Molecular Investigation of Klebsiella pneumoniae from Clinical Companion Animals in Beijing, China, 2017–2019

This work is aimed to elucidate the prevalence and characteristics of antimicrobial resistance, virulence, and molecular typing in Klebsiella pneumoniae from clinical companion animals in Beijing, China. In total, 105 K. pneumoniae (2.0%) isolates were recovered from 5359 samples (dogs, n = 3356; cats, n = 2003). All tested isolates exhibited high resistance to amoxicillin-clavulanate (74.3%). Moreover, resistance rates in dog isolates (2.1%) were significantly higher than in cat isolates (0.9%); however, the rate of multidrug-resistance (MDR) was 57.1% and the MDR prevalence in cats was significantly higher than dogs. Whole-genome sequencing demonstrated plasmids IncX4 and IncFIA (HI1)/FII(K) carried mcr-1 (n = 1) and mcr-8 (n = 1), but blaOXA-181 (n = 1) and blaNDM-5 (n = 4) were harbored in IncX3-type plasmids, and the above genes were in different isolates. The most prevalent sequence types (STs) in companion animals were ST1 (n = 9) and ST37 (n = 9). Compared to National Center for Biotechnology Information (NCBI) data on human K. pneumoniae, resistance genes blaCTX-M and blaTEM were more prevalent in human isolates; however, aac(6′)-Ib-cr and oqxAB showed a higher prevalence in companion animals. Hypermucoviscosity was reported in 9 (8.6%) isolates, whereas 64 isolates (61.0%) were hypervirulent K. pneumoniae (hvKP) via the Galleria mellonella. These findings validate the high risk of K. pneumonia and necessitate its relevant control in pet clinics.


Introduction
Klebsiella pneumoniae is an opportunistic pathogen that colonizes the skin, upper respiratory tract, and digestive tract of healthy asymptomatic subjects [1]. Also, it is a primary cause of diseases in neonates, elderly and immunocompromised humans and animals, including pneumonia, wound infections, urinary tract infections (UTIs), sepsis and meningitis [2]. K. pneumoniae is inherently resistant to penicillin; members of this population, in most cases, acquire resistance to multiple antibiotics. Thus, it is associated with resistance to important antimicrobial agents, thereby a significant threat to public health [3].
In recent years, K. pneumoniae has rapidly become a multidrug-resistant (MDR) pathogen. It develops resistance to third-generation cephalosporins, fluoroquinolones, and aminoglycosides. Of concern, K. pneumoniae has increasingly become resistant to carbapenems by acquiring carbapenemases [4]. The World Health Organization recognizes extended-spectrum β-lactam (ESBL)-producing and carbapenem-resistant K. pneumoniae

Antibiotic Resistance Genes and Virulence Genes
The mobile colistin resistance gene mcr-1 (Kp141) and mcr-8 (Kp24) were detected in K. pneumoniae from tracheal lavage and urine from different cats. Carbapenem resistance

Antibiotic Resistance Genes and Virulence Genes
The mobile colistin resistance gene mcr-1 (Kp141) and mcr-8 (Kp24) were detected in K. pneumoniae from tracheal lavage and urine from different cats. Carbapenem resistance genes bla OXA-181 and bla NDM-5 were harbored in one (Kp3), and four (Kp79, Kp84, Kp165, Kp181) isolates, respectively. Additionally, we analyzed the sequence read archive (SRA) sequences (n = 46) of human K. pneumoniae (HKp) from National Center for Biotechnology Information (NCBI) to demonstrate the similarities, differences, and relevance of the molecular characteristics of K. pneumoniae from humans and companion animals. Resistance genes, mcr and bla NDM , were not present in the genomes of K. pneumoniae isolated from humans; other variants were also absent. bla KPC (95.7%, n = 44/46) was harbored in human isolates, whereas none was found in companion animals. Resistance genes bla CTX-M and bla TEM were more prevalent among K. pneumoniae isolates from humans, compared to companion animals (p < 0.01). Meanwhile, isolates from companion animals showed higher rates of aac(6 )Ib-cr and oqxAB than those from humans (p < 0.01). Moreover, the prevalence of bla CTX-M and rmtB from cats was higher than in dogs (p < 0.05); notably, bla SHV  Similarly, virulence-associated genes enterotoxins (entA/B/E/S), ferrienterochelin receptor (fepA/B/C), fimbriae (fimA/E), outer membrane protein (ompA), and common pili (ecpA/B/C/D/E/R) were harbored by all isolates from humans and companion animals, as revealed using abricate ( Figure 2). The ybtA/E/P/Q/T/U/X, iucA/B, rmpA2, fepD, fyuA, irp1/2, and iutA from human isolates were significantly more prevalent (p < 0.05) than those from companion animals; however, virulence genes of dogs and cats exhibited no difference. Interestingly, the ST23-hvKP (n = 5) isolates harbored ybt, clb, iro, iuc, ent, rmpA, and rmpA2 virulence genes, with the least resistance genes. Moreover, ST37-hvKP isolates harbored common virulence factors ent, fep, fim, ompA, and ecp but simultaneously took along more resistance genes ( Figure 2).

Characterization of Plasmids
After assembling backbone sequences, all contigs and gaps were identified by whole-genome analysis. Among the Inc-type plasmids, IncFIB (67.6%, n = 71/105) was prevalent in K. pneumoniae isolates of companion animals, whereas human isolates were covered by IncFII (100%, n = 46/46). Of these, IncFIA was significantly more prevalent in companion animals than that of humans (p < 0.01), but the prevalence of IncFII and IncR

Discussion
K. pneumoniae is an important host and transmission carrier of clinically important antimicrobial resistance genes in humans and animals [16]. Typically, humans have close contact with pets; therefore, there is a close association between the health of both. However, there is currently a dearth of clinical research on K. pneumoniae from compan- The internal ring is the reference sequence of pNDM_MGR194 (46253bp, accession No. KF220657), pABC239-OXA-181 (51479bp), pECGD-8-33 (33307bp), and p18-29mcr-8.2 (91072bp) respectively, and the outside rings are isolates from this study, which are similar to the reference plasmid.

Discussion
K. pneumoniae is an important host and transmission carrier of clinically important antimicrobial resistance genes in humans and animals [16]. Typically, humans have close contact with pets; therefore, there is a close association between the health of both. However, there is currently a dearth of clinical research on K. pneumoniae from companion animals in China. Herein, we report the prevalence of antibiotic resistance, virulence, molecular typing, and phylogroups in K. pneumoniae from companion animals in Beijing, China. Previous findings demonstrated that the overall prevalence of recovered (2.0%) K. pneumoniae was slightly lower than 3.53% in Italy [17]. K. pneumoniae is the primary pathogen of UTIs, and is usually associated with resistance to the most significant antibiotics [15]. Urine (53.7%) accounted for the largest proportion (1.3%) of samples, from which the largest number of K. pneumoniae isolates (n = 37) was isolated in this study, this was similar with a previous study in Japan [18].
Among the K. pneumoniae isolates from companion animals, the resistance rate in dogs was significantly higher than in cat isolates, but showed no difference in South Korea [19]. This discordance in results could be explained by different types and quantities of tested drugs. An overall MDR rate of 57.1% was slightly higher than that previously reported in Singapore (50.0%) [20]. Otherwise, the MDR prevalence in cats was significantly higher than in dogs in this study (p < 0.05). As a result, this aggravates the threat to human health, as owners are more likely to exhibit intimate behavior with companion animals, increasing the probability of mutual transmission. Colistin had been used as an animal feed contributing to the prevalence of mcr in China; similarly, studies have reported wide prevalence in Vietnam and other South Asian countries. However, the prevalence of mcr-1 in food animals in Europe and America is low [21], which may be attributed to the fact that countries in these regions have not approved colistin use as an antibacterial growthpromoting agent. Elsewhere, a previous study demonstrated low prevalence (<1%) of mcr-1 of K pneumoniae from humans [22]. Moreover, mcr-1 in K. pneumoniae from companion animals; thus, the pet food industry was speculated to be a source of mcr-1 [23]. Since the first description of mcr-8, with IncFII-type plasmid as the carrier, it has been widely disseminated among K. pneumoniae isolates of livestock origin [6]. In the current study, mcr-1 (n = 1) and mcr-8 (n = 1) were carried by plasmids IncX4 and IncFIA (HI1)/FII(K) of K. pneumoniae. To the best of our knowledge, we present the first report on the isolation of the two isolates from tracheal lavage and urine-derived from different cats. Of concern, having found five CRKP, which also appeared in other countries was considered a severe situation [19,20]. In this study, carbapenem resistance genes, bla OXA-181 and bla NDM-5 , were harbored in IncX3-type plasmids of different CRKP but no bla OXA-48 and bla KPC was found in companion animals, while bla OXA-48 was prevalent in other countries [24]. Resistance genes, bla CTX-M and bla TEM , of ESBL, showed a lower prevalence among companion animals, but the prevalence was higher than Japan (34.8%) [18], Italy (21.4%) [17], and other European countries (11.2%) [25]. In addition, CTX-M-genotypes were diverse in different species and countries. Herein, bla CTX-M-65 and bla CTX-M-55 predominated in the isolates of humans and companion animals, a phenomenon similar to findings from a previous study of pets in South Korea [19] and China [26]. Moreover, K. pneumoniae of type CTX-15 was prevalent in companion animals in Japan [18]. These findings demonstrate that the mutations of K. pneumoniae resistance genes are not limited to specific hosts or regions, and highlights the necessity of coordinated control in One Health. Furthermore, the genes aac(6 )Ib-cr and oqxAB of quinolone from companion animals showed higher resistance rates than those from human isolates; the rates were similar to those reported previously [18,27]. Otherwise, the most prevalent STs were ST1 and ST37, which concur with a previous report in China [26], but contrasts from reports in companion animals from Portugal [15] and Japan [18].
Based on the current understanding, virulence factors are encoded by several paragenes and can further increase the severity and/or pathogenicity of K. pneumoniae infection. Researches proved that any three of four siderophore systems (ent, ybt, iuc, and iro) could enhance virulence in murine models [3]. The ybt and iuc of human isolates were significantly more prevalent than that of companion animals. However, virulence factor-encoding genes of dogs and cats exhibited no difference in the current study. Previous reports had indicated that hvKP is mainly associated with ST23 and CRKP primarily belonging to ST11, which are considered the two major clinically important pathogens in China [28]. In this study, hvKP isolates (61.0%) from companion animals were less than the previously reported rates (76.4%) in humans [12]. Unfortunately, five ST23 isolates harbored all siderophore systems, identified as hvKP by G. mellonella model, but contained the least resistance genes. In companion animals, the four CR-hvKP were ST1 (n = 3) and ST16 (n = 1), whereas the CR-hvKP isolates were common genetic types such as ST11 and ST23 [14]. Hypermucoviscosity may be the most dominant virulence factor of K. pneumoniae, but its genetic basis and pathogenic factors are not direct. Due to the presence of one or two of the para-regulatory genes rmpA or rmpA2, this phenotype is usually associated with the overproduction of capsules [29]. Meanwhile, six of nine hypermucoviscous isolates were hvKP, but only one hypermucoviscous isolate in companion animals showed the MDR phenotype. Additionally, MDR-hvKP accounted for 68.3% of MDR isolates, 64.1% of hvKP, and 39.0% of total K. pneumonia isolates. The above findings demonstrate the high-risk of K. pneumonia, creating a huge treatment limitation in pet clinics.
In conclusion, the present study did a large-scale investigation of antimicrobial resistance and molecular genetic analysis in K. pneumoniae from clinical companion animals in Beijing, China. A high prevalence of MDR and hypervirulent K. pneumoniae isolates were found from dogs and cats. The wide distribution of amoxicillin-clavulanic and thirdgeneration cephalosporins in veterinary hospitals may contribute to the ESBL resistance in these isolates. The presence of mcr, bla OXA181 and bla NDM in K. pneumoniae demonstrates that the pathogen is a potential reservoir of colistin and carbapenem resistance genes in pet clinics. Meanwhile, the emergence of MDR-hvKP and epidemic clones elevates the risks of veterinarians; however, the predominant clones of CRKP are scarce in human-related ST clones. These findings emphasize the importance of managing K. pneumonia comorbidities and scientifically conducting antimicrobial susceptibility tests for more accurate treatments. This would reduce the spread of such high-risk clonal lineages to ensure the safety of companion animal practitioners and public health.

Samples Collection and Bacterial Characterization
All samples of companion animals were collected aseptically from the Veterinary Teaching Hospital of China Agricultural University (VTH-CAU), Lpet Veterinary Diagnostic Center (LVDC-Beijing), and North China (Tianjin) Testing Center, between July 2017 and October 2019. Sampling was conducted following the principles of the Beijing Municipality Review of Welfare and Ethics of Laboratory Animals and approved by the China Agricultural University Animal Ethics Committee document (No. AW01017102-2). K. pneumoniae isolates were isolated using 5% sheep blood agar and MacConkey Inositol Adonitol Agar medium (HopeBio, Qingdao, China) containing 100 mg/L carbenicillin following aerobic incubation at 37 • C overnight. The DNA of individual clones with the red centre was extracted by TIANamp Bacteria DNA Kit (Tiangen, Beijing, China) with the protocol as stipulated by the manufacturer. Subsequently, the DNA was used as templates for polymerase chain reaction (PCR) amplification of 16S rDNA gene sequencing (16SrRNA-F: AGAGTTTGATCCTGGCTCAG, 16SrRNA-R: ACGGCTACCTTGTTACGACTT) and conditions consisted 95 • C (10 min), 30 cycles of [95 • C (30 s), 55 • C (30 s), 72 • C (90 s)], and 72 • C (10 min) as previously described [30]. Amplicons were sequenced to reveal bacterial genus using the BLAST algorithm.

Mucoviscosity Assay
Because mucoviscous cells remain in suspension, whilst non-mucoid cells could form pellets after centrifugation, measurement of the turbidity after low-speed centrifugation can serve as an indicator of hypermucoviscosity. In short, bacterial isolates were grown in Lysogeny broth (LB) broth at 37 • C after 6 h incubation with shaking and then centrifuged at 1000× g for 5 min. The optical density at 600 nm (OD600) values of the supernatant were determined and measured. Mucoviscosity isolates were more difficult to pellet, so the supernatants have higher absorbance readings.

Galleria Mellonella Virulence Assay
Here, we tested the in vivo virulence using the G. mellonella infection model to indicate hypervirulence, as previously described [31]. In total, 10 randomly selected G. mellonella larvae approximately 250 to 350 mg for each isolate were purchased from Huiyude Biotech Company, Tianjin, China. Bacteria cells were cultured in a mid-log-phase and pelleted via centrifugation at 3500 rpm, washed twice, and resuspended in 0.01 M phosphatebuffered saline (PBS, pH6.5). Larvae were injected with 10 µL bacterial suspension (with 10 6 colony-forming units, CFU) via the rear left proleg using a micro-sample syringe. PBS and ATCC43816 were used as negative and positive control groups, respectively. After injection, larvae were placed in 90 mm Petri-dishes and kept at 37 • C in the dark. Insects were considered dead when they did not respond to physical stimuli. We monitored death at 6 h intervals during 72 h. Experiments were performed in triplicate.

Whole-Genome Sequencing and Molecular Analysis
To extract the genomic DNA of the isolates, the TIANamp Bacteria DNA Kit was used following the manufacturer's manuals. Indexed DNA libraries were constructed using the KAPA Hyper Prep Kit Illumina platforms (Roche, Basel, Switzerland) following the instruction, then sequenced on the Illumina Hiseq X Ten platform via the 150-bp paired-end strategy (Annoroad, Beijing, China). The draft genomes were assembled using SPAdes (version 3.9.0) [32]. All whole-genome sequencing data for this work are deposited in the GenBank and under BioProject accession no. PRJNA684769. Plasmid types, antibiotic resistance genes, and virulence genes were identified using abricate (https: //github.com/tseemann/abricate, accessed on 13 August 2020), whereas the multilocus sequence typing (MLST) was obtained using the SRST2 toolkit (version 0.2.0) [33]. Capsule serotype (KL) and O-antigen (O) were analyzed using Kleborate (https://github.com/ katholt/Kleborate, accessed on 16 September 2020). All draft genomes were used for core-genome alignments, after which we constructed a phylogenetic tree using parsnp in the Harvest package (version 1.1.2) [34]. The tree was visualized using the online tool Interactive Tree of Life (iTOL, http://itol.embl.de/, accessed on 19 September 2020) with the corresponding features of each isolate. The minimum spanning tree for all STs was generated by BioNumerics version 7.6 (Applied Maths, Sint-Martens-Latem, Belgium) using the BURST algorithm between different backgrounds. To compare the genetic context in the different plasmids, BLAST Ring Image Generator (BRIG) was applied [35]. Additionally, we analyzed the SRA of K. pneumoniae derived from humans in the NCBI database, which was collected in Beijing between July 2017 and October 2019.

Statistical Analysis
Statistical significance was determined using Chi-square (χ 2 ) and Fisher's exact test in SPSS Statistics (version 22, IBM Corporation, Armonk, NY, USA). The level of significance was set at p < 0.05.