Identi ﬁ cation and characterization of vancomycin-resistant Staphylococcus aureus in hospital wastewaters: evidence of horizontal spread of antimicrobial resistance

Antibiotic resistance has become a major threat to human health around the world, but its spread through the aquatic environment has been often overlooked. This study aimed to determine the occurrence of vancomycin-resistant Staphylococcus aureus in hospital wastewaters and their transmission into public water bodies in Kerala, India. A total of 113 S. aureus were isolated from three hospital ef ﬂ uents in Kerala, India. Standard disc diffusion and the strip method were used for antibiotic susceptibility testing and minimum inhibitory concentration detection. Plasmid-mediated vancomycin resistance was con ﬁ rmed by plasmid curing and conjugation; resistant genes were detected by the polymerase chain reaction (PCR). Nearly 76% of S. aureus isolates were resistant to β -lactams, chloramphenicol, macrolides, aminoglycosides, and glycopeptide class of antibiotics. Among the vancomycin-resistant Staphylococcus aureus (VRSA) isolates, the prevalence rates of van A and van B resistance-encoding genes were 46.5 and 59.3%, respectively. Through the broth mating method, vanA gene was successfully transferred from VRSA donor to vancomycin-sensitive S. aureus . The study strongly indicates the contamination of water bodies with antibiotic-resistant bacteria from hospital discharges, their dissemination and possible transfer to microbes in the aquatic environment, posing a serious threat for public health.


INTRODUCTION
Staphylococcus aureus is one of the most common microorganisms frequently associated with various diseases, ranging from mild infections of the skin to life-threatening endocarditis, chronic osteomyelitis, pneumonia, and bacteraemia (Lowy 1998;Murray 2005). During the mid-20th century, the introduction and use of antibiotics such as penicillin and methicillin proved successful against S. aureus infections. However, the bacterium quickly acquired resistance to these antibiotics posing an enormous challenge to both veterinary and human health clinicians (Brouillette & Malouin 2005). Treatment for this bacterium is a concern with the emergence and spread of penicillin-resistant S. aureus and in turn methicillin-resistant S. aureus (MRSA). MRSA has become one of the most common causes of hospital-associated and community-acquired infections, and the global spread of MRSA is a matter of great concern (Grundmann et al. 2006). Besides, community-based MRSA has recently emerged as a potential threat, causing infections in healthy individuals with no risk factors associated with healthcare (Kluytmans-Vandenbergh & Kluytmans 2006). The antibiotic glycopeptide vancomycin has proven effective in the treatment of serious MRSA infections (McGuinness et al. 2017). Moreover, in the last 20 years, S. aureus clinical isolates with reduced sensitivity to vancomycin and less frequently, with maximum resistance to vancomycin, have emerged (Hidayat et al. 2006).

Study area
Three prominent hospitals (H1, H2, and H3) located near water bodies in three districts of Kerala, India (Ernakulam, Kollam, and Kannur) were selected for the study. The study was conducted during August-December 2018-2019. All the hospitals mentioned in the study have a moving bed biofilm reactor (MBBR) system for sewage treatment. The hospitals released their sewage effluents into water bodies that were used for various purposes including inland fishing activities. There were no pharmaceutical industries adjacent to the water bodies.

Sample collection and processing
Sediment and water samples were collected from the sampling points, following standard procedures. Water samples were taken from the outlet pipes of the hospitals in sterilized amber-coloured 500 ml glass bottles labelled as H1, H2, and H3 and transported to the laboratory on ice within 2 h of collection. Using an Ekman Dredge sediment sampler, sediment samples were also collected from the same three sites. From each site, approximately 100-200 g of the sediment sample was collected in 50 ml sterilized tubes. Ten-fold serial dilutions of surface water in 0.9% sterile saline (NaCl) solution were made. The entire sediment sample was dissolved in 100 ml of 0.9% normal saline, followed by subsequent 10-fold dilutions. The diluted samples were spread plated in duplicate plates of Trypticase soy agar (Himedia, India) and incubated at 37°C for 24 h. Colonies were enumerated after incubation, and colonies with various morphologies were randomly selected and transferred at least three times in order to ensure their purity. Traditional biochemical tests were carried out to identify S. aureus isolates, including Gram staining, catalase, oxidase, coagulase, and mannitol utilization tests (Holt et al. 1994).

Multiple antibiotic resistance index study
The multiple antibiotic resistance (MAR) index for each isolate was determined by dividing the number of antibiotics to which the organism was resistant by the total number of antibiotics tested (Kaplan et al. 2005). A general indication of the possible source of an organism is provided by the MAR index.

Plasmid curing and isolation of plasmid
Plasmid curing was performed by growing the isolates that showed high resistance to vancomycin in nutrient broth (Himedia, India) containing 0.1 mg/ml acridine orange for 24 h (Chaudhari & Bajaj 2015). A loopful of broth was inoculated in vancomycin screening agar along with nutrient agar. Growth in both screening agar and nutrient agar indicates chromosomal resistance which is not cured, whereas growth only in nutrient agar shows plasmid-mediated resistance. For the complete confirmation of plasmid curing, 3-6 days of daily sub-culturing in both the agar is preferred. The bacterial plasmid was extracted in accordance with the manufacturer's instructions using a plasmid extraction kit (GeNei, Bangalore). The purified plasmid DNA was dissolved and stored at 4°C in TE buffer.

Transfer of vancomycin resistance by the broth mating procedure
Broth mating was performed as follows (Saha et al. 2008). In Luria-Bertani (LB) broth (Himedia, India), pure colonies of donor and recipient cells were inoculated separately and cultivated overnight at 37°C with shaking. These overnight cultures were diluted in a fresh medium at 1:100 and each was grown to the early exponential phase. The mating combination was prepared by adding 0.1 ml of donor cells to 0.9 ml of recipient cells. The mixture was gently whirled for a few minutes, then incubated at 37°C for 6 h (without shaking). This was followed by plating on LB agar medium (Himedia, India) containing 16 μg/ml vancomycin and 2.5 μg/ml ciprofloxacin. After 48 and 72 h of incubation, colonies were counted. In addition, in the presence of vancomycin plus ciprofloxacin combination, donor and recipient cells were separately plated to assess their inability to grow because the donor was sensitive to ciprofloxacin and the recipient was susceptible to vancomycin.

Molecular characterization of vancomycin resistance-encoding genes
The isolates that phenotypically showed resistance to vancomycin were screened to detect the presence of resistance-encoding genes using appropriate primers (Clark et al. 1993;Saha et al. 2008). Table 1 shows the details of the primers used and the polymerase chain reaction (PCR) conditions. The PCR was carried out in 25 μl reaction mixture containing 15.75 μl of nuclease-free water, 10Â Taq buffer (2.5 μl) (100 mM Tris-HCl, pH 8.3, 20 mM MgCl 2 , 500 mM KCl, and 0.1% gelatin), 200 mM dNTPs (0.5 μl) (dATP, dTTP, dGTP, and dCTP), 10 pmol each of forward and reverse primers (2 μl), 1.0 unit of Taq DNA polymerase (0.25 μl), and 2 μl of the template. All the reagents were purchased from Origin Diagnostics and Research, Bangalore. The PCR was carried out in a My Cycler thermal cycler (Bio-Rad, USA). The PCR products were purified and sequencing was performed with an automated ABI 3100 Genetic analyzer using the ABI BigDYE terminator method (M/s

Identification and characterization
Samples were collected from three direct hospital effluent discharge sites located in the north, central, and south districts of the state and screened for the presence of drug-resistant bacterial pathogens. From the collected hospital effluent samples, a total of 113 S. aureus (69.75%) isolates were identified from the sediment, water and fish/shellfish samples screened (H1 ¼ 59, H2 ¼ 37, H3 ¼ 17). The number of cultivable bacteria in the H1 samples was much higher than that in the H2 and H3 samples. The results of culture on mannitol salt agar showed that all the S. aureus isolates produce positive results for fermentation of mannitol, changing the colour from red to yellow. The occurrence and distribution of multidrug-resistant S. aureus from different locations is shown in Table 2.
Antibiotic susceptibility testing, determination of MIC, and MAR index analysis of selected isolates The pattern of antibiotic resistance among S. aureus strains varied in samples collected from different discharge points. On initial testing, the growth of S. aureus isolates on the MHA screen agar plate with 32-64 μg/ml of vancomycin indicated possible resistance to vancomycin since no inhibition zone was noted around the vancomycin disc. The highest number of isolates was obtained from the H1 site (n ¼ 59). Among them, 48 isolates were found to be resistant to several antibiotics, such as amoxicillin, ampicillin, azithromycin, clindamycin, clarithromycin, chloramphenicol, ciprofloxacin, erythromycin, methicillin, oxacillin, streptomycin, and vancomycin. However, all the isolates were susceptible to gentamicin and ciprofloxacin as determined by the disc diffusion test. Twenty-eight S. aureus from H2 discharge points (n ¼ 37) and 10 isolates from H3 (n ¼ 17) were resistant to almost all classes of antibiotics similar to H1 isolates. The antimicrobial susceptibility pattern of multidrug-resistant S. aureus isolated from direct hospital effluents is illustrated in Table 3. The MICs of vancomycin for S. aureus isolates from different hospital discharge points were found to be 64-128 μg/ml, confirmed as vancomycin-resistant Staphylococcus aureus (VRSA) according to CLSI criteria. However, the MIC value increased to 1,024 μg/ml after sub-culturing the isolates in the presence of vancomycin. The 23 S. aureus isolates from the H1 discharge point were highly resistant to vancomycin and teicoplanin with MIC values ranging from 32 to 128 μg/ml. The 11 VRSA isolates from H2 and two isolates from H3 showed MIC values for vancomycin ranging from 32 to 64 μg/ml; among them, three isolates from H2 were resistant to teicoplanin also. The remaining isolates from H3 showed intermediate resistance to vancomycin with the MIC value of 16 μg/ml. An MAR index revealed that all the multidrugresistant isolates had a high MAR index value of .0.9. These results confirmed that the selected isolates have originated from a clinically high risk source of contamination.

Molecular characterization of vancomycin resistance-encoding genes
A total of 86 (H1 ¼ 48, H2 ¼ 28, H3 ¼ 10) VRSA isolates from the selected sites were screened for vancomycin resistanceencoding genes. The extracted plasmid was used as a template for PCR amplification of vanHAX, vanH, vanA, vanB, vanC, vanX, and mecA with appropriate primers. Amplicons of 1,032 bp for vanA ( Figure 1) and 433 bp for vanB ( Figure 2) were obtained by the PCR. Among the 86 VRSA isolates, 27 isolates from H1 and 13 isolates from H2 showed the existence of vanA gene. The number of vanB-positive isolates was 32 in H1 and 19 in H2, whereas none of the phenotypically resistant VRSA isolates from H3 harboured either vanA or vanB genes. Most of the vanB-positive isolates from the direct hospital effluent co-harboured vanA gene also.  Plasmid analysis and transfer of vancomycin resistance-encoding genes The plasmid profile analysis of VRSA isolates revealed that most of them carried three to four plasmids with different molecular weights ranging from 51.2, 5.5, 5.1 and 1.5 kb. The plasmid-mediated vanA gene was successfully transferred from donor to recipient S. aureus isolate by conjugation. Twenty-eight transconjugant colonies were found on LB agar with appropriate selective antibiotics, and no growth was observed in the selective medium of recipient and donor S. aureus when inoculated separately. The MIC of transconjugant was found to be 32 μg/ml.

Plasmid curing
Plasmid curing was performed with nine VRSA isolates that showed the presence of vanA gene. Following 3-6 days of daily sub-culturing, an examination of the phenotypically resistance status showed that seven plasmids cured VRSA isolates lost one large plasmid of 50 kb size and these isolates became sensitive to glycopeptide as well as macrolide classes of antibiotics.

DISCUSSION
In this study, 113 multidrug-resistant S. aureus isolates were recovered from different hospital discharge points, clearly pointing to the improper treatment of the effluent in these hospitals. Even though all the three hospitals had an MBBR system for effluent treatment in place, it was either dysfunctional or not being used. Several studies have shown that hospitals are one of the major contributors to the emergence and dissemination of multidrug-resistant bacteria into surrounding aquatic environments (ARB) (Kalasseril et al. 2020) and the dissemination of these isolates is greatly influenced by the antibiotic policies of a particular country such as the manufacture of antibiotics, the dispensation of antibiotics, and inappropriate and incorrect dosing of antibiotics (Manyi-Loh et al. 2018). Analysing the findings of previous studies showed that one of the key routes leading to the dissemination of ARB into the environment might be hospital wastewater treatment plants (WWTPs) as suggested by Szczepanowski et al. (2009). The concern about environmental antibiotic pollution in general and that of aquatic ecosystems in particular is increasing worldwide. Effluents from hospitals and even wastewater treatment plants are an important source for the release of antibiotics and antibiotic-resistant bacteria into the environment. Kummerer (2009) concluded that globally 'antibiotics are released in hospital wastewater continuously, daily, and all year round'. The aquatic environment, in particular, can act as a natural reservoir of antibiotic resistance as well as a route for the dissemination of clinical resistance traits (Michael et al. 2013). Due to the improper waste management, large amounts of antibiotics may be released into hospital wastewater due to the excretion of antibiotics used and the incorrect disposal of unused compounds, which can later be released into the environment. The hospital effluents contain a wide range of compounds used for medical, laboratory, and research purposes, as well as human excreta (Biobooster 2016). These biological active compounds are insoluble/soluble organic/inorganic Journal of Water and Health Vol 19 No 5,791 pollutants that have a harmful effect on people and aquatic animals even at extremely low concentrations. Pathogenic microbes such as viruses, bacteria, fungi, protozoans, and helminths are also found in these effluents (Santoro et al. 2015). Studies have shown that once the antibiotics reach the aquatic environment, due to their slow biodegradation, some of these drugs may persist in the environment for longer time due to bioaccumulation and biomagnification. Antibiotics in the natural environment, even in low quantities, can affect the survival, reproduction, and metabolism of aquatic organisms, as well as alter the structure of communities and ecosystem functions (Pereira et al. 2015;Martin-Laurent et al. 2019).
Kerala, India's southernmost state, is bordered on the east by the Western Ghats and on the west by the Lakshadweep Sea. The state comprises of 14 districts. Kerala is considered a pharmaceutical consumer state, with a total drug consumption of over 20,000 crores per year, and antibiotics contributing for 20% of the total drugs taken annually in the state. Although there are regulations for effluent discharge in the state, there appears to be a lack of strict implementation of these regulations and thus, most WWTP units have not been fully functional (The Times of India Report 2018; The New Indian Express 2020), as also indicated by the results obtained in the present study. The wastewater from these systems can easily contaminate the environment, allowing multidrug-resistant bacteria to spread. Extreme temperatures and high relative humidity lead to the survival of such superbugs in hospitals and the emergence of new ones in the environment, which could pose an occupational and food threat not only to fishermen and farmers, but ultimately to the wider population. The Government of Kerala has initiated the Antibiotic Stewardship Program (ASP) to regulate and promote the judicial use of antibiotics in all sectors as part of the 'One-Health' Programme (KARSAP 2018).
The presence and survival of S. aureus in the hospital wastewater discharge, observed in the present study, is thus, not too surprising. Thompson et al. (2013) reported similar results, observing elevated levels of antibiotic-resistant S. aureus in untreated hospital wastewaters and receiving sewage treatment plant of a metropolitan hospital in Australia. Antibiotic resistance among pathogenic bacteria is a well-documented phenomenon which has significant consequences for the treatment of infections. S. aureus has a unique ability to respond rapidly to any new antibiotic by developing a resistance mechanism from penicillin and methicillin to the most recent linezolid and daptomycin (Kaur & Chate 2015). Our findings were comparable with other studies which also found high levels of antibiotic resistance among S. aureus isolated from hospital wastewaters in Australia, North Europe and clinical samples in Damascus (Ali et al. 2014). Interestingly, similar to the findings obtained in our study, Saha et al. (2008) had earlier reported that the S. aureus isolated from an Indian hospital showed 100% tolerance to ampicillin, chloramphenicol, and erythromycin, but was susceptible to ciprofloxacin and gentamicin.
According to the Clinical and Laboratory Standards Institute, S. aureus isolates for which vancomycin MIC is 4-8 μg/ml are classified as vancomycin-intermediate (VISA), and isolates for which vancomycin MICs are greater than 8 μg/ml are classified as vancomycin-resistant isolates (VRSA). The MIC of vancomycin for VRSA isolates from different hospital discharge points was found to be 64-128 μg/ml. However, the MIC value increased to 1,024 μg/ml after sub-culturing the isolates in the presence of vancomycin which indicate that S. aureus isolates were becoming increasingly resistant to vancomycin compared with earlier reports (Ramana & Chaudhury 2012). This is a matter of serious concern and may pose a major problem with its use as the main drug against MRSA infections. MAR index studies revealed that all the multidrug-resistant isolates had a high MAR index value of .0.9. These results indicated that the isolates have originated from a clinically high risk source of contamination where antibiotics are often used and possibly abused, which is similar to many findings from other parts of the world (Habibi et al. 2008).
Of the six major phenotypes of vancomycin resistance, phenotypes of vanA and vanB are both common and transferable. The vanA phenotype is highly resistant to vancomycin and teicoplanin, while vanB phenotype has variable levels of vancomycin resistance, but not teicoplanin. We found both the vancomycin encoding genes, vanA and vanB, in our study among 86 VRSA isolates which showed complete resistance to vancomycin. Chang et al. (2003) reported a clinical isolate of vancomycin-resistant S. aureus with the presence of vanA resistance-encoding gene that showed a high MIC value of 1,024 μg/ml for vancomycin isolated in United States. In India, Chakraborty et al. (2011) also found that VRSA strains which harboured both vanA and vanB genes showed complete resistance to vancomycin.
Besides, a small percentage (31.3%) of phenotypically resistant isolates did not have any common VAN resistance mechanisms. In the present study, some of the VRSA and VISA strains were negative for vancomycin resistance-encoding genes by the PCR. However, the absence of these genes in the present isolates does not rule out that these strains are not VRSA or VISA. There is another hypothesis that proposes that the thickening of cell walls is responsible for the development of vancomycin resistance. The vancomycin resistance mechanism has been studied extensively with the first clinical VRSA strain, Mu50 (Hanaki et al. 1998;Cui et al. 2000). Examination of the Mu50 cell by the biochemical test and transmission electron microscopy indicated that it contains increased levels of peptidoglycan. Vancomycin-resistant staphylococci with thickening of cell walls were also demonstrated by Palazzo et al. (2005). The thickened cell wall of VRSA becomes thinner with the loss of vancomycin through sub-culturing in the absence of antibiotics. This may be the possible mechanism behind the vancomycin-resistant staphylococcal isolates that we identified, even though we were unable to perform the test for cell wall thickening demonstration in these isolates.
Vancomycin resistance was successfully transferred from S. aureus with the vanA gene to vancomycin-sensitive S. aureus by the broth mating procedure. vanA genes are located on mobile plasmids which can promote the transfer of this gene between different bacterial groups. The vanA, pheromone-responsive conjugative plasmids, are anticipated to be critical for the occurrence of VRSA. The presence of sex pheromone in S. aureus, which promotes plasmid transfer in Enterococcus spp., was further demonstrated by Showsh et al. (2001). The vanA phenotype can be transferred to other MRSA strains and through microbial species, with significantly greater potential for spread, even without vancomycin (Hageman et al. 2006). However, the detection of VRSA isolates with vanA and vanB genes in public water bodies in our study indicates that there is high possibility of the intrageneric transfer of vancomycin resistance from S. aureus strain to another.
Plasmid curing was performed with nine VRSA isolates that showed the presence of vanA gene. Following 3-6 weeks of daily sub-culturing, an examination of the phenotypically resistance status showed that seven plasmids cured VRSA isolates lost one large plasmid of 50 kb size and these isolates became sensitive to glycopeptide as well as macrolide classes of antibiotics. It has been suggested that high levels of glycopeptides, vancomycin, and teicoplanin resistance are mediated by the vanA type that may be chromosomally or plasmid located (Perichon & Courvalin 2000). The results obtained agree with the study by Shriram et al. (2013) who proved that vancomycin resistance was mainly plasmid mediated as they became sensitive to low antibiotic concentrations after plasmid curing (Shriram et al. 2013).

CONCLUSION
The data obtained in this study showed the presence of multidrug-resistant S. aureus isolates with resistance-encoding genes in public water bodies illustrating the possible spread of bacterial pathogens from direct hospital wastewaters into natural environments. This study also points to a deeper systemic issue for a state like Kerala, where hospitals are main hotspots for resistant bacteria and its dissemination. A large number of strains was found in samples collected from the direct hospital effluent indicating their survival in the treatment plants. This study demonstrates the urgent need for the effective use of wastewater treatment plants in hospitals as part of ASPs to allow the constant monitoring of hospital wastewater discharges for the presence of antibiotic residues and antibiotic-resistant bacteria into natural aquatic environment.