DiGabriele A D, Steitz T A
Department of Chemistry, Yale University, New Haven, CT 06511.
J Mol Biol. 1993 Jun 20;231(4):1024-39. doi: 10.1006/jmbi.1993.1349.
The structure of d(CGCGAAAAAACG)/d(CGTTTTTTCGCG) was determined at 2.3 A resolution in order to deduce the local structural features that give rise to DNA bending by adenine tracts. Whereas all previously reported B-DNA dodecamers have crystallized isomorphously (spacegroup P2(1)2(1)2(1) with unit cell dimensions of a = 24.5 A, b = 40.3 A, c = 65.9 A), the duplex reported here crystallizes in a different lattice (spacegroup P2(1)2(1)2 with unit cell dimensions of a = 44.8 A, b = 66.1 A, c = 42.9 A). The DNA exhibits a 30 degree bend in the helix axis that is 180 degrees away from the 20 degree bend exhibited by the adenine tract DNA crystal structures that have been previously determined. This bend is 90 degrees away from the bend predicted for an adenine tract by solution and gel experiments. The adenine tract is straight and bending occurs in the G+C-rich regions. Comparison of the various adenine tract DNA crystal structures reveals that in all cases the adenine tracts have nearly identical structures, even though the overall bends of the helix axes are quite different. This implies that the structure of the adenine tract is robust, at least under the conditions of crystallization. The base-pairs in the adenine tracts exhibit a large propeller twist that leads to the formation of bifurcated hydrogen bonds and a narrow minor groove. In the crystal structure of d(CGCGAAAAAACG)/d(CGTTTTTTCGCG), a minor groove spine of hydration is observed that probably stabilizes the straight structure of the adenine tract. This straight structure of the A-tract is not consistent with the results of fiber diffraction, gel experiments, and NMR studies. Although this may imply that the results of solution experiments need to be reinterpreted, the conditions under which the crystals were grown are different from those under which the solution experiments were done. The possibility remains that the 2-methyl-2,4-pentanediol necessary for crystal growth may facilitate formation of the spine of hydration that stabilizes the straight A-tract, although recent NMR results show the presence of the spine of hydration in aqueous solution. We have also extended our previously reported observation that non-self-complementary DNA structures can exist in the crystal lattice in two orientations. A dodecamer brominated on one strand provides experimental evidence that d(CGCAAAAAAGCG)/d(CGCTTTTTTG CG) is positioned in two orientations in the crystal lattice that are related by a 180 degree rotation around the pseudo-dyad axis of the sequence.(ABSTRACT TRUNCATED AT 400 WORDS)
为了推断由腺嘌呤序列引起DNA弯曲的局部结构特征,测定了d(CGCGAAAAAACG)/d(CGTTTTTTCGCG)的结构,分辨率为2.3埃。尽管所有先前报道的B-DNA十二聚体都同晶型结晶(空间群P2(1)2(1)2(1),晶胞尺寸为a = 24.5埃,b = 40.3埃,c = 65.9埃),但此处报道的双链体在不同的晶格中结晶(空间群P2(1)2(1)2,晶胞尺寸为a = 44.8埃,b = 66.1埃,c = 42.9埃)。该DNA在螺旋轴上呈现30度的弯曲,这与先前确定的腺嘌呤序列DNA晶体结构所呈现的20度弯曲相差180度。这个弯曲与溶液和凝胶实验预测的腺嘌呤序列弯曲相差90度。腺嘌呤序列是直的,弯曲发生在富含G + C的区域。对各种腺嘌呤序列DNA晶体结构的比较表明,在所有情况下,腺嘌呤序列具有几乎相同的结构,尽管螺旋轴的整体弯曲有很大差异。这意味着腺嘌呤序列的结构是稳定的,至少在结晶条件下是这样。腺嘌呤序列中的碱基对呈现出大的螺旋桨扭转,导致形成分叉的氢键和狭窄的小沟。在d(CGCGAAAAAACG)/d(CGTTTTTTCGCG)的晶体结构中,观察到一个小沟水化脊,它可能稳定了腺嘌呤序列的直结构。腺嘌呤序列的这种直结构与纤维衍射、凝胶实验和核磁共振研究的结果不一致。尽管这可能意味着溶液实验的结果需要重新解释,但晶体生长的条件与溶液实验的条件不同。仍然存在这样的可能性,即晶体生长所需的2 - 甲基 - 2,4 - 戊二醇可能促进水化脊的形成,从而稳定直的腺嘌呤序列,尽管最近的核磁共振结果表明在水溶液中存在水化脊。我们还扩展了我们先前报道的观察结果,即非自互补DNA结构可以在晶格中以两种取向存在。一条链上溴化的十二聚体提供了实验证据,表明d(CGCAAAAAAGCG)/d(CGCTTTTTTG CG)在晶格中以两种取向定位,这两种取向通过围绕序列的假二元轴旋转180度相关。(摘要截断于400字)