Expression in a cell-free system of normal and variant forms of human antithrombin III. Ability to bind heparin and react with alpha-thrombin.

Human antithrombin III (AT-III) cDNA was cloned into the cell-free expression phagemid vector pGEM-3Zf(+) and site-directed mutagenesis was used to remove nucleotides encoding the signal peptide. AT-III messenger RNA (mRNA) transcripts derived from this construct were translated in an mRNA-dependent rabbit reticulocyte lysate (RRL) system containing (35S)methionine. Immunoprecipitation of the cell-free translation mixture with rabbit polyclonal antibodies to AT-III showed, by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), a 47-Kd polypeptide which is the non-glycosylated mature form of plasma AT-III. Densitometric scanning showed that this polypeptide constitutes greater than 90% of the radiolabeled polypeptides produced in this system. Heparin-Sepharose chromatography resulted in the elution of cell-free derived AT-III as a broad peak between 0.2 and 0.7 mol/L NaCl. The cell-free derived AT-III also reacted with human alpha-thrombin. In 2 minutes approximately 20% of the AT-III was found associated with a higher molecular weight species, consistent with the formation of a 1:1 stoichiometric covalent complex between alpha-thrombin and AT-III. Unfractionated heparin accelerated the rate of formation of such complexes. When Ser394 was mutated to Leu to form the AT-III Denver mutant, the cell-free translation product of this mutation did not show any significant complex formation when reacted with alpha-thrombin. A truncated form of AT-III (Met251-Lys432), containing only the putative thrombin-binding domain, was synthesized independently. This 21-Kd polypeptide did not bind heparin; however, it was cleaved by alpha-thrombin presumably at the reactive center Arg393-Ser394. When Ser394 was mutated to Leu the cell-free translation product of this truncated AT-III mutation did not react with alpha-thrombin at the reactive center. This simple cell-free approach, along with site-directed mutagenesis, should allow for the rapid and accurate mapping of the functional domains of human AT-III.