開發硫基亞硝基化定量蛋白質體學之技術及其應用

Abstract

硫基亞硝基化是藉由一氧化氮專一鍵結於半胱胺酸硫基上的一種可逆性後轉譯修飾，其可調節人類疾病中之許多相關訊息傳遞。相較於其它後轉譯修飾，因硫基亞硝基化含量極低與其不穩定性，使得於蛋白體中亞硝基化位置的鑑定與含量變化的定量極為困難，且方法仍受限制。在此研究中，我們發展一以質譜技術為核心的亞硝基化蛋白體策略，可以鑑定細胞或組織中蛋白質被亞硝基化的修飾位點及其定量分析。 於論文第一部分，我們發展一不可逆烷化生物素標定策略結合質譜技術，可直接鑑定蛋白質上之硫基亞硝基化半胱胺酸位置。利用無內生性一氧化氮合成酶的COS-7細胞株，藉由不可逆烷化生物素標定策略，我們成功改善自由硫基包覆效率，以降低因雙硫鍵置換所造成的偽陽性訊號，及增加蛋白質酵素水解與亞硝基化胜肽的純化效率，提升以質譜技術對胜肽鑑定的專一性及靈敏度。 於論文第二部分，我們開發一資訊輔助免標定定量蛋白體技術結合不可逆烷化生物素標定策略，針對人類癌症組織中的蛋白質被硫基亞硝基化修飾的特定位點進行定量分析。於三次獨立的純化亞硝基化胜肽實驗結果得知，此策略針對CC-M1人類大腸直腸癌細胞株提供了準確性(4%錯誤率)與定量再現性(33%相對標準偏差)。於人類大腸直腸癌組織中，我們分析了11位不同期別的大腸直腸癌腫瘤與鄰近正常組織之個人化亞硝基化蛋白體，並構築了亞硝基化與蛋白體之間的網絡藍圖，利用資訊輔助胜肽比對策略，我們可於11位不同時期的大腸直腸癌病人組織中成功定量到94個亞硝基化蛋白(含199條胜肽與174個亞硝基化特定位置)，相較於腫瘤附近之正常組織，在6位以上的病人腫瘤組織中分別有20與80條亞硝基化胜肽具2倍與0.5倍程度變化，其中包含已知的thioredoxin (Cys73)與peroxiredoxin (Cys148)，以及潛在性新穎標定癌症蛋白Annexin A4 (Cys108)皆在6位以上的病人腫瘤組織中大量表現，並以西方墨點法進一步驗證之。這些潛在標的蛋白皆與癌症與發炎反應有關，且可能透過蛋白質間的交互作用而受誘導型一氧化氮合成酶或thioredoxin調控而被亞硝基化，透過硫基亞硝基化的定量分析和其所造成的生理病理現象之關聯，可幫助我們進一步了解大腸直腸癌之癌症發生，並找尋潛在性治療標的以提升臨床標的用藥的可能性。 於論文第三部分，我們進一步由生物資訊分析角度探討硫基亞硝基化之特性。利用Motif-X及SNOsite軟體比對一氧化氮刺激與癌症內生性亞硝基化胜肽，預測硫基亞硝基化在一級序列和二級結構上具高頻率分佈的結構偏好，結果顯示，硫基亞硝基化半胱胺酸可能位在酸鹼廣泛分佈之序列與α-螺旋上，且疏水性CxV/I/L motif則位在β-褶疊片段上。此外，於大腸直腸癌組織的分析結果中，有18個潛在內生性蛋白具有AxC motif，此結果與先前研究一致，推測其可能為thioredoxin經由trans-nitrosylation作用的標的蛋白。綜合以上結果得知，生物資訊分析可提供我們了解硫基亞硝基化由一級序列至二級結構上的分佈與特性，解開潛在性分子標的之特性。

S-nitrosylation, a reversible post-translational modification (PTM) involving the covalent interaction of nitric oxide (NO) with the thiol group of cysteine residues, plays important role to mediate NO-based signaling in human diseases. Compared to other PTM analysis, identification of the S-nitrosylation site and quantification of its change on the proteome scale is challenging due to its low abundance and labile nature. In this thesis, we developed a mass spectrometry-based S-nitrosoproteomic strategy for site-specific identification and quantification of S-nitrosoproteome in vitro and in vivo. In the first part of thesis, we developed a S-alkylating labeling strategy using the irreversible biotinylation on S-nitrosocysteines for site-specific identification of the S-nitrosoproteome by LC-MS/MS. Using COS-7 cells without endogenous nitric oxide synthase, we demonstrated that the S-alkylating labeling strategy substantially improved the blocking efficiency of free cysteines, minimized the false-positive identification caused by disulfide interchange, and increased the digestion and enrichment efficiency for improved peptide identification using MS analyses. In the second part of thesis, we developed an integrated quantitative proteomic approach combing our S-alkylating biotin switch method and informatics-assisted label-free approach to site-specifically quantify the degree of endogenous S-nitrosylation on human cancerous tissues. This strategy provided accurate (4% error) and reproducible (33% relative S.D.) quantification based on three independently purified S-nitrosylated proteins from the CC-M1 colorectal cancer cell line. Using colorectal cancer (CRC) as a model, we constructed the personalized S-nitrosoproteomic atlas of paired tumor and adjacent normal tissues from 11 patients with different stages of CRC. Using informatics-assisted peptide alignment strategy, 199 S-nitrosylated peptides corresponding to 174 unique S-nitrosylation sites from 94 proteins were quantified in 11 human colorectal cancer patients from Dukes’ A to D stages. Compared to each adjacent normal tissue, 20 and 80 of these quantified S-nitrosylated peptides were up-regulated (≥ 2-fold) and down-regulated (≤ 0.5-fold) with high occurrence in more than 6 colorectal cancer patients, respectively. Some examples, including well-known S-nitrosylated targets, i.e. Cys73 on thioredoxin and Cys148 on peroxiredoxin-4, and potential novel S-nitrosylation site on Cys108 of Annexin A4, were up-regulated and further validated by western blot analysis. These potential targets were involved in cancer and inflammatory response and may be induced by iNOS- or thioredoxin-mediated nitrosylation through protein-protein interaction. In the third part of thesis, we further applied Motif-X algorithm and SNOsite bioinformatics analyses to explore the potential consensus S-nitrosylation motifs. The results revealed motifs with acid-basic amino acids neighboring the central S-nitrosylated cysteine and prefer to locate on α-helix, contract to the hydrophobic motif CxV/I/L located on β-strand. 18 endogenous S-nitrosylated targets in colorectal cancer patients contained the AxC motif, which was consistent with previous literature and suggested as the potential substrate of thioredoxin via trans-nitrosylation. Taken together, these information analysis provides us to clarify the characteristics and distribution of S-nitrosylation from primary sequence to secondary structure. In conclusion, unraveling the property of potential molecular targets and degree of S-nitrosylation-based pathophysiology will allow the better understanding pro-tumorigenesis of colorectal cancer for S-nitrosylation and yield potential therapeutic targets to improve clinical outcomes.