O-Glycome Beam Search Arrays for Carbohydrate Ligand Discovery

O-glycosylation is a post-translational modification of proteins crucial to molecular mechanisms in health and disease. O-glycans are typically highly heterogeneous. The involvement of specific O-glycan sequences in many bio-recognition systems is yet to be determined because of a lack of efficient methodologies. We describe here a targeted microarray approach: O-glycome beam search that is both robust and efficient for O-glycan ligand-discovery. Substantial simplification of the complex O-glycome profile and facile chromatographic resolution is achieved by arraying O-glycans as branches, monitoring by mass spectrometry, focusing on promising fractions, and on-array immuno-sequencing. This is orders of magnitude more sensitive than traditional methods. We have applied beam search approach to porcine stomach mucin and identified extremely minor components previously undetected within the O-glycome of this mucin that are ligands for the adhesive proteins of two rotaviruses. The approach is applicable to O-glycome recognition studies in a wide range of biological settings to give insights into glycan recognition structures in natural microenvironments.


Table S1
Microarray analyses of the VP8* proteins of P[10] and P [19] enteroviruses (at 25 µg/mL), antibodies to blood group antigens and the lectin Ulex europaeus agglutinin I (UEA-I) using the Macromolecule array of 30 mucin-enriched epithelial glycoproteins.

Table S2
Molecular ions and deduced monosaccharide compositions of the major components detected by MALDI-MS analyses and hexose contents in PSM neutral O-glycome fractions obtained by Bio-Gel P4 chromatography.

Table S3
Molecular ions and deduced compositions from MALDI-MS analyses of the NGLs derived from PSM neutral O-glycome fractions.

Table S4
Designations, lipid contents, molecular ions and deduced compositions (from MALDI-MS analyses) of PSM neutral O-glycome NGL fractions.

Table S5
Microarray analyses of the VP8* proteins of enteroviruses P[10] and P[19] (at 50 µg/mL), antibodies to blood group antigens and the lectin Ulex europaeous agglutinin I using the PSM neutral O-glycome primary beam search array.

Table S6
Microarray analyses of the VP8* proteins of the P[10] and P[19] enteroviruses using the secondary beam search array designated 'Fucase Array'.

Table S8
Microarray analyses on a small glucan array set.

Fig. S1
Flow chart of the preparation of O-glycome 'beam search' array.

Fig. S2
Relative binding intensities of the VP8* proteins of rotavirus P[10] and P[19] (at 25 µg/mL), antibodies and the lectin Ulex europaeus agglutinin I in the mucin glycoprotein (primary beam search) array.

Fig. S3
Gel filtration chromatography of the products of reductive alkaline hydrolysis from PSM.

Fig. S4
High performance TLC of the large scale NGL reaction mixtures of PSM neutral-glycan alditol fractions d to k.

Fig. S5
High performance TLC of the pooled PSM neutral O-glycome NGL reaction mixture fractionated on a silica cartridge.

Fig. S6
Beam search array data from exploratory experiments to explore binding of the VP8* protein of rotavirus P[19] to arrayed neutral O-glycome NGL fractions from PSM.

Fig. S7
Microarray analysis of VP8* proteins of P[10] and P[19] enteroviruses using PSM neutral O-glycome NGL secondary beam search array. Table S1. Microarray analyses of the VP8* proteins of P[10] and P [19] enteroviruses (at 25 µg/mL), antibodies to blood group antigens and the lectin Ulex europaeus agglutinin I (UEA-I) using the macromolecule array of 30 mucin-enriched epithelial glycoproteins.

5. Glycan Microarray with "Map"
Array layout Each array slide contained 16-pad subarrays of glycan probes printed at two levels in duplicate (four spots for one probe in a row) as detailed in Material and Methods.
Glycan identification and quality control Antibodies and plant lectins variously used to identify glycan sequences in the macromolecule array and in the primary and secondary NGL arrays are described in Material and Methods and the binding data are in Tables S1, 5 and 6.

Scanner settings
Scanning resolution: 10 m / pixel (this resolution is adequate for the sizes of sample spots) The gpr file was entered into an in-house microarray database using software (designed by Dr Mark Stoll, http://www.beilsteininstitut.de/en/publications/proceedings/glyco-2009) for data processing. No particular normalization method or statistical analysis was used.

Glycan Microarray Data Presentation
Data presentation The microarray binding results are in Fig 2, 3 and 4, Table 1 and in Tables S1, 4, 5, 6, 7 and 8, and Figs 2, 6, and 7 The tables include the list of probes in the arrays, binding intensities shown are at 170 pg or 5 fmol per spot and errors (difference of signal intensities of duplicate spots of each glycan probe).

Data interpretation
No software or algorithms were used to interpret processed data. It is assigned as a hexasaccharide with the type 1-terminating blood group H and an internal type 2 sequence, having been derived from a 3-linked branch at core GalNAcol: Fucα1-2Galβ1-3GlcNAcβ1-3Galβ1-4GlcNAcβ1-Galβ-3OX 2. The sequence-defined arrays: corroborated the involvement of type 1-terminating blood group H in the ligand as in lacto N-fucopentaose 1 (LNF-P1) Fucα1-2Galβ1-3GlcNAcβ1-3Galβ1-4Glc The lower intensity of binding to LNFP-1 indiated that the backbone beyond tetrasaccharde is additionally recognized.