br STAR Methods br Acknowledgments We thank Genevi ve Almouz
Acknowledgments We thank Geneviève Almouzni, Sophie Polo, Ralph Scully, Wael Mansour, and Jeremy Stark for providing cell lines and Wolf Heyer, Haico van Attikum, Martijn Luijsterburg, and Brendan Price for helpful discussions. We thank Ratna Weimer, Bettina Basso, Christel Braun, and Cornelia Schmitt for technical assistance and Anja Waizenegger and Julian Spies for generating the ATRX KO cells and experimental and intellectual input at early stages of the project. Work in the M.L. laboratory is supported by the Deutsche Forschungsgemeinschaft (GRK1657) and the Bundesministerium für Bildung und Forschung (02NUK037C).
Introduction Maintaining fidelity during DNA replication is essential for the survival and propagation of nearly all forms of life. Much of the burden in maintaining genomic fidelity lies on the ability of replicative DNA polymerases to catalyze DNA synthesis with remarkable accuracy . Most DNA polymerases involved in chromosomal replication make one mistake every 106–107 opportunities . This low error rate is very impressive considering that DNA polymerases maintain this remarkable accuracy while performing DNA synthesis with incredibly high catalytic efficiencies approaching diffusion limits of catalysis (108Ms). For example, the high-fidelity bacteriophage T4 DNA polymerase (gp43) incorporates the correct nucleotide, adenosine-2′-deoxyriboside triphosphate (dATP), opposite its correct templating BMS 193885 partner T with an efficiency of ~107Ms. This is achieved through high binding affinity (Kd~10μM) for the correct nucleotide coupled with a fast rate constant (kpol~100s) for incorporation . Historically, both kinetic steps have been attributed to hydrogen-bonding interactions that guide the incorporation of the correct nucleotide oppose its templating partner , , . As expected, inappropriate modifications to the templating base can adversely alter these hydrogen-bonding interactions to subsequently increase the ability for a replicative DNA polymerase to misinsert an incorrect nucleotide. As a consequence, the misreplication of damaged DNA can cause mutagenic events that initiate genetic diseases such as cancer , , . DNA lesions can be classified into three distinct categories based on their physical nature. These categories include bulky lesions such as pyrimidine dimers, miscoding lesions such as 8-oxo-guanine (8-oxo-G), and non-instructional lesions such as abasic sites and double-strand DNA breaks (DSBs) , , . There are several DNA repair pathways that can correct these lesions.  However, under certain conditions, these repair pathways can become overwhelmed, causing a large number of unrepaired lesions to persist. This can lead to an increased opportunity for their inappropriate replication in a process termed translesion DNA synthesis (TLS) , , . Although TLS activity can be error-prone and reduce genomic fidelity, this activity is essential, as most cells would die if unrepaired DNA lesions were not efficiently replicated by specialized DNA polymerases such as pol eta, pol kappa, and pol iota. One commonly formed DNA lesion that can produce devastating effects on cellular function is the abasic site , . Although this DNA lesion lacks hydrogen-bonding information, several DNA polymerases can efficiently by-pass this lesion under in vitro and in vivo conditions , , , . In most instances, dATP is preferentially incorporated opposite this non-instructional lesion, and this unusual phenomenon is termed the “A-rule” of TLS . We previously used the bacteriophage T4 DNA polymerase, gp43, as a model high-fidelity DNA polymerase to understand the molecular forces associated with this preferential incorporation , , , , , . These studies quantified the ability of gp43 to incorporate modified purine analogs and 5-substituted indolyl nucleotides opposite an abasic site. Results from these studies demonstrated that alkylated purine analogs such as N6-methyl-adenosine-2′-deoxyriboside triphosphate (N6-Me-dATP) and O6-methylguanosine-guanosine-2′-deoxyriboside triphosphate (O6-Me-dGTP) were utilized more efficiently than dATP, and this was caused by increases in kpol coupled with decreases in the Kd value for the modified nucleotide . More impressive results were obtained using non-natural indolyl analogs such as 5-nitro-indolyl-2′-deoxyriboside triphosphate (5-NITP) . In this case, analogs possessing increased π-electron surface area were utilized 1000-fold more efficiently than dATP , , , . Based upon these data, we developed the model depicted in Fig. 1a that highlights the importance of nucleobase desolvation toward enhancing the binding affinity of the incoming nucleotide, while increased π-electron density influences the rate of the conformational change step that precedes phosphoryl transfer .