Supporting Information for Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Labeling Reagent that Facilitates Sensitive Fluorescence and ESI- MS Detection Matthew A. Lauber, 1 Ying-Qing Yu, 1 Darryl W. Brousmiche, 1 Zhengmao Hua, 1 Stephan M. Koza, 1 Ellen Guthrie, 2 Paula Magnelli, 2 Christopher H. Taron, 2 Kenneth J. Fountain 1 1 Waters Corporation, 34 Maple Street, Milford, Massachusetts 01757-3696 2 New England Biolabs, 240 County Road, Ipswich, MA 01938-2723 Author for correspondence: Matthew A. Lauber E-mail: Matthew_Lauber@Waters.com Telephone: (508)-482-3017 S-1
Supplemental Experimental and Results Reagents LC/MS grade solvents (water and acetonitrile), formic acid and trifluoracetic acid were purchased from Pierce (Rockford, IL). Ammonium acetate (73594), ammonium formate (70221), anthranilamide (2-AB, 2-aminobenzamide), bovine fetuin (F3004), dimethylformamide (DMF), dimethylsulfoxide (DMSO), glu-fibrinopeptide b (F-3261), guanidine hydrochloride, hexafluoroisopropanol (HFIP), HEPES (2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid) (H3375), human IgG (I4506), monosodium phosphate, n-propylamine, procainamide hydrochloride and TCEP (tris(2-carboxyethyl)phosphine) (C4706) were obtained from Sigma-Aldrich (St. Louis, MO). PNGase F (Glycerol Free) Recombinant (P0709, provided with Glycoprotein Denaturing Buffer and NP-40) were acquired from New England Biolabs (Ipswich, MA). Instant AB labeling reagent (IAB) and 2-AB labeled N-glycans from pooled human IgG were obtained from Prozyme (Hayward, CA). 2-AB labeled triacetyl chitotriose was purchased from Ludger (Oxfordshire, UK). GlycoWorks Rapid PNGase F, GlycoWorks Rapid PNGase F buffer, RapiFluor-MS labeling reagent (RFMS), and RapiGest SF surfactant (RG surfactant) were obtained from Waters (Milford, MA). Cetuximab (Erbitux, ImClone Systems, Bridgewater, NJ) was obtained from Besse Medical (West Chester, OH). Immunoglobulin degrading enzyme from S. pyogenes (IdeS) was obtained from Genovis (Lund, Sweden). Carboxypeptidase B was purchased from Worthington (Lakewood, NJ). Comparison of Labeled Glycans and Measurement of Response Factors (RFMS, IAB, and 2-AB) To compare the response factors of RFMS and IAB labeled glycans, N-glycans from anti-citrinin murine IgG1 were prepared according to the conditions outlined in the previous experimental, except that crude reaction mixtures were directly analyzed by HILIC-FLR-MS in order to avoid potential biases from SPE clean-up procedures. Response factors were determined as ratios of the FA2 N-glycan chromatographic peak area to the mass of glycoprotein from which the glycan originated. Fluorescence chromatograms were used to measure optical detection response factors, and base peak intensity (BPI) chromatograms were used to measure MS response factors. Response factors for this comparison were reported in units of peak area per µg of glycoprotein, and analyses were performed in duplicate. To compare the response factors of 2-AB labeled versus RFMS labeled glycans, equivalent quantities of labeled N-glycans from pooled human IgG were analyzed by HILIC-FLR-MS. A sample of 2-AB labeled N-glycans from pooled human IgG was obtained from Prozyme (Hayward, CA). RFMS labeled N-glycans from pooled human IgG were prepared according to the conditions outlined in the previous experimental. Column loads were calibrated using quantitative standards of 2-AB labeled S-2
triacetyl chitotriose and RFMS derivatized propylamine (obtained in high purity as confirmed by HPLC and 1 H NMR) (See Figures S15 and S16). Response factors were determined as ratios of the FA2 chromatographic peak area to the mole quantity of glycan. Response factors for this comparison were reported in units of peak area per pmole of glycan. Analyses were performed in duplicate. Metrics for the relative performance of labeling reagents were subsequently calculated based on the measured response factors. Ratios of response factors were calculated and reported as percentages in order to display the percent of RFMS signal that is obtained when an alternative labeling reagent is employed. The relative performance of procainamide was calculated using the conclusions from a published comparison of N-glycans, wherein it was found that procainamide provided comparable fluorescence and up to 50 fold greater ESI-MS sensitivity when compared to 2-AB. 1 Comparison of Labeled Glycans and Measurement of Response Factors (2-AB and Procainamide) N-linked glycans from pooled human IgG (ProZyme, Hayward, CA) were fluorescently labeled with either 2-aminobenzamide (2-AB) or procainamide (4-amino-N-(2-diethylaminoethyl) benzamide) reagents. Fluorescent labeling was performed in solutions of 100 µl of glacial acetic acid:dmso (3:7, v/v) mixture with 11 mg procainamide or 5 mg 2-AB, followed by addition of 6 mg of sodium cyanoborohydrate. Labeling reagent solution was added to 2 µg of human IgG N-linked glycans and the resulting mixture was heated for 4 hours at 65 C. All samples were reconstituted in acetonitrile/water (1:1) for injection. Labeled N-glycans were analyzed via HILIC separations combined with fluorescence and mass spectrometric detection using a UHPLC chromatograph (ACQUITY UPLC, Waters, Milford, MA. A 2.1 x 150 mm column packed with 1.7 µm amide-bonded organosilica particles (ACQUITY UPLC Glycan BEH Amide 130Å, Waters, Milford, MA) was employed along with an aqueous mobile phase comprised of 50 mm ammonium formate (ph 4.4) and another of ACN. Samples were separated at 60 C according to the chromatographic method shown in Table S7. Labeled N-glycans were detected using a fluorescence detector (1 Hz scan rate, Gain=1, ACQUITY UPLC FLR, Waters, Milford, MA). Eluting glycans were also detected by positive ion mode electrospray ionization mass spectrometry with a QTof mass spectrometer (Xevo QTof, Waters, Milford, MA) operating with a capillary voltage of 3.2 kv, source temperature of 100 C, desolvation temperature of 350 C, and sample cone voltage of 30 V. Based on this experiment, procainamide labeled N-glycans were found to exhibit approximately 15 times higher MS response factors than 2-AB labeled N-glycans (Figure S6). S-3
HILIC-UV-ESI-Analysis of IdeS Digested Cetuximab Cetuximab was subjected to middle-up glycan characterization 2-3 using HILIC separations combined with UV and ESI-MS detection. Prior to digestion with IdeS, 4 cetuximab was treated with carboxypeptidase B to complete the partial removal of the lysine-c-terminal residues that is typical of the antibody. 5 Formulated cetuximab was mixed with carboxypeptidase B (223 u/mg) at a ratio of 100:1 (w/w), diluted into 20 mm phosphate (ph 7.1), and incubated at a concentration of 1.8 mg/ml for 2 hours at 37 C. The carboxypeptidase B treated cetuximab was then added to 100 units of IdeS and incubated for 30 minutes at 37 C. The resulting IdeS digest was denatured and reduced by the addition of 1M TCEP and solid GuHCl. The final buffer composition for the denaturation/reduction step was approximately 6 M GuHCl, 80 mm TCEP, and 10 mm phosphate (ph 7.1). IdeS-digested cetuximab (0.9 mg/ml) was incubated in this buffer at 37 C for 1 hour. Solutions of reduced, IdeS digested cetuximab (carboxypeptidase b treated) were injected in 1.5 µl volumes onto a 2.1 x150 column packed with 1.7µm wide-pore amide bonded organosilica (300Å) stationary phase. Separations were performed at 60 C using the mobile phases and gradient described in Table S8. Eluting species were detected serially via UV absorbance (214 and 280 nm, 2 Hz) followed by online ESI-MS with a QToF mass spectrometer (Waters Xevo G2 QTof, Milford, MA) operating with a capillary voltage of 3.0 kv, source temperature of 150 C, desolvation temperature of 350 C, and sample cone voltage of 45 V. Mass spectra were acquired at a rate of 2 Hz with a resolution of approximately 20,000 over a range of 500 4000 m/z. Evaluation of Deglycosylation by Gel Electrophoresis The effectiveness of the rapid deglycosylation process was evaluated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). SDS-PAGE is an effective technique for separating proteins based on their size in solution and can often be used to separate the glycosylated and de-glycosylated forms of proteins. 6-7 A diverse set of glycoproteins were deglycosylated according to the 2-step rapid deglycosylation procedure and analyzed by SDS-PAGE along with negative controls containing no PNGase F and positive controls, in which the glycoproteins were subjected to conventional multiple step deglycosylation with SDS based denaturation and PNGase F incubation for 30 minutes at 37 C. The following glycoproteins were assayed for deglycosylation using gel electrophoresis: 30 µg bovine fibrinogen (F8630, Sigma, St. Louis, MO), 15 µg bovine RNaseB (New England Biolabs, Ipswich, MA), 30 µg bovine fetuin (New England Biolabs, Ipswich, MA), 13.5 µg anti-citrinin murine IgG1 (Intact mab Mass Check Standard, Waters, Milford, MA), 10 µg human lactoferrin (L0520, Sigma, St. Louis, S-4
MO), 30 µg chicken ovalbumin (A2512, Sigma, St. Louis, MO), 174 µg bovine holo-transferrin (T1408, Sigma, St. Louis, MO), 30 µg human α-acid glycoprotein (G9885, Sigma, St. Louis, MO), 30 µg bovine asialofetuin (A1908, Sigma, St. Louis, MO) and 180 µg human chorionic gonadotropin (C6322, Sigma, St. Louis, MO). Glycoprotein substrates were deglycosylated under two different conditions: a) with PNGase F (P0709, New England BioLabs, Ipswich, MA) after conventional SDS denaturation; or b) with GlycoWorks Rapid PNGase F after denaturation with GlycoWorks buffer and RapiGest SF surfactant (Waters, Milford, MA). For conventional denaturation, the glycoprotein was incubated at 95 C for 2 min in the presence of 0.5% SDS and 40 mm DTT (Glycoprotein Denaturing Buffer, New England BioLabs, Ipswich, MA), in a total volume of 16 µl. After cooling, NP-40 (to 1% v/v), sodium phosphate ph 7.5 (to 50 mm), and PNGase F (PNGase F (Glycerol Free) Recombinant, 500 units) were added to the solution, increasing the total volume of the mixture to 20 µl. The reaction was incubated at 37 C for 30 minutes. Negative controls were prepared under the same conditions, without the addition of enzyme. The rapid deglycosylation procedure was performed with GlycoWorks Rapid PNGase F, GlycoWorks Rapid PNGase F Buffer, and RapiGest SF surfactant. Glycoproteins were incubated with GlycoWorks Rapid Buffer and 1% (w/v) RapiGest SF at 80 C for 2 minutes, in a total volume of 20 µl. After cooling, 1 µl of GlycoWorks Rapid PNGase F was added and the reaction was incubated at 50 C for 5 minutes. Following enzymatic deglycosylation, all samples were heated to 70 C for 20 minutes to inactivate the enzyme. Each sample was diluted with 30 µl of water and 25 µl 3x Blue Loading Buffer (New England Biolabs, Ipswich, MA) containing 0.14 mm DTT. The samples were heated at 95 C for 2 minutes and 7 µl of each sample was loaded on a Novex 10-20% Tris-Glycine Mini Protein Gel (Life Technologies, Grand Island, NY). The gel was run at 150 V for 1 hour, and then stained with coomassie blue (SimplyBlue SafeStain, Life Technologies, Grand Island, NY). Figure S10A shows the results of this study, where it can be seen that for each of the tested proteins there is a significant decrease in protein apparent molecular weight after they are subjected to the rapid deglycosylation procedure. Moreover, the apparent molecular weight decreases are visually comparable to those observed for proteins deglycosylated by the control method. These results demonstrate that the fast deglycosylation approach facilitated by a unique formulation of PNGase F (GlycoWorks Rapid PNGase F, Waters, Milford, MA) and surfactant (RapiGest SF, Waters, Milford, MA) produces deglycosylation comparable to a conventional approach but in only a fraction of the time required and with conditions that are compatible with rapid labeling and downstream mass spectrometric analyses. S-5
It is worth noting that some glycoproteins must be subjected to reducing conditions in order for complete deglycosylation to be achieved. For instance, samples of holo-transferrin and chorionic gonadotropin were found to present higher apparent molecular bands when subjected to rapid deglycosylation according to the above procedure versus the conventional SDS procedure (Figure S10B). Accordingly, rapid deglycosylation was repeated with the addition of 4 mm TCEP reducing agent. TCEP, being a non-nucleophilic reducing agent, was chosen over thiol reducing agents to ensure compatibility with rapid tagging labeling reactions. TCEP hydrochloride is a strong conjugate acid and was therefore titrated upon dissolution with sodium hydroxide to a neutral ph stock solution before use. In this modified rapid deglycosylation procedure, glycoproteins were incubated with GlycoWorks Rapid Buffer, 1% (w/v) RapiGest SF, and 4 mm TCEP at 80 C for 2 minutes, in a total volume of 20 µl. After cooling, 1 µl of GlycoWorks Rapid PNGase F was added and the reaction was incubated at 50 C for 5 minutes. Results for this reducing, rapid deglycosylation condition are shown in Figure S10B. Evaluation of HILIC SPE Absolute and Relative Recovery The procedure for extracting RFMS labeled glycans after derivatization was evaluated using a test mixture containing N-glycans released and labeled from a 1:1 mixture (by weight) of pooled human IgG and bovine fetuin. The test mixture was prepared according to the conditions outlined in the previous experimental. An aliquot of crude sample was saved prior to the first SPE clean-up procedure. This sample and the one obtained after SPE clean-up (1x SPE) were analyzed by HILIC FLR (Figure 5A versus 5B). The latter sample was once again processed by HILIC SPE (2x SPE) and likewise analyzed by HILIC-FLR. To assess the relative recovery of this process, relative abundances for a selection of labeled N-glycans were then measured for the samples once and twice subjected to HILIC SPE (1x SPE versus 2x SPE) (Table S5, Figure 5). Additionally, measured peak areas for the labeled FA2 glycan were used to estimate the absolute SPE recovery (Table S5). Based on triplicate analysis of two different SPE sorbent batches, the average recovery for the HILIC SPE clean-up was estimated to be 74%. Measurement of the Sample Preparation Yield The percent yield of the entire N-glycan sample preparation was assessed in order to measure the collective efficiency of combining fast deglycosylation, rapid labeling, and HILIC SPE extraction of RFMS labeled glycans. RFMS labeled N-glycans from anti-citrinin murine IgG1 (15 µg) were prepared and analyzed by HILIC-FLR-MS as outlined in previous experimental sections (Figure S12). Using RFMS derivatized propylamine, the RFMS labeled FA2 glycan was quantified (Figures S12A and S12B). A percent yield for the sample preparation was subsequently calculated assuming the theoretical S-6
yield calculation outlined in Figure S12C. Based on duplicate analysis, it was determined that the percent yield of the entire rapid N-glycan sample preparation was approximately 73%. Charge States and CID Fragmentation Pathways of RFMS Labeled Glycans Because the RFMS label has high proton affinity, derivatized glycans preferentially adopt high charge states during positive ion mode electrospray ionization. The predominant charge state for a small neutral glycan, such as an FA2 glycan, is [M+2H] 2+, although it increases to [M+3H] 3+ for larger molecular weight glycans, such as an A3G3S3 glycan (Figure S13). Collision induced dissociation (CID) of these doubly and triply protonated RFMS labeled glycans showed similar fragmentation pathways as those previously reported for 2AB-labeled glycans. For instance, high intensity b and y-ions from glycosidic bond cleavages were predominately observed along with low-intensity, cross-ring fragments. 8 However, unlike 2-AB labeled glycans, RFMS labeled glycans readily produced high signal-to-noise MS/MS spectra from relatively low analyte quantities. An example MS/MS spectrum for an FA2 glycan derivatized with RFMS is illustrated in Figure S14. Synthesis, Purification and Characterization of RFMS Derivatized Propylamine RFMS was reacted with propylamine to produce a material to be used as a quantitative standard. RFMS reagent (75 mg) was dissolved in methylene chloride (3 ml) and added drop-wise to a stirring solution of n-propylamine (70 mg) in methylene chloride (7 ml) over 10 minutes at room temperature. The reaction was stirred for an additional 30 minutes. Solvent was removed by rotoevaporation. The recovered brown powder was washed three times in methylene chloride (10 ml) and then dried for 48 hours under vacuum. The dried material was then dissolved in 1% formic acid (v/v) and chromatographically purified using a 4.6 x 150 mm column packed with 2.5 µm C18 bonded, charged-doped organosilica stationary phase (130Å) (XBridge CSH C18, Waters, Milford, MA) and a UHPLC chromatograph (ACQUITY UPLC H-Class Bio, Waters, Milford, MA). Crude RFMS derivatized propylamine (1 mg) was separated at 30 C using the mobile phases and gradient described in Table S9. Fractions collected from 20 gradients were pooled and dried by centrifugal evaporation to yield a yellow waxy solid. 1 H and 13 C NMR confirmed the identity of the material as the TFA salt of RFMS derivatized propylamine (Figure S15). The purified RFMS derivatized propylamine (4µg) was also assayed for impurities using the reversed phase liquid chromatography (RPLC) separation by which it was purified. Eluting species were detected by UV absorbance (210 nm, 2 Hz) and fluorescence (Excitation 265 nm / Emission 425 nm, 2 S-7
Hz) followed by online ESI-MS with a QTof mass spectrometer (Waters Xevo G2 QTof, Milford, MA) operating with a capillary voltage of 3.0 kv, source temperature of 150 C, desolvation temperature of 600 C, and sample cone voltage of 40 V. Mass spectra were acquired at a rate of 1 Hz with a resolution of approximately 20,000 over a range of 100 2000 m/z. Based on these analyses, the purified RFMS derivatized propylamine was determined to be approximately 99% pure (Figure S16). 1 H NMR Analysis of IAB and RFMS Labeling Reagents IAB labeling reagent (InstantAB, Prozyme, Hayward, CA) was characterized using a Bruker 300M NMR spectrometer (Billerica, MA). The 1 H NMR spectrum observed for IAB labeling reagent was found to be consistent with an NHS-carbamate derivative of 4-aminobenzamide. 1H NMR: (300MHz, CD3CN) δ 8.78(s, 1H), 7.83(d, 2H, J=8.8Hz), 7.53(d, 2H, J=8.7Hz), 6.71(s, 1H), 5.94(s, 1H), 2.80(s, 4H. Use of this reagent, 4-aminobenzamidyl-N-hydroxysuccinimidyl carbamate (4-ABSC), for labeling of N-glycans has been previously disclosed. 9 RFMS labeling reagent (RapiFluor-MS, Waters, Milford, MA) was likewise characterized using a Bruker 300M NMR spectrometer (Billerica, MA). The 1 H NMR spectrum observed for RFMS labeling reagent confirms our proposed structure and indicates RFMS is purified as a 1:1 complex with n- hydroxysuccinimide (NHS). 1H NMR: (300MHz, DMSO-d6) δ 8.79(t, 1H, J=5.7Hz), 8.47(d, 1H, J=8.6Hz), 8.16(d, 1H, J=1.7Hz), 8.10(d, 1H, J=8.5Hz), 8.09(d, 1H, J=9.1Hz), 7.86(dd, 1H, J=9.2, 1.9Hz), 3.40(q, 2H, J=6.5Hz), 2.85(s, 4H), 2.57(s, 4H), 2.50-2.63(m, 6H), 0.98(t, 6H, J=7.0Hz) S-8
Table S1. Chromatographic Gradient for HILIC-Fluorescence-ESI-MS Analysis of Labeled N-Glycans with a 2.1 x 50 mm Column Mobile Phase A: 50 mm ammonium formate, ph 4.4 Mobile Phase B: ACN Time Flow Rate %A %B Curve (ml/min) 0.0 0.4 25 75 6 11.7 0.4 46 54 6 12.2 0.2 100 0 6 13.2 0.2 100 0 6 14.4 0.2 25 75 6 15.9 0.4 25 75 6 18.3 0.4 25 75 6 S-9
Table S2. Chromatographic Gradient for HILIC-Fluorescence-ESI-MS Analysis of Labeled N-Glycans with a 2.1 x 150 mm Column Mobile Phase A: 50 mm ammonium formate, ph 4.4 Mobile Phase B: ACN Time Flow Rate %A %B Curve (ml/min) 0.0 0.4 25 75 6 35.0 0.4 46 54 6 36.5 0.2 100 0 6 39.5 0.2 100 0 6 43.1 0.2 25 75 6 47.6 0.4 25 75 6 55.0 0.4 25 75 6 S-10
Table S3. Fluorescence Wavelengths for HILIC-Fluorescence-ESI-MS Analysis of Labeled N-Glycans. Fluorescence Label Excitation Emission Wavelength Wavelength RFMS 265 nm 425 nm IAB 278 nm 344 nm 2-AB 330 nm 420 nm S-11
Table S4. Chromatographic Gradient for HILIC-Fluorescence-ESI-MS Analysis of Analysis of Intact IgG with Coupled 2.1 x 150 mm Columns (300 mm effective column length). Mobile Phase A: 0.1% TFA, 0.3% HFIP in water Mobile Phase B: 0.1% TFA, 0.3% HFIP in ACN Time Flow Rate %A %B Curve (ml/min) 0.0 0.2 20 80 6 10.0 0.2 50 50 6 11.0 0.2 100 0 6 14.0 0.2 100 0 6 15.0 0.2 20 80 6 25.0 0.2 20 80 6 S-12
Table S5. Recovery of RFMS Labeled Glycans through HILIC SPE Extraction. Results for two different batches of HILIC SPE sorbent are provided. Averages of three replicates are shown. Control SPE Batch 1 % Recovery Glycan Species Peak Area Rel. Abun. Std Dev Peak Area Rel. Abun. Std Dev SPE Batch 1 SPE Batch 2 FA2 46257 7.7 0.1 33895 7.3 0.2 73.3 75.3 FA2G2S1 21726 3.6 0.1 15996 3.5 0.0 A3G3S3 147026 24.5 0.2 118174 25.6 0.3 A3S1G3S3 34809 5.8 0.3 28387 6.1 0.3 S-13
Table S6. Identification of RFMS Labeled Glycans from Anti-Citrinin Murine IgG1 by HILIC-FLR-MS Analysis. Peak No. Glycan Name Experimental Mass Theoretical Mass Mass Error Approximate (M+H) + (M+H) + (ppm) Relative% 1 A1 1425.5837 1425.5891-3.8 0.09 2 FA1 1571.6446 1571.647-1.5 0.17 3 A2 1628.6676 1628.6684-0.5 1.21 4 FA2 1774.7256 1774.7263-0.4 46.02 5 A2G1 1790.7194 1790.7213-1.1 0.66 FA1G1 1733.6976 1733.6998-1.3 <0.04 6 A2G1(iso) 1790.7194 1790.7213-1.1 0.30 7 FA2G1 1936.7772 1936.7792-1.0 18.59 8 FA2G1(iso) 1936.7772 1936.7792-1.0 21.51 9 FA2G1B 2139.8558 2139.8585-1.3 0.04 10 FA2G1B (iso) 2139.8558 2139.8585-1.3 0.21 11 FA2G2 2098.828 2098.832-1.9 8.03 12 FA2G1Ga1 2098.828 2098.832-1.9 0.67 13 FA2G2Ga1 2260.877 2260.8848-3.4 1.08 14 FA2G2Sg1 2405.9174 2405.9223-2.0 0.18 15 FA2G2Sg1(iso) 2405.9174 2405.9223-2.0 0.06 16 FA2G2Ga2 2422.9322 2422.9376-2.2 0.87 17 FA2G2GaSg1 2567.9694 2567.9751-2.2 0.20 18 FA2G2Ga1Sg1(iso) 2567.9694 2567.9751-2.2 0.11 F: Fucose A: Anternary G: Galactose B: Bisecting Ga: alpha-galactose Sg: N-glycolylneuraminic acid iso: structural isomer S-14
Table S7. Chromatographic Method for HILIC-Fluorescence-ESI-MS Analysis of Labeled N-Glycans with a 2.1 x 150 mm Column (Procainamide versus 2-AB Labeled N-Glycans) Fluorescence Wavelengths (2-AB) Ex 330 nm Em 420 nm Fluorescence Wavelengths (Procainamide) Ex 308 nm Em 359 nm Mobile Phase A 50 mm ammonium formate, ph 4.4 Mobile Phase B ACN Gradient 72% to 55% B in 45 min Flow Rate 0.4 ml/min S-15
Table S8. Chromatographic Gradient for HILIC of Reduced, IdeS-Digested Cetuximab using a 2.1 x 150 mm Column Packed with 1.7µm Wide-Pore Amide Bonded Organosilica (300Å) Stationary Phase. Mobile Phase A: 0.1% TFA in water Mobile Phase B: 0.1% TFA in ACN Time Flow Rate %A %B Curve (ml/min) 0.0 0.2 20 80 6 1.0 0.2 30 70 6 21.0 0.2 37 63 6 22.0 0.2 80 20 6 24.0 0.2 80 20 6 25.0 0.2 20 80 6 35.0 0.2 20 80 6 S-16
Table S9. Chromatographic Gradient for HPLC Purification of RFMS Derivatized Propylamine Using a 4.6 x 150 mm Column Packed with 2.5 µm C18 Bonded, Charge-doped Organosilica Stationary Phase (130Å). Mobile Phase A: 0.1% TFA in water Mobile Phase B: 0.1% TFA in ACN Time Flow Rate %A %B Curve (ml/min) 0.0 1.0 80 20 6 10.0 1.0 70 30 6 12.0 1.0 20 80 6 13.0 1.0 80 20 6 25.0 1.0 80 20 6 S-17
A FA2 5 min 10 min 20 min Incubation time prior to rapid labeling 40 min FA2 B Estimated t 1/2 1 st Order 2 nd Order 50 C 100 min 120 min 37 C 200 min 240 min RT 400 min 480 min Figure S1. Estimating the Half-life of N-Glycosylamine Hydrolysis through a Time-Course on Deglycosylation Incubation. (A) Fluorescence traces for RFMS labeled FA2 from anti-citrinin murine IgG1 observed after implementing varying incubation times for deglycosylation (50 C incubations). Fluorescence chromatograms are shown for labeled glycans from anti-citrinin murine IgG1 (from 0.4 µg glycoprotein and a10 µl injection from ACN/DM F diluent) as obtained with a 2.1 x 150 mm column packed with 1.7 µm amide-bonded organosilica (130Å) stationary phase. (B) Approximation of the N- glycosylamine half-life assuming 1 st or 2 nd order reaction kinetics. S-18
A Fluorescence 10x zoom B 200000 180000 Intial 0 Hours 3 Days 72 Hours Initial 10 30 min 10x zoom Fluorescence Peak Areas 160000 140000 120000 100000 80000 60000 40000 20000 After 3 days 10 C 0 A2 FA2 FA2G1 FA2G1(iso) FA2G2 FA2G2Sg1 FA2G2Ga2 10 30 min Figure S2. Evaluating the Stability of the RFMS Glycan Derivatives. Fluorescence chromatograms are shown for labeled glycans from anti-citrinin murine IgG1 (from 0.4 µg glycoprotein and a10 µl injection from ACN/DMF diluent) as obtained with a 2.1 x 150 mm column packed with 1.7 µm amide-bonded organosilica (130Å) stationary phase. (A) Zoomed (10x) views of chromatograms obtained before and after the sample was stored for 3 days at 10 C. (B) Observed fluorescence peak areas for major species in the profile. No significant change was observed in the glycan fluorescence profile during this testing, indicating that the RFMS glycan derivatives are stable. S-19
A B 3.3E+6 0.0E+0 4.0E+6 0.0E+0 3 4 5 6 7 8 9 3.3E+6 2-AB Labeled N-Glycans from Pooled Human IgG 0.0E+0 4.0E+6 RFMS Labeled N-Glycans from Pooled Human IgG FA2 FLR (2.17 pmol) FA2 (2.61 pmol) FA2 FA2 100x zoom MS (BPI) FLR MS (BPI) C Response Factors (Peak Area per pmol of Labeled FA2 Glycan / 1000) 150 100 50 0 105.8 126.9 RapiFluor-MS Labeled FLR MS (BPI) 7.4 0.8 2-AB Labeled 0.0E+0 3 4 5 6 7 8 9 Figure S3. HILIC-FLR-MS of (A) RFMS and (B) 2-AB Labeled N-Glycans from Pooled Human IgG. Fluorescence (FLR) chromatograms are shown in orange and base peak intensity (BPI) MS chromatograms are shown in blue. Labeled glycans (~14 pmol total glycan, 1 µl aqueous injection) were separated using a 2.1 x 50 mm column packed with 1.7 µm amide-bonded organosilica (130Å) stationary phase. The quantities of FA2 glycan were calibrated via two-point external calibrations with quantitative standards (RFMS derivatized propylamine and 2-AB labeled triacetylchitotriose). (C) Response factors for RFMS and 2-AB labeled glycans (measured as the FA2 peak area per picomole of FA2 determined by the external calibration). Fluorescence (FLR) and MS (base peak intensity) response factors are shown in orange and blue, respectively. Analyses were performed in duplicate. S-20
A 3.3E+6 FLR 0.0E+0 4.0E+6 MS (BPI) 0.0E+0 FA2 (2.17 pmol) (Area = 229.6 x 10 3 ) 3 4 5 6 7 8 9 10 FA2 (Area = 275.5 x 10 3 ) RFMS Labeled Human IgG N-Glycans 3 4 5 6 7 8 9 10 B Calibrant Fluorescence (265/425 nm) (pmol) Rep 1 Rep 2 Avg 1.67 177458 177340 177399 6.68 727983 759405 743694 y = 113033x - 11366 Peak Area (Fluorescence - Ex 265 nm / Em 425 nm) 1E+6 8E+5 6E+5 4E+5 2E+5 y = 113033x - 11366 R² = 1 0E+0 0 5 10 Quantity (pmol) higg FA2 232105 227079 229592 2.17 RFMS Label (pmol) Figure S4. Quantitative Analysis of RFMS Labeled N-Glycans from Pooled Human IgG. (A) HILIC- FLR-MS of RFMS labeled N-glycans from pooled human IgG using a 2.1 x 50 mm column packed with 1.7µm amide bonded organosilica (130Å) stationary phase. (B) Quantitation of the RFMS labeled FA2 glycan using an external calibration with RFMS derivatized propylamine. Analyses were performed in duplicate. S-21
A 3.0E+5 FLR FA2 (2.61 pmol) (Area = 19.3 x 10 3 ) B 2-AB Labeled Human IgG N-Glycans Calibrant Fluorescence (330/420 nm) (pmol) Rep 1 Rep 2 Avg 1.04 7046 7347 7197 5.72 43629 43040 43335 y = 7721.8x-834.17 Quantity (pmol) higg FA2 19665 18955 19310 2.61 0.0E+0 3.0E+4 MS (BPI) 0.0E+0 3 4 5 6 7 8 9 10 FA2 (Area = 2.0 x 10 3 ) 3 4 5 6 7 8 9 10 Peak Area (Fluorescence - Ex 330 nm / Em 420 nm) 7E+4 6E+4 5E+4 4E+4 3E+4 2E+4 1E+4 y = 7721.8x - 834.17 R² = 1 0E+0 0 5 10 2-AB Label (pmol) Figure S5. Quantitative Analysis of 2AB Labeled N-Glycans from Pooled Human IgG. (A) HILIC- FLR-MS of RFMS labeled N-glycans from pooled human IgG using a 2.1 x 50 mm column packed with 1.7µm amide bonded organosilica (130Å) stationary phase. (B) Quantitation of the 2-AB labeled FA2 glycan using an external calibration with 2-AB labeled triacetyl chitotriose. Analyses were performed in duplicate. S-22
Figure S6. HILIC-FLR-MS of 2-AB and Procainamide Labeled N-Glycans from Pooled Human IgG. (A) Fluorescence (FLR) chromatograms and (B) base peak intensity (BPI) MS chromatograms for 2-AB labeled N-glycans. (C) Fluorescence (FLR) chromatograms and (B) base peak intensity (BPI) MS chromatograms for procainamide labeled N-glycans. Labeled glycans (200 ng of 2-AB labeled glycans and 40 ng of procainamide labeled glycans) were separated using a 2.1 x 150 mm column packed with 1.7 µm amide-bonded organosilica (130Å) stationary phase. S-23
ph logd RFMS Structure logd Procainamide Structure 1.50-1.73-4.04 5.00-1.58-1.81 6.50-0.69-0.89 7.40 0.16-0.04 RFMS Test Structure O O N N H N N H N H Procainamide Test Structure Figure S7. Approximation of Log D Values for RFMS and Procainamide Labels. Marvin was used for drawing, displaying and characterizing chemical structures, substructures and reactions, Marvin 14.7.7.0, 2014, ChemAxon (http://www.chemaxon.com). S-24
Glycoprotein N-Glycans Deglycosylation Released N-Glycans + Protein Labeling Labeled N-Glycans + Reagents Reaction Byproducts Label Clean-up Labeled Glycans Label HILIC FLR-MS Analysis Fluorescence MS Figure S8. Workflow for the Rapid Preparation of N-glycans Using RFMS. S-25
A 214 A 0.06 0.05 0.04 0.03 0.02 Fc/2 +G0F Fc/2 +Man5 Fc/2 Glycosylated Fc/2 +G1F Fc/2 +Hex6HexNAc3DHex1 Fc/2 +G2F Fd pe + (G2F+NGNA) Fd pe + (G2FGal2) Fd pe + (G2FGal1+NGNA) Fd Glycosylated N-term pe Fd pe + (Hex9HexNAc5DHex1) Fd pe Fc/2 + (G2FGal1) +G0F-GN/G0F 0.01 8 9 10 11 12 13 14 15 16 17 18 Time (min) G0F C Man5 G1F G2F+NGNA G2FGal1 G1F G2F G2FGal2 G2FGal1+NGNA Hex9HexNAc5DHex1 Time 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.50 12.00 B D Identification Man5 (M5) G0F (FA2) G1F (FA2G1) G2F (FA2G2) G2FGal (FA2G2Ga1) G2F+NGNA (FA2G2Sg1) G2FGal2 (FA2G2Ga2) G2FGal+NGNA (FA2G2Ga1Sg1) Identification Fc/2-K + Man5 (M5) Fc/2-K + G0F (FA2) Fc/2-K + G1F (FA2G1) Fc/2-K + G2F (FA2G2) Fd' pe + (G2FGal) (FA2G2Ga1) Fd' pe + (G2F+NGNA) (FA2G2Sg1) Fd' pe + (G2FGal2) (FA2G2Ga2) Fd' pe + (G2FGal+NGNA) (FA2G2Ga1Sg1) RFMS Labeled MWmono,theo Glycan Composition (Da) C63H99N7O37 1545.6080 C73H115N9O41 1773.7190 C79H125N9O46 1935.7719 C85H135N9O51 2097.8247 C91H145N9O56 2259.8775 C96H152N10O60 2404.9150 C97H155N9O61 2421.9303 C102H162N10O65 2566.9678 Hex9HexNAc5DHex1 C117H188N10O76 2949.1154 MWavg,theo MWavg,obs Mass Error (Da) (Da) (Da) 25008.1 25008.8 0.7 25236.3 25237.4 1.1 25398.5 25399.8 1.3 25560.6 25562.0 1.4 27385.5 27386.8 1.3 27530.6 27531.8 1.2 27547.6 27548.2 0.6 27692.7 27693.1 0.4 Fd' pe + (Hex9HexNAc5DHex1) 28075.1 28075.3 0.2 m/zobs (charge state) 773.8141 (2+) 887.8694 (2+) 968.8990 (2+) 1049.9141 (2+) 1130.9503 (2+) 1203.4648 (2+) 1211.9733 (2+) 1284.4969 (2+) 1475.5785 (2+) MWmono,obs (Da) Mass Error (ppm) 1545.6136 3.6 1773.7242 2.9 1935.7834 5.9 2097.8136-5.3 2259.8860 3.8 2404.915 0.0 2421.932 0.7 2566.9792 4.4 2949.1424 9.2 Figure S9. HILIC Profiling of Cetuximab Glycosylation. (A) HILIC-UV of reduced, IdeS-digested cetuximab using a 2.1 x 150 mm column packed with 1.7µm widepore amide bonded organosilica (300Å) stationary phase. Species corresponding to Fc/2 and Fd subunits are labeled in grey and red, respectively. (B) Subunit glycan assignments based on deconvoluted mass spectra. (C) HILIC-FLR of RFMS labeled N-glycans from cetuximab using a 2.1 x 50 mm column packed with 1.7µm amide bonded organosilica (130Å) stationary phase. (D) Mass spectral data supporting the assignments of the RFMS labeled N-glycans. S-26
A - + R - + R - + R - + R - + R - + R - + R - + R B - + R R+TCEP - + holo-transferrin R R+TCEP chorionic gonadotropin (-) Neg Control / Conventional / 2 Steps without PNGase F: SDS + DTT, 95 C, 2 min / NP-40 + Reaction Buffer, 37 C, 30 min (+) Pos Control / Conventional / 2 Steps with PNGase F: SDS + DTT, 95 C, 2 min / NP-40 + Reaction Buffer + PNGase F, 37 C, 30 min (R) 2 Step Rapid Deglycosylation / Reaction Buffer + RapiGest 80 C, 2 min / + PNGase F, 50 C, 5 min (R+TCEP) 2 Step Rapid Deglycosylation / Reaction Buffer + RapiGest + TCEP 80 C, 2 min / + PNGase F, 50 C, 5 min PNGase F inactivated after the deglycosylation step via heat denaturation at 70 C for 20 min. Figure S10. Gel Electrophoresis Assay for Deglycosylation of Glycoproteins. (A) A negative control (-) shows the migration distance and apparent molecular weight of the native glycoproteins, and a positive control (+) shows the migration distance and decreased apparent molecular weight of deglycosylated proteins as obtained by conventional two step deglycosylation using SDS denaturation and a subsequent 30 minute incubation with PNGase F at 37 C. Results demonstrating the complete deglycosylation of these glycoproteins with a rapid procedure involving a two step approach with RapiGest-assisted heat denaturation and a subsequent 5 minute incubation with Rapid PNGase F at 50 C are also shown (R). (B) Two proteins (holo-transferrin and chorionic gonadotropin) were found to require reducing conditions for complete release of glycans. Results observed for these proteins were indicative of complete deglycosylation when they were subjected to 4 mm TCEP along with the rapid deglycosylation conditions (R+TCEP). S-27
6E+8 RFMS FA2 Fluorescence Peak Area 5E+8 4E+8 3E+8 2E+8 1E+8 Specified elution volume 0E+0 0 30 60 90 120 150 180 SPE Elution Volume (µl) Figure S11. Fluorescence peak area as a function of SPE elution volume. The specified elution volume from the protocols employed in this work is 90 µl (three 30µL elution volumes). This elution volume provides an optimized concentration of sample and facilitates direct analysis of the SPE eluate. S-28
2E+6 0E+0 A FA2 Rep #1 1.62 pmol Rep #2 1.66 pmol 100% Theoretical Yield N-Glycan Yield ~73% 3 4 5 6 7 8 9 10 11 12 13 C 1.5x10 7 pg IgG B Peak Area (Fluorescence - Ex 265 nm / Em 425 nm) 1E+6 8E+5 6E+5 4E+5 2E+5 y = 113033x - 11366 R² = 1 0E+0 0 5 10 RFMS Label (pmol) 1 pmol 2 pmol glycan 0.45 pmol FA2 10 µl injection X X X X 150,000 pg 1 pmol IgG 1 pmol total 400 µl = 2.25 pmol glycan pool sample prepared Figure S12. Yield Calculation for the Rapid Preparation of Labeled N-Glycans Using RFMS Derivatization. (A) HILIC-FLR of RFMS labeled N-glycans prepared from anti-citrinin murine IgG1 using a 2.1 x 50 mm column packed with 1.7µm amide bonded organosilica (130Å) stationary phase. (B) Quantitation of the RFMS labeled FA2 glycan using an external calibration with RFMS derivatized propylamine. Analyses were performed in duplicate. (C) Calculation for approximating the 100% theoretical yield for the FA2 glycan released from anti-citrinin murine IgG1. S-29
100 100 RFMS Labeled FA2 [M+2H] 2+ m/z 800 1000 1200 1400 1600 1800 2000 100 % % % 0 m/z 886 887 888 889 890 891 892 0 100 RFMS Labeled A3G3S3 [M+3H] 3+ % 0 m/z 1062 1064 1066 1068 0 m/z 800 1000 1200 1400 1600 1800 2000 Figure S13. Charge States of RFMS Labeled Glycans Observed by ESI(+)-MS. S-30
Figure S14. MS/MS Fragmentation Spectrum of RFMS Labeled FA2. Glycosidic bond cleavage was observed to be the favorable fragmentation pathway. Peaks with symbolic structures were from the b and y ion series. MS/MS spectra were charge deconvoluted and deisotoped using MaxEnt 3. S-31
A 8.191 8.162 7.741 7.727 7.712 7.696 7.560 7.453 7.446 7.422 7.415 Associated or Free NHS 4.684 3.711 3.690 3.669 3.327 3.306 3.285 3.230 3.205 3.180 3.158 2.998 2.975 2.952 1 8 2 H2O 1.418 1.394 1.370 1.346 1.216 1.192 1.168 0.804 0.780 0.755 17 H 3 C 15 9 NH 8 O NH 1 6 4 7 3 5 N 3 5 6 4 7 O O NH 2 10 X N O H N O 12 13 16 10 13 9 12 15 16 17 10 9 8 7 6 5 4 3 2 1 0 1.000 2.137 1.052 1.024 4.170 0.018 1.994 2.210 4.142 2.095-0.001 0.068 2.386 6.483 3.505 B 164.81 156.88 143.85 140.56 139.72 139.02 130.36 126.39 118.52 114.33 112.82 50.28 47.68 41.49 34.63 22.50 10.54 8.02 X and Y TFA residue? Associated or Free NHS c a b 17 H 3 C 15 9 5 e d f NH 8 O 6 b NH f 1 4 7 g 4 3 6 X O O c N O H 7 d O 10 N g N e a NH 5 3 2 12 13 16 13 12 10 9 x 15 17 16 X Y 180 160 140 120 100 80 60 40 20 ppm Figure S15. NMR Spectra for RFMS Derivatized Propylamine. (A) 1 H NMR Spectrum. (B) 13 C NMR Spectrum. S-32
A B C Figure S16. Impurity Assays for Crude and HPLC purified RFMS (RapiFluor-MS) Derivatized Propylamine. (A) RPLC-UV analysis of crude RFMS derivatized propylamine. (B) RPLC-UV of purified RFMS derivatized propylamine. (C) RPLC-fluorescence of purified RFMS derivatized propylamine. S-33
References 1. Klapoetke, S.; Zhang, J.; Becht, S.; Gu, X.; Ding, X., The evaluation of a novel approach for the profiling and identification of N-linked glycan with a procainamide tag by HPLC with fluorescent and mass spectrometric detection. J Pharm Biomed Anal 2010, 53 (3), 315-24. 2. Janin-Bussat, M. C.; Tonini, L.; Huillet, C.; Colas, O.; Klinguer-Hamour, C.; Corvaia, N.; Beck, A., Cetuximab Fab and Fc N-glycan fast characterization using IdeS digestion and liquid chromatography coupled to electrospray ionization mass spectrometry. Methods Mol Biol 2013, 988, 93-113. 3. An, Y.; Zhang, Y.; Mueller, H. M.; Shameem, M.; Chen, X., A new tool for monoclonal antibody analysis: application of IdeS proteolysis in IgG domain-specific characterization. MAbs 2014, 6 (4), 879-93. 4. von Pawel-Rammingen, U.; Johansson, B. P.; Bjorck, L., IdeS, a novel streptococcal cysteine proteinase with unique specificity for immunoglobulin G. EMBO J 2002, 21 (7), 1607-15. 5. Ayoub, D.; Jabs, W.; Resemann, A.; Evers, W.; Evans, C.; Main, L.; Baessmann, C.; Wagner- Rousset, E.; Suckau, D.; Beck, A., Correct primary structure assessment and extensive glycoprofiling of cetuximab by a combination of intact, middle-up, middle-down and bottom-up ESI and MALDI mass spectrometry techniques. MAbs 2013, 5 (5), 699-710. 6. Kung, L. A.; Tao, S. C.; Qian, J.; Smith, M. G.; Snyder, M.; Zhu, H., Global analysis of the glycoproteome in Saccharomyces cerevisiae reveals new roles for protein glycosylation in eukaryotes. Mol Syst Biol 2009, 5, 308. 7. Schwalbe, R. A.; Wang, Z.; Bianchi, L.; Brown, A. M., Novel sites of N-glycosylation in ROMK1 reveal the putative pore-forming segment H5 as extracellular. J Biol Chem 1996, 271 (39), 24201-6. 8. Bosques, C. J.; Collins, B. E.; Meador, J. W., 3rd; Sarvaiya, H.; Murphy, J. L.; Dellorusso, G.; Bulik, D. A.; Hsu, I. H.; Washburn, N.; Sipsey, S. F.; Myette, J. R.; Raman, R.; Shriver, Z.; Sasisekharan, R.; Venkataraman, G., Chinese hamster ovary cells can produce galactose-alpha- 1,3-galactose antigens on proteins. Nat Biotechnol 2010, 28 (11), 1153-6. 9. Banks, D. D.; Gadgil, H. S.; Pipes, G. D.; Bondarenko, P. V.; Hobbs, V.; Scavezze, J. L.; Kim, J.; Jiang, X. R.; Mukku, V.; Dillon, T. M., Removal of cysteinylation from an unpaired sulfhydryl in the variable region of a recombinant monoclonal IgG1 antibody improves homogeneity, stability, and biological activity. J Pharm Sci 2008, 97 (2), 775-90. S-34