Remote mutations alter transition-state structure of human purine nucleoside phosphorylase

Purine nucleoside phosphorylase (PNP) catalyzes the reversible phosphorolysis of purine (2′-deoxy)ribonucleosides to give the corresponding purine base and (2′-deoxy)ribose 1-phosphate as products. Human and bovine PNPs (HsPNP and BtPNP) form distinct transition states despite 87% identity in amino acid sequence. A PNP hybrid was produced by replacing K22 and H104 in HsPNP with the corresponding Glu and Arg residues found in BtPNP. We solved the transition-state structure of E:R-HsPNP (K22E:H104R mutant of HsPNP) using competitive kinetic isotope effects (KIE) and global density functional calculations. An array of PNP transition states was generated from optimized structure candidates with varied C1′-N9, C1′-O^phosphate distances, ribosyl pucker configurations and N7-protonation states. Isotopically labeled [1′-³H], [2′-³H], [1′- ¹⁴C], [9-¹⁵N], [1′-¹⁴C, 9-¹⁵N] and [5′-³H₂]inosines gave intrinsic KIE values of 1.210, 1.075, 1.035, 1.024, 1.065, 1.063 with E:R-HsPNP, respectively. The suite of E:R-HsPNP KIEs match a single structure from the array of PNP transition-state candidates. The transition state of E:R-HsPNP is fully dissociative, N7-protonated hypoxanthine (C1′-N9 distance ≥ 3.0 Å) with partial participation of phosphate (C1′-O ^phosphate distance = 2.26 Å), 2′-C-exo-ribosyl ring pucker and the O5′-C5′-C4′-O4′ dihedral angle near 60°. The transition state of E:R-HsPNP is altered from the fully dissociative D_N*A_N character for HsPNP to a late phosphate-associative character. E:R-HsPNP differs from native HsPNP by only two residues over 25 Å away from the active site. New interactions caused by the mutations increase the catalytic efficiency of the enzyme for formation of a late transition state with increased participation of the phosphate nucleophile. Dynamic coupling motions from the remote mutations to the catalytic sites are proposed.