Elevated activin A levels in inhibin-deficient mice promote the development of gonadal tumors and induce cachexia by reducing muscle, liver, stomach, and fat mass. Because activin A is an important regulator of tissue growth, inhibiting the actions of this TGFβ family ligand may halt or reverse pathology in diseased tissues. In this study, we modified the activin A propeptide to generate a specific activin antagonist. Propeptides mediate the synthesis and secretion of all TGFβ ligands and, for some family members (e.g. TGFβ1), bind the mature growth factor with high enough affinity to confer latency. By linking the C-terminal region of the TGFβ1 propeptide to the N-terminal region of the activin A propeptide, we generated a chimeric molecule [activin/TGFβ1 propeptide (AT propeptide)] with increased affinity for activin A. The AT propeptide was 30-fold more potent than the activin A propeptide at suppressing activin-induced FSH release by LβT2 pituitary gonadotrope cells. Binding of the AT propeptide to activin A shields the type II receptor binding site, thereby reducing Smad2 phosphorylation and downstream signaling. In comparison with the commonly used activin antagonists, follistatin (IC(50) 0.42 nM), soluble activin type II receptor A-Fc (IC(50) 0.47 nM), and soluble activin type II receptor B-Fc (IC(50) 0.91 nM), the AT propeptide (IC(50) 2.6 nM) was slightly less potent. However, it was more specific, inhibiting activin A and activin B (IC(50) 10.26 nM) but not the closely related ligands, myostatin and growth differentiation factor-11. As such, the AT propeptide represents the first specific activin antagonist, and it should be an effective reagent for blocking activin actions in vivo.