Embracing Enzymes

As pharma companies grow more mindful of the environmental impact of their products and supply chains, chemists and engineers are turning their attention to increasing efficiency and reducing waste in API synthesis. One attractive solution relies on nature’s catalysts – enzymes – which can be used as highly specific and selective “biocatalysts.” Notably, biocatalysts follow many of the principles of “green chemistry”, making them an attractive alternative to chemocatalysts.

So why aren’t more companies embracing enzymes? Put simply, the entire workflow of biocatalysis – from sourcing, development, and optimization of an enzyme, to finally delivering a biocatalyzed process – is a complex, multidisciplinary affair. And that’s why, for the pharmaceutical industry to fully leverage the power of biocatalysis, it’s vital to tap into advances across the key areas of enzyme screening, engineering, immobilization, and bioprocessing. 

With the growing complexity of pharmaceutical APIs, the demand for biocatalysts that are active towards novel, non-natural substrates has also increased. However, this is at odds with natural evolution, which has produced an impressive library of powerful yet substrate-specific enzymes. As such, many industrial reactions do not have a process-ready natural biocatalyst because the enzyme’s active site is suboptimal for accommodating the desired substrate.

To overcome this hurdle, scientists developed an approach that mimics natural evolution in a laboratory setting by randomly mutating, screening, and selecting thousands of enzyme variants. Using “directed evolution”, enzyme development scientists can incrementally address inherent hurdles, such as substrate scope, undesirable selectivity and process stability of the natural enzyme starting point. The importance of this technique for the pharmaceutical and wider chemical industry was recognized by the 2018 Nobel Prize in Chemistry – awarded to Frances Arnold.  

Though directed evolution showcased the power of engineered enzymes for industrial biocatalysis, the iterative cycles required to deliver a final process-ready biocatalyst is not always compatible with drug development timelines. Additionally, drug development suffers from a high attrition rate, making the return on investment for designing a new biocatalytic route uncertain. Therefore, for more biocatalyzed processes to be realized, we first need to enhance biocatalytic solutions for all desired routes and shorten enzyme engineering timelines.

Although the enzyme engineering field is synonymous with Arnold’s directed evolution technique, the mounting understanding of protein sequence–structure relationships has allowed for rational design to mature. Rational design approaches rely on computational analysis of sequence alignments and protein structure/dynamics to predict the exact changes required to elicit a desirable effect in the enzyme. The strategy has more success when prior structural and experimental data is available for the specific candidate enzyme or family of enzymes. Although still prone to high failure rate, if successful, enzyme development timelines can be drastically reduced.

At the intersection of both these methods is semi-rational design. By rationally and accurately predicting favorable “hotspots” on the enzyme structure through proprietary computational workflows but also allowing for randomness at hotspots, we can create a smart library that increases our chance of finding vastly improved enzymes in reduced timelines.

Altering an enzyme’s specificity or regioselectivity is usually determined by active site residues, while enzyme solubility and stability is often dictated by surface residues. However, in the pursuit of addressing one limitation, you may often affect another. As such, multiple compensatory mutations are required, which can be a long-process with full-gene random mutagenesis approaches.

Proprietary computational workflows are emerging that can be used to accelerate timelines by scanning millions of possibilities to identify the best enzyme for the target reaction. These workflows can traverse natural sequence space (public or metagenomic databases) and predictively find sequences that are as close as possible to an ideal biocatalyst for the transformation. In silico mutations can also be introduced if required.

But enzyme engineering is only one part of the equation. To deliver a robust biocatalyzed process, reaction engineering and enzyme formulation are equally important. For this reason, expertise should be drawn from a range of different specialists. This approach must operate throughout the entire process, starting with enzyme discovery, design and development through to the initial screening, process intensification, and scale-up stages. Especially as delivering the final industrially viable enzyme often relies on identifying the limitations of the biocatalyst in process conditions coupled with targeted enzyme engineering.

The enzyme development and process chemistry teams are closely linked with the bioprocessing team. Through the optimization of the molecular biology, fermentation, and downstream process, the bioprocessing team aim to reduce the final cost contribution of a biocatalyst. This can be done by maximizing expression in the host-organism, and thus, the specific activity of the final enzyme powder; however, it is also heavily dependent on the development team finding the right combination of plasmid, host, and codon-optimized DNA sequence.

Great headway has been made in biocatalysis – but, in my view, we need many more scientists and engineers working on the solution. The paradigm shift of “predictive biocatalysis” is not too far away, and with multiple academic groups, start-ups, and large corporates innovating across key aspects of the workflow. As such, the need for more efficient and sustainable approaches to pharmaceutical synthesis is not going away; the golden age of biocatalysis beckons.