CpG methylation of half-CRE sequences creates C/EBPα binding sites that activate some tissue-specific genes

CpG methylation of half-CRE sequences creates C/EBPα binding sites that activate some tissue-specific genes

Please download to get full document.

View again

of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 3 | Pages: 6

Extension: PDF | Download: 0

  CpG methylation of half-CRE sequences creates C/EBP α binding sites that activate some tissue-specific genes Vikas Rishi a,1 , Paramita Bhattacharya a,1 , Raghunath Chatterjee a,1 , Julian Rozenberg a,1 , Jianfei Zhao a,1 , Kimberly Glass b ,Peter Fitzgerald c , and Charles Vinson a,2 a Laboratory of Metabolism, National Cancer Institute, National Institutes of Health, Building 37, Room 3128, Bethesda, MD 20892;  b Physics Department,University of Maryland, College Park, MD 20742; and  c Genome Analysis Unit, National Cancer Institute, National Institutes of Health, Bethesda,MD 20892Edited* by Steven L. McKnight, University of Texas Southwestern, Dallas, TX, and approved October 5, 2010 (received for review June 18, 2010) DNA methylation of the cytosine in the CpG dinucleotide istypically associated with gene silencing. Genomic analyses haveidentified low CpG promoters that are both methylated and tran-scriptionally active, but the mechanism underlying the activationof these methylated promoters remains unclear. Here we showthat CpG methylation of the CRE sequence (TGACGTCA) enhancesthe DNA binding of the C/EBP α  transcription factor, a protein criti-cal for activation of differentiation in various cell types. Transfec-tion assays also show that C/EBP α  activates the CRE sequence onlywhen it is methylated. The biological significance of this obser-vation was seen in differentiating primary keratinocyte culturesfrom newborn mice where certain methylated promoters are bothbound by C/EBP α  and activated upon differentiation. Experimentaldemethylation by either 5-azacytidine treatment or DNMT1 deple-tion diminished both C/EBP α  binding and activation of the samemethylated promoters upon differentiation suggesting that CpGmethylation can localize C/EBP α . Transfection studies in cell cul-tures using methylated tissue-specific proximal promoters identi-fied half-CRE (CGTCA) and half-C/EBP (CGCAA) sequences thatneed to be methylated for C/EBP α  mediated activation. In primarydermal fibroblasts, C/EBP α  activates a different set of methylatedtissue-specific promoters upon differentiation into adipocytes.These data identify a new function for methyl CpGs: producingDNA binding sites at half-CRE and half-C/EBP sequences forC/EBP α  that are needed to activate tissue-specific genes. gene regulation  ∣  EMSA  ∣  transcription factor binding site I n mammalian genomes, CpG dinucleotides are rare but dooccur in clusters called CpG islands that are often located inthe proximal promoters of genes, particularly housekeepinggenes (1 – 3). In the early embryo, there is little CpG methylation,but CpG dinucleotides outside of CpG islands typically becomemethylated during the blastula stage of development (4). Promo-ters with methylated CpGs are generally transcriptionally silentas occurs with X-chromosome inactivation and imprinting (4).CpG methylation both recruits repressive complexes (4) andprevents the DNA binding of many transcription factors (TFs)(5). In some cancers, methylation of tumor suppressor gene pro-moters is associated with gene repression (6) lending support tothe suggestion that CpG methylation is a general repressive epi-genetic mark (7). CpG methylation patterns are not as dynamicas previously thought (8), and it is mainly the regions outside of proximal promoters that become demethylated upon cellular dif-ferentiation (9 – 11). Genomic analyses have identified low CpGpromoters that are both methylated and transcriptionally active(8, 12), but the mechanism underlying the activation of methy-lated promoters remains unclear. Results and Discussion C/EBP α  Binds a Methyl CRE.  During a survey to evaluate if CpGmethylation of classical proximal promoter elements affectedDNA binding of nuclear extracts from keratinocytes (Fig. 1  A )or mouse liver ( SI Appendix , Fig. S1  A ), we observed that bothunmethylated and methylated CRE sequences (TGACGTCA) were bound. This was surprising because methylation of theCRE has been reported to inhibit DNA binding of CREB, theprotein known to bind the CRE (13). To identify the protein(s) that bind the unmethylated and methylated CRE, we addedantibodies to B-ZIP proteins, including C/EBP family membersand other proteins known to bind the CRE. These  “ supershift ” experiments indicated that the unmethylated CRE was bound by CREB, c-Jun, JunD, and ATF2 as expected (14). In contrast, themethylated CRE was not bound by CREB but was bound by C/EBP α  and C/EBP β . c-Jun, JunD, and ATF2 showed a modestdecrease in binding. Using pure CREB and C/EBP α  B-ZIPdomains, both EMSA (Fig. 1  B ) and circular dichroism thermaldenaturations ( SIAppendix , Fig. S1  B – C ) recapitulated the nucle-ar extract results: C/EBP α  preferentially binds the CRE whenthe CpG is methylated. Both methylated and unmethylated con-sensus C/EBP sequences (TTGCGCAA) were only bound in nu-clear extracts by C/EBP family members ( SI Appendix , Fig. S1  D ).More detailed studies using EMSA and circular thermal dena-turation show pure C/EBP α  protein preferentially binds whenthe central CpG is methylated ( SI Appendix , Fig. S1  E  and  F  ).EMSA using an equimolar mixture of C/EBP α  and CREBshowed that C/EBP α  preferentially binds to the methylated CREeven in the presence of CREB (Fig. 1 C ). We chose to study C/EBP α  in more detail because it is involved in the activationof differentiation in several cell types (15), including keratino-cytes (16 – 18) and adipocytes (19 – 21), potentially linking C/EBP α binding to methylated promoters with activation of methylatedpromoters upon differentiation (see below).To evaluate whether the preferential binding of C/EBP α  to amethylated CRE is biologically relevant, we did transient trans-fections using primary keratinocytes (Fig. 1  D ). Four copies of theCRE sequence were cloned into a reporter plasmid where allCpG dinucleotides have been deleted from the plasmid backbone(pCpGL) (22). The use of a reporter without any CpGs in theparental plasmid backbone allowed evaluation of the effect of CpG methylation within the insert on transcriptional activation.The CpGs in the pCpGL-4X-CRE plasmid insert were enzyma-tically methylated ( SI Appendix , Fig. S1 G ) and transfected intokeratinocytes where they maintained their CpG methylationstatus ( SI Appendix , Fig. S1  H  ). The methylated plasmid is moreactive and is preferentially inhibited when a dominant-negativeC/EBP protein was expressed, suggesting that the methylation Author contributions: V.R., P.B., R.C., J.R., K.G., and C.V. designed research; V.R., P.B., R.C.,J.R., and J.Z. performed research; V.R., P.B., R.C., J.R., J.Z., K.G., P.F., and C.V. analyzed data;and V.R., P.B., R.C., J.R., J.Z., K.G., and C.V. wrote the paper.The authors declare no conflict of interest.*This Direct Submission article had a prearranged editor.Freely available online through the PNAS open access option. 1 V.R., P.B., R.C., J.R., and J.Z. contributed equally to this work. 2 To whom correspondence should be addressed. E-mail: vinsonc@mail.nih.gov.This article contains supporting information online at www.pnas.org/lookup/suppl/ doi:10.1073/pnas.1008688107/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1008688107 PNAS  ∣  November 23, 2010  ∣  vol. 107  ∣  no. 47  ∣  20311 – 20316      B     I     O     C     H     E     M     I     S     T     R     Y  dependent activation of the methylated CRE plasmid could bedue to the C/EBP family function. Cotransfection of methylatedor unmethylated pCpGL-4X-CRE with an expression plasmidencoding C/EBP α  or CREB showed that the methylated plasmidis activated by C/EBP α , whereas the unmethylated plasmid isactivated by CREB (Fig. 1  D  and  SI Appendix , Fig. S1  I  ). Similarresults were also obtained using an alternative approach toproduce a luciferase reporter plasmid with methylated CREsequences ( SI Appendix , Fig. S1  J  ). These data suggest that CpGmethylation of the CRE sequence is needed to produce a tran-scription factor binding site (TFBS) for C/EBP α . Keratinocyte Differentiation.  The observation that CpG methyla-tion facilitates the DNA binding of a transcription factor involvedin cellular differentiation is particularly interesting consideringthat demethylation using 5-azacytidine (5-aza), a cytosine analo-gue that cannot be methylated, or depletion of CpG methylationmaintenance enzyme (DNA methyltransferase 1, DNMT1), caninhibit differentiation (23 – 25). One way this could happenis if CpG methylation is required for transcriptional activationof certain tissue-specific genes. Thus, we set out to investigate whether C/EBP α  binds to any methylated promoters that are ac-tivated with differentiation. We used calcium to induce differen-tiation of primary keratinocyte cultures from newborn mice andidentified 182 methylated promoters that become activated withdifferentiation (Fig. 2). We determined the methylation status of proximal promoters using methylated DNA immunoprecipitation(MeDIP) (26) and Nimblegen promoter microarrays and identi-fied mRNAs that increased with differentiation using Affymetrix microarrays. No dramatic changes in CpG methylation at prox-imal promoters were observed with differentiation either for allgenes or for differentiation specific genes (Fig. 2  A  and  SI  Appendix , Fig. S2  A ). Plotting CpG methylation in undifferen-tiated cells versus the number of CpGs in the promoter producedtwo distinct groups (Fig. 2  B ). One group is CpG sparse and thesepromoters are predominantly methylated with a steady increasein MeDIP enrichment as the number of CpGs in the promoterincreases (8). The second group has more CpGs but lower MeD-IP enrichment, suggesting that they are either partially methy-lated or completely unmethylated (12). These two groups of promoters are similar to those described in a recent study thatadjusted the MeDIP enrichment for the CpG density of proximalpromoters (27). We also examined the promoters classified intothree groups based on their CpG density (12) ( SI Appendix ,Fig. S2  B ). In undifferentiated keratinocytes, the majority of theactive genes have unmethylated promoters, and some (19%) ac-tive genes have methylated promoters, as expected (Fig. 2 C  and SI Appendix , Fig. S2 C ). With differentiation, the mRNAs for 452genes increased more than 2-fold, and 40% of these genes havemethylated promoters (Fig. 2  D  and  SI Appendix , Fig. S2  D –  H  ).Genes activated with differentiation also showed an increase inbinding of two active marks [phosphorylated RNA polymerase II(p-RNAP) and H3K9acetyl] as assessed by chromatin immuno-precipitation followed by genome-wide promoter analysis (ChIP-chip) ( SI Appendix , Fig. S2  I  ). Immunocytochemistry showedthat the increase in expression of differentiation specific proteins with differentiation was a general property of all cells in theculture ( SI Appendix , Fig. S2  J  –  K  ).To determine if methylation is required for the activation of genes with methylated promoters during keratinocyte differentia-tion, we experimentally demethylated the CpG dinucleotides inthe genome using two methods, namely, 5-aza treatment and siR-NA mediated depletion of DNMT1 ( SI Appendix , Fig. S3  A –  E ).Both demethylation methods caused ∼ 20%  loss of global cytosinemethylation as assayed by HPLC and ELISA ( SI Appendix ,Fig. S3  C –  D ). When we focused on activation of genes with dif-ferentiation, the 5-aza treated cultures preferentially inhibitedthe induction ofthe genes withmethylated promoters. Similar butless statistically significant results were observed using DNMT1-depleted cultures. When both demethylation methods were    C   /   E   B   P      α    C   R   E   B   1   C   /   E   B   P        β    J  u  n   D   A   T   F   2  c  -   J  u  n   C   /   E   B   P      α    C   R   E   B   1   C   /   E   B   P        β    J  u  n   D   A   T   F   2  c  -   J  u  n A    C   /   E   B   P      α    J  u  n   D   C   R   E   B   1   C   /   E   B   P        β    A   T   F   2  c  -   J  u  n + Antibodies+ Antibodies+ Antibodies+ Antibodies    C   R   E   B   1   C   /   E   B   P        β    J  u  n   D   A   T   F   2  c  -   J  u  n   C   /   E   B   P      α DNA + ProteinSupershift(TGACGTCA)(TGAGTCA)(TTGCGCAA)CREAP1C/EBPMethyl CRE(TGA m CGTCA)Free DNA B EMSA (TGACGTCA) C Mixture(50 nM each) l l            45 D Transfection (pCpGL-4X-CRE) Unmethylated CpGMethylated CpG CREBC/EBP α ll            234    R  e   l  a   t   i  v  e   l  u  c   i   f  e  r  a  s  e Free DNA CpGCpGMethyl CpGMethyl CpG 01       C      R      E      B      C      /      E      B      P      α       A   -      C      E      B      P _Free DNA Fig. 1.  C/EBP family members bind to methylatedconsensus CRE site. (  A ) EMSA with primary keratinocytenuclear extracts using antibodies to B-ZIP proteins and28 bp oligonucleotides containing ( i  ) an unmethylatedCRE, ( ii  ) a methylated CRE, ( iii  ) an AP-1 consensus site,or ( iv  ) a C/EBP consensus site identifies that CREB bindsto an unmethylated CRE sequence, whereas C/EBP familymembers bind to a methylated CRE sequence. Only c-Junand JunD bind the AP-1 sequence, and only C/EBP familymembers bindthe C/EBP sequence.( B )EMSA usingthe CREsequence used in  A  and increasing concentrations of pureCREB and C/EBP α  B-ZIP protein domains (5-, 15-, 50-, and150-nM dimer) show CREB preferentially binds the un-methylated CRE, whereas C/EBP α  preferentially binds themethylated CRE. ( C  ) EMSA showing a mixture of CREBand C/EBP α  B-ZIP domains at equimolar concentrations(50-nM dimer). ( D ) pCpGL-4X-CRE, a luciferase reportercontaining four consensus CRE sites, was constructed frompCpGL, a parental luciferase reporter plasmid withoutCpGs in the backbone. Cotransfection of pCpGL-4X-CREinto keratinocytes with CREB coding plasmid increasesthe activity of only the unmethylated reporter, whereasC/EBP α  coding plasmid increases the activity of only themethylated reporter (T test,  p  <  0 . 001 ). 20312  ∣  www.pnas.org/cgi/doi/10.1073/pnas.1008688107 Rishi et al.  compared, statistically significant similar genes with methylatedpromoters were inhibited suggesting that for these promoters,CpG methylation is critical for their activation ( SI Appendix ,Fig. S2  E  and  G ).We next determined the localization of C/EBP α  in proximalpromoters of undifferentiated keratinocytes (Fig. 2  E  and  SI  Appendix , Fig. S3  F  –  I  ). C/EBP α  ChIP-chip showed that C/EBP α preferentially binds promoters containing the known C/EBPconsensus site (28) and that 85% of bound promoters are un-methylated (Fig. 2  E ). However, 15% of bound promoters aremethylated. C/EBP α  ChIP-chip from either 5-aza-treated orDNMT1-depleted undifferentiated keratinocytes showed a pre-ferential loss of binding to similar methylated promoters, suggest-ing that C/EBP α  localization to these promoters is dependent onCpG methylation (Fig. 2  E  and  SI Appendix , Fig. S3 G ), a resultconfirmed by qPCR (Fig. 2  F  ). C/EBP α  bound similar promotersin undifferentiated and differentiated keratinocytes, suggestingthat the recruitment of C/EBP α  to methylated promoters isnot the signal for activation of these promoters with calcium( SI Appendix , Fig. S3  H  ). Experimental demethylation of differen-tiating keratinocytes again preferentially inhibited C/EBP α binding to methylated promoters ( SI Appendix , Fig. S3  I  ), includ-ing the methylated promoters activated with differentiation( SIAppendix ,Fig. S3  J  and  K  ). Thesequence that is most enrichedin promoters where C/EBP α  binding is inhibited after demethy-lation by either 5-aza treatment or DNMT1 depletion is thehalf-CRE motif, suggesting that this methylated sequence may mediate the localization of C/EBP α  to some methylated promo-ters (Fig. 2  E  and  SI Appendix , Fig. S3  L ). The role of a methylatedhalf-CRE in C/EBP α  localization prompted us to determine thegenomic localization of CREB, a transcription factor known tobind half-CRE sequences. CREB ChIP-chip in undifferentiatedkeratinocytes shows that the CRE consensus sequence is prefer-entially bound by CREB as expected, and 98% of bound promo-ters are unmethylated ( SI Appendix , Fig. S4  A ). We observed aslight increase in CREB binding to some methylated promotersafter 5-aza treatment, suggesting that CpG methylation may prevent CREB binding in vivo ( SI Appendix , Fig. S4  B ) as it doesin vitro.The importance of C/EBP α  function in keratinocyte differen-tiation was evaluated using two methods. First, we culturedprimary keratinocytes from transgenic mice expressing a domi-nant-negative protein (A-C/EBP) that inhibits the DNA bindingof C/EBP family members (18). Second, we used siRNA todeplete C/EBP α  and C/EBP β  ( SI Appendix , Fig. S4 C ). GlobalmRNA analysis indicated that both strategies suppressed kerati-nocyte differentiation and preferentially inhibited the activationof genes containing methylated promoters ( SI Appendix ,Fig. S4  D ). The same methylated genes where CpG demethylationinhibited induction with differentiation were also suppressed by C/EBP α  inhibition, suggesting a connection between these twodistinct events ( SI Appendix , Fig. S4  E ). Activity of Methylated Promoters.  We next turned to individualmethylated promoters that are activated with keratinocyte differ-entiation to evaluate whether activation is mediated by C/EBP α binding to methylated DNA sequences. We examined four ker-atinocyte specific promoters that fit five criteria: ( i ) increased Fig. 2.  C/EBP α  binding activates methylated promo-ters in keratinocyte differentiation. (  A ) Scatter plot(log 2 ) of MeDIP enrichment before and after kerati-nocyte differentiation using Nimblegen mouse pro-moter arrays with each dot representing apromoter. ( B ) MeDIP enrichment in undifferentiatedkeratinocytes versus the number of CpG dinucleo-tides in the promoter ( − 1  ; 000  bp to  þ 500  bp). Thered line demarcates two groups of promoters, un-methylated (UM) and methylated (M). ( C  ) Methyla-tion status of promoters for the most abundantlyexpressed genes (top 20%) in undifferentiated(Undiff) keratinocytes. ( D ) Number of UM and M pro-moters whose mRNA expression was doubled withdifferentiation (Diff,  Upper  ). mRNA induction duringdifferentiation is preferentially inhibited (inductiondecreases  > 50% ) for genes with methylated promo-ters by experimental demethylation using either5-aza (*  χ  2 test,  p  <  0 . 01 ), or siRNA to DNMT1. Bothdemethylation methods decreased expression ofthe same methylated promoters (hypergeometrictest,  p  <  0 . 05 ). ( E  ) Number of UM and M promotersbound by C/EBP α  (binding enrichment  > 1 . 2 , T test,  p  <  0 . 03 ) in undifferentiated keratinocytes. Logoshowing a C/EBP consensus site is the most enrichedsequence in promoters bound by C/EBP α . C/EBP α binding is inhibited (< 0 . 8 ) preferentially at methy-lated promoters after treatment with 5-aza (*  χ  2 test,  p  <  0 . 001 ) or siRNA to DNMT1 (*  χ  2 test,  p  <  0 . 02 ).Both demethylation methods decreased C/EBP α  bind-ing to the same methylated promoters (hypergeo-metric test,  p  <  0 . 001 ). The half-CRE (CGTCA) is themost common DNA sequence occurring in promoterswith decreased C/EBP α  binding following either de-methylation method. ( F  ) qPCR of C/EBP α  ChIP DNAfrom undifferentiated keratinocytes for the tissue-specific (Bglap-rs1 and AK043395) and housekeeping(Rps12) promoters with or without 5-aza treatmentcompared to input DNA. All values are expressedasfoldchangenormalizedtoRps12(*Ttest,  p  <  0 . 05 ). Rishi et al. PNAS  ∣  November 23, 2010  ∣  vol. 107  ∣  no. 47  ∣  20313      B     I     O     C     H     E     M     I     S     T     R     Y  transcriptional activity upon differentiation, ( ii ) methylated asdetermined by MeDIP ( SI Appendix , Fig. S4  F  ), ( iii ) bound by C/EBP α , ( iv ) C/EBP α  binding decreased by experimental de-methylation, and (  v ) containing a half-CRE sequence. As a con-trol, we examined two unmethylated housekeeping promoters.For promoters fitting the five criteria above, we determinedthe methylation status of CpGs in individual promoters beforeand after differentiation. We did bisulfite sequencing by clonalanalyses for four promoters (Bglap-rs1, AK043395, Scgb1a1,and Gsta3). The promoters were predominantly methylatedboth before and after differentiation (Fig. 3  A  and  SI Appendix ,Figs. S4 G  and S5). To evaluate the importance of C/EBP α  andCpG methylation on promoter activity, these promoters werecloned into the pCpGL reporter, enzymatically methylated,and then transfected into keratinocytes where CpG methylationof the plasmid was maintained ( SI Appendix , Fig. S4  H  ). Fig. 3shows data for the Bglap-rs1 promoter that has both a half-C/ EBP site and a half-CRE site. C/EBP α  activated the methylatedpromoter but not the unmethylated promoter. CREB, in con-trast, activated the unmethylated promoter but not the methy-lated promoter (Fig. 3  B ). Mutation of either site ( SI Appendix ,Figs. S3  B  and S4  I  ) decreased C/EBP α  activation of the methy-lated Bglap-rs1 reporter and CREB activation of the unmethy-lated promoter. Biochemical studies using circular dichroismthermal denaturation ( SI Appendix , Fig. S6  A –  D ), DNase foot-printing ( SI Appendix , Fig. S6  E ), and EMSA with pure B-ZIPdomains of C/EBP α  and CREB (Fig. 3 C ) recapitulated what was observed in the transient transfections: C/EBP α  bound eithersite better when they were methylated. Mixing C/EBP α  andCREB together with DNA indicated that C/EBP α  binds themethylated half-CRE site even in the presence of equimolarCREB, whereas CREB binds this site when it is unmethylatedeven in the presence of C/EBP α  (Fig. 3  D ). These data suggestthat methylation is essential for C/EBP α  binding to methylatedpromoters, and a subset of these promoters is induced duringkeratinocyte differentiation. Methylation, however, is not the ac-tivation signal, as we do not see changes in methylation at thesepromoters upon calcium-mediated differentiation. The signal(s)that activate keratinocyte-specific gene expression can be poten-tially posttranslational modifications of C/EBP α  (29) that recruitcoactivators causing H3K9 acetylation and phosphorylation of RNAP ( SI Appendix , Fig. S2  J  ). As already shown, when C/EBP α  was overexpressed in undifferentiated keratinocytes, we observeda methylation-dependent promoter activation.Data for five additional methylated promoters activated withdifferentiation (AK043395, Scgb1a1, Gsta3, MMP9, and PSCA)and two housekeeping promoters (Kifc1 and Rps12) are shown in SI Appendix , Figs. S5, S7  A  and  B , and S8  A  and  B . Results for AK043395 are similar to those described for Bglap-rs1: C/EBP α preferentially activates the methylated plasmid via a half-CREsequence and pure C/EBP α  protein binds better to the methy-lated half-CRE sequence. In contrast, the Scgb1a1 and Gsta3promoters are activated by C/EBP α  via a half-CRE sequenceindependent of their methylation status. A plot of the effect of CpG methylation on C/EBP α  binding to different half-CRE se-quences versus transactivation in transient transfections shows apositive correlation: The more methylation enhances C/EBP α binding in EMSA, the more it enhances transactivation (Fig. 3  E ). An estimate of   K   D  of CREB or C/EBP α  and nine studied DNA sequences are shown in  SI Appendix , Fig. S8  K  . The two unmethy-lated housekeeping promoters were almost totally inactivated by methylation supporting the general suggestion that methylation istranscriptionally suppressive ( SI Appendix , Fig. S7  A  and  B ).Primary keratinocytes treated by 5-aza were used for C/EBP α ChIP to examine in vivo whether C/EBP α  preferentially binds themethylated DNA in the presence of the same DNA that is un-methylated ( SI Appendix , Fig. S7  C  and  D ). Bisulfite sequencingof individual clones of genomic DNA from 5-aza-treatedkeratinocytes identified  ∼ 20%  decrease in CpG methylation of the Bglap-rs1 promoter ( SI Appendix , Fig. S7 C ) as was observedglobally by both ELISA and HPLC. C/EBP α  ChIP from the 5-azasamples was enriched for the methylated sequences, demonstrat-ing the importance of CpG methylation for C/EBP α  localizationto these proximal promoters. CREB in contrast preferentially binds to the unmethylated promoters ( SI Appendix , Fig. S7  C and  D ). Adipocyte Differentiation.  C/EBP α  is also known to be importantfor adipocyte differentiation (21), and we investigated if C/EBP α binding to methylated CpG sequences is critical for differentia- -37 -417 A Bisulfite seq C EMSA B pCpGL-Bglap-rs1 (-460 to +40)    U  n   d   i   f   f   i   f   f (1)(2) D Mixture (2)    C   R   E   B   C   /   E   B   P      α    C   R   E   B   C   /   E   B   P      α Free DNA _ AAT CGCAA ACC CGTCA ___ C/EBP α CREB __ Unmethylated (CpG)Methylated (mCpG) Mutant 2   e   l  a   t   i  v  e   l  u  c   i   f  e  r  a  s  e WT 123 Mutant 1    D 12 CpGmCpG Free DNA CpGmCpGCpGmCpG E  B   P      α   p  o  r   t  e  r  s   )   R 0 ___    C   /   E   B   P      α    C   /   E   B   P      α    C   /   E   B   P      α    C   R   E   B      C   R   E   B   C   R   E   B TF456 Bglap-rs14X-CRE    f  o   l   d  c   h  a  n  g  e   i  n   C   /  e  r   U   M   p  r  o  m  o   t  e  r  r 123456712 Scgb1a1MMP9Gsta 3 EMSA (fold change in C/EBP α binding to    T  r  a  n  s   f  e  c   t   i  o  n  a  c   t   i  v  a   t   i  o  n  o   f   M   o M over UM DNA) Fig. 3.  C/EBP α  activates methylated promoters in keratino-cytes. (  A ) Bisulfite sequencing of the cloned keratinocytespecific Bglap-rs1 proximal promoter before and after dif-ferentiation shows that it is predominantly methylated.Filled circles denote methylated CpG and open circlesdenote unmethylated CpG. Half-C/EBP ( − 37  bp) (1) andhalf-CRE ( − 417  bp) (2) sequences are marked. ( B ) Luciferaseactivities of Bglap-rs1 promoter (pCpGL-Bglap-rs1) cotrans-fected with CREB or C/EBP α  coding plasmid show that CREBactivates the unmethylated plasmid, whereas C/EBP α  acti-vates the methylated plasmid (T test,  p  <  0 . 05 ). Promoterswith the mutated half-C/EBP (1) or half-CRE (2) sequencehave reduced activity. All reporter activities are expressedrelative to the unmethylated plasmid. ( C  ) EMSA showingCREB and C/EBP α  (15-, 50-, 150-, and 500-nM dimer) bindingto the unmethylated and methylated half-C/EBP (1) orhalf-CRE (2) sequences from Bglap-rs1. ( D ) Mixture of CREBandC/EBP α (each 500-nMdimer)shows C/EBP α binds prefer-entially to the methylated half-CRE sequence even in thepresence of equimolar CREB. CREB and C/EBP α  cannot bindthe mutant half-CRE site of Bglap-rs1 ( SI Appendix  , Fig. S4 I  ).( E  ) Fold difference in C/EBP α  binding to methylated (M)compared to unmethylated (UM) DNA sequences deter-mined by EMSA versus fold difference in C/EBP α  transacti-vation of a methylated reporter compared to anunmethylated reporter. Seven sequences are plotted andshow a positive correlation; i.e., preferential binding tothemethylatedsequenceproducespreferentialtransactiva-tion of the methylated luciferase reporter. 20314  ∣  www.pnas.org/cgi/doi/10.1073/pnas.1008688107 Rishi et al.  tion (Fig. 4 and  SI Appendix , Fig. S8 C – S10). Dermal fibroblastsfrom newborn mice were differentiated into adipocytes andmRNA expression of   ∼ 200  genes with methylated promotersincreased by more than 2-fold (30). Global mRNA expressionfrom dermal fibroblasts was similar to expression in adipocytecultures differentiated from primary preadipocytes isolatedfrom adult mice inguinal adipose tissue confirming that thisunique differentiation system recapitulates adipocyte differentia-tion ( SI Appendix , Fig. S8 C ). The CpG methylation profile didnot markedly change with differentiation ( SI Appendix ,Fig. S8  D ) and was similar to what was observed in keratinocytes( SI Appendix , Fig. S8  E  and  F  ). Different methylated promoters were induced in adipocytes than in keratinocytes ( SI Appendix ,Fig. S8 G ). CpG demethylation by 5-aza treatment or DNMT1depletion ( SI Appendix , Fig. S9  A – C ) inhibited the differentiationof dermal fibroblasts into adipocytes as measured by Oil-Red Ostaining and global mRNA analysis (Fig. 4  A  and  B  and  SI  Appendix , Fig. S8  H  ). Suppression of C/EBP α  function by A-C/ EBP also inhibited differentiation (Fig. 4  A  and  B ). DNMT1 de-pletion or A-C/EBP almost completely inhibited differentiation,preventing us from observing a difference between the inhibitionof methylated and unmethylated promoters. The 5-aza effect wasmore modest and the methylated promoters were preferentially inhibited, suggesting that methylation is critical for their activa-tion. C/EBP α  localization to methylated promoters was inhibitedby CpG demethylation (Fig. 4 C ) as observed in keratinocytes.These data recapitulate what we observed in keratinocytes: Bothmethylation and C/EBP α  function are needed for differentiation,and C/EBP α  binding to methylated promoters is inhibited by de-methylation (Fig. 4 and  SI Appendix , Figs. S9  D – S10).We next examined Gpd1, a gene with a methylated promoterthat is induced during adipocyte differentiation (Fig. 4  D  and  SI  Appendix , Fig. S10  G  and  H  ). This promoter had the same fivecharacteristics we used to select keratinocyte specific promotersfor further studies. Bisulfite sequencing demonstrated that theGpd1 promoter is primarily methylated before and after differ-entiation (Fig. 4  D ). A luciferase reporter assay showed thatCREB can activate the unmethylated promoter, whereas C/EBP α can activate the methylated promoter. Again, the half-CRE in thepromoter was critical for C/EBP α  activation of the methylatedpromoter (Fig. 4  D ). EMSA also showed that C/EBP α  bindsthe half-CRE sequence better than CREB only when the se-quence is methylated (Fig. 4  D ). Thus, CpG methylation is criticalfor activation of a subset of adipocyte specific genes with methy-lated promoters.Classically, the observation is that tissue-specific genes haveunmethylated promoters in the tissue where they are expressed(31). We observed no change in methylation status of keratino-cyte specific genes following 2 d of keratinocyte differentiation.When we made the same comparison for fibroblasts differen- Fig. 4.  C/EBP α  binding activates methylated promo-ters in fibroblasts. (  A ) Oil-Red-O staining showing ac-cumulation of neutral lipids in newborn dermalfibroblast cultures 8 d after induction for adipogenicdifferentiation. Differentiation is inhibited by 5-azatreatment, depletion of DNMT1, or expression ofA-C/EBP, a dominant negative to C/EBP family mem-bers. ( B ) Number of UM and M promoters whosemRNA expression was doubled with differentiation( Upper  ). mRNA induction during differentiation ispreferentially inhibited (induction decreases  > 50% )for genes with methylated promoters by experimen-tal demethylation using either 5-aza (*  χ  2 test,  p  <  0 . 05 ), or siRNA to DNMT1 or A-C/EBP. ( C  ) Numberof UM and M promoters bound by C/EBP α  (bindingenrichment > 1 . 2 , T test,  p  <  0 . 03 ) in adipocytes. Logoshowing a C/EBP α  consensus site is the most enrichedsequence in these promoters. ( Lower  ) Decreased C/ EBP α  binding (< 0 . 8 ) after demethylation preferen-tially at methylated promoters by 5-aza (*  χ  2 test,  p  <  0 . 05 ) or siRNA to DNMT1. Both demethylationmethods decreased C/EBP α  binding to the samemethylated promoters (hypergeometric test,  p  <  0 . 02 ). Logos containing a CpG are the mostcommon DNA sequences occurring in promoterswith decreased C/EBP α  binding following either de-methylation method. ( D ) Bisulfite sequencing of theadipocyte specific promoter, Gpd1, showing it is pre-dominantly methylated before and after differentia-tion. Filled circles denote methylated CpG and opencircles denote unmethylated CpG. The half-CREsequence is marked by an arrow. Activity of a lucifer-ase reporter containing the Gpd1 promoter (pCpGL-Gpd1), either unmethylated or enzymatically methy-lated, transiently transfected into differentiatedfibroblasts. Cotransfection with expression plasmidsencoding CREB or C/EBP α  enhanced activity of un-methylated and methylated promoters, respectively(T test,  p  <  0 . 05 ). Mutation of half-CRE site reducedactivity of the promoter. EMSA showing CREB andC/EBP α  (15-, 50-, 150-, and 500-nM dimer) bindingto the unmethylated and methylated half-CRE se-quence from Gpd1 promoter. Mixture of CREB andC/EBP α  (150-nM dimer each) shows C/EBP α  binds pre-ferentially to the methylated sequence even in thepresence of equimolar CREB. Rishi et al. PNAS  ∣  November 23, 2010  ∣  vol. 107  ∣  no. 47  ∣  20315      B     I     O     C     H     E     M     I     S     T     R     Y
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks