In 2018, Frances Arnold was among three scientists who were awarded the Nobel Prize in Chemistry. In her Nobel Prize lecture Arnold states, “I wish to go beyond optimizing biological functions that are already known, and instead bring to life whole new chemistries. But how can one create enzymes that catalyze reactions invented by chemists?” To answer this question, Arnold’s lab developed a protein engineering process called directed evolution. Directed evolution is the process of altering amino acids at various sites of an enzyme to create or enhance a protein’s ability to carry out desired chemical reactions. In Arnold’s words, it is taking enzymes “where biology has never gone.” Evolved enzymes can subsequently be integrated into biosynthetic pathways of microorganisms, or cascade reactions to improve the synthesis of pharmaceuticals and other value-added compounds. Enzyme catalysts are also environmentally friendly, an advantage difficult to underscore given our planet’s warming climate.
Using the directed evolution approach described above, Arnold’s lab engineered cytochrome p450 monooxygenases capable of olefin cyclopropanation. Cytochrome p450s and other monooxygenases have since been engineered for nitrene insertion. The paper discussed in this article is work from Rudi Fasan’s lab, a former member of the Arnold lab and now a full professor at Rochester University. The paper, published in Nature Catalysis, reports on a myoglobin variant capable of nitrene insertion for biosynthesis of β-, γ-, and δ-lactam rings. The authors write, “The synthesis of cyclic amides (lactams), which are key structural motifs in many pharmaceuticals, agrochemicals, and other fine chemicals, has presented a major challenge.” Below is a summary of the studies undertaken by the Fasan group to address this challenge.
An H64V, V68A myoglobin variant (Mb*) that synthesizes γ-lactams was engineered. Since the groundbreaking report of their use in cyclopropanation reactions, p450s have been engineered for C-H amination through nitrene transfer. And while Ir and Ru-based organic catalysts have been employed to synthesize γ-lactams, no organic catalysts or biocatalysts have been available for making β- and δ- lactams. Well, not anymore! Using a dioxazolone substrate and an H64V myoglobin variant, the authors were able to synthesize a γ-lactam at 2% yield. Further engineering yielded a double mutant (H64V,V68A or Mb*) with 25x increase in yield and >99% e.e. of the γ-lactam! As a side note, dioxazolone substrates were chosen as substrates because they can be synthesized easily from abundant carboxylate starting materials. The figure below was obtained from the article.
The yield of lactams is enhanced by low pH and the use of acetonitrile as a co-solvent. To limit the acyclic amide byproduct generated in the reaction (presumably resulting from protonation of the nitrene intermediate), the authors ran the anaerobic reactions under alkaline conditions. Various organic solvents were tested as reaction co-solvents to further disfavor acyclic byproduct formation. Both approaches led to an increase in yield (from 50% to 75%), all while retaining excellent e.e. values.
The γ-lactam-generating biocatalyst has a broad substrate scope. Substitutions at the para, ortho, and meta positions of the aryl ring of the dioxazolone substrate were well-tolerated in general. However, replacing the aryl ring with heterocyclic substituents dropped yields and e.e. values in some instances.
Mb* can synthesize β- and δ-lactams from dioxazolone substrates with a similarly broad substrate scope. β-lactams could be synthesized at yields as high as 93% with excellent e.e. values (generally 99%). δ-lactams were also synthesized successfully but contained impurities comprising acyclic amides and β-lactams. However, by replacing the γ-C-H atom with oxygen, the authors were able to eliminate unwanted lactams and increase both the yield and e.e. of the reaction for δ-lactam biosynthesis.
Mechanistic studies reveal that the active site of Mb* plays a significant role in determining enantioselectivity. Mb* catalysis is expected to follow the general mechanism established for p450s: dioxazolone activation (accompanied by CO2 elimination) and nitrene formation is followed by an intramolecular hydrogen atom abstraction (HAA) of the β, γ, or δ C-H atoms. In the radical rebound step, an intramolecular cyclization yields β-, γ-, or δ-lactams. Kinetic isotope effect (KIE) studies revealed, interestingly, that irrespective of deuteration at pro-S or pro-R positions, a single product (the S enantiomer) was the dominant product. From this observation, it was surmised that the active site milieu was critical for enantioselective synthesis (i.e., preferential Si face attack). Additionally, density functional theory (DFT) calculations revealed that the radical rebound step (formation of the C-N bond) displayed a higher energy barrier (12.5 kcal/mol) compared to the HAA step (8 kcal/mol). As a result, the final step was concluded to be the rate-limiting step of the reaction in Mb*. The figure below was obtained from the article.
Mb* was used to synthesize an alkaloid and a pharmaceutical in fewer steps. A phenyl-substituted β-lactam generated by Mb* was used to synthesize (S,S)-(–)-homaline and (S)-dapoxetine. Incorporating Mb& into the synthetic route reduced the number of steps from 11 to 7 and 12 to 8, respectively. As an added advantage, an enantiodivergent variant of the Mb lactam biocatalyst was also engineered.
