Post-transcriptional Regulation of Erythropoietin mRNA Stability by Erythropoietin mRNA-binding Protein*

(Received for publication, November 27, 1996, and in revised form, January 7, 1997)

Eric C. McGary Dagger , Isaac J. Rondon § and Barbara S. Beckman

From the Department of Pharmacology and the Tulane Cancer Center, Tulane University School of Medicine, New Orleans, Louisiana 70112

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

We have previously identified a sequence in the 3'-untranslated region (3'-UTR) of erythropoietin (Epo) mRNA which binds a protein(s), erythropoietin mRNA-binding protein (ERBP). A mutant lacking the ERBP binding site (EpoM) was generated. Hep3B cells were stably transfected with a wild-type Epo (EpoWT) cDNA or EpoM cDNA construct located downstream of a promoter of cytomegalovirus. Following inhibition of transcription, the half-lives of EpoWT and EpoM mRNAs were 7 h and 2.5 h in normoxia, respectively. The EpoM mRNA half-life remained unchanged in hypoxia. EpoWT mRNA half-life increased ~40% in response to a 6-h hypoxic pre-exposure and an additional ~50% when pre-exposed to 12 h hypoxia. The steady-state level of EpoWT mRNA was 4-fold that of EpoM mRNA reflecting the difference in mRNA decay rates in normoxia. The Epo protein level expressed from exogenous EpoM was unchanged in both normoxia and hypoxia. In contrast, the Epo protein level expressed from exogenous EpoWT increased 50% in hypoxia when compared with normoxia. These observations were further supported by chimeric chloramphenicol acetyltransferase and Epo-3'-UTR constructs. We have demonstrated that Epo mRNA stability was modulated in normoxia and further by hypoxia, therefore, providing evidence that Epo is regulated at the post-transcriptional level through ERBP complex formation.


INTRODUCTION

Erythropoietin is a glycosylated protein that plays a central role in erythropoiesis by supporting the proliferation and differentiation of erythroid progenitor cells in the bone marrow (1). Reduced oxygen tension (hypoxia) is the principle physiological stimulus for Epo1 production in the fetal liver (2) and in the kidney after birth (3-5). The human hepatoblastoma cell lines, HepG2 and Hep3B, have been used extensively to study Epo gene regulation because of their ability to secrete Epo in response to hypoxia and cobalt chloride in a physiological manner (6).

Comparison of human and mouse Epo gene sequences has provided information regarding highly conserved cis-regulatory elements which exist within the 5'UTR, the first intron, and 3'- and 5'-flanking regions of the Epo gene (7-9). Analysis of these conserved sequences has provided abundant information regarding tissue- and stimulus-specific gene expression of Epo (10). Using reporter gene constructs, a 3' enhancer element flanking the Epo gene as well as a 5' minimal sequence required for basal promoter activity have each been shown to confer a 4- to 14-fold induction of Epo gene transcription in response to hypoxia in Hep3B cells (11). Furthermore, these elements appear to act synergistically by causing a 40- to 50-fold induction in transcription in response to hypoxia (11, 12). This induction in transcription leads to an increase in the Epo mRNA steady-state level throughout the first 8 h after which period the Epo mRNA level decreases (13). Interestingly, the increase in Epo mRNA level in response to hypoxia cannot be accounted for by a 5- to 10-fold increase in transcription rate in response to hypoxia (14, 15). Furthermore, while Epo mRNA levels decrease during prolonged hypoxia, the Epo protein level continues to increase for up to 24 h following hypoxic induction (6). Taken together, these findings suggest that Epo gene expression may be partly regulated at the post-transcriptional level.

Several genes which appear to be regulated by hypoxia are regulated at the post-transcriptional level (10, 16). Using Northern blot analysis and quantitative polymerase chain reaction (PCR), the half-life of Epo mRNA in normoxia was determined to be 1.5-2 h (13, 14). In experiments using actinomycin D to block transcription, the half-life of Epo mRNA was extended to 7-8 h in both normoxia and hypoxia (14). In addition, Epo mRNA half-lives were also prolonged in Hep3B cells treated with cyclohexamide (14). These observations led Goldberg and co-workers (12, 14) to suggest that Epo mRNA stability may be modulated by a rapidly turning over destabilizing protein.

We have previously reported that a protein(s), erythropoietin mRNA-binding protein, specifically binds to a pyrimidine-rich 120-bp region in the 3'UTR of Epo mRNA (17). We have also demonstrated that in vitro ERBP complex formation was increased using cytoplasmic lysates of various tissues from mice exposed to hypoxia (17). In this study, we investigate the effect of ERBP complex formation on Epo mRNA stability in normoxia and hypoxia. A mutated Epo gene with a deleted ERBP binding region was generated. In addition, both wild-type and mutant Epo constructs lack hypoxic responsive elements such that changes in mRNA stability in hypoxia would only reflect hypoxic modulation of ERBP complex formation. We are fully aware that actinomycin D prolongs the half-life of Epo mRNA (14). However, actinomycin D may be employed in this study to determine the effect of ERBP complex formation on Epo mRNA stability, if ERBP complex formation is not affected by actinomycin D. Moreover, this study does not attempt to determine the physiological half-life of either the Epo wild-type or the mutated Epo mRNA, but rather to compare the differences in their mRNA stabilities. The mRNA half-lives, steady-state mRNA levels, and protein levels of Hep3B cells stably transfected with either the exogenous wild-type Epo gene or the mutant Epo gene and exposed to either normoxia or hypoxia were compared. Furthermore, the effect of ERBP complex formation on the mRNA stability of a heterologous gene, chloramphenicol acetyltransferase (CAT), was examined.


EXPERIMENTAL PROCEDURES

Cell Culture

Human hepatoblastoma (Hep3B) cells were obtained from the American Type Culture Collection and grown in Eagle's minimal essential medium (Life Technologies, Inc.) supplemented with 100 units/ml penicillin, 100 units/ml streptomycin, 0.1 mM nonessential amino acids, 0.2 mM glutamine, 1 mM pyruvate, and 10% fetal bovine serum (Cellgro) and were maintained in a humidified 5% CO2 incubator at 37 °C. Cells exposed to hypoxia were grown in a controlled atmosphere chamber supplied with a constant flow of a hydrated 1% O2, 5% CO2, balanced N2 gas mixture (18). All experiments were performed as cells approached 70-80% confluency.

Constructs and Establishment of Transfected Cell Lines

Human Epo cDNA cloned into pSP70 DNA (Promega) (17) was used as a DNA template in PCR mutagenesis to produce a mutated 3'UTR Epo sequence which lacks the ERBP binding sequence. All nucleotide positions were based on the sequence of accession number M11319[GenBank] in GenBankTM. The 104-bp region between the stop codon (nucleotide 2770-2772) and the NcoI site (nucleotide 2786-2889) containing the wild-type ERBP binding sequence was replaced with a 17-bp sequence (5'-GTTGCATTGTTACAGGA-3'), and the nucleotide sequence of the mutated region was confirmed using the Sequenase® version 1 kit (U. S. Biochemical Corp.). Subsequently, the StuI (nucleotide 2751-2756) and NcoI in EpoWT was replaced with the mutated ERBP binding sequence to create the Epo construct with the mutated 3'UTR (EpoM). An ~1.4-kilobase EcoRI-digested fragment carrying the EpoM or EpoWT cDNA sequence was then cloned into the EcoRI site of the multiple cloning site of a mammalian expression vector, pcDNA3 (Invitrogen). The vector pcDNA3 contains the enhancer-promoter sequences from the immediate early gene of the human cytomegalovirus located upstream of the multiple cloning site. The polyadenylation signal and transcription termination sequence from the bovine growth hormone gene are located downstream of the multiple cloning site, and the vector also contains a neomycin resistance gene for selection of G418-resistant stable cell lines. Two clones with the T7 RNA polymerase promoter upstream of the Epo cDNA sequence were identified and designated as pcDNA3-EpoM and pcDNA3-EpoWT. Primer p221 (5'-CCACTCTGCTTCGGGCTCTGGG-3'), located upstream of the StuI site, and primer p220 (5'-CCTGAGATGTCATTGCTGGCAC-3'), located downstream of the NcoI site, were used to distinguish EpoM from EpoWT.

The 650- and 563-bp StuI and XhoI fragments containing the wild-type Epo 3'UTR and the mutated Epo 3'UTR, respectively, were subcloned between KpnI (the 3'-overhang was flushed with Escherichia coli DNA polymerase I) and XhoI, downstream of the CAT coding sequence, of a mammalian expression vector, pcDNA3-CAT (Invitrogen). A 787-bp fragment containing the CAT coding sequence has been cloned into the HindIII in the multiple cloning site of pcDNA3 to generate pcDNA3-CAT. These CAT and Epo-3'UTR fusion constructs were designated pCAT-EpoWT and pCAT-EpoM.

Hep3B cells were stably transfected with pcDNA3-EpoWT, pcDNA3-EpoM, pcDNA3-CAT, pCAT-EpoWT, or pCAT-EpoM using Lipofectin Reagent® (Life Technologies, Inc.) as described by the manufacturer. Transfected cells were incubated 36-72 h in supplemented minimal essential medium after which time 0.45 mg/ml G418 was added to select transformed cells. Stable transfectants from three different transfections were pooled after 30 days to avoid bias in gene expression due to variable sites of chromosomal integration.

Electromobility Shift Assay

Hep3B cells were grown in the absence or presence of 5 µg/ml actinomycin D (Sigma). Cell lysates used in electromobility shift assays were prepared from Hep3B cells which were pelleted, washed, resuspended, and lysed in 25 mM Tris-HCl (pH 7.9) and 0.5 mM EDTA by four freeze-thaw cycles. Following cell lysis, the samples were centrifuged at 10,000 × g at 4 °C for 10 min, and the supernatant was removed and stored at -70 °C. The protein concentrations of the cell lysates were determined by a Bradford protein assay as described by the manufacturer (Bio-Rad).

Full-length sense-strand radiolabeled RNA run-off transcripts were generated by incubating 5 units of either SP6 or T7 RNA polymerase (Promega) with 1 µg of linearized EpoM or EpoWT DNA template in the presence of 40 mM Tris-HCl (pH 7.9), 6 mM MgCl2, 40 units of RNasin (Promega), 0.4 mM ATP, 0.4 mM GTP, 0.4 mM CTP, 0.25 µM UTP, [32P]UTP (3000 Ci/mmol, ICN), and 1 mM dithiothreitol for 2 h at 37 °C, after which time 1 unit of RNase-free DNase I (Promega) was added and the samples were incubated an additional 30 min at 37 °C. Following phenol-chloroform extraction, the aqueous phase was passed through a 1-ml Sephadex G-50 spin column (Boehringer Mannheim), precipitated with ethanol, resuspended in RNase-free water, and counted in a liquid scintillation counter. An aliquot of 2 µg of the lysate was incubated with 5 × 104 cpm of EpoM or EpoWT run-off transcripts in the presence of 12 mM HEPES (pH 7.9), 10% glycerol, 15 mM KCl, 0.25 mM EDTA, 0.25 mM dithiothreitol, 5 mM MgCl2, and 200 ng/µl E. coli tRNA for 15 min at room temperature. Twenty units of RNase T1 (Life Technologies, Inc.) was added, and reaction mixtures were incubated for an additional 30 min at room temperature prior to electrophoresis in a 7% native polyacrylamide gel.

Actinomycin D Treatment for Determination of mRNA Stability

Hep3B cells stably transfected with either pcDNA3-EpoWT or pcDNA3-EpoM were grown in either normoxia or hypoxia (6 h or 12 h) prior to the addition of 5 µg/ml actinomycin D (Sigma) (14). Following the addition of actinomycin D, cells remained in either normoxia or hypoxia for various periods of time after which total RNA was extracted with RNAzol B (Tel-Test Inc.). Hep3B cells transfected with pCAT-EpoWT, pCAT-EpoM, or pcDNA3-CAT and grown in normoxia were treated with 10 µg/ml actinomycin D. Following the addition of actinomycin D, cells remained in normoxia for various periods of time after which total RNA was extracted with RNAzol B.

Quantitative Reverse Transcriptase-PCR (RT-PCR) of Exogenous Epo mRNA

mRNA levels of exogenous EpoWT, EpoM, CAT, CAT-EpoWT, and CAT-EpoM were determined by co-amplification with glyceraldehyde-3'-phosphate dehydrogenase (GAPDH) mRNA. Primer p475 (5'-GACAGAGCCATGCTGGGAAGAC-3'), located in the 3'UTR of the Epo gene, and primer p471 (5'-CGCCAGTGTGATGGATATCTGC-3'), located in the vector pcDNA3 sequence, were used to amplify mRNA transcribed from the exogenous Epo gene in all experiments. Primer p427 (5'-GAAGGTGAAGGTCGGAGTCAACG-3'), located in exon 1, and primer 428 (5'-TGCCATGGGTGGAATCATATTGG-3'), located in exon 3, were used to amplify GAPDH mRNA. Primer p472 (5'-TCACATTCTTGCCCGCCTGAT-3') and primer p473 (5'-AGGGATTGGCTGAGACGAAAAACAT-3') were used to amplify CAT mRNA from Hep3B cells stably transfected with pcDNA3-CAT vector. Single-step RT-PCR was carried out using rTth DNA polymerase I (19). RT-PCR conditions were 10-100 ng of total RNA, 1 × EZ buffer (50 mM Bicine, 125 mM KOAc, 8% glycerol, pH 8.2) (Perkin-Elmer), 3.5 mM Mn(OAc)2, 200 µM concentration of each deoxyribonucleotide triphosphate, 400 nM p471 and p475, 100 nM p427 and p428, and 5 units of rTth DNA polymerase I (Perkin-Elmer) in a 50-µl reaction. The cycling parameters were 95 °C 1 min, 65 °C 30 min, followed by 35-42 cycles of 95 °C 30 s, 60 °C 30 s, 72 °C 40 s. One-fifth of each reaction was analyzed on 3% NuSieve (FMC) plus 0.5% agarose gel. The amount of PCR products from Epo mRNA or CAT-Epo mRNA was normalized to PCR products from GAPDH mRNA for each time point by densitometry using NIH Image.

Steady-state mRNA levels of exogenous EpoWT and EpoM in normoxia were determined by comparing the lowest detectable levels using serially diluted total RNA. Total mRNA from Hep3B cells transfected with pcDNA3-EpoWT or pcDNA3-EpoM was serially diluted 1:1 from 5 µg/µl to 1.25 ng/µl. Two aliquots of 10 µl of each dilution were co-amplified with primers p471 and p475 for exogenous EpoWT or EpoM mRNA and primers p427 and p428 for GAPDH mRNA for 36 cycles. The end point for GAPDH mRNA was used to monitor the equivalency of the total RNA between EpoWT and EpoM.

Erythropoietin Radioimmunoassay

Untransfected Hep3B cells or those transfected with EpoWT or EpoM were grown to 70-80% confluency and then maintained in normoxia or exposed to hypoxia. After 20 h, the supernatant was removed and the amount of erythropoietin was determined by an erythropoietin radioimmunoassay (20, 21). Recombinant Epo was used in a range of 3.125-400 milliunits/ml to prepare a standard curve (21). The Epo levels in culture medium are expressed in milliunits/ml and were normalized to either total cellular protein or cell number.

Chloramphenicol Acetyltransferase Assay

CAT activity in Hep3B cells stably transfected with pCAT-EpoWT, pCAT-EpoM, or pcDNA3-CAT and grown in normoxia were determined using a non-radioactive FLASH® CAT-Deoxy assay (Stratagene) according to the manufacturer's instructions. Fifty µg of cytoplasmic lysates, prepared as described under "electromobility shift assay," from each cell line was used in each reaction, and the acetylation reaction was visualized using thin-layer chromatography. The percentage of acetylated product was quantified by the relative absorption at 505 nm or by densitometry using NIH image. CAT activity is shown as the mean ± S.D. of the percentage of acetylation of five experiments.

Statistical Analysis

Individual half-lives for each experiment were determined by linear regression lines of the first four time points drawn by the Microsoft Excel program. mRNA half-lives are presented as the mean ± S.D. of all of the experiments (n). Student's unpaired t test was employed to assess the difference between mRNA half-lives of two groups. The mRNA decay rates are illustrated graphically, and each time point is represented as the mean ± S.D. of relative Epo or CAT-Epo mRNA levels of all of the experiments.


RESULTS

Deletion of the ERBP Binding Site

We have previously reported that a 120-nucleotide region in the 3'UTR of Epo mRNA is necessary for ERBP binding (17). To investigate the effect of ERBP binding on the stability of Epo mRNA, the ERBP binding site was deleted using PCR mutagenesis. In the 3'UTR region of EpoM, a 104-bp region between the stop codon and NcoI sites was replaced with a 17-bp sequence, which resulted in an 87-bp deletion and was confirmed by sequencing. Run-off transcripts were prepared from pcDNA3-EpoM or pcDNA3-EpoWT, and electromobility shift assay was performed using cytoplasmic lysates from untransfected Hep3B cells cultured in normoxia. Fig. 1 demonstrates that while ERBP is able to form a complex with radiolabeled EpoWT run-off RNA, EpoM run-off RNA, which lacks the binding sequence for ERBP, is unable to form a complex with ERBP. The specificity of ERBP binding has been demonstrated previously by competition experiments (17).


Fig. 1. ERBP complex formation with EpoM and EpoWT run-off RNA. Radiolabeled EpoWT or EpoM run-off RNA was incubated with cytoplasmic cell lysates from Hep3B cells grown in normoxia. ERBP complex formation was detected by electromobility shift assay as described under "Experimental Procedures." Lanes 1 and 3, negative controls.
[View Larger Version of this Image (48K GIF file)]


Exogenous Epo mRNA Stability in Normoxia

To determine the effect of ERBP complex formation on the stability of Epo mRNA, Hep3B cells were transfected with pcDNA3-EpoWT or pcDNA3-EpoM and treated with actinomycin D to block transcription. The EpoWT cDNA and EpoM cDNA were cloned into the multiple cloning sites, downstream of a constitutive promoter of cytomegalovirus and upstream of the polyadenylation signal and termination sequence of the bovine growth hormone gene of a mammalian expression vector pcDNA3.

The possible effect of actinomycin D on the level of ERBP was first examined. Cytoplasmic lysates were prepared from Hep3B cells after various periods of time following the addition of actinomycin D, incubated with radiolabeled EpoWT run-off RNA, and ERBP complex formation was examined by electromobility shift assay. The level of ERBP binding was unchanged at 8 h following addition of actinomycin D but reduced 20% after 14 h and an additional 30% 24 h following actinomycin D treatment when compared with untreated cells (Fig. 2). Therefore, it appears that the formation of the ERBP complex remained unchanged following 8 h of actinomycin D treatment and decreased slightly at 14 h following treatment.


Fig. 2. Effect of actinomycin D on ERBP complex formation. Cytoplasmic cell lysates were prepared from Hep3B cells grown in the absence and in the presence of 5 µg/ml actinomycin D for 8, 14, and 24 h in normoxia. Formation of the ERBP complex of these cell lysates incubated with radiolabeled EpoWT run-off RNA was detected by electromobility shift assay. Lane 1, EpoWT run-off RNA was incubated in the absence of cytoplasmic cell lysates from Hep3B cells as a negative control.
[View Larger Version of this Image (69K GIF file)]


To distinguish exogenous Epo mRNA from endogenous Epo mRNA, one of the primers for the mRNA decay experiments, p471, was designed to be specific for a sequence located downstream of the Epo gene and upstream of the polyadenylation signal site, in the vector pcDNA3 (Fig. 3A). Fig. 3B demonstrates the specificity of primer p471 for exogenous mRNA of EpoWT and EpoM (lanes 5 and 6).


Fig. 3. Specific amplification of exogenous Epo mRNA. A, the locations of primers for distinguishing exogenous Epo mRNA from endogenous Epo mRNA. B, RT-PCR was performed as described under "Experimental Procedures." Lanes 1 and 3 were negative controls of amplifications with primers p221 and p220 and primers p221 and p471, respectively. Lanes 2 and 4, total RNA from Hep3B cells were amplified with primers p221 and p220 and primers p221 and p471, respectively. Lanes 5 and 6, total RNA from Hep3B cells transfected with pcDNA3-EpoM and pcDNA3-EpoWT was amplified with primers p221 and p471, respectively. Primer p471 was designed to specifically amplify the exogenous Epo mRNA; therefore, no PCR products were detected in lane 4. The marker lane is a 100-bp ladder (Life Technologies, Inc.).
[View Larger Version of this Image (30K GIF file)]


Single-step RT-PCR using rTth DNA polymerase was used to carry out reverse transcription at an elevated temperature and to prevent contamination. The sensitivity of RT-PCR with primer pair p471 and p475 was determined to be 100 molecules using EpoWT run-off transcripts. The absence of DNA contamination in total RNA preparations was confirmed by amplification without the reverse transcription step.

Total RNA was prepared from Hep3B cells transfected with pcDNA3-EpoWT or pcDNA3-EpoM at various time points following addition of actinomycin D. Hep3B cells stably transfected with the pcDNA3-EpoWT or pcDNA3-EpoM construct (Fig. 4A) were pooled to avoid bias in the level of variable gene expression. The amount of total RNA used in each RT-PCR reaction and the cycling number were determined empirically to ensure that the amplification of either Epo mRNA or GAPDH mRNA did not reach plateau. The level of Epo mRNA at each time point was quantified by normalizing it with co-amplified GAPDH mRNA. Fig. 4B illustrates the mRNA decay rates for exogenous EpoWT and EpoM in normoxia. The mRNA half-life of EpoWT, 7.1 ± 1.4 h (n = 6), was significantly (p = 0.00001) different from the mRNA half-life of EpoM, 2.6 ± 0.5 h (n = 6). These results suggest that ERBP binding prolongs the stability of Epo mRNA, and the shorter half-life of EpoM mRNA in normoxia is likely due to the absence of ERBP complex formation.


Fig. 4. mRNA decay for exogenous EpoWT and EpoM in normoxia and 6-h hypoxia. A, constructs of pcDNA3-EpoWT and pcDNA3-EpoM, with corresponding half-lives in either normoxia or hypoxia. B, mRNA decay for exogenous Epo in normoxia following addition of actinomycin D or in hypoxia with 6-h hypoxic treatment prior to addition of actinomycin D. black-square and square , exogenous EpoWT mRNA decay under normoxia and hypoxia with 6-h hypoxic pre-treatment, respectively. bullet  and open circle , EpoM mRNA decay under normoxia and hypoxia with 6-h hypoxic pre-treatment, respectively. Each point represents the average of three amplifications using total RNA prepared from two different actinomycin D chase experiments. Actinomycin D chase and quantification of exogenous Epo mRNA levels were described under "Experimental Procedures."
[View Larger Version of this Image (20K GIF file)]


Steady-state exogenous Epo mRNA levels in pcDNA3-EpoWT or pcDNA3-EpoM transfected Hep3B cells were compared by RT-PCR using serially diluted total RNA. RT-PCR was performed using serially diluted total RNA to determine the lowest quantity of total RNA for which EpoWT mRNA and EpoM mRNA could be amplified. The input total RNA of each reaction was monitored by the GAPDH end point. The end point for GAPDH was the same, 2.5 ng/µl, for total RNA from pcDNA3-EpoWT and pcDNA3-EpoM. The end point for EpoWT mRNA was 19.5 ng/µl, ~4-fold higher than the end point for EpoM, 78 ng/µl (Fig. 5). These results suggest that ERBP binding acts to stabilize EpoWT mRNA in normoxia.


Fig. 5. Steady-state mRNA levels of exogenous Epo in normoxia. An aliquot of 10 µl of serially diluted total RNA from Hep3B cells transfected with pcDNA3-EpoM or pcDNA3-EpoWT were co-amplified with primers p471 and p475, for exogenous Epo mRNA, and primers P427 and p428, for GAPDH mRNA as described under "Experimental Procedures." Only one of the duplicate reactions of each dilution is presented in this figure. The negative of Polaroid® type 55 film is presented. Short arrows represent end point mRNA detection levels for exogenous EpoWT or EpoM mRNA. Long arrows represent end point mRNA detection levels for GAPDH mRNA.
[View Larger Version of this Image (65K GIF file)]


Hypoxic Modulation of Epo mRNA Stability

In this study, only the cDNA sequence of the wild-type or mutated Epo gene was subcloned into pcDNA3. Any changes in EpoWT or EpoM mRNA stability observed in hypoxia would suggest other hypoxically regulated cis-elements or trans-acting regulatory factors which may be involved in modulation of Epo mRNA stability. Hep3B cells stably transfected with pcDNA3-EpoWT or pcDNA3-EpoM were placed in hypoxia for 6 h after which time actinomycin D was added, and the cells returned to hypoxia. Total RNA was extracted at various time points between 0 and 14 h after addition of actinomycin D, and exogenous Epo mRNA levels were determined by RT-PCR as described under "Experimental Procedures." The mRNA decay rates for exogenous EpoWT and EpoM in hypoxia after a 6-h hypoxic treatment are shown in both Fig. 4B and Fig. 6. The half-lives of EpoWT mRNA and EpoM mRNA were 9.8 ± 2.0 h (n = 6) and 3.0 ± 1.1 h (n = 6), respectively, and the difference was highly statistically significant (p = 0.00002). The EpoWT mRNA half-life in hypoxia increased 2.7 h from the EpoWT mRNA half-life in normoxia (p = 0.021) (Fig. 4B). Furthermore, there was no significant change in the EpoM mRNA half-life in hypoxia compared with the EpoM mRNA half-life in normoxia (Fig. 4B). Interestingly, the EpoWT mRNA decay rate appears to decrease between 6 h and 10 h following addition of actinomycin D, corresponding to between 12 h and 16 h of hypoxic exposure. It is likely that following 12 h of hypoxic exposure, in vivo ERBP complex formation might be increased. To investigate this possibility, Hep3B cells stably transfected with pcDNA3-EpoWT or pcDNA3-EpoM were grown in hypoxia for 12 h prior to the addition of actinomycin D, after which the cells were returned to hypoxia and total RNA was extracted at various time points. The mRNA decay rates for exogenous EpoWT and EpoM in hypoxia after 6- or 12-h hypoxic treatment are compared in Fig. 6. The half-life of EpoWT mRNA increased from 9.8 ± 2.0 h (n = 6) to 14.6 ± 3.9 h (n = 3) (p = 0.039) with an additional 6 h of hypoxic pre-treatment. In contrast, EpoM mRNA half-life remained nearly the same with 6 or 12 h of hypoxic pre-treatment.


Fig. 6. mRNA decay for exogenous Epo in hypoxia with 6- and 12-h hypoxic treatment. To demonstrate the difference in Epo mRNA decay rates with varying periods of hypoxic pre-treatment prior to the addition of actinomycin D, the mRNA decay for EpoWT and EpoM with 6-h hypoxic pre-treatment in Fig. 4B are included in this figure. square  and black-square, exogenous EpoWT mRNA decay under hypoxia with 6- and 12-h hypoxic pre-treatment, respectively. open circle  and bullet , EpoM mRNA decay under hypoxia with 6- and 12-h hypoxic pre-treatment, respectively. Each point represents the average of two or three amplifications using total RNA prepared from two different actinomycin D chase experiments. Actinomycin D chase and quantification of exogenous Epo mRNA levels were described under "Experimental Procedures."
[View Larger Version of this Image (14K GIF file)]


Epo Expression Levels

Epo levels in untransfected Hep3B cells and Hep3B cells stably transfected with either pcDNA3-EpoWT or pcDNA3-EpoM and cultured in normoxia or hypoxia were determined using Epo radioimmunoassay (Table I). As expected, Epo produced by Hep3B cells was undetectable under normoxia and increased ~60-fold under hypoxia. In hypoxia, Epo expressed from the endogenous Epo gene in Hep3B cells also contributed to the observed Epo levels in Hep3B cells transfected with pcDNA3-EpoWT or pcDNA3-EpoM. The Epo level expressed from the endogenous Epo gene accounts for the difference in the Epo levels between Hep3B cells transfected with pcDNA3-EpoM and cultured in normoxia and hypoxia. Therefore, the Epo level expressed from the exogenous EpoM gene was identical under normoxia and hypoxia. In contrast, after subtracting the Epo level contributed by the endogenous Epo gene under hypoxia, the Epo level expressed by the exogenous EpoWT gene increased 1.7-fold compared with the Epo level expressed by the exogenous EpoWT gene in normoxia. The Epo level expressed by the exogenous EpoWT gene was ~5-fold higher than that expressed by the exogenous EpoM gene in normoxia. Under hypoxia, the Epo level expressed by the exogenous EpoWT gene was 6-fold higher than that expressed by the exogenous EpoM gene. These data clearly indicate that the effect of ERBP complex formation on the stability of Epo mRNA is also apparent at the protein level.

Table I.

Epo levels from Hep3B cells and transfected Hep3B exposed to either normoxia or hypoxia

Untransfected Hep3B cells and Hep3B cells transfected with pcDNA3-EpoM and pcDNA3-EpoWT were grown to 80% confluency in normoxia followed by continued normoxic exposure or hypoxic exposure for 24 h. Epo levels were determined as described under "Experimental Procedures."


Epo level
Hep3B Hep3B + pcDNA3-EpoM Hep3B + pcDNA3-EpoWT

milliunits/ml
Normoxia negligible 52  ± 5 258  ± 70
Hypoxia 64 ± 7 134  ± 27 510  ± 118

Epo 3'UTR Confers Stability to Chimeric CAT mRNA

To further investigate the possibility that ERBP complex formation confers stability to EpoWT mRNA, we inserted the 3'UTR of either EpoWT or EpoM next to a heterologous CAT sequence (Fig. 7A). The half-life of pCAT-EpoWT was 7.3 ± 1.3 h (n = 3) compared with 3.7 ± 1.7 h (n = 3) for pCAT-EpoM (p = 0.04). The mRNA half-life of pcDNA3-CAT was 3.4 ± 0.9 h (n = 3) which was significantly different from pCAT-EpoWT (p = 0.03) but not from pCAT-EpoM (Fig. 7B). Fig. 7C illustrates a 3-fold increase in CAT activity produced by pCAT-EpoWT compared with that produced by pCAT-EpoM and pcDNA3-CAT (n = 5). These results suggest that the ERBP complex formation not only acts to stabilize Epo mRNA but also chimeric CAT-EpoWT 3'UTR mRNA.


Fig. 7. mRNA decay of CAT-Epo 3'UTR hybrids in normoxia. A, constructs of CAT-Epo 3'UTR hybrids and pcDNA3-CAT. B, mRNA decay of CAT-EpoWT 3'UTR, CAT-EpoM 3'UTR, and CAT. Each point represents the average of two amplifications of total RNA prepared from two different actinomycin D chase experiments. black-square, square , and open circle , mRNA decay of CAT-EpoWT, CAT-EpoM hybrids, and CAT under normoxia, respectively. Actinomycin D chase and quantification of CAT-Epo3'UTR hybrid mRNA levels were determined as described under "Experimental Procedures." C, CAT activity from Hep3B cells transfected with pCAT-EpoWT, pCAT-EpoM, and pcDNA3-CAT. CAT activity was determined with non-radioactive FLASH® CAT-Deoxy assay using 1-deoxychloramphenicol as substrate.
[View Larger Version of this Image (15K GIF file)]



DISCUSSION

The rate at which mRNA decays is a precise and complex process that can determine the level of gene expression. In eukaryotic cells, the process of mRNA degradation is primarily governed by specific cis-acting sequences, trans-acting factors, poly(A) shortening, endonucleolytic cleavage, and translation (22). In response to environmental stimuli, mRNA decay rates can fluctuate and affect the expression of specific genes, thereby providing the cell with the flexibility to adapt to a changing environment.

Epo gene regulation has been studied extensively and compared with other hypoxically regulated genes such as vascular endothelial growth factor (VEGF) and tyrosine hydroxylase (TH) (10, 23). Several studies have provided evidence that the expression of VEGF, TH, and Epo are regulated at transcriptional or post-transcriptional levels upon exposure to hypoxia (10, 24-26). The hypoxic induction of these three genes appears to require the activation of hypoxic inducible factor 1 suggesting that they may all share a common oxygen sensing pathway (10). In addition, the transcription rates of Epo, VEGF, and TH peak or plateau following hypoxic induction, while steady-state mRNA levels increase or remain elevated throughout the duration of hypoxic exposure (14, 15, 25, 27). There is also evidence that hypoxic induction plays a role in the post-transcriptional regulation of TH and VEGF gene expression by increasing the stability of each mRNA in hypoxia (25-27). Similar to ERBP, proteins have been identified which bind to pyrimidine-rich regions in the 3'UTR of both TH mRNA and VEGF mRNA (26, 28, 29). Prior to our study, only the effect of the VEGF 3'UTR mRNA binding protein on VEGF mRNA stability has been demonstrated (26). It has also been suggested that Epo gene expression may be regulated at the post-transcriptional level; however, no direct evidence has been reported to support this hypothesis. In our study, constructs containing only the Epo cDNA sequence were used; therefore, the observed differences in EpoWT mRNA and EpoM mRNA stability in both normoxia and hypoxia were attributed to the hypoxic modulation of ERBP complex formation.

Actinomycin D has been reported to extend Epo mRNA half-life through the modulation of a putative protein(s) with a rapid turning over rate (12, 14). The effect of actinomycin D on prolonging EpoWT mRNA and EpoM mRNA stability was expected to be equivalent if ERBP complex formation was not significantly altered by actinomycin D. We demonstrated that the effect of actinomycin D on ERBP complex formation was minimal up to 14 h following actinomycin D treatment. In determining mRNA half-lives, mRNA levels were measured up to 14 h following the addition of actinomycin D. Therefore, the level of EpoWT mRNA and pCAT-EpoWT mRNA at the last time point may have been only slightly underestimated.

The absence of ERBP complex formation with EpoM run-off RNA demonstrates that a 104-bp region contained the binding site for ERBP. mRNA steady-state levels represent the difference between the rate of mRNA induction and rate of mRNA decay. Exogenous mRNA transcription levels of EpoWT mRNA and EpoM mRNA were under the control of a constitutive promoter from cytomegalovirus. Therefore, in normoxia, the 4-fold increase in EpoWT mRNA steady-state level compared with the steady-state EpoM mRNA level, reflected the difference in exogenous Epo mRNA decay rates. The EpoWT mRNA half-life in normoxia was 7 h which is consistent with previous findings of the half-life for endogenous Epo mRNA in actinomycin D-treated Hep3B cells after hypoxic pre-treatment (14). The shorter half-life of EpoM mRNA, 2.6 h, implies that ERBP complex formation acts to stabilize Epo mRNA. Moreover, this hypothesis is further supported by results of increased CAT activity and prolonged CAT mRNA stability under normoxia in Hep3B cells transfected with pCAT-EpoWT compared with those transfected with either pCAT-EpoM or pcDNA3-CAT.

Although genes of interest with various deleted sequences in the 3'UTR have been used extensively to identify the specific mRNA control sites and to elucidate the mechanisms through which they mediate mRNA stability, such data must be interpreted cautiously. We cannot exclude the possibility that the 87-bp deletion may have contained a "stabilizing" cis-element. Moreover, mRNA-binding proteins may bind to either specific mRNA sequences or intrinsic stem-loop(s) structures. It is interesting that when we examined a computer-predicted secondary structure of Epo mRNA, this pyrimidine-rich ERBP binding region is part of several stems pairing with regions in 5'UTR, coding sequence, and further downstream in the 3'UTR. Specific point mutations to disrupt the pairing in different stem regions will help to delineate whether the secondary structure of Epo mRNA plays a crucial role in ERBP complex formation. A 188-bp region downstream and adjacent to the ERBP binding site has been identified to contain a destabilizing cis-element. Deletion of this region prolonged Epo mRNA half-life to 15 h compared with the endogenous Epo mRNA half-life, 2 h (30). Part of this region forms a stem with the ERBP binding site, and we speculate that ERBP complex formation may play a role in protecting Epo mRNA as the substrate of endonucleolytic cleavage in this region.

The half-lives of exogenous EpoM mRNAs were nearly identical, i.e. 2.5-3 h, when cultured in normoxia and exposed to either 6- or 12-h hypoxia. However, exogenous EpoWT mRNA half-life was increased 38% and an additional 49% over normoxic levels in cells exposed to 6- and 12-h hypoxia, respectively. Rondon et al. (17) have previously reported that while ERBP appears to be hypoxically up-regulated in various mouse tissues, there did not appear to be a difference in ERBP complex formation from cytoplasmic lysates derived from Hep3B cells cultured in either normoxia or hypoxia. However, using modified binding conditions, ERBP complex formation was increased using cytoplasmic lysates from Hep3B cells exposed to hypoxia.2 Therefore, the prolonged half-life of EpoWT mRNA is most likely due to increased ERBP complex formation under hypoxia. In contrast, the half-life of EpoM mRNA, lacking the ERBP binding site, remained unchanged in hypoxia. Furthermore, the hypoxic modulation of Epo mRNA was supported by comparing Epo levels expressed by endogenous Epo, exogenous EpoM, or exogenous EpoWT measured in normoxia or after 20-h hypoxia. In hypoxia, Epo produced by Hep3B cells transfected with pcDNA3-EpoM was twice the amount produced by endogenous Epo in untransfected Hep3B cells, and it was also twice that produced by pcDNA3-EpoM in normoxia. This suggests that the absence of ERBP binding formation is likely responsible for the non-hypoxic responsive expression of the exogenous EpoM gene. However, there was a 1.7-fold difference between the Epo level expressed by the exogenous EpoWT gene under hypoxia and normoxia, likely due to the ~2-fold increase in EpoWT mRNA half-life observed in prolonged hypoxia.

We conclude that the increase in Epo mRNA stability in hypoxia may account for the discrepancy between decreasing Epo mRNA steady-state levels and increasing Epo expression observed in prolonged hypoxia. We have demonstrated that Epo mRNA stability is increased by the binding of ERBP to the 3'UTR of Epo mRNA. Moreover, the increase in Epo mRNA stability in hypoxia is likely modulated by an increase in ERBP complex formation in hypoxia. Our data have provided the first evidence that Epo mRNA stability is regulated at the post-transcriptional level and shed some light on the mechanism of Epo gene regulation.


FOOTNOTES

*   This work was supported in part by National Institutes of Health Grant DK 40501.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Supported by the Louisiana Educational Quality Support Fund Superior Graduate Fellows Award.
§   Present address: Division of Human Retrovirology, Dana Farber Cancer Institute, Harvard Medical School, Boston, MA 02115.
   To whom correspondence should be addressed: Dept. of Pharmacology, Tulane University School of Medicine, 1430 Tulane Ave., New Orleans, LA 70112. Tel.: 504-524-2631; Fax: 504-588-5283.
1   The abbreviations used are: Epo, erythropoietin; ERBP, erythropoietin mRNA-binding protein; bp, base pair(s); UTR, untranslated region; CAT, chloramphenicol acetyltransferase; RT-PCR, reverse transcriptase-polymerase chain reaction; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Bicine, N,N-bis(2-hydroxyethyl)glycine; VEGF, vascular endothelial growth factor; TH, tyrosine hydroxylase.
2   A. Scandurro, and B. S. Beckman, manuscript in preparation.

ACKNOWLEDGEMENTS

We thank S.-Y. Chang at Roche Molecular Systems for valuable advice and discussion, J. Brookins and Y. Tang for technical assistance, and F. Bunn and G. Morris for critical review of this manuscript.


REFERENCES

  1. Jelkmann, W. (1992) Physiol. Rev. 72, 449-488 [Free Full Text]
  2. Zanjani, E. D., Poster, J., Burlington, H., Mann, L. I., and Wasserman, L. R. (1977) J. Lab. Clin. Med. 89, 640-644 [Medline] [Order article via Infotrieve]
  3. Jacobson, L. O., Goldwasser, E., Fried, W., and Plzak, L. (1957) Nature 179, 633-634
  4. Kuratowska, Z., Lewartowski, B., and Michalak, E. (1961) Blood 18, 527-534
  5. Erslev, A. J. (1975) Am. J. Med. 58, 25-30 [Medline] [Order article via Infotrieve]
  6. Goldberg, M. A., Glass, G. A., Cunningham, J. M., and Bunn, H. F. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 7972-7976 [Abstract]
  7. Shoemaker, C. B., and Mitsock, L. D. (1986) Mol. Cell. Biol. 6, 849-858 [Medline] [Order article via Infotrieve]
  8. McDonald, J. D., Lin, K. F., and Goldwasser, E. (1986) Mol. Cell. Biol. 6, 842 [Medline] [Order article via Infotrieve]
  9. Galson, D. L., Tan, C. C., Ratcliffe, P. J., and Bunn, H. F. (1993) Blood 82, 3321-3326 [Abstract]
  10. Bunn, H. F., and Poyton, R. O. (1996) Physiol. Rev. 76, 839-885 [Abstract/Free Full Text]
  11. Blanchard, K. L., Acquaviva, A. M., Galson, D. L., and Bunn, H. F. (1992) Mol. Cell. Biol. 12, 5373-5385 [Abstract]
  12. Blanchard, K. L., Fandrey, J., Goldberg, M. A., and Bunn, H. F. (1993) Stem Cells 11, Suppl. 1, 1-7
  13. Fandrey, J., and Bunn, H. F. (1993) Blood 81, 617-623 [Abstract]
  14. Goldberg, M. A., Gaut, C. C., and Bunn, H. F. (1991) Blood 77, 271-277 [Abstract]
  15. Costa-Giomi, P., Caro, J., and Weinmann, R. (1990) J. Biol. Chem. 265, 10185-10188 [Abstract/Free Full Text]
  16. Wenger, R. H., Rolfs, A., Hugo, H. M., Bauer, C., and Gassmann, M. (1995) J. Biol. Chem. 270, 27865-27870 [Abstract/Free Full Text]
  17. Rondon, I. J., MacMillan, L. A., Beckman, B., Goldberg, M. A., Schneider, T., Bunn, H. F., and Malter, J. S. (1991) J. Biol. Chem. 266, 16594-16598 [Abstract/Free Full Text]
  18. Ueno, M., Seferynska, I., Beckman, B., Brookins, J., Nakashima, J., and Fisher, J. W. (1989) Am. J. Physiol. 257, C743-C749 [Abstract/Free Full Text]
  19. Myers, T. W., and Gelfand, D. H. (1991) Biochemistry 30, 7661-7666 [Medline] [Order article via Infotrieve]
  20. Rege, A. B., Brookins, J., and Fisher, J. W. (1982) J. Lab. Clin. Med. 100, 829-843 [Medline] [Order article via Infotrieve]
  21. Mason-Garcia, M., Beckman, B. S., Brookins, J. W., Powell, J. S., Lanham, W., Blaisdell, S., Keay, L., Li, S., and Fisher, J. W. (1990) Kidney Int. 38, 969-975 [Medline] [Order article via Infotrieve]
  22. Ross, J. (1995) Microbiol. Rev. 59, 423-450 [Abstract]
  23. Goldberg, M. A., and Schneider, T. (1994) J. Biol. Chem. 269, 4355-4359 [Abstract/Free Full Text]
  24. Levy, A. P., Levy, N. S., Wegner, S., and Goldberg, M. A. (1995) J. Biol. Chem. 270, 13333-13340 [Abstract/Free Full Text]
  25. Czyzyk-Krzeska, M. F., Furnari, B. A., Lawson, E. E., and Millhorn, D. E. (1994) J. Biol. Chem. 269, 760-764 [Abstract/Free Full Text]
  26. Levy, A. P., Levy, N. S., and Goldberg, M. A. (1996) J. Biol. Chem. 271, 2746-2753 [Abstract/Free Full Text]
  27. Ikeda, E., Achen, M. G., Breier, G., and Risau, W. (1995) J. Biol. Chem. 270, 19761-19766 [Abstract/Free Full Text]
  28. Czyzyk-Krzeska, M. F., Dominski, Z., Kole, R., and Millhorn, D. E. (1994) J. Biol. Chem. 269, 9940-9945 [Abstract/Free Full Text]
  29. Czyzyk-Krzeska, M. F., and Beresh, J. E. (1996) J. Biol. Chem. 271, 3293-3299 [Abstract/Free Full Text]
  30. Ho, V., Acquaviva, A., Duh, E., and Bunn, H. F. (1995) J. Biol. Chem. 270, 10084-10090 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.