Combining chemical and computational tools for predictive models of microbiome communities

摘要 肠道微生物组对健康和疾病有巨大的影响，积极促进肥胖，糖尿病， 炎症性肠病、心血管疾病和几种知之甚少的神经系统疾病。 我们还没有必要的工具来精确地探测这些微生物群落，尽管这些工具 可以为人类健康带来广泛的益处。阐明单个物种或群体的贡献 将为理解微生物控制的疾病提供合理的基础，并导致新的 治疗为了开展本建议中计划的基础研究，我们将解决三个主要问题： 首先，我们将建立第一套有效和精确调节微生物组的分子工具， 细菌;第二，我们将分析微生物群落的多尺度动态;第三，我们将构建 用于实时监测微生物组群的可摄取生物传感器。虽然抗生素和粪便 移植可以重新配置微生物财团，它们并不精确地针对单个细菌。相反地， 抗微生物肽（AMP）已经进化为选择性地攻击病原菌，但不靶向 微生物组细菌，构成分子工程的理想支架和微生物的潜在来源。 微生物组靶向剂。我们将开发一种新的计算肽设计方法，基于 经典和混合量子力学分子动力学（MD）模拟，创造一个突破性的 评估AMP的动力学和涌现特性。化学合成和大规模筛选 将确认对微生物组物种的预测选择性，机器学习工作流程将连接 单个肽的序列与它们的动力学和活性相关。然后，我们将把合成AMP应用于 通过在细菌聚生体实验期间选择性地去除物种来询问人类微生物组， 在常规或厌氧条件下在生物反应器中进行。我们将把我们的实验与全细胞 代谢网络模型，提供了一个系统生物学的角度来分析物种间的相互作用。 将开发一种集成的可摄入生物传感装置，通过电化学方法监测微生物组。 检测肠道微生物的独特生物标志物这将提供微生物组的首次实时测量 我们的生物反应器将被整合到我们的生物反应器中进行测试，最终用于体内测试。这 这项工作将建立第一套用于微生物组工程的分子和计算工具，并将奠定 基金会，以解决我们对肠道微环境的理解的关键差距， 肠道细菌对疾病病因的影响。基于我们在合成生物学方面的专业知识， 计算机科学，微生物学和电气工程，该项目将提供一个计算- 为肠道微生物组开发肽百科全书的实验框架，与NIH的公共 保健使命和目标。

ABSTRACT The gut microbiome has a tremendous impact on health and disease, actively contributing to obesity, diabetes, inflammatory bowel disease, cardiovascular diseases, and several poorly understood neurological disorders. We do not yet have the necessary tools to precisely probe these microbial communities, though such tools could unlock extensive benefits to human health. Elucidating the contributions of individual species or consortia of bacteria would provide a rational basis for understanding microbiota-controlled disease and lead to novel therapies. To carry out the fundamental research planned in this proposal, we will tackle three major problems: First, we will build the first set of molecular tools that effectively and precisely modulate the microbiome bacteria; second, we will analyze the multiscale dynamics of microbial communities; and third, we will construct an ingestible biosensor for real-time monitoring of microbiome populations. Although antibiotics and fecal transplants can reconfigure microbial consortia, they do not precisely target individual bacteria. Conversely, antimicrobial peptides (AMPs) have evolved to selectively attack pathogenic bacteria but do not target microbiome bacteria, constituting desirable scaffolds for molecular engineering and potential sources of microbiome-targeting agents. We will develop a new computational peptide design methodology, based on classical and hybrid-quantum mechanical molecular dynamics (MD) simulations, to create a groundbreaking assessment of the dynamical and emergent properties of AMPs. Chemical synthesis and large-scale screening will confirm predicted selectivity against microbiome species, and a machine learning workflow will connect sequences of individual peptides to their dynamics and activity. We will then apply the synthetic AMPs to interrogate the human microbiome by selectively removing species during bacterial consortia experiments, to be carried out in bioreactors, under regular or anaerobic conditions. We will pair our experiments with whole-cell metabolic network models, providing a systems biology perspective to the analysis of inter-species interactions. An integrated ingestible biosensing device will be developed to monitor the microbiome by electrochemically sensing unique biomarkers from gut microbes. This will provide the first real-time measurements of microbiome composition and will be integrated to our bioreactors for testing, to ultimately be used for in vivo tests. This work will build the first set of molecular and computational tools for microbiome engineering and will lay the foundation to address critical gaps in our understanding of the gut micro-environment, and of the contributions of gut bacteria to the etiology of disease. Grounded in our demonstrated expertise in synthetic biology, computer science, microbiology, and electrical engineering, this project will provide a computational- experimental framework for developing a peptide encyclopedia for the gut microbiome, in line with NIH's public health mission and goals.