I. 探討蜘蛛毒蛋白對電壓感應開啟式鉀離子通道的作用 II. 抗蟲活性綠豆防禦素VrD1的結構分析和其作用標靶分子生化鑑定

Abstract

本論文分為二部分: I. 探討蜘蛛毒蛋白對電壓感應開啟式鉀離子通道的作用： 電壓感應開啟式鉀離子通道（voltage-gated potassium channels）普遍存在於各類組織，主要負責在細胞膜興奮時開啟，使之再極化，以調節膜內外陽離子濃度，於膜電位的控制上扮演重要角色。電壓感應開啟式鉀離子通道通常由四個次單元體 （subunits）組成，每個次單元體上含有通道形成區域（pore-forming domain（S1-S6））及四個電壓感受過膜片段（voltage-sensing transmembrane segments），後者分別稱作S1-S4。其中 S4 是最主要的電壓感受體（voltage sensor）；目前已知 S3 的 第二片段（S3b）可與鉀離子通道開啟調整毒素，如蜘蛛毒蛋白（Hanatoxin1, HaTx1）作用，造成通道開啟所需電壓之改變。S3b 可能之二級結構（secondary structure）透過相關研究已有相當程度的預測，在本研究進行前一般相信 S3b 含有一段獨立的 alpha-helix。由於 S3b 之三級結構（tertiary structure）在本論文進行前尚未被解出，其結構-功能關係之研究均靠電生理試驗之推理。本研究意圖透過蜘蛛毒蛋白感性鉀離子通道 Kv2.1（drk1）（屬於 shab family）與抗性鉀離子通道 shaker（屬於 shaker family）分別對已知三級結構的蜘蛛毒蛋白 HaTx1 作分子模擬與嵌合實驗之比較，透過推論蜘蛛毒蛋白與鉀離子通道交互作用之分子模型，進一步了解鉀離子通道的作用機制。 後由於 Roderick Mackinnon 所帶領的研究團隊對於電壓感應開啟式鉀離子通道（voltage-gated potassium channel）之電壓感受體（voltage sensor）的研究有一些異於傳統之突破：他們利用古溫泉細菌 Aeropyrum pernix 之電壓感應開啟式鉀離子通道（KvAP）的蛋白質單晶，以 X 光繞射得到結構及配合一系列與電壓感受體相關的電生理實驗研究，提出不同於傳統 Kv channel 結構之模型。此模型與之前的許多真核生物之電壓感應開啟式鉀離子通道的電生理實驗結果未盡相符，而引起廣泛爭論，本論文亦據此對前述 HaTx1 與鉀離子通道作用的分子模型作了進一步的實驗。 以 Kv2.1 S3b 與 HaTx1 結合時之模擬計算為基礎，透過停止流螢光光譜（stopped-flow）實驗，對 S3b 片段與 HaTx1 的作用，進行了分析探討所得之生化動力學數據，可以作為檢驗先前提出的分子機制及探討 S3b 片段之生理可能構形時之重要依據。配合上先前之模擬計算，我們也對此一作用發生時各分子在細胞膜上空間分佈的關係作了深入的推理。透過 HaTx1 與 Kv2.1 S3b 片段結合的動力學實驗，可以計算出結合發生時之 kon、koff 值。比較了使用 Kv1.1 和 Kv2.1 S3b 變異株作為控制組的結果後，可以清楚得知， Kv2.1 S3b 片段與 HaTx1之結合可能確實依照先前分子模擬預測之機制進行。至於 Kv2.1 S3b 不同變異株（hydrophobic mutant/hydrophilic mutant） 之間的比較，我們則得到稍異於分子模擬預測之結果：疏水性與親水性的作用應同等重要，突變任一類鍵結所需殘基均能使蜘蛛毒蛋白與 Kv2.1 S3b 片段之結合能力消失，而非如之前預測的由極性反應擔負較重要之角色。而 HaTx1 與 Kv2.1 S3b 片段結合的動力學實驗所得的 kon 值遠大於由電生理量測推導出來的數值，這除了反應出可能是系統簡化造成的誤差外，也指出二者可能不是在水溶液環境下反應的事實。 綜合以上生化動力學實驗並配合分子模擬預測之結果，加上先前諸多電生理的資料，我們推測負責與 HaTx1 鍵結之 Kv2.1 S3b 片段殘基應位於細胞膜邊界處，靠近磷脂（phospholipids）之hydrophilic head處。但另一方面，我們的磷脂微脂粒與 HaTx1 一同離心之結果顯示，HaTx1 可以部分深入細胞膜。至此，HaTx1 與 Kv2.1 S3b 交互作用之分子模型已大略清晰，接下來最重要的則是進行實驗量測清楚 HaTx1 深入細胞膜的深度，這樣才能確實釐清這整個機制。在此之前，我們仍認為整個 Kv2.1 S3b 片段應以稍稍傾斜的方式坐落於雙層磷脂膜細胞膜之上緣。由於傳統的結構模型指出了外部縫隙（external crevice）的特性，似乎較符合此處的整體空間關係。 II. 抗蟲活性綠豆防禦素VrD1的結構分析和其作用標靶分子生化鑑定 綠豆防禦素VrD1（Vigna radiata defensin 1）是一種由46個胺基酸所組成的植物防禦素，純化自綠豆品系VC 6089A的種皮，經由種子餵食實驗發現其對綠豆豆象（bruchids; Callosobruchus chinensis and C. maculatus）具有毒殺的效果，為第一個被國人發現除了抗真菌、細菌的活性外，同時兼具抗昆蟲活性的植物防禦素; 然而其造成綠豆豆象死亡之活性來源及作用機制目前並不完全清楚。本研究經由序列及結構比對，分別分析了VrD1與對昆蟲具毒性的蠍毒蛋白（scorpion toxins）及多種具有抗菌活性之植物防禦素的異同，推斷VrD1造成綠豆豆象死亡之活性來源應在於其分子一側與蠍毒蛋白。由於兩者具高度保留性之數個鹼性胺基酸，根據過去的結構與序列比對結果，我們推論VrD1可能是使用類似蠍毒中的 short toxins 之相似機制與細胞膜上之鉀或氯離子通道作用，且我們實驗室的電生理實驗結果也已初步排除了作用在氯離子通道的可能性；而分子另一側之數個鹼性胺基酸，則在與多種具抗菌活性的植物防禦素比較時具高度的序列保留性，故應與抗菌活性有關。 為了進一步探討這兩群序列高度保留性的鹼性胺基酸在VrD1抗蟲、抗菌活性上所扮演的角色，本實驗室過去製備了多株VrD1鹼性胺基酸突變株，並進一步以大腸桿菌表現系統（E. coli）表現、純化出數種VrD1的突變蛋白，用於驗證其在抗蟲活性上所扮演的角色。我們嘗試使用生化方法和電生理分析，來尋找VrD1作用的標靶分子。首先利用SELDI-TOF MS來看昆蟲細胞的膜蛋白是否存在能與VrD1交互作用的標靶分子。此外，我們使用Sf-21細胞進行電生理實驗，用鼠腦切片以及同位素標記的毒素和VrD1做競爭結合實驗來進一步尋找其作用的離子通道類型。我研究結果清楚指出，VrD1顯著抑制幾個不同的陽離子通道，包括：A-type current K+-channels，small-conductance Ca2+-activated K+-channels 及N-type Ca2+-channels，這樣的結果與先前的推斷相符。由於VrD1與其標靶分子的作用模式類似於短鏈蠍毒素，配合VrD1鹼性胺基酸突變株測試其對A-type current K+-channels的抑制效果及分子嵌合計算，應可證明 K7，K24，R26這三個胺基酸在VrD1抗蟲活性上扮演重要角色。

This thesis is divided into two parts： I. Investigation on the Interaction of Spider Gating Modifier Peptides with Voltage-gated Potassium Channels: Voltage-gated potassium channels are found in a wide variety of tissues, where their primary role is to respond to the membrane excitation and allow the repolarization phase of an action potential to occur and result in K+ ions efflux. Such channels are normally homotetramers and each subunit contains four voltage-sensing transmembrane segments, namely S1 through S4, whereas S5 and S6 form the pore. Among them, S4 may play the most crucial role in sensing the voltage changes. In addition, it was generally believed that the C-terminus of S3 (S3b) interacts with gating modifier toxin, like Hanatoxin, and thus has influences on the voltage required for gating. The secondary structural arrangement of S3b has been, due to such studies, intensively analyzed and the existence of an independent alpha-helix was then suggested. Due to the lack of complete and high resolution structure in high resolution for S3b, the study of the structural-functional correlation has been examined only using electrophysiology experiments. Upon binding of Hanatoxin 1, the midpoint of the curve for required gating potential of Kv2.1 can be shifted to the right, which means more difficult to open the channels in the same condition. On the contrary, shaker channels do not show similar effects. We have designed a series of experiments to investigate the functional roles of the vicinity around S3, S4 and S3-S4 linker in affecting the gating behavior. Regarding the stereochemistry, the electrostatic properties, as well as the hydrophobicity, and upon the utilization of the molecular simulation and docking techniques, we have derived the most reasonable orientations and binding positions, from which irrational possibilities were prior to that excluded. Furthermore, with the substitution study with shaker residues, the more precise roles of this area in gating have been analyzed. In 2003, research by Mackinnon’s group has led to certain controversial developments in the voltage-sensing theory of Kv channels. Crystal structure of KvAP channel from archaeum Aeropyrum pernix was revealed, combined with a series of electrophysiological experiments and sequence comparisons, a novel ‘voltage-sensor paddle’ model, which was anticipated to be applied on eukaryotic Kv channels. However, such revolutionary idea did challenge the conventional “translocation” model and contradict to a few previously acknowledged experimental results. Such discovery has also brought gross impact on our proposed mechanism which was based upon the conventional translocation model in Kv channels. To better understand the gating mechanism in Kv channels we employed a combination of MD simulations and biochemical methods. Therefore, we have performed the kinetic analysis with stopped-flow to examine our previously proposed hypothesis. Such binding study, in combination with related calculations, provides further possibility to consider in a more decent way the discussion of the reasonable conformation and membrane distribution of S3b segment in the toxin-Kv channel interactions. The binding rate constant kon and release rate constant koff for interactions between hanatoxin and Kv2.1 S3b segments can be calculated through the kinetic analysis with stopped-flow. Upon utilization of Kv2.1 S3b mutants and Kv1.1 S3b as control experiments, it has been indicated that binding of hanatoxin and Kv2.1 S3b may follow the molecular details which were described in our proposed mechanism. However, the comparison between the hydrophobic and hydrophilic interactions required for binding between hanatoxin and Kv2.1 S3b which was observed from rationally designed mutants (hydrophobic part v.s. polar part of residues) suggests that both types of interactions are equally crucial for binding. Mutation of either part of residues can result in the abolishment of binding ability for hanatoxin and Kv2.1 S3b. Polar interactions should not be the only dominant factor to be able to affect such binding as predicted from simulation study. All together, it is reasonable to comprehend that the S3b residues required for binding with hanatoxin should be located at the upper layer of cell membrane, nearby the hydrophilic heads of phospholipids (or interfacial area of external face) with a slight tilting angle. Therefore the conventional ‘translocation’ model may fulfill such requirement better, especially considering the spatial orientations of transmembrane segments around the external crevice. II. Structural Analysis and Biochemical Identification of the Molecular Targets for the Unique Insecticidal Activity of Mungbean Defensin VrD1: VrD1 (Vigna radiata defensin 1) is a member of the plant defensin family, containing 46 amino acids and four pairs of disulfide bonds. Isolation of a cDNA encoding a small cysteine-rich protein designated VrCRP (also known as VrD1) from a bruchid-resistant mungbean revealed the first discovered plant defensin exhibiting both in vitro and in vivo insecticidal and antifungal activities. However, the molecular and structural basis of this unique insecticidal activity of VrD1 is still not fully understood. Based on the structural and sequence alignment, it is suggested that VrD1, in addition to γ-thionins and several amylase inhibitors, is highly homologous to scorpion toxins, especially the short toxins. We have deduced that VrD1 may utilize a newly found cluster of basic residues on one side of VrD1 molecule to achieve its insecticidal function, whereas another cluster of previously identified basic residues located on the other side of the molecule, which is conserved for all γ-thionins, should be used to achieve the antibacterial/antifugual activities for VrD1 and for all other plant defensins. In order to understand the roles of this newly found cluster of conserved basic residues, we have constructed several expression strains for VrD1 mutants and purified these mutant proteins using E.coli system. Base on sequence and structural alignment, we have postulated that VrD1 may utilized a similar interaction mode as short scorpion toxins to act on insect cell membranes with K+-channel or Cl--channels as molecular targets. Preliminary data has excluded Cl--channels for candidates based on electrophysiological experiments. In this study, we are in attempt to apply the biochemical approach and the electrophysiological analysis to unravel the membrane targets for such activity. We first used SELDI-TOF MS first to check the presence of membrane protein(s) as molecular target(s) for the interaction between VrD1 and the insect cells. Due to its resolution limit, the whole-cell recordings with Sf-21 cells and the competition binding assays with rat brain slices as well as isotope-labeled toxins are further applied to examine the channel types for details. Our results from the whole-cell voltage-clamp experiments demonstrated a very large inhibitory effect of VrD1 on the membrane potentials of Sf-21 cells. In addition, the crossing point of the I-V curves suggests a combination of more than one channels involved in this interaction. The competition binding assays showed that VrD1 significantly inhibits the function of several different cation channels, including A-type current K+-channels, small-conductance Ca2+-activated K+-channels and N-type Ca2+-channels. Such results are in line with the comprehension deduced from the previous structural modeling, for which an interaction mode similar to short scorpion toxins has been anticipated. VrD1 may adopt a very similar mechanism as the short scorpion toxins to inhibit membrane potentials via interaction with mainly potassium channels. Moreover, from the mutational study, the importance of several special basic residues (K7, K24, K26) was inferred by comparing the highly conserved residues between scorpion toxins and VrD1 in the corresponding positions.