Technical Note pubs.acs.org/ac
Primary Sequence Confirmation of a Protein Therapeutic Using Top Down MS/MS and MS3 Michaella J. Levy, Ashley C. Gucinski,* and Michael T. Boyne, II∥ U.S. Food and Drug Administration, Center for Drug Evaluation and Research, Office of Testing and Research, Division of Pharmaceutical Analysis, 645 S. Newstead Ave., St. Louis, Missouri 63110, United States S Supporting Information *
ABSTRACT: Mass spectrometry has gained widespread acceptance for the characterization of protein therapeutics as a part of the regulatory approval process. Improvements in mass spectrometer sensitivity, resolution, and mass accuracy have enabled more detailed and confident analysis of larger biomolecules for confirming amino acid sequences, assessing sequence variants, and characterizing post translational modifications. This work demonstrates the suitability of a combined approach using intact MS and multistage top down MS/MS analyses for the characterization of a protein therapeutic drug. The protein therapeutic granulocyte-colony stimulating factor was analyzed using a Thermo Fusion Tribrid mass spectrometer using a multistage top down MS approach. Intact mass analysis identified the presence of two disulfide bonds based on exact mass shifts while a combined collision induced dissociation (CID), higher-energy collisional dissociation (HCD), and electron transfer dissociation (ETD) MS/MS approach obtained 80% protein sequence coverage. Isolating MS/MS fragments for MS3 analysis using HCD or CID increased the sequence coverage to 89%. 95% sequence coverage was obtained by reducing human granulocyte-colony stimulating factor (G-CSF) prior to MS/MS and MS3 analysis to specifically target the residues between the disulfide bonds. The use of this combined intact MS and multistage top down MS approach allows for rapid and accurate determination of the primary sequence of a protein therapeutic drug product.
M
protein. The time intensive workflow as well as the digestion step, which loses important intact sequence information and may also introduce process-related artifacts inherent to bottom up mass spectrometry sample preparation, creates a demand for alternative methods to address these issues.8 One such approach is top down mass spectrometry, which directly fragments an intact protein in the mass spectrometer without any digestion prior to analysis. Large scale application of MSn peptide analysis was made possible by trapping instruments using either collision induced dissociation (CID) or electron transfer dissociation (ETD).7,9,10 MSn has been used to determine specific peptide phosphorylation sites,6,10 as well as the identification of sequence variants of therapeutic proteins.11 Instrument improvements have facilitated increased application of top down approaches for analysis of protein systems.12,13 Historically, the sequence coverage obtained from a top down MS approach dropped rapidly for proteins containing more than ∼100 amino acids; additionally, the fragment ions that could be generated were primarily at the termini14 or
ass spectrometry plays a key role in the regulation of protein therapeutics and monoclonal antibodies providing the foundation for identification and characterization of the active pharmaceutical ingredient, comparability and similarity studies associated with manufacturing changes, surveillance and adulteration studies of gray market products, bioprocess development, and increasingly bioanalysis.1−5 To ensure the quality of these drugs throughout development and licensure, structural characterization using highly sensitive analytical techniques capable of determining structural variations is necessary. Primary sequence confirmation is an initial step in this process used to verify that the genetic sequence was correctly translated and transcribed and that the expected protein is consistently produced across lots. Vital to the application of LC-MS for peptide or protein sequence confirmation is tandem mass spectrometry (MS/MS). Bottom up mass spectrometry is commonly used for peptide or protein primary sequence determination.6,7 This approach subjects the protein(s) of interest to enzymatic digestion followed by a chromatographic separation of peptides prior to MS and/or MS/MS analysis. The bottom up LC-MS/MS approach has been widely utilized by the pharmaceutical industry to generate a unique chromatographic fingerprint and/ or obtain complete protein sequence coverage of a purified This article not subject to U.S. Copyright. Published XXXX by the American Chemical Society
Received: March 23, 2015 Accepted: June 18, 2015
A
DOI: 10.1021/acs.analchem.5b01113 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Technical Note
structurally flexible regions of the protein.15 Electronic dissociation techniques can provide complementary information and de novo sequencing of small peptides,11,16,17 but sequencing of larger proteins still remains a challenge. Alternative activation strategies such as in source decay,18 activated electron capture/transfer,9,19 or photon-based dissociation have been proposed to sequence larger protein molecules.20−25 While the complementary nature of CID, higher-energy collisional dissociation (HCD), and ETD may increase the sequence coverage obtained from a MS/MS experiment, we propose to combine that experiment with an additional stage of fragmentation. Integrating a targeted MS3 experiment can allow for sequence coverage in difficult to sequence regions of the protein, allowing for a more complete sequence characterization of the entire protein therapeutic. Human granulocyte-colony stimulating factor (G-CSF), an 18.8 kDa protein therapeutic used in the treatment of nonlymphomic leukemia, was selected as a model system to demonstrate the suitability of top down MS/MS and MS3 for the confident sequence determination of a protein therapeutic drug product. G-CSF is produced recombinantly in E. coli where the native N-terminal leader sequence is replaced by a single methionine residue. Unlike its native counterpart, recombinant G-CSF is not glycosylated; however, studies have shown no difference in activity between the native and recombinantly produced G-CSF products.26 Historically, the confirmation of protein sequence of G-CSF has relied heavily on bottom up mass spectrometry following digestion by proteolytic enzymes and LC-MS/MS analysis.27−29 Previous reports of top down MS sequencing for G-CSF obtained 25% sequence coverage on an Orbitrap platform when using a single fragmentation method30 and approximately 62% coverage when multiple fragmentation methods were used in combination.31 Our top down MS/MS and targeted MS3 approach was used to analyze G-CSF in order to evaluate the suitability of this approach for primary sequence confirmation of a model protein therapeutic system.
After reduction, the sample was further purified as described above with an Amicon Ultra-0.5 Centrifugal Filter and Zeba spin desalting column. The sample was diluted to ∼5 μM in 40% ACN, 1% FA with 5 mM TCEP in Optima grade water prior to analysis. Mass Spectrometry. Five μM G-CSF was directly infused via a syringe pump into an Orbitrap Fusion Tribrid mass spectrometer (Thermo Scientific, Bremen, Germany). Source voltage was set to 3500 V, sheath gas to 2.0 (arbitrary units), ion transfer tube to 350 °C, and vaporizer temperature to 50 °C. For intact mass analysis, the Orbitrap analyzer was set at a resolution of 60 000 at 400 m/z with a S-lens RF of 60%, an AGC target of 2.0 × 105 with a maximum injection time of 100 ms and 1 microscan. The pressure in the HCD cell was decreased to 4 mTorr to decrease the pressure change between the ion trap and Orbitrap for improved resolution of small proteins. MS3 spectra were collected at an Orbitrap resolution of 60 000 or 120 000 at 400 m/z for reduced or nonreduced GCSF, respectively. The MS/MS precursor window used was wide (240 m/z) to isolate multiple charge states of the intact protein simultaneously. MS3 spectra were generated by isolating the precursor in the quadrupole using an isolation window between 2 and 5 m/z, and the number of microscans varied from 2 to 10 depending on the individual spectrum. Final spectra were generated from 500 summed scans. Data Analysis. Summed spectra were interpreted by using the Xtract function embedded within the Protein Deconvolution 3.0 software (Thermo Scientific, Bremen, Germany) to generate a list of monoisotopic masses. The masses were fit to the known G-CSF sequence using ProSight Lite (Northwestern University, Chicago, IL) with a mass tolerance of 10 ppm error and were verified manually on the basis of predicted fragment masses generated in Protein Prospector v 5.12.4 (University of California San Francisco).
■
RESULTS Top Down MS/MS of G-CSF. The intact mass of the protein was determined (Table 1) and the 5 most abundant
■
METHODS AND MATERIALS Sample Preparation. Granulocyte colony-stimulating factor (G-CSF) was purchased for analysis (Neupogen, Amgen, Thousand Oaks, CA) from a local pharmacy. Prior to analysis, excipients were removed from the formulated product by diluting with Optima grade water (Fisher, Waltham, MA) and applying the protein to a 10 000 Da MWCO Amicon Ultra-0.5 Centrifugal Filter (Millipore, Darmstadt, Germany), which was centrifuged at 10 000g for 5 min at room temperature. The flow through was discarded, and the solution on top of the membrane was diluted to 400 μL with Optima grade water and centrifuged. This step was repeated a total of 3 times. To further desalt the sample, the G-CSF in an approximate volume of 100 μL was applied to a Zeba spin desalting column (7000 MWCO, Thermo Scientific, Rockford, IL) which had been equilibrated previously with Optima grade water. The sample in water was diluted to ∼5 μM in 40% acetonitrile (ACN), 1% formic acid (FA) (Optima grade) prior to MS analysis. For experiments performed using reduced G-CSF, excipients were removed from drug product formulation using an Amicon Centrifugal Filter as described above. The sample was reduced by incubating approximately 40 μM G-CSF in Optima grade water with 2.4 M guanidine hydrochloride and 40 mM tris(2carboxyethyl)phosphine (TCEP) for 1 h at room temperature.
Table 1. Theoretical and Observed Masses of Intact G-CSF
G-CSF reduced GCSF
theoretical monoisotopic mass (Da)a
observed monoisotopic mass (Da)
error (ppm)
18 786.6761 18 790.7074
18 786.7389 18 790.8229
3.3 6.1
a
Theoretical mass calculated using the NIST Mass and Fragment Calculator v1.3.34
charge states (16−20+) were isolated for MS/MS analysis. Top down MS/MS was performed using all three dissociation methods at various collision energies (CID and HCD, 15−50% NCE) and reaction times (ETD, 4−10 ms) to confirm the primary sequence of G-CSF. Representative MS/MS spectra generated from CID at 40% NCE, HCD at 10% NCE, and ETD using a 4 ms reaction time are shown in Supplemental Figures 1A−C, Supporting Information. MS/MS allows for the location of the disulfide bonds to be clearly identified, highlighted by the absence of fragments between the cysteine residues involved in each disulfide bond (Figure 1). The mass of the fragment ions involved in disulfide bonds was shifted by a mass of 2.016 or 4.032 Da from the unmodified sequence to include one or two disulfide bonds, respectively. ETD data were B
DOI: 10.1021/acs.analchem.5b01113 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Technical Note
Figure 1. Fragment map of G-CSF ions generated by using CID (blue), HCD (green), and ETD (red). Cysteines involved in disulfide bonds are shown in yellow. Fragments generated by ETD were fit to allow for the disruption of the disulfide bonds.33
searched against the sequence with and without intact disulfide bonds as previous studies have reported that ETD is capable of disrupting disulfide bonds.32,33 The location of the disulfide bonds between the predicted cysteine residues was consistent with those cited by the manufacturer. The sequence coverage of G-CSF obtained from MS/MS using three complementary dissociation methods, CID, HCD, and ETD, was 79%, excluding the residues between disulfide bonds. Top Down MS3 Analysis. Three MS/MS fragment ions were isolated and fragmented further (MS3) as shown in the workflow in Figure 2. Both b and y ions were selected for MS3 to obtain additional fragment ions that would increase the sequence coverage obtained for G-CSF. MS3 of the b92 ion, present at a relative intensity of ∼14.5% in the MS/MS spectrum, was performed by isolating the ion in the quadrupole with an isolation window of ±3 m/z and fragmented using either CID at 32% NCE or HCD at 25% NCE. The corresponding MS3 spectrum of the b92 ion fragmented by HCD at 25% NCE ([M + 9H]9+, m/z 1089.9) is shown in Figure 2 (bottom panel). Similar MS3 spectra were generated by isolating and fragmenting the y74 ([M + 7H]7+, m/z 1156.9) and y78 ([M + 7H]7+, m/z 1213.5) ions from the HCD at 10% NCE MS/MS spectrum, though present at lower relative abundances (5.8% and 7.1%, respectively), by using HCD. Combining the fragment ions obtained from MS3 analyses with the fragments detected from MS/MS analyses (Figure 1) increased the total observed sequence coverage to 87%, excluding predicted fragments for residues between the disulfide bonds (Figure 3). While fragments were observed for residues between the most N-terminal disulfide bond by ETD, the disulfide bonds between C37−C43 and C65−C75 were retained during both the MS/MS and MS3 experiments when fragmented with CID or HCD; this prevented complete confirmation of amino acid sequences between those disulfide bonds without subjecting the protein to disulfide bond reduction before MS/MS analysis. MS/MS and MS3 of Reduced G-CSF. G-CSF was denatured and reduced prior to MS analysis in order to obtain fragments between the residues within the disulfide bonds. MS/ MS spectra were generated by using CID (50% and 75% NCE), HCD (10, 20, 30% NCE), and ETD with 2, 5, and 10 ms reaction times. The fragment map based on MS/MS experiments is shown in Figure 4. MS3 spectra were generating by isolating and fragmenting the b57 and b92 MS/MS ions by using HCD at 32% NCE and 25% NCE, respectively. Additionally, MS3 spectra of the y47, y74, y78, and y81 ions were generated by using HCD at various normalized collision energies. Fragments were observed at 88% of the positions between the disulfide bonds. The combination of observed fragments from the control and reduced G-CSF yielded a total sequence coverage of 95% (Figure 5). Overall, 8 bonds remained unbroken with 3 of these bonds having a proline N-terminal to the cleavage site
Figure 2. General workflow for the top down sequencing of G-CSF is shown with the intact mass spectra in the top panel with the crystal structure of G-CSF inset. The charge states isolated for MS/MS analysis are boxed in blue. The middle panel shows the MS/MS spectra resulting from fragmentation by HCD 10% NCE with the b92 ion (9+) isolated for MS3 analysis inset. Intact G-CSF is indicated by the red asterisk. The MS3 spectrum resulting from fragmentation by HCD 25% NCE is shown in the bottom panel.
Figure 3. Composite fragment ion map for G-CSF generated from MS/MS and MS3 analysis. Fragments observed only in MS 3 experiments are indicated by dashed lines.
and 3 with a serine N-terminal to the cleavage site, which has been shown to have lower fragmentation efficiency when studied in a pairwise fashion.35,36 As fragmentation N-terminal to proline and serine residues is disfavored, the absence of C
DOI: 10.1021/acs.analchem.5b01113 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
■
Technical Note
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b01113.
■
Figure 4. Fragment map of reduced G-CSF from MS/MS and MS3 analysis. The MS/MS fragments are shown as solid colored lines using the same colors as in Figure 1. The MS3 ions are shown as black dashed lines.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: (314) 539-2167. Fax: (314) 539-2113. Present Address ∥
M.T.B: Biotechlogic, Inc., Glenview, IL.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This article reflects the views of the authors and should not be construed to represent the FDA’s views or policies.
Figure 5. Combined fragment map of G-CSF from control and reduced samples. Fragments from G-CSF are shown in gray, while fragments obtained from MS/MS and MS3 of reduced G-CSF are shown in color, consistent with the colors used in previous figures (MS/MS solid lines; MS3 dashed).
■
(1) Garcia, B. A. J. Am. Soc. Mass Spectrom. 2010, 21, 193. (2) McLafferty, F. W.; Breuker, K.; Jin, M.; Han, X.; Infusini, G.; Jiang, H.; Kong, X.; Begley, T. P. FEBS J. 2007, 274, 6256. (3) Chirino, A. J.; Mire-Sluis, A. Nat. Biotechnol. 2004, 22, 1383. (4) Qian, W. J.; Jacobs, J. M.; Liu, T.; Camp, D. G., 2nd; Smith, R. D. Mol. Cell. Proteomics 2006, 5, 1727. (5) Wang, B.; Gucinski, A. C.; Keire, D. A.; Buhse, L. F.; Boyne, M. T., 2nd Analyst 2013, 138, 3058. (6) Coon, J. J.; Syka, J. E.; Shabanowitz, J.; Hunt, D. F. BioTechniques 2005, 38, 519. (7) Syka, J. E.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528. (8) Krokhin, O. V.; Antonovici, M.; Ens, W.; Wilkins, J. A.; Standing, K. G. Anal. Chem. 2006, 78, 6645. (9) 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. (10) Jedrychowski, M. P.; Huttlin, E. L.; Haas, W.; Sowa, M. E.; Rad, R.; Gygi, S. P. Mol. Cell. Proteomics 2011, 10, M111 009910. (11) Gucinski, A. C.; Boyne, M. T., 2nd Rapid Commun. Mass Spectrom. 2014, 28, 1757. (12) Meng, F.; Cargile, B. J.; Miller, L. M.; Forbes, A. J.; Johnson, J. R.; Kelleher, N. L. Nat. Biotechnol. 2001, 19, 952. (13) Siuti, N.; Kelleher, N. L. Nat. Methods 2007, 4, 817. (14) Breuker, K.; Jin, M.; Han, X.; Jiang, H.; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 2008, 19, 1045. (15) Wysocki, V. H.; Resing, K. A.; Zhang, Q.; Cheng, G. Methods 2005, 35, 211. (16) Rubakhin, S. S.; Sweedler, J. V. Nat. Protoc. 2007, 2, 1987. (17) Yew, J. Y.; Kutz, K. K.; Dikler, S.; Messinger, L.; Li, L.; Stretton, A. O. J. Comp. Neurol. 2005, 488, 396. (18) Zhang, Z.; Shah, B. Anal. Chem. 2007, 79, 5723. (19) Swaney, D. L.; McAlister, G. C.; Wirtala, M.; Schwartz, J. C.; Syka, J. E.; Coon, J. J. Anal. Chem. 2007, 79, 477. (20) Beardsley, R. L.; Reilly, J. P. J. Am. Soc. Mass Spectrom. 2004, 15, 158. (21) Brodbelt, J. S.; Wilson, J. J. Mass Spectrom. Rev. 2009, 28, 390. (22) Cui, W.; Thompson, M. S.; Reilly, J. P. J. Am. Soc. Mass Spectrom. 2005, 16, 1384. (23) Thompson, M. S.; Cui, W.; Reilly, J. P. J. Am. Soc. Mass Spectrom. 2007, 18, 1439. (24) Wilson, J. J.; Brodbelt, J. S. Anal. Chem. 2006, 78, 6855. (25) Shaw, J. B.; Li, W.; Holden, D. D.; Zhang, Y.; Griep-Raming, J.; Fellers, R. T.; Early, B. P.; Thomas, P. M.; Kelleher, N. L.; Brodbelt, J. S. J. Am. Chem. Soc. 2013, 135, 12646.
fragmentation at those positions supports the identification of those residues. This is consistent with the previously reported sequence coverage obtained using bottom up strategies (96%), suggesting that the combined top down MS/MS and MS3 strategy can provide a successful, rapid alternative for primary sequence determination with comparable performance without the need for time intensive enzymatic digestions or chromatographic separation.
■
REFERENCES
CONCLUSIONS
Intact and multistage top down mass spectrometry can provide highly detailed information on a protein including determination of the primary sequence, identity and location of post translational modifications, and identification of some secondary structural components such as disulfide bond linkages. The identification of these sequence and structural elements has become increasingly important to verify batch to batch consistency of protein therapeutics and/or between products after manufacturing or production changes. In this work, we demonstrated the suitability of using a combined intact MS, top down MS/MS, and MS3 approach for confident sequence determination for the protein therapeutic granulocyte-colony stimulating factor, G-CSF, without digestion or extensive sample preparation. Sequence coverage obtained was consistent with traditional bottom up approaches, suggesting that this workflow provides a viable, rapid alternative for primary sequence determination. Recent improvements in mass spectrometer capabilities enabled isolation of lower abundance fragment ions generated from MS/MS in order to obtain MS3 spectra of sufficient quality to identify additional fragment ions, increasing the total protein sequence coverage detected. This study presents a framework that can be applied to other protein systems. Current work in our laboratory focuses on the optimization of MS3 sequencing of larger protein therapeutics, specifically including monoclonal antibodies. D
DOI: 10.1021/acs.analchem.5b01113 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Technical Note
(26) Souza, L. M.; Boone, T. C.; Gabrilove, J.; Lai, P. H.; Zsebo, K. M.; Murdock, D. C.; Chazin, V. R.; Bruszewski, J.; Lu, H.; Chen, K. K.; et al. Science 1986, 232, 61. (27) Ahmed, K. E.; Chen, W. Q.; John, J. P.; Kang, S. U.; Lubec, G. Amino Acids 2010, 38, 1043. (28) Mo, J.; Tymiak, A. A.; Chen, G. Rapid Commun. Mass Spectrom. 2013, 27, 940. (29) Tsuchida, D.; Yamazaki, K.; Akashi, S. J. Am. Soc. Mass Spectrom. 2014, 25, 1747. (30) Holzmann, J.; Hausberger, A.; Rupprechter, A.; Toll, H. Anal. Bioanal. Chem. 2013, 405, 6667. (31) Levy, M. J.; Gucinski, A. C.; Sommers, C. D.; Ghasriani, H.; Wang, B.; Keire, D. A.; Boyne, M. T., 2nd Anal. Bioanal. Chem. 2014, 406, 6559. (32) Cole, S. R.; Ma, X.; Zhang, X.; Xia, Y. J. Am. Soc. Mass Spectrom. 2012, 23, 310. (33) Wu, S. L.; Jiang, H.; Lu, Q.; Dai, S.; Hancock, W. S.; Karger, B. L. Anal. Chem. 2009, 81, 112. (34) Kilpatrick, E. L.; Liao, W. L.; Camara, J. E.; Turko, I. V.; Bunk, D. M. Protein Expression Purif. 2012, 85, 94. (35) Li, W.; Ji, L.; Goya, J.; Tan, G.; Wysocki, V. H. J. Proteome Res. 2011, 10, 1593. (36) Li, W.; Song, C.; Bailey, D. J.; Tseng, G. C.; Coon, J. J.; Wysocki, V. H. Anal. Chem. 2011, 83, 9540.
E
DOI: 10.1021/acs.analchem.5b01113 Anal. Chem. XXXX, XXX, XXX−XXX
