利用指向性圓二色光譜、多片層 X-光繞射以及小角度 X-光散射等 技術探討人工抗菌胜肽 GW-H1 和 GW-Q4 在生物膜上的穿透機制

Abstract

抗菌胜肽在昆蟲、植物及哺乳動物的先天免疫系統及宿主防禦機制上扮演著重要的角色。目前認為，抗菌胜肽造成微生物死亡的機制，多數可能是藉由與微生物細胞膜相互作用後反應引發孔洞形成等幾種穿透機制，來破壞微生物的正常生理。近年來，隨著抗生素過度的濫用，導致微生物抗藥性的迅速增加。因此，發展有效性抗菌胜肽取代抗生素，成為克服微生物抗藥性問題的新型態之治療方式。 我們根據前人設計的人工合成抗菌胜肽，固定長度為20個胺基酸，改變其電荷 (Q)、極性角度 (θ)、疏水性 (H) 及疏水性矩 (MH) 等四種結構性參數 (structural determinants)，得到數種有效抗菌胜肽。其中GW-H1（Q＝＋4、θ＝140°、H＝-0.115、MH＝0.344）及GW-Q4（Q＝＋4、θ＝140°、H＝-0.043、MH＝0.344），對革蘭氏陽性菌及革蘭氏陰性菌（包括弧菌屬病原菌）擁有最強的抗菌效果及最佳的滲透能力，並且對原核生物的細胞膜有選擇性。而相關的 calcein leakage 與圓二色光譜 (circular dichroism, CD) 實驗結果，則可用來推論 GW-H1與GW-Q4所造成的抗菌能力，應該來自於與細菌細胞膜直接作用時所造成的物質滲漏效果。 為瞭解這些具抗菌效力的人工胜肽GW-H1與GW-Q4如何作用在細菌細胞膜上造成穿透，我們利用DOPC:DOPG (3:1) 之微脂體及多片層脂膜作為細菌細胞膜的模型，並運用生物物理相關技術來量測胜肽在膜上的方位和膜厚度的改變。其中，使用指向性圓二色光譜 (oriented circular dichroism, OCD) 可以觀察到 α 螺旋型胜肽在膜上的方位：若是平行膜面，表示胜肽吸附在膜表面；垂直膜面則表示胜肽插入膜內，吸附於孔洞邊緣。脂膜 X-光片層繞射 (lamellar x-ray diffraction) 是用來測量胜肽造成固定相脂質雙層膜厚度的變化；小角度 X-光散射 (small angle x-ray scattering, SAXS) 則是測量胜肽造成水溶液中單層微脂體 (small unilamellar vesicles, SUVs) 厚度的變化。當以上述方法，觀察人工合成抗菌胜肽 GW-H1、GW-Q4 與天然抗菌胜肽 melittin、pleurocidin 在帶電生物膜上的吸附或插入的情形，並配合磷脂質與胜肽比例 (P/L) 之改變，我們可以清楚討論胜肽的作用方位、生物膜厚度改變的程度、微脂體粒徑的變化，進而推導理解抗菌胜肽作用在生物膜上的特性與穿透的可能機制。 實驗結果顯示，人工抗菌胜肽 GW-H1 及 GW-Q4 迥異於天然抗菌胜肽melittin、pleurocidin，並非以已知的形成孔洞的模式與生物膜作用，反而是不斷吸附在生物膜表面，造成膜厚度持續地下降。粒徑分析則顯示，微脂體沒有破裂或融合的現象。根據文獻上指出關於脂膜的物理性質：脂膜分子會受到熱擾動作用力 (thermal fluctuation force) 的影響，產生暫時性的孔洞 (transient pores)。我們因此推論人工抗菌胜肽 GW-H1、GW-Q4 在與生物膜作用時，可能是逐漸吸附聚集於膜表面的胜肽數目愈來愈多，且膜厚度越趨變薄，促使膜表面張力不斷增加，而產生類似熱擾動的效果，於是膜表面出現暫時性孔洞的頻率變高或孔洞本身打開的幅度變大，終致生物膜的障壁功能暫時喪失，膜內外物質得以彼此交會進出。這應該是對於 calcein leakage 實驗中證明膜內外物質確實有進行交換，但粒徑分析卻又顯示膜結構並未真正大幅改變的較佳解釋。

相對的，已知為形成超環面孔洞模式 (toroidal-pore model) 的 melittin，在本研究中被觀察到具有明顯的吸附插入相的轉變（水平或垂直）。透過測量發現 melittin 在 DOPC/PG 膜上形成孔洞的閾值為 (P/L)* = 1/200，相較文獻中的對於中性生物膜 DOPC之閾值 (P/L)* = 1/99 約小一倍，顯示melittin 相對容易結合在帶負電的 DOPG 生物膜中，並附帶證實了陽離子抗菌胜肽與帶負電荷的細菌細胞膜作用的機制最初是先藉由靜電作用力吸附，方得以引發後續的抑菌反應。此外實驗結果亦顯示 pleurocidin 的作用情況與 melittin 相似，因此我們認為 pleurocidin 可能也以形成孔洞的方式 (pore-formation) 對生物膜進行穿透，此結論亦符合相關文獻中電生理實驗的推論。另外，pleurocidin與細胞膜的作用似乎較 melittin 更強，初步實驗結果顯示細胞膜與 pleurocidin 作用後有導致細胞膜完全破壞的可能。 本研究提供了極為先進的實驗平台，為身為本院藥理毒理傳統卻長期以來不易進一步逼近的生物膜和胜肽（如蛇毒和蜘蛛毒）作用模式細節研究展露了可遵循的法則。多種物理參數的量測推導更具體化了判定生物膜和胜肽反應時作用方式的分類標準，而欲掌握更多的反應中物理化學性質和細節的推導則亦透過同步輻射的高解析光源成為可能。只是關於抗菌胜肽的結構參數與抗菌機制的關聯仍是謎團，我們必須仰賴實驗觀測分類推估，無法從這些參數中直接演繹推論人工抗菌胜肽 GW-H1、GW-Q4 與其他天然抗菌胜肽的抗菌機制之異同，這有待未來更多的研究去瞭解。

Antimicrobial peptides (AMPs) play important roles in the host innate defense mechanism in many plants, insects, and mammals. It is believed that AMPs may interact with the microbial membranes and kill the target cells. On the other hands, the extensive use of classical antibiotics has led to the growing emergence of many resistant strains of pathogenic microorganisms in recent years. Therefore, the development of novel therapeutic agents that could overcome the antimicrobial resistance has become a very critical issue. The mechanism for the aforementioned antimicrobial activity has been considered as the membrane-peptide interactions and the subsequent pore- forming that lead to the permeation of biomembranes. Several models have been proposed according to the membrane structure types during pore-formation: “barrel-stave”, “carpet” and “toroidal-pore”. Based on the previous study of our collaborators, a series of cationic α-helical peptides with 20 amide acids has been designed and synthesized according to four structural determinants: charge, polar angle, hydrophobicity, and hydrophobic moment. Two of such de novo designed AMPs, GW-H1 (Q＝＋4, θ＝140°, H＝-0.115, MH＝0.344) and GW-Q4 (Q＝＋4, θ＝140°, H＝-0.043, MH＝0.344), exhibited the most significant antimicrobial activity and selectivity against various Gram-positive and Gram-negative bacteria, including several vibrio strains. Results form the related calcein leakage experiments and circular dichroism spectra are used to infer that the antimicrobial activity of GW-H1 and GW-Q4 should rely on the direct interaction with prokaryotic membranes and the concomitant penetration effect that can lead to the microbial death. In this study, to distinguish the type of membrane-peptide interactions, which will allow us to deduce the properties of such interaction in detail, and to understand the difference in mechanism between artificial and natural AMPs, we apply DOPC/DOPG (3:1) membranes as a bacterial cell membrane system to investigate the physical factors participating in the interaction. Peptides adopted are GW-H1 and GW-Q4 (artificial); melittin and pleurocidin (natural). Both the lamellae and liposomes were used as (apparatus) platforms for membrane. The biophysical techniques applied include the followings. (i) Oriented circular dichroism (OCD). This is used to monitor the peptide orientation: parallel means surface adsorbed, whereas perpendicular means pore wall attached or membrane integrated. (ii) Lamellar X-ray diffraction (LXD): used to measure the change in thickness of membrane bilayer in solid state. (iii) Small angle X-ray scattering (SAXS): used to measure the change in thickness of membrane bilayer of small unilamellar vesicles (SUVs) in solution. The physical measurements are conducted during experiments via observing the peptide orientations, the change in membrane thickness and the change in size of liposomes, for which all as an individual function of peptide-to-lipid molar ratio (P/L). The results show that artificial antimicrobial peptide GW-H1 and GW-Q4 behave in a different manner from the natural antimicrobial peptides melittin and pleurocidin. It is indicated that GW-H1 and GW-Q4 adsorbed onto the biomembrane surface continuously and in parallel, instead of attaching perpendicularly in membrane per se. Therefore the membrane becomes thinner and thinner, even without digging a pore. This coincides with the data on particle size measurement from DLS (Dynamic Light Scattering), suggesting the liposome membrane structure has not been seriously interrupted, damaged or deformed. However, the calcein leakage experiments strongly suggested the exchange of materials through membrane. How can we explain this? According to the literatures, noting that the physical properties of lipid membranes, as membranes lipids are getting closer to each other, they will be influenced by the thermal fluctuation force and moving aparts, a condition that may cause the transient pores to occur on the membrane surfaces. We speculate that the synthetic antibacterial peptide GW-H1 and GW-Q4 should be accumulating on the membranes and force the membrane structure to become more fragile, as the surface tension will be increased during the membranes are thinnened. This will probably cause the transient pores to occur by a higher frequency or to a larger extent in size, which is in a similar way as being influenced by thermal fluctuations. This condition may result in the temporary loss of barrier functions of the biomembrane. In contrast, the natural peptide melittin apparently inserts itself into the membrane as described for the toroidal-pore model. Our results provide clear evidence for such model and working hypothesis, according to the observation for peptide distributions both in parallel and in perpendicular, as well as the change in membrane thickness from both SAXS and LXD data. Besides, the critical concentration for pore-formation (P/L)* of melittin in DOPC/DOPG (3:1) is ~1/200, which is only half value compared with that in pure DOPC (~1/99). This finding is in line with the previous concept that the initial steps of cationic AMPs when binding onto the microbial membrane surface are essentially relying on the electrostatic interactions. In our case, the addition of DOPG into DOPC mixture indeed brings more negative charges to a neutral system. The effect of pleurocidin on membrane is similar to melittin. The pleurocidin may cause the penetration on the microbial membrane by forming pores in toroidal-pore model. This is consistent with the comprehension from electrophysiological data described in literature. Moreover, our preliminary data show that pleurocidin may interact with membranes in an even stronger manner, for which severe disruption of membrane vesicles were observed during experiments, and unexpected huge error bars had to be taken care of. In the present study, we applied the most advanced synchrotron technology to a local tradition of research, which has been since long excellent but then difficult to explore more in a decent way. That is, an extensive scientific focus on the interactions between toxins/peptides and biomembranes. Biophysical measurements of objective parameters enable the detailed study for the aforementioned topic to be approached in a rigid way. Deduction of the high resolution data from operation of synchrotron light sources indeed sheds lights into this new platform of research. Our studies may even provide a standard procedure in methods and methodology for related membrane-peptide research. However, it is still difficult and complicated to comprehend the derivations directly from the structural determinants in designing artificial peptides to the mechanism categorization. This will rely on further approaches in the near future.