A subsite model as proposed by Hiromi Hiromi, K. (1970) Biochem. Biophys. Res. Commun. 40, 1-6 has been applied to various hydrolases including glucoamylase (GA). The model assumes a single enzyme complex, a hydrolytic rate constant which is independent of substrate length, and a ratelimiting hydrolytic step. Recent kinetic studies with GA contradict these assumptions. Here we reevaluate the substrate binding of GA studying the pre-steady-state kinetics with glucose, which is reported here for the first time, and maltose. The association equilibrium constants for glucose and maltose interactions with wild-type and Trp120-->Phe GA from Aspergillus awamori in H2O and D2O buffers were obtained. Kinetic results indicate that a single glucose molecule binds to GA weakly by a single-step mechanism, E + G1<-->EG1, under the conditions studied. Similar fluorescence intensities of the GA-glucose and GA-maltose complexes, the high tryptophan concentration around subsite 1, crystal structures of various inhibitor complexes, pre-steady-state and steady-state modeling, feasibility of condensation reactions, and other evidence strongly suggest that glucose binds at subsite 1. These results conflict with the high subsite 2 and low subsite 1 affinities obtained using Hiromi's model. Using the substrate association constants for glucose and maltose obtained by pre-steady-state kinetics, the affinity of alpha-glucose for subsite 1 is shown to be substantially higher than the apparent affinity of glucose for subsite 2. We propose a GA catalytic mechanism whereby substrate binding is initiated by subsite 1 interactions with the nonreducing end of the oligosaccharide substrate, minimizing nonproductive substrate binding. Through conformational changes, entropic contributions, and increased local concentration, subsite 2 subsequently has enhanced affinity for the second covalently linked glucosyl residue.