大腸桿菌第一型DNA聚合酶對C:C配對錯誤之活體內校對

Abstract

DNA的複製被嚴格的調控，以確保遺傳訊息具有高度保真性。在細胞中，可以透過鹼基選擇(base selection)、聚合酶3’ 端往5’ 端外切酶的校對活性（proofreading）及核酸錯配修復（mismatch repair），將DNA錯誤率降至10-10。本實驗室先前透過大腸桿菌的細胞萃取物，發現第一型DNA聚合酶能夠校對十二種位於引子末端倒數第二個位置的錯誤配對。後續利用MALDI-TOF MS建立了第一型DNA聚合酶的校對活性分析方法，顯示位於引子末端倒數1至4個錯誤配對能被有效校對。以上的發現是基於試管中進行的研究，而細菌體中第一型DNA聚合酶是否具有相似的校對傾向，目前仍不清楚。本研究旨在建立大腸桿菌活體內第一型DNA聚合酶校對活性分析模型。我們使用phagemid製備含C:C錯誤配對的受質，錯誤配對分別位在引子股距離去磷酸根之nick處3’端1至7個核苷酸，依序命名為NCC1至NCC7，且在上游距離16個核苷酸處帶有C:C標誌。將受質轉型到大腸桿菌NM522(WT)中，挑取至少50個菌落抽取DNA後再透過限制酶水解分析校對活性。在NCC5結果中發現背景值上升的現象，推測與細菌的SOS反應有關，後續發現重新合成nick 5’端具磷酸根的受質能夠有效降低背景值，命名NCC1-P至NCC7-P。結果顯示NCC1-P至NCC7-P在NM522中校對活性分別為74.4 %、68 %、69.3 %、0 %、7.3 %、4.7 %及1.3 %。因此距離3’端超過3個核苷酸時，校對活性會顯著下降。此外，NCC3-P的受質轉型至大腸桿菌KA796(polA+)及KA796 D424A (polA exo-)中，校對比例分別為65.3 %及25.3 %。接下來，設計引子股距離nick處同時有2個C:C錯配的受質，命名為NCC47-P、NCC48-P、NCC37-P以及NCC38-P，以了解校對活性的範圍。實驗結果為NCC47上兩個C:C錯誤配對都可被同時校對，比例為18 %；NCC48-P共同校對的比例為14.7 %，只校對倒數第四個錯配的比例為14.7 %；NCC37-P共同校對的比例為20.7 %，只校對倒數第三個錯配的比例為22.7 %；NCC38-P共同校對的比例為8 %，只校對倒數第三個錯配的比例為46 %。本研究探討第一型DNA聚合酶於活體內的校對活性，驗證先前試管中模型之結果，並進一步確認其在細菌體內的實際應用效果，為DNA修復機制提供更全面的見解。

DNA replication is strictly regulated to ensure high fidelity of genome stability. Several mechanisms can reduce DNA replication error rates down to 10-10 in cells, such as base selection, 3' to 5' exonuclease proofreading activity of DNA polymerases, and mismatch repair. Our laboratory’s previous research demonstrated that Escherichia coli DNA polymerase I can proofread twelve mismatches at the penultimate position of primer 3’ terminus by using purified E. coli DNA polymerase I. Subsequently, a MALDI-TOF MS-based assay was developed to analyze the proofreading activity of DNA polymerase I, and the result showed that DNA polymerase I can efficiently proofread mismatches located at the 1 to 4 positions from the primer 3’ terminus. However, these findings were based on in vitro studies, and whether DNA polymerase I exhibits similar proofreading activities in vivo remains unclear. This study aims to establish an in vivo model for analyzing the proofreading activity of DNA polymerase I in E. coli. Since C:C mismatch cannot be processed by any known system in vivo, this study utilizes phagemid substrates containing C:C mismatches at various positions (1 to 7 nucleotides from the 3' end of a dephosphated nick), named NCC1 to NCC7, with additional C:C mismatch as strand markers located at 16 nucleotides upstream to analyze the proofreading activity of pol I. The substrates were transformed into E. coli NM522 (wildtype), and at least 50 colonies were picked for DNA extraction, and C:C mismatch correction activity was scored by restriction enzyme digestion. An increase in background levels of strand loss was observed in NCC5, presumably attributed to the bacterial SOS response. Later, we found that substrates with a 5' phosphate at the nick, NCC1-P to NCC7-P, could effectively reduce background levels of strand loss. The results showed proofreading activities for NCC1-P to NCC7-P in NM522 were 74.4 %, 68 %, 69.3 %, 0 %, 7.3 %, 4.7 %, and 1.3 % respectively, revealing that the C:C mismatches at 1, 2, and 3-nt to the 3’ primer end were well proofread. Additionally, to confirm whether the observed proofreading activity was attributable to DNA polymerase I, NCC3-P was transformed into E. coli KA796 (polA+) and KA796D424A (polA exo-), resulting in proofreading rates of 65.3 % and 25.3 % respectively. The result implied that over half of the C:C proofread observed in vivo was processed by DNA polymerase I. To further explore the range of proofreading activity, we designed substrates with two C:C mismatches at varying distances from the nick (NCC47-P, NCC48-P, NCC37-P, and NCC38-P). The results indicated that NCC47-P could be simultaneously proofread both mismatches at a level of 18%; NCC48-P showed a co-proofreading level of 14.7 %, and 14.7 % for proofreading of the first mismatch only; NCC37-P showed a co-proofreading level of 20.7 %, and 22.7 % for proofreading of the first mismatch only; NCC38-P showed a co-proofreading level of 8 %, and 46 % for proofreading of the first mismatch only. Given these findings, this study suggested a backtracking mechanism for internal mismatch proofreading. In conclusion, this study established an in vivo assay for measuring DNA polymerase I proofreading activity and provided a biological significance for previous in vitro proofreading studies.