Understanding protein folding, evolution and function via molecular simulation

通过分子模拟了解蛋白质折叠、进化和功能

• 批准号: 10011312
• 负责人: Robert Best
• 金额: $ 58.73万
• 依托单位: NATIONAL INSTITUTE OF DIABETES AND DIGESTIVE AND KIDNEY DISEASES
• 依托单位国家: 美国
• 资助国家: 美国
• 起止时间: 至
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10011312
• 关键词: Acceleration; Address; Affect; Agreement; Algorithmic Analysis; Area; Bacterial Proteins; Binding; Capsid Proteins; Charge; Collaborations; Complex; Cytoplasmic Granules; DNA; Data; Development; Diffusion; Disease; Dissociation; Electrostatics; Equilibrium; Evolution; Fluorescence; Fluorescence Resonance Energy Transfer; Free Energy; G-substrate; GYPA gene; Goals; Grain; Histone H1; Hydrogen Bonding; In Vitro; Ions; Kinetics; Length; Liquid substance; Manuscripts; Measures; Membrane Proteins; Methodology; Methods; Modeling; Molecular; Molecular Chaperones; Nucleic Acid Folding; Nucleic Acids; Nucleosomes; Peptides; Phase; Photons; Preparation; Property; Proteins; Publications; Publishing; RNA Folding; Reproduction; Resolution; Ribonuclease H; Ribosomes; Role; SH3 Domains; Series; Solvents; Structure; Surface; Testing; Time; Universities; Viral; Work; amyloid fibril formation; base; connectin; design; experimental study; improved; interest; member; models and simulation; mutant; next generation; novel; phenomenological models; photon-counting detector; protein aggregation; protein folding; protein function; protein misfolding; prothymosin alpha; prototype; screening; simulation; single molecule; theories; tool

The project has addressed the following areas in the past year: 1. Association of highly charged intrinsically disordered proteins. Following up on work done last year characterizing the equilibrium association of the complex between histone H1 and the protein prothymosin alpha, we have now addressed the kinetics of association and dissociation (collaboration with Ben Schuler (University of Zurich) and Birthe Kragelund (University of Copenhagen)). The prothymosin appears to be exchanged by an associative mechanism whereby association of a second prothymosin molecule displaces the first. This finding explains the relatively rapid dissociation kinetics (microsecond time scale) compare to what would be expected based on the diffusion-limited association and picomolar binding. Coarse grained models have been used to describe the exchange mechanism and to calculate a free energy surface for binding. This work is currently being prepared for publication. (R. Best) 2. Role of prothymosin as a chaperone for histone H1 on the nucleosome. Related to (1) above, we have been investigating the ability of prothymosin to facilitate dissociation of histone H1 from the nucleosome. Coarse-grained models have been used to describe the bound ensemble of H1 to the nucleosome, reproducing the experimental FRET efficiencies. Using a novel methodology, we have also computed the effect of prothymosin on the H1 dissociation rates, finding an acceleration of around 2 orders of magnitude, in excellent agreement with experimental estimates (R. Best, D. Mercadante). Publication in preparation. (R. Best) 3. Role of viral capsid protein as nucleic acid folding chaperone (collaboration with B. Schuler). We have developed a very basic coarse-grained model for nucleic acids and protein-nucleic acid interactions in order to examine the effect of a disordered protein on the folding of a model DNA hairpin. The simulations showed that the electrostatic screening effect of the protein, resulting in a collapse of the nucleic acid was sufficient to explain all of the acceleration in observed folding rate (R. Best). (1) 4. All-atom and coarse-grained force field development. 4.1. Improving hydrogen bond potential in all-atom models of RNA folding (collaboration with P. Kuhrova). Our earlier work had shown that the hydrogen bond potential was a possible weak point in all-atom protein force fields and may be responsible for their inability to distinguish folded from misfolded structures. We have now developed a new hydrogen bond potential with improved performanc for RNA folding. (2) 4.2. Local vs global effects on all-atom protein force fields. We have recently shown that certain experimental observables are almost exclusively sensitive to local structural properties of proteins, while others capture more global properties. Both types of observables therefore need to be considered when assessing protein force fields (Collaboration with J. Mittal) (3). 4.3. Improving the coarse-grained Martini force field (collaboration with J. Domanski, P. Telles de Souza). We had earlier shown that the current version of the Martini coarse-grained model could not capture the native state of glycophorin (a prototype for transmembrane helix association) as the most stable bound state. In collaboration with the Martini developers we are testing the next generation of Martini force field, which shows a much improved reproduction of the native structure. (R. Best) 5. Multidomain protein misfolding. Following on earlier work in which we investigated multidomain protein misfolding using coarse-grained models, we have now developed an easy to use tool (TADOSS) to predict misfolding propensity based on a phenomenological model in that work (Collaboration with A. Bateman) (4,5). We are also currently collaborating with Alex Bateman (Cambridge) and Jennifer Potts (York) on how certain multidomain bacterial proteins are able to avoid misfolding despite the adjacent domains having high sequence similarity. (P. Tian, R. Best) 6. Co-translational protein folding. Last year, we developed a coarse-grained model of the ribosome suitable for studying co-translational folding, and applied it to the folding of the I27 domain of titin (6), the src SH3 domain (7) and several other proteins (8). This year, in collaboration with Gunnar von Heijne, we used the model to investigate more directly the relationship between the forces arising from the folding nascent chain and the yield of full length protein obtained in arrest peptide experiments. We also devised a method for obtaining these forces directly from experiment by using a series of different arrest peptides with the same protein constructs (Manuscript under revision in PNAS). We are also collaborating with Susan Marqusee on the co-translational folding of RNase H, and the effect of mutants which make the folding either more or less cooperative. In another collaboration with the group of Sander Tans, we are using our coarse-grained co-translational folding model to understand the effects of the ribosome on the folding and unfolding rates of ADR1a in the ribosome exit tunnel, as probed by single molecule force spectrosccopy (R. Best, P. Tian). 7. Environmental effects on phase separation of intrinsically disordered proteins. We are working to further develop our coarse-grained model for liquid-liquid phase separation (LLPS) of intrinsically disordered proteins published last year, to include such effects as the addition of ions (Hofmeister effects) and co-solvents, which are known to affect the phase diagram for LLPS in vitro. (T. Dannenhofer-Lafage) 8. Modelling sensitivity of single molecule experiments to protein folding transition paths using molecular simulations. Recent single molecule fluorescence experiments have been able to detect transition paths between folded and unfolded states of proteins by combining photon by photon detection with sophisticated maximum likelihood analysis algorithms. However, it is not clear how the inferred transition path durations relate to the actual folding transition path lengths, since they cannot be independently measured. We have used simulations as a model to generate coarse-grained folding trajectories for two proteins (alpha3D, protein G), in which we can unambiguously assign transition paths. We then generated synthetic photon trajectories from these simulations and analyzed them in the same way as the experimental data. We found that the experimentally inferred transition path durations are of the right magnitude, but systematically shorter than the true durations (G. Taumoefolau). Group members involved in each project are listed at the end of each section.

该项目在过去一年中解决了以下领域的问题： 1. 高电荷本质无序蛋白质的关联。继去年表征组蛋白 H1 和蛋白质原胸腺肽 α 之间复合物平衡关联的工作之后，我们现在研究了关联和解离的动力学（与 Ben Schuler（苏黎世大学）和 Birthe Kragelund（哥本哈根大学）合作）。胸腺肽原似乎是通过缔合机制进行交换的，其中第二个胸腺肽原分子的缔合取代了第一个分子。这一发现解释了与基于扩散限制缔合和皮摩尔结合的预期相比相对较快的解离动力学（微秒时间尺度）。粗粒度模型已用于描述交换机制并计算用于结合的自由能表面。这项工作目前正在准备出版。 （R.贝斯特） 2. 胸腺肽原作为核小体上组蛋白 H1 的伴侣的作用。与上述 (1) 相关，我们一直在研究胸腺肽原促进组蛋白 H1 从核小体解离的能力。粗粒度模型已用于描述 H1 与核小体的结合整体，再现实验 FRET 效率。使用一种新颖的方法，我们还计算了胸腺肽原对 H1 解离速率的影响，发现加速了大约 2 个数量级，与实验估计非常一致（R. Best、D. Mercadante）。出版准备中。 （R.贝斯特） 3. 病毒衣壳蛋白作为核酸折叠伴侣的作用（与 B. Schuler 合作）。我们开发了一个非常基本的核酸和蛋白质-核酸相互作用的粗粒度模型，以便检查无序蛋白质对模型 DNA 发夹折叠的影响。模拟表明，蛋白质的静电屏蔽效应导致核酸崩溃，足以解释观察到的折叠速率的所有加速（R. Best）。 (1) 4.全原子、粗粒度力场发展。 4.1.提高 RNA 折叠全原子模型中的氢键潜力（与 P. Kuhrova 合作）。我们早期的工作表明，氢键势是全原子蛋白质力场中可能的弱点，并且可能是它们无法区分折叠结构和错误折叠结构的原因。我们现在开发了一种新的氢键势，具有改进的 RNA 折叠性能。 (2) 4.2.对全原子蛋白质力场的局部与全局影响。我们最近表明，某些实验观测值几乎完全对蛋白质的局部结构特性敏感，而其他实验则捕获更多的全局特性。因此，在评估蛋白质力场时需要考虑这两种类型的可观测值（与 J. Mittal 合作）(3)。 4.3.改进粗粒度 Martini 力场（与 J. Domanski、P. Telles de Souza 合作）。我们之前已经表明，当前版本的 Martini 粗粒度模型无法捕获血型糖蛋白（跨膜螺旋缔合的原型）的天然状态作为最稳定的结合态。我们正在与 Martini 开发人员合作测试下一代 Martini 力场，该力场显示出对原生结构的再现有了很大的改进。 （R.贝斯特） 5.多域蛋​​白错误折叠。继我们使用粗粒度模型研究多域蛋白质错误折叠的早期工作之后，我们现在开发了一种易于使用的工具 (TADOSS)，以根据该工作中的现象学模型预测错误折叠倾向（与 A. Bateman 合作）(4,5)。我们目前还与 Alex Bateman（剑桥）和 Jennifer Potts（约克）合作研究某些多结构域细菌蛋白如何能够避免错误折叠，尽管相邻结构域具有高度的序列相似性。 （P. 田，R. 贝斯特） 6. 共翻译蛋白质折叠。去年，我们开发了一种适合研究共翻译折叠的核糖体粗粒度模型，并将其应用于肌联蛋白 I27 结构域 (6)、src SH3 结构域 (7) 和其他几种蛋白质 (8) 的折叠。今年，我们与 Gunnar von Heijne 合作，使用该模型更直接地研究折叠新生链产生的力与捕获肽实验中获得的全长蛋白质产量之间的关系。我们还设计了一种方法，通过使用一系列具有相同蛋白质结构的不同阻滞肽，直接从实验中获得这些力（手稿正在 PNAS 修订中）。我们还与 Susan Marqusee 合作研究 RNase H 的共翻译折叠，以及使折叠或多或少合作的突变体的影响。在与 Sander Tans 团队的另一项合作中，我们使用粗粒度共翻译折叠模型来了解核糖体对核糖体出口隧道中 ADR1a 折叠和解折叠速率的影响，通过单分子力光谱探测（R. Best，P. Tian）。 7.环境对本质无序蛋白质的相分离的影响。我们正在努力进一步开发去年发表的本质无序蛋白质的液-液相分离 (LLPS) 粗粒度模型，其中包括添加离子（霍夫迈斯特效应）和共溶剂等效应，已知这些效应会影响体外 LLPS 的相图。 （T.丹尼霍夫-拉法奇） 8. 使用分子模拟模拟单分子实验对蛋白质折叠转变路径的敏感性。最近的单分子荧光实验已经能够通过将逐个光子检测与复杂的最大似然分析算法相结合来检测蛋白质折叠和未折叠状态之间的转变路径。然而，尚不清楚推断的过渡路径持续时间与实际折叠过渡路径长度如何相关，因为它们无法独立测量。我们使用模拟作为模型来生成两种蛋白质（alpha3D、蛋白质 G）的粗粒度折叠轨迹，在其中我们可以明确指定转换路径。然后，我们从这些模拟中生成合成光子轨迹，并以与实验数据相同的方式对其进行分析。我们发现，实验推断的过渡路径持续时间大小正确，但系统地短于真实持续时间（G. Taumoefolau）。 每个项目涉及的小组成员都列在每个部分的末尾。