Scaling protein engineering with new gene synthesis technology

Protein engineering – the process of designing and modifying proteins to improve their properties or develop entirely new functions – has the potential to address some of humanity’s greatest challenges¹; from sustainable material production to healthcare and food availability. While technological advancement in areas such as AI and our growing understanding of biological systems have greatly increased our ability to engineer proteins, scaling up of these techniques is required for engineering biology to truly reach its potential. To engineer proteins at scale will require new gene synthesis technology to enable rapid, iterative testing of protein variants – a key bottleneck in current processes.²

Since the first recombinant protein-based therapy was approved over 40 years ago, protein engineering has transformed the world of healthcare, with biologics now representing the fastest-growing sector of the pharmaceutical industry.^3,4 Hundreds of therapies are now approved and the market is predicted to continue to grow to almost $600 billion by 2029.⁵ Proteins as a group of therapies display unique versatility, acting as catalysts, signaling molecules, transporters, scaffolds, receptors, antimicrobial agents and more. By engineering proteins with improved properties, scientists can develop more effective drugs with fewer side effects, as well as optimizing for stability and pharmacokinetics.

Another key application for protein engineering is within agriculture. There is a need to improve crop yields, enhance disease resistance and improve nutritional value while decreasing demand for land, fertilizers and pesticides, a challenge that is becoming more pressing as the world’s population grows: it’s predicted that agricultural production needs to increase by around 50% before 2050.⁶ However, crop improvements through selective breeding and agricultural improvements are yielding diminishing returns due to inefficient practices, slow generation time for crop plants and small gene pools. Protein engineering will be essential to overcome these challenges and meet demands for crop and food production.

The application of protein engineering to industrial processes is also transforming product manufacture. Although less established than medical and agricultural applications for protein engineering, biosynthesis of materials including fuels, plastics and chemicals represents an area of significant growth that is driven by demand for new approaches to materials synthesis that prevent pollution, conserve resources and reduce CO2 emissions. The global industrial biotechnology market is growing at 9.9% per year and is expected to reach over $850 billion by 2030.⁷ It is estimated that 60% of physical products in the global economy, including wood, meat, plastics and fuels could theoretically be produced or production optimized through biological innovations.⁸

Protein engineering involves a variety of techniques to design and functionally test engineered protein variants. Protein design is conducted through both rational methods such as site-directed mutagenesis and de novo design, as well as directed evolution – which involves random mutagenesis and screening of variants for desired traits.

Progress in machine learning and other AI-based tools has given scientists access to a suite of tools to enable rapid, sophisticated protein design. ML-guided protein engineering has rapidly gained attention and investment in recent years. To date, most tools have been used to optimize existing protein structures to achieve desired properties such as improved stability or specificity, for example by accelerating directed evolution by learning from the properties of characterized sequences.⁹

However, more recent developments to build tools that can develop entirely new proteins, not present in nature provide a huge opportunity to explore sequence space and find solutions to biomedical, industrial and agricultural problems that evolution has not been required to solve.¹⁰

Alongside advances in AI and machine learning, exploring optimized and novel proteins requires iterative design and function characterization to ensure that proteins have the desired properties. Whereas AI tools have become relatively fast at developing new protein variants, testing these variants presents a key bottleneck that needs to be addressed to fully reap the benefits that come from advancements in AI-assisted protein design.²

A major bottleneck in the testing of protein variants rests in the capacity for production of gene-length DNA. DNA synthesis and assembly have become centralized services, provided by a few reagent manufacturers. This delays iterative protein variant testing as preparing new sequences involves waiting for DNA to be delivered, the timeframe for which increases with longer and more complex sequences (Figure 1). With huge variation in the length of protein-coding sequences, ranging from a few hundred to many thousand bases, the synthesis of longer DNA poses a particular challenge.

A recent study by the US Department Research Projects Agency illustrates this problem. The study was designed to evaluate the robustness of a biofoundry by challenging the team to generate organisms that would produce defined molecules within a timeframe of 90 days. Six out of ten targets were successfully generated during the test period; however, sourcing DNA was a major bottleneck, taking about half the allotted time.¹¹

The extended timeframe for delivery of longer sequences is largely due to errors incorporated during DNA synthesis. Error incorporation increases exponentially with sequence length, resulting in significant post-synthesis steps being required to correct errors and identify error-free sequences.

Developments in DNA synthesis technology present opportunities to resolve bottlenecks associated with testing protein variants and scaling protein engineering. Evonetix is developing technology that reinvents DNA synthesis with sophisticated gene synthesis capabilities in a benchtop device, putting control directly in the hands of scientists and allowing the synthesis of long, accurate DNA on demand in days rather than weeks or months. This has the potential to significantly speed up the iterative design-build-test-learn cycle essential to the engineering of functional novel and optimized proteins.

In addition, our patented Binary Assembly^® process and thermal control via a unique DNA synthesis and assembly chip allow oligonucleotides to be automatically combined into double-stranded gene-length DNA and error-containing sequences removed, significantly improving the quality of DNA synthesized, reducing post-synthesis steps and further accelerating protein engineering.

Protein engineering is already having a profound effect on healthcare, materials and agriculture. However, there is a need for engineering biology to scale further to address issues such as sustainability, food availability, and emerging health threats. The potential for engineering biology is dependent on access to fast, accurate gene-length DNA that will accelerate the pace of innovation by optimizing development and production timelines.