Posttranscriptional and Posttranslational Regulation of C/EBPδ in G0 Growth-arrested Mammary Epithelial Cells*

Previous work from our laboratory demonstrated that CCAAT/enhancer-binding protein δ (C/EBPδ) functions in the initiation and maintenance of G0 growth arrest in mouse mammary epithelial cells (MECs). In this report, we investigated the posttranscriptional and posttranslational regulation of C/EBPδ in G0 growth-arrested mouse MECs. The results of transcriptional inhibitor studies demonstrated that the C/EBPδ mRNA exhibits a relatively short half-life in G0growth-arrested mouse MECs (t 1 2 ∼35 min). In contrast, C/EBPδ mRNA has a longer half-life in G0 growth-arrested mouse fibroblast cells (t 1 2 >100 min). Oligo/RNase H cleavage analysis and rapid amplification of cDNA ends-poly(A) test both confirmed the short C/EBPδ mRNA half-life observed in MECs and demonstrated that the C/EBPδ mRNA poly(A) tail is relatively short (∼100 nucleotides). In addition, the poly(A) tail length was not shortened during C/EBPδ mRNA degradation, which suggested a deadenylation-independent pathway. The C/EBPδ protein also exhibited a relatively short half-life in G0growth-arrested mouse MECs (t 1 2 ∼120 min). The C/EBPδ protein was degraded in a ubiquitin-dependent manner, primarily in the nucleus, during G0 growth arrest. In conclusion, these studies indicated that the C/EBPδ mRNA and protein content are under tight regulation in G0 growth-arrested mouse MECs, despite the general concept that G0 growth arrest is associated with a decrease in cellular activity.

It is well documented that global gene transcription and translation dramatically decrease as cells exit the cell cycle and enter the quiescent G 0 growth arrest state (1,2). However, a small subset of genes become activated, and these gene products function in the initiation and maintenance of G 0 growth arrest (3,4). Currently, little is known about the regulation and function of G 0 growth arrest-specific genes in cell biology. Recently, an increase in expression and activity of the retinoblastoma (Rb) family member p130 was observed in the initiation and maintenance of G 0 growth arrest (5). Formation of the p130-E2F complex sequesters members of the E2F family to repress the expression of genes necessary for cellular proliferation. Alterations in the structure and function of G 0 growth arrest genes are linked to some types of cancers. For example, germ line mutations found within the von Hippel-Lindau tumor suppressor gene have been linked to hemangioblastomas of the retina and central nervous system and renal carcinogenesis (6). The von Hippel-Lindau protein functions in cell cycle control in a variety of ways, including up-regulating the cyclindependent kinase inhibitor p27 (7).
C/EBPs are directly involved in the regulation of cell fate determination (20 -26). Early reports (20,27) demonstrate that the sequential expression of C/EBP␦, C/EBP␤, and C/EBP␣ is required for optimal adipocyte differentiation. Further studies (28 -30) have identified additional roles for C/EBP␣ in hepatocyte metabolism and granulocyte differentiation. C/EBP␤ also plays an essential role in ovarian granulosa cell biology and the development and differentiation of the mammary gland (22)(23)(24)(25)31). Furthermore, C/EBP⑀ functions in the development and differentiation of neutrophils and eosinophils (32,33).
Control of gene expression can occur at the transcriptional, posttranscriptional, or posttranslational level (34 -37). At the posttranscriptional level, mRNA stability is emerging as a key regulatory mechanism in cell cycle control and DNA damage repair (34, 35, 38 -40). For example, the stability of growth arrest and DNA damage-inducible mRNA increases after exposure to DNA-damaging agents or other growth arrest treatments (40). In addition, the growth arrest-specific gene 5 (gas-5) exhibits a marked increase in mRNA stability in density-arrested NIH 3T3 cells versus exponentially growing and differentiating cells (41). Alterations in the posttranscriptional regulation of genes that function in cell growth control and cell cycle progression can play a crucial role in tumorigenesis (42,43). For example, alterations of trans-acting factors that function in c-myc and c-myb mRNA turnover results in increased c-myc and c-myb mRNA stability, which is linked to acute myeloid leukemia (44). Additionally, an increase in the basic fibroblast growth factor mRNA half-life, due to defects in posttranscriptional regulation, has been implicated in a variety of human tumors (45). Although several reports demonstrate that C/EBP␦ is regulated at the transcriptional level (12,13), posttranscriptional control of C/EBP␦ has not been investigated extensively.
In addition to posttranscriptional control, several genes that play critical roles in cell cycle control are regulated posttranslationally, at the level of protein degradation (46 -49). For example, the rate of p27 protein degradation decreases in response to growth arrest, which results in an accumulation of p27 protein (50,51). Blocking ubiquitination-dependent protein degradation increases p27 protein half-life and demonstrates that the p27 protein is degraded via the ubiquitin/ proteasome pathway (52,53). In addition, increased p53 protein stability occurs during cellular genotoxic stress (54). This is accomplished by N-terminal phosphorylation of the p53 protein, which decreases the degree of ubiquitination and increases protein stability (54). Accumulating evidence indicates that cellular proteins may be degraded by ubiquitin-mediated mechanisms localized to either the nucleus or cytoplasm (55,56). Nuclear localized ubiquitin-mediated degradation appears to provide a rapid mechanism for the disposal of nuclear cell cycle regulatory proteins (57). For example, the tumor suppressor protein product, p53, is degraded within the nucleus via a ubiquitin-proteasome pathway during post-stress recovery (57).
The overall goal of this study was to investigate the posttranscriptional and posttranslational regulation of C/EBP␦ in G 0 growth-arrested mouse MECs. Our laboratory has reported previously (12) that C/EBP␦ exhibits increased transcription and growth suppressor activity in G 0 growth-arrested mouse MECs. However, the posttranscriptional and posttranslational regulation of C/EBP␦ has not been systematically investigated. Previous studies (35) have demonstrated that key cell cycle regulatory proteins are encoded by unstable mRNAs. Because G 0 growth arrest is associated with a period of decreased cellular activity, we hypothesized that both the C/EBP␦ mRNA and protein would exhibit extended half-lives in G 0 growtharrested MECs. Unexpectedly, the results demonstrate that C/EBP␦ mRNA exhibited a novel short mRNA half-life in G 0 growth-arrested mouse MECs (t1 ⁄2 ϳ35 min) and contained a relatively short poly(A) tail of ϳ100 nucleotides. In addition, the C/EBP␦ protein also exhibited a short half-life in G 0 growth-arrested mouse MECs (t1 ⁄2 ϳ120 min). Furthermore, ubiquitination inhibitor studies indicated that C/EBP␦ protein degradation is ubiquitin-dependent and occurs predominantly within the nucleus. The results of these studies demonstrate that the C/EBP␦ mRNA and protein are under tight regulation in G 0 growth-arrested MECs, suggesting that C/EBP␦ plays a key role in mouse MEC growth control.

C/EBP␦ mRNA Exhibits a Short Half-life in G 0 Growtharrested Mouse HC11
MECs-To investigate the posttranscriptional regulation of the C/EBP␦ mRNA in mouse HC11 MECs, we utilized transcriptional inhibitors followed by Northern blot analysis. Briefly, confluent HC11 MECs were G 0 growth-arrested by serum and growth factor withdrawal. After 48 h, HC11 MECs were either maintained in growth arrest medium alone (GAMϪ) or in GAM with the addition of the transcriptional inhibitor, actinomycin D (GAMϩ) for the indicated times. mRNA half-life was analyzed for a panel of cellular mRNAs including cyclophilin (cp), which was used to confirm equal loading. Consistent with previous reports from our laboratory, C/EBP␦ mRNA was detected in G 0 growth-arrested HC11 MECs (Fig. 1A, lanes 1-4). C/EBP␦ mRNA levels rapidly declined with a half-life of ϳ35 min (t1 ⁄2 ϳ35 min) following actinomycin D treatment (Fig. 1A, lanes 5-8 and Fig. 1C). After 60 min of actinomycin D treatment, C/EBP␦ mRNA was undetectable (Fig. 1A, lane 7). Similar results were observed with a second transcriptional inhibitor, DRB (Fig. 1B).
In addition to C/EBP␦, we also investigated the mRNA level of another C/EBP family member, C/EBP␤. Reports from a number of laboratories, including our own, demonstrate that C/EBP␤ plays a significant role in mammary gland growth and differentiation (22,24,25,31). C/EBP␤ mRNA was detected in G 0 growth-arrested HC11 MECs, suggesting that C/EBP␤ also plays a role in G 0 growth arrest (Fig. 1A, lanes 1-4). C/EBP␤ mRNA levels declined with a half-life of ϳ45 min following the addition of actinomycin D (Fig. 1A, lanes 5-8, and Fig. 1C).
In summary, results of transcriptional inhibitor studies indicate that C/EBP␦ mRNA is highly unstable during G 0 growth arrest in HC11 MECs. C/EBP␤, which has been associated previously with cellular proliferation and differentiation, also exhibits a relatively short mRNA half-life in G 0 growth-arrested HC11 MECs. In contrast, gas-1 mRNA is more stable during G 0 growth arrest. Overall, the results suggest that the C/EBP␦ mRNA is undergoing rapid turnover despite the gen-eral decline in global gene expression and biosynthetic activity during G 0 growth arrest.
C/EBP␦ mRNA Half-life during Cell Cycle Re-entry in HC11 MECs-Cell cycle re-entry requires coordination between the inactivation and/or disposal of G 0 -specific proteins and the expression of early G 1 genes, such as c-fos and c-myc. To investigate the posttranscriptional regulation of C/EBP␦ mRNA during early G 1 , mRNA levels were analyzed from HC11 MECs upon addition of complete growth media alone (CGMϪ) or CGM plus actinomycin D (CGMϩ) at the indicated time points. In agreement with previous reports from our laboratory, C/EBP␦ mRNA levels decline with the onset of G 1 and the initiation of the cell cycle (Fig. 2, lanes 1-4) (41). C/EBP␦ mRNA levels also decreased in early G 1 following actinomycin D treatment; however, the cell cycle-induced decline of C/EBP␦ mRNA was delayed compared with the decline observed in G 0 growth arrest (Fig. 2, lanes 5-8).
When G 0 growth-arrested HC11 MECs were induced to reenter the cell cycle by CGM addition, C/EBP␤ mRNA levels increased ϳ10-fold within the first 90 min (Fig. 2, lane 1-4). This induction of C/EBP␤ mRNA levels during early G 1 is consistent with a growth-promoting role for C/EBP␤. The addition of actinomycin D blocked the growth-stimulated induction of C/EBP␤ mRNA, which suggests that C/EBP␤ transcription plays a major role in the increase in C/EBP␤ mRNA levels during early G 1 in HC11 MECs (Fig. 2, lanes 5-8).
gas-1 mRNA levels also declined after G 0 growth-arrested HC11 MECs were induced to re-enter the cell cycle by refeeding with CGM (Fig. 2, lanes 1-4). Interestingly, addition of CGM and actinomycin D stabilized the gas-1 mRNA, resulting in high levels of gas-1 mRNA even at 90 min (Fig. 2, lanes 5-8). In agreement with previous reports, c-fos mRNA was transiently induced following the addition of CGM and the initiation of the cell cycle (Fig. 2, lanes 1-4). The induction of c-fos mRNA was blocked by actinomycin D treatment, indicating that c-fos gene transcription is required for the increase in c-fos mRNA during early G 1 (Fig. 2, lanes 5-8).
The results of the transcriptional inhibitor studies during cell cycle re-entry demonstrate that the C/EBP␦ and gas-1 mRNAs are more stable during the G 0 /G 1 transition compared with G 0 growth arrest. This suggests that both C/EBP␦ and gas-1 mRNA degradation during cell cycle re-entry is dependent on the transcription of gene product(s) important for mRNA decay during the G 0 /G 1 transition. In addition, the increase in C/EBP␤ and c-fos mRNA levels during the G 0 /G 1 transition are inhibited by actinomycin D treatment, indicating that the increase in these immediate early mRNAs is transcriptiondependent.
C/EBP␦ mRNA Is More Stable in G 0 Growth-arrested NIH 3T3 Cells Compared with HC11 MECs-NIH 3T3 cells have been utilized extensively as a model system to investigate mechanisms of cell growth control. We reported previously (11) that C/EBP␦ mRNA is present in NIH 3T3 cells regardless of growth status. To investigate the posttranscriptional control of the C/EBP␦ mRNA in NIH 3T3 cells, transcriptional inhibitor/ Northern blot analysis was performed. In agreement with our previous results, C/EBP␦ mRNA was detected in G 0 growtharrested (GAMϪ) NIH 3T3 cells (Fig. 3A, lanes 1-4). C/EBP␦ mRNA levels declined following actinomycin D treatment, although the rate of decline is slower than that observed in G 0 growth-arrested HC11 MECs (t1 ⁄2 Ͼ100 min for NIH 3T3 cells versus t1 ⁄2 ϳ35 min for HC11 MECs) (Fig. 3A, lanes 5-8 and Fig.  3B). Following cell cycle re-entry, C/EBP␦ mRNA levels decreased by 90 min (Fig. 3A, lanes 9 -12). The addition of actinomycin D stabilized C/EBP␦ mRNA (Fig. 3A, lanes 13-16), paralleling the results from experiments in HC11 MECs.
The extended C/EBP␦ mRNA half-life detected in G 0 growtharrested NIH 3T3 cells suggests that C/EBP␦ is under less stringent control in mouse fibroblast-derived cells compared with mouse mammary epithelial-derived cells. In contrast, posttranscriptional regulation of C/EBP␦ mRNA during cell cycle re-entry is similar between HC11 MECs and NIH 3T3 cells. The posttranscriptional regulation of C/EBP␤, gas-1, and c-fos in both G 0 growth arrest and cell cycle re-entry is comparable between HC11 MECs and NIH 3T3 cells.
C/EBP␦ mRNA Contains a Short Poly(A) Tail-Sequences present within mRNAs influence processing, stability, and transport (34,35,39,63). To investigate the role of the poly(A) tail on C/EBP␦ mRNA stability, we utilized an oligo/RNase H cleavage Northern blot analysis (58). In this analysis, a C/EBP␦-specific oligomer complimentary to the C/EBP␦ mRNA within the 3Ј-untranslated region (UTR) (oligo 1193) was used to form a DNA/RNA heteroduplex that is cleaved by RNase H, producing C/EBP␦ 5Ј and 3Ј mRNA fragments (Fig. 4A). A [␣-32 P]dCTP-labeled C/EBP␦ 3Ј-UTR-specific probe that spanned the oligo/RNase H digestion site was used to detect both of the oligo/RNase H-generated C/EBP␦ fragments: the 5Ј C/EBP␦ cleavage product (ϳ1.4 kb) composed of the C/EBP␦ mRNA 5Ј-untranslated region (UTR), coding sequence, and partial 3Ј-UTR and the 3Ј C/EBP␦ mRNA cleavage product composed of the remaining 3Ј-UTR (ϳ260 bp) plus the length of the poly(A) tail.
Initially, we performed the oligo/RNase H cleavage analysis on RNA from G 0 growth-arrested (GAMϪ) HC11 MECs, and Northern blot analysis detected two cleavage products of ϳ1.4 kb and 370 bp (Fig. 4B, lanes 1-3). The 3Ј-UTR cleavage product contains 260 bp from the 3Ј-UTR and reveals a poly(A) tail of ϳ100 nucleotides. To investigate the mechanism of C/EBP␦ mRNA degradation, mRNA was isolated from actinomycin Dtreated G 0 growth-arrested (GAMϩ) HC11 MECs. After 30 min of actinomycin D treatment, both cleavage products were detected (Fig. 4B, lane 4). However, after 60 min of actinomycin D treatment only the 3Ј oligo/RNase H C/EBP␦ mRNA cleavage product was detected (Fig. 4B, lane 5). These results confirm the short half-life of the C/EBP␦ mRNA in G 0 growth-arrested HC11 MECs.
We next investigated the C/EBP␦ mRNA poly(A) tail length during the G 0 /G 1 transition. G 0 growth-arrested HC11 MECs were induced to re-enter the cell cycle by the addition of CGM alone (CGMϪ) or CGM plus actinomycin D (CGMϩ). Upon addition of CGM, the reduction of C/EBP␦ mRNA oligo/RNase H cleavage products was observed, consistent with a decrease in the C/EBP␦ mRNA content upon the onset of early G 1 (Fig.  4B, lanes 6 -8). After addition of actinomycin D, the reduction of C/EBP␦ mRNA oligo/RNase H cleavage products was slightly FIG. 3. C/EBP␦ mRNA stability during G 0 growth arrest and the G 0 /G 1 transition in NIH 3T3 cells. RNA isolated from untreated and actinomycin D-treated G 0 growth-arrested or serum and growth factor-stimulated (G 0 /G 1 transition) NIH 3T3 cells was analyzed by Northern blot as described in Fig. 1. A, lanes 1-4, RNA from G 0 growth-arrested MECs (GAMϪ); lanes 5-8, RNA from G 0 growth-arrested MECs treated with actinomycin D (GAMϩ); lanes 9 -12, RNA from serum-and growth factor-stimulated (G 0 /G 1 transition) MECs (CGMϪ); lanes 13-16, RNA from serum-and growth factor-stimulated (G 0 /G 1 transition) MECs treated with actinomycin D (CGMϩ). Results are representative of three independent experiments. B, summary of mRNA half-life data obtained from Northern blot/actinomycin D (Act.D) analysis as determine in Fig. 1. delayed compared with GAMϩ, and a more complex pattern of C/EBP␦ mRNA degradation was detected (Fig. 4B, lanes 9 -11). The results of the CGM and actinomycin D experiment suggest that transcription of gene products important for mRNA decay is required for efficient C/EBP␦ mRNA degradation during cell cycle re-entry. Finally, the estimated size of all the C/EBP␦ 3Ј oligo/RNase H cleavage products is consistent with a poly(A) tail length of ϳ100 nucleotides. To confirm the length of the poly(A) tail, we utilized an oligo(dT) in the oligo/RNase H experiment, which generates a C/EBP␦ mRNA product lacking a poly(A) tail (Fig. 4B, lane 12). The results reveal that the mobility of this mRNA product compared with full-length C/EBP␦ mRNA (Fig. 4B, lane 13) is consistent with a poly(A) tail length of ϳ100 nucleotides. Overall, the results demonstrate that the C/EBP␦ mRNA contains a relatively short poly(A) tail that is not shortened during mRNA degradation in HC11 G 0 growth arrest and cell cycle re-entry.

RACE-PAT Analysis Confirms C/EBP␦ mRNA Poly(A) Tail
Length-To investigate further the C/EBP␦ mRNA poly(A) tail length, we utilized a rapid amplification of cDNA ends-poly(A) test (RACE-PAT) (58). Initially, mRNA was obtained from G 0 growth-arrested HC11 MECs; cDNA was synthesized, and PCR was performed utilizing a radiolabeled C/EBP␦ 3Ј-UTR upstream-specific primer and an oligo(dT) downstream primer. The PCR produced multiple products varying in length from 170 to 270 bp (Fig. 5, lanes 1-3), which is consistent with a C/EBP␦ poly(A) tail length of ϳ100 nucleotides. Following addition of actinomycin D, C/EBP␦ mRNA levels declined rapidly as observed previously in the transcriptional inhibitor/Northern blot analysis (Fig. 5, lanes 4 -6). This decline is reflected in the decrease in the amount of RACE-PAT product obtained. RACE-PAT analysis was also performed on mRNA obtained from HC11 MECs refed with CGM, which also indicated that the C/EBP␦ mRNA poly(A) length is approximately ϳ100 nucleotides (Fig. 5, lanes 7-9). Additionally, actinomycin D treatment resulted in a more stable C/EBP␦ mRNA but had no apparent effect on poly(A) tail length (Fig. 5, lanes 10 -12). The results confirm that the C/EBP␦ mRNA contains a poly(A) tail of ϳ100 nucleotides in G 0 growth arrest and during the G 0 /G 1 transition and parallel previous mRNA half-life results from experiments in HC11 MECs (Figs. 1A and 2A).
The 3Ј-Untranslated Region Influences C/EBP␦ mRNA Stability-Numerous studies (34,35,39) have demonstrated that specific sequences within the 3Ј-UTR regulate mRNA stability. To investigate the potential role of the C/EBP␦ 3Ј-UTR in mRNA stability, transcriptional inhibitor/Northern blot analysis was repeated utilizing a stably transfected C/EBP␦ overexpression HC11 MEC line previously developed in our laboratory (13). This HC11 MEC line expresses an exogenous C/EBP␦ mRNA that contains a bovine growth hormone (BGH) 3Ј-UTR in place of the C/EBP␦ 3Ј-UTR. The cells were growth-arrested for 48 h and maintained in GAM in the presence or absence of actinomycin D. The C/EBP␦/BGH 3Ј-UTR mRNA levels were compared with the endogenous C/EBP␦ mRNA levels. Consistent with previous results, the endogenous C/EBP␦ mRNA levels were detected in G 0 growth-arrested cells (Fig. 6, lanes 1-4).
Similarly, the C/EBP␦/BGH 3Ј-UTR mRNA was also detected in G 0 growth-arrested HC11 MECs at all indicated time points (Fig. 6, lanes 1-4). Consistent with previous results, the addition of actinomycin D resulted in a rapid reduction of endogenous C/EBP␦ mRNA levels (t1 ⁄2 ϳ35 min) (Fig. 6, lanes 5-7). Interestingly, the C/EBP␦/BGH 3Ј-UTR mRNA levels did not decline after actinomycin D treatment (Fig. 6, lanes 5-7). These results indicate that the presence of the BGH 3Ј-UTR downstream of the C/EBP␦ coding region stabilizes the C/EBP␦ mRNA compared with the endogenous C/EBP␦ mRNA and suggests that the C/EBP␦ 3Ј-UTR plays a role in mRNA stability during G 0 growth arrest.

C/EBP␦ Protein Exhibits a Short Half-life in G 0 Growtharrested Mouse
MECs-Posttranslational control is a major mechanism by which cells regulate the level of cell cycle control proteins (46 -49). To investigate the posttranslational regulation of the C/EBP␦ protein, HC11 MECs were G 0 growtharrested for 48 h and maintained in GAM in the presence or absence of the translational inhibitor anisomycin. Protein halflife was analyzed for a panel of cellular proteins by Western blot analysis including actin, which was used to confirm equal loading. Consistent with previous reports from our laboratory (11)(12)(13), C/EBP␦ protein was detected in G 0 growth-arrested HC11 MECs (Fig. 7A, lanes 1-6). Following anisomycin treatment, C/EBP␦ protein levels declined with a half-life of ϳ120 min (Fig. 7A, lanes 7-12, and Fig. 7B). As a control, the stability of p27, a growth arrest-specific protein that is regulated predominantly at the posttranslational level, was assessed (50,51). As expected, p27 protein was detected during G 0 growth arrest of HC11 MECs (Fig. 7A, lanes 1-6). In contrast to the C/EBP␦ protein decay kinetics, the p27 protein is relatively stable in G 0 growth-arrested HC11 MECs even after treatment with anisomycin (Fig. 7A, lanes 7-12). The results demonstrate that the C/EBP␦ protein exhibits a shorter half-life compared with p27, which suggests that C/EBP␦ protein is tightly regulated in G 0 growth-arrested mouse MECs.
C/EBP␦ Protein Is Degraded via a Ubiquitin-dependent Pathway in G 0 Growth-arrested Mouse MECs-To investigate the protein degradation pathway utilized by the C/EBP␦ protein in G 0 growth-arrested mouse MECs, we used the ubiquitination inhibitor, MG-132. HC11 MECs were growth-arrested by addition of GAM for 48 h and subsequently maintained in either growth arrest media alone (control), GAM plus a vehicle control FIG. 5. The short C/EBP␦ mRNA poly(A) tail is confirmed by RACE-PAT analysis. RACE-PAT was performed on 5 g of total RNA isolated from untreated and actinomycin D-treated G 0 growth-arrested or serum-and growth factor-stimulated (G 0 /G 1 transition) HC11 MECs at the indicated time points. cDNA was synthesized and subjected to PCR, and products were visualized by PAGE and autoradiography. Lanes 1-3, G 0 growth-arrested MECs (GAMϪ); lanes 4 -6, G 0 growtharrested MECs treated with actinomycin D (GAMϩ); lanes 7-9, serum and growth factor stimulated (G 0 /G 1 transition) MECs (CGMϪ); lanes 10 -12, serum and growth factor stimulated (G 0 /G 1 transition) MECs treated with actinomycin D (CGMϩ). Results are representative of three independent experiments.
As a control, p27 protein levels were monitored in both mouse and human MECs after MG-132 treatment. Previous work (52,53) in vitro and in vivo demonstrates that the ubiquitin-proteasome pathway regulates the p27 protein level. p27 protein was detected in both G 0 growth-arrested control and Me 2 SO samples (Fig. 8, A and B). As expected, a modest increase in p27 protein level is detected after MG-132 treatment in HC11 MECs (Fig. 8A, lanes 8 -10). Taken together, these results demonstrate that the C/EBP␦ protein is degraded via a ubiquitin-proteasome-dependent pathway that is conserved in both mouse MECs.
Ubiquitination of the C/EBP␦ Protein Is Localized Predominantly to the Nuclear Compartment-To determine whether C/EBP␦ protein degradation occurs in the nucleus and/or cytoplasm, nuclear and cytoplasmic protein was analyzed from G 0 growth-arrested HC11 MECs. HC11 MECs were growth-arrested by addition of GAM for 48 h and subsequently maintained in either growth arrest media alone (control), GAM plus a vehicle control (ϩMe 2 SO), or GAM plus the ubiquitination inhibitor (ϩMG-132). Nuclear and cytoplasmic protein fractions were isolated at the indicated times. C/EBP␦ protein was detected in the nuclear protein fraction but not the cytoplasmic protein fraction in both the G 0 growth-arrested control and Me 2 SO samples (Fig. 9, lanes 1-6 and lanes 7-10, respectively). In agreement with previous results (Fig. 8), C/EBP␦ protein content increased after MG-132 treatment (Fig. 9, lanes 11-14). Interestingly, the increase in C/EBP␦ protein is restricted to the nuclear compartment (Fig. 9, lanes 11 and 13).
Nuclear and cytoplasmic p27 protein levels were also monitored in MG-132-treated G 0 growth-arrested HC11 MECs. Similar to the previous study, p27 protein was detected in G 0 growth-arrested control and Me 2 SO samples (Fig. 9, lanes 1-6  and lanes 7-10, respectively). In contrast to C/EBP␦ protein subcellular localization, p27 protein was found in both the nuclear and cytoplasmic compartments. Detection of the p27 protein increased slightly in both compartments after MG-132 treatment (Fig. 9, lanes 11-14).
As a control for our nuclear and cytoplasmic protein fractionation, we monitored the subcellular localization of Bcl-x, which is known to be localized to the cytoplasmic compartment (64). As expected, the majority of the Bcl-x protein content was localized within the cytoplasmic compartment in both G 0 growth-arrested control and Me 2 SO samples (Fig. 9, lanes 1-6  and lanes 7-10, respectively). Importantly, upon MG-132 treatment, no change in Bcl-x localization was detected (Fig. 9, lanes [11][12][13][14]. Taken together, these results demonstrate that the C/EBP␦ protein is absent from the cytoplasmic compartment, which suggests a nuclear localized ubiquitin-mediated degradation pathway. DISCUSSION Although most cells in the adult animal exist in a G 0 growth arrest state, little is known about the regulation and function of genes expressed during G 0 (1-4). This study investigated the posttranscriptional and posttranslational regulation of C/EBP␦ in G 0 growth-arrested mouse MECs in vitro. Previous reports FIG. 7. C/EBP␦ protein exhibits a short half-life in G 0 growtharrested HC11 MECs. Western blot analysis was performed on 50 g of whole cell protein isolated from untreated and anisomycin-treated G 0 growth-arrested HC11 MECs at the indicated time points. Western blots were sequentially probed with C/EBP␦, p27, and actin antibodies. Actin was used as a loading control. A, lanes 1-6, protein from growtharrested MECs (GAMϪ); lanes 7-12, protein from growth-arrested MECs treated with anisomycin (GAMϩ). Results are representative of three independent experiments. B, summary of protein half-life data obtained from Western blot/anisomycin analysis. Signals were quantified, and the relative amount of each protein is expressed as a percentage of the 0-min control time, which was set at 100%. Graphs are plotted as % protein remaining versus time. Filled triangles, protein from anisomycin-treated MECs; filled circles, protein from non-treated MECs.
FIG. 8. C/EBP␦ protein undergoes ubiquitination in G 0 growth-arrested MECs. Western blot analysis was performed on 50 g of whole cell protein isolated from untreated and MG-132 or LLnLtreated G 0 growth-arrested mouse MECs at the indicated time points as described in Fig. 7. A, mouse HC11 MECs treated with MG-132. Lanes 1-4, protein from G 0 growth-arrested MECs (control); lanes 5-7, protein from G 0 growth-arrested MECs treated with vehicle (ϩDMSO); lanes 8 -10, protein from G 0 growth-arrested MECs treated with ubiquitination inhibitor (ϩMG-132). B, mouse HC11 MECs treated with LLnL. Lanes 1-4, protein from G 0 growth-arrested MECs (control); lanes 5-7, protein from G 0 growth-arrested MECs treated with vehicle (ϩDMSO); lanes 8 -10, protein from G 0 growth-arrested MECs treated with ubiquitination inhibitor (ϩLLnL). Results are representative of three independent experiments. from our laboratory (11)(12)(13) have shown that C/EBP␦ gene expression and DNA binding activity increase in G 0 growtharrested MECs. The G 0 -specific increase in C/EBP␦ gene expression is STAT3-dependent (10). In addition, overexpression of C/EBP␦ in MECs accelerated G 0 growth arrest and apoptosis in response to serum and growth factor withdrawal (13). In contrast, reducing C/EBP␦ levels by antisense RNA delayed MEC G 0 growth arrest and apoptosis after serum and growth factor withdrawal (13). In this report, we demonstrated that C/EBP␦ mRNA exhibits a novel short half-life during G 0 growth arrest in mouse MECs (t1 ⁄2 ϳ35 min). Interestingly, mRNAs encoding several important cell cycle control proteins, growth factors, lymphokines, cytokines, and proto-oncogenes also exhibit short half-lives (35). For example, the cytokine interleukin 6, which is important in the inflammatory response, has an mRNA half-life of ϳ20 min (65). We suggest that the short C/EBP␦ mRNA half-life in G 0 growth-arrested MECs allows the cells to respond rapidly to potential growth stimuli. Interestingly, the short half-life of C/EBP␦ mRNA observed in mouse MECs appears to be a property of mammary epithelial derived cell lines. For example, in G 0 growth-arrested NIH 3T3 cells, the C/EBP␦ mRNA half-life is ϳ2-3-fold longer. This suggests that tight regulation of the C/EBP␦ mRNA is important in the initiation and maintenance of G 0 growth arrest in mouse MECs. Similar to our studies in mouse HC11 MECs, the C/EBP␦ mRNA exhibited a relatively short half-life of ϳ40 min in G 0 growth-arrested human MCF-12A MECs (data not shown). This suggests a conservation of mRNA decay kinetics for C/EBP␦ in both mouse and human MEC systems.
Cell cycle re-entry (G 0 /G 1 transition) is associated with dramatic changes in gene expression. Transcription of growth arrest genes is known to decrease with cell cycle re-entry, although mRNAs encoding growth arrest-specific proteins could persist and may delay or interfere with cell cycle re-entry. The disposal of G 0 -specific mRNAs and proteins during MEC cell cycle re-entry is not well characterized. This study sought to determine whether or not the decay kinetics of C/EBP␦ mRNA were similar between G 0 growth arrest and the G 0 /G 1 transition. Results of transcriptional inhibitor/Northern blot studies upon cycle re-entry demonstrate that C/EBP␦ mRNA has a longer half-life during the G 0 /G 1 transition compared with G 0 growth-arrested HC11 MECs. In fact, both C/EBP␦ and gas-1 mRNAs exhibited stabilization during cell cycle reentry in response to transcriptional inhibitors in two mouse MEC lines, HC11 and COMMA D (later data not shown). Posttranscriptional control is known to play a major role in the regulation of gas family members in many cell types (61,66,67). Our results parallel previous work that shows an increase in gas-1 and gas-6 mRNA stability after cell cycle re-entry and treatment with actinomycin D of fibroblastic cell lines (61,66,67). Furthermore, actinomycin D treatment of Schwann cells during cell cycle re-entry stabilized the gas-3 mRNA (68). The difference in C/EBP␦ mRNA half-life in G 0 growth arrest and the G 0 /G 1 transition suggests that there is a specific mRNA degradation pathway for C/EBP␦ during the G 0 /G 1 transition that differs from G 0 growth arrest. In addition, increased stabilization of the C/EBP␦ mRNA suggests that the synthesis of a trans-acting factor(s) or RNA is required to degrade the C/EBP␦ mRNA upon cell cycle re-entry.
The mechanism underlying C/EBP␦ mRNA degradation is currently not known, although numerous studies have demonstrated that the length of the poly(A) tail is a major factor in the stability of eukaryotic mRNAs (i.e. a decrease in poly(A) tail length results in an decrease in mRNA stability) (34,35,39,63). In this report, analysis of poly(A) tail length by oligo/RNase H cleavage and RACE-PAT demonstrated that the C/EBP␦ mRNA has a short poly(A) tail of ϳ100 nucleotides. This is somewhat shorter than the average eukaryotic mRNA that contains a poly(A) tail of ϳ200 nucleotides (69).
Structural elements found within the 5Ј-UTR, coding region, and the 3Ј-UTR are known to be involved in regulating mRNA stability (34,35,39). For example, the 3Ј-UTR of many labile mRNAs, such as cytokine and oncoprotein mRNAs, contain multiple copies of A/U-rich elements (AREs) (34,35,39). These cis-acting elements interact with trans-acting factors to destabilize the mRNA. Analysis of ARE sequences from 12 transcription factor-encoding mRNAs that exhibit early G 1 instability classified two distinct groups of mRNAs: 1) mRNAs with 3Ј-UTRs that contain one or more copies of the well recognized "AUUUA" sequence and 2) mRNAs with 3Ј-UTRs that contain one or more copies of a "non-AUUUA" sequence (70). An example of a non-AUUUA mRNA is c-jun, which contains "U"-rich regions that confer G 1 instability (70). Interestingly, analysis of the C/EBP␦ 3Ј-UTR revealed a single AUUUA element and two U-rich regions (region 1, 18 uracils/32 nucleotides; region 2, 17 uracils/26 nucleotides). This indicates that the C/EBP␦ mRNA has characteristics of both AUUUA and non-AUUUA AREs. Mutational analysis is ongoing to characterize further the role of these instability elements in C/EBP␦ mRNA decay.
Analysis of another C/EBP family member, C/EBP␤, demonstrated similar mRNA decay kinetics as C/EBP␦ during mouse MEC G 0 growth arrest. Like C/EBP␦, C/EBP␤ mRNA displayed a short half-life of ϳ45 min. The results suggest a conserved mRNA decay pathway shared between C/EBP␦ and C/EBP␤ in G 0 growth-arrested mouse MECs. Although the homology between the C/EBP␦ and C/EBP␤ 3Ј-UTRs is ϳ30%, both 3Ј-UTR sequences contain multiple U-rich elements that may regulate mRNA degradation.
Because the C/EBP␦ mRNA was shown to have a short half-life in G 0 growth-arrested MECs, we hypothesized that the C/EBP␦ protein would exhibit a similar short biological half-life (39). A yeast genome-wide analysis has demonstrated that unstable mRNAs encode for unstable proteins (71). Examples include translation initiation factors, termination factors, and proteins of the mating pheromone signal transduction pathway (71). The results in this report established that the half-life of the C/EBP␦ protein is shorter (t1 ⁄2 ϳ120 min) than the tumor suppressor, p27 (t1 ⁄2 Ͼ150 min). The short half-life of the C/EBP␦ protein in G 0 growtharrested MECs suggests that C/EBP␦ function is tightly reg- FIG. 9. C/EBP␦ protein ubiquitination is localized to the nuclear compartment. Western blot analysis was performed on 25 g of nuclear and cytoplasmic protein isolated from untreated or MG-132treated G 0 growth-arrested mouse HC11 MECs at the indicated time points as described in Fig. 7. Lanes 1, 3, and 5, nuclear protein from G 0 growth-arrested MECs (control); lanes 2, 4, and 6, cytoplasmic protein from G 0 growth-arrested MECs (control); lanes 7 and 9, nuclear protein from G 0 growth-arrested MECs treated with vehicle (ϩDMSO); lanes 8 and 10, cytoplasmic protein from G 0 growth-arrested MECs treated with vehicle (ϩDMSO); lanes 11 and 13, nuclear protein from G 0 growth-arrested MECs treated with ubiquitination inhibitor (ϩMG-132); lanes 12 and 14, cytoplasmic protein from G 0 growth-arrested MECs treated with ubiquitination inhibitor (ϩMG-132). Results are representative of three independent experiments. ulated during MEC quiescence, which may allow MECs to respond rapidly to growth signals and re-enter the cell cycle when necessary.
The ubiquitin-proteasome pathway is a major selective decay mechanism of short-lived regulatory proteins (49). Cell cycle regulatory proteins that are degraded by the ubiquitin-proteasome pathway include the tumor suppressors, p21 (t1 ⁄2 ϳ30 min) (72), p53 (t1 ⁄2 ϳ20 min) (73), and p27 (t1 ⁄2 ϳ150 min) (52). The results in this report established that the C/EBP␦ protein is also degraded via the ubiquitin-proteasome pathway in growth-arrested MECs (t1 ⁄2 ϳ120 min). It has yet to be determined whether phosphorylation of the C/EBP␦ protein precedes ubiquitination, which has been observed in the regulation of p27 protein decay.
It has been known for sometime that mammalian proteasome complexes are localized throughout the cell including the nucleus, cytoplasm, and within the endoplasmic reticulum membrane network (55,56). Proteasomes localized within the nucleus have been shown to be responsible for the turnover of short lived proteins important for many critical cellular processes. Some proteins that undergo ubiquitination within the nuclear compartment include the large subunit of RNA polymerase II (74), the progesterone receptor (75), the Xenopus laevis kinase inhibitor, p27Xic1 (76), and the p53 protein (57). It is speculated that cells are able to rid themselves of nuclear proteins that are no longer necessary by ubiquitination within the nucleus (57). Results of this study demonstrate that C/EBP␦ protein ubiquitination is localized to the nucleus. We speculate that nuclear protein degradation provides a mechanistic explanation for the relatively short halflife of the C/EBP␦ protein during MEC G 0 growth arrest and allows for proper cell cycle progression during the G 0 /G 1 transition.
In summary, the data presented establish that the C/EBP␦ mRNA has a short half-life in G 0 growth-arrested MECs. The C/EBP␦ mRNA has a relatively short poly(A) tail (ϳ100 nucleotides) that does not vary in length during decay in G 0 growth arrest or the G 0 /G 1 transition. It is proposed that the C/EBP␦ mRNA is degraded by a mechanism involving endonucleolytic cleavage during G 0 growth arrest. Additionally, the C/EBP␦ protein has a relatively short half-life in G 0 growth-arrested MECs and is degraded by the ubiquitin-proteasome pathway within the nuclear compartment. This study suggests that despite the decrease in cellular activity during G 0 growth arrest, C/EBP␦ mRNA and protein are tightly regulated in MECs. We predict that this tight regulation allows G 0 growtharrested MECs to proliferate in response to growth stimuli. Studies investigating possible instability elements in the C/EBP␦ mRNA 3Ј-UTR and characterization of trans-acting factors important in C/EBP␦ mRNA degradation are currently underway.