Post-Column Make-Up Flow (PCMF) Enhances the Performance of

Subscriber access provided by WEBSTER UNIV

Article

Post-Column Make-up Flow (PCMF) Enhances the Performance of Capillary-Flow PGC-LC-MS/MS-Based Glycomics Hannes Hinneburg, Sayantani Chatterjee, Falko Schirmeister, Terry NguyenKhuong, Nicolle H. Packer, Erdmann Rapp, and Morten Thaysen-Andersen Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b05720 • Publication Date (Web): 27 Feb 2019 Downloaded from http://pubs.acs.org on February 28, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Post-Column Make-up Flow (PCMF) Enhances the Performance of Capillary-Flow PGC-LC-MS/MS-Based Glycomics Hinneburg, Hannes1,2; Chatterjee, Sayantani1,2; Schirmeister, Falko; Nguyen-Khuong, Terry3; Packer, Nicolle H.1,2,4; Rapp, Erdmann5,6; Thaysen-Andersen, Morten1,2* 1

Department of Molecular Sciences, Macquarie University, Sydney, NSW-2109, Australia Biomolecular Discovery and Design Research Centre, Macquarie University, Sydney, NSW-2109, Australia 3 Bioprocessing Technology Institute (BTI), A*STAR, Singapore 138668, Singapore 4 ARC Centre for Nanoscale Biophotonics, Macquarie University, Sydney, NSW-2109 Australia 5 Max Planck Institute for Dynamics of Complex Technical Systems, 39106 Magdeburg, Germany 6 glyXera GmbH, 39106 Magdeburg, Germany

2

*

Corresponding author ([email protected])

ABSTRACT: Deep characterisation of biologically-relevant glycans remains challenging. Porous graphitised carbon–liquid chromatography tandem mass spectrometry (PGC-LC-MS/MS) enables quantitative elucidation of glycan fine structures. However, the early PGC-LC elution of smaller glycans (tri-, tetra- and pentasaccharides) at low organic solvent content hampers their detection. In efforts to improve the glycan profiling sensitivity and accuracy, we present a new capillary-flow PGC-LC-MS/MS-based configuration comprising a post-column make-up flow (PCMF) that supplies an ion-promoting organic solvent to separated glycans prior to their detection by MS. The analytical performance of this setup was systematically evaluated against our existing capillary-flow PGC-LCMS/MS platform (Jensen et al., Nat Protoc (7)1299:2012). Specifically, the ion intensities and signal-to-noise ratios of various classes of non-derivatised glycans from N- and O-glycoproteins and fructooligosaccharide mixtures were compared using methanol (MeOH)-, isopropanol (IPA)-, and acetonitrile (ACN)-based PCMF at various concentrations. In particular, ACN- and IPA-based PCMF dramatically increased the signal response across all glycan types (30-100x), improved the MS/MS spectral quality and reduced the quantitative glycoprofile variation between replicates. In particular, the detection of the early-eluting glycans benefitted from the PCMF. The highest sensitivity gains were achieved with the supplements of 100% ACN and IPA (equating to 57% (v/v) net concentration at the ion source) whilst neither compromising the favourable PGC-LC properties including the high peak capacity and glycan isomer separation nor changing the MS detection behaviour. In conclusion, PCMF-based PGC-LC-MS/MS dramatically improves the glycomics sensitivity, coverage and quantitative accuracy not least for the difficult-todetect early-eluting and low-abundance glycans detached from N- and O-glycoproteins.

Detailing the fine structures of complex carbohydrates (glycans) is critical to advance our understanding of their involvement in glycobiological processes1-2. However, elucidation of glycan structures in biological mixtures is still challenging owing to their chemical complexity and diversity. Several methods, all with distinct advantages and limitations, are available to characterise glycans detached from proteins3-5. For determination of the monosaccharide composition, topology and other key glycan features, porous graphitised carbon–liquid chromatography tandem mass spectrometry with negative ion detection (PGC-LC-

MS/MS) is an informative glycoprofiling technique used across many glycomics laboratories3, 6-11. The retention behaviour and mechanisms of PGC are complex and still not fully understood. It is known that the PGC stationary phase retains charged, polar and non-polar compounds in a “mixed-mode” involving ionic, dipole-dipole, and hydrophobic interactions12-13. This retention characteristics gives PGC a unique capability to separate glycans, at least in part, according to their shape, an extremely useful feature for isomer separation of most biologically-relevant glycans, an

1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

unmatched feature not found in other separation techniques. PGC LC retention can even be used for glycan identification using recently established retention libraries15 although the variable glycan retention (up to minutes) from run-to-run likely caused by PGC column inconsistencies and incomplete sample clean-up represents a challenge for this identification method. Further, it was established that biosynthetically “extreme” glycans i.e. mono-/di-saccharides and very large, branched and/or acidic glycans may go undetected in PGC-LC-MS/MS glycomics experiments due to insufficient or, in the other end, possibly irreversible PGC retention and due to limitations of the mass spectrometer outside the optimal m/z detection window14-15. More than a decade ago, Pabst and co-workers determined the retention behaviour of some common glycan structural features on the PGC-LC-MS/MS platform16. They subsequently showed that both the separation and detection of glycans on this setup are influenced by multiple parameters including the column temperature, the solvent characteristics, the presence of salts and to a lesser extent by the pH of the mobile phase17. Pabst et al. observed that late eluting species, typically the more elongated glycans, are favoured in the glycomics analysis by giving rise to relatively higher signal intensities, in part, due to the higher levels of organic solvents present late in the PGC-LC gradient. Thus, the increasing concentration of organic modifier during a PGC-LC-MS/MS run typically formed by a linear gradient of acetonitrile (ACN) or, less frequently, by methanol (MeOH) or isopropanol (IPA)18 may create an unwanted bias towards larger oligosaccharides of higher molecular mass, and, conversely, may underrepresent the early-eluting glycans often of low molecular mass. Harvey showed via a series of direct infusion experiments that organic solvents such as 50% (v/v) MeOH are beneficial for negative ion detection of glycans19. However, in typical PGC-LC-MS/MS applications, most biologically-relevant glycans elute below 30% (v/v) organic content6, 18, which is below the concentration required for optimal spray stability and glycan ionisation20. More recently, Staples et al. showed that the MS detection of unbranched polyanionic glycans i.e. heparan sulfate eluting under high aqueous conditions in hydrophilic interaction liquid chromatography could be improved by the introduction of a post-column make-up flow (PCMF) of ACN in a nano-chip format21. Supporting the usefulness of PCMF, we recently demonstrated that the spray stability on a nano-flow

Page 2 of 12

PGC-LC-ESI-MS/MS platform benefits from the implementation of an ACN-based PCMF22. To the best of our knowledge, the impact of the PCMF configuration on the glycomics coverage, sensitivity and reproducibility, however, remains to be systematically evaluated. We recently discovered the presence of an unusual class of N-glycans attached to proteins from human tissues23-24. These truncated N-glycans (paucimannosidic glycans), which do not fit into the conventional mammalian N-glycan classes, comprise the trimannosylchitobiose core and further truncated mannosyl-terminating variants of this core (Man1-3GlcNAc2) with variable core fucosylation. Even further truncated Nglycans of the chitobiose type (GlcNAc1-2Fuc0-1) have been reported on mammalian proteins25-26. Similar to the short O-glycans e.g. Tn, T, sialyl T27-28, the short Nglycans are increasingly being associated with pathological conditions and aberrant human physiology including in cancer, inflammation and infection29-31. However, we have experienced that conventional PGC-LC-MS/MS-based glycomics often underrepresents these tri-to-pentasaccharides due to their early LC elution (unpublished observations). The aim of this work was therefore to increase the coverage and sensitivity of our current capillary-flow PGC-LC-MS/MS configuration6 to improve the profiling of all N- and O-glycans typically encountered in biology- and disease-focused glycomics including the early-eluting glycans typically of low mass and other low abundance glycans. To this end, we implemented a simple PCMF within the PGC-LC-MS/MS setup and tested various solvent conditions to achieve the highest possible glycome-wide sensitivity and coverage without compromising the favourable separation properties of the original glycomics profiling platform. Experimental Section Chemicals and reagents Fructooligosaccharides (FOS, PN: F8052), bovine fetuin (PN: F3385, UniProtKB: P12763), bovine ribonuclease B (RNase B, PN: R7884, UniProtKB: P61823), and human immunoglobulin G (IgG, PN: I4506, UniProtKB: P01857) were from Sigma-Aldrich (St. Louis, MO). Human neutrophil elastase (HNE, PN: 342-40, UniProtKB: P08246) and human myeloperoxidase (PN: 426-10LV-1, UniProtKB: P05164) were from Lee BioSolutions (Maryland Heights, MO). Solvents were LC-MS grade (Merck Millipore). Water was purified with a MilliQ-Synthesis system (Merck-Millipore). The glycoproteins were dissolved in phosphate buffered saline (PBS) (5-10 µg/µL) and FOS was dissolved


Page 3 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in water (2 µg/µL, both final concentrations). All analytes were aliquoted and stored at -30°C until analysis. In solution and membrane N- and O-glycan release The glycoproteins were individually heat-denatured for 5 min, 95°C. The protein-linked N-glycans were released in solution from 80 µg of each glycoprotein (100 µg for fetuin) by the addition of 20 U (2 µL) peptide:Nglycosidase F (PNGase F) (Promega, Madison, WI) in PBS at 37°C, 4 h. The liberated N-glycans were desalted using HyperSep HyperCarb (Thermo Scientific, Chelmsford, MA) packed in SPE format in 10 µL (P10) pipette tips containing Empore C18 plugs (Sigma-Aldrich, St. Louis, MO)6. The N-glycans were reduced using 1 M NaBH4 (final concentration) at 50°C, 3 h and desalted by strong cation exchange-C18-SPE and HyperCarb-SPE6 (see Extended experimental section in Supporting Information [SI]). The N-glycans were mixed (1:1:1:2, weight-based ratio of HNE:IgG: RNAseB:fetuin), dried and dissolved in 140 µL water. The O-glycans of bovine fetuin (4 x 20 µg) and N-glycans of human myeloperoxidase (3 x 10 µg) were released from PVDF immobilised proteins as described previously6. In short, PNGase F was used to first remove all N-glycans, which, for fetuin, were not used for analysis in this study. The remaining O-glycans from the same fetuin spots were released by overnight incubation in 0.5 M NaBH4 in 50 mM KOH (final concentrations), 50°C. The O-glycans were mixed, dried and dissolved in 80 µL water. N-glycans of myeloperoxidase were reduced and desalted, see above. From the N- and O-glycan samples, 2 µL (equating to glycans released from ~2.9 µg of total N-glycoprotein and ~0.6 µg of total O-glycoprotein) was used for each injection. The N- and O-glycans were profiled separately by LC-MS/MS as described below. Experimental setups The experimental setups explored in this study (hereafter “Setup 1”, and “Setup 2”) as well as our existing capillary-flow PGC-LC-MS/MS setup (hereafter “Original setup”)6 used with minor modifications across several glycomics laboratories3, 8-11, 16 are depicted in Figure S1 and briefly described below: Setup 1: Direct infusion experiments of FOS (1 ng/µL) in 10 mM aqueous ammonium bicarbonate (ABC, pH 8.0) with various concentrations of MeOH, ACN and IPA (i.e. 0-100%, 10% increments, v/v) at various flow rates (i.e. 1-5 µL/min, 1 µL/min increments). Setup 2 and Original setup: These two capillary-flow PGC-LC-MS/MS setups both used 10 mM aqueous

ABC (solvent A) and 10 mM ABC in 70% (v/v) ACN (solvent B) as the mobile phases for the N- and O-glycan separation on a capillary PGC-LC column (HyperCarb, 3 µm particle size, 180 µm I.D. x 100 mm, Thermo Scientific) heated to 50°C. Both setups used the following LC gradient for N-glycan separation: Start condition of 2.6% (v/v) B for 8 min, a linear increase to 13.5% (v/v) B over 2 min, to 37.3% (v/v) B over 55 min, and to 64% (v/v) B over 10 min, before the column was cleaned in 98% (v/v) B for 6 min and equilibrated in 2.6% (v/v) B. For both setups, a Dionex UltiMate 3000 LC system (Thermo Scientific) was used to deliver a constant column flow rate of 3 µL/min (Setup 2) and 4 µL/min (Original setup). In both setups, the O-glycan analysis used the same solvents and starting conditions as above, but used a different gradient: 2.8-30% (v/v) B over 38 min and then a linear ramp to 98% (v/v) B over 5 min, followed by a 2 min column clean in 98% (v/v) B and column re-equilibration in 2.8% (v/v) B. These conditions prevented carryover between the N- and O-glycan runs (data not shown). Blank samples (water) were nonetheless injected between the experimental conditions. Uniquely for Setup 2 (Figure 1), a PCMF was integrated in the configuration by the post-column installation of a T-piece (Thermo Scientific, catalogue number NC0339106). The A/B mixing and venting tees were made of stainless steel. The PCMF capillary from the LC pump (solvent bottles) was a nanoViper line (Thermo Scientific, I.D. 50 µm/length 750 mm). The capillary connections from the LC column to the Tpiece and from the T-piece to the mass spectrometer were also nanoViper lines with 50 µm/550 mm and 30 µm/100 mm dimensions, respectively. This installation enabled the supply of an additional organic modifier i.e. MeOH, ACN or IPA at a constant flow rate (4 µL/min) to the mobile phase passing through the LC column during the entire LC-MS/MS run. The organic concentrations of the PCMF solvents were varied between 70-100% (v/v) with 10% (v/v) increments to reach various final concentrations of organic modifiers at the ESI-sprayer i.e. 40-57% (v/v). The loading pump of the LC (Dionex UltiMate 3000 LC system), which was configured to capillary flow rates of aqueous/organic solvents, was used to deliver the PCMF solvents at the specified flow rates and concentrations by mixing organic solvent (IPA, MeOH or ACN) with water in appropriate ratios. For the PCMF experiments, the organic solvent concentrations are stated at the ESI source at the beginning of the LC gradient. The actual flow rate of the PCMF delivered by the LC was manually verified over a 5 minute collection period.



Page 4 of 12

The resultant glycomics data was visualised in heat maps and bar graphs using the language and environment for statistical computing R v3.4.2 (https://www.R-project.org/) and Microsoft Excel 2013, respectively. Data points represent the mean of technical replicates; Error bars indicate their standard deviation. Data from the various PCMF conditions (Setup 2) were statistically compared to the Original setup using two-tailed Student’s T-tests; p < 0.05 was used as the minimum significance threshold. For further details of the statistical testing, see SI. Figure 1. Experimental setup (Setup 2) used to test the benefits of PCMF for glycomics. Mixtures of N- and Oglycans were analysed using capillary-flow PGC-LCMS/MS (-) with a PCMF of 70-100% (v/v) MeOH, ACN or IPA (*equating to ~40-60% organic modifier at the ESI source at the beginning of the LC gradient) by supplementing the column flow (3 µL/min) with organic PCMF solvents (4 µL/min). See Figure S1 for an illustration of Setup 1 and the Original setup.

MS parameters, data analysis and statistics All experiments were performed on a Velos linear ion trap (LTQ) (Thermo Scientific) equipped with a HESIII probe (position B depth). The LTQ was calibrated and operated in negative ion polarity mode. Resonance-activation collision induced dissociation (CID) MS/MS of the top five most abundant precursor ions in each MS1 full scan was performed. The MS parameters were kept constant during all analyses, see SI for more. The N- and O-glycan mixtures were analysed in a minimum of technical triplicates at the LC-MS/MS injection level (Setup 2 and Original setup). Experiments relating to Setup 1 were performed without technical replicates. For Setup 1, MS1 full scans were recorded over an infusion time of 1 min for each condition and the resulting mass spectra were then summed for interrogation. Mass spectra were viewed and the signal-tonoise (S/N) ratios of the individual glycan precursor ions were measured using Xcalibur v2.2 (Thermo Scientific) (baseline window: 40, area noise factor: 5, peak noise factor: 10). Representative N- and O-glycans (~50 structures) from the studied glycoproteins were identified based on their reported structures32-33 using multiple lines of orthogonal information i.e. monoisotopic precursor mass, CID-MS/MS fragmentation pattern and relative PGC-LC retention time34. The glycan signal response factors and their relative abundance were determined from the area-under-the-curve (AUC) of monoisotopic extracted ion chromatograms (EICs) of all observed charge states of each monitored glycan using Skyline (64-bit) v3.7.0.11317. See SI for detail.

Safety considerations The organic solvents used in this study should be handled with care. Standard laboratory safety gear should be used to avoid direct exposure, in particular to MeOH, the most hazardous of the PCMF solvents. Results and Discussion Rationale and experimental design PGC-LC-MS/MS is a valuable analytical platform for glycan analysis3, 6, 9, 11, 16, 18. However, reaching a glycome-wide coverage and accurate relative quantitation of all glycans present in biological samples remains challenging, in part, due to the low ionisation potential of particularly glycans that elute early in PGC-LC at low organic solvent concentrations20. Furthermore, spray instability in the water-rich part of the gradient is another significant issue that contributes to low quantitative glycoprofiling reproducibility. The early eluting species include short N-glycans (chitobiose and paucimannosidic glycans23-24) and O-glycans (T and sialyl T27-28, 31) that increasingly are associated with human diseases. Thus, we have realised that the current PGC-LC-MS/MS setup (here “Original setup”6) has shortcomings. To this end, we aimed to develop a sensitive glycome-wide profiling technique compatible with the analysis of most, if not all, N- and O-glycans typically encountered in clinical biospecimens. We hypothesised that the supply of ion-promoting organic modifiers introduced via a PCMF after the PGCLC column would enhance the glycan detection, in particular of the often underrepresented early-eluting glycan analytes. We further hypothesised that the PCMF would also be beneficial for the glycoprofiling reproducibility by providing a stable ESI spray throughout the water-rich part of the PGC-LC gradient as we have recently demonstrated in a proof-of-principal study on a related nano-flow PGC-LC-MS/MS platform22.


Page 5 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


To this end, we systematically investigated the impact of supplying different concentrations of three commonly used organic modifiers i.e. MeOH, ACN, and IPA at a constant flow rate throughout the chromatographic run to the regular PGC-LC column solvents (ACN and 10 mM aqueous ABC). These three organic modifiers were chosen since they are all MScompatible, available at high purity at reasonable cost, and display slightly different physicochemical properties e.g. surface tension, density, viscosity, polarity, dielectric constant, boiling-point, vapour pressure, hydrophilicity/hydrophobicity and proticity35. The effect of these solvents were studied upon direct delivery of oligosaccharides dissolved in the organic modifiers to the mass spectrometer by infusion (Setup 1) or via a PCMF-based PGC-LC-MS/MS setup (Setup 2, see Figure S1 for schematics of the tested setups). Setup 1 was used to determine the organic modifier conditions yielding the highest glycan signal response while minimising the level of interfering (non-glycan) signals and avoiding glycan precipitation. This optimisation step was carried out using FOS comprising six main oligosaccharide species spanning 3-8 hexose units. Setup 2 used a PCMF in PGC-LC-MS/MS-type experiments to systematically test the effect(s) of various organic solvent conditions on the detection of biologically relevant N- and O-glycans from four well-characterised mammalian glycoproteins relative to the Original setup. The N- and O-glycans originating from the four chosen source glycoproteins (i.e. human IgG, HNE, and bovine fetuin and RNase B) were largely non-redundant and spanned all major N-glycan types typically encountered in mammalian glycomics profiling ranging from ultra-short chitobiose-type trisaccharides to branched poly-anionic complex sialoglycans. Organic modifiers enhance FOS ionisation (Setup 1) Direct infusion of equimolar FOS dissolved in different concentrations of MeOH, ACN or IPA demonstrated that an increasing organic solvent content of the PCMF (up to a specific concentration) correlates, as expected, with higher signal intensities, Figure S2. The highest absolute FOS response factors were obtained with 80% IPA when comparing all solvent conditions. In contrast, 70% ACN and 100% MeOH produced the highest FOS signal intensities within their respective series of organic solvent experiments albeit at a slightly lower absolute ion intensity relative to IPA. Noteworthy, reduced and almost abolished signals were observed when infusing FOS in very high organic solvent i.e. 80-100% ACN and 90-100% IPA. This observation may be explained by FOS precipitation in

such aqueous-poor solvent conditions as demonstrated in similar experiments using acetone-rich solvents36. Importantly, we noticed that above a specific level of organic modifier (above ~60% ACN and above ~30% IPA) signals corresponding to interfering (non-FOS) ions were also enhanced. Such contaminants, which may be of solvent or FOS sample origin, were observed at lower levels for the MeOH-rich experiments. Setup 1 was also used to test the effect of various flow rates on the FOS signal response. Flow rates spanning 1-5 µL/min were chosen, a range typically used in capillary-flow PGC-LC-MS/MS setups. The flow rate only slightly impacted the FOS signal intensities (data not shown). Hence, constant column (3 µL/min) and PCMF solvent supply (4 µL/min) flow rates were used in the subsequent Setup 2. The combined flow rate of 7 µL/min was compatible with the vendor recommendation of the used ESI sprayer (5-10 µL/min). The observations from Setup 1 are in accordance with the published literature. Kruve recently found that a high organic content enhanced the signal intensities in negative ion mode for a panel of 17 small acidic compounds37. The same authors also reported a poor repeatability and low sensitivity for pure ACN studies, especially when analysing hydrophilic analytes. They rationalised that this observation was due to the poor hydrogen bond donor properties of ACN rather than a poor organic solubility of the analytes37. Several years before, Huffmann et al. reported a better ionisation efficiency of mainly small acidic molecules and of the trisaccharide raffinose when 100% MeOH was used as the carrier solvent and reported that MeOH provided higher ion response factors than acetone and ACN38. Collectively, the infusion-type experiments demonstrated that the supplement of ACN, MeOH or IPA as a PCMF provides high FOS ion intensities whilst reducing interfering (non-glycan) ions. Thus, in the subsequent PGC-LC-MS/MS experiments (Setup 2) we tested the benefits of supplying 70-100% ACN, MeOH or IPA as a PCMF via a simple T junction (40-57% ESI concentration) on the glycomics analysis. PCMF boosts the glycomics analysis (Setup 2) To determine the impact of supplying organic modifiers on the glycomics performance, N- and O-glycans from four well characterised glycoproteins were analysed on a capillary-flow PGC-LC-MS/MS system with (Setup 2) and without (Original setup) the implementation of a PCMF.



Page 6 of 12

Figure 2. PCMF enhances the performance of PGC-LC-MS/MS-based N-glycomics. The absolute and relative signal intensities of selected N-glycans in a complex mixture were compared between a conventional capillary-flow PGC-LCMS/MS platform (Original setup) and PCMF-based PGC-LC-MS/MS (Setup 2) using 40%, 50% and 57% (v/v) ACN, MeOH or IPA. (A): Heat map depicting the mean of the absolute signal intensities (log10 scale) of a panel of 45 N-glycans monitored across all solvent conditions. See Table S1-S2 in the SI for details of the observed structures and their relative abundances. The N-glycans selected for further quantitative evaluation (Panel B) are boxed. Absolute signal intensities (B) and the relative abundance (C) of the selected subset of N-glycans over the different solvent conditions. Retention times (in blue) of the individual N-glycans are from a representative experiment using 57% ACN-based PCMF. *p < 0.05, **p < 0.01. Data points plot the mean and error bars represent the standard deviation of a minimum of three technical replicates. Comparative S/N values of representative N-glycan structures can be found in Figure 4.

Of the large repertoire of N-glycans known to decorate the four studied glycoproteins, we firstly compared the detection and quantitation of a panel of glycans comprising 45 N-glycans across all experimental solvent conditions, Table S1-S2. A heat map depicting the absolute intensities of the panel of N-glycans demonstrated that PCMF, regardless of the supplied organic modifier and concentration, gives rise to significantly higher glycan intensities relative to the intensities arising from the Original setup, Figure 2A.

Detailed quantitative comparisons across the experimental solvent conditions were performed for a representative subset of 16 N-glycans covering small, medium and large glycan sizes spanning all the major Nglycan classes i.e. paucimannose, high mannose, hybrid, and asialo-/sialo-complex N-glycans, Figure 2BC. The PCMF increased the absolute signal intensities for all monitored glycans relative to the Original setup as measured by the AUC. In particular, ACN and IPA dramatically enhanced all glycan signals (all p < 0.05).


Page 7 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. All glycans, but particularly weakly retained Nglycans, benefit dramatically from PCMF. Dots, representing unique glycan structures, plot their signal gains (in fold change measured by the absolute signal intensity) as a function of their retention times of the various PCMF conditions relative to the Original setup. Trend lines are shown.

The highest signal intensities were almost uniformly obtained for the highest organic modifier concentration (57%, v/v). Relative to the Original setup, the average signal enhancements were 38x/35x/9x, 45x/38x/11x and 46x/41x/11x when using 40%, 50% and 57% ACN/IPA/MeOH, respectively, but varied substantially depending on the structure and/or retention time of the individual glycans. These observations also showed that MeOH-based PCMF consistently gave less improved glycan detection than ACN and IPA. For all PMCF solvents, the small-to-medium sized neutral glycans i.e. the paucimannosidic, hybrid or high mannose glycans eluting relatively early (15-25 min) in the PGC-LC gradient (see retention times, Figure 2C) at lower organic solvent content benefitted the most from the PCMF installation, Figure 3. In particular the signal response of the very early eluting N-glycans i.e. Man2-3GlcNAc2 and GlcNAc2Fuc1 eluting at 17-18 min was greatly improved (50-100x) compared to the Original setup when using ACN and IPA PCMF. The larger bi- and tri-antennary N-glycans carrying 24 sialic acid residues produced signals typically 20-40x above the levels achieved with the Original setup. The preferential signal gain of the weakly retained glycans eluting from the PGC-LC column at high aqueous levels by the use of PCMF indicate a strong link between organic solvent levels and the ion response factors, a relationship previously reported by Pabst et al17. The direct influence of the LC elution time on the relative glycan abundance was confirmed by interrogating the relative abundance of isobaric glycans (with identical masses and chemical compositions) eluting at

different points in the PGC-LC gradient. The ACNbased PCMF differentially impacted the relative abundance of three glycan isomers (Hex4HexNAc3dHex1 NeuAc1, [M-2H]2-, m/z 856.31) eluting at 21.1 min, 27.7 min and 34.8 min, respectively. The less retained isomer showed a limited reduction in the relative abundance (-7%) whereas the two well retained isomers were reduced by 42% and 48%, respectively, compared to their relative levels in the Original setup, Table 1. The trends observed for the ACN-based PCMF were recapitulated in the IPA and MeOH experiments, which confirms that the sensitivity gain provided by the PCMF is dependent on the relative PGC-LC elution time of the glycan. The glycan mass, net charge and three-dimensional structure(s) may also contribute, albeit likely in a less pronounced manner, to the differential signal gain achieved with PCMF-based PGCLC-MS/MS8. Intuitively, the differential signal gains of glycans provided by the PCMF alter the relative glycoprofile compared to the Original setup, an effect particularly apparent for the abundant Man2GlcNAc2Fuc1 eluting at 21.4 min. This glycan was found to comprise ~30% of the entire N-glycome profile when measured by the Original setup whereas its relative abundance increased to ~45% using MeOH-, ACN- and IPA-based PCMF, Table S2. The quantitative glycan profiling accuracy of the PCMF setup was thus investigated for two of the studied glycoproteins i.e. RNase B and HNE relative to other quantitative methods including 2-AB HPLC and MS glycopeptide profiling of the same two glycproteins33, 39. The comparisons demonstrated that PCMF PGC-LC-MS/MS (and the Original setup) in general generates glycan distributions that are in close agreement with profiles produced with other quantitative methods (Figure S3-4 and Table S3-S6). Table 1. Relative abundance (in %) and change in abundance (in % change) of N-glycan isomers in PCMF PGCLC-MS/MS relative to the Original setup. Glycan isomer Origi- PCMF-based PGCdistribution [elution time] nal LC-MS/MS, Rel. % setup, (% change) Rel 57% 57% 57% % ACN MeOH IPA Hex4HexNAc3dHex1NeuAc1 3.2 3.0 3.3 3.0 (isomer 1) [21.1 min] (-7.0) (+0.9) (-8.5) Hex4HexNAc3dHex1NeuAc1 1.4 0.8 0.7 0.7 (isomer 2) [27.7 min] (-41.7) (-51.0) (-52.2) Hex4HexNAc3dHex1NeuAc1 0.09 0.05 0.03 0.03 (isomer 3) [34.8 min] (-47.6) (-66.1) (-67.6) Hex: Hexose; HexNAc: N-acetylhexosamine; dHex: deoxyhexose, NeuAc: N-acetylneuraminic acid.



Page 8 of 12

Collectively, these observations demonstrate that PCMF PGC-LC-MS/MS produces glycan profiles that are quantitatively similar to other analytical techniques thereby validating the use of this method in glycoprotein characterisation. Comparative glycomics aiming to identify qualitative and quantitative glycan differences between two or more biological samples40 is another relevant application for the presented method. Noteworthy, PCMF not only improved the glycan signal intensities, but also decreased the quantitative variations between runs and increased the spray stability. For the former, PCMF utilising 57% (v/v) organic modifier showed the most reproducible glycan signal intensities of the tested conditions; Standard deviations well below 0.5% compared to 0.5-3% levels for triplicate runs were observed for the Original setup. The PCMF setup consistently outperformed the Original setup in terms of spray stability throughout the LC gradient regardless of the organic modifier, Figure S5. IPA-based PCMF showed the highest spray stability. The impact of the PCMF on the PGC-LC peak shape, peak width and glycan isomer separation capacity was tested for the individual solvent conditions, Figure 4 and Figure S6. In general, the peak shapes and peak capacity were comparable amongst the tested solvents and were similar to the peak characteristics of the Original setup. Symmetrical and sharp (~15-20 s FWHM) LC peaks were achieved using all PCMF solvents. For all PCMF experiments, the glycan retention times were, as expected, slightly increased due to the slightly lower column flow rate used in Setup 2. Precipitation of the N-glycans was not observed to be an issue for any of the PCMF experiments (data not shown). The signal-to-noise ratios (S/N) of the observed glycans were also monitored, see Figure 4. As observed for the absolute intensities, the S/N values were significantly higher (1-2 orders of magnitude) for all monitored glycans when using PCMF compared to the Original setup. Similar gains in signal were also achieved for labile acidic glycans as measured in separate experiments by monitoring phosphorylated N-glycans from human myeloperoxidase with and without the 57%ACN based PCMF condition. Similar to the neutral and sialylated glycans, the signal gains were enhanced ~2540x and no detrimental effects of the PCMF on these species such as in-source fragmentation or any notable peak tailing were observed, Figure S7. This preferential gain of analyte signal relative to the intrinsic ion “noise” without fragmenting even labile glycans builds further support for the notion that glycomics benefits substantially from PCMF.

Figure 4. PCMF preserves the PGC-LC-MS/MS separation features. The PGC-LC separation of four types of Nglycans i.e. paucimannose (A), high mannose (B), hybrid (C) and complex sialylated (D) type N-glycans was evaluated with (Setup 2) and without (Original setup) PCMF based on monoisotopic EICs of representative LC-MS/MS runs. The glycan retention times were slightly increased in the PCMF experiments due to the lower effective column flow rate relative to the Original setup.*Non-glycan related peak. Retention times (RT) and S/N values are provided for all glycans. Zooms of selected peaks are included (broken inserts) for a better assessment of the peak characteristics with and without PCMF, see also Figure S6 for more detail.

The charge state distributions were not found to be systematically impacted by the PCMF relative to the charge state distributions observed for the Original setup as assessed for N-glycans detected in two charge states, Figure S8. Further, the fragment ion patterns were near-identical for all conditions, but the individual fragment ions were, in general, of significantly higher absolute intensity and had higher S/N for MS/MS spectra arising from Setup 2 (data not shown). Thus, PCMF enables more confident PGC-LCMS/MS-based glycome identification, in particular of low abundance and early eluting glycan species. PCMF was also shown to significantly enhance the performance of the PGC-LC-MS/MS-based O-glycomics analysis. The 57% concentration of all three solvents was the only PCMF condition tested in these proof-of-principle experiments. For the three small Oglycans reported from bovine fetuin, ~30-200x signal gains were observed for the PCMF setup relative to the Original setup, Figure 5A, Table S7-S8.


Page 9 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(e.g. glycopeptides) also eluting at low organic solvent contents from various LC stationary phases and on a range of instruments having similar source geometry as used in this study. We envisage that this potential will be tested and applications developed in the near future.

Figure 5. PCMF dramatically improves the O-glycomics performance. Absolute signal intensities (A) and the relative abundance (in % out of all structures) (B) of the three monitored O-glycans of bovine fetuin when analysed by PGC-LC-MS/MS with 57% ACN, MeOH and IPA-based PCMF (Setup 2) relative to the Original setup. **Corrected p values < 0.01. Data points and error bars are the mean and standard deviation of technical triplicates.

Similar to the N-glycome, ACN-based PCMF yielded the highest signal gains for O-glycans. The monosialylated O-glycan (Sialyl T) and the larger disialylated counterpart eluting very early from the PGCLC column at 10 min and 14 min, respectively, benefited the most from the PCMF in terms of absolute signal gain. Since the monitored O-glycans were relatively uniform in terms of size and elution time compared to the N-glycan counterparts the relative abundances of the individual O-glycans were not skewed by the PCMF, Figure 5B. Importantly, the technical variation between triplicate runs was reduced from up to 3% to below 0.5% standard deviation as a result of a more stable spray achieved using Setup 2 (Table S5). Given the demonstrated improvement of the glycomics analytical sensitivity, PGC-LC-MS/MS setups utilising PCMF enable high quality and reproducible N- and Oglycoprofiling from less biological starting material. This sensitivity enhancement effectively means that most biological glycomics experiments can be performed with a capillary-flow (>1 µL/min flow rate) PGC-LC-MS/MS rather than demanding a nano-flow setup (

Post-Column Make-Up Flow (PCMF) Enhances the Performance of

Recommend Documents