Alexey Aleksandrov

A promising fast, negative, red-absorbing photochromic system based on an intraprotein intermolecular charge transfer (CT) complex between a flavin cofactor and a substrate-analogue inhibitor MXA (methylthioacetate or methylselenoacetate; X = S, Se) within monomeric sarcosine oxidase (MSOX) displays MXA-independent, barrierless, high-yield femtosecond photoactivation associated with CT-interaction disruption coupled to MXA isomerization. Assessing the origin of the known MXA-dependence of the nanosecond thermal recovery requires a very precise determination of activation energy barriers. Here, a newly developed visible pump–probe setup employing two Ti:sapphire femtosecond lasers spanning the picosecond to microsecond time scale was used. Whereas the room temperature thermal recovery time for MSeA is more than double that for MTA (7.5 ns vs 2.9 ns), the dissociation rate, the quantum yield (0.85), and recombination activation enthalpy (21 kJ·mol–1) are MXA-independent. Quantum mechanical calculations rationalize this value as the intrinsic enthalpic barrier for MXA isomerization. The observed variation in thermal recovery is due to the activation entropy, which is negative (on the order of −15 J·mol–1·K–1) and MXA-dependent. This dependence is attributed to a differently ordered transition state. Implications for the design and application of this class of photochromic systems are discussed.

The initial photoproduct of the natural photoenzyme fatty acid photodecarboxylase involves the flavin anion radical flavin adenine dinucleotide (FAD ^•– ). Using spectrally resolved ultrafast transient absorption spectroscopy, we demonstrate that FAD ^•– photoexcitation in the absence of substrate leads to the formation of the oxidized flavin FAD _ox (the resting state in the catalytic cycle) within 100 femtoseconds. While this feature is similar to that occurring in flavoprotein oxidases, the ensuing photocycle is more complex. Upon excitation at the lowest-energy transition, the ejected electron is initially captured as a hydrated electron ( _e ^– _H ) before transferring to a secondary acceptor in 2.5 picoseconds and returning to the flavin in 37 picoseconds. This implies that _e ^– _H can be generated within a protein environment, an unprecedented finding. This assessment is supported by molecular dynamics simulations showing an expansion of the flavin-binding pocket without substrate, allowing water molecules to fill the void. Our results may pave the way to developing unconventional photocatalytic processes.

Small molecule empirical force fields (FFs), including the CHARMM General Force Field (CGenFF), are designed to have wide coverage of organic molecules and to rapidly assign parameters to molecules not explicitly included in the FF. Assignment of parameters to new molecules in CGenFF is based on a trained bond-angle-dihedral charge increment linear interpolation scheme for the partial atomic charges along with bonded parameters assigned based on analogy using a rules-based penalty score scheme associated with atom types and chemical connectivity. Accordingly, the accuracy of CGenFF is related to the extent of the training set of available parameters. In the present study that training set is extended by 1390 molecules selected to represent connectivities new to CGenFF training compounds. Quantum mechanical (QM) data for optimized geometries, bond, valence angle, and dihedral angle potential energy scans, interactions with water, molecular dipole moments, and electrostatic potentials were used as target data. The resultant bonded parameters and partial atomic charges were used to train a new version of the CGenFF program, v5.0, which was used to generate parameters for a validation set of molecules, including drug-like molecules approved by the FDA, which were then benchmarked against both experimental and QM data. CGenFF v5.0 shows overall improvements with respect to QM intramolecular geometries, vibrations, dihedral potential energy scans, dipole moments and interactions with water. Tests of pure solvent properties of 216 molecules show small improvements versus the previous release of CGenFF v2.5.1 reflecting the high quality of the Lennard-Jones parameters that were explicitly optimized during the initial optimization of both the CGenFF and the CHARMM36 force field. CGenFF v5.0 represents an improvement that is anticipated to more accurately model intramolecular geometries and strain energies as well as noncovalent interactions of drug-like and other organic molecules.

Small molecule empirical force fields (FFs), including the CHARMM General Force Field (CGenFF), are designed to have wide coverage of organic molecules and to rapidly assign parameters to molecules not explicitly included in the FF. Assignment of parameters to new molecules in CGenFF is based on a trained bond-angle-dihedral charge increment linear interpolation scheme for the partial atomic charges along with bonded parameters assigned based on analogy using a rules-based penalty score scheme associated with atom types and chemical connectivity. Accordingly, the accuracy of CGenFF is related to the extent of the training set of available parameters. In the present study that training set is extended by 1,390 molecules selected to represent connectivities new to CGenFF training compounds. Quantum mechanical (QM) data for optimized geometries, bond, valence angle, and dihedral angle potential energy scans, interactions with water, molecular dipole moments, and electrostatic potentials were used as target data. The resultant bonded parameters and partial atomic charges were used to train a new version of the CGenFF program, v5.0, which was used to generate parameters for a validation set of molecules, including drug-like molecules approved by the FDA, which were then benchmarked against both experimental and QM data. CGenFF v5.0 shows overall improvements with respect to QM intramolecular geometries, vibrations, dihedral potential energy scans, dipole moments and interactions with water. Tests of pure solvent properties of 216 molecules show small improvements versus the previous release of CGenFF v2.5.1 reflecting the high quality of the Lennard-Jones parameters that were explicitly optimized during the initial optimization of both the CGenFF and the CHARMM36 force field. CGenFF v5.0 represents an improvement that is anticipated to more accurately model intramolecular geometries and strain energies as well as non-covalent interactions of drug-like and other organic molecules.

Folates comprise a crucial class of biologically active compounds related to folic acid, playing a vital role in numerous enzymatic reactions. One‐carbon metabolism, facilitated by the folate cofactor, supports numerous physiological processes, including biosynthesis, amino acid homeostasis, epigenetic maintenance, and redox defense. Folates share a common pterin heterocyclic ring structure capable of undergoing redox reactions and existing in various protonation states. This study aimed to derive molecular mechanics (MM) parameters compatible with the CHARMM36 all‐atom additive force field for pterins and biologically important folates, including pterin, biopterin, and folic acid. Three redox forms were considered: oxidized, dihydrofolate, and tetrahydrofolate states. Across all protonation states, a total of 18 folates were parameterized. Partial charges were derived using the CHARMM force field parametrization protocol, based on targeting reference quantum mechanics monohydrate interactions, electrostatic potential, and dipole moment. Bonded terms were parameterized using one‐dimensional adiabatic potential energy surface scans, and two‐dimensional scans to parametrize in‐ring torsions associated with the puckering states of dihydropterin and tetrahydropterin. The quality of the model was demonstrated through simulations of three protein complexes using optimized and initial parameters. These simulations underscored the significantly enhanced performance of the folate model developed in this study compared to the initial model without optimization in reproducing structural properties of folate–protein complexes. Overall, the presented MM model will be valuable for modeling folates in various redox states and serve as a starting point for parameterizing other folate derivatives.

Photoreduction of oxidized flavins has a functional role in photocatalytic and photoreceptor flavoproteins. In flavoproteins without light-dependent physiological functions, ultrafast, reversible flavin photoreduction is supposedly photoprotective by nature, and holds potential for nonnatural photocatalytic applications. In this work, we combine protein mutagenesis, ultrafast spectroscopy, molecular dynamics simulations and quantum mechanics calculations to investigate the nonfunctional flavin photoreduction in a flavoenzyme, lysine-specific demethylase 1 (LSD1) which is pivotal in DNA transcription. LSD1 harbors an oxidized flavin adenine dinucleotide (FAD) cofactor and multiple electron-donating residues in the active site. Upon photoexcitation, the FAD cofactor is photoreduced in < 200 fs by electron transfer (ET) from nearby residue(s), and the charge pairs recombine in ca. 2 ps. Site-directed mutagenesis pinpoints a specific tryptophan residue, W751, as the primary electron donor, whereas a tyrosine residue, Y761, despite being located closer to the flavin ring, does not effectively contribute to the process. Based on a hybrid quantum–classical computational approach, we characterize the W751–FAD and Y761–FAD charge-transfer states (CT_W751 and CT_Y761, respectively), as well as the FAD locally excited state (LE_FAD), and demonstrate that the coupling between LE_FAD and CT_W751 is larger than those involving CT_Y761 by an order of magnitude, rationalizing the experimental observations. More generally, this work highlights the role of the intrinsic protein environment and details of donor–acceptor molecular configurations on the dynamics of short-range ET involving a flavin cofactor and amino acid residue(s).

In fatty acid photodecarboxylase (FAP), light‐induced formation of the primary radical product RCOO⋅ from fatty acid RCOO⁻ occurs in 300 ps, upon which CO₂ is released quasi‐immediately. Based on the hypothesis that aliphatic RCOO⋅ (spectroscopically uncharacterized because unstable) absorbs in the red similarly to aromatic carbonyloxy radicals such as 2,6‐dichlorobenzoyloxy radical (DCB⋅), much longer‐lived linear RCOO⋅ has been suggested recently. We performed quantum chemical reaction pathway and spectral calculations. These calculations are in line with the experimental DCB⋅ decarboxylation dynamics and spectral properties and show that in contrast to DCB⋅, aliphatic RCOO⋅ radicals a) decarboxylate with a very low energetic barrier and on the timescale of a few ps and b) exhibit little red absorption. A time‐resolved infrared spectroscopy experiment confirms very rapid, ≪300 ps RCOO⋅ decarboxylation in FAP. We argue that this property is required for the observed high quantum yield of hydrocarbons formation by FAP.

In fatty acid photodecarboxylase (FAP), light‐induced formation of the primary radical product RCOO⋅ from fatty acid RCOO⁻ occurs in 300 ps, upon which CO₂ is released quasi‐immediately. Based on the hypothesis that aliphatic RCOO⋅ (spectroscopically uncharacterized because unstable) absorbs in the red similarly to aromatic carbonyloxy radicals such as 2,6‐dichlorobenzoyloxy radical (DCB⋅), much longer‐lived linear RCOO⋅ has been suggested recently. We performed quantum chemical reaction pathway and spectral calculations. These calculations are in line with the experimental DCB⋅ decarboxylation dynamics and spectral properties and show that in contrast to DCB⋅, aliphatic RCOO⋅ radicals a) decarboxylate with a very low energetic barrier and on the timescale of a few ps and b) exhibit little red absorption. A time‐resolved infrared spectroscopy experiment confirms very rapid, ≪300 ps RCOO⋅ decarboxylation in FAP. We argue that this property is required for the observed high quantum yield of hydrocarbons formation by FAP.

Ongoing climate change is driving the search for renewable and carbon-neutral alternatives to fossil fuels. Photocatalytic conversion of fatty acids to hydrocarbons by fatty acid photodecarboxylase (FAP) represents a promising route to green fuels. However, the alleged low activity of FAP on C2 to C12 fatty acids seemed to preclude the use for synthesis of gasoline-range hydrocarbons. Here, we reveal that Chlorella variabilis FAP ( Cv FAP) can convert n -octanoic acid in vitro four times faster than n -hexadecanoic acid, its best substrate reported to date. In vivo, this translates into a Cv FAP-based production rate over 10-fold higher for n -heptane than for n -pentadecane. Time-resolved spectroscopy and molecular modeling demonstrate that Cv FAP’s high catalytic activity on n -octanoic acid is, in part, due to an autocatalytic effect of its n -heptane product, which fills the rest of the binding pocket. These results represent an important step toward a bio-based and light-driven production of gasoline-like hydrocarbons.