The wide scale usage of OWL for the formalization of real-world ontologies is currently influenced by important limitations which concern both its expressivity and the efficiency of OWL specific reasoning tools. While the expressivity limitations may be overcame by extending the OWL language (e.g. with rules), the reasoning with such heterogeneous knowledge bases is still an open issue. In this paper we propose OWL2Jess, a prototypical tool which enables the transformation of OWL ontologies to Jess rule bases and thus enables OWL models to be extended by means of rules. Facts are derived from an initial OWL file by one XSLT stylesheet, while the RDF(S) and OWL Semantics are pre-defined as Jess rules. By making hidden knowledge explicit, OWL2Jess achieves the knowledge compilation: the implicit subsumption and membership relations can be subsequently identified using the Jess rule engine.