Unprecedented genetic variability of PFam54 paralogs among Eurasian Lyme borreliosis-causing spirochetes

Abstract

Lyme borreliosis (LB) is the most common vector-borne disease in the Northern Hemisphere caused by spirochetes belonging to the Borrelia burgdorferi sensu lato (Bbsl) complex. Borrelia spirochetes circulate in obligatory transmission cycles between tick vectors and different vertebrate hosts. To successfully complete this complex transmission cycle, Bbsl encodes for an arsenal of proteins including the PFam54 protein family with known, or proposed, influences to reservoir host and/or vector adaptation. Even so, only fragmentary information is available regarding the naturally occurring level of variation in the PFam54 gene array especially in relation to Eurasian-distributed species. Utilizing whole genome data from isolates (n = 141) originated from three major LB-causing Borrelia species across Eurasia (B. afzelii, B. bavariensis, and B. garinii), we aimed to characterize the diversity of the PFam54 gene array in these isolates to facilitate understanding the evolution of PFam54 paralogs on an intra- and interspecies level. We found an extraordinarily high level of variation in the PFam54 gene array with 39 PFam54 paralogs belonging to 23 orthologous groups including five novel paralogs. Even so, the gene array appears to have remained fairly stable over the evolutionary history of the studied Borrelia species. Interestingly, genes outside Clade IV, which contains genes encoding for proteins associated with Borrelia pathogenesis, more frequently displayed signatures of diversifying selection between clades that differ in hypothesized vector or host species. This could suggest that non-Clade IV paralogs play a more important role in host and/or vector adaptation than previously expected, which would require future lab-based studies to validate.

Keywords: Borrelia burgdorferi sensu lato; Lyme borreliosis; PFam54 gene array; gene evolution; host‐pathogen interactions; spirochetes.

FIGURE 1

Diversity and prevalence of PFam54 architecture types identified in the 141 Eurasian isolates of B. afzelii, B. bavariensis, and B. garinii analyzed in this study. (a) Frequency of different architecture types present in the studied isolates (n = 141). Architecture types are separated by species (B. afzelii, B. bavarinesis, B. garinii). Colors correspond to tick transmission cycle (dark gray, I. persulcatus; light gray, I. ricinus). For information on the architecture of individual isolates see Table S3. (b) MDS clustering based on presence/absence matrix of orthology groups (OG) in each isolate. Here each point corresponds to a single Borrelia isolate. (c) Saturation curve analysis produced per species per transmission. In panels B and C, colors correspond to Borrelia species (purple, B. afzelii; orange, B. bavariensis; blue, B. garinii) and shapes correspond to the hypothesis tick transmission cycle based on the geographic origin of the isolates (I. persulcatus ▲, I. ricinus ●).

FIGURE 2

Gene gain and loss events of PFam54 paralogs based on the principle of maximum parsimony mapped onto the phylogenetic tree reconstructed based on full lp54 sequences corrected for recombination based on the four‐gamete condition test described in (Gatzmann et al., ; Rollins et al., 2023) and with the PFam54 gene array removed. Phylogenetic reconstruction was performed in MrBayes v.3.2.6 (Huelsenbeck & Ronquist, ; Ronquist et al., 2012) with ploidy set to haploid and a GTR (Tavaré, 1986) substitution model with inverse gamma‐distributed rate variation. Three independent runs were launched and ran for 10 million generations at which point convergence of parameters was checked with Tracer v.1.7.1 (Rambaut et al., 2018). Consensus trees were built using the sumt command from MrBayes using a respective burn‐in of 25%. Convergence to a single topology in all three independent runs was checked manually in FigTree v.1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/). Colors correspond to the hypothesis tick transmission cycle based on the geographic origin of the isolates (I. persulcatus or I. ricinus).

FIGURE 3

Phylogenetic tree of all, unique PFam54 paralogs identified in our analysis. In total, 1302 paralogs were identified which represented 524 unique sequences. Phylogenetic reconstruction was run in MrBayes v.3.2.6 (Huelsenbeck & Ronquist, ; Ronquist et al., 2012) with ploidy set to haploid and a codon substitution model with inverse gamma‐distributed rate variation, the universal genetic code, and assuming equal selection (ω) (Goldman & Yang, ; Muse & Gaut, 1994). Three independent runs were launched and ran for 50 million generations at which point convergence of parameters was checked with Tracer v.1.7.1 (Rambaut et al., 2018). Consensus trees were built using the sumt command from MrBayes using a respective burn‐in of 25%. Convergence to a single topology in all three independent runs was checked manually in FigTree v.1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/). All internal nodes that had probabilities lower than 0.95 are shown in light gray. Orthology groups (OGs), based on the monophyletic clustering of an individual gene copy within the tree either for a single or multiple species, are shown in the outer ring. Hypothesized tick transmission cycle based on the geographic origin of the isolate is shown as either dark gray (I. persulcatus) or light gray (I. ricinus) in the inner ring. Individual paralog placements in the phylogeny are marked outside of these rings. Borrelia species are denoted by tip end color: purple (B. afzelii), orange (B. bavariensis), blue (B. garinii). Roman numerals (I–V) denote the PFam54 Clades as described in (Wywial et al., 2009). Red stars denote branches that were found to show significant instances of diversifying selection as determined by aBSREL v2.2 (Smith et al., 2015) from the HyPhy package (https://www.hyphy.org/) using the universal genetic code and not allowing for multiple hits.