Article pubs.acs.org/ac
Identification and Characterization of Buried Unpaired Cysteines in a Recombinant Monoclonal IgG1 Antibody Taylor Zhang,*,† Jennifer Zhang,† Daniel Hewitt,† Ben Tran,† Xiaoying Gao,‡ Zhihua Julia Qiu,‡ Max Tejada,§ Helene Gazzano-Santoro,§ and Yung-Hsiang Kao*,† †
Protein Analytical Chemistry, ‡BioAnalytical Sciences, and §Biological Technologies, Genentech, California 94080, United States S Supporting Information *
ABSTRACT: The heterogeneity in therapeutic antibodies arising from buried unpaired cysteines has not been well studied. This paper describes the characterization of two unpaired cysteines in a recombinant humanized IgG1 monoclonal antibody (referred to as mAb A). The reversed-phase high-performance liquid chromatography (RP-HPLC) analysis of mAb A samples showed three distinct peaks, indicating the presence of three species. The heterogeneities observed in the RP-HPLC have been determined to arise from unpaired cysteines (Cys-22 and Cys-96) that are buried in the VH domain. The Fab containing free thiols (referred to as “free-thiol Fab”) and the Fab containing the disulfide (referred to as “intact Fab”) of mAb A were generated through limited Lys-C digestion and purified with an ion exchange chromatography method. The binding of free-thiol Fab and intact Fab to its antigen was measured in a cell-based binding assay and an enzyme linked immunosorbent assay. The unpaired cysteines in the Fab of mAb A were found to have no significant impact on the binding to its target. Consistent with these Fab binding data, the enriched intact mAb A containing free thiols was determined to be fully active in a potency assay. The data reported here demonstrate that the redox status of cysteines is potentially a major source of heterogeneity for an antibody.
R
omalizumab (an anti-IgE antibody) contains 0.9 mol of free thiols as measured by the Ellman assay.14 Using a fluorescent assay, Zhang and Czupryn quantified free thiols in various recombinant antibodies produced in CHO (Chinese hamster ovary) cells and found that the free thiol levels in antibodies are significantly higher upon denaturation. 13 These results demonstrated the importance of measuring free thiol levels in antibodies under denaturing conditions since most cysteine residues are buried in the interior of an antibody. Free thiols were also found to be present in human IgG1 of healthy donors.15 Furthermore, new techniques such as isotope16 and fluorescence17 labeling have been developed to pinpoint the exact locations of unpaired cysteines in IgG1. By a differential isotope labeling method, free thiols, in the range of 1−6%, were found at all cysteines involved in the intrachain disulfides across multiple antibodies, and the major variation in the free thiol levels among different antibodies was observed at the cysteines in the variable domain.16 Although the presence of free thiols in IgG1s has been known, there is no report on free thiol variants of full length antibodies in term of their biological functions. The presence of free thiols may result in structural perturbation and introduce changes in hydrophobicity or apparent surface charges in proteins. These changes may then
ecombinant monoclonal antibodies are an important class of biological drugs and have been well characterized.1,2 One key feature in antibody domain architecture is the pairing of cysteines and the interchain and intrachain disulfide linkages between these specifically paired cysteines. The disulfide bonds are important post-translational modifications for antibody biological activity and stability.3−5 The identification and functional characterization of heterogeneity related to cysteines, such as redox status and linkage variations, have been challenging because disulfide reduction or shuffling will result in small or no change in the molecular weight of antibody. In addition, purification methods to enrich these cysteine related variants are often not available. Recently, disulfide bond isoforms of IgG2 with different hinge region cysteine pairings6−8 and antibodies containing trisulfide9 or thioether linkages10 have been reported. These papers, though not focusing on unpaired cysteines (commonly referred to as free thiols), highlighted the structural diversity involving cysteines in antibodies. Except in rare cases,11 most IgG1 molecules have 32 conserved Cys residues that form four interchain and twelve intrachain disulfide bonds. Although the cysteines and the disulfide linkages in an IgG1 antibody are highly conserved, unpaired cysteines have been observed at low levels for many antibodies. Traditionally, the free thiols in proteins have been measured by using the Ellman assay12 or fluorescent labeling assays.13 For example, it was found that every mole of © XXXX American Chemical Society
Received: May 24, 2012 Accepted: July 12, 2012
A
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
The 29-min method began with a two-minute isocratic hold at 36% mobile phase B, followed by an 18-min linear gradient to 45% mobile phase B. The column was washed at 95% mobile phase B for 1 min and equilibrated at 36% mobile phase B for 8 min. Two other columns, Agilent Zorbax StableBond C8 column and Varian PLRP-S column, were also employed using similar gradient methods. The fractions from individual peaks were collected and concentrated through Centricon (30 kD) filters (Millipore, Billerica, MA, USA). Cation-Exchange Chromatography (CEX) Conditions. CEX analysis was carried out on an Agilent 1200 HPLC system and monitored at 280 nm. Dionex ProPac WCX-10 columns (Sunnyvale, CA, USA) were used for analytical (4.0 × 250 mm with temperature controlled at 40 °C) and semipreparative collection (9.0 × 250 mm with temperature controlled at 50 °C). Mobile phase A consisted of 20 mM MES, pH 6.0, and mobile phase B consisted of 20 mM MES pH 6.0 with 0.25 M NaCl. Protein was eluted with a linear gradient from 25% to 55% mobile phase B in 40 min at a flow rate of 0.8 mL/min for analytical analysis and 35% to 45% mobile phase B in 35 min at a flow rate of 2.5 mL/min for semipreparative collection. The fractions corresponding to individual peaks were also collected and concentrated through Millipore Centricon (30 kD) filters for further characterization studies. Free-thiol Fab was purified using a semiprep column with a similar gradient method. Treatments with Oxidants. Samples of mAb A (2 mg/ mL) were prepared in 50 mM Tris buffer, pH 7.2. The samples were incubated with 1 mM oxidized glutathione (GSSG), 1 mM dehydroascorbic acid (DHAA), or 1 mM cystamine, respectively, for 24 h at room temperature or 37 °C. The samples were buffer exchanged into fresh buffer using Centricon (30 kD) filters. mAb A was also incubated with 0.1 mM CuSO4 for 24 h at room temperature, 2 mM EDTA was then added to remove free metal from the sample. The sample was then buffer exchanged into formulation buffer using a Centricon (30 kD) filter. Limited Endoproteinase Lys-C Digestion To Generate the Fab of mAb A. The Fab fragment of mAb A was generated through limited Lys-C digestion procedure similar to the published procedure.11 Briefly, mAb A (1 mg/mL) was mixed with Lys-C at an enzyme to protein ratio of 1:200 (w/w) in 100 mM Tris, pH 8.0, and then the mixture was incubated at 37 °C for 30 min. The reaction was stopped by lowering the pH to 6.0 by addition of 10% TFA. Free Thiol Measurement by Ellman Assay. Free thiol concentrations were determined using Ellman’s reagent. Protein samples were buffer exchanged into a reaction buffer (100 mM KH2PO4, 1 mM EDTA, 8 M urea, pH 8) and adjusted to a concentration that resulted in free thiol concentrations within the standard curve range. One 10 mM DTNB solution and eight levels of cysteine standard solutions between 0−100 μM were prepared in the reaction buffer. To a 96-well plate, 165 μL of sample or standard was added to triplicate wells. The reaction was initiated by the addition of 10 μL of DTNB and incubated for 30 min. After incubation, absorbance was measured at 412 nm using a SpectraMax M2 plate reader (Molecular Devices, Sunnyvale, CA, USA). The concentration of free thiol was calculated using the linear equation obtained from the standard curve. The concentration of the protein was determined using the absorbance at 280 nm obtained from a NanoDrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA). The free thiol was reported as mol free thiol per mol of protein.
be manifested by different chromatographic methods that can be used for the identification and characterization of proteins containing unpaired cysteines. For example, hydrophobic interaction chromatography (HIC) has been utilized to separate the Fabs with and without free thiols14 and has been used to follow the thiol reoxidation upon supplementing CuSO4 to the cell culture medium.18 Additionally, RP-HPLC has been a powerful technique for monitoring disulfide isomers or free thiol variants of proteins produced by E.coli, an expression system in which proteins are often produced as inclusion bodies in the reducing cytoplasmic environment. The recovery of proteins from inclusion bodies usually involves resolubilizing, reoxidizing, and refolding the proteins in vitro.19 Further expanding the chromatography techniques to detect and enrich the free thiol variants in full length Mabs instead of fragments is important for their characterization. mAb A is a recombinant monoclonal IgG1 antibody that binds CD20 expressed on B cells. Here we report the observation of unique chromatographic profiles leading to the identification and characterization of buried unpaired cysteines in the VH domain of mAb A molecule. Since these unpaired cysteines are largely buried in the interior of the VH domain, Ellman’s assay can be used to detect the presence of free thiols only under a denaturing condition. LC-MS peptide mapping identified the free thiols at Cys-22 and Cys-96 in the heavy chain. Furthermore, the Fab fragment of mAb A containing buried unpaired cysteines was enriched using ion exchange chromatography, allowing for the evaluation of its binding to CD20 with various in vitro binding assays. Full length mAb A containing unpaired cysteines in its Fabs was also enriched by cation exchange chromatography (CEX) and tested in the cellbased complement-dependent cytotoxicity (CDC) assay to evaluate the impact of unpaired cysteines on the antibody function.
■
MATERIALS AND METHODS Materials. Trifluoroacetic acid (TFA) and 5,5′-dithio-bis(2nitrobenzoic) acid (DTNB) were purchased from Pierce (Rockford, IL, USA). 2-(N-Morpholino)ethanesulfonic acid (MES), cysteine, glutathione dimer, cystamine hydrochloride, cupric sulfate, urea, dehydroascorbic acid (DHAA), N-ethylmalemide (NEM), and Tris were purchased from SigmaAldrich (St. Louis, MO, USA). Reduced glutathione (GSH) and monobasic potassium phosphate were purchased from J.T. Baker (Phillipsburg, NJ, USA). Acetonitrile (ACN) and isopropanol was purchased from Burdick & Jackson (Muskegon, MI, USA). Ethylenediaminetetraacetic acid (EDTA) was purchased from EM Science (Gibbstown, NJ, USA). nDodecyl-b-d-maltopyranoside (DDM) was purchased from Anatrace (Maumee, OH, USA). mAb A was produced at Genentech using a CHO cell culture process and purified through the typical downstream multistep chromatography process to a high degree of purity intended for therapeutic applications. RP-HPLC Conditions. RP-HPLC analysis was performed on an Agilent 1200 HPLC system (Palo Alto, CA, USA) equipped with a binary gradient pump, autosampler, temperature-controlled column compartment, and a diode array detector. The system included a Pursuit 3 diphenyl reversed phase column (150 × 4.6 mm, 3 μm, Varian, Lake Forest, CA, USA) running at 75 °C and using absorbance at 280 nm for detection. The mobile phase consisted of 0.1% TFA in water (mobile phase A) and 0.1% TFA in ACN (mobile phase B). B
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Liquid Chromatography Electrospray Time-of-Flight Mass Spectrometry (LC-ESI-TOF) Analysis. Mass analyses of antibodies were performed on Agilent 6210 electrosprayionization Time-of-Flight (ESI-TOF) mass spectrometer (Agilent, Santa Clara, CA). Agilent Nano-HPLC Chip (equivalent to Zobax 300SB-C8 analytical column of 43 mm length ×75 μm internal diameter, 5 μm particle size) was used for desalting and separation using a linear gradient from 20% B to 90% B in 6 min. The solvent A was 0.1% formic acid in water, and solvent B was 0.1% formic acid in acetonitrile. The flow rate was at 0.4 μL/min. The injection amount is about 5 ng. The effluent from the column was directed into the electrospray source of an Agilent 6210 ESI-TOF mass spectrometer. The instrument was operated in the positive mode with a capillary voltage of 1900 V. Spectra consisting of multiple charged protein ions were deconvoluted using the Agilent MassHunter Workstation Software. Online RP-HPLC/ESI-MS Analysis of mAb A. mAb A was separated by reversed-phase chromatography using a diphenyl column (described above) for direct online mass spectrometer analysis on a QSTAR Elite Hybrid system (Applied Biosystems/MDS SCIEX). The mass spectrometer was set at an m/z range of 750−4000, source temperature of 350 °C, and declustering potential of 120 V. Spectra were derived from multiply charged ions and deconvoluted using the Analyst QS 2.0/BioAnalyst QS 2.0 software package. Trypsin Enzymatic Digestion and Reduction of Disulfide Bonds. To denature the proteins under nonreducing condition and perform alkylation of free thiol groups, ∼0.5 mg of protein in formulation buffer was mixed with a denaturing buffer consisting of 8 M guanidine hydrochloride, 10 mM NEM, 0.1 M sodium acetate, pH 5.0, to a volume of 0.5 mL. The mixture was then incubated at 37 °C for 3 h. After incubation, the solution was buffer exchanged into 600 μL of 0.1 M Tris, 1 mM CaCl2, pH 7.0, using a NAP-5 column. Acetonitrile (ACN) was added to each sample to a final concentration of 10%. The digestion reactions were carried out by adding trypsin at an enzyme to substrate ratio of ∼1:10 (w/ w) and incubating the mixture at 37 °C overnight. To quench the reactions, TFA was added to the sample digests to a final concentration of 0.2% (v/v). To further reduce the disulfidelinked peptides from the native tryptic peptide digest and thereby create a digest for a reduced peptide map, Tris(2carboxyethyl)phosphine (TECP) was added to a final concentration of 15 mM, and the reduction was performed at 37 °C for ∼30 min prior to the addition of TFA to a final concentration of 0.2% (v/v). LC/ESI-MS/MS Analysis of mAb A Digested with Trypsin. The online LC/ESI-MS/MS analyses of the tryptic digests were carried out using an Agilent Capillary 1200 HPLC system coupled with a Thermo Fisher (San Jose, CA) LTQ Orbitap mass spectrometer equipped with an electrospray ionization source. The antibody digest was injected onto a Jupiter C18 column (5-μm particle size, 300-Å pore size, 1.0 × 250 mm) with column temperature set at 55 °C. Mobile phase A was a mixture of water and TFA in a 1000:1 ratio (v/v), and mobile phase B was a mixture of acetonitrile, water, and TFA in a 900:100:1 ratio (v/v). Peptides were eluted using a linear gradient of 0−45% B at a flow rate of 0.07 mL/min and monitored at 214 nm. The spray voltage was 4.7 kV, and the capillary temperature was 250 °C. The mass spectrometer was operated in the data dependent mode to automatically switch between MS and MS/MS. Survey full scan MS spectra were
acquired from m/z 300 to m/z 2000 in the FT-Orbitrap with a resolution of R = 60,000 at m/z 400. The five most intense ions were fragmented in the linear ion trap using collision-induced dissociation. Reagents and Cell Line for CD 20 Binding Assays. Recombinant CD20 extra-cellular domain (ECD) protein was generated at Genentech (South San Francisco, CA). The full length membrane protein was expressed in E.coli and purified in house using methods described by Ernst et al.20 Affinitypurified and biotinylated goat antihuman kappa chain polyclonal antibody and avidin-horseradish peroxidase (HRP) were purchased from Vector Laboratories, Inc. (Burlingame, CA). Multiarray 96-well plate and sulfo-TAG labeled streptavidin were purchased from Meso Scale Discovery (MSD, Gaithersburg, MD). The human B-lymphoblastoid cell line WIL2 was originally purchased from American Type Culture Collection (ATCC, Rockville, MD) and cultured in RPMI 1640 medium supplemented with 2 mM L-glutamine and 10% fetal bovine serum (FBS). Cell-Based CD20 Binding Assay. The binding of mAb A Fab fragment to cell surface CD20 was assessed in a cell-based binding assay using MSD technology platform. Details of the assay have been described by Lu et al. and are briefly described below.21 Human B-lymphoblastoid cell line WIL2 cells were plated in a 96-well MSD plate at 25,000 cells per well. After blocking the plates with PBS containing 30% FBS, the enriched intact Fab and free-thiol Fab, which were serially diluted in PBS (pH 7.4) containing 10% FBS, were added to the corresponding wells in duplicate. An unrelated recombinant Fab was also included as a negative control. The bound Fab was detected by incubating the plate with biotinylated goat antihuman IgG kappa chain antibody (2.5 μg/mL) followed by Sulfo-TAG Streptavidin (1 μg/mL). The signals, which correlate to the mAb A Fab bound to cell surface CD20, were detected using Sector Imager 6000 reader (Meso Scale Discovery). Dose−response binding curve was generated by plotting the mean electrochemiluminescent signals from duplicates of sample dilutions against the antibody concentrations and fitting the data with a four parameter model using GraphPad Prism (Prism Software Company, Nashville, TN). ELISA-Based CD20 Binding Assay. The binding of mAb A Fab to CD20 ECD was determined in an enzyme linked immunosorbent assay (ELISA). Assay plates were coated with 2 μg/mL of CD20 ECD in 0.1 M carbonate buffer (pH 9.6) at 4 °C overnight, and blocked with PBS containing 3% BSA, 0.05% Tween-20, and 15 ppm Proclin, pH 7.4. Intact Fab and freethiol Fab serial dilutions (a total of 10 concentrations within a range of 2000 nM−0.008 nM) in the assay buffer (PBS, pH 7.4. 0.5% BSA, 0.05% tween-20, 15 PPM Proclin) were added to the CD20 ECD coated wells (100 μL/well). After the plate was incubated at room temperature for 2 h with gentle shaking and then washed three times, 100 μL of biotinylated goat antihuman IgG kappa chain antibody (2.5 μg/mL) was added to each well. The plate was incubated for an additional hour at room temperature and washed three times followed by the addition of 100 μL of Avidin-HRP and incubation at room temperature for an additional hour. The bound Fab was detected by incubating the plates with 3,3′,5,5′-tetramethylbenzidine substrate (TMB, KPL Laboratories, Gaithersburg, MD). The reaction was quenched by adding 100 μL of 1 M phosphoric acid, and the absorbance was read at 450 nm, with reference at 650 nm, using a SpectraMax plate reader (Molecular Devices Corporation, Sunnyvale, CA). Dose− C
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
response binding curve was generated by plotting the absorbance values from sample dilutions against the antibody concentrations and fitting the data with a four parameter model using GraphPad Prism. Complement-Dependent Cytotoxicity (CDC) Assay. The CDC assay measures the ability of an anti CD20 antibody to lyse human B lymphoblastoid (WIL2-S) cells in the presence of human complement. In this assay, varying concentrations of reference standard and sample(s) are diluted in assay diluent are incubated with WIL2-S cells (50,000 cells/well) in the presence of a fixed amount of human complement. The plate(s) are then incubated at 37 °C/5% CO2 in a humidified incubator for 1 to 2 h, At the end of the incubation period, 50 μL of the redox dye alamarBlue is added to each well, and the plate(s) are further incubated for 15 to 26 h. AlamarBlue is a redox dye that fluoresces at an excitation wavelength of 530 nm and an emission wavelength of 590 nm when reduced by live cells. Therefore, the changes in color and fluorescence are proportional to the number of viable cells. The results, expressed in relative fluorescence units (RFU), are plotted against the anti CD20 antibody concentrations.
■
RESULTS Heterogeneity of mAb A Revealed by RP-HPLC. For mAb A, three peaks with a ratio of 68:28:4 were separated with multiple reversed phase columns (diphenyl, C8 and polymer based columns) as shown in Figure 1. The significant heterogeneity observed by RP-HPLC is rather unusual for an IgG1 monoclonal antibody highly purified for therapeutic use, prompting us to launch an investigation to understand the cause of such heterogeneity. The RP-HPLC heterogeneity was apparent at column temperatures between 55 and 80 °C, even though significant tailing was observed at temperatures lower than 70 °C (data not shown). Online RP-HPLC with mass spectrometry showed the charge envelopes for species in peaks 1 and 2 are different (Figure 2A and 2B), suggesting the residual conformations in the two peaks under RP-HPLC conditions are different. However, the deconvoluted mass spectra indicated that the species separated by RP-HPLC have similar molecular weights (Figure 2C and 2D), and multiple molecular weights corresponding to the difference in glycoforms such as G0, G1, G2, and other glycans were observed for each peak. The charge envelope profile of peak 1 shown in Figure 2A exhibits a higher apparent m/z compared with that of peak 2 shown in Figure 2B, indicating that the molecular species in peak 1 has a more compact structure when compared with the species in peak 2.22,23 Peak 3, the smallest peak in RPHPLC chromatogram, also contains a species with a molecular weight similar to those of peak 1 and peak 2. However, the intensity for the species in peak 3 was weak because the abundance of this species is low and the signal intensity was suppressed by TFA. It has been observed that high temperature and organic solvent with TFA as a low pH buffer additive can cause antibody degradation and fragmentation.24,25 To rule out the possibility of on-column degradation artifacts, fractions from peak 1 and peak 2 were collected, concentrated, and then reinjected into the same column. Their retention times did not change indicating that the species in the two peaks are stable, and the separation profile was not due to degradation artifacts. To test whether such heterogeneity is glycoslation related, mAb A was deglycosylated using PNGase F. The resulting deglycosylated mAb A still displayed similar heterogeneity in the RP-HPLC analysis. The nearly identical masses observed in
Figure 1. RP-HPLC profiles for the intact mAb A using three different columns: (A) Varian diphenyl; (B) Zorbax SB C8; and (C) Varian PLRP-S column.
the two RP-HPLC peaks ruled out the possibilities such as Scysteinylation or S-glutathionylation at cysteine residues and pyroglutamic acid formation at the N-terminus. The other possible causes of heterogeneity include deamidation, isoaspartic acid formation, and disulfide related variants. It was interesting to observe that peak 2 partially converted to peak 1 if it was concentrated in the presence of 4 M guanidine HCl, whereas peak 1 was unchanged after the exposure to guanidine HCl (data not shown). CEX Fractions Contained Variants of mAb A with Different RP-HPLC Profiles. To further understand the unique RP-HPLC heterogeneity observed for mAb A, other modes of chromatography such as cation exchange chromatography (CEX) were used to analyze mAb A samples. One advantage of using CEX is that antibody molecules in the fractions collected from CEX remain in their native state, allowing for the evaluation of bioactivity for mAb A species in each fraction. mAb A charge variants were evaluated by using the Dionex ProPac WCX column with MES or ACES buffer at D
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 2. Online LC-MS analysis of the RP-HPLC peaks: (A) peak 1 mass spectrum; (B) peak 2 mass spectrum; (C) peak 1 deconvoluted spectrum; and (D) peak 2 deconvoluted spectrum.
Addition of EDTA to the copper treated sample did not restore the multiple peak profile observed in the RP-HPLC, suggesting that the collapse of minor peaks into main peak was not due to the formation of an adduct between Cu2+ and the antibody. Treatment of mAb A with other oxidants such as DHAA, glutathione dimer (i.e., oxidized glutathione), and cystamine at room temperature for 24 h did not remove the RP-HPLC heterogeneity. It should be noted that DHAA was used to reoxidize hinge region cysteines for thiomAb.26 These results indicated that the free thiols, if present in mAb A as suggested by the data from copper treatment experiment, may reside within a rigid and solvent-inaccessible region, limiting their redox kinetics. Free Thiol Measurements under Denaturing Conditions. The amount of free thiols in mAb A was measured using the Ellman assay under native and denaturing conditions. Omalizumab, which has been found to have unpaired cysteines,14 was used as a control. Similar to a previous report,13 we observed that the free thiol level for mAb A increased significantly under denaturing conditions (Table 1). Only 0.07 mol of free thiols in each mole of mAb A was observed under native conditions, and it increased to 0.99 mol of free thiols per mole of mAb A in the presence of 8 M urea. In our Ellman assay, omalizumab was found to have 1.35 mol of free thiols per mole antibody. The increase in the free thiol level under 8 M urea denaturing condition indicated the
different pHs, ranging from 5.7 to 7.2. As shown in Figure 3A, the major peak in the CEX profile of mAb A exhibited splitting when a pH 6 MES buffer was used. The fractions collected from the front shoulder (peak a) and main peak (peak b) were reinjected into the CEX column, and the retention times for these two fractions were unchanged (Figure 3B and 3C), indicating that these two fractions contain stable mAb A variants. The main peak (peak b) from CEX, when analyzed by the RP-HPLC method, contained mainly peak 1 (Figure 3C, inset). Similarly, the shoulder peak (peak a) in CEX was highly enriched with species eluting as peak 2 in the RP-HPLC assay (Figure 3B, inset). The separation of peaks a and b in CEX occurred only when the buffer pH was between 5.8 and 6.3. For example, at pH 6.9, peak a could not be resolved from peak b, and only a single major peak in the CEX profile of mAb A was observed (data not shown). Treatments of mAb A with Oxidants. It is common to use redox treatment to enrich or eliminate certain disulfide or thiol related isoforms of antibodies.6 Cupric sulfate has previously been added to media to remove free thiol heterogeneity.18 Therefore, we tested whether the multiple species separated in RP-HPLC assay were related to free thiols by incubating mAb A with CuSO4 up to 100 μM at room temperature for 24 h. As shown in Figure 4, the minor peaks (peaks 2 and 3) observed in the RP-HPLC profiles disappeared and were converted into the main peak after copper treatment. E
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 4. RP-HPLC separation for mAb A samples treated with various amount of CuSO4 for 24 h at ambient temperature: (A) control; (B) 5 μM; (C) 50 μM; and (D) 100 μM.
Figure 3. (A) CEX of mAb A showing the splitting main peak and (B) and (C) the two CEX fractions collected from the splitting main peak and reinjected into CEX; the insets are the corresponding RP-HPLC profiles for the two CEX fractions.
Table 1. Free Thiol Contents Measured by Ellman Assay for Several Key Samples
presence of buried free thiol in mAb A. On the contrary, there is little change in the free thiol content under denaturing conditions for omalizumab, which suggests the difference between the free thiols in the two antibodies is their solvent accessibility and/or conformational dynamics. The Ellman’s assay results also showed that the molecular species in the two CEX peaks contained significantly different amounts of free thiols. The CEX shoulder peak (peak a) has 1.7 mol of free thiol per mol of protein, suggesting that one disulfide bond is not formed. The CEX main peak, on the other hand, contains only 0.25 mol of free thiol per mol of protein. The free thiol level was also markedly reduced after CuSO4 treatment, confirming that Cu2+ effectively oxidized free thiols in mAb A. All these results are consistent with the presence of mAb A free thiol variants. Identification of Unpaired Cysteines by LC-MS. To further confirm the presence of unpaired cysteines and determine their locations in mAb A, tryptic peptide maps were employed to characterize both native and reduced mAb A. In this peptide map analysis, N-ethylmaleimide (NEM) was used to tag any free thiols as well as to minimize free thiol reoxidation and disulfide bond scrambling during the denaturation and digestion of the protein. NEM was selected
Mol S−H/Mol mAb samples mAb A mAb A CEX mAb A CEX mAb A Cu2+ mAb A Cu2+ omalizumab a
peak a peak b control treated
native
denatured
0.07 N/Da N/Da N/Da N/Da 1.35
0.99 1.7 0.25 0.92 0.14 1.48
N/D: not determined.
to block free thiols because it is a small and uncharged molecule that can react with thiols in hydrophobic environments.27 Furthermore, unlike other alkylation agents such as iodoacetic acid (IAA) and iodoacetamide (IAM), NEM has high specificity and efficiency to alkylate thiol at neutral or slightly acidic pH. These conditions prevent free thiol scrambling or reoxidation and minimize the reactions that could occur at pHs above 7 between NEM and other nucleophiles in a protein.28 The NEM tagged sample was first analyzed by a microfluidicbased HPLC system in conjunction with ESI-TOF mass spectrometry for intact mass analysis. The masses of major F
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
on the H2 and H10 peptides. To confirm the presence of these unpaired cysteines, the reduced tryptic map of NEM tagged mAb A was examined. A significant portion of the H2 and H10 peptides was found to be tagged with NEM, indicating the presence of free thiol on these peptides. Several other cysteine containing peptides were also found to be tagged with NEM but at much lower levels. By comparing the extracted ion chromatogram from the LC-MS data for each peptide containing unmodified cysteine to that of its corresponding NEM tagged peptide, the level of unpaired cysteine (tagged by NEM) can be estimated (Table 3). The principle of this
species observed are shown in Table 2. In addition to the masses corresponding to intact mAb A with different Table 2. Masses for Major Species Observed from the ESIMS Analysis of NEM Treated mAb A (Intact and Reduced)a major species intact mAb A
G0/G0 G0/G1 G0/G0 with 2 NEM G0/G1 with 2 NEM heavy chain of heavy chain (G0/G0) mAb A heavy chain (G0/G1) heavy chain (G0/G0) with 2 NEM heavy chain (G1/G0) with 2 NEM light chain of mAb A a
observed mass (Da)
theoretical mass (Da)
148452 148614 148702 148864 51004 51167 51254
148455 148616 148706 148868 51004 51166 51255
51417
51416
23238
23238
Table 3. % NEM Tagged Peptide Was Determined by LCMS and Used To Represent the Free Thiol Level in Each Cysteine Containing Peptide % NEM tagged peptide
NEM was found to be attached to heavy chain mainly.
intact mAb A
glycoforms such as G0 and G1, masses arising from a significant amount of mAb A containing G0/G0 or G0/G1 glycoforms with two NEMs attached were also observed, suggesting that a disulfide bond between one pair of cysteines is missing in those mAb A molecules. Additionally, the results from mass spectrometry analysis of the reduced mAb A following NEM treatment showed that only heavy chains, not light chains, were modified with 2 NEMs. This finding strongly suggests that the unpaired cysteines are present only in the heavy chain of mAb A. Tryptic digests from native and reduced mAb A (with and without NEM pretreatment) were separated and analyzed using LC/MS to identify the disulfide bond-containing peptides. A portion of the peptide map profile obtained for native mAb A and the observed masses for the expected disulfide-linked peptides as well as individual reduced or NEM tagged peptides are provided in the Supporting Information (Figure S-1 and Table S-1). Identification of each peptide species was achieved by using the accurate mass results provided from Orbitrap MS and MS/MS analysis. Mass errors between the expected and observed mass were typically less than 5 ppm. All predicted disulfide-linked peptides were recovered, and the observed masses closely matched the expected masses, with the exception of one small disulfide-linked peptide H18-L18 (m/z 757.28), which likely elutes in the void volume. This interchain disulfidebond linkage, Cys225 (heavy chain) linked to Cys213 (light chain), was confirmed in another peptide map obtained under native condition using Lys-C digestion (data not shown). To determine whether alternative disulfide linkages are present, we searched for masses corresponding to all possible combinations of cysteine-containing peptides in the mass spectrometry data. However, no mixed disulfide linkages were found in the search. It has been reported that artifacts may occur at cysteine residues in the tryptic peptide map for a nonreduced protein or when TCEP is used to reduce the disulfide bonds during the peptide map analysis.29 These artifacts are more pronounced at elevated temperatures (e.g., 60 °C). Since the tryptic digestion for mAb A was carried out at 37 °C, we have not observed these reported artifacts. All disulfide-linked peptides were well separated with good recovery in the native peptide map, with the exception of the H2−H10 peptide. The low recovery of this peptide is unexpected and likely due to the presence of unpaired cysteines
free thiol containing peptidea H2 H10 H13 H14 H21 H35 H40 L9 L16
mAb A resolved by RPHPLC
mAb A resolved by CEX
untreated
treated with CuSO4
peak 1
peak 2
peak 3
peak a
peak b
19.7 23.1 1.4 1.9 1.9 8.3 3.1 0.3 1.0
0.6 0 0.5 0.9 2.0 7.2 2.4 0.5 0.9
1.5 0 1.8 1.0 1.8 3.5 2.5 0.2 0.9
49.8 49.3 4.3 0.8 2.3 4.6 2.6 0.7 1.1
80.7 82.5 4.9 1.6 2.8 6.4 3.2 1.0 1.6
35.7 43.7 3.0 1.4 1.3 5.0 7.2 0.3 0.9
3.3 2.4 3.0 1.4 2.0 8.1 5.8 0.1 0.9
a
The peptides are labeled according to either light (L) or heavy chain (H) and number of peptides generated from the enzyme digestion.
method is illustrated in Figure 5 using H10 peptide as an example. Figure 5A displays the extracted ion chromatogram of the reduced H10 peptide and the H10 peptide alkylated by NEM from the tryptic peptide map. Typically, NEM alkylation will make peptides more hydrophobic thus tagged peptides elute later. NEM alkylation introduces a chiral center in the modified cysteine side chain, resulting in diastereomers that often can be resolved in the LC-MS analysis with similar MS/ MS spectra.30 As shown in Figure 5A, the two isomers of NEM tagged H10 elute at approximately 51 min, and the total area of these two peaks was used to calculate the relative amount of the NEM tagged H10 peptide. Figure 5B and 5C show the mass spectra for the reduced H10 and the NEM tagged H10 peptide, respectively. The peak identity was confirmed by the accurate masses observed in the LC-MS analysis. The level of free thiol for the H10 peptide was calculated by dividing the peak area of the NEM-modifided H10 peptide by the total peak area for the modified and unmodified H10 peptide. There are a total of fifteen cysteine-containing tryptic peptides for mAb A. The relative levels of nine free thiol containing peptides are listed in Table 3. Four other cysteine containing peptides (H18, H19, L2, and L5) were found to have very low levels of free thiol. Two small peptides (L18 and H27) are not observed in tryptic map analysis, presumably because they elute in the void. The amounts of free thiols observed in two particular peptides, H2 (containing Cys-22) and H10 (containing Cys-96), are significantly higher than those found in other cysteine containing peptides in mAb A. Cys-22 and Cys-96 are expected to form a disulfide linkage in G
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 5. (A) Extracted ion chromatogram of H10 and H10 tagged by NEM. The peak areas were shown in the chromatogram and were used to calculate the percentage of H10 free thiols (see Table 3) (B) and (C) The corresponding mass spectra of doubly charged species for H10 and H10 tagged by NEM. The double peaks eluting at 51 min are caused by diastereomers.
locations and amount of free thiols in the RP-HPLC fractions, and the results are included in Table 3. There are significant differences among the three RP-HPLC fractions for free thiols on H2 and H10 peptides. For RP-HPLC peak 1, the free thiol levels at both Cys-22 and Cys-96 were minor. However, for peak 2 and peak 3 from RP-HPLC, the percentages of free thiol at both Cys-22 and Cys-96 increased to approximately 50% (in peak 2) and 81% (in peak 3). These results confirm that the reversed phase separation of mAb A is due to the thiol heterogeneity. Furthermore, peak 2 and peak 3 in RP-HPLC are arising from the species containing one and two unformed disulfide linkage in one (peak 2) and two (peak 3) heavy chains (i.e., in one and both Fabs) between Cys-22 and Cys-96. The two fractions collected from CEX (Figure 3) were analyzed by the peptide mapping method and the results are also included in Table 3. Comparison of the free thiol levels obtained for the CEX fractions to those from the RP-HPLC fractions leads to the conclusion that the CEX front shoulder peak (peak a) is enriched with the species containing unpaired Cys-22 and Cys-96 in one of the two heavy chains. The main CEX peak corresponds to the species with no or little unpaired cysteines at these locations. Chromatographic Separation of Free-Thiol Fab and Intact Fab. Papain digestion, then coupled with chromatographic separation, has been widely used to generate Fab and Fc fragments for structural or functional studies.31 However, we observed nonspecific cleavage when using papain to generate Fab and Fc of mAb A. Limited Lys-C digestion has also been shown to be an effective way to generate antibody fragments.32 Thus, it was used to generate mAb A Fab for the antigen binding studies. As shown in Figure 6A, the CEX separation between free-thiol Fab and intact Fab is better than that of the full-length mAb A and free thiol variants as shown in Figure 3. Similar to the free thiol containing full-length mAb, which
the VH domain of an intact antibody. The observation of high level free thiols in H2 and H10 suggests that the disulfide bond is missing between Cys-22 and Cys-96 in a significant amount of mAb A. These unpaired cysteines can be oxidized to form a disulfide by Cu2+ as evidenced by the significant decrease in H2 and H10 free thiol peptides following treatment with CuSO4 (Table 3). The high relative free thiol levels found in H2 and H10 peptides also provide an explanation for the unusual low H2−H10 peptide intensity observed in the native peptide map analysis. Although two small peptides (L18 and H27) were not detected in the tryptic peptide map, they are not believed to carry significant amounts of free thiols since their disulfidelinked partners, H18 and H21 respectively, were both detected and found to have no or low levels of free thiols. Note that the L18 and H18 peptides are involved in the disulfide linkage between the light chain and heavy chain. Unpaired cysteines in these peptides would have resulted in a free light chain, which was not observed by RP-HPLC method or CE-SDS (capillary electrophoresis- sodium dodecyl sulfate) method (data not shown). Identification and Quantitation of Free Thiol in RPHPLC and CEX Fractions. To test if the heterogeneity observed in the RP-HPLC analysis of mAb A is caused by the missing disulfide bond (i.e., the presence of free thiols) in the VH domain of some mAb A molecules, the fractions collected from RP-HPLC were analyzed by ESI-MS analysis after derivatization by NEM and reduction by TCEP. The ESI-MS results showed that the heavy chain of the mAb A in peak 1 was not tagged by NEM, but the heavy chain of the mAb A in peak 2 was tagged by two NEMs, indicating that the RP-HPLC method can separate the intact mAb A from its variants that contain unpaired cysteines. The same tryptic peptide map analysis described above was also performed to determine the H
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 7. Representative binding curves of intact Fab and free-thiol Fab to cell surface CD20. Serial dilutions of intact Fab (square) and free-thiol Fab (triangle) and negative Fab control (circle) were incubated with WIL2 cells on 96-well MSD plates. Bound antibodies were detected with Sulfo-TAG streptavidin. Error bars for each data points were not shown as the size of error bars is approximately the same as the size of the symbols shown.
appears to show only a slightly higher binding activity compared to that of the intact Fab. The binding of the intact Fab and the free-thiol Fab to CD20 ECD was also evaluated in an ELISA (Figure 8). Consistent with the observation in the cell-based assay, the free-thiol Fab showed a slightly higher binding activity than that of the intact Fab. Detectable nonspecific binding signal of the negative Fab control was observed when its concentration was at 667 nM and above; however, that signal was much lower than the signal of intact Fab or the free-thiol Fab. The results from the in vitro
Figure 6. (A): CEX profile for the antibody fragment generated through limit lysC digestion and (B) and (C): free-thiol and intact Fab fractions were collected and analyzed by RP-HPLC.
elutes before the native form in the CEX analysis, the free-thiol Fab elutes before the intact Fab in CEX as well. The free-thiol Fab elutes after the main peak (intact Fab) in RP-HPLC (Figure 6B and Figure 6C), consistent with the separation observed for the full-length mAb A by RP-HPLC, where the free thiol containing mAb A elutes after the main form. The ratio of free-thiol to intact Fab (roughly 1:5) observed in CEX is consistent with the RP-HPLC results for the intact mAb A, showing that approximately 40% of the intact mAb A molecules contain one unformed disulfide. Both free-thiol and intact Fab were collected from CEX for CD20 binding studies. Prior to the binding studies, the presence and absence of free thiols at Cys-22 and Cys-96 in free-thiol Fab and intact Fab, respectively, were confirmed with the same characterization methods described above (i.e., NEM derivatization and peptide map analysis; data not shown). Binding of Fab Variants to CD20. Absence of the disulfide bond between Cys-22 and Cys-96, located in the VH domain of mAb A, can potentially affect its antigen binding. Two in vitro assays were employed to evaluate the binding of Fab variants to cell surface and soluble CD20. As shown in Figure 7, both free-thiol Fab and intact Fab bind to cell surface CD20, whereas an unrelated recombinant Fab (negative control) does not bind CD20. Moreover, the free-thiol Fab
Figure 8. Representative binding curves of intact Fab and free-thiol Fab to CD20 ECD. Serial dilution of intact Fab, free-thiol Fab, and negative Fab control were incubated with CD20 ECD in 96-well plates. Bound antibodies were detected with biotinylated goat antihuman IgG kappa chain antibody. No error bars were shown in this plot as the samples were tested in singlet in this assay. The assay was repeated in a separate run, and the results were confirmed (data not shown). I
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 9. The CDC activities of mAb A with (▲) or without free thiols (□) were compared to the reference standard (○). Samples were tested in two independent assays and representative curves are shown.
approximately 68% of the main peak (main species with no unpaired cysteines), 28% of peak 2 (variant with one unformed disulfide bond), and 4% of peak 3 (variant with two unformed disulfide bonds). The CEX analysis of mAb A after limited LysC digestion showed that approximately 17% of the total Fab fragments contain an unformed disulfide bond. These results suggest that the pairing of two heavy chains in the assembly of mAb A molecule from individual components is a process independent of the burial of these two unpaired cysteines, giving rise to a statistical distribution observed by RP-HPLC. Based on the reversed phase peak area distribution, the unformed disulfide between Cys-22 and Cys-96 is expected to contribute about 0.7 mols of free thiols in every mole of mAb A, a level that is consistent with the free thiol content determined by Ellman assay under denaturing conditions. Under RP-HPLC conditions, an antibody is expected to be at least partially denatured. Presumably, the difference in the disulfide status (reduced vs oxidized) of the Cys22-Cys96 pair in mAb A gives rise to different exposure of hydrophobic residues and results in the resolution observed by RP-HPLC. The RP-HPLC results for mAb A are similar to the data from the studies employing RP-HPLC to monitor IgG2 disulfide variants.7 In contrast, since the mAb A free thiols are inaccessible to solvent under the native condition, the surface charge changes for the intact antibody are likely minor. Consequently, these free thiol related variants are not separated by CEX at pH 6.9 (data not shown) when compared to the more denaturing RP-HPLC method. Also, CEX separation could be confounded by other modifications, further increasing chromatographic complexity. Aside from RP-HPLC and CEX, CE-SDS methods have been instrumental in identifying the IgG2 disulfide linkage variants.36 However, we did not observe separation of free thiol variants in mAb A using a CE-SDS method for the free thiol variants (data not shown). The Buried Nature of Unpaired Cysteines. Treatment of copper(II) eliminated the heterogeneity of mAb A observed by the RP-HPLC method, whereas other oxidants were ineffective in removing this heterogeneity. The difference in the reactivity of free thiols at Cys-22 and Cys-96 in mAb A toward different oxidizing agents is not fully understood at this time. Limited solvent accessibility is likely a major factor for the protection of free thiols from oxidation. However, degree of conformational flexibility around the unpaired cysteines may also be an important factor and allow only oxidizing agents with small size
CD20 binding assay were also consistent with preliminary data (not shown) from a surface plasmon resonance assay indicating that the binding response of free-thiol Fab to CD20 is approximately 15% higher than that of intact Fab. Complement-Dependent Cytotoxicity (CDC) Activity of mAb A Containing Unpaired Cysteines. The mAb A with and without unpaired Cys-22 and Cys-96 was enriched using the CEX method described above (Figure 3). Since the mAb A containing free thiols in both Fab arms was at a lower abundance and the CEX resolution did not allow for the separation of mAb A with the free thiols in one and two Fab arms, the mAb A fraction containing the free thiols is actually a mixture of mAb A with either one or two free-thiol Fabs. In order to determine whether the presence of free thiols had an effect on CDC activity of mAb A, the two fractions collected from CEX containing Mab A with and without the free thiols were compared in the CDC assay relative to the reference standard. This CDC activity assay was shown to be specific (for mAb A) and quantitative in a validation study (data not shown). It is an ideal assay to simultaneously assess the antigen binding and Fc effector function of mAb A. As shown in Figure 9, the curves generated by the two mAb A samples were superimposable with the reference standard, indicating that the presence of free thiols in mAb A has no influence on CDC activity.
■
DISCUSSION Chromatography Heterogeneities from Antibody Free Thiols Variants. RP-HPLC methods have been developed to separate antibody variants. Good column selection and relatively high column temperatures are key optimization parameters.33 It has been reported that IgG1 intact antibodies are typically homogeneous by RP-HPLC analysis with minor peaks often associated with chemical modification such as pyroglutamic acid formation, cystenie capping of free thiols, or C-terminal lysine variants.34,35 Therefore, it is unusual to observe three major peaks for mAb A in RP-HPLC with C8, diphenyl, and polymer based columns. The data described in this report indicate that the observed RP-HPLC heterogeneity results from the presence of variants containing unpaired cysteines on the H2 and H10 peptides (i.e., Cys-22 and Cys-96). The peak area distribution observed in the RP-HPLC analysis of mAb A includes J
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
While impact of free thiols on the mAb A function is not significant, data showing unpaired cysteines have significant impact on antibody activity have been reported. For example, the Fab fragment containing unpaired cysteines from omalizumab has a much lower binding affinity to its antigen (IgE) when compared to the Fab containing the oxidized cysteines.14 Apparently, depending on the sequences of the variable regions in the Fab (i.e., VH and VL domains), the presence of free thiols at the conserved cysteines in the VH domain (Cys-22 and Cys-96) may have very different impact on the Fab structure and its antigen binding affinity. The impacts of unpaired cysteines on other properties of an antibody, such as shelf life stability or serum half-life, are also likely different for different antibodies and should be evaluated carefully for each molecule. Results from our preliminary studies suggest that the unpaired cysteines have minimal impact on mAb A stability (data not shown). In addition, since the unpaired cysteines observed in mAb A are located in the Fab, they are very unlikely to affect the FcRn binding and the clearance rate of mAb A. In conclusion, we reported the identification and characterization of mAb A variants containing free thiols at Cys-22 and Cys-96 in the VH domain. The presence of free thiols results in the heterogeneity observed in various types of chromatographic assays. The CD20 binding studies and CDC assay performed on mAb A showed that the presence of unpaired cysteines in the Fab has little effect on its biological activity. These findings provide insights into cysteine related heterogeneity in antibodies and the impact of such heterogeneity on antibody activities.
to access the unpaired cysteines when the structure opens up in a dynamic and transient fashion. Since copper can readily oxidize the unpaired cysteines in an antibody as discussed above, allowing the exposure to copper during the manufacturing process could be an effective way to eliminate the free thiol heterogeneity in an antibody. Addition of copper in the cell culture process has been shown to effectively minimize the free thiol content in a different IgG1.14 Adding copper in the antibody downstream process, after the protein A purification step, could potentially introduce fragmentation via metal catalyzed hydrolysis.37 In order to have a robust method to eliminate free thiol heterogeneity, further studies are needed to better understand the root cause for the presence of unpaired cysteines in antibodies. It is interesting to note that the same pair of heavy chain cysteines (Cys-22 and Cys-96) has been also observed to be in the reduced form in omalizumab14 and another mAb.38 Our current hypothesis is that the folding rate of the VH domain in these antibodies during their production and assembly in the expression system (e.g., CHO) is faster than the rate of disulfide formation, thereby trapping the unpaired cysteines in a hydrophobic core in the VH domain. Studies are currently ongoing in our lab to test this hypothesis and to determine if amino acid residues or structural elements in the VH domain responsible for the presence of unpaired cysteines can be identified. Impact of Free Thiols on CD20 Binding and CDC Activity. The results from two different binding assays (cellbased binding assay and ELISA) indicated the binding activity to CD20 of free-thiol Fab is only slightly higher than that of the intact Fab. Considering the variability in these binding assays, the observed small increase in the binding activity of free-thiol Fab for CD20 may not be significant in vivo. The data clearly suggest that the unpaired cysteines do not introduce any significant structural perturbations in the mAb A Fab that would result in the loss of binding to its antigen. Since an immunoglobulin fragment can maintain its conformation after removal of the disulfide bridge,39 it is not surprising to observe little effect from the unpaired cysteines on the antigen binding. The unpaired cysteines apparently are well protected in the VH domain as evidenced by the low reactivity with Ellman’s reagent (DTNB) under native condition and the difficulty in reoxidizing the free thiols in mAb A with oxidants other than copper (see above). The unpaired cysteines were also unchanged in mAb A samples prepared under different stress conditions, such as high temperature, UV light exposure, etc. (data not shown), indicating that these free thiols are extremely stable. Since the VH domain free thiols are stable and the buffers used in the two binding assays did not contain any oxidizing components, lacking significant influence of the unpaired cysteines on the CD20 binding obviously is not due to the formation of disulfide bond during the binding assays. Furthermore, the CDC activities for the mAb A with and without free thiols are indistinguishable, indicating that the minor difference observed between intact and free-thiol Fab binding to CD20 apparently is not significant enough to impact the overall antibody function (antigen binding and Fc effector function). Again, the formation of a disulfide bond from the unpaired cysteines during the CDC activity assay can be ruled out because the VH domain free thiols were unchanged when mAb A was incubated with the redox dye (alamarBlue) used in the CDC assay for up to 20 h (RP-HPLC data not shown) and no other oxidizing components were in the buffers for CDC assay.
■
ASSOCIATED CONTENT
* Supporting Information S
Figure S-1 and Table S-1. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (T.Z.) and
[email protected] (Y.H.K.). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank John Stults, Mary Cromwell, and Reed Harris for their helpful discussions. We also thank Xiangdan Wang and Jihong Yang for providing preliminary surface plasmon resonance data and Mena Odocayen for providing CDC data.
■
REFERENCES
(1) Wang, W.; Singh, S.; Zeng, D. L.; King, K.; Nema, S. J. Pharm. Sci. 2007, 96, 1−26. (2) Liu, H.; Gaza-Bulseco, G.; Faldu, D.; Chumsae, C.; Sun, J. J. Pharm. Sci. 2008, 97, 2426−47. (3) van der Neut Kolfschoten, M.; Schuurman, J.; Losen, M.; Bleeker, W. K.; Martinez-Martinez, P.; Vermeulen, E.; den Bleker, T. H.; Wiegman, L.; Vink, T.; Aarden, L. A.; De Baets, M. H.; van de Winkel, J. G.; Aalberse, R. C.; Parren, P. W. Science 2007, 317, 1554−7. (4) Hagihara, Y.; Mine, S.; Uegaki, K. J. Biol. Chem. 2007, 282, 36489−95. (5) Lacy, E. R.; Baker, M.; Brigham-Burke, M. Anal. Biochem. 2008, 382, 66−8. (6) 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, 7496−508. K
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
(7) 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, 16206−15. (8) 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, 3755−66. (9) Pristatsky, P.; Cohen, S. L.; Krantz, D.; Acevedo, J.; Ionescu, R.; Vlasak, J. Anal. Chem. 2009, 81, 6148−55. (10) Tous, G. I.; Wei, Z.; Feng, J.; Bilbulian, S.; Bowen, S.; Smith, J.; Strouse, R.; McGeehan, P.; Casas-Finet, J.; Schenerman, M. A. Anal. Chem. 2005, 77, 2675−82. (11) Gadgil, H. S.; Bondarenko, P. V.; Pipes, G. D.; Dillon, T. M.; Banks, D.; Abel, J.; Kleemann, G. R.; Treuheit, M. J. Anal. Biochem. 2006, 355, 165−74. (12) Ellman, G. L. Arch. Biochem. Biophys. 1959, 82, 70−7. (13) Zhang, W.; Czypryn, M. Biotechnol. Prog. 2002, 509−513. (14) Harris, R. Dev. Biol. (Basel, Switz) 2005, 117−127. (15) Gevondyan, N. M.; Volynskaia, A. M.; Gevondyan, V. S. Biochemistry (Moscow) 2006, 71, 279−84. (16) Xiang, T.; Chumsae, C.; Liu, H. Anal. Chem. 2009, 81, 8101−8. (17) Chumsae, C.; Gaza-Bulseco, G.; Liu, H. Anal. Chem. 2009, 81, 6449−57. (18) Chaderjian, W. B.; Chin, E. T.; Harris, R. J.; Etcheverry, T. M. Biotechnol. Prog. 2005, 21, 550−3. (19) Burgess, R. R. Methods Enzymol. 2009, 463, 259−82. (20) Ernst, J. A.; Li, H.; Kim, H. S.; Nakamura, G. R.; Yansura, D. G.; Vandlen, R. L. Biochemistry 2005, 44, 15150−8. (21) Lu, Y.; Wong, W. L.; Meng, Y. G. J. Immunol. Methods 2006, 314, 74−9. (22) Katta, V.; Chait, B. T. Rapid Commun. Mass Spectrom. 1991, 5, 214−7. (23) Gross, D. S.; Schnier, P. D.; Rodriguez-Cruz, S. E.; Fagerquist, C. K.; Williams, E. R. Proc. Natl. Acad. Sci. 1996, 93, 3143−3148. (24) Ejima, D.; Tsumoto, K.; Fukada, H.; Yumioka, R.; Nagase, K.; Arakawa, T.; Philo, J. S. Proteins 2007, 66, 954−62. (25) Volkin, D. B.; Klibanov, A. M. J. Biol. Chem. 1987, 262, 2945− 50. (26) Junutula, J. R.; Raab, H.; Clark, S.; Bhakta, S.; Leipold, D. D.; Weir, S.; Chen, Y.; Simpson, M.; Tsai, S. P.; Dennis, M. S.; Lu, Y.; Meng, Y. G.; Ng, C.; Yang, J.; Lee, C. C.; Duenas, E.; Gorrell, J.; Katta, V.; Kim, A.; McDorman, K.; Flagella, K.; Venook, R.; Ross, S.; Spencer, S. D.; Lee Wong, W.; Lowman, H. B.; Vandlen, R.; Sliwkowski, M. X.; Scheller, R. H.; Polakis, P.; Mallet, W. Nat. Biotechnol. 2008, 26, 925−32. (27) Smith, D.; Konigsberg, W.; Blumenfeld, O. Biochem. J. 1964, 91, 589−595. (28) Brewer, C.; Riehm, J. Anal. Biochem. 1967, 18, 248−255. (29) Wang, Z.; Rejtar, T.; Zhou, Z. S.; Karger, B. L. Rapid Commun. Mass Spectrom. 2010, 24, 267−275. (30) Goransson, U.; Craik, D. J. J. Biol. Chem. 2003, 278, 48188−96. (31) Moorhouse, K. G.; Nashabeh, W.; Deveney, J.; Bjork, N. S.; Mulkerrin, M. G.; Ryskamp, T. J. Pharm. Biomed. Anal. 1997, 16, 593− 603. (32) Ren, D.; Pipes, G. D.; Hambly, D. M.; Bondarenko, P. V.; Treuheit, M. J.; Brems, D. N.; Gadgil, H. S. J. Chromatogr. A 2007, 1175, 63−8. (33) Dillon, T. M.; Bondarenko, P. V.; Rehder, D. S.; Pipes, G. D.; Kleemann, G. R.; Ricci, M. S. J. Chromatogr., A 2006, 1120, 112−20. (34) Battersby, J. E.; Snedecor, B.; Chen, C.; Champion, K. M.; Riddle, L.; Vanderlaan, M. J. Chromatogr., A 2001, 927, 61−76. (35) Ren, D.; Pipes, G.; Xiao, G.; Kleemann, G. R.; Bondarenko, P. V.; Treuheit, M. J.; Gadgil, H. S. J. Chromatogr., A 2008, 1179, 198− 204. (36) Guo, A.; Han, M.; Martinez, T.; Ketchem, R. R.; Novick, S.; Jochheim, C.; Balland, A. Electrophoresis 2008, 29, 2550−6. (37) Smith, M. A.; Easton, M.; Everett, P.; Lewis, G.; Payne, M.; Riveros-Moreno, V.; Allen, G. Int. J. Pept. Protein Res. 1996, 48, 48−55.
(38) Ouellette, D.; Alessandri, L.; Chin, A.; Grinnell, C.; Tarcsa, E.; Radziejewski, C.; Correia, I. Anal. Biochem. 2010, 397, 37−47. (39) Uson, I.; Bes, M. T.; Sheldrick, G. M.; Schneider, T. R.; Hartsch, T.; Fritz, H. J. Fold Des. 1997, 2, 357−61.
L
dx.doi.org/10.1021/ac301426h | Anal. Chem. XXXX, XXX, XXX−XXX