Critical and Functional Regulation of CHOP (C/EBP Homologous Protein) through the N-terminal Portion*

C/EBP homologous protein (CHOP) is an endoplasmic reticulum stress-inducible protein that plays a critical role in the regulation of programmed cell death; however, the regulation of its function has not been well characterized. We have previously demonstrated that CHOP is regulated by the ubiquitin-proteasome system. In this study, during the process of clarifying the mechanism of the degradation of CHOP, we identified a novel regulation domain of CHOP in its N-terminal portion that is involved in various regulations and functions. The CHOP N-terminal domain is necessary not only for protein degradation but also for its transactivity and interaction with p300. In addition, trichostatin A, a histone deacetylase inhibitor, repressed the degradation of CHOP protein via the N-terminal domain. TRB3, a mammalian tribbles homolog that functions as a repressor of CHOP, also interacted with CHOP via the N-terminal portion and significantly blocked the association of p300 with CHOP. These results suggest that the N-terminal portion of CHOP plays a crucial role in its functional regulation and enable us to identify a novel function of TRB3 as an intracellular antagonist of the p300-binding domain of CHOP.

tive transactivator (12). In recent studies, several CHOP-inducible genes have been shown to be induced during ER stress via this CHOP binding sequence (13,14). We very recently demonstrated that TRB3, a novel ER stress-inducible protein, is induced by CHOP with a novel dimerizing partner, ATF4, which is a transcription factor of ATF/CREB family member, via a novel CHOP binding sequence, CHOP-amino acid response element sites (15).
A large number of transcription factors undergo degradation via a ubiquitin-proteasome-dependent pathway (16,17). A genetic study on Drosophila revealed that Slbo, a Drosophila homolog of C/EBP, is specifically degraded dependent on the expression of tribbles by the ubiquitin-proteasome (18). In humans, we have previously reported that C/EBP family transcription factors CHOP and Ig/EBP (C/EBP␥) are multiubiquitinated and subsequently degraded by the proteasome (19). TRB3, a human ortholog of tribbles, interacted with CHOP but did not promote degradation of CHOP protein. Therefore, the molecular mechanism involved in CHOP degradation is still unclear.
In this study, we identified the amino acid (aa) region required for degradation of the CHOP protein in its N-terminal portion. This region was also shown to be critical for CHOP transcriptional activity and interaction with p300; furthermore, TRB3 antagonized p300 binding to CHOP via this region. Degradation of CHOP protein via this region was suppressed by treatment with trichostatin A (TSA), and therefore this N-terminal domain of CHOP seemed to be crucial for various aspects of its functional regulation.
Cell Culture-Human melanoma cell line A375, human embryonic kidney cell line 293, and human hepatocellular carcinoma cell line HepG2 were cultured as described previously (19).
Reporter Gene Assays-Cells were transfected with luciferase reporter plasmids. After 48 h, lysates were prepared and luciferase assays were performed according to the manufacturer's instructions (Promega). All experiments were performed a minimum of three times, and the values obtained were used to calculate means and standard deviations.
Recombinant Protein Expression-pGEX-6P-1 plasmids encoding the GST proteins alone or GST-TRB3 or GST-FLAG-CHOP fusion proteins were transformed into the BL21 strain of Escherichia coli (Novagen, Madison, WI). Protein expression and purification were performed according to the procedures outlined in the Bulk GST Purification Module (Amersham Biosciences). Recombinant FLAG-CHOP protein was prepared using PreScission Protease (Amersham Biosciences).
Immunoprecipitation and Western Blot Analysis-Cells were transiently transfected and treated as described in the figure legends. The cells were lysed in radioimmune precipitation buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% SDS, 0.5% deoxycholate, and 1% Triton X-100) supplemented with protease inhibitors. The lysates were subjected to immunoprecipitation as described in the figure legends. One to two percent of the lysates or co-immunoprecipitates were subjected to SDS-PAGE (12.5%), transferred onto polyvinylidene difluoride membranes, and probed with antibodies indicated in the figure legends. The immunoreactive proteins were visualized using ECL Western blotting detection reagents (Amersham Biosciences), and light emission was quantified with a LAS1000 lumino image analyzer (Fuji).
Transfection-A375 cells were transfected by a lipofection method using Effectene (Qiagen, Hilden, Germany) according to the manufacturer's instructions. 293 and HepG2 cells were transfected by the Chen-Okayama method (21).

Endogenous CHOP Protein Is Degraded by the Proteasome
System-In normal conditions in multiple tissues and cells, the endogenous CHOP protein level is very low and its expression is induced at the transcriptional level during various extracellular and ER stresses. First, we investigated whether the level of endogenous CHOP protein is regulated through degradation by the proteasome during ER stress. We previously examined the effect of a proteasome inhibitor, MG132, on the endogenous CHOP protein level and showed that its expression was markedly augmented in the presence of MG132 (19). However, as the proteasome inhibitors induce ER stress by inhibiting degradation of unfolded proteins in the ER, there is a possibility that this augmentation was caused at the transcriptional level. Therefore, to directly assess whether CHOP protein is regulated by the proteasome system, we first examined the effect of MG132 on the rate of degradation of endogenous CHOP protein. The degradation rate was investigated by chasing with cycloheximide, a de novo protein synthesis inhibitor, in the presence of tunicamycin, a glycosylation inhibitor, for the indicated periods after pre-induction by treatment with tunicamy-cin for 6 h (Fig. 1, top). During ER stress induced by treatment with tunicamycin, endogenous CHOP protein level was effectively decreased by the degradation; however, NF-IL6 was not, as shown in a previous report (19). Similar results were obtained when CHOP was induced and chased by treatment with other ER stress inducers, A23187 and thapsigargin (data not shown). On the other hand, when chased by MG132, CHOP protein was significantly stabilized (Fig. 1, bottom). This repression of degradation was thought to be caused by the proteasome inhibitory action of MG132, thus confirming that endogenous CHOP induced during ER stress is degraded by the proteasome system.
The N-terminal Region of CHOP Is Critical for Degradation-During ubiquitination, ubiquitins are added to the lysine residues of target molecules through the association of specific ubiquitin ligases, E3. In addition, various modifications, such as phosphorylation, of substrates are necessary for the binding of E3 ligase(s) in some cases. Therefore, to investigate the mechanism of degradation of CHOP in detail, we constructed expression vectors for various mutants of CHOP and observed the expression level of these proteins. As shown in Fig. 2A, CHOP deletion mutants truncated in the N-terminal region (CHOP⌬N70 (aa 71-169) and CHOP⌬N36 (aa 37-169)) were expressed at much higher levels compared with wild-type CHOP ( Fig. 2A). On the other hand, other point and deletion mutants in the potential phosphorylation sites (Ser 79, 82 ) and basic region, respectively, were expressed at levels similar to wild-type CHOP, and the deletion mutants of the leucine zipper domain were expressed at slightly lower levels (data not shown). To assess whether the augmentation of the expression of N-terminal-deleted CHOPs was caused by inhibition of degradation by the proteasome, we next examined the effect of MG132 treatment on the expression of these mutants. As shown in a previous report (19), the accumulation of wild-type CHOP protein was increased by treatment with MG132; however, CHOP⌬N70 and CHOP⌬N36 were not affected (Fig. 2B). These results suggest that the N-terminal region of CHOP is necessary for degradation by the proteasome.
We examined the subcellular localization of wild-type, ⌬N70 (N-terminal deletion), or N70 (N-terminal alone) CHOP by using the GFP fusion system. GFP-CHOP⌬N70 was primarily localized within the nucleus, as was wild-type CHOP; by contrast, GFP alone or GFP-CHOP N70 (aa 1-70) was detected in both the nucleus and cytoplasm (Fig. 2C). This result indicates that the nuclear localization signal of CHOP exists in the C-terminal region (aa 71-169) and the resistance to degradation of the N-terminal deletion mutants does not result from a difference of this subcellular localization compared with that of wild-type CHOP.

Identification of the Region Required for Degradation of CHOP-
To further clarify the importance and function of the CHOP N-terminal region for degradation, we constructed additional N-terminal-truncated mutants (Fig. 3A). CHOP⌬N18 (aa 19 -169) and CHOP⌬19 -26 (aa 1-18, 27-169) were highly expressed and were not affected by MG132 treatment, which was also the case for CHOP⌬N70 and CHOP⌬N36. On the other hand, CHOP⌬37-64 (aa 1-36, 65-169) showed almost the same basal expression level and enhancement of accumulation by MG132 as wild-type CHOP (  showed slightly lower expression levels as compared with the mutants insensitive to MG132, such as CHOP⌬N18, and a slight increase of accumulation by MG132 treatment (Fig. 3B).
These results indicate that the region between aa 10 and 26 in CHOP is required for degradation by the proteasome. As the main part of this region consists of an ␣-helix structure (Fig. 3A) (22), this motif may be significant for the binding to E3 ligase(s) or related molecule(s).
In the ubiquitin-proteasome system, the level of degradation of many factors is controlled by the regulation of phosphorylation or dephosphorylation. These modifications act as switches for direct regulation, such as binding to E3 ligase(s), or indirect regulation, for example, by changing the subcellular localization. As the region between aa 10 and 26 of CHOP, which is required for the degradation, contains some potential sites for phosphorylation (Fig. 3C), we examined the effect of substitution of these amino acids with Ala on the sensitivity to the proteasome inhibitor. CHOP S5,9A, CHOP T12A and CHOP S14,15A protein levels were augmented by MG132 treatment (Fig. 3D). On the other hand, CHOP Y22A protein was accumulated under the normal conditions and was not affected by the proteasome inhibitor; however, the substitution of Tyr 22 by Phe (CHOP Y22F) resulted in almost the same accumulation in response to MG132 treatment as wild-type CHOP (Fig.  3D). These results suggest that the phosphorylation or dephosphorylation of the CHOP N-terminal region is not significant for the constitutive degradation of CHOP but the Tyr 22 amino acid residue plays an essential role in the degradation, probably due to its role in maintaining the N-terminal ␣-helix structure. In summary, we concluded that the N-terminal ␣-helix structure of CHOP is necessary for the degradation.
N-terminal-deleted CHOPs Are Not Polyubiquitinated-In the presence of MG132, CHOP protein was accumulated and multiubiquitinated, indicating that CHOP is constitutively ubiquitinated and degraded by the proteasome (19) (Fig. 2B). As the protein levels of N-terminal-deleted CHOPs were increased in the basal condition and were not affected by the treatment with proteasome inhibitor (Figs. 2 and 3), we examined whether these mutants are polyubiquitinated. Consistent with the protein stabilities, wild-type CHOP and CHOP⌬N9 were highly and slightly, respectively, ubiquitinated upon co-expression with ubiquitin; however, CHOP⌬N18 and CHOP⌬19 -26 were not ubiquitinated under these conditions (Fig. 3E). CHOP Y22A was also not ubiquitinated, but the other substituted mutants shown in Fig. 3C were multiubiquitinated (data not shown). These results also suggest that the region between aa 10 and 26 in CHOP is essential for the polyubiquitination preceding degradation.
The N-terminal Region of CHOP Is Also Critical for Its Transcriptional Activity and Interaction with a Coactivator, p300-We previously reported that the region between aa 10 and 18 in CHOP is very important for transcriptional activity, as shown by analysis using CHOP fusion proteins with the GAL4 DNA binding region (15). The transcriptional activity of GAL4-CHOP⌬19 -26 was also lost in that assay (Fig. 4A). These results suggest that the region aa 10 -26 of CHOP is essential for not only the protein stability but also the transcriptional activity of CHOP. Therefore, to explore the possibility that CHOP associates with primary transcription factor(s) or coactivator(s) via this region, we examined the interaction of CHOP with p300, a transcriptional coactivator having histone acetyl transferase activity. As shown in Fig. 4B, wild-type CHOP strongly associated with p300; however, CHOP⌬⌵18, ⌬19 -26, and ⌬N36 almost did not. In addition, consistent with the transcriptional activities, the interaction of CHOP⌬N9 with p300 was reduced but still significant (Fig. 4B). These results suggest that the N-terminal region of CHOP is critical for the interaction of CHOP with p300 as well.
TRB3 Inhibits the Association of CHOP with p300-Recently, TRB3, a human ortholog of tribbles, was identified as a novel Akt-binding and -regulating protein (23). We previously demonstrated that TRB3 is an ER stress-inducible protein and interacts with CHOP to inhibit its transcriptional activity (15). In addition, the region between aa 10 and 18 in CHOP is necessary for the interaction with TRB3 (15), indicating that this region is overlapping with the p300-binding site. Therefore, we examined the effect of TRB3 expression for the association of CHOP with p300. As shown in Fig. 5A, TRB3 expression dramatically inhibited the binding of CHOP with p300; on the other hand, p300 expression did not affect the binding of CHOP with TRB3. In our previous study, TRB3 also suppressed the transcriptional activity of ATF4, another ER stress-induced transcription factor (15). As shown in Fig. 5B, TRB3 inhibited the p300-ATF4 interaction as well. In an in vitro binding assay, recombinant GST-TRB3 dose-dependently interacted with recombinant FLAG-CHOP and inhibited the association of p300 with that (Fig. 5C). In addition, TRB3 expression suppressed coactivation of the transcriptional activity of wild-type CHOP by p300, whereas neither TRB3 nor p300 affected the transcriptional activity of CHOP ⌬N18 (Fig. 5D). These results suggest that the affinity of the CHOP-TRB3 interaction is probably high as compared with that of the CHOP-p300 interaction and therefore TRB3 inhibits CHOP-dependent transcriptional activation by preventing CHOP-p300 association.
Trichostatin A Represses CHOP Protein Degradation-P300/ CBP coactivators acetylate not only chromatin-conjugated histone but also various molecules such as transcription factor p53. The inhibition of cellular deacetylases leads to a longer half-life of endogenous p53; furthermore, p53 is ubiquitinated and acetylated on similar sites at the C terminus, suggesting that these modifications may compete for the same residues (24,25). As CHOP bound strongly to p300, we next explored whether p300 acetylates CHOP and regulates its ubiquitination and degradation. First, we examined the effect of TSA, a histone deacetylase (HDAC) inhibitor, on the CHOP protein level and found that exogenous wild-type CHOP protein, but not ⌬N18 or ⌬19 -26, accumulated in response to treatment with TSA as well as MG132 (Fig. 6A). The additive or synergistic effect of MG132 and TSA on the CHOP protein accumulation was not observed, suggesting that the effects of these treatments are caused by a similar mechanism. In addition, endogenous CHOP induced by tunicamycin was also accumulated as a result of TSA treatment in HepG2 cells (Fig. 6B) or A375 and 293 cells (data not shown). Furthermore, endogenous CHOP protein degradation was repressed in the presence of TSA (Fig. 6C). These results suggest that CHOP protein is stabilized by inhibition of HDAC activities. The stabilization of CHOP by TSA treatment was thought to be induced through the acetylation of CHOP protein. Therefore, we examined whether CHOP is acetylated by p300 in the presence of TSA or MG132. By Western blotting using an antibody specifically recognizing acetylated lysine residues, acetylated p53 was detected as a positive control; however, acetylated CHOP was not detected at all (Fig.  6D). This result suggests that the accumulation of CHOP protein induced by TSA is not caused through an antagonistic mechanism of acetylation for ubiquitination. On the other hand, CHOP interacted with some HDACs (HDAC1, HDAC5, and HDAC6) (Fig. 6E), and these HDACs may regulate the degradation of CHOP through some mechanism not involving the deacetylation of CHOP itself.

DISCUSSION
CHOP is a transcription factor induced by ER stress and is involved in apoptosis. In this study, we demonstrated that the accumulation of CHOP protein during ER stress is regulated through modulation of its degradation by the proteasome. In addition, the region between aa 10 and 26 of CHOP was essential for its degradation. In the N-terminal region aa 1-70 of CHOP, a lysine residue, which would be necessary for modification by ubiquitins, only exists at position aa 50. CHOP⌬N36 was not accumulated as a result of MG132 treatment despite the absence of Lys 50 , indicating that the enhancement of accumulation caused by N-terminal deletion does not result from the lack of the ubiquitination of lysine residues. Indeed, the substitution of CHOP Lys 50 with Arg (CHOP K50R) resulted in almost the same protein expression level and accumulation upon MG132 treatment as for the wild-type CHOP (data not shown). At the same time, these results indicate that the Nterminal region of CHOP is important for degradation by the proteasome and probably for binding to E3 ligase(s) or related molecule(s).
In this study, we have identified a novel functional domain of CHOP in the region between aa 10 and 26 that is necessary for not only the degradation of the protein but also its transcriptional activity and interaction with p300. Recently, there have been several reports showing that the acetylation of transcription factors can trigger critical regulation of activation or inactivation (26). CHOP associated with p300 but was not acetylated by it. Therefore, at least, p300 seems to not regulate the activity of CHOP via its acetylation. On the other hand, recently p300 was shown to be a crucial factor for degradation of p53 as E4 ligase (27). The N-terminal region of CHOP necessary for degradation was essential for interaction with p300; therefore, p300 may regulate the degradation of CHOP protein as well.
CHOP associated with MDM2, the E3 ligase for p53 (25), and Fbw1, the F-box protein for IB (28) or ␤-catenin (29), in vivo (data not shown). The molecular mechanisms underlying whether these molecules function as the E3 ligase for CHOP are currently under investigation. We showed here that CHOP degradation is repressed by the inhibition of deacetylation activity by TSA treatment via the region aa 10 -26 as well. CHOP interacted with HDACs but was not acetylated upon the treatment with TSA, indicating that the degradation of CHOP is not regulated by its acetylation. Recently, it was revealed that the acetylation of MDM2 causes inactivation of its activity for degradation of p53 (30). Similarly, the E3 ligase for CHOP may be activated through deacetylation by HDAC(s) in untreated cells and may be inactivated through the inhibition of HDAC activity by TSA as well.
We previously showed that TRB3, an ER stress-induced kinase-like protein, associated with CHOP to suppress CHOPdependent transcriptional activation through a proteolysis-independent pathway (15). TRB3 did not interfere with the dimerization of CHOP or with its DNA binding activity. In addition, as TRB3 repressed even the transcriptional activity of a GAL4 fusion protein of CHOP, it was suggested that TRB3 primarily inhibits CHOP transactivation, probably by inhibiting the modification of CHOP required for its transactivation or by interfering with the association of coactivator(s) or by recruiting corepressor(s) to DNA. Indeed, the TRB3 binding region of CHOP, aa 10 -18, is a part of its p300 binding region, aa 10 -26, and TRB3 expression significantly blocked the association of p300 with CHOP. TRB3 suppressed the transcriptional activity of ATF4, another ER stress-induced transcrip- tion factor. In the present study, TRB3 repressed the p300-ATF4 interaction as well. These results suggest a novel function of TRB3 as an intracellular antagonist of the p300-binding domains of both CHOP and ATF4. As TRB3 and other tribbles family members contain the classic substrate-binding domains of a protein kinase but not the ATP-binding and kinase-activa-  DECEMBER 7, 2007 • VOLUME 282 • NUMBER 49 tion domains (31), TRB3 could act as a novel type of decoy kinase-like protein for Akt or other substrates. Similarly, TRB3 could act as a novel type of decoy deacetylase-like protein for p300 as well. This study revealed that various molecules regulate the function of CHOP via effects on its degradation. CHOP is induced by various stress signals, such as ER stress, oxidative stress, hypoxia, and amino acid deprivation. Therefore, it is possible that CHOP signaling is strictly regulated by these stresses via its N-terminal portion. The analysis of these mechanisms may identify potential targets for the regulation of CHOP function.