A-fucosylation detection in antibodies

Abstract

This invention describes a new analytical method to determine the quantity and distribution of fucose per Fc within an antibody preparation.

Claims

1. A method for detecting the presence or absence of fucose residues within a homogenously deglycosylated immunoglobulin G antibody, said method comprising: (a) providing an antibody preparation comprising a glycosylated immunoglobulin G antibody, (b) removing heterogeneous saccharide residues from the glycosylated immunoglobulin G antibody with the enzyme endoglycosidase S and the enzyme endoglycosidase H, thereby obtaining said homogenously deglycosylated immunoglobulin G antibody, (c) removing other heterogenous residues from said homogenously deglycosylated immunoglobulin G antibody with an enzyme, and (d) analyzing said homogenously deglycosylated immunoglobulin G antibody for the presence or absence of fucose residues, wherein the linkage information on the presence or absence of fucose residues within the homogenously deglycosylated immunoglobulin G antibody is maintained.

2. The method of claim 1, wherein said method further comprises purifying said homogenously deglycosylated immunoglobulin G antibody prior to analyzing said homogenously deglycosylated immunoglobulin G antibody for the presence or absence of fucose residues.

3. The method of claim 1, wherein said method further comprises determining the quantity of fucose residues among said homogenously deglycosylated immunoglobulin G antibody.

4. The method of claim 1 or claim 3, wherein said method further comprises determining the distribution pattern of fucose residues among said homogenously deglycosylated immunoglobulin G antibody.

5. The method of claim 1 or claim 3, wherein said method further comprises determining the distribution pattern of fucose residues per homogenously deglycosylated immunoglobulin G antibody.

6. The method of claim 1, wherein removing other heterogeneous residues from the homogenously deglycosylated immunoglobulin G antibody in step (c) is performed by one or more enzymes selected from the group consisting of plasmin and carboxypeptidase B.

7. The method of claim 1, wherein analyzing the homogenously deglycosylated immunoglobulin G antibody for the presence or absence of fucose residues in step (d) is performed by an analysis selected from the group consisting of liquid-chromatography-mass spectrometry analysis, capillary electrophoresis-sodium dodecyl sulfate molecular weight analysis, and electrospray ionization mass spectrometry analysis.

8. A method for detecting the presence or absence of fucose residues within a homogenously deglycosylated immunoglobulin G antibody, said method comprising: (a) providing an antibody preparation comprising a glycosylated immunoglobulin G antibody, (b) removing heterogeneous saccharide residues from the glycosylated immunoglobulin G antibody with the enzyme endoglycosidase S and the enzyme endoglycosidase H, thereby obtaining said homogenously deglycosylated immunoglobulin G antibody, (c) removing C-terminal lysine residues from said homogenously deglycosylated immunoglobulin G antibody with carboxypeptidase B, and (d) analyzing said homogenously deglycosylated immunoglobulin G antibody for the presence or absence of fucose residues by liquid-chromatography-mass spectrometry analysis, capillary electrophoresis-sodium dodecyl sulfate molecular weight analysis, and electrospray ionization mass spectrometry analysis, wherein the linkage information on the presence or absence of fucose residues within the homogenously deglycosylated immunoglobulin G antibody is maintained.

9. The method of claim 8, wherein said method further comprises purifying said homogenously deglycosylated immunoglobulin G antibody prior to analyzing said homogenously deglycosylated immunoglobulin G antibody for the presence or absence of fucose residues.

10. The method of claim 8 or claim 9, wherein said method further comprises determining cooperative fucosylation in said antibody preparation during fermentation.

Description

SHORT DESCRIPTION OF THE FIGURES

(1) FIG. 1. Schematic representation of a carbohydrate moiety attached to Asn-297 of human IgG1-Fc. The sugars in bold define the pentasaccharide core of N-linked glycan structures; the addition of the other sugar residues is variable. In grey is represented a bisecting GlcNAc residue.

(2) FIGS. 2A and 2B. Deglycosylation of intact Fc fragment of antibodies A (wildtype) and C (glycoengineered) monitored by CE-SDS. Electropherograms of non-reduced Fc fragments are shown before and after enzymatic treatment. (FIG. 2A) Fc fragment of antibody C without enzymatic treatment (dashed line) and deglycosylated with PNGase F (dotted line) or Endo S (solid line), (FIG. 2B) Fc fragment of antibody A without enzymatic treatment (dashed line), deglycosylated with PNGase F (dotted line) or deglycosylated with Endo S (solid line).

(3) FIGS. 3A, 3B, and 3C. Positive-ion MALDI-TOF mass spectra of the N-linked oligosaccharides released from Fc fragment of antibody C by consecutive treatment with Endo S and PNGase F or with Endo S and Endo H. (FIG. 3A) Spectrum of glycans released by treatment with Endo S. (FIG. 3B) Spectrum of Endo S-resistant carbohydrates released by subsequent treatment with PNGase F, resulting in an isolated signal at m/z=1663 (possibly corresponding to hybrid- or complex-type structures as schematically depicted). (FIG. 3C) Spectrum of glycans released by subsequent treatment with the hybrid-type structure specific enzyme Endo H (hybrid-type structures corresponding to m/z=1460 released by Endo H treatment are schematically depicted).

(4) FIGS. 4A and 4B. Deglycosylation of the Fc fragment of antibody C monitored by CE-SDS MW analysis (FIG. 4A) and positive-ion MALDI-TOF mass spectrometry (FIG. 4B). (FIG. 4A) Overlay of electropherogram of the non-reduced Fc fragment without glycosidase treatment (dashed line) and treated with a combination of Endo S and Endo H (solid line). (FIG. 4B) Mass spectra of the N-linked oligosaccharides released from the Fc fragment treated with Endo S and Endo H. Hybrid-type structures corresponding to m/z=1460 released by Endo H are schematically depicted.

(5) FIGS. 5A, 5B, and 5C. ESI-MS spectra of Fc fragments after treatment with Endo S and Endo H. (FIG. 5A) Fc fragments of antibody A, (FIG. 5B) Fc fragments of antibody B, (FIG. 5C) Fc fragments of antibody C. Peak 1: Fc-GlcNAc/GlcNAc, Peak 2: Fc-GlcNAc/GlcNAc+Fuc, Peak 3: Fc-GlcNAc+Fuc/GlcNAc+Fuc.

(6) FIGS. 6A and 6B. Deglycosylation of antibody C monitored by CE-SDS (FIG. 6A) and positive-ion MALDI-TOF mass spectrometry (FIG. 6B). (FIG. 6A) Electropherograms of non-reduced IgG are shown before and after enzymatic treatment: Antibody C without enzymatic treatment (dashed line) and deglycosylated with PNGase F (dotted line) or combined treatment with Endo S and Endo H (solid line). (FIG. 6B) Mass spectra of the N-linked oligosaccharides released from entire IgG treated with Endo S and Endo H.

(7) FIG. 7. N-linked oligosaccharide biosynthetic pathway leading to complex- or hybrid-type structures. M1: mannosidase I, 1,2-N-acetylglucosaminyltransferase I, G3: 1,4-N-acetylglucosaminyltransferase III, Gt: 1,4-galactosyltransferase.

(8) FIGS. 8A and 8B. ESI-MS spectra of entire IgGs after treatment with Endo S and Endo H. (FIG. 8A) antibody A, (FIG. 8B) antibody D. Peak 1: Fc-GlcNAc/GlcNAc, Peak 2: Fe-GlcNAc/GlcNAc+Fuc, Peak 3: Fc-GlcNAc+Fuc/GlcNAc+Fuc.

(9) FIGS. 9A and 9B. LC-MS spectra of entire IgGs after treatment with Endo S and Endo H. (FIG. 9A) antibody A, (FIG. 9B) antibody D. Peak 1: Fc-GlcNAc/GlcNAc, Peak 2: Fe-GlcNAc/GlcNAc+Fuc, Peak 3: Fc-GlcNAc+Fuc/GlcNAc+Fuc.

EXAMPLES

Example 1: Methods

(10) Generation of Fc from Human IgG

(11) Four different human IgGs with a different content of a-fucosylated glycans, determined according to Papac et al., 1998 (content in brackets), were used for analysis of the a-fucosylation distribution: wildtype antibody A (2.12%), glycoengineered antibody B (47.0%), glycoengineered antibody C (69.6%), and glycoengineered antibody D (85%).

(12) The proteins were incubated for 72 hours at 25 C. in 50 mM Tris pH 8.0, 150 mM NaCl with 0.42 U plasmin (Roche) per milligram. Cleaved Fc was separated from Fab-fragments using a Protein A affinity column (5 ml HiTrap Protein A HP column, GE Healthcare) equilibrated and washed (5 column volumes (CV)) with buffer A (50 mM Tris pH 8.0, 100 mM glycine, 150 mM NaCl). Fc was eluted by a pH-step using buffer B (50 mM Tris pH 3.0, 100 mM glycine, 150 mM NaCl). Fractions containing Fc were pooled and neutralized by adding 1:40 (v/v) 2 M Tris pH 8.0. Samples were concentrated to a volume of 2.5 ml using ultra concentrators (Vivaspin 15R 10000 MWCO HY, Sartorius) and subsequently applied to a PD-10 desalting column (GE Healthcare) equilibrated with 2 mM MOPS pH 7.4, 150 mM NaCl, 0.02% (w/v) NaN.sub.3. Purified protein was frozen in liquid nitrogen and stored at 80 C.

(13) Release of N-Linked Oligosaccharides from Human Fc

(14) Different enzymes were used for hydrolyzing the N-linked glycans of human IgG. The N-linked oligosaccharides were cleaved from 1 mg of Fc by incubation with 0.005 U recombinant PNGase F (QAbio, Vista Monte, USA). For release of carbohydrates from Fc using non-tagged Endo S (Genovis), samples were incubated with either a molar ratio of 1:20 of Endo S alone or in combination with 0.1 U/mg Endo H (QAbio). All reactions were incubated in 20 mM Tris pH 8.0 at 37 C. for 16 h.

(15) For analyzing carbohydrates spared by Endo S, Fc was purified after Endo S treatment by affinity chromatography using Protein A and subsequently digested with either PNGase F or Endo H, as described above.

(16) Release of N-Linked Oligosaccharides from Entire Human IgG

(17) The N-linked glycans of human IgG were released using different enzymes. The N-linked oligosaccharides were cleaved from 1 mg of IgG by incubation with 0.005 U of recombinant PNGase F (QAbio) in 20 mM Tris pH 8.0 at 37 C. for 16 h. For release of carbohydrates from IgG using non-tagged Endo S (Genovis), samples were applied to a NAP-5 desalting column (GE Healthcare) equilibrated with 20 mM Tris pH 8.0. Eluted sample was concentrated to a final concentration of 4 mg/ml using ultra concentrators (Amicon 5000 MWCO, Millipore) and incubated with a molar ratio of 1:7 of Endo S combined with 0.1 U/mg Endo H (QAbio) at 37 C. for 16 h.

(18) Carboxypeptidase B Treatment

(19) To remove heterogenicity caused by C-terminal lysine residues, after deglycosylation samples were further incubated with carboxypeptidase B (Roche; 1 mg/ml). Therefore 1 l carboxypeptidase B per 50 s human Fc or entire antibody was added to the Endoglycosidase reaction and incubated again for 1 h at 37 C.

(20) MALDI-TOF Mass Spectrometry Analysis of Released Oligosaccharides

(21) Neutral oligosaccharide profiles of the human Fc or entire antibody were analyzed by mass spectrometry (Autoflex, Bruker Daltonics GmbH) in positive ion mode (Papac et al., 1998).

(22) Purification of Deglycosylated Human Fc or Entire Antibody

(23) Fc or entire IgG was separated from enzymes and cleaved carbohydrates by Protein A affinity chromatography using Agilent HPLC 1100 series (Agilent Technologies). Samples were applied to Protein A matrix (Poros 20 A; Applied Biosystems) packed in a guard column 220 mm C-130B (Upchurch Scientific) equilibrated with buffer A (50 mM Tris, 100 mM glycine, 150 M NaCl, pH 8.0). After washing with 5.5 CV of buffer A, human Fc or entire IgG was eluted by a pH-step using buffer B (50 mM Tris, 100 mM glycine, 150 M NaCl, pH 3.0) over 8.3 CV. The fraction containing the protein was neutralized by adding 1:40 (v/v) 2 M Tris pH 8.0.

(24) The purified protein was subsequently further used for either treatment with enzymes to analyze non-cleaved carbohydrates, CE-SDS analysis or ESI-MS.

(25) CE-SDS MW Analysis

(26) Deglycosylation was monitored by CE-SDS-MW analysis, using Beckman PA800 with UV detection. The buffer of 100 g of each Protein A purified sample was exchanged to 20 mM Tris pH 8.0 using spin concentrators (5000 MWCO, Millipore). Non-reduced samples were prepared as described in SDS-MW Analyses Guide using the ProteomeLab SDS-MW Analysis Kit (Beckman Coulter). The final protein concentration was 1 mg/ml. Samples were applied to a preconditioned bare fused silica capillary (50 m ID30.2 cm). Pre-conditioning and separation were performed according to the instruction manual.

(27) Sample Preparation for ESI-MS

(28) The buffer of Protein A purified samples was exchanged to 2 mM MOPS pH 7.4, 150 mM NaCl, 0.02% (w/v) NaN.sub.3 using spin concentrators (5000 MWCO, Millipore). Proteins were frozen in liquid nitrogen and stored at 80 C.

(29) ESI-MS Analysis of Glycan Structures of Human Fc and Entire IgG by Direct Infusion (Off Line Detection)

(30) Desalting by Size Exclusion Chromatography:

(31) 20-50 g (up to 90 l) of Fc after treatment of antibody with the proteases plasmin and carboxypeptidase B and with endo-glycosidases Endo S and Endo H, or entire IgG after treatment with Endo S, Endo H and carboxypeptidase B, were injected onto a Sephadex G25 self-packed ECO SR column (5250 mm; KronLab) equilibrated with 2% formic acid, 40% acetonitrile at a flow rate of 0.5 ml/min for 30 minutes. The injected protein sample was desalted applying an 8 minute isocratic elution with 2% formic acid, 40% acetonitrile at a flow rate of 1 ml/min. The elution of the desalted protein was recorded by UV at 280 nm and the eluting sample (volume about 200-300 l) was collected in a 1.5 ml reaction vial. An aliquot of the desalted sample was manually filled into a metal coated glass needle (Proxeon Biosystems Nano ESI-needles, cat# ES387), inserted into the nanospray source of the mass instrument and sprayed into a ESI-Q-TOF II mass spectrometer from Waters or into a Q-Star Elite mass spectrometer from Applied Biosystems.

(32) MS Parameters for Direct Infusion:

(33) A) Of Plasmin-Treated Samples (Human Fc) on a Q-TOF II Instrument (Waters)

(34) MS spectra were acquired using a capillary voltage of 1000 V, a cone voltage of 30 V in a mass range from 1000-2000 m/z in positive ion mode using a source temperature of 80 C. Desolvation temperature was off. MS data were acquired for approx 2-3 minutes by the respective instrument software.

(35) B) Of Entire Antibody on a MaXis-ESI-MS Instrument (Bruker)

(36) MS spectra were acquired using a NanoMate device as spray interface. The values for data acquisition at the MS instrument were set to 400 Vpp (funnel RF), 120 eV (ISCID energy) and 400 Vpp (Multipol RF) regarding the transfer parameters, 5.0 eV (ion energy) and 300 m/z (low mass) for the quadrupole parameters, 15 eV (collision energy) and 3000 Vpp (collision RF) adjusting the collision cell and 800 Vpp, 160 s for transfer time and 20 s prepulse storage at the ion cooler. Data were recorded at a mass range from 1000-4000 m/z in positive ion mode.

(37) Molar masses of dimeric Fc-fragments and entire antibody, containing different combinations of glycan structures truncated by the endoglycosidases applied, i.e GlcNAc/GlcNAc, GlcNAc+Fuc/GlcNAc and GlcNAc+Fuc/GlcNAc+Fuc, were determined from the respective m/z pattern of the Fc fragment or entire antibody species using an in-house developed software. The relative ratios of the various residually glycosylated dimeric Fc fragments or entire antibodies were calculated with the same in-house software using the sum of peak areas of the m/z spectrum of a distinct glycosylation variant of the dimeric Fc-fragment or entire antibody.

(38) ESI-MS Analysis of Glycan Structures of Entire IgG by LC-MS (on Line Detection)

(39) LC-MS was performed on a Dionex HPLC system (Dionex Ultimate 3000) coupled to a Q-TOF II mass spectrometer (Waters). The chromatographic separation was performed on a ACE C4 column (5 m particle size, 300 A pore size, 130 mm; Advanced Chromatography Technologies). Eluent A was 0.1% formic acid, eluent B was 99.9% acetonitrile and 0.1% formic acid. The flow rate was 100 l/min, the separation was performed at 75 C. and 2 (10 l) of an intact antibody sample treated with Endo S and Endo H, but without plasmin treatment, were used.

(40) TABLE-US-00001 TABLE 1 Parameters for LC-MS. Time (min.) % B remark 0 25 waste 3 25 3.1 25 3.5 25 switch to MS 4.0 25 9.0 50 9.5 100 12.5 100 12.6 25 14.9 25 switch to waste 15.0 255 stop MS-detection

(41) MS spectra were acquired using a capillary voltage of 2700 V, a cone voltage of 80 V in a mass range from 1000-4000 m/z in positive ion mode using a source temperature of 100 C. Desolvation temperature was set to 200 C. MS data were acquired for approximately 11.4 minutes (gradient time 3.5 to 14.9 min) by the respective instrument software.

(42) Molar masses of intact antibody (consisting of two heavy chains and two light chains) containing different combinations of glycan structures truncated by the endoglycosidases applied, i.e GlcNAc/GlcNAc, GlcNAc+Fuc/GlcNAc and GlcNAc+Fuc/GlcNAc+Fuc, were determined from the respective m/z pattern of the antibody species using an in-house developed software. The relative ratios of the various residually glycosylated intact antibodies were calculated with the same in-house software using the sum of peak areas of the m/z spectrum of a distinct glycosylation variant of the intact antibody.

(43) The ratio of non-fucosylated heavy chains was determined by reducing the EndoS and EndoH-treated antibody with TCEP (Tris(2-carboxyethyl)phosphine hydrochloride) and performing an LC-MS analysis as described before, using the same column type and gradient setting but some modified parameters for MS data acquisition. MS parameters were the same as described before, but cone voltage was set to 25 V and mass range was from 600-2000 m/z.

Example 2: Results

(44) Deglycosylation of Fc

(45) N-Glycosidase F, also known as PNGase F, is a highly specific deglycosidase that cleaves between the innermost N-acetylglucosamine of high mannose-, hybrid-, and complex-type N-linked oligosaccharides and the asparagine residue of the glycoprotein to which the glycan is attached (Tarentino et al., 1985). Treatment of the Fc fragments of antibody A and C with PNGase F according to the instructions of the manufacturer was monitored by CE-SDS. Under these conditions PNGase F quantitatively removes the glycan moiety of both analyzed samples, resulting in a mobility shift of the main peak from 3.7910.sup.5 to 3.910.sup.5 (FIG. 2). Endo S cleaves the chitobiose core of N-linked oligosaccharides, leaving the first N-acetylglucosamine residueand an -fucose residue in case of fucosylated carbohydratesattached to the protein. The CE-analysis of a such digested glycoengineered sample revealed that approximately 10% of the protein were still undigested (FIG. 2a, Table 2), as demonstrated by a peak with a mobility of 3.8410.sup.5. Subsequent analysis by PNGase F treatment indicated that the Endo S resistant carbohydrates were entirely of hybrid structure suggesting specificity of this enzyme for complex carbohydrates. This result could be corroborated by the quantitative Endo S digestion of wildtype antibody A which resulted in homogenously deglycosylated protein (FIG. 2b).

(46) TABLE-US-00002 TABLE 2 Peak area of enzyme-treated Fc fragments evaluated by CE-SDS. Peak area [%] Antibody, enzyme Non-cleaved Cleaved A, no enzyme 99.3 0.7 A, PNGase F 1.3 98.7 A, Endo S 1.8 98.2 C, no enzyme 100.0 0.0 C, PNGase F 0.3 99.7 C, Endo S 10.6 89.4

(47) To confirm this hypothesis, Endo S-treated Fc of antibody C was purified by affinity chromatography to remove the enzyme and cleaved carbohydrates, and subsequently incubated with PNGase F to remove the entire glycan moiety. The hydrolyzed carbohydrates were further analyzed by MALDI TOF MS. The obtained spectra showed that Endo S is discriminating (i.e. sparing) either complex- or hybrid-type bisected structures that are corresponding to m/z=1663 (FIG. 3b).

(48) Further experiments were performed to determine whether the discriminated carbohydrates are complex- or hybrid-type bisected structures. After purification by affinity chromatography, the Endo S-treated Fc fragment of antibody C was incubated with PNGase F or Endoglycosidase H (Endo H). Endo H is a recombinant glycosidase that cleaves within the chitobiose core of high mannose- and hybrid-type N-linked oligosaccharides of glycoproteins. It is not able to cleave within complex structures. MALDI TOF MS spectra showed that the carbohydrates discriminated by Endo S are cleaved by Endo H, resulting in a main peak of m/z=1460 (FIG. 3c). These data clearly show that Endo S is not able to release hybrid-type bisected carbohydrates from the asparagine-linked N-acetylglucosamine.

(49) To obtain homogenously deglycosylated material that only varies in its -linked fucose content, a combined treatment of the Fc fragment of antibody C with Endo S and Endo H was performed resulting in a protein that was quantitatively deglycosylated after the first GlcNAc residue as observed by CE-SDS (FIG. 4a). MALDI-TOF MS analysis showed that the hybrid bisected structures (m/z=1460) are released by combination of these two enzymes (FIG. 4b). To confirm that there is no other carbohydrate attached to the N-acetylglucosamine with or without an -linked fucose residue, Endo S- and Endo H-treated Fc fragment of antibody C was incubated with PNGase F. No MALDI TOF spectra could be obtained after this treatment suggesting that no other carbohydrates were remaining that cannot be cleaved by Endo S or Endo H (data not shown).

(50) Determination of the Fucose Distribution in a Fc Preparation

(51) To quantify the distribution of the fucose linked to the N-acetylglucosamine residue attached to the Fc, ESI-MS analyses were performed. After incubation with Endo S and Endo H before separation by affinity chromatography, the Fc domains of antibodies A, B and C (generated by plasmin digestion) were treated with carboxypeptidase B to remove heterogeneity introduced by C-terminal lysine.

(52) ESI-MS spectra revealed Fc fragments with either two, one or no fucose linked to the residual GlcNAc still attached to the protein after EndoS/EndoH treatment (FIG. 5). Distribution of these three fucose species is summarized for the investigated three different IgGs A, B and C (calculated as relative ratio of the sum of peak areas in the m/z-spectra). The results correlate well with the fucose content determined by MALDI-TOF MS (Table 3).

(53) TABLE-US-00003 TABLE 3 Comparison of the a-fucosylation degree determined by mass spectrometry for Fc fragment of antibody A, B and C. ESI-MS MALDI-TOF 2 fucose 1 fucose 0 fucose Non-fuc Non-fuc [%] [%] [%] [%] [%] A 2.12 94 3 3 4.5 B 47.0 29 41 30 50.5 C 69.6 20 40 40 60.0
Deglycosylation of Entire IgG

(54) For deglycosylation of an entire IgG by combined treatment with Endo S and Endo H, cleavage conditions had to be optimized. Deglycosylation with a molar ratio of Endo S to IgG of 1:20, as was used for deglycosylation of the Fc fragment, was insufficient to deglycosylate entire IgG. Increasing the concentration of Endo S to a molar ratio of 1:7 was sufficient to get homogenously deglycosylated material that only varies in its -linked fucose content observed by CE-SDS (FIG. 6a). MALDI-TOF analysis showed that the carbohydrates are released by combined treatment with Endo S and Endo H (FIG. 6b). Using this approach it is possible to analyze the allocation of fucose per IgG without separate generation of the Fc-fragment.

(55) Determination of the Fucose Distribution of Entire IgG

(56) Quantification of the distribution of fucose linked to the innermost N-acetylglucosamine residue of N-linked glycans of entire IgGs was performed using wildtype antibody A (2.12% a-fucosylation) and glycoengineered antibody D (85.0% a-fucosylation). After combined treatment with Endo S and Endo H, both IgGs were incubated with carboxypeptidase B to remove heterogeneity introduced by C-terminal lysine. The antibodies were subsequently purified by affinity chromatography.

(57) Allocation of the core fucose per IgG was determined using two different methods. The pattern of the m/z-spectra obtained by ESI-MS off line detection revealed IgG-species with either two, one or no fucose attached to the residual GlcNAc after EndoS/EndoH treatment (FIG. 8). Distribution of these three fucose species is summarized for the investigated two different IgGs A and D (calculated as relative ratio of the sum of peak areas in the m/z-spectra) (Table 4).

(58) LC-MS analyses were also performed to determine the allocation of fucose per IgG (FIG. 9). For both IgGs, m/z-spectra showed a similar ratio of species with either two, one or no fucose attached as observed in ESI-MS offline detection, (Table 4). Peak areas below 5% are in the detection sensitivity of the methods for entire IgG. Ratio for non-fucosylated heavy chain is presented in Table 4, column Non-fuc [%].

(59) TABLE-US-00004 TABLE 4 Comparison of the a-fucosylation degree and fucose allocation determined by ESI-MS and LC-MS analyses for antibody A and D. ESI-MS LC-MS Non- 0 fu- Non- 2 fucose 1 fucose 0 fucose fuc 2 fucose 1 fucose cose fuc [%] [%] [%] [%] [%] [%] [%] [%] A 92 5 <5 8 94 6 <5 12 D 9 24 67 81 10 24 66 80