The papillomavirus E2 proteins are transcriptional regulators that bind to a consensus DNA sequence ACCG NNNN CGGT. Multiple copies of this binding site are found in the viral genomes. The affinities of the naturally occurring binding sites for the E2 proteins are predominantly dependent upon the sequence of the NNNN spacer. The hierarchies of binding site affinities among the sites present in the viral genomes result in differential occupancy during the viral life-cycle. In turn, this differential binding regulates transcription from viral promoters, including those for the oncogenes E6 and E7. Structural and biochemical studies have shown that E2 proteins bend the DNA to which they specifically bind. Atomic resolution structures of complexes of the bovine papillomavirus strain 1 (BPV-1) E2 protein and DNA show that the protein does not contact the spacer DNA. A direct comparison of the binding of the DNA-binding domains of the E2 proteins from BPV-1 and human papillomavirus strain 16 (HPV-16) to a series of binding sites as a function of the sequence of their central spacer and/or the presence of a nick or gap in one strand of the spacer DNA is presented in this paper. The BPV-1 E2 DNA-binding domain is only moderately sensitive to the nature of the central spacer; less than several fold differences in affinity were observed when the DNA sequence of the spacer was varied and/or a nick or gap was introduced. In contrast, the HPV-16 E2 DNA-binding domain binds to sites containing A:T-rich central spacers with significantly increased affinity. The introduction of a nick or gap into the spacer of these high affinity sequences is very detrimental to HPV-16 E2 binding while comparable nicks or gaps have only small effects in the low affinity sequences. These results suggest that the HPV-16 E2 protein recognizes the structure of the DNA spacer and that the mechanism of DNA-sequence specific binding of the homologous HPV-16 E2 and BPV-1 E2 proteins is significantly different.