De novo antibody sequencing based on mass spectrometry

Although the importance of antibodies in immunity was proposed more than one hundred years ago, comprehensively studying and characterizing antibodies remains a challenge. This thesis demonstrates how mass spectrometry (MS) can be utilized to address this challenge through de novo antibody sequencing on the protein level. In Chapter 1, I reviewed the current state of mass spectrometry-based de novo antibody sequencing. Beginning with an overview of B cell development and antibody diversity, we acknowledged the great diversity of antibodies that safeguard our body against numerous harmful antigens, while also acknowledging the difficulties this poses for their characterization. Recent advancements include partial de novo methods like Ig-seq, which merges single B cell sequencing with bottom-up proteomics to semi-quantitatively explore serological antibody repertoires. Then I described the complete de novo antibody sequencing based on a bottom-up proteomics approach, based on LC-MS/MS. Lastly, I review de novo sequencing based on the middle-down approach as a complementary method to bottom-up, enabling comprehensive antibody repertoire profiling and protein-level sequence insights. In Chapter 2, we introduced a bottom-up approach employing multiple proteases and a dual fragmentation scheme for monoclonal antibody de novo sequencing. This method was successfully examined on a sequence-known mAb Herceptin and a sequence-unknown mAb anti-flag M2. The sequence of anti-flag M2 was validated by expression of this monoclonal antibody in HEK293 cells and comparing its performance in western blot with the input antibody. Moreover, the sequence accuracy was confirmed by remodeling the published crystal structure of the anti-FLAG-M2 Fab. However the de novo sequencing software “Supernovo” that we used in Chapter 2 was limited to monoclonal antibodies. Thereby, in Chapter 3 we introduced a new tool “Stitch” that uses template-based assembly of the peptide reads for de novo antibody sequencing and repertoire profiling. Stitch supports multiple peptide de novo sequencing tools and provides the chance to semi-quantitatively study the V- and C- gene usage and profile the serological antibody repertoire. In Chapter 4, we applied the bottom-up approach that was introduced in Chapter 2 on an anti-MUC1 antibody 139H2 and obtained the sequence in de novo fashion. We used the software PEAKS to de novo sequence peptides and Stitch to assemble the peptide reads into a full sequence. The sequence was successfully validated by comparing the performance of the reverse engineered 139H2 and its Fab fragment to the hybridoma-derived product in Western blot and immunofluorescence microscopy. The sequence enables us to further characterize the binding to the VNTR peptide epitope by surface plasmon resonance (SPR) and solve the crystal structure of the 139H2 Fab fragment in complex with MUC1 VNTR peptide. In Chapter 5, the bottom-up approach we discussed above was employed to an anti-respiratory syncytial virus (RSV) antibody 131-2a, which targets the envelope glycoprotein F. After recombinant expression of 131-2a in HEK293 cells, we validate its sequence accuracy by comparing its performance with the input 131-2a in western blot and ELISA. The precise sequence allows us to further characterize its binding epitope to RSV-post fusion protein by single particle cryo-EM, revealing the molecular basis for 131-2a's specificity to the postfusion conformation of F. Though chapters 2 to 5 underscored the power of the bottom-up approach for monoclonal antibody de novo sequencing, it faces challenges when handling more complex samples. Chapter 6 introduced a hybrid bottom-up and middle-down approach to profile the IgG1 antibody repertoire from a monoclonal gammopathy of undetermined significance (MGUS) patient, successfully obtaining the full sequence from a most abundant clone that has Fab glycosylation. Native MS characterized its Fab glycosylation profile, providing insights into aberrant glycosylation patterns, comparing with the normal Fc glycosylation profile.