Spread of the florfenicol resistance floR gene among clinical Klebsiella pneumoniae isolates in China

Background Florfenicol is a derivative of chloramphenicol that is used only for the treatment of animal diseases. A key resistance gene for florfenicol, floR, can spread among bacteria of the same and different species or genera through horizontal gene transfer. To analyze the potential transmission of resistance genes between animal and human pathogens, we investigated floR in Klebsiella pneumoniae isolates from patient samples. floR in human pathogens may originate from animal pathogens and would reflect the risk to human health of using antimicrobial agents in animals. Methods PCR was used to identify floR-positive strains. The floR genes were cloned, and the minimum inhibitory concentrations (MICs) were determined to assess the relative resistance levels of the genes and strains. Sequencing and comparative genomics methods were used to analyze floR gene-related sequence structure as well as the molecular mechanism of resistance dissemination. Results Of the strains evaluated, 20.42% (67/328) were resistant to florfenicol, and 86.96% (20/23) of the floR-positive strains demonstrated high resistance to florfenicol with MICs ≥512 μg/mL. Conjugation experiments showed that transferrable plasmids carried the floR gene in three isolates. Sequencing analysis of a plasmid approximately 125 kb in size (pKP18–125) indicated that the floR gene was flanked by multiple copies of mobile genetic elements. Comparative genomics analysis of a 9-kb transposon-like fragment of pKP18–125 showed that an approximately 2-kb sequence encoding lysR-floR-virD2 was conserved in the majority (79.01%, 83/105) of floR sequences collected from NCBI nucleotide database. Interestingly, the most similar sequence was a 7-kb fragment of plasmid pEC012 from an Escherichia coli strain isolated from a chicken. Conclusions Identified on a transferable plasmid in the human pathogen K. pneumoniae, the floR gene may be disseminated through horizontal gene transfer from animal pathogens. Studies on the molecular mechanism of resistance gene dissemination in different bacterial species of animal origin could provide useful information for preventing or controlling the spread of resistance between animal and human pathogens. Electronic supplementary material The online version of this article (10.1186/s13756-018-0415-0) contains supplementary material, which is available to authorized users.


Background
Florfenicol, which is only used to treat animal infections, is a derivative of chloramphenicol that is active against chloramphenicol-resistant isolates [1]. Resistance to chloramphenicol occurs mainly through the production of inactivating enzymes called chloramphenicol acetyl transferases (CATs) [2] and chloramphenicol exporters, such as CmlA [3]. Over the past decade, most reports have demonstrated that the bacteria causing animal respiratory diseases show high resistance levels to chloramphenicol but are susceptible to florfenicol [4]. However, the resistance levels and number of bacteria that are resistant to florfenicol have increased due to the widespread use of florfenicol in the treatment of animal diseases [5][6][7]. A study on 1001 bacterial isolates showed that the resistance rates for trimethoprim/sulfamethoxazole and tetracycline were 3.0% and 14.7% in Actinobacillus pleuropneumoniae and 6.0% and 81.8% in S. suis, respectively, while the resistance rate for florfenicol was < 1% for all strains [8]. Other reports have cited different resistance rates. In Australia, 2.0% and 6.0% of A. pleuropneumoniae and Pasteurella multocida strains isolated from pig respiratory infections were resistant to florfenicol, respectively [9]. The resistance rate of E. coli strains from canine urinary tract infections to florfenicol was higher than that of other pathogens: 31.6% (36/114) [5].
The first florfenicol resistance gene, pp-flo (renamed flo), was identified on a plasmid in the fish pathogen Photobacterium damselae subsp. piscicida in 1996 [10]. The floR gene is closely related (97% identity) to the flo gene [11], and their proteins share 47% amino acid sequence identity with the CmlA protein. The floR gene was first reported in 1999 on the chromosome of the worldwide epidemic strain Salmonella enterica serovar Typhimurium DT104 [11]. The primary source of human DT104 infections was thought to be animal populations, with both direct contact and foodborne modes of transmission [12]. The IncC plasmid R55, which was initially described to be capable of conferring non-enzymatic chloramphenicol resistance in the 1970s, was then identified in Klebsiella pneumoniae [13]. Currently, nine florfenicol resistance genes [floR, floRv, floSt, fexA, fexB, pexA, cfr, optrA and estDL136] have been identified. With the exception of cfr and estDL136, which encode a 23S rRNA methyltransferase and a hydrolase, respectively, all of the genes encode exporters [14][15][16][17][18]. The floR gene and its analogs have mainly been identified in gram-negative bacteria, whereas the other resistance genes have mainly been detected in gram-positive bacteria [15][16][17].
Similar to other resistance genes, floR has been identified on both chromosomes and plasmids and has often been associated with mobile genetic elements and genomic islands [19,20]. Mobile genetic elements enable translocation of the floR gene between DNA molecules, such as chromosomes and plasmids. A plasmid carrying the floR gene can spread among bacteria of the same and different species or genera via conjugation or transformation, thereby disseminating resistance [21]. Bacteria generally obtain multiple resistance genes through the horizontal transfer of plasmids carrying resistance genes [22].
K. pneumoniae, which is a member of the Enterobacteriaceae, is an opportunistic pathogen for both animals and humans. This bacterium is pervasive in the natural environment and benignly colonizes the gastrointestinal tracts of healthy humans and animals. However, the bacterium is also capable of causing a wide range of diseases in humans and different animal species [23]. K. pneumoniae strains are a common cause of health-care associated infections including pneumonia, urinary tract infections (UTIs), and bloodstream infections for critically ill and immunocompromised patients. These strains also infect healthy people in community settings, causing severe infections including pyogenic liver abscess, endophthalmitis, and meningitis [24]. For example, in animals, K. pneumoniae strains are well documented to cause mastitis and wounds in cattle [25]; endometritis, cystitis, and liver abscess in horses; tracheitis and wounds in birds; cystitis, phlebitis and otitis externa in dogs; and cystitis in cats [26]. K. pneumoniae has also been associated with classical foodborne disease outbreaks [19]. Notably, the prevalence of antibiotic resistance is increasing among Enterobacteriaceae, including K. pneumoniae [23,27]. In this study, we used multiple genetic approaches to investigate the floR gene in K. pneumoniae isolates of human origin and to further demonstrate the potential transmission of this resistance determinant between animal and human pathogens.

Bacterial strains
The 328 non-duplicate K. pneumoniae strains used in this work were isolated from patient samples at the First Affiliated Hospital of Wenzhou Medical University in Wenzhou, China, from 2010 to 2014. This sample set included all K. pneumoniae strains collected during this time frame. Among these isolates, 55 were isolated in 2010, 109 in 2011, 55 in 2013 and 109 in 2014. The strains were identified using the Vitek-60 microorganism auto-analysis system (BioMerieux Corporate, Craponne, France).

PCR amplification of the floR gene
Total genomic DNA was extracted from the 328 isolates using AxyPrep Bacterial Genomic DNA Miniprep kits (Axygen Scientific, Union City, CA, USA). Template DNA was screened for the floR gene using a PCR method. According to the conserved floR gene-related regions of the K. pneumoniae genome obtained from a pool of strains mainly from this work [28], floR gene screening primers were designed and named P SCR-F and P SCR-R-A/G , which correspond to the cm101 and cm115 primer sequences, respectively [29]. The sequence of the forward primer P SCR-F was 5'-TTTGGTCCGCTCTC AGAC-3′. Two variants of the reverse primer were used due to a single nucleotide polymorphism (A/G) identified in the region where the primer was designed: 5'-CGAGAAGAAGACGAAGAAG-3′ (P SCR-R-A ) and 5'-CGAGAAGAAGGCGAAGAAG-3′ (P SCR-R-G ). These primers yield a product 496 bp in length. PCR amplification was carried out under the following conditions: an initial denaturation of 5 min at 95°C; 35 cycles of denaturation (94°C for 45 s), annealing (58°C for 45 s), and extension (72°C for 90 s); and a final extension step at 72°C for 10 min [29]. The PCR products were purified using a MinElute PCR Purification kit (QIAGEN China, Shanghai, China) and sequenced by Sanger sequencing (in this work, all the PCR products and cloned fragments were sequenced by Sanger sequencing). The nucleotide sequences were analyzed and compared using the BLAST program (http://www.ncbi.nlm.nih.gov/ BLAST).

Antimicrobial susceptibility testing
Antimicrobial susceptibility testing performed via the agar dilution method in accordance with the guidelines of the Clinical and Laboratory Standards Institute (CLSI document M100-S27, 2017) was used to determine the minimum inhibitory concentrations (MICs) [30]. The resistance threshold values (32 μg/mL) for both chloramphenicol and florfenicol were chosen according to the guidelines of CLSI document M100-S27 (2017) [30] and a publication for E. coli [31], respectively. E. coli ATCC 25922 was used as a quality control strain.

Pulsed-field gel electrophoresis (PFGE)
To assess the epidemiology of clinical isolates with floR genes, genomic DNA from K. pneumoniae isolates harboring floR genes was prepared for PFGE typing and digested with 40 U of Xba I (Takara, Dalian, China). The protocol and the Xba I restriction patterns of genomic DNA from the isolates were analyzed and interpreted according to initial criteria [32]. The Bio-Rad Quantity One program was used to analyze the PFGE results, and a minimum spanning tree was constructed using a categorical coefficient with unweighted pair group method with arithmetic mean (UPGMA) clustering [33].

Plasmid DNA extraction and sequencing
For plasmid (pKP18-125) sequencing, the transconjugant KP18/EC600 was incubated overnight in 5 mL of Luria-Bertani broth at 37°C for approximately 16 h to an optimum optical density (OD 600 ) of 1.5 ± 0.2. The plasmid was then extracted using the alkaline lysis method as described previously [34]. Plasmid DNA was sequenced via Illumina HiSeq-2000 and Pacific Bioscience sequencing methods at the Beijing Genomics Institute (Beijing, China). Reads derived from the HiSeq-2000 sequencing were initially assembled de novo using SOAPdenovo software to obtain contigs of the plasmid. Pacific Bioscience sequencing reads of approximately 10-20 kb in length were mapped onto the primary assembly to scaffold the contigs. The gaps were filled either by remapping the short reads from HiSeq-2000 sequencing or by PCR product sequencing of the gaps. Glimmer software was used to predict protein-coding genes with potential open reading frames (ORF) > 150 bp [35]. Gview was used to construct basic plasmid features [36]. BLASTX was used to annotate the predicted protein-coding genes against the non-redundant protein database using an e-value threshold of 1e-5.

Collection and processing of floR gene-related sequences
In addition to the pKP18-125 sequence, other sequences containing the floR gene were obtained from the NCBI nucleotide database using floR, pp-flo, flo, cmlA-like, floRv and floSt as key terms. The resulting sequences were filtered, and only sequences containing a complete floR gene more than 9 kb in length (with approximately 4 kb both upstream and downstream of the floR gene) were retained. Multiple sequence alignments were performed using mafft with the 9-kb floR gene-related fragment of pKP18-125 (KY082186) as a reference [32], and the sequences were clustered with an identity of 80%. The sequence with greatest similarity to the other sequences in each cluster was chosen as a candidate for orthologous analysis. Orthologous groups of genes from the candidate sequences were identified using BLASTP and InParanoid [37]. The sequence retrieval, statistical analyses and other bioinformatics tools used in this study were accomplished using Python and Biopython scripts [38].

Cloning experiments
To identify and clone floR genes, we PCR amplified a fragment including the floR ORF sequence and its upstream 354-bp potential promotor region from strains positive for the floR gene. A set of PCR primers (P ORF-F and P ORF-R ) was designed using the K. pneumoniae plasmid pR55 sequence (JQ010984.1) as a reference. The primer sequences of P ORF-F and P ORF-R were 5'-GTCG AGAAATCCCATGAGTTCA-3′ and 5'-CAGACAGGA TACCGACATTCAC-3′, respectively. The PCR products were eluted from agarose gels and ligated into the pMD18 vector (TaKaRa, Dalian, China). Each recombinant plasmid (pMD18-floR) was transformed into E. coli JM109 using the calcium chloride method, after which the bacterial colonies were grown on Luria-Bertani agar plates supplemented with ampicillin (100 μg/mL). The recombinant plasmids were isolated and digested with restriction enzymes to confirm insertion of a floR fragment of approximately 1600 bp in length. Each cloned floR fragment was analyzed by Sanger sequencing from a purified transformant and was further compared to the reference floR gene (JQ010984.1) using the BLASTN program.

Conjugation experiments
To examine the conjugation potential of resistance gene-harboring pKP18-125, we used rifampin-resistant EC600 as a recipient strain in a biparental mating, which was performed overnight at 37°C on sterile nitrocellulose filters as previously described [39]. The transconjugants were selected on Mueller-Hinton agar plates containing 1200 μg/mL of rifampin and 16 μg/mL of florfenicol [40]. Plasmid DNA was extracted from transconjugants, and the presence of the floR gene was verified by PCR and PCR product sequencing. The plasmid (pKP18-125) of one transconjugant (KP18/EC600) was sequenced in full to verify that the floR gene was encoded on this transferable plasmid.  Table 2).

A transferable plasmid carrying the floR gene
The results of the conjugation experiments for the 23 floR-positive clinical isolates showed that plasmids in   (Fig. 1). KP5 was isolated from a sputum sample of a male patient in the Department of Neurosurgery in March 2014, while KP6 came from a sputum sample of a female patient in the intensive care unit (ICU) in the Department of Brain Surgery in April 2014.
Structure and comparative genomics analysis of floR gene-related regions pKP18-125 was 125,329 bp in length. Annotation determined that the plasmid carried one replicon belonging to incompatibility group FII (IncFII) and harbored 164 coding sequences (CDs). The plasmid genome can be divided into 4 regions according to the functions of the ORFs as follows: the variable region, the conjugation region, the transfer leading region and the replication region. The variable region is approximately 40 kb in length and encodes 42 ORFs, including approximately 20 genes related to drug resistance, 13 recombination-related genes or structures (i.e., integrase and transposase genes and insertion sequences [ISs]) and 9 genes of unknown function.
According to the structure of the mobile genetic elements, this region could be roughly divided into six units, including one class 1 integron and five transposons. The floR gene was located in a transposon-like fragment approximately 9 kb in length (accession number: KY082186) consisting of a conserved gene cluster of lysR-floR-virD2, 5 tnp units and two direct repeats (DRs). In this work, we mainly analyzed the structural characteristics of the 9-kb floR gene-related transposon-like fragment (Fig. 2).
Overall, a total of 105 DNA sequences of approximately 9 kb in length with the floR gene in their center were retrieved from all floR gene-containing sequences in the NCBI nucleotide database. Of these sequences, 45 were from complete or partial bacterial chromosomes, and 60 were from plasmid sequences. Through a multiple sequence alignment, 27 clusters with identities greater than 80% were obtained. According to the similarity of the core sequences adjacent to the floR gene, the sequences of these 27 clusters were orthologously analyzed and further clustered into 4 groups (G1-G4), with each group containing 2 subgroups (e.g., G1a and G1b). Group 1 consisted of only 2 sequences (KY082186 and KT282968) that shared approximately 7 kb in common ( Fig. 1 and Table S1).
Eight representative sequences from the eight subgroups (one from each subgroup) are illustrated in Fig. 2 with their accession numbers. Sequences similar to the representative sequences are shown in Additional file 1: Table S1. The results of this orthologous analysis revealed that an approximately 2-kb sequence encoding lysR-floR-virD2 was conserved and present in the majority of the sequences (79.0%, 83/105). Many of these sequences also shared the same upstream DR and complete or truncated downstream tnp unit (Fig. 1). The 9-kb floR gene-related transposon-like fragment of pKP18-125 in this study showed highest similarity to a 7-kb fragment from the plasmid pEC012 (KT282968). Interestingly, although pKP18-125 was isolated from a K. pneumoniae strain from a patient in South China, pEC012 was found in an E. coli isolate from a chicken in North China [41].

Discussion
In this study, we found that among all the clinical K. pneumoniae isolates detected, 20.42% (67/328) were resistant to florfenicol, of which 7.01% (23/328) carried the floR gene, but 13.41% (44/328) were free of the floR gene. A similar report demonstrated a floR gene positivity rate of only 21.8% (26/119) among 119 florfenicol-resistant gram-negative bacilli from several freshwater Chilean salmon farms [42]. Our MIC results for the 328 strains demonstrated that the floR gene played a key role in the resistance of these bacteria to florfenicol. The floR-positive strains had a much higher resistance rate (23/23, 100%) and much higher MIC values for florfenicol (22/23, 95.65% with MIC values ≥512 μg/mL) than the floR-negative strains, which had a resistance rate of 14.43% (44/305) with only 1.64% (5/305) of the strains having MIC values ≥512 μg/mL. At present, of the nine florfenicol resistance genes, the floR gene is the only known florfenicol resistance gene that has been identified in K. pneumoniae strains of either human or animal origin [43]. Five genes (fexA, fexB, pexA, optrA and cfr) were mainly identified in gram-positive bacteria [15][16][17]. The cfr gene has also been occasionally identified in E. coli or Proteus vulgaris [44,45] and fexA and pexA were once identified in E. coli [44]. The other three genes have only been identified in certain gram-negative bacteria (floRv in Stenotrophomonas maltophilia [46], floSt in Salmonella [47] and estDL136 in E.coli [44]). We hypothesize that other mechanisms, such as exporters and enzymes, in addition to the known florfenicol resistance genes, may also be responsible for florfenicol resistance in gram-negative bacteria including K. pneumoniae.
The floR genes were located on both chromosomes and plasmids amidst various mobile genetic elements, indicating that horizontal transfer of the floR gene occurred among bacteria of different species. The floR gene was identified first on the chromosome of S. typhimurium DT104 (Salmonella typhimurium DT104) and then on a plasmid of E. coli isolate BN10660 [48] and was also identified on the IncC plasmid R55 harbored by K. pneumoniae [13] and on other sources [17,43]. In S. typhimurium DT104, the floR gene was included in a 12.5-kb region with multiple resistance genes. The tetR and tetA tetracycline resistance genes were located downstream of the floR gene and were flanked by two integrons. One integron contained an aadA2 gene and an incomplete sulI resistance gene, and the other harbored a β-lactamase gene and a complete sulI gene [49]. In pKP18-125, the downstream region was a class 1 integron that contained 5 resistance genes (acc(6′), arr2, etc.) and was different from the 12.5-kb region of the S. typhimurium DT104 chromosome. Interestingly, the sequence most similar to the floR-containing fragment on pKP18-125 from a clinical Klebsiella pneumoniae isolate was located on pEC012 (KT282968), a plasmid from an E. coli strain isolated from a chicken [50]. This finding suggests that horizontal transfer of the floR-containing fragment occurred between bacteria of animal and human origins.
Our PFGE analysis revealed that two floR-positive strains (KP5 and KP6) had similar PFGE profiles. They were isolated from the same sample type (sputum) but were found in different hospitalized patients during different time periods. Some K. pneumoniae strains carrying resistance genes were previously reported to have caused outbreaks in European countries, indicating the potential risk of the spread of resistance genes through bacterial outbreaks, especially those caused by bacteria with resistance plasmids [51]. Although the relationship between the two strains carrying floR is still in question, effort should be made to avoid any pathogen outbreaks in hospital environments.

Conclusions
Our study demonstrated that 20.42% (67/328) of the clinical K. pneumoniae isolates were resistant to florfenicol, but only 7.01% (23/328) carried the floR gene. The floR gene was related to a transposon-like sequence and located on a conjugative plasmid. The most similar sequence to the floR-containing fragment on pKP18-125 was that a fragment on pEC012 in an E. coli strain isolated from a chicken. This finding indicates that resistance genes in animal pathogens might be disseminated to human pathogens. The dissemination of resistance genes from animals to humans reflects the risk to human health of antimicrobial agent use in animals. In addition, these results highlight the critical need to consistently implement effective strategies to prevent transmission and infection. Combating antibiotic-resistant bacteria supports patient care, agriculture, economic growth and national security.