AKAP12/Gravin Is Inactivated by Epigenetic Mechanism in Human Gastric Carcinoma and Shows Growth Suppressor Activity

AKAP12/Gravin is inactivated by epigenetic mechanism in human gastric carcinoma and shows growth suppressor activity

Moon-Chang Choi1, Hyun-Soon Jong*,1, Tai Young Kim1, Sang-Hyun Song1, Dong Soon Lee2, Jung Weon Lee1, Tae-You Kim1,3, Noe Kyeong Kim3 and Yung-Jue Bang*,1,3

1National Research Laboratory for Cancer Epigenetics, Cancer Research Institute, Seoul National University College of Medicine, 28 Yongon-dong, Chongro-gu, Seoul 110-799, Korea; 2Department of Clinical Pathology, Seoul National University College of Medicine, Seoul 110-744, Korea; 3Department of Internal Medicine, Seoul National University College of Medicine, Seoul 110-744, Korea

AKAP12/Gravin, one of the A-kinase anchoring proteins Introduction (AKAPs), functions as a kinase scaffold protein and as a dynamic regulator of the b2-adrenergic receptor complex. AKAP12/Gravin, one of the A-kinase anchoring However, the biological role of AKAP12 in cancer proteins (Dell’Acqua and Scott, 1997; Nauert et al., development is not well understood. The AKAP12 gene 1997; Diviani and Scott, 2001; Feliciello et al., 2001), encodes two major isoforms of 305 and 287 kDa was first isolated as a protein recognized by serum (designated AKAP12A and AKAP12B, respectively, in from myasthenia gravis patients (Gordon et al., 1992). this report). We found that these two isoforms are AKAP12 organizes the complex of PKA and PKC independently expressed and that they are probably under (Nauert et al., 1997), and is an important regulator the control of two different promoters. Moreover, both of the b2-adrenergic receptor complex (Shih et al., isoforms were absent from the majority of human gastric 1999; Lin et al., 2000a). AKAP12 expression can cancer cells. The results from methylation-specific PCR be induced by several drugs like phorbol ester (MSP)and bisulfite sequencing revealed that the 5 0 CpG (Gordon et al., 1992; Nauert et al., 1997) and lysopho- islands of both AKAP12A and AKAP12B are frequently sphatidylcholine (Sato et al., 1998), which suggests the hypermethylated in gastric cancer cells. Treatment with participation of AKAP12 in diverse signal transduction DNA methyltransferase inhibitor and/or histone deacety- cascades. lase inhibitor efficiently restored the expression of AKAP12 has been mapped to chromosome 6q24– AKAP12 isoforms, confirming that DNA methylation 25.2, which frequently contains deletions in tumors, is directly involved in the transcriptional silencing of including melanoma (Millikin et al., 1991) and breast AKAP12 in gastric cancer cells. Hypermethylation of cancer (Tibiletti et al., 2000). Interestingly, the down- AKAP12A CpG island was also detected in 56% (10 regulation of AKAP12 expression has been reported in of 18)of primary gastric tumors. The restoration of human prostate cancers in vivo (Xia et al., 2001), AKAP12A in AKAP12-nonexpressing cells reduced suggesting that the inactivation of AKAP12 expression colony formation and induced apoptotic cell death. In may be linked to oncogenesis. Thus, the molecular conclusion, our results suggest that AKAP12A may mechanism of the tumor-specific inactivation of function as an important negative regulator of the survival AKAP12 expression should be defined. pathway in human gastric cancer. The patterns of DNA methylation and chromatin Oncogene (2004) 23, 7095–7103. doi:10.1038/sj.onc.1207932 structure are profoundly altered in neoplasia. Aberrant Published online 19 July 2004 methylation of the CpG islands located in the promoter regions of tumor suppressor genes (TSGs) is now firmly Keywords: AKAP12; alternative promoter; DNA established as a major epigenetic mechanism of gene methylation; histone deacetylation; apoptosis inactivation in tumorigenesis (Jones and Laird, 1999; Esteller, 2002; Jones and Baylin, 2002). Genes silenced by DNA methylation can be restored by treatment with 5-aza-20-deoxycytidine (5-Aza-dC), an inhibitor of DNA methyltransferase (Jones and Taylor, 1980). The src-suppressed C-kinase substrate (SSeCKS), the rodent orthologue of human AKAP12, was originally *Correspondence: Y-J Bang, Department of Internal Medicine, Seoul identified as a gene downregulated in response to Src National University College of Medicine, 28 Yongon-dong, Chongro-gu, and Ras activation (Lin et al., 1995). The overexpression Seoul 110-744, Korea; of SSeCKS suppressed Src-induced oncogenesis by E-mail: [email protected]; H-S Jong, Cancer Research Institute, inhibiting the cellular proliferation, and by reducing Seoul National University College of Medicine, 28 Yongon-dong, Chongro-gu, Seoul 110-799, Korea; E-mail: [email protected] anchorage-independent growth in soft agar and inva- Received 16 November 2003; revised 1 April 2004; accepted 6 May 2004; siveness in Matrigel (Lin and Gelman, 1997). A recent published online 19 July 2004 report showed that SSeCKS regulates blood–brain Epigenetic inactivation of AKAP12 M-C Choi et al 7096 barrier differentiation by inhibiting angiogenesis by and SNU-5 cells expressed three isoforms (305, 287 and reducing VEGF expression, and that the constitutive 250kDa). Two isoforms were detected in SNU-484 (305 expression of SSeCKS may be important for brain and 250kDa) and SNU-638 (287 and 250kDa). homeostasis (Lee et al., 2003). Figure 2a is a schematic diagram of the AKAP12 gene In this study, we identified AKAP12 as a novel structure based on Genbank database (Accession no. epigenetic target gene in gastric cancer, and compara- NT_023451). AKAP12A is composed of four exons (1a, tively evaluated the silencing of the two transcripts of 2, 3 and 4), whereas AKAP12B contains three (1b, 3 and AKAP12. We also examined the tumor suppressor 4). To distinguish between the transcripts of AKAP12A activity of AKAP12A in gastric cancer cells. and AKAP12B, we performed 50-RACE using primers derived from exon 3, which is common to the two transcripts. All RACE products using the mRNA from Results SNU-638 encoded exon 1b, but not exon 1a (data not shown), indicating that SNU-638 cells express only Loss of AKAP12 expression in gastric cancer cell line AKAP12B. Notably, the upstream of exon 1b was not detected by 50-RACE, suggesting the existence of an Northern and Western analyses were performed to internal promoter located at the 50 boundary of the examine the expression of AKAP12 in human gastric coding exon 1b. To characterize the expression profiles cancer cells. HepG2 and three gastric cancer cell lines (SNU-5, -484 and -638) were found to express AKAP12 mRNA and protein, whereas the other cells did not (Figure 1). To determine whether AKAP12 inactivation is due to chromosomal deletion, the genomic DNA at the AKAP12 locus was examined by Southern blot and FISH analysis in gastric cancer cells; however, no genetic defects were detected (data not shown). These cell lines provided a panel of AKAP12-expressing and -nonexpressing cells, and were used to further investigate the mechanism underlying the loss of AKAP12 expression.

Identiﬁcation of distinct transcripts of AKAP12 The two transcripts of AKAP12 encode three isoforms of 305, 287 and 250 kDa (Xia et al., 2001; Gelman, 2002). One transcript variant 1 (designated AKAP12A in this study) encodes the large isoform (305 kDa), and another transcript variant 2 (designated AKAP12B in this study) has the small isoform (287 kDa), whereas the Mr 250kDa isoform is a proteolytic cleavage product of these two major isoforms. As shown in Figure 1, HepG2

Figure 2 Identiﬁcation of differentially expressed AKAP12 Figure 1 Expression of AKAP12A mRNA and protein in gastric transcripts. (a) Schematic diagram of the intron/exon structure of cancer cells. Northern blot of total RNA was hybridized with the AKAP12 and splicing patterns. This representation is based on the probe for exon 3 (common to the two AKAP12 transcripts) and human genome sequence contig NT_023451. (b) Northern blots of subsequently with the b-actin probe as a loading control. The total RNA from HEK293, HepG2, SNU-601 (À), SNU-601 expression of AKAP12 protein was determined by Western blot exposed to 10 mM 5-Aza-dC ( þ ), and SNU-638 cells were analysis. A Western blot with anti-Tubulin antibody was used to hybridized with probes for exon 1a´ and exon 2 (P1 and P2), exon control for protein content 1b (P3), exon 3 (P4), and b-actin

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7097 of the AKAP12 transcripts by Northern analysis, we complete or partial methylation in eight of 11 gastric designed four kinds of probes specific for each cancer cell lines. In contrast, three cell lines (SNU-5, transcript. Probes, P1 and P2, derived from exon 1a -484 and -668) were unmethylated at the AKAP12A and exon 2 were specific for the AKAP12A transcript. CpG island. In the case of AKAP12B, four cell lines P3 derived from exon 1b was specific for the AKAP12B (SNU-5, -16, -638 and -668) were unmethylated at the transcript. P4 derived from exon 3 was common to the AKAP12B CpG island, whereas the remaining seven cell both transcripts. Figure 2b shows that HepG2 expressed lines were either completely or partially methylated. A both AKAP12 transcripts, whereas HEK293 and close correlation was observed between the expression SNU-638 expressed AKAP12A and AKAP12B, respec- and methylation status of each CpG island, except in tively. These results imply that both AKAP12 isoforms SNU-16 and -668 cells (Figures 1 and 3c). were missing in the majority of gastric cancer cells. Restoration of AKAP12 by 5-Aza-dC treatment Aberrant promoter methylation of AKAP12A and AKAP12B in gastric cancer cells We examined whether demethylation could restore AKAP12 expression in gastric cancer cells. In During the genetic mapping of AKAP12, we found that many cases, treatment with the DNA methyltransferase AKAP12A and AKAP12B have CpG islands in their inhibitor 5-Aza-dC in AKAP12-nonexpressing cells promoter regions. The AKAP12A and AKAP12B was found to restore AKAP12 expression (Figure 4a). promoters contained CpG islands of ca. 1.9 kb and Using transcript-specific primers for reverse transcrip- 400 bp, respectively (Figure 3a and b). The CpG island tion (RT)ÀPCR analysis (Figure 4b), we characterized in AKAP12A was found to have a GC content of 65.5% the expression pattern of each isoform induced by and a CpG observed/expected ratio of 0.70, whereas the 5-Aza-dC treatment. Western blot using an antibody CpG island in AKAP12B had a GC content of 65.3% against AKAP12 showed that the expression levels and a CpG : GpC ratio of 0.70, thus satisfying the of the protein correlated well with those of the mRNA criteria for a CpG island. These findings led us to (Figure 4c). To elucidate the kinetic mechanism of investigate whether the hypermethylation of the 5-Aza-dC-mediated AKAP12 induction in detail, AKAP12 CpG islands could regulate AKAP12 expres- SNU-1 was exposed to increasing concentrations of sion in gastric cancer cell lines. Methylation-specific 5-Aza-dC for 4 days (Figure 4d), and the re-expression PCR (MSP) assay was used to assess the methylation of each AKAP12 isoform was observed at both the status of several CpG dinucleotides within the 50 CpG mRNA and protein levels in a dose-dependent manner, islands of both isoforms; examples are illustrated in concurrent with the demethylation of the AKAP12 CpG Figure 3c. MSP analysis for AKAP12A revealed either islands. The methylation patterns of both 17 CpGs of

Figure 3 Methylation analysis of AKAP12A and AKAP12B in gastric cancer cells. (a) CpG island sequence (1875 bp) of AKAP12A. (b) CpG island sequence (444 bp) of AKAP12B. The identified 50 end of each transcript is indicated by a bent arrow. Sequences used for designing primers for MSP (MSP-1 or MSP-2) and for bisulfite sequencing (Seq-1 or Seq-2) are indicated. (c) MSP analysis of AKAP12A (top) and AKAP12B (bottom). Visible bands in ‘U’ lanes are unmethylated DNA products with unmethylation-specific primers and those in ‘M’ lanes are methylated DNA products with methylation-specific primers

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7098 AKAP12A and 23 CpGs of AKAP12B were Aberrant histone deacetylation of AKAP12 in SNU-16 also examined by bisulﬁte sequencing (Figure 4e). cells All CpG sites within both CpG islands of AKAP12 in Interestingly, SNU-16 and -668 cells, which did not SNU-1 were densely methylated. These results obtained from 5-Aza-dC-treated SNU-1 cells revealed that the show AKAP12 expression, were either unmethylated or partially methylated at both AKAP12 CpG islands demethylation of their promoter regions is required (Figures 1 and 3c). In addition, Northern blot analysis to restore AKAP12 expression. Moreover, the methylation status of each CpG island correlated strongly showed that the expression of AKAP12 was not recovered by 5-Aza-dC treatment in these cells with isoform-speciﬁc expression in gastric cancer cells (Figure 4a). A growing body of data indicates the examined. importance of histone deacetylation and DNA methylation, and of corresponding chromatin structural alterations in the process of gene silencing (Cameron et al., 1999; Burgers et al., 2002). To investigate whether aberrant histone deacetylation is involved in AKAP12 silencing in SNU-16 cells, we used 5-Aza-dC and the histone deacetylase inhibitor trichostatin A (TSA), alone or a combination. TSA treatment strongly induced AKAP12B expression in a time-dependent manner, but was unable to restore AKAP12A expression (Figure 5a). A combination of these two inhibitors induced detect- able levels of AKAP12A expression in SNU-16 cells with partial methylation of the AKAP12A CpG islands (Figure 5b).

Aberrant promoter methylation of AKAP12A in primary gastric cancers The frequencies of AKAP12A and AKAP12B methylation were then characterized in a small set of gastric cancer tissues and in matched normal tissues (Figure 6a). The CpG islands of AKAP12A were methylated in 10of 18 (56%) tumor cases, and those of AKAP12B were methylated in two of 18 (11%). The unmethylated band was always present in tumor samples, most of which

Figure 4 Restoration of AKAP12 expression after treating cell lines with 5-Aza-dC. HEK293 (AKAP12A), SNU-638 (AKAP12B), and HepG2 (Both isoforms) served as positive controls. The b-actin transcript and a nonspecific signal were used as loading controls for mRNA and protein, respectively. (a) Northern blot of gastric cancer cells treated with 5-Aza-dC ( þ )or with vehicle alone (À) was hybridized with a probe specific for exon 3, as shown in Figure 2a. (b)RTÀPCR analysis of AKAP12A and AKAP12B. (c) Western blotting with anti-AKAP12 antibody. Note that the anti-AKAP12 antibody recognizes both isoforms. (d) RTÀPCR, Western blot and MSP analysis of SNU-1 cells treated with increasing concentrations of 5-Aza-dC. (e) Comparison of methylation status and isoform-specific expression. Genomic DNA treated with sodium-bisulfite amplified using the methylation- independent primer set, seq-1 and seq-2, as shown in Figure 3. Figure 5 Effect of TSA on AKAP12 expression in SNU-16 cells. Each row of squares represents a single clone, while each square (a) Western blot analysis was performed after treating SNU-16 cells depicts a single CpG site. Filled and open squares represent with 330nM TSA for the indicated times. A nonspecific signal was methylation and unmethylation, respectively. Gene names are used as a protein loading control. (b) Re-expression of AKAP12A indicated at the bottom of each panel after cotreatment with 5-Aza-dC and TSA

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7099

Figure 6 The methylation status of AKAP12 and the expression of AKAP12 in gastric carcinoma. (a) MSP analysis of 18 tumors and their matched normal tissues. SNU-484 and -638 cells served as controls for the methylated bands of AKAP12B and AKAP12A, respectively. T represents a primary gastric cancer sample and N represents its nontumorous counterpart. (b) Bisulﬁte sequencing using the indicated primary specimens. The CpG sites are identical to those shown in Figure 4e. (c) Western blot analysis of eight matched pairs of normal (N) and tumor (T) tissues. SNU-484 and HepG2 served as positive controls for AKAP12A and for both isoforms, respectively

consisted of mixtures of tumor cells and nonmalignant pAKAP12A significantly reduced (Po0.001) the cells. Bisulfite sequencing was used to confirm the tumor number of G418-resistant colonies vs pDAKAP12A specificity of AKAP12 methylation (Figure 6b). All (Figure 7a). To obtain more precise information about CpG sites in the AKAP12A CpG island were heavily the function of AKAP12A, we generated several methylated in tumor sample T2, whereas matched adenoviral vector constructs (Ad), and analysed cell normal tissue was unmethylated, suggesting that this proliferation using an MTT assay 3 days after infection. hypermethylation is a tumor-specific event. Thus, MSP Compared with Ad-empty and -lacZ control virus, analysis is sensitive enough to detect very low levels of Ad-AKAP12A infection strongly inhibited the cellular methylation in the CpG island of AKAP12A. Methyla- growths of AGS (Figure 7b) and MKN28 (data not tion in normal tissues, which is not the result of shown). Ad-AKAP12B also exhibited growth suppres- imprinting, is frequently ‘age-related’ (Issa et al., 2001) sive activity over the control construct (Ad-empty and or represents premalignant changes (Kang et al., 2001). Ad-LacZ). FACS analysis was used to further examine Next, we examined whether promoter hypermethylation the molecular mechanism of growth suppression in- is associated with the downregulation of AKAP12 duced by AKAP12A expression. Cell cycle analysis expression. The downregulation of AKAP12A expres- showed a higher sub-G1 population in Ad-AKAP12A- sion was observed in five of eight gastric tumors by infected AGS than in control virus-infected AGS Western blot analysis (Figure 6c), and three of these five (Figure 7c, top). TUNEL assay and Annexin V staining cancer tissues were methylated at the AKAP12A CpG were used to determine whether the re-expression of island (Figure 6a). AKAP12A was expressed in the AKAP12A induces apoptosis. The transduction of majority of normal gastric tissues, but its expression was Ad-AKAP12A caused DNA fragmentation (Figure 7c, remarkably downregulated in the majority of gastric middle) and a loss of membrane integrity (Figure 7c, carcinomas examined. bottom). Furthermore, the exogenous expression of AKAP12A resulted in the activation of caspase-3 and in Growth suppression by AKAP12A restoration the cleavage of PARP (Figure 7d). These data indicate that the restoration of AKAP12A expression suppresses Based on the tumor-specific loss of AKAP12A expres- cell growth by inducing apoptosis in gastric cancer cells. sion in gastric cancers, we examined whether AKAP12A acts as a tumor suppressor in human gastric cancer cells. The effect of AKAP12A on cancer cell growth was investigated by using a colony formation assay. Plasmid Discussion containing wild-type AKAP12A (pAKAP12A), the deletion construct of pAKAP12A (pDAKAP12A) or Genetic and epigenetic changes in TSGs are inevitable EGFP (pEGFP) were transiently transfected into AGS events in the development of human cancer. cells not expressing AKAP12. The transfection of Since tumor-acquired promoter hypermethylation and

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7100 the alteration of chromatin structure are comparable aberrant DNA methylation. In fact, there are many with genetic mutations or deletions of TSG in cancer, it examples of CpG methylation-mediated transcriptional is important to find target genes that are inactivated by silencing of TSGs, such as p16ink4a, APC, hMLH1, Rb, VHL, MGMT and BRCA1 (Esteller, 2002). In the present study, we found that AKAP12 is a target gene for aberrant DNA methylation in gastric cancer cells. Hypermethylation of AKAP12A was detected in 56% of primary tumors, indicating that the inactivation of AKAP12A is relatively common in gastric tumorigenesis. In our initial study, we observed the loss of AKAP12 expression without genetic deletion in the majority of gastric cancer cells. We then undertook the epigenetic analysis of AKAP12. Two approaches, MSP and bisulfite-sequencing analysis, were used to study the methylation status of the AKAP12 CpG islands. The results showed that the 50 CpG islands of both AKAP12A and AKAP12B were frequently hypermethylated in gastric cancer cell lines. Moreover, pharmaco- logic treatment with DNA methyltransferase inhibitor and/or HDAC inhibitor restored AKAP12 expression in AKAP12-nonexpressing cells, confirming that DNA methylation is directly involved in the transcriptional silencing of AKAP12 in gastric cancer cells. The CpG island of AKAP12A was partially methylated in SNU-620and -719 cells, which did not express this gene. As MSP assays are usually carried out in a nonquantitative manner, samples showing only low levels of DNA methylation may be scored as positive for methylation. Nevertheless, methylation does appear to play a role in AKAP12A repression because demethylation readily reactivates the gene (Figure 4). In the case of SNU-484 with endogenous expression of AKAP12A, 5-Aza-dC treatment did not affect the level of AKAP12A expression, supporting that DNA methylation is directly involved in the silencing of AKAP12A in SNU-620and -719 cells. These cells seemed to have a critical threshold level of the methylation necessary to

Figure 7 Effect of AKAP12A expression on cancer cell growth. (a) Colony formation assay. AGS cells were transfected with an AKAP12A expression vector (pAKAP12A), the deletion construct of pAKAP12A (pDAKAP12A) or pEGFP. A total of 1 Â 105 or 5 Â 104 cells were replated per well and cultured for 11 days before staining with crystal violet. Values are means 7s.d. of two separate experiments, each calculated from quadruplicate plates. (b) Growth inhibition of AGS cells by AKAP12A expression using an adenoviral vector system. Cell growth inhibition was measured quantitatively by MTT assay 3 days after infection. The viability of mock-treated cells was set at 100%. The assay was performed in six replicates and repeated three times; bars, 7s.d. (c) Cell cycle analysis and apoptosis detection in AKAP12A-restored AGS cells. At 48 h after infection at 20MOI with the adenoviral vectors (1: PBS, 2: Ad-empty, 3: Ad-lacZ, 4: Ad-AKAP12A or 5: Ad- AKAP12B), cells were analysed by FACS after PI staining to characterize the cell cycle distribution (top). Apoptosis was detected by TUNEL assay (middle) and by Annexin V staining (bottom). The inset legend shows the percentage of cells under- going apoptosis (Annexin V not shown). Representative data is shown. Similar results were obtained from three (PI staining) and two (TUNEL assay and Annexin V staining) independent experiments. (d) Western blot analysis of cell lysates 24À48 h after adenoviral infection using AKAP12, PARP, caspase-3 and a- Tubulin antibodies

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7101 affect gene expression. Similar cases were reported for knowledge, that the AKAP family is involved in hypermethylated genes, like TMS1 (Conway et al., epigenetic alterations. The high incidence of aberrant 2000), CDH13 (Toyooka et al., 2001) and NES1 (Li AKAP12A methylation and its growth suppression et al., 2001). To clarify this, a further examination using activity identified here strongly suggest that AKA12A the COBRA assay (Xiong and Laird, 1997) may be serves as an important negative regulator of the survival necessary. The inactivation of AKAP12 by other pathway in human gastric cancer. Thus, this gene mechanisms, such as point mutations, will also need to presents a possible target for therapeutic agents, and be examined to further clarify the role of this gene in may be a useful biomarker for gastric cancer and other cancer development. cancers. Alternative promoters represent other means of generating and regulating diverse protein isoforms. It is not clear whether the two transcripts of AKAP12 arise Materials and methods by different translation initiations from the same mRNA or via transcription from different promoters Cell lines and tissues 0 within the same gene. In our study, 5 -RACE and SNU-1, -5, -16, -484, -601, -620, -638, -668 and -719, and Northern blot analysis results suggested that the MKN28, AGS and HepG2 (Korean Cell Line Bank, Seoul, expression of each isoform is likely to be regulated by Korea) were cultured in RPMI 1640supplemented with 10% distinct promoters. In addition, the expression of each fetal bovine serum (FBS) and gentamicin (10 mg/ml). HEK293 isotype was strictly regulated by CpG island hyper- (American Type Culture Collection) was maintained in methylation in each promoter region, supporting the DMEM supplemented with 10% FBS. Tumors and corre- existence of two different promoters. There are similar sponding normal tissues were obtained from 18 patients with examples of alternative promoters being the targets of primary gastric carcinoma, during surgery at the Seoul DNA hypermethylation. RASSF1A and RASSF1C are National University Hospital, Seoul, Korea. produced by transcription from different promoters, but 0 only RASSF1A was found to be silenced by aberrant Northern blot, 5 -RACE, RTÀPCR and Western blot analysis methylation in cancers (Dammann et al., 2000, 2001). In Total and poly(A) RNA were isolated as described previously addition, the RIZ1 TSG, but not RIZ2 regulated by the (Song et al., 2001). Northern blotting was performed internal promoter of the same gene, was inactivated by as described previously (Jong et al., 2002). 50-RACE was DNA hypermethylation in human cancers (Liu et al., carried out on poly(A) RNA using a Marathon cDNA 1997; Du et al., 2001). Amplification Kit (Clontech). RACE products were cloned in Finally, we attempted to introduce AKAP12A ex- pCR2.1 TOPO vector (Invitrogen), and sequenced. The gene- specific primers (GSP) used were as follows: R1, pression in AKAP12-nonexpressing cancer cells to 50ÀTTAAATCCAATATCATTAGCCTGGGACTCAÀ30 (for determine whether the transformed phenotype could RT) and R2, 50ÀGATGTCGTGAACAACCGCTGACTTA be reduced. In the case of rodent SSeCKS, its over- GTAÀ30 (for PCR). RTÀPCR was performed as described expression in NIH3T3 fibroblasts caused G1 arrest (Lin previously (Song et al., 2000). The primer sequences used to et al., 2000b). However, the growth suppressive activity prepare the AKAP12 cDNA probes and for RTÀPCR are of AKAP12 has not been studied previously. In the summarized in Table 1. Western blotting was performed as present study, the exogenous expression of AKAP12A described previously (Park et al., 2003). Polyclonal anti- in AGS cells reduced colony formation, and this AKAP12 antibody was a gift from Dr JD Scott (Vollum inhibitory effect was mediated by apoptotic cell death. Institute, Portland, OR, USA). Antibodies against caspase-3, As shown in Figure 7b (MTT assay) and Figure 7c PARP and a-Tubulin were used, as described previously (Kim et al., 2001). Detection was performed using an ECL system (FACS analysis), the results based on adenoviral (Amersham). expression demonstrated that both AKAP12A and AKAP12B have tumor-suppressive activity. Several Drug treatment and methylation analysis pieces of evidence indicate that the function of AKAP12 may be achieved through direct or indirect interaction Cells were incubated in culture medium with 5-Aza-dC with multiple effector proteins. These include several (Sigma) at a concentration of 10 mM for 4 days, with medium signaling molecules, which participate in cell prolifera- changes on days 1 and 3, whereas SNU-620cells were treated with 1 mM of 5-Aza-dC due to cell death at 10 mM. SNU-16 cells tion and cytoskeletal organization, and which include were treated with 330 mM TSA (Wako) for various periods, as protein kinase C (PKC), protein kinase A (PKA), cyclin described in the legend to Figure 5a. For the synergic study, D1 and calmodulin (Lin et al., 2000b; Diviani and Scott, different concentrations of 5-Aza-dC were used in culture for 4 2001; Lin and Gelman, 2002). AKAP12 was reported to days, and/or the indicated concentrations of TSA detailed in inhibit PKC kinase activity in vitro (Nauert et al., 1997). Figure 5b were added for the last 48 h. MSP and bisulfite- Thus, there is a possibility that the loss of AKAP12 sequencing analysis were performed as described previously expression results in the dysregulation of PKC activity. (Kim et al., 2003). The primer sequences used for these Since the specific role of AKAP12 in signaling pathway analyses are summarized in Table 1. remains unclear, this question should be further explored. Colony formation and growth inhibition assay This is the first report to link aberrant DNA pAKAP12A (a kind gift from Dr JD Scott) was transfected methylation to the absence of AKAP12 expression. into AGS cells seeded on six-well plates using LipofectAMINE More importantly, it is the first to suggest, to our 2000 (Invitrogen) according to the manufacturer’s protocol.

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7102 Table 1 Summary of primer sequence for Northern blot, RT–PCR, busulﬁte sequencing and MSP analysis

Forward primer (50-30) Reverse primer (50-30) Product Accession no. size (bp)

Northern probe P1 TCTTTTAAGGAGTTTGCCGC ATCTTGCTCAGCTACGCCAT 404 NM_005100 P2 GTCTCCTTCATTCGCAGGCT ATCTTGCTCAGCTACGCCAT 378 NM_005100 P3 AGGGCACCTCCGGTTCTC GTGATGGTCCCCAGCATATT 180NM_144497 P4 AGGAGCTCAGCGAGAGTCAG ATGTCAGGTACGCCACCTTC 451 NM_005100

RT–PCR AKAP12A GTCTCCTTCATTCGCAGGCT CATGGCTCCTCCGCACTTCTC 165 NM_005100 AKAP12B AGGGCACCTCCGGTTCTC GGTTCGCTTTCTTTGGATGC 564 NM_144497

Bisulﬁte sequencing Seq-1 TGTTTTTTGAGGTTTTGGGT AACCACCTCTTAACCTCC 228 AL356535 Seq-2 GAGGGTTTTAAGGTGAAGTA CCTAATCTCCTACCTACCAA 302 AL033392

MSP MSP-1 M:GGGTCGTTTTCGTAGTTTTAGTCG CAAAAACGCTACGACGCGCC 102 AL356535 U:TTGGGTTGTTTTTGTAGTTTTAGTTG AACCAAAAACACTACAACACACC 107 MSP-2 M:AAGTTTGCGTTTTCGAAGTTTTGGA TAATTTCGACGAAATCAAAACGAAC 223 AL033392 U:GTTTGTGTTTTTGAAGTTTTGGA TCAACAAAATCAAAACAAACTAAAC 216

The deletion construct of pAKAP12A (pDAKAP12A) was transgene expression control. Adenoviral vector lacking an generated by digesting ca. a 4 kb length containing N-terminal insert (Ad-empty) was used as a negative control. The region from full cDNA (ca. 5.4 kb) causing the removal of transduction efficiency of Ad-lacZ was determined by X-gal AKAP and PKC binding domains. The resulting peptide staining (Invitrogen). encompassed residues 1239-1782 of AKAP12, a fragment containing C-terminal region. pDAKAP12A and pEGFP were used as control plasmids. Cells were trypsinized and plated on Cell cycle analysis and apoptosis assay six-well plates at 24 h post-transfection. They were then AGS cells were seeded in six-well plates, and infected selected with G418 (500 mg/ml), and colonies were counted separately with several adenoviral vector constructs at 20 11 days after transfection. For the growth inhibition study, MOI (multiplicity of infection) in 10% FBS/RPMI, and then AGS cells were plated at a density of 2000 cells/well in 96-well incubated at 371 for 2 h with brief agitation every 15 min. The plates, and cell growth inhibition by infection with adenoviral culture medium was then replaced with normal culture vectors was measured quantitatively by MTT (3-(4,5-di- medium, and the infected cells were returned to a 371 methylthiazol-2-yl)2,5-diphenyl tetrazolium bromide, Sigma) incubator. FACS analysis using propidium iodide (PI) assay, as described previously (Kang et al., 1999). staining, TUNEL reaction with FITC-labeled dUTP and Annexin V-FITC staining combined with PI staining were Recombinant adenoviral vectors performed 48 h after adenoviral infection, as described previously (Kim et al., 2001). AKAP12A cDNA was cloned into pShuttle vector (Clontech) for adenoviral vector construction. To construct pShuttle- AKAP12B, the AKAP12A-specific region on pShuttle-AKA- Acknowledgements P12A was replaced with an AKAP12B-specific sequence. The We are thankful to Dr JD Scott for his kind gifts of the expression cassette in each pShuttle vector containing AKA- AKAP12/Gravin construct and antibody, and to all members P12A, AKAP12B or b-galactosidase cDNA was ligated to of our laboratory, particularly to Jin-Ah Park, for technical Adeno-X viral vector (Clontech). These adenoviral vectors assistance and critical discussions. We also thank Jae-Jung Lee were amplified in HEK293 cells, and then CsCl-purified virus for helpful advice with the preparation of the adenoviral was dialysed against PBS with 10% glycerol. The virus titer constructs. This work was supported in part by grants from the was determined using a standard plaque assay, and the titer Ministry of Science & Technology of Korea through the obtained was confirmed using an Adeno-X Rapid titer kit National Research Laboratory Program for Cancer Epige- (Clontech). Ad-lacZ was used to monitor the efficiency of netics and by 2002 BK21 Project for Medicine, Dentistry, and transduction by the viral vectors, and as a nonspecific Pharmacy.

References

Burgers WA, Fuks F and Kouzarides T. (2002). Trends Genet., Diviani D and Scott JD. (2001). J. Cell Sci., 114, 1431–1437. 18, 275–277. Du Y, Carling T, Fang W, Piao Z, Sheu JC and Huang S. Cameron EE, Bachman KE, Myohanen S, Herman JG and (2001). Cancer Res., 61, 8094–8099. Baylin SB. (1999). Nat. Genet., 21, 103–107. Esteller M. (2002). Oncogene, 21, 5427–5440. Conway KE, McConnell BB, Bowring CE, Donald CD, Warren Feliciello A, Gottesman ME and Avvedimento EV. (2001). ST and Vertino PM. (2000). Cancer Res., 60, 6236–6242. J. Mol. Biol., 308, 99–114. Dammann R, Li C, Yoon JH, Chin PL, Bates S and Pfeifer Gelman IH. (2002). Front. Biosci., 7, d1782–1797. GP. (2000). Nat. Genet., 25, 315–319. Gordon T, Grove B, Loftus JC, O’Toole T, McMillan R, Dammann R, Takahashi T and Pfeifer GP. (2001). Oncogene, Lindstrom J and Ginsberg MH. (1992). J. Clin. Invest., 90, 20, 3563–3567. 992–999. Dell’Acqua ML and Scott JD. (1997). J. Biol. Chem., 272, Issa JP, Ahuja N, Toyota M, Bronner MP and Brentnall TA. 12881–12884. (2001). Cancer Res., 61, 3573–3577.

Oncogene Epigenetic inactivation of AKAP12 M-C Choi et al 7103 Jones PA and Baylin SB. (2002). Nat. Rev. Genet., 3, 415–428. Liu L, Shao G, Steele-Perkins G and Huang S. (1997). J. Biol. Jones PA and Laird PW. (1999). Nat. Genet., 21, 163–167. Chem., 272, 2984–2991. Jones PA and Taylor SM. (1980). Cell, 20, 85–93. Millikin D, Meese E, Vogelstein B, Witkowski C and Trent J. Jong HS, Lee HS, Kim TY, Im YH, Park JW, Kim NK and (1991). Cancer Res., 51, 5449–5453. Bang YJ. (2002). Biochem. Biophys. Res. Commun., 292, Nauert JB, Klauck TM, Langeberg LK and Scott JD. (1997). 383–389. Curr. Biol., 7, 52–62. Kang GH, Shim YH, Jung HY, Kim WH, Ro JY and Rhyu Park JH, Kim TY, Jong HS, Kim TY, Chun YS, Park JW, Lee MG. (2001). Cancer Res., 61, 2847–2851. CT, Jung HC, Kim NK and Bang YJ. (2003). Clin. Cancer Kang SH, Bang YJ, Jong HS, Seo JY, Kim NK and Kim SJ. res., 9, 433–440. (1999). Br. J. Cancer, 80, 1144–1149. Sato N, Kokame K, Shimokado K, Kato H and Miyata T. Kim SG, Kim SN, Jong HS, Kim NK, Hong SH, Kim SJ and (1998). J. Biochem., 123, 1119–1126. Bang YJ. (2001). Oncogene, 20, 1254–1265. Shih M, Lin F, Scott JD, Wang HY and Malbon CC. (1999). Kim TY, Jong HS, Song SH, Dimtchev A, Jeong SJ, Lee JW, J. Biol. Chem., 274, 1588–1595. Kim TY, Kim NK, Jung M and Bang YJ. (2003). Oncogene, Song SH, Jong HS, Choi HH, Inoue H, Tanabe T, Kim NK 22, 3943–3951. and Bang YJ. (2001). Cancer Res., 61, 4628–4635. Lee SW, Kim WJ, Choi YK, Song HS, Son MJ, Gelman IH, Song SH, Jong HS, Choi HH, Kang SH, Ryu MH, Kim YJ and Kim KW. (2003). Nat. Med., 9, 900–906. Kim NK, Kim WH and Bang YJ. (2000). Int. J. Cancer, Li B, Goyal J, Dhar S, Dimri G, Evron E, Sukumar S, Wazer 87, 236–240. DE and Band V. (2001). Cancer Res., 61, 8014–8021. Tibiletti MG, Sessa F, Bernasconi B, Cerutti R, Broggi B, Lin F, Wang H and Malbon CC. (2000a). J. Biol. Chem., 275, Furlan D, Acquati F, Bianchi M, Russo A, Capella C and 19025–19034. Taramelli R. (2000). Clin. Cancer Res., 6, 1422–1431. Lin X and Gelman IH. (1997). Cancer Res., 57, 2304–2312. Toyooka KO, Toyooka S, Virmani AK, Sathyanarayana UG, Lin X and Gelman IH. (2002). Biochem. Biophys. Res. Euhus DM, Gilcrease M, Minna JD and Gazdar AF. (2001). Commun., 290, 1368–1375. Cancer Res., 61, 4556–4560. Lin X, Nelson P and Gelman IH. (2000b). Mol. Cell. Biol., 20, Xia W, Unger P, Miller L, Nelson J and Gelman IH. (2001). 7259–7272. Cancer Res., 61, 5644–5651. Lin X, Nelson PJ, Frankfort B, Tombler E, Johnson R and Xiong Z and Laird PW. (1997). Nucleic Acids Res., 25, Gelman IH. (1995). Mol. Cell. Biol., 15, 2754–2762. 2532–2534.

Oncogene