PDF
Citation
Print
INTRODUCTION
Pasteurella canis, a non-motile, facultatively anaerobic, gram-negative coccobacillus initially classified as Pasteurella multocida, was re-classified in 1985 based on DNA sequence homology [1]. P. canis produces smaller colonies than P. multocida on blood agar plates incubated in 5% CO2 at 37°C for 24 h [2]. Phylogenetic trees based on 16S rRNA sequences and sodA fragment sequences (449–473 bp) from Pasteurella type strains revealed that P. canis is more closely related to P. stomatis and P. dagmatis than to P. multocida subsp. multocida and P. multocida subsp. septica [3]. In biochemical assays, P. canis consistently exhibits positive results for ornithine decarboxylation and indole production, while tests for urease activity are negative [4].
P. canis primarily colonizes the oral cavity of dogs [1]. The Emergency Medicine Animal Bite Infection Study Group [5] performed bacteriological assays of infected wound sites in humans resulting from dog and cat bites, which revealed that P. canis is the most common species isolated from wounds caused by dog bites, whereas P. multocida subsp. multocida and P. multocida subsp. septica are the most commonly isolated from wounds caused by cat bites. P. canis is reportedly involved in human bacteremia [6], soft tissue infection [8], respiratory infection [9], septic arthritis [10], osteomyelitis [11], gastrointestinal infection [12], breast implant infection [13], and peritonitis [14]. Basic investigations of P. canis isolates are scarce, although the prevalence of unique toxin genes (i.e., cytolethal distending toxin [cdt]A-cdtB-cdtC) in P. canis isolates from humans and companion animals has been described [15].
Various genotyping approaches are used to conduct epidemiological investigations of P. canis isolates. One approach is to construct dendrograms based on repetitive element-based fingerprinting using repetitive extragenic palindromic sequence-based PCR, enterobacterial repetitive intergenic consensus-based PCR, randomly amplified polymorphic DNA-based PCR, or M13-based PCR [4, 16]. Another approach is to genotype the virulence-associated genes tadD (encoding tight adherence protein D), ptfA (encoding type IV fimbriae), and ompA (encoding outer membrane protein A) for the molecular characterization of P. canis isolates [16].
Multilocus sequence typing (MLST) is an established method for epidemiological and evolutionary investigations of pathogenic bacteria. MLST is a DNA-based approach for characterizing bacterial isolates based on the well-established principle of multilocus enzyme electrophoresis, in which internal fragments of 400–600 bp from (usually) seven housekeeping enzyme genes are sequenced to determine the genetic relationships among isolates. MLST is well-established for P. multocida isolates [17]. The seven housekeeping enzyme genes for P. multocida MLST were selected based on their widespread distribution throughout the chromosome and the different enzyme functions and include adenylate cyclase (adk) and purine nucleoside phosphorylase (deoD) related to nucleotide biosynthesis, 3-phosphoshikimate 1-carboxyvinyl transferase (aroA) and glutamate dehydrogenase (gdhA) involved in amino-acid biosynthesis, glucose-6-phosphate 1-dehydrogenase (g6pd) involved in the pentose phosphate pathway, malate dehydrogenase (mdh) related to the tricarboxylic acid cycle, and phosphoglucose isomerase (pgi) involved in glycolysis. The P. multocida PubMLST website (https://pubmlst.org/organisms/pasteurella-multocida/) provides two P. multocida MLST schemes: the Multi-host MLST Scheme covers isolates from various hosts (including cattle, sheep, pigs, and birds), whereas the RIRDC MLST scheme was initially developed to investigate avian isolates.
To our knowledge, for the first time, we aimed to establish an MLST scheme for P. canis isolates from human patients and diseased companion animals (primarily dogs) in Japan and Korea. Further, we aimed to provide further insights into the population biology of P. canis.
MATERIALS AND METHODS
P. canis isolates
The study protocol was approved by the ethics committees of Kitasato Institute Hospital (Tokyo, Japan) and Hallym University Hospital (Seoul, Korea). We collected novel Japanese and Korean P. canis isolates separately at the two institutes. Additionally, we obtained isolates reported in previous studies [15, 16, 18]. P. canis isolates from diseased companion animals and infected humans collected in Japan and Korea were identified using PCR-based 16S rRNA sequencing data or matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry analysis results. Background information on the P. canis isolates analyzed in this study is provided in Table 1. We analyzed 39 isolates from companion animals and 22 from humans. The animal isolates included 13 from males, 13 from females, and 13 from animals of unknown sex, aged 2–16 yrs. The human isolates included nine from men, nine from women, and four from patients of unknown sex, aged 3–80 yrs. The animal isolates were collected between 2018 and 2023, and the human isolates between 2017 and 2023 (excluding 2020). Animal isolates were primarily collected from pus/skin/wound (N=13), nose (N=12), ear (N=7), and throat/tooth (N=4); human isolates from pus (N=18) and sputum (N=3). We analyzed 40 Japanese isolates, 18 Korean isolates, and three isolates from the UK and China. One isolate per host was stored at –70°C to –80°C until genotypic analysis.
We retrieved whole-genome sequences (WGSs) of P. canis (N=10), including two Japanese isolates, five Korean isolates, and three isolates from the UK and China, from the National Center for Biotechnology Information (NCBI) (as of September 18, 2024). NCBI recommends using the complete/circular WGS sequence (accession No. NZ_CP085871.1) of Korean isolate HL_NV12211 from dog pus as the P. canis reference genome because the genome sequence of the National Collection of Type Cultures (NCTC) isolate 11621(T) from a dog throat sample is a contig WGS sequence (accession No. NZ_UGTV00000000.1), not a complete/circular one.
Selection of seven housekeeping genes
To establish an MLST scheme, we used the seven housekeeping enzyme genes adk, aroA, deoD, gdhA, g6pd, mdh, and pgi, which are the genes targeted in P. multocida MLST [17]. Supplemental Data Fig. S1, constructed using the graphic genome display on the NCBI website, shows the locations of the seven housekeeping genes in the P. canis HL_NV12211 reference genome, indicating that they are widespread throughout the reference genome. The complete sequence lengths of adk, aroA, deoD, gdhA, g6pd, mdh, and pgi in the HL_NV12211 genome are 645, 1,323, 717, 1,347, 1491, 936, and 1,650 bp, respectively (Supplemental Data Fig. S1). We confirmed that these seven gene sequences in the reference genome (HL_NV12211) are consensus sequences, i.e., allelic profile 1–1–1–1–1–1–1 and sequence type (ST) 1.
Amplification and sequencing of the seven housekeeping genes
Multiple alignments of the seven genes were constructed using 10 WGSs, including the HL_NV12211 reference genome, revealing conserved and variable sequences in these seven genes. We designed PCR primer sets (forward and reverse) using Primer3Plus (https://www.primer3plus.com) [15]. Primer specificity was examined using the nucleotide Basic Local Alignment Search Tool on the NCBI website [15]. Supplemental Data Table S1 provides details on the PCR assays used to amplify internal fragments of the seven genes. To amplify the seven genes for MLST, we used 35 cycles of denaturation at 98°C for 10 secs, annealing at the relevant temperature for 15 secs, and extension at 72°C for 1 min. When the single amplicon of g6pd could not be obtained using the forward primer “Pc_zwf_F,” the alternative forward primer “alt_Pc_zwf_F” (CAGGAGCTGAGTCGTTAGGC; 20-mer) was used along with the reverse primer “Pc_zwf_R.” The annealing temperature for the alternative primer was 54°C, and the amplicon size was 829 bp.
DNA was extracted from the isolates via incubation in Tris-EDTA buffer at 97°C for 10 mins [19]. Isolate PA42 or HL_NV12211 [2] (Table 1) was used as a positive control, and DNase/RNase/protease-free water was used as a negative control in each PCR assay. PCR products were examined using 1.5% agarose gel electrophoresis in a buffer consisting of Tris-acetate (40 mM) and EDTA (1 mM). The same forward and reverse primers were used for PCR amplification and direct sequencing. The seven PCR products were sequenced on an Applied Biosystems 3730xl DNA Analyzer with BigDye Terminator v3.1 (Thermo Fisher Scientific, Waltham, MA, USA). Both strands of each gene fragment were sequenced. The extraction, amplification, and sequencing procedures were conducted at the two institutes independently.
Determination of the fragment sequence, variable site(s), allelic profile, and ST of each isolate
The sequencing data obtained at the Korean institute were transmitted to the Japanese institute, where sequences were assembled from chromatograms generated exclusively by an Applied Biosystems 3730xl DNA Analyzer. For each gene, the different fragment sequences and variable site(s) obtained from the 61 isolates were assigned a distinct allele number. For each locus, we determined the length of the sequenced fragment, number of alleles, number (%) of variable sites, and the ratio of nonsynonymous to synonymous polymorphisms (dN/dS ratio) [20]. Each isolate was defined by an allelic profile consisting of seven integers corresponding to the allele numbers of the seven genes in the order adk, aroA, deoD, gdhA, g6pd, mdh, and pgi. Each unique allelic profile was assigned an ST [20].
Construction of a phylogenetic tree based on the concatenated sequences of the seven housekeeping genes
A phylogenetic tree of the concatenated sequences of the seven housekeeping genes from all isolates was constructed using the neighbor-joining method, with 1,000 bootstrap replicates [21]. Evolutionary analysis was performed using MEGA11 [22]. The corresponding ST and clonal complex (CC) were indicated for each isolate in the tree. P. multocida subsp. multocida (American Type Culture Collection [ATCC], 43137(T); accession No. NZ_CP008918.1) and P. multocida subsp. septica (NCTC, 11995(T); NZ_UGSV00000000.1) were used as outgroups.
Construction of goeBURST diagrams indicating relationships between STs and CCs among all isolates
The goeBURST algorithm [23] implemented in the PHYLOViZ software [24] was used to establish relationships between STs and CCs among the 61 isolates [25]. Numbers of different alleles were indicated between two connected STs. CCs were defined at the single-locus-variant level. We assessed the distribution of STs among different hosts (animal and human) and different countries (Japan, Korea, and others).
ompA genotyping
We used ompA genotyping for molecular characterization of the P. canis isolates because ompA is a virulence-associated gene for which genotyping is well-established [16]. Supplemental Data Table S1 provides information on the PCR assay used to amplify full-length ompA in 30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 30 secs, and extension at 72°C for 1 min. The same forward and reverse primers were used for PCR amplification and direct sequencing. Both strands of ompA were sequenced, and sequences were assembled from the ABI chromatograms.
Comparison of MLST diversity with that of ompA typing
Simpson’s diversity index values (with 95% confidence intervals) were calculated for ompA genotyping and MLST using the Comparing Partitions website (http://www.comparingpartitions.info) [27–29] to compare the diversity of MLST with that of ompA genotyping.
Ethics statement and data availability
The ethics committees of Kitasato Institute Hospital and Hallym University Hospital reviewed and approved the studies in humans and companion animals (approval Nos. 21061 and NON2024-001-001). Background information (including host species, isolation source, collection date, and geographic location) regarding the selected WGSs (accession Nos. NZ_UGTV00000000.1, NZ_CP085791.1, NZ_CP085873.1, NZ_CP085871.1, NZ_BPUX00000000.1, NZ_UATN00000000.1, NZ_CP083262.1, NZ_CP083396.1, NZ_WUMP00000000.1, and NZ_BQFX00000000.1) is available on the NCBI website (Table 2). The partial sequences of the seven housekeeping genes (accession Nos. in Table 3) and the ompA full-length sequences (accession Nos. LC769576–LC769603 and LC842364–LC842373) have been deposited in DDBJ/EMBL/GenBank.
RESULTS
Fragment sequence, variable site(s), allelic profile, and ST of each isolate
Supplemental Data Table S2 shows the basic results of the P. canis MLST scheme assessed. The sequence lengths of adk, aroA, deoD, gdhA, g6pd, mdh, and pgi were 424, 451, 483, 439, 429, 419, and 440 bp, respectively. The seven genes had allele numbers of 16, 13, 15, 18, 22, 19, and 18, respectively. The numbers (percentages) of variable nucleotide sites were 15 (3.5%), 16 (3.5%), 12 (2.5%), 39 (8.9%), 24 (5.6%), 17 (4.1%), and 18 (4.1%), respectively. The dN/dS ratios of all genes were <1.0, except for that of 1.29 for aroA, and most substitutions were synonymous, suggesting that most loci were under stabilizing selective pressure. Supplemental Data Fig. S2 provides detailed information on the sequences and nucleotide substitutions for the seven genes based on consensus sequences in P. canis isolate HL_NV12211.
The allelic profile, ST, and CC for each isolate are shown in Tables 2 and 3. The results revealed a diverse distribution of ST1–ST47. Three isolates, HL_NV12211, PA42, and PA101, belonged to ST1 (allelic profile 1–1–1–1–1–1–1); two isolates, PA44 and PA78, belonged to ST9 (8–7–6–2–13–1–7); three isolates, HL_D3081, PA95, and PA49, belonged to ST14 (4–1–12–5–18–5–7); two isolates, PA30 and HL2121, belonged to ST18 (8–6–1–9–15–11–5); four isolates, HL_D1250, PA23, PA48, and PA81, belonged to ST20 (9–1–1–8–22–10–5); two isolates, PA9 and NV25875, belonged to ST27 (4–9–3–1–13–13–12); two isolates, PA75 and HL1500, belonged to ST33 (13–5–1–14–15–13–2); two isolates, PA38 and PA88, belonged to ST36 (4–1–2–15–16–10–14); and three isolates, PA96, NV24345, and NV26624, belonged to ST37 (12–11–1–11–16–13–6). CC2 (N=2) consisted of ST2 (allelic profile 4–1–1–1–5–1–7) and ST39 (4–1–1–1–5–1–1); CC10 (N=3) consisted of ST10 (8–7–8–10–6–6–4), ST11 (10–7–8–10–6–6–4), and ST16 (8–10–8–10–6–6–4); CC18 (N=3) consisted of ST18 (8–6–1–9–15–11–5) and ST45 (8–1–1–9–15–11–5); CC31 (N=2) consisted of ST31 (13–1–2–15–12–9–15) and ST32 (13–1–2–15–12–9–16); and CC33 (N=3) consisted of ST33 (13–5–1–14–15–13–2) and ST41 (13–6–1–14–15–13–2) (Tables 2 and 3).
Phylogenetic tree based on the concatenated sequences of the seven housekeeping genes
Fig. 1 shows a phylogenetic tree of the concatenated sequences of the seven genes in all isolates constructed using the neighbor-joining method. Apart from P. multocida subsp. multocida ATCC 43137(T) and P. multocida subsp. septica NCTC 11995(T), which were used as outgroups, all isolates were clustered together. Furthermore, we found that isolates belonging to CC2, CC10, CC18, CC31, and CC33 were clustered together in the tree.
goeBURST diagram revealing the distribution of STs and CCs among all isolates
Fig. 2 shows goeBURST diagrams indicating the relationships between STs and CCs in all isolates. Fig. 2A indicates the differential distribution of STs among different hosts (animals and humans), and Fig. 2B shows the differential distribution of STs among different countries (Japan, Korea, and others).
ompA genotype(s)
For the first ompA genotyping analysis conducted at a Japanese institute alone, we used 48 isolates (excluding 13 Korean isolates) (Table 3). Fig. 3 shows a phylogenetic tree constructed based on the ompA amino-acid sequences, using the neighbor-joining method and the maximum-likelihood method with the Whelan and Goldman model. We found four clusters (1–4) in the tree. Identical ST20 isolates (N=4) were clustered in cluster 1; identical ST33 isolates (N=2) were clustered in cluster 2; identical ST9 (N=2) and ST14 (N=3) isolates were clustered in cluster 3; and identical ST1 (N=3) and ST36 (N=2) isolates were clustered in cluster 4, suggesting the validity of the MLST scheme.
Diversity indices of MLST and ompA typing
Simpson’s diversity index values (and 95% confidence intervals) of MLST (N=48) and ompA typing (N=48) were 0.987 (0.975–0.999) and 0.692 (0.612–0.772), respectively, indicating the high discriminatory ability of the MLST scheme established compared with that of ompA typing.
DISCUSSION
Many people have companion animals (including dogs and cats) in their homes in Japan and Korea. In addition, medical hospitals and nursing homes [30, 31] have introduced animal-assisted therapy as a mental health service for patients and older residents. Animals and humans are in constant close contact with the environment. Based on the “One Health” concept (https://www.cdc.gov/onehealth/index.html) [32], which is a comprehensive health control strategy for humans, contact animals, and their environments, bacterial pathogens with virulence factors that may be circulating should be carefully monitored to maintain an environment of total health. P. canis is often isolated from dog bite wounds in humans and may be transmitted from animals to humans via this route. Therefore, we established an MLST protocol specifically for P. canis. Jeong, et al. [33] recently described the prevalence and clinical features of Pasteurella infections in Korea and provided a systematic review and meta-analysis of Pasteurella bacteremia. Their results reflected the need for a better understanding of the rising incidence of Pasteurella infections and the global burden of Pasteurella bacteremia for appropriate case management.
We searched for information on the P. multocida Multi-host MLST Scheme (covering cattle, sheep, pigs, and birds) on the PubMLST website (https://pubmlst.org/bigsdb?db=pubmlst_pmultocida_seqdef&page=schemeInfo&scheme_id=1) (as of July 3, 2024). This website has hosted 375 MLST profiles since its release on May 29, 2009, and is frequently used. The allele numbers of adk, aroA, deoD, gdhA, g6pd, mdh, and pgi in P. multocida are 70, 104, 73, 85, 81, 69, and 102, respectively, whereas, in P. canis, they were 16, 13, 15, 18, 22, 19, and 18, respectively, in the 61 isolates. To clarify the similarities and differences in the evolution of each gene between P. multocida and P. canis, a number of P. canis STs should be determined in the future.
To examine the P. canis population biology, we analyzed 39 companion animal-origin (primarily dogs) and 22 human-origin isolates. goeBURST diagrams revealed diverse ST distributions among different hosts (dogs/humans). Stahel, et al. [34] described the phenotypic and genetic characterization of a P. canis isolate from a rabbit in Switzerland. P. canis bacteremia has been observed in a child after exposure to rabbit secretions [35]. Additionally, P. canis-associated pneumonia has been reported in a non-human primate black-tailed marmoset [36]. P. canis reportedly caused septic arthritis and soft tissue infection in humans after a sheep bite [37]. This microorganism has also been isolated from the oral cavity of a captive California sea lion [38]. Therefore, similar to P. multocida, P. canis can be isolated from various hosts. To validate our MLST scheme and to comprehensively examine the P. canis population biology, we need to collect and examine isolates from multiple hosts other than companion animals and humans in the future.
This study has two major limitations. We collected limited host demographics (i.e., host, sex, age, collection year, isolation source, and isolation country) of the isolates studied. More detailed information (e.g., underlying veterinary or medical situations, clinical diagnosis of infections, therapeutic approaches including antimicrobial administration and/or surgical interventions, and outcomes) should be retrieved to determine the relationships between STs and CCs and their clinical implications/pathogenetic significance. In addition, we could not obtain paired isolates from pets and their owners after receiving pet bites. Using MLST, we can assess the identity or differences in ST(s) of paired isolates in the future.
In conclusion, we could establish an MLST scheme for P. canis using isolates from human patients and diseased companion animals (primarily dogs) in Japan and Korea. Further, our findings provide insights based on P. canis population data (humans vs. animals and Japan vs. Korea) (Fig. 2A and 2B). This MLST scheme may be used to provide details regarding the epidemiology and evolution of P. canis infections on a global scale. The Japanese Veterinary Infection Control Association and the antimicrobial resistance (AMR) working group (consisting of veterinary/medical doctors and laboratory staff), based on the analyses of bacteria isolated from companion animals via blood culture [39], have indicated reduced AMR rates in Staphylococcus intermedius and Escherichia coli isolated from diseased companion animals in a veterinary hospital after the restriction of antimicrobial use [40]. In addition, P. canis collaborative studies between Japanese and Korean researchers [2, 4, 15] have been reported. In the future, we should establish a global research network for P. canis among veterinary/medical doctors and researchers to understand/estimate the rising incidence of P. canis infections and the global burden of P. canis-associated bacteremia.
XML Download