- 學位論文
口蹄疫病毒3A蛋白致內質網改造之機制 與病毒空殼蛋白應用
Mechanism of Foot-and-Mouth Disease Virus 3A Induced-ER Remodeling and Diagnostic Application of Virus-Like Particles
摘要
多數正股RNA病毒入侵細胞後會將宿主胞器進行改造，目的是為了保護病毒RNA與集中資源，營造出對病毒最優化的複製環境。其中包括冠狀病毒、登革熱病毒、腸病毒和口蹄疫病毒。然而，至今人們對於病毒改造胞器的機制多半不甚了解。腸病毒與口蹄疫病毒皆屬於Picornaviridae家族，且有相對大量的研究基礎，值得口蹄疫病毒研究借鏡。已知腸病毒3A蛋白可與COPI因子GBF1作用，可能與改造Golgi或ERGIC（ER-to-Golgi intermediate compartment）相關，改造的胞器稱為replication organelles（ROs）。但文獻指出口蹄疫病毒增殖需要的是COPII因子而非COPI，且改造胞器推測為內質網。因此，本研究主題之一是深度探究口蹄疫病毒3A蛋白與COPII機制，企圖建立一個假說說明口蹄疫病毒是如何將宿主的內質網進行改造。 對應GBF1於COPI機制所扮演的角色，我們大膽假設並證實口蹄疫病毒3A蛋白可與COPII因子Sec12以及Sar1作用。證明的方法為Co-immunoprecipitation assay和Colocalization test。另外，使細胞表現不同長度的3A蛋白並接合螢光蛋白GFP (Green fluorescent protein)或APEX 2後，以螢光顯微鏡或穿透式電子顯微鏡發現3A的第42至92氨基酸（3A全長153氨基酸）就足以將內質網改造，雖然改造功能弱於全長3A。其中「42-59」與「76-92」氨基酸同與Sar1的pocket（第198氨基酸附近）作用。總結，我們的假說為3A與Sar1以及內質網上的Sec12作用，藉此加速Sec12活化Sar1。兩個active Sar1同時與3A上的42-59及72-92氨基酸區域結合，3A貼在內質網膜上藉由這方式將膜外凹成vesicle。該vesicle推測為口蹄疫病毒RO的前身。此機制不需要COPII因子內外coat蛋白參與，顯然與COPII傳統機制迥異。雖然，這model仍需要更精確的其他實驗佐證，但根據我們的實驗結果以及各方的文獻資料，我們認為這model是可能性最高、最完美的詮釋。 除了研究口蹄疫病毒非結構蛋白3A外，我們企圖建立一個無病毒參與的血清檢測試劑，以取代只能於負壓實驗室操作的血清中和試驗（Serum Neutralization test, SNT）。實際構想為以病毒空殼蛋白（Virus-Like Particles, VLPs）為抗原、單源抗體（Monoclonal Antibodies, MAbs）為capture antibody和tracer，建構blocking ELISA (bELISA)。口蹄疫病毒中O血清型別影響全球最為廣泛，包括台灣。其病毒表面上已知有5個中和抗原決定位，當中只有site 1為linear epitope，其它則需要經過蛋白折疊甚至需多個蛋白互相擠壓折疊才能構成。可想而知，常規做法以合成peptide去鑑定單源抗體的結合位並不可行。另外，製備中和決定位突變之病毒在台灣又不恰當。所以，為了從已製備的單源抗體庫存中篩選出site 2 MAb，我們建立了一套以VLP為基準的鑑定平臺。構想為表現wild-type與site 2突變的兩種VLPs，篩選出可與正常VLP作用但無法與site 2 mutated VLP作用的MAb。VLP製備方法參考Polacek等人的論文：表現病毒的P1蛋白與少量的3C酵素，P1經過3C切割後自行組裝成VLP。總之，藉此方法成功篩選出10個site 2 MAbs。 接下來，以同樣方式鑑定出S11B MAb辨識site 3。由於血清中抗體除了5種中和抗體外，還有不具備中和能力的抗體。為了能反映出血清的中和能力，我們原先希望偵測血清中各site的抗體成分。實際作法為以VLP為抗原，搭配不同site MAb為tracer（包括site 1、site 2和site 3），結果以site 3 MAb為tracer所得的值與SN titer有最好的相關性（R2 = 0.8071, n=63）。但同時我們也注意到抗體所佔據體積幾乎涵蓋整個病毒基本重複結構（protomer），代表抗體之間難免有steric effect，這同時意謂著site 3 MAb所測得（阻擋掉）的抗體不只有site 3抗體。以MAb為tracer是不可能精確測得該類抗體的濃度。另一方面，由於VLP容易崩解，我們同時也採用了未切割過的P1蛋白為抗原建構檢測系統。雖然，P1無法呈現site 3，其抗原完整性比VLP差，但最終P1-based bELISA搭配site 1 Q10E MAb為tracer仍有不錯的SN titer相關性（R2 = 0.7680, n=63）。 總之，本論文分成兩大主題。其一，研究口蹄疫病毒非結構蛋白3A如何改造細胞胞器，其機制牽涉到兩個未被發現的作用關係（3A-Sar1和3A-Sec12）。我們相信這機制在口蹄疫病毒RO形成機制中是最為關鍵的環節。另一主題著重在結構蛋白、VLP和抗體。我們建立了一套以VLP為基準的MAb鑑定平臺，以及以VLP或P1為抗原的bELISA檢測系統。希望能在產業和應用端做出貢獻。
並列摘要
Most positive-stranded RNA viruses modify host organelles to protect the viral RNA (vRNA) and concentrate resources. The modified environment optimizes viral replication, such as that of coronavirus, dengue virus, enterovirus, and foot-and-mouth disease virus (FMDV). However, the mechanism of a virus reorganizing cellular endomembranes is barely understood. Enterovirus and FMDV both belong to the Picornaviridae family; therefore, the relatively large body of research on enterovirus can provide a reference for FMDV. Enteroviral 3A protein can interact with the COPI factor GBF1, which is possibly associated with modifying the Golgi and ERGIC (ER-to-Golgi intermediate compartment) into replication organelles (ROs). However, previous research indicates that FMDV replication depends on the COPII factor instead of COPI, and it should be the ER (endoplasmic reticulum) that is modified. Therefore, one of our goals was to profoundly investigate the FMDV 3A protein and the COPII pathway to build a model to explain how FMDV remodels the ER into ROs. As the counterpart of GBF1 in the COPI pathway, we hypothesized that FMDV 3A interacted with the COPII factors Sec12 and Sar1, which was further proved by co-immunoprecipitation assay and colocalization test in 3A expressing cells. In addition, the region of aa 42–92 for 3A (full length: 153 aa), despite showing relatively low performance compared to its full length, is sufficient for modifying the ER into vesicle-like structures, according to fluorescence microscopy and transmission electron microscopy images of cells expressing truncated 3A fusing to GFP (green fluorescent protein) and APEX2, respectively. Furthermore, the regions of aa 42–59 and aa 76–92 can bind to the pocket on Sar1 (around aa 198). In short, we hypothesize that FMDV 3A enhances Sar1 activation by co-interacting with Sar1 and Sec12. Two active Sar1 interact with aa 42–59 and aa 76–92 of 3A, whereby 3A can directly curve the ER membrane into protrusive vesicle-like structures. These vesicles might be the precursors of FMDV RO. This mechanism is independent of COPII inner and outer coat proteins and distinct from the traditional COPII pathway. Although it still requires further experiments to prove, we think the model we raised is the most reasonable and ideal interpretation for our data and previous references. In addition to studying the FMDV non-structural protein 3A, we aimed to develop a diagnostic tool for seroconversion of FMDV to replace the serum neutralization test (SNT), which can only be conducted in a negative pressure laboratory. The format of blocking ELISA was designed with FMD virus-like particles (VLPs) as diagnostic antigens and monoclonal antibodies (MAbs) as capture antibodies and tracers. There are five neutralization sites in serotype O, which is prevalent worldwide, including in Taiwan. Only site 1 is a linear epitope; others are conformational structures that demand protein folding and mutual interaction between subunits. Synthetic peptides are thus unsuitable for mapping and screening most neutralizing antibodies. In addition, the generation of site-mutated artificial FMDV is not allowed in Taiwan. Therefore, we established a platform based on VLP to screen out site MAbs from a well-prepared MAb pool. Based on the study by Polacek and colleagues, we expressed viral P1 protein (or site 2 mutated P1) with a relatively low amount of 3C protease within the cells. Processed P1 (or mutated P1) self-assembled into VLPs (or site 2 mutated VLPs (mVLPs)). Ten site 2 MAbs were screened for mutations that abrogated or inhibited the binding. Next, the site 3 MAb, S11B, was characterized by the same strategy as that of the site 2 MAb. Given that those serum antibodies are composed of five neutralizing antibodies and “non-neutralizing” antibodies, our first goal was to detect each neutralizing antibody by indicating MAbs. Therefore, bELISAs based on VLP paired with the site 1, site 2, and site 3 MAbs were performed. The PI value from VLPs paired with S11B MAb showed the highest correlation to SN titer (R2 = 0.8071, n=63). On the other hand, we found that the site 3 MAb as a tracer would detect (block) not only site 3 antibodies. Because of the large bulk of antibodies compared to repeating unit protomers of virions, it was almost impossible to detect specific antibodies by bELISA with MAb. Although the antigenicity of P1, which failed to present site 3, is not an authentic representation of virions compared to VLPs, it can avoid the problem of dissociation from VLPs. Hence, unprocessed P1 was also used as an antigen for bELISA, and it demonstrated a good performance (R2 = 0.7680, n=63) when paired with the site 1 MAb Q10E. In summary, this thesis is divided into two parts. One is to study how FMDV non-structure 3A protein modifies the organelles. The modification is involved two undiscovered interactions (3A-Sar1 and 3A-Sec12). We believe the mechanism we raised is most critical in RO formation. The other topic is about structure protein, VLP, and antibodies. We established a platform based on VLP to characterize MAb’s binding sites and a detection system based on VLP or P1 for evaluating vaccinated animal serum, hoping we can contribute to industrial and practical areas.