Exploring the sequence-function space of microbial fucosidases

Abstract

Microbial α-L-fucosidases catalyse the hydrolysis of terminal α-L-fucosidic linkages and can perform transglycosylation reactions. Based on sequence identity, α-L-fucosidases are classified in glycoside hydrolases (GHs) families of the carbohydrate-active enzyme database. Here we explored the sequence-function space of GH29 fucosidases. Based on sequence similarity network (SSN) analyses, 15 GH29 α-L-fucosidases were selected for functional characterisation. HPAEC-PAD and LC-FD-MS/MS analyses revealed substrate and linkage specificities for α1,2, α1,3, α1,4 and α1,6 linked fucosylated oligosaccharides and glycoconjugates, consistent with their SSN clustering. The structural basis for the substrate specificity of GH29 fucosidase from Bifidobacterium asteroides towards α1,6 linkages and FA2G2 N-glycan was determined by X-ray crystallography and STD NMR. The capacity of GH29 fucosidases to carry out transfucosylation reactions with GlcNAc and 3FN as acceptors was evaluated by TLC combined with ESI-MS and NMR. These experimental data supported the use of SSN to further explore the GH29 sequence-function space through machine-learning models. Our lightweight protein language models could accurately allocate test sequences in their respective SSN clusters and assign 34,258 non-redundant GH29 sequences into SSN clusters. It is expected that the combination of these computational approaches will be used in the future for the identification of novel GHs with desired specificities.

Fig. 1. Sequence similarity network (SSN) of the GH29 fucosidase family.

A The coloured SSN of GH29 family after cluster analysis. B The distribution of functionally characterised GH29s in different clusters. Red nodes represent enzymatically characterised GH29s, purple nodes represent structurally characterised GH29s while green nodes represent new GH29 enzymes characterised in this work.

Fig. 2. Enzymatic characterisation of GH29 fucosidases.

A Kinetic parameters of GH29 fucosidases on CNP-Fuc. ***, p < 0.001; ****, p < 0.0001. B Fucosylated oligosaccharides used in this study. Monosaccharide symbols follow the Symbol Nomenclature for Glycans (SNG) system. C Substrate specificity of GH29 fucosidases. The data are presented in base-2 logarithm function. D HPAEC-PAD analysis of GH29 enzymatic reaction products with 6FN. The data were analysed with Prism. Standards were Fuc (red, 100 µM), 6FN (green, 100 µM), GlcNAc (blue, 100 µM). The black lines correspond to the enzymatic reactions with 6FN incubated with the different GH29 fucosidases tested or with buffer only. See Supplementary Fig. S2 for HPAEC-PAD analysis of GH29 enzymatic reaction products with all other substrates tested (2’FL, 3FL, BgA, BgB, BgH, LeA, sLeA, LeX, sLeX, LeY, pNP-Fuc and pPGM).

Fig. 3. LC-FD-MS/MS analysis of BaGH29^26A reaction with FA2G2.

A Reaction was performed without enzyme. B Reaction was performed with enzyme. Glycan products are annotated next to peaks on the chromatograms.

Fig. 4. Crystal structure of BaGH29^26A.

A Crystal structure of BaGH29^26A in complex with Fuc. Boxout shows ligand bound WT BaGH29^26A in light blue and unbound D218N in grey. The bound Fuc residue is shown in green. The catalytic acid base and nucleophile residues are highlighted in orange and magenta, respectively. Hydrogen bonding interactions are indicated with black dashed lines. B Proposed rotation of active site Tyr57 in the presence of substrate molecules suggested by alignment to AlfC, show bound to 6FN in yellow (PDB 6OHE) and bound to fucose in pink (PDB 6O1A). BaGH29^26A is shown in light blue.

Fig. 5. STD-NMR analysis of the interaction between BaGH29^26A and FA2G2.

A Binding epitope mapping of FA2G2 as bound to BaGH29^26A from STD NMR experiments. Protein contact strength reflects relative values of saturation transfer after normalisation to the most intense one (the methyl group of GlcNAc(A)) obtained from STD initial slopes (full STD NMR build-up curves and initial slopes for each proton can be found in Fig. S5). B STD NMR difference (black) and reference (red) spectra of the FA2G2/BaGH29^26A D218A sample, acquired at 2 s saturation time. Only isolated protons that were unambiguously assigned could be analysed for binding epitope determination and are labelled on the spectra (impurities are marked with *). The STD NMR analysis supports that the enzyme preferentially recognises the reducing end, with sugar rings of Fuc and GlcNAc(A) showing the strongest STD intensities.

Fig. 6. Transfucosylation activity of GH29 fucosidases.

A TLC analysis of GH29 transfucosylation reactions with GlcNAc as acceptor and pNP-Fuc as donor; ATCC_03833^3A was used as control. Lanes 1 to 6 correspond to standards: Fuc (lane 1), pNP-Fuc (lane 2), GlcNAc (lane 3), 4FN (lane 4), 3FN (lane 5) and 6 is 6FN (lane 6). Lane 7 is the control reaction with ATCC_03833^3A. Lanes 8 to 22 are the GH29 reactions with PgGH29^1B (lane 8), RiGH29^2A (lane 9), SmGH29^1B (lane 10), SsFuc^1B (lane 11), TfFuc1^8A (lane 12), LaGH29^3A (lane 13), BaGH29^26A (lane 14), FbGH29^26A (lane 15), RsGH29^3A (lane 16), Afc1^45B (lane 17), SgGH29^9A (lane 18), BsGH29^44B (lane 19), StGH29^nc (lane 20), NyGH29^4A (lane 21) and E1_10125^1B (lane 22). The upper grey dotted line corresponds to the 6FN standard and the lower grey dotted line corresponds to the 4FN substrate control. B 600 MHz ¹H NMR spectra of BaGH29^26A reaction and standards of 3FN, 4FN and 6FN. The mid field region displays distinctive signals showing the presence of 6FN and trace levels of 3FN in BaGH29^26A. C TLC and TLC-ESI-MS analysis of BaGH29^26A transfucosylation reactions with 6FN as acceptors and pNP-Fuc as donor. Lanes 1 to 5 and 9 correspond to standards: Fuc (lane 1), pNP-Fuc (lane 2), GlcNAc (lane 3), 4 is 4FN (lane 4), 3FN (lane 5) and 6FN (lane 9). Lanes 6 to 8 are the BaGH29^26A reactions with GlcNAc (lane 6), 3FN (lane 7) and 6FN (lane 8). The upper black dotted line corresponds to the 6FN standard, and the lower black dotted line corresponds to the 4FN standard. Glycan symbols follow the SNFG.

Fig. 7. Sequence representations of GH29 family.

ProtT5 was employed to map each of 34,258 non-redundant GH29 sequences into a 1024-dimension representation followed by SSN cluster ID allocation. These representations were further projected onto 2-dimension space (x- and y-axes) using UMAP for visualisation and colour-coded by predicted SSN ID clustering and taxonomy as shown in (A) and (B), respectively. Each dot on the map represents a sequence. The original (A) is supplied in supplementary data 6 for maximum resolution. The number in each region in panel B) corresponds to SSN clustering ID.