Transforming data to understanding and knowledge for the sequencing of therapeutic oligonucleotides by mass spectrometry

Abstract

Recently, the numbers of oligonucleotides approved to the market for medicines has increased1 and take an important place in new drugs for undruggable diseases. For the safety of the patients, the drug, as well as any impurities from the synthesis, must be characterised.
Oligonucleotides can be analysed by negative ion electrospray ionisation (ESI) mass spectrometry with collision-induced dissociation (CID) where different product ions characteristic of the sequence are obtained. To understand and interpret the data, an expert analyst is needed to reconstruct the full sequence. Depending on the complexity and the length of the oligonucleotide and impurities present, it can take days, weeks, or months to sequence them manually. This can directly impact the time for the drug to be approved to the market. Some software packages exist to help the interpretation of the data but are limited due to a lack of detailed understanding how oligonucleotides dissociate in the gas phase inside the mass spectrometer.
To predict the dissociation of oligonucleotides, it is important to understand how simple sequences dissociate. To probe that, a library of unmodified and modified 21 mer oligonucleotides (phosphorothioate backbone rather than phosphodiester backbone) were synthesised to understand the impact of the different nucleobases present in the sequences, their localisation, and the impact of the different phosphate backbones on the dissociation. This library was analysed by direct infusion and RP-IP-LC-MS to obtain different charge states that were dissociated (with or without isolation) by isCID and CID. After dissociation the different charge states give different information and sequence coverages due to their conformation in the gas phase. During the assignment of the product ions, IMS has been used to help identify the charge states of the diverse product ions. After assignment of the product ions, the data were compared for the phosphodiester sequences, the phosphorothioate sequences, and then between the phosphodiester and phosphorothioate sequences.
The results from the library by high resolution negative ion ESI and CID mass spectrometry in the gas phase has shown that the dissociation is driven by the proton affinity of the nucleobases, the secondary structure by Watson-Crick pairing, the number of each nucleobase present in the sequence, their position, and their neighbours, depending on the charge state dissociated. By analysing, understanding, and comparing the same sequences with and without phosphorothioate phosphate backbone, new rules and observations have been obtained which are affected by modifying the phosphate backbone where the proton affinity of the nucleobases is impacted.
The use of this library will help to improve the current software to be more accurate and to predict the dissociation of more complex sequences depending to the charge state dissociated. Furthermore, the observations and rules obtained here could be used to develop a database to be incorporated into machine learning to predict the dissociation of known oligonucleotide structures and then characterise unknown oligonucleotides automatically. It would help the pharmaceutical industry to develop and characterise therapeutic oligonucleotides quicker, as well as their impurities, used for undruggable diseases. This will affect the lives of millions of patients worldwide to have access to their medicine faster.

More information

Identifiers

ORCID for Fabien Hannauer: orcid.org/0000-0002-8277-9817

ORCID for John Langley: orcid.org/0000-0002-8323-7235

ORCID for Eugen Stulz: orcid.org/0000-0002-5302-2276

Catalogue record

Export record