Modeling and prediction of binding affinities between the human amphiphysin SH3 domain and its peptide ligands using genetic algorithm-Gaussian processes

In this article, we discuss the application of the Gaussian process (GP) and other statistical methods (PLS, ANN, and SVM) for the modeling and prediction of binding affinities between the human amphiphysin SH3 domain and its peptide ligands. Divided physicochemical property scores of amino acids, involving significant hydrogen bond, electronic, hydrophobic, and steric properties, was used to characterize the peptide structures, and quantitative structure-affinity relationship models were then constructed by PLS, ANN, SVM, and GP coupled with genetic algorithm-variable selection. The results show that: (i) since the significant flexibility and high complexity possessed in polypeptide structures, linear PLS method was incapable of fulfilling a satisfying behavior on SH3 domain binding peptide dataset; (ii) the overfitting involved in training process has decreased the predictive power of ANN model to some extent; (iii) both SVM and GP have a good performance for SH3 domain binding peptide dataset. Moreover, by combining linear and nonlinear terms in the covariance function, the GP is capable of handling linear and nonlinear-hybrid relationship, and which thus obtained a more stable and predictable model than SVM. Analyses of GP models showed that diversified properties contribute remarkable effect to the interactions between the SH3 domain and the peptides. Particularly, steric property and hydrophobicity of P(2), electronic property of P(0), and electronic property and hydrogen bond property of P(-3) in decapeptide (P(4)P(3)P(2)P(1)P(0)P(-1)P(-2)P(-3)P(-4)P(-5)) significantly contribute to the binding affinities of SH3 domain-peptide interactions.     

Proteome-Wide Inference of Src Homology 3 Domain-Binding Peptides

Src homology 3 (SH3) domain is a versatile protein structure module that participates in mediating various protein?Cprotein binding events by specifically recognizing proline-rich region of diverse plasmatic proteins. Reliable and fast inference of SH3-binding partners over the human proteome are fundamentally important for our understanding of the molecular functions and biological implications underlying SH3-mediated signaling network. Herein, we employ an atom-realistic protocol to perform proteome-wide inference of SH3-binding peptides using the information gained from both the primary sequence of affinity-known peptides and the interaction properties deriving from SH3?Cpeptide complex structures. It is revealed that the binding affinity and specificity of peptides to SH3 domain are co-contributed from electrostatic, steric and hydrophobic effects, and the hydrophobicity and electrostatic property at P₂, P₀ and/or P_?3 play an essential role in determining the binding. In addition, SH3 domain exhibits a broad specificity of recognizing its ligands and thus a large number of protein candidates that might be the potential interacting partners of SH3 domain are extracted from the human proteome, from which several samples are suggested to be the highly promising SH3 binders.     

Quantitative Structure–Activity Relationship Model for Prediction of Protein–Peptide Interaction Binding Affinities between Human Amphiphysin-1 SH3 Domains and Their Peptide Ligands

Protein–protein interaction plays a critical role in signal transduction and many other key biological processes. The present study evaluated four parameters selected from among 554 physiochemical variables of 20 natural amino acids listed in AAindex, namely, hydrophobicity, electronic properties, steric properties, and hydrogen-bond properties. Human amphiphysin-1 Src homology 3 (SH3) domain-binding decapeptides were the object of analysis. A quantitative structure–activity relationship model of the SH3 domain-binding peptides was constructed using multivariate linear regression. The results showed that the four parameters ably characterize the structure of SH3 domain-binding decapeptides, have definitive physicochemical properties and a low level of computational complexity, are accessible, and may be used in integrated prediction models for other protein–peptide interactions.     

Design of novel quinazolinone derivatives as inhibitors for 5HT7 receptor

To study the pharmacophore properties of quinazolinone derivatives as 5HT₇ inhibitors, 3D QSAR methodologies, namely Comparative Molecular Field Analysis (CoMFA) and Comparative Molecular Similarity Indices Analysis (CoMSIA) were applied, partial least square (PLS) analysis was performed and QSAR models were generated. The derived model showed good statistical reliability in terms of predicting the 5HT₇ inhibitory activity of the quinazolione derivative, based on molecular property fields like steric, electrostatic, hydrophobic, hydrogen bond donor and hydrogen bond acceptor fields. This is evident from statistical parameters like q² (cross validated correlation coefficient) of 0.642, 0.602 and r² (conventional correlation coefficient) of 0.937, 0.908 for CoMFA and CoMSIA respectively. The predictive ability of the models to determine 5HT₇ antagonistic activity is validated using a test set of 26 molecules that were not included in the training set and the predictive r² obtained for the test set was 0.512 & 0.541. Further, the results of the derived model are illustrated by means of contour maps, which give an insight into the interaction of the drug with the receptor. The molecular fields so obtained served as the basis for the design of twenty new ligands. In addition, ADME (Adsorption, Distribution, Metabolism and Elimination) have been calculated in order to predict the relevant pharmaceutical properties, and the results are in conformity with required drug like properties.     

Docking and 3D QSAR study of thiourea analogs as potent inhibitors of influenza virus neuraminidase

Surflex-Dock was applied to study interactions between 30 thiourea analogs and neuraminidase (NA). The docking results showed that hydrogen bonding and electrostatic interactions were highly correlated with the activities of neuraminidase inhibitors (NIs), followed by hydrophobic and steric factors. Moreover, there was a strong correlation between the predicted binding affinity (total score) and experimental pIC₅₀ (correlation coefficient r = 0.870; P < 0.0001). A three dimensional holographic vector of atomic interaction field (3D-HoVAIF) was employed to construct a QSAR model. The r ², q ² and r ² _test values of the optimal QSAR model were 0.849, 0.724 and 0.689, respectively. From the QSAR model, it could be seen that electrostatic, hydrophobic and steric interactions were closely related to inhibitory activity, which was consistent with the docking results. Based on the docking and QSAR results, five new compounds with high predicted activities were designed.     

Models of the pharmacophoric pattern and affinity trend of methyl 2-(aminomethyl)-1-phenylcyclopropane-1-carboxylate derivatives as σ1 ligands

Sigma-1 (σ₁) affinities of methyl 2-(aminomethyl)-1-phenylcyclopropane-1-carboxylate (MAPCC) derivatives were modelled by the genetic algorithm with linear assignment of hypermolecular alignment of datasets (GALAHAD) and the comparative molecular field analysis (CoMFA)/comparative molecular similarity indices analysis (CoMSIA) methods. GALAHAD was used for deriving the 3D pharmacophore pattern that encompasses the most potent σ₁ ligands within this series. Five MAPCC derivatives with a high σ₁ affinity were used for deriving this model. The obtained model included a nitrogen atom, the hydrophobes and the hydrogen bond acceptor features; it was able to identify other potent σ₁ ligands. On the other hand, CoMFA and CoMSIA methods were used for deriving quantitative structure–activity relationship (QSAR) models. All QSAR models were trained with 17 compounds, after which they were evaluated for predictive ability with additional five compounds. The best QSAR model was obtained by using CoMSIA, including steric, electrostatic and hydrophobic fields, and had a good predictive quality according to both internal and external validation criteria. In general, the models described herein provide meaningful information relevant for the rational design of new σ₁ ligands.