Activation of the human DNA polymerase beta promoter by a DNA-alkylating agent through induced phosphorylation of cAMP response element-binding protein-1.

Treatment of cells with the DNA-alkylating agent N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) induces expression of the endogenous mammalian DNA polymerase β (β-pol) gene and of the cloned promoter in transient expression studies. The lone cAMP response element (CRE) in the core promoter, along with functional protein kinase A, is critical for the MNNG-induced up-regulation. Recently, we described a kinetic mechanism for transcriptional regulation of the β-pol promoter in vitro and found that CRE-binding protein (CREB) from MNNG-treated cells differentially up-regulates the promoter by stimulating formation of closed preinitiation complex (RPc). Here, using a CRE-dependent chimeric β-pol promoter, we purified the RPc assembled with nuclear extract from MNNG-treated and control HeLa cells. Comparison of proteins in the purified RPc samples revealed that the MNNG induction is associated with a strong increase in the Ser133-phosphorylated form of recombinant CREB (CREB-1). CREB depletion of the nuclear extracts diminished transcriptional activity, and addition of purified Ser133-phosphorylated CREB-1 restored activity, whereas unphosphorylated CREB-1 did not. Addition of phosphorylated CREB-1 to the control cell extract mimicked the MNNG-induced up-regulation of transcriptional activity. These results indicate that phosphorylation of CREB-1 is the probable mechanism of activation of the β-pol promoter after treatment of cells with the DNA-alkylating agent MNNG.

. These cloned ␤-pol promoters lack typical TATA and CCAAT elements, are GϩC-rich, and have distinct binding elements for Sp1 and activating transcription factor (ATF)/ cAMP response element-binding protein, referred to as CREB throughout. The CRE site in the ␤-pol promoters is required for full promoter activity (7), and a purified CREB from bovine testis binds specifically to the conserved CRE site of the ␤-pol promoters and stimulates promoter activity in vitro (7,10). The human and bovine ␤-pol promoters also are known to have a functional binding element for the YY1 family of initiation site binding proteins (11).
␤-Pol gene expression is induced after exposure of cells to the DNA-alkylating agent MNNG. This induction required transcription (12), and use of a transfected ␤-pol core promoter fusion gene revealed transcriptional up-regulation of the ␤-pol promoter after MNNG treatment. This response is mediated through the CRE of the ␤-pol promoter (13), and up-regulation of promoter activity is dependent upon the protein kinase A (PKA) signal transduction pathway (14,15).
Using an in vitro transcription assay system, we found that the rate of transcript formation from a chimeric human ␤-pol promoter was much higher with nuclear extract (NE) from MNNG-treated HeLa cells than with extract from control HeLa cells. Further results indicated that this up-regulation was dependent upon CREB. The role of CREB from normal and MNNG-treated cells on the transcription initiation process has been described; CREB from MNNG-treated cells supports recruitment of more RP c than CREB from control cells (16). The present investigation was conducted to isolate RP c assembled from control and MNNG-treated HeLa cells and to characterize the CREB family member(s) present and its possible modification as a function of cellular MNNG treatment.

EXPERIMENTAL PROCEDURES
Treatment of Cells with MNNG and Preparation of NE-HeLa cells (S3, from ATCC) were grown as a monolayer in 150-mm culture dishes in Dulbecco's modified Eagle's medium with 10% (v/v) fetal bovine serum. After cells were grown to 80% confluence, the medium was replaced with fresh medium containing 30 M MNNG. The MNNG stock solution was prepared in dimethyl sulfoxide. The final concentration of dimethyl sulfoxide in the medium was 0.1% (v/v). For 32 P metabolic labeling experiments, cells were treated simultaneously with 0.1 mCi/ml [ 32 P]orthophosphate. After 5 h of MNNG treatment, cells were scraped and washed twice with cold phosphate-buffered saline (pH 7.4). Normal cells received 0.1% (v/v) dimethyl sulfoxide solution and were treated in the same way without MNNG. NEs from normal and MNNGtreated cells were prepared by the procedure of Shapiro et al. (17) as described previously (10,16). CREB-depleted NE (NEd) was prepared as described previously (10) and showed no detectable CRE binding activity in electrophoretic mobility shift assays (10).
Preparation of Biotinylated pSH15 Template-The pSH15 template, a derivative of the human ␤-pol promoter (16), was restricted with EcoRI, and overhanging ends were filled with Klenow fragment, replacing dTTP with biotin-11-dUTP. The standard reaction mixture for biotinylation (in a final volume of 100 l) contained 100 g of pSH15 template (EcoRI-restricted); a 500 M concentration each of dATP, dCTP, dGTP; 40 M biotin-11-dUTP; 1 ϫ TMD buffer (50 mM Tris-HCl (pH 7.2), 10 mM MgCl 2 , 0.2 mM dithiothreitol); and 10 units of Klenow fragment. The reaction mixture was incubated at 25°C for 15 min. DNA was extracted with phenol-chloroform, precipitated with ethanol, and finally purified on a Sephadex G-50 column to remove unincorporated biotin-11-dUTP.
Purification of Preinitiation Complex-Transcriptionally active RP c was prepared from control and MNNG-treated cells by magnetic Streptavidin-agarose bead selection. One g of biotinylated pSH15 template (pSH15 (biotin)) was incubated for 30 min at 22°C with 30 g of NE from control or MNNG-treated cells in 25 l of transcription buffer containing 20 mM Hepes (pH 7.9), 6.5 mM MgCl 2 , 65 mM KCl, 2 mM dithiothreitol, and 10% (v/v) glycerol. Then 2 l of magnetic Streptavidin-agarose beads (10 mg/ml Dyna beads M-280 from Dynal Inc.) was added to the above mixture, and the incubation was continued for an additional 30 min on a rotary mixer. The RP c assembled on pSH15 (biotin) template was separated from the rest of the NE proteins utilizing a Dynal MPC E/E-1 magnet (Dynal Inc.), then washed four times with 300 l of transcription buffer, suspended in 24 l of the same buffer containing 20 units of RNasin (Promega), and used directly for transcription assays. For electrophoretic mobility shift assay and Southwestern and Western blot analysis, the RP c purification was scaled up 25-fold. Transcription factors assembled in RP c were released from the pSH15 (biotin) template in 100 l of transcription buffer containing 1 M KCl. The mixture was shaken for 10 min on a rotary shaker, and released RP c proteins (transcription factors) were recovered by applying a magnetic field. Aliquots of 25 l were stored at Ϫ70°C for further analysis as described below.
In Vitro Run-off Transcription Assay-The in vitro run-off transcription assay system was described previously (10,16). The standard reaction mixture (in a final volume of 25 l) contains 20 mM Hepes (pH 7.9), 6.5 mM MgCl 2 , 65 mM KCl, 10% (v/v) glycerol, 20 units of RNasin, 30 g of NE, and 1 g of pSH15 (PvuII-or EcoRI-restricted, generating 180-or 284-nucleotide run-off products, respectively). The mixture was incubated for 60 min at 22°C before transcription was initiated by the addition of 1 l of NTP solution containing a final concentration of 500 M each of ATP, GTP, and UTP and 25 M CTP with 7 Ci of [␣-32 P]CTP (800 Ci/mmol). Three min after the addition of NTPs, 0.25% (w/v) sarkosyl was added, and incubation was continued for an additional 42 min. Transcription was stopped with 75 l of stop solution containing 25 mM EDTA (pH 8), 0.2 M Tris-HCl (pH 7.9), and 2% (w/v) SDS. The RNA was extracted with phenol-chloroform, precipitated with ethanol, and resolved on a 6% polyacrylamide, 8 M urea gel. The run-off product was quantitated on a PhosphorImager (Molecular Dynamics).
Southwestern Blot Analysis-For determination of the CREB(s) of RP c by Southwestern blot analysis, 25 l of purified RP c was resolved on 8% SDS-PAGE. Proteins were electrophoretically transferred to ProBlott membrane (Applied Biosystems). The membrane was blocked for 2 h with 5% (w/v) non-fat dry milk dissolved in 1 ϫ TNE buffer (5 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA (pH 8) and 2 mM dithiothreitol). After washing, the membrane was hybridized for 2 h with 32 P-labeled CRE oligonucleotide (approximately 10 5 -10 6 cpm/ml). The membrane was washed three times (30 min each) with 1 ϫ TNE buffer to remove unhybridized oligonucleotides, then exposed to x-ray film for autoradiography.
Western Blot Analysis-NE or purified RP c proteins resolved on SDS-PAGE and transferred to ProBlott membrane were blocked 12 h at 4°C with 5% (w/v) non-fat dry milk dissolved in 1 ϫ phosphate-buffered saline (80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , 100 mM NaCl (pH 7.5)). The membrane was washed and incubated for 2 h at 22°C in 1 ϫ phosphatebuffered saline containing 1% (w/v) non-fat dry milk and 0.1% (v/v) Tween 20 with CREB-1 (C-21) polyclonal antibody (Santa Cruz Biotechnology) for total CREB-1 estimation or with anti-phospho-CREB polyclonal antibody for Ser 133 -phosphorylated CREB-1 estimation. After incubation, the membrane was washed three times for 15 min each with the same buffer and incubated with anti-rabbit IgG (Bio-Rad) for 1 h at 22°C. The CREB-1 signal was detected by ECL Western blotting detection reagents (Amersham Corp.).
In Vitro Phosphorylation and Dephosphorylation of CREB-The recombinant CREB-1 was expressed by transforming BL21 cells with the pET-3A-CREB plasmid (18). The purification procedure was the same as described by Gonzalez et al. (18) with some modifications. After harvesting, the cells were sonicated, and the clear lysate was precipitated with 40% (w/v) ammonium sulfate. A heparin column purification step was included after ammonium sulfate precipitation. The active fractions, eluted over a 0.5-1 M KCl gradient, were pooled, dialyzed, and passed over a DNA-cellulose column. After washing the column, bound CREB-1 was eluted with 0.5 M KCl and purified on a CRE oligonucleotide affinity column. The purified CREB-1 was phosphorylated in a 100-l reaction mixture containing 5 g of CREB-1 (100 pmol), 25 mM Tris-HCl (pH 7.5), 20 mM MgCl 2 , 1 mM ATP, 6 mg/ml dithiothreitol, 10% (v/v) glycerol, and 100 units of the catalytic subunit of PKA (Sigma). The sample was incubated at 37°C for 30 min. To prepare dephosphorylated CREB-1, PKA-phosphorylated CREB-1 (2 pmol) was incubated with 2 units of (Boehringer Mannheim) at 37°C for 1 h. Phosphorylated and dephosphorylated CREB-1 was purified on an ATF/CREB oligonucleotide affinity column, dialyzed, and stored at Ϫ70°C for further analysis. To dephosphorylate CREB from control and MNNG-treated cells, 50 g of NE was incubated with 20 units of calf intestinal phosphatase at 37°C for 1 h.

Recruitment of Closed Preinitiation Complex Is Stimulated by MNNG Treatment-
In previous experiments, we used a fixed saturating concentration of promoter DNA (12 nM) to examine transcription by NE from control or MNNG-treated HeLa cells (16). We found an ϳ10-fold greater overall rate of run-off transcript formation with NE from MNNG-treated cells than with NE from control cells (16). A similar result is illustrated in Fig. 1A where the pSH15 concentration dependence of run-off transcript formation was examined. The RP c was assembled with NE from control or MNNG-treated cells, and after a preincubation to allow RP c assembly, single-cycle runoff transcript formation was measured (Fig. 1A, left panel). To compare the rate of RP c assembly for the two NEs, the preincubation was performed for different periods with a limiting amount of promoter. Aliquots from the preincubation mixture were taken at different times, and the amount of RP c assembled was measured by single-cycle run-off transcript formation. The data are plotted in Fig. 1A (right panel). The computerderived curves fitting the data indicated that the rate constants of RP c formation with control cell NE and MNNG-treated cell NE were similar, yet the amount of RP c formed with the MNNG-treated cell NE was much greater than that formed with the control cell NE. From these results it appeared that the amount, rather than the rate constant, of RP c assembly was stimulated by MNNG treatment of cells.
Purification of Transcriptionally Active RP c -In our previous studies, a NE was used for RP c assembly. Under these conditions, nuclear factors or other nuclear proteins not involved in RP c assembly are also present in the incubation mixture, potentially complicating analysis of factors assembled on the promoter. Since RP c dissociation is very slow under the conditions used (10), we applied a technique with immobilized promoter DNA to purify preassembled RP c (19,20). A similar approach was used by Zawel et al. (21), among others, to describe the fate of transcription factors during RNA polymerase II transcription initiation and elongation. Biotinylated pSH15 was incubated with NE from control or MNNG-treated cells to assemble RP c , which was then purified by use of magnetic streptavidin-agarose beads as described in "Experimental Procedures." In vitro transcription with this purified RP c revealed that the correctly initiated single cycle run-off transcript (284 nucleotides) was formed with RP c assembled with NE from both control and MNNG-treated cells (Fig. 1C). The amount of transcript formed by purified RP c from MNNG-treated cells was greater than that with RP c from control cells. Since transcription was initiated with purified RP c and was limited to a single cycle, these results indicated that the amount of RP c assembled was stimulated by MNNG treatment. The results, therefore, corroborated the findings in Fig. 1A.
CREB-1 in RP c -Previously, we reported that pSH15 promoter activity is dependent upon CRE and that the enhanced transcriptional activity of the MNNG-treated cell NE is mediated through CREB (16). Thus, CREB purified from MNNGtreated cells was transcriptionally more active than CREB from control cells (16). However, it was unclear which member of the CREB superfamily is the target of MNNG treatment. In the present study, using purified RP c assembled from NE of control and MNNG-treated cells, experiments were performed to identify and characterize the RP c -associated CREB. DNA binding analysis of proteins in purified RP c used Southwestern blotting with a 32 P end-labeled CRE oligonucleotide as a probe. The same amount of NE was used in the preparation of RP c from either control or MNNG-treated cells. Results from these experiments are presented in Fig. 2A; lanes 1 and 2 contained proteins from control and MNNG-treated cells, respectively. In lanes 3 and 4 recombinant CREB-1 and ATF-1, respectively, were run as controls. A 43-kDa polypeptide, similar in size to CREB-1, was the major CRE DNA-binding species detected. Also, the amount of 32 P-CRE oligonucleotide binding activity was similar with purified RP c proteins from control and MNNG-treated cell NE ( Fig. 2A). Protein-DNA interactions were also studied by incubation of proteins isolated from purified RP c , assembled from NE of control and MNNG-treated cells, with a 32 P end-labeled oligonucleotide probe and subsequent resolution of the protein-DNA complexes by electrophoretic mobility shift assay (data not shown). The CRE oligonucleotide binding was specific, as an excess of unlabeled CRE oligonucleotide competed binding, whereas a nonspecific oligonucleotide did not. The amount of CRE oligonucleotide binding activity with protein isolated from each RP c was similar by this gel shift assay. The presence of 43-kDa CREB-1 in RP c was confirmed further by Western blot analysis. CREB-1 specific antibody detected a 43-kDa protein in both the NE (lanes 1 and  2) and in purified RP c (lanes 3 and 4) assembled from control or MNNG-treated cells (Fig. 2B). It appeared from these results that CREB-1 was present in a similar amount in purified RP c assembled from each NE.
MNNG-induced Phosphorylation of 43-kDa CREB-1-CREB purified from MNNG-treated cells is known to have greater transcriptional activity than that from control cells (16). It is also known that the MNNG-induced transcriptional up-regulation is mediated through the PKA signal transduction pathway and that an intact CRE in the ␤-pol promoter is required for the response (15; for review, see Ref. 22). CREB is known to be phosphorylated by PKA in vivo, and this phosphorylation is critical for transcriptional activation of some CREB-dependent genes (23). Phosphorylation of CREB does not affect its affinity for strong CRE sites (for review, see Ref. 24) such as the site in the ␤-pol promoter. We examined phosphorylation of CREB in response to MNNG treatment of cells metabolically labeled with [ 32 P]orthophosphate with or without simultaneous treatment with MNNG. RP c was eventually purified from NE from 32 P-labeled control and MNNG-treated cells, and RP c DNAbinding proteins were purified further on a CRE oligonucleotide affinity column. The affinity column-bound proteins were resolved by SDS-PAGE (Fig. 3). The results indicated that phosphorylation of the 43-kDa CREB-1 was much higher in RP c from MNNG-treated cells (duplicate lanes 3 and 4) than in RP c from control cells (duplicate lanes 1 and 2).
Previous studies have shown that the PKA signal transduction pathway is required for MNNG-induced ␤-pol promoter up-regulation (14,15) and that PKA phosphorylates CREB-1 at Ser 133 (23). In the present study, we examined whether the increased CREB-1 phosphorylation after MNNG treatment (Fig. 3) involved Ser 133 . We used an anti-phospho-CREB polyclonal antibody (25), which is specific for recognition of Ser 133 phosphorylation, to probe the phosphorylated form of CREB-1 in NE by Western blot analysis. The antibody produced a CREB-1 signal that was much stronger in NE from MNNGtreated cells (Fig. 4, lane 2) than from control cells (Fig. 4, lane 1). To examine further the MNNG-induced phosphorylation of CREB-1 and to confirm the specificity of the anti-phospho-CREB polyclonal antibody, NEs were treated with calf intestinal phosphatase and probed with the polyclonal antibody. The results showed a strongly diminished signal in the NE from MNNG-treated cells after calf intestinal phosphatase treatment (Fig. 4, lane 4; compare with lane 2). These results indicate that MNNG treatment induces CREB-1 phosphorylation at Ser 133 . To confirm further that the protein recognized by anti-phospho-CREB polyclonal antibody was CREB-1, NEd was prepared and used to perform the Western blot analysis. That antibody did not show a phosphorylated CREB-1 signal with NEd (Fig. 4, lanes 5 and 6) suggested that all of the phosphorylated CREB-1 was removed from the NE by CREB oligonucleotide affinity column depletion. The depletion of CREB from the NE was confirmed by electrophoretic mobility shift assay (data not shown). With NE from control cells (Fig. 4, lane 1) the very low or undetectable signal observed with this antibody further suggests that phospho-CREB-1 was present in a limiting amount in control cells.
Recombinant CREB-1 Phosphorylated at Ser 133 Mimics the MNNG Effect and Restores Transcriptional Activity in NEd-Recombinant CREB-1 was purified and phosphorylated with PKA as described under "Experimental Procedures." PKA phosphorylation of CREB-1 at Ser 133 was measured by Western blot analysis with anti-phospho-CREB polyclonal antibody (Fig. 5A, lane 5). Unphosphorylated CREB-1 (recombinant or phosphatase-treated) is not cross-reactive with this antibody (Fig. 5A, lanes 4 and 6). After observing that phosphorylation of RP c -associated CREB-1 is induced by MNNG treatment, we examined the effect of CREB-1 on transcriptional activity of pSH15 (Fig. 5B). A dose-dependent increase in transcript formation was observed. The NE from control cells supplemented with 10 ng of phosphorylated CREB-1 had transcriptional activity similar to that of NE from MNNG-treated cells (striped bar). On the other hand, the same NE supplemented with 10 ng of unphosphorylated CREB-1 showed promoter activity (open bar) that was the same as control NE alone. These results are consistent with the idea that phosphorylated CREB-1 is limiting in the control cell NE and that phosphorylation of CREB-1 is stimulated by MNNG treatment, in turn resulting in upregulation of the ␤-pol promoter activity. To test this idea further, we first prepared NEd and then assembled RP c with either pSH15 or the wild type ␤-pol promoter (p␤P8) in the presence or absence of unphosphorylated or Ser 133 -phosphorylated CREB-1. p␤P8 was included in this experiment for comparison. The results from these experiments are shown in Fig.  6, and run-off transcripts generated with pSH15 (180 nucleotides) and p␤P8 (210 nucleotides) are indicated by arrows. Run-off transcript formation with control NE (lane 1) was higher than with NEd (lane 2) with both templates (panels A and B). This observation was consistent with previous findings that NEd exhibits only a low basal level of transcriptional activity (16). Addition of unphosphorylated CREB-1 to NEd (lane 3, panels A and B) did not restore promoter activity. However, addition of PKA-phosphorylated CREB-1 restored transcriptional activity of both promoters. These results demonstrate that phosphorylated CREB-1, but not unphosphorylated CREB-1, is required for ␤-pol promoter transcriptional activity.

DISCUSSION
The present study describes characterization of a CREB family member that is involved in ␤-pol gene regulation, as well as a mechanism for transcriptional activation of ␤-pol gene expression by an alkylating agent in vivo. Previous work, with the cloned human ␤-pol promoter, indicated that the PKA signal transduction pathway plays a required role in transcriptional activation after exposure of cells to MNNG (12,13,15) and that the CRE site of the ␤-pol promoter is required for the MNNG response (15). Recently, we extended these observa-  5 and 6). CREB phosphorylated at Ser 133 was estimated by anti-phospho-CREB polyclonal antibody. The phosphorylated CREB-1 signal is marked with an arrow. Data are representative of two independent experiments. tions using an in vitro transcription assay with HeLa NE and a chimeric ␤-pol promoter (10) and found that NE from cells exposed to MNNG is transcriptionally more active than that from control cells. This transcriptional activation is mediated by CREB (16). Using a kinetic model of transcriptional initiation, it was shown that CREB from MNNG-treated cells stimulates formation of more RP c than CREB from control cells. The present studies were designed to purify RP c assembled with pSH15 promoter with NE from control and MNNGtreated cells and then to characterize CREB in the two complexes. We first confirmed that the amount of RP c assembled, instead of the rate of assembly, was increased after MNNG treatment. The apparent rate of RP c assembly was not altered by MNNG treatment. Immobilized promoter DNA was used to purify RP c . A larger amount of single cycle run-off transcript was formed by purified RP c from MNNG-treated cells than from the same number of control cells, indicating that more RP c was formed from MNNG-treated cells. Purified RP c was used as starting material to characterize the CREB family members present. RP c assembled from both NEs had a 43-kDa protein, similar to CREB-1 in mass. The identity of this protein was confirmed by DNA binding analysis and Western blotting (Fig.  2). We found that the amount of CREB in RP c from MNNGtreated cells was similar to that in RP c from control cells. These results are consistent with previous findings (15,16) and suggested that a physical modification of CREB secondary to MNNG treatment may account for the transcriptional up-regulation.
Phosphorylation of CREB-1 by PKA may be indispensable for CREB transcriptional activation (23,26,27). The phosphorylation of CREB-1 by PKA does not affect the protein's dimerization or DNA binding affinity for strong CRE sites (26). However, in some cases PKA phosphorylation enhances binding of CREB-1 to asymmetrical CREs, such as that in the tyrosine aminotransferase gene (TGACGCAG) (28; for reviews, see Refs. 29 and 30). Since the MNNG transcriptional response is mediated through the PKA pathway (15), and CREB purified from MNNG-treated cells is transcriptionally activated (16), it was reasonable to propose that the transcriptional activation is due to increased phosphorylation. We examined this possibility by metabolically labeling cells with [ 32 P]orthophosphate and simultaneously treating them with MNNG. 32 P-Labeled CREB was purified from RP c and resolved by SDS-PAGE. A 43-kDa protein identical in mass to CREB-1 was much more strongly phosphorylated after exposure of cells to MNNG than in control cells. Using an anti-phospho-CREB polyclonal antibody that is specific for the Ser 133 -phosphorylated form of CREB-1, we further confirmed that phosphorylation of NE CREB-1 at Ser 133 was strongly increased by MNNG treatment.
To test the functional activity of CREB and to attempt to mimic the MNNG response, we used a mixed activator approach in which PKA-phosphorylated recombinant CREB-1 was added to the NE from control cells. A very small amount of added phosphorylated CREB-1 (10 ng) (but not unphosphorylated CREB-1) was able to stimulate promoter activity. The role of phosphorylated CREB-1 in transcriptional up-regulation of the ␤-pol promoter was examined further with NEd. Phosphorylated CREB-1 could restore transcriptional activity of NEd, but unphosphorylated CREB-1 could not. The results indicate that exposure of cells to the alkylating agent MNNG induces phosphorylation of CREB-1, which in turn recruits more RP c onto the promoter. This effect results in a stimulation of the overall amount of transcript formation. These observations are similar to those proposed for transcriptional activation of c-Jun by phorbol ester-induced phosphorylation of the transactivation domain (31). The importance of PKA-phosphorylated CREB-1 in vivo was described in a transgenic mouse model study by Struthers et al. (32), where expression of a phosphorylation-deficient mutant CREB impaired transcriptional activation of the growth hormone gene and produced growth abnormalities. The enhanced transactivation property of phosphorylated CREB also has been examined by nuclear microinjection in fibroblasts; the results showed that phosphorylated CREB, but not unphosphorylated CREB, stimulated CREBdependent gene expression (33). The transcriptional activation of target genes by the Ser 133 -phosphorylated form of CREB has been proposed to be due to increased association with the phospho-CREB-binding protein (CBP) (34). Further, in transient expression studies, overexpression of CBP can stimulate cAMP-responsive promoter activity in a CREB phosphorylationdependent manner (35). CREB phosphorylation, however, has not always been found to mediate activation of CREB-dependent promoters. In several studies (33,36,37) with in vitro transcription system, no difference in transactivation properties was found with phosphorylated and unphosphorylated CREB. In some reports, it was suggested that the failure to observe phosphorylated CREB-dependent activation of genes was due to rapid dephosphorylation of CREB by nuclear phosphatases (37,38). Recent studies of Brindle and Montminy (24) suggested that, in addition to Ser 133 -phosphorylated CREB, other proteins are required to mediate CREB⅐CBP complex formation and transcriptional activation of target genes. Data from our studies indicate that phosphorylation of CREB at Ser 133 by MNNG treatment of cells, perhaps in association with CBP, correlates with ␤-pol gene activation in the HeLa cell system. Explanations for the differences between these results and those from some other studies are unknown at the moment but could be related with differences in promoter sequence context and as well as details of extract preparation.
In summary, we have made use of in vitro transcription systems to understand a potential mechanism for up-regulation of the ␤-pol promoter after mammalian cells are treated with a monofunctional DNA-alkylating agent. Since this upregulation appears to require DNA alkylation damage (39), it should be interesting to study the linkage between DNA damage and the signal transduction system.