Mechanism of Light Induced Radical Pair ... - ACS Publications

Jun 12, 2018 - Department of Theoretical Chemistry, Institute of Chemistry, University ... Department of Food Sciences, Medical University of Gdansk, ...
0 downloads 0 Views 3MB Size
Subscriber access provided by University of Wyoming Libraries

Article

Mechanism of Light Induced Radical Pair Formation in Coenzyme B12-Dependent Ethanolamine Ammonia-Lyase Abdullah Al Mamun, Megan J Toda, Piotr Lodowski, Maria Jaworska, and Pawel M. Kozlowski ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.8b00120 • Publication Date (Web): 12 Jun 2018 Downloaded from http://pubs.acs.org on June 19, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Mechanism of Light Induced Radical Pair Formation in Coenzyme B12-Dependent Ethanolamine Ammonia-Lyase

Abdullah Al Mamun,1 Megan J. Toda,1 Piotr Lodowski,2 Maria Jaworska,2 and Pawel M. Kozlowski* 1,3

1

2

Department of Chemistry, University of Louisville, Louisville, Kentucky 40292, United States

Department of Theoretical Chemistry, Institute of Chemistry, University of Silesia in Katowice, Szkolna 9, PL-40 006 Katowice, Poland 3

Visiting professor, Department of Food Sciences, Medical University of Gdansk, Al. Gen. J. Hallera 107, 80-416 Gdansk, Poland

*Phone: (502) 852-6609. Fax: (502) 852-8149. E-mail: [email protected]

1 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 39

Abstract Coenzyme B12 (Adenosylcobalamin = AdoCbl)-dependent enzymes catalyze complex molecular transformations by employing radical chemistry. The initial step in the native catalytic cycle, upon substrate binding, involves homolytic cleavage of the Co-C bond of AdoCbl to form Co(II)/Ado• radical pair (RP). Formation of Co(II)/Ado• is subsequently coupled with H-atom abstraction from the substrate. Interestingly, these same RP can be generated upon light absorption without presence of a substrate. Herein, the photochemistry associated with the mechanism of Co-C bond photo-cleavage inside the (AdoCbl)-dependent ethanolamine ammonia-lyase (EAL) was investigated using a combined time-dependent density functional theory and molecular mechanics (TD-DFT/MM) approach. Excited state potential energy surfaces (PESs), constructed as a function of axial bond lengths, were used to understand the photo-cleavage of the Co-C bond and to elucidate the mechanism of photodissociation for AdoCbl inside the enzyme. The S1 PES is characterized by two minima regions namely, metalto-ligand charge transfer (MLCT) and ligand field (LF) states, which are key minima regions along the reaction pathway. There are two possible routes for photolysis of AdoCbl inside EAL named Path A and Path B. Path B is slightly more energetically favorable than Path A and involves the elongation of Co-Nax bond followed by the elongation of both axial bonds Co-C and Co-Nax. To further understand the effect of environment on the formation of RP, the photochemical data for (AdoCbl)-dependent EAL was also compared with base-on and base-off AdoCbl in solution.

Keywords: Photochemistry, enzymatic catalysis, Co-C bond cleavage, radical pair, potential energy surface, QM/MM calculations

2 ACS Paragon Plus Environment

Page 3 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1. Introduction Coenzyme B12 (Adenosylcobalamin = AdoCbl, Figure 1) acts as cofactor for various enzymatic reactions including carbon skeletal rearrangement, heteroatom eliminations, and intermolecular amino group migration.1-4 It is a highly complex organometallic molecule containing a 5`-deoxy-5`-adenosyl (Ado) group as the upper axial ligand and a 5,6dimethylbenzimidazole (DBI) ligand as the lower axial base. The key feature of AdoCbl cofactor is a Co-C σ bond between Co(III) and Ado ligand.5 It is generally accepted that the Co-C bond of enzyme-bound AdoCbl is not significantly activated without the presence of a substrate.6-7 Binding of substrate triggers cleavage of the Co-C bond leading to the formation of Co(II) and Ado• radical pair (RP). Subsequently, the Ado radical abstracts a hydrogen atom from the substrate (Sub) to generate a stable Co(II)/Sub• RP (Scheme 1a), which is the first detectable intermediate via electron paramagnetic resonance (EPR) spectroscopy.8-12 The cleavage of the Co-C bond is rapid and it is significantly enhanced inside the enzymatic environment when compared to the homolysis of free AdoCbl in solution.13-15 Furthermore, cleavage of the Co-C bond and H-atom abstraction steps are coupled, as revealed by a kinetic isotope effect (KIE) study.6, 16-20 The most remarkable aspect of (AdoCbl)-dependent catalysis is the ~1012 observed rate enhancement inside the enzyme.5, 12, 21-24 However, the mechanism of Co-C bond activation and the factors responsible for the observed rate acceleration (~1012) in the enzymatic environment, remain elusive. AdoCbl cofactor also possesses complex photolytic properties which are mediated by its low-lying excited states.25-36 The photodissociation of the Co-C bond occurs in the presence of light, when the corresponding wavelength ranges between 530 – 300 nm.37-38 Specifically, AdoCbl undergoes photolysis upon excitation at wavelengths shorter than ~600 nm which generates cob(II)alamin and corresponding Ado radical (Scheme 1b).29,

33, 39

In contrast to

methylcobalamin (MeCbl), where the quantum yield of Co-C bond cleavage is wavelength dependent,33, 40-41 photolysis of AdoCbl is wavelength independent.30-31 The specific mechanism of Co-C bond photodissociation depends on a number of factors. The most decisive is the nature of the cofactor’s environment, whether in solution or complexed within an enzyme.42-45 For instance, in strongly acidic conditions the protonated DBI axial base of coenzyme B12 is replaced 3 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 39

by water forming the base-off conformation,46 which has very different photolytic properties in comparison to its base-on analogue.47 The ability to probe photolytic cleavage of the Co-C bond for enzyme-bound AdoCbl is of particular relevance in enzymatic catalysis (Scheme 1b).11-12, 39, 48-49 In (AdoCbl)-dependent enzymes, homolysis of the Co-C bond is triggered by a substrate, which is kinetically coupled to the subsequent H-atom abstraction from substrate.19, 50 In contrast, photolysis of AdoCbl inside the enzymatic environment, creates cob(II)alamin and Ado RP (Co(II)/Ado RP) without coupling to the hydrogen abstraction step from the substrate. The photodissociation of the Co-C bond in enzyme-bound coenzyme B12, has been demonstrated for (AdoCbl)-dependent glutamate mutase (GLM),51-52 (AdoCbl)-dependent ethanolamine ammonia-lyase (EAL).53-54 A recently discovered and structurally characterized photoreceptor CarH, also uses the photolysis of AdoCbl to regulate the biosynthesis of carotenoids, where photolysis of the Co-C bond initiates a large scale conformational change of the CarH protein and subsequently promotes disassembly of CarH tetramer that is directly attached to the DNA.55-57 The photolytic properties of B12 cofactors and photo-generation of the Co(II)/Ado RP have been investigated using variety of experimental techniques.30-35, 40, 43, 47, 51-54 These studies have been carried out in solution as well as in enzyme-bound cofactors. For (AdoCbl)-dependent GLM, it has been shown that the protein environment appears to influence the excited state of the cofactor rather than the ground state.51-52 Significant changes have been observed in the rate constant for the recombination of post-homolysis products in the enzymatic environment when compared to the solution medium. It has been suggested that the protein acts as a sort of cage to prevent diffusion of the Ado radical. This is evidenced by an increase of the recombination rate of the RP in GLM.51-52 As a result of this, the quantum yield for longer lived RP is reduced as more RP tend toward geminate recombination. In studies of AdoCbl-dependent methylmalonyl CoA mutase (MCM), it was proposed that the post-homolysis product is stabilized.24, 58 Several explanations for this stabilization have been proposed based on spectroscopic techniques

52, 58-59

but there is still not a clear consensus for how this works. In the case of (AdoCbl)-dependent EAL, it has been suggested that binding of the substrate analogue to the active site does not increase the quantum yield of RP from the photodissociation of Co-C bond cleavage.54 The result also implied that no structural changes occurred in protein configuration to cleave the Co-C bond

4 ACS Paragon Plus Environment

Page 5 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

upon substrate binding.53-54 Thus, the stabilization of photo-generated Co(II)/Ado RP was explained based on the contribution of protein configuration during Co-C bond cleavage.53 Despite all of these experimental efforts there is still much to be learned about the role of the enzymatic environment in regards reactivity and the stabilization of the RP in AdoCbl-dependent enzymes. In addition to these experimental studies, significant theoretical work has been completed to explore photolytic properties of cobalamins37-38,

41, 45-46

. Specifically, Density Functional

Theory (DFT) and Time-Dependent DFT (TD-DFT) have been applied to elucidate their electronically excited states. Reliable structural models with various axial ligands have been used to construct potential energy surfaces (PESs) as a function of axial bond lengths. Photolytic properties of cobalamins are mediated by low-lying excited states and the crucial step is the RP generation from the first excited state (S1). It was found that the photodissociation initiates from the metal-to-ligand charge transfer state (MLCT) at essentially unchanged axial bonds and proceeds through the ligand field (LF) state at elongated axial bond distances to form the Co(II)/Ado RP. Two possible pathways (Path A and Path B) were identified on S1 surfaces for the radical photo-generation. Specifically, Path A, where the Co-C bond elongates first, is active for the base-on form of AdoCbl, whereas Path B is active for the base-off form (with water as the lower axial ligand), where the Co-O bond elongates first. Also, the semi-classical Landau-Zener theory has been applied to explore the possibility of the involvement of triplet states.46 While the studies described above provided valuable insight about the photochemical and photophysical properties of the isolated AdoCbl cofactor in solution, the photolytic properties have not been investigated theoretically inside (AdoCbl)-dependent enzymes such as EAL. Let’s recall that EAL belongs to a class of B12-dependent enzymes found in bacteria and archaea, which uses AdoCbl to initiate the radical reaction.60-62 In presence of the substrate, ethanolamine (EA), homolysis of the Co-C initiates the formation of Co(II)/Ado RP.7 In contrast, a similar RP can be generated upon radiation by visible light and thus it has been suggested that photolysis of the Co-C bond inside an enzyme can mimic thermal cleavage of the Co-C bond in enzymatic catalysis. The main purpose of this present study is to explore the photolysis of (AdoCbl)dependent enzyme EAL, through careful examination of its ground and low-lying excited states via a combined DFT/MM and TD-DFT/MM approach, respectively. To the best of our

5 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 39

knowledge, this is the first photochemical study of (AdoCbl)-dependent EAL that employs the use of TD-DFT/MM methodology.

2. Computational Details The (AdoCbl)-dependent EAL enzyme was selected to explore how the Co-C bond is photo-cleaved based on the availability of a high-resolution crystal structure with coenzyme B12 analogues and substrates and photochemical data for direct comparison with calculations. The EAL enzyme from E. coli was crystalized at a resolution of 2.25 Å (PDB id: 3ABS)62 with two asymmetric units. The asymmetric units are identical, and the reactive parts, the cofactor and substrates, are independent to their respective subunits and do not interact. One asymmetric unit of EAL is composed α and β subunits with adeninylpentylcobalamin (AdePeCbl) instead of AdoCbl cofactor and EA as a substrate (Figure S1, Supporting Information). The x-ray structure is available only for the substrate-bound (AdoCbl)-dependent EAL. However, the EA substrate is small, and it has been suggested that binding of EA does not lead to significant structural changes.11, 53, 63 It is important to note that large structural changes are not the only thing to take into consideration when building a proper model for this study. In EAL there are other subtle conformational changes that can occur and contribute to the various aspects of AdoCbldependent catalysis. For instance, the crystal structure that was modified for this study contained a glutamate residue (E287) that is in van der Waals contact with the Ado group.62 This residue E287 appears to have a role in native catalysis as well as in the photoexcitation and the low lying excited states of EAL in the presence and absence of substrate EA.39,

64

Thus, removal of

substrate EA and the minimization of the structure without the substrate and the inclusion of E287 in the low layer model ensure that our model is a fair description of the photochemical cleavage of Co-C bond in EAL and are not significantly affected by the residue E287.

2.1 QM/MM Setup

6 ACS Paragon Plus Environment

Page 7 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

To prepare the QM/MM input, the crystal structure of EAL (PDB id: 3ABS) was obtained from the Protein Data Bank and contains two asymmetric units (Figure S1a). In the reported crystal structure, the AdePeCbl structural analogue of coenzyme B12 is bound to the protein at the interface of two subunits, where one subunit covers the upper part while second covers the lower part of the cobalamin.62, 65-66 Only a single asymmetric unit was used in the calculations (Figure S1b, Supporting Information). The AdePeCbl binds in the base-on configuration where the lower axial ligand is DBI in the presence of substrate EA (Figure 2a). The structure was then protonated using PropKa 3.0 software67 and the protonation states of the titratable residues were determined by manual inspection. The AdePeCbl was modified by adding a ribose moiety to the purine ring to restore the actual structure of AdoCbl cofactor (Figure 2b). Finally, substrate EA was removed from the structure and two water molecules were added to fill the empty cavity (Figure 2b). The substrate was removed in order to ensure that CoC bond activation would be based on light and not substrate binding. The modified EAL structure was then minimized with the MM level of theory using amber force field in UCSF Chimera.68 The minimized structure (Figure 2b), was divided into three layers. The corrin ring, imidazole (Im) part of the DBI base with the side chain replaced by hydrogen, cobalt atom and the Ado moiety, were added to the high layer system (Figure S2a, Supporting Information). The remaining part of coenzyme B12 with the additional water molecules was placed into the middle layer and the rest of the protein, including the crystal water, was placed into the low layer system. Atoms within 20 Å from the cobalt center were kept unfrozen and the rest of the protein was frozen. The model used in calculations contained a total of 11988 atoms, with 3550 unfrozen and 8438 frozen. All reported QM/MM calculations were performed using the Gaussian 09 software.69

2.2 DFT/MM Calculations The high layer part of the system, where the Co-C bond cleavage takes place, was computed using DFT with BP86 functional (please see below for explanation of functional choice). To maintain consistency with our previous calculations, the TZVP basis set was used for the H and TZVPP was used for the Co, C, N, and O in QM layer. The middle layer of the system was computed using semi empirical (PM6) level of theory while the protein part of the system 7 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 39

was treated with the Amber force field (FF). Amber parameters for the AdoCbl cofactor were obtained from Marques et al.70 Mechanical embedding was chosen for the ONIOM calculation.71 QM(DFT)/PM6/MM(Amber) calculations were performed to optimize the structure of the EAL and default optimization criteria was used in the calculations. Then, using the optimized geometry, the ground state (S0) potential energy curves (PECs) were constructed as a function of Co-C bond distance. Figure 3 contains PECs for (a) AdoCbl inside EAL and (b) the base-on model complex in solution denoted, Im-[CoIII(corrin)]-Ado+. Figure S3, contains PECs for (a) AdoCbl inside EAL and (b) the base-off model complex in solution denoted, H2O-[CoIII(corrin)]Ado+. The 3D S0 potential energy surfaces (PESs) for AdoCbl inside EAL, Im-[CoIII(corrin)]Ado+, and H2O-[CoIII(corrin)]-Ado+, were also constructed by optimizing the geometries using the QM(DFT)/PM6/MM(Amber) level of theory as a function of Co-C and Co-NIm bond lengths (Figure 4). The BP86 functional was chosen for the QM layer portion of the calculations. Several benchmark studies have been carried out for Co-C bond dissociation15,

72

as well as for

electronically excited states.73-76 Based on those calculations, we have confidence that the BP86 functional is appropriate level of theory. While, DFT and TD-DFT calculations possess inherent tendency toward certain errors, one cannot provide the corresponding error bars. In our previous studies of isolated AdoCbl in solution, we have shown that pure functionals produce less error compared to the hybrid functionals in DFT when estimating bond dissociation energies.71 This is why we have chosen BP86, a pure functional, for this study. It provides good agreement with experimentally determined bond dissociation energies13, 15, 23, 72 as well as good agreement with relevant structural details like bond lengths and angles from crystal structures (Table S1). In this way we are confident in the results produced by the methodology we employed.

2.3 TD-DFT/MM Calculations The optimized ground state (S0) geometry was then used for the TD-DFT/MM calculations to obtain the low-lying singlet and triplet states. By using the S0 geometry, we applied single point TD-DFT calculations to construct the S1 PESs (Figure 4) as a function of Co-C and Co-NIm bond distances. The detailed orbital analysis was also performed to characterize the nature of electronic excitations. Due to the problems associated in ONIOM 8 ACS Paragon Plus Environment

Page 9 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

calculations to combine the orbitals in high layer and medium layer, additional single point calculations were performed to obtain the molecular orbitals. The MLCT region of the S1 surface was then characterized to have the excitations from the Co d-orbitals to the corrin π* whereas in the LF region of the S1 surface the electronic excitations occurs from the Co d and π corrin orbitals to the dz2 orbitals of cobalt. In many applications of TD-DFT methodology, a concern arises regarding potential underestimation of long-range charge-transfer (LR CT) excitations. This underestimation is usually the result of poor overlap between the relevant occupied and virtual orbitals involved in LR excitations. In the case of isolated AdoCbl, the LR CT problem would potentially occur with electronic excitations involving the Ado and corrin moieties. To ensure that TD-DFT/MM energy surfaces were not affected by the LR CT problem, careful analysis of corresponding electronic excitations was performed by applying several different tests and diagnostics. For isolated AdoCbl in solution the Λ diagnostic was applied.77 The Λ parameter has a value from 0 to 1 and for GGA functionals, this should be greater than 0.4 otherwise, excitations are likely to be significantly underestimated. The nature of low-lying excited states was also investigated by applying different QM and MM partitions because Ado and corrin ligands are spatially separated. Specifically, the Ado ligand was further partitioned into ribose and purine rings, and only the ribose ring was placed in high layer with the cobalt atom, corrin ring and the Im part of the axial base. The rest of the cofactor, including the purine ring, was then placed into medium layer and the protein part into the low layer. We are referring to this structure as RibCbl (Figure S2b). TD-DFT/MM calculations for RibCbl, revealed that the PES corresponding to the S1 state (Figure S4) looks very similar to the corresponding S1 PES where the full geometry of the Ado group was used in the high layer model system (Figure 4). The insensitivity with respect to the QM/MM partition reflects the nature of electronic transitions which are primarily localized within cobalt and coring ring. Finally, the analysis of LR CT was performed using a densitybased matrix (DCT)78. It is a strong diagnostic tool for the LR CT problem in TD-DFT calculations. The matrix is the spatial distance between the barycenter’s of electron depletion (R+) and electron augmentation (R-) zone during the electronic transitions (DCT = |R+-R-|). Excited states involving the charge transfer in greater than 6.0 Å DCT length were excluded from

9 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 39

calculated spectra due to possible LR CT character. Gaussian 16 was used for the DCT calculations79. A more detailed analysis and discussion will be presented in the next section.

3. Results and Discussion 3.1 Ground State Geometry The modified structure of substrate-free (AdoCbl)-dependent EAL, where the ribose part of the Ado was reconstructed, has been optimized using the procedure described above (Figure 2b). Frequency calculations confirmed that this was the optimal geometry, as no imaginary frequencies were present. The structural parameters of the optimized geometry were compared with the experimental values from the crystal structures of EAL (PDB id: 3ABS). In addition, the results were also compared with experimental data from the high-resolution crystal structure of AdoCbl (Figure S5) as well as with the geometrical parameters obtained from DFT/BP86 calculations for the base-on and base-off model complexes (Figure S5) in solution from previous computational studies (Table S1). When comparing the crystal structure of EAL to the optimized geometry, some structural differences were observed regarding the corrin ring and the axial bond distances. The axial Co-C and Co-NIm bond lengths in the optimized structure of AdoCbl in EAL were 2.05 Å and 2.37 Å, respectively (Table S1). The experimental values for these bond lengths in the enzyme were 2.01 Å for the Co-C and 2.61 Å for the Co-NIm bonds (Table S1). These structural changes were most likely induced due to the replacement of the AdePeCbl moiety with AdoCbl. To further explore the differences associated with the lower axial base in EAL, the CoNIm bond distance was compared to the results obtained from the calculations of the base-on and base-off forms of AdoCbl model in solution (Figure S5). In the case of the base-on form where the lower axial ligand is DBI, the Co-NIm bond is 2.19 Å compared to 2.37 Å in EAL. In the base-off form where DBI is replaced by water, the Co-O bond is 2.43 Å. These results indicate that the axial bond distances are sensitive to environment, whether the cofactor is in solution or within an enzyme.

10 ACS Paragon Plus Environment

Page 11 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

The most significant differences observed between AdoCbl inside EAL and the base-on and base-off forms of AdoCbl in solution were the conformations of the ribose ring of the Ado ligand. The structure of the Ado ligand consists of two units, purine and the ribose ring (Figure 1). The relative orientation of the ribose with respect to the purine ring can be explained based on the glycosyl rotation angle (χCN) (Figure S5). The ribose conformation can be explained using pseudorotation phase (P) and pseudorotation amplitude (ϴm) (for definition of χCN, P and ϴm see Figure S6). Note that the pseudorotation values consisting of -90° to 90° corresponds to the 3՛ – endo conformation of the ribose whereas the P with 90° to 270° corresponds to the 2՛ – endo conformation.80-81 To understand the influence of the enzymatic environment on the ribose conformation, the structural data of the ribose, extracted from (AdoCbl)-dependent EAL optimized structure, was compared to optimized geometries of base-on and base-off forms of isolated AdoCbl (Table S2). Due to the absence of the ribose ring in the x-ray crystallographic data of EAL,62 it was not possible to make direct comparison with the optimized structural model of substrate-free (AdoCbl)-dependent EAL. However, it was found that the ribose ring in all the structures were in 3՛ –endo conformation. Interestingly, the 3՛ –endo ribose conformation obtained as a result of DFT/MM geometry optimization for EAL, is the same in another (AdoCbl)-dependent enzyme, namely GLM, where reported crystal structure contains a mixture of two conformers 2՛ –endo and 3՛ –endo, respectively.82-83 The most significant difference observed was with respect to the glycosyl rotation angle (χCN) when compared with the x-ray structure of AdoCbl and the calculated base-on and base-off forms. In the case of AdoCbl inside EAL, the χCN = -72.46° whereas in the crystal structure of isolated AdoCbl χCN = 68.00°. On the other hand, in the optimized models of base-on and base-off AdoCbl corresponding χCN angles were -13.69° and -92.33°, respectively. Although our model does not contain the substrate we believe that these significant differences observed in glycosyl rotation angles when comparing the structures of AdoCbl in solution to AdoCbl inside EAL can help to understand the labilization of the Co-C bond in the presence of substrate. There is still much to be understood about how the protein environment contributes to this. We feel as though the results from our electronic structure calculations corroborate the idea that the glycosyl rotation angle of the ribose moiety is an important factor to

11 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 39

be considered when elucidating the mechanism of homolysis of the Co-C bond in AdoCbldependent enzymatic catalysis.

3.2 Electronically Excited States Along Co-C Bond To provide the preliminary insight into the photodissociation of AdoCbl inside EAL, the S0 energy curve (Figure 3a) was generated by systematically stretching the Co-C bond with a step size of 0.05 Å within the range of 1.80 Å to 2.60 Å. The S0 geometry of AdoCbl inside EAL was optimized at each point. Then, single point TD-DFT calculations were carried out to obtain the manifold of low-lying singlet and triplet excited states. For each optimized geometry, 30 excited states were calculated up to the energy values of 3.1 eV in EAL and compared with the excited states obtained for the isolated base-on and base-off forms of AdoCbl (Figure 3 and Figure S3) from previous studies. Overall, energy curves computed inside EAL (Figure 3a) are not significantly different when compared with the isolated base-on (Figure 3b) and base-off (Figure S3b) forms of AdoCbl, although some differences are noticeable. The most apparent similarity in all the cases is the repulsive triplet state (σ  σ*). This state drops in energy and at bond lengths of approximately 2.4 Å, it becomes the lowest energy state. As discussed before, this is due to the single determinant wave function used in TD-DFT calculations. This state would correctly level off in a multi-reference wave function approach.37, 41, 46 Eight lowest singlet excited states were characterized based on the orbital analysis as summarized in Table 1 and corresponding figures for these orbitals can be found in Figure S7. For the low-lying singlet excited state S1 inside EAL enzyme, the first vertical excitation was found at 1.97 eV, whereas for the base-on and base-off forms of isolated AdoCbl, the calculated excitation energies did not exceed 2.30 eV and 2.38 eV, respectively. In the enzymatic environment, the energy of vertical excitations are significantly smaller when compared to those values in solution and this observation is consistent with experimental results. The S1 state involves electronic transitions that are 59% dxz/dz2 + π  π*, 23% dyz+ π  π* and 11% dxz+π π* (Table 1).



The S1 electronic state can be characterized as arising from the MLCT state

because all the mentioned transitions are predominantly from the metal d orbital to the corrin π* orbital. This is consistent with previous theoretical studies for alkyl cobalamins. The next excited state S2 can be characterized as 64% dyz+ π



π*, 21% dxz/dz2+ π



π*, and 4% dxz/dz2+ π



π* 12

ACS Paragon Plus Environment

Page 13 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

transitions. In excited state S3, the major contribution is an excitation from HOMO-5 to LUMO which has 88% dxy  π* character. This is in good agreement with the results of calculations for base-on model complex of AdoCbl but for the base-off form the antibonding σ* character was observed in excited state S3. Antibonding σ* character which is an electron excitation from HOMO-1 to LUMO+1 in EAL, is observed in the S4 state and up. The major component of transition contributions in the S4 excited state is 7% dyz+π  dz2, 12% dxz/dz2+ π  π*, 65% dxz+ π 

π*, and 7% dxy  π*. Significant amount of σ* character occurs in S5 state where the transition

contribution is 67% πAdo  dxy+n+π*, and 26% dxz+π  dz2. The other significant difference with base-on and base-off AdoCbl, is the absence of pure πAdo  π* transition in presented results of calculations. The energies of excited states reported here (Table 1) are typically lower than the energies in base-on and base-off AdoCbl in solution. This significant reduction in excitation energies agreed well with the idea of lowering of energies induced by the enzymatic environment.24, 52, 58, 84

3.3 Potential Energy Surfaces as a Function of Axial Bond Lengths Cobalamins have been the subject of several structural investigations using TD-DFT and the most significant structural changes upon electronic excitations were noted for axial bonding, while very negligible changes were associated with the corrin ring. Similar behavior was also observed experimentally for cyanocobalamin (CNCbl). Using XANES spectroscopy,85 it was found that the axial bonds namely, Co-CN and Co-NIm, underwent the most significant changes while the corrin ring structure was maintained. Thus, in order to simulate corresponding spectra for direct comparison with experiment, it is sufficient to only adjust the axial bond lengths in models. Previous theoretical studies have revealed that, to explore the photolytic properties of cobalamins, it is crucial to construct a PES as function of both Co-C and Co-Naxial bond distances.37-38, 41, 46 For (AdoCbl)-dependent EAL, the S0 PES (Figure 4a) was constructed using DFT/MM by systematically elongating the Co-C and Co-NIm axial bonds using step-size of 0.10 Å. The manifold of low-lying singlet S1-Sn excited states was obtained from TD-DFT/MM calculations starting with the S0 PES. The details regarding DFT/MM as well as TD-DFT/MM methodology can be found in the Computational Section.

13 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 39

The major feature of the S0 PES is the single energy minimum, which is consistent with the base-on and base-off forms of free AdoCbl. The corresponding lowest energy region is around Co-C of 2.05 Å and Co-NIm bond at 2.37 Å (S0, Figure 4a). On the other hand, the S1 PES is characterized by two energy minima separated by a seam. The first minimum localized at CoC equals 2.10 Å and Co-NIm at 2.30 Å (I(S1min), Figure 5a) corresponds to a MLCT state. The second minimum localized at Co-C bond length of 2.50 Å and Co-NIm bond length of 2.90 Å (IIIB(S1min), Figure 5a), is characterized as LF state. The transition contributions in the minimum of the MLCT region, I(S1min), are 52% dxz+ π



π* and 33% dyz+π



π* (Figure S8). For the

other MLCT minima (IIA(S1min), Figure 5a), that is associated with Path A, the HOMO to LUMO (dxz/dz2 + π  π*) transition is dominant (Figure 6 and Figure S8). The excitations at IIB (MECP) are HOMO to LUMO+1 and have been characterized as dyz+π



dz2 (Figure 6). In the

LF state, the HOMO to LUMO transition is dominant. The excitation contribution for the LF state is 100% dyz+π



dz2 at IIIB(S1min) (Figure 6). The corresponding ground state geometries

for each of these minima regions (I(S1min), IIA(S1min), and IIIB(S1min) are depicted in Figure S9. Apart from the character of the excitations, the main difference for the MLCT (I (S1min)) and LF (IIIB (S1min)) minima is in their relative energy. Namely, the LF region is 1.0 kcal/mol lower than the MLCT, which is consistent with experiment.32, 34, 42 The analogous PES for S1 excited states of base-on (Figure 4b) and base-off AdoCbl (Figure 4c) are also characterized by two minima. Comparing all the cases, some similarities and differences were observed. The MLCT minimum in base-on AdoCbl is lower in energy than the LF minimum, whereas in the case of base-off AdoCbl, the MLCT is higher in energy than the LF minimum. In base-on AdoCbl, the seam separating the minima is more pronounced which indicates the high barrier between MLCT and LF minima. Alternatively, the S1 PES of base-off AdoCbl and the S1 PES of AdoCbl inside the enzyme correspond to a very similar topology, where the LF region is energetically more stabilized. It appears that the enzymatic environment plays the role to stabilize the LF electronic state.

3.4 Analysis of Long-Range Charge Transfer Excitations As has been already pointed out in the Computational Section, TD-DFT calculations often suffer from the potential underestimation of the LR CT-type excitations. Taking into 14 ACS Paragon Plus Environment

Page 15 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

account that the BP86 functional was used in calculations, such underestimation could have a potentially negative impact on the proper description of the S1 PES. Specifically, for isolated AdoCbl, the LR CT-type excitations are typically associated with electronic transitions between the Ado and corrin ligands. To make sure that this is not the case, careful analysis of electronically excited states has been carried out. For the S1 surface, the corresponding orbital contributions were analyzed and additional TD-DFT/MM calculations with different QM/MM partitions were performed. The analysis was carried out for the S1 PES based on TD-DFT/MM calculations with respect to the Ado



corrin transitions. When comparing the S1 surface where

the upper ligand was only Rib (Figure S4), to the S1 surface with the full Ado ligand (Figure 4a), it was shown that the topology of the two surfaces look very similar, although the relative energies of the MLCT and LF minima are slightly different. For the AdoCbl in EAL S1 PES (Figure 4a), the sections where axial bonds were noticeably longer than the equilibrium geometry, additional analysis of LR CT transitions was also performed using the DCT diagnostic tool. The excited states involving charge transfer that were greater than 6.0 Å DCT length were not included in the construction of the S1 surface. Taking all the above considerations into account we have confidence that the S1 surface for (AdoCbl)-dependent EAL based on TDDFT/MM methodology was correctly constructed.

3.5 Pathways for Photodissociation Photolysis of AdoCbl inside EAL to generate Co(II)/Ado RP should primarily occur from the LF state, similar to the base-on or base-off forms of AdoCbl (Figure 7). It is also expected that the photodissociation mechanism for AdoCbl in EAL should not significantly differ from free AdoCbl in solution because for both scenarios the cleavage of the Co-C bond occurs from the LF region on the S1 state. From the LF region, regardless of environment, either geminate recombination or radical pair separation is possible, although the energetics and the quantum yield of photoproducts associated with enzyme-bound AdoCbl should be different than AdoCbl in solution. In contrast to base-on AdoCbl in solution, where the MLCT state is energetically more stable, inside EAL the LF state is stabilized by 1.2 kcal/mol more than the MLCT state, which is in line with the base-off AdoCbl in solution (Figures 5 and 6). The main challenge in describing the photodissociation mechanism is connecting these two states via a minimum 15 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 39

energy path. Previous theoretical studies for various cobalamins have revealed that, based on energetic grounds, two possible photodissociation pathways, referred as Path A and Path B, can be distinguished.37, 41, 46 Path A involves the initial elongation of the Co-C bond followed by CoNaxial, while Path B involves the lengthening of Co-NIm bond prior to the elongation of Co-C. Likewise, the photodissociation mechanism of AdoCbl inside EAL can also be associated with two paths. Both pathways start from the energy minimum of MLCT state and proceed through the seam to the LF state, from which final photo-cleavage takes place. Path A initiates from the MLCT minima I and proceeds to the IIA of the MLCT minima and then overcomes the barrier of the seam to go to the LF state. Path B also initiates from the MLCT minima I and proceeds through the seam IIB to go into the LF minima at IIIB (Figure 5a). To determine which pathway is active, the energetic barriers between the minima and the seam were considered. The energy barriers were plotted as a function of energy vs the Q value Q = R + R For Path A, the energy barrier is 3.48 kcal/mol at Q value 3.37 and for Pathway B it is 2.48 kcal/mol at Q value 3.47 (Figure 5b). Considering the energy profiles of both paths, we identified that Path B is more energetically favorable and should be considered as the active path for the photodissociation. Previous studies have shown that in the case of base-on AdoCbl in solution, Path A is active due to the low energy barrier where photodissociation initiates from MLCT state and proceeds through the elongation of Co-C bond to LF state37, whereas in the base-off AdoCbl, Path B is active for photodissociation and this is more in line with our study in the enzymatic environment. In both cases, i.e. for base-off form of free AdoCbl in solution as well as for AdoCbl in the enzymatic environment, the LF state is more stabilized than MLCT state.

3.6 Implications of AdoCbl Photochemistry Inside Enzyme When comparing the S1 PESs of AdoCbl inside EAL to isolated AdoCbl in solution it is of note that the AdoCbl S1 PES adopts features from both of the S1 PESs for base-on and baseoff forms of isolated AdoCbl (Figure 4). The topology of the S1 PES for AdoCbl and base-on AdoCbl in solution are very similar as both contain two energy minima in approximately the 16 ACS Paragon Plus Environment

Page 17 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

same locations. However, those two surfaces are not identical from an energetic point of view. The LF state is lower in energy for AdoCbl inside EAL than in the base-on form. When comparing the S1 PES of AdoCbl in EAL to the base-off form, the topologies are not as similar, but the energetics are more in line with each other. It seems as though the S1 PES of AdoCbl in EAL can be considered as a superposition of the base-on and base-off S1 PESs of the isolated cofactor. The optimum energy paths for photolysis in the base-on and base-off forms of AdoCbl in solution are different, with Path A being active for the base-on and Path B active for the base-off form (Figure 7). For photolysis of AdoCbl inside EAL, the minimum energy path is a mixture of the optimum paths for the two forms of isolated AdoCbl. Dissociation ultimately occurs through the LF region, as is the case of AdoCbl in solution. Overall, based on energetics, Path B is more favorable than Path A by 1.0 kcal/mol for AdoCbl in EAL (Figure 5b). Path B begins with only the elongation of the Co-NIm of the axial base followed by simultaneous elongation of both axial bonds, Co-C and Co-NIm until the LF minimum (IIIB) is reached. Although Path B is slightly more energetically favorable based on our calculations we cannot rule out that Path A may also be active as well.

3.7 Comparison with Experiment Alkyl cobalamins, such as MeCbl or AdoCbl, upon excitation with visible or near UV light, generally exhibit rapid internal conversion (IC) to the lowest excited electronic state followed by cleavage of the Co-C bond.31,

45

Geminate recombination limits the ultimate

photolysis yield. Following the experimental work of Sension et al., and other groups including Warncke, et al., and Scrutton, et al, our current understanding of photochemical behavior of AdoCbl can be summarized as depicted in Scheme 2. Excitation at 400 or 520 nm results in rapid IC to the lowest excited electronic state (S1) and the nature of the S1 state observed in the experiments depends on the environment of AdoCbl. Our TD-DFT/MM calculations indicate that in the case of (AdoCbl)-dependent EAL this initial excitation state can be associated with the S4 state as the transition dipole moment is noticeably larger (i.e., order of magnitude) when compared to other energetically similar low-lying excited states (Note that the S2 state has a noticeable value of oscillator strength, although it has low energy versus experiment). The S4 17 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 39

state, with an excitation wavelength of 530 nm, has the most significant contribution (65%) coming from HOMO-4



LUMO electronic excitation and can be described as a dxz+ π



π*

transition (see Table 1). Calculated excitation energy to the S4 state corresponds well to the maximum of experimental α/β-band with λmax of 527 nm.54 The rapid conversion from the initial excited (S4) to a species that resembles the cob(II)alamin on the S1 PES involves several photoproduct intermediates, which can be readily characterized with our TD-DFT/MM calculations (Scheme 2). The first intermediate, denoted as AdoCbl(=MLCT), is the MLCT state. There is subsequent relaxation to another intermediate, AdoCbl(=LF), which is the LF state. These changes can be associated with the gradual elongation of axial bond distance. The LF state is primarily responsible for homolytic cleavage of the Co-C bond, which produces the Co(II)/Ado RP. At this point, according to experiment, there is a competition between cage escape and geminate recombination with noticeable differences when these processes are compared between solvent and enzyme.31,32-34,

54

This

competition dictates the quantum yield of photo-products.32 When AdoCbl is in water, 75-80% of RP formed in the LF region (AdoCbl(=LF) Scheme 2) tend toward geminate recombination on a time-scale less than 1 ns while the others proceed towards diffusion and cage escape.31-32, 36, 86 The tendency toward geminate recombination is increased in the enzymatic environment to 9095%.31, 39, 49, 51-52, 54 Alternatively, in solution, the geminate RP exists within a solvent cage and the RP can diffuse apart from the jointly occupied solvation shell in a process known as cage escape. In an enzymatic environment, the quantum yield is significantly lower than the isolated cofactor in solution because of the higher recombination yield of Co(II)/Ado RP to Co(III).51-53 All this can be monitored experimentally by UV-vis spectroscopy, but it is difficult to model computationally without proper analysis of kinetic equations. Additional explanation requires consideration of the excited state photoproduct relaxation to the ground state and the differences associated with whether or not the cofactor is in solution or in an enzyme. For isolated AdoCbl in solution two possible channels for deactivation have been proposed. One involves the corrin ring distortion after detachment of axial lower ligand (base or water) while the second involves electronic de-excitations of Co(II). Based on current TD-DFT/MM calculations we postulate that only the second one is operational in the case of (AdoCbl)-dependent EAL. In the base-off form of AdoCbl the intermediate product is in the LF

18 ACS Paragon Plus Environment

Page 19 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

excited state (Figure 7c). From this state conversion to the ground state can occur by bending the corrin ring, similar as in the base-off form of methylcobalamin.46 This mechanism is possible when cobalt is in the Co(III) state with the axial methyl or adenosyl ligand attached. Deactivation takes place with Co-C shortening and the corrin ring bending. In the base-on form the deactivation mechanism is different. Co-N and Co-C bonds are significantly elongated. On the axial axis NaxLCoLCax system acquires the character of a radical pair Co•L•Cax, where Co(II) is in the excited state of low energy, which can undergo fast deactivation to the ground state,87 followed by recombination of Co• and R• and subsequent connecting of the axial base. To summarize, we believe there are at least three important things to take into account when considering the formation and recombination of RP in AdoCbl in solution and in an enzyme including, ligand field stabilization, cage effect, and the mechanism of deactivation to S0. Based on energetic grounds, it is easier to generate RP in solution than in an enzyme because the LF is stabilized in the protein environment. This means that the energetics associated with the LF in the enzyme are lower than in solution. However, the rate of recombination is faster in the solution medium.52 In solution, cage effect is a larger contributor to the recombination yield and the quantum yield of photoproducts in that the solution medium allows more freedom for diffusive loss of the RP from the solvent cage.32 There is less opportunity for diffusion in an enzyme.39, 48 There are two mechanisms of deactivation where the geminate RP can recombine to the S0. The first involves corrin ring distortion41, 46 that is more in line with base-off cobalamins in solution.46 The second involves deactivation of Co(II) species to the S0 from the LF by deexcitations from the (dyz)1(dz2)2 excited state returning to a Co(II) (dyz)2(dz2)1.37, 45 For the case of AdoCbl in EAL, it would seem that de-excitations of Co(II) from the LF is the active deactivation route.

3.8 New Perspective for Photochemistry of AdoCbl Inside EAL For AdoCbl inside EAL, the bond dissociation energetic barrier of ~30 kcal/mol reported for isolated coenzyme B12, is overcome to form the Co(II)/Ado RP. Warncke et al. has suggested that this photolysis is a good mimic for cleavage of the Co-C bond in EAL.53-54 To further connect photochemistry of (AdoCbl)-dependent EAL in relation to the native reaction, Warncke

19 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 39

et al. hypothesized that such connection would require the explicit introduction of the protein structural changes that can be associated with Co-C bond cleavage and RP separation.53 Such protein structural changes can be represented by using a protein configuration coordinate that is orthogonal to the coordinate representing Co-C bond stretch. Consequently, a simple twodimensional free energy PES representative of S0 was proposed and was characterized by two minima, Co(III)Ad and Co(II)Ad•. This model was the basis for the identification of two paths, Path 1 and 2, for RP formation in EAL. Path 1 is associated with photolysis and initiates in the Co(III)Ad minimum region and proceeds as Co-C lengthens while the protein configuration is unchanged. Path 2 corresponds to native thermal cleavage of the Co-C bond and is a function of two coordinates, the Co-C bond length and protein configuration. Path 2 initiates from the Co(II)Ad region until reaching the Co(II)Ad• region where dissociation takes place. Based on our calculations, we have determined that photolysis of AdoCbl inside EAL is more in line with Path 2 on the Warncke model in that photolysis involves two minima regions not one. Taking the above information and the outcome of present TD-DFT/MM calculations into consideration we can offer a different interpretation in regards to the photochemistry of AdoCbl inside EAL. First, photolysis of AdoCbl does not occur on the S0 PES. It is apparent from inspection of Figure 4, that photolysis of AdoCbl, regardless of environment, cannot be described by considering changes in only one coordinate. Upon excitation, the most significant structural changes that occur are in axial bonding whereas corrin macrocycle is only slightly perturbed. This cooperativity between the axial bond distances has also been shown experimentally with XANES measurements85 and with several other theoretical studies.84,

88

Based on this evidence, the most important coordinates to consider in AdoCbl photolysis are the axial bond lengths. While we agree that two coordinates are necessary to describe the formation of Co(II)/Ado RP in EAL via photolysis, we suggest that the second coordinate should be the Co-Naxial bond distance instead of protein configuration. Based on Figure 4, which includes S1 PESs as a function of axial bond lengths, it is apparent that regardless of environment, solvent or enzyme, the S1 PES of AdoCbl is always characterized by two energy minima associated with MLCT and LF electronic states where photolysis occurs from the LF state. Thus, the electronic properties associated with AdoCbl are the primary component of photolysis which can be fully described in terms of the cofactor and its

20 ACS Paragon Plus Environment

Page 21 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

axial bonding. In our opinion, protein structural changes do not play a significant role in the photolysis of EAL. This is also apparent on the Warncke model where the protein configuration coordinated is depicted as unchanged in the photolysis pathway. Rather, we believe that the enzymatic environment has the important role of stabilizing the LF state on the S1 surface. The other important role of EAL enzyme in terms of photochemistry may be associated with specific geometric constraints imposed on AdoCbl. In contrast, solution does not impose these geometric constraints and upon dissociation of the Ado ligand, the RP can be separated freely from the corrin moiety as evidenced by the higher quantum yield of RP as compared to the quantum yield in the enzyme.54 While it is true that in the native reaction there are conformational changes that effect the reactivity of AdoCbl inside EAL,39, 89 it is important for us to re-iterate that our model is not an exact replication of the native reaction. Instead, our model is representative of photochemistry. Protein contributions were explored in our model through molecular mechanics (MM), not as a specific coordinate on the PESs. However, based on our combined QM/MM study, we have observed that the protein environment is affecting the energetics associated with photolysis as compared to AdoCbl in solution. Inside EAL the LF, where Co-C bond cleavage occurs, is energetically lower compared to the MLCT state. In contrast, in solution the LF is higher in energy than the MLCT. These observations indicate that the protein environment affects the S1 state and this is being accounted for in our model system.

Summary and Conclusions Simple mechanistic details of AdoCbl photochemistry inside EAL are presented here. The photochemical data is consistent with the previous theoretical studies of isolated AdoCbl, where it was illustrated that there are significant similarities in the topologies of S0 and S1. The S1 surface is characterized by two minima, namely MLCT and LF. The LF is 1.0 kcal less in energy than the MLCT state and hence it is more stabilized, which is consistent with previous experimental work in (AdoCbl)-dependent enzymes. Two possible paths for photodissociation were identified to explain the mechanism of photolysis. Path A is along the Co-C bond distance and is inactive due to the higher energetic barrier. Path B is more energetically feasible as it can surpass the lower energy barrier from the MLCT to LF state. To understand the effect of 21 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 39

environment on the formation of Co(II)/Ado RP, the photochemical data for AdoCbl in EAL was also compared with base-on and base-off AdoCbl in solution. Some similarities and differences were noted in the topology of S1 surface. The S1 surface for AdoCbl in EAL was more in line with the isolated base-off cofactor which is confirmed with the experimental findings. Although the photochemistry of EAL is not physiological necessary for bacteria, these findings will shed light on the still elusive mechanistic details associated with (AdoCbl)-dependent enzymes. In addition to this, we expect that our study can help to understand systems where light is physiologically required such as in AdoCbl-dependent CarH.

Author Information Corresponding Author Email for P.M.K: [email protected] Notes The authors declare no competing financial interest. Supporting Information: Crystal structure PDB images, model structures of AdoCbl and RibCbl cofactor for calculations, PECs for AdoCbl inside EAL and [CoIII(corrin)]−Ado+ model complex, PESs for RibCbl inside EAL, glycosyl rotation angle details, S0 geometries corresponding to key points on the S1 PES, relevant molecular orbitals for lowest vertical singlet electronic transitions. This information is available free of charge at the ACS Publications website.

Acknowledgments: This work was supported by the National Science Centre Poland (UMO2015/17/B/ST4/ 03733). Abdullah Al Mamun would like to acknowledge Brady D. Garabato for his assistance with performing diagnostic calculations. A.A.M, M.J.T, and P.M.K., wish to acknowledge the Cardinal Research Cluster (CRC) at the University of Louisville for providing access to high performance computing facilities.

References:

22 ACS Paragon Plus Environment

Page 23 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

(1) Banerjee, R., Chemistry and Biochemistry of B12. John Wiley & Sons: New York, 1999 (2) Dolphin, D., B12. John Wiley & Sons: New York, 1982. (3) Banerjee, R.; Ragsdale, S. W., The Many Faces of Vitamin B12: Catalysis by CobalaminDependent Enzymes. Annu. Rev. of Biochem. 2003, 72, 209-247. (4) Brown, K. L. Chemistry and Enzymology of Vitamin B12. Chem. Rev. 2005, 105, 20752150. (5) Finke, R. G.; Martin, B. D. Coenzyme AdoB12 vs AdoB12-homolytic Co-C Cleavage Following Electron Transfer: A Rate Enhancement Greater than or qual to 1012. J. Inorg. Biochem. 1990, 40, 19-22. (6) Padmakumar, R.; Padmakumar, R.; Banerjee, R. Evidence that Cobalt−Carbon Bond Homolysis is Coupled to Hydrogen Atom Abstraction from Substrate in Methylmalonyl-CoA Mutase. Biochemistry 1997, 36, 3713-3718. (7) Toraya, T. Radical Catalysis in Coenzyme B12-Dependent Isomerization (Eliminating) Reactions. Chem. Rev. 2003, 103, 2095-2128. (8) Warncke, K.; Utada, A. S. Interaction of the Substrate Radical and the 5'-deoxyadenosine-5'methyl Group in Vitamin B12 Coenzyme-dependent Ethanolamine Deaminase. J. Am. Chem. Soc. 2001, 123, 8564-8572. (9) Wetmore, S. D.; Smith, D. M.; Bennett, J. T.; Radom, L. Understanding the Mechanism of Action of B12-dependent Ethanolamine Ammonia-lyase: Synergistic Interactions at Play. J. Am. Chem. Soc. 2002, 124, 14054-14065. (10) Sandala, G. M.; Smith, D. M.; Radom, L. Modeling the Reactions Catalyzed by Coenzyme B12-dependent Enzymes. Acc. Chem. Res. 2010, 43, 642-651. (11) Bender, G.; Poyner, R. R.; Reed, G. H. Identification of the Substrate Radical Intermediate Derived from Ethanolamine During Catalysis by Ethanolamine Ammonia-lyase. Biochemistry 2008, 47, 11360-11366. (12) Canfield, J. M.; Warncke, K. Active Site Reactant Center Geometry in the Co(II)-product Radical Pair State of Coenzyme B12-dependent Ethanolamine Deaminase Determined by using Orientation-selection Electron Spin-echo Envelope Modulation Spectroscopy. J. Phys. Chem. B 2005, 109, 3053-3064. (13) Hay, B. P.; Finke, R. G. Thermolysis of the Cobalt-carbon Bond of Adenosylcobalamin. 2. Products, Kinetics, and Cobalt-carbon Bond Dissociation Energy in Aqueous Solution. J. Am. Chem. Soc. 1986, 108, 4820-4829. (14) Hay, B. P.; Finke, R. G. Thermolysis of the Cobalt-carbon Bond in Adenosylcorrins. 3. Quantification of the Axial Base Effect in Adenosylcobalamin by the Synthesis and Thermolysis of Axial Base-free Adenosylcobinamide. Insights into the Energetics of Enzyme-assisted Cobaltcarbon bond homolysis. J. Am. Chem. Soc. 1987, 109, 8012-8018. (15) Jensen, K. P.; Ryde, U. Theoretical Prediction of the Co−C Bond Strength in Cobalamins. The J. Phys. Chem. A 2003, 107, 7539-7545. (16) Semialjac, M.; Schwarz, H. Computational Investigation of Hydrogen Abstraction from 2aminoethanol by the 1,5-dideoxyribose-5-yl Radical: A Model Study of a Reaction Occurring in the Active Site of Ethanolamine Ammonia Lyase. Chem. - Eur. J. 2004, 10, 2781-2788. (17) Marsh, E. N.; Ballou, D. P. Coupling of Cobalt-carbon Bond Homolysis and Hydrogen Atom Abstraction in Adenosylcobalamin-dependent Glutamate Mutase. Biochemistry 1998, 37, 11864-11872.

23 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 39

(18) Licht, S. S.; Lawrence, C. C.; Stubbe, J. Thermodynamic and Kinetic Studies on CarbonCobalt Bond Homolysis by Ribonucleoside Triphosphate Reductase: The Importance of Entropy in Catalysis. Biochemistry 1999, 38, 1234-1242. (19) Bandarian, V.; Reed, G. H. Isotope Effects in the Transient Phases of the Reaction Catalyzed by Ethanolamine Ammonia-lyase: Determination of the Number of Exchangeable Hydrogens in the Enzyme-cofactor Complex. Biochemistry 2000, 39, 12069-12075. (20) Jones, A. R.; Hay, S.; Woodward, J. R.; Scrutton, N. S. Magnetic field effect studies indicate reduced geminate recombination of the radical pair in substrate-bound adenosylcobalamin-dependent ethanolamine ammonia lyase. J. Am. Chem. Soc. 2007, 129, 15718-27. (21) Vlasie, M. D.; Banerjee, R. Tyrosine 89 Accelerates Co-Carbon Bond Homolysis in Methylmalonyl-CoA Mutase. J. Am. Chem. Soc. 2003, 125, 5431-5435. (22) Huhta, M. S.; Ciceri, D.; Golding, B. T.; Marsh, E. N. A Novel Reaction Between Adenosylcobalamin and 2-methyleneglutarate Catalyzed by Glutamate Mutase. Biochemistry 2002, 41, 3200-3206. (23) Halpern, J.; Kim, S. H.; Leung, T. W. Cobalt-Carbon Bond Dissociation Energy of Coenzyme B12. J. Am. Chem. Soc. 1984, 106, 8317-8319. (24) Brooks, A. J.; Vlasie, M.; Banerjee, R.; Brunold, T. C. Spectroscopic and Computational Studies on the Adenosylcobalamin-dependent Methylmalonyl-CoA Mutase: Evaluation of Enzymatic Contributions to Co-C Bond Activation in the Co3+ Ground State. J. Am. Chem. Soc. 2004, 126, 8167-8180. (25) Endicott, J. F.; Netzel, T. L. Early Events and Transient Chemistry in the Photohomolysis of Alkylcobalamins. J. Am. Chem. Soc. 1979, 101, 4000-4002. (26) Chagovetz, A. M.; Grissom, C. B. Magnetic Field Effects in Adenosylcob(III)alamin Photolysis: Relevance to B12 Enzymes. J. Am. Chem. Soc. 1993, 115, 12152-12157. (27) Schwartz, P. A.; Frey, P. A. 5'-Peroxyadenosine and 5'-peroxyadenosylcobalamin as Intermediates in the Aerobic Photolysis of Adenosylcobalamin. Biochemistry 2007, 46, 72847292. (28) Jones, A. R.; Russell, H. J.; Greetham, G. M.; Towrie, M.; Hay, S.; Scrutton, N. S. Ultrafast Infrared Spectral Fingerprints of Vitamin B12 and Related Cobalamins. J. Phys. Chem. A 2012, 116, 5586-5594. (29) Jones, A. R. The Photochemistry and Photobiology of Vitamin B12. Photochem. Photobiol. Sci. 2017, 16, 820-834. (30) Walker, L. A.; Jarrett, J. T.; Anderson, N. A.; Pullen, S. H.; Matthews, R. G.; Sension, R. J. Time-Resolved Spectroscopic Studies of B12 Coenzymes:  The Identification of a Metastable Cob(III)alamin Photoproduct in the Photolysis of Methylcobalamin. J. Am. Chem. Soc. 1998, 120, 3597-3603. (31) Walker, L. A.; Shiang, J. J.; Anderson, N. A.; Pullen, S. H.; Sension, R. J. Time-Resolved Spectroscopic Studies of B12 Coenzymes:  The Photolysis and Geminate Recombination of Adenosylcobalamin. J. Am. Chem. Soc. 1998, 120, 7286-7292. (32) Yoder, L. M.; Cole, A. G.; Walker, L. A.; Sension, R. J. Time-Resolved Spectroscopic Studies of B12 Coenzymes:  Influence of Solvent on the Photolysis of Adenosylcobalamin. J. Phys. Chem. B 2001, 105, 12180-12188. (33) Cole, A. G.; Yoder, L. M.; Shiang, J. J.; Anderson, N. A.; Walker, L. A.; Banaszak Holl, M. M.; Sension, R. J. Time-Resolved Spectroscopic Studies of B12 Coenzymes:  A Comparison of

24 ACS Paragon Plus Environment

Page 25 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

the Primary Photolysis Mechanism in Methyl-, Ethyl-, n-Propyl-, and 5‘Deoxyadenosylcobalamin. J. Am. Chem. Soc. 2002, 124, 434-441. (34) Sension, R. J.; Harris, D. A.; Cole, A. G. Time-Resolved Spectroscopic Studies of B12 Coenzymes: Comparison of the Influence of Solvent on the Primary Photolysis Mechanism and Geminate Recombination of Methyl-, Ethyl-, n-Propyl-, and 5'-deoxyadenosylcobalamin. J. Phys. Chem. B 2005, 109, 21954-21962. (35) Shiang, J. J.; Cole, A. G.; Sension, R. J.; Hang, K.; Weng, Y.; Trommel, J. S.; Marzilli, L. G.; Lian, T. Ultrafast Excited-State Dynamics in Vitamin B12 and Related Cob(III)alamins. J. Am. Chem. Soc. 2006, 128, 801-808. (36) Chen, E.; Chance, M. R. Continuous-Wave Quantum Yields of Various Cobalamins are Influenced by Competition Between Geminate Recombination and Cage Escape. Biochemistry 1993, 32, 1480-1487. (37) Garabato, B. D.; Lodowski, P.; Jaworska, M.; Kozlowski, P. M., Mechanism of Co-C Photodissociation in Adenosylcobalamin. Phys. Chem. Chem. Phys. 2016, 18, 19070-19082. (38) Kozlowski, P. M.; Garabato, B. D.; Lodowski, P.; Jaworska, M., Photolytic Properties of Cobalamins: a Theoretical Perspective. Dalton Trans. 2016, 45, 4457-4470. (39) Jones, A. R.; Hardman, S. J. O.; Hay, S.; Scrutton, N. S. Is There a Dynamic Protein Contribution to the Substrate Trigger in Coenzyme B12-Dependent Ethanolamine Ammonia Lyase? Angew. Chem., Int. Ed. 2011, 50, 10843-10846. (40) Shiang, J. J.; Walker, L. A.; Anderson, N. A.; Cole, A. G.; Sension, R. J. Time-Resolved Spectroscopic Studies of B12 Coenzymes:  The Photolysis of Methylcobalamin Is Wavelength Dependent. J. Phys. Chem. B 1999, 103, 10532-10539. (41) Lodowski, P.; Jaworska, M.; Andruniow, T.; Garabato, B. D.; Kozlowski, P. M. Mechanism of Co-C Bond Photolysis in the Base-on Form of Methylcobalamin. J. Phys. Chem. A 2014, 118, 11718-11734. (42) Harris, D. A.; Stickrath, A. B.; Carroll, E. C.; Sension, R. J. Influence of Environment on the Electronic Structure of Cob(III)alamins: Time-Resolved Absorption Studies of the S(1) State Spectrum and Dynamics. J. Am. Chem. Soc. 2007, 129, 7578-7585. (43) Wiley, T. E.; Arruda, B. C.; Miller, N. A.; Lenard, M.; Sension, R. J. Excited Electronic States and Internal Conversion in Cyanocobalamin. Chin. Chem. Lett. 2015, 26, 439-443. (44) Rury, A. S.; Wiley, T. E.; Sension, R. J. Energy Cascades, Excited State Dynamics, and Photochemistry in Cob(III)alamins and Ferric Porphyrins. Acc. Chem. Res. 2015, 48, 860-867. (45) Lodowski, P.; Jaworska, M.; Andruniow, T.; Garabato, B. D.; Kozlowski, P. M. Mechanism of the S1 excited state Internal Conversion in Vitamin B12. Phys. Chem. Chem. Phys. 2014, 16, 18675-18679. (46) Lodowski, P.; Jaworska, M.; Garabato, B. D.; Kozlowski, P. M. Mechanism of Co-C Bond Photolysis in Methylcobalamin: Influence of Axial Base. J. Phys. Chem. A 2015, 119, 39133928. (47) Peng, J.; Tang, K.-C.; McLoughlin, K.; Yang, Y.; Forgach, D.; Sension, R. J. Ultrafast Excited-State Dynamics and Photolysis in Base-Off B12 Coenzymes and Analogues: Absence of the trans-Nitrogenous Ligand Opens a Channel for Rapid Nonradiative Decay. J. Phys. Chem. B 2010, 114, 12398-12405. (48) Jones, A. R.; Woodward, J. R.; Scrutton, N. S. Continuous Wave Photolysis Magnetic Field Effect Investigations with Free and Protein-bound Alkylcobalamins. J. Am. Chem. Soc. 2009, 131, 17246-17253.

25 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 39

(49) Jones, A. R.; Levy, C.; Hay, S.; Scrutton, N. S. Relating Localized Protein Motions to the Reaction Coordinate in Coenzyme B12-dependent Enzymes. FEBS J. 2013, 280, 2997-3008. (50) Zhu, C.; Warncke, K. Kinetic Isolation and Characterization of the Radical Rearrangement Step in Coenzyme B12-Dependent Ethanolamine Ammonia-lyase. J. Am. Chem. Soc. 2010, 132, 9610-9615. (51) Sension, R. J.; Cole, A. G.; Harris, A. D.; Fox, C. C.; Woodbury, N. W.; Lin, S.; Marsh, E. N. Photolysis and Recombination of Adenosylcobalamin Bound to Glutamate Mutase. J. Am. Chem. Soc. 2004, 126, 1598-1599. (52) Sension, R. J.; Harris, D. A.; Stickrath, A.; Cole, A. G.; Fox, C. C.; Marsh, E. N. Timeresolved Measurements of the Photolysis and Recombination of Adenosylcobalamin Bound to Glutamate Mutase. J. Phys. Chem. B 2005, 109, 18146-18152. (53) Robertson, W. D.; Wang, M.; Warncke, K. Characterization of Protein Contributions to Cobalt-carbon Bond Cleavage Catalysis in Adenosylcobalamin-dependent Ethanolamine Ammonia-lyase by using Photolysis in the Ternary Complex. J. Am. Chem. Soc. 2011, 133, 6968-6977. (54) Robertson, W. D.; Warncke, K. Photolysis of Adenosylcobalamin and Radical Pair Recombination in Ethanolamine Ammonia-Lyase Probed on the Micro- to Millisecond Time Scale by Using Time-Resolved Optical Absorption Spectroscopy. Biochemistry 2009, 48, 140147. (55) Bridwell-Rabb, J.; Drennan, C. L. Vitamin B12 in the Spotlight Again. Curr. Opin. Chem. Biol. 2017, 37, 63-70. (56) Jost, M.; Fernandez-Zapata, J.; Polanco, M. C.; Ortiz-Guerrero, J. M.; Chen, P. Y.; Kang, G.; Padmanabhan, S.; Elias-Arnanz, M.; Drennan, C. L. Structural Basis for Gene Regulation by a B12-dependent Photoreceptor. Nature 2015, 526, 536-541. (57) Kutta, R. J.; Hardman, S. J.; Johannissen, L. O.; Bellina, B.; Messiha, H. L.; Ortiz-Guerrero, J. M.; Elias-Arnanz, M.; Padmanabhan, S.; Barran, P.; Scrutton, N. S.; Jones, A. R. The Photochemical Mechanism of a B12-dependent Photoreceptor Protein. Nat. Commun. 2015, 6, 7907. (58) Brooks, A. J.; Vlasie, M.; Banerjee, R.; Brunold, T. C. Co-C Bond Activation in Methylmalonyl-CoA Mutase by Stabilization of the Post-homolysis Product Co2+ Cobalamin. J. Am. Chem. Soc. 2005, 127 , 16522-8. (59) Brooks, A. J.; Fox, C. C.; Marsh, E. N. G.; Vlasie, M.; Banerjee, R.; Brunold, T. C. Electronic Structure Studies of the Adenosylcobalamin Cofactor in Glutamate Mutase. Biochemistry 2005, 44 , 15167-15181. (60) Chang, G. W.; Chang, J. T. Evidence for the B12-dependent Enzyme Ethanolamine Deaminase in Salmonella. Nature 1975, 254, 150-151. (61) Sagermann, M.; Ohtaki, A.; Nikolakakis, K. Crystal Structure of the EutL Shell Protein of the Ethanolamine Ammonia Lyase Microcompartment. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 8883-8887. (62) Shibata, N.; Tamagaki, H.; Hieda, N.; Akita, K.; Komori, H.; Shomura, Y.; Terawaki, S.; Mori, K.; Yasuoka, N.; Higuchi, Y.; Toraya, T. Crystal Structures of Ethanolamine Ammonialyase Complexed with Coenzyme B12 Analogs and Substrates. J. Biol. Chem. 2010, 285, 2648426493. (63) Wang, M.; Warncke, K. Entropic Origin of Cobalt–Carbon Bond Cleavage Catalysis in Adenosylcobalamin-Dependent Ethanolamine Ammonia-Lyase. J. Am. Chem. Soc. 2013, 135, 15077-15084. 26 ACS Paragon Plus Environment

Page 27 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

(64) Chen, Z. G.; Ziętek, M. A.; Russell, H. J.; Tait, S.; Hay, S.; Jones, A. R.; Scrutton, N. S. Dynamic, Electrostatic Model for the Generation and Control of High‐Energy Radical Intermediates by a Coenzyme B12‐Dependent Enzyme. ChemBioChem 2013, 14, 1529-1533. (65) Abend, A.; Bandarian, V.; Nitsche, R.; Stupperich, E.; Retey, J.; Reed, G. H. Ethanolamine Ammonia-lyase has a "Base-on" Binding Mode for Coenzyme B12. Arch. Biochem. Biophys. 1999, 370, 138-141. (66) Ke, S. C.; Torrent, M.; Museav, D. G.; Morokuma, K.; Warncke, K. Identification of Dimethylbenzimidazole Axial Coordination and Characterization of (14)N Superhyperfine and Nuclear Quadrupole Coupling in Cob(II)alamin Bound to Ethanolamine Deaminase in a Catalytically-Engaged Substrate Radical-Cobalt(II) Biradical State. Biochemistry 1999, 38, 12681-12689. (67) Olsson, M. H. M.; Søndergaard, C. R.; Rostkowski, M.; Jensen, J. H. PROPKA3: Consistent Treatment of Internal and Surface Residues in Empirical pKa Predictions. J. Chem. Theory Comput. 2011, 7, 525-537. (68) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E., UCSF Chimera—A visualization system for exploratory research and analysis. J. Comput. Chem. 2004, 25 , 1605-1612. (69) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J. V.; Izmaylov, A. F.; Sonnenberg, J. L.; Williams; Ding, F.; Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson, T.; Ranasinghe, D.; Zakrzewski, V. G.; Gao, J.; Rega, N.; Zheng, G.; Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Throssell, K.; Montgomery Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J. J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Millam, J. M.; Klene, M.; Adamo, C.; Cammi, R.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J. Gaussian 09, Wallingford, CT, 2009. (70) Marques, H. M.; Ngoma, B.; Egan, T. J.; Brown, K. L. Parameters for the Amber Force Field for the Molecular Mechanics Modeling of the Cobalt Corrinoids. J. Mol. Struct. 2001, 561, 71-91. (71) Vreven, T.; Byun, K. S.; Komáromi, I.; Dapprich, S.; Montgomery, J. A.; Morokuma, K.; Frisch, M. J. Combining Quantum Mechanics Methods with Molecular Mechanics Methods in ONIOM. J. Chem. Theory Comput. 2006, 2, 815-826. (72) Kozlowski, P. M.; Kumar, M.; Piecuch, P.; Li, W.; Bauman, N. P.; Hansen, J. A.; Lodowski, P.; Jaworska, M. The Cobalt–Methyl Bond Dissociation in Methylcobalamin: New Benchmark Analysis Based on Density Functional Theory and Completely Renormalized Coupled-Cluster Calculations. J. Chem. Theory and Comput. 2012, 8, 1870-1894. (73) Kornobis, K.; Kumar, N.; Wong, B. M.; Lodowski, P.; Jaworska, M.; Andruniów, T.; Ruud, K.; Kozlowski, P. M. Electronically Excited States of Vitamin B12: Benchmark Calculations Including Time-Dependent Density Functional Theory and Correlated ab Initio Methods. J. Phys. Chem. A 2011, 115, 1280-1292. (74) Solheim, H.; Kornobis, K.; Ruud, K.; Kozlowski, P. M. Electronically Excited States of Vitamin B12 and Methylcobalamin: Theoretical Analysis of Absorption, CD, and MCD Data. J. Phys. Chem. B 2011, 115, 737-748.

27 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 39

(75) Kornobis, K.; Ruud, K.; Kozlowski, P. M. Cob(I)alamin: Insight Into the Nature of Electronically Excited States Elucidated via Quantum Chemical Computations and Analysis of Absorption, CD and MCD Data. J. Phys. Chem. A 2013, 117, 863-876. (76) Kornobis, K.; Kumar, N.; Lodowski, P.; Jaworska, M.; Piecuch, P.; Lutz, J. J.; Wong, B. M.; Kozlowski, P. M. Electronic structure of the S1 state in methylcobalamin: Insight from CASSCF/MC‐XQDPT2, EOM‐CCSD, and TD‐DFT calculations. J. Comput. Chem. 2013, 34, 987-1004. (77) Peach, M. J.; Benfield, P.; Helgaker, T.; Tozer, D. J. Excitation energies in density functional theory: an evaluation and a diagnostic test. J. Chem. Phys. 2008, 128, 044118. (78) Le Bahers, T.; Adamo, C.; Ciofini, I. A Qualitative Index of Spatial Extent in ChargeTransfer Excitations. J. Chem. Theory Comput. 2011, 7, 2498-2506. (79) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J. V.; Izmaylov, A. F.; Sonnenberg, J. L.; Williams; Ding, F.; Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson, T.; Ranasinghe, D.; Zakrzewski, V. G.; Gao, J.; Rega, N.; Zheng, G.; Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Throssell, K.; Montgomery Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J. J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Millam, J. M.; Klene, M.; Adamo, C.; Cammi, R.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J. Gaussian 16, Wallingford, CT, 2016. (80) Khoroshun, D. V.; Warncke, K.; Ke, S. C.; Musaev, D. G.; Morokuma, K. Internal Degrees of Freedom, Structural Motifs, and Conformational Energetics of the 5'-deoxyadenosyl Radical: Implications for Function in Adenosylcobalamin-Dependent Enzymes. A Computational Study. J. Am. Chem. Soc. 2003, 125, 570-579. (81) Altona, C.; Sundaralingam, M. Conformational Analysis of the Sugar Ring in Nucleosides and Nucleotides. A New Description using the Concept of Pseudorotation. J. Am. Chem. Soc. 1972, 94, 8205-8212. (82) Gruber, K.; Reitzer, R.; Kratky, C. Radical Shuttling in a Protein: Ribose Pseudorotation Controls Alkyl-Radical Transfer in the Coenzyme B12 Dependent Enzyme Glutamate Mutase. Angew. Chem. Int. Ed. 2001, 40, 3377-3380. (83) Rommel, J. B.; Kastner, J. The Fragmentation-recombination Mechanism of the Enzyme Glutamate Mutase Studied by QM/MM Simulations. J. Am. Chem. Soc. 2011, 133, 10195-10203. (84) Conrad, K. S.; Jordan, C. D.; Brown, K. L.; Brunold, T. C. Spectroscopic and Computational Studies of Cobalamin Species with Variable Lower Axial Ligation: Implications for the Mechanism of Co–C Bond Activation by Class I Cobalamin-Dependent Isomerases. Inorg. Chem. 2015, 54, 3736-3747. (85) Miller, N. A.; Deb, A.; Alonso-Mori, R.; Garabato, B. D.; Glownia, J. M.; Kiefer, L. M.; Koralek, J.; Sikorski, M.; Spears, K. G.; Wiley, T. E.; Zhu, D.; Kozlowski, P. M.; Kubarych, K. J.; Penner-Hahn, J. E.; Sension, R. J. Polarized XANES Monitors Femtosecond Structural Evolution of Photoexcited Vitamin B12. J. Am. Chem. Soc. 2017, 139, 1894-1899. (86) Chen, E.; Chance, M. R., Nanosecond Transient Absorption Spectroscopy of Coenzyme B12. Quantum Yields and Spectral Dynamics. J. Biol. Chem. 1990, 265, 12987-94.

28 ACS Paragon Plus Environment

Page 29 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

(87) Garabato, B. D.; Kumar, N.; Lodowski, P.; Jaworska, M.; Kozlowski, P. M. Electronically Excited States of Cob(II)alamin: Insights from CASSCF/XMCQDPT2 and TD-DFT Calculations. Phys. Chem. Chem. Phys. 2016, 18, 4513-26. (88) Reig, A. J.; Conrad, K. S.; Brunold, T. C. Combined Spectroscopic/Computational Studies of Vitamin B12 Precursors: Geometric and Electronic Structures of Cobinamides. Inorg. Chem. 2012, 51, 2867-2879. (89) Russell, H. J.; Jones, A. R.; Hay, S.; Greetham, G. M.; Towrie, M.; Scrutton, N. S. Protein Motions Are Coupled to the Reaction Chemistry in Coenzyme B12-Dependent Ethanolamine Ammonia Lyase. Angew. Chem. Int. Ed. 2012, 51, 9306-9310.

Figure 1. (a) General molecular structure of cobalamins (b) Simplified model of the AdoCbl cofactor used in the QM region, where R represents the Ado ligand, also shown

29 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Page 30 of 39

Scheme 1. (a) General mechanism of Co-C homolysis in AdoCbl-dependent enzymes to generate Co(II) and substrate radical (orange box). Substrate rearrangement in AdoCbl-dependent EAL (blue box). (b) Cleavage of Co-C bond by light in enzymes without substrate

30

ACS Paragon Plus Environment

Page 31 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

ACS Catalysis

Figure 2. (a) Crystal structure of EAL (PDB id: 3ABS) with AdePeCbl cofactor where adenine-9-yl-pentyl (AdePe) is the upper axial ligand. Substrate EA is depicted in purple. (b) Optimized structure of EAL with reconstructed AdoCbl cofactor where AdePe is replaced with the Ado ligand. Substrate EA was removed and replaced by two water molecules.

31

ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Page 32 of 39

Figure 3. Potential energy curves for the ground (black) and lowest, vertical singlet (red) and triplet (blue) excited states as functions of Co-C bond length for (a) AdoCbl inside EAL and (b) Im−[CoIII(corrin)]−Ado+ base-on model complex in water. 32

ACS Paragon Plus Environment

Page 33 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Figure 4. Potential energy surfaces for the S0 with vertical projections of the S1 state plotted as functions of axial bond lengths along with the corresponding S0 and S1 contour PESs with color codes for (a) AdoCbl inside EAL (b) Im−[CoIII(corrin)]−Ado+ base-on model complex in water and (c) H2O−[CoIII(corrin)]−Ado+ base-off model complex in water

33 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 39

Figure 5. (a) Scheme of photoreaction of AdoCbl inside EAL on the S1 PES with minima regions, separated by a seam, where MLCT minimum is denoted as I (S1min) and the LF minimum is denoted as IIIB (S1min) and (b) minimum energy paths (Path A and B) between MLCT and LF state plotted as a function of energy versus Q value.

34 ACS Paragon Plus Environment

Page 35 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Figure 6. HOMO and LUMO molecular orbitals involved in electronic excitations corresponding to selected points on the S1 PES (Figure 5a) along Path A and Path B. 35 ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 39

Figure 7. Potential energy surfaces of S1 state as function of axial bond lengths with photoreaction pathways depicted by red arrows for (a) AdoCbl inside EAL (b) Im−[CoIII(corrin)]−Ado+ base-on model complex in water and (c) H2O−[CoIII(corrin)]−Ado+ base-off model complex in water .

36 ACS Paragon Plus Environment

Page 37 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

ACS Catalysis

Scheme 2. Mechanism of AdoCbl photolysis in EAL and post-homolysis photophysical events (cage escape, internal conversion), as described in Section 3.7

37

ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 39

Table 1. Eight lowest vertical singlet electronic transitions and orbital characterization based on the single point TD-DFT/MM calculations of AdoCbl inside EAL E[eV] S1

S2

S3

S4

S5

S6

S7

S8

1.97

2.12

2.26

2.33

2.51

2.59

2.66

2.68

f

λ[nm]

DCT

%

.0038

629.4

1.029

23

177-179

H-1→L

dyz+π → π*

59

176-179

H-2→L

dxz/dz2+ π → π*

11

174-179

H-4→L

dxz+ π → π*

64

177-179

H-1→L

dyz+π → π*

21

176-179

H-2→L

dxz/dz2+ π → π*

4

176-182

H-2→L+3

dxz/dz2+ π → π*

88

173-179

H-5→L

dxy → π*

5

174-179

H-4→L

dxz+ π → π*

7

177-180

H-1→L+1

dyz+π → dz2

12

176-179

H-2→L

dxz/dz2+ π → π*

65

174-179

H-4→L

dxz+ π → π*

7

173-179

H-5→L

dxy → π*

26

177-180

H-1→L+1

dyz+π → dz2

67

178-181

H→L+2

πAdo → dxy-n

32

178-181

H→L+2

πAdo → dxy-n

55

177-180

H-1→L+1

dyz+π → dz2

3

174-179

H-4→L

dxz+ π → π*

70

172-179

H-6→L

dxz/dz2+ πAdo+ πRib → π*

25

171-179

H-7→L

dxz/dz2+ πAdo+ π → π*

15

177-181

H-1→L+1

dyz+π → dz2

7

176-180

H-2→L+1

dxz/dz2+ π → dz2

52

170-179

H-8→L

dxz+ πAdo+ π → π*

10

171-179

H-7→L

dyz+ πAdo+ π → π*

5

172-179

H-6→L

dxz/dz2+ πAdo+ πRib→ π*

.0254

.0027

.0385

.0035

.0081

.0024

.0022

587.5

548.1

530.8

490.5

490.4

466.8

459.2

1.247

0.913

1.331

3.971

5.428

5.942

3.695

Character

Exp. [nm] (eV)

527 (2.35)

490(2.53)

38 ACS Paragon Plus Environment

Page 39 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Table of Contents Graphic:

39 ACS Paragon Plus Environment