We study in this paper optimal signaling for downlink transmission with mutual interference. Restricting each receiver to a single user detector, we show that computing the achievable rate region amounts to solving a family of nonconvex optimization problems. Recognizing the intrinsic connection between the signal power at the intended receiver and the interference power at the unintended receiver, we convert the original family of non-convex optimization problems into a new family of convex optimization. Closed-form solutions to this new family of optimization problems are derived. Interestingly, with multiple-antenna transmitters and single-antenna receivers, the optimal downlink transmission with interference always reduces to a simple beamforming scheme. Finally, we note that similar procedure can be devised for systems with more than two pairs of mutually interfering users.