Deep Population Genomics Reveals Systematic and Parallel Evolution at a Lipopolysaccharide Biosynthetic Locus in Xanthomonas Pathogens That Infect Rice and Sugarcane

Abstract

The advent of high-throughput sequencing and population genomics has enabled researchers to investigate selection pressure at hypervariable genomic loci encoding pathogen-associated molecular pattern (PAMP) molecules like lipopolysaccharide (LPS). Xanthomonas is a model and a major group of phytopathogenic bacteria that infect hosts in tissue-specific manner. Our in-depth population-based genomic investigation revealed the emergence of major lineages in two Xanthomonas pathogens that infect xylem of rice and sugarcane is associated with the acquisition and later large-scale replacement by distinct type of LPS cassettes. In the population of the rice xylem pathogen, Xanthomonas oryzae pv. oryzae (Xoo) and sugarcane pathogens Xanthomonas sacchari (Xsac) and Xanthomonas vasicola (Xvv), the BXO8 type of LPS cassette is replaced by a BXO1 type of cassette in Xoo and by Xvv type LPS cassette in Xsac and Xvv. These findings suggest a wave of parallel evolution at an LPS locus mediated by horizontal gene transfer (HGT) events during its adaptation and emergence. Aside from xylem pathogens, two closely related lineages of Xoo that infect parenchyma of rice and Leersia hexandra grass have acquired an LPS cassette from Xanthomonas pathogens that infect parenchyma of citrus, walnut, and strawberries, indicating yet another instance of parallel evolution mediated by HGT at an LPS locus. Our targeted and megapopulation-based genome dynamic studies revealed the acquisition and dominance of specific types of LPS cassettes in adaptation and success of a major group of phytopathogenic bacteria. IMPORTANCE Lipopolysaccharide (LPS) is a major microbe associated molecular pattern and hence a major immunomodulator. As a major and outer member component, it is expected that LPS is a frontline defense mechanism to deal with different host responses. Limited studies have indicated that LPS loci are also highly variable at strain and species level in plant-pathogenic bacteria, suggesting strong selection pressure from plants and associated niches. The advent of high-throughput genomics has led to the availability of a large set of genomic resources at taxonomic and population levels. This provides an exciting and important opportunity to carryout megascale targeted and population-based comparative genomic/association studies at important loci like those encoding LPS biosynthesis to understand their role in the evolution of the host, tissue specificity, and also predominant lineages. Such studies will also fill major gap in understanding host and tissue specificity in pathogenic bacteria. Our pioneering study uses the Xanthomonas group of phytopathogens that are known for their characteristic host and tissue specificity. The present deep phylogenomics of diverse Xanthomonas species and its members revealed lineage association and dominance of distinct types of LPS in accordance with their origin, host, tissue specificity, and evolutionary success.

Keywords: LPS; Xanthomonas citri; Xanthomonas oryzae; genomics; grasses; host specificity; parenchyma; pathovars; phages; phylogeny; population; tissue specificity; xylem.

FIG 1

Phylogenetic tree of Xanthomonas type strains. (A) Fast-tree was used to reconstruct the core-gene tree, which was then visualized using iTOL software. Clade I is depicted in yellow, while clade II is depicted in green. The presence of BXO8 type lipopolysaccharide (LPS) cassettes is indicated by pink arrowheads. (B) Easyfig and the BLASTn algorithm were used to compare the LPS synteny of Xoo-BXO8, Xaxn-NCPPB 796, Xvv-NCPPB 795, and Xsac F10 or X. sacchari F10. The gene locations are represented by arrows, and the degree of homology between pairs of genes in two LPS cassettes is represented by shading. The etfA and metB genes are represented by gray arrows, while IS elements are shown in black.

FIG 2

Distribution of LPS cassette in the Xoo population. (A) PhyML was used to reconstruct the core-gene tree, which was then visualized using iTOL software. Yellow represents BXO1 type LPS, while pink represents BXO8 type LPS cassette found in each strain. The number of substitutions per site is indicated by the bar. (B) Genetic comparison of LPS cassettes of the BXO1 and BXO8 types LPS cassette. The location and direction of genes are represented by arrows, and homology between two genes is represented by similar colors. The etfA and metB genes are represented by gray arrows, while IS elements are shown in black.

FIG 3

Population phylogenomics of X. vasicola. (A) Fast-tree was used to reconstruct the core-gene tree, which was then visualized using iTOL software. The different lineages in the X. vasicola population are represented by the inner color ring. The presence of BXO8 type LPS cassettes (pink) and Xvv type LPS cassettes (sky blue) is indicated in the outermost ring. (B) Comparison of BXO8 type LPS cassette (pink) with Xvv NCPPB 795 (pink) and Xvv SAM119 (sky blue). The etfA and metB genes are represented by gray arrows, while IS elements are shown in black.

FIG 4

Phylogenetic tree of X. albilineans and X. sacchari populations. (A) Population study based on core genes and visualized with iTOL software, lineage I (dark purple) and X. sacchari, lineage II (light purple). The presence of BXO8 type LPS cassette (pink), Xvv type (sky blue), and Xoc type (green color) LPS cassettes in both populations is represented by the outermost arrows. (B) Comparison of LPS cassettes and genetic organization in two populations with BXO8 type LPS cassette. Pink represents BXO8 type LPS (BXO8 and X. sacchari F10), while sky blue and green represent Xvv (X. albilineans GPE PC73) and Xoc (X. albilineans REU209). The etfA and metB genes are represented by gray arrows, while IS elements are shown in black.

FIG 5

Distribution of Xoc and Xol lineages and LPS cassettes in the Xoo population. (A) PhyML was used to reconstruct the core-gene tree, which was then visualized using iTOL software. Lineages are represented by different node colors. The green node represents the Xoc and Xol population, with * denoting parenchyma pathovar and #Xoo denoting Xoo-xylem pathovar. The outermost arrows indicate the presence of LPS cassettes (yellow, BXO1 type LPS cassette; green, Xoc type LPS cassette; pink, BXO8 type LPS cassette) in each strain. The number of substitutions per site is indicated by the bar. (B) Comparison of Xoc type LPS cassette (X. oryzae pv. oryzicola BXOR1 and X. oryzae pv. leersia NCPPB 4346) and BXO8 type LPS cassette the etfA and metB genes are represented by gray arrows, while IS elements are shown in black.

FIG 6

Distribution of the Xoc type LPS in Xcc population. (A) PhyML was used to reconstruct the core-gene tree, which was then visualized using iTOL software. The outermost symbols represent different pathotypes, with green arrows representing strains with Xoc type LPS cassette with A* pathotype (red circles) and Aw pathotype (blue circles). (B) Easyfig analysis: Xoc type LPS and BXO8 type LPS cluster comparison with X. citri pv. citri DAR84832. The arrows indicate gene location, and the shaded lines indicate the degree of homology between pairs of genes in two LPS cassettes.

FIG 7

Schematic view representing the major type of LPS cassettes in Xanthomonas population. On the left are names of different tissues (indicating color boxes), and on the right are different hosts to which each type of LPS cassette is linked. Each LPS cassette has metB and etfA flanking genes. The wzm and wzt marker genes are shown next to each other in the same color scheme, indicating homology. The star signifies IS elements and hypothetical genes. In the present study, the metB locus region of each cassette is found to be the most hypervariable and is responsible for virulence and transport. The other biosynthesis locus is in the etfA region and contains genes that are similar to the ancestral one (the BXO1 type LPS cassette is an exception).