Pre-computed free energy maps for rapid structure-based ligand design

描述（由申请人提供）：在拟议的研究中，SilcsBio， LLC将进一步开发一种新方法，即通过配体竞争饱和进行位点识别（SILCS），该方法已经为SilcsBio， LLC带来了多项商业销售，以解决客户对SILCS实用性的要求和担忧。配体设计计算方法的成功商业应用需要一个平台，该平台既可以提供指导设计过程的定性数据，也可以快速生成定量数据，以便对特定配体的可能性进行严格评估。这两个属性构成了SILCS技术的核心。但是，有必要扩大信息系统信息系统技术，使其能够为各种各样的计算软件包所用，并易于用于数据库筛选。最重要的是，虽然客户报告使用SILCS技术在预测能力方面有了很大的提高（将领先时间从12个月缩短到6个月，为“不可药物”目标识别独特的支架，更好地匹配评分与实际性能），但客户要求适度提高准确性，以便对新疗法的开发产生重大影响。对于给定的靶标，SILCS包括一个一次性的预先预处理步骤，在没有任何药物样配体的情况下，将目标蛋白、RNA或任何感兴趣的大分子靶标浸入小有机溶质的水溶液中，并进行详尽的分子动力学（MD）模拟。从这些模拟中，根据严格的自由能标准，包括蛋白质灵活性，生成了目标整个表面上不同化学功能类的3D概率分布。然后将这些概率分布转换为基于玻尔兹曼分布的自由能，产生“网格自由能（GFE）”。在纯水溶液中有机溶质的概率分布被归一化，使得GFEs包括溶质和蛋白质表面的脱溶以及溶质与蛋白质的相互作用的能量贡献。GFE分布，称为FragMaps，可以定性地识别蛋白质表面上与不同功能基团（“热点”）相互作用有利的区域，从而指导配体设计。GFEs也可以用来估计配体的相对自由结合能，从而促进对特定设计结果的定量评估。由于SILCS方法是基于目标分子的一次性预处理模拟，这一过程可以在商用计算机集群上几天完成，基于FragMaps的定量自由能结合估计可以在几秒钟内对大量化合物进行，从而允许评估大范围的配体修饰。SBIR的具体目标包括：1)将SILCS技术和GFE FragMaps扩展到计算化学软件包可访问的格式；2)提高SILCS GFE FragMaps的准确性；3)扩展SILCS GFE FragMaps用于数据库筛选。该提案第一阶段的成功将基于以下里程碑：1)SILCS计算平台：能够使用公共和商业上可用的计算化学软件生成、读取和可视化SILCS GFE FragMaps； 2) SILCS准确性：改进已知配体-蛋白质复合物的相对结合能预测；3)SILCS数据库筛选。使用基于SILCS gfe的药效团模型从数据库筛选中识别已知配体的能力。SILCS技术的这些实质性改进将通过SilcsBio现有的营销渠道实现商业化，其中包括与计算化学软件供应商合作，使该技术可用于制药公司内部使用，SilcsBio作为合同研究组织（CRO）执行基于结构的配体设计。CRO的工作包括向客户提供和解释GFE FragMaps，重要的是，该产品可以在客户不透露其IP的情况下提供，因为FragMaps是在没有配体的情况下生成的。I期提案的成功完成预计将为II期SBIR奠定基础，通过以下方式改进和扩展SILCS平台：1)将该技术扩展到无法进入周围水环境的靶配体结合位点，例如许多gpcr中常见的那些位点；2)将SILCS扩展为可销售的产品，供制药公司内部使用；3)开发SilcsBio成为成功CRO所需的必要基础设施。

DESCRIPTION (provided by applicant): In the proposed study, a novel approach which has already generated multiple commercial sales for SilcsBio, LLC, Site-Identification by Ligand Competitive Saturation (SILCS), will be further developed to address customer requests and concerns around SILCS utility. Successful commercial application of computational methods for ligand design require a platform that provides both qualitative data to direct the design process and rapid production of quantitative data to allow for rigorous evaluation of specific ligand possibilities Both of those attributes form the core of SILCS technology. However, it is necessary to extend the SILCS technology to be accessible to a wide range of computational software packages and to be readily used in database screenings. Most importantly, while customers report large improvements in predictive capacity using SILCS technology (reduced time to lead from 12 months to 6 months, identification of unique scaffolds for an 'undruggable" target, better matches of scoring to actual performance), modest improvements in accuracy are requested by customers in order to have a large impact on the development of new therapeutics. For a given target, SILCS involves a one-time up-front preconditioning step where the target protein, RNA or any macromolecule target of interest, in the absence of any drug-like ligands, is immersed in an aqueous solution of small organic solutes and subjected to exhaustive molecular dynamics (MD) simulations. From these simulations 3D probability distributions of different chemical functional classes on the entire surface of the target are generated based on rigorous free energy criteria, including protein flexibility. These probability distributions are then converted to free energies based on a Boltzmann distribution yielding "Grid Free Energies (GFE)." The probability distributions are normalized with respect to the organic solutes in pure aqueous solution, such that the GFEs include energetic contributions from desolvation of both the solutes and the protein surface, as well as interactions of the solutes with the protein. The GFE distributions, termed FragMaps, may then be used in a qualitative fashion to identify regions on the protein surface that interact favorably with differet classes of functional groups ("hot spots"), thereby directing ligand design. The GFEs may also be used to estimate ligand relative free energies of binding thereby facilitating quantitative evaluation of specific design outcomes. As the SILCS method is based on one-time preconditioning simulations of the target molecule, a process that can be completed in several days on commodity computer clusters, the quantitative free energy binding estimates based on the FragMaps can be performed on large numbers of compounds in a matter seconds allowing for wide ranges of ligand modifications to be evaluated. Specific goals for the proposed SBIR include 1) extending SILCS technology and GFE FragMaps to formats accessible to computational chemistry software packages, 2) improving the accuracy of SILCS GFE FragMaps and 3) extending SILCS GFEs for use in database screening. Success of the Phase I aspect of the proposal will be based on the following milestones: i) SILCS computational platform: Ability to generate, read and visualize SILCS GFE FragMaps using publically and commercially available computational chemistry software, 2) SILCS accuracy: Improved prediction of relative binding energies of known ligand-protein complexes, and 3) SILCS database screening: Ability to identify known ligands from database screening using pharmacophore models based on SILCS GFEs. Commercialization of these substantial improvements to the SILCS technology will occur through SilcsBio's existing marketing channels which include partnering with computational chemistry software vendors to make the technology available for licensing for in-house use by pharmaceutical companies and SilcsBio acting as a contract research organization (CRO) performing structure-based ligand design. CRO work includes supplying and interpreting GFE FragMaps to customers, a product that can, importantly, be supplied without customers revealing their IP as the Fragmaps are generated in the absence of ligands. Successful completion of this Phase I proposal is anticipated to lay the foundation for a Phase II SBIR to improve and expand the SILCS platform by 1) extending the technology to occluded target ligand binding sites not accessible to the surrounding aqueous environment, such as those commonly found in many GPCRs, 2) extending SILCS into a product that will be marketable for in-house use by pharmaceutical companies and 3) developing the necessary infrastructure required for SilcsBio to be a successful CRO.