Ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCO) of a thermophilic cyanobacterium, Thermosynechococcus vulcanus, was cloned and expressed in Escherichia coli. The purified enzyme had higher thermostability than RuBisCOs isolated from mesophilic cyanobacteria. Prediction of the tertiary structure was performed using the software Molecular Operating Environment (MOE). The predicted structure did not give any clue about the basis of thermostability. Then, the molecular docking of substrates and inhibitors in the catalytic site were carried out to test analogs for consistency of ribulose 1,5-bisphosphate, a RuBisCO substrate. The analogs were searched in the Kyoto Encyclopedia of Genes and Genomes (KEGG), and 99 compounds were selected for the docking. The mol files from LIGAND Database in KEGG were changed to a three dimensional (3D) structure for use in docking simulation. The docking simulation was performed on ASEDock of MOE, and the SiteFinder command suggested about 20 candidates for the docking site of the compounds. Based on the homology of these candidate sites with the xylulose 1,5-bisphosphate (XBP)-binding site of RuBisCO isolated from Synechococcus PCC 6301, one site was selected for the docking simulation. The 40 compounds with the highest docking energies included synthetic organic substances that had never been demonstrated to be inhibitors of RuBisCO. The total docking energies were -102 kcal/mol, -104 kcat/mol, -94.0 kcal/mol, and -57.7 kcal/mol for ribulose 1,5-bisphosphate (RuBP), etidronate, risedronate, and citrate respectively. Kinetic analysis of RuBisCO revealed a K-m value of 315 mu M for RuBP, and K-i values of 1.70, 0.93, and 2.04 mM for etidronate, risedronate, and citrate respectively. From these values, the binding energies were estimated to be -4.85, -3.84, -4.20, and -3.73 kcal/mol for RuBP, etidronate, risedronate, and citrate respectively. The differences between the values estimated from experimental data and by simulation may mainly depend on the dissimilarity of the environment for the protein and ligands between the experiments and the simulation. The results obtained here suggested a few new inhibitors, which might be useful as tools for studying the relationship between the structure and the function of RuBisCO.