Molecular modeling of the catalytic domain of serine/threonine phosphatase-1 with the Zn2+ and Mn2+ di-nuclear ion centers in the active site.

Catalytic domains of the metalloenzymes protein phosphatases (PPP) 1, 2A and 2B (PP1, PP2A and PP2B, respectively) are homologous to approximately 45%, with the residues in the enzymatic centers strictly conserved. PP1, PP2A and PP2B are abundant in cells and they dephosphorylate serine and/or threonine residues in a variety of proteins serving as cellular phospho switches. The active enzymes work as invariant catalytic subunits PP1c, PP2Ac and PP2Bc, respectively, complexed with diverse regulatory subunits, dependent on the enzymes' specific location and biological function. The crystal structures of PP1c and PP2B (calcineurin) heterotetramer calcineurinA x calcineurinB x FKBP x FK506 have been determined. A comparison of the catalytic subunits of both enzymes indicates their significant structural homology and virtual identity within the catalytic centers, each including a set of conservative amino acids, two metal ions and a phosphate; thus confirming a hypothesis on their common enzymatic mechanisms. The elongated substrate cleft at the active centre is kinked by approximately 120 degrees at the active center in its middle and thus divided into a pre-phospho-Ser/Thr (ligand N-terminal) and a post-phospho-Ser/Thr (ligand C-terminal) section. In PP1c the N-terminal section is highly acidic while in PP2Bc is not. This feature is likely pertinent but not sufficient to the enzymes' selectivity, which is also controlled by regulatory subunits, diverse in various tissues. The metalloenzymes in general and PPP in particular are hard to deal with using theoretical simulations due to parameterization problems for the metal cations. In fact, there are only a few PP1c simulations reported, with the metal di-cations treated quite crudely. This is a preliminary work, in which we introduce and test against some experimental evidence a concept of pseudomolecules of proper geometry, composed of double metal (2Zn2+ or 2Mn2+) cation, and the OH- nuclephile incorporated into the PP1c catalytic site. Both models are associated with either the phosphate (a free enzyme) or the phosphorylated dodecapeptide RRRRPpTPAMLFR, an active fragment (residues 29-40) of a regulatory subunit DARPP-32 inhibitor (PP1c-inhibitor complex); four models total. We have parameterized both pseudomolecules within the AMBER force field. Subsequently, using molecular dynamic in water, we have found the free PP1c subunits to be less stable than the complexed ones and we have speculated on possible reasons for this feature.