Based on an extensive collection of sigma 70 associated regulatory mechanisms, a grammatical model has been constructed that define the functional positions and combinations of sites within DNA regulatory regions. The syntactic rules and the dictionary implemented in a Prolog program were coupled to consensus matrices used as "sensors" to integrate a syntactic recognizer. A systematic comparison between the syntactic recognizer and the standard weight matrix methodology is presented using 12 regulatory proteins and the whole collection of about 130 sigma 70 DNA regulatory regions. On the average an increased sensitivity of 5 to 10 fold is obtained with this novel approach.