Journal of Proteomics & Bioinformatics- Open Access www.omicsonline.com Research Article JPB/Vol.2/June 2009 Analysis of Binding Properties of Phosphoinositide 3-kinase Through In silico Molecular Docking

Phosphoinositide 3-kinases (PI3-kinases) are increasingly considered to have a key role in intracellular signal transduction in health and disease. Particularly the enzymes plays vital role in wide range of cancer such as breast, ovarian, myeloid leukemia, prostate, Small Cell Lung cancer (SCLC) etc., Compounds such as Wortmannin, LYS2002 are the inhibitors of PI3-kinases but these compounds shown adverse side effects . Hence five natural flavanoids having inhibitory effects on PI3-kinase namely Andrographolide, Kaempferol, Luteolin, Quercetin and Gingerol were taken for in silico prediction of binding affinities of the protein PI3- kinase. Our reports can be used to develop new inhibitors with better binding affinities towards the protein PI3-kinase protein. For the binding analysis the catalytic subunit of the protein PI-3 Kinase p110? was taken for the study as it considered being a potential target in cancer treatment.


Introduction
Phosphoinositide 3-kinases (PI 3-kinases or PI3Ks) are a family of related enzymes that are capable of phosphorylating the 3 position hydroxyl group of the inositol ring of phosphatidylinositol (PtdIns). PTEN/PI3K/AKT constitutes an important pathway regulating the signaling of multiple biological processes such as apoptosis, metabolism, cell proliferation and cell growth (Carnero A et al., 2008). Genomic mutations, alterations of the PI3K-AKT regulatory network, underlie such diseases as cancer, glucose intolerance (diabetes mellitus), schizophrenia, and/or autoimmune diseases (Noguchi et al ., 2008). In particular the PI 3-kinases generate and convey signals that have an important role in cancer (Stein 2001). PI3-kinases are ubiquitously expressed, are activated by a high proportion of cell surface receptors, especially those linked to Tyrosine kinases, and influence a bewildering variety of cellular functions and events. The majority of the research on PI 3-kinases has focused on the Class I PI 3-kinases. Class I PI 3-kinases are composed of a catalytic subunit known as p110. Many literature studies has proven that PI 3-Kinases to be the most signific ant contributor to activation of cancer in human such as ovarian cancer (Bellacosa et al., 1995;Yuan et al., 2000;Shayesteh et al., 1999), breast cancers (Nakatani et al., 1999), myeloid leukaemia (Vanhaesebroeck et al., 1999), glioblastoma, prostatic, endometrial and endometroid ovarian cancer [Ali et al., 1999).Apart from these frequent and early involvement of the PI3-kinase pathway was observed in lung cancer specifically small cell lung cancer (SCLC) (Pierre et al., 2004;Moore et al., 1998). A number of compounds such as wotmannin (Powis et al ., 1994), demethoxyviridin (Woscholski et al., 1994), LY294002 (a morpholino derivative of the broad-spectrum kinase inhibitor quercetin (Vlaho and showed less potent for this group of enzymes. More over Inhibitors of PI 3-Kinase have unacceptable toxicity if administered continuously in protein trafficking and in DNA repair and cell cycle checkpoint control is likely to be undesirable. The potential toxicity of PI 3-kinase inhibitors can probably best be limited by compounds extracted from natural source. Flavonoids provide a large number of interesting natural compounds that are consumed daily and exhibit more or less potent and selective effects on some signaling enzymes as well as on the growth and proliferation of certain malignant cells in vitro (Laurence et al., 1999). In silico molecular docking is one of the most powerful techniques to discover novel ligands for receptors of known structure and thus play a key role in structure-based drug design (Brooijmans et al., 2003).Investigators often use docking computer programs to find the binding affinity for molecules that fit a binding site on the receptor. Hence here we have taken Insilico molecular docking to analyze the binding properties of the enzyme PI 3-kinase with the flavanoids.

Flavanoids Taken for Binding Analysis with PI3 Kinase
Natural flavanoids such as Andrographolide from Andrographis paniculata, Gingerol from Zingiber officinale , Kaempferol from tea, broccoli, Delphinium, Witch-hazel, grapefruit etc. , Luteolin from Chromolaena odorata and Quercetin from Allium cepa were taken. All these compounds were shown to exhibit anticarcinogenic, anti diabetic and antimicrobial effects and their references were shown in Table I. For all the four compounds namely andrographolide, kaempferol, luteolin and quercetin except gingerol literature proof has been available to shown inhibitory effects towards PI 3-kinase except for Glycerol . Hence the compounds were taken for the study of binding affinities towards the protein PI 3-kinase, despite their lack of strict specificity, the study provided valuable bases for the prediction of natural compounds could be specific inhibitors of PI 3-kinase. Through this we could predict that these compounds exhibit diverse effects by inhibiting PI 3-kinases

Materials and Methods
Bioinformatics online databases such as pubmed, PDB and Pubchem, were used. PubMed database developed by the National Center for Biotechnology Information (NCBI) at the National Library of Medicine (NLM) is designed to provide access to citations from biomedical journals. From PubMed we have collected literatures on PI 3-kinases, and flavanoids.
Understanding the interactions between proteins and ligands is crucial for the pharmaceutical and functional food industries. The experimental structures of these protein/ligand complexes are usually obtained, under highly expert control, by time-consuming techniques such as X-ray crystallography or NMR. These techniques are therefore not suitable for routinely screening the possible interaction between one receptor and thousands of ligands. To overcome this limitation, computational algorithms (i.e. docking algorithms)

Docking
A number of powerful software programs, e.g. AutoDock, HEX, GOLD, FlexX, DOCK, Glide, Surflex, LigandFit, have been developed over the past several decades to carry out docking calculations, and good success in both binding mode and binding affinity prediction has often been achieved in selected test cases. We used a new shape-based method, LigandFit, for accurately docking ligands into protein active sites. The method employs a cavity detection algorithm for detecting invaginations in the protein as candidate active site regions. A shape comparison filter is combined with a Monte Carlo conformational search for generating ligand poses consistent with the active site shape. Candidate poses are minimized in the context of the active site using a gridbased method for evaluating protein-ligand interaction energies. The method appears quite promising, reproducing the X-ray structure ligand pose within an RMSD of 2A. A high-throughput screening study applied to the thymidine kinase receptor is also presented in which LigandFit, when combined with LigScore, an internally developed scoring function, yields very good hit rates for a ligand pool seeded with known actives (Venkatachalam et al., 2003). Thus docking analysis of Gingerol, kaempferol, luteolin, andrographolide and Quercetin with PI3 Kinase was carried out by Ligand Fit of Discovery studio (Version 1.7, Accelry's Software Inc.). The software allows us to virtually screen a database of compounds and predict the strongest binders based on various scoring functions. It explores the ways in which these five molecules and the enzyme PI3 Kinase fit together and dock to each other well, like pieces of a three-dimensional jigsaw puzzle. The collection of Gingerol, kaempferol, luteolin, andrographolide and Quercetin and PI3 Kinase complexes was identified via docking and their relative stabilities were evaluated using their binding affinities.

Ligand Preparation
The three dimensional structures of anticancer compounds like Gingerol, kaempferol, luteolin, andrographolide and Quercetin were downloaded in .sdf format from Pubchem database. Hydrogen Bonds were added and the energy was minimized using CHARMm force field. Molecular weight, log P and number of Hydrogen-bond donors and acceptors for the active principles were noted (shown in Table III). All the five molecules were satisfied Lipinski's drug properties and their two dimensional structures were shown in Figure 3.

Protein Selection
Sequences of Phosphoinositide 3-kinases catalytic subunit alpha isoform were retrieved from swissprot for various species in FASTA Format for multiple sequence alignment and for phylogenetic analysis using ClustalW. Phylogenetic analysis revealed that Mus musculus and Bovine were closely related to Human (Shown in Fig 1), but the three dimensional structures were available only for Hu-man and Sus scrofa. Hence their structures were retrieved and compared for further analysis.
There are several PDB structures available for the same protein and they are listed in table II along with their resolution and length. The PDB structure which was chosen for our study has a good resolution of 2.00 when compared to other structures. To predict the binding mechanism accurately, PDB structure (PDB ID: 1E7U) of Sus scrofa PI3 Kinase was chosen for the interaction analysis which is of 961 aminoacids. The PDB structure was also compared using the DALI server to find the structural alignment using the RMSD value as shown in Fig 2. As the RMSD score for the three dimensional structures of human PI3 kinase and Sus scrofawere below 2.00A, the structure from 0 Sus scrofa could be taken for further analysis.

Protein Preparation
The ligands and crystallographic water molecules were removed from the protein; and the chemistry of the protein was corrected for missing hydrogen. Crystallographic disorders and unfilled valence atoms were corrected using alternate conformations and valence monitor options. Following the above steps of preparation, the protein was subjected to energy minimization using the CHARMm force field.

Docking Studies
The active site of the protein was first identified and it is defined as the binding site resulted in a cavity size of 3475 point units. There is evidence that wortmannin alkylates a lysine residue at the putative ATP binding site of p110α (Wymann et al.,1996). LY294002, in contrast, is a pure competitive inhibitor of ATP. The X-ray structure of wortmannin, LY294002 and several broad-spectrum kinase inhibitors, including quercetin in complex with p110, confirms the mechanism of inhibition and offers a basis for designing more specific compounds (Walker et al.,2000). Thus Binding sites were defined based on the ligands already present in the PDB file (i.e. ATP binding site region) which were followed      by site sphere definition. Here site 1 was chosen as the binding site and the site sphere size was set to (434.375 A° 3 , Partition level 1).The determination of the ligand binding affinity was calculated using LigScore and PLP1,JAIN and Dock score were used to estimate the ligand-binding energies. Apart from these, other input parameters for docking were set as default options.

Results and Discussion
Molecular Docking continues to holds great promise in the field of Computer based drug design which screens small molecules by orienting and scoring them in the binding site of a protein. As a result novel ligands for receptors of known structure were designed and their interaction energies were calculated using the scoring functions (Irwin et al., 2002). Number of reports citing successful application of CADD in developing specific drugs in different therapeutic areas is expanding rapidly. A very interesting example which can also serve as a proof of principle of the in silico approach involves a type I TGF β receptor kinase inhibitor. The same molecule (HTS-466284/LY-364947), a 27 nM inhibitor, was discovered independently using virtual screening by Biogen IDEC (Singh et al., 2003) and traditional enzyme and cellbased high-throughput screening (Sawyer JS et al., 2003). Another in silico modeling drug development program led to clinical trials of a novel, potent, and selective anti-anxiety, anti-depression 5-HT 1A agonist in less than 2 years from the start and requiring less than 6 months of lead optimization and synthesis of only 31 compounds (Becker et al., 2006). It is estimated that docking programs currently dock 70 -80% of ligands correctly (Congreve et al., 2005).

Validation of Docking Results
To ensure that the ligand orientation obtained from the docking studies were likely to represent valid and reasonable binding modes of the inhibitors, the LigandFit program docking parameters had to be first validated for the crystal structure's Active site (PDBid 1E7U). Protein Utilities and Health protocol of Discovery's studio was used to find out the active sites in the structure and it was found that the active site contains amino acids such as ASP950 TYR867 MET804 GLU880 LYS808 VAL882 SER806 ILE831 ILE879 ASP964 LYS833 TRP812. Results of docking showed that LigandFit determined the optimal orientation of the docked inhibitor, exactly to these active sites.  Here top ranked ligands were taken for binding affinity studies. The validation process consisted of two parts: (i) Hydrogen bond details of the top-ranked docked pose and (ii) prediction of Binding energy between the docked ligand and the enzyme using various score calculated using Discovery studio (DJD, 2 LigScore2, 3 LigScore1, 3 PLP, 45 PMF, 46 and JAIN47 scores were taken for the analysis.

Hydrogen Bond Details
A close view of the binding interactions of PI3 kinase with the flavanoids Andrographolide, Kaempferol, Querce-tin, Luteolin and Gingerol were shown in Fig. 4. Ligand is coloured in Yellow (in balland-stick drawing) where as amino acids involved in Hydrogen bonding where shown in blue colour.
As shown in Fig. 4A, there are five hydrogen bonds (shown as green dotted lines) formed between the compound Andrographolide and PI3 Kinase. Those residues involved in forming hydrogen bonds with the enzyme were: (    ted lines in Fig.4B) and the residues involved in forming the hydrogen bonds from the enzyme were: Lys-807 and ASP-950. Kaempferol forms two hydrogen bonds (shown as green dotted lines in Fig.4C) and here the residues involved in forming the hydrogen bonds from the enzyme were: Asp-950 and Asp 964. Table 4 showed the detailed information of the hydrogen bonds. Luteolin and Gingerol forms three hydrogen bonds (shown as green dotted lines in Fig.4D and Fig 4E) and the residues involved in forming the hydrogen bonds from the enzyme were: Lys-807 and Asp-950. The detailed atoms in forming the hydrogen bonds are given in Table IV, V,VI,VII and VIII for each flavanoids separately, which may provide useful information for in-depth understanding binding mechanism of the compound to the active site of the protein.

Docking Score and RMSD Values
As a result of docking there were 10 different conformations were generated for andrographolide, Quercetin, Kaempferol, luteolin and for gingerol. But only for top ranked docked complex the scores were copied from the table browser view of Discovery studio for binding affinity analysis. Table IX shown the different score values of top ranked ligands. The score values include Ligscore1&2 (Protein-Ligand Affinity Energy) ( Krammer et al., 2005), PLP1, PLP2 (Steric and H-bonding intermolecular function, Higher PLP scores indicate stronger receptor-ligand binding (larger pK i values)) (Gehlhaar et al., 1995(Gehlhaar et al., ,1999, JAIN(sum of five interaction terms namely Lipophilic interactions,Polar attractive interactions ,Polar repulsive interactions ,Solvation of the protein and ligand ,An entropy term for the ligand) ( Jain 1996), PMF(developed based on statistical analysis of the 3D structures of protein-ligand complexes, scores are calculated by summing pairwise interaction terms over all interatomic pairs of the receptor-ligand complex, A higher score indicates a stronger receptor-ligand binding affinity) (Muegge 2006;Muegge et al., 1999) and Dockscore(Candidate ligand poses are evaluated and pri-oritized according to the DockScore function) . The determination of the ligand binding affinity was calculated using the shape-based interaction energies of the ligand with the protein. The two scoring methodologies namely LigScore and PLP1 were used to estimate the ligand-binding energies. Larger score value indicates better ligand-binding affinity.

Conclusion
The Protein-Ligand interaction plays a significant role in structural based drug designing. In the present work we have taken the enzyme PI3 Kinase and the drugs to explore the binding mechanism of flavanoids to the PI3 kinase enzyme. They are Andrographolide, Gingerol, Kaempferol, luteolin and Quercetin. When the enzyme docked to the five anticancer compounds the scores obtained were shown in Table.: Andrographolide:  It was also predicted that the compound gingerol showed good binding affinities towards the protein when compared to others. For all the four compounds like kaempferol, Quercetin, luteolin and Andrographolide literature proofs were available to indicate that they inhibit PI3-Kinase but for gingerol there is no such a proof is available. Here through in silico approach it was predicted that the compound gingerol also shown to inhibit PI3-Kinase as it had good Ligscore and PLP1 when compared to Quercetin and Luteolin. Hydrogen bond formation also makes important contributions to the interactions between ligand and the enzyme. Here a maximum of four hydrogen bonds were formed between the protein and the ligand Andrographolide followed by three hydrogen bonds were formed between the enzyme and the ligand Gingerol and luteolin. Thus the concept of protein-Ligand interaction helps in analyzing the binding properties of the protein PI3-Kinase with its inhibitors. The study report also concluded that the residues Lys 802, Lys-807, Asp-950, Asp 964 plays an important role in binding mechanism. Hence drugs such as Luteolin and Gingerol which were shown similar binding mechanism and good docking score to quercetin could be the lead one to target the PI3 Kinase. Our results provide insight into the structural requirement for the activity of the inhibitor and the most favorable binding mode of the top ranking compounds will be useful in designing new derivatives of Luteolin and gingerol as PI3 kinase inhibitors similar to the quercetin derivative of LYS2002.

Future Perspectives
Understanding the interactions between proteins and ligands is crucial for the pharmaceutical and functional food industries. The experimental structures of these protein/ligand complexes are usually obtained, under highly expert control, by time-consuming techniques such as X-ray crystallography or NMR. Molecular modeling and molecular docking methods still have a long way to run before producing completely reliable results. This could be achieved by NMR screening remains a multifaceted and unique technique that is sensitive to both structure and dynamics and that can monitor the binding of low molecular weight ligands to biological macromolecules in the early stages of drug discovery due to its ability to detect even very weak binders.