With the benzamidine group as the invariant, FXa inhibitor 16 was then employed as the starting point to systematically introduce substituents into both the amidine-containing (proximal) and carboxamide-containing (distal) phenyl rings to optimize the binding interaction with the FXa active site. Three examples of proximal phenyl ring substitutions are highlighted in Table 2 (entries 17-19). While the majority of substitutions to either phenyl ring resulted in a complete loss of FXa inhibitory activity, incorporation of an amine (17) or hydroxyl moiety (18) into the 6-position of the proximal phenyl ring led to a maximal 70-fold increase in FXa inhibitory activity. 10 These data suggested that a hydrogen bond donating group at the 6-position of the proximal phenyl ring was critical for enhanced FXa inhibitory activity. This hypothesis was supported by the 6-methoxy analogue (19) which was nearly 500-fold less active than its 6-hydroxy counterpart (18).

Finally, a number of heterocyclic groups were introduced at the 3-position of the distal phenyl ring (R3′) to further refine the SAR at this site. As outlined in entries 20-24 of Table 2, this position can accommodate a variety of heterocyclic moieties with different steric, electronic, and conformational properties. Of this group, the functionalities which lead to the most potent inhibitors share a conformational bias toward a nonplanar orientation with the distal phenyl ring. Nonplanarity of these two groups appears to be more important to FXa inhibitory activity than the electronics or basicity of the 3-position substituent (16 vs 15, 21 vs 22, and