Nuclease sensitivity of the human growth hormone-chorionic somatomammotropin locus in pituitary and placenta suggest different mechanisms for tissue-specific regulation

ELSEVIER Molecular and Cellular Endocrinology 118 ( 1996) l55- 162

Nuclease sensitivity of the human growth hormone-chorionic somatomammotropin locus in pituitary and placenta suggest

different mechanisms for tissue-specific regulation

Barbara E. Nickel, Peter A. Cattini” Department of’ Physiology, Uniwrsity of’ Munitoba, 730 Willium Avenue, Winnipeg, Manitobu. C‘LUKILIU R.IE 3J7

Received I8 December 1995; accepted 5 February 1996

Abstract

The five human growth hormone (GH) and chorionic somatomammotropin (CS) genes are located at a single locus on chromosome 17. These genes share extensive nucleotide sequence similarity ( - 94%) even in their flanking DNA, yet GH-N is expressed efficiently in the pituitary under the control of the pituitary-specific factor GHF-l/Pit-l and the remaining CS-A, CS-B, CS-L and GH-V genes are transcriptionally active in the placenta. Despite this specificity in vivo, a truncated CS-A promoter can bind GHF-l/Pit-l and allow CS-A promoter activity in pituitary cells in vitro. With a view to assessing whether the placental genes of the GHjCS locus possess a different chromatin structure in the pituitary and are, thus, less transcriptionally active than the GH-N gene, we have compared the DNAase I sensitivity of GH/CS in isolated pituitary and placenta cell nuclei. Our data indicate that these genes are equally sensitive in isolated human pituitary nuclei. By contrast, the CS-A, CS-B and CS-L genes were significantly (P < 0.05) more sensitive than the GH-N gene in isolated human placenta nuclei. Although just not significant, the GH-V gene was slightly more sensitive than the GH-N gene. This pattern was also seen with nuclei from human choriocarcinoma BeWo and JEG-3 cells, which express low and extremely low levels of CS RNA, respectively, but was distinct from the pattern observed in the non placental human cervical carcinoma HeLa cell line. These data indicate that the inactivity of the CS genes in the pituitary does not correlate with a ‘closed’ chromatin structure. However, they are consistent with a role for a more ‘open’ chromatin conformation in placenta-specific expression, but not necessarily high levels of transcriptional activity.

Kewords: Growth hormone and chorionic somatomammotropin gene family; DNAase I sensitivity; Pituitary; Placenta; Chromatin

1. Introduction

Human growth hormone (GH) and chorionic somatomammotropin (CS) are members of the physiolog- ically and clinically important GH gene family that also includes prolactin [l]. While GH is essential for postna- tal growth, development and homeostatic control, CS has been implicated in receptor-mediated [2-51, stimu- lation of mammary gland differentiation, maternal pan- creatic islet 0 cell function during pregnancy and fetal growth [6,7]. The five human GHjCS genes are located at a single locus of 47 kilobases (kb) on chromosome 17 [l]. These genes share extensive nucleotide sequence similarity ( N 94%) even in their flanking DNA and

* Tel.: + I 204 7X9 3735: Fax: + 1 204 774 9517.

have clearly evolved by gene duplication [l]. In spite of this homology, control of the GHjCS locus is complex involving multiple regulatory elements and correspond- ing factors. While GH-N is expressed efficiently in the pituitary due to the presence of the pituitary-specific factor variously called GHF-1 or Pit-l (GHF-l/Pit-l) [S-lo], the remaining four genes, CS-A, CS-B. CS-L and GH-V, are expressed in villus syncytiotrophoblast of the placenta [ 1 , 1 1 - 161. Placenta-specific regulation, of at least the CS-B gene, requires multiple regulatory factors [17-241. Despite this specificity in vivo, the CS-A, CS-B and GH-V promoters are able to bind GHF-l/Pit-I in their first 150 base pairs (bp) of 5’- flanking DNA [25,26]. Further. we showed that binding of GHF-l/Pit-l to the proximal region - 97/ - 66 of the CS-A promoter is a requirement for efficient expression in rat pituitary tumor cells after gene transfer

0303-7207,‘96/Sl5.00 0 1996 Eisevier Science Ireland Ltd. All rights reserved

PII SO303-7207(96)03778- I

156 B.E. Nic,kel. P.A. Cirttini :I Moleculur mti Cellular Enrlocrinolo,yy 118 (1996) 15% 162

[26,27]. This raised the question as to why the CS-A promoter is active in pituitary cells in vitro but not in vivo? [28]. Here, we have considered the possibility that the placental genes of the GHjCS locus in the pituitary possess a different chromatin structure and are, thus, less accessible to transcription factors than the GH-N gene.

We have compared the sensitivity of GH-N, CS-A, CS-B, CS-L and GH-V genes to DNAase I treatment in isolated pituitary and placenta nuclei. Our data indicate that while these genes are equally sensitive in isolated human pituitary nuclei, the placental CS genes were significantly more sensitive than the pituitary GH-N gene in isolated human placenta nuclei. The study was also extended to include human placental tumor cell lines (JEG-3 and BeWo) which express relatively low levels of endogenous CS and GH-V compared to term placenta [16,29,30]. In spite of these low levels of hormone synthesis, the pattern of GHjCS gene sensitivity was similar to that observed in term placenta. These data are discussed in terms of transcriptional activity versus the chromatin structure incorporating the GH/ CS gene family in pituitary and placenta cells.

2. Materials and methods

2.1. Tissue und cell lines

Human pituitary tissue was obtained from Drs. I. Worsely and R.P.C. Shiu of the Human Pituitary Repository in the Protein and Polypeptide Laboratory at the University of Manitoba and placental tissue from normal term deliveries at the Health Sciences Centre, Winnipeg, Manitoba. Human choriocarcinoma JEG-3 and BeWo cells were obtained from the American Type Culture Collection and grown in monolayer in RPM1 1640 supplemented with 10% fetal bovine serum in a humidified atmosphere of 95% air and 5% carbon diox- ide. Cells were grown on 150-mm culture dishes and harvested when they were 80%) confluent.

2.2. Isolution of nuclei

All steps were done at 4°C unless otherwise indicated, with 1 mM phenylmethylsulphonyl fluoride and 2 /lg/ml aprotinin included in all buffers used. For pituitary and placenta, frozen tissue was scissor minced and then homogenized using a motor driven homogenizer, in 4 volumes of tissue nuclei isolation buffer (TNIB: 10 mM HEPES pH 7.9, 1.5 mM magnesium chloride, 10 mM potassium chloride, 10 mM sodium butyrate, 0.1% NP-40) containing 25 mgjml iodoacetamide. The homogenate was filtered through four layers of cheesecloth and centrifuged at 1500 x g for IO min. The pellet was resuspended in 4 volumes TNIB

with iodoacetamide and again centrifuged. The nuclei were resuspended in nuclei isolation buffer (NIB: 50 mM Tris-HCl pH 7.5, 2 mM magnesium chloride, 25 mM potassium chloride, 30 mM sodium butyrate, 0.25 M sucrose), and homogenized using a Dounce homogenizer with a B pestle. Nuclei were checked visually by phase contrast microscopy. Following centrifugation as above, nuclei were resuspended in 3 ml NIB. The DNA concentration was determined by measuring the optical density at 260 nm (1 OD,,, = 50 pg/ml DNA) of 5 ~1 of nuclei in 995 ~1 of 5 M urea/2 M sodium chloride.

Frozen cells (BeWo, JEG-3, HeLa) from ten 150-mm plates were thawed in 20 ml DNIB containing 1% Triton X-100 and 25 mg/ml iodoacetamide. A glass homogenizer with a loose fitting teflon pestle was used to resuspend the cells. Cells were centrifuged at 1500 x g for 10 min. The pellet was resuspended in 20 ml NIB containing 1% Triton X-100 and 25 mg/ml iodoacetamide, then homogenized five times using a Dounce homogenizer with the B pestle. Following centrifugation at 1500 x g for 10 min, nuclei were resuspended in 10 ml NIB containing 25 mg/ml iodoacetamide using the Dounce homogenizer. Nuclei were checked by phase contrast microscopy and then resuspended in NIB (about 3 ml) after centrifugation at 1500 x g for 10 min. The DNA concentration was determined as indicated above.

2.3. DNAase I sensitivity analysis

Nuclei were diluted to 20 OD,,, units/ml using NIB and calcium chloride added to a final concentration of 0.1 mM. DNAase I (Sigma D-4527, St. Louis, MO) was prepared at 10 units/p1 in 0.15 M sodium chloride and aliquots stored at - 20°C. Prior to use, the DNAase I was diluted with 0.15 M sodium chloride. Aliquots of nuclei (400 jtl) were preincubated for 2 min in 1.5 ml microfuge tubes and then DNAase I (0 to 40 units in 20 /tl) was added and incubated for 10 min at 37°C. The reaction was stopped with 80 ~1 of 62.5 mM EDTA, 1 .O M sodium chloride and 35.2% sodium-sarcosyl. The level of endogenous nuclease activity was assessed through reactions done in the absence of DNAase I that were stopped after zero and a IO-min incubation. DNA was then isolated as previously described [29] and resuspended in 10 mM Tris-HCl pH 7.5, 1 mM EDTA. DNA (10 pug) was digested for 18 h at 37°C with BarnHI (2.5 units/pg DNA), fractionated by lo/o agarose gel electrophoresis and then transfernred to nitrocellulose. The blots were then probed with a radiolabelled 5’-flanking probe consisting of the region -496/ + 1 of the GH-N gene generated by EcoRI/ BarnHI digestion of phGH-1 [31], or a coding region probe represented by the GH-N cDNA, as described previously [29]. Detection even under stringent conditions was ensured by the greater than 90% sequence

B.E. Nickel, P.A. Cattini / Molecular and Cellular Endocrinolog~~ II8 (19%) 155 162 151

homology between each of these regions in the five GHjCS genes [l]. Densitometry was done on the appro- priate sized bands detected in autoradiographs. The liver-specific human apolipoprotein A-I (apoA-I) gene was used as a control for an inactive gene and the coding region was detected using a 2.2 kb Pst 1 fragment of the apoA-I gene which was described in Karathanasis et al. [32] (compliments of J. Breslow, Rockefeller University, New York). DNA (10 lug) was digested for 18 h with EcoRI and Hind111 and analyzed as described above.

2.4. Statistics

Statistical analysis of the data was done using the unpaired t-test and a value was considered statistically significant if P was determined to be < 0.05.

3. Results

Initial attempts to assess the DNAase I sensitivity of the GH/CS gene family members in pituitary cell nuclei made use of detection conditions essentially as used previously [33]. DNA was isolated from DNAase I treated nuclei and digested with BarnHI, elec- trophoresed, transferred to nitrocellulose and probed with proximal ( - 496/ + 1) GH-N 5’-flanking sequences after radiolabelling. The pattern obtained after autoradiography is shown in Fig. 1 (5’-flanking region probe). It suggested that the GH-N fragment was more sensitive to nuclease than the CS-B and CS-A/GH-V bands, since these bands were still visible in the presence of 2 units of DNAase I. However, it was difficult to resolve the CS-A and GH-V bands. In contrast, the CS-L as well as the GH-N fragments are clearly sensitive to DNAase I treatment, but the comparison was impeded by the much larger size of the CS-L band (8.4 versus 2.881.8 kb for the other GHjCS gene family members), due to the presence of coding sequences as well as 5’-flanking DNA in the CS-L fragment detected [l]. In light of these difficulties, the GH-N cDNA was used as an alternative probe to detect and compare the coding regions of each member of the GHjCS gene family. Under these conditions, a narrower range of fragment sizes was detected and each GHjCS gene fragment was resolved readily (Fig. 1, coding region probe). The pattern obtained suggested that all five GHjCS genes were equally sensitive to DNAase I digestion in pituitary cell nuclei.

We also assessed the DNAase I sensitivity of the GHjCS genes in term placental cell nuclei using the coding region probe. By contrast to the situation in the pituitary, the GH-N gene (and the GH-V gene) appeared less sensitive to nuclease treatment than the CS genes (Fig. 2). While human choriocarcinoma JEG-3

and BeWo cells display placental-specific expression of the GHjCS gene family [16], the level of synthesis is considerably lower than in term placenta [1,29,30]. We examined the DNAase I sensitivity of the GHjCS genes in JEG-3 and BeWo cells as well as the non placental (and non pituitary) human cervical carcinoma HeLa cell line. The pattern of nuclease sensitivity seen with nuclei isolated from both choriocarcinoma cell lines was similar to that observed from term placenta; the GH-N gene was less sensitive to nuclease than the CS genes (Fig. 2). This was distinct from the pattern observed with isolated HeLa cell nuclei, in which all of the GHjCS genes appeared equally sensitive to DNAase I treatment (Fig. 2). Thus, the pattern more closely re- sembled that seen with pituitary nuclei (Fig. l), however, much higher levels of nuclease appeared to be required in HeLa cells (Figs. 1 and 2).

The visual observations made from autoradiographs were confirmed by densitometry. The densities of indi-

units DNAase I

cs-L

GH-N CS-B

z-f; _

5 RankIng Ragkn Probe

units-I

8.4

2.8

2.3

:*il .

Coding Region Probe

Fig. 1. Autoradiographs showing DNAase 1 sensitivity of GH/CS gene fragments containing upstream flanking DNA or the entire coding region in pituitary cells. Nuclease sensitivity of major GH/‘CS gene-containing fragments were generated by BanzHI digestion of DNA (10 ,~g) that was isolated from pituitary cell nuclei treated with increasing amounts (units) of DNAase I as shown. Fragments were detected with radiolabelled 5’-flanking region ( - 496,’ + 1 GH-N) or coding region (GH-N cDNA) probes after DNA blotting. The idcn- tity of each CHKS gene-containing fragment and its approximate size is indicated.

158 B.E. Nickel, P.A. Cattini / Molecular and CeNular Endocrinology 118 (1996) 155-162

units DNAm+e I

0.0 0.1 0.2 0.5 1.0 2.0

CS-L CS-S CS-A OWN OH-V

units DNAam I 0.0 0.2 0.5 1.0 2-o 5.0

w CS-A GH-W

GH-V

BeWo

unmONA@wI

0.0 1.0 2.5 5.0 10.0 20.0

E Q a

Fig. 2. Autoradiographs showing the DNAase I sensitivity of fragments containing the entire coding regions of the GHjCS gene in placental cells. Nuclease sensitivity of major GH/CS gene-containing fragments were generated by BamHI digestion of DNA (IO pug) that was isolated from human term placenta, choriocarcinoma JEG-3 and BeWo cells, as well as non placental/non pituitary cervical carcinoma HeLa cell nuclei treated with increasing amounts (units) of DNAase I as shown. Fragments were detected with a radiolabelled coding region (GH-N cDNA) probe after DNA blotting. The identity of each GHjCS gene-containing fragment is shown and their approximate sizes are as indicated in Fig. I (coding region probe).

vidual gene-containing bands were plotted against increasing units of DNAase. The amount of DNAase I required to see a 50% decrease in band intensity was extrapolated from the graph to provide a value for sensitivity. To allow comparison, the value for the GH-N gene in each tissue/cell type was arbitrarily set to 1.0 and the sensitivity of the remaining genes was expressed relative to this value (Table 1). There was no significant difference between the nuclease sensitivity of the GH-N gene and the remaining members of the GHjCS gene family in pituitary or non pituitary/non placenta HeLa cell nuclei. In contrast, there was a significant increase in sensitivity of the CS genes relative to the GH-N gene in placenta (CS-L, P < 0.02; CS-A, P < 0.01; and CS-B, P < 0.001) as well as placental tumor JEG-3 (CS-L, P < 0.01; CS-A, P < 0.05; and CS-B, P < 0.03) and BeWo cells (CS-L, P < 0.005; CS-A, P < 0.02; and CS-B, P <c 0.002). Although just not significant (P = 0.054), there was also a slight but consistent increase in DNAase I sensitivity of the GH-V versus the GH-N gene in placenta. To obtain a value for an inactive gene in pituitary and placental tissue/cells, DNA from DNAase I-treated pituitary and placental cell nuclei was cut with EcoRI/ HindIII, probed with a fragment of the liver-specific apoA-I gene, and the DNAase I sensitivity assessed.

These results are also included in Table 1 for comparison. The nuclease sensitivity of the apoA-I was comparable in pituitary and placental cells, but the value was not drastically less than that determined for the GH-N gene in either of these cell types (Table 1).

4. Discussion

There is relatively little information on chromatin structure incorporating the GH/CS locus. A two domain model was suggested for position-dependent regulation of the GHjCS family using a digest and a probe to detect the upstream flanking DNA and proximal promoter regions [33]. The ‘left-hand’ domain containing GH-N was activated in the pituitary and the ‘right- hand’ domain containing CS-A and CS-B was activated in the placenta. This model predicted that the GH-V gene, which is flanked by CS-A and CS-B genes, would be expressed in the placenta and this was subsequently confirmed [ 1,12,13,15,16]. A further consequence of this model was that the CS-L gene, which lies immediately downstream of the GH-N gene, was likely to be expressed in the pituitary. Indeed, the CS-L gene was shown to be relatively more sensitive to DNAase I digestion than CS-B in this tissue [33], a result we

B.E. Nickel, P.A. Catfini 1 Molecular and Cellular Endocrinology 118 (1996) 155162 159

Table 1 DNAase I sensitivity of individual members of the GH/CS gene family relative to the GH-N gene

Tissue/cell GH-N cs-L CS-A GH-V CS-B apoA-I

Pituitary 1.00 i: 0.29 1.28 i 0.56 1.02 + 0.02 1.00 * 0.10 1.10 * 0.28 0.71 k 0.32 Placenta (term) 1.00 i: 0.12 5.75 & 1.17 2.41 + 0.27 1.48 * 0.13 3.43 * 0.25 0.66 & 0.63 JEG-3 (placenta) 1.00 i: 0.05 7.12 + 1.32 2.27 i 0.40 1.04 * 0.03 6.40 k 1.62 0.81 + 0.04 BeWo (placenta) 1.00 i: 0.11 2.82 & 0.27 1.67 f 0.13 1.11 f 0.04 2.07 + 0.09 0.71 * 0.30 HeLa (cervix) 1 .oo 1.74 1.30 1.14 1.48 3.50

A value for 50% digestion of each gene-containing fragment was extrapolated from plots of band density, determined from autoradiographs by densitometry, versus units of DNAase I. The sensitivity to DNAase I of each gene assessed is given relative to GH-N whose value has been arbitrarily set to 1.0, and is expressed as mean rt standard error of the mean from three determinations: except for HeLa cells where the mean is given from two experiments. Values greater or less than 1 .O indicate fold increase or decrease in nuclease sensitivity relative to the GH-N gene, respectively. A value for the liver-specific apoA-I gene in each tissue/cell is provided to facilitate comparisons of GHCS gene sensitivity between tissues.

confirmed in the present study using the same digestion and similar probe conditions (Fig. 1, 5’-flanking region probe). However, the available evidence indicates that the CS-L gene is transcriptionally active in the placenta [1,1.X16]. This discrepancy with the CS-L gene was explained by the relatively higher molecular weight (and hence ‘target size’) of the CS-L band investigated [33]. However, it is also worth noting that the CS-L fragment detected contains not only upstream flanking sequences but, unlike the other GHjCS family members, also the coding region [I]. The data on the chromatin configuration (open or closed) of the CS-A and GH-V genes was also compromised by the difficulty in resolving the bands containing these genes (Fig. 1, 5’-flanking region probe) [33]. In contrast, a comparison of the GH-N and CS-B genes was possible and suggests that the GH-N upstream flanking DNA, in- cluding the proximal promoter region, is slightly more sensitive than the equivalent sequences (2-3 kb) of the CS-B gene (Fig. 1, S-flanking region probe). It is possible that this reflects a less open chromatin configuration for the CS-B gene or, perhaps, local changes in chromatin structure in regions of these upstream sequences. However, the lack of DNAase I sensitivity is certainly not as evident as observed, for example, with the GH-N gene coding region in placental cells (Fig. 2). Given the variability and stated problems with assessing the placental members of the GHjCS gene family using the 5’-flanking region probe, we resolved to re- duce these difficulties by using a coding region probe in conjunction with a restriction endonuclease, that gener- ates GHjCS gene-containing fragments readily resolved by electrophoresis and with a more comparable range (6 kb) of sizes. Using these conditions, our data suggest that there is no significant difference in the sensitivity of the GH-N versus the CS-A, CS-B, CS-L and GH-V genes in pituitary cells (Fig. 1). The results also suggest that the chromatin containing the GHjCS genes is slightly more sensitive than that associated with the liver-specific apoA-I gene which, presumably, repre-

sents a closed and inactive gene (Table 1). This is consistent with previous observations which suggested that GHjCS chromatin was only slightly more sensitive to DNAase I digestion than inactive chromatin [33].

Previously, we proposed at least three possible expla- nations for the CS-A promoter activity in pituitary cells in vitro but not in vivo [28]. The first possibility was that it might reflect a difference in the properties of human versus rat GHF-l/Pit-l, in the sense that human GHF- 1 /Pit- 1, unlike rat GHF- 1 /Pit- 1, could distinguish the CS-A and GH-N genes by binding to, and/or inducing, the GH-N promoter [28]. This possibility was considered unlikely because of the high degree of nucleotide sequence similarity (96%) shared by human and rat GHF-l/Pit-l [34]. A second possibility was that the placental genes of the GHjCS locus in the pituitary possess a ‘closed’/inactive chromatin structure compared to the GH-N gene. Our data suggest that this is not the case, since both GH and CS genes are equally sensitive in pituitary cells (Fig. 1 and Table 1). By this criteria the GH/CS locus is ‘open’ and, thus, the GH-N, GH-V, CS-A, CS-B and CS-L genes appear to be available for transcription. Presumably, this accessibil- ity extends to GHF-l/Pit-l binding, since the GHF-I/ Pit-l elements are located in the first 100 (GH-N, CS-A, CS-B and CS-L) or 135 bp (GH-V) of 5’-flanking sequences [l&25,26], which corresponds to less DNA than is associated with a single nucleosome. This would, in turn, support a third possibility that a mechanism exists to specifically repress the CS-A promoter, as well as the CS-B, CS-L and GH-V genes, in the pituitary. In support of this possibility, we showed that sequences located 1.7-2.1 kb upstream of the placental (CS-A, -B, -L and GH-V) but not pituitary (GH-N) members of the GHjCS family, which contain two sites of protein-DNA interaction (PSF-A and PSF-B), can efficiently repress CS-A promoter activity in pituitary but not placental cells after gene transfer [28]. Of course our data do not exclude the possibility that regional condensation involving sequences related to the placen-

160 B.E. Nic,kel, P.A. Cuttini i Molecular and Cellular Endocrinology 118 (1996) 155-162

tal genes occurs in the pituitary, in spite of the locus being ‘open’.

By contrast to the results with pituitary chromatin, the placental CS-A, CS-B and CS-L genes were significantly more sensitive to DNAase I digestion than the pituitary GH-N gene in term placenta cell nuclei (Fig. 2 and Table 1). Thus, unlike the situation in the pituitary, these results would appear to suggest that nuclease sensitivity, and thus an open chromatin structure, correlates with transcriptional activity; the CS genes are sensitive and active while the GH-N gene is relatively less sensitive and inactive [16,29]. Although not quite significant, there was also a slight but consistent increase in the sensitivity of the GH-V gene relative to the GH-N gene (Table 1). This correlates with the low level of GH-V RNA expression in the placenta. This was shown by an analysis of cDNA clones from a term placenta library [I] and relative levels of GH-V versus CS transcripts indicated by reverse transcriptase-polymerase chain reaction (RT-PCR) with placental RNA [16]. It is possible that the slight increase in sensitivity of the GH-V gene relative to the GH-N gene reflects its location between the highly transcriptionally active CS- A and CS-B genes in the term placenta [l]. The location of DNAase I hypersensitive sites supports the idea that extensive alterations in chromatin structure of the GH/ CS locus are required for placenta-specific expression and regulation [30,3.5]. In a study using both placental and non placental cell lines, most of the hypersensitive sites identified were restricted to cells of placental origin 1301.

In comparison to term placenta, human choriocar- cinema BeWo and JEG-3 cells express low (detectable by RNA blotting) and extremely low (detectable by RT-PCR) levels of CS RNA, respectively [16,29,30]. In spite of this, a similar pattern of DNAase I sensitivity of the CS genes relative to GH-N gene was observed in all of these placental cell types and was distinct from the pattern observed with human cervical carcinoma HeLa cell nuclei (Fig. 2 and Table 1). Much higher levels of nuclease were required for equivalent digestion in HeLa cells, which is consistent with the presence of a ‘closed’ and inactive GHjCS locus. A correlation between nuclease sensitivity and at least low levels of transcriptional activity was maintained in both JEG-3 and BeWo cell nuclei, since the CS genes were significantly more sensitive than the GH-N gene (Table 1). These data also suggest that an ‘open’ chromatin configuration for the GHjCS genes is not synonymous with high levels of transcriptional activity. The reason for this is unclear. One possible explanation is methylation. However, a study using the methylation sensitive isoschizomers MspI and HpaII failed to detect any differences consistent with the patterns of tissue-specific expression of the GHjCS genes [36]. The results did not exclude a role for methylation in the developmental

regulation of the placental GHjCS genes. Another possible explanation is the absence, low levels or modifica- tion of necessary transcription factors. This is not supported by transient transfection experiments. Hy- brid reporter genes directed by CS promoter sequences and/or CS enhancer elements are highly expressed even in JEG-3 cells [19,22]. Furthermore, there was no difference in the pattern of placenta-specific hypersensitive sites detected in the GHjCS locus in nuclei isolated from JEG-3 versus BeWo cells [30], despite the distinc- tive levels of CS RNA expression in these cell lines [29]. Whatever the mechanism (methylation and/or a post- transcriptional process like RNA stability) responsible for the difference in CS production by term placenta versus choriocarcinoma cells, it would likely reflect the properties or developmental stage of the trophoblast concerned. During gestation, the GH-V/CS-producing villus syncytiotrophoblast, a non dividing and multinu- cleated structure, develops via fusion of transitioning cytotrophoblasts [37-401. Choriocarcinoma cells are of- ten characterized as cytotrophoblast-like, being highly proliferative and largely mononuclear [41,42].

In summary, the failure of the CS genes to be expressed in pituitary cells does not correlate with a less nuclease sensitive/closed chromatin structure relative to that assessed for the GH-N gene. However, our data are consistent with an increased sensitivity of the CS genes to DNAase I treatment thus, a role for a more open placenta-specific expression.

Acknowledgements

in placental cells and, chromatin structure in

The authors are grateful to Drs. I. Worsley and R.P.C. Shiu at the University of Manitoba for human pituitary tissue, and Dr. Breslow at the Rockefeller University for the plasmid (pSV2AI) containing a 2.2 kb Pst 1 fragment of the human apolipoprotein A-I gene. We would also like to thank Dr. R.M. Surabhi for critical evaluation of this manuscript. This work was supported by a grant from the Medical Research Council of Canada (MT-10853). PAC is the recipient of a Medical Research Council of Canada Scientist Award.

References

111

PI

Chen, E.Y., Liao, Y-C., Smith, D.H., Barrera-Saldana, H.A., Gelinas, R.E. and Seeburg, P.H. (1989) The growth hormone locus: nucleotide sequence, biology and evolution. Genomics 4, 479 - 497. Freemark, M. and Comer, M. (1989) Purification of a distinct placental lactogen receptor, a new member of the growth hormone/prolactin receptor family. J. Clin. Invest. 83, 883 - 889.

[31

[41

B.E. Nickel, P.A. Cattini i Molecular and Cellular Endocritzology 118 (1996) 155-162

Urbanek, M., MacLeod, J.N., Cooke, N.E. and Liebhaber, S.A. (1992) Expression of a human growth hormone (hGH) receptor isoform is predicted by tissue-specific alternative splicing of exon 3 of the hGH receptor gene transcript. Mol. En- docrinol. 6, 2799287. Freemark, M., Kirk, K., Pihoker, C., Robertson, M.C., Shiu. R.P.C. and Driscoll, P. (1993) Pregnancy lactogens in the rat conceptus and fetus: circulating levels, distribution of binding, and expression of receptor messenger ribonucleic acid. En- docrinology 133, 1830- 1842.

[5] Urbanek, M., Russell, J.E., Cooke, N.E. and Liebhaber, S.A. (1993) Functional characterization of the alternatively spliced, placental human growth hormone receptor. J. Biol. Chem. 268, 19025- 19032.

[61

[71

PI

[91

UOI

[Ill

[I31

u41

u51

V61

1171

[181

Handwerger. S. (1991) Clinical counterpoint: the physiology of placental lactogen in human pregnancy. Endocr. Rev. 12, 329- 336. Brelje. T.C., Scharp, D.W., Lacy, P.E., Ogren, L., Talamantes, F., Robertson, M.. Friesen, H.G. and Sorenson, R.L. (1993) Etfect of homologous placental lactogens, prolactins. and growth hormones on islet B-cell division and insulin secretion in rat, mouse, and human islets: implication for placental lactogen regulation of islet function during pregnancy. Endocrinol- ogy 132, 879m 887. Lefevre, C., Imagawa, M., Dana, Grindlay, J.. Bodner, M. and Karin, M. (1987) Tissue specific expression of the human growth hormone gene is conferred in part by the binding of a specific trarzs-acting factor. EMBO J. 6, 971-981. Bodner, M., Castrillo, J-L., Theill, L.E., Deerinck, T., Ellisman, M. and Karin. M. (1988) The pituitary-specific transcription factor GHF-I is a homeobox-containing protein. Cell 55, 505 518. Ingraham, H.A., Chen, R., Mangalam, H.J., Elsholtz, H.P., Flynn, SE., Lin, C.R., Simmons, D.M., Swanson, L. and Rosenfeld. M.G. (1988) A tissue-specific transcription factor containing a homeodomain specifies a pituitary phenotype. Cell 55, 519-529. McWilliams, D. and Boime, 1. (1980) Cytological localization of placental lactogen messenger ribonucleic acid in syncytiotrophoblast layers of human placenta. Endocrinology 107, 761 765. Frankenne, F., Rentier-Delrue, F., Scippo, M-L., Martial, J. and Hennen, G. (1987) Expression of the growth hormone variant gene in human placenta. J. Clin. Endocrinol. Metab. 64, 6355637. Liebhaber, S.A., Urbanek, M., Ray, J., Tuan, R.S. and Cooke, N.E. (1989) Characterization and histologic localization of human growth hormone-variant gene expression in the placenta. J. Clin. Invest. 83, 198551991. Klassen. M.E.. Nachtigal, M.W. and Cattini, P.A. (1989) Hu- man chorionic somatomammotropin gene expression in primary placental cell cultures. Placenta 10, 321-329. MacLeod, J.N., Lee, A.K., Liebhaber, S.A. and Cooke, N.E. (1992) Developmental regulation and alternative splicing of the placentally expressed transcripts from the human growth hormone gene cluster. J. Biol. Chem. 267, 14219- 14226. Lytras. A., Bock M E.. Dodd, J.G. and Cattini, P.A. (1994) Detection of placental growth hormone variant and chorionic somatomammotropin RNA expression in human trophoblastic neoplasms by reverse transcriptase-polymerase chain reaction. Endocrinology 134, 2461-2467. Fitzpatrick, S.L., Walker, W.H. and Saunders, G.F. (1990) DNA sequences involved in the transcriptional activation of a human placental lactogen gene. Mol. Endocrinol. 4, 181551826. Walker, W.H., Fitzpatrick. S.L. and Saunders, G.F. (1990) Human placental lactogen transcription enhancer. J. Biol. Chem. 265, 12940 - 12948.

]l91

]201

PII

PI

~231

~241

P51

161

Nachtigal, M.W., Nickel, B.E., Bock, M.E., Lytras, A., Pa- sumarthi, S.K.B. and Cattini, P.A. (1993) Expression of human placental hormone genes in growing and growth-inhibited tran- sitional trophoblasts. In: Trophoblast Cells. Pathways for Ma- ternal-Embryonic Communication (Soares, M., Talamantes, F. and Handwerger, S., eds.), pp. 2844289, Springer-Verlag. New York. Jacquemin, P., Oury, C., Peers, B., Morin. A., Belayew, A. and Martial, J.A. (1994) Characterization of a strong tissue-specific enhancer downstream from the three human genes encoding placental lactogen. Mol. Cell. Biol. 14, 93 103. Jiang, S.-W. and Eberhardt. N.L. (1994) The human chorionic somatomammotropin gene enhancer is composed of multiple DNA elements that are homologous to several SV40 enhan- sons. J. Biol. Chem. 269, 10384410392. Lytras, A. and Cattini. P.A. (1994) Human chorionic somatomammotropin gene enhancer activity is dependent on the blockade of a repressor mechanism. Mol. Endocrinol. 8. 478- 489. Jiang. S.-W. and Eberhardt, N.L. (1995) Involvement of a protein distinct from transcription enhancer factor-l (TEF-I) in mediating human chorionic somatomammotropin gene enhancer function through the CT-IIC enhanser in choriocar- cinema and COS cells. J. Biol. Chem. 270. 13906- 13915. Jiang, S.-W., Shepard, A.R. and Eberhardt, N.L. (1995) An initiator element is required for maximal human chorionic somatomammotropin gene promoter and enhancer function. J. Biol. Chem. 270, 3683 3692. Lemaigre, F.P.. Peers, B., Lafontaine D.A., Mathy-Hartert. M., Rousseau, G.G.. Belayew!, A. and Martial, J.A. (1989) Pituitary-specific factor binding to the human prolactin, growth hormone. and placental lactogen genes. DNA 8, 1499159.

[26] Nickel, B.E.. Nachtigal, M.W., Bock. M.E. and Cattini P.A. (199 I) Dilferential binding of rat pituitary-specific nuclear factors to the 5’-flanking region of pituitary and placental members of the human growth hormone gene family. Mol. (‘ell. Biochem. 106. I81 187.

~271

PI

~291

[301

[311

~321

[331

1341

Nachtigal, M.W., Nickel, B.E., Klassen. M.E.. Zhang, W.. Eberhardt, N.L., and Cattini. P.A. (1989) Human chorionic somatomammotropin and growth hormone gene expression in rat pituitary tumour cells is dependent on proximal promoter sequences. Nucleic Acids Res 17. 4327 -4337. Nachtigal, M.W., Nickel, B.E. and Cattini. P.A. (1993) Pitu- itary ~ specific repression of placental members of the human growth hormone gene family. J. Biol. Chem. 268, 847338479. Nickel, B.E. and Cattini, P.A. (1991) Tissue-specific expression and thyroid hormone regulation of the endogenous placental growth hormone variant and chorionic somatomammotropin genes in a human choriocdrcinonia cell line. Endocrinology 128. 2353 2359. Jimenez, G., Ford, A.M.. Enver. T. and Boronat, A. (1993) Multiple changes in chromatin structure precede the transcriptional activation of the human growth hormone locus in placental cells. Mol. Cell. Endocrinol. 96, 53- 60. Cattini, P.A., Anderson. T., Baxter. J.D., Mellon, P. and Eher- hardt, N.L. (1986) The human growth hormone gene is nega- tively regulated by triiodothyronine vvhen transfected into rat anterior pituitary cells. J. Biol. Chem. 261, 13367- 13372. Karathanasis. S.K., Zannis, V.I. and Breslou, J.L. (1983) Isola- tion and characterization of the human apolipoprotein A-I gene. Proc. Natl. Acad. Sci. USA 80, 6147 6151. Barsh. G.S.. Seeburg, P.H. and Gelinas. R.E. (1983) The human growth hormone gene family: structure and evolution of the chromosomal locus. Nucleic Acids Res. I I, 3939-3958. Lew. A.M and Elsholtz. H.P. (1991) Cloning of the human cDNA for transcription factor Pit-l. Nucleic Acids Res. 19. 6329.

162 B.E. Nickel, P.A. Cattini / Molecular and Cellular Endocrinology 118 (1996) 155-162

[35] Jones, B.K., Monks, B.R., Liebhaber, S.A. and Cooke, N.E. (1995) The human growth hormone gene is regulated by a multicomponent locus control region. Mol. Cell. Biol. 15, 7010- 7021.

[36] Hjelle, B.L., Phillips, J.A., and Seeburg, P.H. (1982) Relative levels of methylation in human growth hormone and chorionic somatomammotropin genes in expressing and non-expressing tissues. Nucleic Acids Res. 10, 345993474.

[37] Hoshina, M., Hussa, R., Pattillo, R. and Boime, I. (1983) Cytological distribution of chorionic gonadotropin subunit and placental lactogen messenger RNA in neoplasms derived from human placenta. .I. Cell Biol. 97, 1200%1206.

[38] Boime, I., Boothby, M., Darnell, R.B. and Pohcastro, P. (1986) Structure and expression of human placental hormone genes. In: Molecular and Cellular Aspects of Reproduction (Dhindsa, D.S.

and Bahl, O.P., eds.), pp. 267-280, Plenum Press, New York. [39] Ohlsson, R. (1989) Growth factors, protooncogenes and human

placental development. Cell Differ. Dev. 28, 1-16. [40] Ringler, G.E. and Strauss III, J.F. (1990) In vitro systems for the

study of human placental endocrine function, Endocr. Rev. 11, 105-123.

[41] Nachtigal, M.W., Bock, M.E., Kowaluk, B.A. and Cattini, P.A. (1992) Chorionic gonadotropin and c-rnyc expression in growing and growth-inhibited intermediate trophoblasts. Placenta 13, 371-383.

[42] Nickel, B.E., Bock, M.E., Nachtigal, M.W. and Cattini, P.A. (1993) Differential expression of human placental growth hormone variant and chorionic somatomammotropin genes in choriocarcinoma cells treated with methotrexate. Mol. Cell. Endocrinol. 91, 159- 166.

Documents

Nuclease sensitivity of the human growth hormone-chorionic somatomammotropin locus in pituitary and placenta suggest different mechanisms for tissue-specific regulation