Modeling and Docking Study of GABA-AT Protein in Mycobacterium Tuberculosis -  A Computational Approach

 

Sriroopreddy R¹, Raghuraman.P¹, Reena Rajkumari. B²*

¹Department of Biotechnology, School of Bio Sciences and Technology, VIT University, Vellore – 632014, Tamil Nadu, India.

²Department of Integrative Biology, School of Bio Sciences and Technology, VIT University, Vellore – 632014, Tamil Nadu, India.

 

ABSTRACT:

Tuberculosis, a transmittable killer disease caused by Mycobacterium tuberculosis whose pathogenesis is poorly understood. The treatment available for the anti-mycobacterial activity has poor observance because of the uprising of the drug resistance strains resulted by the prolonged treatment giving the chance to the microbe to develop resistance. Hence, there is an urgent need to identify novel compounds from marine sources which are having antimicrobial activity. GABA-AT is an important enzyme involved in GABA shunt pathway of TCA cycle which has a role in virulence of Mycobacterium tuberculosis as it is a PLP dependent enzyme. In this present study, the 3D structure of GABA-AT was generated by using SWISSMODEL and validated using various methods. The model generated was validated and secondary structural analysis was predicted using YASPIN. Thereafter, the active site was predicted and further used the residues for docking. The acquired model was docked with the known inhibitors of GABA-AT and marine compounds having anti-tuberculosis activity. The comparative analysis of protein ligand interaction between known inhibitor and marine compounds were studied.  Among them, Heteronemin has the best binding energy (-9.19Kcal/mol) which indicates that the marine compound can be used for inhibiting the activity of GABA-AT in Mycobacterium tuberculosis and can be proposed as valuable drug candidate.

 

 

1.INTRODUCTION:

Tuberculosis (TB) is one of the major diseases which has been known to human kind from ages and has been prevailing till date. In every year, Mycobacterium tuberculosis kills around 2 million people, and has been declared as a global emergency in 1993.¹ As per now, there are many vaccines such as BCG’s available in market for tuberculosis. Also, many drugs are being developed targeting many enzymes associated with this disease condition. Mycobacterium tuberculosis can develop resistance for the available drugs, called resistance TB.

 

So now-a-days the medication is been given as a combination of two or more drugs for TB treatment, which leads harmful side effects to the victim. 4 aminobutyrate aminotransferase (GABA-AT) is acts as an essential enzyme for the survival of mycobacterium tuberculosis in the host as it involves amino acid degradation pathway. This pathway also known as GABA shunt pathway is part of a TCA cycle which bypasses the glutamate into the cytosol.² GABA –AT is found to be functional only in the presence of the pyridoxal phosphate. Hence, our current study focuses on the GABA-AT inhibition in mycobacterium tuberculosis.

 

A comparative study between the reported inhibitors and few marine compounds were performed to determine efficient novel drug. The known inhibitors such as, Gabaculine³, Phenelzine, Phenylethylidenehydrazine (PEH), Rosmaric acid, Vigabatrin, Valproic acid⁴ and Amino oxyacetic acid which were reported previously in literature. The selected marine compounds such as, Bipinnapterolide, Heteronemin, Hymenidin, Litosterol, Nephalsterol B, Parguesterol A, Parguestrol B and Puupehenone have an anti TB activity as reported in invitrostudies.⁵

 

In our study, we have demonstrated the importance of the 4- aminobutyrate aminotransferase in structural and functional aspect to reveal the insights of secondary structural elements and targeted binding sites in GABA protein of Mycobacterium tuberculosis. This study might be helpful for development of targeted drug therapyin tuberculosis.

 

2. MATERIALS AND METHODS:

2.1 Homology modelling:

The amino acid sequence of 4 aminobutyrate aminotransferase (ID : P9WQ79) was retrieved from the UNIPROT database (http://www.uniprot.org/). The target protein is of 449 amino acids long. The target sequence was subjected for homologous template identification using the BLASTp, an online program.⁶ The search performed using the target protein against the proteins in the protein data bank. The best similarity match with the template identity of 83% (PDB ID: 3R4T). The 3D model structure was done using SWISS-MODELLER (http://swissmodel.expasy.org/).⁷ The 3D models for GABA-AT were obtained and each of the models was subjected to QMEAN Server for further analysis for model quality.

 

2.2 Model validation and Energy minimization:

The final model validation was done using PROCHECK, VERIFY 3D, WHAT IF and ERRAT by SAVES (Structural Analysis And Verification Server) (http://services.mbi.ucla.edu/SAVES/), PROSA .The energy minimization was done by the NOMAD REF (http://lorentz.immstr.pasteur.fr/nomad-ref.php).

 

2.3 Secondary structure prediction:

The secondary structure of the protein was predicted for its structural pattern in terms of helix, sheets and coils. The secondary structure prediction was done using YASPIN server (http://zeus.few.vu.nl/).⁸ The result of the secondary structure prediction provides a brief idea about the folding pattern of the sequence.

 

2.4 Selection of ligands:

Gabaculine, Phenelzine, Phenylethlidenehydrazine, Rosmarinic acid, Valproic acid and Vigabatrin were inhibits the function of 4 aminobutyrate aminotransferase has been reported in literature. The selected marine compounds such as, Bipinnapterolide, Heteronemin, Hymendin, Litosterol, Nephasterol, Parguestrol A, Parguestrol B and Puuphenone. The structures of those compounds were obtained from Pubchem database (https://pubchem.ncbi.nlm. nih.gov/compound) and their structural class has been described in Table 1.

 

Table 1.Marine compounds having inhibitory activity against Mycobacterium tuberculosis and their structural class.

Sr. No	Marine Compounds	Structural class
1.	Bipinnapterolide	Shikimate-sesquiterpene
2.	Heteronemin	Sesterterpene
3.	Hymenidin	Sesterterpene
4.	Litosterol	Steroid
5.	Nephalsterol B	Steroid
6.	Parguestrol A	Steroid
7.	Parguestrol B	Steroid
8.	Puuphenone	Shikimate-sesquiterpene

 

2.5 Active site prediction:

The active site binding analysis was done using CASTp calculation.⁹ The validated model was subjected to active site prediction. Out of the obtained results, the one with the highest area and volume with the default value of 1.4 angstroms for the probe radius was selected as a binding site of ligand in the protein.

 

2.6 Molecular docking study:

Molecular docking studies refers to the prediction of the preferred orientation of ligand to the protein molecule which bound together to form a stable complex. The scoring gives the strength of association of interacting residues in the docking site. The docking studies were carried out for using AUTODOCK 4.2 using Lamarkian genetic algorithm to find the global optimized conformation.¹⁰ Polar Hydrogens were added and the kollaman charges were added. The grid box was fixed around the predicted active site residues. The grid box dimension was set to be 56 x 66 x 60 with space 0.519nm. The default parameters were used for docking studies. After completion of 10 runs, the best conformation of ligand was selected based on the lowest binding energy for analysis

 

 

 

Figure 1.Modelled Structure of protein 4 aminobutyrate aminotransferase

3. RESULTS AND DISCUSSION:

3.1 Homology modeling and validation:

Homology modeling is a most reliable method for 3D structure prediction for proteins without a crystal structure. The 3D structure of GABA-AT protein was generated using the template 3R4T chain A with 83% identity and query coverage of 99% using SWISS MODELLER and illustrated as Figure 1.

 

The Q mean score of this model is 0.765 which is obtained from the Q mean server estimating the absolute model quality. The Ramachandran plot acquired through the PROCHECK analysis validates the model with 86% the amino acid residues are in phi-psi distribution of the plot which represents that the model is valid for further analysis and the residues present in the disallowed region occur in loop region of the model as shown in Figure 2.

 

Overall quality factor was obtained from ERRAT analysis with a percentage of 93.848% as indicated in (Figure 3). As the overall average quality factor the protein which has more than 91% makes the model as a considerable one. The structure resolution from the modelled protein is in the range of 2.5 to 3A⁰. From Table 2, the VERIFY 3D analysis shows that 80% of the amino acids have scored >=0.2 in the 3D/1D plot (Figure 4). The constructed model was energy minimized using NOMAD ref server.

 

 

 

 

Figure 2.Ramachandran plot for the modelled structure of GABA-AT.

 

 

Figure 3. ERRAT results with a quality factor of 93.868. Generally

 

 

Table 2. Assessment and validation results of the model.

Protein	Template PDB ID	Identity	Query coverage	Q -Mean score5	Procheck				Prosa Z - score6	Verify 3D Score (AA in 3D/1D)6	ERRAT Score6
					A1	B2	C3	D4
GABA-AT	3R4T	83%	99%	0.765	86%	12.9%	0.8%	0.3%	-9.26	80%	93.848%

¹Percentage of most favoured regions. ² Percentage of residues in additional allowed regions. ³ Percentage of residues in generously allowed regions.  ⁴ Percentage of residues in disallowed regions.⁵ Model quality estimation.⁶ Overall model quality.

 

 

Figure 4. Verify 3D analysis

 

3.2 Active site prediction analysis:

Active site prediction was done to identify the ligand binding packet of the protein which may play a major role in drug designing. Using the CASTp results, the functional region with highest volume and solvent accessible surface was considered to be the best binding site of the protein. The residues lying in this packets are R26/ 423/ 156, I65/231/296, N124, S125, V30/ 34/ 67/ 130/ 264, H151/ 154, Q265, G34/35/64/126/155/235/299/416/ 419, T37/164/170/266/398/414, A127/152/395, Y153/170, K166/294, M168/417, P169/301, F174/324, E229/234/399, L300/302/413, C415.

3.3 Secondary structure prediction:

Secondary structure prediction is a significant step to elucidate the 3D structure and  thefunction of the protein. Secondary structure prediction consist of α helix (A), β sheet (E) and the coil (C) based on the local conformations in the folding pattern of the protein. In the YASPIN server results, the predicted 12 helices, 17 β sheets and 26 coils were identified in the protein and denoted in Figure 5.

 

 

 

 

 

Figure 5.Secondary structure prediction by theYASPIN server. Here, AA: Target sequence, Pred: Predicted secondary structure (H=helix, E=strand, - = coil), Conf: Confidence (0=low, 9=high)

3.4 Protein ligand interaction analysis

Molecular docking program was carried out to predict the binding modes of the ligand into the binding packet of the modelled protein using Genetic algorithm. The known inhibitors which were reported and the marine compounds selected for docking study prefer the same groove. Among them, Gabaculine is the best known inhibitor and Heteronemin is the best marine compounds were screened based on the binding energy. The binding energies of the known inhibitors are in the range of -3.47 to -6.36 as shown in Table 3. whereas the binding energies that of the marine compounds are in the range of-5.81 to -9.19 as shown in Table 4.

\

Table 3.The binding energies and the interacting residues of the protein with the known inhibitors using docking studies.

S. No	Known inhibitors	Binding energies Kcal/mol	Residues involved in interaction
			H-Bond	Vander waal interactions	Electrostatic Interaction
1.	Gabaculine	-6.36	LYS294, ASP262, TYR153	VAL130, ALA127, GLY126, GLY155	GLU229, HIS154, VAL264
2.	Aminooxyacetic acid	-4.11	LYS171, ARG183, ALA165	-	LYS166, LYS294, THR164, ALA162
3.	Phenelzine	-4.39	TYR153, GLU229, GLU234	HIS154, ASP262, GLY126, VAL130	LYS294, GLY235, GLN265, GLY233
4.	Phenylethylidenehydrazine	-4.22	GLU229, GLN265	VAL130, GLY155, GLY126, ALA127, HIS154, VAL264, ASP262	LYS294, GLY233, GLY235, GLU234, TYR153
5.	Vigabratin	-5.25	ASP262, TYR153, LYS294	GLN265, GLY155	ALA127, GLY126, HIS154, VAL264, GLU229
6.	Valproic acid	-3.47	LYS294, GLY126	SER125, ASN124, LEU302, PRO301, GLY299, LEU300, ALA293	-
7.	Rosmarinic acid	-4.66	LYS294*, ALA293, ILE296	VAL264, PRO301, VAL67	ALA297, LEU300, GLY299, GLY295, ASN124, GLY64, GLY126, ALA127

* indicates the sigma interactions between the residue and the ligand.

 

 

 

Table 4.The binding energies and interacting residues of the protein with marine compounds

S. No	Marine compounds	Binding energies Kcal/mol	Residues involved in interactions
			H-bond	Vanderwaal interactions	Electrostatic Interaction
1.	Bipinnapterolide	-6.61	LYS294	ASN124, SER125, VAL264, PRO301, ALA293, LEU302, ILE296, VAL67, GLY64	GLY126, GLY299, LEU300
2.	Heteronemin	-9.19	SER125, GLY126, ALA127, ARG156	VAL67, TYR170, GLY155, ILE65, GLU234, VAL264, LYS294, ALA293, ASN124, LEU302, PRO301, LEU300, GLY299	TYR153
3.	Hymendin	-5.81	TYR153, ASP262	VAL130, VAL264	ALA127, LYS294, GLY126, LEU302, ASN124, LEU300, SER125, HIS154GLU229
4.	Litosterol	-6.88	HIS154	PRO301, LEU302, LEU300, GLY299, ASN124, VAL67, GLY64, SER125, GLY126, ARG156, VAL264	LYS294, GLY155, ALA127
5.	Nephalsterol	-6.62	VAL40, VAL42	ARG26, VAL36, THR37, LEU38, ILE411, LEU413	PHE41
6.	Parguestrol A	-7.04	-	TYR153, VAL264, LYS294, ALA293, LEU302, GLY299, LEU300, PRO301, ASN124, GLY126, ALA127	ARG156, SER125
7.	Parguestrol B	-6.48	LEU38	VAL36, ARG26, VAL34, VAL30, PHE41, VAL42, LEU413	THR37, VAL40
8.	Puuphenone	-7.34	LEU302, LEU300	VAL67, ARG156, TYR153, ALA127, GLY126, VAL264, ALA293, ILE296	LYS294, GLY299, ALA297, PRO301

 

From the 2D plot obtained, a total of 10 residues were found to have the interaction with Gabaculine such as, LYS294, ASP262, TYR153, VAL130, ALA127, GLY126, GLY155, GLU229, HIS154, VAL264 whereasHeteronemin has 18 binding residues which are as follows SER125, GLY126, ALA127, ARG156, VAL67, TYR170, GLY155, ILE65,GLU234, VAL264, LYA294, ALA293, ASN124, GLY299, LEU300, PRO301, LEU302, TYR153. The 2D plot of these compound is shown in Figure 6.

 

Figure 6. (A) 2D Plot of Heteronemin and the interacting residues. (B) 2D Plot of Gabaculine and the interacting residues.

 

 

Figure 7. (A) Binding site of Gabaculine with GABA-AT Protein. (B) Binding site of Heteronemin with GABA-AT.

 

When investigated the binding pocket we can observe that heteronemin showed as proper binding to the GABA-AT protein with respect to that of the known inhibitor i.e. gabaculine.  The docked complexes viewed in Discovery studio visualizer 4.0 shows us the pattern of binding of the gabaculine and the heteronemin with GABA protein and also we can observe a proper hydrogen bonding in the heternemin as shown in Figure 7.

 

4. CONCLUSION:

The current analysis provides a brief scenario of structural and functional importance of 4-aminobutyrate aminotransferase protein of Mycobacterium tuberculosis. This modelled protein and docking analysis shows that apart from the known inhibitors of 4-aminobutyrate aminotransferase (GABA-AT), the effective marine compound Heteronemin showed promising results for the drug development against the Mycobacterium tuberculosis. This study possibly provide a key role in monitoring and optimizing the therapeutics of Tuberculosis.

 

5. REFERENCES:

1.       Kobayashi K, Miyazawa S, Terahara et al. Gabaculine: γ aminobutyrate aminotransferase inhibitor of microbial origin. Tetrahedron Letters 1976; 17(7): 537-540.

2.       Potashman M, Duggan M. Perspective covalent Modifiers: An orthogonal approach to drug design. Journal of Medical chemistry 2009; 52: 1231-1246.

3.       Khalid AES, Piotr B, Xiaoyu S, Tony LP, Jordan KZ, Mark TH. Marine natural products as Antituberculosis Agents. Tetrahedron 2000; 56: 949-953.

4.       Gram L.  Experimental studies and controlled clinical testing of valproate and vigabatrin. ActaNeurologicacandinavica 1988; 78(4): 241-270.

5.       Ana  martins, Gongalo A, Catia R, Ana almeida, Ricardo P, Patricia C, Helena V. Anti tuberculosis activity present in a uique marine bacteria collection from Portuguese deep sea hydrothermal vents. Journal of Marine biology and Oceanography 2013; 2:3.

6.       Altschul S, Gish W, Miller W et al. Basic local alignment tool. Journal of Molecular Biology 1990; 215(3): 403-10.

7.       Bordoli L, Kiefer F, Arnold K et al. Protein structure homology modeling using SWISS-MODEL workspace. Nature protocols 2009; 4:1-13.

8.       Lin K., Simossis V.A., Taylor W.R. and Heringa J. A Simple and Fast Secondary Structure Prediction Algorithm using Hidden Neural Networks. Bioinformatics. 2005, 21(2):152-9.

9.       Joe D, Zheng Q, Jeffery T, Andrew B, Yaron T, Jie L. CASTp: Computed atlas of surface topography of proteins with structural and topographical mapping of functionally annotated resuides. Nucleic Acid Research 2006; 34:116-118.

10.     Forli, Stefano, et al. "Computational protein-ligand docking and virtual drug screening with the AutoDock suite." Nature protocols 11.5 (2016): 905-919.

 

Accepted on 02.09.2017          © RJPT All right reserved

Research J. Pharm. and Tech 2018; 11(4):1283-1288.