人類冠狀病毒229E的宿主識別與抗原漂變之分子基礎

Abstract

新冠病毒（COVID-19）大流行已經導致全球超過600萬人死亡，促使人們進行大量研究以對抗嚴重急性呼吸綜合症冠狀病毒2型（SARS-CoV-2）。相比之下，已經發現了超過50年且只引起輕微症狀的最早識別的季節性冠狀病毒HCoV-229E在學界受到較少的關注。在本研究中，我們首先探討HCoV-229E之親緣關係。我們選擇了八個病毒株來依時間順序展示基因演化。結果顯示於HCoV-229E棘蛋白（S）的受體結合區域（RBD）和N端區域（NTD）存在許多突變。另外，在N-糖基化序列中有六個突變，表明糖基組成亦存在變異。接著，我們專注於其中兩個病毒株，即P100E 與Seattle，以分別代表“舊病毒株”和“新病毒株”。與P100E棘蛋白序列相比，Seattle病毒株棘蛋白中被觀察到了共72個點突變。通過模擬預測發現這些突變有可能對蛋白質結構產生破壞性影響，此猜想在對Seattle病毒株棘蛋白進行負染色的電子顯微鏡（NSEM）觀察時得到了確認，其中正常折疊的棘蛋白數量大幅減少。通過液相色譜-串聯質譜（LC-MS/MS）進行的N-糖基分析揭示了明顯的糖基組成差異，P100E在N62和N930位置表現出更多的高甘露糖型糖基，而Seattle株則含有四分之一以上的複雜型糖基。我們假設Seattle株棘蛋白的結構完整性受損，導致較多的複雜型醣基存在。為了驗證此假設，我們引入了GlycoSHIELD以定量分析醣基化的三維分子結構，進而提供與結構和演化相關的證據。另一方面，目前已知的229E S PDB結構僅有無法讓受體結合的RBD-all-down狀態。為了瞭解其宿主識別機制，我們選用相對穩定的P100E S進行cryo-EM分析，研究229E S與人類胺肽酶N （hAPN）形成之複合物。根據計算出的庫倫電位圖，該複合物由兩個採用了未曾報導過之RBD-up構型的229E棘蛋白與hAPN同源二聚體結合，後者為細胞表面常見的寡聚化狀態。為了更接近生理狀態，我們再次利用GlycoSHIELD結合所有醣基化分析結果和cryo-EM解析之複合物結構，構建一個完全醣基化的模型。該模型提供了有關HCoV-229E病毒入侵體內細胞的分子機制。本研究建立的分析流程提供了一種有效而全面的方法，用於探究抗原漂變、醣基化和分子結構之間的交互作用，並且能適用於其他蛋白質。此外，它亦可作為一系統性的分析策略，以預測蛋白質表面潛在的廣泛性中和抗體（bnAbs）抗原決定位。

The COVID-19 pandemic has resulted in over six million deaths worldwide, prompting extensive research efforts to combat SARS-CoV-2. In contrast, HCoV-229E, the earliest-identified seasonal coronavirus that has been present for over 50 years and only causes mild symptoms, receiving comparatively less attention from scientific studies. In this study, we initially examined the phylogeny of HCoV-229E and selected eight strains to demonstrate the genetic evolution chronologically. Significant mutations were identified, particularly in the receptor binding domain (RBD) and N-terminal domain (NTD) of the HCoV-229E spike protein (S). Moreover, six mutations were found in the N-glycosylation sequon, indicating variations in glycan profile. We then took P100E strain and Seattle strain to represent the "old strain" and the "new strain". A total of 72 point mutations were observed in the Seattle strain compared to the P100E sequence. In-silico ∆∆G prediction suggested destabilizing effects of the mutations, which were later confirmed by negative staining EM (NSEM) of the Seattle strain, where the number of well-folded S proteins decreased. N-glycan analysis via LC-MS/MS revealed marked differences in the glycan composition, with P100E showing a higher abundance of high-mannose type glycans at N62 and N930, while Seattle exhibited over ¼ of the complex glycans. We hypothesized that the structural integrity of the Seattle strain was compromised, leading to the presence of further processed glycans. To investigate this hypothesis, GlycoSHIELD was used to quantitatively analyze the glycosylated 3D molecular architectures, providing structurally and evolutionarily relevant results. In addition, all the current 229E S PDB structures only display the RBD-all-down state, which prevents the receptor binding. To uncover the molecular basis of its host recognition, we conducted cryo-electron microscopy (cryo-EM) analysis of P100E S in complex with human aminopeptidase N (hAPN), revealing a structure consisting of two 229E spikes adopting an unprecedented RBD-up conformation, bound to a hAPN homodimer. To emulate the physiological state, we utilized GlycoSHIELD again to combine all glycosylation profiles and the cryo-EM-solved complex structure, constructing a fully glycosylated model anchored on a lipid membrane. This model provides insights into HCoV-229E virus entry in vivo. The established workflow serves as an efficient and comprehensive approach to investigate the interplay between antigenic drift, glycosylation, and molecular structure, with potential application to other proteins. It also provides a systematic way to probe the possible targets for broadly neutralizing antibodies (bnAbs).