Introduction
HSD3B gene
The enzyme 3β-hydroxysteroid dehydrogenase/isomerase (3β-HSD) regulates an essential step in the biosynthesis of steroid hormones, including mineralocorticoids,
glucocorticoids, and sex steroids (STRAUSS et al. 1996). Multiple isoforms of 3β-HSD have been identified in humans and rodents. These isoforms are encoded by distinct genes and are expressed in a tissue-specific pattern (SIMARD et al. 2005). Humans possess two 3β-HSD isoforms, type I (HSD3B1) and type II (HSD3B2), which share 93.5% identity in their amino acid sequences (RHEAUME et al. 1991). HSD3B2 is exclusively expressed in the classic steroidogenic tissues such as the adrenal gland, ovary, and testis, whereas HSD3B1 is expressed in the placenta and peripheral tissues, including the mammary gland, prostate, and skin (RHEAUME et al. 1991; GINGRAS et al.
1999; GINGRAS AND SIMARD 1999).
Physiological function of HSD3B in placenta
Progesterone is essential for the establishment and maintenance of pregnancy in
mammals. The biosynthesis of progesterone from cholesterol is catalyzed by two steroidogenic enzymes: cytochrome P450 side-chain cleavage enzyme (P450scc;
encoded by CYP11A1) and 3β-HSD. During the early stages of human pregnancy, the corpus luteum of the ovary is active in secreting progesterone for the implantation and maintenance of pregnancy (STRAUSS et al. 1996). After about 8–10 weeks, large amounts of progesterone are produced by the placenta to sustain the pregnancy. The expression of HSD3B1 in the placenta is involved in the biosynthesis of progesterone;
thus, it is needed for successful pregnancy. Little is known about the molecular mechanisms that regulate HSD3B1 expression in the placenta. Previous studies have demonstrated that stimulation with cAMP and PMA (phorbol 12-myristate 13-acetate) can increase HSD3B1 mRNA in human choriocarcinoma JEG-3 cells (TREMBLAY AND
BEAUDOIN 1993); however, no response elements have been identified in the promoter of the HSD3B1 gene. An initial study investigating the regulatory sequences of the HSD3B1 promoter has demonstrated that some cis-acting elements located in the first
intron regulate the basal expression of the HSD3B1 gene (GUERIN et al. 1995).
Furthermore, it has been shown that a distal cis-regulatory element located at -2570/-2518 mediates the specific expression of the HSD3B1 gene in the placenta (PENG et al.
2004). This element contains a specific binding site for transcription enhancer factor-5, which is highly expressed in the human placenta, and an overlapping GATA-binding
site for a GATA-like protein.
GATA2/3
GATA transcription factors have conserved zinc finger domains that bind to the consensus DNA motif [(A/T)GATA(A/G)] to regulate the transcription of downstream genes. In mammals, six GATA factors (GATA1~6) are expressed in a tissue-specific manner and involved in the regulation of cell-specific gene expression. GATA1, GATA2 and GATA3 regulate development and differentiation of hematopoietic cell lineage (TSAI et al. 1994; PANDOLFI et al. 1995; FUJIWARA et al. 1996), while GATA4, GATA5 and GATA6 are involved in cardiac development and differentiation of endodermal lineage (MOLKENTIN et al. 1997; MORRISEY et al. 1998; MOLKENTIN et al.
2000).
Earlier studies have demonstrated that GATA2 and GATA3 are the isoforms predominantly expressed in placental trophoblast cells, regulating the expression of a number of genes such as placental lactogen I, proliferin, and syncytin (MA et al. 1997;
CHENG AND HANDWERGER 2005; BAI et al. 2013). Placentas lacking GATA2 or GATA3 both showed lower placental lactogen I expression, and greater reduction in proliferin expression in GATA2 mutant placenta, which results in less neovascularization (MA et al. 1997). Thus, they are both involved in placental development and synthesis of
placental hormones.
Aim of this study
HSD3B1 is essential for progesterone production to sustain pregnancy in
placenta. Little is known about the regulation of HSD3B1 transcription in the placenta. It has been demonstrated that a distal cisregulatory element located at
-2570/-2518 mediates the specific expression of the HSD3B1 gene in the placenta (PENG et al. 2004). This region contains binding sequence for transcription enhancer factor-5 and GATA-like protein. Nevertheless, the regulation of HSD3B1
transcription is less characterized. In this study, we sought to identify the novel cis-element(s) essential for HSD3B1 transcription in the choriocarcinoma JEG-3 cells.
Result
Promoter activity of human HSD3B1 gene in JEG-3 cells
To identify novel regulatory elements involved in the placental expression of human HSD3B1 gene, the 5′-flanking sequence from -2226 to +337 was cloned upstream of the luciferase reporter gene. In addition, the luciferase reporter construct containing the promoter fragment (-1859/+56) from human HSD3B2 gene was also generated. In human JEG-3 trophoblast cells, the HSD3B1 -2226/+337 fragment demonstrated a marked increase in transcriptional activity (26 fold) over the basic vector, whereas the -1859/+56 fragment of the HSD3B2 promoter did not exhibit luciferase activity (Fig. 19A), indicating the specific expression of HSD3B1 in the placenta. This finding suggests that sequences between -2226 to +337 contain functional elements for the regulation of HSD3B1 expression in JEG-3 cells. To further localize the regulatory cis-elements in this region, a series of deletions were introduced in the -2226/+337 sequence. As shown in Fig. 19B, the -238/+337 construct had the highest promoter activity in JEG-3 cells. Furthermore, the deletion of 262 bp from -500 to -238 resulted in a drastic increase in transcriptional activity by 4.5 fold, suggesting the presence of negative regulatory elements in this region. These data demonstrate that the -238/+337 basal promoter region contains functional cis-elements responsible for
Enhancement of HSD3B1 promoter activity by GATA2 and GATA3
A database search of the -238/+337 promoter region identified two putative GATA-binding sites at -106/-99 and -52/-45, which are referred to as Gu and Gp, respectively. To determine the role of these GATA-binding sites in the regulation of HSD3B1 transcription, mutated GATAbinding sequences were introduced into the
-238/+337 promoter construct (Fig. 20A). As shown in Fig. 20B, mutations of Gu and Gp significantly reduced promoter activity by 86% and 45%, respectively, in JEG-3 cells. In addition, the simultaneous mutation of both GATA sites resulted in a 92%
reduction in promoter activity.
GATA2 and GATA3 have previously been reported to regulate the expression of a number of genes in trophoblast cells and placental tissues (BAI et al. 2013). To determine whether GATA proteins affect HSD3B1 promoter activity, the -238/Luc construct was co-transfected with GATA expression plasmids in HeLa cells. As shown in Fig. 21A, co-transfection of the -238/Luc construct with GATA2 or GATA3 significantly increased transcriptional activity, whereas co-transfection with GATA4 did not affect reporter activity. Mutation of the Gu or Gp GATA-binding site resulted in the loss of GATA2- and GATA3-stimulated promoter activity (Fig. 21C and 21D).
Expression of HSD3B, GATA 2, and GATA3 in the placenta
As shown in Fig. 22A, we confirmed the expression of HSD3B protein in JEG-3 cells by western blot analysis. HSD3B2 is the isoform expressed in human adrenals (RHEAUME et al. 1991); thus, a high amount of HSD3B protein was detected in H295R human adrenocortical carcinoma cells (Fig.22A). In addition, Fig. 22A shows that GATA2 and GATA3 proteins were present at high levels in JEG-3 cells but at low to undetectable levels in H295R, HepG2, HeLa, or HEK293T cells. This is consistent with earlier studies showing the high expression of GATA2 and GATA3 in JEG-3 cells but not in H295R cells (STEGER et al. 1994; FLUCK AND MILLER 2004). Cellular fractionation experiments demonstrated that GATA2 and GATA3 proteins were exclusively present in the nuclear fraction of JEG-3 cells (Fig. 22B).
The expressions of HSD3B, GATA 2, and GATA3 were examined in placental tissues by western blot analysis. Mouse Hsd3b6 is orthologous to human HSD3B1, which is the only isoform expressed in the placenta and skin (ABBASZADE et al. 1997).
Consistent with previous studies, we found that Hsd3b protein was highly expressed in mouse placenta on E10.5 and gradually decreased from E11.5–E12.5 (Fig. 22C).
Similar to Hsd3b, protein levels of GATA2 and GATA3 were high in mouse placenta on E10.5 and declined thereafter. However, GATA3 protein level was markedly decreased on E11.5. The protein expressions of HSD3B and GATA2 were also high in
human placental tissues; conversely, GATA3 protein was detected at a very low level (Fig. 22C).
Analysis of GATA2 and GATA3 binding to the GATA elements
We performed electrophoretic mobility shift assay (EMSA) to examine the potential nuclear proteins involved in binding to the GATA-binding sites. Synthetic oligonucleotides containing the GATA site Gu (-115/-86) and Gp (-65/-36) were used as probes. Nuclear extracts from JEG-3 cells formed DNA-protein complexes with Gu probe (Fig. 23A, lane 2) that were competed by a 100-fold excess of unlabelled Gu (lane 3) or Gp (lane 4) probe. Similarly, protein complexes formed from JEG-3 nuclear extracts and Gp probe were competed by an excess of unlabelled Gu (lane 10) or Gp (lane 11) probe. Probes with mutation of the GATA-binding sequences (mu and mp) shown in Fig. 10 did not compete with the complexes (lane 5, 6, 12, and 13), indicating that the GATA sequence is required for protein binding. With Gu or Gp as the probe, a supershift was observed by the antiserum to GATA2 or GATA3, but not by the control serum (Fig. 23B). This suggests the involvement of GATA2 and GATA3 in the DNA-protein complex observed in EMSA. Furthermore, we used ChIP assays to confirm the presence of GATA2 or GATA3 at the endogenous promoter. As shown in Fig. 23C, GATA2 occupancy was observed in the GATA-binding region of the HDS3B1 promoter,
whereas GATA3 was not found to be associated with this region.
Effects of GATA2 or GATA3 knockdown on HSD3B1 expression
To determine the roles of GATA2 and GATA3 in regulating HSD3B1 expression, we used shRNA to knock down GATA2 or GATA3 in JEG-3 cells. As shown in Fig. 6A and B, knockdown of GATA2 markedly reduced HSD3B1 mRNA levels by ~80% and protein levels by ~70% in JEG-3 cells. In contrast, we observed that GATA3 knockdown resulted in a 2-fold increase in the mRNA and protein levels of HSD3B1 (Fig. 24C and 24D). The results suggest that GATA2 and GATA3 may play distinct functional roles in HSD3B1 transcription.
Discussion
Human HSD3B1 is involved in the production of placental progesterone for maintenance of pregnancy. Although several cis-acting elements have been implicated in the transcription of HSD3B1, the transcriptional mechanisms that regulate its basal expression in the placenta remain unknown. Here, we found that the proximal -238/+337 sequence of HSD3B1 exhibited high transcriptional activity in human choriocarcinoma JEG3 cells. In this region, two GATA elements were identified at -106/-99 (Gu) and -52/-45 (Gp). Mutation of the GATA-binding sites at Gu or Gp greatly reduced promoter activity in JEG-3 cells. Our findings in this study strongly suggest that the proximal GATA motifs are important for trophoblast-specific expression of the HSD3B1 gene, and GATA2 and GATA3 play a role in the regulation of HSD3B1.
Our study showed that both GATA2 and GATA3 proteins were abundantly expressed in JEG-3 trophoblast cells and placental tissues (Fig. 22). GATA2 and GATA3 exhibited similar temporal expression patterns to Hsd3b in mouse placenta.
Furthermore, overexpression of GATA2 or GATA3 significantly stimulated HSD3B1 promoter activity in non-trophoblast cells (HeLa cells), whereas GATA4 had no effect on HSD3B1 promoter activity (Fig. 21). Mutations of GATA-binding sites significantly impaired GATA2- or GATA3-induced HSD3B1 promoter activity. In addition, EMSA
experiments demonstrated the specific binding of GATA2 and GATA3 proteins to the GATA sequences at -106/-99 and -52/-45. These data suggest that GATA2 and GATA3 can regulate the transcription of HSD3B1 through the GATA-binding sequences.
In EMSA, we observed that the two GATA probes (Gu and Gp) seemed to bind with a stronger affinity to GATA2 than to GATA3 (Fig. 23B). ChIP-qPCR assays further verified the binding of GATA2 at the GATA-binding regions of the HSD3B1 promoter in JEG-3 cells. GATA2 knockdown in JEG-3 cells resulted in a sharp reduction in the expression of HSD3B. Together, these findings reveal that GATA2 is the predominant protein recruited to the GATA motifs at -106/-45 to stimulate HSD3B1 gene expression in JEG-3 cells. In addition to the placenta, the expression of HSD3B1 has been identified in peripheral tissues including the mammary gland, prostate, and several human cancer cell lines (RHEAUME et al. 1991; GINGRAS et al. 1999; GINGRAS AND
SIMARD 1999). The LNCaP human prostate cancer cells have been shown to express low levels of endogenous HSD3B1 (GINGRAS AND SIMARD 1999; CHANG et al. 2013).
GATA2 ChIP-seq data derived from LNCaP cells are available in the GEO (Gene Expression Omnibus) database under accession number GSE69043 (ZHAO et al. 2016).
Analysis of the ChIP-seq datasets for LNCaP cells revealed GATA2 ChIP enrichment in the regulatory regions of the HSD3B1 locus between -267 and +104 (Fig. 25),
covering the two GATA-binding sites identified in this study. This is consistent with our ChIP-qPCR results of GATA2 in JEG-3, strongly indicating that GATA2 may be necessary for the basal transcription of the HSD3B1 gene.
In contrast to GATA2, the levels of HSD3B1 protein and mRNA were increased by GATA3 knockdown in JEG-3 cells, indicating that GATA3 may negatively regulate the expression of HSD3B1. EMSA experiments showed that both GATA2 and GATA3 were able to bind to the GATA elements at -106/-45 in vitro (Fig. 23B), suggesting that GATA2 and GATA3 may occupy identical GATA-binding regions of the HSD3B1 locus but exert distinct activities in HSD3B1 expression. Similar findings demonstrating that GATA factors exhibit different functions through shared chromatin sites were also observed in the transcriptional regulation of the Gata2 gene. During mouse erythroid differentiation, GATA1 and GATA2 directly regulate Gata2 transcription. GATA1 represses Gata2 transcription by displacement of GATA2 from certain GATA-binding sites in the upstream region (-77, -3.9, -2.8, and -1.8 kb) and an intron region (+9.5 kb) of the Gata2 locus (GRASS et al. 2003; MARTOWICZ et al. 2005; GRASS et al. 2006).
GATA2 is transcriptionally active in these regions, indicating that it positively autoregulates transcription. The process of “GATA switch” was also found in the development of the placenta. GATA3 was reported to repress the Gata2 gene by
occupying the -3.9 kb and +9.5 kb regions of the Gata2 locus in undifferentiated trophoblast stem cells (RAY et al. 2009). The displacement of GATA3 by GATA2 at these regions induces Gata2 transcription during trophoblast differentiation (RAY et al.
2009). If a similar GATA switch occurs in the HSD3B1 locus, the concentration of GATA2 and GATA3 may be an important determinant in HSD3B1 gene regulation. In the present study, we showed that GATA3 protein was scarce in human placental tissues (Fig. 22C). This finding is consistent with previous study, which indicated that GATA3 expression was strong in immature placenta and less expression in mature placenta (MIRKOVIC et al. 2015). In contrast, GATA2 protein was present at high levels, suggesting the binding of GATA2 to the GATA sites in the promoter to induce HSD3B1 transcription.
Although both GATA3 and GATA2 were abundant in JEG-3 cells, GATA3 binding to the GATA sites at -106/-45 was not observed in ChIP assays, indicating that this region may not be involved in GATA3-mediated HSD3B1 repression. Piao et al.
(PIAO et al. 1997) identified a GATA3-binding motif in a silencer element that represses the transcriptional activity of the HSD17B1 promoter in choriocarcinoma cells. By using a transcription factor binding site search program, we identified another putative GATA element present between -558 and -547 of the HSD3B1 promoter (Fig. 26).
Mutation of the GATA sequence in this region increased the promoter activity in JEG-3 cells, indicating that this motif is a potential silencer of the HSDJEG-3B1 gene. However, the specific association of GATA proteins to this GATA element was not detectable by EMSA analysis. Whether other GATA3-binding sites are present in the HSD3B1 locus and participate in the regulation of HSD3B1 transcription remain to be determined.
The expressions of CYP11A1 and HSD3B1 are required for the synthesis of progesterone in the human placenta. A previous study identified a GATA element at -475/-470 that is required for the activation of the Cyp11a1 promoter in mouse trophoblast giant cells (SHER et al. 2007). The study demonstrated via EMSA that GATA2 is the predominant protein bound to this GATA site (SHER et al. 2007). In this study, we found that two GATA-binding motifs located at -106/-45 are essential for the activation of the HSD3B1 promoter in human trophoblast cells. Our data demonstrate that GATA2 binds to the GATA motifs and acts as the critical transcriptional activator of the HSD3B1 gene in trophoblast cells. In addition, GATA3 seems to exert suppressive effects on HSD3B1 gene transcription. Together, GATA2 and GATA3 may play a critical role in the regulation of placental progesterone production.
In this study, we searched for novel regulatory elements within the 2.2 kb
proximal promoter of HSD3B1. We identified two GATA elements located at -106/-99 and -52/-45 that are essential for the functional activity of the HSD3B1 promoter in
JEG-3 cells. Our data revealed that GATA2 and GATA3 are predominant proteins bound to the GATA elements and may play distinct functional roles in the expression of HSD3B1 in the placenta.
Figure 18. Transcriptional activity of the human HSD3B1 promoter
(A) The human HSD3B1 (B1: -2226/+337) or HSD3B2 (B2: -1859/+56) luciferase constructs were generated. (B) A series of 5′-deleted HSD3B1 promoter-luciferase constructs were transiently transfected into JEG-3 cells. Cells were harvested after 24 h of transfection and assayed for luciferase activity. Results are presented as the activity relative to the empty vector (n = 3). Bars with different letters are significantly different from each other (P < 0.05).
Figure 19. Enhancement of the transcriptional activity of the HSD3B1 promoter by two putative GATA sites
(A) Schematic of the promoter region in the HSD3B1 gene. “+1” is the transcription start site indicated by the arrow. Two putative GATA-binding sites are indicated as Gu and Gp. Their sequences are shown with mutated nucleotides in lowercase letters. (B) JEG-3 cells were transiently transfected with the -238/+337 HSD3B1 promoter-luciferase construct containing wild-type or mutated GATA sites (either or both sites).
Cells were harvested after 24 h of transfection and assayed for luciferase activity.
Results are presented as the activity relative to the wild-type construct (n = 4). Bars with different letters are significantly different from each other (P < 0.05).
Figure 20. Effects of GATA factors on the transcriptional activity of the HSD3B1 promoter
(A) HSD3B1 promoter-luciferase constructs containing wild-type or mutated GATA sites were co-transfected into HeLa cells with the expression vector for GATA2, GATA3, or GATA4 as indicated in (A), (C), and (D), respectively. Cells were harvested after 24 h of transfection and assayed for luciferase activity. Results are presented as the activity relative to the empty vector. Data represent mean ± SEM. *P
< 0.05, **P < 0.01, ***P < 0.001 compared with the vector control. (B) Expressions of GATA proteins were detected by western blot analysis with GATA3 and anti-Myc for GATA2 and GATA4. N, non-transfection; V, expression vector.
Figure 21. Expression of HSD3B, GATA2, and GATA3 in cell lines and tissues (A) Cell lysates from the indicated cell lines were subjected to western blot analysis.
(B) Nuclear and cytoplasmic fractions from JEG-3 cells were examined by western blot analysis. The markers used were PARP for the nucleus and GAPDH for the cytoplasm. (C) Protein extracts from two human placental tissues, mouse placenta of E10.5–E12.5, and mouse lung were analyzed by western blot analysis. GAPDH was used as a loading control.
Figure 22. Analysis of GATA2 and GATA3 binding to sites at -106/-45 in the HSD3B1 promoter
EMSA was performed with JEG-3 nuclear extract using two radiolabelled probes containing the putative GATA-binding site Gu (-115/-86) or Gp (-65/-36). (A) Competition experiments were performed by incubation with a 100-fold excess of the
unlabelled oligonucleotides Gu, Gp, mutated Gu (mu), mutated Gp (mp), or non-specific sequence (NS). (B) The anti-GATA2 or anti-GATA3 antibody was added to the binding reaction, and the respective isotype antibody was used as a control. Supershift bands are indicated by the arrowhead. (C) JEG-3 cells were subjected to ChIP assay with the indicated antibodies or control IgG and analyzed by qPCR using a primer set covering the Gp site. Representative results of two independent experiments are shown.
Data are mean ± SEM of triplicate qPCR analyses.
Figure 23. Effects of GATA2 or GATA3 knockdown on HSD3B1 expression JEG-3 cells were transfected with shRNA specific for GATA2, GATA3, or LacZ
Figure 23. Effects of GATA2 or GATA3 knockdown on HSD3B1 expression JEG-3 cells were transfected with shRNA specific for GATA2, GATA3, or LacZ