| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Molecular Medicine (B.Z., L.B., S.A.M., G.W.), Karolinska Institutet, S-17176 Stockholm, Sweden; Instabilité et Alterations des Génomes (B.Z., M.F., P.G.), 06107 Nice, France; and Institut National de la Santé et de la Recherche Médicale U45, Hôpital E. Herriot (M.F.), 69437 Lyon, France
Address all correspondence and requests for reprints to: Günther Weber, Ph.D., Department of Molecular Medicine, Tumor Biochemistry Unit, CMM, L8:02, Karolinska Hospital, S-17176 Stockholm, Sweden. E-mail: gunther.weber{at}cmm.ki.se.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The highly conserved protein, menin (6, 8, 9, 10, 11), is widely expressed as a 68-kDa protein (12, 13). It is predominantly nuclear (13) but has also been observed in the cytoplasm (8, 12, 14) and specifically associated with telomeres in meiotic cells (15). Various proteins have been shown to interact with menin, most of them identified as transcription factors (for review, see Ref. 16) that may be inhibited by binding to menin. Recent investigations suggest that menin may control transcription either directly or indirectly. This may be the case for the insulin gene (17) and also for genes involved in gastric cancer (18).
Little is known about how the MEN1 gene itself is regulated. Studies on cell cycle-dependent expression of menin have led to contradictory results (12, 19, 20). RNA in situ hybridization of sporadic parathyroid tumors has shown no difference in expression between normal and tumor tissue (21). Moreover in cell lines from MEN1 patients, the expression of wild-type protein does not differ from healthy controls (12). These results indicate that cells attempt to compensate for allelic loss by up-regulating the expression of menin.
Several alternatively spliced transcripts have been identified in both human and mouse (5, 8, 22) and are displayed as two bands in multiple tissue Northern blot analysis. All splice variants identified so far differ in their first, untranslated exon and encode the same protein. RNA in situ hybridization on mouse tissues has revealed a differential spatiotemporal expression pattern of the splice variants (5). In human a larger transcript (4.2 kb) may be expressed in a more restricted manner than the ubiquitously expressed 2.9-kb transcript (2).
The mechanisms behind the transcriptional regulation of the MEN1 gene are yet unknown. We have therefore analyzed its 5'-flanking region to identify its upstream regulatory elements. The high structural and functional similarity of human and murine MEN1 gene prompted us to perform a comparative study to facilitate the detection of common regulatory denominators.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Human embryo kidney (HEK) 293, EcR-293, and E2-M1 cells were grown in DMEM or Optimem (Invitrogen, Paisley, UK) with 10% fetal bovine serum and antibiotics (100 U penicillin and 100 µg/ml streptomycin). NIH 3T3 mouse fibroblasts cells were grown in DMEM with 10% calf serum and antibiotics in concentration as above.
Generation of a cell line with inducible MEN1 expression
EcR-293 (Invitrogen) is a derivative of the HEK 293 line and constitutively expresses the ecdysone and retinoic acid receptors from the pVgRXR vector. The coding sequence of MEN1 was cloned into the Not I site of the pIND(SP1) vector (Invitrogen) and transfected into EcR-293 using Fugene (Roche, Indianapolis, IN). One day after transfection, cells were grown in the presence of hygromycin B (200 µg/ml; Roche) and cloned. After induction by Ponasterone A (Invitrogen), the expression of menin was quantified by Western blot analysis against the antiserum M23C as described (12). The stable cell line E2-M1 expressed the highest inducible amount of menin and was used further on.
SMART 5' rapid amplification of cDNA ends (RACE)
Total RNA was extracted from mouse testis and 3T3 mouse embryo cells using Trizol (Invitrogen). Two micrograms RNA were reversed transcribed from a start primer as position 1703–1683 on mouse genomic DNA (accession no. AF024513) by the use of Thermo Script II reverse transcriptase (Invitrogen) according to the manufacturer’s protocol. For second-strand synthesis and incorporation of the SMART oligo II (CLONTECH, Palo Alto, CA), Super Script reverse transcriptase (Invitrogen) was used as recommended by the supplier. The cDNA was then amplified with the SMART RACE kit according to the manufacturer’s protocol (CLONTECH) using a primer at position 1376–1354 (accession no. AF024513) in addition to the mix delivered with the kit. PCR fragments were purified from a 1% of agarose gel and subcloned into pGEMT-easy (Promega, Madison, WI) and transformed into JM109. Bacterial clones were screened by PCR reaction with primers located at 1328–1350 (accession no. AF024513) and in the vector. Positive clones were sequenced using Big Dye terminators (Perkin-Elmer, Norwalk, CT).
Nuclease protection assay
A genomic fragment was amplified from positions 635 to 930 referring to the published genomic sequence (accession no. AF 024513). The reverse primer contained a recognition site for T7 RNA polymerase. The antisense RNA probe was transcribed with T7 RNA polymerase using the Maxiscript kit (Ambion, Austin, TX) using 
-[32P]-uridine 5-triphosphate (800 Ci/mmol) for labeling. The nuclease protection assay was performed using the Multiple-NPA kit (Ambion). Briefly, 10 µg total RNA was hybridized with probe of 10,000 cpm and digested for 30 min at 42 C and a 100-fold dilution of the provided mixture of nuclease S1, RNase A, and T1. The digested sample was separated in an 8% polyacrylamide gel, side by side with a sequencing reaction of the plasmid from which the probe was generated and detected by phosphor imager analysis (BAS-1000; Fujifilm, Tokyo, Japan). The assay was similarly performed with a 5'[32P]-labeled oligonucleotide (10,000 cpm at 6,000 Ci/mmol) as the probe, hybridizing at 37 C and digesting at 25 C at a 400-fold dilution of the nuclease mixture.
Plasmids for transient transfection
A promoterless luciferase reporter vector, pGL2-basic and pGL3-promoter vector (Promega) were used in the course of these studies. The ß-galactosidase expression plasmid pEF1/LacZ (Invitrogen) was used as internal control. Fragments from the 5'-flanking region were obtained by PCR using BAC clone 272B14 for mouse and BAC clone 137c/7 for human as template. The primers used for amplification contained a restriction site for Kpn I for cloning into the pGL2-basic or pGL3- promoter vectors. Positives clones were screened by PCR and than sequenced to confirm their identity.
Ten-base pair deletions or point mutations were introduced by the QuikChange Site-directed mutagenesis kit according to the manufacturer (Stratagene, La Jolla, CA). All plasmids were verified by DNA sequencing.
Transient transfection and promoter analysis
All measurements of promoter activity were performed as triplicates and in addition repeated in an independent experiment. The 3T3 cells were transiently transfected using Lipofectamine Plus (Invitrogen) according to the manufacturer’s protocol. Briefly, cells were seeded in a 6-well plate in 2 ml growth medium without antibiotics and transfected at 60% of confluence. In each transfection, 1.3 µg DNA (50 ng pEF1/LacZ and 1.25 µg of the reporter construct), 12 µl Reagent Plus, and 8 µl Lipofectamine were used. EcR-293 and E2-M1 cells were transfected with 4.5 µl TransIT-293 (Mirus Corp., Madison, WI) and 2 µg of the reporter construct, plus 50 ng pEF1/LacZ. The cell lines were then cultivated an additional 48 h, washed in PBS (pH 7.4), and lysed in 200 µl per one well of 1x reporter lysis buffer (Promega). The activities were measured in assay systems according to the manufacturer, and the luciferase activity was normalized against ß-galactosidase.
The in silico analysis for transcription factor binding sites was performed with MatInspector version 2.2 (http://genomatix.de/matinspector) (23). Repeats were investigated with GRAIL (http://compbio.ornl.gov/grailexp/) (24) and RepeatMasker (http://ftp.genome.washington.edu/ cgi-bin/RepeatMasker).
RNA interference and cotransfection with reporter constructs
Small interfering RNA oligonucleotide duplexes (siRNA) (25) with 2-nt 3' overhangs of 2'-deoxythymidines were purchased from Dharmacon (Boulder, CO) and Proligo (Hamburg, Germany). The siRNA sequence targeting MEN1 was at position 179–197 and 381–399 relative to the start codon. The siRNA against lamin A/C (position 608–630) (2) or vascular endothelial growth factor (position 51–69) were used in the control experiments. The transfection efficiency was monitored by Western blot analysis against lamin A/C or use of a Cy3-labeled siRNA against luciferase (Dharmacon) in a parallel experiment.
ECR-293 cells were grown in 12-well plates in Optimem (Invitrogen), 10% fetal calf serum. At a 40% confluence, they were cotransfected with 0.5 µg reporter plasmid, 30 ng pEF1/LacZ, and siRNA at a final concentration of 100 nM, using 4 µl Trans-IT-TKO (Mirus) as the transfection agent. For inhibition of histone deacetylase, Trichostatin A (Sigma, St. Louis, MO) was added 12 h after transfection at a final concentration of 100 nM. Thirty-six hours after transfection, cells were extracted with reporter lysis buffer, and analyzed for luciferase activity and in Western blot analysis as above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our study was initiated on the mouse for which no full-length transcript start has previously been described. Recently four variants of the Men1 transcripts were identified (5). Three of them contained alternative spliced forms of the first exon, and the remaining variant maintained the first intron. The major spliced transcript corresponded to the human splice variant e1B of which the transcription start site has previously been determined. To identify the transcription start site in the mouse, we performed SMART-RACE on RNA isolated from mouse testis and NIH 3T3 fibroblasts, starting from exon 2. The products were cloned into plasmid pGEMT-easy, and several clones were sequenced to determine the transcription start site. All products were derived from the nonspliced form, and the transcription start site was localized on position 885 (data not shown) on the published genomic sequence (accession no. AF 024513). The identified position is identical with the transcription start site for the e1B splice version in human, and in the following we will refer to these positions as +1 for the respective species (Fig. 1
).
|
). A control experiment with a 5'-labeled oligonucleotide probe spanning from -31 to +39 gave the same result (data not shown).
We also performed this analysis on HEK 293 cells that were transfected with murine reporter vectors. A construct from -924 to +248 was used in addition to the reporter construct as above. Nuclease protection analysis produced the same signals as on NIH 3T3 cells, identical for both reporter constructs. RNA from HEK 293 cells transfected with a control vector did not give any signal, indicating that the murine probe did not protect the endogenous human transcript (Fig. 1).
The in silico analysis of the region revealed a 60% conservation between the two species up to 250 bp upstream the transcription start. However, the guanosine-cytosine content differed considerably (66% in human vs. 54% in mouse) and a CpG island was predicted for human (-194/+73) but not for mouse. Binding sites for the SP1 transcription factor and nuclear factor (NF)1 were located in both promoters but at different positions. Additional conserved potential transcription factor-binding sites were not found, except for a CCAAT-box at -178 and -174, respectively. Overlapping with this box, a recognition site for nuclear factor 
B (NFB) is predicted for human (-182 to -171) but not for mouse.
The untranslated region downstream the transcription start, including the first intron, is conserved by 72%, and contains a CpG island for both species (+320 to +821 in human, +242 to +825 in mouse).
No significant conservation was found upstream -250. A notable common feature was the presence of repetitive elements in both species. Two repeats of the B2 type are found in the mouse at -895 to -688 and -566 to -372. In human, a continuous stretch of four Alu repeats from -1497 to -534 is interrupted by two simple repeats, a polyA repeat at -813 to -786 and a TATATG repeat at -785 to -709.
Comparison of the promoter activity in mouse and human
To identify common features of transcription regulation, we assayed the promoter activity of mouse and human fragments from the 5'-flanking region in parallel. For mouse a series of 5'-deletion fragments was generated within 924 and +248 and for human within -1416 and +217. All constructs were assayed in NIH 3T3 and in EcR-293, which is a derivative of HEK 293. The results are presented in Fig. 2.
|
The longest murine fragment (-924/+248) showed higher activity in NIH 3T3 and a slightly elevated but not substantially different activity in EcR-293 when compared with the basic promoter (Fig. 2A). In EcR-293 the 5'-deletion of 0.4 kb (fragment -505/+248) only slightly reduced the maximum activity, to a level similar as the smaller constructs. In striking contrast, however, the same fragment did not show any activity above background when assayed in NIH 3T3. By further deletion of 255 bp, most of the activity was recovered.
The human promoter region showed similar features but differed with respect to their behavior in the cell lines (Fig. 2B). In EcR-293 cells, the deletion of the region from -1416 to -985 reduced its promoter activity to background level. The activity was partially recovered by deletion of further 335 bp (construct: -652/+217) and obtained its maximum after deletion of additional 213 bp (-437/+217). Furthermore, 5' deletions to positions -192 decreased the activity by 20% and 95% after deletion to -102.
In NIH 3T3 cells, the largest human construct displayed a relatively higher activity than in EcR-293, and the deletion down to -985 caused a clear but less dramatic loss of activity as in EcR-293. A low but significant activity was maintained after deletion to -102. In both cell lines, no activity above the negative control was seen when the 5'-deletions were extended to -64.
The deletion from -192 to -102 in the human promoter caused a stronger reduction of activity than the corresponding region in the mouse. Both promoters contain a putative CCAAT-box at -178 and -174, respectively, and a weak recognition site for NFB on the reverse strand at -176 (GGGAT) was predicted in the human promoter. To monitor the significance of this region, we sequentially deleted 10-bp stretches from position -192 to -103 in the human construct (-192/+217). Additionally, we introduced point mutations into (-192/+217) at the CCAAT box at -176 (CCAAT to CCTTT). When assayed in EcR-293 none of these mutations significantly reduced the promoter activities (data not shown).
Characterization of the upstream regulatory regions in the MEN1 promoter
Based on the observations reported above, we hypothesized that the MEN1 promoter contained both an inhibitory and a stimulatory region upstream the basic promoter. To assay whether the effect of the inhibitory region could be transferred to other promoters, we cloned the murine fragment from -505 to -253 and the human fragment from -985 to -613, upstream the Simian virus 40 promoter in pGL3. The murine construct was assayed in NIH 3T3 and the human in EcR-293. In both cases the fragments significantly reduced the activity of the Simian virus 40 promoter (Fig. 3), indicating a true silencing effect of this region.
|
The murine fragment between positions -925 and -505 was assayed in the same manner and displayed promoter activity (data not shown). However, nuclease protection analysis (Fig. 1) had shown that the total length construct initiated at the same position as the basic promoter. This indicated that the upper fragment was not an active promoter in the total length construct.
Feedback regulation of menin to the human promoter
Previous analysis (12, 21) had indicated that the loss of a functional MEN1 allele is compensated by up-regulation of the wild-type allele, suggesting a feedback regulation. To assay whether menin affected the activity of its own promoter, we established two systems in which the amount menin could be altered. We restricted this investigation to the human promoter.
For the overexpression of menin, the cell line E2-M1 (EcR-293-MEN1) was established by stable transfection of EcR-293 with an inducible expression vector. After treatment with ponasterone, E2-M1 overexpressed menin by about 10-fold (Fig. 4A).
|
). RNAi against unrelated genes did not alter the amount of menin.
The effect on the human promoter activity is shown in Fig. 4B for the constructs -192/+217 and -1416/+217, which showed the most prominent, and contrasting, response in the assays. On induction of menin, the activity of the human basic promoter (-192/+217) was significantly reduced, whereas the larger construct (-1416/+217) remained unaffected. In the parental line EcR-293, the constructs showed no change of activity after treatment with Ponasterone (data not shown), thus excluding an unspecific effect of the inducing hormone.
To investigate whether the effect was caused by a feedback at the putative NFB and CAAT boxes, we investigated the in vitro mutagenized constructs of (-192/+217) as described above. Because they responded to the overexpression of menin in the same manner as the unmutated construct, we excluded that these sites were responsible for the regulation (data not shown).
After down-regulation of menin by RNAi, the two constructs showed the opposite behavior. Contrary to the smaller construct that was not affected, the full-length construct was up-regulated by 2-fold (Fig. 4B). The effect was obtained with two different siRNA against MEN1, and was not observed when control genes were targeted by RNAi. This indicated that the reduction of menin was the true cause for the up-regulation. Reporter constructs solely containing the upstream regulatory region were assayed under the same conditions and did not show any induction of activity (data not shown).
Menin has previously been shown to repress junD-activated transcription by a histone deacetylase (HDAC)-dependent manner (27). To monitor whether such a mechanism was responsible for the increase of promoter activities after down-regulation of menin, we inhibited HDAC by Trichostatin A. The activities of all reporter constructs were up-regulated by 4- to 5-fold and additive to the effects caused by RNAi (data not shown). Because the effects or RNA interference and Trichostatin A were independent from each other, we concluded that the effect of down-regulation of menin is not mediated by HDAC.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The transcription start site we identified so far is identical in both species. The immediate upstream region has promoter activity and is conserved by 60%. In both species the region upstream the promoter serves as a cell-specific, inhibitory element. Further upstream, a stimulatory region is present in both species. In spite of these common elements in transcription regulation, there are functional differences: the regions from man and mouse can completely inhibit the promoter or be entirely ineffective but require a different cellular context. We presently cannot conclude whether the different behavior of the human and mouse promoter solely reflects differences in the regulation between the two species or whether tissue-specific effects might contribute. An obstacle for functional studies on MEN1 has been the lack of human cell lines from tissues in which MEN1 usually is manifested. So far, initial experiments on BON-1, a human insulinoma line, show the same activity profile and the effects of the regulatory regions of the human promoter as in EcR-293 cells.
|
Human and mouse promoters contain an inhibitory region, which, however, does not show any sequence conservation. The only common denominator is a short interspersed (SINE)/Alu repetitive element; in fact, the entire human inhibitory region consists of SINEs plus simple repeats. SINEs are able to work as silencers (32, 33) and activators (34) of transcription.
The upstream stimulatory region seems to contain the most powerful regulatory element of the promoter, capable of overriding the silencing effect of the region downstream, neutralizing the feedback further downstream caused by overexpression of menin and strongly up-regulating the expression at deficiency of menin. Again, the region contains repetitive elements in both species, and the human region is entirely composed of them. An Alu element with promoter-stimulating activities has been reported before; however, it contained binding sites for a well-known transcription factor (34). In our study, we did not discover any putative binding sites that might explain our observations. Altogether, a repetitive region that inducibly regulates transcription, without apparent transcription factor sites involved, has to our knowledge not been reported before.
SINE element may contain promoters for RNA polymerase II (35). In our study, the evidence speaks against an alternative promoter as the explanation for our data. One should keep in mind that in human tissues, alternative transcripts have been observed (2, 22), which may start further upstream the start point for the most abundant exon 1. These transcripts do not appear in kidney and have not been found in 293 cells (Bylund, L., unpublished data), which derive from kidney. Cell lines that contain the alternative transcripts may provide additional information about the transcription regulation of MEN1.
At least in vitro two different regions within the promoter may contribute to a feedback regulation but have quite a different potential. The activity of the region downstream -192 was slightly reduced at overexpression of menin but remained unaltered when the endogenous transcript was down-regulated by RNAi. Because the full-length real-life promoter did not show this response, this feedback may have little physiological relevance. Northern blot analysis on E2-M1 has confirmed that the overexpression of menin did not affect the endogenous transcript (Bylund, L., unpublished data). However, we should not outrule that the regulatory potential of this region might be used in a different cellular context.
In presence of the more distal regions, the promoter showed no effect on overexpression of menin, i.e. the regulatory activity seemed saturated at physiological concentrations of menin. On reduction of menin, a strong compensatory activity was induced, and it makes sense that the reduced expression of a tumor suppressor gene is not tolerated. This feedback mechanism confirms and explains previous observations on parathyroid tumors in which the MEN1 transcript level was maintained despite losses of heterozygosity on 11q13 (21). Also, a transcriptional feedback can explain similar observations on cell lines lymphoblastoid cell lines from MEN1 patients (12), confirming that the feedback is not restricted to endocrine cells.
We have not identified any specific transcription factor sites that may aid us to understand the promoter or its feedback mechanism. The only conserved putative site, a CCAAT-box, was not essential. A plausible explanation for the feedback would have been that menin binds to, and inhibits, transcription factors that are important for the transcription of its own promoter. The only candidate we found was NFB (30), which contained a putative binding site in the human promoter. However, this site could be removed without consequences for the feedback.
Inhibition of histone deacetylase remarkably up-regulated the promoter activity, which indicated a crucial impact of the chromatin structure to the promoter. We have outruled a HDAC-dependent mechanism as the cause for the feedback because the effects of Trichostatin A and the down-regulation of menin were additive. Still, this does not exclude any by other regulatory mechanisms of chromatin remodeling. Considering the apparent lack of unique parts in both regulatory regions, it is easier to imagine that modifications of the chromatin structure control the activity, rather than the binding or the inhibition of a transcription factor at a yet-unknown site. In this context it may be of interest that the inhibitory region contains a polymorphic simple repeat (D11S4946), which may influence the chromatin structure and promoter activity. In general, the promoter region of MEN1 and its regulation deserve attention especially in regard to the 5–10% of MEN1 families that have shown no mutations within the coding region.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: HDAC, Histone deacetylase; HEK, human embryo kidney; MEN1, multiple endocrine neoplasia type I; NFB, nuclear factor B; RACE, rapid amplification of cDNA ends; RNAi, RNA interference; SINE, short interspersed repetitive element; siRNA, small interfering RNA oligonucleotide duplex.
Received February 20, 2003.
Accepted May 9, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B proteins and inhibits NF-B-mediated transactivation. Oncogene 20:4917–4925[CrossRef][Medline]
-subunit gene 5'-flanking DNA and identification of a potential element as an Alu repetitive sequence. Biochim Biophys Acta 1493:302–318[Medline]
This article has been cited by other articles:
 |
|
 |
|
  E. Mensah-Osman, Y. Zavros, and J. L. Merchant Somatostatin stimulates menin gene expression by inhibiting protein kinase A Am J Physiol Gastrointest Liver Physiol, October 1, 2008; 295(4): G843 - G854. [Abstract] [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 |
