- Email: [email protected]
Activity and response to serum of the mammalian thyroid hormone deiodinase 3 gene promoter: identification of a conserved enhancer
Molecular and Cellular Endocrinology 206 (2003) 23 /32 www.elsevier.com/locate/mce
Activity and response to serum of the mammalian thyroid hormone deiodinase 3 gene promoter: identification of a conserved enhancer Arturo Hernandez *, Donald L. St. Germain Departments of Medicine and Physiology, Dartmouth Medical School, Lebanon, NH, USA Received 23 April 2003; accepted 25 June 2003
Abstract Dio3 is an imprinted gene that codes for the type 3 deiodinase (D3), a conserved selenocysteine-containing enzyme that degrades thyroid hormones (TH) into inactive metabolites. D3 is partially responsible for the very low TH levels that are found in utero as it is highly expressed in mammalian uterus, placenta, and fetal tissues. For this reason, it is presumed that D3 protects the developing fetus from high, adult-like levels of TH. To further understand the regulation of D3 expression, we analyzed the promoter activity of 6 kb of 5?-flanking regions of the human and mouse Dio3 , by transient transfection studies. Both mouse and human Dio3 promoters are markedly responsive to serum and, to a lesser extent, to phorbol esters and fibroblast growth factor, but only in a cell line in which the endogenous Dio3 mRNA is also responsive to those factors. In addition, we identified a conserved enhancer located 3? of the gene containing putative AP-1 and serum response elements, and that further increases the serum responsiveness of the Dio3 promoter in a cell-specific manner. The cell-specific serum response of the Dio3 promoter and the identified enhancer may play critical roles in the regulation of gene expression at this imprinted locus. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 3 deiodinase; Dio3 gene promoter; Serum response element; Thyroid hormone; Growth factor
1. Introduction Type 3 deiodinase (D3) is a selenoenzyme that inactivates thyroid hormone (TH) by deiodinating the inner ring of both triiodothyronine (T3) and thyroxine (T4) (Bianco et al., 2002; St. Germain and Galton, 1997). In the adult, D3 activity is low and expression is limited to a few tissues such as skin, uterus and brain (Bates et al., 1999; Bianco et al., 2002; St. Germain and Galton, 1997). In contrast, D3 expression is very high in decidual tissue (Galton et al., 1999), placenta (Koopdonk-Kool et al., 1996; Roti et al., 1982) and the fetus (Galton et al., 1999; Huang et al., 2003; Kaplan and Yaskoski, 1980). Since TH levels are much lower in the developing fetus than in the adult, this pattern of expression during gestation suggests that D3 plays a * Corresponding author. Address: Departments of Medicine and Physiology, Dartmouth Medical School, One Medical Center Drive, Borwell Building, Room 720W, Lebanon, NH 03756, USA. Tel.: /1603-650-2582; fax: /1-603-650-6130. E-mail address: [email protected] (A. Hernandez).
role in protecting the developing embryo from maternal levels of TH. In certain cell culture models, we and others have described a marked up-regulation of D3 expression by the action of serum, phorbol esters and various growth factors, including basic fibroblast growth factors (bFGF) (Herna´ndez and Obregon, 1995; Courtin et al., 1991; Herna´ndez et al., 1998b; Hernandez and St. Germain, 2002; Pallud et al., 1999). The important role of bFGF in angiogenesis (Friesel and Maciag, 1995) and the high D3 activity reported in human infant hemangiomas, a tumor enriched in blood vessels (Huang et al., 2000) indicate a possible relationship between D3 expression and angiogenic processes. The essential role of growth factors during implantation (Morita et al., 1994; Stewart and Cullinan, 1997) and embryonic development and the establishment of an active vascular network during early placentation (Starkey, 1993) coincides with the high D3 expression found in the tissues involved in these processes (Galton et al., 1999; Koopdonk-Kool et al., 1996). These observations suggest that
0303-7207/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0303-7207(03)00239-9
24
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
these growth factors or others might play a role in the induction of D3 in vivo. The human and mouse genes coding for the D3 (DIO3 and Dio3 , respectively) have been localized to human chromosome 14q32 and mouse chromosome 12F1, respectively (Herna´ndez et al., 1998a), in regions that are known to contain imprinted genes. Indeed, we and others, have recently shown that Dio3 expression occurs preferentially from the paternal allele in the mouse fetus (Hernandez et al., 2002; Tsai et al., 2002). In addition, antisense mRNAs are transcribed in this locus from the opposite strand (Hernandez et al., 2002). Characterization of Dio3 has shown that all the coding and 3?-untranslated regions are contained in a single exon. The Dio3 gene promoter, located immediately 5? of this exon, is rich in basal promoter elements and includes a TATA box that is critical for function (Herna´ndez et al., 1999). In the present work, we describe the cell-specific promoter activity of the 5?flanking region (FR) of the mammalian gene and its response to growth factors and serum, and identify a conserved enhancer element located downstream. These findings add to our understanding of the regulation of gene expression from this imprinted locus.
2. Materials and methods 2.1. Generation of promoter /reporter constructs Human and mouse Dio3 promoter constructs were fashioned by standard subcloning procedures in pXP2 (Nordeen, 1988) and pGL3 (Promega Corp., Madison, WI) plasmids containing the luciferase reporter gene. DNA fragments from human and mouse D3 genomic clones previously described (Herna´ndez et al., 1998a) were used. Serial deletions of the human D3 promoter were generated using the Erase-a-Base System (Promega) and characterized by sequencing. Mutagenesis of the response elements identified in the enhancer region (see below) was accomplished by PCR using modified specific primers carrying the mutations followed by subcloning of the products. Mutations were confirmed by sequencing using an automated sequencer with fluorescent dye terminators (PE Applied Biosystems, Foster City, CA). 2.2. Cell culture BeWo and COS-7 cells were obtained from American Type Culture Collection (Manassas, VA). BVS-1 is a D3-expressing cell line generated in this laboratory as described (Hernandez and St. Germain, 2002) from the vascular stroma of rat brown adipose tissue. BeWo and BVS-1 cells were cultured in high glucose DMEM supplemented with 10% fetal bovine serum (FBS, Life
Technologies, Gaithersburg, MD) and streptomycin and penicillin (50 mg/ml). COS-7 cells were cultured in the same medium supplemented with 10% newborn calf serum. Amphibian XTC-2 and XL-2 cell lines, derived from Xenopus laevis , were kindly provided by Dr J. Tata (National Institute for Medical Research, The Ridgeway, London) and were cultured in 0.6X Leibovitz-15 medium (Sigma, St. Louis, MO) supplemented with 10% FBS (Life Technologies). 2.3. Transfection experiments All cell types were plated at a density of approximately 2 /105 cells per well in six-well culture dishes the day before transfection. Typically, cells were transfected overnight using the FuGene6 reagent (Roche) as directed by the supplier with 1/2 mg of plasmid DNA and 0.5 mg of a control b-galactosidase expression vector to correct for transfection efficiency. Medium was changed the morning after and cells were harvested after an additional 24 h period (about 40 h after the cells were transfected). In experiments studying the response of different promoter constructs to serum and/or growth factors, cells were plated and kept in medium supplemented with 1.5% FBS during and after the transfection. At 4 h before harvesting, 12% FBS, bFGF (10 ng/ml) or TPA (0.5 mM) were added to the appropriate wells. No significant variation ( B/10%) in b-galactosidase activity (use as a control of transfection efficiency) was observed in cells treated with serum or growth factors for the exposure time used. Luciferase and b-galactosidase activities were determined in cell lysates using assay reagents from Promega. Light emission was quantified using an EG & G Berthold microplate luminometer LB 96V (Wallac, Gaithersburg, MD). 2.4. Northern analysis Total RNA from all cell types was extracted in guanidinium-HCl as described (Herna´ndez et al., 1998a,b), using ethanol precipitation. For Northern analysis 15 mg of total RNA was denatured and electrophoresed on a 2.2 M formaldehyde/1% agarose gel in 1 / MOPS buffer and transferred to nylon membranes as described (Sambrook et al., 1989). A X. laevis D3 cDNA (St. Germain et al., 1994) and a 1100 bp fragment of a rat D3 cDNA (Croteau et al., 1995), corresponding to most of the translated region of the D3 mRNA, were labeled with a-[32P]-dCTP using random primers (Pharmacia /Biotech) and used as probes for the RNA obtained from the amphibian and mammalian cell lines, respectively. Filters were hybridized for 20 /24 h at 42 8C (in 50% formamide, 5 / SSC, 2 / Denhardt’s, 0.1% SDS), washed four times in 2 / SSC/0.2% SDS at room temperature for 15 min and then twice in 0.1 /
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
SSC/0.2% SDS at 65 8C for 20 min. Filters were autoradiographed for 2 days and then stripped and hybridized with cyclophilin or actin as a control to correct for differences between lanes in the amount of RNA loaded.
3. Results 3.1. Activity of the mouse and human D3 promoter and 5flanking regions To gain further understanding of the events regulating the transcription of the mammalian gene coding for the D3, we analyzed the promoter activity of 6 kb of the Dio3 5?-FR. Genomic fragments of different sizes, spanning 193/5984 bp were subcloned in front of the luciferase reporter gene and transfected into several cell lines that do or do not express endogenous D3. Results show robust promoter activity for all constructs in all cell lines, varying from a 100- to 3000-fold increase in promoter activity when compared with the activity of the empty vector (Fig. 1). When using human COS-7
25
cells, that do not express endogenous D3, relatively modest promoter activity was found that did not vary significantly between different constructs. Promoter activities were two to tenfold higher in cells that do express high levels of endogenous D3, such as the rat BVS-1 and Xenopus XL-2 cell lines. However, in these cell lines, promoter constructs spanning 193 /1443 bp of 5?-FR were markedly more active than larger constructs. Interestingly, in both BVS-1 and XL-2 cells, a sharp decrease in promoter activity is observed when 301 bp of additional 5? sequence is added to the construct starting at /1443 (Fig. 1, compare /1443 and /1744 constructs). The finding that the promoter activity is highest in constructs containing up to position /1443, suggests that this region is the critically important for transcription in D3-expressing cells. Of note, this region is well conserved between the mouse and human DIO3 genes (83% homology) and is highly G/C enriched. In order to study in more detail which sequences in this conserved region might be a potential target for transcription factors, we used a DNA exonuclease-based protocol to perform linear deletions in a plasmid containing 1486 bp of the human DIO3 5?-FR driving
Fig. 1. Promoter activity of the mouse Dio3 5?-FR. pXP2 plasmid constructs containing genomic sequences spanning up to 6 kb of the mouse Dio3 5?-FR were transfected into BVS-1, XL-2 and COS-7 cell lines and analyzed for luciferase activity as described in Section 2. Bars represent the mean9/S.D. of two different experiments, each performed with triplicate cultures.
26
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
a luciferase reporter gene. We generated 18 human DIO3 promoter constructs with inserts starting at different positions, from /42 to /1486. The 3? end of all constructs was located at position /72, which is just before the native translational start codon. Their promoter activity was analyzed by transient transfections experiments in three different cell lines (Fig. 2). In these experiments, the human DIO3 promoter was also more active in D3-expressing cells (XTC-2 and BVS-1) than in cells that do not express D3 (BeWo). A fragment containing only 72 bp of the human 5?-FR was sufficient to generate strong promoter activity in D3-expressing cells as compared to the empty vector (400- to 1300-fold increase). Several important cell-specific variations in promoter activity can be observed as a result of small differences in sequence. Notably, the most dramatic change in promoter activity occurs in the region located between positions /415 and /272. The addition of 129 bp of 5? sequence to the /272 promoter construct results in a sharp decrease in promoter activity in D3expressing cells (XTC-2 and BVS-1), but produces a marked increase in luciferase activity in BeWo cells (Fig. 2). Further addition of 14 bp of 5? sequence to the /401 construct results in inhibition of promoter activity in BeWo cells, but the opposite effect is observed in XTC-2 and BVS-1 cells. These results indicate a role for cellspecific factors in regulating transcription by binding DNA sequences within this region.
3.2. Response to serum and growth factors It has been shown that D3 mRNA is markedly induced by serum, growth factors and phorbol esters in different cell culture models (Courtin et al., 1991; Herna´ndez and Obregon, 1995), including BVS-1 cells (Hernandez and St. Germain, 2002). The induction of endogenous D3 mRNA by these factors was examined in the cells used for DIO3 promoter studies. Endogenous D3 mRNA levels are markedly induced by serum, bFGF and TPA in the rat BVS-1 cells (Fig. 3A), but these agents had no effect on D3 expression in the amphibian cell lines XTC-2 and XL-2 (Fig. 3B and C). BeWo cells do not express any detectable basal or induced D3 mRNA (Data not shown). We then analyzed whether the activity of the Dio3 promoter was induced by any of these agents in BVS-1 cells. These agents did not induce the promoter activity of the house-keeping gene phosphoglycerate-kinase (PGK), used as a negative control (Fig. 4). In contrast, a mouse Dio3 promoter construct containing 505 bp of 5?-FR showed increased promoter activity when transfected into BVS-1 cells after a 4 h treatment with TPA or bFGF (2.5- and 2-fold increase, respectively). A more pronounced induction in promoter activity (6.5-fold) was noted in cells treated with serum (Fig. 4). The serum response of several mouse and human DIO3 promoter deletion constructs was then analyzed
Fig. 2. Deletion analysis of 1.5 kb of the human DIO3 5?-FR. pXP2 plasmid constructs with serial deletions in 1.5 kb of the human DIO3 5?-FR were transfected into different cell lines and analyzed for luciferase activity as described in Section 2. Bars represent the mean9/S.D. of two different experiments, each performed with triplicate cultures.
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
27
Fig. 5. Serum response of deletion constructs of the mouse Dio3 and human DIO3 promoters. Several mouse and human promoter constructs were transfected into BVS-1 cells and the promoter activity analyzed for serum response. Cells were treated with serum for 4 h prior to harvesting. Results are expressed as a percentage of the promoter activity measured in untreated cells. The absolute promoter activities of the human constructs are shown in Fig. 2, while those from mouse are comparable to that of the shortest construct shown in Fig. 1. Bars represent the mean9/S.D. of two different experiments, each performed with triplicate cultures. Serum response of the PGK and human /42 constructs are the only ones not statistically significant.
Fig. 3. Dio3 expression in D3-expressing cell lines in response to growth factors. Twenty micrograms of total RNA isolated from BVS1, XTC-2 and XL-2 cell lines were subjected to Northern analysis as described in Section 2. Blots were autoradiographed for 2 days.
in BVS-1 cells. A response to serum is noted in all constructs that manifest significant promoter activity (Fig. 5), including a mouse fragment starting at position /192 and a human fragment containing only 72 bp of 5?-FR. A human fragment with only 42 bp of 5?-FR showed no response to serum (Fig. 5), although this short fragment is not enough to promote significant transcription (see Fig. 2 for absolute promoter activity). Thus, the serum response of the mouse and human DIO3 promoters map to sequences constituting the basal promoter, though the higher serum response observed in larger constructs might indicate that other sequences further upstream may also have a role. Analysis of the sequence of the basal promoter did not reveal any consensus serum or AP-1 response elements that might mediate this response. 3.3. Cell specific serum response and identification of a conserved enhancer
Fig. 4. Response of the mouse Dio3 promoter to serum and growth factors in BVS-1 cells. A pXP2 plasmid construct containing 519 bp of the mouse Dio3 5-FR was transfected into BVS-1 cells and promoter activity in response to serum and growth factors was determined as described in Section 2. The PGK promoter was used as a negative control. Cells were treated with serum, bFGF and TPA for 4 h prior to harvesting. Bars represent the mean9/S.D. of two different experiments performed in triplicate cultures. *, P B/0.001.
We tested the response to serum of the Dio3 promoter in cells (BeWo, XTC-2) displaying no induction of the endogenous D3 mRNA upon serum treatment. Also, in order to identify other sequences within the Dio3 locus that might be important in mediating the promoter response to serum, we examined the ability to regulate the serum response of the Dio3 promoter of several genomic fragments that lie outside the proximal pro-
28
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
moter region. To these purposes, we performed studies in BeWo, XTC-2 and BVS-1 cells. These cells were transfected with a plasmid construct containing the PGK promoter as a control and with a construct spanning 505 bp of the Dio3 5?-FR. A third construct was also transfected. This one consisted of 505 bp of the Dio3 5?-FR that had attached at its 5? end a mouse genomic fragment that was found in preliminary studies to enhance the serum response of the Dio3 promoter in BVS-1 cells. This enhancer was a 550 bp Kpn I/Bgl II restriction fragment that is located approximately 6 kb 3? of the Dio3 transcription start. The results from these experiments are shown in Fig. 6. No response to serum of the PGK promoter, used as a negative control, is found in any of the cell lines tested (Fig. 6, first pair of bars in A, B and C). A marked serum response of the Dio3 promoter is found in BVS-1 cells (panel C, pair of bars in the middle), whereas no serum response is detected in BeWo or XTC-2 cells
(Panels A and B, pair of bars in the middle) where there is also no induction of endogenous D3 mRNA by serum. The presence of the identified enhancer acting in cis of the Dio3 promoter resulted in an additional threefold increase of the serum response in BVS-1 cells (panel C, bars at the bottom). In addition, this enhancer conferred serum responsiveness to the Dio3 promoter when transfected in BeWo cells, but had no effect on promoter activity in the amphibian XTC-2 cells (bars at the bottom of panels A and B, respectively). When transfected into BVS-1 cells, one copy of this enhancer also conferred serum responsiveness to an heterologous SV40 promoter (twofold induction), while three copies resulted in a 13-fold increase in promoter activity in response to serum (Fig. 6, panel D). These results indicate that the serum response of the Dio3 promoter is cell-specific and is only present in cells in which the endogenous D3 mRNA is also induced by serum. The absence of any response in promoter activity in the D3-
Fig. 6. Factors influencing mouse Dio3 promoter activity in response to serum. A, B and C, Dio3 promoter and enhancer-dependent serum response in different cells lines. Promoter constructs containing a control promoter (PGK), the mouse Dio3 promoter (519 bp of 5?-FR) or the latter with the 550 bp genomic fragment identified as an enhancer in cis , were transfected into BeWo, XTC-2 and BVS-1 cells and their response to serum examined. D, Response to serum in BVS-1 cells of a heterologous promoter (SV40) by itself or in the presence of one or three copies of the identified enhancer in cis . Cells were treated with serum for 4 h prior to harvesting. Results are expressed as a percentage of the promoter activity measured in untreated cells. Bars represent the mean9/S.D. of two (panels A, C and D) or six (Panel C) different experiments, each performed with triplicate cultures. *, P B/0.001.
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
expressing XTC-2 cells demonstrates that regulation of D3 expression in amphibian cells might be significantly different from that in mammals, or that their transcription factors do not interact with the mouse promoter. Finally, the 3? fragment identified as having promoterenhancing properties upon serum treatment is not promoter-specific.
3.4. The enhancer is conserved and features functional AP-1 and serum response elements The relative location within the Dio3 locus of the genomic fragment with the enhancer properties described above is depicted in Fig. 7. Analysis of the corresponding mouse and human sequences reveal a 180 bp region showing 83% homology (Fig. 7). Within this homologous stretch of sequence we identified the presence of consensus AP-1 and serum response elements that were conserved between mouse and human (Fig. 7). In particular, the sequence for the serum response element is a perfect match to that represented in the transcription factor database (Wingender et al., 2000). To determine whether these putative AP-1 and serum response elements are functional and responsible for the serum response properties conferred by the enhancer fragment, we performed mutation analysis using the mouse 550 bp fragment. Triple and double point mutations were introduced in this fragment in order to disrupt significantly the consensus AP-1 and serum
29
response elements, respectively. These mutations are indicated at the top of Fig. 8. To test the effects of these mutations, we generated plasmid constructs containing the enhancer fragment, in which a mutated AP-1 site, a mutated serum response element, or both were present. These fragments were then introduced at the 5? end of a heterologous promoter (SV40), transfected into BVS-1 cells and their response to serum was measured and compared to that of the wild-type enhancer. Mutations of the AP-1 site, the serum response element, or both, resulted, respectively, in a 55, 45 and 80% inhibition of the serum response found in the wild type enhancer (Fig. 8). These results demonstrate that both response elements are essential for the enhancer properties displayed by the described mouse 550 bp 3? genomic fragment in response to serum.
4. Discussion The Dio3 gene product inactivates TH. It is highly expressed in decidual tissue, placenta and in the developing fetus (Galton et al., 1999; Koopdonk-Kool et al., 1996) and in human infant hemangiomas (Huang et al., 2000). Although none of the developmental factors responsible for Dio3 induction in vivo has been identified, growth factors and/or cytokines are potential candidates, since some of them have been shown to markedly induced Dio3 expression in various cell culture models (Courtin et al., 1991; Herna´ndez and
Fig. 7. Location and sequence of the 3? Dio3 enhancer. Location of the enhancer with respect to the Dio3 exon is indicated (top). Mouse and human sequences for the conserved region are shown and compared. Location of the conserved AP-1 and serum response elements are indicated.
30
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
Fig. 8. Serum responsiveness is impaired when putative AP-1 and serum response elements of the 3? enhancer region are mutated. Disruptive mutations in the AP-1 and SRE response elements were performed and the effect of each of the mutations or both on the serum responsiveness of a chimeric enhancer/SV40 promoter was examined in BVS-1 cells. Results are expressed as a percentage of the serum response conferred to the SV40 promoter by the wild-type enhancer. Bars represent the mean9/S.D. of two different experiments, each performed in triplicate cultures. *, P B/0.001 respect to the control.
Obregon, 1995; Hernandez and St. Germain, 2002). These factors include epidermal growth factor, fibroblast growth factor, phorbol esters and serum (Courtin et al., 1991; Herna´ndez et al., 1998b). In glial cells, induction of the Dio3 gene by growth factors appears to be mediated by the extracellular signal regulated kinases (ERKs) (Pallud et al., 1999). In the present work, we have used transient transfection experiments in several cell lines, some that do and some that do not manifest endogenous Dio3 expression, to analyze the activity of the human and mouse Dio3 promoters. Our results show that the first 1.5 kb of 5?-FR are most important for promoter activity in those cells with endogenous Dio3 expression. Notably, this genomic region is well conserved between mouse and human (83%) and has a high G/C content (80%). Finer deletion analysis has identified several short sequences within this region that are responsible for the marked cell-specific differences in promoter activity. These sequences are good candidates for the binding of cell-specific transcription factors and will be the subject of detailed further investigation. We have shown herein that the mammalian Dio3 promoter is markedly responsive to serum and, more modestly, to bFGF and phorbol esters (Fig. 4). This observation is consistent with previous reports on the
regulation of Dio3 mRNA levels by the same factors (Courtin et al., 1991; Herna´ndez et al., 1998b; Hernandez and St. Germain, 2002; Pallud et al., 1999). However, this response is only observed in cells in which the endogenous Dio3 also responds to these factors, indicating that there are cell-specific transcription factors mediating the response to serum, bFGF and TPA. The serum response is observed with both the mouse and human promoters, and the deletion analysis suggests that sequences involved in this response map to the basal promoter region. In this regard, it is remarkable that a human construct containing only 72 bp of 5?-FR displays strong promoter activity in D3-expressing cell lines and its activity increases more than threefold in response to serum. Sequence analysis of this 72 bp of promoter sequence does not reveal any consensus serum response element. However, it is possible that the serum response might be mediated by conserved, overlapping CAAT and GC boxes centered at position /54. These response elements have been shown to mediate the transcription of ERK2, a kinase that is rapidly induced by growth factors (Sugiura and Takishima, 2000). More specifically, the CAAT binding factor has been demonstrated to be critical in mediating the serum response of the thrombospondin 1 gene promoter by binding to a consensus CAAT box (Framson and Bornstein, 1993).
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
Further investigations are needed to identify precisely which sequences of the Dio3 promoter mediate the response to serum. In addition, we have identified a genomic fragment located 3? to the Dio3 exon that further enhances the serum response when acting in cis of the Dio3 promoter. This fragment also confers responsiveness to serum of the Dio3 promoter when transfected in BeWo cells, in which the Dio3 promoter alone is not responsive to serum. It also confers serum responsiveness to a heterologous promoter, as three copies of the fragment in cis dramatically increases serum-dependent promoter activity. This suggests that the serum responsiveness of this fragment is not specific to D3-expressing cells or promoter-specific. Within this fragment, we have identified 180 bp that are highly conserved between mouse and human. Sequence analysis revealed conserved putative AP-1 and serum response elements within this sequence. Binding of serum response and AP-1 factors to these consensus elements is very likely, since they received the highest scores when screened against a transcription factor database (Wingender et al., 2000). To demonstrate that these elements are functional and responsible for the enhancer properties of this 550 bp genomic fragment, we have shown that five point mutations disrupting both elements result in 80% inhibition of the serum response. This enhancer is located approximately 6 kb 3? of the Dio3 start of transcription. Although numerous response elements located 3? of the promoters that they regulate have been described, the complexity of the Dio3 locus adds uncertainty regarding the significance and role of this enhancer. The mouse Dio3 gene is preferentially expressed from the paternal allele (Hernandez et al., 2002; Tsai et al., 2002). In addition, antisense transcripts are expressed from this locus (Hernandez et al., 2002), a feature that is not unusual in imprinted genes (Reik and Walter, 2001). Furthermore, the antisense gene (Dio3as ) overlaps with the Dio3 promoter region and exon, but its 5? region and promoter have not been identified (Hernandez et al., submitted for publication) and there is yet no information about its imprinting status. These observations raise the possibility that the identified enhancer belongs to the antisense gene. Preliminary results in our lab support this hypothesis. Although this enhancer has no significant promoter activity by itself, it confers strong antisense promoter activity to genomic sequences located further 3? (Hernandez et al., submitted for publication). Another interesting possibility is that the enhancer is part of the imprinting mechanism, and acts on either the Dio3 or the Dio3as promoters depending on the imprinting status of the particular allele. This type of mechanistic model has been proposed for other imprinted loci (Reik and Walter, 2001).
31
In summary, we have shown that both basal D3 promoter activity and the response to serum and growth factors are cell specific. We have identified a genomic fragment with enhancer properties that contains functional AP-1 and serum response elements that may play a role in Dio3 expression. Additional studies are needed to identify the factors responsible for the serum and growth factors response of the Dio3 promoter and to determine the precise role of the identified enhancer in regulating gene expression in vivo in this imprinted locus .
References Bates, J.M., St. Germain, D.L., Galton, V.A., 1999. Expression profiles of the three iodothyronine deiodinases, D1, D2 and D3 in the developing rat. Endocrinology 140, 844 /851. Bianco, A.C., Salvatore, D., Gereben, B., Berry, M.J., Larsen, P.R., 2002. Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases. Endocr. Rev. 23, 38 /89. Courtin, F., Liva, P., Gavaret, J.M., Toru-Delbauffe, D., Pierre, M., 1991. Induction of 5-deiodinase activity in astroglial cells by 12-O tetradecanoylphorbol 13-acetate and fibroblast growth factors. J. Neurochem. 56, 1107 /1113. Croteau, W., Whittemore, S.L., Schneider, M.J., St. Germain, D.L., 1995. Cloning and expression of a cDNA for a mammalian type III iodothyronine deiodinase. J. Biol. Chem. 270, 16569 /16575. Framson, P., Bornstein, P., 1993. A serum response element and a binding site for NF-Y mediate the serum response of the human thrombospondin 1 gene. J. Biol. Chem. 268, 4989 /4996. Friesel, R.E., Maciag, T., 1995. Molecular mechanisms of angiogenesis: fibroblast growth factor signal transduction. FASEB J. 9, 919 /925. Galton, V.A., Martinez, E., Hernandez, A., St. Germain, E.A., Bates, J.M., St. Germain, D.L., 1999. Pregnant rat uterus expresses high levels of the type 3 iodothyronine deiodinase. J. Clin. Invest. 103, 979 /987. Herna´ndez, A., Obregon, M.J., 1995. Presence of growth factorsinduced type III iodothyronine 5-deiodinase in cultured rat brown adipocytes. Endocrinology 136, 4543 /4550. Hernandez, A., St. Germain, D.L., 2002. Dexamethasone inhibits growth factor-induced type 3 deiodinase activity and mRNA expression in a cultures cell line derived from rat neonatal brown fat vascular-stromal cells. Endocrinology 143, 2652 /2658. Herna´ndez, A., Park, J., Lyon, G.J., Mohandas, T.K., St. Germain, D.L., 1998a. Localization of the type 3 iodothyronine deiodinase (DIO3) gene to human chromosome 14q32 and mouse chromosome 12F1. Genomics 53, 119 /121. Herna´ndez, A., St. Germain, D.L., Obrego´n, M.J., 1998b. Transcriptional activation of type III inner ring deiodinase by growth factors in cultured rat brown adipocytes. Endocrinology 139, 634 /639. Herna´ndez, A., Lyon, G.J., Schneider, M.J., St. Germain, D.L., 1999. Isolation and characterization of the mouse gene for the type 3 iodothyronine deiodinase. Endocrinology 140, 124 /130. Hernandez, A., Fiering, S., Martinez, E., Galton, V.A., St. Germain, D.L., 2002. The gene locus encoding the iodothyronine deiodinase type 3 (Dio3) is imprinted in the fetus and expresses antisense transcripts. Endocrinology 143, 4483 /4486. Huang, S.A., Tu, H.M., Harney, J.W., Venihaki, M., Butte, A.J., Kozakewich, H.P., Fishman, S.J., Larsen, P.R., 2000. Severe hypothyroidism caused by type 3 iodothyronine deiodinase in infantile hemangiomas. N. Eng. J. Med. 343, 185 /189.
32
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
Huang, S.A., Dorfman, D.M., Genest, D.R., Salvatore, D., Larsen, P.R., 2003. Type 3 iodothyronine deiodinase is highly expressed in the human uteroplacental unit and in fetal epithelium. J. Clin. Endocrinol. Metab. 88, 1384 /1388. Kaplan, M.M., Yaskoski, K.A., 1980. Phenolic and tyrosyl ring deiodination of iodothyronines in rat brain homogenates. J. Clin. Invest. 66, 551 /562. Koopdonk-Kool, J.M., De Vijlder, J.J.M., Veenboer, G.J.M., RisStalpers, C., Kok, J.H., Vulsma, T., Boer, K., Visser, T.J., 1996. Type II and type III deiodinase activity in human placenta as a function of gestational age. J. Clin. Endocrinol. Metab. 81, 2154 / 2158. Morita, Y., Tsutsumi, O., Taketani, Y., 1994. In vitro treatment of embryos with epidermal growth factor improves viability and increases the implantation rate of blastocysts transferred to recipient mice. Am. J. Obstet. Gynecol. 171, 406 /409. Nordeen, S.K., 1988. Luciferase reporter gene vectors for analysis of promoters and enhancers. Biotechniques 6, 454 /457. Pallud, S., Ramauge, M., Gavaret, J.M., Lennon, A.M., Munsch, N., St. Germain, D.L., Pierre, M., Courtin, F., 1999. Regulation of type 3 iodothyronine deiodinase expression in cultured rat astrocytes: role of the Erk cascade. Endocrinology 140, 2917 /2923. Reik, W., Walter, J., 2001. Genomic imprinting: parental influence on the genome. Nat. Rev. Genet. 2, 21 /32. Roti, E., Braverman, L.E., Fang, S., Alex, S., Emerson, C.E., 1982. Ontogenesis and placental inner ring thyroxine deiodinase and amniotic fluid 3,3?,5?-triiodothyronine concentration in the rat. Endocrinology 111, 959 /963.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. A Laboratory Manual, vol. 1, second ed.. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. St. Germain, D.L., Galton, V.A., 1997. The deiodinase family of selenoproteins. Thyroid 7, 655 /668. St. Germain, D.L., Schwartzman, R., Croteau, W., Kanamori, A., Wang, Z., Brown, D.D., Galton, V.A., 1994. A thyroid hormone regulated gene in Xenopus laevis encodes a type III iodothyronine 5-deiodinase. Proc. Natl. Acad. Sci. USA 91, 7767 /7771 Correction: Proc. Natl. Acad. Sci. USA 91, 11282.. Starkey, P.M., 1993. The decidua and factors controlling placentation. In: Redman, C.W.G., Sargent, I.L., Starkey, P.M. (Eds.), The Human Placenta. Blackwell Scientific Publications, Boston, pp. 362 /413. Stewart, C.L., Cullinan, E.B., 1997. Preimplantation development of the mammalian embryo and its regulation by growth factors. Develop. Genet. 21, 91 /101. Sugiura, N., Takishima, K., 2000. Regulation of the gene promoter for extracellular signal-regulated protein kinase 2 by transcription factors NF-Y and Sp3. Biochem. J. 347, 155 /161. Tsai, C.-E., Lin, S.-P., Ito, M., Takagi, N., Takada, S., FergusonSmith, A.C., 2002. Genomic imprinting contributes to thyroid hormone metabolism in the mouse embryo. Cur. Biol. 12, 1221 / 1226. Wingender, E., Chen, X., Hehl, R., Karas, H., Liebich, I., Matys, V., Meinhardt, T., Pru¨ß, M., Reuter, I., Schacherer, F., 2000. TRANSFAC: an integrated system for gene expression regulation. Nucleic Acid Res. 28, 316 /319.
Activity and response to serum of the mammalian thyroid hormone deiodinase 3 gene promoter: identification of a conserved enhancer Arturo Hernandez *, Donald L. St. Germain Departments of Medicine and Physiology, Dartmouth Medical School, Lebanon, NH, USA Received 23 April 2003; accepted 25 June 2003
Abstract Dio3 is an imprinted gene that codes for the type 3 deiodinase (D3), a conserved selenocysteine-containing enzyme that degrades thyroid hormones (TH) into inactive metabolites. D3 is partially responsible for the very low TH levels that are found in utero as it is highly expressed in mammalian uterus, placenta, and fetal tissues. For this reason, it is presumed that D3 protects the developing fetus from high, adult-like levels of TH. To further understand the regulation of D3 expression, we analyzed the promoter activity of 6 kb of 5?-flanking regions of the human and mouse Dio3 , by transient transfection studies. Both mouse and human Dio3 promoters are markedly responsive to serum and, to a lesser extent, to phorbol esters and fibroblast growth factor, but only in a cell line in which the endogenous Dio3 mRNA is also responsive to those factors. In addition, we identified a conserved enhancer located 3? of the gene containing putative AP-1 and serum response elements, and that further increases the serum responsiveness of the Dio3 promoter in a cell-specific manner. The cell-specific serum response of the Dio3 promoter and the identified enhancer may play critical roles in the regulation of gene expression at this imprinted locus. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 3 deiodinase; Dio3 gene promoter; Serum response element; Thyroid hormone; Growth factor
1. Introduction Type 3 deiodinase (D3) is a selenoenzyme that inactivates thyroid hormone (TH) by deiodinating the inner ring of both triiodothyronine (T3) and thyroxine (T4) (Bianco et al., 2002; St. Germain and Galton, 1997). In the adult, D3 activity is low and expression is limited to a few tissues such as skin, uterus and brain (Bates et al., 1999; Bianco et al., 2002; St. Germain and Galton, 1997). In contrast, D3 expression is very high in decidual tissue (Galton et al., 1999), placenta (Koopdonk-Kool et al., 1996; Roti et al., 1982) and the fetus (Galton et al., 1999; Huang et al., 2003; Kaplan and Yaskoski, 1980). Since TH levels are much lower in the developing fetus than in the adult, this pattern of expression during gestation suggests that D3 plays a * Corresponding author. Address: Departments of Medicine and Physiology, Dartmouth Medical School, One Medical Center Drive, Borwell Building, Room 720W, Lebanon, NH 03756, USA. Tel.: /1603-650-2582; fax: /1-603-650-6130. E-mail address: [email protected] (A. Hernandez).
role in protecting the developing embryo from maternal levels of TH. In certain cell culture models, we and others have described a marked up-regulation of D3 expression by the action of serum, phorbol esters and various growth factors, including basic fibroblast growth factors (bFGF) (Herna´ndez and Obregon, 1995; Courtin et al., 1991; Herna´ndez et al., 1998b; Hernandez and St. Germain, 2002; Pallud et al., 1999). The important role of bFGF in angiogenesis (Friesel and Maciag, 1995) and the high D3 activity reported in human infant hemangiomas, a tumor enriched in blood vessels (Huang et al., 2000) indicate a possible relationship between D3 expression and angiogenic processes. The essential role of growth factors during implantation (Morita et al., 1994; Stewart and Cullinan, 1997) and embryonic development and the establishment of an active vascular network during early placentation (Starkey, 1993) coincides with the high D3 expression found in the tissues involved in these processes (Galton et al., 1999; Koopdonk-Kool et al., 1996). These observations suggest that
0303-7207/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0303-7207(03)00239-9
24
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
these growth factors or others might play a role in the induction of D3 in vivo. The human and mouse genes coding for the D3 (DIO3 and Dio3 , respectively) have been localized to human chromosome 14q32 and mouse chromosome 12F1, respectively (Herna´ndez et al., 1998a), in regions that are known to contain imprinted genes. Indeed, we and others, have recently shown that Dio3 expression occurs preferentially from the paternal allele in the mouse fetus (Hernandez et al., 2002; Tsai et al., 2002). In addition, antisense mRNAs are transcribed in this locus from the opposite strand (Hernandez et al., 2002). Characterization of Dio3 has shown that all the coding and 3?-untranslated regions are contained in a single exon. The Dio3 gene promoter, located immediately 5? of this exon, is rich in basal promoter elements and includes a TATA box that is critical for function (Herna´ndez et al., 1999). In the present work, we describe the cell-specific promoter activity of the 5?flanking region (FR) of the mammalian gene and its response to growth factors and serum, and identify a conserved enhancer element located downstream. These findings add to our understanding of the regulation of gene expression from this imprinted locus.
2. Materials and methods 2.1. Generation of promoter /reporter constructs Human and mouse Dio3 promoter constructs were fashioned by standard subcloning procedures in pXP2 (Nordeen, 1988) and pGL3 (Promega Corp., Madison, WI) plasmids containing the luciferase reporter gene. DNA fragments from human and mouse D3 genomic clones previously described (Herna´ndez et al., 1998a) were used. Serial deletions of the human D3 promoter were generated using the Erase-a-Base System (Promega) and characterized by sequencing. Mutagenesis of the response elements identified in the enhancer region (see below) was accomplished by PCR using modified specific primers carrying the mutations followed by subcloning of the products. Mutations were confirmed by sequencing using an automated sequencer with fluorescent dye terminators (PE Applied Biosystems, Foster City, CA). 2.2. Cell culture BeWo and COS-7 cells were obtained from American Type Culture Collection (Manassas, VA). BVS-1 is a D3-expressing cell line generated in this laboratory as described (Hernandez and St. Germain, 2002) from the vascular stroma of rat brown adipose tissue. BeWo and BVS-1 cells were cultured in high glucose DMEM supplemented with 10% fetal bovine serum (FBS, Life
Technologies, Gaithersburg, MD) and streptomycin and penicillin (50 mg/ml). COS-7 cells were cultured in the same medium supplemented with 10% newborn calf serum. Amphibian XTC-2 and XL-2 cell lines, derived from Xenopus laevis , were kindly provided by Dr J. Tata (National Institute for Medical Research, The Ridgeway, London) and were cultured in 0.6X Leibovitz-15 medium (Sigma, St. Louis, MO) supplemented with 10% FBS (Life Technologies). 2.3. Transfection experiments All cell types were plated at a density of approximately 2 /105 cells per well in six-well culture dishes the day before transfection. Typically, cells were transfected overnight using the FuGene6 reagent (Roche) as directed by the supplier with 1/2 mg of plasmid DNA and 0.5 mg of a control b-galactosidase expression vector to correct for transfection efficiency. Medium was changed the morning after and cells were harvested after an additional 24 h period (about 40 h after the cells were transfected). In experiments studying the response of different promoter constructs to serum and/or growth factors, cells were plated and kept in medium supplemented with 1.5% FBS during and after the transfection. At 4 h before harvesting, 12% FBS, bFGF (10 ng/ml) or TPA (0.5 mM) were added to the appropriate wells. No significant variation ( B/10%) in b-galactosidase activity (use as a control of transfection efficiency) was observed in cells treated with serum or growth factors for the exposure time used. Luciferase and b-galactosidase activities were determined in cell lysates using assay reagents from Promega. Light emission was quantified using an EG & G Berthold microplate luminometer LB 96V (Wallac, Gaithersburg, MD). 2.4. Northern analysis Total RNA from all cell types was extracted in guanidinium-HCl as described (Herna´ndez et al., 1998a,b), using ethanol precipitation. For Northern analysis 15 mg of total RNA was denatured and electrophoresed on a 2.2 M formaldehyde/1% agarose gel in 1 / MOPS buffer and transferred to nylon membranes as described (Sambrook et al., 1989). A X. laevis D3 cDNA (St. Germain et al., 1994) and a 1100 bp fragment of a rat D3 cDNA (Croteau et al., 1995), corresponding to most of the translated region of the D3 mRNA, were labeled with a-[32P]-dCTP using random primers (Pharmacia /Biotech) and used as probes for the RNA obtained from the amphibian and mammalian cell lines, respectively. Filters were hybridized for 20 /24 h at 42 8C (in 50% formamide, 5 / SSC, 2 / Denhardt’s, 0.1% SDS), washed four times in 2 / SSC/0.2% SDS at room temperature for 15 min and then twice in 0.1 /
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
SSC/0.2% SDS at 65 8C for 20 min. Filters were autoradiographed for 2 days and then stripped and hybridized with cyclophilin or actin as a control to correct for differences between lanes in the amount of RNA loaded.
3. Results 3.1. Activity of the mouse and human D3 promoter and 5flanking regions To gain further understanding of the events regulating the transcription of the mammalian gene coding for the D3, we analyzed the promoter activity of 6 kb of the Dio3 5?-FR. Genomic fragments of different sizes, spanning 193/5984 bp were subcloned in front of the luciferase reporter gene and transfected into several cell lines that do or do not express endogenous D3. Results show robust promoter activity for all constructs in all cell lines, varying from a 100- to 3000-fold increase in promoter activity when compared with the activity of the empty vector (Fig. 1). When using human COS-7
25
cells, that do not express endogenous D3, relatively modest promoter activity was found that did not vary significantly between different constructs. Promoter activities were two to tenfold higher in cells that do express high levels of endogenous D3, such as the rat BVS-1 and Xenopus XL-2 cell lines. However, in these cell lines, promoter constructs spanning 193 /1443 bp of 5?-FR were markedly more active than larger constructs. Interestingly, in both BVS-1 and XL-2 cells, a sharp decrease in promoter activity is observed when 301 bp of additional 5? sequence is added to the construct starting at /1443 (Fig. 1, compare /1443 and /1744 constructs). The finding that the promoter activity is highest in constructs containing up to position /1443, suggests that this region is the critically important for transcription in D3-expressing cells. Of note, this region is well conserved between the mouse and human DIO3 genes (83% homology) and is highly G/C enriched. In order to study in more detail which sequences in this conserved region might be a potential target for transcription factors, we used a DNA exonuclease-based protocol to perform linear deletions in a plasmid containing 1486 bp of the human DIO3 5?-FR driving
Fig. 1. Promoter activity of the mouse Dio3 5?-FR. pXP2 plasmid constructs containing genomic sequences spanning up to 6 kb of the mouse Dio3 5?-FR were transfected into BVS-1, XL-2 and COS-7 cell lines and analyzed for luciferase activity as described in Section 2. Bars represent the mean9/S.D. of two different experiments, each performed with triplicate cultures.
26
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
a luciferase reporter gene. We generated 18 human DIO3 promoter constructs with inserts starting at different positions, from /42 to /1486. The 3? end of all constructs was located at position /72, which is just before the native translational start codon. Their promoter activity was analyzed by transient transfections experiments in three different cell lines (Fig. 2). In these experiments, the human DIO3 promoter was also more active in D3-expressing cells (XTC-2 and BVS-1) than in cells that do not express D3 (BeWo). A fragment containing only 72 bp of the human 5?-FR was sufficient to generate strong promoter activity in D3-expressing cells as compared to the empty vector (400- to 1300-fold increase). Several important cell-specific variations in promoter activity can be observed as a result of small differences in sequence. Notably, the most dramatic change in promoter activity occurs in the region located between positions /415 and /272. The addition of 129 bp of 5? sequence to the /272 promoter construct results in a sharp decrease in promoter activity in D3expressing cells (XTC-2 and BVS-1), but produces a marked increase in luciferase activity in BeWo cells (Fig. 2). Further addition of 14 bp of 5? sequence to the /401 construct results in inhibition of promoter activity in BeWo cells, but the opposite effect is observed in XTC-2 and BVS-1 cells. These results indicate a role for cellspecific factors in regulating transcription by binding DNA sequences within this region.
3.2. Response to serum and growth factors It has been shown that D3 mRNA is markedly induced by serum, growth factors and phorbol esters in different cell culture models (Courtin et al., 1991; Herna´ndez and Obregon, 1995), including BVS-1 cells (Hernandez and St. Germain, 2002). The induction of endogenous D3 mRNA by these factors was examined in the cells used for DIO3 promoter studies. Endogenous D3 mRNA levels are markedly induced by serum, bFGF and TPA in the rat BVS-1 cells (Fig. 3A), but these agents had no effect on D3 expression in the amphibian cell lines XTC-2 and XL-2 (Fig. 3B and C). BeWo cells do not express any detectable basal or induced D3 mRNA (Data not shown). We then analyzed whether the activity of the Dio3 promoter was induced by any of these agents in BVS-1 cells. These agents did not induce the promoter activity of the house-keeping gene phosphoglycerate-kinase (PGK), used as a negative control (Fig. 4). In contrast, a mouse Dio3 promoter construct containing 505 bp of 5?-FR showed increased promoter activity when transfected into BVS-1 cells after a 4 h treatment with TPA or bFGF (2.5- and 2-fold increase, respectively). A more pronounced induction in promoter activity (6.5-fold) was noted in cells treated with serum (Fig. 4). The serum response of several mouse and human DIO3 promoter deletion constructs was then analyzed
Fig. 2. Deletion analysis of 1.5 kb of the human DIO3 5?-FR. pXP2 plasmid constructs with serial deletions in 1.5 kb of the human DIO3 5?-FR were transfected into different cell lines and analyzed for luciferase activity as described in Section 2. Bars represent the mean9/S.D. of two different experiments, each performed with triplicate cultures.
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
27
Fig. 5. Serum response of deletion constructs of the mouse Dio3 and human DIO3 promoters. Several mouse and human promoter constructs were transfected into BVS-1 cells and the promoter activity analyzed for serum response. Cells were treated with serum for 4 h prior to harvesting. Results are expressed as a percentage of the promoter activity measured in untreated cells. The absolute promoter activities of the human constructs are shown in Fig. 2, while those from mouse are comparable to that of the shortest construct shown in Fig. 1. Bars represent the mean9/S.D. of two different experiments, each performed with triplicate cultures. Serum response of the PGK and human /42 constructs are the only ones not statistically significant.
Fig. 3. Dio3 expression in D3-expressing cell lines in response to growth factors. Twenty micrograms of total RNA isolated from BVS1, XTC-2 and XL-2 cell lines were subjected to Northern analysis as described in Section 2. Blots were autoradiographed for 2 days.
in BVS-1 cells. A response to serum is noted in all constructs that manifest significant promoter activity (Fig. 5), including a mouse fragment starting at position /192 and a human fragment containing only 72 bp of 5?-FR. A human fragment with only 42 bp of 5?-FR showed no response to serum (Fig. 5), although this short fragment is not enough to promote significant transcription (see Fig. 2 for absolute promoter activity). Thus, the serum response of the mouse and human DIO3 promoters map to sequences constituting the basal promoter, though the higher serum response observed in larger constructs might indicate that other sequences further upstream may also have a role. Analysis of the sequence of the basal promoter did not reveal any consensus serum or AP-1 response elements that might mediate this response. 3.3. Cell specific serum response and identification of a conserved enhancer
Fig. 4. Response of the mouse Dio3 promoter to serum and growth factors in BVS-1 cells. A pXP2 plasmid construct containing 519 bp of the mouse Dio3 5-FR was transfected into BVS-1 cells and promoter activity in response to serum and growth factors was determined as described in Section 2. The PGK promoter was used as a negative control. Cells were treated with serum, bFGF and TPA for 4 h prior to harvesting. Bars represent the mean9/S.D. of two different experiments performed in triplicate cultures. *, P B/0.001.
We tested the response to serum of the Dio3 promoter in cells (BeWo, XTC-2) displaying no induction of the endogenous D3 mRNA upon serum treatment. Also, in order to identify other sequences within the Dio3 locus that might be important in mediating the promoter response to serum, we examined the ability to regulate the serum response of the Dio3 promoter of several genomic fragments that lie outside the proximal pro-
28
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
moter region. To these purposes, we performed studies in BeWo, XTC-2 and BVS-1 cells. These cells were transfected with a plasmid construct containing the PGK promoter as a control and with a construct spanning 505 bp of the Dio3 5?-FR. A third construct was also transfected. This one consisted of 505 bp of the Dio3 5?-FR that had attached at its 5? end a mouse genomic fragment that was found in preliminary studies to enhance the serum response of the Dio3 promoter in BVS-1 cells. This enhancer was a 550 bp Kpn I/Bgl II restriction fragment that is located approximately 6 kb 3? of the Dio3 transcription start. The results from these experiments are shown in Fig. 6. No response to serum of the PGK promoter, used as a negative control, is found in any of the cell lines tested (Fig. 6, first pair of bars in A, B and C). A marked serum response of the Dio3 promoter is found in BVS-1 cells (panel C, pair of bars in the middle), whereas no serum response is detected in BeWo or XTC-2 cells
(Panels A and B, pair of bars in the middle) where there is also no induction of endogenous D3 mRNA by serum. The presence of the identified enhancer acting in cis of the Dio3 promoter resulted in an additional threefold increase of the serum response in BVS-1 cells (panel C, bars at the bottom). In addition, this enhancer conferred serum responsiveness to the Dio3 promoter when transfected in BeWo cells, but had no effect on promoter activity in the amphibian XTC-2 cells (bars at the bottom of panels A and B, respectively). When transfected into BVS-1 cells, one copy of this enhancer also conferred serum responsiveness to an heterologous SV40 promoter (twofold induction), while three copies resulted in a 13-fold increase in promoter activity in response to serum (Fig. 6, panel D). These results indicate that the serum response of the Dio3 promoter is cell-specific and is only present in cells in which the endogenous D3 mRNA is also induced by serum. The absence of any response in promoter activity in the D3-
Fig. 6. Factors influencing mouse Dio3 promoter activity in response to serum. A, B and C, Dio3 promoter and enhancer-dependent serum response in different cells lines. Promoter constructs containing a control promoter (PGK), the mouse Dio3 promoter (519 bp of 5?-FR) or the latter with the 550 bp genomic fragment identified as an enhancer in cis , were transfected into BeWo, XTC-2 and BVS-1 cells and their response to serum examined. D, Response to serum in BVS-1 cells of a heterologous promoter (SV40) by itself or in the presence of one or three copies of the identified enhancer in cis . Cells were treated with serum for 4 h prior to harvesting. Results are expressed as a percentage of the promoter activity measured in untreated cells. Bars represent the mean9/S.D. of two (panels A, C and D) or six (Panel C) different experiments, each performed with triplicate cultures. *, P B/0.001.
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
expressing XTC-2 cells demonstrates that regulation of D3 expression in amphibian cells might be significantly different from that in mammals, or that their transcription factors do not interact with the mouse promoter. Finally, the 3? fragment identified as having promoterenhancing properties upon serum treatment is not promoter-specific.
3.4. The enhancer is conserved and features functional AP-1 and serum response elements The relative location within the Dio3 locus of the genomic fragment with the enhancer properties described above is depicted in Fig. 7. Analysis of the corresponding mouse and human sequences reveal a 180 bp region showing 83% homology (Fig. 7). Within this homologous stretch of sequence we identified the presence of consensus AP-1 and serum response elements that were conserved between mouse and human (Fig. 7). In particular, the sequence for the serum response element is a perfect match to that represented in the transcription factor database (Wingender et al., 2000). To determine whether these putative AP-1 and serum response elements are functional and responsible for the serum response properties conferred by the enhancer fragment, we performed mutation analysis using the mouse 550 bp fragment. Triple and double point mutations were introduced in this fragment in order to disrupt significantly the consensus AP-1 and serum
29
response elements, respectively. These mutations are indicated at the top of Fig. 8. To test the effects of these mutations, we generated plasmid constructs containing the enhancer fragment, in which a mutated AP-1 site, a mutated serum response element, or both were present. These fragments were then introduced at the 5? end of a heterologous promoter (SV40), transfected into BVS-1 cells and their response to serum was measured and compared to that of the wild-type enhancer. Mutations of the AP-1 site, the serum response element, or both, resulted, respectively, in a 55, 45 and 80% inhibition of the serum response found in the wild type enhancer (Fig. 8). These results demonstrate that both response elements are essential for the enhancer properties displayed by the described mouse 550 bp 3? genomic fragment in response to serum.
4. Discussion The Dio3 gene product inactivates TH. It is highly expressed in decidual tissue, placenta and in the developing fetus (Galton et al., 1999; Koopdonk-Kool et al., 1996) and in human infant hemangiomas (Huang et al., 2000). Although none of the developmental factors responsible for Dio3 induction in vivo has been identified, growth factors and/or cytokines are potential candidates, since some of them have been shown to markedly induced Dio3 expression in various cell culture models (Courtin et al., 1991; Herna´ndez and
Fig. 7. Location and sequence of the 3? Dio3 enhancer. Location of the enhancer with respect to the Dio3 exon is indicated (top). Mouse and human sequences for the conserved region are shown and compared. Location of the conserved AP-1 and serum response elements are indicated.
30
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
Fig. 8. Serum responsiveness is impaired when putative AP-1 and serum response elements of the 3? enhancer region are mutated. Disruptive mutations in the AP-1 and SRE response elements were performed and the effect of each of the mutations or both on the serum responsiveness of a chimeric enhancer/SV40 promoter was examined in BVS-1 cells. Results are expressed as a percentage of the serum response conferred to the SV40 promoter by the wild-type enhancer. Bars represent the mean9/S.D. of two different experiments, each performed in triplicate cultures. *, P B/0.001 respect to the control.
Obregon, 1995; Hernandez and St. Germain, 2002). These factors include epidermal growth factor, fibroblast growth factor, phorbol esters and serum (Courtin et al., 1991; Herna´ndez et al., 1998b). In glial cells, induction of the Dio3 gene by growth factors appears to be mediated by the extracellular signal regulated kinases (ERKs) (Pallud et al., 1999). In the present work, we have used transient transfection experiments in several cell lines, some that do and some that do not manifest endogenous Dio3 expression, to analyze the activity of the human and mouse Dio3 promoters. Our results show that the first 1.5 kb of 5?-FR are most important for promoter activity in those cells with endogenous Dio3 expression. Notably, this genomic region is well conserved between mouse and human (83%) and has a high G/C content (80%). Finer deletion analysis has identified several short sequences within this region that are responsible for the marked cell-specific differences in promoter activity. These sequences are good candidates for the binding of cell-specific transcription factors and will be the subject of detailed further investigation. We have shown herein that the mammalian Dio3 promoter is markedly responsive to serum and, more modestly, to bFGF and phorbol esters (Fig. 4). This observation is consistent with previous reports on the
regulation of Dio3 mRNA levels by the same factors (Courtin et al., 1991; Herna´ndez et al., 1998b; Hernandez and St. Germain, 2002; Pallud et al., 1999). However, this response is only observed in cells in which the endogenous Dio3 also responds to these factors, indicating that there are cell-specific transcription factors mediating the response to serum, bFGF and TPA. The serum response is observed with both the mouse and human promoters, and the deletion analysis suggests that sequences involved in this response map to the basal promoter region. In this regard, it is remarkable that a human construct containing only 72 bp of 5?-FR displays strong promoter activity in D3-expressing cell lines and its activity increases more than threefold in response to serum. Sequence analysis of this 72 bp of promoter sequence does not reveal any consensus serum response element. However, it is possible that the serum response might be mediated by conserved, overlapping CAAT and GC boxes centered at position /54. These response elements have been shown to mediate the transcription of ERK2, a kinase that is rapidly induced by growth factors (Sugiura and Takishima, 2000). More specifically, the CAAT binding factor has been demonstrated to be critical in mediating the serum response of the thrombospondin 1 gene promoter by binding to a consensus CAAT box (Framson and Bornstein, 1993).
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
Further investigations are needed to identify precisely which sequences of the Dio3 promoter mediate the response to serum. In addition, we have identified a genomic fragment located 3? to the Dio3 exon that further enhances the serum response when acting in cis of the Dio3 promoter. This fragment also confers responsiveness to serum of the Dio3 promoter when transfected in BeWo cells, in which the Dio3 promoter alone is not responsive to serum. It also confers serum responsiveness to a heterologous promoter, as three copies of the fragment in cis dramatically increases serum-dependent promoter activity. This suggests that the serum responsiveness of this fragment is not specific to D3-expressing cells or promoter-specific. Within this fragment, we have identified 180 bp that are highly conserved between mouse and human. Sequence analysis revealed conserved putative AP-1 and serum response elements within this sequence. Binding of serum response and AP-1 factors to these consensus elements is very likely, since they received the highest scores when screened against a transcription factor database (Wingender et al., 2000). To demonstrate that these elements are functional and responsible for the enhancer properties of this 550 bp genomic fragment, we have shown that five point mutations disrupting both elements result in 80% inhibition of the serum response. This enhancer is located approximately 6 kb 3? of the Dio3 start of transcription. Although numerous response elements located 3? of the promoters that they regulate have been described, the complexity of the Dio3 locus adds uncertainty regarding the significance and role of this enhancer. The mouse Dio3 gene is preferentially expressed from the paternal allele (Hernandez et al., 2002; Tsai et al., 2002). In addition, antisense transcripts are expressed from this locus (Hernandez et al., 2002), a feature that is not unusual in imprinted genes (Reik and Walter, 2001). Furthermore, the antisense gene (Dio3as ) overlaps with the Dio3 promoter region and exon, but its 5? region and promoter have not been identified (Hernandez et al., submitted for publication) and there is yet no information about its imprinting status. These observations raise the possibility that the identified enhancer belongs to the antisense gene. Preliminary results in our lab support this hypothesis. Although this enhancer has no significant promoter activity by itself, it confers strong antisense promoter activity to genomic sequences located further 3? (Hernandez et al., submitted for publication). Another interesting possibility is that the enhancer is part of the imprinting mechanism, and acts on either the Dio3 or the Dio3as promoters depending on the imprinting status of the particular allele. This type of mechanistic model has been proposed for other imprinted loci (Reik and Walter, 2001).
31
In summary, we have shown that both basal D3 promoter activity and the response to serum and growth factors are cell specific. We have identified a genomic fragment with enhancer properties that contains functional AP-1 and serum response elements that may play a role in Dio3 expression. Additional studies are needed to identify the factors responsible for the serum and growth factors response of the Dio3 promoter and to determine the precise role of the identified enhancer in regulating gene expression in vivo in this imprinted locus .
References Bates, J.M., St. Germain, D.L., Galton, V.A., 1999. Expression profiles of the three iodothyronine deiodinases, D1, D2 and D3 in the developing rat. Endocrinology 140, 844 /851. Bianco, A.C., Salvatore, D., Gereben, B., Berry, M.J., Larsen, P.R., 2002. Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases. Endocr. Rev. 23, 38 /89. Courtin, F., Liva, P., Gavaret, J.M., Toru-Delbauffe, D., Pierre, M., 1991. Induction of 5-deiodinase activity in astroglial cells by 12-O tetradecanoylphorbol 13-acetate and fibroblast growth factors. J. Neurochem. 56, 1107 /1113. Croteau, W., Whittemore, S.L., Schneider, M.J., St. Germain, D.L., 1995. Cloning and expression of a cDNA for a mammalian type III iodothyronine deiodinase. J. Biol. Chem. 270, 16569 /16575. Framson, P., Bornstein, P., 1993. A serum response element and a binding site for NF-Y mediate the serum response of the human thrombospondin 1 gene. J. Biol. Chem. 268, 4989 /4996. Friesel, R.E., Maciag, T., 1995. Molecular mechanisms of angiogenesis: fibroblast growth factor signal transduction. FASEB J. 9, 919 /925. Galton, V.A., Martinez, E., Hernandez, A., St. Germain, E.A., Bates, J.M., St. Germain, D.L., 1999. Pregnant rat uterus expresses high levels of the type 3 iodothyronine deiodinase. J. Clin. Invest. 103, 979 /987. Herna´ndez, A., Obregon, M.J., 1995. Presence of growth factorsinduced type III iodothyronine 5-deiodinase in cultured rat brown adipocytes. Endocrinology 136, 4543 /4550. Hernandez, A., St. Germain, D.L., 2002. Dexamethasone inhibits growth factor-induced type 3 deiodinase activity and mRNA expression in a cultures cell line derived from rat neonatal brown fat vascular-stromal cells. Endocrinology 143, 2652 /2658. Herna´ndez, A., Park, J., Lyon, G.J., Mohandas, T.K., St. Germain, D.L., 1998a. Localization of the type 3 iodothyronine deiodinase (DIO3) gene to human chromosome 14q32 and mouse chromosome 12F1. Genomics 53, 119 /121. Herna´ndez, A., St. Germain, D.L., Obrego´n, M.J., 1998b. Transcriptional activation of type III inner ring deiodinase by growth factors in cultured rat brown adipocytes. Endocrinology 139, 634 /639. Herna´ndez, A., Lyon, G.J., Schneider, M.J., St. Germain, D.L., 1999. Isolation and characterization of the mouse gene for the type 3 iodothyronine deiodinase. Endocrinology 140, 124 /130. Hernandez, A., Fiering, S., Martinez, E., Galton, V.A., St. Germain, D.L., 2002. The gene locus encoding the iodothyronine deiodinase type 3 (Dio3) is imprinted in the fetus and expresses antisense transcripts. Endocrinology 143, 4483 /4486. Huang, S.A., Tu, H.M., Harney, J.W., Venihaki, M., Butte, A.J., Kozakewich, H.P., Fishman, S.J., Larsen, P.R., 2000. Severe hypothyroidism caused by type 3 iodothyronine deiodinase in infantile hemangiomas. N. Eng. J. Med. 343, 185 /189.
32
A. Hernandez, D.L. St. Germain / Molecular and Cellular Endocrinology 206 (2003) 23 /32
Huang, S.A., Dorfman, D.M., Genest, D.R., Salvatore, D., Larsen, P.R., 2003. Type 3 iodothyronine deiodinase is highly expressed in the human uteroplacental unit and in fetal epithelium. J. Clin. Endocrinol. Metab. 88, 1384 /1388. Kaplan, M.M., Yaskoski, K.A., 1980. Phenolic and tyrosyl ring deiodination of iodothyronines in rat brain homogenates. J. Clin. Invest. 66, 551 /562. Koopdonk-Kool, J.M., De Vijlder, J.J.M., Veenboer, G.J.M., RisStalpers, C., Kok, J.H., Vulsma, T., Boer, K., Visser, T.J., 1996. Type II and type III deiodinase activity in human placenta as a function of gestational age. J. Clin. Endocrinol. Metab. 81, 2154 / 2158. Morita, Y., Tsutsumi, O., Taketani, Y., 1994. In vitro treatment of embryos with epidermal growth factor improves viability and increases the implantation rate of blastocysts transferred to recipient mice. Am. J. Obstet. Gynecol. 171, 406 /409. Nordeen, S.K., 1988. Luciferase reporter gene vectors for analysis of promoters and enhancers. Biotechniques 6, 454 /457. Pallud, S., Ramauge, M., Gavaret, J.M., Lennon, A.M., Munsch, N., St. Germain, D.L., Pierre, M., Courtin, F., 1999. Regulation of type 3 iodothyronine deiodinase expression in cultured rat astrocytes: role of the Erk cascade. Endocrinology 140, 2917 /2923. Reik, W., Walter, J., 2001. Genomic imprinting: parental influence on the genome. Nat. Rev. Genet. 2, 21 /32. Roti, E., Braverman, L.E., Fang, S., Alex, S., Emerson, C.E., 1982. Ontogenesis and placental inner ring thyroxine deiodinase and amniotic fluid 3,3?,5?-triiodothyronine concentration in the rat. Endocrinology 111, 959 /963.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. A Laboratory Manual, vol. 1, second ed.. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. St. Germain, D.L., Galton, V.A., 1997. The deiodinase family of selenoproteins. Thyroid 7, 655 /668. St. Germain, D.L., Schwartzman, R., Croteau, W., Kanamori, A., Wang, Z., Brown, D.D., Galton, V.A., 1994. A thyroid hormone regulated gene in Xenopus laevis encodes a type III iodothyronine 5-deiodinase. Proc. Natl. Acad. Sci. USA 91, 7767 /7771 Correction: Proc. Natl. Acad. Sci. USA 91, 11282.. Starkey, P.M., 1993. The decidua and factors controlling placentation. In: Redman, C.W.G., Sargent, I.L., Starkey, P.M. (Eds.), The Human Placenta. Blackwell Scientific Publications, Boston, pp. 362 /413. Stewart, C.L., Cullinan, E.B., 1997. Preimplantation development of the mammalian embryo and its regulation by growth factors. Develop. Genet. 21, 91 /101. Sugiura, N., Takishima, K., 2000. Regulation of the gene promoter for extracellular signal-regulated protein kinase 2 by transcription factors NF-Y and Sp3. Biochem. J. 347, 155 /161. Tsai, C.-E., Lin, S.-P., Ito, M., Takagi, N., Takada, S., FergusonSmith, A.C., 2002. Genomic imprinting contributes to thyroid hormone metabolism in the mouse embryo. Cur. Biol. 12, 1221 / 1226. Wingender, E., Chen, X., Hehl, R., Karas, H., Liebich, I., Matys, V., Meinhardt, T., Pru¨ß, M., Reuter, I., Schacherer, F., 2000. TRANSFAC: an integrated system for gene expression regulation. Nucleic Acid Res. 28, 316 /319.