INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

c-myc is a proto-oncogene playing a dual role in the control of cell proliferation-differentiation and apoptosis. Its overexpression contributes to oncogenic transformation of various cell tissues, leading to multiple neoplasms, in particular to hematopoietic tumors in humans and mice (5, 25). Several mechanisms have been shown to generate c-myc activation, including gene amplification, chromosomal translocation, proviral insertion, and retroviral transduction (for reviews see references 25 and 31). However, the correlation between tumor occurrence and increased transcription, marked mRNA stabilization, or elevated level of c-Myc proteins is far from clear. Although it is classically assumed that c-Myc protein abundance is determined by transcriptional control, it has been shown that the c-myc gene is strongly subjected to posttranscriptional regulation (reviewed in reference 12). Numerous studies have demonstrated the role of c-myc mRNA stability in the control of gene expression (for review, see reference 30), and instability determinants have been identified in their 3' untranslated region (3' UTR [3, 20, 33]) as well as in coding exons (23, 43, 44). Translational control also plays an important role: the c-myc 5' UTR has been proposed to be a modulator of translational efficiency (28, 32). It has been recently shown to contain an internal ribosome entry site (IRES) capable of directing the internal initiation of protein synthesis in a cap-independent manner (27, 36). In the human c-myc gene, the IRES lies in the 408 nucleotides (nt) upstream from the AUG start codon of the major c-Myc protein (c-Myc2). It is thus present in the c-myc transcripts initiated from either the P0, P1, or P2 promoter and is able to promote the cap-independent synthesis of the two c-Myc proteins c-Myc1 and c-Myc2, initiated at the two alternative CUG and AUG codons, respectively (15, 27). Various pathologies, including breast cancer (11) and multiple myeloma (41), have been correlated to c-myc aberrant translational upregulation, which is proposed to result from overexpression or activation of the cap-binding protein eIF-4^E. Overexpression of this protein relieves the translational repression imposed by highly structured 5' UTRs (22). However, c-Myc overexpression, at least in multiple myeloma, is clearly eIF-4^E independent and would thus most probably result from aberrant activation of the c-myc IRES.

To date, several IRESs have been discovered in cellular mRNAs, most of them encoding proteins that are involved in the control of cell growth and/or apoptosis (1, 16-18, 27, 34, 36, 39). It has been suggested recently that the presence of the IRES might allow translation of these cellular mRNAs in situations where the cap-dependent translation machinery is inactive, for instance in response to stress (34, 40) and apoptosis (16, 17, 35) and during cell cycle G₂-to-M transition (6, 29). However, these experiments have mostly been performed with transformed or immortalized cell lines where IRES could be aberrantly upregulated (13, 41), thus precluding a knowledge of their possible regulatory role in physiological conditions.

We have investigated this question by analyzing the activity of c-myc IRES in view of its possible involvement in the well-known role of c-myc in controlling cell proliferation and tumorigenesis. Using the bicistronic vector strategy, we compared activities of c-myc and encephalomyocarditis virus (EMCV) IRESs in various transiently transfected cell lines and transgenic mice during embryogenesis and in adult tissues. Our results revealed that the c-myc IRES was active in all the transiently transfected cells and was subjected to strong developmental control in transgenic mice.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmid construction. The bicistronic vectors of the MyCAT and NuCAT series were constructed from the previously described plasmids pCVC and pHCVC (containing a hairpin) which have two tandem chloramphenicol acetyltransferase (CAT) genes (18). Part of the c-myc cDNA (551 nt corresponding to the 408 nt of the myc P2 5' UTR plus the 143 nt downstream from the AUG codon) was introduced into the intercistronic region between the two CAT genes. The two resulting plasmids were called pBI-MyCAT and pHP-MyCAT (Fig. 1A). Two other plasmids were constructed to determine the 3' border of the IRES. These plasmids contained the myc P2 5' UTR (408 nt), in which the AUG codon of myc was directly fused to a coding part of the nucleolin gene, resulting in a chimeric NuCAT gene (pSVNC83) (9). The 5' UTR-NuCAT fusion was obtained by PCR amplification using oligonucleotides T7 and 5' TTTCCATGGTCGCGGGAGGCTGCTGC 3' on plasmid pSCT-MyCAT-P2 (27). This PCR fragment was digested at the NcoI site and was fused to the CAT open reading frame in pBI-MyCAT and pHP-MyCAT (in place of MyCAT fusion). These CAT-myc-NuCAT-containing plasmids were called pBI-NuCAT and pHP-NuCAT, respectively (Fig. 1A).

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Transfection of COS-7 cells with tandem CAT bicistronic vectors. (A) Schematic representation of the tandem CAT bicistronic vectors. The CMV promoter is represented by an arrow, the tandem CAT genes are represented by black boxes, the c-myc sequence fused to CAT is represented by a grey box, and the nucleolin sequence fused to CAT is represented by a white box. The line between the boxes corresponds to the c-myc 5' UTR (containing the IRES). In the plasmids pHP-MyCAT and pHP-NuCAT, the 5' hairpin is represented between the intron (broken line) and the first-cistron CAT (left black box). The line between the boxes corresponds to the c-myc 5' UTR (containing the IRES and the two initiation codons C [CUG] and A [AUG]). (B) Western immunoblotting of transiently transfected COS-7 cells. COS-7 cells were transiently transfected by the bicistronic plasmids described for panel A, and cell extracts were analyzed by Western immunoblotting with anti-CAT antibody as previously described (27). Lane ct corresponds to mock-transfected cells. Lane 1, pBI-MyCAT; lane 2, pHP-MyCAT; lane 3, pBI-NuCAT; lane 4, pHP-NuCAT. The first cistron (CAT) and the second cistron (MyCAT or NuCAT) are indicated by arrows. MyCAT and NuCAT appear as a doublet which corresponds to the initiation of translation at the CUG and AUG codons of the c-myc mRNA, the location of which is shown in panel A.

Cell transfection. Cell lines were obtained from the American Type Culture Collection (ATCC) or European Collection of Animal Cell Culture (ECACC). NIH 3T3 is a mouse immortalized fibroblast cell line. C2C12 (ECACC no. 91031101) is a mouse muscle myoblast line. ABAE is an adult bovine aortic endothelial primary cell line (7). CHO is a Chinese hamster ovary carcinoma cell line. HeLa (ATCC no. CCL2) is a human uterus carcinoma cell line of epithelial origin. COS-7 (ATCC no. CRL 1654) is a monkey kidney cell line transformed by the simian virus 40 large T antigen. Jurkat (ATCC no. TIB-152) is a human acute leukemia T-cell line. Skin fibroblasts come from a human primary culture cell line (38). ECV304 (ECACC no. 92091712) is a spontaneously transformed human endothelial cell line. 293 (ATCC no. CRL-1573) is a human kidney epithelial cell line transformed with adenovirus 5. SK-Hep-1 (ATCC no. HTB 52) is a human liver adenocarcinoma cell line of endothelial origin; SK-N-AS (ECACC no. 94092302) and SK-N-BE (ECACC no. 95011815) are human neuroblastoma cell lines. Saos2 (ATCC no. HTB-85) is a p53^/ human osteosarcoma cell line of epithelial origin. Cells were cultivated according to ATCC or ECACC instructions.

Generation of transgenic mice. REL and RMyL transgenic mice were obtained by injecting fragments containing the CMV-LucR-EMCV IRES-LucF and CMV-LucR-c-myc-IRES-LucF sequences, respectively, into one of the pronuclei of (C57BL/6 × CBA)² fertilized eggs (4). Transgenic embryos and adult mice were identified by PCR and Southern blot analysis of placental or tail DNA using a LucF 500-bp-long probe amplified with the LucF-specific oligonucleotides 5'-CAGTATGAACATTTCGCAGCC-3' and 5'-CTGAAGGGACTGTAAAAACAGC-3'. Stable lines were maintained by successive crosses with (C57BL/6 × CBA)F₁ mice. The transgene copy number was determined by PhosphorImager scanning of Southern blots. The intensity of the bands corresponding to tandem head-to-tail integration was expressed relative to the intensity of the bands corresponding to 5' and 3' flanking regions.

Reverse transcription (RT)-PCR analysis. The cDNAs were synthesized as previously described, using 5 µg of DNase-treated total RNA (38). The PCR was performed with a primer couple (5'-GATTACCAGGGATTTCAGTCG-3' and 5'-CTGAAGGGACTGTAAAAACAGC-3' or 5'-CCACATATTGAGCCAGTAGC-3' and 5'-CCATGATAATGTTGGACGAC-3') to amplify the LucF or LucR DNA sequence, respectively. The PCRs were carried out using 0.3 U of Goldstar Taq DNA polymerase (Eurogentec) in a final volume of 30 µl, with 1 µl of cDNA. The reaction was performed on a Perkin-Elmer apparatus under the following conditions: 94°C for 3 min and then 25 cycles at 94°C for 30 s, 58°C for 45 s, 72°C for 45 s, and finally 72°C for 5 min. Amplification results (one-third of the reactions) were analyzed on 2% agarose gels (Tris-borate-EDTA), followed by ethidium bromide staining.

Luciferase activity analysis. The two luciferase activities were measured in cell or tissue extracts, and IRES activity was determined by calculating the LucF-to-LucR ratios. The ratio of IRES efficiency between different cell types was compared ex vivo by calibrating the pCRMyL or pCREL (c-myc or EMCV IRES activity, respectively)-to-pCRHL (background activity without an IRES) ratio. For the in vivo analysis (i) 11-day-old (E11) embryos and placenta; (ii) E16 heart, limb bud, tail, brain, liver, and placenta; and (iii) adult tissue fragments were frozen in liquid nitrogen and stored at 80°C. They were homogenized in 200 µl of passive lysis buffer (Promega) using thurax and were centrifuged for 20 min at 4,500 rpm at 4°C (Biofuge Pico centrifuge; Heraeus). The supernatant was centrifuged for 15 min at 13,000 rpm at 4°C, and the last supernatant was used for luminescence dosage (30 µl). LucR and LucF activities were measured using the Dual Luciferase kit from Promega.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

c-myc IRES activity is not influenced by elements located downstream from the AUG start codon. In a previous report, researchers characterized the human c-myc IRES as requiring RNA elements located between nt 408 and 140 upstream from the c-myc mRNA AUG codon (27). However, the different constructs used in that report contained, in addition to the 5' UTR, 143 nt of the translated region downstream from the AUG (Fig. 1A, pBI-MyCAT). To evaluate their possible role in the IRES activity, a bicistronic construct with two tandem CAT genes was designed in which the 143 nt of the c-myc coding sequence were replaced by 280 nt from the nucleolin coding sequence (Fig. 1A, pBI-NuCAT). These bicistronic constructs and their counterparts bearing a 5' hairpin (pHP-MyCAT and pHP-NuCAT) were used for transient transfection of COS-7 cells and expression of the first and second cistrons analyzed by Western immunoblotting with anti-CAT antibodies (Fig. 1B). The results clearly showed that the second cistron was expressed just as efficiently from the constructs pMyCAT and pNuCAT, demonstrating that the coding sequence downstream from the c-myc AUG start codon was not involved in the activity of the c-myc IRES.

c-myc IRES is active in tissue culture cells with varying efficacy. To test the activity of the c-myc IRES in different cell types, we used a bicistronic vector expressing two highly sensitive luciferases, LucR and LucF, under the control of the cytomegalovirus (CMV) promoter. This vector was successfully used in a previous report to demonstrate the existence of the two vascular endothelial growth factor IRESs (18). The human c-myc IRES (the 408-nt-long fragment upstream from the AUG codon) was inserted between the two Luc genes (pCRMyL, Fig. 2A). LucR, the first cistron, provides the level of cap-dependent translation; it is expected to reflect the activity of CMV sequences in the transfected cells and thus to be proportional to the amount of bicistronic mRNA. LucF, the second cistron, is expressed proportionally to the IRES activity. The use of such an assay, based on dual highly sensitive reporters, allows rapid and concomitant quantitation of both reporter gene activities. As positive or negative control, we used similar vectors containing either the EMCV IRES (pCREL) or a hairpin (pCRHL) between the luciferase genes, respectively (18).

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Analysis of FGF-2 IRES activity in transiently transfected cells. (A) Schematic representation of the bicistronic LucR-I-LucF vectors. Plasmid construction is described in Materials and Methods. They contain a CMV promoter controlling the expression of a bicistronic LucR-LucF mRNA. A synthetic intron is present 5' of LucR, and a poly(A) site is present 3' of LucF (18). In the construct pCRMyL, the complete c-myc P2 leader (408 nt) has been fused to the AUG codon of the LucF open reading frame. In the construct pCRHL, a hairpin has been introduced between the two luciferase genes. The construct pCREL contains the EMCV IRES between the two luciferase genes. (B) Fourteen different cell types were transiently transfected with pCRMyL, pCRHL, or pCREL DNA. Cells were harvested 48 h after transfection, and luciferase activities present in cell extracts were measured. c-myc IRES activity was obtained by calculating the ratio 100 × (LucF/LucR). To calibrate the data from the different cell types, the LucF/LucR ratio of pCRMyL was divided by the ratio of the negative control pCRHL. Experiments were repeated three to six times, and results are expressed as means ± standard error (SE). The names of the different cell types (described in Materials and Methods) are indicated. Fib., fibroblast.

Generation of transgenic mice expressing bicistronic LucR-IRES-LucF constructs. To determine the tissue specificity of the c-myc IRES in the absence of any bias due to cell transformation or cell culture conditions, we decided to study the expression of the LucR-Myc-IRES-LucF construct in vivo. For that purpose, transgenic mice were produced that contained the LucR-IRES-LucF DNA fragment (Fig. 2A, CRMyL, and Materials and Methods). Two independent lines (RMyL-22 and RMyL-28) were established. Southern blots showed that these transgenic lines contained 1 or 2 and 28 to 35 copies of the transgene, respectively (Fig. 3A and B, lanes 5 and 6 and 7 and 8). Two other transgenic lines were derived in parallel, which contained the construct with the viral EMCV IRES in place of c-myc IRES (Fig. 2A, CREL). These two lines, RELA and RELB, both contained 12 to 15 copies of the construct (Fig. 3B, lanes 1 and 2 and 3 and 4).

View larger version (47K): [in this window] [in a new window]   FIG. 3.   Southern blotting analysis. (A) Schema of the tandem repeat insertion of the bicistronic constructs in the mouse genome. The complete digestion of genomic DNA with the EcoRI enzyme (unique site in the transgene) generates a 4.3-kb (RMyL) or 4.7-kb (REL) fragment corresponding to the complete tandem repeat (N copies) and two fragments of unknown sizes (both hybridizing with the probe which encompasses the EcoRI site) corresponding to the 5' and 3' flanking regions. The number of tandem copies has been evaluated by calibrating the signal given by the complete tandem repeat to the signals given by the flanking regions. (B) Twenty micrograms of genomic DNA was digested with the EcoRI enzyme and was hybridized to a LucR-specific PCR fragment (see Materials and Methods). Two offspring of each different strain (RELA and RELB; RMyL-22 and RMyL-28) were tested in parallel. (Only 10 µg of sample 5 was digested.) Arrows indicate the complete tandem repeat fragments corresponding to the CREL and CRMyL transgenes. The flanking region fragments are indicated by asterisks. Other bands of higher molecular weight correspond to partially digested DNA. The intensity of the bands was measured with a phosphorimager, and the copy number of the transgene was evaluated as explained for panel A.

c-myc IRES is active in transgenic embryos but silent in adult tissues. Luciferase activities were measured in total E11 embryos and placenta of the 4 transgenic lines, in 5 E16 tissues and placenta, and in 13 adult organs. As shown in Fig. 4 and Table 2, where the values correspond to representative experiments obtained in the two transgenic lines, the activity of c-myc IRES was high in the E11 embryo, whereas it was hardly detectable in the placenta. In contrast the activity of EMCV IRES was high in both the E11 placenta and embryo (Fig. 4A).

View larger version (41K): [in this window] [in a new window]   FIG. 4.   Activities of c-myc and EMCV IRESs in transgenic embryos. Embryos from the different transgenic lines were prepared for E11 (A) and E16 (B) embryos, and luciferase activities were measured in different tissues (named on the left), as described in Materials and Methods. IRES activities were measured by calculating the ratios [100 × (LucF/LucR)] and are represented by histograms (black boxes for RMyL-28 and -22 and white boxes for RELB and -A). For values, see Table 2 and reference 8. Experiments were repeated at least three times, and results are expressed as means ± SE. Abbreviations: nd, not determined (the experiment was not performed); ud, undetected activity (LucR activity was not superior to three times the background value in the corresponding organ).

View larger version (34K): [in this window] [in a new window]   FIG. 5.   c-myc and EMCV IRES activities in transgenic adult tissues. Different organs (named on the left) were prepared from adult transgenic mice, and luciferase activities were measured as described for Fig. 4. IRES activities are represented by histograms corresponding to the equation 100 × (LucF/LucR) (Table 2). These values correspond to representative experiments that were repeated at least three times on the two independent lines expressing each bicistronic construct, pCRMyL (c-myc) and pCREL (EMCV). Abbreviations: nd, not done; ud, undetectable LucR activity.

View larger version (50K): [in this window] [in a new window]   FIG. 6.   Detection of the bicistronic Luc mRNA in transgenic adult tissues. Total RNA was extracted from RMyL-22 or RELA transgenic brain (B) or testis (Te), and the level of LucF and LucR cistrons was analyzed by RT-PCR (as described in Materials and Methods). The negative control (C) corresponds to a PCR experiment performed using RNA from the brain of a nontransgenic mouse. The positive control (C+) was performed with an in vitro-transcribed bicistronic REL mRNA. The bands corresponding to amplification of LucR (465 nt) and LucF (366 nt) cDNAs are indicated by arrows. This experiment enabled us to check the LucF/LucR ratios at the RNA level in the different organs and to show that they are similar in REL and RMyL transgenic mice.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We report here that the IRES of a proto-oncogene, c-myc, is repressed in vivo in adult transgenic mice. In contrast it is active in E11 and E16 embryos (with a strong tissue specificity for E16 embryos) and in all the cultured, proliferating cells tested. Strikingly, the EMCV IRES, albeit with different efficiencies, exhibits similar levels of activity both ex vivo and in vivo in embryos and adult mice. The IRES silencing observed in vivo is thus specific to c-myc IRES and points to a physiological role of this IRES, which, if deregulated, might contribute to tumorigenesis.

We provide here a direct comparison of activities of c-myc and EMCV IRESs in human and nonhuman cell types in a highly calibrated quantitation system. Two previous studies have separately reported cell type-specific variations of these two IRESs (2, 37). First, Borman et al. compared the behavior of seven picornaviral IRESs, either from type I or type II (19), in six primate or nonprimate cell lines (2). The cells could be divided broadly into two groups: cells including HeLa, FRhK4, and SK-HepG2, which were permissive to all the picornaviral IRESs; and cells corresponding to SK-N-BE, BHK21, and Neuro-2A, which were permissive for type II IRESs (such as EMCV) but were refractory to translation from type I IRESs. Accordingly, we have also found that the EMCV IRES is active in both HeLa and SK-N-BE cells. The apparent differences observed between Borman's experiments and ours reflect only the different methods used to measure IRES-dependent translation: whereas in Borman's study, the IRES activities are expressed relative to the cell type in which the IRES is the most efficient (taken as 100%), we calibrate the IRES activity in a certain cell type relative to the background value given by a bicistronic construct devoid of an IRES between the two cistrons.

Second, Stoneley et al., using bicistronic constructs, studied c-myc IRES activity in HeLa cells (37). In agreement with our study, they reported that Hela cells showed a high level of IRES-dependent translation. Furthermore, by comparing CMV- or simian virus 40-initiated bicistronic transcripts, they demonstrated that the level of c-myc IRES activity was dependent upon the amount of bicistronic mRNAs found in the cell: the more they were expressed, the lower the IRES activity. This suggests that noncanonical factors are required to activate IRES-dependent translation and that they might be titrated by an excess of bicistronic mRNAs in the transiently transfected cells.

We have shown by using a large variety of transiently transfected cultured cells that EMCV and c-myc IRESs are both efficient in the five species tested (human, monkey, cow, hamster, and mouse). However, the EMCV IRES is globally more efficient than that of c-myc, with a maximum of 110 AU for EMCV versus 30 AU for the c-myc IRES. This point is well illustrated by the analysis of ABAE, 293, CHO, and 3T3 cell lines, where the EMCV IRES is extremely potent while the c-myc IRES is weak. However, this does not hold true for all cell lines, since in Jurkat lymphoma cells, for instance, both IRESs present similar activities (around 14 AU). These differences, obtained in conditions of mRNA overexpression, reveal the cell types where the trans-acting factors required by the c-myc IRES are the most limiting.

The different activities observed for EMCV and c-myc IRES could rely on particular features due to their viral and cellular origins, respectively. This hypothesis is in agreement with a recent analysis of the activity of another cellular IRES, that of fibroblast growth factor 2 (FGF-2) (8). We found that within the same cell lines, the FGF-2 IRES also presented a broad spectrum of activity, with variations similar to those observed with the c-myc IRES, ranging from 3 to 36 AU. However, the highest efficiency was observed in the p53^/ osteosarcoma Saos2 cell line, in which c-myc IRES was two times less active, and the lowest efficiency was observed in Jurkat cells, where the c-myc IRES was five times more efficient. Taken together, these data point out the tissue specificity of each IRES under study. Strikingly, the c-myc IRES is active in transformed cells and especially in neuroblastoma and leukemia cell lines: as their normal nontransformed counterparts express c-myc mRNA abundantly (26), it is possible that an aberrant activity level of the c-myc IRES in these cells contributes to their progression towards tumorigenesis.

The c-myc IRES shows a much more drastic regulation of its activity in vivo than ex vivo. This contrasts with the EMCV IRES: we show that the viral IRES is active throughout development in the whole E11 embryos as well as in most E16 embryonic and extraembryonic tissues, while c-myc IRES activity is restricted to the E11 embryo proper and to the E16 brain and heart. In adults, the c-myc IRES is silenced, while the EMCV IRES is fully active with a broad spectrum of activity. Whereas the wide activity of EMCV IRES confirms previous data obtained with chicken embryos (14) and transgenic mice (21), the developmental regulation of c-myc IRES activity extends the conclusions reached from analysis of other cellular IRESs, such as the FGF-2 IRES in transgenic mice (8) and Antennapedia and Ultrabithorax IRESs in Drosophila (42). FGF-2 IRES activity was low but significant in most organs of adult transgenic mice and remarkably high in their brains (8). This contrasts with the complete silencing of the c-myc IRES observed in adult tissues and highlights a specific regulatory mechanism for the c-myc IRES. This regulation most probably relies on the presence of spatiotemporally controlled trans-acting factors. Such factors, which are not in limiting amount in vivo when present, since we obtained similar results with two transgenic lines (RMyL-22 and RMyL-28) that express very different levels of pCRMyl transcripts, could be enhancers of IRES activity in embryonic tissues (37, 40). Alternatively, which is not exclusive, they could be inhibitors of c-myc IRES activity in adult tissues.

What could be the physiological relevance of the c-myc IRES regulation observed in vivo? c-myc is widely expressed during mouse embryogenesis, where active cellular proliferation takes place. Its expression is required during development, since c-myc gene invalidation causes lethality during the period between 9.5 and 10.5 days of gestation in homozygotes (10). Strikingly, the pathologic abnormalities include the heart and the neural tube, two developing structures in which we have shown the IRES to be active. Thus the regulation of IRES-dependent translation might play an important role in the control of c-myc expression during development.

c-myc gene expression, active in the embryo, is downregulated in the adult, where most cells are in a nonproliferating state. Aberrant c-myc overexpression in transgenic mice leads to tumor formation (24). Previously, the downregulation of c-myc expression, needed to prevent abnormal cell proliferation and/or differentiation, has been shown to take place mainly at the level of mRNA stability (30). Such a mechanism involves elements located in both the coding and 3' UTR of the c-myc mRNA sequences (3, 20, 23, 33, 43, 44). The results that we provide in this study indicate that the 5' UTR of this mRNA, in addition, plays a role in the regulation of c-myc expression in vivo. The translational silencing mediated by the IRES (due to the absence of activating factors or the presence of inhibiting factors) could ensure the complete repression of c-myc expression in fully differentiated tissues. This translational repression could result from the fact that differentiated tissues have reduced levels of proteins required for cell proliferation or increased levels (or activities) of proteins involved in cell growth arrest; such proteins could be involved in the positive or negative control of IRES-mediated translation. Physiologically, the transient activation of the IRES (resulting from a change of IRES-regulating factor activity) would allow quiescent cells to induce rapidly the synthesis of c-Myc proteins in response to various signals inducing proliferation, differentiation, or apoptosis, independently of an increase in the level of mRNA. The transgenic mice that we have described in this study represent a powerful tool to investigate such a hypothesis.