| Nucleic Acids Research | Pages |
Overproduction of histone H1 variants in vivo increases basal and induced activity of the mouse mammary tumor virus promoter
Introduction
Materials And Methods
Plasmid constructs
Construction of stable cell lines
Cell culture
Ribonuclease (RNase) protection assays
Results
Construction of cell lines
Effect of H1 overproduction on MMTV promoter activity
Effect of H1 overexpression on transiently transfected MMTV reporter genes
Effect of core histone acetylation on MMTV promoter activity
Discussion
Acknowledgements
References
Overproduction of histone H1 variants in vivo increases basal and induced activity of the mouse mammary tumor virus promoter
Received March 11, 1999; Revised and Accepted June 21, 1999
ABSTRACT BALB/c 3T3 cell lines containing integrated copies of the MMTV promoter driving a reporter gene were constructed. Expression vectors in which either of two H1 variants, H10 or H1c, were under control of an inducible promoter were introduced into these lines. Surprisingly, overproduction of either variant resulted in a dramatic increase in basal and hormone-induced expression from the MMTV promoter. H1 overproduction also slowed the loss of MMTV promoter activity associated with prolonged hormone treatment. Transiently transfected MMTV reporter genes, which do not adopt a phased nucleosomal arrangement, do not display increased activity upon H1 overproduction. Thus the effects observed for stable constructs most likely represents a direct effect of H1 on a chromatin-mediated process specific to the nucleosomal structure of the integrated constructs. Induction of increased levels of acetylated core histones by treatment with trichostatin A also potentiated MMTV activity and this effect was additive to that caused by H1 overproduction. However, the effects of TSA treatment, in control or H1-overproducing cells, were eliminated by inhibiting protein synthesis. TSA treatment does not necessarily potentiate MMTV promoter activity by increasing core histone acetylation within the MMTV promoter but perhaps by altering the synthesis of an unlinked transcriptional regulator.
INTRODUCTION
The organization of eukaryotic DNA into chromatin has important consequences for the transcription process (1-7). In vitro studies indicate that, in general, nucleosomes have a repressive effect on transcription by preventing access of transcription factors to their recognition sites. Genetic studies in yeast clearly demonstrate that the core histones and their covalent modifications play a major role in transcriptional regulation in vivo (8). In higher eukaryotes the presence of linker histones, such as H1 and H5, adds additional complexity (9-11). Inclusion of H1 during in vitro assembly of pol II or pol III promoters into chromatin often results in repression of transcription beyond that conferred by core nucleosomes alone (12-16). Linker histones facilitate the condensation of nucleosomal DNA into 30 nm fibers (17,18) and in this capacity are considered to function as global transcriptional repressors in vivo (19). H1s might also modify the accessibility of transcriptional regulators by limiting nucleosome mobility (20,21) or by reducing transient exposure of DNA from the nucleosome surface (22,23).
Numerous biochemical studies indicate that linker histones are under-represented in transcriptionally active chromatin or the manner in which they bind is altered (11,24,25). The removal or remodeling of H1 may be an important aspect of the activation of some genes. Several molecules including high mobility group proteins (26,27), nucleoplasmin (28) and prothymosin [alpha] (29) have been proposed to modulate the interaction of histone H1 with chromatin. Direct competition between linker histones and certain classes of transcription factors seems plausible in light of the similar structure of their respective domains involved in DNA/chromatin binding (30-32), and evidence for this has been presented (33). Alternatively, any of the recently described chromatin remodeling complexes (34,35) may function in part by displacing H1 or altering the mode of binding to chromatin, although this has not been directly demonstrated.
Several recent studies indicate that H1 has specific effects on the expression of some genes. The involvement of H1 in the selective repression of oocyte versus somatic 5S genes in Xenopus is well documented (15,16). Recently it was shown that differential nucleosome positioning on the oocyte and somatic 5S genes favors binding of H1 to the former and TFIIIA to the latter (36,37). In vivo studies in Xenopus (38-40), Tetrahymena (41) and mouse (42,43) support the view that H1 can modulate the expression of a subset of genes. This is probably accomplished by modification of chromatin architecture but the details and mechanisms are unclear.
We developed a system for the overexpression of individual H1 variants and mutagenized constructs in homologous mouse 3T3 cells (44). We achieved high levels of overproduction and demonstrated significant differences with respect to cell cycle progression and gene expression following the overexpression of two variants, H10 and H1c (42). Transformants overproducing H10 exhibited transient inhibition of both G1 and S phase progression and significantly reduced expression of all genes tested. In contrast, overproduction of H1c to comparable levels had no effect on cell cycle progression and actually resulted in increased transcript levels from some genes. We recently demonstrated that these differences in gene expression are due to differences in the globular domain of these variants (43).
The mouse mammary tumor virus (MMTV) is a well-characterized system in which transcriptional activation is intimately linked to chromatin structure (45-55). In vivo, the hormone-responsive promoter of the MMTV-long terminal repeat (MMTV-LTR) is organized in a phased array of six positioned nucleosomes (46,47). Prevailing models for activation of the MMTV promoter suggest a bimodal process initiated by the binding of the hormone-receptor complex to hormone response elements (HRE) in a positioned nucleosome (Nuc B) adjacent to the transcriptional start site (48). This binding is facilitated by the rotational and translational positioning of Nuc B relative to the HREs (49,50) and results in a remodeling of chromatin that involves a reconfiguring of Nuc B and the recruitment of nuclear factor 1 (NF1) (45,48,51-54). In a second step, hormone-dependent recruitment of additional factors results in the establishment of a preinitiation complex (PIC) and ultimately activation of transcription. Most importantly, for this study, evidence has been presented that removal or reorganization of H1 may be an important component of the activation of this promoter (55). Although a great deal has been learned from studies in which the MMTV promoter is reconstituted into nucleosomes in vitro, some aspects of the transcriptional regulation, including NF1 binding, are not reproduced in these structures (52). In this study we investigated the effect of in vivo overproduction of H1 variants on the activity of the MMTV promoter. We obtained the surprising result that overproduction of either H10 or H1c results in dramatically increased levels of expression from this promoter. We also present evidence that core histone acetylation contributes to the regulation of the MMTV promoter in an indirect manner.
MATERIALS AND METHODS
Plasmid constructs
All plasmids were constructed using standard molecular biological techniques (56) by joining sequences derived from plasmids obtained commercially (see Fig. 1 for schematics). Plasmid pMMTVGFPhyg contains the MMTV-LTR promoter driving the coding region of the mammalian codon-optimized, red-shifted, `enhanced' green fluorescent protein (GFP) reporter gene. The chloramphenicol acetyltransferase (CAT) gene of pMAMneoCAT (Clontech) was excised with NheI and XhoI and replaced with the GFP gene from pEGFP-C1 (Clontech). The resulting plasmid was cut with HindIII and XmnI and the larger fragment was ligated to the large HindIII-XmnI fragment from pTK-Hyg (Clontech) to generate pMMTVGFPhyg. Plasmid pBPVMMTVCAT contains the MMTV-LTR gene driving the CAT gene and was constructed by fusing the 3.5 kb HindIII fragment from pMAMneoCAT to the 8 kb EcoRI-HindIII fragment from pdBPVMMTneo vector [American Type Culture Collection (ATCC)]. Plasmid pGRpur contains the coding sequences for the human glucocorticoid receptor (hGR) driven by the constitutive Rous sarcoma sirus (RSV)-LTR promoter and was constructed by ligating the large XmnI-NdeI fragment from pRShGR (ATCC) to the large PvuII-NdeI fragment from pPUR (Clontech). Construction of expression vectors pMTH10Aneo, pMTH1cAneo and pMTAneo (control expression vector not carrying any H1 coding sequences) has been described previously (42).
Figure 1. Structure of the plasmids used. Plasmid pMMTVGFPhyg carries the `enhanced' green fluorescent protein (GFP) reporter gene driven by the glucocorticoid inducible MMTV-LTR promoter with the gene for hygromycin (hyg) resistance as the selectable marker. Plasmid pGRpur carries the constitutive RSV LTR promoter driving the expression of the human glucocorticoid receptor (hGR) gene and the gene for puromycin (pur) resistance as the selectable marker. Plasmid MTH1Aneo is the prototype for the H1 expression plasmids. These plasmids carry a gene for the wild-type mouse histone H10 or H1c variants, or their mutant forms, driven by the heavy-metal inducible mouse metallothionein I (MT) promoter and employs the gene for neomycin (neo) resistance as a selectable marker.
Construction of stable cell lines
Plasmids pMMTVGFPhyg and pRShGRpur were co-transfected into mouse BALB/c 3T3 fibroblasts (clone A31 from ATCC) using the Calcium Phosphate Transfection System (Gibco BRL) according to the manufacturer's instructions. Cells were selected with 100 µg/ml puromycin and 200 µg/ml hygromycin (Clontech) for several weeks. Stable, double-resistant, single-cell-derived colonies were isolated using cloning cylinders (Bel Art Products) and put through two more rounds of selection with the antibiotics to ensure clonal purity. Individual colonies were then expanded and characterized for dexamethasone (DEX)-inducible GFP expression by fluorescence microscopy and by ribonuclease protection assays (RPAs) performed on total cytoplasmic RNA. Several independent cell lines showing high levels of DEX-induced transcription from the MMTV promoter were further characterized by Southern blots, which revealed that the pMMTVGFPhyg construct had integrated at multiple sites in the cellular genome. One stable cell line, MMTVGFP4, was independently transfected with various H1 expression plasmids and selected with 400 µg/ml G418 sulfate (Gibco BRL). From these transfections, stable G418-resistant colonies were obtained and screened for zinc-inducible histone H1 variant overexpression as previously described (42). Several independent cell lines capable of expressing GFP upon DEX induction and exogenous H1 variants upon zinc induction were obtained and used for subsequent studies. Cell lines are denoted by shorthand for the transfected H1 expression vector; for example cell line MTH10 was transfected with pMTH10Aneo. H1 variant overexpressing cell lines containing integrated copies of the MMTV-CAT reporter gene were constructed using a similar strategy.
Cell culture
All experiments were performed with cells initiated from primary stocks of stable cell lines stored in liquid nitrogen. Cells were routinely maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated calf serum (Gibco BRL) at 37°C under 5% CO2. For the overexpression of H1 histone variants during exponential growth conditions, cells were seeded at a low density; typically, <10% of the total surface area of the flask was covered with cells after the seeded cells had reattached. Twelve hours after the initial seeding, these exponentially growing cells were treated with 50 µM ZnCl2 for 12 h. This was followed by 72 h of induction with 100 µM ZnCl2. The medium was replaced every 24 h. Control cultures, where applicable, were treated exactly the same, except that no ZnCl2 was added. Cells were harvested before they approached confluency. For the overexpression of H1 histone variants under conditions of density arrest, cells were seeded at higher densities and allowed to attain growth arrest due to contact inhibition. Cultures were then induced with ZnCl2 exactly as described above for the induction of exponentially growing cells. After a total of 84 h of induction with ZnCl2, exponentially growing or density arrested cells were induced for various periods of time with 2 × 10-7 M DEX.
Ribonuclease (RNase) protection assays
RPAs were carried out essentially as described previously (42,43). Briefly, total cytoplasmic RNA was harvested from the different cell lines by the method of Gilman (57) and total cell counts for each sample were determined. After ethanol precipitation, the RNA was resuspended in calculated amounts of RNA storage solution (Ambion) to normalize its concentration on a per cell basis. To ensure correct normalization of the RNA samples, aliquots were routinely analyzed on 1% agarose gels and stained with ethidium bromide to visualize the levels of 18S rRNA, which were quantitated by laser scanning densitometry (Molecular Dynamics). 32P-labeled RNA probes were generated by in vitro transcription using the MAXIscriptTM kit (Ambion) and gel purified as described by the manufacturer. To quantitate steady-state MMTV transcript levels, RPAs were carried out on RNA from equal numbers of cells using the RPA IITM, RPA IIITM or HybSpeed RPATM kits (Ambion) according to the manufacturer's instructions. The RNase-treated samples were run on 5% polyacrylamide, 8 M urea gels. The intensity of the protected fragment corresponding to transcripts from the MMTV-LTR was quantitated by phosphorimaging on a Molecular Dynamics Storm Phosphorimager. Control experiments to ensure probe excess and a linear range of responses for the RPAs were carried out routinely.
RESULTS
Construction of cell lines
The generation of cell lines to assay the effect of in vivo H1 overproduction on the activity of the MMTV promoter was done in two steps. A mammalian codon-optimized, red-shifted variant of the GFP gene was placed under control of the MMTV promoter and inserted into a vector carrying a selectable hygromycin resistance gene to generate pMMTVGFPhyg (see Fig. 1 for plasmid constructs). This vector was co-transfected with a plasmid (pGRpur) bearing a human glucocorticoid receptor gene and a selectable puromycin resistance gene into mouse BALB/c 3T3 fibroblasts. We isolated stable single colony transformants, resistant to both drugs, and screened for isolates that expressed the reporter gene in a DEX-dependent manner. One of these isolates, MMTVGFP4, was used for most of the studies presented here. Importantly, as assayed by genomic Southern blots, this isolate maintained the integrated MMTV-GFP reporter gene at a constant copy number during propagation for >2 years. We, and others (58), have experienced instability in copy number of some integrated MMTV reporter constructs during propagation. Isolate MMTVGFP4, which has also maintained a consistent DEX response, was then transfected independently with pMTH10Aneo and pMTH1cAneo overexpression constructs (Fig. 1) and with the expression vector lacking H1 sequences (MTAneo). We isolated neomycin-resistant transformants and screened for isolates that could be induced to overproduce the respective exogenous H1 variants. Total histones were then extracted from crude chromatin preparations by H2SO4 solubilization and separated by HPLC. As we observed previously (42,43), the normal H1 variant ratio and the total amount of chromatin-bound H1 histone can be significantly perturbed in these cell lines (Table 1).
Table 1. H1 variant stoichiometries following induction of exponentially growing cultures of H1 variant overexpressing cell linesa
| Reporter gene | Overexpression construct | Molecules per nucleosomeb | ||||
| H10 | H1b | H1d/e | H1c | Total | ||
| MMTV-GFP | MTA | 0.14 | 0.12 | 0.34 | 0.20 | 0.80 |
| MMTV-GFP | MTH10 | 0.86 | 0.17 | 0.22 | 0.10 | 1.35 |
| MMTV-GFP | MTH1c | 0.15 | 0.11 | 0.18 | 0.75 | 1.19 |
| MMTV-CAT | MTA | 0.19 | 0.11 | 0.32 | 0.16 | 0.78 |
| MMTV-CAT | MTH10 | 0.90 | 0.08 | 0.23 | 0.11 | 1.32 |
| MMTV-CAT | MTH1c | 0.09 | 0.07 | 0.18 | 0.85 | 1.19 |
bValues were estimated as previously described (42).
Effect of H1 overproduction on MMTV promoter activity
We assayed the effect of in vivo H1 overproduction on expression from the MMTV-GFP reporter construct (Fig. 2). Exponentially growing cultures were treated for 4 days with ZnCl2 to induce overproduction of H1 variants and then treated with DEX for 4 h. Steady-state levels of transcripts from the MMTV-GFP reporter gene were determined by RPA (Fig. 2A and B). The results were clear and somewhat surprising. Overproduction of either H10 or H1c resulted in a dramatic increase in both basal and DEX-induced expression of MMTV-GFP. This result was unexpected as previous reports suggested that loss of H1 accompanied MMTV promoter activation (55).
Figure 2. Overexpression of histone H1 variants enhances expression from the MMTV-GFP reporter gene. Exponentially growing or density arrested cultures of the indicated cell lines were treated, as described in Materials and Methods, in the presence or absence of ZnCl2, for 84 h. Cultures were then treated for 4 h with or without DEX prior to isolation of total cytoplasmic RNA. (A) Representative RPA analysis of transcripts from the MMTV-GFP reporter gene. RNA was from exponentially growing cultures. The arrowhead indicates the protected fragment corresponding to initiation from the MMTV promoter. The lane labeled P contains a diluted aliquot of the full-length RNA probe which was not treated with RNase. The lane labeled C contains probe hybridized to 5 µg torulla yeast RNA and treated with RNase. Total chromatin-bound histones were isolated from parallel cultures and analyzed by HPLC (Table 1). (B and C) Quantitation of the effect of H1 overexpression on the expression of the MMTV-GFP reporter gene. Exponentially growing (B) or density arrested (C) cultures were treated with ZnCl2 +/- DEX as described above. Error bars represent the standard deviation from the results of RPA analysis of RNA from at least three independent experiments.
In the experiments described above, the deposition of exogenous H1s and the induction of the MMTV promoter were done in exponential cultures in which a significant portion of the cells are undergoing DNA synthesis. A distinction has been made between dynamic and pre-emptive competition between chromatin structure and transcriptional activators in that the latter requires ongoing replication to provide access of the activators to their DNA binding sites (6,14,59). Previous studies indicated that DEX-induction of the MMTV promoter does not require replication (46,48). However, we have detected differences in the nucleosome repeat length between exponentially growing and density arrested cells in control and H1 overproducing lines (submitted for publication). We therefore investigated the effect of H1 overproduction in density arrested cells on the expression of the MMTV promoter. Under these conditions we also observed a dramatic increase in basal and hormone-induced MMTV promoter activity upon overproduction of either H10 or H1c (Fig. 2C).
Although we present data from a single stable transformant for each construct, we routinely analyze several independent isolates. Slight quantitative differences may be noted between individual isolates transformed with the same construct. This usually reflects differences in the level of overexpression of the exogenous H1 construct. In addition to MMTVGFP4, we also used other isolates as the host line for transfection of H1 overexpression constructs. These included additional independently isolated lines containing the MMTV-GFP reporter gene, lines containing integrated copies of the MMTV promoter driving the bacterial CAT gene and lines containing the MMTV-GFP reporter gene on a stably replicating episomal bovine papilloma virus vector. In all cases, we observed the same result; overproduction of H10 or H1c resulted in dramatically increased levels of expression from the reporter gene. Results from representative isolates expressing the MMTV-CAT reporter gene are displayed in Figure 3 and Table 1. The effects we observe are dependent upon induction of overexpression of the exogenous construct and are not seen in parallel untreated cultures. Identical treatments of a transformant containing the expression vector without H1 sequences does not display this behavior. Thus the effects we observe are not due to clonal variation. We routinely isolate total cytoplasmic RNA from equal numbers of cells for each transformant and RPAs are normalized to a `per cell' basis. As previously reported (43) we observe little effect of overexpression of exogenous H1 variants on rRNA levels (data not shown).
Figure 3. Overexpression of histone H1 variants enhances expression from the MMTV-CAT reporter gene. Cell lines containing the MMTV-CAT reporter gene, on plasmid pBPVMMTVCAT, and H1 overexpression constructs were isolated essentially as described for the MMTV-GFP lines. Although the pBPVMMTVCAT plasmid contains BPV sequences, Southern blot analysis revealed that the sequences were stably integrated into the genome. Exponentially growing cultures of the indicated cell lines were treated, as described in Materials and Methods, in the presence or absence of ZnCl2, for 84 h. Cultures were then treated for 4 h with or without DEX prior to isolation of total cytoplasmic RNA. A representative RPA analysis of the protected fragment corresponding to transcripts from the MMTV-CAT reporter gene is displayed. Total chromatin-bound histones were isolated from parallel cultures and analyzed by HPLC (Table 1).
It is well documented that the chromatin remodeling and activation of the MMTV promoter in response to hormone is transient (58). Upon prolonged exposure to hormone, the promoter becomes refractory to activation and re-establishes a closed conformation. We tested the effect of prolonged DEX exposure on the response of our MMTV-GFP cell lines (Fig. 4). Control cultures displayed the previously reported decrease in expression from the MMTV promoter. However, the H1 overexpressing cell lines maintained high levels of MMTV expression during prolonged hormone treatment. In these lines, the levels of MMTV-GFP transcripts decrease somewhat with prolonged hormone treatment. However, even after 96 h these levels are higher than that of control cells following 4 h of DEX treatment. The persisting levels of MMTV transcripts are due to continuing activity of the promoter in the presence of DEX, as removal of the hormone results in the transcript levels returning to basal levels within 24 h (data not shown). We conclude that overproduction of H1 in vivo potentiates the initial activation of the MMTV promoter and also inhibits or slows the re-establishment of the inert state upon prolonged hormone treatment.
Figure 4. Overexpression of H10 or H1c prevents rapid down-regulation of transcription from the MMTV promoter in the continuous presence of glucocorticoids. Density arrested cultures were treated for the indicated times with DEX. Culture medium was changed daily. (A) Representative RPA analysis performed as described in Materials and Methods. (B) Quantitation of transcript levels. Error bars represent the standard deviation from the results of RPA analysis of RNA from at least three independent experiments.
Effect of H1 overexpression on transiently transfected MMTV reporter genes
The activation of the MMTV promoter by the hormone-GR complex is currently believed to be a bimodal process (48). In this model, binding of GR to GRE elements results in a remodeling of chromatin which facilitates binding of NF1, which is occluded prior to activation. In a second step, a hormone and GR-dependent recruitment of transcription factors, including TFIIB, establishes a PIC. The bimodal nature of hormone-GR activation of the MMTV promoter is illustrated by comparing the behavior of transiently transfected versus stably integrated (or stably replicating episomal) reporter genes (60). Transiently transfected genes exhibit constitutive nuclease sensitivity and NF1 loading in the absence of hormone, suggesting that these genes have not acquired the repressive chromatin structure. However, upon addition of hormone transcriptional activity is elevated by recruitment of additional transcriptional factors and establishment of a PIC. If the effect of H1 overproduction on MMTV promoter activity is due to alterations in chromatin structure prior to the establishment of the PIC then transiently transfected templates would not be affected by increased levels of H1.
We transiently transfected reporter genes, in which the MMTV promoter was fused to the bacterial CAT gene, into cell lines containing integrated MMTV-GFP reporter genes. Using gene-specific RNA probes we simultaneously assayed transcripts from the two types of template using an RPA (Fig. 5). The data clearly show that, in contrast to the stably integrated genes, transiently transfected MMTV reporter genes do not display increased activity upon H1 overproduction. This is an important observation as transiently transfected constructs are believed not to adopt the phased nucleosome arrangement characteristic of stably integrated genes (60). Thus the effects we observe for our stable constructs most likely represent a direct effect of H1 on a chromatin-mediated process specific to the nucleosomal structure of the integrated constructs. The transcript levels from the transiently transfected templates were increased somewhat by hormone treatment, probably reflecting the recruitment of transcription factors as described above. That this response is the same across all cell lines and is not affected by H1 stoichiometry suggests that levels of GR and/or the rate of DEX entry do not differ significantly among these cell lines.
Figure 5. H1 ovexpression does not affect expression of transiently transfected MMTV reporter genes. The indicated cell lines, which contain integrated MMTV-GFP reporter genes, were treated with or without 50 µM ZnCl2 for 12 h, followed by 100 µM ZnCl2 for another 36 h under exponential growth conditions. These cells were then transfected with plasmid pMMTV-CAT, which contains the CAT reporter gene driven by the MMTV LTR, using the Calcium Phosphate Transfection System (Gibco BRL). Forty-eight hours post-transfection, the cells were treated with or without 2 × 10-7 M DEX for 4 h. Total cytoplasmic RNA was isolated and the MMTV-driven GFP and CAT transcript levels were quantitated by RPA using riboprobes specific for each reporter gene. Note that the samples not treated with ZnCl2 also show high levels of H1 variant overexpression from the metallothionein promoter, which is apparently activated under the conditions of the calcium phosphate transfection. This was confirmed by HPLC (data not shown).
Effect of core histone acetylation on MMTV promoter activity
A moderate increase in acetylation of core histones potentiates the activation of the MMTV promoter (61). This was concluded on the basis of the effect of treatment of cell lines, similar to our own, with the deacetylase inhibitor trichostatin A (TSA). We wished to confirm whether this same effect would be seen in our lines and, if so, the relationship to the H1-mediated activation of the MMTV promoter. We treated exponentially growing cultures of control and H1 overproducing cell lines with various concentrations of TSA and assayed the effect of TSA treatment on MMTV promoter activity following a 4 h exposure to hormone (Fig. 6). Treatment of control MTA cells with TSA resulted in a 5- to 10-fold increase in basal (Fig. 6B) and DEX-induced (Fig. 6C) MMTV promoter activity, in agreement with the previous report (61). Treatment of H1 variant overexpressing cell lines with TSA resulted in increased levels of MMTV transcription 4- to 5-fold greater than that due to H1 overproduction alone, i.e. the effects were additive. These results suggest that H1 overproduction and histone acetylation potentiate MMTV promoter activity by independent mechanisms. This conclusion is valid only if the effect of TSA treatment is near saturating levels. We generated a dose-response curve with levels as high as 500 ng/ml TSA (data not shown; Fig. 7). We observed a plateau in the effect on MMTV promoter activity at concentrations >50 ng/ml. We did not observe a reduction in activity, as was reported previously in other cell lines (61), within this concentration range.
Figure 6. Effect of TSA treatment on expression from the MMTV LTR promoter. (A) Exponentially growing cultures were treated with ZnCl2 for 84 h, followed by treatment for 12 h in the presence or absence of 5 or 50 ng/ml TSA. Cultures were then treated for 4 h in the presence or absence of DEX, in the continued presence or absence of TSA, prior to RNA isolation. Steady-state levels of transcripts from the MMTV-GFP reporter gene were determined by RPA. (B) Quantitation of the effect of TSA treatment on the basal level of expression from the MMTV promoter. Values are relative to the MTA cell line in the absence of TSA treatment. (C) Quantitation of the effect of TSA treatment on DEX-induced expression from the MMTV promoter. Values are relative to the MTA cell line in the absence of TSA treatment. Open bars, MTA; striped bars, MTH10; filled bars, MTH1c0.
Figure 7. Effect of cycloheximide on the expression of the MMTV promoter. Exponentially growing cultures were treated with ZnCl2 for 84 h, followed by treatment for 8 h +/- 250 ng/ml TSA +/- 50 µg/ml cycloheximide (CHX), as indicated. DEX was added to all cultures during the final 4 h. Steady-state levels of transcripts from the MMTV-GFP reporter gene were determined by RPA. Numbers above the lanes indicate the band intensity relative to that of MTA in the absence of TSA or CHX.
We found that the effects of TSA treatment and H1 overproduction on MMTV promoter activity differed in sensitivity to the protein synthesis inhibitor cycloheximide (Fig. 7). Exponential cultures were treated in the presence or absence of TSA and CHX. The increase in MMTV promoter activity attributable to H1 overproduction was not affected by cycloheximide treatment (compare lanes 5 to 6 and 9 to 10). However, the effects of TSA treatment, whether in control or H1 overproducing cells were eliminated by inhibiting protein synthesis (compare lanes 3 to 4, 7 to 8 and 11 to 12).
We investigated the effect of cycloheximide treatment on core histone acetylation by separating the modified forms on Triton-acid-urea (TAU) gels. If histone acetylases are more sensitive to cycloheximide treatment than deacetylases, perhaps due to a shorter protein half-life, then such treatment should affect the overall level of histone acetylation. We find that cycloheximide treatment does not affect the distribution of acetylated isoforms (Fig. 8A). We also analyzed parallel samples from those used in the RPA analysis of Figure 7 (Fig. 8B). This gel reveals that the TSA treatment protocols used result in a significant increase in hyperacetylated forms of core histones, most notably H4. Furthermore, cycloheximide treatment did not affect the distribution of acetylated isoforms in TSA-treated samples. As these gels measure bulk histone modifications, it is possible that a specific modification of core histones at the MMTV is cycloheximide sensitive. We feel this is unlikely, in part due to the observation that treatment with cycloheximide alone does not affect MMTV promoter expression (Fig. 7). An alternative interpretation of these results is that TSA treatment does not necessarily potentiate MMTV promoter activity by a direct effect of increasing core histone acetylation within the MMTV promoter but perhaps by altering the synthesis of an unlinked transcriptional regulator.
Figure 8. Cycloheximide treatment does not affect core histone acetylation. Exponentially growing cultures were treated with ZnCl2 for 84 h, followed by treatment for 8 h +/- 250 ng/ml TSA +/- 50 µg/ml cycloheximide (CHX), as indicated. DEX was added to all cultures during the final 4 h. Total histones were acid extracted and subjected to electrophoresis on TAU gels (72). (A) The gel demonstrates that treatment with CHX alone does not affect core histone acetylation. (B) The gel contains material isolated from parallel samples to those used for the RPA analysis in Figure 7.
DISCUSSION
The major observation of these studies is that in vivo overproduction of either of two H1 variants, H10 or H1c, results in a dramatic increase in the basal and hormone-induced expression of a reporter gene driven by the MMTV promoter. This was unexpected for two reasons. In previous studies, overproduction of the H1c variant was associated with increased expression of several genes, but we only observed gene repression upon H10 overproduction (42,43). In addition, data from immunoprecipitation experiments suggested that reduced amounts of H1 were associated with MMTV chromatin following activation (55). We offer several possible explanations all of which are predicated on the assumption that overproduction of H1 in vivo alters chromatin architecture.
Although the nucleosomes surrounding the MMTV promoter are considered positioned, recent high resolution in vivo mapping indicates that nucleosomes in the Nuc B and Nuc A region are present in a frequency-biased population of non-overlapping translational frames (47). It has been suggested that some frames are more amenable to facilitating cooperative binding of the GR (47,52,54). Increased levels of H1 in vivo might result in an alteration of positioning to favor those frames. The basal level of MMTV activity, in the absence of hormone treatment, may be due to a subset of nucleosome positions in which the NF1 binding site is an accessible region. We observe dramatically increased levels of basal MMTV activity upon H1 overproduction, which may be explained in the same way. H1 has been demonstrated to modulate nucleosome mobility and affect nucleosome positioning in vitro (20,21,62). In two recent studies of the selective repression of Xenopus 5S genes, H1 was shown to direct nucleosome positioning such that TFIIIA binding was excluded from the oocyte promoter and favored by the somatic promoter (36,37).
It should be noted that treatment protocols which result in high levels of overproduction of H1 variants (Table 1) result in an increase in the total amount of H1 per nucleosome to ~1.5 times that found in untreated or control cells. However, a detailed analysis (submitted for publication) revealed only subtle changes in chromatin structure under these conditions. These cultures are fully viable and no evidence of aberrant chromatin structure was detected. The accumulation of chromatin-bound H1 to these levels implies that some nucleosomes would contain two molecules of H1. Recent models for an asymmetric localization of H1 in nucleosomes allow for the presence of two binding sites per nucleosome and direct evidence for a second site has been presented (63). Presumably, under conditions in which H1 stoichiometry is approximately one per nucleosome, this second site is unoccupied or occupied by other proteins. H1 has been shown to compete with or at least to bind to the same site in nucleosomes as transcription factors (33) and other DNA binding proteins, notably HMGs (26,27) and the methylated DNA binding protein MeCP2 (64). Overexpression of H1 might displace other chromatin proteins that serve to silence MMTV chromatin in the absence of hormone or during prolonged hormone treatment. The observation that overproduction of either variant prevents or slows the proposed reclosing of MMTV chromatin during extended DEX treatment is consistent with this possibility. Finally, the effects of H1 need not be confined to their nucleosomal binding sites. Binding of linker histones has been shown to induce changes in core histone-DNA interactions in the nucleosome (65,66). This suggests a mechanism, independent of the effects on nucleosome mobility, by which H1 might influence GR or transcription factor binding.
Overproduction of the H10 or H1c variant to high levels results in reduced amounts of the other variants, including those of the predominant H1d/e class. It might be argued that the loss of these variants is responsible for the observed increase in MMTV promoter activity. We think this unlikely for the following reasons. The H1c and H1e variants are 80% identical in total amino acid sequence and differ at only two positions in the central globular domain (67,68). A recent analysis of the in vitro binding of mouse H1 subtypes to nucleosomally organized MMTV promoter sequences revealed a high degree of similarity among all five somatic variants in binding affinity to mononucleosomes and in the ability to aggregate polynucleosomes (69). Finally, in protocols in which we observed a significant enhancement of MMTV promoter activity (Figs 2 and 3), the amounts of H1d/e were only moderately reduced, although the total amount of H1 per cell was increased (Table 1). In fact, the most consistent parameter associated with increased MMTV promoter activity is the total H1 stoichiometry and probably reflects a general property associated with H1 function such as limiting nucleosome mobility or promoting formation of higher order chromatin structure.
The results with transiently transfected templates (Fig. 5) provide the strongest evidence that the increase in MMTV promoter activity by the in vivo overproduction of H1 is mediated by alterations in the MMTV chromatin structure. Transiently transfected MMTV templates do not adopt the phased nucleosomal arrangement of identical stably integrated templates and display numerous differences in their regulatory responses (reviewed in 60). Our results indicate that the activity of transiently transfected templates is not affected by H1 stoichiometry. Thus the effects we observe for our stable constructs probably most likely reflect a direct effect of H1 on the nucleosomal or higher order structure of the integrated templates.
It is important to note that the observations described here are not necessarily incompatible with the previous observation that decreased amounts of H1 are associated with activation of the MMTV promoter (55). Prevailing models for activation of the MMTV promoter suggest a multistep process initiated by the binding of the hormone-receptor complex to HREs in a positioned nucleosome (Nuc B) adjacent to the transcriptional start site (48,51-54). This binding is facilitated by the rotational and translational positioning of Nuc B relative to the HREs (49,50) and leads to a chromatin remodeling event. This remodeling, which may involve mammalian homologs of the SWI/SNF complex (70), results in a reconfiguring of Nuc B and the recruitment of NF1. In a subsequent step, hormone-dependent recruitment of additional factors results in the establishment of a PIC and ultimately activation of transcription. It is possible that, as described above, increased amounts of H1 promote a more efficient binding of GR to the HRE elements. This does not preclude the subsequent loss or reorganization of H1 during the chromatin remodeling or PIC establishment phases of the activation process.
Recently it was reported that a moderate increase in core histone acetylation enhances transcription of the MMTV promoter in the absence of hormone and potentiates transactivation by glucocorticoids (61). We investigated the relationship of this activation to that we observe upon H1 overproduction. The results were quite clear (Fig. 6). Core histone hyperacetylation by treatment of cells with the deacetylase inhibitor TSA and H1 overproduction activate MMTV promoter activity in an additive fashion. This suggests that, in this system, H1 overproduction and core histone acetylation may each activate MMTV promoter activity by acting at separate steps in the process. The effects of TSA treatment and H1 overproduction on MMTV promoter activity differed in the sensitivity to the protein synthesis inhibitor cycloheximide (Fig. 7). The increase in MMTV promoter activity attributable to H1 overproduction was not affected by cycloheximide treatment; however, the effects of TSA treatment, whether in control or H1 overproducing cells were completely eliminated. We demonstrated that cycloheximide treatment does not affect the bulk distribution of modified core histone isoforms in TSA-treated or untreated cells (Fig. 8). If TSA acts solely by inhibiting deacetylase activity and thereby increasing core histone acetylation at or near the MMTV promoter, this action should be unaffected or enhanced by the inhibition of new protein synthesis. A more plausible explanation is that TSA increases the expression of an unlinked transcriptional regulator that promotes MMTV expression. These results indicate that caution must be exercised in the interpretation of data based solely on the effects of TSA treatment.
Many studies have suggested that a relative depletion of H1 from promoter regions is characteristic of gene activation (reviewed in 11). Yet, we observe a very dramatic activation of MMTV transcription rather than repression upon increasing the H1 stoichiometry in vivo. This behavior may be reflective of the unusual mode of regulation of the MMTV promoter. The organization of this promoter into a phased array of positioned nucleosomes is critical for facilitating the synergistic binding of hormone receptors and NF1 during activation (71). The multistep nature of the activation offers several opportunities for H1-mediated chromatin architecture to influence promoter activity in a positive manner. As described previously, this does not preclude loss or rearrangement of H1 at a later step or after initiation of transcription. In fact, all of our data are consistent with a model in which the loss of H1 may be an important feature of the transient nature of the response of the MMTV promoter to hormone. Re-establishment of a hormone-responsive state may require rebinding of H1, which might occur more efficiently in our overproducing cell lines as compared to cells with a normal stoichiometry of H1.
ACKNOWLEDGEMENTS
The authors are grateful to Dr Donald Sittman for valuable discussions and critical reading of the manuscript. We thank Carla Smith and Melissa Jones for technical assistance. This work was supported by grant MCB-9305308 from the National Science Foundation, an institutional grant from the University of Mississippi Medical Center and a donation from the F.D. Wade Research Fund.
REFERENCES
*To whom correspondence should be addressed. Tel: +1 601 984 1849; Fax: +1 601 984 1501; Email: dbrown{at}biochem.umsmed.edu
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Belikov, C. Astrand, and O. Wrange Mechanism of Histone H1-Stimulated Glucocorticoid Receptor DNA Binding In Vivo Mol. Cell. Biol., March 15, 2007; 27(6): 2398 - 2410. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Song and M. A. Gorovsky Unphosphorylated H1 Is Enriched in a Specific Region of the Promoter when CDC2 Is Down-Regulated during Starvation Mol. Cell. Biol., March 1, 2007; 27(5): 1925 - 1933. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Chen, M. Dundr, C. Wang, A. Leung, A. Lamond, T. Misteli, and S. Huang Condensed mitotic chromatin is accessible to transcription factors and chromatin structural proteins J. Cell Biol., January 3, 2005; 168(1): 41 - 54. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Wang, Y. Peng, Y. Ma, and N. Jahroudi Histone H1-like protein participates in endothelial cell-specific activation of the von Willebrand factor promoter Blood, September 15, 2004; 104(6): 1725 - 1732. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Freidkin and D. J. Katcoff Specific distribution of the Saccharomyces cerevisiae linker histone homolog HHO1p in the chromatin Nucleic Acids Res., October 1, 2001; 29(19): 4043 - 4051. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Thiriet and J. Hayes Assembly into chromatin and subtype-specific transcriptional effects of exogenous linker histones directly introduced into a living Physarum cell J. Cell Sci., January 3, 2001; 114(5): 965 - 973. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Gunjan, B. T. Alexander, D. B. Sittman, and D. T. Brown Effects of H1 Histone Variant Overexpression on Chromatin Structure J. Biol. Chem., December 31, 1999; 274(53): 37950 - 37956. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Gunjan, D. B. Sittman, and D. T. Brown Core Histone Acetylation Is Regulated by Linker Histone Stoichiometry in Vivo J. Biol. Chem., January 26, 2001; 276(5): 3635 - 3640. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. C. Banks, L. J. Deterding, K. B. Tomer, and T. K. Archer Hormone-mediated Dephosphorylation of Specific Histone H1 Isoforms J. Biol. Chem., September 21, 2001; 276(39): 36467 - 36473. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||