Inactivation of p57KIP2 by regional promoter ... - Nature

Oncogene (2002) 21, 2741 ± 2749 ã 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Inactivation of p57KIP2 by regional promoter hypermethylation and histone deacetylation in human tumors Takefumi Kikuchi1, Minoru Toyota*,1,2, Fumio Itoh1, Hiromu Suzuki1, Toshiro Obata1, Hiroyuki Yamamoto1, Hideki Kakiuchi1, Masanobu Kusano1, Jean-Pierre J Issa3, Takashi Tokino2 and Kohzoh Imai1 1

First Department of Internal Medicine, Sapporo Medical University, Sapporo 060-8543, Japan; 2Department of Molecular Biology, Cancer Research Institute, Sapporo Medical University, Sapporo 060-8556, Japan; 3Department of Leukemia, MD Anderson Cancer Center, Houston, Texas, TX 77030, USA

To clarify the role of DNA methylation in the silencing of the expression of cyclin-dependent kinase inhibitor p57KIP2 seen in certain tumors, we investigated the methylation status of its 5' CpG island in various tumor cell lines and primary cancers. Dense methylation of the region around the transcription start site was detected in 1 out of 10 colorectal, 2 out of 8 gastric, and 6 out of 14 hematopoietic tumor cell lines and in 5 out of 35 (14%) gastric, 6 out of 20 (30%) hepatocellular, and 2 out of 18 (11%) pancreatic cancers; 7 out of 25 (28%) acute myeloid leukemia cases also showed methylation of the p57KIP2 gene, which strongly correlated with the CpG island methylator phenotype (P50.001). Detailed mapping revealed that dense methylation of the region around the transcription start site (7300 to +400), but not of the edges of the CpG island, was closely associated with gene silencing. 5-aza-2'-deoxycytidine, a methyltransferase inhibitor, restored expression of p57KIP2, and chromatin immunoprecipitation using anti-histone H3 and H4 antibodies showed histone to be deacetylated in cell lines where p57KIP2 was methylated at the transcription start site. Regional methylation and histone deacetylation thus appear to be crucially involved in the silencing of p57KIP2 expression in human tumors. Oncogene (2002) 21, 2741 ± 2749. DOI: 10.1038/sj/ onc/1205376 Keywords: cell cycle; gene expression; CIMP; chromatin; acetylation Introduction Cell cycle abnormalities are a hallmark of malignant tumor cells (Elledge, 1996; Sherr, 2000). The aected transitions within the cell cycle are regulated by the balanced activities of cyclin-dependent kinases (CDKs) and CDK inhibitors. Accumulation of CDKs during

*Correspondence: M Toyota; E-mail: [email protected] Received 15 October 2001; revised 29 January 2002; accepted 31 January 2002

G1 phase results in sequestration of Cip/Kip family CDK inhibitors such as p21WAF1/CIP1, p27KIP1 and p57KIP2, which complement the eects of the E2F transcriptional program by facilitating cyclin E-cdk2 activation at the G1-S transition. Of particular interest to us is p57KIP2, which is located on chromosome 11p15, a region often deleted in human tumors, and is the gene thought to be responsible for BeckwithWiedemann syndrome (BWS) Hatada et al., 1996; Matsuoka et al., 1995; Zhang et al., 1997). When introduced into cells, p57KIP2 suppresses cell transformation by inhibiting CDKs and interacting with proliferating nuclear antigen (PCNA) (Watanabe et al., 1998). Conversely, cells lacking p57KIP2 show increased cell proliferation and decreased dierentiation (Yan et al., 1997; Zhang et al., 1997). Taken together, these ®ndings suggest that p57KIP2 may serve as a tumor suppressor gene, though mutated forms of p57KIP2 have rarely been detected in human tumors (Tokino et al., 1996). Diminished expression of p57KIP2 has recently been associated with a variety of tumors, including gastric (Shin et al., 2000a), hepatocellular (Schwienbacher et al., 2000), esophageal (Matsumoto et al., 2000) and bladder cancers (Oya and Schulz, 2000). This suggests that, in addition to genetic changes, epigenetic mechanisms might also be involved in altering expression of the gene. In that regard, a considerable amount of evidence implicates DNA methylation in the silencing of genes in tumors (Baylin et al., 2001; Jones and Laird, 1999; Tycko, 2000) via a pathway involving a number of transcription repressors, including methyl binding domain proteins (MBDs) and histone deacetylases (HDACs). Indeed, Shin et al. (2000b) recently reported p57KIP2 to be hypermethylated in a subset of gastric cancer cell lines. Still, the precise mechanism and functional consequences of p57KIP2 gene silencing in human tumors remains unclear. To better understand the role of aberrant methylation of p57KIP2 in human tumors, we examined the methylation status of its 5' CpG island in a large panel of primary cancers and cell lines. Detailed analysis of the entire CpG island revealed that aberrant methylation in the area of the transcription start site is associated with hypoacetylation of histones

p57KIP2 methylation in human tumors T Kikuchi et al

2742

H3 and H4 in a variety of human tumors, suggesting that DNA methylation and histone deacetylation play key roles in the silencing of p57KIP2 expression.

Results Aberrant methylation of the 5' CpG island of p57KIP2 Sequencing of the genome in the region of p57KIP2 has shown the gene to contain a CpG island that spans all four of its exons (Tokino et al., 1996). In addition, Blast analysis of exon 1 matched the BAC sequences (GenBank: AC013791) and revealed the island to extend about 1.5 kb upstream from the transcription start site (Figure 1a). To explore the role of p57KIP2 methylation in human tumors, we ®rst used bisul®tePCR to examine the methylation status of the region around the transcription start site in a series of human tumor cell lines (Figure 1b). Aberrant methylation was detected in 1 out of 10 colorectal cancer cell lines and 2 out of 8 gastric cancer cell lines; out of 14

Figure 1 Analysis of p57KIP2 methylation in various cancer cell lines. (a) Schematic representation of a p57KIP2 CpG island: exons are indicated by solid bars; CpG sites are indicated by vertical bars. Regions examined by bisul®te-PCR and bisul®te sequencing are indicated above. The arrow indicates the transcription start site. (b) Representative bisul®te-PCR analysis; M, methylated alleles. The cell lines analysed are shown on the top. The methylated alleles were examined using a densitometer, and the percentage methylation is shown below the gel. (c) Methylation analysis of p21WAF1 and p27KIP1; the genes analysed are shown on the left Oncogene

hematopoietic tumor cell lines, six showed dense methylation (50 ± 90%) and ®ve showed sparse methylation (2 ± 14%); ®ve hepatocellular and two pancreatic cancer cell lines showed no methylation (Figure 1b). Aberrant methylation of two other Cip/Kip family CDK inhibitors, p21WAF1 and p27KIP1, was not detected in any of the cell lines analysed (Figure 1c). The methylation status of each CpG dinucleotide was assessed in more detail by bisul®te sequencing of the DNA from one methylated (Raji) and two unmethylated (DLD-1 and Colo205) cell lines (Figure 2). Two Sp1, one AP1 and one TFIID binding site (TATA box) were identi®ed in the region analysed. In the Raji cells, methylation was particularly dense at the Sp1 and TFIID binding sites: of 36 CpG dinucleotides from 12 alleles corresponding to the SP1 sites, 34 (94%) were methylated; and out of 12 CpG dinucleotides corresponding to the TFIID site, 11 (92%) were methylated. In addition, of the 12 CpG dinucleotides corresponding to the AP1 binding site, seven (58%) were methylated. Analysis of the methylation pro®les of 12 Raji clones exhibiting a G/A polymorphism at position 7170 of p57KIP2 enabled us to con®rm that p57KIP2 methylation is biallelic. The 5' CpG island of p57KIP2 is relatively large, and for cancer cells in which the gene has been silenced, it is not clear whether the island is uniformly methylated throughout or whether regional hypermethylation suppresses gene expression in a subset of cases. To answer this question, we examined the methylation status of p57KIP2 using six primer sets that together spanned the entire CpG island (Figure 3a,b). No repetitive sequences were observed in the six regions analysed. Surprisingly, even cell lines in which p57KIP2 was not methylated in regions C, D and E did show methylation in other regions (MKN7, MKN74, AZ521, JRST, NUGC3, SW480, Colo201, Colo205, Colo320, BM314, DLD-1, Panc-1, MIAPaCa-2, KMS12PE, RPMI8225, K562). All of the cell

Figure 2 Bisul®te-sequencing of the 5' region of p57KIP2. In total, 60 CpG sites between locations 7300 and +80, relative to the transcription start site, were examined. Methylated and unmethylated alleles are shown as solid and open circles, respectively. An arrow indicates the transcription start site. Putative transcription factor binding sites predicted by TESS, a promoter prediction program (http://www.hgsc.bcm.tmc.edu), are indicated above


2743

Figure 3 Methylation mapping of the p57KIP2 CpG island. (a) The locations of the six primer sets (A-F) used to examine methylation of entire CpG island. (b) Representative results for bisul®te-PCR analysis. The cell lines examined are indicated above the gels. The regions and restriction enzymes used are shown on the left. The level of methylation of each CpG site was evaluated densitometrically and shown below the gels. Arrowheads indicate the methylated alleles (M)

lines showing aberrant methylation as in Figure 1b (region C) were densely methylated throughout the entire CpG island (MKN28, KatoIII, HT-29, HO3238, HS-Sultan and Raji). Consequently, the cell lines examined could be divided into two distinct groups: those methylated in regions A, B and F but not in regions C, D or E (group 1) and those densely methylated over the entire CpG island (group 2). Aberrant methylation of regions C and D of the p57KIP2 CpG island diminishes or abolishes the gene's expression We next carried out a series of RT ± PCR analyses using cDNAs from 22 tumor cell lines to verify whether aberrant methylation could account for the silencing of p57KIP2 (Figure 4a). Expression of p57KIP2 was detected in normal tissues and a majority of the cell lines examined. However, three cell lines (HT-29, MKN28 and KatoIII) expressed only low levels of p57KIP2, and three (HO3238, HS-Sultan, and Raji) did not express the gene at all. Methylation mapping showed that all six cell lines in which p57KIP2 expression was diminished or absent belonged

to group 2, whereas the lines that showed strong expression belonged to group 1 (Figure 5). Thus, aberrant methylation of the region around the transcription start site (regions C and D) was closely associated with silencing p57KIP2 expression. Pretreating cells with 5-aza-2'-deoxycytidine (5-aza-dC), a methyltransferase inhibitor, restored expression of p57KIP2, con®rming the role of methylation in silencing p57KIP2 (Figure 4b). On the other hand, treatment with Trichostatin A (TSA), a histone deacetylase inhibitor, did not restore gene expression (Figure 4b). Histone acetylation, p57KIP2 methylation and gene expression It was recently shown that methylation-dependent gene silencing is associated with altered chromatin structure, including hypoacetylation of histone (Bird and Wole, 1999). To assess the level of histone acetylation associated with p57KIP2 silencing, we performed chromatin immunoprecipitation (ChIP) assays using anti-histone H3 and H4 antibodies. Thereafter, PCR Oncogene


2744

Figure 4 Analysis of p57KIP2 expression in various tumor cell lines and tissues. (a) Expression of p57KIP2; total RNA from the cell lines was reverse transcribed (RT+) and ampli®ed by PCR. The cell lines and tissues examined are indicated above the gels. Each PCR reaction is accompanied with negative control (RT7). GAPDH was ampli®ed to con®rm the integrity of the RNA. BM, normal bone marrow; Ly, normal lymphocyte. Group 1, cell lines showing no methylation in regions C, D and E; Group 2, cell lines showing methylation of all regions analysed. (b) Expression of p57KIP2 before (No treat) and after treatment with 5-aza-dC (AzadC) or TSA

dC+TSA (Figure 6b) showed that TSA alone or in combination with 5-aza-dC signi®cantly enhanced histone acetylation at p57KIP2. By contrast, histone acetylation was increased only slightly in cells treated with 5-aza-dC alone. Figure 5 Schematic diagram summarizing the methylation of the p57KIP2 CpG island. The cell lines were divided into two distinct groups based on the methylation status of the six regions examined. The percentage of methylated alleles in each group was determined as in Figure 3, and the average percentage of methylated alleles at each site is shown below the circles. The regions (A ± F) and restriction enzymes used are indicated at the top. The expression status of p57KIP2 is shown on the right

was carried out using primers designed to amplify the region whose methylation was correlated with loss of expression (locations 772 to +89 from the transcription start site). Following this protocol, PCR products were detected only in cell lines in which region C was not methylated (MKN45, RPMI8226 and K562) (Figure 6a), indicating histone in the 5' region of p57KIP2 to be hypoacetylated in methylated cell lines (MKN28, HS-Sultan and Raji). Moreover, ChIP analysis using the immunoprecipitated DNA from MKN28 cells treated with 5-aza-dC, TSA or 5-azaOncogene

Aberrant methylation of p57KIP2 in primary tumors To determine whether aberrant methylation of p57KIP2 is involved in primary human cancers, methylation analysis was carried out in a panel of primary tumors from various tissue types. In normal tissues, methylation of region C was barely detectable (Figure 7a ± e). By contrast, of 35 gastric cancers examined, 5 (14%) showed aberrant methylation (Figure 5a), as did 6 out of 20 (30%) hepatocellular carcinomas (Figure 7b). In two cases, the normal gastric mucosa adjacent to the tumors also showed dense methylation, but in all others methylation of region C was cancer speci®c. None of the 37 colorectal cancer cases showed methylation of p57KIP2 (Figure 5c), but 2 out of 18 pancreatic cancers (11%) and 7 out of 25 (28%) AML cases did so (Figure 5d,e). The AML cases examined had been previously analysed for methylation of multiple CpG islands; methylation of p57KIP2 was detected exclusively in those exhibiting


presence of CIMP (P50.001, Fisher's exact test, two sided).

2745

Discussion

Figure 6 Acetylation status of histones H3 and H4 in tumor cell lines examined by ChIP. (a) The region around transcription start site (772 to +89) was ampli®ed by PCR after chromatin immunoprecipitation using anti-acetylated histone H3 and H4 antibodies. Reactions were controlled in two ways: ampli®cation of DNA before precipitation (Input) and ampli®cation of the 5' region of GAPDH using DNA precipitated with anti-acetylated histone antibody. The cell lines examined are indicated above the gel. Antibodies used for immunoprecipitation are shown below. Also shown are negative controls carried out using aliquots precipitated with no antibody (No Ab). Bands produced by the ChIP-PCR products of p57KIP2 and GAPDH were quanti®ed by densitometry, and the ratios of the signals from p57KIP2 and GAPDH are shown. (b) Acetylation status of histone H3 after treatment of methyltransferase and/or histone deacetylase inhibitor. MKN28 cells were treated with 0.1 mM 5-aza-dC, 300 nM TSA or 0.2 mM 5-aza-dC+300 nM of TSA. ChIP assays were performed using anti-acetyl histone H3 antibody followed by PCR ampli®cation

characteristics of the CpG island methylator phenotype (CIMP) (Toyota et al., 1999a, 2001). In addition, 6 out of 6 (100%) CIMP+ cases showed methylation of p57KIP2, whereas only 1 out of 19 (5%) CIMP7 case did so, which indicated a signi®cant correlation between methylation and the

p21WAF1, p27KIP1 and p57KIP2 belong to the CIP/ KIP family of CDK inhibitors, which play key roles in cell cycle regulation. Impairment of p53, a transactivator of p21WAF1, silences expression of aected genes (el-Deiry et al., 1993), while diminished expression of p27KIP1 is associated with various tumors and with increased degradation of protein via the ubiquitin proteasome pathway (Pagano et al., 1995). Although decreased expression of p57KIP2 has also been reported (Matsumoto et al., 2000; Oya and Schulz, 2000; Schwienbacher et al., 2000; Shin et al., 2000a), the precise molecular mechanism responsible remains unknown. We have shown herein that p57KIP2 is aberrantly methylated in human gastric, hepatocellular and hematopoietic tumors. This methylation of p57KIP2 appears to be gene speci®c, as methylation is not primarily responsible for the inactivation of p21WAF1 or p27KIP1, though they also have 5' CpG islands (Figure 1c). Moreover, the frequency of p57KIP2 methylation varied among tumor types. Similar tissue-speci®c methylation has been reported for p15INK4B, which is frequently methylated in hematopoietic but not colorectal tumors (Herman et al., 1996), and for BRCA1, which is speci®cally methylated in breast and ovary tumors (Esteller et al., 2000). It is also possible that other CIP/KIP family members (e.g., p27KIP1) are involved in the tumorigenesis in tissues where p57KIP2 is not methylated (Loda et al., 1997). Alternatively, p57KIP2 may be inactivated by other mechanisms, such as mutation and deletion. In fact, chromosome 11p15, where p57KIP2 is located, is deleted in 20 ± 70% of human tumors of the stomach, liver, breast and lung and in AML (Bepler and GarciaBlanco, 1994; Krskova-Honzatkova et al., 2001; Sheu et al., 1999; Shin et al., 2000a; Winqvist et al., 1993). Mutation of p57KIP2 has rarely been detected (Shin et al., 2000a); instead, DNA methylation and gene deletion appear to be the major mechanisms by which p57KIP2 is inactivated. Several lines of evidence suggest that p57KIP2 is an imprinted gene (Hatada and Mukai, 1995; Matsuoka et al., 1996), though the role of DNA methylation in the imprinting of p57KIP2 is unknown. Our bisul®te-PCR analysis showed that with the exception of the gastric mucosa adjacent to the lesions in two cases, dense methylation of p57KIP2 does not occur in normal tissues. Consistent with that ®nding, Chung et al. (1996) reported that the 5' region of p57KIP2 is not highly methylated in normal tissues. Nevertheless, methylation may still be an important regulator of p57KIP2 expression. For example, embryonic ®broblasts from mice in which DNMT1 was disrupted by homologous recombination exhibit higher levels of p57KIP2 expression than ®broblasts from wild type Oncogene


2746

Figure 7 Analysis of p57KIP2 methylation in primary gastric (a), hepatocellular (b), colorectal (c) and pancreatic (d) cancers and AML (e). The methylation status of region C was examined using bisul®te-PCR, followed by digestion with EcoRI, electrophoresis and staining with ethidium bromide. The per cent methylation quanti®ed densitometrically is shown below the gels. N, paired normal tissues; T, tumors; M, methylated alleles; CEM, AML cell line; BM, normal bone marrow; Ly, normal lymphocyte

mice (Jackson-Grusby et al., 2001). Further study will be required to clarify the precise function of methylation in the imprinting of p57KIP2. Aberrant methylation in tumors may be random with stochastic selection during tumor progression, or it may be a dynamic and progressive process (Cameron et al., 1999b; Wong et al., 1999). Our ®ndings indicate that the pattern of p57KIP2 methylation in tumor cell lines can be divided into two groups: one with extensive methylation of the entire CpG island and one methylated only at the edges of the island (regions A, B and F) but not at the heart (regions C, D and E). Interestingly, it appears that methylation of CpG islands downstream of the transcription initiation site does not suppress gene expression (Jones, 1999; Salem et al., 2000). The spread of methylation into the region around the transcription start site is thus critical for the silencing of the gene, which suggests that alteration of cis or transacting factors that normally protect this region from such aberrant modi®cation may be key to its methylation (Mummaneni et al., 1998). Although the link between methylation and gene silencing is apparent from our data, the precise mechanism by which the methylated promoter is silenced is not. Recent studies, however, suggest histone deacetylation may play a key role (Bird and Wole, 1999; Cameron et al., 1999a). Consistent with those results, histone precipitated from cell lines expressing p57KIP2 was hyperacetylated, while that from cell lines Oncogene

not expressing p57KIP2 was hypoacetylated. On the other hand, whereas treatment with a histone deacetylase inhibitor, TSA, restored acetylation of histone in a methylated cell line (Figures 4b and 6b), expression of p57KIP2 was not restored unless methylation was inhibited. This suggests that methylation of cytosine suppresses gene transcription in both a histone deacetylation dependent and independent manner. The nature of the molecular mechanism by which DNA methylation silences a gene in the presence of acetylated histone remains unknown. However, one important element may be the structure of the chromatin associated with DNA methylation, which appears to be regulated by protein complexes such as NURD/MBD3 and Sin3A/MeCP2 (Bird and Wole, 1999), as well as by a recently identi®ed complex that includes MBD1 as one component (Fujita et al., 2000; Ng et al., 2000). The fact that DNA methylation mediates stable suppression of transcription (Bird, 2002) is consistent with our ®nding that expression of p57KIP2 is not restored by acetylation of histone. Further study will be necessary to clarify which MBDs and HDACs are involved in silencing p57KIP2. We previously reported that some of the genes frequently methylated in tumors are also methylated in normal aged tissues (Issa, 2000; Toyota et al., 1999a). Other genes, by contrast, are methylated exclusively in a subset of tumors that exhibit the CpG island methylator phenotype (Toyota et al., 1999a,b, 2001;


Ueki et al., 2000). In the present study, we found a strong correlation between methylation of p57KIP2 in AML and the presence of the CpG island methylator phenotype, which evolves in AML via a distinct molecular pathway leading to hypermethylation of multiple CpG islands and is predictive of a poor prognosis (Toyota et al., 2001). p57KIP2 thus appears to be a target gene associated with the CpG island methylator phenotype, along with others that include p16INK4A, hMLH1, THBS1, CACNA1G and COX-2 (Toyota et al., 1999a,c, 2000). In summary, the data presented suggest that p57KIP2 is inactivated by aberrant methylation in a variety of tumors. That diminished or abolished p57KIP2 expression is closely associated with regional methylation and histone deacetylation suggests the use of methyltransferase and histone deacetylase inhibitors to attenuate epigenetic silencing may represent a new therapeutic approach to the treatment of certain tumors.

Materials and methods Cell lines and specimens Ten colorectal cancer (Caco2, RKO, SW48, SW480, Colo201, Colo205, Colo320, BM314, DLD-1, and HT29), eight gastric cancer (MKN7, MKN28, MKN45, MKN74, NUGC3, KatoIII, AZ521 and JRST), two pancreatic cancer (Panc-1 and MIAPaCa-2), ®ve hepatocellular cancer (CHC20, CHC4, HLE, HUH7, and CHC32), and 14 hematopoietic tumor cell lines (HO3238, KMS12PE, RPMI8226, K562, KR12, Daudi, BALL-1, Jurkat, HS-sultan, RPMI1788, SCC3, Raji, IM9, and ALT2) were obtained either from the American Type Culture Collection (Manassas, VA, USA) or from the

Japanese Collection of Research Bioresources (Tokyo, Japan). The cells were cultured in the appropriate medium until harvested for extraction of DNA using standard procedures or extraction of total RNA using Isogen (Nippon Gene, Tokyo, Japan). In some cases, four cell lines (HT-29, MKN28, HS-Sultan and Raji) were treated with 5-aza-dC (SIGMA) for 96 h prior to extracting the RNA. Specimens from 37 cases of primary colorectal cancer, 35 cases of primary gastric cancer, 20 cases of primary hepatocellular carcinoma, 18 cases of pancreatic cancer and 25 cases of AML were also analysed. All of the colorectal cancers and AML cases and some of the gastric cancers used in this analysis were described previously (Toyota et al., 1999a,b, 2001).

2747

Bisulfite-PCR and sequencing Bisul®te treatment was carried as described previously (Clark et al., 1994). Two-microliter aliquots were ampli®ed in a 50 ml reaction mixture consisting of 16MSP buer (67 mM TrisHCl (pH 8.8), 16.6 mM (NH4)2SO4, 6.7 mM MgCl2, and 10 mM beta-mercaptoethanol) containing 2.5 mM each dNTP mixture, 0.5 mM each primer and 1 U of Taq polymerase (TaKaRa, Japan). In total, six regions (A-F) were ampli®ed using primers that amplify both methylated and unmethylated alleles equally (p57A-p57F). PCR products were then digested with restriction enzymes that recognize the methylated CpG sites retained after bisul®te treatment (Xiong and Laird, 1997). After digestion, the DNA was precipitated with ethanol, electrophoresed in 3% agarose gels and stained with ethidium bromide. The intensity of the methylated alleles was calculated by densitometry using a Lane and Spot Analyser 6.0 (Atto, Japan). Signals above 5% were considered to re¯ect the presence of methylation, as signals below 5% were not reproducible. The primer sequences, PCR parameters and restriction enzymes used are summarized in Table 1. To sequence bisul®te-PCR products, fragments ampli®ed using primers p57-BF and p57-CR were cloned into pCR2 vector using a TOPO-cloning KIT (Invitrogen). DNA sequencing

Table 1 Primers used for Bisul®te-PCR, RT ± PCR and ChIP Primers

Sequence

p57-A

F: 5'-TAGTAGGAAATAATGGTTTTTTTTGG-3' R: 5'-AAAATAAAAAATCTACCCRAAATCA-3' F: 5'-GTTAGTTGGYGTAGGAGGTTTA-3' R: 5'-RCCRACTCCTTTATCTACAAAC-3' F: 5'-GGTTGGGYGTTTTATAGGTTA-3' R: 5'-ACCTAACTATCCGATAATAAACTCTTC-3' F: 5'-GTTTGYGTAGTTTYGGGTTATGTT-3' R: 5'-ATTCTAATCCTCRACRTTCAACTC-3' F: 5'-AGGAGTTTTTYGTTGATTGTTGTATT-3' R: 5'-ACCCCCRAAAACTAAAAAAACC-3' F: 5'-GTAAGAGGTTGYGGTGAGTTAAGTG-3' R: 5'-AAACCCAACRCCCTTCCAAC-3' F: 5'-GGYGGGGYGGTTGTATATTAG-3' R: 5'-CRACCCRAAATCCCCTATTATC-3' F: 5'-GAGGGTAGTYGTTGGGTTTT-3' R: 5'-CTTAAAATACTCCRCCTACCTAAC-3' F: 5'-CTG ACC AGC TGC ACT CGG GGA TTT C-3' R: 5'-GCC GCC GGT TGC TGC TAC ATG A-3' F: 5'-GTATAAAGGGGGCGCAGGCGGGCT-3' R: 5'-TGGTGGACTCTTCTGCGTCGGGTTC-3' F: 5'-TCGGTGCGTGCCCAGTTGAACC-3' R: 5'-ATGCGGCTGACTGTCGAACAGGAG-3'

p57-B p57-C p57-D p57-E p57-F p21WAF1 p27KIP1 p57RT-F p57RT-R p57ChIP-F P57ChIP-R GAPDHChIP-F GAPDHChIP-R

Condition: 8C (cycles)

Size (bp)

Enzyme

53(3),51(4),49(5),47(26)

193

TaqI

53(3),51(4),49(5),47(26)

163

58(3),56(4),54(5),52(26)

155

53(3),51(4),49(5),47(26)

196

HhaI TaqI EcoRI HhaI BstUI

58(3),56(4),54(5),52(26)

196

61(3),59(4),57(5),55(26)

170

60(3),58(4),56(5),54(26)

168

HpaI RsaI TaqI HhaI BstUI

55(3),53(4),51(5),49(26)

191

BstUI

60(35)

341

62(35)

161

62(32)

246

Y=C or T, R=A or G Oncogene


2748

was then carried out using an ABI377 automated sequencer; at least 10 clones were sequenced for each cell line. RT ± PCR Samples of total RNA (5 mg) were treated for 30 min with DNaseI, which was then inactivated using a DNA Free Kit (Ambion) according to the manufacturer's instructions. Reverse transcription was performed using Superscript II (Life Technologies), after which PCR was performed using primers p57RT-F and p57RT-R (Table 1). To assess the integrity of the RNA, GAPDH was ampli®ed as described previously (Toyota et al., 1999c). Aliquots (10 ml) of the PCR products were electrophoresed in 2% agarose gels and stained with ethidium bromide.

immune complexes were recovered using protein A-agarose, puri®ed by phenol/chloroform extraction, precipitated with ethanol, and resuspended in distilled water. Samples containing 1/100 of the immunoprecipitated DNA were used for PCR, which was carried out in a solution containing 16PCR buer (TaKaRa), 0.5 mM each primer (p57ChIP-F and p57ChIP-R (Table 1)), 0.25 mM dNTP mixture and 2.5 U of Taq polymerase (TaKaRa). As a positive control, the 5' region of GAPDH was ampli®ed using primers GAPDHChIP-F and GAPDHChIP-R (Table 1). PCR products were analysed by agarose gel electrophoresis.

Chromatin immunoprecipitation

Abbreviations CDK, cyclin dependent kinase; RT ± PCR, reverse transcriptase PCR; TSA, trichostatin A; HDAC, histone deacetylase; CIMP, CpG island methylator phenotype

To cross link DNA with chromatin, 16106 cells were incubated for 10 min in 1% formaldehyde at 378C. The cells were then harvested, washed with phosphate buered saline (PBS), resuspended in lysis buer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.0, with protease inhibitor), and the DNA with chromatin was broken into 200-1000 bp fragments by sonication. Samples containing 1/100 of the resultant solution were used as an internal control for the amount of DNA (Input). The remainder was immunoprecipitated for 16 h at 48C using anti-acetylated histone H3 and antiacetylated histone H4 antibodies as probes (Upstate Biotechnologies, Lake Placid, NY, USA). The precipitated

Acknowledgments This study is supported by the Grant-in-Aid for Scienti®c Research on Priority Areas from the Ministry of Eduction, Culture, Sport, Science, and Technology (M Toyota, F Itoh, T Tokino and K Imai). T Kikuchi is a research fellow from the Japanese Society for the Promotion of Science. H Suzuki is a postdoctoral fellow from the Japanese Society for the Promotion of Science. The authors thank Dr William F Goldman for editing the manuscript.

References Baylin SB, Esteller M, Rountree MR, Bachman KE, Schuebel K and Herman JG. (2001). Hum. Mol. Genet., 10, 687 ± 692. Bepler G and Garcia-Blanco MA. (1994). Proc. Natl. Acad. Sci. USA, 91, 5513 ± 5517. Bird A. (2002). Genes Dev., 16, 6 ± 21. Bird AP and Wole AP. (1999). Cell, 99, 451 ± 454. Cameron EE, Bachman KE, Myohanen S, Herman JG and Baylin SB. (1999a). Nat. Genet., 21, 103 ± 107. Cameron EE, Baylin SB and Herman JG. (1999b). Blood, 94, 2445 ± 2451. Chung WY, Yuan L, Feng L, Hensle T and Tycko B. (1996). Hum. Mol. Genet., 5, 1101 ± 1108. Clark SJ, Harrison J, Paul CL and Frommer M. (1994). Nucleic Acids Res., 22, 2990 ± 2997. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein B. (1993). Cell, 75, 817 ± 825. Elledge SJ. (1996). Science, 274, 1664 ± 1672. Esteller M, Silva JM, Dominguez G, Bonilla F, Matias-Guiu X, Lerma E, Bussaglia E, Prat J, Harkes IC, Repasky EA, Gabrielson E, Schutte M, Baylin SB and Herman JG. (2000). J. Natl. Cancer Inst., 92, 564 ± 569. Fujita N, Shimotake N, Ohki I, Chiba T, Saya H, Shirakawa M and Nakao M. (2000). Mol. Cell. Biol., 20, 5107 ± 5118. Hatada I and Mukai T. (1995). Nat. Genet., 11, 204 ± 206. Hatada I, Ohashi H, Fukushima Y, Kaneko Y, Inoue M, Komoto Y, Okada A, Ohishi S, Nabetani A, Morisaki H, Nakayama M, Niikawa N and Mukai T. (1996). Nat. Genet., 14, 171 ± 173. Herman JG, Jen J, Merlo A and Baylin SB. (1996). Cancer Res., 56, 722 ± 727. Oncogene

Issa JP. (2000). Curr. Top. Microbiol. Immunol., 249, 101 ± 118. Jackson-Grusby L, Beard C, Possemato R, Tudor M, Fambrough D, Csankovszki G, Dausman J, Lee P, Wilson C, Lander E and Jaenisch R. (2001). Nat. Genet., 27, 31 ± 39. Jones PA. (1999). Trends Genet., 15, 34 ± 37. Jones PA and Laird PW. (1999). Nat. Genet., 21, 163 ± 167. Krskova-Honzatkova L, Cermak J, Sajdova J, Stary J, Sedlacek P and Sieglova Z. (2001). Leuk. Res., 25, 45 ± 53. Loda M, Cukor B, Tam SW, Lavin P, Fiorentino M, Draetta GF, Jessup JM and Pagano M. (1997). Nat. Med., 3, 231 ± 234. Matsumoto M, Furihata M, Ohtsuki Y, Sasaguri S and Ogoshi S. (2000). Anticancer Res., 20, 1947 ± 1952. Matsuoka S, Edwards MC, Bai C, Parker S, Zhang P, Baldini A, Harper JW and Elledge SJ. (1995). Genes Dev., 9, 650 ± 662. Matsuoka S, Thompson JS, Edwards MC, Bartletta JM, Grundy P, Kalikin LM, Harper JW, Elledge SJ and Feinberg AP. (1996). Proc. Natl. Acad. Sci. USA, 93, 3026 ± 3030. Mummaneni P, Yates P, Simpson J, Rose J and Turker MS. (1998). Nucleic Acids Res., 26, 5163 ± 5169. Ng HH, Jeppesen P and Bird A. (2000). Mol. Cell. Biol., 20, 1394 ± 1406. Oya M and Schulz WA. (2000). Br. J. Cancer, 83, 626 ± 631. Pagano M, Tam SW, Theodoras AM, Beer-Romero P, Del Sal G, Chau V, Yew PR, Draetta GF and Rolfe M. (1995). Science, 269, 682 ± 685. Salem CE, Markl ID, Bender CM, Gonzales FA, Jones PA and Liang G. (2000). Int. J. Cancer, 87, 179 ± 185.


Schwienbacher C, Gramantieri L, Scelfo R, Veronese A, Calin GA, Bolondi L, Croce CM, Barbanti-Brodano G and Negrini M. (2000). Proc. Natl. Acad. Sci. USA, 97, 5445 ± 5449. Sherr CJ. (2000). Cancer Res., 60, 3689 ± 3695. Sheu JC, Lin YW, Chou HC, Huang GT, Lee HS, Lin YH, Huang SY, Chen CH, Wang JT, Lee PH, Lin JT, Lu FJ and Chen DS. (1999). Br. J. Cancer, 80, 468 ± 476. Shin JY, Kim HS, Lee KS, Kim J, Park JB, Won MH, Chae SW, Choi YH, Choi KC, Park YE and Lee JY. (2000a). Exp. Mol. Med., 32, 79 ± 83. Shin JY, Kim HS, Park J, Park JB and Lee JY. (2000b). Cancer Res., 60, 262 ± 265. Tokino T, Urano T, Furuhata T, Matsushima M, Miyatsu T, Sasaki S and Nakamura Y. (1996). Hum. Genet., 97, 625 ± 631. Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB and Issa JP. (1999a). Proc. Natl. Acad. Sci. USA, 96, 8681 ± 8686. Toyota M, Ahuja N, Suzuki H, Itoh F, Ohe-Toyota M, Imai K, Baylin SB and Issa JP. (1999b). Cancer Res., 59, 5438 ± 5442. Toyota M, Ho C, Ohe-Toyota M, Baylin SB and Issa JP. (1999c). Cancer Res., 59, 4535 ± 4541.

Toyota M, Kopecky KJ, Toyota MO, Jair KW, Willman CL and Issa JP. (2001). Blood, 97, 2823 ± 2829. Toyota M, Shen L, Ohe-Toyota M, Hamilton SR, Sinicrope FA and Issa JP. (2000). Cancer Res., 60, 4044 ± 4048. Tycko B. (2000). J. Clin. Invest., 105, 401 ± 407. Ueki T, Toyota M, Sohn T, Yeo CJ, Issa JP, Hruban RH and Goggins M. (2000). Cancer Res., 60, 1835 ± 1839. Watanabe H, Pan ZQ, Schreiber-Agus N, DePinho RA, Hurwitz J and Xiong Y. (1998). Proc. Natl. Acad. Sci. USA, 95, 1392 ± 1397. Winqvist R, Mannermaa A, Alavaikko M, Blanco G, Taskinen PJ, Kiviniemi H, Newsham I and Cavenee W. (1993). Cancer Res., 53, 4486 ± 4488. Wong DJ, Foster SA, Galloway DA and Reid BJ. (1999). Mol. Cell. Biol., 19, 5642 ± 5651. Xiong Z and Laird PW. (1997). Nucleic Acids Res., 25, 2532 ± 2534. Yan Y, Frisen J, Lee MH, Massague J and Barbacid M. (1997). Genes Dev., 11, 973 ± 983. Zhang P, Liegeois NJ, Wong C, Finegold M, Hou H, Thompson JC, Silverman A, Harper JW, DePinho RA and Elledge SJ. (1997). Nature, 387, 151 ± 158.

2749

Oncogene