J Biol Chem, Vol. 275, Issue 10, 7430-7437, March 10, 2000

A Conserved Mechanism for Controlling the Translation of -F1-ATPase mRNA between the Fetal Liver and Cancer Cells^*

Miguel López de Heredia, José M. Izquierdo, and José M. Cuezva^§

From the Departamento de Biología Molecular, Centro de Biología Molecular "Severo Ochoa," Universidad Autónoma de Madrid, 28049 Madrid, Spain

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To characterize the mechanisms governing the biogenesis of mitochondria in cancer, we studied the mitochondrial phenotype and the mechanisms controlling the expression of the  subunit of the mitochondrial H⁺-ATP synthase (-F1-ATPase) gene in the rat FAO and AS30D hepatomas. When compared with normal adult rat liver, the relative cellular content of the mitochondrial -F1-ATPase and glutamate dehydrogenase, as well as of mitochondrial DNA, was severely reduced in both cell lines. A paradoxical increase in the cellular abundance of -F1-ATPase mRNA was observed in cancer cells. Run-on transcription assays and the estimation of mRNA half-lives revealed that the increased abundance of -F1-ATPase mRNA results from the stabilization of the transcript in cancer. In vitro translation assays revealed a specific inhibition of the synthesis of the -precursor when translation reactions were carried out in the presence of extracts derived from cancer cells. The inhibitory effect was recapitulated using an RNA chimera that contained the 3'-untranslated region of -F1-ATPase mRNA. Hepatoma extracts also contained an increased activity of the developmentally regulated translation-inhibitory proteins that bind the 3'-untranslated region of -F1-ATPase mRNA. The results indicate that the expression of this gene in hepatoma cells is controlled by the same mechanisms that regulate its expression in the liver during fetal development.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

A common feature of the phenotype of many cancer cells is their abnormal bioenergetics, i.e. an increased glycolytic capacity accompanied by an impaired respiration (1-5). The abnormal respiratory capacity of cancer cells has been ascribed to a marked deficit in the cellular content of mitochondria (4, 5). However, the molecular mechanism whereby the program of mitochondrial biogenesis is altered in cancer is unknown. In this regard, it has been reported that poorly differentiated hepatomas have a large reduction in their mitochondrial content (4, 6) despite showing a paradoxical increase in the cellular abundance of transcripts encoding mitochondrial proteins (6). The paradoxical up-regulation of oxidative phosphorylation transcripts has also been noted in estrogen-induced hepatocarcinogenesis (7, 8), in other cancer cells (9, 10), as well as in cellular lines transformed with viral and cellular oncogenes (9, 11, 12).

Mitochondrial biogenesis results from the coordinated expression of the nuclear and mitochondrial genomes (13). In the fetal liver it has been described that the mitochondrial content of the hepatocyte is very much reduced when compared with the adult (14-16). Remarkably, the fetal liver shows a paradoxical increase in the cellular representation of oxidative phosphorylation transcripts (5, 15-18). During development of the liver, the expression of the nuclear -F1-ATPase¹ gene of oxidative phosphorylation is exerted at the levels of the stability (16) and translation (15, 19) of the transcript. Concurrently, the control of the expression of the mitochondrial genome is also exerted at the same two post-transcriptional levels (17, 18). Regulation of translation of the -F1-ATPase gene in the liver involves two mechanisms that use the 3'-UTR of the transcript as target element (19). The 3'-UTR of -F1-ATPase mRNA is a translational enhancer sequence required for mRNA translation and, in addition, provides the cis-acting sequence where developmentally regulated inhibitory proteins bind for controlling the expression of the mRNA (19).

Recently, we suggested that the abnormal biogenesis of mitochondria in hepatomas might represent a reversion to a fetal program of expression of oxidative phosphorylation genes (5). This would imply the activation of the mechanisms controlling the stability and translation of essential mRNAs required for organelle biogenesis in hepatomas. In the present investigation we have studied the feasibility of this hypothesis. To this end, we have analyzed the mitochondrial phenotype of the rat FAO and AS30D hepatoma cell lines. Both hepatomas have a reduced mitochondrial content when compared with the adult rat liver. The results indicate the operation in cancer cells of the same post-transcriptional mechanisms that control the expression of the -F1-ATPase gene in the liver during fetal development. Remarkably, hepatoma extracts also contained an increased activity of translation-inhibitory proteins that bind the 3'-UTR of -F1-ATPase mRNA. The results provide the first mechanistic indications that explain the paradoxical abnormal biogenesis of mitochondria in cancer cells. It is suggested that the reduction in the mitochondrial content of cancer cells derives from the dilution of mitochondrial constituents during cellular proliferation as a result of the translational arrest experienced by essential mitochondrial components required for the biogenesis of mitochondria.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Animals and Tumor Cell Lines-- Adult female albino Wistar rats weighing 200 g were fed standard laboratory chow and water ad libitum. Fetuses were obtained by hysterectomy from cervically dislocated pregnant rats on day 22 of gestation (20). Adult and fetal tissues were collected after decapitation and exsanguination of the animals.

Cells were routinely cultured at 37 °C in plastic dishes with 93% air and 7% CO₂. Rat hepatoma AS30D cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal calf serum, 2 mM glutamine, 100 units/ml penicillin, 0.1 mg/ml streptomycin, and 400 µM non-essential amino acids. Monolayer cultures of FAO hepatoma cells were grown in F12 Coon's modified media supplemented with 5% heat-inactivated fetal calf serum, 100 units/ml penicillin, and 0.1 mg/ml streptomycin.

Preparation of Liver and Cell Homogenates-- Liver homogenates were prepared as described previously in detail (14). Homogenates from the tumor cell lines were prepared after washing the collected cells with ice-cold phosphate-buffered saline. Briefly, the washed cells were resuspended and swelling occurred in 3 volumes of a medium containing 10 mM Tris-HCl, pH 7.4, for 10 min. Cells were homogenized by 30 strokes of a 0.0005-inch clearance pestle (Kontes). After cellular lysis had occurred, 1 volume of a medium containing 80 mM Tris-HCl, pH 7.4, 1.75 M sucrose, 7 mM EGTA, 700 µM phenylmethylsulfonyl fluoride, 70 µg/ml leupeptin, 35 µg/ml pepstatin A, 119 µg/ml aprotinin, and 700 µM benzamidin was added, and the cells were further homogenized by another 10 strokes with the same pestle. Liver and cell homogenates were centrifuged at 800 × g for 10 min to remove nuclei and unbroken cells. Homogenates were stored frozen at 70 °C until used.

Cytoplasmic extracts from cell lines and rat tissues were prepared by centrifugation of post-mitochondrial supernatants at 180,000 × g for 1 h (19). The resulting supernatants were dialyzed against 10 mM Tris-HCl, pH 7.4 (molecular weight cut-off of the membrane, 6,000-8,000). When necessary, the dialyzed extracts were concentrated using Centricon-10 concentrators (Amicon Inc.) and stored at 70 °C until used (19).

DNA Isolation and Southern Blot Hybridization-- Total DNA was extracted from adult rat liver and hepatoma cells after digestion with RNase and proteinase K as described previously in detail (16, 18). Total cellular DNA (10 µg) was digested with BamHI. The digested DNAs were resolved on 0.8% agarose gels, transferred, and fixed onto nylon membranes (GeneScreen, NEN Life Science Products) (16, 18). The membranes were incubated with [³²P]dCTP-labeled DNA probes. The rat liver DNA probes used in this study were -F1-ATPase cDNA, for a nuclear-encoded gene, and specific DNA probes for the mitochondrial encoded ATPase 6-8 and 12 S rRNA genes (16, 18). Conditions for hybridization and membrane washing have been previously described in detail (16, 18). For stripping labeled DNA probes, membranes were incubated in sterile water at 90-100 °C for 20 min. Membranes were exposed to x-ray films and analyzed by laser densitometric scanning.

RNA Isolation and Northern Blot Hybridization-- Total RNA was obtained from adult rat liver and hepatoma cells (21). RNA samples were fractionated by electrophoresis on formaldehyde, 1.4% agarose gels (20). Denatured RNAs were transferred to GeneScreen membranes by vacuum transfer. The DNA probes were labeled by nick translation. The rat liver cDNA probes used in this study were as follows: -F1-ATPase (19, 22), -F1-ATPase (23), and GAPDH (24) for transcripts encoded in the nuclear genome. For transcripts encoded in the mitochondrial genome, the ATPase 6-8 and 12 S rRNA DNAs were used (25). For normalization of RNA loading in the gels, an oligonucleotide probe that hybridizes to the 18 S rRNA from rat liver was used (26). Hybridization conditions, membrane washing, stripping of labeled DNA probes, and the exposure and quantification of hybridization signals were as detailed above.

Isolation of Nuclei and Run-on Transcription Assays-- Nuclei were isolated from rat liver and hepatoma cells as described previously in detail (16, 27). The only modifications introduced in the protocol for the isolation of nuclei from hepatoma cells were the inclusion of 0.1% Triton X-100 to the lysis medium and the utilization of a 1.8 M sucrose cushion for the preparation of the nuclei. Run-on experiments were carried out with 2-3 × 10⁷ nuclei, following the modifications previously reported (16, 27). For specific detection of radioactive nascent RNA transcripts, 10-20 × 10⁶ cpm of the isolated RNA were hybridized for 60 min at 60 °C and for 72 h at 42 °C with the following plasmid DNAs immobilized on nylon filters: cytosolic phosphoenolpyruvate carboxykinase (PEPCK) (28), -F1-ATPase (19), -actin (29), and pBS (30). Immobilization of plasmid DNAs onto membranes and conditions for membrane washing were those previously described (16, 27). Membranes were exposed to x-ray films and analyzed by laser densitometric scanning.

Determination of RNA Half-lives in Cell Cultures-- Aproximately 5 × 10⁶ cells, for both the AS30D and FAO cell lines, were grown in the conditions described above with the following exceptions: the growth media were supplemented with 10 µg/ml actinomycin D, to inhibit DNA transcription, and contained 2% of heat-inactivated fetal calf serum. At the indicated time intervals, total cellular RNA was isolated and subsequently processed as described under Northern blot hybridization. RNA half-lives were calculated assuming a first-order rate kinetics (16, 17).

In Vitro Synthesis of RNA Transcripts-- The in vitro synthesis of RNAs was performed as described previously (19). For transcripts used in translation assays, the plasmids, linearized with HindIII, were transcribed using the mCap RNA capping kit (Stratagene) following the manufacturer's instructions. The plasmids (1 µg) used were as follows: pJMI--F1, to generate the full-length -F1-ATPase mRNA, and pARF-3'UTR, to generate the ARF-3'UTR mRNA (19). For the generation of RNA riboprobes, the plasmid pJMI3'UTR--F1 was digested with SacI, DraI, and HindIII to generate, respectively, the 3'-1, 3'-2, and wild-type riboprobes of the 3'-UTR of -F1-ATPase mRNA (19). Radiolabeled RNA probes were prepared, without adding cap structures, in the presence of 0.05 mM cold rUTP plus 50 µCi of [-³²P]UTP (400 Ci/mmol) (Amersham Pharmacia Biotech).

In Vitro Translation and Stability of RNA Transcripts-- In vitro synthesized mRNAs (20 µg/ml) were used as templates for synthesis of the encoded protein using nuclease-treated rabbit reticulocyte lysates (Amersham Pharmacia Biotech) (19). The 30-min reactions were performed in the presence of 20 µCi of L-[³⁵S]methionine (>1,000 Ci/mmol) (Amersham Pharmacia Biotech), 40 units of RNasin (Roche Molecular Biochemicals), and either in the presence or absence of 25 µg of the indicated cell extract. The products were analyzed by SDS-PAGE and further processed for fluorography (19).

To determine the stability of -F1-ATPase mRNA in in vitro translation assays, the radiolabeled transcript was added to translation reaction mixtures. After 0, 10, 20, and 30 min of reaction, a 5-µl aliquot of the translation mixture was withdrawn to determine the radioactivity remaining in trichloroacetic acid-precipitated material, following standard protocols.

RNA EMSA and UV Cross-linking Assays-- Band shift (EMSA) and UV cross-linking assays were carried out with the modifications previously described in detail (19). Briefly, 35 µg of protein from the extracts were incubated with radiolabeled probes (1-5 × 10⁵ cpm) at 30 °C for 10 min prior to the addition of 20 units of RNase T1 (Roche Molecular Biochemicals). In UV cross-linking assays the reaction mixtures were exposed to 254 nm of UV light (Stratalinker 1800; Stratagene) for 6 min on ice. The RNA-protein complexes were resolved on either 5% polyacrylamide, 0.5× Tris borate/EDTA native gels (EMSA) or SDS-12% PAGE (UV cross-linking). For competition studies, an excess of unlabeled RNA was added 10 min before the addition of the radiolabeled RNA (19). After electrophoresis, the gels were dried and exposed to x-ray films.

To characterize the proteins involved in the formation of the RNA-protein complexes from the band shift assays, extracts and riboprobes were UV cross-linked before the EMSA. Subsequently, the bands of interest were cut from the dried gels, hydrated, and further fractionated on SDS-12% PAGE.

Other Methods-- Western blots of mitochondrial proteins were carried out as described previously in detail (20). The primary antibodies used were as follows: rabbit anti--F1-ATPase from rat liver (31) and rabbit anti-glutamate dehydrogenase from bovine liver (Biogenesis, UK). The secondary antibody used was goat anti-(rabbit IgG)-peroxidase conjugate (Nordic Immunology, The Netherlands). Protein concentrations were determined with the Bradford reagent (Bio-Rad Protein assay) using bovine serum albumin as standard. Statistical analysis was performed by the Student's t test.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

FAO and AS30D Hepatomas Have a Low Mitochondrial Content-- In this study we have determined the relative mitochondrial content of two hepatoma cell lines derived from rat liver (AS30D and FAO) and compared it with that of normal adult rat liver. It has been described that the FAO hepatoma has a differentiated phenotype qualitatively resembling that of the normal hepatocyte (32, 33), whereas the AS30D hepatoma has a poorly differentiated phenotype (34, 35). Analysis of the cellular proteins in the adult liver and AS30D and FAO hepatomas revealed a closer protein profile between FAO and AS30D cells (Fig. 1).

View larger version (51K): [in this window] [in a new window]   Fig. 1.   FAO and AS30D hepatomas show a reduced cellular content of -F1-ATPase and glutamate dehydrogenase when compared with adult rat liver. Left panel, 50 µg of protein from adult rat liver (A.L.) and the AS30D and FAO hepatoma homogenates were fractionated on SDS-12% PAGE. A representative Coomassie Blue-stained gel is shown. Molecular mass markers (200, 116, 97, 66, 45, 31, 21, and 14 kDa) are shown on the left. Upper and middle panels show representative Western blots probed with polyclonal anti--F1-ATPase (-F1) and anti-glutamate dehydrogenase (GDH) antibodies. Lower panel shows the quantification of the relative cellular content of immunoreactive -F1-ATPase and GDH proteins in adult liver (black bars), AS30D (white bars), and FAO (gray bars) homogenates. Data are expressed in arbitrary units (a.u.)/mg cellular protein and are the means ± S.E. of 4-9 different samples. *, p < 0.05; **, p < 0.005 when compared with adult liver by Student's t test.

The mitochondrial content in liver and hepatoma cells was assessed by the quantification of (i) the relative cellular content of the specific mitochondrial proteins -F1-ATPase and GDH (Fig. 1) and (ii) the relative mitochondrial DNA content of the cell (Fig. 2). Both the -F1-ATPase and GDH content in hepatoma cell lines were found to be significantly reduced when compared with the normal adult liver (Fig. 1). It should be noted that within the two hepatomas, the AS30D showed a higher reduction in both the -F1-ATPase and GDH contents (Fig. 1). In line with the reduction observed in the cellular content of mitochondrial proteins, it was found that the relative mtDNA content of the hepatomas was also significantly reduced when compared with that of the normal adult liver (Fig. 2). However, the mtDNA content was not significantly different between AS30D and FAO hepatomas (Fig. 2).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   FAO and AS30D hepatomas have a reduced mtDNA content when compared with adult rat liver. Total cellular DNA from adult liver (A.L.) and the AS30D and FAO hepatomas was isolated, digested with BamHI, and processed as described under "Experimental Procedures." The membranes were hybridized subsequently with the nuclear and mitochondrial DNA probes for the rat nuclear -F1-ATPase (n--F1) and mitochondrial ATPase 6-8 (mt-A6-8) and 12 S (mt-12S) genes. The relative cellular content of mtDNA (arbitrary units, a.u.) is expressed as the ratio of the hybridization signal of 12 S or ATPase 6-8 to that of -F1-ATPase. The results shown are the means ± S.E. of 2-4 different preparations from adult liver (black bars), AS30D (white bars), and FAO (gray bars) hepatomas. *, p < 0.05; **, p < 0.01; ***, p < 0.005 when compared with adult liver by Student's t test.

Hepatomas Have a Paradoxical Increase in the Abundance of mRNAs of Oxidative Phosphorylation-- In order to analyze the molecular mechanism(s) that might be responsible for the aberrant program of the biogenesis of mitochondria in hepatomas (Figs. 1 and 2), we examined first the expression level of several nuclear and mitochondrially encoded RNAs involved in oxidative phosphorylation. The steady-state levels of the mRNAs encoding the  and  subunits of the F1-ATPase complex, normalized to the 18 S rRNA signals, were found to be significantly higher than in the normal adult liver (Fig. 3). Remarkably, it was found that the steady-state level of the -F1-ATPase mRNA was higher in AS30D than in FAO hepatomas (Fig. 3). In contrast to the observed up-regulation of nuclear-encoded transcripts, the steady-state level of the ribosomal 12 S rRNA was found to be increased slightly (35%) in the AS30D hepatoma, whereas that of the ATPase 6-8 mRNA showed no significant changes when compared with the normal rat liver in any of the hepatomas studied (Fig. 3). Consistent with the high glycolytic phenotype of the AS30D hepatoma (34, 35), it was found that the glycolytic GAPDH mRNA was much more represented in AS30D than in the FAO hepatoma or in the normal adult liver (Fig. 3).

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Steady-state RNA levels of oxidative phosphorylation in adult rat liver and hepatoma cells. 30 µg of total cellular RNA from adult liver (A.L.) and the AS30D and FAO hepatomas were analyzed by Northern blot hybridization procedures with the indicated rat DNA probes for the nuclear-encoded (18S for 18 S rRNA; -F1 and -F1 for - and -F1-ATPase and GAPDH) and mitochondrially encoded (12S for 12 S rRNA and A6-8 for ATPase 6-8) genes. For normalization purposes, the relative cellular content of each RNA (arbitrary units, a.u.) is expressed as the ratio of the hybridization signal of the RNA to that of the corresponding 18 S rRNA hybridization signal. The results shown are the means ± S.E. of 2-6 different RNA preparations from adult liver (black bar), AS30D (white bar), and FAO (gray bar) hepatomas. *, p < 0.05; **, p < 0.025; ***, p < 0.005 when compared with adult liver by Student's t test.

Transcription Rates of Oxidative Phosphorylation Transcripts in Hepatomas-- The higher cellular representation of the nuclear-encoded -F1-ATPase mRNA in hepatomas (Fig. 3) led us to examine whether this resulted from an increase in the transcription rate of the corresponding gene. Nuclear run-on assays in isolated nuclei from normal adult rat liver and from the AS30D and FAO hepatomas showed profound differences in the relative transcription rate of the gluconeogenic PEPCK gene, a marker of the differentiated phenotype of the normal hepatocyte (Fig. 4). Nuclei of FAO cells showed a 15-fold reduction in PEPCK transcription when compared with normal liver (Fig. 4). However, the synthesis of this transcript in FAO nuclei was still 10-fold above that found in the nuclei of AS30D cells (Fig. 4). The PEPCK transcription rates observed (normal liver > FAO > AS30D) are in agreement with the phenotype and degree of differentiation of the hepatoma cell lines (32, 35). In contrast to the transcription rate of the PEPCK gene, the transcription rates of both the -F1-ATPase and -actin genes in the nuclei of normal adult liver were found to be not significantly different from the transcription rates in nuclei of the AS30D and FAO cells (Fig. 4).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Transcription rates of the -F1-ATPase gene in isolated nuclei of rat liver and hepatoma cells. Nuclei were isolated from adult rat liver (A.L. and the black bar), AS30D (white bar), and FAO (gray bar) cells. Nuclear run-on assays were carried out to determine the relative transcription rates of the gluconeogenic PEPCK, -F1-ATPase (-F1), and -actin genes. In each assay, the hybridization signals obtained were corrected by the nonspecific hybridization of radiolabeled RNA to plasmid DNA (pBS). Upper panel, representative autoradiograms hybridized with radiolabeled nascent RNA transcripts. Lower panel, the histogram illustrates the relative transcription rates (arbitrary units, a.u./106 cpm) of the genes after densitometric scanning of the resulting hybridization signals. The results are the means ± S.E. of 2-4 independent experiments. *, p < 0.025 when compared with adult liver by Student's t test.

Increased Stability of Oxidative Phosphorylation Transcripts in Hepatomas-- Developmental changes in the stability of the mRNAs involved in oxidative phosphorylation have been recently demonstrated to be responsible for the paradoxical increase in availability of these transcripts in the fetal liver (16, 17). To examine the operation of such a mechanism in cancer cells, we next estimated the half-life of several nuclear-encoded and mitochondrially encoded transcripts in the FAO and AS30D cells. Previously, we have reported the estimated half-life of these transcripts in the neonatal and adult rat liver (16, 17). The results obtained illustrated a longer half-life of the nuclear-encoded -F1-ATPase mRNA in both hepatomas when compared with the normal adult liver (Fig. 5 and also see Fig. 6 in Ref. 16). In fact, it was found that the stability of -F1-ATPase mRNA decreased in the following order: neonatal liver (8 h), AS30D hepatoma (7 h), FAO hepatoma (5 h), and adult liver (2 h). Similarly, the hepatomas also revealed an increased half-life for the mitochondrially encoded ATPase 6-8 mRNA when compared with that found in the adult liver (Fig. 5 and also see Fig. 1 in Ref. 17). For this transcript, it was found that its half-life varied in the following order: neonatal liver (4 h), FAO hepatoma (5 h), AS30D hepatoma (3 h), and adult liver (2 h). In contrast, the hybridization signals for both the nuclear-encoded GAPDH mRNA and the mitochondrially encoded 12 S rRNA revealed no significant decrease by actinomycin D treatment in any of the hepatomas studied (Fig. 5). Although this situation prevented the estimation of the half-life for the latter transcripts, it should be noted that the 12 S rRNA half-life in hepatoma cells also appeared to be significantly increased when compared with normal adult (4 h) and neonatal (9 h) liver (see Fig. 5 and Refs. 16 and 17). Since the transcription rates of the -F1-ATPase gene were not significantly altered in hepatomas (Fig. 4), it is suggested that the higher cellular availability of -F1-ATPase mRNA observed in cancer cells (Fig. 3) is caused by the modification of the turnover of the transcript (Fig. 5) as a result of carcinogenesis.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   RNA half-lives of oxidative phosphorylation transcripts in hepatomas. DNA transcription was inhibited in cell cultures by the addition of actinomycin D (10 µg/ml) to the culture medium. Total cellular RNA was isolated at various times (0, 1/2, 1, 2, 4, 6, 8, and 10 h) after the addition of the inhibitor and analyzed by Northern blot hybridization procedures. Upper panels, nuclear-encoded (GAPDH and -F1-ATPase (-F1)) and mitochondrially encoded (ATPase 6-8 (A6-8) and 12 S rRNA (12S)) RNAs were determined with the probes indicated on the left. A representative experiment is shown. The graphs illustrate the calculation of RNA half-lives (t1/2) of the mitochondrially encoded (mt-encoded) and nuclear-encoded (n-encoded) RNAs, assuming first-order rate kinetics for the experiment shown above. The ln of steady-state RNA levels (arbitrary units, a.u./µg of total RNA) is represented at various times after the addition of actinomycin D. Open symbols and dashed lines illustrate RNA levels in AS30D cells. Closed symbols and solid lines illustrate RNA levels in FAO cells.  and , 12 S rRNA;  and , ATPase 6-8 mRNA;  and , GAPDH mRNA; and  and , -F1-ATPase mRNA.

Hepatoma Extracts Promote a Specific Repression of -F1-ATPase mRNA Translation-- The lower cellular representation of the -F1-ATPase protein (Fig. 1), despite an increased abundance of its mRNA in hepatoma cells (Fig. 3), suggested the operation of a specific inhibitory mechanism controlling the translation of -F1-ATPase mRNA (19) in cancer cells. Therefore, we next explored whether the translational efficiency of an in vitro generated -F1-ATPase mRNA could be modified by the presence of exogenously added extracts of hepatoma cells in in vitro translation assays (19). The addition of AS30D and FAO extracts to the reticulocyte lysate promoted a significant reduction in the amount of the synthesized -precursor when compared with assays in which no extract was added (Fig. 6) or those in which normal adult liver extracts were added (Fig. 6). The slight reduction in translational efficiency of -F1-ATPase mRNA observed with adult liver extracts is in line with previous findings (19). Interestingly, the translation-repression activity of hepatoma extracts was essentially the same as that promoted by fetal liver extracts (Fig. 6 and see Ref. 19). The specificity of the inhibitory effect of the fetal and hepatoma extracts on -F1-ATPase mRNA translation is illustrated by the lack of significant inhibition of the translation of p52 (Fig. 6), derived from a polycistronic RNA used as control (19). Since it was observed that the decay of -F1-ATPase mRNA in translation assays was the same, irrespective of the extract tested (data not shown), the results indicated that the expression of -F1-ATPase mRNA in hepatoma cells is compromised by the existence of a specific mechanism that controls its translation.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Effect of hepatoma extracts on the translational efficiency of -F1-ATPase mRNA. Upper panels, 500 ng of in vitro synthesized -F1-ATPase mRNA or control B RNA (Amersham Pharmacia Biotech) was added to nuclease-treated rabbit reticulocyte lysates. Where indicated, 25 µg of protein from fetal (F.L.) or adult (A.L.) rat liver or AS30D or FAO extracts was added to determine the effect of the various extracts on the translational efficiency of -F1-ATPase mRNA. Translation was carried out for 30 min at 30 °C. The [35S]methionine-labeled products were fractionated on SDS-12% PAGE. Representative fluorograms are shown. The migration of the synthesized  subunit precursor (p) and p52, derived from control B RNA, are indicated. Molecular mass markers (200, 97, 69, 46, and 30 kDa) are indicated on the left of each fluorogram. There is no differential effect of the extracts on the synthesis of p52. Lower panel, the histogram illustrates the quantification of the synthesis of the  subunit precursor (p) relative to the synthesis of  subunit precursor obtained in parallel assays without extracts (None). Black bar, adult liver; bar with vertical lines, fetal liver; white bar, AS30D; bar with diagonal lines, FAO. The results shown are means ± S.E. of 4-12 different extracts. *, p < 0.01; **, p < 0.0025; ***, p < 0.0005 when compared with adult liver extracts by Student's t test.

The 3'-UTR of -F1-ATPase mRNA Is a cis-Acting Element Involved in the Control of Its Translation-- The 3'-UTR of -F1-ATPase mRNA is an essential cis-acting element that controls the translation of the transcript (19). To gain further insight into the translational regulatory mechanism of -F1-ATPase mRNA in cancer cells, we next tested the effect of hepatoma extracts on the translational efficiency of a reporter chimeric construct containing the 3'-UTR of -F1-ATPase mRNA (Fig. 7). Remarkably, when translation of the ARF-3'UTR construct was carried out in the presence of hepatoma extracts, a recapitulation of the inhibitory effect of fetal liver or adult kidney extracts (19) was observed on the synthesis of the ARF reporter (Fig. 7). As illustrated previously with the full-length -F1-ATPase mRNA (Fig. 6), the inhibitory effect was not observed when translation of the chimera was carried out in the presence of adult liver extracts (Fig. 7). These findings indicate that the 3'-UTR of -F1-ATPase mRNA is also a target cis-acting element involved in translational regulation of the expression of -F1-ATPase mRNA in cancer cells.

View larger version (42K): [in this window] [in a new window]   Fig. 7.   Effect of hepatoma extracts on the translational efficiency of a chimeric RNA containing the 3'-UTR of -F1-ATPase mRNA. Upper panel, a representative fluorogram illustrates the translation product derived from the chimeric RNA (ARF-3'UTR mRNA) in the presence or absence of different extracts. Zero (- - -) or 500 ng of ARF-3'UTR were translated in nuclease-treated rabbit reticulocyte lysates. Where indicated, 25 µg of protein from fetal (F.L.) or adult liver (A.L.), adult rat kidney (A.K.) or FAO or AS30D hepatoma extracts were added. The migration of the synthesized ARF is indicated by an arrowhead. Molecular mass markers (21 and 14 kDa, from top to bottom) are shown on the left. Lower panel, the histogram illustrates the quantification of the synthesized ARF in the presence of each extract relative to the synthesis of ARF in parallel assays without extract (None). Black bar, adult liver; bar with vertical lines, fetal liver; white bar, AS30D; bar with diagonal lines, FAO; and cross-hatched bar, adult kidney. The results shown are the means ± S.E. of 3-8 different extract preparations. *, p < 0.025; **, p < 0.01; ***, p < 0.005 when compared with adult liver by Student's t test.

Hepatomas Showed an Increased Activity of 3'FBPs-- Developmental and tissue-specific regulated translation of -F1-ATPase mRNA has been shown to depend on the existence of specific proteins (3'FBPs) with binding activity for the 3'-UTR of -F1-ATPase mRNA (19). Thus, it was of interest to analyze whether extracts from rat hepatomas contained an increased abundance of 3'FBPs and, in that case, whether they represent the same cellular proteins and bind the same sequence element as those recently described in the fetal rat liver (19).

UV cross-linking assays of the 3'-UTR of -F1-ATPase mRNA with extracts derived from normal fetal and adult rat liver and kidney, as well as from hepatomas, illustrated the existence of a set of 3'FBPs (p129, p89, p61, p59, p54, and p51) that provided specific RNA-protein complexes (Fig. 8A). In line with previous findings (19), the binding of hepatoma proteins to the 3'-UTR of -F1-ATPase mRNA is specific because of the following: (i) it is competed by an excess amount of the unlabeled 3'-UTR riboprobe (data not shown), (ii) the binding activity is not affected by an excess amount of an unrelated riboprobe (data not shown), and (iii) there is no significant binding of the extracts to a 3'-UTR -F1-ATPase-deleted riboprobe (3'-2 in Fig. 8A). The 3'-2 riboprobe lacked the binding site of the previously described 3'FBPs of normal rat tissues (19), thus indicating that formation of the RNA-protein complexes with hepatoma extracts also requires the sequence element located at the most 3' of the 3'-UTR.

View larger version (61K): [in this window] [in a new window]   Fig. 8.   Hepatoma extracts have an increased activity of 3'FBPs. A, UV cross-linking assays. The sense 3'-UTR of -F1-ATPase mRNA riboprobe (3'UTR), as well as a riboprobe with a deletion of the binding site of 3'FBPs (3'-2) were synthesized. 32P-Labeled probes were incubated with extracts (35 µg) derived from fetal liver (F.L.), adult liver (A.L.), AS30D, FAO, and adult kidney (A.K.), subjected to UV cross-linking, and digested with RNase T1. The RNA-protein complexes were resolved on SDS-12% PAGE. The migration of molecular mass markers (200, 97, 69, and 46 kDa) is indicated on the left. Formation of specific RNA-protein complexes was observed only with the full-length riboprobe of the 3'-UTR of -F1-ATPase mRNA. A representative autoradiogram is shown. B, EMSA. The 32P-labeled full-length 3'-UTR riboprobe was incubated as described above and analyzed by gel retardation in non-denaturing polyacrylamide gels. A representative experiment illustrates the migration of two major specific RNA-protein complexes. The open arrowhead indicates the migration of the RNA-protein complex previously identified in fetal rat liver and adult kidney samples. The closed arrowhead identifies the RNA-protein complex formed with AS30D extracts. C, mapping the binding site on the 3'-UTR of -F1-ATPase mRNA for the formation of the RNA-protein complex observed in EMSAs with AS30D extracts. Different non-labeled (competitor) and radiolabeled (probes) riboprobes of the 3'-UTR of -F1-ATPase mRNA (3'UTR, full-length 3'-UTR; 3'-2, deletion of the last 26 nucleotides of the 3'-UTR; and 3'-1, deletion of the last 92 nucleotides of the 3'-UTR) were generated and incubated with 35 µg of protein of AS30D extracts. The amount of the competitor riboprobes used were 60 and 400 ng, for the 1st and 2nd lane over each competitor, respectively. The black arrowhead shows the migration of the RNA-protein complex of AS30D extracts. D, identification on SDS gels of the proteins forming RNA-protein complexes in EMSA. After incubation of the labeled 3'UTR riboprobe with the different extracts, samples were UV cross-linked before the EMSA. After electrophoresis, the bands of interest were cut from the dried gels and fractionated on SDS-12% PAGE. The migration of molecular mass markers of 200, 97, 69, and 46 kDa are indicated on the left.

The RNA-protein complexes involved in repressing -F1-ATPase mRNA translation in fetal rat liver and adult kidney extracts are those found in the 50-60-kDa range (19). Interestingly, the FAO hepatoma contained an increased cellular representation of the p51, p54, p59, and p61 proteins, when compared with that in normal adult liver, that only showed a low content and/or activity of the p54 and p59 (Fig. 8A). The AS30D hepatoma showed an increased representation of p61 and p59 and negligible or no levels of p54 and p51 (Fig. 8A). In addition, extracts from adult and fetal liver, as well as from the hepatomas, showed RNA-protein complexes (p129 and p89) absent in adult kidney (Fig. 8A). These RNA-protein complexes have not been described previously, and their binding activity has not been correlated with regulation of -F1-ATPase mRNA translation (19). According to the tissue-specific subcellular presentation of -F1-ATPase mRNA (31),² it is suggested that these proteins might represent a set of tissue-specific proteins involved in the cytoplasmic presentation of the mRNA in rat liver.

It has been documented that the set of translation inhibitory 3'FBPs from the fetal rat liver and the adult kidney, when bound to its target sequence on the 3'UTR, provided a characteristic RNA-protein complex in non-denaturing band shift assays (19). Under this condition, the FAO hepatoma revealed a similar RNA-protein complex as that present in fetal liver and adult kidney extracts (see the open arrowhead in Fig. 8B). In agreement with the lack of significant repression of -F1-ATPase mRNA translation by the adult liver extracts (Fig. 6 and Ref. 19), there was no significant formation of this RNA-protein complex when using adult extracts (Fig. 8B). Interestingly, extracts derived from the AS30D hepatoma showed a main RNA-protein complex that migrated slightly ahead of that found in fetal liver, adult kidney, and the FAO hepatoma (see closed arrowhead in Fig. 8B). The formation of the RNA-protein complex that is observed in the AS30D hepatoma (Fig. 8B) also required a full-length 3'UTR (Fig. 8C). In fact, there was no formation of the RNA-protein complex when 3'UTR-deleted riboprobes were used in the assay (Fig. 8C), and only the full-length 3'UTR was able to significantly compete the formation of such complex (Fig. 8C). These findings suggested, in agreement with the findings obtained by UV cross-linking (Fig. 8A), that the 3'FBPs of AS30D extracts bind the same sequence element on the 3'-UTR of -F1-ATPase mRNA as that bound by the 3'FBPs of fetal liver and FAO extracts.

The characteristic RNA-protein complex formed in EMSAs with fetal liver, adult kidney, and FAO hepatoma extracts (Fig. 8B) is built by p51 and p54, as revealed after fractionation of each UV cross-linked complex on denaturing gels (Fig. 8D). In contrast, p59 is the main protein of the RNA-protein complex formed with AS30D extracts (Fig. 8D). Altogether, these results suggested the existence of a conserved mechanism for controlling the translation of -F1-ATPase mRNA between the fetal liver and cancer cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We recently hypothesized that the paradoxical reduction in the complement of mitochondria in cancer and transformed cells might result from a reversion to a fetal program of expression of oxidative phosphorylation genes in these cells (5). In the present investigation we demonstrate that two hepatomas derived from rat liver have a severe reduction in their mitochondrial complement when compared with normal adult rat liver. The down-regulation of the mitochondrial cellular complement of the hepatomas occurs in a paradoxical situation of an increased abundance of the nuclear transcript encoding the mitochondrial -F1-ATPase protein. A similar finding has been observed in human hepatocarcinomas when compared with normal human liver.³ These observations are similar to that previously described in the fetal rat liver (15, 16) and in different cell lines in culture that have been transformed with viral and cellular oncogenes (9, 12). These findings thus suggest the existence of common determinants between development and carcinogenesis for the control of the biogenesis of mitochondria.

The findings indicate that the increased abundance of -F1-ATPase mRNA in hepatomas is not the result of an increase in the transcription rate of the gene but is rather mediated by an increase in the mRNA half-life. A similar observation explains the augmented cellular representation of the transcript in the fetal liver (16). In hepatomas, a specific translation repression mechanism operates for the expression of -F1-ATPase mRNA that has the 3'-UTR of the transcript as cis-acting element of regulation. These findings also agree with the mechanism controlling the expression of the -F1-ATPase gene during fetal liver development (19). Finally, hepatoma extracts have an increased activity of 3'FBPs. This set of 3'FBPs have quite the same molecular mass and binding requirements on the 3'-UTR of -F1-ATPase mRNA as the set that inhibits the translation of -F1-ATPase mRNA in the fetal liver (19). Altogether, the results presented strongly argue that the expression of the -F1-ATPase gene in hepatomas, and perhaps in other cancer cells, results from a reversion in cancer to the program of expression of the gene that operates in the liver during intrauterine development (5). This finding suggests that the observed reversion could also be applied to the expression of other genes involved in oxidative phosphorylation and, therefore, in the biogenesis of mitochondria.

The nuclear-encoded -F1-ATPase mRNA has been recently shown to be localized in defined cytoplasmic structures of the hepatocyte that appear to be responsible for controlling the cytoplasmic metabolism of the transcript (31, 36), as well as the delivery of the encoded precursor protein to mitochondria (36, 37). Most described cis-acting sequences involved in the localization and regulation of mRNA turnover are placed within the 3'-UTR of the mRNAs (38-41). The 3'-UTR of -F1-ATPase mRNA is a short 150-base pair AU-rich sequence that also contains the determinants for mRNA translation (19). Due to the short 3'-UTR of -F1-ATPase mRNA, we suggest the existence of a certain degree of overlap in the function of cis-acting sequences involved in controlling its metabolism (localization, stability, and translation). In fact, the findings in hepatomas provide an additional example, to those previously documented during development (15, 16, 19), in which changes in the stability and in the cellular representation of the transcript are coupled to changes in its translation rate. In this regard, we suggest that proteins binding the 3'-UTR of -F1-ATPase mRNA participate in defining both activities of the cytoplasmic fate of the transcript. In other words, the translation masking of the mRNA affects its stability, and concomitantly its cellular representation, because both mechanisms may illustrate the same cellular process that has, as effector molecules, the regulated proteins that bind the 3'-UTR of the transcript (19).

Remarkably, the findings in hepatomas also provide an additional example in which the stability of oxidative phosphorylation transcripts encoded in both the nuclear and mitochondrial genomes varied in parallel (16, 17). These findings thus reinforce the idea that there should be a common mechanism for controlling mRNA decay rates in both genetic compartments of the cell, a mechanism that may be different from that described in HepG2 cells (42). On the other hand, the observation that the activation of cytoplasmic translation controls the expression of mitochondrially encoded mRNAs (18) during organelle differentiation (5, 14) suggests that cytoplasmic translation also controls the expression of mitochondrially encoded mRNAs in cancer cells.

The operation of mechanisms of stabilization of oxidative phosphorylation mRNAs and subsequent translation masking for the transcripts generated in both genetic units could most likely be responsible for the progressive decline in the mitochondrial complement of cancer cells as a result of cellular proliferation. The reduction of mtDNA in cancer cells could result directly from the dilution of mitochondrial constituents during cellular proliferation. However, it is also possible that essential components of the replication machinery of mtDNA are also subjected to translational regulation. In this regard, it is interesting to note that both DNA polymerase  (43, 44) and mitochondrial transcription factor A (45) seem to be subjected also to post-transcriptional regulation.

In agreement with our previous proposal for the control of -F1-ATPase mRNA translation in the fetal rat liver (19), we suggest that the increased activity of 3'FBPs in hepatomas could be responsible for the inefficient translation of -F1-ATPase mRNA in cancer. Binding of the repressor proteins to the most 3'-sequence of the 3'-UTR may sterically hamper the 5'-3' communication required for an efficient reinitiating of -F1-ATPase mRNA translation (19). Remarkably, the proteins that inhibit -F1-ATPase mRNA translation in the fetal liver may be the same as those present in FAO hepatomas (Fig. 8, B and D). In fact, there is compelling evidence in the literature that illustrates a switch from the adult-type to fetal-type isozyme patterns in hepatic tumors (46). Indeed, it is possible that the 3'FBP observed in AS30D hepatomas (Fig. 8, B and D) could result from covalent modification of one of those present in fetal liver and the FAO hepatoma. The phosphorylation of 3'-UTR binding proteins has been described to control the rate of translation of certain mRNAs (47-49).

The fetal liver and hepatomas share phenotypic characteristics in addition to those herein described related to the biogenesis of mitochondria (3-5, 46). It is reasonable to assume that these analogies arise from common genetic determinants between developing liver and carcinogenesis. Hypoxia is a prevailing and required condition during mammalian development (5, 50, 51) and also a prominent feature of malignant tumors (52). Hypoxia promotes an increased transcriptional expression of genes involved in the glycolytic pathway (26, 51, 53-55), a phenotypic characteristic of both the fetal liver and cancer cells (3-5, 46). Hypoxia also promotes an up-regulation in the cellular representation of certain mRNAs by controlling the stability of the transcripts (55-57). In several of these cases, the control of mRNA stability is exerted by hypoxia-regulated RNA-binding proteins (57-61). It is possible that the control of the stability and subsequent up-regulation of -F1-ATPase mRNA levels observed in the fetal rat liver (16) and in the hepatomas (this study) could share the same mechanism and signaling molecule.

Finally, recent findings indicated the necessity of a functional H⁺-ATP synthase (62) to execute apoptosis. Apoptosis is a genetically encoded program of cell death that is indispensable for normal development of the organism (63) and that can also contribute to lessen the progression of certain disorders (64, 65). Mitochondria play a central role as executioners of apoptotic cell death (66, 67). Remarkably, a frequent characteristic of cancer cells is their resistance to apoptotic stimulus (65, 68). In line with these findings, it appears that cancer cells have developed alternative mechanisms to control the cellular content of this essential mitochondrial protein. In fact, it has been described that the -F1-ATPase protein has an increased turnover in the Zadjela hepatoma (6). Whereas this finding illustrates a mechanism of post-translational regulation for controlling the cellular content of the protein, our findings illustrate a mechanism that is regulated at the level of translation. The final result of the operation of any of these mechanisms is the reduction in the content and/or activity of mitochondria in cancer cells. Recent findings indicated that hypoxia provides a physiologically selective pressure for the clonal expansion of cancer cells that have acquired mutations in genes that are involved in the control of apoptosis (69). Due to the pivotal role played by mitochondria as sensors and executioners of apoptosis (66, 67), it seems reasonable to suggest that a low mitochondrial content also provides cancer cells with a selective advantage to escape and/or become resistant to apoptosis. In this regard, the elucidation of the mechanism that controls the binding activity of 3'-FBPs could provide an alternative strategy to manipulate essential cell components required in apoptotic cell death.

	ACKNOWLEDGEMENTS

Drs. L. G. Bagetto (Lyon, France) and M. Pastor-Anglada (Barcelona, Spain) are gratefully acknowledged for providing the AS30D and FAO hepatomas, respectively. We thank Drs. P. L. Pedersen (Baltimore) and P. Cantatore (Bari, Italy) for the generous supply of cDNAs and mtDNAs of oxidative phosphorylation used as probes in this study. We are grateful to Dr. J. Satrústegui for critical review of the manuscript. We thank M. Chamorro and D. Jelenic for technical and secretarial assistance, respectively.

	FOOTNOTES

^* This work was supported in part by Grant PB97-0018 from the Dirección General de Enseñanza Superior e Investigación Científica y Técnica, Grant 08.3/003/97 and 08.1/0006/1998 from the Comunidad de Madrid, Spain, and by an institutional grant from Fundación Ramón Areces, Spain.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.

Recipient of a predoctoral fellowship from the Gobierno Vasco, Spain.

^§ To whom correspondence should be addressed: Centro de Biología Molecular "Severo Ochoa", Universidad Autónoma de Madrid, 28049 Madrid, Spain. Tel.: 34-91-397-4866; Fax: 34-91-397-4799; E-mail: jmcuezva@cbm.uam.es.

² J. Ricart and J. M. Cuezva, unpublished observations.

³ M. López de Heredia, C. Ugalde, M. Chamorro, and J. M. Cuezva, manuscript in preparation.

	ABBREVIATIONS

The abbreviations used are: -F1-ATPase, subunit of the mitochondrial H⁺-ATP synthase; 3'-UTR, 3'-untranslated region; 3'FBPs, 3'UTR -F1-ATPase mRNA-binding proteins; ARF, ADP ribosylation factor 1; EMSA, electrophoretic mobility shift assay; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GDH, glutamate dehydrogenase; A6-8, mitochondrial ATPase subunits 6 and 8; mtDNA, mitochondrial DNA; 12 S, mitochondrial 12 S rRNA; nt, nucleotides; PAGE, polyacrylamide gel electrophoresis; PEPCK, phosphoenolpyruvate carboxykinase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES