nst Life Sciences and Bioinformatics ISSN: 0974-9179

http://www.nstjournal.com Open Access

Research Paper

In-Silico Modeling and Investigation of ATP Binding Pocket of

An Algal Oil Producing Enzyme

Raghunath Satpathy*, Rajesh ku.Guru, Rashmi R Behera and Aparajita P

Department of Biotechnology, MIRC LAB, MITS Engineering College, Rayagada, Odisha, India – 765 017.

(*Corresponding Author E-mail: [email protected])

(Received March 11, 2010; Revised and Accepted March 31, 2010)

ABSTRACT
Acetyl-CoA carboxyl transfersae (ACCase) is the key enzyme for the synthesis of oils in algae. The
particular enzyme catalyses the reaction to form the Malonyl -CoA by binding with the substrate like ATP
and Coenzyme A. The present work is an in-silico approach which deals with the Homology based
modeling and identification of ATP binding residues of the enzyme ACCase of the biofuel producing algae
Cyanidium caldarium. MODELLER program has been used for 3D modeling. The reliability of the model
was verified on basis of PROCHEK, ERRAT, and, DOPE result. Then prediction of binding pocket and
docking study show that the residues ILE 110, ARG113, LEU 296 are responsible for binding to ATP. The
structural motif analysis for the protein sequence by EMBOSS show that Prosite motif present in the
sequence is AGRR (110-113), which are in an agreement with previous findings. The stability of the
docked structure is also checked by energy minimization by Chimera.

Key words: Homology modeling; algae; docking; biofuel; binding site.

INTRODUCTION
The development of alternative energy sources such as shows that the residues like 110,113,296 are responsible
biofuels are one of the most exciting and challenging for the binding to ATP.
area. So the effort has been put to produce biodiesel
from the algal species [1].The red algae Cyanidium
caldarium is one of the most preferable organism to MATERIALS AND METHODS
produce the oil [2]. Considering the metabolic pathways
of oil production in the algae, the Acetyl-CoA carboxyl Sequence retrieval and 3D model building
transfersae (ACCase) enzyme plays a major role for oil
production. It catalyzes the production of Malonyl Co- The sequence for the ACCase enzyme was retrieved from
enzyme a using the substrate Acetyl -CoA and ATP [3].As SWISSPROT database having ID O19903 [7]. Then with
the three dimensional structure of the enzyme is not this query sequence a BLAST [8] search was performed
present in PDB (Protein Data Bank), so an in-silico against PDB (Protein Databank) to retrieve the
approach has been taken to predict the three corresponding template for the ACCase enzyme. The
dimensional structure by homology modeling method model was built by homology modeling and for this
[4]. By this approach a valid structural model has been MODELLER 9v5 [9] program. The MODELLER
produced with the available template having 46% program uses an automated approach to comparative
similarity. After model generation and validation an protein structure modeling by satisfaction of spatial
automated version of pocket finding search by CastP restraints (Fig.1). Briefly, the core modeling procedure
server [5] docking by Hex 5.0 version and structural begins with an alignment of the sequence to be modeled
motif analysis by EMBOSS tool [6] has been used to (target) with related known 3D structures (templates).
investigate the binding residues responsible for binding This alignment is usually the input to the program. The
of ATP. Furthermore the in-silico structural analysis output is a 3D model for the target sequence containing
all main chain and side chain non-hydrogen atoms [10].
_______________________________________________________________________________________________________________________
Journal of Natural Science and Technology - Life Sciences and Bioinformatics, 2010, Vol. 2: Page No. 147-152 -147-
nst Life Sciences and Bioinformatics Raghunath Satpathy et al.
http://www.nstjournal.com
RESULTS AND DISCUSSIONS
Model validation

The MODELLER generated structure was verified for the Homology modeling of ACCase enzyme
missing side chains by SCWRL4 tool [11] further verified
by PROCHECK [12]. The PROCHECK programs The sequence of 324 amino acid protein AcetylCo-
provides the information about the stereo chemical enzyme carboxylase carboxyl transferase subunit alpha
quality of a given protein structure. The PROCHECK was was obtained from SWISSPROT database. After the
used to generate Ramachandrans plot and the quality of sequence searched in PDB (protein data bank) by
the structure was computed in terms of % of residues in BLAST the template PDB ID 2F9I [17] chain A was
favorable regions, % of non Proline Glycine residues chosen as suitable one as it is having 46 % identity with
etc.The quality of structure was also further accessed by the query sequence. Then MODELLER was used to
using ERRAT [13]. ERRAT is a protein structure generate the three dimensional structure (Fig.2) and the
verification algorithm that is especially well-suited for structural features were derived (Table 1). Overall six
evaluating the progress of crystallographic model models were obtained the suitable model was chosen by
building and refinement. The program works by lowest molpdf value.
analyzing the statistics of non-bonded interactions
between different atom types. A single output plot is
produced that gives the value of the error function vs. Refining and Validation of the Model
position of a 9-residue sliding window. By comparison
with statistics from highly refined structures, the error After getting the model the missing side chains of the
values have been calibrated to give confidence limits. model was checked and managed by SCWRL4 software.
This is extremely useful in making decisions about The output PDB file from the software was then verified
reliability. Then the backbone alignment and RMSD by backbone comparison between the template and
(root mean square deviation) study was performed by refined model. The RMSD of backbone was calculated as
using PYMOL tool [14]. PyMOL is useful molecular 0.210 which is very much reliable.
modeling software allowing the visualization of three
dimensional molecular structures as well as for the The modeler generated DOPE score was computed and
backbone alignment between the model and template. analyzed. From the comparative chart (Fig.3) the peak is
showing that there is no defect in the loop regions in the
residues. So in the present case the loop refinement
method is not considered for the model. To verify further
In-silico conserved motif analysis and binding the predicted structures, the coordinates of the predicted
pocket analysis of the model structure was fed into the ERRAT Protein Verification
Server. The overall quality factor was obtained as 78.344
The refined validated model was then subjected to CastP which are very much satisfactory (Fig.4). The stereo
server to observe the possible pockets in the chemical quality of the structure was checked by
protein.CastP server is a automated on line tool that PROCHECK tool. The Φ and Ψ distributions of the
predicts the possible pockets for along with the residue Ramachandran plots of non-Glycine, non-Proline
positions and surface area.Also to find out the conserved residues are summarized (Fig 5). Altogether 92.3% of the
motif responsible for binding then protein sequence was residues were in favored regions (Table 2) so it can
searched in EMBOSS tool which scan a protein sequence consider as a good model for further analysis. The
with motif from Prosite database. overall G-factor used was computed as- 0.14.

Docking study and energy minimization Docking study

Docking was performed between the validated structure Docking of the ATP with the modeled enzyme was
of ACCase enzyme and the energy minimized ligand by performed using HEX tool. The grid was fixed at 0.6 and
using HEX 5.0 tool [15]. HEX is an interactive molecular the default FFT (Fast Fourier Transformation) mode was
graphics program for protein-ligand docking, assuming chosen. The algorithm exhaustively searches the entire
the ligand is rigid, and it can superpose pairs of rotational and translational space of the ligand with
molecules using only knowledge of their 3D shapes. It is respect to the receptors. The ATP binding with the
still the only docking and superposition program to use enzyme was obtained as energy value -228.72.Then the
spherical polar Fourier (SPF) correlations to accelerate energy minimization of the only model and the model
the calculations [16].The binding energy was computed complex with ATP was computed by Chimera tool [18]
and the ligand binding pattern was observed. Then a show the potential energy as -13981 .757 and -15605.920
compared energy minimization was performed in the respectively that indicates the ligand ATP is lodge in its
model alone and model with the docked ligand by proper binding site. From the docked structure the
Chimera tool for 100 steps to observe the stability in residues that facilitates for binding of the above ligands
terms of potential energy.

_______________________________________________________________________________________________________________________
Journal of Natural Science and Technology - Life Sciences and Bioinformatics, 2010, Vol. 2: Page No. 147-152 -148-
nst Life Sciences and Bioinformatics Raghunath Satpathy et al.
http://www.nstjournal.com

CastP server that also showing the pocket 3 having Ryagada to provide us the MIRC lab for computing
residues 110-CB –ILE 110-O-ILE, 113-CB-ARG, 296- facility.
CD2-LEU are responsible for binding (Fig.6). Cross
verified result by EMBOSS tool show that the prosite
motif of the sequence is from residue 110-113(AGRR).
REFERENCES

1. Edward M. et al., Nature, 454, 841-845, 2008.

CONCLUSION 2. Masamichi Akimoto et al., Journal of chemical
Engineering of Japan, 28, 349-352, 1995.
Acetyl Co-enzyme carboxyl transferase is the main 3. Qiang Hu et al., The Plant Journal, 54, 621–639,
enzyme which catalyses the first step of the lipid 2008.
synthesis in algae Cyanidium caldarium. The storage 4. Van Hannen EJ et al., Eur J Phycol, 37,203-208,
lipids in the algae correspond to the biofuel producing 2002.
capability. In the present work a homology based 5. T. Andrew Binkowski et al., Nucleic Acids Res., 31,
modeling for the above enzyme has been constructed 3352–3355, 2003.
using MODELLER software. And after various types of 6. Rice P et al., Trends in Genetics, 16, 276-277, 2000.
model validation software the result suggest that the 7. GloecknerG., J. Mol. Evol., 51, 382-390, 2000.
model is reliable. The stable structure is further 8. Altschul, S.F. et al ,J. Mol. Biol. 215, 403-410, 1990,
subjected to Docking with ATP the key ligand of the 9.A. Sali & T.L. Blundell, J. Mol. Biol., 234,779-815,
enzyme action. The docking result shows that almost 1993.
similar part of the enzyme is responsible for binding with 10. Eswar N et al., Mol. Biol, 426,145-159, 2008.
both the ligand also the residue. The predicted pocket 11. Georgii G. Krivov, Proteins Structure, Function, and
result of the enzyme obtained from Cast P server was Bioinformatics, 77, 778-795, 2009.
found is consistent with the predicted result by EMBOSS 12. Laskowski R et al, J. Appl. Cryst., 26, 283-291, 1993.
tool for the conserved motif. The work basically relates 13 V C. Colovos & T. O. Yeats, Protein Sci, 2, 1511-1519,
to the in-silico analysis of the fatty acid synthetic 1993.
pathway leading to accumulation of oils and fats and an 14. www.pymol.org.
understanding about how the pathway could be utilized 15. D.W. Ritchie, PROTEINS. Structure, Function and
for the commercial production of algal biofuels. Genetics., 52, 98-106, 2003.
16. L. Mavridis & D.W. Ritchie, J. Chem. Inf. Model., 47,
1787-1796, 2008.
ACKNOWLEDGEMENT 17. Bilder P, et al., Biochemistry, 45, 1712-1722, 2006.
18. Pettersen EF et al., J Comput Chem., 25, 1605-1612,
We are thankful to CEO, Director and Dean of 2004.
Majhighariani Institute of Technology & Science,

Table - 1

Structural information of the Modeled ACCase enzyme

Structural features Information

No. of residues 324
No. of atoms 2572
No. of Hydrogen bonds 225
No. of helices 17
No. of strands 15
No. of turns 28

_______________________________________________________________________________________________________________________
Journal of Natural Science and Technology - Life Sciences and Bioinformatics, 2010, Vol. 2: Page No. 147-152 -149-
nst Life Sciences and Bioinformatics Raghunath Satpathy et al.
http://www.nstjournal.com

Table - 2

Ramachandaran plot calculation on 3D model of ACCase enzyme computed with PROCHECK program

% residues in favorable regions 92.3

% residues in additional residue regions 5.2

% residues in generously regions 1.7

% residues in disallowed regions 0.7

% of non Proline and non Glycine residues 100

Figure - 1

Working principle of MODELLER (Sali and Blundell 1993)

Figure - 2

Rasmol view of final 3D structure of the ACCase. Yellow color indicates beta sheets and Red one is the alpha helix.

_____________________________________________________________________________________________________________________
Journal of Natural Science and Technology - Life Sciences and Bioinformatics, 2010, Vol. 2: Page No. 147-152 -150-
nst Life Sciences and Bioinformatics Raghunath Satpathy et al.
http://www.nstjournal.com

Figure - 3

Comparative DOPE value of template and model obtained from MODELLER output

Figure - 4

ERRAT structural quality

Figure - 5

The Ramachandran’s plot for the model protein

_____________________________________________________________________________________________________________________
Journal of Natural Science and Technology - Life Sciences and Bioinformatics, 2010, Vol. 2: Page No. 147-152 -151-
nst Life Sciences and Bioinformatics Raghunath Satpathy et al.
http://www.nstjournal.com

Figure – 6

The docked structure of ATP (Red color) in the pocket residues of the enzyme. (Shown in Blue color)

Citation: Raghunath Satpathy et al., nst Life Sciences and Bioinformatics Vol. 2: 147-152 (2010)

License statement: This is an open-access article, which permits unrestricted use, distribution, and reproduction in any medium,
for non-commercial purposes, provided the original author and source are credited.

_______________________________________________________________________________________________________________________
Journal of Natural Science and Technology - Life Sciences and Bioinformatics, 2010, Vol. 2: Page No. 147-152 -152-