CAZyme evolution and discovery: Ultrahigh throughput screening of carbohydrate-active enzymes in modular assays modular based on coupled reactions

Enzymes are at the core of the white biotechnologies that will create greener, more efficient processes. The environmental challenge of switching our carbon source from fossil to renewable will require many more enzymes than are currently available, so enzyme discovery and evolution are necessary. Finding new catalysts in 'libraries' (person-made in directed evolution or environmental in functional metagenomics) is a numbers game: hits are rare, so technologies that can screen large numbers of library members will be more successful. The selection criterion in such combinatorial campaigns also matters: the more similar the assay reaction is to the actual application, the more likely is the screening campaign to yield useful enzymes. We have developed a system in which more than 10 million variants can be tested in a day for breakdown of natural substrates (and not for model substrates), by using coupled reaction systems, so we hope to find the right "needle in a haystack", more quickly. In a first application of coupled assays to carbohydrate-active enzymes, we will evolve enzymes used for the degradation of the main components of wood (carbohydrates, mainly in the form of cellulose) and also try to discover new enzymes in metagenomic samples collected in hot springs, the Antarctic and several domestic settings (local compost heaps, ponds, soils). The results of directed evolution will be interpreted with a sequencing technology that we have recently developed ('UMIC-Seq': Nat Commun 2020, 11 (1), 6023), that allows us to obtain full-length readouts of > 10,000 sequence per round of evolution (at a price of less than 1 penny per sequence). Equipped with this insight we hope to understand better how cooperative interactions in the enzymes enable evolution to proceed (or make it difficult in other cases), based on the reconstruction of sequence landscapes, in which areas of high activity are to be explored.

This proposal is based on the application of coupled enzyme assays in water-in-oil-emulsion droplets that are handled in microfluidic devices to achieve a throughput of >10e7 per day for natural (rather than model) substrates to three main targets:

(1) Directed evolution of two CAZymes, namely Xyl43A (from GH43) Xyn10A (from GH10) Paenibacillus barengoltzii starting from libraries that are made by error-prone PCR, recombination (USERec) and insertion deletions (TRIAD). The analysis of evolution trajectories of the variants of HiXyl43A and PbXyn10 retrieved over multiple rounds of directed evolution will take advantage UMI-linked nanopore long-reads (UMIC-seq). These long reads will allow us to determine the co-variation of residues, so that epistatic relationships can be uncovered. Insight into networks of epistatic relationships will provide insight into the 'evolvability' of protein scaffolds.

(2) Functional metagenomics of CAZymes. Metagenomic libraries will be screened to identify new CAZymes or families. Given that our sort is based on functionality alone, we expect that hits will emerge that cannot be predicted by comparison based on sequence homology. It may be possible to uncover new families and we expect that multifunctional, promiscuous enzyme will be identified, which may uncover hitherto undiscovered functional relationships between enzymes that so far are only classified by the main activity and compared by sequence analysis. We also expect that libraries screened with coupled assays for native substrates will give different hits compared to previous low throughput campaigns that used model substrates (such as pNP-sugars).

(3) Evolve three coupling enzymes currently used in our coupled assay system. To provide better signal/noise rations, less background, versatility and robustness, coupling enzymes will be evolved for higher activity under difficult assay conditions, generating more specific, more active and more useful reagents.