Computational Analysis for Residue-Speci c CDK2-Inhibitor Bindings

Cyclin-dependent kinase 2 (CDK2) is a key macromolecule in cell cycle regulation. In cancer cells, CDK2 is often overexpressed and its inhibition is an effective therapy of many cancers including breast carcinomas, leukemia, and lymphomas. Quantitative characterization of the interactions between CDK2 and its inhibitors at atomic level may provide a deep understanding of protein-inhibitor interactions and clues for more effective drug discovery. In this study, we have used the computational alanine scanning approach in combination with an efficient interaction entropy method to study the microscopic mechanism of binding between CDK2 and its 13 inhibitors. The total binding free energy from the method shows a correlation of 0.76?0.83 with the experimental values. The free energy component reveals two binding mode in the 13 complexes, namely van der Waals dominant, and electrostatic dominant. Decomposition of the total energy to per-residue contribution allows us to identify five hydrophobic residues as hot spots during the binding. Residues that are responsible for determining the strength of the binding were also analyzed.     

Inhibition of CK2 Activity by TCDD via Binding to ATP-competitive Binding Site of Catalytic Subunit: Insight from Computational Studies

Alternative mechanisms of toxic effects induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin(TCDD), instead of the binding to aryl hydrocarbon receptor(AhR), have been taken into consideration. It has been recently shown that TCDD reduces rapidly the activity of CK2(casein kinase II) both in vivo and in vitro. It is found that TCDD has high molecular similarities to the known inhibitors of CK2 catalytic subunit(CK2α). This suggests that TCDD could also be an ATP-competitive inhibitor of CK2α. In this work, docking TCDD to CK2 was carried out based on the two structures of CK2α from maize and human, respectively. The binding free energies of the predicted CK2α-TCDD complexes estimated by the molecular mechanics/Poisson-Boltzmann surface area(MM/PBSA) method are from -85.1 kJ/mol to -114.3 kJ/mol for maize and are from -96.1 kJ/mol to -118.2 kJ/mol for human, which are comparable to those estimated for the known inhibitor and also ATP with CK2α. The energetic analysis also reveals that the van der Waals interaction is the dominant contribution to the binding free energy. These results are also useful for designing new drugs for a target of overexpressing CK2 in cancers.     

Assessing the performance of the g_mmpbsa tools to simulate the inhibition of oseltamivir to influenza virus neuraminidase by molecular mechanics Poisson–Boltzmann surface area methods

The molecular mechanics Poisson–Boltzmann surface area (MM-PBSA) method for GROMACS (g_mmpbsa) is an open-source tool that is capable of reading the trajectories generated by GROMACS and calculating the binding free energy using the MM-PBSA method. However, there are multiple force fields available for users to choose from in the GROMACS software, and there are also different solvent water models to combine with the chosen force fields. These different combinations of parameters may significantly impact the results of g_mmpbsa calculation. Unfortunately, the exact combination of force field and solvent water that can well calculate the free energy of the receptor–ligand binding in GROMACS has not been explored yet. To resolve the above issues, this study mainly explored the molecular dynamics (MD) simulations by GROMACS with the six commonly used force fields and three solvent water models, in combination with g_mmpbsa, to calculate the binding free energies of the influenza virus neuraminidase and its mutants with inhibitor oseltamivir carboxylate and compared the present results with previous published results of Amber software from ours and other researchers. Finally, we provided an optimized calculation model, as well as suggestions that may serve as advice and guidance for future computer-aided designs of drug molecules.     

分子动力学模拟研究质子化态在HIV-1 Protease-Indinavir复合物中的作用

I 型人体免疫缺陷病毒(HIV-1)蛋白酶中Asp25/Asp25'的质子化对于理论研究HIV-1 蛋白酶和抑制剂的作用机制
以及氨基酸变异对抗药性的影响有重要意义. 分别对Protease-Indinavir (PR-IDV)复合物的六种可能的质子化态进行了
5 ns 的分子动力学模拟, 分析了不同状态对动力学特征和结构的影响, 用molecular mechanics/Possion-Boltzman surface
area (MM-PBSA)方法计算了PR 和IDV 在各种状态下的结合自由能. 计算结果说明A 链Asp25 的OD2 的质子化是最
为可能的状态. 对PR-IDV 复合物中起到媒介作用的水分子与PR-IDV 复合物形成的氢键进行了分析, 分析结果说明不
同的质子化态对水分子在PR-IDV 复合物中所起的媒介作用没有影响, 这一结果与我们先前对PR-BEA369 复合物的研
究不同. 我们的研究结果为更高效的PR 抑制剂的设计以及PR 氨基酸变异对药物抗药性的研究提供了理论上的指导.     

谷氨酰胺结合蛋白的分子动力学模拟和自由能计算

谷氨酰胺结合蛋白(Glutamine-binding protein, GlnBp)是大肠杆菌透性酶系统中一个细胞外液底物专一性结合蛋白, 对于细胞外液中谷氨酰胺(Gln)的运输和传递至关重要. 本文运用分子动力学(Molecular dynamics, MD)模拟采样, 考察了GlnBp关键残基与底物Gln之间的相互作用和GlnBp两条铰链的功能差别; 并采用MM-PBSA方法计算了GlnBp与底物Gln的结合自由能. 结果表明: Ph13, Phe50, Thr118和Ile69与底物Gln的范德华相互作用和Arg75, Thr70, Asp157, Gly68, Lys115, Ala67, His156与底物Gln的静电相互作用是结合Gln的主要推动力; 复合物的铰链区85～89柔性大, 对构象开合提供了结构基础; 而铰链区181～185柔性小, 其作用更多是在功能上把底物Gln限制在口袋中; 自由能预测值与实验值吻合. 本研究很好地解释了GlnBp结构与功能的关系, 为进一步了解GlnBp的开合及转运Gln的机制提供了重要的结构信息.     

疏水蛋白吸附碳纳米管的分子动力学模拟

采用分子动力学方法模拟了水溶液中Ⅱ型疏水蛋白HFBI在单壁碳纳米管(SWNTs)表面的吸附过程, 考察了3种不同的HFBI初始取向, 并计算了结合自由能, 从累计240 ns的模拟轨迹中得到了不同的吸附结构. 结果表明, 当HFBI完全通过疏水面与SWNTs作用时, 其结合自由能最有利吸附, 且吸附最稳定. 另外, 由于HFBI含有4个二硫键, 因此吸附过程几乎不改变其二级结构.     

Effect of the R119G mutation on human P5CR structure and its interactions with NAD: Insights derived from molecular dynamics simulation and free energy analysis

Pyrroline-5-carboxylate reductase (P5CR), an enzyme with conserved housekeeping roles, is involved in the etiology of cutis laxa. While previous work has shown that the R119G point mutation in the P5CR protein is involved, the structural mechanism behind the pathology remains to be elucidated. In order to probe the role of the R119G mutation in cutis laxa, we performed molecular dynamics (MD) simulations, essential dynamics (ED) analysis, and Molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) binding free energy calculations on wild type (WT) and mutant P5CR-NAD complex. These MD simulations and ED analyses suggest that the R119G mutation decreases the flexibility of P5CR, specifically in the substrate binding pocket, which could decrease the kinetics of the cofactor entrance and egress. Furthermore, the MM-PBSA calculations suggest the R119G mutant has a lower cofactor binding affinity for NAD than WT. Our study provides insight into the possible role of the R119G mutation during interactions between P5CR and NAD, thus bettering our understanding of how the mutation promotes cutis laxa.     

HIV-1蛋白酶与抑制剂作用的结合自由能计算

HIV-1蛋白酶是治疗艾滋病的重要靶标酶之一. 采用分子动力学模拟, 运用MM-PBSA方法计算了HIV-1蛋白酶与三个抑制剂BE4, BE5和BE6的结合自由能, 结果表明抑制剂P₁/ 位置的苄基上双氟原子的不同位置对结合自由能产生不同的影响. 通过能量分解的方法考察了HIV-1蛋白酶的主要残基与三个抑制剂间的相互作用与识别, 结果表明三个抑制剂以相同的作用模式与HIV-1蛋白酶结合, 计算结果与实验结果基本吻合.     

Investigation of hydroxyl-mediated spectral interference from easily ionized elements in laser-enhanced ionization spectrometry in the 281–285 nm range

The spectral background from 281 to 285 nm in the laser-enhanced ionization (LEI) spectrum of aqueous samples containing easily ionized elements (EIE) at concentrations similar to those found in blood was investigated. A complex, structured spectral background was observed, which appears in the presence of Na or K, but does not match the spectral signature of either element. The same behavior was also observed for Li. It was established that this background originates from an energy transfer between laser-excited hydroxyl (OH) molecules and ground-state EIEs. The intensity of this spectral feature was found to increase with EIE concentration and applied electrode voltage. This unexpected source of spectral interference may complicate the determination of trace metals by LEI in the presence of EIEs, since it can not be prevented by simply avoiding interference from atomic lines.     

Computational study of the molecular hydrogen physisorption in some of the corannulene derivatives as a carbon nanostructure

Substitution of the peripheral H atoms in the corannulene molecule as a carbon nanostructure by OH, CH₃, NH₂ and NO₂ groups on the molecular hydrogen physisorption was evaluated at MP2/6-31G(d) level of theory. Two orientations of hydrogen were used on the concave and convex sides of corannulene. It was seen that binding to the concave face is favored relative to the convex face. The average binding energy was calculated and corrected for the basis set superposition error (BSSE) using the counterpoise method. Results showed that binding energy varies depending upon the site and side of absorption. The electronic density, charge transfer and spatial prohibition of the substituted groups affects the binding energy. The increment of the electronic density because of the substitution of electron donor groups facilitates hydrogen adsorption and leads to larger binding energies than when H atoms are substituted by electron acceptor groups. Substitution of H atoms with each of the considered groups leads to decreasing of the HOMO–LUMO energy gap and so decreasing of the kinetic stability and increasing of the reactivity. The energy gap and binding energy for corannulene derivatives decreases in the order of: CH₃ > OH > NH₂ > NO₂.     

用1H NMR相对化学位移计算CyDTA金属配合物乙羧基键合几率的方法

用~１ＨＮＭＲ相对化学位移计算ＣｙＤＴＡ金属配合物乙羧基键合几率的方法宋瑞方,李菲,刘广民,毛友钢（吉林大学理论化学研究所分子光谱与分子结构开放实验室,长春,１３００２３）关键词ＭＣｙＤＴＡ,乙羧基键合分布几率,标准自由能胺羧金属配合物一般为正八面体?..     

In silico analysis of carbohydrate-binding pockets in the lectin genes from various species of Canavalia

Legumes are endowed with an opulent class of proteins called lectins that can detect tenuous variations in carbohydrate structures and bind them reversibly with high affinity and specificity. The genus Canavalia, in the family of Leguminosae, is considered to be an affluent source of lectin. An effort has been made to analyse the sequences encoded by the lectin gene and its carbohydrate binding pockets from three species of Canavalia, including C. virosa, C. rosea, and C. pubescens. Crude seed extract showed highest haemagglutination titer against buffalo RBCs and has high affinity to mannose and trehalose. Amplification of the lectin gene by gene-specific primers showed the presence of an 870 bp amplicon. Physicochemical characterization using various bioinformatic tools showed that the isoelectric point was below 7, suggesting that lectin molecules were acidic. A high aliphatic index and high instability index were observed, which indicated that lectin molecules were stable towards a wide range of temperatures. The occurrence of N-glycosylation sites at two sites was also identified in all three species. Prediction of secondary structure showed that approximately 59.05 %, 56.76 % and 54.88 % of the elements were random coils in the case of C. virosa, C. pubescens and C. rosea, respectively. Comparative modelling of the proteins and docking of hypothetical models with sugar moieties that inhibited the agglutination activity suggested that asparagine, serine, alanine, valine, tyrosine and threonine were the major residues involved in hydrogen bonding and other stacking interactions. This can further provide insights on its prospective antibiosis property.     

Alpha7 nicotinic acetylcholine receptor agonists: prediction of their binding affinity through a molecular mechanics Poisson-Boltzmann surface area approach

A group of agonists for the alpha7 neuronal nicotinic acetylcholine receptors (nAChRs) was investigated, and their free energies of binding DeltaG(bind) were calculated by applying the molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) approach. This method, based on molecular dynamics simulations of fully solvated protein-ligand complexes, allowed us to estimate the contribution of both polar and nonpolar terms as well as the entropy to the overall free energy of binding. The calculated results were in a good agreement with the experimentally determined DeltaG(bind) values, thereby pointing to the MM-PBSA protocol as a valuable computational tool for the rational design of specific agents targeting the neuronal alpha7 nAChR subtypes.     

Application of Empirical and Quantum-chemical Computational Methods in the Determination of the Free Conformational Energy of Substituents in 1,3-Dioxanes

The advantages and disadvantages of empirical and quantum-chemical methods for the determination of the free conformational energy of methyl and phenyl substituents at the C₍₄₎ and C₍₅₎ atoms of the ring in the molecules of 1,3-dioxanes are analyzed.     

槲皮素与腺嘌呤氢键作用的最佳位点

本文优化得到了16个由槲皮素与腺嘌呤形成的氢键复合物的稳定结构,并计算了它们的结合能.研究发现,在气相和水相中,槲皮素均通过qu1位点与腺嘌呤作用形成稳定的氢键复合物.比较了腺嘌呤与槲皮素形成的氢键复合物、腺嘌呤与胸腺嘧啶形成的Watson-Crick碱基对的相对稳定性.在气相条件下Watson-Crick碱基对更稳定,在水相条件下腺嘌呤与槲皮素形成的氢键复合物更稳定,说明水相条件下腺嘌呤与槲皮素之间的相互作用强于与胸腺嘧啶之间的相互作用.基于标准反应Gibbs自由能变的计算结果估算了水相条件下腺嘌呤与槲皮素形成的氢键复合物和Watson-Crick碱基对的相对平衡浓度.     

Theoretical Investigating Mechanisms of Drug-Resistance Generated by Mutation-Induced Changes in Influenza Viruses

Influenza A (A/H\begin{document}$ x $\end{document}N\begin{document}$ y $\end{document}) is a significant public health concern due to its high infectiousness and mortality. Neuraminidase, which interacts with sialic acid (SIA) in host cells, has become an essential target since its highly conserved catalytic center structure, while resistance mutations have already generated. Here, a detailed investigation of the drug resistance mechanism caused by mutations was performed for subtype N9 (A/H7N9). Molecular dynamics simulation and alanine-scanning-interaction-entropy method (ASIE) were used to explore the critical differences between N9 and Zanamivir (ZMR) before and after R294K mutation. The results showed that the mutation caused the hydrogen bond between Arg294 and ZMR to break, then the hydrogen bonding network was disrupted, leading to weakened binding ability and resistance. While in wild type (A/H7N9\begin{document}$ ^{ \rm{WT}} $\end{document}), this hydrogen bond was initially stable. Meanwhile, N9 derived from A/H11N9 was obtained as an R292K mutation. Then the relative binding free energy of N9 with five inhibitors (SIA, DAN, ZMR, G28, and G39) was predicted, basically consistent with experimental values, indicating that the calculated results were reliable by ASIE. In addition, Arg292 and Tyr406 were hot spots in the A/H11N9\begin{document}$ ^{ \rm{WT}} $\end{document}-drugs. However, Lys292 was not observed as a favorable contributing residue in A/H11N9\begin{document}$ ^{ \rm{R292K}} $\end{document}, which may promote resistance. In comparison, Tyr406 remained the hotspot feature when SIA, ZMR, and G28 binding to A/H11N9\begin{document}$ ^{ \rm{R292K}} $\end{document}. Combining the two groups, we speculate that the resistance was mainly caused by the disruption of the hydrogen bonding network and the transformation of hotspots. This study could guide novel drug delivery of drug-resistant mutations in the treatment of A/H\begin{document}$ x $\end{document}N9.     

Computational Approach to Explore the B/A Junction Free Energy in DNA

Protein–DNA interactions induce conformational changes in DNA such as B‐ to A‐form transitions at a local level. Such transitions are associated with a junction free energy cost at the boundary of two different conformations in a DNA molecule. In this study, we performed umbrella sampling simulations to find the free energy values of the B–A transition at the dinucleotide and trinucleotide level of DNA. Using a combination of dinucleotide and trinucleotide free energy costs obtained from simulations, we calculated the B/A junction free energy. Our study shows that the B/A junction free energy is 0.52 kcal mol^?1 for the A‐philic GG step and 1.59 kcal mol^?1 for the B‐philic AA step. This observation is in agreement with experimentally derived values. After excluding junction effects, we obtained an absolute free energy cost for the B‐ to A‐form conversion for all the dinucleotide steps. These absolute free energies may be used for predicting the propensity of structural transitions in DNA.     

Computational chemistry in drug lead discovery and design

The main contributions of our group during the last 15 years developing and using biomolecular simulation tools in drug lead discovery and design, in close collaboration with experimental researchers, are presented. Special emphasis has been given to methodological improvements in the following areas: (1) target homology modeling incorporating knowledge about known ligands to accurately characterize the binding site; (2) designing alternative strategies to account for protein flexibility in high-throughput docking; (3) development of stochastic- and normal-mode-based methods to de novo design structurally diverse protein conformers; (4) development and validation of quantum mechanical semi-empirical linear-scaling calculations to correctly estimate ligand binding free energy. Several successful cases of computer-aided drug discovery are also presented, especially our recent work on viral targets.