Calmodulin Controls Liver Proliferation via Interactions with C/EBPβ-LAP and C/EBPβ-LIP

Abstract

A truncated isoform of C/EBPβ, C/EBPβ-LIP, is required for liver proliferation. This isoform is expressed at high levels in proliferating liver and in liver tumors. However, high levels of C/EBPβ-LIP are also observed in non-proliferating livers during acute phase response (APR). In this paper we present mechanisms by which liver regulates activities of C/EBPβ-LIP. We found that calmodulin (CaM) inhibits the ability of C/EBPβ-LIP to promote liver proliferation during APR through direct interactions. This activity of CaM is under negative control of Ca²⁺, which is reduced in nuclei of livers with APR, whereas it is increased in nuclei of proliferating livers. A mutant CaM, which does not interact with C/EBPβ-LIP, also fails to inhibit the growth promotion activity of C/EBPβ-LIP. Down-regulation of CaM in livers of LPS-treated mice causes liver proliferation via activation of C/EBPβ-LIP. Overexpression of C/EBPβ-LIP above levels of CaM also initiates liver proliferation in LPS-treated mice. In addition, CaM regulates transcriptional activity of another isoform of C/EBPβ, C/EBPβ-LAP, and might control liver biology through the regulation of both isoforms of C/EBPβ. In searching for molecular mechanisms by which C/EBPβ-LIP promotes cell proliferation, we found that C/EBPβ-LIP releases E2F·Rb-dependent repression of cell cycle genes by a disruption of E2F1·Rb complexes and by a direct interaction with E2F-dependent promoters. CaM inhibits these growth promotion activities of C/EBPβ-LIP and, therefore, supports liver quiescence. Thus, our findings discover a new pathway of the regulation of liver proliferation that involves calcium-CaM signaling.

Introduction

Liver proliferation is controlled by a complex cooperation of several signal transduction pathways (1–4). A member of C/EBP³ family, C/EBPβ, is one of the key proteins that regulates liver proliferation (5, 6). Although C/EBPβ is an intronless gene, a single C/EBPβ mRNA produces three isoforms: full-length protein, liver activator protein (LAP), and liver inhibitor protein (LIP) (7, 8). C/EBPβ-LIP lacks activation domains and works as a dominant negative molecule by neutralizing activities of full-length C/EBP proteins (7, 9). C/EBPβ-LIP also interacts with other transcription factors and displays its activity through these interactions (10, 11). Partial hepatectomy increases expression of C/EBPβ mRNA at early times, resulting in elevation of all three isoforms of C/EBPβ (5, 12, 13). The alternative translation of C/EBPβ-LIP isoform requires activation of signal transduction pathways which increase initiation of translation from the third AUG codon. One of these pathways is controlled by RNA-binding protein CUGBP1. It has been shown that CUGBP1 directly binds the 5′ region of C/EBPβ mRNA and increases translation of C/EBPβ-LIP (12, 13). We have recently identified mechanisms by which CUGBP1 increases translation of C/EBPβ-LIP. These mechanisms include activation of CUGBP1 by cyclin D3-cdk4/6, which enhances interactions of CUGBP1 with translation initiation complex eIF2 and leads to recruitments of ribosomes to translate C/EBPβ mRNA (14, 15). The CUGBP1-eIF2-dependent activation of C/EBPβ-LIP has been observed in three biological situations; that is, in livers proliferating after partial hepatectomy (PH), in the liver during acute phase response (APR), and in livers of old mice (9, 13–15). Further studies have shown that elevation of C/EBPβ-LIP leads to the acceleration of proliferation of certain cultured cells (16). The elevation of C/EBPβ-LIP also leads to aggressive forms of breast cancer (17). Luedde et al. (18) have shown that C/EBPβ-LIP accelerates liver proliferation after PH by activation of PCNA and cyclin A.

Calmodulin (CaM) is a calcium-binding protein that is a common sensor for intracellular calcium signaling (19). CaM has no enzymatic activity and functions mainly as the translator of calcium signaling. There are several pathways by which CaM translates calcium signaling: that is, CaM-dependent phosphatases, CaM-dependent kinases, the transcription corepressors Cabin1, and histone deacetylase (19–21). In addition to these pathways, CaM directly interacts with transcription factors (calmodulin binding transcription activators) and might control growth and differentiation of several tissues (22). Several recent reports have suggested that CaM might regulate cell proliferation via different mechanisms. It has been shown that insulin-mediated stimulation of fibroblasts proliferation involves activation of calcium-CaM-CaM kinase II pathway (23). Choi et al. (24) have found that CaM regulates proliferation of vascular smooth muscle cells via interactions with cyclin E (26). Calmodulin also interacts with cyclin-dependent kinase inhibitor p21 and controls nuclear localization of p21 (27, 28).

C/EBPβ-LIP is increased in non-proliferating livers during APR (9, 13) and in livers of old mice, which is characterized by reduced proliferative capacities (14, 29, 30). Given the ability of C/EBPβ-LIP to accelerate liver proliferation after PH (18), we suggested that livers with APR have developed a mechanism that blocked growth promotion activities of C/EBPβ-LIP. In this paper we have examined this hypothesis using LPS-mediated activation of APR in mouse livers. We found that C/EBPβ-LIP promotes proliferation via interaction with and disruption of Rb·E2F complexes and that CaM blocks these growth promotion activities of C/EBPβ-LIP in livers of LPS-treated nice. The down-regulation of CaM in LPS-treated mice initiates liver proliferation by a release of growth promotion activities of C/EBPβ-LIP.

EXPERIMENTAL PROCEDURES

Antibodies and Reagents

Antibodies against C/EBPα (14AA), C/EBPβ (C-19), Rb (C-15), E2F1 (KH95), and E2F4 (C-20) were purchased from Santa Cruz Biotechnology. Antibodies to calmodulin and β-actin were from Millipore and Sigma, respectively. Antibodies to total Rb, to ph-Ser-612-Rb, and to ph-Ser-811-Rb were from Millipore. True-Blot secondary antibodies and IP beads were from Ebioscience. siRNAs to C/EBPβ and calmodulin were from Dharmacon. LPS and BrdUrd were from Sigma. The BrdUrd uptake assay kit and Fura-2 were from Invitrogen.

Generation of p3XFLAG-C/EBPβ-LIP-Δ(264–296) Mutant

Mutations were constructed by using the QuikChange^TM XL site-directed mutagenesis kit from Stratagene. A plasmid p3XFLAG-C/EBPβ-LIP was used as a template, and PCR amplification was performed in the presence of a forward primer, GCGGAGAACGAGCGGTCTAGAGGATCCCGG, and a reverse primer, CCGGGATCCTCTAGACCGCTCGTTCTCCGC. HEK293 cells were co-transfected with p3XFLAG-C/EBPβ-LIP-Δ(264–296) and pAd-Track-CaM. The presence of C/EBPβ-LIP-Δ(264–296) in CaM IP was examined by Western blotting using FLAG-horseradish peroxidase from Sigma.

Animals and Experiments with LPS, C/EBPβ, and CaM siRNAs

All research protocols for animal experiments were approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine (protocol #AN1349). 2–4-month-old mice were used for experiments described in this paper. The FLAG-C/EBPβ-LIP and calmodulin siRNA with or without C/EBPβ siRNA were delivered into mice by tail vein injection using the “in vivo-jetPEI transfection reagent” from PolyPlus Transfection. LPS was injected intraperitoneally on the next day. Because the proliferation of hepatocytes might be initiated by C/EBPβ-LIP in non-synchronized manner, BrdUrd was injected every day to provide higher BrdUrd incorporation. Physiological saline (0.9% NaCl) was used as a control. Liver samples were collected and kept at −80 °C. Data in the paper represent a summary of three experiments with 3–4 animals at each time point after LPS injection.

Cell Culture and Transient Transfection

HEK293 and Hep3B2 cell lines (from ATCC) were cultured in monolayers in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Hyclone) with 100 units/ml penicillin-streptomycin (Invitrogen). The cells were grown at 37 °C in a humidified incubator with 5% CO₂. Cells were transfected with pAdTrack-C/EBPβ-LIP, pAdTrack-CaM, or siRNA to CAM using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. 16 h after transfections, the cells were washed with phosphate-buffered saline and collected in phosphate-buffered saline. Protein extracts were isolated as described below.

Isolation of Proteins and Western Blotting Analysis

Cytoplasmic and nuclear extracts were isolated as described in our previous papers (30, 31). Briefly, cell pellets were resuspended in buffer A (25 mm Tris-HCl, pH 7.5, 50 mm KCl, 2 mm MgCl₂, 5 mm dithiothreitol, and inhibitors of phosphatases) and homogenized by shearing the suspension through an insulin syringe. After centrifugation, the supernatant (cytoplasm) was frozen and kept in the −80 °C freezer. The pellet (nuclei) was treated with buffer B containing 20 mm Tris-HCl, pH 7.5, 0.42 m NaCl, 1.5 mm MgCl₂, 25% sucrose, 1 mm dithiothreitol, and inhibitors of phosphatase. After 30 min incubation on ice, nuclei were spun down at 10,000 rpm for 10 min, and the supernatant (nuclear extract) was frozen and kept in −80 °C freezer. The procedure of isolation of cytoplasmic and nuclear extracts from liver was similar to that described above except that liver was homogenized in Buffer A by using a homogenizer. 50–100 μg of proteins were fractionated by SDS-PAGE using 4–20% gels, electrophoretically transferred to nitrocellulose membrane, and then probed with the indicated antibodies. The signals were detected by ECL Western blotting detection reagents (Amersham Biosciences). Protein loading was verified by the re-probe of the membranes with antibodies to β-actin and by Coomassie Stain.

Two-dimensional Examination of the Proteins That Differentially Interact with C/EBPβ-LIP

GST-C/EBPβ-LIP was incubated with 1000 μg of nuclear proteins isolated from liver at 8 h after PH and with proteins isolated from livers at 8 h after LPS administration. After intensive wash with phosphate-buffered saline, the proteins were eluted and separated by two-dimensional gel electrophoresis on strips pH 3–10 using Protean II (Bio-Rad) and after electrophoresis in 4–20% polyacrylamide gel. The gels were stained with silver, and the spots with differential intensities were cut out and sequenced. The differential interaction of Rb and CaM with C/EBPβ-LIP was further confirmed by a GST pulldown assay and co-IP approaches.

Co-IP Approach

HEK293 or Hep3B2 cells were transfected with pAd-Track-LIP or pAd-Track-CaM, respectively. 500 μg of total cell lysates were used for immunoprecipitation. The co-IP studies were performed as described in our publications (15, 30, 31). In the experiments with increasing concentrations of EDTA or/and Ca²⁺, cell lysates with CaM transfections were preincubated with 2 and 4 mm EDTA or with 2 and 4 mm CaCl₂ on ice for 1 h. Then, protein extracts from HEK293 cells transfected with C/EBPβ-LIP were added to these mixtures, and C/EBPβ-LIP was precipitated with antibodies to C/EBPβ. The rabbit IgG Trueblot beads were added to the mixture and incubated at 4 °C for overnight. The immunoprecipitates were resolved on a 4–20% SDS-PAGE followed by immunoblotting with the indicated antibodies. For the immunoprecipitation of C/EBPβ from liver nuclear extracts, 500 μg of nuclear extracts from mouse livers were incubated with 4 mm EDTA or CaCl₂ overnight. The rabbit IgG Trueblot beads and antibodies to C/EBPβ were added and incubated for 2 h. The presence of CaM in C/EBPβ IP was examined by Western blotting. All data in the manuscript represent the results of the 3–4 repeats.

EMSA Assay

Conditions for the EMSA (or gel shift) assay were described in our previously papers (30, 31). Briefly, the ³²P-labeled DHFR, E2F, and cdc2 oligomers were used as the probes. For examination of E2F·Rb complexes, nuclear extracts from HEK293 cells transfected with C/EBPβ-LIP or C/EBPβ-LIP and CaM were incubated with probes in the binding reactions containing salmon DNA as a competitor. Antibodies to E2F1, Rb, E2F4, and p130 were added before the probe addition. For examination of the direct interactions of C/EBPβ-LIP with E2F consensuses, poly(dI-dC) was used as nonspecific competitor. It has been shown that poly(dI-dC) inhibits E2F binding but does not affect interactions of C/EBP proteins with DNA (30, 31). DNA-protein complexes were separated by non-denaturing 6% polyacrylamide gel electrophoresis in 0.5× Tris borate EDTA buffer. After electrophoresis, the gel was dried and exposed to x-ray film. Each EMSA experiment was repeated 3–4 times with different transfections of the proteins.

Chromatin-IP Studies

A chromatin-IP assay was performed as described in our previous papers using a Chip-It kit (30, 31). Briefly, chromatin solutions were prepared from HEK293 cell transfected with FLAG-C/EBPβ-LIP, with empty FLAG vector, with CaM, or mutant calmodulin CaMΔ. E2F1, Rb, or FLAG were immunoprecipitated from the solutions. DNA was isolated and used for PCRs with primers covering E2F sites within the B-myb and DHFR promoter. The sequences of primers for the B-myb promoter were 5′-CCGGACTGACACGTGAGC-3′ (forward) and 5′-GTCAGCGTGTCAGCAGGTC-3′ (reverse). The sequences of DHFR primers were 5′-CTGCACAAATGGGGACGAG-3′ (forward) and 5′-CCATGTTCTGGGACACAGC-3′ (reverse).

BrdUrd Uptake

BrdUrd was injected every day after LPS treatments. Liver sections were fixed in 10% formalin. BrdUrd staining was performed using a BrdUrd uptake assay kit from Invitrogen according to the manufacturer's protocol. Examination of BrdUrd uptake was performed using three animals per each time point after LPS injection.

[Ca²⁺] Measurement

Ca²⁺ concentration was measured by using calcium-sensitive indicator Fura-2 (32, 33). Levels of Fura-2 fluorescence intensity of nuclear extracts containing 1 μm Fura-2 were measured at 340-nm and 380-nm excitations (excitation scan) and 510-nm emission wavelengths using an F-4500 fluorescence spectrophotometer.

Colony Formation Assay

The ability of C/EBPβ-LIP to promote cell proliferation was examined using colony formation assay. This approach has been previously development in our laboratory, and it is one of the best procedures for the analyses of cell proliferation/inhibition (31). This approach allows verification of the expression of transfected proteins in the experimental plates in the end of experiments. The HEK293 or Hep3B2 cells were plated at low density when each colony represents a single cell in the beginning of experiments. These cells were transfected with AdTrack-C/EBPβ-LIP plasmid, which expresses C/EBPβ-LIP and green fluorescent protein (GFP) from different promoters (see Fig. 6A). Therefore, each green cell also expresses C/EBPβ-LIP. Control cells were transfected with an empty AdTrack plasmid. The number of cells in each colony was examined at days 2, 3, and 4 after transfections. The colonies with two and more cells were considered as proliferating cells, whereas single cells are considered as growth-arrested cells. The rate of cell proliferation was calculated by counting the percentage of proliferating cells transfected with empty vector and with C/EBPβ-LIP plasmid. 150–200 cells were used for these calculations. After calculations, protein extracts were isolated from the experimental plates and used for Western blotting analyses as is shown in Fig. 6D. Data in the paper present results of three-four independent experiments.

FIGURE 6.

Ectopic expression of C/EBPβ-LIP promotes proliferation of HEK293 cells. A, shown is the structure of pAdTrack-C/EBPβ-LIP plasmid (upper) and a typical picture of the colony growth assay (see “Experimental Procedures” for more details). GFP, green fluorescent protein; CMV, cytomegalovirus. B, ectopic expression of C/EBPβ-LIP promotes proliferation of HEK293 cells. pAdTrack-C/EBPβ-LIP was transfected into HEK293 and Hep3B2; cells and the proliferation of green cells were examined by colony formation assay at day 4 after transfections. Bar graphs show the results as a summary of three independent experiments. The percentage of proliferating cells is shown. V, vector. C, overexpression of CaM in HEK293 cells reduces the ability of C/EBPβ-LIP to promote cell proliferation. HEK293 cells were transfected with C/EBPβ-LIP, CaM, and with C/EBPβ-LIP + CaM. The percentage of proliferating cells is shown. Bar graphs represent a summary of three independent experiments. D. CaM blocks growth promotion activities of C/EBPβ-LIP via direct interactions. The experiments were performed as described above. C/EBPβ-LIP was co-transfected with WT CaM and with CaMΔ. The bottom image shows Western blotting with Abs to C/EBPβ-LIP using protein extracts isolated from the experimental cells. L, LIP, C, CaM.

RESULTS

Identification of Proteins That Differentially Interact with C/EBPβ-LIP in Nuclear Extracts from Proliferating Livers and from Livers with APR

Previous studies have shown that C/EBPβ-LIP is increased in livers proliferating after PH and in livers during APR (9, 13, 15). We have first examined if the elevation of C/EBPβ-LIP is comparable in these settings. APR was initiated by injections of LPS as described (13). Western blotting confirmed previous findings and revealed a significant increase of LIP in these two biological situations (Fig. 1A). Calculations of the levels of C/EBPβ-LIP revealed that expression of C/EBPβ-LIP, as a ratio to β-actin, is 6–8-fold increased in livers after PH and in livers with APR (Fig. 1A, bar graphs).

FIGURE 1.

C/EBPβ-LIP differentially interacts with calmodulin and Rb in proliferating livers and in livers with APR. A, elevation of C/EBPβ-LIP in livers proliferating after PH and in livers of LPS-treated mice is shown. Nuclear extracts from livers of control mice (con), from mice after partial hepatectomy (PH8), and from mice after LPS treatments (LPS8) were examined by Western blotting with Abs to C/EBPβ. The filter was re-probed with β-actin. Bar graphs: protein levels of C/EBPβ-LIP were calculated as ratios to β-actin and as ratios to C/EBPβ-LAP. Summary of three independent experiments is shown. FL, full-length. B, two-dimensional (2D) gel electrophoresis of the GST-C/EBPβ-LIP pulldown samples is shown. GST-C/EBPβ-LIP was incubated with nuclear extracts isolated from liver at 8 h after PH and with nuclear extracts isolated from livers of LPS treated mice. The proteins were separated by two-dimensional gel electrophoresis and stained with silver. Locations of proteins preferentially interacting with C/EBPβ-LIP in proliferating livers are shown by red circles and by red arrows. Positions of proteins preferentially interacting with C/EBPβ-LIP in livers of LPS-treated mice are shown by blue arrows and circles. C, shown is a list of proteins differentially interacting with C/EBPβ-LIP in proliferating livers and in livers of LPS-treated mice.

Because the elevation of C/EBPβ-LIP after PH promotes liver proliferation but identical levels of C/EBPβ-LIP after LPS injection do not initiate liver proliferation, we suggested that the biological activities of C/EBPβ-LIP are regulated in these settings by differentially interacting proteins. To examine this hypothesis, we have isolated and sequenced proteins that differentially bind to C/EBPβ-LIP in livers proliferating after PH and in livers with APR. For this goal GST-C/EBPβ-LIP was incubated with protein extracts of proliferating livers (8 h after PH) and with protein extracts isolated from livers with APR (8 h after LPS injection). We have observed several spots with different intensities on two-dimensional gels (Fig. 1B). The spots with significant differences in intensity were sequenced in the Protein Chemistry Core Laboratory at Baylor College of Medicine. We found that Rb, Grp78, and Hsp70 strongly interact with GST-C/EBPβ-LIP in nuclear extracts from proliferating livers, whereas mouse ATP synthase, cytochrome b₅, major mouse urinary protein, and CaM strongly interact with GST-C/EBPβ-LIP in extracts from livers with APR (Fig. 1C). We have further focused our studies on the interactions of C/EBPβ-LIP with Rb and CaM as Rb has been shown to be a mediator of C/EBPβ-LIP activities (34–36) and because CaM regulates activities of transcription factors (22).

The Reduction of Ca²⁺ in Nuclei of LPS-treated Mice Increases the Interaction of CaM with C/EBPβ-LIP and with C/EBPβ-LAP

Because the interaction of CaM with GST-C/EBPβ-LIP is increased without alterations in protein levels of CaM and Rb (see Figs. 2, B and C), we have looked for the mediators that might regulate this interaction. It has been shown that the interaction of CaM with proteins depends on the concentration of Ca²⁺ (19). Therefore, we examined the interactions of C/EBPβ-LIP and CaM under different concentrations of Ca²⁺. C/EBPβ-LIP and CaM were overexpressed in cultured cells and incubated with each other in buffers containing increasing concentrations of Ca²⁺ and EDTA (2 and 4 mm each). Fig. 2A shows that the elimination of Ca²⁺ by EDTA significantly increases interactions of CaM with C/EBPβ-LIP, whereas the increase of concentration of Ca²⁺ significantly reduces this interaction. Calculations of a ratio of CaM to C/EBPβ-LIP within C/EBPβ IPs show a 2–3-fold increase of the interaction by EDTA and around a 4-fold reduction of interaction by high concentrations of Ca²⁺. We next determined concentrations of Ca²⁺ in nuclear extracts isolated from control livers and from livers at 4 and 8 h after injections of LPS and found that the concentration of Ca²⁺ is reduced in nuclei of LPS-treated mice (Fig. 2B).

FIGURE 2.

Ca²⁺ regulates interactions of C/EBPβ-LIP and CaM. A, removal of Ca²⁺ by EDTA enhances interactions of C/EBPβ-LIP and CaM, whereas the addition of Ca²⁺ reduces interactions of these proteins. HEK293 cells were transfected with pAd-Track-LIP or pAd-Track-CaM plasmids. The protein extracts were isolated and incubated with each other in buffers containing different concentrations of Ca²⁺ (2 and 4 mm) and EDTA (2 and 4 mm). C/EBPβ was immunoprecipitated, and levels of CaM and C/EBPβ-LIP were determined in C/EBPβ IPs by Western blotting. Bar graphs show the amounts of CaM in C/EBPβ IPs calculated as ratios to C/EBPβ-LIP. B, the decrease of Ca²⁺ in nuclear extracts of LPS-treated mice increases interactions of C/EBPβ-LIP and CaM. The bar graph (upper) shows concentrations of Ca²⁺ as fluorescence intensity calculated as a ratio to protein content. Western blotting (Input) of nuclear extracts with antibodies against CaM shows the level of CaM used for Co-IP. C/EBPβ-IP, C/EBPβ was immunoprecipitated from nuclear extracts, and the IPs were examined by Western blotting with anti-CaM and C/EBPβ. Because heavy chains of IgGs migrate in the positions close to the full-length (FL) and C/EBPβ-LAP isoforms, the section with C/EBPβ-LIP is shown. Bottom image. nuclear extracts were preincubated with 4 mm Ca²⁺, and C/EBPβ-LIP was immunoprecipitated. Nuclear extract isolated from liver at 8 h after LPS treatment (first lane) was used as a positive control without adding Ca²⁺. C, CaM was immunoprecipitated from nuclear extracts, and the IPs were examined by Western blotting using C/EBPβ and CaM antibodies. The upper panel shows loading of the C/EBPβ proteins (input). IgG shows the loading control. D, C/EBPβ was immunoprecipitated from nuclear extracts, and the IPs were examined by Western blotting with antibodies to Rb. The upper panel shows Western blotting of nuclear extracts with Abs to Rb (input). CRM, cross-reactive molecule. The filter was stained with Coomassie. IgG, heavy chains and light chains are shown as the loading controls. E, elevation of Ca²⁺ in nuclei during liver regeneration inhibits interactions of C/EBPβ-LIP and CaM. The bar graph shows concentrations of Ca²⁺ at different time points after PH as fluorescence intensity calculated as a ratio to protein content. Western blotting was performed with nuclear extracts isolated from quiescent livers (0 h) and from livers at 4, 8, 36, and 48 h after partial hepatectomy. F, the increase of Ca²⁺ in nuclei of regenerating livers inhibits interactions of C/EBPβ-LIP and CaM. C/EBPβ was immunoprecipitated from nuclear extracts with or without preincubation with EDTA. The IPs were examined by Western blotting using CaM antibodies. The membrane was stripped and re-probed with C/EBPβ antibodies. Signals of C/EBPβ-LIP are shown.

We next determined if endogenous CaM and C/EBPβ-LIP interact with each other in the liver. The studies of endogenous C/EBPβ-LIP in the liver are complicated by the fact that both C/EBPβ-LAP and C/EBPβ-LIP are expressed from the single mRNA and by the fact that there are no available antibodies to C/EBPβ-LIP to distinguish this isoform from C/EBPβ-LAP. Therefore, our experiments with antibodies to C/EBPβ (C19, Santa Cruz) show interactions of CaM with both isoforms of C/EBPβ. We first immunoprecipitated C/EBPβ from nuclear extracts of livers harvested at 0, 4, and 8 h after LPS injection, and CaM was examined in these IPs. We found that the interactions of the endogenous C/EBPβ with CaM are increased in LPS-treated mice at 4 and 8 h. To determine whether Ca²⁺ plays a causal role in the interactions of C/EBPβ-LIP and CaM, we have preincubated nuclear extracts with Ca²⁺ and found that the restoration of Ca²⁺ concentration in nuclear extracts inhibits interactions of CaM with C/EBPβ-LIP (Fig. 2B). Immunoprecipitation of CaM and Western blotting with Abs to C/EBPβ confirmed that the association of CaM with C/EBPβ-LIP is increased in LPS-treated mice (Fig. 2C). These studies have shown that both C/EBPβ-LIP and C/EBPβ-LAP isoform bind to CaM.

Our GST pulldown experiments demonstrated that the interactions of C/EBPβ-LIP with Rb are reduced in nuclear extracts from LPS-treated mice compared with interactions in regenerating livers (Fig. 1). Therefore, we next examined in interactions of endogenous C/EBPβ-LIP with Rb are altered in livers of LPS-treated mice. For this goal, C/EBPβ was immunoprecipitated, and Rb was examined in these IPs. Fig. 2D shows that the association of C/EBPβ-LIP and C/EBPβ-LAP with Rb is reduced at 4 and 8 h after LPS treatment. Taken together, examination of early steps of APR showed that the interactions of C/EBPβ-LIP with Rb are reduced, whereas the interactions of C/EBPβ-LIP with CaM are increased in the liver.

The Increase of Ca²⁺ in Nuclei of Proliferating Livers Leads to the Inhibition of Interactions of C/EBPβ-LIP with CaM

We next determined if endogenous C/EBPβ-LIP interacts with CaM in livers proliferating after partial hepatectomy, PH. Examination of Ca²⁺ in nuclei of regenerating livers showed an increase of Ca²⁺ at 4–48 h after PH (Fig. 2E, bar graphs). Levels of CaM after PH were not changed significantly in nuclei of livers at 4–48 h after PH, whereas levels of C/EBPβ-LIP and C/EBPβ-LAP were increased (Fig. 2E). We next precipitated C/EBPβ-LIP from nuclear extracts of regenerating livers and examined CaM and C/EBPβ-LIP in these IPs. We found that CaM is not detectable in C/EBPβ IPs from regenerating livers (Fig. 2F, upper image). The elevation of Ca²⁺ seems to be a major cause of the lack of interactions as the elimination of Ca²⁺ from nuclear extracts by EDTA increases interactions of C/EBPβ-LIP and CaM at 8–48 h after PH. Thus, these studies demonstrated that the increase of Ca²⁺ in nuclei of regenerating livers prevents the interaction of CaM with C/EBPβ-LIP.

Identification of Regions of CaM and C/EBPβ-LIP, Which Are Required for the Interactions

For the investigations of the effects of CaM on activities of C/EBPβ-LIP, we have generated a mutant CaMΔ that does not interact with C/EBPβ-LIP. CaM consists of two domains linked by a spacer, which is critical for the “active” conformation of CaM (19). Therefore, we have generated the mutant CaM (linked to V5 tag) with the deletion of six amino acids within the spacer region (Fig. 3A). Co-IP studies have shown that the mutant CaMΔ does not bind to C/EBPβ-LIP; whereas WT CaM interacts with C/EBPβ-LIP (Fig. 3B).

FIGURE 3.

Identification of regions of CaM and C/EBPβ-LIP that are required for interactions. A, the structure of the mutant CaM is shown. B, the mutant CaM Δ76–81 (CaMΔ) does not interact with C/EBPβ-LIP. C/EBPβ-LIP and WT/Mut V5-CaM were co-transfected into Hep3B2 cells. C/EBPβ-LIP was immunoprecipitated, and the IPs were probed with Abs to V5. C, shown is generation of C/EBPβ-LIP mutant, which does not interact with CaM. A diagram shows the structure of the mutant C/EBPβ-LIP. The deleted 32 amino acids (aa) are shown in red. D, C/EBPβ-LIP mutant does not interact with CaM. FLAG-tagged WT C/EBPβ-LIP and C/EBPβ-LIP-Δ264–296 constructs were co-transfected with CaM. CaM was immunoprecipitated, and C/EBPβ-LIP was determined by Western blotting with Abs to FLAG-tag. E, CaM inhibits transcriptional activity of C/EBPβ-LAP isoform. C/EBPβ-LAP was co-transfected with C3-luc reporter (rep) plasmid and with increasing amounts of WT CaM or CaMΔ. The activity of C3-luc promoter was calculated as a ratio to protein. Transcription factor Smad2 and the Smad2 reporter construct were used as negative controls.

To determine regions of C/EBPβ-LIP that interact with CaM, we generated several FLAG-linked C/EBPβ-LIP-truncated molecules and examined these mutants for the interactions with CaM. However, the small deletions were not efficient to block these interactions. We found that deletion of a long region from the C terminus completely blocks interactions of C/EBPβ-LIP with CaM (Fig. 3, C and D). The deleted region is 32 amino acids long and includes three leucine residues from the zipper region of C/EBPβ-LIP. The identification of the leucine zipper of C/EBPβ-LIP as the CaM-interacting region raised a possibility that CaM might also affect transcriptional activity of the C/EBPβ-LAP isoform, which contains the zipper region and which also interacts with CaM (see Fig. 2). Therefore, we have examined if CaM regulates transcriptional activity of C/EBPβ-LAP. For this goal, C/EBPβ-LAP was co-transfected with C/EBP-dependent C3-luc reporter plasmid and with increasing amounts of a plasmid coding for CaM. Fig. 3E shows that CaM inhibits transcriptional activity of C/EBPβ-LAP. This inhibition is specific as CaM does not affect the translational activity of SMAD2 and because CaMΔ mutant does not inhibit C/EBPβ-LAP. These studies suggested that CaM might regulate activities of both isoforms of C/EBPβ, LAP and LIP. Therefore, we next examined if interactions of CaM with C/EBPβ-LAP and C/EBPβ-LIP control activities of these proteins in the liver.

Down-regulation of CaM in Livers of LPS-treated Mice Causes Liver Proliferation via Activation of C/EBPβ-LIP and C/EBPβ-LAP

During liver regeneration after PH, C/EBPβ-LIP, and C/EBPβ-LAP are elevated (Fig. 2E) and are required for proper liver proliferation (5, 6). However, a similar elevation of C/EBPβ isoforms after LPS treatments is not sufficient to initiate liver proliferation (see Fig. 1 and Ref. 13). Because CaM interacts with C/EBPβ-LIP and C/EBPβ-LAP in LPS-treated mice, we suggested that the lack of liver proliferation in LPS-treated mice might be due to the increased interactions of C/EBPβ with CaM. To test this suggestion, we have inhibited CaM by siRNA in the liver of mice treated with LPS and examined liver proliferation after LPS administration. Because previous studies showed expression of C/EBP proteins within 24 h after LPS injection, we initially characterized expression of C/EBPβ and C/EBPα at 48 h after LPS injection. This time point was selected because further studies showed liver proliferation at 48 h after LPS injections and inhibition of CaM. C/EBPα was also included in these studies because it is a strong inhibitor of liver proliferation and it is reduced after LPS treatments similar to reduction after PH (5, 13). We found that levels of C/EBPβ isoforms are elevated at 24 h and slightly reduced at 48 h after LPS injections (Fig. 4, A and B). C/EBPα levels are reduced at 24 h but are returned to normal levels at 48 h.

FIGURE 4.

Inhibition of CaM in livers of LPS-treated mice causes liver proliferation via activation of C/EBPβ. A and B, expression of C/EBPα and C/EBPβ in LPS-treated mice at 4, 8, 24, and 48 h is shown. The expression of C/EBPα and C/EBPβ was examined by Western blotting (A), and the levels of these proteins were calculated as ratios to β-actin (B, bar graphs). C, inhibition of CaM and C/EBPβ by siRNAs in the livers of LPS-treated mice is shown. Nuclear extracts from livers of LPS-treated mice were isolated at different time points after injections of LPS (shown on the top) from control animals and from animals co-injected with siRNAs to CaM and to C/EBPβ. These extracts were examined by Western blotting with antibodies to CaM and C/EBPβ. Light and dark exposures for C/EBPβ are shown. The filters were re-probes with Abs to β-actin. Positions of C/EBPβ isoforms LAP and LIP as well as cross-reactive molecule (CRM) are shown. D, liver proliferation is increased in LPS-treated animals with reduced levels of CaM. Shown is BrdUrd staining of livers from LPS-treated mice after injections of vehicle (LPS), siCaM (LPS+siCaM), and siCaM+siC/EBPβ (LPS+siCaM+siC/EBPβ). E, bar graphs show the percentage of BrdUrd-positive hepatocytes.

We next inhibited expression CaM and examined liver proliferation after treatments of the mice with LPS. Liver proliferation was examined by measuring BrdUrd uptake. Fig. 4C shows that siRNA to CaM inhibits expression of CaM to 50–60%. Under these conditions livers of LPS-treated mice start proliferation and incorporate BrdUrd at 48 and 72 h after LPS treatments (Fig. 4, D and E). We found that up to 8–10% of hepatocytes proliferate in LPS-treated and siCaM-injected mice. To determine whether this proliferation is mediated by C/EBPβ isoforms, we simultaneously inhibited CaM and C/EBPβ by specific siRNAs. As one can see in Fig. 4C, siRNA to C/EBPβ almost completely inhibits expression of C/EBPβ-LIP and C/EBPβ-LAP isoforms. We found that the inhibition of C/EBPβ abolishes proliferation of the liver that was initiated by siRNA to CaM and LPS injections (Fig. 4, D and E). Thus, these studies showed that livers of LPS-treated mice do not proliferate due to CaM-mediated inhibition of C/EBPβ-LIP and perhaps C/EBPβ-LAP and that down-regulation of CaM is sufficient for the initiation of C/EBPβ-dependent liver proliferation.

Overexpression of C/EBPβ-LIP above CaM Levels Initiates Liver Proliferation in LPS-treated Mice

Because the endogenous C/EBPβ-LAP and C/EBPβ-LIP are expressed from the same mRNA, it is not possible to inhibit C/EBPβ-LIP without inhibition of C/EBPβ-LAP (see Fig. 4C). Therefore, the data with siRNA to C/EBPβ did not distinguish the contribution of C/EBPβ-LIP and C/EBPβ-LAP in the promotion of liver proliferation in mice with reduced levels of CaM. To examine the role of C/EBPβ-LIP in liver proliferation after LPS injection, we overexpressed FLAG-C/EBPβ-LIP above CaM levels and examined if this might promote liver proliferation under conditions of APR. FLAG-C/EBPβ-LIP was injected in mice as described under “Experimental Procedures,” and animals were treated with LPS for 24, 48, and 72 h. It is known that C/EBPβ-LIP affects the C/EBPα and C/EBPβ-LAP activities through dimerization as well as via repression of the C/EBPα promoter (6, 13). Therefore, we initially calculated levels of C/EBPβ-LAP and C/EBPα in livers of mice injected with FLAG-C/EBPβ-LIP. Western blotting analyses showed that FLAG-C/EBPβ-LIP is expressed at high levels at all examined time points. Calculations of ratios of total LIP (FLAG-C/EBPβ-LIP + endogenous C/EBPβ-LIP) to signals of CaM revealed that these ratios are 6–8-fold higher in FLAG-C/EBPβ-LIP injected livers at 48 and 72 h after LPS treatments (Fig. 5A). The ratios of total C/EBPβ-LIP to C/EBPβ-LAP are also significantly increased and reached 1.5 and 2.0. These studies showed that the injections of FLAG-C/EBPβ-LIP resulted in the significant excess of C/EBPβ-LIP above levels of CaM and C/EBPβ-LAP. It has been previously shown that C/EBPβ-LIP directly binds to C/EBPα promoter and represses expression of C/EBPα in the liver (13). Therefore, we have examined if the injected FLAG-C/EBPβ-LIP is biologically active by measuring levels of C/EBPα. We found that FLAG-C/EBPβ-LIP inhibits expression of C/EBPα (Fig. 5A). Thus, these studies revealed that injection of FLAG-C/EBPβ-LIP resulted in elevation of protein levels of C/EBPβ-LIP to levels that are above CaM and that FLAG-C/EBPβ-LIP is biologically active and inhibits expression of C/EBPα. Examination of BrdUrd uptake in these mice showed that up to 5–6% of hepatocyte are BrdUrd-positive in LPS-treated and FLAG-C/EBPβ-LIP-injected mice, whereas no BrdUrd uptake is detected in LPS-treated mice that were transfected with an empty vector (Fig. 5, B and C). Thus, these studies show that overexpression of C/EBPβ-LIP above levels of CaM in LPS treated mice is sufficient to initiate liver proliferation.

FIGURE 5.

Overexpression of FLAG-C/EBPβ-LIP in LPS-treated mice causes proliferation of the liver. A, expression of FLAG-C/EBPβ-LIP, endogenous CaM, C/EBPβ, and C/EBPα was examined by Western blotting. The ratios of total C/EBPβ-LIP to CaM and to C/EBPβ-LAP and the levels of C/EBPα are shown in bar graph pictures. V, vector. B and C, proliferation of the livers was examined by BrdUrd uptake. Bar graphs show a summary of three independent experiments.

C/EBPβ-LIP Promotes Proliferation of HEK293 Cells

Examination of the mechanisms of C/EBPβ-LIP-mediated cell proliferation in the liver is complicated because it is not possible to distinguish C/EBPβ-LIP from C/EBPβ-LAP. Therefore, we performed a search for tissue culture systems in which C/EBPβ-LIP accelerates proliferation. For these studies, C/EBPβ-LIP was cloned into pAdTrack vector, which also expresses green fluorescent protein from an independent cytomegalovirus promoter; therefore, each transfected cell expresses C/EBPβ-LIP and green fluorescent protein (see Fig. 6A). To screen cultured cells for the ability of C/EBPβ-LIP to promote cell proliferation, we used colony formation assay. In this assay cells are transfected at very low density so that there are only single cells on the plates in the beginning of experiments. In several days the proliferating cells form cell clusters containing two and more cells, whereas inhibited cells stay as single cells. A typical picture of green colonies is shown in Fig. 6A. We found that HEK293 and Hep3B2 cells have quite different responses to the ectopic expression of C/EBPβ-LIP. As one can see in Fig. 6B, ectopic expression of C/EBPβ-LIP in HEK293 cells significantly increases the amounts of proliferating cells at day 4; however, the significant portion of Hep3B2 cells transfected with C/EBPβ-LIP does not proliferate and is rather inhibited by C/EBPβ-LIP. The comparison of proliferating cells transfected with an empty vector and with C/EBPβ-LIP showed that C/EBPβ-LIP accelerates proliferation of HEK293 cells (Fig. 6C). Given the established cell line for the growth promotion activities of C/EBPβ-LIP, we asked if CaM might block the activity of C/EBPβ-LIP. We have co-transfected CaM with C/EBPβ-LIP and found that CaM abolishes the growth promotion activity of C/EBPβ-LIP (Fig. 6D). This effect of CaM is mediated through direct interactions with C/EBPβ-LIP as the CaMΔ mutant is not able to reduce growth promotion activities of C/EBPβ-LIP.

HEK293 Cells Contain Abundant Rb·E2F1 Complex

Because C/EBPβ-LIP preferentially interacts with Rb in proliferating livers (Fig. 1), we suggested that the different biological activities of C/EBPβ-LIP in HEK293 and Hep3B2 cells might be associated with differences in expression of E2F·Rb complexes. We found that the amounts of Rb are higher in HEK293 cells than in Hep3B2 cells and that electrophoretic mobility of Rb differs in these cells (Fig. 7A), suggesting that Rb might be differentially phosphorylated. Although the phosphorylation of Rb leads to dissociation of E2F·Rb complexes (37, 38), Inoue et al. (39) found that phosphorylation of Rb at Ser-612 does not block its interactions with E2F1 and that Ser-612-ph-Rb·E2F1 complexes repress the E2F-dependent promoters. Western blotting analysis with phospho-specific Abs of Rb showed that the Rb-Ser-612-ph isoform is abundant in HEK293 cells, but it is not detectable in Hep3B2 cells. Examination of another isoform of Rb, Ser-811-ph, showed that the amounts of this isoforms are identical in the tested cells. We also found that expression of E2F1 and E2F4 is higher in HEK293 cells compared with Hep3B2 cells. Examination of CaM showed approximately identical levels of CaM in HEK293 and Hep3B2 cells.

FIGURE 7.

E2F·Rb family complexes are different in HEK293 and Hep3B2 cells. A, expression of Rb and E2F proteins in HEK293 and Hep3B2 cells is shown. Nuclear extracts of HEK293 and Hep3B2 cells were examined by Western blotting with antibodies shown on the left. A Coomassie stain of a parallel gel shows loading of the protein. B, E2F1·Rb complexes are abundant in HEK293 cells. EMSA was performed with nuclear extracts (NE) from HEK293 cells using the DHFR probe covering the E2F site within the DHFR promoter (31). Antibodies to Rb, p130, E2F1, and E2F4 (shown on the top) were incorporated in the binding reactions. Positions of Rb·E2F1 complexes, nonspecific band (NS), and free probe are shown on the right. C, shown is the Ser-612-ph isoform of Rb forms complexes with E2F1 in HEK293 cells. The EMSA was performed as described above. Antibodies to total Rb, ph-Ser-612, and ph-Ser-811 were included in the binding reactions. D, Hep3B2 cells contain abundant E2F-p130 complexes. EMSA was performed with nuclear extracts from Hep3B2 cells as described above. Antibodies to E2F and Rb family proteins were incorporated in the binding reactions. Note: antibodies to p130 neutralize the complexes. Positions of p130-E2F complexes, E2F4, and free probe are shown on the left.

We next examined the compositions of E2F·Rb complexes in HEK293 and in Hep3B2 cells by EMSA approach. Three major complexes were detected in HEK293 cells (Fig. 7B). Incorporation of specific antibodies into the binding reactions revealed that the slower migrating band represents the E2F1·Rb complex. Because Rb is phosphorylated at Ser-612 in HEK293 cells, we asked if this isoform of Rb is involved in the formation of Rb·E2F1 complexes. EMSA showed that Abs to Ser-612-ph supershifted the E2F1·Rb complex (Fig. 7C). Quite different compositions of the Rb·E2F complexes are observed in Hep3B2 cells (Fig. 7D). In these cells free E2F4 and E2F4p130 represent the major complexes that interact with the DHFR probe. Taken together, these studies show that HEK293 and Hep3B2 cells express different levels of E2F, Rb, and p130 proteins and contain different E2F·Rb complexes.

C/EBPβ-LIP Activates the B-myb Promoter by a Release of Rb-dependent Repression of the Promoter

Given the interaction of C/EBPβ-LIP with Rb (Fig. 2), we suggested that C/EBPβ-LIP might activate E2F-dependent promoters that are repressed by Rb. B-myb promoter was used to test this suggestion as it has been shown to be repressed by Rb·E2F complexes (37). The WT and mutant B-myb promoters (Fig. 8A) were co-transfected with C/EBPβ-LIP and with an empty vector into HEK293 cells. The cells were starved for 48 h, and the activity of the promoters was examined. We found that expression of C/EBPβ-LIP causes activation of the WT B-myb promoter. This activation is mediated via elimination of Rb·E2F-dependent repression as C/EBPβ-LIP does not change the activity of the mutant B-myb promoter, which is no longer under control of the Rb·E2F complexes (Fig. 8B). Thus, these studies showed that C/EBPβ-LIP activates the B-myb promoter by the release of Rb·E2F1 repression. We have next examined if CaM inhibits this activity of C/EBPβ-LIP. We have co-transfected WT CaM and the mutant CaMΔ with C/EBPβ-LIP and B-myb promoter into HEK293 cells. These studies showed that WT CaM blocks C/EBPβ-LIP-mediated release of the repression of the B-myb promoter, whereas the mutant CaMΔ does not change the C/EBPβ-LIP-mediated de-repression of the promoter (Fig. 8B). Because all three members of Rb family might repress the B-Myb promoter (37, 38), we have performed the second set of experiments in which we examined if the competition between CaM and Rb regulates the E2F-dependent promoters. These studies showed that ectopic expression of Rb in HEK293 cells further represses the B-myb promoter and that C/EBPβ-LIP releases this repression (Fig. 8C). WT CaM blocks the ability of C/EBPβ-LIP to release Rb-mediated repression of the B-myb promoter; however, the CaMΔ mutant is not able to do this.

FIGURE 8.

C/EBPβ-LIP activates the B-myb promoter via a release of E2F1·Rb mediated repression. A, a diagram shows the sequence of the B-Myb promoter and the general strategy of the experiments. Mut, mutant. B, WT CaM, but not CaMΔ, blocks the ability of C/EBPβ-LIP to de-repress the B-myb promoter. Luciferase-linked WT and Mut B-myb promoters were co-transfected into HEK293 cells with empty vector (control), with vector expressing C/EBPβ-LIP, and with vectors expressing WT CaM or CaMΔ. Cells were starved for 48 h, and luciferase activity was examined. Bottom image, levels of endogenous Rb and transfected proteins were determined in the extracts from the experimental plates. CRM, cross-reactive molecule. C, competition between Rb and CaM regulates the ability of C/EBPβ-LIP to activate the B-myb promoter. WT and mutant B-myb promoters were co-transfected with Rb, C/EBPβ-LIP, WT CaM, and CaMΔ. V, vector. The activity of the promoters was examined 16 h after transfections. Bottom image, levels of transfected proteins were determined in extracts from the experimental plates. CRM, cross-reactive molecule.

C/EBPβ-LIP Activates the E2F-dependent Promoters by Two Mechanisms

To determine the precise mechanisms by which C/EBPβ-LIP de-represses the E2F-dependent promoters, we examined if C/EBPβ-LIP disrupts the E2F1·Rb complexes and if CaM might inhibit this activity of C/EBPβ-LIP. EMSA studies show that the ectopic expression of C/EBPβ-LIP reduces the amounts of E2F1·Rb complexes to less than 10% and that WT CaM blocks this activity of C/EBPβ-LIP (Fig. 9, A and B). The mutant CaMΔ has much weaker ability to protect the E2F1·Rb complexes.

FIGURE 9.

Two mechanisms of C/EBPβ-LIP-mediated activation of E2F-dependent promoters. A, CaM blocks C/EBPβ-LIP-mediated disruption of the E2F1·Rb complexes via direct interactions with C/EBPβ-LIP. EMSA was performed with nuclear extracts from HEK293 cells transfected with an empty vector (V), C/EBPβ-LIP (LIP), or co-transfected with C/EBPβ-LIP and WT CaM or CaMΔ. The red arrow shows an additional band that is observed in C/EBPβ-LIP-transfected cells. NS, nonspecific band. B, CaM protects E2F1·Rb complexes from C/EBPβ-LIP mediated disruption. The bar graph presents a calculation of intensity of E2F1·Rb complex as a ratio to a nonspecific band. C, C/EBPβ-LIP binds to the DHFR promoter. EMSA was performed with nuclear extracts isolated from C/EBPβ-LIP-transfected cells. Salmon DNA was used as a nonspecific competitor. Abs to C/EBPβ were incorporated in the binding reaction as shown on the top. D, C/EBPβ-LIP directly binds to E2F-dependent promoters. Bacterially expressed, purified GST-C/EBPβ-LIP, and C/EBPα (positive control) were incubated with three DNA probes covering E2F consensuses in the promoters of DHFR, cdc2, and E2F1 genes. Poly(dI-dC) competitor was used in these studies. Antibodies to C/EBPα and C/EBPβ were incorporated in the binding reactions as shown on the top. Positions of C/EBPβ-LIP, C/EBPα, supershifts (SS), and free probes are shown. E, CaM reduces interactions of C/EBPβ-LIP with the DHFR promoter. HEK293 cells were transfected with C/EBPβ-LIP, C/EBPβ-LIP + WT CaM, and CaMΔ. Nuclear extracts were isolated and used for EMSA with the DHFR probe. Antibodies to C/EBPβ were incorporated in the binding reactions. Positions of C/EBPβ-LIP DNA complexes are by red arrows. F, C/EBPβ-LIP interacts with E2F-dependent promoters in vivo and displaces Rb·E2F1 complexes from these promoters; CaM blocks this activity of C/EBPβ-LIP. Chromatin-IP studies were performed with chromatin solutions isolated from HEK293 cells after transfections with FLAG-C/EBPβ-LIP, with empty vector, and with FLAG-C/EBPβ-LIP + CaM and FLAG-C/EBPβ-LIP + CaMΔ. E2F1, Rb, and FLAG were immunoprecipitated from chromatin solutions. PCR reactions were performed with primers covering E2F sites within the B-myb and DHFR promoters. M, markers; In, 1/100 input; B, beads.

In the course of EMSA experiments, we observed the appearance of additional bands with nuclear extracts of cells transfected with C/EBPβ-LIP (Fig. 9, A and C, red arrows), suggesting that C/EBPβ-LIP might directly bind to the DHFR probe. To examine this possibility, we incorporated antibodies to C/EBPβ-LIP into binding reactions with nuclear extracts from HEK293 cells transfected with C/EBPβ-LIP. Fig. 9C shows that antibodies to C/EBPβ supershifted/neutralized the three additional bands. These results suggested that C/EBPβ-LIP might directly bind to the DHFR promoter. Because salmon DNA competitor (used in EMSA for E2F binding) is not optimal for the interactions of C/EBP proteins with DNA, we performed further studies of the interactions of C/EBPβ-LIP with the three E2F DNA probes (covering E2F binding sites within DHFR, cdc2, and E2F1 promoters) using dI-dC competitor, which completely inhibits E2F binding but does not affect interactions of C/EBP proteins with DNA (30, 31). We previously found that several E2F sites contain C/EBP consensus and that C/EBPα binds to these consensuses in vitro and in vivo (31). Therefore, we used C/EBPα as a positive control. EMSA studies showed that both C/EBPα and bacterially expressed, purified to homogeneity C/EBPβ-LIP bind to all tested E2F-dependent promoters (Fig. 9D). We have next examined if CaM might inhibit the interaction of C/EBPβ-LIP with the E2F consensuses. Fig. 9E shows that WT CaM dramatically reduces the interactions of C/EBPβ-LIP with the DHFR promoter, whereas the mutant CaMΔ does not significantly affect the interactions of C/EBPβ-LIP with the DHFR promoter.

To examine if C/EBPβ-LIP interacts with E2F-dependent promoters in vivo and displaces E2F1·Rb complexes from the promoters, we performed chromatin-IP analysis. Because endogenous C/EBPβ-LIP cannot be distinguished from C/EBPβ-LAP by available antibodies, we transfected FLAG-tagged C/EBPβ-LIP alone or with WT CaM and CaMΔ into HEK293 cells and examined the occupation of endogenous B-myb and DHFR promoters by E2F1, Rb, and FLAG-C/EBPβ-LIP. In control cells B-myb and DHFR promoters are occupied by the Rb·E2F1 complex, which represses these promoters (Fig. 9F). The expression of FLAG-C/EBPβ-LIP causes a reduction of the complexes on the B-myb and DHFR promoters, whereas FLAG-C/EBPβ-LIP is abundant on the promoters, suggesting that it displaces the Rb·E2F1 complexes from the E2F-dependent promoters (Fig. 9F). WT CaM blocks the C/EBPβ-LIP-dependent displacements of Rb·E2F1 complexes, whereas the mutant CaMΔ is not able to block this activity of C/EBPβ-LIP (Fig. 9F). Thus, these studies show that C/EBPβ-LIP binds to the E2F-dependent promoters in vivo and that this binding leads to the removal of E2F1·Rb complexes from these promoters.

DISCUSSION

Liver is a unique tissue that is able to regenerate itself after surgical resections (1–4). The transition of the liver from quiescence to proliferation requires the orchestrated re-organization of a number of pathways including changes in activity of C/EBP family proteins (2, 4). C/EBPβ is expressed at high levels in the liver, and it is one of the critical regulators of liver growth and differentiation (5, 6). The biological functions of C/EBPβ-LAP have been intensively investigated; however, very little is known about biological functions of the truncated isoform C/EBPβ-LIP. In this paper we have identified molecular mechanisms by which C/EBPβ-LIP promotes liver proliferation and mechanisms by which liver controls this activity. C/EBPβ-LIP is increased in several biological situations; however, its growth promotion activity is displayed only in livers after partial hepatectomy. In searching for the proteins that might control activities of C/EBPβ-LIP, we have identified CaM as a protein that interacts with C/EBPβ and blocks growth promotion activities of C/EBPβ-LIP. CaM has been previously implicated in the translation of calcium signaling mainly through four mechanisms (19). Data in our paper suggest an additional mechanism by which CaM translates calcium signaling. The reduction of Ca²⁺ after LPS treatment leads to the increased interactions of CaM with C/EBPβ-LIP in the liver. It is important to note that previous studies have shown that treatment of rats with LPS increases intracellular calcium in Kupffer cells and in endothelial cells but not in the hepatocytes (40). In addition, the LPS-mediated elevation of calcium in cultured hepatic microphages takes place within first the 2 min and then returns to normal levels (41). We have measured Ca²⁺ in the liver nuclear extracts starting from 4 h after LPS treatments and found a reduction of Ca²⁺ in the total nuclear extracts. Our data suggest that this reduction of Ca²⁺ leads to the increased interactions of CaM with C/EBPβ-LIP and that these interactions block growth promotion activities of C/EBPβ-LIP.

Although the ability of C/EBPβ-LIP to accelerate cell proliferation has been reported by several groups (16–18), molecular mechanisms underlying this activity have been not shown. Data in our paper suggest pathways by which C/EBPβ-LIP accelerates cell proliferation (Fig. 10B). Our data show that C/EBPβ-LIP promotes cell proliferation through two mechanisms. One mechanism includes the disruption of E2F1·Rb complexes and de-repression of the E2F-dependent promoters. The second mechanism includes a direct interaction of C/EBPβ-LIP with E2F consensuses on the E2F-dependent DHFR, cdc2, and cyclin E promoters. We suggest that such interactions of C/EBPβ-LIP displace E2F·Rb complexes from the promoters and release the repression of the promoters. It is also important to note that, in addition to the release of E2F·Rb-mediated repression of E2F-dependent targets, C/EBPβ-LIP down-regulates expression of C/EBPα, which is a strong inhibitor of liver proliferation. This pathway might also contribute to the promotion of proliferation. We have identified CaM as the protein that regulates activities of C/EBPβ in the liver and, therefore, might regulate liver proliferation. We suggest that CaM displays the main effects on the liver proliferation via inhibition of activities of C/EBPβ-LIP and perhaps C/EBPβ-LAP. Under conditions of APR, direct interactions of CaM with C/EBPβ-LAP and C/EBPβ-LIP block the activities of the C/EBPβ proteins and prevent liver proliferation (Fig. 10A). Previous studies show that CaM might regulate cell cycle progression through interactions with p21 and with cyclin E in other tissues (24, 25, 28). It would be interesting to examine if CaM interacts with these proteins in the liver and if these possible interactions might be involved in the regulation of liver proliferation. Because Ca²⁺ regulates interactions of CaM and C/EBPβ, we suggest that CaM is one of the important regulators of liver proliferation as a mediator of calcium signaling.

FIGURE 10.

A hypothetical model for CaM-mediated control of liver proliferation via interactions with C/EBPβ proteins. A, in livers with APR, CaM interacts with C/EBPβ-LIP and C/EBPβ-LAP and inhibits biological activities of these proteins. B, inhibition of CaM with siRNA or overexpression of C/EBPβ-LIP above CaM levels promotes liver proliferation in LPS-treated mice. The release of C/EBPβ-LIP from complexes with CaM leads to neutralization of Rb functions by two pathways; that is, by the disruption of E2F1·Rb complexes and by displacement of these complexes from the E2F-dependent promoters.