Activation of Protein Kinase C Induces Nuclear Translocation of RFX1 and Down-regulates c-myc via an Intron 1 X Box in Undifferentiated Leukemia HL-60 Cells*

Treatment of human promyelocytic leukemia cells (HL-60) with phorbol 12-myristate 13-acetate (PMA) is known to decrease c-myc mRNA by blocking transcription elongation at sites near the first exon/intron border. Treatment of HL-60 cells with either PMA or bryostatin 1, which acutely activates protein kinase C (PKC), decreased the levels of myc mRNA and Myc protein. The inhibition of Myc synthesis accounted for the drop in Myc protein, because PMA treatment had no effect on Myc turnover. Treatment with PMA or bryostatin 1 increased nuclear protein binding to MIE1, a c-myc intron 1 element that defines an RFX1-binding X box. RFX1 antiserum supershifted MIE1-protein complexes. Increased MIE1 binding was independent of protein synthesis and abolished by a selective PKC inhibitor, which also prevented the effect of PMA onmyc mRNA and protein levels and Myc synthesis. PMA treatment increased RFX1 in the nuclear fraction and decreased it in the cytosol without affecting total RFX1. Transfection of HL-60 cells with myc reporter gene constructs showed that the RFX1-binding X box was required for the down-regulation of reporter gene expression by PMA. These findings suggest that nuclear translocation and binding of RFX1 to the X box cause the down-regulation of myc expression, which follows acute PKC activation in undifferentiated HL-60 cells.

The c-myc proto-oncogene encodes nuclear proteins, which heterodimerize with Max and regulate the expression of genes implicated in cell growth (size), metabolism, differentiation, apoptosis, tumorigenesis, and genomic stability (1)(2)(3)(4). Activation of the c-myc gene is a crucial oncogenic determinant in a wide variety of human cancers (1,3). Uncontrolled expression of the normal Myc proteins is associated with a wide variety of animal and human tumors, with almost a third of breast and colon carcinomas having elevated c-myc expression (1). Mammalian cells express the following multiple Myc polypeptides: Myc1, Myc2, and MycS, which are produced by initiation of translation at different codons (5). Normal cell function depends on tightly modulated Myc protein levels. Myc proteins and myc mRNA turnover rapidly (t1 ⁄2 Ͻ 30 min) in eukaryotic cells, and multiple redundant systems regulate myc transcription and translation (1)(2)(3). Myc and MycS proteins are degraded by the tightly regulated ubiquitin-proteasome system (6,7), and stabilization of Myc has been suggested to be caused by certain cancer-associated mutations (6). Overexpression of Myc in mammalian cells blocks differentiation, predisposes to malignant transformation, and can initiate apoptosis (8 -10).
Initiation of c-myc transcription at the two major promoters (P1 and P2) is under the control of several protein factors and DNA elements (2). Furthermore, c-myc was the first eukaryotic gene shown to be regulated at the level of transcription elongation (11,12). Premature transcription termination near the first exon/intron junction depends on initiation at the predominant P2 promoter and explains the early phase of c-myc downregulation following induction of differentiation (2,(11)(12)(13). For example, in human promyelocytic leukemia HL-60 cells, induction of differentiation along either the monocytic/macrophage pathway by phorbol 12-myristate 13-acetate (PMA) 1 and perhaps by 1,25-dihydroxyvitamin D 3 or along the granulocytic pathway by retinoic acid or Me 2 SO blocks c-myc transcription near the first exon/intron border (11)(12)(13)(14)(15). Protein kinase C ␤ plays a critical role in the differentiation response to PMA, retinoic acid, and 1,25-dihydroxyvitamin D 3 (16 -18). Cotransfection of Burkitt's lymphoma cells with a c-myc reporter gene construct together with myc gene fragments suggested that the 5Ј half of the first intron contained sequences that competed for one or more putative negative regulatory factors (19). Remarkably somatic mutations in a 20-bp c-myc intron 1 element (called MIE1 or MIF-1) abolished nuclear protein binding to the element and were associated with c-myc activation in Burkitt's lymphoma cell lines (20). Burkitt's lymphoma mutations also appeared to be clustered in two additional protein binding elements (MIE2 and MIE3) that were just downstream of MIE1 (21). The functional significance of the intron 1 elements in myc expression remains to be established. Deletion of MIE1 and MIE2 had no effect on c-myc-driven reporter gene expression, and deletion of MIE3 modestly increased reporter gene activity in transfected cells (21). Recently five tandem repeats of MIE1 were shown to suppress the activity of the SV40 promoter in hepatocarcinoma cell lines (22,23).
RFX proteins are the chief component of nuclear complexes previously referred to as MDBP, MIF-1, NF-X, EF-C, or EP protein (21,24,27,31). RFX family members (RFX1-5) have a highly conserved winged-helix DNA-binding domain (32). RFX proteins homo-and heterodimerize with one another and up-or down-regulate transcription of target genes in a DNA contextdependent manner (27,(33)(34)(35). RFX1, which appears to be ubiquitously expressed in mammalian cells, has an N-terminal activation domain and a C-terminal repression domain that overlaps the dimerization domain (27,28,31). The functional regions can neutralize one another resulting in a nearly inactive transcription factor (34). Association of RFX1 with other family members, with non-RFX proteins such as c-Abl, or with other DNA-bound proteins apparently determines whether it has enhancer or silencer activity, although the determinants of the activity are not understood (33)(34)(35)(36).
In this study, we measured the rates of synthesis and degradation of Myc proteins following the treatment of undifferentiated HL-60 cells with PMA or Bryo. Bryo, like PMA, binds to the zinc finger C1 domain of conventional (␣, ␤, and ␥) and novel (␦, ⑀, , , and ) isoforms of PKC, which turns on its kinase function and concomitantly predisposes it to ubiquitinylation and degradation by the 26 S proteasome (37)(38)(39). In contrast to PMA, Bryo fails to induce differentiation of HL-60 cells and prevents the induction of HL-60 differentiation by PMA (40). Bryo, which is in clinical trials as an anti-cancer agent, apparently down-regulates PKC more rapidly and efficiently than PMA (41). Our results indicate that a brief activation of PKC in undifferentiated HL-60 cells decreased the Myc protein by blocking its synthesis without affecting its turnover. Studies of HL-60 cells transfected with c-myc luciferase reporter constructs suggest that the RFX-binding X box of intron 1 is essential for the down-regulation of myc by PKC. We also show, for the first time, that PMA treatment induced nuclear translocation of RFX1. Our findings suggest that nuclear translocation and binding of RFX1 to the X box contributes to the down-regulation of myc expression following acute activation of PKC.

MATERIALS AND METHODS
Cell Culture and Protein Concentration-HL-60 cells were grown in RPMI 1640 containing 15% FBS, 100 units/ml penicillin G, and 0.1 mg/ml streptomycin. The medium was diluted with fresh medium three times per week, and cell density was kept below 1 million per ml. The cells were collected by centrifugation and incubated in RPMI 1640 containing the additions indicated in the figure legends. Protein concentration was measured by the BCA method (Pierce) with bovine serum albumin as a standard.
Plasmids-pMPCAT (42), which contains a 3.2-kb HindIII-SacI fragment (nucleotide Ϫ2238 to ϩ936) of human c-myc upstream of the CAT gene, was used to produce the luciferase reporter constructs. A 2-kb fragment of the c-myc (nucleotide Ϫ1058 to ϩ936 relative to P1) was excised from pMPCAT with KpnI and EcoRV and inserted into the pGL3 control vector at the KpnI and HindIII (blunt) sites in place of the SV40 promoter. The 2-kb c-myc promoter was upstream of the SV40 enhancer and the firefly luciferase gene. The same procedure was used to subclone the Myc promoter mutants from pMPCAT⌬287 and pMP-CAT⌬220 (21). pMP-Luc⌬14, which lacked the 14-bp intron 1 X box (nucleotide 3004 -3017) was produced by overlap extension polymerase chain reaction mutagenesis with the following forward and reverse primers, respectively: 5Ј-TTT TCT CAG ATG GGG CTG GGG TGG GGG GTA and 5Ј-CCC AGC CCC ATC TGA GAA AAG TGT CAA TAG. Each of the constructions was validated by sequencing, carried out on double-stranded DNA with dye-terminator chemistry, and the products were resolved using an ABI Prism 377 automated sequencer.
Electroporation and Dual-luciferase Assay-HL-60 cells were collected by centrifugation and rinsed once with antibiotic-free RPMI 1640, and 20 million cells were suspended with 0.8 ml of this medium. Electroporation was done at room temperature at 350 V and 960 microfarads with a Gene Pulser (Bio-Rad) and 15 g of the indicated c-myc reporter vector and 15 g of pRL-TK (Promega) to control for transfection efficiency. After electroporation the cells were placed on ice for 30 min prior to dilution with 20 ml of RPMI 1640 containing 10% FBS without antibiotics. To assay luciferase the cells were harvested by centrifugation, rinsed once with room temperature phosphate-buffered saline, and suspended with passive lysis buffer (Promega). The cells were lysed by three freeze-thaw cycles using liquid nitrogen and a room temperature water bath. The protein concentration of the samples was measured, and each was diluted to 3 g/l with lysis buffer. Luciferase activity (60 g) was determined with the Dual-Luciferase™ Reporter Assay System (Promega) as recommended by the manufacturer. Statistical analysis was done by two-tailed Student's t test.
Western Blot Analyses-For Myc protein analysis HL-60 cells were lysed with hot (Ͼ95°C) SDS lysis buffer, which contained 1% (w/v) SDS, 2 mM EDTA, 2 mM EGTA, and 10 mM Tris-HCl, pH 7.4. Samples containing 30 g of protein were fractionated by SDS-PAGE on a 10% gel (Myc proteins) or 7% gel (MIBP1 or RFX1). Proteins were electrophoretically transferred to a polyvinylidene difluoride membrane (Millipore Corp.), and the membrane was blocked for 1 h at room temperature with 5% (w/v) nonfat dry milk in TBS. TBS contained the following (per liter): 8 g of NaCl, 0.2 g of KCl, and 3 g of Tris base and was adjusted to pH 7.4 at room temperature, and in the case of TBST, 0.05% (w/v) Tween 20. Membranes were incubated overnight at 4°C with 2 g/ml OP10 primary antibody in TBS containing 1% nonfat dry milk. For Western analysis of MIBP1, RFX1, and NFB, the membrane was incubated overnight at 4°C with a mouse monoclonal antibody to NFB (200-fold dilution, F-6, Santa Cruz Biotechnology, Inc.), rabbit antiserum to MIBP1 (250-fold dilution), rabbit antiserum to RFX1 (1000-fold dilution), or the respective preimmune serum in TBS containing 1% nonfat dry milk. The preparation and specificity of anti-MIBP1 and anti-RFX1 were described previously (22). Membranes were rinsed and processed with horseradish peroxidase-conjugated goat antimouse IgG (Transduction Laboratories) or with horseradish peroxidaseconjugated goat anti-rabbit IgG (BIOSOURCE) and a chemiluminescent substrate as described (7). Films were scanned and analyzed with a model GS-670 imaging densitometer using Molecular Analyst software (Bio-Rad).
Immunoprecipitations and Pulse Labeling of Myc-For immunoprecipitation a sample of the SDS lysate (usually 0.5 mg of protein) was diluted 10-fold with immunoprecipitation buffer, which contained the following: 70 mM NaCl, 50 mM NaF, 1 mM EDTA, 1 mM EGTA, 0.2 mM sodium orthovanadate, 0.2 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, 0.5% Nonidet P-40, and 10 mM Tris-HCl, pH 7.4. 2 g of the C-8 anti-Myc monoclonal (C-8; Santa Cruz Biotechnology) or 5 l of anti-MIBP1, anti-RFX1, or preimmune sera were added, and the solution was rotated overnight at 4°C with 30 l of a 50% suspension of protein A-agarose (Life Technologies, Inc.) present during the last hour. Immunoprecipitates were collected by centrifugation and washed twice with each of two buffers, which differed only in NaCl concentration; the first buffer contained the following: 1 mM EDTA, 1 mM EGTA, 500 mM NaCl, 0.5% Nonidet P-40, 1% Triton, and 10 mM Tris-HCl, pH 7.4; the second buffer lacked NaCl. Finally immunoprecipitates were washed twice with 10 mM Tris-HCl, pH 7.4, extracted with 20 l of 2ϫ SDS sample loading buffer, and incubated for 5 min in a boiling water bath. Proteins were fractionated by SDS-PAGE (10% gel), and the dried gel was fluorographed with Eastman Kodak Co. PPB film to visualize Pulse-Chase Labeling of Myc-HL-60 cells (10 7 cells per condition) were rinsed twice with phosphate-buffered saline and once with Met/ Cys-free culture medium. Pulse-labeling was usually done for 60 min with medium containing 10% of the usual concentration of Cys and Met and 0.15 mCi of [ 35 S]Met/Cys. Labeling was stopped by addition of 10 mM each of Met and Cys. After the indicated chase interval the cells were lysed with hot SDS buffer as described above, and Myc was immunoprecipitated from a sample containing 0.5 mg of protein as described above with the OP10 antibody. Half-lives were determined by nonlinear regression curve fitting to a single exponential decay equation.
c-myc Northern Blot Analysis-Total RNA was extracted by the acidified guanidinium thiocyanate-phenol chloroform method with Trizol as recommended by the manufacturer (Life Technologies, Inc.) and quantified by absorbance at 260 nm. RNA samples (10 g) were size-fractionated by electrophoresis on 1% agarose gel containing the following: 20 mM MOPS, pH 7.4, 1 mM EDTA, 5 mM sodium acetate, 0.2 M formaldehyde, and 0.5 g/ml ethidium bromide. RNA samples contained 50% formamide. The gel was illuminated with a UV lamp and photographed to compare the quality and quantity of the rRNA. RNA was transferred to Duralon (UV) membranes (Stratagene) by downward capillary transfer with the Turboblotter (Schleicher and Schuell) and cross-linked to the membrane with a Stratalinker 1800 (Stratagene). Membranes were prehybridized with 6 ml of QuikHyb (Stratagene) for 10 min at 68°C in a roller-bottle oven. 32 P-Labeled cDNA probe (3-5 Ci/ml; Ͼ109 cpm/g) was mixed with 0.1 ml of denatured salmon sperm DNA (10 mg/ml) and added to the roller bottle. After 2 h at 68°C the membrane was washed twice at room temperature for 15 min with twice-concentrated sodium chloride sodium citrate (SSC) containing 0.1% SDS and twice at 60°C for 15 min with SSC containing 0.1% SDS. SSC contained 8.8 g of NaCl and 4.4 g of sodium citrate per liter and was adjusted to pH 7.0 with HCl. A 1.4-kilobase pair ClaI-EcoRI fragment of pHSR-1 of human c-myc (ATCC number 41010) was agarose gel-purified and labeled using [␣-32 P]dCTP and the Klenow fragment of DNA polymerase I (Life Technologies, Inc.). c-myc transcript was quantified by autoradiography with Konica PPB film and an intensifying screen for Ͻ24 h at Ϫ70°C. Autoradiograms were scanned and analyzed with a model GS-670 imaging densitometer using Molecular Analyst software (Bio-Rad).
Nuclear Extracts-The cells (3 ϫ 10 7 per condition) were rinsed and subjected to hypotonic lysis without mechanical disruption in buffer A, which contained the following: 10 mM Hepes-Tris, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 0.5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 20 nM calyculin A. The cells were incubated on ice for 10 min and observed by phase-contrast microscopy to determine that Ͼ95% lysis had occurred. A nuclear pellet was obtained by centrifugation (16,000 ϫ g for 10 s) and extracted with buffer B for 20 min on ice with intermittent dispersal by pipetting. Buffer B contained the following: 20 mM Hepes-Tris, pH 7.9, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 0.5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 100 nM calyculin A. Particulate material was removed by centrifugation for 15 min at 16,000 ϫ g at 4°C, and the supernatant was used for electrophoretic mobility shift assay (EMSA). Nuclear extracts were diluted 3-fold with buffer C, which contained 20 mM Hepes-Tris, pH 7.9, 20% glycerol, 100 mM KCl, 0.5 mM dithiothreitol, 0.2 mM EDTA, and the following phosphatase inhibitors: 2 nM calyculin A, 1 mM Na 3 VO 4 , and 5 mM sodium pyrophosphate. Nuclear extracts (ϳ1 g/l protein) were used immediately for EMSA and then frozen in liquid N 2 and stored at Ϫ80°C.
EMSA and Supershift Assays-Binding reactions, which contained nuclear extract (2 g of protein), 12 l of buffer C, and 1 l of poly(dI-dC) (1 g/l), were incubated for 10 min at 25°C in the absence or presence of the indicated competitor double-stranded oligonucleotide or 1 l of rabbit antiserum to MIBP1, RFX1, or the respective preimmune serum. [ 32 P]-Labeled double-stranded MIE1 (0.1 ng; ϳ20,000 cpm) was added, and incubation was continued for 30 min. After the addition of gel loading buffer (2 l), which contained 250 mM Tris-HCl, pH 7.4, 0.2% bromphenol blue, and 40% glycerol, the entire reaction was loaded onto a 4% acrylamide gel that had been prerun for 1 h at 100 V at 4°C. Electrophoresis was for 1 h at 100 V and 4°C, and the gel was dried and autoradiographed. The gel buffer contained the following: 380 mM glycine, 50 mM Tris, and 2 mM EDTA and had a pH value of 8.5.
Complementary strands of the oligonucleotides (Life Technologies, Inc.), having the sequences indicated in Fig. 5B, were mixed in annealing buffer (20 mM Tris-HCl, pH 7.4, 50 mM NaCl, and 1 mM EDTA), incubated at 65°C for 10 min, and allowed to cool slowly (Ͼ4 h) to room temperature. The double-stranded oligonucleotides were 5Ј end-labeled with [ 32 P] using T4 kinase (Life Technologies, Inc.) and PAGE purified on a 20% gel that had been prerun for 1 h at 150 V.
Reagents-The following three antibodies that recognize c-Myc were used: OP10 (Calbiochem), which is a monoclonal to the Myc epitope tag (amino acids 410 -419); a rabbit polyclonal to c-Myc (catalog number 06 -340; Upstate Biotechnology); and C-8 (Santa Cruz Biotechnology, Inc.), which is a monoclonal produced by immunization with full-length human c-Myc. [␣-32 P]dCTP (3,000 Ci/mmol) and [ 35 S]Met/Cys (Ͼ1000 Ci/mmol; EXPRE 35 S 35 S) was from PerkinElmer Life Sciences. Bryo, PMA, and Bis were dissolved in dimethyl sulfoxide and added to culture medium from 1000-fold concentrated stock solutions.

Activation of PKC Decreases c-myc mRNA and Protein Levels and Myc2
Synthesis-Treatment of HL-60 cells with Bryo or PMA for 4 h markedly decreased the steady state level of the 64-kDa Myc2 protein (Fig. 1A). The decrease in Myc2 protein depended on the concentration of Bryo or PMA. At 1 nM neither compound affected the Myc2 level, and at 10 nM each compound almost maximally decreased Myc2 (Fig. 1A). PMA maximally decreased Myc2 by 82%, whereas Bryo maximally decreased Myc2 by 58% (Fig. 1B). Bis, a selective inhibitor of PKC (43), completely prevented PMA or Bryo from decreasing Myc2 (Fig.  2, top panel). Treatment of the cells with Bis alone moderately increased Myc2, which may be caused by the inhibition of residual PKC activity in cells not treated with PMA or Bryo (Fig. 2). In addition to decreasing the Myc2 level, treatment with 20 nM PMA or Bryo for 4 h decreased the level of c-myc mRNA (Fig. 2, middle panel). Bis prevented either PMA or Bryo from decreasing the c-myc transcript in the cells (Fig. 2). These results indicate that the decreases in myc mRNA and protein were produced by the activation of PKC. The 67-kDa Myc1 and 45-50-kDa MycS proteins are much less abundant in HL-60 cells than Myc2. 2 Although we quantified MycS protein levels and synthesis rates, we present only Myc2 data, because all of the treatments described had essentially the same effects on MycS and Myc2. 2 To determine whether the activation of PKC affected Myc translation, the cells were treated with PMA or Bryo for 1 h and pulse-labeled with [ 35  Myc2 labeling (Fig. 3A). Treatment with 20 nM PMA decreased the pulse labeling of Myc2 to 45 Ϯ 3% control (n ϭ 3). Simultaneous treatment with Bis prevented PMA from decreasing Myc2 labeling (Fig. 3B). Treatment with Bis alone slightly increased Myc2 labeling (Fig. 3B), in agreement with the effects of Bis on the Myc2 level (Fig. 2). Treatment with PMA had no effect on general protein synthesis, which was determined by the rate of incorporation of [ 35 S]Met/Cys into protein. 2 These results show that a 1-h treatment with PMA or Bryo decreased the rate of Myc2 synthesis in HL-60 cells. Treatment of the cells with 20 nM PMA for 1 h decreased myc mRNA to 42 Ϯ 3% control (n ϭ 3) (Fig. 3C). These findings show that a 1-h treatment with PMA produced similar decreases in myc mRNA and Myc2 synthesis. Treatment of the cells with 20 nM Bryo for 1 h decreased myc mRNA levels and Myc2 synthesis similarly to PMA. 2 Down-regulation of Myc Depends on the X Box of MIE1-To determine whether the previously-identified MIEs were involved in the down-regulation of Myc by PMA, we transfected HL-60 cells with a c-myc luciferase reporter vector, which contained a 2-kb c-myc cDNA upstream of the SV40 enhancer and luciferase gene (Fig. 4A). The 2-kb Myc cDNA consisted of 1057 bp upstream of exon 1, exon 1 (554 bp), and the first 387 bp of intron 1. Two Myc deletion mutants, pMP-Luc⌬287, which lacked all three MIEs, and pMP-Luc⌬220, which lacked MIE2 and MIE3, were used to determine whether one or more of the MIEs affected reporter gene expression (Fig. 4A). HL-60 cells were transfected with the pMP-Luc wild type or mutant vectors, and 18 h later half of the cells were treated with 20 nM PMA for 6 h. PMA treatment significantly decreased Mycdriven luciferase expression by 44 Ϯ 2% (p Ͻ 0.005) in cells transfected with pMP-Luc and by 40 Ϯ 4% (p Ͻ 0.02) in cells transfected with the deletion mutant that lacked MIE2 and MIE3 (Fig. 4B). In cells transfected with pMP-Luc deletion mutant that lacked all three MIEs, PMA treatment had no significant effect on luciferase expression (p ϭ 0.19) (Fig. 4B). These findings are consistent with the hypothesis that only MIE1 is required for the down-regulation of Myc-driven luciferase expression by PMA. Next we deleted only the 14-bp X box to determine whether it was required for the down-regulation of Myc. PMA had no significant effect on reporter gene expression in cells transfected with pMP-Luc⌬14 (p ϭ 0.44) (Fig. 4B). Luciferase expression in the untreated cells was essentially the same for each of the pMP-Luc constructs. These findings indicate the down-regulation of reporter gene expression by PMA required the c-myc intron 1 X box.
Activation of PKC Increases Protein Binding to MIE1 DNA-Treatment of HL-60 cells with 20 nM PMA or Bryo for 1 h increased MIE1 binding activity in the nuclear fraction (Fig.  5A). Two MIE1-protein complexes with slightly different elec-

FIG. 4. Effects of c-myc intron 1 elements on the down-regulation of reporter gene expression by PMA treatment of transfected HL-60 cells.
A, diagram of human c-myc-firefly luciferase construct (pMP-Luc) and deletion mutants lacking all three MIEs (pMP-Luc⌬287) or lacking MIE2 and MIE3 (pMP-Luc⌬220). pMP-Luc⌬14 lacked the 14-bp X box of MIE1. Nucleotide positions are relative to the P1 transcription start site. B, change in firefly luciferase expression produced by a 6-h treatment of transfected HL-60 cells with 20 nM PMA. After electroporation with the indicated vector, the cells were incubated for 18 h in 20 ml of RPMI 1640 containing 10% FBS before adding PMA to half the cells from each electroporation. The cells were cotransfected with pRL-TK vector, and the Dual-Luciferase™ Reporter Assay System (Promega) was used to measure Myc-driven firefly luciferase and Ranilla luciferase activity as a control for transfection efficiency. Values are mean Ϯ S.E. for five to seven experiments. The percentage change in luciferase activity produced by PMA was significantly different from pMP-Luc to pMP-Luc⌬287 (p ϭ 0.003) and to pMP-Luc⌬14 (p ϭ 0.004) but not to pMP-Luc⌬220 (p ϭ 0.419). trophoretic mobilities were observed by EMSA (Fig. 5). Treatment with Bis eliminated the increases in both MIE1-protein complexes produced by PMA or Bryo (Fig. 5A), as expected if the increases depended on the activation of PKC. The specificity of the binding to MIE1 was determined by addition of competitor oligonucleotides with the indicated sequences (Fig.  5B). A 10-fold excess of unlabeled MIE1 was sufficient to completely block the MIE1 protein binding (Fig. 5C). A 100-fold excess of duplex MIE2 and MIE3 oligonucleotides relative to the 32 P-labeled MIE1 probe had no effect on binding (Fig. 5C). Therefore, the MIE1-protein complexes were specific for MIE1. A 10-fold excess of the duplex BL1ϩ2 (Burkitt's lymphoma mutation 1 ϩ 2) oligonucleotide, a mutant with two substitutions in the 3Ј half of the MIE1 X box (Fig. 5B), had no effect on [ 32 P]MIE1 binding, and a 100-fold excess of the X box mutant only partially reduced binding (Fig. 5C). This finding confirms the role of the X box in nuclear protein binding to MIE1 as reported previously (20,24). These results show that a brief treatment of undifferentiated HL-60 cells with PMA or Bryo is sufficient to increase specific MIE1 binding activity in the nuclear fraction.
Treatment of the cells with PMA for 0.5, 1, 2, 3, and 4 h increased MIE1 binding activity. 2 However, a 48-h treatment of the cells with 0.1 M PMA, which induced differentiation, as indicated by the attachment and elongation of cells on the culture surface, had no effect on MIE1 binding. 2 A 48-h treatment with 0.1 M Bryo, which failed to induce attachment and differentiation, also had no effect on MIE1 binding. 2 These findings indicate that MIE1 binding activity returned to the basal level between 4 and 48 h of PMA treatment, and they suggest that there is no difference in MIE1 binding activity between undifferentiated and differentiated HL-60 cells, as observed by Ehrlich and co-workers (24).
Presence of RFX1 and MIBP1 in MIE1 DNA-Protein Complexes-Supershift analysis of MIE1-protein complexes was carried out with antisera to RFX1 and MIBP1 (Fig. 6). MIBP1 is a 160-kDa protein that is present in MIE1 complexes and apparently associates with RFX1 (22). The formation of both MIE1-protein complexes depended on the addition of the nuclear extract as expected (Fig. 6A). A 10-fold excess of unlabeled duplex MIE1, but not the BL1ϩ2 MIE1 mutant, abolished the supershifted complexes indicating that 32 P MIE1 binding was specific (Fig. 6A). Antiserum to RFX1 supershifted both of the complexes, but the MIBP1 antiserum supershifted only the slower mobility (Complex 1) (Fig. 6A). This finding suggests that both of the MIE1-protein complexes contained RFX1, but only the slower mobility complex contained MIBP1, as reported for MIE1-protein complexes from HeLa cells (22). Western blot analysis of nuclear extracts from undifferentiated HL-60 cells confirmed that the antisera specifically recognized proteins with the expected electrophoretic mobility of RFX1 or MIBP1, respectively (Fig. 6B). The preimmune sera were not reactive with MIBP1 or RFX1 (Fig. 6B).
PMA Treatment Increases Nuclear Accumulation of RFX1 but Not MIBP1-Treatment of HL-60 cells with PMA for 1 h increased RFX1 protein in the nuclear extract as determined by Western blot (Figs. 6B and 7A). PMA increased nuclear fraction RFX1 2.5 Ϯ 0.2-fold (n ϭ 4, p ϭ 0.001). NFB was also increased in the nuclear fraction by PMA treatment (Fig. 7A), which is known to induce NFB translocation from the cytoplasm to the nucleus (44). PMA treatment decreased RFX1 in the cytosol and had no effect on total cell RFX1 (Fig. 7A). PMA treatment had no effect on the level of MIBP1 in the nuclear fraction (Fig. 7B). The increase in nuclear RFX1 was independent of protein synthesis. Blockade of protein synthesis with cycloheximide had no effect on the accumulation of RFX1 in the nuclear fraction (Fig. 7B). Blockade of protein or RNA synthesis with cycloheximide or actinomycin D, respectively, had no effect on increased MIE1 binding activity produced by PMA (Fig. 7C). These findings indicate that the increases in nuclear extract RFX1 and MIE1 protein binding were independent of protein synthesis.  (Fig. 8). DISCUSSION It has been known for some time that premature termination of transcription plays a critically important role in the downregulation of c-myc in the early phase of the response to differentiation-inducing compounds in undifferentiated HL-60 cells (11)(12)(13)(14)(15). During the later phase of differentiation of HL-60 cells; however, a loss of transcriptional initiation occurs (45). In this report we observed that acute activators of conventional and novel isoforms of PKC, namely Bryo and PMA, rapidly and markedly decreased the steady state level of Myc protein and mRNA in undifferentiated HL-60 cells (Figs. 1 and 2). Treatment with PMA or Bryo for 1 h strongly inhibited the rate of Myc synthesis (Fig. 3). Because the half-life of the Myc protein was unaffected by PMA (Fig. 8), the inhibition of Myc synthesis explained the decrease in the Myc level. For this inference to be correct the Myc protein must have a relatively short half-life in untreated HL-60 cells, which it does (ϳ20 min) (Fig. 8).
Our findings indicate that PMA or Bryo markedly increased nuclear protein binding to MIE1 as determined by EMSA (Fig.  5). Protein binding depended on the MIE1 X box, because substitution of two nucleotides in the 3Ј half of the X box strongly decreased the ability of the BL1ϩ2 oligonucleotide to compete with MIE1 (Fig. 5C). Supershift analysis with antiserum to RFX1 showed that it was present in both of the MIE1-protein complexes that were resolved by EMSA (Fig. 6).
Only one of the complexes contained MIBP1, which is a 160-kDa uncharacterized MIE1-binding protein from HeLa cells (22). Increased MIE1 protein binding depended on PKC activation but was independent of protein synthesis (Figs. 5 and 7). Bis, a selective inhibitor of PKC (43), which prevented the down-regulation of myc mRNA and protein by PMA or Bryo, also abolished their effects on MIE1 binding activity (Fig. 5A). The mechanism by which acute activation of PKC increased protein binding to MIE1 appears to be indirect and due at least in part to the nuclear translocation of RFX1 (Fig. 7). Thus we have been unable to detect the 32 P-labeled RFX1 following immunoprecipitation from 32 P-labeled cells. 2 Although 32 P labeling of MIBP1 was readily detected, PMA treatment had no effect on the labeling. 2 Nuclear translocation of RFX1 could explain the PMA-or Bryo-evoked increase in the MIE1 complex that contained MIBP1, because this complex also contained RFX1 (Fig. 6A, Complex 1). In agreement with this idea, MIBP1 and RFX1 coimmunoprecipitated from HeLa cell nuclear extracts in the absence of MIE1 (22). No RFX nuclear localization signals have been reported, and there appear to be no reports of a dynamic change in the subcellular localization of an RFX family protein. One possible explanation for nuclear translocation of RFX1 is the phosphorylation of an RFX1-associated and -cotranslocated protein in response to PKC activation.
Although RFX proteins are most well known as essential transactivators of MHC class II genes and as a cellular transactivators of pathogenic viruses such as hepatitis B virus (27,29,30), RFX1 appears to be ubiquitously expressed in mammalian cells (25,27,28), including undifferentiated HL-60 cells as determined by Western blot analysis (Figs. 6 and 7) and FIG. 7. Effect of PMA on the level of RFX1 in nuclear extracts, cytosol, and total cell lysate. A, Western blot analysis of nuclear extract, cytosol, and total cell lysate from untreated and PMA-treated cells. After the PMA treatment (20 nM for 1 h in RPMI 1640 containing 2% FBS) a sample of the treated and untreated cells was extracted with Ͼ95°C SDS lysis buffer to obtain the total cell lysate. Another sample of the cells was subjected to hypotonic lysis, and the cytosol was collected after centrifugation to remove nuclei. Nuclear extracts were prepared as described under "Materials and Methods." Cytosol was concentrated with Centricon 10 concentrators, and proteins (30 g of nuclear extract, 60 g of cytosol, or 100 g of total lysate) were sizefractionated by SDS-PAGE (7% gel) and subjected to Western blot analysis with the indicated antiserum. B, Western blot analysis of nuclear extracts (30 g of protein) from cells that were incubated with 10 g/ml cycloheximide (CHX) or 2.5 g/ml actinomycin D (AD) for 10 min prior to the addition of 20 nM PMA as indicated. 1 h later nuclear extracts were prepared. C, EMSA was carried out on nuclear extracts that were prepared from cells that were treated as described in B. Northern blot analysis (28). The present results suggest that RFX1 binding to the X box of intron 1 has silencer activity toward Myc expression and are consistent with the apparent silencer function of tandem MIE1 repeats toward SV40 promoter activity in hepatocarcinoma cell lines (22,23). Apparently the interaction of RFX proteins with other DNA-bound proteins determines whether it has enhancer or silencer activity, although the determinants of the activity are unknown (36).
A recent report by Pan and Simpson (18) of the suppression of c-myc following a 2-day treatment of HL-60 cells with 1,25dihydroxyvitamin D 3 implicated MIE1 and activation of PKC, in agreement with our study using acute PKC activators. However, they suggested that the homeobox HOXB4 protein was a major MIE1-binding protein and that the 1,25-dihydroxyvitamin D 3 treatment down-regulated c-myc by increasing the level of HOXB4 (18). It is not known if HOXB4 complexes with RFX1 or other MIE1-binding proteins. The following evidence supports the view that RFX1 plays a role in the rapid downregulation of c-myc following acute activation of PKC: 1) downregulation of myc-driven reporter gene expression by PMA was abolished by deletion of the RFX-binding X box (Fig. 4); 2) acute treatment with a PKC activator, PMA or Bryo, increased protein binding to the X box of MIE1 (Fig. 5); 3) RFX1 was present in the MIE1-protein complexes (Fig. 6); 4) a selective inhibitor of PKC prevented PMA or Bryo from increasing MIE1 binding activity and decreasing Myc expression (Figs. 2, 3, and 5); and 5) PMA treatment increased RFX1 in the nuclear fraction and decreased it in the cytosol (Fig. 7). It is noteworthy that PMA and Bryo only transiently increased MIE1 protein binding. 2 We observed no difference in MIE1 binding activity between undifferentiated and differentiated (2-day PMA treatment) HL-60 cells as reported previously (24). 2 The lack of increased MIE1 binding activity in differentiated cells is consistent with the role of transcription termination near the first exon/intron junction in the early response to a stimulus of differentiation (14,45). Hence, acute induction of nuclear translocation and binding of RFX1 to the intron 1 X box produced by PMA correlates with the rapid onset of the blockade of transcription elongation following the addition of PMA or other stimuli of differentiation, in contrast to the later phases of Myc downregulation, which involve a loss of transcriptional initiation (14,45). Bryo treatment down-regulated endogenous c-myc similarly to PMA (this report); however Bryo fails to induce a differentiation response and antagonizes differentiation produced by PMA (40). PKC activation by Bryo is shorter in duration than that produced by PMA, because Bryo more rapidly and efficiently down-regulates PKC (41). Additional studies are needed to identify critical differences in the cellular responses to PMA and Bryo, which are subsequent to the rapid down-regulation of c-myc.
In conclusion, the present findings implicate nuclear translocation of RFX1 and an intron 1 X box in the early phase of the down-regulation of c-myc produced by acute PKC activators in undifferentiated HL-60 cells. Considering the pivotal role of Myc overexpression in malignant tumors (1,2), biochemical understanding of Myc regulation by RFX1 should help to devise novel strategies for silencing Myc.