An effect of DNA sequence on nucleosome occupancy and removal

Abstract

A barrier phases nucleosomes at the yeast (Saccharomyces cerevisiae) GAL1-GAL10 genes. Here we separate nucleosome positioning from occupancy and show that the degree of occupancy of these phased sites is predictably determined by the underlying DNA sequences. As this occupancy is increased (by sequence alteration), nucleosome removal upon induction is decreased, as is mRNA production. These results explain why promoter sequences have evolved to form nucleosomes relatively inefficiently.

Figure 1. Chromatin architecture at the GAL1/10 locus prior to and following induction

(a) The wild type locus. The distribution of nucleosomes prior to induction, as well as the fractional occupancy of each implied site, is shown in blue. The corresponding values following induction are shown in yellow. The data was obtained using the assay of Bryant et al.. The UASg bears four Gal4 binding sites shown in cyan, and the RSC/nucleosome complex is indicated by a cyan oval. Nucleosomes flanking the UASg are shown as green ovals. The shaded areas −1 and −2 indicate the sites of positioned (phased) nucleosomes in the GAL1 promoter. The 5′ ends of the GAL1 and GAL10 genes are shown as horizontal black bars, and the GAL1 TATA box by the small vertical blue bar that lies between sites −1 and −2. The distribution shown in blue is unchanged by addition or deletion of the activator Gal4, but nucleosome removal as shown in yellow requires Gal4 and the inducer galactose (2% in this case, added to cells growing in raffinose). The envelopes encompassing the curves indicate the range of experimental error, calculated as described in Bryant et al.. (b) As in (a) except that the “superbinder” sequence of Supplementary Table 1 has been substituted for the wild type sequence at site −1 as indicated by the horizontal magenta bar. (c) As in (a) except that the superbinder sequence has been substituted for the wild type sequence at site −2.

Figure 2. The effects of increasing nucleosome-forming propensities on nucleosome occupancies at site −1

(a) The bottom line indicates the positions of GC and AT/TT/TA dinucleotide elements in a 133 bp sequence designed to form nucleosomes with high efficiency. Each successive TA element is separated by ten base pairs, as is each successive GC element. At the top is the array of these sequence elements found in the wild type (WT) sequence at position −1 in Figure 1. Each successive sequence (starting at the top) was modified by sequential substitutions of 20 bp, resulting in the distribution of TA and GC elements as indicated. The predicted nucleosome-forming propensities of these sequences increase from top to bottom (see text). (b) Nucleosome occupancies were determined, prior to (blue) and some 30 min following (yellow) induction, of wild type, superbinder and five other sequences (M1–M5) substituted at site −1 (see (a)).

Figure 3. Nucleosome removal and mRNA production

(a) Cells bearing the substitution mutant of Figure 1c, (which bears the superbinder at site −2), and wild type (WT) cells, growing in 2% glucose, were transferred to medium containing 0.1% glucose and 2% galactose. Nucleosome occupancy at the WT site −2 (yellow) and the superbinder-substituted site −2 (blue) was assayed at the indicated time points following induction. (b) Aliquots of cells used for the experiment of (a) were assayed for GAL1 mRNA using QPCR as described.