Effect of Genomic Long-Range Correlations on DNA Persistence Length: From Theory to Single Molecule Experiments

详细信息    查看全文

• 作者：Julien Moukhtar ; Cendrine Faivre-Moskalenko ; Pascale Milani ; Benjamin Audit ; Cedric Vaillant ; Emeline Fontaine ; Fabien Mongelard ; Guillaume Lavorel ; Philippe St-Jean ; Philippe Bouvet ; Franoise Argoul
• 刊名：Journal of Physical Chemistry B
• 出版年：2010
• 出版时间：April 22, 2010
• 年：2010
• 卷：114
• 期：15
• 页码：5125-5143
• 全文大小：783K
• 年卷期：v.114,no.15(April 22, 2010)
• ISSN：1520-5207

文摘

Sequence dependency of DNA intrinsic bending properties has been emphasized as a possible key ingredient to in vivo chromatin organization. We use atomic force microscopy (AFM) in air and liquid to image intrinsically straight (synthetic), uncorrelated (hepatitis C RNA virus) and persistent long-range correlated (human) DNA fragments in various ionic conditions such that the molecules freely equilibrate on the mica surface before being captured in a particular conformation. 2D thermodynamic equilibrium is experimentally verified by a detailed statistical analysis of the Gaussian nature of the DNA bend angle fluctuations. We show that the worm-like chain (WLC) model, commonly used to describe the average conformation of long semiflexible polymers, reproduces remarkably well the persistence length estimates for the first two molecules as consistently obtained from (i) mean square end-to-end distance measurement and (ii) mean projection of the end-to-end vector on the initial orientation. Whatever the operating conditions (air or liquid, concentration of metal cations Mg²⁺ and/or Ni²⁺), the persistence length found for the uncorrelated viral DNA underestimates the value obtained for the straight DNA. We show that this systematic difference is the signature of the presence of an uncorrelated structural intrinsic disorder in the hepatitis C virus (HCV) DNA fragment that superimposes on local curvatures induced by thermal fluctuations and that only the entropic disorder depends upon experimental conditions. In contrast, the WLC model fails to describe the human DNA conformations. We use a mean-field extension of the WLC model to account for the presence of long-range correlations (LRC) in the intrinsic curvature disorder of human genomic DNA: the stronger the LRC, the smaller the persistence length. The comparison of AFM imaging of human DNA with LRC DNA simulations confirms that the rather small mean square end-to-end distance observed, particularly for G+C-rich human DNA molecules, more likely results from a large-scale intrinsic curvature due to a persistent distribution of DNA curvature sites than from some increased flexibility.