Myc interacts with Max and Miz1 to repress C/EBPδ promoter activity and gene expression

To investigate the role of Myc as a repressor of C/EBPδ gene transcription we first determined Myc and C/EBPδ protein levels in actively cycling (growing (GR)) and growth arrested (GA), nontransformed HC11 mammary epithelial cells. The results confirmed that Myc protein levels are elevated in growing HC11 cells and reduced in growth arrested HC11 cells (Figure 1AB). Conversely, C/EBPδ protein levels are virtually undetectable in growing HC11 cells and C/EBPδ protein levels are induced in growth arrested HC11 cells (Figure 1AB). Cyclin D1, a labile G1/S marker, is elevated in growing HC11 cells and reduced in growth arrested cells, paralleling Myc protein levels and confirming HC11 growth (cell cycle) status in these experiments [46, 47]. Myc binding partners Miz1 and Max are also present at relatively constitutive levels in growing and growth arrested HC11 cells (Figure 1AB). Induction of C/EBPδ gene transcription requires the transcriptional activator Sp1 [7, 48, 49]. Sp1 protein levels, however, are unaffected by cell cycle status (Figure 1AB). These results demonstrate that Myc and C/EBPδ protein levels are directly influenced by growth status and that Myc and C/EBP protein levels are inversely correlated in nontransformed HC11 mammary epithelial cells.

To investigate Myc repression of C/EBPδ promoter activity, HC11 cells were transfected with increasing amounts of Myc (5~50 ng) plus a C/EBPδ promoter luciferase construct (Figure 1C). Myc repressed C/EBPδ promoter activity in a dose-dependent manner, even at dose levels as low as 5 ng (Figure 1C). Expression of Myc constructs was confirmed by Western blot analysis of cell lysates (Figure 1D). To map the domains of Myc essential for the repression of C/EBPδ promoter activity, Myc deletion mutants corresponding to Myc box 1 (ΔMB1, 45-63), Myc box 2 (ΔMB2, 129-143), basic region (ΔBR, 355-367), helix-loop-helix (ΔHLH, 368-410) and leucine zipper (ΔLZ, 411-439) were constructed (Figure 1E). The full length Myc construct and the Myc MB1 deletion mutant both repressed C/EBPδ promoter activity to ~50% of the empty vector control. These results indicate that the Myc MB1 deletion mutant is nearly as effective as the full length Myc construct in repressing C/EBPδ promoter activity and therefore, the Myc MB1 region is not required for Myc repression of C/EBPδ promoter activity (Figure 1F). In contrast, the Myc MB2, BR, HLH and LZ deletion mutants all resulted in C/EBPδ promoter activity that was similar to the empty vector control (Figure 1F). These results demonstrate that the MB2, BR, and the HLH/LZ regions are required for Myc repression of C/EBPδ promoter activity. Western blots demonstrated that the protein levels of the individual transfected Myc constructs were approximately equal; indicating that differences in C/EBPδ promoter activity were not due to variations in the expression of the transfected Myc constructs (Figure 1G).

Myc represses gene promoters by interacting with DNA bound transcription factors including Sp1 and Miz1 [44]. To identify the Myc interacting protein implicated in Myc repression of the C/EBPδ promoter we transfected HC11 cells with V5-tagged Myc expression constructs and performed co-immunoprecipitations to assess Myc/Miz1 and Myc/Sp1 interactions in HC11 cell lysates. The results demonstrated that Myc interacts with Miz1, but not Sp1, supporting a role for Myc/Miz1 repression of the C/EBPδ promoter (Figure 2A). We next used ChIP assays to investigate the association between Myc and Miz1 and the C/EBPδ proximal promoter (P200) in Growing ("Gr", C/EBPδ non-expressing) and growth arrested ("GA", C/EBPδ expressing) HC11 cells. The ChIP results demonstrated that both Myc and Miz1 associate with the C/EBPδ proximal (P200) promoter in HC11 cells under Growing ("Gr", C/EBPδ non-expressing) conditions (Figure 2B). Miz1 remains associated with the C/EBPδ proximal (P200) promoter under growth arrest ("GA", C/EBPδ expressing) but Myc is not associated with the C/EBPδ proximal (P200) promoter in growth arrested (GA) HC11 cells (Figure 2B). Regardless of the growth conditions, neither Miz1 nor Myc is associated with the distal C/EBPδ promoter region located 1.8 kb upstream of the C/EBPδ transcription start site (P1.8K) (Figure 2B). These results are consistent with the presence a Myc/Miz1 complex in association with the C/EBPδ proximal promoter during active cell proliferation when C/EBPδ gene transcription is repressed and the absence of Myc in association with the C/EBPδ proximal promoter during growth arrest when C/EBPδ gene transcription is highly induced [6]. The negative ChIP results from the distal C/EBPδ promoter region 1.8 kb upstream of the C/EBPδ transcription start site (P1.8K) indicate that the Myc repressive complex is localized to C/EBPδ proximal (P200) region.

To determine if Myc/Miz1 interaction is required for Myc repression of C/EBPδ promoter activation we obtained a mutant Myc construct that is deficient in Miz1 binding (MycV394D, Val³⁹⁴ → Asp, generous gift from Dr. M. Eilers). To validate the MycV394D Miz1 binding defect co-immunoprecipitation assays were performed on HC11 cells transfected with V5 tagged Myc wild type (wt) or the MycV394D (Miz1 binding deficient) constructs. The results demonstrated that the Myc wt construct (myc) binds to Miz1, but the Myc V394D construct ("394") does not (Figure 2C). To assess the functional significance of Myc/Miz1 binding on Myc repression of C/EBPδ promoter activity HC11 cells were transfected with Myc wt or the MycV394D Miz1 binding deficient mutant plus a C/EBPδ proximal (P200) promoter-luciferase construct. The results demonstrated that the Miz1 binding deficient MycV394D mutant construct was relatively ineffective in repressing C/EBPδ promoter activity compared to the Myc wt construct (Figure 2D). Western blots documented the expression of transfected Myc constructs (Figure 2E). These results demonstrate that optimal Myc repression of C/EBPδ promoter activity requires Miz1.

Previous reports have demonstrated that Miz1 functions as a transcriptional activator and that Myc represses Miz1 target gene activation [39]. Miz1 is associated with the C/EBPδ proximal promoter during growth arrest (Figure 2B above) when C/EBPδ is actively transcribed [6], suggesting that Miz1 may function as a transcriptional activator of C/EBPδ transcription. To investigate Miz1 transcriptional activation of the C/EBPδ promoter HC11 cells were co-transfected with a Miz1 expression construct plus a C/EBPδ proximal promoter-luciferase (P200) construct. The results demonstrate that Miz1 expression does not increase C/EBPδ promoter activity in proliferating (growing, Gr), or in growth arrested (GA) HC11 cells (Figure 3A). As expected, C/EBPδ promoter activity is higher in growth-arrested vs growing HC11 cells [6, 7]. Western blot analysis of HC11 cell lysates demonstrate the increased levels of Miz1 in HC11 cells transfected with the Miz1 expression construct and confirm the presence of Myc protein levels in growing (Gr) cells and the absence of Myc in growth arrested (GA) cells (Figure 3B).

Although Miz1 over expression had no effect on C/EBPδ promoter activity, reducing Miz1 levels by Miz1 siRNA treatment had a profound effect on C/EBPδ promoter activity in growing (Gr) HC11 cells. C/EBPδ promoter activity was induced ~2.5 fold in Miz1 siRNA treated; growing (Gr) HC11 cells (Figure 3C). Reducing Miz1 levels, however, had no effect on C/EBPδ promoter activity in growth arrested (GA) HC11 cells (Figure 3C). These results indicate that reducing Miz1 levels increases C/EBPδ promoter activity in proliferating cells, presumably by reducing Miz1/Myc repression (Figure 2D). Interestingly, C/EBPδ promoter activity is not altered by reducing Miz1 levels in growth arrested HC11 cells (Figure 3C). These results indicate that Miz1 does not increase C/EBPδ promoter activity during growth arrest. Western blot analysis of HC11 cell lysates confirmed that siRNA treatment was highly effective in reducing Miz1 protein levels (Figure 3D). To verify that Miz1 can function as a transcriptional activator in the proper cell context HEPG2 cells were transfected with a Miz1 expression construct plus a low-density lipoprotein receptor (LDLR) promoter-luciferase construct essentially as described by Tjian and co-workers [50]. The results demonstrated that Miz1 functions as a transcriptional activator of the LDLR promoter in HEPG2 cells (Figure 3E). These results demonstrate that Miz1 functions exclusively in Myc repression of C/EBPδ promoter activity under growing (proliferating) conditions but Miz1 does not activate C/EBPδ promoter activity in growth arrested HC11 nontransformed mammary epithelial cells.

Myc functions as a transcriptional repressor by binding to DNA-bound Miz1 [28]. Miz1 binds to highly divergent proximal promoter transcription initiator (Inr) elements [28]. To investigate Miz1 binding to the C/EBPδ promoter we performed electromobility shift assays (EMSAs) using recombinant mouse Miz1 and fluorescent labeled C/EBPδ proximal promoter fragments. The initial results confirmed Miz1 binding to the -140 to +30 C/EBPδ proximal promoter (Figure 4A). To localize the region of Miz1 binding we performed EMSAs using C/EBPδ promoter fragments deleted from the 3' and 5' ends. Miz1 binding was retained in all C/EBPδ proximal promoter fragments deleted from the 3' end, indicating that Miz1 binding was localized within -140 to -70 region of the C/EBPδ proximal promoter (Figure 4B). Deletions from the 5' of the C/EBPδ proximal promoter indicated that the Miz1 binding was localized within the -110 to -80 region of the C/EBPδ proximal promoter (Figure 4C, lanes f, g). To further investigate Miz1 binding EMSAs were performed with recombinant Miz1 protein and short (~30 bp) C/EBPδ proximal promoter fragments spanning the following regions of the C/EBPδ proximal promoter: -127 to -100 bp (Probe "h"), -100 to -70 bp (Probe "i") and +1 to +30 (Probe "j", a negative EMSA control). The results demonstrated weak Miz1 binding to the -127 to -100 region ("h"), strong Miz1 binding to the -100 to -70 region ("i"), and no detectable Miz1 binding to the +1 to +30 region ("j") (Figure 5A). We next performed individual competition assays with the same 3 C/EBPδ promoter fragments and the C/EBPδ -140 to +30 proximal promoter fragment. The results demonstrated that the -100 to -70 ("i") C/EBPδ promoter fragment was the most effective in reducing Miz1 binding to the C/EBPδ -140 to +30 proximal promoter fragment (Figure 5B). The -127 to -100 C/EBPδ promoter exhibited a limited capacity to compete with the C/EBPδ -140 to +30 proximal promoter fragment for Miz1 binding, consistent with the weak binding to this region demonstrated in Figure 5A. Although Inr sequences are highly degenerate, a candidate Inr sequence is located at -85 to -93 (5'-CCCCAGTCCCT-3') of the C/EBPδ proximal promoter, within the -100 to -70 region [6].

Myc Associated protein X (Max) is a ubiquitously expressed, long lived (t 1/2 > 24 hours) helix loop helix/leucine zipper (HLH/bZIP) protein that heterodimerizes with Myc and is required for Myc transcriptional activation and repression [51‐53]. Using ChIP assays, we assessed the association between Max and the C/EBPδ promoter under growing and under growth arrest conditions. The results indicated that Max is associated with the C/EBPδ promoter under both growing (GR, C/EBPδ expression repressed) and growth arrest (GA, C/EBPδ expression highly induced) conditions (Figure 6A). To determine if Max is required for Myc repression of C/EBPδ promoter activity HC11 cells were treated with Max siRNA and C/EBPδ promoter driven luciferase activity assessed. The results indicated that Max siRNA treatment reduces Myc repression of C/EBPδ promoter activity and this reduction in Myc repression is comparable to Miz1 siRNA treatment (Figure 6B). These results demonstrate that Max is constitutively associated with the C/EBPδ promoter and plays a key role in Myc repression of C/EBPδ promoter activity.

IL-6 and Oncostatin M (OSM) induce STAT3 activation (phosphorylation) and phosphorylated STAT3 (pSTAT3) activates C/EBPδ transcription in growth arrested cells [7, 8, 54]. OSM activates pSTAT3, but pSTAT3 does not fully activate C/EBPδ expression in proliferating (growing) cells due to Myc repression [43]. Myc repression of OSM induced endogenous C/EBPδ expression is attenuated by Myc siRNA treatment [43]. To investigate the role of Max and Miz1 in Myc repression of endogenous C/EBPδ expression HC11 cells were transfected with Max and Miz1-specific siRNAs. Endogenous C/EBPδ expression was assessed by western blot of whole cell lysates from actively proliferating (growing) vector control, Max and Miz1 siRNA treated HC11 cells treated with OSM. The results demonstrated that Max and Miz1 siRNA treatments reduced endogenous Max and Miz1 protein levels and the individual reductions in Max and Miz1 protein levels were associated with increased OSM-induced C/EBPδ protein levels compared to scrambled or "Junk" siRNA treated HC11 cells (Figure 6C). These results indicate that Max and Miz1 function in repression of endogenous C/EBPδ gene expression.

RuvBl1 (Pontin, TIP49) and RuvBl2 (Reptin, TIP48) are members of the highly conserved AAA+ (ATPases associated with diverse cellular activities) superfamily with functions in chromatin remodeling and transcriptional regulation [45]. We hypothesized that RuvBl1 (Pontin, TIP49) and RuvBl2 (Reptin, TIP48) may contribute to Myc repression of C/EBPδ promoter activity as both proteins interact with Myc Box II and enhance Myc transcriptional repression and Myc mediated transformation [25, 34, 53, 55]. In addition, RuvBl1 (Pontin) and RuvBl2 (Reptin) are overexpressed in a variety of human cancers [45]. To investigate the role of RuvBl1 and RuvBl2 in Myc repression of C/EBPδ promoter activation RuvBl1 and RuvBl2 expression constructs were transfected into proliferating HC11 cells and C/EBPδ promoter activity assessed by luciferase assay. The results demonstrated that both RuvBl1 and RuvBl2 repress C/EBPδ promoter activity in a dose-dependent manner (Figure 7A). In addition to assessing the repressive effects of RuvBl1 and RuvBl2 individually, the repressive effect of co-transfecting RuvBl1 and RuvBl2 on C/EBPδ promoter activity was also investigated. The results demonstrated that co-expression of RuvBl1 and RuvBl2 was more effective in repressing C/EBPδ promoter activity than expression of either RuvBl1 and RuvBl2 alone (Figure 7A). Western blot analysis documented the expression of the transfected constructs and the positive correlation between increased RuvBl1 and RuvBl2 expression levels and increased repression of C/EBPδ promoter activity (Figure 7B).

HC11 mouse mammary epithelial cells were grown in complete growth media (CGM) containing RPMI 1640 medium plus 5% fetal bovine serum (FBS), 10 μg/ml bovine insulin, 10 ng/ml epidermal growth factor, 100 U/ml penicillin, 100 μg/ml streptomycin and 500 ng/ml Fungizone. Growth arrest was induced by 24~48 hrs serum and growth factor withdrawal (growth arrest medium, GAM, 0.1% FBS).

Mouse C/EBPδ proximal promoter sequence flanking -127 bp to transcriptional start site (P-127, containing Sp1, STAT3 and CREB binding sites) was constructed in the pGL2 basic luciferase reporter vector [7, 60]. Myc and MycV394D mutant constructs in pBabe-puro vector were a generous gift from Dr. Martin Eilers (Institute for Molecular Biology and Tumor Research, University of Marburg, Germany). Myc and MycV394D were then amplified by PCR from pBabe-puro vector using primers specific for Myc. The primer sequences for Myc wild type and MycV394D cloning are as follows: 5'-CGCGGATCCGCGATGCCCCTCAACGTTAGCTTC-3' (forward primer) and 5'-GCTCTAGACGCGCACAAGAGTTCCGTAGCTG-3' (reverse primer). Myc deletion constructs MycΔ45-63(MB1), MycΔ129-143(MB2), MycΔ355-367(BR), MycΔ368-410(HLH) and MycΔ411-439(LZ) were constructed by site-specific mutagenesis as previously described [61, 62]. Myc-, Myc deletion- and V394D- pcDNA3.1-V5-His expression constructs were verified by sequencing. The Miz1 full length cDNA construct in pCMV6 vector was purchased from Origene.

HC11 cells were plated in 12-well plates, grown to 50% confluence in CGM and transfected using the enhanced Lipofectamine transfection protocol (Invitrogen, Carlsbad, CA) as previously described [60]. Co-transfections were performed with 0.3 ug C/EBPδ promoter luciferase reporter construct, 1 ng Renilla luciferase reporter construct (transfection efficiency control), and 5~50 ng of expression constructs or vector controls. For growth arrest experiments, transfected cells were washed 2× with PBS and cultured in GAM for 24-48 hours. Cells were harvested and assayed for firefly and renilla luciferase activities using the Dual-Luciferase Reporter Assay kit with luciferase detection by Hewlett-Packard Lumicount microplate luminometer as previously described [43]. C/EBPδ promoter activities were normalized to renilla luciferase activity. Results shown are the average-fold changes from 3 independent experiments with duplicates. Co-immunoprecipitation experiments were performed as described [43, 61, 62]. HC11 cell lysates used in co-immunoprecipitation assays were prepared by transfecting Myc or V394D Myc mutant expression constructs (1 μg) (Lipofectamine) into HC11 cells. HC11 Miz1 and Max siRNA transfections were performed using the Amaxa Nucleofector (Amaxa, Inc., Cologne, Germany). Briefly, HC11 cells were suspended in Amaxa Nucleofector Solution V supplemented with 50 pmol Miz1 or Max Smartpool siRNAs (Dharmacon, Inc., Lafayette, CO) and the nucleofection was performed using cell-type specific protocol (T-20). HC11 cells nucleofected with non-specific scrambled siRNAs were used as controls. Transient siRNA nucleofection protocols were optimized and protocols achieving >80% specific gene knockdown as verified by western blot were used in all experiments.

Western blots were performed on whole cell lysates as previously described [61, 62]. Co-immunoprecipitation assays were performed with HC11 cell lysates isolated by NP-40 lysis, primary antibody immunoprecipitation, Protein A-Agarose bead pull down, elution and analysis by SDS PAGE as previously described [61]. Co-immunoprecipitations were performed 2-3 times and representative results presented.

ChIP experiments were performed using the Chromatin Immunoprecipitation (ChIP) Assay Kit (Sigma) as previously described [3, 43]. Briefly, HC11 cells were cross-linked with 1% formaldehyde, washed 3× with cold PBS (4°C), and the nuclear pellets were collected by centrifuge. Nuclear pellets were then resuspended in 300 μl DNA shearing buffer containing protease inhibitor cocktail, sonicated on ice to approximately 200~1000 bp (verified by standard agarose gel analysis), centrifuged at 14000 rpm for 10 minutes to pellet cell debris and the supernatants were collected and diluted 1:1 in dilution buffer and used for DNA immunoprecipitation. 10 ul diluted supernatant was used as input control. One μg of Myc or Miz1 specific IgG immunoprecipitated protein-DNA complexes were isolated and protein-DNA crosslinks reversed (65°C, 2 hours). After purification, immunoprecipitated DNA was analyzed by PCR using primers specific for proximal and distal mouse C/EBPδ promoter [60]. Primer sequences are as follows: P200 (region -226 to -24 of the mouse C/EBPδ promoter containing STAT3 and SP1 binding sites), 5'-GCGTGTCGGGGCCAAATCCA-3'(forward primer), 5'-TTTCTAGCCCCAGCTGACGCGC-3'(reverse primer); P1.8K (region -1856 to -1676 of the promoter) as control, 5'-TGCTTCTATGGCATCCAG-3'(forward primer), 5'-GAGGGGCTGTGGAATATT-3'(reverse primer).

Full length Miz1 cDNA was cloned into pGEX-4T-1 vector (Miz1-GST). The Miz1-GST plasmid was transformed to BL21 (DE3) competent cell (Stratagene). The Miz1-GST protein was purified by affinity binding using Glutathione sepharose beads (GE Healthcare) following the manufacturer's protocol. Miz1 protein was confirmed by western blot with detection using Miz1 and GST antibodies (Santa Cruz, Biotechnology).

DNA probes (a to g) were generated by PCR using mouse C/EBPδ promoter (1.7 kb fragment) as template. Primer sequences are available upon request. Double stranded oligos used to produce Probes h, j, and i were purchased (Sigma). Probes used in EMSA reactions were 5' end-labeled with 6-FAM (6-Carboxyfluorescein, Sigma). EMSAs were performed by incubating labeled probes (20 ng) with purified Miz1 protein in binding buffer (10 mM Tris pH 7.9, 4 mM MgCl_2, 5% glycerol, 0.1 mM DTT, 20 ng/μl poly(dI:dC) and 0.2% NP-40) for one hour at room temperature. To perform EMSA competition assays unlabelled probes were pre-incubated with Miz1 in binding buffer for 10 min prior to addition of the labeled probe. The concentration of unlabeled probes used was 5-25-fold molar excess over labeled probe. Following incubation, samples were loaded onto a 4.5% native acrylamide gel (pre-run for one hour) and electrophoresed for one hour at 100 V. Gels were scanned using the Typhoon 9410 imager (GE healthcare).