Ralstonia solanacearum requires PopS, an ancient AvrE-family effector, for virulence and To overcome salicylic acid-mediated defenses during tomato pathogenesis.

UNLABELLED: During bacterial wilt of tomato, the plant pathogen Ralstonia solanacearum upregulates expression of popS, which encodes a type III-secreted effector in the AvrE family. PopS is a core effector present in all sequenced strains in the R. solanacearum species complex. The phylogeny of popS mirrors that of the species complex as a whole, suggesting that this is an ancient, vertically inherited effector needed for association with plants. A popS mutant of R. solanacearum UW551 had reduced virulence on agriculturally important Solanum spp., including potato and tomato plants. However, the popS mutant had wild-type virulence on a weed host, Solanum dulcamara, suggesting that some species can avoid the effects of PopS. The popS mutant was also significantly delayed in colonization of tomato stems compared to the wild type. Some AvrE-type effectors from gammaproteobacteria suppress salicylic acid (SA)-mediated plant defenses, suggesting that PopS, a betaproteobacterial ortholog, has a similar function. Indeed, the popS mutant induced significantly higher expression of tomato SA-triggered pathogenesis-related (PR) genes than the wild type. Further, pretreatment of roots with SA exacerbated the popS mutant virulence defect. Finally, the popS mutant had no colonization defect on SA-deficient NahG transgenic tomato plants. Together, these results indicate that this conserved effector suppresses SA-mediated defenses in tomato roots and stems, which are R. solanacearum's natural infection sites. Interestingly, PopS did not trigger necrosis when heterologously expressed in Nicotiana leaf tissue, unlike the AvrE homolog DspEPcc from the necrotroph Pectobacterium carotovorum subsp. carotovorum. This is consistent with the differing pathogenesis modes of necrosis-causing gammaproteobacteria and biotrophic R. solanacearum. IMPORTANCE: The type III-secreted AvrE effector family is widely distributed in high-impact plant-pathogenic bacteria and is known to suppress plant defenses for virulence. We characterized the biology of PopS, the only AvrE homolog made by the bacterial wilt pathogen Ralstonia solanacearum. To our knowledge, this is the first study of R. solanacearum effector function in roots and stems, the natural infection sites of this pathogen. Unlike the functionally redundant R. solanacearum effectors studied to date, PopS is required for full virulence and wild-type colonization of two natural crop hosts. R. solanacearum is a biotrophic pathogen that causes a nonnecrotic wilt. Consistent with this, PopS suppressed plant defenses but did not elicit cell death, unlike AvrE homologs from necrosis-causing plant pathogens. We propose that AvrE family effectors have functionally diverged to adapt to the necrotic or nonnecrotic lifestyle of their respective pathogens.

INTRODUCTION

Plant-pathogenic bacteria cause destructive diseases that limit crop production worldwide. Many Gram-negative phytopathogenic bacteria use a type III secretion system (T3SS) to inject effector proteins into host cells. These generally modulate host immunity and physiology for pathogenesis (1–4). Individual effectors rarely contribute measurably to virulence but rather function as a consortium (5). Because of their redundancy and subtle biological activities, the functions of individual type III (T3) effectors remain largely unknown.

Plant immune systems have evolved complex signaling responses to defend against microbial invasion. Plants use specific protein receptors to detect conserved features of pathogen products such as flagellin, chitin, lipopolysaccharide, and elongation factor Tu (EF-Tu); perception of these microbe-associated molecular patterns (MAMPs) triggers basal immunity (1). In addition, plant R genes specifically recognize pathogen effectors. The plant hormone salicylic acid (SA) is a major defense signal molecule (6), and upon recognition of a pathogen, SA production induces basal immune responses such as callose deposition (7, 8). Accumulation of SA in plants also induces expression of pathogenesis-related (PR) defense genes to resist microbial infection and sometimes triggers rapid host cell death (9, 10). Studies of a few genera in the gammaproteobacteria have revealed how pathogen T3 effectors disrupt immune signaling and suppress SA-mediated defenses (5, 11, 12).

The AvrE family of effectors is well conserved across agriculturally important phytobacteria, including enterobacteria, xanthomonads, and pseudomonads (gammaproteobacteria) and Ralstonia spp. (betaproteobacteria) (13). Effectors in this family, which includes AvrE, DspE, and WtsE, induce host cell death and suppress defense signaling (12–14). AvrE from Pseudomonas syringae pv. tomato DC3000 and its ortholog DspE from Erwinia amylovora promote pathogen growth and overcome plant immunity by inhibiting SA-mediated defense responses (12). Despite their broad relevance to the interactions of plant-pathogenic bacteria with their hosts (13), little is known about AvrE-like effectors outside plant-pathogenic gammaproteobacteria.

The bacterial wilt pathogen R. solanacearum is responsible for diseases of many crops in tropical and subtropical climates worldwide. This bacterium enters plant roots from the soil and colonizes the host vasculature, which eventually leads to wilt and plant death (15). R. solanacearum requires a T3SS for root and stem invasion and colonization (16), and T3SS-deficient strains are essentially unable to wilt host plants (16). The R. solanacearum genome encodes an extensive effector repertoire (2, 17). Mutants lacking individual effectors generally do not have virulence defects (18), likely because the effectors have redundant functions (3, 5). The defense-suppressing functions of the individual effectors during the infection cycle remain unknown.

We previously used gene expression analysis to define the R. solanacearum in planta transcriptome, the set of bacterial genes expressed during growth in wilting tomato plant stems (19). An orthologous gene encoding an AvrE-family effector was expressed in planta in two ecologically and phylogenetically distinct strains. This locus (RRSL_03375 in strain UW551 and RSp1281 in strain GMI1000) encodes a secreted T3 effector in the AvrE/DspE/HopR protein family (20), herein named PopS. Relative to expression in rich culture medium, UW551 and GMI1000 upregulate popS in planta 14- and 8-fold, respectively (19). Expression of popS is dependent on HrpB, the transcriptional activator of the T3SS and its effectors (21–23). Most effector genes are upregulated in planta via HrpB (19, 21, 24).

This study characterizes the role of PopS throughout the tomato infection process. We determined that this effector, which has ancient roots in the R. solanacearum species complex, is required for normal host colonization and virulence on multiple Solanum spp. crop hosts. PopS was dispensable for virulence on a weed, Solanum dulcamara, suggesting that it has species-specific virulence activity within the genus Solanum. PopS forms a unique clade in the AvrE family of effectors. Although it is highly divergent from its closest orthologs, we found that PopS retains the function of suppressing SA-mediated plant defenses. In contrast, PopS did not cause plant cell death or necrosis as do the AvrE-family proteins of necrosis-causing pathogens, such as DspE and WtsE from Pectobacterium carotovorum subsp. carotovorum and Pantoea stewartii subsp. stewartii. Together, our phylogenetic, virulence, and gene expression data suggest that PopS suppresses SA-mediated host defenses but lacks the ability to cause cell death, which may have helped this pathogen adapt to its nonnecrotic lifestyle.

RESULTS

PopS is a conserved, vertically inherited T3 effector in the R. solanacearum species complex.

The AvrE effector family is widely present among plant-pathogenic bacteria (13). PopS most closely resembles HopR in Pseudomonas spp. and Xanthomonas spp. (also known as XopAM; http://www.xanthomonas.org/t3e.html) (13), which were, respectively, 25 to 26% and 27 to 28% identical to PopS at the amino acid level. To understand the phylogenetic relationships among these proteins, we developed a maximum-likelihood phylogenetic tree in MEGA5 based on protein sequences of available PopS, AvrE, DspE, and HopR effector orthologs (Fig. 1; also, see Table S1 in the supplemental material) (25). The resulting robust tree, based on 58 amino acid sequences, demonstrated that PopS forms a unique clade in the AvrE/DspE/HopR effector family and that this large family of effectors forms three distinct clusters: AvrE/DspE, HopR, and PopS (Fig. 1A).

FIG 1 

PopS, an ancient core T3 effector in plant-pathogenic Ralstonia species, forms a distinct clade of the AvrE/DspE/HopR effector family. Phylogenetic trees were based on comparative analysis of whole genomes of 11 representative sequenced strains. (A) PopS, HopR (XopAM) from Xanthomonas spp. and Pseudomonas syringae pathovars, AvrE from P. syringae pathovars, and DspE/WtsE from enterobacteria; (B) PopS from sequenced plant-pathogenic Ralstonia; (C) R. solanacearum species complex. For panels A and B, amino acid sequences were aligned using CLUSTAL-W, from which a maximum-likelihood phylogenetic tree was created with MEGA5. The percentage of replicate trees in which the individual orthologs clustered together in the bootstrap test (200 replicates) is noted at each branch. For panel C, a phylogenetic tree of the R. solanacearum species complex was derived from a MUM index (MUMi) distance matrix of whole-genome sequences using the neighbor-joining clustering method.

The R. solanacearum species complex is divided into four subgroups called phylotypes (26). A single copy of popS is found in the genomes of all 12 sequenced strains in the R. solanacearum species complex, which includes the blood disease bacterium (BDB) and the clove pathogen Ralstonia syzygii, both fastidious pathogens with limited host ranges (17, 27–33). A maximum-likelihood phylogenetic tree of genes encoding the R. solanacearum AvrE-like proteins indicated that popS sequences share the phylogeny of the species complex as a whole (Fig. 1B).

To test the hypothesis that popS has been vertically inherited over the evolution of the R. solanacearum species complex, we developed a neighbor-joining tree based on a MUMi distance matrix for whole-genome phylogenetic analysis (Fig. 1C) (17, 34). The MUMi distance matrix and phylogenetic tree of the sequenced plant-pathogenic Ralstonia strains mirror the previously determined phylogeny of the species complex (17). Moreover, the phylogenetic tree of PopS itself mirrors the whole-genome MUMi tree of the species complex (Fig. 1B and C) (26). There is no indication of interstrain movement of popS via horizontal transfer, suggesting that PopS is an ancient core effector in the species complex. This conserved, vertically inherited gene is potentially useful for typing strains and identifying their phylogenetic positions in the R. solanacearum species complex.

R. solanacearum T3 effector PopS is required for full virulence on several hosts.

To study the virulence function of this effector, we disrupted popS in strain UW551 (phylotype II, sequevar 1) via allelic exchange to create strain UW551 popS::Kmr (referred to here as the popS mutant). This mutant grew indistinguishably from the wild type (WT) in culture medium (data not shown), indicating that popS is not required for in vitro growth. The popS mutant retained wild-type ability to grow on sucrose as the sole carbon source, indicating that insertion of the Kmr cassette did not disrupt expression of the scrK sucrose kinase gene (RRSL_03374) immediately downstream of popS. Quantitative reverse transcriptase PCR (qRT-PCR) analysis revealed that UW551 WT, but not the popS mutant, accumulates popS transcript when grown in minimal medium, confirming that popS is not expressed in the mutant strain (data not shown). To test the hypothesis that PopS contributes to bacterial wilt virulence, we used a naturalistic soil soak virulence assay to compare wilt disease progress of UW551 WT and the popS mutant on susceptible and moderately resistant tomato (Solanum lycopersicum cv. Bonny Best and H7996, respectively), susceptible potato (S. tuberosum cv. Russet Norkotah), and a natural weed host, S. dulcamara (bittersweet nightshade) (19). Briefly, pots containing unwounded plants were soaked with bacterial suspensions, and disease progress was rated daily. The popS mutant was delayed in virulence on potato (P < 0.05; repeated-measures analysis of variance [ANOVA]) and on both susceptible and resistant tomato (P < 0.05 and P < 0.005, respectively; repeated-measures ANOVA) (Fig. 2A, B, and C). The popS mutant had a larger virulence defect on the moderately resistant H7996, suggesting that PopS plays a larger role in virulence on resistant plants (Fig. 2A and B).

FIG 2 

Virulence of UW551 popS mutant on diverse plant hosts. R. solanacearum wild-type strain UW551 (black) and the popS mutant (red) were inoculated by pouring a bacterial suspension onto unwounded roots of wilt-susceptible tomato (A) and wilt-resistant tomato (B) (Bonny Best and H7996, respectively), potato (cv. Russet Norkotah) (C), and Solanum dulcamara (bittersweet nightshade) (D). Symptoms were rated daily using a disease index scale of 0 to 4 (0, healthy; 1, 1 to 25% of leaves wilted; 2, 26 to 50% of leaves wilted; 3, 51 to 75% of leaves wilted; 4, 76 to 100% of leaves wilted). Each point represents the mean disease index from three independent experiments, each containing 16 plants per treatment (A, B, and D), or one experiment (C). The popS mutant was significantly less virulent than the wild type on susceptible (P < 0.005; repeated-measures ANOVA) and resistant tomato and potato (P < 0.05) but not on bittersweet nightshade (P = 0.2)

In contrast, loss of popS did not affect R. solanacearum virulence on S. dulcamara, an epidemiologically important weed host (35) (P = 0.2; repeated-measures ANOVA) (Fig. 2D). Lowering the assay temperature from a tropical 28°C to 24°C, which might favor this temperate host plant, did not change this result (data not shown), further evidence that popS is dispensable for wilt on S. dulcamara. Thus, this effector has species-specific activity, since it was necessary for wild-type disease progress on crop hosts but not on a related weed host.

PopS contributes to colonization of susceptible and resistant tomato stems.

To dissect the mechanisms by which PopS contributes to R. solanacearum virulence, we compared the rates at which strain UW551 WT and the popS mutant wilted and colonized the stems of wilt-resistant tomato plants. Virulence and colonization rates for a completely T3SS-deficient hrcC mutant were also measured. To distinguish the stem colonization process from root invasion, tomato stems were directly inoculated through a cut petiole with WT UW551, the popS mutant, or the hrcC mutant, and bacterial colonization was quantified over time.

After direct petiole inoculation, the popS mutant was slightly delayed in virulence compared to WT UW551 (Fig. 3A) (P < 0.001; repeated-measures ANOVA). The popS mutant also colonized resistant H7996 tomato stems significantly more slowly than its wild-type parent (P < 0.03; Mann-Whitney test), although its population size reached wild-type levels by 96 h postinoculation (Fig. 3). Complementing the popS mutant by adding a single copy of popS under the control of its native promoter restored the ability of the popS mutant to both wilt and colonize tomato stems (P < 0.05; Mann-Whitney test) (Fig. 3). These results suggest that PopS is required for bacterial success in planta after the early stages of root infection. This result is congruent with a previous observation that a popS mutant of phylotype I strain GMI1000 had reduced fitness in eggplant leaves (36), although there are significant biological differences between the apoplast and xylem tissue.

FIG 3 

R. solanacearum requires the T3SS and effector PopS for virulence and colonization of tomato stems. (A) Stems of the resistant tomato strain H7996 were inoculated through cut petioles with 40,000 cells of WT UW551 (open black circles), a UW551 popS mutant (solid red squares), a UW551 popS mutant plus miniTn7T::popS (UW551 popS-complemented strain; open red squares), and a UW551 hrcC mutant (open black triangles). Symptoms were rated daily using a disease index scale of 0 to 4 (0, healthy; 1, 1 to 25% of leaves wilted; 2, 26 to 50% of leaves wilted; 3, 51 to 75% of leaves wilted; 4, 76 to 100% of leaves wilted). Each point represents the average disease index (n = 10). The popS and hrcC mutants were significantly different from the WT and popS-complemented strains (P < 0.001; repeated-measures ANOVA). (B) H7996 stems were inoculated through cut petioles with 40,000 cells of WT UW551 (white), a UW551 popS mutant (gray), or a UW551 hrcC mutant (black). Values are average bacterial population sizes (CFU/g stem) of four plants, which were determined at 24, 48, 72, and 96 hpi. Error bars indicate standard errors; asterisks represent statistically significant differences between wild-type UW551 and the popS or hrcC mutant (Mann-Whitney test). (C) Plant stems from susceptible tomato (cv. Bonny Best) were inoculated through cut petioles with 40,000 cells of WT UW551 (white), the popS mutant (gray), the popS-complemented strain (gray with black lines), or the hrcC mutant (black). Columns represent the average bacterial population sizes (CFU/g stem) of four to five individual plants, which were determined 48 h postinoculation. Error bars indicate standard errors; asterisks represent statistically significant differences (Mann-Whitney test) between the WT and the popS or hrcC mutant or the popS-complemented strain.

As expected (16), the T3SS-deficient hrcC mutant was avirulent and did not effectively colonize either susceptible or resistant tomato stems (Fig. 3). The hrcC mutant never reached populations greater than 1.5 × 107 CFU/g stem on either host. Population sizes of the hrcC mutant declined gradually over the 4 days of the assay to 1.6 × 105 CFU/g stem.

The popS mutant induced higher SA defenses in plant roots.

Following infection by pathogens, plant tissues accumulate SA, which induces expression of several PR defense genes (6, 9). Specifically, tomato plants upregulate the SA-mediated PR genes PR-1a and PR-1b in response to infection by R. solanacearum (37). Effectors AvrE in P. syringae pv. tomato DC3000 and DspE in E. amylovora both suppress plant defenses mediated by SA (12). Because UW551 PopS shares 23% amino acid identity with DspE and AvrE, we tested the hypothesis that it similarly suppresses SA-mediated host defense gene expression.

We measured expression of PR-1a and PR-1b in roots of moderately resistant H7996 tomato inoculated with GMI1000, UW551, the UW551 popS mutant, or a water control. Twenty-four hours postinoculation, plants inoculated with wild-type strains GMI1000 or UW551, respectively, increased expression of PR-1a by 2.4- and 4.9-fold and PR-1b by 2.5- and 3.2-fold (Fig. 4A and B). This is consistent with our previous finding that UW551 triggers a faster response in H7996 than GMI1000 (37). Plants inoculated with the UW551 popS mutant had much higher levels of PR-1a (15.8-fold increase) and PR-1b (13.3-fold increase) than those inoculated with WT (Fig. 4A and B). This result indicates that PopS functions to suppress expression of host plant SA-mediated defense genes. Complementation of the popS mutant with the wild-type popS locus restored the ability of the mutant to suppress tomato SA-mediated defenses (see Fig. S1 in the supplemental material).

FIG 4 

PopS is required to overcome SA-mediated defense induction. (A and B) Expression of tomato SA-induced PR defense genes increases in response to a popS mutant of R. solanacearum. Quantitative reverse transcriptase PCR was used to measure expression of plant defense genes in roots of the resistant tomato strain H7996 24 h after inoculation with R. solanacearum phylotype I strain GMI1000, phylotype II strain UW551 (WT), or the UW551 popS mutant. Expression of PR-1a (A) and PR-1b (B) was normalized to that of the tomato GAPDH gene, and the change in expression was determined using the ΔΔC method comparing pathogen-treated plants to water-inoculated control plants. Results reflect two replicates, each including 10 to 13 pooled roots per treatment; error bars indicate standard errors. (C) Expression of tomato defense genes PR-1a, PR-1b, ACO5, and pin2 was measured by qRT-PCR 6 h after soil soak treatment of Money-maker tomato plants with 0.75 mM SA; RNA was extracted from midstem tissue, and expression levels are shown relative to those of water-treated controls. Results shown are the averages of two replicates, each including 5 pooled stem samples per treatment; bars indicate standard errors. (D and E) Treating plants with SA exacerbates the virulence delay of the popS mutant. Average symptom development of susceptible tomato plants (cv. Bonny Best) that were soil soak inoculated with approximately 1 × 108 CFU/g soil of R. solanacearum strain UW551 (WT) (open bars) or the UW551 popS mutant (filled bars). Six hours preinoculation, roots of the plants were drenched with either water (D) or 0.75 mM sodium salicylate (E) (10 plants per strain per treatment). A representative of two replicates is shown. By the end of the assay, all plants treated with water and WT UW551 or the popS mutant or with SA and WT UW551 were completely wilted, and 40% of plants treated with SA and the UW551 popS mutant were asymptomatic (P = 0.0336; Student’s t test). Of these, 75% were colonized with >1010 CFU/g and 25% contained no detectable bacteria. (F and G) Stems of susceptible cv. Money-maker tomato plants (F) or a SA-deficient NahG transgenic derivative of Money-maker (G) were inoculated through cut petioles with 40,000 cells of wild-type R. solanacearum strain UW551 (white), a popS effector mutant (gray), or a T3SS-deficient hrcC mutant (black). Columns represent the average bacterial population sizes (CFU/g stem) of 5 plants per treatment per time point, determined by dilution plating ground stem tissue 24 and 48 hpi; error bars indicate standard errors. Asterisks represent statistically significant differences between wild-type UW551 and the popS mutant or wild-type UW551 and the hrcC mutant (Mann-Whitney test).

SA-treated tomato plants have increased resistance to a popS mutant.

Because roots upregulated PR-1a and PR-1b in response to the UW551 popS mutant and SA induces tomato PR defense gene expression (6, 38, 39), we predicted that pretreating tomato plants with SA would specifically increase their resistance to the UW551 popS mutant. We primed the SA defenses by soaking the soil of unwounded susceptible tomato plants (cv. Bonny Best) with 7.5 mM sodium salicylate (for an estimated soil concentration of 0.75 mM SA/g soil) 6 h before inoculating the plants with either UW551 WT or the popS mutant. As predicted, plants pretreated with 7.5 mM sodium salicylate upregulated the SA-triggered defense genes PR-1a and PR-1b relative to water-treated control plants (Fig. 4C). SA treatment did not trigger expression of ethylene- or jasmonic acid-dependent defense genes ACO5 or pin2, respectively (Fig. 4C), suggesting that PR-1a and PR-1b induction is specific to SA.

SA treatment delayed wilt symptom development in plants inoculated with UW551 WT by 2 days compared to water-treated controls (Fig. 4D and E). By the end of the assay, SA-treated tomato plants inoculated with WT UW551 wilted all tomato plants (Fig. 4E). This demonstrates that SA triggers defenses that increase plant resistance to R. solanacearum.

Interestingly, pretreatment with SA significantly exacerbated the popS mutant virulence defect. SA-treated plants that were inoculated with the popS mutant never reached WT levels of disease (P = 0.0336; Student’s t test) (Fig. 4E). In fact, 40% of SA-treated tomato plants inoculated with the UW551 popS mutant remained asymptomatic. Quantification of bacterial populations in these plants showed that three of the four asymptomatic plants harbored large R. solanacearum populations (average, 1010 CFU/g stem) but the remaining plant contained no detectable R. solanacearum cells. This observation was consistent across replicates (data not shown). In response to the popS mutant, SA-primed roots had decreased rates of initial stem infection and also delayed symptom development. These SA treatment experiments offer further evidence that R. solanacearum uses PopS to overcome SA-induced defenses.

SA-deficient NahG tomato plants restore the colonization defect of the popS mutant.

If the function of PopS is to repress SA-mediated defenses, then reduced levels of SA in planta should allow a popS mutant to be more successful. We tested this hypothesis using transgenic Money-maker tomato plants expressing nahG, which encodes a bacterial salicylate hydroxylase that degrades salicylic acid and reduces SA-mediated defenses (40–42). We measured growth of WT UW551, the popS mutant, and the T3 secretion-deficient hrcC mutant in petiole-inoculated stems of wilt-susceptible cv. Money-maker and an isogenic SA-deficient NahG transgenic line. Both the popS and hrcC mutants were significantly delayed in colonization of nontransgenic Money-maker (P < 0.05 and P = 0.004, respectively; Mann-Whitney test) (Fig. 4F); after 48 h, the popS mutant grew to 5.6 × 108 CFU/g stem, compared to 1.5 × 109 CFU/g stem for the wild-type strain (Fig. 4F). These results demonstrated that Money-maker and the susceptible cultivar Bonny Best respond similarly to these R. solanacearum strains (Fig. 3C). The hrcC mutant grew equally poorly in both tomato lines, indicating that absence of SA alone is not enough to restore the stem growth defect of a completely T3SS-deficient strain (Fig. 4F and G). However, the UW551 popS mutant grew as well as its wild-type parent in the NahG tomato plant stems (Fig. 4G). Thus, an SA-deficient plant host could restore the popS mutant’s colonization rate to wild-type levels, offering independent evidence that a direct or indirect function of PopS is to suppress SA-mediated plant defenses.

PopS does not elicit cell death in Nicotiana benthamiana.

Some AvrE orthologs possess not only the ability to suppress SA-mediated defenses but also to cause cell death when they are expressed transiently in leaf tissue of Nicotiana benthamiana (14, 43). AvrE-like proteins contain conserved WXXXE motifs; at least two of these motifs are required to trigger cell death or for virulence (14, 43). A multiple alignment of PopS and other AvrE orthologs revealed that PopS contains all conserved tryptophans shown to be important for function in other AvrE family members (Fig. 5A) (14, 43). To determine if R. solanacearum PopS elicits plant cell death, we transiently expressed a C-terminally hemagglutinin (HA)-tagged PopS (PopS-HA) via Agrobacterium tumefaciens leaf infiltration of both tobacco and N. benthamiana. As a positive control, we infiltrated leaves to transiently express cell death-eliciting DspE from the necrotrophic soft rot bacterium P. carotovorum subsp. carotovorum (DspE) (44). As expected, DspE caused visible cell death at 24 hours postinfection (hpi) and complete tissue collapse at 48 hpi, but PopS did not elicit detectable cell death in N. benthamiana or Nicotiana tabacum (Fig. 5B, C, and D). An N- and C-terminally green fluorescent protein (GFP)-tagged version of PopS also did not elicit cell death (data not shown). Expression of PopS-HA in N. benthamiana was verified with Western blot analysis (Fig. 5E).

FIG 5 

Transient expression of PopS in N. benthamiana leaves did not induce cell death. (A) Sequence alignment based on T-Coffee analysis of AvrE orthologs (PopS, AvrE, WtsE, and DspE). AvrE ortholog sequences from strains (NCBI sequence reference) analyzed include R. solanacearum UW551 (ZP_00944047.1), P. syringae pv. tomato DC3000 (NP_791204.1), P. stewartii subsp. stewartii (AAG01467.2), P. carotovorum subsp. carotovorum WPP14 (ZP_03833468.1). Conserved tryptophans known to be important for virulence or cell death activity (14, 43) are highlighted in red. (B to D) Agrobacterium tumefaciens-mediated transient expression in N. tabacum (B and C) and N. benthamiana (D). Leaves were infiltrated with A. tumefaciens pGWB14::popS (PopS-HA), A. tumefaciens pGWB2::dspE (DspE from P. carotovorum subsp. carotovorum [Pcc] WPP14) as a positive control, A. tumefaciens pGWB14 (empty vector control), or buffer as a negative control (neg.). Plant symptoms were imaged 48 h postinoculation. In panel C, trypan blue staining shows cell death caused by DspE. Each infiltration was repeated for at least three biological replicates. (E) Western blot analysis of PopS-HA (pGWB14::popS) or negative control (pGWB14; empty) from N. benthamiana leaf tissue.

DISCUSSION

It is well established that T3-secreted effectors are essential for R. solanacearum virulence (16), but the biological roles of specific effectors remain underexplored. In this study, we characterized the function of PopS, an AvrE family effector that is present throughout the R. solanacearum species complex. This effector family is widely conserved among plant-pathogenic bacteria, but its members make various contributions to pathogenesis (12–14, 45). In enteric plant pathogens such as E. amylovora and P. carotovorum, disrupting dspE renders the pathogen completely avirulent (44–46). In contrast, avrE mutants of P. syringae pv. tomato have no detectable colonization or virulence defects, although AvrE apparently works with other effectors, such as HopM1, to suppress host immunity and facilitate pathogenesis (12). PopS falls in the middle of this functional spectrum, because popS mutants are significantly delayed in virulence and plant colonization but can still cause bacterial wilt disease. The virulence and colonization defects of the popS mutant suggest that none of R. solanacearum’s more than 70 putative effectors is fully redundant with PopS activity (2, 22, 47). Nonetheless, a completely T3SS-deficient hrcC mutant was much less able to colonize plants than the popS mutant, confirming that additional T3-secreted effectors contribute to this process. As shown for P. syringae pv. tomato, multiple effector polymutants may identify those effectors that promote colonization and wilt in the absence of PopS (5, 13). Overall, the AvrE family’s wide conservation and consistent role in virulence suggest that this effector has ancient origins in the evolutionary history of bacterial plant pathogens.

We determined that PopS contributes measurably to R. solanacearum virulence on several different hosts in the genus Solanum. Most strains of this broad-host-range pathogen have multiple effector families (e.g., GALA and AWR), whose homologs together potentiate virulence on solanaceous crop hosts such as tomato and eggplant (48–50). For example, individual GALA-family effectors are not required for full virulence on solanaceous hosts, but deleting three or more GALA effector genes delays wilt on tomato and eggplant (48, 49). PopS is a single-copy effector present in all members of the species complex, and our data indicate that it is needed for success on two agriculturally important Solanum hosts. Notably, PopS was dispensable for virulence on S. dulcamara, a common weed that can shelter and disseminate R. solanacearum (51). This difference suggests that PopS can have plant species-specific activity. As a result of selection pressures in natural ecosystems, wild hosts like S. dulcamara may have evolved to avoid PopS activity by modifying or eliminating the PopS target. Further studies are needed to define the specific mechanisms that permit S. dulcamara to resist PopS.

The popS mutant had the largest virulence defect on moderately wilt-resistant H7996 tomato, which upregulated its SA-induced PR defense genes to a greater degree in response to the popS mutant than in response to wild-type UW551. No such difference was observed in the response of roots of susceptible cv. Bonny Best (data not shown). We previously found that after infection by UW551, H7996 upregulates SA-mediated defense gene expression faster than Bonny Best (37). Consistent with this previous observation, we detected no differences in expression of PR-1a and PR-1b in Bonny Best roots inoculated with wild-type UW551 or the popS mutant (data not shown). We suspect that the larger virulence defect of the popS mutant on H7996 is directly correlated to the magnitude and timing of the defense signaling in H7996. This hypothesis was supported by our finding that susceptible tomato plants were more resistant to infection by the popS mutant when roots were pretreated with SA, which induces PR gene expression. It seems likely that PopS, like other AvrE-like effectors (12–14), also suppresses SA-induced immune responses, such as callose deposition, that are triggered by recognition of MAMPs. The specific R. solanacearum MAMPs are undetermined, but purified R. solanacearum exopolysaccharide (EPS), a conserved and essential virulence factor, triggers increased PR gene expression in quantitatively resistant H7996 but not in susceptible Bonny Best (37). Further studies are needed to determine how R. solanacearum’s T3 effectors suppress MAMP- and EPS-triggered plant defenses.

The tomato pathogenesis-related protein PR-1a is triggered by SA (52). PR-1b has been described in the literature as ethylene responsive (53–55). Based on this, we previously used PR-1b as a marker of ethylene pathway activation (37). However, PR-1b has also been described as SA responsive (56, 57) and there is some evidence that it is upregulated under both conditions (52, 58, 59). We therefore directly tested the effect of SA exposure on expression of this gene in H7996 tomato. This experiment revealed that under our conditions, both PR-1a and PR-1b are upregulated around 35-fold in response to SA treatment (Fig. 4C).

The importance of PopS for tomato plant stem colonization and wilt is consistent with our previous observation that many R. solanacearum T3SS genes are highly expressed at midstage disease in planta (19). Further, SA-induced defenses are not expressed in tomato stems until R. solanacearum reaches 108 CFU/g stem (37). Together these results affirm that T3 effectors are active not only at low pathogen cell densities early in colonization, as previously suggested (60, 61), but also at a later stage in the disease cycle. Between initial root infection and the end-stage collapse and death of the plant, R. solanacearum primarily inhabits the xylem elements, which are composed of nonliving tracheids; this raises the question of where T3SS effectors might act during midstage wilt disease. It has been suggested that bacteria in xylem elements inject effectors into the living xylem parenchyma cells that are adjacent to tracheids and accessible through the pits in xylem cell walls (19).

As expected (16), the T3SS-deficient UW551 hrcC mutant could not reach the 108-CFU/g cell densities in stems required for bacterial wilt symptom development (62) and hrcC populations declined in tomato stems over time. This suggests that the T3SS is important not only for growth but also for persistence in planta. Both animal- and plant-pathogenic Pseudomonas spp. use the T3SS to persist in host tissue (63–66). Unlike the popS mutant, growth of the hrcC mutant was not restored in SA-deficient plants. We suspect that the constraints that limit success of the hrcC mutant include an inability to overcome basal immunity (1) or manipulate host physiology (4).

To the best of our knowledge, this is the first study to explore the defense-suppressing functions of an R. solanacearum effector in roots and stems, which are the important niches for R. solanacearum during natural pathogenesis. Using PopS as an example, we propose a model for effector functions during the bacterial wilt disease cycle where R. solanacearum uses T3 effectors for (i) root invasion and colonization, (ii) suppression of root defenses, (iii) stem colonization and growth, and (iv) induction of wilt symptoms. Further studies using adjustable promoters or inducible deletion mutations could reveal when these virulence traits are required during the disease process.

Three independent lines of evidence supported our conclusion that PopS suppresses SA-mediated defenses. First, a popS mutant strain elicited 3- to 4-fold-higher expression of SA-triggered PR genes in tomato than WT UW551. Second, pretreating tomato plants with SA increased the magnitude of the popS mutant’s virulence defect, as would be expected if that defect resulted from an inability to modulate SA-triggered plant defenses. Third, PopS was dispensable for colonization of SA-defective NahG transgenic tomato, as would be predicted if the popS mutant’s colonization defect was caused by SA-mediated plant defenses. These results suggest that AvrE-family effectors generally function to suppress SA-mediated defenses, in beta- as well as gammaproteobacteria. Although PopS shares 23% amino acid sequence identity with its closest AvrE and DspE orthologs, protein sequence alignments revealed many scattered, moderately conserved regions, including the three conserved tryptophans important for the virulence activity in other gammaproteobacterial AvrE homologs (14, 43). Given the diversity of hosts that can be manipulated by AvrE-family proteins, it seems likely that this effector family interacts with a broadly conserved element of the plant defense system that indirectly or directly impacts SA-mediated responses.

Most surprisingly, we determined that PopS does not elicit cell death when transiently expressed in leaf tissue, unlike other AvrE homologs from hemibiotrophic and necrotrophic pathogens. Of the many AvrE-containing plant pathogens studied to date, R. solanacearum is the only one that causes a nonnecrotic wilt. As a biotroph, it multiplies to high cell densities in the xylem without causing necrosis. We speculate that as AvrE, WtsE, DspE, and PopS diverged from a common ancestor, they adapted to the pathogenic lifestyles (necrotrophy, hemibiotrophy, and biotrophy) of each bacterium (Fig. 6). T3 effectors from hemibiotrophic Pseudomonas induce cell death in host tissue more often than effectors from biotrophic R. solanacearum (67), which suggests that in general, T3 effectors may function differently based on a pathogen’s lifestyle. Necrotrophs such as P. carotovorum subsp. carotovorum kill host tissue upon plant contact and during multiplication. P. carotovorum subsp. carotovorum DspE elicits cell death but does not suppress SA-mediated defenses (43, 44) (Fig. 6), which is consistent with the observation that SA is not a major signal associated with necrotroph infection. AvrE-possessing hemibiotrophs (e.g., P. syringae pv. tomato and P. stewartii subsp. stewartii) multiply in living host tissue but elicit necrosis during pathogenesis. Their AvrE homologs (AvrE and WtsE, respectively) not only elicit plant cell death when heterologously introduced into plant tissue but also indirectly or directly suppress SA-mediated defenses, which are known to defend plants against biotrophs and hemibiotrophs (Fig. 6) (12, 14; S. Y. He, personal communication). It remains to be determined if P. syringae HopR, the closest ortholog of PopS, causes cell death like AvrE, although it also suppresses SA-mediated defenses (13). R. solanacearum does not cause necrosis during infection and wilt, placing it on the biotrophic end of the spectrum. PopS does not elicit plant cell death but suppresses SA-mediated plant defenses, consistent with the biology of R. solanacearum. Functional studies of chimeric AvrE homologs and heterologous complementation across pathogens could reveal the specific domains that contribute to the distinct and common biological activities of this conserved effector.

FIG 6 

Functional phylogeny of the AvrE effector family. The phylogenetic tree is adapted from Fig. 1. DspE from the necrotroph P. carotovorum subsp. carotovorum (Pcc) causes cell death when transiently expressed in plant tissue but does not suppress SA-mediated defenses (43, 44). WtsE and AvrE from the hemibiotrophs P. stewartii subsp. stewartii (Pnss) and P. syringae pv. tomato (Pst), respectively, cause cell death when heterologously introduced into plant tissue and suppress SA-mediated defenses against virulence (12, 14). R. solanacearum (Rs) is a nonnecrotic wilt pathogen. We found that PopS does not elicit cell death when expressed in plant tissue but does suppress SA-mediated defenses during tomato wilt. This model suggests that these effectors have undergone adaptation to the pathogenic lifestyles of their respective microbes.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains used in this study are listed in Table 1. Escherichia coli was grown at 37°C in Luria-Bertani medium (68). R. solanacearum was cultivated at 28°C on rich Casamino Acids-peptone-glucose (CPG) medium (pH 7.0) (69). When required, media were supplemented with kanamycin (Km) (25 µg/ml), gentamicin (Gm) (15 µg/ml), tetracycline (Tc) (15 µg/ml), or rifampin (Rif) (25 µg/ml).

TABLE 1 

Strains and plasmids used in this study

Strain or plasmid	Genotype and characteristics	Source or reference
Escherichia coli Top10	F− mcrA Δ(mrr-hsdRMS-mcrBC) φ80lacZΔM15 ΔlacΧ74 recA1araD139 Δ(ara-leu)7697 galU galK rpsL endA1 nupG λ− Smr	Life Technologies
Agrobacterium tumefaciens GV3101	Gmr Rifr	75
Ralstonia solanacearum
UW551	Wild-type geranium isolate; phylotype II, sequevar 1	76
RSW19	UW551 popS::Kmr Kmr	This study
RSW35	UW551 popS::Kmr pUC18-miniTn7T-Gm-GW::popS Kmr Gmr	This study
RSW36	UW551 hrcC::Gmr, type III secretion deficient, Gmr	This study
Plasmids
pSTBlue-1	Cloning vector, Apr Kmr	EMD Bioscience
pSUP202	Cloning vector, Apr Kmr Tcr Cmr	71
pUCGM	Apr Gmr	77
pENTR/D-TOPO	Cloning vector, Kmr; Gateway (Life Technologies)	Life Technologies
pGWB2	Expression vector, 35S promoter, Kmr	78
pGWB14	Expression vector, 35S promoter, C-terminal HA, Kmr	78
pUC18-miniTn7T-Gm-GW	Complementation vector, Gmr; Gateway (Life Technologies)	79
pTNS1	Helper plasmid for transposition; Apr	79
ECW34	pSUP202::popS::Kmr Apr Kmr Cmr	This study
ECW35	pSUP202::hrcC::Gmr Apr Gmr Cmr	This study
pENTR/D-TOPO::popS	~6.5-kb fragment containing UW551 popS with its native promoter cloned into pENTR/D-TOPO; Kmr	This study
pUC18-miniTn7T-Gm-GW::popS	~6.5-kb fragment containing UW551 popS with its native promoter cloned into pUC18-miniTn7T-Gm-GW; Gmr	This study
pGWB14::popS	~5.2-kb gene encoding UW551 PopS cloned into pGWB14; Kmr	This study
pGWB2::dspE	~4.9-kb gene encoding Pectobacterium carotovorum subsp. carotovorum DspE	43

Recombinant DNA techniques and mutagenesis.

Genomic and plasmid DNA was isolated by standard protocols (68). E. coli and R. solanacearum were transformed as previously described (68). PCR primer sequences are listed in Table S2 in the supplemental material. To disrupt the R. solanacearum UW551 locus RRSL_03375 (popS), flanking regions from popS and a kanamycin resistance (Kmr) cassette were amplified from UW551 genomic DNA and pSTBlue-1, respectively, using Phusion high-fidelity DNA polymerase (Finnzymes, Vantaa, Finland). The Kmr cassette was inserted between the popS-flanking fragments by splicing by overlap extension (SOE)-PCR (70). Similar methods were employed to create a hrcC mutant (hrcC::Gmr), except that a Gmr cassette amplified from vector pUCGM was inserted in the regions flanking hrcC. The resulting SOE-PCR product was gel purified, phosphorylated with T4 polynucleotide kinase (Promega, Madison, WI), and ligated into the EcoRV site of cloning vector pSUP202 (71) to create pSUP202-popS::Kmr. Wild-type UW551 was transformed with pSUP202-popS::Kmr, and double recombinant mutants were selected for Kmr but screened for Tcs to ensure proper allelic exchange. To complement the R. solanacearum popS::Kmr mutant, popS and its upstream native promoter were amplified from the UW551 genome and directly inserted into pENTR/D-TOPO (Life Technologies, Carlsbad, CA) following the manufacturer’s instructions. UW551 popS was transferred to complementation vector pUC18-miniTn7T-Gm-GW via LR Gateway cloning as described by the manufacturer’s protocol (Life Technologies, Carlsbad, CA). R. solanacearum popS::Kmr mutant competent cells were transformed with pUC18-miniTn7T-Gm-GW and Tn7 helper plasmid pTNS1 as previously described (72). Mutations were confirmed with PCR.

Phylogenetic analysis.

Phylogenetic trees of PopS and AvrE/DspE/HopR orthologs were created with MEGA5 (25). The amino acid or gene sequences of PopS orthologs from NCBI or MAGE databases were aligned with CLUSTAL-W, from which a maximum-likelihood phylogenetic tree was created. The R. solanacearum strains used in this analysis are listed in Table S2 in the supplemental material. The percentage of replicate trees in which individual orthologs clustered together in the bootstrap test (200 replicates) was calculated and noted at each branch. MUM index (MUMi) distances for whole-genome analysis were computed for each pair of sequenced genomes of R. solanacearum using the web server (http://genome.jouy.inra.fr/mumi/index.cgi). Briefly, the MUMi estimated the genomic distances by considering divergence of the R. solanacearum core genome as well as a gain/loss of DNA segments (34). The R. solanacearum species complex tree was created from a MUMi distance matrix using neighbor-joining cluster analysis.

Plant assays.

To evaluate pathogen virulence, pots containing individual unwounded plants were soaked with a water suspension of UW551 or the popS mutant to create a final inoculum density of 1 × 108 CFU/g soil. Hosts included 17-day-old susceptible tomato plants (cv. Bonny Best), 14- to 16-day-old moderately resistant Hawaii 7996 tomato plants, 21-day-old potato plants (cv. Russet Norkotah) grown from minitubers, and 18- to 19-day-old bittersweet nightshade plants (Solanum dulcamara). Each treatment contained 14 plants and was repeated three times, except potato, which represents one biological replicate. Inoculated plants were incubated at 28°C with a 12-h light cycle. To evaluate the impact of salicylic acid on disease development, susceptible tomato (cv. Bonny Best) plants were pretreated with either 50 ml water or 50 ml 7.5 mM sodium salicylate in water, followed 6 h later by soil soak inoculation with either UW551 WT or the popS mutant. Disease was rated daily on a disease index scale from 0 to 4 (0, no wilt symptoms; 1, 1 to 25% of leaves wilted; 2, 26 to 50% of leaves wilted; 3, 51 to 75% leaves of wilted; and 4, 76% to 100% of leaves wilted). Each assay from two independent experiments included 10 to 14 plants per treatment. The effect of salicylic acid on defense gene expression in tomato was measured by extracting RNA as described below from Money-maker tomato stems 6 h after soil soak treatment with either 7.5 mM SA or sterile water as described above.

To measure bacterial colonization, plant stems were inoculated by applying a droplet of bacterial suspension to the cut petiole of the first true leaf. Host plants used for colonization were susceptible Bonny Best and resistant H7996 tomato. Bacterial colonization was quantified daily by grinding and dilution plating stem segments on appropriate antibiotic plates. Four or five plants were sampled for bacterial colonization each day per treatment in two biological replicates. To evaluate the impact of plant salicylic acid levels on stem colonization, we measured bacterial colonization of wilt-susceptible tomato cultivar Money-maker and transgenic Money-maker expressing NahG (42). The experiment included five plants per treatment in two biological replicates.

Plant defense gene expression.

Seeds of Bonny Best or H7996 tomato were surface sterilized with 50 ml 10% bleach for 10 min, followed by an ethanol wash in 50 ml 70% ethanol for 5 min. Ethanol and bleach washes were performed in 50-ml conical tubes, and seeds were incubated on a shaker at 200 rpm at room temperature. Seeds were then rinsed 5 to 7 times with sterile water to remove residual ethanol. For uniform germination, seeds were stored at 4°C overnight in the dark in water and then germinated on 1% water agar for 48 h at room temperature in the dark. Germinated seedlings were transferred to plates containing 1% agar and 0.5× Murashige and Skoog basal salts medium plus Gambourg’s vitamins (MS medium) (MP Biomedicals, Santa Ana, CA) and incubated for 2 days at 28°C with a 12-h light cycle. Root tips were inoculated with 2 µl containing 2 × 106 CFU/ml of either GMI1000, WT UW551, or the popS mutant. Plant root tissue was harvested 2.54 cm from the inoculation point 24 hpi, immediately frozen in liquid nitrogen, and stored at −80°C. Results are averages of data from 7 to 12 plants per treatment.

Tomato RNA was extracted and purified from pooled tissue samples ground in liquid nitrogen using an RNeasy minikit (Qiagen, Valencia, CA) following the manufacturer’s instructions, except that initial flowthrough was applied to column twice. RNA was eluted in 30 µl RT-grade water. RNA purity and quality were evaluated on a NanoDrop (Thermo Scientific, Wilmington, DE) and an Agilent bioanalyzer picochip (Agilent Technologies, Santa Clara, CA), respectively. Samples with high quality (as defined by RNA integrity numbers greater than 8.0) and high purity (as defined by A260/230 and A260/280 of >1.9) were used for analysis.

Plant defense gene expression from tomato roots and stems was measured using quantitative reverse transcriptase PCR (qRT-PCR). One microgram of total RNA per sample was reverse transcribed using Superscript III reverse transcriptase first-strand synthesis Supermix (Life Technologies, Carlsbad, CA) with oligo(dT) and random hexamer primers, following the manufacturer’s protocol. qRT-PCR was performed in duplicate with 1× PowerSYBR green master mix (Life Technologies, Carlsbad, CA), 400 nM forward and reverse primers, and 50 ng cDNA template for a final volume of 25 µl. The reaction conditions were as follows: 10 min polymerase activation and 40 cycles of 95°C for 15 s and 57°C for 1 min. Relative gene expression was quantified for the tomato defense genes PR-1a and PR-1b using previously described primers (52) and normalized to that of a reference gene ( GAPDH). Relative expression of treatment compared to control was defined using the ΔΔC method (73).

Transient expression of PopS in leaf tissue.

The popS gene was amplified as described above and inserted into Gateway vector pENTR/D-TOPO following the manufacturer’s instructions (Life Technologies, Carlsbad, CA). The popS gene was inserted into expression vector pGWB14 for C-terminal HA fusion protein expression using LR cloning (Life Technologies, Carlsbad, CA). The resulting plasmid, pGWB14::popS, was confirmed with sequencing and transformed into Agrobacterium tumefaciens, followed by selection with the appropriate antibiotics. Leaves from ~30-day-old N. benthamiana and N. tabacum plants were infiltrated with either buffer control, A. tumefaciens pGWB2::dspE (positive control) (43), A. tumefaciens pGWB14 (empty vector control), or A. tumefaciens pGWB14::popS. Leaf symptoms were observed and captured by scanning leaves at 48 h postinoculation. To verify the visible cell death elicited by DspE, plant leaves were stained with trypan blue as previously described (74). Each treatment was carried out in triplicate over three independent experiments.

AvrE-family effector sequences used to generate the phylogenetic tree in Fig. 1

Table S1, DOCX file, 0.1 MB.

Primers used in this study.

Table S2, DOCX file, 0.1 MB.

Complementation with the wild-type popS locus restored the ability of the popS mutant to suppress SA-mediated defenses in roots of H7996 tomato plants. Tomato plants were grown on MS medium as described in Materials and Methods, and root tips were treated with 1 µl of either water or 2 × 105 CFU of wild-type (WT) R. solanacearum or the popS::Kmr or the complemented popS mutant (popS::Kmr miniTn7::popS). After 24 h, the bottom 2 cm of root was harvested, and total RNA was extracted from a pool of 10 to 11 root tips for each treatment, using a Direct-Zol RNA miniprep kit (Zymo Research, Irvine, CA). cDNA was synthesized by reverse transcription with Invitrogen Superscript III Supermix (Life Technologies, Carlsbad, CA), and qRT-PCRs were performed using PowerSYBR green master mix (Life Technologies, Carlsbad, CA) with an input of 75 ng cDNA template. The change in induction of the SA-responsive defense gene PR-1a is shown as the ratio of expression level in plants treated with bacteria to that in water-treated control plants, calculated by relative quantification using the ΔΔC method. Gene expression was normalized to an endogenous control GAPDH gene. Download

Figure S1, DOCX file, 0.1 MB