Middle-Down Analysis of Monoclonal Antibodies ... - ACS Publications

Feb 18, 2014 - are known to potentially introduce artifacts in the original sample. For this ... as artifact introduction and lengthy sample preparati...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/ac

Middle-Down Analysis of Monoclonal Antibodies with Electron Transfer Dissociation Orbitrap Fourier Transform Mass Spectrometry Luca Fornelli,†,∥ Daniel Ayoub,†,∥ Konstantin Aizikov,‡ Alain Beck,§ and Yury O. Tsybin*,† †

Biomolecular Mass Spectrometry Laboratory, Ecole Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland Thermo Fisher Scientific GmbH, 28199 Bremen, Germany § Centre d’Immunologie Pierre Fabre, 74160 St. Julien-en-Genevois, France ‡

S Supporting Information *

ABSTRACT: The rapid growth of approved biotherapeutics, e.g., monoclonal antibodies or immunoglobulins G (IgGs), demands improved techniques for their quality control. Traditionally, proteolysis-based bottom-up mass spectrometry (MS) has been employed. However, the long, multistep sample preparation protocols required for bottom-up MS are known to potentially introduce artifacts in the original sample. For this reason, a top-down MS approach would be preferable. The current performance of top-down MS of intact monoclonal IgGs, though, enables reaching only up to ∼30% sequence coverage, with incomplete sequencing of the complementarity determining regions which are fundamental for IgG’s antigen binding. Here, we describe a middle-down MS protocol based on the use of immunoglobulin G-degrading enzyme of Streptococcus pyogenes (IdeS), which is capable of digesting IgGs in only 30 min. After chemical reduction, the obtained ∼25 kDa proteolytic fragments were analyzed by reversed phase liquid chromatography (LC) coupled online with an electron transfer dissociation (ETD)-enabled hybrid Orbitrap Fourier transform mass spectrometer (Orbitrap Elite FTMS). Upon optimization of ETD and product ion transfer parameters, results show that up to ∼50% sequence coverage for selected IgG fragments is reached in a single LC run and up to ∼70% when data obtained by distinct LC−MS runs are averaged. Importantly, we demonstrate the potential of this middle-down approach in the identification of oxidized methionine residues. The described approach shows a particular potential for the analysis of IgG mixtures.

W

has to prove similar physicochemical characteristics, functional properties, and clinical efficiency to those of the innovator product.5 One of the most important criteria required is for the amino-acid sequence to be strictly identical.6 Bottom-up liquid chromatography−tandem mass spectrometry (LC−MS/MS) sequencing of mAbs is widely used in pharmaceutical laboratories. Although bottom-up methods provide the most structural information, they suffer a number of drawbacks such as artifact introduction and lengthy sample preparation.7,8 Topdown MS sequencing constitutes an interesting alternative and might become a method of choice as it is a fast and convenient way to obtain useful protein sequence information.9−11 It has the advantage of implying limited sample preparation and therefore minimal artifact introduction due to sample processing. The advent of high-resolution MS instruments coupled to electron transfer dissociation (ETD)12 or electron capture dissociation (ECD)13,14 enabled easier access to topdown analysis. In recent papers our team and others showed that top-down MS allows the sequencing of IgG terminal regions as well as variable domains and the characterization of major glycoforms.15−17 Up to 30% sequence coverage was

ith the tremendous progress in the development of protein engineering technologies, the use of recombinant protein therapeutics has expanded significantly in recent years with monoclonal antibodies (mAbs) becoming the fastest growing class of human therapeutics.1 Currently, there are more than 40 approved mAbs and mAbs derivative products and 30 others in advanced clinical investigations. MAbs are indicated for the treatment of a variety of diseases including cancer.2 The success of mAbs as therapeutics is mainly attributed to their specificity to targets and their favorable pharmacokinetics. All approved therapeutic mAbs for clinical use belong to the immunoglobulin G (IgG) class. IgGs are tetrameric glycoproteins with molecular weights near 150 kDa. They consist of four polypeptide chains: two heavy chains (Hc) of ∼50 kDa each and two light chains (Lc) of ∼25 kDa each, linked together by disulfide bonds to form the characteristic Yshaped complex.3 mAbs are required to be well characterized structurally to ensure their safety, efficiency, batch-to-batch consistency, and stability. With some approved therapeutic mAbs coming off patents, biosimilar antibodies are starting to be filed for approval. Biosimilar antibodies are “generic” versions of “innovator” (or “originator”) antibodies produced through different manufacturing processes and from different clones.4 For a biosimilar to be approved by the regulation agencies, it © 2014 American Chemical Society

Received: November 13, 2013 Accepted: February 18, 2014 Published: February 18, 2014 3005

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry

Article

Figure 1. Schematics of the proposed middle-down workflow. Sample preparation is performed within 1 h, and it is followed by LC−MS/MS and ProSight PC analysis. The presented scheme can be applied to single IgGs or mixtures.

achieved using ETD fragmentation on Orbitrap17 or ECD fragmentation on ion cyclotron resonance (ICR)16 Fourier transform (FT) instruments. However, the current state of the art does not allow overcoming this 30% limit primarily due to the gas phase retention of highly structured areas in the IgG, mainly in correspondence of the immunoglobulin domains and the disulfide bond protected areas. By analogy to top-down and bottom-up approaches, the terms middle-up and middle-down were introduced.18 Middleup refers to the mass measurement of large fragments or subunits of a protein after limited proteolysis while middledown also includes the MS/MS-based sequencing of these subunits.6,18 The terms top-down and middle-down are sometimes misused in the literature when referring to mass measurements of intact IgG or protease generated IgG subunits respectively, without performing MS/MS. Subunits of mAbs can be obtained through the chemical reduction of disulfide bonds thus yielding free heavy chains (50 kDa) and light chains (25 kDa), or by limited proteolysis in nondenaturing conditions with cleavage of the hinge region of the heavy chain yielding Fab (∼50 kDa) or (Fab)′2 (∼100 kDa) and Fc (∼50 kDa) fragments, Figure 1. Chemical reduction in denaturing conditions of these proteolysis-generated subunits results in three ∼25 kDa fragments: the light chain and two half heavy chains (the Fc/2 and the Fd). These 25 kDa fragments better match the performance characteristics of the state-of-theart LC−MS/MS methods and techniques compared to intact mAbs targeted with top-down approaches.3,6 Several proteases such as papain,19 Lys-C,20 and pepsin21 have been described for cleaving in the IgG hinge region under nondenaturing and controlled conditions. These proteases suffer however from limited specificity leading to nonspecific cleavages, complicating therefore the data analysis and interpretation.22 Recently, IdeS (Immunoglobulin G-degrading enzyme of Streptococcus pyogenes) has been reported to specifically cleave between the two consecutive glycine residues under the hinge region.23 It has the advantage of being rapid (30 min), low material consuming,

and active in mAbs formulation buffers.22 The reduced complexity of IdeS-generated ∼25 kDa subunits allows an accurate profiling of N-glycans site by site and, compared to intact IgG, the improved identification of various IgG microvariants (proteoforms) such as C-terminal lysine cleavage, cyclization of N-terminal glutamine, and others.6,24 In this paper, we investigated the utility of ETD coupled with high-resolution high-field Orbitrap FTMS25 for the sequencing and characterization of approved therapeutic monoclonal antibodies in a middle-down fashion, Figure 1. The limited sample processing when using IdeS to generate medium-sized subunits allows minimizing artifact introduction. On the other hand, their relatively small size and the reduction of the disulfide bonds give access to the fragmentation of the S−S protected areas that are not efficiently sequenced by top-down MS. To benchmark the efficiency of the presented method in monitoring post-translational modifications, we chose the oxidation of labile amino-acid side chains as a case study. Oxidation is one of the major challenges for improving the stability profile in the development of monoclonal antibodies.26 Oxidation of the methionine sulfur to the sulfoxide form is one of the common modifications known to occur in mAbs during the manufacturing, formulation, and/or storage process.27 It could decrease bioactivity and stability of IgGs which result in reduced serum half-life and limited shelf life. Occurring mainly in Fc, it leads to decreased binding to the neonatal Fc receptor (FcRn) and loss of protection for catabolism.28 Hydrophobic interaction chromatography and ion exchange are the most commonly used LC techniques to detect the presence of oxidized mAbs proteoforms.26,27 Bottom-up peptide mapping is then used to identify oxidation sites. We employed mild oxidation conditions to generate oxidized IgG proteoforms that we analyzed by middle-down MS. Overall, the results reported here indicate that IdeS-based middle-down ETD MS can constitute a complementary or even an alternative method to bottom-up oxidation assessment in IgG quality control. 3006

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry

Article

Figure 2. Total ion current chromatogram of IdeS-digested Adalimumab. The three subunits (from left to right, Fc/2, Lc, and Fd) are baseline resolved. Broadband Orbitrap FTMS mass spectra recorded with 15 000 resolution at m/z 400 show the charge state distributions of the three subunits, insets. Arrows indicate the center of the employed precursor ion isolation window for ETD MS/MS. The bottom left inset displays a baseline resolved single charge state of the Fc/2 subunit with G0F glycosylation, recorded by LC−MS with 120 000 resolution at m/z 400.



MATERIALS AND METHODS Chemicals. Water, acetonitrile (ACN), trifluoroacetic acid (TFA), and isopropanol (IPA) were purchased in LC−MS purity grade. Water and ACN were obtained from Fluka Analytical (Buchs, Switzerland), formic acid (FA) from Merck (Zug, Switzerland), IPA from Thermo Fisher Scientific (Switzerland), and guanidinium chloride (GdnCl) from Carl Roth (Germany). Tris(2-carboxyethyl)phosphine, tert-butyl hydroperoxide, hydrogen peroxide, and iron chloride were purchased from Sigma Aldrich (Buchs, Switzerland). Chemical Mild Oxidation of IgGs. Therapeutic monoclonal antibodies of the IgG1 class, Adalimumab (Humira, Abbot Laboratories), Bevacizumab (Avastin, Genentech/ Roche), and Trastuzumab (Herceptin, Genentech) were obtained as the European Medicines Agency approved versions and formulations, available commercially to the general public. To benchmark PTM identification by middle-down ETD MS, Adalimumab was oxidized prior to IdeS proteolysis. Two previously described protocols were employed.26 The first one required the incubation of the IgG with 1 mM of hydrogen peroxide and 60 μM of FeCl3 at 37 °C for 19 h. In the second, the IgG was treated with 1.4% of tert-butyl hydroperoxide for 19 h at room temperature. IdeS Digestion. IdeS (FabRICATOR, Genovis, Lund, Sweden) digestion of IgGs was performed in formulation buffers. One unit of IdeS was added to each microgram of IgG and left to react for 30 min at 37 °C. Then, IgGs were denatured and reduced by incubation with 6 M GdnCl and 30 mM TCEP at room temperature for 30 min. Finally, the reaction was quenched by acidifying the solution to 1% TFA.

For analysis, samples were diluted with 0.1% FA in water to a final concentration of 1 μg/μL. Liquid Chromatography−Mass Spectrometry. The chromatographic separation of IgG proteolytic fragments was performed using an Ultimate 3000 LC system (Thermo Scientific, Amsterdam, The Netherlands) under UPLC conditions. A combination of reversed phase C4 trap-column (Acquity UPLC PrST C4 VanGuard precolumn, 2.1 mm × 5 mm, particle size 1.7 μm, pore size 300 Å, Waters, BadenDättwil, Switzerland) and C4 column (Acquity UPLC PrST C4, 1 mm × 150 mm, particle size 1.7 μm, pore size 300 Å, Waters) was employed to ensure online IgG fragment desalting and separation. For each injection, 1 μg of digestion product was loaded on the column, heated at 65 °C. After initial loading at 5% solution B (organic phase), a gradient of solution B from 15 to 45% in 15 min was used at a flow rate of 100 μL/min. Solution A consisted of 0.1% of FA in water, whereas solution B was composed of 39.9% IPA, 60% ACN, and 0.1% FA. The LC column outlet was online coupled with the electrospray (ESI) source of the mass spectrometer. MS experiments were performed on an ETD-enabled hybrid linear ion trap highfield Orbitrap FT mass spectrometer (LTQ Orbitrap Elite FTMS, Thermo Scientific, Bremen, Germany). Separate experiments were dedicated to record broadband mass spectra and ETD tandem mass spectra. Instrumental parameters were set as follows: S-lens rf level was set to 70%, the temperature of heated transfer capillary was 350 °C, the microESI source (IonMax source, Thermo Scientific) was used with a 3.7 kV potential, and the sheath gas was set to 20 and auxiliary gas to 10 arbitrary units. All the mass spectra were acquired using ion 3007

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry

Article

as displayed in the figure insets, present over the selected m/z window (i.e., m/z 200−2000) extended charge state envelopes centered around charge states 24+ to 26+, with maximum charge state exceeding 30+. A similar LC separation has been achieved for fragments derived from Bevacizumab and Trastuzumab, although, as expected, retention times of Lc and Fd fragments, containing IgG variable regions, are different among the three antibodies (Supporting Information, Figure S1). Importantly, the applied chromatographic conditions, with the UPLC C4 column heated at 65 °C, ensured high elution reproducibility (data not shown). The three fragments have masses in the range of 23−25 kDa. In the case of Adalimumab, for instance, theoretical average masses (neutral) of nonmodified fragments are the following: Fc/2, 23 887.08 Da; Lc, 23 412.17 Da; and Fd, 25 458.63 Da. We observed and determined by MS 1 two classical modifications, well described in the literature, located on the Fc/2 fragment (of all the three antibodies), which is Nglycosylated and presents often C-terminal lysine clipping. Major glycosylations could be detected: in the case of Adalimumab, the most intense peak corresponds to Fc/2 with G0F glycosylation, followed by G1F with an intensity of about one-third of the base peak.32 The mass accuracy achieved with the adopted MS settings (see Material and Methods) at 15 000 resolution (at m/z 400) was ∼1 Da. With regard to mass accuracy, on the LC time scale it was also possible to fully resolve the isotopic distribution of the fragments using a resolution setting of 120 000 (at m/z 400), as shown in Figure 2 for Fc/2 fragment with G0F glycosylation of Adalimumab. Nevertheless, as a general consideration about the current state of available instrumentation, it is fundamental to recall that the advantage of isotopically resolved distributions of polypeptides relies mainly on the determination of the monoisotopic mass. According to calculations based on model proteins, the dynamic range needed for simultaneously detecting the most intense peak in the isotopic distribution and the monoisotopic one is of ∼1032 for an intact IgG and ∼6 × 104 for the ∼29 kDa carbonic anhydrase. Therefore, considering the molecular weight of the IgG fragments, close to that of carbonic anhydrase, the utility of isotopically resolving the here studied fragments is currently limited given that no current FTMS mass analyzer exceeds 4 orders of magnitude in the spectral dynamic range. In addition, further complications are due to the restricted number of mass spectra recorded during a single LC experiment, the possible coelution of protein adducts (e.g., Na+ adducts), and presence of proteoforms with small differences in mass. As an example, Chen et al. showed that the commonly employed method for monoisotopic mass estimation based on averagine (i.e., average amino acid with the formula C4.9384H7.7583N1.3577O1.4773S0.0417)33 can lead to an error of ∼0.3 Da for isotopically resolved 15 kDa protein RNase A.34 ETD MS/MS of IgG Fragments. ETD of IdeS-produced IgG subunits was performed in two different fashions. Proteolytic subunits derived from Bevacizumab and Trastuzumab were subjected to ETD in a “proteomic” fashion, which corresponds to the isolation of precursor ions in a single charge state. These experiments proved the efficiency of ETD and facilitated tuning the instrumental parameters for further studies described below. Conversely, ETD MS/MS of Adalimumab was aimed at maximizing the sequence coverage, for direct comparison with top-down MS results, and was thus performed exploiting the high chromatographic reproducibility of subunit elution. The

detection in the Orbitrap FTMS, in the m/z range 200−2000. For broadband and tandem mass spectrometry, we both reduced the gas (N2) “delta pressure” in the Orbitrap detector region to 0.1 × 10−10 Torr and applied “HCD trapping”, which is a temporary ion storage in the HCD cell before ion transmission to the Orbitrap mass analyzer through the Ctrap.29 Broadband mass spectra were recorded with either 15 000 or 120 000 resolution at 400 m/z, with a target value for the automatic gain control (AGC) of 1 million charges in either MS or MS/MS modes. For ETD experiments, precursor ions were isolated in the high-pressure chamber of the LTQ and subsequently subjected to ETD MS/MS. The AGC target value for fluoranthene radical anions was set to 7−8 × 105 charges, with an anion maximum injection time of 50 ms. ETD duration (i.e., ion−ion interaction time) was progressively increased from 3 to 9 ms in consecutive experiments. Product ion detection in the Orbitrap mass analyzer was performed with 120 000 resolution at 400 m/z (eFT enabled). All Orbitrap FTMS scans were recorded averaging 10 microscans to improve the signal-to-noise ratio (SNR). Isolation windows for ETD of IgG fragments included one charge state per precursor ion (isolation width, 15 Th) in the case of Bevacizumab and Trastuzumab or multiple charge states for Adalimumab (isolation width, 100 Th and wider). Data Processing and Tandem MS Analysis. Data were analyzed both as single LC−MS/MS runs and after additional data processing aimed at improving SNR of tandem mass spectra. In the latter case, time-domain (transient) signals recorded in separate LC−MS/MS experiments were processed as previously described for top-down LC−MS/MS of mAbs.17 Briefly, Orbitrap FTMS transient signals were first recorded in MIDAS .dat format;30 these were then grouped according to the IgG fragment type and duration of ETD MS/MS, averaged, and finally subjected to time-to-frequency conversion with the FT procedure. The resulting standard Thermo .RAW files could be then opened and processed with commercial XCalibur software (Thermo Scientific) and were thus ready for the data analysis. For each ETD duration, a summed mass spectrum for each IgG fragment was obtained; in addition, a total tandem mass spectrum was built by averaging all the transients (i.e., transients derived from different ETD duration experiments) available for a single IgG fragment. Data analysis was performed using Xtract and ProSightPC 3.0 (Thermo Scientific).31 First, Xtract was used for tandem mass spectra deconvolution, peak centroiding, and peak picking. Then, cleavage sites were assigned with ProSightPC using 15 ppm tolerance. Both methionine and tryptophan were considered as possible oxidation sites. Complementary c- and ztype product ions were searched separately from b- and y-type product ions.



RESULTS AND DISCUSSION LC−MS of IdeS-Derived IgG Fragments. Monoclonal antibodies digested with IdeS were subjected to chemical denaturation and disulfide bond reduction prior to LC−MS analysis. Notably, no alkylation of reduced thiols of cysteine residues was necessary, as the MS analysis was performed immediately after sample preparation and IgG fragments were maintained under acidic conditions (see Material and Methods), which helped preventing the reformation of disulfide bridges. Figure 2 shows a typical total ion current (TIC) chromatogram of the three proteolytic fragments, namely, Fc/2, Fd, and Lc, of Adalimumab. These are baseline separated and, 3008

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry

Article

Figure 3. Fragmentation maps of IdeS digested Adalimumab after spectral SNR improvement by averaging of transients from different LC runs. Product ions assigned from different ETD data sets are represented according to the color legend. Both c- and z-type as well as y-type product ions were assigned. CDRs are highlighted in yellow, whereas the N-glycosylation site on the Fc/2 fragment is indicated in green. Right bottom inset shows summary of sequence coverage for the corresponding fragments of Adalimumab.

instrument was operating in MS/MS mode only during selected time windows corresponding to the elution times of the three fragments. In this case, a larger isolation window was used, for including ∼5 highly charged precursors and increasing ETD efficiency. Isolation windows were selected on the base of the charge state envelopes of each fragment (i.e., using the information previously obtained by MS1) and centered as indicated by the arrows in the insets of Figure 2. Note, in topdown and middle-down experiments, the sequence coverage is calculated as the ratio of assigned cleavage sites to the total number of possible backbone cleavage sites, whereas in bottomup experiments, the identified peptides are used without regard to the number of cleavage sites assigned after evaluation of MS/ MS data. As expected, the highest sequence coverage was obtained with the acquisition parameters previously optimized for Adalimumab analysis. The corresponding fragmentation map is represented in Figure 3. As reported in the figure, a final coverage of almost 70% was obtained for Fc/2 and Lc fragments, whereas a slightly lower coverage, of nearly 60%, was obtained for the heavier Fd fragment, presumably because of different charge localization and lower charge-overmass unit ratio relative to Fc/2. The final sequence coverage was calculated accounting all the cleavage sites assigned through the analysis of different experimental data sets, derived from experiments where the ETD duration spanned from 3 to 9 ms. For each data set, 4 to 10 LC runs were acquired. In each LC run, up to 90 transients (microscans) per fragment were recorded (the number varies among the fragments, being dependent on the elution time frame). In total, ∼3000 transients (microscans) were averaged. Importantly, all the complementarity determining regions (CDRs) were sequenced, and the position of the glycosylation site on the Fc/2 was

confirmed, as well as the Lys-clipping on the same fragment, Figure 3. For Bevacizumab and Trastuzumab, the averaging of a similar number of transients resulted in slightly lower sequence coverage (Table S1, Supporting Information). Nevertheless, sequencing of CDRs and confirmation of the glycosylation site as well as Lys-clipping were obtained (Figure S2, Supporting Information). Finally, it is noteworthy that the sequence coverage achieved with a single LC run (related ETD mass spectra are reported in Figure S3, Supporting Information) is considerably high for ∼25 kDa polypeptides with sharp elution peaks (of 20−40 s): almost 50% sequence coverage is reached for Fc/2 and Lc and ∼30% for Fd of Adalimumab (Table S2, Supporting Information). Importantly, the advantage in terms of sequence coverage obtainable operating the mass spectrometer under reduced HCD cell pressure and with HCD trapping enabled is dramatic for a single ETD LC−MS/MS experiment, as summarized in Table S2 and visually reported in Figure S4, Supporting Information. Recently, Rose et al. described an improvement of ETD efficiency by modification of the HCD collision cell of the LTQ-Orbitrap FTMS to enable in-cell ETD.35 Earley et al. described a front-end reagent anion source for ETD allowing for multiple fills of C-trap with ETD product ions.36 These instrumental improvements of ETD MS/MS aiming to enhance ETD efficiency should allow reaching higher sequence coverage in a single LC−MS/MS run with fewer scans needed to be averaged. Middle-Down LC−MS/MS Analysis of Oxidized IgG. The chromatographic separation of fragments derived from IdeS digestion of oxidized Adalimumab resulted in baselineresolved elution peaks. Figure 4 illustrates a comparison between fragments of IgG oxidized with tert-butyl hydro3009

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry

Article

Figure 4. Middle-down MS-based analysis of IgG oxidation. Top panel: comparison between TIC chromatograms of control (black line) and oxidized (red line) Adalimumab. The oxidized Fc is eluting earlier than the nonmodified counterpart. Middle panel: expanded view of broadband mass spectra showing the isotopically resolved Fc/2 subunit (mass spectra recorded with 120 000 resolution at m/z 400). This panel, made by overlapping mass spectra of “control” and “oxidized” IgGs, illustrates the presence of both singly and doubly oxidized Fc/2 subunits. Bottom panel: 3 ms ETD Orbitrap FTMS mass spectrum (single LC run) of oxidized Fc (isolation window 120 Th). The left inset shows a product ion confirming the position of oxidized Met16. The right inset is a scheme of identified product ions, with the positions of oxidized Met indicated by arrows.

peroxide (red line) and those of the “native” IgG (black line). It is apparent how only the Fc/2 shows a different chromatographic behavior in the two cases, Figure 4 top panel. Effectively, whereas Lc and Fd where not oxidized (as determined by MS1 analysis), Fc/2 presented one or two oxidations, with the doubly oxidized species being about 4-fold more abundant than the mono-oxidized species, Figure 4 middle panel. Note that in the control experiment, no oxidation was detected which proves that no artifact oxidation was induced by the IdeS sample preparation (data not shown). By a single 3 ms ETD LC−MS/MS run it was possible to determine the sites of oxidation, Figure 4 bottom panel. These are located at two methionine residues, namely, Met16 and Met192, as schematically indicated on the right inset in Figure 4 bottom panel and further detailed in Figure S5, top panel, Supporting Information. A more extended sequence coverage was then obtained by averaging 300 transients (microscans) deriving from ETD experiments with a duration of 3 ms and 300 other

transients (microscans) from 5 ms ETD experiments, Figure S5, bottom panel, Supporting Information. These results can be compared to the higher-energy collision dissociation (HCD)37 MS/MS results reported in the literature for similar ∼25 kDa antibody subunits.38 In this study, the two oxidized forms of the Fc/2 were also detected by MS1. However, HCD MS/MS yielded 8% sequence coverage of the Fc/2 subunits with only one distant b fragment (b48) assigning the oxidation at Met30 and nine distant y fragments assigning oxidation at Met206. Our results show that ETD fragmentation increased significantly sequence coverage to more than 30% in one run with 3 ms ETD to 47.1% when averaging ∼10 runs of 3 and 5 ms ETD. Note that these percentages are influenced by the contemporary analysis of the two forms of oxidized Fc/2. Cleavage sites adjacent to the oxidized methionines and other potential oxidation sites (tryptophanes) are obtained giving more confident assignment to the oxidation sites. 3010

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry

Article

Figure 5. TIC chromatogram of an equimolar mixture of IdeS cleaved monoclonal antibodies Adalimumab, Bevacizumab, and Trastuzumab. The inset shows the Orbitrap FTMS mass spectrum of coeluting Lc of Bevacizumab and Fd of Trastuzumab.

Middle-down Analysis of IgG Mixtures. As displayed in Figure 5, we applied the above-described LC−MS method for the analysis of an equimolar mixture of monoclonal IgGs digested with IdeS. Adalimumab, Bevacizumab, and Trastuzumab were first pooled together and subjected to IdeS digestion, reduction, and subsequent LC−MS analysis (with 15 000 resolution at m/z 400). This validation experiment was aimed at verifying the actual possibilities offered by reversed phase UPLC for the future analysis of either IgG mixtures of biological origin (e.g., polyclonal antibodies) or so-called therapeutic “IgG cocktails”. While Fc/2 fragments share a very high sequence homology (over 90%) and are therefore not separated, most of the other fragments can be isolated. Two fragments, Lc of Bevacizumab and Fd of Trastuzumab, are coeluting, but this does not impair the recording of high SNR broadband mass spectra for the two (Figure 5, left inset). Single charge state precursor ion isolation and sequencing with ETD MS/MS should thus be possible as indicated by the data presented here.

informative ETD mass spectra for middle-down mass spectrometry without the need of averaging a high number of scans or transients. The particular advantage of the approach described here consists of the limited sample preparation reducing potential introduction of artifacts. The described method proved to be efficient in detecting and assigning IgG oxidation sites, which makes it a quick and easily implemented technique for mAbs′ oxidation assessment. Finally, we demonstrate that the 25 kDa molecular weight range of the IgG fragments under study should enable an efficient structural analysis of simple, up to three IgGs considered here, antibody mixtures. We envision application of the developed approach for structural analysis of more complex mAbs mixtures, as required by modern drug discovery strategies.



ASSOCIATED CONTENT

S Supporting Information *



Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

CONCLUSIONS To improve structural analysis of monoclonal antibodies, IgGs, we introduce a middle-down mass spectrometry approach consisting of quick (about 1 h) IgG sample preparation protocol followed by 25 kDa polypeptide analysis with reversed phase LC-based protein separation online coupled to highresolution Orbitrap FTMS with ETD capability. The reported results show that the application of the described middle-down approach doubles the IgG sequence coverage previously obtained with top-down MS. The recent and continuous development of MS/MS instrumentation targeting ETD efficiency improvement should enable the acquisition of highly



AUTHOR INFORMATION

Corresponding Author

*Address: EPFL ISIC LSMB, BCH 4307, 1015 Lausanne, Switzerland. E-mail: yury.tsybin@epfl.ch. Author Contributions ∥

L.F. and D.A. contributed equally to this work.

Notes

The authors declare no competing financial interest. 3011

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012

Analytical Chemistry



Article

(21) Gadgil, H. S.; Bondarenko, P. V.; Pipes, G.; Rehder, D.; McAuley, A.; Perico, N.; Dillon, T.; Ricci, M.; Treuheit, M. J. Pharm. Sci. 2007, 96, 2607−2621. (22) Chevreux, G.; Tilly, N.; Bihoreau, N. Anal. Biochem. 2011, 415, 212−214. (23) Ryan, M. H.; Petrone, D.; Nemeth, J. F.; Barnathan, E.; Bjorck, L.; Jordan, R. E. Mol. Immunol. 2008, 45, 1837−1846. (24) Wang, B.; Gucinski, A. C.; Keire, D. A.; Buhse, L. F.; Boyne, M. T. Analyst 2013, 138, 3058−3065. (25) Michalski, A.; Damoc, E.; Lange, O.; Denisov, E.; Nolting, D.; Muller, M.; Viner, R.; Schwartz, J.; Remes, P.; Belford, M.; Dunyach, J. J.; Cox, J.; Horning, S.; Mann, M.; Makarov, A. Mol. Cell. Proteomics 2012, 11, O111 013698. (26) Boyd, D.; Kaschak, T.; Yan, B. J. Chromatogr., B 2011, 879, 955−960. (27) Teshima, G.; Li, M. X.; Danishmand, R.; Obi, C.; To, R.; Huang, C.; Kung, J.; Lahidji, V.; Freeberg, J.; Thorner, L.; Tomic, M. J. Chromatogr., A 2011, 1218, 2091−2097. (28) Kuo, T. T.; Aveson, V. G. mAbs 2011, 3, 422−430. (29) Rosati, S.; Rose, R. J.; Thompson, N. J.; van Duijn, E.; Damoc, E.; Denisov, E.; Makarov, A.; Heck, A. J. Angew. Chem. 2012, 51, 12992−12996. (30) Senko, M. W.; Canterbury, J. D.; Guan, S.; Marshall, A. G. Rapid Commun. Mass Spectrom. 1996, 10, 1839−1844. (31) Zamdborg, L.; LeDuc, R. D.; Glowacz, K. J.; Kim, Y. B.; Viswanathan, V.; Spaulding, I. T.; Early, B. P.; Bluhm, E. J.; Babai, S.; Kelleher, N. L. Nucleic Acids Res. 2007, 35, W701−W706. (32) Nallet, S.; Fornelli, L.; Schmitt, S.; Parra, J.; Baldi, L.; Tsybin, Y. O.; Wurm, F. M. New Biotechnol. 2012, 29, 471−476. (33) Senko, M. W.; Beu, S. C.; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 1995, 6, 229−233. (34) Chen, Y. F.; Chang, C. A.; Lin, Y. H.; Tsay, Y. G. Anal. Biochem. 2013, 440, 108−113. (35) Rose, C.; Russell, J.; Ledvina, A.; McAlister, G.; Westphall, M.; Griep-Raming, J.; Schwartz, J.; Coon, J.; Syka, J. P. J. Am. Soc. Mass Spectrom. 2013, 24, 816−827. (36) Earley, L.; Anderson, L. C.; Bai, D. L.; Mullen, C.; Syka, J. E. P.; English, A. M.; Dunyach, J.-J.; Stafford, G. C.; Shabanowitz, J.; Hunt, D. F.; Compton, P. D. Anal. Chem. 2013, 85, 8385−8390. (37) Olsen, J. V.; Macek, B.; Lange, O.; Makarov, A.; Horning, S.; Mann, M. Nat. Methods 2007, 4, 709−712. (38) Zhang, J.; Liu, H.; Katta, V. J. Mass Spectrom. 2010, 45, 112− 120.

ACKNOWLEDGMENTS We thank Kristina Srzentić, Ü nige A. Laskay, and Alexander A. Makarov for motivating discussion and technical support. We express our sincere gratitude to Thermo Fisher Scientific Inc. for providing us access under license to Orbitrap transient signals. The work was supported by the Swiss National Science Foundation (Projects 200021-125147 and 128357) and the European Research Council (ERC Starting Grant 280271 to Y.O.T.).



REFERENCES

(1) Walsh, G. Nat. Biotechnol. 2010, 28, 917−924. (2) Debaene, F.; Wagner-Rousset, E.; Colas, O.; Ayoub, D.; Corvaia, N.; Van Dorsselaer, A.; Beck, A.; Cianferani, S. Anal. Chem. 2013, 85, 9785−9792. (3) Beck, A.; Wagner-Rousset, E.; Ayoub, D.; Van Dorsselaer, A.; Sanglier-Cianferani, S. Anal. Chem. 2013, 85, 715−736. (4) Beck, A.; Diemer, H.; Ayoub, D.; Debaene, F.; Wagner-Rousset, E.; Carapito, C.; Van Dorsselaer, A.; Sanglier-Cianférani, S. TrAC, Trends Anal. Chem. 2013, 48, 81−95. (5) Weise, M.; Bielsky, M. C.; De Smet, K.; Ehmann, F.; Ekman, N.; Giezen, T. J.; Gravanis, I.; Heim, H. K.; Heinonen, E.; Ho, K.; Moreau, A.; Narayanan, G.; Kruse, N. A.; Reichmann, G.; Thorpe, R.; van Aerts, L.; Vleminckx, C.; Wadhwa, M.; Schneider, C. K. Blood 2012, 120, 5111−5117. (6) Ayoub, D.; Jabs, W.; Resemann, A.; Evers, W.; Evans, C.; Main, L.; Baessmann, C.; Wagner, E.; Suckau, D.; Beck, A. mAbs 2013, 5, 699−710. (7) Ren, D.; Pipes, G. D.; Liu, D. J.; Shih, L. Y.; Nichols, A. C.; Treuheit, M. J.; Brems, D. N.; Bondarenko, P. V. Anal. Biochem. 2009, 392, 12−21. (8) Krokhin, O. V.; Antonovici, M.; Ens, W.; Wilkins, J. A.; Standing, K. G. Anal. Chem. 2006, 78, 6645−6650. (9) Fornelli, L.; Parra, J.; Hartmer, R.; Stoermer, C.; Lubeck, M.; Tsybin, Y. O. Anal. Bioanal. Chem. 2013, 405, 8505−8514. (10) Kellie, J. F.; Tran, J. C.; Lee, J. E.; Ahlf, D. R.; Thomas, H. M.; Ntai, I.; Catherman, A. D.; Durbin, K. R.; Zamdborg, L.; Vellaichamy, A.; Thomas, P. M.; Kelleher, N. L. Mol. Biosyst. 2010, 6, 1532−1539. (11) Tran, J. C.; Zamdborg, L.; Ahlf, D. R.; Lee, J. E.; Catherman, A. D.; Durbin, K. R.; Tipton, J. D.; Vellaichamy, A.; Kellie, J. F.; Li, M. X.; Wu, C.; Sweet, S. M. M.; Early, B. P.; Siuti, N.; LeDuc, R. D.; Compton, P. D.; Thomas, P. M.; Kelleher, N. L. Nature 2011, 480, 254−U141. (12) Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528−9533. (13) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1998, 120, 3265−3266. (14) Zubarev, R. A.; Horn, D. M.; Fridriksson, E. K.; Kelleher, N. L.; Kruger, N. A.; Lewis, M. A.; Carpenter, B. K.; McLafferty, F. W. Anal. Chem. 2000, 72, 563−573. (15) Tsybin, Y. O.; Fornelli, L.; Stoermer, C.; Luebeck, M.; Parra, J.; Nallet, S.; Wurm, F. M.; Hartmer, R. Anal. Chem. 2011, 83, 8919− 8927. (16) Mao, Y.; Valeja, S. G.; Rouse, J. C.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2013, 85, 4239−4246. (17) Fornelli, L.; Damoc, E.; Thomas, P. M.; Kelleher, N. L.; Aizikov, K.; Denisov, E.; Makarov, A.; Tsybin, Y. O. Mol. Cell. Proteomics 2012, 11, 1758−1767. (18) Zhang, Z.; Pan, H.; Chen, X. Mass Spectrom. Rev. 2009, 28, 147−176. (19) Yan, B.; Valliere-Douglass, J.; Brady, L.; Steen, S.; Han, M.; Pace, D.; Elliott, S.; Yates, Z.; Han, Y.; Balland, A.; Wang, W.; Pettit, D. J. Chromatogr., A 2007, 1164, 153−161. (20) 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, 165−174. 3012

dx.doi.org/10.1021/ac4036857 | Anal. Chem. 2014, 86, 3005−3012