Reverse thiophosphorylase activity of a glycoside phosphorylase in the synthesis of an unnatural Manβ1,4GlcNAc library

β-Mannosides are ubiquitous in nature, with diverse roles in many biological processes. Notably, Manβ1,4GlcNAc a constituent of the core N-glycan in eukaryotes was recently identified as an immune activator, highlighting its potential for use in immunotherapy. Despite their biological significance, the synthesis of β-mannosidic linkages remains one of the major challenges in glycoscience. Here we present a chemoenzymatic strategy that affords a series of novel unnatural Manβ1,4GlcNAc analogues using the β-1,4-d-mannosyl-N-acetyl-d-glucosamine phosphorylase, BT1033. We show that the presence of fluorine in the GlcNAc acceptor facilitates the formation of longer β-mannan-like glycans. We also pioneer a “reverse thiophosphorylase” enzymatic activity, favouring the synthesis of longer glycans by catalysing the formation of a phosphorolysis-stable thioglycoside linkage, an approach that may be generally applicable to other phosphorylases.

The reaction mixture was allowed to warm to RT and stirred overnight. TLC analysis revealed complete consumption of the starting material (100 % EtOAc).

BT1033 synthetic activity screen with unnatural mannose-1-phosphates
Reactions were prepared in duplicate on a 30 µL scale containing 40 mM NaOAc pH 5.5 buffer, 10 mM sugar phosphate donor, 1 mM acceptor, 15 mM MgCl2 and BT1033 (2-3.3 mg/mL). Samples were incubated at 37 °C for 24 h. Specific reaction conditions for each analogue are listed in Table S1. 3 µL of each sample was diluted in 30 µL of MeCN and analysed by HILIC-LC-MS to validate product formation. The remaining sample was diluted 1:1 in 40 mM NaOAc pH 5.5 buffer and then further diluted 1:1 in EtOH and stored at -20 °C for 18 h. Samples were centrifuged at 17 000 x g for 5 min and the supernatant collected. 1 mL of H2O was added to each sample and the samples were lyophilised. The samples were labelled with the ITag (procedure in section 7) and analysed by C18-LC-MS. (Figure S3-S10, Table S1).

Investigating thioglycolygase activity of WT BT1033 and BT1033 D101A
Reactions were prepared on a 30 µL scale containing 40 mM NaOAc pH 5.          The data for BT1033 turnover of GlcNAc-N3 11 and SH-GlcNAc-N3 33 at 0.2 mg/mL (the first 2 bars shown here) are presented in Figure 3C in the main text.

ITag labelling of phosphorylase reaction products.
Lyophilised samples were resuspended in 24 µL of H2O, then 0.5 -1 µL of Itag solution (250 mM in DMSO) was added to each sample. Then 10 µL of a stock solution containing 500 mM CuSO4 . 5 H2O and 500 mM THPTA in H2O, was added to each sample. Samples were mixed by vortexing and then 5 µL of a 1 M sodium ascorbate solution was added to each sample. Samples were further mixed by vortexing and then incubated for 2 h at 20 °C 300 rpm. 1 mL of H2O was added to each sample and the samples were lyophilised. Samples were resuspended in 100 µL of H2O. and analysed by C18-LC-MS.
Relative conversion of starting material to products was determined by analysing the peak intensities of starting material and product species, and using the following equation:

Determination of BT1033 specific activity
For the acceptor standard curves, stock solutions of 5 mM acceptor (11 or 33) were prepared in HPLCgrade H2O. The stock solution for SH-GlcNAc-N3 33, also contained TCEP reducing agent (2.5 mM). A two-fold serial dilution of the respective stock solutions was performed to produce a series of standards ranging from 5 mM to 0.15625 mM, for each acceptor. The standards were prepared in triplicate. 10 µL of each standard was incubated at 37 °C for 30 min and then lyophilised.  To investigate specific activity, a series of reactions was assembled on a 10 µL scale containing 10 mM Man-1P 2, 3 mM acceptor 11 or 33, 10 mM MgCl2, 40 mM NaOAc buffer pH 5.5 and 1 µL of BT1033 (stock concentrations ranging from 13 mg/mL -0.025 mg/mL). For reactions containing SH-GlcNAc-N3 33, TCEP reducing agent was added to a final concentration of 1.5 mM. Reactions were incubated at incubated at -20 °C. Samples were then lyophilised, following the addition of 1 mL of HPLC-grade H2O.
Lyophilised samples were resuspended in 24 µL of H2O, then 0.5 µL of Itag solution (250 mM in DMSO) was added to each sample. Then 10 µL of a stock solution containing 500 mM CuSO4 . 5 H2O and 500 mM THPTA in H2O, was added to each sample. Samples were mixed by vortexing and then 5 µL of a 1 M sodium ascorbate solution was added to each sample. Samples were further mixed by vortexing and then incubated for 2 h at 20 °C 300 rpm. 1 mL of H2O was added to each sample and the samples were lyophilised. Samples were resuspended in 100 µL of H2O. and analysed by C18 and the absolute concentration of acceptor was obtained from the standard curve ( Figure S34).
Conversion (%) to product was calculated as follows: Conversion (%) was plotted against enzyme concentration ( Figure S35). Specific activity (Table S5) was calculated at the enzyme concentration that resulted in ~ 15% conversion (i.e. the initial change in conversion).  Figure S35. BT1033 titration curve for reactions with GlcNAc-N3 (A) or SH-GlcNAc-N3 (B).

Molecular modelling
The BT1033 amino acid sequence was obtained from UniProt and the PDB file was obtained from the AlphaFold Protein Structure Database (UniProt entry: Q8A8Y4). Uhgb_MP amino acid sequence and PDB file was obtained from RCSB Protein Data Bank (PDB) (entry 4UDJ). Molecular graphics and analyses were performed with UCSF Chimera. 9 NCBI BLAST web interface (https://blast.ncbi.nlm.nih.gov/Blast.cgi) was used for sequence alignment of BT1033 to Uhgb_MP ( Figure S36). Figure S36. An amino acid sequence alignment of BT1033 and Uhgb_MP. The conserved catalytic Asp is highlighted in turquoise.