Transcriptional Profiling of Three Pseudomonas syringae pv. actinidiae Biovars Reveals Different Responses to Apoplast-Like Conditions Related to Strain Virulence on the Host.

Elodie Vandelle,Alice Regaiolo,Davide Danzi,Maria-Rita Puttilli,Teresa Colombo,Vanessa Maurizio,Annalisa Polverari,Tommaso Libardi

doi:10.1094/mpmi-09-20-0248-r

Elodie Vandelle, Alice Regaiolo + Show 6 more

Open Access

https://doi.org/10.1094/mpmi-09-20-0248-r

Copy DOI

Abstract

Pseudomonas syringae pv. actinidiae is a phytopathogen that causes devastating bacterial canker in kiwifruit. Among five biovars defined by genetic, biochemical, and virulence traits, P. syringae pv. actinidiae biovar 3 (Psa3) is the most aggressive and is responsible for the most recent reported outbreaks; however, the molecular basis of its heightened virulence is unclear. Therefore, we designed the first P. syringae multistrain whole-genome microarray, encompassing biovars Psa1, Psa2, and Psa3 and the well-established model P. syringae pv. tomato, and analyzed early bacterial responses to an apoplast-like minimal medium. Transcriptomic profiling revealed i) the strong activation in Psa3 of all hypersensitive reaction and pathogenicity (hrp) and hrp conserved (hrc) cluster genes, encoding components of the type III secretion system required for bacterial pathogenicity and involved in responses to environmental signals; ii) potential repression of the hrp/hrc cluster in Psa2; and iii) activation of flagellum-dependent cell motility and chemotaxis genes in Psa1. The detailed investigation of three gene families encoding upstream regulatory proteins (histidine kinases, their cognate response regulators, and proteins with diguanylate cyclase or phosphodiesterase domains) indicated that cyclic di-GMP may be a key regulator of virulence in P. syringae pv. actinidiae biovars. The gene expression data were supported by the quantification of biofilm formation. Our findings suggest that diverse early responses to the host apoplast, even among bacteria belonging to the same pathovar, can lead to different virulence strategies and may explain the differing outcomes of infections. Based on our detailed structural analysis of hrp operons, we also propose a revision of hrp cluster organization and operon regulation in P. syringae.[Formula: see text] Copyright © 2021 The Author(s). This is an open access article distributed under the CC BY-NC-ND 4.0 International license.

Highlights

To find transcriptomic differences among different P. syringae pv. actinidiae biovars, we designed a multistrain microarray containing the whole set of annotated sequences from four P. syringae pv. actinidiae strains belonging to the three main biovars best characterized at the time: NCPPB3739 (Psa1), ICMP19073 (Psa2), ICMP18884 (Psa3), and CRA-FRU8.43 (Psa3)
Given the importance of the type III secretion system (TTSS) in bacterial pathogenicity, and based on the differential expression of hrp and hrc genes in the P. syringae pv. actinidiae biovars described above, we investigated the correlation between hrp and hrc gene expression and virulence in more detail
PscA was upregulated in Psa2 in minimal medium at 4 h, it was still expressed at higher levels in Psa3 in both media at both time points. This low level of pscA expression may contribute to the higher c-di-GMP levels in Psa2 which, in turn, would suppress motility-related genes such as flhA, fliE, and fliN (Wang et al 2019). We found that these genes were expressed at lower levels in Psa2 than Psa3 and, Psa2 showed a lower swarming motility capacity compared with the other P. syringae pv. actinidiae biovars

Summary

Methods

The sequences and gene annotation data were collected in July 2015 from the NCBI GenBank and RefSeq databases and the University of Udine If multiple genome annotations were available for a given strain, a unique reference transcriptome was created by merging the data to avoid artificial redundancy at the strain level, giving the following priority for each transcript: RefSeq > GenBank > University of Udine (Pruitt et al 2005). The remaining set of probes targeting a 60-mer string shared by two or more transcripts (here dubbed “ambiguous probes”, n = 4,141) reflected residual transcript redundancy retained in our comprehensive microarray design originating from both biological (multiple strains sharing related sequences) and artificial (multiple collections of genome annotation data) sources. We favored full representation of annotated sequences to minimize hybridization biases due to strain-specific variation in transcript sequences and to conservatively treat inherent uncertainty of annotated sets of transcripts

Results

Discussion

Conclusion