A 2016 paper by workers at GSK, published in J. Med. Chem., really grabbed my attention. GSK appear to have achieved the difficult feat of discovering a molecule that exhibits monokinase selectivity! Intriguingly, the initial hit was discovered not through screening the GSK kinase inhibitor set, nor through the HTS screening of the GSK compound collection (~2 million compounds). The novel benzoxazepinone series was actually found by screening GSK’s DNA-encoded library.
In its simplest form a DNA-encoded library is constructed with chemical building blocks that are individually ‘tagged’ with DNA fragments. The DNA serves as an amplifiable genetic barcode. Millions (or billions) of compounds can be produced in weeks using ‘routine’ combichem approaches (Fig 1). The resulting libraries can be screened in a single test tube against the target. Hits are read/identified by PCR amplification of DNA barcodes – Fantastic! A smart and efficient way of getting around the arduous deconvolution involved with combinatorial libraries.
Figure 1 taken from a fantastic review of GSK’s contribution to the field of DNA-encoded libraries (Med. Chem. Commun., 2016 016, 7, 1898–1909)
An advantage of screening DNA-encoded libraries is that they enable a more thorough exploration of chemical space (4 – 5 orders of magnitude more!) than is achievable by traditional HTS methods. To give you an idea of the space explored, just one of the many DNA-encoded libraries that GSK screened to find the benzoxazepinone hit had a ‘total warhead diversity of approx. 7.7 billion’.
Figure 2 taken from paper (J. Med. Chem. 2016, 59, 2163−2178)
The benzoxazepinone series was selected for further investigation because it appeared to be far removed from what is perceived as the ‘same old’ kinase motif. Fortunately for the team on the project, this line of enquiry furnished a novel chemotype with tractable SAR. The initial hit matter was progressed to a lead that had desirable properties for a kinase inhibitor (i.e. excellent potency, exquisite kinase specificity, oral bioavailability etc. etc.) and also appears to be competitive with ATP whilst making no interactions with the hinge of the kinase at all! (a.k.a. Type III kinase inhibitor). The atypical binding mode of the inhibitor was elucidated using a raft of techniques including photoaffinity labeling, hydrogen-deuterium exchange analysis and co-crystallography.
DNA-encoded libraries are an accessible screening platform to both academic groups (whilst undertaking my Master’s project I saw the creation of (very similar) PNA-encoded libraries) and industry. The reason for an academic group to screen via this method is obvious – the combinatorial platform enables them to amass vast compound libraries that would have taken a substantial cash injection and multiple (hu)man hours to produce. For industry, the DNA-encoded libraries represent another screening platform to churn out novel hit matter, casting the net deep into chemical space.
The construction and screening of DNA-encoded libraries is a relatively young field (~ 10 years). Reports of high-affinity hits for biological targets are becoming more common. Ultimately the quality of the library and its output will depend on the diversity/novelty of each building block and the diversity of reactions employed in its construction. For sure DNA-encoded libraries (and other encoded libraries) represent a step forward from the usual solid-phase mix and split OBOC libraries of the past. Improvements in methodology (encoding and selection methods), library construction (arrays of 3D- and diverse building blocks, reaction diversity etc.), library curation and data analysis are all possible growth areas in the future.
GSK’s paper also serves as a useful reminder that choosing an unusual starting point – especially in the Kinase field where many privileged kinase motifs exist in the public domain – may guide you into hitherto unexplored chemical space. The benefits of a proprietary starting point are obvious – the DNA-encoded library screening platform employed in this paper was critical to the discovery of novel, selective and potent hit matter.
For recent reviews read ,  and .
Blog written by Scott Henry Henderson
 Harris et al., J. Med. Chem. 2016, 59, 2163
 Zimmermann and Neri., Drug Discovery Today. 2016, 21, 11, 1828
 Goodnow, Dumelin and Keefe., Nature Reviews Drug Discovery (2016) [doi:10.1038/nrd.2016.213]
 Arico-Muendel., Med. Chem. Commun., 2016 016, 7, 1898–1909