Flavin reduction potential tuning by substitution and bending

Computations on a series of flavin derivatives with the B3LYP hybrid density functional and 6-311G(2d,2p) basis set indicate that the ionization potentials (IPs) of the anionic semiquinone and hydroquinone states serve as accurate predictors of the flavin one- and two-electron reduction potentials. The relation between the semiquinone IPs and the two-electron reduction potentials, Em, is ΔIP_sq-/ΔEm = 3.69 ± 0.18 meV/mV, with a very similar relation obtained for the flavin derivatives' Kohn-Sham highest occupied molecular orbital (KS-HOMO) energies, E ^KS-HOMO vs. Em. Interestingly, these good correlations between vertical IPs and E_ms are observed even though the second reduction step, Fl_sq ⇌ Flred- involves significant conformational changes for a number of the derivatives. In fact, the flavin derivatives can be divided roughly into two categories. Those derivatives with high Ems are either planar in the anionic reduced state, or else a negligible amount of energy (< kT) is required for planarizing them. Flavin derivatives of low potential however possess significant conformational energy (> kT), and tend to have larger butterfly bends. The flavin parent compound, lumiflavin, represents the dividing point between these two categories, a result with possibly interesting biological implications for the conformational control of flavin redox potentials by enzymes. B3LYP/6-311G(2d,2p) calculations on lumiflavin constrained to various butterfly bend angles show that the oxidized and semiquinone states (both anionic and protonated at N₅) resist bending, with the oxidized state being by far the stiffest. On the other hand, the optimum geometry of the fully reduced state is bent by 15.9° in the anionic state and 24.4° in the neutral state. Full MP2 geometry optimizations confirm the reduced flavin butterfly bend, however the bend angles are larger than the DFT results: 28.7° and 32.6° for the anionic and neutral states, respectively. The relation of the N₅ and N₁₀ pyramidalization to the flavin butterfly bend is discussed. The results indicate that a protein-enforced flavin conformation should have significant and differing effect on each of the one electron reduction steps.