【作者】 楚夏昆；

【导师】 汪劲；

【作者基本信息】 吉林大学 ， 理论物理学， 2014， 博士

【摘要】 生物分子的动力学过程在细胞中是普遍存在的，同时也是最基本的生命活动。生物分子的动力学过程一般都是柔性的，通常伴随有大的构象变化。人们已经认识到柔性或者构象转变在许多生物过程，包括蛋白质折叠，蛋白质结合和蛋白质变构等，中扮演着举足轻重的作用。生物分子通过与其他分子结合来发挥其生物功能，而结合过程中的构象动力学对生物分子的功能实现也是至关重要的。但是，对柔性生物分子动力学全面的物理解释一直以来都是分子生物学中一个巨大的挑战。在本文中，我们通过量化能量地貌拓扑结构的方法将理论和实验联系在了一起，解决了这个难题。1.我们通过探索态密度的方法量化了蛋白质折叠的能量地貌。我们计算了三个能量地貌拓扑结构的量度：能量间隙，能量粗糙度和熵。这三项分别表示了折叠漏斗的倾斜度，漏斗表面的粗糙度和漏斗的体积大小。通过这三项，我们计算了一个无量纲的比值，它描述了能量地貌的拓扑结构。我们发现能量地貌的拓扑结构可以预测出折叠的热力学稳定性相对于陷阱的比值和动力学速率。能量地貌的拓扑形状越像漏斗，那么折叠的热力学稳定性越高，折叠的动力学速率也越快。我们研究了对于不同体积大小的蛋白质和具有相同体积大小但是不同拓扑结构的蛋白质的拓扑阻挫和能量阻挫对蛋白质折叠的影响。我们发现这种能量地貌拓扑结构同折叠热力学和动力学单调相关的现象在上述所有情况中都存在。也就是说，我们证明了蛋白质折叠的能量地貌是决定蛋白质折叠热力学和动力学的关键因素。我们的工作连接了理论和实验。2.利用态密度，我们量化了15个同源二聚体的柔性生物分子识别过程中有效结合和折叠，以及全局的结合-折叠能量地貌的拓扑结构。通过有效的结合和折叠的能量地貌的拓扑结构之间的关系，我们成功地将这15个同源二聚体划分为协同的结合-折叠耦合的两态识别机制和非协同的折叠先于结合的三态识别机制。这个结果与之前的理论和实验保持一致。我们发现非天然相互作用通过调节能量地貌的拓扑结构来调节识别机制。通过量化全局的结合-折叠的能量地貌的拓扑结构，我们发现结合-折叠的全局热力学稳定性与陷阱的比值和能量地貌拓扑结构量，以及全局动力学速率和能量地貌拓扑结构量强烈相关。因此，我们证明了能量地貌拓扑结构决定着柔性识别的热力学和动力学。我们同样发现了识别的动力学与温度之间是“U”型的关系，同时会有一个动力学交叉温度来区分e指数和非e指数的动力学行为。这些结果都与能量地貌的拓扑结构有关。我们的方法可以定量地将理论预测同实验测量联系在一起。3.通过发展一种基于结构的粗粒化的包含用德拜-休克尔模型描述的静电相互作用的模型，我们研究了组蛋白伴侣Chz1与其目标组蛋白H2A.Z-H2B结合的过程。未结合的Chz1是典型的天然无规蛋白质。我们发现Chz1折叠的构象转变只发生在主过渡态之后，并且越过主过渡态之后耦合的结合和折叠行为会通过两条平行的路径进行。我们发现链间的静电相互作用起着“引导力”的作用来加速结合的过程。有趣的是，增加静电相互作用的强度，会导致形成最终复合体的速率变慢。这是因为强的链内静电相互作用使得Chz1坍缩形成一系列紧密的非天然结构。结合会由于这些能量地貌表面上的陷阱而变慢。我们提供了一种静电相互作用控制的柔性分子识别机制。我们的发现导致了一种“局部坍缩或陷阱”和“飞掷”协同在一起的动力学结合机制，我们的结果提供了一种新的理解天然无规蛋白质结合中静电相互作用的方式。更多还原

【Abstract】 Biomolecular dynamics is prevalent and fundamental in cellar activity.Biomolecular dynamics is often flexible and associated with large conformationalchanges. It has been recognized that the flexibility or conformational transition playsa major role in many biomolecular process, including protein folding, protein bindingand protein allostery. Biomolecules realize their function by binding to the partners.The conformational dynamics in binding is also critical for the biomolecular function.However, the physical and global understanding for the flexible biomoleculardynamics is still challenging. In this thesis, we meet this challenge by quantifying theenergy landscape topography and establishing the connections between theoreticalpredications and experiment measurements.1. We quantify the protein folding energy landscapes by exploring the density ofstates. We calculate three energy landscape topography quantities: energy gap, energyroughness and entropy, corresponding to the slope, bumpiness and size of the foldingfunnels, respectively. We show that the dimensionless ratio between gap, roughnessand entropy can accurately describe the energy landscape topography. We find that theenergy landscape topography can predict the folding thermodynamic stability againsttrapping and the kinetic rates. More funneled energy landscapes lead to more stablethermodynamics and faster kinetics. We investigate the role of topological andenergetic roughness for protein of different sizes and for protein of same size, but withdifferent structural topologies. We find that the monotonic correlations between theenergy landscape topography and folding thermodynamics as well as kinetics arepresent in all the cases. In short, we demonstrate that the folding energy landscape isthe underlying factor to determine the folding thermodynamics and kinetics. Thiswork bridges the gap between theories and experiments. 2. Using density of states, we quantify the effective binding and folding, as wellas the whole global binding-folding energy landscape topography in flexiblebiomolecular recognition for15homodimers. Based on the interplay betweentopography of the effective binding and folding energy landscape topography, the15homodimers can be successfully classified into two-state cooperative “coupledbinding-folding” and three-state non-cooperative “folding prior to binding” scenario.The results are consistent with the previous theoretical and experimentalinvestigations. We find that the non-native interactions modulate the associationmechanism through the underlying binding and folding energy landscapes. Byquantifying the whole global binding-folding energy landscapes, we find the strongcorrelation between the landscape topography measure and the thermodynamicstability versus trapping, as well as the kinetic rates. Therefore, we demonstrate thatenergy landscape determines the thermodynamics and kinetics of the flexiblebiomolecular recognition. We also find “U-shape” temperature-dependent kineticbehavior and a dynamical cross-over temperature of dividing exponential andnonexponential kinetics for two-state homodimers. The findings are controlled by thetopography of the underlying energy landscapes. Our results provide the quantitativebridge between the landscape topography and experimental measurements.3. By developing a structure-based coarse-grained model, in whichDebye-Hückel model is implemented for describing the electrostatic interactions, weinvestigate the histone chaperone Chz1binding to its target histone variantH2A.Z-H2B. Free Chz1is a typical Intrinsically Disordered Protein (IDP). We findthat the folding conformational changes in Chz1only happens after the majortransition states and then Chz1undergoes coupled binding-folding through twoparallel pathways. We find that the inter-chain electrostatic interactions serve as“steering forces” to facilitate the association. Interestingly, we find increasing thestrength of electrostatic interactions leads to decreasing rate of formation of the finalcomplex. It is due to that the strong intra-chain electrostatic interactions collapse theChz1into non-native compact structures. Binding is slowed by the escape of the trapson the energy landscapes. Our studies provide an ionic-strength-controlled flexible binding-folding mechanism. The findings lead to a cooperative binding mechanism of“local collapse or trapping” and “fly-casting” together and a new understanding of theroles of electrostatic interactions in IDPs’ binding.更多还原