Developmental changes in histone macroH2A1-mediated gene regulation

Abstract

macroH2A histone variants have been implicated to function in gene silencing by several studies, including ones showing a preferential association of macroH2A on the inactive X chromosome. To examine macroH2A function in vivo, we knocked out macroH2A1. macroH2A1 knockout mice are viable and fertile. A broad screen of liver gene expression showed no evidence of defects in X inactivation but did identify genes that have increased expression levels in macroH2A1 knockouts. macroH2A1-containing nucleosomes are enriched on the coding and/or upstream regions of these genes, suggesting that their increased expression levels are a direct effect of the absence of macroH2A1. The concentrations of macroH2A1 nucleosomes on these genes are low in the livers of newborn mice, and the macroH2A1 knockout had little effect on the expression levels of these genes in newborn liver. Our results indicate that an increase in liver macroH2A1 during the transition from newborn to young-adult status contributes to a decrease in the expression levels of these genes. These genes cluster in the area of lipid metabolism, and we observed metabolic effects in macroH2A1 knockouts. Our results indicate that the function of macroH2A1 histones is not restricted to gene silencing but also involves fine tuning the expression of specific genes.

FIG. 1.

Absence of macroH2A1 proteins in the macroH2A1 knockout. (A) Diagram of macroH2A. The H2A region is ∼65% identical to conventional H2A. +++ indicates a highly basic region that may bind DNA. macroH2A1.1 and 1.2 differ in a single ∼30-amino-acid segment in the nonhistone region. (B) Western blots with antibodies against nonhistone regions of macroH2A1.1 and 1.2 (left) and macroH2A2 (right) are shown; note that this macroH2A2 serum has some cross-reaction with macroH2A1.1 and 1.2, so all three macroH2A variants show. macroH2A1 variants are much more abundant in mouse liver than macroH2A2 (7). Total nuclear extracts were prepared from normal mouse liver (lanes N) and macroH2A1 knockout mouse liver (lanes KO). Loadings were equalized using core histones (see stained gel below the macroH2A2 blot).

FIG. 2.

Northern blot analysis of Serpina7 and Lpl expression in macroH2A1 knockout mice. Total mouse liver RNA was analyzed on Northern blots for Serpina7 and Lpl expression. The blots were rehybridized with β actin to show equal loadings. KO, RNA from female macroH2A1 knockout mouse liver; N, RNA from normal female mouse liver.

FIG. 3.

Purification of macroH2A1-containing nucleosomes by thiol-affinity chromatography. Mouse liver mono- and oligonucleosomes were prepared, and macroH2A1-containing chromatin fragments were purified by selective thiol affinity chromatography (5). S, starting material; E, material eluted from thiopropyl Sepharose with β-mercaptoethanol. The mercaptoethanol-eluted chromatin is highly enriched for macroH2A1-containing nucleosomes. Results for normal mouse liver are shown on the left, and results for macroH2A1 knockout liver are on the right. Protein compositions were analyzed by sodium dodecyl sulfate gel electrophoresis (top panels). DNA was analyzed in agarose gels (middle panels). The bottom panel is a Western blot of the starting material and a pool of the eluted fractions stained with antibodies against macroH2A1.1 and 1.2; these lanes were loaded for equal amounts of H3 and H4.

FIG. 4.

Distribution of macroH2A1 on genes that have altered expression levels in macroH2A1 knockout mouse liver. Results for adult liver (2-month-old mice) are on the left, and those for newborn liver (3-day-old mice) are on the right. macroH2A1-containing chromatin fragments were purified from normal mouse liver (Fig. 3). The relative macroH2A1 concentration on a sequence was estimated by using real-time PCR to calculate the relative concentration of the sequence in the macroH2A1-enriched, thiopropyl-eluted fraction in comparison to that of the starting material. A value of 1 on the scales on the left of the diagrams indicates a macroH2A1 concentration equal to that of the starting material prepared from adult liver. A few sites tested with the genes of the adult samples were not tested with the newborn samples. The horizontal bars on the bottom of the charts indicate the transcribed regions of the genes.

FIG. 5.

Developmental changes in the macroH2A1 content of mouse liver nuclei. Nuclei were isolated from the livers of newborn (3-day-old) and adult (2-month-old) mice. Total nuclear extracts were analyzed on Western blots using antibodies against macroH2A1.1 and 1.2. Gels were loaded for equal content of core histones; see Coomassie blue-stained lanes below the blot. Lane N, extract from newborn livers; lane A, extract from adult livers.

FIG. 6.

Glucose tolerance is altered in macroH2A1 knockout mice. Mice were fasted overnight and injected with 2 g glucose/kg of body weight intraperitoneally, and blood glucose concentrations were measured at 0, 15, 30, 60, 90, and 120 min. The error bars indicate the standard errors of the means. *, P < 0.05 (two-tailed t test); **, P < 0.001 (two-tailed t test). Numbers of mice: 63 normal males, 49 knockout males, 44 normal females, and 47 knockout females.