Practical solutions for overcoming artificial disulfide scrambling in the non-reduced peptide mapping characterization of monoclonal antibodies

ABSTRACT

Non-reduced peptide mapping provides essential data for characterizing therapeutic monoclonal antibodies by isolating disulfide connections between specific cysteines. However, conventional digestive strategies used throughout the biopharmaceutical industry have been shown to cause unintentional rearrangement of disulfide connections (disulfide scrambling), thus generating connectivity profiles that do not accurately represent the protein being analyzed. Common misconceptions (e.g. avoiding basic-pH digestion to prevent disulfide scrambling) have led to the development of alternative reagents and conditions that can alleviate this issue, but yield problematic digestion profiles. Herein, we systematically and comprehensively examine the primary considerations for accurate non-reduced peptide mapping, and provide effective, practical solutions to minimize undesired behavior while still yielding high-quality digests. Additionally, we present a method that exploits intentional disulfide scrambling as a reference tool to demonstrate the robustness of our proposed strategies. We also introduce maleimide as a cysteine-alkylating reagent and demonstrate its benefits over industry-leading analogs such as N-ethylmaleimide in terms of compatibility with regulatory reports.

Introduction

Monoclonal antibodies (mAbs) have become increasingly popular as therapeutics for a wide range of illnesses, including cancer, autoimmune disorders, and infectious diseases.¹ mAbs are distinguished from traditional small molecule medications by their specificity toward designed targets, owing to their complex three-dimensional (3D) structures,² which are maintained through covalent disulfide bonds between cysteine residues.³ With improper disulfide connectivity (disulfide scrambling), proteins may exhibit incorrect folding and stability issues, and become unable to bind the intended antigens.⁴ Correct disulfide linkage is therefore a critical quality attribute of mAbs.⁵ However, ensuring the completeness of this correct linkage can be analytically challenging.

Mass spectrometry (MS) is currently the most practical technique for detailed protein analysis because of its ability to differentiate species with unique masses.⁶ Many common post-translational modifications in mAbs can be identified through both top-down (intact MS)⁷ and bottom-up (peptide mapping)⁸ approaches because the native and modified forms possess different masses. However, intra-protein disulfide scrambling cannot be distinguished from proper disulfide connectivity through top-down MS analysis because no overall difference in mass exists between variants with rearranged disulfide bonds. The preferred method to completely and explicitly characterize local disulfide connectivity is a bottom-up peptide approach under non-reducing conditions.⁹ However, performing mAb digestions under traditional non-reducing conditions has been extensively demonstrated to rearrange native disulfide connections, thus inducing new artificial disulfide scrambling,¹⁰ which cannot be distinguished from similar scrambling present within the undigested protein sample in the data analysis. Limiting disulfide rearrangement requires detailed knowledge of the mechanisms underlying artificial disulfide scrambling and is a major focus of the work herein.

Furthermore, because the levels of disulfide scrambling, whether preexisting¹¹ or digest-induced,¹² are often low within mAb digests, the detection and identification of each peptide can be challenging even with a sensitive mass spectrometer. Our previous work described a simple method for confident detection and identification of nearly all possible disulfide scrambling permutations for a given mAb, in which test samples are compared against a positive control digest of the same mAb with purposefully induced disulfide scrambling.¹³

Because of the labile reactivity of the sulfhydryl group (−SH) of cysteine, alkylating prior to digestion is the preferred method to minimize numerous side-reactions (e.g., oxidation, disulfide scrambling).¹⁴ Current popular sulfhydryl alkylating reagents contain iodo-substituting groups such as iodoacetamide and iodoacetic acid, which react with free cysteine via the S_N2 nucleophilic substitution mechanism.¹⁵ These reagents have benefits of high-selectivity toward sulfhydryls and room temperature reactivity, although they are generally limited to basic pH conditions,¹⁶ may over-alkylate under certain conditions¹⁷ or decompose when exposed to light.¹⁸ They are excellent choices for reduced peptide mapping of mAbs because all disulfide bonds are intentionally broken, and complete alkylation can be easily achieved.

For non-reduced peptide mapping, cysteine alkylation remains recommended because native mAb samples commonly contain low levels of unbonded free thiol cysteines,¹⁹ despite that mAb cysteines should ideally be fully paired and disulfide bonded to maintain the intended 3D structure to perform biological functions. Free thiols shielded within the folded structure cannot be alkylated without first denaturing the mAb to increase solvent-accessibility. However, after denaturation, the free thiols gain movement flexibility and consequently can approach and displace native disulfide connections through a thiol-disulfide exchange mechanism, thereby forming new disulfide connections that are not present in the original protein sample (i.e., disulfide scrambling).²⁰ Subsequent digestion would then generate a peptide map that is not fully representative of the protein being analyzed. To minimize this phenomenon, an alkylating reagent that can react with free thiols faster than disulfide rearrangement can occur must be chosen. Previous reports have indicated that iodo-based reagents fall short in this regard.²¹

Another common cysteine alkylating reagent is N-ethyl maleimide (NEM), which reacts with free thiols via the Michael addition mechanism. The flat maleimide ring structure is a particularly reactive Michael acceptor because of the strain caused by its conjugated double bonds, which restrict rotation into a more relaxed state. Thiol-maleimide coupling is often referred to as “click” chemistry because of its speed, convenience, and efficiency.²² NEM can alkylate cysteine under a wide pH range (pH 3–9) although at basic pH it may begin to react with amino groups (e.g., lysine side chain, peptide N-terminus). Using NEM under acidic pH conditions has been reported to prevent artificial disulfide scrambling in non-reduced mAb digests.²³ Our previous work has detailed the profound extensiveness of these effects in digestions comparing NEM and iodoacetamide.¹³ However, despite the utility of commercial products such as Promega’s AccuMAP Low pH Protein Digestion Kit, performing tryptic digests under acidic conditions is not ideal because many nonspecific or missed cleavages may occur, often yielding crowded and messy chromatograms.

Herein, we investigated individual factors contributing to artificial disulfide scrambling to optimize an efficient non-reduced digestion protocol that generates a simple, predictable tryptic peptide map with an accurate disulfide connectivity profile. We also describe a markedly improved positive control digest that exploits multiple combined principles of artificially induced disulfide scrambling. In addition, our recent discovery of glycine as a trifluoroacetic acid (TFA) mobile phase additive has been shown to boost the MS signal intensity of peptides by more than an order of magnitude on average, without compromising chromatography, and was applied to all the data reported.²⁴ These combined techniques greatly increase the confidence of identifying disulfide scrambling within mAb digests.

Materials and methods

Multiple non-reduced digests of cemiplimab (Regeneron Pharmaceuticals; IgG4, MW 143.8 kDa) were performed in parallel, each with slightly different conditions chosen to demonstrate the effects on the resulting disulfide scrambling profile. Generally, 100 µg of cemiplimab was denatured with 8 M urea (Sigma; U4883) in 0.1 M Tris-HCl pH 7.5 (Invitrogen Life Technologies; 15567027) at 50°C for 10 min, then digested with 1:20 E/S trypsin (Promega; V5111) at 37°C for 4 h after dilution of the urea concentration to 1.6 M with additional 0.1 M Tris-HCl pH 7.5, and was finally quenched to pH 2 with TFA (Thermo Fisher Scientific; 28904) to stop trypsin digestion. During the initial denaturing step, the urea solution was pre-mixed with the chosen alkylating reagent, composed of either 4 mm iodoacetamide (Sigma; A3221) for method A, 4 mm NEM (Sigma; 04259) for method B, no alkylation reagent for method C, or 4 mm maleimide (Sigma; 129585). Additional digests were performed according to the recommended procedure of the AccuMAP Low pH Protein Digestion Kit (Promega; VA1040), which was specifically designed to prevent disulfide scrambling and minimize other digestion-induced post-translational modifications. Generally, 50 µg of cemiplimab was denatured with 8 M urea in 1.4× AccuMAP Low pH Reaction Buffer at 37°C for 30 min, followed by a pre-digestion step using AccuMAP Low pH Resistant rLys-C at 37°C for 1 h, and the remaining digestion using AccuMAP Modified Trypsin and additional AccuMAP Low pH Resistant rLys-C was performed at 37°C for 3 h (pH 5.3). Finally, quenching was performed with TFA. One sample contained 8 mm NEM during the urea denaturing step for method D, and the other contained no alkylating reagent for method E. An additional 100 µg sample was denatured with 8 M urea in 0.1 M Tris-HCl pH 8.5 (Rockland; MB-027-1000) at 50°C for 10 min without alkylation, then digested with 1:20 E/S trypsin at 37°C for 4 h after dilution of the urea concentration to 1.6 M with additional 0.1 M Tris-HCl (pH 8.5), and was finally quenched to pH 2 with TFA for method F.

For peptide mapping analysis of cemiplimab, 5 µg aliquots of each tryptic digest were injected onto an ACQUITY UPLC I-Class system (Waters) equipped with a peptide BEH C18 column (Waters; 186003556). Peptide elution was subsequently performed with a linear gradient that increased from 0.1% mobile phase B to 40% mobile phase B over 80 min at a flow rate of 0.25 mL/min with a column temperature of 40°C. Mobile phase A was 0.05% TFA in water, and mobile phase B was 0.05% TFA in acetonitrile (Thermo Fisher Scientific; A955–4). All eluted peptides were monitored at a wavelength of 215 nm with a photodiode array detector. A mixing tee was set up to combine the post-column LC flow with a 0.1 M glycine (J. T. Baker; 4059–02) solution, administered with a syringe pump at 5 µL/min (2 mm final glycine concentration). The resulting flow was electro-sprayed and analyzed with a Q-Exactive Plus hybrid mass spectrometer (Thermo Fisher Scientific; IQLAAEGAAPFALGMBDK).

Results

Non-reduced digests with various conditions and cysteine alkylating reagents

Non-reduced trypsin digests of cemiplimab were performed using common cysteine alkylating reagents, as well as digests with no alkylation to purposefully maximize the occurrence of disulfide scrambling through the thiol-disulfide exchange mechanism. Basic-pH digestion without alkylation (method C) was effectively used as a positive control by forming nearly all possible scrambled disulfide permutations of cemiplimab, thereby generating each artificial peptide in relatively high abundance and enabling easy analysis of the conventional alkylated digest samples (methods A and B) through alignment of extracted ion chromatograms (EICs). The acidic-pH digest alkylated with NEM (method D) was used as a negative control for disulfide scrambling because those conditions were demonstrated to prevent such artifacts in our previous study. In addition, to establish that the native cemiplimab protein indeed contained the free thiols required for the thiol-disulfide exchange scrambling mechanism, we report the observed MS responses of the alkylated cysteine-containing peptides from method B (Supplementary Table 1).

The UV chromatograms of the digests described above are shown in Figure 1. All digests performed at pH 7.5 produced near-identical peak profiles despite the different alkylation choices, demonstrating that consistent tryptic peptide maps can be generated with various reagent classes under non-reduced conditions. All major UV peaks represented desired and predictable tryptic peptides, with only a few minor missed cleavage peptides (Supplementary Table 2), thus indicating the excellent quality of the chosen conditions. The digests performed under acidic conditions (pH 5.3) with Promega’s AccuMAP reagents and recommended protocol revealed many of the same peptides as the basic-pH digests but were accompanied by several major tryptic missed cleavage peptides and possible enzyme autolysis peptides (Supplementary Table 3), yielding more complex chromatograms because of the inefficiency of the digestive enzymes at this pH.

Figure 1.

UV chromatograms (215 nm) of non-reduced tryptic digests of cemiplimab under various conditions. Methods A, B and C involved digestion at pH 7.5 with various alkylation reagents (iodoacetamide, NEM and no alkylation, respectively). Methods D and E involved digestion at pH 5.3 with Promega’s AccuMAP Low pH Protein digestion Kit with NEM and no alkylation, respectively. Trypsin missed cleavages, non-specific cleavages and protease autolysis peaks are labeled with an asterisk (*).

MS detection of all possible scrambled disulfide connections in digests of cemiplimab

A typical homodimeric IgG4 mAb (such as cemiplimab) contains 16 unique cysteines (11 in the heavy chain and 5 in the light chain), which are usually divided among 15 unique linear peptides in a reduced trypsin digestion, with the hinge region peptide containing two cysteines (…PPCPPCPA …). To simplify the disulfide scrambling analysis in this study, we omitted the hinge peptide because of the complexity that each scrambled disulfide would possess as a result of the presence of multiple cysteines. The remaining 14 linear peptides (shown in the table in Figure 2) could theoretically form disulfide connections with any other, resulting in a total of 105 unique disulfide permutations. Excluding the seven expected native disulfide connections leaves 98 possible unique scrambled disulfides. Using our positive control digest (method C), we successfully detected 94 of the 98 possible peptides (95.9%) via MS (Figure 2). To our knowledge, no previous study in the literature has identified such a comprehensive set of scrambled disulfides from a mAb digestion.

Figure 2.

Relative MS response heat map for all possible unique scrambled disulfide connections within cemiplimab after trypsin digestion under various conditions and alkylating reagents. All digests were performed at 37°C for 4 h total. Each mAb domain contains two cysteine-containing peptides, labeled with “a” and “b” in the tables above (e.g., VHa and VHb). All major charge states and isotopic peaks were included in the EIC peak integration. “Not detected” indicates that the disulfide peptide was not identified even in the positive control digests with no alkylation (methods C and F).

Artificial disulfide scrambling due to thiol-disulfide exchange using various digestion conditions and cysteine alkylating reagents

For each prepared digest, the alkylating reagent was pre-mixed into the urea denaturant solution, so that no time would elapse between the protein unfolding and availability for alkylation to occur. Therefore, only the molecular kinetics would determine whether a free thiol was alkylated and therefore deactivated from further chemistry or was able to rearrange with other existing disulfides via thiol-disulfide exchange. Using the EICs of scrambled disulfide peptides from the positive control as references, it was simple to detect the presence of the same peptides in the other digests through matching of the retention times and accurate masses. Although iodoacetamide (method A) was not expected to fully prevent scrambled disulfides from forming, an astonishing 77 of the 98 possible artifactual peptides (78.6%) were nonetheless detected at various levels (Figure 2), thus demonstrating that the reaction kinetics of iodoacetamide coupling with free thiols are insufficient to minimize the rearrangement of disulfide bonds. In contrast, the basic-pH digestion performed with NEM (method B) produced only trivial MS responses for a few scrambled disulfides. Quantitatively, the average relative abundance of all scrambled disulfides for the positive control, iodoacetamide and NEM digests was 1.3%, 0.1%, and 0.002%, respectively. This negligible amount indicates that NEM is extremely effective (compared to iodoacetamide) toward preventing disulfide scrambling artifacts from forming, by reacting with free thiols much faster than the rearrangement of native disulfide bonds, even under basic conditions.

In comparison, the acidic-pH digest using NEM (method D) yielded no detectable MS response for any scrambled disulfide (except a trivial amount of one peptide), as expected according to previous literature suggesting that acidic conditions would prevent such artifacts. However, unexpectedly, the acidic-pH digest with no alkylation (method E) in fact yielded substantial disulfide scrambling, with an abundance profile similar to that of the basic-pH digestion alkylated with iodoacetamide (0.1% on average for all possible peptides). Although this value was lower than that of the basic-pH digestion with no alkylation (1.3%), it indicates that digestion under acidic conditions is not the only meaningful consideration for preventing disulfide scrambling and that the thiol-disulfide exchange mechanism still occurs if free thiols remain unalkylated.

Head-to-head reaction kinetics of various cysteine alkylation reagents

To explicitly demonstrate the difference in kinetics between iodoacetamide and NEM, two additional digests were prepared with both alkylating reagents pre-mixed together in the urea denaturant solution, and using the same basic-pH digestion conditions as those in method A/B. The MS responses of a common IgG4 cysteine-containing peptide were extracted for both alkylated forms, as well as for an artificial scrambled disulfide involving the same linear peptide (Figure 3). One sample contained a 1:1 mixture of each reagent at 4 mm, and the other sample contained 40 mm iodoacetamide and 2 mm NEM. The sample with the 1:1 mixture showed nearly identical results to the sample containing only NEM. In the other sample, even despite a 20-fold excess of iodoacetamide, the vast majority of the resulting alkylated peptide still existed in the NEM-labeled form (42-fold more), thereby demonstrating that free thiols preferentially react much faster with NEM than iodoacetamide, regardless of concentration. Additionally, the amount of the scrambled disulfide generated was minimal whenever NEM was present, and the abundance of this artifactual peptide became substantial only in the digest containing iodoacetamide alone.

Figure 3.

MS responses for a common IgG4 mAb cysteine-containing tryptic peptide alkylated with either NEM or iodoacetamide (IAM), as well as a scrambled disulfide involving the same peptide. Samples of cemiplimab were denatured, alkylated and digested under non-reducing conditions at pH 7.5. All major charge states and isotopic peaks were included in the EIC peak integrations.

Artificial disulfide scrambling due to β-elimination with variable digestion pH and time

Beyond direct thiol-disulfide exchange due to the presence of preexisting free thiols, other mechanisms can cause artificial disulfide scrambling during a mAb digestion, even after free thiols have been successfully alkylated. The most commonly accepted mechanism is β-elimination, in which a hydroxide anion abstracts a proton from the carbon two atoms away (the β-position) from the sulfur of a disulfide peptide cysteine, thereby displacing the other peptide into a persulfate form, which then decomposes into free cysteine (Supplementary Figure S1). This new free thiol peptide can further exchange with a native disulfide peptide, forming a scrambled disulfide.²⁵ Because this mechanism is initiated by hydroxide anions (⁻OH), it occurs more rapidly at higher pH.

To demonstrate the effects of β-elimination on the formation of artificial disulfide scrambling, we performed a set of digests similarly to method B, while varying the digestion pH and time. Each basic-pH digest was initially treated with NEM at pH 7.5 to ensure that all free thiols were quickly alkylated, and therefore any significant disulfide scrambling detected would necessarily be ascribable to mechanisms other than thiol-disulfide exchange from preexisting free thiols. The acidic digest at pH 5.3 was prepared according to method D. A clear trend emerged in which the abundances of two chosen representative scrambled disulfides markedly increased with both digestion pH and time (Figure 4). As expected, the acidic-pH digest yielded negligible disulfide scrambling, even after 22 h because of the significantly lower concentration of hydroxide ions (2–3 orders of magnitude) than was present in the basic-pH digests. In comparison, the mild conditions of method B (pH 7.5, 4 h) also yielded only trace quantities of disulfide scrambling. Only under harsher conditions (i.e., pH >7.5 and time >4 h) did the effect of β-elimination significantly contribute to disulfide scrambling.

Figure 4.

MS responses for scrambled disulfides in tryptic digests of cemiplimab under variable pH and time. All samples were initially alkylated with NEM to ensure that any significant scrambling formation would be due to alternative mechanisms such as β-elimination. All major charge states and isotopic peaks were included in the EIC peak integrations.

Generating an improved disulfide scrambling positive control sample

Scrambled disulfides, whether preexisting or artificially induced, often exist in low abundance within a given digest. Characterizing such disulfide peptides by MS/MS can be especially challenging due to low MS signals and the potential complexity of their fragmentation patterns. To aid in the explicit identification of each scrambled disulfide peptide, artificial scrambling formation can be maximally exploited by combining thiol-disulfide exchange and β-elimination simultaneously. A cemiplimab sample digested without alkylation and also under harsher basic conditions (method F, pH 8.5) generated a peptide mixture that contained, on average, nearly fivefold more of each scrambled disulfide permutation than observed in the previously described positive control sample under method C (Figure 2). Thus, with a greater MS signal, each peptide is easier to characterize with a more detailed MS/MS spectrum.

A beneficial alternative to NEM as a cysteine alkylating reagent for non-reduced peptide mapping

Finally, although NEM is clearly an effective alkylating reagent for minimizing artificial disulfide scrambling during non-reduced trypsin digestion, it nonetheless possesses multiple undesired attributes. As shown in the UV chromatogram in Figure 1, the use of NEM commonly results in two large reagent peaks (representing NEM and hydroxylated NEM), which tower above all relevant peptide peaks and are also sufficiently hydrophobic to potentially overlap with peptides, thus making the chromatogram unacceptable for therapeutic regulatory reports. In comparison, although iodoacetamide also shows a large reagent UV peak, it elutes before the peptides and therefore can be simply cropped from the displayed chromatogram. Another result of NEM’s hydrophobicity, and consequently limited aqueous solubility, is that preparing a stock solution may be difficult or impossible if high concentrations are desired. Thus, we recommend maleimide as an alternative alkylating reagent that provides all the same reactivity benefits of NEM without these disadvantages, yet remains unused in the literature. Maleimide possesses the same thiol-reactive ring structure as NEM, but the nitrogen is substituted with hydrogen instead of an ethyl group, thus significantly decreasing its hydrophobicity. It is available commercially and inexpensively from common vendors. Maleimide demonstrated the same ability as NEM to minimize artificial disulfide scrambling during mAb digestion under conditions identical to method B (Supplementary Figure S2). Additionally, its reagent UV peaks have shorter retention times that are out of the range of reported peptides (Supplementary Figures S3 and S4), and it dissolves quickly in water at high concentrations (e.g., 0.5 M). We have observed only minimal maleimide ring hydrolysis (2% on average) from these mild digestion conditions. Maleimide may therefore be an ideal substitute for NEM for non-reduced mAb digests while also satisfying the criteria of UV chromatograms for therapeutic regulatory reports.

Discussion

Our results provide valuable insights into the mechanisms and prevention of artificial disulfide scrambling during non-reduced trypsin digestion of mAbs. The findings highlight the critical role of the alkylating reagent in minimizing disulfide scrambling, and underscore the importance of optimizing digestion conditions to prevent such artifacts while also generating desirable tryptic peptide maps. Upon following all the considerations described herein, any detectable disulfide scrambling can be concluded as preexisting within a given mAb sample and likely other disulfide-containing proteins as well.

NEM is a highly effective alkylating reagent for preventing disulfide scrambling because its reaction kinetics with free thiols are significantly faster than thiol-disulfide exchange. Contrary to common belief throughout the industry, this is even true under the basic-pH conditions needed for efficient trypsin digestion. Thus, alternative conditions and reagents that use acidic pH are unnecessary. Alkylation using iodoacetamide, however, resulted in extensive disulfide scrambling, yielding a wide range of detectable levels for a majority of all the possible disulfide permutations. The slower reaction kinetics of iodoacetamide compared to NEM were explicitly evident when both reagents were mixed and added simultaneously to a mAb sample, which resulted in a digest containing peptides almost exclusively labeled with NEM.

Apart from the thiol-disulfide exchange mechanism, β-elimination was explored as a potential contributor to disulfide scrambling as well. Our results showed that the abundance of scrambled disulfides increased with both higher pH and longer digestion times, aligning with the known mechanism of β-elimination, which is initiated by hydroxide anions and occurs continuously during the digestion through decomposition of native disulfide peptides. Therefore, the production of scrambled disulfides due to β-elimination is not limited by the initial free thiol quantities of the protein, in contrast to scrambling from direct thiol-disulfide exchange. Mild basic conditions (pH 7.5) only yielded significant disulfide scrambling in the 22 h digest sample, thus we recommend avoiding overnight non-reduced digestions. As expected, the acidic digests with NEM alkylation yielded negligible disulfide scrambling, even after 22 h, confirming that lower pH conditions effectively prevent β-elimination due to the decreased hydroxide anion concentration. However, surprisingly, the acidic-pH digest with no alkylating reagent yielded a disulfide scrambling profile similar to that of the basic-pH digest alkylated with iodoacetamide, demonstrating that thiol-disulfide exchange still occurs under acidic conditions, and emphasizing the general necessity of alkylation for non-reduced digestions.

The identification of each scrambled disulfide peptide was facilitated by a positive control digestion that intentionally induced high levels of scrambling, combining conditions that exploit both thiol-disulfide exchange and β-elimination in a single sample, by digesting without alkylation and at high pH. The improved MS responses allowed better MS/MS fragmentation and clear retention time peak alignment with peptides from all comparator digests, which enabled confident assignment and peak selection.

Although alkylating with NEM and digesting with trypsin under acidic conditions successfully minimizes artificial disulfide scrambling, the peptide profile will likely contain several missed cleavages, nonspecific cleavages, and enzyme autolysis peptides. Fortunately, we have shown that digesting with mild basic conditions similarly yields negligible disulfide scrambling, with the added benefit of a desirable and efficient tryptic profile, if alkylation is performed with a maleimide-based reagent such as NEM. The UV chromatogram for this digest displays only a few minor missed cleavage peaks and is otherwise extremely high-quality, except for the large hydrophobic reagent peaks of NEM. To resolve this issue, maleimide was tested as an alternative alkylating reagent. Because it has the same cysteine-reactive ring structure of NEM but lacks the hydrophobic N-ethyl group, it prevents disulfide scrambling formation just as effectively as NEM, and its reagent UV peaks have significantly shorter retention times that elute earlier than the tryptic peptides, as required for therapeutic regulatory reports. Additionally, maleimide ring hydrolysis was found to be minimal under our recommended mild basic digestion conditions. This makes maleimide a practical choice for non-reduced digestion of mAbs.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

The authors declare the following competing financial interest(s): A.K., Y.M., and N.L. are full-time employees and shareholders of Regeneron Pharmaceuticals, Inc.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19420862.2024.2420805