Tilocca Antonio, Gamba Aldo, Vanoni Maria Antonietta, Fois Ettore
Dipartimento di Scienze Chimiche, Fisiche e Matematiche, Università dell'Insubria at Como, Via Lucini 3, I-22100 Como, Italy.
Biochemistry. 2002 Dec 3;41(48):14111-21. doi: 10.1021/bi020309q.
Large-scale Car-Parrinello molecular dynamics simulations of D-alanine oxidation catalyzed by the flavoenzyme D-amino acid oxidase have been carried out. A model of the enzyme active site was built by starting from the enzyme X-ray structure, and by testing different subsystems comprising different sets of aminoacyl residues. In this process, the stability of the enzyme-substrate complex was taken as a measure of the accuracy of the model. The activated transfer of the amino acid alpha-hydrogen from the substrate to the flavin N5 position was then induced by constraining a suitable transfer reaction coordinate, and the free energy profile of the reaction was calculated. The evolution of electronic and structural properties of both enzyme-bound substrate and flavin cofactor along the reaction path is consistent with a hydride-transfer mechanism. The calculated free energy barrier for this process (13 kcal/mol) is in excellent agreement with the activation energy value derived from the experimentally determined rate constant for the corresponding enzyme-catalyzed reaction. The electronic distribution of the reduced flavin shows that the transferred electrons tend to be centered near the C4a position rather than delocalized over the flavin pyrimidine ring. This feature is mechanistically relevant in that such an electronic distribution may promote the subsequent enzyme-catalyzed reduction of molecular oxygen to yield hydrogen peroxide via a postulated flavin 4a-peroxide intermediate. These results also show that a first-principles molecular dynamics approach is suitable to study the mechanism of complex enzymatic processes, provided that a smaller, yet reliable, subsystem of the enzyme can be identified, and special computational techniques are employed to enhance the sampling of the reactive event.
我们对黄素酶D-氨基酸氧化酶催化D-丙氨酸氧化反应进行了大规模的Car-Parrinello分子动力学模拟。从酶的X射线结构出发，通过测试包含不同氨基酸残基组的不同子系统，构建了酶活性位点模型。在此过程中，酶-底物复合物的稳定性被用作模型准确性的衡量标准。然后，通过约束合适的转移反应坐标，诱导氨基酸α-氢从底物向黄素N5位置的活化转移，并计算反应的自由能分布。沿着反应路径，与酶结合的底物和黄素辅因子的电子和结构性质的演变与氢化物转移机制一致。该过程计算得到的自由能垒（13千卡/摩尔）与从相应酶催化反应的实验测定速率常数推导出的活化能值非常吻合。还原型黄素的电子分布表明，转移的电子倾向于集中在C4a位置附近，而不是在黄素嘧啶环上离域。这一特征在机理上具有相关性，因为这种电子分布可能通过假定的黄素4a-过氧化物中间体促进随后酶催化的分子氧还原以产生过氧化氢。这些结果还表明，只要能够识别酶的一个较小但可靠的子系统，并采用特殊的计算技术来增强对反应事件的采样，第一性原理分子动力学方法就适用于研究复杂酶促过程的机制。
