Analysis of chromatin structure of rat [alpha]1-acid glycoprotein gene; changes in DNase I hypersensitive sites after thyroid hormone, glucocorticoid hormone and turpentine oil treatment
Analysis of chromatin structure of rat [alpha] 1 -acid glycoprotein gene; changes in DNase I hypersensitive sites after thyroid hormone, glucocorticoid hormone and turpentine oil treatmentToru Matsukawa, Hiroki Kawasaki, Makoto Tanaka and Yoshiki Ohba*
Faculty of Pharmaceutical Sciences, Kanazawa University, 13-1 Takara-machi, Kanazawa, Ishikawa 920, Japan
Received March 7, 1997;Revised and Accepted May 6, 1997
ABSTRACT
Transcription of the rat [alpha]1-acid glycoprotein (AGP) gene is activated by glucocorticoid, thyroid hormone (T3) and cytokines. Following these treatments, the chromatin structure of this gene was analyzed by means of digestion with DNase I or micrococcal nuclease. Four DNase I hypersensitive sites were observed in the 5'-upstream region of the rat AGP gene of liver cells. They were designated HS1, HS2, HS3 and HS4 (3' -> 5'). After T3 treatment the sensitivity of HS1 and HS2 increased and after dexamethasone (Dex) treatment that of all four sites did so. Three new sites appeared after turpentine oil treatment, while the sensitivities of HS3 and HS4 increased. We conclude that transcriptional activation of the gene by T3 and Dex have very similar mechanisms, but that at the inflammation stage they become slightly different. The increase in sensitivity at HS1 and HS2 after T3 treatment in vivo was successfully reproduced in a cell-free system by in vitro treatment with T3. HS1, HS2 and HS3 were also sensitive for micrococcal nuclease.
INTRODUCTION
The rat [alpha]1-acid glycoprotein (AGP) gene is a single copy gene (1 ) and one of the acute phase protein genes which is expressed in liver cells in the acute phase response (2 ,3 ). Cytokines such as IL-1 and IL-6 activate synthesis of AGP (4 ,5 ). Glucocorticoids also induce AGP (4 ,5 ) and we found that expression of the rat AGP gene was also controlled by thyroid hormone [3,3',5-L-triiodothyronine (T3)] (6 ). The glucocorticoid response element (GRE) of the rat AGP gene is localized in the 5'-upstream region of the gene (7 ) and a thyroid hormone response element (TRE) has been identified in the first intron (6 ). It still remains unclear, however, how ligand-receptor complexes activate mRNA synthesis. Steroid hormone receptors have been shown to interact in vitro directly with components of the initiation complex, but the physiological significance of these interactions also remains unclear (8 ). Finally, there have been indications of the existence of co-activators that act as bridging factors between steroid hormone receptors and the transcription initiation complex (9 ), but the character of these co-activators is not yet well understood.
Chromatin structure plays an important role in regulating gene expression (10 -12 ) and the DNase I hypersensitivity assay is the most common method to analyze the structures of chromatin. DNase I hypersensitive sites (HS) are fundamental biological elements as they are ubiquitous among eukaryotes and can be divided into many classes which differ in form and function. Furthermore, many HSs have been mapped at a number of specific positions for known functions, including promoters, upstream activation sequences and enhancers of inducible genes (13 ,14 ). The acute phase protein genes are controlled by many factors, as mentioned above, but the chromatin structures of these genes have not been thoroughly investigated. We used the rat AGP gene as a model system to study changes in chromatin structure after treatment with these factors.
MATERIALS AND METHODS
Materials
Wister male rats were purchased from Sankyo Laboratory Service Co. (Toyama, Japan). DNase I and micrococcal nuclease were provided by Worthington Biochemical (Freehold, NJ). HindIII, BamHI, RNase A and proteinase K were purchased from Toyobo Co. Ltd (Tokyo, Japan) or Takara Shuzo Co. Ltd (Ohtsu, Japan) and [[alpha]-32P]dCTP from DuPont/NEN Research Products (Boston, MA). All other chemicals were obtained from Wako Pure Chemical Corp. (Osaka, Japan) or Nacalai Tesque Inc. (Kyoto, Japan).
Removal of thyroid and adrenal glands from rats and hormone treatment
Thyroids or adrenal glands were removed from 4-week-old Wister rats, who were then fed for 7 days to let them recover. Their diet was supplemented with either 1% lactate calcium or 0.5% NaCl solution. T3 (2 mg/kg body wt) was injected into rats i.p. as described (6 ). A synthesized glucocorticoid, dexamethasone (Dex), was dissolved in ethanol and then mixed with an equal volume of water. Dex (2 mg/kg body wt) was injected i.p. and livers were removed from rats 12 h after injection. Turpentine oil (2 ml/kg body wt) was injected under the back skin of the rats.
Digestion and purification of DNA
Liver and spleen nuclei were prepared from ~6-week-old rats as described previously (6 ). For DNase I digestion, purified nuclei were suspended in 60 mM KCl, 15 mM NaCl, 15 mM Tris-HCl, pH 7.5, 6 mM MgCl2, 0.05 mM CaCl2, 10% glycerol, 0.5 mM DTT at a concentration of 108 nuclei/ml and incubated on ice for 10 min with or without DNase I. Alternatively, the nuclei were digested with micrococcal nuclease as described elsewhere (15 ), with some modifications, i.e. 1 mM PMSF, 1 mM DTT and 0.1 mM EDTA were added to the digestion buffer and the nuclei were incubated for 5 min. After addition of 10 mM EDTA, the nuclei were centrifuged and suspended in TE (10 mM Tris-HCl, pH 8.0, 1 mM EDTA), to which were added 1% SDS, 100 mM NaCl, 5 mM EDTA and 0.3 mg/ml proteinase K, followed by incubation at 37oC for 12 h. Proteins were removed by successive treatments with phenol, phenol/chloroform (1:1) and chloroform/isoamyl alcohol (24:1), all accompanied by gentle mixing and each treatment repeated twice. After precipitation with ethanol, the DNA was dissolved in TE to which 100 [mu]g/ml RNase A and 200 U HindIII or BamHI were added. The solution was then incubated at 37oC for 3 h. Proteins were removed again as described above and purified DNA obtained. This purified DNA (100 [mu]g) was further treated with 200 U HindIII or BamHI and, after the protein had been removed as described above, DNA samples for gel electrophoresis were obtained.
Incubation of nuclei with T3
Liver nuclei from 6-week-old rats were purified as described above, washed in 0.25 M sucrose, 20 mM Tris-HCl, pH 7.8, 1.1 mM MgCl2, 1 mM PMSF and 1 mM DTT and suspended in the same buffer at a concentration of 108 nuclei/ml, to which was added 10-12 or 10-10 M T3 or solvent. The nuclei were incubated for 30 min at 20oC.
Probe labeling and Southern blotting
BamHI (at +381 bases from one of the transcription initiation sites; 16 )-HindIII (+952) and PstI (-64)-BamHI (+381) fragments of the rat AGP gene were labeled with [[alpha]-32P]dCTP using the Multiprime DNA labeling system (Amersham, Little Chalfont, UK) and used as probes.
Southern blot analysis was performed by the method described by Sambrook et al. (17 ) with some modifications. DNA (20 [mu]g/lane) digested by either DNase I or micrococcal nuclease was electrophoresed respectively on a 1 or 1.5% agarose gel (14 * 13 cm) in 40 mM Tris-acetate, 1 mM EDTA at 60 V for ~6 h, then blotted onto a BA85 nitrocellulose filter (Schleicher & Schuell, Dassel, Germany) by means of overnight capillary transfer and indirectly end-labeled by hybridization with a 32P-labeled probe at 42oC for 16 h. The filters were washed twice with 0.2* SSC at 65oC for 30 min.
RESULTS
Changes in DNase I hypersensitive sites of the rat AGP gene induced after T3 injection
Soon after removal of the thyroid or adrenal glands, AGP mRNA and protein levels increased rapidly, then declined gradually. Five days after removal of the adrenal or the thyroid glands the mRNA level had decreased to half or an eighth respectively that of normal rats and these low levels were maintained for some time (data not shown). T3 or Dex was injected 7 days after removal of the glands.
In order to analyze HSs formed on the rat AGP chromatin by T3 treatment, nuclei were isolated from the liver 5 h after T3 injection. The isolated nuclei were then treated with DNase I, after which the DNA was isolated and treated with HindIII. Following electrophoresis, the DNA fragments were hybridized with a 32P-labeled probe consisting of a BamHI-HindIII fragment. As shown in Figure 1 , on incubation with 8 U/ml DNase I (lanes 5 and 6) five radioactive bands were detected consisting of 1.1, 1.3, 3.1, 6.6 and 17 kb fragments, where the 17 kb fragment was a HindIII-HindIII fragment and the others were DNase I accessible sites and designated HS1, HS2, HS3 and HS4 respectively. HS1 was located at ~0.2 kb, HS2 at 0.4 kb, HS3 at 2.2 kb and HS4 at 5.7 kb upstream from one of the transcription initiation sites of the rat AGP gene (16 ; lower panel in Fig. 1 ). These sites could still be detected on the chromatin derived from thyroidectomized and non-T3-treated rats, while the intensity of the bands was enhanced, especially on HS1 and HS2, after treatment with T3. When the enzyme was reduced to 4 U (lanes 3 and 4), only HS3 and HS4 could be detected, indicating that these two sites were more sensitive to DNase I than HS1 and HS2. It should be noted that no detectable band was observed at the TRE site.
DISCUSSION
In the rat AGP gene of control liver nuclei four HSs, HS1, HS2, HS3 and HS4, were observed and these HSs were identified in all the experiments performed. Because no clear band was observed in spleen nuclei (Fig. 1 ), it was concluded that all HSs were tissue-specific hypersensitive sites. Furthermore, because the sensitivities of the four sites were very low in chromatin prepared from adrenalectomized rats (Fig. 3 ), where the amount of rat AGP mRNA declined to an eighth of that of normal rats (data not shown), formation of these HSs seemed to be necessary for expression of the gene. All sites were inducible and no constitutive sites (13 ) were found in rat AGP chromatin.
The HS patterns after T3 or Dex treatment remained more or less the same (Figs 1 -3 ). All four bands were observed and the sensitivities of HS1 and HS2 always increased. Furthermore, in hyperthyroid rats the sensitivities of HS3 and HS4 increased as much as they did after Dex treatment. These results suggest that the regulation mechanisms induced by T3 and Dex are very similar. Although the receptors and their hormone response elements were different, the hormones produced almost the same changes in chromatin structure. It was noted that the TRE and GRE sites were insensitive to DNase I, regardless of hormone treatments and this suggests that no detectable change occurred at either site.
In addition to the four bands, three new HSs, THS1, THS2 and HS1', were observed after turpentine oil treatment. Furthermore, the sensitivity of HS4 greatly increased, but that of HS2 did not change (Fig. 4 ). These results suggest that the activation mechanism induced by inflammation is different from that induced by T3 and Dex. Among the turpentine oil-specific sites, THS2 and HS1' were located near the previously reported binding sites of inflammation-specific transcription factors. THS2 was close to a cis-element, from -5300 to -5150 bp, known as the IL-6 response element (20 ). Several proteins (21 ) and a transcription factor, NFIL-6 (22 ), bind to this region. HS1' was located near the NFIL-6 binding site (-118 to -68 bp) (23 ). In the mouse albumin gene NFIL-6 reportedly binds to the distal HS region of the gene (24 ). It is thus likely that inflammation-specific transcription factors were involved in formation of the turpentine oil-specific HSs.
HS1, HS2 and HS3 were also sensitive to micrococcal nuclease (Fig. 6 ). This indicates that they are located at the linker regions of nucleosomes or at gaps in the nucleosome structure. This finding agrees with that of previous reports (14 ). Because neither the sensitivity nor the location of the HS1 and HS2 regions for micrococcal nuclease changed after T3 treatment, it can be assumed that neither movement nor displacement of nucleosomes in these regions occurred during transcriptional activation by T3.
After purified nuclei were treated with T3, HSs were induced within 30 min at the same positions as observed in vivo (Fig. 5 ). Since the physiological T3 concentration, 10-10 M, was high enough to induce the changes of chromatin structure, this result indicates that de novo protein synthesis was not required for induction of HSs. As discussed above, since no changes were observed in nucleosome arrays in the HS regions it is likely that association, dissociation and/or modification of a transcription factor(s) cause a change(s) in the chromatin structure at HS1 and HS2 and that they play a role in induction of AGP. Several transcription factors have been reported to bind to regions near HS1 [-125 to -65 bp (25 ) and-180 to -120 bp (26 )]. We also identified several protein binding sites in a region near HS2 (unpublished observations), so that these binding proteins must be related to formation of HS1 and HS2 and to stimulation of transcription. With the in vitro system described here the molecular mechanism of induction after hormone treatment can be investigated in detail.
ACKNOWLEDGEMENTS
This work was supported by Grants-in-Aid for Science Research (C) from the Ministry of Education, Science, Sports and Culture of Japan to T.M. and Y.O.
