A deazariboflavin chromophore kinetically stabilizes reduced FAD state in a bifunctional cryptochrome

An animal-like cryptochrome derived from Chlamydomonas reinhardtii (CraCRY) is a bifunctional flavoenzyme harboring flavin adenine dinucleotide (FAD) as a photoreceptive/catalytic center and functions both in the regulation of gene transcription and the repair of UV-induced DNA lesions in a light-dependent manner, using different FAD redox states. To address how CraCRY stabilizes the physiologically relevant redox state of FAD, we investigated the thermodynamic and kinetic stability of the two-electron reduced anionic FAD state (FADH−) in CraCRY and related (6–4) photolyases. The thermodynamic stability of FADH− remained almost the same compared to that of all tested proteins. However, the kinetic stability of FADH− varied remarkably depending on the local structure of the secondary pocket, where an auxiliary chromophore, 8-hydroxy-7,8-didemethyl-5-deazariboflavin (8-HDF), can be accommodated. The observed effect of 8-HDF uptake on the enhancement of the kinetic stability of FADH− suggests an essential role of 8-HDF in the bifunctionality of CraCRY.


Theoretical part of the midpoint potential measurement of FAD by xanthine/xanthine oxidase method
When the oxidation of xanthine (X) by xanthine oxidase (XO) supplies electrons to FAD and a reference dye without accumulating their one-electron reduced forms, the Nernst equation is applied to each redox active compound, as follows 1 .
By substituting constant values into Equation S3, Equation S4 is obtained for the midpoint potential analysis of the FADox/FADH − couple in millivolts.
12.8 ln The rate constants of k1, k2, and k3 in Supplementary Scheme S1 are related to the concentrations of redox forms of FAD as follows.
Considering that the total concentration of FAD is the summation of the concentrations of FADH − , FADH • , and FADox, D[FADox] should be described by Equation S10.
However, the k2 value in our tested proteins is much smaller than k1 and/or k3 as discussed in Results.
For At64, At64-DExLoop, and CraCRY-HDF, the absorption changes unique to the reoxidation were observed to be fitted not with a single exponential component but with a line because of their slow reoxidation kinetics.When we analyzed these data, we used Equations S13, S14, and S15, which were derived from the linear approximation of Equations S8, S11, and S12, respectively.
Based on the obtained absorption changes, the Beer-Lambert law (A = εcl) and Equations S8, and S11-15, we calculated the rate constants of k1 and k3 and the time-dependent concentrations of

)
By solving Equations S5 and S6, we could describe the time development of the concentrations of FADH − and FADH • (D[FADH − ] and D[FADH • ], in which D means the changes in the concentration of a reactant over a specific time interval), as shown in Equations S8 and S9.
(a) The representative UV/vis absorption spectra measured before and 1 h after the reaction indicated that FAD and Safranin T were gradually converted to their two-electron reduced states without concentrating FADH • .The insets show the decay of the absorption at 450 nm and 510 nm unique to the fully oxidized FAD and Safranin T forms, respectively, during the measurement time of 1 h.(b), (c), (d) Three independent experimental data were analyzed by the plots (12.8 ln ([FADox]t/[FADH − ]t) vs. 12.8 ln ([Dox]t/[Dred]t)).The plots were approximated by a line for each experiment.The slope values of the lines described in the figures were close to 1, validating the analyses.Supplementary Figure 3. Midpoint potential measurements of FAD in Xl64 by the xanthine/xanthine oxidase method.(a) The representative UV/vis absorption spectra measured before and 1 h after the reaction indicated that FAD and Safranin T were gradually converted to their two-electron reduced states without concentrating FADH • .The insets show the decay of the absorption at 450 nm and 510 nm unique to the fully oxidized FAD and Safranin T forms, respectively, during the measurement time of 1 h.(b), (c), (d) Three independent experimental data were analyzed by the plots (12.8 ln ([FADox]t/[FADH − ]t) vs. 12.8 ln ([Dox]t/[Dred]t)).The plots were approximated by a line for each experiment.The slope values of the lines described in the figures were close to 1, validating the analyses.determination of the k1 + k3 value from the fitting of the FADH • concentration with Eq. 11.The FADH − concentration at each time point was extrapolated from the rate constant.Finally, the FADox concentration was obtained using absorbance at 450 nm, after subtraction of the contributions of FADH • and FADH − at this wavelength.As the total FAD concentration remains nearly constant with a standard deviation of less than 5% for all the samples, the deconvolution of FAD concentrations was reasonably performed.SupplementaryFigure 7. Structural comparison of the FAD binding site within 5 Å of FAD among Xl64, At64, and CraCRY.The used structure of Xl64 was the model generated previously 4 , and the structures of At64 and CraCRY were taken from the crystal structures 6,7 of PDB ID: 3FY4 and 6FN2, respectively.FAD is shown in yellow.The conserved residues among the proteins are shown in gray lines.The residues showing some varieties among the proteins are shown in sticks.The green, cyan, and magenta sticks correspond to the residues of At64, Xl64, and CraCRY, respectively.
In these equations, reduction potentials of FAD and the reference dye represented as EFAD and ED, respectively, are related to their concentration ratios between oxidized and two-electron reduced states ([FADox]t and [FADH − ]t for FAD, [Dox]t and [Dred]t for the dye, where t indicates the time at which the absorbance is measured) and their midpoint potentials (Em,FAD and Em,D).R is the gas constant (8.314J K −1 mol −1 ), F is the Faraday constant (96,485 C mol −1 ), and T is the absolute temperature (298 K in the experiment).Supposing that their midpoint potentials are