Various weighted ${{l}_{p}}$-norm inequalities in convolutions are derived by a simple and general principle whose ${{l}_{2}}$ version was obtained by using the theory of reproducing kernels. Applications to the Riemann zeta function and a difference equation are also considered.