Molecular Detection of Bartonella Species in Rodents Residing in Urban and Suburban Areas of Central Thailand

Bartonella spp. are Gram-negative zoonotic bacteria transmitted to humans via various blood-sucking arthropods. Rodents have been identified as reservoir hosts of several zoonotic pathogens, including Bartonella spp. In Thailand, studies of Bartonella spp. in rodents from urban areas are limited; thus, a study in this area is necessary. The objectives of this study were to detect Bartonella spp. in rodents in Thailand and to compare the species’ distribution across different areas. In total, 70 blood samples from rodents in urban and suburban areas were tested for Bartonella spp. using a conventional polymerase chain reaction that targeted the citrate synthase (gltA) gene. All Bartonella-positive sequences were analyzed using polymorphism in order to build a phylogenetic tree. Approximately 38% of the rodents studied contained Bartonella DNA. Both Rattus exulans (Pacific rat) and R. tanezumi (Asian house rat) contained Bartonella spp. Four species of Bartonella were detected in blood samples: B. tribocorum, B. phoceensis, B. grahamii, and B. rattimassiliensis. In addition, eight Pacific rats contained the B. kosoyi–B. tribocorum complex. Bartonella phoceensis and B. tribocorum–B. kosoyi complexes were found in a specific habitat (p < 0.05). Interestingly, only seven haplotypes were identified in the sequences analyzed, and only haplotype A was found in both rodent species. Finally, a monitoring program for zoonotic Bartonella infection, especially the B. kosoyi–B. tribocorum complex, B. phoceensis, B. grahamii, and B. rattimassiliensis should be established, especially in high-risk areas.


Introduction
Bartonella spp. are Gram-negative intraerythrocytic bacteria [1] that have been rearranged into the alpha-proteobacteria [2] and are transmitted by blood-sucking arthropods [3]. Several species of Bartonella have been confirmed as zoonotic pathogens, including B. henselae, B. clarridgeiae, B. elizabethae, B. grahamii, B. koehlerae, B. quintana, B. washoensis, and B. vinsonii [4]. New members of the genus Bartonella are being found continuously [5], and more than 35 species have had their whole-genome sequences identified. To date, around 6 of the 20 rodent-adapted Bartonella spp. are zoonotic pathogens [6] that are important in medical and veterinary medicine [7]. Several types of animals are confirmed as hosts transmitting Bartonella spp., including cats [4,8], dogs [1,4], cattle [9,10], and rodents [11,12]. Furthermore, Bartonella spp. have been detected in rodents in several continents, including Asia [13], Africa [14], Europe [7], Americas [12,15], and Australia [16]. Some rodents have been suspected of being a source of zoonotic pathogens transmitted to humans [17], such as Rickettsia spp., Leptospira spp., Coxiella burnetii, Orientia tsutsugamushi, and Bartonella spp. [18]. From a public health viewpoint, there could be an increase in human cases due to infection by rodent-borne Bartonella spp. from outdoor activities [6] and other indirect contact. Changing land use and the sharing of habitat among rodents, animals, and humans have been identified as causes of zoonotic infection [19,20]. Several Bartonella spp. that have been associated with specific species of rodent could be the causative agents of Bartonella-related diseases in humans, such as endocarditis, lymphadenopathy, and some neurological abnormalities [11,21].
In 2010, Thai febrile patients showed evidence of zoonotic species of Bartonella based on a molecular detection [26,30]. In Thailand, the study of zoonotic Bartonella spp. has been mainly conducted in companion animals [31][32][33] and their ectoparasites [34]; however, studies involving rodents are limited, and additional studies in urban areas are needed [35]. With regard to the Thai government's "One Health" approach to humans, animals, and vectors, further studies on Bartonella are also necessary [34]. The current study therefore aims to survey the prevalence of Bartonella infection in rodents and to compare the species of Bartonella in areas of different characteristics (urban and suburban environments).

Sample Size
The sample size was calculated using an equation for the infinite population proportion [36] and a prevalence (p), taken from a previous study in Thailand [19]. Bangkok and Nakhon Sawan provinces were defined as urban and suburban areas, respectively. Hence, the previous proportion (p = 4.38%) was re-calculated from data on settlement and rain-fed areas derived from the previous study. For the sample size calculation, the maximum tolerated error (d) and alpha (α) were set at 5%. In total, the calculated sample size was 65 rodents based on the following equation:

Sample Collection
The sampling was approved by the Kasetsart University Institutional Animal Care and Use Committee, Bangkok, Thailand, under the Ethical Review Board of the Office of the National Research Council of Thailand (NRCT; approval ID: ACKU63-VET-048). This study was a cross-sectional survey of rodents trapped from fields in two provinces (Bangkok and Nakhon Sawan), as shown in Figure 1. The rodents were trapped between 2011 and 2013. All trapped rodents were classified based on external morphological characteristics. Three milliliters of blood were collected from each sample using aseptic cardiac puncture and were kept in a sterile EDTA tube. Euthanasia was conducted using chloroform inhalation after blood collection. All blood samples were stored at −20 • C until DNA extraction. Additionally, the standard procedures applied by the laboratory in this study followed the verification of the Institutional Biosafety Committee (IBC), Faculty of Veterinary Medicine, Kasetsart University. followed the verification of the Institutional Biosafety Committee (IBC), Faculty of Veterinary Medicine, Kasetsart University.

DNA Extraction from Whole Blood
Two hundred microliters of whole blood was extracted for genomic DNA testing using a commercial extraction kit (FavoPrep TM Blood DNA Extraction Mini Kit, Favorgen Biotech Corporation, Pingtung, Taiwan). The extraction protocol was performed according to the manufacturer's instructions, and 100 µL of nuclease-free water was used as the elution solution. The extracted DNA was kept at −20 °C until a polymerase chain reaction (PCR) was performed.

DNA Extraction from Whole Blood
Two hundred microliters of whole blood was extracted for genomic DNA testing using a commercial extraction kit (FavoPrep TM Blood DNA Extraction Mini Kit, Favorgen Biotech Corporation, Pingtung, Taiwan). The extraction protocol was performed according to the manufacturer's instructions, and 100 µL of nuclease-free water was used as the elution solution. The extracted DNA was kept at −20 • C until a polymerase chain reaction (PCR) was performed.

Bartonella Detection Using Polymerase Chain Reaction
Bartonella spp. were detected using conventional PCR. BhCS.781p (5 -GGGGACCAGC TCATGGTGG-3 ) and BhCS.1137n (5 -AATGCAAAAAGAACAGTAAACA-3 ) were used to target the citrate synthase (gltA) gene of Bartonella spp. [37]. The targeted 379 bp fragment was suspected to be a Bartonella-positive blood sample, and the conditions of amplification were controlled using a thermocycler (Mastercycler ® Nexus Gradient, Eppendorf, Hamburg, Germany). In total, 25 µL of PCR mixture (0.2 mM of each dNTPs, 1X of Taq reaction buffer with MgSO 4 , 4 pmol/µL of each primer, 0.04 U/µL of Taq DNA polymerase, 0.8% of dimethyl sulfoxide, and 3 µL of DNA template) were run for PCR detection and commercial Taq DNA polymerase was used (Taq DNA Polymerase, Applied Biological Materials (ABM ® Inc., Richmond, BC, Canada). The running of PCR for BhCS.781p and BhCS.1137n was performed as follows: 5 min at 95 • C for initial denaturation, 35 repeated cycles of denaturation (95 • C, 20 s), annealing (51 • C, 30 s), and elongation (72 • C, 2 min). The last elongation was conducted at 72 • C for 5 min. Bartonella henselae strain Houston-1 DNA and nuclease-free water were used as the positive and negative controls, respectively. The amplified products were kept at 4 • C until gel electrophoresis.

Gel Electrophoresis and Purification
Twenty microliters of the amplified product was run in 1.5% agarose gel under a 0.5X tris-acetate EDTA (TAE) buffer for 45 min at 100 V, and UltraPower TM Nucleic Acid Stain (BioTeke Corporation, Wuxi, China) was used for DNA staining. DNA visualization was performed using an ultraviolet illuminator (Gel Doc InGenius, SYNGENE, Frederick, MD, USA) and a 100-bp DNA ladder was used as the DNA size marker (Enzynomics, Daejeon, South Korea). The 379 bp band was cut and purified using a DNA purification kit (Gel and PCR Purification System, BioFACTTM, Daejeon, South Korea). The running protocol followed the manufacturer's instructions. Forty microliters of eluted purified DNA fragments were sent to a commercial sequencing unit using Sanger's sequencing technology (Macrogen ® , Seoul, Korea).

Analysis of DNA Sequence
The obtained DNA sequences were trimmed using the Chromatogram Explorer Lite version 5.0.2 software (http://www.dnabaser.com, accessed on 20 November 2021) under the default low-quality end trimming conditions (75% of good bases, 18 bases of window length, and 25 quality value (QV) of good base). The trimmed DNA chromatograms were edited using the SnapGene ® Viewer version 5.3.2 software (https://www.snapgene.com/ snapgene-viewer, accessed on 20 November 2021) and analyzed using BLASTn (https: //blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 20 November 2021). A phylogenetic tree was constructed using the neighbor-joining method based on a proper substitution model with 1000 bootstrapping replications in the Molecular Evolutionary Genetics Analysis (MEGA) version X software (https://www.megasoftware.net, accessed on 20 November 2021). Bartonella positive sequences were analyzed for polymorphism based on the number of variable sites (VS), the proportion of G + C content (GC), the number of haplotypes (h), the average number of nucleotide differences (k), haplotype diversity (Hd), and nucleotide diversity (π) using the DNA Sequence Polymorphism (DnaSP) version 6.12.03 software (http://www.ub.edu/dnasp, accessed on 20 November 2021). Then, the sequences were analyzed based on a median-joining network using the Population Analysis with Reticulate Trees (PopART) version 1.7 software (http://popart.otago.ac.nz/index.shtml, accessed on 20 November 2021) with the default setting (epsilon = 0). All Bartonella-matched sequences were submitted to GenBank with accession numbers: OK381826-OK381850.

Statistical Analysis
The data were presented using descriptive statistics (mean and standard deviation). Prevalence was calculated, and the Wilson score interval method was used to estimate the 95% confidence interval of prevalence [38]. Associated factors were analyzed using Chi-square or Fisher's exact test. All statistical analyses were performed using the 95% confidence interval (CI), and p < 0.05 was considered the significant level. All statistical analyses were performed using the R programming language version 4.0.2 [39].

Phylogenetic Tree and Polymorphism Based on gltA Sequences
Sequencing of the gltA fragments identified five species of Bartonella; the phylogenetic tree of the Bartonella positive sequences is presented in Figure 2. The tree had two main species complexes, consisting of a B. phoceensis complex (n = 14) and a B. kosoyi-B. tribocorum complex (n = 9). However, two sequences in the B. phoceensis complex were separate from the others. In the same way, in the B. tribocorum-B. kosoyi complex, one sequence was separate from the others. The polymorphism information of the partial gltA sequences that matched the B. kosoyi-B. tribocorum complex and B. phoceensis is presented in Table 3. Additionally, the variable position (singleton and parsimony-informative sites) is presented in Table 4. For these two complexes, the median-joining network is illustrated in Figure 3. Seven different haplotypes among 25 gltA sequences (n = 13 for R. exulans and n = 14 for R. tanezumi) showed π = 0.06471 ± 0.00617, Hd = 0.687 ± 0.071, and k = 19.61. Only haplotype A (12 sequences) showed different species of rodent that had B. phoceensis. However, other haplotypes (B-G) revealed specific Bartonella spp. Comparing the with and without reference sequences, those that matched B. kosoyi and B. tribocorum were clearly different from the references (CP031843 and HG969192). In contrast, the diversity difference between the with and without reference groups in B. phoceensis was not different from the previous finding.

Discussion
In this study, the trapped rodents belonged to two species: R. exulans and R. tanezumi. In Thailand, various species of rodent were identified, including Rattus spp., Bandicota spp., Leopoldamys spp., Mus spp., and Niviventer spp. [40]. Rattus spp. was found in suburban areas in the current study. Of these, R. tanezumi and R. exulans were reported as major species of Rattus in Thailand [40]. Interestingly, R. tanezumi is a synanthropic rodent species mostly found in suburban environments, including residential and agricultural areas [41]. In urban areas, some species of Rattus were also reported such as R. norvegicus and R. rattus [35]. Due to the increase in human-rodent contact, attention has been focused on various emerging diseases caused by novel pathogens [42]. Several rodent-borne pathogens can cause various human diseases, such as hantavirus, Borrelia spp., Toxoplasma gondii, Yersinia pestis, Bartonella spp., Leptospira spp., and Coxiella burnetii [43].
Many rodents in the current study had Bartonella spp. in their blood samples. Importantly, rodents have been mentioned as a major source of Bartonella infection in humans [44]. Rodent-borne Bartonella spp. have been discovered globally, and the rodentadapted Bartonella spp. have high diversity [6]. Partial sequences of the gltA gene revealed five species of Bartonella: B. tribocorum, B kosoyi, B. phoceensis, B. grahamii, and B. rattimassiliensis. Of these, three species have been reported as human pathogens [30,45]. The overall prevalence of Bartonella spp. in the current study differed from other studies in Malaysia [46] and China [13]; however, the prevalence of a study in Chile [12] was similar to the prevalence of the current study. Most studies in Thailand have reported Bartonella spp. being frequently isolated from R. rattus [20,47,48]. The prevalence of Bartonella spp. in R. tanezumi in the current study was similar to that found in a study in Singapore [49]; however, it contrasts with many other studies [17,50,51]. Bartonella rattimassiliensis and B. phoceensis were positive in R. tanezumi blood samples, which is a result similar to those of studies in Malaysia [46], Indonesia [52], Vietnam [50], and Singapore [49]. In addition, two zoonotic Bartonella spp. (B. tribocorum and B. grahamii) were detected in R. exulans blood samples, which was similar to other studies in Thailand [17,48,53]. Nevertheless, in the current study, the prevalence of Bartonella spp. in R. exulans differed from other studies [17,[47][48][49][50][53][54][55]. Particularly in urban habitats, Bartonella infection risk in humans increases from contact among humans, rodents, and ectoparasites [56]. In Thailand, rodent lice and fleas have been reported to carry Bartonella spp. and to circulate these pathogens in the rat population [47].

Discussion
In this study, the trapped rodents belonged to two species: R. exulans and R. tanezumi. In Thailand, various species of rodent were identified, including Rattus spp., Bandicota spp., Leopoldamys spp., Mus spp., and Niviventer spp. [40]. Rattus spp. was found in suburban areas in the current study. Of these, R. tanezumi and R. exulans were reported as major species of Rattus in Thailand [40]. Interestingly, R. tanezumi is a synanthropic rodent species mostly found in suburban environments, including residential and agricultural areas [41]. In urban areas, some species of Rattus were also reported such as R. norvegicus and R. rattus [35]. Due to the increase in human-rodent contact, attention has been focused on various emerging diseases caused by novel pathogens [42]. Several rodent-borne pathogens can cause various human diseases, such as hantavirus, Borrelia spp., Toxoplasma gondii, Yersinia pestis, Bartonella spp., Leptospira spp., and Coxiella burnetii [43].
Many rodents in the current study had Bartonella spp. in their blood samples. Importantly, rodents have been mentioned as a major source of Bartonella infection in humans [44]. Rodent-borne Bartonella spp. have been discovered globally, and the rodent-adapted Bartonella spp. have high diversity [6]. Partial sequences of the gltA gene revealed five species of Bartonella: B. tribocorum, B kosoyi, B. phoceensis, B. grahamii, and B. rattimassiliensis. Of these, three species have been reported as human pathogens [30,45]. The overall prevalence of Bartonella spp. in the current study differed from other studies in Malaysia [46] and China [13]; however, the prevalence of a study in Chile [12] was similar to the prevalence of the current study. Most studies in Thailand have reported Bartonella spp. being frequently isolated from R. rattus [20,47,48]. The prevalence of Bartonella spp. in R. tanezumi in the current study was similar to that found in a study in Singapore [49]; however, it contrasts with many other studies [17,50,51]. Bartonella rattimassiliensis and B. phoceensis were positive in R. tanezumi blood samples, which is a result similar to those of studies in Malaysia [46], Indonesia [52], Vietnam [50], and Singapore [49]. In addition, two zoonotic Bartonella spp. (B. tribocorum and B. grahamii) were detected in R. exulans blood samples, which was similar to other studies in Thailand [17,48,53]. Nevertheless, in the current study, the prevalence of Bartonella spp. in R. exulans differed from other studies [17,[47][48][49][50][53][54][55]. Particularly in urban habitats, Bartonella infection risk in humans increases from contact among humans, rodents, and ectoparasites [56]. In Thailand, rodent lice and fleas have been reported to carry Bartonella spp. and to circulate these pathogens in the rat population [47].
The previously reported prevalence of Bartonella spp. infection in rodents varied from 6% to 100% [18]. There are several factors related to the Bartonella prevalence rate in rodents, including habitat characteristics, body mass of rodents, age, rodent species, climate, rodent behavior, movement pattern, sampling method, and detection technique [11,18,35,46,57,58]. Additionally, varied levels of prevalence and Bartonella diversity were mentioned as common in rodents that had caused the heterogenous distribution of pathogens [59]. Even if there were no associated factor related to Bartonella infection in the studied rodents, the comparison between species of rodent and Bartonella revealed specificity. However, there was a higher infection rate in the urban area (Bangkok). Furthermore, the cool season had a higher rate of Bartonella infection than the warm season. As the preferred habitat of each rodent species differed, the detected Bartonella spp. might have been affected by characteristics of the habitat in the current study. Moreover, the size and structure of the rat population, referred to as "ecological factors", related to Bartonella prevalence and diversity, particularly in the urban environment [56,59,60].
Inferring detected species, 11 of the 25 rodents carried zoonotic Bartonella spp. One R. exulans was carrying B. grahamii, which has been defined as a human pathogen and Ctenophthalamus nobilis has been suspected of being a possible vector [24]. Additionally, B. tribocorum, a zoonotic species [27], was also detected in one R. exulans. Interestingly, rodents have adapted to promote Bartonella spp. transmission to humans; in addition, Polyplax spinulosa and Xenopsylla cheopis have been mentioned as vectors of this species [27]. In other results, one R. tanezumi provided evidence of B. rattimassiliensis, which has also been identified as a zoonotic species carried by P. spinulosa, Haemophysalis longicornis, and Hoplopleura pacifica [61]. Bartonella phoceensis was found in both R. tanezumi and R. exulans. Bartonella phoceensis has not been reported as causing human infection, and P. spinulosa has been suggested as a vector carrying this non-zoonotic species [61]. Interestingly, eight of the nine B. tribocorum-positive R. exulans also matched a sequence of B. kosoyi with the same similarity and query coverage percentage but with differences in nucleotide substitution positions. The finding of B. kosoyi in R. exulans was similar to that in a study in Myanmar [55]. To date, there is no explanation of the relationship regarding gltA between B. tribocorum and B. kosoyi. However, the genome of B. kosoyi is closely related to B. elizabethae [62], which has been defined as a zoonotic species of Bartonella [23]. In addition, no evidence has been reported of B. kosoyi infection in humans, even though it has been isolated in rodents elsewhere [55,62,63].
The citrate synthase gene is widely used for Bartonella detection [64,65]. The citrate synthase gene was targeted for Bartonella detection in the current study, and seven haplotypes of partial gltA sequences were revealed. Several genetic events including mutation, demography, and recombination were factors regulating haplotype diversity [11]. Trimmed gltA sequences (approximately the 327 base pair) have been acclaimed for taxonomic classification in the genus Bartonella [66] and for distinguishing among subspecies and species [2,37,65,66]. However, an additional RNA-polymerase beta subunit (rpoB) gene has been noted to increase identification efficacy, especially for the classification of new species [66]. Remarkably, gltA sequencing was suggested as a common method for Bartonella diversity study in wild animals [6], although homologous recombination was an important point of this gene [67,68]. Compared with sequences in the NCBI database, the gltA sequences of Bartonella spp. have been continuously updated, and this has facilitated more species to be distinguished [47,65]. Furthermore, the citrate synthase gene showed synonymous amino substitutions, and it has been emphasized that gltA was an important gene for critical functions [65]. Importantly, sequencing in the current study aimed to detect species of Bartonella using a reliable gltA marker. All sequences had values of similarity percentage > 96%, indicating a full match for the species, based on the recommendations from the La Scola study [66].
Unfortunately, the detection of Bartonella spp. in the ectoparasites of the captured rodents was not included in the current study. However, there should be further study of infectious ectoparasites and ectoparasites in habitat environments in order to elucidate the dynamics of Bartonella spp. circulation in rodent populations. A complete explanation of the gltA gene and haplotype requires the analysis of the whole sequence of the gltA gene of Bartonella. A higher-level method should be used to fill this gap in knowledge, and full gene cloning and sequencing techniques should be considered for future studies.

Conclusions
In this study, the overall prevalence was 38.57% in rodents inhabiting areas of central Thailand (43.33% in R. exulans and 35% in R. tanezumi). Importantly, three zoonotic species were detected in the rodents' blood samples (B. tribocorum, B. grahamii in R. exulans, and B. rattimassiliensis in R. tanezumi). Furthermore, B. phoceensis was identified as the major Bartonella spp. In this rodent population. Remarkably, Bartonella phoceensis and the complex of B. tribocorum-B. kosoyi were significant in the suburban (Nakhon Sawan province) and urban (Bangkok province) areas, respectively. Comparing polymorphism in Bartonellapositive and matched reference sequences found that the complex of B. tribocorum-B. kosoyi had more differences in nucleotide sequences than B. phoceensis. Seven haplotypes of the sequences analyzed were identified; however, only haplotype A showed infection in both R. exulans and R. tanezumi. The authors suggest monitoring zoonotic species of Bartonella infection in humans, particularly in workers in contact with rodents. Furthermore, updating the knowledge on Bartonella-related diseases should be supported in risk areas.