Isolation, characterization and molecular phylogeny of multiple metal tolerant and antibiotics resistant bacterial isolates from river Ganga, Varanasi, India

Abstract The present study was focused on the isolation of multiple metal tolerant and antibiotics resistant bacterial strains from water samples of five different Ghats of River Ganga, Varanasi, India. These strains were biochemically characterized and their phylogenetic relatedness was assumed using amplified ribosomal DNA restriction analysis fingerprinting and 16S ribosomal gene sequencing. The presence of heterogeneous groups of bacteria belonging to alpha, beta, gamma proteobacteria, and bacilli was noticed. Some of the bacterial strains like Pseudomonas, Serratia, Enterobacter, and Proteus vulgaris were mainly found at the Dashashwamedh Ghat and the Assi Ghat showing minimum inhibitory concentration 200–300 mg/L for copper, nickel, lead, and chromium. Comamonas sp. mainly isolated from the Samne Ghat and the Rajendra Prasad Ghat was able to grow at very high concentration of lead viz. 400 mg/L. Some of the strains showed multidrug resistant property against 10 different antibiotics which are of most serious concern because these drugs are frequently used against various bacterial infections.


PUBLIC INTEREST STATEMENT
Mixing of sewage along with industrial effluents into the Ganga river perturb water quality. Polluted water became a habitat of various pathogenic bacteria having dual properties in rendering multiple heavy metals tolerant capacity as well as antibiotics resistance. The present study was focused on the isolation of multiple metal tolerant and antibiotics resistant bacterial strains from water samples of five different Ghats of river Ganga, Varanasi, India. These strains were biochemically characterized and their phylogenetic relatedness was assumed using amplified ribosomal DNA restriction analysis fingerprinting and 16S ribosomal gene sequencing. Collaboration of physiological, biochemical, and molecular techniques could provide complete information about bacterial identification and characterization. Presence of harmful bacteria in the river Ganga invites serious attention from human health perspective since Ganga water is also used for drinking, holy bathing and irrigation purposes.
Introduction of antimicrobial agents such as antibiotics into the river and streams from industrial pollution as metals, emerged as an ever increasing health hazard due to spread of antibiotic resistant organism in water environment (Baquero, Martínez, & Cantón, 2008;Levy, 1997). Ability of different species of bacteria in acquiring multidrug resistance (MDR) has led to severe complication for human beings (Tenover & McGowan, 1996). A large proportion of MDR was found to contain plasmid borne mobile resistance genes that can transmit through bacterial populations (Kumarasamy et al., 2010).
Molecular techniques have initiated a new era of microbiology and may prove to be helpful in determining the inhabitants of freshwater microbial communities (Mukherjee, Kumar, Kumar, & Chakraborthy, 2013). Sequences based on 16S ribosomal gene reveal numerous cultured as well as uncultured microorganisms in diverse environments as 16S ribosomal RNA (16S rRNA) molecules contain both highly conserved as well as variable regions (Woese, 1987). In the recent few years, amplified ribosomal DNA restriction analysis (ARDRA) has become a useful tool for assessment of microbial community structure changes as it is a simple, fast and reliable method for evaluation of bacterial diversity in different environments (Martínez-Murcia, Acinas, & Rodriguez-Valera, 1995).
In the present communication, the water quality of river Ganga has been analyzed using physicochemical properties. Isolation and biochemical characterization of multiple metal tolerant and antibiotics resistant bacteria have been investigated. Molecular characterization of isolates based on 16S rDNA gene sequences and ARDRA fingerprints were performed. The significance of the proposed study has been ascertained by applying statistical parameters such as pearson correlation/regression and principal component analysis (PCA) analysis.

Sample collection and physicochemical analysis
The areas selected for the present study are based on the human use and pollution. Water samples were collected from the bank of five Ghats (5 km area) viz. Dhashashwamedh ghat, Harishchandra ghat, Samne ghat, Rajendra Prasad Ghat and Assi Ghat. Water samples (1 L) were collected monthly during the first week of each month from January 2014 to December 2014 from 15 cm depth at 8.00-10.00 am near bank of each Ghat in three replicate in 1 L acid washed glass bottles. Temperature, pH, and conductivity were measured on the spot using a thermometer, a pH-meter and a conductivity meter. Samples were then transported to the laboratory under sterile condition for isolation of bacteria and analysis of dissolved oxygen (DO), biochemical oxygen demand (BOD), total hardness (TH), total alkalinity (TA), calcium (Ca 2+ ), magnesium (Mg 2+ ), sulfate (SO 4 − ), total dissolved solids (TDS), total nitrogen (TN) and total phosphorus (TP), the latter estimated using the American Public Health Association [APHA] (1998) method. Estimation of TN content was performed by following the micro Kjeldahl technique using a Gerhardt automatic analyser (model KB8S Kjeldatherm, Germany). Titrimetric methods were for testing DO, BOD and TA whereas a gravimetric method was used for TDS (Vogel, 1978). Turbidometric method was used for estimation of SO 4 −2 (Vogel, 1978). An ethylenediaminetetraacetic acid titrimetric method was used to estimate Ca 2+ , Mg 2+ , and TH (Vogel, 1978). An atomic absorption spectrophotometer (Model 2380, Perkin Elmer, Inc. Norwalk, CT, USA), was used for determination of Cr, Cd, Ni, Pb, Zn, and Cu. A Pearson correlation matrix was constructed by using SPSS software version 16.0 in order to study correlation among physicochemical parameters at five Ghats of river Ganga Varanasi, India.

Isolation and maintenance of bacterial isolates
In order to get all types of bacterial strains, water samples from each Ghats were concentrated through centrifugation at 10,000 rpm for 15 min. Pellet was dissolved in 100 μL sterilized distilled water and spread on nutrient agar (NA) media containing 0.5% peptone, 0.3% beef extract/yeast extract, 1.5% agar, 0.5% NaCl, and incubated at 37°C. Numerous colonies were obtained on NA plates. The single bacterial colonies showing unique morphology were isolated and subcultured at least five times to ascertain purity of the isolates. The purity of the isolates was monitored routinely by microscopic analysis. Purified bacterial isolates were subjected to the biochemical and molecular identification tests.

Minimum inhibitory concentration of metals
The isolated pure bacterial cultures were checked for their respective minimum inhibitory concentrations (MICs) towards all the metal ions. Different concentrations of metals ions such as Zn,Cd,Cu,Ni,Pb,and Cr (0,25,50, and 100 mg/L) were added separately against each bacterial isolates incubated in solid NA medium and then kept for 24-48 h at 37°C in incubator. Those bacterial isolates, in which growth was visible up to 100 mg/L were further grown at higher metal concentration ranging from 150,200,250,300,350 to 400 mg/L. At that metal concentration where bacterial isolates failed to grow on plates even after 7 days of incubation were considered as the MIC for that particular bacterial strain (Narasimhulu, Rao, & Venu Vinod, 2010).

Antibiotics resistance assay
Bacterial strains were checked for antibiotics resistance pattern against ampicillin (Amp), chloramphenicol (Chl), nalidixic acid (Nal), tetracyclin (Tet), vancomycin (Van), ciprofloxacin (Cip), imipenem (Imi), erythromycin (Ery), streptomycin (Str), and polymixin (Pol) (Himedia, India) according to Kirby-Bauer disk diffusion method (Bauer, Kirby, Sherris, & Turck, 1966). A volume of 5 mL of nutrient broth was prepared for inoculating bacterial strains which were incubated overnight at 37°C. The pure broth cultures were then swabbed over the sterilized 20 mL Mueller Hinton pre-made agar plates. Disks of 6 mm diameter were prepared from sterilized Whatman filter paper. The 10 antibiotics at concentrations of 30 mg/L were delivered in 10 disks which were then dispensed on each of the swabbed plates at appropriate distances from each other. Plates were then incubated at 37°C for 24 h. Zones of inhibition were measured in diameter across the centre of each zone in millimeters.

Physiological characterization of isolated bacteria
Gram staining was performed for differentiating Gram-positive and Gram-negative bacteria based on the differences in bacterial cell wall peptidoglycan layer (thick layer present in gram positive).

Citrate test
A volume of 100 mL Simmon citrate agar medium (15 g/L agar, 0.2 g/L ammonium dihydrogen phosphate, 0.08 g/L bromothymol blue, 0.8 g/L disodium ammonium phosphate, 0.2 g/L magnesium sulfate heptahydrate, 5 g/L sodium chloride, 2 g/L trisodium citrate and 15 g/L agar) pH 7.0 was prepared and then autoclaved. The sterilized Simmon citrate agar medium was poured in different glass tubes and placed obliquely to prepare slants. After that, the slants were streaked with different bacterial isolates and kept for 48 h in the incubator. The change in color from dark-green to blue was noticed after 48 h (Claus, 1989;Jawetz, 1989).

Catalase test
A small inoculum of bacterial isolates was transferred to a clean dry glass slide and few drops of hydrogen peroxide solution (3%) were added. Rapid evolvement of oxygen bubbles gave the positive result for catalase (Facklam & Elliott, 1995).

Oxidase test
One to two drops of Kovács oxidase reagent (1% tetra-methyl-p-phenylenediamine dihydrochloride, in water) (Sigma-Aldrich, USA) were placed on freshly grown bacteria on Nutrient agar media (Kovacs, 1956). Color change to dark purple gave positive test for oxidase.

Indole test
About 100 mL of tryptone broth was prepared and autoclaved. 7 mL of the medium was added in 20 test tubes. The bacterial colonies were inoculated in the test tubes exclusive of blank. These test tubes were kept at room temperature for 48 h. After 48 h 10 drops of Kovac's reagent was added and kept for the color change (Isenberg & Sundheim, 1958).

Urease test
The bacterial isolates were inoculated in the urea agar media containing pH indicator phenol red and kept under 37°C under incubator for 24-48 h. Color change to deep pink color gave the positive reaction for the presence of urease (Seeliger, 1955).

Cloning and sequencing
The amplified 16S ribosomal RNA was purified from agarose (1.2% w/v) using QIA quick Gel extraction Kit according to manufacturer's instruction. The purified PCR products were cloned in pGEM-T vector and transformed into competent E. coli DH5α cells using standard protocols. The plasmid insert were amplified using vector specific primer SP6 (5′-GATTTAGGTGACACTATA-3′) and T7 (5′-TAATACGACTCACTATAG-3′′). PCR was performed in 25 μl aliquots containing 20-50 μg DNA templates, 0.4 μM of each primers, 1.5 μM MgCl 2 , 200 μM dNTPs and 1U/μl Taq Polymerase. The PCR conditions comprised an initial denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 56°C for 30 s and extension at 72°C for 50 s with final extension at 72°C for 20 min. The PCR products were analyzed on a 1.5% agarose gel stained with ethidium bromide in 1X TBE buffer and visualized under gel documentation system. The bands were cut and eluted using Qiagen quick Gel Extraction kit and eluted amplified products sent directly for sequencing (Sci Genome Cochin, Kerala, India).

ARDRA
ARDRA was performed prior to sequencing in order to determine number of genotypes and genetic diversity among bacterial strains. Generated amplified fragments of 16S ribosomal RNA were digested with restriction endonuclease AluI, HaeIII (Arya, Joshi, Gupta, Kumar, & Raturi, 2015) and AfaI/MspI (Zeng, Liu, Li, Yu, & Chen, 2007) in accordance with manufacturer's instruction. The restriction products were resolved in 2% agarose (w/v solution of agarose in TBE buffer containing 1 μg/ mL of ethidium bromide). Gels were visualized using Bio Rad Gel Documentation system. All reactions were repeated at least three times for each restriction enzyme to ascertain the reproducibility of banding pattern.

Phylogenetic analysis of ARDRA fingerprints
The presence and absence of distinct and reproducible bands in each of the individual DNA fingerprinting pattern generated by AluI, HaeIII, and AfaI/MspI profiles were converted into binary data, and the pooled binary data were used to construct composite dendrogram respectively. The Biodiversity Pro software (Version 2) was used to perform the phylogenetic analysis using Bray-Curtis Cluster analysis.

Principal component analysis
DNA fingerprints obtained using all the 4 restriction endonucleases viz. AluI, HaeIII, and AfaI/MspI were utilized for the PCA. The software Sigma plot 11 was used to generate the graphical representation of the values generated by Biodiversity Pro software version 2.0.

Sequence analysis and construction of phylogenetic tree
Sequences of all the 20 isolates were validated by comparing them with sequences available in a public database (Ribosomal Database Project release 9.43 [RDP]) (Cole et al., 2005). The 16S rDNA sequences obtained were then subjected to nucleotide basic algorithm search tool (Blastn) (http:// blast.ncbi.nlm.nih.gov/Blast.cgi) and aligned with the most similar bacterial species and then submitted to the NCBI database using the submission tool Sequin. Sequences exhibiting close relatedness in a BLAST search were used for construction of phylogenetic tree by neighbor joining algorithm method with the aim to study their evolutionary history (Saitou & Nei, 1987). The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1,000 replicates) showed above the branches (Felsenstein, 1985). The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Jukes-Cantor method (Jukes & Cantor, 1969) and are in the units of the number of base substitutions per site. All positions containing gaps and missing data were eliminated. There were a total of 120 positions in the final data-set. Evolutionary analyses were conducted by using MEGA 6 software (Tamura, Stecher, Peterson, Filipski, & Kumar, 2013).

Results and discussion
In the present study, attempts have been made to isolate different metal tolerant as well as antibiotic resistant bacterial strains from the banks of Ghats of river Ganga and characterize them using various biochemical and modern molecular techniques. In order to analyze water quality, physicochemical characteristics of Ganga water near Ghats was performed and found to be higher than maximum permissible limit set by Indian Standard Specification for Drinking Water (IS: 10500) throughout the year (Table 1). Out of five selected Ghats, Assi ghat and Dashashwamedh ghat were found to be most polluted Ghats. Water quality of the Assi ghat has drastically deteriorated in the recent past as Assi Nala (Sewage) is directly connected to the Assi Ghat which is the mixing point of untreated waste water, domestic water and industrial pollutants with the river water. But in Dashashwamedh ghat, extreme level of human interference is possibly the major reason of increased water pollution. Level of nutrients viz. total nitrogen and total phosphorus were found to be higher in the Harishchandra ghat which is mainly due to the combustion of human bodies on this particular ghat. Other factors like total alkalinity and total dissolved solids were also high in the same ghat because washing of clothes is also very frequent. The level of heavy metals was found much beyond the safe limit at all the Ghats of river Ganga. According to Indian Standard Specifications for Drinking Water IS: 10500 safe limit (Indian Standard Specifications For Drinking Water IS: 10500, 2012), Cd concentrations beyond 0.01 mg/L becomes toxic, Cr above limit of 0.05 becomes carcinogenic and Pb beyond 0.1 could cause blood cancer in human. Atmospheric deposition of metals emitted from vehicles, direct anthropogenic activities and discharge of untreated industrial effluents as well as domestic sewages into the river could be the possible reason of metal load in river Ganga (Pandey et al., 2010). The higher level of Cu in water might be due to corrosion of pipes, fittings, and utensils.
A marked correlation between physico-chemical properties of water collected from different Ghats was observed after performing the statistical analysis using Pearson's formula of correlation and regression (Table 2). Metals like Zn, Fe, and Ni showed significant correlation with different abiotic factors. At higher concentrations, metals exert toxic effects by forming complexes with organic compounds and toxic effects of metals can change by changing metals structure (Das et al., 2013).

Table 1. Physicochemical characteristics of 5 Ghats of river Ganga Varanasi
Notes: ± Represents mean standard deviation with three replicates; note that the values are minimum and maximum;

Parameters
Site 1   A total of 115 culturable bacteria were isolated on the basis of their morphological and biochemical properties. Of these, 20 were selected for further studies due to their ability to tolerate toxicity of six different metals (Zn, Cd, Cu, Ni, Pb, and Cr) and their resistance against 10 different antibiotics. Bacterial isolates viz. Enterobacter sp., Serratia sp., Pseudomonas fluorescens and Pseudomonas pseudoalcaligenes were able to grow at very high concentrations of various heavy metals such as Cu, Ni, Pb, and Cr. The minimum growth inhibitory concentrations of these metals were in between 300 and 400 mg/L (Table 3). Comamonas has recovered mainly from Samne ghat and Rajendra Prasad ghat where the concentration of Pb was considerably high. Industrial effluents are thought to be the major source of high Pb in these sites. The abundance of Comamonas in these extremely Pb polluted sites might be due to their genomic adaptation and subsequent increased tolerance to Pb. Occurrence and activity of Comamonas strains have already been reported in Pb contaminated soils (Konopka et al., 1999). Stoppel and Schlegel, (1995) also reported metal tolerant strain of Comamonas sp. Although many Comamonas strains has been isolated from clinical samples and regarded as opportunistic pathogens (Willems et al., 1991). The persistent presence of Serratia, Pseudomonas pseudoalcaligenes, Enterobacter sp., Proteus vulgaris and Shewanella sp. in the water sample collected from the Assi ghat and Dashashwamedh ghat was observed throughout year. Along with high MIC towards heavy metals, these isolates also showed MDR against Amp, Chl, Nal, Tet, Van, Imi, Cip, Ery, Pol, and Str (Table 4). Zone of inhibition of different antibiotics (Amp, Chl, Nal, Tet, Van, Imi, Cip, Ery, Pol, and Str) against sensitive bacterial strains were found in the range of 5.02 to 11.23 mm. Except Serratia, Pseudomonas pseudoalcaligenes, Comamonas, and Proteus vulgaris, all the other strains were sensitive towards imipenem, a β-lactam carbapenem (Yong et al., 2002). Maximum zone of inhibition produced by imipenem in Aeromonas hydrophila, and Ochrobactrum intermedium viz.11.03 and 11.23 mm, respectively. Therefore, the ability of these bacterial strains to overcome the toxic effects of different drugs poses a serious problem because these drugs are frequently used against bacterial infection. Imipenem is a high end broad spectrum drug mainly used to combat nosocomial Table 3. Antibiotics resistance patterns of 20 bacterial isolates Notes: "S"-Sensitive "R"-Resistant; Amp-Ampicillin, Chl-Chloramphenicol, Nal-Nalidixic acid, Tet-Tetracyclin, Van-Vancomycin, Imi-Imipenem, Cip-Ciprofloxacin, Ery-Erythromycin, Pol-Polymixin, Str-Streptomycin.

Bacterial isolates Conc. (mg/L) Amp Chl Nal Tet Van Imi Cip Ery Pol Str
Providencia infections. In the Assi ghat, city municipal waste and untreated sewage are the pervasive source of water pollution and also a suitable environment for the growth of human pathogenic bacteria. These bacteria are known for colonizing the respiratory and urinary tract and have also the ability to cause meningitis, arthritis and wound infection. In Dashashwamedh ghat, uncontrolled human interferences and other various activities like boat landing have created water logging condition where harmful bacteria grow in copious amount. Aeromonas, Kurthia, Ochrobactrum, and Brucella were the usual bacterial strains in all the sites throughout year. Morphological and biochemical characteristics of bacterial strains are mentioned in Table 5.
The advent of tools in molecular biology has revolutionized the process of strain identification and saved money, time and energy too. In this study ARDRA was used as a model to examine the capacity of restriction-based techniques for bacterial strain identification, and the possibility of generating phylogenetic information from ARDRA-based dendrograms. Approx. 1.3 kb 16S ribosomal RNA amplified product was observed. ARDRA concerning the 16S ribosomal RNA amplified product of bacterial isolates digested by AluI, HaeIII and AfaI/MspI were reproducible yielded 2-5, 3-7, and 2-4 bands respectively. Band pattern of each isolates represented particular ARDRA genotypes. AluI showed similar banding pattern for Aeromonas hydrophila and Aeromonas sp., Shewanella decolorationis, and Shewanella, Comamonas terrigena and Comamonas sp., Pseudomonas fluorescens, and Pseudomonas pseudoalcaligenes as shown in Figure 1a. Also double restriction endonuclease digestion AfaI/MspI was failed to differentiate most of the bacterial isolates and yielded lowest number of genotypes showing in Figure 2a. However, HaeIII discriminated all the strains on the basis of band pattern and generated species specific patterns as shown in Figure 3a which was further confirmed by sequencing data that all isolates belong to different species. Therefore, it was found that out of all the restriction endonuclease, HaeIII could able to detect interspecies and interstrain variability of cultivable fresh water bacterial strains Restriction endonuclease viz. HaeIII has also been used for bacterial identification and diversity assessment by many workers (Heyndrickx, Vauterin, Vandamme, Kersters, & De Vos, 1996;Sun et al., 2008)    In order to study evolutionary relatedness, phylogenetic assessment of isolated bacterial strain was performed using ARDRA fingerprints. The AluI fingerprints based phylogenetic tree revealed two major clusters i.e. clusters I and cluster II shown in Figure 1b. The cluster I was divided into subcluster IA and IB. Subcluster IA was again divided into subcluster IA.a and IA.b where cluster IA.a contained Stenotrophomonas as a single genotype. In subcluster IA.b Exiguobacterium sp. and Rummellibacillus sp. belongs to bacilli group formed a close clustering. Cluster IB contained Brucella sp. and Ochrobacterium intermedium formed a tight cluster belongs to alphaproteobacteria group. Cluster II was bigger and divided into subclusters IIA, IIB and IIC. Subcluster IIA again comprised of IIA.a, IIA.b, and IIA.c. IIA.a represented Comamonas sp. and Comamonas terrigena belongs to betaproteobacteria. Subcluster IIA.b contained Kurthia sp. and Kurthia gibsonii belongs to bacilli group. Subcluster IIA.c comprised of both Shewanella sp. and Shewanella decolorationis belongs to gammaproteobacteria. Cluster IIB represented gammaproteobacteria members like Pseudomonas fluorescens and Pseudomonas pseudoalcaligenes formed tight cluster. Cluster IIC divided into subclusters IIC.a and IIC.b. Subclusters IIC.a and IIC.b comprised of gammaproteobacteria members. Enterobacter sp. represented outgroup so did not include in clustering.

Bacterial isolates Zn (mg/L) Cd (mg/L) Cu (mg/L) Ni (mg/L) Pb (mg/L) Cr (mg/L)
AfaI/MspI fingerprints based phylogenetic tree showed two major clusters i.e. cluster I and cluster II as in Figure 2b. Cluster I comprised Rummellibacillus sp. as a single genotype. Cluster II was a major cluster divided into subclusters IIA, IIB, and IIC. Cluster IIA showed tight cluster between Stenotrophomonas sp. and Exiguobacterium sp. Although cluster IIB gave clear picture of close cluster between bacilli group members viz. Kurthia sp. and Kurthia gibsonii. Further IIC divided into many subclusters like IIC.a, IIC.b, IIC.c, and IIC.d where intermixing was served very prominently between alpha, beta, gammaproteobacteria and bacilli members.  Dendrogram based on HaeIII fingerprints revealed the presence of 2 clusters out of which the cluster II was the largest one and divided into many subclusters represented members of all the groups like alpha, beta, gammaproteobacteria and bacilli. Whereas cluster I comprised Enterobacter sp. as an outgroup shown in Figure 3b.   The dendrogram generated after combining all DNA fingerprints obtained from ARDRA using AluI, HaeIII and AfaI/MspI showed two clusters i.e. cluster I and cluster II as depicted in Figure 4a. Cluster I revealed members of alphaproteobacteria. Cluster II was a major cluster divided into cluster IIA, IIB, IIC and IID. Cluster II A further subdivided into subcluster IIA.a, IIA.b and IIA.c. Members of betaproteobacteria showed a tight clustering in cluster II A.a. Cluster IIA.b comprised members of bacilli group viz. Kurthia sp. and Kurthia gibsonii whereas in cluster IIA.c included members of gammaproteobacteria. Cluster IIB comprised Pseudomonas fluorescens and Pseudomonas pseudoalcaligenes. Cluster IIC represented Stenotrophomonas sp. as a single genotype. Cluster IID subdivided into subcluster IID.a, IID.b and IID.c. Subcluster IID.a included bacillli group members whereas subcluster IID.b and IID.c comprised members of gammaproteobacteria.
Based on the result of ARDRA fingerprinting and phylogenetic characterization of bacterial isolates, we could say that ARDRA is a robust means which can generate species specific pattern for rapid assessment of microbial diversity. Representative strain of each ARDRA fingerprints could be chosen for 16S ribosomal RNA sequencing for obtaining information about what type of microorganism present in the sample. The reliability of ARDRA technique in congruence with the phylogenetic analysis for microbial diversity study was also proved by many workers (Arzu, Nilgun, Birgul, & Cumhur, 2012;Baik et al., 2008). Vaneechoutte and Heyndrickx (2001) mentioned ARDRA technique as a new approach for bacterial identification, phylogeny and taxonomy study. Moyer, Tiedje, Dobbs, and Karl (1996) compared phylogenetic trees based on ARDRA profiles and 16S rDNA sequences and therefore concluded that using different types of restriction endonucleases could be able to yield 76-100% success in obtaining phylogenetic affiliations accurately. Although, more studies on efficiency of ARDRA with other restriction endonucleases required with respect to fresh water bacterial community.  Furthermore, the fingerprints obtained using all the 4 restriction endonucleases viz. AluI, HaeIII and AfaI/MspI were utilized for the PCAPCA in order to check whether phylogenetic tree were in coherence with PCA or not. PCA revealed the presence of seven major clusters and two minor clusters showing in Figure 4b. In the major clusters, Aeromonas sp., Aeromonas hydrophila, Serratia marcescens, Providencia sp. and Proteus vulgaris represented members of group gammaproteobacteria settled into one cluster. Members of betaproteobacteria viz. Comamonas sp. and Comamonas terrigena occupied single cluster, whereas members of bacilli groups were represented into two separate clusters. Members of gammaproteobacteria viz. Pseudomonas fluorescens, pseudomonas pseudoalcaligenes, Shewanella sp. and Shewanella decolorationis were occupied two major clusters. In the major clusters, members of alphaproteobacteria viz. Brucella sp. and Ochrobactrum intermedium comprised in separate cluster. In minor clusters Enterobacter sp. as well as Stenotrophomonas sp. represented outgroup. Thus the phylogenetic tree and the PCA gave a clear picture of evolutionary lineage of bacterial isolates.
The phylogenetic tree was constructed by using neighbor-joining method ( Figure 5). Bootstrap values (%) were based on 1,000 replicates and more than 50% bootstrap values were shown at the branch point. The sequenced 20 bacterial strains grouped into four major classes; alpha, beta, gammaproteobacteria and bacilli. The dominant group was gammaproteobacteria having 60% bacterial isolates, only 10% isolates belongs to alphaproteobacteria, 10% were betaproteobacteria, and 20% were bacilli. Therefore, it is worth mentioning that combination of physiological, biochemical and molecular methods gives clear cut picture about the existence of multiple metal tolerant and antibiotics resistant bacterial community in fresh water system.

Conclusion
Mixing of sewage along with industrial effluents into the Ganga river perturb water quality. Polluted water became a habitat of various pathogenic bacteria having dual properties in rendering multiple heavy metals tolerant capacity as well as antibiotics resistance. Collaboration of physiological, biochemical and molecular techniques could provide complete information about bacterial identification and characterization. Presence of harmful bacteria in the river Ganga invites serious attention from human health perspective since Ganga water is also used for drinking, holy bathing and irrigation purposes.