TL;DR
Researchers have developed a streamlined and faster method (SP-MEGD combined with Fusion assembler) for sequencing antibodies, crucial for developing treatments against viruses like SARS-CoV-2. This new approach significantly improves throughput and accuracy, even for complex mixtures of antibodies and previously unsequenced antibodies.
What Problem Does This Paper Address?
The rapid development of effective antibody therapies, especially for rapidly evolving viruses like SARS-CoV-2, hinges on accurately and quickly determining antibody sequences. Traditional methods for antibody sequencing, such as hybridoma technology, are time-consuming (months) and costly. While mass spectrometry (MS)-based de novo sequencing offers a faster alternative, it has historically suffered from high costs, low throughput, and challenges with data completeness and accuracy, particularly when dealing with mixtures of antibodies or antibodies with unknown sequences. This paper aims to overcome these limitations by presenting a more efficient, robust, and accurate MS-based workflow for antibody sequencing.
Methods
The core of this work lies in a novel sample preparation method called SP-MEGD (Single-Pot and Multi-Enzymatic Gradient Digestion). This method streamlines the process by:
- Gradient Digestion: Utilizing a combination of five proteases (trypsin, chymotrypsin, pepsin, elastase, and Asp-N) in a single pot over a 6-hour period, with samples collected at 2-hour intervals. This "gradient" approach aims to generate a diverse range of peptide fragments, increasing the chances of overlapping sequences needed for de novo assembly.
- Optimized Lysis Buffer: Identifying Guanidine hydrochloride (Gu.HCl) as the most effective lysis buffer for achieving high protein sequence coverage.
- Dual Fragmentation Strategy: Employing both High-Energy Collision Dissociation (HCD) and Electron-Transfer High-Energy Collision Dissociation (EThcD) fragmentation techniques during LC-MS/MS analysis. These methods are complementary, with HCD providing broad fragmentation and EThcD being particularly effective for higher charged states, together maximizing sequence coverage.
The resulting peptide data is then processed using a trained Casanovo model for de novo peptide sequencing. Finally, these peptides are assembled into complete antibody sequences using the Fusion assembler, a parameter-independent algorithm previously developed by the authors. This assembler is designed to efficiently reconstruct antibody sequences from mixtures.
Datasets
The study primarily uses antibody samples for training and validation. Specifically:
- S2P6LH antibody: Used extensively for optimizing the SP-MEGD method and evaluating its performance. Its known sequence served as a reference.
- Anti-CD8 and Anti-CD4 antibodies: Used as test cases to demonstrate the workflow's ability to sequence commercial antibodies with unknown sequences.
- Mixtures of COVID-19 neutralizing antibodies (S2P6LH, BD5514LH, and BD5840LH): Used to showcase the method's capability in simultaneously discriminating and sequencing multiple antibodies.
The Casanovo model was trained on a large dataset of human proteome spectra from MassIVE (2,154,269 spectra from 1,114,503 unique peptides).
Key Insights and Ideas
- Gradient Digestion for Diversity: The gradient digestion approach, collecting samples at different time points, is key to generating a rich peptide library. This diversity is crucial for de novo sequencing, where overlapping fragments are essential for accurate assembly.
- SP-MEGD as a "One-Pot Wonder": Streamlining sample preparation into a single pot with optimized reagents significantly reduces sample loss and processing time compared to traditional multi-step methods.
- Fusion Assembler for Complex Mixtures: The Fusion assembler's ability to handle mixtures of antibodies is a significant advancement, allowing for the simultaneous sequencing of multiple antibody molecules.
- Correcting "Authentic" Sequences: The study highlights how their method can even identify and correct potential mutations in antibody sequences, ensuring the determination of the true, authentic sequence.
Key Results
- High Sequence Accuracy: The SP-MEGD and Fusion workflow achieved over 99% sequence accuracy for mixtures of three COVID-19 neutralizing antibodies.
- Complete Coverage: The method demonstrated complete sequence coverage (100%) for both heavy and light chains of the S2P6LH antibody, including critical CDR regions.
- Low Sample Input: The SP-MEGD method is effective with as little as 50 µg of antibody sample, making it suitable for precious samples.
- Efficient Sequencing of Unknown Antibodies: The workflow successfully sequenced commercial anti-CD8 and anti-CD4 antibodies for which no public sequences were available, and the resulting sequences were functionally validated.
- Outperforms Commercial Kits: SP-MEGD demonstrated superior performance compared to a commercial enrichment kit (EasyPept-Micro) in terms of sequence coverage.
Limitations and Future Work
- Distinguishing Similar Amino Acids: The paper notes a slight decrease in accuracy when distinguishing between leucine (L) and isoleucine (I) in complex mixtures, suggesting this remains a challenge.
- Further Validation: While intact mass analysis and functional validation were performed, further biochemical validation was suggested for some sequences (e.g., anti-CD8 heavy chain CDR3 region) to confirm accuracy.
- Potential for Drug Screening: The authors suggest their method could be a powerful tool for small-molecule drug screening in future applications, implying a broader utility beyond just antibody sequencing.
- Advancing Serology: The high-quality spectra obtained are seen as a stepping stone towards a new era of advanced serology, where antibody sequences can be directly derived from bodily fluids.
Summary updated