One area of our research is devoted to the study of the known bio-organometallic cofactors, which include the enzymatically active forms of vitamin B₁₂, adenosylcobalamin and methylcobalamin.

While structurally related to the well-characterized hemes, these cofactors employ macrocyclic ligands (termed corrin and hydrocorphin, respectively) that are substantially more reduced than the porphyrin ring, offering them considerably more conformational freedom. The mechanistic significance of this increased flexibility in the catalytic cycles of B₁₂- and NiF₄₃₀-dependent enzymes remains a subject of intense debate.

Current reseach efforts are focused on Coenzyme-B12, aimed at obtaining mechanistic insight into the electronic properties of the cofactor when bound to proteins. One of such systems involve the synthesis of the Co-C(Ado) bond that is so important for functionality. The general class of enzymes that carry out this reaction are known as adenosyltransferases (ATRs). Furthermore, another class of enzymes harness the Co-C(Ado) bond to mediate radical rearragenment reactions in biological systems. These are known as AdoCbl-dependent enzymes.

AdenosylTransferases (ATRs)

Three clases of enzymes are known that carry out the adesnosylation of cofactors. These are known as the CobA, PudO, and EutT types based on the biological systems in which they were first characterized. These familes have different sequence identity and bind their substrates (ATP and pre-adenosylated cofactor) with different protein topologies. The catalytic cycle of two of these enzymes has been captured with X-ray crystallography (CobA and PduO, homolog to the enzyme found in humans hATR), while the latter, EutT, still awaits direct observation.

These ATRs must overcome a challenging thermodynamic barrier to reduce the Cobalt center to the Co(I) state in the biological setting, a nessesary step in preparing the Co-C(Ado bond). The reduction potential of the free cofactor (Co(II) to Co(I) -610mV vs SHE) is too negative for the suspected biological reductants (-440 mV vs SHE for FlavodoxinA). This "super-reduced" form of the cofactor is poised to attach the 5' carbon of the ATP co-substrate to form the Co-C(Ado) bond and release inorganic phosphate.

Our techniques are suited at probing both the metal center present in the cofactor, as well as the conjugated corring ring that ligates it. Both EPR and MCD show a uniquely perturbed axial environment for biological relevant forms of the cofactor (A paramagnetic Co(II) species), not found in the absensce of protein or ATP. Additionally, given the constraining enviroment in the protein binding pocket, resonance Raman provides insight into the changes induced by the protein.

Our initial spectroscopic studies were supported by the crystal stucture of the LrPduO system, which highlighting the role of the protein matrix for catalysis, and shows direct evidence of the intermediate in question. Although similar spectroscopic signals are obtained across all ATRs studied in our group, the unique amino acid sequence and active site topology present for each family highlights possible differences in the way that the stabilization of the 4-coordinate intermediate occurs.

In the LrPduO system, a hydrophobic pocket is found in the enzyme for the exclusion of the axial ligand, as well as various hydrogen bonds between the corrin ring sidechains aligning the cofator in place. In addition, key residues are found that constrain the geometry of the active site, and thus are important for stabilization. Given the spectroscopic signals obtained for this system, a good reference point exists in which to interpret the observed results for other classes of ATRs. This combined approach should provide a comprehensive understanding how these systems accomplish the same rection.

AdoCbl-Dependent Enzymes

Organisms utilize AdoCbl-dependent isomerases to catalyze a variety of 1,2-rearrangements via a radical mechanism. The first step in this mechanism is a homolytic cleavage of the Co-C(Ado) bond to form an organic adenosyl radical and a reduced cob(II)alamin. One of our research goals is to understand the mechanism by which these proteins can increase the rate of this homolysis by many orders of magnitude.

AdoCbl-dependent isomerases can be loosely grouped into three classes. Class I enzymes catalyze rearrangements in which the migrating group is a carbon skeleton; class II and class III enzymes help either an –OH or –NH2 group migrate. Another important distinction between the classes of enzymes is their method of binding the cofactor. Enzymes in classes I and III bind a histidine residue directly to the Co center, displacing the intramolecular dimethylbenzimidazole (DMB) base that is axially ligated in the free cofactor. Class II enzymes do not displace the DMB. One outstanding question in B12 research, then, is whether or not all classes of enzymes work via the same mechanism.

Generally speaking, a reaction is catalyzed when the free energy gap between the reactant and transition states is lowered. This can happen either by destabilizing the ground state of the reactant or by stabilizing the transition state. Previous work in the group on methylmalonyl-CoA mutase (MMCM) and glutamate mutase (GM), both Class I enzymes, has shown that there is very little change to the electronic structure of the intact cofactor when it binds to these protein. However, the Abs, CD, and MCD spectra of the free and enzyme-bound Co(II)Cbl show significant blue-shifting of some transitions in the latter. With the help of our computational models, these were identified as metal to ligand charge transfer (MLCT) transitions, while the relatively unperturbed peaks were identified as either ligand field (metal-based) or π à π* (corrin-based) transitions. This model allowed us to come to the conclusion that uniform stabilization of the Co d-based MOs in the Co(II)Cbl species in the protein make it more energetically accessible than in solution. As the homolysis is an endergonic reaction step, we expect a similar stabilization of the transition state.

Now that we are fairly sure these enzymes work by transition state stabilization, we can try to discover the exact mechanism that the protein uses to do it. A likely possibility is that MMCM and GM tune the electron-donating properties of the coordinating histidine, which contributes to a formally antibonding axial Co-N interaction. This could occur, for example, by uptake of a proton by either the histidine residue or a preserved aspartate capable of hydrogen bonding with it. This is a promising hypothesis, but if it is confirmed it will open up questions about class II enzymes, which would be unable to tune the axial Co-N interaction in such a way. We plan to follow up on this question by spectroscopically and computationally investigating class II enzymes, including ethanolamine ammonia-lyase (EAL). Comparison of our results from EAL to those from MMCM and GM will provide key insight into how these two classes differ.