Original Article | Published:

Zinc-finger protein 331, a novel putative tumor suppressor, suppresses growth and invasiveness of gastric cancer

Oncogene volume 32, pages 307317 (17 January 2013) | Download Citation


Zinc-finger protein 331 (ZNF331), a Kruppel-associated box zinc-finger protein gene, was identified as a putative tumor suppressor in our previous study. However, the role of ZNF331 in tumorigenesis remains elusive. We aimed to clarify its epigenetic regulation and biological functions in gastric cancer. ZNF331 was silenced or downregulated in 71% (12/17) gastric cancer cell lines. A significant downregulation was also detected in paired gastric tumors compared with adjacent non-cancer tissues. In contrast, ZNF331 was readily expressed in various normal adult tissues. The downregulation of ZNF331 was closely linked to the promoter hypermethylation as evidenced by methylation-specific PCR, bisulfite genomic sequencing and reexpression by demethylation agent treatment. DNA sequencing showed no genetic mutation/deletion of ZNF331 in gastric cancer cell lines. Ectopic expression of ZNF331 in the silenced cancer cell lines MKN28 and HCT116 significantly reduced colony formation and cell viability, induced cell cycle arrests and repressed cell migration and invasive ability. Concordantly, knockdown of ZNF331 increased cell viability and colony formation ability of gastric cancer cell line MKN45. Two-dimensional gel electrophoresis and mass spectrometry-based comparative proteomic approach were applied to analyze the molecular basis of the biological functions of ZNF331. In all, 10 downstream targets of ZNF331 were identified to be associated with regulation of cell growth and metastasis. The tumor-suppressive effect of ZNF331 is mediated at least by downregulation of genes involved in cell growth promotion (DSTN, EIF5A, GARS, DDX5, STAM, UQCRFS1 and SET) and migration/invasion (DSTN and ACTR3), and upregulation of genome-stability gene (SSBP1) and cellular senescence gene (PNPT1). A novel target of ZNF331 (DSTN) was functionally validated. Overexpression of DSTN in BGC-823 cells increased colony formation and migration ability. In conclusion, our results suggest that ZNF331 possesses important functions for the suppression of gastric carcinogenesis as a novel functional tumor-suppressor gene.


Gastric cancer is one of the leading causes of cancer-related death in China. Although the molecular mechanisms of gastric carcinogenesis remain unclear, epigenetic silencing of tumor-related genes by promoter hypermethylation has recently emerged as an important mechanism of tumorigenesis. The promoter hypermethylation profile differs in each cancer type and within each gene, providing tumor type- and gene-specific hypermethylation profiles that may involve in the corresponding molecular mechanism of tumorigenesis. The identification of a novel gene targeted by promoter hypermethylation may provide insights into the mechanisms for the inactivation of the tumor-suppressive pathways and is important for the identification of tumor markers in gastric cancer.1, 2 Recently, using suppression subtraction hybridization,3 we have identified < it>zinc-finger protein 331 (ZNF331) as a candidate tumor-suppressor gene. We aimed to clarify the epigenetic regulation and biological functions of ZNF331 in gastric cancer in this study.

ZNF331, also known as ZNF361, ZNF463 and Rita, is a Kruppel-associated box zinc-finger protein gene consisting of a Kruppel-associated box-A box and a zinc-finger domain with 12 Cys2His2 zinc fingers and locates on chromosome 19q13.4, 5 The role of ZNF331 in tumor development remains largely unknown. The functions of zinc-finger proteins are extraordinarily diverse, including DNA recognition, RNA packaging, transcriptional activation, regulation of apoptosis, and so on. Structural studies of zinc-finger proteins have shed new insights into their extraordinary diversity of structure and function.6 Proteins containing the classical Cys2His2 zinc finger are among the most abundant in eukaryotic genomes. Many of these proteins function by recognition of specific DNA sequences. Because all the Kruppel-associated box zinc-finger genes so far have been shown to act only as transcriptional repressors,7, 8, 9 ZNF331 is considered to be a transcriptional repressor.10 The engineered chimeric fusion proteins of Kruppel-associated box domains combined with different DNA-binding domains of oncogenic transcription factors have been shown to target their respective oncogenes and specifically suppress malignant growth.11 The application of artificial zinc-finger protein transcription factors to repress the expression of cancer-related genes, such as human telomerase reverse transcriptase, has presented a new promising strategy for inhibiting the growth of human cancer cells.12 These findings demonstrate important tumor-suppressor functions of zinc-finger proteins and their promising application in cancer therapy. In this study, we investigated the gene-expression pattern of ZNF331, molecular characterization of the promoter of ZNF331 and its tumor-suppressor function in gastric cancer.


Correlation of ZNF331 mRNA expression level with methylation status in gastric cancer cell lines

Three transcription variants of ZNF331 have been identified in human tissues previously.4 The three variants are transcribed from different loci and share the same open reading frame (ORF) (Figure 1a), suggesting that they possess the same function and might be transcriptionally regulated in a tissue-specific manner. RT–PCR results indicated that variants 1 and 2 of ZNF331 were ubiquitously expressed in various normal human tissues, including gastric tissue, whereas variant 3 was found only in the reproductive tissue, testis (Figure 1b). These results demonstrated that constitutive expression of ZNF331 in normal tissues mainly involves variants 1 and 2 but not variant 3. So we further analyzed the transcription of variants 1 and 2 in a series of cancer cell lines. RT–PCR results indicated that variant 1 was frequently silenced or downregulated in the tested gastric cancer cell lines (12/17, 71%). Expression of variant 2 was detected only in 1 out of the 17 cell lines (SNU719) by RT–PCR (Figure 1c). Real-time quantitative RT–PCR (qRT–PCR) indicated that expression of variant 2 in SNU719 was over 200 times higher than in the other cells. Dissimilarly, ZNF331 was less frequently silenced or downregulated in colon, esophageal, liver and other human cancer cell lines (data not shown). These results implied that loss of ZNF331 expression, especially variant 2, may be mainly involved in gastric carcinogenesis.

Figure 1
Figure 1

Transcriptional silencing/downregulation of ZNF331 is associated with DNA methylation. (a) Schematic for the three transcription variants of ZNF331, which are transcribed from different transcription start sites while containing the same ORF. (b) Ubiquitous mRNA expression of ZNF331 variants 1 and 2 in normal human adult tissues. (c) ZNF331 variant 1 and variant 2 expression were silenced or reduced in many gastric cancer cell lines. (d) By methylation-specific PCR, full methylation of ZNF331 variant 2 was observed in most silenced cell lines, and partial or full methylation of ZNF331 variant 1 was detected in cell lines with reduced or silenced expression. MSP, methylation-specific PCR; M, methylated; U, unmethylated. (e) The mRNA expression of ZNF331 was restored after treatment with demethylation agent 5-Aza as evidenced by RT–PCR (left) and real-time qRT–PCR (right). Bisulfite genomic sequencing results indicated that methylation level of ZNF331 was reduced after 5-Aza treatment, in concordance with its restored expression.

Sequence analysis using both CpG island searcher (http://cpgislands.usc.edu/) and CPGPLOT (http://www.cbib.u-bordeaux2.fr/pise/cpgplot.html) revealed CpG islands around the transcription start sites of both variants 1 and 2, indicating potential epigenetic regulation of ZNF331 expression by DNA methylation. To evaluate the association of silencing or downregulation of ZNF331 variants 1 and 2 with methylation status, methylation-specific PCR was carried out. Heavy methylation of ZNF331 variant 2 was observed in all silenced cell lines, whereas partial or full methylation of variant 1 was detected in cancer cells with reduced or silenced expression (Figure 1d). These results indicated a correlation between ZNF331 mRNA expression and its methylation status in gastric cancer cell lines.

To further confirm that DNA methylation is indeed responsible for the silencing or downregulation of ZNF331, six gastric cancer cell lines and one colon cancer cell line (HCT116) were treated with the demethylation agent 5-Aza-2′-deoxycytidine (5-Aza). Expression of ZNF331 variants 1 and 2 was dramatically restored or upregulated after treatment in all these cancer cell lines. Bisulfite genomic sequencing indicated lower level of methylation in 5-Aza-treated cells as compared with untreated cells (Figure 1e). These results confirmed that DNA methylation mediated the transcriptional silencing or downregulation of ZNF331.

Genetic deletion or mutation of ZNF331 was not detected in gastric cancer cell lines

To find out whether genetic alteration also contribute to the silencing of ZNF331, we performed sequence screening on the exons, exon–intron junctions and >2-kb upstream of the transcription start sites of all transcripts of ZNF331 using PCR and direct sequencing. Results showed no genetic mutation/deletion of ZNF331 in all the seven gastric cancer cell lines, including MKN28, MKN45, AGS, Kato III, SNU16, SNU719 and NCI87, suggesting that silencing or downregulation of ZNF331 might not be regulated by genetic alteration.

Downregulation of ZNF331 in primary gastric cancer tissues

ZNF331 was downregulated in gastric tumor samples as compared with their adjacent normal samples as determined by immunohistochemistry (Figure 2a). ZNF331 protein was located mainly in the nuclei of normal cells, but predominantly in the cytoplasm of cancer cells, suggesting that the subcellular localization of ZNF331 may also be regulated distinctly between cancerous and normal cells as a transcription repressor. We counted the cells with positive nuclear staining and found that ZNF331 was significantly downregulated in gastric tumor samples as compared with their adjacent normal samples (5.8% vs 39.6%, P<0.001) (Figure 2b).

Figure 2
Figure 2

Immunohistochemistry examination of ZNF331 expression in primary gastric tumor and adjacent normal samples. (a) Representative results of immunohistochemical staining for ZNF331 in gastric tumor and adjacent normal tissues. (b) Quantitation of ZNF331 expression by counting cells with positive nuclear staining.

The association between clinicopathological features and the expression of ZNF331 was evaluated in 224 primary gastric cancers by immunohistochemistry on tissue array slides. Low ZNF331 nuclear expression is associated with male gender (P<0.05), late TNM staging (P<0.05) and diffuse type of gastric cancer (P<0.001). There was no correlation between the expression of ZNF331 and other clinicopathological features such as age and Helicobacter pylori status (Supplementary Table 2). Moreover, as shown in the Kaplan–Meier survival curves (Supplementary Figure 1), gastric cancer patients with low ZNF331 expression had significantly shorter survival than those with high ZNF331 expression (P<0.05, log-rank test) However, ZNF331 expression was not associated with the outcome in gastric cancer patients by multivariate Cox regression analysis (Supplementary Table 3).

Expression of ZNF331 suppressed cell growth and induced cell cycle arrest

The frequent silencing or downregulation of ZNF331 in gastric cancer cell lines suggests that it is likely a tumor-suppressor gene. In order to test this point, we constructed and transfected a pcDNA3.1–ZNF331 expression vector into the gastric cancer cell line MKN28 and the colon cancer cell line HCT116, which shows no or low level of ZNF331 expression. Expression of ZNF331 after stable transfection was evidenced using RT–PCR and Western blot (Figure 3a). Expression of ZNF331 suppressed the growth of both MKN28 (P<0.01) and HCT116 cells (P<0.001) by MTT assay (Figure 3b). This was confirmed by colony formation assay. The colonies formed by ZNF331-transfected cells were significantly less and smaller in MKN28 (P<0.05) and HCT116 (P<0.001) than empty vector-transfected cells (Figure 3c).

Figure 3
Figure 3

Effect of ZNF331 expression on cell growth and cell-cycle phase distribution. (a) Expression of ZNF331 in MKN28 and HCT116 cells after transfection with ZNF331-expressing vectors was confirmed using RT–PCR and Western blot. (b) Representative results of MTT viability assay from empty vector- and ZNF331-transfected MKN28 and HCT116 cells. (c) Representative results of colony-formation assay from MKN28 and HCT116 cells. (d) Representative cell cycle analysis results from empty vector- and ZNF331-transfected MKN28 and HCT116 cells. (e) Western blot examination of proliferating cell nuclear antigen in control and ZNF331-transfected MKN28 and HCT116 cells.

To determine the molecular mechanism by which ZNF331 inhibits cell growth, we analyzed the effect of ZNF331 on cell cycle distribution by flow cytometry after propidium iodide staining. Ectopic expression of ZNF331 led to a significant decrease in the number of the S-phase cells of MKN28 (P<0.001) and HCT116 (P<0.05) (Figure 3d). Western blot analysis of protein extracts derived from ZNF331-transfected cells showed a clear decrease in proliferating cell nuclear antigen expression (Figure 3e) as compared with vector-transfected cells, confirming the inhibitory effect of ZNF331 on cell cycle progression. Concomitant with this inhibition, there was a significant increase in the number of cells accumulating in the G1 (P<0.05) and G2 phases (P<0.01) in MKN28 and in the G1 phase only in HCT116 (P<0.05) (Figure 3d). We also examined the contribution of apoptosis to the observed growth inhibition of ZNF331-transfected cells using flow cytometry with Annexin V:FITC and propidium iodide staining. However, no significant induction of apoptosis by ZNF331 was observed in both cell lines. These results indicated that expression of ZNF331 inhibited cancer cell growth by suppressing cell proliferation and inducing cell cycle arrest.

ZNF331 reduced migration and invasion ability of cancer cells

To investigate the effects of ZNF331 on cancer cell migration and invasive ability, the monolayer scratch healing assay and Martrigel invasion assay were performed. Ectopic expression of ZNF331 markedly slowed cell migration at the edges of scratch wound of MKN28 and HCT116 (Figure 4a). Quantitative analyses at 48 h confirmed a significant reduction in wound closure in ZNF331-transfected cells compared with empty vector-transfected control cells (Figure 4a). In addition, ZNF331 also significantly impaired the invasiveness of both MKN28 and HCT116 cells (Figure 4b). These results suggest that ZNF331 has an important inhibitory role in the invasiveness of cancer cells.

Figure 4
Figure 4

Effects of ZNF331 on MKN28 and HCT116 cell migration and invasion. (a) Representative result of scratch-healing assay. (b) Representative result of invasion assay.

Knockdown of ZNF331 increased cell proliferation in MKN45 gastric cancer cells

In order to further validate the inhibitory effect of ZNF331 on cancer cell growth, we knocked down ZNF331 expression in the ZNF331-expressing gastric cancer cell line MKN45 by short hairpin RNA vector transfection. Knockdown efficiency was evaluated by both real-time qRT–PCR and Western blot (Figure 5a). Knockdown of ZNF331 increased cell growth in MKN45 cells by MTT cell viability assay (Figure 5b). In keeping with this, knockdown of ZNF331 significantly increased colony formation ability of MKN45 cells (Figure 5c). These results further indicated that ZNF331 acts with a tumor-suppressor function in cancer cells.

Figure 5
Figure 5

Knockdown of ZNF331 increased cell proliferation. (a) Evaluation of knockdown of ZNF331 in MKN45 cells after transfection with sh-ZNF331 (knockdown) and sh-Scrambled (control) using qRT–PCR and Western blot. (b) Representative results of MTT viability assay. (c) Representative results of colony-formation assay.

Identification of genes modulated by ZNF331

To elucidate the molecular mechanism underlying the inhibitory effect of ZNF331 on cell growth and invasiveness, protein-expression profiles in ZNF331-transfected MKN28 cells and vector-transfected MKN28 cells were analyzed using 2D gel assay. In all, 33 spots were found differentially expressed with fold changes more than two and they were subjected to further analysis by the combined mass spectrometry (MS) and MS/MS analysis (Figure 6). Ten spots were well recognized (Figure 7a) (Table 1). When compared with control vector-transfected cells, ZNF331 exerted its anti-growth effect by downregulation of six genes involved in cell growth and proliferation, including DEAD (Asp-Glu-Ala-Asp) box proteins (DDX5, −4.6-fold), eukaryotic translation initiation factor 5A (EIF5A, −3.5-fold), glycyl-tRNA synthetase (GARS, −3.5-fold), signal-transducing adaptor molecule (STAM, −3.3-fold), ubiquinol-cytochrome-c reductase 1 (UQCRFS1, −3.1-fold) and SET nuclear oncogene (SET, −4.8-fold). The anti-invasive activity of ZNF331 was mediated by downregulation of pro-metastatic genes, destrin (DSTN, −6-fold) and ARP3 actin-related protein 3 homolog (ACTR3, −2.8-fold). Further, ZNF331 induced the human PNPase gene (PNPT1) expression by 4.3-fold, a pro-cellular senescence regulator and single-stranded DNA-binding protein 1 (SSBP1) by 4.6-fold, which has an important role in the maintenance of genome stability (Table 1). These changes on expression profiling were validated using Western blot (Figure 7b) and real-time qRT–PCR (Figure 7c) in MKN28 and HCT116 cells. In addition, ZNF331 downregulated the expression of DDX17 (the gene accompanying DDX5 in cell growth regulation) and c-Myc (a downstream target of both PNPT1 and STAM) as demonstrated by real-time qRT–PCR in both cell lines.

Figure 6
Figure 6

Identification of differentially expressed proteins modulated by ZNF331 expression. (a) 2D gel electrophoresis images of detergent extracts from vector- and ZNF331-transfected MKN28 cells. The first dimension was run on linear 13 cm IPG strip, pH 3–10. The second dimension was run on 10% SDS–PAGE. The major spots identified by MS and MS/MS analysis were labeled. (b) Histogram of fold-changes of the differentially expressed genes in ZNF331-transfected cells vs vector-transfected cells.

Figure 7
Figure 7

Molecular mechanism of the inhibitory activity of ZNF331 on gastric cancer growth. (a) Enlarged images of 10 differentially expressed protein spots of interest. Selected regions of Two-dimensional gel electrophoresis (2-DE) gels illustrate differentially expressed proteins between vector- and ZNF331-transfected MKN28 cells. Spots of interest are shown in circles. SET has two isoforms and the spots 1 and 2 were both identified to be SET proteins by MS/MS ion search. (b) Validation of differentially expressed genes using Western blot. (c) Validation of differentially expressed genes by real-time qRT–PCR. (d) Schematic diagram for the molecular mechanism of the anti-tumorigenesis functions of ZNF331 deriving from 2-DE, Western blot and real-time qRT–PCR.

Table 1: Differentially expressed proteins in ZNF331-transfected MKN28 cells vs vector-transfected MKN28 cells

The correlations between ZNF331 and its downstream targets, as well as their association with the inhibition of gastric cancer growth, were shown in Figure 7d. These results unveiled the molecular mechanism by which ZNF331 inhibited the growth and invasiveness of cancer cells.

Functional validation of a novel ZNF331 downstream target DSTN

We further selected DSTN, a novel downstream target of ZNF331, for further functional validation. DSTN is required for cell migration and invasion in colon cancer cells (Estornes et al.13). Ectopic expression of DSTN in BGC-823 cells (Figure 8a) significantly increased colony formation ability as compared with empty vector-transfected BGC-823 cells (Figure 8b). In addition, scratch-healing assay indicated that DSTN significantly increased migration ability of BGC-823 cells (Figure 8c). These results indicated that ZNF331 inhibited gastric cancer growth, at least in part, by downregulation of DSTN.

Figure 8
Figure 8

Functional validation on DSTN, a novel downstream target of ZNF331. (a) Expression of DSTN in BGC-823 cells after transfection with DSTN-expressing vectors was confirmed by RT–PCR and Western blot. (b) Reexpression of DSTN reduced the colony-formation ability in BGC-823 cells. (c) Reexpression of DSTN inhibited cell-migration ability in BGC-823 cells as evidenced by scratch-healing assay.


In this study, we identified and characterized a putative tumor-suppressor gene, ZNF331, in gastric cancer. The three transcript variants of ZNF331 share the same ORF, which suggests they exert the same function. Transcript variants 1 and 2 were readily expressed in normal human tissues but frequently silenced or downregulated in gastric cancer cell lines, whereas variant 3 was not found in normal tissue except the reproductive tissue (testis) (Figure 1b) and thus may not be involved in tumorigenesis. In addition, ZNF331 was also downregulated in gastric tumor samples compared with their adjacent normal samples by immunohistochemistry. Low ZNF331 expression was further revealed to be associated with advanced stage (P<0.05) and diffuse type (P<0.001) of gastric cancers. Moreover, low ZNF331 expression was correlated with poor survival of gastric cancer patients by Kaplan–Meier survival analysis (P<0.05). These results suggested that the downregulation of ZNF331 may be involved in tumor progression during gastric cancer development. Promoter methylation was demonstrated to mediate the transcriptional silence of cancer cell lines at variants 1 and 2 of ZNF331. Demethylation treatment by 5-Aza successfully restored the expression of ZNF331, further confirming its regulation by DNA methylation. In addition, no genetic mutation or deletion of ZNF331 was found in all the seven gastric cancer cell lines tested. Collectively, these results revealed the epigenetic regulation mechanism on the expression of ZNF331 in gastric cancer cells.

In cell biology experiments, the full-length ORF was cloned into the expression vector pcDNA3.1 for gain-of-function study, whereas in the knockdown experiment mediated by short hairpin RNA transfection, the ORF was targeted. Functional studies revealed that restoration of ZNF331 in cancer cell lines MNK28 and HCT116 significantly inhibited colony formation and cell proliferation. Concordantly, knockdown of ZNF331 increased proliferation of MKN45 cells as evidenced by cell viability and colony-formation assays. Flow cytometry analysis indicated that the suppression of cell growth by ZNF331 was mediated by cell-cycle phase arrest, without significant induction of apoptosis. Expression of ZNF331 decreased the number of both MKN28 and HCT116 cells in S phase (Figure 3d), inferring that ZNF331 blocked cell proliferation. There was an increase in the number of MKN28 cells in the G2/M phase, suggesting that G2 arrest was also involved. Furthermore, ectopic expression of ZNF331 reduced the migration and invasion ability of these cancer cells. These results indicate for the first time that ZNF331 functions as a tumor suppressor in gastric cancer.

The molecular mechanisms by which ZNF331 exerts its anti-growth and anti-invasive functions in cancer cells were defined using two-dimensional gel electrophoresis protein profiling assay. Genes that were significantly altered in protein levels were then validated using RT–PCR and immunoblotting. We demonstrated that the suppression of cell growth and proliferation by ZNF331 was mediated by downregulation of genes involved in cell growth, including DDX5, DDX17, EIF5A, GARS, STAM, UQCRFS1 and SET oncogene (Figure 7d). The DEAD (Asp-Glu-Ala-Asp) box proteins DDX5 and DDX17 have important roles in ribosome biogenesis and cell proliferation. Overexpression of DDX5 and DDX17 has been found in colon, breast and prostate cancers, suggesting that DDX5/DDX17 promotes tumorigenesis as proto-oncoproteins.14 Co-silencing of both DDX5 and DDX17 causes perturbation of nucleolar structure and cell proliferation, whereas knocking down of either gene does not affect cell proliferation.15 Thus, downregulation of both DDX5 and DDX17 by ZNF331 contributes to its anti-proliferation effect. In addition, inhibition of EIF5A hypusination has been reported to impair melanoma growth;16 GARS, the glycyl-tRNA synthetase gene, has an important role in the synthesis of proteins;17 STAM, the signal-transducing adaptor molecule, is involved in signaling for cell growth through induction of the oncogene c-Myc;18 UQCRFS1 is known as the complex III of the mitochondrial respiratory chain passing electrons from reduced ubiquinol to cytochrome-c during the process of synthesis of ATP.19 The amplification of UQCRFS1 has been reported in breast, ovarian and gastric cancers,20, 21, 22 which has an important role for dividing and proliferating cancer cells. The SET nuclear oncogene is involved in the regulation of cell growth23 and functions as a specific inhibitor of a gene suppressing tumor metastasis, non-metastatic cell 1.24 Collectively, suppression of DDX5/DDX17, EIF5A, GARS and STAM (followed by c-Myc downregulation), UQCRFS1 and SET may mediate the inhibitory activity by ZNF331 as a repressor on cancer cell growth.

Cell cycle analysis indicated an increase in G2-phase cells in ZNF331-transfected MKN28 (Figure 3d). Members of the DEAD box proteins have been reported to be involved in the regulation of cell division.25 A recent study has indicated that the proliferation-associated nuclear antigen DDX5 has an important role in mediating the G2-/M-phase arrest.26 We found that DDX5, a member of the DEAD box family proteins, was significantly downregulated by ZNF331 in MKN28 (−4.6-fold change) (Table 1). Thus, the G2-phase block by ZNF331 may be regulated through the downregulation of DDX5.

The anti-invasion function of ZNF331 appeared to be associated with the downregulation of DSTN and ACTR3 (Figure 7d). It has been reported that DSTN is required for cell migration and invasion.13 ACTR3 is known to be a major constituent of the ARP2/3 complex. The frequency of expression of ARP2/3 by the stromal cells increased with the atypia of the colorectal neoplasms. It contributes to the increased motility of both neoplastic and stromal cells, and thus provides suitable conditions for invasion.27 Thus, downregulation of DSTN and ACTR3 contributes to the reduced migration and invasion ability induced by ZNF331.

Two genes (PNPT1 and SSBP1) were found to be upregulated following ZNF331 reexpression. PNPT1, the gene encoding human PNPase, is thought to contribute to cellular senescence through its RNA-degrading activity in the cytosol.28, 29 It was also found to induce growth arrest by downregulating expression of c-Myc in human melanoma cells.30 Therefore, upregulation of PNPT1, followed by the downregulation of c-Myc, may contribute to the growth-inhibitory activity induced by ZNF331. SSBP1 is a housekeeping gene involved in mitochondrial biogenesis.31 It has an important role in the DNA-damage response and maintenance of genome stability,32 inferring that ZNF331 can increase genomic stability by upregulating SSBP1.

In conclusion, we characterized a novel tumor suppressor ZNF331 regulated epigenetically by DNA methylation in gastric cancer. ZNF331 suppressed growth and invasiveness of cancer cells through repressing genes involved in cell proliferation (DSTN, DDX5/DDX17, EIF5A, GARS, STAM, UQCRFS1 and SET) and migration/invasion (DSTN and ACTR3), as well as inducing a cellular senescence gene (PNPT1) and a genome-stability gene (SSBP1) (Figure 7d). ZNF331 possesses important functions for the suppression of gastric cancer.

Materials and methods

Cancer cell lines

Seventeen gastric cancer cell lines (MKN28, MKN45, NCI87, BGC-823, AGS, Kato III, YCC1, YCC 2, YCC 3, YCC 6, YCC 7, YCC 9, YCC 10, YCC 11, YCC 16, SNU1, SNU16 and SNU719) and one colon cancer cell line (HCT116) were obtained from American Type Culture Collection (Manassas, VA, USA). The immortalized normal human gastric epithelial cell line GES-1 was provided by School of Oncology, Beijing University. Cells were cultured in RPMI 1640, DMEM or McCoy medium (Gibco BRL, Rockville, MD, USA) supplemented with 10% fetal bovine serum (Gibco BRL).

Semi qRT–PCR and real-time qRT–PCR analyses

Total RNA was extracted from cell pellets using QiaZol reagent (Qiagen, Valencia, CA, USA). cDNA was synthesized from 2 μg of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). Semi-qRT–PCR and real-time qRT–PCR analyses were performed using primers listed in Supplementary Table 1.

Demethylation treatment

Cells were seeded at a density of 1 × 105 cells/ml in 100-mm dishes and grew for 24 h. Cells were then treated with 2 μM 5-Aza (Sigma-Aldrich, St Louis, MO, USA) for 48 h, with culture medium containing 5-Aza replenished every day. Then cells were harvested and gene expression was analyzed using RT–PCR.

DNA extraction, methylation-specific PCR and bisulfite genomic sequencing

Genomic DNA from cancer cells and tissue samples were isolated using the QIAamp DNA mini kit (Qiagen, Hilden, Germany). A total of 1 μg of DNA was modified by sodium metabisulfite as previously described.33 The bisulfite-modified DNA was then amplified by methylation-specific PCR using primer pairs that specifically amplify either methylated or unmethylated sequences of the ZNF331 gene. Bisulfite genomic sequencing was performed to assess the methylation levels of nine CpG sites spanning +92 to +285 of variant 1 of ZNF331 in cells with or without 5-Aza treatment. Nucleotide sequences of the primers used in methylation-specific PCR and bisulfite genomic sequencing were listed in Supplementary Table 1.

Immunohistochemistry and tissue array analysis

Paired primary tumor and adjacent non-tumor samples were obtained from 10 gastric cancer patients after surgical resection and from 224 gastric tumors in tissue array slides. Tissue types (tumor or normal) were assessed by histological staining. Immunohistochemistry was performed on 5-μm paraffin sections using anti-ZNF331 antibodies (Abcam, Cambridge, UK).34 ZNF331 staining in the nucleus and cytoplasm was evaluated by scanning of the whole section and counting >1000 representative cells. The percentages of cells with positive nuclear staining were used as immunohistochemical scores. The signal intensities and the percentages of positive staining were both graded as 0, 1, 2 and 3 independently. Then the score of a sample was defined as ‘high’ with ‘intensity+percentage’ 3, otherwise as ‘low’ (‘intensity+percentage’ <3). The ethics committee of the Chinese University of Hong Kong approved of this study, and written consents were obtained from all patients involved.

DNA mutation analysis

Mutation of ZNF331-coding sequences were determined by direct DNA sequencing using an ABI prism 3100 genetic analyzer (Applied Biosystems). Sequence homology to published database was analyzed with the BLAST program at the internet site of National Centre for Biological Information.

Construction of expression vector for ZNF331 and DSTN

The expression vectors (pcDNA3.1–ZNF331 and pcDNA3.1–DSTN) encoding the full-length ORFs of human ZNF331 and DSTN genes were constructed. Briefly, DNA sequences corresponding to the ORFs of ZNF331 or DSTN were generated by RT–PCR. The PCR products were confirmed by direct DNA sequencing and cloned into the mammalian expression vector pcDNA3.1 (Invitrogen, Carlsbad, CA, USA). The sequences of the constructs were further confirmed by sequencing. The ZNF331 silencing cancer cell lines MKN28 and HCT116 were transfected with pcDNA3.1–ZNF331 or empty vector, whereas DSTN-downregulated BGC-823 cells were transfected with pcDNA3.1–DSTN or empty vector, using Lipofectamine 2000 (Invitrogen). Stably transfected cells were established under selection with neomycin (G418) (Invitrogen).

Knockdown of ZNF331

A set of vectors carrying short hairpin RNAs against ZNF331 was purchased from Origene (Rockville, MD, USA). The ZNF331-expressing gastric cancer cell line MKN45 was transfected, with vectors carrying scrambled sequence (sh-scrambled) as negative control. Knockdown efficiency was evaluated 3 days after transfection by real-time qRT–PCR. The one with the highest knockdown efficiency was further used to establish stable knockdown cells under selection with puromycin (Invitrogen) for colony-formation assay and cell-viability assay.

Colony-formation assay

For overexpression assay, MKN28 and HCT116 cells (2 × 105 cells/well) were plated in a 12-well plate and transfected with pcDNA3.1–ZNF331 or empty vector. BGC-823 cells were transfected with pcDNA3.1–DSTN or empty vector. For knockdown assay, MKN45 cells were transfected with sh-ZNF331 or sh-scrambled. Two days after transfection, cells were subsequently split at 1:20 ratio on six-well plates with G418 or puromycin. After culturing for 14–21 days, cells were fixed with 70% ethanol and stained with crystal violet solution. Colonies with >50 cells per colony were counted. All experiments were conducted three times in triplicates.

Cell-viability assay

Cell proliferation of stably transfected cells was examined using the Vybrant MTT Cell Proliferation Assay Kit (Invitrogen) according to the manufacturer’s instructions. The experiment was conducted in three independent triplicates and results were shown as the means±s.d.

Flow cytometry

The stably transfected MKN28 and HCT116 cells with pcDNA3.1–ZNF331-expressing or pcDNA3.1 empty vector were fixed in 70% ethanol and stained with 50 μg/ml propidium iodide (BD Pharmingen, San Jose, CA, USA). The cells were then sorted by FACSCalibur (BD Biosciences, Franklin Lakes, NJ, USA), and cell-cycle profiles were analyzed by ModFit 3.0 software (Verity Software House, Topsham, ME, USA). Apoptosis was determined by dual staining with Annexin V:FITC and propidium iodide (Invitrogen). The Annexin V-positive cells were counted as apoptotic cells.

Scratch-healing assay

Cell migration was assessed using scratch-healing assays. Briefly, MKN28 and HCT116 cells stably transfected with pcDNA3.1–ZNF331. BGC-823 cells stably transfected with pcDNA3.1–DSTN or empty vectors were cultured in six-well plates. When the cells grew up to 90% confluence, three scratch wounds across each well were made using a P-200 pipette tip. Fresh medium supplemented with reduced (5%) fetal bovine serum was added, and the wound-closing procedure was observed for 48 h. Photographs were taken at 0, 24 and 48 h, respectively.

Invasion assay

Invasion assay was performed using BD BioCoat Growth Factor Reduced MATRIGEL Invasion Chamber (BD Biosciences) according to the manufacturer’s instructions. Complete culture medium (supplemented with 10% fetal bovine serum) was used as chemoattractant. The insert membranes were stained with crystal violet and the permeating cells were counted under inverted microscope and photographed. Three independent experiments were performed and results were shown as the means±s.d.

Two-dimensional gel electrophoresis and image analysis

Samples containing 150 mg protein were dissolved in 250 μl rehydration solution (8 M urea, 2% CHAPS, 0.4%. DTT, 0.5% IPG buffer, 0.002% bromophenol blue) and subjected to two-dimensional gel electrophoresis as previously described by us.35 The isoelectric focussing was carried out using IPG-PhorII apparatus (Amersham, Piscataway, NJ, USA). All Immobiline DryStrip gels strips (13 cm, pH 3–10, nonlinear) were used according to the manufacturer's instructions. Two-dimensional SDS–PAGE was performed in tris-glycine buffer at a constant current setting of 15 mA/gel for 30 min and 30 mA/gel thereafter. Gels were stained by a modified silver-staining method, which was compatible with the MS analysis, as previously described.35 Gel images were digitalized using a scanner (GS-800 calibrated densitometer, Bio-Rad, Hercules, CA, USA) and analyzed using the software PDQuest (version 8.0, Bio-Rad) with manual editing. Spots on different gels were matched and analyzed. Statistical analysis was performed using the independent-samples t-test.

Protein identification using MS

Differential protein spots were excised from stained gels, and subjected to protein identification by undertaking the in-gel digestion approach and using MALDI TOF/TOF MS. Briefly, the gel pieces were destained, reduced with 1.75% DTT, alkylated with 350 mM iodoacetamide and digested with modified porcine trypsin overnight (sequencing grade; Promega, Madison, WI, USA). The tryptic peptides were harvested, cleaned up with C18 ZipTips (Millipore Corp., Billerica, MA, USA) and subjected to MALDI-TOF/TOF MS (Ultraflex-III, Bruker Daltonics, Bremen, Germany) with α-Cyano 4-hydroxy cinnamic acid as the matrix. The MS and MS/MS spectra were automatically processed with the FlexAnalysis program (version 3.0, Bruker Daltonics) with the default parameters. The MS spectrum data were searched via the MASCOT search engine to obtain the protein identity by undertaking the peptide mass fingerprinting approach and the MS/MS ion search approach. For the search parameters, one missed cleavage in trypsin digestion was allowed; partial oxidation of methionine, phosphorylation of serine/threonine/tyrosine and iodoacetamide modification of cysteine residues were selected. The error-tolerance values of the parent peptides and the MS/MS ion masses were 50 ppm and 0.1 Da, respectively. A search result having an expectation value <0.05 was considered statistically significant. For a gel spot, an identification result was considered valid when both peptide mass fingerprinting and MS/MS ion search identified the same protein as the statistically significant hit from the Swiss-Prot database, or MS/MS ion search identified at least two tryptic peptides with sequences from the same protein as the statistically significant hits.

Western blot analysis

Total protein was extracted and protein concentration was then measured by the DC protein-assay method of Bradford (Bio-Rad). A total of 40 μg of protein from each sample were separated by 10% SDS–PAGE gel and transferred to an equilibrated polyvinylidene difluoride membrane (Amersham Biosciences, Buckinghamshire, UK). After incubation with specific primary antibody at 4 °C overnight and then the secondary antibody, the proteins were detected by enhanced chemiluminescence (Amersham Corporation, Arlington Heights, IL, USA). The primary antibodies anti-ZNF331, anti-DSTN and anti-EIF5A were purchased from Abcam. Antibodies against PNPT1 and DDX5 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies against proliferating cell nuclear antigen and GAPDH were purchased from Cell Signaling Technology (Danvers, MA, USA).

Statistical analysis

All measurements or variables were shown as means±s.d. Results of colony-formation assays, flow-cytometry analyses, cell growth, migration and invasion assays were analyzed by analysis of variance. Survival analyses were assessed using the Kaplan–Meier method and a Cox proportional hazards regression model. A difference was considered statistically significant when P-value is <0.05. All statistical analyses were conducted using the SPSS program (version 17.0; SPSS, Chicago, IL, USA).


  1. 1.

    , , , . Profile of aberrant CpG island methylation along multistep gastric carcinogenesis. Lab Invest 2003; 83: 519–526.

  2. 2.

    , , , , , et al. Frequent CpG island methylation in precursor lesions and early gastric adenocarcinomas. Oncogene 2004; 23: 4646–4654.

  3. 3.

    , , , , , et al. The tumor suppressor UCHL1 forms a complex with p53/MDM2/ARF to promote p53 signaling and is frequently silenced in nasopharyngeal carcinoma. Clin Cancer Res 2010; 16: 2949–2958.

  4. 4.

    , , , , , . A KRAB zinc finger protein gene is the potential target of 19q13 translocation in benign thyroid tumors. Genes Chromosomes Cancer 1999; 26: 229–236.

  5. 5.

    , , , , , et al. Isolation, characterization, and mapping of a novel human KRAB zinc finger protein encoding gene ZNF463. Biochim Biophys Acta 2001; 1518: 190–193.

  6. 6.

    , , . Zinc finger proteins: new insights into structural and functional diversity. Curr Opin Struct Biol 2001; 11: 39–46.

  7. 7.

    , , , , , . Kruppel-associated boxes are potent transcriptional repression domains. Proc Natl Acad Sci USA 1994; 91: 4509–4513.

  8. 8.

    , , , , . Repression of transcriptional activity by heterologous KRAB domains present in zinc finger proteins. FEBS Lett 1995; 369: 153–157.

  9. 9.

    , , , , . The Kruppel-associated box-A (KRAB-A) domain of zinc finger proteins mediates transcriptional repression. Proc Natl Acad Sci USA 1994; 91: 4514–4518.

  10. 10.

    , , , , , . A 3.4-kbp transcript of ZNF331 is solely expressed in follicular thyroid adenomas. Cytogenet Genome Res 2003; 101: 113–117.

  11. 11.

    , , , , . An engineered PAX3-KRAB transcriptional repressor inhibits the malignant phenotype of alveolar rhabdomyosarcoma cells harboring the endogenous PAX3-FKHR oncogene. Mol Cell Biol 2000; 20: 5019–5031.

  12. 12.

    , , , , , et al. Repression of human telomerase reverse transcriptase using artificial zinc finger transcription factors. Mol Cancer Res 2010; 8: 246–253.

  13. 13.

    , , , , , et al. Differential involvement of destrin and cofilin-1 in the control of invasive properties of Isreco1 human colon cancer cells. Int J Cancer 2007; 121: 2162–2171.

  14. 14.

    . Multi-talented DEAD-box proteins and potential tumor promoters: p68 RNA helicase (DDX5) and its paralog, p72 RNA helicase (DDX17). Am J Transl Res 2010; 2: 223–234.

  15. 15.

    , , . Redundant role of DEAD box proteins p68 (Ddx5) and p72/p82 (Ddx17) in ribosome biogenesis and cell proliferation. Nucleic Acids Res 2007; 35: 3590–3601.

  16. 16.

    , , , , , et al. Inhibition of eukaryotic translation initiation factor 5A (eIF5A) hypusination impairs melanoma growth. Cell Biochem Funct 2007; 25: 109–114.

  17. 17.

    , , . Glycyl-tRNA synthetase. Biol Chem Hoppe Seyler 1996; 377: 343–356.

  18. 18.

    , , , , , et al. STAM, signal transducing adaptor molecule, is associated with Janus kinases and involved in signaling for cell growth and c-myc induction. Immunity 1997; 6: 449–457.

  19. 19.

    , , , , , . Structure, sequence and location of the UQCRFS1 gene for the human Rieske Fe-S protein. Gene 1995; 155: 207–211.

  20. 20.

    , , , . Ubiquinol cytochrome c reductase (UQCRFS1) gene amplification in primary breast cancer core biopsy samples. Gynecol Oncol 2004; 93: 54–58.

  21. 21.

    , , , , , . CA125 and UQCRFS1 FISH studies of ovarian carcinoma. Gynecol Oncol 2003; 90: 29–36.

  22. 22.

    , , , , , et al. Comprehensive analysis of 19q12 amplicon in human gastric cancers. Mod Pathol 2006; 19: 854–863.

  23. 23.

    , , , , , et al. PCOTH, a novel gene overexpressed in prostate cancers, promotes prostate cancer cell growth through phosphorylation of oncoprotein TAF-Ibeta/SET. Cancer Res 2005; 65: 4578–4586.

  24. 24.

    , , , , . Tumor suppressor NM23-H1 is a granzyme A-activated DNase during CTL-mediated apoptosis, and the nucleosome assembly protein SET is its inhibitor. Cell 2003; 112: 659–672.

  25. 25.

    , . RNA splicing. Alive with DEAD proteins. Nature 1991; 349: 463–464.

  26. 26.

    , . The DEAD-box RNA helicase DDX3 interacts with DDX5, co-localizes with it in the cytoplasm during the G2/M phase of the cycle, and affects its shuttling during mRNP export. J Cell Biochem 2011; 113: 985–996; e-pub ahead of print 27 October 2011; doi: 10.1002/jcb.23428.

  27. 27.

    , , , , , et al. Involvement of Arp2/3 complex in the process of colorectal carcinogenesis. Mod Pathol 2004; 17: 461–467.

  28. 28.

    , , , , , et al. Identification and cloning of human polynucleotide phosphorylase, hPNPase old-35, in the context of terminal differentiation and cellular senescence. Proc Natl Acad Sci USA 2002; 99: 16636–16641.

  29. 29.

    , , , , , . Defining the domains of human polynucleotide phosphorylase (hPNPaseOLD-35) mediating cellular senescence. Mol Cell Biol 2005; 25: 7333–7343.

  30. 30.

    , , , , , et al. Down-regulation of Myc as a potential target for growth arrest induced by human polynucleotide phosphorylase (hPNPaseold-35) in human melanoma cells. J Biol Chem 2003; 278: 24542–24551.

  31. 31.

    , , , , , et al. Chromosomal localization of mitochondrial transcription factor A (TCF6), single-stranded DNA-binding protein (SSBP), and endonuclease G (ENDOG), three human housekeeping genes involved in mitochondrial biogenesis. Genomics 1995; 25: 559–564.

  32. 32.

    , , , . SOSS complexes participate in the maintenance of genomic stability. Mol Cell 2009; 35: 384–393.

  33. 33.

    , , , , , et al. A genome-wide approach to identifying novel-imprinted genes. Hum Genet 2008; 122: 625–634.

  34. 34.

    , , , , , et al. Increased expression of survivin in gastric cancer patients and in first degree relatives. Br J Cancer 2002; 87: 91–97.

  35. 35.

    , , , , , et al. Comparative proteomic analysis of mesenchymal stem cells derived from human bone marrow, umbilical cord, and placenta: implication in the migration. Proteomics 2009; 9: 20–30.

Download references


This project was supported by the National Basic Research Program of China (973 Program, 2010CB529305), Research Grants Council RGC CERG CUHK (473008), CUHK Focused Investment Grant (1903026) and RFCID (10090942, 11100022).

Author information

Author notes

    • J Yu
    • , Q Y Liang
    •  & J Wang

    These authors contributed equally to this work


  1. Institute of Digestive Disease and Department of Medicine and Therapeutics, Li Ka Shing Institute of Health Sciences, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong, China

    • J Yu
    • , Q Y Liang
    • , J Wang
    • , S Wang
    • , T C W Poon
    • , M Y Y Go
    •  & J J Y Sung
  2. Cancer Epigenetics Laboratory, Department of Clinical Oncology, The Chinese University of Hong Kong, Hong Kong, China

    • Y Cheng
    •  & Q Tao
  3. State Key Laboratory of Biomembrane and Membrane Biotechnology, School of Medicine, National Engineering Laboratory for Anti-tumor Therapeutics, Tsinghua University, Beijing, China

    • Z Chang


  1. Search for J Yu in:

  2. Search for Q Y Liang in:

  3. Search for J Wang in:

  4. Search for Y Cheng in:

  5. Search for S Wang in:

  6. Search for T C W Poon in:

  7. Search for M Y Y Go in:

  8. Search for Q Tao in:

  9. Search for Z Chang in:

  10. Search for J J Y Sung in:

Competing interests

The authors declare no conflict of interest.

Corresponding authors

Correspondence to J Yu or J J Y Sung.

Supplementary information



two-dimensional gel electrophoresis




ARP3 actin-related protein 3 homolog


DEAD (Asp-Glu-Ala-Asp) box proteins 5




ukaryotic translation initiation factor 5A


glycyl-tRNA synthetase


methylation specific PCR


the human PNPase gene


quantitative RT–PCR


reverse transcript PCR


single-stranded DNA binding protein


signal transducing adaptor molecule


Ubiquinol-cytochrome-c reductase, Rieske iron-sulfur polypeptide


zinc-finger protein

About this article

Publication history







Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Further reading