To create a molecule that would block the purine binding site of PNP, investigators made two changes in guanine, a purine. They expected that compounds carrying the resulting guanine derivative would bind tightly to PNP. But the strategy failed. Analyses of the chemical bonds in the complexes formed by the union of PNP with that and related guanine derivatives explained why. Guanine itself (a) forms two hydrogen bonds (purple bands) with the amino acid asparagine residing in position 243 of PNP. ("R" is rest of the compound.) Making one of the proposed changes in guanine -- replacing nitrogen in position nine (blue sphere) with carbon (green), thus generating 9-deazaguanine (b) -- enhanced affinity for PNP by causing a strong hydrogen bond (red) to replace the weaker one. Separately making the other proposed change in guanine (c) -- substituting an amino group (orange) for the hydrogen bound to carbon in position eightÑalso improved guanine's affinity for PNP somewhat. It did so by causing an extra hydrogen bond to appear (between the amino group and threonine 242). Making both changes at once (d) yielded a poor inhibitor because the dual alteration led to a repulsive clash between threonine 242 and the amino group (red arrow). The exercise thus revealed that 9-deazaguanine would actually bind best.