An Engineered Distant Homolog of Pseudomonas syringae TTSS Effector From Physcomitrella patens Can Act as a Bacterial Virulence Factor

Abstract

Pseudomonas syringae pv. phaseolicola is the causative agent of halo blight in common bean (Phaseolus vulgaris). Similar to other pathogenic gram-negative bacteria, it secrets a set of type III effectors into host cells to subvert defense mechanisms. HopQ1 (for Hrp outer protein Q) is one of these type III effectors contributing to virulence of bacteria. Upon delivery into a plant cell, HopQ1 undergoes phosphorylation, binds host 14-3-3 proteins and suppresses defense-related signaling. Some plants however, evolved systems to recognize HopQ1 and respond to its presence and thus to prevent infection. HopQ1 shows homology to Nucleoside Hydrolases (NHs), but it contains a modified calcium binding motif not found in the canonical enzymes. CLuster ANalysis of Sequences (CLANS) revealed that HopQ1 and alike proteins make a distinct group of putative NHs located distantly from the classical enzymes. The HopQ1 - like protein (HLP) group comprises sequences from plant pathogenic bacteria, fungi, and lower plants. Our data suggest that the evolution of HopQ1 homologs in bacteria, fungi, and algae was independent. The location of moss HopQ1 homologs inside the fungal clade indicates a possibility of horizontal gene transfer (HGT) between those taxa. We identified a HLP in the moss Physcomitrella patens. Our experiments show that this protein (referred to as PpHLP) extended by a TTSS signal of HopQ1 promoted P. syringae growth in bean and was recognized by Nicotiana benthamiana immune system. Thus, despite the low sequence similarity to HopQ1 the engineered PpHLP acted as a bacterial virulence factor and displayed similar to HopQ1 virulence properties.

Keywords: HopQ1; Physcomitrella patens; Pseudomonas syringae; evolution; horizontal gene transfer; type three effector.

FIGURE 1

The results of CLANS clustering of nucleoside hydrolase homologs at 1E-30 E-value threshold. HopQ1 amino acid sequence was used as a query to search for homologs using jackhmmer tool (https://www.ebi.ac.uk/Tools/hmmer/search/jackhmmer; sequence significance E-value - 0.01; hit significance E-value - 0.03). After two iterations, CD-HIT (http://weizhongli-lab.org/cdhit_suite/cgi-bin/index.cgi?cmd=cd-hit) (Huang et al., 2010) clustering at 90% sequence identity cut-off and manual reduction of the number of hits, the remained sequences were analyzed by CLANS (https:// toolkit.tuebingen.mpg.de/#/tools/clans) program from the same toolkit. The CLANS program allows to join sequences with high homology and closest evolutionary connection into groups (clans). The five largest clans are depicted including HopQ1 homologs (HLPs). A reduced representation (sequences grouped at 70% percentage identity with CD-HIT) was used to visualize the connections in CytoScape (Shannon et al., 2003). Node color corresponds to taxonomic classification, node size is proportional to the number of sequences grouped (logarithmic scale). Shading of the edges corresponds to the best BLAST E-value (stronger similarities equal darker edges). The initial visualization was obtained in CytoScape using edge-weighted, spring-based layout and the final positions were manually adjusted for clarity. The HopQ1 clan is surrounded by a green line.

FIGURE 2

Bayesian phylogenetic tree of HLPs. The tree was rooted with classical nucleoside hydrolase RihA from Escherichia coli. Branches representing main taxa were highlighted: bacteria in green, algae in orange, Dothideomycetes in teal, Sordariomycetes in gray, Cystobasidiomycetes in purple, mosses in brown, Leotiomycetes in navy. Each taxon is marked with black bracket. The analysis was carried out in PhyloBayes-MPI v. 1.5 (Lartillot et al., 2009). Three chains of 40000 iterations were run in parallel, two chains of best convergence were chosen for the consensus tree construction.

FIGURE 3

Assessment of virulence properties of TTSS-PpHLP. Bean leaves were inoculated with Pseudomonas syringae pv. tomato DC3000Δ28E (approximately 10⁵ cfu mL^-1) strains expressing HopQ1 or TTSS-PpHLP. Immediately prior to infiltration, bacteria were mixed in a 1:1 ratio. Two and 6 days post-inoculation (dpi), two leaf disks per plant were cut out from the infiltrated zones, ground in sterile 10 mm MgCl₂, diluted, and plated on LB medium. Bacterial strains were distinguished by a selectable marker. The CI (competitive index) was calculated as the ratio of bacteria expressing TTSS-PpHLP to bacteria expressing wild-type HopQ1 isolated from plant leaf and normalized to the input titers of the bacteria. Asterisks indicate that the index is significantly different from 1, as established using Student’s t-test (P < 0.001). Pluses represent the means. The experiment was performed three times with similar results.

FIGURE 4

PpHLP is recognized by plant immune system. Nicotiana benthamiana plants were inoculated with PsyB728a wild-type strain (A) or strains carrying pBBR1-MCS2 derivatives, which express TTSS-mCherry (B), TTSS-RihA (C), HopQ1 (D), or TTSS-PpHLP (E). Disease symptoms developed only in control plants, that is plants treated with the wild-type PsyB728a, PsyB728a carrying plasmids encoding TTSS-mCherry or TTSS-RihA. The photographs were taken 7 days post-inoculation. The experiment was performed twice, with similar results.