Identification of Low Abundant Isomeric N-Glycan Structures in

Subscriber access provided by - Access paid by the | UCSB Libraries

Article

Identification of low abundant isomeric N-glycan structures in biological therapeutics by LC/MS Jia Zhao, Siyang Li, Chen Li, Shiaw-Lin Wu, Wei Xu, Yuetian Chen, Mohammed Shameem, Douglas Dennis Richardson, and Huijuan Li Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b00636 • Publication Date (Web): 11 Jun 2016 Downloaded from http://pubs.acs.org on June 12, 2016

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 free 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 accessible to all readers and 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.

Analytical Chemistry 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 36

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

Identification of low abundant isomeric N-glycan structures in biological therapeutics by LC/MS Jia Zhao1, Siyang Li2, Chen Li2, Shiaw-Lin Wu2, Wei Xu1, Yuetian Chen1, Mohammed Shameem1, Douglas Richardson3, Huijuan Li1 1.

Protein Mass Spectrometry, Sterile Product and Analytical Development, Bioprocess Development, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA 2

BioAnalytix, Inc., Cambridge, MA, USA.

3

Bioprocess Technology and Expression, Bioprocess Development, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA

Abstract An effective LC-MS based method for on-line characterization of low abundant structural isomers of N-linked glycans in biological therapeutics was developed. N-linked glycans of a recombinant monoclonal antibody were released by PNGase F and labeled with 2aminobenzamide (2-AB) fluorescent tag. The labeled glycans were analyzed by on-line ultraperformance liquid chromatography-hydrophilic interaction liquid chromatography (UPLC-HILIC) coupled with mass spectrometry (MS). The glycan structure was characterized by MSn fragmentation in negative ion mode followed by identification of the signature D ions. The assignment included monosaccharide sequence and linkage information. The developed method successfully characterized structural isomers of A1G1F (assigned as terminal sialic acid attached in the 1,6 branch at 2,3 position), and A1G1F’ (assigned as terminal sialic acid attached in the 1,3 branch at 2,3 position). Moreover, using the same approach, previously unknown low abundant species were identified unambiguously. One such structural isomer at low level,terminal GlcNAc of G1F+GlcNAc, was identified to be linked at the 1,6 branch. Additionally, another low level structural isomer, previously assigned as Man8 glycan, was found to be Man7+Glc glycan as its 1,3 branch containing three mannoses and one terminal glucose. The identification was further confirmed by a purified α-1,2-endomannosidase enzyme to generate the cleavage of α-1,3 linked terminal di-saccharides (Man+glucose). Using this approach, different lots or different CHO produced mAbs was thoroughly examined and found that the newly identified “Man8” (Man7+Glc) was also present in different batches and in some commercially available therapeutic mAbs. 1 ACS Paragon Plus Environment


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

Page 2 of 36

1. Introduction N-linked glycosylation is one of the most common post-translational modifications (PTMs) in therapeutic proteins such as monoclonal antibodies (mAbs). In comparison to other types of biomolecules, such as proteins and peptides, glycans possess many additional structural features such as variable sequence, linkage and branching of the constituent monosaccharide. These attributes complicate the structural elucidation of the glycoproteins. N-linked glycans have significant effects on the structure, stability and biological functions of mAbs.1-4 It is expected by regulatory agents that N-linked glycans of mAbs should be monitored through the entire development process and thoroughly characterized before biological license application (BLA)5, 6. Methods applied for glycan characterization in general require the release of glycans from proteins and fluorescence labeling of the released glycans. The separation methods play critical role in the quantitation of released glycans. High-performance anion-exchange chromatography (HPAEC)7, 8, reversed-phase liquid chromatography9, 10, normal phase (NP)/hydrophilic interaction chromatography (HILIC)11-14 and capillary electrophoresis (CE)15-17 are widely used in glycan quantitation. NP/HILIC utilized a polar stationary phase column and an increasing aqueous gradient to elute glycans. This separation method permits a good balance of sensitivity, resolution, reproducibility 18-20and is also compatible with mass spectrometry (MS). Further improvements in peak capacity and sample throughput can be achieved by replacing a 3 µm column with a sub-2µm glycan column in HILIC analysis with ultra-performance liquid chromatography (UPLC) 12. UPLC-HILIC analysis of 2-aminobenzamide (2-AB) labeled N-linked glycans has become a method of choice for monitoring the glycosylation profile of therapeutic proteins. While most peak identification in HILIC analysis of glycans performed in the pharmaceutical industry is based on known standards. The lack of certain standards and coelution of some species make the characterization and quantitation quite challenging. Identification of low abundant glycan peaks in HILIC analysis has been often executed by collecting peaks and off-line characterization using mass spectrometry in combination with enzyme treatment or chemical derivatization. However, the process is time consuming and labor intensive. The application of liquid chromatography coupled to mass spectrometry (LC/MS) to in-depth analysis of glycosylation of therapeutic proteins has grown in recent years given its sensitivity and robustness.21 Glycans are identified by mass spectrometry from ion fragments generated in two ways: (a) glycosidic cleavages resulting from a bond rupture between two adjacent sugar residues; and (b) cross-ring cleavages in which any two bonds on the same sugar unit are broken. Cross-ring fragment ions are commonly observed in high-energy CID methods as demonstrated in the tandem TOF/TOF approach. 22-24 Glycosidic cleavage ions are mainly observed in lowenergy fragmentation methods and mainly provide sequence and limited branching information. Studies have shown that in contrast to the B- and Y-type fragments under positive ion mode, 2 ACS Paragon Plus Environment

Page 3 of 36

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


low or high energy CID fragmentation of glycans under negative ion mode generate cross-ring and C-type fragments and mainly follow a single pathway of fragmentation mechanism.25, 26 The C-type fragments can provide sequence information of the monosaccharide residues, while many of the cross-ring fragments under negative ion mode are diagnostic ions for certain structural features and could be used to distinguish structure and linkage isomers. One important feature in the negative mode MS2 fragmentation is the observation of D ions that result from two glycosidic cleavages. Although there are other more abundant mass fragments in the spectra, D-type ions are unique to specific isomers. D-type fragment ions generated by cleaving the core GlcNAc residues and the 3- antenna are unique to each isomeric form and their identities can be confirmed by one stage of collisional fragmentation. Detailed fragmentation mechanism and several diagnostic ions have been discussed in literatures25-31. In this paper, we applied a method of on-line LC-MS characterization of 2-AB labeled glycans in negative mode. The design of UPLC-HILIC with fluorescence and on-line mass spectrometry detection enabled the analysis to provide orthogonal resolutions for structural isomers of glycans. The added MS detection further expanded the method capability to resolve and identify unknown glycan species. Compared to the conventional approach for structure elucidation using enzyme treatment or chemical derivatization (e.g. permethylation), this method is faster and more straightforward. Three low abundant glycan isomers released from a mAb are discussed here: G1F+GlcNAc, A1G1F and Man8. Linkage assignments using on-line LC-MS in negative ion mode with MSn approaches are described with illustrated figures. The isomeric structures of each species were elucidated by identifying the signature “D” ions, which were further confirmed by the identities generated in subsequent MS3 spectra. A previously assigned Man8 glycan based on mass and consistent UPLC retention time with standard Man8 was found to be three mannoses and one glucose at the 1,3 branch. It is further confirmed by α-1,2endomannosidase enzyme, that the terminal residue at α-1,3 branching in “Man8” is a glucose.

3 ACS Paragon Plus Environment


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

Page 4 of 36

2. Experimental section Native N-glycans and 2-AB labeled N-glycans from a mAb (IgG4) were provided by Merck (Kenilworth, NJ). IgG from human serum was purchased from Sigma-Aldrich (St. Louis, MO). N Glycanase (PNGase F) was purchased from Prozyme (Hayward, CA) and New England Biolab, (Lpswich, MA). HPLC grade water, acetonitrile, glacial acetic acid, formic acid and ammonium bicarbonate were obtained from Fisher Scientific (Pittsburgh, PA). SignalTM 2-AB labeling kit and GlycoCleanTM S Cartridges were purchased from Prozyme (Hayward, CA).

Release of glycans by PNGaseF For Merck mAb, 1 mg of antibody was diluted in 200 µL of Tris buffer (from Prozyme). 10 µL of glycanase (Prozyme) was added to the mixture and incubated at 37oC overnight (16 to 20 hours). For human IgG or Ribonuclease B, 1mg protein was dissolved in 100 µL buffer (100 mM Ammonium bicarbonate, pH 8). A 4 µL aliquot of PNGase F (New England Biolabs) was added to the protein buffer and incubated overnight at 37 °C. For Human IgG, after an aliquot of PNGase F was added and incubated for 4 hours, a 2nd dose of 4 µL PNGase F was added and then incubated overnight at 37 °C. The enzymatic digestion was quenched by adding formic acid (to the final concentration 1 %). The released N-glycans were purified using 10 kDa molecular weight cutoff filters (PALL Life Science, Ann Arbor, MI) and lyophilized via vacuum centrifugation. The use of 1 mg as the starting protein amount has been validated for precise and reproducible recovery and measurement of observable glycan peaks at 0.1% level using the column id and flow rate described in the following section. Glycan label with 2-AB and separation of 2-AB labeled glycans Released glycans were subjected to 2-AB labeling with SignalTM 2-AB-Labeling kit. 2-AB glycans were separated on a Waters BEH glycan hydrophilic interaction chromatography column (100 mm x 2.1 mm, 1.7 µm) connected to a UPLC instrument (Dionex Ultimate 3000 or Waters Acquity UPLC) online coupled with fluorescence detection (Dionex RF 2000 or Waters Acquity UPLC FLR). The excitation and emission wavelength parameters were set to 330 nm/420 nm for 2-AB glycans with fluorescence gain fixed at 16. The mobile phases used were as follows: (A) 100 mM ammonium formate (pH 4.5) and (B) 100% acetonitrile. The system was operated at either 200 µL/min (used for on-line MS characterization) or 400 µL/min (used for LC-fluorescence profiling) flow rate with 60 °C column temperature. For 200 µL/min gradient, the elution gradient was maintained at 200 µL/min using 72 %B for 4min and then decrease from 72 %B to 62 %B for 30min. The flow rate was decreased to 100 µL/min using 100 %A for 3 min to wash the column and re-equilibrated at an initial condition with 200 µL/min flow rate for an additional 10 min prior to subsequent injection. For 400 µL/min, the elution gradient was maintained at 400 µL/min using 72 %B for 2 min and then decrease from 72 %B to 62 %B for 30min. The flow rate was decreased to 250 µL/min using 100 %A for 3 min to wash the column and re-equilibrated at initial condition with 400 µL/min flow rate for an additional 10 min prior 4 ACS Paragon Plus Environment

Page 5 of 36

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 subsequent injection. Samples were prepared in 75:25 acetonitrile/water and maintained at 5 °C prior to injection. A volume of 30-50 µL (corresponding to glycans released from 30-50 µg mAb protein) was injected on the column for LC/MS analysis. On-line LC-MS of HILIC separated 2-AB labeled Glycans The outlet of the chromatographic separation was coupled to an Orbitrap Elite MS or Orbitrap Velos Pro MS (ThermoFisher scientific, San Jose, CA) equipped with an HESI-II electrospray ion source (ThermoFisher scientific) with a flow rate of 200 µL/min. The heated ESI source was set to 320 °C, sheath gas flow rate of 25 arb, Aux Gas Flow Rate of 10 arb, spray voltage of 3.2 kV and capillary temp of 300 °C. Full MS were acquired in FTMS under negative ion mode with resolution of 60,000 with scan range of m/z 380-1650. MS2 were acquired in ion trap with normalized collision energy of 35. Source fragmentation energy was set to 30.0 V. MS/MS data representing glycan structures were conducted by manual inspection. Endomannosidase treatment The endomannosidase enzyme was produced in Pichia Pastoris32. 2AB labeled glycans released from 100 µg Merck mAb molecule were incubated with 10 µL endomannosidase (0.53 mg/mL) in sodium acetate buffer at pH 5 overnight at 37 oC. A glycan sample without enzyme treatment was incubated side by side as a control. The treated and untreated samples were analyzed by LC-MS.

3. Results and discussion Glycans released from monoclonal antibodies are a mixture of many glycoforms. MS is a powerful tool for glycan characterization. MS analysis without separation gives a reasonably good glycoform profile but is not able to distinguish structural isomers which have same mass. To fully utilize the orthogonal high resolving power of both HILIC UPLC and MS analysis, we have developed a method to connect high sensitivity and high resolution MS instrument with UPLCHILIC separation. Figure 1 shows the zoomed fluorescence chromatogram of HILIC UPLC-MS analysis of 2-AB labeled N-linked glycans released from a mAb. We choose Waters BEH Glycan HILIC UPLC column with 1.7 µm particle size as it provides faster separation and better resolution compared to the conventional HPLC column. As shown in Figure 1, structural isomers, such as G1F and G1F’, can be separated by HILIC UPLC and identified by MS. For species coeluted in HILIC UPLC, they can be easily identified and differentiated by their unique masses using the on-line MS detection. As for structure isomers, their linkages are also assigned using the on-line LC-MS (negative ion mode) with MSn approaches. The critical features in the MS2 fragmentation spectrum are D ions generated from two glycosidic cleavages33,34. Although there are other high abundant fragment ions in the fragmentation spectrum, they are not sufficiently unique for the isomer assignment. In contrast, the critical D ions, from the loss of the core GlcNAc residues and the 3- antenna portion are unique to each isomeric form for the 5 ACS Paragon Plus Environment


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

Page 6 of 36

specific isomer assignment. As shown in Fig.S1 in supporting information, three isomeric structures from standard Man8 glycan from Ribonuclease B were separated by HILIC UPLC. Identification of important characteristic masses, D, D-18 and D’ ions allow us to distinguish the Man8 branching difference in these three isoforms in a relatively simple procedure. In this paper, we applied this approach on characterization of three low abundant isomeric structures (A1G1F, G1F+GlcNAc, Man8) from a human therapeutic mAb. These three glycans were highlighted in red in Figure 1. Their abundances are all less than 1% of the total glycans, as determined by fluorescence signal. The structures of these low abundant species were shown in Figure 1.

Figure 1 Zoomed fluorescence chromatogram of HILIC UPLC-MS linked glycans released from a mAb. Peaks labeled in red are structural isomers discussed in this paper. Key: green circle: acetylglucosamine (GlcNAc); yellow circle: galactose; red triangle: acid

3.1

analysis of 2-AB labeled Nexamples of low abundant mannose; blue square: Nfucose; pink diamond: sialic

Linkage assignment of G1F+GlcNAc

G1+GlcNAc is a low abundant complex type glycan found in monoclonal antibodies, its abundance is less than 0.5% in the total glycans from the Merck mAb. Little is known about its structure and linkage. Two isomer structures were observed for 2AB labeled G1F+GlcNAc. Figure 2A shows the two isomers, a major and a minor form co-eluted. Figure 3A shows the low energy CID MS2 spectrum of the co-eluted isoforms of G1F+GlcNAc. The critical features in the 6 ACS Paragon Plus Environment

Page 7 of 36

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


MS2 fragmentation spectrum are D ions generated from the cleavage of core GlcNAc residues and the 3-antenna portion. As shown in the Figure 3A, D ion with m/z 729.28 corresponds to the major form with additional GlcNAc attached at the 1,6 branch. D ion with m/z 526.45 corresponds to the minor form with additional GlcNAc attached at the 1,3 branch. Some separation of these two D ions was observed: D ion of m/z 729.29 from the major isoform eluted slightly earlier in MS2 spectrum than the D ion of m/z 526.45. To further confirm the D ion structures, each of the D ions was isolated for further fragmentation (MS3) to confirm the indeed expected structures. Figure 3B shows the MS3 of D ion from major form with additional GlcNAc at 1,6 branch. Glycosidic cleavages including D’ ion (m/z 381.99) and other fragments (y2, c2) were detected. The minor form isomer with GlcNAc attached at the 1,3 branch was eluted slightly later, with its D ion structure further confirmed by MS3 with detection of glycosidic fragmentation (y2, c2) and cross-ring fragmentation (0,4A2 and 1,3A2) (Figure 3C). Additionally, other possible linkages of G1F+GlcNAc such as bisecting linkages were also investigated. In this regard, human IgG (known to contain bisecting GlcNAc) were analyzed using the same approach. The chromatographic profile of the bisecting G1F+GlcNAc isoforms was shown in Figure 2B. Two isoforms of G1F+GlcNAc from human IgG were observed and they eluted earlier than G1F+GlcNAc peak from the Merck mAb. These two isomers from human IgG correspond to bisecting G1F+GlcNAc structure with galactose attached on 1, 3 (minor form) or 1,6 branch (major form). MS2 fragmentation of the minor form by CID is shown in Figure 4A. D ion with loss of bisecting GlcNAc residue and H2O (D-221,) was observed.35, 36 This ion (D-221, m/z 508.04) was further fragmented in MS3 and confirmed the D ion structure with the observations of b1, b2, c2 and c2 fragment ions (Figure 4B).

Figure 2 The separation of structural isomers of G1F+GlcNAc. Top panel (a) is the selected ion chromatogram (SIC) of G1F+GlcNAc isoforms of Merck mAb. The bottom panel (b) is the SIC of the two bisecting G1F+GlcNAc isomers from human IgG.



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

Page 8 of 36

(A)

(B)

(C) Figure 3 (A) CID MS2 of G1F+GlcNAc species of Merck MAb (B) CID MS3 of D ion of major glycoform with m/z 729.28 (C) CID MS3 of D ion of minor glycoform with m/z 526.45


Page 9 of 36

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


(A)

(B) Figure 4 (A) CID MS2 of G1F+GlcNAc with bi-secting GlcNAc from Human IgG (B) CID MS3 of D221 ion (m/z 506.04) from G1F+GlcNAc with bi-secting GlcNAc.

3.2

Linkage assignments of A1G1F and A1G1F’

A1G1F is a complex type glycan, its abundance is

Identification of Low Abundant Isomeric N-Glycan Structures in

Recommend Documents