Fig 1.
Bartonella evolution experiment in rodent hosts.
(a) The design included 20 serial passages of Bartonella krasnovii OE1-1 A2 through individuals of one of the two host species (G. andersoni in green or G. pyramidum in purple) or alternating between them, with five independent rodent lines per treatment. (b) In each infection cycle, captive Bartonella-negative rodents were inoculated intradermally with bacteria. Bacteria were cultured from red blood cells sampled 15 days later to create the inocula for the next passage. A portion of each inoculum was archived for further study. Yellow rods and shading indicate bacteria in animal infections or samples. Figures incorporate artwork from Biorender.com.
Fig 2.
Bartonella evolution is dominated by mutations in mononucleotide simple sequence repeats (SSRs) in the badE trimeric autotransporter adhesin (TAA) gene.
(a) Dynamics of competition between bacteria with new mutations in each of 15 experimental populations (L1-L15) over 20 rodent infection cycles. The relative abundance of each mutation was determined from metagenomic sequencing data. (b) Number of mutations in each of 10 clones isolated at the end of the evolution experiment from each of 15 populations. Each panel shows results for the five populations that were passaged only in G. andersoni (L1, L4, L7, L10, L13), the five passaged only in G. pyramidum (L2, L5, L8, L11, L14), or the five passaged alternately in the two rodent species (L3, L6, L9, L12, L15). Sequencing data was insufficient to call mutations in the three clones marked ND (not determined). In a and b, mutations that altered the lengths of two mononucleotide SSRs in badE are shown in color. All other mutations are displayed in grey. For full details see S1 File. (c) Effects of badE SSR mutations. Key portions of the ancestral gene’s nucleotide and amino acid sequences are shown in the first two rows. SSR mutations and how they lead to frameshifts and premature stop codons that inactivate the gene are shown in the three bottom rows. Stop codons are indicated by red bars above and below nucleotide sequences. Dotted lines represent portions of the reading frames that are not displayed.
Fig 3.
Evolved SSR mutation in badE alters infection dynamics.
The B. krasnovii OE1-1 A2 ancestor strain or an evolved strain, differing only by a mutation in badE SSR2, were inoculated into three pairs of female and three pairs of male G. andersoni and G. pyramidum twins (one sibling was inoculated with the ancestor and the other with the evolved strain). Bars show mean ± s.e. of the log-transformed bacterial titer in blood samples collected on the specified days after inoculation. Points show values for individual hosts. *Significant difference between the mutant and ancestor groups. See the main text for statistics.
Fig 4.
Conservation of mononucleotide SSRs in TAA genes most similar to the mutated badE gene in Negev Desert Bartonella and their relatives.
The maximum likelihood phylogenetic tree of Bartonella strains on the left was constructed from a concatenated amino acid alignment of single-copy orthologs, with B. bovis 91–4 as the outgroup. Portions of the tree with very similar strains are collapsed into triangles. Every node in the tree not collapsed into a triangle was supported in all 100 bootstrap trees. The TAA gene in each Bartonella genome with the highest similarity to the first 200 amino acids of the B. krasnovii OE1-1 A2 badE gene that mutated during the evolution experiment was identified, and this region was aligned. The columns from this alignment that include or are adjacent to the two simple sequence repeats that mutated (SSR1 and SSR2) are shown on the right.
Fig 5.
Mononucleotide simple sequence repeats (SSRs) in the genomes of Negev Desert Bartonella.
(a) Number of mononucleotide SSRs with lengths ≥9 bases in the genomes of 38 strains of Bartonella belonging to four species that were isolated from rodents in the Negev Desert dunes (S3 File). Only SSRs overlapping protein-coding genes are included. Points are values for individual strains. Bars are averages within each species. (b) Fraction of mononucleotide SSRs in each species that are repeats of each base. Bases are specified on the coding strand of the gene containing the SSR. (c) Distribution of mononucleotide SSRs at different normalized positions within protein-coding open reading frames in each species. (d) Venn diagram showing the number of SSRs that occurred in at least two genomes from each of the four Bartonella species and how they are shared between these species. (e) Protein-coding genes exhibiting variation in the lengths of a mononucleotide SSR that result in frameshifts within the Negev Desert Bartonella strains. Numbers in boxes are the lengths of SSRs, with their backgrounds shaded by the repeated base as in b. A red X indicates a frameshift caused by the SSR in that gene. White boxes indicate that a gene was not identified in a genome. Black boxes indicate that the gene is present but does not have a mononucleotide SSR of ≥6 bases at that location. Asterisks indicate conserved genes of unknown function that were named in this study.
Table 1.
Highlighted genes with mononucleotide SSRs in Negev Desert Bartonella genomes.