Complete Genome Sequencing of Three Clade-1 Xanthomonads Reveals Genetic Determinants for a Lateral Flagellin and the Biosynthesis of Coronatine-Like Molecules in Xanthomonas

The Gram-negative bacterial genus Xanthomonas constitutes a large group of mostly plant-associated bacteria, collectively causing diseases on hundreds of host plants, including economically important crops and ornamental plants. Thanks to increased genome sequencing efforts, the number of Xanthomonas genome sequences has increased tremendously in recent years, including strains of scientifically and/or economically relevant species and pathovars (Mansfield et al. 2012). However, understudied non-model organisms have also been systematically sequenced, thus correcting taxonomically mis-assigned strains and leading to the description of novel species (Bansal et al. 2021; Dia et al. 2021; López et al. 2018; Mafakheri et al. 2022; Martins et al. 2020; Vicente et al. 2017).

Currently, the List of Prokaryotic names with Standing in Nomenclature (LPSN; https://www.bacterio.net, assessed on 8 September 2022) includes 34 validly described Xanthomonas species (Parte et al. 2020). Previous work based on partial sequencing of the gyrB gene from poorly characterized pathovars and from unidentified xanthomonads, however, suggested the existence of additional species, which were at that time described as species-level clade strains, SLC 1 to SLC 7 (Parkinson et al. 2009). Strains from SLC 1 and 4 have been sequenced since then, corresponding to the proposed species “Xanthomonas cannabis” and “Xanthomonas badrii,” respectively (GenBank accession numbers CP051651 and JSZE00000000; Jacobs et al. 2015; Zain and Roberts 1977).

Xanthomonads are classified into two major groups based on sequence comparisons (Ferreira-Tonin et al. 2012; Gonçalves and Rosato 2002; Hauben et al. 1997; Parkinson et al. 2007). SLC 1 to 4 belong to clade 2, whereas SLC 5 to 7 belong to clade 1, also known as the clade of early-branching species (Parkinson et al. 2007, 2009). Most of the clade-1 strains have been isolated from monocots, such as asparagus, banana, sugarcane, hyacinths, forage grasses, and small-grain cereals (Jacques et al. 2016). However, there are a few exceptions: X. theicola causes canker on tea plants, X. bonasiae and X. youngii were found on canker-symptomatic ornamental plants, and strains of X. translucens have also been isolated from pistachio trees (Giblot-Ducray et al. 2009; Koebnik et al. 2021; Mafakheri et al. 2022). SLC-5 strain NCPPB 2654 (=CFBP 8443) was isolated from bean (Phaseolus vulgaris), SLC-6 strain NCPPB 2983 (=CFBP 8444, LMG 702) originates from New Zealand flax (Phormium tenax), and seven SLC-7 strains have been collected from monocots such as banana (Musa × paradisiaca), ginger (Zingiber officinale), or sugarcane (Saccharum officinarum), but also from cotton (Gossypium sp.) (Parkinson et al. 2009).

The SLC-6 bacterium X. campestris pv. phormiicola (Takimoto 1933) Dye 1978 was described as a bacterial pathogen of New Zealand flax (Phormium tenax Forst.), causing water-soaked lesions with chlorosis (Takimoto 1933). The same bacterium was also found to induce hypertrophy on potato tuber tissues at the inoculated site (Tamura et al. 1987). The hypertrophic active substance isolated from the bacterial culture filtrate produced visible chlorotic lesions on the leaves of New Zealand flax, ryegrass, and tobacco plant (Tamura et al. 1992). Biochemical and biophysical parameters (retardation factor on silica gel thin-layer chromatography and 1H nuclear magnetic resonance spectrum) of the purified toxin coincided with those of coronatine, and it was concluded that the toxin produced by X. campestris pv. phormiicola is coronatine (Tamura et al. 1992). In a parallel study, liquid cultures of X. campestris pv. phormiicola were found to contain two analogues of coronatine lacking the cyclopropane ring structure, but no trace of coronatine (Mitchell 1991). Using nuclear magnetic resonance spectroscopy, mass spectrometry, and gas chromatography of hydrolysis products, these two compounds were identified as N-coronafacoyl-l-valine and N-coronafacoyl-l-isoleucine (Mitchell 1991). These were the first two reports of coronatine-like phytotoxins outside of the genus Pseudomonas.

To obtain insight into the genetics of coronatine production in Xanthomonas and to expand genomic information on clade-1 xanthomonads, we determined the complete genome sequences of the pathotype strain CFBP 8444^PT (ICMP 4294^PT, LMG 702^PT, NCPPB 2983^PT) and of the species-level clade strains CFBP 8443 (NCPPB 2654) and CFBP 8445 (NCPPB 1131).

From lyophilized stocks available at CIRM-CFBP (https://cirm-cfbp.fr/), single colonies of the three strains were isolated and grown on peptone-sucrose (10 g/liter of peptone, 10 g/liter of sucrose, 1 g/liter of glutamic acid) agar media. Genomic DNA was extracted using the Genomic DNA buffer set and Genomic-tips following the manufacturer's instructions (Qiagen, Valencia, CA, U.S.A.).

Multiplex library preparation in pools of eight strains, including strains CFBP 8443, CFBP 8444, and CFBP 8445, and simultaneous sequencing of eight strains on one SMRTCell was conducted with the PacBio Sequel technology on the GENTYANE genotyping platform (INRAE, Clermont-Ferrand, France).

Sequence reads were de novo assembled with Flye 2.7 using the “–plasmids –iterations 2” parameters (Kolmogorov et al. 2019), yielding a single chromosome for each of the three strains. Genome sequences were functionally annotated using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP version 6.2) (Li et al. 2021). The clade-1 xanthomonads genome project is available at NCBI as BioProject PRJNA865934. Annotation details are given in Table 1.

TABLE 1. BioSamples, GenBank accession numbers, and annotation details for the clade-1 xanthomonads genome BioProject PRJNA865934

Strain	CFBP 8443	CFBP 8444	CFBP 8445
Species-level clade	SLC 5	SLC 6	SLC 7
NCBI BioSample	SAMN30147989	SAMN30148166	SAMN30161540
NCBI GenBank accession number (chromosomes)	CP102592	CP102593	CP102594
Coverage	200×	244×	208×
Size (bp)	5,238,480	5,314,063	4,869,984
G+C content (%)	68.96	68.64	69.25
Predicted genes	4,368	4,506	4,192
Predicted coding sequences	4,303	4,441	4,122
Predicted RNA genes	65	65	70
Predicted pseudogenes	44	72	30

View as image HTML

To clarify the phylogenetic relationships within clade 1, including the three newly sequenced SLC strains, pairwise average nucleotide identities (ANIs) were calculated on a Galaxy-implemented version of FastANI (Jain et al. 2018). This analysis included all publicly available genome sequences from clade 1, except for X. translucens, for which only eight pathotype strains were chosen as representatives of the genetic diversity of this species due to the large number of sequenced X. translucens strains (Goettelmann et al. 2022). The evolutionary history was inferred using the UPGMA method (Sneath and Sokal 1973), as implemented in MEGA11 (Stecher et al. 2020; Tamura et al. 2021). This analysis involved 75 genome sequences and confirmed the taxonomic position of the three SLC strains within clade 1 (Fig. 1; Supplementary Table S1).

Download as PowerPoint

Using a 95% cutoff value as an ANI species delimiter (Goris et al. 2007; Richter and Rosselló-Móra 2009), SLC-5 strain CFBP 8443 was found to be the only representative for this novel species. SLC-6 strain CFBP 8444 and its clone NCPPB 2983 in the U.K. strain collection represent a sister species of X. bonasiae (Mafakheri et al. 2022). Three nearby strains in the dendrogram, GW, SI, and SS, which were isolated as candidate biocontrol agents from the microbiome of perennial ryegrass (Lolium perenne), were found to be phylogenetically close to the species X. bonasiae (Li et al. 2020). However, using digital DNA-DNA hybridization as a species delimiter suggests that strains GW, SI, and SS form another sister species of X. bonasiae (Supplementary Table S2) (Meier-Kolthoff et al. 2013, 2022). We therefore consider strain CFBP 8444 as the representative of a novel species, corresponding to SLC 6.

SLC-7 strain CFBP 8445 is a clone of NCPPB 1131, for which a draft genome sequence was published previously (Studholme et al. 2011). These strains clustered with another genetically similar strain, LMG 8989, which was isolated from a Citrus plant (Bansal et al. 2020), in contrast to CFBP 8445 and NCPPB 1131, which were isolated from banana. Recently, nonpathogenic strains were isolated from healthy rice seeds. Genome sequencing revealed that two of the strains, PPL560 and PPL568, belong to a novel species, for which the name “Xanthomonas indica” was proposed (Rana et al. 2022). ANI and digital DNA-DNA hybridization analyses demonstrate that these two strains belong to the same species-level clade as the four former strains, and thus, SLC 7 could be renamed “X. indica” (Fig. 1; Supplementary Tables S1 and S2). Hence, this species-level clade comprises bacteria that can colonize both monocots and dicots. Whether such promiscuity also holds true for SLC 5 and SLC 6 strains awaits the characterization of additional isolates.

This comprehensive comparison of clade-1 strains suggests the existence of at least four additional novel species, as exemplified by strains 569 (origin unknown), DD13 (from sugarcane), LMG 9002 (from orange), F4 (origin unknown), 3498 (origin unknown), LMG 8992 (from orange), and NCPPB 1128 (from common bean) (Fig. 1; Supplementary Tables S1 and S2) (Aritua et al. 2015; Bansal et al. 2020). Notably, for many isolates, it is not clear whether they are pathogenic on their host of isolation or on other plants, or if they are regular constituents of the plant's microbiome, perhaps even with protective features. This genome-wide comparison challenges previous views of the phylogenetic relationships within clade-1 and supports recent findings based on comparing 466 core genes (Rana et al. 2022). Comparison of 16S rRNA sequences grouped X. sacchari and “X. sontii” in one subclade and X. albilineans, X. hyacinthi, X. theicola, and X. translucens in another subclade (Bansal et al. 2021; Hauben et al. 1997). In contrast, comparisons based on 16S–23S rDNA intergenic spacer sequences or on a portion of the gyrase B gene suggested that X. albilineans, “X. pseudalbilineans,” X. sacchari, and “X. sontii” form one subclade, and X. hyacinthi, X. theicola, and X. translucens belong to another subclade (Gonçalves and Rosato 2002; Khenfous-Djebari et al. 2019; Koebnik et al. 2021; Parkinson et al. 2009). Our genome-wide ANI comparison clearly separates X. albilineans and “X. pseudalbilineans” from the rest of the clade-1 strains, which correlates with the reduced genome size of these two species (average 3.7 Mbp for 20 strains) in comparison with the other 55 strains in our comparison (4.9 Mbp) (Pieretti et al. 2009, 2015).

The phytotoxin coronatine (COR) consists of two moieties, the polyketide coronafacic acid (CFA) and the cyclopropyl amino acid coronamic acid (CMA), which are conjugated by an amide bond. COR and COR-like molecules are produced by several pathovars of Pseudomonas syringae but also some plant-pathogenic enterobacteria and streptomycetes (Bender et al. 1996; Fyans et al. 2015; Slawiak and Lojkowska 2009). In P. syringae, COR production is under the control of the unconventional two-component system CorRPS, which mediates COR biosynthesis at 18°C, whereas COR is not detectable at 28°C (Smirnova et al. 2002). COR and COR-like molecules have been identified as structural mimics of the plant hormone jasmonoyl-L-isoleucine (Bender et al. 1999; Bown et al. 2017), which affects plant signaling pathways with importance for plant development and defense (Gimenez-Ibanez et al. 2015). Notably, COR was identified as the virulence factor in P. syringae that is responsible for suppressing stomatal innate immune defense (Melotto et al. 2006).

Because X. campestris pv. phormiicola was reported to synthesize coronatine-like molecules, we searched the three new genome sequences for homologs of the nine cfa genes (coronafacic acid biosynthesis), seven cma genes (coronamic acid biosynthesis), the cfl gene (coupling factor), and the three regulatory cor genes from Pseudomonas syringae pv. tomato strain DC3000 using TBLASTN. Homologs were only detected in strain CFBP 8444 for all nine Cfa sequences (between 66 and 92% sequence identity), with the Cfl sequence (64% sequence identity), and with the three Cor sequences (between 58 and 67% sequence identity) (Fig. 2). Of the Cma sequences, only CmaA and CmaT had hits, with 32% sequence identity in both cases, which were, however, not associated with coronamic acid biosynthesis.

Download as PowerPoint

We then further looked at whether related genes are found in other bacteria, using the CFBP 8444 Cfa, Cfl, and Cor amino acid sequences as queries. As expected, we found homologs for the whole biosynthetic gene set in species of Pseudomonas, in several enterobacteria (e.g., Brenneria, Pectobacterium, Dickeya, and Lonsdalea), and in Streptomyces. However, the regulatory cor genes were only found in Pseudomonas. Surprisingly, we also identified close homologs of all these genes in two additional strains of Xanthomonas, annotated as X. arboricola strain 3058 and X. campestris strain 3075. According to the BioProject descriptions at NCBI (PRJNA583332, PRJNA583333), these strains originated from a study of plant-associated saprophytic bacteria and their role in plant health and plant-pathogen interactions. However, because these two Xanthomonas strains share a genome-wide ANI of 98.7%, they belong to the same species. Based on analyses at the Type Strain Genome Server (https://tygs.dsmz.de/) (Meier-Kolthoff and Göker 2019), both strains are predicted to belong to a hitherto undescribed species within clade 2 of Xanthomonas.

The regulatory cor genes were also identified in the genome sequence of X. theicola, strain CFBP 4691. However, none of the cfa or cfl genes was found. Because the Cor regulatory system serves as a thermosensor in P. syringae, we speculate that biosynthesis of a COR-like substance is likely controlled in a similar way in Xanthomonas, but not in enterobacteria or Streptomyces, whereas other factors might be under thermal control in X. theicola.

Xanthomonas has been proposed as a name for a group of non-sporing, rod-shaped, uni- or rarely biflagellate, or non-motile Gram-negative bacteria, forming abundant yellow, slimy colonies on nutrient agar and potato, and mostly digesting starch and producing acid in lactose but not in salicin (Dowson 1939). Genomics has since confirmed the presence of a canonical gene cluster for flagella biosynthesis in almost all sequenced strains, regardless of the species. However, some strains have experienced deletion mutations in the flagellar gene cluster, which rendered them non-motile on diagnostic media (Darrasse et al. 2013; Jacobs et al. 2015). All three genomes (Table 1) contained the canonical flagellar gene cluster, which has a genomic organization very similar to the one in clade-2 strains (Supplementary Fig. S1).

Surprisingly, strain CFBP 8444 was found to possess a second gene cluster for the possible biosynthesis of alternative or additional flagella (Fig. 3). This gene cluster is unique in the genus Xanthomonas. The only other Xanthomonas that was found to have a similar system, strain XNM01, has been misclassified and belongs to the sister genus Pseudoxanthomonas (Gowda et al. 2022). BLASTP searches revealed additional close homologs in other strains of Pseudoxanthomonas, in one strain of Pseudomonas, in several beta proteobacteria (mainly from the order Burkholderiales, but also from Neisseriales, Nitrosomonadales, and Rhodocyclales), and in one strain of alpha proteobacteria (Sphingomonas). This atypical gene cluster contains a flagellin gene, named fliC or lafA, that is sometimes annotated as lateral flagellin. Overall, sequence similarity between the predicted CFBP 8444 flagellar proteins and their homologs is not very strong, with less than 70% sequence identity for the LafA (FliC) protein and on average 64% sequence identity to the most similar system in strain XNM01.

Download as PowerPoint

Vibrio parahaemolyticus is another bacterium that has two types of flagella, which are both associated with distinct cell types, the swimmer cell and the swarmer cell (McCarter and Silverman 1989). The swimmer cell is a short rod that harbors a single polar flagellum, which is produced when the bacterium is grown in liquid media. On semi-solid media, swarmer cells of V. parahaemolyticus are elongated and synthesize, in addition to the polar flagellum, numerous lateral flagella that are responsible for translocation over surfaces. Transition between swimmer and swarmer cells is mediated by contact with surfaces and involves the polar flagellum as a tactile sensor. In addition, iron limitation serves as a second signal that is required for swarmer cell differentiation (McCarter and Silverman 1989).

Mesophilic aeromonads are another group of bacteria that express a polar flagellum in all culture conditions, and certain strains produce lateral flagella on semisolid media or on surfaces (Canals et al. 2006). In addition to motility, lateral flagella of Aeromonas species have been found to mediate epithelial cell adherence and biofilm formation (Gavín et al. 2002). Later, homologs of the lateral flagellin were found in many beta proteobacteria, such as Burkholderia dolosa (Roux et al. 2018). Notably, a lafA deletion mutant was more motile under swimming conditions due to an increase in the number of polar flagella but was not affected in biofilm formation, host cell invasion, or murine lung colonization or persistence over time. Shewanella putrefaciens is another species that possesses two separate flagellar systems (Kühn et al. 2022). Using functional fluorescence tagging, it was found that screw thread-like motility is mediated by the primary, polar flagellum and that the lateral flagella support spreading through constricted environments such as polysaccharide matrices.

Lateral-type flagellins were also found in the alpha proteobacterium Sphingomonas (Maruyama et al. 2015). Sphingomonas sp. strain A1, originally identified as a non-motile and aflagellate bacterium, possesses two sets of flagellar genes, consisting of 35 and 46 genes, respectively. The smaller set encodes the lateral-type flagellin, whereas the larger set includes two flagellar genes typical for polar flagella (Maruyama et al. 2015). Strain A1 cells became motile when they were cultured on semisolid media, due to the presence of a single flagellum at the cell pole, which surprisingly consisted of both lateral and non-lateral flagellins (Maruyama et al. 2015). Immunogold labeling later demonstrated that strain A1 produces two types of flagellar filaments, one formed by three flagellins (lateral and non-lateral flagellins), and the other formed only by lateral flagellins (Kobayashi et al. 2016). Interestingly, lateral-type flagellins were found at the proximal end of the chimeric filaments, next to the hook basal body, whereas the polar flagellins were detected farther apart from the cell surface (Kobayashi et al. 2016).

Given the low sequence conservation between the CFBP 8444 noncanonical flagellar genes and those in other proteobacteria, it will be interesting to study this phenomenon further. It remains to be explored whether this strain of Xanthomonas produces chimeric polar flagella or lateral flagella under certain conditions, and if so, which environmental clues trigger a switch between different types of flagella.

The first complete genome sequence for three undescribed species will stimulate further work on the underexplored clade-1 of xanthomonads. Strikingly, this work revealed the presence of a novel, noncanonical flagellar gene cluster that may produce lateral flagella, challenging the view of Xanthomonas as a unipolar flagellated bacterium. This work also presents the genetic basis for the production of COR-like molecules in the genus Xanthomonas. Because coronatine has been shown to stimulate reopening of plant leaves’ stomata during the infection process, it will be interesting to study the contribution of these molecules to pathogenicity. Notably, strain CFBP 8444, which synthesizes these COR-like molecules, does not produce type 3 effectors, which, in other xanthomonads, suppress stomatal immunity (Liu et al. 2022; Raffeiner et al. 2022; Wang et al. 2021).

Acknowledgments

We thank Cécile Dutrieux and Audrey Lathus for assistance with strain preservation. We also acknowledge the ISO 9001-certified IRD i-Trop HPC (South Green Platform) at IRD Montpellier for providing high-performance computing resources.