Common Proteins Bind mRNAs Encoding Erythropoietin, Tyrosine Hydroxylase, and Vascular Endothelial Growth Factor

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.

246, 436–440 (1998)

RC988639

Common Proteins Bind mRNAs Encoding Erythropoietin, Tyrosine Hydroxylase, and Vascular Endothelial Growth Factor1 Aline B. Scandurro2 and Barbara S. Beckman3 Department of Pharmacology, Tulane Cancer Center, Tulane University School of Medicine, New Orleans, Louisiana 70112

Received March 23, 1998

The hypoxia-inducible genes erythropoietin (Epo), tyrosine hydroxylase (TH), and vascular endothelial growth factor (VEGF) are regulated post-transcriptionally by proteins binding to specific regions located in the 3* untranslated region (UTR) of their mRNAs. To determine whether trans-factors binding to this region in all three of these RNAs are similar, we generated riboprobes containing the 3* UTR of erythropoietin, tyrosine hydroxylase, and vascular endothelial growth factor mRNA and assayed them by electrophoretic mobility shift assay (EMSA) and UV cross-linking experiments. Each riboprobe formed similar shifted protein complexes using human hepatoma cell (Hep3B) cytoplasmic lysates in the EMSA. Hep3B proteins bound to each probe could be cross-competed by the specific unlabeled Epo, TH, or VEGF riboprobes. By contrast, a non-specific 3* UTR riboprobe did not compete for binding with the Epo, TH, or VEGF RNA shifted protein complexes. UV cross-linking studies revealed proteins of similar molecular weights for the Epo, TH, and VEGF RNA shifted protein complexes. Taken together, these results suggest a common posttranscriptional regulatory mechanism for hypoxia-inducible genes. q 1998 Academic Press

Oxygen-sensing mechanisms are among the most basic of all environmental sensing systems. In vertebrates, detection of changes in oxygen tension is essential to avoid the consequences of tissue hypoxia (1). Important responses to chronic hypoxia in mammals include in1 This work was supported in part by National Institutes of Health Grant DK40501. 2 Recipient of an Individual National Research Service Award (DK09170-02). 3 To whom reprint requests should be addressed at Department of Pharmacology, Tulane University School of Medicine, 1430 Tulane Ave., New Orleans, LA 70112. Fax: (504) 588-5283. E-mail: alibscan@ mailhost.tcs.tulane.edu.

0006-291X/98 $25.00

creased red blood cell production (1), catecholamine synthesis (2), and angiogenesis (1). These processes are principally mediated by Epo,4 TH and VEGF, respectively. The genes encoding Epo (1), TH (3) and VEGF (4) are up-regulated by low oxygen tension and thus serve as models for studying the genetic mechanisms of hypoxia response pathways. In the case of Epo, a physiologic drop in oxygen tension leads to a 50 to 100fold induction of gene expression (5,6). Epo mRNA transcription increases only 10-fold as a result of hypoxia; thus it has been postulated, that post-transcriptional stabilization of the normally labile Epo mRNA may account for the observed increased Epo levels (6). Investigations in this laboratory have identified, a complex of proteins, erythropoietin mRNA binding protein (ERBP), in cytoplasmic lysates of Hep3B cells which specifically binds to a 120 nucleotide (nt) region in the 3* UTR of Epo mRNA (7). Additionally, a stabilizing role has been suggested for this region from studies in which deletion of this 120 nt region leads to an unchanged mRNA halflife in response to hypoxia (6 hrs) compared to a forty percent increase in half-life observed for the wild-type mRNA (8). Oxygen-sensing is likely to be mediated posttranscriptionally by specific binding of ERBP to the 3* UTR of the Epo mRNA. Goldberg and Schneider suggested that a 9 nt sequence (CCTCCCTCA) found in the 3*UTR of Epo and VEGF is important in oxygen-sensing (1). This 9mer is contained within the 120 nt ERBP binding site. Recently, the 3*UTR region of VEGF was shown by EMSA to bind a protein from hypoxia-stimulated pheochromocytoma cells (PC12) (9). An hypoxia-inducible 3*UTR 4

Abbreviations used: Epo, erythropoietin; ERBP, erythropoietin mRNA binding protein; HIPBS, hypoxia-inducible protein binding site; HI-RPB, hypoxia-inducible RNA-protein binding; kDa, kiloDalton; nt, nucleotide; PCR, polymerase chain reaction; SDS-PAGE, sodium dodecyl sulphate-polyacrylamide gel electrophoresis; TH, tyrosine hydroxylase; UTR, untranslated region; VEGF, vascular endothelial growth factor.

436

Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.

AID

BBRC 8639

/

6952$$$561

05-04-98 20:38:17

bbrcg

AP: BBRC

Vol. 246, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 1. (A) Map of Epo 3*UTR and Epo riboprobes. Shown is a diagram of the 3* untranslated region of the Epo gene (nt numbers correspond to GenBank Accession Number X02157). The previously mapped ERBP binding site and the corresponding riboprobes generated for this study as described in the Experimental Procedures are also included. (B) Characterization of Epo RNA-ERBP complex formation. Hep3B cytoplasmic lysates from near confluent cultures were prepared by freeze-thaw lysis in 25 mM Tris (pH 7.9) and 0.5 mM EDTA, followed by centrifugation at 15,000 1 g at 47C for 10 min. Increasing amounts of lysate 1-10 mg was added in the gel shift assay as described in the Experimental Procedures. Proteinase K (20 mg) and RNase T1 (20 U) treatments for 30 min at 377C preceded the gel shift assay. (C) Competition experiment. Competitor unlabeled riboprobes (10-, 50-, 100-fold molar excess) Epo (SB) or (100-fold) Epo (NB) were added to the gel shift mixture 10 min prior to radiolabeled riboprobes. Gel shift assay performed as above in the presence of competitor unlabeled riboprobes as indicated. (D) Hypoxia upregulates ERBP-Epo RNA shifted complex. Gel shift assay as above with lysates prepared from Hep3B cells grown for 24 hrs in 21% O2 (N) or 1% O2 (H) as indicated. Shown is a representative autoradiogram demonstrating upregulation by hypoxia (1.4-fold) of the ERBP-Epo RNA shifted complex as determined by densitometric analysis. In each case, arrow indicates ERBP shifted complex.

TH RNA binding protein has also been described in PC12 cells (2). Mutational analysis of the 3*UTR region of the TH RNA identified crucial pyrimidine-rich sequences within the 3*UTR region in the binding of this protein (10). Several hypoxia-inducible genes including Epo and VEGF contain a consensus hypoxia-inducible protein binding site (HIPBS), (U/C)(C/U)CCCU. Subsequently, more detailed mapping of the human and rat 3*UTR of VEGF mRNA has revealed at least five hypoxia-inducible RNA-protein binding sites (HI-RPB IV)(11). Of the five, HI-RPB II is adjacent to the 9mer and HIBPS. A definitive analysis of cis - acting mRNA stability determinants for hypoxia-inducible genes awaits further study. The present study provides evi-

dence that similar trans -acting factors bind to the 3*UTR of the mRNAs for the hypoxia-inducible genes, Epo, TH and VEGF, and suggests a common post-transcriptional mechanism for oxygen-sensing. EXPERIMENTAL PROCEDURES Cell Culture Human hepatoblastoma (Hep3B) cells obtained from the American Type Culture Collection (ATCC, Rockville, MD) were routinely cultured in Eagle’s minimal essential medium (MEM) supplemented with 10% fetal bovine serum (FBS), 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 U/ml penicillin G, and 100 mg/ml streptomycin in a humidified atmosphere of 21% O2 , 5% CO2 , 74%

437

AID

BBRC 8639

/

6952$$$562

05-04-98 20:38:17

bbrcg

AP: BBRC

Vol. 246, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 2. (A) Epo, TH, and VEGF 3 *UTR bind a similar complex of proteins in Hep3B cell lysates. Hep3B lysates were incubated for 10 min with Epo (SB), Epo (NB), TH and VEGF radiolabeled probes as indicated and band-shift assay. As expected no shift was observed when radiolabeled Epo (NB) RNA transcript was used in the reaction. (B) Competition of shifted RNA complexes. Inhibition of formation of the TH and VEGF RNA-protein complex by competitor unlabeled (10-, 50-, 100- fold molar excess) Epo (SB) or TH (100-fold) or VEGF (100-fold) and not Epo (NB) (100-fold), as indicated.

N2 at 377C. Hypoxic conditions were obtained by placing cells in a chamber (Billups-Rothenburg, Del Mar, CA) designed to maintain low oxygen tension (1% O2 , 5% CO2 , 94% N2).

XbaI and KpnI digestion and religation downstream from the Sp6 RNA polymerase promoter within pGEM-1. The resulting TH transcript after in vitro Sp6 RNA polymerization (Promega) was approximately 317 nt long.

Preparation of Lysates

Vascular endothelial growth factor. A 489 bp cDNA fragment containing the rat VEGF 3 *UTR was amplified by PCR with primers containing the T7 RNA polymerase promoter as described above (a generous gift from Dr. Andrew P. Levy and Dr. Mark A. Goldberg). The resulting VEGF RNA transcript synthesized was 458 nt long.

Cytoplasmic lysates of Hep3B cells were prepared by freezing and thawing through repetitive cycles as described (7). The protein concentration of each lysate was determined by the Bradford assay (BioRad) using bovine serum albumin as standard.

Electrophoretic Mobility Shift Assay

RNA Probes Erythropoietin. Radiolabeled and unlabeled riboprobes were synthesized in vitro using bacteriophage RNA polymerases according to manufacturer’s recommendations (Promega). Two riboprobes spanning the Epo*UTR were prepared (Fig. 1A). Briefly, T7 RNA polymerase promoter containing Epo DNA fragments were generated by PCR. The 5* primers were designed to contain the T7 RNA polymerase recognition sequence in addition to Epo DNA sequences. The riboprobe designated Epo (SB), contains the published ERBP-specific binding site (corresponding to nt 768 to 891 from GenBank Accession Number X02157) and was prepared with the following primers: T7 polymerase promoter/5*-GGATCCTAATACGACTCACTATAGGGGTGTGTCCACCTGGGCAT and 3*-AGTGTCCATGGGACAGGC. Epo (NB, non-ERBP binding) was transcribed as a negative control and corresponds to sequences (nt 878 to 1002) adjacent to but not including the ERBP-specific binding site. It was prepared as per Epo (SB) using the following primers: T7 polymerase promoter/5*-GGATCCTAATACGACTCACTATAGGGGTCCCATGGACACTCCAG and 3*CTGAATGCTT CCTGCTCT. Following PCR and agarose gel purification, the isolated DNAs were transcribed in vitro with T7 RNA polymerase (Promega). Both riboprobes resulted in 120 nt transcripts after T7 RNA polymerization. Tyrosine hydroxylase. The entire rat TH gene cloned into the pGEM-1 vector (Promega, Madison, WI) was obtained from Dr. Kent E. Vrana (Wake Forest University). The vector was modified for the purposes of this study. The complete 3*UTR of TH remained after

Two to ten micrograms of cytoplasmic lysate was incubated with 51104 cpm of RNA transcript and specific or non-specific competitor RNA as described previously (7). After drying, the polyacrylamide gel was exposed overnight to Hyperfilm MP (Amersham) or to a phosphorimaging screen (0.5-2 hrs, Fuji). Cold competitor RNA transcripts were added 10 min prior to labeled probe in the assay.

UV Cross-Linking Studies The gel shift assay mixture was placed on ice and exposed to UV light (254nm) with the settings at 125mJ for 10 min (or as indicated) at a distance of 6 cm in a BioRad GS Gene Linker. Gel shift mixtures were treated with RNase T1 as indicated above. Laemmli sample buffer was added to the samples and subsequently samples were denatured at 100 7C for 3 min prior to loading on an SDS-PAGE. After SDS-PAGE analysis the gel was dried and exposed to a phosphorimager board (Fuji) or Hyperfilm MP (Amersham). The molecular weight of protein bands was determined after scanning of the autoradiograph by the Molecular Analyst program (BioRad ver. 2.0) with a prestained marker (cat.#161-0318, BioRad) as standard.

RESULTS AND DISCUSSION Characterization of ERBP-Epo RNA binding. To characterize factors responsible for the stabilization of

438

AID

BBRC 8639

/

6952$$$562

05-04-98 20:38:17

bbrcg

AP: BBRC

Vol. 246, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 3. (A) UV cross-linking studies. The gel shift assay mixture was placed on ice and exposed to UV light (254 nm) with the settings at 125mJ for 1, 2, 5, or 10 min. After UV-treatment mixtures were treated with RNase T1 as indicated above before loading on an 8% SDSpolyacrylamide gel, drying and exposure to phosphorimaging screen. Molecular weights are indicated in kDa on the left side. Arrows indicate shared 115, 78, 67, 49, 45 and 30 kDa cross-linked proteins. (B) The gel shift assay mixture was UV cross-linked for 10 min. as above after competition with 100-fold molar excess of either Epo (SB) or Epo (NB) unlabeled riboprobe and separation on a 12% SDS-polyacrylamide gel. Arrow indicates specific ca. 30 kDa Epo RNA binding protein.

mRNA of the hypoxia responsive genes, two-120 nt transcripts of the erythropoietin 3* UTR were transcribed in vitro by T7 RNA polymerase (Fig. 1A). As was previously reported (7) and shown in Fig. 1B binding of Hep3B cytoplasmic proteins to Epo RNA was specific for the ERBP specific binding region (nt 768 to 891) present in the Epo (SB) riboprobe. By comparison, ERBP-Epo RNA complex formation was not observed with the Epo (NB) riboprobe (nt 878 to 1002) corresponding to the region downstream from the ERBP binding site. Incubation with increasing amounts of cytoplasmic lysate (1-10 mg) resulted in greater ERBP -Epo RNA complex formation. Incubation with lysate previously treated with proteinase K or probes treated with RNase T1 greatly diminished complex formation indicating that complex formation is both protein and RNA-dependent. Longer incubation of the lysate or probe with proteinase K or RNase T1 (respectively) completely abolished complex formation (data not shown). In addition to the characteristic Epo RNA-ERBP complex formation a slower migrating band is occasionally observed. This complex may be related to the primarily shifted complex and seems to appear with increased protein concentrations and low salt conditions. The Epo RNA-ERBP complex could be competed with increasing amounts (10-, 50- or 100- fold molar excess) of unlabeled Epo(SB) RNA but not with 100-fold molar excess of the unlabeled Epo(NB) riboprobe (Fig. 1C). Increased complex formation was consistently observed with cytoplasmic Hep3B lysates prepared from cells exposed to hypoxia (1%, 4-24h) (Fig. 1D) (2,4).

whether the 3*UTR RNA binding proteins for the three hypoxia-inducible genes are similar, 3*UTR riboprobes for Epo, TH and VEGF were prepared and examined for their ability to bind proteins from Hep3B cytoplasmic lysates. As indicated in Fig. 2A, all three radiolabeled riboprobes resulted in the characteristic Epo RNAERBP shift when combined with Hep3B lysates (7,12). Epo, TH and VEGF RNA-protein complexes appear to share the characteristic Epo RNA-ERBP band, although some differences were noted. TH RNA-protein complexes preferentially form a slower migrating band. Likewise the complexes formed with VEGF RNA included the shared characteristic migrating species but additionally there was an even faster migrating shifted complex that is unique to VEGF RNA. This complex could not be disrupted with excess unlabeled Epo (SB) (see fig. 2B) or TH transcripts (data not shown). The differences in the gel shift pattern might result from differences in the proteins recruited to the proteinRNA complex, which may confer specificity to an otherwise global genetic sensing system. Evidence for common factors binding the 3*UTR of these genes is further supported by the observation that the Epo (SB) riboprobe shifted protein complex could be inhibited with increasing amounts (10-, 50- and 100-fold molar excess) of unlabeled TH and VEGF riboprobes. As expected each of the unlabeled transcripts competed its own riboprobe more efficiently (Fig. 2B). Competition experiments also demonstrated that Epo (SB), TH and VEGF but not the Epo (NB) riboprobe cross-competed with each other (data not shown).

Epo, TH, and VEGF 3*UTR bind a similar complex of proteins from Hep3B cell lysates. To determine

UV cross-linking studies. To examine whether the Epo, TH and VEGF riboprobes bound similar molecular

439

AID

BBRC 8639

/

6952$$$562

05-04-98 20:38:17

bbrcg

AP: BBRC

Vol. 246, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

weight proteins, gel band shift reaction mixtures were UV cross-linked as described in the Experimental Procedures. The approximate molecular weight of the proteins bound to each of the Epo (SB), TH or VEGF radiolabeled transcripts was determined after UV crosslinking of the complexes and separation on an 8% SDS polyacrylamide gel. As is represented in Fig. 3A there were similar molecular weight RNA binding proteins (ca. 115, 78, 67, 49, and 45 kDa) associated with each of the riboprobes. Shorter film exposure to the gel also revealed a shared lower molecular weight protein (ca. 30 kDa). Observed differences among these riboprobes may reflect affinity differences of each protein for each riboprobe. Incubation of the reaction with the unlabeled riboprobe Epo (SB) but not the unlabeled Epo (NB) riboprobe demonstrated that a ca. 30 kDa protein specifically bound the 3*UTR (Fig. 3B). Evidence presented here suggests the involvement of common factors for mRNA-protein complex formation in hypoxia-inducible gene expression. Importantly, this global sensing mechanism occurs at the post-transcriptional and not at the transcriptional level. Sensing at this level may represent an initial rapid response facilitating adaptation to a hypoxic environment. Similarly, mitogenic stimuli lead to common post-transcriptional regulatory events of cytokine mRNAs orchestrated by a related set of 3* UTR AU-rich nucleotide binding factors (13). Taken together, these studies suggest a novel target to exploit in the genetic manipulation of cells and underscores the importance of further understanding post-transcriptional mechanisms.

ACKNOWLEDGMENTS We thank Kent Vrana, Andy Levy, Mark Goldberg, and Gideon Dreyfuss for kindly and freely providing their reagents. We appreciate the helpful suggestions given by Jawed Alam, Tim Hammond, Chuck Hemenway, Jim Kaysen, and Laura Levy.

REFERENCES 1. Goldberg, M. A., and Schneider, T. J. (1994) J. Biol. Chem. 269, 4355–4359. 2. Czyzyk-Krzeska, M. F., Dominski, Z., Kole, R., and Millhorn, D. E. (1994) J. Biol. Chem. 269, 9940–9945. 3. Czyzyk-Krzeska, M. F., Furnari, B. A., Lawson, E. E., and Millhorn, D. E. (1994) J. Biol. Chem. 269, 760–764. 4. Levy, A. P., Levy, N. S., Wegner, S., and Goldberg, M. A. (1995) J. Biol. Chem. 270, 13333–13340. 5. Goldberg, M. A., Glass, G. A., Cunningham, J. M., and Bunn, H. F. (1987) Proc. Natl. Acad. Sci. 84, 7972–7976. 6. Goldberg, M. A., Gaut, C. C., and Bunn, H. F. (1991) Blood 77, 271–277. 7. Rondon, I. J., MacMillan, L. A., Beckman, B. S., Goldberg, M. A., Schneider, T., Bunn, H. F., and Malter, J. S. (1991) J. Biol. Chem. 266, 16594–16598. 8. McGary, E. C., Rondon, I. J., and Beckman, B. S. (1997) J. Biol. Chem. 272, 8628–8634. 9. Levy, A. P., Levy, N. S., and Goldberg, M. A. (1996) J. Biol. Chem. 271, 2746–2753. 10. Czyzyk-Krzeska, M. F., and Beresh, J. E. (1996) J. Biol. Chem. 271, 3293–3299. 11. Levy, N. S., Goldberg, M. A., and Levy, A. P. (1997) Biochim. Biophys. Acta 1352, 167–173. 12. Rondon, I. J., Scandurro, A. B., Wilson, R. B., and Beckman, B. S. (1995) FEBS Letters 359, 267–270. 13. Hamilton, B. J., Nagy, E., Malter, J. S., Arrick, B. A., and Rigby, W. F. (1993) J. Biol. Chem. 268, 8881–8887.

440

AID

BBRC 8639

/

6952$$$562

05-04-98 20:38:17

bbrcg

AP: BBRC