Metal Contaminants Occurrence and Microbial Diversity From the Perspectives of Dual Tolerance by the Bacterial Isolates Along Altitudinal Zonation of River Teesta – a Pristine River System in Eastern Himalayas


 The present study was conducted along River Teesta in Eastern Himalaya. Water and sediment samples were collected from six sampling points during the monsoon and winter seasons along the course of the river. The background information was collected by analysing physico-chemical parameters and the heavy metal concentrations. Culturable bacterial diversity using culture-based and molecular based 16S rRNA approaches characterized bacterial isolates to 5 major phyla, majority belonging Proteobacteria, Actinobacteria and Firmicutes followed by Bacteroidetes, Alpha-Proteobacteria, Beta-Proteobacteria, Gamma-Proteobacteria, Actinobacteria, Bacilli, Flavobacteriia, Deinococci, Sphingobacteriia and Cytophagia. The total of 245 characterized isolates belonged to 69 genera. Diversity indices were calculated for each site. There were variations in the community structure of culturable bacteria along the river stretch with some common and unique groups. Heavy metal tolerance and antibiotic resistance profiles showed some isolates to be tolerant to high levels of heavy metals and multiple antibiotic indicating a major concern. The antibiotic resistance diversified along the human impacted downstream sites. The present report on bacterial diversity and the associated metal and antibiotics tolerance is the first of its kind on Teesta river, the only major river system flowing through the state of Sikkim and parts of North Bengal.


Introduction
Microbial ecosystem is one of the key components of aquatic ecosystems playing important roles in various processes like biogeochemical cycling, microbial food webs, degradation of pollutants and energy ows (Guan et al. 2018). The discharge of wastewater from urban areas and industrial e uents has in uenced the bacterial community composition, species abundance and their function thereby disrupting the nutrient cycling processes and negatively affecting the energy ow and loss of biodiversity. It is an established fact that the structure and function of aquatic ecosystems are facing potential threat posed by the changing global environment while the responsiveness of microbial community to the environmental variations still remains the topic of investigation and furthermore, microbial diversity in river systems is less commonly studied than the marine and lake ecosystems (Zeglin 2015). Spatial and temporal heterogeneity in bacterial diversity has been observed in various rivers driven by several environmental factors like temperature, nutrient uptake, organic matter concentration and composition, hydrology and various land use patterns (Gessner and Chauvet 1994;Valett et al. 1997;Dodds et al. 2000;Mulholland et al. 2008;Boyero et al. 2011). From the review conducted by Zeglin (2015) it was found that metal concentration in the river system was the most prominent drivers of microbial ecosystem whereas nutrients and land use were least de ned. The same study also concluded that temporal variation (93%) was more prevalent than among stream variation (86%) or longitudinal variation (83%) in de ning the composition and structure of microbiota in lotic systems. Rivers are contaminated not only with agents like pesticides, heavy metals or pharmaceuticals but also potentially allochthonous microorganisms (Breton-Deval et al. 2019). Therefore, it is also necessary to study the response of microbes at the structure and community level and throw insights on the response of microbes to the pollutants.
The Indian rivers are not just economically important but also considered to be sacred and holds important place in Indian tradition and culture (Semwal and Akolkar, 2006). In the recent past years, with fast growing population, upsurging industrialization and various agricultural practices there has been unregulated ow of waste water into the river thereby making it vulnerable to pollution. Studies have attributed anthropogenic activities resulting in heavy metal contamination of the riverine ecosystem Patel et al. 2018;Zheng et al. 2018;Bhuyan et al. 2019). Elevated levels of toxic heavy metals like Hg, As, Pb, Cd, Cr, Zn and Cu have been reported from various Indian rivers (Hussain et al. 2017;Singh et al. 2017;Patel et al. 2018;Rahman and Singh 2018). Heavy metals stress causes shift in bacterial community structure from native to the one dominated by more resistant and tolerant bacteria. In response to toxic concentrations of heavy metals, many aquatic organism including microorganisms can develop tolerance (Klerks and Weiss, 1987). Heavy metal tolerant bacteria have been isolated from polluted river system showing tolerance to copper, lead, nickel, cobalt, chromium, cadmium and mercury (Rajkumar et al. 2012;Sair et al. 2018). Studies have revealed that co-selection takes place in environmental bacteria with metal and antibiotics resistance Kalayu and Ahemad, 2014). Bacteria in metal-con taminated environments have been reported to be a rich source for antibiotics resistance phenotypes than in control areas (Baker-Austin et al. 2006) and rivers are the major means of dissemination of antibiotic resistance genes which is a major public health concern.
The present study was undertaken to assess the spatial-temporal variation in the culturable bacterial diversity in surface water and sediment samples of the river Teesta, a pristine river system originating in the Eastern Himalayas which is subjected to various human intervention including several food, beverage and pharmaceutical companies and also to investigate the co-occurrence of heavy metal and antibiotic resistance among the isolates.

Materials And Methods
Study area and sampling River Teesta, a prominent river of the Eastern Himalayan originates in the Teesta Khangse glacier located in northeastern part of Sikkim in Northeast India. The total catchment area is 12,159 km 2 and is 414 km long with the greater part in India (151 km in Sikkim,19 km along the border between Sikkim and West Bengal, 29 km in West Bengal), 94 km along the Indo-Bangladesh border, and the nal 121 km in Bangladesh. In terms of socio-economic development Sikkim has ourished in the past decade with various food, beverage and pharmaceutical industries that have been established along the banks of the river. In addition, Teesta has been used for power generation and has ve operational hydro projects located in Sikkim and West Bengal. This has led to the destruction of vegetation and natural habitats in these areas. The sampling points and site description are depicted in Fig.1 and Table 1. Water and sediment samples were collected in presterilized containers during monsoon and winter season in the year 2018-2019 from six sampling sites along the river with the altitude ranging from 5183m to 293m above mean sea level. Site T1 is a tributary site between site S2 and S3. The samples were brought to the laboratory on ice condition.

Background study of the sampling sites
The physico-chemical parameters and the heavy metal concentration in the river surface water and sediment samples were determined following the standard APHA (1998) protocols.

Bacterial isolation and enumeration
Isolation of bacteria from both the sediment and water samples was done by the standard spread plate technique. In brief, 1 gram of sediment sample was suspended in 10 ml of 0.85% saline and kept in shaking condition (150 rpm) for 1 hour. The samples were then serially diluted up to 10 -5 and 0.1 ml from each 10 -3 , 10 -4 and 10 -5 dilutions were spread on Luria Bertani agar plates (composition in grams/litre: tryptone-10, yeast extract-5, NaCl-10 and agar -15). Similarly, water samples were diluted in 0.85% saline up to 10 -4 fold dilution and subsequently 0.1 ml each from 10 -2 , 10 -3 and 10 -4 were plated on Reasoner's 2A or R2A agar plates (composition in grams/ litre: acicase-0.50, yeast extract-0.50, proteose peptone-0.50, dextrose-0.50, starch soluble-0.50, dipotassium hydrogen phosphate-0.30, magnesium sulfate-0.024, sodium pyruvate-0.30 and agar-15.0). All the media and glasswares were sterilized by autoclaving at 121ºC for 20 minutes. The inoculated plates were incubated at 28 ºC for 24-48 hours except for the samples from site S1 which was incubated at 10 ºC for two weeks. After incubation, colonies were selected based on their characteristic color, morphology and texture and the selected colonies were puri ed by streak plate technique. The cultures were cryopreserved in the cryovials with 50% glycerol at -20 ºC. Enumeration of total bacterial load was done by the standard plate count (SPC) technique in the form of colony forming units (CFU). The bacterial load is expressed as CFU/ml in case of water samples and CFU/grams in sediment samples.
Bacterial DNA isolation and 16Sr DNA ampli cation Genomic DNA was extracted using the Bacterial genomic DNA kit by HiMedia, India. The genomic DNA was ampli ed targeting the 16S rRNA gene using the universal bacterial primers 27F: 5'AGAGTTTGATCCTGGCTCAG 3' and 1492R: 5'TACGGYTACCTTGTTACGACT T 3' (Weisburg et al. 1991). The ampli cation was carried out on a GeneAmp PCR system 9700 (Applied Biosystems, USA). Each 30 μl reaction mixture consisted of 15.75 μl of milipore water, 3μl of 10X buffer with 17.5mM MgCl 2 , 3 μl of each forward and reverse primer of 10mM concentration, 2 μl of dNTP mixture, 0.25 μl of Taq polymerase and 3 μl of template DNA. The ampli ed PCR products were electrophoresed on 1.2% agarose gel and puri ed using QIA Quick Gel Extraction Kit (Qiagen, Hilden, Germany) and sequenced at Macrogen Inc, Seoul, Korea.

Sequence alignment and phylogenetic analysis
Identi cation of the bacterial isolates were carried out by calculating 16S rRNA gene sequence similarity against all known procaryotic species in the Eztaxon server (Chun et al. 2007) which involves two steps. First step is the Basic Local Alignment Search Tool (BLAST) which is a sequence similarity search program was used to compare our query sequence to the sequence database of the National Centre for Biotechnology Information (Altschul et al. 1990) and secondly, pairwise sequence alignment to nd out sequence similarity between the query sequence and hit sequences obtained in the rst step. This algorithm uses 98.65% similarity threshold as species boundary. Before phylogenetic analysis, the sequences obtained from sanger sequencing were trimmed by the help of Molecular Evolutionary Genetics Analysis software, MEGA 7.0 (Tamura et al, 2011) and the sequences were aligned by using the clustal W inbuilt in MEGA 7.0 where the most similar sequences with best alignment score are aligned rst followed by more distant groups and a phylogenetic tree was constructed using the neighbor-joining method (Saitou and Nei, 1987) with 1000 bootstrap replications to assess nodal support in the tree (Felsenstein, 1981). The nucleotide sequences were submitted to NCBI and the accession numbers were obtained.

Diversity analysis
With a view to compare the culturable bacterial diversity in both the water and sediment samples, samples were analysed both spatially and temporally in order to understand the effect of altitude gradient, human intervention and seasonal uctuations. Diversity analysis was assessed by calculating Shannon-Weiner diversity index (H), Simpson's dominance index (D), evenness index (E) and richness index by Margalef.

Determination of MIC (Minimum Inhibitory Concentration) of the bacterial isolates against heavy metals
MIC of the bacterial isolates was determined against four heavy metals Pb, Zn, Cu and Cd by spot plate technique. Stock solutions of the metals were obtained from the metal salts of ZnSO 4 . H 2 O, Pb(NO 3 ) 2 , CuSO 4 .5H 2 O and Cd(NO 3 ) 2 purchased from Sigma Aldrich to obtain stock solution of 10,00 ppm. Heavy metal containing plates were prepared by supplementing Luria Bertani agar with various concentrations of heavy metals ranging from 50 to 2000 ppm in case of Pb, Zn and Cu and Cd was used at the concentration ranging from 1-100 ppm. The isolates were grown overnight in Luria Bertani broth and 10µl of the culture corresponding to OD 0.5 at 600nm was spotted on the metal containing plates and incubated at 28 ºC for 24 h. MIC was determined as the lowest concentration of metal that inhibited growth after 24 h. Antibiotic susceptibility test for the isolates The antibiotic susceptibility of the bacterial isolates was conducted using agar diffusion technique on Mueller-Hinton agar (MHA) plates following the guidelines of the Clinical Laboratory standard Institue (CLSI) 2013. In the present study total 15 antibiotics was used which are as follows Meropenem (10mcg), Imipenem (10mcg), Erythromycin (10mcg), Cipro oxacin (5mcg), Nalidixic acid (30mcg), Netillin (30mcg), Novobiocin (30mcg), Neomycin (30mcg), Tigecycline (15mcg), Lincomycin (15mcg), Rifampicin (5mcg), Polymyxin B (300U), Co-Trimoxazole (25mcg), Ceftazidime (30mcg), Trimethoprim (5mcg) using standard antibiotic discs (HiMedia, India). The multiple antibiotic resistance (MAR) index for the test isolates was calculated as a/b where a represents the number of antibiotics the isolates is resistant to and b represents the total number of antibiotics used against the isolate (Krumperman 1983). A MAR index value of >0.2 indicates that the isolate is multiple antibiotic resistant whereas a MAR index value of < or = 0.2 indicates a very low or negligible antibiotic resistance.
Statistical analysis PAST 4.07b software was used for conducting all the statistical analysis including the calculation of diversity indices and preparing graphs. Pearson's correlation was employed to understand the relationship between various physicochemical properties and the culturable microbial taxa in both the water and sediment samples.

Results And Discussion
Physico-chemical parameters of river water and sediment samples The mean values along with the standard deviation of the various physico-chemical parameters are presented in the Table 2. The temperature of the river water ranged from 2ºC to 19 ºC where the lowest temperature was recorded at site S1 during winter. The mean values of all the parameters were found to be below the permissible limit of BIS 2012.
However, BOD was slightly higher at site S3 followed by site S5 and T1. The determined DO at S1 is lower than the other sites because of the high altitude with low atmospheric oxygen.
The estimation of Pb, Zn, Cu and Cd were carried out in both the water and sediment samples. The mean concentration of the metals for the analysed samples are presented in Table 3. The concentration of Pb in the water samples from the sites S4, S5 and T1 slightly crossed the permissible limit of BIS during both monsoon and winter seasons whereas other heavy metals were found to be below the threshold range of BIS. In the sediment samples, the concentrations of heavy metals were found be several orders of magnitude higher than the water samples. The mean concentration of Pb in the downstream sites S4, S5 and T1 ranged from 12.5 -27.4 ppm which surpassed the limit of BIS which is 11 ppm. The concentration of Zn ranged from 1.8 ppm to 85.6 ppm and the concentration was found to exceed the BIS limit of 16 ppm in the monsoon season as compared to winter season. this elevated levels of heavy metals in the downstream sites could be the effect of industrialization.
Mann-Whitney U test, was employed due to skewness of data to nd out for any seasonal variations in the heavy metal concentrations in monsoon and winter season. The test revealed no signi cant differences in the heavy metal concentrations in monsoon and winter season for water samples. In case of sediment samples we see signi cant difference in concentration of Zn (t df(10) = 7.477, p<0.01) and Cu (t df(10) =3.533, p<0.01) between the two seasons which showed that both the concentration of Zn and Cu decreased signi cantly in winter season at various sites . However, no signi cant differences were observed in concentrations of Pb and Cd between the two seasons.

Total viable count
Out of the water and sediment samples, rich bacterial load was shown by the sediment samples which was expressed as CFU/g than the water samples depicted in Table 4. The mean bacterial load was highest in site S4 and lowest in S1 in both the water sample and sediment respectively. Less bacterial load at S1 could be explained by the high altitude of S1, cold and oligotrophic environment. From ANOVA, the variation in bacterial load between different sites was highly signi cant in both the water sample (F df(5) = 35.16, p <0.01) and sediment (F df(5) = 5.42, p<0.01)as well. However, the ttest showed no signi cant seasonal variation in the mean bacterial load in both the water samples (t df(10) =1.954, p = .079) and sediment samples (t df(10) = -.889,p = .395). In the present study, bacterial load was higher in the downstream sites which could be explained by the waste water discharge close to the sites thereby increasing the organic matter which supports bacterial growth. On the basis of colony morphology total 129 and 116 bacterial isolates were obtained both from water and sediment samples during monsoon and winter season respectively.
Culturable bacterial diversity of river water and sediment samples Based on the colony morphology and 16S rRNA gene sequencing a total of 129 isolates (water-86 and sediment-43) and 116 isolates (water-61 and sediment-55) were recovered in monsoon season and winter season respectively making a total of 245 isolates. The bacterial isolates belonged to 5 major phyla with the higher percentage of the bacterial isolates belonging to major phyla like the Proteobacteria, Actinobacteria and Firmicutes followed by Bacteroidetes and Deinococcus-Thermus. These phyla were represented by various classes belonging to Alpha-peoteobacteria, Betapeoteobacteria, Gamma-proteobacteria, Actinobacteria, Bacilli, Flavobacteriia, Deinococci, Sphingobacteriia and Cytophagia. Representative comparison of different phyla and classes in both water and sediment samples are shown in Fig 2 and Fig 3. In the water samples Proteobacteria and Actinobacteria were dominating phyla in all the sites during both monsoon and winter season. Firmicutes were found to be prevalent during monsoon season than in winter at all the sites. Deinococcus-Thermus was a unique phylum that was detected only at site S5 which represented 5% of the total phyla at that particular site. The common classes present in water samples during both the seasons are the Alphaproteobacteria, Beta-proteobacteria, Gamma-proteobacteria and Bacilli. Flavobacteriia and Deinococci was detected only during monsoon season whereas Sphingobacteriia and Cytophagia were two unique classes found only at site S1.
In the sediment samples, Proteobacteria, Actinobacteria and Firmicutes are the phyla commonly present in all the sites during both monsoon and winter season. Deinococcus-Thermus was present only at site S4 in monsoon season whereas in winter season Bacteroidetes was present only at site S1. At the class level, Alpha-proteobacteria, Betaproteobacteria and Gamma-proteobacteria, Bacilli and Actinobacteria are commonly present. Flavobacteriia was detected only at site S1 in both the seasons whereas Deinococci was found only in monsoon season present only at site S4.

Bacterial diversity at the genus level
From all the samples with 244 total isolates, a total of 69 genera were identi ed after comparing the 16S rRNA sequence with the known sequences deposited in the public database EzBioCloud, https://www.ezbiocloud.net/identify. The spatio-temporal comparison of the various genera isolated from water and sediment samples are provided in the Fig 3. The isolates exhibited >97% to the previously deposited sequences. High bacterial diversity was displayed by both the water and sediment samples. The recovered bacterial genera were higher in the water samples as compared to the sediment samples.
In case of water samples, out of 138 identi ed genera some of the genera prevalent in the water samples in more than two sites are Pseudomonas (13%) followed by Microbacterium (5.79%), Bacillus (5%), Acinetobacter (5%), Brevundimonas (4.34%), Serratia (3.6%) Arthrobacter (2.8%), Janibacter (2.8%), and that altogether accounted for 42.33% of the total isolates identi ed in the water samples and the rest 57.67% isolates represented genera that were uniquely present at less than two sites. Genera like Flavobacterium, Duganella, Pedobacter, Deefgea, Arcicella were present only at site S1 belonging to the phylum Bacteroidetes and Proteobacteria. Genera distinct to site S2 are Haemophilus, Ideonella, Roseateles and Variovorax whereas Phycicoccus, Nocardioides, Asticcacaulis and Vogesella were unique only to site S3. Myroides and Comamonas were found to be present only at site S4 and Priestia, Deinococcus, Phenylobacteium and Brucella were unique to site S5 and similarly Demacoccus, Agromyces and Gordonia were unique genera present only at site T1. The dowmstream sites impacted by human intervention were represented more by the opportunistic pathogens like Serratia, Acinetobacter, Pseudomonas, Microbacterium, Gordonia. and Brucella. Water is an easy means of dissemination of these pathogens and these bacteria can easily come in contact with the humans when the water is used for various purposes including recreation.
In the sediment samples, total 95 genera were identi ed amongst which Pseudomonas (11.57%) was prevalent in more than two sites followed by Microbacterium (10.5%), Acinetobacter (8.4%) and Arthrobacter (6.3%) accounting for 36.9% of the total isolates in the sediment samples and the remaining 60.1 % accounted for the less prevalent genera. Cryobacterium was found only in the site S1 which is characterised by high altitude and cold temperature throughout the year. Comamonas was identi ed only from the samples of site S2 and Deinococcus was found to be present only at site S4 and Hydrogenophaga, Paracoccus and Ensifer were unique genera found at site T1. Shannon index values showed that the bacterial diversity was comparatively higher in the water samples as compared to the sediment samples which is also apparent from the Shannon rarefaction curve (Fig. 4). In case of water samples, bacterial diversity as indicated by the Shannon index increased as we go downstream but however no such observation was found in the case of sediment samples. The Simpson's index ranged from 0.77-1 and 0.89-1 respectively in monsoon and winter season in the water sample. In the sediment samples the values ranged from 0.6-0.95 in monsoon and 0.75-0.96 in winter season. The Margalef's richness index for water samples ranged from 2.16-4.50 in monsoon and 1.37-2.89 in winter. According to the Margalef's index calculated for the water samples high bacterial richness was observed at site S5 (downstream site) and least at site S1 (uppermost site). Similarly, in the case of the sediment samples the values ranged from 1.12-2.92 in monsoon and 1.37-3.22 in winter samples with the bacterial richness least in the upstream sites and highest in the downstream sites except for site S5 which showed less Margalef value in winter.
This richness at particular sites in the river could possibly be due to its ability to grow at particular environmental conditions which are enriched by inputs from human intervention. In addition, the culturable bacterial diversity are limited to those bacteria which can grow in the provided media and given conditions therefore the relative abundance provided by the culturable bacteria does not provide a clear picture of the microbial community like the culture independent approaches (Venkatachalam et al. 2015). The summary of various diversity indices in water and sediment samples are given in Table 5. Shannon diversity index was used to study any seasonal variation and the t-test analysis showed no signi cant variation in diversity between seasons (monsoon vs winter) in both the water samples (t df (

Relationship between species richness and bacterial load with the various physico-chemical parameters
The species diversity and bacterial load in the water samples showed a positive correlation with each other (r=.690, p<0.05). Among all the physico-chemical parameters only temperature and Cu concentration showed a positive correlation with the species diversity (Table 6)  the abundance of Bacilli. In one study from metal polluted area, it was found that Bacillus sp. was the most dominant and frequently isolated species which indicates its potential to survive in metal polluted areas and could be an indicator of metal contaminated areas (Zampieri et al 2016).
In the case of sediment samples, Pb and Cd showed signi cant positive correlation with the bacterial load and Pb further showed positive correlation with Deinococci. Increased heavy metals have showed higher relative abundance of abundances of Deinococci in metal exposed areas. Similarly, Zn and Cu were found to be strongly positively correlated and both metals demonstrated moderate positive correlation with β-Proteobacteria (Fig. 5).
Heavy metal tolerance of the isolates In this study, total 107 bacterial isolates were tested for their tolerance to four heavy metals Pb, Zn, Cu and Cd and the MIC values were recorded. MICs are crucial in understanding the response of the bacterial isolates to the heavy metals and could be associated with the heavy metal concentration and toxicity (Zampeiri et al 2016).
The tolerance of the isolates to the heavy metals showed varying trend depending upon the sampling sites (Table 8).
The MIC of Pb ranged from 500-2000 ppm where about 82% of the isolates belonging to the downstram sites showed MIC values 2000 ppm. The MIC values of Zn ranged from 100-1000 ppm, 6.5% of total isolates displayed MIC of 100 ppm, 38% showed MIC of 250 ppm, 42% showed MIC of 500 ppm and remaining 13% showed MIC of 1000 ppm. In the case of Cu the MIC of the isolates ranged from 250-1000 ppm with 25% showing MIC value of 250, 67% displayed MIC of 500 ppm and the remaining 8% showed MIC of 1000 ppm. Considering the background values for Cd determined in the study sites which ranged from 0.02-1.2, concentartion od Cd was used at lower concentration ranging from 1-100 ppm. The MIC of the isolated ranged from 2.5-100 ppm where 2% isolated shoed MIC between 2.5-5 ppm, 26% of the isolates showed MIC of 10 ppm, 38% had MIC of 50 ppm and the remaining 44% showed MIC of 100 ppm. The degree of tolerance of the isolates to the heavy metals were as follows Pb>Zn>Cu>Cd.
Antibiotic resistance pro le of the bacterial isolates All 107 isolates were tested for antimicrobial resistance against 15 antibiotics belonging to various classes of antibiotics. The number of isolates showing resistance to various classes of antibiotics are summarised in Table 9 and percentage wise graphical representation is shown in Fig 6. which also indicates that the frequency of the antibiotics to which bacterial isolates are resistant to increases as we move from upper stretch to lower stretch which has more human impact in terms of anthropogenic input in to the river. Considering the number of antibiotics an isolate is resistant to, MAR (multiple antibiotic resistant) index was calculated for each isolate. The MAR index range for each site and the percentage of MAR bacteria at each site is presented in the Table 10 .
The MAR index at site S1 ranged from 0.13-0.2, at S2 it ranged from 0-0.46, followed by S3 that ranged from 0.2-0.53, at S4 it ranged from 0.13-0.53, TS5 showed 0.2-0.46 and nally at T1 it ranged from 0.13-0.53. An increase in the number of isolates showing multiple antibiotic resistance was observed at the downstream sites (Table 10). Highest percentage of isolates 92% and 95% was observed at sites S4 and S5 respectively. S4 is a site characterised by human settlement and several pharmaceutical companies along the banks of the river and S5 is a site downstream to S4 The percentage of MAR bacteria increased downstream along the length of the river (Table 10).

Conclusion
This study showed that the levels of heavy metas like Pb and Zn at certain human impacted areas to be elevated as compared to the other sites. The bacterial load showed signi cant variations among the sites with higher load at the downstream sites possibly explained by the anthropogenic organic matter input from adjoining industrial and domestic areas. Bacterial diversity was higher in the water samples however our study found that there was no effect of season on the diversity. Heavy metals alter the bacterial composition and function and here we isolated bacterial strains showing tolerance to both heavy metals and antibiotics and the MAR bacteria increased in the downstream sites.

Declarations
Site ID The values represent the average of three replicates with standard deviation    Pearson's correlation coe cient rank between and among the physico-chemical properties of water (A) and sediment (B) and bacterial taxa abundances. The values in the box represent the level of signi cance (p<0.05).

Figure 6
Graphical representation of the percentage of isolates resistant to various antibiotics at each sites