C/EBP-β Regulates Endoplasmic Reticulum Stress–Triggered Cell Death in Mouse and Human Models

Endoplasmic reticulum (ER) stress elicits the unfolded protein response (UPR), initially aimed at coping with the stress, but triggering cell death upon further stress. ER stress induces the C/EBP-® variant Liver-enriched Activating Protein (LAP), followed by the dominant-negative variant, Liver Inhibitory Protein (LIP). However, the distinct role of LAP and LIP in ER stress is unknown. We found that the kinetics of the ER stress-induced expression of LIP overlapped with that of the cell death in mouse B16 melanoma cells. Furthermore, inducible over-expression of LIP augmented ER stress-triggered cell death whereas over-expression of LAP attenuated cell death. Similar results were obtained in human 293T cells. Limited vasculature in tumors triggers hypoxia, nutrient shortage and accumulation of toxic metabolites, all of which eliciting continuous ER stress. We found that LAP promoted and LIP inhibited B16 melanoma tumor progression without affecting angiogenesis or accelerating the cell cycle. Rather, LAP attenuated, whereas LIP augmented tumor ER stress. We therefore suggest that C/EBP-® regulates the transition from the protective to the death–promoting phase of the UPR. We further suggest that the over-expression of LAP observed in many solid tumors promotes tumor progression by attenuating ER stress–triggered tumor cell death.


Introduction
The endoplasmic reticulum is the site of post-translational processing (refolding, glycosylation and oxidative disulfide bond formation) of secreted and cytoplasmic membrane proteins. The ER is also the site of membrane biogenesis and helps maintain intracellular calcium ion and lipid homeostasis. Accumulation of client proteins or misfolded proteins, overload of free cholesterol, perturbations in calcium ion homeostasis, oxidative stress and xenobiotic toxins rapidly induce ER stress, which triggers an evolutionarily conserved cellular response, termed the Unfolded Protein Response (UPR) [1]. Initial UPR is aimed at coping with the stress by inducing ER chaperones and attenuating general protein translation. Persistent ER stress induces the expression of the C/EBP homologous protein (CHOP), which initiates the cell death machinery [2]. The mechanisms that regulate the transition from the protective phase to the death-promoting phase of the UPR are not known.
ER stress induces the expression of the transcription factor C/ EBP-b [3,4] but there are contradictory reports on the role of C/ EBP-b in ER stress. Embryonic fibroblasts from C/EBP-b-deficient mice are less sensitive to ER stress-mediated cell death [5,6]. In contrast, other cell types from these mice exhibit enhanced sensitivity to cell death triggered by growth factor deficiency, exposure to toxins or carcinogens and other stresses [7][8][9][10][11]. Other studies found that C/EBP-b is pro-apoptotic [5,[12][13][14][15][16]. Because the common C/EBP-b mRNA encodes both transcriptional activators (LAP* and LAP) and a dominant-negative, truncated form (LIP) [17], these opposing observations may be related to the LIP:LAP ratio. Furthermore, the LIP:LAP ratio is not constant, increasing in time, as the ER stress continues [18]. Therefore, further studies are required to determine the specific role of LAP and LIP in the cellular response to ER stress. ER stress has been associated with many diseases, where it leads to cell death [1,19]. Because neovascularization lags behind tumor cell proliferation, solid tumors are under continuous ER stress due to nutrient shortage, accumulation of cytotoxic metabolites and hypoxia. However, the cells adapt to this stress through the constitutive activation of the protective phase of the UPR (reviewed in: [2,[20][21][22]). So far, the mechanisms that mediate this adaptation are not fully characterized. The various C/EBP-b isoforms are over-expressed in many types of tumors [8,[23][24][25][26][27][28][29][30][31][32], but the specific role of LAP and LIP in tumor progression has not been studied. Recently, C/EBP-b was shown to support the survival of neurons under hypoxic conditions, which also trigger ER stress [33]; however, in these studies, both LAP and LIP were either over-expressed or were deleted, and therefore their individual function is not known.
To study the role of C/EBP-b LAP and LIP in ER stress and in tumor progression, we utilized B16 melanoma clones that inducibly over-express the dominant-negative LIP, thereby inhibiting endogenous LAP activity, as well as other clones that inducibly over-express LAP. Our findings suggest that C/EBP-b has a key role in regulating the transition from protective to death promoting UPR; LAP attenuates and LIP augments cell death, and LAP contributes to tumor progression by attenuating ER stress and subsequent cell death.

LAP Attenuates and LIP Augments ER Stress-Triggered Cell Death
To study the role of LAP to LIP ratio in ER stress-triggered cell death, we used the specific inducers of ER stress tunicamycin (Tm) and thapsigargin (Tg) and murine B16 melanoma clones that inducibly express either LAP or LIP [34]. Clone F10.9-3 constitutively expresses the Tet repressor and carries a vector with a bidirectional inducible promoter encoding wt LAP and GFP. Immunoblotting with an antibody directed against the common Cterminus of the various C/EBP b isoforms revealed that doxycycline induced only LAP in Clone F10.9-3 ( Figure S1). Similarly, Clone F10.9-4 expressed only LIP, which is the dominant-negative form of LAP ( Figure S1). Initially, we triggered ER stress in un-induced cells and compared the progression of cell death and the expression of endogenous LAP and LIP. Cell detachment and death started 4-6 h after addition of tunicamycin, as determined quantitatively by staining of the remaining attached cells with Crystal Violet ( Figures 1A & B, solid lines). Basal LAP and LIP were undetectable by immunoblotting. As previously reported [18], LAP expression was apparent already at 3 h, whereas LIP appeared at 6 h ( Figure 1C). Thus, initiation and further progression of cell death was temporally associated with endogenous LIP levels.
To further evaluate the role of LAP and LIP in determining the cell fate under ER stress, we studied the impact of their overexpression on the progression of cell death, as determined by Neutral Red staining. Over-expression of LIP in the F10.9-4 cells prior to addition of tunicamycin abolished the 4 h ''grace period'', leading to a statistically significant increase of cell death by 4 h (P,0.009, N = 4), as well as at all later time points (P,0.01, N = 4; Figure 1A, dashed line). No cell death was seen upon LIP expression in the absence of ER stress (data not shown). Overexpression of LAP had no significant effect in the first 4 h, but significantly attenuated cell death at all other time points (P,0.01-0.03; N = 4; Figure 1B, dashed line). These observations demonstrated that LAP attenuates and LIP augments ER stresstriggered cell death. Since both LAP and LIP were induced past the first 5 h, we evaluated their combined effect on cell viability by siRNA knockdown. Silencing of both LAP and LIP expression significantly augmented the tunicamycin-triggered death of F10.9-3 cells as determined by viability staining with Neutral Red (P,0.0001; N = 9; Figure 1D). Because LIP lacks a transactivation domain, it probably augmented cell death by inhibiting ER stress-induced LAP.
Microscopic observation of Crystal Violet-stained cultures further demonstrated that LIP over-expression augmented thapsigargin or tunicamycin-triggered cell death at 24 h ( Figure 2A). In contrast, over-expression of LAP attenuated thapsigargin or tunicamycin-triggered cell death, as determined at 24 and 48 h, respectively ( Figure 2B). The impact of C/EBP-b on tunicamycin-induced cell death was further determined quantitatively by viability staining with Neutral Red. Over-expression of LIP significantly augmented tunicamycin-induced cell death (P,0.0001, N = 9), whereas over-expression of LAP significantly attenuated cell death (P,0.0001, N = 9; Figure 2C, D, respectively). A similar increase in ER stress-triggered cell death was obtained in three other independent F10.9 LIP-expressing sub-clones following induction with doxycycline. Similarly, ER stress-triggered cell death was attenuated in four other independent F10.9 LAP-expressing sub-clones following doxycycline (data not shown). The enhancement of cell death by LIP was dosedependent and apparent upon induction with $100 ng/ml doxycycline ( Figure S2A). Similarly, the attenuation of cell death was dose-dependent and apparent upon induction of LAP with $100 ng/ml doxycycline ( Figure S2B).
Microscopic evaluation of Annexin V staining of tunicamycinchallenged cultures revealed lack of apoptotic cells within 24 h of treatment ( Figure S3A). Rather, tunicamycin triggered cell rounding and detachment from the plate. The detached cells were dead or dying, as determined by propidium iodide staining ( Figure S3B). Some of the tunicamycin-treated cells remained attached to the plate and although they were propidium iodide negative, they appeared to be dead, as they were condensed and rounded ( Figure S3C, left panels). Unlike control cells, the LAPover expressing F10.9-3 cells survived the tunicamycin treatment, they remained attached to the plate and appeared viable ( Figure  S3C, right panels). Furthermore, they propagated normally after removal of the ER stress triggers (data not shown). Based upon these observations we concluded that the ER stress-induced LAP attenuated cell death, whereas LIP augmented cell death, probably by inhibiting or re-directing the pro-survival activity of LAP.
To determine how general is the role of C/EBP-b in ER stress, we studied the impact of C/EBP-b over-expression and knockout in additional cell types. Over-expression of LAP in human 293T cells by transfection with an expression vector designed to encode LAP only (pcLAP, Figure S4A) rendered the cells significantly more resistant to tunycamycin-triggered cell death (P,0.0007, N = 4, Figure S4B). In contrast, over-expression of LIP in human 293T cells by transfection with an expression vector encoding only the LIP open reading frame (pcLIP, Figure S4A) rendered the cells significantly more sensitive to tunycamycin-triggered cell death (P,0.003, N = 4, Figure S4B).
Knockdown of C/EBP-b in human HeLa cells by siRNAs specific to C/EBP-b led to efficient silencing of its expression at 48 h, as determined by RT-PCR and immunoblotting with an antibody directed against a common C-terminal peptide of C/ EBP-b ( Figure S4C; left and right panels, respectively). The silencing of C/EBP-b increased the sensitivity of HeLa cells to ER stress-triggered cell death ( Figure S4D), leading to a statistically significant decrease in the cell counts (P,0.0003, N = 4, Figure  S4E). Over-expression of LAP in human HeLa cells by transfection with pcLAP rendered the cells significantly more resistant to tunicamycin-triggered cell death (P,0.001, N = 6, Figure S4F). In contrast with all other experiments, overexpression of LIP in HeLa cells by pcLIP also rendered the cells more resistant by 13% to tunicamycin-triggered cell death (P,0.002, N = 5). Yet, all other results suggest a general role for C/EBP-b in modulating ER stress-triggered cell death.

Attenuation of Cell Death by LAP Prevails Over Its Inhibitory Effect on the Cell Cycle
To study the mechanism by which LAP and LIP regulate the cell fate upon ER stress, we first determined the impact of LAP over-expression on the cell cycle. As previously reported [35], treatment of the B16 melanoma clone F10.9-3 with tunicamycin abolished DNA synthesis, leading to cell cycle arrest, as determined by flow cytometry of propidium iodide-stained cells ( Figure 3 compare the two left panels). Over-expression of LAP in F10.9-3 cells for 24 h prior to treatment with tunicamycin or  thapsigargin did not reverse the cell cycle arrest. In fact, LAP lowered the extent of cell proliferation in control cells, reducing the number of cells in S phase by about 25% and increasing the percentage of cells in the G0/G1 phase ( Figure 3, top panels). The observed LAP-mediated inhibition of cell proliferation was consistent with a previous study on the impact of LAP (LAP-2) on cell cycle progression [26]. Hence, we concluded that the prosurvival activity of LAP must have been the outcome of its effect on cell death rather than on the cell cycle. Furthermore, the attenuation of cell death by LAP under stress prevailed over its inhibitory effect on unstressed cell proliferation. We therefore focused on the impact of LAP and LIP on cell death mechanisms.

LAP Attenuated Necrotic Cell Death by Lowering CHOP Expression
Although tunicamycin triggered rounding, detachment and death of the F10.9 cells, we were unable to detect either annexin V-positive cells ( Figure S3A) or cleaved caspase 3 in the cell extracts ( Figure 4A). This result was also in accordance with the absence of sub G0 cell population, as revealed by flow cytometry ( Figure 3). Rather, we found that ER stress triggered the release of High-Mobility Group Box 1 (HMGB1) to the culture supernatant, indicating a mechanism involving necrotic cell death [36]. Furthermore, over-expression of LAP attenuated the release of HMGB1 ( Figure 4B). Therefore, we concluded that LAP attenuates the ER stress-triggered necrosis of F10.9 cells. The lack of annexin V staining minimizes the possibility that necrosis was initiated by an early stage of a caspase 3-independent apoptotic mechanism. To further study the mechanism by which LAP modulates ER stress, we measured the effect of its over-expression on the early and late ER stress markers Binding Protein (BiP; GRP78) and CHOP in tunicamycin-treated F10.9-3 cells. Immunoblot analysis revealed that over-expression of LAP did not lower the ER stress-induced chaperone BiP. In contrast, LAP reduced the level of the death executor CHOP ( Figure 4C). These results indicated that LAP does not affect the early protective phase of the UPR. Rather, LAP attenuates the late, death-promoting phase of the UPR.

LAP Promotes Tumor Progression by Attenuating Cell Death
To study the role of LAP in tumor progression, we inoculated mice subcutaneously either with F10.9-4 cells or with F10.9-3 cells and induced LIP or LAP expression, respectively, by including doxycycline (Doxy; 1 mg/ml) in the drinking water. Palpable tumors were detected in the mice after 8 days and differences in tumor mass were apparent by day 13. Therefore tumors in one group of mice were isolated and examined at day 8 and at day 13 in a second group of mice. The inhibition of endogenous LAP by over-expression of LIP led to a significant attenuation of tumor growth as determined at day 13 (P,0.005, N = 9). In contrast, over-expression of LAP significantly increased the tumor mass by fourfold over a 5 day period (P,0.001, N = 10; Figure 5).
To distinguish between the possible effects of LAP on tumor cell proliferation and cell death we stained tumor sections with the proliferation marker Ki-67 [37]. The percent of Ki67 (dividing) cells in tumors isolated at day 8 was not significantly affected by overexpression of LIP (P = 0.570, N = 4; Figure 6A, top panels). In accordance with our results in vitro (Figure 3), over-expression of LAP in the tumor resulted in a statistically significant reduction of Ki67-positive cells (P,0.03, N = 4; Figure 6A, bottom panels). Hence, the increase in the tumor mass upon induction of LAP must have been due to reduction in cell death, which apparently   Figure 6B, block arrows). Therefore, we focused our study on the impact of LAP and LIP on the extent of tumor cell death. Immunostaining of tumor sections isolated at day 8 revealed a rather low frequency (#3%) of cleaved caspase 3-positive cells in control tumors and in tumors expressing either LAP or LIP. Furthermore, expression of either LIP or LAP had no statistically significant effect on the fraction of the cleaved caspase 3-positive cells (P = 0.64, N = 3 and P = 0.52, N = 3 for LIP and LAP, respectively; Figure 7). Hence, similarly to the in vitro study (Figure 4), spontaneous B16 melanoma cell death in vivo was probably mediated by a caspase-3 independent mechanism.

LAP Attenuates Hypoxia-and Nutrient Deprivation-Triggered Cell Death
LAP could attenuate tumor cell death indirectly by inducing angiogenic factors or by modulating intrinsic tumor cell functions.
Staining of frozen tumor sections with antibodies to the endothelial cell marker CD31 did not reveal a significant effect of LAP on the extent of tumor vascularization ( Figure 8). Hence, we concluded that LAP attenuated mechanisms that trigger tumor cell death or rendered the cells more resistant to such mechanisms. We therefore studied the impact of LAP on B16 cell survival upon triggering ER stress by the tumor-related stressors hypoxia and nutrient deprivation. To this end, we employed extreme hypoxic conditions (1% oxygen) or media completely lacking nutrients (serum-free Earle's Balanced Salt Solution (EBSS)) in a series of in vitro studies. As expected, hypoxia or nutrient deprivation led to massive ER stress-triggered cell death (Figure 9). Over-expression of LAP significantly attenuated the extent of cell death (P,0.0001, N = 4 for hypoxia; P,0.0002, N = 4 for starvation; determined by measuring the relative Crystal Violet staining intensity; Figure 9). These results were in accordance with the impact of LAP on the specific inducers of ER stress tunicamycin and thapsigargin.

LAP Attenuates Tumor ER Stress
Hypoxia, nutrient deprivation and accumulation of toxic metabolites lead to tumor cell death by triggering ER stress. Therefore, we examined the impact of LAP inhibition or over-expression on ER stress in the tumors. To this end, we immunostained sections obtained from tumors of similar size at day 8 for two ER stress markers, TRIB3 and HERPUD [38,39]. Inhibition of endogenous LAP by overexpression of LIP significantly increased the expression of TRIB3 (P,0.0003, N = 3) and HERPUD (P,0.0005, N = 5), whereas overexpression of LAP reduced the expression of these markers (P,0.01, N = 3 for TRIB3; P,0.01, N = 4 for HERPUD) ( Figure 10). The reduced expression of the CHOP-induced death executor TRIB3 indicated that the tumor progression and the inhibition of ER stresstriggered cell death in vitro share a common mechanism, mediated by C/EBP-b.

Discussion
Tumor progression depends on the ability of its cells to adapt to ER stress due to lagging development of the neovasculature, which lead to nutrient limitation, impaired clearance of toxic metabolites and hypoxia. The combination of these constraints with the inherent hypermutability of cancer cells selects for cells resistant to ER stress-triggered cell death. Such resistance may be achieved by constitutive activation of the protective phase of the UPR [40,41]. Despite significant advances in recent years [22], many features of tumor cell resistance to ER stress are unknown. The present study demonstrates the critical role of resistance to ER stress in tumor progression and shows that C/EBP-b modulates the ER stresstriggered cell death. We suggest that selection of LAP overexpressing cells, as reported in many types of cancers, enables the growing tumor to thrive despite continuous exposure to an ER stress-inducing environment. Although further studies are re-quired to confirm the mode of action of C/EBP-b in the tumor, it is tempting to speculate that LAP or certain LAP-induced genes may serve as potential therapeutic targets for cancer and for other ER stress-mediated diseases.
The reported induction of C/EBP-b by ER stress [3] provided a general physiological framework for these studies on top of the tumor-related pathophysiology. Earlier studies with C/EBP-bdeficient cells and mice yielded conflicting results on the role of C/ EBP-b in cell survival and death. Over-expression of LAP has been described in several cancer cell lines and was implicated in oncogenic transformation [25,42]. Other studies found that C/ EBP-b was pro-apoptotic [43,44]. Such conflicting observations are not unexpected, as the single C/EBP-b mRNA is alternatively translated to both transcriptional activators (LAP* and LAP) and to their natural, dominant-negative inhibitor (LIP). Interactions of LAP and LIP with other transcription factors may also explain these conflicting results, as seen by us upon over-expression of LIP in HeLa cells. However, by using B16 melanoma clones that individually express either inducible LAP or LIP ( Figure S1), we were able to determine the specific impact of LAP and LIP on the cell fate under ER stress in vitro and in tumors in vivo. Induction of LIP results in suppression of endogenous LAP activity. Induction of LAP, on the other hand, mimics the over-expression seen in many tumors.
Our finding that LAP attenuated whereas LIP augmented ER stress-triggered cell death in vitro (Figures 1, 2 &S2) correlated with the observed fourfold increase in the size of LAP-expressing tumors during days 8-13 and the mirror image fourfold decrease in size of LIP-expressing tumors ( Figure 5). The protective effect exerted by LAP in cells exposed to nutrient deprivation or hypoxia further supports this correlation (Figure 9). The inhibition of the expression of death executor CHOP by LAP under ER stress ( Figure 4) provides one mechanism by which LAP attenuated ER stress-triggered cell death in vitro. Since the death executor TRIB3 is induced by CHOP, our finding that LAP attenuated and LIP augmented the expression of TRIB3 in the tumor ( Figure 10A) suggests that LAP promotes tumor progression by attenuating CHOP-mediated cell death. By performing knockdown and overexpression studies in two other human cell types ( Figure S4) we demonstrated the general role of LAP and LIP in modulating the resistance to ER stress-triggered cell death. A previous study demonstrated that C/EBP-b null mice (lacking both LAP and LIP) are completely refractory to generation of skin tumors upon exposure to a carcinogen [8]. Our findings further show that C/ EBP-b promotes the progression of an established tumor by attenuating cell death, implicating ER stress in the process.
Our findings suggest that changes in the LAP-LIP ratio control the transition from the early, protective UPR to its late, deathpromoting phase. Others [18] and we observed that ER stress initially triggered the expression of LAP, whereas extended ER stress led to expression of both LAP and LIP (Figure 1). C/EBP-b was found to be constitutively bound to the CHOP promoter at the common C/EBP-ATF4 site, inhibiting CHOP expression in rat cells [45]. Our findings that LAP attenuated the expression of CHOP and its downstream target TRIB3 (Figures 4C &10A) suggest that LAP rather than LIP is the constitutive transcriptional inhibitor of CHOP. Thus, LIP may relieve the transcription inhibition of LAP by forming a futile LAP-LIP heterodimer and thereby triggering cell death. Alternatively or additionally, such a LAP-LIP heterodimer may trigger cell death in a manner analogous to the proposed action of the LAP-CHOP heterodimer [5]. Regardless of the exact mechanism, this study demonstrates that LAP inhibits the death-promoting phase of the UPR whereas in most cases LIP promotes ER stress-triggered cell death. Establishing the impact of LAP and LIP on gene expression under ER stress will further clarify the mechanism by which LAP attenuates cell death.

Ethics Statement
The Weizmann Institute Animal Care and Use Committee approved all studies in mice, according to institutional guidelines and Israeli law.

Tumor Generation and Immunohistochemistry
C57/BL6 mice were inoculated subcutaneously with F10.9-3 or F10.9-4 cells (100 ml medium, containing 5610 5 cells mixed with 100 ml Matrigel). Doxycycline (1 mg/ml) was added to the drinking water for the induction of LAP and LIP. Mice (10 per group) were euthanized at day 8 or 13, and tumors were resected, weighed, photographed and either fixed in 10% formaldehyde or frozen before sections were prepared. The significance of differences in the tumor mass was determined by Student's t-test of independent samples. The paraffin sections were immunostained with antibodies to Ki-67, cleaved caspase 3, TRIB3 and HERPUD, followed by incubation with biotin-conjugated antibodies and Cy3-conjugated streptavidin. Cell nuclei in tumor slices and in cultures were stained with Hoechst 33258 (5 mg/ml; excitation at 365 nm and emission at 470 nm). For CD31 immunostaining, tumor tissues were embedded in OCT and were frozen at 280uC for cryosectioning. After fixation in 100% acetone (220uC, 20 min), slides were immunostained as described above.

Quantitative Measurment of the Immunostains
The percentage of proliferating cells was determined by counting ki67-positive nuclei and dividing by the total number of (Hoechst 33258-stained) nuclei. The fraction of cleaved caspase 3 cells was similarly determined. The staining intensity of TRIB3 and HERPUD was determined by the histogram function of The Photoshop program (red channel) and normalized to the staining intensity of the nuclei (blue channel) in the same field. The

Cell Staining, Fluorescence Microscopy and Flow Cytometry
To evaluate the extent of cell death, cultures were stained first with Hoechst 33258 (5 mg/ml, 15 min) washed and stained with propidium iodide (2.5 mg/ml, 15 min) and then visualized by fluorescence microscopy. To evaluate cell morphology, cultures were fixed and stained with 5% crystal violet in 66% methanol. Total cell counts were determined by counting manually, by using the ImageJ program (NIH) or by reading the crystal violet stained cultures in 96 well plates with an ELISA reader at 570 nm. Alternatively, staining intensity was determined in photographs by the histogram function of Photoshop. For quantitative measurement of viable cells, cultures in 96 well plates were incubated with Neutral Red (70 mg/L in 0.1 ml DMEM-10, 37uC, 30-60 min), washed 3X with PBS, re-suspended in lysis buffer (0.1 ml) and OD was measured with an ELISA reader at 540 nm. A calibration curve was prepared by seeding known amounts of cells onto 96well plates. After 6 h the cells were stained with Neutral Red and the OD was read as described above. For each experiment, statistical analyses were performed on the indicated number of independent experiments using Student's t-test of the independent samples. Each independent experiment consisted of 9 replicates. Apoptosis was evaluated by staining with the rh annexin-V/FITC kit (Bender MedSystems, Vienna, Austria). Cell cycle analysis was performed by collecting the cultured cells, fixing in 70% ethanol on ice, treatment with RNase A (100 mg/ml, 30 min, 37uC), staining with propidium iodide and flow cytometry.

Immunoblotting of Cellular Proteins
Cells were washed three times with ice-cold phosphate buffered saline (PBS). Cell pellets were resuspended in two packed cell volume of lysis buffer (20 mM Hepes, pH 7.9, 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM Phenyl Methyl Sulfonyl Flouride (PMSF), 0.5 mM dithiothreitol (DTT), 25% glycerol). The resuspended pellets were then frozen and thawed twice and kept on ice for 20 min. The lysates were centrifuged (14,000 rpm, 20 min.) and the supernatants containing the cellular proteins were collected, frozen in liquid nitrogen and stored at 280uC. Protein concentration was determined by a BCA Protein assay reagent kit (Pierce, Rockford USA) using bovine serum albumin as a standard. Protein samples were boiled in SDS-PAGE sample buffer containing 25 mM DTT, and the supernatants were resolved by gradient SDS-PAGE (7.5-15% acrylamide). Proteins were then transferred onto a nitrocellulose membrane, which was incubated with the indicated antibodies. Second antibody conjugates were visualized by the Super Signal Detection Kit (Pierce).

Construction of pcLAP and pcLIP Expression Vectors
To construct a mammalian expression vector that expresses only LAP (pcLAP), the complete human C/EBP-b ORF was inserted into pcDNA4 (Life Technologies, Carlsbad CA). Codon 1 (LAP*initiating ATG) was point-mutated to CTG using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA), the forward oligo 59 TG-TGCTGGAATTCCCTGCAACGCCTGGTGG and the reverse oligo 59 ACGGGAAGCCCGCCGCCAGGCCTGCGCCGCC-GC. Codon 199 (ATG) was then point-mutated as above with the forward oligo 59 GCGGCGGCGCAGGCCTGGCGGCGGGC-TTCCCGT and the reverse oligo 59 ACGGGAAGCCCGCCGC-CAGGCCTGCGCCGCCGC to prevent expression of LIP. The LIP ORF (codons 199-346) was inserted into pcDNA4 to generate pcLIP, which expresses only the human LIP isoform of C/EBP-b. Human 293T cells were transfected as in E and after 24 h were trypsin-digested and seeded in poly(D-lysine)-coated 96 well plates (20,000 cells/well). After additional 24 h the cells were either untreated (Ctrl) or challenged with tunicamycin (2 mg/ml). The plates were left for 24 h and the number of viable cells was determined by the Neutral Red assay. Percent survival represent the ratio of tunicamicyn-treated to diluent-treated cells. Data are mean6SD of four replicates. Similar transfection efficiencies were obtained with the three vectors and cell counts were very similar in the control cultures not treated with tunicamycin (Ctrl). (C) HeLa cells (2610 5 /well) were cultured in 6-well plates for 24 h and then transfected with a control siRNA or C/EBP-b-specific siRNA. After 24 h, RNA was isolated and analyzed by RT-PCR (left panels), and cell extracts were prepared and analyzed by immunoblotting with antibodies directed to the common Cterminus of C/EBP-b (right panels). The gels and blots are representatives of three replicate experiments. (D) HeLa cells were seeded and transfected as described in (C). After 36 h, diluent (Ctrl) or Tunicamycin (Tm) were added and after an additional 31 h the cultures were stained with Crystal Violet and photographed. The photographs are representatives of four replicate experiments. Bar = 0.5 mm. (E) Cell counts following the treatments described in (D). The data are average of counting four fields. (F) Human HeLa cells (200,000) were transfected with 1 mg of either control pcDNA4 vector, pcLAP vector (mutated to encode LAP only), or pcLIP vector (encoding LIP only). After 24 h the cells were seeded in 96 well plates (20,000 cells/well) and after additional 24 h diluent or tunicamycin (4 mg/ml) were added. After 48 h cell viability was determined by Neutral Red staining. Percent viability was determined by the ratio of tunicamycin to diluent-treated cells. Data are mean6SD of six replicates. Found at: doi:10.1371/journal.pone.0009516.s004 (2.03 MB TIF)