The cold desert of southeastern Himalayan region, Kargil district (KD), India (34° 32′ North latitude and 76° 07′ East longitude) was selected for soil sample collection at high altitude (10,500 ft above sea level). The annual temperature varies from −20 to 5°C. Compost (CM) was prepared from paddy straw collected from the premises of Division of Microbiology, IARI, New Delhi, India. At the time of collection, compost heap was in the thermophilic phase with temperature ranging between 60° and 65°C. Cow dung (CD) was collected from dairy farm situated in Todapour village, New Delhi, India. All samples were kept at 4°C during transportation and therafter at −20°C in the laboratory.

The soil samples from Kargil District (KD), Compost (CM), and Cow dung (CD) were analyzed based on utilization of carbon, nitrogen, phosphorus and sulfur substrate using the BIOLOG phenotypic microarray PM MicroPlates (Biolog, Inc., Hayward, California, USA). Phenotype MicroPlates PM1-2, PM3B and PM4A measure the carbon, nitrogen and phosphorus/sulfur metabolism, respectively. The samples were suspended in sterile saline water (0.8% w/v) and incubated under shaking condition (100 rpm) for 3 h at room temperature for proper mixing. The mixture was allowed to settle and aliquots were diluted in 15 ml of inoculation fluid (IF-0; Biolog) to adjust the optical density at 420 nm of 0.2 which were then used as the inoculum for the Microplates (Weber et al., 2007). For PM3B and PM4A plates, 15 ml inoculation fluid was added to the mixture of sodium succinate and ferric citrate as carbon sources. The 100 X additive solutions were prepared by dissolving 5.402 g of sodium succinate and 0.49 mg ferric citrate in sterile saline water. A total of 150 μl from the 100 X solution were added to the 15 ml of IF-0 inoculating fluid before preparing the sample suspensions. Thereafter, 100 μl of sample suspension was inoculated into each PM well. All PM plates were used in triplicate for each sample and were incubated at 30°C. After 72 h of incubation, each well was inoculated with 100 μl of the Biolog redox dye mix. For color development, the plates were incubated at 37°C until violet color appeared. Absorbance readings were taken at 590 nm with a plate reader (Micro Plate. Reader) (Weber et al., 2007).

The absorbance data were normalized to observe the average good color development (AWCD) as mentioned by (Garland, 1999) and further selected for analysis.

The normalized absorbance for each well was calculated as: AWCD=ΣODi/95

Where OD_i is the absorbance value from each well, corrected subtracting the blank well (A1) values from each plate well and 95 is the number of total substrates.

Substrate richness (S) in selected sample was calculated as the number of wells with a corrected absorbance greater than 0.25 (Siragusa et al., 2009).

The Shannon index (H) or “diversity” was calculated to understand the common ecological metric and the microbial communities shift over space and time. The Shannon–Weaver index was calculated as follows: H=−Σpiln(pi)

Where p_i is the ratio of the activity on each substrate (OD_i) to the sum of activities on all substrates ΣOD_i.

The phenotypic expression level and t-test based significance was considered by using Multiple array viewer (MeV 4.9.0) software.

A total of 247 glycosyl hydrolase Family 1 β-glucosidase amino acid sequences were retrieved from NCBI Genbank. Protein sequences were aligned using PROMALS3D (http://prodata.swmed.edu/promals3d/promals3d.php) to predict conserved regions and secondary structure based on available homologous 3D structures in database. All sequences were edited and trimmed to remove short, ambiguously aligned regions. The trimmed alignment was analyzed and graphical representation of conserved region was generated through weblogo3 interface (http://weblogo.threeplusone.com/create.cgi) (Crooks et al., 2004). Based on lowest degeneracy, equal GC content and Tm value, the following primers were designed: BGLTF (27): 5′ ACI YTI TAY CAY TGG GAY YTI CCI CAR 3′; BGLTR (27): 5′ RTA ICC YTC IGC CCA YTC RAA RTT RTC 3′; BGL21F (21): 5′ TAY CAY TGG GAY YTI CCI CAR 3′; BGL21R (21): 5′ RTA ICC YTC IGC CCA YTC RAA 3′; BGL17F (21): 5′ GCG CCT AYC AYT GGG AYY TIC C 3′; BGL17R (17): 5′ CCY TCI GCC CAY TCR AA 3′. Inosine (I) was used to decrease the degeneracy. The efficiency of these primer sets was confirmed by in silico PCR through a NCBI primer BLAST search with different microorganism name, microorganism group or taxonomy ID.

DNA extraction from soil sample of Kargil District (KD), Compost (CM), and Cow dung (CD) was performed using the Power Soil DNA Isolation Kit (MoBio Laboratories Inc., Carlsbad, USA) as per the manufacturer's instruction.

PCR amplification was performed using high fidelity Pfu DNA polymerase (Fermentas) with proofreading activity to avoid Taq polymerase based low replication fidelity. Reactions were carried out in a BioRed PCR system with metagenomic DNA as template. PCR condition was optimized for all parameters of PCR by gradient PCR (50–58°C), different MgCl₂ concentration (1–2.5 mM) and primer concentration (0.1–0.4 mM). The final PCR was conducted by using primers sets forward primer BGL17F (5′ GCG CCT AYC AYT GGG AYY TCC 3′) and reverse primer BGLTR (5′ RTA CCY TCG CCC AYT CRA ART TRT C 3′). The reaction mixture contained: 10 μl: 10X Pfu DNA polymerase buffer with MgCl₂, 0.3 mM: dNTP mix (10 mM), 0.1 μM each: Primers, 2 units: Pfu DNA polymerase, 10–50 ng: metagenomic DNA and MQ water make the final volume up to 100 μl. The PCR program was as follows: initial incubation at 95°C for 5 min, followed by 40 cycles (95°C for 50 s, 55°C for 1 min, and 72°C for 90 s), and then by a final extension at 72°C for 10 min. The amplified products were analyzed by electrophoresis at 60 V for 1 h in 1.2% agarose gel in 1X TAE buffer.

The pGEM-T Easy vector (Promega) was used to construct the three metagenomic libraries of partial β-glucosidase gene. Since, Pfu DNA polymerase is unable to add “A” overhang, “A” tail was generated in the PCR product. The PCR product was purified using the Qiaquick PCR purification kit (Qiagen, Valencia, CA, USA) as per manufacture's instruction. The purified PCR products of three environmental samples were first cloned in pGEM-T Easy vector. The pGEM-T and β-glucosidase construct was transformed into competent cells of E. coli strain (DH5α) and colonies were screened by blue/white colony selection on the LB plates containing X-gal (80 μg/ml), IPTG (0.5 mM) and ampicillin (100 mg/ml).

All positive white colonies were further inoculated to the 5 ml LB broth for plasmid isolation. QIAprep Spin Miniprep Kit (Qiagen, Valencia, CA, USA) was used for plasmid extraction as per manufacture's instruction. Clones containing putative β-glucosidase genes were screened by amplification of the isolated plasmids with M13F and M13R primers.

In order to avoid sequencing several identical β-glucosidase genes, Restriction Fragment Length Polymorphism (RFLP) analysis was performed by treating the PCR products (amplified with M13 primers) with four base pair-cleaving restriction endonuclease AluI (MBI Fermentas) at 37°C for 12 h, and the resulting electrophoretic mobility at 3% agarose was used to group the isolates.

The PCR products were sequenced by the Sanger sequencing method at Xcelris Labs Ltd, Ahmedabad, India. The obtained sequences were checked for chimera and homology were analyzed by the BLASTx search in the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov) and were submitted to NCBI Genbank. The nucleotide sequence was further translated into amino acid sequence (http://web.expasy.org/translate). All sequences of the same environmental niches were aligned using Clustal W version 1.8. The evolutionary distances were calculated and phylogenetic dendrogram was constructed by neighbor-joining method and tree topologies were evaluated by performing bootstrap analysis of 1000 data sets using MEGA 5.0 (Molecular Evolutionary Genetics Analysis).

The metabolic profiling of the indigenous microbial population of three samples from the Himalayan region, Kargil district (KD), India, Cow dung (CD) and Compost (CM) was analyzed by BIOLOG based substrate carbon (n = 190), nitrogen (n = 95), phosphorus (n = 59) and sulfur (n = 35) utilization pattern. Figure 1 shows clear differences among the substrate utilization patterns in microbial communities of diverse environmental samples, as evident from AWCD, richness, and diversity values. The P values were ≤ 0.05 which showed the significance of the experiment. For carbon and nitrogen sources, all the three AWCD and richness values were greater for Cow dung (CD), followed by CM and KD.

The normalized data were further utilized for hierarchal clustering to explore the possible linkages between substrate utilization patterns and three different samples. On the basis of 1.0 distance threshold, all 190 carbon sources were grouped into 5 major clusters (Supplementary Figure 1). Himalayan soil and cow dung sample were grouped closely which showed a common carbon utilization pattern. The clustering of nitrogen sources showed a very diverse utilization pattern, as only one major cluster was formed. The phosphorus and sulfur substrates were grouped into 3 clusters. According to the t-test based, on –log₁₀ (p) values, 86% (n = 164) carbon sources were utilized by the microbial community of all three samples, while the rest of the 14% (n = 26) registered non–significant utilization. Similarly, for nitrogen, phosphorus and sulfur substrates, 97, 90, and 86% substrates were utilized significantly by all the samples, respectively.

The alignment of 247 glycosyl hydrolase Family 1 β-glucosidase amino acid sequences was analyzed and graphical representation of conserved region was generated (Figure 2). The alignment showed characteristic features of GH1 with secondary structure of a classical (α/β)₈ sheets. The presence of two catalytic glutamate residues being 200 residues apart from each other and located at the C-terminal ends of β-strands 4 (acid/base) and 7 (nucleophile). Other characteristic identical and conserved residues are also shown in Figure 2. The peptide sequences [T/C]-[L/I/V/M/A]-[Y/F]-H-W-D-L-P-Q and D-[N/L]-[F/L]-[E/S]-W-[A/L/G/V/S]-[E/N/F/W/M/L]-G-[Y/I/E/L/F] were found to be the most conserved throughout the consensus sequence according to the maximum length of the peptide. Based on the maximum probability peptide sequence TLYHWDLPQ and DNFEWAEGY were selected to design forward and reverse degenerate primers, respectively. The selected region was pretended to amplify approx. 876 bp of PCR product.

The metagenomic DNA was extracted from all three samples and used as a template for amplifying the partial β-glucosidase gene. The synthesized primer sets were used to optimize the PCR condition for amplification of PCR products with the desired length. Gene fragments, about 876 bp for β-glucosidase gene, were amplified from the primer pair BGL17F/BGLTR (Supplementary Figure 2).

The purified PCR products of three environmental samples were separately cloned in pGEM-T Easy vector for metagenomic library construction. The positive clones were selected based on blue/white colony screening. After random confirmation of 80 clones from each library by PCR with primers M13F and M13R, a total of 60 clones was sequenced from three metagenomic libraries.

The identical clones were selected by Restriction Fragment Length Polymorphism (RFLP) analysis of 90 PCR products obtained by M13 primers from selected clones by restriction endonuclease AluI (Supplementary Figure 3). On the basis of restriction pattern, 20 clones were selected from each metagenomic library for further sequencing.

A total of 20 sequences of β-glucosidase genes were obtained from each metagenomic library. All sequences were identified as glycosyl hydrolase family 1 β-glucosidase (BGL) gene by BLASTx search and contained the two catalytic glutamic acid residues. The sequences were submitted to NCBI GenBank under accession numbers, KT898053- KT898072 for compost (CM), KT898073- KT898092 for cow dung (CD) and KT898093- KT898112 for Himalayan soil (KD) (Table 1). Based on the NCBI-BLASTp analysis, 20, 5, and 35% sequences had low identities (< 70%) with known GH1 family β-glucosidase sequences in GenBank from the Himalayan soil, cow dung and compost sample, respectively (Figure 3).

The phylogenetic diversity of deduced protein sequences was analyzed by building the unrooted protein-level phylogenetic tree for BGL sequence OTUs (Operational Taxonomic Units) using the 20 divergent sequences from each clone library with their respective neighbor sequences retrieved from NCBI GenBank (Figure 4). Metagenomic sequences from Himalayan soil (KD) showed similarities with 19 belonging to 6 groups, including Deinococcus, α-Proteobacteria, β-Proteobacteria, γ-Proteobacteria, the Bacteroidetes and Firmicutes. The Compost (CM) OTUs also showed 19 genera comprising 6 groups, including Chlamydia, Bacteroidetes, α-Proteobacteria, γ-Proteobacteria, Firmicutes and Actinobacteria. The Cow dung (CD) sample showed least diversity which arranged into 14 genera with only three groups - Bacteroidetes, Firmicutes and Actinobacteria (Figure 5).

The microbial metabolic activities of any particular environmental system depend on the utilization of specific nutrients. Hence, the functional diversity of environmental niches can be compared through the analyses of substrate utilization profiles (Torsvik and Øvreås, 2002). Phenotypic microarray provided by the BIOLOG may be the best way to compare the utilization of diverse group of essential substrates including carbon, nitrogen, phosphorus and sulfur (Stefanowicz, 2006). Specific substrate utilization patterns were generated by the microbial communities of three different environmental samples. The AWCD values showed that among all three samples, Himalayan soil showed least active utilization of substrates, as compared with two other samples. Cow dung and compost represent microbially induced active environmental systems as compared to the cold desert Himalayan soil. However, the diversity indices for carbon utilization was highest for the microbial community samples from the Himalayan soil, illustrating that microbial abundance is high, but not metabolically active. Carbon was utilized more efficiently by all the three environmental systems, as compared with other substrates. It is already reported that the microbial ecosystem is mainly governed based by the carbon cycle (Paerl and Pinckney, 1996). The carbon, phosphorus and sulfur substrates were grouped into 5, 3, and 3 clusters based on their combined utilization by all three samples (Supplementary Figure 1). Interestingly, nitrogen sources were utilized diversely by the microbial community and no visible clustering was observed even at 1.0 distance threshold. Nitrogen mineralization or transformation is carried out by different groups of microorganisms in environmental systems, which maybe responsible for lack of discernible groups in the different samples (Francis et al., 2007). This further ilustrates that nitrogen may not be a selective factor for distinguishing among the microbial populations from the three samples, although carbon based utilization pattern revealed that all three environments share selective and common metabolic profiles.

Potential producers of β-glucosidase were identified by the conventional culture based method, during previous investigations (Tiwari et al., 2014, 2015). Although, the cultural diversity is more relevant due to its applications, the unseen potential of β-glucosidase gene embedded in uncultured microbial population of diverse environmental niches needs exploration. Therefore, the GH1 family β-glucosidase protein sequences were searched extensively and based on the most conserved region, degenerate primers were designed. Although, the putative amplified region is not representative of the full length gene sequence, it encodes the sequence belongiong to the complete conserved domain sequence of the BglB superfamily.

The primer pair was used successfully to amplify the PCR product of desired size (approx. 876 bp) from all three environmental systems. Furthermore, the metagenomic library was constructed to perform the diversity and distribution studies. After the selection of positive clones, RFLP based screening and sequencing, twenty OTUs (Operational Taxonomic Units) from each metagenomic library were selected for diversity studies. Himalayan soil (KD), compost (CM) and cow dung (CD) was having 6, 11, and 10 novel OTUs based on the identities of ≤ 75% with known GH1 β-glucosidase protein sequences. The low identities implying that these β-glucosidases may be as yet undiscovered. In all three environmental niches, GH1 β-glucosidase was distributed in Bacteroidetes, Deinococcus, α, β, and γ-Proteobacteria, Firmicutes, Actinobacteria and Chlamydiae groups belonging to the total of 39 different genera. This result validated the potential of the designed degenerate primers and confirmed that primers can work over a broad range of GH1 β-glucosidases and facilitate future metagenomic studies of diverse ecological niches.

The highest diversity index and number of microbial groups were detected in the Himalayan soil sample. Himalayan samples were obtained from a cold desert where no microbial transformation was possible due to the extreme cold climate conditions; this was supported by the results of the phenotypic microarray experiment revealing lesser metabolic activity as compared with other samples. But, geographically the environment contained a hidden organic pool, as recently evidenced by Hood and co-workers (Hood et al., 2015). They informed that the mountain glaciers are storing high organic carbon content and releasing the carbon into the environment due to microbial transformation, as and when the ice shield melts due to the climate change. This concluded that cold desert environment is having low active, but high abundant microbial population.

The lowest diversity was displayed by cow dung samples, where only three groups, including Bacteroidetes, Firmicutes and Actinobacteria were present. The selective pressure on microbial load while passing through the animal gut may be the possible reason for this low diversity (Callaway et al., 2010).

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.