HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) All Versions of this Article: 16/10/1319 01-0827fjev1    most recent Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by ZARDO, G. Articles by CAIAFA, P. Search for Related Content PubMed PubMed Citation Articles by ZARDO, G. Articles by CAIAFA, P.

Inhibition of poly(ADP-ribosyl)ation induces DNA hypermethylation: a possible molecular mechanism¹

GIUSEPPE ZARDO^*,2, ANNA REALE^*, CLAUDIO PASSANANTI, SRIHARSA PRADHAN, SERENA BUONTEMPO^*, GIOVANNA DE MATTEIS^*, ROGER L.P. ADAMS and PAOLA CAIAFA^*3

Departments of
^* Cellular Biotechnologies and Haematology, University of Rome ‘La Sapienza’, Rome, Italy;
CNR, ITBM, Rome, Italy;
New England Biolabs, Beverly, Massachusetts, USA; and
Institute of Biomedical and Life Sciences, University of Glasgow, UK

³Correspondence: Dipartimento di Biotecnologie Cellulari ed Ematologia, sezione di Biochimica Clinica, Università ‘La Sapienza’, V. le Regina Elena 324, 00161 Roma, Italia. E-mail: caiafa{at}bce.med.uniroma1.it

SPECIFIC AIMS

Nothing is known about the mechanism(s) whereby the methylation pattern of CpG island regions are protected from methylation during replication and in chromatin in normal cells while they become methylated during tumorigenesis. The aim of the present study is to provide clues to understanding the molecular mechanism by which a block of poly(ADP-ribosyl)ation allows new methyl groups to be introduced onto DNA, particularly onto CpG island regions.

PRINCIPAL FINDINGS

1. Effect of inhibition of poly(ADP-ribose) polymerases (PARPs) on Dnmt1 and p21 gene expression
Experiments have been carried out on L929 mouse fibroblast cells at different cell cycle phases to determine whether inhibition of PARPs modulates expression of the Dnmt1 gene. PARPs were inhibited by treating cells with 2 mM 3-aminobenzamide (3-ABA) for 24 h. Northern blot and RT-PCR analysis show that inhibition of poly(ADP-ribose) polymerase(s) increases the expression of Dnmt1. Densitometric analysis confirms that after treatment with 3-ABA, the amount of Dnmt1 expressed in fibroblasts at the G1/S border is the same as or greater than that normally expressed in S phase control cells; the level of expression increases up to fourfold (Fig. 1 a, c). The influence of poly(ADP-ribosyl)ation on Dnmt1 gene expression is gene specific since parallel experiments have shown that treating cells with 3-ABA does not influence the expression of ß-actin or p21 and c-fos genes. Results from untreated control cells agree with those in the literature showing that during the G0 phase, p21 gene expression is inversely correlated with that of the Dnmt1 gene.

View larger version (50K): [in this window] [in a new window]   Figure 1. Effect of inhibition of poly(ADP-ribosyl)ation on Dnmt1, p21, and c-fos gene expression. Experiments were carried out on L929 mouse fibroblast cells synchronized in G0, G1/S border, and late S phases. Cells growing exponentially were divided into subcultures that were treated in different ways so as to have cells synchronized at different stages of the cell cycle. To obtain inhibition of PARPs, cells were treated with 2 mM 3-ABA for different periods. Untreated cells were used as control (-). Cells were synchronized in G0 phase by starvation for 36 h in GMEM with the addition of 0.5% newborn calf serum and without glutamine. G0 (+) corresponds to cells treated with 3-ABA during the 36 h of starvation. Cells were synchronized at the G1/S border by replacing the starvation medium with the normal growing medium for 9 h and incubating with 3 mM hydroxyurea for 15 h. G1/S (+) and G1/S (++) correspond to cells treated with 3-ABA for 24 h (after replacing the starvation medium) and for 60 h (during and after replacing the starvation medium). Cells were synchronized at late S phase starting from cells synchronized in G0 phase. The starvation medium was replaced with normal medium and cells were cultivated for another 16 h. Late S (+) and late S (++) correspond to cells treated with 3-ABA respectively for 22 h (6 h before replacing the starvation medium plus 16 h in normal medium growth) and for 52 h (during and after replacing the starvation medium). Cytometric analysis was used to monitor different cell cycle phases. a) RT-PCR analyses of Dnmt1, p21, c-fos were performed using ß-actin as a control for mRNA levels. Total cellular RNA was quantified using a spectrophotometer and 20 ng of total RNA was used (final volume of 25 µl) for RT-PCR analyses. Every RT-PCR was performed using One-Step RT-PCR system according to the following experimental scheme: 30 min at 50°C; 2.5 min at 90°C (experimental conditions for reverse transcriptase reaction step); 40 cycles of 1 min at 94°C, 50 s at the annealing temperature that is specific for the DNA fragment being amplified, 1 min at 70°C; and finally 6 min at 70°C. Dnmt1 (acc. no: X14805): annealing temperature: 62°C, amplified fragment size: 467 bp, primers: 5'-GTGAAACGCCCAAAGAAGG-3' sense/5'-TTCCCTTTGTTCCCAGGGCT-3' antisense; p21 (acc. no: U24173): annealing temperature: 55°C, amplified fragment size: 640 bp, primers: 5'-GTCAGAGTCTAGGGGAATTG-3' sense/5'-TAAGACACACAGAGTGAGGG-3' antisense; ß-actin (acc. no: AW212307): annealing temperature: 55°C, amplified fragment size: 435 bp, primers: 5'GGCATAGAGGTCTTTACGG-3' sense/5'-CACAGGCATTGTATGGACTC-3' antisense; c-fos (acc. no. 50399),: annealing temperature: 52°C, amplified fragment size: 319 bp primers: 5'-GGTCCTTTTCTTCTATAG-3' sense/5'GTTTTTCCTTCTCTTTCAG-3' antisense. Number of RT-PCR cycles to perform was chosen according to the minimum number of cycles necessary to obtain a quantifiable amplification of mRNA samples obtained from untreated control cells in G1/S border. b) immunoblot showing the amount of DNMT1, p21, and PCNA present in 40 µg of nuclear lysates from cells synchronized at the G1/S border or in late S phase. Blots were developed with purified rabbit polyclonal antibodies against the amino-terminal domain of DNMT1, anti-p21 (F-5) and anti-PCNA (PC10) monoclonal antibodies and detected using ECL Chemiluminescent detection system. Sp1 protein was used as internal control and quantified by specific antibodies (1C6). Proteins detected by immunoblot assay were quantified by densitometric scanning of the bands. Control cells (-) and cells treated with 3-ABA for 60 h (cells synchronized at the G1/S border) or 52 h (cells synchronized in late S phase) as indicated on top (++). c) Dnmt1 quantitative densitometric analysis of RT-PCR: data represent the average of 2 independent experiments.

2. Effect of inhibition of PARPs on DNMT1, PCNA, and p21 cellular levels
The decision to study the cellular level of p21 and proliferating cell nuclear antigen (PCNA) proteins was influenced by reports that DNMT1 and p21 compete for the same binding site on PCNA during DNA replication. Using rabbit polyclonal antibodies against the amino-terminal domain of DNMT1, we verified that after treatment of cells with 3-ABA, increased expression of the Dnmt1 gene was associated with twice the amount of enzyme present in cells whereas the quantity of PCNA and p21 was not affected (Fig. 1b ).

3. Levels of PCNA and p21 that coimmunoprecipitate with DNMT1
The focal point could be the competition between DNMT1 and p21 for the common binding domain on PCNA, as the DNMT1-PCNA complex can be dissociated in vitro by competition with p21 or its specific peptide fragments. These data led us to carry out coimmunoprecipitation experiments with anti-DNMT1 antibodies. The presence of PCNA and p21 in DNMT1 immunoprecipitates was analyzed using anti-PCNA and anti-p21 antibodies (Fig. 2 ).

View larger version (33K): [in this window] [in a new window]   Figure 2. Levels of PCNA and p21 that coimmunoprecipitate with DNMT1: results obtained from cells synchronized at the G1/S border. Immunoblot of PCNA and p21 coprecipitated from 0.5 mg of cellular lysates by using anti-amino-terminal domain of DNMT1 rabbit polyclonal antibodies cross-linked to Dynabeads protein A. SDS/PAGE and transfer of proteins to nitrocellulose membranes were performed according to standard procedures. Immunostaining of PCNA and p21 coprecipitated with DNMT1 was carried out using the same antibodies reported to detect these proteins in nuclear lysates. Immunoprecipitation experiments were performed three times, with similar results.

The amount of DNMT1 present in G0 phase cells even after treatment with 3-ABA was too low to allow characterization of immunoprecipitates. Densitometric analysis of immunoprecipitates from cells at the G1/S border has shown that inhibition of PARPs increases by 2.7 ± 0.5-fold the amount of PCNA that coimmunoprecipitates with DNMT1 whereas the amount of immunoprecipitated p21 decreases by half (Fig. 2) . These data indicate that the inhibition of PARPs, by increasing DNMT1 expression favors the formation of a PCNA-DNMT1 complex as soon as in the G1/S border. PCNA clamps the DNA helix in a circular trimeric complex, where each of the three units is capable of binding to either DNMT1 or p21. Here we have assumed that after treatment of cells with 3-ABA, the ratio of p21-PCNA-DNMT1 changes inside the trimer and that some p21 can remain bound to PCNA despite the increased amount of bound DNMT1. The high level of DNMT1 gets the better of p21 during the competition for the binding domain of PCNA.

4. Effect of inhibition of poly(ADP-ribosyl)ation on c-fos gene expression
Experiments were performed to establish whether poly(ADP-ribosyl)ation modulates DNMT1 expression through another chain of events capable of activating the expression of DNMT1. The proto-oncogene c-fos was chosen because in c-fos transfected cells, the expression of c-fos is increased 25-fold and the amount of DNMT1 present is 10-fold greater than in control cells, resulting in a parallel 20% increase in the number of methyl groups present in the DNA. However, treatment of cells with 3-ABA does not affect c-fos proto-oncogene expression (Fig. 1a ) other than the fact that the increased expression and level of DNMT1 is due to this protein.

CONCLUSIONS AND SIGNIFICANCE

Methods have already been described to induce DNA hypomethylation, such as treatment of cells with 5-azacytidine, but treatment of cells with 3-ABA is the first treatment shown to induce in vivo DNA hypermethylation.

Our previous data show that after inhibition of poly(ADP-ribose) polymerases, DNMTs are able to methylate cytosines on DNA. A block of poly(ADP-ribosyl)ation introduces an anomalous hypermethylated pattern onto genomic DNA; further experiments show that this inhibition changes the methylation pattern of a plasmid transfected in its unmethylated form. As for the CpG islands, inhibition of PARPs leads to disappearance of ‘HpaII tiny fragments’ after digestion of genomic DNA with methylation-dependent HpaII restriction enzyme and, at least for the Htf9 promoter region, allows new methyl groups to position themselves on DNA. This is of interest considering that an anomalous DNA methylation of CpG islands induces the silencing of correlated genes and the mechanism involved in protecting these DNA regions against methylation is unknown. To provide an explanation of how the inhibition of poly(ADP-ribosyl)ation leads to DNA hypermethylation, we performed experiments on cells at different cell cycle phases. Our results show that competitive inhibition of PARPs via treatment of cells with 3-ABA increases the mRNA (four times) and protein levels (once) of the major maintenance DNA methyltransferase (DNMT1) vs. control cells. This favors expression of the enzyme at the G1/S border to such an extent that the level becomes similar or higher than in S phase control cells, the phase in which DNMT1 is typically expressed. It has recently been shown that antisense constructs, which down-regulate cellular DNMT1 levels, induce a rapid and corresponding increase in the level of the cell cycle regulator p21. Considering that inhibition of PARPs increases the expression of DNMT1 without affecting that of p21 and that these two proteins compete for the same binding site on PCNA, the results presented here provide an explanation as to how the inhibition of poly(ADP-ribosyl)ation leads to DNA hypermethylation. Although it is well known that DNMT1 is primarily expressed in S phase in normal cells, the increased expression of this enzyme after PARP inhibition induces an early formation of the DNMT1-PCNA complex at the G1/S border. The precocious presence of a DNMT1-PCNA complex may modify the unmethylated state of the promoter regions in housekeeping genes (CpG islands) present in early replicating DNA. However, additional mechanism(s) cannot be ruled out to explain the involvement of poly(ADP-ribosyl)ation in the control of the methylation process. Poly(ADP-ribosyl)ated isoforms of PCNA and p21 have been described and it would be important to ascertain whether this postsynthetic modification can modulate the competition between p21 and DNMT1 for the PCNA binding domain.

Our reported data suggest 3-ABA-induced up-regulation of DNMT1 and the ensuing premature formation of PCNA-DNMT1 complex in G1/S border may be part of the molecular mechanism by which inhibition of poly(ADP-ribose) polymerases introduces anomalous hypermethylation onto DNA; Fig. 3 .

View larger version (117K): [in this window] [in a new window]   Figure 3. Schematic model suggesting that 3-ABA-induced up-regulation of DNMT1 and the ensuing premature formation of active PCNA-DNMT1 complex in G1/S border may be part of the molecular mechanism through which inhibition of poly(ADP-ribose) polymerases introduces anomalous hypermethylation onto DNA.

If we consider the possibility that cellular transformation could be due to a hyperexpression of DNMT1 in a phase of the cell cycle in which the enzyme is down-expressed, our results could lead to a new line of research aimed at clarifying the mechanism through which anomalous hypermethylation of CpG islands occurs during neoplasia.

FOOTNOTES

¹ To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.01-0827fje; to cite this article, use FASEB J. (June 21, 2002) 10.1096/fj.01-0827fje

² Current address: Brain Tumor Research Center, University of California, San Francisco, USA.

This article has been cited by other articles:

				M. Y. Kim, T. Zhang, and W. L. Kraus Poly(ADP-ribosyl)ation by PARP-1: `PAR-laying' NAD+ into a nuclear signal Genes & Dev., September 1, 2005; 19(17): 1951 - 1967. [Abstract] [Full Text] [PDF]

				A. Hermann, R. Goyal, and A. Jeltsch The Dnmt1 DNA-(cytosine-C5)-methyltransferase Methylates DNA Processively with High Preference for Hemimethylated Target Sites J. Biol. Chem., November 12, 2004; 279(46): 48350 - 48359. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) All Versions of this Article: 16/10/1319 01-0827fjev1    most recent Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by ZARDO, G. Articles by CAIAFA, P. Search for Related Content PubMed PubMed Citation Articles by ZARDO, G. Articles by CAIAFA, P.