The ligand-structure-selective binding of oligonucleotide by cobalt complexes Huili Chen a, * , Wei Gao a , Fengmin Zhang b , Chunjiao Dou a , Pin Yang a a Institute of Molecular Science, Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Shanxi University, Wucheng Road No. 92, Taiyuan 030006, PR China b Testing Center, Test and Service Center for Material Microanalysis and Property of Jiangsu Province, Yangzhou University, Wenhui East Road No. 48, Yangzhou 225009, Jiangsu Province, PR China article info Article history: Received 27 September 2009 Accepted 9 December 2009 Available online 16 December 2009 Keywords: Oligonucleotide DNA Cobalt complex Intercalation NMR abstract The binding of two Co(III) complexes [Co(phen) 2 (DPQ)] 3+ and [Co(phen) 2 (HPIP)]Cl 3 [HPIP = 2-(2-hydroxy- phenyl) imidazo [4,5-f][1,10] phenanthroline, DPQ = dipyrido[3,2-f:2 0 ,3 0 -h]quinoxaline] to the normal base-paired decanucleotide d(CCTAATTAGG) 2 was studied by 2D NMR. The results indicate that the width of intercalating ligand has a large effect on the selectivity of binding site. For [Co(phen) 2 (HPIP)]Cl 3 , the complex binds the decanucleotide at C 2 T 3 :G 9 A 8 and A 4 A 5 :T 7 T 6 by intercalation from the minor groove, while [Co(phen) 2 (DPQ)] 3+ intercalates into T 3 A 4 :T 7 A 8 region from the minor groove. The conclusion was further proved by molecular modeling. Ó 2009 Elsevier B.V. All rights reserved. In the past decade, many octahedral Co(III) polypyridyl-type complexes have been synthesized as probes for DNA secondary structure, and their DNA-binding properties have been investi- gated mostly by UV–vis absorption spectra, emission spectra, vis- cosity measurements, circular dichroism spectra and gel electrophoresis experiments, etc. [1–12]. However, these methods are based on the properties of whole DNA molecules and could not provide the information about binding site. Two-dimensional NMR is powerful tool to detect DNA interactions and locate accurate binding sites, such as major/minor groove binding mode and base pairs involved in the complex intercalation [13–15]. In our previous work, we initially reported the binding of [Co(phen) 2 (DPQ)] 3+ to hexanucleotide d(GTCGAC) 2 by 2D NMR and molecular simulations [16]. However, as DNA model, hexamer is too short to form an integrated duplex. Herein, as an extension of our work, we studied the binding of the complex [Co(phen) 2 - (DPQ)] 3+ and [Co(phen) 2 (HPIP)] 3+ to the decanucleotide d(CCTAAT- TAGG) 2 by 2D NMR (Fig. 1). The results suggest that the two complexes, with different width of the intercalating ligand, bind to the decanucleotide at different sites. [Co(phen) 2 (DPQ)]Cl 3 , with narrower ligand DPQ, intercalates into T 3 A 4 :T 7 A 8 region from the minor groove, while [Co(phen) 2 (HPIP)] 3+ , with wider ligand HPIP, intercalates into C 2 T 3 :G 9 A 8 and A 4 A 5 :T 7 T 6 region from the minor groove. The 1 H NMR resonances of the free oligonucleotide acid have been assigned in the previous work [13]. The imino resonances spectrum indicates that only the terminal residues failed to form stable base pairs (Supplementary material Fig. S1). It is therefore believed that the oligonucleotide is predominantly present as a stable duplex in the experimental condition. Addition of [Co(phen) 2 (HPIP)] 3+ or [Co(phen) 2 (DPQ)] 3+ to d(CC- TAATTAGG) 2 induced large chemical shift changes for the ligand HPIP (especially H12, H13, H14, H15, H16) and DPQ (H11, H12, H13), while small shifts were observed for the ancillary ligand phen (Fig. 2). This behavior is consistent with preferential oligonu- cleotide binding of the ligand HPIP or DPQ by intercalation [13– 15]. However, the chemical shift changes from the oligonucleotide are different for the two complexes binding. For [Co(phen) 2 (HPIP)] 3+ binding, a large upfield (0.771 ppm) of T 3 Me in the major groove indicates the strong shied from the intercalating ligand, and reveals the binding site at T 3 :A 8 region. While for [Co(phen) 2 (DPQ)] 3+ binding, significant upfield shifts were ob- served for the particular protons located in the minor groove in T 3 A 4 /T 7 A 8 region, such as A 4 H1 0 (0.07 ppm), T 7 H1 0 (0.09 ppm), T 7 H2 00 (0.08 ppm) and A 8 H2 00 (0.10 ppm), which imply that the complex intercalates into T 3 A 4 :T 7 A 8 from the minor groove. NOESY spectra of d(CCTAATTAGG) 2 bound to the complexes (1:1) was recorded respectively at 20 °C with mixing time of 300 ms (Fig. 3 and Supplementary material Tables S3–S4). In addition to the expected intraduplex sequential NOEs from the oli- gonucleotide, a number of intermolecular NOE cross-peaks be- tween complexes to d(CCTAATTAGG) 2 were observed (Fig. 3). For [Co(phen) 2 (HPIP)] 3+ binding, the intense NOE cross-peaks between phen and sugar are observed, including H3/8–G 9 H1 0 (A), H3/8– T 3 H1 0 (B), H5/6–T 3 H1 0 (E), H5/6–A 8 H1 0 (I), H4/7–A 8 H1 0 (J), H4/7–T 3 H1 0 (K). Because the sugar H1 0 protons are located in the decanucleotide minor groove, these NOEs indicate that the phen 1387-7003/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.inoche.2009.12.012 * Corresponding author. Tel.: +86 3517010699; fax: +86 3517011022. E-mail address: huilichen@sxu.edu.cn (H. Chen). Inorganic Chemistry Communications 13 (2010) 310–313 Contents lists available at ScienceDirect Inorganic Chemistry Communications journal homepage: www.elsevier.com/locate/inoche