Molecular Detection and Isolation of Bartonella Species in Bats and Their Ectoparasites Along the China–Myanmar Border

Abstract

Bartonella are parasitic pathogens that infect many mammals, including humans, and cause significant diseases. This study investigates the presence, genetic diversity, and tissue tropism of Bartonella in bats and their ectoparasites along the China–Myanmar border. Bats and ectoparasites were collected from Yingjiang, Ruili, and Gengma Counties. Nested PCR (nPCR) and quantitative real-time PCR (qPCR) were used to detect and quantify Bartonella in bat tissues. Bartonella was isolated using brain–heart infusion broth and tryptone soy agar medium containing 5% sheep blood (TSA containing 5% sheep blood), and DNA sequences were analyzed with Clustal W and MEGA X. In total, 601 bats from 11 species (four families and seven genera) and 32 ectoparasites (two orders, three families, and four genera) were collected. The qPCR results revealed Bartonella detection rates of 22.96% (138/601) in bats and 62.50% (5/8) in ectoparasites. Using nPCR to detect the Bartonella gltA and rpoB genes in bats, ectoparasites, and strains isolated from bat blood samples, yielding 58 and 10 strains, respectively. When comparing bats, ectoparasites, and isolated strains to other Bartonella in GenBank, the gltA gene was 74.21%–100.00% at the nucleotide level of similarity and 75.70%–100.00% at the amino acid level. In comparison, the rpoB gene was 79.58%–100.00% at the nucleotide level of similarity and 89.71%–100.00% at the amino acid level. By phylogenetic analysis except for Bartonella sp. and uncultured Bartonella sp., we found a clade that was less than 96.0% at the nucleotide level of similarity in the gltA gene and less than 95.4% at the nucleotide level of similarity in the rpoB gene. Based on the threshold values for the delineation of new species of Bartonella, we believe that a new species of Bartonella prevalent in bats was discovered in this study, which we named “Candidatus Bartonella dianxisis”. Otherwise, the average copy number of Bartonella in bat tissues (blood, spleen, heart, brain, kidney, lung, liver, and rectum) ranged from 1.15 × 10⁴ to 6.87 × 10⁴ copies/μL, with the highest levels observed in blood and spleen. Our findings highlight the genetic diversity of Bartonella in bats and ectoparasites along the China–Myanmar border and underscore potential public health risks associated with these pathogens.

1. Introduction

Bartonella is a genus of gram-negative bacteria that primarily infects endothelial cells and erythrocytes. It belongs to the phylum Bacteria, class Alphaproteobacteria, order Hyphomicrobiales, and family Bartonellaceae. Bartonellosis is an emerging and re-emerging infectious disease with a broad spectrum of clinical manifestations. Symptoms can range from mild ones like fever, headache, and malaise to more severe effects such as hallucinations. The bacteria are primarily transmitted among mammals by arthropods and are responsible for several human illnesses, including cat-scratch disease (CSD), trench fever, endocarditis, and lymphadenitis [1].

There are currently 39 known species and three subspecies of Bartonella worldwide [2] (https://www.bacterio.net/genus/bartonella, assessed on October 24, 2024), with numerous unnamed and unidentified species still to be described. Seventeen Bartonella species have been linked to human illness [1], including B. henselae [3], B. tribocorum [4], B. quintana [5], B. grahamii [6], B. rochalimae [7], and B. elizabethae [8]. The primary natural hosts of Bartonella species are mammals from several taxonomic orders, including Carnivora (e.g., cats, dogs, and foxes) [9], Perissodactyla (e.g., horses), Artiodactyla (e.g., cows, sheep) [10, 11], Primates (e.g., rhesus monkeys) [12], Rodentia [13], and Chiroptera [14]. The order Chiroptera, which includes bats, is the second largest group in the class Mammalia, after Rodentia. With 1462 species classified into 21 families and 264 genera, bats represent over 25% of all mammal species (https://www.mammalwatching.com/resources/global-mammal-checklist/, assessed on October 24, 2024). As the only group of mammals capable of flight, bats are uniquely suited to carry and transmit a wide variety of pathogens, making them important vectors for Bartonella and other zoonotic diseases [15].

Bats have a wide geographical distribution worldwide and are often associated with emerging pathogens due to their unique immune mechanisms [16]. While much research has focused on bats as reservoirs of viral pathogens, less attention has been given to bacterial pathogens such as Bartonella [17]. More than 100 species of bats worldwide have been identified as hosts of Bartonella [18]. It is believed that bats were the original hosts of Bartonella and may have played a vital role in the early geographic spread of the genus [19]. A case report described a Bartonella infection of bat origin that caused endocarditis in a male patient from the USA [20, 21]. Studies of bat-derived Bartonella in Georgia revealed a high genetic to strains isolated from human samples in Poland and dog samples in Thailand [17]. These findings highlight that Bartonella is widely distributed in bats, likely due to factors such as host specialty and coevolution. The bacteria have been isolated from bats in various Countries, For example, Bai et al. [16] isolated and characterized Bartonella rousetti from Rousettus aegyptiacusa in Nigeria; Nabeshima et al. [22] isolated and studied the Bartonella in Miniopterus fuliginosus from Japan and found that Japan contains two endemic Bartonella of Miniopterus; In addition, a study of Bartonella in bats from Thailand isolated bacteria belonging to the Bartonella species from five bat species (including Chaerephon plicatus, Hipposideros armiger, Hipposideros fulvus, Hipposideros larvatus, and Taphozous melanopogon) [23]. Bat ectoparasites, such as bat flies, ticks, fleas, and mites, are believed to play a significant role in the transmission of Bartonella among bats [17]. Therefore, studying the interactions between Bartonella, bats, and their ectoparasites is essential for understanding the ecology and transmission of this pathogen.

The ssrA gene is a small RNA molecule found in bacteria that plays a key role in protein modification [24]. The ssrA genes in Bartonella are valuable tools for the rapid identification and classification of these bacteria [25]. Since the 1980s, various molecular techniques, such as DNA–DNA hybridization, multilocus sequence analysis (MLSA), and single-gene assays, have been developed to classify Bartonella species [26]. Traditional PCR methods, however, have limitations, including high detection limits and low sensitivity and specificity. To overcome these challenges, this study employed quantitative real-time PCR (qPCR) to detect the ssrA gene, a single-copy prokaryotic-specific marker, allowing for improved sensitivity and specificity in Bartonella detection. Scola et al. [27] carried out a comparative analysis of seven different genes in Bartonella and showed that genes encoding RNA polymerase subunits (rpoB) and citrate synthase (gltA) genes for intraspecific differentiation proved to be the most efficient. Therefore, we also amplified and characterized the species-specific gltA and rpoB genes of Bartonella to understand the genetic evolution of Bartonella on bats and their ectoparasites in our region by nPCR. We investigated the presence of Bartonella in bats from three counties along the China–Myanmar border (Yingjiang, Ruili, and Gengma), analyzed the genetic diversity of the Bartonella, and quantitatively assessed their presence in bat blood, brain, heart, lung, liver, spleen, kidney, and intestinal tissues. Additionally, we isolated and characterized Bartonella strains.

2. Materials and Methods

2.1. Medical Ethics Approvals

This research was approved by the Medical Ethics Committee of Dali University under number 2021-PZ-177. All animals were treated according to the Guidelines of Regulations for the Administration of Laboratory Animals (Decree No. 2 of the State Science and Technology Commission of the People's Republic of China, 1988) and the Guidelines for Treating Animals Kindly from Ministry of Science and Technology of the People's Republic of China. All efforts were made to minimize discomfort to the animals.

2.2. Sample Collection

Bat samples were collected from Yingjiang County, Ruili City, and Gengma County, located along the China and Myanmar border, between July 2022 and August 2023. (Figure 1) Local villagers captured and culled bats in stages due to concerns about the potential damage caused by bat colonies to crops and fruit trees. After a certain number of bats were collected at various locations, they were stored in a refrigerator at 4°C before being transported back to the local CDC laboratory, where sample preparation was completed within 5 h of collection. During the dissection of executed bats in the autopsy laboratory, the harvested bats were first morphologically identified and numbered, and each bat was examined for the presence of ectoparasites, which were immediately extracted with forceps and stored according to the number in 2 mL cryogenic vials (DNase free) (CORNING, Shanghai, China), and the bats were dissected shortly after the harvest, and the blood clots in their thoracic cavities were collected using a pipette Pasteur (Servicebio, Wuhan, China) under aseptic conditions and stored in 1.5 mL microtube (DNase free; Servicebio, Wuhan, China). The bats' heart, lung, liver, spleen, kidney, rectum, and brain tissues were then collected in correspondingly numbered 2 mL cryogenic vials (DNase free), and all samples were stored in a removable −80°C refrigerator at our research center before being transported back to the testing center for analysis.

Figure 1.

Map of sampling locations. The right panel shows the geographical map of China and Myanmar, while the left panel highlights the counties and cities situated along the China–Myanmar border region where sampling was conducted.

2.3. DNA Extraction of Bats

In the nucleic acid extraction laboratory, 1 g of each sample (blood, heart, liver, spleen, lung, kidney, rectum, and brain tissues) was clipped into small pieces using autoclaved surgical scissors. The samples were placed into GeneReady Animal PIII pulverizing tubes (Life Real, Guangzhou, China). Each tube was treated with 600 μL of phosphate buffer solution (PBS) and located into the GeneReady Ultimate Grinder (Life Real, Guangzhou, China) for mechanical grinding. Using a pipette, 300 μL of supernatant from the tissue suspension was transferred to the DNA extraction kit (TIANGEN, Beijing, China) and PBS was used as a blank control in the experimental manipulation. Nucleic acids were extracted using a fully automated nucleic acid extraction and purification system (BIOER, Hangzhou, China) according to the manufacturer's instructions. The extracted nucleic acid samples were aliquoted into two portions and stored at −80°C refrigerator for further analysis.

2.4. qPCR and Nested PCR (nPCR)

2.4.1. qPCR Detection of Bartonella sp. in Bats

To detect Bartonella nucleic acids in bat tissues, a qPCR assay targeting the ssrA gene was performed, based on the work of Diaz et al. [25]. The details of the reaction system, cycling conditions, and oligonucleotides for qPCR can be found in Supporting Information 1: Table S1. A sample was considered positive for Bartonella if the circle threshold (Ct) was ≤38, based on a standard curve from plasmid standards. To prevent contamination and false positives, ddH₂O was used as a blank control. Reaction preparation, sample pipetting, and amplification were conducted in separate areas.

2.4.2. nPCR of Bartonella sp. in Bats

Samples that tested positive by qPCR were examined further, with the target gltA gene amplified using nPCR in accordance with the methods of Norman et al. [28] and Birtles and Raoult [29]. Semi-nested primers developed by Renesto et al. [30] and André et al. [31] were also used to amplify the target rpoB gene in positive samples. The final amplicon lengths were 357 bp and 603 bp, respectively. Specific descriptions of the oligonucleotides, cycling conditions, and reaction systems for nPCR are shown in Supporting Information 1: Table S1. During the experiment, ddH₂O was used as a blank control, do not have positive control and the rest of the notes were similar to those for qPCR. The amplification products were separated on a 1.5% agarose gel (TransGen Biotech, Beijing, China), which was stained with the nucleic acid dye SuperRed (Biosharp, Beijing, China), and then visualized with a gel imaging system. Target bands were excised, purified with a purification kit (OMEGA Bio-Tek, Norcross, GA, USA), and sequenced by Sangon Biological Engineering Technology and Services Company Ltd. (Sangon Biotech) (Shanghai, China) Company.

2.4.3. Molecular Identification of Bat Species

Bat species were initially identified morphologically in the field, followed by molecular confirmation in the laboratory using mitochondrial cytochrome b (mt-Cytb) amplification. Primers used were Cytb-F (5′-ATGATATGAAAAACCATCGTTG-3′) and Cytb-R (5′-TTTCCNTTTCTGGTTTACAAGAC-3′) [32–34]. The PCR reaction consisted of 25 µL containing 12.5 µL of 2 × Rapid Taq Master Mix, 1 µL each of Cytb-F and Cytb-R (10 µM), 9.5 µL of RNase-free ddH₂O, and 1 µL of DNA template. The thermocycling program included predenaturation at 94°C for 3 min, followed by 40 cycles of denaturation at 94°C for 30 s, annealing at 51.9°C for 15 s, and extension at 72°C for 1 min and 20 s. A final extension at 72°C for 5 min completed the reaction. Amplicons (approximately equal to 1140 bp) [33], and the products were subjected to agarose gel electrophoresis and observed using a gel imaging system. Samples with correctly positioned electrophoretic bands were then selected and sent to Sangon Biotech for sequencing.

2.4.4. DNA Extraction, Molecular Identification, and Detection Bartonella sp. of Ectoparasites in Bat

According to Kuang et al. [35], we first removed the samples from cryogenic vials at −80°C, then removed single stored bat ectoparasites from cryogenic tubes, rinsed twice using Minimum Essential Medium (MEM, Thermo Fisher, USA), discarded the rinsing solution, and placed the worms into grinding tubes, then added 300 µL of fresh MEM, and then ground using a grinder. Based on morphological identification and location information, the 32 individual worm grindings obtained were combined into eight reaction pools. Accordingly, 28 ectoparasites of Nycteribiidae were mixed into six pools, three ectoparasites of Streblidae were mixed into one pool and one tick in a separate pool (Table 1). A total volume of 300 μL of the pools mixing homogenate was pipetted into the DNA Extraction Kit (TIANGEN, Beijing, China). Nucleic acids were extracted using a fully automated nucleic acid extraction and purification system (BIOER, Hangzhou, China). The extracted nucleic acid samples were divided into triplicates and stored separately in a −80°C refrigerator for further analysis.

Table 1.

Distribution and detection of Bartonella in bat ectoparasites from the China–Myanmar border region.

Extracorporeal parasite species			Location	Number of ectoparasites	Pools (N)	Positivity
Family	Genus	Species				n/N	%
Nycteribiidae	Penicillidia	Penicillidia sp.	Gengma county	5	1	0/8	0.00
		Penicillidia sp.		5	1	0/8	0.00
	Nycteribia	Nycteribia allotopa		4	1	1/8	12.50
	Phthiridium	Phthiridium sp.	Yingjiang county	5	1	1/8	12.50
		Phthiridium sp.		5	1	1/8	12.50
	Penicillidia	Penicillidia sp.		4	1	1/8	12.50
Streblidae	Brachytarsina	Brachytarsina amboinensis		3	1	1/8	12.50
Ixodidae	Ixodes	Ixodes collaris		1	1	0/8	0.00
Total				32	8	5/8	62.50

In molecular identification, mitochondrial cytochrome oxidase subunit I (COI) gene amplification was conducted using primers LCO1490 (5′-GGTCAACAAATCATAAAGATATTGG-3′) and HCO2198 (5′-TAAACTTCAGGGTGACCAAAAAATCA-3′) [36]. After performing nPCR on the grouped reaction pools, we obtained positive sample sequences for the gltA and rpoB genes for further analysis.

2.5. Sequence Comparison and Phylogenetic Analysis

Purified products were sequenced by Sangon Biotech using gltA and rpoB gene primers. The resulting sequences were assembled using the DNAstar Lasergene software package, manually edited, and trimmed to generate the final sequences. These sequences, derived from bats, arthropods, and pathogenic bacteria isolated in this study, were compared to gltA and rpoB gene sequences of Bartonella species available in the GenBank database using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi, assessed on October 24, 2024).

The gltA and rpoB gene reference sequences were downloaded from GenBank and a phylogenetic tree was constructed using the maximum likelihood method in MEGA X software [37]. First, the sequences obtained in this study were compared with the reference sequences in GenBank using Clustal W for the gltA gene. Then, the tree was constructed in MEGA X using the T92 + G model (Tamura 3-parameter + Gamma distribution) and Brucella abortus strains (CP081403.1 and CP081431.1) as the outgroup reference. The rpoB gene was then manipulated in the same way, and the most appropriate model (T92 + G) was chosen to build the tree and B. abortus strains (CP081443.1 and CP081377.1) as the outgroup reference. All sequences from this study were submitted to GenBank with the following accession numbers: CytB: PQ460893–PQ460958, PV667499–PV667524; COI: PQ461921–PQ461925, PV658813; gltA: PV695984–PV696039; rpoB: PV696040–PV696049.

2.6. Plasmid Construction and Standard Curves Plotting

A representative Bartonella strain was selected for cloning a fragment of its ssrA gene into the pEASY-T1 vector (TransGen Biotech, Beijing, China). The T-vector product was transformed into DH5-α E. coli cells, and positive single colonies were identified using a blue–white screening assay. Selected colonies were cultured in LB broth supplemented with 0.5% ampicillin (Soybean Casein Digest Agar; HuanKai Microbial, Guangzhou, China), and the target gene insertion was confirmed by sequence analysis. Plasmids containing the cloned fragments were extracted using the EasyPure Plasmid MiniPrep Kit (TransGen Biotech, Beijing, China), aliquoted, and stored at −20°C. The plasmid concentration was measured with a UV spectrophotometer (Life Real, Hangzhou, China) and converted into copy number using the formula are as follows:

$$\begin{array}{c}\text{Copies}/\mu \mathrm{L}=\frac{\text{Plasmind\hspace{0.17em}concentration\hspace{0.17em}}\left(\text{ng}/\mu \mathrm{L}\right)\times {10}^{-9}\times 6.02\times {10}^{23}}{660\times \text{DNA\hspace{0.17em}length}}.\end{array}$$   

A standard curve was established based on the plasmid copy number. A representative positive standard was serially diluted 10-fold and used as the template for a quantitative assay. Each concentration gradient was tested in triplicate, and the average Ct value was calculated. The standard curve was plotted with the logarithm of the plasmid copy number on the x-axis and the corresponding Ct values on the y-axis. The slope and correlation coefficient of the curve were determined to assess assay performance.

2.7. Tissue Tropism of Bartonella sp. in Bats

Quantification of Bartonella DNA in the blood and tissues of Bartonella-positive specimens was performed using the qPCR method described in this study to investigate Bartonella tissue tropism in bats.

2.8. Bartonella Cultivation and Identification

2.8.1. Isolation of Bartonella sp. by Primary Culture of Blood

Our isolation method improved upon the method described by Liu et al. [38]. First, dilutions were configured by adding 5% (50 mg/mL) amphotericin B solution to brain–heart Infusion broth medium (BHI; HuanKai Microbial, Guangzhou, China) after autoclaving (121°C, 15 min), the BHI/AmpB ratio was set to 9:1. After being removed from the −80°C refrigerator, the blood was gently mixed and a 2 mL cryogenic vial was used to hold the inoculum solution. According to the numbering system, the blood homogenate and diluent were mixed in a 1:4 ratio to create an inoculum solution for spare use. Remove tryptone soy agar containing 5% sheep blood (TSA medium containing 5% sheep blood; Hopebio, Shandong, China), number it, and aspirate 100 μL of the inoculum solution. Inoculate the plate medium using the three-zone (or four-zone) inoculation ring method, then, once the inoculum had been absorbed by the medium, the plate was inverted and placed in a constant temperature incubator at 37°C with 5% CO₂ for 30 days. Colony growth was observed daily, and the progression and size of colonies were recorded.

Bartonella colonies typically appear within 3–7 days but may require up to 1 month to develop fully [39]. Colonies are grayish-white or tan, small, round slightly transparent, and raised, with smooth or rough edges [40]. Primary cultures often contain stray bacteria or colonies with varying morphologies. Suspected Bartonella colonies were marked and isolated by streaking onto new TSA medium containing 5% sheep blood plates for subsequent incubation. This process was repeated for 1–6 generations to obtain pure colonies containing a single bacterial strain.

2.8.2. Identification of Colonies

Although Bartonella is a gram-stain-negative bacterium, it does not easily adhere to gram stain dye. Thus, this study was conducted using Giemsa stain, which provides clear coloring and strong staining power for a better staining effect. The original solution (Solarbio Science & Technology [Beijing] Co., Ltd.) contains azure and eosin, which color acidophilic particles pink and basophilic particles blue–violet. Due to its unique characteristics, the Barton body combines with the azure in the Giemsa stain to reveal a blue–purple, rod- or ball-rod-shaped bacterial form under the microscope. So colonies were homogenized onto slides to form a monolayer, fixed, and stained using conventional Giemsa staining. Stained slides were under a light microscope using oil immersion at a magnification of 1000x (Axio Imager.A2, ZEISS, Germany Jena).

Using an inoculation needle, pick a single colony from the above purified colony plate and add it to a sterile microtube (DNase free) containing 100 μL of PBS. Mix well and in a water bath at 100°C for 10 min. Then centrifuge in an ultrahigh-speed centrifuge pre-cooled to 4°C at 12000 rpm for 10 min. Aspirate the supernatant into a new small microtube and store at −20°C. It can then be used as a template, and the gltA and rpoB genes can be amplified and identified.

2.9. Statistical Analysis

Statistical analyses were performed using SPSS 22.0 software. Chi-square and rank-sum tests were conducted to analyze the relationship between bat species, gender, body weight, and Bartonella infection rate. Multifactorial logistic regression analysis was used to identify risk factors associated with Bartonella detection rate (positive vs., negative). The model included all bat species with sample sizes of at least 10. Variables analyzed included bat species, gender, and body weight. Variables with p  < 0.1 were included in the model. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to assess the strength of the associations. Tissue tropism results were compared across groups using analysis of variance (ANOVA) to evaluate differences in Bartonella distribution among tissues. Statistical significance was set at α = 0.05, with p < 0.05 considered significant.

3. Results

3.1. Characteristics of Bartonella in Bat Samples

A total of 601 bats from four families, seven genera, and 11 species were collected from Yingjiang, Ruili, and Gengma Counties along the China–Myanmar border (Figure 1). The most common species were Miniopterus schreibersi (34.61%, 208/601) and H. armiger (19.80%, 119/601). The overall detection rate of Bartonella detected by qPCR targeting the ssrA gene was 22.96% (138/601, 95% CI: 19.60%−26.30%). Species-specific prevalence rates were as follows: M. schreibersi (12.02%, 25/208), H. armiger (49.58%, 59/119), Rousettus leschenaultia (6.98%, 6/86), Cynopterus sphinx (27.03%, 20/74), Rousettus amplexicaudatus (19.51%, 8/41), Miniopterus fuliginosus (7.14%, 2/28), H. larvatus (60.00%, 12/20), Eonycteris spelaea (15.79%, 3/19), Megaerops niphanae (66.67%, 2/3), and Rhinolophus blythi (100.00%, 1/1).

Insectivorous bats had a significantly higher prevalence of Bartonella (25.99%, 98/377) compared to fruit-eating bats (17.86%, 40/224; χ² = 5.260, p = 0.022). Regional prevalence rates were Yingjiang (52.98%, 80/151), Ruili (14.69%, 31/211), and Gengma (11.30%, 27/239). Prevalence in Yingjiang was significantly higher than in both Ruili (χ² = 60.688, p < 0.001) and Gengma (χ² = 80.759, p < 0.001), while the difference between Ruili and Gengma was not statistically significant (χ² = 1.150, p = 0.283; Table 2, Figure 2).

Table 2.

Distribution and detection of Bartonella in bats from the China–Myanmar border region (qPCR for ssrA gene and nPCR for gltA gene).

Bat species	Composition ratio (%)	Locations
		Yingjiang county		Ruli county		Gengma county		Total
		qPCR (%)	nPCR (%)	qPCR (%)	nPCR (%)	qPCR (%)	nPCR (%)	qPCR (%)	nPCR (%)
Miniopterus schreibersi	34.61 (208/601)	—	—	—	—	12.02 (25/208)	3.37 (7/208)	12.02 (25/208)	3.37 (7/208)
Hipposideros armiger	19.80 (119/601)	50.00 (59/118)	11.86 (14/118)	—	—	0.00 (0/1)	0.00 (0/1)	49.58 (59/119)	11.76 (14/119)
Rousettus leschenaultii	14.31 (86/601)	—	—	6.98 (6/86)	5.81 (5/86)	—	—	6.98 (6/86)	5.81 (5/86)
Cynopterus sphinx	12.31 (74/601)	66.67 (8/12)	16.67 (2/12)	24.19 (12/62)	8.06 (5/62)	—	—	27.03 (20/74)	9.46 (7/74)
Rousettus amplexicaudatus	6.82 (41/601)	—	—	19.51 (8/41)	17.07 (7/41)	—	—	19.51 (8/41)	17.07 (7/41)
Miniopterus fuliginosus	4.66 (28/601)	—	—	—	—	7.14 (2/28)	3.57 (1/28)	7.14 (2/28)	3.57 (1/28)
Hipposideros larvatus	3.33 (20/601)	60.00 (12/20)	20.00 (4/20)	—	—	—	—	60.00 (12/20)	20.00 (4/20)
Eonycteris spelaea	3.16 (19/601)	—	—	15.79 (3/19)	10.53 (2/19)	—	—	15.79 (3/19)	10.53 (2/19)
Megaerops niphanae	0.50 (3/601)	—	—	66.67 (2/3)	33.33 (1/3)	—	—	66.67 (2/3)	33.33 (1/3)
Rhinolophus blythi	0.33 (2/601)	—	—	—	—	0.00 (0/2)	0.00(0/2)	0.00(0/2)	0.00 (0/2)
Macroglossus sobrinus	0.17 (1/601)	100.00 (1/1)	0.00 (0/1)	—	—	—	—	100.00 (1/1)	0.00 (0/1)
Total	100.00 (601/601)	52.98 (80/151)	13.25 (20/151)	14.69 (31/211)	9.48 (20/211)	11.30 (27/239)	3.35 (8/239)	22.96 (138/601)	7.99 (48/601)

Figure 2.

Prevalence of Bartonella among different bat species. Dots represent the prevalence rates of Bartonella for each bat species, while bar graphs indicate the 95% confidence intervals (95% CIs). Bat species with sample sizes fewer than 10 were excluded from the analysis.

Multifactorial logistic regression analysis identified bat species as the main factor influencing Bartonella prevalence. The odds of Bartonella infection in H. armiger were 5.285 times that in E. spelaea (p = 0.011) and 6.687 times higher in H. larvatus than in E. spelaea (p = 0.025; Supporting Information 2: Table S2 and Supporting Information 3: Figure S1).

3.2. Characteristics of Bartonella in Bat Ectoparasites

A total of 31 bat flies and one bat tick were collected in Yingjiang and Gengma Counties, belonging to two orders, three families, and four genera. These ectoparasites were grouped into eight pools for molecular analysis (Table 1). After nPCR for the gltA gene, a total of five reaction pools were detected with positive amplification of Bartonella (rpoB gene was not analyzed here because of unsuccessful amplification), with an overall prevalence of 62.50% (5/8, 95% CI: 24.50%−91.50%). Four out of six reaction pools of Nycteribiidae tested positive for Bartonella DNA amplification, with a prevalence of 50% (4/8). And then it also detected in one Streblidae reaction pool, but not in ticks.

3.3. Detection of Bartonella sp. in Bats and Ectoparasites

3.3.1. Sequence Comparison and Phylogenetic Analysis Based on the gltA Gene of Bartonella sp. in Bats, Ectoparasites, and Strains Isolated From Bats

We compared the 56 valid sequences obtained with the nucleotide sequences of Bartonella in GenBank and found that the identity ranged from 74.21% to 100.00% (Supporting Information 4: Table S3), revealing at least 18 genotypes (Figure 3). It is worth noting that from the analysis of 56 sequences obtained from all samples, we found that some genotypes had a high level of similarity: G1 is a strain isolated from a blood sample of H. armiger, and its sequence is almost identical to that of uncultured Bartonella sp. from bats in Hubei Province, China (99.37% identity, e.g., GenBank accession numbers MK876231). G2 is a strain isolated from H. larvatus, which is almost identical to Bartonella sp. detected in M. emarginatus in the USA (99.37% identity, e.g., GenBank accession numbers MK140365). There are two G3 sequences from M. schreibersi were identical to uncultured Bartonella sp. from M. schreibersi in the USA (100.00% identity, e.g., GenBank accession numbers MK140192). The three M. schreibersi sequences and one M. fuliginosus sequence in G4 are 99.69%–100.00% identical to Bartonella sp. from N. allopota collected from bats in China (e.g., GenBank accession numbers OR525745). There are three sequences from H. armiger in G5 were also completely identical to uncultured Bartonella sp. from H. armiger in Vietnam (100.00% identity, e.g., GenBank accession numbers KP100356). G6 includes four sequences from bats (three from H. armiger and one from H. larvatus) and one sequence from an ectoparasite of bats (Phthiridium sp. pool), which had 99.69% identity with Bartonella sp. from H. armiger in Thailand (e.g., GenBank accession numbers KY232212).

Figure 3.

Phylogenetic classification of Bartonella species based on bat gltA gene sequences. The phylogenetic tree was constructed using the maximum likelihood method under the T92 + G model (Tamura 3-parameter + gamma distribution). Bootstrap values were calculated with 1000 iterations, and values greater than 70% are displayed on the tree.

In addition, the G7 dataset comprises eight sequences: seven from bats (four from H. armiger, one from H. larvatus, and one from C. sphinx), and one from bat ectoparasites (Phthiridium sp. pool), which contains a Bartonella sp. strain isolated from E. helvum in Kenya that is 94.65%–95.28% identical to the reference strain (GenBank accession number HM363765). There are two sequences, G8 and G9, one from Penicillidia sp. pool and one from M. schreibersi, which clustered together with Bartonella sp. from M. fuliginosus and its ectoparasite N. allopota from Japan (100.00% identity, e.g., GenBank accession numbers LC522033 and LC483822). G10 includes one sequence from the N. allotopa pool collected from bats and one sequence from a strain isolated from H. armiger, both of which were identical to uncultured Bartonella sp. from ectoparasites (E. sundaica) of bats in Yunnan, China (100.00% identity, e.g., GenBank accession numbers OP433676).

Furthermore, G12 and G13 contain six sequences (three from C. sphinx, two from R. amplexicaudatus, and one from E. spelaea) that cluster with uncultured Bartonella sp. detected in ectoparasites of bats from Malaysia and China (98.12%–100.00% identity, e.g., GenBank accession numbers MZ388461 and OP433673). G14 consists of 2twosequences from H. armiger that are almost identical to uncultured Bartonella sp. from the ectoparasite P. conspicua of bats from the USA (99.06% identity, e.g., GenBank accession numbers MK140327). The two sequences of G15 and G16 (hosts are R. amplexicaudatus and C. sphinx) were grouped with uncultured Bartonella sp. of Thaumapsylla sp., an ectoparasite of Chinese bats, and M. niphanae in Vietnam, with identities of 100.00% and 96.86%, respectively (e.g., GenBank accession numbers OP433686 and KP100352). G17 has one sequence from M. schreibersi and one sequence from a bat ectoparasite (B. amboinensis pool), which also had high identity with Bartonella sp. on N. allopota collected from bats in Yunnan, China (identity of 94.65%–99.69%, e.g., GenBank accession numbers OR525748). The nine sequences of G18 are relatively new, and the hosts are mainly fruit bats (including M. niphanae, C. sphinx, R. amplexicaudatus, and R. leschenaultii), which have the highest identity with Bartonella symbiont of Eucampsipoda theodori from Madagascar, at around 90%.

The remaining new sequences are classified as G11, which has a low level of identity with Bartonella in GenBank and the highest identity with uncultured Bartonella sp. from S. caffer found in Africa (87.42%–89.62% identity, e.g., GenBank accession numbers MF774324). According to phylogenetic analysis of the gltA gene, it forms a new branch in the phylogenetic tree. Furthermore, Becker et al.'s [41] research on Bartonella in vampire bats shows that the relatedness between Bartonella in bats and ruminants is low, despite being classified as a branch. Therefore, based on the conventional 95% sequence identity threshold for bacterial species classification according to Goris et al. [42] and the critical value of less than 96.0% nucleotide identity in the gltA gene according to Scola et al. [27], we believe that a new species of Bartonella prevalent in bats has been discovered in this study, which we named “Candidatus Bartonella dianxisis.”

3.3.2. Sequence Comparison and Phylogenetic Analysis Based on the rpoB Gene of Bartonella sp. in Bats, Ectoparasites, and Strains Isolated From Bats

This study obtained a total of 10 rpoB gene sequences of Bartonella. Nucleotide sequence alignment of Bartonella was performed in GenBank, and the identity ranged from 79.58% to 100.00% (Supporting Information 5: Table S4). There were a total of five genotypes (Figure 4). Through sequence alignment and phylogenetic analysis, we found that four genotypes had a high level of identity with Bartonella in GenBank: G1 is a sequence isolated from H. armiger, which is almost identical to uncultured Bartonella sp. from P. monoceros collected from bats in Hubei Province, China (99.43% identity, e.g., GenBank accession numbers MZ208753). G2 has one sequence from H. armiger with 99.62% identity to uncultured Bartonella sp. detected in R. norvegicus in Guangzhou, China (e.g., GenBank accession numbers MZ208753). G4 has two sequences, one from H. larvatus and one from a strain isolated from this host, which clustered with uncultured Bartonella sp. from unidentified bats in Yunnan, China, with an identity of 99.81%–100.00% (e.g., GenBank accession numbers PQ810519). There is one sequence from G5, which came from H. armiger, showed a high level of identity with uncultured Bartonella sp. from E. sundaica collected from bats in Yunnan Province, China (99.43% identity, e.g., GenBank accession numbers OP433716).

Figure 4.

Phylogenetic classification of Bartonella species based on bat ectoparasites rpoB gene sequences. The phylogenetic tree was constructed using the maximum likelihood method under the T92 + G model (Tamura 3-parameter + gamma distribution). Bootstrap values were calculated with 1000 iterations, and values greater than 70% are displayed on the tree.

The remaining five sequences are grouped in G3, whose hosts mainly include H. armiger, R. amplexicaudatus, and R. leschenaultii. The identity with Bartonella in GenBank is relatively low, but the highest identity was found with Bartonella sp. from E. helvum in Ghana, Africa (83.78%–90.46% identity, e.g., GenBank accession numbers KM215201). Based on rpoB gene phylogenetic analysis, it formed a separate branch in the phylogenetic tree, which is consistent with the results of the previous gltA gene study. Therefore, based on the conventional 95% sequence identity threshold in bacterial species classification and the study by Scola et al. [27], which is below the 95.4% threshold for rpoB gene nucleotide identity, the findings of this study on “Candidatus Bartonella dianxisis” are confirmed.

3.4. Tissue Tropism of Bartonella sp. in Bats

A qPCR standard curve was established using a series of plasmid dilutions ranging from 1.00 × 10⁴ to 1.00 × 10⁸ copies/µL. The Ct values for these plasmid dilutions were plotted, demonstrating a stable linear relationship between plasmid concentrations and their corresponding average Ct values (Supporting Information 6: Figure S2 and Supporting Information 7: Table S5). The minimum detectable copy number for the Bartonella-positive standard plasmid was 1.00 × 10² copies/µL, highlighting the superior sensitivity of the qPCR method—100-fold higher than cPCR (Supporting Information 8: Figure S3 and Supporting Information 9: Figure S4).

A total of 18 bat samples, each with adequate blood and tissue samples, were analyzed for Bartonella copy numbers using qPCR. The mean copy numbers (ranked highest to lowest) were as follows: 6.87 × 10⁴ copies/μL (95% CI: 4.56 × 10⁴–9.70 × 10⁴) in blood, 5.26 × 10⁴ copies/μL (95% CI: 2.10 × 10⁴–6.89 × 10⁴) in the spleen, 4.08 × 10⁴ copies/μL (95% CI: 2.03 × 10⁴–3.66 × 10⁴) in heart, 3.49 × 10⁴ copies/μL (95% CI: 1.54 × 10⁴–4.53 × 10⁴) in the kidney, 3.41 × 10⁴ copies/μL (95% CI: 2.40 × 10⁴–4.10 × 10⁴) in the lung, 3.15 × 10⁴ copies/μL (95% CI: 1.66 × 10⁴–6.25 × 10⁴) in the brain, 1.27 × 10⁴ copies/μL (95% CI: 0.26 × 10⁴–3.97 × 10⁴) in liver, and 1.15 × 10⁴ copies/μL (95% CI: 1.11 × 10⁴–2.30 × 10⁴) in rectum (Supporting Information 10: Table S6). One-way ANOVA revealed that Bartonella copy numbers were significantly higher in blood, spleen, kidney, and heart tissues compared to other tissues. Lung and brain tissues had intermediate copy numbers, while the liver and rectum had the lowest (Figure 5).

Figure 5.

Tissue tropism of Bartonella in positive samples. The Bartonella load was measured across various tissues, including blood, spleen, heart, brain, kidney, lung, liver, and intestine, from 18 positive samples. Data are presented as mean ± standard error in copes/μL. Significance testing were conducted using one-way ANOVA (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001). The red dotted line indicates the detection limits of qPCR established in this study.

3.5. Isolation and Characterization of Bartonella sp.

A total of 42 Bartonella-positive bat blood samples (Ct < 38) from Yingjiang County were selected for culture. After inoculating these samples on TSA medium containing 5% sheep blood, colony growth was monitored daily. Colonies began to appear as early as 3 days and up to 30 days postinoculation. Initially, colonies were small, raised, off-white to transparent, with smooth or rough edges and shiny or waxy surfaces. Over time, amphotericin B resistance decreased, and stray bacterial colonies emerged. Subsequent purification produced larger colonies with varying morphologies-some cauliflower-like, others diffuse, nonadherent, or slightly adherent to the medium. When colonies were scraped, a rounded depression was observed, indicative of Bartonella growth into the medium (Supporting Information 11: Figure S5A). From these cultures, three strains were isolated, including two from H. armiger and one from H. larvatus. Using Giemsa staining, it observed under a light microscope using oil immersion at a magnification of 1000x, the isolated Bartonella appeared as tiny bluish-purple, rod-shaped, or bulbous structures. The organisms measured 1–3 μm in length and were observed as single entities, in pairs, or chains forming V- or Y-shapes (Supporting Information 11: Figure S5B).

4. Discussion

Bats (order Chiroptera) are natural hosts for a wide range of bacteria and viruses globally, such as Rhabdoviridae, Paramyxoviridae, Coronaviridae, Bartonella sp., Leptospira sp., Enterobacteriaceae and so on [43]. Their capacity to migrate across borders and even continents contributes significantly to the spread of zoonotic diseases. Previous studies have demonstrated that Bartonella species carried by bats can infect humans and cause disease [16, 20]. The diversity of arthropod species associated with bats further suggests that Bartonella may be transmitted through blood-sucking ectoparasites acting as vectors [16, 44, 45]. Understanding the interactions among Bartonella, bats, and their vectors is,therefore, crucial.

The overall detection rate of Bartonella in this study was 22.96%, which was lower than the reported rates of 64.00% in Bangladesh [46], 26.00% in Japan [47], 25.50% in Thailand [48], and 24.10% in Peru [49], but higher than that the 16.70% reported in Zambia [50], 10.70% in Costa Rica [51], and 3.39% in South Africa and Swaziland [52]. Within China, the prevalence observed was lower than in central (32.70%) [53], northern (25.20%) [54], and southwestern regions (56.40%) [55] but higher than in eastern China (11.10%) [56]. Such differences may be attributed to variations in geographic location, bat species diversity, and environmental conditions [45].

In this study, the detection rate varied among bat species. While M. schreibersi, R. leschenaultii, and R. blythiwere commonly sampled, their Bartonella detection rates were 12.02%, 6.98%, and 0%, respectively. Notably, higher detection rates were observed in M. niphanae (66.67%), H. larvatus (60.00%), and H. armiger (49.58%). These findings are consistent with a study from Vietnam, where M. niphanae had a detection rate of 50.00%, H. larvatus 60.00%, and H. armiger 33.33% [57]. However, discrepancies exist when comparing the detection rates of M. schreibersi in Taiwan (42.90%) [56], Italy (25.00%), and Hungary (33.30%) [58], likely due to regional differences. Furthermore, while a study from Yunnan Province reported a prevalence of 56.90% for R. leschenaultii [55], our findings of 6.98% for this species highlight the influence of sample sizes and geographic variability. This study represents the first investigation into Bartonella prevalence in bats along the China–Myanmar border, offering valuable insights into species-specific infection rates and emphasizing the need for expanded sampling and geographic coverage in future studies.

Bat species is shown as the primary factor influencing Bartonella prevalence. In this study, prevalence rates were highest in M. niphanae (66.67%), H. larvatus (60.00%), and H. armiger (49.58%). Logistic regression analyses revealed no significant correlations between Bartonella prevalence and factors such as bat weight or sex. These findings reinforce the species-specific nature of Bartonella infections. Although global studies have documented Bartonella in bats and their ectoparasites [19, 52, 59], limited research has focused on southwest China. In this study, Bartonella was detected in 62.50% of bat flies including Nycteribiidae and Streblidae. This rate exceeds the 19.80% detection rate reported in Brazilian bat ectoparasites [60] and the 54.05% prevalence in Southwest China [35] but is lower than the 66.70% reported in Malaysia [61]. Differences are probably due to the small sample sizes, regional factors, and species diversity. Kuang et al. [35] detected Bartonella in mites, fleas, ticks, and wingless bat flies, highlighting ectoparasites' role in spreading Bartonella. Similarly, our study identified Bartonella in bat flies. The high infection rates observed in insectivorous bats suggest a correlation between host-vector interactions and feeding behaviors [62]. According to phylogenetic analysis, it revealed the sequences of ectoparasites were grouped together with the sequences obtained from bats, supporting the concept of a “host-vector-pathogen” transmission cycle [63, 64].

The ssrA gene is one of the most commonly used target genes for rapid detection of Bartonella. In contrast, genes encoding RNA polymerase subunits (rpoB) and citrate synthase (gltA) provide strong evidence for Bartonella typing [27]. Phylogenetic analysis revealed that Bartonella in bats and their ectoparasites in this study had a diverse genotype, with specific Bartonella genotypes associated with bat species and ectoparasite species, though some Bartonella genotypes could infect multiple hosts. Previous studies in Yunnan Province, China, have shown that the Bartonella found in bats differ significantly from those found in cats, dogs, primates, and rodents. This suggests that the Bartonella found in bats form an independent group [55]. Research by McKee et al. [65] also supports the view that adaptation to bat hosts is the primary factor in the evolution of Bartonella species associated with bats. This suggests that the Bartonella genotypes obtained are commonly transmitted and evolve among bats and their ectoparasites in Asian and Southeast Asian countries, but does not support frequent transmission between bats and other hosts. This is consistent with the findings of Becker et al. [41] in North America.

Our research into two genes revealed that some branches of the evolutionary tree do not cluster with known Bartonella species. This finding is consistent with a previous study of bat ectoparasites [35], indicating the potential presence of new, unknown Bartonella species in bats and their ectoparasites. For instance, Lilley et al. [14] identified “Candidatus Bartonella hemsuletiensis” from M. daubentonii in Finland. Furthermore, Lin et al. [20] isolated 'Candidatus Bartonella mayotimonensis' from patient samples with endocarditis. Lilley et al. [66] conducted a survey of this Bartonella species in bats in North America, confirming that it may originate from bats. Since the gltA and rpoB genes both have good discriminatory power [29, 67], the limited first-generation sequencing results enabled the preliminary identification of Bartonella. Based on the conventional 95% sequence identity threshold for bacterial species classification by Goris et al. [42] and the critical values proposed by Scola et al. [27] of less than 96.0% nucleotide identity for the gltA gene and less than 95.4% nucleotide identity for the rpoB gene, we believe that a new species of Bartonella prevalent in bats has been discovered in this study. As the study area is located in western Yunnan Province, China, on the border with Myanmar, the new species has been named “Candidatus Bartonella dianxisis.” However, further in-depth studies on “Candidatus Bartonella dianxisis” are needed in terms of multigene locus sequencing, genetic evolutionary genomics, and pathogenic biology.

Our findings revealed a specific tissue tropism of Bartonella, with higher bacterial loads observed in the blood and spleen compared to other tissues (p < 0.05). This aligns with previous studies in small mammals and underscores the importance of these tissues in Bartonella biology.

Using an optimized culture method based on Liu et al. [38], we isolated three Bartonella strains. The isolation rate of 6.25%, which is lower than those reported in Zambian bats (16.70%) [50] and Nigerian bats (15.50%) [68]. This may be due to differences in geographical environments, species, and isolation methods between different studies, leading to varying results. Future studies on the strains isolated in this study should expand the number of housekeeping genes, such as 16S – 23S rRNA, ribC, ITS, and ftsZ, and obtain the whole genome through high-throughput sequencing to strengthen strain characterization.

5. Conclusion

In summary, we studied the prevalence, genetic diversity, and tissue tropism of Bartonella in bats and their ectoparasites in the China–Myanmar border region, as well as its isolation. The prevalence of Bartonella was mainly determined by bat species, and a high prevalence rate was also observed in their ectoparasites, suggesting that research on the host-vector-pathogen relationship should be strengthened. Bartonella in bats and their ectoparasites exhibited high genetic diversity. Our analysis revealed that a new Bartonella infected in bats may be occurring, which we named “Candidatus Bartonella dianxisis.” Additionally, we isolated three Bartonella strains and, through quantitative studies, observed that Bartonella exhibits broad tissue tropism. The pathogenicity of most bat-transmitted Bartonella strains to humans remains unclear to date. These findings highlight the need for further research to elucidate the zoonotic potential and interspecies transmission risks of bat-transmitted Bartonella.

Data Availability Statement

The nucleotide sequences from Bartonella species, hosts, and ectoparasites described in this study have been deposited into the NCBI GenBank database (http://blast.ncbi.nlm.nih.gov/Blast.cgi, assessed on October 24, 2024) under accession numbers: PQ460893–PQ460958, PV667499–PV667524, PQ461921–PQ461925, PV658813, PV695984–PV696039, and PV696040–PV696049. The data are provided in the manuscript and supporting documents to this article and can be obtained from the corresponding author upon reasonable request.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding

This study received financial support from the National Natural Science Foundation of China (Grants U2002218 and 81874274), the Yunnan Health Training Project of High-Level Talents (Grant L-2017027), the Doctoral Research Start-up Fee Project of Dali University (Grant KYBS2018004), and the project for the Cross-border Control and Quarantine Innovation Group on Zoonosis of Dali University (Grant ZKPY2019302).

Supporting Information

Additional supporting information can be found online in the Supporting Information section.

Supporting Information 1

Table S1. Oligonucleotides, cycling conditions, and reaction systems used in quantitative real-time PCR and nested PCR assays based on ssrA, gltA, and rpoB genes for Bartonella molecular characterization

Supporting Information 2

Table S2. Logistic regression results for the effect of bat specie.

Supporting Information 3

Figure S1. Logistic regression results for the effect of bat species.

Supporting Information 4

Table S3. Comparisons of nucleotide identity and amino acid identity of Bartonella gltA gene (357 bp) from bat in this study.

Supporting Information 5

Table S4. Comparisons of nucleotide identity and amino acid identity of Bartonella rpoB gene (603 bp) from bat in this study.

Supporting Information 6

Figure S2. Standard curve.

Supporting Information 7

Table S5. Based on three repetitive qPCR experiments, the Ct values, slope, and amplification efficiency for each experiment.

Supporting Information 8

Figure S3. Sensitivity test results.

Supporting Information 9

Figure S4: Results of conventional PCR against diluted standard plasmids.

Supporting Information 10

Table S6. Bartonella species copies in tissues of bats.

Supporting Information 11

Figure S5: Isolated culture of Bartonella.