Theoretical studies of ligand-receptor interactions in ion channels

人类基因组编码超过400种形成离子通道的蛋白质。电压门控的K+、Na+和Ca2+通道是控制疼痛和治疗心血管、神经、自身免疫和其他疾病的药物靶点。这些通道在结构上以不同的闭合、打开和失活状态存在。药物通常优先与某些状态结合。这种依赖状态的作用很难通过高通量配体筛选来研究，而高通量配体筛选是药物发现的主要实验方法。分子模型有助于理解状态依赖性药物作用的原子机制。2009-2013年，我们在PNAS USA， Chem Reviews, Trends Pharmacol Sci， J Biol Chem等期刊上发表了13篇nserc资助论文。我们将继续进行药物和毒素离子通道的理论研究。我们的方法包括5个阶段。1)以K+和Na+通道在开放、关闭和失活状态下的现有x射线结构为模板，建立具有重要医学意义的离子通道的同源性模型。2)分析已发表的配体构效关系及配体结合位点的突变分析。3)揭示实验数据中的悖论，关注缺乏结构性解释或提出可疑数据的数据。4)对接同源模型中的代表性配体，考虑配体与渗透阳离子直接相互作用的可能性（Zhorov & Tikhonov, 2013, Trends Pharmacol Sci）。对接实验通常预测一个能量合理的配体结合模式的集合。我们选择那些最符合实验数据的。我们将使用在ZMM程序中实现的蒙特卡罗最小化方法，这是我多年来一直在阐述的，作为主要的计算工具。ZMM已得到验证，例如（Garden & Zhorov, 2010）。自制软件的一个优点是它可以用来解决新问题。需要结构解释的已发表的实验数据并不缺乏。最近细菌Na+通道的x射线结构为毒素和医学上重要药物的Na+和Ca2+通道模型提供了可靠的模板。我也为我的研究计划提出了一个新的方向。突变分析是用来确定配体结合位点，但结构的解释结果是困难的。事实上，配体的作用可能由于配体结合位点的残基突变（直接效应）或远超过该位点（变构效应）而改变。我们将阐述一种区分直接效应和变构效应的方法。驱动假说是，突变改变了离子通道段之间依赖状态的接触，改变了开放/关闭/失活状态的种群，从而可能变构地影响药物作用。为了区分这些影响，我们将划分通道能量，选择涉及强段间契约的残基，探索这些接触在不同通道状态下如何变化，并根据这些结果分析突变数据。这种方法将有助于解释为什么Na+和Ca2+通道中保守天冬酰胺的突变（不面向孔）会影响许多孔靶向配体的作用。结构相似的配体PAP-1和Psora-4对Kv1通道阻断的突变分析导致了配体结合模型的冲突（Zimin et al. 2010; Marzian et al. 2013）。我们将分析潜在的突变，并通过这些高效配体重新解决Kv1通道阻断的机制。区分突变的直接和变构效应对建立其他配体通道模型很重要。我们的研究将有助于基本知识和新药物的开发，这将有助于维护和改善加拿大人的健康。

The human genome encodes over 400 proteins that form ion channels. Voltage-gated K+, Na+, and Ca2+ channels are drug targets for pain control and treatment of cardiovascular, neurological, autoimmune and other disorders. The channels exist in structurally different closed, open and inactivated states. Drugs often bind preferentially to certain states. Such state-dependent action is difficult to study by high-throughput ligand screening, the major experimental approach in drug discovery. Molecular modeling helps understand atomistic mechanisms of state-dependent drug action. In 2009-2013 we have published 13 NSERC-funded papers in PNAS USA, Chem Reviews, Trends Pharmacol Sci, J Biol Chem and other journals. We will continue theoretical studies of ion channels with drugs and toxins. Our approach includes 5 stages. 1) Building homology models of medicinally important ion channels using as templates available X-ray structures of K+ and Na+ channels in the open, closed, and inactivated states. 2) Analyzing published data on structure-activity relations of ligands and mutational analysis of ligand-binding sites. 3) Revealing paradoxes in the experimental data and focusing on data that lack structural interpretations or suggest questionable ones. 4) Docking representative ligands in the homology models with considering a possibility that ligands may directly interact with permeant cations (Zhorov & Tikhonov, 2013, Trends Pharmacol Sci). 5) Docking experiments usually predict an ensemble of energetically reasonable ligand-binding modes. We select those that are best consistent with experimental data. We will use the Monte Carlo-minimization method realized in the ZMM program, which I am elaborating over many years, as a major computational tool. ZMM has been validated, e.g., in (Garden & Zhorov, 2010). An advantage of the home-made software is that it can be adapted to address new problems. There is no shortage of published experimental data that need structural interpretations. Recent x-ray structures of bacterial Na+ channels provide reliable templates to model Na+ and Ca2+ channels with toxins and medically important drugs. I also propose a new direction for my research program. Mutational analysis is used to determine ligand-binding sites, but structural interpretation of results is difficult. Indeed, a ligand action may change due to mutation of a residue in the ligand-binding site (a direct effect) or far beyond this site (an allosteric effect). We will elaborate an approach to discriminate the direct and allosteric effects. The driving hypothesis is that mutations, which change state-depended contacts between ion channel segments, change populations of the open/closed/inactivated states and thus may allosterically affect drug action. To discriminating these effects we will partition the channel energy, select residues involved in strong intersegment contracts, explore how such contacts change in different channel states, and analyze mutational data in view of these results. This approach will help explain why mutations of conserved asparagines in Na+ and Ca2+ channels, which do not face the pore, affect action of many pore-targeting ligands. Mutational analyses of Kv1 channels block by structurally similar ligands PAP-1 and Psora-4 resulted in conflicting ligand-binding models (Zimin et al. 2010; Marzian et al. 2013). We will analyze underlying mutations and readdress the mechanism of Kv1 channel block by these highly potent ligands. Discriminating direct and allosteric effects of mutations will be important for building other ligand-channel models. Our studies will contribute to basic knowledge and development of new drugs that will help maintain and improve health of Canadians.