【作者】 刘新；

【导师】 钱小红； 应万涛；

【作者基本信息】 中国人民解放军军事医学科学院 ， 药物分析学， 2007， 博士

【摘要】 蛋白质组学研究已从静态的定性分析逐步过渡到动态变化的定量研究。为了实现蛋白质组学的定量分析，各种定量方法相继被开发，例如基于双向凝胶电泳的定量技术，基于非标记的定量技术，基于质量标签的定量技术，基于液质联用的稳定同位素标记定量技术等多类定量策略。由于基于液质联用的稳定同位素标记技术在定量的准确度，适用范围等方面都超过了前三类定量方法，因而被广泛采用，其商业化的试剂不断被推出。但是由于商业化的标记方法普遍存在试剂昂贵（iTRAQ），标记位点少（ICAT），只能用指定的质谱仪进行数据采集（依赖性强），分析软件只能用于商业公司自身的质谱仪（通用性差）等缺陷，其应用范围受到限制。本研究目的是开发一整套具有自主知识产权的基于稳定同位素标记和液质联用的蛋白质组学定量新策略，实现高通量，高准确性，高灵敏度的蛋白质组学定量分析，用来克服商业化试剂存在的不足，摆脱对于商业化标记试剂的依赖性。本研究首先建立和优化了一种针对肽段还原性氨基的乙酰化稳定同位素标记方法。该方法经条件优化后具有以下优点：乙酰化稳定同位素试剂用量少，副反应少，能标记样品中全部蛋白质或肽段，标记后的蛋白质或肽段的质量差可以预测，标记方法简单，标记过程反应条件温和，标记基团在多维色谱分离条件中稳定，标记基团在质谱分析中稳定，不会产生额外的碎片离子，标记方法能够应用于包括人体在内的所有生物样本并且试剂廉价易得。但是单纯的乙酰化标记方法标记位点是还原性的氨基端，每个肽段至少标记上一个乙酰基团（即至少掺入3个氘原子），尽管氘原子数量的增加可以避免标记肽段间同位素峰重叠问题，但是随着掺入的氘原子数量的增加而引起的反相色谱中的同位素效应也会增加。为了减少由于同位素效应引起的标记肽段色谱保留时间滞后，从而造成质谱分析的误差，构建了基于纳升级反相色谱分离、在线点靶和MALDI—TOF／TOF—MS质谱仪联用的定量策略。通过标准蛋白质混合物的定量实验验证，表明该策略完全可以实现准确定量分析，具有实际应用价值。在规模化蛋白质鉴定分析中，电喷雾离子源（ESI）质谱和基质辅助激光解析电离源（MALDI）质谱相比，前者使用更为广泛。因此，为了能够将乙酰化标记方法与在线色谱分离和纳升级电喷雾离子源的傅立叶变换离子回旋共振质谱仪联用，同时提高乙酰化标记方法的质谱检测灵敏度，本研究又发展了一种胍基化修饰乙酰化稳定同位素标记方法。该方法除了具有上述乙酰化标记方法的优点之外，由于胍基化修饰的乙酰化的标记方法保证了每个肽段只会被一个乙酰基所标记（即只有3个氘原子的掺入），因此很好的控制了氢氘同位素试剂在反相色谱中的同位素效应，同时提高了标记肽段在质谱中的灵敏度。通过标准蛋白质混合物体系的定量分析实验，表明该策略定量分析准确，可重复性强。在胍基化修饰的乙酰化标记方法建立的基础上，我们根据标记方法和所使用的傅立叶变换离子回旋共振质谱仪数据采集的特点，自主开发了一套自动化数据定量分析软件MSAQ，用于规模化的定量计算。为了进一步评价本研究中建立的基于胍基化修饰的乙酰化稳定同位素标记，液相色谱—傅立叶变换离子回旋共振质谱和自主开发的定量分析软件MSAQ联用的定量蛋白质组策略在生物样本中应用的可行性，我们将该策略在三种不同的复杂生物样本体系中进行了实际应用，获得了满意的结果。首先，我们将该策略应用于大肠杆菌N末端肽组学研究中。在整体蛋白质水平胍基化后的氢代氘代乙酰化标记，既提高了肽段在质谱中检测灵敏度，又实现了对蛋白质N末端肽的特异性同位素标记，因而易于从一级质谱图中识别N末端肽同时进行串联质谱测序分析。77种大肠杆菌蛋白质的N—末端序列被确定，其中46种蛋白质N—末端序列信息和Swiss-Prot数据库完全一致，另外31种蛋白质的甲硫氨酸是否去除的信息在数据库中没有详细注释，而通过我们实验对该31种蛋白质甲硫氨酸是否去除进行了精确测定，另外确证了9种蛋白质的N末端是以信号肽去除的方式存在。实验结果表明，本研究建立的定量蛋白质组策略不仅能对蛋白质的N末端进行规模化的特异性识别和测序，还能对N末端的翻译后修饰（如甲酰甲硫氨酸和信号肽去除问题）进行精确测定。该策略与传统的Edman降解具有灵敏度高、特异性强和通量化的优点，因而在蛋白质N末端测序领域有着更加广泛的应用前景。其次，我们将该策略应用于代谢蛋白质组学研究中。对由四氯化碳（CCl₄）造成的大鼠肝损伤组织CYP450蛋白质表达量进行了定量研究，17种CYP450蛋白质被定量分析，其中2E1的表达量显著下调，该结果验证了由于2E1介导的CCl₄的代谢产生活泼的自由基，从而导致2E1蛋白比之其他CYP450蛋白质更容易受到CCl₄毒理性的影响的论断。同时除了2E1表达量显著变化外，其他几种CYP450蛋白质表达量同时也发生了显著性的变化，说明本研究中所使用的规模化的定量分析手段，不仅可以对于已有研究结果进行确认，同时还可以揭示其他CYP450蛋白质表达量的变化。最后，我们将该策略应用于血浆定量蛋白质组学研究。血浆蛋白质组成与细胞、组织与器官的生理病理状态密切相关。由于血浆蛋白质含量动态范围非常宽(＞10¹²)，因而，如果能够对血浆蛋白质进行高通量化、高准确度和高灵敏度的定量分析，是对本研究中所建立的整套定量分析策略最好的评价。（1）我们对正常血浆样本进行了1：1标记实验，结果显示理论值和实验值的误差小于5％。（2）对处于肝炎不同阶段的血浆样本进行了定量研究，总共对1025种血浆蛋白质进行了定量分析，其中具有2倍以上显著性差异的蛋白质有185种。包括了在血浆中浓度只有20ng／ml的Heparin cofactor 2 precursor，和浓度小于10pg／ml Angiotensinogen precursor，该结果表明本标记方法实现了动态范围达9个数量级的血浆蛋白质定量分析（albumin，35mg／mL-Angiotensinogen precursor，＜10pg／mL）。（3）为了进一步验证定量结果的可靠性，对其中一种差异显著的蛋白质Fibronectin进行了western验证。分析显示western定量结果和质谱定量结果一致。综上所述，本研究建立的基于胍基化乙酰化稳定同位素标记，液相色谱—傅立叶变换离子回旋共振质谱和自主开发的定量分析软件MSAQ联用的定量蛋白质组学新策略，既能够对来自于原核生物的样本进行分析，又能够对来自真核生物的样本进行分析，既能够分析组织样本又能够分析体液样本。在对上述三个复杂生物体系蛋白质表达的定量分析中，从多维分离系统角度研究表明，无论对基于一维SDS—PAGE凝胶的蛋白质分离还是对基于液相色谱系统的蛋白质分离，本研究中所使用的定量策略都具有很好的兼容性；从定量结果角度研究表明，本研究中所使用的定量策略满足了高通量化，高准确度，高灵敏度和高自动化的定量蛋白质组学研究需求。证明该定量策略在蛋白质组学研究中将具有广阔的应用前景!更多还原

【Abstract】 In recent years, the focus of the proteome research moves towards quantitative rather than solely qualitative analysis. In order to realize the proteome quantitative analysis, many methods have been developed, such as the quantitative strategies based on 2-DE, label-free, mass-coded and stable isotopic labeling. Nowadays, taking the quantitative accuracy and feasibility into account, stable isotopic tagging technique coupled with online liquid chromatography （LC） and mass spectrometry （MS） has been adopted widely. While lots of commercial stable isotope reagents have been provided, there are still some limitations in commercial ones, such as expensive reagent （iTRAQ）, few labeling sites （ICAT）, dependence on the designated mass spectrometer, and limited availability of quantitative analysis software.The aim of our research is to develop an integrated analysis platform: stable isotopic labeling, on-line LC with MS and automated software, which achieves the high-throughput, high-accuracy and high-sensitivity quantitative proteome analysis. As a result, the novel strategy successfully avoided the limitations of the commercial reagents and overcomed the dependences on the commercial ones.In our research, a d₀/d₃-acetyl stable isotopic labeling quantitative strategy was established and optimized first of all. The tagging sites are the reductive amino of the peptides. There are several major advantages in d₀/d₃-acetyl labeling method, such as tagging all the proteins or peptides, the predicable mass difference between the labeled peptides, simple labeling procedure, mild reactive conditions, good stability of the labeling group in LC, no extra ions in MS, wide compatibility including human sample and cheap reagent. After labeling, each peptide will be incorporated into one acetyl group at least （no less than 3 deuterium atoms）. Although the incorporation of higher number of deuterium atoms will decrease the overlap between the labeled peptides, the isotopic effect caused by deuterium atom in reverse phase liquid chromatography （RPLC） will be severe. So, a compatible quantitative strategy is provided: d₀/d₃-acetyl stable isotopic labeling, nano-RPLC, online spotting and matrix-assisted laser desorption/ionization time of flight mass spectrometry （MALDI-TOF/TOF-MS）. The quantitative result of the protein mixture demonstrated the feasibility and accuracy of the above strategy in comparative proteome research.In the large-scale proteome research, compared with MALDI MS, electrospray ionization mass spectrometry （ESI-MS） is more comprehensively applied. To make d₀/d₃-acetyl stable isotopic labeling fit for LC-ESI-MS, we exploited another novel quantitative method: guanidination and acetylation stable isotopic labeling （GA labeling）. Besides keeping the merits of acetylation labeling, GA labeling excellently controlled the deuterium atoms isotopic effect. Guanidination blockedε-amino and then makes sure each peptide was incorporated with only one acetyl group （3 deuterium atoms）, which ensured the same retention times of the labeled peptides. The quantitative analysis of protein mixture indicated that the above strategy has good accuracy, good sensitivity and good reproducibility.Base on GA labeling characteristics and high resolution data of the hybrid linear ion trap fourier-transform ion cyclotron resonance （LTQ-FTICR） MS, a quantitative software ’MSAQ’ was developed for automated mass data processing.To evaluate the feasibility of the above integrated analysis platform - GA labeling, FT-MS for data acquisition and a graphic user interface program called ’MSAQ’ for automatic data processing, we applied it on three complex biological samples for proteome analysis. Firstly, the whole strategy was applied on the research of proteins N-terminal peptides. Afterε-amino group of lysine residues of the proteins was blocked by guanidination reaction, the resultant proteins were equally divided into two parts and labeled by d₀-acetic anhydride and d₆-acetic anhydride respectively. At last, the labeled proteins were mixed at 1:1, digested by trypsin and analyzed by MS. Only N-terminal peptide of the protein showed a pair of characteristic isotopic peaks with a mass shift of 3.0188 Da, and other peptides were still singlet isotopic peaks. After N-termini of four model proteins were recognized and sequenced, 77 N-terminal peptides from Escherichia coli proteins were also successfully identified. Among them, 46 N-terminal sequences are consistent with database, 31 N-terminal sequences renewed the database annotation, and 9 N-termini whose signal peptides removed were also identified and sequenced. Compared with Edman degradation, our strategy has three advantages: high sensitivity, high specificity and high throughput. Such a result adequately shows the feasibility of the whole strategy in N-termini special recognition and sequencing.Secondly, the whole strategy was applied on the quantitative analysis of CYP450 proteins during the injury of rat hepatic induced by carbon tetrachloride （CCl₄）. 17 CYP450 proteins were quantified accurately. Among them, 2E1 was down-regulated dramatically, which was coincident with the previous results. Besides, other CYP450 proteins were also changed significantly. The results indicated that GA labeling not only validates the previous results but also disclosed the alterations of other CYP450 proteins, which would explain the response of these metabolic proteins during the liver injury in a full-scale and systematical angle.Thirdly, the whole strategy was applied on the quantitative analysis of human plasma proteins. Human plasma contains thousands of distinct proteins and its concentration dynamic range is over 12 magnitude orders. For example, 22 most abundant human plasma proteins, such as HSA, IgG, transferring,α2-macroglobulin, andα1-antitrypsin, account for 99% of the protein amount, and then less than 1% protein amount covers more than million proteins.Whether or not the whole strategy can realize the high-throughput, high-sensitivity, high accuracy quantitative analysis of human plasma proteins is a key point of its feasibility in proteomics.At first, 1:1 labeled human plasma proteins was analyzed. The error between the expected value and the experiment value was less than 5%, which illuminated the quantitative accuracy. Afterwards, samples from 5 different phases of human hepatitis plasma were compared and quantified. The 1025 plasma proteins were identified and quantified. Among them, 185 plasma proteins varied evidently, including 20ng/ml heparin cofactor 2 precursor （concentration < 20ng/ml） and angiotensinogen precursor （concentration < 10pg/ml）, which proved our method achieves quantitative analysis of 9 magnitude orders. To test the quantitative accuracy, one dramatically different protein, fibronectin, was validated by western-blot. The immunology quantitative result coincided well with the GA labeling result.In summary, we developed the integrated quantitative strategy, which could be used in samples not only from prokaryote and eukaryote, but also from tissue and body fluid. At the same time, our method holds good compatibility either with LC or PAGE separation system. Through the above 3 proteome applications, the feasibility of our technique in high-throughput, high-sensitivity and high-accuracy proteome quantitative research was further verified. So we believe our whole quantitative strategy will own its splendid perspective!更多还原