Angiogenin (ANG) protein is a member of ribonuclease family induced by other angiogenic factors such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (aFGF), acid fibroblast growth factor (bFGF) and epidermal growth factor (EGF) [1]. ANG plays important role in number of pathogenic conditions including cancer and neurodegenerative diseases by modulating cell growth and survival properties [2]. In contrast to being up regulated in cancer, ANG is down regulated in neurodegenerative diseases such as amyotrophic lateral sclerosis, Parkinson and Alzheimer disease [3-5]. Role of ANG in neurodegenerative diseases have been documented due to function lose mutations found in patients [6,7]. ANG also facilitates the production of tiRNA (tRNA-derived stress induced RNA) [8,9] which suppress overall protein translation. tiRNA is important in stress mechanisms under adverse environments utilized by cell [10]. ANG and RNASE4 transcription is regulated by RNA polymerase III and a CCCTC binding factor (CTCF)-depending intragenic chromatin loop [11]. In malignancies, ANG is important proangiogenic and its higher expression is associated with a non-nuclear Maspin expression [12].

Studies have shown the critical role of ANG in tumor microenvironment for angiogenesis [13]. It was reported that ANG targets p53 and B-cell lymphoma 2 to accelerate cancer development through anti-apoptotic effects [14]. Sheng et al. explained a mechanism by which ANG stimulates rRNA transcription through an epigenetic activation of rDNA promotor [15]. Under androgen stimulation, ANG undergoes nuclear translocation in androgen-dependent prostate cancer cells, where it binds to rDNA promoter and stimulates rRNA transcription [16]. Ribosomal protein is mediated by ANG [2]. The P13k/Akt/mTOR signaling pathways is thought to be a central mediator in signal transduction pathways, which is frequently activated in different cancer types [17]. It was reported that ANG activates nitric oxide synthase (NOS) by interacting with cellular nucleus. Similarly NOS activity is stopped by blocking the P13k/Akt kinase signaling transduction cascade showing the importance of this pathway and ANG for NOS activity [18]. The experiments suggest that cross talk between ANG and protein kinase B/Akt signaling pathways could be essential for ANG-induced angiogenesis in vitro and in vivo [2,4].

ANG protein plays vital role in different pathways of human. All the reported mutations in this gene were selected for this study and using different bioinformatics and computational biology approaches analyzed their structure and their impacts and effects on its structure, stability and functionality.

There was a case study conducted for the investigation of MEN1 protein family but on low scale [19]. Therefore, until there is no such type of case study conducted on ANG. This case study will reveal facts about disease causing mutations and describing the information about disease severity with respect to mutation in ANG. This study will be helpful to conduct research on angiogenin at clinical scale to investigate which mutation leads to severity of disease. This case study also may helpful for mutation targeting in drug designing and helping in finding the mutation with neutral effect.

Interaction and appearance of ANG genes

GeneMANIA is a user friendly online tool for the prediction of gene function. It classifies the functionally similar genes to a query gene list by using accessible proteomics and genomics data.

Mutational screening

Already reported mutation in ANG i.e. F12S, P20S, Q36L, Y38H, K41E, D46G, S52N, R55K, C63W, K64I, I70V, K84E, P136L, V137I and H138R were selected for this study. The selected mutation sequences were retrieved from UNIPROT [20]. In “Pathology and Biotech” section of UNIPROT database, data was available, showing that either the mutation is indel or substitution and also describes mutation involvement in disease.

Structural and functional impact prediction of SNPs

Effect of single amino acid mutation was studied by PredictSNP server which is used to predict either mutation is deleterious or neutral with percentage of expected accuracy [21]. The surface accessibility of selected mutations was examined by NetSurfp. NetSurfp not only predicts the accessible surface area and secondary structure of amino acids but also check the reliability for each prediction, in the form of a Z-score. The Z-score is related to the surface prediction and not the secondary structure. I-Mutant2.0 [22] is a support vector machine based web server. It was used for automatic prediction of protein stability based on single site mutation. It can predict the protein stability changes (more stability or de-stability) corresponding all possible mutations of a residue. I-Mutant2.0 gave DDG value (free energy change value). PROVEAN [23] (Protein Variation Effect Analyzer) was used to predict whether the single amino acid substitution, deletion or insertion affect the function of protein. It is very useful for analyzing protein variants and for identifying functionally important nonsynonymous mutations.

PolyPhen is a web server used for the prediction of probable impression of amino acid substitutions on the function and stability of human proteins via evolutionary comparisons. It provides functional annotations of single-nucleotide polymorphisms (SNPs) on the basis of gene transcripts maps coding, damaging effects, abstract protein sequences and attributes of structures [24].

Mutation3D is an online tool used for functional analysis and visualization of amino acid substitutions on three-dimensional structure of protein. This tool is intended to be utilized for the identification of groups of amino acid substitutions appearing from somatic cancer mutations across several patients in direction to determine fuel downstream suggestions and functional hotspots [25].

Protein 3d modeling and detection of hydrogen bonding and clashes

MODELLER was used to predict the 3D protein structure by using homology modeling. The end user offers a sequence alignment to be predicted with identified correlated structures, it automatically computes a model comprising all non-hydrogen atoms and implements homology protein structure modeling by consummation of spatial restraints [26]. Chimera was used to scan the 3D (three-dimensional) structure of specific protein and hence, changed the original amino acid with the mutated one to grasp the impact of the mutation. The outcome was graphic model depicting the mutation [27]. It is also used for localization of the mutated region and describing mutation in coil region are helix region and show hydrogen bonds, clashes and contacts of the mutant residue.

Disorder tendency analysis

PONDR is online available software used to predict disordered region in protein. Protein disorders were identified by individual predictor tool available at PONDR. PONDR-VLXT, PONDR-VL3 and PONDR-VSL2 were used for long as well as short protein disorder region of ANG [28]. For the confirmation of selected mutations in conserved region of ANG, Conserved Domain analysis of all selected mutations was carried out using CDD [29].

Interaction and appearance of ANG genes

GENEMANIA analysis revealed that ANG has active role in endoribonuclease activity, RNA catabolic process, endonuclease activity, Nucleic acid phosphodiester bond hydrolysis and ribonuclease activity. ANG gene possess physical interactions with RNH1 (ribonuclease/angiogenin inhibitor 1), RNASE2 (ribonuclease A family member 2) and RNASE7 (ribonuclease A family member 7) with 67.64% confidence score. ANG co-expressed and co-localized with ADH6 (alcohol dehydrogenase 6), ASGR2 (asialoglycoprotein receptor 2), (SERPINA1) serpin family A member 1, SERPINA3 (serpin family A member 3), GC (GC vitamin D binding protein) and HGD (homogentisate 1, 2-dioxygenase). RNASE-10 (ribonuclease A family member 10), RNASE12 (ribonuclease A family member 12), PTEN (phosphatase and tensin homolog), RNASE3 (ribonuclease A family member 3), RNASE7 (ribonuclease A family member 7) and RNASE8 (ribonuclease A family member 8) sharing common protein domains are shown in Figure 1.

Figure 1: The interaction of ANG gene with other gene families and elaborating its major function involvement.

Structural impact prediction of SNPs

Netsurp analysis revealed that the surface accessibility of mutated ANG was different as compared to wild-type ANG. Relative Surface Accessibility index (RSA) of a protein from 0-1 is a measure of residue solvent exposure. Higher RSA value depicts higher surface accessibility of protein. Comparison between RSA wild-type and RSA mutant is shown in Figures 2 and 3. Surface Accessibilities of F12S, Y38H, K41E, S52N, K64I, I70V, P136L, V137I and H138R were decreasing while P20S, D46G, R55K, C63W and K84E were increasing. PredictSNP server revealed that F12S, Q36L, C63W, K64I, P136L, V137I and H138R were deleterious with>70% expected accuracy while P20S, Y38H, K41E, D46G, S52N, R55K, I70V and K84E were neutral shown in Figure 4. I-Mutant protein stability prediction server provided the Free Energy change value (DDG) varying from positive to negative numbers. Positive values show that stability of protein is increasing whereas negative values confirm the decreasing stability of protein. DDG value of Q36L and H138R were positive while DDG values of all the other mutations were negative representing the decreased stability of protein with single amino acid substitution. Similarly, the results obtained from PROVEAN suggested that F12S, P20S, Q36L, Y38H, C63W, K64I, P136L and H138R were deleterious with -4.003, -2.990, -6.500, -3.179, -10.690, -7.995 and -7.449 (score higher than threshold -2.5) while K41E, D46G, S52N, R55K, I70V. K84E and V137I were found neutral with scores -1.003, -2.491, -0.957, 0.220, -0.476, -1.103 and -0.986 respectively (score less than threshold -2.5).

Figure 2: The comparison between RSA values of both wild-type and mutated.

Figure 3: The Free energy change value DDG by I-Mutant Server. DDG value of Q36L and H138R showing positive value and remaining showing less than zero values that representing the stability of mutated protein decreasing with single amino acid substitution.

Figure 4: Single amino acid substitution is either deleterious or neutral with expected accuracy in percentage.

Polyphen analysis revealed that F12S, P20S, Q36L, C63W, K64I, P136L, V137I and H138R have probably damaging effect on protein while Y38H, K41E, D46G, S52N, R55K, I70V and K84E have low damaging effect and considered as Benign shown in Table 1. Score ranges from 0-1, if score is higher than 0.5 then mutation has probably damaging effect and lower score revealed that mutation is Benign.

Table 1: showing the SNPs impact on protein function. Score rages from 0-1, if score is higher than 0.5 then mutation has damaging effect and lower the score reveal that mutation is Benign.

Mutation 3D analysis revealed that mutated amino acid was present either in cluster region or not. Red color sphere represents that mutation in cluster region while blue color sphere represents presence of mutated amino acid in covered region of angiogenin. Mutation in cluster region is more sensitive as compared to covered region. The results revealed that mutations i.e. Q36L, Y38H, I70V, P136L, V137I and H138R were in cluster region while remaining mutations were present in covered region of protein except F12S and P20S. All the mutations were present in RNase A Domain from amino acid 26 to 140.

Protein 3d modeling and detection of hydrogen bonding and clashes

Angiogenin Protein 3D both wild-type and mutated models were predicted with the help of homology modelling by using MODELLER. Each mutation model represented almost 100% accuracy/confidence score. The Qmean score of predicted models of all mutations were in the range of 0-1, which signifies reliability of the mutated model data shown in Table 2. Mutated angiogenin model's evaluation by Procheck for stereo chemical quality showed that structures predicted by MODELLER are the best quality models and all models have>90% of residues in favored regions. These models were also validated using Swiss model and ProsA-web which gave a negative Z-score value representing a good quality of predicted models. Higher the negative value higher the quality of predicted models. RMSD values are calculated using Chimera, where RMSD value closest to zero shows the maximum deviation of mutated and normal angiogenin structure shown in Table 2. Chimera investigation manually checked whether the selected mutation was present in coiled region or helix region. In mutated models F12S, P20S, K41E, D46G, C63W and K64I mutation were found in coiled region while Q36L, Y38H, S52N, R55K, I70V, K84E, P136L, V137I and H138R were in helix region as shown in Figures 5 and 6. The mutation in helix region of protein is more sensitive than coiled region. UCSF Chimera exposed hydrogen bonds, clashes and contacts of the mutant residue as shown in Figure 7.

Figure 5: The selected mutation presence either in cluster region or non-cluster region.

Figure 6: The disorder tendency of mutants. Higher the tendency of disorder touching the value to 0.5, lower the value means disorder tendency with low confidence score.

Figure 7: Selected mutation present in coiled region or helix region.

Table 2: Representing the predicted mutated model assessment values.

Disorder tendency analysis

PONDR result showed disorder tendency by using three different disorder predictors as shown in Figure 6. Analysis of Conserved Domain of Angiogenin Protein CDD confirmed a domain of Pancreatic Ribonuclease A from amino acid to 26 to 140 with 2.24e-55. Pancreatic ribonuclease A possesses many Catalytic sites but mutation containing catalytic sites are from amino acid 33 to 38, 64 to 72 and 136 to 140. Disorder Tendency was checked which revealed the disorder tendency value as shown in Figure 6 higher than 0.5 describing the probability of disorder causing in protein.

This study hypothesized that selected missense mutations in angiogenin family Q36L, C63W, K64I, P136L, V137R and H138R have functional, structural and conformational effects with high confidence score in wild- type protein. Furthermore, these mutations i.e. Q36L, C63W, K64I, P136L, V137R and H138R disturb the stability and surface accessibility of protein. Consequently, these mutations are pathogenic in a high rate of disorder tendency, while F12S, P20S, Y38H, K41E, D46G, S52N, R55K, I70V and K84E also pathogenic to wild-type angiogenin, disturbing the structure and function but with low confidence score. These investigations may prove supportive to conduct clinical work on ANG family and to find out which mutation is responsible for the disease. This case study could be helpful for mutation targeting in drug designing and finding the mutation with neutral effect.