AmericanChemicalSociety.com

The significance of oxidation reactions are readily observed in both industry and biology, with examples ranging from conversion of inert hydrocarbons into fuels (e.g., methane → methanol), synthesis of fine chemicals (e.g., oxidation of long-chain alkanes for surfactants), and hormone biosynthesis (e.g., androgens to estrogens).  While great advances have been made with regards to chemical catalysts for effecting such conversions, biological catalysts remain the “gold standard” for performing oxidation reactions with high degrees of regio- and stereospecificity.  Heme biocatalysts are particularly attractive given their propensity towards dioxygen activation; specifically, P450 cytochromes are generally sought for commercial exploitation.  A key challenge to capturing P450 activity in vitro involves (and remains) recapitulating the in vivo electron transfer machinery, which consists of multiple complex proteins and (expensive, e.g., NADPH) cofactors. 

Our research involves exploiting artificial electron transfer systems for electrode-driven oxidation reactions using biological heme complexes.  We have been working specifically with P450 from Bacillus megaterium (BM3) which contains hydroxylase and flavin-containing reductase domains on a single polypeptide.  Our focus has been on replacing NADPH with the electrochemically recyclable hydride transfer complex Cp*Rh(bpy)(H₂O)Cl₂ (Rh).  As the protein-bound flavins hinder efficient reaction with Rh, we sought to exploit mutagenesis to generate enzyme variants displaying improved activity with Rh.  While a generic directed evolution approach such as error-prone PCR could be used to generate a library with thousands of variants, we hoped to narrow the possibilities by selecting several sites for targeted, simultaneous saturation mutagenesis.  Previous work by Dunford and Girvan (2009) identified three residues in the BM3 reductase domain that were crucial for recognition and binding of NADPH:  S965, R966, and K972.  Indeed, mutation of these residues increased binding of the related cofactor NADH over 8000-fold, demonstrating that these residues can markedly impact cofactor recognition.  Thus, focusing on these residues we have applied two different methodologies to achieve simultaneous saturation mutagenesis.  The first involves a “megaprimer” scheme described by Sanchis and Reetz (2008) whereby a large DNA fragment (500-3000 bp) encompassing the mutation sites is generated by PCR.  This fragment containing the library is then used in whole-plasmid PCR followed by ligation and transformation.  While theoretically straightforward, this methodology has not generated libraries with sufficient variability.  Thus, concurrent with attempts to refine this scheme, we have recently turned to a more traditional approach involving whole-plasmid PCR with primers that begin at the mutation sites.  Using this scheme we have been able to generate DNA fragments of the appropriate length and we are currently evaluating the diversity of this library through sequence analysis. 

The challenges involved with harnessing BM3 activity in vitro lead us to expand our view and consider alternatives.  Thus, we turned to virus particles as versatile scaffolds for polyvalent display of heme centers that may be more electrochemically accessible, and therefore catalytically active.  The platform afforded by bacteriophage Qb permitted us to display three motifs, His₆, His₆-His₆, and Cys-His₆ via a co-expression methodology at ratios of 1.1:1, 1.1:1 and 2.3:1 for wild-type to modified coat protein.  Ni-NTA affinity chromatography resulted in retention times that increase according to Qb-His₆ < Qb-Cys-His₆ < Qb-His₆-His₆.  Two of the three constructs, Qb-Cys-His₆ and Qb-His₆-His₆, bind heme (absorbances at 416 and 418 nm, respectively) and were found to be electrochemically active.  Both constructs yield similar E_1/2 values and pH dependences, however a standard rate constant k^o could only be measured for Qb-Cys-His₆ (83 s^-1) as electron transfer for Qb-His₆-His₆ was too rapid to estimate.  Experiments with rotated-disk electrodes yielded significant activity of the constructs towards dioxygen reduction.  Despite this activity, in the presence of substrates no oxidized products were detected, suggesting that the heme-virus constructs were not generating sufficiently active ferryl species.  Nonetheless, it is likely possible to fine-tune/alter this activity by selectively mutating the coat protein; for example, incorporating hydrogen bonding residues adjacent to the heme-binding moieties may modulate catalytic activity.  Indeed, this further extends the potential applications:  that the particles interact with both metallocycles and immobilized metals, one can easily look towards applications such as, for example, immobilizing the particles onto a surface while simultaneously binding a cofactor for catalysis.  The potential benefit is further magnified given the polyvalent nature of the scaffold employed. 

Funds provided by the ACS PRF UNI have been used almost exclusively to support student workers, all of whom have benefited greatly from their experiences in the lab (e.g., sufficient data for publications and posters).  Indeed, the plethora of student help has allowed us to expand our research efforts to include the aforementioned heme-virus systems as P450 mimics, which has proved itself to be a lucrative field of research.  We hope to further develop and characterize these constructs with an emphasis on generating a range of biocatalysts displaying activities that correlate with heme-protein interactions.