RNA-Binding Proteins HuR and PTB Promote the Translation of Hypoxia-Inducible Factor 1{alpha}

The levels of hypoxia-inducible factor 1 ${alpha}$ (HIF-1 ${alpha}$ ) are tightlycontrolled. Here, we investigated the posttranscriptional regulationof HIF-1 ${alpha}$ expression in human cervical carcinoma HeLa cells respondingto the hypoxia mimetic CoCl₂. Undetectable in untreated cells,HIF-1 ${alpha}$ levels increased dramatically in CoCl₂-treated cells,while HIF-1 ${alpha}$ mRNA levels were unchanged. HIF-1 ${alpha}$ translation waspotently elevated by CoCl₂ treatment, as determined by de novotranslation analysis and by monitoring the polysomal associationof HIF-1 ${alpha}$ mRNA. An internal ribosome entry site in the HIF-1 ${alpha}$ 5' untranslated region (UTR) was found to enhance translationconstitutively, but it did not further induce translation inresponse to CoCl₂ treatment. Instead, we postulated that RNA-bindingproteins HuR and PTB, previously shown to bind HIF-1 ${alpha}$ mRNA, participatedin its translational upregulation after CoCl₂ treatment. Indeed,both RNA-binding proteins were found to bind HIF-1 ${alpha}$ mRNA in aCoCl₂-inducible manner as assessed by immunoprecipitation ofendogenous ribonucleoprotein complexes. Using a chimeric reporter,polypyrimidine tract-binding protein (PTB) was found to bindthe HIF-1 ${alpha}$ 3'UTR, while HuR associated principally with the 5'UTR.Lowering PTB expression or HuR expression using RNA interferencereduced HIF-1 ${alpha}$ translation and expression levels but not HIF-1 ${alpha}$ mRNA abundance. Conversely, HIF-1 ${alpha}$ expression and translationin response to CoCl₂ were markedly elevated after HuR overexpression.We propose that HuR and PTB jointly upregulate HIF-1 ${alpha}$ translationin response to CoCl₂.

INTRODUCTION

Top

INTRODUCTION

The hypoxia-inducible factor (HIF) functions as a master regulatorof angiogenesis, a process whereby new blood vessels are formedto ensure adequate oxygen perfusion of tissues. Under conditionsof low oxygen tension (hypoxia), HIF activates the transcriptionof genes bearing hypoxia response elements in their promoterregions. HIF-regulated genes include many cancer-related genes,such as vascular endothelial growth factor (VEGF), the glucosetransporter GLUT1, erythropoietin, β-catenin, and manyother genes implicated in cellular functions, such as proliferation,angiogenesis, glucose metabolism, and survival (64). Consequently,understanding the regulation of HIF function and searching fortherapies to inhibit HIF activity have been the focus of intensestudy in recent years (56).

Expressed in two main isoforms (HIF-1 and HIF-2), HIF is a heterodimericcomplex consisting of a transiently expressed, oxygen-regulated ${alpha}$ -subunit (HIF-1 ${alpha}$ and HIF-2 ${alpha}$ ) and a constitutively expressed βsubunit (HIF-1β). The HIF-1 ${alpha}$ gene has been proposed to beconstitutively transcribed through the action of transcriptionfactors such as AP-1, AP-2, NF-1, and NF- ${kappa}$ B (10). Under normaloxygen conditions (normoxia, ranging from 2% to 14% O₂, dependingon the tissue [60]), HIF-1 ${alpha}$ protein levels are essentially undetectable,but in response to hypoxia, HIF-1 ${alpha}$ expression is rapidly andpotently elevated. The posttranslational stabilization of theHIF-1 ${alpha}$ protein is the best-understood mechanism to increase itsexpression levels. Under normoxic conditions, HIF-1 ${alpha}$ has a veryshort half-life ( $~$ 5 min), achieved through hydroxylation at conservedprolyl and asparagyl residues that are targeted for degradationby the von Hippel-Lindau protein (pVHL), the recognition factorof a ubiquitin E3 ligase complex (5, 30, 55). During hypoxia,these hydroxylases are quickly inhibited, in turn causing arapid and robust stabilization of HIF-1 ${alpha}$ (37).

Additional mechanisms to regulate HIF-1 ${alpha}$ expression during hypoxiahave begun to emerge. HIF-1 ${alpha}$ expression is also strongly regulatedat the posttranscriptional level, primarily through the actionof trans-acting factors (noncoding RNAs and RNA-binding proteins)that interact with the HIF-1 ${alpha}$ mRNA and regulate its decay andtranslation rates. Among the noncoding RNAs, the naturally occurringantisense HIF transcript has been shown to bind to the HIF-1 ${alpha}$ mRNA and to regulate its expression posttranscriptionally (51,59), and microRNAs have been postulated to influence HIF-1 ${alpha}$ expressionin response to hypoxia (11). The polypyrimidine tract-bindingprotein (PTB) was shown to bind to the internal ribosome entrysite (IRES) located in the 5' untranslated region (UTR) of theHIF-1 ${alpha}$ mRNA (and in many other viral and cellular mRNAs) (23,46) and to promote the translation of HIF-1 ${alpha}$ during hypoxia (34,54). In fact, the translational control of HIF-1 ${alpha}$ expressionis emerging as a major regulatory step, estimated to accountfor $~$ 40 to 50% of the elevation in HIF-1 ${alpha}$ protein levels by short-termhypoxia (1% O₂, 3 h) (6, 22). In this regard, while hypoxiacan inhibit general protein synthesis and the translationalregulator mammalian target of rapamycin (mTOR), the cap-mediatedtranslation of a subset of mRNAs, including HIF-1 ${alpha}$ mRNA, is induced;this effect was proposed to be linked to protein kinase C ${alpha}$ activity(22).

Under normoxic conditions, numerous other stimuli, such as cytokines,vascular hormones, iron chelators, and growth factors, havealso been shown to modulate HIF-1 ${alpha}$ expression (reviewed in references30 and 67). Nonhypoxic inducers of HIF-1 ${alpha}$ expression appear tofunction primarily by enhancing HIF-1 ${alpha}$ biosynthesis in a varietyof cells. For example, vasoactive hormones (such as angiotensinII, thrombin, and endothelin) were found to increase HIF-1 ${alpha}$ expressionthrough a combination of upregulated HIF-1 ${alpha}$ gene transcriptionand heightened HIF-1 ${alpha}$ translation (47). The latter effect waslinked to the activation of phosphatidylinositol 3-kinase (PI3K)by vasoactive hormones, leading to the activation of the downstreameffectors mTOR, p70 S6 kinase (S6K), and the ribosomal proteinS6 (rpS6). Phosphorylated rpS6, in turn, has been widely believedto promote the translation of a specific subset of transcriptsbearing 5'-oligopyrymidine tracts (5'TOP), among them HIF-1 ${alpha}$ mRNA (3, 47). This hypothesis was recently proven incorrect,as S6K knockout mice and rpS6 knockin mice (carrying mutationsin all rpS6 phosphorylation sites) displayed normal translationalcontrol of TOP mRNAs (48, 52; reviewed in reference 53). Inaddition, mTOR-independent, Akt-dependent pathways of increasedHIF-1 ${alpha}$ translation have also been reported (49). Treatment withanother nonhypoxic agent, the androgen dihydrotestosterone,has also been proposed to influence HIF-1 ${alpha}$ expression posttranscriptionally,via the formation of ribonucleoprotein (RNP) complexes comprisingHIF-1 ${alpha}$ mRNA and the RNA-binding protein HuR (57). A member ofthe Hu/Elav family, HuR has been implicated in stabilizing andregulating the translation of many target mRNAs (4, 18, 29);however, the specific posttranscriptional influence of HuR uponthe expression of HIF-1 ${alpha}$ in dihydrotestosterone-treated cellswas not examined directly (57).

Our own studies also identified the HIF-1 ${alpha}$ mRNA as a putativetarget of HuR, since it contained several hits of a motif presentin HuR target transcripts (38). Given the links between HuRand cancer (19, 39), and between cancer and HIF-1 ${alpha}$ (50, 56, 62),we were interested in testing experimentally if HuR contributedto the regulation of HIF-1 ${alpha}$ expression. Here, we investigatethe response of human cervical carcinoma HeLa cells to the hypoxiamimetic cobalt chloride (CoCl₂). Like other metal ions, cobalt(II)can mimic the response to hypoxia in cells under normal oxygentension, elevating HIF-1 ${alpha}$ levels and implementing a HIF-1-regulatedgene expression program (41); however, the two stimuli differin that hypoxia triggers an antiapoptotic response, but CoCl₂and other mimetics do not (12). Using CoCl₂ in the present studieswas advantageous because it enabled many experiments that couldnot be performed under the restrictions of hypoxia chambers.Moreover, CoCl₂ is a physiologic stress in its own right, asmetal exposure represents a significant environmental and occupationalrisk for humans today (9). Our results show that HuR and PTBbind the HIF-1 ${alpha}$ mRNA and promote HIF-1 ${alpha}$ translation under hypoxia-likeconditions.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Cell culture, treatment, transfections, and infections. Human cervical carcinoma HeLa cells were cultured in Dulbecco'smodified essential medium (Invitrogen), and human lung carcinomaA549 cells were cultured in Ham's F-12K medium (Invitrogen),each supplemented with 10% fetal bovine serum and antibiotics.Cells were treated at $~$ 80% confluence. Hypoxia was achieved byplacing cells in a hypoxia incubator (1% O₂, 5% CO₂, 94% N₂;Thermo Forma). CoCl₂ (Sigma) was used at 200 µM for 2.5h (unless otherwise indicated) in complete medium containing5% fetal bovine serum. Small interfering RNAs (siRNAs; all fromQiagen) were used at a final concentration of 30 nM. siRNAstargeting HuR (pooled before use) were AATCTTAAGTTTCGTAAGTTA(HuR U1), TTCCTTTAAGATATATATTAA (HuR U2), and AAGTGCAAAGGGTTTGGCTTT(HuR H4); PTB-directed siRNA was S100301490; control (Ctrl)siRNAs were AATTCTCCGAACGTGTCACGT (c1) and AllStars negativecontrol siRNA (c2). Adenovirus AdHuR (custom-made; containingthe entire HuR coding region) and control AdCtrl were generatedand amplified, and titers were determined by ViraQuest, Inc.;they were used at the indicated PFU.

Plasmid pEGFP-HIF(5') was generated by cloning the HIF-1 ${alpha}$ 5'UTR(bp 12 to 352) into the multiple cloning site of pEGFP-N1 (Clontech)at the HindIII restriction site; pEGFP-HIF(3') was generatedby cloning the HIF-1 ${alpha}$ 3'UTR (bp 2793 to 3934) into the multiplecloning site of pEGFP-C1 (Clontech) after HindIII and BamH1digestion. The HIF-1 ${alpha}$ 5'UTR was amplified by PCR using cDNA preparedfrom HeLa cells as a template and the oligonucleotide primersGCCGGCTAGCTCGTCTGAGGGGACAGGA (NheI site underlined) and GCCGCTCGAGGGTGAATCGGTCCCCGCGATGT(XhoI site underlined). The resulting product was subsequentlycloned into the NheI and XhoI sites of the bicistronic reporterplasmid pBIC (20) to generate pBIC-HIF-1 ${alpha}$ . The promoterless pBIC-HIF-1 ${alpha}$ (–CMV)plasmid was generated by restriction digestion of pBIC-HIF-1 ${alpha}$ with BamHI and BglII to remove the cytomegalovirus (CMV) promoter,followed by religation of the plasmid backbone.

Western blotting and co-IP assays. Whole-cell lysates (WCE) were prepared using radioimmunoprecipitation(RIPA) buffer. Cytoplasmic and nuclear protein fractions (CEand NE, respectively) were isolated with NE-PER nuclear andcytoplasmic extraction reagents (Pierce) following the manufacturer'sspecifications. Proteins were resolved by 8%, 10%, or 12% sodiumdodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)and transferred onto polyvinylidene difluoride membranes (Invitrogen).Incubations with primary mouse monoclonal antibodies recognizingHuR (Santa Cruz), β-actin (Abcam), ${alpha}$ -tubulin (Santa Cruz),or HIF-1 ${alpha}$ (BD Biosciences Pharmingen), as well as with goat polyclonalanti-PTB antibody (Abcam), were followed by incubations withthe appropriate secondary antibodies (Amersham). Protein bandswere detected by ECL-Plus (Amersham) or SuperSignal West Femtomaximum sensitivity substrate (Pierce). For coimmunoprecipitation(co-IP) analysis, whole-cell lysates (5 mg/ml) prepared in RIPAbuffer were either left untreated or incubated with RNase A(100 µg/ml) and RNase T₁ (1,000 U/ml) for 10 min at 30°C.IP was performed using 20 µg antibody (mouse immunoglobulinG [IgG] from BD Biosciences Pharmingen, anti-HuR antibody fromSanta Cruz, or anti-PTB from Abcam) for 16 h at 4°C; IPsamples were then assayed by Western blotting.

β-Galactosidase assay, CAT ELISA, and neomycin ELISA for IRES analysis. Transiently transfected cells were washed in 1 ml phosphate-bufferedsaline and harvested in 300 µl chloramphenicol acetyltransferase(CAT) enzyme-linked immunosorbent assay (ELISA) kit lysis bufferaccording to the manufacturer's protocol (Roche Molecular Biochemicals).β-Galactosidase (β-Gal) enzymatic activity was determinedby spectrophotometric assay using o-nitrophenyl-β-D-galactopyranosideas previously described (40). CAT levels were determined usingthe CAT ELISA kit according to the protocol provided by themanufacturer (Roche Molecular Biochemicals). The relative IRESactivity was calculated as the CAT/β-Gal ratio. The levelsof neomycin phosphotransferase (NEO; NPTII) were determinedusing the NPTII ELISA kit following the manufacturer's protocol(Agdia).

Reverse transcription-quantitative PCR (RT-qPCR) for IRES analysis. Total RNA was isolated from transfected cells using the RNAspinMini RNA isolation kit according to the manufacturer's instructions(GE Healthcare). cDNA was generated using an oligo(dT)₁₈ primerand the Bulk 1st-Strand synthesis kit according to the protocolprovided by the manufacturer (GE Healthcare). The synthesizedcDNA was used as the template for quantitative PCR using theQuantiTect SYBR green PCR kit (Qiagen) and analyzed on an ABIPrism 7000 detection system using the ABI Prism 7000 SDS software.Quantitative PCRs were carried out using primer pairs (forwardand reverse in each case) ACTATCCCGACCGCCTTACT and CTGTAGCGCTGATGTTGAAto detect the β-Gal mRNA and GCGTGTTACGGTGAAAACCT and GGGCGAAGAACTTGTCCATAto detect the CAT mRNA.

Analysis of the expression of HIF-1 ${alpha}$ transcriptional targets. For the detection of HIF-1 ${alpha}$ transcriptional targets VEGF andGLUT1, total RNA was prepared from HeLa cells and mRNA levelswere measured by RT-qPCR analysis. The primer pair CCTAAGGATCTCTCAGGAGCACAGand TCAGGTTTGGAAGTCTCATCCAG was used for GLUT1. Transcript levelswere normalized to those of the housekeeping glyceraldehyde-3-phosphatedehydrogenase (GAPDH) mRNA; GAPDH and VEGF primer sequencesare listed above.

Polysome analysis. HeLa cells were either left untreated or treated with CoCl₂for 2.5 h. Translation was inhibited with 0.1 mg/ml cycloheximide(10 min), and then cells were scraped and lysed in 300 µlPEB lysis buffer (0.3 M NaCl, 15 mM MgCl₂, 15 mM Tris-HCl, pH7.6, 1% Triton X-100) on ice for 10 min. Nuclei were pelletedat 10,000 x g for 10 min, and 250 µg of protein of theresulting supernatant was fractionated through a 10 to 50% linearsucrose gradient, as described elsewhere (15). Eleven fractionswere obtained using a fraction collector (Brandel), and theirquality was monitored at 254 nm using a UV-6 detector (ISCO).RNA in each fraction was extracted with TRIzol (Invitrogen);following RT, real-time qPCR was performed using SYBR green(Applied Biosystems) and the following primer pairs: AACCCATTCCTCACCCATCAand TCCACCTCTTTTGGCAAGCA for HIF-1 ${alpha}$ and ACATCAAGAAGGTGGTGAAGCAGGand CCAGCAAGGATACTGAGAGCAAGAG for GAPDH.

IP of RNP complexes. Endogenous mRNA-protein complexes were precipitated as previouslydescribed (32). Briefly, HeLa cytoplasmic lysates were preparedin polysome lysis buffer (100 mM KCl, 5 mM MgCl₂, 10 mM HEPES,pH 7.0, 0.5% Nonidet P-40 [NP-40], 1 mM dithiothreitol) containing100 units/ml RNase Out (Invitrogen) and a protease inhibitorcocktail (Roche). One mg of lysate was incubated (1 h, 4°C)with 100 µl of a 50% (vol/vol) suspension of protein A-Sepharosebeads precoated with 30 µg each of mouse anti-HuR (SantaCruz), goat anti-PTB (Abcam), goat anti-TIAR or anti-TIA-1 (SantaCruz), G3BP (BD Transduction Laboratories), and mouse IgG1 orgoat IgG1 (BD Pharmingen). The beads were washed five timeswith NT2 buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mMMgCl₂, and 0.05% NP-40). For RNA analysis, the beads were incubatedwith 100 µl NT2 buffer containing 20 units of RNase-freeDNase I (15 min, 30°C), washed twice with 1 ml NT2 bufferand further incubated in 100 µl NT2 buffer containing0.1% SDS and 0.5 mg/ml proteinase K (15 min, 55°C) to digestthe proteins bound to the beads. RNA was extracted using phenoland chloroform, precipitated in the presence of glycoblue, andused in RT reactions in the presence of random hexamers, oligo(dT)primer, and SuperScript II reverse transcriptase (Invitrogen).Conventional PCR (28 cycles) and qPCR amplification were performedusing the following specific primer pairs: AACCCATTCCTCACCCATCAand TCCACCTCTTTTGGCAAGCA for HIF-1 ${alpha}$ , TCTACCTCCACCATGCCAAGT andGATGATTCTGCCCTCCTCCTT for VEGF, ACGTAAACGGCCACAAGTTC and AAGTCGTGCTGCTTCATGTGfor enhanced green fluorescent protein (EGFP), and ACATCAAGAAGGTGGTGAAGCAGGand CCAGCAAGGATACTGAGAGCAAGAG for GAPDH.

Analysis of de novo translation. Nascent translation of HIF-1 ${alpha}$ and GAPDH was studied by incubatingHeLa cells with 1 mCi L-[³⁵S]methionine and L-[³⁵S]cysteine(Easy Tag Express; NEN/Perkin-Elmer, Boston, MA) per 60-mm platefor 15 min. Cells were lysed in RIPA buffer (10 mM Tris-HCl[pH 7.4], 150 mM NaCl, 1% NP-40, 1 mM EDTA, 0.1% SDS, and 1mM dithiothreitol), and IP reactions were carried out in 1 mlTNN buffer (50 mM Tris-HCl [pH 7.5], 250 mM NaCl, 5 mM EDTA,0.5% NP-40) for 16 h at 4°C using anti-HIF-1 ${alpha}$ (BD Pharmingen),IgG1 (BD Pharmingen), or anti-GAPDH (Santa Cruz Biotechnology)antibodies. Following extensive washes in TNN buffer, the IPsamples were resolved by SDS-PAGE, transferred onto polyvinylidenedifluoride filters, and visualized with a PhosphorImager (MolecularDynamics). The signals were quantified and presented as a percentageof the signals in CoCl₂-treated cells; the values of radiolabeledsignals were adjusted to the amount of label in that sample.

RESULTS

Top

RESULTS

The hypoxia mimetic CoCl₂ specifically elevates HIF-1 ${alpha}$ translation and protein levels. In keeping with earlier literature (36), exposure of HeLa cellsto either hypoxia (1% O₂) or to the widely used hypoxia mimeticCoCl₂ (200 or 500 µM) for 2.5 h did not change HIF-1 ${alpha}$ mRNAlevels significantly, as assessed after RT of total RNA followedby conventional PCR or real-time qPCR amplification (Fig. 1A and B).By contrast, HIF-1 ${alpha}$ protein levels increased dramatically incells treated with CoCl₂ (200 µM; the dose used in allsubsequent experiments) and showed a much less pronounced induction( $~$ 20-fold lower) after incubation in 1% O₂ (Fig. 1C).

View larger version (26K):
[in this window]
[in a new window]

FIG. 1. CoCl₂ treatment does not affect HIF-1 ${alpha}$ mRNA levels but potently elevates HIF-1 ${alpha}$ protein abundance. (A and B) HeLa cells were either left untreated or treated with hypoxia (1% O₂, 2.5 h) or with the hypoxia mimetic CoCl₂ (200 or 500 µM for 2.5 h), whereupon total RNA was extracted and used to measure the levels of the HIF-1 ${alpha}$ and housekeeping GAPDH mRNAs by RT followed by either conventional PCR (A) or real-time qPCR (B). Conventional PCR products were visualized by electrophoresis in ethidium bromide-stained agarose gels, while qPCR results are shown as the mean values for HIF-1 ${alpha}$ mRNA levels (normalized to GAPDH mRNA levels) plus the standard errors of the means from three independent experiments. (C) HeLa cells were treated for 2.5 h with CoCl₂ (200 µM) or with 1% O₂, whereupon the levels of HIF-1 ${alpha}$ , positive control (CoCl₂-inducible) p53, and loading control β-actin were assessed by Western blot analysis. Representative data are shown; 60 µg (for HIF-1 ${alpha}$ ) and 10 µg (for p53 and β-actin) of protein in whole-cell lysates was loaded per lane. (8%), HIF-1 ${alpha}$ signals were visualized after electrophoresis through 8% acrylamide SDS-containing gels (labeled 8%); p53 and β-actin signals were visualized after 12% SDS-PAGE (labeled as 12%).

Given the robust increase in HIF-1 ${alpha}$ protein levels by CoCl₂ treatment,we postulated that in addition to possibly stabilizing the HIF-1 ${alpha}$ protein (like hypoxia does), CoCl₂ treatment also elevated HIF-1 ${alpha}$ translation. To test this hypothesis, we monitored the rateof nascent HIF-1 ${alpha}$ translation by performing a brief (15-min)incubation with L-[³⁵S]methionine and L-[³⁵S]cysteine in HeLacells that had been treated with 1% O₂ or CoCl₂ for 2.5 h orhad been left untreated. Following cell lysis, nascent HIF-1 ${alpha}$ was visualized by IP of radiolabeled material using either anti-HIF-1 ${alpha}$ or IgG antibodies. This assay revealed a markedly elevated denovo translation of HIF-1 ${alpha}$ in CoCl₂-treated cells compared withuntreated cells; by contrast, hypoxia (1% O₂, 2.5 h) had onlya modest stimulatory effect on HIF-1 ${alpha}$ translation ( $~$ 2% of whatwas seen after CoCl₂ treatment) (Fig. 2A). De novo translationof the housekeeping protein GAPDH, tested in parallel, remainedunchanged among treatment groups (Fig. 2A).

View larger version (56K):
[in this window]
[in a new window]

FIG. 2. CoCl₂ treatment promotes HIF-1 ${alpha}$ translation. (A) Nascent HIF-1 ${alpha}$ production was monitored following a brief (15-min) incubation of HeLa cells with L-[³⁵S]methionine and L-[³⁵S]cysteine after either no treatment (Untr.), treatment with CoCl₂ (200 µM, 2.5 h), or hypoxia (2.5 h in a 1% O₂ incubator). Following IP using either anti-HIF-1 ${alpha}$ antibody, anti-GAPDH antibody, or control IgG, the incorporation of radiolabeled amino acids into the newly synthesized HIF-1 ${alpha}$ and GAPDH proteins was assessed by electrophoresis through 8% or 12% acrylamide SDS-containing gels, respectively. The radiolabeled signals were visualized and quantified using a PhosphorImager; the incorporation of radiolabeled amino acids into nascent polypeptides is shown as the percentage of the signal in CoCl₂-treated cultures. (B) Top, the relative association of the HIF-1 ${alpha}$ mRNA with polysomes was tested by preparing cytoplasmic lysates from cells treated as for panel A, fractionating them through sucrose gradients, and collecting 11 fractions for analysis. Bottom, RNA was extracted from each fraction, and the levels of HIF-1 ${alpha}$ mRNA in each fraction from each population (untreated or CoCl₂ treated) was measured by RT-qPCR (left graph). The levels of GAPDH mRNA were also measured in each fraction and plotted (right graph). The data shown are representative of three independent experiments. The direction of sedimentation (top and bottom of the gradient), as well as the components of the translational machinery in each fraction (no ribosome components, ribosome subunits, LMWP, and HMWP) are indicated.

Further evidence that HIF-1 ${alpha}$ was translated more actively inthe presence of CoCl₂ was obtained from experiments that examinedthe association of HIF-1 ${alpha}$ mRNA with the translation machinery.The relative distribution of the HIF-1 ${alpha}$ mRNA on sucrose gradientsused to fractionate polysomes according to their size was usedas a measure of the HIF-1 ${alpha}$ mRNA engagement with the translationalapparatus. Compared with the distribution of HIF-1 ${alpha}$ mRNA in untreatedcells, CoCl₂ treatment shifted the distribution of HIF-1 ${alpha}$ mRNAtowards heavier fractions of the gradient, revealing an enhancedassociation of HIF-1 ${alpha}$ mRNA with larger polysomes (Fig. 2B) thatwas consistent with the elevation in HIF-1 ${alpha}$ de novo protein biosynthesis(Fig. 2A). A control housekeeping mRNA (GAPDH) was routinelyfound to shift towards heavier polysomes. This observation contrastedwith the prediction that global translation would decrease afterCoCl₂ treatment. However, in this study system, we did not observea loss in overall protein translation rates after 2.5 h in 1%O₂ or CoCl₂ (see Fig. S1 in the supplemental material), althoughtranslational elongation has been reported to slow down duringhypoxia due to alteration of eEF2 activity (8).

It is worth noting that the GAPDH coding region is significantlyshorter (1.0 kb) than the HIF-1 ${alpha}$ coding region (2.5 kb). Accordingly,an actively translated HIF-1 ${alpha}$ mRNA would be predicted to accommodatemany more ribosomes within its coding region and, hence, formheavier polysomes than those seen with GAPDH mRNA. Instead,HIF-1 ${alpha}$ polysomes peaked at a size that cosedimented with GAPDHpolysomes (approximately fraction 7). It was only after CoCl₂treatment that HIF-1 ${alpha}$ polysomes showed the expected heavier size(approximately fraction 10), revealing a robust enhancementin HIF-1 ${alpha}$ translation, while GAPDH polysomes showed a comparativelyless pronounced shift (approximately fraction 8) after CoCl₂treatment (Fig. 2B). Taken together, these results indicatethat exposure to CoCl₂ promotes the translation of the HIF-1 ${alpha}$ mRNA, linked to the increase in HIF-1 ${alpha}$ protein levels.

IRES-dependent translation is not responsible for increasing HIF-1 ${alpha}$ translation in response to CoCl₂ treatment. To begin to study the mechanisms mediating the CoCl₂-inducedtranslation of HIF-1 ${alpha}$ mRNA, we treated cells with pateamine A,a small molecule that inhibits cap-dependent translation. Asshown in Fig. 3A, pateamine A treatment potently suppressedHIF-1 ${alpha}$ expression, suggesting that CoCl₂ induced HIF-1 ${alpha}$ translationin a cap-dependent manner. However, since pateamine A couldalso function by inhibiting translation driven by IRES, andgiven earlier reports that an IRES within the 5'UTR of the mouseHIF-1 ${alpha}$ promoted HIF-1 ${alpha}$ translation (34, 46, 54), we sought tostudy if the human HIF-1 ${alpha}$ 5'UTR also harbored IRES activity and,if so, if HIF-1 ${alpha}$ IRES activity was enhanced in response to CoCl₂treatment. The human HIF-1 ${alpha}$ 5'UTR was cloned into the β-Gal-CATbicistronic reporter system (pBIC) (Fig. 3B), in which expressionof the 5' cistron (β-Gal) is driven by cap-dependent translation,whereas expression of the 3' cistron (CAT) is driven by IRES-dependenttranslation. HEK 293 cells were transfected with either theempty reporter plasmid pBIC or the pBIC-HIF-1 ${alpha}$ construct. Wefound that the HIF-1 ${alpha}$ 5'UTR could enhance expression of the 3'CAT cistron (Fig. 3C); importantly, expression of the 3' CATcistron mediated by the insertion of the HIF-1 ${alpha}$ 5'UTR sequencewas not due to either cryptic promoter activity (Fig. 3D) orsplicing of the bicistronic mRNA (Fig. 3E). These data indicatethat the human HIF-1 ${alpha}$ 5'UTR exhibits IRES activity.

View larger version (27K):
[in this window]
[in a new window]

FIG. 3. The IRES in the HIF-1 ${alpha}$ 5'UTR does not mediate the enhanced HIF-1 ${alpha}$ translation in response to CoCl₂ treatment. (A) Cells were treated with pateamine A (100 nM) for 4 h, whereupon they were either left untreated or treated with CoCl₂ (200 µM, 2.5 h). The levels of HIF-1 ${alpha}$ , p53, and β-actin were then assessed by Western blot analysis. (B) Schematic representation of the bicistronic reporter constructs used in this study. (C) HEK 293 cells were transfected with either pBIC or pBIC-HIF-1 ${alpha}$ , and 24 h later β-Gal and CAT protein levels were measured. The relative IRES activity was expressed as the CAT/β-Gal ratio. The activity of pBIC was set as 1. Graph depicts the means + standard errors of the means (SEM) of triplicate samples. (D) HEK 293 cells were transfected with either pBIC-HIF-1 ${alpha}$ or the promoterless pBIC-HIF-1 ${alpha}$ (–CMV), and 24 h later the levels of β-Gal, CAT, and NEO (which served as a transfection control, as it is expressed from a separate promoter on the bicistronic reporter plasmid) were determined. The graph depicts the means + SEM of triplicate samples. (E) Total RNA was isolated 24 h after transfection of HEK 293 cells with either pBIC or pBIC-HIF-1 ${alpha}$ ; cDNA was generated by reverse transcription, and the levels of β-Gal and CAT cistrons were determined by qPCR. The values are expressed as CAT relative to β-Gal {2^{–[Ct(CAT) – Ct(β-Gal)]}} and the ratio for pBIC-transfected cells was set as 1. The graph depicts the means +SEM of triplicate samples. (F) Eighteen hours after transfection with the pBIC-HIF-1 ${alpha}$ bicistronic reporter, HeLa cells were treated with 200 µM CoCl₂ for 2.5 h and protein extracts were harvested for analysis of β-Gal and CAT levels. Relative IRES activity was expressed as the CAT/β-Gal ratio. The activity of HIF-1 ${alpha}$ IRES in untreated cells was set as 1. The graph depicts the means +SEM of two independent experiments performed in triplicate. Simultaneously, expression levels of endogenous HIF-1 ${alpha}$ and GAPDH were determined by Western blot analysis to assay the effectiveness of CoCl₂ treatment.

To determine if HIF-1 ${alpha}$ IRES activity was enhanced during CoCl₂treatment, 18 h after transfection of the pBIC-HIF-1 ${alpha}$ bicistronicreporter, HeLa cells were treated with CoCl₂ (200 µM,2.5 h). Surprisingly, CoCl₂ treatment did not enhance HIF-1 ${alpha}$ IRES activity above basal levels, despite the strong inductionof endogenous HIF-1 ${alpha}$ (Fig. 3G). These findings suggest that increasedtranslation of HIF-1 ${alpha}$ during CoCl₂ treatment is not mediatedby an enhancement of HIF-1 ${alpha}$ IRES activity.

HuR and PTB can bind endogenous and biotinylated HIF-1 ${alpha}$ mRNA. Besides the reported IRES-dependent increase in mouse HIF-1 ${alpha}$ translation after hypoxia, RNA-binding proteins have also beenshown to enhance HIF-1 ${alpha}$ translation. Therefore, we next examinedthe possibility that RNA-binding proteins elicit the CoCl₂-mediatedincrease in HIF-1 ${alpha}$ translation. The $~$ 4-kb HIF-1 ${alpha}$ mRNA containsseven computationally predicted hits of a previously identifiedHuR motif (38) in the HIF-1 ${alpha}$ coding region and 3'UTR, in agreementwith earlier findings that HuR forms a complex with the HIF-1 ${alpha}$ mRNA in leukemia and prostate carcinoma cell extracts (57).However, using biotinylated partial transcripts spanning theentire HIF-1 ${alpha}$ mRNA, we observed that cytoplasmic HuR presentin lysates from CoCl₂-treated cells bound preferentially theHIF-1 ${alpha}$ 5'UTR and was also capable of binding the coding region(see Fig. S2 in the supplemental material). Unexpectedly, PTBassociated with all of the biotinylated fragments tested invitro, indicating that PTB had a general affinity for each ofthe transcripts (see Fig. S2 in the supplemental material),all of which contained polypyrimidine tracts of different lengths,and suggesting that the specificity of binding of PTB to RNAscould not be studied adequately by this in vitro assay. Therefore,we decided to narrow down the region to which HuR and PTB boundthe HIF-1 ${alpha}$ mRNA by constructing and assaying chimeric reporterscontaining the EGFP coding region along with either the 5'UTRor the 3'UTR of HIF-1 ${alpha}$ (Fig. 4). Following transfection of theseconstructs into HeLa cells, we performed RNP-IP assays usinganti-HuR, anti-PTB, or (control) IgG, and tested the abundanceof EGFP, EGFP-HIF(5'), or EGFP-HIF(3') mRNAs in the IP materialsby using RT followed by qPCR. This analysis revealed that HuRpreferentially bound the 5'UTR-containing EGFP chimeric transcript,while PTB preferentially bound the 3'UTR-containing EGFP chimerictranscript (Fig. 4, graph).

View larger version (23K):
[in this window]
[in a new window]

FIG. 4. HIF-1 ${alpha}$ heterologous reporter analysis. Top, schematic of the HIF-1 ${alpha}$ mRNA, indicating the 5'UTR, coding region (CR), and 3'UTR. Bottom, schematic representation of the reporter plasmids generated for the analysis. HeLa cells were transfected with constructs pEGFP, pEGFP-HIF(5'), and pEGFP-HIF(3') (3 µg of plasmid per transfection mixture); 48 hours later, cells were harvested and cytoplasmic lysates were used in IP reactions with either mouse IgG or anti-PTB1. Following DNase treatment, RT-qPCR analysis was used to measure the amount of EGFP or chimeric EGFP mRNAs present in the IP reaction mixtures. The graph shows the means + standard errors of the means from three independent experiments.

Given the evidence that HuR and PTB associated with chimericRNAs carrying the HIF-1 ${alpha}$ 5'UTR and 3'UTR, respectively, we soughtto test if HuR and PTB also bound the endogenous HIF-1 ${alpha}$ mRNA.Employing lysates from untreated HeLa cells, RNP-IP assays revealeda $~$ 10-fold enrichment in HIF-1 ${alpha}$ mRNA when using an anti-HuR antibodycompared with control IP reactions using IgG (Fig. 5A). Whenusing lysates from CoCl₂-treated cells, the association of HIF-1 ${alpha}$ mRNA with HuR was a striking >150-fold higher than that seenin control IgG IP reactions (Fig. 5A). A substantial enrichmentin PTB-bound HIF-1 ${alpha}$ mRNA was also observed, particularly afterCoCl₂ treatment ( $~$ 80-fold more abundant in PTB IP than in IgGIP). As a positive control, the binding of VEGF mRNA to HuRwas tested, revealing $~$ 60-fold-increased binding after exposureto CoCl₂; PTB also associated with the VEGF mRNA (as previouslyreported [21]), showing an enrichment of $~$ 14-fold (Fig. 5A).As a negative control, the abundance of the housekeeping (nontarget)GAPDH mRNA in the HuR IP or PTB IP was very low (comparableto the IgG IP) and was unchanged by CoCl₂ treatment (Fig. 5A).The relative enrichment of these mRNAs in each RNP-IP reactionwas also tested by RT followed by conventional PCR amplificationand was visualized on agarose gels (Fig. 5B). These RNP complexeswere also readily detected in another cell type, the human lungcarcinoma cell line A549, after hypoxia for 2.5 h, indicatingthat these associations were not specific to HeLa cells (Fig.5C). Other RNA-binding proteins tested showed little or no enrichmentcompared with IgG (like G3BP or TIA-1 [Fig. 5C]) or showed nohypoxia-inducible binding (TIAR [Fig. 5C]). Collectively, ourfindings indicate that PTB and HuR can bind to the 3'UTR and5'UTR of HIF-1 ${alpha}$ mRNA, and their association with the endogenousHIF-1 ${alpha}$ mRNA increases following CoCl₂ treatment.

View larger version (17K):
[in this window]
[in a new window]

FIG. 5. Binding of endogenous RNA-binding proteins to endogenous mRNAs. (A) Lysates from HeLa cells that were either left untreated or treated with CoCl₂ (2.5 h) were used for RNP-IP analysis (see Materials and Methods). The association of endogenous HuR and PTB with the mRNAs shown was tested by RT-qPCR and is expressed as the relative enrichment compared with the levels of each mRNA in IgG samples. (B) RNP-IP samples were prepared as described for panel A, and the relative enrichment in HIF-1 ${alpha}$ , VEGF, and GAPDH mRNAs was tested by RT followed by conventional PCR amplification (28 cycles). PCR products were visualized by eletrophoresis in ethidium bromide-containing agarose gels. (C) Lysates from the lung carcinoma cell line A549 were prepared following either no treatment (normoxia) or hypoxia (1% O₂) for 2.5 h. The enrichment of endogenous HIF-1 ${alpha}$ mRNA in RNPs containing the RNA-binding proteins HuR, TIAR, TIA-1, or G3BP was tested by RNP-IP followed by RT-qPCR as explained for panel A. Graphs in panels A and C show results from three independent experiments (means + standard errors of the means).

PTB promotes the binding of HuR to HIF-1 ${alpha}$ mRNA. Since earlier studies indicated that PTB stimulated the translationof HIF-1 ${alpha}$ in hypoxic cells (54), we hypothesized that PTB performedthis function as part of a multiprotein RNP complex with theHIF-1 ${alpha}$ mRNA, possibly involving HuR. First, we examined if HuRformed direct protein-protein associations with PTB in an RNA-independentmanner. Indeed, HuR was found to coimmunoprecipitate with PTB(Fig. 6A) and, conversely, PTB was found to coimmunoprecipitatewith HuR (Fig. 6B). However, digestion with RNases A and T₁inhibited these associations, indicating that HuR and PTB formedcomplexes with shared RNAs but the two proteins did not interactdirectly.

View larger version (52K):
[in this window]
[in a new window]

FIG. 6. Absence of a protein-protein interaction between PTB and HuR. For co-IP analysis, whole-cell lysates were prepared from HeLa cells and then they were either left untreated (-) or incubated with RNase A and RNase T₁ for 10 min (+). IP was carried out using either IgG or antibodies recognizing the proteins indicated, and coimmunoprecipitated proteins were detected by Western blotting (WB) analysis of HuR (A) and PTB (B). Input, 2% of the material used for IP; H.C., heavy immunoglobulin chain; L.C., light immunoglobulin chain.

To assess if HuR and PTB were functionally linked as HIF-1 ${alpha}$ translationalinducers, we began by testing the influence of PTB upon HIF-1 ${alpha}$ translation. We reduced PTB expression levels in HeLa cellsby using RNA interference, achieved by transfection of siRNAtargeting the PTB mRNA (see Materials and Methods). As shownin Fig. 7A, PTB was effectively silenced, particularly in cytoplasmicextracts; in keeping with previous studies using hypoxia (54),the silencing of PTB strongly inhibited HIF-1 ${alpha}$ expression inCoCl₂-treated HeLa cells (Fig. 7A). To test if the levels ofPTB affected HuR function, we examined the abundance of HuR-HIF-1 ${alpha}$ mRNA complex RNPs in cultures with normal or silenced PTB levels.PTB indeed appeared to be necessary for maximal binding of HuRin CoCl₂-treated cells, since in PTB siRNA populations, HuR-HIF-1 ${alpha}$ mRNA complex RNPs were diminished to approximately one-halfof the RNP levels seen in control siRNA populations (Fig. 7B).These differences in HuR-HIF-1 ${alpha}$ mRNA complex abundance were assessedby RT followed by both real-time qPCR (Fig. 7B, top) and conventionalPCR analyses (Fig. 7B, bottom). No changes in the steady-statelevels of HIF-1 ${alpha}$ mRNA were observed, as these remained unalteredamong the transfection groups (Fig. 7C). Together, these findingsindicate that binding of HuR to the HIF-1 ${alpha}$ mRNA decreases whenPTB levels are reduced and support the view that PTB and HuRcooperate in promoting binding to HIF-1 ${alpha}$ mRNA.

View larger version (41K):
[in this window]
[in a new window]

FIG. 7. PTB silencing decreases HIF-1 ${alpha}$ levels and lowers HuR binding to HIF-1 ${alpha}$ mRNA. (A) HeLa cells were transfected with PTB-targeting or control (Ctrl.) siRNAs; 72 h later, cells were either left untreated (Untr.) or treated with CoCl₂ (2.5 h), and the expression of PTB, HIF-1 ${alpha}$ , and loading control β-actin was assessed by Western blot analysis in CE and NE. After silencing, the remaining PTB levels were $~$ 5% in CE and $~$ 20% in NE. (B) PTB-silenced (or control) HeLa cells were treated as described for panel A, and the levels of endogenous HuR-HIF-1 ${alpha}$ mRNA RNP complexes were tested by RNP-IP analysis followed by RT-qPCR (top left); the data are shown as relative enrichment of HIF-1 ${alpha}$ in the HuR IP compared with the IgG IP and normalized to background GAPDH mRNA levels in the IP materials (which were unchanged among the samples [data not shown]). Conventional RT-PCR analysis was also performed in order to visualize the relative amplification of PCR products from each RNP-IP reaction using ethidium bromide-stained agarose gels (bottom left). (C) In the same cultures, total HIF-1 ${alpha}$ mRNA levels were tested by RT-qPCR and normalized to GAPDH mRNA levels. Graphs represent the means and standard errors of the means from three independent experiments.

HuR promotes the binding of PTB to HIF-1 ${alpha}$ mRNA and increases HIF-1 ${alpha}$ protein abundance after CoCl₂ treatment. Next, we tested if modulating HuR levels influenced HIF-1 ${alpha}$ expression.First, we overexpressed HuR by transfecting cells with a vectorthat expressed HuR linked to a tandem affinity purification(TAP) tag (HuR-TAP) (33). By 72 h after transfection, HuR overexpressionin untreated cells had a negligible effect on HIF-1 ${alpha}$ production;by contrast, the CoCl₂-triggered rise in HIF-1 ${alpha}$ expression wasenhanced in the presence of elevated HuR levels (Fig. 8A). Again,the positive effect of HuR overexpression on HIF-1 ${alpha}$ abundancewas most clearly observed in CE, likely because changes in translationare reflected first in the cytoplasm. The positive influenceof HuR upon HIF-1 ${alpha}$ expression levels in CoCl₂-treated cultureswas due, at least in part, to the enhanced association of HIF-1 ${alpha}$ mRNA with actively translating polysomes. In CoCl₂-treated cultures,the distribution of HIF-1 ${alpha}$ mRNA along sucrose gradients shiftedtowards heavier polysome fractions in the presence of elevatedHuR levels, both when considering individual fractions (Fig.8B, inset) and when pooling fractions by translational activityinto untranslated (nonpolysomal [NP]) fractions, moderatelytranslated (low-molecular-weight polysomes [LMWP]) fractions,or actively translated (high-molecular-weight polysomes [HMWP])fractions (Fig. 8B). In additional experiments, a more accentuatedoverexpression of HuR was achieved by infection with an adenoviralvector (AdHuR). This intervention also elevated HIF-1 ${alpha}$ proteinexpression levels. HIF-1 ${alpha}$ expression was modestly elevated whenusing low PFU of AdHuR (Fig. 8C, left) but was strongly elevatedat high AdHuR PFU (Fig. 8C, right).

View larger version (41K):
[in this window]
[in a new window]

FIG. 8. HuR overexpression increases the association of HIF-1 ${alpha}$ mRNA with actively translating polysomes and elevates HIF-1 ${alpha}$ abundance. (A) HeLa cells were transfected with a plasmid to overexpress a tagged version of HuR (pHuR-TAP) (see Materials and Methods) or with the corresponding control plasmid (pTAP); 72 h later, cells were either left untreated (Untr.) or treated with CoCl₂ (2.5 h), and the expression of HuR-TAP, endogenous HuR, HIF-1 ${alpha}$ , and the loading control β-actin was assessed by Western blot analysis in WCE, CE, and NE. In the 12% gel, 5 µg was loaded per lane; in the 8% gel, 40 µg was loaded per lane for WCE, 80 µg for CE, and 3 µg for NE. (B) HeLa cells that were transfected and treated as explained for panel A were fractionated through sucrose gradients, and the presence of HIF-1 ${alpha}$ mRNA in each fraction was tested by RT-qPCR (inset). The bar graph represents the fractions pooled according to translational activity: untranslated (NP, nonpolysomal) fractions, moderately translated (LMWP) fractions, or actively translated (HMWP) fractions. Data represent the averages of two independent experiments showing similar results. (C) Lysates were prepared from HeLa cells that had been infected with AdHuR 24 h earlier in order to overexpress HuR and then either left untreated or treated with CoCl₂ for an additional 2.5 h. The expression of HuR, HIF-1 ${alpha}$ , β-actin, and ${alpha}$ -tubulin was assessed by Western blot analysis in cells infected with low PFU (left) or high PFU (right, showing images after short and long exposures). Note that HIF-1 ${alpha}$ protein levels were tested in whole-cell lysates, where differences are less pronounced than in CE.

Second, we reduced HuR expression by RNA interference. As shown,silencing of HuR by transfecting HeLa cells with HuR-directedsiRNAs (see Materials and Methods) lowered HuR abundance andstrongly diminished the CoCl₂-induced HIF-1 ${alpha}$ levels (Fig. 9A);as with PTB silencing (Fig. 7A), these effects were observedmore prominently in the cytoplasm (Fig. 9A, CE). This interventionalso reduced the binding of PTB to the HIF-1 ${alpha}$ mRNA (Fig. 9B)but did not affect HIF-1 ${alpha}$ mRNA levels (Fig. 9C). Together withthe effect of PTB silencing (Fig. 7), our results support acooperative association of HuR and PTB with the HIF-1 ${alpha}$ mRNA,since reducing HuR levels diminishes the binding of PTB to HIF-1 ${alpha}$ mRNA, while reducing PTB levels lowers HuR binding to the HIF-1 ${alpha}$ mRNA (Fig. 7B and 9B).

View larger version (38K):
[in this window]
[in a new window]

FIG. 9. HuR silencing decreases HIF-1 ${alpha}$ levels and lowers PTB binding to HIF-1 ${alpha}$ mRNA. (A) HeLa cells were transfected with HuR-targeting or control (Ctrl.) siRNAs; 72 h later, cells were either left untreated (Untr.) or treated with CoCl₂ (2.5 h), and the expression of HuR, HIF-1 ${alpha}$ , and the loading control β-actin was assessed by Western blot analysis in CE and NE. Silencing reduced HuR levels to 10 to 15% of the levels seen in the control siRNA groups. (B) HuR-silenced (or control) HeLa cells were treated as described for panel A, and the levels of endogenous PTB-HIF-1 ${alpha}$ mRNA RNP complexes were tested by RNP-IP analysis followed by RT-qPCR (top left). The data are shown as relative enrichment of HIF-1 ${alpha}$ in PTB IP compared with IgG IP and normalized to background GAPDH mRNA levels in the IP materials (which were unchanged among the samples [data not shown]). Conventional RT-PCR analysis was also performed in order to visualize the relative amplification of PCR products from each RNP-IP reaction using ethidium bromide-stained agarose gels (bottom left). (C) In the same cultures, total HIF-1 ${alpha}$ mRNA levels were tested by RT-qPCR and normalized to GAPDH mRNA levels. Graphs represent the means and standard errors of the means from three independent experiments.

Lowering HIF-1 ${alpha}$ expression levels after HuR or PTB silencing is associated with a reduction in HIF-1 ${alpha}$ nascent translation and with decreased expression of HIF-1 ${alpha}$ transcriptional target genes. The decreases in HIF-1 ${alpha}$ levels after PTB or HuR silencing (Fig.7 and 9, respectively) were due, at least in part, to the reducedassociation of HIF-1 ${alpha}$ mRNA with actively translating fractionsof the cytoplasm in the silenced cultures. As shown, the globaltranslational profiles of CoCl₂-treated cells did not differsignificantly among the silenced groups (Fig. 10A). However,the HIF-1 ${alpha}$ mRNA profile curves shifted leftward, albeit modestly,indicating a reduction in translational activity; pooled populationsagain showed that the HIF-1 ${alpha}$ mRNA was more abundant in LMWP fractionsand less abundant in HMWP fractions (Fig. 10B). Importantly,nascent HIF-1 ${alpha}$ translation was significantly reduced in HeLacells in which HuR or PTB was silenced; the incorporation of³⁵S-labeled amino acids into de novo-synthesized HIF-1 ${alpha}$ was $~$ 15%and 22%, respectively, relative to the incorporation measuredin the control siRNA-transfected cultures (Fig. 10C). Incorporationof ³⁵S-labeled amino acids into the housekeeping protein GAPDHwas not dependent on PTB or HuR levels (see Fig. S3 in the supplementalmaterial).

View larger version (35K):
[in this window]
[in a new window]

FIG. 10. HuR or PTB silencing decreases HIF-1 ${alpha}$ translation and lowers expression levels of HIF-1 ${alpha}$ transcriptional targets VEGF and GLUT1. (A) HeLa cells were treated with CoCl₂ after silencing for 48 h with the siRNAs indicated and then the polysome profiles were studied. (B) HIF-1 ${alpha}$ mRNA levels were measured in each fraction, and the relative abundance of the HIF-1 ${alpha}$ mRNA was measured as described in the legend for Fig. 2B. The bar graph represents the fractions pooled according to translational activity: untranslated (NP, nonpolysomal, comprising fractions which lacked ribosome components, and fractions containing ribosome subunits), moderately translated (monosomes and LMWP), or actively translated (HMWP). Data represent the averages of two independent experiments showing similar results. (C) De novo HIF-1 ${alpha}$ translation was monitored in HeLa cells that were transfected with control (Ctrl.) siRNA, HuR siRNA, or PTB siRNA. Cultures were subjected to a brief (15-min) incubation with L-[³⁵S]methionine and L-[³⁵S]cysteine and collected. Following IP reactions using anti-HIF-1 ${alpha}$ antibody or IgG, the incorporation of radiolabeled amino acids into the newly synthesized HIF-1 ${alpha}$ protein was assessed by electrophoresis through 8% SDS-containing gels. Signals were visualized and quantified using a PhosphorImager, and the incorporation of ³⁵S-labeled amino acids into nascent HIF-1 ${alpha}$ is indicated as a percentage of the incorporation seen in cells transfected with control siRNA and treated with CoCl₂. GAPDH nascent translation was not altered by HuR silencing, PTB silencing, or CoCl₂ treatment (see Fig. S3 in the supplemental material). (D) Forty-eight hours after transfection with either HuR- or PTB-directed siRNAs, HeLa cells were left untreated or treated with CoCl₂ (200 µM) and the levels of the transcriptional targets VEGF and GLUT1 were measured by RT-qPCR analysis 8 h later. Transcript levels were normalized to those of housekeeping GAPDH mRNA and are presented as fold differences in mRNA levels in CoCl₂-treated cells relative to untreated cells. Data represent the means and standard errors of the means from three independent experiments.

The reduction in HIF-1 ${alpha}$ expression levels in cells with silencedHuR or PTB (Fig. 7, 9, and 10C) led to a diminution in the levelsof the HIF-1 ${alpha}$ transcriptional targets VEGF and GLUT1 (Fig. 10D).Together with the HuR overexpression experiments, our findingssupport a model whereby HuR and PTB positively influence thetranslation of HIF-1 ${alpha}$ in response to CoCl₂ treatment. These regulatoryprocesses in turn influence HIF-1 ${alpha}$ -directed gene expression programs.

DISCUSSION

Top

DISCUSSION

We report here that HIF-1 ${alpha}$ translation increases after cellsare challenged with the hypoxia mimetic CoCl₂, and we show thatthis effect depends, at least in part, on the association ofHIF-1 ${alpha}$ mRNA with the RNA-binding proteins HuR and PTB. PTB waspreviously shown to bind to the mouse HIF-1 ${alpha}$ IRES, thereby increasingHIF-1 ${alpha}$ translation (54). While we also found IRES activity withinthe human HIF-1 ${alpha}$ 5'UTR (Fig. 3), this IRES was not responsiblefor promoting HIF-1 ${alpha}$ translation after CoCl₂ treatment. Instead,modulation of the levels of RNA-binding proteins HuR and PTBdid influence HIF-1 ${alpha}$ translation after CoCl₂ treatment, withoverexpression of HuR elevating HIF-1 ${alpha}$ biosynthesis and silencingof PTB or HuR reducing it. Moreover, HuR-HIF-1 ${alpha}$ mRNA complexeswere diminished in PTB-depleted cells and, likewise, PTB-HIF-1 ${alpha}$ mRNA complexes were decreased in HuR-depleted cells, furtherrevealing a functional cooperation between PTB and HuR duringthe CoCl₂-induced translation of HIF-1 ${alpha}$ .

Treatment with the hypoxia mimetic CoCl₂ does not affect steady-state HIF-1 ${alpha}$ mRNA levels but elevates HIF-1 ${alpha}$ protein levels. In agreement with earlier literature, neither hypoxia nor CoCl₂seemed to enhance HIF-1 ${alpha}$ gene transcription or elevate HIF-1 ${alpha}$ mRNA stability, since HIF-1 ${alpha}$ mRNA levels remained unchanged afterthese treatments (Fig. 1). Formal measurement of the HIF-1 ${alpha}$ mRNAhalf-life using actinomycin D further confirmed that there wereno CoCl₂-dependent changes (not shown). Instead, the potentinduction of HIF-1 ${alpha}$ protein expression likely results from acombination of two distinct processes: increased HIF-1 ${alpha}$ translationand decreased HIF-1 ${alpha}$ proteolysis. CoCl₂ treatment can induceHIF-1 ${alpha}$ protein levels through its inhibitory effect on prolylhydroxylases and thereby block the proteasome-dependent degradationof HIF-1 ${alpha}$ (7, 13, 61). In addition, CoCl₂ treatment can alsopromote HIF-1 ${alpha}$ translation, as reported here. While bona fidehypoxia is arguably a more physiologic stress, gaining the molecularunderstanding of these processes and elucidating the functionof HIF-1 ${alpha}$ RNPs using hypoxia would have been hampered by spaceand methodological limitations imposed by the use of hypoxicchambers/incubators. However, considering the similarities betweenthe mechanisms of action of CoCl₂ and hypoxia, we anticipatethat hypoxia-regulated HIF-1 ${alpha}$ expression will also prove to havea significant translational component, likely modulated by PTBand HuR.

CoCl₂ treatment increases binding of HuR to HIF-1 ${alpha}$ mRNA and HuR-dependent HIF-1 ${alpha}$ translation. The mechanisms whereby CoCl₂ treatment increased HuR bindingto HIF-1 ${alpha}$ mRNA are unclear, but neither the whole-cell abundanceof HuR nor its cytoplasmic levels (Fig. 8A and data not shown)were elevated in response to CoCl₂. These findings were in contrastto what we and others have observed regarding this shuttlingprotein, as its cytoplasmic concentration increases markedlyfollowing exposure to numerous other stress agents, likely aresult of HuR increased cytoplasmic export or reduced nuclearimport. In addition, a variety of damaging agents induced theaggregation of cytoplasmic, HuR-containing stress granules (reviewedin reference 28). By contrast, CoCl₂ treatment neither elevatedthe cytoplasmic HuR levels nor induced HuR aggregation intostress granules (see Fig. S4 in the supplemental material).These observations support the notion that HuR binding to theHIF-1 ${alpha}$ mRNA is influenced by posttranslational processes thatoccur in the absence of net changes in nucleocytoplasmic distribution.Several regulatory mechanisms have been described which influence,directly or indirectly, the abundance of HuR-containing RNPcomplexes. For example, the translational suppression imposedby microRNA miR-122 upon the CAT-1 mRNA was relieved followingdisplacement by HuR in response to stress conditions (2), theactivity of the checkpoint kinase Chk2 influenced HuR bindingto the SIRT1 mRNA after oxidative stress (1), and HuR methylationby coactivator-associated arginine methyltransferase 1 affectedHuR binding to tumor necrosis factor alpha mRNA in lipopolysaccharide-treatedcells (35). In addition, PI3K and the mitogen-activated proteinkinase p38 influenced HuR function (44, 58), while the activityof mitogen-activated protein kinase extracellular signal-regulatedkinase (ERK) promoted HuR binding to target p21 mRNA in cellstreated with prostaglandin A₂, as ERK inhibition diminished[HuR-p21 mRNA] complexes but not the cytoplasmic abundance ofHuR (66). The latter effect might be implicated in the CoCl₂-triggeredincrease in HuR binding to HIF-1 ${alpha}$ mRNA, since CoCl₂ treatmentactivates ERK (65). A more in-depth analysis of the mechanismsresponsible for the CoCl₂-induced HuR-HIF-1 ${alpha}$ mRNA complexes iswarranted.

CoCl₂ treatment promotes PTB binding to HIF-1 ${alpha}$ mRNA and PTB-dependent HIF-1 ${alpha}$ translation. Similar questions arise regarding the influence of CoCl₂ uponPTB binding to HIF-1 ${alpha}$ mRNA. What are the cellular effectors ofthe CoCl₂-enhanced formation of PTB-HIF-1 ${alpha}$ mRNA complexes? Thesignaling pathways that are activated in response to CoCl₂ treatmenthave not been fully delineated, but exposure to cobalt activatesthe production of reactive oxygen species and induces PI3K activitylinked to increased HIF-1 ${alpha}$ expression; the latter events aresignificant because PI3K/mTOR/p70 S6 kinase signaling has beenlinked to increased HIF-1 ${alpha}$ translation (6, 10). However, whetherCoCl₂ regulates the function of PTB via any of the reportedkinase pathways that control HIF-1 ${alpha}$ translation, dependentlyor independently of mTOR (3, 22, 47, 49), remains to be investigated.It is also possible that cobalt itself directly elicits structuralchanges in the HIF-1 ${alpha}$ mRNA, as reported for other metal ions,such as iron and magnesium, which can affect mRNA tertiary structure(17, 27, 63). Thus, the possibility remains that cobalt directlyalters the accessibility of RNA-binding proteins to the HIF-1 ${alpha}$ mRNA and thereby facilitates translation.

Cooperative binding of HuR and PTB. Although HuR was found to bind the HIF-1 ${alpha}$ 5'UTR and PTB was foundto bind the HIF-1 ${alpha}$ 3'UTR, the specific region(s) of associationof HuR and PTB on the human HIF-1 ${alpha}$ mRNA also awaits more detailedstudy. It was surprising to discover that the HIF-1 ${alpha}$ 3'UTR didnot associate with HuR (Fig. 4), despite the predicted existenceof three HuR motif hits (38) and a previous report of HuR-HIF-1 ${alpha}$ 3'UTR complexes (57). HuR was shown to enhance the translationof numerous target mRNAs, including those that encode prothymosin ${alpha}$ , p53, cytochrome c, GLUT1, and CAT-1 (2, 16, 26, 33, 42). Paradoxically,HuR was also shown to inhibit the translation of other targetmRNAs, such as those encoding cytokines, p27, and IGF-IR (25,31, 43). In the cases in which these effects were characterizedin molecular detail, HuR was found to associate via the 5'UTRof target mRNAs which had IRESs (31, 43). In this regard, givenour findings described here that HuR bound the IRES-containing5'UTR of HIF-1 ${alpha}$ , we anticipated that HuR would trigger the inhibition,not the promotion, of HIF-1 ${alpha}$ translation. Finally, the functionalinteraction between HuR and PTB upon the translation of HIF-1 ${alpha}$ is reminiscent of those described for Apaf-1 and the human rhinovirusRNA, whose translation is potentiated through binding of PTBto the RNA-binding protein Unr (upstream of N-ras) (24, 45).

In sum, these studies have identified HuR and PTB as two RNA-bindingproteins that function jointly to stimulate HIF-1 ${alpha}$ translationin response to CoCl₂ treatment. The translational control ofHIF-1 ${alpha}$ expression complements the previously reported influenceof CoCl₂ on HIF-1 ${alpha}$ protein stability (7, 13, 61). The concomitantincrease in HIF-1 ${alpha}$ biosynthesis and inhibition of HIF-1 ${alpha}$ degradationunderscore the need for rapid and efficient increases in HIF-1 ${alpha}$ abundance following environmental changes. The expression ofa growing number of additional genes has been found to be regulatedsimultaneously at multiple levels, transcriptional, posttranscriptional,translational, and posttranslational. Such extensive regulatorynetworks afford the appropriate degree of redundancy, checks,and safeguards to ensure that critical proteins such as HIF-1 ${alpha}$ are expressed in the necessary abundance, time, and locationto maintain cellular homeostasis.

ACKNOWLEDGMENTS

We thank S. D. Baird and B. Frank for their assistance withthis study.

This research was supported by the Intramural Research Programof the National Institute on Aging, National Institutes of Health.

FOOTNOTES

* Corresponding author. Mailing address: LCMB, NIA, NIH, 5600 Nathan Shock Dr., Baltimore, MD 21224. Phone: (410) 558-8443. Fax: (410) 558-8386. E-mail: myriam-gorospe{at}nih.gov

Published ahead of print on 29 October 2007.

Supplemental material for this article may be found at http://mcb.asm.org/.

REFERENCES

Top