Roy H, Kupferschmid M, Bell J A
Center for Biophysics, Rensselaer Polytechnic Institute, Troy, New York 12180-3590.
Protein Sci. 1992 Jul;1(7):925-34. doi: 10.1002/pro.5560010711.
We have performed a computational simulation of the aggregation and chaperonin-dependent reconstitution of dimeric prokaryotic ribulose bisphosphate carboxylase/oxygenase (Rubisco), based on the data of P. Goloubinoff et al. (1989, Nature 342, 884-889) and P. V. Viitanen et al. (1990, Biochemistry 29, 5665-5671). The aggregation is simulated by a set of 12 differential equations representing the aggregation of the Rubisco folding intermediate, Rubisco-I, with itself and with aggregates of Rubisco-I, leading up to dodecamers. Four rate constants, applying to forward or reverse steps in the aggregation process, were included. Optimal values for these constants were determined using the ellipsoid algorithm as implemented by one of us (Ecker, J.G. & Kupferschmid, M., 1988, Introduction to Operations Research, Wiley, New York, pp. 315-322). Intensive exploration of simpler aggregation models did not identify an alternative that could simulate the data as well as this one. The activity of the chaperonin in this system was simulated by using this aggregation model, combined with a model similar to that proposed by Goloubinoff et al. (1989). The model assumes that the chaperonin can bind the folding intermediate rapidly, and that the chaperonin complex releases the Rubisco molecule slowly, permitting time for its spontaneous folding while interacting with the chaperonin. This is followed by self-association of the folded Rubisco monomer to yield the active dimeric Rubisco. A modification of the model that simulates temperature effects was also constructed. The most important results we obtained indicate that the chaperonin-dependent reconstitution of Rubisco can be simulated adequately without invoking any catalysis of folding by the chaperonin. In addition, the simulations predict values for the association rate constant of Rubisco-I with the chaperonin, and other variables, that are subject to experimental verification.
基于P. Goloubinoff等人（1989年，《自然》342卷，884 - 889页）以及P. V. Viitanen等人（1990年，《生物化学》29卷，5665 - 5671页）的数据，我们对二聚体原核生物核酮糖二磷酸羧化酶/加氧酶（Rubisco）的聚集以及伴侣蛋白依赖性重构进行了计算模拟。聚集过程通过一组12个微分方程进行模拟，这些方程表示Rubisco折叠中间体Rubisco-I自身之间以及与Rubisco-I聚集体之间的聚集，直至形成十二聚体。其中包括四个适用于聚集过程中正向或反向步骤的速率常数。这些常数的最优值是使用我们其中一人（Ecker, J.G. & Kupferschmid, M.，1988年，《运筹学导论》，Wiley出版社，纽约，第315 - 322页）所实现的椭球算法确定的。对更简单聚集模型的深入探索并未找到能像这个模型一样很好地模拟数据的替代方案。通过将这个聚集模型与一个类似于Goloubinoff等人（1989年）提出的模型相结合，模拟了该系统中伴侣蛋白的活性。该模型假定伴侣蛋白能够快速结合折叠中间体，并且伴侣蛋白复合物缓慢释放Rubisco分子，从而在与伴侣蛋白相互作用的同时为其自发折叠留出时间。随后，折叠后的Rubisco单体进行自我组装，产生有活性的二聚体Rubisco。还构建了一个模拟温度效应的模型修改版本。我们获得的最重要结果表明，在不调用伴侣蛋白对折叠的任何催化作用的情况下，能够充分模拟伴侣蛋白依赖性的Rubisco重构。此外，模拟预测了Rubisco-I与伴侣蛋白的缔合速率常数以及其他变量的值，这些值有待实验验证。
