The Prevalence of Colistin Resistant Strains and Antibiotic Resistance Gene Profiles in Funan River, China

Anthropogenic activities near urban rivers may have significantly increased the acquisition and dissemination of antibiotic resistance. In this study, we investigated the prevalence of colistin resistant strains in the Funan River in Chengdu, China. A total of 18 mcr-1-positive isolates (17 Escherichia coli and 1 Enterobacter cloacae) and 6 mcr-3-positive isolates (2 Aeromonas veronii and 4 Aeromonas hydrophila) were detected, while mcr-2, mcr-4 and mcr-5 genes were not detected in any isolates. To further explore the overall antibiotic resistance in the Funan River, water samples were assayed for the presence of 15 antibiotic resistance genes (ARGs) and class 1 integrons gene (intI1). Nine genes, sul1, sul2, intI1, aac(6′)-Ib-cr, blaCTX-M, tetM, ermB, qnrS, and aph(3′)-IIIa were found at high frequencies (70–100%) of the water samples. It is worth noting that mcr-1, blaKPC, blaNDM and vanA genes were also found in water samples, the genes that have been rarely reported in natural river systems. The absolute abundance of selected antibiotic resistance genes [sul1, aac(6′)-Ib-cr, ermB, blaCTX-M, mcr-1, and tetM] ranged from 0 to 6.0 (log10 GC/mL) in water samples, as determined by quantitative polymerase chain reaction (qPCR). The sul1, aac(6′)-Ib-cr, and ermB genes exhibited the highest absolute abundances, with 5.8, 5.8, and 6.0 log10 GC/mL, respectively. The absolute abundances of six antibiotic resistance genes were highest near a residential sewage outlet. The findings indicated that the discharge of resident sewage might contribute to the dissemination of antibiotic resistant genes in this urban river. The observed high levels of these genes reflect the serious degree of antibiotic resistant pollution in the Funan River, which might present a threat to public health.


INTRODUCTION
Multi-drug resistant (MDR) Gram-negative pathogens are resistant to almost all antibiotics, including cephalosporins, quinolones, aminoglycosides and carbapenems, making treatment difficult. Colistin is considered the last line of defense against MDR Gram-negative pathogens, playing an important role in the treatment of severe bacterial infections (Zavascki et al., 2007). Unfortunately, the recent emergence of plasmid-mediated colistin resistance genes in carbapenemresistant Enterobacteriaceae presents a serious new threat to human health. The plasmid-mediated colistin resistance gene mcr-1 was first discovered Liu et al. (2016). Soon afterward, another mobile phosphoethanolamine transferase gene, termed mcr-2, was discovered in porcine and bovine Escherichia coli isolates in Belgium . Recently, Yin et al. (2017) discovered a novel mcr subtype, mcr-3, encoded on an IncI2 plasmid in an E. coli isolated from a pig in China. The mcr-4 and mcr-5 genes were detected in Europe almost simultaneously (Borowiak et al., 2017;Carattoli et al., 2017). Although there have been numerous reports of colistin resistance genes in animals and humans, fewer studies have focused on mcr-bearing isolates from aquatic environments.
Due to the continual release of antibiotic residues and antibiotic resistant bacteria (ARB) into the environment from hospitals, livestock facilities, and sewage treatment plants (STP), antibiotic resistant genes (ARGs) are regarded as environmental contaminants (Pruden et al., 2006;Zurfluh et al., 2017). The occurrence and dissemination of antibiotic resistance in pathogenic and zoonotic bacteria pose a potential threat to human health (Rosenberg Goldstein et al., 2012;Neyra et al., 2014). Moreover, an increasing number of bacteria are resistant to multiple antibiotics, and are able to transfer their resistant determinants among different bacterial species and genera in aquatic environments (Akinbowale et al., 2006). Urban rivers may provide an ideal setting for the acquisition and dissemination of antibiotic resistance because they are frequently impacted by anthropogenic activities. Although antibiotic resistance is a major and developing public health concern, the surveillance of this phenomenon in urban rivers is remarkably limited.
The Funan River, a major urban river in Chengdu used for agricultural activities (e.g., irrigation and cultivation) as well as recreational activities (e.g., swimming and fishing), was used as the model in this study to analyze the magnitude of antibiotic resistance in urban rivers.

Sampling of River Water
To investigate the prevalence of colistin resistant strains, 30 water samples (2 L) were collected from the Funan River near densely populated areas in September 2017. To further explore the antibiotic resistance of bacteria throughout the Funan River, 10 water samples (2 L) were collected from representative locations along the river (Figure 1). The representative locations included river intersections, streams near parks, and sewage outlets near residential areas, the hospital, and the municipal wastewater treatment plant (WWTP). The site near the residential sewage outlet is designated RWW and the sample near the municipal wastewater treatment plant is designated WWTP. Sites P1, P2, and P3 are close to various parks and HWW1 and HWW2 are close to the hospital sewage outlet. Site RI is located adjacent to the intersection of a tributary and the mainstream of the river. Sites UWP and DWP are upstream and downstream of Wetland Park, respectively. Water samples were collected from each site, immediately placed on ice, and transported to the laboratory within 4 h. The samples were then maintained at 4 • C until investigation.

Bacterial Isolation
A total of 30 water samples were concentrated by vacuum filtration through 0.22 µm filter membranes. The membranes were washed and the collected material was suspended in 10 ml of sterile PBS. A volume of 1 ml thereof was added to 9 ml of Brain Heart Infusion (BHI) broth with polymyxin B at a final concentration of 4 µg/mL. After incubation at 37 • C overnight, 100 µl culture samples were streaked onto MacConkey agar plates. Fifty colonies were picked from each MacConkey agar plates and subsequently grown in BHI broth with 4 µg/mL polymyxin B for 18-24 h. Isolates were screened for the presence of mcr-1, mcr-2, mcr-3, mcr-4, and mcr-5 by PCR. Next, mcr-positive isolates were purified by subculturing. The mcr-positive isolates were identified using 16S rRNA gene sequencing and the BD Phoenix-100 Automated Microbiology System (BD Diagnostic Systems, Sparks, NV, United States).

Total DNA Extraction and Detection of ARGs
To further explore the extent of antibiotic resistance throughout the Funan River, water samples were collected from 10 locations (Figure 1). Total DNA was extracted using the Water DNA kit (OMEGA, United States) from the bacteria sample trapped by 0.22 µm pore filter (2 L samples). Standard PCR performed as listed in Table 1 was used to detect 15 ARGs (sul1, sul2, bla CTX−M , bla VIM , bla KPC , bla NDM , qnrS, aac(6 )-Ib-cr, vanA, mecA, ermB, ermF, tetM, aph(3 )-IIIa and mcr-1) and the class 1 integron gene (intI1). Negative and positive controls were used for each set of PCR primers. PCR amplification reactions were conducted in 20 µl volumes containing 1× PCR Master Mix (Tsingke, China), 1.0 µl template DNA, and 0.5 µM of each primer. After amplification, 5 µl samples of the PCR products were loaded on a 1.0% agarose gel containing GoldView, and separated electrophoretically in 1 × TAE buffer at 120 V for 20 min and visualized.

Quantitative Polymerase Chain Reaction
To compare the abundance of ARGs for different sampling sites, the gene copy numbers of the sul1, aac(6 )-Ib-cr, ermB, bla CTX−M , and tetM genes were quantified using qPCR assays. These genes confer resistance to five major classes of antibiotics: sulphonamides, aminoglycosides, macrolides, β-lactams, and tetracyclines. The levels of mcr-1 and 16S rRNA genes were also quantified. To quantitate the amounts of these genes, the levels were compared to the levels in standard samples prepared from plasmids containing these specific genes, as described previously (Chen and Zhang, 2013). The standard samples were diluted to yield a series of 10-fold concentrations and were subsequently used to generate qPCR standard curves. The R 2 values were higher than 0.990 for all standard curves. The 20 µl qPCR mixtures contained 10 µL of SYBR premix Ex Taq TM (TaKaRa, Dalian, China), 0.5 µM of each forward and reverse primer, and 1 µl of template DNA. The final volume was adjusted to  20 µl by addition of DNase-free water. The IQ TM 5 real-time PCR system was employed for amplification and quantification, using the following protocol: 30 s at 95 • C, 40 cycles of 5 s at 95 • C, 30 s at the annealing temperature, and extension for another 30 s at 72 • C. For detection, simultaneous fluorescence signal was scanned at 72 • C, followed by a melt curve stage with temperature ramping from 65 to 95 • C. Details of the qPCR primers of the target genes and the annealing temperatures are given in Table 2. The method design was adopted from prior research (Thornton and Basu, 2011). The copy numbers of the selected ARGs were normalized against the 16S rRNA gene copy number. Therefore, the copy number unit is described as copies/16S.

Statistical Analysis
Statistical analysis was performed using SPSS 17.0 (IBM, United States). One-Way ANOVA was employed to analyze the results and values of P < 0.05 were considered statistically significant.

RESULTS AND DISCUSSION
The Prevalence of mcr-Positive Isolates in the Funan River The screening of 1500 isolates for mcr yielded a total of 24 mcrpositive isolates. They included 18 mcr-1 positive isolates (17 Escherichia coli and 1 Enterobacter cloacae) and 6 mcr-3 positive isolates (2 Aeromonas veronii and 4 Aeromonas hydrophila). mcr-2, mcr-4, or mcr-5 were not observed in any of the isolates. Many reports have described the presence in mcr-1 in animal-and human-derived Enterobacteriaceae isolates isolated worldwide (Du et al., 2016;Liu et al., 2016;Malhotra-Kumar et al., 2016;Shen et al., 2016), but only two previous studies identified mcr-1 in waterborne Enterobacteriaceae. One study reported detection of the mcr-1 gene in 1 out of 74 Enterobacteriaceae isolated from 21 rivers and lakes in Switzerland that produced extended spectrum β-lactamases (ESBLs) (Zurfuh et al., 2016). In a separate study, similar to our results, Zhou et al. (2017) isolated 23 mcr-1-positive isolates from environmental water sources in Hangzhou, indicating that mcr-1-carrying Enterobacteriaceae may be common in lakes and rivers in China. Data addressing the prevalence of mcr-3 is limited. Recently, a novel mcr variant, mcr-3, was first discovered on an IncI2 plasmid from a strain of E. coli isolated from a pig in China . Since then, mcr-3-positive strains have been identified in humans and food (Ling et al., 2017;Liu L. et al., 2017). Worryingly, mcr-3 has been detected on the chromosome of Aeromonas veronii, and these chromosomally encoded mcr-3 determinants can become plasmid-bound and transferable (Cabello et al., 2017;Ling et al., 2017). Recently, Shen et al. (2018a) presented evidence that mcr determinants originated from aquatic environments, including mcr-3 harboring Aeromonas spp. Because Aeromonas species are prevalent in aquatic environments, the occurrence of colistin resistant isolates in urban rivers is of great concern as these strains may contribute to the potential dissemination of mcr determinants.

Prevalence of Antibiotic Resistance Genes in the Funan River
In this study, the prevalence of ARGs in water samples was investigated by sampling various sites along the Funan River. The sul1, qnrS, tetM, and intI1 genes were detected in samples from all 10 sampling sites (100%). Additionally, aac(6 )-Ib-cr, sul2, aph(3 )-IIIa, ermB, and bla CTX−M were detected at high rates of 90%, 90%, 90%, 80% and 70%, respectively. Many studies have reported the presence of these genes in aquatic environments (Hu et al., 2008;D'Costa et al., 2011;van Hoek et al., 2011;Lin et al., 2015;Makowska et al., 2016). Interestingly, the aph(3 )-IIIa gene has rarely been reported in river water microorganisms, but has been reported in clinical specimens (Tuhina et al., 2016). The detection of the aph(3 )-IIIa gene was high in this study, suggesting contamination of the Funan River with resistant bacteria carrying the aph(3 )-IIIa gene.
Genes conferring resistance to the last line of antibiotics, including mcr-1, bla NDM , bla KPC and vanA genes, were detected at rates of 30%, 20%, 10%, and 10%, respectively. bla VIM was not detected at any site. The mcr-1 gene was detected in 30% of samples, suggesting the Funan River could act as a reservoir for the mcr-1 gene. The bla NDM , bla KPC and vanA genes were detected near the WWTP (Figure 1). Although mcr-1 is found frequently in human and animal settings, there is only limited data for urban rivers (Marathe et al., 2017;Ovejero et al., 2017;Yang et al., 2017). Similarly, Marathe et al. detected bla NDM and bla KPC genes in the sediments of an Indian river (Marathe et al., 2017). Although a bla VIM positive carbapenem-resistant strain was isolated from a river in Switzerland (Zurfluh et al., 2013), here is a lack of data on bla VIM in the non-clinical environment. The vanA gene is associated with vancomycin resistance and has been found in wastewater biofilms and in drinking water biofilms in Mainz (Schwartz et al., 2003). Although these genes have rarely been identified in natural aquatic environments, given the dangerous infections that can arise from ARB (and which subsequently create intractable challenges for clinical treatment), further observation of the prevalence of these genes in aquatic environments is required.

Abundance of ARGs
Concerning the absolute abundance of ARGs in the Funan River, ARGs were detected at levels that ranged from 0 to 6.0 log 10 GC/mL (Figure 2). The sul1, aac(6 )-Ib-cr, and ermB genes were the dominant ARGs in the Funan River with mean absolute abundances of 4.8, 4.1, and 3.4 log 10 GC/mL, respectively. The sul1 gene exhibited the most prominent average abundance in water samples. Previous studies reported that sul1 is abundant in numerous water areas, including the Tordera River Basin (Proia et al., 2016) and the Haihe River (Luo et al., 2010). Although the mcr-1 gene was not detected in water samples at some sites, three sites (RWW, HWW1, and HWW2) displayed 2.0-2.7 log 10 GC/mL. Notably, the highest detected level of mcr-1 (2.7 log 10 GC/mL) was higher than that in previous reports about the Haihe river (2.6 log 10 GC/mL) . The absence of mcr in some samples may indicate that no mcr-1 positive strains were present in the water samples or that the levels of mcr-1 were below the detection limit. Site RWW is located near the residential sewage outlet, suggesting the presence of mcr-1 was related to human activity. Consistently, mcr-1 was detected at HWW1 and HWW2, adjacent to the hospital sewage outlets, suggesting the spread of mcr-1 from hospitals to urban river, although colistin is not used widely in human medicine. The mcr-1 abundance at RWW (2.7 log 10 GC/mL) was slightly higher than that at HWW1 (2.6 log 10 GC/mL) and at HWW2 (2.3 log 10 GC/mL). Similarly, the prevalence of mcr-1-positve E. coli from healthy individuals (0.7-6.2%) is higher than the prevalence for inpatients (0.4-2.9%) (Shen et al., 2018b). It is striking that mcr is the only gene that was absent from sites other than RWW and HWW. The reasons for high rate of fecal carriage of mcr in humans in China may reflect the rapid emergence of plasmid-encoded mcr-1 within many MDR E. coli carried by humans and also be related to the significant diversity and genetic flexibility of MGEs harboring mcr-1 (Zhong et al., 2018).
At RWW, RI, and WWTP, the absolute abundances of certain ARGs (sul1, aac(6 )-Ib-cr, and ermB) were significantly higher than those at other sampling sites (P < 0.05). At P3 and DWP, the absolute abundances of most ARGs were significantly lower than the levels detected at the other sites (P < 0.05). RWW was associated with the highest absolute abundance of the six ARGs (mcr-1, sul1, aac(6 )-Ib-cr, ermB, bla CTX−M , and tetM) (Figure 2). Samples near the wastewater treatment plant (WWTP) and densely populated areas exhibited a relatively greater content of resistant genes. Wastewater discharge may contribute to the spread of ARGs into the environment, thereby affecting the bacterial communities of the receiving river (Marti et al., 2013;Xu et al., 2015). Our results indicate that human activities influence the dissemination of resistance genes in the Funan River. Remarkably, the absolute abundances of most ARGs were low at the DWP sampling point, located downstream of the wetland park. This is consistent with a decrease in the ARGs levels of the effluents from a constructed wetland with a free surface flow (Liu et al., 2014).
As shown in Figure 2, the relative abundances of each ARG are only partly correlated with their absolute abundance. That is, although the absolute abundances of most ARGs at RWW, RI and WWTP were relatively high, their relative abundances were comparatively low. These differences may be related to the differences in the proportion of resistant bacteria to total bacteria at each site (Tao et al., 2014).

CONCLUSION
This study describes 18 mcr-1-positive strains and 6 mcr-3positive strains isolated from the Funan River, of which 87.5% were found to be MDR. The sul1, sul2, intI1, aac(6 )-Ibcr, bla CTX−M , tetM, ermB, qnrS and aph(3 )-IIIa genes were abundant in the Funan River. Interestingly, the mcr-1, bla KPC , bla NDM , and vanA genes were detected, although these four resistance genes have rarely been found in natural river systems. Notably, the mcr-1 gene was detected at a rate of 30%. Our results suggest urban activities may increase the prevalence of antibiotic resistance genes and demonstrate the current presence of drugresistance pollution in the Funan River. The processes by which the dissemination of ARGs occurs in urban rivers should be the focus of future studies.

AUTHOR CONTRIBUTIONS
AZ designed the study. HT, DL, XX, and PL carried out the sampling work. HT, YY, XT, and JG performed the experiments. AZ, HT, RX, LK, and CL analyzed the data. AZ, HT, YL, and HW drafted the manuscript. All authors have read and approved the final manuscript.