Activity Characterization of New Carbohydrate Processing Enzymes (PhD2)

Glycoside hydrolases (GHs) are carbohydrate active enzymes (CAZymes) that catalyse the hydrolysis of the vast diversity of glycans found in nature. To perform this task, GHs have evolved into an equally diverse group of enzymes that is classified by sequence similarity in the manually-curated Carbohydrate Active Enzyme (CAZy) database. GHs are widely used in industry and research, and discovering new ones is a priority for both commercial and academic purposes. Vast amounts of uncharacterised sequences have accumulated in CAZy and other databases thanks to genomic and metagenomic sequencing. Accurately predicting their function before heterologous expression would benefit GH discovery. For machine learning models to be trained for this purpose, functional annotation of understudied regions of the CAZy sequence space should be prioritised. Rational exploration of these areas has proven successful in the past.

In this thesis, we explore GH119 and GH123, two families with few characterised members prior to this study. The families were selected for further exploration after successful functional identification of one of their phylogenetically related proteins. Using rational sequence space exploration, we reveal that, unlike other amylolytic families, GH119 is monofuntional, despite its diversity in modular architecture. We confirm the family's catalytic residues through site-directed mutagenesis and propose a new GH-T clan for family GH119 and the phylogenetically related family GH57, which share catalytic residues, mechanism and an uncommon (β/α)₇-barrel fold, as suggested by GH119 structural models. On the other hand, the discovery of a distant relative of family GH123, the β-N-acetylgalactosaminidase NaNga, reveals a cluster of related, non-classified GH sequences that comprise a potential subfamily of this reported monospecific family. We resolved the three-dimensional structure of NaNga, which reveals a unique pocket topology linked to a sequence conservation pattern detected across the new cluster. These findings suggested that the subfamily may harbour new specificities or even new activities. Despite our efforts, NaNga's natural substrate could not be found. However, a recent publication confirms that family GH123 is multispecific as predicted.

These findings show that rational exploration of uncharted CAZy sequence space regions can yield significant discoveries, enriching GH classification and CAZy's functional annotation. These advancements can contribute to accelerated discovery of new and improved biocatalysts for a faster industrial green transition and tackling of biomedical and environmental challenges.