Bioorganometallic chemistry. 13. Regioselective reduction of NAD(+) models, 1-benzylnicotinamde triflate and beta-nicotinamide ribose-5'-methyl phosphate, with in situ generated [CpRh(Bpy)H](+): structure-activity relationships, kinetics, and mechanistic aspects in the formation of the 1,4-NADH derivatives.

Cofactor regeneration; i.e., regiospecific conversion of NAD(+) to 1,4-NADH, has been extensively studied and is a crucial component in the eventual use of 1,4-NADH in a variety of bioorganic synthesis processes involving the formation of chiral organic compounds. We have studied the reduction of a model NAD(+) compound, 1-benzylnicotinamide triflate, 1a, using CpRh(bpy)(H(2)O), 2 (Cp = eta(5)-C(5)Me(5), bpy = 2,2'-bipyridyl), as the catalyst precursor and sodium formate (HCO(2)Na) as the hydride source in 1:1 H(2)O/THF and have found exclusive 1-benzyl-1,4-dihydronicotinamide regioselectivity, as was observed previously for natural NAD(+) that provided 1,4-NADH (see: Steckhan et al. Organometallics 1991, 10, 1568). Moreover, a variety of 3-substituted derivatives of 1-benzylpyridinium triflate, in addition to the -C(O)NH(2) group (1a), were also studied to ascertain that this 3-functionality (e.g., -C(O)NHCH(3), -C(S)NH(2), -C(O)CH(3), -C(O)OCH(3), and -CN, 1b,d-g) coordinates to a CpRh(bpy)H complex to direct the concerted, regioselective transfer of the hydride group from the rhodium to the 4-ring position of the NAD(+) model; all coordinating 3-substituents had relative rates in the 0.9-1.3 range with substrate 1a set to 1.0. If in fact the 3-substituent presented a steric effect [-C(O)NH(CH(2)CH(3))(2), 1c] or was a nonbinding group (-CH(3), 1h; -H, 1i), no catalytic hydride transfer was observed even with the more electrophilic 2 and 6 ring positions being readily available, which further implicated the crucial coordination of the NAD(+) model to the CpRh metal ion center. We also found that the 1-benzyl substituent on the nitrogen atom exerted a substantial electron-withdrawing effect, in comparison to the electron-donating 1-methyl substituent, and favorably affected the rate of the regioselective reduction (rate enhancement of 1-benzyl/1-methyl = 2.0). The kinetics of the regioselective reduction of 1a were studied to show that the initial rate of reduction, r(i), is affected by the concentrations of the substrate, 1a, precatalyst, 2, and the hydride source, HCO(2)Na, in 1:1 H(2)O/THF: d[1-benzyl-1,4-dihydronicotnamide]/dt = k(cat)[1a][2][HCO(2)Na]. Furthermore, we wish to demonstrate that a previously synthesized aqueous NAD(+) model, beta-nicotinamide ribose-5'-methyl phosphate, 3, shows a similar regioselectivity for the 1,4-NADH analogue, while the initial rate (r(i)) for the regioselective reduction of 3 and NAD(+) itself was found to be comparable in water but faster by a factor of approximately 3 in comparison to 1a in 1:1 H(2)O/THF; the solvent, THF, appeared to inhibit the rate of reduction in 1a by presumably competing with the substrate 1a for the CpRh metal ion center. However, in H(2)O, the initial kinetic rate for substrate 3 was not affected by its concentration and implies that, in H(2)O, CpRh(bpy)H formation is rate determining. We assume that binding of 3 and NAD(+) to the CpRh metal ion center is also a pertinent step for 1,4-dihydro product formation, the experimental rate expression in H(2)O being d[1,4-dihydro-beta-nicotinamide ribose-5'-methyl phosphate]/dt = k(cat)[2][HCO(2)Na]. What we have discovered, for the first time, is evidence that the regioselective reduction of NAD(+) to 1,4-NADH by CpRh(bpy)H is a consequence of the amide's ability to coordinate to the CpRh metal center, thereby constricting the kinetically favorable six-membered ring transition state for plausible concerted hydride transfer/insertion to C4 to regioselectively provide the 1,4-NADH derivative; CpRh(bpy)H can be categorized as a biomimetic enzymatic hydride via its ability to bind and regioselectively transfer hydride to C4, exclusively. Clearly, the pyrophosphate and adenosine groups associated with the structure of NAD(+) are not essential in the rate of hydride transfer to C4, with NAD(+) model 3 having a similar initial rate (r(i)) of reduction as NAD(+) itself in water. Finally, a catalytic cycle will be proposed to account for our overall observations.