RIKEN Center for Computational Science, 7-1-26, Minatojima-minami-machi, Chuo-ku, Kobe, Hyogo 650-0047, Japan.
Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan.
J Phys Chem B. 2024 Sep 12;128(36):8641-8650. doi: 10.1021/acs.jpcb.4c03475. Epub 2024 Aug 28.
Within the protein interior, where we observe various types of interactions, nonuniform flow of thermal energy occurs along the polypeptide chain and through nonbonded native contacts, leading to inhomogeneous transport efficiencies from one site to another. The folded native protein serves not merely as thermal transfer medium but, more importantly, as sophisticated molecular nanomachines in cells. Therefore, we are particularly interested in what sort of "communication" is mediated through native contacts in the folded proteins and how such features are quantitatively depicted in terms of local transport coefficients of heat currents. To address the issue, we introduced a concept of inter-residue thermal conductivity and characterized the nonuniform thermal transport properties of a small globular protein, HP36, using equilibrium molecular dynamics simulation and the Green-Kubo formula. We observed that the thermal transport of the protein was dominated by that along the polypeptide chain, while the local thermal conductivity of nonbonded native contacts decreased in the order of H-bonding > π-stacking > electrostatic > hydrophobic contacts. Furthermore, we applied machine learning techniques to analyze the molecular mechanism of protein thermal transport. As a result, the contact distance, variance in contact distance, and H-bonding occurrence probability during MD simulations are found to be the top three important determinants for predicting local thermal transport coefficients.
在蛋白质内部,我们观察到各种类型的相互作用,热能沿着多肽链和非键合的天然接触非均匀流动,导致从一个位置到另一个位置的非均匀传输效率。折叠的天然蛋白质不仅作为热传递介质,而且更重要的是作为细胞中复杂的分子纳米机器。因此,我们特别感兴趣的是通过折叠蛋白质中的天然接触来介导什么样的“通信”,以及如何通过热流的局部传输系数来定量描述这些特征。为了解决这个问题,我们引入了一个概念,即残基间热导率,并使用平衡分子动力学模拟和格林-库伯公式来表征小球蛋白 HP36 的非均匀热传输特性。我们观察到,蛋白质的热传输主要是沿着多肽链进行的,而非键合天然接触的局部热导率按氢键>π-堆积>静电>疏水接触的顺序降低。此外,我们应用机器学习技术来分析蛋白质热传输的分子机制。结果表明,在 MD 模拟过程中,接触距离、接触距离的方差和氢键的出现概率是预测局部热传输系数的三个最重要的决定因素。