Roles of host factor protein subnetworks in regulating steps of filovirus infection

RP 03项目摘要/摘要 丝状病毒严重依赖于细胞蛋白来促进复制，并且容易受到细胞因子的抑制。 抗病毒系统研究项目3（RP 03）测试了宿主蛋白复合物在病毒复制中的作用， 建立了一种方法来评估大型数据集在病毒感染中的生物重要性。这项工作建立 我们的全基因组siRNA，CRISPR-Cas9，BioID，酵母双杂交（Y2 H），共亲和纯化加质量 光谱学筛选数据已经确定了在EBOV和MARV中发挥作用的100多个宿主因子 复制的传统上，对这种大的命中集的后续工作一直很缓慢，受到足够的免疫缺陷检测的阻碍。 吞吐量和信息读出以提供所需的优先级。此外，命中倾向于在 孤立，不考虑功能或细胞关联的每次击中的相关性。在这里，我们应用一个先进的 计算机算法，Prize Collecting Steiner Forest（PCSF）算法，通过已知的， 高置信度、已发表的蛋白质-蛋白质相互作用（PPI）网络。重叠病毒蛋白相互作用 在Y2 H和蛋白质组学工作中检测到的发现揭示了与共同病毒蛋白相互作用的宿主蛋白质簇。 此外，使用来自不同的基因本体（GO）术语覆盖最高概率蛋白质函数， 数据库，揭示集群的宿主蛋白质相关的可能细胞功能，与肌动蛋白调控和RNA 加工是最多的代表，但也与蛋白质修饰有关的途径， 泛素化、类小泛素化或磷酸化是明显的（如在RP 01和RP 02中所见）。基于这些 新的发现，我们提出的假设，主机因素驻留在子网络相关的共同病毒 蛋白质、功能或两者在特定的病毒复制步骤中起相同的作用。在这里，我们测试这个假设， 应用一种新的、统计学上高功率的光学合并筛选平台， 通过测量病毒蛋白质和病毒RNA表达水平以及亚细胞水平， 将病毒功能相关性与已知宿主PPI的相关性相关联的染色模式。这种方法 允许通过鉴定受影响步骤的机械测定来有效地对用于评估的因素组进行优先排序 在病毒感染中，然后确定负责感染结果的宿主和病毒蛋白的区域。 这项工作从代表肌动蛋白和RNA加工网络的高优先级验证线索开始， 评估我们现有的高严格性网络，然后通过新的数据进行扩展和丰富， RP 01和RP 02。我们的团队由在执行和分析大型主机方面具有良好记录的专家组成 因子遗传和蛋白质组学筛选，并进行病毒机制分析。通过广泛的互动 在每一组之间，我们期望获得对已鉴定的宿主因子蛋白的作用的机理性了解 这些网络将为设计针对这些人的干预措施提供基于知识的选择 EBOV和MARV的相互作用。

RP03 Project Summary/Abstract Filoviruses critically depend on cellular proteins to facilitate replication and are susceptible to inhibition by cellular antiviral systems. Research Project 3 (RP03) tests the roles of host protein complexes in virus replication and establishes an approach to evaluate large datasets for biological importance in virus infection. The work builds on our genome-wide siRNA, CRISPR-Cas9, BioID, yeast two-hybrid (Y2H), and co-affinity purification plus mass spectrometry screening data that has identified 100s of host factors that play roles in EBOV and MARV replication. Traditionally, follow up on such large hit sets has been slow, hampered by assays of sufficient throughput and informative read-out to provide needed prioritization. Additionally, hits tend to addressed in isolation, disregarding the relatedness of each hit by function or cellular association. Here, we apply an advanced computer algorithm, the Prize Collecting Steiner Forest (PCSF) algorithm to associate hit proteins by known, high confidence, published, protein-protein interaction (PPI) networks. Overlaying virus protein interactions detected in Y2H and proteomics work revealed clusters of host proteins that interact with a common virus protein. Furthermore, overlaying highest probability protein function using Gene Ontology (GO) terms from different databases, revealed clusters of host proteins related by likely cellular function, with actin regulation and RNA processing being the most over-represented but also pathways related to protein modifications through ubiquitinylation, sumoylation or phosphorylation being evident (as seen in RP01 and RP02). Based on these novel findings, we propose the hypothesis that the host factors residing in subnetworks related by common virus proteins, function or both play the same role in a specific virus replication step. Here, we test this hypothesis by applying a novel, statistically high powered optical pooled screening platform that phenotypically evaluates infection outcome by measuring both virus protein and virus RNA expression levels together with subcellular staining patterns to associate viral functional relatedness to relatedness by known host PPI. This approach allows efficient prioritization of groups of factors for evaluation by mechanistic assays that identify affected steps in virus infection and then defining regions of the host and virus proteins responsible for the infection outcome. The work starts with high priority validated leads representing the actin and RNA processing networks, then evaluates our existing high stringency network, which is then expanded and enriched through new data fed from RP01 and RP02. Our team consists of experts with strong track records in performing and analyzing large host factor genetic and proteomic screens, and performing virus mechanistic analysis. Through extensive interaction between each group, we expect to gain mechanistic insight into roles for identified host factor protein subnetworks, which will provide knowledge-based choices for design of interventions targeting these interactions, for both EBOV and MARV.