|Year : 2021 | Volume
| Issue : 2 | Page : 106-114
In silico structural characterization of Cytochrome c oxidase Subunit 1: A transmembrane protein from Aedes aegypti
Jhansi Venkata Nagamani Josyula, Srinivasa Rao Mutheneni
Applied Biology Division, CSIR-Indian Institute of Chemical Technology, Tarnaka, Hyderabad-500007, Telangana, India
|Date of Submission||22-Mar-2019|
|Date of Acceptance||15-May-2019|
|Date of Web Publication||13-Jan-2022|
Dr. Srinivasa Rao Mutheneni
Pharmacology & Toxicology Division, CSIR-Indian Institute of Chemical Technology, Tarnaka, Hyderabad–500007, Telangana
Source of Support: None, Conflict of Interest: None
Background & objectives: The present study proposed a series of computational techniques such as homology modelling, molecular simulation, and molecular docking to be performed to explore the structural features and binding mechanism of Cytochrome c oxidase subunit I (COX1) protein with known inhibitors.
Methods: Elucidation of the three-dimensional structure of COX1 protein was carried out by using MODELLER software. The modelled protein was validated using GROMACS, structural qualitative tools and web servers. Finally the model was docked with carbon monoxide (CO) and nitric oxide (NO) using Auto Dock Tools.
Results: The three-dimensional structure of mitochondrial transmembrane protein COX1 was built using homology modelling based on high-resolution crystal structures of Bos taurus. Followed by inserting the lipid bilayer, molecular dynamics simulation was performed on the modelled protein structure. The modelled protein was validated using qualitative structural indices. Known inhibitors such as carbon monoxide (CO) and nitric oxide (NO) inhibit their active binding sites of mitochondrial COX1 and the inhibitors were docked into the active site of attained model. A structure-based virtual screening was performed on the basis of the active site inhibition with best scoring hits. The COX1 model was submitted and can be accessible from the Model Archive site through the following link https://www.modelarchive.org/doi/10.5452/ma-at44v.
Interpretation & conclusion: Structural characterization and active site identification can be further used as target for the planning of potent mosquitocidal compounds, thereby assisting the information in the field of research.
Keywords: Cytochrome c oxidase subunit I (COX1); Gremlin; Homology modelling; Modeller; Molecular Dynamics simulations; Molecular docking studies
|How to cite this article:|
Josyula JV, Mutheneni SR. In silico structural characterization of Cytochrome c oxidase Subunit 1: A transmembrane protein from Aedes aegypti. J Vector Borne Dis 2021;58:106-14
|How to cite this URL:|
Josyula JV, Mutheneni SR. In silico structural characterization of Cytochrome c oxidase Subunit 1: A transmembrane protein from Aedes aegypti. J Vector Borne Dis [serial online] 2021 [cited 2022 Jan 29];58:106-14. Available from: https://www.jvbd.org/text.asp?2021/58/2/106/331415
| Introduction|| |
Cytochrome c oxidase (COX) is a major membrane protein complex (molecular weight 204,000 Da) located on the inner membrane of mitochondrion. COX catalyzes the reduction of four electrons from ferrocytochrome c to oxygen and later converting them to water in the final step of the respiration. Cytochrome c oxidase protein complex also acts as a major regulation site for oxidative phosphorylation. The COX is active as a dimer composed of two identical parts. It contains three copper ions, which are necessary for its function and it also has zinc, magnesium and two iron atoms bound in heme groups.
COX is comprised of 13 protein subunits of which three are encoded by mitochondrial DNA and ten are encoded by the nucleus. A recent study reveals that COX activity was regulated by the nuclear encoded subunits. The nuclear-encoded subunits play a major role in the development and physiology of multicellular organisms, as revealed by lethality and pleiotropic phenotypes that caused by mutations in subunit VIc in drosophila. The three biggest subunits of Cytochrome C oxidase are COX I, II and III which are encoded by the mitochondrial genome form the functional core of the enzyme complex,.
Cytochrome c oxidase subunit I (COX1) is the terminal catalyst of the electron transport chain and it is involved in electron transport and proton translocation across the membrane,. The subunits I and II contain metal binding sites, whereas subunit III stabilizes these catalytic centres are involved in O2 transport pathway and the remaining 10 subunits stabilizing functions,. COX1 is found in all heme-copper respiratory oxidases because of the presence of the bimetallic centre. The bimetallic centre is formed by heme-a3 and copper B, both ligated to six conserved histidine residues outside of four transmembrane spans. Electrons originating from COX1 have transferred via the copper A (CuA) centre of subunit II and heme A of subunit 1 to the bimetallic centre formed by heme-a3 and copper B,. This Cytochrome c oxidase subunit-1 shows the transport of protons against an electrochemical gradient, using energy.
Mitochondrial genes have frequently been used as molecular markers in evolutionary studies. The COX1 region is highly efficient for discriminating the vertebrate and invertebrate species, whereas, it is not suitable for fungal and plants species. In DNA bar-coding studies COX1 is used as a molecular marker for species identification, classification and evolutionary studies because COX1 is the largest subunit and the protein sequence contains highly conserved functional domains and variable regions,. Cytochrome c oxidase subunit 1 of Aedes aegypti consists of 512 amino acids. The A. aegypti COX1 peptide shows high similarity in BLAST searches: 93% identity with Anopheles gambiae, 92% to Anopheles quadrimaculatus and 88% to Drosophila yakuba. The sequence alignment of COX1 shows that 12 amino acid residues are unique to the A. aegypti, of which six are located in the COOH-terminal region, the most variable region.
Computational chemistry is very helpful for providing atomic resolution models of COX1 by homology modelling techniques, without an experimentally known three- dimensional structure of the protein. Homology modelling generates a model which is used to design particular or highly selective target ligands. The goal of the present study is to build the three-dimensional structure of COX1 through in silico modelling and molecular docking studies to assess the active sites of the protein in order to accelerate the development of insecticidal compounds towards a target.
| Material & Methods|| |
The COX1 protein sequence data were retrieved from NCBI and UniProt. The mosquito Aedes aegypti amino acid sequence of Cytochrome c oxidase subunit I (COX1_ AEDAE) obtained in FASTA format with accession no. B0FWC7.
Gremlin co-evolutionary analysis
GREMLIN a statistical model was used to identify COX1 protein family that integrate both conservation and co-evolution patterns using pseudo likelihood approach,.
Template identification and 3D model building
Basic local alignment search tool (BLAST) from NCBI was used to ascertain the homologous proteins with kenned structures was employed as template protein utilizing the BlastP algorithm. COX1 had shown 75% similar identities with expect value (E) cut off of 0.000 utilizing compositional adjust matrix with no gaps of mitochondrial encoded COX 1 (mt: Cox1) gene crystal protein structure of 1V54 and 1OCC (Bovine Heart Cytochrome C Oxidase) of Bos Taurus. The template was aligned with the target sequence and ascertains the conserved regions. Iterations of BlastP templates 1V54 and 1OCC were used to carry out comparative modelling of COX1. The aligned sequences were taken for model construction utilizing MODELLER (version: 9.18),.
Transmembrane helix prediction and consensus sequence alignment
To define the trans-membrane regions of the protein, COX1 model were consigned utilizing TOPCONS (consensus prediction of membrane protein topology and signal peptides). Crystal structures of 1V54, 1OCC at resolution 1.8 Å and 2.8 Å was obtained from Protein Data Bank as templates for COX1 transmembrane (TM) helix prediction. Several prediction methods were applied including MEMSTAT3.0, PHOBIUS, TMHMM, STRIDE, TOPCONS, and PSIpred were used to assign putative TM helix segments of COX1. Followed by predicting the helix using these servers, the consensus method was applied and assessed by counting the number of servers that predicted residues as TM helix.
Molecular dynamics simulation
The modelled structure was refined by energy minimization and molecular docking studies. The model protein was refined by GROMACS (version: 5.1.4), a molecular dynamics simulation software. Molecular energy minimizations were carried for COX1 modelled protein packed with lipid bilayer in an aqueous environment using optimized potentials for liquid simulation under all-atom force field (OPLS-AA force field) to generate the topology of the protein. The modelled protein energy was minimized to discard high energy inter-molecular interactions. The system was stimulated at temperature 300K NVT [constant number (N), volume (V) and temperature (T)] over 100ps using modified Berenson thermostat. Similarly, 100ps in NPT (constant number (N), pressure (P) and temperature (T)) equilibration were applied using ensemble run with the Parrinello–Rahman pressure coupling.
The modelled protein stability was examined by GROMACS analysis and the results were evaluated by using root mean square deviation (RMSD), root mean square fluctuation (RMSF). RMSD shows the crystal structure deviation whereas RMSF shows deviation from the mean structure over the dynamic ensemble. The trajectories were visualized by using a 2D plotting grace tool (www.plasma-gate.weizmann.ac.il/ Grace/). Graphical plots were drawn using R statistical package.
Evaluation and validation
Procheck was used to evaluate stereochemistry of the structure (e.g. bonds, bond angles, dihedral angles, and non-bonded atom-atom distances). The quality of modelled protein structure depends on the distribution of dihedral angles φ (phi), ψ (psi) and ω (omega) of Ramachandran plot. In addition, ERRAT, Verify3D, 3Dimensional structural superposition and ProSA (protein structure analysis) were used to validate the quality of the 3D model.
Active site prediction
Identification of active site binding pockets of COX1 protein was explored by using CastP and Site Hound web servers.
Structural homology searches for the exploration of homologous proteins
The COX1 model was submitted to the DALI server to compare the 3D protein structures database. The analysis also compares any biologically related entities similar to the protein structure revealed by X-ray crystallography and NMR.
Receptor and ligand preparation
To assess the binding sites of COX1 protein, the virtual selection was conducted using AutoDock Vina and Auto Dock 4.0. In this study, the ligands were fully flexible while the protein was kept rigid. The COX1 protein model was generated by using Modeller and this model was used as a receptor for the molecular docking. The ligands, carbon monoxide (CO) and nitric oxide (NO) were obtained from PubChem compound database. Downloaded 2D structures of ligand were converted into protein data bank files using Accelrys Discovery Studio 2017 software package. PDB coordinates of the COX1 protein model and ligand molecules were optimized by using Discovery Studio 2017 software.
Auto Dock Tools v.1.5.6 was used for preparing input files for AutoDock Vina and Auto Dock 4.0,. The energy was minimised in receptor and ligands, the modelled protein was placed in a grid box with 74*70*72 for carbon monoxide and 60*64*70 for nitric oxide at 1.0 Å spacing using Auto grid. For all docking parameters, the standard protocol was used along with the addition of merge non-polar hydrogen and Gasteiger charges. Docking was performed using the implemented empirical free energy function with the Lamarckian Genetic Algorithm (LGA). Carbon monoxide and nitric oxide were first docked into the active binding cluster, and the resulting interactions were compared with the Clusters (1-3) for the attachment of active site residues using the same grid box. Cluster analysis was performed based on active sites docking results such as binding energy interactions, Hydrogen bonding interactions of closely interacting amino acids and the docking results using root mean square (RMS) with a tolerance of 0.5 Å.
Analysing and output visualisation using Discovery Studio
Best pose with lowest binding energy was considered for each ligand according to their docking score. The conformation with the lowest binding energy was chosen for further analysis. According to docking scores, the docking poses were ranked and visualized using Pymol viewer and Discovery Studio.
| Results and Discussion|| |
Homology modelling and in silico screening play a pivotal role to understand the structure and active binding sites of COX1 protein, which is crucial for designing small molecules and peptides for targeting COX1 protein. The COX1 protein sequence of Aedes aegypti was retrieved from UniProtKB. BLAST was performed using B0FWC7 as query sequence, a crystal transmembrane protein structure of bovine heart cytochrome c oxidase in the fully oxidized state (PDB ID: 1V54 A) and chain A cytochrome c oxidase of Bos Taurus MT-CO1 (PDB ID: 1OCC A) showed 40% identity, 95% query coverage at the E-value of 0.0 was obtained as templates for COX1. Gremlin co-evolution and covariance model evaluates templates (1OCC A and 1V54 A) to obtain the best template for COX1 modelling. The accurate contact was observed with 1V54 A chain with the coverage 0.9981 [Figure S1] [Additional file 2] and probability 100% at higher HHΔ score of 0.042 as the best template. The HH search of GREMLIN was used for prediction of positional alignment confidence of COX1 with the closest match of a 1V54 PDB structure. The two-dimensional representation of the three-dimensional protein structure shows the residues are in contact with COX1. Analysis of COX1 secondary structure shows that the alpha-helix, beta-sheet and turn regions consist of 73.5%, 69.5%, 7.2% residues. The amino acid distribution pattern of COX1 showed that high amount of hydrophilic residues with an average hydropathicity of 0.720. Similarly, COX1 is stable with 15 positively charged amino acids and 23 negatively charged amino acids with the instability index of 30.46 and extinction coefficient of 106340 M-1 cm-1 at 280 nm in water.
[Figure 1] describes the phylogenetic tree of COX1 in different species. The cladogram shows that the evolution with other structurally-related proteins using fast minimum evolution method with the maximum sequence difference of 0.85 using Grishin distance. The COX1 protein is nearer to the even-toed ungulates Chain A of bovine heart cytochrome c oxidase structure.
|Figure 1: Slanted Cladogram of COX1 Protein model shows coevolution with other mosquito protein of Cytochrome c oxidase subunit 1 and other structural related species.|
Click here to view
High accuracy 3D model of COX1 was achieved by using homology modelling. The COX1 belongs to transmembrane (TM) protein complex hence, sequence based homology modelling is not strong enough for model building. In order to facilitate the TM regions of COX1 model, the secondary structure was predicted by using Chou & Fasman prediction server. Prediction of TM helices, 1V54, 1OCC and 1AR1 PDB structures was used as templates. The length of helix distribution was examined and consensus observed at 70%, 80%, 90% and 100% predicts alpha helices are present in TM regions. Various algorithms such as TOPCONS, OCTOPUS, Philius, PolyPhobius, SCAMPI and SPOCTOPUS were used to predict the length and helices size. These algorithms predicted 12 alpha-helices which are present in the TM regions of mitochondrial membrane [Figure 2]. The identities were normalised by aligned length and consensus regions are in grey colour shown in [Figure 2].
|Figure 2: Location of the 12 helical transmembrane segments of COX1 model by different servers.|
Click here to view
The COX1 protein structure was build using Modeller v 9.18 and a large number of spatial restraints was generated from the template structure (1V54) by comparing the target and template proteins bond lengths and angle between aligned residues. The COX1 models were built by Modeller and reviewed with two types of scores such as DOPE Score and GA341 Score. The better COX1 model shows the lowest scores, the DOPE score is -76165.73 and GA341 shows 1.0. The DOPE energy profile plots of COX1 model protein and template were shown in [Figure 3].
Methods for validation of predicted model
1. Structure Quality Assessment: The quality of the model was assessed by PROCHECK. The 3D protein model quality was evaluated using Ramachandran plots, it is an easy way to understand the distribution of torsion angles Φ and Ψ of non-Glycine and non-Proline residues [Figure 4]. COX1 model shows that 96.4% of residues (i.e. 439) are present in favoured regions of Ramachandran plot. However, only one residue was present in the disallowed region. The quality of covalent bonds and bond angle distance was checked by PROCHECK and the overall average G-factor for the modelled protein is -0.67 [Table S1][Additional file 1].
2. Validation by ERRAT: To verify the structure of the 3D model protein, the structure was submitted to ERRAT server. The bonded and non-bonded interactions in the structure were assessed based on error values provided by the ERRAT. The overall quality factor of the COX1 model is 83.39 it implying that the error value of almost all residues is below the cut-offs and it shows that the modelled protein is highly accurate [Figure S2] [Additional file 3]. [Figure 5] describes that the X-axis consists of residues and Y-axis is an error. The two lines indicate 95% and 99% confidence which is possible to reject regions that exceed the error value.
|Figure 5: Three dimensional structure of Cytochrome C Oxidase Sub-unit 1 protein model of Aedes aegypti|
Click here to view
3. VERIFY-3D: To verify three-dimensional processed protein model by assigning a structural class based on alpha, beta, loop, polar and non-polar environments. The packing quality of each residue was evaluated by Verify 3D and 80.35% of the residues had shown average 3D-1D score >= 0.2. This indicates that the 3D protein structure has good packing quality and a valid folded conformation [Figure S3] [Additional file 4].
4. Validation by ProSA: ProSA was used for further refinement and validation of modelled protein structure. [Figure 6]A depicts the ProSA Z-Score for the modelled protein was -4.88 determined by NMR spectroscopy and X-ray crystallography with respect to length of residues. [Figure 6]B shows that the energy function of amino acid sequence position in modelled protein. The positive value shows that the erroneous parts of the protein and the negative value shows the stable protein conformation. Overall, the residue contains more negative energy peaks than the positive peaks. The former peaks correspond to protein membrane-spanning regions which indicate that the protein is of good quality.
|Figure 6: (A) Z-score of -4.88 predicted model showing black dot (B) Local quality score based on knowledge based energy.|
Click here to view
5. TM-align: Further, the model was validated through TM-align, to identify the best structural alignment between protein pairs that combine with the TM-score rotation matrix and dynamic programming. The structural alignments of IV54 and COX1 obtained TM-score is 0.99354, RMSD 0.88 and sequence identical 0.737. The TM-score is implying that the structural similarity of COX1 and 1V54 are almost identical folds were observed.
6. Structural analysis: The 3D structural superposition (3dSS) server is an interactive server to superpose 3D Protein template structures with COX1 model using STAMP. The 3dSS server identified invariant and complete water molecules of the model. The RMSD values for model and template are 0.226 and 0.235 with a sequence identity of 74.26% [Table 1].
|Table 1: 3D structural super position of model protein, 1V54 and 10CC proteins|
Click here to view
Molecular dynamic simulation
MD simulation was carried out in water, after the protein was inserted into lipid bilayer in order to attain the protein is in a natural environment. The backbone of the model is the RMSD value which was analysed to understand the conformational variations of the model. The protein structure model showed the average RMSD trajectory around 0.15-0.20nm, this indicates that the model is stable [Figure 7]A.
|Figure 7: (A) RMSD during 2 ps/step NVT production simulation of the backbone atoms of the modelled protein over a time period of 1ns. (B) RMS fluctuations (a measure of back bone deviation) of backbone atoms (N, Cα and C atoms) during the molecular dynamics simulations.|
Click here to view
The root mean square fluctuation (RMSF) plot shows the COX1 model loops and secondary structure stability [Figure 7]B. The RMSF with respect to the average MD simulation conformation describing flexibility differences among the residues in the form of peaks that represent areas of the protein that are fluctuating most during the simulation. The N-terminal of the modelled protein much fluctuates than other protein areas because the secondary structural element of alpha-helices are more than the unstructured part of the protein. The RMSD and RMSF values show that the protein has a stable conformation. After MD simulation, the three-dimensional structure of COX1 model protein was shown in [Figure 5].
Predicted binding sites for inhibitor
The substrate binding sites in COX1 model was identified using CASTp 3.0 and Site Hound servers. The active binding pockets along with rank, energy and amino acids are represented in [Table 2].
|Table 2: Ranking energy of COX1 binding sites using CASTp and Site Hound|
Click here to view
In silico screening
Molecular docking is widely used for the calculation of protein-ligand interactions. Docking was used to predict the potential ligand-binding sites on the protein and also used to study the characteristics of a binding pocket of the model, which in turn helps to develop novel insect and pesticides. To assign the best binding pose, Auto dock uses the free binding energy to maintain binding affinity between protein and ligand complex. Each binding conformation was separated into clusters according to RMSD tolerance of 0.5 Å. Autodock elucidates the parameters such as intermolecular energy which includes Vander Waals, Hydrogen bonding, Desolvation, Electrostatic energies and inhibition constant of protein and ligand. The total Electrostatic energy of COX1 model was 5.33E+04 kcal/mol.
Carbon monoxide (CO) and nitric oxide (NO) were docked at the active site of the COX1. The binding affinities of COX1 were ranked according to the ligand conformations. The CO shows the best binding affinity of -1.87 kcal/mol whereas NO showed the lowest binding affinity of -2.11 in docking confirmation with COX1. The binding affinities were tabulated in [Table 3]. From the study, it is observed that both CO and NO showed interactions within the active site clusters 1 and 2. The active binding pocket of COX1 showed interactions with CO and form hydrogen bonds with the amino acid residues of cluster 1 and 2 are Gln 427, Tyr 370, Arg 437, His 367, Arg 438 and Arg 179, Lys 263, Lys 264, Leu 182, Thr 269 respectively [Figure 5]. In addition to this, the COX1 amino acids Thr 235, His 377, Phe 424, Phe 267, Thr 266, Glu 265, His 255, Asp 363, Val 372, and Thr 266 form hydrophobic interactions with CO. Docking results of NO showed that the formation of hydrogen bonds with COX1 amino acids Trp 235 and Arg 438 of cluster 1. Similarly Pro 181, His 255, Tyr 128, Leu 270 and Ile 273 involved in hydrophobic interactions with cluster 1 and 2 of NO. The 2D representation of protein ligand interactions of COX1 with CO and NO are shown in [Figure 8].
|Figure 8: Docked images of Carbon monoxide and Nitric oxide binding to COX1 protein cavities and 2D representation of hydrophobic, hydrophilic interactions and Hydrogen bonding interactions using Ligplot Software.|
Click here to view
|Table 3: In silico screening of COX1 protein with inhibitors such as Carbon monoxide and Nitric oxide.|
Click here to view
The Cluster 1 is located in the extracellular side of mitochondrial transmembrane hence, CO and NO able to inhibit the COX1 [Table 4]. The CO and NO shows degradation activity because of the large substrate channel passage allows multiple conformations to fit in the active site cavity 1 and 2 of COX1 and it can be assumed that CO and NO has the potential to inhibit COX1. NO shows the lowest binding affinity with 7 multiple conformations [Figure S4][Additional file 5] representing that, it inhibits COX1 better than CO as it has the lowest binding affinity with 6 multiple conformations.
| Conclusion|| |
A computational network approach was utilized to determine the three-dimensional structure of COX1 protein from Aedes aegypti, due to unavailability of crystallographic structure. Homology modelling is a useful approach for building protein structures and understanding the binding mechanism for structure-based drug design techniques. This approach is strongly dependent on virtual screening of a model and docking results. The validation of COX1 model was assessed by the various tools and provided good support for further functional analysis of experimentally derived structures. Molecular dynamics simulations revealed the stability of COX1 model and docking studies showed that the binding conformations of the docking complex and key amino acids that interact with well kenned inhibitors of COX1 modelled structure. The rational structure of COX1 generated by Modeller has confessed to allowing better guidance on different substrate binding predictions based on in silico screening studies. This structural model can be used as a putative inhibitor for the development of novel insecticide compounds for effective control of Aedes aegypti vector population.
Conflict of interest: None
| References|| |
Capaldi RA. Structure and function of cytochrome c oxidase. Annu Rev Biochem
Yoshikawa S, Shimada A. Reaction mechanism of cytochrome c oxidase. Chemical Reviews
2015; 115: 1936
Kadenbach B, Hüttemann M, Arnold S, Lee I, Bender E. Mitochondrial energy metabolism is regulated via nuclear-coded subunits of cytochrome c oxidase. Free RadicBiol Med
Tsukihara T, et al
. The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 A. Science
1996; 272: 1136–1144.
Das J, Miller ST, Stern DL. Comparison of diverse protein sequences of the nuclear-encoded subunits of cytochrome C oxidase suggests conservation of structure underlies evolving functional sites. Mol Biol Evol
2004; 21: 1572–82.
Szuplewski, S, Terracol R. The cyclope gene of Drosophila encodes a cytochrome c oxidase subunit VIc homolog. Genetics
Yoshikawa S, Muramoto K, Shinzawa-Itoh K. Proton-pumping mechanism of cytochrome C oxidase. Annu Rev Biophys
Gennis RB. “Site-directed mutagenesis studies on subunit I of the aa3-type cytochrome c oxidase of Rhodobactersphaeroides: a brief review of progress to date”. Biochimica et Biophysica Acta
Morlais I, Severson DW. Complete Mitochondrial DNA Sequence and Amino Acid Analysis of the Cytochrome C Oxidase Subunit I (COI) from Aedes aegypti. DNA Sequence
Yoshikawa S, Muramoto K, Shinzawa-Itoh K. Proton-pumping mechanism of cytochrome C oxidase. Annu Rev Biophys
Yoshikawa S, Muramoto K, Shinzawa-Itoh K, Mochizuki M. Structural studies on bovine heart cytochrome c oxidase. Biochim Biophys Acta
Capaldi RA, Malatesta F, Darley-Usmar. VM Structure of cytochrome c oxidase. Biochimica et Biophysica Acta
Holm L, Saraste M, Wikstrom M. Structural models of the redox centres in cytochrome oxidase. The EMBO Journal
1987; 6(9): 2819–2823.
Castresana J, Lübben M, Saraste M, Higgins DG. Evolution of cytochrome oxidase, an enzyme older than atmospheric oxygen. The EMBO Journal
Trivedi S, Aloufi AA, Ansari AA, Ghosh SK. Role of DNA barcoding in marine biodiversity assessment and conservation: an update. Saudi J Biol Sci
Tavares ES, Baker AJ. Single mitochondrial gene barcodes reliably identify sister-species in diverse clades of birds. BMC Evol Biol
2008; 8: 81.
De Boer HJ, Ouarghidi A, Martin G, Abbad A, Kool ADNA. Barcoding reveals limited accuracy of identifications based on folk taxonomy. PLoS One
Beard CB, Hamm D.M. and Collins, F.H. (1993) “The mito-chondrial genome of the mosquito Anopheles gambiae:
DNA sequence, genome organization, and comparisons with mito-chondrial sequences of other insects”. Insect Molecular Biology
Gennis, R.B. “Site-directed mutagenesis studies on subunit I of the aa3-type cytochrome c oxidase of Rhodobacter sphaeroides:
a brief review of progress to date”. Biochimica et Biophysica Acta
Rodrigues MS, Morelli KA, Jansen AM. Cytochrome c oxidase subunit 1 gene as a DNA barcode for discriminating Trypanosoma cruzi
DTUs and closely related species. Parasite Vectors 1992; 10:
Morlais I, Severson DW. Complete mitochondrial DNA sequence and amino acid analysis of the Cytochrome c oxidase subunit I (COI) from Aedes aegypti. DNA Sequence
Lunt DH, Zhang DXcz, Szymura, J.M. and Hewitt, G.M. “The insect cytochrome oxidase I gene: evolutionary patterns and conserved primers for phylogenetic studies”. Insect Molecular Biology
1996; 5: 153–165.
Ovchinnikov S, Kamisetty H, Baker D. Robust and accurate prediction of residue-residue interactions across protein interfaces using evolutionary information. Elife
2014; 3: e02030.
Kamisetty H, Ovchinnikov S, Baker D. Assessing the utility of coevolution-based residue–residue contact predictions in a sequence-and structure-rich era. Proceedings of the National Academy of Sciences
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J MolBiol 1990; 215: 403-410.
Sali A, Overington JP. Derivation of rules for comparative protein modelling from a database of protein structure alignments. Protein Sci
1994; 3: 1582–1596.
Eswar N, Eramian D, Webb B, Shen MY, Sali A. Protein structure modeling with MODELLER. Methods Mol Biol
Jones DT, Taylor WR, Thorton JM. A model recognition approach to the prediction of all-helical membrane–protein structure and topology. Biochemistry
1994; 33: 3038–3049.
Li W, Cowley A, Uludag M, Gur T, McWilliam H, Squizzato S, et al
. The EMBL-EBI bioinformatics web and programmatic tools framework. Nucleic Acids Research
Melen K, Krogh A, von Heijne G. Reliability measures for membrane protein topology prediction algorithms. J Mol Biol
Heinig M, Frishman D. STRIDE: a web server for secondary structure assignment from known atomic coordinates of proteins. Nucleic Acids Research
Bernsel A, Viklund H, Hennerdal A, Elofsson A. TOPCONS: consensus prediction of membrane protein topology. Nucleic Acids Res
2009; 37: W465–468.
McGuffin LJ, Bryson K, Jones DT. The PSIPRED protein structure prediction server. Bioinformatics
Pronk S1, Páll S, Schulz R, Larsson P, Bjelkmar P, Apostolov R, et al
. GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics
Berendsen HJC, Postma JPM, van Gunsteren WF, Hermans J. Interaction models for water in relation to protein hydration. Intermolecular Forces
Parrinello M, Rahman A. Polymorphic transitions in single crystals: A new molecular dynamics method. Journal of Applied Physics
1981; 52: 7182–7190.
R Development Core Team. (2010). R version 3.3.3
Sumathi K, Ananthalakshmi P, Roshan MNA, Sekar K. “3-Dimensional Structural Superposition”. Nucleic Acids Res
Wiederstein M, Sippl MJ. ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins. Nucleic Acids Research
2007; 35: W407–410.
Hernandez M, Ghersi D, Sanchez R. SITEHOUND-web: a server for ligand binding site identification in protein structures. Nucleic Acids Res
2009; 37: W413-416.
Holm L, Rosenström P. Dali server: Conservation mapping in 3D. Nucleic Acids Research
Morris GM, Goodsell DS, Huey R, Olson AJ. Distributed automated docking of flexible ligands to proteins: parallel applications of AutoDock 2.4. J Comput Aided Mol Des
Trott O, Olson AJ. AutoDockVina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8]
[Table 1], [Table 2], [Table 3], [Table 4]