doi:10.1016/j.jmb.2004.11.066 J. Mol. Biol. (2005) 346, 411–422

Transcriptional Regulator CTCF Controls HumanInterleukin 1 Receptor-associated Kinase 2 Promoter

Igor Kuzmin1*, Laura Geil1, Lauren Gibson2, Tiziana Cavinato2

Dmitry Loukinov3, Victor Lobanenkov3 and Michael I. Lerman2

1Basic Research ProgramSAIC-Frederick, Inc., FrederickMD 21702, USA

2Laboratory of ImmunobiologyNCI Frederick, Frederick, MD21702, USA

3Section of Molecular PathologyLIP, NIAID, NIH, RockvilleMD 20852, USA

0022-2836/$ - see front matter q 2004 E

Abbreviations used: 6FAM, 6-carChIP, chromatin immunoprecipitatielectrophoretic mobility shift assay;1 receptor associated protein kinaseMal/TIRAP, MyD88 adapter-like ordifferentiation factor 88; NF-kB, nucacetate; TAMRA, 6-carboxytetrametToll/IL-1 receptor; TLR, Toll-like reTNFR associated factor 6; YopP, YerE-mail address of the correspond

kuzmin@mail.ncifcrf.gov

Immune responses to invading pathogens are mediated largely through afamily of transmembrane Toll-like receptors andmodulated by a number ofdownstream effectors. In particular, a family of four interleukin 1 receptor-associated kinases (IRAK) regulates responsiveness to bacterial endotoxins.Pharmacological targeting of particular IRAK components may bebeneficial for treatment of bacterial infections. Here, we studied transcrip-tional regulation of the human IRAK2 gene. Analysis of the IRAK2promoter region reveals putative binding sites for several transcriptionalfactors, including ZIP (EGR1 and SP1), CTCF and AP-2beta. Deletion of theZIP or AP-2 sites did not significantly affect IRAK2 promoter activity innaıve and endotoxin-treated mononuclear cells, in dormant and activatedJurkat T-cells, in lung and kidney cells. In contrast, we found that CTCFplays a major role in IRAK2 transcription. An electrophoretic mobility shiftassay of the DNA fragments containing the IRAK2 CpG island, revealed asingle high-affinity binding site for the transcriptional regulator and achromatin insulator protein, CTCF. This assay revealed a CTCF-bindingsite within the mouse Irak2 promoter. The presence of the CTCF protein inhuman IRAK2 promoter was confirmed by chromatin immunoprecipita-tion assay. Specific residues that interacted with the CTCF protein, wereidentified by methylation interference assay. In all cell lines analyzed,including cells of lung, renal, monocytic and T-cell origin, the IRAK2luciferase reporter construct, containing an intact CTCF-binding site,showed strong promoter activity. However, IRAK2 promoter activity wasdecreased dramatically for the constructs with a mutated CTCF-bindingsite.

q 2004 Elsevier Ltd. All rights reserved.

Keywords: CTCF; IRAK2; Yersinia

*Corresponding author

Introduction

Immune responses to invading pathogens are mediated through a family of transmembrane Toll-likereceptors (TLR) that share a common motif with the Drosophila Toll receptor1 and are able to recognizevarious bacterial lipopolysaccharide endotoxins (LPS). These receptor-dependent signals converge on theadapter protein MyD88, which binds to the cytoplasmic receptor domain.1–8 MyD88 then recruits members

lsevier Ltd. All rights reserved.

boxyfluorescein; AP-1, activating protein 1; APPB, amyloid b-protein precursor;on assay; CTCF, CCCTC binding factor; DMS, dimethylsulfate; EMSA,IKK, IkB kinase; IL-1R, interleukin 1 receptor; IL-2, interleukin 2; IRAK2, interleukin2; LPS, lipopolysaccharide endotoxin; Mal, MyD88 adapter-like protein;Toll-IL1 receptor domain-containing adapter protein; MyD88, myeloidlear factor kB protein; PHA, phytohemagglutinin; PMA, phorbol 12-myristate 13-hylrhodamine; TF, transcription factor; TGF-b, transforming growth factor beta; TIR,ceptors; TLR4, toll-like receptor 4; TNFR, tumor necrosis factor receptor; TRAF6,sinia enterocolitica outer protein P; ZIP, zinc finger protein binding region.ing author:

412 CTCF and IRAK2 Promoter

of the IL-1R-associated kinase (IRAK) family to thereceptor complex. IRAK molecules: IRAK1, IRAK2,IRAK-M and IRAK49–13 can relay the NF-kB-activating signal via TRAF6 and TGF-b-activatedkinase 1 to the IKK complex, which ultimatelymediates NF-kB activation.1–5 In addition, the Mal/TIRAP adapter may signal TLR4-responsive NF-kBactivation independently from MyD88.14,15

It was shown that many members of IRAKfamily are regulated in response to LPSexposure. In particular, IRAK1 protein degradationand IRAK-Mup-regulationarewelldocumented.16–18

Mutations of IRAK-4 were shown to decreaseresponsiveness of patients to LPS and bacterialinfections.19 In contrast, the steady-state level ofIRAK2 protein does not change after exposure toLPS.16

However, IRAK2 was recently implicated in pro-apoptotic signaling in response to LPS exposure inthe course of Yersinia infections. A pro-apoptoticsignal elicited by bacterial infectionmay bifurcate atthe IRAK level.20 In LPS/YopP-treated mousemacrophages, IRAK1 and TRAF6 serve asmediators of macrophage survival upon LPS treat-ment. Inversely, the pro-apoptotic signal relayexploited by Yersinia enterocolitica YopP potentiallyinvolves MyD88 and IRAK2, which target the Fas-associated death domain protein—caspase 8apoptotic pathway. Dominant-negative MyD88 ordominant-negative IRAK2 diminishes LPS-inducedapoptosis in YopP-transfected macrophages,suggesting involvement of MyD88 and IRAK2 insignaling cell death. In contrast, dominant-negativeIRAK1 and TRAF6 did not provide protection butaugmented LPS-mediated apoptosis in the absenceof YopP. To add another level of complexity, themurine Irak2 is transcribed from two differentpromoters and encodes four alternatively splicedisoforms, two of which are inhibitory.21 Unlike itsmouse counterpart, human IRAK2 transcribed froma single promoter and does not produce splicevariants.

These data suggest that pharmacological down-regulation of the IRAK2 protein could provide

Figure 1. Bisulfite sequencingmethylation map of the IRAKlines. Methylated and unmethylated CpG dinucleotides aresequenced plasmid clone is represented by a single row betwfour PCR fragments). The region scanned for CTCF binding srepeats, which flank the IRAK2 CpG island, are designated byin red. The IRAK2 exon 1, including the first ATG codon is s

novel treatments for inflammation, sepsis andautoimmunity. Here, we describe isolation andanalysis of the human IRAK2 promoter and identifya multifunctional transcriptional factor, CTCF, as asingle critical factor indispensable for transcriptionof the IRAK2 gene.

Results

IRAK-2 exon 1 is located within the CpG islandand protected from DNA methylation

IRAK-2 is situated downstream from the VHLtumor suppressor gene, and both genes are tran-scribed from the same DNA strand. We first wantedto check if aberrant methylation of the VHL tumorsuppressor gene affected the methylation status ofIRAK-2. We studied two renal carcinoma cell lines:UMRC6 cells that express VHL and LAF cells wherethe VHL gene was silenced by methylation.

We analyzed the methylation status of 115 CpGdinucleotides in UMRC6 and LAF cells, startingfrom CpG NT 90,532 and ending at NT 92,258(GenBank accession number AC022383), bybisulfite sequencing. This contiguous sequenceincluded the entire IRAK2 CpG island, harboringexon 1 of the IRAK2 gene, and two Alu repeats,located immediately upstream and downstreamfrom the island. We found that IRAK2 exon 1 andthe GC-rich sequences flanking it were protectedfrom methylation, and this protected region wascentered in heavily methylated Alu repetitivesequences (Figure 1). Bisulfite sequencing revealednearly identical methylation patterns across thewhole analyzed area in both the UMRC6 and LAFcell lines.

The transcriptional start site for the humanIRAK2 gene was defined around nucleotide K9by 5 0-race.11,21 We reasoned that the IRAK2promoter should be located within the methyl-ation-protected region upstream of the transcrip-tion start site. Therefore, we delineated IRAK2

2 CpG island and flanking regions in UMRC6 and LAF cellshown as squares in green and yellow, respectively. Eacheen black vertical lines (five to nine clones for each of theites by EMSA is underlined in black. The positions of Alua blue bar. The position of the CTCF binding site is shownhown in brown.

Table 1. IRAK2 promoter deletion analysis

Promoter segmentLuciferase activity (% full

promoter activity)

K270 toC1 100.0G16.3K148 to C1 70.7G13.3K80 to C1 10.7G1.8K270 to K78 18.6G8.1PGL-3-basic 5.1G3.2

CTCF and IRAK2 Promoter 413

promoter sequences within 270 nucleotidesupstream of the transcription start site.

Deletion analysis of the IRAK2 promoter

We designed a set of luciferase reporter con-structs that included a “full promoter” construct(nucleotides K270 through C1) and severaldeletion constructs including nucleotides K148 toC1, K80 to C1 and K270 to K78. Deletion ofnucleotides K270 through K148 did not signifi-cantly affect activity of the promoter, while furtherdeletion of nucleotides K148 to K81 decreasedpromoter activity approximately fivefold. Deletionof the proximal region K77 to C1 decreasedactivity of full promoter twofold (Table 1).

Prediction of putative TF binding sites

We then analyzed putative transcription factorbinding sites within the full promoter usingWiconsin Package software (Accelrys, Inc., CA).

We identified a perfect 9 bp ZIP site22 betweennucleotidesK154 andK163 (Figure 2). An identicalsite was first found in the IL-2 promoter, where itplays an important role in IL-2 induction duringT-cell activation. The ZIP sequence (GGGGTGGAG)consists of overlapping sites for EGR-1 and Sp1factors, which bind to the ZIP sequence com-petitively. In the IL-2 promoter of activated T-cells,inducible EGR-1 substitutes constitutivelyexpressed Sp1.22 However, our Northern analysisindicated that transcription of IRAK2 did notchange in PHA-PMA-stimulated Jurkat T-cells orin LPS-treated mononuclear THP-1 cells (data notshown). Concurrently, the activity of ZIP-deleted

K148 toC1 luciferase reporter constructs in kidneyand T-cells did not change significantly when theputative ZIP site was eliminated (Table 1). There-fore, the biological significance of this sequenceremains unclear.Immediately downstream from ZIP we found an

element of a consensus binding sequence for thetranscription and insulation factor CTCF.23 It waslocated in positions K116 through K104(CNNNNNNNCCCTC) in a region critical forIRAK2 transcription (Figure 2).An AP-2beta binding site (Figure 2) was pre-

dicted in position K76 to K68 (SCCNNNGGC).24

Deletion of the region containing AP-2betadecreases IRAK2 promoter activity only slightly(Table 1). No other binding site was predicted in theregion K68 to C1.Other numerous short TF consensus sequences

were identified within and upstream of the CTCFbinding site. However, their biological significanceremains in question, since deletion of the regionbetween nucleotides K270 and K148 affects IRAK2promoter activity insignificantly.

Verification of CTCF binding to the IRAK-2promoter in vivo

We verified CTCF binding to the IRAK-2promoter in UMRC6 and LAF cells by anti-CTCFChIP followed by real-time PCR. Using twodifferent antibodies against CTCF, we achieved atleast 100-fold enrichment of the IRAK-2 promotersequences in both UMRC6 and LAF after-ChIPfractions, compared to the CTCF-negative sequenceof the VHL exon 3 (Figure 3). The actual enrichmentvalue may be even higher, because in mostexperiments we were not able to detect VHL exon3-specific DNA in anti-CTCF ChIP preparations.Both VHL exon 3 and IRAK2 promoter DNAsequences were equally detectable in UMRC6 andLAF “naked DNA” preparations.

EMSA identification of CTCF binding sites

Using electrophoresis mobility-shift assay(EMSA), we screened the methylation-protected

Figure 2. The sequence of theIRAK2 promoter. Predicted tran-scription binding sites are shownin frames. Contact guanine basesof the CTCF binding site aremarked by asterisks (*). Sequencesof BglI and RsrII restriction sitesare underlined. An arrow indicatesthe position of the transcriptionstart. The first ATG codon of theIRAK2 mRNA is shown in bold.

Figure 3. Enrichment of the IRAK2 promoter sequencesafter anti-CTCF ChIP using UMRC6 and LAF cell lines.

Figure 4. EMSA analysis of the methylation-protectedregion around IRAK2 exon 1. (a) EMSA of the IRAK2DNA fragments (from IRAK2-1 through IRAK2-14) andAPPB CTCF binding fragment. (b) Electrophoresis andethidium bromide staining of the EMSA fragments. (c)Competition of the 32P-labeled IRAK2 fragment 6 (3 ng)and different amounts of cold APPB fragment for CTCFbinding.

414 CTCF and IRAK2 Promoter

region for the presence of CTCF binding sites.We generated 14 overlapping fragments, about 100–120 bp each, covering the entire protected area(w1 kb) and parts of the adjacent methylatedflanking regions (Figure 1). For a positive control,we produced a fragment of the APPB promotercontaining a strong known CTCF binding site.25

We identified one high-affinity binding site nearthe 5 0 end of the IRAK-2 CpG island within thefragment IRAK2-6 (Figure 4). The predicted CTCFbinding site was located in the same fragment. Theamount of CTCF bound to IRAK2-6 and controlAPPB-binding sites was similar. To avoid missingother functional CTCF recognition sites at thefragment junctions, an additional set of fragmentswith overlaps of at least 100 was designed. EMSA ofthese genomic fragments did not reveal anyadditional CTCF binding site (data not shown).The results obtained using the 35S-labeled full-length CTCF protein were essentially identical withthose achieved with the DNA-binding domainalone (data not shown).

Specific CTCF binding to fragment 6 was demon-strated using labeled DNA fragment 6. Addition ofthe cold CTCF-specific APPB fragment abolishedCTCF binding to fragment 6 (Figure 4(c)).

CTCF methylation interference assay

To determine the contact guanine residues of theIRAK2 CTCF binding site experimentally, we con-ducted methylation interference assays using the32P-labeled fragment IRAK2-6 (see Materials andMethods). We did not detect any protected guaninebase on the anti-sense strand (data not shown) ofthe IRAK2-6 fragment. We found strong guaninemethylation interference in six guanine residues on

the sense strand (Figure 5). The same CTCF-bindingguanine bases were revealed using longer DNAtarget fragments, indicating that no additionalCTCF binding site exists in the vicinity of theIRAK2-6 fragment (data not shown). These sixcontact guanine bases (K124 to K115) overlappedwith the predicted CTCF binding site (K116 toK104; Figure 2).

Mutations of the CTCF binding site but not theAP-2beta site abolish IRAK2 promoter activity

Genomic sequences upstream of the IRAK2 exon1 were cloned into a promoter reporter plasmid,pGL3-basic, in front of the luciferase reading frameas described in Materials and Methods (plasmidpGL3-bas-IRAK2-CTCF-wt). These sequencesincluded nucleotides from C1 to K270, startingfromA (C1) in the first ATG codon of the publishedIRAK2 cDNA sequence (nucleotide ten in GenBankaccession number AF026273). The six contactguanine residues were located in positions K115,K118, K119, K120, K121 and K124. The reporterplasmid carrying the mutated CTCF-binding site(pGL3-bas-IRAK2-CTCF-mut1) was identical withpGL3-bas-IRAK2-CTCF-wt, with the exception of

CTCF and IRAK2 Promoter 415

three nucleotides: T instead of G in positions K118and K121, and C in place of G in position K120(Figure 6(a)). Another mutation was created bydigestion of pGL3-IRAK2-CTCFmut1 plasmid withBsiWI, fill-in with phage T4 DNA polymerase andre-ligation to produce construct pGL3-IRAK2-CTCFmut2. EMSA of the mutated DNA fragmentindicated no CTCF binding (Figure 6(c)).

We analyzed the impact of CTCF binding on thetranscription of IRAK2 gene in two clear cell renalcarcinoma cell lines, UMRC6 and LAF. In addition,we tested one lung carcinoma (A549) and twoleukemia cell lines: Jurkat T-cells and monocyticTHP-1. We compared firefly luciferase activity aftertransfection with pGL3-bas-IRAK2-CTCF-wt orpGL3-bas-IRAK2-CTCF-mut plasmids. In the cellstransfected with the construct carrying mutatedCTCF binding site, the firefly luciferase activitydecreased on average three to sevenfold. The mostdramatic, six to sevenfold, decrease of promoteractivity was found in LAF, UMRC6 and THP-1 cells(Figure 6(b)).

The activity of the reporter construct containingthe mutated AP-2beta-binding site did not differsignificantly from the wt full-length promoterconstruct (Figure 6(b)).

The CTCF binding site is present in the mouseIrak2 promoter

We obtained the Irak2 promoter sequence fromthe UCSC mouse genome browser†. We analyzednucleotides 114,159,457–114,160,111 comprisingexon 1 of the Irak2 cDNA (GenBank accessionnumber AY162378) and 500 nucleotides upstream.This sequence was obtained from mouse chromo-some 6 database assembled in May 2004. Transcrip-tion factor binding site analysis did not predictCTCF binding in this region. We then performedEMSA, which included genomic sequences (nucleo-tides K167 to C62) around the first ATG codon ofthe Irak2 mRNA (GenBank accession numberAY162378) where the A in ATG was assignednumber C1. This sequence was divided into threeoverlapping fragments (see Materials andMethods). EMSA revealed a strong CTCF bindingto fragments 1 and 2 (Figure 7(a)). Competitivebinding confirmed the presence of CTCF–DNAcomplexes (Figure 7(b)). To delineate CTCF-bindingsite(s) precisely, we performed methylation inter-ference assay by end-labeling the PCR primers usedfor synthesis of the fragments 1 and 2. Only thesense strand contained protected guanine bases(Figure 7(c)) located in the region of overlapbetween fragments 1 and 2 (positions K47 toK37). Sequence comparison between mouse andhuman IRAK2 CTCF binding sites revealed con-siderable conservation (Figure 7(d)).

In addition, a high-affinity binding site for EGR1in positions K141 to K133 (GGTCGGGTG) was

† http://genome.uscs.edu/

predicted by computer analysis. However, we didnot verify this site experimentally.Therefore, both mouse and human promoters

apparently include EGR1 and CTCF binding siteslocated in similar positions.

Discussion

In this study, we defined and analyzed thepromoter of the human IRAK2 gene. We firstdelineated DNA sequences protected from methyl-ation in IRAK2-expressing cell lines. Computationalanalysis revealed several putative TF-binding siteswithin the methylation protected upstream of theIRAK2 transcription start, including ZIP (Sp1 andEGR1), CTCF and AP-2beta sites.We studied the significance of the different

promoter regions in kidney, lung, monocytic andT-cell lineages. We found that the IRAK2 transcrip-tion rate does not change after T-cell activation byPHA and PMA, in spite of the presence of the ZIPsite. Moreover, the ZIP site seems to be dispensablefor IRAK2 transcription in all cell types analyzed.The putative AP-2beta binding site did not contri-bute significantly to the IRAK2 promoter activity. Incontrast, we found that deletion of the promotersegment, which contains the predicted binding sitefor CTCF, inhibits IRAK2 transcription dramatically.Presence of the CTCF protein within the IRAK2promoter in vivo was confirmed by anti-CTCF ChIPassay.ChIP assays usually identify proteins within a

0.5–1 kb DNA region but do not allow preciselocalization of protein binding site(s). CTCF is aversatile protein, which employs different combi-nations of its 11 zinc fingers to bind an extremelywide array of target sequences.26,27 Although noother CTCF binding site was predicted by computeranalysis of the IRAK2 locus, we could not excludethe possibility that other or additional sites couldcontribute to the observed CTCF-mediated chro-matin immunoprecipitation. Therefore, we ana-lyzed the entire sequence around the IRAK2 CpGisland. The predicted CTCF binding was confirmedby EMSA, and no additional site was found withinthis 1 kb region.CTCF contact guanine nucleotides were identi-

fied by methylation interference. The sequencecontaining contact guanine nucleotides overlappedwith the predicted CTCF binding site. We thendemonstrated that mutation of the CTCF contactguanine bases causes significant down-regulationof the IRAK2 promoter, while mutation of theAP-2beta site does not change promoter activity.In addition, we identified a CTCF binding site in

the mouse Irak2 promoter in a position similar tothat of the human CTCF element. These resultsimply that both promoters are evolutionarilyconserved.Apart from transcriptional activation of IRAK2,

CTCF may protect these GC-rich promoters frommethylation. This CTCF function can be

Figure 5. Methylation interference of the CTCF binding to the fragment IRAK2-6. Lane 1, the DNA G-sequencingladder obtained from the sense strand of fragment 6 bound to CTCF; lane 2, the sequencing ladder from unboundfragment 6. The corresponding DNA sequence is shown at the right. Six interfering methylated guanine residues aremarked with black diamonds.

416 CTCF and IRAK2 Promoter

demonstrated in future by knocking-out the CTCFbinding site in mice.

The CTCF protein is multi-functional. It plays amajor role in transcriptional activation of the APPBpromoter;25 silencing of c-myc and chicken lyso-zyme genes;26,28 insulation of chicken, mouse andhuman b-globin gene;29,30 insulation of DM1 myo-tonic dystrophy locus31 and imprinting control ofthe H19 region.32–35 CTCF also forms 5 0 boundariesof the human apolipoprotein B36 and c-myc genes.37,38

It was shown recently that CTCF is up-regulated inT-cells resistant to HIV-1 infection;39 however, itsrelevant target genes were not defined. HumanIRAK2 is the first gene implicated in immuneresponse that is controlled by CTCF.

It was suggested recently that down-regulation ofIRAK2 would suppress apoptosis of the hostmacrophages and could be beneficial for treatmentof Yersinia infections.20 Such down-regulation canbe achieved by selective targeting of the CTCF zinc

Figure 6. Activity of normal, AP-2beta-mutated and CTCF-mutatedIRAK2 promoter constructs indifferent cell lines. (a) A partialIRAK2 promoter sequence corre-sponding to the wild-type and themutated CTCF binding sites inreporter constructs. CTCF contactguanine nucleotides determined bymethylation interference are in boldand underlined, mutated nucleo-tides are shown in lower case.Nucleotides that belong to thepredicted CTCF consensusbinding sequence are underlined.(b) Activity of the wild-type and themutated IRAK2 promoter in differ-ent cell lines. The activity of thewild-type promoter is shown asopen rectangles. The activity of theconstructs carrying AP-2betamutation is shown by crosshatchedrectangles. The activity of con-structs carrying CTCF mutations 1and 2 are shown as black andhorizontally hatched rectangles,respectively. (c) EMSA analysis ofthe wild-type and mutated IRAK2CTCF binding sites.

CTCF and IRAK2 Promoter 417

fingers, which interact specifically with IRAK2promoter. The proof of principle of this therapeuticapproach that selectively targets zinc fingers of theDNA-binding proteins was demonstrated recentlyfor the estrogen receptor.40 The DNA-bindingability of the zinc fingers within the estrogenreceptor DNA-binding domain was shown to besusceptible to chemical inhibition by electrophilicdisulfide benzamide and benzisothiazolonederivates, which selectively block binding of theestrogen receptor to its responsive elements andsubsequent transcription. Under these conditions,the vulnerable C-terminal zinc finger was prefer-ably disrupted, suppressing estrogen receptor-mediated breast carcinoma progression. If futureexperiments verify that the IRAK2 protein isessential for promotion of apoptosis in multiple

cell types, specific inhibitors of CTCF–IRAK2interactions could be developed for treatment.

Materials and Methods

Methylation analysis of the IRAK2 locus

Analysis of methylation was performed using thebisulfite sequencing approach. A contiguous genomicregion including the IRAK2 CpG island and its flankinggenomic sequences (from nucleotides 90,415 to 92,370,GenBank accession number AC022383) was amplified asfour separate fragments by single or nested PCRs frombisulfite-treated genomic DNA. PCR amplification wasconducted using the FailSafe PCR System (EpicentreTechnologies, WI) and the following conditions: 94 8C, 30seconds; 58 8C, 30 seconds; 72 8C, 30 seconds; 35 cycles.

Figure 7. CTCF binding to the mouse Irak2 promoter. (a) Upper panel: EMSA of the Irak2 promoter region. Lowerpanel: electrophoresis and ethidium bromide staining of the EMSA fragments. (b) Competition of the 32P-labeled Irak2fragment 2 (3 ng) and different amounts of cold APPB fragment for CTCF binding. (c) Methylation interference of theCTCF binding to mouse fragment 2. Lane 1, the DNA G sequencing ladder obtained from the sense strand of thefragment 2 bound to CTCF; lane 2, the sequencing ladder from the unbound fragment 2. The corresponding DNAsequence is shown on the left. Seven methylated contact guanine residues are marked with black diamonds. (d) Thesequence alignment of the human and mouse IRAK2 CTCF binding sites.

418 CTCF and IRAK2 Promoter

The DNA bisulfite treatment procedure was asdescribed.41 Primers were as follows: fragment 1-BIS-81D (TAGTTTTTTTTTTTATATTTAGATTGTAAG) andBIS-81R (CTCATCAAATACTCCATTAATTACC), nestedPCR-BIS-82D (TATGAAAGTAAGGGTTTTTGTAGGT)and BIS-82R (CTATATACCAAACAACTCCCTAAAT);fragment 2-BIS-83D (ATTTAGGGAGTTGTTTGGTATATAG) and BIS-83R (ACACAAATCRTCCAACACCCAAAA), nested PCR-BIS-84D (GGTAATTAATGGAGTATTTGATGAG) and BIS-84R (ACTAATAAATATAACAAACCATAACAC); fragment 3-BIS-85D (GTGTTATGGTTTGTTATATTTATTAGT) and BIS-85R (AAATCATAAATCCTTCTAATATACACT); fragment 4-BIS87D(GGTGTTGTTATTTTTTGGGGGG) and BIS-87R (CACATCTAATCTCAAAACAAATTAATA),nestedPCR-BIS-88D(AGTGTATATTAGAAGGATTTATGATTT) and BIS-88R(AATACTATAATCAAAACAATATACACTA). The finalconcentration of primers in the PCR reaction was 200 nM.PCR fragments were sub-cloned into pCR2.1-TOPOplasmid (Invitrogen, CA) and from five to nine clonesfor each fragment were sequenced.

Mobility-shift DNA-binding assay

35S-labeled CTCF proteins for electrophoretic mobility-shift assays were produced using a TnT phage T7 coupledreticulocyte lysate system (Promega, WI) and CTCF-encoding plasmids under phage T7 promoter, accordingto the manufacturer’s instructions. The plasmid pET 7.1contains the full-length CTCF cDNA and pET zf11encoded 11-Zn-finger CTCF DNA-binding domain.33

PCR (94 8C, 30 seconds; 72 8C, 30 seconds; 35 cycles)was conducted using the FailSafe PCR System (EpicentreTechnologies, WI). EMSADNA fragments were producedby PCR using the primers and FailSafe premixesdescribed in Table 2. Positive control fragments of similarsize were amplified by nested PCR from HeLa genomicDNA using primers specific for the APPB genepromoter.25 For the first PCR in FailSafe premixes D orH, we used primers APPB-1D (GCGGGGCTCAGAGCCAGGCGAGT) and APPB-1R (AGGCGGCGCCAAGGGCGCTGCA); nested PCR was performed in FailSafe pre-mix D with primers APPB-2D (CAGCTGATCCGGCC

CTCF and IRAK2 Promoter 419

CACCCCGCT) and APPB-2R (TGTGGGCGCGGGGCGCGAGGG). Each EMSA reaction was carried out in10 ml of phosphate-buffered saline (pH 7.4) containing5 mM MgCl2, 0.1 mM ZnSO4, 1 mM DTT, 0.1% (w/v)Igepal CA-630 (Sigma, MO), 10% (v/v) glycerol, 150 ng ofdouble-stranded poly[(dI)–(dC)] (Amersham-PharmaciaBiotech, NJ), 3 ng of target DNA fragment and 1 ml of theappropriate 35S-labeled CTCF TnT T7 coupled reticulo-cyte lysate mix. After 0.5–one hour of incubation at roomtemperature, the assay mixes were run on 6% NovexDNA retardation gels (Invitrogen, CA). The gels werefixed in 45% (v/v) methanol and 10% (v/v) glacial aceticacid for at least one hour, dried and autoradiographed.Fragments of mouse Irak2 promoter were produced

from mouse genomic DNA by PCR using the followingprimers: mIRAK2-D1 (TGGTCTACTCCTGACACTGGTCCAGG) and mIRAK2-R1 (TGGGTAAGCGCGCAAGCTCCCCCGA) for fragment 1; mIRAK2-D2(TGAATGCCACTGGGAGGCGCTCGCT) and mIRAK2-R2 (GACTGAGCCTGTGGCGGAGGGCC) for fragment 2;and mIRAK2-D3 (TCGGGGGAGCTTGCTCGCTTACCCA) and mIRAK2-R3 (TCGATATTGCGGCACAGGTCGTCCAG) for fragment 3.For mobility-shift competition experiment, human

fragment 6 and mouse fragment 2 were labeled asdescribed in the next section.

Methylation interference assay

To produce fragments labeled with 32P at one end thatcontain the putative CTCF binding sites, we conductedPCR with primers IRAK2 6D, IRAK2 6R, mIRAK2-D1,mIRAK2-R1, mIRAK2-D2 or mIRAK2-R2 labeled usingphage T4 kinase (Invitrogen, CA) according to themanufacturer’s protocol. The fragments were purifiedusing ProbeQuant G-50 Micro Columns (Amersham-Pharmacia Biotech, NJ). 200 ml of DMS reaction buffer(50 mM sodium cacodilate (pH 8.0), 1 mM EDTA) and

Table 2. Primers and DNA fragments used for IRAK2 EMSA

IRAK2 fragment Primer name Sequen

IRAK2-1 IRAK2 1D ggcgtgIRAK2 1R caggag

IRAK2- 2 IRAK2 2D ttggccaIRAK2 2R aaaccc

IRAK2-3 IRAK2 3D ccaccgIRAK2 3R catgcat

IRAK2-4 IRAK2 4D ggagtaIRAK2 4R cgggaa

IRAK2-5 IRAK2 5D cctgccgIRAK2 5R cccagc

IRAK2-6 IRAK2 6D tccgattIRAK2 6R gcgggg

IRAK2-7 IRAK2 7D cccgccIRAK2 7R gtcccgg

IRAK2-8 IRAK2 8D acccagIRAK2 8R ggcaca

IRAK2-9 IRAK2 9D cctgggIRAK2 9R tcccctc

IRAK2-10 IRAK2 10D gtgagtIRAK2 10R aacccg

IRAK2-11 IRAK2 11D caagctcIRAK2 11R gcacctt

IRAK2-12 IRAK2 12D acctcggIRAK2 12R gtaccca

IRAK2-13 IRAK2 13D ctctgagIRAK2 13R gttgtgg

IRAK2-14 IRAK2 14D ggtggcIRAK2 14R cccaag

1 ml of DMS were added to 5–10 ml of the fragment (w106

cpm) and incubated at room temperature. After fiveminutes, 40 ml of DMS stop buffer (1.5 M sodium acetate(pH 7.0), 1 M 2-mercaptoethanol), 1 ml of 10 mg/ml oftRNA (Sigma, MI) and 600 ml of 100% ethanol wereadded. The fragments were precipitated after incubationin a bath of solid CO2/ethanol for ten minutes. The pelletwas re-suspended in 250 ml of 0.3 M sodium acetate,1 mM EDTA on ice followed by addition of 750 ml ofethanol. After repeated precipitation in ethanol, the pelletwas re-suspended in TE buffer at w20,000 cpm/ml.For each 32P-labeled fragment, ten preparative EMSA

binding reactions were assembled. Nine reactions con-tained 2!105 cpm each of the 32P-labeled, DMS-treatedDNA fragment IRAK2-6, instead of the unlabeled frag-ment used for EMSA. The tenth (control) reactioncontained the labeled target fragment, which was nottreated with DMS. Each reaction also included 5 ml of TnTCTCF mix (unlabeled), in place of 1 ml of the 35S-labeledCTCF TnT mix.The reactions were run on a 6% ten-well Novex DNA

retardation gel. The gel was autoradiographed for two tosix hours at 4 8C. The bands containing CTCF/DNAcomplex or unbound DNA were excised separately, andDNAwas eluted at 37 8C for four hours into 10 mM Tris–HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA. To remove salt,DNA fragments were re-dissolved in 0.3 M sodiumacetate (pH 7.0) and re-precipitated in ethanol at leastthree times. Finally, DNA was re-suspended in 100 ml of1 M piperidine. Methylated guanine bases were cleavedat 90–95 8C for 30 minutes, and the piperidine solutionwas evaporated under vacuum. The DNA pellet was re-dissolved in 100 ml of water and vacuum-dried threemoretimes. DNA was then dissolved in formamide loadingbuffer (80% (v/v) formamide, 10 mM NaOH, 1 mMEDTA, 0.1% (w/v) xylene cyanol FF and 0.1% (w/v)bromphenol blue). Equal numbers of counts from thesamples corresponding to the CTCF/DNA complex and

ce Buffer

cgccaccgtgcccggc Gttcgagactagcctggccaaggctagtctcgaactcctg Gtacagccgggcgcggtgcgcccggctgtagggttt Acatgcgctcatcaaatactcctttgatgatgagcgcatgatgcatg Bttctgccgcagcggcaggctgcggcagaattcccg Cgccggccagagaatcggactctggccggcgctggg Gaagtgggcggggcgggccgcccacttccccgc Bcgcgggacgactgggttcgtccgcgccggagc Agctcgtccagcaccaggtgctggacgacctgtgcc Acccgggccgcactcacgcggcccggggagggga Aggtgcccgagggagcttgcctcgggcacccgggtt Bcaggtgacaggccgaggtcctgtcacctgaaggtgc Bccgttctctggactcagagtccagagaacggtgggtac Bgctggatgagaggccaccctctcatccagcccacaac Baagtgacagcaccgtgggt

420 CTCF and IRAK2 Promoter

free DNA were run concurrently on a 6% or 12%polyacrylamide/urea sequencing gel. The gel was fixedin 5% (v/v) acetic acid/5% (v/v) methanol, dried andautoradiographed.

CTCF chromatin immunoprecipitation assay

Anti-CTCF ChIP was performed using a ChromatinImmunoprecipitation Assay kit (Upstate, NY) and anti-CTCF antibodies sc-15914 (Santa Cruz Biotechnology,Inc., CA) or 06-917 (Upstate, NY) according to themanufacturer’s instructions. The content of CTCF-immunoprecipitated DNA was analyzed by real-timePCR using the iCycler iQ Real-Time detection system(Bio-Rad Laboratories, CA). Amplification was con-ducted using FailSafe PCR System (Epicentre Tech-nologies, WI) and consisted of three minutesdenaturation at 95 8C followed by 40 cycles of denatura-tion for 15 seconds at 95 8C and 30 seconds extension at60 8C. The reactions were performed in FailSafe buffer Kfor the IRAK2 primers and buffer D for the VHL primers.Primers used for detection of the IRAK2 promoter DNAwere RT/IL1RK-D (GGCCTGCTACATCTACCAGC),RT/IL1RK-R (CGAACTCCATCCAGTCCC) and theprobe oligonucleotide was RT/IL1RK-P ([6FAM]-CCCTCCTGGGTGCTGGACGACCTG-[TAMRA]). Fordetection of VHL exon 3 sequences, we used primersRT360-D (AAGAAGGCATTGGCATCTG), RT360-R(TCACGGATGCCTCAGTCTT) and the probe oligo-nucleotide RT360-P ([6FAM]-AGCGGTTGGTGACTTGTCTGCCTCCTG-[TAMRA]). Each experiment con-sisted of six PCR reactions. For each IRAK2 and VHLtarget, three reactions were performed: two reactionsincluded ChIP-template DNA and one served as negativecontrol (without DNA). The amount of the IRAK2promoter-specific DNA sequences in anti-CTCF ChIPpreparations was compared to that of the VHL exon3-specific DNA.

Luciferase reporter assays

The IRAK2 promoter region was amplified by PCR(95 8C, 30 seconds; 72 8C, 30 seconds; 35 cycles) usingFailSafe buffer D (Epicentre Technologies,WI), fromcosmid 3 containing exon 1 of the IRAK2 gene.42 The“full” IRAK2 promoter was amplified using primersMluI/IRAK2-D (AAAAACGCGTGGAAGTGGGGTCAGAAGGCCTCCTGC) and BglII/IRAK2-R (AAAAAGATCTGGCACGCTACGGGGCCGGCTCCG), whichincluded MluI and BglII restriction sites for cloning intopGL3-basic luciferase reporter vector (Promega, WI) toproduce pGL3-bas-IRAK2-CTCF-wt. The reporter plas-mid containing the IRAK2 promoter carrying a mutatedCTCF binding site was constructed in two steps. First, thepart of the promoter downstream from the CTCF bindingsite was amplified using primers MluI/BsiWI/IRAK2-D(AAAAACGCGTAAAACGTACGCTGGTGCCCTCCCGCCCCGC) and BglII/IRAK2-R and sub-cloned intoMluI/BglII sites of the pGL3-basic plasmid to produceplasmid pGL3-bas-BsiWI/BglII. Second, the upstreampart of the IRAK2 promoter was amplified using theprimers MluI/IRAK2-D and BsiWI/IRAK2-R (AAAACGTACGAAGCGGCCACGGGCGGCACGACGG).The primers MluI/BsiWI/IRAK2-D and BsiWI/IRAK2-Rhad BsiWI restriction sites in place of the CTCF bindingsite. Using this feature, the upstream part of the mutatedIRAK2 promoter was cloned into MluI/BsiWI sites of thepGL3-bas-BsiWI/BglII to produce the plasmid pGL3-bas-IRAK2-CTCFmut1.

The plasmid containing nucleotides K148 to K1 of theIRAK2 promoter was constructed by partial digestion ofthe pGL3-bas-IRAK2-CTCF-wt with BglI, digestion withMluI and re-ligation. The plasmid containing nucleotidesK80 to K1 was prepared by digestion of pGL3-bas-IRAK2-CTCF-wt with RsrII/MluI and re-ligation. Theplasmid containing nucleotides K270 to K78 was madeby digestion of pGL3-bas-IRAK2-CTCF-wt with RsrII/BglII and re-ligation. The AP2-minus construct wasdesigned using MluI/IRAK2-D and BglII/AP2/IRAK2-R primers. The downstream fragment was amplifiedusing BglII/AP2/IRAK2-D primer (AAAAAGATCTAGAGAAGCCGCAGCCCGCAGT) and HindIII/IRAK2-Rprimer (AAAAAGATCTGGCACGCTACGGGGCCGGCTCCG), cut with BglII/HindIII and subcloned intoBglII/HindIII sites of the AP2-minus construct.The clear cell renal carcinoma cell-line UMRC6 was as

described.43 The LAF cell line was established afterexplantation of a clear cell renal carcinoma. Lungcarcinoma A549 and acute T-cell leukemia (Jurkat) werepurchased from ATCC, VA. The acute monocyticleukemia (THP-1) cell line was provided by Dr T.Yoshimura.Each reporter luciferase assay was performed in a

minimum of four transfection reactions, each conductedin one well of a 24-well cell culture cluster. SyntheticRenilla luciferase reporter vector phRL-TK (Promega, WI)was used for normalization of the transfection efficiency.Lipofectamine 2000 transfections were performed accord-ing to the manufacturer’s instructions. For each well, weused 1 ml of Lipofectamine 2000 (Invitrogen, CA) and 1 mgof pGL3-bas-CTCF(wt or mut)/phRL-TK mix (in the ratio10 : 1, w/w). Luciferase activity was measured using theDual Luciferase Reporter Assay System (Promega, WI)and MicroLumat Plus LB 96V microplate luminometer(Berthold Technologies GmbH and Co.) according to themanufacturer’s instructions.Treatments of Jurkat T-cells by PHA-PMA and THP-1

cells by LPS were as described.16,22 Northern analysis ofIRAK2 expression was performed using 1 kb IRAK2 ORFas a probe.

Acknowledgements

The content of this publication does not necess-arily reflect the views or policies of the Departmentof Health and Human Services, nor does mention oftrade names, commercial products, or organizationsimply endorsement by the U. S. Government. Wethank Dr David Symer for technical help with theiCycler iQ Real-Time detection system. We aregrateful to Dr Teizo Yoshimura for THP-1 cells.This work has been funded, in part, with Federalfunds from the National Cancer Institute, NationalInstitutes of Health, under contracts NO1-CO-12400and NO1-CO-56000.

References

1. Zhang, G. & Ghosh, S. (2001). Toll-like receptor-mediated NF-kB activation: a phylogeneticallyconserved paradigm in innate immunity. J. Clin.Invest. 107, 13–19.

CTCF and IRAK2 Promoter 421

2. Aderem, A. & Ulevitch, R. J. (2000). Toll-like receptorsin the induction of the innate immune response.Nature, 406, 782–787.

3. Anderson, K. V. (2000). Toll signaling pathways in theinnate immune response. Curr. Opin. Immunol. 12,13–19.

4. Brightbill, H. D. & Modlin, R. L. (2000). Toll-likereceptors: molecular mechanisms of the mammalianimmune response. Immunology, 101, 1–10.

5. Medzhitov, R. & Janeway, C. (2000). The toll receptorfamily and microbial recognition. Trends Microbiol. 8,452–456.

6. Dziarski, R., Wang, Q., Miyake, K., Kirschning, C. J. &Gupta, D. (2001). MD-2 enables Toll-like receptor 2(TLR2)-mediated responses to lipopolysaccharideand enhances TLR2-mediated responses to Gram-positive and Gram-negative bacteria and their cellwall components. J. Immunol. 166, 1938–1944.

7. Zhang, F. X., Kirschning, C. J., Mancinelli, R., Xu, X. P.,Jin, Y., Faure, E. & Mantovani, A. (1999). Bacteriallipopolysaccharide activates nuclear factor-kappaBthrough interleukin-1 signaling mediators in culturedhuman dermal endothelial cells and mononuclearphagocytes. J. Biol. Chem. 274, 7611–7614.

8. Yang, R. B., Mark, M. R., Gurney, A. L. & Godowski,P. J. (1999). Signaling events induced by lipopoly-saccharide-activated Toll-like receptor 2. J. Immunol.163, 636–643.

9. Croston, G. E., Cao, Z. & Goeddel, D. V. (1995). NF-kBactivation by interleukin-1 (IL-1) requires an IL-1receptor associated protein kinase activity. J. Biol.Chem. 270, 16514–16517.

10. Cao, Z., Henzel, W. J. & Gao, X. (1996). IRAK: a kinaseassociated with the interleukin-1 receptor. Science,271, 1128–1131.

11. Muzio, M., Ni, P., Feng, P. & Dixit, V. M. (1997). IRAK(Pelle) family members IRAK-2 and MyD88 asproximal mediators of IL-1 signaling. Science, 278,1612–1615.

12. Wesche, H., Gao, X., Li, X., Kirschning, C. J., Stark,G. R., Cao, Z. & Matsumoto, K. (1999). IRAK-M is anovel member of the pelle/interleukin-1 receptor-associated kinase (IRAK) family. J. Biol. Chem. 274,19403–19410.

13. Li, S., Strelow, A., Fontana, E. J. & Wesche, H. (2002).IRAK-4: a novel member of the IRAK family with theproperties of an IRAK-kinase. Proc. Natl Acad. Sci.USA, 99, 5567–5572.

14. Fitzgerald, K., Palsson-McDermott, E. M., Bowie,A. G., Jefferies, C. A., Mansell, A. S., Brady, G. et al.(2001). Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction. Nature, 413, 78–83.

15. Horng, T., Barton, G. M. & Medzhitov, R. (2001).TIRAP: an adapter molecule in the Toll signalingpathway. Nature Immunol. 2, 835–841.

16. Hu, J., Jacinto, R., McCall, C. & Li, L. (2002).Regulation of IL-1 receptor-associated kinases bylipopolysaccharide. J. Immunol. 168, 3910–3914.

17. Escoll, P., del Fresno, C., Garcia, L., Valles, G.,Lendinez, M. J., Arnalich, F. & Lopez-Collazo, E..(2003). Rapid up-regulation of IRAK-M expressionfollowing a second endotoxin challenge in humanmonocytes and monocytes isolated from septicpatients. Biochem. Biophys. Res. Commun. 311, 465–472.

18. Noubir, S., Hmama, Z. & Reiner, N. E. (2004). Dualreceptors and distinct pathways mediate interleukin 1receptor associated kinase degradation in response tolipopolysaccharide. J. Biol. Chem. 279, 25189–25195.

19. Picard, C., Puel, A., Bonnet, M., Ku, C. L., Bustamante,

J., Yang, K. et al. (2003). Pyogenic bacterial infectionsin human with IRAK-4 deficiency. Science, 299,2076–2079.

20. Ruckdeschel, K., Mannel, O. & Schrottner, P. (2002).Divergence of apoptosis-inducing and preventingsignals in bacteria-faced macrophages throughmyeloid differentiation factor 88 and IL-1 receptor-associated kinase members. J. Immunol. 168,4601–4611.

21. Hardy, M. P. & O’Neil, L. A. (2004). The murineIRAK2 gene encodes four alternatively spliced iso-forms, two of which are inhibitory. J. Biol. Chem. 279,27699–27708.

22. Skerka, C., Decker, E. L. & Zipfel, P. F. (1995). Aregulatory element in the human interleukin 2 genepromoter is a binding site for the zinc finger proteinSp1 and EGR-1. J. Biol. Chem. 270, 2250–22506.

23. Klenova, E. M., Nicolas, R. H., Paterson, H. F., Carne,A. F., Heath, C. M., Goodwin, G. H. et al. (1993). CTCF,a conserved nuclear factor required for optimaltranscriptional activity of the chicken c-myc gene isan 11-Zn-finger protein differentially expressed inmultiple forms. Mol. Cell. Biol. 13, 7612–7624.

24. Moser, M., Imhof, A., Pscherer, A., Bauer, R.,Amselgruber, W., Sinowatz, F. et al. (1995). Cloningand characterization of a second AP-2 transcriptionfactor: AP-2 beta. Development, 121, 2779–2788.

25. Vostrov, A. A. & Quitschke, W. W. (1997). The zincfinger protein CTCF binds to the APBb domain of theamyloid b-protein precursor promoter. J. Biol. Chem.272, 33353–33359.

26. Filippova, G. N., Fagerlie, S., Klenova, E. M., Myers,C., Dehner, Y., Goodwin, G. et al. (1996). Anexceptionally conserved transcriptional repressor,CTCF, employs different combinations of zinc fingersto bind diverged promoter sequences of avian andmammalian c-myc oncogenes. Mol. Cell. Biol. 16,2802–2813.

27. Ohlsson, R., Renkawitz, R. & Lobanenkov, V. (2001).CTCF is a uniquely versatile transcription regulatorlinked to epigenetics and disease. Trends Genet. 17,520–527.

28. Lobanenkov, V. V., Nicolas, R. H., Adler, V. V.,Paterson, H., Klenova, E. M., Polotskaja, A. V. &Goodwin, G. H. (1990). A novel sequence-specificDNA binding protein which interacts with threeregularly-spaced direct repeats of the CCCTC-motifin the 5 0-flanking sequence of the chicken c-myc gene.Oncogene, 5, 1743–1753.

29. Bell, A. C., West, A. G. & Felsenfeld, G. (1999). Theprotein CTCF is required for the enhancer blockingactivity of vertebrate insulators. Cell, 98, 387–396.

30. Farrell, C. M., West, A. G. & Felsenfeld, G. (2002).Conserved CTCF insulator elements flank the mouseand human beta-globin loci. Mol. Cell. Biol. 22,3820–3831.

31. Filippova, G. N., Thienes, C. P., Penn, B. H., Cho,D. H., Hu, Y. J., Moore, J. M. et al. (2001). CTCF-binding sites flank CTG/CAG repeats and form amethylation-sensitive insulator at the DM1 locus.Nature Genet. 28, 335–343.

32. Szabo, P., Tang, S. H., Rentsendorj, A., Pfeifer, G. P. &Mann, J. R. (2000). Maternal-specific footprints atputative CTCF sites in the H19 imprinting controlregion give evidence for insulator functions. Curr.Biol. 10, 607–610.

33. Kanduri, C., Pant, V., Loukinov, D., Pugacheva, E., Qi,C. F., Wolffe, A. et al. (2000). Functional association of

422 CTCF and IRAK2 Promoter

CTCF with the insulator upstream of the H19 gene isparent of origin-specific and methylation-sensitive.Curr. Biol. 10, 853–856.

34. Hark, A. T., Schoenherr, C. J., Katz, D. J., Ingram, R. S.,Levorse, J. M. & Tilghman, S. M. (2000). CTCFmediates methylation-sensitive enhancer-blockingactivity at the H19/Igf2 locus. Nature, 405, 486–489.

35. Bell, A. C. & Felsenfeld, G. (2000). Methylation of aCTCF-dependent boundary controls imprintedexpression of the Igf2 gene. Nature, 405, 482–485.

36. Antes, T. J., Namciu, S. J., Fournier, R. E. & Levy-Wilson, B. (2001). The 5 0 boundary of the humanapolipoprotein B chromatin domain in intestinal cells.Biochemistry, 40, 6731–6742.

37. Gombert, W. M., Farris, S. D., Rubio, E. D., Morey-Rosler, K. M., Schubach, W. H. & Krumm, A. (2003).The c-myc insulator element and matrix attachmentregions define the c-myc chromosomal domain. Mol.Cell. Biol. 23, 9338–9348.

38. Lutz, M., Burke, L. J., LeFevre, P., Myers, F. A., Thorne,A. W., Crane-Robinson, C. et al. (2003). Thyroidhormone-regulated enhancer blocking: cooperationof CTCF and thyroid hormone receptor. EMBO J. 22,1579–1587.

39. Kartvelishvili, A., Lesner, A., Szponar, M. & Simm, M.(2004). Microarray analysis of differentially expressedgenes in cells resistant to HIV-1. Immunol. Letters, 93,79–86.

40. Wang, L. H., Yang, X. Y., Zhang, X., Mihalic, K., Fan,Y. X., Xiao, W. et al. (2004). Suppression of breastcancer by chemical modulation of vulnerable zincfingers in estrogen receptor. Nature Med. 10, 40–47.

41. Kuzmin, I., Geil, L., Ge, H., Bengtsson, U., Duh, F.-M.,Stanbridge, E. J. & Lerman, M. I. (1999). Analysis ofaberrant methylation of the VHL gene by transgenes,monochromosome transfer, and cell fusion. Oncogene,18, 5672–5679.

42. Kuzmin, I., Stackhouse, T., Latif, F., Duh, F.-M., Geil,L., Gnarra, J. et al. (1994). One megabase yeastartificial chromosome and 400-kilobase cosmid-phage contigs containing the von Hippel-Lindautumor suppressor and Ca2C transporting adenosinetriphosphatase isoform 2 genes. Cancer Res. 54,2486–2491.

43. Grossman, H. B., Wedemeyer, G. & Ren, L. Q. (1985).Human renal carcinoma: characterization of five newcell lines. J. Surg. Oncol. 28, 237–244.

Edited by M. Yaniv

(Received 10 July 2004; received in revised form 17 November 2004; accepted 23 November 2004)