Staphylococcus aureus Panton-Valentine Leukocidin Causes Necrotizing Pneumonia

See allHide authors and affiliations

Science  23 Feb 2007:
Vol. 315, Issue 5815, pp. 1130-1133
DOI: 10.1126/science.1137165


The Staphylococcus aureus Panton-Valentine leukocidin (PVL) is a pore-forming toxin secreted by strains epidemiologically associated with the current outbreak of community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) and with the often-lethal necrotizing pneumonia. To investigate the role of PVL in pulmonary disease, we tested the pathogenicity of clinical isolates, isogenic PVL-negative and PVL-positive S. aureus strains, as well as purified PVL, in a mouse acute pneumonia model. Here we show that PVL is sufficient to cause pneumonia and that the expression of this leukotoxin induces global changes in transcriptional levels of genes encoding secreted and cell wall–anchored staphylococcal proteins, including the lung inflammatory factor staphylococcal protein A (Spa).

The combined actions of many virulence factors enable Staphylococcus aureus to cause disease (1, 2). Depending on these factors and on the immune status of the host, staphylococci can cause diseases ranging from superficial skin infections to deep-seated and systemic conditions such as osteomyelitis, septic shock, and necrotizing pneumonia. Staphylococcal necrotizing pneumonia can affect young, immunocompetent patients. This disease, characterized by leukopenia, hemoptysis, and extensive necrosis of the lung tissue, is caused by S. aureus strains that produce Panton-Valentine leukocidin (PVL) (3). S. aureus PVL-positive strains are often methicillin-resistant (MRSA) and, in the USA, they are the predominant cause of community-associated infections (4).

PVL is a bi-component, pore-forming exotoxin (5) that targets cells of the immune system such as polymorphonuclear neutrophils (PMNs). The active form of PVL requires the assembly of two polypeptides, LukS-PVand LukF-PV, into a heterooligomeric pore. Although PVL has potent cytolytic and inflammatory activities in vitro (6, 7), its role in necrotizing pneumonia has not been demonstrated. To analyze the molecular pathogenesis of PVL-expressing S. aureus strains, we have established a murine model of acute primary pneumonia.

We infected mice with strains isolated from necrotizing (PVL-positive) or nonnecrotizing (PVL-negative) staphylococcal pneumonia patients (table S1). PVL-positive strains caused murine necrotizing pneumonia with manifestations resembling human disease (fig. S1). In the PVL-positive strains, the lukS-PV and lukF-PV genes are organized as an operon within a phage (φSLT, or other similar phages) that could potentially contribute to the virulence of these strains. To define the role of PVL, we developed several isogenic strains (8). A PVL-negative, transformable S. aureus strain was lysogenized with φSLT or with a mutated φSLT in which the PVL operon (luk-PV) was deleted. We complemented the PVL-negative strains with a plasmid containing the luk-PV operon under the control of its own promoter (table S1).

Mice infected with PVL-positive strains showed symptoms of severe illness: lethargy, hunched posture, ruffled fur, and significant weight loss. Lungs from infected mice were examined 48 hours postinoculation. Tissue sections from lungs infected with PVL-positive strains revealed a strong recruitment of neutrophils and significant inflammation in the lung parenchyma, bronchial epithelial damage, tissue necrosis, and hemorrhage (Fig. 1 and table S2). The lungs infected with PVL-negative strains showed normal lung structures, despite some leukocyte infiltration. By contrast, when the PVL-negative strains were complemented with a plasmid encoding PVL, we observed massive tissue damage and 35 to 80% mortality within 24 hours after inoculation (Fig. 1 and fig. S2). We stained lung sections using antibodies against LukS-PV (fig. S3) and showed that the toxin was detected in tissues infected with PVL-positive bacteria.

Fig. 1.

Expression of PVL enhances the virulence of isogenic S. aureus strains. (A, B, E, F, I, and J) Lung histology of mice infected with PVL-positive and PVL-negative strains or inoculated with PVL toxin. The sections are representative of at least three separate experiments. Scale bar, 100 μm. (C, G, and K) Line graphs indicate weight loss in grams. (C) Parental versus PVL phage; *P < 0.001. (G) Parental versus ΔPVL phage; no statistical difference observed. (K) Animals inoculated with 3 μgof LukS+F-PV versus 5 μg LukS-PV; *P < 0.01 on day 1. (D, H, and L) Mouse survival. (D) Parental and PVL plasmid; (H) PVL phage, ΔPVL phage, and ΔPVL phage PVL plasmid; (L) 20 μgor 10 μg of LukS+LukF-PV versus 3 μgLukS+LukF-PV or 5 μgLukS-PV, *P <0.0001.

Administration of increasing equimolar amounts of native LukS-PV and LukF-PV resulted in concentration-dependent localized lesions, weight loss, and, at concentrations higher than 3 μg, high rates of mortality (Fig. 1). The protein-inoculated mice recovered the lost weight after 24 hours, whereas those infected with PVL-positive bacteria were still ill at 48 hours (Fig. 1C compared with Fig. 1K); these findings demonstrated that an active bacterial infection is required to cause severe morbidity.

Previous studies have demonstrated that PVL-positive strains had increased adherence to injured airway epithelium (3, 9). To examine the expression of surface proteins in these strains, we examined by SDS–polyacrylamide electrophoresis (SDS-PAGE) cell wall extracts and supernatants from cultures taken at both the exponential growth phase and the stationary growth phase (Fig. 2). Samples from a PVL-positive strain showed an enhanced expression of at least two cell wall–anchored polypeptides identified as SdrD and protein A (Spa) by N-terminal sequencing and Western analysis (Fig. 2, C to E, and fig. S4). The SdrD and Spa overexpression was not observed in a strain carrying the phage with the deleted luk-PV operon, which indicated that this effect was not mediated by products encoded by other phage genes or by its insertion in the chromosome of S. aureus. PVL-negative strains complemented with the luk-PVoperon in a multicopy plasmid (table S1) also showed an increased expression of Spa during both logarithmic and stationary growth phases (Fig. 2F). A group of polypeptides with apparent molecular masses between 32 and 47.5 kD (Fig. 2B, dots) were present in the supernatants from the PVL-negative, but not the PVL-positive, strains. Some of the secreted polypeptides were identified as proteases by using zymograms (Fig. 2G). Thus, expression of the luk-PV operon resulted in an altered regulation of cell wall–anchored and secreted protein production.

Fig. 2.

PVL alters the expression pattern of cell wall–anchored and secreted proteins. (A to E) Lanes 1, 2, and 3 represent extracts from parental, ΔPVL phage, and PVL phage strains respectively; samples were isolated from bacterial cultures grown at exponential (EXP) and stationary (STAT) phases of growth. (A) Lysostaphin extracts analyzed by SDS-PAGE. The arrows point to overexpressed proteins. (B) Exoproteins from culture supernatants harvested from bacterial cultures analyzed by SDS-PAGE. Arrows point to overexpressed proteins; dots indicate exoproteins reduced or absent in PVL-positive strains. (C and D) Western blot analysis of lysostaphin extracts and supernatants using a monoclonal antibody against Spa. Complete gels shown in fig. S4 (8). *Samples were diluted 1:100. (E) Western blot analysis of lysostaphin extracts using polyclonal antibodies against SdrD. SdrD is detected as two polypeptides because of the proteolytic cleavage of an N-terminal subdomain. (F) Western blot analysis of supernatants from parental PVL plasmid stationary phase (7-hour) cultures (lane 1) and ΔPVL phage PVL plasmid (lane 2) using monoclonal antibodies against Spa. (G) Zymogram analysis of exoproteins from cultures grown at stationary phase. Lanes 1, 2, and 3 represent proteins extracted from parental, ΔPVL phage, and PVL phage, respectively.

Spa is a known virulence factor in mouse models of S. aureus infections, including pneumonia (10, 11); therefore, we examined the role of Spa in necrotizing pneumonia. Animals infected with the spa-deleted isogenic strains had less severe symptoms of disease (Fig. 3A). However, the lungs from animals infected with spa-deleted, PVL-positive strains showed localized lesions with massive leukocyte infiltration (fig. S5), which demonstrated that PVL alone was sufficient to cause pneumonia. Complementation of Spa-positive strains with PVL rendered them lethal, whereas the Spa-negative, PVL-plasmid strains did not cause mortality (Fig. 3B). These data suggested that PVL and Spa may act together to cause the overwhelming inflammation and tissue damage that are seen in necrotizing pneumonia.

Fig. 3.

Spa enhances the virulence of PVL-positive strains. (A) Line graph indicates weight loss in grams, *P <0.01 ΔSpa versus ΔSpa PVL phage or ΔSpa PVL plasmid. (B) Percent survival of animals infected with ΔSpa, ΔSpa PVL plasmid, ΔSpa PVL phage, PVL plasmid, and ΔPVL phage PVL plasmid, *P <0.001 ΔSpa, ΔSpa PVL phage and ΔSpa PVL plasmid versus PVL plasmid or ΔPVL phage PVL plasmid.

PVL-positive strains expressed Spa during both exponential and stationary growth phases (Fig. 2). To analyze the effect of PVL on spa transcription, PVL was introduced into mutants deficient in spa regulators. The Spa production was abolished in a sarS-deletion mutant (fig. S6), whether PVL was present or not, which indicated that PVL acts upstream of SarS. To evaluate the transcriptional profile of a PVL-positive strain compared with the PVL-negative strain, we used microarray analysis: 28 genes showed a distinct expression pattern during exponential growth (table S3), whereas, during the stationary phase, 133 genes showed differential expression (table S4). The agr transcripts and several exoproteins were repressed, whereas genes encoding for cell wall–anchored proteins and the spa activator sarS (12) were up-regulated (Fig. 4A). Elevated expression of the spa and sdrD transcripts correlated with the enhanced production of Spa and SdrD observed in Western blot analysis (Fig. 2 and fig. S4). The repression of exoprotein transcripts paralleled the absence of the 32- to 47.5-kD exoproteins in the supernatants of PVL-positive strains (Fig. 2). This pattern of transcription implicated PVL in an interaction with a factor (or factors) that controls gene expression during the transition from the logarithmic growth to stationary phase.

Fig. 4.

S. aureus PVL-positive strains show an altered transcriptional profile. (A) Fold increase or decrease levels of transcript from selected genes. Total RNA extracted from cultures grown to stationary phase. Genes were considered to be induced or repressed in the PVL phage if they were transcribed at least twice or half as much as those of ΔPVL phage. The shown transcripts encode agrA-C, accessory gene regulator system; sarS, staphylococcal accessory regulator S; spa, staphylococcal protein A; sdrD, serine-aspartate repeat protein D; sdrC, serine-aspartate repeat protein C; clfB, clumping factor B; hla, alpha toxin; ssp, a representative of serine proteases sspBand sspC; spl, a representative of splA-F proteases. (B) A schematic overview of the interactions between regulators involved in cell wall–anchored and secreted protein genes (full and broken lines indicate positive and negative regulation, respectively) based on previously published data. Numbers next to the gene name indicate fold change based on microarray analysis (upward arrow indicates up-regulation, downward arrow indicates down-regulation). The down-regulation of RNAIII (the effector of the agr system) results in the down-regulation of secreted protein genes (hla, hlgC, hlgB, and proteases) and the up-regulation of sarS and cell wall–anchored proteins (spa, sdrD, sdrC, clfB). In addition, the up-regulation of sarS results in the up-regulation of spa.

The regulatory model proposed here was inferred by using strains derived from RN6390, a strain harboring a deletion in rsbU, which encodes a regulator necessary for the activity of the stress sigma factor σB. Strains with a reduced σB (13) activity display, among other traits, a decreased production of cell wall–anchored proteins and an increased production of exoproteins. We subsequently generated isogenic PVL-positive and PVL-negative strains in the SH1000 (14) rsbU+ background and observed overexpression of Spa during the stationary phase of growth (fig. S7A). The SH1000-derived PVL-positive strain was more virulent than its PVL-negative isogenic pair (fig. S7, C and D). When compared with RN6390, the SH1000 lineage showed an increased expression of Spa, but produced ≥20% PVL, although some variability was seen (fig. S7A) (15). This suggested that RsbU/σB partially regulates PVL.

We show here that PVL is a significant S. aureus virulence factor and that PVL-positive strains can cause murine necrotizing pneumonia with manifestations that resemble those observed in human patients. Our results demonstrate that the expression of the genes that encode PVL (lukS-PVand lukF-PV) or direct inoculation with native toxin is sufficient to induce pneumonia in mice. The expression of the luk-PVoperon also resulted in an altered expression of multiple proteins, including the tightly regulated (16, 17) proinflammatory factor Spa.

In PVL-positive strains, many secreted proteins are down-regulated (Fig. 4B), similarly to data reported by Vojtov et. al. (18), who demonstrated that two staphylococcal superantigens, toxic shock syndrome toxin-1 (TSST-1) and enterotoxin B (SEB), strongly repressed production of secreted proteins. It is possible that these toxins act similarly to PVL, interacting with unknown factors that interfere with regulatory networks.

Several genes encoding putative and known microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) (including SdrD) are up-regulated in the PVL-positive strains (Fig. 4A). The up-regulation of MSCRAMMs may lead to enhanced tissue adherence and colonization of PVL-expressing strains, thereby contributing to the virulence potential of these strains (19, 20).

Spa is highly expressed in PVL-positive strains. Our in vivo data underscore the documented role of Spa as a proinflammatory factor in pneumonia (11). Increased production of Spa, coupled with the ability of PVL to lyse PMNs and macrophages (6), could lead to a vicious cycle of cell recruitment, lysis, and release of inflammatory mediators (7), resulting in overwhelming tissue inflammation and necrosis.

Here, we show not only that PVL is a key virulence factor in pulmonary infections but also that expression of the luk-PV genes interferes with global regulatory networks, which may also enhance virulence. A detailed analysis of such dysregulation will be useful to identify targets for the potential development of novel therapies to treat S. aureus infections.

Supporting Online Material

Materials and Methods

Figures S1 to S8

Tables S1 to S5


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article