Mechanism of histone H1-stimulated glucocorticoid receptor DNA binding in vivo

Abstract

Xenopus oocytes lack somatic linker histone H1 but contain an oocyte-specific variant, B4. The glucocorticoid receptor (GR) inducible mouse mammary tumor virus (MMTV) promoter was reconstituted in Xenopus oocytes to address the effects of histone H1. The expression of Xenopus H1o [corrected] (H1) via cytoplasmic mRNA injection resulted in H1 incorporation into in vivo assembled chromatin based on (i) the appearance of a chromatosome stop, (ii) the increased nucleosome repeat length (NRL), and (iii) H1-DNA binding assayed by chromatin immunoprecipitation (ChIP). The H1 effect on the NRL was saturable and hence represents H1-binding to a specific site. A subsaturating level of H1 enhanced the hormone-dependent binding of GR to the glucocorticoid response elements (GREs) and the hormone-dependent MMTV transcription while it reduced the access to DNA as revealed by micrococcal nuclease (MNase) analysis. These H1 effects were lost at higher levels of H1. ChIP and MNase analysis revealed a hormone-dependent dissociation of H1 from the activated chromatin domain. The proposed mechanism of H1-induced GR binding is based on two effects: (i) a GR-induced asymmetric distribution of H1 in favor of inactive chromatin and (ii) an H1-induced reduction in DNA access. These effects results in increased concentration of free GR and, hence, in increased GR-GRE binding.

FIG. 1.

(A) The MMTV reporter construct pMMTV:M13. The black arrow indicates the primer used for DMS methylation protection. Black dots show protected guanines in DMS in vivo footprinting. The GRE probe (−218/−54) was used in MNase experiments. (B) Estimation of intranuclear H1A content. Autoradiography of SDS-PAGE of Xenopus oocyte nuclear extract after injection of 0, 0.7, 1.4, 2.9, and 5.8 ng of H1 mRNA and 4 ng of GR mRNA and incubation with 3 μCi/ml of [¹⁴C]lysine. Right: translated H1A protein is plotted versus injected mRNA as quantified by comparison to GR standard. (C) The experimental design followed unless otherwise indicated. This protocol generated an H1/N ratio of ∼1 when injecting 0.27 ng H1A mRNA and 3 ng ssDNA.

FIG. 2.

(A) An H1-dependent chromatosome stop is detected by MNase digestion. Groups of 10 oocytes were injected with H1 mRNA as illustrated in Fig. 1C. Oocyte homogenate was digested with MNase (5, 15, and 45 U in lanes 1, 2, and 3, respectively, at 25°C for 5 min). DNA was resolved in a 3.6% agarose gel and hybridized with ³²P-labeled pMMTV:M13 DNA. Scans of lanes 2 and 3 are shown to the right with stippled lines for calculated sizes (in bp). (B) Effect of histone H1 on NRL. The procedure was the same as described for panel A but with the injection of 4 ng ssDNA and using a 1.6% agarose gel. Scans of lane 1 versus lane 4 and lane 1 versus lane 13 are shown. (C) Effect of histone H1 on NRL at low temperature. The procedure was the same as described for panel A but with H1 mRNA injected to achieve an H1/N ratio of ∼1.3 and digestion of 9 U MNase at 15°C for 5 min. Arrowheads mark the MNase ladders. Right: plot of DNA length versus nucleosome number. Open and black circles show the data points for without and with H1 chromatin (−H1 and +H1), respectively. The lines show linear regressions which indicate NRLs of ∼164 and ∼176 bp for −H1 and +H1 chromatin, respectively. (D) Effect of H1 on DNA topology. Injections were as described for panel A and illustrated in Fig. 1C but using 3 ng ssM13mp18. DNA was separated on a 1% agarose gel. Right: profiles of scanned lanes show the distributions of topoisomers, and the stippled line indicates the midpoint for cells without H1. (E) An H1-induced transition in topology effect at an H1/N ratio of ∼2.6. The procedure was the same as described for panel D but with double samples of pools of 10 oocytes (open circles) and average values (black dots).

FIG. 3.

(A) Binding of GR, NF1, and Oct 1 to hormone-activated MMTV promoter is enhanced at an apparent H1/N ratio of ∼1. DMS methylation protection analysis of oocytes injected as illustrated in Fig. 1C with mRNA mixes to achieve the indicated H1/N ratios and containing 3.68, 0.76, and 3.22 ng per oocyte of GR, NF1, and Oct1 mRNAs, respectively, is shown. Duplicate pools of eight oocytes were analyzed. Methylation protection of indicated bands (open circles) was quantified and related to the sum of indicated reference bands (black dots), with the absence of hormone level (−TA) set to 100%. Right: diagrams of DMS methylation; black dots are individual samples and bars are mean values. (B) Binding of GR in the absence of NF1 and Oct1 is enhanced at an apparent H1/N ratio of ∼1. DMS methylation protection analysis of triplicate pools of eight oocytes was as illustrated in Fig. 1C and described for panel A but used 2.8 ng of GR mRNA. The results in the diagrams are the averages of triplicates with standard deviations shown as error bars. (C) Constitutive binding of NF1 to the inactive MMTV promoter is not affected by an apparent H1/N ratio of 1 to 2. DMS methylation protection analysis of duplicate pools of eight oocytes injected and assayed as described for panel A but with 2.1 ng GR mRNA and increasing amounts of NF1 and Oct1 mRNAs, i.e., 0, 0.23, and 1.4 ng NF1 and 0, 1.6, and 6.7 ng of Oct1 mRNA per oocyte, respectively. No hormone was added. (D) The H1-dependent enhancement of GR binding is more prominent at low glucocorticoid hormone concentration. DMS methylation protection analysis of duplicate pools of eight oocytes was as described for panel A but with an mRNA mix for GR, NF1, and Oct1 mRNA of 3.9, 0.39, and 3.7 ng, respectively, in the absence (−, dark gray bars) or presence (+, light gray bars) of 0.35 ng H1 mRNA to generate an H1/N ratio of ∼1.3. Hormone (TA) was added at the concentrations indicated.

FIG. 4.

Maximal H1-stimulated GR binding by DMS methylation protection analysis in vivo (A) occurs at half-maximal H1 binding to chromatin as defined by the H1-induced effect on NRL by MNase digestion in situ (B and C). Oocytes were injected and hormone treated as illustrated in Fig. 1C with mRNA mixes to achieve indicated H1/N ratios, i.e., 0, 0.11, 0.18, 0.25, 0.37, 0.56, and 0.76 ng H1 mRNA, and also containing 2.3, 0.35, and 1.8 ng per oocyte of GR, NF1, and Oct1 mRNAs, respectively, and then injected with 3 ng ssDNA MMTV reporter in the absence (A) or presence (B and C) of 90 nCi of [α-³³P]dCTP to in vivo label the DNA. Triplicate pools of 10 oocytes were analyzed by DMS methylation protection for each pool (A), and 100% methylation is defined by the DMS methylation in the absence of H1 and hormone (−TA, stippled line), while the black line describes the plot of apparent H1/N ratio versus GR binding monitored as methylation protection in the presence of hormone (+TA). The results in the diagrams are the averages of triplicates, with standard deviations indicated by error bars. (B) Ten oocytes from each pool without hormone treatment were taken for MNase analysis of the NRL. Isolated DNA was resolved in a 1.8% agarose gel. MNase ladders were visualized by PhosphorImager analysis of dried agarose gels.

FIG. 5.

(A) Histone H1 enhances basal MMTV transcription only at very high H1/N ratios. S1-nuclease analysis of MMTV-driven transcription of groups of 8 oocytes injected with 2.8 ng of GR mRNA and increasing amounts of H1 mRNA to achieve the apparent H1/N ratios indicated was done according to the procedures illustrated in Fig. 1C. No hormone was added (TA). (B) Hormone-activated MMTV transcription is enhanced at an apparent H1/N ratio of ∼1 and is partially inhibited at higher levels. S1-nuclease analysis of MMTV transcription of duplicate pools of eight oocytes injected with mRNA mixes containing 3.7, 0.8, and 3.2 ng of GR, NF1, and Oct1 mRNA, respectively, and amounts of H1 mRNA to achieve the apparent H1/N ratios indicated was done according to the procedures illustrated in Fig. 1C. (C) An apparent H1/N ratio of ∼1 enhances hormone-dependent MMTV transcription. The experiment was as described for Fig. 4B but used triplicate pools of eight oocytes and injection with mRNA mixes containing 1.2, 0.9, and 2.3 ng of GR, NF1, and Oct1 mRNA, respectively.

FIG. 6.

Oocyte-specific linker histone B4 neither competes with H1 binding nor has an effect on the MMTV transcription. (A) Oocytes were injected (inj.) with the indicated amounts of B4 mRNA, and the increase in B4 content is shown by a Western blot probed with anti-B4 (αB4) on dissected nuclei, quantified below. (B) No effect on hormone-activated MMTV transcription was induced by an ∼2.7-fold increase in histone B4. S1-nuclease analysis of MMTV transcription was carried out with triplicate pools of six oocytes injected with 1.8, 0.34, and 1.4 ng of GR, NF1, and Oct1 mRNA, respectively, and increasing amounts of B4 mRNA, followed by 3 ng of single-stranded pMMTV:M13 DNA to achieve the B4/N ratios indicated, estimated as shown in Fig. 1B. (C) ChIP analysis: B4 does not compete for H1 binding to chromatin. Oocytes injected with 0.6 ng HA-H1 mRNA to generate an ∼1.3 H1/N ratio and with or without 3.9 ng B4 mRNA and 3 ng of single-stranded pMMTV:M13 DNA to render an apparent B4/N ratio of ∼8.2. IP, immunoprecipitation.

FIG. 7.

(A) Histone H1 dissociates from the promoter and the transcribed region upon hormone activation. Oocytes were injected as described for Fig. 3A and with 0.6 ng of HA-H1 mRNA, followed by ChIP analysis in duplicate. The average amount of DNA relative to input (gray staples) precipitated for samples not treated with hormone was 1.25% at the promoter region, 1.34% at the TK reporter gene, and 0.57% at the vector region. These precipitated amounts were normalized to 1 in the diagram; black dots represent double samples of two independent immunoprecipitations (IPs). (B) Oocytes were injected with GR, NF1, and Oct1 mRNA mix as described for Fig. 3A and also with H1 mRNA to obtain an apparent H1/N ratio of ∼1.3. Hormone treatment (TA) was as indicated: +, present; −, absent. For MNase digestion and analysis, see Fig. 3B. The filter was hybridized with ³²P-labeled GRE probe (Fig. 1A) and then, after stripping, rehybridized with ³²P-labeled M13 probe.

FIG. 8.

Chromatin is more resistant to MNase in situ digestion at an H1/N ratio of ∼1 to 1.3. Pools of 15 oocytes were injected as illustrated in Fig. 1C with H1 mRNA to achieve the indicated H1/N ratio and with 3 ng of single-stranded M13 mp18 DNA. The oocyte homogenates were subsequently digested with 0, 1.9, and 3.8 U of MNase at 18°C for 5 min. Stop mix with an 882-bp ³³P-labeled PCR fragment was added to each sample and served as an internal standard (Ref.). Isolated DNA was digested with HindIII/BglII to generate a 720-bp DNA fragment. This was resolved on a 1.6% agarose gel, transferred, and hybridized with ³²P-labeled HindIII/BglII M13 fragment. Protection is defined as digested HindIII/BglII fragment/nondigested (%), where −H1 oocyte pools were normalized to 100%. Black dots represent the two concentrations of MNase digestion.

FIG. 9.

The proposed mechanism of H1-enhanced binding of GR to the GRE in vivo. Hormone-activated GR induces a selective dissociation of H1 from the active chromatin domain. An asymmetric H1 equilibrium is formed that selectively reduces the DNA access in the inactive chromatin domain. This reduces the amount of GR that is occupied by searching through the nonspecific DNA and hence increases the free GR pool that is available for interaction with the GRE.