Targeting cancer stem cell biology by enzymatic inhibition

Project Overview

Figure 1: Schematic of the let-7 miRNA regulatory pathway.

Figure 2: Structures of Cid1, the yeast cytoplasmic terminal uridylyl transferase, in two Apo and a ...

Cancer stem cell biology is controlled by the let-7 miRNA regulatory network (see Figure 1). Let-7 miRNAs suppress the expression of a number of key oncogenes including the RAS family, MYC, Bcl-2, HER2 and Lin28A. Let-7 miRNA maturation is controlled by terminal uridylyl transferase (TUT) enzymes in the cytoplasm which mono- or oligo-uridylate pre-let-7 miRNAs. Mono-uridylation leads to miRNA maturation and oncogene suppression. Oligo-uridylation is promoted by the cofactor Lin28A and leads to pre-let-7 degradation by the exonuclease DIS3L2. If we can inhibit the human TUTs then we may be able to develop a new therapeutic strategy that boosts endogenous miRNA levels. There is good evidence that a basis for patient stratification exists based on TUT over-expression and/or the presence of Lin28A. Data also indicate that reduction in TUT activity and the boosting of let-7 expression leads to chemo- and radio-sensitisation.

We solved the structure and defined the mechanism of a yeast homologue of the human TUTs which shows a high degree of conservation with the human enzymes (Yates et al., 2012, 2015) (Figure 2). We are now engaged in a drug discovery programme in which we have had success in identifying hits and showing efficacy of TUT inhibitors against the human enzymes. We are now expanding our programme to determine structures of the human enzymes in Apo and functional states and to develop further the drug discovery programme we have started using a combination of structural, biophysical, functional and cellular approaches. The project is also now part of an active commercial collaboration, providing further resources, and new opportunities for application of our findings. You would therefore join this project at a critical and exciting stage.

Our work builds on an excellent network of collaborators including industrial partners and (in Oxford) the Norbury lab based at the Dunn School that discovered terminal uridylylation, the Target Discovery Institute, and the Higgins lab in the Department of Oncology.

Training Opportunities

X-ray crystallography, cryo-EM and structure determination; fragment screening; biophysical interaction analysis; assay development and high-throughput screening; inhibitor development.


Protein Science & Structural Biology and Cancer Biology


Project reference number: 849

Funding and admissions information


Name Department Institution Country Email
Professor Robert Gilbert Structural Biology Oxford University, Henry Wellcome Building of Genomic Medicine GBR
Dr Monica Mantri Division of Structural Biology University of Oxford GBR

Yates LA, Durrant BP, Fleurdépine S, Harlos K, Norbury CJ, Gilbert RJ. 2015. Structural plasticity of Cid1 provides a basis for its distributive RNA terminal uridylyl transferase activity. Nucleic Acids Res., 43 (5), pp. 2968-79. Read abstract | Read more

Terminal uridylyl transferases (TUTs) are responsible for the post-transcriptional addition of uridyl residues to RNA 3' ends, leading in some cases to altered stability. The Schizosaccharomyces pombe TUT Cid1 is a model enzyme that has been characterized structurally at moderate resolution and provides insights into the larger and more complex mammalian TUTs, ZCCHC6 and ZCCHC11. Here, we report a higher resolution (1.74 Å) crystal structure of Cid1 that provides detailed evidence for uracil selection via the dynamic flipping of a single histidine residue. We also describe a novel closed conformation of the enzyme that may represent an intermediate stage in a proposed product ejection mechanism. The structural insights gained, combined with normal mode analysis and biochemical studies, demonstrate that the plasticity of Cid1, particularly about a hinge region (N164-N165), is essential for catalytic activity, and provide an explanation for its distributive uridylyl transferase activity. We propose a model clarifying observed differences between the in vitro apparently processive activity and in vivo distributive monouridylylation activity of Cid1. We suggest that modulating the flexibility of such enzymes-for example by the binding of protein co-factors-may allow them alternatively to add single or multiple uridyl residues to the 3' termini of RNA molecules. Hide abstract

Yates LA, Norbury CJ, Gilbert RJ. 2013. The long and short of microRNA. Cell, 153 (3), pp. 516-9. Read abstract | Read more

MicroRNAs (miRNAs) are versatile regulators of gene expression in higher eukaryotes. In order to silence many different mRNAs in a precise manner, miRNA stability and efficacy is controlled by highly developed regulatory pathways and fine-tuning mechanisms both affecting miRNA processing and altering mature miRNA target specificity. Hide abstract

Yates LA, Lumb CN, Brahme NN, Zalyte R, Bird LE, De Colibus L, Owens RJ, Calderwood DA, Sansom MS, Gilbert RJ. 2012. Structural and functional characterization of the kindlin-1 pleckstrin homology domain. J. Biol. Chem., 287 (52), pp. 43246-61. Read abstract | Read more

Inside-out activation of integrins is mediated via the binding of talin and kindlin to integrin β-subunit cytoplasmic tails. The kindlin FERM domain is interrupted by a pleckstrin homology (PH) domain within its F2 subdomain. Here, we present data confirming the importance of the kindlin-1 PH domain for integrin activation and its x-ray crystal structure at a resolution of 2.1 Å revealing a C-terminal second α-helix integral to the domain but found only in the kindlin protein family. An isoform-specific salt bridge occludes the canonical phosphoinositide binding site, but molecular dynamics simulations display transient switching to an alternative open conformer. Molecular docking reveals that the opening of the pocket would enable potential ligands to bind within it. Although lipid overlay assays suggested the PH domain binds inositol monophosphates, surface plasmon resonance demonstrated weak affinities for inositol 3,4,5-triphosphate (Ins(3,4,5)P(3); K(D) ∼100 μM) and no monophosphate binding. Removing the salt bridge by site-directed mutagenesis increases the PH domain affinity for Ins(3,4,5)P(3) as measured by surface plasmon resonance and enables it to bind PtdIns(3,5)P(2) on a dot-blot. Structural comparison with other PH domains suggests that the phosphate binding pocket in the kindlin-1 PH domain is more occluded than in kindlins-2 and -3 due to its salt bridge. In addition, the apparent affinity for Ins(3,4,5)P(3) is affected by the presence of PO(4) ions in the buffer. We suggest the physiological ligand of the kindlin-1 PH domain is most likely not an inositol phosphate but another phosphorylated species. Hide abstract