A generalised mass spectrometry method leverages electrophilicity to rapidly rank reaction outcomes in experimental screens. The new approach attaches a nucleophilic tag to the product molecules, creating a common chemical fingerprint which is then easily detected in the subsequent analysis. This simplified detection step evaluated samples more than 60 times faster than conventional analytical methods and addresses a key bottleneck in high throughput experimentation and compound library generation.
High throughput screening has dramatically accelerated the rate of drug discovery and optimisation. But, while automation successfully streamlined compound preparation, the efficiency of the concomitant analysis remains limited by the need for bespoke assays and chromatographic methods for each reaction campaign.
Previous work by Daniel Blair at St Jude Children’s Research Hospital in the US has already stripped back the complexity of this analytical step in an effort to accelerate progress. Last year the group developed a generalised mass spectrometry method that used the fragmentation fingerprint of the starting material to compare different reaction outcomes over a full screen. ‘When you’re trying to analyse most molecules the same way, you want to look for something that’s going to be broadly applicable across a wide swathe of chemical matter,’ he explains. ‘One starting material would define all of its products in the manifold. But, as with any methodology, there are always gaps.’ For instance, the method only works for reaction campaigns with a common starting material.
However, Blair’s team has now extended this logic one step further and developed an orthogonal method which instead leverages chemical properties – namely electrophilicity – to generalise the analysis across multiple different starting materials.
The core principle still depends upon the generation of a common fragmentation fingerprint across all reactions, something electrophilicity alone can’t produce. However, by exploiting this common feature, the team was able to install additional structural units, carefully chosen to dominate the final fragmentation pattern and facilitate direct comparison between samples. Specifically, they focused on sulfur-centred nucleophilic tags as the product label, paralleling the interactions of cysteine residues on target proteins with electrophilic small molecule drugs.
To translate this concept into a proof-of-concept, the team screened conditions for a Buchwald–Hartwig reaction between an acrylamide and six different amines, evaluating 384 unique reactions in total. Each mixture was conjugated to a thiol probe post-workup and the resulting adducts were analysed in less than 20 minutes – roughly 60 times faster than standard LCMS analysis – enabling the researchers to quickly identify the best conditions for each amine substrate tested. ‘Whether or not one compares the parent electrophile on its own by LCMS, or the conjugated product using [this method], our ability to discern differential reaction outcomes was equally as effective,’ says Blair. ‘This was very reassuring and shows that the conjugation step is very efficient at converting the product to the adduct.’
Crucially though, the conjugation step itself was also generalisable. The team’s collection of sulfur-centred tags proved compatible with most other classes of electrophile and cross-coupling reaction and the predictable fragmentation of these added labels provided a consistent point of reference across more than 90% of substrates.
‘This is a really exciting and innovative development,’ says Perdita Barran, director of the Michael Barber Centre for Collaborative Mass Spectrometry at the University of Manchester. ‘The simplicity of the readout could have a lot of benefit to synthetic chemists who are doing combinatorial chemistry and making large compound libraries. I hope it will make this type of method more accessible to [them].’
With this bottleneck seemingly fixed, Blair is already turning his attention to the next. ‘We’re really excited about hyphenating all of those techniques with ways that cut out the next slow step, which is finding functions for those molecules,’ he says. ‘I think deploying this where we can pick out the winners in a biological context is where we’re really going next.’
