Comparative Analysis of Chloramphenicol-Resistant Enterococcus faecalis Isolated from Dairy Companies in Korea

Although chloramphenicol is currently banned from use in livestock, other phenicols, such as florfenicol and thiamphenicol, have been used for the treatment of bacterial infections in domestic cattle in Korea. This study compares the characteristics of chloramphenicol-resistant Enterococcus faecalis isolated from the bulk tank milk of four major dairy companies in Korea. Although the distribution of multidrug resistance patterns showed no significant differences between the four companies, 85 chloramphenicol-resistant Enterococcus faecalis isolates showed a significantly high number of resistances against five or six antimicrobial classes (37.6%, respectively) (p < 0.05). When analyzing the distribution of phenicol resistance genes, 31 (36.5%) isolates only carried the catA gene, and two (6.3%) isolates from company A only carried the cfr gene. No isolates carried the catB or fexA genes. Regarding the distribution of other resistance genes, both the tetL and tetM (45.9%), ermB (82.4%), and both aac(6″)-Ie-aph(2″)-la and ant(6′)-Ia genes (30.6%) showed a high prevalence, and the optrA and poxtA genes were observed separately, each in only two (2.4%) isolates. Our results confirm that the dissemination of chloramphenicol-resistant Enterococcus faecalis and some antimicrobial resistance genes show significant differences between dairy companies. Therefore, our results support that each dairy company should undertake effective surveillance programs to better understand and minimize the emergence of resistance on a multidisciplinary level.


Introduction
Chloramphenicol, a broad-spectrum antimicrobial in human and veterinary medicine in use since the 1950s [1], inhibits translation in bacteria, by inhibiting peptidyl transferase activity, which is mediated by binding to several proteins in the 50S ribosomal subunit [2]. However, since chloramphenicol has been shown to cause serious bone marrow suppression and fetal aplastic anemia [1,3], its use in humans and animals has been subsequently limited. It is currently banned from use in food-producing animals in many countries and regions, including the European Union [1], China [4], the United States [5], and Korea [6].
Enterococci are known as a genus of gut-equivalent microorganisms commonly found in animals and humans, but considered to be environmental mastitis-causing pathogens that can enter into milk and milk products via unhygienic food production and handling conditions [7][8][9]. Moreover, enterococci have been recognized as notoriously opportunistic pathogens that frequently acquire antimicrobial resistance determinants and potential virulence factors [10,11]. Resistance to chloramphenicol is mainly caused by the production of inactivating chloramphenicol acetyltransferase (CAT) [1], the genes of which are widely disseminated on plasmids and capable of contributing to multidrug resistance (MDR) by conjugative transfer [12,13]. Therefore, the occurrence of chloramphenicol-resistant Enterococcus faecalis related MDR from dairy products and farms is often reported [14][15][16][17][18]. This study compares the phenotypic and genotypic characteristics of chloramphenicolresistant E. faecalis isolated from the bulk tank milk of four major dairy companies in Korea to provide an insight into the potential emerging problems of Enterococcus spp. in milk.

Bacterial Isolation
A total of 1584 batches of bulk tank milk were aseptically collected from 396 farms affiliated with four dairy companies in Korea in the summer and winter seasons. The isolation and identification of Enterococcus spp. was performed following standard microbiological protocols published by the Ministry of Food and Drug Safety (Korea) (2019) [19]. Briefly, a one mL milk sample was cultured in nine mL of buffered peptone water (BPW; BD Biosciences, San Jose, CA, USA). Pre-enriched BPW was then mixed with Enterococcosel broth (BD Biosciences, San Jose, CA, USA) at a 1:10 ratio. After incubation at 37 • C for 18-24 h, each medium was streaked onto Enterococcosel agar (BD Biosciences, San Jose, CA, USA). Confirmation of E. faecalis was performed via polymerase chain reaction (PCR), with the primer specifically targeted to the ddl1 gene, as described previously [20]. As a result, a total of 301 E. faecalis isolates were tested in this study.

Antimicrobial Susceptibility Testing
Antimicrobial susceptibility analysis was performed by broth microdilution according to the Clinical and Laboratory Standards Institute guidelines (2018) [21], using the commercially available Sensititre ® panel KRVP2F (TREK Diagnostic Systems, West Sussex, UK), in accordance with the manufacturer's instructions. The antimicrobial classes tested were aminoglycosides (gentamicin, kanamycin, and streptomycin), glycopeptides (vancomycin), ionophores (salinomycin), lipopeptides (daptomycin), macrolides (erythromycin, and tylosin tartrate), oxazolidinones (linezolid), penicillins (ampicillin), phenicols (chloramphenicol and florfenicol), quinolones (ciprofloxacin), and tetracyclines (tetracycline and tigecycline). E. faecalis ATCC 29212 was used as a quality control. If isolates from the same farm showed the same antimicrobial susceptibility pattern, only one isolate was randomly selected. MDR was defined as acquired non-susceptibility to at least one agent in three or more antimicrobial classes [22].

Statistical Analysis
The SPSS 25 statistical package (IBM corp., Armonk, NY, USA) was used for performing statistical analyses. Pearson's chi-squared test and Fisher's exact test with Bonferroni correction were conducted to analyze the differences between the dairy companies, as previously reported [35]. Differences were considered significant at p < 0.05.

Prevalence of Chloramphenicol-Resistant E. faecalis
Among the 301 E. faecalis isolates, 85 (28.2%) chloramphenicol-resistant E. faecalis isolates were found among the four dairy companies (Table 1). Although company D showed the highest prevalence of E. faecalis, the prevalence of chloramphenicol-resistant E. faecalis was significantly higher in isolates from company A (61.5%) than company D (12.9%) (p < 0.05).

Distribution of Antimicrobial Resistance Genes
The distribution of resistance genes in 85 chloramphenicol-resistant E. faecalis isolates is shown in Table 3. In the distribution of phenicol resistance genes, although 52 (61.1%) isolates expressed no resistance genes, 31 (36.5%) isolates carried only the catA gene without a significant difference between dairy companies. Two (6.3%) isolates from company A only carried the cfr gene, and no isolates carried both the catB and fexA genes. In the tetracycline resistance genes, the distribution of the tetM gene alone (43.5%), and both tetL and tetM genes (45.9%) was common but showed a significant difference between dairy companies (p < 0.05). In the macrolide resistance genes, the prevalence of the ermB gene (82.4%) was the highest, and in the oxazolidinone resistance genes, four (4.7%) isolates only carried the optrA or poxtA genes. In the distribution of aminoglycoside resistance genes, the ant(6 )-Ia gene alone (25.9%) and both the aac(6")-Ie-aph(2")-la and ant(6 )-Ia genes (30.6%) were common, without a significant difference between dairy companies. Table 3. Antimicrobial resistance genes in 85 chloramphenicol-resistant Enterococcus faecalis from bulk tank milk in dairy companies.

Discussion
The prevalence of chloramphenicol-resistant enterococci in dairy products has been reported in many countries, including Switzerland (45.9%), Poland (32.91%), South Africa (13%), and Turkey (10.7%) [9,15,18,36]. In this study, the total prevalence of chloramphenicolresistant E. faecalis isolates (28.2%) was also shown similar to other countries, but there were significant differences in the distribution of chloramphenicol-resistant E. faecalis between the four dairy companies studied (12.9%-61.5%). In Korea, although chloramphenicol is no longer used in cattle, other phenicols, such as florfenicol and thiamphenicol, have been continuously used for the treatment of bacterial infections in domestic cattle [19]. In particular, florfenicol is commonly recommended for the treatment of bacterial pneumonia and associated respiratory infections caused by Haemophilus somnus, Mannheimia (Pasteurella) haemolytica, and Pasteurella multocida in cattle [37]. Florfenicol inhibits protein synthesis by binding to ribosomal subunits of susceptible bacteria and shares antimicrobial binding sites with chloramphenicol [1]. Because antimicrobial usage generally results in antimicrobial resistance [5], the varying usage of florfenicol between the dairy companies studied could be attributed to the differences seen in the distribution of chloramphenicol resistance.
In this study, 83 (97.6%) chloramphenicol-resistant E. faecalis isolates demonstrated MDR, especially against five or six classes, which showed a significantly higher prevalence. Chloramphenicol resistance mechanisms in bacteria comprise a reduction in membrane permeability, mutations in 23S rRNA, and proliferation of CATs [1]. However, another mechanism of chloramphenicol inactivation is performed by efflux pumps, which are regulated by translation attenuation [1,38]. Efflux pumps can contribute to the internal environment by removing toxins or antimicrobial agents [10,39]; therefore, the efflux of chloramphenicol could simultaneously confer resistance to this antimicrobial as well as others [10,39].
As previously described, the most frequently encountered mechanism of resistance is enzymatic inactivation by acetylation of chloramphenicol via CATs, encoded by the catA and catB genes, which are widespread among Gram-positive and Gram-negative bacteria [1,12]. Although the catB gene has been found in Gram-negative bacteria, the catA gene is commonly found in Gram-positive bacteria [1]. Jamet et al. [16] reported that 16 of 20 (80%) chloramphenicol-resistant E. faecalis isolates from French cheese carried the catA gene. Hummel et al. [14] also reported that all chloramphenicol-resistant E. faecalis isolates from milk, whey, and cheese in their study carried the CAT gene. However, only 36.5% of chloramphenicol-resistant E. faecalis carried the catA gene in this study, indicating that other MDR-related mechanisms, such as efflux pumps, may be involved in chloramphenicol resistance to E. faecalis in Korea.
Moreover, chloramphenicol resistance can result from changing the binding site of chloramphenicol in cells through a mutation in 23S rRNA [1], facilitated by the phenicol resistance gene cfr [40]. The cfr gene is also known as the MDR gene [40]; therefore, several researchers have highlighted increasing concerns with regard to this gene in public health [31,[40][41][42]. In addition, the cfr gene is thought to be linked to the spread of linezolid resistance [41]. Although Elghaieb et al. [43] and Ahmed et al. [44] reported no cfr gene in E. faecalis isolates taken from dairy milk, two cfr-positive, chloramphenicol-resistant E. faecalis isolates were detected in this study. However, these two isolates showed no resistance to linezolid.
Although the distribution of the catA gene showed no significant differences between dairy companies, some resistance genes showed a significant difference. The distribution of the tetM gene alone and both tetL and tetM genes, which are related to tetracycline resistance, is common, but showed significant differences between the four companies. Miller et al. [41] reported that the tetL gene encodes efflux pumps, which are plasmid-borne determinants. Mobile resistance genes carried via plasmids can be easily transferred by conjugation, which allows the sharing of genetic information such as antimicrobial and virulence genes [45]. In this study, isolates from company B showed the highest prevalence of both tetL and tetM genes (75.0%), and as a result dissemination of tetracycline-resistant pathogens and MDR, including chloramphenicol may be expected to increase further in company B.
In the distribution of macrolide resistance genes, prevalence of the ermB gene (82.4%) was the highest, but also showed significant differences between dairy companies (p < 0.05)isolates from company A and C showed 90.6% and 95.2% ermB gene prevalence, respec-tively, but only 56.3% of isolates from company D carried the ermB gene. Moreover, resistance against macrolides was also shown to be higher in isolates from company A (93.8%) and C (95.2%) than company D (62.5%). The mechanisms of macrolide resistance in enterococci are described as two types: (i) inducible resistance associated with methylation of 23S rRNA by a methylation enzyme, and (ii) resistance associated with an efflux pump [41,46]. Although the prevalence of erythromycin-resistant E. faecalis from dairy products was reported differently between countries including Korea (85.4%), France (70.9%), Switzerland (60.1%), and Poland (18.5%), most erythromycin-resistant E. faecalis isolates were positive for the ermB gene [15,16,47,48]. The ermB gene, which is considered to be the most widespread macrolide resistance gene among enterococci in livestock and food, induces target modification via the action of methylase, and either conjugative plasmids such as pAMβ1, pRE25, and pUW1965, or transposons, such as Tn917, Tn1545, Tn5384, and Tn5385, which carry the ermB gene [14,15]. This study suggests that there is a need to further investigate the possibility of mobile resistance gene transfer by sequencing plasmids and detecting genes related to transposon families [49,50].
In this study, eight (9.4%) chloramphenicol-resistant E. faecalis isolates from three dairy companies showed co-resistance to linezolid. Surprisingly, four (4.7%) isolates also showed co-resistance to florfenicol and harbored the linezolid resistance genes optrA or poxtA. Linezolid is considered to be a last resort treatment for infections caused by MDR gram-positive pathogens, including vancomycin-resistant Enterococcus spp., methicillin-resistant Staphylococcus spp., and Streptococcus pneumoniae [51]. Recently, the ABC-F transporter gene, poxtA, has been the focus of increasing resistance to phenicols and oxazolidinones [32,52,53]. In addition, a novel gene, optrA, confers resistance against linezolid, tedizolid, and phenicols by encoding an ATP-binding cassette transporter [30]. So far, linezolid resistance genes in E. faecalis from milk and dairy products have not been reported in Korea. Therefore, as previously reported, there is a need for continuous surveillance to monitor the emergence of linezolid-resistant E. faecalis in dairy products, as a matter of public health [43].
In this study, among the four dairy companies studied, most chloramphenicol-resistant E. faecalis isolates carried virulence genes including ace, asa1, cad1, efaA, and gelE, although no significant difference was demonstrated between the companies. The presence of virulence genes may contribute to the severity of pathogenesis and accelerate the transfer of antimicrobial resistance genes, which is of public health concern [18].

Conclusions
In a comparative analysis of E. faecalis isolated from four major dairy companies in Korea, the dissemination of chloramphenicol-resistant E. faecalis and some antimicrobial resistance genes showed a significant difference between companies, although the prevalence of MDR showed no significant difference. Therefore, our results support that each dairy company should undertake effective surveillance programs to better understand and minimize the emergence of resistance on a multidisciplinary level.

Data Availability Statement:
In regard to the data presented in this study, we included it in the article.

Conflicts of Interest:
The authors declare no conflict of interest.