Tumour cells develop drug resistance when mutations accumulate in their genome that drive phenotypic diversity. Normally, cytosine deaminases APOBEC3A-H1 are part of antiviral defence, mutating C-to-U on single-stranded DNA. However, many drug-resistant tumours exhibit mutagenic signatures associated with increased activity of specifically APOBEC3A and APOBEC3B (A3A/B), which recognise only a TC-motif in a minimal 4-nucleotide recognition sequence.2 Therefore, inhibition of A3A/B working alongside existing front-line anticancer therapies could potentially extend their efficacy. Within the genome, the loops of DNA-hairpin structures are known hotspots for mutagenesis.3
Here we reveal through X-ray structures of A3A in complex with modified DNA stem-loop (hairpin) inhibitors and substrates the structural basis of potent inhibition of A3A by hairpin-DNA bearing 2’-deoxy-5-fluorozebularine (Ki=7.8 ± 0.8 nM) (and other moieties4). We have recently explored design space around the stem and the loop of the hairpin DNA to (i) increase hydrophobicity for membrane transfer without transfection reagents, (ii) prevent digestion by cellular nucleases through selective phosphorothiolation of phoshate linkers , (iii) perturb loop dynamics for greater potency, (iv) reduce cytotoxicity to negligible levels, and (v) modify the inhibitor moiety itself to better mimic the transition-state.
Importantly, nuclease-resistant derivatives of these inhibitors maintain in-vitro potency against A3A and potently inhibit mutagenic activity of A3A in cancer.5