Genomic characterization and drug resistance of Bordetella pseudohinzii first isolated from wild niviventer

Abstract

Background

Niviventer, a rodent species widely distributed in Asian forests, serves as a significant reservoir for pathogens. Bordetella pseudohinzii(B. pseudohinzii), a recently identified Bordetella species with unclear pathogenic potential, poses challenges in species identification and understanding of its pathogenicity, its biological traits and antibiotic resistance are not well understood.

Methods

B. pseudohinzii(strains 21F10, 22F12, and 27F25) were isolated from lung tissue of wild niviventer rodents in Guizhou, China. Initial identification was performed using Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) and 16 S rRNA gene sequencing. A phylogenetic tree based on the 16 S rRNA gene sequences was constructed using the neighbor-joining method implemented in MEGA 11. Whole-genome sequencing (WGS) was conducted on all three strains, and strain 21F10 underwent hybrid assembly of second- and third-generation sequencing to achieve high-quality sequences. Average Nucleotide Identity (ANI) and digital DNA-DNA hybridization (dDDH) were used as gold standards for strain identification, with thresholds set at 95% and 70%, respectively. Gene annotation was performed using nine databases, including KEGG, VFDB, CARD, PHI, COG, and NR. Antimicrobial susceptibility testing was carried out using the drug-sensitive plate method.

Results

Initial MALDI-TOF MS identification misclassified the strains as B. avium and B. hinzii. However, PCR amplification of the 16 S rRNA gene (primers 27 F and 1492R) revealed that the strains were identified as B. hinzii (identity > 99%). Further analysis of the 16 S rRNA gene sequences obtained from WGS showed identities greater than 99% with both B. pseudohinzii and B. hinzii. Phylogenetic analysis of the 16 S rRNA gene sequences showed that the strains were closely related to B. hinzii, followed by B. pseudohinzii. Ultimately, the ANI values of all three strains with B. pseudohinzii were greater than 95%, and dDDH values exceeded 70%, confirming the strains as B. pseudohinzii. Strain 21F10 exhibited notable findings in terms of virulence factors and antibiotic resistance genes. Antimicrobial susceptibility testing revealed significant resistance to several cephalosporins (cefoxitin, cefuroxime, cefotaxime, cefazolin, and ceftiofur). The 16 S rRNA and WGS of strain 21F10 have been deposited in GenBank and Genome Sequence Archive (GSA)under accession numbers PQ881859 and CRA022358, respectively.

Conclusion

The first isolation of B. pseudohinzii from the lung tissue of wild niviventer was reported, and the limitations of traditional methods for identifying B. pseudohinzii were demonstrated. We highlight the superiority of WGS for accurate species identification. The findings reveal a complex pathogenic profile and notable antibiotic resistance, providing important insights for the future prevention and treatment of B. pseudohinzii infections in humans, as well as underscoring the need for monitoring B. pseudohinzii in rodent populations.

Clinical trial number

Not applicable.

Introduction

Niviventer, a small rodent primarily found in mountainous and forested regions across Asia, is notably prevalent in Yunnan and Guizhou provinces of China. Furthermore, niviventer has been identified as a host for a diverse array of pathogenic bacteria and parasites [1,2,3].

Bordetella pseudohinzii (B. pseudohinzii), a recently identified member of the Bordetella genus, has garnered attention due to its enigmatic taxonomic status and its potential role as a pathogen [4]. The Bordetella genus includes well-known pathogenic species like B. pertussis, the causative agent of whooping cough, and B. parapertussis, which causes a milder form of the disease [5]. Unlike its more studied counterparts, B. pseudohinzii remains poorly characterized, with its biological traits and pathogenic potential still under scrutiny. Originally isolated from rodents, the reports focused mainly on laboratory mice, its distinction from B. hinzii complicates accurate species identification [6,7,8].

Current evidence suggests that B. pseudohinzii primarily infects rodent hosts, with mice serving as a significant reservoir. Although its pathogenicity is not fully elucidated, the bacterium’s presence in animals raises concerns about zoonotic transmission and risks to immunocompromised individuals, such as those with cystic fibrosis [9, 10]. Notably, clinical reports have documented lung infections caused by B. pseudohinzii in immunocompetent hosts [6]. The bacterium’s antibiotic resistance and pathogenic capabilities remain under investigation, highlighting the need for further research to evaluate its risks. Transmission among cohabiting mice has been observed, although the risk of human infection remains uncertain [11, 12].

Globally, B. pseudohinzii prevalence data are limited, and in China, data on its prevalence and pathogenicity in wild rodents are severely lacking, focusing mostly on laboratory mice [11,12,13]. Research in Malaysia has identified B. pseudohinzii in diseased rats, emphasizing the importance of monitoring this pathogen in wild rodent populations [14].

We first isolated B. pseudohinzii from wild niviventer in Guizhou, China, this study aims to conduct phenotypic resistance testing and whole-genome sequencing (WGS). By exploring its genetic evolution and resistance mechanisms, this research will enhance our understanding of B. pseudohinzii’s genetic features, antimicrobial resistance, and potential pathogenic mechanisms, thereby informing future public health strategies and preventive measures.

Methods and materials

Source of strains

Wild niviventer specimens were collected following the Technical Standards for Vector Surveillance—Rodents issued by the Chinese Ministry of Health [15]. Free-living individuals were captured in rural areas of Qiandongnan Autonomous Prefecture, Guizhou Province, China. All specimens were sourced from natural habitats, without involvement of privately owned or captive-bred animals. This study was approved by the Ethics Committee of the Guizhou Provincial Center for Disease Control and Prevention and adhered to ethical guidelines for the use of wildlife. Traps were set overnight and retrieved the following morning. Experienced operators performed cervical dislocation to euthanize live mice (weight < 200 g), ensuring a rapid and painless process, this method of euthanasia is recommended in animal euthanasia guidelines in China and the United States [16, 17]. This research complied with all relevant wildlife protection laws. The species, size, and sex of the specimens were identified according to the Handbook of Important Medical Animals in China [18]. After identification, the mice were promptly dissected in the local disease prevention and control center’s laboratory under sterile conditions. Lung tissues (approximately 500 mg) were excised using surgical scissors and immediately inoculated into 1.5 ml of brain heart infusion liquid containing 20% glycerol for preservation. The specimens were homogenized and plated onto Columbia blood agar plates, which were then incubated at 37 °C for 48 h to observe microbial growth. Single colonies were isolated for further purification on additional Columbia blood agar plates.

Identification of strains by MALDI-TOF MS and 16 S rRNA gene sequencing

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) was employed for the identification of microorganisms in the samples, instrument and model: Autof MS 1000 (manufactured by Antu Bio Co., China), the confidence scoring system is as follows: a score of 9.5–10.0 indicates reliable species-level identification; a score of 9.0–9.5 indicates reliable genus-level identification with species-level reference; a score of 6.0–9.0 indicates genus-level reference; and a score of 0.0–6.0 indicates unreliable results. This mass spectrometry-based proteomics technique offers several advantages, such as convenience, speed, and accuracy, over traditional biochemical methods [19, 20]. Purified colonies were initially analyzed using MALDI-TOF MS to identify potentially significant isolates for further study.

For molecular identification, genomic DNA was extracted from the strains using a nucleic acid extraction kit (Hangzhou Baiyi Technology Co., Ltd.). PCR amplification was performed using primers 27 F (5′-AGTTTGATCMTGGCTCAG-3′) and 1492R (5′-AGTTTGATCMTGGCTCAG-3′). The total reaction volume was 50 µL, which included: 25 µL of Premix Taq, 2 µL of each primer (27 F and 1492R), 2 µL of template DNA, and 19 µL of water. The PCR amplification conditions were as follows: initial denaturation at 94 °C for 10 min; 40 cycles of denaturation at 94 °C for 45 s, annealing at 52 °C for 60 s, and extension at 72 °C for 60 s; followed by a final extension at 72 °C for 10 min. The resulting PCR products were sequenced, and the obtained sequences were compared with sequences in the NCBI BLAST database. Homologous sequences with high similarity from GenBank were downloaded for subsequent phylogenetic analysis.

Whole-genome sequencing

A hybrid assembly strategy combining second-generation and third-generation (Oxford Nanopore) sequencing data was employed to leverage the high accuracy of second-generation sequencing and the long-read capabilities of third-generation sequencing, yielding a complete and accurate genome assembly of the strain.

Second-Generation Sequencing was performed using advanced technologies provided by the Beijing Genomics Institute. The process began with DNA fragmentation via enzymatic digestion to generate fragments of 200–500 bp in length suitable for sequencing. Adapter ligation was then performed on both ends of the DNA fragments. The adapters consist of a DNA sequence that includes amplification primers, sequencing primers, and barcode sequences. The barcode allows for distinguishing between different samples and ensuring sequencing accuracy. Following adapter ligation, PCR amplification was carried out to construct the sequencing library. The amplification was performed using primers with dual-barcode sequences, and the products were purified before sequencing. The sequencing was conducted on the BGISEQ platform, which utilizes DNA nanoball (DNB) amplification technology. This method forms DNA nanoballs through rolling circle amplification and generates high-quality raw sequencing data in FASTQ format.

Third-generation sequencing was performed using Oxford Nanopore’s platform, following the proprietary library preparation protocols. DNA fragmentation was carried out using Oxford Nanopore’s DNA fragmentation method, and adapters were ligated to the DNA fragments. Library quality control was conducted using the Agilent Bioanalyzer to assess fragment size distribution, ensuring the library was suitable for sequencing. Sequencing was performed on the Oxford Nanopore platform using electrophoretic migration to load DNA samples into nanopore chips for real-time single-molecule sequencing. Data acquisition and preliminary processing were carried out using the Oxford Nanopore MinKNOW software, and raw sequencing data were obtained in FASTQ format.

The hybrid assembly strategy combined both second- and third-generation sequencing data using the Micro IBS Analyzer software (Beijing MicroFuture Technology Co., Ltd.) for pathogen genome assembly and identification. First, second-generation sequencing data underwent quality control using Trimmomatic (v0.39) to remove low-quality reads. Third-generation sequencing data were filtered and corrected using the PBJelly tool. Next, an initial assembly of the second-generation data was performed using SPAdes, while the third-generation data were initially assembled with Canu (v2.1.1). Finally, the hybrid assembly was refined by integrating both datasets with Pilon (v1.23), which allowed for genome correction and the generation of high-quality, clean data in FASTA format.

Genetic analysis

16 S rRNA gene sequences of the relevant strains were downloaded from the NCBI database and the LPSN database (https://lpsn.dsmz.de/) to ensure they originated from standard reference strains. The phylogenetic tree was constructed using the Neighbor-Joining method implemented in MEGA 11. Bootstrap values, expressed as percentages, are indicated at the nodes of the branches. Circular genome maps were generated using CGView online software (https://stothardresearch.ca/cgview/). For bacterial species identification, Average Nucleotide Identity (ANI) and digital DNA-DNA hybridization (dDDH) were used as gold standards. Thresholds for ANI and dDDH were set at 95% and 70%, respectively, with values above these thresholds indicating the strains belong to the same species [21,22,23]. Genomic homology was analyzed using ANI to compare the genomic sequences of the strains against reference databases, providing a measure of genetic relatedness (https://www.ezbiocloud.net/). Additionally, dDDH analysis was performed to further confirm bacterial identity by quantifying the genomic similarity between the unknown strains and reference strains (https://ggdc.dsmz.de/).

Genome annotation was carried out using several databases, including the Comprehensive Antibiotic Resistance Database (CARD), Kyoto Encyclopedia of Genes and Genomes (KEGG), Clusters of Orthologous Groups (COG), Non-Redundant Protein Database (NR), Pathogen-Host Interaction (PHI), Swiss-Prot, Virulence Factor Database (VFDB), Metacyc, and the Carbohydrate-Active Enzyme Database (CAZY). These databases were accessed and annotated through the Beijing Micro Future Pathogen Microbiological Information Analysis System platform.

Drug resistance phenotype

In accordance with Clinical and Laboratory Standards Institute (CLSI) guidelines, with Escherichia coli ATCC 25,922 serving as the quality control strain, the antibacterial activity of antibiotics was assessed using the drug-sensitive plate method (Customized AST plate CHN5FGNF, Customized AST plate CHN6FGNF). The minimum inhibitory concentration (MIC) of bacterial solutions was measured to evaluate in vitro antibacterial activity. A total of 29 antibiotics were tested.

Results

Preliminary identification of strains

The isolated and purified strains 21F10, 22F12, and 27F25 were cultured on a Columbia plate at 37℃ under aerobic conditions for 72 h, and colonies appeared as white, convex, circular shapes with smooth surfaces, lacking hemolytic zones (Fig. 1A). The results of the Gram staining were negative, and they showed the same short rod shape under the microscope (Fig. 1B). MALDI-TOF MS analysis identified strain 21F10 as B. avium (identification score: 7.01), strain 22F12 as Achromobacter denitrificans (identification score: 6.299), and strain 27F25 as B. hinzii (identification score: 6.048).

The results of 16 S rRNA gene sequencing using PCR amplification with primers 27 F and 1492R were compared against the NCBI BLAST database. The sequences of strains 21F10, 22F12, and 27F25 showed 99.93%, 99.93%, and 99.79% identity, respectively, to B. hinzii LMG13501^T (AF177667). These results suggest that strains 21F10, 22F12, and 27F25 are likely B. hinzii. Following WGS, more complete 16 S rRNA sequences were obtained for strains 21F10, 22F12, and 27F25, each comprising 1519 base pairs (NCBI accession number: PQ881859). Pairwise comparisons of the 16 S rRNA sequences using NCBI BLAST revealed 100% similarity between strains. Further comparisons indicated that the 16 S rRNA sequences of all three strains exhibited 100% similarity with B. hinzii LMG13501^T (AF177667), 99.87% similarity with B. pseudohinzii 8-296-03^T (JHEP02000033), and 99.28% similarity with B. avium ATCC35086^T (AF177666).

These findings indicate that 16 S rRNA sequencing alone cannot definitively distinguish whether the strains are B. hinzii or B.pseudohinzii.

Fig. 1

(A) Colony morphology after 72 h of growth on the Columbia plate. (B) Morphology of strain under microscope (100x oil lens)

Phylogenetic relationships based on 16 S rRNA gene sequences

We downloaded all publicly available, valid 16 S rRNA gene sequences from Bordetella species listed in the LSPN, totaling 15 sequences, and included Allopoulusillimonas ginsengisoli DCY25^T (EF672088) as an outgroup. A phylogenetic tree was constructed using the Neighbor-Joining method. The results showed that strains 21F10, 22F12, and 27F25 were most closely related to B. hinzii LMG13501^T (AF177667), forming a distinct branch together, followed by B. pseudohinzii 8-296-03^T (JHEP02000033) (Fig. 2). Notably, these strains also exhibited close relationships with B. parapertussis ATCC15311^T (U04949), B. bronchiseptica NBRC13691^T (AB680479), and B. pertussis ATCC9797^T (U04950) (Fig. 2). These results reaffirm that strains 21F10, 22F12, and 27F25 are closely related to B. hinzii and B. pseudohinzii, and suggesting that their pathogenic potential should not be overlooked.

Fig. 2

Phylogenetic tree(Neighbour-joining)with 16 S rRNA gene sequences of strains 21F10, 22F12, 27F25 and 16 other Bordetella species

WGS characterization of strains

We obtained second-generation WGS data for strains 21F10, 22F12, and 27F25. Although the assembly quality was adequate for strain identification, it was insufficient for a complete genome analysis. Therefore, strain 21F10 underwent third-generation sequencing, which allowed successful assembly of both second- and third-generation sequencing data into a complete genome sequence.

The WGS data for strain 21F10 resulted in a genome with 99.94% integrity, 100% splicing quality, and a contamination level of 2.81%. The genome contains 63 tRNAs and three copies each of 23 S rRNA, 16 S rRNA, and 5 S rRNA genes. The total genome length is 3,548,721 base pairs, with a GC content of 65.42%. The longest sequence, N50, and N75 values were all 3,548,721 base pairs, with no gaps present. According to the evaluation criteria for high-throughput genetic data as outlined by Bowers et al. and Yunfeng Duan et al., this genome sequence is considered complete and of high quality [24, 25]. The WGS data for strain 21F10 has been uploaded to Genome Sequence Archive (GAS accession number: CRA022358). We mapped the complete gene. Prokka annotation was used to annotate genes, with a total of 3511 genes annotated. The map features a circular representation of the coding regions, color-coded by functional categories, including non-coding RNAs (tmRNA, tRNA, rRNA), GC content and CDS, highlighting functional categories such as coding sequences, tRNAs, and rRNAs, only a few gene names and their locations are shown (Fig. 3).

Fig. 3

Complete gene map of strain 21F10

Comparison of WGS data

We searched for whole-genome sequences of 15 Bordetella species on NCBI, 11 of which included complete WGS data. The ANI values for strains 21F10, 22F12, and 27F25 compared with B. pseudohinzii (GCF 001698185.1) were 97.45%, 97.24%, and 97.20%, respectively, all exceeding the 95% threshold for species delineation. Additionally, the ANI values between strains 21F10, 22F12, and 27F25 were above the 95% threshold, confirming their identity as B. pseudohinzii (Fig. 4). In contrast, the ANI values between these strains and B. hinzii (GCF 006770325.1) were 92.90%, 93.11%, and 93.07%, demonstrating that they are distinct from B. hinzii, although closely related (Fig. 4). Furthermore, ANI values between B. parapertussis (GCF 004008295.1), B. pertussis (GCF 004008975.1), and B. bronchiseptica (GCF 900636925.1) all exceeded 95%, indicating that it is difficult to distinguish between these species based solely on ANI analysis (Fig. 4).

For dDDH analysis, the DDH values between strain 21F10 and strains 22F12, 27F25, and B. pseudohinzii (GCF 001698185.1) were 99.00%, 98.80%, and 78.30%, respectively, all above the 70% threshold, further confirming their identity as B. pseudohinzii (Table 1). Notably, the DDH value between strain 21F10 and B. hinzii (GCF 006770325.1) was only 23.50%, a significant difference (Table 1).

The above results demonstrate that ANI and dDDH analyses conclusively identify strains 21F10, 22F12, and 27F25 as B. pseudohinzii, rather than B. hinzii.

Fig. 4

Heat map of ANI between strain 21F10 and 11 Bordetella species

Table 1 DDH values between 21F10 and 22F12, 27F25 and 11 Bordetella species

Gene function analysis

The complete gene sequence of strain 21F10 was annotated in 9 databases, 66, 2962, 2983, 3250, 218, 1459, 289, 790 and 87 genes were annotated in CARD, KEGG, COG, NR, PHI, Swiss-Prot, VFDB, Metacyc and CAZY databases, respectively.

A total of 2962 orthologous protein-coding genes were mapped to 37 KEGG metabolic pathways. The mapping highlights metabolic pathways critical for bacterial survival and adaptation, such as amino acid and carbohydrate metabolism. The predominant pathways were the global and overview maps, representing 41.76% of the annotations, followed by amino acid metabolism at 7.73% and another carbohydrate metabolism category at 7.29%. These pathways are crucial for sustaining bacterial metabolic functions (Fig. 5A).

In the NR database, the gene sequence of strain 21F10 was translated into amino acid sequences, revealing 3250 annotated genes. Notably, B. pseudohinzii was the most frequently annotated, comprising 72.28% of the annotations, followed by B. sp at 10.03%, and B. hinzii accounts for 7.08% (Fig. 5B).These results also indicated that the strain was B. pseudohinzii.

Fig. 5

(A) KEGG analysis of strain 21F10. (B) NR analysis of strain 21F10

A total of 218 pathogen-host interaction-related genes were annotated using the PHI database (with at least 40% identity). Among these, 42 genes exhibited greater than 60% identity, including one Hfq gene from B. pertussis with 100% identity. These genes are associated with bacterial metabolic adaptability, antioxidant capabilities, and environmental sensing, suggesting that these strains may survive and proliferate in complex and fluctuating environments. Additionally, they may possess strong cell membrane regulation, chemotaxis, and stress response functions (details are provided in Table S1). These genes were categorized into eight groups, with ‘reduced virulence’ (147 genes) being the most prominent, followed by ‘unaffected pathogenicity’ (30 genes), ‘increased virulence (hypervirulence)’ (19 genes), ‘loss of pathogenicity’ (14 genes), ‘lethal’ (3 genes), ‘effector (plant avirulence determinant)’ (3 genes), and ‘chemistry target: sensitivity to chemical’ (1 gene) and ‘chemistry target: resistance to chemical’ (1 gene) (Fig. 6A).

A total of 122 virulence genes were identified with at least 40% identity, and 64 of these genes showed greater than 60% identity. Of these, 58 are known to be associated with Bordetella species (details are provided in Table S2). Virulence factors were classified into eight functional categories: adherence, antiphagocytosis, invasion, iron uptake system, regulation, secretion system, stress protein, and toxin.With the highest number of annotated genes associated with flagella, fimbriae, LOS, and LPS. These factors predominantly contribute to invasion, adherence, toxin production, and secretion systems (Fig. 6B).

COG analysis annotated 2983 genes, categorized into 26 functional groups. The major annotated functions included ‘amino acid transport and metabolism’, ‘translation, ribosomal structure and biogenesis’,‘cell wall/membrane/envelope biogenesis’, ‘transcription’, ‘energy production and conversion’, ‘general function prediction only’, and ‘inorganic ion transport and metabolism’ (Fig. 6C). These findings align closely with the KEGG metabolic pathway analysis, highlighting the involvement of numerous genes in fundamental metabolic processes.

Fig. 6

(A) Types distribution of PHI genes of strain 21F10. (B) VFDB analysis of strain 21F10. (C) COG analysis of strain 21F10

In the CARD database, we accepted results with at least 40% identity, resulting in the annotation of 13 antibiotic families and 41 antibiotic resistance genes (66 fragments) in strain 21F10 (details are provided in Table S3). These genes were linked to four primary resistance mechanisms: antibiotic efflux, antibiotic inactivation, antibiotic target alteration, and antibiotic target replacement. These included 13 resistance gene families, such as resistance-nodulation-cell division antibiotic efflux pump, small multidrug resistance antibiotic efflux pump, ATP-binding cassette antibiotic efflux pump, AAC(3), and others. Notably, 24 resistance genes related to the resistance-nodulation-cell division antibiotic efflux pump were annotated, with OprM, Pseudomonas aeruginosa CpxR, Enterobacter cloacae acrA, Escherichia coli acrA being the most prevalent. These genes render the strain resistant to commonly used antibiotics like cephalosporins, complicating treatment strategies(Fig. 7).

Annotations for strain 21F10 in SwissProt, MetaCyc, and CAZy databases are detailed in the supplementary material (Table S4, S5, and S6, respectively).

Fig. 7

Drug-resistant genes of strains 21F10 annotated in CARD database (identity > 40%). ‘aep’ antibiotic efflux pump; ‘(ABC)’ ATP-binding cassette; ‘grgc’ glycopeptide resistance gene cluster; ‘MFS’ major facilitator superfamily; ‘pmr pt’pmr phosphoethanolamine transferase; ‘(RND)’ resistance-nodulation-cell division; ‘(SMR)’ small multidrug resistance; ‘sr sul’ sulfonamide resistant sul; ‘trdr dfr’ trimethoprim resistant dihydrofolate reductase dfr; ‘uprp’ undecaprenyl pyrophosphate related proteins; ‘Ab AbaF’ Acinetobacter baumannii AbaF; ‘Enc acrA’ Enterobacter cloacae acrA; ‘Esc acrA’ Escherichia coli acrA; ‘Pa CpxR’ Pseudomonas aeruginosa CpxR; ‘Pa emrE’ Pseudomonas aeruginosa emrE

Drug resistance

Antibiotic susceptibility testing was conducted on strain 21F10 across a panel of 29 antibiotics. The results revealed that this strain exhibits resistance to five cephalosporins: cefoxitin, cefuroxime, cefotaxime, cefazolin, and ceftiofur, and showed sensitivity to 24 other antibiotics (Table 2).

Table 2 Drug susceptibility test results

Discussion

Our study was the first to isolate B. pseudohinzii from the lung tissues of wild niviventer in China. We employed multiple identification techniques, including MALDI-TOF MS and 16 S rRNA gene sequencing, and conducted comprehensive WGS and analysis. Additionally, we performed drug susceptibility testing to further characterize the strain.

It is well known that classic Bordetella sp., such as B. pertussis and B. parapertussis, poses a serious threat to human beings, but as non-classic Bordetella sp., B. hinzii was originally thought to exist mainly in birds and poultry [26]. It is also capable of infecting rabbits and laboratory mice [7, 27, 28], but there are increasing reports that it may cause respiratory infections and other clinical cases, especially in immunocompromised people [29,30,31,32]. Also, as a non-classic Bordetella sp., human understanding of B.pseudohinzii is not enough, and the distribution of B.pseudohinzii in the host needs to be further discovered, and the pathogenicity and drug resistance of B.pseudohinzii are not clear, only in vitro model studies have shown that B.pseudohinzii attaches to respiratory cilia, destroys cilia function within 4 h, and causes epithelial injury within 24 h [33], let al.one the epidemiological characteristics. Existing studies have shown that B.pseudohinzii was likely to infect the host through respiratory tract [11, 12], and related clinical cases have been reported [6].

In this study, we elucidate the intricate genetic and phenotypic characteristics of B.pseudohinzii isolated from the lung tissues of niviventer in Guizhou, China. Our results underscore the limitations of traditional identification methods such as MALDI-TOF MS and 16 S rRNA gene sequencing, which frequently confound B. pseudohinzii with B. hinzii. The sequence obtained after amplification of nucleic acid with primers was compared on BLAST, and the result was not B. pseudohinzii but B. hinzii, on the one hand, the similarity between them was very high, and the length of 16 S rRNA amplification (primers 27 F and 1492R) was not enough, on the other hand, there was no 16 S rRNA data of B.pseudohinzii in BLAST database, so it was very important that we uploaded complete 16 S rRNA sequence of B.pseudohinzii to GenBank. Subsequently, we extracted sufficiently long 16 S rRNA sequences from the WGS data, but the results of the sequence alignment and phylogenetic tree still showed confusion between B. pseudohinzii and B. hinzii. Finally, the strain was accurately identified as B. pseudohinzii based on both ANI (> 95%) and dDDH (> 70%) values. Additionally, the gene sequence of the strain was translated into an amino acid sequence using the NR database, where the highest proportion of annotations was identified as B. pseudohinzii (72.28%), while B. hinzii accounted for only 7.08%. These results further support the identification of the strain as B. pseudohinzii. WGS provides high-resolution insights, allowing for precise species identification and the discovery of unique resistance genes, demonstrating its superior reliability for accurate identification through ANI and dDDH. Furthermore, the larger DDH value differences between B. pseudohinzii and B. hinzii suggest that dDDH may offer more distinct advantages compared to ANI. Additionally, the identification of B. parapertussis based on ANI and DDH values between B. pertussis and B. bronchiseptica reveals some confusion. Ivanov YV et al. reported differences in the hosts of B. pseudohinzii and B. hinzii, with B. pseudohinzii predominantly found in mice and B. hinzii mainly in poultry [7].

The genome annotation of B. pseudohinzii reveals a repertoire of genes associated with virulence, including those contributing to increased virulence, potential loss of pathogenicity, and sensitivity to chemical agents. The functional annotation, informed by COG and KEGG pathway analyses, indicates a strong metabolic capability essential for bacterial survival. Notably, we identified 122 putative virulence genes categorized into functions such as adherence, invasion, and toxin production, with flagella and fimbriae being prominent. PHI analysis revealed 19 genes that may lead to enhance pathogenicity, which is worthy of attention and in-depth study. The annotation of 41 antibiotic resistance genes, particularly those conferring resistance through mechanisms like antibiotic efflux and target alteration, highlights the challenge of managing infections caused by B. pseudohinzii. Drug susceptibility tests confirm resistance to multiple cephalosporins, underscoring the need for alternative therapeutic strategies.

Our findings contribute to a deeper understanding of the pathogenic and resistance profiles of B. pseudohinzii, emphasizing the importance of employing advanced genomic techniques for precise microbial characterization and effective treatment planning, reasonable methods should be adopted for the identification of different strains of Bordetella. The surveillance and assessment of B. pseudohinzii infection in wildlife, particularly rodents, holds significant scientific value. In the next step, we will focus on the actual pathogenicity of B. pseudohinzii in animal models, as well as the exploration of rapid diagnostic methods for B. pseudohinzii, the lung tissues of wild mice will be collected for isolation and culture to analyze epidemiological characteristics.

The limitations of this study include the absence of animal or cell experiments for gene function validation and the insufficient sample size to support the epidemiological distribution characteristics of Bordetella pseudohinzii in wild Niviventer species.

Conclusion

The first isolation of B. pseudohinzii from the lung tissue of wild niviventer was reported, in Guizhou, China, exhibits distinct genetic and phenotypic characteristics compared to B. hinzii, challenging the accuracy of traditional identification methods such as MALDI-TOF MS and 16 S rRNA gene sequencing. WGS proved superior, providing clear differentiation through ANI and DDH. Our genome annotation identified 122 putative virulence genes and 66 antibiotic resistance genes, highlighting the pathogen’s significant virulence potential and complex resistance mechanisms, including multiple cephalosporin resistances. The findings reveal a complex pathogenic profile and notable antibiotic resistance, providing important insights for the future prevention and treatment of B. pseudohinzii infections in humans, as well as underscoring the need for monitoring B. pseudohinzii in rodent populations.