Distinguishing Oligosaccharide Isomers Using Far-Infrared Ion Spectroscopy: Identification of Biomarkers for Inborn Errors of Metabolism

Distinguishing isomeric saccharides poses a major challenge for analytical workflows based on (liquid chromatography) mass spectrometry (LC–MS). In recent years, many studies have proposed infrared ion spectroscopy as a possible solution as the orthogonal, spectroscopic characterization of mass-selected ions can often distinguish isomeric species that remain unresolved using conventional MS. However, the high conformational flexibility and extensive hydrogen bonding in saccharides cause their room-temperature fingerprint infrared spectra to have broad features that often lack diagnostic value. Here, we show that room-temperature infrared spectra of ion-complexed saccharides recorded in the previously unexplored far-infrared wavelength range (300–1000 cm–1) provide well-resolved and highly diagnostic features. We show that this enables distinction of isomeric saccharides that differ either by their composition of monosaccharide units and/or the orientation of their glycosidic linkages. We demonstrate the utility of this approach from single monosaccharides up to isomeric tetrasaccharides differing only by the configuration of a single glycosidic linkage. Furthermore, through hyphenation with hydrophilic interaction liquid chromatography, we identify oligosaccharide biomarkers in patient body fluid samples, demonstrating a generalized and highly sensitive MS-based method for the identification of saccharides found in complex sample matrices.


Figure S1 | IRMPD MS/MS spectra of the (a) [M+Li] + , (b) [M+Na] + , (c) [M+Cs] + , (d) [M+NH 4 ] + , (e) [M+Cl]and (f) [M-H]ions of D-glucose (blue), D-mannose (red) and
. The most important peaks are labeled with their m/z value. Irradiation wavelengths and precursor ion masses are mentioned above each spectrum.  Table S1. - D-glucose, D-mannose, D-galactose, D-lactose, D-maltose, D-cellobiose, D-globotriose, D-raffinose, D-maltotriose, D-tetraglucoside and D-maltotetraose reference standards were obtained from Sigma-Aldrich Chemie (Steinheim, Germany). Purified D-glc-glc-glc-man was obtained from the urine of a MOGS-CDG patient as described in Ref. 1  . After an initial time of 1 minute at 100% A, a gradient was run to 50% A in 8 minutes, followed by a gradient to 11.1% A in 2 minutes. After a hold of 1 minute at 11.1% A, the column was re-equilibrated for 4 minutes at 100% A. An injection volume of 10 µl and a flow rate was used and the column was held at a temperature of 55 °C. Retention times of all reference standards reported in Table S5 were also measured using this separation method.

Chemicals
Here, an injection volume of 5 µl was used.

Infrared Ion Spectroscopy
Solutions of reference standards and the diluted urine sample were introduced at 180 µL h -1 flow rates to the electrospray source using a syringe pump. Fractions of LC eluent were collected using a two-position six-port switching valve controlled by the quadrupole ion trap and an 80 μL sample loop. In this set-up, the LC-eluent is diverted to waste by the switching valve until the peak of interest elutes. At that time, the valve is switched, storing the fraction of eluent in the sample loop. The valve switches back after the peak is eluted and the fraction is slowly infused by the syringe pump (120-180 µL h -1 ).
Ions of interest were mass-isolated and stored for 180 ms in the ion trap to be irradiated with two macropulses of FELIX, which produced 300-1000 cm -1 radiation in the form of 5-10 μs macropulses of 10-80 mJ at 10 Hz repetition rate (bandwidth ~0.4% of the central frequency). Our IRIS experiments are based on infrared multiple-photon dissociation (IRMPD) spectroscopy: resonant absorption of IR laser light raises the internal energy of the ions and, after absorption of multiple photons, unimolecular dissociation can occur. The fractional amount of fragmentation is a measure for the degree of IR absorption at each wavelength point and an IR spectrum can be reconstructed by relating the precursor and fragment ion intensities [IRIS intensity = ln(ΣI(precursor + fragment ions)/I(precursor ion))] as a function of wavelength. 3 When no fragment ions are observed (because the fragments are below the low mass cut-off of the ion trap) the IR spectrum is plotted as a precursor depletion spectrum [IRIS intensity = ln(I(precursor, without irradiation)/I(precursor)]. 3 To obtain IR spectra, FELIX was scanned in 4 or 5 cm -1 steps from ~10-34 µm and a fragmentation spectrum after irradiation was obtained as an average of 6-12 scans. The IRIS intensity was linearly corrected for frequency-dependent laser pulse energy variation. 3 No smoothing was applied.

Quantum chemical calculations
In order to understand the general nature of the vibrations in the 300-1000 cm -1 IR region, quantum-chemical calculations were performed using the Gaussian16 software package 4 . The lowest energy conformer was selected by performing a conformational search using a workflow described previously. 5 Geometry optimizations and (harmonic) IR frequency calculations were performed using density functional theory at the B3LYP/6-31++G(d,p) level of theory. Computational frequencies were scaled by 0.975. Among the lowest-energy structures resulting from the conformational search, the conformer was assigned that showed the best match between experimental and computational spectrum.

4-methylphenyl 2,3,4-tri-O-acetyl-1-thio-D-xylopyranoside (S1)
To an ice-cold solution of D-xylose (20.1 g, 134 mmol) in DCM (100 mL) and TEA (150 mL), was added DMAP (1.64 g, 13.4 mg). To this solution Ac 2 O (90 mL, 950 mmol) was added dropwise. The solution was allowed to warm up to rt and was stirred overnight. The reaction was quenched with H 2 O (200 mL), the layers were separated. The aqueous phase was extracted with DCM, the combined org layers were washed with 1 M HCl and sat. aq. NaHCO 3 , dried over MgSO 4 and concentrated in vacuo. The product was obtained as a thick oil in 97% yield (41.7 g, 130 mmol) with an 1:0.13 α:β ratio. Analytical data of major α anomer given.

D-xylose-(ɑ1-3)-D-xylose-(ɑ1-3)-D-glucopyranose (2)
Acetylated trisaccharide 12 (200 mg, 272 µmol) was dissolved in AcOH:H 2 O (100 mL, 8:2) and the resulting solution was heated to 95 o C overnight. The solution was allowed to cool to rt, the solvent was removed in vacuo. The residue was dissolved in pyridine (5 mL) and Ac 2 O (385 µL, 4.08 mmol) was added as well as a catalytic amount of DMAP (3.3 mg, 27 µmol). The reaction was stirred overnight and subsequently concentrated in vacuo. The residue was taken up in EtOAc, washed with sat. aq. CuSO 4 , sat. aq. NaHCO 3 and sat. aq. NaCl. The organic layer was dried over MgSO 4 and concentrated in vacuo. The resulting residue was purified with column chromatography (40-70% EtOAc in heptane) resulting in a white foam contaminated with an inseparable peracetyled glucose (R f = 0.51 (8:2 EtOAc:heptane)). The mixture was dissolved in MeOH (5 mL) and solid K 2 CO 3 (10 mg) was added. The solution was stirred for 2 h, neutralized with Amberlyst (15 hydrogen form), filtered and concentrated in vacuo. The mixture was purified by column chromatography (0-20% H 2 O in MeCN) and subsequent gel filtration chromatography (Biorad P2 gel). The product was lyophilized to obtain 2 as a fluffy white solid in 35% yield (43 mg, 97 µmol).