The adherens junctions control susceptibility to Staphylococcus aureus α-toxin

Significance

Staphylococcus aureus is a major cause of invasive bacterial infection. One prominent virulence factor is α-toxin, a protein that injures the cell by forming a damaging pore across the cell membrane. We conducted a genetic screen to identify host factors that control susceptibility to α-toxin. We discovered that several components of the adherens junction complex modulate α-toxin cytotoxicity. By eliminating expression of the junctional protein plekstrin-homology domain containing protein 7 (PLEKHA7), cells gained the ability to recover from α-toxin injury and mice lacking PLEKHA7 exhibited improved healing from S. aureus skin infection and enhanced survival of pneumonia. Our data suggest that targeting nonessential host epithelial junction components can reduce S. aureus morbidity by enhancing cellular resilience to α-toxin injury.

Abstract

Staphylococcus aureus is both a transient skin colonizer and a formidable human pathogen, ranking among the leading causes of skin and soft tissue infections as well as severe pneumonia. The secreted bacterial α-toxin is essential for S. aureus virulence in these epithelial diseases. To discover host cellular factors required for α-toxin cytotoxicity, we conducted a genetic screen using mutagenized haploid human cells. Our screen identified a cytoplasmic member of the adherens junctions, plekstrin-homology domain containing protein 7 (PLEKHA7), as the second most significantly enriched gene after the known α-toxin receptor, a disintegrin and metalloprotease 10 (ADAM10). Here we report a new, unexpected role for PLEKHA7 and several components of cellular adherens junctions in controlling susceptibility to S. aureus α-toxin. We find that despite being injured by α-toxin pore formation, PLEKHA7 knockout cells recover after intoxication. By infecting PLEKHA7^−/− mice with methicillin-resistant S. aureus USA300 LAC strain, we demonstrate that this junctional protein controls disease severity in both skin infection and lethal S. aureus pneumonia. Our results suggest that adherens junctions actively control cellular responses to a potent pore-forming bacterial toxin and identify PLEKHA7 as a potential nonessential host target to reduce S. aureus virulence during epithelial infections.

The bacterium Staphylococcus aureus is not only one of the most important human pathogens resulting in considerable morbidity and mortality (1, 2) but also can be found as a transient skin resident, intermittently colonizing a sizable portion of the healthy population (3). S. aureus infections manifest in a diverse array of clinical presentations, but related to its transitory epithelial niche, S. aureus predominantly results in skin and soft tissue infections (4, 5). Through local infections bacteria can gain access to deeper tissue and disseminate hematogenously to cause invasive disease such as endocarditis, osteomyelitis, deep tissue abscesses, sepsis, and pneumonia (1). In the face of increasing antibiotic resistance, the widespread prevalence of methicillin-resistant S. aureus (MRSA) strains both in hospitals and communities across the globe presents a growing threat to human health worldwide (5, 6). Given the growing difficulty of treating these common and frequently life-threatening infections, understanding host–pathogen interactions that mediate S. aureus pathogenesis is imperative.

Chief among the arsenal of S. aureus virulence factors, α-toxin (or α-hemolysin) is a critical determinant for pathogenesis in a wide variety of experimental infections, particularly during epithelial infections such as skin abscesses and pneumonia (7–10). After secretion as a soluble monomer, α-toxin oligomerizes on the targeted host cell surface via interactions with its high-affinity metalloprotease receptor, a disintegrin and metalloprotease 10 (ADAM10), forming a 1–3-nm pore that spans the cellular membrane lipid bilayer (11, 12). Originally described solely for its ability to induce lysis of erythrocytes, it is now appreciated that α-toxin exerts pleiotropic effects on a diverse set of host cells (13). In addition to inducing cell death, at sublytic concentrations α-toxin has been described to alter a wide variety of cellular processes, including cell signaling, proliferation, immunomodulation, autophagy, and others (13–17).

Importantly, S. aureus uses α-toxin to remodel host epithelia and alter tissue integrity. Engagement of α-toxin with ADAM10 leads to intracellular ion flux across the toxin pore, which enhances the proteolytic activity of ADAM10 through an unknown mechanism (18). ADAM10 is essential for tissue morphogenesis and remodeling and acts on a multitude of extracellular substrates (19), one of which is the adherens junction protein E-cadherin (20). It has been proposed that α-toxin–enhanced ADAM10 cleavage of E-cadherin dismantles the adherens junctions to disrupt the integrity of cell–cell contacts in epithelial tissues during infection to contribute to S. aureus pathogenesis (18, 21). However, the molecular components that govern intracellular responses elicited by α-toxin in the targeted host cell remain largely undefined.

To advance understanding of how S. aureus α-toxin modulates host cell biology, we conducted a high-throughput genetic screen using human cells (22, 23) to discover novel host factors required for α-toxin cytotoxicity. Our screen unexpectedly revealed that multiple components of the cellular adherens junctions modulate susceptibility to α-toxin, suggesting a previously unidentified role for the junctions as critical mediators of α-toxin cytotoxicity—not merely its target. The most significant hit following ADAM10 was plekstrin-homology domain containing protein 7 (PLEKHA7), a cytoplasmic accessory member of the adherens junction complex (24). In PLEKHA7-deficient cells, α-toxin pore formation occurs, but remarkably cells exhibit enhanced recovery from α-toxin injury. Furthermore, we establish the important contribution of PLEKHA7 for S. aureus pathogenesis in vivo using MRSA USA300 LAC mouse models of both a self-resolving skin and soft tissue infection and lethal pneumonia.

Results

Human Haploid Genetic Screen Reveals Novel Host Factors Required for α-Toxin Cytotoxicity.

To discover host factors required for α-toxin cytotoxicity, we conducted a live/dead genetic screen by intoxicating haploid human cells (HAP1) carrying knockout alleles in essentially all genes through insertional mutagenesis (22, 23). A large-scale library of ∼100 million HAP1 mutagenized cells was treated with recombinant α-toxin, and gene-trap insertion sites were identified in the pool of surviving, toxin-resistant cells. The gene trap insertion sites in the toxin-selected population were compared with a control library that was not exposed to α-toxin (Dataset S1). We identified 46 genes that were significantly enriched in the toxin-treated library (P < 0.05) spanning several functional biological categories, suggesting that these genes are essential for α-toxin intoxication (Fig. 1A and Dataset S1).

Fig. 1.

A haploid genetic screen for cellular host factors required for α-toxin cytotoxicity identifies several components of the adherens junctions. (A) A haploid human cell genetic screen identified genes that confer resistance to α-toxin when inactivated by retroviral gene-trap insertion. Each bubble represents a gene, and bubble size corresponds to the number of independent gene-trap insertion events observed in the toxin-selected population. Genes are ranked on the y axis according to the significance of enrichment of gene-trap insertions in the α-toxin selected library compared with the nonselected control population (Fisher’s exact test, corrected for false discovery rate). Functional groups of significant hits share a common color as indicated and are grouped together horizontally (with other genes in random order along the x axis). Gene names are displayed until P = 0.05. (B) An abridged diagram of the adherens junctional complex. Cell–cell adherens junction proteins found significantly enriched in our α-toxin screen are depicted in red, and associated proteins are shown in gray. (C) Viability of WT HAP1 cells or knockout (Δ) HAP1 clones for the indicated genes following treatment with α-toxin (indicated as percentage of nonintoxicated controls; data are mean + SEM; n = 3 biological replicates; all differences relative to WT cells are P = <0.01, unpaired t test).

The validity of our screen was confirmed by identification of ADAM10, the α-toxin cell surface metalloprotease receptor (12), which emerged as the most enriched gene in the toxin-selected library (Fig. 1A, P = 7.36 × 10⁻¹⁰⁰). Moreover, the screen identified the chaperone protein TSPAN33 (25), which reportedly mediates surface localization of ADAM10 (Fig. 1A, P = 1.09 × 10⁻²⁵). Lipid biogenesis genes, such as sterol response element binding proteins, were also enriched (Fig. 1A) and have been attributed roles in cellular innate immune responses against pore-forming toxins (26, 27).

Our screen discovered several genes not previously implicated in α-toxin cytotoxicity. PLEKHA7, the second most enriched gene in our screen (Fig. 1A, P = 1.84 × 10⁻⁸¹), is a cytoplasmic, accessory component of adherens junctions (24). PLEKHA7 localizes to the zonula adhaerens belt of the apical–junctional complex in epithelial cells, where it directly interacts with p120 catenin, paracingulin, and afadin (24, 28–31). Unlike E-cadherin or p120 catenin, PLEKHA7 is not required for adherens junctions formation, and previous studies suggest a role for PLEKHA7 in regulating the stability of the adherens junctions (24, 28–31). In accord with adherens junctions mediating α-toxin injury, our screen also identified several other junctional components including α-catenin, afadin, and N-cadherin (Fig. 1 A and B). Enrichment of inactivating insertions in multiple adherens junctions complex genes in the toxin-resistant library suggested the novel hypothesis that components of the cellular adherens junctions can control susceptibility to S. aureus α-toxin. To test this hypothesis, we used CRISPR/Cas9 gene editing to generate HAP1 knockout (Δ) subclone cell lines for multiple adherens junction genes revealed by our screen (Fig. S1). In accord with our screen results, we find that cells individually lacking the junctional proteins PLEKHA7, afadin, α-catenin, and N-cadherin are all significantly less susceptible to α-toxin cytotoxicity than WT controls (Fig. 1C). Given that PLEKHA7 was the most important junctional gene modulating susceptibility to α-toxin (Fig. 1C), we explored its role further.

Fig. S1.

Validation of HAP1 knockout subclone cell lines generated by CRISPR/Cas9 gene editing. (A) CRISPR/Cas9 target guide sites in HAP1 ΔADAM10, ΔPLEKHA7, Δα-catenin, ΔAfadin, and ΔN-cadherin HAP1 subclones were sequenced and compared with WT predicted sequence. CRISPR guide target sequences are indicated in bold, underlined text. Insertions and/or deletions caused by the observed editing event(s) are noted. (B) Western blot analysis of whole-cell lysates from WT, ΔADAM10, ΔPLEKHA7, Δα-catenin, ΔAfadin, and ΔN-cadherin HAP1 subclones as shown in A probed with the indicated antibodies.

The Adherens Junction Component PLEKHA7 Controls Susceptibility to α-Toxin.

HAP1 cells lacking PLEKHA7 are markedly more resistant to α-toxin cytotoxicity than WT cells at a wide range of toxin concentrations (Fig. 2A). Furthermore, stable expression of a human PLEKHA7-FLAG construct in trans fully complements α-toxin cytotoxicity in ΔPLEKHA7 cells (Fig. 2A). To determine if PLEKHA7 deletion reduces susceptibility in the context of endogenously expressed levels of α-toxin, we treated WT and ΔPLEKHA7 cells with cell-free supernatants from MRSA strain USA300 LAC and an isogenic mutant MRSA strain lacking the α-toxin gene, USA300 LAC hla::ermB. This treatment confirmed that ΔPLEKHA7 cells are also more resistant to endogenous α-toxin (Fig. S2). We find that PLEKHA7 deletion does not confer a generalized injury resistance, as ΔPLEKHA7 cells are not more resistant to cytotoxicity caused by either bacterial pore-forming toxins (streptolysin O, perfringolysin O) or the potassium ionophore nigericin (Fig. S3). PLEKHA7’s role in modulating α-toxin susceptibility on well-differentiated epithelial cells was further confirmed by testing the effect of α-toxin on a ΔPLEKHA7 cell line in the Madin–Darby Kidney (MDCK) background. MDCK cells are a widely used in vitro model of a simple epithelium and form a polarized monolayer when grown on transwell filters (32). Consistent with our observations made using HAP1 cells, we find that ΔPLEKHA7 MDCK epithelial monolayers are more resistant to α-toxin cytotoxicity than WT MDCK monolayers (Fig. S4).

Fig. 2.

The adherens junction protein PLEKHA7 controls susceptibility to α-toxin cytotoxicity despite pore formation and cellular injury. (A) WT HAP1 cells, knockout (Δ) HAP1 clones, and ΔPLEKHA7 cells stably expressing human PLEKHA7-FLAG were treated with indicated α-toxin concentrations or media only, then subsequently stained with crystal violet (Left), or viability was quantified after 24 h of α-toxin treatment (percentage viability shown relative to cell type-specific media controls; data are mean ± SEM; n = 3 biological replicates). (B) Diagram of deletion human PLEKHA7-FLAG constructs stably expressed by lentiviral transduction in ΔPLEKHA7 HAP1 cells. (C) ΔPLEKHA7 HAP1 cells stably expressing the indicated constructs in B were treated with α-toxin and 18 h later stained using a fluorescence-based LIVE/DEAD assay, where green fluorescence indicates live cells by esterase activity and red fluorescence demonstrates loss of plasma membrane integrity. (Scale bar, 70 µm.) (D) Western blot analysis of ADAM10 and GAPDH expression in whole-cell lysates from WT, ΔADAM10, and ΔPLEKHA7 HAP1 cells. (E) Cell surface expression of ADAM10 on WT, ΔADAM10, and ΔPLEKHA7 HAP1 cells as measured by flow cytometric analysis. Data are representative of three independent experiments. (F) Change in [K+] in extracellular media following α-toxin treatment of WT, ΔADAM10, and ΔPLEKHA7 HAP1 cells (shown relative to cell type-specific nonintoxicated controls; data are mean ± SEM; n = 3 biological replicates; comparison of WT and ΔPLEKHA7, P = 0.069, unpaired t test). (G) Western blot analysis of p38 and phosphorylated p38 in whole-cell lysates from left (nonintoxicated) and right (α-toxin treated) WT, ΔADAM10, and ΔPLEKHA7 HAP1 cells. Data are representative of three independent experiments. (H) Time course of intracellular [ATP] in WT, ΔADAM10, and ΔPLEKHA7 HAP1 cells following α-toxin treatment (shown as percentage of cell type-specific nonintoxicated controls; data are mean ± SEM; n = 3 biological replicates).

Fig. S2.

ΔPLEKHA7 cells are more resistant to MRSA bacterial supernatants in an α-toxin–dependent manner. (A) WT, ΔADAM10, ΔPLEKHA7, and ΔPLEKHA7 HAP1 cells stably expressing human PLEKHA7-FLAG cell lines were incubated overnight with stationary phase bacterial cell-free supernatants derived from equivalent density cultures of WT USA300 LAC or the USA300 LAC hla::ermB α-toxin isogenic mutant. Two days after treatment, viable adherent cells were fixed and monolayers visualized by crystal violet stain. (B) Secreted proteins from stationary phase bacterial cell-free supernatants applied to cells as shown in A were concentrated by TCA precipitation, separated by SDS/PAGE, and total protein visualized by Coomassie blue staining to confirm similar exoprotein profiles for the two strains. The absence of a major band in the hla::ermB USA3000 LAC isogenic mutant strain at the predicted size of α-toxin is denoted by an asterisk.

Fig. S3.

PLEKHA7 deletion does not confer resistance to other bacterial pore-forming toxins or monovalent cation ionophores. (A–C) WT and ΔPLEKHA7 HAP1 cells were treated for 24 h with media as a control or the indicated concentrations of (A) Streptolysin O, (B) Perfringolysin O, and (C) nigericin. Two days after treatment, surviving cells were fixed and monolayers visualized by crystal violet stain.

Fig. S4.

PLEKHA7 modulates susceptibility to α-toxin cytotoxicity in a polarized, epithelial monolayer. (A) Sequencing reads from the CRISPR guide target site in ΔPLEKHA7 diploid canine MDCK subclone reveals inactivating CRISPR/Cas9 genome editing events in both PLEKHA7 alleles. CRISPR guide target site sequence is indicated in bold, underlined text. (B) Western blot analysis of PLEKHA7 expression in whole-cell lysates from WT and ΔPLEKHA7 MDCK cells. (C) Number of SYTOX-positive (dying or dead, membrane-compromised) nuclei observed per field of view following α-toxin treatment of WT MDCK polarized epithelial monolayers or ΔPLEKHA7 MDCK polarized epithelial monolayers. Data are mean ± SEM; n = 13 fields of view. Data shown are representative of three biological replicates. (D) Confocal microscopy images of WT (Upper Left) or ΔPLEKHA7 (Upper Right) MDCK epithelial monolayers treated with α-toxin. Filamentous actin, red; SYTOX-positive cells, green; nuclei stained with DAPI, blue. The Lower panel depicts a confocal microscopy 3D image reconstruction of a dying, extruding SYTOX-positive nucleus in a WT MDCK monolayer following α-toxin treatment. (Scale bars, 10 μm.)

Next, using PLEKHA7 deletion mutant constructs stably expressed in ΔPLEKHA7 cells (Fig. S5) and subsequently treated with α-toxin, we defined the PLEKHA7 residues necessary and sufficient for restoring α-toxin cytotoxicity to be restricted to the first 538 N-terminal amino acids (Fig. 2 B and C). This region of PLEKHA7 encompasses residues that mediate its interaction with afadin (29)—another junctional protein enriched in our screen (Fig. 1A) that modulates α-toxin susceptibility (Fig. 1C).

Fig. S5.

Relative protein expression of PLEKHA7-FLAG truncation constructs. Western blot analysis of whole-cell lysates from PLEKHA7-FLAG constructs stably expressed by lentiviral transduction in ΔPLEKHA7 HAP1 cells, as shown in Fig. 2 B and C, probed with the indicated antibodies. Data are representative of three independent experiments.

PLEKHA7-Deficient Cells Exhibit Enhanced Resilience to α-Toxin Injury.

To interrogate the mechanism underlying PLEKHA7 modulation of α-toxin cytotoxicity, we first assessed whether PLEKHA7 deletion alters expression of the toxin receptor ADAM10. Western blot of whole-cell lysates demonstrates that ΔPLEKHA7 cells express comparable levels of ADAM10 as WT controls (Fig. 2D). To determine whether ADAM10 is localized at the plasma membrane in the absence of PLEKHA7, we quantified surface-available ADAM10 by flow cytometry and found no significant decrease in cell surface ADAM10 expression between WT and ΔPLEKHA7 cells (Fig. 2E). Consistent with these findings, time-lapse video microscopy reveals that WT and ΔPLEKHA7 cells, but not ΔADAM10 cells, develop cytopathic effects following intoxication. However, we observe that individual ΔPLEKHA7 cells recover from this initial injury and survive intoxication, in contrast to WT cells (Movie S1).

We next sought to determine whether ΔPLEKHA7 cells are more resistant to α-toxin cytotoxicity because of a defect in α-toxin pore formation spanning the targeted host cell surface. To do so, we assessed functional outcomes of pore formation and subsequent cellular injury in α-toxin–treated ΔPLEKHA7 cells using several distinct assays. Rapid efflux of intracellular potassium is an early effect of α-toxin pore formation (33). We observe an increase in extracellular potassium relative to nonintoxicated controls following intoxication of WT and ΔPLEKHA7 cells, but not ΔADAM10 cells (Fig. 2F). Another functional consequence of α-toxin injury known to be dependent on pore formation is the activating phosphorylation of the cellular stress response kinase p38 (34). Consistent with cellular injury occurring in the absence of PLEKHA7, we observe α-toxin–dependent p38 activation in WT and ΔPLEKHA7 cells but not ΔADAM10 cells (Fig. 2G). Because intracellular ATP depletion is a hallmark of α-toxin damage and ATP repletion is associated with enhanced recovery from α-toxin injury (16, 35), we quantified changes in intracellular ATP levels following intoxication. Shortly after α-toxin treatment, both WT and ΔPLEKHA7 cells quickly deplete intracellular ATP. At later time points, however, only ΔPLEKHA7 cells restore intracellular ATP (Fig. 2H). From these studies, we conclude that PLEKHA7 controls susceptibility to α-toxin in a step downstream of ADAM10 recognition and α-toxin pore formation. PLEKHA7 deletion does not strictly prevent damage caused by α-toxin, but rather cells lacking PLEKHA7 exhibit enhanced resilience to and recovery from α-toxin injury.

PLEKHA7 Contributes to the Severity of MRSA Skin and Pneumonia Infections in Vivo.

Given α-toxin’s critical role in S. aureus pathogenesis during skin and lung infections and the expression of PLEKHA7 in epithelial tissues at the adherens junctions, we hypothesized that PLEKHA7 may contribute to S. aureus pathogenesis during an in vivo infection. To test the role of PLEKHA7 during MRSA infection, we made use of previously unpublished PLEKHA7^−/− whole-body transgenic mice (Fig. S6). Consistent with a recently described PLEKHA7-deficient rat (36), PLEKHA7^−/− mice are viable and fecund, exhibiting no gross developmental defects.

Fig. S6.

Generation and validation of PLEKHA7^−/− mice. (A) Diagram of the targeting vector used to generate PLEKHA7^−/− mice. Genotyping primer binding sites are indicated in red. (B) Genotyping of PLEKHA7 ^+/+ (WT), PLEKHA7^−/−, and PLEKHA7^−/+ heterozygous mice using a WT reverse primer, and either a WT forward primer or a forward primer complementary to the neomycin cassette, generating a 200 bp larger product. (C) Confocal microscopy images of PLEKHA7 expression in murine gastric epithelium of WT (Left panels) or PLEKHA7^−/− (Right panels) animals. Filamentous actin, red; PLEKHA7, green; nuclei stained with DAPI, blue. (Scale bars, 25, 10, and 5 μm for the Top, Middle, and Lower panels, respectively.) (D) Photograph of 6-wk-old PLEKHA7^−/− female mice. PLEKHA7^−/− animals are healthy and fecund, with no gross developmental phenotypes. (E) Weight (in grams) of 6–8-wk-old WT and PLEKHA7^−/− matched littermates; data are mean ± SEM; n ≥ 13 animals in each group.

We first examined the contribution of PLEKHA7 during a self-limiting MRSA skin and soft tissue infection. In this model, S. aureus is superficially introduced into the ear pinnae using a shallow needle, resulting in a necrotic, inflammatory lesion that resolves with tissue loss (37). Confirming the importance of α-toxin for this infection, USA300 LAC hla::ermB isogenic mutant infections result in significantly decreased lesion size compared with WT USA300 LAC (Fig. S7). To assess the contribution of PLEKHA7 to disease during a skin and soft tissue infection, we infected WT and PLEKHA7^−/− mice with WT USA300 LAC and followed disease development in individual animals over time. We observe that both WT and PLEKHA7^−/− mice developed similar-sized necrotic lesions at early time points (Fig. 3 A and B) indistinguishable by histopathology and bacterial burden (Figs. S8 and S9). However, despite an initial similarity between lesions in WT and PLEKHA7^−/− mice, at day 14 postinfection, we found that PLEKHA7^−/− mice had resolved the lesions with significantly less tissue loss than WT controls (Fig. 3B and Movie S2). Paralleling our in vitro observations at the cellular level, these data support an enhanced recovery phenotype for PLEKHA7^−/− mice during a superficial MRSA skin infection.

Fig. S7.

α-toxin contributes to the severity of a self-resolving MRSA skin and soft tissue infection. (A) Western blot analysis of α-toxin expression in cell-free bacterial supernatant lysates from strains WT USA300 LAC, hla::ermB USA300 LAC, and hla::ermB-phla USA300 LAC complemented with α-toxin. (B) Representative images of WT USA300 LAC and hla::ermB USA300 LAC ear skin and soft tissue lesions within individual animals at day 14. (Scale bar, 5 mm for all panels.) (C) Bacterial density measurements from WT animals infected with WT USA300 LAC or hla::ermB USA300 LAC at the indicated time points. n = 6 animals per group, per time point. Data are mean ± SEM. (D) Lesion size (in mm²) at day 14 in mice infected with WT or hla::ermB USA300 LAC. Data are mean ± SEM, representing n = 5 animals in each group, representative of two independent experiments.

Fig. 3.

PLEKHA7 contributes to the severity of MRSA skin and pneumonia infections in vivo. (A) Images of ear skin and soft tissue infection progression within individual animals over time from representative WT and PLEKHA7^−/− mice following infection with MRSA strain USA300 LAC. (Scale bar, 5 mm for all panels.) (B) Lesion size (mm²) in WT and PLEKHA7^−/− mice at day 5 and day 14 in superficial ear infection with USA300 LAC. Data are mean ± SEM from two independent experiments, representing n = 15 animals in each group. P = 0.0116, unpaired t test. (C) Core body temperature of USA300 LAC-infected WT and PLEKHA7^−/− mice as shown in D at 6 h and 24 h postinfection (P = 0.0047, unpaired t test). At 6 h, n = 14 animals in each group; at 24 h, n = 11 WT and n = 14 PLEKHA7^−/− animals. (D) Survival analysis of PLEKHA7^−/− mice and WT controls after infection with S. aureus USA300 LAC. n = 14 animals in each group; P = 0.0037, log-rank test. (E) USA300 LAC bacterial density measurements from infected lungs of WT and PLEKHA7^−/− mice at 6 h (n = 6 animals per group) and 24 h postinfection (n = 6 WT animals and 10 PLEKHA7^−/−; data are mean ± SEM). (F) Survival analysis of PLEKHA7^−/− mice and WT controls after infection with a threefold higher inoculum of S. aureus USA300 LAC than in D. n = 23 animals in each group; P = 0.0247, log-rank test.

Fig. S8.

USA300 LAC bacterial burden in self-resolving MRSA skin and soft tissue infections of WT and PLEKHA7^−/− mice. (A) USA300 LAC bacterial density measurements from infected ear tissue of WT and PLEKHA7^−/− animals at the indicated time points. At 6 h postinfection, n = 3 animals per group; at day 1 and day 3 postinfection, n = 6 animals per group. Data are mean ± SEM. (B) Lesion size (in mm²) at day 14 postinfection of WT and PLEKHA7^−/− animals in superficial ear infection model with USA300 LAC. Data are mean ± SEM pooled from two independent experiments, representing n = 17 WT and n = 18 PLEKHA7^−/− matched littermates. P < 0.001, unpaired t test.

Fig. S9.

Histopathology of MRSA skin infection in WT and PLEKHA7^−/− mice. (A) Representative images of WT and PLEKHA7^−/− animal MRSA ear skin and soft tissue lesions taken at day 3 (A and B) and day 5 (C and D) postinfection. Histopathology studies of USA300 LAC skin lesions reveal full-thickness lesions filled with serocellular material and a massive influx of degenerate neutrophils (A–D, Insets). Ear tissue edges adjacent to the necrotic, eosinophilic coagulum display moderate re-epithelialization with mild to moderate epidermal acanthosis and a moderate population of viable neutrophils. Lesions were uniformly indistinguishable when examined by a veterinary pathologist (D.M.B.) who was blinded to the experimental groups (n = 6 ears per time point per group).

Upon observing that a self-limiting MRSA skin infection in PLEKHA7^−/− mice results in significantly reduced pathology, we sought to assess the importance of PLEKHA7 during an acute, lethal MRSA pneumonia (7, 8). Although both groups developed comparable hypothermia following infection, only PLEKHA7^−/− animals showed an enhanced, significant recovery of core temperature by 24 h postinfection, and more animals fully recovered without antibiotic treatment (Fig. 3 C and D). We find that PLEKHA7^−/− mice survived USA300 LAC pneumonia significantly better than WT controls (Fig. 3D). To explore the strength of this phenotype, we challenged mice with a threefold higher inoculum and measured similar bacterial loads in lung tissue homogenates at 6 h and 24 h postinfection in both groups (Fig. 3E). Despite this higher inoculum, PLEKHA7^−/− mice exhibit an increased mean time to death relative to WT control animals (Fig. 3F), highlighting the contribution of PLEKHA7 to MRSA pneumonia virulence.

Discussion

S. aureus can transiently and asymptomatically colonize the human skin epithelium and also cause significant morbidity and mortality, predominantly through skin and soft tissue infections that can progress to dangerous systemic disease. Due to the critical importance of S. aureus α-toxin for both pathogenesis and its interface with host epithelia, we conducted a genetic screen to identify novel host mediators of α-toxin cytotoxicity. In addition to the toxin receptor ADAM10, our screen identified the intracellular junctional protein PLEKHA7 and several other members of the epithelial adherens junction complex as host factors that modulate α-toxin injury.

Our data suggest a previously unidentified biological role for the adherens junctions in controlling cellular injury caused by a potent bacterial cytotoxin. It is well established that many diverse pathogens and their virulence factors have evolved to target the cellular junctions to facilitate attachment, entry, and invasion across tissue barriers (38–40). Indeed, α-toxin has been shown to alter epithelial barrier integrity by enhancing ADAM10 cleavage of the ectodomain of E-cadherin (18, 21). It is increasingly appreciated, however, that microbial interactions with the host epithelium and cell–cell junctions are not restricted to mechanical disruption of barrier function (39, 41). Mounting evidence illustrates that bacteria can actively modulate other important aspects of junction function, such as altering cell polarity (42–44) and intracellular signals emanating from the junctions (45–47). Our findings build importantly on these observations, revealing not only that S. aureus α-toxin does target the junctions but also that α-toxin acts through adherens junctions components to mediate its cytotoxicity. Considering that S. aureus is an epithelial colonizer, our data suggest the likely existence of more subtle biological interplay between S. aureus and the host adherens junctions at sublytic concentrations of α-toxin to facilitate bacterial modification of its replicative niche during colonization.

Our in vitro studies demonstrate that PLEKHA7 modulates susceptibility to α-toxin in both a human cell line and an in vitro model of a simple, polarized epithelium. Surprisingly, further mechanistic investigations revealed that PLEKHA7 controls susceptibility to α-toxin downstream of functional pore formation, suggesting that pore formation by itself is not sufficient to cause cell death. Rather, we determined that PLEKHA7-deficient cells are more resilient than WT cells and better recover from injury caused by α-toxin, ultimately exhibiting enhanced survival from intoxication. This conclusion is supported by time-lapse video microscopy revealing individual cells recovering from intoxication, as well as population-level assays quantifying cellular viability and the kinetics of intracellular ATP repletion in α-toxin–treated PLEKHA7-deficient cells.

We speculate that adherens junctions may regulate cytotoxicity through controlling resolution of pores and cellular membrane repair, or alternatively may act to transmit prodeath intracellular signals or localize injury caused by pore-forming toxins. PLEKHA7 is known to link microtubules to the adherens junctions and regulate stability of the junctions (24, 30), which may serve to coordinate these hypothesized functions. Our data support a new biological role for intracellular components of the adherens junctions in regulating cellular injury in response to α-toxin, a paradigm that warrants future investigation.

The relevance of PLEKHA7 for determining the outcome of in vivo MRSA bacterial infections was demonstrated in two relevant infection models, a self-resolving skin infection (37) and a lethal pneumonia (7, 8). Some canonical adherens junctions proteins such as E-cadherin and p120 catenin are essential for junction formation, and systemic knockouts are embryonic lethal (48, 49). In contrast, our previously unpublished PLEKHA7^−/− mice and a recently described PLEKHA7^−/− rat model (36) are healthy and fecund, exhibiting no gross developmental or epithelial pathology. From this we infer that PLEKHA7 is not an essential junctional protein in vivo but rather may serve to regulate some previously unidentified aspect of junction function under specific conditions. We find that systemic PLEKHA7 deletion in vivo attenuates the pathogenicity of the clinically relevant MRSA USA300 LAC strain in mouse models of both a self-resolving skin infection as well as a lethal pneumonia. In both epithelial infection models, we observe an initial similarity in pathology between the WT- and PLEKHA7^−/−-infected animals; however, at later time points, PLEKHA7^−/−-infected animals recover better than WT-infected controls. These results suggest that targeting nonessential components of the host adherens junctions could potentially reduce MRSA morbidity by enhancing resilience to and recovery from α-toxin injury. The increasing prevalence of drug-resistant MRSA strains underscores the urgent need to develop host cellular targets of S. aureus virulence, which may have future utility as adjunctive therapy.

Materials and Methods

Haploid Human Cell Genetic Screen.

HAP1 cells were mutagenized with a retroviral gene trap to cause inactivating mutations throughout the genome, and a haploid genetic screen was performed as previously described (22, 23). For a complete description of the haploid genetic screen, see SI Materials and Methods.

Genome Engineering and PLEKHA7 Cloning.

Clustered regularly interspaced short palindromic repeats (CRISPR) sequence-targeting sequences were designed using the Zhang Lab CRISPR design tool (crispr.mit.edu), and oligos corresponding to the guide RNA sequences were directly cloned into the Zhang laboratory-generated Cas9-expressing plasmid px458 using the Gibson Assembly Reaction (NEB). A complete description of HAP1 and MDCK genome engineering, guide target sequence oligos, and PLEKHA7 construct cloning is presented in SI Materials and Methods.

Bacterial Strains and Culture.

The MRSA strain USA300 LAC was kindly provided by Fabio Bagnoli. A detailed description of the generation of the α-toxin isogenic mutant strain hla::ermB and the complemented hla::ermB-phla is provided in SI Materials and Methods. Bacteria were grown in tryptic soy broth at 37° and prepared as indicated for animal infections.

Generation and Validation of PLEKHA7^−/− Transgenic Mice.

A detailed description of the generation and validation of Plekha7(LacZ) mutant mice is provided in SI Materials and Methods.

Murine S. aureus Superficial Skin and Pneumonia Infection Models.

Murine models of MRSA superficial skin and pneumonia infections were carried out as previously described (7, 8, 37) with minor modifications. Inoculum preparation, infection conditions, and postinfection procedures are fully presented in SI Materials and Methods.

Other Procedures.

Detailed descriptions of all other procedures are available in SI Materials and Methods. Animal experiments were carried out with the approval of the Institutional Animal Care and Use Committees of Stanford University School of Medicine and the RIKEN Center for Developmental Biology.

SI Materials and Methods

Eukaryotic Cell Culture.

HAP1 cells and isogenic knockout subclone lines were grown in Iscove's modified Dulbecco's Medium (IMDM) (HyClone) supplemented with 10% (vol/vol) FCS, 1× penicillin/streptomycin (Gibco), and 2 mM l-glutamine. HEK-293T cells were used for lentivirus generation and cultured in DMEM (HyClone) supplemented with 10% FCS, 1× penicillin/streptomycin (Gibco), and 2 mM l-glutamine. MDCK II cells (kindly provided by W. James Nelson, Stanford University, Stanford, CA) were grown in DMEM (HyClone) supplemented with 10% FCS and 1× penicillin/streptomycin. Polarized MDCK monolayers were cultured by seeding cells at confluent density onto 12 mm, 0.4-µm pore polycarbonate tissue culture inserts (transwell filters; Corning Costar) and maintained as previously described (42).

Haploid Human Cell Genetic Screen.

HAP1 cells were mutagenized with a retroviral gene trap to cause inactivating mutations throughout the genome, and a haploid genetic screen was performed as previously described (22, 23). Recombinant, purified S. aureus α-toxin was resuspended in PBS at 0.5 mg/mL (Sigma, lot 111M4048V). Approximately 10⁸ gene-trap mutagenized cells were treated with 0.5 μg/mL α-toxin for 48 h. Following selection, surviving HAP1 colonies were expanded and pooled. Genomic DNA from the surviving, expanded cell population was isolated using the QIAamp DNA mini kit (Qiagen). Gene-trap insertion sites were recovered by linear amplification of genomic DNA sequences flanking the proviral DNA insertions and mapped to the human genome by deep sequencing. For each gene, enrichment of inactivating gene-trap insertions in the α-toxin selected pool over an unselected control dataset was determined by a one-sided Fisher’s exact test and corrected for false discovery rate as previously described (23).

Genome Engineering.

CRISPR sequence targeting oligos were designed using the Zhang Lab CRISPR design tool (crispr.mit.edu). Oligos corresponding to the guide RNA sequences in Table S1 were synthesized (Integrated DNA Technologies). Guide RNA oligos were directly cloned into Zhang laboratory-generated Cas9-expressing plasmid px458 as described (50), a gift from Feng Zhang, Broad Institute of Massachusetts Institute of Technology, Cambridge, MA (Addgene plasmid 48138), using the Gibson Assembly Reaction (New England BioLabs). HAP1 or MDCK cells were transfected with guide RNA encoding px458 plasmids as described in Table S1 using Lipofectamine 2000 according to the manufacturer’s guidelines (Life Technologies). For the α-catenin CRISPR knockout cell line generation, the blasticidin resistance gene encoding plasmid pTIA (kindly provided by Thijn Brummelkamp, Netherlands Cancer Institute, Amsterdam, The Netherlands) was transfected along with the α-catenin guide containing px458 plasmid to allow for blasticidin screening-based selection of edited clones.

Table S1.

CRISPR/Cas9 genome editing guide sequences and primers

Gene	Guide RNA sequence	Genotyping sequencing primers
ADAM10	5′ TGATGATGGCGTACTTGGTC 3′	5′ GATCAGCTCAGGGATGTGG 3′ and 5′ CTGTAGTGAGATAAAGAGGAG 3′
PLEKHA7	5′ GGAGAAAACTGTCCCGTCAC 3′	5′ GACAATACCCATCATGCAC 3′ and 5′ CAACAGATATATGTGCCAGGG 3′
α-catenin	5′ GAACCATGTTGCCTCGCTTC 3′	5′ CCAGCAATATATGAAGGTGC 3′ and 5′ TTGGAATAGAGGAGTAAGTG 3′
Afadin	5′ GAGGACAGCATTCGCATATC 3′	5′ CATCTCTCTACATTAGTCTCAG 3′ and 5′ CGTTACTTATCTTTGGAGAAAT 3′
N-cadherin	5′ AGCACAGTGGCCACCTACAA 3′	5′ AGCTTTCTAATCCACAGTGTG 3′ and 5′ TTTCAGATACACAGTATAACCC 3′
PLEKHA7, canine	5′ ACCAACAGACCACAGCATTCAGG 3′	5′ CCACACTTCCACATCTTCCC 3′ and 5′ GTGGTCTCTTGCAGCCATTC 3′

At 48 h posttransfection, cells were single-cell sorted using a BD InFlux cell sorter at the Stanford Shared FACS facility based on GFP expression into 96-well tissue plates containing cell culture growth media. Single-cell subclones were expanded and genomic DNA isolated for sequence-based genotyping of the targeted exonic sites using the sequencing primers listed in Table S1. Subclones containing frame-shift or large indels were selected for further experimentation, and when possible, gene knockout was confirmed by protein immunoblotting. ΔPLEKHA7 subclones in the diploid MDCK cell line were confirmed by TOPO cloning genomic DNA PCR products and sequencing multiple colonies to verify the editing events in each allele. Multiple independent subclones for each gene knockout were isolated and tested to confirm the reported phenotypes.

PLEKHA7 Cloning and Lentiviral Transduction of Cells.

Using Gateway cloning and LR Clonase II (Life Technologies), human PLEKHA7 cDNA (Thermo Scientific, accession no. BC071599) was inserted into the pLenti CMV Puro DEST vector (w118-1). pLenti CMV Puro DEST (w118-1) was a gift from Eric Campeau, University of Massachusetts Medical School, Worcester, MA (Addgene plasmid 17452) and has been previously described (51). Lentivirus was produced and used to transduce the CRISPR/Cas9-generated ΔPLEKHA7 cell line. Two days posttransduction, stable cell lines were selected by treatment with 1 µg/mL puromycin (InvivoGen). To generate PLEKHA7 full-length and deletion mutant constructs, the following primers were used to generate PCR products from the human PLEKHA7 cDNA and then cloned directly into pLenti CMV Puro DEST using the Gibson Assembly Reaction (New England BioLabs). Reverse primers were designed to incorporate a C-terminal 1× FLAG tag sequence. Constructs were generated with the following primers to generate PLEKHA7 constructs: PLEKHA7 full-length, forward 5′ GACTCTAGTCCAGTGTGGTG 3′ and reverse 5′ ATCCAGAGGTTGATTGTCGAG 3′; PLEKHA7 Δ1–284, forward 5′ TGTGGTGGAATTCTGCAGATACCATGTCTCGATCG TCACTGAAGAG 3′ and reverse 5′ ATCCAGAGGTTGATTGTCGAG 3′; PLEKHA7 Δ1–500, forward 5′ TGTGGTGGAATTCTGCAGATACCATGCGAGCCAGCCACCTGAAG AT 3′ and reverse 5′ ATCCAGAGGTTGATTGTCGAG 3′; PLEKHA7 Δ800–1121, forward 5′ GACTCTAGTCCAGTGTGGTG 3′ and reverse 5′ CGGCCGCCACTG TGCTGGATTTACTTATCGTCGTCATCCTTGTAATCGGCTCTTCTGTGTTGCTCCT 3′; PLEKHA7 Δ698–1121, forward 5′ GACTCTAGTCCAGTGTGGTG 3′ and reverse 5′ CGGCCGCCACTGTGCTGGATTTACTTATCGTCGTCATCCTTGTAATCGATGCTCAGTTTGACGTCAG 3′; PLEKHA7 1–163, forward 5′ GACTCTAGTCCAGTGTGGTG 3′ and reverse 5′ CGGCCGCCACTGTGCTGGATTTACTTATCGTCGTCATCCTTGT AATCAACATTGGGGTTCCTCCGAA 3′; and PLEKHA7 1–284, forward 5′ GACTCTAGTC CAGTGTGGTG 3′ and reverse 5′ CGGCCGCCACTGTGCTGGATTTACTTATCGTC GTCATCCTTGTAATCCAGCACCTGTGCAGCCTGGT 3′. To generate the internal deletion construct PLEKHA7 Δp120 (538–696), two PCR products were generated and then assembled into pLenti CMV Puro DEST using the Gibson Assembly Reaction. The N-terminal fragment was generated using the following primers: forward 5′ GACTCTAGTCCAGTGT GGTG 3′ and reverse 5′ GGGGCTGCCGTGCCGGAACT 3′. The C-terminal fragment was generated using the following primers: forward 5′ AGTTCCGGCACGGCAGCCCCAT CTTCTGTGAACAAGACAG 3′ and reverse 5′ ATCCAGAGGTTGATTGTCGAG 3′.

Intoxication and Cell Viability Assays.

HAP1 cells were seeded at 30,000 cells per well in 96-well tissue culture plates 1 d before intoxication with 0.5 µg/mL α-toxin. At 24 h postintoxication, viability was determined by CellTiter-Glo luminescent cell viability assay (Promega) according to the manufacturer’s guidelines and read with a microplate reader (Tecan). Viability is plotted as a percentage compared with untreated control wells corresponding to each cell line. For macroscopic visualization of cellular viability following intoxication, HAP1 cells were seeded at 300,000 cells per well in 24-well tissue culture plates the day before intoxication and then treated with the indicated concentrations of α-toxin, other toxins (streptolysin O, kindly provided by Andres Lebensohn, Stanford University School of Medicine, Stanford, CA, and perfringolysin O, kindly provided by David Bewslow, Stanford University School of Medicine, Stanford, CA), or the ionophore nigericin (Sigma). Two days after α-toxin treatment, viable, adherent cells were fixed with 4% (vol/vol) formaldehyde in PBS and stained with crystal violet. For microscopic visualization of cellular viability following intoxication, HAP1 cells were seeded at 50,000 cells per well onto Permanox chamber slides (LabTek) and treated the following day with 5 µg/mL α-toxin. At 24 h following intoxication, cells were washed 1× with PBS, then treated with the LIVE/DEAD viability/cytotoxicity kit according to the manufacturer’s guidelines (Life Technologies), and visualized directly with a Zeiss LSM 700 confocal microscope.

WT and Δ PLEKHA7 MDCK cells were seeded onto 0.4 µm 12-well transwell filter cell culture inserts (Corning) at 1.0e⁶ cells per transwell insert and cultured under polarization conditions for 4 d (42). On day 5 postseeding, basolateral media was changed to media containing 2.5 µg/mL α-toxin, and apical media was replaced with plain DMEM containing 1 µm SYTOX 488 (Life Technologies). At 3 h postintoxication, the apical media was washed 5× with plain DMEM to remove excess SYTOX dye before fixation with 2% (vol/vol) paraformaldehyde in 100 mM phosphate buffer (pH 7.4) for 10 min at room temperature. Whole transwells filters were collected and monolayers counterstained with DAPI and 594 phalloidin (Life Technologies) in blocking buffer with PBS with 3% (wt/vol) BSA, 1% saponin, and 1% Triton X-100. Monolayers were imaged with a Zeiss LSM 700 confocal microscope.

Immunoblotting.

Cell pellets were lysed with hot Lamelli SDS sample buffer containing 5% (vol/vol) β-mercaptoethanol and boiled for 10 min. To probe for secreted S. aureus α-toxin expression, WT USA300 LAC or isogenic mutant cultures were grown overnight to stationary phase, diluted to equivalent OD₆₀₀, and supernatants filtered before dilution in 2× hot Lamelli SDS sample buffer containing 5% β-mercaptoethanol.

To visualize total secreted bacterial exoproteins, stationary phase, cell-free supernatants from WT USA300 LAC and hla::ermB USA300 LAC were concentrated by trichloroacetic acid (TCA) precipitation. Cells were separated from 10 mL of equivalent density overnight cultures by centrifugation, the supernatants were passed through a 0.2-μm filter, and then one volume of 100% TCA (wt/vol) was added to four volumes of cell-free supernatant. After incubation for 10 min at 4 °C, total protein was pelleted by centrifugation and washed with 200 μL cold acetone and repeated for two acetone washes. Protein pellets were dried at room temperature for 30 min before resuspension in 2× Lamelli SDS sample buffer containing 5% β-mercaptoethanol and separation by SDS/PAGE as described below, and then total protein was visualized by Coomassie brilliant blue stain (Bio-Rad).

Lysates were separated by SDS/PAGE using the Mini-Protean system (Bio-Rad) on 4–15% (vol/vol) polyacrylamide gradient gels (Bio-Rad). Proteins were semiwet-transferred onto PVDF membranes (Bio-Rad) using the Bio-Rad Transblot protein transfer system. Membranes were blocked by incubating with PBS buffer containing 5% (wt/vol) nonfat milk and 0.01% Tween-20 for 1 h at room temperature. Membranes were incubated overnight at 4 °C with primary antibodies (from the following sources and dilutions) in blocking buffer as described below. Primary antibodies were detected using HRP-conjugated secondary antibodies, anti-mouse, and anti-rabbit (GeneTex) by incubating membranes with 1:5,000 in blocking buffer for 1 h at room temperature. Antibody-bound proteins were visualized by film using the West Pico and Extended Duration chemiluminescence peroxide solutions (Thermo).

Antibodies.

ADAM10 was detected using a rabbit monoclonal antibody (GeneTex, EPR5622, 1:3,000). Human and murine PLEKHA7 was detected with previously described (28) rabbit polyclonal antibody against recombinant C-terminal fragments of PLEKHA7 (1:2,000 Western blot and 1:200 immunofluorescence). Canine PLEKHA7 was detected by Western blot with a polyclonal guinea pig antibody against recombinant C-terminal fragments of canine PLEKHA7 used at 1:1,000 (antibody a kind gift from Sandra Citi, University of Geneva, Geneva, Switzerland). Afadin was detected using a rabbit polyclonal antibody (Sigma, A0349, 1:2,000). N-cadherin was detected using a mouse monoclonal antibody (EMD Millipore, clone 13A9, 1;5,000). α-catenin was detected using a rabbit polyclonal antibody kindly provided by W. James Nelson, Stanford University, Stanford, CA, at 1:5,000. p38 MAPK was visualized with rabbit polyclonal antibody (Cell Signaling Technology, 9212, 1:1,000). Phosphorylated p38 MAPK (Thr180/Tyr182) was visualized using a rabbit polyclonal antibody (Cell Signaling Technology, 9211, 1:1,000). GAPDH was visualized with rabbit polyclonal antibody (Cell Signaling Technology, 5174, 1:5,000). S. aureus α-toxin was visualized with a rabbit antibody against H35L mutant α-toxin provided by Fabio Bagnoli at 1:1,000. Expression of FLAG-tagged PLEKHA7 constructs were visualized using rabbit polyclonal antibody anti-DYKDDDK (Cell Signaling Technology, 2368s, 1:2,500).

Live-Cell Time-Lapse Microscopy.

WT or knockout HAP1 cells were seeded on Lab-TeK II two-chambered coverglass (Nalge Nunc International) slides for simultaneous imaging of two conditions. Imaging was performed using a Nikon TE2000E microscope equipped with high numerical aperture objectives for imaging at low light levels to minimize phototoxicity and a stage enclosed in an incubator system with temperature and CO₂ enrichment controls. A motorized, computer-controlled stage was used to sequentially monitor multiple sites of the same specimen over time by transmitted light differential interference contrast (DIC). For DIC imaging, a Hamamatsu high-resolution ORCA-285 digital camera was used. A z-stack of images was collected at each time point and processed with OpenLab software (Improvision), and the resulting movies were assembled in Quicktime Pro.

Flow Cytometry.

HAP1 cells were harvested from tissue culture plates by incubating cells in PBS with 1 mM EDTA for 10 min at 37 °C. Live cells were suspended in 100 µL FACS buffer [5% (vol/vol) FCS, 0.1% Azide, and 1 mM EDTA in 1× PBS] and stained for 30 min on ice with ADAM10-PE conjugated antibody (anti-human CD156c, clone SHM14, BioLegend) according to the manufacturer’s guidelines. Cells were analyzed using a BD LSRII cytometer with DIVA 6 acquisition software and gated using FlowJo vX.0.7. Events (at least 7,000 cells in the final gate) were gated on live cells and singlets. Cells were gated for PE-positive signal using unstained cells as a negative control.

Cellular Assays of α-Toxin Pore Formation.

Efflux of intracellular potassium upon intoxication was assessed by seeding HAP1 cells at 2.0e⁶ cells per well in a six-well tissue culture plate 1 d before intoxication. Cells were either treated with media alone or with 5 µg/mL α-toxin in 500 µL total volume. At 30 min postintoxication, extracellular media was collected from α-toxin–treated cells and media control-treated cells, spun at 10,000 × g for 10 min, and passed through a 0.4-µm filter (Millipore) to remove any cellular debris. Potassium concentration in extracellular media was quantified using an Xpand Chemistry Analyzer (Siemens) by the Stanford University Veterinary Service Center diagnostic facility. Changes in extracellular potassium levels were normalized as change from media control potassium levels for each individual cell line tested.

Intracellular ATP concentration was assessed over a time-course postintoxication by seeding HAP1 at 30,000 cells per well in 96-well tissue culture plates 1 d before intoxication with 0.75 µg/mL α-toxin in complete media. Total intracellular [ATP] was determined at the indicated time points postintoxication by CellTiter-Glo luminescent cell viability assay (Promega) according to the manufacturer’s guidelines and read with a microplate reader (Tecan). Cells were treated in a reverse time course of intoxication in technical quadruplicates, allowing for simultaneous lysis of all conditions at the end of the experiment to facilitate uniform luminescence reading across multiple conditions. Intracellular ATP is plotted as a percentage compared with untreated control wells corresponding to each cell line and each time point posttreatment with α-toxin or media alone.

Generation and Validation of PLEKHA7^−/− Transgenic Mice.

The PLEKHA7(LacZ) mutant mice (accession no. CDB0750K, www.clst.riken.jp/arg/mutant%20mice%20list.html) were generated as described (www.clst.riken.jp/arg/Methods.html). The targeting vector of Plekha7 was cloned from a mouse genomic library (BACPAC). The fragment of eighth exon was replaced by the Neo selection cassette. Targeted ES clones were microinjected into ICR eight-cell stage embryos, and injected embryos were transferred into pseudopregnant ICR females. The resulting chimeras were bred with C57BL/6 mice, and heterozygous offspring were identified by PCR using the following primers: forward P1 5′ ACATGAACGCCTGGGTCAGG 3′ and reverse P3 5′ GCCAGTGAGATGGTCCAGTT 3′ for the WT allele, and forward P2 5′ ATGGAAGGATTGGAGCTACG 3′ and reverse P3 for the targeted allele, yielding 627 bp and 811 bp products, respectively. PLEKHA7(LacZ) and WT mice (Jackson Laboratory) were maintained and bred in accordance with the protocols approved by the Institutional Animal Care and Use Committees of Stanford University School of Medicine and the RIKEN Center for Developmental Biology.

Tissue samples from WT and PLEKHA7^−/− murine stomachs were processed for confocal immunofluorescence microscopy as previously described, with minor modification. Tissue samples were fixed in 2% paraformaldehyde in 100 mM phosphate buffer (pH 7.4) for 1 h. Tissue was embedded in 4% (wt/vol) agarose, and 100–200-µm sections were generated using a Vibratome (Leica). Tissue sections were permeabilized in PBS with 3% BSA, 1% saponin, and 1% Triton X-100 before staining with antibodies described above. Samples were imaged with a Zeiss LSM 700 confocal microscope.

Bacterial Strains and Culture.

The WT MRSA strain USA300 LAC was kindly provided by Fabio Bagnoli. To generate hla:ermB mutant, an insertional lesion of the bursa aurealis minitransposon in the hla gene of S. aureus Newman was transduced to S. aureus USA300 LAC WT using bacteriophage 80 as previously described (52). For selection of the colonies harboring the trasposon mutation, tryptic soy agar (Sigma) plates containing 100 μg/mL erythromycin were used. The insertion was verified by DNA sequencing, and lack of Hla expression was checked by Western blotting. For plasmid complementation, full-length hla incorporating the endogenous promoter was cloned into PstI and BamHI sites of plasmid pOS1 (53) and transformed into Escherichia coli Dh5α. The resulting plasmid phla was introduced into the S. aureus RN4220 strain by electroporation. Plasmid isolated from this RN4220 strain was then introduced by electroporation into the hla::ermB USA300 LAC mutant strain. Complementation of hla gene was verified by DNA sequencing, and Hla expression was checked by Western blotting. The complemented strain was named hla::ermB-phla. Bacteria was grown in tryptic soy broth (Sigma) at 37° and prepared as indicated below for animal infections.

Superficial Skin and Soft Tissue Murine Model of S. aureus Infection.

We used a superficial skin and soft tissue murine infection model of MRSA infection as previously described (37). Briefly, 2 d before infection, S. aureus USA300 LAC or the USA300 LAC hla::ermB was streaked onto trypic soy agar (Sigma), and the following day an overnight culture in tryptic soy broth was initiated from a single bacterial colony. On the day of infection, overnight bacterial cultures were subcultured at 1:100 (vol/vol) in 50 mL of fresh tryptic soy broth on a shaker at 200 rpm for 2–3 h until an optical absorbance density (OD₆₀₀) of between 0.75 and 0.9 was reached. Bacteria were pelleted by centrifugation, washed 1× in sterile PBS, and resuspended in PBS at a density corresponding to 1.0 × 10¹¹ cfu/mL. Inoculum density was determined by serial dilution on tryptic soy agar to enumerate viable staphylococci. We dispensed 10 µL of this inoculum onto the tip of a sterile allergy-testing needle (Morrow Brown Allergy Diagnostics) for infecting a single murine ear. WT PLEKHA7^+/+ and PLEKHA7^−/− mice (males and females, 6–8 wk old, 6–12 per group) were lightly anesthetized by isoflurane inhalation before epicutaneous challenge. Before infection, mouse ears were cleansed with 70% (vol/vol) isoproponal, allowed to dry, and then the ventral epidermis of the ear was pricked 20 consecutive times at the same tissue site with a S. aureus-coated needle.

To enumerate bacterial density in infected ear tissues, animals were euthanized by CO₂ inhalation at the indicated time points, the infected ear pinna was excised and mechanically homogenized, and then homogenates were serially plated on tryptic soy agar. Disease progression was followed in individual animals over time by consecutively imaging isoflurane-anesthesized animals every 2 d postinfection for up to 14 d. A fixed camera position was used to standardize image size across animals and time points throughout multiple experiments, and an internal image ruler was used for validation and scaling. Lesion sizes were calculated from raw images by measuring the pixel area of the lesion area in ImageJ (NIH) to convert image pixel area to mm². For immunohistochemical analysis, infected ear tissue was gathered at indicated time points postinfection and fixed with 10% (vol/vol) neutral-buffered formalin. Samples were paraffin-embedded, sectioned, and stained with hematoxylin and eosin by the Department of Comparative Medicine at Stanford University. Blinded slides were interpreted by a veterinary pathologist (D.M.B.).

Pneumonia Model of S. aureus Infection.

We used a murine model of S. aureus pneumonia as previously described (7, 8). Briefly, S. aureus USA300 LAC was grown as described above. After subculturing, bacteria were pelleted by centrifugation, washed 1× in sterile PBS, and the pellet was resuspended in PBS at a density corresponding to ca. 2.5 × 10¹¹ cfu/mL. A total of 30 µL of this suspension was delivered by intranasal inoculation into the left nare of each infected animal, corresponding to a per animal inoculum of ca. 2–3 × 10⁸ cfu. In a higher inoculum challenge shown in Fig. 3 E and F, an infectious dose of ca 8 × 10⁸ cfu per animal was administered in a 30 µL PBS suspension. Inocula were determined by serial dilution plating on tryptic soy agar. Before infection, 6–8-wk-old male and female WT PLEKHA7^+/+ and PLEKHA7^−/− mice (12–25 per group, per experiment) were anesthetized by i.p. injection of 100 mg/kg ketamine and 10 mg/kg xyazline in sterile PBS. Animals were held upright for 1 min postinfection. Animals were monitored every 6 h for 72 h and euthanized after 72 h or earlier if moribund. Core body temperature and weight of all animals was monitored before infection and every 6 h postinfection by use of a rectal thermometer. To enumerate bacterial density in infected lung tissues, infected animals were euthanized by CO₂ inhalation at the indicated time points, and the left lung was excised and mechanically homogenized, and homogenates were serially plated on tryptic soy agar.

Ethics Statement.

All animal experiments were performed in accordance with NIH guidelines, the Animal Welfare Act, and US federal law. Animal experiments were carried out with the approval of the Institutional Animal Care and Use Committee of Stanford University. Animals were housed in Stanford University School of Medicine animal facilities, which are fully staffed with 24-h veterinary personnel and accredited by the Association of Assessment and Accreditation of Laboratory Animal Care International.

Statistical Analyses.

Unless otherwise indicated, unpaired Student’s t test was used for statistical calculations involving two group comparisons (*P < 0.05, **P < 0.01). Statistical significance of mouse mortality studies in the pneumonia model was assessed by the Mantel–Cox log-rank test.