乙酰辅酶A合成酶的金属中心结构、功能及其催化分子机理研究

【作者】 刘奕；

【导师】 谭相石；

【作者基本信息】 复旦大学 ， 化学生物学， 2012， 博士

【摘要】 乙酰辅酶A合成酶/一氧化碳脱氢酶(acetyl-CoA synthase/carbon monoxide dehydrogenase,简称ACS/CODH)是一类来源于厌氧微生物的双功能金属蛋白酶。它是一个由两个α亚基和两个β亚基组成的四聚体蛋白,其分子量为310kD。该酶分子内部含有一个长达133A的疏水通道将七个复杂的金属簇(2个A-cluster、2个B-cluster、2个C-cluster和1个D-cluster)连接起来。A-cluster坐落于α亚基,催化乙酰辅酶A的合成反应：CH3-Co3+FeSP+CO+CoA(?)CH3-C(O)-CoA+Co1+FeSP+H+(反应1)B、C和D-cluster(两个p亚基交界面)位于p亚基,其中C-cluster是催化CO和CO2可逆氧化还原反应的金属活性中心：CO2+2e-+2H+←→CO+H20(反应2)含有CODH酶的微生物与植物(光合作用)一起共同维护着地球表面的大气碳循环。拒统计,每年含有CODH酶的微生物所消耗掉的CO多达108吨。乙酰辅酶A是物质和能量代谢的重要物质,在体内能源物质代谢中发挥着枢纽性的作用。一些厌氧微生物如产乙酸菌、产甲烷菌等主要通过乙酰辅酶A合成通路(也称为Wood-Ljungdahl通路),依靠CO/CO2和H2作为碳源和能源合成乙酰辅酶A,进行物质和能量的代谢。拒统计,微生物通过该过程每年可产生多达1011吨的乙醇和109吨的甲烷。由于ACS/CODH酶催化反应的重要性、酶结构的复杂性以及金属中心的多样性,使得该酶近三十年内一直成为生物化学家和无机化学家的研究热点。ACS酶的活性中心A-cluster含有一个铁硫立方烷([Fe4S4] cubane)、个双核镍中心([NipNid])以及一个将两者连接起来的半光氨酸残基。A-cluster进行催化反应需要首先被两个外部电子还原激活至活性状态Ared-act态。过去十年间,针对该酶的结构、功能和催化反应动力学的研究取得了巨大进展,并针对Ared-act态的构型提出了多种模型。但是,关于该金属中心各个亚结构单元的具体功能,以及它们之间的相互作用的机制还很不清楚。因此,在本文第二章和第三章中对这一问题进行了深入而系统的研究。首先,为了探测A-cluster中第509位半光氨酸“桥”的功能,在第二章中我们设计并在厌氧条件下制备了四种单点突变体蛋白(包括C509A、C509V、C509H和C509S)以及一种三点突变体蛋白Abridge (C509A/S511A/H516A)。通过生物化学和生物物理的手段,如紫外可见光谱(UV/Vis)和电子顺磁共振光谱(EPR)等对各种突变体蛋白的金属中心性质进行了表征。利用截流光谱仪(stopped-flow)对蛋白中铁硫立方烷的还原反应以及各蛋白的甲基转移反应进行了详细的动力学研究。我们的实验结果显示C509对于ACS蛋白的催化活性不可缺少,但组氨酸的咪唑基团可以部分替代C509的硫醇基团的功能。我们首次证明了C509在A-cluster中的功能是调节铁硫立方烷和双核镍之间的相互联系,并影响两者的氧化还原电势(redox potential)。另外,前人的研究表明A-cluster中铁硫立方烷的还原反应比ACS的甲基转移反应慢100倍,该结果和人们普遍认为铁硫立方烷是充当电子传递过程的中间体的观点相违背。因此,我们在第三章中探测了A-cluster中铁硫立方烷的功能。我们采用了两种策略来破坏A-cluster中铁硫立方烷的形成：一种方法是将编码形成铁硫立方烷肽段的基因截掉,只保留ACS蛋白碳端的136个氨基酸残基的基因片段,进而表达纯化得到一个片段蛋白α15。第二种方法是将参与铁硫立方烷配位的三个半光氨酸残基(cys506、cys518和cys528)全部突变为丙氨酸,进而得到一个三点突变体蛋白A cubane.我们制备了镍重组的α15蛋白的单晶,解析了它的晶体结构至分辨率2.35A,并测定了其金属Ni中心的K边X-射线扩展精细吸收光谱(EXAFS).上述结果显示该蛋白具有不同于野生型ACS蛋白的双核镍中心。α15和A cubane无甲基转移活性并不呈现出NiFeC"的EPR信号,说明了铁硫立方烷对于ACS蛋白的催化活性是不可缺少的。更有趣地,当我们将Ni重组的α15蛋白与一种含有电子离域环的有机分子苯硫酚混合后,获得了苯硫酚结合的α15蛋白。该蛋白具有完全不同于α15蛋白的紫外可见吸收光谱,并具有了缓慢的甲基转移的活性,说明了苯硫酚可部分替代铁硫立方烷的功能。综合前面的结果,我们提出了ACS进行催化反应的金属中心结构模型为{R-X-Nip}(R为铁硫立方烷,或者苯硫酚等具有离域电子能力的分子；X为cys509的硫醇基团,或者类似的具有“桥”连作用的氨基酸残基,如His)。该金属中心可形成一个超共轭的体系(hyper conjugation system),并且电子的离域作用能稳定Nip处于低价活性状态。我们首次证明了铁硫立方烷在A-cluster中的功能是作为Nip的氧化还原性质的调节体(modulator)而并非作为Nip与外部还原介质的电子传递中间体(electron-transfer redox intermediary)。总之,A-cluster的三个结构单元必须作为一个密不可分的整体发挥催化功能,三者缺一不可。金属蛋白约占整个自然界所有蛋白的二分之一,它们的金属中心是自然界中许多重要生理反应的催化活性位点。金属蛋白由于有金属离子或金属辅基的帮助,具有功能高效性与多样性等特点,特别适合蛋白质工程改造。近些年来,人们尝试利用金属蛋白及其金属活性中心进行设计,以期构建新颖的人造金属蛋白来模拟天然蛋白的结构和功能。在本论文第四章中,我们利用蛋白质设计的思路和蛋白质工程的方法来实现ACS片段蛋白α15向镍超氧化物歧化酶(Ni-SOD)的结构-功能转换。首先,我们通过比较ACS蛋白和Ni-SOD蛋白的金属中心发现：ACS蛋白的Nid位点和还原态(reduced) Ni-SOD的单核Ni中心具有相似的NiⅡN2S2的平面四边形的配位环境。ACS的Nid在整个催化过程中保持NiⅡ的还原非活性状态,而Ni-SOD的Ni中心在催化循环中会在NiⅡ和NiⅢ之间转换。当Ni-SOD处于氧化态时,第1位的组氨酸残基的侧链咪唑基团会与Ni中心配位,而ACS的Nid位点则缺少这样的第5位轴向配体,这很可能是导致两种含镍蛋白功能迥异的原因。在第三章中我们报道了ACS的片段蛋白α15的晶体结构,结果显示它含有完整的Nid金属位点。首先,我们尝试向α15蛋白溶液中加入咪唑分子,但是实验显示加入咪唑前后的αα15蛋白均无SOD活性,说明了游离的咪唑分子不能充当轴向配体。接着,我们以α15蛋白为母体蛋白,设计并制备了两个单点突变体蛋白S594H和F598H,希望在金属Ni中心引入His轴向配体。它们在KO2氧化条件下均呈现出五配位低自旋Ni3+的EPR信号,并且活性实验显示它们都具有了SOD酶活,其活力单位分别为340U和2850U(约为野生型Ni-SOD酶活的1/16),说明突变引入的轴向配体确实发挥了功效。除此以外,我们还通过等温量热滴定(ITC)和竞争性抑制实验测定了α15蛋白、S594H和F598H突变体蛋白Ni的结合能力；利用色氨酸荧光淬灭法和ANS荧光探针法探测了Ni诱导的蛋白的构象变化；利用电化学循环伏安法测定了各蛋白的金属Ni中心的氧化还原电势,数据显示改造蛋白的中点电势均在歧化超氧负离子需满足的电势范围内。接着,为了更好地模拟Ni-SOD金属Ni中心的配位微环境,尤其是Tyr9提供的稳定底物超氧负离子的氢键,我们设计并构建了一系列双点突变体蛋白,包括S594Y/F598H. S594E/F598H和S594H/F598Y。最后,我们利用分子动力学模拟(MD)和密度泛函计算(DFT)的方法模拟了各突变体蛋白的金属中心结构,以及催化循环过程中的各种过渡态结构。综合上述实验结果,我们提出了具有SOD酶活的改造蛋白进行催化所依据的out-sphere机理。目前,我们实验室正在对改造蛋白进行蛋白质晶体学尝试,希望获得改造蛋白的结构从而更好地理解它们的催化机理,以利于进一步的蛋白设计。CODH蛋白在大肠杆菌表达时通常以包涵体沉淀的形式,人们一直未能获得可溶性的CODH重组蛋白。在本文的最后一章,我们尝试通过多种原核表达质粒,对来源于嗜热菌Moorella thermoacetica的CODH的表达与纯化进行了探索,并通过两种策略拿到了可溶性的CODH重组蛋白：(1)ACS和CODH共表达；(2)ACS与CODH结合部分的结构域与CODH共表达。此外,我们还尝试利用共转化表达可参与铁硫立方烷重组相关蛋白的质粒,来促进CODH中复杂金属中心的自组装。该研究结果为将来更进一步地研究CODH的结构与功能及其多个金属活性中心之间的相互关系奠定了基础。更多还原

【Abstract】 Acetyl-coenzyme A synthase (ACS)/carbon monoxide dehydrogenase (CODH) is a bifunctional metalloenzyme found in anaerobic archaea Morrella thermoacetica. It is a α2β2tetramer with a310kD molecular weight, containing seven metal clusters (two A-cluster, two B-cluster, two C-cluster and one D-cluster) connected by a133A molecular tunnel network. The active site A-cluster locates in the a subunit and catalyzes the synthesis of acetyl-coenzyme A (Eq.1): CH3-Co3+FeSP+CO+CoA(?) CH3-C (O)-CoA+Co1+FeSP+H+(Eq.l) The B, D, and C-cluster locate in the β subunit, in which catalyzes the reversible reduction of CO2to CO by the C-cluster (Eq.2): CO2+2e-+2H+(?)CO+H2O (Eq.2)CODH in microorganisms, in combine with plants’photosynthesis, plays a critical role in global carbon cycle. It is estimated that CODH activity accounts for the annual removal-108tons of CO from the environment. As a key substance in material and metabolism, acetyl-coenzyme A plays a central role in energy metabolism in vivo. Anaerobic microorganisms, such as acetogenic and methanogenic bacteria, catalyze the synthesis of acetyl-coenzyme A using CO/CO2as carbon sources and H? as energy sources, respectively. It is estimated that～1011tons of acetate and～109tons of methane are produced globally through this process every year. In the last thirty years, ACS/CODH has become a research focus by biochemical and inorganic scientists, due to complexity and diversity for its metal centers as well as importance for its catalytic reactions.The active site A-cluster, which is located in a the subunit of ACS, consists of an [Fe4S4] cluster coordinated through a bridging cysteinate to a dinuclear nickel sub-component. The A-cluster should be converted to an activated Ared-act state through reduction by two external electrons before every catalytic cycle. In the last ten years, great progress concerning the structure, function and kinetic mechanism of metalloenzyme ACS/CODH has been made, and several state forms of the A-cluster have been identified or proposed to understand the reductive activation of the A-cluster (Ared-act).However, the roles of the sub-components of the active site of the A-cluster and the mechanisms of their interactions remain elusive. Thus, we have carried out intensive and systematic researches for the issue in chapter2and chapter Firstly, to probe the role of bridging C509in A-cluster, four single-substitution mutants (C509A, C509V, C509H and C509S) and one triple-substitution mutant (C509A/S511A/H516A) named△bridge were designed and purified in a glove box under anaerobic atmosphere. The relevant properties of these variants were evaluated in detail by biochemical/biophysical approaches, such as U V/Vis spectra and EPR, etc. The kinetics of the variants’[Fe4S4]2+cubane reduction and methylation were measured by stopped-flow experiments. Our results reveal that the bridging thiolate of C509, which could be partly substituted by imidazole of histidine, is crucial for the methylation activity of ACS. In addition, we proposed that the C509could mediate an interaction between the [Fe4S4] cubane and the [NipNid] sub-site, as well as influence the redox potential of both sub-components.Previously, stopped-flow kinetics of [Fe4S4]+cubane reduction and methylation of the A-cluster monitored by us revealed that the reduction rate of cubane is100-fold slower than that of the methylation. This result contradicts and disproves the mechanistic role for [Fe4S4] as an electron transfer redox intermediary between the [NipNid] and external redox agents. Thus, the mechanistic role of the [Fe4S4] cubane in the A-cluster remains intangible to date. In chapter3, we adopted two strategies to structurally modify the A-cluster by genetics. One strategy is to truncate the gene segments encoded [Fe4S4] cubane, yielding a15kD protein named α15with the C-terminal136residues of ACS. The other is to prepare a triple-substitution mutant called△cubane, in which the three cysteine residues (cys506, cys518and cys528) that coordinate to the iron in the cubane were replaced by alanines, such that the cubane could not form. The crystal structure of Ni-containing α15was crystallized and refined to2.35A, and its Ni-K edge XAFS spectrum were measured as well. The results of these experiments showed that α15has a novel di-nickel center distinct from the one in wild-type ACS. Both α15and△cubane showed no methyl transfer activity and lost the NiFeC EPR signals, indicating that the [Fe4S4] cubane is essential for the catalytic activity of ACS. Next, we treated the Ni-containing α15with benzenethiol, an organic molecule with electron-delocalization ring, to obtain a benzenethiol-bound variant. Unexpectedly, this variant exhibited a slow but detectable level of methyl group transfer activity, indicating that benzenethiol has a "rescue" effect on the methylation activity instead of [Fe4S4] cubane. These results combined with previous studies, have led us to propose that the {R-X-Nip}(R:[Fe4S4], bezenthiol; X:thiolate of Cys509or other bridging residues) might form a "hyper conjugation system", and that the electron delocalization effect could stabilize the activated low valence Nip state. The [Fe4S4] cubane might function as a modulator of redox properties of the Nip, but not as an electron-transfer redox intermediary between the Nip and external redox agents. In summary, the three sub-components of the A-cluster form a co-operative unit to conduct the function of ACS, and each one is indispensible.Metalloproteins account for near half of all proteins in nature, and their metal-binding sites are always responsible for catalyzing numerous important biological processes. Metalloproteins are suitable for modification and functional design due to the efficiency and diversity of their metal cofactors. An ultimate test is to use our knowledge to design new metalloproteins that reproduce the structures and functions of native metalloproteins. In chapter4, we have converted the ACS truncated protein α15into a nickel superoxide dismutase (Ni-SOD) mimic protein through functional design and protein engineering.ACS and Ni-SOD are two known Ni redox metalloenzymes catalyzing distinct biological reactions. Initially, comparing the metal centers in these two enzymes reveal that the distal nickel site (Nid) of ACS and reduced Ni-SOD display similar square-planar NiⅡN2S2coordination environments. The Nid of ACS has been suggested to be redox inactive remaining NiⅡ state, while the Ni-SOD accesses both the NiⅡ and NiⅢ during its catalytic cycles. The lack of axial ligand in Nid of ACS in contrast to His1in oxidant Ni-SOD may cause the huge differences in their physiological functions. In chapter3, we have reported the crystal structure of a truncated ACS protein α15containing only the Nid segments of ACS. However, this variant likewise lacks an axial ligand and shows no SOD activity in either the presence or absence of imidazole. To introduce an axial ligand in Nid site, we firstly prepared two single-substitution mutants (S594H and F598H) based on α15.The SOD activity measurements indicated that both variants obtained obvious SOD activity, corresponding to2850U for F598H and340U for S594H, respectively. Their EPR results in presence of KO2revealed a low-spin five coordination signal of Ni3+, clearly proving that our design of introducing an axial ligand is successful. In addition, the Ni bind affinity of the variants was measured by ITC and glycine competition titration.; the Trp fluorescence emission and ANS fluorescence experiments suggest a Ni-induced structure rearrangements exist in the variants; the cyclic voltammetry results showed that the redox potential of Ni centers in the active variants were all in the ranges of the reaction catalyzed by SOD enzymes. Secondly, to better mimic the secondary coordination sphere of the metal center in Ni-SOD, in particular for H-bond supplied by the conserved residue Tyr9, we designed four double-substitution mutants (S594Y/F598H, S594E/F598H and S594H/F598Y) to introduce a putative H-bond. Furthermore, we used MD and DFT methods to monitor the structures of metal centers in the variants as well as the transition states in the catalytic cycles.In conclusion, we proposed an out-sphere mechanism exists in the active mimic SOD proteins in this study. The crystallographic studies of the mutated proteins are still under way in our lab to better understand their catalytic mechanism, as well as to guide our design in future.There is no efficient method to obtain recombinant CODH by E.coli due to its high insolubility. In the last chapter, we tried several prokaryotic expression plasmids to express and purify recombinant CODH from thermophilic bacteria Moorella thermoacetica. Finally, we have successfully obtained soluble CODH by two methods:(1) co-expression of ACS and CODH;(2) co-expression of the binding domain of ACS to CODH and CODH. In addition, we also tried to co-express another plasmid, which expressed a series of proteins to help assemble the iron-sulfur clusters in CODH. These results lay a solid basis for further studies on the structure-function relationship of CODH as well as the interaction of its metal clusters.更多还原