醣胜肽定序之關鍵質譜技術的建立與其有效應用於高精確性醣蛋白質體分析

Abstract

蛋白質醣化修飾就像是細胞表面一層美麗又危險的外衣，聯絡與調控著細胞與細胞間的繁雜作用，例如：細胞辨識、免疫細胞自我調節功能等，但在癌化的細胞中，又可能扮演著癌細胞擴散的推手。因此，瞭解蛋白質上醣化修飾對於單一蛋白質乃至於整個醣蛋白質體，都是相當重要。蛋白質醣化過程是經由一連串醣轉酶分工合作的結果，故其最終產物往往具有很大的異質性，了解醣化位置不僅對於瞭解蛋白質功能上具有重要的意義，對於分析技術也是一大挑戰。 隨著蛋白質體技術與質譜儀的發展，串聯式質譜術是個不可缺的重要工具，其中以四級柱串聯時間飛行式質譜儀 (Quadrupole/Time-of-flight, Q-TOF)以及LTQ-Orbitrap最常見，本篇論文主要目標為藉由這兩種質譜儀建立高精確性定位醣化位置的方法，針對已知蛋白序列，Q-TOF 上的母離子發掘法 (Precursor ion discovery)，提供二次質譜術篩選特定醣結構的功能，並可提供可靠的醣序列與胜肽組合，其中胜肽的資訊來自於胜肽分子含有一個Ｎ-乙醯葡萄糖胺的Y1離子；對於未知序列乃至於整個醣蛋白質的定序，必須回歸到胜肽碎片離子的確認。兩個關鍵技術在本篇論文中提出，此二技術皆屬目標式質譜術，分為兩個步驟：一為在層析時間上找尋醣胜肽出現的時間點，二為使用三次直譜術進行特定胜肽序列的確認。在Q-TOF 系統上，液相層析搭配能量調控質譜術(LC-MSE)可提供快速尋找醣胜肽的方法，進一步利用新一代含有兩個碰撞反應池的Ｑ-TOF系統，可獲得類似三次質譜術的結果，即可得到胜肽碎片離子的訊息；LTQ-Orbitrap可提供真正三次質譜術，但如何找到Y1離子進行三次質譜術，還是須經由有效Y1離子的找尋，進而進行目標式三次質譜術的分析。此二法皆可成功地應用於常用的醣蛋白標準品的序列解析；方法確立後，並成功地應用於解決帶有特殊雙唾液酸的蛇毒蛋白上未知的醣胜肽來源；推演至醣蛋白質體的規模，兩種常見的生物材料血清與癌細胞中帶有大量且關鍵的醣蛋白，經由針對醣質結構特異性所選用的特定方法萃取醣胜肽，個別應用於所建立的兩個質譜術，不僅可用於評估與解析複雜且未知的醣蛋白質體上，亦成功地提供精確定位後這些帶有特殊末端醣結構蛋白質的身分，可提供後續功能上的探討。最後推演到複雜度最高的生物個體，在發育過程中，斑馬魚魚卵帶有特殊氮及氧連結的醣蛋白結構，經由物理方式可將魚卵分為三個部份，並利用質譜術針對其每一部份的醣質體、蛋白質體與醣蛋白質體進行詳細的研究，其中在卵膜以及與受精卵的間質液中發現大量氧連接的醣胜肽片斷，此外氮連接特殊醣結構胜肽片斷也首次被發現出現在卵黃蛋白上。 此篇論文所提出的先進質譜術不僅可應用於單一醣蛋白醣化位置的分析，亦可實際應用於複雜樣品中醣化位置的解析。

Protein glycosylation, likely a sweet but dangerous coat on the cell surface, plays essential roles in cell-cell recognition, communication and adhesion in normal situation. However, aberrant glycosylation may result in many severe effects, such as cancer cell metastasis. Glycosylation process is a successive action of transferases, which cause the heterogeneity and uncertainty of glycan structures that may be attached to any glycosylation site. To resolve site-specific glycosylation pattern on glycoproteins is a considerable analytical challenge but is essential to provide the much needed structural details for functional studies. Experimental approach in which glycans are cleaved off may facilitate identification and sequencing of the deglycoslated peptide but information on the site-specific glycan is lost. With the development of proteomic technologies and mass spectrometers, multi-stage tandem mass spectrometry (MS) becomes an indispensible tool for glycosylation mapping. Two hyphenated mass spectrometers, Quadrupole/Time-of-flight (Q-TOF) and LTQ-Orbitrap, are the most common tools in proteomic field. The major aim of this thesis work was to develop a practical mass spectrometry-based platform for site-specific glycosylation mapping. For the single glycoprotein with known sequnce, Precursor ion discovery (PID) function on Waters Q-TOF system provided not only a simple way for glycopeptides scouting but also a reliable combination of glycan and peptide sequence in which information on peptide was derived from the prominent Y1 ion corresponding to the peptides core carrying an N-acetylglucosamine residue. For unknown proteins or glycoproteome, further sequencing of this peptide core is necessary. Two key techniques developed and presented here are both based on a general workflow of two sequential steps: 1) survey scouting of candidate glycopeptides or Y1 ions on the first survey, and 2) targeted multi-stage mass spectrometry for confirmation of peptide sequence. In Q-TOF system, LC-MSE provides a simple way for first survey of glycopeptides. Subsequently, the identified peptide sequences are validated by targeted pseudo-MS3, which can only be implemented on Q-TOF system equipped with dual-collision cells. The LTQ-Orbitrap mass spectrometer can provide a true MS3 function. However, the most critical consideration for glycopeptide sequencing is ability to correctly identify the right Y1 ion for MS3 analysis. Therefore, the first step of the workflow on LTQ-Orbitrap is to assign the accurate Y1 ion from HCD (higher collision energy dissociation), followed by a targeted MS3 analysis. These two-tier workflows established were first validated by using the single well-known glycoprotein standard, fetuin. Then, the origin of unexpected proteolytic glycopeptides from the purified glycoprotein with O-acetylated disialylated N-glycan from snake venom was confirmed by both methods. For real glycoproteomic studies, the glycopeptides from two biological materials, the serum and a cancer cell line, were enriched according to its respective glycosylation profile and successfully analyzed by each MS method. The protein identities and glycan composition of most glycosylated proteins were confirmed in both cases. Finally, for sample of highest complexity, i.e. an intact fertilized zebrafish eggs, both MS3 approaches were used for more comprehensive glycoproteomic analysis. To better understand the developmental glycobiology of this important model, it is imperative to delineate the spatial-temporal distribution of the different classes of glycans and their respective protein carriers. Large amount of endogenously digested O-glycopeptides were only observed in the chorion/perivitelline compartment but not the yolk or embryos. From the latter fractions, vitellogenin was identified as the protein carrier of the unique complex type N-glycan structures.