人類第二型拓樸異構酶β亞型之結構分析及其與抗癌藥物之交互作用

Abstract

第二型拓樸異構酶 (type II topoisomerases; Top2s) 能夠在DNA上造成暫時的雙股螺旋斷裂，藉此改變DNA的拓樸結構。因此，此酵素可以解決細胞在各個週期所面臨的DNA拓樸結構問題，尤其在細胞分裂時，Top2s是負責分開交纏的子代基因體的關鍵酵素。許多臨床治療上使用的抗癌藥物，如etoposide (VP-16)、doxorubicin、amsacrine (m-AMSA) 及mitoxantrone 等Top2標的藥物 (Top2-targeting agents)，能夠穩定此酵素在其催化過程中，和DNA短暫形成之切割複合體 (Top2 cleavage complex; Top2-cc)，因而導致癌細胞染色體DNA的斷裂並促使細胞死亡。人類細胞中的兩種Top2亞型酵素：Top2α及2β，均為這些藥物的目標，並由於快速生長的癌細胞高度依賴此類酵素，Top2標的藥物 (Top2-targeting drugs) 已廣泛的被運用在多種癌症的化療當中，並行之有年。然而，這類藥物具有心臟毒性，及會引起白血病，特別是急性骨髓性白血病 (acute myeloid leukemia; AML) 等副作用，使得用藥上仍有諸多疑慮。而新一代藥物的設計與開發，卻長期受限於目前仍缺乏藥物與酵素交互作用的分子細節。 在本研究中，利用X-射線蛋白質晶體學，我們成功解析了人類Top2β-亞型酵素與DNA及抗癌藥物etoposide所形成之切割複合體的晶體結構，此結果清楚地揭示藥物、酵素、DNA三者交互作用的分子結構基礎。其中可見etoposide藥物分子嵌入了由酵素造成的DNA斷裂處，因此能阻止酵素將DNA接回，並穩定整個Top2-DNA切割複合體。此結構也顯示etoposide的嵌入造成了酵素活性中心胺基酸在空間位置上的分離，顯示在此藥物的作用下，Top2-DNA切割複合體是被穩定在一個非活化態的蛋白構型。同時，鑒於已知的Top2標的藥物彼此間化學結構式差異相當大，為了快速檢視其他的藥物與Top2-DNA切割複合體交互作用的模式，我們接下來利用浸潤 (soaking) 將蛋白晶體中的etoposide置換成其他藥物，並藉此成功得到了與doxorubicin、m-AMSA及mitoxantrone結合的人類Top2β-DNA切割複合體。結構解析的結果顯示，m-AMSA及mitoxantrone和etoposide擁有相似的Top2抑制機制，並享有共同的藥物結合區，但對應不同藥物，此結合區中的氨基酸殘基會以不同的構型與這三種藥物各自形成專一並緊密的交互作用。並且，這些交互作用均可合理解釋已知的結構藥物活性之關係 (structural-activity relationship) 及產生藥物抗性的原理，顯示利用浸潤的方式我們可成功得到具生理意義之藥物、Top2及DNA三者形成的切割複合體。另一方面，doxorubicin結合的結構顯示了一個與其他藥物全然不同的結合方式。此結果雖然無法解釋已知的藥物特性，但明確暗示了doxorubicin應具有其獨特的Top2抑制機制。 總和而言，藉由解析不同抗癌藥物與人類Top2切割複合體之高解析度晶體結構，本研究為Top2標的藥物的設計與開發提供了許多重要資訊。尤其，目前已知Top2標的藥物有機會引起白血病的副作用，是由於藥物作用到Top2β-亞型所致。本研究藉由比對Top2α及2β在藥物結合區之胺基酸的異同，也提供了設計亞型專一性藥物的關鍵訊息，有望能大幅降低使用此類藥物時所帶來的副作用。

Type II topoisomerases (Top2s) are essential enzymes responsible for the timely resolution of DNA topological problems by temporally cleaving both strands of a DNA duplex to allow the passage of another. Agents that perturb the Top2-mediated DNA cleavage/rejoining process consist of a group of successful clinically active anticancer drugs, including etoposide, doxorubicin, amsacrine (m-AMSA) and mitoxantrone. Bothe isoforms of human Top2, 2α and 2β, can be targeted by these agents, in which the enzymes are trapped on DNA in the form of covalent enzyme-DNA adduct, termed Top2 cleavage complex (Top2-cc). The drug-induced accumulation of Top2-cc on DNA leads to fragmentation of genomic DNA and cell death. Despite their potent anticancer activities, a wider application of Top2-targeting drugs is hampered by deleterious side effects and the emergence of drug-resistant cells. Among them, the therapy-related leukemia is a thorny side effect particularly raised by Top2-based chemotherapy, and is considered induced by Top2β-targeting of drugs, which introduces double-strand break on regulatory region of genes and leads to genome rearrangement. This calls for an isoform-specific Top2-targeting strategy that may suppress such life-threatening side effect of this sort of drugs. To facilitate development of next generation Top2-targeting agents, it is essential to understand the structural basis of drug action in detail. Therefore, in the present study, the crystal structure of an etoposide-bound human Top2β cleavage complex was determined, which reveals a DNA cleavage site-specific drug insertion and a concomitant decoupling of active site residues, thus explaining how the drug blocks the rejoining of broken DNA ends. In addition, we established a post-crystallization drug replacement procedure to simplify the structural analysis of other Top2-targeting drugs, by which the structures of m-AMSA-, doxorubicin- and mitoxantrone-stabilized hTop2β cleavage complexes were successfully determined. While m-AMSA and mitoxantrone also targeted to DNA cleavage sites as expected, however, doxorubicin bound at an unusual location aside from the typical drug-binding pocket. For those structures derived from drug-replacing procedure, the structures bound by m-AMSA and mitoxantorne nicely explain reported drug-resistant mutation and structural-activity relationships of the two drugs. In contrast, the binding mode of doxorubicin is not consistent with the known properties, but nevertheless implies doxorubicin may adopt a unique inhibiting mechanism different from other Top2-targeting agents. By recognizing the conformational landscapes of the drug-binding pockets and those drug-interacting residues that are different between human Top2α and 2β, we propose the guidelines for design of isoform-specific Top2-targeting agents.