在文章Interfering peptides targeting protein–protein interactions: the next generation of drugs?（https://doi.org/10.1016/j.drudis.2017.10.016）中介绍基于蛋白复合物PDB结构已知的情况下设计interfering peptides（IPs）存在如下几种技术方案：
When the structure of the complex can be solved, IPs can be rationally designed based on the direct observation of the natural sequences that mediate PPIs in the proteins. It has been observed that only a few hot segments involving residues located at the PPI seem responsible for the interaction between the partners, and the analyses of the 3D structures in interaction using structural bioinformatics methods can help to identify peptides from one partner that could bind to the other partner and thus interfere with the PPI. LoopFinder, PeptiDerive or searching for helix interfaces in protein–protein interactions (HIPPs) are examples of in silico approaches exploiting this observation. PeptiDerive systematically splits the chains in the interaction as series of fragments of 10 amino acids and identifies those corresponding to hot segments. It has been used successfully to design agonists of the MD2–TLR4 interaction. It is interesting to note that the linear peptide identified originally was not active but inserting a disulfide bond stabilized its variant. Consequently, PeptiDerive now specializes in the identification of hot segments compatible with the use of peptides that can be closed by a disulfide bridge. The HIPP approach specializes in the identification of helical segments that can undergo further modifications such as stapling, whereas LoopFinder focuses on loops that can be cyclized. Submicromolar inhibitors of stonin2 and Eps15, designed using LoopFinder, have recently been reported.
其中介绍的策略包含1）LoopFinder查找loop区域，2）PeptiDerive查找对于PPI能量贡献最大的区域（hotspot segments），3）helix区域（helix interfaces in protein-protein interactions，HIPPs）。此文主要介绍基于Rosetta PeptiDerive的技术策略和实例：
原理介绍主要基于文章Peptiderive server: derive peptide inhibitors from protein–protein interactions（10.1093/nar/gkw385），其过程图如下：
Figure 1. Rosetta PeptiDerive的过程介绍
输入为蛋白复合物的PDB结构文件，1）在对PDB进行相应的clean和能量最小化操作后，2）将对应蛋白链按照指定的长度（默认为10个氨基酸）滑动切分序列，3）切分的片段，利用Rosetta Energy Function进行能量贡献度计算；4）选择贡献度最大的片段进行线性或成环设计。
实例文章参考Rationally designed macrocyclic peptides as synergistic agonists of LPS-induced inflammatory response (http://dx.doi.org/10.1016/j.tet.2014.07.026)，此文基于TLR4-MDM2（3FXI）的PDB结构进行多肽设计，利用Rosetta PeptiDerive得到的贡献度最大的segments为DDDYSFCRA（TLR4作为受体，MDM2作为配体进行设计的，竞争性抑制MDM2与TLR4的结合），针对性的设计YH1和YH3，其中可以看到利用半胱氨酸（C）进行了成环的操作，YH2和YH4作为对照，实验验证YH1和YH3有一定的作用效果。
Figure 2. TLR4-MDM2的IPs设计
具体的命令操作执行，本文是基于linux上安装的Rosetta可执行文件进行实际代码运行查找hotspot segments，另外可用在线版[ROSIE] Peptiderive Protocol (jhu.edu)进行分析：
PeptideDeriver.mpi.linuxclangrelease -s 3FXI_rec.pdb #此pdb只在pymol去除了一些水和离子，未做其他操作，其结果输出为3FXI_rec_0001.peptiderive.txt，不过耗时有点长，大概4个小时左右，随后将计算步骤增加一步Rosetta提供的clean操作
clean_pdb.py 3FXI_rec.pdb ignorechain #pdb文件的清洗和优化
relax.mpi.linuxclangrelease -s 3FXI_rec_ignorechain.pdb #结构优化
PeptideDeriver.mpi.linuxclangrelease -s 3FXI_rec_ignorechain.pdb #30来分钟产出结果，会发现它的结果比之前的较为精炼，主要是PDB文件经过clean后，有很大简化
备注：binding energy的计算公式可参考[rosie] Peptiderive Server Documentation (jhu.edu)介绍或者基于文章附件介绍；Rosetta energy function的能量计算可参考介绍Rosetta基础3: Rosetta能量函数简介 - 知乎 (zhihu.com)。
网友评论