Computational studies of membrane transport proteins

次级活性转运蛋白是一类膜蛋白，其利用预先存在的分子浓度梯度作为能量来源，用于逆其浓度梯度转运另一种底物，例如营养物或神经递质。这个过程需要蛋白质改变构象，以便在称为交替进入的循环中暴露到膜一侧或另一侧上的底物结合位点的通路。每种生物都表达几十种不同的二级转运蛋白，这些蛋白表现出一套不同的结构，尽管总是具有某种形式的内部结构对称性。在过去的十年中，从三维结构中获得了前所未有的突破性见解。然而，每个膜转运蛋白的机制的详细理解需要在许多构象状态下的结构的知识，包括识别的底物或底物的结合区域。此外，这些结构需要被放置到一个动态的整体背景下的热力学景观分离的动力障碍。在过去的一年里，我们小组的研究继续在许多膜蛋白中研究这些问题。 转运蛋白和其他膜蛋白可以影响其周围膜的形态，这一点越来越被人们所认识。不太了解的是，这些变形的程度和自由能成本可能在蛋白质的不同功能状态之间变化，因此，它们可能对定义其机制有重要贡献。我们认为三聚体钠-天冬氨酸同向转运体GltPh，一个重要的神经元转运体，EAAT类，其机制需要已知的最剧烈的结构变化之一的同系物。与Faraldo-Gomez实验室（NHLBI）合作，我们对GltPh进行了分子模拟，这表明当原聚体向内时，它们会引起深度，长距离，但相互独立的膜变形。使用一种新的模拟方法，我们估计这种膜扰动的自由能成本是相当大的，提出了重要的新问题，膜在神经元谷氨酸摄取的作用，特别是在拥挤的环境中，如视网膜EAAT蛋白（参考文献1）。 膜相互作用对于水溶性蛋白质也是至关重要的，例如，通过与带电脂质的钙依赖性静电相互作用。与沃德实验室（NINDS）合作，我们使用结构建模来预测钙结合位点和膜联蛋白A11（参考文献2）羧基末端结构域中静电势的变化，膜联蛋白A11是一种与神经退行性疾病额颞叶痴呆有关的蛋白质。膜联蛋白A11的氨基末端片段与RNA颗粒相关，其在神经元周围长距离运动的机制仍有待确定。结构模型提供了一个可测试的机制预测的钙依赖性附着与溶酶体通过膜联蛋白A11，一个过程，这将使RNA颗粒搭便车到当地的转录位点。 对于许多转运蛋白和其他膜蛋白，如通道，近年来已经显示出前所未有的结构数据量，部分原因是使用冷冻电子显微镜的研究。一个特别令人感兴趣的情况是大而多样的瞬时受体电位（TRP）家族的离子通道，其中超过100个结构已在短短几年内报道。这些过多的数据需要一种系统的方法来分析共同的特征，如途径和结合位点。我们以前开发的结构比对程序，允许大量的膜蛋白结构的比较。在与Swartz实验室（NINDS）的合作中，我们将这些程序适用于TRP通道结构的特定情况，从而能够进行全面，系统的调查，导致多个可验证的假设（参考文献3），并为系统分析其他膜蛋白家族奠定了基础，这些家族具有大量结构。 总之，我们今年的出版物反映了与实验室密切合作利用计算方法的持续努力，并推动了对神经元过程中生物医学重要蛋白质机制的理解，包括转运蛋白，通道和其他膜相关蛋白。

Secondary active transporters are a class of membrane proteins that utilize pre-existing molecular concentration gradients as an energy source for translocating another substrate, such as a nutrient or a neurotransmitter, against its concentration gradient. This process requires the protein to change conformations so as to expose a pathway to the substrate binding site(s) on one or other side of the membrane, in a cycle known as alternating access. Every organism expresses dozens of different secondary transporter proteins, and these exhibit a diverse set of architectures, albeit always with some form of internal structural symmetry. Unprecedented, ground-breaking insights have been garnered from three-dimensional structures obtained in the last decade. Nevertheless, a detailed understanding of the mechanism of each membrane transport protein requires knowledge of its structure in many more conformational states, including identification of the binding regions for the substrate or substrates. Moreover, those structures need to be placed into a context of dynamic ensembles on a thermodynamic landscape separated by kinetic barriers. Studies from our group over the last year have continued to investigate these issues in many membrane proteins. That transporters and other membrane proteins can influence the morphology of their surrounding membrane is increasingly recognized. Less appreciated is that the extent and free-energy cost of these deformations likely varies among different functional states of a protein, and thus, that they might contribute significantly to defining its mechanism. We considered the trimeric sodium-aspartate symporter GltPh, a homolog of an important class of neuronal transporters, the EAATs, whose mechanism entails one of the most drastic structural changes known. In collaboration with the Faraldo-Gomez lab (NHLBI), we carried out molecular simulations of GltPh which indicated that when the protomers become inward-facing, they cause deep, long-ranged, and yet mutually-independent membrane deformations. Using a novel simulation methodology, we estimated that the free-energy cost of this membrane perturbation is substantial, raising important new questions about the role of the membrane in neuronal glutamate uptake, especially in crowded environments such as that experienced by retinal EAAT proteins (Ref. 1). Membrane interactions are also critical for water-soluble proteins, for example, through calcium-dependent electrostatic interactions with charged lipids. In collaboration with the Ward lab (NINDS), we used structural modeling to predict calcium binding sites and resultant changes in electrostatic potential in the carboxy-terminal domain of annexin A11 (Ref. 2), a protein implicated in the neurodegenerative disease Frontotemporal Dementia. The amino-terminal segment of annexin A11, is associated with RNA granules, whose mechanism of long-distance movement around neurons remained to be identified. The structural models provided a testable mechanistic prediction of calcium-dependent attachment with lysosomes through annexin A11, a process that would enable to RNA granules to hitchhike to local sites of transcription. For many transporters and other membrane proteins such as channels, recent years have shown an unprecedented amount of structural data, in part due to studies using cryo-electron microscopy. One case of particular interest is the large and diverse transient receptor potential (TRP) family of ion channels, of which over one hundred structures have been reported in just a few years. This plethora of data requires a systematic approach to enable analysis of common features such as pathways and binding sites. We have previously developed structure alignment procedures that allow comparison of large numbers of membrane protein structures. In collaboration with the Swartz lab (NINDS), we adapted these procedures for the specific case of TRP channel structures, enabling a comprehensive, systematic survey that led to multiple, testable hypotheses (Ref. 3), and laid the groundwork for systematic analyses of other membrane protein families for which large numbers of structures become available. In summary, our publications this year reflect ongoing efforts to utilize computational approaches in close collaboration with experimental laboratories, and drive understanding of the mechanism of biomedically-important proteins in neuronal processes, including transporters, channels and other membrane-associated proteins.