Diversity of quorum sensing pherotypes amongst ecotypes of plant-associated Bacillus subtilis sensu lato isolates.

Abstract

Ecologically-adapted populations, or ecotypes, are species forms that are functionally adapted to particular habitat niches. Bacterial ecotypes are challenging the significance and implications of ecological adaptation in terms of prokaryote taxonomy, community ecology, and biocontrol applications. Plant-associated members of the Bacillus subtilis sensu lato group—in particular B. subtilis and B. velezensis—are ecologically specialised to perform numerous functions that are beneficial to plant and soil health. Underpinning these beneficial activities is the ability to colonise plant surfaces by biofilm formation. Biofilms are a result of co-ordinated social behaviour amongst microorganisms. In the B. subtilis sensu lato this sociality is governed by the ComQXPA quorum-sensing cassette, which uses the ComX pheromone for intercellular communication leading social behaviours like biofilm formation. This peptide pheromone contains a post-translational modification on a conserved tryptophan residue. This modification is highly variable between populations, resulting in discrete ComX variants which form communication groupings known as pherotypes. The limitation of communication to within a pherotype may constitute a form of ecological adaptation designed to protect the products of co-operative behaviour and restrict their benefit to the producing population. The present study aimed to explore ecotypes and pherotypes amongst a subset of plantassociated B. subtilis and B. velezensis isolates. These isolates originated from phylloplane and rhizosphere samples from seven crop species grown across the KwaZulu-Natal province, South Africa, and had demonstrated biocontrol potential in previous studies. An exploratory study set out to apply in silico approaches to determine gene-sequence-based variation amongst representative strains of the B. subtilis sensu lato available in the GenBank database. Nine housekeeping gene targets (viz., 16S rRNA, cheA, dnaJ, groEL, gyrA, gyrB, polC, purH, and rpoB) were evaluated for suitability to resolve clustering of closely-related B. subtilis sensu lato. Four of these gene sequences (viz., 16S rRNA, dnaJ, gyrA, and rpoB) were identified as candidates for a Multilocus Sequence Analysis (MLSA) scheme to distinguish between members of the B. subtilis sensu lato group. Putative pherotype variation amongst these reference strains was explored in silico using comQXP gene sequence data. The suitability of a comQXP PCR-RFLP protocol with potential for rapid pherotyping amongst B. subtilis sensu lato was evaluated in silico using simulated comQXP amplicons. This necessitated the design of a PCR primer set targeting the quorum-sensing gene region of B. velezensis. Four restriction enzyme candidates namely, BtsCI, Fnu4HI, Cac8I, and Hpy166II, were identified for further study. Ecotyping amongst the B. subtilis and B. velezensis isolates was carried out using a four-gene (viz., 16S rRNA, dnaJ, gyrA, and rpoB) MLSA. This concatenated sequence dataset was applied to ecotype simulation (ES) analysis to corroborate putative ecotype sub-clusters in the MLSA phylogeny. Two DNA fingerprinting approaches—Repetitive Element Palindromic PCR (Rep-PCR) and Random Amplified Polymorphic DNA PCR (RAPD-PCR)—were also evaluated for their potential to identify putative ecotypes within the isolate subset. This investigation of pherotypes amongst the putative ecotype groupings examined isolate comQ sequence data as well as the comQXP PCR-RFLP, and also applied a srfA-LacZ reporter gene assay to examine isolate stimulation of seven known pherotype tester strains (viz., 168, RO-B- 2, RO-C-2, RO-E-2, RO-FF-1, RO-H-1, and RS-D-2). The MLSA of isolate gene sequences presented distinct sequence clusters suggestive of ecotype populations amongst the two Bacillus species which were corroborated by ES analysis. The MLSA and ES determined two putative ecotypes within the B. subtilis isolates, and six within the B. velezensis isolates. Ecotype groupings were found to contain isolates from different crop species and locations, and four B. velezensis ecotypes were distinct from included B. subtilis sensu lato reference strains. PCR fingerprinting identified strain-level variances amongst the isolates, and were able to differentiate plant-associated B. velezensis from closelyrelated B. amyloliquefaciens, but did not define groupings entirely consistent with the ES and MLSA phylotypes. The MLSA, ES, and PCR fingerprinting delineated a group of five isolates (viz., B81, CT-R67, bnd136, bng221, and sqo271) to be grouped with reference strain Bacillus sp. JS in the MLSA. This grouping is of interest as Bacillus sp. JS is a strain demonstrating biocontrol capability and formed a clade distinct to B. subtilis sensu stricto in gene sequence phylogenies. This Bacillus sp. JS grouping was further confirmed by ES to constitute a single ecotype, and both Rep-PCR and RAPD-PCR OPG-16 distinguished this grouping based on fingerprint profiles. Investigation of reference strain variation in the comQ gene sequence determined significant levels of sequence variation amongst reference strains evaluated in silico, and showed divergence in some strains from known pherotype counterparts. Furthermore, sequence clusters were resolved in B. velezensis that suggested unique pherotype variants amongst reference strains, and showed comQ similarity amongst the five isolates related to Bacillus sp. JS. This trend was observed for the comQ phylogeny applying isolate and pherotype sequences, which resolved two distinct clusters each amongst the B. velezensis and B. subtilis isolates, with only two isolates (viz., bnd134 and bnd156) grouping with known pherotypes. The B. velezensis clades demonstrated significant comQ sequence variance from the pherotype reference strains based on clustering distances in the phylogeny. Of the four enzyme candidates identified for potential use in a comQXP PCR-RFLP protocol, two were found to be applicable: BtsCI resolved the grouping of the Bacillus sp. JS clade, while Fnu4HI was found to distinguish between more closely-related isolates in B. velezensis. The profile groupings for both of these enzymes demonstrated most of the isolates to be distinct from known reference pherotype strains, and the technique proved able to corroborate groupings in the comQ phylogeny. The pherotyping srfA-LacZ reporter gene assay failed to support the pherotype groupings defined by the comQ phylogeny, but demonstrated that the environmental isolates were capable of eliciting substantial responses in the pherotype tester strains, particularly amongst the B. velezensis isolates. Eight isolates did not elicit significant responses in any of the tester strains, while most matched to multiple pherotypes on the basis of tester strain response to conditioned medium from isolates. These findings demonstrated that putative ecotype variation was present amongst the B. subtilis and B. velezensis isolates that were resolvable by MLSA, and that these ecotype groups comprised isolates from different crop types and location sites. Some of these ecotypes bore only distant relation to B. subtilis sensu lato reference strain counterparts. The investigation of PCR fingerprinting methods for ecotyping purposes found that the primer sets applied were not able to consistently corroborate putative ecotype groupings defined by MLSA and ES. Pherotyping investigations demonstrated that there was evidence of gene-sequence-based pherotype variation amongst the isolates within these putative ecotype groupings. The pherotype clades resolved with the comQ phylogeny and with comQXP PCR-RFLP were not corroborated by srfA-LacZ reporter gene pherotyping assays, but the data suggested that assay optimisation to suit environmental strains is warranted in future studies. This study presents the first instance of ecotyping and pherotyping amongst plant-beneficial B. velezensis. The B. velezensis isolates were largely distinct in ecotype and pherotype from the plant-associated model strain FZB42, which further promotes that these isolates demonstrate localised adaptation. The Bacillus sp. JS-related isolate grouping is of interest as these isolates comprised a single putative ecotype resolvable by MLSA, ES, and Rep-PCR and RAPD-PCR OPG-16. This grouping was distantly-related to B. subtilis sensu stricto, and this is interesting from a taxonomic standpoint, as is the fact that that Bacillus sp. JS is reported to have plantbeneficial capabilities. The existence of uniquely South African plant-associated B. subtilis sensu lato ecotypes is valuable in agricultural management approaches targeting beneficial microbes, as these organisms likely possess adaptations allowing them to compete with extant bacterial communities and thrive within the South African agricultural climate.