CUTL1 Is Phosphorylated by Protein Kinase A, Modulating Its Effects on Cell Proliferation and Motility*

CUTL1, also known as CDP (CCAAT Displacement Protein), Cut, or Cux-1, is a homeodomain transcription factor known to play an essential role in development and cell cycle progression. Previously, we identified CUTL1 as modulator of cell motility and invasiveness. Here we report that protein kinase A (PKA), known to inhibit tumor progression in various tumor types, directly phosphorylates CUTL1 at serine 1215 in NIH3T3 fibroblasts. The PKA-induced phosphorylation results in decreased DNA binding affinity of CUTL1 and diminished CUTL1-mediated cell cycle progression and cell motility. Furthermore, the expression of several CUTL1 target genes involved in proliferation and migration, such as DNA polymerase A and DKK2, was modulated by PKA-induced phosphorylation. These data identify CUTL1 as a novel target of PKA through which this protein kinase can modulate tumor cell motility and tumor progression.

Protein kinase A (PKA) 2 belongs to the family of serine/threonine kinases whose activity is implicated in the regulation of a wide variety of cellular functions, such as cell metabolism, secretion, proliferation, differentiation, and neoplastic transformation (1). PKA consists of two separate subunits, the catalytic (C) and regulatory (R) subunits that interact to form an inactive holoenzyme complex (2). Activation of PKA is achieved by binding of the second messenger cAMP to the R subunit, which consequently induces a conformational change in the R subunit and leads to the dissociation of the holoenzyme into its constituent subunits (1,2). The free active C subunit is then able to phosphorylate a variety of cytoplasmic and nuclear protein substrates, including transcription factors (1). cAMP and protein kinase A have been shown to play a major role in the modulation of tumor progression. However, these effects appear to be dependent on the cellular context (3). cAMP is known to have antiproliferative effects by antagonizing ERK activation via PKA in a variety of cell systems including fibroblasts, glial cells, and lymphocytes (4). Furthermore, cAMP has been shown to inhibit cell motility and invasiveness of several epithelial cancers (5,6). In several neuronal, endocrine, and other epithelial cells, however, cAMP is able to stimulate growth by activating ERKs (7)(8)(9). CUTL1, also known as CDP (CCAAT displacement protein), belongs to a family of homeobox transcription factors involved in the regulation of cell growth and differentiation (10). It is evolutionarily conserved and contains four DNA binding domains, three of which are known as Cut repeats and one as a Cut homeodomain (11). CUTL1 has been described as a transcriptional activator as well as a transcriptional repressor. Its activity has been associated with cellular proliferation and cell cycle progression (12)(13)(14) as well as modulation of genes involved in terminal differentiation (15,16). Accordingly, knockout studies with the murine homologue Cux-1 revealed reduced growth, retarded differentiation of the lung epithelia, hair follicle defects, reduced male fertility, and deficient T and B cell function (17)(18)(19). In contrast, mice transgenic for Cux-1 showed organomegaly and multiorgan hyperplasia (20).
CUTL1 has been demonstrated to be a phosphorylation target of several kinases that modulate its DNA binding affinity: phosphorylations by cyclin A/cdk1 and by protein kinase C at various sites have been shown to inhibit DNA binding activity (21,22). Studies in flies indicate that the Drosophila homologue of CUTL1, Cut, is a major determinant of cell type specification downstream of the notch pathway (23) and may also interact with the wingless signaling pathway (24).
Recent data indicate that CUTL1 plays a major role in tumorigenesis and tumor progression. Although initial data based on loss of heterozygosity studies suggested CUTL1 as a candidate tumor suppressor gene (25), a recent report suggests instead that CUTL1 could play a role in promoting breast cancer: Goulet et al. (26) describe a tissue-specific CUTL1 isoform that appears to be strongly expressed in some breast tumors and, when overexpressed, inhibited tubule formation of breast cancer cells in vitro. In addition, we have demonstrated that CUTL1 strongly promotes cell motility and invasiveness in vitro and in vivo in a variety of cell systems and that CUTL1 expression is associated with a less differentiated phenotype and worsened prognosis in breast cancer (27).
Based on our previous data, we sought to investigate whether the transcriptional activity of CUTL1 is modulated by upstream signaling pathways implicated in the regulation of cell proliferation and motility. We could demonstrate that CUTL1 is phosphorylated by protein kinase A, leading to a decrease in DNA binding affinity and transcriptional activity. Furthermore, we were able to show that CUTL1 phosphorylation by PKA affected expression of several of its downstream target genes and inhibited CUTL1-mediated proliferation and migration.
Flow Cytometry-Fluorescence-activated cell sorter analysis was performed as described previously (29).
Reverse Transcription PCR-Reverse transcription was performed using the Superscript first strand synthesis kit (Invitrogen). PCR was performed for 30 cycles. Primer sequences are available on request. Quantitative real-time PCR analyses using the comparative C T method were performed on an ABI PRISM 7700 sequence detector system using the SYBR Green PCR Master Mix kit (PerkinElmer, Applied Biosystems) according to the manufacturers' instructions. Following initial incubation at 50°C for 2 and 10 min at 95°C, amplification was performed for 40 cycles at 95°C for 15 s and 60°C for 1 min. Specific primer pairs were determined with the PrimerExpress program (Applied Biosystems). The murine cyclophilin A gene was used as the internal standard. Primer sequences are available on request.
Luciferase Reporter Assays-Cells were transiently transfected using GeneJuice (Merck Biosciences) according to the manufacturer's instructions and harvested 24 h after transfection. Because the internal control plasmid is itself often repressed by CDP/Cux, as a control for transfection efficiency the purified ␤-galactosidase protein (Sigma) was included in the transfection mix as previously described (14,30). The luciferase assays were performed using the Luciferase Assay System (Promega, Madison, WI) according to the manufacturer's instructions. ␤-galactosidase activity was detected using 2-nitrophenyl ␤-D-galactopyranoside (ONPG; Sigma) in substrate buffer (1 mM MgCl 2 , 1 mg/ml ONPG, 75 mM sodium phosphate buffer, pH 7.4, 60 mM ␤-mercaptoethanol) at 420 nm.
Wound Healing Assays-An artificial "wound" was created using a 10-l pipette tip on confluent cell monolayers in 6-well culture plates. Photographs were taken at 0 and 8 h. Quantitative analysis of the wound closure was calculated by counting the number of cells/10 5 m 2 wound area at 8 h.
Two-chamber Migration Assays-Cell migration was determined by using the modified two-chamber migration assay (8-m pore size; BD Biosciences) according to the manufacturer's instructions. 2.5 ϫ 10 4 cells were seeded in serum-free medium into the upper chamber and migrated toward 10% fetal calf serum as a chemoattractant in the lower chamber for 18 h. Cells in the upper chamber were carefully removed using cotton buds, and cells at the bottom of the membrane were fixed and stained with crystal violet 0.2%/methanol 20%. Quantification was performed by counting the stained cells.
Video Time-lapse Microscopy-Time-lapse imaging of migrating cells was performed on an inverted Zeiss Axiovert 135TV microscope (X-Y-Z motorized stage/piezo Z control; Zeiss, Jena, Germany) over 24 h at 37°C/10% CO 2 . Images were obtained with a Hamamatsu Orca ER firewire CCD camera every 7 or 20 min and analyzed using image analysis software (AQM Advance6 and Tracker; Kinetic Imaging Ltd, Bromborough, UK). Migration of each cell was analyzed by measuring the distance traveled by a cell nucleus over the 24-h time period. Statistical analysis was performed using the Mathematica 4.2 software (Wolfram Research, Long Hanborough, UK). The average migration speed in m/min was calculated by analyzing at least 24 cells/group.
Electromobility Shift Assays-Gel shift assays were performed as described previously (31). The consensus and mutant CDP binding oligonucleotides were obtained from Santa Cruz Biotechnology. 4 g of nuclear extract from cells transiently transfected with CUTL1-(831-1336) WT or S1215A mutant were incubated in binding buffer (125 mM NaCl, 50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 25 mM EDTA, 25% glycerol, 5 mM dithiothreitol, 6 g/ml of bovine serum albumin, 5 g/ml of poly(dI-dC)). For supershift assays, 6 g of rabbit anti-CUTL1/CDP antibody (Santa Cruz supershift reagent) was included. The end-labeled probe was then added to each reaction and subsequently loaded onto a 4% non-denaturing polyacrylamide gel. Dried gels were scanned in a PhosphorImager (Storm 860; Amersham Biosciences).

RESULTS
CUTL1 Is Phosphorylated by PKA-By screening CUTL1 for putative phosphorylation sites using Scansite (scansite.mit.edu), we found several putative phosphorylation sites for PKA and protein kinase B, also known as Akt. One of these sites (MKRRHSSVS at position 1210-1218) is evolutionarily conserved across species and was predicted to be a target site for PKA (phosphorylation of serine 1215) and for protein kinase B (phosphorylation of serine 1216). Initially, we wanted to test whether CUTL1 is phosphorylated by protein kinase B at serine 1216. Therefore, we used an antibody that was designed to recognize phosphorylated serine/threonine motifs of protein kinase B substrates. We and others have found, however, that this antibody has broader specificity for several phospho-serine/threonine motifs containing upstream basic residues (32). We found that the antibody recognized CUTL1 phosphorylation induced by dibutyryl-cAMP or forskolin as PKA stimulators (Fig. 1a). Activation of PKB did not result in CUTL1 phosphorylation (data not shown). Stimulation with cAMP induced a rapid phosphorylation of CUTL1, detectable by Western blot after 15 min with a peak at 45 min (Fig. 1b).
Because PKA-independent effects of cAMP and forskolin have been described (1), we aimed to examine direct involvement of PKA in the cAMP/forskolin-induced phosphorylation. As a first step, we used the pharmacological PKA inhibitor H89. Coincubation with H89 decreased the cAMP-induced phosphorylation markedly (Fig. 2a). In addition, cotransfection with dominant negative PKA or with the inhibitory peptide PKI (28) inhibited cAMP-induced phosphorylation (Fig. 2b), indicating that the CUTL1 phosphorylation is mediated by PKA and is not a PKA-independent phosphorylation event induced by cAMP and forskolin. Immunoprecipitation with an antibody designed to detect phosphorylated PKA substrates revealed a detectable CUTL1 immunoblot signal upon cAMP stimulation (Fig. 2c), showing that the endogenous CUTL1 protein is phosphorylated following activation of PKA. To further rule out possible involvement of other serine/threonine kinases that might be able to phosphorylate similar motifs, we used several drugs such as the mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK) inhibitor UO126 acting upstream of p90RSK, the mTOR inhibitor rapamycin acting upstream of p70S6K, and the phosphatidylinositol 3-kinase inhibitor LY294002. None of these inhibitors was able to decrease cAMP-induced CUTL1 phosphorylation to a significant extent, indicating that involvement of the serine/threonine kinases p90RSK and p70S6K is unlikely and that protein kinases controlled by phosphatidylinositol 3-kinase are also not involved (Fig. 2d).
PKA Phosphorylates CUTL1 at Ser-1215-To investigate which residue serves as the actual phosphorylation site for PKA, we generated serine to alanine mutants by site-directed mutagenesis at codons 1215 and 1216 as well as threonine to alanine mutants at codons 45 and 1454, which were also predicted as putative PKA sites by Scansite. Using these constructs, we determined that cAMP stimulation is unable to induce any detectable phosphorylation of the S1215A mutant (Fig. 3a).
To further confirm the PKA-mediated phosphorylation of CUTL1 at Ser-1215, we generated a phospho-specific peptide antibody against the motif around the phosphorylated Ser-1215. Using this antibody, we saw cAMP-inducible phosphorylation that was only detectable in WT CUTL1 but was absent in CUTL1 S1215A (Fig. 3b). This antibody detected cAMP-induced phosphorylation both in Western blots using immunoprecipitated Myc-tagged CUTL1 and in cAMP-stimulated whole-cell lysates with endogenous CUTL1, as demonstrated in Fig. 3c.
PKA Decreases CUTL1-mediated Proliferation and Migration-To study the transcriptional effects of CUTL1, we have previously suppressed CUTL1 expression by vector-based transfection of specific shRNA in NIH3T3 fibroblasts, which express high endogenous CUTL1 levels. Stable CUTL1 RNAi clones that showed no detectable CUTL1 expression by Western blot were used for further studies in comparison with empty vector-transfected control clones (27). Using this cell system as well as other cell lines, we previously showed that CUTL1 expression is strongly associated with increased cell migration and invasiveness in vitro and in vivo (27). Furthermore, CUTL1 has been shown to enhance proliferation by stimulating S-phase progression in a variety of cell lines (13).
To test the biological significance of the CUTL1 modulation by PKA, we first examined the effect of CUTL1 on proliferation in the presence or absence of PKA activation. Suppression of CUTL1 by shRNA significantly reduced proliferation as measured by cell counting (Fig. 4a), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays, and bromodeoxyuridine enzyme-linked immunosorbent assay (data not shown). cAMP significantly decreased proliferation in the presence of endogenous CUTL1. However, it did not further affect the reduced proliferation in the absence of CUTL1 (Fig. 4a). The CUTL1  Cells were incubated with H89 for 1 h before stimulation with dibutyryl-cAMP. b, dominant negative PKA abolishes dibutyryl-cAMP-induced phosphorylation of CUTL1. Cells were cotransfected with a dominant PKA construct (dnPKA) or PKA inhibitory peptide (PKI) before stimulation with dibutyryl-cAMP. c, endogenous CUTL1 can be detected by immunoprecipitation using an antibody directed against phosphorylated PKA substrates after stimulation with dibutyryl-cAMP. d, S6K, RSK, and phosphatidylinositol 3-kinase are not involved in cAMP-induced CUTL1 phosphorylation. Cells were coincubated with various inhibitors for 1 h before stimulation with cAMP.
shRNA-mediated decrease in proliferation was paralleled by a decrease in S-phase progression as demonstrated by fluorescence-activated cell sorter analysis (Fig. 4b). cAMP reduced S-phase progression in wildtype cells to a significant extent but had only minor effects on S-phase progression in the absence of CUTL1. Based on these data, PKA appears likely to exert at least a significant part of its anti-proliferative effects through CUTL1.
Besides its effect on CUTL1 phosphorylation, PKA is known to decrease ERK activity in several cell systems. In the NIH3T3 cells used, PKA activation reduced basal ERK activity (Fig. 4c). However, ERK activation by epidermal growth factor (EGF) did not affect cAMP-induced phosphorylation of CUTL1 at Ser-1215 as detected by the phosphospecific CUTL1 antibody (Fig. 4d). This suggests that activation of ERK does not affect the transcriptional activity of CUTL1 by decreasing PKA-induced phosphorylation of CUTL1 at Ser-1215.
Given our previous results on CUTL1 as an enhancer of cell migration (27), we examined as a next step the effect of PKA activation on CUTL1-induced cell migration. For that purpose, we employed three different assays examining different aspects of cell motility: modified Boyden chamber assays, wound healing assays, and time-lapse video microscopy, which measures cell migration independent of a proliferation bias. As demonstrated above for proliferation, addition of cAMP significantly decreased the migration of NIH3T3 control cells. In contrast, the effect of cAMP on the migration of CUTL1 RNAi cells was markedly less pronounced. These results were obtained by two-chamber migration assays with a chemotactic gradient from serum-free to serum-containing medium (mean migrated cells/ field Ϯ S.E.: control ϩ cAMP: 52 Ϯ 7 cells, control: 75 Ϯ 9 cells, CUTL1 RNAi ϩ cAMP: 23 Ϯ 5 cells, CUTL1 RNAi: 25 Ϯ 5 cells) (Fig.  5a), by time-lapse microscopy (Fig. 5b) and by wound healing assays (Fig. 5c) and were also confirmed in HT1080 cells (data not shown). These data suggest that PKA mediates its anti-migratory effects to a significant extent through CUTL1.
PKA Decreases CUTL1 Transcriptional Activity-To investigate the effect of PKA on CUTL1 transcriptional activity at a single-gene level, we examined the transcriptional regulation of the known CUTL1 target gene DNA polymerase A (14) and several other CUTL1 target genes, which we previously identified by microarray experiments (27), in the FIGURE 3. Protein kinase A phosphorylates CUTL1 at serine 1215. a, cAMP is unable to phosphorylate the S1215A CUTL1 mutant. Cells were transfected with either wild-type or S1215A mutated CUTL1 before stimulation with dibutyryl-cAMP. b, cAMP stimulates CUTL1 wild-type, but not S1215A mutant, CUTL1 phosphorylation as detected by anti-phospho-CUTL1 in Western blots of immunoprecipitated CUTL1 and whole-cell lysates. c, cAMP stimulates phosphorylation of endogenous CUTL1 as detected by anti-phospho-CUTL1 antibody. compared with CUTL1 RNAi cells; **, p Ͻ0.05 compared with control cells. c, EGF induces and cAMP decreases ERK phosphorylation in NIH3T3 cells. NIH3T3 cells were stimulated with EGF and/or dibutyryl-cAMP for 15 min. Activating phosphorylation of ERK was detected via immunoblot using a specific phospho-ERK antibody. A parallel blot was probed with anti-ERK antibody as control. Data are representative of three independent experiments. d, EGF does not alter cAMP-induced CUTL1 phosphorylation at serine 1215. NIH3T3 cells were stimulated with EGF and/or cAMP for 15 min, and phosphorylation of CUTL1 at serine 1215 was detected with the phospho-specific CUTL1 antibody. A parallel blot was probed with anti-CUTL1 antibody as loading control. Data are representative of three independent experiments. presence or absence of PKA stimulation. To enhance the transcriptional activation of CUTL1, we used the C-terminal fragment of CUTL1, amino acids 831-1336, which mimics the product of cleavage of fulllength CUTL1 by cathepsin L, which has enhanced transcriptional activity (33). By reporter assay, we could confirm the negative effect of PKA on overexpressed CUTL1 activity at the promoter level for the CUTL1 target DNA polymerase A, which is known to be involved in DNA synthesis and S-phase progression (Fig. 6a). In addition, we examined by real-time PCR the expression levels of several endogenous CUTL1 target genes previously identified by microarray experiments, such as DKK2, DAB2, and DPYSL3 (27), with or without PKA activation. The majority of the genes tested showed a marked decrease of the CUTL1-induced up-regulation by PKA. In the absence of CUTL1 the effect of PKA was much less pronounced (Fig. 6b). To examine PKAdependent alterations in the affinity of CUTL1 for its DNA binding site, we performed electrophoretic mobility shift assay studies in the presence or absence of PKA activity by analyzing overexpressed CUTL1 WT or the non-phosphorylatable S1215A mutant with or without PKA inhibitor H89. CUTL1 binding activity increased markedly upon PKA inhibition in CUTL1 WT nuclear extracts. In CUTL1 S1215A nuclear extracts, the basal binding activity was increased as compared with WT extracts, and addition of H89 was not able to further increase the binding affinity, indicating that PKA decreases CUTL1 binding activity via phosphorylation at Ser-1215 (Fig. 6c). To assess the PKA-dependent effect of the S1215A mutant on migration, we performed time-lapse migration experiments with transiently transfected CUTL1 WT or S1215A mutants in NIH3T3 cells. PKA activation decreased cell migration only in cells transfected with CUTL1 WT, but not in cells transfected with CUTL1 S1215A (Fig. 6d). This further confirms that PKA inhibits CUTL1 transcriptional activity by phosphorylation of Ser-1215, with subsequent effects on cell motility.
In addition to the phosphorylation at Ser-1215 by PKA, previous reports described the adjacent residues Ser-1237 and Ser-1270 as targets for cyclin A/cdk1 (22), phosphorylation of which caused inhibition of DNA binding. Therefore, we were interested whether phosphorylation of the adjacent serines 1237 and 1270 modulates PKA action. Using luciferase assays for the CUTL1 target DNA polymerase A, we found that double mutants S1237D/S1270D mimicking phosphorylations at these residues led to reduced transcriptional activity of CUTL1, as expected. PKA activation by forskolin reduced the transcriptional activity of S1237D/S1270D CUTL1 still significantly, reaching levels similar to wild-type CUTL1. This suggests that phosphorylation events in the region between residue 1215 and 1270 induced by PKA or cyclin A/cdk1 have similar effects on the transcriptional activity of CUTL1. They might be induced by PKA and/or cyclin A/cdk1, depending on the cellular context. Phosphorylation at Ser-1237/Ser-1270, however, does not abrogate the ability of PKA to phosphorylate CUTL1 at Ser-1215.

DISCUSSION
In this study, we have demonstrated that the homeodomain transcription factor CUTL1 is phosphorylated at serine 1215 by protein kinase A. This phosphorylation results in reduced DNA binding affinity and decreased CUTL1-mediated proliferation and migration, which was paralleled by a reduced transcription of CUTL1 target genes.
Several other protein kinases have been shown to modulate CUTL1 transcriptional activity by phosphorylation at distinct residues. Based on the observation that Cut-specific DNA binding was increased following phosphatase treatment, Coqueret et al. (21) found that PKC efficiently phosphorylated Cut repeats 1-3 at residues Thr-415, Thr-804, and Ser-987, which inhibited DNA binding. In addition, cyclin A/cdk1 was shown to bind to CUTL1 and modulate its DNA binding activity in vitro and in vivo (22): phosphorylation of serines 1237 and 1270 caused inhibition of DNA binding. In cotransfection studies, cyclin A/cdk1 inhibited CUTL1 stable DNA binding and prevented repression of the p21 WAF1 reporter. In contrast, mutant CUTL1 proteins in which serines 1237 and 1270 were replaced with alanines were not affected by cyclin A/cdk1. The fact that serine 1215, the target residue for PKA identified in this report, and serines 1237 and 1270, the target residues for cyclin A-ckd1, are in close proximity suggests that phosphorylations in this region induce conformational changes that lead to reduced transcriptional activity. All three residues are located between the conserved C-terminal cut repeat 3 and the homeodomain, both of which have recently been reported as the major determinants of CUTL1 activity after proteolytic cleavage of the N-terminal region (34). Despite similar effects on the transcriptional activity of CUTL1 induced by phosphorylations at both Ser-1237/Ser-1270 and Ser-1215, our data indicate that the phosphorylation at Ser-1237/Ser-70 by cyclin A/cdk1 does not abrogate the effect of PKA on Ser-1215, underlining the importance of the phosphorylation event induced by PKA.
Protein kinase A is known to play an important role in tumorigenesis and cancer progression. However, its effect on tumor cell proliferation and migration is dependent on the cellular context. Although in certain cell systems PKA activity induced by cAMP is able to stimulate cell growth by enhancing ERK activity, in many other cell types, including the cells used here, cAMP inhibits ERK activity (4,36) and inhibits cell proliferation. A number of mechanisms for the growth-inhibitory effect of cAMP have been proposed, only some of which involve cross-talk with the ERK pathway. First, PKA blocks Raf-1 activation by direct serine phosphorylation (37)(38)(39)(40)(41). Second, PKA may activate the Rap1 FIGURE 6. cAMP decreases CUTL1 transcriptional activity. a, luciferase assay of DNA polymerase A promoter luciferase construct in NIH3T3 cells with or without transiently transfected CUTL1-(831-1336) Ϯ dibutyryl-cAMP/IBMX for 6 h. Data are representative for three independent experiments and are shown as mean Ϯ S.E. *, p Ͻ0.05 compared with empty vector-transfected control cells; **, p Ͻ0.05 compared with CUTL1-transfected cells without cAMP. b, quantitative real-time reverse transcription PCR of selected CUTL1 target genes identified by previous microarray experiments Ϯ dibutyryl-cAMP/IBMX for 6 h in NIH3T3 control and CUTL1 RNAi cells, normalized to cyclophilin expression. c, electrophoretic mobility shift assay of NIH3T3 cells transfected with CUTL1-(831-1336) WT or S1215A mutant Ϯ PKA inhibitor H89 in the presence of 10% donor calf serum. Data are representative for three independent experiments. d, video time-lapse microscopy of NIH3T3 cells transiently transfected with CUTL1 WT or CUTL1 S1215A mutant Ϯ dibutyryl-cAMP/IBMX. Cells were cotransfected with pEGFP, and fluorescent cells were tracked. *, p Ͻ0.05 compared with CUTL1 WT cells without PKA activation. e, luciferase assay of DNA polymerase A promoter luciferase construct in NIH3T3 cells with or without transiently transfected CUTL1-(831-1336) (CUTL) or CUTL1-(831-1336) S1237D/S1270D double mutant (S1237/1270D) Ϯ forskolin for 6 h. Data are representative for three independent experiments and are shown as mean Ϯ S.E. *, p Ͻ0.05 compared with CUTL1-(831-1336)-transfected cells without forskolin; **, p Ͻ0.05 compared with CUTL1-(831-1336)-S1237/1270D-transfected cells without forskolin.
GTPase via Src, thereby inhibiting ERK in some cell contexts (4). Third, PKA may inhibit Akt activity and thereby activate the FoxO4 transcription factor and induce the CDK inhibitor p27 Kip1 (42). Based on our data, we propose that in addition, PKA activity decreases proliferation and cell motility, and possibly tumor progression, at least in part by phosphorylating CUTL1 and thereby inhibiting its transcriptional activity.
Detailed study of cAMP effects on carcinoma cell motility by Mercurio and co-workers (43) and others has shown that ␣6␤4 integrin-mediated invasion required suppression of cAMP levels by activation of a phosphodiesterase. This was required for RhoA translocation from the cytosol to membrane ruffles (44). Elevated cAMP levels have also been reported to block RhoA activation (45,46). Although it has been suggested that these effects are due, at least in lymphocytes, to direct phosphorylation of RhoA by PKA at Ser-188 (47), our data suggest that control of the transcriptional activity of CUTL1 following phosphorylation by PKA could be of major significance in controlling cell motility. It is possible that inhibitory effects of PKA on the ERK MAP kinase pathway might also contribute to the reduction in cell migration seen in cells treated with cAMP, possibly in a synergistic manner to the CUTL1 pathway. However, we found that ERK activation induced by EGF did not result in a decreased phosphorylation of CUTL1 at Ser-1215. Therefore, a direct effect of the ERK pathway on the phosphorylation of CUTL1 at Ser-1215 is unlikely.
The abilities of tumor cells to proliferate and migrate are essential hallmarks of cancer and prerequisites of local tumor progression and metastatic spread. The data presented here demonstrate that the transcription factor CUTL1, which has previously been shown as modulator of cell motility and invasive behavior in vitro and in vivo, is inhibited by PKA, a major regulatory pathway for tumor progression. We propose that the inhibition of CUTL1-mediated tumor progression by PKA accounts to a significant extent for the anti-tumoral effects of PKA seen in several tumor types such as leukemias, sarcomas, and gliomas expressing high levels of CUTL1 (35, 48 -50). For those tumor types, targeting CUTL1 activity by administering PKA-activating agents may represent a promising treatment strategy.