Amino Acid Limitation Induces Expression of CHOP, a CCAAT/Enhancer Binding Protein-related Gene, at Both Transcriptional and Post-transcriptional Levels*

In mammals, plasma concentrations of amino acids are affected by nutritional or pathological conditions. Here we examined the role of amino acid limitation in regulating the expression of CHOP, a CCAAT/enhancer binding protein (C/EBP)-related gene. CHOP protein is capable of interacting with other C/EBPs to modify their DNA binding activities and may function as a negative regulator of these transcription factors. Our data show that leucine limitation in human cell lines leads to induction of CHOPmRNA and protein in a dose-dependent manner.CHOP mRNA induction is rapidly reversed by leucine replenishment. Elevated mRNA levels result from both an increase in the rate of CHOP transcription and an increase in theCHOP mRNA stability. Using a transient expression assay, we show that a promoter fragment, when linked to a reporter gene, is sufficient to mediate the regulation of CHOPexpression by leucine starvation in HeLa cells. In addition, we found that decreasing amino acid concentration by itself can induceCHOP expression independently of a cellular stress due to protein synthesis inhibition. Moreover, CHOP expression is induced at leucine concentrations in the range of those observed in blood of protein-restricted animals suggesting that amino acids can participate, in concert with hormones, in the regulation of gene expression.

Cells regulate gene expression in response to changes in the external environment. Metabolite control of gene expression has been well documented in prokaryotes and lower eukaryotes. Specific mechanisms have evolved to allow these organisms to quickly metabolize various molecules based on their availability in the external medium (1,2). However, much less is known about the response of multicellular organisms to nutrient variations. The control of gene expression differs in many aspects from those operating in single cell organisms and involves complex interactions of hor-monal, neuronal, and nutritional factors. It has been shown that major (carbohydrates, fatty acids, sterols) or minor (minerals, vitamins) dietary constituents participate, in concert with many hormones, in the regulation of gene expression in response to nutritional changes (3)(4)(5)(6)(7). There is considerably less information available concerning the control of mammalian gene expression by amino acids. However, it has been shown that starvation of one essential amino acid causes a specific increase in mRNA abundance of certain genes including c-myc, c-jun, ornithine decarboxylase (8), asparagine synthetase (9), the mammalian equivalent of ribosomal protein L-17 (10), the insulin-like growth factor binding protein gene (11). Moreover, Marten et al. (12) have shown that the abundance of several different mRNAs is affected by amino acid starvation. In this study the greatest induction in response to amino acids starvation was exhibited by the CHOP gene. However, little is known about the molecular mechanisms involved in gene regulation by amino acids. It has only been shown that the induction of asparagine synthetase gene by amino acid starvation involves both transcriptional and post-transcriptional mechanisms (9). These authors have characterized cisacting elements involved in transcriptional regulation of that gene in response to amino acid starvation.
CHOP (also called gadd153) is a mammalian gene whose expression is also induced in all tested cells by a wide variety of stresses and agents (13)(14)(15)(16). CHOP encodes a small nuclear protein related to the CCAAT/enhancer-binding protein (C/ EBP) 1 family of transcription factors. Members of the C/EBP family have been implicated in the regulation of processes relevant to energy metabolism (17), cellular proliferation, differentiation, and expression of cell type-specific genes (18 -20). By forming heterodimers with the members of the C/EBP family, CHOP protein can influence gene expression as both a dominant negative regulator of C/EBP binding to one class of DNA targets and by directing CHOP-C/EBP heterodimers to other sequences (21)(22)(23)(24)(25)(26).
In mammals, plasma concentrations of glucose and free amino acids are markedly affected by nutritional or pathological conditions (27,28). Carlson et al. (15) have shown that CHOP mRNA expression is induced by glucose deprivation in mammalian cell lines, suggesting a close relationship between nutrient variation and CHOP expression. In the present study we have examined the role of amino acids in the regulation of CHOP expression. We demonstrate that amino acid limitation, in conditions which do not inhibit protein synthesis, can induce * This work was supported in part by grants from the Institut National de la Recherche Agronomique and the Fondation pour la Recherche Médicale. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  CHOP expression. Particularly, we show that leucine starvation induces CHOP expression through both transcriptional and post-transcriptional mechanisms. The implication of these findings are discussed in a general context of the control of mammalian gene expression by amino acids in various nutritional conditions.

MATERIALS AND METHODS
Cell Culture and Treatment Conditions-Cells were cultured at 37°C in Dulbecco's modified Eagle's medium F12 (DMEM/F12) (Sigma) containing 10% (HeLa and HepG2) or 20% (Caco-2) fetal bovine serum. When indicated, DMEM/F12 lacking leucine was used. For other amino acid or glucose starvation experiments, MEM medium (Life Technologies, Inc.) was used. For amino acid starvation experiments 10% dialyzed calf serum was used.
RNA Isolation and Northern Blot Analysis-Total RNA was prepared as described previously (29). Northern blots were performed according to the procedure of Sambrook et al. (30). The membranes were UV cross-linked and then prehybridization was carried out for 2 h at 55°C in 50% formamide, 6 ϫ SSC, 5 ϫ Denhardt's reagent, 0.5% SDS, 0.1 mg/ml sonicated salmon sperm DNA, and 10 g/ml yeast tRNA. The human CHOP cDNA (BH1), generously provided by Dr. N. J. Holbrook (31), was used as a probe. BH1 plasmid was linearized by PstI, and 32 P-riboprobes were synthesized (30) using T7 RNA polymerase (Promega). Hybridization was carried out for 16 h at 55°C. The membranes were washed for 15 min at 55°C successively in 2 ϫ SSC containing 0.1% SDS, 0.5 ϫ SSC containing 0.1% SDS, 0.1 ϫ SSC containing 0.1% SDS. Labeled bands were detected by autoradiography. Autoradiogram signals were quantified by using a densitometric scanner (Appligene) and NIH image software. To control for variation in either the amount of RNA in different samples or loading errors, all blots were rehybridized with an oligonucleotide probe corresponding to 18 S RNA. All densitometric values for CHOP mRNA were normalized to 18 S RNA values obtained on the same blot. Relative CHOP mRNA was determined as the ratio of CHOP mRNA and 18 S RNA.
DNA Transfection and CAT Assay-HeLa cells (5 ϫ 10 5 ) were plated in 60-mm diameter dishes and transfected by the calcium phosphate coprecipitation method as described previously (32). Ten micrograms of CAT plasmid were transfected into the cells along with 2 g of pCMV-␤Gal, a plasmid carrying the bacterial ␤-galactosidase gene fused to the human cytomegalovirus immediate-early enhancer/promoter region, as an internal control. Cells were exposed to the precipitate for 16 h, washed twice in phosphate-buffered saline, and then incubated with DMEM/F12 containing 10% fetal calf serum. Twenty-four hours after transfection, cells were amino acid-starved for the desired time and then collected for CAT assay (33). The protein concentration of the cell extracts was determined using the BCA method (34). ␤-Galactosidase activity was measured as described by Hall et al. (35) and used to calibrate transfection efficiency. Relative CAT activity was given as a percentage of pSV2CAT activity. All values are the means calculated from the results of at least three independent experiments.
Primer Extension-Total cellular RNA from transfected cells was isolated as described above. A 20-base pair oligonucleotide (5Ј-CAACG-GTGGTATATCCAGTG-3Ј), complementary to the DNA sequence located 11-30 base pairs downstream from the translation initiation site of the cat gene, was end-labeled with T 4 polynucleotide kinase (Eurogentec) and then used for primer extension as described previously (36).
Nuclear Run-on Transcription Assays-In vitro transcription experiments in isolated HeLa cell nuclei were carried out essentially as described by Liu et al. (37). RNA was labeled with [ 32 P]UTP and then hybridized to filter-bound cDNAs of CHOP (31), ribosomal S26 protein (38), and pBluescript DNA (Stratagene). Hybridization with labeled RNA was performed at 45°C for 24 h. The filters were washed twice for 15 min in 5 ϫ SSC plus 0.2% SDS at 45°C, followed by three washes in 2 ϫ SSC plus 0.2% SDS at 45°C. Radioactive dots were visualized and quantified by using a PhosphorImager (Bio-Rad) and the MOLECULAR ANALYST software.
Protein Synthesis Measurements-HeLa cells were incubated for 16 h in DMEM/F12 containing 420, 140, 70, 35, or 0 M leucine. During the last 3 h of incubation, 0.5 Ci/ml [ 35 S]methionine were added. The medium was then removed, and the cells were incubated for 30 min in cold 5% trichloroacetic acid. The wells were washed once with trichloroacetic acid and three times with water. The radioactivity incorporation into trichloroacetic acid-precipitable material was measured by liquid scintillation counting after protein solubilization in 0.1 M NaOH plus 0.5% SDS. Results are given as a percentage of methionine incorporation in cells incubated in DMEM/F12 control medium.
Western Blot Analysis-Cells were lysed in a SDS-containing buffer (0.1 M Tris-HCl, pH 6.8, 1% SDS) and immediately boiled for 5 min. Proteins were resolved by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membrane. CHOP and the ubiquitous nuclear protein TLS were detected in D. Ron's laboratory as previously described (39,40).

Induction of CHOP mRNA Expression by Leucine Limitation-
To understand the regulation of gene expression by amino acids at a molecular level, we have studied the regulation of CHOP expression in response to leucine limitation because (i) leucine is an essential amino acid that is poorly utilized by cells during a 16-h incubation period (data not shown), (ii) leucine, which is transported by system L, is rapidly equilibrated through the cell membrane (41,42), and (iii) Marten et al. (12) have shown that leucine depletion strongly induces CHOP expression. To test the possibility that leucine concentration can influence CHOP expression, HeLa, HepG2, or Caco-2 cells were incubated for 16 h in medium containing different concentrations of leucine. As shown in Fig. 1A, CHOP mRNA levels were very low in each cell type in control medium containing 420 M leucine and were inversely proportional to the leucine concentration in the medium, ranging from 15-to 30-fold over the control value. These results indicate that the expression of CHOP mRNA in human cells is regulated in response to changes in leucine concentration. Fig. 1B shows   FIG. 1. Effect of leucine limitation on CHOP mRNA and protein expression. A, subconfluent HeLa, HepG2, and Caco-2 cells were incubated for 16 h in DMEM/F12 containing the indicated leucine concentrations. 420 M leucine correspond to DMEM/F12 control medium. Total RNA was extracted, and Northern blots were prepared as described under "Materials and Methods." The blots were hybridized with a labeled probe corresponding to CHOP. The CHOP mRNA migrates as a single 0.9-kilobase pair transcript. The same membranes were rehybridized with an 18 S probe to normalize for RNA loading. The quantification of these data is shown below the signal for CHOP and 18 S on RNA blots. B, subconfluent HeLa cells were incubated for 16 h in DMEM/F12 containing the indicated leucine concentrations. Whole cell lysates were prepared and probed for the presence of CHOP by Western blot analysis as described under "Materials and Methods." The blot was then probed with an anti-TLS antibody as an internal control.
that the increase in CHOP mRNA levels results in the increase in the CHOP protein. Kinetic analysis of CHOP mRNA level in HeLa cells exposed to medium lacking leucine indicated that mRNA was detectable 2 h after starvation, and a maximum level was reached after 10 -12 h ( Fig. 2A). To determine whether the induction of CHOP expression by leucine starvation is reversible by leucine replenishment, 420 M leucine was added to the culture medium of HeLa cells incubated for 16 h in leucine-free medium. Fig. 2B clearly shows that leucine addition resulted in a rapid loss of CHOP mRNA expression with levels declining over 90% by 1 h following the addition of leucine.
Inhibition of Protein Synthesis Is Not Responsible for Induction of CHOP mRNA Expression-To determine whether leucine limitation affects protein synthesis, HeLa cells were incubated in medium containing different concentrations of leucine and then [ 35 S]methionine incorporation in the acidprecipitable fraction was measured (Fig. 3A). Cells incubated in medium lacking leucine showed a 40% reduction of methionine incorporation into total protein together with a drastic increase in CHOP mRNA level (Fig. 3B, lane b). However, cells incubated in medium containing 35 or 70 M leucine gave no significant reduction of the global protein synthesis, whereas CHOP mRNA expression was significantly increased (Fig. 3B,  lanes c and d). These observations are consistent with the idea that inhibition of protein synthesis is not responsible for the induction of CHOP mRNA expression.
The Induction of CHOP Expression by Leucine Starvation Involves Both Transcriptional and Post-transcriptional Mechanisms-Leucine starvation could increase CHOP mRNA expression either by increasing the rate of transcription or by stabilizing existing transcripts, or through both mechanisms. Nuclear run-on experiments provided evidence that the rate of CHOP transcription was increased by leucine starvation (Fig.   4A). Four hours of leucine starvation increased dramatically the transcription of CHOP (21-fold), while the transcription of the S26 ribosomal gene remained unchanged. To determine whether leucine starvation can affect the half-life of CHOP mRNA, HeLa cells were first incubated for 16 h in medium lacking leucine and then incubated with actinomycin D (4 g/ml) in the presence or absence of 420 M leucine, and total mRNA was extracted from cells at various times. As shown in Fig. 4B, addition of leucine resulted in a rapid decline in CHOP mRNA levels. In starved cells, the CHOP mRNA half-life was increased about 3-fold compared with cells incubated in the control medium. These findings indicate that leucine starvation elevates CHOP mRNA levels both by increasing the rate of CHOP transcription and by enhancing the stability of CHOP mRNA. To assess the importance of protein synthesis for the increase of CHOP mRNA expression during leucine starvation, cells were leucine-starved and treated with cycloheximide for 4 h. As shown in Fig. 4C, cycloheximide present during leucine starvation prevented the accumulation of CHOP mRNA. This result indicates that the increase in CHOP mRNA during leucine starvation is dependent on de novo protein synthesis.
Regulation of CHOP Promoter Activity by Leucine Starvation-To analyze the role of CHOP promoter in transcription activation by leucine starvation, a chimeric gene (pCHOP-CAT) containing the 5Ј-flanking sequence from nucleotides Ϫ954 to ϩ91 fused to the cat gene (31) was transiently transfected in HeLa cells. The data presented in Fig. 5A (summarized in the graph of Fig. 5B) show that CAT activity expressed under the control of the CHOP promoter was induced 7-fold by 16 h of leucine starvation, whereas CAT activity expressed from the pSV2CAT construct used as a control was not induced. These results gave direct evidence that regulation of CHOP transcription by leucine starvation is mediated through the promoter sequence situated between nucleotide position Ϫ954 and ϩ91. Similar increased levels of CAT activity were also observed with transfection of pCHOP-CAT into HepG2 and Caco-2 cells (data not shown). To correlate CAT activity and amounts of CAT mRNA transcribed under leucine-starved and nonstarved conditions, primer extension experiments were performed. As shown in Fig. 6, under leucine starvation, the amounts of CAT mRNA initiating at the correct start site of the promoter were much higher (lane b) than those transcribed in normal conditions (lane a), and the levels of CAT mRNA derived from pSV2CAT remained unchanged (lanes c and d).
These results show that the degree of induction of pCHOP-CAT mRNA expression (6 -7-fold) is in agreement with the degree of induction determined in CAT assays and indicate that, under our experimental conditions, leucine starvation does not affect significantly translation of the CAT mRNA.
To determine whether the CHOP promoter-driven CAT induction is consistent with that described for the endogenous CHOP mRNA, we examined the characteristics of the CHOP promoter activity in response to leucine limitation. Fig. 7A shows that the transcriptional activity from CHOP promoter was enhanced by a decrease in leucine concentration in a dose-dependent manner. Furthermore, kinetic analysis of the cat gene expression revealed that maximal CAT activity induction was reached 16 h after starvation (Fig. 7B).
Specificity of the CHOP Promoter Response to Amino Acid Starvation-Starvation of other amino acids was tested for their abilities to influence CHOP promoter-driven CAT expression in HeLa cells (Fig. 8). The most potent amino acids increasing CAT activity level appeared to be methionine, lysine, arginine, phenylalanine, and threonine. They produced about the same induction level of CAT activity as that obtained with leucine (5-8-fold). Glutamine, aspartate, asparagine, cysteine, proline, and glutamate had minor but consistent increasing effects on CAT activity (2-to 3-fold). In contrast, alanine and serine had no significant effect on the level of CAT activity. These results provide evidence that the degree of effectiveness for each amino acid on the CHOP promoter activity varied widely. Moreover, this part of CHOP promoter also responds to glucose deprivation. In these experimental conditions, the induction in CAT activity due to glucose deprivation appeared to be not additive with that for leucine starvation. DISCUSSION In mammals, plasma concentrations of amino acids are affected by nutritional or pathological conditions. The experiments reported in this paper were designed to investigate the role of amino acids in the control of gene expression. A study performed by Marten et al. (12) showed that in a rat hepatoma cell line, removal of one amino acid in the culture medium induced an increase in the expression of several genes. Among these genes, CHOP expression exhibited the greatest induction in response to amino acid starvation. Nevertheless, molecular mechanisms involved in the regulation of CHOP mRNA expression have not been elucidated to date. To understand the regulation of gene expression by amino acids at a molecular level, we have studied the regulation of CHOP expression in response to leucine limitation.
The main effect of amino acid limitation on cellular function is the inhibition of protein synthesis. We show that low leucine concentrations (35 and 70 M) can induce CHOP expression but do not significantly inhibit total protein synthesis. However, this does not preclude the possibility that low leucine concentrations could affect the synthesis of particular proteins. These findings demonstrate that the regulation of CHOP expression by amino acid limitation is not a consequence of a cellular stress due to protein synthesis inhibition.
Since no general accumulation of mRNAs in amino acidstarved cells has been observed, mammalian cells must have a specific mechanism(s) that enables them to alter one specific pattern of gene expression in response to amino acid deprivation. Accumulation of asparagine synthetase, c-myc, c-jun, and c-fos mRNA have been reported to be induced transcriptionally and/or post-transcriptionally by amino acid starvation (8,43,44). We show that regulation of CHOP expression by leucine limitation has both transcriptional and post-transcriptional components. Our results clearly establish that the stability of CHOP mRNA is very low in the presence of leucine and is markedly increased in the absence of leucine. However, the mechanisms affecting CHOP mRNA stability in leucine-starved cells remain to be characterized. Furthermore, the induction of CHOP mRNA expression is sensitive to cycloheximide treatment suggesting that signaling pathways activated by leucine starvation involve synthesis of essential regulatory protein(s). We also show that starvation of other amino acids like lysine, methionine, arginine, phenylalanine, or threonine increases strongly CHOP promoter activity. These results suggest that gene regulation by leucine may be an example of a more general regulatory mechanism by which CHOP expres- sion would be controlled by the levels of amino acids. In yeast, the general control response to amino acid starvation is mediated through translational control of the positive transcription factor GCN4 which in turn modulates expression of numerous genes (2,45). Our results are in agreement with the existence in mammal cells of such a regulatory protein(s) involved in a general regulatory mechanism of gene expression by amino acid starvation. This hypothesis remains to be demonstrated. However, it is also possible, as suggested by Wang et al. (39), that what is being sensed is not the level of amino acids as such but rather some perturbation that arises when amino acid levels become limiting, for example the synthesis of abnormal proteins.
Our present results show that the 5Ј-flanking region of the human CHOP gene contains cis elements involved in the regulation of the CHOP transcription by leucine starvation. The promoter of the human asparagine synthetase gene has been shown to contain a 7-base pair region (5Ј-CATGATG-3Ј), designated amino acid response element (AARE), which mediates the transcriptional activation of the gene in response to amino acid starvation (9). Sequence analysis indicates that the CHOP promoter region contains several sequences homologous to the AARE, but their functional role remains to be demonstrated. Moreover, numerous regulatory elements that are likely to function in controlling the expression of this gene in response to cellular stress have been identified (31). Promoter deletion analyses have shown that several cis elements are involved in transcriptional activation of CHOP by UV irradiation or oxidant treatment (46). However, the cis elements involved in CHOP regulation by amino acid limitation remain to be identified.
The CHOP protein has been shown to heterodimerize with members of the C/EBP family (23). McKnight et al. (17) have hypothesized that C/EBP transcription factor family could play an important role in the control of energy metabolism. Through its interaction with C/EBPs, CHOP may participate in the regulation of downstream effector gene transcription during cellular response to amino acid limitation. It has been reported that C/EBP is involved in the transcriptional regulation of the carbamoyl-phosphate synthetase gene and two other urea cycle enzyme genes (47)(48)(49). Therefore, CHOP could play a crucial role in the regulation of nitrogen metabolism under amino acid control, although the cause and effect relationships have to be demonstrated.
In mammals, the plasma concentration of free amino acids shows striking alterations according to the nutritional or pathological conditions. For example, blood amino acid concentrations drop in animals fed with a low protein diet or starved (27,50). Under such extreme nutritional conditions, cells could undergo a limitation for essential amino acids. Indeed, Strauss et al. (11) have hypothesized that induction of IGFBP-1 gene expression in the liver of protein-restricted animals may be partially explained by a limitation for essential amino acids. We show that CHOP induction by amino acid limitation can take place (i) in all human cell lines tested (HepG2, CaCo-2, HeLa cells) and (ii) at a leucine concentration (70 M) in the range of those observed in the blood of protein-restricted animals. Therefore, leucine limitation related to those observed in nutritional situations may be a factor contributing to the induction of CHOP gene expression. Further work will be necessary to determine whether changes in blood amino acid concentrations could play an important role, in concert with hormones, in the modulation of gene expression.