Modeling-based control of the crystallizability of proteins in biotechnological processes

到目前为止，蛋白质在技术规模上结晶的适当条件只能通过经验来确定。本研究的目的是用一种基于模型的方法来取代传统的精心设计的经验方法。在第一个资助期内，对典型蛋白质短链乳杆菌酒精脱氢酶(LbADH)的特定突变株进行了中子和X射线衍射术以及分子动力学模拟。结果表明，合理的蛋白质工程可以改善蛋白质的结晶，这是可以从力学上解释的。在后续项目过程中，将首先将所获得的知识转移到相关酶--开菲尔乳杆菌酒精脱氢酶(LkADH)。然后，将通过结晶学和MD模拟来研究促进LbADH结晶的突变是否以类似的方式影响LkADH结晶过程中的分子间力。LkADH及其突变体结晶行为的实验结果将用于理论方法的验证和推广。在此基础上，首次利用对蛋白质结晶及其可控性的理论认识，对两种尚未结晶的蛋白质--二羟基酸脱水酶和烯-还原酶进行从头结晶研究。同源模型将被用来确定可以形成潜在晶体接触的氨基酸位置。通过应用合理的蛋白质工程，随后将诱导晶体接触处相互作用的增强。所开发的分子动力学模拟方法将被用于预测结晶行为。在实验验证之后，对蛋白质结晶的分子洞察力将得到扩展。实验评估将通过使用高通量筛选方法(HTS)表征蛋白质的相行为以及使用高频流变仪分析其在溶液中的流变特性来执行，以全面了解溶解蛋白质的粘弹性行为与其结晶倾向之间的关系。最后，将开发一种过程分析技术(PAT)方法，通过使用UV/VIS光谱选择性地量化目标蛋白质浓度的变化，从而实现对蛋白质结晶过程的实时监测。TUM和KIT的联合研究项目涉及分子尺度(蛋白质工程和MD模拟)以及微观尺度(相行为和流变学)。此外，还包括了宏观尺度(PAT)的分析，以便能够完整地从根本上阐明技术蛋白质结晶的基本工程原理。

Up to now, proper conditions for crystallization of proteins on a technical scale can only be identified empirically. Aim of the present research project is to substitute the conventional elaborate empirical approach by a modeling-based approach. During the first funding period, neutron and X-ray diffractometry and MD simulations were applied to specifically generated mutants of the exemplary protein Lactobacillus brevis alcohol dehydrogenase (LbADH). It was shown that rational protein engineering led to an improvement of protein crystallization which was mechanistically explainable. In the course of the follow-up project, a transfer of the gained knowledge to a related enzyme, Lactobacillus kefir alcohol dehydrogenase (LkADH), will be performed first. Then, it will be investigated via crystallographic and MD simulations whether mutations that had improved the crystallization of LbADH affect the inter-molecular forces during crystallization of LkADH in a similar manner. The experimental results of the crystallization behavior of LkADH and selected mutants will be used for the validation and extension of the theoretical methods. Based on this work, the gained theoretical insights into protein crystallization and its controllability are to be utilized for the first time for the de novo crystallization of two proteins that have not yet been crystallized, a dihydroxy-acid dehydratase and an ene-reductase. Homology models will be used to determine amino acid positions that can form potential crystal contacts. By applying rational protein engineering, an intensification of the interactions at the crystal contacts is subsequently to be induced. The developed MD simulation methods will be applied to predict the crystallization behavior. After experimental validation, the molecular insights into protein crystallization will be extended. The experimental evaluation will be performed by characterizing the phase behavior of proteins using high-throughput screening methods (HTS) as well as by analyzing their rheological characteristics in solution using high-frequency rheometry in order to gain a comprehensive understanding of the relationship between the viscoelastic behavior of dissolved proteins and their crystallization propensity. Finally, a process-analyzing technology (PAT) method will be developed in order to enable a real-time monitoring of the protein crystallization process via selective quantification of changes in the concentration of the target protein using UV/Vis spectroscopy. The joint research project of TUM and KIT addresses the molecular scale (protein engineering and MD simulations) as well as the microscopic scale (phase behavior and rheology). Furthermore, an analysis on the macroscopic scale (PAT) is included in order to enable an integral fundamental elucidation of basic engineering principles of technical protein crystallization.