Anal. Chem. 2010, 82, 6751–6755
Resolving Disulfide Structural Isoforms of IgG2 Monoclonal Antibodies by Ion Mobility Mass Spectrometry Dhanashri Bagal,† John F. Valliere-Douglass,‡ Alain Balland,*,‡ and Paul D. Schnier*,† Molecular Structure, Amgen, Thousand Oaks, California 91320, and Process and Product Development, Amgen, Seattle, Washington 98119 Recombinant monoclonal antibodies are an important class of therapeutic agents that have found widespread use for the treatment of many human diseases. Here, we have examined the utility of ion mobility mass spectrometry (IMMS) for the rapid characterization of disulfide variants in intact IgG2 monoclonal antibodies. It is shown that IMMS reveals 2 to 3 gas-phase conformer populations for IgG2s. In contrast, a single gas-phase conformer is revealed using IMMS for both an IgG1 antibody and a Cys232 f Ser mutant IgG2, both of which are homogeneous with respect to disulfide bonding. This provides strong evidence that the observed IgG2 gas-phase conformers are related to disulfide bond heterogeneity. Additionally, IMMS analysis of redox enriched disulfide isoforms allows assignment of the mobility peaks to established disulfide bonding patterns. These data clearly illustrate how IMMS can be used to quickly provide information on the higher order structure of antibody therapeutics. The overall structure of the immunoglobulin G (IgG) family is organized in 12 subdomains, each closed by an intrachain disulfide bond.1 Heavy chain (HC) and light chain (LC) are connected by interchain disulfide bonds to form a covalent complex of the form (HC-LC)2. IgG1, IgG2, and IgG4 isotypes share greater than 90% sequence homology in their constant domains but differ significantly in the hinge region. IgG1 and IgG4 hinge core sequences are very similar with two cysteines on each heavy chain involved in interheavy chain connection, whereas IgG2 is unique in presenting four cysteine residues in the hinge region, notably two consecutive residues, Cys-232 and Cys-233 (amino acid numbering of Kabat et al.),2 that have no equivalent in any other immunoglobulin subclass. Researchers at Amgen recently reported that these residues confer distinctive structural features to the human IgG2 isotype resulting in the formation of disulfide-related structural isoforms.3,4 Three distinct structural isoforms (IgG2-A, IgG2-B, and IgG2-A/B)3,4 specific to human IgG2s were revealed by chro* To whom correspondence should be addressed. E-mail:
[email protected] (A.B.);
[email protected] (P.D.S.). † Molecular Structure. ‡ Process and Product Development. (1) Padlan, E. A. Adv. Protein Chem. 1996, 49, 57–133. (2) Kabat, E. A.; Wu, T. T.; Perry, H. M.; Gottesman, K. S.; Foeller, C. Sequences of Proteins of Immunological Interest, 5th ed.; U.S. Public Health Service, NIH: Washington, DC., 1991. 10.1021/ac1013139 2010 American Chemical Society Published on Web 07/21/2010
matographic and electrophoretic methods including capillary electrophoresis with sodium dodecyl sulfate (CE-SDS),3,5 reversedphase high performance liquid chromatography (RP-HPLC),4 and cation exchange chromatography (CEX).3,6 IgG2-A corresponds to the classical model with independent Fab and Fc domains connected by the hinge (Figure 1a).7 IgG2-B is a symmetrical form with HC and LC covalently linked to the hinge by disulfide bridges (Figure 1b). IgG2-A/B is an asymmetrical form intermediate between A and B. Detailed analysis of each IgG2 kappa structural isoforms showed that the different interchain disulfide bond arrangements involved only four residues: the cysteine in constant region one of the heavy chain (CH1), Cys-127 (Kabat numbering), the cysteine at the C-terminus of the light chain, Cys-214, and two cysteines in the upper hinge region, specific to the IgG2 subclass, Cys-232 and Cys-233. The precise cysteine connectivity of each structural form was established by partial reductionalkylation and mass spectrometry (MS)8 and Edman sequencing MS.9 Modeling of the IgG2 sequence based on the three-dimensional antibody structure places the four cysteines in close spatial proximity, supporting the concept that a variable arrangement of these residues could generate IgG2 structural isoforms. Two specific cysteine-to-serine mutants were designed at positions 232 and 233 to disrupt potential disulfide rearrangements.10 These mutants both exhibited no significant difference in expression and potency characteristics when compared to wild type IgG2 but proved structurally homogeneous with respect to the disulfide bonding of the IgG2-A type.10 (3) Wypych, J.; Li, M.; Guo, A.; Zhang, Z.; Martinez, T.; Allen, M. J.; Fodor, S.; Kelner, D. N.; Flynn, G. C.; Liu, Y. D.; Bondarenko, P. V.; Ricci, M. S.; Dillon, T. M.; Balland, A. J. Biol. Chem. 2008, 283, 16194–16205. (4) Dillon, T. M.; Ricci, M. S.; Vezina, C.; Flynn, G. C.; Liu, Y. D.; Rehder, D. S.; Plant, M.; Henkle, B.; Li, Y.; Deechongkit, S.; Varnum, B.; Wypych, J.; Balland, A.; Bondarenko, P. V. J. Biol. Chem. 2008, 283, 16206–16215. (5) Guo, A.; Han, M.; Martinez, T.; Ketchem, R. R.; Novick, S.; Jochheim, C.; Balland, A. Electrophoresis 2008, 29, 2550–2556. (6) Zhang, Y.; G. A.; Novick, S.; Jochheim, C.; Boyce, J. M.; Gerhart, M.; Qin, X.; Gombotz, W. I. Bioprocessing J. 2003, (Nov-Dec), 37–43. (7) Milstein, C.; Frangione, B. Biochem. J. 1971, 121, 217–225. (8) Martinez, T.; Guo, A.; Allen, M. J.; Han, M.; Pace, D.; Jones, J.; Gillespie, R.; Ketchem, R. R.; Zhang, Y.; Balland, A. Biochemistry 2008, 47, 7496– 7508. (9) Zhang, B.; Harder, A. G.; Connelly, H. M.; Maheu, L. L.; Cockrill, S. L. Anal. Chem. 2009, 82, 1090–1099. (10) Allen, M. J.; Guo, A.; Martinez, T.; Han, M.; Flynn, G. C.; Wypych, J.; Liu, Y. D.; Shen, W. D.; Dillon, T. M.; Vezina, C.; Balland, A. Biochemistry 2009, 48, 3755–3766.
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010
6751
structure with a high degree of detail. Although MS is not routinely used to characterize higher order structural elements in IgGs, MS coupled with hydrogen/deuterium exchange was recently demonstrated as a method to characterize the conformational dynamics of IgG1 antibodies in solution.17 Ion mobility mass spectrometry (IMMS) has shown great promise as an intact protein separation and analysis methodology to probe higher order structural elements including the overall size/shapeofbiopolymersandlargemacromolecularassemblies.18-28 Recently, Waters Corporation commercialized an ion mobility mass spectrometer (Synapt) based on traveling waves (T-Wave).29 In the T-Wave implementation of ion mobility, ion separation occurs when a sequence of dc pulses push ions through the mobility cell in the presence of an inert gas at relatively high pressure.29,30 The ability of an ion to “surf” the T-wave depends on its collision cross section (CCS). Ions with compact structures are pushed through the mobility cell faster than ions with more elongated structures. In this work, we present evidence that T-Wave IMMS can be used to separate disulfide variants of intact IgG2 antibodies. Attractive features of the method include high sensitivity (µg sample consumption), minimal sample preparation, and fast analysis time (minutes). EXPERIMENTAL SECTION Human monoclonal antibodies mAb#1 (IgG2), mAb#2 (IgG1), and mAb#3 (IgG2) were produced recombinantly in Chinese hamster ovary (CHO) cells and purified at Amgen. All additional reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise specified. For control experiments, constant region two (CH2) domain N-glycans were removed by adding 1500 U of PNGase F (New England Biolabs, Ipswich, MA) per 100 µg of protein and incubating at 37 °C for 16 h. Disulfide isoforms IgG2-A and IgG2-B were selectively enriched using methods established by Dillon et al.4 Briefly, to enrich isoform B, IgG2s were incubated in 200 mM Tris buffer
Figure 1. Illustration of the hinge region disulfide bonding pattern of human (a) IgG2-A and (b) IgG2-B antibodies.
The ability to rapidly detect and characterize IgG2 isoforms is of great interest, as it may help to facilitate the transition of new IgG2 molecules from discovery into development and ultimately commercialization. In recent years, mass spectrometry has played an increasingly important role in the analytical characterization of IgG therapeutics. Mass spectrometry is now widely used to confirm the intact molecular weight of IgGs,11,12 establish their glycosylation profile,13,14 and confirm15 or establish16 the primary (11) Gadgil, H. S.; Pipes, G. D.; Dillon, T. M.; Treuheit, M. J.; Bondarenko, P. V. J. Am. Soc. Mass Spectrom. 2006, 17, 867–872. (12) Brady, L. J.; Valliere-Douglass, J.; Martinez, T.; Balland, A. J. Am. Soc. Mass Spectrom. 2008, 19, 502–509. (13) Damen, C. W. N.; Chen, W.; Chakraborty, A. B.; van Oosterhout, M.; Mazzeo, J. R.; Gebler, J. C.; Schellens, J. H. M.; Rosing, H.; Beijnen, J. H. J. Am. Soc. Mass Spectrom. 2009, 20, 2021–2033. (14) Olivova, P.; Chen, W.; Chakraborty, A. B.; Gebler, J. C. Rapid Commun. Mass Spectrom. 2008, 22, 29–40. (15) Ren, D.; Pipes, G. D.; Hambly, D.; Bondarenko, P. V.; Treuheit, M. J.; Gadgil, H. S. Anal. Biochem. 2009, 384, 42–48. (16) Bandeira, N.; Pham, V.; Pevzner, P.; Arnott, D.; Lill, J. R. Nat. Biotechnol. 2008, 26, 1336–1338.
6752
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010
(17) Houde, D.; Arndt, J.; Domeier, W.; Berkowitz, S.; Engen, J. R. Anal. Chem. 2009, 81, 2644–2651. (18) Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F. J. Am. Chem. Soc. 1995, 117, 10141–10142. (19) Bohrer, B. C.; Merenbloom, S. I.; Koeniger, S. L.; Hilderbrand, A. E.; Clemmer, D. E. Annu. Rev. Anal. Chem. 2008, 1, 293–327. (20) Ruotolo, B. T.; Giles, K.; Campuzano, I.; Sandercock, A. M.; Bateman, R. H.; Robinson, C. V. Science 2005, 310, 1658–1661. (21) Kaddis, C. S.; Loo, J. A. Anal. Chem. 2007, 79, 1778–1784. (22) Leary, J. A.; Schenauer, M. R.; Stefanescu, R.; Andaya, A.; Ruotolo, B. T.; Robinson, C. V.; Thalassinos, K.; Scrivens, J. H.; Sokabe, M.; Hershey, J. W. B. J. Am. Soc. Mass Spectrom. 2009, 20, 1699–1706. (23) Kim, H. I.; Kim, H.; Pang, E. S.; Ryu, E. K.; Beegle, L. W.; Loo, J. A.; Goddard, W. A.; Kanik, I. Anal. Chem. 2009, 81, 8289–8297. (24) Atmanene, C. D.; Wagner-Rousset, E.; Malissard, M.; Chol, B.; Robert, A.; Corvaı¨a, N.; Dorsselaer, A. V.; Beck, A.; Sanglier-Cianferani, S. Anal. Chem. 2009, 81, 6364–6373. (25) Ruotolo, B. T.; Hyung, S.-J.; Robinson, P. M.; Giles, K.; Bateman, R. H.; Robinson, C. V. Angew. Chem., Int. Ed. 2007, 46, 8001–8004. (26) Hilton, G. R.; Thalassinos, K.; Grabenauer, M.; Sanghera, N.; Slade, S. E.; Wyttenbach, T.; Robinson, P. J.; Pinheiro, T. J. T.; Bowers, M. T.; Scrivens, J. H. J. Am. Soc. Mass Spectrom. 2010, 21, 845–854. (27) Schenauer, M. R.; Meissen, J. K.; Seo, Y.; Ames, J. B.; Leary, J. A. Anal. Chem. 2009, 81, 10179–10185. (28) Thalassinos, K.; Grabenauer, M.; Slade, S. E.; Hilton, G. R.; Bowers, M. T.; Scrivens, J. H. Anal. Chem. 2008, 81, 248–254. (29) Pringle, S. D.; Giles, K.; Wildgoose, J. L.; Williams, J. P.; Slade, S. E.; Thalassinos, K.; Bateman, R. H.; Bowers, M. T.; Scrivens, J. H. Int. J. Mass Spectrom. 2007, 261, 1–12. (30) Shvartsburg, A. A.; Smith, R. D. Anal. Chem. 2008, 80, 9689–9699.
(pH 8.0) with cysteine and cystamine at concentrations of 6 and 1 mM, respectively. Isoform-A was enriched by incubating the IgG2 under the same conditions but with the addition of 1.0 M guanidinium chloride (GuHCl) to the buffer. The samples were protected from light at 2-8 °C for approximately 48 h. The IgG2 mAb#3 was subjected to Cysf Ser mutagenesis at position 232 as described by Allen et al.10 Site-directed mutagenesis was performed using the QuickChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). The mutant was stably transfected into a serum-free suspension-adapted CHO cell line.31 Following production, the mAb was purified by protein A affinity chromatography. Incorporation of the expected mutation was verified on the purified molecule by size exclusion chromatography (SEC) MS.12 For IMMS analysis, IgG samples were buffer exchanged and concentrated using Vivaspin 30 kDa molecular weight cutoff filters (GE Healthcare, Buckinghamshire, UK). For experiments performed under native conditions, the IgGs were diluted using 160 mM ammonium acetate (pH not adjusted) to a final working concentration of ∼3 µM. Mass spectrometry experiments were performed using a hybrid ion mobility quadrupole time-of-flight MS (Synapt, Waters Inc., Milford, MA) equipped with a nanoelectrospray ionization (ESI) source using metal coated borosilicate glass capillaries (nanoflow probe tips, long thin walled, Waters Corporation). Solution flow rates of ∼75 nL/min and an ESI capillary voltage of ∼1.3 kV were used for all experiments. The source temperature was 50 °C, and the pressure of the vacuum/backing region was 3.5 mbar. Each ion mobility mass spectrum was acquired from m/z 4000-8000 every 2 s; approximately 12 counts per scan were observed. The signal was typically averaged for approximately 10 min. Gentle source conditions were used to minimize gas phase unfolding of the protein (sample cone: 40 V, trap voltage: 10 V, transfer lens: 12 V, bias: 25; cone gas: 30 L/Hr; PTRAP(Ar): 0.0175 mbar). Nitrogen was used as the mobility carrier gas, and the following parameters were found to give optimal ion mobility separation (PIMS: 0.5 mbar, wave velocity: 300 m/s, wave height: 9.8 V). The instrument was mass calibrated using a 50 µg/µL CsI solution. Waters’ raw data files were translated to Matlab binary files using software developed in-house. Data processing and plotting were performed in Matlab (The MathWorks Inc. Natick, MA) and Igor Pro (WaveMetrics Inc., Lake Oswego, OR).
Figure 2. Ion mobility TOF mass spectra of (a) mAb#1 and (c) mAb#2. The normalized ion mobility intensities are graphed as contour plots with 9 levels from 4% to 100% intensity. Extracted arrival time distributions for the 26+ charge states of (b) mAb#1 and (d) mAb#2.
RESULTS AND DISCUSSION Antibodies belonging to the IgG2 subclass exist as a group of distinct isoforms with different disulfide connectivities between the Fab domain and the hinge region of the molecule. The goal of this study was to investigate the ability of IMMS to successfully separate these discrete IgG2 isoforms. A nano-ESI ion mobility mass spectrum for a solution of 3 µM mAb#1 (theoretical MW for the most abundant glycoform (G1F/ G1F): 149821.50 Da) in 160 mM ammonium acetate is shown Figure 2a. A narrow charge state distribution centered on the 24+ ion is observed. Ion mobility spectra for all MAbs analyzed were acquired with minimal acceleration voltages (sample cone: 40 V, trap: 10 V, transfer: 12 V). These gentle tuning conditions were found to provide optimal resolution in the ion mobility dimension
(vide infra). However, with these conditions, the mass spectral peaks are relatively broad making it impossible to observe the individual glycoforms of the IgG. The glycoforms for these IgG molecules can, however, be readily resolved using higher source voltages (data not shown). Robinson and co-workers have previously observed that optimal conditions for the mass and mobility dimensions are often not compatible on the Synapt instrument when analyzing large protein assemblies.32 Ion mobility separation of mAb#1 (Figure 2a) reveals two distinct conformer populations. For example, the mobility spectrum for the 26+ charge state (Figure 2b) shows two distinct peaks with a 1.1 ms difference in drift times (9.3 and 10.4 ms, respectively). Similar distributions are observed for each charge state. However, with the tuning conditions employed here, the best mobility resolution was achieved for the 27+, 26+, and 25+ charge states. To elucidate if the heterogeneity of the sugar chains on the antibody influences the observed gas-phase conformers, deglycosylated mAb#1 was analyzed. The same distribution of gas-phase conformers is revealed in the ion mobility mass spectrum of mAb#1 with the sugars released (Figure S1a, Supporting Information). This clearly demonstrates that these multiple conformers are not caused by sugar heterogeneity but instead are related to conformational differences in the protein chains. The arrival time distributions for each charge state virtually overlaps with that of the glycosylated protein. For example, the major peaks in the arrival time distribution plot of the 26+ charge state of deglycosylated mAb#1 are only shifted by -0.13 ms compared to
(31) Rasmussen, B.; Davis, R.; Thomas, J.; Reddy, P. Cytotechnology 1998, 28, 31–42.
(32) Ruotolo, B. T.; Benesch, J. L. P.; Sandercock, A. M.; Hyung, S.-J.; Robinson, C. V. Nat. Protoc. 2008, 3, 1139–1152.
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010
6753
that of the glycosylated IgG (Figure S1b, Supporting Information). This indicates that the glycosylated and deglycosylated antibodies have nearly identical gas-phase collision cross sections. The relative abundance of each gas-phase conformer can be roughly estimated by fitting the arrival time distribution to a sum of log-normal functions and integrating the relative area of each peak. The log-normal function is an asymmetric Gaussian function whose logarithm is normally distributed.33 The log-normal function was used because it was empirically found to best fit the experimental peak shape of this ion mobility data. For the 26+ charge state, the normalized area of peaks 1 and 2 from the best fits are 42% and 58%, respectively. Similar areas are observed for other charge states. These abundances are similar to the relative peak areas observed in electropherograms of IgG2s separated with capillary electrophoresis. We and others have shown CE-SDS to be a resolving technique for the separation of IgG2 structural isoforms.3,5,35 However, this correlation alone does not signify that the resolved gas-phase conformers are definitively due to disulfide variants. To elucidate whether the gas-phase conformers observed for IgG2 molecules are related to their disulfide connectivity, we analyzed an IgG1 antibody, mAb#2 (theoretical MW for the most abundant glycoform (G1F/G0F):148408.0 Da), as a control (Figure 2c). The most significant difference between human IgG1 and IgG2 subclasses is the primary structure of the hinge region, resulting in the absence of disulfide related isoforms in the IgG1.3,4 In contrast to the IgG2 mobility data, for each charge state of mAb#2, the arrival time profile is relatively narrow and consists of a single uniform distribution (Figure 2c). For example, the arrival time profile for the 26+ charge state of mAb#2 (Figure 2d) shows a single peak at 8.7 ms. This arrival time profile is representative of the distribution observed for each of the charge states. This suggests that the multiple conformers observed of mAb#1 are due to disulfide variants in the antibody. To further demonstrate that the observed gas-phase conformers are indeed IgG2 disulfide variants, individual disulfide isoforms were selectively enriched in a refolding experiment using redox chemistry employing cysteine/cystamine. Dillon et al. previously demonstrated that isoform IgG2-A and IgG2-B can be redoxenriched by refolding with and without 1 M GuHCl, respectively.4 Under redox conditions in buffer alone, IgG2-A is refolded to IgG2B. Liu et al. demonstrated that a slow conversion of IgG2-A to IgG2-B also occurs in vivo.34Isoform conversion toward IgG2-A requires in vitro refolding in presence of low levels of chaotropic reagents.4 Figure 3a,b shows the arrival time distributions for the +26 charge state for redox enriched mAb#1 in the presence of guanidine (isoform A) and redox enriched mAb#1 in the absence of guanidine (isoform B), respectively. In contrast to IMMS for the untreated Mab#1 antibody (Figure 3, dashed line), a single abundant conformer is observed for each of these enriched isoforms. As a control, the IgG1 antibody, mAb#2, was also subjected to the same redox refolding protocol with or without guanidine. All treated IgG1 samples have identical arrival time (33) Brown, R. Personal Eng. Instrum. News 1991, 8, 51–54. (34) Liu, Y. D.; Chen, X.; Enk, J. Z.; Plant, M.; Dillon, T. M.; Flynn, G. C. J. Biol. Chem. 2008, 283, 29266–29272. (35) Lacher, N. A.; Wang, Q.; Roberts, R. K.; Holovics, H. J.; Aykent, S.; Schlittler, M. R.; Thompson, M. R.; Demarest, C. W. Electrophoresis 2010, 31, 448– 458.
6754
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010
Figure 3. Arrival time distributions for the 26+ charge states of IgGs treated with redox reagents (cystamine, cysteine). (a,b) mAb#1 and (c,d) mAb#2 (control). Samples represented in (a) and (c) had 1 M GuHCl added to the buffer. The dashed line shows the arrival time distribution for the 26 + charge state of untreated mAb#1.
distribution profiles as the untreated IgG1 molecule (Figure 3c,d), demonstrating that this refolding protocol does not affect the overall tertiary structure of the antibody. Comparing the untreated IgG2 IMMS trace with the enriched isoform distributions (Figure 3a,b) identifies peaks 1 (9.3 ms) and 2 (10.4 ms) in the mobility spectra as isoform A and isoform B, respectively. The relatively late arrival time of isoform B indicates that this form of the antibody has a larger gas-phase collision cross section compared to isoform A. The IMMS resolution of the isoforms correlates with the previously observed capillary electrophoresis separation.5,8,35 With CE, IgG2 isoforms were resolved into two peaks, with IgG2-A migrating more rapidly than the IgG2-B isoform. The intermediate IgG2-A/B forms were split between the two peaks, IgG2-A/B1 migrating with -A and IgG2-A/B2 with -B.8 While the disulfide connectivities of IgG2 isoforms have been well characterized,3,4,8,9 only limited information is available describing the overall tertiary structure of human IgGs. To investigate if the relative ordering of gas-phase collision cross sections (IgG1 ≈ IgG2-A < IgG2-B) correlates with IgG solution structures, collision cross sections were calculated for two IgG structures using Waters’ CCS software.36 The calculated cross section of an IgG antibody, with a disulfide bonding pattern consistent with the B isoform, is 8385 Å2 (unpublished results). This value is 3% smaller than the calculated cross section of an IgG1 antibody (protein data bank code 1HZH, 8653 Å2).37 (36) Williams, J. P.; Lough, J. A.; Campuzano, I.; Richardson, K.; Sadle, P. J. Rapid Commun. Mass Spectrom. 2009, 23, 3563. (37) Saphire, E. O.; Parren, P. W.; Pantophlet, R.; Zwick, M. B.; Morris, G. M.; Rudd, P. M.; Dwek, R. A.; Stanfield, R. L.; Burton, D. R.; Wilson, I. A. Science 2001, 293, 1155–1159.
Figure 4. Ion mobility mass spectra of (a) mAb#3 and (b) Cys f Ser Mab#3 mutant. The extracted arrival time distributions for the 25+ charge states of these molecules are shown in (c) and (d), respectively.
These limited calculations using individual static structures do not explain the relatively late arrival time of the IgG2-B isoform and suggest that the overall three-dimensional structure of these gas-phase ions may be quite different than the solution structure. While the IMMS results demonstrate that ion mobility separates covalent disulfide mediated IgG2 structural isoforms, the overall three-dimensional structure of these gasphase ions is presently not clear. To investigate the generality of these ion mobility results, measurements were also performed on a different IgG2 antibody (mAb#3) which also exists as an ensemble of disulfide mediated isoforms (Figure 4a). Ion mobility separation of mAb#3 reveals two abundant gas-phase conformer populations and one minor conformer, for each charge state. These three populations are readily apparent in the arrival time distribution for the 25+ charge state (Figure 4c) which shows three distinct peaks with drift times of 9.5, 10.3, and 11.9 ms. The relative abundance is determined by fitting the arrival time distribution to a sum of log-normal functions of each conformer and is roughly estimated to be 46%, 49%, and 5%. Figure 4b, shows the ion mobility mass spectrum for a mutant form of this antibody with a single point mutation introduced at amino acid 232 (Cys f Ser) by site-directed mutagenesis. A single narrow peak is observed for the arrival time distribution of each charge state. The arrival time profile of the 25+ charge state is shown in Figure 4d as an illustrative example demonstrating a homogeneous population, with a single peak at 9.3 ms. This peak corresponds to the gas-phase conformer with the more compact structure (isoform A, vide supra). This is consistent with previous studies demonstrating that this mutant is homogeneous with respect to disulfide bonding and of the IgG2-A type.10,38 This result also unambiguously confirms our assertion that the multiple IMMS peaks observed in this study of selected IgG2s can be correlated with the disulfide bonding patterns in these molecules. (38) Lightle, S.; Aykent, S.; Lacher, N.; Mitaksov, V.; Wells, K.; Zobel, J.; Oliphant, T. Protein Sci. 2010, 19, 753–762.
CONCLUSIONS The results of the present study demonstrate that ion mobility as a shape-selective separation methodology can be used to detect disulfide heterogeneity in large (150 kDa) intact IgG2 antibodies. Two to three gas-phase conformers are observed by ion mobility for IgG2 antibodies. These gas-phase conformers were maintained with deglycosylated IgG2s. Analysis of redox refolded IgG2s as well as an IgG2 with a Cys f Ser single point mutation clearly demonstrates that the observed gas-phase conformers are related to disulfide variants. Ion mobility is fast (millisecond measurements), sensitive (nanomole), and amenable to high throughput automation. IMMS is a powerful new methodology for the characterization of intact antibodies and may be useful to routinely fingerprint higher order structure of these protein biopharmaceuticals in the near future. We are currently extending these measurements to investigate the utility of IMMS for the analysis of IgG2s containing lambda light chains as well as to directly characterize the binding of antigen targets to individual disulfide isoforms of IgG2 antibodies. ACKNOWLEDGMENT We are grateful to Allen Sickmier and Leszek Poppe for insightful discussions, Keith Richardson (Waters Inc.) for providing the Waters’ CCS software, Mike Berke, Rick Stanton, and Mikhail Toupikov for help with the data analysis software, and Mike Treuheit, Dean Pettit, Peter Grandsard, and Philip Tagari for championing and supporting this work. We also thank our Amgen colleagues, whose names are listed in the references, for their expert contributions to the collective knowledge built recently on IgG2 isoforms. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review May 19, 2010. Accepted July 6, 2010. AC1013139
Analytical Chemistry, Vol. 82, No. 16, August 15, 2010
6755
