Characterization of charge variants, including post-translational modifications and proteoforms, of bispecific antigen-binding protein by cation-exchange chromatography coupled to native mass spectrometry

Abstract

Charge variant characterization of biologics is critical to ensure that product meets the required quality and regulatory requirements to ensure safety and efficacy of the biotherapeutic. Charge variants arise from post-translation modifications (PTMs) during upstream processing and due to enzymatic and non-enzymatic chemical reactions that occur during downstream processing and storage. Some of these modifications may impact therapeutic potency, efficacy, or immunogenicity of a biotherapeutic. The traditional workflow for characterizing charge variants that involves fraction enrichment is time-consuming and labor-intensive. This approach can be especially challenging if the product is manufactured at low concentrations (e.g., ≤ 2 mg/mL). Recent advances in pH-based elution for ion-exchange chromatography utilizing volatile buffers have enabled rapid native mass-spectrometry-based identification of PTMs and proteoforms associated with protein therapeutics. In this study, we develop a novel workflow to rapidly and unambiguously characterize modifications associated with a new class of biotherapeutics known as bispecific antigen-binding protein (BsABP), including low-level modifications. A cation-exchange separation was optimized using volatile buffers to provide online hyphenation for native mass spectrometry to profile modifications and proteoforms present at the native level of a biotherapeutic, such as deamidation, O-glycosylation, amino acid substitution, N-linked glycosylation and oxidation. Furthermore, a limited proteolysis method was developed to specifically inform about modifications in the different domains of the bispecific antibody. Using this approach, we could efficiently identify PTMs in unstressed, thermally and photo-stressed samples, and provide information about the impact of downstream purification in clearing out modified BsABP species. Furthermore, peptide mapping was performed to identify and confirm modifications at the amino acid residue level. The developed workflow is less time-consumable and reduces sample processing- and analysis-related artifacts compared to traditional approaches.

Graphical Abstract

Graphical abstract figure for characterization of charge variants of bispecific antigen-binding protein (BsABP) molecule

1. Introduction

Microheterogeneity in biotherapeutics is influenced by post-translational modifications (PTMs) that happen during biosynthesis, purification, or accumulate during storage conditions^1–4. Post-translational modifications (PTMs) provide therapeutic proteins with various biochemical and biophysical characteristics^5–7. PTMs of monoclonal antibodies (mAbs) are widely and thoroughly studied. PTMs, including glycosylation, deamidation, fragmentation (i.e., cleavage or/and truncation), isomerization, oxidation, etc., can have an impact on product quality if present above a specific limit and can lead to decreased potency due to loss of binding or immunogenicity; as a result, certain PTMs as considered as critical quality attributes (CQAs)^8–11.

First step in achieving CQAs is identifying quality target product profile (QTPP), wherein quality attributes that need to be present to have desired product quality and potential steps in the process that can impact this quality attribute are identified¹². Hence, analytical methods are deployed to characterize and quantify these modifications and ensure their presence within desired concentration limits.

Generally, different PTMs will change surface charges on the protein surface, which alters the population distribution of the charged variants^{13, 14}. Traditionally, ion-exchange chromatography methods utilize high ionic strength non-volatile salts to separate charge variants. Because such mobile phases are non-volatile, it is difficult to directly couple these techniques to mass spectrometry to identify charge variants. A time-consuming and labor-intensive fraction collection, cleanup, and enrichment approaches are often necessary prior to MS analysis¹⁵. This characterization process can take up to months and could induce modifications due to degradation during storage or buffer exchange. Hence there is a need for an approach to provide rapid identification and quantitation of modifications for protein therapeutics.

The use of volatile salts for cation exchange has recently enabled its coupling to mass spectrometry to determine the molecular weights related to charge variants present in biotherapeutics^16–19. These methods can provide valuable information on a diversity of modifications, requiring less time and experimental work than the traditional offline fraction enrichment studies approach. The majority of studies have focused on characterizing charge variants in monoclonal antibodies, with very few investigating the applicability of such methods to other biotherapeutic modalities, such as bispecific antigen-binding proteins (BsABP)²⁰.

Immuno-oncology therapies, such as bispecific antigen binding protein (BsABP), are being developed to combat various oncology disorders. The single-chain BsABP in this study is engineered to be aglycosylated and only one-third the mass of traditional mAbs. It is made up of two ScFv domains, each containing variable heavy (V_H) and variable light (V_L) chains, which are connected by linker peptides consisting of Gly-Gly-Gly-Gly-Ser amino acid repeats; one domain binds to antigen-specific to cancer (Antigen-binding domain 1) while other binds to antigen responsible for immune response (Antigen-binding domain 2). Due to its high potency at lower doses, the BsABP in this study was formulated at a concentration significantly lower than most monoclonal antibodies (mAbs), which made the characterization of the charge variant with the traditional fraction enrichment approach challenging.

In this work, we designed a workflow, as described in Figure 1, to characterize the charge heterogeneity associated with bispecific antigen-binding protein. First, we developed an approach for charge variant analysis of BsABP using volatile salt buffer, where each variant is eluted based on the pH of the buffer in the gradient. By optimizing the MS source conditions of a Biopharma Q-Exactive Plus mass spectrometer, we obtained native mass spectra associated with various charge isoforms, with deconvoluted masses used to predict the associated modifications. Next, we developed a limited Lys-C digestion method in which a lysine residue between the two ScFv domains was cleaved to enable chromatographic resolution and MS analysis to determine PTMs associated with different domains of BsABP. Finally, LC-MS/MS-based peptide mapping was used to confirm the identities of modifications at the amino acid level.

Figure 1:

Developed workflows for charge variant characterization of bispecific antigen-binding protein (BsABP) molecule.

2. Methods and reagents

2.1. Reagents and materials

The highest purity reagents were acquired for analysis. MS grade ammonium acetate was acquired from Sigma-Aldrich (St Louis, MO). Ammonium hydroxide and formic acid were used to adjust the buffer’s pH; the buffers for mobile phases were obtained from Fischer (Madison, WI) and Sigma-Aldrich (St Louis, MO), respectively. Lys-C enzyme for limited digestion was obtained from New England Biolabs (Ipswich, MA). The Promega (Madison, WI) low pH digestion kit was used for the peptide mapping procedure. O-glysosidase was obtained from New England Biolabs (Ipswisch, MA)

2.2. Samples

BsABP molecule and downstream purified sample for BsABP molecule was formulated at Amgen Inc. (Cambridge, MA) at 36.36 μM. BsABP drug substance was thermally-stressed between 55°–60°C by incubating on a heat block at an for 10 days. BsABP drug substance was photo-stressed in a photo chamber from Caron (part# 7540, Marietta, Ohio) at 8 Klux for 24 hours.

2.3. Method information

2.3.1. Native cation-exchange chromatography coupled to mass spectrometry (CEX-MS).

Cation-exchange chromatography was performed on a Waters UPLC (Milford MA) equipped with a binary pump and column heater. A strong cation exchange column from YMC (Andover, MA) bio pro-SPF-IEX (100 mm × 4.6 mm ID with a particle size of 5 μm) was used at room temperature. Solvent A consists of 50 mM ammonium acetate, pH 5.5, and solvent B’s composition is 150 mM ammonium acetate, pH 10.2; the solvents were prepared fresh before the analysis. The gradient was optimized to separate the charge variants of the bispecific antibody sample at a flow rate of 0.5 mL/min. For analysis, the column was equilibrated with 100% A for 3 min, a 2% B/min gradient applied for 35 min, then washed with 100% buffer B for 5 min and equilibrated with 100% A for 5 min. BABB samples were injected at a final amount of 30 μg without any buffer exchange.A Biopharma Orbitrap Q-Exactive Plus (Thermo, Bremen, Germany) mass spectrometer was used in LC-MS analyses. The ion source conditions were optimized to provide complete desolvation of protein ions. Native electrospray ionization (ESI) conditions allow the protein to retain its tertiary or quaternary structure, resulting in ions bearing fewer charges than the denaturing ESI conditions. high mass range (HMR) mode was enabled by setting trapping gas at 1.3 to acquire spectra in the 1800− 7500 m/z range. Spray voltage was set at 3.8 kV, S-lens at 150, source heat temperature at 270 °C, and auxiliary heat temperature at 220 °C. The optimal in-source CID condition was between 100–110 eV, and the AGC target was set at 3e⁶. For O-glycosidases, BsABP at 36.36 μM was treated with 5 μL of O-glycosidase at a concentration of 40,000 units/μL for a duration of 3 hours. Subsequently, the reaction was quenched by the addition of 4% v/v of 10% acetic acid. The analysis of the samples was performed as per abovementioned method.

2.3.2. Limited Digestion of BsABP samples

BsABP drug material was diluted with 25 mM Tris-HCl pH 7.5 to a final concentration of 9.2 μM. The Lys-C enzyme was reconstituted to a final concentration of 1 mg/mL with 25 mM Tris-HCl pH 7.5 and added to the protein solution for an enzyme/substrate ratio of 1:10 for 4 hours of digestion at 37°C was followed by quenching with 1% v/v 10% acetic acid. UPLC, column, gradient, column temperature, injection load, and mass spectrometry conditions were the same as described in the previous section about native ion-exchange chromatography. The mobile phase was made of solvent A, consisting of 20 mM ammonium acetate, pH 4.5, and solvent B composed of 150 mM ammonium acetate, pH 9.5.

2.3.3. Peptide mapping of BsABP samples

Samples for peptide mapping were prepared utilizing buffers and enzymes from the Promega (Madison, WI) low pH digestion kit. In the experimental procedure, a total of 25 μL of protein was combined with 40 μL of proprietary accumap denaturing buffer. To facilitate denaturation and reduction, 2 μL of 100 mM TCEP (Tris(2-carboxyethyl)phosphine) was added, and the mixture was allowed to incubate for a duration of 30 minutes. Following denaturation and reduction, the samples were subjected to alkylation using 2 μL of 300 mM iodoacetamide at pH 8.0 for 30 minutes. Once the alkylation process was completed, the samples were treated with Lys-c enzyme at a ratio of 1:5 (enzyme to protein) for a period of 1 hour. Subsequently, the samples underwent treatment with trypsin at a ratio of 1:5 (enzyme to protein) for 3 hours. To quench the enzymatic reactions, a 20% solution of trifluoroacetic acid (TFA) was employed, reaching a final volume of 2% v/v. The digests were analyzed by reversed-phase liquid chromatography on a Waters UPLC equipped with a binary pump and column heater set at 65 °C. A BEH C18 column (130 Å, 1.7 μm, 2.1 × 150 mm) from Waters (Milford, MA) was used for peptide separation with buffer A as 0.1% formic acid in water and buffer B as 0.1% formic acid in acetonitrile. The linear gradient was established at 0.25 mL/ min and a 0.3% change of buffer B/min. A Biopharma Orbitrap Q-Exactive Plus (Thermo, Bremen, Germany) was used to acquire data-dependent MS/MS spectra. Instrument settings are listed in Table S1.

2.4. MS data analysis

2.4.1. CEX-MS data of the intact BsABP and its limited Lys-C digest

For all the native MS and LC-MS of limited digest experiments in characterization of BsABP, deconvolution of mass spectra was performed using an internal Amgen Inc. software Magtran (v 4.03)^{17, 21, 22,23}. The m/z scan range for deconvolution was set between 1000–7000. S/N threshold was set at 1.05% of the most intense peak in each mass spectrum, and mass accuracy was set at 1 ppm. The deconvoluted mass obtained for each charge variant was then subtracted from that obtained for the main species. The mass difference indicates a post-translational modification of the native protein and is attributed to either domain of BsABP from the results of the limited digestion samples.

2.4.2. Peptide mapping data

All peptide map results were analyzed using an internal Amgen Inc. software MassAnalyzer (v4.03)^{17, 21–23}. MS/MS spectra were manually examined to identify the peptides with modified amino acid residues. Relative abundances of modifications (as a percentage) was calculated based on the extracted ion chromatogram (EIC) peak areas of the unmodified/native and modified peptides using equation 1:

$$\%\mathit{\text{Modification}}=\frac{\text{Extracted}\text{ion}\text{chromatogram}\text{(EIC)}\text{peak}\text{area}\text{for}\text{modified}\text{peak}}{\text{EIC}\text{(peak}\text{area}\text{of}\text{native}\text{peak}\text{+}\text{Peak}\text{area}\text{for}\text{modified}\text{peak)}}\times 100\%$$ Equation 1:

3. Results and discussion

3.1. Optimization of native CEX-MS.

For the analysis and separation of BsABP charge variants, the cation-exchange chromatography was performed using volatile buffers and elution for different charge variants was based on pH gradients delivered by binary pumps of liquid chromatography system, when the pH of solvent approached the pI (isoelectric point) of the charged variant, this variant eluted from the column. Volatile salt-based mobile phases enable the appropriate resolution in the separation of charge variants while being compatible with online mass spectrometry-based characterization. We selected an ammonium acetate-based mobile phase system at a concentration ranging from 50 to 150 mM, as it can provide a sufficient range of buffering capacity to enable the separation of charge variants and compatibility with mass spectrometry whilst retaining its native structure. A critical concern with ammonium acetate-based volatile buffers is the formation of ammonium adducts of the analytes, which can interfere with deconvolution, leading to high mass errors and difficulties predicting modifications. Thus, we optimized the in-source CID energy level (at 100 eV) to aid in removing adducts bound to native ions formed at the source, as seen in Figure S1. Based on our evaluations, we chose to set the MS resolution at 17,500 to increase the sensitivity, improve the statistics of ion measurements, and minimize the transient decay.

3.2. Modifications determined for charge variant analysis under native conditions

Five acidic species and one basic species were observed in the BsABP chromatogram using the developed gradient, as shown in Figure 2A and Figure 2B. Two acidic species with mass differences of 0.5 and 1.5 Da (Figure 2D, Figure 2A, and Figure 2C, were predicted to be deamidated at different sites, consistent with a pI decrease of ~0.07 units and a mass increase of 0.98 Da expected for asparagine deamidation along with earlier elution of deamidated species^24–26. Apart from deamidation, three different modifications are observed in the acidic region, two of which showed mass differences of 995 and 915 Da, as seen in Figure 2D in Figure 2A and Figure 2C (annotations !! and ¶). Upon treatment with O-glycosidase, peaks corresponding to these two mass differences were not observed, as seen in Figure S2, indicating the presence of O-glycosylation species at serine or threonine residues. There are several structures for O-glycans at these two masses, and two proposed representative O-glycan structures are shown in Figure S3. We hypothesize that the presence of O-glycan can affect the pI of the protein, causing it to elute in the chromatogram’s acidic, main, or basic region. Aside from that, we see a mass difference of + 58 Da, as shown in Figure 2D in Figure 2B, and Figure 2C (see annotation *!), which we predict is due to the presence of an amino acid substitution from glycine to aspartic acid, causing it to elute earlier. Extended discussions can be found in the section on peptide mapping. For the basic species, the deconvoluted mass was twice the mass observed from the main species, indicating the presence of a dimer that has higher pI due to a higher surface area compared to the monomer.

Figure 2:

Modifications observed in bispecific antigen-binding protein (BsABP) molecule by native CEX-MS. (A) A zoomed-in view of the total ion chromatogram (TIC) of the BsABP molecule. (B) A full-scale view TIC of the BsABP molecule. (C) Mass spectra of the variants that were eluted at different times. (D) List of the mass differences with predicted modifications for each of the observed BsABP species.

Various PTMs can be induced or stoichiometrically elevated under various stress conditions, which may provide information concerning possible degradation pathways²⁷. BsABP drug substance was subjected to thermal-stress and photo-stress and examined using the presented here CEX-MS method. Thermally stressed material showed an increase in deamidated species peaks compared to non-stressed drug substance, with an additional (+1 Da) peak that may be related to an additional deamidation site. The results of peptide mapping experiments, as discussed later in the section, correlated with the data acquired using the developed CEX-MS technique as the residue N313 was observed as an extra deamidated site in the thermally stressed sample, as shown in Table 1, Figure 3A, Figure 3C (annotation §§), and Figure 3D. An early eluting peak, as seen in Figure 3A, Figure 3C (annotation !*), and Figure 3D, was observed to contain two species with masses related to fragments of the two domains of BsABP due to hydrolysis at elevated temperatures. Two peaks were identified in the basic region. One has a mass of the intact BsABP, which we predict corresponds to the isomerization of aspartic acid (Figure 3A, Figure 3C (annotation **), and Figure 3D). Elution as a basic species could be due to the charge/pI difference caused by isomerization²⁶. The other basic peak has a mass difference of −18 Da from the intact mass, as seen in Figure 3A, Figure 3C (annotations!!), and Figure 3D. We predict this shift to be caused by succinimide formation, an intermediate for asparagine conversion to either iso-aspartic acid or aspartic acid, as succinimide intermediate is traditionally observed in basic species^28–30.

Table 1:

Results of the peptide mapping analysis of BsABP molecule and the forced-degraded material related to the extent of modification.

Sample type of BsABP	Modifications observed	Amino acid residue	Antigen-binding Domain	% Modifications
Drug substance	Deamidation	N360	2	0.96
		N363		2.34
	Amino acid substitution	G467	2	0.10
Thermal Stress	Deamidation	N360	2	10.85
		N363		14.19
		N313		0.78
	Succinimide formation	D62	1	0.87
		D140		0.31
		D172		3.27
		D204		0.34
	Isomerization	D54		21.17
Photo-stressed	Dioxidation	W52	1	4.00
		W490	2	3.36
		W495	2	0.68
		M83	1	3.25
		M233		0.69
		M342	2	0.93
	Isomerization	D54	1	4.77

Figure 3:

Modifications observed in the forced-degraded BsABP molecule by native CEX-MS. (A) Zoomed-in view of the TIC corresponding to the thermally-stressed BsABP molecule. (B) Zoomed-in view of the TIC of the photo-stressed BsABP molecule. (C) Mass spectra of the variants that were eluted at different retention times. (D) List of the mass differences between the unmodified BsABP species and the detected BsABP species with predicted modifications for each of the observed species in the evaluated stressed BsABP molecule.

Oxidation on particular amino acids, including methionine, tryptophan, or histidine on the complementary determining region (CDR) of antibodies, may impact antigen-binding and product quality⁹, which is important to monitor. CEX-MS analysis of the photo-stressed BsABP sample showed the presence of a peak in the acidic region with a mass difference of +16 and +32 Da due to oxidation Figure 3B, Figure 3C (annotation ╞) with predominant +32 Da species indicating higher presence of double oxidation. Although the specific mechanism causing acidic elution of oxidized species is unknown, we postulate that early eluting oxidized species could be generated by tryptophan or histidine oxidation due to charge or pI difference caused by oxidized carbon atoms, as it was corroborated by peptide mapping experiments showing higher tryptophan oxidation. Furthermore, we observed an isomerization demonstrated in Figure 3B, Figure 3C (annotation ‡), and Figure 3D, as previously observed in thermal stress was also seen in photo-stressed material. Peptide mapping identified the sites specific to oxidation, as shown in Table 1.

When designing purification processes for protein therapeutics, it is essential to understand how changes are introduced or eliminated during the purification process. Such information will help design a purification process that results in protein therapies with a lower content of undesired PTMs and degradation products. To better understand the differential effects of the purification process on the levels of PTMs, the CEX-MS profiles of the BsABP from the affinity column purification step were compared to the profile of the final drug substance, as seen in Figure 4B and Figure 4D. Two peaks were observed in the affinity-captured material that was not in the final drug substance, as observed in Figure 2, which corresponded to the presence of N-linked sialylated glycans (Figure S4). This observation is consistent with the expectation for sialylated species to elute earlier in the applied CEX gradient¹⁹. In subsequent purification stages, these modified species of the BsABP were further cleared away as we did not observe their presence in the final drug substance.

Figure 4:

Modification observed for the affinity-purified BsABP molecule by native CEX-MS. (A) Full view TIC of affinity-purified BsABP molecule. (B) Zoomed-in view of the TIC of the affinity-purified BsABP molecule. (C) Mass spectra corresponding to the peaks detected in the CEX-MS-based BsABP characterization experiments. (D) List of the mass differences with observed modifications of the species detected in the affinity-purified BsABP molecule

Analysis of triplicate samples revealed consistent chromatographic and total ion chromatogram (TIC) profiles, with similar retention times for each triplicate injection observed at less than 5% coefficient of variation (CV). Deconvoluted mass of each variant and mass differences were similar for all the charge variant species, indicating the repeatability of the developed method.

3.3. Limited Digestion CEX-MS

Limited proteolysis is used to probe the conformational features of proteins, affording information about the residues in the exposed and flexible regions more readily accessible to enzymes than the structured regions³¹. Recently, limited digestion with carefully selected enzymes has found its application in the biopharma field for subunit analysis of monoclonal antibodies and different types of biotherapeutics and variant stability evaluation^32–36. We expected that limited digestion with an appropriate enzyme and optimization of the enzyme-to-substrate ratio would result in the cleavage of the linker and separation of the two antigen-binding domains of the BsABP. Limited digestion of the BsABP was carried out using Lys-C as the enzyme of choice. A significant cleavage occurred at one of the C-terminal of lysine residue at position 251 and 271within the linker peptide connecting the two domains of the BsABP molecule under non-reducing and non-denaturing conditions, resulting in the separation of two domains of BsABP, i.e., antigen-binding domain-1 and antigen-binding domain-2. The developed novel CEX-MS method was used to analyze limited digested samples, detect modifications specific for two domains of the BsABP and validate results observed at an intact level.

The unstressed, thermally stressed, and photo-stressed samples were treated by limited Lys-C digestion and analyzed using the developed CEX-MS method. In Figure 5A, Table 5B, and Table S2, the unstressed sample showed the presence of amino acid substitution and deamidation in antigen-binding domain-2. At the same time, O-glycosylation was observed for antigen binding domain-1, as shown in Figure 5A (black trace). The thermally stressed sample resulted in the detection of new peaks with masses corresponding to isomerization and succinimidation of antigen-binding domain 1, as seen in Figure 5A (red trace). Photo-stressed material showed the presence of oxidation observed for the two domains of the BsABP molecule. Isomerization was observed for antigen-binding domain-1 of the BsABP molecule. Modifications predicted for new peaks observed in limited digest samples correlated with results observed from the intact molecule, as demonstrated in Figure 5A (blue trace). This approach allows for the reliable assignment of the PTMs in the two domains of BsABP. All analyses were performed in triplicate, and they exhibited similar chromatography and mass-spectrometry profiles, indicating the repeatability of the developed method.

Figure 5:

Domain-specific modifications in the limited digestion samples of the BsABP molecule and the forced-degraded product observed by native CEX-MS. (A) An overlay of native (black), thermal-stress (red), and photo-stress (blue) BsABP molecule, along with the modifications presented for each domain. (B) The summary of annotation symbols and corresponding modifications

3.4. Peptide mapping

Peptide mapping was performed to confirm the results from native BsABP and limited digest samples and to identify the amino acid residues undergoing modifications. Reversed-phase chromatography was implemented to trypsin-digested peptides, and separated peptides were further analyzed for accurate mass identification using Orbitrap mass analyzer along with conformation of amino acid residues where modifications take place by MS/MS fragmentation performed in HCD (high energy collisional dissociation) cell.

Modifications were confirmed by the determination of the retention time difference between the native (non-modified) and modified peptide species and verification of the amino acid residue location based on MS/MS data for selected modifications, including deamidation, oxidation, and amino acid substitution. Isomerization of aspartic acid residue to iso-aspartic (iso-Asp) residue does not induce any mass shift, nor does HCD-based fragmentation provide any information about the location of iso-Asp formation from the b and y sequence ions. Hence, identification was based on the retention time shift of the modified peptide, wherein isomerized species eluted at a different time compared to the native peptide to confirm the presence of such modification. Table 1 lists the amino acid-specific modifications correlated with modifications observed from CEX-MS characterization of intact and limited digestion of BsABP. Figure 6 shows MS/MS confirmation of glycine to aspartic acid substitution observed for antigen-binding domain-2. A mass shift of +58.01 m/z units was observed starting from y₃ and consecutive larger y sequence ions. Our analysis revealed higher levels of tryptophan dioxidation product, dihydroxypyroloindole (+31.98 Da), compared to methionine dioxidation product, i.e., methionine sulfone (+31.98 Da), in the photo-stressed sample; we hypothesize that this difference in oxidation states of the sulfur atom in case of methionine and the carbon atom in case of tryptophan may lead to variation in overall charge, resulting in elution of oxidized species as acidic variants when the photo-stressed sample was analyzed by native CEX-MS. We were unable to identify sites for O-glycosylation because the overall heterogeneity of BsABP could lead to difficulty in the separation of glycosylated and non-glycosylated species using reversed-phase chromatography. Alternative methods, such as hydrophilic liquid chromatography or capillary electrophoresis coupled with mass spectrometry, could be explored to identify sites of O-glycosylation.

Figure 6:

MS/MS conformation of amino acid substitution for BsABP molecule. MS/MS spectra with annotated y sequence ions are shown for the modified peptide (A, red) and native/unmodified peptide (B, black).

4. Conclusion

An efficient workflow for rapid and unambiguous identification of PTMs and storage-related chemical modifications of the BsABP molecule has been successfully developed. The method can detect low-level variants such as amino acid substitution and O-glycosylation, as well as deamidation, isomerization, oxidation, and succinimide formation. Limited digestion followed by native ion-exchange chromatography revealed modifications specific to two domains of the BsABP molecule, verified by peptide mapping that led to pinpointing of amino acid residues involved in the observed modifications to the two domains of the BsABP molecule. Applying this workflow can save a considerable amount of time during the product characterization phase compared to a traditional approach. Moreover, the developed workflow reduces sample artifacts (i.e., modifications induced during sample preparation and analysis) generated by the conventional approach. The presented here native cation-exchange chromatography interfaced with MS method can also be employed during the process development to characterize the influence of different biosynthesis or purification steps on removing or incorporating modifications. The developed workflow can be optimized for various protein therapeutics, such as monoclonal antibodies, Fc-fusion proteins, and enzyme replacement therapies, to provide critical information on CQAs.

• A workflow for charge variant characterization of novel therapeutic bispecific antigen binding protein was designed

• A highly sensitive native mass spectrometry-based approach was developed to identify predominant and low-level post-translation enzymatic and non-enzymatic chemical modifications of the bispecific antigen-binding protein

• The developed workflow achieved rapid and unbiased characterization of charge variants with lower sample consumption and faster analysis at a lower cost compared to traditional workflows

Abbreviations

BsABP

Bispecific antigen-binding protein

CEX-MS

Cation-exchange chromatography coupled to mass spectrometry

CQAs

Critical quality attributes

CV

coefficient of variance

EIC

Extracted ion chromatogram

HCD

High energy collision dissociation

HMR

High mass range

MS

Mass spectrometry

pI

Isoelectric point

PTMs

Post-translation modifications

QC

Quality control

RP-HPLC-MS

reverse phase chromatography coupled to mass spectrometry

ScFv

Single-Chain fragment variable

TIC

Total ion chromatogram

V_H

Variable heavy chain

V_L

Variable light chain