CDX2, a homeobox transcription factor, upregulates ... - Nature

Oncogene (2003) 22, 7942–7949

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

CDX2, a homeobox transcription factor, upregulates transcription of the p21/WAF1/CIP1 gene Yun-Qing Bai1,2, Satoshi Miyake1, Takehisa Iwai2 and Yasuhito Yuasa*,1 1

Department of Molecular Oncology, Graduate School of Medicine and Dentistry, Tokyo Medical and Dental University, Tokyo 113-8519, Japan; 2Department of Surgery, Graduate School of Medicine and Dentistry, Tokyo Medical and Dental University, Tokyo 113-8519, Japan

The CDX2 homeobox transcription factor plays key roles in intestinal development and homeostasis. CDX2 is downregulated during colorectal carcinogenesis, whereas overexpression of CDX2 results in growth inhibition and differentiation of colon carcinoma and intestinal cells. However, the means by which CDX2 functions remain poorly understood. p21/WAF1/CIP1 is one of the cyclindependent kinase inhibitors. In addition to its role in cell cycle control, p21 plays critical roles in differentiation and tumor suppression. The overlapping in both the expression and function of CDX2 and p21 in the small intestine and colon strongly suggests a link between these two genes. By means of luciferase reporter and electrophoretic mobility shift assays, we show here that CDX2 transactivated and physically interacted with the promoter of p21 in a p53independent manner. Moreover, overexpression of CDX2 increased the mRNA expression of p21 in HT-29 colon carcinoma cells, as demonstrated by reverse transcription– polymerase chain reaction. These data suggest that p21 is a transcriptional target of CDX2. Our results may thus provide a new mechanism underlying the functions of CDX2. Oncogene (2003) 22, 7942–7949. doi:10.1038/sj.onc.1206634 Keywords: CDX2; transcription; p21/WAF1/CIP1

Introduction Human CDX2 is a member of the caudal-related homeobox gene family (McGinnis and Krumlauf, 1992; Mallo et al., 1997). During early development in Drosophila, caudal is involved in anterior–posterior patterning (Macdonald and Struhl, 1986). The expression of the rodent Cdx2 homeobox gene is tissue specific and present from the early embryo to the adult (James et al., 1994), and thus it is likely that Cdx2 plays roles in both the establishment and maintenance of the intestinal *Correspondence: Y Yuasa, Department of Molecular Oncology, Graduate School of Medicine and Dentistry, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan; E-mail: [email protected] Received 2 January 2003; revised 31 March 2003; accepted 7 April 2003

epithelial phenotype. Previous studies suggested that CDX2 is a tumor suppressor. Firstly, heterozygous Cdx2 knockout mice develop multiple colonic polyps (Chawengsaksophak et al., 1997). Secondly, the expression of CDX2 is downregulated during colorectal carcinogenesis (Ee et al., 1995; Mallo et al., 1997). Thirdly, CDX2 overexpression seems to inhibit growth and/or promote differentiation of colorectal cancer cells (Mallo et al., 1998) or undifferentiated intestinal epithelial cells (Suh and Traber, 1996). Moreover, the ectopic expression of CDX2/Cdx2 has been reported to be associated with intestinal metaplasia formation of the stomach (Bai et al., 2002; Silberg et al., 2002; Mutoh et al., 2002). Nevertheless, the molecular mechanisms underlying these roles of CDX2 remain poorly understood. The homeobox is a highly conserved 180-bp DNA sequence encoding a 60-amino-acid motif termed the homeodomain. The homeodomain, with a helix–turn– helix structural conformation, is the sequence-specific DNA-binding domain of a family of transcriptional regulatory proteins (McGinnis and Krumlauf, 1992). Homeoproteins of the caudal family bind DNA via an AT-rich sequence whose consensus is A/CTTTATA/G (reviewed by Freund et al., 1998). Previous studies showed that CDX2 binds cis-elements present in the gene promoters of enterocytic markers such as sucraseisomaltase, lactase-phlorizin hydrolase, phospholipaseA, apolipoprotein B, carbonic anhydrase 1, and calbindin D9K (Freund et al., 1998). Homeobox genes such as Hox-C8 and HNF-1 have also been demonstrated to be transcriptional targets of CDX2 (Freund et al., 1998). In addition, expression of the gut-enriched Krupel-like factor (GKLF) gene, an epithelial-specific transcription factor that functions as a suppressor of cell proliferation, is dependent on CDX2 in human colon cancer cell line RKO (Dang et al., 2001). Importantly, CDX2 can regulate the expression of the liver intestinecadherin (LI- cadherin) gene, indicating its key role in mediating CDX2 function in intestinal cell fate determination (Hinoi et al., 2002). Clearly, further investigations are needed to clarify the complex network involved in CDX2. p21/WAF1/CIP1 (hereafter referred to as p21) was one of the first cyclin-dependent kinase (cdk) inhibitors identified, and it can bind and inhibit G1 cyclin/Cdk

CDX2 up-regulates transcription of p21 Y-Q Bai et al

7943

complexes (Harper et al., 1993; Xiong et al., 1993). p21 can induce growth suppression in vivo through cyclin D1-Cdk4 and cyclin E-Cdk2 complexes (Lin et al., 1996), and overexpression of p21 results in cell cycle arrest in G1 (Xiong et al., 1993; Luo et al., 1995) and tumor cell growth suppression (El-Deiry et al., 1993). In addition to its role in cell cycle control, p21 is also believed to inhibit DNA replication through its ability to bind proliferating cell nuclear antigen (PCNA), which is required for both replicative DNA synthesis and DNA repair. However, p21 has no inhibitory effect on the DNA repair function of PCNA (Li et al., 1994; Waga et al., 1994). Unexpectedly, p21 plays critical roles in differentiation, such as the role in control of epithelial self-renewal and commitment to differentiation (Di Cunto et al., 1998). Finally, p21 has been proved to be an unsuspected tumor suppressor, which can mediate at least some of the tumor-suppressing effects of p53 (reviewed by Dotto, 2000). Since firstly CDX2 and p21 are both associated with differentiation induction and tumor growth suppression, and secondly their expression overlaps in the small intestine and colon (Gartel et al., 1996; Polyak et al., 1996; Mallo et al., 1997; Silberg et al., 2000), a possible link between these two genes is suggested. In the present study, we first identified nine CDX-core binding sequences, TTTAT, in the sense or, ATAAA, the reverse orientation (Suh et al., 1994) within the human p21 promoter (2699 to þ 45 bp). By means of luciferase reporter assay, electrophoretic mobility shift assay (EMSA) and reverse transcription–polymerase chain reaction (RT–PCR), we then demonstrated that CDX2 activated transcription of the p21 gene through direct physical interaction with the p21 promoter.

Results Transactivation of the p21 promoter by CDX2 To determine whether or not the p21 gene is a transcriptional target of CDX2 and, if so, whether or not p53 is involved, SAOS-2 (p53/) and U-2OS (p53 þ / þ ) osteosarcoma cells were cotransfected with the p21Luc reporter construct and expression vector pcDNA3. 1CDX2. These two cell lines were selected because they are widely used for p53-related studies and lack endogenous CDX2 expression (data not shown). According to the dual-luciferase reporter assay, cotransfection of pcDNA3.1CDX2 resulted in 11.1- and 3.3-fold transcriptional activation of the 2744-bp p21 promoter construct in SAOS-2 and U-2OS, respectively (Figure 1). Then the ability of CDX2 to activate p21 transcription was further studied in gastrointestinal cancer cell lines GT3TKB and HT-29. The endogenous CDX2 expression was below the detectable level in GT3TKB (Bai et al., 2000) and was also very low in HT-29 cells. As a result, overexpression of CDX2 induced p21 activity by 6.4- and 2.6-fold in GT3TKB and HT-29 (Figure 1), respectively. In contrast, no transcriptional activation of p21-Luc was observed after CDX1 transfection in the

Figure 1 Effect of CDX2 on the p21 promoter. (a) Reporter construct p21-Luc, in which the p21 promoter drives expression of the firefly luciferase gene, was cotransfected with the expression vector pcDNA3.1CDX2 or the empty vector in different cells, as indicated. The renilla reporter plasmid (1 : 10 ratio with target reporter) was also included in all experiments to normalize variation in transfection efficiency. Cells were harvested after 48 h and assayed for luciferase activity. All values were equalized on the basis of the activity observed upon cotransfection with a control renilla expression vector, and expressed as ratio of CDX2expressing vector to empty vector. These results represent the means and standard deviation of three independent experiments. (b) p21-Luc, p21-Luc/D1 or p21-Luc/D2 were cotransfected with the expression vector pcDNA3.1CDX2 or the empty vector in HT-29 cells. p21-Luc/D1 and p21-Luc/D2 were constructed by deleting 1900 to 658 bp or 2504 to 212 bp of the p21-Luc construct, respectively, as described under ‘Materials and methods’. Numbers on the promoter refer to positions relative to the transcription start site of the human p21 gene ( þ 1). Note that no CDX binding site remains upstream of position 1900 bp. The position of p21-WT, which was used as probe for subsequent EMSA, containing three clustered CDX binding sites is also indicated (471 to 434 bp). The activity of p21-Luc was arbitrarily set at 100%. Variations in transfection efficiency were corrected by normalization for expression of renilla luciferase. All transfections were repeated three times in duplicate

four cell lines described above (data not shown), indicating specific transactivation by CDX2 of the p21 promoter. Since p53 is null in SAOS-2 and mutated in HT-29 cells, CDX2 transactivated the p21 promoter in a p53-independent manner. To determine whether or not the activation of the p21 promoter by CDX2 is dependent on CDX binding sites, deletion mutants were made as described in ‘Materials and methods’. As a result, deletion of putative CDX binding sites on the 50 side from 1900 to 658 bp (p21Luc/D1) or all the binding sites (p21-Luc/D2) in the p21 promoter resulted in 35 and 79% reduction in the p21 promoter activity, respectively (Figure 1b), suggesting that these binding sites are crucial for CDX2 to transactivate the p21 promoter. CDX2 binds to the p21 promoter To examine the physical interaction between the p21 promoter and the CDX2 homeodomain protein, we Oncogene


7944

performed EMSA. The oligonucleotides, p21-WT and p21-Mut, used in this study are shown in Figures 1b and 2a. The reasons for choosing these oligonucleotides were as follows. (1) Deletion mutant analysis indicated that the portion from 658 to 212 bp contains important CDX2 binding sites as described above (Figure 1b). (2) Three CDX-core binding sequences, TTTAT, are located within the DNA fragment of less than 30 bp in the p21 promoter, suggesting that these oligonucleotides may represent the most possible region for the interaction with CDX2. (3) Another homeodomain protein, HOXA10, has been demonstrated to bind to this region in the p21 promoter (Bromleigh and Freedman, 2000). p21-WT was labeled, incubated in the presence of nuclear extract from COS-7 cells transfected with CDX2, and then migrated through a 5.5% nondenaturing polyacrylamide gel. The autoradiograph in Figure 2b reveals the position of the rapidly migrating unbound probe at the base of the gel and several DNA/protein complexes of slower mobility that were formed after incubation with the COS-7 nuclear extract (lane 2), whereas no band shift occurred in lane 1, in which no

nuclear extract was added. The DNA/protein complexes were competed for by 50–500-fold excess unlabeled wildtype oligonucleotide p21-WT (lanes 3–5), but were not competed for by even 500-fold unlabeled excess mutanttype oligonucleotide p21-Mut (lanes 6–8) or nonspecific oligonucleotide (data not shown), suggesting that the CDX-core binding sites in p21-WT are critical for their interaction with CDX2. Gel shifts with nuclear extract prepared from CDX1-transfected or parental COS-7 cells yielded no complex formation (data not shown). To further determine the specificity of the interaction between p21-WT and CDX2, we incubated the radiolabeled p21-WT probe and nuclear extract from CDX2-transfected COS-7 cells in the absence or presence of a polyclonal CDX2 antibody. The CDX2 antibody recognized the nuclear protein bound to the p21-WT probe, resulting in reduced gel mobility of C1 and C2, and a supershift (lane 9). In contrast, no supershift was observed for the binding reaction carried out in the presence of nonimmune serum (lane 10). As the upper two DNA/protein complexes mostly supershifted upwardly, we reasoned that these two bands represent the p21-WT/CDX2 complexes (indicated by C1 and C2). The C1 (lower band) and C2 (upper band) complexes may be formed from binding of CDX2 monomer and dimer on the p21-WT DNA element, respectively, similar to those of the previous report (Suh et al., 1994). These results suggest that homeodomain protein CDX2 interacts physically with the p21-WT ciselement. P21 expression was increased by CDX2 overexpression in colon carcinoma cell line HT-29

Figure 2 EMSA of the physical interaction between the CDX2 homeoprotein and p21 promoter. (a) Synthetic oligonucleotides p21-WT (wild type) and p21-Mut (mutated) are shown. Note that the CDX-core binding sequences TTTAT (in bold) in p21-WT were changed into GCGTG (in bold) in p21-Mut. The complementary oligonucleotides were annealed and end-labeled with 32P. (b) Nuclear extract (5 mg) from COS-7 cells transfected with CDX2 was incubated with the 32P-labeled p21-WT oligonucleotide probe alone (lane 2) or in the presence of 50, 100, or 500 unlabeled p21WT (lanes 3–5) or p21-Mut (lanes 6–8) oligonucleotide. In lane 1, only 32P-labeled p21-WT oligonucleotide was added. Samples were loaded on a 5.5% acrylamide gel. In addition, gel supershifts were carried out with nuclear extract from CDX2-transfected COS-7 cells and the 32P-labeled p21-WT probe. The reaction mixture was incubated with 2.5 ml of CDX2 antibody (lane 9) or nonimmune serum (lane 10) for 15 min at room temperature. The arrows indicate the free probe (FP), complex 1 (C1), complex 2 (C2), and supershift (SS). Oncogene

Based on the findings of the reporter assay and EMSA, we further examined whether or not CDX2 can transactivate transcription of the endogenous p21 gene. Preconfluent HT-29 cells were transiently transfected with the CDX2 expression vector or the empty vector as a control. Total RNA was isolated 24 h after transfection from either transfected cells or parental cells cultured in parallel. Semiquantitative RT–PCR analysis showed that p21 mRNA expression was increased in HT-29 cells transfected with CDX2 compared with those transfected with the empty vector and parental cells (Figure 3, upper panel). Higher expression of CDX2 in the CDX2-transfected cells was also confirmed (Figure 3, middle panel), whereas the internal standard GAPDH mRNA level remained constant (Figure 3, lower panel). We confirmed the nucleotide sequence of the RT–PCR products by sequencing (data not shown). To confirm the findings of semiquantitative RT–PCR, we performed LightCycler real-time PCR. Specificity of real-time PCR was documented with agarose gel electrophoresis and resulted in a single product with desired length (p21, 417 bp; GAPDH, 457 bp). In addition, a LightCycler melting curve analysis was performed, which resulted in single product specific melting temperatures as follows: p21, 88.81C (Figure 4b), and GAPDH, 88.61C (Figure 4d). Representative amplification of p21 and GAPDH is shown in Figure


7945

Figure 3 Semiquantitative RT–PCR analysis of p21 expression in HT-29 cells. HT-29 cells were transfected with the CDX2 expression vector or the empty vector as a control. Total RNA was isolated 24 h after transfection from either the transfected cells or parental cells, and then cDNA was synthesized. Standard PCR reactions were performed with 30 cycles for p21 (upper panel) and CDX2 (middle panel), and 21 cycles for GAPDH (lower panel). These cycles were observed to be within the logarithmic phase of the amplification. The PCR products (5 ml) were electrophoresed in 2% agarose gels containing 0.5 mg/ml ethidium bromide. The gels were illuminated with UV light and then photographed. Lane 1, parental cells; lane 2, empty vector-transfected cells; lane 3, CDX2transfected cells

4a and c, respectively, the results of which are listed in Table 1. LightCycler real-time PCR was performed in duplicate in two independent experiments, and the results are shown in Figure 4e. The p21 mRNA expression level was expressed as the ratio of the concentrations of p21 and GAPDH. As shown in Figure 4e, the p21 expression level in CDX2-transfected cells was 3.7- and 2.4-fold higher than those in the empty vector-transfected and parental HT-29 cells, respectively. In contrast, the p21 mRNA level in CDX1transfected cells was similar to those in the empty vector-transfected and parental HT-29 cells in a parallel study (data not shown). These results suggest that CDX2, but not CDX1, can transactivate the transcription of the p21 gene.

Discussion The intestinal epithelium is continuously renewed from stem cells anchored near the bottom of invaginations called intestinal crypts or glands. With the exception of Paneth cells, the epithelial cells migrate upwards toward the crypt mouth where proliferation stops, and differentiation is turned on and proceeds along the villi in the small intestine or the upper part of the glands in the colon. Homeobox transcription factor CDX2 seems to have critical functions in the intestinal development, differentiation, and maintenance of the intestinal phenotype (Silberg et al., 2000). However, the means by which it does so remain poorly defined. Here, we demonstrated that the p21 gene is a transcriptional target of CDX2, and our results thus provided a new mechanism whereby CDX2 functions. Previous studies showed that overexpression of CDX2 in HT-29 colon carcinoma cells resulted in a decrease in cell growth rate by half (Mallo et al., 1998). Moreover, conditional overexpression of Cdx2 in IEC-6, an

undifferentiated intestine cell line, led to arrest of proliferation for several days followed by a period of growth, resulting in multicellular structures containing a well-formed columnar layer of cells with multiple morphological characteristics of intestinal epithelial cells (Suh and Traber, 1996). On the other hand, exit from the cell cycle is a prerequisite for terminal differentiation, and p21 expression is induced during terminal differentiation in vitro and in vivo (reviewed by Gartel and Tyner, 1999), suggesting that p21 plays a critical role in the differentiation process. Thus, it is interesting to speculate that CDX2 functions as a proliferation inhibitor and differentiation promoter through its transactivation of p21 in addition to other previously reported genes, such as GKLF (Dang et al., 2001) and LI-cadherin (Hinoi et al., 2002). Of note, it remains unknown whether or not other family members of cdk inhibitor or other factors related to the cell cycle are involved. Our findings are supported by the following previous observations. (1) The expression patterns of Cdx2 and p21 overlap each other along the vertical intestinal axis in mice (Gartel et al., 1996; Silberg et al., 2000). In the normal proximal intestine, both genes are strongly expressed at the crypt–villus junction into villus, whereas in the normal colon the expression of Cdx2 and p21 is limited to the upper crypt and surface epithelium. (2) The expression of both CDX2 and p21 increased as Caco-2 colon carcinoma cells differentiated, which differentiate spontaneously after reaching confluency (Gartel et al., 1996; Lorentz et al., 1997). (3) During colorectal tumorigenesis, CDX2 and p21 were downregulated and stronger expression of these two genes was demonstrated in portions with higher differentiation in tumors (Ee et al., 1995; Polyak et al., 1996; Mallo et al., 1997; Bukholm and Nesland, 2000). Since mutations in the p21 gene are extremely rare in human tumors (Gartel and Tyner, 1999), decreased expression of CDX2 might contribute to the downregulation of p21 in addition to methylation in its promoter (Allan et al., 2000) or inactivation of its upstream genes like p53 (Gartel and Tyner, 1999). These results strongly suggest that CDX2 and p21 may act in concert in the control of epithelial self-renewal and commitment to differentiation, and in carcinogenesis as well. Expression of p21 is controlled through both p53dependent and p53-independent mechanisms. After DNA damage, p21 transcription is activated by p53 (Brugarolas et al., 1995). However, a variety of agents that promote growth arrest and differentiation also activate p21 transcription through p53-independent pathways by means of different transcriptional factors (Gartel and Tyner, 1999). Interestingly, overexpression of homeobox gene gax or HOXA10 also induced cell proliferation inhibition or differentiation via p53-independent upregulation of p21 (Smith et al., 1997; Bromleigh and Freedman, 2000). In our study, a p53independent mechanism is suggested because overexpression of CDX2 activated transcription of exogenous and endogenous p21 in p53 null or mutated cells. Oncogene


7946

Figure 4 LightCyler real-time PCR analysis of p21 expression in HT-29 cells. Cell transfection, RNA isolation, and cDNA synthesis were performed in the same way as for semiquantitative RT–PCR. LightCycler PCR was performed as described in Materials and methods. Representative amplification of p21 and GAPDH is shown in (a) and (c), respectively, in which the real-time SYBR Green I fluorescence history versus cycle number of various templates is indicated. The calculated results in (a) and (c) are listed in Table 1. The Fit Point Method was performed for determination of concentration and CP in this study. LightCycler melting curve analyses demonstrated single product specific melting temperatures for p21 (b) and GAPDH (d), as indicated by w. LightCyler real-time PCR was performed in duplicate in two independent experiments, and the results are shown in (e). The p21 mRNA level was expressed as the ratio of the concentrations of p21 and GAPDH

Table 1 p21 and GAPDH mRNA expression after CDX2 transfection in HT-29 cellsa p21b Cells CDX2-transfected pcDNA3.1-transfected Parental

GAPDHb

Concentration

Crossing point

Concentration

Crossing point

127.2 48.0 70.8

25.9 27.2 26.7

1143 1269 1121

13.5 13.3 13.5

a

These are the results of the LightCycler reactions shown in Figure 4a and c. Concentrations and crossing points were calculated as described in Materials and methods.

b

Nevertheless, previous studies suggest that other transcription factors may cooperate with CDX2 in transactivation of p21. GATA-4, -5 and -6 zinc-finger and hepatocyte nuclear factor-1a (HNF-1a) homeodomain transcription factors are expressed in the intestinal Oncogene

epithelium and synergistically activate the promoters of intestinal genes, such as LPH or SI (Krasinski et al., 2001; van Wering et al., 2002). Moreover, HNF-1a, GATA-4 and Cdx2 can interact physically and activate SI promoter activity in cotransfection experiments,


7947

where GATA-4 requires the presence of HNF-1a and Cdx2 (Boudreau et al., 2002). These data suggest that zinc-finger/homeodomain interactions are an important pathway for cooperative activation of gene transcription. Finally, GATA-6 can induce G1 cell cycle arrest via upregulation of p21 (Perlman et al., 1998), and Sp1 and Sp3, members of the zinc-finger family, activate p21 transcription (Gartel et al., 2000; Koutsodontis et al., 2001). Unlike CDX2, CDX1 has pro-oncogenic potential. Firstly, its expression is restricted to the proliferative compartment of the intestinal epithelium (Subramanian et al., 1998). Secondly, CDX1 overexpression stimulates the proliferation of IEC6 cells (Soubeyran et al., 1999). Thirdly, Cdx1 is upregulated through oncogenic activation of the Ras and Wnt/b-catenin pathways (Lorentz et al., 1999; Lickert et al., 2000). However, the role of CDX1 in carcinogenesis remains controversial. For example, the expression of CDX1 was found to be downregulated during colorectal carcinogenesis (Mallo et al., 1997). Recently, Cdx1 was found to function as a negative regulator in the transcription of the p21 gene in colon adenocarcinoma SW480 cells (Moucadel et al., 2002). Our data did not show the upregulation of p21 expression by CDX1 either. These data indicate the difference between CDX2 and CDX1 in the transcriptional regulation of the p21 gene. In summary, we have shown that the CDX2 homeodomain protein transactivated the transcription of p21 cdk inhibitor via its binding to the promoter of p21. These data suggest that p21 is probably one of the mediators for CDX2 to act in both normal and pathologic processes. Our results thus may shed new insights into the mechanism underlying the functions of CDX2.

Materials and methods Plasmids The human p21/WAF1/CIP1 promoter luciferase reporter (p21-Luc) construct was described previously (Irwin et al., 2000). This reporter construct contains the p21 promoter from 2699 to þ 45 bp in the pGL2-basic vector that lacks eucaryotic promoter and enhancer sequences. Using restriction enzyme AflII, putative CDX binding sites on the 50 side from 1900 to 658 bp of the p21-Luc construct were excised, resulting in p21-Luc/D1 containing three clustered CDX binding sites located from 471 to 434 bp. Further, p21Luc/D2 without any CDX binding site was created by PstI, which deleted the region from 2504 to 212 bp (Figure 1b). CDX2 and CDX1 expression constructs have been described previously (Bai et al., 2002), which were created by subcloning human CDX2 or CDX1 complementary DNA (cDNA) inserts containing the full coding region into a mammalian expression vector, pcDNA3.1 (Invitrogen, Groningen, Netherlands). Cell culture, transfection and luciferase reporter assay Human osteosarcoma cell lines SAOS-2 and U-2OS, human gastric cancer cell line GT3TKB and human colon adenocarcinoma cell line HT-29 were grown at 371C in a 5% CO2, 95% air atmosphere in Dulbecco’s modified Eagle’s medium

containing 10% fetal bovine serum (GIBCO BRL) and 50 mg/ml kanamycin. The cells were plated at 5 104 cells per well (24-well plate) 1 day before transfection. After incubation for 24 h, the cells (50–80% confluent) were transfected with 350 ng pcDNA3.1CDX2 or pcDNA3.1CDX1, 100 ng p21-Luc reporter plasmid or its deleted constructs and 10 ng pRL-SV40 vector using TransIT-LT1 transfection reagent (Mirus, Madison, WI, USA) according to the protocol of the manufacturer. The pcDNA3.1 empty vector was used as a negative control. The pRL-SV40 vector containing the SV40 early promoter upstream of renilla luciferase was cotransfected as an internal control. Cells were harvested 48 h after transfection, and luciferase activity was measured with a Dual-Luciferase Reporter Assay System (Promega) as described by the manufacturer in Lumicounter 700 (Microtech Niti-On, Chiba, Japan). Each transfection was performed in triplicate and experiments were repeated three times. The results were expressed as fold activation, that is, the ratio of normalized luciferase activity of the CDX2 expression construct to that of the empty vector. Electrophoretic mobility shift assay To prepare nuclear extracts, COS-7 cells were transfected with CDX2 expression vector pcDNA3.1CDX2 and the empty pcDNA3.1 vector using a previously described procedure (Bai et al., 2002). Cells were harvested 72 h after transfection by scraping and then resuspended in 4 ml precooled phosphatebuffered saline (PBS), followed by centrifugation at 500 g and 41C for 5 min. The cells were resuspended in five packed cell volumes of cold low salt buffer (10 mm HEPES, pH 7.9, 10 mm KCl, 1.5 mm MgCl2, 0.5% Nonidet P-40, 1 mm dithiothreitol and 0.5 mm phenylmethylsulfonyl fluoride (PMSF)), and incubated on ice for 10 min. The cells were then homogenized with 10 slow up-and-down strokes in a glass Dounce homogenizer using a type B pestle, followed by centrifugation at 13 000 g and 41C for 5 min. The supernatant was removed and the nuclear pellet was washed once with 80 ml low salt buffer. The nuclear pellet was resuspended in 40 ml high salt buffer (20 mm HEPES, pH 7.9, 1.5 mm MgCl2, 420 mm NaCl, 0.2 mm EDTA, 1 mm dithiothreitol, and 0.5 mm PMSF) and then incubated for 30 min on ice with gentle agitation. After centrifugation (13 000 g for 10 min at 41C), the supernatant was recovered, mixed with an equivalent volume of glycerol buffer (20 mm HEPES, pH 7.9, 100 mm KCl, 20% glycerol, and 0.2 mm EDTA). After the protein concentration had been determined using the Bio-Rad Protein Assay reagent (BioRad, Richmond, CA, USA) according to the directions of the manufacturer, the nuclear extract was stored at 801C. Synthetic wild-type oligonucleotides p21-WT 50 TCGTTTTTATAATTTATGAATTTTTATGTATTAATG-30 (435 bp to 471 bp) and 50 -ACATTAATACATAAAAAT TCATAAATTATAAAAACG-30 (434 to 470 bp) within the p21 promoter, and mutated oligonucleotides p21-Mut 50 -TC GTTGCGTGAAGCGTGGAATTGCGTGGTATTAATG-30 and 50 -ACATTAATACCACGCAATTCCACGCTTCACGCAACG-30 were used for EMSA (Figure 2a). Note that the CDX-core binding sequences TTTAT (in bold) in p21-WT were changed to GCGTG (in bold) in p21-Mut. Annealing was performed as follows. Synthetic complementary oligonucleotides (10 nmol) were mixed in annealing buffer (20 mm TrisHCl, pH 7.5, 10 mm MgCl2, and 50 mm NaCl), the final volume being 200 ml, incubated 5 min at 941C, and then cooled down slowly to room temperature. Annealed oligonucleotide probe (10 pmol) was end-labeled using 10 U of T4 polynucleotide kinase and 5 ml of 7000 Ci/mmol [g-32P] adenosine triphosphate (Amersham Pharmacia Biotech). The labeled probe was Oncogene


7948 purified by size-exclusion chromatography through a 0.5-ml column of Sephadex G-25 (Amersham Pharmacia Biotech). Nuclear extract (5 mg) was incubated in the presence of 10 fmol labeled probe with 1 mg poly dI:dC (Amersham Pharmacia Biotech) with or without 100–500-fold excess unlabeled probe in 20 mm HEPES, pH 7.9, 50 mm KCl, 2 mm MgCl2, 10% glycerol, 300 mg/ml BSA, 0.1 mm EDTA, and 0.5 mm dithiothreitol for 15 min at room temperature. DNA–protein complexes were resolved on a 5.5% TBE (90 mm Tris borate and 2 mm EDTA, pH 8.3) polyacrylamide gel and then exposed to radiograph film for autoradiograph detection. Gel supershifts were carried out with nuclear extracts and the gel shift method described above in the presence of 2.5 ml of the antibody against CDX2, as described previously (Bai et al., 2002). RNA extraction and cDNA synthesis HT-29 cells were seeded at 1.25 106 per 10-mm plate. After incubation for 24 h, the cells were transfected with 10 mg pcDNA3.1CDX2 or the empty pcDNA3.1 vector using TransIT-LT1 transfection reagent (PanVera, WI, USA) according to the protocol of the manufacturer. In addition, cells without transfection were cultured in parallel as a control. Total RNA was extracted 24 h after transfection using an RNeasy Mini Kit (QIAGEN, Hilden, Germany) according to the protocol recommended by the manufacturer. RNA concentration was determined by spectrophotometry, and absence of RNA degradation was confirmed by agarose gel electrophoresis. The isolated RNA (1 mg) was preincubated with 0.4 mg 12–18 mer oligo(dT) at 701C for 10 min, and then with 10 mm dNTP, 0.1 M DTT, and 1 ml Superscript II (RNase H [] reverse transcriptase; Life Technologies Inc., Gaithersburg, MD, USA) at 421C for 1 h (Bai et al., 2000). Semiquantitative PCR analysis The synthesized cDNA was then amplified by the PCR method. The PCR reactions for p21 and CDX2 were performed with both 30 and 35 cycles, in a 25 ml mixture comprising 1 ml reverse transcriptase reaction mixture, 5% dimethyl sulfoxide, 2.5 ml 10 PCR buffer, 4 ml 1.25 mm dNTP (Pharmacia), 10 pmol each oligonucleotide primer pair, and 0.5 U Taq DNA polymerase (Biotech International Ltd., Bentley, Australia). Each PCR cycle consisted of 941C for 1 min, 60–631C for 2 min, and 721C for 1 min, followed by a final extension at 721C for 10 min. All these primers were designed to be located in different exons to identify any genomic DNA contamination. The primers used for CDX2 amplification were the same as previously described (Bai et al., 2000) and those for p21 were as follows, 50 -CAAGCTCTACCTTCCCACGG-30 (sense) and 50 -GCCAGGGTATG TACATGAGG-30 (antisense). The GenBank accession number of the human p21 genomic DNA used in this study is 2276311. As an internal control for RT–PCR analysis, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) tran-

scripts were amplified from the same cDNA samples with the following primer pair, 50 -GACCACAGTCCATGCCATCAC30 (sense) and 50 -GTCCACCACCCTGTTGCTGTA-30 (antisense) for 21 cycles, which were observed to be within the logarithmic phase of amplification. The PCR products (5 ml) were electrophoresed in 2% agarose gels containing 0.5 mg/ml ethidium bromide. The gels were illuminated with UV light and then photographed. LightCycler real-time PCR Based on the findings of semiquantitative PCR, we further performed LightCycler real-time PCR. First, conditions for real-time PCRs were optimized with regard to Taq DNA polymerase, forward and reverse primers, MgCl2 concentrations and various annealing temperatures. RT–PCR amplification products were separated by 2% agarose gel electrophoresis, and LightCycler melting curve analyses were also performed. The optimized conditions were transferred on the following protocol. For LightCycler reaction, a mastermix of the following reaction components was prepared to the indicated final concentrations: 8 ml H2O, 2.4 ml MgCl2 (3 mm), 1 ml forward primer (0.5 mm), 1 ml reverse primer (0.5 mm), 2 ml 10 PCR buffer, 3.2 ml dNTP (0.2 mm), 0.4 ml Taq DNA polymerase (2 U), and 1 ml SYBR Green I (0.5 , TAKARA Biomedicals, Kyoto, Japan). LightCycler glass capillaries were filled with the LightCycler mastermix (19 ml) and 1 ml cDNA was added as a PCR template. The capillaries were closed, centrifuged, and placed into the LightCycler rotor. The following run protocol was used: denaturation program (951C for 1 min), amplification and quantification program repeated 45 times for p21 and 40 times for GAPDH (951C for 0 s, 601C for 5 s for p21 or 671C for 5 s for GAPDH, 721C for 20 s with a single fluorescence measurement), melting curve program (65–951C with a heating rate of 0.11C per second and a continuous fluorescence measurement) and finally a cooling step to 401C. The crossing point (CP) is defined as the point at which the fluorescence rises appreciably above the background fluorescence. In this study, the Fit Point Method was performed for determination of concentration and CP using LightCycler Software version 3.3 (Roche Molecular Biochemicals). The standard curve was constructed with fivefold (for p21) or 10-fold (for GAPDH) serial dilutions of the cDNA from the CDX2-transfected HT29 cells, and the original mRNA concentrations of p21 and GAPDH were arbitrarily designated as 125 and 1000, respectively. Acknowledgements We wish to thank Dr WG Kaelin Jr for providing the plasmid, and Drs H Yamamoto, Y Yanagisawa and Y Akiyama for valuable discussions. This study was supported in part by Grants-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports and Culture of Japan.

References Allan LA, Duhig T, Read M and Fried M. (2000). Mol. Cell. Biol., 20, 1291–1298. Bai Y-Q, Akiyama Y, Nagasaki H, Yagi OK, Kikuchi Y, Saito N, Takeshita K, Iwai T and Yuasa Y. (2000). Mol. Carcinog., 28, 184–188. Bai Y-Q, Yamamoto H, Akiyama Y, Tanaka H, Takizawa T, Koike M, Yagi OK, Saitoh K, Takeshita K, Iwai T and Yuasa Y. (2002). Cancer Lett., 176, 47–55. Oncogene

Boudreau F, Rings EHHM, van Wering HM, Kim RK, Swain GP, Krasinski SD, Moffett J, Grand RJ, Suh ER and Traber PG. (2002). J. Biol. Chem., 277, 31909–31917. Bromleigh VC and Freedman LP. (2000). Genes Dev., 14, 2581–2586. Brugarolas J, Chandrasekaran C, Gordon JI, Beach D, Jacks T and Hannon GJ. (1995). Nature, 377, 552–557.


7949 Bukholm IK and Nesland JM. (2000). Virchows Arch., 436, 224–228. Chawengsaksophak K, James R, Hammond VE, Ko¨ntgen F and Beck F. (1997). Nature, 386, 84–87. Dang DT, Mahatan CS, Dang LH, Agboola IA and Yang VW. (2001). Oncogene, 20, 4884–4890. Di Cunto F, Topley G, Calautti E, Hsiao J, Ong L, Seth PK and Dotto GP. (1998). Science, 280, 1069–1072. Dotto GP. (2000). Biochim. Biophys. Acta, 1471, M43–M56. Ee HC, Erler T, Bhathal PS, Young GP and James R J. (1995). Am. J. Pathol., 147, 586–592. 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. Freund JN, Domon-Dell C, Kedinger M and Duluc I. (1998). Biochem. Cell Biol., 76, 957–969. Gartel AL, Goufman E, Najmabadi F and Tyner AL. (2000). Oncogene, 19, 5182–5188. Gartel AL, Serfas MS, Gartel M, Goufman E, Wu GS, El-Deiry WS and Tyner AL. (1996). Exp. Cell Res., 227, 171–181. Gartel AL and Tyner AL. (1999). Exp. Cell Res., 246, 280–289. Harper JW, Adami GR, Wei N, Keyomarsi K and Elledge SJ. (1993). Cell, 75, 805–816. Hinoi T, Lucas PC, Kuick R, Hanash S, Cho KR and Fearon ER. (2002). Gastroenterology, 123, 1565–1577. Irwin M, Marin MC, Phillips AC, Seelan RS, Smith DI, Liu W, Flores ER, Tsai KY, Jacks T, Vousden KH and Kaelin Jr WG. (2000). Nature, 407, 645–648. James R, Erler T and Kazenwadel J. (1994). J. Biol. Chem., 269, 15229–15237. Koutsodontis G, Tentes I, Papakosta P, Moustakas A and Kardassis D. (2001). J. Biol. Chem., 276, 29116–29125. Krasinski SD, van Wering HM, Tannemaat MR and Grand RJ. (2001). Am. J. Physiol. Gastrointest. Liver Physiol., 281, G69–G84. Li R, Waga S, Hannon GJ, Beach D and Stillman B. (1994). Nature, 371, 534–537. Lickert H, Domon C, Huls G, Wehrle C, Duluc I, Clevers H, Meyer BI, Freund JN and Kemler R. (2000). Development, 127, 3805–3813. Lin J, Reichner C, Wu X and Levine AJ. (1996). Mol. Cell. Biol., 16, 1786–1793. Lorentz O, Cadoret A, Duluc I, Capeau J, Gespach C, Cherqui G and Freund JN. (1999). Oncogene, 18, 87–92.

Lorentz O, Duluc I, Arcangelis AD, Simon-Assmann P, Kedinger M and Freund JN. (1997). J. Cell Biol., 139, 1553–1565. Luo Y, Hurwitz J and Massague J. (1995). Nature, 375, 159–161. Macdonald PM and Struhl G. (1986). Nature, 324, 537–545. Mallo GV, Rechreche H, Frigerio J-M, Rocha D, Zweibaum A, Lacasa M, Jordan BR, Dusetti NJ, Dagorn J-C and Iovanna JL. (1997). Int. J. Cancer, 74, 35–44. Mallo GV, Soubeyran P, Lissitzky JC, Andre F, Farnarier C, Marvaldi J, Dagorn J-C and Iovanna JL. (1998). J. Biol. Chem., 273, 14030–14036. McGinnis W and Krumlauf R. (1992). Cell, 68, 283–302. Moucadel V, Totaro MS, Dell CD, Soubeyran P, Dagorn JC, Freund JN and Iovanna JL. (2002). Biochem. Biophys. Res. Commun., 297, 607–615. Mutoh H, Hakamata Y, Sato K, Eda A, Yanaka I, Honda S, Osawa H, Kaneko Y and Sugano K. (2002). Biochem. Biophys. Res. Commun., 294, 470–479. Perlman H, Suzuki E, Simonson M, Smith RC and Walsh K. (1998). J. Biol. Chem., 273, 13713–13718. Polyak K, Hamilton SR, Vogelstein B and Kinzler KW. (1996). Am. J. Pathol., 149, 381–387. Silberg DG, Sullivan J, Kang E, Swain GP, Moffett J, Sund NJ, Sackett SD and Kaestner KH. (2002). Gastroenterology, 122, 689–696. Silberg DG, Swain GP, Suh ER and Traber PG. (2000). Gastroenterology, 119, 961–971. Smith RC, Branellec D, Gorski DH, Guo K, Perlman H, Dedieu JF, Pastore C, Mahfoudi A, Dene`fle P, Isner JM and Walsh K. (1997). Genes Dev., 11, 1674–1689. Soubeyran P, Andre F, Lissitzky JC, Mallo GV, Moucadel V, Roccabianca M, Rechreche H, Marvaldi J, Dikic I, Dagorn J-C and Iovanna JL. (1999). Gastroenterology, 117, 1326– 1338. Subramanian V, Meyer B and Evans GS. (1998). Differentiation, 64, 11–18. Suh E, Chen L, Taylor J and Traber PG. (1994). Mol. Cell. Biol., 14, 7340–7351. Suh E and Traber PG. (1996). Mol. Cell. Biol., 16, 619–625. van Wering HM, Huibregtse IL, van der Zwan SM, de Bie MS, Dowling LN, Boudreau F, Rings EHHM, Grand RJ and Krasinski SD. (2002). J. Biol. Chem., 277, 27659–27667. Waga S, Hannon GJ, Beach D and Stillman B. (1994). Nature, 369, 574–578. Xiong Y, Hannon G, Zhang H, Casso D, Kobayashi R and Beach D. (1993). Nature, 366, 701–704.

Oncogene