| 摘要: | 蛋白質的磷酸化是細胞訊息傳遞網路中相當普遍的轉譯後修飾，不正常的磷酸化表現會使訊息傳遞異常，並且通常與人類疾病的發生或進展有關。磷酸?純化及高性能質譜技術的進步已被應用於大規模細胞與臨床樣本的磷酸化蛋白質體分析。然而，目前在微量磷酸化蛋白質體學技術，大都還需要數百微克的蛋白質樣品，限制其分析臨床、稀少微量樣本(小於200微克蛋白質起始量，例如穿刺)的應用，開發具分析微量樣品能力之高靈敏度磷酸蛋白質體學方法為重要議題。近年來，微流體裝置已被用於蛋白質體樣品製備，但其在磷酸化蛋白質體學上的應用仍有相當大的挑戰及有待開發。因此具備的。
為了達到靈敏、簡易的散彈式磷酸化蛋白質體學流程，本篇論文中我們針對微量樣本開發了一個簡化的樣品製備流程，以1~200微克的肺癌PC9細胞株作為模型。首先，評估相容於胰蛋白?裂解與層析質譜的介面活性劑脫氧膽酸鈉(SDC)的優化條件，從而省略蛋白質沉澱和水解後脫鹽步驟。在不除鹽的情況下，使用 1% 脫氧膽酸鈉 (SDC) 作為裂解緩衝液和 加入50% 異丙醇 (IPA) 在磷酸?純化緩衝液中導致磷酸胜?減少 1.5 倍。通過甲酸酸化濃度從 6%增加到 50%，專一性從 46% 增加到 84%，磷酸?數目從 1725 增加到 2255。此外，乳酸助於防止非特異性結合，將純化磷酸?的特異性從 84% 提高到 92%。 最後，我們使用 0.5% TFA 代替 50% FA 以減少樣品量。磷酸?數目和專一性結果相似（0.5% TFA鑑定到 2071條，專一性 87%，50% FA鑑定到 2139條，專一性 88%）。其次，用蛋白質 LoBind 管 (500μL) 替換 StageTip (20μL) 可減少蛋白質吸附，並通過減少純化?使用的二氧化鈦 (TiO2) 的量來防止非專一性結合，磷酸?數目提高了 1.5 倍 ( 559 至 889 條磷酸?)，專一性從 72% 提高到 87%。此外，蛋白質與胰蛋白?的比例從 1:50 優化到 1:10，以將消化效率從 16 小時加速到 2 小時。最後，結合數據非依賴性採集(DIA)方法，評估此樣品製備方法的靈敏度。我們從1-10?g細胞中鑑定出 1,093-4,478 個磷酸?和 1,306-6,560 個磷酸位點，而library-based DIA 方法將鑑定覆蓋率提高了 5.3 倍（5,804-14,968磷酸?和 1,441-6,560)。總結來說，本篇研究成功開發了一種簡化、快速和穩定的磷酸蛋白質體學分析策略。與最近發表的從 1μg 裂解物中鑑定 600 種磷酸?（1011 個磷酸位點）的最佳鑑定結果相比。我們的結果顯示出優越的鑑定結果。此方法可應用於臨床或稀有樣品的磷酸蛋白質體學分析，研究異常磷酸化與疾病潛在機制的相關性。
;Protein phosphorylation is the most ubiquitous post-translational modification (PTM) for regulation of cellular signaling networks. Aberrant protein phosphorylation results in abnormal signaling cascades and is often associated with initiation or progression of human diseases. Advancements in phosphopeptide enrichment and high-performance mass spectrometry-based shotgun have demonstrated large-scale phosphoproteomics analysis for various applications from cell line to clinical specimens. However, current sample preparation workflow in the phosphoproteomics analysis often requires samples at a hundred-microgram level of protein, which limits its implementation for micro-scale sample, i.e. < 200?g protein starting amount (i.e. < 200?g protein starting amount [1]) to study clinical or rare samples such as biopsy. Recently, microfluidic platforms have been developed for low-loss proteome sample preparation, while the application of phosphoproteomics still remains an enormous challenge. Therefore, a highly sensitive sample processing pipeline of phosphoproteomics is an urgent need and is worth development.
To achieve a sensitive and simplified shotgun phosphoproteomic workflow, in this thesis, we developed a streamlined sample preparation workflow for microscale sample (microgram). The PC9 cell line from 1-200?g was used as a model in this study. Firstly, optimized conditions for the detergent sodium deoxycholate (SDC) were evaluated to be compatible with tryptic digestion and LC-MS/MS, which allows bypassing the protein precipitation and post-digestion desalting steps. Without desalting, the use of 1% sodium deoxycholate (SDC) as lysis buffer and 50% isopropanol (IPA) in phosphopeptide enrichment buffer caused 1.5-fold less phosphopeptides. By more acidification concentration from 6% to 50% Formic acid, resulting in the specificity increased from 46% to 84% and phosphopeptides identification increased from 1725 to 2255. The use of lactic acid helped to prevent non-specific binding and improved the specificity from 84% to 92% with phosphopeptide identification similar to control (1932 for control and 1923 for adding lactic acid). In the end, we use 0.5% TFA to replace 50% FA for reducing sample volume. The phosphopeptide ID and specificity was similar between two acids (2071, 87% for 0.5% TFA and 2139, 88% for 50% FA). Secondarily, minimizing protein absorption by replacing the the StageTip (20?L) with Protein LoBind Tube (500?L) and prevent non-specific binding was achieved by reducing the amount of titania oxide (TiO2) beads from 1.5mg to 0.625mg for 7.5?g peptide, which enhanced specificity from 72% to 87% and 1.5-fold identification of 559 to 889 phosphopeptides. Moreover, the protein-to-trypsin ratio was optimized from 1:50 to 1:10 to accelerate the tryptic digestion efficiency from 16 hours to 2 hours. The total time spent of the entire workflow was successfully reduced to 4hrs. Finally, the sensitivity of this majorized workflow was examined by using a series dilution sample with data dependent acquisition (DDA) and data independent acquisition (DIA). For phosphopeptide identification, we identified 1,093-4,478 phosphopeptides and 1,306-6,560 phosphosites from 1-10 ?g lysate, while the library-based DIA approach improved the identification coverage up to 5.3-fold (5,804-14,968 phosphopeptides and 1,441-6,474 phosphosites. In summary, this study successfully developed a streamlined and rapid and robust strategy for minimalistic phosphoproteomics analysis. Compared with recently published best identification of 600 phosphopeptides (1011 phosphosites) from 1?g lysate. Our results showed superior identification results. This pipeline can be applied to clinical or rare samples for phosphoproteomic profiling, which may provide a better understanding on the correlation of the aberrant phosphorylation and disease’s underlying mechanism. |
