Normal and mutant forms of human antithrombin-III (AT-III) were synthesized in a cell-free system in order to identify putative functional domains required for heparin binding and complex-formation with alpha-thrombin. Heparin-Sepharose chromatography resulted in the elution of approx. 70% of cell-free-derived normal AT-III-(1-432)-polypeptide as a peak between 0.2 M- and 0.7 M-NaCl. The cell-free-derived normal AT-III also reacted with alpha-thrombin. Approx. 15% of this AT-III formed covalent complexes with alpha-thrombin in 2 min. Unfractionated heparin accelerated the rate of formation of such complexes. Two truncated forms of AT-III (amino acid residues 219-432 and 251-432), containing only the putative thrombin-binding domain, were synthesized independently in this cell-free system. These truncated AT-III polypeptides did not bind heparin and were unable to form stable covalent complexes with alpha-thrombin. However, both of these AT-III polypeptides were cleaved by alpha-thrombin, presumably at the reactive centre Arg-393-Ser-394. The formation of the disulphide bond between Cys-247 and Cys-430 in AT-III-(219-432)-polypeptide had no effect on the results obtained. Mutations in full-length AT-III at Cys-430 had no effect on the ability of AT-III to bind heparin. There was, however, a slight decrease in the formation of stable inhibitory complexes with alpha-thrombin. A cell-free-derived AT-III mutant, devoid of amino acid residues 41-49, which comprise heparin-binding region 1 of AT-III, had slightly decreased heparin binding compared with cell-free-derived normal AT-III-(1-432)-polypeptide. This mutant AT-III polypeptide was unable, however, to form a stable complex with alpha-thrombin. We conclude therefore that the N-terminal domain of AT-III is essential for both heparin binding and complex-formation with alpha-thrombin, but not for the cleavage of AT-III at its reactive centre by alpha-thrombin.