The expression of retinoblastoma and Sp1 is increased by low concentrations of cyclin-dependent kinase inhibitors Silvia Pen ˜ uelas*, Cristina Alemany*, Ve ´ ronique Noe ´ and Carlos J. Ciudad Department of Biochemistry and Molecular Biology, School of Pharmacy, University of Barcelona, Spain We examined the effect of suboptimal concentrations of cyclin-dependent kinase inhibitors, which do not interfere with cell proliferation, on retinoblastoma expression in hamster (Chinese hamster ovary K1) and human (K562 and HeLa) cells. To achieve this, we used the chemical inhibitors roscovitine and olomoucine (which inhibit CDK2 prefer- entially), UCN-01 (which also inhibits CDK4/6) and p21 (as an intrinsic inhibitor). All chemical inhibitors and over- expression of p21 strongly induced retinoblastoma protein expression. UCN-01-mediated retinoblastoma expression was caused by an increase in both the levels of retino- blastoma mRNA and the stability of the protein. The expression of the transcription factor Sp1, a retinoblastoma- interacting protein, was also enhanced by all the cyclin-dependent kinase inhibitors tested. However, Sp1 expression was caused by an increase in the levels of Sp1 mRNA without modification in the stability of the pro- tein. By using luciferase experiments, the transcriptional activation of both retinoblastoma and Sp1 promoters by UCN-01 was confirmed. Bisindolylmaleimide I, at concen- trations causing a similar or higher inhibition of protein kinase C than UCN-01, provoked a lower activation of retinoblastoma and Sp1 expression. Finally, the effects of cyclin-dependent kinase inhibitors on dihydrofolate reduc- tase gene expression were evaluated. Treatment with UCN- 01 increased cellular dihydrofolate reductase mRNA levels, and dihydrofolate reductase enzymatic activity was enhanced by UCN-01, roscovitine, olomoucine and p21, in transient transfection experiments. These results support a mechanism for the self-regulation of retinoblastoma expression, and point to the need to establish the appropriate dose of cyclin-dependent kinase inhibitors as antiprolifera- tive agents in anticancer treatments. Keywords: retinoblastoma gene product; Sp1; UCN-01; roscovitine; dihydrofolate reductase. Cyclin-dependent kinases (CDKs) are key regulators of cell cycle progression. They constitute the catalytic subunits of holoenzymes formed in combination with regulatory sub- units named cyclins. Thirteen CDKs [1,2] and at least 25 cyclins [2] have been reported to date. Cyclin expression varies during the cell cycle and the cyclin/CDK holoenzyme is activated by phosphorylation of specific residues in the CDK catalytic subunit by the cdk-activating kinase [3,4]. CDKs are involved in transcriptional control [5], mitotic progression [6], DNA repair (CDK7) [7], differentiation of brain neurons (CDK5) [8] and play a crucial role in the progression of cells from G1 to S phase by regulating the phosphorylation state of the retinoblastoma gene product (Rb). The tumor suppressor Rb is a nuclear protein of 928 amino acids [9] that is present in distinct phosphorylation states depending on the phase of the cell cycle [10,11]: it is nonphosphorylated when newly synthesized; hypophos- phorylated in early G1; and hyperphosphorylated in late G1, S, and G2/M phases. In mitosis, a protein phosphatase 1-like protein removes all phosphates from phosphorylated Rb to reset the phosphorylation status of Rb in early G1. The hypophosphorylated form is involved in the growth inhibitory potential of Rb [12,13], which has been related to its capacity to bind and block the activity of the family of transcription factors, E2F [14,15], thus inhibiting the expression of genes that contain the E2F response element in their promoters, e.g. dihydrofolate reductase (DHFR) [16–18], DNA polymerase alpha [19], thymidine kinase [20,21], histone H2A [22], proliferating cell nuclear antigen [23], B-myb [24,25], cyclin D [26,27], cyclin E [28], cyclin A [29,30] and cdc2 [31,32]. Rb is phosphorylated by the action of various combinations of cyclin/CDK complexes, such as cyclin D/CDK4-CDK6 in early G1 and cyclin E/CDK2 in late G1 and G1/S phases. After mitosis, Rb returns to its nonphosphorylated state [10,11]. Cyclin/CDK complexes can be regulated by small inhibitory proteins, known as intrinsic CDK inhibitors, which suppress cell growth. The INK4 CDK inhibitors (p15, p16, p18 and p19) inhibit CDK4 and CDK6, whereas the family of p21, p27 and p57 inhibit or sequester the different known cyclin/CDK complexes [33–35]. Numerous human cancers present abnormalities in some components of the Rb pathway as a result of CDK hyperactivation, decrease in endogenous Correspondence to C. J. Ciudad, Departamento de Bioquı ´ mica y Biologı ´ a Molecular, Division IV, Facultad de Farmacia, Universidad de Barcelona, Avenue Diagonal 643, E-08028 Barcelona, Spain. Fax: + 34 93 402 4520, Tel.: + 34 93 403 4455, E-mail: cciudad@farmacia.far.ub.es Abbreviations: APRT, adenine phosphoribosyltransferase; BSM-I, bisindolylmaleimide I; CDK, cyclin-dependent kinase; CHO, Chinese hamster ovary; DHFR, dihydrofolate reductase; IC 50 ,50%inhibitory concentration; PKC, protein kinase C; Rb, retinoblastoma gene product. *Note: The two first authors contributed equally to this work. (Received 17 June 2003, revised 10 September 2003, accepted 7 October 2003) Eur. J. Biochem. 270, 4809–4822 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03874.x CDK inhibitors or Rb gene mutations. Therefore, the use of pharmacologic CDK inhibitors offers great potential for the treatment of many neoplasms [36]. In this regard, chemical CDK inhibitors, such as UCN-01, roscovitine and olomou- cine, have been developed. Roscovitine and olomoucine are more specific towards CDK2, whereas UCN-01 shows a similar 50% inhibitory concentration (IC 50 )forCDK2and CDK4/6. Flavopiridol (another CDK inhibitor), UCN-01 and roscovitine are currently used in clinical studies on cancer therapy [2,37]. CDK inhibition leads to cell cycle arrest [38–41], apoptotic cell death [39,41,42], differentiation [43–45] and inhibition of angiogenesis [46]. CDK inhibitors also modify the transcript levels of 2–3% of the genes in Saccharomyces cerevisiae, as measured by array methods [47]. Given that the modulation of CDK activity is an attractive target for cancer chemotherapy, we studied the changes produced by low concentrations of different CDK inhibitors at the molecular level, with a special focus on their natural substrate, retinoblastoma. In addition to the primary function of Rb as a transcrip- tional co-repressor in cell cycle regulation, this tumor suppressor protein can function as a transcription co-activator through its physical interaction with selective transcriptional factors such as hBRm, C/EBP, AP2 and Sp1 [48–51]. Rb activates a set of gene promoters, e.g. c-fos, c-myc, transforming growth factor-b1 [52,53], transforming growth factor-b2 [54], c-jun [55], Cyclin D1 [26], thymidine kinase and dihydrofolate reductase (DHFR) [53] that control the cell cycle through stimulation of Sp1 mediated transcription. Furthermore, the Rb promoter contains potential binding sites for transcription factors such as ATF-2, Sp1 and RBF-1, through which these proteins may regulate Rb expression. Sp1 and Rb are especially inter-related at the transcrip- tional level and in their degradation fate. Sp1 is a ubiquitous transcription factor involved in the activation of a large number of genes. Its activity can be modulated during differentiation [56,57], cell growth [58,59], and development [60]. Sp1 and Rb interact physically, forming a complex that enhances the transcriptional activation of Sp1 [51]. Rb has also been described as a transcriptional activator of the p21 gene in epithelial cells through Sp1 and Sp3 transcription factors [61,62]. The transcriptional interaction of Rb, Sp1 and p21 implies an auto-loop of regulation between Rb and CDK activities. Sp1 can be phosphorylated by a cyclinA/CDK complex that probably includes CDK2, as phosphorylation is inhibited by olomoucine. Dephosphorylated Sp1 shows decreased DNA binding and transcriptional activity [63,64]. Moreover, Rb and Sp1 proteins are degraded by the same proteolytic enzyme, SPase, a nuclear and cytosolic protease with cathepsin B- and L-like proteolytic activity [65]. The levels of SPase vary along the cell cycle, correlating with Rb degradation, suggesting that SPase regulates Rb [66]. Taken together, these data prompted us to analyze the changes in the expression (transcription, mRNA and protein levels), of both Rb and Sp1 after CDK inhibition by the chemical inhibitors UCN-01, roscovitine and olomoucine, and by the intrinsic inhibitor p21. As DHFR activity is enhanced by the association between Rb and Sp1, we used DHFR as a target model to study the final effects of CDK inhibitors. We report that the expression of retinoblastoma and Sp1 is increased by low concentrations of CDK inhibitors in Chinese hamster ovary (CHO) K1 and human cells by a mechanism involving transcriptional activation and, also in the case of Rb, by an increase in its stability. Materials and methods Materials UCN-01 was kindly provided by H. Nakano (Kyowa Hakko Co., Tokyo, Japan). Roscovitine and olomoucine were purchased from Calbiochem. Bisindolylmaleimide I (BSM-I) was obtained from Sigma-Aldrich. Stock solutions were prepared in dimethylsulfoxide and maintained at )20 °C. Cell culture Conditions for the monolayer culture of CHO cells were as described previously [67]. CHO K1 and CHO DG44 cells [68] were grown in Ham’s F12 medium supplemented with 7% (w/v) fetal bovine serum (both from Gibco) and maintained at 37 °C in a humidified 5% (v/v) CO 2 - containing atmosphere. Human K562 and HeLa cells were grown under the same culture conditions. When determin- ing the activity of DHFR by the deoxyuridine method, cells were incubated in F12 selective DHFR medium (–GHT) lacking glycine, hypoxanthine and thymidine, the final products of DHFR activity. Flow cytometry analysis Cell cycle phase distribution upon incubation with CDK inhibitors was monitored by flow cytometry. Nuclei were stainedwith25lgÆmL )1 propidium iodide (Sigma-Aldrich) and analyzed on a Becton-Dickinson flow cytometer. mRNA levels mRNA levels were determined by quantitative RT-PCR using total cell lysates as the starting material for the RT reaction, as described in Noe ´ et al. [69]. K1 cells were plated in 35 mm-diameter dishes and, after several periods of incubation with 5 · 10 )8 M UCN-01, they were collected in 500 lL of F-12 medium. The cells were then centrifuged and washed with NaCl/P i , and the final pellet was resuspended in 11.25 lL of diethylpyrocarbonate-treated water. The cells were lysed at 80 °C for 5 min. cDNA was synthesized in 20 lL of reaction mixture containing 125 ng of random hexamers (Roche), 10 m M dithiothreitol, 20 U of RNasin (Promega), 0.5 m M dNTPs (AppliChem), 4 lLof5· RT buffer, 200 U of M-MLV reverse transcriptase (BRL- Gibco) and the cell lysate. The reaction mixture was incubated at 37 °C for 60 min. Five microlitres of the cDNA mixture was used for PCR amplification. PCR reactions were typically carried out as follows. A standard 50 lL mixture contained 5 lL of the cDNA mixture, 4 lLof10· PCR buffer (Mg 2+ -free), 1.5 m M MgCl 2 ,0.2m M dNTPs, 2.5 lCi of [ 32 P]dATP[aP] 4810 S. Pen ˜ uelas et al. (Eur. J. Biochem. 270) Ó FEBS 2003 (3000 CiÆmmol )1 ; Amersham Ibe ´ rica), 1.5 U of Taq poly- merase (Ecogen) and 500 ng of each of the four primers. For the determination of mRNA levels, the primers were: 5¢-CGCCAAACTTGGGGGAAGCA-3¢ and 5¢-GAACC AGGTTTTCCGGCCCA-3¢ for DHFR; 5¢-GTGCCAAT GGCTGGCAGATCA-3¢ and 5¢-ACCATCCTGCTGCA CTTGGGC-3¢ for Sp1; 5¢-CTCCACACACTCCAGTT AGGA-3¢ and 5¢-CTGATTTAAGCATGGATTCCA-3¢ for Rb; and 5¢-CGCAGTTTCCCCGACTTCCC-3¢ and 5¢-GGCAGCGCACATGGTTCCTC-3¢ for adenine phos- phoribosyltransferase (APRT), which was used as an internal control. The reaction mixture was separated into two phases by a solid paraffin wax layer (melting temperature 58–60 °C; Fluka), which prevents complete mixing of PCR reactants until the reaction has reached a temperature that minimizes nonspecific annealing of primers to nontarget DNA. The lower solution contained the cDNA, the MgCl 2 , the dNTPs, the [ 32 P]dATP[aP] and one half of the buffer, and the upper solution contained the four primers, the Taq enzyme and the remaining buffer. PCR was performed for 30 cycles, in the case of DHFR, and for 22 cycles, in the case of Sp1 and Rb, after a 1 min denaturation step at 94 °C. Each cycle consisted of dena- turation at 92 °C for 30 s; primer annealing at 59 °Cfor 1 min for DHFR and Sp1, and at 55 °C for 1 min for Rb; and primer extension at 72 °C for 1 min. Five microlitres of each PCR sample was electrophoresed in a 5% (w/v) polyacrylamide gel. The gels were dried and the radioactive bands visualized by autoradiography. Results were quanti- fied by image analysis using the BIO -1 D , version 99.03, software from Vilbert-Lourmat. The DHFR, Rb or Sp1 mRNA levels were expressed as the ratio of the intensities of the DHFR, Rb or Sp1 signals and APRT signals. Nuclear extracts Nuclear extracts from K1 cells were prepared as described in Ciudad et al. [70]. Protein concentrations were determined by the Bio-Rad protein assay based on the Bradford method [71], using bovine serum albumin as a standard (Sigma), and extracts were frozen in liquid N 2 andstoredat)80 °C. Total extracts Whole extracts were obtained from K1 or K562 cells according to the method of Kraus et al.[72].Cellswere collected in ice-cold F-12 medium and centrifuged at 1000 g for 5 min. The cell pellet was gently resuspended in 5 mL of hypotonic buffer (15 m M NaCl, 60 m M KCl, 0.5 m M EDTA, 1 m M phenylmethanesulfonyl fluoride, 1 m M 2-mercaptoethanol, 15 m M Tris/HCl, pH 8). After centri- fugation (1000 g, 5 min), the cell pellet was resuspended in 100 lL of a buffer containing deoxycholate (100 m M NaCl, 10 m M NaH 2 PO 4 , pH 7.4, 0.5% sodium deoxycholate, 0.1% SDS, 1% Triton X-100) and centrifuged at 13 000 g for 10 min. The resulting supernatant corresponded to the whole extract. The entire procedure was carried out at 4 °C. Five microlitres of the extract was used for determining the protein concentration by using the Bradford assay (Bio- Rad). The extracts were frozen in liquid N 2 andstoredat )80 °C. Western blot analysis Twenty micrograms of nuclear extract from CHO K1 or human K562 cells was resolved on SDS/5% polyacrylamide gels [71] and transferred to poly(vinylidene difluoride) membranes (Immobilon P; Millipore) using a semidry electroblotter. The membranes were probed with anti-Rb [C-15, against amino acids 914–928; IF-8, against amino acids 300–380 (both from Santa Cruz Biotechnology); and G3-425, against amino acids 332–344 (BD-Pharmingen)] or anti-(Sp1 PEP 2) (Santa Cruz Biotechnology). Detection of p21 was by using anti-SX118 (BD-Pharmingen). Signals were detected by secondary horseradish peroxidase-conju- gated antibody and enhanced chemiluminescence, as recommended by the manufacturer (Amersham). Blots were reprobed with C-21 antibody against Oct-1, or with antibody A2066 against actin (Sigma), in the case of experiments with p21, to normalize the results. Half-life of retinoblastoma and Sp1 The stability of Rb and Sp1 proteins was assessed by calculating their half-life from the concentration of protein remaining at different time-points after addition of cyclo- heximide to the cell culture. CHO K1 control cells, or those treatedwithUCN-01,5· 10 )8 M for 48 h, were incubated with 50 lgÆmL )1 cycloheximide for various periods of time. Total protein extracts were prepared and analyzed by Western blot using Rb and Sp1 specific antibodies. The results are expressed as the number of cells collected. Cloning of the retinoblastoma promoter Human genomic DNA from HeLa cells was used to isolate, by PCR, a clone containing a 630 bp fragment correspond- ing to the Rb promoter region. The PCR fragment was generated using two Rb-specific primers whose sequences were taken from GenBank accession number L11910. For the forward primer, the specific sequence (shown below in upper case text) was preceded by an arbitrary sequence (shown below in lower case text) that included an NheI restriction site (underlined). The reverse primer followed a similar structure but contained a XhoI restriction site (underlined) in the arbitrary sequence. The numbers indi- cated after the primer sequences correspond to the distance, in nucleotides (nt), from the translational start site: Rbprm- for, 5¢-tcaagtcag gctagcGTTCCGCACCTATCAGCGC TCC-3¢ (630 nt); Rbprm-rev, 5¢-cagtgctgc ctcgagGACGC CTTTCGCGGCGGGAGC-3¢ (1 nt). The PCR product was sequenced using Big Dye v2.0 (PE Biosystems). After digestion with NheIandXhoI, the PCR fragment from the Rb promoter was cloned into the NheI/XhoIsitesof the reporter luciferase vector pGL3-basic (Promega). The resulting construct was named prmRB-luc. Transfections and luciferase assay HeLa or K562 cells were seeded into six-well plates, the day before transfection, at a density of 7.5 · 10 4 cells/well in HAM F-12 containing 5% fetal bovine serum. Eighteen hours later, the CDK inhibitors were incubated with the cells and, 18 h later, transfections were performed using Ó FEBS 2003 Rb and Sp1 regulation by CDK inhibitors (Eur. J. Biochem. 270) 4811 FuGENE6 (Roche Molecular Biochemicals). For each well, 3 lL of FuGENE6 in 100 lL of serum free HAM F-12 medium was incubated at room temperature for 5 min. The mixture was added to 125 ng of the Rb promoter (prmRB- luc) or 250 ng of Sp1 promoter (prmSp1-luc) constructs. The DNA–lipid mixture was incubated at room tempera- ture for 15 min. The mixture was added to the cells for 30 h. Luciferase activity was assayed 48 h after treatment with the CDK inhibitors and 30 h after transfection of the constructs. Cells were lysed in 500 lL of freshly diluted 1· reporter lysis buffer (Promega). The lysate was centrifuged at 10 000 g for 2 min to pellet the cell debris and the supernatants were transferred to a fresh tube. A 10 lL aliquot of the cell extract was added to 25 lLofthe luciferase assay substrate (Promega) and the luminescence of the samples was read immediately on a TD-20/20 Luminometer, in which light production (relative light units) was measured for 10 s. Luciferase activity was normalized to cellular protein concentration, determined using the Bio-Rad protein assay reagent in accordance with the manufacturer’s protocol. Transfections, cotransfections and the DHFR transient activity assay CHO K1 cells were cotransfected with increasing amounts (1, 3 and 5 lg) of a eukaryotic expression vector for p21 (pCMV-Cip1), together with 0.4 lgofBPV-Neo,byusing the calcium phosphate method [73]. After 24 h of expres- sion, selection with GeneticinÒ (800 lgÆmL )1 ) was applied. Three weeks later, the surviving colonies were pooled. Transient expression experiments were carried out in dhfr-deficient cells (CHO-DG44) by transfecting a dhfr minigene in the presence and absence of CDK inhibitors. When using p21, the expression vector corresponding to this protein was co-transfected together with the dhfr minigene. All transient transfections were also performed by the calcium phosphate method. The plasmid providing basal DHFR activity was p410, corresponding to a dhfr minigene driven by its minimal promoter [74]. After 24 h of expres- sion, the medium was replaced with –GHT medium (DHFR selective medium) and the resulting DHFR activity was determined by the incorporation of radioactive deoxy- uridine to cellular DNA, as described by Noe ´ et al.[51]. The chemical CDK inhibitors UCN-01, roscovitine and olomoucine were added to the medium immediately after transfection and maintained during the expression and labeling times of the assay. Results Effects of the CDK inhibitor UCN-01 on retinoblastoma and Sp1 proteins CHO K1 cells were incubated with increasing concentra- tions (10 to 100 n M ) of UCN-01, for 48 h. Nuclear extracts were then prepared and analyzed by Western blot. UCN-01 resulted in an increase of the total amount of Rb protein in a dose-dependent manner, with a maximum ( 10-fold) at 5 · 10 )8 M (Fig. 1A). The results were normalized using the signal obtained upon reprobing the same blots with an antibody against Oct-1. To help define the phosphorylation states of Rb in these cells, we performed Western blot analysis with nuclear extracts from CHO cells in the different phases of the cell cycle. Three bands were observed: a nonphos- phorylated form of Rb, which is the only band present in starved cells; a hypophosphorylated form, with inter- mediate mobility, which appears when cells are in G1; and a low-mobility hyperphosphorylated form, which is pre- sent mainly in S and G2/M phases (Fig. 1C). Only the nonphosphorylated form of Rb was detected upon incubation with high concentrations of UCN-01 (10 )6 M ) (Fig. 1C), revealing that this compound inhibits the CDK–cyclin complexes that phosphorylate Rb in these cells. At this concentration, K1 cells were arrested. However, 50 n M UCN-01, at which the maximum increase of Rb was observed, did not affect cell prolifer- ation (data not shown). As UCN-01 is also able to inhibit protein kinase C (PKC) activity (IC 50 ¼ 7 · 10 )9 M )we investigated the different effects of UCN by using the PKC inhibitor, BSM-I (IC 50  10 )8 M ). In CHO cells, 10 )6 M BSM-I produced an increase in Rb expression that represented 40% of the increase caused by 50 n M UCN-01 (Fig. 1D). We also aimed to determine the levels of Sp1 protein in UCN-01-treated cells. To achieve this, the blots used for determining Rb protein levels were reprobed with a specific antibody (PEP 2) against Sp1. Three bands of Sp1 were detected in nuclear extracts from CHO K1 cells, as reported previously [75]. The levels of this transcription factor increased, in a dose-dependent manner, when K1 cells were incubated for 48 h with UCN-01 (Fig. 1B), peaking at 5 · 10 )8 M UCN-01. Mechanisms explaining the increase in Rb and Sp1 protein caused by CDK inhibitors in CHO cells The increase in total levels of Rb and Sp1 proteins caused by UCN-01 may be a result of the enhanced transcription and stability of Rb and Sp1. To assess transcription, we determined Rb and Sp1 mRNA levels after incubation of CHO cells with 50 n M UCN-01 for different periods of time. The CDK inhibitor increased Rb mRNA levels by 3.5-fold at 20 h of incubation (Fig. 2A) and Sp1 mRNA levels by fivefold at 6 h (Fig. 2B). In both cases, the signal obtained for APRT mRNA was used to normalize the results. To test whether the effect of UCN-01 was also caused by an increase in the stability of Rb and Sp1 proteins, we determined the decay of these proteins after inhibiting protein synthesis by cycloheximide. CHO cells were incubated with 50 n M UCN-01 for 48 h, which yields maximal expression of Rb and Sp1, and control and UCN- 01-treated cells were then incubated with 50 lgÆmL )1 cycloheximide for different periods of time. Whole protein extracts were prepared and used to determine the levels of Rb and Sp1 by Western blot, as described in the Materials and methods. The half-life of the Rb protein increased from 8.7 h to 14.2 h upon treatment with UCN-01 (Fig. 3A,B), which corresponds to a 62% increase in the stability of the protein, whereas the difference in Sp1 stability between control and UCN-01 treated cells was not significant (data not shown). 4812 S. Pen ˜ uelas et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Therefore, the effect of UCN-01 on the levels of Rb protein may be caused by an increase in the synthesis of Rb and by a decrease in its degradation. However, the effect on Sp1 could be accounted for by the increase observed in Sp1 mRNA. Effects of roscovitine, olomoucine and p21 in CHO cells To examine whether the effects caused by UCN-01 were shared by other CDK inhibitors, we extended the analysis to roscovitine and olomoucine. These inhibitors belong to the C2,N6,N9-substituted adenine family and mainly inhibit CDK2 activity, as their IC 50 values for CDK4/6 are  100 and 1000 times higher than for CDK2, respectively. Cells were incubated with increasing concen- trations of these two chemical inhibitors, and the levels of Rb and Sp1 protein were determined. The range of concentrations of olomoucine used was higher than for roscovitine, given its higher IC 50 for CDK2. Both inhibitors resulted in an increased total amount of the proteins Rb and Sp1 (Fig. 4A,B). We also tested the effect of overexpression of the intrinsic CDK inhibitor p21 in pooled permanent trans- fectants. CHO K1 cells were co-transfected with increas- ing amounts of an expression vector for p21 together with BPV-Neo and, upon selection with Geneticin, the pools were used to determine the protein levels of Rb and Sp1. The levels of both proteins increased, even in the transfectants obtained with 1 lgofp21(Fig.4C).The overexpression of p21 was confirmed in these transfect- ants. (Fig. 4D). These results extended the original observations for UCN-01 and confirmed that the inhibition of CDK activity increases the expression of Rb and Sp1. Fig. 1. Effects of UCN-01 on the levels of Rb and Sp1 protein. (A), (B) Dose–response of the effect of UCN-01 on the levels of retinoblastoma gene product (Rb) and Sp1 proteins, respsectively. K1 cells (10 6 cells per 100 mm diameter dish) were incubated with the indicated concentrations of UCN-01 for 48 h. Nuclear extracts were prepared and resolved by SDS/PAGE. Rb protein was detected by Western blotting using a 1 : 100 dilution of the C-15 antibody against Rb and enhanced chemiluminescence. Sp1 protein was detected using PEP 2 antibody. The same blots were reprobed with a 1 : 100 dilution of the C-21 antibody against Oct-1 to normalize the results. Quantification of the signal is shown in the bottom panel. Rb and Sp1 protein levels were determined from the ratio of the intensities between the Rb or the Sp1 and Oct-1 signals, respectively. Results represent the mean ± SEM of three experiments. (C) Phosphorylated forms of Rb in K1 cells. Nuclear extracts (20 lg) from cells in each phase of the cell cycle (G0, G1, S and G2/M), exponentially growing cells (Exp) or cells treated with 10 )6 M UCN-01 for 24 h, were resolved by SDS/PAGE. A Western blot was performed, as described above for (A). The phosphorylated forms of Rb are indicated by arrows. Rb (nonphosphorylated form), Rb-P (hypophosphorylated form), Rb-PP (hyperphosphorylated form). (D). Determination of Rb protein levels with 50 n M UCN-01 or 1000 n M bisindolylmaleimide I by Western blot analysis using C-15 antibody in CHO K1 cells. Ó FEBS 2003 Rb and Sp1 regulation by CDK inhibitors (Eur. J. Biochem. 270) 4813 Effects of CDK and PKC inhibition on Rb and Sp1 expression in human cells Next, we determined whether the increased expression of Rb and Sp1 protein upon incubation with UCN-01, roscovitine and olomoucine were also produced in human cells. It was observed that low concentrations of UCN-01 (50 n M ), roscovitine (100 n M ) or olomoucine (500 n M ) caused an increase in the expression of Rb protein, both in nuclear and total extracts from K562 cells (Fig. 5A,C,E). Sp1 expression was also increased by low concentrations of these three inhibitors in K562 cells (Fig. 5B,D). The changes in transcriptional activity caused by UCN-01, roscovitine and olomoucine in K562 cells were also determined by using luciferase assays. The three CDK inhibitors caused an increase in transcription Fig. 2. Rb (A) and Sp1 (B) mRNA levels upon incubation with UCN-01. CHO K1 cells (5000 cells) were incubated with 5 · 10 )8 M UCN-01 for the indicated periods of time. The mRNA levels for Rb and Sp1 were then determined from cell lysates by RT-PCR in quantitative conditions using specific primers. The top panels correspond to rep- resentative autoradiographs of the amplified products and the quan- tification of the bands is shown in the bottom panels. Rb and Sp1 mRNA levels were determined from the ratio of the intensities between the Rb or Sp1 and the APRT signals, which was used as an internal control of the reaction. Results are the mean ± SEM of three experiments. Fig.3. EffectofUCN-01onRbhalf-life.CHO K1 cells (1000 cells per 35 mm diameter dish) were incubated with 5 · 10 )8 M UCN-01 for 48 h, followed by the addition of 50 lgÆmL )1 cycloheximide to the culture medium. At different time-points, cells were collected and used to prepare total protein extracts. The total amount of the extract was resolved by SDS/PAGE and the Rb protein levels were determined by Western blot, as described in the legend to Fig. 1A. (A) Semi-log plot of the levels of Rb protein as a function of chase time. A representative experiment of the five performed is shown. The remaining Rb protein levels are expressed as the percentage of the Rb protein present at 0 h of CHX, for either the control or the UCN-01-treated cells. (B) Effect of UCN-01 on Rb stability. The half-life was calculated using the exponential curve fit method depicted in the legend to Fig. 3A. Results represent the mean ± SEM from five experiments. 4814 S. Pen ˜ uelas et al. (Eur. J. Biochem. 270) Ó FEBS 2003 upon transfection of reporter constructs Rb-luc and Sp1- luc in K562 (Fig. 6A,B,C,D). Transient transfection with a p21 expression vector also caused an activation of Rb- and Sp1 promoters (data not shown). In addition, UCN01 also increased Rb and Sp1 transcription in HeLa cells (Fig. 6E,F). Cell cycle distribution was determined in K562 to study whether the effect of the three compounds on Rb expression was related to CDK inhibition. UCN-01, at 50 n M , caused a change in the distribution of the cell cycle towards the G1 phase, at the 24 h time-point, without affecting cell proliferation. At higher concentrations, this effect persisted, whereby an increase of cells in G2/M was observed (Fig. 5F). Roscovitine produced a displacement of the distribution of the cell cycle towards the G2/M phase that started at the low concentration of 100 n M . Roscovitine and olomoucine caused similar effects, and the results cor- responding to roscovitine are shown in Fig. 5F. The CDK-independent effect of UCN01 was studied upon incubation with BSM-I, a PKC inhibitor. In K562 cells, concentrations of BSM-I that have a similar PKC inhibitory ability as UCN-01, based on their IC 50 values, increased Rb and Sp1 expression to a lower level than UCN-01 (Fig. 5A,B). These concentrations of BSM-I increased Rb and Sp1 promoter activity in luciferase assays (Fig. 6A,B) to  30% of the increase observed with UCN-01. Effects of UCN-01 on DHFR activity Taking into account that Sp1 is a powerful activator of DHFR transcription and that Rb stimulates Sp1 transcrip- tional activity, we aimed to assess whether the increases in Rb and Sp1 proteins caused by CDK inhibitors affected DHFR expression. To achieve this, we determined the levels of DHFR mRNA upon incubation of CHO K1 cells with UCN-01, and DHFR activity in transient transfections after incubation with UCN-01, roscovitine and olomoucine and in co-transfections with p21. After incubation with 50 n M UCN-01 for different periods of time, the levels of DHFR mRNA transiently increased to a maximum at 36 h (Fig. 7A). The effect of UCN-01 on DHFR activity was analyzed in transiently transfected dhfr-deficient cells (CHO DG44) using a dhfr minigene (p410). DHFR activity increased to a maximum at 50 n M UCN-01; at higher concentrations of the inhibitor, the activity decreased (Fig. 7B). Roscovitine and olomoucine also enhanced DHFR activity, to a maximum at 100 n M roscovitine and Fig. 4. Effect of roscovitine, olomoucine and p21 on the levels of Rb and Sp1 protein. (A), (B) Effect of roscovitine or olomoucine on retinoblastoma gene product (Rb) and Sp1 protein levels. CHO K1 cells (500 000 per 100 mm diameter dish) were treated with increasing concentrations of the CDK2 inhibitors, roscovitine (A) or olomoucine (B) for 48 h. Twenty micrograms of nuclear extracts from either control or treated cells was resolved by SDS/PAGE. Rb and Sp1 proteins were detected by Western blot using C-15 antibody against Rb, PEP 2 antibody against Sp1, or C-21 antibody against Oct-1, which was used as a control to normalize the results. (C) Effect of the overexpression of p21 on Rb and Sp1 protein levels. Nuclear extracts (20 lg) were obtained from CHO K1 cells stably transfected with different amounts (1, 3 and 5 lg) of a eukaryotic expression vector for p21. Western blot analysis was performed with C-15 antibody against Rb, PEP 2 antibody against Sp1, or C-21 antibody against Oct-1, which was used as a control to normalize the results. (D) Determination of the overexpression of p21 in permanent transfectant cells using SX118 antibody against p21 and A2066 antibody against actin to normalize the result. Ó FEBS 2003 Rb and Sp1 regulation by CDK inhibitors (Eur. J. Biochem. 270) 4815 500 n M olomoucine, but the activity decreased thereafter (Fig. 7C,D). In addition, the effect of overexpression of p21 on DHFR activity was also analyzed in transient co-transfection experiments. DHFR activity increased, depending on the amount of co-transfected p21 (Fig. 7E). Discussion We aimed to explore the effects of the inhibition of CDKs on the expression of their natural substrate, retinoblastoma. This has special interest given that some CDK inhibitors, Fig. 5. Effect of UCN-01, roscovitine, olomoucine or bisindolylmaleimide I on Rb and Sp1 expression in human cells. (A), (B), (C), (D) K562 cells (10 6 cells per 100-mm diameter dish) were incubated with the indicated concentrations of bisindolylmaleimide I, UCN-01, roscovitine (RSC) and olomoucine (OLM) for 48 h. Nuclear extracts were prepared and resolved by electrophoresis on a 5% polyacrylamide/SDS gel. Western blots were performed to detect Rb protein using G3-245 antibody against Rb (A,C), or Sp1 protein using PEP 2 antibody (B,D). The same blots were reprobed with C-21 antibody against Oct-1 to normalize the results. Quantification of the signal is shown on the bottom panel. Results represent the mean ± SEM of three experiments. (E) Cells were treated with the indicated concentrations of the CDK inhibitors and total extracts were prepared as described in the Materials and methods. The levels of Rb protein were determined by Western blot using G3-245 antibody. (F) Cell cycle distribution of K562 cells upon incubation with CDK inhibitors. Cells were treated with either UCN01 for 24 h or roscovitine for 48 h, at the indicated concentrations, and the percentage of cells in each phase of the cell cycle was analyzed by flow cytometry. 4816 S. Pen ˜ uelas et al. (Eur. J. Biochem. 270) Ó FEBS 2003 like UCN-01 and roscovitine, are undergoing clinical trials for use in anticancer treatment, as high concentrations of CDK inhibitors have antiproliferative effects. We used the chemical CDK inhibitors UCN-01, roscovi- tine and olomoucine, as each shows a distinct specificity towardCDKs.p21wasalsousedasanintrinsicCDK inhibitor. UCN-01 inhibits CDK4/6 and CDK2 with IC 50 values of 0.032 l M and 0.030 l M , respectively. However, roscovitine and olomoucine have an IC 50 for CDK4/6 that is  100-fold higher than for CDK2 (0.7 l M for CDK2 and > 100 l M for CDK4/6, for roscovitine, and 7 l M for CDK2 and > 1000 l M forCDK4/6,forolomoucine, respectively) [76,77]. Thus, CDK2 can be selectively inhi- bited by using low concentrations of these inhibitors, according to their IC 50 values. A first conclusion of this work is that upon incubation with submaximal concentrations of CDK inhibitors, the total amount of Rb protein increases in a dose-dependent manner. In the case of UCN-01, the maximal effect was observed at 50 n M , a concentration that did not interfere with cell proliferation. However, this inhibitor was able to arrest the cells and to prevent Rb phosphorylation (Fig. 1C) when used at high concentrations (10 )6 M ), in agreement with previous observations [78]. The amount of Rb protein was also increased by the CDK intrinsic inhibitor, p21, with the same broad spectrum of action as UCN-01, and by roscovitine and olomoucine, which inhibit CDK2 activity. UCN-01 was originally described as a PKC inhibitor and shows a low IC 50 for this kinase (0.007 l M ). Therefore, we explored the possible contribution of the PKC inhibitiory activity of UCN-01 to the increase in Rb expression, to characterize the extent of the CDK-independent effect. Indeed, inhibition of PKC by using BSM-I, a more selective inhibitor of PKC, also triggered the expression of Rb protein and Rb transcription. Therefore, the action of Fig. 6. Transcriptional activity of the retinob- lastoma gene product (Rb) and Sp1 upon treatment with UCN-01 and bisindolylmalei- mide I in human cells. prmRb-LUC (125 ng) or prmSp1-LUC (250 ng) were transiently transfectedintoK562cellsandassayedfor luciferase activity after 48 h of treatment with the indicated concentrations of bisindolylma- leimide I or UCN-01 (A), (B) and roscovitine or olomoucine (C), (D). Transfections were performedinduplicate,andtheresultsrepre- sent the mean ± SEM of two experiments. Luciferase activity was normalized to micro- grams of protein for each sample. (E), (F) As for (A), (B), but incubation was with increas- ing concentrations of UCN-01 in HeLa cells. Ó FEBS 2003 Rb and Sp1 regulation by CDK inhibitors (Eur. J. Biochem. 270) 4817 UCN01 is caused by overlapping effects. However, this increase in Rb expression is lower than that caused by UCN-01, when similar inhibition of PKC was achieved. In addition, 100 n M roscovitine and 500 n M olomoucine, which increase Rb expression, do not inhibit PKC (IC 50 values of roscovitine and olomoucine for PKC are > 100 l M and > 1000 l M , respectively [76,77]). Regarding roscovitine and olomoucine, the levels of Rb protein increase at concentrations that inhibit CDK2 but not CDK4/6. Thus, inhibition of CDK2 alone, which prevents hyperphosphorylation of Rb, is sufficient to trigger the increase in expression of Rb. The effects on Rb and Sp1 have also been demonstrated in human cells both at the level of protein expression and transcriptional activity. The changes in the cell cycle distribution in human K562, produced by the low concen- trations of these compounds, show an effect on CDK activity, and it is precisely at these submaximal concentra- tions where the clearest effects on RB and Sp1 expression are seen. Given that Rb has various phosphorylation states depending on the phase of the cell cycle, two mechanisms may explain the increase in the total amount of Rb upon CDK inhibition. Rb expression may be enhanced by a negative effect caused by the hyperphosphorylated form, by a positive effect caused by accumulation of the hypophosphorylated form or by a combination of the two mechanisms, in keeping with the self-regulation of the Rb gene by its own gene product, as proposed elsewhere. On the one hand, Hamel et al. [79], using transfections of Rb in differentiated P19 cells, thus overexpressing nonphosphorylated Rb, and Gill et al. Fig. 7. Effect of CDK inhibitors on mRNA levels and dhfr activity. (A) Dihydrofolate reductase (DHFR) mRNA levels after incubation of K1 cells with UCN-01. One thousand K1 cells were plated in 35-mm diameter dishes and treated with 5 · 10 )8 M UCN-01, for the indicated periods of time, on the following day. Cells were then harvested and lysed, and DHFR mRNA levels were determined by quantitative RT-PCR using specific primers in exonic sequences of the dhfr and aprt genes. The top panel of the figure corresponds to a representative autoradiograph of the amplified products, and quantification of the bands is shown in the bottom panel. DHFR mRNA levels were determined from the ratio of the intensities between the DHFR and APRT signals. Results represent the mean ± SEM of three experiments. (B), (C), (D) Effect of UCN-01, roscovitine and olomoucine on DHFR activity upon transient transfection with a DHFR minigene. DG44 cells (225 000 cells per 35 mm diameter dish) were plated and, 18 h later, transfected with 1 lg of plasmid p410 (dhfr minigene) using the calcium phosphate method. Immediately after the transfection, UCN-01 (B), roscovitine (C) or olomoucine (D) were added at the indicated concentrations, and maintained in the culture medium throughout the expression experiments. After 24 h of expression, the medium was replaced with 1 mL of F12 selective DHFR medium (to renew the cyclin- dependent kinase inhibitors), and 2 lCi of 6[ 3 H]deoxyuridine was then added. Cells were collected 24 h later, and the incorporated radioactivity was measured in a scintillation counter. DHFR activity is expressed as c.p.m. incorporated to DNA. Results represent the mean ± SE of three experiments. (E) Effect of the overexpression of p21 on DHFR activity upon transient co-transfection with a dhfr minigene. CHO-DG44 cells (225 000 cells per 35 mm diameter dish) were co-transfected with 1 lgofthedhfr minigene p410 plus 1 or 2 lg of an eukaryotic expression vector for p21, using the calcium phosphate method. DHFR activity was determined as described in the legend to Fig. 6B. Results represent the mean ± SEM of three experiments. 4818 S. Pen ˜ uelas et al. (Eur. J. Biochem. 270) Ó FEBS 2003 [...]... that the increased level of total Rb protein is caused by two mechanisms On the one hand, there is an increase in transcription, as shown by the luciferase experiments and in the levels of Rb mRNA; and, on the other hand, the stability of Rb protein is increased As stated in the Introduction, Rb interacts with a variety of proteins We selected Sp1 to study the possible effect of CDK inhibitors for the. .. similarities in the protein expression of Sp1 and Rb in response to low concentrations of CDK inhibitors First, the expression of Sp1 strongly increases with the chemical inhibitors UCN-01, roscovitine and olomoucine, and with the overexpression of the intrinsic inhibitor p21 Second, the levels of Sp1 mRNA increase upon incubation with UCN-01, reflecting an elevated transcription, as demonstrated in the transient... with Sp1- luc in human cells However, there are also some differences, namely (a) in the case of Rb, there is an increase in both the levels of mRNA and the stability of the protein, while the degradation rate of Sp1 does not vary and thus the increase in the protein levels seems to depend only on the increase in transcription, and (b) the time dependency of the increase in mRNAs is shorter for Sp1 than... phosphorylates Sp1 and enhances Sp1mediated transcription [63], and high concentrations of olomoucine reduce Sp1 transcriptional activity [64], the phosphorylated form of Sp1 may be needed for the transcriptional activation of a variety of Sp1- controlled genes A slight decrease in the phosphorylated form of Sp1, caused by CDK inhibitors, may trigger Sp1 synthesis to re-establish its normal levels Given that Sp1. .. was increased by UCN-01, roscovitine, olomoucine and p21 in assays of transient transfections This increase takes place at concentrations of the inhibitors where the expression of Rb and Sp1 starts to increase At higher concentrations of these inhibitors, the effect on DHFR is no longer noticeable In summary, we describe that CDK inhibitors, when used at submaximal concentrations, enhance expression of. .. the following reasons, namely that (a) the presence of GC-boxes in the Rb promoter may allow regulation through Sp1, (b) Sp1- mediated transcription is stimulated by Rb through a physical complex between the two proteins, (c) Sp1 can be phosphorylated by a cyclinA/ CDK complex and (d) Sp1 is degraded by SPase, a protease that is also active on Rb and regulates this protein in the cell cycle There are... combination with other chemotherapeutic agents Lower than optimal concentrations of CDK inhibitors can increase the activity of potential target proteins, such as DHFR, which is inhibited by treatment with methotrexate Moreover, overexpression of Rb decreases the susceptibility of cells to therapy with agents that increase apoptosis [83] Thus, the dose adjustment of potential chemotherapy combinations is required... hypophosphorylated form is the major species Sandig et al [82] reported that overexpression of the CDK inhibitor, p16, which prevents the phosphorylation of Rb by CDK4/6, down-regulates Rb The hypophosphorylated form of Rb may thus activate Rb transcription Our results are in keeping with those observations, as low concentrations of CDK inhibitors trigger a response of increased Rb transcription Of particular... than Rb mRNA, and that the Rb promoter contains GC boxes, Sp1 may contribute to the rise in Rb expression Finally, we used the dhfr gene as a target model to study the effect of CDK inhibitors, on a late response gene, upon stimulation to proliferate The dhfr gene was selected because it is activated mainly by Sp1, especially when the latter is associated with Rb [51] DHFR mRNA was increased by UCN-01... J.P., Blow, J.J., Inagaki, N., Iganaki, M., Delcros, J.G & Moulinoux, J.P (1997) Biochemical and cellular effects of roscovitine, a potent and selective inhibitor of the cyclin-dependent kinases cdc2, cdk2 and cdk5 Eur J Biochem 243, 527–536 40 Park, D.S., Farinelli, S.E & Greene, L.A (1996) Inhibitors of cyclin-dependent kinases promote survival of post-mitotic neuronally differentiated PC12 cells and . The expression of retinoblastoma and Sp1 is increased by low concentrations of cyclin-dependent kinase inhibitors Silvia Pen ˜ uelas*, Cristina. UCN-01-mediated retinoblastoma expression was caused by an increase in both the levels of retino- blastoma mRNA and the stability of the protein. The expression of the