冷凍電子顯微鏡解析分枝桿菌噬菌體 Maliketh 之結構

Abstract

近年臺灣非結核分枝桿菌 (NTM) 感染呈上升趨勢，其中以具高度抗藥性的膿瘍分枝桿菌 (Mycobacterium abscessus) 為臨床治療上的重大挑戰。現行多重抗生素治療不僅療程冗長且副作用顯著，使噬菌體療法重新受到關注。然而，目前針對感染分枝桿菌之噬菌體結構資訊有限，限制了治療與診斷的開發。本研究選用由中央研究院生物化學研究所 Lowary 實驗室自臺灣花蓮分離之分枝桿菌噬菌體 Maliketh，基因組大小為 41,901 bp，屬 G1 亞群，在臺大醫院的初步實驗中顯示有感染膿瘍分枝桿菌的潛能。

本研究旨在以冷凍電子顯微鏡 (cryo-EM) 解析分枝桿菌噬菌體 Maliketh 之高解析度三維結構，並建立原子級結構模型，以擴充分枝桿菌噬菌體結構資料庫。研究首先利用非致病性的恥垢分枝桿菌 mc2155 菌株 (Mycobacterium smegmatis mc2155) 大量增殖噬菌體，經密度梯度超高速離心進行純化，並製備玻璃狀冰樣本。影像由 Titan Krios 電子顯微鏡進行採集並於 CryoSPARC 平台處理。噬菌體拆分為衣殼 (capsid)、連接體 (connector)、尾管 (tail tube) 與底板 (baseplate) 等組件分別重構冷凍電顯電子密度圖，以克服整體構型彈性造成之解析限制。各區域最終解析度介於約 2.6–3.2 Å。後續使用基於機器學習的軟體 ModelAngelo 建構初始模型，使用 COOT 與 Phenix 軟體精修模型，最終成功建立首個 G 群分枝桿菌噬菌體 Maliketh 之完整原子級結構模型。

本研究所得之高解析 Maliketh 結構模型為噬菌體感染分枝桿菌之宿主辨識機制提供結構線索。透過與結構資料庫比對，本研究鎖定三個具有潛在醣親和性的蛋白質結構域，並成功表達與純化這些結構域以進行功能驗證。利用醣鏈陣列 (glycan array) 分析其與分枝桿菌細胞壁相關醣類之特異性結合能力。結果顯示位於底板的一個蛋白質功能域對常見於分枝桿菌細胞壁的分枝型阿拉伯呋喃糖 (arabinofuranose) 表現親和性，推測該結構域可能參與噬菌體感染過程中的宿主細胞辨識。

本研究拓展了對 G 群分枝桿菌噬菌體之結構理解，並初步分析噬菌體蛋白與宿主細胞壁醣鏈間的交互作用。高解析度的噬菌體結構資訊與模型為後續噬菌體工程改造、治療效能提升與快速診斷偵測技術應用提供重要結構基礎。

In recent years, nontuberculous mycobacterial (NTM) infections have increased in Taiwan, with Mycobacterium abscessus representing a major clinical challenge due to its intrinsic multidrug resistance. Current multidrug antibiotic therapy requires prolonged treatment and are frequently associated with significant adverse effects, renewing interest in bacteriophage therapy as an alternative therapeutic strategy. However, high-resolution structural information on mycobacteriophages remains limited, posing obstacles to the advancement of phage-based therapeutic and diagnostic development. In this study, we investigated Maliketh, a mycobacteriophage isolated in Hualien, Taiwan by the Lowary laboratory at the Institute of Biological Chemistry, Academia Sinica. Maliketh possesses a 41,901-bp genome and is classified as a subcluster G1 phage. Preliminary experiments performed at National Taiwan University Hospital suggest that Maliketh can infect M. abscessus.

The aim of this study was to determine the high-resolution structure of the mycobac-teriophage Maliketh using cryogenic electron microscopy (cryo-EM) and to establish an atomic-level structural model, thereby expanding the structural database of mycobacterio-phages and providing a foundation for downstream applications. Maliketh phages were amplified using the non-pathogenic host strain Mycobacterium smegmatis mc2155 and purified by cesium chloride density-gradient ultracentrifugation. Vitrified specimens were prepared for imaging, and micrographs were acquired on a Titan Krios transmission elec-tron microscope and processed using the CryoSPARC platform. To overcome resolution loss resulting from overall conformational flexibility, the phage was segmented into its principal structural components, including the capsid, connector, tail tube and baseplate, and cryo-EM density maps of each component were reconstructed separately. Final re-constructions reached resolutions of approximately 2.6–3.2 Å. Initial atomic models were generated using the machine-learning-based program ModelAngelo and subsequently re-fined manually in COOT and Phenix. These procedures enabled the construction of the first complete atomic models of a cluster G mycobacteriophage, Maliketh.

The high-resolution structural model of Maliketh obtained in this study provides structural insights into the mechanism of host recognition by mycobacteriophages. Struc-tural comparison against database entries identified three protein domains with potential carbohydrate-binding capability, which were subsequently expressed and purified for functional validation. Glycan array assays were performed to evaluate the specificity of these domains toward mycobacterial cell wall–associated glycans. The results revealed that one domain located within the baseplate exhibits affinity toward branched arabino-furanose structures, a glycan commonly found in the mycobacterial cell wall, suggesting that this domain may participate in recognizing host cell during phage infection.

This study expands the current understanding of the structural features of G-cluster mycobacteriophages and provides preliminary insight into interactions between phage proteins and mycobacterial cell wall glycans. The high-resolution structural information reported here will support future efforts in phage engineering to enhance therapeutic effi-cacy and enable the development of rapid diagnostic tools.