Automated stopped-flow library synthesis for rapid optimisation and machine learning directed experimentation.

For the discovery of new candidate molecules in the pharmaceutical industry, library synthesis is a critical step, in which library size, diversity, and time to synthesise are fundamental. In this work we propose stopped-flow synthesis as an intermediate alternative to traditional batch and flow chemistry approaches, suited for small molecule pharmaceutical discovery. This method exploits the advantages of both techniques enabling automated experimentation with access to high pressures and temperatures; flexibility of reaction times, with minimal use of reagents (µmol scale per reaction). In this study, we integrate a stopped-flow reactor into a high-throughput continuous platform designed for the synthesis of combinatory libraries with at-line reaction analysis. This approach allowed ∼900 reactions to be conducted in an accelerated timeframe (192 hours). The stopped flow approach used ∼10% of the reactants and solvents compared to a fully continuous approach. This methodology demonstrates a significantly improved synthesis success rate of smaller libraries by simplifying the implementation of cross-reaction optimisation strategies. The experimental datasets were used to train a feed-forward neural network (FFNN) model providing a framework to guide further experiments, which showed good model predictability and success when tested against an external set with fewer experiments. As a result, this work demonstrates that combining experimental automation with machine learning strategies can deliver optimised analyses and enhanced predictions, enabling more efficient drug discovery investigations across the design, make, test and analysis (DMTA) cycle.

Rapid approach to complex boronic acids.

The compatibility of free boronic acid building blocks in multicomponent reactions to readily create large libraries of diverse and complex small molecules was investigated. Traditionally, boronic acid synthesis is sequential, synthetically demanding, and time-consuming, which leads to high target synthesis times and low coverage of the boronic acid chemical space. We have performed the synthesis of large libraries of boronic acid derivatives based on multiple chemistries and building blocks using acoustic dispensing technology. The synthesis was performed on a nanomole scale with high synthesis success rates. The discovery of a protease inhibitor underscores the usefulness of the approach. Our acoustic dispensing-enabled chemistry paves the way to highly accelerated synthesis and miniaturized reaction scouting, allowing access to unprecedented boronic acid libraries.

Acoustic Droplet Ejection Enabled Automated Reaction Scouting.

Miniaturization and acceleration of synthetic chemistry are critically important for rapid property optimization in pharmaceutical, agrochemical, and materials research and development. However, in most laboratories organic synthesis is still performed on a slow, sequential, and material-consuming scale and not validated for multiple substrate combinations. Herein, we introduce fast and touchless acoustic droplet ejection (ADE) technology into small-molecule chemistry to transfer building blocks by nL droplets and to scout a newly designed isoquinoline synthesis. With each compound in a discrete well, 384 random derivatives were synthesized in an automated fashion, and their quality was monitored by SFC-MS and TLC-UV-MS analysis. We exemplify a pipeline of fast and efficient nmol scouting to mmol- and mol-scale synthesis for the discovery of a useful novel reaction with great scope.

Automated and Accelerated Synthesis of Indole Derivatives on a Nano-Scale.

Automated, miniaturized and accelerated synthesis for efficient property optimization is a formidable challenge for chemistry in the 21st century as it helps to reduce resources and waste and can deliver products in shorter time frames. Here, we used for the first-time acoustic droplet ejection (ADE) technology and fast quality control to screen efficiency of synthetic reactions on a nanomole scale in an automated and miniaturized fashion. The interrupted Fischer indole combined with Ugi-type reactions yielded several attractive drug-like scaffolds. In 384-well plates, a diverse set of interrupted Fischer indole intermediates were produced and reacted to the tricyclic hydantoin backbone by a 2-step sequence. Similarly, preformed Fischer indole intermediates were used to produce divers sets of Ugi products and the efficiency was compared to the in-situ method. Multiple reactions were resynthesized on a preparative millimole scale, showing scalability from nano to mg and thus synthetic utility. An unprecedented large number of building was used for fast scope and limitation studies (68 isocyanides, 72 carboxylic acids). Miniaturization and analysis of the generated big synthesis data enabled deeper exploration of the chemical space and permitted gain of knowledge that was previously impractical or impossible, such as the rapid survey of reactions, building block and functional group compatibility.

Identification of Mineralocorticoid Receptor Modulators with Low Impact on Electrolyte Homeostasis but Maintained Organ Protection.

The mechanism-based risk for hyperkalemia has limited the use of mineralocorticoid receptor antagonists (MRAs) like eplerenone in cardio-renal diseases. Here, we describe the structure and property-driven lead generation and optimization, which resulted in identification of MR modulators ( S)-1 and ( S)-33. Both compounds were partial MRAs but still demonstrated equally efficacious organ protection as eplerenone after 4 weeks of treatment in uni-nephrectomized rats on high-salt diet and aldosterone infusion. Importantly, and in sharp contrast to eplerenone, this was achieved without substantial changes to the urine Na+/K+ ratio after acute treatment in rat, which predicts a reduced risk for hyperkalemia. This work led to selection of ( S)-1 (AZD9977) as the clinical candidate for treating MR-mediated cardio-renal diseases, including chronic kidney disease and heart failure. On the basis of our findings, we propose an empirical model for prediction of compounds with low risk of affecting the urinary Na+/K+ ratio in vivo.

Lipophilicity prediction of peptides and peptide derivatives by consensus machine learning.

Lipophilicity prediction is routinely applied to small molecules and presents a working alternative to experimental log P or log D determination. For compounds outside the domain of classical medicinal chemistry these predictions lack accuracy, advocating the development of bespoke in silico approaches. Peptides and their derivatives and mimetics fill the structural gap between small synthetic drugs and genetically engineered macromolecules. Here, we present a data-driven machine learning method for peptide log D 7.4 prediction. A model for estimating the lipophilicity of short linear peptides consisting of natural amino acids was developed. In a prospective test, we obtained accurate predictions for a set of newly synthesized linear tri- to hexapeptides. Further model development focused on more complex peptide mimetics from the AstraZeneca compound collection. The results obtained demonstrate the applicability of the new prediction model to peptides and peptide derivatives in a log D 7.4 range of approximately -3 to 5, with superior accuracy to established lipophilicity models for small molecules.

Increasing the delivery of next generation therapeutics from high throughput screening libraries.

The pharmaceutical industry has historically relied on high throughput screening as a cornerstone to identify chemical equity for drug discovery projects. However, with pharmaceutical companies moving through a phase of diminished returns and alternative hit identification strategies proving successful, it is more important than ever to understand how this approach can be used more effectively to increase the delivery of next generation therapeutics from high throughput screening libraries. There is a wide literature that describes HTS and fragment based screening approaches which offer clear direction on the process for these two distinct activities. However, few people have considered how best to identify medium to low molecular weight compounds from large diversity screening sets and increase downstream success.

Discovery of 4-morpholino-pyrimidin-6-one and 4-morpholino-pyrimidin-2-one-containing Phosphoinositide 3-kinase (PI3K) p110ß isoform inhibitors through structure-based fragment optimisation.

The discovery of 4-morpholino-pyrimidin-6-one and 4-morpholino-pyrimidin-2-one-containing inhibitors of Phosphoinositide 3-kinases (PI3K) p110ß isoform is reported. Structure-based optimisation of the original fragment hit resulted in lead compounds with improvements in ligand efficiency, lipophilicity efficiency, p110ß potency and selectivity over p110α.

"@-Tides": the 1,2-dihydro-3(6H)-pyridinone unit as a beta-strand mimic.

The cyclic amino acid surrogate 1 was designed to mimic the extended conformation of a peptide unit and to provide hydrogen bond donor and acceptor functions conducive to beta-sheet formation. A convenient synthesis of this unit and solution and solid-phase methods for its incorporation into an oligomer alternating with peptide units have been devised. The resulting "@-tides", as these oligomers have been designated, show a high propensity for self-association in comparison to oligopeptides; insights into the structure and dynamical properties of their antiparallel dimers have been obtained by NMR.