We can derive from the Poisson summation formula the modularity of the Theta function, which results in the functional equation. In his book on the Riemann Zeta function, Patterson mentions also that there is a way back. How can this be done?

Equivalently, how can the modularity of the Theta function (which is Poisson summation for a one parameter family of functions only) be used to derive the Poisson summation formula for general test functions?