- Email: [email protected]
HIF1A Gene Transcription Is Dependent on a Core Promoter Sequence Encompassing Activating and Inhibiting Sequences Located Upstream from the Transcription Initiation Site and cis Elements Located within the 5′UTR
Biochemical and Biophysical Research Communications 261, 534 –540 (1999) Article ID bbrc.1999.0995, available online at http://www.idealibrary.com on
HIF1A Gene Transcription Is Dependent on a Core Promoter Sequence Encompassing Activating and Inhibiting Sequences Located Upstream from the Transcription Initiation Site and cis Elements Located within the 59UTR E. Minet,* ,1 I. Ernest,* G. Michel,† I. Roland,* J. Remacle,* M. Raes,* and C. Michiels* ,1 *Laboratoire de Biochimie et Biologie Cellulaire, Faculte´s Universitaires de la Paix, 61 rue de Bruxelles, 5000 Namur, Belgium; and †Laboratoire de Chimie Structurale, Faculte´s Universitaires de la Paix, 61 rue de Bruxelles, 5000 Namur, Belgium
Received June 9, 1999
Hypoxia inducible factor-1 (HIF-1) is a transcription factor composed of two subunits, HIF-1a and ARNT, which is activated under hypoxia. HIF-1a mRNA is expressed constitutively in a wide variety of cell types, whereas in some others HIF1A gene expression is upregulated by hypoxia. In this report, we show that in endothelial cells (HMEC-1) the HIF-1a mRNA expression level is the same in both normoxia and hypoxia. Deletion analysis experiments of the HIF1A promoter showed that in hypoxia HIF1A gene expression is upregulated through a short sequence located next to the transcription initiation site. We also show that in hypoxia another sequence located upstream from the 11 initiation site plays an inhibitory role on HIF1A transcription in HMEC-1 but not in hepatoma cells and brings back this expression level to that observed in normoxia. Finally, we demonstrate that HIF1A gene transcription is dependent on Sp1 binding sites and that the 5*UTR sequence also contains other important cis-acting elements. © 1999 Academic Press
The human HIF1A gene encodes for the HIF-1a subunit of the hypoxia-inducible factor-1 (HIF-1) (1, 2). HIF-1 is a heterodimeric transcription factor composed of the two sub-units HIF-1a and ARNT (3). It is involved in cellular adaptation to hypoxia by inducing the expression of genes such as VEGF, EPO, II NOS, GLUT1 and genes coding for proteins involved in gly1 To whom correspondence should be addressed. Fax: 132 (0)81 72 41 35. E-mail: [email protected]. Abbreviations used: HIF-1, hypoxia inducible factor-1; ARNT, aryl hydrocarbon nuclear translocator; HRE, hypoxia response element; HMEC-1, human microvascular endothelial cells; EPO, erythropoietin.
0006-291X/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
colysis (2). HIF-1 activation is mediated among others by phosphorylation and by inhibition of HIF-1a subunit degradation by the proteasome (4). In addition to this post-translational regulation, transcriptional regulation of the HIF1A gene in hypoxia has been reported. Indeed there is a transcriptional activation of HIF1A expression by hypoxia in cell lines such as hepatoma (Hep3B, HepG2) (3). HIF1A expression is also upregulated by hypoxia and by PMA in rat cardiomyocytes, in brain and kidney (5–7), this occurs without stabilization of the HIF-1a messenger (8). On the other hand, in both hypoxia and normoxia, the HIF1A expression is highly constitutive in HeLa, LN229, L929 cells and in others cell lines (8). The HIF1A gene promoter that has been recently sequenced, belongs to the TATA-less promoter family (1). Constitutive transcription of this type of promoters is often dependent on Sp1, NF-1 binding sites and/or Inr (Initiator) sequences which are located upstream or even downstream from the 11 transcription start site. The HIF1A promoter is composed of a GC rich sequence containing several Sp1 putative binding sites. Some of these Sp1 binding sites are located downstream from the transcription initiation site. This promoter also contains several putative HRE (hypoxia responsive element) which are HIF-1 cis acting elements (1). Here we provide evidence that in endothelial cells (HMEC-1) (9) but not in hepatoma cells, the level of expression of the HIF1A gene is maintained constant in hypoxia due to the presence of a sequence acting as a cis-repressor despite the presence of an upregulating element. We also show that the constitutive HIF1A transcription is partly mediated by Sp1 binding sites and cis-acting elements located within the 59UTR.
534
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
MATERIALS AND METHODS HMEC-1 cell culture. Human microvascular endothelial cells-1 (HMEC-1) (9) were grown in MCDB 131 (Gibco) containing 15% fetal calf serum and 10 ng/ml EGF, 1 mg/ml hydrocortisone, 10 mM glutamine, 50 U/ml penicillin G, 50 ng/ml amphotericin B. HepG2 cell culture. Hepatoma cells were grown in DMEM (Gibco) containing 10% fetal calf serum, 50 U/ml penicillin G, 50 ng/ml amphotericin B. Northern blot analysis. Total RNA was isolated from hypoxic and normoxic HMEC-1 cells by the guanidium thiocyanate method. 15 mg of total RNA from each sample of HMEC-1 cells were loaded on a 1% agarose–formaldehyde gel and transferred to Hybond-N 1 nylon membrane (Amersham). Two probes were used. The first one, called 59P, is a 240-bp probe corresponding to nucleotides 29 to 269 of the HIF-1a cDNA (Accession No. U22431). This probe has also more than 70% identity with the 59 end of the HIF-2 mRNA. The second one, called 39P, is a 450-bp probe corresponding to nucleotides 2059 to 2509 of the HIF-1a cDNA (accession n°: U22431), this probe has no homology with the HIF-2 mRNA sequence. As internal control, we used a 18S RNA probe. All these probes were labeled by random priming (Amersham) using [g- 32P]dCTP. Hybridization was performed at 42°C in a 50% formamide medium. Sequence analysis software. To find putative cis acting elements in the HIF1A 59UTR and promoter, we used the Transfac II analysis software in the MatInspector Professional web site (http:// genomatix.gsf.de). Parameters used are as follow: Core similarity 1.00, Matrix similarity 0.80, vertebrate. Cloning of the HIF-1a promoter. The HIF-1a promoter (1) was amplified by PCR using human genomic DNA (Boehringer-Mannheim) as template and the primers HPH5 (59-gaacagagagcccagcagagttgggcgg39) and SPH3 (59-CTTCCATGGTGAATCGGTCCCCGCGATG-39). Amplification was performed in a Perkin–Elmer 9700 thermal cycler using the Vent polymerase (Westburg Biolabs) and conditions were as follows: denaturation 5 min 99°C, cycles (353) 96°C 10 s– 65°C 10 s–75°C 50 s. The amplification product, called H800, was then cloned into the pBSKII 1 plasmid (Stratagene) and characterized by cycle sequencing (Perkin–Elmer, Big Dye Terminator). This plasmid was called pBSKII 1 H800. The H800 PCR product was restricted with the NcoI restriction endonuclease (Promega) and cloned into the pGL3-Enhancer vector (Promega) digested with SmaI and NcoI (Promega). The resulting recombinant plasmid was called pH800 and was characterized by cycle sequencing (Perkin–Elmer, Big Dye Terminator). HIF-1a promoter deletions. Deletions were performed using the Exonuclease III (Erase-a-Base system, Promega). Five micrograms of pGLE (PHIF800) was restricted with KpnI (Promega) which generates an EXO III 39 resistant overhang and with NheI (Promega) which generates an EXO III 59 sensitive overhang upstream from the HIF-1a promoter. The plasmid was then incubated with ExoIII in the digestion mix at 4°C (digestion mix: 5 mg plasmid in 10 ml of water, 6 ml of 103 ExoIII buffer (660 mM Tris–HCl, pH 8.0, 6.6 mM MgCl 2), water was added to each to a final volume of 60 ml, and 400 units of EXO III). Every minute 6 ml of the digestion mix was removed and mixed with 7.5 ml of S1 nuclease mix (7 ml of water, 5 ml of S1 buffer (0.3 M potassium acetate, pH 4.6, 2.5 M NaCl, 10 mM ZnSO 4, 50% glycerol), 2.5 Units of S1 nuclease). Each aliquot was kept on ice. Afterward all the samples were placed at room temperature for 30 min. The S1 nuclease removes the remaining single stranded tails. Finally, 1 ml of S1 nuclease stop buffer (0.3 M Tris Base, 0.05 M EDTA) was added to each aliquot and they were incubated at 70°C for 10 min in order to inactivate the S1 nuclease. The plasmids were purified by ethanol precipitation and resuspended in 9 ml of water. The ends of the linear deleted plasmids were flushed using the Klenow DNA polymerase. Therefore, 1 ml of Klenow buffer (200 mM Tris–HCl, 100 mM MgCl 2) was added to the plasmids with 0.2 U of Klenow DNA polymerase and 1 ml of a dNTP
mix (0.125 mM for each dNTP). The solution was then incubated at 37°C for 5 min and the reaction was stopped at 65°C. 40 ml of a ligase mix (31 ml of water, 2.5 ml ligase buffer (Promega), 2.5 ml 50% PEG, 0.25 ml 100 mM DTT, 1 U T4 DNA ligase) were added to each plasmid sample and the solutions were incubated at room temperature during 1 h. XL-1Blue SURE CaCl 2 competent cells were transformed and plated. The positive clones were screened and characterized by PCR screening, plasmid minipreparation followed by ScaI restriction (Promega), and cycle sequencing (Perkin–Elmer ABI PRISM, Big Dye Terminator). Other plasmid vectors. The pGL 3 x HRE vector contains three copies of the EPO HRE (10) downstream from the luciferase gene for which the expression is driven by a heterologous SV40 promoter. The pHdelAaTII vector is the pH800 vector deleted in the HIF1A 59UTR from position 1174 to 1196. The pD5.11. is the pHdelAatII vector deleted in the 59 region of the HIF1A promoter from position 2541 to 2249. Transfection experiments. HMEC-1/HepG2 transfection were performed in a 96 well plate at 90% confluency. The vectors were co-transfected with the control vector pRL-SV40 (Promega) as transfection efficiency control. Prior to transfection, the culture medium was removed and replaced by 200 ml of fresh medium without serum. Three micrograms of DOTAP transfection reagent (BoehringerMannheim) (11) was mixed with 600 ng of assay plasmid and 200 ng of control plasmid to a final volume of 15 ml in 20 mM Hepes buffer. The transfection solution is added to the cells for 3 h. The cell medium was replaced with fresh culture medium containing 2% serum. Twenty-four hours later, luciferase activity is measured. Luciferase activity was quantified in a luminometer using the dualluciferase-reporter system (Promega). For this, the culture medium was removed and 20 ml of PLB 1x (Passive Lysis Buffer, Promega) were added to each well. Firefly luciferase activity was measured after addition of a mix containing the beetle luciferin substrate (LAR I: Luciferase Assay Reagent I, Promega) to the cell lysate. Then renilla luciferase activity was measured by adding a mix containing coelenterazine (Stop & Glo: Renilla Assay Reagent, Promega). Experiments were performed in triplicates. Results are expressed as means of the ratios between firefly luciferase activity and renilla luciferase activity.
RESULTS AND DISCUSSION Northern blot analysis. Northern blot analysis was performed providing a semi-quantitative assay of the HIF-1a mRNA in HMEC-1. HMEC-1 were incubated in normoxia (20% O 2) and in hypoxia (1% O 2) during 2 h 30 min. Using the 59P probe, we showed that in HMEC-1, the HIF-1a mRNA level did not vary in hypoxia in comparison to normoxia (Fig. 1). On the other hand, a low increase of HIF-2 mRNA level was observed in hypoxia. Similar results were obtained using the 39P probe which is specific for HIF-1. HIF-1 activation is induced under hypoxia by posttranslational (2, 4) mechanisms. Northern blot analyses have shown that in a majority of cellular types, there is no variation in HIF1A gene transcription in hypoxia compared to normoxia (8). We report here that in HMEC-1, the HIF-1a messenger is constitutively expressed at the same level in both normoxia and hypoxia. However in the human Hep3B cell line, bovine pulmonary artery endothelial cells, but also, in rat brain and kidney, the HIF-1a mRNA level increases in hypoxia (5– 8). The presence of a transcriptional up-
535
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
tween positions 1180 and 1210 is located a cluster of different putative transcription factor binding sites, with among them: c-Ets-1, NF-kB, and NF-1. Since the HIF1A promoter is a TATA-less promoter (1), these elements all together suggest that the HIF1A transcription could be dependent on cis acting elements located upstream and downstream from the transcription initiation site.
FIG. 1. Comparative Northern blot analysis of HIF-1a mRNA in normoxia and hypoxia in HMEC-1 cells. HMEC-1 were incubated 2h30 in normoxia or in hypoxia. HIF-1a mRNA was detected in northern blot experiments using two probes. The first one called 59P probe hybridized on the 59 end of the HIF-1a mRNA and on the HIF-2 mRNA. The second one, called 39P probe, hybridized on the 39 end of the HIF-1a mRNA. 18S mRNA was used as internal control.
regulation mechanism for HIF-1a seems thus to be species and/or cell-type specific. HIF1A promoter sequence analysis. To identify putative cis-acting elements located within the HIF1A promoter and in the 59UTR, we used the MatInspector analysis software. Results are shown in Fig. 2. In the first 180 bp of the HIF1A promoter sequence, we found putative Sp1, AP-2, HIF-1 binding sites but also binding sites for AP-1 which is a transcription factor activated by phorbol ester and by hypoxia (12). In the 287-bp 59UTR, 2 putative Sp1 binding sites but also AP-1, and NF-kappaB binding sites were found. Be-
Functional analysis of the HIF1A promoter. To characterize the elements involved in the HIF1A gene transcription, a 0.8-kb fragment containing the entire HIF1A 59UTR (287 bp) and the promoter (11 to 2541) was cloned in the pGL3-Enhancer reporter vector using the SmaI and NcoI cloning sites. Then progressive deletions were generated from the 59 to the 39 end of the insert using the exonuclease III. We obtained a population of reporter plasmids containing deletions spanning the whole HIF1A promoter and 59UTR. All these constructions were transfected in HMEC-1 and in HepG2. The cells were then incubated 24 h in normoxia with or without CoCl 2 (13) an agent used to provoke a chemical hypoxia. As shown in Fig. 3A, the first significant transcriptional activity was detected for the plasmid p15C both in normoxia and in hypoxia. This plasmid contains the entire 59UTR sequence and a very short sequence containing at least the initiation site (1). For the next three constructions spanning the 1287 to 2105 region (p8D, p19E, p9F) (Fig. 3A), a low increase in reporter activity was observed in normoxia. However, a strong induction in the reporter activity was demonstrated in hypoxia. But, thereafter the reporter activity for the plasmid p20I (promoter region 1287 to 2201) (Fig. 3A)
FIG. 2. Putative cis elements located within the HIF1A gene 59UTR and the promoter. The HIF1A gene 59 region sequence was analyzed using the MatInspector professional software (http//:genomatix.gsf.de). Underlined sequences represent cis elements producing the highest similarity score. 536
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Functional analysis of the HIF1A promoter. The different constructs were cotransfected together with pRL-SV40 into HMEC-1 (A) and into HepG2 (B). The cells were then incubated 24 h either in normoxia or in the presence of 150 mM CoCl2 inducing a chemical hypoxia. Then the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates. The figures to the left of each construct indicates the length of the promoter numbered 59 from the transcription start site.
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Control for the chemical hypoxia (CoCl 2) efficiency. As positive control of chemical hypoxia induced by CoCl 2, a pGL3 x HRE vector was cotransfected with the pRL-SV40 control vector in HMEC-1 cells. pGL3-SV40 vector was used as a negative control. Cells were incubated 24 h in normoxia or in the presence of 150 mM CoCl 2 to induce chemical hypoxia. Then, the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates.
and the subsequent ones (p7H to pH800) became identical in both hypoxia and normoxia. These results indicate that the HIF1A core promoter sequence is very short, from 11 to 2200 bp, and that the HIF1A promoter fragment spanning the 11 to 2105 bp contains critical cis acting elements involved in the upregulation of the HIF1A gene transcription in hypoxia. This also suggests that within the sequence spanning the 2105 to 2201 bp, a transcriptional downregulator is involved in the decrease of the HIF1A gene expression under hypoxia, bringing back the HIF1A expression level to the one observed in normoxia. Sequence analysis revealed that a putative AP-1 binding site is located within the position 229 to 223, which could be the transcriptional activator because it is known that AP-1 is activated under hypoxia (12). Moreover, an AP-2 cis acting element could play the role of the transcriptional repressor because it is sometimes found in silencers (14). This mechanism could explain why according to the cell type, the HIF1A gene expression is either constitutive in normoxia and in hypoxia like in HMEC-1 or induced in hypoxia. Indeed, if the negative regulator is inactive in cells such as hepatoma, then it may be possible that a HIF1A gene transcriptional upregulation in hypoxia can be observed. However, if the cells contain both the activator and the repressor, then no increase in HIF1A gene transcription is observed in hypoxia. HepG2 hepatoma cells were transfected with the different plasmid constructs in order to confirm this hypothesis. HepG2 cells were then incubated in normoxia or in the presence of CoCl 2. A transcriptional activation also appeared in the presence of CoCl 2 for the p8D construct as for HMEC-1 (Fig. 3B). Interestingly, this CoCl 2-inducer activation is maintained through the whole promoter sequence and no down regulation was observed between the position 2105 and 2201. This result reinforce the hypothesis that the cell type-dependent expression of the HIF1A gene is modulated in hypoxia by
at least two cis elements, the first acting as an activator, the second as a repressor. CoCl 2 chemical hypoxia. We checked that the conditions of chemical hypoxia used here are able to activate HIF-1 using a luciferase reporter gene for which the expression is regulated by 3 HRE sequences (10). This vector, called pGL3 x HRE was cotransfected with the control vector pRL-SV40 in HMEC-1 and these cells were incubated 24 h in the presence of 150 mM CoCl 2. As shown in Fig. 4, the reporter expression was induced in hypoxia but not in normoxia. We conclude that the chemical hypoxia was able to activate HIF-1. Involvement of the 59UTR in HIF1A transcription regulation. Previous results pointed out that the HIF1A gene 59UTR may play a role in HIF1A transcription. The sequence analysis indicated that there are putative nondegenerated protein binding sites within the 59UTR. It is known that in TATA-less promoter cis acting elements are often found downstream from the transcription initiation site (15). Indeed, we detected a significant transcriptional activity in plasmids containing only the 59UTR and the very beginning of the promoter sequence (plasmid 15C). We thus investigated the transcriptional activity of the sequence between position 1120 and 1210. At least 3 clustered putative cis elements in addition to Sp1 binding sites were pointed out by sequence analysis. The region spanning the 1176 to 1198 nucleotides corresponding to the NF-kB/c-Rel, c-ETS-1, next to NF-1 binding sites, was deleted. These new plasmids were called pHdelAatII and pD5.11 according to the length of their 59 end. Both plasmids were transfected in HMEC-1 and were incubated 24 h in normoxia or in hypoxia (150 mM CoCl 2). The luciferase reporter activity was recorded. As shown in (Fig. 5) the level of expression of the reporter plasmid is two to four times weaker for these two plasmids compared to pH800. These results indicate that an enhancer element
538
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. Involvement of the HIF1A 59UTR in HIF1A transcription. HMEC-1 were transfected with pGL3-E, pH800 vector, pHdelAatII vector or pD.5.11. vector. The pHdelAaTII vector is deleted in the HIF1A 59UTR between position 1176 to 1198 and the pD5.11. vector is deleted between positions 2541 to 2249 and positions 1176 to 1198. Cells were incubated 24 h in normoxia or in the presence of 150 mM CoCl 2 to induce chemical hypoxia. Then, the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates. The figures to the left of each construct indicates the length of the promoter numbered 59 from the transcription start site.
within the 59UTR (1180 to 1200) is involved in the HIF1A gene transcription. Indeed, it is well known that in TATA-less promoter, cis acting elements are frequently found downstream from the transcription initiation site (15). As examples of such genes we can quote: gamma interferon (14), the chemokine receptor BLR1 (15), phosphofructokinase P1 (16). Negative and positive cis acting elements are also located in the first exon of the prostaglandin H synthase-2 gene (17). We still have to define which transcription factors bind to the 59UTR region of the HIF1A gene.
The Sp1 factor constitutively activates the HIF1A transcription. Several putative Sp1 binding sites are located within the HIF1A promoter. It is known that in TATA-less promoters, transcription is partially dependent on Sp1 factors (16). We tested mithramycine A, a Sp1 DNA binding inhibitor, on the transcriptional activity of pH800 and p15C in HMEC-1. An inhibition of the reporter expression was observed for pH800, but not for p15C which contains the 59UTR of the HIF1A gene and a short promoter sequence without Sp1 binding sites (Fig. 6). This result suggests that at least some of Sp1 binding sites are active within the HIF1A promoter on the contrary to the Sp1 binding sites located within in the 59UTR. CONCLUSION
FIG. 6. HIF1A transcription is dependent on Sp1 transcription factor. HMEC-1 were transfected with pGL3-E, pH800 or p15C vector. These constructs were cotransfected with the control vector pRL-SV40. The cells were incubated 24 h in the presence of mithramycine A 1 mM. Mithramycine A is a Sp1 DNA binding inhibitor. Then the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates.
In endothelial HMEC-1 cells, the HIF1A expression is highly constitutive in both hypoxia and normoxia. This expression is driven by a short promoter sequence (11 to 2200 bp), at least by Sp1 binding sites and by cis-acting elements located within the 59UTR region. Moreover, in hypoxia the expression level of the HIF1A gene is maintained identical to the one observed in normoxia due to the presence of a cis-repressor acting as an antagonist on an upregulating element located next to the transcriptional initiation site. The transcription factors involved in these regulations are still to be defined. ACKNOWLEDGMENTS E. Minet is a research assistant and C. Michiels is a research associate of the FNRS (Fonds National de la Recherche Scientifique, Belgium). G. Michel is a fellow of FRIA (Industrial and Agronomic
539
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Research Fund, Belgium). This work was partly supported by FRFC and SSTC. This text presents results of the Belgian Programme on Interuniversity Poles of Attraction initiated by the Belgian State, Prime Minister’s Office, Science Policy Programming. Scientific responsibility is assumed by the authors. The authors thank the Re´gion Wallonne for financial support. We are grateful to Dr. F. Candal (CDC, Atlanta, GA) for providing HMEC-1.
REFERENCES 1. Iyer, N. V., Leung, S. W., and Semenza, G. L. (1998) Genomics 52, 159 –165. 2. Semenza, G. L. (1998) Curr. Opin. Genet. Dev. 8, 588 –594. 3. Wang, G. L., Jiang, B.-H., Rue, E. A., and Semenza, G. L. (1995) Proc. Natl. Acad. Sci. USA 92, 5510 –5514. 4. Salceda, S., and Caro, J. (1997) J. Biol. Chem. 272, 22642–22647. 5. Ladoux, A., and Frelin, C. (1997) Biochem. Biophys. Res. Commun. 240, 552–556. 6. Wiener, C. M., Booth, G., and Semenza, G. L. (1996) Biochem. Biophys. Res. Commun. 225, 485– 488. 7. Palmer, L. A., Semenza, G. L., Stoler, M. H., and Johns, R. A. (1998) Am. J. Physiol. 274, L212–L219.
8. Wenger, R. H., Kvietikova, I., Roles, A., Gassmann, M., and Marti, H. H. (1997) Kidney Int. 51, 560 –563. 9. Ades, E. W., Candal, F. J., Swerlick, R. A., George, V. G., Summers, S., Bosse, D. C., and Lawley, T. J. (1992) J. Invest. Dermatol. 99, 683– 690. 10. Kvietikova, I., Wenger, R., Marti, H. H., and Gassmann, M. (1997) Kidney Int. 51, 564 –566. 11. Schmedtje, J. F., Liu, W. L., Ji, Y., Thompson, T. M., and Runge, M. S. (1996) Biochem. Biophys. Res. Commun. 220, 687– 691. 12. Rajpurohit, R., Koch, C. J., Tao, Z., Teixeira, C. M., and Shapiro, I. M. (1996) J. Cell Physiol. 168, 424 – 432. 13. Bunn, H. F., and Poyton, R. O. (1996) Physiol. Rev. 76, 839 – 885. 14. Ye, J., Cipitelli, M., Dorman, L., Ortaldo, J. R., and Young, H. A. (1996) Mol. Cell. Biol. 16(9), 4744 – 4753. 15. Wolf, I., Pevzner, V., Kaiser, E., Bernhardt, G., Claudio, E., Siebenlist, U., Forster, R., and Lipp, M. (1998) J. Biol. Chem. 273(44), 28831–28836. 16. Johnson, J. L., and McLachlan, A. (1994) Nucleic Acids Res. 22, 5085–5092. 17. Tazawa, R., Xu, X. M., Wu, K. K., and Wang, L. H. (1994) Biochem. Biophys. Res. Commun. 203, 190 –199.
540
HIF1A Gene Transcription Is Dependent on a Core Promoter Sequence Encompassing Activating and Inhibiting Sequences Located Upstream from the Transcription Initiation Site and cis Elements Located within the 59UTR E. Minet,* ,1 I. Ernest,* G. Michel,† I. Roland,* J. Remacle,* M. Raes,* and C. Michiels* ,1 *Laboratoire de Biochimie et Biologie Cellulaire, Faculte´s Universitaires de la Paix, 61 rue de Bruxelles, 5000 Namur, Belgium; and †Laboratoire de Chimie Structurale, Faculte´s Universitaires de la Paix, 61 rue de Bruxelles, 5000 Namur, Belgium
Received June 9, 1999
Hypoxia inducible factor-1 (HIF-1) is a transcription factor composed of two subunits, HIF-1a and ARNT, which is activated under hypoxia. HIF-1a mRNA is expressed constitutively in a wide variety of cell types, whereas in some others HIF1A gene expression is upregulated by hypoxia. In this report, we show that in endothelial cells (HMEC-1) the HIF-1a mRNA expression level is the same in both normoxia and hypoxia. Deletion analysis experiments of the HIF1A promoter showed that in hypoxia HIF1A gene expression is upregulated through a short sequence located next to the transcription initiation site. We also show that in hypoxia another sequence located upstream from the 11 initiation site plays an inhibitory role on HIF1A transcription in HMEC-1 but not in hepatoma cells and brings back this expression level to that observed in normoxia. Finally, we demonstrate that HIF1A gene transcription is dependent on Sp1 binding sites and that the 5*UTR sequence also contains other important cis-acting elements. © 1999 Academic Press
The human HIF1A gene encodes for the HIF-1a subunit of the hypoxia-inducible factor-1 (HIF-1) (1, 2). HIF-1 is a heterodimeric transcription factor composed of the two sub-units HIF-1a and ARNT (3). It is involved in cellular adaptation to hypoxia by inducing the expression of genes such as VEGF, EPO, II NOS, GLUT1 and genes coding for proteins involved in gly1 To whom correspondence should be addressed. Fax: 132 (0)81 72 41 35. E-mail: [email protected]. Abbreviations used: HIF-1, hypoxia inducible factor-1; ARNT, aryl hydrocarbon nuclear translocator; HRE, hypoxia response element; HMEC-1, human microvascular endothelial cells; EPO, erythropoietin.
0006-291X/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
colysis (2). HIF-1 activation is mediated among others by phosphorylation and by inhibition of HIF-1a subunit degradation by the proteasome (4). In addition to this post-translational regulation, transcriptional regulation of the HIF1A gene in hypoxia has been reported. Indeed there is a transcriptional activation of HIF1A expression by hypoxia in cell lines such as hepatoma (Hep3B, HepG2) (3). HIF1A expression is also upregulated by hypoxia and by PMA in rat cardiomyocytes, in brain and kidney (5–7), this occurs without stabilization of the HIF-1a messenger (8). On the other hand, in both hypoxia and normoxia, the HIF1A expression is highly constitutive in HeLa, LN229, L929 cells and in others cell lines (8). The HIF1A gene promoter that has been recently sequenced, belongs to the TATA-less promoter family (1). Constitutive transcription of this type of promoters is often dependent on Sp1, NF-1 binding sites and/or Inr (Initiator) sequences which are located upstream or even downstream from the 11 transcription start site. The HIF1A promoter is composed of a GC rich sequence containing several Sp1 putative binding sites. Some of these Sp1 binding sites are located downstream from the transcription initiation site. This promoter also contains several putative HRE (hypoxia responsive element) which are HIF-1 cis acting elements (1). Here we provide evidence that in endothelial cells (HMEC-1) (9) but not in hepatoma cells, the level of expression of the HIF1A gene is maintained constant in hypoxia due to the presence of a sequence acting as a cis-repressor despite the presence of an upregulating element. We also show that the constitutive HIF1A transcription is partly mediated by Sp1 binding sites and cis-acting elements located within the 59UTR.
534
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
MATERIALS AND METHODS HMEC-1 cell culture. Human microvascular endothelial cells-1 (HMEC-1) (9) were grown in MCDB 131 (Gibco) containing 15% fetal calf serum and 10 ng/ml EGF, 1 mg/ml hydrocortisone, 10 mM glutamine, 50 U/ml penicillin G, 50 ng/ml amphotericin B. HepG2 cell culture. Hepatoma cells were grown in DMEM (Gibco) containing 10% fetal calf serum, 50 U/ml penicillin G, 50 ng/ml amphotericin B. Northern blot analysis. Total RNA was isolated from hypoxic and normoxic HMEC-1 cells by the guanidium thiocyanate method. 15 mg of total RNA from each sample of HMEC-1 cells were loaded on a 1% agarose–formaldehyde gel and transferred to Hybond-N 1 nylon membrane (Amersham). Two probes were used. The first one, called 59P, is a 240-bp probe corresponding to nucleotides 29 to 269 of the HIF-1a cDNA (Accession No. U22431). This probe has also more than 70% identity with the 59 end of the HIF-2 mRNA. The second one, called 39P, is a 450-bp probe corresponding to nucleotides 2059 to 2509 of the HIF-1a cDNA (accession n°: U22431), this probe has no homology with the HIF-2 mRNA sequence. As internal control, we used a 18S RNA probe. All these probes were labeled by random priming (Amersham) using [g- 32P]dCTP. Hybridization was performed at 42°C in a 50% formamide medium. Sequence analysis software. To find putative cis acting elements in the HIF1A 59UTR and promoter, we used the Transfac II analysis software in the MatInspector Professional web site (http:// genomatix.gsf.de). Parameters used are as follow: Core similarity 1.00, Matrix similarity 0.80, vertebrate. Cloning of the HIF-1a promoter. The HIF-1a promoter (1) was amplified by PCR using human genomic DNA (Boehringer-Mannheim) as template and the primers HPH5 (59-gaacagagagcccagcagagttgggcgg39) and SPH3 (59-CTTCCATGGTGAATCGGTCCCCGCGATG-39). Amplification was performed in a Perkin–Elmer 9700 thermal cycler using the Vent polymerase (Westburg Biolabs) and conditions were as follows: denaturation 5 min 99°C, cycles (353) 96°C 10 s– 65°C 10 s–75°C 50 s. The amplification product, called H800, was then cloned into the pBSKII 1 plasmid (Stratagene) and characterized by cycle sequencing (Perkin–Elmer, Big Dye Terminator). This plasmid was called pBSKII 1 H800. The H800 PCR product was restricted with the NcoI restriction endonuclease (Promega) and cloned into the pGL3-Enhancer vector (Promega) digested with SmaI and NcoI (Promega). The resulting recombinant plasmid was called pH800 and was characterized by cycle sequencing (Perkin–Elmer, Big Dye Terminator). HIF-1a promoter deletions. Deletions were performed using the Exonuclease III (Erase-a-Base system, Promega). Five micrograms of pGLE (PHIF800) was restricted with KpnI (Promega) which generates an EXO III 39 resistant overhang and with NheI (Promega) which generates an EXO III 59 sensitive overhang upstream from the HIF-1a promoter. The plasmid was then incubated with ExoIII in the digestion mix at 4°C (digestion mix: 5 mg plasmid in 10 ml of water, 6 ml of 103 ExoIII buffer (660 mM Tris–HCl, pH 8.0, 6.6 mM MgCl 2), water was added to each to a final volume of 60 ml, and 400 units of EXO III). Every minute 6 ml of the digestion mix was removed and mixed with 7.5 ml of S1 nuclease mix (7 ml of water, 5 ml of S1 buffer (0.3 M potassium acetate, pH 4.6, 2.5 M NaCl, 10 mM ZnSO 4, 50% glycerol), 2.5 Units of S1 nuclease). Each aliquot was kept on ice. Afterward all the samples were placed at room temperature for 30 min. The S1 nuclease removes the remaining single stranded tails. Finally, 1 ml of S1 nuclease stop buffer (0.3 M Tris Base, 0.05 M EDTA) was added to each aliquot and they were incubated at 70°C for 10 min in order to inactivate the S1 nuclease. The plasmids were purified by ethanol precipitation and resuspended in 9 ml of water. The ends of the linear deleted plasmids were flushed using the Klenow DNA polymerase. Therefore, 1 ml of Klenow buffer (200 mM Tris–HCl, 100 mM MgCl 2) was added to the plasmids with 0.2 U of Klenow DNA polymerase and 1 ml of a dNTP
mix (0.125 mM for each dNTP). The solution was then incubated at 37°C for 5 min and the reaction was stopped at 65°C. 40 ml of a ligase mix (31 ml of water, 2.5 ml ligase buffer (Promega), 2.5 ml 50% PEG, 0.25 ml 100 mM DTT, 1 U T4 DNA ligase) were added to each plasmid sample and the solutions were incubated at room temperature during 1 h. XL-1Blue SURE CaCl 2 competent cells were transformed and plated. The positive clones were screened and characterized by PCR screening, plasmid minipreparation followed by ScaI restriction (Promega), and cycle sequencing (Perkin–Elmer ABI PRISM, Big Dye Terminator). Other plasmid vectors. The pGL 3 x HRE vector contains three copies of the EPO HRE (10) downstream from the luciferase gene for which the expression is driven by a heterologous SV40 promoter. The pHdelAaTII vector is the pH800 vector deleted in the HIF1A 59UTR from position 1174 to 1196. The pD5.11. is the pHdelAatII vector deleted in the 59 region of the HIF1A promoter from position 2541 to 2249. Transfection experiments. HMEC-1/HepG2 transfection were performed in a 96 well plate at 90% confluency. The vectors were co-transfected with the control vector pRL-SV40 (Promega) as transfection efficiency control. Prior to transfection, the culture medium was removed and replaced by 200 ml of fresh medium without serum. Three micrograms of DOTAP transfection reagent (BoehringerMannheim) (11) was mixed with 600 ng of assay plasmid and 200 ng of control plasmid to a final volume of 15 ml in 20 mM Hepes buffer. The transfection solution is added to the cells for 3 h. The cell medium was replaced with fresh culture medium containing 2% serum. Twenty-four hours later, luciferase activity is measured. Luciferase activity was quantified in a luminometer using the dualluciferase-reporter system (Promega). For this, the culture medium was removed and 20 ml of PLB 1x (Passive Lysis Buffer, Promega) were added to each well. Firefly luciferase activity was measured after addition of a mix containing the beetle luciferin substrate (LAR I: Luciferase Assay Reagent I, Promega) to the cell lysate. Then renilla luciferase activity was measured by adding a mix containing coelenterazine (Stop & Glo: Renilla Assay Reagent, Promega). Experiments were performed in triplicates. Results are expressed as means of the ratios between firefly luciferase activity and renilla luciferase activity.
RESULTS AND DISCUSSION Northern blot analysis. Northern blot analysis was performed providing a semi-quantitative assay of the HIF-1a mRNA in HMEC-1. HMEC-1 were incubated in normoxia (20% O 2) and in hypoxia (1% O 2) during 2 h 30 min. Using the 59P probe, we showed that in HMEC-1, the HIF-1a mRNA level did not vary in hypoxia in comparison to normoxia (Fig. 1). On the other hand, a low increase of HIF-2 mRNA level was observed in hypoxia. Similar results were obtained using the 39P probe which is specific for HIF-1. HIF-1 activation is induced under hypoxia by posttranslational (2, 4) mechanisms. Northern blot analyses have shown that in a majority of cellular types, there is no variation in HIF1A gene transcription in hypoxia compared to normoxia (8). We report here that in HMEC-1, the HIF-1a messenger is constitutively expressed at the same level in both normoxia and hypoxia. However in the human Hep3B cell line, bovine pulmonary artery endothelial cells, but also, in rat brain and kidney, the HIF-1a mRNA level increases in hypoxia (5– 8). The presence of a transcriptional up-
535
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
tween positions 1180 and 1210 is located a cluster of different putative transcription factor binding sites, with among them: c-Ets-1, NF-kB, and NF-1. Since the HIF1A promoter is a TATA-less promoter (1), these elements all together suggest that the HIF1A transcription could be dependent on cis acting elements located upstream and downstream from the transcription initiation site.
FIG. 1. Comparative Northern blot analysis of HIF-1a mRNA in normoxia and hypoxia in HMEC-1 cells. HMEC-1 were incubated 2h30 in normoxia or in hypoxia. HIF-1a mRNA was detected in northern blot experiments using two probes. The first one called 59P probe hybridized on the 59 end of the HIF-1a mRNA and on the HIF-2 mRNA. The second one, called 39P probe, hybridized on the 39 end of the HIF-1a mRNA. 18S mRNA was used as internal control.
regulation mechanism for HIF-1a seems thus to be species and/or cell-type specific. HIF1A promoter sequence analysis. To identify putative cis-acting elements located within the HIF1A promoter and in the 59UTR, we used the MatInspector analysis software. Results are shown in Fig. 2. In the first 180 bp of the HIF1A promoter sequence, we found putative Sp1, AP-2, HIF-1 binding sites but also binding sites for AP-1 which is a transcription factor activated by phorbol ester and by hypoxia (12). In the 287-bp 59UTR, 2 putative Sp1 binding sites but also AP-1, and NF-kappaB binding sites were found. Be-
Functional analysis of the HIF1A promoter. To characterize the elements involved in the HIF1A gene transcription, a 0.8-kb fragment containing the entire HIF1A 59UTR (287 bp) and the promoter (11 to 2541) was cloned in the pGL3-Enhancer reporter vector using the SmaI and NcoI cloning sites. Then progressive deletions were generated from the 59 to the 39 end of the insert using the exonuclease III. We obtained a population of reporter plasmids containing deletions spanning the whole HIF1A promoter and 59UTR. All these constructions were transfected in HMEC-1 and in HepG2. The cells were then incubated 24 h in normoxia with or without CoCl 2 (13) an agent used to provoke a chemical hypoxia. As shown in Fig. 3A, the first significant transcriptional activity was detected for the plasmid p15C both in normoxia and in hypoxia. This plasmid contains the entire 59UTR sequence and a very short sequence containing at least the initiation site (1). For the next three constructions spanning the 1287 to 2105 region (p8D, p19E, p9F) (Fig. 3A), a low increase in reporter activity was observed in normoxia. However, a strong induction in the reporter activity was demonstrated in hypoxia. But, thereafter the reporter activity for the plasmid p20I (promoter region 1287 to 2201) (Fig. 3A)
FIG. 2. Putative cis elements located within the HIF1A gene 59UTR and the promoter. The HIF1A gene 59 region sequence was analyzed using the MatInspector professional software (http//:genomatix.gsf.de). Underlined sequences represent cis elements producing the highest similarity score. 536
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Functional analysis of the HIF1A promoter. The different constructs were cotransfected together with pRL-SV40 into HMEC-1 (A) and into HepG2 (B). The cells were then incubated 24 h either in normoxia or in the presence of 150 mM CoCl2 inducing a chemical hypoxia. Then the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates. The figures to the left of each construct indicates the length of the promoter numbered 59 from the transcription start site.
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Control for the chemical hypoxia (CoCl 2) efficiency. As positive control of chemical hypoxia induced by CoCl 2, a pGL3 x HRE vector was cotransfected with the pRL-SV40 control vector in HMEC-1 cells. pGL3-SV40 vector was used as a negative control. Cells were incubated 24 h in normoxia or in the presence of 150 mM CoCl 2 to induce chemical hypoxia. Then, the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates.
and the subsequent ones (p7H to pH800) became identical in both hypoxia and normoxia. These results indicate that the HIF1A core promoter sequence is very short, from 11 to 2200 bp, and that the HIF1A promoter fragment spanning the 11 to 2105 bp contains critical cis acting elements involved in the upregulation of the HIF1A gene transcription in hypoxia. This also suggests that within the sequence spanning the 2105 to 2201 bp, a transcriptional downregulator is involved in the decrease of the HIF1A gene expression under hypoxia, bringing back the HIF1A expression level to the one observed in normoxia. Sequence analysis revealed that a putative AP-1 binding site is located within the position 229 to 223, which could be the transcriptional activator because it is known that AP-1 is activated under hypoxia (12). Moreover, an AP-2 cis acting element could play the role of the transcriptional repressor because it is sometimes found in silencers (14). This mechanism could explain why according to the cell type, the HIF1A gene expression is either constitutive in normoxia and in hypoxia like in HMEC-1 or induced in hypoxia. Indeed, if the negative regulator is inactive in cells such as hepatoma, then it may be possible that a HIF1A gene transcriptional upregulation in hypoxia can be observed. However, if the cells contain both the activator and the repressor, then no increase in HIF1A gene transcription is observed in hypoxia. HepG2 hepatoma cells were transfected with the different plasmid constructs in order to confirm this hypothesis. HepG2 cells were then incubated in normoxia or in the presence of CoCl 2. A transcriptional activation also appeared in the presence of CoCl 2 for the p8D construct as for HMEC-1 (Fig. 3B). Interestingly, this CoCl 2-inducer activation is maintained through the whole promoter sequence and no down regulation was observed between the position 2105 and 2201. This result reinforce the hypothesis that the cell type-dependent expression of the HIF1A gene is modulated in hypoxia by
at least two cis elements, the first acting as an activator, the second as a repressor. CoCl 2 chemical hypoxia. We checked that the conditions of chemical hypoxia used here are able to activate HIF-1 using a luciferase reporter gene for which the expression is regulated by 3 HRE sequences (10). This vector, called pGL3 x HRE was cotransfected with the control vector pRL-SV40 in HMEC-1 and these cells were incubated 24 h in the presence of 150 mM CoCl 2. As shown in Fig. 4, the reporter expression was induced in hypoxia but not in normoxia. We conclude that the chemical hypoxia was able to activate HIF-1. Involvement of the 59UTR in HIF1A transcription regulation. Previous results pointed out that the HIF1A gene 59UTR may play a role in HIF1A transcription. The sequence analysis indicated that there are putative nondegenerated protein binding sites within the 59UTR. It is known that in TATA-less promoter cis acting elements are often found downstream from the transcription initiation site (15). Indeed, we detected a significant transcriptional activity in plasmids containing only the 59UTR and the very beginning of the promoter sequence (plasmid 15C). We thus investigated the transcriptional activity of the sequence between position 1120 and 1210. At least 3 clustered putative cis elements in addition to Sp1 binding sites were pointed out by sequence analysis. The region spanning the 1176 to 1198 nucleotides corresponding to the NF-kB/c-Rel, c-ETS-1, next to NF-1 binding sites, was deleted. These new plasmids were called pHdelAatII and pD5.11 according to the length of their 59 end. Both plasmids were transfected in HMEC-1 and were incubated 24 h in normoxia or in hypoxia (150 mM CoCl 2). The luciferase reporter activity was recorded. As shown in (Fig. 5) the level of expression of the reporter plasmid is two to four times weaker for these two plasmids compared to pH800. These results indicate that an enhancer element
538
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. Involvement of the HIF1A 59UTR in HIF1A transcription. HMEC-1 were transfected with pGL3-E, pH800 vector, pHdelAatII vector or pD.5.11. vector. The pHdelAaTII vector is deleted in the HIF1A 59UTR between position 1176 to 1198 and the pD5.11. vector is deleted between positions 2541 to 2249 and positions 1176 to 1198. Cells were incubated 24 h in normoxia or in the presence of 150 mM CoCl 2 to induce chemical hypoxia. Then, the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates. The figures to the left of each construct indicates the length of the promoter numbered 59 from the transcription start site.
within the 59UTR (1180 to 1200) is involved in the HIF1A gene transcription. Indeed, it is well known that in TATA-less promoter, cis acting elements are frequently found downstream from the transcription initiation site (15). As examples of such genes we can quote: gamma interferon (14), the chemokine receptor BLR1 (15), phosphofructokinase P1 (16). Negative and positive cis acting elements are also located in the first exon of the prostaglandin H synthase-2 gene (17). We still have to define which transcription factors bind to the 59UTR region of the HIF1A gene.
The Sp1 factor constitutively activates the HIF1A transcription. Several putative Sp1 binding sites are located within the HIF1A promoter. It is known that in TATA-less promoters, transcription is partially dependent on Sp1 factors (16). We tested mithramycine A, a Sp1 DNA binding inhibitor, on the transcriptional activity of pH800 and p15C in HMEC-1. An inhibition of the reporter expression was observed for pH800, but not for p15C which contains the 59UTR of the HIF1A gene and a short promoter sequence without Sp1 binding sites (Fig. 6). This result suggests that at least some of Sp1 binding sites are active within the HIF1A promoter on the contrary to the Sp1 binding sites located within in the 59UTR. CONCLUSION
FIG. 6. HIF1A transcription is dependent on Sp1 transcription factor. HMEC-1 were transfected with pGL3-E, pH800 or p15C vector. These constructs were cotransfected with the control vector pRL-SV40. The cells were incubated 24 h in the presence of mithramycine A 1 mM. Mithramycine A is a Sp1 DNA binding inhibitor. Then the cells were lysed for luciferase assays. Data represent the ratio between test firefly luciferase activity and renilla luciferase activity (RF/R). Results are presented as means 6 SD for triplicates.
In endothelial HMEC-1 cells, the HIF1A expression is highly constitutive in both hypoxia and normoxia. This expression is driven by a short promoter sequence (11 to 2200 bp), at least by Sp1 binding sites and by cis-acting elements located within the 59UTR region. Moreover, in hypoxia the expression level of the HIF1A gene is maintained identical to the one observed in normoxia due to the presence of a cis-repressor acting as an antagonist on an upregulating element located next to the transcriptional initiation site. The transcription factors involved in these regulations are still to be defined. ACKNOWLEDGMENTS E. Minet is a research assistant and C. Michiels is a research associate of the FNRS (Fonds National de la Recherche Scientifique, Belgium). G. Michel is a fellow of FRIA (Industrial and Agronomic
539
Vol. 261, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Research Fund, Belgium). This work was partly supported by FRFC and SSTC. This text presents results of the Belgian Programme on Interuniversity Poles of Attraction initiated by the Belgian State, Prime Minister’s Office, Science Policy Programming. Scientific responsibility is assumed by the authors. The authors thank the Re´gion Wallonne for financial support. We are grateful to Dr. F. Candal (CDC, Atlanta, GA) for providing HMEC-1.
REFERENCES 1. Iyer, N. V., Leung, S. W., and Semenza, G. L. (1998) Genomics 52, 159 –165. 2. Semenza, G. L. (1998) Curr. Opin. Genet. Dev. 8, 588 –594. 3. Wang, G. L., Jiang, B.-H., Rue, E. A., and Semenza, G. L. (1995) Proc. Natl. Acad. Sci. USA 92, 5510 –5514. 4. Salceda, S., and Caro, J. (1997) J. Biol. Chem. 272, 22642–22647. 5. Ladoux, A., and Frelin, C. (1997) Biochem. Biophys. Res. Commun. 240, 552–556. 6. Wiener, C. M., Booth, G., and Semenza, G. L. (1996) Biochem. Biophys. Res. Commun. 225, 485– 488. 7. Palmer, L. A., Semenza, G. L., Stoler, M. H., and Johns, R. A. (1998) Am. J. Physiol. 274, L212–L219.
8. Wenger, R. H., Kvietikova, I., Roles, A., Gassmann, M., and Marti, H. H. (1997) Kidney Int. 51, 560 –563. 9. Ades, E. W., Candal, F. J., Swerlick, R. A., George, V. G., Summers, S., Bosse, D. C., and Lawley, T. J. (1992) J. Invest. Dermatol. 99, 683– 690. 10. Kvietikova, I., Wenger, R., Marti, H. H., and Gassmann, M. (1997) Kidney Int. 51, 564 –566. 11. Schmedtje, J. F., Liu, W. L., Ji, Y., Thompson, T. M., and Runge, M. S. (1996) Biochem. Biophys. Res. Commun. 220, 687– 691. 12. Rajpurohit, R., Koch, C. J., Tao, Z., Teixeira, C. M., and Shapiro, I. M. (1996) J. Cell Physiol. 168, 424 – 432. 13. Bunn, H. F., and Poyton, R. O. (1996) Physiol. Rev. 76, 839 – 885. 14. Ye, J., Cipitelli, M., Dorman, L., Ortaldo, J. R., and Young, H. A. (1996) Mol. Cell. Biol. 16(9), 4744 – 4753. 15. Wolf, I., Pevzner, V., Kaiser, E., Bernhardt, G., Claudio, E., Siebenlist, U., Forster, R., and Lipp, M. (1998) J. Biol. Chem. 273(44), 28831–28836. 16. Johnson, J. L., and McLachlan, A. (1994) Nucleic Acids Res. 22, 5085–5092. 17. Tazawa, R., Xu, X. M., Wu, K. K., and Wang, L. H. (1994) Biochem. Biophys. Res. Commun. 203, 190 –199.
540