Protein force fields have been undergoing continual development since the first complete parameter sets were introduced nearly four decades ago. The functional forms that underlie these models have many common elements for the treatment of bonded and nonbonded forces, which are reviewed here. The most widely used force fields to date use a fixed-charge convention in which electronic polarization effects are treated via a mean-field approximation during partial charge assignment. Despite success in modeling folded proteins over many years, the fixed-charge assumption has limitations that cannot necessarily be overcome within their potential energy equations. To overcome these limitations, several force fields have recently been derived that explicitly treat electronic polarization effects with straightforward extensions of the potential energy functions used by nonpolarizable force fields. Here, we review the history of the most popular nonpolarizable force fields (AMBER, CHARMM, OPLS, and GROMOS) as well as studies that have validated them and applied them to studies of protein folding and misfolding. Building upon these force fields are more recent polarizable interaction potentials, including fluctuating charge models, POSSIM, AMOEBA, and the classical Drude oscillator. These force fields differ in their implementations but all attempt to model electronic polarization in a computationally tractable manner. Despite their recent emergence in the field of protein folding, several studies have already applied these polarizable models to challenging problems in this domain, including the role of polarization in folding free energies and sequence-specific effects on the stability of α-helical structures.

自从大约40年前引入第一个完整的参数集以来，蛋白质力场一直在不断发展。这些模型基础的功能形式具有许多用于处理结合力和非结合力的常见元素，在此进行介绍。迄今为止，使用最广泛的力场使用的是固定电荷惯例，其中在部分电荷分配过程中通过均场近似来处理电子极化效应。尽管多年来成功地对折叠蛋白进行了建模，但固定电荷的假设仍然存在局限性，无法在其势能方程中加以克服。为了克服这些限制，最近已经获得了几种力场，它们通过不可极化力场使用的势能函数的直接扩展来明确地处理电子极化效应。在这里，我们回顾了最流行的非极化力场（AMBER，CHARMM，OPLS和GROMOS）的历史，以及对其进行了验证并将其应用于蛋白质折叠和错折叠研究的研究。在这些力场的基础上，出现了更多可极化的相互作用势，包括波动的电荷模型，POSSIM，AMEOBA和经典的Drude振荡器。这些力场的实现方式不同，但是所有力场都试图以计算上容易处理的方式对电子极化进行建模。尽管它们最近出现在蛋白质折叠领域，