How does trp repressor bind to its operator?

Three explanations have been advanced to account for the unexpected absence of direct hydrogen bonds and presence of a buried water layer in the co-crystal-line complex of Escherichia coli trp repressor with DNA. We present results of physical and biochemical measurements that address the testable predictions of each model. We find that the DNA oligomer used for co-crystallization binds to the repressor with high affinity and specificity, and 1:1 stoichiometry, consistent with other evidence that this sequence represents the true operator target for a single repressor dimer. A proposed alternative DNA sequence binds weaker and with higher stoichiometry, consistent with a cooperative binding mode. The operator DNA in solution has a B-form helical structure in the presence and absence of repressor. Affinity of repressor for operator is altered under the conditions used for cocrystal growth.