Possible mechanistic paths for self-inhibition of rubisco have been theoretically characterized by using analytical gradients with both AM1 semiempirical and HF/3-21G level calculations. Starting from the framework of an enediol moiety previously obtained from characterizations of the saddle point of index 1 (SPi-1) for the carboxylation and oxygenation reactions, the formation of xylulose, 3-ketoribitol, and 3-ketoarabinitol inhibitors is made possible by following specific intramolecular hydrogen rearrangements. The xylulose is attained from the SPi-1 describing intramolecular enolization via an intermediate made by protonation of the hydroxyl group Linked to the third carbon (C3) of the model substrate 3,4-dihydroxy-2-pentanone. One of these two hydrogens can migrate toward C3 with the correct stereochemistry to form xylulose inhibitor. 3-Ketoribitol and 3-ketoarabinitol inhibitors can be obtained after the enediol moiety is formed. Another minimum energy structure is found which is derived from the enediol via a SPi-1 corresponding to a retroenolization. This process finishes by forming a protonated hydroxyl group at C2. From this, and following the SPi-1, the 3-ketoribitol and the 3-ketoarabinitol inhibitors can be formed. The carbon and oxygen frameworks of the stationary geometries characterized in vacuo fit well at the active site of rubisco, except perhaps for xylulose inhibitor. This geometric overlap with an experimentally determined transition state analog suggests that the active site can accommodate the interconversion chemistry found with the present approach which leads to self-inhibition. This is compatible with the hypothesis that the loss of activity is due to the products of substrate isomerization formed during catalysis.