Homing endonucleases are remarkably specific enzymes. However, they can tolerate base-pair substitutions at certain positions within their recognition region (Fig. 2), allowing for the degeneracy of their homing sites without decreasing specificity. For instance, in the 31-bp recognition region of Pl-Scel, only 11 bases are essential for DNA cleavage activity as revealed by mutation studies on the DNA (Gimble and Wang 1996). This level of recognition is sufficient to ensure a single cut in the yeast genome (Bremer et al. 1992). I-Crel recognizes a pseudo-palindromic 22-bp substrate (Fig. 2) for which inverted repeats of the left or right half-sites are cleaved with similar activities. Estimations of substrate specificity indicate that I-Crel will recognize only one out of a billion random sequences of 22 bp (Argast et al. 1998).
Although free-standing homing endonucleases or the endonuclease domains of inteins are relatively small (200-250 amino acids), they can recognize long DNA substrates. The two P-sheets that form the saddle structure provide an extensive binding surface that can interact over long DNA distances. The pi and P2 strands in I-Crel (Jurica et al. 1998) and sheet 9 in Pl-Scel (Moure et al. 2002) interact with nine consecutive base pairs. The P-strands provide hydrogen-bond donors and acceptors that interact with bases edges in a sequence-specific manner and are critical to preserve the bending of the DNA into the actives sites. For instance, in Pl-Scel, distortions of the bound DNA in both the splicing and endonuclease domains occur at locations where base-specific interactions are made (Moure et al. 2002). The different lengths of the P-strands coupled to the low sequence homology of their primary sequences add flexibility to the protein-DNA interface. The undersaturation of the DNA-binding surface contributes to the flexibility of DNA interactions and might explain the recognition of degenerate substrates. In structural studies carried out on I-Crel and I-Msol, which recognize almost identical sequences, it was found that the divergence of the binding surface was greater than expected from the alignment of the primary sequences and that different sets of amino acids were involved in recognizing almost identical DNA substrates (Chevalier et al. 2003). The same principle may also apply to other homing endonucleases that recognize homologous substrates.
Was this article helpful?