Ripoll D R, Scheraga H A
Baker Laboratory of Chemistry, Cornell University, Ithaca, New York 14853-1301.
Biopolymers. 1990;30(1-2):165-76. doi: 10.1002/bip.360300116.
The conformational space of the membrane-bound portion of melittin has been searched using the electrostatically driven Monte Carlo (EDMC) method with the ECEPP/2 (empirical conformational energy program for peptides) algorithm. The former methodology assumes that a polypeptide or protein molecule is driven toward the native structure by the combined action of electrostatic interactions and stochastic conformational changes associated with thermal movements. The algorithm produces a Monte Carlo search in the conformational hyperspace of the polypeptide using electrostatic predictions and a random sampling technique, combined with local minimization of the energy function, to locate low-energy conformations. As a result of 8 test calculations on the 20-residue membrane-bound portion of melittin, starting from six arbitrary and two completely random conformations, the method was able to locate a very low-energy region of the potential with a well-defined structure for the backbone. In all of the cases under study, the method found a cluster of similar low-energy conformations that agree well with the structure deduced from x-ray diffraction experiments and with one computed earlier by the build-up procedure.
使用静电驱动蒙特卡罗（EDMC）方法和ECEPP/2（肽的经验构象能量程序）算法搜索了蜂毒肽膜结合部分的构象空间。前一种方法假设多肽或蛋白质分子通过静电相互作用和与热运动相关的随机构象变化的联合作用被驱动到天然结构。该算法使用静电预测和随机采样技术在多肽的构象超空间中进行蒙特卡罗搜索，并结合能量函数的局部最小化，以定位低能量构象。对蜂毒肽20个残基的膜结合部分进行了8次测试计算，从6个任意构象和2个完全随机构象开始，该方法能够定位势能的一个非常低能量区域，其主链结构明确。在所有研究的案例中，该方法都发现了一组相似的低能量构象，这些构象与X射线衍射实验推导的结构以及早期通过构建程序计算的结构非常吻合。
