Exploring the sequence-function space of microbial fucosidases | JoVE Visualize

Area of Science:

Enzymology
Glycobiology
Bioinformatics

Background:

Microbial alpha-L-fucosidases are enzymes that cleave alpha-L-fucosidic linkages and can facilitate transglycosylation.
These enzymes belong to the glycoside hydrolases (GHs) superfamily, specifically GH29 in the carbohydrate-active enzyme database.
Understanding the sequence-function relationships within GH29 fucosidases is crucial for enzyme discovery and application.

Purpose of the Study:

To explore the sequence-function landscape of GH29 fucosidases.
To functionally characterize selected GH29 fucosidases and determine their substrate specificities.
To develop and apply computational models for predicting enzyme function and classifying novel GH29 sequences.

Main Methods:

Sequence Similarity Network (SSN) analysis to select diverse GH29 fucosidases for characterization.
High-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) and liquid chromatography-fluorescence detection-tandem mass spectrometry (LC-FD-MS/MS) for substrate specificity analysis.
X-ray crystallography, STD NMR, TLC, ESI-MS, and NMR for structural determination and transfucosylation activity assessment.
Machine learning models, including lightweight protein language models, for sequence-based clustering and classification of GH29 sequences.

Main Results:

Fifteen GH29 fucosidases were functionally characterized, revealing specificities for various alpha-linked fucosylated oligosaccharides and glycoconjugates, aligning with SSN clusters.
The structural basis for the specificity of Bifidobacterium asteroides GH29 fucosidase towards alpha1,6 linkages and FA2G2 N-glycan was elucidated.
Transfucosylation activities were confirmed using GlcNAc and 3FN as acceptors.
Machine learning models accurately assigned 34,258 non-redundant GH29 sequences into SSN clusters, demonstrating the power of computational approaches.

Conclusions:

SSN analysis is a valuable tool for exploring the sequence-function space of GH29 fucosidases.
The integration of experimental characterization with machine learning models enables accurate prediction and classification of enzyme functions.
These combined computational and experimental strategies hold significant promise for the discovery of novel glycoside hydrolases with tailored specificities.