Increased Production of Outer Membrane Vesicles by Salmonella Interferes with Complement-Mediated Innate Immune Attack

Abstract

Bacterial outer membrane vesicles (OMVs) enriched with bioactive proteins, toxins, and virulence factors play a critical role in host-pathogen and microbial interactions. The two-component system PhoP-PhoQ (PhoPQ) of Salmonella enterica orchestrates the remodeling of outer membrane lipopolysaccharide (LPS) molecules and concomitantly upregulates OMV production. In this study, we document a novel use of nanoparticle tracking analysis to determine bacterial OMV size and number. Among the PhoPQ-activated genes tested, pagC expression had the most significant effect on the upregulation of OMV production. We provide the first evidence that PhoPQ-mediated upregulation of OMV production contributes to bacterial survival by interfering with complement activation. OMVs protected bacteria in a dose-dependent manner, and bacteria were highly susceptible to complement-mediated killing in their absence. OMVs from bacteria expressing PagC bound to complement component C3b in a dose-dependent manner and inactivated it by recruiting complement inhibitor Factor H. As we also found that Factor H binds to PagC, we propose that PagC interferes with complement-mediated killing of Salmonella in the following two steps: first by engaging Factor H, and second, through the production of PagC-enriched OMVs that divert and inactivate the complement away from the bacteria. Since PhoPQ activation occurs intracellularly, the resultant increase in PagC expression and OMV production is suggested to contribute to the local and systemic spread of Salmonella released from dying host cells that supports the infection of new cells. IMPORTANCE Bacterial outer membrane vesicles (OMVs) mediate critical bacterium-bacterium and host-microbial interactions that influence pathogenesis through multiple mechanisms, including the elicitation of inflammatory responses, delivery of virulence factors, and enhancement of biofilm formation. As such, there is a growing interest in understanding the underlying mechanisms of OMV production. Recent studies have revealed that OMV biogenesis is a finely tuned physiological process that requires structural organization and selective sorting of outer membrane components into the vesicles. In Salmonella, outer membrane remodeling and OMV production are tightly regulated by its PhoPQ system. In this study, we demonstrate that PhoPQ-regulated OMV production plays a significant role in defense against host innate immune attack. PhoPQ-activated PagC expression recruits the complement inhibitor Factor H and degrades the active C3 component of complement. Our results provide valuable insight into the combination of tools and environmental signals that Salmonella employs to evade complement-mediated lysis, thereby suggesting a strong evolutionary adaptation of this facultative intracellular pathogen to protect itself during its extracellular stage in the host.

Keywords: C3b; Factor H; PagC; PhoPQ; Rck; S. Typhimurium; Salmonella; complement resistance; outer membrane vesicles.

FIG 1

Quantitative analysis of OMV production by Salmonella under PhoPQ-activating conditions (5.8L). All S. Typhimurium strains used were flagella mutants (fliC fljB). (A) Relative fold OMV production by 2 × 10⁹ CFU/ml bacteria was calculated relative to the OMVs produced by parental strain S. Typhimurium fliC fljB (labeled as WT) normalized to 1. NTA analysis showed that the OMV production was significantly downregulated in ΔphoP, ΔpagC, and ΔpagP mutants. (B) A constitutively active phoP mutant (PhoP^C) and the ΔphoP mutant induced to express PagC from plasmid ppagC showed a hypervesiculating phenotype. Deletion of pagC in PhoP^C and the ΔphoP mutant containing the empty vector (pRS1) decreased OMV production by ∼3-fold. (C) The ΔpagC mutants of S. Typhimurium and S. Typhi showed significant reductions in OMV production compared with their respective WT parental strains. The phenotype of the S. Typhi strain could be restored by complementing the ΔpagC mutants with plasmid ppagC induced to express PagC. The data represent one of three separate, reproducible experiments, expressed as mean ± SEM. Statistical significance was calculated using one-way ANOVA multiple-comparison test with significance set at a P value of <0.05 (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; and ns, P > 0.05).

FIG 2

Relative fold OMV production by Salmonella grown under PhoPQ-activating conditions (5.8L). (A) Relative fold OMV production by the ΔpagC mutant was significantly reduced compared with WT bacteria or Δrck or ΔompX mutants. Deletion of rck or ompX had no significant effect on OMV production. (B) OMVs produced by the ΔpagC Δrck ΔompX triple mutant expressing PagC, Rck, or OmpX from corresponding plasmids showed that OMV production is upregulated only by the PagC-expressing plasmid in the ΔpagC Δrck ΔompX triple mutant. Plasmid pRS1 is the control empty vector. The data represent one of three separate, reproducible experiments, expressed as mean ± SEM. (C) Relative fold OMV production by ΔpagC Δrck ΔompX triple mutant containing empty vector pMMB207 or pail induced to express Ail from Y. pestis (100 μM isopropyl-β-d-thiogalactopyranoside [IPTG]) showed no significant change in OMV production. Statistical significance was calculated using one-way ANOVA multiple-comparison test and set at a P value of <0.05 (****, P < 0.0001; and ns, P > 0.05).

FIG 3

Serum resistance assay for Salmonella mutants grown under PhoPQ-activating conditions. (A) Survival of Salmonella deletion mutants in 25% NHS incubated at 37°C for 1 h. Deletion of ΔpgtE, Δrck (used as positive controls), ΔphoP, and ΔpagC made the bacteria susceptible to complement-mediated killing. (B) A comparison of serum resistance between Salmonella ΔpagC, Δrck, and ΔompX single, double, and triple mutants. (C) Quantitative RT-PCR-based analysis of gene expression ratio of pagC, rck, and ompX (relative to housekeeping gene rpoB) in wild-type Salmonella grown in LB or under PhoPQ-activating conditions (5.8L) showed the expression of pagC increased significantly under PhoPQ-activating conditions. (D) Resistance to serum-mediated attack was restored in single Salmonella mutants complemented with plasmids induced to express PagC, Rck, and OmpX in the corresponding mutants. Plasmids expressing Rck or PagC complemented best the triple mutant compared with mutants containing empty vector (pRS1). The data represent one of three separate, reproducible experiments, expressed as mean ± SEM. Statistical significance was calculated using one-way ANOVA multiple-comparison test and set at a P value of <0.05 (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; and ns, P > 0.05).

FIG 4

Serum protection assay by OMV produced under PhoPQ-activating conditions. (A) Survival of WT and ΔpagC Salmonella incubated with 25% NHS in the presence of sterile culture medium, OMV-depleted culture supernatants, or OMV-containing culture supernatants showed that bacteria incubated with OMV-containing culture supernatants from PagC-expressing bacteria (WT, ΔpagC + ppagC, and PhoP^C) were significantly more resistant to serum-mediated attack than bacteria incubated with sterile culture medium or culture supernatant from the ΔpagC mutant. Bacterial survival in sterile culture medium was comparable to the survival in OMV-depleted spent culture supernatants. (B) Survival of WT and ΔpagC Salmonella incubated with 25% NHS in the presence of purified OMVs at indicated concentrations (0 to 10⁵ OMVs), isolated from bacteria grown under PhoPQ-activating (5.8L; solid line) or nonactivating (7.6H; dashed line) conditions. OMVs from bacteria expressing PagC (WT) grown under PhoPQ-activating conditions were more protective than OMVs from bacteria grown under nonactivating conditions. The data represent one of three separate, reproducible experiments, expressed as mean ± SEM. Statistical significance was calculated using Student’s t test and set at a P value of <0.05 (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; and ns, P > 0.05).

FIG 5

Blocking of complement activation by OMVs from PagC-expressing bacteria. (A) Survival of WT Salmonella incubated with 25% NHS treated with 5 mM EGTA or EDTA, in the presence or absence of 1 × 10⁹ purified OMVs from WT or ΔpagC mutant. OMVs from PagC-expressing WT bacteria block the alternative pathway of the complement more efficiently than the OMVs from the ΔpagC mutant. Statistical significance was calculated using Student’s t test. (B) Schematic representation of C3 and its cleavage products C3b and iC3b (molecular weight [MW] of the α and β chain and the cleaved or uncleaved products is indicated in the brackets). (C) WT OMVs (10⁸ to 10¹⁰ OMVs) incubated with 5% NHS, when immunoblotted with anti-C3 antibody, showed bands for C3b inactivation at 68 kDa (iC3bα′) and 46 kDa (iC3bα″) in OMV pellet and supernatant fractions, respectively. No C3b inactivation bands were found when ΔpagC OMVs were incubated with NHS, suggesting that WT OMVs bind to C3b and inactivate C3b into iC3b in a dose-dependent manner. The Western blot image was spliced to remove an empty lane in between pellet and supernatant fractions. The panel on the right represents relative protein band densities (WT versus ΔpagC OMVs) as analyzed by NIH ImageJ software from 3 separate, reproducible experiments, expressed as mean ± SEM. Statistical significance was calculated using Student’s t test. (D) WT or ΔpagC mutant incubated with C3/C4-depleted human serum (gray bars) survived significantly better than the bacteria incubated with NHS (black bars). Presence of OMV-containing culture supernatant (+ OMV) contributed to bacterial survival when incubated with NHS, compared with bacteria incubated in the presence of OMV-depleted culture supernatants (− OMV). (E) Incubation of the ΔpagC Δrck ΔompX ΔpgtE mutant with NHS in the presence of OMV-containing culture supernatant prevented C3b deposition on bacterial cell surfaces. OMVs from the quadruple mutant expressing PagC prevented C3 deposition most significantly. The bar graph below the blot represents relative protein band densities as analyzed by NIH ImageJ software from 3 separate, reproducible experiments, expressed as mean ± SEM. For iC3bα′, relative protein densities were calculated for bacteria incubated in sterile culture medium (−) with no OMVs versus bacteria incubated with OMV-containing spent culture supernatant. Statistical significance was calculated using the one-way ANOVA multiple-comparison test. The data represent one of three separate, reproducible experiments, expressed as mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; and ns, P > 0.05).

FIG 6

Factor H interaction with OMVs. (A) 1 × 10⁹ Purified OMVs from WT, ΔpagC, Δrck, ΔpagC Δrck ΔompX, or ΔpagC Δrck ΔompX ΔpgtE mutant bacteria incubated with purified Factor H (FH; 2 μg) showed FH binding by OMVs from PagC- or Rck-expressing bacteria, as detected by Western blotting using anti-FH antibody. The bar graphs below each blot represent relative protein band density as analyzed by NIH ImageJ software from 3 separate, reproducible experiments, expressed as mean ± SEM. The relative protein density was calculated relative to WT bacteria (left) or the ΔpagC Δrck ΔompX ΔpgtE + pRS1 (empty vector)-containing mutant (right). (B) F-NTA analysis showed a significantly high percentage of OMVs from the PagC- or Rck-expressing quadruple mutant bound to FH compared with OMVs from the pRS1-containing quadruple mutant. The % OMVs bound to FH were calculated as the number of OMVs estimated by F-NTA versus the number of OMVs estimated by scatter-NTA × 100. (C) Incubation of WT, ΔpagC mutant, or Δrck mutant bacteria with normal mouse serum or FH mutant (W1206R) serum. Both PagC and Rck contribute to bacterial survival in serum in an FH-dependent manner. Statistical significance was calculated using Student’s t test (A and C). (D) Far-Western analysis of FH binding to recombinant proteins PagC, Rck, and FimA (1 μg each, indicated by arrowheads) showed purified FH (1 μg/ml) bound only to PagC and Rck, with the former showing greater affinity for FH. FimA being only partially reduced shows two bands in the blot stained by Ponceau S. Molecular weight standards are indicated on the right. The graph on the right represents relative protein densities as analyzed by the NIH ImageJ software, averaging 3 separate, reproducible experiments (mean ± SEM). The protein band densities for Rck and FimA were calculated relative to the PagC protein. Statistical significance was calculated using the one-way ANOVA multiple-comparison test (B and D). The data represent one of three separate, reproducible experiments, expressed as mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; and ns, P > 0.05).