Enzyme biocatalysis plays an important role in pharmaceutical synthesis as it can afford precise control over stereochemistry while achieving many conversions under physiological conditions. These benefits can both lower manufacturing costs and reduce environmental impact by eliminating solvent and heavy-metal catalyst steps in a synthetic strategy. Some recent examples of this include the enzyme-catalyzed synthesis of the antidiabetic medication sitagliptin, and the biomanufacturing of the antimalarial medication, artemisinin. Efficient enzyme discovery, engineering, and validation methods are required as the foundation of these efforts. As organisms occupy diverse environments, enzymes involved in identical conversions often possess dramatically different catalytic capacity and can harbor auxiliary domains which are critical for performance. As a result, it is has been routinely shown that broadening search efforts by drawing from diverse source organisms is the most efficient strategy for exploring catalytic fitness landscapes. Due to the cost and time associated with accessing, through DNA synthesis, large numbers of enzyme variants, researchers typically focus on testing a subselection of variants which have been previously characterized or those that are easy to obtain. This approach is fundamentally limited as: 1) the overwhelming majority of available biocatalysts have not been studied in detail due to cultivation bottlenecks and 2) there are no robust methods of predicting enzyme activity from primary sequence. To address this limitation, we propose to create a platform in which all computationally-identified enzyme variants in a database search can be immediately isolated, engineered, and delivered to end users through a publicly available, bioinformatics and LIMS-driven automation platform. Our strategy is ~1,000x less expensive while being up to 300x faster than current DNA synthesis theoretical limits. The platform is built upon a massive metagenomic library, the largest reported in the world, which overcomes the primary limitation of studying enzymes from cultivated sources. This collection contains several orders of magnitude greater enzyme diversity than can be found in culture collections. To populate the database, we will apply our patented high-efficiency sequencing method, to our metagenomic library, generating an unprecedented data set containing an N50 of >50kb. End-users will be able to search and identify enzyme variants in this dataset and both native sequences and combinatorial libraries of variants can be retrieved, produced, and automatically delivered to end users in less than a week and for far less cost than DNA synthesis. Overall, this system comprises an end-to-end, rapid biocatalyst discovery, engineering, and delivery system that will be a powerful resource for end users in basic research and industrial biotechnology.
The antimalarial medication, artemisinin, and the enzyme-catalyzed synthesis of the antidiabetic medication sitagliptin are two prominent successes of bio-based manufacturing of therapeutics. In order for biomanufacturing to succeed consistently, enzymes in the synthesis pathway must be extremely fast and efficient in order to produce the target molecules at cost-effective yields, yet the available choices for many of these pathway enzymes are very limited and slow. We propose to sequence the largest and most diverse metagenomic library reported to date and build a web-accessible platform that will provide researchers access to orders of magnitude more enzyme diversity to choose from when building out biomanufacturing pathways.
