copLAB gene prevalence and diversity among Trinidadian Xanthomonas spp. black-rot lesion isolates with variable copper resistance profiles

Agricultural sites involved in the study with varying intensities of cultivation

Leaf samples of cabbage and cauliflower showing symptoms of black-rot infection were collected from two fields at each location indicated in Fig. 1. The farming districts of Maloney and Aranguez have >30-year history of agricultural land use, whereas Navet and, Bon Air have a shorter history (<15 years). Each agricultural district has been associated with crucifer cultivation, with a greater incidence of monoculture in the Navet district.

Sample collection and bacterial isolation

Leaf samples were stored at 4 °C, and bacteria were isolated within 24 h according to Lugo et al. (2013). Isolates similar to or with matching morphologies to Xanthomonas (yellow, convex shaped colonies) were selected after 48 h incubation on nutrient agar (NA). These isolates were labelled as the lesion-associated population. Further investigation into the distribution of copper-resistant bacteria in the Aranguez district was carried out in 2017 to determine the distribution of variant BrA1 copLAB genes between bacteria at an active and abandoned crucifer field. Phylloplane-associated and soil-borne bacteria were isolated from plant material and soil respectively from an active cauliflower field and abandoned plot approximately 300 m away. Isolates were obtained by plating dilution washes on NA. Morphologically unique colonies were selected after 48 h and labelled as the environmental reference population. In total 45 lesion-associated, and 199 environmental population isolates were retained and stored at −80 °C in 25% glycerol/Nutrient Broth.

Taxonomic assignment of bacterial isolates in the lesion-associated and environmental reference population

Lesion-associated isolates displaying similar morphologies to Xanthomonas were identified as XCC using primers (XCF-CGATTCGGCCATGAATGACT, XCR-CTGTTGATGGTGGTCTGCAA) targeting a highly conserved region of the Xcc hrpF gene (Park et al., 2004) or as a member of the Xanthomonas genus (Xsp) using primers targeting the hrpB6 gene (RST2-AGGCCCTGGAAGGTGCCCTGGA, RST3-ATCGCACTGCGTACCGCGCGCGA) (Leite et al., 1994), or using RT 16 F/R primers targeting the gumL gene (Pandey et al., 2016). Isolates with expected amplicons for XCF/R and RT 16 F/R primers were labelled as Xcc while those with RST2/3 and or RT 16 F/R bands were labelled as Xsp. Strains without amplification with any primer were not used in further analysis. Classification using this scheme was corroborated using partial 16s rRNA sequencing (Macrogen, Korea), which was also applied to environmental isolates. Sequences with QV >16 were selected and aligned using BLAST (BLASTN, https://blast.ncbi.nlm.nih.gov/) against the NCBI 16s rRNA database. Only copper-resistant environmental isolates were identified using 16s rRNA sequencing. While species-level identification was obtained for most isolates (Coverage >90% and %ID >98%), only identified genera were considered for consistency across the environmental dataset. WGS characterised local Xcc and Xanthomonas melonis (Xmel) isolates (Ramnarine, Jayaraman & Ramsubhag, 2022) were included as controls. Partial 16s rRNA Sanger-generated sequences can be accessed at https://doi.org/10.5281/zenodo.6795859.

Copper sensitivity profiling

Copper sensitivity screening was carried out according to Ramnarine, Jayaraman & Ramsubhag (2022) using CuSO₄.5H₂O concentrations ranging from 0–2.4 mM amended into Mannitol Glutamate Yeast Agar. Briefly, the copper sensitivity of isolates was categorized based on the MIC of CuSO₄.5H₂O of isolates after 48 h incubation on amended media at 28 °C. Copper-sensitive isolates only grew in <0.6 mM CuSO₄.5H₂O, copper tolerant strains survived concentrations between 0.6–0.8 mM while copper resistant isolates survived >0.8 mM. Amended media was prepared using a filter sterilized CuSO₄.5H₂O stock solution both prepared fresh on the day of use. A total of 48 h cultures of bacterial isolates were suspended in sterile water and 5 µL spot plate onto solid amended agar in triplicate.

copLAB gene detection in lesion-associated and environmental isolates

Primers were designed for the BrA1 variant copLAB gene sequences (Behlau et al., 2017) using reference homologs (similarity >= 60%) from the Xanthomonas and Stenotrophomonas genera available from GenBank. These were aligned using ClustalW in BioEdit v7.0.5.3 (Hall, 1999)⁠. Variant-specific primers targeting unaligned regions unique to the BrA1 variants were manually designed and screened for specificity using PrimerBlast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/).⁠ Those predicted to amplify variant genes unambiguously were further screened using gDNA of Xcc strain BrA1, and the amplicons were Sanger sequenced (Macrogen, Seoul, Korea) before cross-referencing to the draft BrA1 genome contig containing the variant genes (GCA_002806765.1). Primers targeting aligned regions from reference homologs (similarity >= 60%) from the Xanthomonas and Stenotrophomonas genera available from GenBank were designed to target previously reported sequences. Primer pairs were screened for specificity using PrimerBlast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/).⁠ These primers were in-silico verified to amplify those from plasmids, characterised genes from copper resistant Xsp and those from Stenotrophomonas chromosomes with identical synteny to copLABMGF copper resistant operons. All primers were synthesized at IDT, sequences and amplicon length are listed in Table 1.

Total genomic DNA was isolated from all lesion-associated and copper-resistant environmental isolates according to Wilson (2001) and quality was assessed using standard gel electrophoresis methods (1% agarose). PCR amplification using designed primers (Table 1) and molecular reagents from Bioland Scientific were carried out in 25 µL reactions as outlined in Ramnarine, Jayaraman & Ramsubhag (2022).

copLAB gene sequencing and evolutionary analysis of variants

Amplicons generated from primers targeting both variant and previously reported genes were gel purified using the Wizard® SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA) and, sequenced on an ABI3730XL platform (Macrogen, Seoul, Korea). An overall QV score threshold of <16 was set for sequence rejection and end trimming. Pairwise blast using the BLASTN and TBLASTN (https://blast.ncbi.nlm.nih.gov/) algorithms were used to verify the homology with annotated copLAB reference sequences against the entire NCBI GenBank database. Final curated sequences, GenBank references, and full gene sequences from the Xcc BrA1 and other local Xsp genomes⁠ (Table S1) were submitted to ngphylogeny (https://ngphylogeny.fr/) for evolutionary analysis. Using this server, multiple alignment and curation were carried out using MAFFT (Katoh & Standley, 2013) and BMGE (Criscuolo & Gribaldo, 2010) respectively. Alignments were then submitted to FastTree (Price, Dehal & Arkin, 2010) for maximum likelihood phylogenetic reconstruction with 1,000 bootstrap branch support (Felsenstein, 1985). The bootstrap consensus tree for each dataset was visualized and edited in TreeGraph 2 (Stöver & Müller, 2010)⁠. Alignments of copLAB sequences were also analysed in MEGA 11 (https://www.megasoftware.net/) to identify variable sites. Sanger-generated sequences for amplified copLAB genes can be accessed at https://doi.org/10.5281/zenodo.6795859.

Statistical inference

Statistical analysis was carried out on lesion associated and environmental isolate datasets using RStudio (2023.03.0+386; R Studio Team, 2023) and R version 4.2.3 (R Core Team, 2023). Statistical tests included Anova, TukeyHSD post-hoc test and Pearson’s Chi-squared Test from the R stats package (v4.2.3).

copLAB gene prevalence in copper-resistant lesion-associated Xanthomonas spp

The 45 lesion-associated isolates were distributed by agricultural district as follows: Aranguez = 12, Maloney = 5, Navet = 24, Bon Air = 4. These are represented in Table 2 which includes species-level identification, copper sensitivity profiles and max CuSO₄.5H₂O MIC. Most isolates from this population were copper resistant (78%) but only 31% were Xcc. There were <7% tolerant and only one sensitive isolate obtained but profiles of five isolates could not be determined. Most (80%) of the copper-resistant isolates grew in the presence of 2.4 mM CuSO₄.5H₂O with the lowest MIC for this category being 1.2 mM. Among the Xanthomonas species there were significant differences in CuSO₄.5H₂O MIC means (P = 0.0176) but only between Xsp and Xcc (P = 0.0134). However, no significant difference was observed between MIC’s, cop genotype and sample location.

Figure 2 only represents copper resistant isolates (35) which survived four CuSO₄.5H₂O concentrations represented in Table 2, 1.2 mM (2), 1.6 mM (4), 2 mM (1) and 2.4 mM (28). Only isolates from Aranguez (11) contained all three variant BrA1 copLAB genes (Variant, Fig. 2). Previously reported copLAB genes (Traditional, Fig. 2) were found in Xsp, Xmel and Xcc strains from Maloney, Bon Air and Navet. The BrA1 Variant copLAB primers were able to amplify individual copLAB genes in WGS characterised local Xcc. Primers targeting Xanthomonad copLAB genes amplified those in characterised Xmel strains described in Ramnarine, Jayaraman & Ramsubhag (2022). No significant difference in MIC’s were observed between variant and previously reported cop genotypes. However, there was a significant difference in cop genotypes (P = <2e−16) by location, specifically when comparing Aranguez to the other three districts.

Two Xcc and one Xsp from Navet and, two Xsp from Maloney were characterised as copper resistant, but PCR amplification was not able to detect any copLAB gene. Only one sensitive (Xcc CaNP1C) and three tolerant (Xcc Ar1PC2 and CNP4A and Xmel CaNP1D) strains were isolated and were all copLAB PCR negative. All 45 isolates and associated isolation location and PCR prevalence data are given in Table S2. In brief, most of the 45 Xanthomonas isolates were copper resistant (80%) up to 2.4 mM CuSO₄.5H₂O. From these isolates, only Xcc strains from the Aranguez district contained the BrA1 variant copLAB gene cluster. Previously reported copLAB genes were also identified in copper resistant isolates from other agricultural districts using the primers designed in this study. However, four copper resistant strains, including Xcc, were PCR negative for copLAB genes. Significant differences in CuSO₄.5H₂O MIC were observed between Xcc and Xsp strains and with the BrA1 variant cop genotype in Aranguez compared to the other districts.

BrA1 variant copLAB gene prevalence among copper resistant environmental phylloplane and soil associated bacteria from the Aranguez district

Differences in the proportions of isolates characterised as copper sensitive, tolerant and resistant in the environmental population are summarised in Fig. 3. Full isolate profiles are given in Table S3. Greater proportions of resistant isolates were obtained from the phylloplane and soil of the cultivated field, with the highest proportions from the latter environment (57%) (Fig. 3B). While copper-resistant isolates were found in all environments and higher proportions were observed among soil isolates (Fig. 3B). Copper tolerant and sensitive isolate proportions were higher from the phylloplane and soil of the abandoned plot respectively (20% vs. 48% and 20% vs. 49%). Overall, of the 199 isolates screened, 78 were copper sensitive, 46 tolerant and 75 were resistant. Pearson’s Chi-squared test showed a significant association between copper sensitivity and, the sampling environment and location (P = 0.0006). Using Pearson residuals, a positive correlation was observed between copper tolerant bacterial strains and the phylloplane environment of the abandoned field. This was also seen with copper resistant strains and the soil environment of the cultivated field. In summary, there were statistically significant differences in copper sensitive isolates from the phylloplane and soil of a cultivated and abandoned field. From this there were greater positive correlations with copper resistant isolates and the cultivated field.

The 75 copper-resistant isolates from the environmental population were screened for the variant copLAB genes and identified using 16s rRNA gene sequencing (Table 3). Interestingly, no copper-resistant Xsp were identified within this sub-group of environmental isolates. Notable is the higher diversity in soil and phylloplane from the cultivated field and the presence of Stenotrophomonas spp. isolates (Table 3) in all locations except the phylloplane of the abandoned field. More isolates from this genus were obtained from the cultivated field which also had a greater number of Pseudomonas spp., Achromobacter spp. and unidentified isolates. However, no copper-resistant isolate in the environmental population contained the BrA1 variant copLAB gene.

Evolutionary analysis of copLAB genotypes

Sequences from local X. sp. isolates were similar (>96%) to previously reported cop sequences from Xanthomonas plasmids or Stenotrophomonas spp. chromosomes. Furthermore, these sequences were not clustered with coh chromosomal homologs further establishing confidence in their designation as cop sequences. In line with established similarity patterns in global strains, copA (Fig. 4B) and copB (Fig. 4C) genes from local isolates appear to be more closely related than copL (Fig. 4A). In each phylogeny, three distinct groupings of PCR amplified copLAB genes from local isolates were observed. Group 1 and 3 copL sequences, were most distantly placed when compared to all other sequences. Both groups consisted of genes ~93.5% similar to each other. Isolates in these groups were all from Navet. Group 1 strains had a CuSO₄.5H₂O MIC of 400 ppm with Group 3 being 600 ppm. Notably, these identical Group 1 sequences were obtained from two Xanthomonas species (Xcc and Xsp). When compared to X. euvesicatoria pLMG930.1, four variable sites in Group 1 copL sequences from both Xcc and Xsp were identified. Group 3 sequences were identical for the aligned region but did not cluster with a previously reported sequence. However, when compared to Group 1, there were 22 variable sites in the aligned sequences in Group 3. Interestingly, in the BrA1 variant Group 2, four variable sites were detected when compared to the BrA1 draft genome copL sequence. However, these variations were only present in Xcc Ca3B and Cf6A1. Overall, the copL sequences in all three groups contained 147 variable sites.

The copA sequences formed two distinct groups, with Group 1 consisting of the BrA1 variant genes. While Group 2 comprised Xsp and Xcc isolates from Navet, isolates in both groups had a CuSO₄.5H₂O MIC of 600 ppm. The Xcc CaNP6B sequence in this group almost entirely (99%) aligned with 99.8% similarity to multiple copA homologs in Stenotrophomonas spp. and Xanthomonas spp. plasmids. The Xsp sequences in Group 2 were identical to a partial copA CDS from X. perforans (MK616431) while local Xcc variant sequences clustered with the BrA1 variant and a Stenotrophomonas sp. copA gene. These variant sequences in Group 1 did not show the same trend as copL (Fig. 4A Group 2) as all sequences were identical. Furthermore, when compared to other sequences in Group 2, Xcc CaNP6B shared 15 variable sites with the Xsp sequences in this group (which were all identical). Overall, both groups shared 153 variable sites.

Previously reported copB sequences clustered into two closely related groups, 2 and 3 (Fig. 4C). Group 2 contained a single sequence from Xcc CaNP6B which was similar to sequences from the local Xmel CaNP6A genome but identical to multiple Stenotrophomonas spp. homologs (Similar to Fig. 4B, Group 2). Notably, the copL and copA genes from this Xcc strain did not show this clustering pattern. Group 3 sequences clustered with the well-referenced copL sequence from X. euvesicatoria citrumelonis (HM579937) showing 99.5% similarity at 99% coverage. Isolates from all groups had a CuSO₄.5H₂O MIC of 600 ppm. In Group 1, the BrA1 variant sequences were all identical (as seen in Fig. 4B Group 1). Despite their separate group designation, the closely placed Xcc sequence and Xsp sequences in Groups 2 and 3 (Fig. 4C) were compared. Xcc CaNP6B copB shared 46 variable sites with Group 3 sequences. Within Group 3, however, two groupings based on four variable sites were observed. These consisted of PNP25 and PNP34 and, PNP54 and PNP62. Overall, the sequences from all three groups shared 210 variable sites. As noted with the previous Xcc BrA1 study, variant sequences from local strains clustered separately from others. These variant sequences also clustered with and were identical to cop homologs from Stenotrophomonas sp. WZN-1 (CP021768.1). Variant copLAB sequences are represented in Fig. 4A (Group 2) and Figs. 4B and 4C (Group 1), respectively in blue boxes.