It is still in search of a clinical relevance. In our study, 10 cfr-positive isolates were detected from 839 E. coli isolated between 2010 and 2012, and all the positive isolates were from swine farms where florfenicol is extensively used to prevent and cure diseases caused by a variety of bacterial pathogens in China. PFGE analysis of the 10 cfr-positive E. coli revealed that these strains were genetically divergent. Most of the strains form a distinct phylogenetic group in the dendrogram based on genetic similarity, which suggested that the spread of the cfr gene in E. coli was not due to clonal dissemination but horizontal transfer. S1 nuclease PFGE and hybridization 3,4,5-Trimethoxyphenylacetic acid showed that the cfr gene was located on an 
approximately 30 kb plasmid in all but two cfrpositive strains. Strains 2ZX7S and 8ZG1D harbored cfr-carrying plasmids of different sizes. The failure of the conjugation assays using cfr-positive strains suggests that the cfr gene was carried by nonconjugative plasmids in these strains. Recently cfr was identified on a conjugative plasmid in E. coli, which may further accelerate the dissemination of the cfr gene among different Gram-negative bacteria. Although we were not successful in obtaining transformants with a single plasmid after several attempts, RFLP and hybridization profiles of the plasmids showed that 7 cfr-carrying plasmids yielded the same-sized cfr-harboring fragment of,18 kb after Amikacin hydrate digestion with EcoRI, and all the 7 plasmids are,30 kb in size with the same genetic environments confirmed by PCR mapping and inverse PCR, implying that the 7 plasmids are likely identical and originated from the same source. Interestingly, compared with the 9 strains in which a single cfrcarrying plasmid was harbored, strain 8ZG1D and the transformant 8ZG1D-21 have two cfr-positive plasmids. Considering the presence of IS26 flanking the cfr gene, IS-mediated recombination may account for the transfer of cfr between plasmids in the strain. Interestingly, the result of antimicrobial susceptibility testing showed that 8ZG1D-21 was moderately resistant to florfenicol and sensitive to chloramphenicol, suggesting that the cfr gene may confer low-level resistance to chloramphenicols in E. coli. Four different genetic environments were detected surrounding the cfr gene, all of which have two copies of IS26 of the same orientation flanking the cfr gene. Previous studies have suggested that insertion element IS26 can mediate the transfer of cfr gene. In our study, structural comparison of the genetic environments showed that part of the segment between IS26 shares high homology with plasmid pEC-01 from E. coli LYP-C-BCTb11 and chromosomal fragment from P. vulgaris PV-01, further suggesting that IS26 may have played an important role in the transfer of the cfr gene. Furthermore, inverse PCR performed on all of the cfr-positive strains can obtain an amplicon, and subsequent sequencing analysis showed that the pair of intact IS26 flanking the cfr gene can loop out the intervening sequence through homologous recombination, which can further accelerate the transfer of cfr gene. To conclude, we present the first study on the prevalence of the cfr gene in E. coli from food producing animals. The identified cfrpositive E. coli strains were limited to pigs, coinciding with the extensive use of florfenicol for swine production. PFGE analysis showed that the cfr-positive E. coli strains were genetically diverse; however, plasmid and genetic environment analysis suggested that most of these strains harbored the same cfr-carrying plasmid.