STOCHASTIC BIOCHEMICAL SIMULATION OF ACTIN-BASED FIBROBLAST CELL SPREADING

这个子项目是许多研究子项目中利用 资源由NIH/NCRR资助的中心拨款提供。子项目和 调查员(PI)可能从NIH的另一个来源获得了主要资金， 并因此可以在其他清晰的条目中表示。列出的机构是 该中心不一定是调查人员的机构。 细胞运动在许多生理过程中起着重要的作用。基于肌动蛋白的细胞运动是驱动细胞运动、细胞迁移和细胞形态响应细胞外和细胞内信号的机制之一。我们的合作者，哥伦比亚大学的Michael Sheetz教授和他的同事，最近开发了定量成像方法来测量哺乳动物成纤维细胞在接触纤维连接蛋白等细胞外基质蛋白时的扩散情况(Dubin-Thaler等BiPhys)。J.2004，86：1794-1806)。我们开发了一种计算方法来模拟这一过程，以了解细胞信号如何通过基于肌动蛋白细胞骨架的运动机制调节细胞传播。这个项目的目标是了解描述肌动蛋白细丝网络受调控的组装和拆解的信号调控和基于肌动蛋白的生化反应是否可以用来获得细胞扩散的定量模型。我们使用随机方法来模拟肌动蛋白细胞骨架的动态生长，它推动细胞膜前进，导致细胞扩散。我们的计算机模拟程序的主要算法是基于吉莱斯皮的生化反应随机模拟的第一种方法。基于肌动蛋白的细胞铺展过程中涉及的生化反应包括肌动蛋白细丝聚合、肌动蛋白细丝分支和肌动蛋白细丝封顶。详细模拟了基于这些生化反应的肌动蛋白细胞骨架的生长以及由细胞骨架生长引起的细胞膜的运动。在蒙特卡罗模拟的每个循环中，每个肌动蛋白细丝都要计算三个基于肌动蛋白的生化反应的速率，并对基于所有反应速率的概率分布进行采样，以确定下一步发生哪个反应。在此模拟过程中，随着细胞的不断扩散，会产生越来越多的肌动蛋白细丝，根据输入的分子浓度，肌动蛋白细丝的数量可以达到5万条以上。因此，基于肌动蛋白的细胞扩散的模拟需要大量的计算能力才能达到具有生物学意义的细胞扩散持续时间，通常约为2-3分钟。图1说明了模拟程序使用的随机算法。仿真程序使用C语言开发，由GCC 3.4编译，并在Intel Pentium4和Intel Itanium2平台上的Redhat Enterprise Linux4上进行了测试。基于此模拟的性质，我们需要在TeraGrid SDSC IA64 Linux集群上申请开发(快速)分配。由于开发分配允许多达30,000个服务单元，因此我们计划使用48个节点(96个Itanium2处理器)13天来运行我们的模拟。我们可以使用SecureShell连接直接访问TeraGrid，以传输源文件和数据文件并执行命令行工作。

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Cell motility plays important roles in many physiological processes. Actin-based cell motility is one of the mechanisms that drive cell movement, cell migration and cell morphology in response to extracellular and intracellular signals. Our collaborators, Professor Michael Sheetz and his colleagues at Columbia University, have recently developed quantitative imaging approaches to measure the spreading of mammalian fibroblast cells when they contact extracellular matrix proteins such as fibronectin (Dubin-Thaler, et al Biophys. J. 2004, 86: 1794-1806). We have developed a computational method to model this process in order to understand how cellular signals regulate cell spreading through the motility machinery based on actin cytoskeleton. The goal of this project is to understand whether the signal-regulated and actin-based biochemical reactions that describe the regulated assembly and disassembly of actin filament networks can be used to obtain the quantitative model of cell spreading. We use a stochastic approach to simulate the dynamic growth of actin cytoskeleton which pushes cell membrane forward and results in cell spreading. The main algorithm of our computer simulation program is based on Gillespie's First Method of stochastic simulation of biochemical reactions. The biochemical reactions involved in the simulation of actin-based cell spreading process include actin filament polymerization, actin filament branching and actin filament capping. The growth of actin cytoskeleton based on these biochemical reactions and the movement of cell membrane caused by cytoskeleton growth are simulated in details. Within each loop of Monte Carlo simulation, every actin filament is subject to the calculation of the rates of three actin-based biochemical reactions and the probability distribution based on all reaction rates is sampled to determine which reaction to occur at next step. During this simulation, more and more actin filaments are created as cell keeps spreading, and based on the input molecular concentrations, the number of actin filaments can reach more than 50,000. Therefore the simulation of actin-based cell spreading needs significant amount of computing power to be able to reach biologically meaningful time duration of cell spreading, usually about 2 to 3 minutes. Figure 1 illustrates the stochastic algorithm used by the simulation program. The simulation program has been developed using C++ program language, compiled by GCC 3.4, and tested in Redhat Enterprise Linux 4 in both Intel Pentium4 platform and Intel Itanium2 platform. Based on the nature of this simulation, we need to apply for a Development (Expedited) Allocation on TeraGrid SDSC IA64 Linux Cluster. Since Development Allocation allows up to 30,000 Service Units, We plan to use 48 nodes (96 Itanium2 processors) for 13 days to run our simulation. We can access the TeraGrid using SecureShell connection directly to transfer source and data files and perform command line work.