抗菌肽广泛存在于高等动物、昆虫、节肢动物、被囊动物和植物中，是真核生物天然免疫系统的第一道防线。抗菌肽通常对革兰氏阴性和革兰氏阳性菌、酵母和真菌表现出广泛的活性，部分抗菌肽也有助于治疗癌症。最重要的是，抗菌肽对传统抗生素的耐药病原体具有活性，这为解决细菌耐药性这一全球性危机、以及开发新型抗菌药物提供了希望。目前已有一些抗菌肽应用于临床前和临床试验中。但它们在分子水平上的抗菌机制仍不清楚，远非“桶板”、“地毯”、“环形孔”等经典杀菌模型所能解释，还需要进行深入细致的研究。本论文研究的是两类抗菌谱广、杀菌活性强、在化学合成上有明显效益优势的短肽Gomesin和Temporin。

Gomesin是从巴西蜘蛛血细胞中分离出来的β-折叠短肽，实验上光学显微镜观察表明该肽并不是通过形成稳定的跨膜孔来破坏细菌膜，而是诱导脂质堆积，随后膜在该位置附近突然破裂，因此从实验上推测其抗菌机制为“地毯”模型。然而由于实验分辨率的限制，其作用过程的机制仍是未知的。而Temporin 则是从欧洲红蛙的皮肤分泌物中得到的α-螺旋短肽，是一类极短的抗菌肽，带电量极低。根据实验观察，科学家们提出了多种模型解释其抗菌机制，包括“地毯”、“肽-脂质动态超分子孔”和“漏缝”模型，但它们与膜相互作用的分子细节并不清晰。本论文采用分子动力学模拟的方法系统地研究了这两类长度较短、活性较强且具有不同二级结构的抗菌肽与细胞膜的相互作用，在分子层面阐明了它们独特的杀菌机制。主要工作如下：

（1）论文的第一个工作首次采用无水粗粒化分子动力学模拟的方法研究了β-折叠结构的外周型抗菌肽Gomesin与由中性和阴性离子磷脂组成的囊泡和平面双分子层脂质膜作用的分子机制，并与成孔肽蜂毒肽诱导的膜形态和性质变化进行比较。模拟结果表明疏水面较小的楔形肽 Gomesin只浅浅地吸附在膜的外层，并引起外层膜的突出和折叠，随后膜在突起腋窝处突然撕裂；而具有均等亲疏水表面的圆柱形蜂毒肽通过诱导囊泡出芽、膜内陷和稳定孔的方式破坏膜。

（2）论文的第二个工作采用全原子和粗粒化分子动力学模拟结合的方法，研究了Temporin B和L诱导平面双分子层形态的变化，发现了它们特殊的抗菌机制。这些肽在膜表面折叠成α-螺旋，并且浅穿透膜的外层，使得它们诱导形成膜孔的倾向性较低，但却具有强烈的挤压拉出脂质的能力。在相对较高的肽浓度下，Temporin B和L的强疏水性促使它们在膜表面聚集成团簇，这些聚集体吸引大量脂质离开膜表面，以释放其它分散的肽与膜结合所引起的张力。挤压出的脂质逐渐拉长并与其它部分融合，形成管状突起。随着脂质的运动，水分子填充突起的空腔，并帮助维持管状结构。与此同时，无肽的内层膜保持完整。在某些特定肽浓度下Temporin B聚集并以“桶板”的方式插入膜，诱导形成瞬时跨膜孔，并允许水分子通过。

Antimicrobial peptides are widely found in higher animals, insects, arthropods, tunicates and plants. They are the first line of defense for the natural immune system of eukaryotes. These peptides generally show a wide range of activity against gram-negative and gram-positive bacteria, yeasts and fungi, and some may also help treat cancer. Most importantly, antimicrobial peptides are active against pathogens resistant to traditional antibiotics. They provide hope for solving the global crisis of bacterial resistance and developing new antimicrobial drugs. Several antimicrobial peptides have been used in preclinical and clinical trials. However, a detailed understanding of their mechanisms of action at the molecular level is still lacking. Classical antimicrobial mechanism models such as “barrel-starve”, “carpet” and “toroidal” cannot well describe the real antibacterial action process, specific studies are still needed. In this thesis, two kinds of short peptides, Gomesin and Temporin, which have broad antibacterial spectrum, strong bactericidal activity and obvious advantages in chemical synthesis, were studied.

Gomesin is a β-hairpin short peptide isolated from the hemocytes of the Brazilian spider. Optical microscopy showed that the peripheral peptides did not destroy the bacterial membrane by forming stable membrane pores, but induced lipid accumulation and sudden membrane rupture near this domain. The antibacterial mechanism of Gomesin was speculated to be a “carpet” model. However, due to the limited experimental resolution, the molecular mechanism of its action process is still unknown. Temporins are α-helical short peptides derived from the skin secretions of the European red frog. They are one of the shortest antibacterial peptides with extremely low charge. Several mechanism models have been proposed according to experimental observations, including “carpet”, “dynamic peptide-lipid supramolecular pores” and “leaky slit”. But the molecular details of their interaction with membranes are also in suspense. In order to solve the above two problems, this thesis employed molecular dynamics simulation method to systematically study the interaction between these antimicrobial peptides with short length, high antibacterial activity but different secondary structure and cell membrane. Their unique bactericidal mechanisms were elucidated. The main work of this thesis are as follows:

(1) The first work of this thesis investigated the molecular mechanism of the β-hairpin structured peripheral antimicrobial peptide Gomesin with vesicles and planar bilayers composed of zwitterionic and anionic phospholipids using implicit solvent molecular dynamic simulation method at coarse-grained level. Changes in membrane morphologies and properties induced by Gomesin and the pore-forming peptide Melittin were also compared. The simulation results showed that Gomesin, a wedge-shaped peptide with a small hydrophobic surface only shallowly penetrates into the membrane, causing the protrusion and folding of the upper leaflet, followed by a sudden membrane laceration at the axillary of the protrusion. Whereas, cylinder-shaped Melittin with equal hydrophobic and hydrophilic surfaces disrupts membranes by inducing vesicle budding, invagination, and stabilizing pores.

(2) The second work of this thesis adopted a combination of all-atom and coarse-grained molecular dynamics simulations to study the antimicrobial behavior of Temporin B and L. The results showed that they perform bactericidal activities in a way different from many other antimicrobial peptides. These peptides fold into α-helices on the membrane surface and shallowly penetrate the bilayer. They have low tendency to induce pore formation, but have strong abilities to squeeze and pull lipids out. At relatively high peptide concentrations, the high hydrophobicity of Temporin B and L promotes them to aggregate on the membrane surface. These clusters attract lipids away from the membrane surface at multiple sites to release the tension caused by the binding of other dispersed peptides to the membrane. The extruded lipids gradually elongate and fuse with others to form tubular protrusions. As the lipids move, water molecules enter the cavity between the protrusion and the mother membrane maintaining the tubular structure. At the same time, the inner leaflets remain intact. At a specific peptide concentration, the aggregated Temporin B insert into the membrane in a “barrel-stave” manner, inducing the formation of transient transmembrane pores and allowing water molecules to pass through.

The results of our work demonstrate that amphiphilicity of antimicrobial peptides is a key factor affecting the mechanism of cell death. Molecular dynamics simulations shed light on the bactericidal mechanisms of various antimicrobial peptides. This thesis provides significance guidance on drug delivery and the design of novel antibacterial drugs.