Sawle Lucas, Ghosh Kingshuk
Department of Physics and Astronomy, University of Denver, Denver, Colorado 80208, USA.
J Chem Phys. 2015 Aug 28;143(8):085101. doi: 10.1063/1.4929391.
A general formalism to compute configurational properties of proteins and other heteropolymers with an arbitrary sequence of charges and non-uniform excluded volume interaction is presented. A variational approach is utilized to predict average distance between any two monomers in the chain. The presented analytical model, for the first time, explicitly incorporates the role of sequence charge distribution to determine relative sizes between two sequences that vary not only in total charge composition but also in charge decoration (even when charge composition is fixed). Furthermore, the formalism is general enough to allow variation in excluded volume interactions between two monomers. Model predictions are benchmarked against the all-atom Monte Carlo studies of Das and Pappu [Proc. Natl. Acad. Sci. U. S. A. 110, 13392 (2013)] for 30 different synthetic sequences of polyampholytes. These sequences possess an equal number of glutamic acid (E) and lysine (K) residues but differ in the patterning within the sequence. Without any fit parameter, the model captures the strong sequence dependence of the simulated values of the radius of gyration with a correlation coefficient of R(2) = 0.9. The model is then applied to real proteins to compare the unfolded state dimensions of 540 orthologous pairs of thermophilic and mesophilic proteins. The excluded volume parameters are assumed similar under denatured conditions, and only electrostatic effects encoded in the sequence are accounted for. With these assumptions, thermophilic proteins are found-with high statistical significance-to have more compact disordered ensemble compared to their mesophilic counterparts. The method presented here, due to its analytical nature, is capable of making such high throughput analysis of multiple proteins and will have broad applications in proteomic studies as well as in other heteropolymeric systems.
提出了一种通用形式，用于计算具有任意电荷序列和非均匀排除体积相互作用的蛋白质及其他杂聚物的构型性质。采用变分方法来预测链中任意两个单体之间的平均距离。所提出的分析模型首次明确纳入了序列电荷分布的作用，以确定不仅在总电荷组成上不同，而且在电荷修饰上也不同（即使电荷组成固定）的两个序列之间的相对大小。此外，该形式体系具有足够的通用性，允许两个单体之间的排除体积相互作用发生变化。针对30种不同的聚两性电解质合成序列，将模型预测结果与Das和Pappu的全原子蒙特卡罗研究结果[《美国国家科学院院刊》110, 13392 (2013)]进行了对比。这些序列具有相等数量的谷氨酸（E）和赖氨酸（K）残基，但序列内的模式不同。在没有任何拟合参数的情况下，该模型以相关系数R(2) = 0.9捕捉到了回转半径模拟值对序列的强烈依赖性。然后将该模型应用于真实蛋白质，以比较540对嗜热和嗜温蛋白质直系同源对的未折叠状态尺寸。假设在变性条件下排除体积参数相似，并且仅考虑序列中编码的静电效应。基于这些假设，发现嗜热蛋白质与其嗜温对应物相比，具有更高统计显著性的更紧凑无序集合。本文提出的方法由于其分析性质，能够对多种蛋白质进行这种高通量分析，并将在蛋白质组学研究以及其他杂聚体系中有广泛应用。
