Next-generation DNA sequencing (NGS) has made it feasible to sequence large number of microbial genomes and advancements in computational biology have opened enormous opportunities to mine genome sequence data for novel genes and enzymes or their sources. In the present communication in silico mining of microbial genomes has been carried out to find novel sources of nitrilases. The sequences selected were analyzed for homology and considered for designing motifs. The manually designed motifs based on amino acid sequences of nitrilases were used to screen 2000 microbial genomes (translated to proteomes). This resulted in identification of one hundred thirty-eight putative/hypothetical sequences which could potentially code for nitrilase activity. In vitro validation of nine predicted sources of nitrilases was done for nitrile/cyanide hydrolyzing activity. Out of nine predicted nitrilases, Gluconacetobacter diazotrophicus, Sphingopyxis alaskensis, Saccharomonospora viridis, and Shimwellia blattae were specific for aliphatic nitriles, whereas nitrilases from Geodermatophilus obscurus, Nocardiopsis dassonvillei, Runella slithyformis, and Streptomyces albus possessed activity for aromatic nitriles. Flavobacterium indicum was specific towards potassium cyanide (KCN) which revealed the presence of nitrilase homolog, that is, cyanide dihydratase with no activity for either aliphatic, aromatic, or aryl nitriles. The present study reports the novel sources of nitrilases and cyanide dihydratase which were not reported hitherto by in silico or in vitro studies.

1. Introduction

Advancement in the DNA sequencing technologies has led to sequencing of large number of genomes and the enormous sequence data are available in the public domain. The fourth-generation DNA sequencing has made it possible to sequence a bacterial genome within a few hours at a reasonably low cost [14]. As of today 5293 prokaryotic and 22 eukaryotic genomes have been completely sequenced and the sequence data are easily accessible in databases such as NCBI, GOLD, and IMG/ER. It is evident from previous studies that not all the gene/protein sequences in the databases are functionally characterized, which make these repositories a rich source for the discovery of novel genes and proteins [5, 6]. Genome mining has emerged as an alternate approach to find novel sources of desired genes/proteins as the conventional screening methods which involve isolation of microbes and their screening for desired products are time consuming, tedious, and cost intensive [7, 8].

Microbial nitrilases are considered to be the most important enzymes in the nitrilase superfamily that find application in the synthesis of fine chemicals, production of some important acids, and drug intermediates and in green chemistry [913]. Besides their wide applications nitrilases are prone to certain limitations, for example, their inactivation or inhibition by the acidic product, extremes of pH, temperature, and organic solvent [14, 15]. These limitations are being addressed either by the isolation of microorganisms from the extreme habitat or by enrichment techniques for specific substrate using conventional microbiological procedures [6] prone to limitation as mentioned above. The present communication focuses on in silico screening of publicly available bacterial genomes for nitrilase genes and in vitro validation of the predicted novel sources of nitrilases.

2. Material and Methods

2.1. Genome Screening Using Homology and Motif Based Approach

Primary screening of microbial genomes (data given as supplementary material in Supplementary Material available online at https://doi.org/10.1155/2017/7039245) was done using homology based approach. Tblastn and blastp were used to screen the sequenced genomes with query sequence to identify the presence and position of similar genes in the genome. Computationally predicted proteins from the bacterial genomes with keyword “nitrilase/cyanide dihydratase” were also downloaded using advanced search options in the IMG/ER database. Sequences with low (30%) and high similarity (80%) were discarded. Nitrilase gene in contigs showing the presence of nitrilase homologs was downloaded from IMG/ER. GenMark S tool was used to predict the ORFs in each contig, and the output was downloaded selecting protein sequence as output option. Amino acid sequences less than 100 amino acids were considered to be as false positive (FP) and were discarded. Small amino acid sequence database was created which was further subjected to local blast, to confirm the presence of nitrilase homolog in the contigs of the individual genome.

On the other hand, protein based manually designed motifs (MDMs) were used to screen the bacterial genome to search for the presence of conserved motifs using MAST (Motif Alignment and Search Tool) at MEME (Multiple Em for Motif Elicitation) suite. The motifs used are already described in our previous communication [12]. Motifs identified in sequences less than hundred amino acids were rejected, considered to be false positive (FP). Sequences above 100 amino acids were taken to be as true positive (TP).

2.2. Study of Physiochemical Properties and Phylogenetic Analysis of Predicted Nitrilases

Physiochemical data of the in silico predicted nitrilases were generated from the ProtParam software using ExPASy server and compared to the values deduced from the previous nitrilase study [16]. Some important physiochemical properties such as number of amino acids, molecular weight (kda), isoelectric point (pI), computing pI/Mw and the atomic compositions, values of instability index, aliphatic index, and grand average of hydropathicity (GRAVY) were calculated. A comparative chart was drawn between previously characterized and predicted nitrilases.

An output file of multiple aligned sequences using Clustal W for both previously characterized and predicted nitrilases was used to generate the Neighbor Joining (NJ) tree using MEGA 6 version. Phylogenetic tree was generated in order to predict the sequences as aliphatic or aromatic with previously characterized nitrilases.

2.3. Nitrilase Activity Assay

Culture of some of the bacteria predicted to have nitrilase gene (Shimwellia blattae, Runella slithyformis, Geodermatophilus obscurus, Nocardiopsis dassonvillei, Streptomyces albus, Flavobacterium indicum, Saccharomonospora viridis, Sphingopyxis alaskensis, and Gluconacetobacter diazotrophicus) was procured from Microbial Type Culture Collection (MTCC); Chandigarh Escherichia coli BL21 (DE3) from Invitrogen was used as negative control as this organism does not have nitrilase gene. These cultures were grown in the laboratory using different media (Table 1) for the production of nitrilase activity following the procedures described earlier [1719]. Nitrilase activity was assayed in 1.0 mL reaction mixture containing nitrile as substrate (1–10 mM) and 0.1 mL resting cells. After 30 min of incubation at 30°C the reaction was quenched with 0.1 M HCl and the amount of ammonia released was estimated using nitrilase assay, that is, modified phenate-hypochlorite method described by Dennett and Blamey [20]. One unit of nitrilase activity was defined as the amount of enzyme required to release 1 μmole of ammonia per min under the assay conditions.

3. Results

3.1. Genome Screening Using Conserved Motifs and Homology Search

As many as 138 candidate sequences were identified using tblastn and blastp at IMG/ER on both gene and protein level. Identification of potentially coding nitrilase genes was done using homology based approach (blastp and tblastn) allowing the identification of nitrilase sequences. To identify newer sources of nitrilases, candidate sequences bearing unassigned functions (hypothetical, uncharacterized, or putative) were selected from the translated genomes (Table 2). The identified sequences shared 30–50% sequence identity to biochemically characterized Rhodococcus rhodochrous J1 nitrilase which was taken as query sequence. Catalytic residues were found to be conserved in all the predicted proteins. Nine predicted and translated sequences were further chosen for their in silico and in vitro validation based on the manually designed motifs (MDMs) (Tables 3 and 4) identified from previous study [12].

3.2. Physiochemical Parameters and Phylogenetic Analysis

In silico identified nitrilases were analyzed for their physiochemical properties using ProtParam, an online tool at the ExPASy proteomic server. The selected candidates values for various properties were found to be very much similar to those with earlier published data by Sharma and Bhalla [16] as mentioned in Table 5. Average values deduced for aliphatic and aromatic nitrilases from earlier characterized proteins were taken as standard for the comparison of a predicted set of nitrilase. The values of the same were found to be very much similar to those with earlier published data by Sharma and Bhalla [16] as mentioned in Table 5. The total number of amino acids ranged from 260 amino acids (Nocardiopsis dassonvillei) to 342 amino acids (Shimwellia blattae) with different molecular weight. Isoelectric point ranged between 4.8 and 5.8 which is found to be closer to the consensus value, that is, the average data value from previously characterized aliphatic or aromatic nitrilases.

Neighbor Joining (NJ) tree using MEGA 6 shows the phylogenetic analysis with in silico predicted sequences from completely sequenced microbial genomes with that of previously characterized nitrilase sequences. They were distinguished either as aliphatic or aromatic according to their position in the phylogenetic tree (Figure 1).

3.3. In Vitro Validation of Some In Silico Predicted Nitrilases

To validate for nitrile transforming activity of nine predicted novel sources of nitrilases, these were tested against common aliphatic, aromatic, and aryl nitriles and potassium cyanide (KCN). Gluconacetobacter diazotrophicus, Sphingopyxis alaskensis, Saccharomonospora viridis, and Shimwellia blattae were found to be more specific for aliphatic nitriles. On the other hand, Geodermatophilus obscurus, Nocardiopsis dassonvillei, Runella slithyformis, and Streptomyces albus exhibited nitrilase activity for aromatic nitriles. Flavobacterium indicum was the only organism which showed no activity for either aliphatic, aromatic, or aryl nitriles but was specific towards the degradation of the potassium cyanide (KCN) (Table 6). On the other hand, negative control, that is, Escherichia coli BL21 (DE3), showed no activity for any of the nitriles/substrates tested.

4. Discussion

Annotation of sequenced genomes to identify new genes has become integral part of the research in bioinformatics [2124]. The present investigation has revealed some novel sources of nitrilases. Homology and conserved motif approach screened microbial genomes and proteins predicted as nitrilase or cyanide dihydratase or carbon-nitrogen hydrolase in 138 prokaryotic bacterial genomes. Manually designed motifs (MDMs) also differentiated the in silico predicted nitrilases as aliphatic or aromatic [12] as the designed motifs are class specific. All the four motifs identified were uniformly conserved throughout the two sets of aliphatic and aromatic nitrilases as mentioned in Table 4.

The sequences belonged to the nitrilase superfamily, showing the presence of the catalytic triad Glu (E), Lys (K), and Cys (C) to be conserved throughout. Phylogenetic analysis using the MEGA 6.0 version for the aliphatic and aromatic set of protein sequences revealed two major clusters. Neighbor Joining (NJ) tree used for phylogenetic analysis revealed that in silico predicted proteins (this study) and previously identified nitrilases as aliphatic and aromatic [16] were found to be grouped in their respective clusters (Figure 1).

Aliphaticity and aromaticity of in silico predicted and characterized nitrilases were differentiated based on their physiochemical properties. The physicochemical properties of the predicted set of nitrilase were deduced using the ProtParam subroutine of Expert Protein Analysis System (ExPASy) from the proteomic server of the Swiss Institute of Bioinformatics (SIB), in order to predict aromaticity or aliphaticity. Several of the parameters (number of amino acids, molecular weight, number of negatively charged residues, extinction coefficients, and grand average of hydropathicity) listed in Table 5 are closer to the consensus values reported for aromatic and aliphatic nitrilases, supporting that the predicted set of nitrilase has aromatic or aliphatic substrate specificity (Table 5).

In silico predictions were verified by in vitro validation of the predicted proteins. Common nitriles (aliphatic, aromatic, and aryl nitriles) and potassium cyanide (KCN) were tested to check for the nitrile/cyanide transforming ability of the predicted proteins. Out of nine predicted proteins eight were found active for different nitriles, whereas Flavobacterium indicum was found to hydrolyze toxic cyanide (KCN) into nontoxic form (Table 6). The present approach contributed to finding novel sources of desired nitrilase from microbial genome database.

5. Conclusion

Genome mining for novel sources of nitrilases has predicted 138 sources for nitrilases. In vitro validation of the selected nine predicted sources of nitrilases for nitrile/cyanide hydrolyzing activity has furthered the scope of genome mining approaches for the discovery of novel sources of enzymes.

Conflicts of Interest

The authors declare that they have no conflicts of interest.


The authors are thankful to the Department of Biotechnology (DBT), New Delhi, for the continuous support to the Bioinformatics Centre, Himachal Pradesh University, Summer Hill, Shimla, India.

Supplementary Materials

List of organisms with completely sequenced genomes avaliable at NCBI and IMG/ER.

  1. Supplementary Material