Phylogeny of a relapsing fever Borrelia species transmitted by the hard tick Ixodes scapularis

Abstract

The discovery of Borrelia species that were related to the agents of relapsing fever but were transmitted by hard ticks rather than soft ticks challenged previous taxonomies based largely on microbe-host specificities and geographic considerations. One of these newly-identified organisms is the Borrelia miyamotoi sensu lato strain LB-2001 from North America and transmitted by Ixodes scapularis. This or related strains have been identified as the cause of human disease, but comparatively little is known about their biology or genetics. Using recently acquired chromosome sequence of LB-2001 together with database sequences and additional sequences determined here, I carried out comparisons of the several species of Borrelia, including those in the two major clades: the relapsing fever group of species and the Lyme disease group of species. Phylogenetic inference at the species level was based on four data sets: whole chromosomes of ∼1Mb each, and concatenated sequences of 19 ribosomal protein genes, 3 conserved nucleic acid enzymes (rpoC, recC, and dnaE), and 4 contiguous genes for nucleotide salvage on a large plasmid. Analyses using neighbor-joining, maximum likelihood, and Bayesian methods were largely concordant for each of the trees. They showed that LB-2001 and related hard tick-associated organisms, like Borrelia lonestari, are deeply positioned within the RF group of species and that these organisms did not, as some earlier estimations had suggested, constitute a paraphyletic group. The analyses also provided further evidence that major changes in host ranges and life cycles, such as hard to soft ticks or vice versa, may not correlate well with overall sequence differences. The genetic differences between LB-2001 and B. miyamotoi sensu stricto justify provisional use of the "sensu lato" designation for LB-2001.

Keywords: Genomics; Lyme disease; Spirochetes; Tick-borne disease.

Figure 1

Scatter plot with linear regression of the CG dinucleotide content on G+C nucleotide content of chromosomes of relapsing fever group Borrelia spp. (red) and Lyme disease group Borrelia spp. (blue).

Figure 2

Dot matrix plots of aligned LB-2001 chromosome sequence against B. hermsii strain DAH chromosome sequence (left panel) and B. burgdorferi strain B31 chromosome sequence. Lines shown are similarities on direct strands (red) or on opposite strands (blue). Sequences were aligned using the default settings of the MAFFT alignment program.

Figure 3

Phylogram of ungapped aligned chromosome sequences of LB-2001 and four other Borrelia species by BioNJ protocol for observed differences (NJ) or by maximum likelihood (ML) with the General Time Reversible model. Nodes with bootstrap values of >70% support after 1000 replicates for NJ and 100 replicates for ML are shown. The scale bar represents nucleotide substitutions per site.

Figure 4

Phylogram of codon-aligned ungapped concatenated nucleotide sequences of 19 ribosomal protein genes of LB-2001 and 11 other Borrelia species as inferred by neighbor-joining (NJ; 1000 replicates), maximum likelihood (ML; 100 replicates), and Bayesian analysis (see section 2.3). The evolutionary model for ML and Bayesian analysis was General Time Reversible (GTR) with a discrete gamma model of 4 categories and a shape parameter of 0.25; it began with a preliminary BioNJ tree. Node values are percent. The scale bar represents nucleotide substitutions per site.

Figure 5

Phylogram of codon-aligned ungapped concatenated nucleotide sequences for chromosome-encoded rpoC, recC, and dnaE (panel A) and plasmid-encoded thyX, nrdI, nrdE, and nrdF (panel B) of LB-2001 and other Borrelia species, as inferred by neighbor-joining (NJ; 1000 replicates), maximum likelihood (ML; 1000 replicates), and Bayesian analysis (see section 2.3). For models for both were GTR with discrete gamma model of 4 categories and a shape parameter of 0.29. The scale bar represents nucleotide substitutions per site.

Figure 6

Alignment of N-termini of selected unprocessed Vsp proteins of LB-2001, B. miyamotoi sensu stricto HT31, B. hermsii DAH, B. turicatae Oz1, and B. anserina Es, as well as the homologous OspC protein of B. burgdorferi B31. Underlined residues are the consensus signal peptidase 2 sites in each protein.

Figure 7

Phylograms of ungapped aligned chromosome sequences of selected relapsing fever group Borrelia species and strains (left; 891,745 sites) and Lyme disease group species and strains (right; 893,252 sites) by BioNJ neighbor joining protocol for observed differences. Nodes with bootstrap values of >70% support after 1000 replicates. The phylograms are to the same scale. The bar represents nucleotide substitutions per site.