Aquatic environmental pollution is a global threat to natural biodiversity and human health (Nicolopoulou-Stamati et al., 2016). With the ever-increasing human population, the use of pesticides and other agro-input has been increased to achieve global food security (Ehrlich and Harte, 2015). Apart from regular use in the agricultural sector, pesticides are used in household purposes to control vector-borne diseases (VBD) like malaria, especially in developing countries (Van den Berg et al., 2012), including India. But the indiscriminate use of pesticides in agriculture and household ultimately results in a significant amount of their residues in an environment which eventually washes out to natural streams like rivers, wetlands, and finally to sea, causing a huge disruption in the natural aquatic biodiversity. It has been reported that pesticide residue found in wetlands, rivers, and sea in substantial amounts can cause the innumerable ill effect to aquatic organisms like endocrine disruption, growth reduction, etc. (Carvalho et al., 2009). Uses of organochlorine (OC) pesticides are banned in many countries for a long, but they are still found in the natural streams, sediments, aquatic flora, and fauna due to their high persistency. Several aquatic ecosystems including rivers, wetlands, and sea were being continuously monitored regularly for assessing the level of OC and their risk assessment on aquatic life (DiScenza et al., 2017; Fair et al., 2018; Unyimadu et al., 2018). Due to their high lipophilicity properties, they are being easily adsorbed on the river sediments and by the aquatic fauna. High concentrations of OC have been reported in the fish, Drapane africana (2237–6368 μg/kg) as compared to Mochokus niloticus (1006–3288 μg/kg) in the Niger River of Nigeria (Unyimadu et al., 2018). Gene networks involved in the reproduction and immune function of largemouth bass, are reported to disrupt in OC contaminated sites of the Lake Apopka (Martyniuk et al., 2016). Other than OC, the frequently used pesticides found in the natural stream belong to the group organophosphate, carbamates, pyrithroids, etc. Chlorpyrifos and other organophosphate pesticides are reported to impair immune function and structural integrity of the fish Cyprinus carpio L. through oxidative stress and apoptosis (Jiao et al., 2017). Lambda-cigalothrine, a pyrithroid insecticide has been reported to cause the alleviation of free amino acid content in the muscle, liver, and brain of fishes in the Alazani River (Dzhikiia et al., 2011).

Because of the toxic effect of pesticidal residues on the native flora and fauna, several physical and chemical methods are being used to eliminate these toxic xenobiotics from natural environments, like landfills, recycling, pyrolysis, etc., However, bioremediation using different microorganisms was found to be the most feasible technique, proved to be worked in many contaminated sites. It was reported that the presence of organophosphate hydrolase (OPH) enzyme in these microorganisms has the capacity of detoxifying them through cleavage of phosphate ester bonds (P-O, P-F, P-CN, and P-S) (Schofield and DiNovo, 2010). The yeast (Saccharomyces cerevisiae) capable of hydrolyzing the poorly hydrolyzed P-S class of organophosphate by integrating gene, encoding the wild-type OPH (enhanced variant enzyme S308L-OPH) into the ribosomal operon of the same (Makkar et al., 2013). Similarly, bacterial strain, Bacillus pumilis, was reported to be capable of bioremediation of methyl parathion, a P-S type OP pesticide, due to the presence of opdA gene (Ali et al., 2012). Several microorganisms belonging to the genera Bacillus, Brevibacillus, Ochrobactrum, Pseudomonas, Serratia, and Sphingobium are reported to be capable of degrading various pyrithroid pesticides through metabolic activity (Cycon and Piotrowska-Seget, 2016). Several genes with pyrithroid degrading ability were reported viz. estP in Klebsiella sp. ZD112 (Wu et al., 2006), pye3 from the metagenome of soil (Li et al., 2008), pytH in Sphingobium sp. JZ-1 (Wang et al., 2009), pytZ and pytY (from a genomic library of Ochrobactrum anthropi YZ-1) (Zhai et al., 2012). Not only bacteria, but several fungi were also reported with the ability of pesticide remediation through the catabolic or co-metabolic process. Fungi belonging to the genera Aspergillus, Candida, Cladosporium, Tricoderma, etc., were reported to degrade different pyrethroid pesticides like cyfluthrin, bifenthrin, deltamethrin, etc. (Cycon and Piotrowska-Seget, 2016).

The natural aquatic ecosystem holds and perishes a good amount of microbial population, which can degrade these toxic xenobiotic residues in situ. Therefore, to identify these potential microorganisms and their functional role, metagenomic studies have been conducted extensively in recent years. Metagenomic studies revealed a sediment microbiome in the Deepwater Horizon oil spill and the impact of oil deposition on microbial communities in surface sediments (Mason et al., 2014). Metagenome-assembled genomes (MAGs) revealed the microbial communities of two thermal pools in Kamchatka, Russia (Wilkins et al., 2019). Similarly, the key bacterial species were characterized in the Daphnia magna microbiota using shotgun metagenomics (Cooper and Cressler, 2020). Antibiotic Resistance Genes (AMRs) from the sediments of River Yamuna have been identified from the metagenomic study (Das et al., 2020). Using the metagenomics approach, potential microbial community involved in the biodegradation of phenanthrene, diesel, and hexadecane in the mangrove sediment were identified and the degrading bacteria belonged to genera Bacillus sp., Pseudomonas sp., Acinetobacter sp., and Staphylococcus sp. (Tiralerdpanich et al., 2018). The metagenomic approach was also utilized to demonstrate the role of IS1071, an insertion element flanks with xenobiotic degradation, in the formation and distribution of bacterial catabolic pathway gene cluster (Dunon et al., 2018). The biodegradation pathway of DDT, HCH, and atrazine in freshwater and marine sediments was analyzed through a metagenomics approach. The study identified 69 genera capable of degrading these persistent pesticides with major populations belonging to genus Plesiocystis sp., Anaerolinea sp., Jannaschia sp., and Mycobacterium sp., and found the presence of different genes viz. atzB, hdg, and hdt which encode for ethylaminohydrolase, dehalogenase, and hydratase, respectively (Fang et al., 2014). Carbamate pesticide degrading enzyme was also reported by functional metagenomic analysis of rumen samples of Holstein dairy cows (Ufarte et al., 2017). Similarly, the metagenomics approach was utilized to understand the underlying mechanism of biodegradation in situ and to predict the degradation potential of microbes in the soil of Queensland, Australia (Jeffries et al., 2018). However, there were not many studies on the occurrence of the native microbial population in the Ganga and Yamuna river ecosystem having bioremediation potential. With this background, the present study accentuates the identification and relative abundance of potential bioremediation microbes capable of degrading pollutants in these river sediments through the metagenomics approach. Furthermore, the study also aims to investigate the diversity and relative abundance of these bioremediation microbes in the different polluted and non-polluted sites of river Ganga and Yamuna.

The present metagenomics study identified the potential bioremediation microbes in the sediments of river Ganga and Yamuna and investigates the correlation between microbial relative abundance and diversity with polluted and non-polluted sites of these rivers. The lists of identified bioremediation microbes with target pollutants are given in Supplementary Tables S4, S5. In recent times, bioremediation of several toxic chemicals reported throughout the world. Sediments from rivers and lagoons of the Southern Gulf of Mexico were used for the biodegradation of hexadecane (Garcia-Cruz et al., 2018). The proteins of Brevibacillus brevis respond to Triphenyl phosphate and provide novel insights into the biodegradation process of bacteria under adverse environmental conditions (Wei et al., 2018). A previous study revealed that the bacterial isolates of a stream in Manaus-Amazon could bioremediate chromium-polluted ecosystems. These bacteria were classified in 10 genera viz. Micrococcus sp., Proteus sp., Acinetobacter sp., Acidovorax sp., Bacillus sp., Enterobacter sp., Comamonas sp., Alicycliphilus sp., Serratia sp., and Vagococcus sp. (Teles et al., 2018).

The present research was designed to identify bioremediation bacteria and fungi species of commercial importance and their relations with different riverine ecosystems using high-throughput sequencing and computational metagenomics. The sensitive taxonomical classification was used for the identification of microbial community biodiversity and their functions (Mendes et al., 2017). A large number of bioremediation bacteria and fungal species (Bacteria: 45 genera and 92 species; fungi: 13 genera and 24 species) (Supplementary Tables S1, S2) were found including pesticide degrading domains in the Ganga and Yamuna river sediments.

To investigate the evolutionary correlations among the identified bioremediation bacterial genome, a phylogenetic tree was constructed which showed the majority of the species are similar throughout the evolution. The phylogenetic tree of all the identified bioremediation species was configured in four different clusters. In Cluster 1, Pseudomonas fluorescens and Aeromonas hydrophila lies in one clade with the highest bootstrap value. The Bacillus tequilensis and Bacillus aerophilus are phylogenetically closely related in Cluster 3. The phylogenetic analysis based on the 16S rRNA gene of 181 type strains of Bacillus species and related taxa constituted nine clusters. The study revealed that Bacillus was not a monophyletic group. B. subtilis was in Group 1. Group 4 corresponds to thermophiles, Group 6 to halophilic or halotolerant bacilli, and Group 8 to alkaliphilic bacilli (Wang and Sun, 2009).

Relative abundance study showed that bioremediation bacteria of different genera were equally distributed among the three locations; however, few species dominate in one location over others, viz. Flavobacterium aquatile, Methylibium petroleiphilum, Pantoea agglomerans, Pseudomonas citronellolis were highly abundant in Kanpur location of river Ganga as compared to other locations. These location-specific changes of microbial diversity in the river sediments might be due to differential physicochemical properties and pollution level of the collected sediments. The present results also suggested that the relative abundance of bioremediation microbial communities was significantly correlated with polluted sites at river Ganga and Yamuna. The study revealed that polluted sites at New Delhi stretch of Yamuna and Kanpur stretch of Ganga were having a higher relative abundance of bioremediation bacteria and fungi as compared to less polluted Farakka sites.

Farakka stretch of the river Ganga was found to be less polluted (Samanta, 2013) whereas, the Kanpur stretch of Ganga and the New Delhi stretch of the river Yamuna were reported to be severely polluted due to the release of untreated metropolitan swages, factory effluents, etc. (Sehgal et al., 2012; Malik and Tauler, 2013). High concentrations of malathion (2.618 μg/l) and γ-HCH (0.259 μg/l) were detected in the surface water collected from the Ganga riverine ecosystem at Kanpur stretch. From the potable water samples in Delhi, Organochlorine pesticides, mainly isomers of hexachlorohexane, dichloro-diphenyl-trichloroethane, endrin, endosulfan, aldrin, dieldrin, and heptachlore were identified. It had been found α and β isomers of endosulfan residues in the Yamuna river (Agarwal et al., 2015). In the present investigation, the differential physicochemical parameters might be responsible for differences in the relative abundance of bioremediation microbes at different locations in the river Ganga and Yamuna. In the present study, a higher relative abundance of Bordetella petrii in Kanpur and Farakka was found as compared to the Yamuna. For that reason, it was postulated that pollution makes differences in the relative abundance of bioremediation bacteria species among different locations in the river Ganga and Yamuna. The β-diversity analysis also revealed that diversity indices of bioremediation bacteria and fungi of Farakka samples are different from the polluted sites of New Delhi.

Pesticide degrading bacteria identified in the present study are important as naturally occurring microbes can be manipulated to degrade highly toxic and carcinogenic compounds. Several microbes like Acephate degrading bacteria, Enterobacter asburiae, Bacillus cereus, Pantoea agglomerans was identified from sediments of the river Ganga and Yamuna, and these bacteria are reported earlier for pesticide decomposition by Ramya et al., 2016. Fipronil degrading Acinetobacter calcoaceticus and Acinetobacter oleivorans (Uniyal et al., 2016) were also found from the present sediment metagenomics at all locations. Pseudomonas putida and Rhodococcus sp. which degrade both the endosulfan isomers through oxidative and hydrolytic pathways were also found (Singh et al., 2017). Endosulfan degrading bacterial strains Pusillimonas sp. JW2 and Bordetella petrii NS were reported from polluted water and soil environments (Kong et al., 2018).

Bioremediation fungi Burkholderia jiangsuensis was reported to possess methyl parathion hydrolase (MPH) gene (bjmpd) (Liu et al., 2014). The present experiments found a significantly higher relative abundance of Burkholderia jiangsuensis in the Farakka stretch, whereas the same was not found at New Delhi stretch of river Yamuna. Cupriavidus necator was found to be involved in the degradation of aromatic compounds through 2, 3-dioxygenase pathway (Berezina et al., 2015). Cupriavidus necator was also found significantly higher quantities in Farakka locations in the present investigation. Highly competent chlorpyrifos degrading strain of Cupriavidus taiwanensis (Wang et al., 2015) was also found in higher proportion at Farakka locations.

Bioremediation fungi, Aspergillus terreus, found in the present study was reported to expresses Feruloyl esterases (FAEs) which is the key enzyme for biodegradation of lignocelluloses and hydrolysis of ester linkages between hemicellulose and lignin (Zhang et al., 2015). The study revealed that chlorfenvinphos could be degraded by Aspergillus fumigatus, Penicillium citrinum, Aspergillus terreus, and Trichoderma harzianum (Oliveira et al., 2015), and lots of these species were reported from the metagenomics samples. Phanerochaete chrysosporium, Aspergillus awamori, Aspergillus flavus, Trichoderma viride found to be involved in heavy metal bioremediation like Pb, Cd, Cr, and Ni (Joshi et al., 2011).

Aspergillus nidulans isolated from arsenic-contaminated soil was found to have the arsenic adsorption potential to be 84.35% (Maheswari and Murugesan, 2009). Aspergillus versicolor was reported to degrade antibiotic triclosan in simulated wastewater and semi-synthetic media (Tastan and Donmez, 2015). Exophiala xenobiotica was found to be dominant black yeast present in the car gasoline tanks, able to use toluene as a carbon source (Isola et al., 2013). Fusarium solani was reported to tolerate caffeine concentration of 8 g/L (Nayak et al., 2013). It is reported that Trametes versicolor (3000 UL^–1) could produce laccase and Pleurotus ostreatus (2700 UL^–1) showed the ability to degrade PAHs, phenanthrene (PHE) and pyrene (PYR) (Rosales et al., 2013). In our study, Aspergillus terreus, Trichoderma harzianum, and Trametes versicolor found higher quantities in Kanpur locations compared to Farakka and New Delhi stretches. Aspergillus versicolor and Fusarium solani were discovered in higher quantities in the New Delhi stretch of river Yamuna compared to the river Ganga.

PCA has been widely used to interpret significant bioremediation species from the experimental data set (Zitko, 1994; Vega et al., 1998; Shrestha and Kazama, 2007). The present analysis showed that the relative abundance of bacterial and fungal species of bioremediation potential is better correlated with the polluted sites of the river as compared to less polluted sites i.e., Farakka stretch. The study also revealed the location-wise clustering of data for Kanpur and New Delhi stretches of river Ganga and Yamuna due to a high correlation between pollution sites and relative abundance of microbes. Bioremediation fungal species showed a more prominent effect through a strong correlation with polluted sites viz. K2 and K3.

In the present metagenome, urea ABC transporter ATP binding protein Urtd gene in Alcaligenes faecalis, Bacillus cereus, Rhodococcus qingshengii, UrtE gene in Alcaligenes faecalis and Bacillus cereus bacteria were found in Farakka sample. In Kanpur stretch, UrtA gene was found in Ochrobactrum anthropi, UrtD, UrtE gene in Serratia marcescens. All members of this protein family ABC transporter ATP-binding subunits or gene clusters are expressed, producing the urea permease and connected with urea transport and metabolism. The role of these genes or domain in the bioremediation of urea can be supported by the previous finding (Valladares et al., 2002) where it was suggested that the involvement of a transporter of this superfamily in urea scavenging by some cyanobacteria in natural environments.

Another domain cd08419 which is a C-terminal substrate-binding of LysR-type transcriptional regulator (CbbR) of RubisCO operon, involved in carbon dioxide fixation was identified (Supplementary Table S3). The versatile function of various metal ion transport, amino acid transport, nickel, nitrate, phosphate, molybdenum, cobalt, Mn2⁺/Zn2⁺ transport, zinc/cadmium/mercury/lead-transporting ATPase, etc., found in the present metagenome were likely to play an active role in bioremediation process. In bacteria Pseudomonas putida, P450 was reported to be involved in Hexachlorobenzene and Pentachlorobenzene metabolism earlier (Scott et al., 2008). The Oxidoreductase gene was also reported in Pseudomonas sp. LBr, Agrobacterium strain T10 participates in Glyphosphate biodegradation (Scott et al., 2008). In the present study, enzymes from two different microbes, Burkholderia xenovorans and Flavobacterium aquatile were discovered which are actively involved in pesticide bioremediation.

The Metagenomic sequences have been submitted to the NCBI-SRA database under the Accession Nos: Kanpur Samples K1 (SRP190174), K2 (SRP190175), K3 (SRP189880), Farakka Samples, F1 (SRP191076), F2 (SRP191079), F3 (SRP191075), and New Delhi Samples, ND1 (SRP191073), ND2 (SRP191080), and ND3 (SRP191499) respectively.

This is the first report on the identification of bioremediation microbes in sediments of the river Ganga and Yamuna in India, using a metagenomic approach. The present study unraveled the in-depth insights on the relative abundance of important native bacteria and fungi convoluted in bioremediation in these rivers and their functional properties for possible biotechnological applications. The identified Nitrobacter humburgensis could reduce heavy metals like zinc, cadmium, mercury, and lead. Similarly, Rhodobacter sphaeroides could reduce arsenic, Shewanella putrefaciens for iron, Bacillus cereus for molybdenum and Alcaligenes faecalis for PAHs. The identified fungus like Aspergillus flaves could reduce lead, cadmium, chromium, and nickel, and Aspergillus nidulans for arsenic, etc., It was found that these bioremediation bacteria and fungi were more abundant in the sediments of a polluted stretch of the river. Several identified protein domains, which are actively involved in environmental bioremediation processes, could be explored for genome-scale engineering for industrial applications in the future.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.

BKB designed the metagenomics study and also collected samples for this research. HJC, BP and AKR performed bioinformatics data analysis. DJS and RR performed statistical analysis. BKB, BP, HJC, BD, AKR, DJS, PKP, ARR, AR, BKD, and TM wrote the manuscript. All authors participated in revision of the final manuscript for submission.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.