Selective Inactivation of a Human Neuronal Silencing Phosphatase by a Small Molecule Inhibitor

The C-terminal domain (CTD) of eukaryotic RNA polymerase II serves as a binding scaffold for numerous regulatory factors involved in transcription, RNA processing, histone modificationm, and various other cellular events.1 During transcription, the long, unstructured, and largely repetitive CTD undergoes dynamic phosphorylation and dephosphorylation cycles, creating different phosphorylation patterns (termed the “CTD code”) that temporally and spatially orchestrate the recruitment of those regulatory factors. Though inherently disordered, the CTD is evolutionarily conserved from yeast to human.2 The CTD usually consists of 26–52 tandem heptapeptide repeats with the consensus sequence Y1S2P3T4S5P6S7.2 Phosphorylation of the CTD (with serines at positions 2 and 5 being the primary phosphorylation sites) is a major mechanism by which cells regulate gene expression. The CTD specific kinases and phosphatases function as house-keeping regulatory factors for global transcription.3 Recently, it has been shown that certain CTD regulatory factors can also epigenetically modulate the expression level of a specific group of genes.4–6 As an example, a newly discovered class of phosphatases, the human small C-terminal domain phosphatases (Scp's), specifically dephosphorylates phosphorylated Ser5 (phospho.Ser5) of the tandem heptad repeats of the CTD.7 Interestingly, Scp's have also been shown to epigenetically silence the expression of a specific set of neuronal genes in neuronal stem cells and non-neuronal cells by acting as co-repressors in REST/NRSF complex.4 Inhibition of Scp's in P19 stem cells by the dominant negative mutants (which retain the overall structure but have abolished phosphatase activity) or microRNA miR-124 (which directly targets the untranslated region of Scp genes and suppresses their expression) allows neuronal gene expression and induces neuronal differentiation.4,8 Given the demonstrated role of Scp's in limiting inappropriate expression of neuronal specific genes in pluripotent cells and the fact that their down-regulation leads to neuronal differentiation, Scp's serve as promising new targets for small molecule inhibitors to regulate neuronal stem cell development and to promote neuronal differentiation. However, one of the greatest challenges associated with phosphatase inhibitor identification is the cross inhibition of other phosphatases due to poor selectivity,9 which usually stems from small, uncharacteristic active sites of phosphatases. For Scp inhibitors, selectivity is of great concern as two close family members in the Scp/Fcp family, Fcp110 and Dullard,11 play essential roles in cell survival as well as proper development.11–13 The crystal structures of Scp/Fcp family members Fcp1 and Scp1 were initially solved by Ghosh et al.14 and Kamenski et al.15 Unlike the traditional cysteine-based or dimetal-dependent phosphatases, the Scp/Fcp family members belong to a unique family of phosphatases that rely on the DxDx(T/V) motif and Mg2+ to catalyze the phosphoryl-transfer,7,11 as found in the haloacid dehydrogenase (HAD) superfamily.16 In fact, the overall core fold of Scp's resembles the core domains of other HAD family members, despite low sequence similarity.17 The HAD superfamily can be further divided into three subfamilies (Supplementary Figure S1) according to the presence and the location of a second domain known as the “cap domain”.18,19 Both type I and II subfamily members utilize the cap domain to shield the active site from bulk solvent and to achieve substrate recognition.16 The type III HAD superfamily, including the Scp/Fcp family, do not have a cap domain, requiring specificity achieved through alternative strategies, such as recruitment of other regulatory proteins. This feature poses an additional challenge for inhibitor design. In fact, no specific inhibitors have been reported to date for type III HAD family members. We have previously solved the complex structure of Scp1 bound to its substrate peptide,20 captured snapshots of the phosphoryl-transfer reaction at each step and the formation of the phospho-aspartyl intermediate using X-ray crystallography, and established the reaction mechanism of Scp's.17 These crystallographic and biochemical studies not only provided insights into the catalytic mechanism of Scp/Fcp but also hinted a novel strategy of specific recognition of substrates by Scp's. The complex structure of Scp1 bound to its substrate peptide revealed a hydrophobic binding pocket that is specific for Scp's, suggesting that a specific Scp inhibitor might be obtained through targeting this pocket.20 In the present study, we exploited this possibility and identified rabeprazole as the first reported lead compound for Scp inhibition (Ki = 5 ± 1 µM). This small molecule shows no inhibition toward Fcp1 or Dullard, nor toward bacteriophage λ Ser/Thr phosphoprotein phosphatase (λPPase). This extraordinary selectivity can be explained through analysis of our high resolution structure of Scp1 complexed with the compound, which shows, as expected, the compound binds specifically to the unique hydrophobic binding pocket of Scp's. The structure highlights the chemical functional groups that make essential contributions to binding. To the best of our knowledge, this is the first selective lead compound for the Scp/Fcp phosphatase family, as well as the type III HAD family.