Fragments vs CDC14 phosphatases

Practical Fragments has periodically written about protein tyrosine phosphatases (PTPs), which remove phosphate moieties from tyrosine side chains in proteins. Despite decades of attention, progress towards selective inhibitors has been slow due to both the similar active sites and their highly charged nature. A new paper in J. Med. Chem. by Zhong-Yin Zhang and colleagues provides some hope.

 

The researchers were interested in CDC14 phosphatases, so-called dual-specificity phosphatases that can dephosphorylate phosphoserine and phosphothreonine in addition to phosphotyrosine. Two members of this family, hCDC14A and hCDC14B, are widely expressed in humans, but their role in cancer is ambiguous, with some studies suggesting they are oncoproteins while others suggest they may have a protective function. Clearly a chemical probe would be useful.

 

The researchers started by considering non-hydrolyzable phosphotyrosine mimetics, specifically those replacing the central oxygen with a difluoromethyl moiety; we wrote about this bioisostere back in 2013. Eight fragments were made and assessed at 1 mM in aqueous buffer to demonstrate they did not aggregate. They were then tested in functional assays against a panel of ten PTPs, and compound 9 turned out to be quite potent and selective for hCDC14A. Subsequent experiments showed it to have similar activity against hCDC14B, and Lineweaver-Burk plots revealed it to be a competitive inhibitor of both, as expected.

 

Although no hCDC14A structures have been reported, modeling the compound into a published structure of hCDC14B gave some insights into the binding mode and selectivity. In particular, hCDC14B has a larger active site than some other PTPs, thus explaining why the tricyclic compound 9 could fit. Further analysis suggested the possibility of growing the compound towards a hydrophobic pocket, so the researchers synthesized a small set of molecules, of which compound 15 turned out to be the most potent.

 

Compound 15 was tested against 16 PTPs and found to be quite selective against hCDC14A and hCDC14B, with IC₅₀ values 5 µM or worse against the others. Mutagenesis studies in the hydrophobic pocket were consistent with the proposed binding mode. Despite the presence of the highly charged difluorophosphonate moiety, compound 15 showed activity in cells at low micromolar concentration and had some oral bioavailability in mice.

 

Although better cell activity is probably necessary to make a truly useful chemical probe, this is a nice start. Researchers at AbbVie have taken a competitive inhibitor of a different PTP into the clinic, so perhaps we will start to see more successes against these challenging enzymes.