Figure legends
Figure 1 Symptoms caused by Dickeya pathogens. Symptoms (white arrows) caused by Dickeya solani and Dickeya dianthicola on potato stems (a) and hyacinths (b). The five symptom classes (0 to 4) used to compare the aggressiveness ofDickeya solani and Dickeya dianthicola on potato tubers (c).
Figure 2 Dickeya and Pectobacterium prevalence from potato fields exhibiting blackleg symptoms. Percentage ofPectobacterium-containing and Dickeya-containing fields (a) and that of D. solani-containing fields and D. dianthicola-containing fields (b) were calculated each year from 2004 to 2015 (with the exception of 2006). (c) Number and relative abundance of the Pectobacterium, D. dianthicolaand D. solani isolates collected in 19 symptomatic fields sampled from 2013 to 2016. A hierarchical clustering paired group method delineated four pathogen population groups, i.e. Pectobacteriumonly (I: 3 fields), D. solani and Pectobacterium (II: 9 fields), D. dianthicola and Pectobacterium (III: 3 fields), and D. dianthicola, D. solani and Pectobacterium(IV: 4 fields).
Figure 3 Symptom incidence and fitness of Dickeya dianthicola and D. solani in potato plants. In a, mean value and standard error (SE) of the percentage (%) values of plants exhibiting blackleg symptoms, which were measured after inoculation of each of the five D. solani strains (3337, IPO2222, RNS05-1-2A, Ds0432.1, PPO9019) and five D. dianthicola strains (RNS11-47-1-1A, CFBP1888, CFBP2982, CFBP2015, MIE34) on 15 plants. The p-value of the Kruskal-Wallis test comparing of the symptomatic classes between the two species is indicated below the graph. In b, mean value and SE between two replicates of the percentages (%) of plants exhibiting blackleg symptoms which was measured on 15 plants inoculated by D. dianthicola and D. solani populations and their mix. The p-values of the pairwise comparisons (Post-hoc Tukey tests) of the symptomatic classes are indicated below the graph, when p ≤ 0.1. In c, competitive index (CI) values between D. solani and D. dianthicola populations were calculated in eight emerging lesions of co-infected plants and revealed a competitive advantage of D. dianthicola: the CI median (= 2.4 10-5) is represented by a thick bar; p-values resulting from Kruskal-Wallis tests testing difference from one are indicated. Legend: * for 0.05<p≤0.1; ** for 0.01<p≤0.05 and *** for p ≤0.01.
Figure 4 Aggressiveness and fitness of Dickeya dianthicola and D. solani in potato tubers. Rotting assays were conducted by inoculating each tuber by either 107colony-forming units (CFU) (in a, b and c) or 105 CFU of pathogens (in d, e andf). In a, b, d and e, each disease severity index (DSI) value was calculated using symptom classes observed on 10 tubers. In a and d, mean values and standard errors (SE) of DSI values were measured for each of the fiveD. solani strains (3337, IPO2222, RNS05-1-2A, Ds0432.1, PPO9019) and five D. dianthicola strains (RNS11-47-1-1A, CFBP1888, CFBP2982, CFBP2015, MIE34) using five (a) and two (d) independent experiments. The p-values of the Kruskal-Wallis tests comparing symptom classes are indicated below the graphs. In band e, mean values and standard errors of DSI values were measured twice for each of the experimental populations of D. dianthicola, D. solani and the mixture of the two. The p-values of the pairwise Tukey tests comparing the symptomatic classes are indicated, when p ≤ 0.1. In c and f, the competitive index (CI) between D. solani and D. dianthicola populations was calculated in 10 co-infected symptomatic tubers: the median values (indicated by a thick bar) reached 5.7 (c) and 5.8 (f) and the CI values were statistically different from one (Kruskal-Wallis test; p = 5 10-5 and p = 0.05 respectively), revealing a competitive advantage of D. solani. Legend: * for 0.05<p≤0.1; ** for 0.01<p≤0.05 and *** for p ≤0.01.
Figure 5 Expression of the pectate lyase genes pelA,pelD and pelE. The expression level of the pelA,pelD and pelE genes was evaluated in each of the fiveD. solani strains (3337, IPO2222, RNS05-1-2A, Ds0432.1, PPO9019) and five D. dianthicola strains (RNS11-47-1-1A, CFBP1888, CFBP2982, CFBP2015, MIE34), grown in three conditions: a rich culture medium in the absence of pectin (exponential growth phase) and symptomatic tubers (at 5 days post infection) and emerging lesions in stems. Relative expression was measured four times and normalized using the rpoB and yafS gene expression. In the graphs, the mean values and standard error (SE) of gene expression from all strains of a given species are indicated, as well as p-values of pairwise comparisons by Tukey tests. Legend: * for 0.05<p≤0.1; ** for 0.01<p≤0.05 and *** for p ≤0.01.
Figure 6 Population genomics of Dickeya solani. Ina, scan of 67 D. solani genomes revealed that the genome position 2,930,940, (according the D. solani 3337 genome), in thevfmB gene, was the most balanced non-synonymous variation, with alternated VfmBPro (71%) and VfmBSer(29%) alleles. In b, single nucleotide polymorphism (SNP)-based tree of D. solani strains using PHYLOViZ; the name of the French isolates is indicated in blue font; the VfmBSer and VfmBPro strains used in the plant assays are underlined. In c, dynamics of the VfmBSer allele (%) among the D. solani isolates along the sampling period 2005-2015 in France.
Figure 7 Characterization of the VfmB alleles. In a, using the Escherichia coli AidB protein (the PDB accession is 3DJL) as a model, the Phyre2-predicted structure of the VfmB protein showed conformational differences in the beta-sheet associated with the VfmBPro and VfmBSer alleles; the Pro55 and Ser55 positions are in red color in the VfmB representations obtained using the EzMol web server. In b, mean values and standard error (SE) of disease severity index (DSI) values, which were calculated by recording symptom classes on 10 tubers infected by 107 CFU of each of the D. solani strains carrying either VfmBPro (IPO2222, MIE35, AM3a and 3337) or VfmBSer (Ds0432.1, RNS10-27-2A, Sp1a and M21a); the p-value of the Kruskal-Wallis test comparing the four VfmBPro and four VfmBSer strains using symptom classes is indicated below the graph. In c, comparative transcriptome of D. solani 3337 (VfmBPro) and Ds0432.1 (VfmBSer) recovered from soft-rot lesions in potato tubers; upregulated virulence genes were enriched in Ds0432.1 (VfmBSer) as illustrated by the pelB, pelC, pelC, hcpA and prtA genes (closed circles).
Figure 8 Aggressiveness, symptom incidence and fitness assays of two Dickeya solani experimental populations expressing the VfmB alleles. Soft-rot assays were conducted by inoculating each tuber by either 107 colony-forming units (CFU) (a) or 105 CFU (b) of the VfmBProand VfmBSer D. solani populations and a mixture of the two. In a and b, each disease severity index (DSI) value (measuring aggressiveness) was calculated using symptomatic classes observed on 10 tubers; the assays were performed in triplicate. The p-values of the pairwise Tukey tests comparing the symptom classes are indicated, when p ≤ 0.1. In c, percentage (%) of plants exhibiting blackleg symptoms was measured on plants inoculated by the VfmBPro and VfmBSer D. solanipopulations and a mixture of the two. These assays measuring symptom incidence were performed in triplicate. Kruskal-Wallis test revealed no difference between the three treatments using symptom classes (p = 0.86). In a, b and c, competitive index (CI) values of VfmBPro and VfmBSer populations were calculated using allele counts based on shotgun sequencing of bacterial populations recovered from co-infected tissues. Median values of CI are represented as thick lines and reached 1.7 (8 repeats in a), 2.0 (3 repeats in b) and 14.0 (16 repeats in c) in the three treatments. Statistical differences between CI values and 1 (Kruskal-Wallis test) revealed that neither D. solaniVfmBPro or D. solani VfmBSer had significant advantage in tubers inoculated with a high load (p = 0.3 ina), but D. solani VfmBPro was more competitive in tubers inoculated with a low load (p = 0.04 inb) and in stems (p = 9 x 10-3 in c). Legend: * for 0.05<p≤0.1; ** for 0.01<p≤0.05 and *** for p ≤0.01.