Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

High prevalence and diversity of Bartonella in small mammals from the biodiverse Western Ghats

Abstract

Bartonella species are recognized globally as emerging zoonotic pathogens. Small mammals such as rodents and shrews are implicated as major natural reservoirs for these microbial agents. Nevertheless, in several tropical countries, like India, the diversity of Bartonella in small mammals remain unexplored and limited information exists on the natural transmission cycles (reservoirs and vectors) of these bacteria. Using a multi-locus sequencing approach, we investigated the prevalence, haplotype diversity, and phylogenetic affinities of Bartonella in small mammals and their associated mites in a mixed-use landscape in the biodiverse Western Ghats in southern India. We sampled 141 individual small mammals belonging to eight species. Bartonella was detected in five of the eight species, including three previously unknown hosts. We observed high interspecies variability of Bartonella prevalence in the host community. However, the overall prevalence (52.5%) and haplotype diversity (0.9) was high for the individuals tested. Of the seven lineages of Bartonella identified in our samples, five lineages were phylogenetically related to putative zoonotic species–B. tribocorum, B. queenslandensis, and B. elizabethae. Haplotypes identified from mites were identical to those identified from their host species. This indicates that these Bartonella species may be zoonotic, but further work is necessary to confirm whether these are pathogenic and pose a threat to humans. Taken together, these results emphasize the presence of hitherto unexplored diversity of Bartonella in wild and synanthropic small mammals in mixed-use landscapes. The study also highlights the necessity to assess the risk of spillover to humans and other incidental hosts.

Author summary

Zoonotic bacterial infections cause both economic and health burdens to humans, especially in developing countries. Bartonella is a diverse group with several species that are infectious to humans. There is limited knowledge of Bartonella diversity and subsequent risk for spillover in several tropical countries, including India. This study, set in a biodiversity hotspot with high human density, reveals a high prevalence and diversity of Bartonella in the most common, and synanthropic small mammals. Several lineages identified are phylogenetically related to zoonotic species known to cause infections in humans. We suggest that there is unexplored diversity of Bartonella in small mammals that reside in human-modified landscapes in India, demonstrating an urgent need for further investigation. By identifying the natural reservoirs of pathogens, their ecology, and their relationship and proximity to humans, we will be able to proactively survey for these under-detected bacterial infections and minimize the burden.

Citation: Ansil BR, Mendenhall IH, Ramakrishnan U (2021) High prevalence and diversity of Bartonella in small mammals from the biodiverse Western Ghats. PLoS Negl Trop Dis 15(3): e0009178. https://doi.org/10.1371/journal.pntd.0009178

Editor: Amanda Bastos, University of Pretoria, SOUTH AFRICA

Received: October 5, 2020; Accepted: January 25, 2021; Published: March 11, 2021

Copyright: © 2021 Ansil et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting Information files. All genetic sequences generated in this study are deposited in the National Center for Biotechnology Information (NCBI) and available under accession numbers MT787669-MT787733, MT790780-MT790833, and MT785315-MT785361.

Funding: This work was supported by the NCBS/TIFR internal plan fund awarded to UR and we acknowledge the support of the Department of Atomic Energy (DAE), Government of India, under Project Identification No. RTI 4006. ABR was supported by the SPM research fellowship awarded by CSIR-HRDG, Government of India. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

The genus Bartonella comprises many cosmopolitan species capable of infecting a wide range of mammals, including humans [1]. Of the approximately 46 defined species, at least 18 are zoonotic [2,3]. Small mammals such as rats, squirrels, and shrews act as reservoirs for several of these bacterial species [1], while hematophagous arthropods, such as ticks, mites, and fleas transmit these pathogens between reservoirs and incidental hosts [4]. Infections in humans include cat scratch fever (Bartonella henselae), trench fever (B. quintana), and carrion’s disease (B. bacilliformis) [5–7]. Symptoms of these infections present as swollen lymph nodes (lymphadenopathy) with prolonged fever, presence of bacteria in red blood cells (intraerythrocytic bacteremia), inflammation of the optic nerve and retina (neuroretinitis), and often infections of the heart’s inner lining (endocarditis) [2].

Despite the evidence of disease burden in humans, bartonellosis remains under-detected in tropical countries such as India. Few studies investigate Bartonella-associated infections in humans in India [8–10]. For instance, in the Government General Hospital, Chennai, in South India (2005–2006), patients admitted with endocarditis were diagnosed with B. quintana [10]. Similarly, in tertiary care hospitals in North India, B. henselae was detected in patients admitted with fever, lymphadenopathy, and optic neuritis [8,9]. All these studies reported a high prevalence (8–23%) of Bartonella indicating that bartonellosis is an important existing infection. Other than these studies, there is no information available on the prevalence or the burden of this infection in India. Even more poorly documented are the diversity, natural maintenance, incidental hosts, vectors, and mode of transmission of these pathogens. We seek to bridge this gap by investigating the presence of Bartonella species in small mammal hosts in India.

As a group, small mammals primarily consist of non-volant mammals such as rodents and shrews that usually weigh less than 1 kg [11]. They host several medically important zoonotic pathogens, such as viruses (Hantaviruses, Lassa virus) bacteria (Lyme disease, leptospirosis), protozoans (toxoplasmosis), and helminths (rat lungworm) [12,13]. Their high birth and growth rates and synanthropic nature facilitate high rates of contact with humans and provide opportunities for spillover [14]. Their cosmopolitan distribution and increased connectivity, especially because of human facilitated dispersal, enables the influx and efflux of host and pathogen gene flow, posing a serious challenge for public health [15].

The potential risk of zoonoses from small mammals is very significant in a country like India, which is home to 101 species of rodents and 30 species of shrews [16]. Along with the high richness of small mammals, India is also experiencing high rates of deforestation and associated land-use change [17]. Natural habitats in the country are often interspersed with agricultural fields, plantations, and habitations with high human density [18,19]. This spatial structuring leads to increased human-livestock-wildlife interactions [20], which alongside high density of generalist and synanthropic small mammals poses a higher risk of zoonotic infections [12,21]. Therefore, identifying pathogens in their natural reservoirs and vectors is imperative for understanding their transmission dynamics, particularly in rapidly changing landscapes.

In this study, we investigated the presence of Bartonella in small mammal species and associated mites from a mixed-use landscape in the biodiverse Western Ghats in the southern Indian state of Karnataka. We examined the prevalence and haplotype diversity of Bartonella species in the small mammal host community. Studies from other tropical countries find high Bartonella prevalence in small mammal communities [22]. They also find that among different groups of small mammals in these communities, synanthropic species tend to have high prevalence and diversity [22,23]. In accordance with these patterns from other tropical small mammal communities, we expected to find (a) high overall prevalence and (b) high prevalence and diversity in synanthropes. We used PCR-based screening, multi-locus sequencing, and phylogenetic methods to investigate these predictions and understand the phylogenetic affinities of Bartonella in small mammals and their ectoparasites. Using this information, we aim to identify natural reservoir hosts and vectors of these emerging pathogens to ultimately better understand the risk of zoonotic spillover.

Methods

Ethical statement

This study was approved by the National Centre for Biological Sciences (NCBS) Institutional Animal Ethics Committee (UR-5/2014 and NCBS-IAEC-2016/10-[M]). Sample collection and processing in this study adhered to strict biosafety guidelines and approved by the Institutional Biosafety Committee (TFR: NCBS:23_IBSC/2017). Field sampling in Kadumane Tea Estate was conducted under the provision of a Memorandum of Understanding (MoU) signed between Kadumane Estate Company and NCBS.

Small mammal trapping and sample collection

This study was part of a larger project investigating small mammals and associated bacterial communities in mixed-use landscapes in the Western Ghats. Small mammal trapping was conducted during the dry months (January–May) of 2016, 2017, and 2018 in different land-use types present in the Kadumane Tea Estate, located in the state of Karnataka in South India. Land-use types in the study area included forest fragments, grasslands, abandoned plantations, tea plantations, and human habitations (Fig 1). Nine sites were selected across these land-use types for small mammal trapping. The number of sites in each land-use type was decided based on the available area. Each site was sampled for small mammals using a grid framework, comprising 100 medium-sized Sherman traps (H.B. Sherman Traps, Inc., Tallahassee FL, USA) arranged in 10 lines. Each grid was sampled for four consecutive days, for a total of 400 trap nights per site. In forests and abandoned plantations, 20 additional traps were placed (a total of 480 trap nights per site) on trees at different heights to capture arboreal small mammals. Traps were baited with peanut butter and checked twice a day (morning and evening) for captures. Species were identified with the help of taxonomic keys and field guides [24,25].

thumbnail
Download:
Fig 1. Map of the study area and sampling sites.

(A) World map with India marked inside the red box. (B) Map of India with the Karnataka state and the Kadumane landscape marked. (C) The enlarged land-use land cover map of Kadumane. Red circles indicate sampling locations. The shapefile for the world was downloaded from DIVA-GIS (http://www.diva-gis.org/gdata) and the shapefile for India from Community Created Maps of India (http://projects.datameet.org/maps/). The land-use map of Kadumane was created by Nature Conservation Foundation. All layers are in WGS84 geographic coordinate system and were processed and plotted using QGIS3.10 (https://qgis.org/en/site/), an open-access software.

https://doi.org/10.1371/journal.pntd.0009178.g001

Individuals captured were humanely sacrificed by overexposure to isoflurane (Baxter International, Deerfield, IL, USA) and dissected for blood and spleen collection. Cardiac blood was collected on Whatman FTA cards (GE Life Sciences, Pittsburgh, PA, USA) and stored using silica gel as a desiccant. Spleen samples were collected in RNAlater (Sigma-Aldrich, St. Louis, MO, USA) and stored at 4°C for one week before being transferred to a -20°C freezer at the National Centre for Biological Sciences (NCBS) in Bangalore. From the small mammals captured, individuals from which either spleen or blood was collected were used. Hematophagous mites were collected by combing each individual over a sheet of white paper placed in a tray and stored in 100% ethanol. An average of 5 mites (ranging from 2–10) was collected from each individual and identified morphologically using a stereomicroscope. The shape of the genital ventral plate, anal plate, and mouthpart morphology were used as identification features based on illustrations provided in previous studies and taxonomic keys [26–28].

PCR-based Bartonella screening

Approximately 25 mg of spleen was homogenized using Silicon-Carbide Sharp Particles in Mini-Beadbeater-96 bead beater (BioSpec Products, Bartlesville, OK, USA) and DNA was extracted using ZR-Duet DNA/RNA Mini-Prep Plus Kit—D7003 (Zymo Research Corporation, Irvine, CA, U.S.A.). For mite DNA extraction, 5–10 individuals were pooled by host species and the land-use type and processed as above. Considering the microscopic size and large number, pooling was an ideal method to yield enough DNA for Bartonella detection. For DNA extraction from dried blood spots, 3 mm diameter circles (n = 3–5 circles) were cut out from FTA cards and extracted using a QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). DNA extracts from the spleen, dried blood spots, and pooled mites were subjected to an initial screening PCR targeting an 825 bp region of the rpoB locus (beta subunit of bacterial RNA polymerase) of Bartonella [29]. The 25 μl PCR reaction mixture was comprised of 2 μl of template DNA, 12.5 μl of 1X HotStarTaq Master Mix (QIAGEN, Hilden, Germany), 1 μl each of 5 μM primers, and 8.5 μl of nuclease-free water. For screening dried blood spots, 4 μl of template DNA was used to increase PCR detection. Nuclease free water and DNA extraction controls were used as negative controls for PCR reactions. Previously identified positive Bartonella samples confirmed by sequencing were used as positive controls. PCR products were visualized on a 1.5% agarose gel stained with GelRed Nucleic Acid Gel Stain (Biotium Inc., Fremont, CA, USA). Amplified products were purified using AMPure XP magnetic beads (Beckman Coulter, IN, USA) and sequenced at the NCBS Genomics facility. Sequences generated were inspected using Geneious 11.1.4 (Biomatters, Ltd., Auckland, New Zealand) to assess quality and to trim primer binding sites and low-quality ends. These sequences were examined for similarity with Bartonella using the BLAST algorithm at NCBI [30]. In addition, a ~900 bp region of the ftsZ (cell division protein) and a 369 bp region of the 16S rRNA (16S ribosomal RNA) loci were amplified for each rpoB-positive sample [31,32]. PCR products for ftsZ and 16S were visualized, purified, and sequenced as described earlier. PCR primer sequences and annealing temperature for all three loci are provided in the S1 Table. GenBank accession numbers of sequences generated in this study are provided in the S2 Table.

Prevalence estimation

From the PCR screening results, the number of positives, negatives, and the total number of individuals were summarized and counted for each species using the ‘dplyr’ package in R [33]. The presence of Bartonella in either the spleen or the blood sample was considered a positive individual. From these data, species-specific and overall prevalence was calculated as the proportion of positive individuals in the total number of individuals tested. We estimated 95% confidence intervals around the prevalence for the four most common species. These confidence intervals are the binomial probabilities of positives (success) calculated from the sample (n) which include positives and negatives (success and failure). We used Wilson’s method [34] included in R Package ‘HMisc’ [35] for calculating these values. Confidence interval estimation was not possible for other species as they did not meet the minimum sample size requirement (10 individuals). The result of the same was plotted using ggplot2 [36].

Haplotype diversity and phylogenetic analysis

A consensus of forward and reverse sequences was generated for the partial rpoB gene to construct a haplotype network and calculate haplotype diversity. Only sequences with an initial quality score greater than 70% were used in the analysis. By setting this cutoff, we were able to remove low-quality mixed sequences from our sequence datasets. The sequence quality scores increased to 99–100% upon trimming primer binding sites, and low-quality ends. We also verified that no degenerate bases were present in the alignment created for the haplotype network analysis. Aligned rpoB sequences were imported to POPART [37] and an integer neighbor joining network was created. Names of the corresponding host species were used as traits and mapped to the alignment during the analysis. Haplotype diversity was calculated using the R package ‘pegas’ [38]. The corresponding host of each haplotype was identified from the haplotype network and used to construct a presence-absence matrix. A haplotype accumulation curve was created based on this presence-absence matrix and extrapolated to estimate the potential number of haplotypes that could be detected from different small mammal species present in the landscape. R package ‘vegan’ [39] and ‘iNEXT’ [40] were used for these analyses.

In order to construct the phylogeny, rpoB sequences of other known Bartonella species were downloaded from the NCBI database and a 773 bp alignment was created using MUSCLE in MEGA 7.0.26 [41]. Concatenated alignments of rpoB, ftsZ, and 16S sequences (1,861bp) were created following the same procedure. Alignments were codon-optimized (except 16S) and tested for the appropriate nucleotide substitution models using PartitionFinder2 [42]. As Bayesian methods only allow partitioned heterogenous substitution models, we used BEAST 1.10.0 [43] for phylogenetic analysis. Brucella melitensis was selected as an outgroup and each phylogeny was run for 108 Markov chain Monte Carlo (MCMC) cycles. Effective sample sizes and parameter convergence of MCMC sampling were assessed using Tracer v1.6 [44]. The resulting trees were used to generate a consensus tree (after 25% burnin), which was visualized and annotated using FigTree 1.4.3 [45]. Internal nodes of the tree are marked and colored based on the Posterior Probability (PP).

Results

Small mammal assemblage and Bartonella prevalence

Blood or spleen samples were collected from 141 individuals belonging to eight species captured across five different land-use types (Table 1). Rattus satarae, Rattus rattus, Mus booduga, and Mus musculus were the most common species in the captures. Among these, R. satarae (n = 67) was the most abundant, followed by M. booduga (n = 35), and M. musculus (n = 24). R. satarae was captured in forests and abandoned plantations, while R. rattus (n = 10) was only captured in human habitation. Although M. booduga was abundant in grasslands (n = 28), a few individuals were also trapped in forests (n = 2) and human habitations (n = 5). M. musculus was trapped largely in grasslands (n = 21) and rarely in forests (n = 3). Golunda ellioti (n = 1) was only trapped in grassland. Species such as Funambulus tristriatus (n = 1) and Platacanthomys lasiurus (n = 1) were rare and trapped only in forests. In addition to R. rattus and M. booduga, Suncus murinus (n = 2), was also trapped in human habitation. No small mammals were trapped in tea plantations. From these small mammals, ~250 mites were collected and identified to the family Macronyssidae, a family that includes tropical rat mites. All the mites collected from different small mammal species were morphologically identical. This could be due to the cryptic life stage morphology of these taxonomically complex groups. Therefore, to avoid erroneous taxonomic assignments, we report them as ’mites’ henceforth.

thumbnail
Download:
Table 1. Prevalence of Bartonella in small mammals captured in different land-use types in the study area.

https://doi.org/10.1371/journal.pntd.0009178.t001

Bartonella was present in 52.48% of the small mammals tested (Table 1). Among the most common four species, R. satarae had the highest prevalence of Bartonella at 85.07% (n = 67) (Fig 2). Synanthropic species M. musculus had a prevalence of 58.33% (n = 24), followed by M. booduga at 2.85% (n = 35), a species abundant in grasslands but often present in forests and human habitations. All the R. rattus, another well-recognized synanthrope, screened were PCR-negative for Bartonella. The only individual of F. tristriatus caught was positive while one of the two S. murinus individuals was positive. Screening results obtained from dried blood spots were consistent with the spleen samples (Table 2). Of the seven mite pools screened, one pool from M. musculus (n = 3) and another one from M. booduga (n = 1) was positive. Other mite pools from R. satarae, G. ellioti, and P. lasiurus were negative. A list of all individual small mammals used in this study and Bartonella screening results for spleen and dried blood spots are provided in S1 File.

thumbnail
Download:
Fig 2. Prevalence of Bartonella in the four most common small mammal species.

This figure summarizes our predictions about the high prevalence of Bartonella in the small mammal community. The dots represent calculated prevalence and the bar represents 95% confidence intervals. The sample size of the other species was too low to calculate confidence intervals. The silhouette images of animals were downloaded from PhyloPic (http://phylopic.org), an open-access database that stores reusable silhouette images of organisms.

https://doi.org/10.1371/journal.pntd.0009178.g002

thumbnail
Download:
Table 2. Tissue specific prevalence of Bartonella in different species of small mammals in the study area.

https://doi.org/10.1371/journal.pntd.0009178.t002

Haplotype diversity and host association of Bartonella

Of the 74 PCR-positives, 65 rpoB sequences (63 sequences from small mammals and two sequences from pooled mite samples) met quality thresholds and were used to construct the haplotype network and calculate haplotype diversity. Analysis of 657 bp of these rpoB sequences revealed 13 haplotypes, and haplotype diversity of 0.9 (Fig 3A). All of the haplotypes were host-specific, and within two host species, multiple haplotypes were observed. Five haplotypes each were observed in R. satarae (with two distinct lineages RS1 and RS2) and synanthropic species M. musculus (with another two distinct lineages MM1 and MM2). Both of these species had a high prevalence among the common species found in the landscape. Among the five haplotypes identified from M. musculus, four were closely related, and three of them (satellite haplotypes) originated from a central haplotype. The sequence from the pooled mites from M. musculus also belonged to this central haplotype. The fifth haplotype from M. musculus (MM2) was distinct and topologically close to RS2 lineage. Sequences from M. booduga, F. tristriatus, and S. murinus formed three independent haplotypes MB1, FT1, and SM1, respectively. The sequence obtained from mites collected from M. booduga also belonged to the MB1 haplotype. The results of the haplotype network were used to create a haplotype accumulation curve (Fig 3B). The shape of the curve indicates that there is a significant turnover of Bartonella haplotypes in the small mammals sampled. The extrapolated region of the haplotype accumulation curve revealed three more possible haplotypes, with the curve reaching the asymptote at around 160 positive individuals. Additional sampling of 100 positive individuals is estimated to capture all the haplotypes present in the small mammals. This indicates that we captured the majority (81%) of the haplotype diversity at the sampling sites.

thumbnail
Download:
Fig 3. Haplotype diversity.

(A) Haplotype network showing the relationship among Bartonella sequences from small mammal host species in Kadumane, South India. An integer neighbor joining network was constructed using 657 bp of rpoB using PopArt. The numbers between nodes indicate the number of nucleotide differences between haplotypes. (B) The haplotype accumulation curve (solid line) shows the cumulative haplotype richness (rpoB) sampled across 63 positive small mammals. The extrapolated richness (dotted lines) saturates around the estimated requirement to sample 160 positive individuals to detect all possible haplotypes in this community.

https://doi.org/10.1371/journal.pntd.0009178.g003

Phylogenetic position of small mammal-associated Bartonella

The phylogenetic analysis based on 773 bp of the rpoB locus revealed seven distinct lineages of Bartonella. The 65 sequences from this study (GenBank accession no. MT787669-MT787733) were polyphyletic and phylogenetically related to multiple Bartonella species. (Fig 4). Sequences from R. satarae formed two independent monophyletic lineages, with the first one (RS1) being a sister clade to B. queenslandensis (PP = 1). The second lineage (RS2) formed a sister lineage to B. phoceensis (PP = 1). Sequences from M. musculus also formed two lineages, the first one (MM1) as a sister clade (PP = 1) to sequences from M. booduga and its mites (MB1). Sequence from F. tristriatus (FT1) was distinct and formed an outgroup to this M. musculus—M. booduga clade (MM1 and MB1). These three lineages formed a monophyletic clade and were nested between B. tribocorum and B. elizabethae (PP = 1). The sequence from the pooled mites collected from M. musculus was also part of the above-mentioned clade (S1 Fig). The second lineage from M. musculus (MM2) formed an outgroup to Candidatus B. thailandensis—B. sylvatica lineage (PP = 1). Finally, the single sequence from S. murinus (SM1) was nested between RS1—B. queenslandensis lineage and B. tribocorum, but with low posterior probability support (PP <0.7).

thumbnail
Download:
Fig 4. Bayesian phylogenetic inference of Bartonella based on 773 bp of rpoB sequences.

Sequences from this study, zoonotic Bartonella and other Bartonella are colored purple, brown, and black respectively. Nodes are colored based on posterior probability and monophyletic clades are collapsed for easy visualization. The tree was rooted with one representative of Brucella melitensis. The original phylogenetic tree and accession numbers (MT787669-MT787733) are provided in the S1 Fig and S2 Table respectively. The silhouette images of animals were downloaded from PhyloPic (http://phylopic.org), an open-access database that stores reusable silhouette images of organisms.

https://doi.org/10.1371/journal.pntd.0009178.g004

The multi-locus phylogeny based on 1861 bp of rpoB, ftsZ, and 16SrRNA showed a concordant topology to the rpoB phylogeny with few minor deviations (S2 Fig). Divergence of the RS2 lineage into two was one such deviation, although this had low posterior probability support (PP <0.5). Another difference was that the SM1 lineage formed a sister lineage to B. tribocorum (PP = 0.7) but was nested between B. tribocorum and B. queenslandensis in rpoB phylogeny. Similarly, the FT1 lineage formed an outgroup to the B. tribocorum—SM1 lineage (PP = 0.8) whereas the same was an outgroup to the MM1-MB1 lineage in rpoB phylogeny. However, these minor deviations did not alter the phylogenetic topology or interpretations. Moreover, the posterior probability support for many lineages in the multi-locus phylogeny was considerably lower, hence the rpoB phylogeny was primarily used to draw phylogenetic conclusions in this study.

Discussion

We investigated the prevalence, diversity, and phylogenetic affinity of Bartonella in a small mammal community in the Western Ghats biodiversity hotspot. We report Bartonella from five species of small mammals and small mammal-associated mites. This is the first time that Bartonella was documented in three of these species—Rattus satarae, Mus booduga, and Funambulus tristriatus. The results from spleen and dried blood spots were nearly identical. We expected a similar prevalence in dried blood spots as Bartonella species are known to present in the blood of natural hosts [46]. Also, the larger volume of template DNA (4 μl as opposed to 2 μl used for spleen samples) from these samples may have increased the sensitivity of PCR detection. We speculate this could be the reason we have nearly identical results from spleen and dried blood spots. Every negative control used for each PCR run was negative, indicating there was no issue with contamination. Similar patterns are observed with other tropical pathogens using dried blood spots as a sensitive alternative to other gold standard sample types [47,48].

High prevalence of Bartonella in the small mammal community

The results of the study confirmed our predictions of overall prevalence in the host community and high prevalence in synanthropic species. The high prevalence in the host community (52.48) is consistent with other studies from tropical countries [22,49]. However, the overall prevalence estimate may be affected by the oversampling of common host species [50], such as the large sample size of R. satarae and M. musculus. Targeted sampling of rare species could address this bias [51]. For individual species, prevalence patterns varied. Among the four common species found in the landscape, the highest prevalence was observed in R. satarae followed by M. musculus. The former is known to prefer forested habitats including plantations and forest edges [52,53]. The latter is a widely distributed synanthrope [54] captured in forests and grasslands also known to show preference to a wide range of habitat types [52]. While the detection of Bartonella in R. satarae was novel, the high prevalence was not surprising since other members of Rattus genus are known to host Bartonella with high prevalence [49]. In contrast, R. rattus, another member from this genus and a well-known synanthrope [53] with previous reports of Bartonella infection [22], were negative. This result could be due to the low sample size of R. rattus in this study [55]. The high prevalence of Bartonella in R. satarae and M. musculus are of concern as they are most common in forested and abandoned plantations, often close to human habitations [56]. As human-altered habitats tend to allow generalist and synanthropic species to reach high abundance, they can enhance the transmission risk of Bartonella and other zoonotic pathogens to several incidental hosts, including humans [57,58]. Of the four synanthropic species in this study (R. rattus, M. musculus, M. booduga, and Suncus murinus), Bartonella was detected in three species (M. musculus, M. booduga, and S. murinus). Two of these species (M. booduga and S. murinus) were captured in human habitations. This indicates the presence of Bartonella in human habitation. However, low sample sizes for these species from human habitations precludes us from making any comparisons with other land-use types. Future studies with more focused sampling to capture synanthropic species could help discern any patterns in prevalence if present.

Under detected Bartonella diversity in the Western Ghats small mammals

Even though Bartonella was detected in only five small mammal species, the haplotype network analysis revealed high diversity. As expected, the bulk of this diversity was contributed by R. satarae and M. musculus, the two species detected with high prevalence in the community. Understanding the existing genetic diversity of Bartonella in this community is essential as the two host species (M. musculus and S. murinus) are widespread synanthropes and are also known to carry zoonotic species of Bartonella [4]. Furthermore, it is crucial from a public health perspective, as diversity may impact the evolution of high virulence, novel infections, and drug-resistant strains [59]. Bartonella haplotypes identified from mites were not novel and were identical to their respective hosts (M. musculus and M. booduga). Hematophagous ectoparasites are known to acquire these bacteria from their hosts and also transmit them to other incidental hosts [4]. However, the vector competence of these arthropods from this study is unknown. Our interpretation of diversity here remains limited due to the smaller sequence coverage used in this study relative to the Bartonella genome. Therefore, generating whole-genome data for these genetic variants will provide comprehensive insights into existing diversity and its implications for public health [60].

Due to lower rank abundance in the host community, our sample sizes for species such as F. tristriatus and S. murinus (and mite samples) are low. This made it difficult for us to compare the genetic diversity across species and land-use types. Interestingly, the haplotype accumulation curves suggest that a larger sample size of positive individuals might capture a few more novel Bartonella haplotypes from small mammals. This further substantiates our hypothesis that uncaptured Bartonella diversity exists in small mammals in the mixed-use landscapes of the Western Ghats. As a biodiversity hotspot with high human density and mammalian richness, pathogen diversity and potential zoonoses are predicted to be high for this landscape [61,62]. Large spatial scale sampling across the Western Ghats is necessary to understand the true Bartonella diversity and further validate these predictions. Such surveillance efforts targeting host and vector communities will help lay the foundation for understanding geographical risk, host associations, and natural transmission cycles of these pathogens. Additionally, investigating these pathogens in humans and comparison of data with other hosts and vectors will hold a clue to many febrile illnesses prevailing in forested areas in this biodiversity hotspot.

Potential zoonotic lineages of Bartonella in the small mammal community

Both the haplotype network and the phylogenetic results support the existence of multiple Bartonella lineages in small mammals. The most intriguing is the existence of divergent lineages in a single host species. Earlier studies have observed multiple lineages with independent genetic affinities in single host species, particularly in the genus Rattus [22,49]. Bai et al. [63] demonstrated this genetic heterogeneity of Bartonella sequences from Rattus with 23 unique variants falling into six different lineages. In our study, one of the Bartonella lineages from R. satarae (RS1) showed phylogenetic similarity to B. queenslandensis and the other (RS2) to B. phoceensis. Frank et al. [3] reported genetic similarity of Bartonella sequences reported from a febrile patient in Thailand to B. queenslandensis, a species previously isolated from Australian and Asian rodents [22]. Sequences from synanthropic species S. murinus showed phylogenetic relatedness to B. queenslandensis and B. tribocorum. Similarly, Bartonella lineages derived from M. musculus, M. booduga, and F. tristriatus were phylogenetically related to B. tribocorum and B. elizabethae. Both B. tribocorum and B. elizabethae were previously detected in small mammals and febrile illness patients in Asia as well as Europe [64,65]. Hence, the phylogenetic relatedness of five Bartonella lineages from this study (RS1, MM1, MB1, FT1, and SM1) to these zoonotic species indicates that these lineages could also potentially be zoonotic. Since the Bartonella sequences from pooled mites were also part of these zoonotic lineages, these arthropods may be involved in the transmission. Whole-genome analysis and pathogenicity prediction using computational tools might substantiate these phylogenetic signals from these zoonotic lineages [66]. Furthermore, additional sampling of vectors and identification of their species will aid in understanding the transmission interfaces of Bartonella across species and land-use types.

In summary, we attempt to understand the prevalence, haplotype diversity, and genetic affinity of Bartonella in India. Several lineages of Bartonella identified from small mammals including synanthropic species are phylogenetically related to known zoonotic Bartonella species. The detection of mites carrying these haplotypes substantiates the possibility of mite-mediated transmission to other incidental hosts including humans, heightening the risk of bartonellosis in the landscape. The presence of diverse Bartonella lineages in common rodent species, such as R. satarae and M. musculus highlights suspected reservoirs that future research must monitor. Zoonotic lineages identified from widespread synanthropic species need to be further investigated to understand the urban and peri-urban cycles of these pathogens. Despite the small spatial coverage, our results are significant in the context of emerging zoonotic infections and public health. Our results on these under-detected emerging pathogens reiterate the importance of strengthened surveillance in such areas, especially given the importance of global trade, travel, and pathogen movements across continents.

Supporting information

S1 Table. Primers used for multi-locus sequencing in this study.

https://doi.org/10.1371/journal.pntd.0009178.s001

(DOCX)

S2 Table. GenBank accession numbers of sequences reported from this study.

https://doi.org/10.1371/journal.pntd.0009178.s002

(DOCX)

S1 Fig. Bayesian phylogenetic inference of Bartonella based on 773 bp of rpoB sequences.

Sequences from this study, zoonotic Bartonella, and other Bartonella are colored purple, brown, and black respectively. The tree was inferred by gene partitioned analysis using BEAST 1.10.0 with 108 MCMC iterations. The posterior probability for the nodes is colored based on the values and monophyletic clades are collapsed for easy visualization. The tree was rooted with one representative of Brucella melitensis. The accession numbers for the sequences are provided in the S2 Table. The silhouette images of animals were downloaded from PhyloPic (http://phylopic.org), an open-access database that stores reusable silhouette images of organisms.

https://doi.org/10.1371/journal.pntd.0009178.s003

(PDF)

S2 Fig. Bayesian phylogenetic inference of Bartonella based on 1861 bp of multi-locus (rpoB, ftsZ, and 16SrRNA) sequences.

Sequences from this study, zoonotic Bartonella and other Bartonella are colored purple, brown, and black respectively. The tree was inferred by gene partitioned analysis using BEAST 1.10.0 with 108 MCMC iterations. The posterior probability for the nodes is colored based on the values, and monophyletic clades are collapsed for easy visualization. The tree was rooted with one representative of Brucella melitensis. The accession numbers for the sequences are provided in S2 Table. The silhouette images of animals were downloaded from PhyloPic (http://phylopic.org), an open-access database that stores reusable silhouette images of organisms.

https://doi.org/10.1371/journal.pntd.0009178.s004

(PDF)

S1 File. Bartonella screening results.

This file lists all individuals used in this study, their screening results, and the land-use type where the individuals were captured.

https://doi.org/10.1371/journal.pntd.0009178.s005

(XLSX)

Acknowledgments

We thank Ashwin Viswanathan, Aparna Lajmi, and Chinta Sidharthan for their assistance in the analysis; Meghana Natesh, Bhanu Sridharan, and Megan Aylward for valuable editorial inputs; S. Vijayakumar, K. Kamaraj, and C. Karnan for their help in sample collection. ABR thanks V.V. Robin and Varadharajan Sundaramurthy for academic discussions and comments on the results. ABR and UR are grateful to Kadumane Estate and Mr. K.M. Cariappa, Senior Manager, Kadumane Estates Company, for granting permission and field accommodation; Nature Conservation Foundation, India for providing the land-use map of the study area; NCBS NGS facility for sequencing.

References

  1. 1. Jiyipong T, Jittapalapong S, Morand S, Rolain J-M. Bartonella species in small mammals and their potential vectors in Asia. Asian Pac J Trop Biomed. 2014;4: 757–767.
  2. 2. Okaro U, Addisu A, Casanas B, Anderson B. Bartonella Species, an Emerging Cause of Blood-Culture-Negative Endocarditis. Clin Microbiol Rev. 2017;30: 709–746. pmid:28490579
  3. 3. Frank HK, Boyd SD, Hadly EA. Global fingerprint of humans on the distribution of Bartonella bacteria in mammals. PLoS Negl Trop Dis. 2018;12: e0006865. pmid:30439961
  4. 4. Tsai Y-L, Chang C-C, Chuang S-T, Chomel BB. Bartonella species and their ectoparasites: Selective host adaptation or strain selection between the vector and the mammalian host? Comp Immunol Microbiol Infect Dis. 2011;34: 299–314. pmid:21616536
  5. 5. Lamps LW, Scott MA. Cat-Scratch Disease. Pathol Patterns Rev. 2004;121: S71–S80. pmid:15298152
  6. 6. Atenstaedt RL. Trench fever: the British medical response in the Great War. J R Soc Med. 2006;99: 564–568. pmid:17082300
  7. 7. Maguina C, Garcia PJ, Gotuzzo E, Cordero L, Spach DH. Bartonellosis (Carrión’s Disease) in the Modern Era. Clin Infect Dis. 2001;33: 772–779. pmid:11512081
  8. 8. Chaudhry R, Mukherjee A, Menon V. Isolation & characterization of Bartonella sp. from optic neuritis patients. Indian J Med Res. 2012;136: 7. pmid:22885257
  9. 9. Chaudhry R, Kokkayil P, Ghosh A, Bahadur T, Kant K, Sagar T, et al. Bartonella henselae infection in diverse clinical conditions in a tertiary care hospital in north India. Indian J Med Res. 2018;147: 189. pmid:29806608
  10. 10. Balakrishnan N, Menon T, Fournier P-E, Raoult D. Bartonella quintana and Coxiella burnetii as Causes of Endocarditis, India. Emerg Infect Dis. 2008;14: 1168–1169. pmid:18598654
  11. 11. Barnett A, Dutton J. Small mammals (excluding bats). Expedition field techniques. 2nd ed. Expedition Advisory Centre, Royal Geographical Society, London; 1995.
    • 12. Mills JN. Biodiversity loss and emerging infectious disease: An example from the rodent-borne hemorrhagic fevers. Biodiversity. 2006;7: 9–17.
    • 13. Rael RC, Peterson AC, Ghersi-Chavez B, Riegel C, Lesen AE, Blum MJ. Rat Lungworm Infection in Rodents across Post-Katrina New Orleans, Louisiana, USA. Emerg Infect Dis. 2018;24: 2176–2183. pmid:30457534
    • 14. Firth C, Bhat M, Firth MA, Williams SH, Frye MJ, Simmonds P, et al. Detection of Zoonotic Pathogens and Characterization of Novel Viruses Carried by Commensal Rattus norvegicus in New York City. mBio. 2014;5: 1–16. pmid:25316698
    • 15. Gardner-Santana LC, Norris DE, Fornadel CM, Hinson ER, Klein SL, Glass GE. Commensal ecology, urban landscapes, and their influence on the genetic characteristics of city-dwelling Norway rats (Rattus norvegicus). Mol Ecol. 2009;18: 2766–2778. pmid:19457177
    • 16. Sharma G, Kamalakannan M, Venkataraman K. A checklist of Mammals of India with their distribution and conservation status. Zool Surv India. 2015; 111.
    • 17. Reddy CS, Dutta K, Jha CS. Analysing the gross and net deforestation rates in India. Curr Sci. 2013;105: 10.
    • 18. Kale MP, Chavan M, Pardeshi S, Joshi C, Verma PA, Roy PS, et al. Land-use and land-cover change in Western Ghats of India. Environ Monit Assess. 2016;188: 387. pmid:27256392
    • 19. Cincotta RP, Wisnewski J, Engelman R. Human population in the biodiversity hotspots. Nature. 2000;404: 990–992. pmid:10801126
    • 20. Hassell JM, Begon M, Ward MJ, Fèvre EM. Urbanization and Disease Emergence: Dynamics at the Wildlife–Livestock–Human Interface. Trends Ecol Evol. 2017;32: 55–67. pmid:28029378
    • 21. Morand S, Bordes F, Blasdell K, Pilosof S, Cornu J-F, Chaisiri K, et al. Assessing the distribution of disease-bearing rodents in human-modified tropical landscapes. McCallum H, editor. J Appl Ecol. 2015;52: 784–794.
    • 22. Gundi VAKB Kosoy MY, Myint KSA Shrestha SK, Shrestha MP Pavlin JA, et al. Prevalence and Genetic Diversity of Bartonella Species Detected in Different Tissues of Small Mammals in Nepal. Appl Environ Microbiol. 2010;76: 8247–8254. pmid:21037303
    • 23. Tsai Y-L, Chuang S-T, Chang C-C, Kass PH, Chomel BB. Bartonella Species in Small Mammals and Their Ectoparasites in Taiwan. Am J Trop Med Hyg. 2010;83: 917–923. pmid:20889892
    • 24. Ellerman JR. A Key to the Rodentia Inhabiting India, Ceylon, and Burma (Based on Collections in the British Museum) Part II. J Mammal. 1947;28: 357.
    • 25. Johnsingh A, Manjrekar N, editors. Mammals of South Asia. Universities Press; 2015.
      • 26. Knee W. Keys to the Families and Genera of Blood and Tissue Feeding Mites Associated with Albertan Birds. Can J Arthropod Identif. 2006 [cited 4 Dec 2020].
      • 27. Bhuyan P, Nath A. Record of Tropical Rat Mite, Ornithonyssus bacoti (Acari: Mesostigmata: Macronyssidae) from Domestic and Peridomestic Rodents (Rattus rattus) in Nilgiris, Tamil Nadu, India. J Arthropod-Borne Dis. 2016;10: 98–101. pmid:27047977
      • 28. Mullen GR, OConnor BM. Mites (Acari). Medical and Veterinary Entomology. Elsevier; 2019. pp. 533–602. https://doi.org/10.1016/B978-0-12-814043-7.00026–1
        • 29. Renesto P, Gouvernet J, Drancourt M, Roux V, Raoult D. Use of rpoB Gene Analysis for Detection and Identification of Bartonella Species. J Clin Microbiol. 2001;39: 430–437. pmid:11158086
        • 30. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic Local Alignment Search Tool. J Mol Biol. 1990;215: 403–410. pmid:2231712
        • 31. Zeaiter Z, Liang Z, Raoult D. Genetic Classification and Differentiation of Bartonella Species Based on Comparison of Partial ftsZ Gene Sequences. J Clin Microbiol. 2002;40: 3641–3647. pmid:12354859
        • 32. Paziewska A, Harris PD, Zwolińska L, Bajer A, Siński E. Recombination Within and Between Species of the Alpha Proteobacterium Bartonella Infecting Rodents. Microb Ecol. 2011;61: 134–145. pmid:20740281
        • 33. Wickham H, François R, Henry L, Muller K. dplyr: A Grammar of Data Manipulation. R package version 0.7.6. R Core Team; 2016.
        • 34. Agresti A, Coull BA. Approximate is Better than “Exact” for Interval Estimation of Binomial Proportions. Am Stat. 1998;52: 119–126.
        • 35. Harrell F Jr. ‘Hmisc’. R package version 4.2–0. CRAN; 2019.
          • 36. Wickham H. ggplot2: R package Version 3.3.2. CRAN; 2011. Available: http://doi.wiley.com/10.1002/wics.147
            • 37. Leigh JW, Bryant D. POPART: full-feature software for haplotype network construction. 2015. Available: http://doi.wiley.com/10.1111/2041-210X.12410
            • 38. Paradis E. pegas: an R package for population genetics with an integrated-modular approach. 2010. Available: https://academic.oup.com/bioinformatics/article-lookup/doi/10.1093/bioinformatics/btp696
            • 39. Oksanen J, Blanchet G, Friendly M, Kindt R, Legendre P, McGlinn D, et al. vegan. R package version 2.5–6. CRAN; 2019.
              • 40. Hsieh TC, Ma KH, Chao A. iNEXT: an R package for rarefaction and extrapolation of species diversity (Hill numbers) Version 2.0.20. CRAN; 2016. Available: https://onlinelibrary.wiley.com/doi/abs/10.1111/2041-210X.12613
                • 41. Kumar S, Stecher G, Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol. 2016;33: 1870–1874. pmid:27004904
                • 42. Lanfear R, Frandsen PB, Wright AM, Senfeld T, Calcott B. PartitionFinder 2: New Methods for Selecting Partitioned Models of Evolution for Molecular and Morphological Phylogenetic Analyses. Mol Biol Evol. 2016;34: 260. pmid:28013191
                • 43. Drummond AJ, Rambaut A. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol. 2007;7: 214. pmid:17996036
                • 44. Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. Posterior Summarization in Bayesian Phylogenetics Using Tracer 1.7. Susko E, editor. Syst Biol. 2018;67: 901–904. pmid:29718447
                • 45. Rambaut A. FigTree 1.4. 2 Software. Institute of Evolutionary Biology, Univ. Edinburgh; 2014.
                  • 46. Harms A, Dehio C. Intruders below the Radar: Molecular Pathogenesis of Bartonella spp. Clin Microbiol Rev. 2012;25: 42–78. pmid:22232371
                  • 47. Smit PW, Peeling RW, Garcia PJ, Torres LL, Pérez-Lu JE, Moore D, et al. Dried Blood Spots for qPCR Diagnosis of Acute Bartonella bacilliformis Infection. Am J Trop Med Hyg. 2013;89: 988–990. pmid:24043691
                  • 48. Smit PW, Elliott I, Peeling RW, Mabey D, Newton PN. An Overview of the Clinical Use of Filter Paper in the Diagnosis of Tropical Diseases. Am J Trop Med Hyg. 2014;90: 16.
                  • 49. Neves ES, Mendenhall IH, Borthwick SA, Su YCF, Smith GJD. Detection and genetic characterization of diverse Bartonella genotypes in the small mammals of Singapore. Zoonoses Public Health. 2018;65: e207–e215. pmid:29235263
                  • 50. Inumaru M, Murata K, Sato Y. Prevalence of avian haemosporidia among injured wild birds in Tokyo and environs, Japan. Int J Parasitol Parasites Wildl. 2017;6: 299–309. pmid:28971016
                  • 51. Zelmer DA. Estimating Prevalence: A Confidence Game. J Parasitol. 2013;99: 386–389. pmid:23016891
                  • 52. Molur S, Singh M. Non-volant small mammals of the Western Ghats of Coorg District, southern India. J Threat Taxa. 2009;1: 589–608.
                  • 53. Varudkar A, Ramakrishnan U. Commensalism facilitates gene flow in mountains: a comparison between two Rattus species. Heredity. 2015;115: 253–261. pmid:25944468
                  • 54. Zeder MA. Pathways to Animal Domestication. In: Gepts P, Famula TR, Bettinger RL, Brush SB, Damania AB, McGuire PE, et al., editors. Biodiversity in Agriculture: Domestication, Evolution, and Sustainability. Cambridge: Cambridge University Press; 2012. pp. 227–259. https://doi.org/10.1017/CBO9781139019514.013
                    • 55. Simon RC, Schtll WB. Tables of sample size requirements for detection of fish infected by pathogens: three confidence levels for different infection prevalence and various population sizes. J Fish Dis. 1984;7: 515–520.
                    • 56. Kumar A, Chellam R, Choudhury BC, Mudappa D, Vasudevan K, Ishwar NM, et al. Impact of rainforest fragmentation on small mammals and herpetofauna in the Western Ghats, South India. Wildlife Institute of India; 2002 p. 159. Available: https://archive.org/details/smallmammals
                      • 57. Suzán G, Armién A, Mills JN, Marcé E, Ceballos G, Ávila M, et al. Epidemiological considerations of rodent community composition in fragmented landscapes in Panama. J Mammal. 2008;89: 684–690.
                      • 58. Morand S, Blasdell K, Bordes F, Buchy P, Carcy B, Chaisiri K, et al. Changing landscapes of Southeast Asia and rodent-borne diseases: decreased diversity but increased transmission risks. Ecol Appl. 2019;29: e01886. pmid:30986339
                      • 59. Read AF. The Ecology of Genetically Diverse Infections. Science. 2001;292: 1099–1102. pmid:11352063
                      • 60. Van Goethem N, Descamps T, Devleesschauwer B, Roosens NHC, Boon NAM, Van Oyen H, et al. Status and potential of bacterial genomics for public health practice: a scoping review. Implement Sci. 2019;14: 79. pmid:31409417
                      • 61. Jones KE, Patel NG, Levy MA, Storeygard A, Balk D, Gittleman JL, et al. Global trends in emerging infectious diseases. Nature. 2008;451: 990–993. pmid:18288193
                      • 62. Johnson PTJ, Ostfeld RS, Keesing F. Frontiers in research on biodiversity and disease. Seabloom E, editor. Ecol Lett. 2015;18: 1119–1133. pmid:26261049
                      • 63. Bai Y, Kosoy MY, Lerdthusnee K, Peruski LF, Richardson JH. Prevalence and Genetic Heterogeneity of Bartonella Strains Cultured from Rodents from 17 Provinces in Thailand. Am J Trop Med Hyg. 2009;81: 811–816. pmid:19861616
                      • 64. Kosoy M, Bai Y, Sheff K, Morway C, Baggett H, Maloney SA, et al. Identification of Bartonella Infections in Febrile Human Patients from Thailand and Their Potential Animal Reservoirs. Am J Trop Med Hyg. 2010;82: 1140–1145. pmid:20519614
                      • 65. Vayssier-Taussat M, Moutailler S, Féménia F, Raymond P, Croce O, Scola BL, et al. Identification of Novel Zoonotic Activity of Bartonella spp., France. Emerg Infect Dis. 2016;22: 7. http://dx.doi.org/10.3201/eid2203.150269
                      • 66. Che D, Hasan M, Chen B. Identifying Pathogenicity Islands in Bacterial Pathogenomics Using Computational Approaches. Pathogens. 2014;3: 36–56. pmid:25437607