Treatment of a self-complementary duplex with Pt anticancer drugs leads to formation of an unusual type of hairpin-like oligonucleotide, Pt(A2){5'd(AiT2G3G4*G5*T6A7CgC9Ci0Ai iT^'} (A2 = cw-(NH3)2or ethylenediamine (en) and G*'s are platinated at N7). In previous NMR studies, several residues exhibited abnormal upfield-and downfield-shifted 'H and 31P signals, and one aromatic *H signal, G4*H8, could not be located. In the present study, we found the G4*H8 signal to be broad and shifted ~2 ppm upfield into the HI' spectral region. This shift change is much larger than the theoretical maximum (< 1 ppm) predicted for an unstrained structure, suggesting the platination site is strained. Distance geometry (DG) structures were calculated from NMR data in order to elucidate the structural basis for the unique properties of these species, including the unusual NMR shifts. At the platination site, the Pt(en) moiety is located in the major groove, and the platinum-bound G's both possess some unusual features; G4* has an N sugar conformation and G5* has a syn conformation. The syn Gj* base is oriented perpendicular to the Pt coordination plane. In contrast, the G4* base is almost coplanar with this plane; the unusual orientation forces G4*H8 into a close (<3 A) clash with the five-membered ring of G3*. This proximity explains the substantial upfield shift observed for G4*H8. The broadness of the G4*H8 signal could be explained by minor vacillations about the Pt-N7(4) bond, which would sweep G4*H8 across different shielding regions of the anisotropic G3* base. The A7 base in DG models is tucked inside the hairpin loop, with A7H8 close to G5*H8. A7 is correctly oriented to explain (a) the G4*H2' shift into the methyl region of the 'H spectrum, (b) the strong A7H8 to G3*H8 NOE, and (c) the downfield-shifted A7H8 signal (caused by Gs* deshielding). The unprecedented downfield 31P shift of A7pCg is a result of an anti,anti conformation about the f,a torsion angles induced by a distortion in the backbone needed to allow G4*C9 Watson-Crick base pairing at the top of the stem. A B-DNA-like helical stem was found with base pairing between the first four bases of the 3' and 5' ends (AjTu, T2A11, G3Ct0, and G4*C9). These features are supported by 2D N0ESY-in-H20 data. Such hairpin-like structures, induced by the need to balance DNA and Pt structural demands, could form in palindromic regions of DNA and could be instrumental in platinum drug activity.

Metal ligation by N-donor ligands in biological systems involves heterocyclic ligands almost exclusively.1 The two most common classes of such ligands are macrocycles (e.g., porphyrins) and lopsided bases (B, e.g., imidazoles). Ligation by two or more of the same type of B is widespread. For two cis B's, the corresponding atoms of each ligand can be on the same side or opposite sides of the N-M-N plane, giving the head-to-head (HH) and the head-to-tail (HT) orientations, respectively. The most lopsided B's are five-six bicyclic ligands, Bs_6 (purines, benzimidazoles). The vast majority of cij-bis(B5_6) complexes are HT.2-* The very few reported HH c/í-bisfBs-é) complexes are square-planar.2,5-8 These HH cii-bis(Gua)Pt(II) complexes

The solution and solid state structures of two octahedral Ru(II) complexes, cis,cis,cis-RuCl(2)(Me(2)SO)(2)(py)(Me(3)Bzm) (Me(3)Bzm = 1,5,6-trimethylbenzimidazole, py = pyridine) (1) and cis,cis,cis-RuCl(2)(Me(2)SO)(2)(Me(3)Bzm)(2) (2), were compared. 2, the subject of a preliminary report, is described in more detail here. 1 has two possible geometric isomers with py trans to Cl in one (position "a") and trans to Me(2)SO in the other (position "b"), Me(3)Bzm occupying the other position in each isomer. The X-ray structure of 1 revealed that py is at "a". Since Me(3)Bzm is lopsided, each Me(3)Bzm has two possible orientations related by a rotation of approximately 180 degrees about the Ru-N3 bond; there are two possible atropisomers for each geometric isomer of 1 and four for 2. For 1, the solid state structure shows that Me(3)Bzm adopts the orientation with H2 (H on C between the two N's) pointing between the two cis Cl ligands, the same disposition as Me(3)Bzm "b" in 2 in the solid. For 1, the py signals (two broad py alpha and beta signals, a sharp gamma signal) in CDCl(3) show that py "a" is rotating on the NMR time scale and that only one atropisomer is present. This interpretation was supported by ROESY and EXSY (1)H NMR spectra. The (1)H NMR shift pattern and the NOE data can be understood best if Me(3)Bzm "b" remains primarily in the orientation found in the solid. The solution data for 1, with the nonlopsided and sterically less demanding py ligand, provide insight into the more complicated properties of 2. For 2, there is a marked dispersion of (1)H NMR signals of Me(3)Bzm "a" between the two atropisomers, which have nearly equal stability. One atropisomer is a head-to-head (HH) and the other a head-to-tail (HT) species. Me(3)Bzm "a" flips between the two species. Thus, ligand "a" is fluxional in both complexes. The dispersion of Me(3)Bzm "a" signals is due to the effect of Me(3)Bzm "b" anisotropy. For 1 and both atropisomers of 2, Me(3)Bzm "b" prefers one orientation, which appears to be the most hindered orientation. We postulate that the H2 of Me(3)Bzm "b" is electrostatically attracted to the two cis halides, accounting for this surprising result. Crystallographic details for 1 are as follows: C(19)H(29)Cl(2)N(3)O(2)RuS(2), P2(1)/c, a = 10.947(1) Å, b = 9.046(1) Å, c = 24.221(2) Å, D(calcd) = 1.580 g cm(-)(3), Z = 4, R = 0.026 for 4627 independent reflections.

In examining NMR methods to assess orientation and fluxional motions of planar N-donor heterocyclic coordinated ligands (L's) in solution, we introduce the use of exchange-NOE NMR data as a powerful method for defining (i) the extent of L rotation about the metal-N bond, (ii) the direction of L rotation, and (iii) even the halves of C 2 -symmetrical L's that interchange during dynamic processes. The full value of the approach depends on a strategy in which the complexes studied are chiral and similar except that one has a lopsided L (Me 3 Bzm ) 1,5,6trimethylbenzimidazole in [Re 2 O 3 Cl 4 (Me 3 Bzm) 4 ] ( 1)) and the other has a C 2 -symmetrical L (3,5-lut ) 3,5-lutidine in [Re 2 O 3 Cl 4 (3,5-lut) 4 ] (2)). Each Re is part of a nearly linear OdRe-O-RedO grouping and has a "terminal" L (L t ) and a stacked L (L s ). The fluxional inversion of the two chiral dimers involves rotations of ∼180°about the Re-O-Re bonds and of ∼90°about all four Re-L bonds. The exchange-NOE data for 2 show that the half of L t away from the dimer center interchanges with the half of L s close to the center, with the L plane rotating past the OdRe-O bonds, not the N-Re-Cl bonds. Thus, the exchange-NOE data help to establish the direction of L rotation. In 1, Me 3 Bzm t has the rare head-to-head (HH) orientation with respect to the partner. This partner Me 3 -Bzm s stacks with the Me 3 Bzm s from the other Re in the common head-to-tail (HT) orientation. Compelling evidence that the predominant solution conformer of 1 has the HH,HT,HH structure includes unusual chemical shift dispersions and a strong interligand NOE cross-peak. This is the only case in which cis,bis imidazole-ring-ligated untethered ligands have been found to be predominantly HH in solution. This predominance can be attributed to the electrostatic attraction of the δ + N 2 C proton for the negative core of the molecule (bridging O, cis Cl on same Re, and two cis Cl's on the other Re).