Résumé
The complete genome of Xanthomonas albilineans, the causal agent of sugarcane leaf scald, was recently sequenced (1). Interestingly, this pathogen lacks a type Ill secretion system (T3SS) of the Hrp1 and Hrp2 injectisome families that is commonly found in plant pathogenic bacteria. However, X. albilineans possesses a T3SS belonging to the injectisome family SPI-1 (Salmonella Pathogenicity Island-1). This SPI-1 injectisome family mainly includes T3SSs from human and insect bacterial pathogens or symbionts, but X. albilineans is not known to be insect-transmitted. Occurrence of the SPI-1 T3SS was confirmed in 11 out of 12 strains of X. albilineans from worldwide locations, but its absence in one strain isolated from diseased sugarcane suggested that this system was not required by X. albilineans to spread within sugarcane xylem vessels and to cause disease. This hypothesis was further validated by functional analysis of two T3SS SPI-1 insertional knockout mutants of X. albilineans. These mutants were able to produce leaf symptoms and to colonize the sugarcane stalk as efficiently, both spatially and in intensity, as wild type X. albilineans. Although it appeared useless in xylem spread, the T3SS SPI-1 encoded by X. albilineans is expected to be functional because all genes encoding the structural injectisome components and the translocase proteins are conserved. No genes encoding known T3SS effectors are conserved in X. albilineans. However, the T3SS SPI-1 gene cluster of X. albilineans encodes 14 hypothetical proteins that are candidate T3SS effectors. One of these hypothetical proteins harbors an F-box-like domain, and such a domain was described in Hrp type III effectors of Ralstonia solanacearum (2). The translocase protein SipD encoded by X. albilineans contains a unique additional sequence, suggesting a unique function of the T3SS SPI-1 of X. albilineans. Selection and phylogenetic analyses using a collection of 11 strains representative of the genetic diversity of X. albilineans revealed purifying selection acting on the genes encoding the T3SS SPI-1 injectisome components. Furthermore, recombination and phylogenetic analyses provided clear evidence of recombination events between two copies of the same T3SS SPI-1 gene cluster, one integrated in the genome and an ancestral one most likely harboured by a lost plasmid. Interestingly, three PCR DNA fragments which are more than 85% identical to the sequence of the T3SS SPI-1 of X. albilineans were previously described in a lineage of Xanthomonas axonopodis pv. phaseoli (3), the causal agent of common bacterial blight of bean. Occurrence in this lineage of 8 of the 14 genes encoding the structural injectisome components and 1 of 3 genes encoding translocase proteins was demonstrated by PCR and sequencing. No hypothetical protein (candidate T3SS effector) encoded by the T3SS SPI-1 of X. albilineans was detected in X. axonopodis pv. phaseoli by PCR. The T3SS SPI-1 present in both X. albilineans and X. axonopodis pv. phaseoli was not detected either by PCR or by Dot blot in a collection of 130 additional plant pathogenic bacteria belonging to the genera Xanthomonas, Pseudomonas, Erwinia, Clavibacter, Pectobacterium, Acidovorax, Dickeya, Herbaspirilum and Ralstonia. The evolutionary history of the T3SS SPI-1 gene cluster will be discussed in light of a putative interaction between X. albilineans and an insect or a putative role of this T3SS during the interactions between X. albilineans and sugarcane. (Résumé d'auteur)