Characterization of Resistance and Virulence of Pasteurella multocida Isolated from Pet Cats in South China

Pasteurella multocida, one of the major zoonotic pathogens, may cause localized cellulitis, bacteremia, septic shock, and other symptoms in people. Accidental bites or scratches in close contact between humans and cats are one of the causes of P. multocida infection in humans. The prevalence of diseases due to P. multocida in humans may be rising as a result of the popularity of cat cafes in China. However, the epidemiology of P. multocida in cat-themed cafes in South China is yet to be reported. P. multocida in 92 healthy cats from four cafes with pet-cat themes was examined in their tonsils and oral cavities. The antibiotic susceptibility of each isolate was evaluated by using the disk-susceptibility testing method. The 19 P. multocida were selected for whole-genome sequencing (WGS), together with the genome data of eight previously described clinical strains isolated from humans, and the analysis of antimicrobial resistance (AMR) determinants, capsular genotyping, MLST genotyping, and virulence gene profiles was carried out. The results showed that 48.91% of cats (45/90) were positive for P. multocida. All isolated P. multocida stains were highly resistant to erythromycin (15 μg) (100%) and nonsensitive to fluoroquinolones (5 μg) (37.38%), while they were all susceptible to penicillin (10 untis), tetracycline (30 μg), doxycycline (30 μg), and chloramphenicol (30 μg). According to WGS analysis, those with the gyrA resistance gene were all insensitive to fluoroquinolone antibiotics. Virulence gene profiles showed that the genes pfhA, nanH, and plpB were practically all present in cats compared to humans. To conclude, increased antibiotic resistance, along with virulence factors, complicates the treatment of P. multocida infection. Thus, clinical treatment for P. multocida infection should be performed more cautiously.


Introduction
As a Gram-negative and facultatively anaerobic coccobacillus, Pasteurella multocida is one of the main zoonotic pathogens that can harm humans and livestock [1]. In animals, P. multocida is linked to chronic and acute diseases, including poultry cholera [2], swine atrophic rhinitis and suppurative bronchopneumonia [3], bovine respiratory disease [4], and respiratory disorders in varied hosts. However, many pets, such as dogs and cats, have P. multocida in their natural microflora [5]. Some research has shown that P. multocida is a proportion of the normal flora in the upper respiratory tract and nasopharynx of cats [6]. Therefore, cats are more commonly considered as healthy hosts of P. multocida [7,8]. P. multocida may be spread when a cat bites, scratches, or even licks an open wound on a person's skin. In humans, infection occurs more frequently in those who are immunosuppressed or have major co-morbidities [9], but it has also been reported in healthy individuals [10,11]. In recent years, the rapidly increasing number of pet-cat-themed cafes may have fueled the problem of P. multocida-related infections as intimate contact with cats has been a major factor in the case reports of P. multocida infection in humans. P. multocida infection has been linked to localized cellulitis; bacteremia [12]; and, on rare occasions, even septic shock and empyema [13]. A few cases of peritonitis linked to peritoneal dialysis have been reported [14][15][16].
Much of the current P. multocida study has concentrated on livestock and its virulence. However, using broad-spectrum antibiotics can contribute to the formation of multi-drug resistant bacteria. Moreover, fewer scientists are investigating the P. multocida resistance genes in pets. During our study, we reported the epidemiology of P. multocida in cat-themed cafes in south China with the purpose of evaluating the occurrence of P. multocida, identifying antimicrobial susceptibility profiles and resistance genes, and revealing virulence gene patterns in P. multocida.

Results
A total of 242 strains were obtained from 93 cats in the themed cafes for our study. Among these cats, we found that 48.38% were positive for P. multocida isolation (45/93). Additionally, the P. multocida took the largest percentage of the 242 isolates, followed by Neisseria zoodegmatis and Staphylococcus felis (Figure 1), suggesting that the flora of P. multocida are common in cats. The antibiotic susceptibility of all identified P. multocida was investigated by using agar disk diffusion tests, as recommended by document (M45 3rd edition, 2016) [17] CLSI. All 45 P. multocida isolates were shown to be resistant to erythromycin (15 µg) Erythromycin; this implied that erythromycin (15 µg) Erythromycin had an insignificant effect on these P. multocida stains. The susceptibility rates for other antimicrobial agents such as moxifloxacin (5 µg), levofloxacin (5 µg), and ceftriaxone (30 µg) moxifloxacin, levofloxacin, and ceftriaxone were 62.22% (28/45), 62.22% (28/45), and 80.00% (36/45), respectively. Additionally, the results of the antimicrobial susceptibility tests showed that every isolate of P. multocida was sensitive to penicillin (10 untis), tetracycline (30 µg), doxycycline (30 µg), and chloramphenicol (30 µg) ( Table 1) .  For the purpose of better understanding the antimicrobial resistance genes, the phylogenetic relationship, and the virulence genes of P. multocida isolates, whole-genome sequencing data were used.
The antimicrobial resistance genes were examined through the CARD database in order to better comprehend the mechanisms underlying these P. multocida isolates' antimicrobial resistance. The CARD data revealed that, with the exception of the resistance gene gyrAgyrA, all isolates carried the same antimicrobial resistance genes ( Figure 2). All sequences contained the resistance genes fabI, CRP, EF-Tu, rsmA, msbA, hmrM, and PBP3, but only pm7, pm32, pm33, pm98, pm101, pm106, pm111, pm112, and pm113 strains contained the mutations in gyrA. The strains harboring the gyrA resistance gene were nonsusceptible to levofloxacin (5 µg) and moxifloxacin (5 µg), which was consistent with the results of the antimicrobial susceptibility testing. On the basis of the MLST genotyping results, three P. multocida (pm18, 98, and 111) could not be classified by the MLST databases. In addition, there were six P. multocida (pm92, 94, 95, 99, 107, and 112) that could not be perfectly matched to the house-keeping gene est; therefore, their MLST typing indicated the closest ST264, and we referred to that as ST264*. Other MLST sequence types were diverse, comprising two ST3, three ST9, one ST11, one ST15, four ST25, one ST39, one ST54, one ST130, two ST155, one ST242, and two ST357. In order to find the correlation between P. multocida isolated from cats and humans, we gathered 8 whole-genome sequencing (WGS) data sets from P. multocida obtained from humans, which were found in GenBank (Accession numbers: CP026744, CP023516, CP023972, LS483473, NBTJ02000002, UGSW01000003, UGSV01000001, and UGSU01000002). The phylogenetic tree showed that pm98, together with UGSV01000001, was closely linked to strains obtained from humans. The results also revealed that the same ST typing was on the same branch ( Figure 3, Image A).
Image A Image B The pathogenicity of P. multocida is often linked to a variety of virulence factors [18]. Additionally, the potential virulence genes in the P. multocida genome were found by using the PmGT virulence genotyping database. In the epidemiological research of P. multocida, the following 24 genes that encode virulence factors have usually been analyzed [19]. The genes ptfA, fimA, hsf-1, hsf-2, pfhA, pfhB, and tadD code for fimbriae and other adhesins. ToxA and pmHAS encode toxin and hyaluronidase, respectively. Both exbB, exbD, tonB, hgbA, hgbB, fur, and tbpA encode iron acquisition proteins. SodA and sodC encode superoxide dismutase, while nanB and nanH encode sialidases. Additionally, the genes ompA, ompH, oma87, and plpB encode the outer membrane proteins [19][20][21][22][23]. During our study, we examined the virulence genes of the 19 P. multocida acquired from cats and the eight P. multocida isolated from humans. As shown in Image B (Figure 3), we found cats almost universally possess the virulence genes pfhA, nanH, and plpB, in contrast to humans. Additionally, all cat isolates expressed the virulence gene sodA, except pm98. The heatmap also demonstrated that the same MLST type of P. multocida isolated from cats or people had the same virulence profile. Furthermore, the virulence profile of strain pm17 showed that it carried all 24 virulence genes.

Discussion
Pasteurella multocida is found naturally in cats' oral cavities [6] and can transmit from cats to humans, resulting in infection; this phenomenon is not rare [12][13][14][15][16]24,25]. Due to the popularity of cat-themed cafes in China, close contact between customers and cats is going to promote the spread of P. multocida between these two groups, which may increase the frequency of human P. multocida infection cases. Further research is therefore required to analyze and explore these potentially infectious P. multocida strains.
The prevalence of P. multocida identified during our study (48.38%) was higher than that reported by Ferreira [7], who found 10.47% of positive samples in cats' gingival mucosa (20/191). The positive rate of P. multocida isolated from the oral mucosa of cats, according to Freshwater, was 89.97% (368/409) [26], which was higher than our finding. The differences might be explained by the different living conditions of cats. During our study, the cats were derived from cats housed in coffee shops in a home environment comparable to Freshwater [26], under similar circumstances, and with similar separation rates. The variation in P. multocida separation rates might also be influenced by the cats' own oral environment. The distribution of flora in the mouths of the cats might alter, depending on geography.
The virulence factors are important for P. multocida both in overcoming the host immune response and in boosting its survival in the host environment [27,28]. The pfhA gene, which encodes filamentous haemagglutinins, plays an important role in the early bacterial colonization of the upper respiratory tract [29]. Much research has shown that the P. multocida strains that were isolated from various hosts varied in their pfhA isolation rates. For example, compared to a far lower frequency of P. multocida pfhA isolation from cattle [29], the isolation rate of P. multocida pfhA from cats ranged from 0% to 23.8% [7,18,20]. Interestingly, we found that the P. multocida virulence gene pfhA was expressed in nearly all cats during the current study. The nanH gene is involved in epithelial surface colonization, similar to the function of pfhA, and its prevalence varies depending on host origin and geographical location [20,29]. During our study, the frequency of nanH was similar to pfhA, which may imply that they may act synergistically to promote bacterial colonization and survival in the oral cavity of cats. This could mean that people who visit these cat cafes and who get bitten by a cat are more likely to have P. multocida infection.
As for the treatment of P. multocida, antibiotics are currently recommended. Antibiotics, including penicillin, fluoroquinolone, and the second or third generation of cephalosporin, are commonly used to treat P. multocida infection [30]. Mu and Yang [14] dealt with the statistics and analysis of antibiotic-sensitive test results in the cases of the last three decades. Their research revealed that P. multocida had a very low rate of antibiotic resistance, and all the isolates were sensitive to ceftriaxone and fluoroquinolone. Based on our antibiotic-sensitive test results, the sensitive rates of ceftriaxone (30 µg) and fluoroquinolones (5 µg) were 80.00% and 62.22%, respectively. In particular, the resistance rate to erythromycin (15 µg) was found to be 100%. The data indicated an increasing prevalence of antibiotic-resistant P. multocida. As a result, the administration of antimicrobials for P. multocida infections should always be based on the results of antibiotic susceptibility testing.
In our study, we discovered that eight fluoroquinolone-resistant strains had mutations in the gyrA D87G gene. Fluoroquinolone prevents bacterial reproduction by inhibiting DNA gyrase (gyrA and gyrB) and topoisomerase IV [31]. Point mutations in the QRDR, which determines quinolone resistance, are closely linked to increased resistance to fluoroquinolone. A chromosomal point mutation in type IV topoisomerase (parC and parE) or DNA gyrase (gyrA and gyrB) is known as a QRDR mutation [32]. In our research, eight strains had an identical point mutation in gyrA. Additionally, fluoroquinolone insensitivity could also occur in strains without gyrA mutations. HmrM, for instance, belongs to the MATE family of multidrug efflux pumps. Based on earlier research, the hmrM gene majorly contributed to norfloxacin resistance in H. influenzae [33]. Thus, resistance to fluoroquinolone was considered to be multifactorial.
During our study, all isolates were resistant to erythromycin (15 µg). WGS analysis found that CRP, the regulator of the MdtEF-TolC efflux pump, was presented in all isolates. The MdtEF-TolC efflux pump is a member of the HAE-RND subfamily [34]. RND efflux pumps in Gram-negative bacteria can expel several kinds of antimicrobials from cells, which directly leads to multidrug resistance y [35]. It has also been discovered that MdtEF may extrude erythromycin under anaerobic conditions [34]. During our research, the wholegenome sequencing of P. multocida uncovered a variety of mechanisms that led to antibiotic resistance and demonstrated that P. multocida led to antibiotic resistance by modifying antibiotic target sites or by regulating the flow of antibiotics into or out of bacterial cells.

Sample Collection and Bacterial Isolate Identification
The samples were taken by using sterile swabs from the tonsils and oral cavities of 93 cats from four cat-themed cafes in different parts of Guangzhou, South China in 2022. All of the cafe cats in our study were in good health. They were energetic and had a voracious appetite. No respiratory or gastrointestinal illnesses existed, they had no fever, and they had not taken any antibiotics or other medications in the last three months. The cafes had enough room for them to live there. After taking the samples, the swabs were put in nutrient broth mediums and cultivated for 24 h at 35 • C in a 5 percent CO 2 incubator. Then, broth medium was inoculated onto blood agar, which was incubated for 18 h. All isolates were directly identified by using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS).

Antimicrobial Susceptibility Testing
Antimicrobial susceptibility was tested on all P. multocida strains that were collected. The disk-susceptibility testing method was used to determine the susceptibility profile as suggested by the Clinical and Laboratory Standards Institute (CLSI), and the breakpoint was interpreted according to the CLSI document (M45 3rd edition, 2016) [17]. The antimicrobial agents tested included amoxicillin-cavulanate (20/10 µg), ampicillin (10 µg), penicillin

Whole-Genome Sequencing
A total of 19 P. multocida strains that showed antimicrobial nonsusceptibility to more than two antibiotic classes were chosen for whole-genome sequencing (WGS). The total genomic DNA of P. multocida was extracted by using a Steady Pure Bacterial Genomic DNA Extraction Kit (Accurate Biotechnology, Hunan, China) and stored at −20 • C until use. The Illumina Novaseq6000 platform at Sangon Biotech (Shanghai, China) was chosen for WGS with a 150 bp paired-end strategy. After assembling the reads, unqualified contigs covering fewer than 10 lengths or lengths lower than 500 bp were removed from further study .

Genomic Analysis
In order to predict open reading frames (ORFs) and related functional genes, we employed Rapid Annotation using Subsystem Technology (RAST) (https://rast.nmpdr. org/rast.cgi, accessed on 2 July 2022) to annotate the whole nucleotide sequence of P. multocida. We used PmGT (http://vetinfo.hzau.edu.cn/PmGT/home.php, accessed on 19 July 2022) to identify virulence factors encoding genes of P. multocida strains based on the WGS data [36]. Public databases for molecular typing and microbial genome diversity (https://pubmlst.org/bigsdb?db=pubmlst_pmultocida_seqdef&page=sequenceQuery, accessed on 25 July 2022) were used to determine multilocus sequence typing (MLST) from the assembled genome. We predicted resistomes from nucleotide data based on homology by using the resistance gene identifier (RGI) of the comprehensive antibiotic resistance database (CARD) (https://card.mcmaster.ca/analyze/rg, accessed on 25 July 2022). For the purpose of choosing antimicrobial-resistance genes from CARD, we set the threshold at > 60% identity of the matching region. On all of the chosen P. multocida, a phylogenetic tree analysis was carried out by using MEGA 11.0 and the NDtree database (https://cge.food.dtu.dk/services/NDtree/, accessed 19 July 2022) from the Center for Genomic Epidemiology [37][38][39]. Additionally, we performed heatmaps of the genes encoding virulence factors and antibiotic resistance by using TBtools (version 1.09876) [40].

Conclusions
In conclusion, we presented antimicrobial susceptibility profiles and resistance genes, as well as virulence genes, in P. multocida in cafe cats in south China. Our study demon-strated that P. multocida strains isolated from cafe cats were resistant and nonsusceptible to the commonly used antibiotics erythromycin (15 µg) and fluoroquinolones (5 µg), respectively. In addition, we speculated that the high expression of relevant virulence genes in P. multocida from cafe cats may make it easier for the species to colonize and survive in cells. Therefore, more caution should be used when administering clinical treatment for P. multocida infection. Additionally, the relevant health authorities should be concerned about the emergence of P. multocida strains that are nonsusceptible to antibiotics and that express a high level of various virulence factors, and solutions such as limiting antibiotic abuse should be suggested in response. As for consumers, close contact with cafe cats should be followed by washing hands thoroughly, and dining in the same area as the cafe cats should be avoided.