Poly(dA-dT) promoter elements increase the equilibrium accessibility of nucleosomal DNA target sites

Abstract

Polypurine tracts are important elements of eukaryotic promoters. They are believed to somehow destabilize chromatin, but the mechanism of their action is not known. We show that incorporating an A(16) element at an end of the nucleosomal DNA and further inward destabilizes histone-DNA interactions by 0.1 +/- 0.03 and 0.35 +/- 0.04 kcal mol(-1), respectively, and is accompanied by 1.5- +/- 0.1-fold and 1.7- +/- 0.1-fold increases in position-averaged equilibrium accessibility of nucleosomal DNA target sites. These effects are comparable in magnitude to effects of A(16) elements that correlate with transcription in vivo, suggesting that our system may capture most of their physiological role. These results point to two distinct but interrelated models for the mechanism of action of polypurine tract promoter elements in vivo. Given a nucleosome positioned over a promoter region, the presence of a polypurine tract in that nucleosome's DNA decreases the stability of the DNA wrapping, increasing the equilibrium accessibility of other DNA target sites buried inside that nucleosome. Alternatively (if nucleosomes are freely mobile), the presence of a polypurine tract provides a free energy bias for the nucleosome to move to alternative locations, thereby changing the equilibrium accessibilities of other nearby DNA target sites.

FIG. 1

DNA constructs used. (A) Construct 601.3. The boundaries of the nucleosomal DNA as mapped on our earlier study of the slightly longer construct 601.2 are indicated by the black vertical bars; relative locations of specific restriction enzyme recognition sites are shown. The other sequences in the 601.3 series (B and C) incorporate various alterations from 601.3 that are represented as shaded boxes. (B) 601.3(A₁₆End) and 601.3(Random End). The shaded box extends from bp 1 to 16 and represents poly(dA-dT) DNA [601.3(A₁₆End)] or random pGEM3z DNA [601.3(Random End)]. (C) 601.3(A₁₆Mid) and 601.3(Random Mid) DNA. The shaded box extends from bp 37 to 52 and represents poly(dA-dT) DNA [601.3(A₁₆Mid)] or random pGEM3z DNA [601.3(Random Mid)].

FIG. 2

Sucrose gradient purification and reanalysis by sucrose gradient and native gel electrophoresis. Nucleosomes are reconstituted by gradual salt dialysis and separated from naked DNA on 5 to 30% (wt/vol) sucrose gradients (A). Purified nucleosomes are further analyzed on a second sucrose gradient (A) and by native gel electrophoresis (B) +, naked 601.3 DNA; □, preparative run of reconstituted 601.3 nucleosomes; ○, preparative run of reconstituted 601.3(A₁₆End) nucleosomes; ◊, preparative run of reconstituted 601.3(A₁₆Mid) nucleosomes; ×, reanalysis of gradient purified 601.3 nucleosomes. (B) Native gel analysis. W indicates the location of the loading wells; R indicates the mobility of the reconstituted nucleosomes; D indicates the mobility of naked DNA. Lane M, 100-bp DNA marker; lane 1, naked 601.3 DNA; lane 2, purified 601.3 nucleosomes; lane 3, purified 601.3(A₁₆End) nucleosomes; lane 4, purified 601.3(A₁₆Mid); lane 5, purified 601.3(Random End) nucleosomes; lane 6, purified 601.3(Random Mid) nucleosomes. Phosphorimager analysis of the gel reveals contamination by free DNA and other nonnucleosomal aggregates to be ≤0.5%. This small level of naked DNA does not contribute to the observed kinetics because it is digested to completion within the first time point, which we omit from the kinetic analysis.

FIG. 3

Native gel analysis of competitive reconstitution assays. Radiolabeled tracer competes with a large excess of unlabeled natural nucleosome core particle DNA for limiting quantities of histone octamer in dialysis from concentrated NaCl. Lane M, 100-bp DNA marker; lane 1, 601.3; lane 2, 601.3(A₁₆End); lane 3, 601.3(A₁₆Mid); lane 4, 601.3(Random End); lane 5, 601.3(Random Mid); lane 6, the 256-bp EcoRI fragment of the well-characterized natural nucleosome-positioning sequence from sea urchin 5S rRNA gene (34); lane 7, 601.2; D, mobility of naked DNA; R, mobilities of reconstituted nucleosomes. The diverse mobilities represent reflect a range of positionings on the different molecules. The 5S derivative yields several distinct nucleosomal positions, whereas 601.2 and the 601.3 series show one predominant position. The raw data show that 601.2 and the 601.3 series compete much more effectively for the limiting histone octamer than does the 5S sequence (greater ratio of counts in band R versus counts in band D). The contrast in lane 6 was increased to a larger degree than the rest of the gel due to the presence of fewer counts in that lane (see Materials and Methods). Note that this lane is included only for comparison with other studies; the use of this (or any) reference molecule does not influence the ΔΔG°s obtained for comparison between differing DNA samples measured in the same competitive environment.

FIG. 4

Representative kinetic analysis, probing site exposure at the HhaI site, 76 to 79 bp pairs from the 5′ end of the predominant core particle position. (A to D) Denaturing polyacrylamide gel analysis of the time course of digestion. Lanes 1 through 7 in all digestion gels are samples removed at 0, 0.5, 1, 2, 5, 10, and 15 min from reaction initiation. In each case, the substrate (S; 152 nucleotides [nt] for all 601.3 constructs) is converted over time to two products (82 nt [P1] and 72 nt [P2] for all 601.3 constructs). The sizes of S, P1, and P2 expected from the DNA sequence are confirmed against the 100-bp DNA markers in lane M. (A) Naked 601.3 DNA, digested with HhaI at 0.1 U ml⁻¹; (B) 601.3 nucleosomes, digested with HhaI at 2,000 U ml⁻¹; (C) 601.3(A₁₆End) nucleosomes, digested with HhaI at 2,000 U ml⁻¹; (D) 601.3(A₁₆Mid) nucleosomes, digested with HhaI at 2,000 U ml⁻¹. (E to H) Quantitative analyses of the time course of digestion from the data in panels A to D, respectively. The fraction of DNA remaining uncut is corrected for a small initial extent of nucleosome dissociation (which did not correlate with the presence or absence of A₁₆ elements, in contrast to another case in which nucleosome stability was dependent on the acetylation state of the histones [1]) and is plotted versus time. The superimposed lines represent the results of fits to a single exponential decay. See Materials and Methods for further discussion of the kinetic analysis. Note that a 20,000-fold-lower enzyme concentration was used for the digestion on naked DNA.

FIG. 5

Effects of poly(dA-dT) elements on position-dependent equilibrium accessibilities of nucleosomal DNA target sites. Results for poly(dA-dT)-containing constructs are determined relative to those for reference construct 601.2, which were measured on an absolute scale (2). See Fig. 1 for nucleosomal locations of the different restriction sites. Open bars, 601.2 reference; shaded bars, 601.3(A₁₆End); hatched bars, 601.3(A₁₆Mid). Note the log scale for K_eq^conf.