Longitudinal Study of Bacterial Infectious Agents in a Community of Small Mammals in New Mexico

Abstract

Background and Objectives: Vector-borne bacterial diseases represent a substantial public health burden and rodents have been recognized as important reservoir hosts for many zoonotic pathogens. This study investigates bacterial pathogens in a small mammal community of the southwestern United States of America. Methods: A total of 473 samples from 13 wild rodent and 1 lagomorph species were tested for pathogens of public health significance: Bartonella, Brucella, Yersinia, Borrelia, Rickettsia spp., and Anaplasma phagocytophilum. Results: Three animals were positive for Yersinia pestis, and one Sylvilagus audubonii had a novel Borrelia sp. of the relapsing fever group. No Brucella, Rickettsia, or A. phagocytophilum infections were detected. Bartonella prevalence ranged between 0% and 87.5% by animal species, with 74.3% in the predominant Neotoma micropus and 78% in the second most abundant N. albigula. The mean duration of Bartonella bacteremia in mark-recaptured N. micropus and N. albigula was 4.4 months, ranging from <1 to 18 months, and differed among Bartonella genogroups. Phylogenetic analysis of the Bartonella citrate synthase gene (gltA) revealed 9 genogroups and 13 subgroups. Seven genogroups clustered with known or previously reported Bartonella species and strains while two were distant enough to represent new Bartonella species. We report, for the first time, the detection of Bartonella alsatica in North America in Sylvilagus audubonii and expand the known host range of Bartonella washoensis to include Otospermophilus variegatus. Interpretation and Conclusion: This work broadens our knowledge of the hosts and geographic range of bacterial pathogens that could guide future surveillance efforts and improves our understanding of the dynamics of Bartonella infection in wild small mammals.

Keywords: Bartonella; bacteria; small mammals; vector-borne pathogens; woodrat.

FIG. 1.

Bartonella prevalence in Neotoma micropus over time. Circles show estimated prevalence and gray outline shows 95% confidence intervals. The dashed line is the predicted fit for the binomial regression. Numbers above show the number of animals sampled.

FIG. 2.

Modeled probability of Bartonella infection in sampled N. micropus individuals including terms for sex, weight, an interaction between sex and weight, and a random effect for individual ear tag. Panels show the predicted probability of infection accounting for the effect of sex, averaging over weight (A); the effect of weight, with sexes combined (B); and the interaction of sex and weight (C).

FIG. 3.

Phylogenetic tree and host range of Bartonella genogroups. The neighbor-joining tree was produced from a 351 bp alignment of the gltA gene. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 pseudoreplicates) are shown as colored circles at each branch. Evolutionary distances were computed using the Tamura-Nei method and are in the units of the number of base substitutions per site. Relative abundance and counts of sequences from each genogroup are summarized to the right of the tree based on the host species and the host family from which the sequence was obtained.

FIG. 4.

Relative abundance of Bartonella genogroups A–I in host species and families.

FIG. 5.

Resampling history, infection course, and genotypic characterization of sequentially recovered Bartonella samples from 29 woodrats (NM = N. micropus and NA = N. albigula) captured three and more times during 21 months of the study. The genogroups of Bartonella recovered from bacteremic woodrats at a sample month are shown with different colors. The classification ‘‘pos, no gltA’’ means that the sample was positive by ITS and real-time ssrA tests, but no gltA sequence was obtained to determine the Bartonella genogroup present. ITS, intergenic transcribed spacer.

FIG. 6.

Infection duration for any Bartonella genogroup (left panel) and for separate genogroups (right panel). Individual points for infection durations of each genogroup in an individual are shown as open circles, the median is a thick black line on the box plot, and the mean is an open diamond. Numbers above the box plots show the counts of infections for each genogroup. Some individuals had multiple infections for the same genogroup over their timeline, for example, they were infected with A1, then H, then A1 again.

FIG. 7.

Phylogenetic of B. grahamii and genogroups D1 and D2. The neighbor-joining tree was produced from a 351bp alignment of 41 gltA sequences. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown as colored circles at each branch. Evolutionary distances were computed using the Tamura-Nei method and are in the units of the number of base substitutions per site. DO = Dipodomys ordii; NM = N. micropus; PL = Peromyscus leucopus.