Multi-Enzyme Peptide Mapping of Site-Selective Trastuzumab Modification by an Engineered Thermostable Microbial Transglutaminase

Abstract Peptide mapping by liquid chromatography-tandem mass spectrometry (LC–MS/MS) is a key analytical strategy for residue-level characterization of antibody conjugates, providing modification-site assignments and apparent relative modification estimates that complement intact-mass and chromatographic drug-to-antibody ratio (DAR) measurements. The development of homogeneous antibody-drug conjugates (ADCs) relies on precise control of the conjugation site and DAR. For site-selectively modified ADCs, in which amino acid-level homogeneity is an important quality attribute, peptide mapping is therefore not only a confirmatory assay but also a central component of structural characterization. However, its reliability depends on enzymatic coverage and sample-handling conditions that affect the recovery of hydrophobic Fc-derived peptides, particularly in conjugates generated at high modification levels. In this study, we evaluated a multi-enzyme LC–MS/MS peptide-mapping workflow for antibody conjugates generated by microbial transglutaminase (MTG)-mediated Fc modification, using trastuzumab conjugated with dansylcadaverine (DNC) as a model system. An engineered thermostable MTG variant carrying five amino acid substitutions (D3C/S101P/G157S/G250R/G283C) was used to prepare a highly modified trastuzumab-DNC conjugate for evaluation of complementary protease coverage, residue-level modification-site assignment, and recovery of hydrophobic Fc-derived peptides. The engineered MTG was also compared with the commercial MTG formulation Activa using glycosylated trastuzumab under the specified nominal enzyme-loading conditions. Because this comparison was not normalized on the basis of catalytic activity under the antibody-conjugation conditions, it was treated as a formulation-based reference and was not used to establish intrinsic superiority of either enzyme preparation. Complementary digestion with trypsin, chymotrypsin, and Glu-C provided convergent evidence that DNC modification was predominantly localized to heavy-chain Q298 and enabled evaluation of other detectable glutamine-containing peptides. We further found that post-digestion addition of guanidine hydrochloride (GnHCl) during quenching increased the recovery of hydrophobic Fc-derived peptides, most notably the Q298-containing peptides. Omission of GnHCl during quenching reduced the recovery of several hydrophobic Fc-derived peptides and could bias peptide-map-based apparent modification estimates under the tested conditions. Under the optimized workflow, Q298 consistently showed the highest peptide-map-based apparent modification level, whereas apparent modification levels at other unambiguously assigned glutamine residues remained low under the tested conditions. However, these peak-area-based values do not represent absolute molar occupancy, and low-abundance off-site modifications below the analytical detection threshold cannot be excluded. These findings demonstrate that multi-enzyme peptide mapping combined with denaturant-assisted quenching provides a practical analytical workflow for residue-level characterization of highly modified, site-selective antibody conjugates. The workflow supports predominant modification at heavy-chain Q298 under the tested conditions and highlights the importance of sample-preparation optimization in peptide mapping-based characterization of ADCs and related antibody conjugates. Because DNC is a model primary amine, application of this workflow to therapeutic ADC linker-payloads will require substrate-specific optimization and analytical qualification.