Glutamine synthetase enzymes isolated from pea seeds and from Escherichia coli are observed to behave differently in experiments designed to probe reaction mechanism. Although both enzymes were found to bind and release substrates in random order mechanisms (Wedler, F.C. (1974) J. Biol. Chem, 247, 5080-5087), isotopic exchanges with partial reaction systems indicative of a gamma-glutamylphosphate intermediate are catalyzed only by the pea seed enzyme. The E. coli system fails to catalyze any exchanges at appreciable rates unless all substrates are present. This negative result implies either an absolute conformational requirement for bound substrates or that the putative complex (E-Glu-P-MgADP) is exceedingly tight. To test the latter, a nonreactive structural analog of gamma-glutamyl-phosphate, namely 3-(phosphonoacetylamido)-L-alanine (PA2LA), has been synthesized. With the E. coli enzyme PA2LA was found to bind no more tightly than L-glutamate and is strictly competitive versus L-glutamate (Ki = 3 mM). Thus, failure to catalyze partial exchange reactions indicative of gamma-Glu-P is probably not attributable to tight complex formation. The binding of PA2LA with the pea seed enzyme apparently involves a two-step process: a rapid, reversible step in which PA2LA binds 10-fold more tightly than L-glutamate, followed by a slow (but reversible) process involving very tight PA2LA binding, apparently with enzyme isomerization promoted by nucleotide. The specificity of the two enzymes toward L-methionine-SR-sulfoximine, Met(O)(NH), was also different...