Steady-state and time-resolved emission spectroscopy were used to study the interaction of Escherichia coli purine nucleoside phosphorylase (PNP) with its specific inhibitors, viz. formycin B (FB), and formycin A (FA) and its N-methylated analogues, N(1)-methylformycin A (m(1)FA), N(2)-methylformycin A (m(2)FA) and N(6)-methylformycin A (m(6)FA), in the absence and presence of phosphate (P(i)). Complex formation led to marked quenching of enzyme tyrosine intrinsic fluorescence, with concomitant increases in fluorescence of FA and m(6)FA, independently of the presence of P(i). Fluorescence of m(1)FA in the complex increased only in the presence of P(i), while the weak fluorescence of FB appeared unaffected, independently of P(i). Analysis of the emission, excitation and absorption spectra of enzyme-ligand mixtures pointed to fluorescence resonance energy transfer (FRET) from protein tyrosine residue(s) to FA and m(6)FA base moieties, as a major mechanism of protein fluorescence quenching. With the non-inhibitor m(2)FA, fluorescence emission and excitation spectra were purely additive. Effects of enzyme-FA, or enzyme-m(6)FA, interactions on nucleoside excitation and emission spectra revealed shifts in tautomeric equilibria of the bound ligands. With FA, which exists predominantly as the N(1)-H tautomer in solution, the proton N(1)-H is shifted to N(2), independently of the presence of P(i). Complex formation with m(6)FA in the absence of P(i) led to a shift of the amino-imino equilibrium in favor of the imino species, and increased fluorescence at 350 nm; by contrast, in the presence of P(i), the equilibrium was shifted in favor of the amino species, accompanied by higher fluorescence at 430 nm, and a higher affinity for the enzyme, with a dissociation constant K(d)=0.5+/-0.1 microM, two orders of magnitude lower than that for m(6)FA in the absence of P(i) (K(d)=46+/-5 microM). The latter was confirmed by analysis of quenching of enzyme fluorescence according to a modified Stern-Volmer model. Fractional accessibility values (f(a)) varied from 0.31 for m(1)FA to 0.70 for FA, with negative cooperative binding of m(1)FA and FB, and non-cooperative binding of FA and m(6)FA. For all nucleoside ligands, the best model describing binding stoichiometry was one ligand per native enzyme hexamer. Fluorescence decays of PNP, FA and their mixtures were best fitted to a sum of two exponential terms, with average lifetimes (<tau>) affected by their interactions. Complex formation resulted in a 2-fold increase in <tau> of FA, and a 2-fold decrease in <tau> of enzyme fluorescence. The amplitude of the long-lifetime component also increased, confirming the shift of the tautomeric equilibrium in favor of the N(2)-H species. The findings have been examined in relation to enzyme-nucleoside binding deduced from structural studies.