Ž .Molecular Brain Research 71 1999 23–31www.elsevier.comrlocaterbres

Research report

ž /Downregulation of DNA cytosine-5- methyltransferase is a late event inNGF-induced PC12 cell differentiation

Jing Deng, Moshe Szyf )

Department of Pharmacology and Therapeutics, McGill UniÕersity, 3655 Drummond Street, Montreal, PQ, Canada H3G 1Y6

Accepted 13 April 1999

Abstract

DNA methylation patterns are a critical component of the epigenetic machinery that controls the expression of genetic programs inŽ .vertebrates. DNA methyltransferase gene dnmt1 encodes the enzyme catalyzing the methylation of DNA during replication. We tested

the hypothesis that the expression of dnmt1 is regulated with the developmental state of neuronal cells. We show that DNAŽ .methyltransferase Dnmt1 activity is sharply reduced 4 days after induction of differentiation of PC12 cells with NGF. Similarly, the

adult brain expresses reduced levels of Dnmt1 activity. We propose that the level of Dnmt1 is downregulated to adjust the activity of theDNA methyltransferase to a different role in mature post-mitotic neurons. Both the abundance of dnmt1 mRNA as well as the Dnmt1polypeptide are downregulated. Downregulation of dnmt1 parallels other indicators of withdrawal from the cell cycle such as induction of

Ž .p21, and downregulation of the S phase maker PCNA proliferating cell nuclear antigen . The temporal pattern of downregulation ofŽ .dnmt1 in nerve growth factor NGF -induced PC12 cells is different from myotube differentiation where downregulation of DNA

methyltransferase and demethylation is an early event and was proposed to play a causal role in differentiation. We propose that NGFdifferentiation of PC12 cells represents a different paradigm of involvement of DNA methylation in terminal differentiation. q 1999Elsevier Science B.V. All rights reserved.

Keywords: DNA methylation pattern; DNA methyltransferase; Nerve growth factor; PC12 differentiation; Myotube differentiation; Terminal differentiation

1. Introduction

Vertebrate DNA is modified by methylation at thew xdinucleotide sequence CpG 40 . Not all CpG sites are

modified and the distribution of the methyl groups exhibitssite and tissue specificity resulting in cell specific DNA

w xmethylation patterns 42,55 . DNA methylation patternsw xare formed during development 7,45 and maintained in

w xsomatic cells 55 . Several lines of evidence support thehypothesis that DNA methylation is a critical componentof the epigenetic layer of information of the genomew x40,41,52 . DNA methylation has been proposed to beinvolved in different modalities of differential gene expres-

w xsion such as parental imprinting 13,25 and X-inactivationw x2,8,9,29 . One possible mechanism through which DNA

) Corresponding author. Fax: q1-514-398-6690; E-mail:mszyf@pharma.mcgill.ca

methylation can modulate gene expression is silencinggene activity either directly, by inhibiting binding of tran-

w xscription factors to regulatory sequences 3 , or indirectly,by signaling the binding of methylated-DNA binding fac-

w xtors 11,22,31,41 . Two methylated DNA binding proteinsthat can repress transcription in a methylation dependentmanner have been recently characterized, MeCP2 and

w xMeCP1 11,22,31 . The carboxy terminal half of MeCP2contains a repressor domain which associates with thetranscription repressor mSin3A and histone deacetylasew x20,32 . Deacetylation of histones possibly results in theirtight interaction with DNA, forming an inactive chromatinwhich does not allow the entry of transcription activatorsw x20,32 .

The DNA methyltransferase Dnmt1 catalyzes the DNAw xmethylation reaction 1,24 and is critical for both the

generation of DNA methylation patterns during develop-w xment and their maintenance in somatic cells 27 . It has

been previously proposed that regulation of DNA methyl-transferase expression plays an important role in the con-

0169-328Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0169-328X 99 00147-3

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–3124

w xtrol of the integrity of DNA methylation patterns 47,48 .Different modalities of regulation of dnmt1 expressionhave been previously documented. First, expression ofDNA methyltransferase is downregulated posttranscrip-tionally in growth arrested cells and is induced upon

w xentrance into the S-phase of the cell cycle 49 . Second, theRas–Jun oncogenic signaling pathway upregulates DNAmethyltransferase through AP-1 regulatory sequences in

w xthe DNA methyltransferase gene 30,38,43,44 . Regulationof DNA methyltransferase expression during developmen-tal switches might play an important role in these pro-

w xcesses 51 . To understand the role that dnmt1 might playin these switches, one has to test the hypothesis thatchanges in dnmt1 expression occur during a defined devel-opmental switch. A previous study has shown that expres-sion of dnmt1 is rapidly downregulated upon onset ofmuscle differentiation in vitro which is followed by a

w xgenome wide demethylation 21,28,34 . This study sug-gests that reduction in DNA methyltransferase expressionis required to accomplish the transient hypomethylationobserved in differentiation and therefore downregulation ofdnmt1 expression might play a critical role in the initiationof differentiation. It is not clear, however, whether thismodel is unique or whether it is a representative of a moregeneral mechanism.

Although maintenance methylation is required only dur-ing the synthetic phase of the cell cycle, significant expres-sion of Dnmt1 has been previously detected in post-mitotic

w xbrain cells 16 . It is still unclear what is the role thatDnmt1 might play in a post-mitotic tissue, but it is possiblethat DNA methyltransferase is required to maintain theplasticity of gene expression programs. To understand thedifferent roles that Dnmt1 might play in either mitotic,differentiating or mature post-mitotic CNS cells, it is im-portant to first characterize the expression profile of Dnmt1at these stages.

A well established model to study the molecular eventsoccurring during differentiation of neurons is rat pheochro-

Ž .mocytoma cells PC12 differentiation induced by nerveŽ . w xgrowth factor NGF 17 . NGF can induce a reversible

differentiation of PC12 cells to sympathetic neuron-likew xcells 14 . This system has been extensively utilized to

delineate the different molecular signals involved in devel-w xopment of these cells 6,15,18,33,36,37,46 and is therefore

an ideal model to test the hypothesis that alterations indnmt1 expression are involved in neuronal developmentand to determine the specific stages where these changes

w xoccur. Differentiation of PC12 involves three phases 36 .The first two phases involve mitogenic activity whereasthe last phase is cytostatic resulting in the full maturationof neurons. Thus, this system allows us to determine

w xwhether similar to myotube differentiation 50 , downregu-lation of Dnmt1 is required for initiation of differentiationor whether alterations in dnmt1 expression are a late eventrequired for adjusting the expression of dnmt1 to its newrole in a post-mitotic mature neuron.

2. Materials and methods

2.1. Cell culture and rat tissues

Ž .PC12 cells ATCC were grown in collagen coatedplates in RPMI 1640 supplemented with 10% fetal calf

Žserum and 1% antibiotics. NGF Prince laboratories,.aG7211 was dissolved in a 1% BSArPBS solution and

was added to the cultures at a final concentration of 50ngrml. The culture medium was replaced with fresh NGFcontaining medium every 48 h. Embryonic and adult tis-sues were prepared from time-gestation pregnant or adultSprague–Dawley rats respectively.

2.2. Western blot analysis and total proteins staining

Ž .1 Cell extract preparation: PC12 cells were harvestedat different time points following NGF treatment. Wholecell extracts were prepared by lysing the cells in 2 volumes

Žof the lysis buffer in ice for 10 min: 50 mM Tris–HCl pH.7.5 , 150 mM NaCl, 1% Np-40, 0.5% sodium deoxy-

cholate, 1 mM EDTA, 5 mgrml leupeptin, 1 mgrmlaprotinin, 0.7 mgrml pepstatin, 50 mgrml TLCK, 100mgrml PMSF, 1 mM Na VaO . The lysates were then3 4

spun for 30 min at 12 000 RPM at 48C and the super-natants were collected. Nuclear extracts were prepared as

w xpreviously described 49 .Ž .2 Western blot analysis of DNA methyltransferase: To

visualize the Dnmt1 protein, 50 mg of nuclear extractswere fractionated on a 5% SDS-PAGE and transferred

Žonto a nitrocellulose membrane Westran, Schleicher and.Schuell . The membrane was reacted with an antibody

directed against the catalytic domain of Dnmt1 MetCat2Ž .10 mgrml, 1:2000 dilution for 1 h at room temperaturew x39 which was followed by incubation with a secondary

Žhorseradish peroxidase conjugated anti-rabbit IgG Sigma.A0545, 1:5000 dilution . Following washing of the sec-

ondary antibody, the signal was visualized using ECLWestern blotting following the protocol recommended by

Ž .the manufacturer Amersham . The membrane was thenŽ .stripped of the antibody with 62.5 mM Tris pH 6.7 , 2%

SDS and 100 mM b-mercaptoethanol at 658C for 30 minand reacted again with the same concentration of MetCat2antibody that had been preincubated with the MetCat2peptide for 1 h in ice as a control, and the signal obtainedwas visualized as above. To determine that stripping didnot remove the Dnmt1 protein and that on a second probeMetCat2 antibody lacking preincubation with MetCat2peptide would generate a signal, the membrane was re-

Žacted with the MetCat2 antibody that was not preincu-.bated with MetCat2 after the signal was stripped and a

Ž .similar signal was obtained data not shown .Ž . Ž3 Western blot analysis of p21 and PCNA proliferat-

.ing cell nuclear antigen : To visualize the p21 protein, 50mg of whole cell extracts were analyzed in a 4–15%

Ž .gradient gel Bio-Rad , transferred to nitrocellulose mem-brane. The membrane was reacted with p21 antibody

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–31 25

Ž .Santa Cruz, aSC-397, 100 mgrml, 1:1000 dilution andŽthe secondary antibody anti-rabbit IgG Sigma A0545,

.1:5000 . The signal was revealed with ECL as recom-Ž .mended by the manufacturer Amersham . The blot was

stripped as above and reacted with anti-PCNA antibodyŽ .Santa Cruz, PC10, 200 ugrml, 1:2000 dilution .

Ž .4 Total protein staining of western membranes: Afterwestern blot analysis, the amount of total protein trans-ferred to the same membrane were determined by amido

Ž .black staining. 0.1% wrv Amido black solution wasprepared in 25% isopropanol, 10% acetic acid and 65%H O. The membranes were stained for 30 min with gentle2

shaking and destained by the same solution but Amidoblack for three times; each time was 15 min. The mem-branes were then scanned by laser densitometer for totalproteins.

2.3. DNA methyltransferase actiÕity assay

Nuclear extracts were prepared as above from NGFinduced or untreated PC12 cells as well as rat embryonicand adult tissues. 3 mg of nuclear extract were assayed forDNA methyltransferase activity in a 30 ml reaction mix-ture using 0.1 mg of a synthetic 33-base pair hemimeth-

w xylated oligonucleotide as a substrate 49 and 1 mCi ofw3 x Ž .S- H-methyl adenosyl-L-methionine 78.9 Cirmmol as

the methyl donor in a buffer containing 20 mM Tris–HClŽ .pH.7.4 , 10 mM EDTA, 25% glycerol, 0.2 mM PMSF,and 20 mM b-mercaptoethanol. Following a 3 h incuba-tion at 378C, the samples are incubated at 658C for 10 min,precipitated by 10% TCA, filtered through GFrC filtersŽ . w3 xFisher and the rate of incorporation of H -methyl wasdetermined by liquid scintillation counting.

Ž .Fig. 1. The enzymatic activity of DNA methyltransferase is reduced in NGF-induced PC12 cell differentiation and rat tissue development. A Nuclearextracts were prepared from control and NGF-treated PC 12 cells at different time points. 3 mg of nuclear extracts were assayed for DNA

w3 x w3 xmethyltransferase activity using tritiated H -methyl-S-adenosylmethione as the methyl donor and the incorporation of H -methyl into a hemimethylatedŽ .DNA substrate was determined. The results are presented as an average of three different determinations"S.D. B Neurite outgrowths of NGF-treated

PC12 cells was measured as an indicator of differentiation. Only cells bearing at least one neurite which length is at least 2 times the diameter of the cellbody were included. The numbers are presented as percent of cells bearing neurites. Three independent plates were counted for each specific time point.

Ž .The results are presented as an average of three determinations"S.D. C Nuclear extracts were prepared from rat embryonic and adult tissues and theŽ .DNA methyltransferase activity was determined as in A .

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–3126

2.4. Plasmid construction and RNase protection analysis

The plasmid encoding the 5X dnmt1 riboprobe wasconstructed as follows: RNA was prepared from exponen-

Ž .tially growing PC12 cells using RNAzol B Tel Testfollowing the manufacturer’s protocol. The fragment cov-ering exon 1 to exon 4 was amplified by RT-PCR from 1mg of PC12 cell RNA using Superscript reverse transcrip-tase at 378C for 1 h, which was followed by amplification

Ž . Xwith Promega Taq using the sense primer 74 : 5 -GTGC-X Ž .CTGCACTTGCCTC-3 and the antisense primer 264 :

5X-GAACGGAACACTCTCTCT-3X for 40 cycles of 958Cfor 0.5 min, 528C for 0.5 min and 728C for 1.5 min. The

ŽPCR products were ligated into pCR 2.1 vector Invitro-.gen and the identity of the insert was verified by sequenc-

ing. The plasmid was linearized with restriction enzyme

BamHI and the riboprobe was transcribed using T7 RNAw 32 xpolymerase and a- P -UTP.

The 5X RNA riboprobe was then hybridized at 508C for18 h with 10 mg of the different RNA samples. Thehybridized RNA was digested with RNase A and T1 as

w xdescribed by Krieg and Melton 23 , and the protectedfragments were resolved on an 8 M urea sequencing gel

Ž .next to a Century RNA ladder Ambion .

3. Results

3.1. The enzymatic actiÕity of Dnmt1 is reduced after thelong-term treatment by NGF

We first determined whether the level of DNA MeTaseactivity in the nucleus is altered following NGF-induced

Ž . Ž .Fig. 2. Dnmt1 is downregulated during NGF-induced differentiation of PC12 cells. A 50 ug of nuclear extracts prepared from either NGF-treated NGFŽ . Ž .or control Ctrl PC12 cells at the indicated times in days, d were fractionated on a 5% SDS-PAGE and subjected to a Western blot analysis using a

Ž .Dnmt1 catalytic domain specific antibody MetCat2 . The signal was visualized using standard ECL protocols. To exclude the possibility that the absenceof the faster band in earlier samples is an artifact of the masking of the faster band by the very strong signal seen at earlier times, we show a shorterexposure of the Western blot in the middle panel. The 200 KD Dnmt1 and the smaller isoform are indicated by arrows. The specificity of MetCat2antibody was demonstrated by the disappearance of the bands when the antibody was preincubated with the MetCat2 peptide on ice for 1 h prior to

Ž . Ž .application aMetCat2qpeptide . B The amount of signal corresponding to the DNA methyltransferase was normalized to the level of total proteins asŽ . Ž .determined by Amido black staining see Section 2 and quantified by scanning O.D. arbitrary units . C, control cells. N, NGF-treated cells.

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–31 27

differentiation of PC12 cells and defined the time point atwhich it occurs. As shown in Fig. 1A, DNA MeTaseactivity was essentially unchanged up to 2 days followingNGF treatment. The first signaling response to NGF such

as triggering of the Ras signaling pathway occurs withinw xminutes after application of NGF 5,33,53 and neurite

outgrowth is recognizable at 2 days. Thus, in distinctionw xfrom myotube differentiation 28 , downregulation of DNA

Ž . XFig. 3. The expression of dnmt1 mRNA is downregulated during PC12 differentiation. A 5 RNase protection assay. The riboprobe used in this assayŽ .covers the first four exons of dnmt1 as shown in B . 10 mg of total RNA prepared from untreated and NGF-treated PC12 cells was hybridized with this

Ž .probe and an internal control probe for 18s ribosomal RNA Ambion in the same tube. The full length protected fragment is 332 bp. The other protectedŽ .fragments initiated from exon 1 are indicated. The bottom panel shows the signal obtained with the 18s RNA probe. B dnmt1 riboprobe. A physical map

X Ž . Ž .of the 5 region of rat dnmt1 is shown in the top line. The dnmt1 exons boxed and vector sequences line included in the riboprobe are indicated in theŽ .bottom line. Multiple initiation sites identified previously Deng and Szyf, unpublished data are indicated by horizontal arrows and P1, P2, P3. The

Ž .position of the primers 74,264 used to amplify the riboprobe is indicated by arrows. The position of the binding site for T7 RNA polymerase is indicatedŽ . Ž . Ž . Ž . ŽT7 . B BamHI. C Quantification of dnmt1 mRNA. The autoradiogram shown in A was scanned by scanning laser densitometer Scanalytics

.one-dimensional analysis . The optical density of the protected transcripts initiated from exon 1 was normalized to the signal obtained for 18 s ribosomalRNA which served as an internal control.

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–3128

methyltransferase expression is not an early event andprobably does not play a crucial role in initiating neuronaldifferentiation. The level of DNA methyltransferase activ-ity in the cells was downregulated 2 to 3-folds at day 4when a large percentage of the cells have already devel-

Ž .oped neurite outgrowths Fig. 1B . This data showing thatthe reduction in the level of DNA methyltransferase occursin parallel with neurite outgrowth and does not precedeneurite outgrowth. This is consistent with the hypothesisthat downregulation of DNA methyltransferase activitydoes not play a causal role in initiation of differentiation.

To determine whether similar downregulation of DNAmethyltransferase activity occurs during late stages of de-velopment in vivo, we determined the level of DNAmethyltransferase activity in the developing brain at em-bryonic 13–18 days and compared it with the adult brain.As observed in Fig. 1C, the embryonal rat brain expresseshigh levels of DNA methyltransferase activity which isdramatically reduced in the adult rat brain. However, theadult brain which is mostly a post-mitotic tissue stillexpresses measurable levels of DNA methyltransferaseactivity relative to other post-mitotic tissues such as heartand liver.

3.2. Western blot analysis of Dnmt1 in differentiated PC12cells

To detect whether the reduction in the level of enzy-matic activity of DNA methyltransferase reflects a change

in the abundance of Dnmt1, we extracted nuclear proteinsfrom treated and untreated PC12 cells and performed aWestern blot analysis. As shown in Fig. 2, similar to theDNA methyltransferase activity in NGF-treated cells,Dnmt1 polypeptide abundance is unchanged up to twodays following NGF induction. A clear downregulation ofthe protein occur at 4 days post-induction that is enhanced5.5 days after initiation of NGF treatment. In addition to aquantitative change in the abundance of Dnmt1, a qualita-tive change is observed following NGF induction. A lowermolecular weight polypeptide that is shorter than the fullsize Dnmt1 is enhanced at 5.5 days whereas the Dnmt1polypeptide is downregulated. We have previously demon-strated that during NGF induced differentiation alterna-

w xtively spliced forms of Dnmt1 are expressed 12 . It ispossible that the increase in the relative abundance of thelower molecular weight Dnmt1 is encoded by an alterna-tively spliced form of DNA methyltransferase. However,further experiments are required to identify the molecularidentity of this isoform.

3.3. NGF downregulates the expression of Dnmt1 mRNA

To determine whether the expression of dnmt1 is down-regulated at the mRNA level we performed an RNaseprotection assay using a riboprobe that bears the first 4

Žexons of dnmt 1. We have previously shown Deng and.Szyf, unpublished data that transcription of rat brain

dnmt1 occurs at a cluster of sites in the CG rich area of

Ž .Fig. 4. Expression of cell cycle regulators following NGF treatment. A Expression of p21 and Dnmt1. Whole cell extracts were prepared from untreatedŽ . Ž .0 and NGF-treated PC12 cells at different time points 2 days, 2d to 6 days, 6d . 50 mg of each sample were subjected to fractionation on a 4–15%gradient SDS-PAGE and a Western blot analysis using an anti-p21 antibody. The signal was visualized with ECL using protocols recommended by the

Ž . Ž .manufacturer Amersham . The amount of signal corresponding to p21 O.D. arbitrary units was normalized to the level of total proteins as determined byŽ . Ž .Amido black staining. Dnmt1 expression was determined as in Fig. 2. B Expression of proliferating cell nuclear antigen PCNA . The Western blot

described above was stripped and reacted with an anti-PCNA antibody. The expression of PCNA was determined as above.

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–31 29

exon 1 as well as at exon 3 and 4 downstream to apromoter that has been previously characterized in the

Ž . w xmouse Fig. 3B 43 . To normalize the total RNA level, an18s riboprobe was hybridized with different RNA samplesin the same tube. As observed in Fig. 3A and quantified inFig. 3C, the dnmt1 transcript initiating at P1 is reduced at4 days and is strongly inhibited 5 days post-NGF-induc-tion.

3.4. Downregulation of Dnmt1 following NGF inductionparallels the induction of p21 and inhibition of PCNA

NGF-responsive neurons withdraw from the cell cyclew xduring terminal differentiation 54 . The decrease in Dnmt1

following NGF induction might be one component of theprogram arresting the cell cycle progression of terminallydifferentiated neurons. We therefore determined the timecourse of induction by NGF of p21, an inhibitor of cyclin

w xD1 kinase activity and cell proliferation 4,37 . Expressionof p21 and Dnmt1 have been previously shown to be

w xinversely related in SV40 transformed cells 10 . As ob-served in Fig. 4A p21 and Dnmt1 expression have areciprocal relationship. Similarly, the decrease in expres-sion of PCNA, an auxiliary factor for DNA replication and

ŽDNA repair, parallels the downregulation of Dnmt1 Fig..4B . Thus downregulation of Dnmt1 is a late event in the

differentiation of PC12 cells which parallels the cell cyclearrest of terminally differentiated neurons.

4. Discussion

The experiments described above show that the latephase of NGF induced neuronal differentiation is associ-ated with downregulation of DNA methyltransferase activ-ity, the abundance of the Dnmt1 protein and dnmt1 mRNA.The kinetics of downregulation of Dnmt1 parallels othercell arrest signals such induction of p21 and downregula-tion of the S phase marker PCNA, and may not be requiredfor initiation of PC12 cell differentiation. Downregulationof Dnmt1 occurs in normal adult tissues in comparison tothe situation in the embryo including the brain as demon-strated in Fig. 1. Therefore, we suggest that the downregu-lation of Dnmt1 observed in terminally differentiated PC12cells might reflect the adaptation of Dnmt1 levels ofexpression to the requirements of a post-mitotic terminallydifferentiated cell. Our data demonstrates that Dnmt1 ex-pression is developmentally regulated and that cells havedeveloped mechanisms to coordinate the level of expres-sion of Dnmt1 with other cell cycle parameters. Thesemechanisms might have been developed to prevent thebiological consequences of inappropriate levels of Dnmt1on the physiology of mature neurons. However, the biolog-ical consequences of ectopic high expression of Dnmt1 onneurons remains to be addressed. Another unresolved ques-tion is what is the role of Dnmt1 in a post-mitotic neuron.

Does Dnmt1 play a role in modulating genome functionand allowing the plasticity of gene expression programs inneurons? To address these questions, the effects of inhibi-tion of Dnmt1 on normal neuronal physiology have to bedetermined.

The profile of changes in Dnmt1 expression observed inthis study is different from what has been observed inother models of differentiation such as myoblast and F9cells. In this system genome wide demethylation coincideswith a rapid downregulation of Dnmt1. Dnmt1 inhibitionby either expression of a Dnmt1 antisense construct ortreatment with 5-azaCdR results in induction of myogenic

w xdifferentiation of 10T 1r2 cells 19,26,50 . It is possiblethat the initiation of neuronal PC12 differentiation repre-sents a different paradigm of differentiation that does notinvolve demethylation or an early inhibition of Dnmt1.This is consistent with previous published observationsw x35 and our non-published data that inhibition of DNAmethylation does not induce differentiation of PC12 cells.Dnmt1 expression might be required for the proliferativeinitial stages of neuronal differentiation. This hypothesiscould be supported by previous observations suggestingthat inhibition of DNA methyltransferase with 5-azaCdR

w xprevents neuronal differentiation with NGF 35 . PC12 andw xthe myotube differentiation model previously studied 50

might represent different stages of development that arecommon to both cell lineages or the different paradigms ofinvolvement of Dnmt1 might be unique to each cell lin-eage respectively.

Even though further experiments will be required toaddress these questions, the experiments described in thismanuscript demonstrate that neuronal development is asso-ciated with a regulated change in the expression of Dnmt1positioning the regulation of this enzyme as an importantcomponent of neuronal maturation.

Acknowledgements

This study was supported by a grant to M.S. fromNCIC.

References

w x1 R.L. Adams, E.L. McKay, L.M. Craig, R.H. Burdon, Mouse DNAmethylase: methylation of native DNA, Biochim. Biophy. Acta 561Ž .1979 345–357.

w x2 C. Beard, E. Li, R. Jaenisch, Loss of methylation activates Xist inŽ .somatic but not in embryonic cells, Genes Dev. 9 1995 2325–2334.

w x3 P.B. Becker, S. Ruppert, G. Schutz, Genomic footprinting revealsŽ .cell type-specific DNA binding of ubiquitous factors, Cell 51 1987

435–443.w x4 N. Billon, G.L. Van, B.B. Rudkin, The CDK inhibitor p21

WAF1rCip1 is induced through a p300-dependent mechanism dur-ing NGF-mediated neuronal differentiation of PC12 cells, Oncogene

Ž .13 1996 2047–2054.w x5 B. Binetruy, T. Smeal, M. Karin, Ha-Ras augments c-Jun activity

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–3130

and stimulates phosphorylation of its activation domain, Nature 51Ž .1991 122–127.

w x6 A. Bonni, M.E. Greenberg, Neurotrophin regulation of gene expres-Ž .sion, Can. J. Neurol. Sci. 24 1997 272–283.

w x7 M. Brandeis, M. Ariel, H. Cedar, Dynamics of DNA methylationw x Ž .during development Review , Bioessays 15 1993 709–713.

w x8 L. Carrel, C.M. Clemson, J.M. Dunn, A.P. Miller, P.A. Hunt, J.B.Lawrence, H.F. Willard, X-inactivation analysis and DNA methyla-tion studies of the ubiquitin activating enzyme E1 and PCTAIRE-1

Ž .genes jin human and mouse, Hum. Mol. Genet. 5 1996 391–401.w x9 L. Carrel, H.F. Willard, An assay for X-inactivation based on

differential at the fragile X locus, FMR1, Am. J. Med. Genet. 64Ž .1996 27–30.

w x10 L.S. Chuang, H.I. Ian, T.W. Koh, H.H. Ng, G. Xu, B.F. Li, HumanŽ .DNA– cytosine-5 methyltransferase–PCNA complex as a target for

Ž .p21 WAF1, Science 277 1997 1996–2000.w x11 S.H. Cross, R. Meehan, X. Nan, A. Bird, A component of the

transcriptional repressor MeCP1 shares a motif with DNA methyl-Ž .transferase and HRX proteins, Nat. Genet. 16 1997 256–259.

w x12 J. Deng, M. Szyf, Multiple isoforms of DNA methyltransferase areencoded by the vertebrate cytosine DNA methyltransferase gene, J.

Ž .Biol. Chem. 273 1998 22869–22872.w x13 A. Efstratiadis, Parental imprinting of autosomal mammalian genes,

Ž .Curr. Opin. Genet. Dev. 4 1994 265–280.w x14 K. Fujita, P. Lazarovici, G. Guroff, Regulation of the differentiation

w x w xof PC12 pheochromocytoma cells Review 249 refs , Environ.Ž .Health Perspect. 80 1989 127–142.

w x15 L. Golloapudi, K.E. Neeet, Different mechanisms for inhibition ofcell proliferation via cell cycle proteins in PC12 cells by nerve

Ž .growth factor and staurosporine, J. Neurosci. Res. 49 1997 461–474.

w x16 K. Goto, M. Numata, J. Komura, T. Ono, T.H. Bestor, H. Kondo,Expression of DNA methyltransferase gene in mature and immatureneurons as well as proliferating cells in mice, Differentiation 56Ž .1994 39–44.

w x17 L.A. Greene, A. Tischler, Establishment of a noradrenergic clonalline of rat adrenal pheochromocytoma cells which respond to nerve

Ž . Ž .growth factor, Proc. Natl. Acad. Sci. U.S.A. 73 7 1976 2424–2428.

w x18 N. Hagag, S. Halegoua, M. Viola, Inhibition of growth factor-in-duced differentiation of PC12 cells by microinjection of antibody to

Ž .ras p21, Nature 319 1986 680–682.w x Ž .19 P.A. Jones, Alter gene expression with 5-azacytidine, Cell 40 1985

485–486.w x20 P.L. Jones, G.J.C. Veenstra, P.A. Wade, D. Vermaak, S.U. Kass, N.

Landsberger, J. Strouboulis, A.P. Wolffe, Methylated DNA andMeCP2 recruit histone deacetylase to repress transcription, Nat.

Ž .Genet. 19 1998 187–191.w x21 J.P. Jost, Y.C. Jost, Transient DNA demethylation in differentiating

mouse myoblasts correlates with high activity of 5-methylde-Ž .oxycytidine excision repair, J. Biol. Chem. 269 1994 10040–10043.

w x22 S.U. Kass, N. Landsberger, A.P. Wolffe, DNA methylation directs atime-dependent repression of transcription initiation, Curr. Biol. 7Ž .1997 157–165.

w x Ž .23 P.A. Krieg, D.A. Melton, Methods Enzymol. 155 1987 397–415.w x24 S. Kumar, X. Cheng, S. Klimsauskas, S. Mi, J. Posfai, R.J. Roberts,

Ž .G.G. Wilson, The DNA cytosine-5 methyltransferases, NucleicŽ . Ž .Acids Res. 22 1 1994 1–10.

w x25 M. Lalande, Parental imprinting and human disease, Annu. Rev.Ž .Genet. 30 1996 173–195.

w x26 A.B. Lassar, B.M. Paterson, H. Weintraub, Transfection of a DNAlocus that mediates the conversion of 10T1r2 fibroblasts to my-

Ž .oblasts, Cell 47 1986 649–656.w x27 E. Li, T.H. Bestor, R. Jaenisch, Targeted mutation of the DNA

Ž .methyltransferase gene results in embryonic lethality, Cell 69 1992915–926.

w x28 Y. Liu, L. Sun, J.P. Jost, In differentiating mouse myoblasts DNA

methyltransferase is posttranscriptionally and posttranslationally reg-Ž .ulated, Nucleic Acids Res. 24 1996 2718–2722.

w x29 S. Luo, J.C. Robinson, A.L. Reiss, B.R. Migeon, DNA methylationof the reagile X locus in somatic and germ cells during fetaldevelopment: relevance to the fragile X syndrome and X-inactiva-

Ž .tion, Somatic Cell Mol. Genet. 19 1993 393–404.w x30 A.R. MacLeod, J. Rouleau, M. Szyf, Regulation of DNA methyla-

Ž .tion by the Ras signaling pathway, J. Biol. Chem. 270 199511327–11337.

w x31 X. Nan, F.J. Campoy, A. Bird, MeCP2 is a transcriptional repressorŽ .with abundant binding sites in genomic chromatin, Cell 88 1997

471–481.w x32 X. Nan, H.-H. Ng, C.A. Jonson, C.D. Laherty, B.M. Turner, R.N.

Eisenman, A. Bird, Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex,

Ž .Nature 393 1998 386–389.w x33 N.F. Ng, E.M. Shooter, Activation of p21ras by nerve growth factor

in embryonic sensory neurons and PC12 cells, J. Biol. Chem. 268Ž .1993 25329–25333.

w x34 Z. Paroush, I. Keshet, J. Yisreali, H. Cedar, Dynamics of demethyl-ation and activation of the alpha-actin gene in myoblasts, Cell 63Ž .1990 1229–1237.

w x35 S.P. Persengiev, D.L. Kilpatrick, Nerve growth factor induced dif-ferentiation of neuronal cells requires gene methylation, NeuroRe-

Ž . Ž .port 8 1 1997 227–231.w x36 N. Peunova, G. Enikolopov, Nitric oxide triggers a switch to growth

Ž .arrest during differentiation of neuronal cells, Nature 375 199568–73.

w x37 W. Poluha, C.M. Schonhoff, K.S. Harrington, M.B. Lachyankar,N.E. Crosbie, D.A. Bulseco, A.H. Ross, A novel nerve growthfactor-activated pathway involving nitric oxide, p53, and p21WAF1regulates neuronal differentiation of PC12 cells, J. Biol. Chem. 272Ž .1997 24002–24007.

w x38 S. Ramchandani, P. Bigey, M. Szyf, Genomic structure of theŽ .human DNA methyltransferase gene, Biol. Chem. 379 1998 535–

540.w x39 S. Ramchandani, A.R. MacLeod, M. Pinard, H.E. Von, M. Szyf,

Inhibition of tumorigenesis by a cytosine–DNA, methyltransferaseantisense oligodeoxynucleotide, Proc. Natl. Acad. Sci. U.S.A. 94Ž .1997 684–689.

w x40 A. Razin, A.D. Riggs, DNA methylation and gene function, ScienceŽ .210 1980 604–610.

w x41 A. Razin, H. Cedar, DNA methylation and gene expression, Micro-Ž .biol. Rev. 55 1991 451–458.

w x42 A. Razin, M. Szyf, DNA methylation patterns. Formation andw x w x Ž .function Review 99 refs , Biochim. Biophys. Acta 782 1984

331–342.w x43 J. Rouleau, G. Tanigawa, M. Szyf, The mouse DNA methyltrans-

ferase 5X-region. A unique housekeeping gene promoter, J. Biol.Ž .Chem. 267 1992 7368–7377.

w x44 J. Rouleau, A.R. MacLeod, M. Szyf, Regulation of the DNAmethyltransferase by the Ras-AP-1 signaling pathway, J. Biol. Chem.

Ž .270 1995 1595–1601.w x45 R. Shemer, S. Eisenberg, J.L. Breslow, A. Razin, Methylation

patterns or the human APOA-1rC-111rA-IV gene cluster in adultand embryonic tissues suggest dynamic changes in methylation

Ž .during development, J. Biol. Chem. 266 1991 23676–23681.w x46 J. Szeberenyi, H. Cai, G.M. Cooper, Effect of a dominant inhibitory

Ha-ras mutation on neuronal differentiation of PC12 cells, Mol. CellŽ .Biol. 10 1990 5324–5332.

w x47 M. Szyf, K. Avraham-haetzni, A. Reifamn, J. Shlomai, F. Kaplan,A. Oppenheim, A. Razin, DNA methylation pattern is determined bythe intracellular level of the methylase, Proc. Natl. Acad. Sci. U.S.A.

Ž .81 1984 3278–3282.w x48 M. Szyf, DNA methylation patterns: an additional level of informa-

Ž .tion?, Biochem. Cell Biol. 69 1991 764–767.w x49 M. Szyf, V. Bozovic, G. Tanigawa, Growth regulation of mouse

( )J. Deng, M. SzyfrMolecular Brain Research 71 1999 23–31 31

Ž .DNA methyltransferase gene expression, J. Biol. Chem. 266 199110027–10030.

w x50 M. Szyf, J. Rouleau, J. Theberge, V. Bozovic, Induction of myo-genic differentiation by an expression vector encoding the DNAmethyltransferase cDNA sequence in the antisense orientation, J.

Ž .Biol. Chem. 267 1992 12831–12836.w x51 M. Szyf, DNA methylation properties: consequences for pharmacol-

Ž .ogy, Trends Pharmacol. Sci. 15 1994 233–238.w x52 P.H. Tate, A.P. Bird, Effects of DNA methylation on DNA-binding

Ž .proteins and gene expression, Curr. Opin. Genet. Dev. 3 1993226–231.

w x53 K.W. Wood, C. Sarnecki, T.M. Roberts, J. Blenis, Ras mediatesnerve growth factor receptor modulation of three signal-transducing

Ž .protein kinases: MAP kinase, Raf-1 and RSK, Cell 68 19921041–1050.

w x54 G. Yan, E. Ziff, NGF regulates the PC12 cell cycle machinerythrough specific inhibition of the Cdk kinases and induction of

Ž .cyclin D1, J. Neurosci. 15 1995 6200–6212.w x Ž .55 J. Yisraeli, M. Szyf, in: A. Razin, H. Cedar, A.D. Riggs Eds. ,

DNA Methylation: Biochemistry and Biological Significance,Springer-Verlag, New York, 1984, pp. 353–378.