A genomic element within the third intron of the human oxytocin receptor gene may be involved in transcriptional suppression

Molecular and Cellular Endocrinology 135 (1997) 129 – 138

A genomic element within the third intron of the human oxytocin receptor gene may be involved in transcriptional suppression Yoshifumi Mizumoto 1, Tadashi Kimura *, Richard Ivell IHF Institute for Hormone and Fertility Research, Uni6ersity of Hamburg, Grandweg 64, 22529 Hamburg, Germany Received 11 September 1997; accepted 13 October 1997

Abstract The human oxytocin receptor (OTR) gene comprises a large (\ 10 kb) third intron between the regions encoding the transmembrane domains six and seven. It has been shown for other genes that transcriptional control elements may reside within such introns, and that these may correlate with changes in the methylation status of the DNA. Methylation mapping indeed indicated that within this third intron there was a region which appeared to be hypermethylated in non-expressing tissues, but relatively hypomethylated in the myometrium of the cycle and at term, when the OTR gene is upregulated. We then employed in vitro nuclear protein-DNA binding assays to evaluate the importance of this region in the control of the human OTR gene. As source of nuclear proteins we have compared a non-expressing tissue, human peripheral blood leucocytes, with human myometrium from the cycle (low expression) and from term pregnancy (high expression). It could be shown that a specific motif of ca. 10–15 nucleotides close to the middle of the third intron specifically binds nuclear proteins correlating with the down-regulated state of the gene. The accumulated data suggest that this intronic element is specifically binding nuclear protein(s) associated with a suppression of OTR gene activity. © 1997 Elsevier Science Ireland Ltd. Keywords: Oxytocin receptor; DNA methylation; Transcription factors; Human; Gene regulation; Myometrium

1. Introduction The nonapeptide hormone oxytocin (OT) mediates a variety of acute reproductive functions, including milk let-down and in particular uterine contractions at birth (reviewed in Ivell and Russell, 1995). It has been shown that, besides increasing amplitude and frequency of pulsatile oxytocin release from the posterior pituitary (Dawood et al., 1979; Fuchs et al., 1991), a significant role is played by the upregulation of the myometrial

* Corresponding author. Present address: Department of Obstetrics and Gynecology, Osaka University Medical School, 2-2 Yamadaoka Suita, Osaka 565, Japan. 1 Present address: Department of Obstetrics and Gynecology, National Defence Medical College, 3-2 Namiki, Tokorozawa, Saitama, 359 Japan.

oxytocin receptor (OTR) at term. This upregulation is demonstrable in several species both at the level of the receptor protein (Fuchs et al., 1984), as well as of its specific gene transcript (Kimura et al., 1996). In fact, it has been shown that there is over a 300-fold increase in the human OTR mRNA within the myometrium of the term uterus, compared with that of the non-pregnant cycle (Kimura et al., 1996). Since obstetric problems associated with preemptive or protracted labour account for more than 30% of births, a better understanding of the molecular mechanisms governing the regulation of the oxytocin receptor and its gene are of considerable clinical as well as fundamental relevance. In general, transcription is regulated by the differential binding of DNA sequence-specific nuclear proteins (transcription factors) within or in the neighbourhood of a gene (reviewed in Mitchell and Tjian, 1989). Such

0303-7207/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 3 0 3 - 7 2 0 7 ( 9 7 ) 0 0 1 9 5 - 0

130

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

binding occurs not only within the 5% upstream region (promoter) of the gene, but also within the proteincoding region and within introns, (Hambor et al., 1993; Henkel and Brown, 1994; Brooks et al., 1994; Betz et al., 1994; Goomer et al., 1994). Cloning of the OTR gene from the human (Inoue et al., 1994) and other species (Bathgate et al., 1995; Rosen et al., 1995; Kubota et al., 1996) has made it possible to search for specific DNA motifs (cis elements) and their cognate transcription factors, particularly within the promoter region of the gene. Indeed, current studies indicate several novel cis elements in the 5% region of the human OTR gene which appear to bind nuclear proteins differentially between term and nonpregnant myometrium (Kimura, unpublished). However, studies in sheep, where the OTR is highly upregulated also in the non-pregnant endometrium at estrus, suggest that specific transcriptional suppression may also play a significant role in OTR gene regulation. Specifically, endometrial fragments explanted from the mid-cycle uterus, at a time when the OTR is relatively downregulated, spontaneously upregulate in culture (Sheldrick et al., 1993), suggesting that in vivo there may be an active suppression of the OTR gene by systemic factors. Also in a marsupial, which has separate gravid and non-gravid uteri during pregnancy, there appears to be a specific suppression of the receptors to below basal levels at term in the non-gravid uterus (Parry et al., 1997). Thus gene suppression may be as important a factor in the regulation of the OTR gene as specific transcriptional activation. The OTR gene in all species so far studied includes a large ( \10 kb) intervening sequence (intron) interrupting the protein-coding region between the sequences corresponding to transmembrane domains 6 and 7. To date there is no information either on the structure and sequence of this intron, nor on any potential role in respect to transcriptional regulation. Since the methylation pattern of DNA has been shown to correlate with the transcriptional status of a gene (Eden and Cedar, 1994; Bird, 1992), we have analysed the differential methylation pattern within this large third intron using methylation-dependent isoschizomeric restriction enzymes, comparing genomic DNA from non-pregnant and term human myometrium, and from non-expressing peripheral blood leucocytes. Furthermore, we have characterized the nuclear proteins from these tissues for their ability to bind to the intronic DNA, and have shown that there is a defined sequence motif (cis element) within the third intron of the human OTR gene, whose proteinbinding properties correlate with the downregulation of OTR in the non-pregnant myometrium, suggesting that this may be a candidate suppressor element.

2. Materials and methods

2.1. Tissue collection Heparinized peripheral blood was obtained from a single male volunteer, and the mononuclear cells separated by standard 5% dextran sedimentation. Samples of non-pregnant myometrium were obtained from patients undergoing hysterectomy because of uterine leiomyomata. Term myometrial samples were obtained at hysterectomy due to severe obstetric hemorrhage from the uterus following delivery. The terms of the Helsinki declaration were observed in all cases. Myometrial samples were washed in ice-cold normal saline, frozen in liquid nitrogen, and stored at − 80°C. Two independent samples of each type of myometrial sample were collected and analysed in all subsequent assays.

2.2. Southern hybridization of genomic DNA The 6.5 kb EcoRI–EcoRI fragment located within the third intron of the human OTR gene (Inoue et al., 1994) was further cleaved into four subfragments (in order: 1.7 kb EcoRI–HindIII, 0.8 kb HindIII–SmaI, 2.5 kb SmaI–KpnI, 1.5 kb KpnI–EcoRI). Each fragment was labelled by incorporation of [a 32P]dCTP (Megaprime DNA labelling system, Amersham-Buchler, Braunschweig, Germany) to high specific activity. A 30 mg sample of high molecular weight genomic DNA, prepared by standard procedures (Sambrook et al., 1989) from peripheral leucocytes, non-pregnant and term myometrium, were first digested with EcoRI restriction endonuclease. Then 10 mg aliquots were further digested with either HpaII or MspI restriction endonucleases. The digested samples were electrophoresed through a 1.2% agarose gel in 0.5×TBE, and transferred by capillary blotting to charged nylon membranes (Hybond-N + , Amersham-Buchler) using 0.4 N NaOH. The resulting southern blots were hybridized independently with all four intronic subfragments (see above) as probes using QuikHyb hybridization solution (Stratagene, Heidelberg, Germany) at 65°C for 2 h. Blots were washed at high stringency and exposed to autoradiographic film (Kodak X-Omat R, Kodak, New York) at − 70°C overnight.

2.3. Northern hybridization Poly(A)-enriched RNA was prepared from the myometrial tissue samples by oligo(dT)-latex bead saturation as described previously (Kimura et al., 1996). This method assures absolutely identical amounts of poly(A)-RNA in each electrophoresed sample. The equivalent of 0.5 mg mRNA per lane was electrophoresed in 1.0% agarose gels using the MOPS/

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

formaldehyde buffer system (Sambrook et al., 1989), and transferred to Hybond-N membranes by capillary transfer. Northern hybridization with an OTR-specific cDNA probe was performed as described previously (Kimura et al., 1996), except that hybridization made use of QuikHyb solution (Stratagene) at 65°C for 2 h.

2.4. DNA sequence analysis The 1.5 kb KpnI – EcoRI intronic subfragment, which is localized at the 3% end of the 6.5 kb EcoRI– EcoRI fragment from the third intron of the human OTR gene (Inoue et al., 1994), was sequenced on both strands by a primer-walking method using the T7 sequencing kit (Pharmacia, Freiburg, Germany). Homology to known transcription factor binding sites was analysed using the TFSEARCH computer programme established by Akiyama at the Kyoto University Medical School (accessed by internet).

2.5. Preparation of nuclear extracts Nuclear proteins were extracted from peripheral blood leucocytes by the procedure of Dyer and Herzog (1995), all steps being carried out at 4°C. A total of 1.5 ×107 cells were suspended in 150 ml of sucrose buffer I (0.32 M sucrose, 3 mM CaCl2, 2 mM Mg-acetate, 0.l mM EDTA, 10 mM Tris – HCI, pH 8.0, 1 mM dithiothreitol (DTT), 0.5mM phenylmethylsulfonylfluoride (PMSF), 0.5% nonidet-P40 (NP-40)) and then centrifuged at 500× g for 5 min to pellet the nuclei. The nuclei were resuspended in 30 ml of low-salt buffer (20 mM HEPES, pH 7.9, 25% glycerol, 1.5 mM MgCl2, 20 mM KCl, 0.2 mM EDTA, 0.5 mM DTT, 0.5 mM PMSF). To extract the nuclear proteins this suspension was supplemented with 30 ml of high-salt buffer (20 mM HEPES, pH 7.9, 25% glycerol, 1.5 mM MgCl2, 0.8 M KCl, 0.2 mM EDTA, 1% NP-40, 0.5 mM DTT, 0.5 mM PMSF), and gently mixed by rotation for 20 min. A 240 ml volume of dilution buffer (25 mM HEPES, pH 7.6, 25% glycerol, 0. 1 M EDTA, 0.5 mM DTT, 0.5 mM PMSF) was then added and the mixture centrifuged at 14 000×g for 15 min. The supernatant (nuclear extract) was stored at −80°C. Nuclear extracts from myometrial tissue were prepared by the method of Deryckere and Gannon (1994). A 1 g sample of non-pregnant or term myometrium, frozen at − 80°C, was pulverized in liquid nitrogen and transferred to a 15 ml Dounce tissue homogenizer (Wheaton, Millville, NJ). The thawed powder was homogenized in 15 ml of solution A (0.6% NP-40, 150 mM NaCl, 10 mM HEPES, pH 7.9, 1 mM EDTA, 0.5 mM PMSF). After 20 strokes, the homogenate was centrifuged for 20 s at 1500 rpm and 4°C. The supernatant was incubated on ice for 5 min and then centrifuged again at 4000 rpm for 10 min. The pelleted

131

nuclei were resuspended in 0.5 ml of solution B (25% glycerol, 20 mM HEPES, pH 7.9, 0.42 M NaCl, 1.2 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 1× ‘Complete’ proteinase inhibitor (Boehringer-Mannheim)) and incubated on ice for 30 min for high-salt extraction. The lysed nuclei were centrifuged at 14 000 rpm for 5 min, and the supernatant (nuclear extract) stored at − 80°C. The protein concentration in the nuclear extracts was determined using the BioRad protein assay kit (BioRad, Mu¨nchen, Germany).

2.6. Electrophoretic mobility shift assays (EMSA) Using the primer combinations listed in Table 1, we prepared ten different, partially overlapping DNA fragments by PCR, with the 1.5 kb KpnI–EcoRI fragment as template (Fig. 1). The 50 ml PCR reaction mixtures contained 16 mM (NH4)2SO4, 67 mM Tris–HCl, pH 8.0, 1.5 mM MgCl2, 0.01% Tween 20, 0.2 mM dNTP, 2 mM each of the sense and antisense primers as listed, 5 U Taq polymerase (Promega, Madison, WI), and 0.5 ng DNA template. Following a 5 min denaturation step at 95°C, PCR products were amplified for 30 cycles (denaturing 1 min at 95°C, annealing 1 min at 60°C, elongation 1 min at 72°C) with a final extension for 10 min at 72°C. The amplified DNA fragments were electrophoresed in a 3% agarose gel, excised and purified using the QIAEX II gel extraction kit (Qiagen, Hilden, Germany). A total of 300 ng of each fragment were Table 1 Sequences of PCR primers used to produce the ten DNA subfragments (F1 – F10) from the 1.5 kb Kpn1 – EcoR1 restriction fragment from intron 3 of the human QTR F1

Sense Antisense

TAGGCTGGGTCTTGTCTAAG ACAGAATGTGCAAGAGTGCC

F2

Sense Antisense

CAGATGCTTTGAGACTGTGG ACAGAATGTGCAAGAGTGCC

F3

Sense Antisense

GCCACTCTTGCACATTCTGT GAAGCATGAAATTAGCAGGC

F4

Sense Antisense

GTAGCCCTTCCAACATACTG TATAACAGCATCACCCCATC

F5

Sense Antisense

ACAATATGATGGGGTGATGC TAGTAAGCTCCTACTCACCC

F6

Sense Antisense

TTTTGAGGGGTGAGTAGGAG TGATAGGCGCCACCTGTTTC

F7

Sense Antisense

AAACAGGTGGCGCCTATCAG TCTGGAAGAAGACAAGCCAG

F8

Sense Antisense

AAACAGGTGGCGCCTATCAG ATGGGTTTCTCTGACAGCAC

F9

Sense Antisense

TTGGAGCAGTGCTGTCAGAG CAATATACACTACAGGGTGC

F10

Sense Antisense

TCTTTTCTAAGAAGGCACCC GGAATTCAGATGTAAGTCCC

132

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

Fig. 1. Structural organization and restriction map of the human oxytocin receptor gene, indicating positions of PCR-derived fragments (F1–F10; see Table 1), relative to the 1.5 kb KpnI–EcoRI restriction fragment. Also shown is the full sequence of the 75 bp F2 fragment, together with the position of the wild-type and mutant 35 bp double-stranded subfragments used for competition in electrophoretic mobility shift assays (only the upper strand is indicated here relative to the 75 bp sequence, the lower strand is complementary). Also illustrated are the relative positions of the 20-mer double-stranded oligomers (oligo 1–5) used similarly as specific competitors. E, EcoRI; H, HindIII; S, SmaI; K, KpnI; M, MspI/HpaII.

end-labelled to a specific activity \106 cpm/mg, using 50 mCi [g32P]ATP (Amersham-Buchler) and T4 polynucleotidyl kinase. A 104 cpm quantity of each DNA probe and 2 mg of nuclear proteins were incubated in 100 mM NaCl, 20 mM Tris – HCl, pH 7.9, 2 mM MgCl2, 1 mM EDTA, 10% glycerol, 0.1% NP-40, 1 mM DTT, 50 mg bovine serum albumin, and l mg poly(dI/dC) in a total volume of 20 ml at 22°C for 30 min. The reactions were loaded onto a 4% non-denaturing polyacrylamide gel (19:1 acrylamide:bisacrylamide)

in 0.5 × TBE buffer and electrophoresed for 2 h at 200 V. The resulting gels were dried under vacuum and exposed overnight to autoradiography film at −80°C. For the competitive inhibition assay, a molar excess (as indicated) of the competing PCR fragment or doublestranded oligonucleotides (Table 2) were mixed into the reaction 10 min prior to the addition of the labelled probes. Competing double-stranded DNA was produced by specific PCR as indicated in Table 1, or by annealing of the respective sense and antisense oligonu-

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

133

cleotides (Table 2). A non-specific competing fragment was created by excising part of the multiple cloning site (MCS) from the intact pBluescript II.SK plasmid (Stratagene) using the restriction enzymes KpnI and SacI.

3. Results

3.1. Northern hybridization of mRNA from non-pregnant and term myometrium Northern hybridization was performed on samples of non-pregnant and pregnant myometrium to verify the relative levels of endogenous oxytocin receptor gene expression (Fig. 2). These are similar samples to those used for the extraction of genomic DNA (see below) and nuclear proteins. It is quite evident that there is a considerable upregulation of the OTR gene in the myometrial sample from the pregnant uterus at term, by comparison with the non-pregnant sample. Neverthless, there is a detectable signal for OTR mRNA in the non-pregnant sample, indicating that in this tissue the gene is in an activated state, but relatively less upregulated.

3.2. Methylation status of the OTR gene third intron in expressing and non-expressing tissues In order to localize potential sites of regulatory activity within the 10 kb of intron 3, methylation status of the genomic DNA was assessed, comparing expressing and non-expressing tissues. A 10 mg quantity of genomic DNA from peripheral blood leucocytes, where the OTR gene is not activated, and from non-pregnant and pregnant myometrium (tissues obtained from two patients in each category) were digested with the methylation-sensitive HpaII endonuclease following complete digestion with EcoRI (Fig. 3). In parallel a control sample of leucocyte genomic DNA was exhaustively digested with EcoRI and the methylation-insensitive HpaII-isoschizomer MspI. Ethidium bromide staining of the resulting gel demonstrated even loading and completely homogenous digestion for all samples. Table 2 20-mer oligonucleotides synthesized and used as competitors in the electrophoresis mobility shift assays Oligo Oligo Oligo Oligo Oligo

l 2 3 4 5

5%-GGCACTCTTGCACATTCTGT-3% 5%-GTCTCCCACCCCAACGGCAC-3% 5%GCCTCCCGGCCCATGTCTCC-3% 5%-GACTGTGGGCTTCTGCCTCC-3% 5%-CAGATCGTTTGAGACTGTGG-3%

Only the sense-strand oligonucleotides are shown; the complementary oligonucleotides were also made and annealed to the sense-strands to produce the required double-stranded competitors.

Fig. 2. Northern hybridization of poly(A)-enriched RNA (0.5 mg per lane) from non-pregnant and term human myometrium, and from Hela cells used here as negative control.

Southern analysis was carried out using four non-overlapping probes derived from the 6.5 kb EcoRI–EcoRI internal fragment of the third intron from the human OTR gene. Comparing the hybridization signals, only the 1.5 kb KpnI–EcoRI fragment, located at the 3% end of the 6.5 kb intron fragment indicated a differential pattern of methylation between peripheral leucocytes, non-pregnant and pregnant myometrium (Fig. 3). From the sequence of the 1.5 kb KpnI–EcoRI fragment (Fig. 4), there are two HpaII/MspI restriction sites close together at 1.25 and 1.31 kb distance from the EcoRI site. This accounts for the : 1.3 kb band seen upon MspI digestion in Fig. 3, which upon closer inspection reveals itself as a doublet. In addition there is a second hybridizing band at : 2.0 kb, presumably the product of cleavage at the MspI site within the labelled fragment and one lying further 5% beyond the KpnI site (Fig. 3). This band shows a weaker hybridization signal because only a small portion of the radiolabelled probe cross-hybridizes with this fragment. Since only the KpnI–EcoRI probe gave rise to a differential methylation pattern, then only the internal HpaII/MspI doublet is of concern, and this is best visualized via the intensity

134

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

of the 1.3 kb band. Upon cleavage of all three tissues by HpaII, which is methylation-sensitive, the 1.3 kb product is clearly of weaker intensity than that resulting from MspI digestion. This shows that this site is markedly methylated in vivo. Secondly, however, this 1.3 kb fragment indicates a consistently higher signal intensity in the sample from the non-pregnant and term

myometrium, than from peripheral leucocytes (arrow), suggesting that in the myometrium this MspI site is relatively hypomethylated by comparison to the nonexpressing tissue. This conclusion is also drawn from the precursor product at 3.3 kb, which is absent in the MspI cleavage, and more intense in the leucocyte sample (arrow) than in the myometrial samples. The least signal intensity in the 1.3 kb product and the greatest intensity in the 3.3 kb precursor fragment, indicating maximum methylation at this 3% MspI doublet site, is found for the DNA from peripheral blood. This site thus appears to be specifically less methylated in those tissues where the oxytocin receptor gene is upregulated.

3.3. In 6itro nuclear protein–DNA binding

Fig. 3. Restriction analysis of genomic DNA extracted from peripheral blood leucocytes (L), from non-pregnant myometrium (N) and from myometrium at term of pregnancy (P). The genomic DNA was first cleaved with either EcoRI alone, or together with the methylation-sensitive enzyme HpaII, or with its methylation-insensitive isoschizomer MspI, as indicated. The restriction digests were then electrophoresed, Southern blotted onto nylon membranes and hybridized against different probes derived from the third intron of the human oxytocin receptor gene. This figure shows the results using the 1.5 kb KpnI – EcoRI fragment as probe. The small arrows indicate the differentially hypomethylated HpaII restriction fragment at 1.3 and 3.3 kb. (A) Ethidium bromide stained gel to validate even digestion and loading of samples. (B) Hybridization of the transferred restriction fragments. (C) Restriction map of the 6.5 kb EcoRI DNA fragment to explain the methylation-dependent, partial digestion of the genomic DNA. E, EcoRI site; K, KpnI site; M, MspI/HpaII site.

The pattern of differential hypomethylation (see above) pointed to the possible involvement of the 1.5 kb KpnI–EcoRI intronic fragment in transcriptional regulation. In order to assess this in greater detail, electrophoretic mobility shift experiments were performed to analyse the complement of nuclear proteins extracted from the three tissue types. Altogether ten partially overlapping subfragments, ranging in size from 75 to 244 bp, were prepared by PCR (Fig. 1) and radiolabelled. Nine of these fragments (no. 1 and 3–10) failed to show any significant and differential binding to the nuclear extracts from the three tissues: blood leucocytes, nonpregnant and pregnant myometrium (not shown). However, using the smaller 75 bp fragment (Fig. 5), which includes one of the differentially methylated HpaII restriction sites identified above (Fig. 4), four protein–DNA complexes are observed (labelled a, b, g and d) when using nuclear proteins from nonpregnant myometrium (Fig. 5A,lane 5), but not with nuclear extract from term myometrium, where only the d complex is evident (Fig. 5A lane 8). And with the nuclear extract from peripheral blood leucocytes, only the g complex can be detected (Fig. 5A,lane 2). The band retardation observed with this fragment was indeed sequence-specific, since it could be competed by an excess of unlabelled fragment F2 (Fig. 5A, lane 3), but not by the non-specific fragment (MCS). Within this 75 bp fragment 2 there is a region located towards the 3% end with high homology (96.8%) to the binding motif for the transcription factor v-Myb (Grotewold et al., 1994). In order to define more precisely the actual site of specific nuclear protein binding, and also to determine whether a Myb-related factor could be involved, we carried out specific competition assays using overlapping 20 mer double-stranded oligonucleotides (Table 2, Fig. 1), covering the whole 75 bp region. The resulting electrophoretic mobility shift assay using nuclear proteins from non-pregnant myometrium (Fig. 5C) indicated that only with oligonucleotide no. 2 was a competition possible (Fig.

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

135

Fig. 4. Nucleotide sequence of the ca 1.5 kb KpnI–EcoRI restriction fragment, indicating the locations of the HpaII sites, the v-Myb-related motif, and the 75 bp F2 fragment (boxed).

5C, lane 4), and then preferentially with the complex a, though also the b and g complexes were reduced in intensity, in some experiments more markedly than in others (not shown). We then constructed a series of double-stranded oligonucleotides, with strategic nucleotides within this approximate region mutated, as shown in Fig. 1. In mutant 2, the core-sequence of the v-Myb binding site (C/TAACGG) is disrupted (CAGCGA). These various oligonucleotides were used

as competitors in a standard retardation assay using the 75 bp fragment 2 as probe, and nuclear extract from non-pregnant myometrium (Fig. 5B). As expected, on addition of the wild type competitor, the a, b and d complexes are diminished proportionately at all concentrations (Fig. 5B, lanes 3–6). The g complex was also reduced, but only at the highest concentrations of competing oligonucleotide. Addition of the different mutant competitors at a 600-fold molar excess indicated com-

136

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

Fig. 5. Electrophoretic mobility shift assays using the F2 fragment as radioactive probe and nuclear extracts derived from peripheral blood leucocytes (L), non-pregnant myometrium (N) and term myometrium (P). (A) Complex formation and competition using the unlabelled F2 fragment or a non-specific fragment derived from a plasmid multiple cloning site (MCS), both at a 100-fold molar excess. (B) Complex formation using the nuclear extract from non-pregnant myometrium only, and competition with increasing amounts (100 – 600-fold molar excess) of the 35 bp wild type subfragment (see Fig. 1), as well as with a 600-fold molar excess of the four mutant fragments (M1 – M4). (C) Complex formation using the nuclear extract from non-pregnant myometrium only, and competition with a 600-fold molar excess of the 20-mer double stranded oligonucleotides (oligo 1 –5; see Fig. 1) as competitors. The four different complexes formed are indicated by greek letters in order of increasing mobility (a– d).

petition for complexes a and b by mutants 1 and 2, but not by mutants 3 and 4. Thus the protein component(s) of the various tissue-specific complexes do not appear to be related to the Myb family. Since the mutation in mutant 1 is near to the 5% end of this :35 bp region, it would imply that this region is likely to represent the cis element interacting with whatever nuclear protein(s) in complexes a and b are specifically expressed in the non-pregnant myometrium, i.e. the relatively suppressed situation with respect to OTR gene expression. Interestingly, the less defined complex d, present in both types of myometrium, but not in peripheral lymphocytes, is by the 75 bp F2 fragment, only poorly by the 35 bp wild type oligonucleotide, and by 20 mutants M3 and M2, suggesting that the cis element recognized here is different from, though overlapping with that for the a and b complexes.

4. Discussion The present study set out to assess a possible role in transcriptional regulation for the large third intron region which is present in the OTR gene in all mammals so far analysed. Our analysis has shown that within this intron there appears to be one region which

shows a differential pattern of methylation when comparing genomic DNA from peripheral blood leucocytes, a non-expressing tissue, with non-pregnant and pregnant myometrium, low and very highly expressing tissues, respectively. Since even term myometrium comprises both expressing and non-expressing cells (Kimura et al., 1996), clear-cut differences in methylation patterns are not to be expected. Differential methylation localizes to a pair of HpaII/MspI sites close together within the KpnI–EcoRI restriction fragment at the 3% end of the 6.5 kb EcoRI–EcoRI intronic restriction fragment, and thus near to the middle of the whole intron. This region appears to be relatively hypomethylated in the myometrium when the OTR gene is moderately or highly upregulated, and more or less hypermethylated in non-expressing peripheral blood leucocytes. The same region was shown by gel retardation assays to form four complexes, a, b, g and d, specifically with nuclear proteins from the non-pregnant myometrium. One of these complexes (g) was also formed with nuclear proteins from blood leucocytes. There was only binding of nuclear proteins from term myometrium to form the d complex; all other complexes were absent. Further competition analysis using smaller and mutated oligonucleotides reduced the specific binding motif to a short sequence of about 10–15

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

nucleotides, with the sequence -GTCTCCCACCCCAAC-. This sequence is unlike any motif published to date, and thus the identity of the protein(s) binding to this region remains obscure. Taking the results together suggests that this region of the third intron may be involved in specific regulation of OTR gene transcription, possibly mediating the interaction of silencer or suppressor proteins with the genome. Whereas numerous positively acting transcription factors or enhancers have been described and their role in upregulating gene transcription characterized, suppressor factors are less well known, and there is no clear consensus regarding mechanisms of action. Several mechanisms have been described for such molecules (Rhodes et al., 1994), ranging from quenching or active repression, in which the cis element is prevented from interacting with a positive activator or one of its coactivators, to direct suppression or silencing, in which the components of the DNA – protein complex negatively interact with the basal transcription complex. More recently, a third mechanism has been proposed whereby DNA methylation, and presumably also suppressor protein – DNA complexes, may stabilize a negative nucleosome partitioning pattern (Kass et al., 1997) and prevent transcription. Whether such mechanisms can occur within intron sequences removed from the transcriptional initiation complex is not clear. DNA methylation appears to interfere with gene expression in at least three different ways. Firstly, it can directly influence the binding of transcription activators or repressors. A number of sequence-specific DNAbinding proteins have been described whose activity is disturbed by DNA methylation. We had originally surmised that tissue-specific hypomethylation in the third intron might be a prerequisite for binding of specific activator proteins. This appears not to be the case, the hypomethylated region we have identified shows no formation of nuclear protein. DNA complexes specifically with extracts from term myometrium by comparison with non pregnant myometrium, although the d complex does appear to be expressed in both myometrial tissues, but not in the peripheral blood leucocytes. It is possible that an inhibitory factor is binding directly to the methylated region of the identified cis element, much like AP-1 in the intron of the mts1 gene (Tulchinsky et al., 1996), and thus preventing upregulation of OTR gene expression in the non-pregnant myometrium. In this context it is interesting to speculate on the identities of the protein components within the a, b, g, and d complexes, of which only the g complex is found also in blood leucocyte nuclei. Secondly, DNA methylation can affect the structure of chromatin. A recent study shows that histone H1, which is known to be associated with general transcriptional inhibition, binds more strongly to methylated than to unmethylated templates (Levine et al., 1993).

137

As mentioned above, specific DNA–protein complexes may also favour chromatin conformations which are inhibitory to active transcription. Thirdly, transcription may be inhibited by non-specific methylcytosine-binding proteins (Boyes and Bird, 1992). Within the region identified in the third intron as correlating with the downregulation of the OTR gene, there is a very good consensus for the binding motif of the v-Myb transcriptional activator (Grotewold et al., 1994). However, effective competition for specific complex formation with an oligonucleotide, in which this motif is mutated, strongly implies that this or similar transcriptional regulators are not involved. It seems likely that the factor(s), such as that in the g complex, binding to the third intron are involved in specific suppression or downregulation of the OTR gene in leucocytes and in the non-pregnant myometrium, and that this might be associated with hypermethylation of the specific intron region concerned. Equally, it is plausible that in a methylation-dependent manner this factor may be replaced by other factors, such as that in the d complex, when the gene is being upregulated. Normally, one could test such effects using transfection assays in appropriate cell-lines where the endogenous OTR gene is highly upregulated. Unfortunately, such cell-lines are at present unknown. Thus future research must resort to a purification of the relevant factor(s) from differentially expressing tissues. Only when such molecules have been fully characterized, can we begin to understand their possible role and mode of action in controlling parturition.

Acknowledgements We gratefully acknowledge the financial support of the Deutsche Forschungsgemeinschaft (grant Iv7/8-1) as well as of the CIBA-Geigy Research Foundation (Japan). T.K. was supported by a fellowship from the Alexander-von-Humboldt Foundation, Bonn. We are also grateful to Drs Ross Bathgate and Norbert Walther for helpful discussions.

References Bathgate, R., Rust, W., Balvers, M., Hartung, S., Morley, S., Ivell, R., 1995. Structure and expression of the bovine oxytocin receptor gene. DNA Cell Biol. 14, 1037 – 1048. Betz, A.G., Milstein, C., Gonzales-Fernandez, A., Pannell, R., Larson, T., Neuberger, M.S., 1994. Elements regulating somatic hypermutation of an immunoglobulin K gene: critical role for the intron enhancer/matrix attachment region. Cell 77, 239 –248. Bird, A., 1992. The essentials of DNA methylation. Cell 70, 5–8. Boyes, J., Bird, A., 1992. Repression of genes by DNA methylation depends on CpG density and promoter strength: evidence for involvment of a methyl-CpG binding protein. EMBO J. 11, 327 – 333.

Y. Mizumoto et al. / Molecular and Cellular Endocrinology 135 (1997) 129–138

138

Brooks, A.R., Nagy, B.P., Taylor, S., Simonet, W.S., Taylor, J.M., Levy-Wilson, B., 1994. Sequences containing the second-intron enhancer are essential for transcription of the human apolipoprotein B gene in the livers of transgenic mice. Mol. Cell. Biol. 14, 42243 – 42256. Dawood, M.Y., Ylikorkala, O., Trivedi, D., Fuchs, F., 1979. Oxytocin in maternal circulation and amniotic fluid during pregnancy. J. Clin. Endocrinol. Metab. 49, 429–434. Deryckere, F., Gannon, F., 1994. A one-hour minipreparation technique for extraction of DNA-binding proteins from animal tissues. Biotechniques 16, 405. Dyer, R.B., Herzog, N.K., 1995. Isolation of intact nuclei for nuclear extract preparation from a fragile B-lymphocyte cell line. Biotechniques 19, 192 – 195. Eden, S., Cedar, H., 1994. Role of DNA methylation in the regulation of transcription. Curr. Opin. Genet. Dev. 4, 255–259. Fuchs, A.-R., Fuchs, F., Husslein, P., Soloff, M., 1984. Oxytocin receptors in the human uterus during pregnancy and parturition. Am. J. Obstet. Gynecol. 150, 734–741. Fuchs, A.-R., Romero, R., Keefe, D., Parra, M., Oyarzun, E., Behnke, E., 1991. Oxytocin secretion and human parturition: pulse frequency and duration increase during spontaneous labor in women. Am. J. Obstet. Gynecol. 165, 1515–1523. Goomer, R.S., Holst, B.D., Wood, I.C., Jones, F.S., Edelman, G.M., 1994. Regulation in vitro of an L-CAM enhancer by homeobox genes HoxD9 and HNF-1. Proc. Natl. Acad. Sci. USA 91, 7985 – 7989. Grotewold, E., Drummond, B.J., Bowen, B., Peterson, T., 1994. The myb-homologous P gene controls phlobaphene pigmentation in maize floral organs by directly activating a flavonoid biosynthetic gene subset. Cell 76, 543–553. Hambor, J.E., Mennone, J., Coon, M.E., Hanke, J.H., Kavathas, P., 1993. Identification and characterization of an Alu-containing, T-cell-specific enhancer located in the last intron of the human CD8a gene. Mol. Cell. Biol. 13, 7056–7070. Henkel, G., Brown, M.A., 1994. PU.1 and GATA: components of a mast cell-specific interleukin 4 intronic enhancer. Proc. Natl. Acad. Sci. USA 91, 7737–7741.

.

Inoue, T., Kimura, T., Azuma, C., et al., 1994. Structural organization of the human oxytocin receptor gene. J. Biol. Chem. 269, 32451 – 32456. Ivell, R., Russell, J.A., 1995. Oxytocin: Cellular and Molecular Approaches in Medicine and Research. Plenum Press, New York. Kass, S.U., Landsberger, N., Wolfe, A.P., 1997. DNA methylation directs a time-dependent repression of transcription initiation. Curr. Biol. 7, 157 – 165. Kimura, T., Takemura, M., Nomura, S., et al., 1996. Expression of oxytocin receptor in human pregnant myometrium. Endocrinology 137, 780 – 785. Levine, A., Yeivin, A., Ben-Asher, E., Aloni, Y., Razin, A., 1993. Histone H1-mediated inhibition of transcription initiation of methylated transcripts in vitro. J. Biol. Chem. 268, 21754–21759. Kubota, Y., Kimura, T., Hashimoto, K., et al., 1996. Structure and expression of the mouse oxytocin receptor gene. Mol. Cell. Endocrinol. 124, 25 – 32. Mitchell, P.J., Tjian, R., 1989. Transcriptional regulation in mammalian cells by sequence specific DNA binding proteins. Science 245, 371 – 378. Parry, L.J., Bathgate, R.A.D., Shaw, G., Renfree, M.B., Ivell, R., 1997. Evidence for a local fetal influence on myometrial oxytocin receptors during pregnancy in the tammar wallaby (Macropus eugenli ). Biol. Reprod. 56, 200 – 207. Rhodes, K., Rippe, R.A., Umezawa, A., Nehls, M., Brenner, D.A., Breindl, M., 1994. DNA methylation represses the murine a1(I) collagen promoter by an indirect mechanism. Mol. Cell. Biol. 14, 5950 – 5960. Rosen, F., Russo, C., Banville, D., Zingg, H.H., 1995. Structure, characterization and expression of the rat oxytocin receptor gene. Proc. Natl. Acad. Sci. USA 92, 200 – 204. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory Press, New York. Sheldrick, E.L., Flick-Smith, H.C., Dos Santos Cruz, G.J., 1993. Oxytocin receptor binding activity in cultured ovine endometrium. J. Reprod. Fert. 98, 521 – 528. Tulchinsky, E.M., Georgiev, G.P., Lukanidin, E.M., 1996. Novel AP-1 binding site created by DNA-methylation. Oncogene 12, 1737 – 1745.