Back

The Covalent Binding of Splicing Inhibitors occurs via a Zinc- Assisted Enzymatic-like mechanism

Rozza, R.; Parise, A.; Aupic, J.; Spinello, A.; Pena, V.; Magistrato, A.

2025-09-18 biochemistry
10.1101/2025.09.17.676740 bioRxiv
Show abstract

Branch-site recognition is a pivotal event in spliceosome assembly, and branch-site antagonists disrupt this process, inducing a splicing rewiring that underlies much of their antitumor activity. Antagonists from the spliceostatin family covalently bind the SF3b complex nearby the branch-site pocket, forming an adduct with Cys26 of the PHF5A zinc finger while remarkably preserving Zn2+ coordination - an unconventional mode of covalent inhibition. Yet the molecular features that activate this normally protected ZnCys4 cysteine and initiate spliceostatins epoxide ring opening have remained obscure. Here, we combine extensive classical and QM/MM molecular dynamics simulations to map the full reactive trajectory of spliceostatin attachment to PHF5A. We identify two alternative noncovalent spliceostatin binding poses and show that only its thermodynamically favoured conformation pre-organizes the epoxide for nucleophilic attack. The local distortion of the ZnCys4 coordination sphere weakens the Zn-Cys26 bond, enabling water-Cys26 exchange and generating a highly reactive nucleophilic thiolate. Covalent bond formation is subsequently accelerated by an Asp34-Lys29 proton relay that activates the epoxide leaving group. Free-energy calculations confirm that the overall reaction is fast and strongly exergonic. Together, these findings provide a complete mechanistic framework for zinc-assisted inhibition of early spliceosomes by branch-site antagonists and advance our fundamental understanding of zinc finger reactivity.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.