A type VI secretion system of Pseudomonas aeruginosa targets a toxin to bacteria

Abstract

The functional spectrum of a secretion system is defined by its substrates. Here we analyzed the secretomes of Pseudomonas aeruginosa mutants altered in regulation of the Hcp Secretion Island-I-encoded type VI secretion system (H1-T6SS). We identified three substrates of this system, proteins Tse1-3 (type six exported 1-3), which are coregulated with the secretory apparatus and secreted under tight posttranslational control. The Tse2 protein was found to be the toxin component of a toxin-immunity system and to arrest the growth of prokaryotic and eukaryotic cells when expressed intracellularly. In contrast, secreted Tse2 had no effect on eukaryotic cells; however, it provided a major growth advantage for P. aeruginosa strains, relative to those lacking immunity, in a manner dependent on cell contact and the H1-T6SS. This demonstration that the T6SS targets a toxin to bacteria helps reconcile the structural and evolutionary relationship between the T6SS and the bacteriophage tail and spike.

Figure 1. Overview and results of an MS-based screen to identify H1-T6SS substrates

(A) Gene organization of P. aeruginosa HSI-I. Genes manipulated in this work are shown in color. (B) Activity of the H1-T6SS can be modulated by deletions of pppA and clpV1. Western blot analysis of Hcp1–V in the cell-associated (Cell) and concentrated supernatant (Sup) protein fractions from P. aeruginosa strains of specified genetic backgrounds. The genetic background for the parental strain is indicated below the blot. An antibody directed against RNA polymerase α (α-RNAP) is used as a loading control in this and subsequent blots. (C) Deletion of pppA causes increased p-Fha1–V levels. p-Fha1–V is observed by Western blot as one or more species with retarded electrophoretic mobility. (D) Spectral count ratio of C1 proteins detected in R1 and R2 of the comparative semi-quantitative secretome analysis of ΔpppA and ΔclpV1. The position of Hcp1 in both replicates is indicated. Proteins within the dashed line have SC ratios of < 2-fold and constitute 89% of C1 proteins.

Figure 2. Two VgrG-family proteins are regulated by retS and secreted in an H1-T6SS-dependent manner

(A) Overview of genetic loci encoding C2 proteins identified in R1 and R2 (green). RetS regulation of each ORF as determined by Goodman et al. is provided (Goodman et al., 2004). Genes not significantly regulated by RetS are filled grey. (B and C) Western blot analysis demonstrating that secretion of VgrG1–V (B) and VgrG4–V (C) is triggered in the ΔpppA background and is H1-T6SS (clpV1)-dependent. All blots are against the VSV-G epitope (α-VSV-G).

Figure 3. The Tse proteins are tightly regulated H1-T6SS substrates

(A) Tse secretion is under tight negative regulation by pppA and is H1-T6SS-dependent. Western analysis of Tse proteins expressed with C-terminal VSV-G epitope tag fusions from pPSV35 (Rietsch et al., 2005). Unless otherwise noted, all blots in this figure are α-VSV-G. (B) H1-T6SS-dependent secretion of chromosomally-encoded Tse1–V measured by Western blot analysis. (C) Hcp1 secretion is independent of the tse genes. Western blot analysis of Hcp1 localization in control strains or strains lacking both vgrG1 and vgrG4, or the three tse genes. (D) The tse genes are not required for formation of a critical H1-T6S apparatus complex. Chromosomally-encoded ClpV1–GFP localization in the specified genetic backgrounds measured by fluorescence microscopy. TMA-DPH is a lipophilic dye used to visualize the position of cells. (E) The production and secretion of Tse proteins is dramatically increased in ΔretS. Western blot analysis of Tse levels from strains containing chromosomally-encoded Tse-VSV-G epitope tag fusions prepared in the wild-type or ΔretS backgrounds. Note – under conditions used to observe the high levels of Tse secretion in ΔretS, secretion cannot be visualized in ΔpppA as was demonstrated in (B).

Figure 4. The Tse2 and Tsi2 proteins are a toxin-immunity module

(A) Tse2 is toxic to P. aeruginosa in the absence of Tsi2. Growth of the indicated P. aeruginosa strains containing either the vector control (−) or vector containing tse2 (+) under non-inducing (− IPTG) or inducing (+ IPTG) conditions. (B) Tse2 and Tsi2 physically associate. Western blot analysis of samples before (Pre) and after (Post) α-VSV-G immunoprecipitation from the indicated strain containing a plasmid expressing tsi2 (control) or tsi2–V. The glycogen synthase kinase (GSK) tag was used for detection of Tse2 (Garcia et al., 2006).

Figure 5. Heterologously expressed Tse2 is toxic to prokaryotic and eukaryotic cells

(A) Tse2 is toxic to S. cerevisiae. Growth of S. cerevisiae cells containing a vector control or a vector expressing the indicated tse under non-inducing (Glucose) or inducing (Galactose) conditions. (B) Tsi2 blocks the toxicity of Tse2 in S. cerevisiae. Growth of S. cerevisiae harboring plasmids with the indicated gene(s), or empty plasmid(s), under non-inducing or inducing conditions. (C, D and E) Transfected Tse2 has a pronounced effect on mammalian cells. Flow cytometry (C) and fluorescence microscopy (D) analysis of GFP reporter co-transfection experiments with plasmids expressing the tse genes or tsi2. The percentage of rounded cells following the indicated transfections was determined (E) (n > 500). Control (ctrl) experiments contained only the reporter plasmid. Bar graphs represent the average number from at least three independent experiments (± SEM). (F and G) Expression of tse2 inhibits the growth of E. coli (F) and B. thailandensis (G). E. coli (F) and B. thailandensis (G) were transformed with expression plasmids regulated by inducible expression with IPTG (F) or rhamnose (G), respectively, containing no insert, tse2, or both the tse2 and tsi2 loci. Growth on solid medium was imaged after one (F) or two (G) days of incubation.

Figure 6. Immunity to Tse2 provides a growth advantage against P. aeruginosa strains secreting the toxin by the H1-T6SS

(A) Tse2 secreted by the H1-T6SS of P. aeruginosa does not promote cytotoxicity in HeLa cells. LDH release by HeLa cells following infection with the indicated P. aeruginosa strains or E. coli. P. aeruginosa strain PA14 and E. coli were included as highly cytotoxic and non-cytotoxic controls, respectively. The PAO1 strain used by our laboratory displays decreased cytotoxicity relative to other laboratory P. aeruginosa strains due to low expression of the type III secretion system (Rietsch et al., 2004). (B and C) Results of in vitro growth competition experiments in liquid medium (B) or on a solid support (B and C) between P. aeruginosa strains of the indicated genotypes. The parental strain is ΔretS. The ΔclpV1 and Δtsi2-dependent effects were complemented as indicated by + clpV1 and + tsi2, respectively (see methods).