A metabolic code for cell signaling and polypharmacology.

 描述(由申请人提供)：药物通常与多个靶点相互作用(多元药理学)，不仅解释它们的副作用，而且解释它们的疗效。这样做的目的是 建议从分子进化的角度解释多重药理学的机制，并利用这一见解来推断新的信号网络和设计新的药物先导。 一个激动人心的想法是，生物信号网络已经进化成使用基本上固定的内源性信号分子(5-羟色胺、乙酰胆碱、雌激素等)的小词汇表。这导致蛋白质具有一系列退化的小分子结合位点，药物通过多种药理学“发现”这些结合位点。 为了读取这一代谢密码，我们开发了一种稳健的方法(SEA)来测量两种蛋白质何时拥有相似的配体。使用它，我们已经证明了近500个非GPCR的合成配体与150个GPCR的合成配体相似，并且我们已经证明我们可以准确地预测现有药物的新的副作用。 我认为，1)蛋白质-代谢相互作用解释了配基相似性网络的成功，2配基相似性关联在很大程度上不能用其他生物信息学网络来解释，3)配基相似性可以用来预测可以被单个合成配体激活的靶点集合。 为了测试，我将1)使用配基相似性网络来预测和实验测试一组新的序列无关的靶标与常见的内源信号代谢产物相互作用。我将2)量化配基相似网络与序列相似网络、共表达网络和蛋白质-蛋白质相互作用网络之间的功能互补和交集。我将通过实验测试合成化合物和内源性代谢物通过共表达优先激活配体相似的靶点，并共同注释生物功能、表型和疾病。

 DESCRIPTION (provided by applicant): Drugs typically interact with multiple targets (polypharmacology), explaining not only their side effects but also their efficacy. The aim of this proposal is to explain the mechanism for polypharmacology in terms of molecular evolution and exploit this insight to infer new signaling networks and design new drug leads. A motivating idea is that biological signaling networks have evolved to use a small vocabulary of essentially fixed, endogenous signaling molecules (serotonin, acetylcholine, estrogen, etc.). This causes proteins to have a degenerate repertoire of small molecule binding sites, which drugs 'discover' through polypharmacology. To read this metabolic code, we have developed a robust method (SEA) to measure when two proteins share similar ligands. Using it, we have shown that synthetic ligands of close to 500 non-GPCRs resemble those of 150 GPCRs and we have shown that we can accurately predict novel side-effect of existing drugs. I argue that 1) protein-metabolic interactions explain the success of ligand similarity networks, 2 ligand similarity associations largely cannot be explained by other bioinformatics networks, and 3) ligand similarity can be used to predict sets of targets that can be activated by a single synthetic ligand. To test, I will 1) use ligand similarity networks to predict and experimentally test that novel set of sequence un-related targets interact with common endogenous signaling metabolites. I will 2) quantify the functional complementarity and intersection between ligand- similarity networks and sequence-similarity, co-expression, and protein-protein interaction networks. I will 3) experimentally test that synthetic compounds and endogenous metabolites jointly activate ligand similar targets prioritizing by co-expression, and co-annotated for biological function, phenotype, and disease.