This paper presents an efficient beamforming technique for secondary cognitive radio network deployed in the coverage region of a cellular network. In cognitive network, it must be ensured that secondary users (SUs) do not cause any harmful interference to the primary users (PUs). To keep this in mind, we have formulated a new beamforming problem for the secondary network considering some new constraints in addition to the only constraint of the conventional Minimum Variance beamformer (MVB), to maintain minimum array gain towards the direction of the PUs and provide a solution by using Lagrange multiplier method and complex-matrix calculus. Our proposed solution for the optimal array weight vector has been able to put deeper nulls towards the direction of the PUs which has been shown in the simulation result.