Predicting the functional and structural consequences of nsSNPs in human methionine synthase gene using computational tools

ABSTRACT

Methionine synthase encoded by the MTR gene is one of the key enzymes involved in the SAM (S- Adenosyl Methionine) cycle catalyzing the conversion of homocysteine to methionine. Methionine plays an important role in the DNA, RNA, protein, phospholipids, and neurotransmitters methylation. It also maintains serum homocysteine level and indirectly regulates de novo nucleotide synthesis and repair. The current study predicted the functional consequences of nsSNPs in human MTR gene using SIFT, PolyPhen2, PROVEAN, SNAP2, PMut, nsSNPAnalyzer, PhD-SNP, SNPs&GO, I-Mutant, MuPro, and iPTREE-STAB. The PTM sites within the protein were predicted using ModPred and the phylogenetic conservations of amino acids & conserved domains of protein were predicted using ConSurf and NCBI conserved domain search tool respectively. The protein 3D structure was generated using SPARKS-X and analyzed using RAMPAGE. Structural deviation was analyzed using TM-Score. STRING analysis was preformed to predict protein-protein interactions. D621G, G682D, V744L, V766E, and R1027W were predicted to be the most deleterious nsSNPs in MTR. R1027 was predicted to having the three PTM sites and G682 & V744 were predicted as highly conserved residues. D621G, G682D, V744L, V776E, and R1027W were predicted to be within conserved domains of methionine synthase. The G682D, V744L, V776E,  and R1027W were predicted to alter protein 3D structure. STRING predicted that methionine synthase interacting with 10 different proteins. The present study predicted D621G, G682D, V744L, V766E, and R1027W as functionally and structurally significant nsSNPs in human MTR gene. The present study can provide the significant information for further experimental analysis.

Abbreviations: cblG: methylcobalamin deficiency G; MTR: 5-methyl tetrahydrofolate-homocysteine methyl transferase; MS: methionine synthase; SAM: S-adenosyl methionine; nsSNPs: non-synonymous single nucleotide polymorphisms; OMIM: online mendelian inheritance in man; NCBI: national center for biological information; SIFT: sorting intolerant from tolerant; PolyPhen2: polymorphism phenotyping 2; PROVEAN: protein variation effect analyzer; SNPs&GO: single nucleotide polymorphisms and gene ontology; PhD-SNP: predictor of human deleterious single nucleotide polymorphisms; RI: reliability index; PTM: post translational modification; SPDBV: Swiss PDB viewer; PDB: protein data bank; RMSD: root mean square deviation; STRING: search tool for the retrieval of interacting proteins

KEYWORDS:

• Computational analysis
• human MTR gene polymorphism
• single nucleotide polymorphisms
• folate metabolism
• deleterious prediction

Introduction

Methionine synthase (MS) in human is encoded by the gene, 5-methyl tetrahydrofolate-homocysteine methyl transferase (MTR), located on chromosomal location 1q43. Human MTR gene contains 33 exons while the human methionine synthase protein contains 1265 amino acid residues (Leclerc et al. 1996; Li et al. 1996; Watkins et al. 2002). Methionine synthase (MS) is one of the key enzymes involved in the SAM (S- Adenosyl Methionine) cycle catalyzing the conversion of homocysteine to methionine. Regeneration of methionine is carried out in two steps in a ping pong reaction in which the methyl group from 5-Methly tetrahydrofolate first transferred to Cobalamin (Vitamin B12) by converting 5-methly tetrahydrofolate to tetrahydrofolate. Then methyl Cobalamin transfers its methyl group to homocysteine to regenerate methionine in SAM cycle (Matthews et al. 2003; Wolthers and Scrutton 2007). Methionine in the SAM cycle subsequently produces SAM which is a very important source of active methyl group in many biological reactions including DNA, RNA, protein, phospholipids, and neurotransmitters (Mattson and Shea 2003; de Jonge et al. 2009; Blom and Smulders 2011; Desai and Chauhan 2016b). By this means methionine synthase plays an essential role in perpetuation of SAM cycle, even without continuous influx of methionine and helps in maintaining serum homocysteine level (Li et al. 1996). Since it is one of the few enzymes that use 5-methyl tetrahydrofolate as a substrate, the 5-methyl tetrahydrofolate level is also indirectly maintained by methionine synthase. The 5-methyl tetrahydrofolate is very important for de novo generation of thymine (Hesse and Hoefgen 2003). Methionine synthase is important for maintaining methionine level and preventing homocysteine accumulation. Increased blood homocysteine level have been associated with increased risk of developing cardiovascular disease, infertility, late pregnancy complications such as pre-eclampsia, abruption placenta, intrauterine growth retardation, preterm birth & intrauterine fetal death, birth defects, neural tube defects, Down syndrome, and development of some types of cancer (Eskes 2000; Nelen 2001; Carmel and Jacobsen 2002; Hague 2003; Steegers-Theunissen et al. 2004; Tamura and Picciano 2006; Farcas et al. 2009; Li et al. 2015).

Figure 3. rs57752780 (V744L) SNP structure analysis by SPDBV: H-bonding interactions in native (left) show interactions with Met740 (2.958 Å), Leu747 (2.948 Å), and Ile748 (2.820 Å); while mutant (right) show interactions with Met740 (2.971 Å), Leu747 (2.901 Å), and Ile748 (2.824 Å).

Figure 4. rs74710714 (V776E) SNP structure analysis by SPDBV: H-bonding interactions in native (left) show interactions with Gly828 (2.741 Å) and Ile826 (3.072 Å); while mutant (right) show interactions with Ala819 (3.029 Å), Ile826 (3.069 Å), and Gly828 (2.721 Å).

Single nucleotide polymorphisms (SNPs) are the variations in the single nucleotide occurring more frequently than 1% at a specific position in the genome. The SNPs can be within coding or non coding regions or it can be within intergenic regions of the human genome. Among them the non synonymous SNPs within the coding region can have the major impact on the phenotype as it leads to the alterations in the amino acid residue in the corresponding protein and plays a major role in biological variations (Carninci et al. 2005; Datta et al. 2015). Reports suggested that >50% of the polymorphisms leading to the inherited genetic disorders are due to nsSNPs (Ramensky et al. 2002; Datta et al. 2015). Thus, analyzing the association between the genetic mutations and their phenotypic impact can be helpful in understanding the genetic causes of various complex inherited diseases (Datta et al. 2015; Patel et al. 2015; Desai and Chauhan 2016a, 2017a, 2018).

To date the deleterious nsSNPs of the human MTR gene have not been predicted using in silico analysis. Hence, the present study explored the understanding of the association between the genetic variations and its phenotypic effect using the in silico analysis. The current study analyzed the functional consequences of the nsSNPs using SIFT, PolyPhen2, PROVEAN, SNAP2, PMut, nsSNPAnalyzer, PhD-SNP, and SNPs&GO. The impact of the nsSNPs on the protein stability was analyzed using I-Mutant, MuPro and iPTREE-STAB. The post translational modification sites within the protein were analyzed using ModPred. The Phylogenetic conservation of amino acid residues and the conserved domains of the human methionine synthase protein were analyzed using ConSurf and NCBI conserved domain search tool respectively. The 3D structure of native and mutant protein was generated using SPARKS-X and analyzed using RAMPAGE Ramachandran plot analysis. The Swiss PDB Viewer (SPDBV) was used to analyze the 3D structure of the protein as well as to locate the SNPs on the protein structure. STRING analysis was performed to know protein–protein interactions.

Results

SNP retrieval

The NCBI-dbSNP showed a total of 7,224 SNPs reported in the human MTR gene. Out of which only 445 are missense SNPs and were used for the further analysis. The remaining SNPs includes 5,930 intronic, 10 in 3ʹ splice site, 486 in 3ʹ UTR, 9 in 5ʹ splice site, 329 in 5ʹ UTR, 229 in coding synonymous, 29 were frame shift and 17 nonsense.

Prediction of functional consequences of nsSNPs of human MTR gene

Out of 445 SNPs subjected for the prediction of functional consequences using SIFT server, only 6 (V229I, D621G, G682D, V744L, V776E and R1027W) predicted damaging, 11 predicted tolerated and rest were not reported in SIFT. All the 17 nsSNPs predicted either damaging or tolerated by SIFT were further validated using PolyPhen2, PROVEAN, SNAP2 and PMut. PolyPhen2 server predicted 4 SNPs (G682D, V744L, V776E, and R1027W) to be probably damaging and the remaining 13 to be benign (Table 1). PROVEAN predicted 8 nsSNPs (R52Q, P62L, D621G, G682D, V744L, V776E, D919G, and R1027W) to be deleterious and rest 9 to be neutral. SNAP2 predicted 7 SNPs (R52Q, D621G, G682D, V744L, V776E, D919G, and R1027W) having effect on protein function and remaining 10 as neutral while PMut predicted 4 SNPs (D621G, D919G, G939R, and R1027W) as pathological and rest 13 as neutral (Table 2).

Table 1. Prediction of damaging nsSNPs of human MTR gene.

		SIFT		PolyPhen2
SNPs	Variants	Prediction	Score	Prediction	Score	Sensitivity	Specificity
rs1274958	R52Q	Tolerated	0.08	Benign	0.287	0.91	0.89
rs61736442	P62L	Tolerated	0.08	Benign	0.003	0.98	0.44
rs61736441	V229T	Damaging	0.04	Benign	0.149	0.92	0.87
rs74767314	T279P	Tolerated	0.29	Benign	0.001	0.99	0.15
rs116252762	D314N	Tolerated	0.22	Benign	0.234	0.91	0.88
rs61736440	D621G	Damaging	0.01	Benign	0.159	0.92	0.87
rs113914406	G682D	Damaging	0.00	Probably damaging	1.000	0.00	1.00
rs57752780	V744L	Damaging	0.00	Probably damaging	1.000	0.00	1.00
rs60599724	G745A	Tolerated	0.93	Benign	0.005	0.97	0.74
rs74710714	V776E	Damaging	0.00	Probably damaging	1.000	0.00	1.00
rs1805087	D919G	Tolerated	0.09	Benign	0.159	0.92	0.87
rs113042166	G939R	Tolerated	0.40	Benign	0.000	1.00	0.00
rs116836001	R1027W	Damaging	0.02	Probably damaging	0.981	0.75	0.96
rs111465715	S1073P	Tolerated	0.14	Benign	0.007	0.96	0.75
rs61736326	R1164H	Tolerated	0.07	Benign	0.001	0.99	0.15
rs61739582	N1222S	Tolerated	0.24	Benign	0.008	0.96	0.76
rs11799647	I1259V	Tolerated	0.69	Benign	0.004	0.97	0.59

Table 2. Prediction of functional nsSNPs of human MTR gene.

	PROVEAN		SNAP2			PMut
Variants	Prediction (cutoff = −2.5)	PROVEAN score	Prediction	Score	Accuracy (%)	Prediction	Reliability
R52Q	Deleterious	−3.543	Effect	82	91	Neutral	0
P62L	Deleterious	−5.056	Neutral	−15	57	Neutral	3
V229T	Neutral	−0.893	Neutral	−94	97	Neutral	7
T279P	Neutral	−0.422	Neutral	−87	93	Neutral	7
D314N	Neutral	−2.304	Neutral	−8	53	Neutral	4
D621G	Deleterious	−5.292	Effect	67	80	Pathological	4
G682D	Deleterious	−6.767	Effect	71	85	Neutral	0
V744L	Deleterious	−3.000	Effect	58	75	Neutral	1
G745A	Neutral	−0.381	Neutral	−80	87	Neutral	6
V776E	Deleterious	−5.443	Effect	83	91	Neutral	5
D919G	Deleterious	−3.871	Effect	3	53	Pathological	5
G939R	Neutral	2.058	Neutral	−83	93	Pathological	0
R1027W	Deleterious	−3.421	Effect	69	80	Pathological	8
S1073P	Neutral	−0.701	Neutral	−85	93	Neutral	7
R1164H	Neutral	2.135	Neutral	−55	78	Neutral	5
N1222S	Neutral	−2.435	Neutral	−32	66	Neutral	2
I1259V	Neutral	−0.248	Neutral	−61	82	Neutral	7

Prediction of disease associated nsSNPs

All the 17 nsSNPs of human MTR gene predicted using SIFT server were further analyzed using nsSNPAnalyzer, PhD-SNP, and SNPs&GO. Total 5 SNPs (P62L, G682D, V776E, R1027W, and N1222S) showed disease association and rest 12 predicted neutral by nsSNPAnalyzer server. Of the 17 nsSNPs analyzed using PhD-SNP, 9 were identified (R52Q, T279P, D621G, G682D,  V744L, V776E, R1027W, S1073P and N1222S) as disease associated and 8 were neutral, while SNPs&GO identified 5 (R52Q, D621G, G682D, V744L and V776E) nsSNPs as disease associated and 12 as neutral (Table 3).

Table 3. Prediction of disease associated nsSNPs of human MTR gene.

	nsSNPAnalyzer		SNPs&GO		PhD-SNP
Variants	Phenotype	Sift_Score	Prediction	RI	Prediction	RI
R52Q	Neutral	0.23	Disease	2	Disease	5
P62L	Disease	0.03	Neutral	1	Neutral	1
V229T	Neutral	1.00	Neutral	5	Neutral	6
T279P	Neutral	0.40	Neutral	2	Disease	0
D314N	Neutral	0.24	Neutral	2	Neutral	3
D621G	Neutral	0.08	Disease	3	Disease	7
G682D	Disease	0.00	Disease	9	Disease	6
V744L	Neutral	0.00	Disease	6	Disease	0
G745A	Neutral	1.00	Neutral	7	Neutral	9
V776E	Disease	0.00	Disease	9	Disease	3
D919G	Neutral	0.41	Neutral	5	Neutral	6
G939R	Neutral	0.64	Neutral	7	Neutral	5
R1027W	Disease	0.02	Neutral	3	Disease	1
S1073P	Neutral	0.11	Neutral	6	Disease	7
R1164H	Neutral	0.05	Neutral	5	Neutral	1
N1222S	Disease	0.01	Neutral	5	Disease	6
I1259V	Neutral	1.00	Neutral	9	Neutral	8

Prediction of protein stability

The impact of all the 17 nsSNPs on the protein stability was analyzed using I-Mutant, MuPro and iPTREE-STAB. All the nsSNPs were found to decrease protein stability except S1073P by I-Mutant. Similarly, all the nsSNPs were predicted decreasing protein stability by iPTREE-STAB server except T279P & R1027W. While, MuPro server analyzed five SNPs (T279P, D314N, G682D, G745A and V776E) as increasing protein stability and rests were decreasing protein stability (Table 4).

Table 4. Prediction of change in protein stability due to nsSNPs in human MTR gene.

	I-Mutant			MuPro		iPTREE-STAB
Variants	Stability	RI	DDG	Effect	Confident Score	The descreminated direction of thermal stability change	Predicted value of thermal stability change (kcal/mol)
R52Q	Decrease	8	−0.94	Decrease	−0.69343781	Negative (destabilizing)	0.5236
P62L	Decrease	5	−0.87	Decrease	−0.15284842	Negative (destabilizing)	1.4567
V229T	Decrease	9	−2.64	Decrease	−1	Negative (destabilizing)	−0.5370
T279P	Decrease	6	−1.59	Increase	0.62683938	Positive (Stabilizing)	−3.8413
D314N	Decrease	4	−0.29	Increase	0.049654273	Negative (destabilizing)	0.2320
D621G	Decrease	7	−1.00	Decrease	−0.62120061	Negative (destabilizing)	−1.5851
G682D	Decrease	8	−1.92	Increase	0.36835608	Negative (destabilizing)	−0.9729
V744L	Decrease	2	−0.58	Decrease	−0.084190515	Negative (destabilizing)	−0.1300
G745A	Decrease	3	−0.16	Increase	0.503335	Negative (destabilizing)	−0.1405
V776E	Decrease	7	−1.40	Increase	0.015835468	Negative (destabilizing)	−2.5073
D919G	Decrease	1	−0.91	Decrease	−1	Negative (destabilizing)	−3.3860
G939R	Decrease	4	−0.38	Decrease	−0.28149996	Negative (destabilizing)	−1.4050
R1027W	Decrease	7	−0.53	Decrease	−0.84324914	Positive (Stabilizing)	0.0475
S1073P	Increase	5	−0.29	Decrease	−0.31719131	Negative (destabilizing)	0.0697
R1164H	Decrease	9	−1.30	Decrease	−0.62633143	Negative (destabilizing)	−1.0376
N1222S	Decrease	6	−1.74	Decrease	−0.09579365	Negative (destabilizing)	0.9422
I1259V	Decrease	8	−1.34	Decrease	−0.50886901	Negative (destabilizing)	−1.4700

The prediction of nsSNPs by all 11 tools (SIFT, PolyPhen2, PROVEAN, SNPA2, PMut, nsSNPAnalyzer, SNPs&GO, PhD-SNP, I-Mutant, MuPro and iPTREE-STAB) are shown in Figure 1. The five nsSNPs (D621G, G682D, V744L, V776E, and R1027W) found the functionally most significant nsSNPs.

Figure 1. The number of in silico tools (SIFT, PolyPhen2, PROVEAN, SNAP2, PMut, nsSNPAnalyzer, SNPs&GO, PhD-SNP, I-Mutant, MuPro and iPTREE-STAB) predicted various SNPs as deleterious. The functionally most significant SNPs predicted were D621G, G682D, V744L, V776E, and R1027W.

Prediction of post-translational modification sites

Serine at 1073 was predicted to having site for O-linked glycosylation. Amino acids R52, D314, D919 and R1164 were predicted having site for proteolytic cleavage. Amino acid arginine at 1027 was predicted to having sites for ADP-ribosylation, methylation, and proteolytic cleavage.

Conservation of amino acids

The amino acid arginine at protein position 52 and glycine at position 682 were predicted to be highly conserved (score 9), buried and structural residues while V229, V744, and G745 were predicted highly conserved (score 9), exposed and functional residues. R1164 and N1222 were also predicted as highly conserved (score 8), exposed and functional residues (Table 5).

Table 5. Phylogenetic conservation of amino acid residue of human methionine synthase.

Sr. No.	Position	Sequence	Score	Conservation Score	b/e	Function
1	52	R	−0.831	9	b	s
2	62	P	−0.164	6	b
3	229	V	−0.873	9	e	f
4	279	T	0.483	3	e
5	314	D	0.451	3	e
6	621	D	−0.044	5	e
7	682	G	−0.907	9	b	s
8	744	V	−0.924	9	e	f
9	745	G	−0.812	9	e	f
10	776	V	−0.307	6	b
11	919	D	0.389	3	e
12	939	G	1.571	1	b
13	1027	R	1.268	1	e
14	1073	S	0.816	1	e
15	1164	R	−0.640	8	e	f
16	1222	N	−0.602	8	e	f
17	1259	I	0.152	4	b

Sr. No.: serial number; b: buried; e: exposed, s: structural residue; f: functional residue.

Prediction of the conserved domains in the protein

The human methionine synthase protein (NP_000245) consists of 1265 amino acids. The four conserved domains were predicted in the human methionine synthase protein. The N terminal domain, cobalamin-dependent methionine synthase I (homocysteine binding and methyl transferase activity), consisting of 21st to 340th amino acid residues. The second MeTr subgroup of pterin-binding enzyme domain is (folate binding domain) between 371st and 627th amino acid residues. The third domain from 671st to 897th amino acid residues having B12 binding activity and the last C terminal domain is vitamin B12-dependent methionine synthase activation domain containing 965th to 1246th amino acids.

Protein 3D structure prediction, validation, and molecular dynamics

The 3D structure of the native and mutant (D621G, G682D, V744L, V776E, and R1027W) methionine synthase protein was generated using SPARKS-X web server. It gave the 10 best structures out of which the first structure with highest Z-score generated using 4cczA as template was selected for the further analysis. The 4cczA is a crystal structure of human 5-methyltetrahydrofolate-homocysteine methyl transferase with homocysteine and folate-binding domain and showing 100% identity and 100% positives with human MS protein.

The 3D structure of native and mutant (D621G, G682D, V744L, V776E, and R1027W) proteins generated by SPARKS-X was further analyzed using RAMPAGE Ramachandran plot analysis. The root-mean-square deviation (RMSD) values for all five mutant models (D621G, G682D, V744L, V776E, and R1027W) were calculated using TM-Score server. The Z score value, results of RAMPAGE analysis and RMSD of mutant and native protein are given in Table 6.

Table 6. Z-Score of the protein structures and results of RAMPAGE & TM-Score analysis.

	SPARKS-X	RAMPAGE			TM-Score server
Variants	Z-Score	Favored region (%)	Allowed region (%)	Outliner region (%)	TM-Score	RMSD (Å)
Native	27.42	1200 (95.0)	51 (4.0)	12 (1.0)	–	–
D621G	27.46	1209 (95.7)	42 (3.3)	12 (1.0)	0.5665	0.45
G682D	27.23	1208 (95.6)	42 (3.3)	13 (1.0)	0.5174	0.43
V744L	27.41	1198 (95.0)	48 (3.8)	15 (1.2)	0.5559	0.45
V776E	27.43	1205 (95.4)	43 (3.4)	15 (1.2)	0.5427	0.44
R1027W	27.39	1216 (96.3)	33 (2.6)	14 (1.1)	0.5171	0.42

The 3D structures of native and mutant protein were analyzed using Swiss PDB Viewer (SPDBV). Structural analysis of native and all five mutant proteins suggested significant alteration in H-bonding interactions of amino acids in four mutants (G682D, V744L, V776E, and R1027W) as compared to native (Figures 2–5). In native protein, Gly 682 showed H-bond of 2.750 and 3.071 with A678 and L679 respectively. In mutant protein, Asp 682 showed H-bond of 2.825, 3.072, and 2.895 with A678, L679 and R738 respectively. In native protein, Val 744 showed H-bond of 2.958, 2.948, and 2.820 with M740, L747 and I748 respectively. In mutant protein, Leu at 744 showed H-bond of 2.971, 2.901, and 2.824 with M740, L747, and I748, respectively. In native protein, Val 776 showed H-bond of 3.072 and 2.741 with I826 and G828 respectively. In mutant protein, Gly at 776 showed H-bond of 3.029, 3.069, and 2.721 with A819, I826, and G828, respectively. In native protein, Arg at 1027 showed H-bond of 2.932 with C1104. In mutant protein, Arg at 1027 showed H-bond of 2.942 with C1104. The D621G didn’t show any alteration in the H-bonding in native and mutant.

Figure 2. rs113914406 (G682D) SNP structure analysis by SPDBV: H-bonding interactions in native (left) show interactions with Ala678 (2.750 Å) and Leu679 (3.071 Å); while mutant (right) show interactions with Ala678 (2.825 Å), Leu679 (3.072 Å), and Arg738 (2.895 Å).

Figure 5. rs116836001 (R1027W) SNP structure analysis by SPDBV: H-bonding interactions in native (left) show interactions with Cys1104 (2.932 Å); while mutant (right) show interactions with Cys1104 (2.942 Å).

The graphical representation of nsSNP location in different protein domain and summary of deleterious prediction of human methionine synthase protein encoded by MTR gene is shown Figure 6.

Figure 6. The graphical representation of SNP location in different protein domain and summary of deleterious prediction of human MTR gene.

Protein–protein interactions prediction

STRING database was used for the prediction of functional interaction between the methionine synthase and other proteins in the cell. STRING results predicted the functional association of methionine synthase protein with AHCY (adenosylhomocysteinase), CBS (cystathionine beta synthase), CTH (cystathionine gamma lyase), MAT1A (methionine adenosyltransferase I), MAT2A (methionine adenosyltransferase II), MTHFD1 (methylene tetrahydrofolate dehydrogenase I), MTHFD1L (methylene tetrahydrofolate dehydrogenase I like), MTHFR (methylene tetrahydrofolate reductase), MTRR (5-methyl tetrahydrofolate-homocysteine methyl transferase reductase or MSR: methionine synthase reductase) and SHMT1 (serine hydroxymethyltransferase I) (Figure 7).

Figure 7. Methionine synthase protein-protein interactions network shown by STRING.

Discussion

The MTR gene encodes a key enzyme involved in the folate/homocysteine metabolism pathway, methionine synthase. To date more than 7,000 SNPs are reported in human MTR gene in NCBI-dbSNP, all may not have deleterious impact on protein function or structure. Non-synonymous SNPs are the most common form of the mutation which affects the protein biological function by altering the amino acids of the encoded protein. Most of the nsSNPs of the human MTR gene are still uncharacterized for their potential to cause disease. Hence, the present study was conducted to predict the impact of the nsSNPs of human MTR gene on the protein biological function, stability and structure using various bioinformatic tools. The methods used in the current study provide the clues on the effect of variations at molecular level by means of the various aspects as well as the parameters describing pathogenicity of particular amino acid substitution. Different algorithms use different sets of the sequences for the alignments; hence, the prediction capabilities may differ for every method. To predict the pathogenic effect of nsSNPs using single bioinformatic tool may not be reliable (Abdelraheem et al. 2016). Hence, the consequences of the nsSNPs were analyzed using multiple tools. In the present study, five different in silico prediction tools (SIFT, PolyPhen2, PROVEAN, SNAP2 and PMut) were used to predict the functional effect of nsSNPs of MTR gene. Three different tools were used (nsSNPAnalyzer, SNPs&GO and PhD-SNP) to predict disease-causing nsSNP. Three different computational algorithms (I-Mutant, MuPro and iPTREE-STAB) were used to predict the nsSNPs-affecting protein stability. Out of 445 missense mutations subjected for SIFT, only 6 (V229I, D621G, G682D, V744L, V776E, and R1027W) were predicted damaging and among them five SNPs (D621G, G682D, V744L, V776E and R1027W) were predicted deleterious by at least three tools out of four (PolyPhen2, PROVEAN, SNAP2 and PMut) used for the prediction of functional consequences of nsSNPs. Moreover, these five nsSNPs were predicted disease causing by at least two tools out of three (nsSNPAnalyzer, SNPs&GO, and PhD-SNP) used for the prediction of disease-associated SNPs. D621G, G682D, V744L, V776E, and R1027W were also predicted to decrease the protein stability. By comparing outputs of 11 different tools (SIFT, PolyPhen2, PROVEAN, SNAP2, PMut, nsSNPAnalyzer, SNPs&GO, PhD-SNP, I-Mutant, MuPro, and iPTREE-STAB) D621G, G682D, V744L, V776E, and R1027W were suggested as functionally significant nsSNPs.

These five nsSNPs were further analyzed functionally and structurally by seven different in silico prediction tools; ModPred, ConSurf web server, NCBI conserved domain search tool, SPARKS-X, RAMPAGE, TM-Score, and Swiss PDB Viewer. ModPred used to predict the post translation modification (PTM) sites in methionine synthase protein, which predicted three different PTM sites (ADP-ribosylation, methylation and proteolytic cleavage) at R1027. The results of the ConSurf suggested that G682 and V744 are highly conserved (conservation score: 9) & exposed and hence, the important functional residues. The results of NCBI conserved domain search tool indicates that D621G presents within MeTr subgroup of pterin binding enzyme. G682D, V744L, and V776E present within the B12 binding domain of methionine synthase suggesting that these may affect the interaction of methionine synthase with vitamin B12. R1027W presents within the vitamin B12-dependent methionine synthase activation domain indicating that this may affect the methionine synthase activation. To check the effect of MTR gene mutations on protein structural and binding interaction with vitamin B12, the 3D structures of native as well as all five mutants (D621G, G682D, V744L, V776E, and R1027W) were generated using SPARKS-X and analyzed using RAMPAGE. The results of RAMPGE Ramachandran plot analysis suggested the good quality of protein 3D models and hence, used for the further analysis. The prediction of TM-Score suggested the structural deviations of all five mutant (D621G, G682D, V744L, V776E, and R1027W) proteins as compared to native. Further, structural analysis of native and mutant proteins also suggested alteration in H-bonding interactions in mutants (G682D, V744L, V776E, and R1027W) as compared to native residues.

The one SNP i.e. D919G (MTR c.2756A>G) has been extensively studied (Brandalize et al. 2010; Zampieri et al. 2012; Coppedè et al. 2013; Victorino et al. 2014; Li et al. 2015; Sukla et al. 2015; Moustafa et al. 2016; Desai and Chauhan 2017b). However, the results were contradictory in different studies. By means of in silico deleterious prediction done in the present study, the D919G (MTR c.2756A>G) was not predicted highly deleterious. The D919 was predicted as a variable residue with conservation score of 3. However, the D919 was predicted to be involved in proteolytic cleavage during post translational modification. To the best of our knowledge, none of the study showed the genetic risk of D621G, G682D, V744L, V776E, and R1027W with any known disease condition. Watkins and co-workers amplified each of the 33 exons of MTR gene from genomic DNA from a panel of 21 patients with methylcobalamin deficiency G (cblG) and identified 13 novel mutations among which 5 were missense (A410P, S437Y, S450H, H595P, and I804T) and two recurrent mutations P1173L and H920D (Watkins et al. 2002). However, none of missense SNPs identified in their study was predicted in the present study.

Moreover, the present study also predicted the interactions of methionine synthase with other proteins. The results predicted the association of MTR with AHCY, CBS, CTH, MAT1A, MAT2A, MTHFD1, MTHFD1L, MTHFR, MTRR (or MSR), and SHMT1. Experimental analysis by Wolthers and Scrutton (2007) also observed the association of methionine synthase activation domain with methionine synthase reductase (MSR) and mutagenesis L1071 and K987 in protein activation domain weakens this interaction. Similarly, other polymorphisms (like R1027W, S1073P, R1164H, and N1222S) activation domain may also weakens this interaction.

The present study concluded that D621G, G682D, V744L, V776E and R1027W were predicted the most deleterious and functionally significant nsSNPs in human MTR gene. R1027 predicted to be having the post translational modification sites while G682 and V744 were predicted the highly conserved residues. G682D, V744L, V776E, and R1027W also analyzed to alter protein structure. The present study can provide the significant information for further experimental analysis.

Materials and methods

SNP retrieval

The data on the human MTR gene was retrieved from the web-based data sources like Online Mendelian Inheritance in Man (OMIM: http://www.ncbi.nlm.gov/omim) and the National Center for Biological Information (NCBI: http://www.ncbi.nlm.gov/). The rs IDs of nsSNPs and FASTA amino acid sequences of the human MTR gene (NP_000245) were obtained from the NCBI.

Prediction of functional consequences of nsSNPs of human MTR gene

The functionally deleterious nsSNPs of human MTR gene were predicted using five different computational tools namely SIFT (Sorting Intolerant from Tolerant; http://sift.jcvi.org/), PolyPhen2 (Polymorphism Phenotyping 2; http://genetics.bwh.harvard.edu/pph2/), PROVEAN (Protein Variation Effect Analyzer; http://provean.jcvi.org/index.php), SNAP2 (https://rostlab.org/services/snap/) and PMut (http://mmb.pcb.ub.es/pmut2005/). All are web-based tools predicting the functional consequences of amino acid substitution. SIFT classifies the nsSNPs into damaging and tolerated (Ng and Henikoff 2006). PolyPhen classifies them into probably damaging, possibly damaging, and benign (Ramensky et al. 2002). PROVEAN classifies the nsSNPs into deleterious and neutral (Choi et al. 2012). SNAP2 classifies the nsSNPs into effect and neutral (Yachdav et al. 2014). PMut classifies them into pathological and neutral variants (Ferrer-Costa et al. 2005).

Prediction of disease associated nsSNPs

All the nsSNPs of human MTR gene were further analyzed using nsSNPAnalyzer, SNPs&GO and PhD-SNP to differentiate the disease-associated nsSNPs. The nsSNP Analyzer (http://snpanalyzer.uthsc.edu/), SNPs&GO (single nucleotide Polymorphism Database& Gene Ontology; http://snps-and-go.biocomp.unibo.it/snps-and-go/), and PhD-SNP (Predictor of human deleterious single nucleotide polymorphisms; http://snps.biofold.org/phd-snp/phd-snp.html) are the web based tools which predict the disease associated variants among the protein (Bowie et al. 1991; Ng and Henikoff 2001; Thomas et al. 2003; Capriotti et al. 2006; Calabrese et al. 2009).

Prediction of change in the protein stability

The impact of all the single nucleotide variations of human MTR gene on the protein stability was analyzed using three different tools: I-Mutant, MuPro, and iPTREE-STAB. I-Mutant 2.0 (http://folding.biofold.org/cgi-bin/i-mutant2.0), MUpro (http://mupro.proteomics.ics.uci.edu/) and iPTREE-STAB (http://210.60.98.19/IPTREEr/iptree.htm) are the web based tool, which predicts the change in the protein stability upon the point mutation. I-Mutant and MuPro classifies the nsSNPs into decrease or increase protein stability (Capriotti et al. 2005; Cheng et al. 2006), while iPTREE-STAB classifies them into negative (destabilizing) or positive (stabilizing) nsSNPs (Huang et al. 2007).

Prediction of post-translational modification sites

The post translational modification sites within the human methionine synthase protein sequence were predicted using ModPred (http://www.modpred.org/) server. It consists of 34 ensembles of logistic regression models that are trained separately with a cluster of 126,036 non-redundant experimentally confirmed sites for the 23 different polymorphisms and achieved from ad-hoc literature as well as public databases (Pejaver et al. 2014).

Phylogenetic conservation of amino acid residues

The phylogenic conservation of the amino acid residues in the human Methionine Synthase protein was analyzed using ConSurf (http://ConSurf.tau.ac.il/) server. It calculates the conservation score (ranging from 1 to 9) for each amino acid residue. The score 1–3 indicates variable residues, 4–6 indicates average conserved residues, and 7–9 indicates highly conserved residues (Ashkenazy et al. 2010).

Prediction of the conserved domains in the protein

The conserved domains of the human MS protein were analyzed using the NCBI Conserved Domain Search tool (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). It uses the FASTA amino acid sequence of the particular protein as a query to predict their conserved domains and motifs (Marchler-Bauer et al. 2015).

Protein 3D structure prediction, validation, and molecular dynamics

The 3D structure of native and mutant protein was generated using SPARKS-X (http://sparks-lab.org/yueyang/server/SPARKS-X/) server (Yang et al. 2011). This was further analyzed by Ramachandran plot analysis using RAMPAGE (http://mordred.bioc.cam.ac.uk/~rapper/rampage.php) (Lovell et al. 2002). TM-Score (http://zhanglab.ccmb.med.umich.edu/TM-score/) was used to check similarities between native and mutant models (Xu and Zhang 2010). SPDBV was used to visualize the 3D structure of the native and mutant proteins and to locate nsSNPs on protein structure.

Protein–protein interactions prediction

The interactions of methionine synthase with other proteins were predicted using STRING (search tool for the retrieval of interacting proteins; https://string-db.org/). The STRING predicts protein–protein interaction by means of, either direct or indirect, associations among a known protein and other proteins by utilizing its database of 5,214,234 proteins of 1113 organisms (Szklarczyk et al. 2011). For STRING prediction methionine synthase and Homo sapiens were used as the input.

Disclosure of interest

Authors declares that they have no conflict of interest

Acknowledgments

The authors are very thankful to Charutar Vidya Mandal, Vallabh Vidyanagar and SICART, Vallabh Vidyanagar for providing platform for present study. We are also very thankful to Director, Ashok and Rita Patel Institute of Integrated Studies and Research in Biotechnology and Allied Sciences (ARIBAS) for providing facilities, technical guidance, motivation, and valuable suggestions during the research work. INSPIRE division of Department of Science and Technology (DST, New Delhi) thankfully acknowledged for providing the fellowship to Ph.D scholar, Mansi Desai.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Notes on contributors

Mansi Desai

Performed methodology, Study design, data analysis, and manuscript writing: MD; Contributed to study design, data interpretation and manuscript revision: JBC.