Direction-of-arrival (DOA) estimation in the presence of mutual coupling and coherent signals is a hard task for arbitrary sensor arrays including uniform circular array (UCA). While the coherent sources can be resolved using spatial smoothing algorithms for uniform linear and rectangular arrays, it cannot be applied to UCA. In this paper, a new technique is proposed for DOA estimation in UCA using a single snapshot. Joint-sparse recovery algorithm is proposed where the source signal spatial directions and coupling coefficients are embedded into a joint-sparse signal. A dictionary is defined according to restricted isometry and compressed sensing is employed for both DOA and coupling coefficient estimation. It is shown that the proposed method performs better than the alternative sparse recovery techniques.