The photochemical mechanism of a B12-dependent photoreceptor protein

The coenzyme B12-dependent photoreceptor protein, CarH, is a bacterial transcriptional regulator that controls the biosynthesis of carotenoids in response to light. On binding of coenzyme B12 the monomeric apoprotein forms tetramers in the dark, which bind operator DNA thus blocking transcription. Under illumination the CarH tetramer dissociates, weakening its affinity for DNA and allowing transcription. The mechanism by which this occurs is unknown. Here we describe the photochemistry in CarH that ultimately triggers tetramer dissociation; it proceeds via a cob(III)alamin intermediate, which then forms a stable adduct with the protein. This pathway is without precedent and our data suggest it is independent of the radical chemistry common to both coenzyme B12 enzymology and its known photochemistry. It provides a mechanistic foundation for the emerging field of B12 photobiology and will serve to inform the development of a new class of optogenetic tool for the control of gene expression.

. Three time profiles were calculated for k 1 = 10 (ms) -1 , (black), k 2 = 3 (ms) -1 (red) and k 3 = 1 (ms) -1 (blue). The four individual signal and reference datasets are overlaid for each time profile from dark to lighter tones. They show that species with lifetimes longer than the time difference between two sequential probe pulses result in an inverse transient absorption signal prior to the subsequent laser excitation. These signals also overlay with the transient absorption signals that evolve after the subsequent excitation pulse. The time window covered by the pump-probe experiment is typically two orders of magnitude smaller than the lifetime of the long-lived species. These contributions therefore manifest as a constant component in each data acquisition window. Therefore, it is necessary to use a constant function over all acquisition wavelengths in the global fit analysis. See also the methods section of the main article. determine the SAS of |B the branching ratio,  was varied between 0.0128 (red) and 0.0462 (green) in 0.0018 steps, with the optimum spectrum achieved at  = 0.022 (black). c. To determine the SAS of |C the branching ratio, , was varied between 0.0320 (red) to 0.1441 (green) in 0.00118 steps, with the optimum spectrum achieved at  = 0.084 (black). S19

Modelling of the ultrafast transient absorption data
Global fitting of transient absorption data results in DADS that then can be linearly combined based on a model to generate SAS. The fs -ns data following the photoexcitation of CarH-GS required the sum of three exponentials for a good quality fit. Therefore, only models that consist of a system of differential equations with three eigenvalues were applied. Three different models were used which consist of one ground state, |GS , one excited state, |A , and two intermediates, |B and |C .
A schematic drawing of each is shown in Supplementary Fig. 18.
Model 1: Model 2: Inversion of these linear equations yields the SADS, and ultimately the SAS are generated by adding a fraction of the CarH-GS spectrum (c 0 ). This therefore requires knowledge about c 0 and the branching fractions,  and  (model 1) or  and  (model 2), used in each model. These parameters cannot be determined by a fit to the data. For each model, conversion to the SAS are therefore given as follows: Model 0: Model 1: Model 2: As can be seen in eqs. 10, 13, and 16, the SAS of the first intermediate is given by the sum of all three DADS plus an appropriate amount, c 0 , of the ground state spectrum in all three models.
Different values of c 0 were therefore screened to assess the effect on the resulting shape of the SAS of |A . To achieve a reasonable SAS the following two criteria were used: i) the SAS must not  Fig. 19d). Therefore, model 0 was considered an inappropriate description of the data.
In model 1 the branching ratios  and  also need to be chosen using the same constrains set out above for c 0 . The SAS of |B only depends on the branching ratio  (eq. 14). Supplementary Fig.   20b shows the SAS of |B for different values of , tuned from 0.0320 to 0.1916 in 0.0084 steps.  < 0.0320 resulted in negative features of the SAS.  < 0.1076 still shows highly unlikely spectral kinks at ~ 400, 425, 445, 525, 555, and 590 nm, whereas  > 0.1076 the SAS started to resemble the CarH-GS spectrum ( Supplementary Fig. 20d). Although the SAS of |B at  = 0.1076 represents perhaps the optimum spectrum for this model, it still shows a high similarity to the CarH-GS spectrum. The SAS of |C in model 1 depends on both branching ratio  and  (eq. 15), where  is already determined by generating the SAS of |B . Fixing  = 0.1076 and tuning  from 0.283 to 0.774 in 0.026 steps gives SAS without negative features for |C ( Supplementary Fig. 20c).  < 0.283 the SAS still shows a highly unlikely spectral kink at ~ 590 nm. Above this value the SAS starts to resemble the spectrum of CarH-GS ( Supplementary Fig. 20d). Therefore,  = 0.5157 results in the most likely SAS of |C , which closely resembles the pure spectrum of the MLCT state observed previously following the photoexcitation of methylcobalamin. 4 However, the similarity of