, 8]. Such infections, especially those involving non-cutaneous sites, are often chronic or relapsing and difficult to eradicate permanently even with appropriate antibiotic treatment [9–13]. Chronic S. aureus infections may persist for years, as in chronic rhinosinusitis [14], or even decades, as in cystic fibrosis respiratory infection [15–17].

An important contributor to S. aureus persistence is the organism’s capacity as a facultative intracellular pathogen [18–23]. Multiple studies have shown that S. aureus is able to penetrate and survive within professional and non-professional phagocytes alike [13, 20, 24–27]. (Here, we will use the term “invasion” to generally describe the entry and early survival of pathogenic bacteria into any host cell type, regardless of its phagocytic activity.) Although in some instances invading bacteria rapidly lyse host cells, in others S. aureus can become metabolically semi-dormant and remain intracellular for extended periods [20, 28, 29]. The transition to an enduring intracellular state has been shown to protect S. aureus from the host immune response and many antibiotics, establishing a reservoir that perpetuates chronic disease [13, 14, 20, 24, 25, 30]. Of particular interest is the ability of S. aureus to infect macrophages, which both shields bacteria from host innate immunity and potentially provides a mechanism for bacterial dissemination [22, 31–33].

Prior work has offered insight into some of the S. aureus genes and pathways contributing to intracellular invasion and survival using various laboratory models or in vivo disease states [26, 31, 34–37]. Although such studies have been illuminating, a global understanding of the pathogen-encoded factors enabling S. aureus to invade and persist within human cells and the host cell responses to such invasion has yet to emerge. Diversity in both host cell- and pathogen-encoded factors likely influences which mechanisms are employed in different contexts, as suggested by basal differences in the secretome [28] and invasive capacity [27] among S. aureus strains, and by differences in infection efficiency and intracellular lifestyle outcomes that correlate with host cell properties [28, 38, 39], but this variability has not yet been systematically explored.

Here, we investigated genes and pathways important for S. aureus intracellular pathogenesis using Dual RNA-Seq [40], a technique that enables simultaneous cataloging of gene expression in an intracellular bacterium and its eukaryotic host cell. We ascertained transcriptional changes occurring in pathogen and host cells at two timepoints: immediately after invasion (3 h) and after intracellular infection has stabilized (24 h). To ascertain the generalizability of our findings and to identify differences in transcriptional responses in different contexts, we conducted studies employing two phylogenomically dissimilar S. aureus reference strains from lineages commonly causing human infection [41] (ATCC29213 [Sequence Type 5, Clonal Complex 5] and JE2 [Sequence Type 8, Clonal Complex 8, derived from the USA300 lineage that is the dominant MRSA (methicillin-resistant S. aureus) clade in the United States [42, 43]) infecting two distinct cultured human cell types involved in S. aureus disease (THP-1-derived macrophages and CCL-185 lung epithelial cells) [20, 44, 45].

Methods

Strains, cell lines, and growth conditions

S. aureus strain ATCC29213 was obtained from the American Type Culture Collection (ATCC), and strain JE2 from the Biodefense and Emerging Infections Research Resources Repository (BEI Resources). Bacteria were maintained at 37 °C using LB broth.

THP-1 and CCL-185 cells were obtained from ATCC and grown according to the supplier’s recommendations at 37 °C in a humidified 5% (v/v) CO2 air atmosphere using medium supplemented with 0.1 mg/ml L-glutamine, 0.1 mg/ml streptomycin, 100 U/ml penicillin (PAA) and 20% (v/v) Nu-Serum Serum Replacement (Corning). Basal medium for THP-1 was RPMI 1640 medium (Gibco), whereas F-12 Nutrient Mix medium was used for CCL-185 (Gibco).

THP-1 cells were differentiated into macrophages before bacterial invasion as described previously [37].

Bacterial invasion of human cell lines

Selection for bacteria having entered host cells was based on previously described lysostaphin protection assays [46]. Bacteria were expanded overnight in LB, washed twice in Dulbecco’s phosphate-buffered saline (DPBS) (Difco) to remove soluble toxins, and cell density assessed by optical density at 600 nm (OD600). All invasion experiments were performed in triplicate at 37 °C with 5% atmospheric CO₂, using 24 well tissue culture plates. A multiplicity of infection for each cell line that empirically maximized numbers of successfully internalized bacteria while minimizing host cell death during the course of the experiments was employed. Bacteria were therefore combined with cultured cells in basal medium, using a multiplicity of infection (MOI) of 150 for differentiated THP-1 and a MOI of 800 for CCL-185. MOI was calibrated to achieve a population-level average of one successfully internalized bacterium per host cell: for THP-1 cells, we found that an MOI of 150 S. aureus was optimal and resulted in ~ 0.97 bacteria internalized per cell, whereas for CCL-185 an MOI of 800 was used and achieved ~ 0.27 intracellular bacteria per host cell, with higher bacterial burdens found to adversely affect immediate host cell viability. Aliquots of the initial bacterial inoculum were plated to determine the count of viable organisms applied. After 1 h, basal medium was replaced with medium containing 50 µg/ml lysostaphin (Sigma) to kill uninternalized bacteria. All invasion experiments were performed in triplicate.

The numbers of viable and dead cultured host cells were assessed prior to S. aureus infection, then at 3 h and 24 h after infection using trypan blue exclusion assays as described elsewhere [47], with measurements performed at least in triplicate. The number of viable intracellular bacteria at the same time points post-infection was separately determined by plating serial dilutions of cell lysate as described elsewhere [37]. Viable intracellular bacteria recovered at 24 h were normalized to the number of viable intracellular organisms counted at 3 h from the same inoculum to normalize for differences invasion efficiency [37].

Dual RNA-seq

The full series of three experimental replicates performed for each combination of S. aureus strain and host cell line (3 h and 24 h post-infection), as well as three replicates of uninfected cell lines and bacteria collected immediately prior to infection (time 0 h) were subjected to high throughput sequencing.

Medium was aspirated and cells washed with DPBS prior to RNA extraction. RNA was extracted using GenElute Total RNA Purification Kit (Sigma-Aldrich) according to manufacturer’s instructions, with the addition of 1% lysostaphin to lysis buffer and 10% lysostaphin to the TE buffer used for resuspension of the bacterial pellet. DNA was eliminated from RNA purifications using On-Column DNase I Digestion Set (Sigma-Aldrich) per manufacturer’s instructions. RNA was recovered in 50ul and stored at -80 °C until library preparation. Next-generation sequencing library preparation was performed using Stranded Total RNA Library Prep with Ribo Zero Kit (Illumina). Libraries were sequenced using a NextSeq2000 implementing 200 cycle paired-end SBS chemistries (Illumina), with a 750pM loading concentration.

S. aureus CRISPRi knockdown

CRISPRi knockdown, construction of the GFP-targeted control vector, and data analysis were conducted as described previously [37]. Oligonucleotides used for knockdown of targeted genes and gene-targeted real-time PCR, as well as the level of knockdown achieved, are provided in Supplementary Table 1.

Growth rates of CRISPRi knockdown strains and controls were assessed from a minimum of three replicates as described elsewhere [37], except that doubling times were determined in basal medium appropriate for the specific pairing of knockdown strain and host cell line of interest.

Data analysis

Primary analysis of bacterial RNA-seq data was conducted using EDGE-pro v1.3.1 [48] with default parameters, except that the -M flag was set to a value of 800 to allow for mapping of higher molecular weight library fragments. Appropriate reference genomes for ATCC29213 (GenBank CP007454.1) and JE2 (GenBank CP020619.1) were used for analysis.

Homologous S. aureus genes in strains ATCC29213 and JE2 were identified on the basis of 80% similarity threshold as described elsewhere [17]. Primary analysis of human RNA-seq data was performed through Salmon v1.8.0 [49] with the gcBias option activated, and using the hg38 reference genome.

Differential expression analyses were performed using DESeq2 v1.49.2 [50] using the set of three experimental replicates for each condition. Genes were required to exhibit an absolute fold change of ≥ 2 and a Benjamini-Hochberg adjusted significance of p ≤ 0.05 (i.e., significance after correction for multiple testing) to be considered differentially expressed. Principal component analysis of DESeq2-normalized read counts was performed using the prcomp function of the R ‘stats’ package. Data visualizations were generated using the ggplot2 [51] and plotly [52] R packages.

Gene ontology (GO) enrichment analysis was conducted by analyzing sets of upregulated and downregulated genes separately [53]. Bacterial GO analysis was conducted using FUNAGE-Pro v2 [54] in conjunction with the reference genomes indicated above, and human analyses using DAVID (v2021, using v2023q4 knowledgebase) [55], with Benjamini-Hochberg corrected p < 0.05 considered significant in both analyses.

Comparisons of bacterial phenotypes under experimental conditions were assessed using unpaired two-tailed T-tests with p < 0.05 considered significant, with exception of testing intracellular index which used Z-test for proportions to compare all normalized measurements for a given condition against the null hypothesis of no change (i.e., equal to value of 1), with p < 0.05 considered significant.

Assessment of bacterial doubling times was conducted using two-sided, unpaired Wilcoxon rank sum tests.

Clinical trial number

Not applicable. This in vitro study does not meet the definition of a clinical trial.

Results

Infection outcomes for bacteria and host cells

We initially characterized the functional impact of infection on host cells and the S. aureus strains invading them. Intracellular bacterial counts were assessed 3 h after their application to host cells (Fig. 1A). Statistically equivalent proportions of ATCC29213 and JE2 were recovered after infecting CCL-185. Compared to CCL-185, THP-1 derived macrophages yielded greater proportions of internalized bacteria from both strains (ATCC29213 p = 4.52 × 10^{− 4}, JE2 p = 1.03 × 10⁻⁵, unpaired two-tailed T test), likely reflecting active phagocytic uptake [56]. More JE2 internalized into THP-1 macrophages than ATCC29213 (p = 4.5 × 10⁻⁵, unpaired two-tailed T test). The proportion of surviving intracellular S. aureus was also assessed 24 h after cell entry [37] (Fig. 1B). At 24 h the fractions of ATCC29213 and JE2 surviving in CCL-185 were significantly reduced relative to their counts at 3 h (ATCC29213 Z-test for proportions Z-score = -2.05, p = 0.04; JE2 Z-score = -2.83, p = 0.005) and statistically equivalent between the two bacterial strains, indicating loss of viable intracellular bacteria over time in that host cell type. In contrast, intracellular bacterial counts of both strains remained unchanged in THP-1 macrophages between timepoints.

Fig. 1.

Proportion of intracellular S. aureus after host cell infection. Results in both panels are stratified by S. aureus strain and host cell line. Each point represents an individual experiment (n = 12), with horizontal line indicating median values. (A.) The proportion of viable intracellular bacteria ascertained at 3 h post-infection were normalized to the number of viable bacteria in the initial inoculum. Comparisons between conditions are indicated by black bars at top, with significant differences (p<5 × 10⁻⁴, unpaired two-tailed T test) marked by an asterisk. (B.) Intracellular survival index represents the counts of surviving intracellular S. aureus assessed 24 h post-infection after normalization to the counts of viable intracellular bacteria determined 3 h post-infection. Significant deviations from an index value of 1 (p < 0.05) are indicated by an asterisk. NS, not significant

Counts of intracellular bacteria must be interpreted relative to host cell status after infection. We therefore used trypan blue staining to quantify living and dead host cells at each time point. Uninfected CCL-185 continued to replicate after seeding, resulting in significantly increased viable cell counts at 24 h (p = 0.03, unpaired two-tailed T test, Fig. 2A). CCL-185 counts did not significantly differ from the uninfected control after infection with ATCC29213, denoting negligible impact to host cell fitness. In contrast, infection of CCL-185 with JE2 significantly reduced host cells by 24 h (p = 0.002, unpaired two-tailed T test). Intact, dead host cells were minimal (0% to 1.2%, Supplementary Table 2) and statistically equivalent across all time points and conditions tested, indicating that JE2 either lysed host cells or reduced their growth rates. A different pattern was seen for infections of terminally differentiated THP-1-derived macrophages (Fig. 2B), which remained stable but did not replicate in infection-free experimental conditions. Infection with either ATCC29213 or JE2 significantly reduced macrophage counts relative to uninfected controls by 3 h (p = 0.002 and p = 0.05, respectively, unpaired two-tailed T test), with further reductions at 24 h (p = 4.5 × 10⁻⁵ and p = 2.2 × 10⁻⁶, respectively, unpaired two-tailed T test). JE2 impacted THP-1 cell counts more severely at 24 h than ATCC29213 (p = 0.014, unpaired two-tailed T test). Dead host cells were again minimal (0.1% to 0.6% Supplementary Table 2) and statistically equivalent across assayed conditions, indicating loss by cell lysis.

Fig. 2.

Host cell viability after S. aureus infection. Host cell counts at various timepoints before (0 h) and after (3 and 24 h) after S. aureus infection, normalized to cell counts at 0 h. Results are shown separately for CCL-185 epithelial cells (A.) and THP-1-derived macrophages (B.) for infection with the two S. aureus strains and uninfected controls. Values in infected cells that are significantly different from time matched controls (p < 0.05, unpaired two-tailed T test) are indicated by an asterisk

Taken together, these data indicate different models of infection for specific combinations of host cell lines and bacterial strains used in this study. In THP-1, absolute counts of intracellular bacterial remained stable in the context of ongoing host cell lysis: This indicates an increasing number of intracellular bacteria within each host cell, a finding that could be consistent with either ongoing cycles of bacterial release and re-infection and/or intracellular bacterial replication. In CCL-185 epithelial cells, viable intracellular ATCC29213 counts remained stable without negatively impacting host cell counts, suggesting attenuated virulence and persistence, whereas infection with JE2 reduced both counts of CCL-185 cells and ultimately the fraction of viable intracellular bacteria, consistent with greater bacterial virulence and active host cell lysis.

Landscape of S. aureus transcriptional changes during host cell infection

We used dual RNA-seq to ascertain shifts in S. aureus gene expression during host cell infection. We examined both S. aureus strains immediately prior to infection (0 h), and after invasion of both host cell types at early (3 h) and later (24 h) timepoints (Fig. 3A).

Fig. 3.

Differentially expressed genes identified for experimental comparisons. Scatterplots depict comparative log₂-fold gene expression changes identified in (A.) bacterial and (B.) human host cells for each of the indicated pairwise contrasts, stratified by timepoints (0 to 3 h and 3 to 24 h). X- and Y-axes in each subplot indicate the specific data sets that are being compared. Data point size is scaled proportional to the magnitude of gene expression change and colored according to their relative directionality in the two comparator conditions. Linear regression is indicated by gray dashed line, with 95% confidence interval shown by gray shading and R² correlation value reported at the top of each plot

We performed principal component analysis (PCA) to evaluate the global transcriptional state of S. aureus strains under these different conditions (Fig. 4A, Supplementary Fig. 1), considering the set of 2,232 “core genes” common to both S. aureus strains. The first principal component accounted for 54.8% of the variance in bacterial expression changes, and corresponded most closely to host cell type (i.e., no infection, THP-1 cell infection, or CCL-185 infection). Although the second principal component (7.5% variance) did not have an obvious identifiable biological correlate, the third (5.5% variance) best distinguished S. aureus strains, indicating distinct transcriptional responses of JE2 and ATCC29213 across experimental conditions. When infecting either host cell type, strain transcriptional profiles changed most markedly between the time prior to infection and the earliest timepoint after infection, with more subtle distinctions between 3 h and 24 h post-infection.

Fig. 4.

Principal component analysis of bacterial and human transcriptional profiles. Patterns of bacterial (A.) and human host cell (B.) transcription are depicted for the first three principal components (PCs) with the proportion of variance summarized by each indicated in parenthesis. Corresponding two-dimensional pairwise plots are provided for bacteria (Supplementary Figure 1) and host cells (Supplementary Figure 2)

Conserved changes in S. aureus gene expression during human cell infection

We compared core gene expression changes during the largest shifts (0 to 3 h) to identify differentially expressed genes (DEGs) that were consistently observed across S. aureus strains during infection of both host cell types.

Small subsets of S. aureus genes were consistently upregulated (n = 20) or downregulated (n = 23), indicating a core set of DEGs with stereotyped patterns of expression during initial intracellular infection of human cells (Table 1). Gene ontology (GO) enrichment analysis (Fig. 5A) of consistently upregulated DEGs indicated significant (Benjamini-Hochberg corrected p < 0.05) functional enrichment for zinc, cobalt, and nickel ion transport, and de novo pyrimidine biosynthesis. Virulence factors sortase B and fibronectin-binding protein A were also identified. Consistently downregulated genes did not show significant enrichment for any GO categories, but were largely involved in metabolism (n = 11, including 4 factors involved in redox and 4 in anaerobic metabolism). Notably, three canonical virulence factors (clpX, isaB, and delta hemolysin) were downregulated.

Table 1.

Conserved gene expression changes across S. aureus strains during infection of both host cell types

JE2 locus	ATCC29213 locus	Description	Expression change
B7H15_00640	CH52_05150	iron-siderophore ABC transporter permease	upregulated
B7H15_06280	CH52_13060	orotate phosphoribosyltransferase	upregulated
B7H15_06275	CH52_13065	orotidine 5’-phosphate decarboxylase	upregulated
B7H15_01055	CH52_04775	PTS glucose EIICBA component	upregulated
B7H15_13700	CH52_06670	ABC transporter ATP-binding protein	upregulated
B7H15_06575	CH52_12780	phosphatidate cytidylyltransferase	upregulated
B7H15_13905	CH52_06485	fibronectin-binding protein A	upregulated
B7H15_01265	CH52_04545	PrsW family intramembrane metalloprotease	upregulated
B7H15_02950	CH52_03060	50 S ribosomal protein L10	upregulated
B7H15_12660	CH52_07560	N-acetyltransferase	upregulated
B7H15_13705	CH52_06665	ABC transporter ATP-binding protein	upregulated
B7H15_13710	CH52_06660	nickel ABC transporter permease	upregulated
B7H15_02955	CH52_03055	50 S ribosomal protein L7/L12	upregulated
B7H15_13725	CH52_06645	staphylopine dehydrogenase CntM	upregulated
B7H15_13985	CH52_06430	lantibiotic ABC transporter ATP-binding protein	upregulated
B7H15_13990	CH52_06425	ABC transporter permease	upregulated
B7H15_06675	CH52_12690	peptidase M16	upregulated
B7H15_05890	CH52_00215	SrtB family sortase	upregulated
B7H15_06670	CH52_12695	peptidase M16	upregulated
B7H15_07095	CH52_12320	glycine/betaine ABC transporter	upregulated
B7H15_09165	CH52_10690	threonine–tRNA ligase	downregulated
B7H15_14180	CH52_06260	Clp protease ClpX	downregulated
B7H15_00710	CH52_05090	diacetyl reductase ((S)-acetoin forming)	downregulated
B7H15_01750	CH52_04140	MurR/RpiR family transcriptional regulator	downregulated
B7H15_04865	CH52_01160	sodium: proton antiporter	downregulated
B7H15_14695	CH52_05815	immunodominant staphylococcal antigen IsaB	downregulated
B7H15_06815	CH52_12550	MerR family transcriptional regulator	downregulated
B7H15_14915	CH52_05610	ATP phosphoribosyltransferase regulatory subunit(hisZ)	downregulated
B7H15_11280	CH52_08825	delta-hemolysin	downregulated
B7H15_08880	CH52_10945	LLM class flavin-dependent oxidoreductase	downregulated
B7H15_02185	CH52_03735	NAD(P)-dependent oxidoreductase	downregulated
B7H15_09750	CH52_10165	phosphoenolpyruvate carboxykinase (ATP)	downregulated
B7H15_03705	CH52_02455	ABC transporter permease	downregulated
B7H15_01205	CH52_04595	formate acetyltransferase	downregulated
B7H15_14950	CH52_05575	polyisoprenoid-binding protein	downregulated
B7H15_01210	CH52_04590	pyruvate formate-lyase-activating enzyme	downregulated
B7H15_03700	CH52_02460	ABC transporter ATP-binding protein	downregulated
B7H15_00945	CH52_04890	aldehyde dehydrogenase	downregulated
B7H15_12240	CH52_07950	2,5-diketo-D-gluconic acid reductase	downregulated
B7H15_11325	CH52_08785	ammonia permease	downregulated
B7H15_01860	CH52_04040	ring-cleaving dioxygenase	downregulated
B7H15_07395	CH52_12050	regulatory protein MsaA	downregulated
B7H15_06925	CH52_12480	hypothetical protein	downregulated

Fig. 5.

Clustered heatmap of GO enrichment analysis. Comparative conditions within each panel are indicated below respective columns. Only conditions with one or more significant results are displayed. Blue or red coloration indicate significant GO enrichment in downregulated or upregulated DEGs, respectively. Color intensity indicates the Benjamini-Hochberg adjusted p-value for each GO term. Panels depict bacterial (A.) or human host cell (B.) GO enrichment analyses. All conditions correspond to comparison of the timepoint prior to infection (0 h) against the timepoint 3 h post-infection, unless otherwise indicated

Strain-specific changes in S. aureus gene expression during human cell infection

We next explored S. aureus DEGs that consistently differed between the two bacterial strains when infecting both cell lines, again considering changes that occurred at the earliest timepoint post-infection (Supplementary Table 3). Genes consistently upregulated in JE2 that remained unaltered in ATCC29213 (n = 21) were functionally enriched for multiple GO categories (n = 8) including ATP-related processes, leucine biosynthesis, and 3-isopropylmalate dehydratase activity, while downregulated JE2-specific genes (n = 30) showed enrichment for acetoin catabolism alone. Although ATCC29213 similarly had genes that were consistently upregulated (n = 9) and downregulated (n = 23) during infection that were unaltered in JE2, neither such group showed functional enrichment. However, virulence genes (n = 5), especially those involving fibrinogen binding (n = 3), were prominently upregulated by ATCC29213, and metabolic genes (n = 12, ten involving amino acid metabolism) constituted the majority of downregulated DEGs. No genes were consistently overexpressed in one strain and consistently downregulated in the other, or vice versa.

S. aureus gene expression changes between strains in different host cell types

We next compared core gene expression profiles between S. aureus strains to identify conserved and strain-specific DEGs (Fig. 3) during infection of specific host cell types, again considering changes at 0 to 3 h.

When bacteria were infecting THP-1 (Supplementary Table 4), 33 genes were overexpressed in both strains, and 45 were consistently downregulated. We excluded features that were consistently altered during infection of both cell lines (i.e., Table 1) to identify changes unique to macrophage invasion (Supplementary Table 5). Upregulated, macrophage-specific DEGs (n = 14) did not show GO category enrichment, although consistently downregulated genes (n = 22) were significantly enriched for glycerol catabolism. Other expression changes during macrophage infection were specific to one S. aureus strain or the other. After excluding constitutional differences in expression changes between strains (i.e., Supplementary Table 3), upregulated JE2 DEGs (n = 101, Supplementary Table 6) showed no GO category enrichment but were notable for type VII (n = 7) and Sec (n = 3) secretion system components, while downregulated DEGs (n = 73) were enriched for glycerol catabolism. DEGs specific to ATCC29213 (Supplementary Table 6) showed no GO category enrichment in upregulated (n = 27) or downregulated (n = 45) genes.

During CCL-185 infection 96 genes were upregulated and 106 were downregulated across both S. aureus strains (Supplementary Table 7). Upregulated genes specific to epithelial cell infection (n = 75, Supplementary Table 8) were enriched for 22 GO terms, most significantly putrescine transport, rRNA binding, and ribosomal structural constituents. Downregulated factors (n = 83) showed no significant GO enrichment, but primarily reflected metabolic functions (n = 40). As with macrophage infection, multiple DEGs were specific to particular S. aureus strains in CCL-185 (Supplementary Table 9), after excluding genes having consistent expression differences between the strains (i.e., Supplementary Table 3).

Upregulated factors specific to JE2 (n = 65) showed enrichment for 14 GO features, most significantly inorganic phosphate transmembrane transport and fatty acid biosynthesis, whereas downregulated JE2-specific DEGs (n = 96) showed no functional enrichment. Upregulated ATCC29213-specific genes (n = 32) also lacked GO term enrichment, but downregulated DEGs (n = 114) spanned 20 GO categories, most significant being biosynthesis of O antigen and siroheme.

S. aureus gene expression changes within individual strains in different host cell types

We next compared DEGs for individual S. aureus strains when they were infecting different host cell types, regarding the full complement of annotated genes in each strain. We again evaluated changes occurring between the earliest two timepoints (0 to 3 h), where such shifts were greatest.

Of the 2,748 transcripts in JE2 (Supplementary Table 10), the majority (n = 2,041) were not significantly altered during infection of either host cell line. 110 genes, enriched for 27 GO categories, were upregulated during infection of both host cell types, while downregulated DEGs common to both host cell types (n = 110) were enriched for 6 GO categories, particularly acetoin catabolism and ferric iron binding. Other JE2 DEGs were specific to host cell type. Those exclusively upregulated during macrophage infection (n = 82) were not enriched for GO categories but primarily had metabolic functions (n = 28, including 14 involved in amino acid metabolism) and were notable for secretion systems (n = 9, including accessory Sec and type VII).

Downregulated DEGs (n = 76) showed GO enrichment for DNA-binding transcription factor activity, protein-DNA complex, and extracellular region categories. JE2 DEGs specifically upregulated during epithelial cell infection (n = 134) encompassed 23 GO categories, while downregulated DEGs (n = 191) showed no GO enrichment. Four genes were modulated in opposite directions during JE2 infection of different cell lines: acetyltransferase, cobalt ABC transporter permease, and a hypothetical protein were upregulated during macrophage invasion and downregulated during epithelial infection, while PTS sugar transporter subunit IIA showed the opposite pattern.

ATCC29213 (Supplementary Table 11) encodes 2,501 transcripts, of which 1,942 were unaltered during either host cell infection. Only 41 genes were consistently upregulated during infection of both cell types, encompassing 8 GO categories relevant to transport and homeostasis of cobalt and various metal ions, cell adhesion, and de novo pyrimidine biosynthesis.

ATCC29213 genes always downregulated during host cell infections (n = 74) involved 16 GO categories, including amino acid metabolism, peptidoglycan turnover, and imidazoleglycerol-phosphate synthesis (involved in de novo purine biosynthesis). ATCC29213 DEGs that were specific to host cell types were more numerous. Genes upregulated during macrophage infection (n = 66) showed enrichment of GO categories (n = 10), including plasminogen activation, and biotin and queuosine biosynthesis, while downregulated genes (n = 62) were enriched across GO categories (n = 7) including phosphate transport and lactose and glycerol metabolism.

Upregulated ATCC29213 DEGs specific to CCL-185 epithelial cell infection (n = 122) were enriched for GO categories (n = 39) including metal ion transport, de novo pyrimidine biosynthesis, and queuosine biosynthesis, while downregulated DEGs (n = 193) did not show enrichment for GO categories but primarily involved metabolic functions (n = 92). A single gene, the antiholin LrgB, was expressed in opposing directions during infection of the two cell lines, being upregulated in CCL-185 infections and downregulated in macrophage infection.

S. aureus gene expression differences between early and late infection

We separately surveyed bacterial gene expression within individual S. aureus strains to describe the more limited changes occurring between early (3 h) and late (24 h) stages of infection in both host cell lines. Twenty-five JE2 genes were significantly altered during THP-1 infection (Supplementary Tables 12, 18 downregulated and 7 upregulated), which were largely hypothetical or uncharacterized (n = 16). Only two JE2 DEGs changed in CCL-185 (Supplementary Table 13): upregulation of gamma-hemolysin subunit A and downregulation of riboflavin biosynthesis protein RibBa. Similarly, during ATCC29213 infection of THP cells (Supplementary Table 14), 11 genes were upregulated at late infection (primarily those involving siderophore biosynthesis, n = 5), while only one hypothetical gene was downregulated.

Late infection of CCL-185 by ATCC29213 (Supplementary Table 15) was marked by upregulation of 7 genes and downregulation of 3, without consistent functions. In neither strain were DEGs at this later infection phase altered in both host cell lines, and only a single core (betaine-aldehyde dehydrogenase) gene showed a conserved change across strains between 3 h and 24 h, being consistently upregulated during late macrophage infection (Supplementary Tables 16 and Supplementary Table 17).

Landscape of transcriptional changes in host cells during S. aureus infection

We conducted parallel analyses to examine transcriptional changes occurring in human cell lines during S. aureus infection (Fig. 3B). PCA (Fig. 4B, Supplementary Fig. 2) indicating that macrophages underwent more dramatic changes in transcription than epithelial cells. The first principal component separated data by host cell type and accounted for 95.1% of the data variance. The second principal component (1.1% variance) separated most THP-1 macrophage data by timepoint. While JE2 infection of macrophages generated an initial transcriptional signature that was relatively durable over the experimental period, ATCC29213 infection elicited more successive changes in macrophage expression between early and late infection.

Conserved and cell-type specific changes in human gene expression during S. aureus infection

We again compared and contrasted host cell DEGs under specific conditions, focusing on those identified in the period from 0 to 3 h, which encompassed the largest shifts in gene expression for both human cell lines.

Expression changes in 12 genes were common to the two cell lines during infection by both S. aureus strains, all of which were upregulated and showed significant GO enrichment (Fig. 5B) for cellular response to lipopolysaccharide (Table 2).

Table 2.

Conserved gene expression changes across host cell types during infection by both S. aureus strains

Gene ID	Gene name	Gene description	Expression change
ENSG00000023445	BIRC3	baculoviral IAP repeat containing 3	upregulated
ENSG00000073756	PTGS2	prostaglandin-endoperoxide synthase 2	upregulated
ENSG00000090339	ICAM1	intercellular adhesion molecule 1	upregulated
ENSG00000099625	CBARP	CACN subunit beta associated regulatory protein	upregulated
ENSG00000108342	CSF3	colony stimulating factor 3	upregulated
ENSG00000145911	N4BP3	NEDD4 binding protein 3	upregulated
ENSG00000163734	CXCL3	C-X-C motif chemokine ligand 3	upregulated
ENSG00000164400	CSF2	colony stimulating factor 2	upregulated
ENSG00000169429	CXCL8	C-X-C motif chemokine ligand 8	upregulated
ENSG00000185101	ANO9	anoctamin 9	upregulated
ENSG00000259374	NDUFB4P11	NADH: ubiquinone oxidoreductase subunit B4 pseudogene 11	upregulated
ENSG00000284713	SMIM38	small integral membrane protein 38	upregulated

More DEGs were consistent within host cell lines when infected by either S. aureus strain. THP-1 macrophages showed 434 upregulated DEGs with significant enrichment for 46 GO terms including several relevant to immune response, and 21 downregulated DEGs without enrichment for specific GO terms (Supplementary Table 18). Infected CCL-185 cells consistently upregulated 66 genes, enriched for lipopolysaccharide response and neutrophil chemotaxis (n = 3), and 7 downregulated DEGs with no significant GO enrichment (Supplementary Table 19).

Other human DEGs were specific to the host cell types being infected (Supplementary Table 20). Infection of THP-1 macrophages by either S. aureus strain resulted in consistent upregulation of 415 DEGs and downregulation of 21 that were unaltered in CCL-185 cells.

While the downregulated DEGs did not show significant GO enrichment, upregulated genes were enriched for multiple functions (n = 29), predominantly relating to inflammation and immune response. DEGs specific to CCL-185 were less numerous, with 15 upregulated and 7 downregulated genes that remained unaltered in THP-1 cells, with neither category showing GO enrichment.

Changes in host cell DEGs specific to S. aureus strains

We enumerated host cell DEGs that were specific to infection by one S. aureus strain or the other, again examining expression changes 3 h post-infection.

In THP-1 (Supplementary Table 19), 486 genes upregulated during ATCC29213 infection were unaltered during infection with JE2, correlating with GO enrichment categories (n = 37) primarily involving immune response, while 231 genes were downregulated exclusively during ATCC29213 infection without enrichment for GO categories. JE2 infection caused specific upregulation of 2,099 genes that were unaltered during ATCC29213 infection, comprising assorted GO categories (n = 12), and downregulation of 87 genes enriched for functions (n = 5) mostly relevant to chromatin structure and organization.

Compared with THP-1 cells, fewer strain-specific changes were observed in CCL-185 (Supplementary Table 20). Epithelial cells showed 27 genes upregulated and 11 downregulated during ATCC29213 infection that were unaltered during JE2 infection, with neither category showing significant GO enrichment. 63 genes were specifically upregulated during JE2 infection, similarly without functional enrichment, while the 24 JE2-asociated downregulated genes were enriched for proton transport and proton-motive ATP synthesis (n = 2).

Host cell gene expression differences between early and late infection

We next compared transcriptional changes occurring in host cells during the transition between early (3 h) and later infection (24 h).

Examining changes that were consistently induced by both S. aureus strains, in THP-1 derived macrophages (Supplementary Table 21), expression of 21 genes increased between early and late infection, while 20 decreased, with neither category significantly enriched for GO categories. Over the same period, CCL-185 (Supplementary Table 22) had 5 genes that were consistently upregulated later in infection and 37 downregulated, with the latter category enriched for GO categories (n = 9) including those relating to innate immunity and interferon response.

We also cataloged DEGs that arose during infection with one bacterial strain or the other over the later timeframe. In THP-1 macrophages (Supplementary Table 21), 740 genes were specifically upregulated during ATCC29213 infection and 318 were downregulated, encompassing 24 and 21 GO terms, respectively. JE2 infection specifically upregulated 979 genes in THP-1 during this time period, enriched for GO categories (n = 2) relevant to G protein signaling and chemical stimulus perception, and downregulated 90 genes which lacked significant GO enrichment. Interestingly, 15 genes lacking any obvious functional theme were regulated in opposite directions by macrophages during infection by the different strains. For CCL-185 (Supplementary Table 22), ATCC29213 infection resulted in specific upregulation of 21 genes and downregulation of 98, while JE2 infection upregulated 30 genes and downregulated 36: none of these groups of DEGs showed significant enrichment for GO categories.

Experimental evaluation of S. aureus factors contributing to intracellular pathogenesis

To experimentally assess the involvement of S. aureus genes with the greatest transcriptional changes in intracellular infection, we generated isogenic CRISPR interference (CRISPRi) [57] knockdown mutants and assessed their effects on pathogenesis empirically by measuring the invasive capacity of mutant strains when applied to host cells. The ten most significant genes identified from each S. aureus strain when infecting either host cell type at 3 h were selected for testing of invasion phenotypes using the same pairing of strain and cell line at that timepoint (Supplementary Table 1). Fourteen of the targeted factors could be successfully knocked down in ATCC29213, two of which (CH52_04575 and CH52_13065) were marked by significantly altered expression during infection of both host cell lines. Similarly, 14 targeted genes could be knocked down in JE2. Two genes were prioritized in common between the two S. aureus strains (B7H15_00640/CH52_05150 [sirB] and B7H15_04440/CH52_01815 [vwb]), albeit inconsistently during infection of different host cell types. We determined that the doubling times of CRISPRi knockdown strains did not differ significantly from those of controls when cultured in invasion medium (Supplementary Table 23), indicating that any differences in invasive capacity could not be attributed to differences in underlying growth during the brief period of incubation with host cells.

CRISPRi knockdown (Fig. 6) produced significant differences (p < 0.05, two-tailed T-test) in measured host cell invasion at 3 h post-infection relative to strain-matched controls for the majority of conditions tested (n = 26 of 32, 81%), indicating an impact of those genes on S. aureus host cell invasion phenotypes. However, there was little correlation between the directionality of gene expression change seen during infection and the phenotypic effect after knockdown. When testing 27 conditions involving genes showing significant upregulation by RNA-seq, only 5 (18.5%) showed the anticipated reduction in host cell invasion when gene expression was blunted by CRISPRi knockdown whereas 18 (67%) paradoxically showed elevated invasive capacity. Conversely, knockdown in 2 of the 5 conditions (40%) involving genes having reduced expression during pathogenesis expectedly resulted in increased host cell invasion while 2 others (40%) suppressed invasion.

Fig. 6.

Infection phenotypes of S. aureus isogenic knockdowns. Panels display invasion capacity of various isogenic CRISPRi knockdowns constructed in S. aureus strains and tested in cell lines as indicated. Results in each panel are shown relative to strain-matched controls (carrying a GFP-targeted CRISPRi vector, median value shown with dashed line), assessed at 3 h post-infection. Horizontal lines indicate median values for specified knockdown strains. Values significantly different from controls are indicated by asterisks (p < 0.05, unpaired two-tailed T-test). Gene targets tested in both host cell lines are indicated in bold. Common gene names are indicated in parentheses, where applicable. Two genes common between the S. aureus strains (vwb and sirB) are indicated in red. Identification of gene targets as significantly upregulated or downregulated by dual RNA-seq is indicated below panels

We conclude that most candidate genes selected for CRISPRi exert a measurable effect on S. aureus intracellular pathogenesis, but that overall enhancement or reduction of that phenotype does not uniformly match expectations based on the direction of transcriptional change.

Discussion

To better characterize and explore the variation in transcriptional responses occurring for invading S. aureus strains and host cell types, we utilized Dual RNA-seq to concurrently define gene expression changes occurring for two S. aureus strains undergoing intracellular infection of two distinct human cell lines. In comparing across multiple bacterial strains, host cell types, and time points, this study builds on prior studies of S. aureus applying Dual RNA-seq during intracellular infection of cultured human endothelial cells [36] and in vivo murine infection models [38, 58, 59], or RNA-seq during co-culture with epithelial cells [60], and provides a more systematic framework to identify the DEGs distinguishing between bacterial strains and defined host cell types during intracellular S. aureus pathogenesis.

Consistent with prior reports using alternative approaches, we found that rates of infection and outcomes of bacterial intracellular pathogenesis could vary given the combination of host cell [39] and S. aureus strain [27] used. Bacterial intracellular uptake was highest in THP-1 macrophages, likely reflecting active phagocytosis [56]. Infection of THP-1 by both strains showed evidence of host cell cytotoxicity and cycles of re-infection and/or intracellular replication over the period of observation, as described elsewhere [39], with JE2 evidencing greater virulence than ATCC29213. In contrast, during infection of CCL-185, JE2 caused host cell death without strong evidence of re-infection whereas ATCC29213 exhibited an attenuated state without measurable detriment to host cell growth.

Some bacterial genes showed consistent shifts in regulation across all combinations of S. aureus strain and host cell type, illustrating conserved changes in core functions needed for S. aureus to transition to intracellular pathogenesis. Genes related to metal ion homeostasis were consistently upregulated as reported in other studies [36, 58, 60], illustrating the needs of infecting bacteria to scavenge those nutrients from the intracellular environment [61]. Factors related to pyrimidine biosynthesis were also upregulated, reinforcing the role of this pathway in S. aureus in intracellular infection [62]. The central role of virulence factors sortase B and fibronectin-binding protein A, both previously implicated in host cell invasion [63, 64], are similarly highlighted during the early invasion period. Consistently downregulated genes affected more diverse pathways, and involved redox chemistry, core metabolism, and anaerobic respiration, consistent with alterations related to anaerobic or microaerophilic growth conditions and shifts in carbon source utilization necessitated by intracellular status. Consistent downregulation of two well-described virulence factors, protease clpX [65] and delta hemolysin [66], suggest active self-attenuation of S. aureus to reduce host cell killing in those successfully transitioning to a more durable intracellular state, while decreased expression of a third virulence factor, isaB, is expected to facilitate autophagic uptake by host cells [67], possibly aiding initial bacterial entry. Transporters comprised a large proportion of both upregulated and downregulated DEGs across all conditions, evidencing fundamental shifts in the uptake of nutrients and other substrates after S. aureus enters host cells.

We also identified reproducible, strain-specific transcriptional alterations distinguishing S. aureus strains, regardless of infected host cell type. Such DEGs largely encoded metabolic and transporter functions, indicating key lineage-dependent differences in expression of shared genetic content. Additionally, JE2 downregulated multiple virulence factors while ATCC29213 upregulated those relating to fibrinogen binding: given basal differences in the virulence of the two strains, these shifts may effectively equilibrate their pathogenicity to a level that is optimal for intracellular invasion.

Other S. aureus DEGs were specific to infection of particular host cell types. Here, too, DEGs largely encoded transporters and metabolic functions, suggesting niche metabolic or virulence factors requirements unique to the intracellular environment and innate resistance properties of that host cell type [38]. In THP-1, invading strains upregulated queuosine and biotin biosynthetic pathways, illustrating the importance of these metabolites during macrophage infection. Specific stress response and virulence factors were also elevated, including staphylokinase, which may signify active suppression of macrophage-specific antimicrobial peptides [68]. S. aureus DEGs that were consistently downregulated in macrophages reflected changes in substrate transport, transcriptional regulation, and central metabolism, particularly glycerol catabolism. In comparison, epithelial cell infection induced upregulation of S. aureus translational machinery, suggesting generalized augmentation of protein synthesis, as well as changes in metabolic and substrate transport genes. Upregulated virulence genes specific to epithelial cell invasion were notable for oatA, which drives lysozyme resistance [69], and well-described adhesins [70, 71] which may facilitate attachment to such non-phagocytic host cells. Downregulated DEGs during epithelial cell invasion involved central metabolism, amino acid biosynthesis, and particular stress response factors, among other features.

DEGs specific to individual S. aureus strains were also identified during the infection of the different host cell types. Both up- and downregulated genes of this class overwhelmingly encoded metabolism, virulence, transcriptional regulation, and transport functions, consistent with context-dependent metabolic and substrate dependencies existing within particular host cell types.

Although most gene expression changes occurred during the initial stages of infection, a small subset of significant DEGs arose between the 3 and 24 h time points, indicating unique adaptive responses dependent on the specific combination of S. aureus strain and host cell involved. In macrophages, most late-phase DEGs in JE2 were hypothetical proteins, however, nearly half of DEGs upregulated in ATCC29213 between 3 and 24 h pertained to siderophore biosynthesis, emphasizing the importance of iron scavenging for the maintenance of intracellular macrophage infection for that strain. No consistent functions in the genes modulated during late infection were apparent for either strain during CCL-185 infection, suggesting less dramatic modulation of bacterial physiology in that host cell.

To understand host transcriptional response to S. aureus infection we conducted similar analyses with human cell line data. Unsurprisingly, there were marked differences in gene expression between host cell types following S. aureus infection: macrophages underwent more dramatic changes in gene expression during infection than did epithelial cells, consistent with the well-documented macrophage transcriptional activation programs that respond to bacteria [72], whereas changes in epithelial cells were far more muted. We recorded a number of changes in host cell regulation that were specific to particular strain of infecting S. aureus and that likely reflect underlying differences in strain pathophysiology and baseline virulence [27]. Such strain-specific host cell responses were most prominent in THP-1 derived macrophages, potentially reflecting differences in the relative virulence of those strains and their proportional impact on subsequent macrophage physiology and activation. During early infection ATCC29213 induced more genes associated with the classical immune response than JE2, and also caused significant shifts in gene regulation 24 h after infection that were absent after JE2 exposure. JE2-specific changes in that cell line largely involved chromatin remodeling, consistent with stimulus-dependent macrophage activation pathways [73] that were not induced by ATCC29213. One interpretation of these findings would be that JE2 has evolved mechanisms to suppress or avoid particular immune responses that facilitate its carriage and transmission among individuals [42, 43], whereas ATCC29213 evokes an unrestrained macrophage response. Exploration of such differences in host cell response and functional characterization of host cell DEGs relevant to S. aureus infection, while outside the scope of the present work, will be valuable for future studies.

We generated CRISPRi knockdowns of the most statistically significant bacterial DEGs identified by Dual RNA-seq for each of the specific pairings of S. aureus strain and host cell type to experimentally assess their influence on intracellular pathogenesis. Most knockdowns (81%) significantly altered measurements of intracellular pathogenesis when tested in the cell line where the gene was initially identified, suggesting their involvement in the infection process. Interestingly, the phenotypes of CRISPRi mutants did not consistently match predictions from gene expression analysis in multiple instances: silencing genes that were upregulated during infection could yield improved intracellular pathogenesis phenotypes, and conversely, silencing downregulated genes sometimes impaired S. aureus intracellular invasion. Such discrepancies are consistent with our prior studies of intracellular pathogenesis genes identified by transposon sequencing (Tn-seq) [74], and indicate that perturbing relevant genes individually does not always recapitulate phenotypes effected by larger networks of coordinated gene expression changes. It is accordingly thought that multiple, epistatic interactions underlie many complex S. aureus phenotypes [75, 76]. Even so, several bacterial genes that were highly expressed during invasion reduced that phenotype when knocked down, fitting an intuitive model of their being required for intracellular pathogenesis. Such factors notably included vwb [77], oatA [69], and virulence regulator secY [78], as well as genes involved in purine (purA) [79] and pyrimidine (pyrF) [80] biosynthesis.

While providing valuable data regarding the changes in host and bacterial gene expression occurring during intracellular pathogenesis, our study has limitations. First, cultured cell lines do not have identical properties to their analogous tissues in vivo [28], and moreover, tumor-derived host cells used here may differ from primary cells in multiple respects [81]. Thus, although our experimental design allowed for well controlled experiments and provided sufficient input material for robust data generation, our reliance on in vitro systems introduces differences from natural infection states. Second, it is known that S. aureus can infect a wide variety of host cells [13, 20, 24–27] and that individual strains can exhibit marked differences in their capacities for intracellular pathogenesis [27]. Our study necessarily considers only a subset of those combinations. Further work characterizing a larger and more diverse matrix of both S. aureus strains and host cells will be necessary to fully explore the range of transcriptional responses occurring during intracellular pathogenesis. Third, the high MOI necessary for infecting host cells with sufficient bacteria for robust detection of pathogen transcripts likely generate non-physiological conditions, potentially introducing artifacts that are absent from natural infection states. It is therefore possible that transcriptional signatures cataloged here reflect, at least in part, acute stress responses rather than stable intracellular adaptation programs and attendant host cell responses. Fourth, some of the differences identified by Dual RNA-Seq may reflect the effects of differing growth states or density of bacteria under the different conditions examined, and not explicitly reflect changes attributable to the infection process. Fifth, studies involving sequencing of RNA obtained from any cell populations necessarily represents an aggregation of results over its entirety, and we are therefore unable to identify subpopulations that may have markedly different phenotypic or transcriptomic properties [82, 83]. Last, we found that most transcriptional changes occur rapidly after bacterial invasion of host cell lines, especially in the infecting S. aureus. Studies examining transcriptional changes at multiple, earlier time intervals could therefore provide information about the temporal progression of such changes as infection initially becomes established. Conversely, examining timepoints at later stages of infection, as organisms establish durable non-replicative intracellular states could lead to greater understanding of S. aureus persister formation [82, 83].

Conclusions

In summary, this study provides a detailed view of the changes in bacterial and host cell gene expression that occur during S. aureus intracellular pathogenesis of host cells that are relevant to chronic human infection. In both bacteria and host cells, the vast majority of transcriptional reprogramming occurred rapidly (within 3 h). For both bacteria and host, our work illustrates a combination of genes undergoing consistent shifts in transcriptional response during infection and DEGs that are specific to particular conditions. These analyses provide a framework for defining patterns of expression that underly specific host-pathogen interactions, and highlight conserved bacterial elements that could potentially be targeted by novel therapeutic strategies to disrupt intracellular infection. For example, disruption of bacterial metal ion homeostasis using systemic gallium treatment has shown promise in treating chronic lung Pseudomonas aeruginosa infections [84], and may show similar utility in treating persistent S. aureus populations, given the consistent overexpression of such pathways during intracellular pathogenesis.

Authors' contributions

EAH conceived the study. EAH, JDL, KP performed the experiments and acquired data. AW, DRL, SJS analyzed and interpreted the data. DRL and SJS wrote the manuscript with input from all authors. All authors read and approved the final manuscript.

Funding

This work was supported by grants from the Cystic Fibrosis Foundation (SINGH24R0 to SJS, https://www.cff.org/), and NIH (NIDDK P30 DK089507 to SJS, NIAMS K23 AR080209 to DRL, https://www.nih.gov/). The funders of this work did not play any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data availability

Sequence data generated in support of this study are available from the NCBI Sequence Read Archive under accession PRJNA1292219. Mapped read counts for transcripts under specific experimental conditions are provided for bacteria (Supplementary Tables 3, 6, 9, 10, 13, 14, 15, and 16) and human (Supplementary Table 17, 18, 20, 21) analyses.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.