Analytical and structural characterisation methods such as X-ray crystallography, cryo-electron microscopy and nuclear magnetic resonance (NMR)3,4 can be used to assess protein higher order structure (HOS) and obtain information on folding and disulphide bonds. However, with regard to speed, specificity, quantitation and amount of material required, mass spectrometry coupled to liquid chromatography (LC-MS) is the most efficient method for confirming the proper formation of disulphide bonds; identifying the presence of improperly formed disulphides; and for measuring the extent to which specific cysteines in a protein sequence may not be fully oxidised to disulphides. The discussion below focuses on therapeutic antibodies of the immunoglobulin G (IgG) class, which contain many disulphide bonds within and between chains that stabilise the tertiary structure of such molecules.
Confirmation of proper disulphide bond formation
It is important, particularly for therapeutic proteins such as antibody drugs, to confirm that disulphide bonds have formed properly. This is accomplished by digestion with a specific enzyme (eg, trypsin) and identification of cysteine‑containing disulphide-linked peptides.4-6 Reduction of disulphides, followed by enzymatic digestion, is used for further confirmation, as pairs of disulphide-linked peptides would no longer be present, but the individual peptides are.
Identification of improperly formed disulphides can be accomplished in a similar fashion.7,8 It is quite simple, albeit computationally rather intensive, to calculate the masses of all possible permutations of disulphide-linked peptides that could be generated by enzymatic digestion and search the LC‑MS data for such signals – specialised software greatly facilitates this task. The process that generates data from disulphide‑linked peptides is shown schematically in Figure 2.
Figure 2: Schematic representation of the process of identifying, using mass spectrometry, the cysteines linked with disulphides in a sequence. In the normal arrangement of disulfides (A), pairs of peptides, generated by enzymatic digestion and which contain disulphide-linked cysteines (red-red and blue-blue pairs) remain linked and their unique masses can be used to confirm the assignments. An abnormal disulphide arrangement (B) can also be identified from the unique masses of proteolytic peptides thus linked (red-blue and blue-red pairs). Peptides that do not contain cysteines (shown in green) are the same in both situations. Upon reduction, signals from the disulphide‑linked peptides disappear and the individual, previously linked, peptides are detected.
In addition to identifying peptides linked with disulphides, LC-MS is also used to discover unexpected (and undesirable) modifications such as trisulphides, various cysteine thiol oxidation products and the formation of mixed disulphides with cysteine and glutathione molecules that may occur during cell culture.9,10
LC-MS of intact antibodies
The molecular mass measurement across a broad chromatographic peak can confirm the presence of a mixture of species differing by a few mass units, due to varying amounts of cysteine free thiol (sulphhydryl) groups in the molecule (such mass differences are due to the loss of two hydrogens each time two cysteines form a disulphide). The portion of the molecules with fewer disulphide bonds will typically unfold more readily under the denaturing conditions of reversed-phase chromatography and will elute later (as a shoulder to the main peak) than the more tightly folded molecules in which all disulphides have formed. This is reflected in the differences in the measured mass of the main antibody peak across the elution profile (Figure 3). Despite these mass differences being relatively small, they usually increase as more unfolded species elute that contain more cysteine-free thiol groups.
Figure 3: Elution of an antibody from a reversed phase high-performance liquid chromatography (HPLC) column. Species with all cysteines linked with disulphides (with calculated molecular mass 147621.5) elute earlier than species with some of the cysteines present in the reduced (sulphhydryl) state, the latter eluting as a shoulder to the main peak.
This can be confirmed using spectrophotometric methods, in which a free cysteine thiol is modified with a small molecule that carries a chromophore, which absorbs (or fluoresces) at a different wavelength from that of the protein’s intrinsic absorbance (or fluorescence). In addition to the original reagent (Ellman’s reagent) used for such determinations,11 other molecules with larger extinction coefficients for more accurate and sensitive determinations are commercially available.
Identification and quantitation of cysteines partially oxidised to disulphides
Intact mass measurements, as well as spectrophotometric methods, measure total free thiol in an antibody. However, it is often desirable to identify specific cysteines that have not formed disulphide bonds and to determine the relative ratio of oxidised sulphhydryl groups (ie, those in a disulphide bond) to free sulphhydryl groups. This can be accomplished by differential alkylation of cysteines.12-14 Briefly, the antibody is treated under denaturing conditions with a thiol reactive reagent such as iodoacetamide, iodoacetic acid, N-ethyl maleimide, etc, resulting in irreversible alkylation of the side chains of all cysteines that have not formed disulphides. Subsequently, disulphides are reduced and the newly formed thiols are alkylated with a different alkylating agent. The antibody is then digested with an enzyme and the cysteine‑containing peptides are identified. Usually, the majority have their cysteines modified by the alkylating reagent used after disulphide reduction, but some will show a mixture of alkylation with both reagents, as one portion of the particular cysteine had originally been oxidised to a disulphide and one portion had not.
Figure 4: The same cysteine-containing peptide from two different lots of an antibody drug, generated by cysteine alkylation with ‘heavy’ iodoacetamide (I-13C2H2–13CONH2), followed by disulphide reduction with D,L-dithiothreitol (DTT), followed by cysteine alkylation with ‘light’ iodoacetamide (I-CH2-CONH2), followed by enzymatic digestion. Any of the particular cysteine that had not formed a disulphide was modified with the ‘heavy’ iodoacetamide, while the balance of the cysteine, which had formed disulphide, modified with the ‘light’ iodoacetamide after disulphides were reduced. From the abundances (top number on each peak) the ratio of the oxidised-todisulphide cysteine and non-oxidised cysteine (free thiol) can be obtained for each lot.
However, quantitation from the relative mass spectrometric responses of such differentially alkylated peptides may not be accurate, as chemically different alkylating agents may, and usually do, affect peptide ionisation efficiency and hence the mass spectrometric signal intensities. This can be addressed by alkylating cysteines, before and after disulphide reduction, with two molecules that behave identically chemically in the mass spectrometer, but with one of them synthesised with stable isotopes. For example, cysteines can be alkylated before any disulphides are reduced with ‘heavy’ iodoacetamide that contains non-exchangeable 2H (deuterium) and 13C atoms (I-13C2H2–13CONH2) and after disulphide reduction with ‘light’ iodoacetamide (I-CH2-CONH2), which is not isotopically enriched. Pairs of peptides generated in this manner behave chemically (and chromatographically) identically and ionise with the same efficiency in the mass spectrometer, yet owing to the heavier isotopes present in one they have masses that differ by exactly 4.0193u (two 13C and two 2H atoms versus two C and two H atoms). Their peaks in the mass spectrum would be sufficiently well separated to enable accurate measurement of relative ion intensities, which can then be directly linked to their relative abundances and, therefore, to the ratio of free thiol versus disulphide for the particular cysteine. An example of the mass spectrum of such a ‘light’ and ‘heavy’ peptide pair (same amino acid sequence, one with cysteine modified with isotopically enriched ‘heavy’ iodoacetamide and the other with standard ‘light’ iodoacetamide) is shown in Figure 4.
The utility of this approach is illustrated in Table 1, which shows the cysteines of two lots of an IgG2 antibody and the percentage of each cysteine that was determined to have been originally reduced (sulphhydryl) or in a disulphide. It is worth noting that in this case a significant percentage of cysteines had not formed disulphides and the spectrophotometric determination of total free thiol was quite consistent with the mass spectrometric measurements. Differences in stability between the two lots could also be attributed to the differences in the respective content of unoxidised cysteines.
Conclusion
Properly formed disulphide bonds between cysteines play a critical role in the stability and activity of therapeutic antibodies and other biologically active proteins. With regard to therapeutic antibody drugs, it is essential – indeed a regulatory requirement – that the oxidation state of cysteines be determined; that disulphide bonds be fully characterised; and that any such changes between different lots of the same antibody drug be identified and evaluated. Mass spectrometry is a necessary and often sufficient method for accomplishing this.
About the author
Dr Ioannis Papayannopoulos has been involved in mass spectrometry R&D for several decades, focusing primarily on proteins. He has held senior scientific and management roles in the biopharmaceutical industry and in academia, at such companies as Celldex Therapeutics, AstraZeneca, Biogen and at the Koch Institute for Integrative Cancer Research at MIT, and he is currently principal scientist at Dragonfly Therapeutics, a biotechnology company in Massachusetts. Dr Papayannopoulos received his PhD degree in Organic Chemistry from the Massachusetts Institute of Technology, under the supervision of the late Professor Klaus Biemann.
References
- Feige M, et al. CHAPTER 1.1:Disulfide Bonds in Protein Folding and Stability , in Oxidative Folding of Proteins: Basic Principles, Cellular Regulation and Engineering, 2018, pp. 1-33.
- Liu H, May K. mAbs 4, 17-23 (2012).
- Mobli M, King GF. Toxicon 6, 849-854 (2010)
- Li X, et al. Analytical Biochemistry 436, 93-100 (2013)
- Liu H, et al. Analytical Chemistry 88, 5080-5087 (2016)
- Na S, et al. Molecular Biosystems 11, 1156-1164 (2015)
- Wang Y, et al. Analytical Chemistry 83, 3133-3240 (2011)
- Gu S, et al. Analytical Biochemistry 400, 89-98 (2010)
- Moritz B, Stracke JO. Electrophoresis 38, 769-785 (2017)
- Guan X, et al. mAbs 10, 572-582 (2018)
- Ellman GL. Archives of Biochemistry and Biophysics 82, 70–77 (1959)
- Xiang T, et al. Anlytical Chemistry 81, 8101-8108 (2009)
- Papayannopoulos IA, et al. “Quantitation of free protein thiols by differential cysteine alkylation with stable isotopes”, 66th ASMS conference, June 3-7, 2018, San Diego, CA.
- Robotham AC, Kelly JF. mAbs 11, 757-7066 (2019)
