Spectroscopic Properties of Lumiflavin: A Quantum Chemical Study

Abstract

In this work, the electronic structure and spectroscopic properties of lumiflavin are calculated using various quantum chemical methods. The excitation energies for ten singlet and triplet states as well as the analysis of the electron density difference are assessed using various wave function‐based methods and density functionals. The relative order of singlet and triplet excited states is established on the basis of the coupled cluster method CC2. We find that at least seven singlet excited states are required to assign all peaks in the UV/Vis spectrum. In addition, we have studied the solvatochromic effect on the excitation energies and found differential effects except for the first bright excited state. Vibrational frequencies as well as IR, Raman and resonance Raman intensities are simulated and compared to their experimental counterparts. We have assigned peaks, assessed the effect of anharmonicity, and confirmed the previous assignments in case of the most intense transitions. Finally, we have studied the NMR shieldings and established the effect of the solvent polarity. The present study provides data for lumiflavin in the gas phase and in implicit solvent model that can be used as a reference for the protein‐embedded flavin simulations and assignment of experimental spectra.

Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical level crossing analysis

“…Pseudo-secular couplings b i were estimated from quantum chemical calculations as differences between the principal values of the corresponding HFC tensors. These calculations were performed on a truncated model of FMN, namely the lumiflavin, which was previously 64 benchmarked by some of the present authors. Here we have studied lumiflavin in its radical anion (Lumi ●− ) form and the tryptophan residue in its radical cation (TrpH ●+ ) form.…”

Understanding flavin electronic structure and spectra

“…For example, B3LYP which incorporates 20% HF exchange, overestimates the excitation energy by 0.26 eV, whereas the deviation is greater than 1 eV using the BHLYP functional which includes 50% HF exchange. Wavefunction‐based methods such as coupled cluster at second order (CC2) and algebraic diagrammatic construction at second order ADC(2) yield excitation energies of 3.06 and 2.94 eV with triple‐ζ type basis sets, which are in much better agreement with the experimental value 25,34 . Interestingly, application of multireference methods, for example, symmetry‐adapted cluster‐configuration interaction (SAC‐CI) have led to underestimates of the experimental energy, at 2.46 eV 124 .…”

“…Furthermore, such information serves as a reference for calculations on embedded flavins and supports experimental assignments. Gas‐phase calculations were used by several groups to benchmark quantum chemical methods for OX flavin 23–25 and to elucidate electronic differences characterizing other oxidation states of flavin 26,27 . However, gas phase approaches cannot describe the response of the correlated polarizable π system to solvation 28 and protein environments 29 …”

“…Optimization of the geometrical parameters (bond lengths and angles) is one of the most important steps toward obtaining the minimum energy configuration of a molecule. Based on quantum chemical methods, optimized geometries have been reported for flavin and its derivatives by several groups, using the TD‐DFT methods for excited states 24,25,28,113 . Among geometrical parameters obtained, the bond‐length alternation (BLA) change between the S 0 and S 1 geometries 118 or between different redox states of flavin 119,120 informs on the degree of electron delocalization, as smaller BLA indicates greater electron delocalization 121 .…”

“…In the protein context, TD‐DFT was again found to overestimate the excitation energy 127 by amounts of 0.06 eV (B3LYP) and 0.54 eV (BHLYP) 126 . Application of the CAM‐B3LYP functional, which can describe charge transfer (CT) states reasonably well 25,128 results in excitation energies that are high by 0.03 eV in AppA BLUF protein and 0.25 eV in Slr1694 BLUF protein, albeit with different basis set sizes 129,130 . Meanwhile the CC2 method produces an overestimate of 0.24 eV.…”