Erwinia carotovora Quorum Sensing System Regulates Host-Specific Virulence Factors and Development Delay in Drosophila melanogaster

Abstract

Multihost bacteria have to rapidly adapt to drastic environmental changes, relying on a fine integration of multiple stimuli for an optimal genetic response. Erwinia carotovora spp. are phytopathogens that cause soft-rot disease. Strain Ecc15 in particular is a model for bacterial oral-route infection in Drosophila melanogaster as it harbors a unique gene, evf, that encodes the Erwinia virulence factor (Evf), which is a major determinant for infection of the D. melanogaster gut. However, the factors involved in the regulation of evf expression are poorly understood. We investigated whether evf could be controlled by quorum sensing as, in the Erwinia genus, quorum sensing regulates pectolytic enzymes, the major virulence factors needed to infect plants. Here, we show that transcription of evf is positively regulated by quorum sensing in Ecc15 via acyl-homoserine lactone (AHL) signal synthase ExpI and AHL receptors ExpR1 and ExpR2. We also show that the load of Ecc15 in the gut depends upon the quorum sensing-mediated regulation of evf Furthermore, we demonstrate that larvae infected with Ecc15 suffer a developmental delay as a direct consequence of the regulation of evf via quorum sensing. Finally, we demonstrate that evf is coexpressed with plant cell wall-degrading enzymes (PCWDE) during plant infection in a quorum sensing-dependent manner. Overall, our results show that Ecc15 relies on quorum sensing to control production of both pectolytic enzymes and Evf. This regulation influences the interaction of Ecc15 with its two known hosts, indicating that quorum sensing signaling may impact bacterial dissemination via insect vectors that feed on rotting plants.IMPORTANCE Integration of genetic networks allows bacteria to rapidly adapt to changing environments. This is particularly important in bacteria that interact with multiple hosts. Erwinia carotovora is a plant pathogen that uses Drosophila melanogaster as a vector. To interact with these two hosts, Ecc15 uses different sets of virulence factors: plant cell wall-degrading enzymes to infect plants and the Erwinia virulence factor (evf) to infect Drosophila Our work shows that, despite the virulence factors being specific for each host, both sets are coactivated by homoserine lactone quorum sensing and by the two-component GacS/A system in infected plants. This regulation is essential for Ecc15 loads in the gut of Drosophila and minimizes the developmental delay caused by the bacteria with respect to the insect vector. Our findings provide evidence that coactivation of the host-specific factors in the plant may function as a predictive mechanism to maximize the probability of transit of the bacteria between hosts.

Keywords: Drosophila; Ecc15; bacterial infections; homoserine lactones; host-pathogen interactions; insect development; invertebrate-microbe interactions; quorum sensing.

FIG 1

Signaling pathways regulating PCWDE and evf production in Erwinia spp. At low cell density, when the concentration of AHL signaling molecules is low, ExpR1 and ExpR2 induce transcription of rsmA, repressing expression of both PCWDE and evf. As cell density increases, AHLs accumulate and when the concentration threshold is reached these signal molecules bind to ExpR1 and ExpR2 receptors, inhibiting their DNA binding ability. As a result, rsmA transcription is no longer induced. The GacS/A two-component system is also active at high cell density and promotes transcription of rsmB, a noncoding RNA that has high binding affinity to RsmA and inhibits the remaining available RsmA. Inhibition of RsmA results in increased production of both PCWDE and hor and, consequently, evf, leading to full induction of virulence. We show here that evf is regulated by quorum sensing and the GacS/A system via hor. While evf is not necessary to infect the plant host, we also show that there is coexpression of evf and PCWDE during plant infection. Gray and black lines indicate inactive and active pathways, respectively. Arrows indicate activation, while intersecting lines indicate repression.

FIG 2

Production of pectate lyase and expression of evf are dependent on both quorum sensing and the GAC system. (A) Pectate lyase activity in cell-free supernatants of WT Ecc15 and expI and gacA mutants at 6 h of growth in LB plus 0.4% PGA, n = 10. (B) Potato maceration quantification (grams) in potatoes infected with WT Ecc15 and expI, gacA, and evf mutants, 48 h postinfection, n = 8. (C) Pevf::gfp expression in WT Ecc15, expI and gacA mutants, and expI+AHLs at 6 h of growth in LB + Spec, n = 5. Complementation with AHLs (expI+AHLs) was performed with a mixture of 1 μM 3-oxo-C6-HSL and 3-oxo-C8-HSL. Growth curves of the strains used are shown in Fig. S1. Error bars represent standard deviations of the means. For each panel, results from a representative experiment from three independent experiments are shown (the results from the other two experiments are shown in Fig. S2). Results of statistical analysis taking the data of all three experiments are shown in Fig. S2. a.u., arbitrary units.

FIG 3

evf regulation by quorum sensing is dependent on ExpR receptors and hor. (A) P_evf::gfp expression without (white bars) or with (gray bars) addition of exogenous AHLs in the Ecc15 WT and in expI, expI expR1 expR2, and hor mutants at 6 h of growth in LB + Spec, n = 5. (B) P_evf::gfp expression in the Ecc15 WT strain and expI and gacA mutants containing a plasmid with the P_evf::gfp fusion (white bars) or with both P_lac::hor and P_evf::gfp fusions (gray bars) at 6 h of growth in LB + Spec, n = 5. (C) P_hor::gfp expression in WT Ecc15 and expI and gacA mutants at 6 h of growth in LB + Spec, n = 5. Complementation with AHLs was performed with a mixture of 1 μM 3-oxo-C6-HSL and 3-oxo-C8-HSL. Error bars represent standard deviations of the means. For each panel, results from a representative experiment from three independent experiments are shown (results from the other two experiments are shown in Fig. S5). Results of statistical analysis taking the data of all three experiments are shown in Fig. S5. a.u., arbitrary units.

FIG 4

Ecc15 loads are higher in D. melanogaster larvae orally infected with the WT than in those infected with mutants impaired in evf expression. D. melanogaster L3 stage larvae were infected with WT Ecc15 and evf, expI, and gacA mutants for 30 min and then transferred to fresh media. Following the infection period, CFU levels of Ecc15 were measured at the specified time points. Each dot represents CFU of one single larvae (5 larvae per time point). The time point indicated as 0 h after infection corresponds to 30 min of confined exposure to 200 μl of an OD₆₀₀ of 200. Results from a representative experiment from three independent experiments are shown (results from the other two experiments are shown in Fig. S5). Results of statistical analysis of the comparisons of data from the entire infection period for each condition tested in all three experiments are shown in Fig. S6.

FIG 5

Ecc15 causes a developmental delay in D. melanogaster larvae that is dependent on evf, quorum sensing, and the GAC system. (A and C) L3 stage Drosophila larvae pupariation time after exposure to (A) WT Ecc15 and evf, expI, and gacA mutants or (C) WT Ecc15 overexpressing Evf [p(evf⁺)], compared with noninfected larvae. (B and D) Average developmental time (in hours) with standard deviation (data correspond to the results shown in panels A and C, respectively). Overexpression of Evf was lethal, as larva exposed to WT Ecc15 overexpressing Evf (C) died without reaching the pupa stage. NA, not applicable. Results from a representative experiment of three independent experiments are shown (results from the other two experiments are shown in Fig. S7). The statistical groups represented in panels B and D were determined using a linear mixed-effect model taking into consideration the data from the three experiments. A Tukey HSD test was applied for multiple comparisons using the estimates obtained from the model.

FIG 6

Coactivation of evf and pelA expression in plant infections. Potatoes were infected for 6, 24, or 48 h with a WT Ecc15 strain or an expI mutant carrying an evf or pelA gfp reporter plasmid with a constitutive mCherry promoter. At each specified time point, pelA or evf expression, CFU levels, and weight of macerated plant tissue were determined. (A and B) P_evf::gfp expression (A) and P_pelA::gfp expression (B) in potatoes infected with WT Ecc15 or expI mutant. (C) CFU of WT Ecc15 or expI mutant expressing mCherry and carrying the P_evf::gfp (green boxes) or the P_pelA::gfp (red boxes) reporter plasmids. (D) Potato maceration (in grams) in potatoes infected with WT Ecc15 or expI mutant. Each dot represents an independent potato, n = 5. Results are from a representative experiment of two independent experiments (results from the second experiment are shown in Fig. S8). Results of statistical analysis of comparisons of data from the entire infection period for each condition tested using the data from both experiments are shown in Fig. S8.