There is increasing evidence that sensitization of the androgen receptor (AR) signaling pathway contributes to the failure of androgen ablation therapy for prostate cancer, and that direct targeting of the AR may be a useful therapeutic approach. To better understand how AR function could be abrogated in prostate cancer cells, we have developed a series of putative dominant-negative variants of the human AR, containing deletions or mutations in activation functions AF-1, AF-5, and/or AF-2. One construct, AR inhibitor (ARi)-410, containing a deletion of AF-1 and part of AF-5 of the AR, had no intrinsic transactivation activity but inhibited wildtype AR (wtAR) in a ligand-dependent manner by at least 95% when transfected at a 4:1 molar ratio. ARi-410 was an equally potent inhibitor of gain-offunction AR variants. Ectopic expression of ARi-410 inhibited the proliferation of AR-positive LNCaP cells, but not AR-negative PC-3 cells. Whereas ARi-410 also marginally inhibited progesterone receptor activity, this was far less pronounced than the effect on AR (50% vs. 95% maximal inhibition, respectively), and there was no inhibition of either vitamin D or estrogen receptor activity. In the presence of ligand, ARi-410 interacted with wtAR, and both receptors translocated into the nucleus. Whereas the amino-carboxy terminal interaction was not necessary for optimal dominant-negative activity, disruption of dimerization through the ligand binding domain reduced the efficacy of ARi-410. In addition, although inhibition of AR function by ARi-410 was not dependent on DNA binding, the DNA binding domain was required for dominant-negative activity. Taken together, our results suggest that interaction between ARi-410 and the endogenous AR in prostate cancer cells, potentially through the DNA binding and ligand binding domains, results in a functionally significant reduction in AR signaling and AR-dependent cell growth.