OneG: A Computational Tool for Predicting Cryptic Intermediates in the Unfolding Kinetics of Proteins under Native Conditions

Understanding the relationships between conformations of proteins and their stabilities is one key to address the protein folding paradigm. The free energy change (ΔG) of unfolding reactions of proteins is measured by traditional denaturation methods and native hydrogen-deuterium (H/D) exchange methods. However, the free energy of unfolding (ΔGU) and the free energy of exchange (ΔGHX) of proteins are not in good agreement, though the experimental conditions of both methods are well matching to each other. The anomaly is due to any one or combinations of the following reasons: (i) effects of cis-trans proline isomerisation under equilibrium unfolding reactions of proteins (ii) inappropriateness in accounting the baselines of melting curves (iii) presence of cryptic intermediates, which may elude the melting curve analysis and (iv) existence of higher energy metastable states in the H/D exchange reactions of proteins. Herein, we have developed a novel computational tool, OneG, which accounts the discrepancy between ΔGU and ΔGHX of proteins by systematically accounting all the four factors mentioned above. The program is fully automated and requires four inputs: three-dimensional structures of proteins, ΔGU, ΔGU* and residue-specific ΔGHX determined under EX2-exchange conditions in the absence of denaturants. The robustness of the program has been validated using experimental data available for proteins such as cytochrome c and apocytochrome b562 and the data analyses revealed that cryptic intermediates of the proteins detected by the experimental methods and the cryptic intermediates predicted by the OneG for those proteins were in good agreement. Furthermore, using OneG, we have shown possible existence of cryptic intermediates and metastable states in the unfolding pathways of cardiotoxin III and cobrotoxin, respectively, which are homologous proteins. The unique application of the program to map the unfolding pathways of proteins under native conditions have been brought into fore and the program is publicly available at http://sblab.sastra.edu/oneg.html阐明蛋白质构象与其稳定性之间的关联，是破解蛋白质折叠范式的关键之一。传统变性实验方法与天然态氢氘（hydrogen-deuterium, H/D）交换实验方法，均可用于测定蛋白质解折叠反应的自由能变化（ΔG）。然而，尽管两种方法的实验条件高度匹配，蛋白质的解折叠自由能（ΔGU）与交换自由能（ΔGHX）却并不一致。这一异常现象可归因于以下一项或多项因素：(i) 蛋白质平衡态解折叠反应中顺反脯氨酸异构化（cis-trans proline isomerisation）的影响；(ii) 解链曲线（melting curves）基线校正不当；(iii) 存在可能逃过解链曲线分析的隐蔽中间体（cryptic intermediates）；以及(iv) 蛋白质氢氘交换反应中存在高能亚稳态（higher energy metastable states）。 本研究开发了一款全新的计算工具OneG，通过系统整合上述四大影响因素，解决了蛋白质ΔGU与ΔGHX之间的数值差异问题。该程序完全自动化，仅需四项输入数据：蛋白质的三维结构、ΔGU、ΔGU*，以及无变性剂（denaturants）条件下于EX2交换（EX2-exchange conditions）环境中测定的残基特异性ΔGHX（residue-specific ΔGHX）。 我们采用细胞色素c（cytochrome c）与脱辅基细胞色素b562（apocytochrome b562）的公开实验数据，验证了该程序的稳健性。数据分析结果显示，实验方法检测到的蛋白质隐蔽中间体，与OneG针对这些蛋白预测得到的隐蔽中间体高度吻合。 此外，借助OneG，我们证实了同源蛋白质心脏毒素III（cardiotoxin III）与眼镜蛇毒素（cobrotoxin）的解折叠通路（unfolding pathways）中，分别存在隐蔽中间体与亚稳态。本程序可用于绘制天然态条件下蛋白质的解折叠通路，这一独特应用价值已得到充分彰显，且该程序已于http://sblab.sastra.edu/oneg.html 公开可用。