Tissue factor (TF) is the protein cofactor for factor VIIa (FVIIa), the first serine protease of the clotting cascade. Previous studies using alanine mutagenesis have identified TF residues Lys(165) and Lys(166) as important for factor X (FX) activation, hypothesizing either that these residues interact with phospholipid head groups or that they directly or indirectly promote macromolecular substrate binding. In the recently reported x-ray crystal structure of the isolated extracellular domain of TF, both Lys(165) and Lys(166) are solvent-exposed and predicted to be near the phospholipid surface in intact TF. We hypothesized that these residues may in fact be ideally positioned to interact with the 4-carboxy-glutamate-rich domain (Gla domain) of FX. We therefore predicted that mutations at Lys(165) and Lys(166) should have no effect on the activation of Gla domainless FX. To test this hypothesis, we mutated both residues Lys(165) and Lys(166) of TF to Ala, Glu, or Gln and examined the ability of these double mutants to support FVIIa-mediated activation of FX, Gla domainless FX, and factor IX (FM). Each TF mutant was equivalent to wild-type TF in both FVIIa binding and promotion of FVIIa amidolytic activity, However, all three mutants were markedly deficient in supporting FIX and FX activation, with FX activation rates decreased more than FIX activation rates. In both reactions, the TF mutants exhibited different extents of activity: Gln(165)-Gln(166) > Ala(165)-Ala(166) > Glu(165)-Glu(166). In sharp contrast, all three TF mutants were equivalent to wild-type TF in supporting activation of Gla domainless FX by FVIIa. Interestingly, the deficiency of the mutants in FX activation was less pronounced when Gla domainless FVIIa was used in place of native FVIIa, Together, these findings suggest that TF residues Lys(165) and Lys(166) contribute to a binding site for the Gla domain of FX (and perhaps other substrates) and that this interaction may be facilitated by the presence of the Gla domain of FVIIa.