Channel allocation problem is a major challenge in wireless local area networks (WLANs), especially in dense deployments of access points (APs) where congestion of the unlicensed spectrum bands (i.e., ISM bands) could undermine achieved network performance. This paper analyses the possibility to alleviate congestion of the ISM band by allowing some APs to use additional channels located in licensed bands in an opportunistic manner whenever licensee services (i.e., primary users) are not affected. Availability of these additional channels in licensed bands is assumed not to be the same for all the APs. Based on this assumption, we formulate the problem for the channel assignment as a Binary Linear Programming (BLP) problem, which allows us to obtain an optimal solution despite an elevated execution time. We also develop a heuristic method based on building a Minimum Spanning Tree (MST) graph attending to interference conditions that is able to find nearoptimal solutions with a shorter execution time. Results are provided to assess the benefits of such a proposal under different WLAN deployment situations and primary channel availability conditions.