Impact of horizontal gene transfer on emergence and stability of cooperative virulence in Salmonella Typhimurium

Abstract

Intestinal inflammation fuels the transmission of Salmonella Typhimurium (S.Tm). However, a substantial fitness cost is associated with virulence expression. Mutations inactivating transcriptional virulence regulators generate attenuated variants profiting from inflammation without enduring virulence cost. Such variants interfere with the transmission of fully virulent clones. Horizontal transfer of functional regulatory genes (HGT) into attenuated variants could nevertheless favor virulence evolution. To address this hypothesis, we cloned hilD, coding for the master regulator of virulence, into a conjugative plasmid that is highly transferrable during intestinal colonization. The resulting mobile hilD allele allows virulence to emerge from avirulent populations, and to be restored in attenuated mutants competing against virulent clones within-host. However, mutations inactivating the mobile hilD allele quickly arise. The stability of virulence mediated by HGT is strongly limited by its cost, which depends on the hilD expression level, and by the timing of transmission. We conclude that robust evolution of costly virulence expression requires additional selective forces such as narrow population bottlenecks during transmission.

Fig. 1. Virulence can emerge through HGT in a population of cheaters in vivo, but is unstable.

a Experimental system to measure maintenance of cooperative virulence by HGT. Donors contain pVir encoding hilD, but cannot produce a functional TTSS-1 (invG mutant), making them avirulent. Recipients contain all genes for a functional TTSS-1 but do not have a functional copy of hilD (cheaters), preventing TTSS-1 expression and virulence. Transfer of pVir from the donor to the recipient forms a transconjugant that contains both functional TTSS-1-encoding genes and a copy of hilD from pVir, allowing TTSS-1-mediated virulence. Transconjugants can transfer pVir to additional recipients. b–e pVir is transferred to cheater recipients and allows cooperative virulence to emerge. Ampicillin pretreated mice were sequentially infected orally with donors (14028S ΔinvG ΔhilD ΔssaV; Cm^R, Amp^R) harboring pVir^Low (blue; n = 8), pVir^High (orange; n = 11), or P2 lacking hilD (control; white; n = 11), and recipients (14028 S ΔhilD ΔssaV; Kan^R, Amp^R). Each replicate is shown and bars indicate the median. Source data are provided as a Source Data file. b–d Statistics compare pVir^Low (blue asterisks) and pVir^High (orange asterisks) to the control on each day; Kruskal–Wallis test with Dunn’s multiple test correction (p > 0.05 not significant and not indicated, *p < 0.05, **p < 0.01, ***p < 0.001. Dotted line represents the detection limit. b Plasmid transfer was measured by selective and/or replica plating. The proportion of transconjugants is calculated by dividing the transconjugant population by the sum of recipients and transconjugants. c Inflammation was measured by a Lipocalin-2 ELISA on fecal samples. d Total population determined by summing all subpopulations. Donor, recipient, and transconjugant populations are presented in Fig. S3. e Transconjugants carrying pVir^Low (blue) and pVir^High (orange) were analyzed by colony western-blot and compared using a two-tailed Mann–Whitney U test (p > 0.05 not significant and not indicated, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). The percentage of colonies that expressed SipC are reported out of the total transconjugant population. Bars indicate the median. The black dotted line indicates the conservative detection limit, which is dependent on the number of colonies on the plate (values can therefore appear below the detection limit).

Fig. 2. Successful infections in new hosts depend on the proportion of transmitted cooperators.

a Experimental scheme for transmission experiments. Feces from mice in Fig. 1 collected on day 2 and day 10 post infection were suspended in PBS and given to new ampicillin (Amp.) pretreated mice. b–g Mice orally given fecal resuspensions with S.Tm harboring pVir^Low (blue; dark shade for day 2 transmission; light shade for day 10 transmission; n = 8 for both groups) are compared to pVir^High (orange; dark shade for day 2 transmission (n = 11); light shade for day 10 transmission (n = 10)) using a two-tailed Mann–Whitney U test (p > 0.05 (ns), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). All data points are shown and medians are represented by bars. Source data are provided as a Source Data file. b–d Mice given feces from day 2 post infection. e–g Mice given feces from day 10 post infection. b, e Inflammation was quantified using a LCN2 ELISA. The dotted lines indicate the detection limit. c, f The shedding population was enumerated by summing all populations determined by selective plating. Donor, recipient, and transconjugant populations are presented in Fig. S4. d, g MacConkey plates containing colonies of transconjugants were analyzed for expression of SipC as a proxy for TTSS-1 expression using a colony western blot; the percentage of colonies that expressed SipC are reported out of the total transconjugant population. The black dotted line indicates the conservative detection limit for the colony blot, which is dependent on the number of colonies on the plate (values can therefore appear below the detection limit).

Fig. 3. Genetic drift favors cooperative virulence.

Ampicillin pretreated mice were orally infected with evolved transconjugant clones isolated from day 7 or day 10 from mice in Fig. 1. Three cooperating clones (solid circles; TTSS-1⁺ clones) and cheating clones (hollow circles; TTSS-1⁻ clones) were randomly chosen for each of pVir^Low (blue) or pVir^High (orange). Each clone was infected into 2–3 mice (~5 × 10⁷ CFU inoculum), leading to a total of 6–7 mice per group. All clones were whole-genome sequenced (mutations and indels are summarized in Tables S1–4): TTSS-1⁺ pVir^Low (Z2296 (3 mice), Z2306 (2 mice), Z2310 (2 mice); n = 7), TTSS-1⁺ pVir^High (Z2238 (3 mice), Z2246 (2 mice), Z2253 (2 mice); n = 7), TTSS-1^- pVir^Low (Z2298, Z2301, Z2305; 2 mice per clone; n = 6), TTSS-1^- pVir^High (Z2239 (3 mice), Z2243 (2 mice), Z2311 (2 mice); n = 7). All data points are shown and medians are indicated by bars. Comparisons are made between TTSS-1⁺ and TTSS-1^- clones (for each of pVir^Low and pVir^High) using a two-tailed Mann–Whitney U test (p > 0.05 (ns), *p < 0.05, **p < 0.01, ***p < 0.001). Source data are provided as a Source Data file. a Inflammation was quantified using a LCN2 ELISA. The dotted lines indicate the detection limit. b The shedding population was enumerated on MacConkey agar. c MacConkey plates containing colonies were analyzed for expression of SipC as a proxy for TTSS-1 expression using a colony western blot; the percentage of colonies that expressed SipC are reported out of the total transconjugant population. The black dotted line indicates the conservative detection limit for the colony blot, which is dependent on the number of colonies on the plate (values can therefore appear below the detection limit).

Fig. 4. HGT can increase the duration of, but does not stabilize, cooperative virulence.

a Experimental system to determine the role of HGT in the stabilization of cooperative virulence. In both the mobile pVir and non-mobile pVir scenarios, the cooperator contains pVir^Low and a functional TTSS-1, making it virulent. In the mobile scenario, pVir can be transferred to cheaters. In the non-mobile scenario, transfer of pVir is blocked because of an incompatible plasmid, P2, in the cheater strain. b–g Ampicillin pretreated mice were orally infected with ~10² CFU of cheater (14028 ΔhilD ΔssaV; Kan^R, Amp^R) immediately followed by ~10² CFU of pVir^Low donor (14028 ΔhilD ΔssaV pVir; Cm^R encoded on pVir, Amp^R). The donor contained a functional invG allele, making it virulent (ssaV is deleted in both strains). Mice (n = 8 until day 4; n = 5 until day 10 for both groups) were given either a cheater with no plasmid (i.e., same strain used in Fig. 1; mobile pVir; black) or a cheater with P2 (incompatible with pVir; non-mobile pVir; pink). Dotted lines indicate detection limits. Medians are indicated by lines. Two-tailed Mann–Whitney U tests (p > 0.05 (ns), *p < 0.05, **p < 0.01, ***p < 0.001) are used to compare the mobile pVir and non-mobile pVir scenarios on each day. Donor, recipient, and transconjugant populations are presented in Fig. S6. Source data are provided as a Source Data file. b The pVir-bearing population was determined by selective plating on Cm-supplemented MacConkey agar. c the pVir-bearing population is reported as a percentage of the total population. Dashed line indicates 100% plasmid spread. d Inflammation was measured by a Lipocalin-2 ELISA on fecal samples. e Total S.Tm populations determined by the sum of Cm- and Kan-supplemented MacConkey agar. f, g Cm-supplemented MacConkey agar plates containing colonies from feces collected on day 10 post infection were analyzed for SipC expression as a proxy for TTSS-1 expression using a colony western blot and represented as the percentage of colonies that produced SipC are reported out of the total S.Tm population (f) or out of the pVir-containing population for the mobile pVir scenario (g).