| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
INSERM, U-458, Hôpital R. Debré (C.S., L.P., T.M.), 75019 Paris; and INSERM, U-129, Institut Cochin de Génétique Moléculaire, Université René Descartes (C.S., O.C., R.B., P.A., P.C.), 75014 Paris, France
Address all correspondence and requests for reprints to: Dr. Christine Perret, INSERM, U-129, ICGM, Université René Descartes, 24 rue du Faubourg Saint Jacques, 75014 Paris, France. E-mail: perret{at}cochin inserm.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A number of in vitro studies have revealed how gene expression is regulated by vitamin D. Promoter studies have shown that there is a complex interplay between the hormone- responsive elements and other cis-acting elements that leads to the recruitment of several synergizing receptors or trans-acting factors. These all contribute to the induced response (7, 8, 9, 10, 11, 12). There are also protein-protein interactions between the VDR and transcription factors, or coactivators, that lead to cell-specific alterations in transcription in response to the hormone (see Refs. 13, 14 for reviews).
The calbindin-D 9k (CaBP9k) gene is expressed in the intestinal epithelium (mainly in the duodenum and cecum) of the rat and mouse and, to a lesser extent, in the lung and uterus (15, 16, 17). It is also expressed in the mouse kidney (18, 19). Its transcription is under the control of 1,25-(OH)2D3 in rat and mouse intestine, as it is in the mouse kidney and lung (17, 19, 20). This control is transcriptional (21) and requires the VDR, as shown by the drop of expression of CaBP9k in VDR-null mice (20, 22). Some vitamin D-responsive genes described to date have been shown to contain a well defined vitamin D-responsive DNA element (VDRE) (see Ref. 13 for a review). Such a VDRE has been proposed for the rat CaBP9k gene (from -489 to -445) (23). However, this putative 9k/VDRE is very unlike the consensus VDRE sequence, consisting of two AGGTCA motifs separated by three nucleotides (24, 25). Alone, it confers only a poor responsiveness to 1,25-(OH)2D3 in transfection studies (23, 26, 27, 28). Carlberg’s group has proposed that the strength of the response might be greatly enhanced by adding thyroid hormone; they suggest that heterodimerization of VDR and thyroid hormone receptor (TR) is implicated in the hormonal response (29). These data have recently been contradicted by Raval-Pandya et al., who used a double hybrid approach to demonstrate that no direct interaction is possible between VDR and TR (27), and by Thompson et al., using gel-shift and transfection studies (28). However, all of these studies have been performed in vitro or ex vivo, and ex vivo studies are hampered by the fact that no established intestinal cell line expressing the CaBP9k gene is available. Our previous transgenic mouse analyses indicate that 4580 bp of the rat CaBP9k regulatory sequence can result in correct transgene expression in the duodenum in a vitamin D-dependent manner (30), although we did not detect this responsiveness to vitamin D using transfection assays (26). The construct that was microinjected [9k/-4580-chloramphenol acetyltransferase (CAT)] included two major intestinal deoxyribonuclease I-hypersensitive sites, HS1 at -3500 and HS4 around the transcription initiation site (30, 31). In contrast to the 9k/-4580-CAT transgene, a construct containing only the minimal promoter, which includes HS4 (9k/-117-CAT), does not confer intestinal expression on the reporter transgene (32). Juxtaposing HS1 with the minimal promoter (9k/HS1-117-CAT) produced a combination able to drive transgene expression in the intestine in a vitamin D-dependent manner, although these sequences lack the putative 9K/VDRE described by Darwish et al. (23, 32).
We therefore analyzed the role of this putative 9K/VDRE in vivo in transgenic mice by site-directed mutagenesis of this element in its whole promoter (9k/-4580 sequences). Analysis of transgene expression shows that, as expected, mutation of the putative 9k/VDRE does not change the pattern of transgene expression in the tissues, nor does it impair induction of the transgene by 1,25-(OH)2D3 in the intestine. These data indicate that the putative 9K/VDRE is not needed for the vitamin D responsiveness of the rat CaBP9k gene in vivo. We also observed an effect of the transgene insertion site on vitamin D induction of the transgene in the different tissues targeted: intestine, kidney, and lung. This finding provides evidence of a modulation of the 1,25-(OH)2D3 responsiveness, depending on the site in the host chromatin at which the transgene has been integrated, which varies from tissue to tissue.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
PCR analysis of genomic DNA
PCR analysis was carried out on genomic DNA isolated from the
tails of transgenic mice. The two primers used to amplify the rat
transgene CaBP9k sequences were -1057
5'-GAAGCACCATTCCCCT-3' -1041 and -251 5'-CGCACAGC/TCCACATTTC-3'
-267. Standard conditions were employed using 0.5 µl genomic DNA for
a 50-µl reaction.
Synthesis of VDR and RXR proteins and gel mobility shift
assays
The expression vectors mRXR
-pSG5 (from Prof. P. Chambon) and
human (h) VDR-pSG5 were transcribed and translated in rabbit
reticulocyte lysates (24). The mouse osteopontin and rat osteocalcin
VDRE probes have been described previously (24), and the rat
calbindin-D 9k VDRE (from -489 to -445) probe was that
used by Darwish et al. (23). Oligonucleotide labeling,
binding reactions, and polyacrylamide gels were previously described
(24). The antibodies used were antibody 9A7
directed against hVDR
(from JW Pike) and anti-mRXR from Prof. P. Chambon.
Animal treatments
Mice were made vitamin D deficient using the strontium chloride
method for 10 days. They were treated with
1,25-(OH)2D3 at 1800 h
and killed the next morning at 1000 h (30, 32).
CAT assays
Various tissues were dissected out from F1
transgenic mice, and cell lysates were prepared (30). CAT activity was
measured according to TLC standard protocols, with 5–300 µg protein
and reaction times of up to 4 h to keep the enzyme activity in a
linear range.
Northern blots
Total RNAs were prepared using the guanidium thiocyanate single
step procedure (35) and were analyzed by Northern blotting (30).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). We then tested the ability of
9k/VDRE to compete for a complex obtained with the mouse osteopontin
VDRE. The shift obtained with the osteopontin VDRE was due to both VDR
and RXR (Fig. 1B), as shown by the absence of a complex when VDR-
or RXR-producing lysates were used and by the use of specific
antibodies. A 20-fold molar excess of osteopontin VDRE competitor was
sufficient to compete with the binding of the RXR/VDR heterodimer to
the mouse osteopontin VDRE, but a 50-fold excess of osteocalcin VDRE
was required. A 500-fold excess of the putative 9k/VDRE only partially
competed with binding of the RXR-VDR heterodimer to the osteopontin
VDRE (Fig. 1B).
|
Production of 9kVDRE-mutated transgenic mice
We mutated the 5'-motif of the putative 9k/VDRE (5'-GGGTGT CGG
AAGCCC-3') as it is the most similar to the canonical motif RGGTSA,
specific for nuclear receptors (R = G or A and S = C or G).
We mutated the third and fifth nucleotides (both G to A) of the motif
because our results obtained by PCR-mediated random site selection had
identified the DNA sequences that bind the RXR/VDR heterodimers (24).
The A was never selected, in either the third or the fifth position. An
A in the third position alters the half-site response element to a
glucocorticoid response element (39). An EcoRV restriction
site was also engineered into the mutated fragment so that the correct
construct could be easily confirmed (Fig. 3A). The mutated 9K/VDRE thus
contained a 5'-GGATAT CGG AAGCCC-3' sequence (Fig. 3A).
|
). Four founders were obtained for
9k/-4580VDREmut-CAT, with 2, 5, 20, and 200 copies of the transgene,
respectively (Fig. 2). The lines and the mice used in this study are
indicated in Fig. 2. We confirmed the presence of the mutated VDRE in
the 9k/-4580-CAT transgene integrated into the mouse genome by
EcoRV restriction analysis (Fig. 3B
).
|
) (30). These mice had duodenal
activities in the same range as those obtained with the 9k/-4580-CAT
wild-type mice (see Fig. 4). The
9k/-4580VDREmut-CAT mice from line 120 had barely detectable CAT
activity, similar to the 9k/-4580-CAT mice from line 84 (both lines
carried two copies of the transgene; see Fig. 2), and were not used in
the following study. The patterns of CAT activity along the
cephalocaudal axis of the intestine were the same in the
9k/-4580VDREmut-CAT mice from lines 18, 44, and 129 (Fig. 4B) and in
the 9k/-4580-CAT mice carrying the wild-type sequence (Fig. 4A).
However, the CAT activity in the colon of VDREmut mice was generally
higher than that in wild-type transgenic mice. The significance of this
result is not clear. Similarly, the patterns of transgene expression in
the kidneys and lungs of VDREmut and wild-type 9k/-4580-CAT mice were
identical (Fig. 4, A and B). The control for the experimental
mutagenesis analysis was a 9k/-4580Cdx2mut-CAT transgenic mouse
line carrying a mutated Cdx2 motif. Cdx2, a homeoprotein, is a key
factor in controlling the differentiation of intestinal epithelium
(40). We have recently shown that a two-base mutation in the distal
Cdx2 motif of the 9k/-4580 sequence dramatically decreases intestinal
expression in transgenic mice (32) (see Table 1). However, mutation of
the distal Cdx2 motif did not alter transgene expression in the kidney
and lung, as indicated by the ratio between kidney and duodenal
activities in 9k/-4580-Cdx2mut-CAT transgenic mice (Fig. 4C and Table 1). This is in contrast to its dramatic effect on transgene expression
in the intestine, but agrees with the fact that Cdx2 expression is
restricted to the adult intestine (41).
|
|
were made vitamin D deficient by the strontium
chloride method. The mice were killed 16 h after a single
injection of vehicle or 25 ng
1,25-(OH)2D3 (Table
1).
The 9k/-4580-CAT mice carrying the wild-type sequence (only line 94
was analyzed in the present study) responded strongly to
1,25-(OH)2D3, with a
23.6-fold increase in CAT activity (Table 1), in agreement with our
previous study (30). In this study we also analyzed the mice from line
38 harboring the 9k/-4580-CAT wild-type sequence. As shown in Table 1,
this line responded 5.4-fold to vitamin D (30). Two of the 3 lines of
transgenic mice carrying the mutated VDRE showed induction of the
transgene by 1,25-(OH)2D3;
there were significant increases in duodenal CAT activity in lines 18
(21.4-fold) and 44 (5.3-fold; Table 1). Duodenal CAT messenger RNAs
(mRNAs) were also analyzed. We found 2 mRNA species (1.5 and 1.8 kbp)
in transgenic mice (Fig 5). The vitamin
D-deficient 9k/-4580-CAT mice had lower concentrations of CAT mRNA and
CaBP9k mRNA, and
1,25-(OH)2D3 caused a
marked increase in both messengers (Fig. 5A). The amounts of duodenal
CAT mRNAs in VDREmut line 44 mice increased similarly to those of the
endogenous CaBP9k gene in response to
1,25-(OH)2D3 (Fig. 5B).
Duodenal CAT mRNAs were not detectable in line 18 mice under our
conditions using Northern blots (data not shown), probably because of
the low copy number of this line (see Fig. 2 and Table 1). The hormonal
induction of the transgene seemed to depend on the insertion site. The
CAT activity in mice from line 129 was not significantly stimulated
(P = 0.16) by
1,25-(OH)2D3 (Table 1).
There were increases in CAT mRNA in 3 of the 5 mice tested (Fig. 5C),
in agreement with this result. We postulated that the high transgene
copy number of line 129 (200 copies; see Fig. 2) limited its response
to vitamin D by titrating an essential element(s) needed for
transcription. However, Northern blots indicated that the regulation of
the endogenous CaBP9k gene is not altered in line 129 mice
(Fig. 5C); thus, the insertion site of the transgene was probably
responsible for the poor response of the transgene in the duodenum of
line 129 mice to vitamin D. Previously, we reported that the
9k/HS1-117-CAT transgenic lines revealed a dependence on the insertion
site for the vitamin D responsiveness of the transgene (32). These
results are recapitulated in Table 1, where only 2 mouse lines (lines
20 and 43) showed a significant induction of the transgene by
1,25-(OH)2D3 (5.6- and
26.3-fold, respectively).
|
).
1,25-(OH)2D3 increased
transgene activity 4.9-fold, but this was not statistically significant
(P = 0.07; Table 1) and provided no conclusion about
the role of Cdx2 in the response to vitamin D. We have thus shown that the transgene in two of the three lines carrying a mutated VDRE transgene still responds to 1,25-(OH)2D3. This implies that the putative 9k/VDRE is not necessary for the in vivo control by vitamin D of rat CaBP9k gene expression in the duodenum. Our results also show that the induction of the transgene by 1,25-(OH)2D3 in 9k/-4580-CAT mice is similar to that of the endogenous gene. This shows that the 4580 bp of regulatory sequences contain the cis-acting elements needed for the control of the rat CaBP9k gene by 1,25-(OH)2D3. However, these regulatory sequences lack elements able to render this hormonal response independent of the integration site of the transgene in the host nucleus.
Analysis of the responses of the transgene in the kidney and lung
to vitamin D
As the endogenous CaBP9k gene is also controlled by
1,25-(OH)2D3 in mouse
kidney and lung, we analyzed the expression of the transgene in these
tissues in mouse lines carrying the 9k/-4580-CAT, the
9k/-4580VDREmut, the 9k/-4580Cdx2mut, or the 9k/HS1-117-CAT
constructs (Table 1 and Fig. 5). No CAT activity was detected in the
kidney and lung of the 9k/HS1-117-CAT mice (Table 1), in accordance
with the fact that the -1011 to -117 regulatory sequences are
required for expression in these tissues (30). The CAT activities in
the kidney and lung of the 9k/-4580-CAT mice (line 94) were not
regulated by 1,25-(OH)2D3
(Table 1). The same result was obtained for the 9k/-4580Cdx2mut-CAT
mice (Table 1 and Fig. 5D). The transgenic mice carrying the mutated
VDRE transgene, lines 18 and 44 (whose transgene responded to vitamin D
in the duodenum; Table 1 and Fig. 5B), revealed no increase in CAT
activity in the kidney and lung in response to
1,25-(OH)2D3 (Table 1).
Surprisingly, mice from line 129, which responded weakly to vitamin D
in the duodenum (Table 1 and Fig. 5C), showed significant induction of
CAT activity by
1,25-(OH)2D3 in the kidney
(2.8-fold; P = 0.001) and in the lung (3.7-fold;
P = 0.003; Table 1). These results were confirmed by
analysis of kidney CAT mRNA in line 129 mice; this mRNA was induced by
vitamin D in the same way as endogenous CaBP9k mRNA (Fig. 5E). Thus,
the transgene in the kidney and lung of mice with regulatory elements
carrying a mutated VDRE responds to vitamin D, implying that the
putative 9K/VDRE is not necessary for control of the rat
CaBP9k gene by vitamin D, in either the duodenum or the
kidney and lung. The transgene in the kidney and lung of only one
transgenic line of the five tested responded to vitamin D. Thus, the
response of the transgene in the kidney and lung to vitamin D depends
greatly on the insertion site and does not parallel the induciveness
seen in the intestine.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We investigated the role of a putative 9k/VDRE (23) in transgenic mice,
by producing mutations within the 9k/-4580 transgene, as we had
previously demonstrated that these 9k/-4580 regulatory elements
contain all of the information needed to direct correct vitamin
D-dependent expression in the intestine (30). The present results
demonstrate no role for this putative 9k/VDRE in the control of
transgene expression by vitamin D, and thus imply that this element is
not involved in the activation of the rat CaBP9k gene by
1,25-(OH)2D3 in vivo.
This result is significant, because the putative 9k/VDRE has been used
to suggest new concepts of vitamin D activation, implicating TR-VDR
heterodimers (29). The discrepancy between our results and those of the
groups of Carlberg and DeLuca probably reflect differences in the
strategies used to characterize the VDRE of the CaBP9k gene.
They used in vitro and ex vivo studies, whereas
we used in vivo methods. Transgenesis has been used in
several studies to assess the physiological implications of hormone
response elements. For example, two thyroid response elements (TREs)
were identified in vitro in the -cardiac
myosin heavy chain (MHC) gene promoter
(42). A targeted mutation in either TRE in MHC/CAT transgenic mice
has shown that the one that binds TR in vitro most
effectively is the less potent in vivo (42).
Such a hormonal response that can be seen only using an in vivo approach evokes a complex pathway of activation, illustrated here by the effects of the transgene integration site on the existence, the level and tissue specificity of the response of this transgene to vitamin D. We have shown that the transgene is unresponsive to vitamin D in some lines and/or some tissues of transgenic mice. This heterogeneity had already been observed for the responses to vitamin D of 9k/-4580-CAT wild-type transgenes between 2 transgenic mouse lines (30). Such an effect of the integration site has been also observed in transgenic mice carrying the VDRE-mutated 9k/-4580-CAT sequences or carrying the 9k/HS1-117-CAT sequences with a large internal deletion (Refs. 30, 32 and the present data). A similar repressive effect of the insertion site has been observed for the androgenic response induced by a ß-glucuronidase promoter on a luciferase reporter transgene (43). With a first construct including 3.8 kb of promoter sequences containing the androgen-responsive elements, characterized in vitro, 1 transgenic line of 6 responded to androgen; 4 of 13 were androgen responsive when 1.8 kb of intronic androgen-dependent DNA hypersensitive sites region was added, and 6 of 6 were androgen responsive when this intronic region was extended to 6.4 kbp. However, even in this latter case, the extent of transgene induction by androgen varied from 8- to 2300-fold between the transgenic lines. The researchers deduced that 2 types of element are necessary for the response of the ß-glucuronidase promoter to androgen: the classical androgen-responsive elements and an element(s) that makes the response to hormone independent of the integration site (43). Thus, our results also show that the 9k/-4580 regulatory sequences of the rat CaBP9k gene are able to make the transgene respond to vitamin D (as 4 of 7 mouse lines respond to vitamin D), but lack elements able to counteract repression by the transgene environment or, alternatively, lack elements able to create a permissive environment for vitamin D-induced transcription of the transgene.
The response of the transgene to vitamin D in the kidney and lung is also intriguing. In contrast to the mouse gene, the rat CaBP9k gene is not expressed in the kidney. For still greater complexity, the expression of the CaBP9k gene in the lung is not hormonally regulated in the rat (44), whereas it is controlled by vitamin D in the mouse (20). Our results show that 9k/-4580 sequences from the rat CaBP9K gene can express the transgene in the kidney and lung of the mouse in a vitamin D-regulated manner. Moreover, it has already been observed that the vitamin D-dependent control of the transcription of the CaBP9k gene in the kidney differs from that in the intestine, because the kinetics of expression are not the same (19), and CaBP9k mRNAs are differently affected in kidney and intestine of VDR-null mice (20, 22, 45). We show now that the response of the transgene in the kidney and lung to vitamin D depends greatly upon its site of integration, as only one of five lines tested responded to vitamin D. This transgenic line was the only that produced a poor response to vitamin D in the intestine, suggesting that the requirements for a transgene environment allowing the vitamin D response are tissue specific.
Although the CaBP9k gene was one of the first targets of vitamin D characterized (46), the mechanisms involved in its hormonal control remain unclear. Even if we have not identified any VDRE, we have restricted its (or their) localization to two modules of regulatory sequences, which are necessary for the vitamin D response, but they do not contain any consensual VDRE of the DR3 type (24, 32). Thus, the hormonal response involves a nonconventional pathway. One explanation for this complexity is that a VDRE(s) could be located in the distal module (from -3731 to -2894) of regulatory sequences. It has been demonstrated for some glucocorticoid-responsive unit or retinoic acid-responsive element that such a distal location leads to complex responses to the hormones (47, 48). In the CaBP9k promoter, a vitamin D-responsive unit could exist, composed of several cis-acting elements binding VDR or tissue-specific trans-activators, cooperating to produce a full response in vivo. Thus, the CaBP9k gene remains a promising model to make some advances in the understanding of the genomic action of vitamin D.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received July 26, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, and a member of the CTF/NF-1 family of transcription
factors mediates the vitamin D responsiveness of the c-fos promoter.
Mol Cell Biol 16:584–592[Abstract]
-cardiac myosin heavy chain gene promoter. J Biol Chem 268:4331–4336This article has been cited by other articles:
 |
|
 |
|
  Y. Song and J. C. Fleet Intestinal Resistance to 1,25 Dihydroxyvitamin D in Mice Heterozygous for the Vitamin D Receptor Knockout Allele Endocrinology, March 1, 2007; 148(3): 1396 - 1402. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D. Kutuzova, S. Akhter, S. Christakos, J. Vanhooke, C. Kimmel-Jehan, and H. F. DeLuca Calbindin D9k knockout mice are indistinguishable from wild-type mice in phenotype and serum calcium level PNAS, August 15, 2006; 103(33): 12377 - 12381. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Wang, A. Klopot, J.-N. Freund, L. N. Dowling, S. D. Krasinski, and J. C. Fleet Control of differentiation-induced calbindin-D9k gene expression in Caco-2 cells by cdx-2 and HNF-1{alpha} Am J Physiol Gastrointest Liver Physiol, November 1, 2004; 287(5): G943 - G953. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Peleg, M. Uskokovic, A. Ahene, B. Vickery, and Z. Avnur Cellular and Molecular Events Associated with the Bone-Protecting Activity of the Noncalcemic Vitamin D Analog Ro-26-9228 in Osteopenic Rats Endocrinology, May 1, 2002; 143(5): 1625 - 1636. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. C. Li, M. J. G. Bolt, L.-P. Cao, and M. D. Sitrin Effects of vitamin D receptor inactivation on the expression of calbindins and calcium metabolism Am J Physiol Endocrinol Metab, September 1, 2001; 281(3): E558 - E564. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Muller, J. G. J. Hoenderop, C. H. van Os, and R. J. M. Bindels The epithelial calcium channel, ECaC1: molecular details of a novel player in renal calcium handling Nephrol. Dial. Transplant., July 1, 2001; 16(7): 1329 - 1335. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
