Genomic and functional analysis of phage-mediated horizontal gene transfer in Pseudomonas syringae on the plant surface

Abstract

Many strains of Pseudomonas colonise plant surfaces, including the cherry canker pathogens, Pseudomonas syringae pathovars syringae and morsprunorum. We have examined the genomic diversity of P. syringae in the cherry phyllosphere and focused on the role of prophages in transfer of genes encoding Type 3 secreted effector (T3SE) proteins contributing to the evolution of virulence. Phylogenomic analysis was carried out on epiphytic pseudomonads in the UK orchards. Significant differences in epiphytic populations occurred between regions. Nonpathogenic strains were found to contain reservoirs of T3SE genes. Members of P. syringae phylogroups 4 and 10 were identified for the first time from Prunus. Using bioinformatics, we explored the presence of the gene encoding T3SE HopAR1 within related prophage sequences in diverse P. syringae strains including cherry epiphytes and pathogens. Results indicated that horizontal gene transfer (HGT) of this effector between phylogroups may have involved phage. Prophages containing hopAR1 were demonstrated to excise, circularise and transfer the gene on the leaf surface. The phyllosphere provides a dynamic environment for prophage-mediated gene exchange and the potential for the emergence of new more virulent pathotypes. Our results suggest that genome-based epidemiological surveillance of environmental populations will allow the timely application of control measures to prevent damaging diseases.

Keywords: Pseudomonas syringae; bacterial evolution; genomic diversity; horizontal gene transfer (HGT); phyllosphere; prophage.

Fig. 1

Diverse Pseudomonas isolates were recovered from cherry and showed distinct regional three differences. (a) A multilocus sequence typing‐based phylogenetic tree was built using five housekeeping genes (acnB, gltA, gyrB, pgi and rpoD) comparing isolates with genomes from NCBI. The tree was rooted to Pseudomonas putida, and the scale is substitutions per site. Isolates recovered from cherry in this work have the reduced naming system in the inner wheel and are defined within the rings 2–5 of the heatmap. Other Pseudomonas strains are labelled by assembly name, and known pathogens of stone fruits are highlighted in red (cherry), purple (other Prunus) and pink (other Rosaceae). The Pseudomonas syringae species complex phylogroups (PG) 1–13 are highlighted in blue and phylogroups (P) are noted on the dendrogram (black numerals) and outside the outer circle. The heatmap rings going outwards: (1) average nucleotide identity (ANI) 95% species clusters labelled for new isolates (used in comparative analysis in B); (2) region of isolation (WM, West Midlands; SW, Southwest; SE, Southeast); (3) orchards within these regions coloured in the same base colour as region; (4) cherry variety (SH, Sweetheart; Pe, Penny; La, Lapins; Ko, Kordia); (5) cherry tissue (green, leaf; brown, woody). (b) Balloon plot showing the frequency of each ANI species cluster dependent on region, cherry variety or tissue. Clusters are labelled from A to T corresponding to the groups in A. Dots are proportional to the number of isolates in each species group. The total number of isolates is listed at the bottom.

Fig. 2

Core genome phylogenetics shows that the cherry phyllosphere supports diverse clades of Pseudomonas syringae, varying in Type 3 secreted effector (T3SE) and toxin repertoires. Strain identifiers (inner wheel) are highlighted if they were sampled in this study (blue), or are known pathogens of cherry (red), other Prunus (purple) or Rosaceae (pink), including P. syringae pvs syringae (Pss), morsprunorum race 1 and race 2 (Psm R1 and Psm R2), Ps. avii and P. cerasi (P. ce), as marked between heatmap rings 4 and 5. Phylogroups (P) are also identified between rings 4 and 5. The phylogenetic branches are coloured by average nucleotide identity (ANI) groups (≥ 95%). Heatmap going outwards: (1) region of isolation (REG); (2) orchards (ORC) within these regions coloured in the same base colour as region; (3) cherry variety (VAR) (Pe, Penny; SH, Sweetheart; La, Lapins; Ko, Kordia); (4) cherry tissue (TISS), (green, leaf; brown, woody); (5) presence of the canonical hrp gene‐encoded Type 3 secretion system (T3SS); (6) number of known T3SE genes; (7) presence of toxins (TOX) (from inner to outer circle: i, syringolin A; ii, syringomycin; iii, Syringopeptin). Major groups of interest are labelled on the tree with numbers referring to the phylogroup. The scale shows amino acid substitutions per site.

Fig. 3

Pathogenicity test of 14 representative isolates on cherry leaves. Cherry leaves were infiltrated with each isolate, and symptoms were scored after 7 d. Strain identity and phylogroup are shown. Cherry pathogens Pss9644 and Psm R1‐5244 were included for comparison. Symptom scoring was as in Hulin et al. (2018b). 0, no symptoms; 1, isolated flecking; 2, browning of < 50% of inoculation site; 3, browning > 50% of the site; 4, 100% browning; 5, 100% browning and spread from the zone of infiltration. Mock inoculation with 10 mM MgCl₂ caused no symptom development. Strains are coloured based on their Pseudomonas syringae phylogroup as recorded on the x‐axis. The box plots created in R (ggplot2) show median, 25^th and 75^th percentiles and outliers. Statistical groupings after ANOVA analysis are highlighted with different letters. This experiment was performed twice, and both datasets were combined and presented here.

Fig. 4

hopAR1 gene is located in prophage sequences common to known cherry pathogens and new isolates from the phyllosphere. (a) Presence of hopAR1 annotated onto the core genome phylogeny as displayed in Fig. 2 showing phylogroups (P) labelled on the tree and the identity of strains, from the cherry phyllosphere (light blue), known pathogens of cherry (red), other Prunus (purple) and Rosaceae (pink). The three heatmap circles going outwards record: 1, hopAR1 gene presence (blue); 2, presence of prophage (Pp) sequences surrounding hopAR1 and if these prophages were predicted to be active, ambiguous, inactive or the region was too short for accurate prediction; 3, the most homologous phage taxon predicted by Prophage Hunter. (b) A maximum‐likelihood phylogeny of the HopAR1 protein from 84 strains with full‐length HopAR1 alleles. Strains are highlighted as in (a). Circles at each tip are coloured based on the phylogroup of each strain as in phylogeny in (a). Bootstrap support values are labelled at inner nodes where they are below 100. Phylogroups (P) and cherry pathogen lineages are labelled. (c) Alignment of the 74 216 bp region including the hopAR1 gene in Psm R1‐5244 and 1‐12B. Similarity between the two sequences is shown on the above plot of nucleotide identity over sliding windows of 20 nucleotides. The plot is coloured by a scale from red (low) to green (high) identity ranging from 0% to 100%. The prophage region is within the dashed rectangle. The key denotes whether genes were annotated as phage or bacterial by Phaster and other key regions including the predicted attachment (att) sites. Int, integrase.

Fig. 5

Excision and circularisation of the prophage harbouring hopAR1 from Psm R1 and epiphytic phylogroup four isolates. (a) Schema of prophage excision from the host chromosome and circularisation occurring within bacterial cultures without induction treatments. Phage and bacterial attachment sites (attP and attB) are marked. Excision was detected using primers Excise F and R located in the bacterial chromosome; no product was expected in the absence of excision. Circularisation was detected using Circularise F and R, amplifying across the attP site. (b) Polymerase chain reaction products indicate prophage excision and circularisation occurring from Psm R1 strains 5244 and 5300. Psm race 2 (R2‐leaf) lacking this hopAR1 prophage was used as a bacterial control. N, no template control; sizing hyperLadder 1 kb (Bioline, London, UK). (c) Prophage excision from the host chromosome and circularisation in phylogroup 4 strains 1‐10F and 1‐12B.

Fig. 6

Induction of the hopAR1‐encoding prophage in Psm R1 is dependent on a functional phage integrase gene. (a, b) Detection of polymerase chain reaction products for hopAR1, and the phage endolysin gene from purified phage preparations after induction of Psm R1‐5244 and R1‐5300, following mitomycin C (MMC) and Chloroform (Chlo), UV radiation and MgSO₄ wash and UV radiation treatments. (c) A diagram of the gene knockout strategy. (d) Absence of induction after disruption of the integrase gene in Psm R1‐5244 prophage. Polymerase chain reaction analysis was performed for hopAR1 and phage endolysin genes after induction and purification as in (a). Amplification from bacterial DNA of 5244 and 5300 is shown as positive controls. N, no template control, sizing hyperLadder 1 kb (Bioline).

Fig. 7

Transfer of the hopAR1‐encoding prophage from Psm R1‐5244_Gm ^R into Pseudomonas syringae phylogroup 10 strain 3‐7F_Rif on cherry leaves following UV radiation. (a) Schema of leaf inoculation with Psm R1‐5244_Gm ^R (donor) and Ps 3‐7F_Rif (recipient), irradiation with UV and bacterial isolation. (b) Three Rif‐r Gm‐r resistant colonies were checked by polymerase chain reaction for the hopAR1 and endolysin genes, which are present in the prophage genome (and for hopY1, only present in 3‐7F, Supporting Information Fig. S6B). (c) Alignment of the tRNA‐cys region in Psm R1‐5244 and Ps 3‐7F to show site of integration into Ps 3‐7F. The tRNA‐cys and flanking regions are similar. The DNA alignment is colour‐coded with the key presented (green, phage; blue, bacterial; pink, hopAR1; white, flanking genes; yellow, attachment (att) sites).