In response to DNA damage, a set of unlinked E. coli genes are expressed (Kenyon et al., 1982; Little, 1983; for a review see Little and Mount, 1982). The genes of the SOS regulatory system are controlled in part at the level of transcription by the direct binding of the lexA gene product to the promoters. Five binding sites are well characterized. Two sites are linked to LexA, one is linked to each of recA (Little et al., 1981; Brent and Ptashne, 1981; Uhlin et al., 1982), uvrA (the same site as for ssb) (Sancar et al., 1982a; Brandsma et al., 1983; Backendorf et al., 1983), and uvrB (Sancar et al., 1982b). Two others have been reasonably well identified: at sulA (=sfiA) (Cole, 1983) and on the plasmid cloDF13 (van den Elzen et al., 1982). Several plasmid promoters may have two deeply overlapping LexA sites (Ebina et al., 1981; van den Elzen et al., 1982; Morlon et al., 1983). Since it is possible that one of these is not functional, which would confuse the analysis, we did not use these sites. Since there are two adjacent sites upstream from the lexA gene, the range was limited to 20 bases. This is approximately the region protected by LexA protein from digestion by DNaseI (Little et al., 1981; Brent and Ptashne, 1981). For both the Rsequence and Rfrequency calculations, we assumed that LexA repressor binds to its operators symmetrically (Little and Mount, 1982), and that the center of the symmetry is between bases 0 and 1 (Fig. 3). For the 14 example sequences, Rsequence = 21.1 bits per site. The nucleotide composition used for this and all remaining recognizers was from E. coli chromosomal DNA (LIB2): A = T = 21260, C = G = 21644 (Stormo et al., 1982a). Hg = 1.99994 bits/base.
The damage-inducible (din) genes are spread around the E. coli genome (Little and Mount, 1982), so the size of E. coli DNA determines G. There are at least 11 chromosomal genes under lexA control (Little and Mount, 1982), giving a minimum estimate for the number of sites , and an upper bound on Rfrequency of 18.4.