Open access peer-reviewed chapter

Characterization of Whole and Fragmented Wild-Type Porcine IgG

Written By

Claudia Nelson, Raymond Bacala, Baylie Gigolyk, Evelyn Ang, Haley Neustaeter, Emy Komatsu, Oleg Krokhin, Dave Hatcher and Hélène Perreault

Submitted: 15 August 2018 Reviewed: 29 November 2018 Published: 08 January 2019

DOI: 10.5772/intechopen.82792

From the Edited Volume

Recent Advances in Analytical Chemistry

Edited by Muharrem Ince and Olcay Kaplan Ince

Chapter metrics overview

1,340 Chapter Downloads

View Full Metrics


Glycoproteomic analyses of tryptic (glyco)peptides from wild-type (WT) porcine IgG were performed. In a first protocol, intact antibody was digested with trypsin, followed by glycopeptide enrichment and liquid chromatography-tandem MS (HPLC–MS/MS). This procedure allowed to detect N-glycopeptides observed previously (Lopez, P. G. et al., Glycoconj. J. 2016, 33 (1), 79), plus other non-reported N-glycopeptides. The method provided useful information but did not allow to discern between Fab (antigen-binding region) and Fc (constant region, fragment crystallizable) peptides/glycopeptides. In a second scheme, glycoproteomic analysis was attempted for Fab and Fc fragments obtained by papain and Fabulous™ hydrolysis. Usually employed for milligram amounts of antibodies, the papain and Fabulous™ protocols were adapted to 200 μg of WT IgG. Fab and Fc fragments were separated by size-exclusion (SEC) HPLC. Fractions collected were reanalyzed by gel electrophoresis (SDS-PAGE). Bands were excised, and fragments digested in-gel, followed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) MS and HPLC/MS–MS. In the protocol no glycopeptide enrichment was involved, that is, whole tryptic digests were analyzed. Fc N-glycopeptides were identified, and greater numbers of non-glycosylated peptides were tabulated. Very few peptides overlapped between Fc and Fab, as most peptides were clearly from Fc or Fab. HPLC-MS/MS detected more sialylated glycoforms than MALDI-TOF-MS. Sections of Fab and Fc were assigned de novo, through a database search or manually.


  • porcine IgG
  • papain
  • enzymatic fragmentation
  • Fabulous™
  • glycoproteomics

1. Introduction

There have been reports on the mass spectrometric (MS) analysis of pig immunoglobulins (IgG) in relationship with use in a xenotransplantation context [1, 2, 3, 4]. These studies have explored the amino acid composition and glycosylation of pig IgG according to glycoproteomic [2, 3] and glycomic [1] workflows involving the enzymatic digestion of whole antibodies. Glycoproteomic workflows resulted in the identification of many peptides that could be matched with the conserved gamma portion of the heavy chains of some of the 11 subtypes of pig IgG [5], including N-glycopeptides EEQFNSTYR and EAQFNSTYR [3]. No specific information was given on the variable portions of neither Fab nor Fc components, as most of such assignments would have had to be attributed de novo. The conserved Fc glycosylation site is often described as the only IgG site glycosylated at 100%, in spite of the fact that 10–15% of wild-type antibodies have glycosylation also in their variable region [6], and reports have shown that even higher glycosylation levels (up to 30–40%) can exist in the variable regions of polyclonal IgGs [7, 8].

For more specificity, the analysis of antibodies by MS can take great advantage of enzymatic fragmentation with papain [9, 10] or new enzymes produced by recombinant methods and available on the market [11]. This type of procedure has not been reported for the fragmentation of porcine IgG, to the authors’ knowledge. For instance, procedures have been published for mouse [12, 13], chicken [14], rabbit [15], sheep [16], and human [17, 18] IgGs. The two groups of antibody fragments of primary interest are the antigen-binding fragments (Fab) and class-defining crystallizable fragments (Fc). The hinge region of immunoglobulins (IgGs) is readily accessible to proteolytic attack by enzymes [9, 10], and cleavage at that point produces F(ab’)2 or Fab fragments and the Fc fragment. Papain is a nonspecific thiol-endopeptidase and has a sulfhydryl group in its active site, which must be reduced for activity. When IgG molecules are incubated with papain in the presence of a reducing agent, one or more peptide bonds in the hinge region are split, producing three fragments of similar size: two Fab fragments and one Fc fragment [9, 10].

Fabulous™ enzyme is a recombinant cysteine protease that under reduced conditions digests in the hinge region of antibodies from many species and subclasses, including human, mouse, rat, and goat, yielding Fab and Fc fragments [11]. As a reducing agent is present during digestion, it is likely that interchain thiols will be reduced. Fabulous™ and papain have typically been used for the production of relatively large amounts of antibody fragments (10 mg of starting material), whereas methods adapted to MS require much less, about 50–200 μg. There is a need for downscaling these workflows, especially for porcine IgG, which has not been previously studied by fragmentation followed by MS.

Reports on post-fragmentation MS analyses of antibodies have demonstrated that detailed fragment characterization allows for the identification of more glycosylation sites than bottom-up approaches [19, 20]. It is also important to study amino acid sequences of the variable portions of IgG for therapeutic purposes [21], and the information obtained after fragmentation is much more specific than data generated through the tryptic digestion of whole intact antibodies.

The interest of the present study is to compare results from two main workflows aimed at characterizing wild-type porcine IgG’s glycosylation and amino acid sequence features. In the first workflow, porcine IgG is digested with trypsin, followed by glycopeptide enrichment. Reversed-phase high-performance liquid chromatography (RPLC) coupled with electrospray ionization (ESI) MS and tandem MS (MS/MS) is used to characterize the glycopeptide-rich fraction. The second workflow involves subjecting porcine IgG to fragmentation by one of two enzymes, papain or Fabulous™. The steps necessary between fragmentation and MS included size-exclusion chromatography (SEC), sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and tryptic digestion. This is the first attempt to characterize porcine IgGs in small amounts (sub-mg) using a combination of these methods. Two different MS techniques were used for the analysis of tryptic products of antibody fragments: matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) MS and ESI/MS–MS coupled with RPLC.

As the tryptic digestion products of whole wild-type porcine IgG antibodies have been characterized by MALDI-MS [2, 3] and ESI-MS and MS/MS (new results presented in this report), data from these different workflows serve as comparative benchmarks between intact and fragmented IgG samples. Overlaps and differences allow to identify peptides and glycopeptides as originating from either the Fc or Fab portions, and database searches [22] can verify if these sequences are already available in the literature or need to be determined de novo. Porcine IgG is a complex mixture of subtypes, and the complementarity of MALDI- and LC/ESI-MS–MS brings a considerable amount of information to document the identification of these IgG components.


2. Experimental

2.1 Materials

Wild-type porcine IgGs were obtained from Université de Nantes (Dr. Jean-Paul Soulillou’s Laboratory). Trypsin Ultra™ was purchased from Promega (Wisconsin, USA). The Fabulous™ enzyme was kindly provided by Genovis (Cambridge, MA). Dihydroxybenzoic acid (DHB), sinapinic acid, ammonium bicarbonate, dithiothreitol (DTT), L-cysteine, iodoacetamide (IAA), and trifluoroacetic acid (TFA) were purchased from Sigma-Aldrich (St. Louis, MO). Mini-Protean TGX precast gels (4–15%), Precision Plus™ protein standard, 2-mercaptoethanol, and 4x Laemmli sample buffer were obtained from Bio-Rad (Hercules, CA). Imperial™ protein stain, tris base (2-amino-2-(hydroxymethyl)-1,3-propanediol) and immobilized papain-cross linked and 6% in beaded agarose supplied as 50% glycerol in sodium acetate pH 4.5 were purchased from Thermo Scientific (Rockford, IL). Strata-X C-18 cartridges were obtained from Phenomenex (Torrance, CA). Acetonitrile (ACN) was purchased from EMD-Millipore (Darmstadt, Germany). Sodium phosphate dibasic anhydrous was purchased from McArthur Chemical Co. Ltd. (Montreal, Canada). Hydrochloric acid was purchased from Anachemia (Vancouver, Canada) and deionized water was obtained from an adapted filtration system (Millipore).

2.2 Tryptic digestion of whole porcine IgG

Porcine IgG (200 μg) was dissolved in 100 mM ammonium bicarbonate (pH~8) and vortexed. A DTT solution (10 mM) was added to the sample, which was then vortexed and left to react at 56°C for 40 min, then cooled to room temperature. After 500 mM IA was added, and the sample was left to react in the dark for 45 min. The excess of IA was quenched with the addition of 100 mM DTT, and the sample was left to react for 10 min in the dark. Trypsin was added and proteolysis took place at 37°C for ~18 h. To deactivate trypsin, the sample was frozen and dried under vacuum. Glycopeptide enrichment was then performed using a ProteoExtract™ glycopeptide enrichment kit (Millipore-Sigma, Etobicoke, ON) according to the manufacturer’s procedure [23].

2.3 Papain digestion of porcine IgG

Just before use, 20 mM sodium phosphate digestion buffer was prepared with a 10 mM cysteine content, and the pH was adjusted to 7 Bead-immobilized papain slurry (20 μL, 50%) was added to an Eppendorf™ tube. The beads were washed twice with 160 μL of digestion buffer and then re-suspended in the buffer. Porcine IgG (200 μg) was dissolved in the digestion buffer. This was added to the tube containing the washed immobilized papain and digestion buffer. The sample was incubated for ~24 h at 37°C. Constant mixing of the bead slurry was maintained during the whole incubation. The bead-immobilized enzyme was separated from the digest by centrifugation and 20 μL of 10 mM Tris–HCl, pH 7.5 was added before centrifugation. The supernatant, which contained the IgG fragments, was removed.

2.4 Fabulous™ digestion for porcine IgG

The IgG sample (200 μg) was added to 200 units of Fabulous™ enzyme in 200 μL of 10 mM Tris, 50 mM cysteine buffer. The sample was vortexed and incubated at 37°C for 1 h.

2.5 Fractionation of IgG fragments by HPLC using a SEC column

The digestion mixtures were injected into a preconditioned SEC-300 4.6 × 300 mm silica-based column (Phenomenex, Torrance, CA). The mixtures were eluted with a mobile phase of 0.1% TFA, 0.1% formic acid in 20% ACN at a flow rate of 0.3 mL/min (manufacturer’s recommendation). The HPLC system used was a Waters 1525 binary pump equipped with a Waters 2707 autosampler and a Waters 2998 photodiode array detector (Milford, MA). Fractions were collected, dried, and re-suspended for MALDI-MS analysis.

2.6 Separation of IgG fragments by SDS gel electrophoresis

Once fractionated by SEC, Fab and Fc components were separated on a Mini-Protean™ Tetra cell system (Bio-Rad). Bio-Rad TGX™ gels used had 10 wells and a density gradient of 4–15%. Wells were washed individually four times with running buffer (10 × tris-glycine-SDS buffer diluted 1 × with water) prior to the loading samples. Each sample fraction containing Fab, Fc, or both Fab and Fc had its own lane on the gel. Each gel was loaded with 15 μL of each fraction (in water) in 11.3 μL of 4 × Laemmli sample buffer, without adding 2-mercaptoethanol. Well 1 was loaded with 10 μL of Precision Plus Protein Kaleidoscope™ standard. Intact-reduced IgG (15 μL, ~14 μg) was loaded into well 2. IgG fragments (~14 μg) were loaded in other lanes. Running buffer was poured in the cell system and the voltage was set at 150 V. Samples were allowed to migrate for 40 min, until the dye front reached the bottom of the gel. The gel was removed from the cell and was rinsed four times with Millipore water, and sufficient Imperial™ protein stain was added. IgG fragments absorbed the stain overnight, and the stain was decanted and replaced with Millipore water until gel bands became visible.

2.7 In-gel tryptic digest of IgG Fab and Fc fragments

Non-reduced Fab and Fc bands were excised from the gel. Tryptic digestion was performed on each single cut out band. Bands were cut into small pieces and placed into 1.5 mL Eppendorf tubes. The digestion buffer was 50 mM ammonium bicarbonate in water. Each tube contained one lane worth of gel. Gel pieces were washed with wash buffer (1:1 digestion buffer-ethanol) until all protein stain was removed. They were then incubated in absolute ethanol for 10 min. Gel pieces were then rewashed with digestion buffer for 20 min and then incubated again in absolute ethanol for 20 min, which was removed from the gel by vacuum centrifugation. Trypsin solution was added and the tubes were placed on ice where the gel was allowed to swell. Thereafter, excess trypsin solution was discarded. Gel pieces were covered with digestion buffer and incubated at 37°C for ~18 h. Glycopeptides and peptides were extracted on C18 cartridges with buffer A, 0.5% acetic acid; extraction buffer B, 5:3 30%; ACN, 0.5% acetic acid; and extraction buffer C, 100% ACN. Samples were dried down for further analysis.

2.8 Peptide-N-glycosidase F (PNGase F) removal of glycans from trypsin digested Fab and Fc fragments

A solution of PNGase F (4 μL, 10 units/μL) was added to a solution of tryptic glycopeptides. The sample was vortexed and set at 37°C for ~18 h. After the digest, glycans were separated from the de-glycosylated peptides on a C18 cartridge. The cartridge was conditioned with 5 x 1 mL (ACN + 0.1% TFA), then 5 × 1 mL of (H2O + 0.1% TFA). The sample was loaded onto the cartridge and allowed to equilibrate for about 1–2 min. Glycans were eluted with 3 mL H2O + 0.1% TFA and collected in two fractions. Peptides were eluted with 1.5 mL of 50:50 ACN:H2O + 0.1% TFA. Solvent was evaporated from the fractions.

2.9 Preparation of samples for MALDI-MS analysis

Glycopeptide fractions were mixed directly with 5 μL of DHB in 30:70 ACN: water. Samples (1 μL) were then spotted onto the stainless steel target and allowed to dry, for reflector positive mode MALDI-MS. For the Fab and Fc fragments from HPLC fractions, 5 μL of 0.1% TFA and 5 μL of sinapinic acid in 0.1% TFA, 30:70 ACN: water was added, and 1 μL was spotted onto the target already pre-spotted with 0.5 μL of sinapinic acid in ethanol. Spots were then allowed to dry for linear positive mode MALDI-MS.

2.10 MALDI-TOF-MS analysis

These analyses were performed on an UltraFleXtreme™ mass spectrometer (Bruker, Billerica, MA) equipped with LID-LIFT™ technology for tandem MS experiments. For positive ionization mode, a nine-peptide calibration mixture with masses ranging from 500 to 5000 Da was used. In linear positive mode, the instrument was calibrated using tryptic peptides of cytochrome C, mass ranging from 10,000 Da to 160,000 Da.

2.11 Preparation of samples for HPLC-ESI-MS analysis

For the Fab and Fc tryptic digest fragments, 100 μL of water was added to the pooled C18 cartridge fractions of each Fab and Fc tryptic digests. Samples were sonicated and then ready for HPLC–MS analysis.

2.12 RPLC-MS/MS analysis of Fab and Fc tryptic digests

Both digestion mixtures (50 μL) were in turn injected into a preconditioned Acquity BEH C18 (1.7 μm, 2.1 × 50 m) silica-based, reverse phase column, on a Waters Acquity UPLC system (Waters, Mississauga, ON). The flow rate was 0.25 mL/min, and a linear increase from zero to 28% ACN in water with 0.1% formic acid was used. Mass spectrometric detection was performed on a Waters G2 Synapt ESI-MS instrument. Positive ionization mode was used. The analyzer mode was set to high resolution, with a capillary voltage of 3.00 kV and a cone voltage of 25 V. The Progenesis™ software was used to handle and search databases for these HPLC-MS/MS analyses.

2.13 Analysis of whole porcine IgG tryptic digest by RPLC-MS/MS

Separations were conducted on a LC Ultra system (Eksigent, Dublin, CA). A 100 μm × 200 mm analytical column packed with 3 μm Luna C18 (Phenomenex, Torrance, CA) was used, at 500 nL/min flow rate. A 300 μm × 5 mm PepMap 100 trap-column (Thermo Fisher, San Jose, CA) was used to protect the analytical column. The elution gradient was as described above. A Triple TOF 5600 mass spectrometer (ABSciex, Concord, ON) was used in standard MS/MS data-dependent acquisition mode. Mass spectra (250 ms) were collected (m/z 370–1250) and followed by up to 20 MS/MS measurements on the most abundant parent ions (400 counts/s threshold, +2 to +5 charge state, m/z 100–1500 mass range for MS/MS, 100 ms each). Database search was performed using the Global Proteome Machine (GPM) proteomics data analysis system [24].


3. Results and discussion

3.1 Mass spectrometric analysis of tryptic digests of whole porcine IgG samples

In previous studies, wild-type porcine IgG was digested with trypsin, and glycopeptides/peptides were fractionated on a C18 cartridge. Glycopeptide fractions were analyzed by MALDI-MS/MS [2, 3, 30]. In a new separate experiment, the results of which are presented here, all tryptic products were then enriched for glycopeptides using an EMD-Millipore ProteoGlycan™ kit [23]. The glycopeptide-enriched fraction was injected into a RPLC-MS/MS system, using data-dependent MS/MS acquisition. Figure 1a shows the MS/MS total ion chromatogram (TIC) obtained from the elution, whereas Figure 1b represents the elution of peptides with m/z 204 ions as product in their MS/MS spectra, identified as glycopeptides [25]. Although enrichment was performed, there were still non-glycosylated peptides present in the sample. The m/z 204 trace shows that glycosylated peptides eluted early in the analysis and that the most abundant compounds eluting in (a) were most probably non-glycosylated peptides.

Figure 1.

(a) Total HPLC/MS ion chromatogram obtained from the MS/MS spectra of porcine IgG tryptic peptides and glycopeptides after enrichment on EMD-Millipore Proteoglycan™. (b) Selected MS/MS fragment ion chromatogram at m/z 204, indicating the elution times of glycopeptides [25].

A database search using GPM [24] helped to identify some non-glycosylated peptides, while for glycopeptides the extracted m/z 204 chromatogram was used to manually identify as many glycopeptides as possible. Table 1 gives a listing of all peptides detected and sequenced with GPM [24] or manually. For glycopeptide sequencing, MS/MS spectra were treated with CycloBranch [26] after removing abundant glycopeptide ions. As a few sequences are available in UniprotKB [22] for the Fc gamma portion of porcine IgG (K7ZLAZ, L8B180, L8B0S7, L8B0S2, L8B0Z4), and more are available [3] from previously published DNA sequences [5], some peptides show more than one identification source in Table 1. Only one entry was found for a porcine IgG Fab portion, P01846, which corresponds to the lambda (λ) constant region. Figure A1 compares the published porcine IgG Fc heavy chain sequences relating to Table 1 assignments.

Peptide sequencem/z (M + H)+Error (ppm)m/z glycoformGlycoformIdentification source*
NFSTYR787.3702231.14 2392.28 2701.08 2863.10G0F G1F G1FS G2FSIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b, IgG6a-b,L8B0S7,L8B0S2,K7ZLA7,L8B180, L8B0Z4
LIYQATNR978.543.37IPR 007110 (Ig C1-set)
EAQFNSTYR1115.49−19.42414.05 2560.12 2722.15 2884.24 3029.35 3191.50G0 G0F G1F G2F G1FS G3FIgG6a,L8B180,L8B0Z4
EEQFNSTYR1173.5−14.62471.01 2618.09 2634.07 2780.13 3087.23 2942.19 3104.23 3249.28 2833.04G0 G0F G1 G1F G1FS G2F G3F G2FS M9-NIgG1a-b,IgG2a-b,IgG4a-b,IgG5ab,IgG6b, L8B0S7,L8B0S2,K7ZLA7
H2N-300-QNFSVFR1212.132656.09 2819.14 2981.20G0F G1F G2FUnavailable
Unknown1467.282911.26 3073.31G0F G1FUnavailable
QLIYQTNSRPTGVPSR1816.971.98IPR 007110 (Ig-like)

Table 1.

Masses and sequences of peptides found in a glycopeptide-enriched fraction from porcine IgG (wild-type) tryptic digestion products. Red: Labeled manually; black: Sequenced and/or matched by GPM [24]. Glycopeptides, m/z values in bold not observed, only glycoforms.

L8B180,L8B0Z4,L8B0S2,L8B0S7,K7ZLA7 porcine IgG heavy chain entries (Fc) [22]; P01846, porcine IgG lambda constant domain (Fab); K7ZJP7, IgM heavy chain; P01786, mouse IgG heavy chain; P01857, human IgG heavy chain; IgGs. from [3]; IPR, GPM entries.

Points of interest arising from this table are (i) the presence of a large number of non-glycosylated, even after enrichment, from the Fab and Fc portions of the antibody and (ii) the detection of seven distinct glycosylated peptides, some with complex glycoforms and one with high-mannose glycoforms, thus of the N-type. Indeed, all MS/MS spectra of these glycopeptides showed characteristic fragment ions corresponding to the (M + H)+ ions of the bare peptides, next to the characteristic (M + H + 83)+ and (M + H + 203)+ ions [27], with the latter being predominant.

Figure 2 shows four MS/MS spectra of glycopeptides, starting with two of the complex G0F forms of peptides of (a) constant region EAQFNSTYR [3] and (b) a variable region sequence partially determined as H2N-(300)-QNFSVFR by the CycloBranch software [26]. In these cases, the mass difference between the protonated precursors and (M + H)+ bare peptide fragments is 1444. The m/z 204 and 366 fragment ions are predominant as glycopeptide signature. In Figure 2c, a sialylated complex glycoform is featured, the G1FS form of peptide EEQFNSTYR [3]. In previous studies of wild-type pig IgG, the presence of N-glycolyl neuraminic acid (NeuGc) is featured exclusively (i.e., no N-acetyl neuraminic acid (NeuAc)) has been reported [2, 3]. The fragmentation of NeuGc-containing species produces distinctive m/z 308 and 290 fragments, as observed in (c). Figure 2d shows the fragmentation of a high-mannose glycoform of a peptide of undetermined sequence. The bare peptide (M + H)+ ions appear at m/z 1069, with +83 ions (m/z 1152) and + 203 ions (m/z 1272).

Figure 2.

Tandem mass spectra of doubly or triply charged ions of glycopeptides enriched from the tryptic products of porcine wild-type IgG. (a) G0F glycoform of EAQFNSTYR, (b) G0F of undetermined peptide, (c) G1FS of EEQFNSTYR, (d) Man-9 of undetermined peptide.

For known peptide sequences such as in Figure 2a and c, it is possible to find most bare peptide y and b ions, although they appear with very low abundance and are not accounted for by the search engine, due to the domination of all glycopeptide signature ions. There was an attempt by the authors to sequence all unknown glycopeptide sequences, with partial success, as indicated in red in Table 1.

Overall results suggest that N-glycosylation occurs in the Fc but also in the variable regions of the Fab and/or Fc domains of porcine IgG. Each glycosylated peptide detected indicated patterns linked to N-glycosylation, while there was no obvious detection of O-glycans. This may be due to the conditions used to enrich the glycopeptides with the EMD-Millipore Kit, which were optimized for N-glycosylated peptides [23, 28]. Results obtained with this first workflow will be compared with those generated with a more elaborated procedure involving fragmentation enzymes as discussed below.

3.2 Fragmentation of porcine wild-type IgG with papain and fabulous™ followed by SEC

Antibody samples were first fragmented on immobilized papain, and thus it is expected that only IgG-related products will be present in the mixture. As shown by Figure 3a, all antibody was fragmented (intact antibody would have appeared at ca. 4.5 min). In general, when IgG is incubated with papain with a reducing agent, one or more peptide bonds in the hinge region are split, producing three fragments of similar size: two Fab fragments and one Fc fragment. The Fc may remain intact based on conditions and enzyme used [9, 10]. The cleavages occur at cysteines around position 271 (in Figure A1), about 10 amino acids from the IdeS cleavage site [29]. In human IgG subtypes (IgG1-4), there are on average of three cysteines in the range of hinge region positions 265–275 to make the numbers correspond with those of Figure A1, where papain cleavages can be initiated [29]. Porcine IgG has similar cysteine motifs in these positions; however, Figure A1 shows different lengths of amino acid chains in the hinge region, which did not seem to prevent fragmentation. Interestingly in this papain-fragmentation experiment, the Fc did not remain intact, but the Fab did. Figure 3a shows significant separation although not at the baseline, but which still allowed the collection of Fab and Fc fractions.

Figure 3.

Size exclusion chromatograms obtained for 200 μg of wild-type porcine IgG fragmented by (a) immobilized papain and (b) Fabulous™.

In order to further isolate Fab and Fc segments, collected fractions were analyzed in turn by SDS-PAGE. This allowed identification by mass, with confirmation by linear mode MALDI-TOF-MS (not shown) and in-gel tryptic digestion of the bands (analyzed by reflector mode MALDI-TOF-MS and HPLC-MS, next section). The first major peak in Figure 3a was identified as originating from the Fab, and the second peak was identified as the Fc. The intact Fab fragment is larger and thus elutes prior to the split Fc, while the opposite is usually true in the case of human IgG [15].

Fabulous™ is a recombinant cysteine protease which under reducing conditions digests in the hinge region of antibodies from many species and subclasses, including human, mouse, rat, and rhesus monkey, yielding Fab and Fc fragments [11]. Some specific fragmentation site information is available for human IgG1, mouse IgG1, and rabbit IgG, but no information at all pertains to the fragmentation of porcine IgG. Looking at all porcine IgG subtype sequences (see alignment in Figure A1), they have a motif similar to that of rabbit IgG (KP270I/CPP) [11], that is, with a potential fragmentation site between isoleucine I and cysteine C: CP270I/CPG or CP270I/CPA.

In general, when IgG molecules are incubated with Fabulous™ in the presence of a reducing agent, one or more peptide bonds in the hinge region are split next to a cysteine, producing two Fab fragments and one Fc. As the reducing agent is present during the digestion reaction, it is likely that all interchain thiols will be reduced.

In this experiment with Fabulous™ using 100 μg of pig IgG, a small portion of the antibody was not fragmented; otherwise the Fc was split into two halves and the Fab fragments remained intact. The HPLC-SEC chromatogram is shown in Figure 3b. The first major peak was indicative of the Fab fragment, and the second peak eluting after identified as the Fc, as verified later by SDS-PAGE and further tryptic digestion. As noticed in the chromatograms of Figure 3, elution times are different, that is, longer in portion b. This is due to gradual deterioration of the column, as both experiments depicted in a and b were performed several months apart.

The fact that Fabulous™ (28,724 Da [11]) was free in solution and not immobilized as in the case of papain involves that it would elute in the SEC chromatogram (Figure 3b), most probably in the second peak with the Fc. If this proprietary enzyme has a sequence similar to that of papain, subsequent proteolysis by trypsin is likely to occur extensively, as many lysine and arginine residues are present in papain’s sequence [31].

3.3 SDS-PAGE separation of porcine IgG fragments

For papain-produced fragments, a nonreducing gel experiment was performed on the previously collected SEC fractions (Figure 4a). This experiment allowed confirming the identity of the fragments. A reducing SDS-PAGE experiment was also performed on intact pig IgG to serve as a control (Figure 4b). Bands were then excised from the gel, followed by in-gel tryptic digestion. This helped to single out heavy and light chains for differentiating between Fab and Fc according to their respective known glycopeptides and peptides [2, 3]. In Figure 4a, lanes 6–9 correspond to a fraction containing both Fab and Fc collected at the junction of SEC peaks in Figure 3a. A single 50 kDa Fab band and two 25 kDa bands were obtained. The Fab band is positioned just below the 50 kDa marker band and clearly below the Fc band in Figure 4b, and the Fab should have a lower molecular weight than intact Fc [32]. Both bands appearing in the Fc region were digested independently with trypsin as denoted by “upper Fc band” and “lower Fc band.” It appears from further results that they belong to different IgG subtypes. For Fabulous™ fragments, the same experiment was conducted and results are shown in Figure 4b. According to gel separations, papain and Fabulous™ had very similar fragmentation effects on porcine IgG.

Figure 4.

(a) Fab and Fc SEC fractions from papain-fragmented porcine IgG on nonreducing SDS-PAGE and (b) from Fabulous™-fragmented porcine IgG, (c) whole porcine IgG runs under reducing conditions.

3.4 MALDI-MS of in-gel tryptic-digested Fc bands

Figure 5a shows the tryptic products for the upper Fc band (papain generated) in Figure 4a. The main glycopeptide observed has three main glycoforms at m/z 2472, 2618, and 2780 as [M + H]+ ions. In Figure 5b (lower Fc band), two series of glycopeptides are observed, that is, the same as above and another one at m/z 2414, 2560, and 2722. The amino acid sequences of these glycopeptides were verified by MALDI-MS/MS as EEQFNSTYR and EAQFNSTYR, respectively [2, 3], with glycoforms as indicated in the figure. PNGase de-glycosylation of these two samples led to the spectrum of Figure 5c, where peptides at m/z 1117 (EAQFDSTYR) and 1175 (EEQFDSTYR) were detected, while the analysis of released glycans was not successful. Similar MALDI-MS and MS/MS results were obtained for the Fabulous™-generated porcine IgG that combined upper and lower Fc fragments digested with trypsin. Besides glycopeptides from the Fc region, other Fc ɤ-domain peptides were observed, as identified from the database [22] inTable 2.

Figure 5.

Reflector positive mode MALDI-TOF-MS spectra of (a) tryptic digestion products from pig IgG Fc fragment (higher 25 kDa band) and (b) lower Fc fragment. Both bands were excised from the gel run under non-reducing conditions shown in Figure 4a. (c) De-glycosylated peptides after removal of glycans to give peptides at m/z 1117 and 1175.

Peptide sequencem/z (M + H)+Error ppmm/z glyco-formGlyco-formSource*
DLPAPITR882.5−5.78IgG2ba-b,IgG4a-b,IgG6a-b, L8B180,L8B0S7,L8B0Z4
EEQFNSTYR1173.50−14.62471.60 2779.62 2779.63G0 G0F G1FIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b, IgG6b,L8B0S7,L8B0S2,K7ZLA7
SNGQPEPEGNYR1347.6−1.48IgG1a,IgG6a-b,L8B180,L8B0S2, K7ZLA7,L8B0Z4
VSSQNIQDFPSVLR1589.842.9K7ZJP7 (IgM HC const. region)
YAASSYLALSASDWK1632.8−24.5P01846 UniprotKB (Ig λ const. region)
VVSVLPIQHQDWLK166213.7IgG1a-b,IgG3,IgG6b,L8B180,K7ZLA7, L8B0Z4
STGKPTLYNVSLVLSDT1794.8−60.4K7ZJP7 (IgM HC const. region)
TTPPQQDVDGTYFLYSK1959.9−1.07IgG2a-b,IgG3,IgG4a-b,IgG6a,L8B180, L8B0S7,L8B0Z4
AAPTVNLFPPSSEELGTNK1971.9−36.6P01846 UniprotKB (Ig λ const. region)

Table 2.

Masses and sequences of peptides found by MALDI-MS in porcine IgG (wild-type) tryptic digestion products of combined Fc bands obtained by papain and Fabulous™ fragmentation. Peptides and glycopeptides in red were sequenced manually after MS/MS and/or tentatively identified by mass fingerprinting.

L8B180,L8B0Z4,L8B0S2,L8B0S7,K7ZLA7, porcine IgG heavy chain entries (Fc) [22]; P01846, porcine IgG lambda constant domain (Fab); K7ZJP7, IgM heavy chain; P01786, mouse IgG heavy chain; P01857, human IgG heavy chain; IgGs, from [3]; IPR, GPM entries.

3.5 In-gel tryptic-digested fab bands analyzed by MALDI-MS

The bands just below 50 kDa from both papain and Fabulous™ digestion processes were excised from the gel and digested with trypsin. MALDI-MS spectra of the products were obtained, and display no obvious glycopeptides as found in the Fc. Some peaks could be identified as light and heavy chain Fab peptides. Fab peptide masses and sequences available are listed in Table 3.

Sequencem/z ESIm/z MALDIError ppmIdentification source
FSGAISGNK880.45880.39−70.8P01846 (Ig λ const. region)
YAASSYLALSASDWK1632.791632.8536.7P01846 (Ig λ const. region)
AAPTVNLFPPSSEELGTNK1972.001972.1890.2P01846 (Ig λ const. region)

Table 3.

Fab tryptic peptides from porcine IgG fragmented with papain and fabulous™. Second column (m/z ESI): Already identified in Table 1. Error: Measured using the MALDI m/z values against calculated values. Sequences in red are assigned from mass only based on Table 1.

Overall, MALDI-TOF-MS experiments allowed suggesting that efficient separation of Fab and Fc fragments from each other occurred, using the SEC/SDS-PAGE procedure. There was a minimal number of overlapping peptides between Fab and Fc MALDI spectra (Tables 2 and 3). In Table 3, peptides in red had already been sequenced and appeared in Table 1; peptides whose m/z values appear in the second column were also observed in Table 1, although not sequenced. Overlapping peptides between Tables 2 and 3 are at m/z 2509 (unknown), 2884 (glycoform of m/z 1115), and m/z 1677 (IgM). The presence of IgM in the sample had been noticed in an earlier report [3].

3.6 Tryptic products of Fabulous™-fragmented porcine IgG analyzed by HPLC/MS

Each sample of digestion products from Fc and Fab bands were analyzed twice by HPLC/MS, and results are summarized in Table 4. In Figure A2, total ion chromatograms (TIC) of the pig Fab and Fc tryptic products showed differences in the retention times of some peaks as expected. Peaks observed at the same retention times (e.g., 4.41–4.46, 10.35, and 23.10 min) and common to all injections corresponded to singly charged ions of small compounds of m/z ˂ 700 and were not considered in the analysis, as they probably were non-peptide contaminants. Peaks at 5.96 min also present in both chromatograms corresponded to a mass 1660 peptide from the Fab, AGGTTVTQVETTKPSK. Non-glycosylated peptides in Table 4 were identified through MS/MS and database search, while glycopeptides were assigned by mass only in reference to entries in Table 1.

Mass (calc.)m/z (M + H) + (exp.)m/z error (ppm)Normalized abundanceSequenceSources of identification*
Pig Fab wt (avg)Pig Fc wt (avg)
881.492882.500−5.780.0410.6DLPAPITRIgG2ba-b,IgG4a-b, IgG6a-b,L8B180,L8B0S7,L8B0Z4
1524.8221525.830−2.621005.820216.6LVESGGGLVQPGGSLRL8B180, L8B0S7
1578.7441579.75247.704.7439.3EEQFNSTYR +2GlcNAcIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b,IgG6b,L8B0S7,L8B0S2,K7ZLA7
1642.8511643.859−0.921827.2614.7AGGTTVTQGVETTKPSK[C-term] neutral loss
1958.9241959.932−1.073.42082.2TTPPQQDVDGTYFLYSKIgG4a-b, IgG6a, L8B180, L8B0S7, IgG2a, IgG2b, IgG3
2413.9602414.968−11.420.02793.9G0 of EAQFNSTYRIgG6a,L8B180,L8B0Z4
2558.9882559.995−19.690.025.6G0F EAQFNSTYRIgG6a,L8B180,L8B0Z4
2617.0362618.044−2.663.51257.9G0F EEQFNSTYRIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b,IgG6b,L8B0S7,L8B0S2,K7ZLA7
2779.0922780.100−1.360.01327.0G1F EEQFNSTYRIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b,IgG6b,L8B0S7,L8B0S2,K7ZLA7
2941.1662942.1745.900.0266.4G2F EEQFNSTYRIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b,IgG6b,L8B0S7,L8B0S2,K7ZLA7
3086.1823087.190−1.330.0159.4G1FS EEQFNSTYRIgG1a-b,IgG2a-b,IgG4a-b,IgG5a-b,IgG6b,L8B0S7,L8B0S2,K7ZLA7

Table 4.

Tryptic peptides from pig IgG Fab and Fc fragment analyzed by HPLC-MS/MS.

L8B180,L8B0Z4,L8B0S2,L8B0S7,K7ZLA7, porcine IgG heavy chain entries (Fc) [22]; P01846, porcine IgG lambda constant domain (Fab); K7ZJP7, IgM heavy chain; P01786, mouse IgG heavy chain; P01857, human IgG heavy chain; IgGs, from [3]; IPR, GPM entries.

As Fabulous™ was possibly still present in the mixture at the time of tryptic digestion of the Fc gel band, analogous papain tryptic digestion products were sought for by m/z, but not found. Only one papain peptide with sequence YIDETNK (m/z 882.4203) could be present according to its nominal mass; however, the observed m/z 882.4998 peptide ions are closer to the calculated mass of IgG DLPAPITR, 882.4939. As seen in Table 4, many peptides remain un-sequenced/unidentified, and others are identified as either Fab or Fc peptides of porcine IgG. Interestingly, there were some very good matches with human and mouse IgG, which is not surprising given relatively high levels of homology between mammal species in general. However even with all the heavy chain sequences available in the literature for porcine IgG (see footnote of Table 4 and Figure A1), none of them contained the human- and murine-assigned peptides. It is also interesting to observe that most peptides are predominant in either the Fab or Fc injections and that overlapping peptides are clearly more present in one sample than in the other. Accordingly, all Fab P01846 peptides (m/z 880, 1175, 1274, 1375, 1533, 1590, and 1661) were more abundant in the Fab than the Fc sample, although all these peptides showed some level of overlap. The same reasoning is applicable to Fc peptides, including the N-glycoforms of EAQFNSTYR and EEQFNSTYR. It is also possible to know the origin of most unidentified peptides in Table 4 by comparing their normalized abundances. Other N-glycopeptides reported in Table 1 were not detected in these Fab and Fc RPLC/MS experiments, possibly due to the lack of glycopeptide enrichment and to sample loss during the SEC and SDS-PAGE procedures.

3.7 Comparative aspects of both workflows

The initial RPLC-MS/MS experiment on the enriched whole IgG digest produced the most useful information in terms of glycopeptides. Besides the latter, many peptides were identified as originating from the Fc or Fab portion, while for other peptides origin remained unknown. This first workflow allowed detecting N-glycopeptides from the Fab portion of the antibody, although their sequences could not be completely assigned. The second workflow necessitated more intensive work than the first one. Size-exclusion chromatography or SDS-PAGE was not sufficient on its own to properly separate Fab and Fc, and a combination of both was preferable. Even then, RPLC/MS results show partial overlaps for many peptides, although most peptides were much more abundant in one sample (Fc or Fab) than in the other. In terms of ionization modes used in the second workflow, ESI yielded more details on sialylated glycopeptides, but did not allow the observation of Fab N-glycopeptides. As analyses by MALDI were the result of depositing all tryptic products of a gel band onto one target spot, it is probable that the signals of sialylated glycopeptides or low concentration Fab glycopeptides were overtaken competitively by peptides that are more abundant.



The authors thank Dr. Jean-Paul Soulillou and his team at the Université de Nantes, France, for providing pig IgG samples. Thanks to Genovis for donating Fabulous™ enzyme.


Conflict of interest

None of the authors has a conflict of interest.


Figure A1.

Alignment of sequences available in the literature for the porcine IgG heavy chains. IgGn: from Ref. [3]. Others: UniprotKB [22].

Figure A2.

HPLC/MS total ion chromatograms obtained for the tryptic products of wild-type porcine IgG Fabulous™ fragments, (a) Fc and (b) Fab.


4. Conclusions

Porcine immunoglobulins constitute a complex ensemble of biomolecules, with several subtypes whose amino acid sequences are not clearly assigned and described in the literature (see Figure A1). In this study, three state-of-the-art mass spectrometers were used to characterize tryptic peptides, offering a considerable amount of complementary information owing to the great sensitivities of these instruments, given the small amounts of IgGs used. Indeed, fragmentation of porcine IgG into its Fc and Fab portions was achieved for the first time using papain and Fabulous™ on 200 μg or less of antibody. Fragments needed separation by both SEC and SDS-PAGE before analysis by MALDI-TOF-MS and HPLC/MS. These separation procedures did not eliminate the overlapping of Fab and Fc tryptic peptides entirely; however, there was a significant level of discrimination. This workflow resulted in better knowledge about the origin of tryptic peptides from IgG.

During the course of this work, the several sources of sequence information on porcine IgG found in Figure A1 were used to verify HPLC-MS/MS, MALDI-TOF-MS, and HPLC-MS data. Most Fc peptides were from the gamma constant region, where some Fab peptides were identified as belonging to the constant portion of the lambda chain. The HPLC-MS/MS method of tryptic peptide without previous Fab-Fc fragmentation was the most efficient in terms of useful data generated per amount of sample used, although many peptides could not be related to either Fc or Fab. This study highlights the need for detailed pertinent sequence information for porcine IgG, which is not a commonly studied set of biomolecules. Future work will involve the quantification of IgG subtypes according to unique peptide sequences that are already known in each subtype.


  1. 1. Burlak C et al. N-linked glycan profiling of GGTA1/CMAH knockout pigs identifies new potential carbohydrate xenoantigens. Xenotransplantation. 2013;20(5):277-291
  2. 2. Buist M et al. Features of N-glycosylation of immunoglobulins from knockout pig models. Journal of Analytical and Bioanalytical Techniques. 2016;7:333-342
  3. 3. Lopez PG et al. Characterization of N-glycosylation and amino acid sequence features of immunoglobulins from swine. Glycoconjugate Journal. 2016;33(1):79-91
  4. 4. Reynard O et al. Anti-EBOV GP IgGs lacking α1-3-Galactose and Neu5Gc prolong survival and decrease blood viral load in EBOV-infected Guinea pigs. PLoS One. 2016;11:e0156775. Accepted, in revision. (Accepted, in revision)
  5. 5. Butler JE et al. Porcine IgG: Structure, genetics, and evolution. Immunogenetics. 2009;61(3):209-230
  6. 6. Vidarsson G, Dekkers G, Rispens T. IgG subclasses and allotypes: From structure to effector functions. Frontiers in Immunology. 2014;5:520
  7. 7. Dwek RA, Lellouch AC, Wormald MR. Glycobiology: 'The function of sugar in the IgG molecule'. Journal of Anatomy. 1995;187(Pt 2):279-292
  8. 8. Wormald MR et al. Variations in oligosaccharide-protein interactions in immunoglobulin G determine the site-specific glycosylation profiles and modulate the dynamic motion of the Fc oligosaccharides. Biochemistry. 1997;36(6):1370-1380
  9. 9. Andrew SM, Titus JA. Fragmentation of immunoglobulin G. Current Protocols in Immunology. 2001;Chapter 2:Unit 2.8
  10. 10. Andrew SM, Titus JA. Fragmentation of immunoglobulin G. Current Protocols in Cell Biology. 2003;Chapter 16:Unit 16.4
  11. 11. Application Site for Fabulous™ Enzyme, on Genovis Inc. Web Site [Internet]. Available from: [Accession Date: 20-08-2018]
  12. 12. Demignot S, Garnett MC, Baldwin RW. Mouse IgG2b monoclonal antibody fragmentation. Preparation and purification of Fab, Fc and Fab/c fragments. Journal of Immunological Methods. 1989;121(2):209-217
  13. 13. Adamczyk M, Gebler JC, Wu J. Papain digestion of different mouse IgG subclasses as studied by electrospray mass spectrometry. Journal of Immunological Methods. 2000;237(1-2):95-104
  14. 14. Suzuki N, Lee YC. Site-specific N-glycosylation of chicken serum IgG. Glycobiology. 2004;14(3):275-292
  15. 15. Antibody Fragmentation Methods Explained on the Thermo-Fisher Inc. Web Site [Internet]. Available from: [Accession Date: 10-08-2018
  16. 16. Al-Abdulla I, Casewell NR, Landon J. Single-reagent one-step procedures for the purification of ovine IgG, F(ab')2 and Fab antivenoms by caprylic acid. Journal of Immunological Methods. 2014;402(1-2):15-22
  17. 17. Yu D, Ghosh R. Membrane bioreactor separator system for integrated IgG fragmentation and Fab purification. Journal of Immunological Methods. 2010;359(1-2):37-41
  18. 18. Yu D, Ghosh R. Integrated fragmentation of human IgG and purification of Fab using a reactant adsorptive membrane bioreactor separator system. Biotechnology and Bioengineering. 2009;104(1):152-161
  19. 19. Bondt A et al. Immunoglobulin G (IgG) Fab glycosylation analysis using a new mass spectrometric high-throughput profiling method reveals pregnancy-associated changes. Molecular & Cellular Proteomics. 2014;13(11):3029-3039
  20. 20. Hafkenscheid L et al. Structural analysis of variable domain glycosylation of anti-citrullinated protein antibodies in rheumatoid arthritis reveals the presence of highly sialylated glycans. Molecular & Cellular Proteomics. 2017;16(2):278-287
  21. 21. Schmelter C et al. Peptides of the variable IgG domain as potential biomarker candidates in primary open-angle glaucoma (POAG). Human Molecular Genetics. 2017;26(22):4451-4464
  22. 22. Protein Database, Uniprotkb [Internet]. Available from: [Accession date: 01-11-2018]
  23. 23. Product Application Site for Glycopeptide Enrichment Kits, on Millipore Inc. Web Site [Internet]. Available from: [Accession Date: 15-07-2018]
  24. 24. The Global Proteome Machine Proteomics Data Analysis, Reuse and Validation for Biological and Biomedical Research [Internet]. Available from: [Accession Date: 21-11-2018]
  25. 25. Pompach P et al. Semi-automated identification of N-Glycopeptides by hydrophilic interaction chromatography, nano-reverse-phase LC-MS/MS, and glycan database search. Journal of Proteome Research. 2012;11(3):1728-1740
  26. 26. Novak J et al. CycloBranch: De novo sequencing of nonribosomal peptides from accurate product ion mass spectra. Journal of the American Society for Mass Spectrometry. 2015;26(10):1780-1786
  27. 27. Krokhin O et al. Site-specific N-glycosylation analysis: Matrix-assisted laser desorption/ionization quadrupole-quadrupole time-of-flight tandem mass spectral signatures for recognition and identification of glycopeptides. Rapid Communications in Mass Spectrometry. 2004;18(18):2020-2030
  28. 28. Komatsu E et al. Characterization of immunoglobulins through analysis of N-glycopeptides by MALDI-TOF MS. Methods. 2016;104:170-181
  29. 29. Wenig K et al. Structure of the streptococcal endopeptidase IdeS, a cysteine proteinase with strict specificity for IgG. Proceedings of the National Academy of Sciences of the United States of America. 2004;101(50):17371-17376
  30. 30. Bodnar E et al. An integrated approach to analyze EG2-hFc monoclonal antibody N-glycosylation by MALDI-MS. Canadian Journal of Chemistry. 2015;93(7):754-763
  31. 31. Specific Accession Number for Papain on Uniprotkb Protein Database [Internet]. Available from: [Accession Date: 10-08-2018]
  32. 32. Janeway CA, Travers P, Walport M. The structure of a typical antibody molecule. In: Immunobiology: The Immune System in Health and Disease. 5th ed. New York: Garland Science; 2001

Written By

Claudia Nelson, Raymond Bacala, Baylie Gigolyk, Evelyn Ang, Haley Neustaeter, Emy Komatsu, Oleg Krokhin, Dave Hatcher and Hélène Perreault

Submitted: 15 August 2018 Reviewed: 29 November 2018 Published: 08 January 2019