ABI innovation: Computational method for exploring the mysteries of cell-penetrating peptides

该项目旨在开发一种新的计算方法和网络服务器，用于预测自然产生的和合成的多肽的细胞穿透能力，并分析它们与细胞膜的相互作用。细胞穿透肽(CPPs)是二十年前发现的，从那时起就被用于将各种大分子输送到细胞和组织中。然而，多肽通过细胞膜的分子机制仍然是一个谜。为了阐明CPP直接跨膜的机理，将发展一种创新的理论方法。这种方法将量化多肽进入细胞的过程的能量学，包括多肽与膜的结合，它们折叠成α-螺旋和β-片状结构，以及通过扩散或通过诱导膜变形(如变薄、弯曲或孔形成)穿过脂质双层。该项目预计将产生重大的科学影响，为发现和设计新的国家方案提供独特的方法和第一个公共网络服务器。这种方法对于分析具有潜在有益或有害生理效应的更广泛的生物活性多肽和蛋白质与膜的相互作用也是有价值的。这种通用工具的开发和应用将促进膜多肽和蛋白质领域的科学知识的发展。一个易于使用的网络工具将使从事生物物理学、分子和细胞生物学、生物纳米技术和药物设计领域的广大学生、教师和研究人员受益。该项目包括密歇根大学的学生参与计算资源的开发，以及使用视频、图像和膜中多肽和蛋白质的有形分子模型教授来自社会经济弱势群体的K-12学生。新方法将结合多肽形成二级结构的热力学模型和脂双层的各向异性溶剂模型，该模型考虑了不同生物膜中特定的极性分布、曲率、弹性模数和跨膜电位。所建议的方法将在公共网络服务器上进行广泛的测试和实施，从而能够：(1)量化膜-多肽的相互作用、能量和结合模式，包括多肽对脂质双层的变形(PPM 3.0)；(2)对膜结合的β-折叠和α-螺旋进行建模，包括对其全原子三维结构的产生和优化(FMAP 2.0)；以及(3)通过计算和评估它们在脂质双层中的可能转运机制、途径和能量障碍(CPPpred)来预测潜在的CPPs。CPPpred服务器将生成多肽的全原子三维结构，估计它们的膜结合模式和转位途径上的能量屏障，并以它们的氨基酸序列和特定膜的参数为输入来评价多肽的转位机制。该方法的高计算效率将使其应用于对多肽数据库的大规模计算筛选，以识别潜在的CPP。

This project aims to develop a new computational method and a web server for prediction of the cell-penetrating ability of naturally-occurring and synthetic peptides and analysis of their interactions with cellular membranes. Cell-penetrating peptides (CPPs) were discovered two decades ago and since then were used for delivery of various macromolecules into cells and tissues. However, the molecular mechanisms of passage of peptides through cell membranes still remain a mystery. To shed light on the mechanisms of direct membrane crossing by CPPs, an innovative theoretical method will be developed. This method will quantify the energetics of the processes underlying peptide entry into cells, including binding of peptides to membranes, their folding into alpha-helical and beta-sheet structures, and crossing the lipid bilayer by diffusion or by inducing membrane deformations, such as thinning, curving, or pore formation. The project is expected to have a significant scientific impact by providing unique methodology and the first public web server for discovery and design of novel CPPs. This method will be also valuable for analysis of interactions with membranes of a much wider spectrum of biologically active peptides and proteins with potentially beneficial or harmful physiological effects. The development and application of such general tool will advance scientific knowledge in the field of membrane peptides and proteins. An easy-to-use web tool will benefit a broad community of students, teachers and researchers engaged in the fields of biophysics, molecular and cell biology, bio-nanotechnology, and drug design. The project includes participation of students from the University of Michigan in development of the computational resources, as well as teaching K-12 students from socio-economically disadvantaged groups using videos, images, and tangible molecular models of peptides and proteins in membranes.The new method will combine a thermodynamic model of secondary structure formation by peptides and an anisotropic solvent model of the lipid bilayer that accounts for specific polarity profiles, curvature, elastic moduli, and transmembrane potential in different biological membranes. The proposed methodology will be extensively tested and implemented on a public web server that will allow the following: (1) to quantify membrane-peptide interactions, energy and binding modes, including deformations of the lipid bilayer by peptides (PPM 3.0); (2) to perform modeling of membrane-bound beta-sheets and alpha-helices, including generation and optimization of their all-atom three-dimensional structures (FMAP 2.0); and (3) to predict potential CPPs by calculating and evaluating their possible translocation mechanisms, pathways and energy barriers in the lipid bilayer (CPPpred). The CPPpred server will generate all-atom three-dimensional structures of peptides, estimate their membrane binding modes and energy barriers along the translocation pathway, and evaluate mechanisms of peptide translocation using their amino acid sequence and parameters of specific membranes as input. The high computational efficiency of the method will allow its application for large-scale computational screening of peptide databases to identify potential CPPs.