Human milk oligosaccharides function as prebiotics for beneficial bacteria that occupy the gut of breast fed infants. Certain species of Bifidobacteria are equipped with the glycosidases necessary for oligosaccharide digestion and have demonstrated the ability to metabolize HMOs. The mechanism and specificity of HMO consumption of each subspecies has not been fully characterized. In this study, the digestion of a glycan array (over 100 human milk oligosaccharide species extracted from a pooled milk sample) is rapidly profiled using sensitive, high-throughput nano-LC Chip/TOF mass spectrometry-based methods for a series of glycosidases. The resulting digestion glycoprofile provides insight to the specificity of the enzymes used by gut bacteria to selectively metabolize milk oligosaccharides.
Pooled milk was first defatted via centrifugation and depleted of protein by ethanol precipitation.
The isolated HMOs were then reduced to alditol form using NaBH4. Solid phase extraction on graphitized carbon cartridges (GCC) was used for desalting and enrichment. HMOs were eluted with 40% acetonitrile in 0.05% trifluoroacetic acid (v/v). SPE fractions were dried and reconstituted to approximately 1.5 mg/mL. The reduced HMO pool was then digested by a β-galactosidase isolated from B. longum subsp. infantis ATCC15697 (Blon_2016) for 1 hour at 37°C, pH 4.5. The digestion mixture was then filtered using Millipore C18 ZipTip pipette tips. The HMO pool was profiled using nano-LC Chip/TOF mass spectrometry on a PGC chip both before (undigested control) and after digestion.
Over 200 oligosaccharides were detected in the pooled milk sample before digestion, to which monosaccharide composition was assigned based on accurate mass (typically within ±5 ppm). An annotated in-house HMO library containing 75 neutral and anionic species was used to rapidly assign specific structure using the reproducible LC retention time and CID fragmentation patterns. Using chip-TOF mass spectrometry, the digestion of individual HMO structures was quantitated with respect to the undigested control by normalzing the absolute peak intensity of the digested sample to that of the control. The total oligosaccharide abundance decreased in the digested sample by 32%, almost all of which can be attributed to the digestion of neutral, non-fucosylated species, whose abundance decreased by 80%. Isomers of m/z 710.23 Lacto-N-tetraose and lacto-N-neotetraose, and isomers of m/z 1075.41 lacto-N-hexaose, lacto-N-neohexaose and para lacto-N-hexaose were consumed to the greatest extents (the abundance decreased by more than 90% in the digested sample). The formation of peaks with m/z 548.23 and m/z 751.30 in the digested sample indicate the loss of one or two galactose residues from the structures listed above, respectively, demonstrating that the β-galactosidase cleaves both β1-3 and β1-4 galactoses in the terminal positions. When the above core structures are fucosylated, they are digested more selectively. Monofucosylated isomers lacto-n-fucopentaose I, II, and III were not cleaved by the enzyme, however monofucosylated structures of increasing degrees of polymerization with un-decorated terminal galactoses were almost completely digested, suggesting that the presence of a fucose residue, either linked to the same GlcNAc as the terminal galactose or linked to the terminal galactose itself, blocks the cleavage of that terminal galactose on both type 1 and type 2 chains.The mass spectrometry-based characterization of the enzymatic digestion of the glycan array described above by additional galactosidases, fucosidases, and sialidases is currently underway.