An Integrated Computational and Experimental Approach to Reveal Design Principles for Responsive Nanomaterials from Lipidated Disordered Proteins

非技术摘要。脂质修饰蛋白质是一类新兴的生物材料，在材料科学和医疗保健领域有着广泛的应用。然而，为了实现它们的前景，关键是要全面了解在一系列溶液条件下管理这些杂化生物材料的形成和性能的基本原理。这个合作项目将结合实验和模拟，通过改变脂化蛋白质的组成和温度来创建具有用户定义特性的纳米颗粒的设计蓝图。机器学习算法将被用来迭代地整合来自模拟和实验的反馈，并推导出指导纳米生物材料设计的预测规则，这些规则适用于从提供化疗药物到纳米材料模板合成的特定应用。项目负责人还将为来自不同背景的学生，包括妇女和STEM中代表性较低的少数民族，开发一个基于实验和计算队列的综合研究体验计划。在化学、生物、材料科学和物理的界面上进行研究培训，将使这些学员能够推进知识前沿，加速材料创新，并为美国S在全球生物经济中的领导地位做出贡献。尽管在过去十年中取得了进展，但由脂质和蛋白质组成的杂化材料的设计在很大程度上仍然是一个特别的过程，阻碍了该领域的进展。这一限制的出现是因为脂肪-蛋白质生物材料的设计空间的绝对大小阻碍了对设计规则的经验性说明。因此，为了有效地揭示基本的设计原则，需要整合实验、模拟和机器学习算法的新方法。这一合作项目利用了研究团队在生物合成以及脂化蛋白质的计算和实验表征方面的互补专业知识。该项目将研究脂质修饰的内在无序蛋白质聚合物(Lipo-IDPps)，它将脂类的分层组织与IDPps的温度响应行为结合起来，形成具有温度依赖特性的纳米和介观组装。使用这一模型，该团队将开发预测性设计规则，以编程Lipo-IDPP的热响应和分子语法(其构建块的物理化学、其初级序列、拓扑和两亲性结构)的分层组装。将追求两个目标，每个目标都使用建模、合成和表征一系列具有精确遗传编码语法的Lipo-IDPP的闭环策略。使用一种巧妙地结合实验、模拟和机器学习的迭代反馈的综合方法，该团队将：(1)基于其分子语法开发Lipo-IDPPs的温度响应预测模型；(2)确定根据生理相关温度定制其组装的结构层次的大分子蓝图。将使用多尺度实验(光谱、散射和显微镜)和计算(原子模拟和粗粒度模拟)方法来表征一系列Lipo-IDPP的材料性质随温度的变化。将使用机器学习方法将实验和计算结果结合到一个模型中，将分子属性映射到观察到的材料属性。优化的模型将深入了解不同分子语法成分对这类材料的可编程温度响应组装的生物物理贡献，并可用于制定严格的预测规则，用于在生物相关环境中设计具有所需性质的Lipo-IDPps。阐明支配Lipo-IDPPs多尺度组织的设计原则将使具有遗传可编程分子语法和性质的响应材料的合理合成成为可能。将来自硅胶和实验表征技术的迭代反馈与适用于其他混合材料的数据分析相结合，将加快生物材料的设计和发现。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Non-Technical Summary. Proteins modified with lipids are an emerging class of biomaterials with diverse applications in materials science and healthcare. However, to realize their promise, it is critical to develop a comprehensive understanding of foundational principles governing the formation and properties of these hybrid biomaterials under a range of solution conditions. This collaborative project will integrate experiments and simulations to create a blueprint for the design of nanoparticles with user-defined properties by changing the composition of lipidated proteins and temperature. Machine learning algorithms will be used to iteratively integrate feedback from simulations and experiments and to derive predictive rules that guide the design of nano-biomaterials for specific applications ranging from the delivery of chemotherapeutics to the templated synthesis of nanomaterials. The project leaders also will develop an integrated experimental and computational cohort-based research experience program for students from diverse backgrounds, including women and underrepresented minorities in STEM. Research training at the interface of chemistry, biology, materials science, and physics will enable these trainees to advance the frontiers of knowledge, accelerate materials innovation, and contribute to U.S.’s leadership in the global bioeconomy.Technical Summary. Despite advancements in the past decade, the design of hybrid materials comprising lipid and protein building blocks remains a largely ad hoc process, impeding progress in the field. This limitation arises because the sheer size of the design space of lipid-protein biomaterials prohibits empirical elucidation of design rules. Thus, to efficiently reveal foundational design principles, new approaches that integrate experiments, simulations, and machine learning algorithms are needed. This collaborative project leverages the research team’s complementary expertise in biosynthesis as well as computational and experimental characterization of lipidated proteins. The project will investigate lipid-modified intrinsically disordered protein polymers (Lipo-IDPPs), which combine the hierarchical organization of lipids with temperature-responsive behavior of IDPPs to form nano- and meso-assemblies with temperature-dependent characteristics. Using this model, the team will develop predictive design rules for programming the thermo-response and hierarchical assembly of Lipo-IDPPs in their molecular syntax (the physicochemistry of their building blocks, their primary sequence, topology, and amphiphilic architecture). Two objectives will be pursued, each of which uses a closed-loop strategy of modeling, synthesis, and the characterization of a series of Lipo-IDPPs with precise genetically encoded syntax. Using an integrated approach that judiciously combines iterative feedback from experiments, simulations, and machine learning, the team will: (1) develop a predictive model of Lipo-IDPPs’ thermo-response based on their molecular syntax and (2) identify a macromolecular blueprint for tailoring the structural hierarchy of their assemblies as a function of physiologically relevant temperatures. The material properties of a series of Lipo-IDPPs as a function of temperature will be characterized using multiscale experimental (spectroscopy, scattering, and microscopy) and computational (atomistic and coarse-grained simulations) approaches. Machine learning methods will be used to combine experimental and computational results into a model for mapping molecular attributes to observed material properties. The optimized model will provide insights into the biophysical contribution of different components of molecular syntax to the programmable temperature-responsive assembly of this class of materials and can be used to formulate rigorous and predictive rules for the inverse design of Lipo-IDPPs with desired properties in biologically relevant milieus. Elucidating the design principles governing the multiscale organization of Lipo-IDPPs will enable the rational synthesis of responsive materials with genetically programmable molecular syntax and properties. And the integration of iterative feedback from in silico and experimental characterization techniques with data analytics, which is applicable to other hybrid materials, will accelerate biomaterials’ design and discovery.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.