Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2008 Jun 24.
Published in final edited form as: Oncogene. 2007 Mar 12;26(38):5553–5563. doi: 10.1038/sj.onc.1210351

Cancer cells express aberrant DNMT3B transcripts encoding truncated proteins

KR Ostler 1, EM Davis 1, SL Payne 1,3, BB Gosalia 1, J Expósito-Céspedes 1, MM Le Beau 1,2, LA Godley 1,2
PMCID: PMC2435620  NIHMSID: NIHMS48740  PMID: 17353906

Abstract

Cancer cells display an altered distribution of DNA methylation relative to normal cells. Certain tumor suppressor gene promoters are hypermethylated and transcriptionally inactivated, whereas repetitive DNA is hypomethylated and transcriptionally active. Little is understood about how the abnormal DNA methylation patterns of cancer cells are established and maintained. Here, we identify over 20 DNMT3B transcripts from many cancer cell lines and primary acute leukemia cells that contain aberrant splicing at the 5′ end of the gene, encoding truncated proteins lacking the C-terminal catalytic domain. Many of these aberrant transcripts retain intron sequences. Although the aberrant transcripts represent a minority of the DNMT3B transcripts present, Western blot analysis demonstrates truncated DNMT3B isoforms in the nuclear protein extracts of cancer cells. To test if expression of a truncated DNMT3B protein could alter the DNA methylation patterns within cells, we expressed DNMT3B7, the most frequently expressed aberrant transcript, in 293 cells. DNMT3B7-expressing 293 cells have altered gene expression as identified by microarray analysis. Some of these changes in gene expression correlate with altered DNA methylation of corresponding CpG islands. These results suggest that truncated DNMT3B proteins could play a role in the abnormal distribution of DNA methylation found in cancer cells.

Keywords: DNA methylation, DNMT3B, aberrant mRNA splicing, epigenetics

Introduction

Cytosine methylation and histone modifications play an important role in the transcriptional regulation of cellular genes (Robertson, 2005). DNA is methylated at the 5-C position of cytosines that are part of CpG dinucleotides. In normal cells, repetitive DNA is highly methylated and transcriptionally silenced, effectively inactivating transposable elements that could mediate genomic rearrangements. Open chromatin associated with actively transcribed genesis hypomethylated. DNA methylation is used to control gene expression in a variety of normal cellular processes, including X chromosome inactivation, genomic imprinting and aging.

Cancer cells are characterized by abnormal patterns of DNA methylation (Robertson, 2005). Repetitive DNA sequences are hypomethylated and transcriptionally active, and the altered chromatin structure of these regions is thought to contribute to the formation of some of the chromosomal rearrangements seen in cancer cells. Additionally, some gene promoters are hyper-methylated in tumor cells, resulting in transcriptional silencing of tumor suppressor genes without accompanying inactivating mutations.

Three DNA methyltransferase (DNMT) enzymes have been identified in eukaryotic cells (Bestor, 2000; Rountree et al., 2001). DNMT1 is generally considered to be a maintenance methylase, although it also has de novo methylase activity (Bestor, 2000; Robertson, 2002). Both DNMT3A and DNMT 3B catalyse de novo methylation of DNA sequences (Li, 2002). Each of these three enzymes is essential for life, since homozygous knockout alleles of Dnmt1 and Dnmt3b cause embryonic lethality in mice, and mice with homozygous knockout alleles of Dnmt3a die several weeks after birth (Li et al., 1992; Okano et al., 1999). Dnmt1 −/− embryonic stem cells display extensive demethylation of endogenous retroviral DNA (Li et al., 1992), and murine embryonic stem cells lacking Dnmt3b demonstrate hypomethylation of minor satellite sequences (Okano et al., 1999). Patients with a rare autosomal recessive syndrome, the immunodeficiency, centromere instability, facial anomalies (ICF) syndrome, have germline mutations in the DNMT3B gene (Hansen et al., 1999; Xu et al., 1999; Shirohzu et al., 2002), and lymphocytes from affected individuals display hypomethylation of repetitive DNA sequences. Furthermore, mice expressing Dnmt3b alleles similar to those found in ICF syndrome are small with abnormal craniofacial development and hypomethylation of repetitive elements, suggesting that these alleles encode hypomorphic proteins (Ueda et al., 2006).

Numerous studies have implicated both DNMT1 and DNMT3B in the altered distribution of DNA methylation in cancer cells (Beaulieu et al., 2002; Robert et al., 2003). A human colon cancer cell line, HCT116 deficient in DNMT1 following targeted recombination, demonstrates demethylation of pericentromeric satellite sequences (Rhee et al., 2000), whereas HCT116 cells lacking both DNMT1 and DNMT3B contain demethylated satellite 2 and Alu repetitive sequences (Rhee et al., 2002) and demonstrate increased chromosomal instability (Karpf and Matsui, 2005). Additional studies of the doubly targeted cells suggest that at least some histone modifications occur before alterations in DNA methylation (Bachman et al., 2003). Mice expressing a hypomorphic Dnmt1 protein develop T-cell lymphomas (Gaudet et al., 2003), and when this deficiency is crossed into other tumor models, tumor incidence increases (Eden et al., 2003; Yamada et al., 2005). Dnmt3b deficiency inhibits the formation of macroadenomas in the murine Apc Min/+colon cancer model, indicating a requirement for Dnmt3b activity in the transition from microadenoma to tumor (Lin et al., 2006).

Experiments designed to understand the mechanism by which cancer cells have altered patterns of DNA methylation have not yielded a simple explanation. Mutations of the DNMT genes are rare in cancer cells (Kanai et al., 2003). Several groups have examined the expression of DNMT transcripts and have not found a correlation with DNA methylation levels in cancer cells (Robertson et al., 1999; Saito et al., 2002; Ehrlich et al., 2006). However, expression of transcripts that originate from a promoter within intron 4, ΔDNMT3B1-7, correlates with the DNA methylation states of the p16 and RASSF1A promoters in non-small-cell lung cancers (Wang et al., 2006a). Interestingly, several DNMT3B transcripts are predicted to encode proteins lacking critical parts of or the entire catalytic domain and therefore would produce catalytically inactive proteins: DNMT3B4 and DNMT3B5 encode proteins lacking the final two methyltransferase domains (Robertson et al., 1999), and three of the recently described transcripts from non-small-cell lung cancers, ΔDNMT3B5-7, also encode truncated DNMT3B proteins (Wang et al., 2006a, b).

To characterize the expression of the DNMT genes within cancer cells, we systematically amplified and sequenced the DNMT cDNAs from cancer cell lines. Here, we show that expression of abnormally spliced DNMT3B transcripts is common in cancer cell lines and in primary leukemia cells. We have identified over 20 aberrant DNMT3B transcripts from cancer cells, many of which contain intron sequences. Although only a minority of the DNMT3B transcripts in cancer cells displays aberrant transcription, truncated DNMT3B proteins are detectable in the nuclear protein extracts of cancer cell lines by Western blot analysis. To test if expression of a truncated DNM T3B protein can affect DNA methylation and gene expression, we engineered 293 cells to express the most frequently identified aberrant DNMT3B transcript, DNMT3B7. We identified genes whose expression level changed in the presence of DNMT3B7 using microarray-based expression profiling and showed that some gene expression changes correlate with altered DNA methylation of corresponding CpG islands. These results suggest that expression of truncated DNMT3B proteins could affect DNA methylation patterns and gene expression in cancer cells.

Results

Cancer cells express aberrant DNMT3B transcripts

To examine potential mechanisms for the abnormal patterns of DNA methylation in cancer cells, we amplified the cDNAs encoding the three DNMT enzymes from several cancer cell lines using a high-fidelity polymerase. Although amplicons derived from the DNMT1 and DNMT3A cDNAs were wildtype in sequence (data not shown), polymerase chain reaction (PCR) amplification of DNMT3B cDNA from exon 9 to exon 13 produced the two expected amplification products (products A and B in Figure 1a) as well as an unexpected amplicon (product C in Figure 1a). Sequence analysis demonstrated that this novel transcript contained an aberrant splicing event from exon 9 to the 3′ end of intron 10, resulting in an insertion of 94 base pairs that is normally part of intron 10, located just 5′ to exon 11. We have named this transcript DNMT3B7 (Genbank Accession Number DQ321787; Figure 1b).

Figure 1.

Figure 1

Open in a new tab

Cancer cells express aberrant DNMT3B transcripts as demonstrated by reverse-transcription PCR (RT–PCR). (a) DNMT3B cDNA was amplified from exon 9 to exon 13 in several cancer cell lines. DNA sizing is shown at the left. Product A is derived from a normal DNMT3B transcript that lacks exon 10, either DNMT3B2 or DNMT3B3. Product B is derived from DNMT3B1, which contains exon 10. Product C is an abnormally migrating species, the DNMT3B7 transcript. Amplification of the GAPD cDNA served as a loading control, demonstrating equal amounts of input cDNA from each cDNA source (bottom panel). (b) Alternative splicing of the DNMT3B gene. The protein-encoding exons of the DNMT3B gene are indicated at the top of the figure. Six major splice forms of the DNMT3B gene have been described, DNMT3B1-6. The exons contained in a particular transcript are shown in solid black rectangles. Exons excluded through alternative splicing are indicated by rectangular outlines. Premature translational stop codons are indicated by stop signs. The structure of the 5′ ends of the DNMT3B4 and DNMT3B5 transcripts are not known (Robertson et al., 1999). The most widely expressed aberrant DNMT3B transcript in cancer cells, DNMT3B7, is indicated in the middle of the figure. The retained intron sequence is indicated with a solid gray rectangle. The ΔDNMT3B1-4 transcripts identified recently in non-small-cell lung cancer differ from each other with respect to alternative splicing of exons 7 and 10 (numbering according to exons in DNMT3B1-3 transcripts), and ΔDNMT3B5-6 differ from each other with respect to alternative splicing of exon 7, indicated by hatched bars (Wang et al., 2006a, b).

To determine whether aberrant DNMT3B7 transcripts were expressed in cancer cell lines of diverse origins, we expanded our screen to include 25 established cancer cell lines derived from both hematopoietic malignancies (11 cell lines) as well as solid tumors (14 cell lines), and30 primary acute leukemia samples (27 acute myeloid leukemia samples and three acute lymphoblastic leukemia samples) (Table 1 and Supplementary Figure 1). DNMT3B transcripts involving aberrant splicing events at the 5′ end of the gene could be detected in all of the samples tested, except in HepG2 and Alexander cells, derived from hepatocellular carcinomas. Alexander cells are known to express DNMT3B4, which encodes a catalytically inactive DNMT3B isoform (Saito et al., 2002). All of the primary leukemia samples expressed a transcript containing aberrant splicing between exons 9 and13 as originally observed, confirming that the expression of these novel mRNAs was not due to an artifact of in vitro culture. Notably, there was expression of at least one of the three wild-type DNMT3B transcripts in all of the tumor cell line-derived cDNAs as well as in all of the primary leukemia samples, in keeping with previous data that complete loss of DNMT3B activity is incompatible with viability (Li et al., 1992; Okano et al., 1999).

Table 1.

Expression of DNMT3B transcripts in cancer cell lines and primary acute leukemias

Tumor type Cell type Transcripts encoding catalytically active DNMT3B proteins Transcripts encoding catalytically inactive DNMT3B proteins
DNMT3B1 DNMT3B2, 3, 6 DNMT3B7, 12, 13 Othera
Myeloid leukemia UoC-M1 X X X
KG-1 X X
K562 X X
HL60 X X
U937 X X
Lymphoid leukemia Jurkat X X
CEM X X
REH X X
SUPB13 X X
Raji X X X
Follicular lymphoma FL18 X X
Breast cancer HCC1937 X X
MCF-7 X X X
MDA-MB-231 X X
SK-BR-3 X X
Cervical cancer HeLa X X X X
Mesothelioma MSTO 211 H X X X
NCI H28 X X X X
Head and neck cancer PCI13 X X X
Small-cell-lung cancer H82 X X
Colorectal cancer LoVo X X X X
Gastric cancer Kato-III X X X X
Hepatocellular carcinoma Alexander X X X
HepG2 X
Glioma Hs 683 X X
Primary myeloid leukemias With normal cytogenetics
N1-1 X X X
N3-1 X X
N4-1 X X X
N5-1 X X
N6-1 X X
N7-1 X X
J1 X X
J3 X X
J4 X X
J5 X X
With inv(16)
I2-1 X X
I3-1 X X X
I4-1 X X
I5-1 X X
I6-1 X X
With -5/del(5q)
D1-1 X X X
D2-2 X X
D3-1 X X
D4-1 X X
D5-1 X X
D6-2 X X
With -7/del(7q)
7-1 X X
7-4 X X X
7-5 X X
7-6 X X
7-7- X X
With del(13q)
J7 X X
Primary lymphoblastic leukemias Complex cytogenetics
J2 X X X
With del(20q)
J6 X X
With t(4;11)
J8 X X
Open in a new tab

Expression of a particular transcript within each sample is indicated by an X. The PCR conditions used to analyse these samples could not distinguish among the DNMT3B2, DNMT3B3, and DNMT 3B6 transcripts. The DNMT3B7, -12 and-13 transcripts are also indistinguishable by this analysis.

a

Other: DNMT3B4, -5, -8–11 or 14–30 (see Supplementary Figure 1).

Most of the aberrantly spliced DNMT3B transcripts contain sequences that are normally intronic and lack various exons, and all of them encode truncated DNMT3B proteins containing novel amino acids but lacking the catalytic C terminus. Supplementary Table 1 lists the properties of each aberrant transcript. We noted alternative splicing of several 5′ exons, including exon 5 (Xu et al., 1999). The 5′ half of each transcript could be amplified using a primer derived from exon 1A (data not shown), suggesting that the promoter originally identified for DNMT3B is used to generate the aberrant transcripts (Yanagisawa et al., 2002).

To quantitate the levels of aberrant DNMT3B transcripts in cancer cells precisely, we performed quantitative reverse transcription PCR (QRT–PCR) of DNMT3B transcripts. Two assays were designed: the first assay assessed the levels of aberrant DNMT3B transcripts containing intron 10 sequences by placing the forward primer in exon 9, and the reverse primer and Taqman probe within the retained intron sequence. The second assay assessed total DNMT3B transcript levels, by placing the forward primer in exon 12, and the reverse primer and Taqm an probe within exon 13, because neither exon 12 nor exon 13 is subject to alternative splicing. QRT–PCR of eight normal human tissues (Clontech, Mountain View, CA, USA, Stratagene, La Jolla, CA, USA) demonstrated no detectable aberrant DNMT3B transcripts, whereas 2–5% of DNMT3B transcripts in cancer cell lines derived from both solid and hematopoietic tumors contained intron 10 sequences (Table 2).

Table 2.

Expression of DNMT3B transcripts containing intron 10 sequence

Normal cells % Cancer cells % s.d. Transfected cells % s.d.
Normal peripheral blood U MCF7 3.5 0.03 293 cells U
Normal liver U MDA-MB-231 4.8 0.03 Vector transfected 293 Cells U
Normal breast U K562 3.0 0.06 293-DNMT3B7 line 1 42.7 0.50
Normal cervix U Jurkat 3.7 0.07 293-DNMT3B7 line 2 54.2 0.50
Normal bone marrow U H526 2.0 0.02
Normal kidney U
Normal lung U
Normal fibroblasts U
Open in a new tab

Abbreviations: U, undetectable; s.d., standard deviation. Percentage (%), amount of intron 10 containing DNMT3B transcripts relative to total DNMT3B transcripts, determined by quantitative real-time reverse transcription PCR. The assay for intron 10-containing transcripts accurately measures ⩾180 molecules, and the assay for total DNMT3B transcripts accurately measures ⩾200 molecules.

Cancer cells express truncated DNMT3B proteins

Although aberrant transcripts represent a minority of DNMT3B transcripts, we next tested whet her truncated DNMT3B proteins were detectable by Western blotting in protein extracts from cancer cells. All of the aberrant DNMT3B transcripts contain premature stop codons and are predicted to produce truncated DNMT3B proteins lacking the two strongest nuclear localization signals. Therefore, we tested both cytoplasmic and nuclear protein extracts from SK-BR-3 cells (a breast cancer cell line) and HeLa cells (a cervical cancer cell line). The full-length DNMT3B (96 kDa) as well as the truncated DNMT 3B7 (40 kDa) protein were observed in the nuclear protein fractions by Western blot analysis using an N-terminal-specific anti-DNMT3B antibody (Figure 2a, upper panel). The signals were competed away with the respective antigenic peptides for each of the antibodies tested (Figure 2a, middle panel), confirming that the observed signal s were DNMT3B-derived species. Similar results were obtained with two other anti-DNMT3B antibodies as well as nuclear protein extracts from MDA-MB-231 andK562 cells (data not shown).

Figure 2.

Figure 2

Open in a new tab

Identification of truncated DNMT3B proteins by Western blotting. (a) Identification of truncated DNMT3B proteins in extracts from cancer cell lines. Upper panel, cytosolic and nuclear extracts (60 μg each) from SK-BR-3 and HeLa were probed for DNMT3B by Western blotting. The positions of full-length DNMT3B and truncated DNMT3B7 are indicated to the right. The positions of the molecular weight markers are given at the left. C, cytoplasmic extract; N, nuclear extract. Middle panel, parallel blot to that shown in the upper panel probed with the same antibody plus the antigenic peptide. Bottom panel, equal loading of cytosolic extracts and integrity of nuclear extracts as demonstrated by Western blotting using GAPDH. (b) Expression of DNMT3B7 in stable 293 cell lines. Upper panel, cytosolic and nuclear extracts (40–50 μg each) from vector-transfected or DNMT3B7-transfected293 cells demonstrate expression of DNMT3B7 in the nuclear fraction. Middle panel, parallel blot to that shown in the top panel probed with the same antibody plus the antigenic peptide. Bottom panel, equal loading of cytosolic extracts and integrity of nuclear extracts as demonstrated by Western blotting using GAPDH.

To determine if low levels of a truncated DNMT3B protein could alter DNA methylation levels, we isolated 293 cells (human embryonic kidney cells) that stably express DNMT3B7 (Table 2 and Figure 2b, upper panel). Although QRT–PCR indicated that 40–50% of DNMT3B transcripts within transfected cells encoded DNMT3B7 (Table 2), Western blot analysis demonstrated that the truncated protein represented a minority of DNMT3B protein within transfected cell s (Figure 2b, upper panel), as we observed in cancer cells. Therefore, DNMT3B7-expressing 293 cells could serve as a reasonable model of the relative amounts of full-length versus truncatedDNMT3B proteins found in cancer cells. We characterized the phenotype of two independently derived clones expressing DNMT3B7 compared to vector-transfected cells.

DNMT3B7-expressing 293 cells demonstrate gene expression changes that correspond with altered DNA methylation within some CpG islands

We used microarray analysis to compare the gene expression changes found in DNMT3B7-expressing cells versus vector-transfected cell s. Three independently isolated RNA samples from DNMT3B7-expressing line 1 and line 2 cells were used as probes hybridized to Affymetrix Gene Chip Human Genome U133 Plus 2.0 oligonucleotide arrays. These samples were compared to four independently isolated RNA samples from vector-transfected cells. We searched for genes whose expression levels differed statistically between the two DNMT3B7-expressing lines and the vector-transfected cells, but not between the two DNMT3B7-expressing lines themselves. Fifty-one genes fulfilled these criteria (Figure 3): 27 genes were under expressed in the DNMT3B7-expressing cells, and24 genes were overexpressed. Interestingly, more than half of the genes whose expression changed are located on chromosomes 1, 16 and X, and75% of the overexpressed genes are located on those chromosomes.

Figure 3.

Figure 3

Open in a new tab

Gene expression changes in DNMT3B7-expressing 293 cells. A heat map shows the 51 genes whose expression changed with DNMT3B7 expression (red, overexpression; blue, under expression). Four samples of vector-transfected cells were compared to three samples each of DNMT3B7-expressing cells, and the average-fold change for each probe is listed. Gene names are in bold for the genes whose expression changes were validated by semiquantitative RT–PCR (data not shown). The chromosomal locus of each gene is given at the right, and the genes located on chromosomes 1, 9 and16, and the X chromosome are indicated by highlighting. Some genes (e.g., PRKCB1 and MID1) are listed more than once in the figure, because the microarray contained multiple probes for these genes.

We reasoned that at least some of the gene expression changes that we observed were likely to involve changes in DNA methylation of corresponding CpG islands. Genes whose expression decreased with DNMT3B7 expression would be expected to have increased CpG island/promoter methylation, whereas genes whose expression increased with DNMT3B7 expression would be expected to have less DNA methylation of their CpG islands/promoters. We isolated genomic DNA from transfected cell s and treated it with sodium bisulfite, which chemically converts unmethylated cytosine to uracil, but does not alter methylated cytosine. We analysed the DNA methylation state of particular CpG islands by PCR amplification and sequencing in vector-transfected versus DNMT3B7-expressing 293 cells. We observed changes in the DNA methylation state of the CpG islands/promoters of four genes whose expression was altered in DNMT3B7-expressing cells, in parallel to what is observed in cancer cells (Figure 4).

Figure 4.

Figure 4

Open in a new tab

293 cells overexpressing DNMT3B7 demonstrate gene expression changes that correspond with altered DNA methylation within some CpG islands as determined by sodium bisulfite analysis. Methylated CpG dinucleotides are represented by filled-in black circles, and unmethylated CpG dinucleotides are represented by open circles. Each numbered row represents an individual clone, and the CpG dinucleotide number is given across the top of each section. The number of identical clones is given in parentheses after a representative row. (a) Analysis of the methylation state of 18 individual CpG dinucleotides from the portion of the CDH1 CpG island that is located just 5′ to the gene’s transcriptional start. Hypermethylation of particular CpG dinucleotides in both line 1 and line 2 were statistically significant and are indicated by daggers. (b) Analysis of the methylation state of 12 individual CpG dinucleotides from the portion of the MAGEA3 CpG island that is located overlapping with the gene’s transcriptional start. (c) Analysis of the methylation state of two portions of the PLP2 CpG island. At the left, the figure shows the methylation state of 19 individual CpG dinucleotides from a part of the CpG island located just 5′ to the gene’s transcriptional start, and at the right, the figure indicates the methylation state of 24 additional CpG dinucleotides located just 3′ to the translational start. Hypomethylation of one particular CpG residue in the 3′ portion of the CpG island is indicated with an asterisk. (d) Analysis of the methylation state of seven individual CpG dinucleotides from the portion of the SH2D1A CpG island that is located within exon 2.

The E-cadherin (CDH1) gene is hypermethylated and transcriptionally repressed in gastric and breast cancers, myeloid malignancies, and other tumors (Cowin et al., 2005; Aggerholm et al., 2006; Chan, 2006). We observed hypermethylation of the CDH1 CpG island (Figure 4a), corresponding to a 2.19-fold decrease in gene expression (Figure 3). One CpG dinucleotide (No. 2, Figure 4a) is contained within the known Sp1-binding site located closest to the gene’s transcriptional start.

The X-linked MAGEA3 gene is hypomethylated and overexpressed in melanoma (Sigalotti et al., 2002), and the methylation status of the regulatory regions of SH2D1A, the X-linked lymphoproliferative disease gene, correlates with tissue-specific expression (Parolini et al., 2003). Hypomethylation of the CpG islands associated with the MAGEA3 and SH2D1A genes also correlated with increased expression, 2.54- and -3.05- fold, respectively (Figures 4b, d and 3). In addition, we observed hypo methylation of the CpG island of PLP2, a gene located at Xp11.2, which exhibited 2.85-fold increased expression in DNMT3B7-expressing cells (Figures 4c and 3). Interestingly, most of the DNA methylation changes were located in the part of the CpG island just 5′ to the transcriptional start site of the gene, but there was also some demethylation that extended well past the gene’s translational start. We confirmed that increased gene expression from the X chromosome occurred from the active X chromosome, rather than from re-activation of gene expression from the inactive X chromosomes, using an informative SNP in the X-linked MID1 gene in 293 cells (a T/C allele at position 10247528; SNP rs16986145) (data not shown).

Discussion

In this paper, we demonstrate that cancer cells express numerous aberrant splice variants of the DNMT3B gene, all of which are predicted to encode truncated proteins lacking the catalytic C-terminus. Western blotting of cancer cell extracts demonstrates that truncatedDNMT3B proteins are present in nuclear protein extracts, despite low levels of aberrant DNMT3B transcripts. When expressed in 293 cells, DNMT3B7, the most frequently observed truncated protein, causes altered gene expression with corresponding changes in the DNA methylation states of several CpG islands. These findings suggest that truncated DNMT3B proteins could influence the DNA methylation state of cancer cells.

Our extensive analysis of 25 cancer cell lines indicates that aberrant DNMT3B transcription is extremely widespread in cancer. The only tumor type in which we did not observe novel DNMT3B transcripts was hepatocellular carcinoma. Interestingly, Alexander cells, derived from hepatocellular carcinomas, are known to express DNMT3B4, which encodes a catalytically inactive DNMT3B isoform (Saito et al., 2002). Furthermore, recently described transcripts from non-small-cell lung cancers, ΔDNMT3B5-7, are also predicted to encode truncated DNMT3B proteins lacking the catalytic domain (Wang et al., 2006a, b). Therefore, it appears that many cancer cells express full-length DNMT3B catalytically active proteins as well as truncated DNMT 3B proteins that are predicted to be catalytically inactive.

We used microarray analysis to indicate which genes showed altered transcription in DNMT3B7-expressing cells, and some of these changes in gene expression correlated with DNA methylation of corresponding CpG islands. Almost all of the changes in DNA methylation levels were stronger in the 293 cells that expressed higher DNMT3B7 levels (line 2), suggesting that subtle changes in levels of catalytically inactive DNMT3B proteins could have significant effects in cells over the many generations of cell divisions that occur during tumor formation and growth.

Our studies show that the most common of the truncated proteins, DNMT3B7, is concentrated in the nucleus, suggesting that the nucleus is the major site of its activity. DNMT3B7 may localize to the nucleus via the retention of one weak nuclear localization signal or via binding to a protein that could shuttle it into the nucleus. We can envision several models by which DNMT3B7 could affect DNA methylation, and consequently, gene expression:

  1. DNMT3B7 may interfere with the normal DNA methylation machinery by binding one or more of the known DNMT3B binding partners. Several proteins have been shown to bind DNMT 3B, including SUMO-1 and UBC-9, two components of the sumoylation pathway (Kang et al., 2001); h-CAP-C and hCA P-E, two components of the condensin complex; KIF4A, a chromokinesin homolog; hSNF2H, an ATP-dependent chromatin remodeling enzyme; HDAC1, a histone deacetylase; and SIN3A, a transcriptional corepressor (Geiman et al., 2004a). DNMT3B is also known to interact with HDAC2 (Geiman et al., 2004b) andDNMT1 (Kim et al., 2002), although the precise binding regions have not been defined. In addition, the PHD domain of murine Dnmt3b, which is located in the N-terminus, mediates binding to Suv39h1 andHP1 proteins, both of which are components of the histone methylation machinery (Geiman et al., 2004b).

  2. DNMT3B7 may bind DNA directly and affect the activity of active DNMTs.

  3. DNMT3B7 may affect DNA methylation as outlined above, which could in turn lead to alterations in histone modifications. Mice lacking histone H1 showed alterations in DNA methylation levels and consequent gene expression changes in relatively few genes, often in imprinted genes or genes located on the X chromosome (Fan et al., 2005). In addition, histone H3 and H4 acetylation levels increased after patients received 5-azacytidine, a global hypomethylating agent (Gore et al., 2006).

Interestingly, the genes that were overexpressed in DNMT3B7-expressing cells were over-represented on chromosomes 1, 16 and X. Cells from patients with ICF Syndrome contain dramatically hypomethylated repetitive DNA sequences of the satellite 2 repeats concentrated at the pericentromeric regions of chromosomes 1 and16 and of the satellite 3 repeats found near the centromere of chromosome 9 (Ehrlich, 2003) as well as hypomethylation of the LINE-1 elements located on the inactive X chromosome (Hansen, 2003). We did not observe significant hypomethylation of repetitive elements in our DNMT3B7-expressing 293 cells, possibly because they have relatively under methylated repetitive sequences at baseline (data not shown).

Aberrant processing of mRNAs in tumor cells has been described for the MDM2, RasGRP4, SLP-65, TLE1, TLE4, MTA1 and CD44 genes and is emerging as an important characteristic of tumor cells (Bartel et al., 2002; Kumar et al., 2002; Reuther et al., 2002; Yang et al., 2002; Jumaa et al., 2003; Venables, 2004; Watermann et al., 2006) and is observed in other diseases, such as myotonic dystrophy (Charlet-B et al., 2002). In aberrant splicing, portions of exons, portions of introns or both are retained within transcripts that fail to be purged by the cellular pathways designed to scavenge abnormal mRNAs, such as nonsense-mediated decay (Culbertson, 1999; Wilkinson and Shyu, 2002). Often, the aberrant mRNAs produce truncated proteins that in some cases allow them to function in a dominant-negative manner. The aberrant transcripts do not represent intermediates of the normal splicing reaction, because they are found in the cytoplasm and are translated. The cellular abnormalities that lead to aberrant mRNA processing and possible defects in nonsense-mediated decay are not clear, but represent an exciting opportunity for future studies.

In summary, our paper is the first to demonstrate that a wide variety of cancer cells exhibit aberrant splicing of the DNMT3B gene, producing transcripts that encode truncated protein s lacking the C-terminal catalytic domain. Our data support a model in which aberrations in mRNA splicing in cancer cells give rise to truncated DNMT3B proteins, resulting in changes in DNA methylation patterns and gene expression. These results suggest that the abnormal patterns of DNA methylation present in nearly all cancer cells may be regulated in part by the presence of catalytically inactive DNMT3B proteins. Further work in which levels of aberrant DNMT3B transcripts and/or truncated DNMT3B proteins are varied within cancer cells may help to confirm this model. Future experiments could also address whether truncated DNMT3B proteins affect epigenetic modifications in addition to DNA methylation, such as histone alterations. Overall, our work supports growing evidence that aberrant gene splicing in cancer cells affects cellular phenotype.

Materials and methods

Cell lines and primary acute leukemia cell samples

Hematopoietic cell lines were grown in RPMI1640, 10mM HEPES, 10% fetal bovine serum and100 U/ml penicillin/100 μg/ml streptomycin. Adherent cell lines were grown in the recommended media. Viable primary acute leukemia cells were cryopreserved in liquid nitrogen after patients gave informed consent to participate in an IRB-approved research protocol.

Reverse transcription, PCR amplification and sequencing

Viable primary leukemia cells were isolated by Ficoll density centrifugation. Total RNA was made using STAT-60 (Tel-Test Inc., Friendswood, TX, USA) or Trizol (Invitrogen, Carlsbad, CA, USA), and reverse transcription was performed using SuperScriptII (Invitrogen). PCR amplifications were performed as described in the Supplementary Information.

Construction of DNMT3B7 expression plasmid

To construct the DNMT3B7 expression plasmid, the DNMT3B7 cDNA was amplified from Raji cells, ligated into pcDNA3.1+ (Invitrogen), and sequenced.

Stable transfections in cultured cells

Transfections were performed using 2 μg of the desired plasmid and Effect amine (Qiagen, Valencia, CA, USA). Stable transfectants were selected by adding 400 μg/ml G418 (Invitrogen) to the media 48 h after transfection and picked after reculturing for 3 weeks.

Western blotting

Protein extracts were made 48 h after transfection, separated by SDS–PAGE electrophoresis and transferred to nitrocellulose. Western blotting was performed as described in the Supplementary Information.

Sodium bisulfite treatment and PCR amplification

Genomic DNA was treated with sodium bisulfite (Clark et al., 1994), and PCR amplifications were performed as described in the Supplementary Information.

Microarray analysis

Cells from at least five plates were combined and used to collect total RNA using STAT-60 (Tel-Test Inc.), followed by RNeasy Mini column purification (Qiagen). In collaboration with the University of Chicago Functional Genomics Facility, each RNA sample was split into at least three samples, and microarray analysis was performed as described in the Supplementary Information.

Supplementary Material

supplementary

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

Acknowledgments

We thank Hongwei Zou and Wai Hui for technical assistance; Dr Gregory L Shipley of the Quantitative Genomics Core Laboratory at The University of Texas at Houston Health Science Center for developing and running the real-time PCR assays; Dr Xinmin Li and Jean Shi for assistance with microarray analysis; Dr Shang Lin for assistance with statistical analysis; and Drs Charles Rudin, Michael Nishimura, Suzanne Conzen, and Michael Thirman for gifts of cell lines. We thank Drs Stephen B Baylin, Kevin Shannon, Anthony Fernald, Yanwen Jiang, John Joslin and Dr Zhijian Qian for their contributions to the data presented and their thoughtful comments regarding this work. This work was funded by a Howard Hughes Postdoctoral Fellowship, the University of Chicago Section of Hematology/Oncology Fund-a-Fellow Program, a Cancer Research Foundation Young Investigator Award, an American Society of Clinical Oncology Young Investigator Award, an American Cancer Society Institutional Research Grant IRG-58-004-44, a Schweppe Foundation Career Development Award and The Kimmel Scholar Award(LA Godley), and by NIH grant CA 40046 (MM Le Beau).

References

  1. Aggerholm A, Holm MS, Guldberg P, Olesen LH, Hokland P. Promoter hypermethylation of p15INK4B, HIC1, CDH1, and ER is frequent in myelodysplastic syndrome and predicts poor prognosis in early-stage patients. Eur J Haematol. 2006;76:23–32. doi: 10.1111/j.1600-0609.2005.00559.x. [DOI] [PubMed] [Google Scholar]
  2. Bachman KE, Park BH, Rhee I, Rajagopalan H, Herman JG, Baylin SB, et al. Histone modifications and silencing prior to DNA methylation of a tumor suppressor gene. Cancer Cell. 2003;3:89–95. doi: 10.1016/s1535-6108(02)00234-9. [DOI] [PubMed] [Google Scholar]
  3. Bartel F, Taubert H, Harris LC. Alternative and aberrant splicing of MDM2 mRNA in human cancer. Cancer Cell. 2002;2:9–15. doi: 10.1016/s1535-6108(02)00091-0. [DOI] [PubMed] [Google Scholar]
  4. Beaulieu N, Morin S, Chute IC, Robert M-F, Nguyen H, MacLeod AR. An essential role for DNA methyltransferase DNMT3B in cancer cell survival. J Biol Chem. 2002;277:28176–28181. doi: 10.1074/jbc.M204734200. [DOI] [PubMed] [Google Scholar]
  5. Bestor TH. The DNA methyltransferases of mammals. Hum Molec Genet. 2000;9:2395–2402. doi: 10.1093/hmg/9.16.2395. [DOI] [PubMed] [Google Scholar]
  6. Chan AO. E-cadherin in gastric cancer. World J Gastroenterol. 2006;12:199–203. doi: 10.3748/wjg.v12.i2.199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Charlet-B N, Savkur RS, Singh G, Philips AV, Grice EA, Cooper TA. Loss of the muscle-specific chloride channel in type 1 myotonic dystrophy due to misregulated alternative splicing. Molecular Cell. 2002;10:45–53. doi: 10.1016/s1097-2765(02)00572-5. [DOI] [PubMed] [Google Scholar]
  8. Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 1994;22:2990–2997. doi: 10.1093/nar/22.15.2990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cowin P, Rowlands TM, Hatsell SJ. Cadherins and catenins in breast cancer. Curr Opin Cell Biol. 2005;17:499–508. doi: 10.1016/j.ceb.2005.08.014. [DOI] [PubMed] [Google Scholar]
  10. Culbertson MR. RNA surveillance: unforeseen consequences for gene expression, inherited genetic disorders and cancer. Trends Genet. 1999;15:74–80. doi: 10.1016/s0168-9525(98)01658-8. [DOI] [PubMed] [Google Scholar]
  11. Eden A, Gaudet F, Waghmare A, Jaenisch R. Chromosomal instability andtumors promoted by DNA hypomethylation. Science. 2003;300:455. doi: 10.1126/science.1083557. [DOI] [PubMed] [Google Scholar]
  12. Ehrlich M. The ICF syndrome, a DNA methyltransferase 3B deficiency and immunodeficiency disease. Clin Immunol. 2003;109:17–28. doi: 10.1016/s1521-6616(03)00201-8. [DOI] [PubMed] [Google Scholar]
  13. Ehrlich M, Woods CB, Yu MC, Dubeau L, Yang F, Campan M, et al. Quantitative analysis of associations between DNA hypermethylation, hypomethylation, and DNM T RNA levels in ovarian tumors. Oncogene. 2006;25:2636–2645. doi: 10.1038/sj.onc.1209145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fan Y, Nikitina T, Zhao J, Fleury TJ, Bhattacharyya R, Bouhassira EE, et al. Histone H1 depletion in mammals alters global chromatin structure but causes specific changes in gene regulation. Cell. 2005;123:1199–1212. doi: 10.1016/j.cell.2005.10.028. [DOI] [PubMed] [Google Scholar]
  15. Gaudet F, Hodgson JG, Eden A, Jackson-Grusby L, Dausman J, Gray JW, et al. Induction of tumors in mice by genomic hypomethylation. Science. 2003;300:489–492. doi: 10.1126/science.1083558. [DOI] [PubMed] [Google Scholar]
  16. Geiman TM, Sankpal UT, Robertson AK, Chen Y, Mazumdar M, Heale JT, et al. Isolation and characterization of a novel DNA methyltransferase complex linking DNMT3B with components of the mitotic chromosome condensation machinery. Nucleic Acids Res. 2004a;32:2716–2729. doi: 10.1093/nar/gkh589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Geiman TM, Sankpal UT, Robertson AK, Zhao Y, Robertson KD. DNMT3B interacts with hSNF2H chromatin remodeling enzyme, HDACs 1 and2, and components of the histone methylation system. Biochem Biophys Res Commun. 2004b;318:544–555. doi: 10.1016/j.bbrc.2004.04.058. [DOI] [PubMed] [Google Scholar]
  18. Gore SD, Baylin S, Sugar E, Carraway H, Miller CB, Carducci M, et al. Combined DNA methyltransferase and histone deacetylase inhibition in the treatment of myeloid neoplasms. Cancer Res. 2006;66:6361–6369. doi: 10.1158/0008-5472.CAN-06-0080. [DOI] [PubMed] [Google Scholar]
  19. Hansen RS. X inactivation-specific methylation of LINE-1 elements by DNMT3B: implications for the Lyon repeat hypothesis. Hum Mol Genet. 2003;12:2559–2567. doi: 10.1093/hmg/ddg268. [DOI] [PubMed] [Google Scholar]
  20. Hansen RS, Wijmenga C, Luo P, Stanek AM, Canfield TK, Weemaes CMR, et al. The DNMT3B DNA methyltransferase gene is mutated in the ICF immunodeficiency syndrome. Proc Natl Acad Sci USA. 1999;96:14412–14417. doi: 10.1073/pnas.96.25.14412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Jumaa H, Bossaller L, Portugal K, Storch B, Lotz M, Flemming A, et al. Deficiency of the adaptor SLP-65 in pre-B-cell lymphoblastic leukaemia. Nature. 2003;423:452–456. doi: 10.1038/nature01608. [DOI] [PubMed] [Google Scholar]
  22. Kanai Y, Ushijima S, Nakanishi Y, Sakamoto M, Hirohashi S. Mutation of the DNA methyltransferase (DNMT) 1 gene in human colorectal cancers. Cancer Lett. 2003;192:75–82. doi: 10.1016/s0304-3835(02)00689-4. [DOI] [PubMed] [Google Scholar]
  23. Kang ES, Park CW, Chung JH. Dnmt3b, de novo DNA methyltransferase, interacts with SUMO-1 and Ubc9 through its N-terminal region and is subject to modification by SUMO-1. Biochem & Biophys Res Comm. 2001;289:862–868. doi: 10.1006/bbrc.2001.6057. [DOI] [PubMed] [Google Scholar]
  24. Karpf AR, Matsui S. Genetic disruption of cytosine DNA methyltransferase enzymes induces chromosomal instability in human cancer cells. Cancer Res. 2005;65:8635–8639. doi: 10.1158/0008-5472.CAN-05-1961. [DOI] [PubMed] [Google Scholar]
  25. Kim G-D, Ni J, Kelesoglu N, Roberts RJ, Pradhan S. Co-operation and communication between the human maintenance and de novo DNA (cytosine-5) methyltransferases. EMBO J. 2002;21:4183–4195. doi: 10.1093/emboj/cdf401. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kumar R, Wang R-A, Mazumdar A, Talukder AH, Mandal M, Yang Z, et al. A naturally occurring MTA1 variant sequesters oestrogen receptor-alpha in the cytoplasm. Nature. 2002;418:654–657. doi: 10.1038/nature00889. [DOI] [PubMed] [Google Scholar]
  27. Li E. Chromatin modification and epigenetic reprogramming in mammalian development. Nat Rev Genet. 2002;3:662–673. doi: 10.1038/nrg887. [DOI] [PubMed] [Google Scholar]
  28. Li E, Bestor TH, Jaenisch R. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell. 1992;69:915–926. doi: 10.1016/0092-8674(92)90611-f. [DOI] [PubMed] [Google Scholar]
  29. Lin H, Yamada Y, Nguyen S, Linhart H, Jackson-Grusby L, Meissner A, et al. Suppression of intestinal neoplasia by deletion of Dnmt3b. Mol Cell Biol. 2006;26:2976–2983. doi: 10.1128/MCB.26.8.2976-2983.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Okano M, Bell DW, Haber DA, Li E. DNA methyltranferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell. 1999;99:247–257. doi: 10.1016/s0092-8674(00)81656-6. [DOI] [PubMed] [Google Scholar]
  31. Parolini O, Weinhausel A, Kagerbauer B, Sassmann J, Holter W, Gadner H, et al. Differential methylation pattern of the X-linked lymphoproliferative (XLP) disease gene SH2D1A correlates with the cell lineage-specific transcription. Immunogenetics. 2003;55:116–121. doi: 10.1007/s00251-003-0557-x. [DOI] [PubMed] [Google Scholar]
  32. Reuther GW, Lambert QT, Rebhun JF, Caligiuri MA, Quilliam LA, Der CJ. RasGRP4 is a novel ras activator isolated from acute myeloid leukemia. J Biol Chem. 2002;277:30508–30514. doi: 10.1074/jbc.M111330200. [DOI] [PubMed] [Google Scholar]
  33. Rhee I, Bachman KE, Park BH, Jair KW, Yen RW, Schuebel KE, et al. DNMT1 and DNMT3b cooperate to silence genes in human cancer cells. Nature. 2002;416:552–556. doi: 10.1038/416552a. [DOI] [PubMed] [Google Scholar]
  34. Rhee I, Jair K-W, Yen R-WC, Lengauer C, Herman JG, Kinzler KW, et al. CpG methylation is maintained in human cancer cells lacking DNMT1. Nature. 2000;404:1003–1007. doi: 10.1038/35010000. [DOI] [PubMed] [Google Scholar]
  35. Robert M-F, Morin S, Beaulieu N, Gauthier F, Chute IC, Barsalou A, et al. DNMT1 is required to maintain CpG methylation and aberrant gene silencing in human cancer cells. Nat Genet. 2003;33:61–65. doi: 10.1038/ng1068. [DOI] [PubMed] [Google Scholar]
  36. Robertson KD. DNA methylation and chromatin – unraveling the tangled web. Oncogene. 2002;21:5361–5379. doi: 10.1038/sj.onc.1205609. [DOI] [PubMed] [Google Scholar]
  37. Robertson KD. DNA methylation and human disease. Nat Rev Genet. 2005;6:597–610. doi: 10.1038/nrg1655. [DOI] [PubMed] [Google Scholar]
  38. Robertson KD, Uzvolgyi E, Liang G, Talmadge C, Sumegi J, Gonzales FA, et al. The human DNA methyltransferases (DNMTs) 1, 3a and3b: coordinate mRNA expression in normal tissues and overexpression in tumors. Nucl Acids Res. 1999;27:2291–2298. doi: 10.1093/nar/27.11.2291. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rountree MR, Bachman KE, Herman JG, Baylin SB. DNA methylation, chromatin inheritance, and cancer. Oncogene. 2001;20:3156–3165. doi: 10.1038/sj.onc.1204339. [DOI] [PubMed] [Google Scholar]
  40. Saito Y, Kanai Y, Sakamoto M, Saito H, Ishii H, Hirohashi S. Overexpression of a splice variant of DNA methyltransferase 3b, DNMT3b4, associated with DNA hypomethylation on pericentromeric satellite regions during human hepatocarcinogenesis. Proc Natl Acad Sci USA. 2002;99:10060–10065. doi: 10.1073/pnas.152121799. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Shirohzu H, Kubota T, Kumazawa A, Sado T, Chijiwa T, Inagaki K, et al. Three novel DNMT3B mutations in Japanese patients with ICF Syndrome. Am J Med Genet. 2002;112:31–37. doi: 10.1002/ajmg.10658. [DOI] [PubMed] [Google Scholar]
  42. Sigalotti L, Coral S, Nardi G, Spessotto A, Cortini E, Cattarossi I, et al. Promoter methylation controls the expression of MAGE2, 3 and4 genes in human cutaneous melanoma. J Immunother. 2002;25:16–26. doi: 10.1097/00002371-200201000-00002. [DOI] [PubMed] [Google Scholar]
  43. Ueda Y, Okano M, Williams C, Chen T, Georgopoulos K, Li E. Roles for Dnmt3b in mammalian development: a mouse model for the ICF syndrome. Development. 2006;133:1183–1192. doi: 10.1242/dev.02293. [DOI] [PubMed] [Google Scholar]
  44. Venables JP. Aberrant and alternative splicing in cancer. Cancer Res. 2004;64:7647–7654. doi: 10.1158/0008-5472.CAN-04-1910. [DOI] [PubMed] [Google Scholar]
  45. Wang J, Walsh G, Liu DD, Lee JJ, Mao L. Expression of ΔDNMT3B variants and its association with promoter methylation of p16 and RASSF1A in primary non-small cell lung cancer. Cancer Res. 2006a;66:8361–8366. doi: 10.1158/0008-5472.CAN-06-2031. [DOI] [PubMed] [Google Scholar]
  46. Wang L, Wang J, Sun S, Rodriguez M, Yue P, Jang SJ, et al. A novel DNMT3B subfamily, ΔDNMT3B, is the predominant form of DNMT3B in non-small cell lung cancer. Int J Oncol. 2006b;29:201–207. [PubMed] [Google Scholar]
  47. Watermann DO, Tang Y, Zur Hausen A, Jager M, Stamm S, Stickeler E. Splicing factor Tra2-beta1 is specifically induced in breast cancer and regulates alternative splicing of the CD44 gene. Cancer Res. 2006;66:4774–4780. doi: 10.1158/0008-5472.CAN-04-3294. [DOI] [PubMed] [Google Scholar]
  48. Wilkinson MF, Shyu A-B. RNA surveillance by nuclear scanning? Nat Cell Biol. 2002;4:144–147. doi: 10.1038/ncb0602-e144. [DOI] [PubMed] [Google Scholar]
  49. Xu G-L, Bestor TH, Bourc’his D, Hsieh C-L, Tommerup N, Bugge M, et al. Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature. 1999;402:187–191. doi: 10.1038/46052. [DOI] [PubMed] [Google Scholar]
  50. Yamada Y, Jackson-Grusby L, Linhart H, Meissner A, Eden A, Lin H, et al. Opposing effects of DNA hypomethylation on intestinal and liver carcinogenesis. Proc Natl Acad Sci USA. 2005;102:13580–13585. doi: 10.1073/pnas.0506612102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Yanagisawa Y, Ito E, Yuasa Y, Maruyama K. The human DNA methyltransferases DNMT3A and DNMT3B have two types of promoters with different CpG contents. Biochim Biophys Acta. 2002;1577:457–465. doi: 10.1016/s0167-4781(02)00482-7. [DOI] [PubMed] [Google Scholar]
  52. Yang Y, Li L, Wong GW, Krilis SA, Madhusudhan MS, Sali A, et al. RasGRP4, a new mast cell-restricted ras guanine nucleotide-releasing protein with calcium-and diacylglycerol-binding motifs. J Biol Chem. 2002;277:25756–25774. doi: 10.1074/jbc.M202575200. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementary

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

NIHMS48740-supplement-supplementary.pdf (293.3KB, pdf)

RESOURCES