Wang Suyun, Schaefer Henry F
Center for Computational Chemistry, University of Georgia, Athens, Georgia 30602-2525, USA.
J Chem Phys. 2006 Jan 28;124(4):044303. doi: 10.1063/1.2162538.
The amino group in the nucleic acid bases frequently interacts with other bases or with other molecular systems. Thus any nonplanarity of the amino group may affect the molecular recognition of nucleic acids. Ab initio Hartree-Fock (HF) and second-order Moller-Plesset perturbation (MP2) levels of theory have been used to obtain the equilibrium geometries of the C(l) and C(s) structures for five common nucleic acid bases. The energy barriers between the C(l) and C(s) structures have also been predicted. A series of correlation consistent basis sets up to cc-pCVQZ and aug-cc-pVQZ has been used to systematically study the dependence of the amino group nonplanarity. The equilibrium geometries of the nucleic acid bases with an amino group, including adenine, guanine, and cytosine, are examined carefully. At the MP2 level of theory, larger basis sets decrease the extent of nonplanarity of the amino group, but the decrease slows down when the QZ basis sets are used, demonstrating the intrinsic property of nonplanarity for guanine. For adenine and cytosine the situation is less clear; as the HF limit is approached, these two structures become planar. Addition of core correlation effects or diffuse functions further decreases the degree of nucleic acid base nonplanarity, in comparison to the original cc-pVXZ (X=D, T, and Q) basis sets. The aug-cc-pVXZ basis shows smaller degrees of nonplanarity than the cc-pCVXZ sets. The aug-cc-pVXZ basis is less size dependent than the cc-pVXZ and cc-pCVXZ sets in the prediction of the amino-group-related bond angles and dihedral angles and energy barriers for adenine, guanine, and cytosine. The cc-pCVQZ and aug-cc-pVQZ MP2 results may be regarded as benchmark predictions for the five common bases. The predicted classical barriers to planarization are 0.02 (adenine), 0.74 (guanine), and 0.03(cytosine) kcal mol(-1).
核酸碱基中的氨基常常与其他碱基或其他分子体系相互作用。因此，氨基的任何非平面性都可能影响核酸的分子识别。已使用从头算哈特里 - 福克（HF）和二阶莫勒 - 普列塞特微扰（MP2）理论水平来获得五种常见核酸碱基的C(l)和C(s)结构的平衡几何构型。还预测了C(l)和C(s)结构之间的能垒。使用了一系列直至cc-pCVQZ和aug-cc-pVQZ的相关一致基组来系统研究氨基非平面性的依赖性。仔细研究了含氨基的核酸碱基（包括腺嘌呤、鸟嘌呤和胞嘧啶）的平衡几何构型。在MP2理论水平上，更大的基组会减小氨基的非平面程度，但当使用QZ基组时，这种减小会放缓，这表明了鸟嘌呤非平面性的固有特性。对于腺嘌呤和胞嘧啶，情况不太明确；随着接近HF极限，这两种结构会变得平面化。与原始的cc-pVXZ（X = D、T和Q）基组相比，加入核心相关效应或弥散函数会进一步降低核酸碱基的非平面程度。aug-cc-pVXZ基组显示出比cc-pCVXZ基组更小程度的非平面性。在预测腺嘌呤、鸟嘌呤和胞嘧啶的与氨基相关的键角、二面角和能垒时，aug-cc-pVXZ基组比cc-pVXZ和cc-pCVXZ基组对大小的依赖性更小。cc-pCVQZ和aug-cc-pVQZ的MP2结果可被视为这五种常见碱基的基准预测。预测的平面化经典能垒分别为0.02（腺嘌呤）、0.74（鸟嘌呤）和0.03（胞嘧啶）千卡/摩尔（-1）。
