Barsky D, Kool E T, Colvin M E
Biology and Biotechnology Research Program, Lawrence Livermore National Lab, Livermore, CA 94550, USA.
J Biomol Struct Dyn. 1999 Jun;16(6):1119-34. doi: 10.1080/07391102.1999.10508321.
Although DNA polymerase fidelity has been mainly ascribed to Watson-Crick hydrogen bonds, two nonpolar isosteres for thymine (T) and adenine (A)--difluorotoluene (F) and benzimidazole (Z) --effectively mimic their natural counterparts in polymerization experiments with pol I (KF exo-) [JC Morales and ET Kool. Nature Struct Biol, 5, 950-954, 1998]. By ab initio quantum chemical gas phase methods (HF/6-31G* and MP2/6-31G**) and a solvent phase method (CPCM-HF/6-31G**), we find that the A-F interaction energy is 1/3 the A-T interaction energy in the gas phase and unstable in the solvent phase. The F-Z and T-Z interactions are very weak and T-Z is quite unstable in the solvent. Electrostatic solvation energy calculations on F, Z and toluene yield that Z is two times, and F and toluene are five times, less hydrophilic than the natural bases. Of the new "base-pairs" (F-Z, T-Z, and F-A), only F-A formed an A-T-like arrangement in unconstrained optimizations. F-Z and T-Z do not freely form planar arrangements, and constrained optimizations show that large amounts of energy are required to make these pairs fit the exact A-T geometry, suggesting that the polymerase does not require all bases to conform to the exact A-T geometry. We discuss a model for polymerase/nucleotide binding energies and investigate the forces and conformational range involved in the polymerase geometrical selection.
尽管DNA聚合酶的保真度主要归因于沃森-克里克氢键,但胸腺嘧啶(T)和腺嘌呤(A)的两种非极性等电子体——二氟甲苯(F)和苯并咪唑(Z)——在与DNA聚合酶I(KF exo-)进行的聚合实验中能有效模拟它们的天然对应物[JC·莫拉莱斯和ET·库尔。《自然结构生物学》,5,950 - 954,1998]。通过从头算量子化学气相方法(HF/6 - 31G和MP2/6 - 31G)以及溶剂相方法(CPCM - HF/6 - 31G*),我们发现气相中A - F相互作用能是A - T相互作用能的1/3,且在溶剂相中不稳定。F - Z和T - Z相互作用非常弱,T - Z在溶剂中相当不稳定。对F、Z和甲苯的静电溶剂化能计算表明,Z的亲水性是天然碱基的二分之一,F和甲苯是天然碱基的五分之一。在新的“碱基对”(F - Z、T - Z和F - A)中,只有F - A在无约束优化中形成了类似A - T的排列。F - Z和T - Z不会自由形成平面排列,约束优化表明,要使这些碱基对符合精确的A - T几何结构需要大量能量,这表明聚合酶并不要求所有碱基都符合精确的A - T几何结构。我们讨论了一个聚合酶/核苷酸结合能模型,并研究了聚合酶几何选择中涉及的力和构象范围。