The later steps of the proposed mechanistic pathway for the reaction catalyzed by dehydroquinate synthase have been probed by using three substrate analogues. Each of these analogues is structurally prohibited from undergoing the ring-opening reaction that necessarily precedes the carbon-carbon bond-forming step in the overall conversion of the substrate 3-deoxy-D-arabino-heptulosonate 7-phosphate (1) to dehydroquinate (2). Two of the analogues (the 2-deoxy cyclic compound 3 and the carbacyclic material 4) are locked into a cyclic form, mimicking the pyranose form of the substrate DAHP. The third analogue, 5, contains no carbonyl group at C-2 and may thus resemble the open-chain form of DAHP. Analogues 3 and 4 each bind to the enzyme and are competitive inhibitors having Ki values of 35 and 0.12 microM, respectively. More importantly, however, incubation of these analogues with the enzyme leads to the catalytic production of Pi along with the corresponding exomethylene compounds that are analogous to the enol ether IV postulated for the normal synthase reaction. In contrast to these results, the acyclic analogue 5 is neither a substrate nor an inhibitor of the enzyme. These data suggest that the enzyme recognizes and acts upon the alpha-pyranose form of the natural substrate. The ready release of the exomethylene products from the processing of analogues 3 and 4 is consistent with the suggestion of Bartlett and his group that the enzyme may release the enol ether intermediate IV into solution, where the ring opening and cyclization occur nonenzymically. The use of 3 stereospecifically labeled with deuterium at C-7 allows the sterochemical course of the beta-elimination of phosphate to be established. This step proceeds with syn stereochemistry, which fits the pattern of enzyme-catalyzed elimination from substrates where the proton is lost from a position alpha to a ketone, an aldehyde, or a thiolester. Since the overall stereochemical course of the transformation mediated by dehydroquinate synthase had been shown to be inversion, the present finding of a syn elimination suggests that the transition state for the subsequent intramolecular aldol reaction has a chairlike geometry.