Petri Nets are a formalism for describing coordinated interactions between independent entities, called transitions, by means of synchronized token exchange through places. Tuplespaces are an abstraction of the shared memory concept providing means for loosely-coupled coordination between individual components. In the tuplespace model, coordination between applications is achieved by production and blocking consumption of tuples which are communicated over tuplespace buffers. On the basis of the conceptual similarities of Petri nets and tuplespaces, we investigate in this paper whether tuplespaces are a suitable platform for execution of applications modeled as Petri net-based workflows, motivate an extension to the original tuplespace interface to facilitate efficient synchronization of control flow, and highlight its benefits by providing quantitative and qualitative evaluations of an application using the traditional tuplespace interface versus an application using our extension.