Nucleosome mapping across the CFTR locus identifies novel regulatory factors.

0.69 MB·English

by  Yigit, Erbay

Published online 15 January 2013 Nucleic Acids Research, 2013, Vol. 41, No. 5 2857–2868
doi:10.1093/nar/gks1462
Nucleosome mapping across the CFTR locus
identifies novel regulatory factors
Erbay Yigit3, Jared M. Bischof1,2, Zhaolin Zhang1,2, Christopher J. Ott1,2,
Jenny L. Kerschner1,2, Shih-Hsing Leir1,2, Elsy Buitrago-Delgado3, Quanwei Zhang4,
Ji-Ping Z. Wang4, Jonathan Widom3,y,z and Ann Harris1,2,*
1Human Molecular Genetics Program, Children’s Memorial Research Center, 2Department of Pediatrics,
Northwestern University Feinberg School of Medicine Chicago, IL 60614, USA, 3Department of Molecular
Biosciences and 4Department of Statistics, Northwestern University, Evanston, IL 60208, USA
Received July 31, 2012; Revised December 12, 2012; Accepted December 14, 2012
ABSTRACT glucocorticoid receptor as a novel trans-acting
factor that regulates CFTR expression in vivo.
Nucleosome positioning on the chromatin strand
plays a critical role in regulating accessibility of
DNA to transcription factors and chromatin modify- INTRODUCTION
ing enzymes. Hence, detailed information on nu-
Recent progress in generating maps of open chromatin
cleosome depletion or movement at cis-acting genome wide (1,2) has greatly accelerated the discovery
regulatoryelementshasthepotentialtoidentifypre- of regulatory mechanisms both for individual genes and
dictedbindingsitesfortrans-actingfactors.Usinga for coordinated transcriptional networks. However,
novel method based on enrichment of mononu- advances in the technologies that enable de novo charac-
cleosomal DNA by bacterial artificial chromosome terization of trans-acting factors that interact with these
hybridization, we mapped nucleosome positions by regions of open chromatin are still in development (3–5).
Until recently, prediction of transcription factor binding
deep sequencing across 250 kb, encompassing the
sitesingenomicDNAwaslargelydependentontheuseof
cystic fibrosis transmembrane conductance regula-
in silico approaches followed by in vivo verification by
tor (CFTR) gene. CFTR shows tight tissue-specific
chromatin immunoprecipitation (ChIP) using specific
regulation of expression, which is largely deter-
antibodies for individual factors. As the regions of open
mined by cis-regulatory elements that lie outside
chromatin identified by DNase-seq (DNase I hypersensi-
the gene promoter. Although multiple elements are tivity mapping followed by deep sequencing) and FAIRE
known, the repertoire of transcription factors that (formaldehyde-assisted identification of regulatory
interact with these sites to activate or repress elements) can encompass >1kb of DNA sequence, it
CFTR expression remains incomplete. Here, we remains a challenge to characterize the cis-regulatory
show that specific nucleosome depletion corres- elementsthatareresponsibleforthechromatinlandscape.
ponds to well-characterized binding sites for One approach that can assist in this effort is to fine map
the location of nucleosomes across the region of open
known trans-acting factors, including hepatocyte
chromatin, as nucleosome positioning is known to play
nuclear factor 1, Forkhead box A1 and CCCTC-
a critical role in accessibility of the DNA strand to inter-
binding factor. Moreover, the cell-type selective
acting proteins (6,7). The dynamic competition between
nucleosome positioning is effective in predicting
DNA unwinding, nucleosome positioning and transcrip-
binding sites for novel interacting factors, such
tion factor binding is not yet fully understood (8,9).
as BAF155. Finally, we identify transcription fac- However, there are many cases, for example, the
tor binding sites that are overrepresented in androgen receptor (4,10), where loss of a specific nucleo-
regions where nucleosomes are depleted in a somecanbedirectlycorrelatedwithoccupancyofaregu-
cell-specific manner. This approach recognizes the latory element by one or more of its cognate protein
*To whom correspondence should be addressed. Tel:[phone]; Fax:[phone]; Email: [email protected]
Present address:
Christopher J. Ott, Dana-Faber Cancer Institute, Harvard Medical School, 450 Brookline Ave, Dana 510D, Boston, MA 02215, USA.
yDeceased.
zThis work is dedicated to Jon Widom, an inspiring colleague who is greatly missed.
(cid:2)TheAuthor(s)2013.PublishedbyOxfordUniversityPress.
ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionNon-CommercialLicense(http://creativecommons.org/licenses/
by-nc/3.0/),whichpermitsunrestrictednon-commercialuse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
2858 NucleicAcidsResearch,2013,Vol.41,No.5
partners. Here, we combine the techniques of DNase-seq supplemented with 10% FBS. Primary skin fibroblasts
and a novel method to fine map nucleosome positions by (GM08333) were grown in MEM (Invitrogen) supple-
bacterial artificial chromosome (BAC) enrichment of mented with 15% FBS. NHBE cells, a mixture of
mononucleosomal DNA and deep sequencing, to identify primary human bronchial and tracheal epithelial cells
trans-acting factors regulating cystic fibrosis transmem- (Lonza, CC-2541), were cultured in BEGM (Lonza) as
brane conductance regulator (CFTR) gene expression. perthemanufacturer’sinstructions.Twoindependentbio-
TheCFTRgenespans189kbofgenomicDNAatchromo- logical samples were analysed, NHBE1 (7F3506) and
some7q31(11),althoughregulatoryelementsforthelocus NHBE2 (8F3407). Primary human fetal epididymis cells
maptoregionsfarupstreamanddownstreamofthegene, wereculturedinCMRL1066mediumwith10%FCS(20).
in addition to within introns [reviewed in (12)]. Hence, to For glucocorticoid receptor binding experiments,
fully understand the CFTR cis-regulatory sequences and 16HBE14o- cells at 100% confluence were treated with
trans-acting factors that control cell-specific expression of 100nM dexamethasone (Sigma-Aldrich, St. Louis, MO,
thegene,aregionof>250kbmustbeevaluated.Theavail- USA) in ethanol or the same volume of ethanol alone
abilityofaCFTRBACclonethatincludesthemajorityof for 3h and then harvested for ChIP or RNA extraction.
thecriticalregions(13)greatlyfacilitatedthisanalysis. CFTR mRNA was assayed by quantitative reverse tran-
CFTR shows tight tissue-specific regulation of expres- scriptase–polymerase chain reaction (RT–qPCR) as
sion, which is largely determined by cis-regulatory described previously (21).
elements that lie outside the gene promoter (14). We
recentlyshowedthattheactivelocusexistsinaloopedcon- Isolation of nuclei, MNase digestion and DNA purification
formation that brings distal regulatory elements, located
Cells were harvested with trypsin or non-enzymatic dis-
within cell-type–specific DNase I hypersensitive sites
sociation solution (NHBE), washed twice in chilled
(DHS) into close proximity with the gene promoter (15). phosphate-bufferedsalineat4(cid:2)Candpelletedbycentrifu-
Moreover, we showed that the regulatory mechanisms in
gation at 200g for 1min. Cells were then resuspended in
airwayepithelialcells,whichgenerallyexpresslowlevelsof
ice-cold NP-40lysis buffer [10mMTris, pH 7.4, 10mM
CFTRincomparisonwithintestinalepithelialcells,achieve
NaCl, 3mM MgCl , 1mMCaCl 0.5% Nonidet P-40,
thisthroughdifferentcis-actingelements(15,16).However, 2 2,
0.15mM spermine, 0.5mM spermidine supplemented
in either cell type, the identification of the critical trans-
with one complete ethylenediaminetetraacetic acid
actingfactors that activate these elements remains incom-
(EDTA)-free protease inhibitor cocktail/50ml] for 5min
plete.Severalmethodshaverecentlybeendevelopedtoac-
on ice. Nuclei were pelleted at 19g for 10min in a
curately map transcription factor binding sites including pre-cooled swinging bucket rotor at 4(cid:2)C. Nuclei were
directmethods,suchashigh-resolutioninvivofootprinting
gently re-suspended in micrococcal nuclease (MNase) di-
by DNase-seq, (3,4) and indirect methods provided by
gestion buffer (10mMTris, pH 7.4, 60mMKCl,
mapping nucleosomes in vivo (17). Here, we use a novel
15mMNaCl,0.15mMspermine,0.5mMspermidinesup-
method(Yigitetal.,2012,inreview)tosearchforcell-type-
plemented with one complete EDTA-free protease inhibi-
–specificdepletionofindividualnucleosomesatthecoresof
tor cocktail/ 50ml) and pelleted again as described earlier
regions of open chromatin that correspond to well-
inthetext.Formicrococcalnucleasedigestion,nucleiwere
characterized regulatoryelements associated withDHS in
resuspended in 1ml MNase buffer and then adjusted to
the CFTR gene. We find specific loss of nucleosomes that 4(cid:3)107 nuclei/ml in MNase buffer, on ice. MNase diges-
correspond to well-characterized binding sites for known tionwasperformedat37(cid:2)Cinthepresenceof1mMCaCl
trans-acting factors. Next, we identify predicted binding 2
until the 147-bp (core nucleosome) and 167-bp
sites for factors that have not previously been implicated
(chromatosome) bands were clearly identifiable. The di-
in CFTR regulation, which interact with nucleosome
gestion was stopped with a final concentration of
depleted regions in vivo and demonstrate occupancy of
10mMEDTA and 1% sodium dodecyl sulphate (SDS).
these factors in relevant cell types by ChIP. Finally, we
After the digestion, the nucleosomes were treated with
use a bioinformatic approach to identify transcription RNase A at 37(cid:2)C for 30min, and proteins were digested
factor binding sites that are overrepresented in regions
by proteinase K treatment. DNA was extracted with
where a nucleosome is lost in a cell-specific manner in a
phenol/chloroform, precipitated with ethanol and resus-
bronchial epithelial cell line (16HBE14o-) in comparison
pended in 1(cid:3)TE buffer.
with skin fibroblasts. This approach predicts sites for
several factors including the glucocorticoid receptor
Gel extraction and purification of mononucleosome DNA
(GR),andweshowbyChIPthatdexamethasone-induced
occupancyofmultipleGRelements(GREs)causesrepres- Digested chromatin DNA was separated on 3% agarose
sionofCFTRtranscription. gel (NusieveGTG, Lonza) in 1(cid:3)TBE buffer until 147-
and 167-bp DNA bands were clearly identifiable. The
147-bp DNA band was carefully excised from the
MATERIALS AND METHODS agarose gel, and it was eluted in crush and soak buffer
(300mM sodium acetate, 1mM EDTA) with gentle over-
Cell culture
night agitation. The excess agarose was removed by cen-
The human colon carcinoma cell lines Caco2 (18) and trifugation,andsupernatantcontainingnucleosomeDNA
16HBE14o- bronchial epithelial cells (19), were grown was cleaned up by QIAquick PCR Purification Kit
in Dulbecco’s modified Eagle’s medium (Invitrogen) (Qiagen).
NucleicAcidsResearch,2013,Vol.41,No.5 2859
Preparation of SOLiD library After enrichment, DNA was amplified by PCR (18
cycles) using P1 and P2 primers, and PCR products
About6–7mgofDNAwasendrepairedandligatedtothe
were gel extracted. Enriched paired-end libraries were
forward (P1) and reverse (P2) adaptors according to the
combined in equal molar amount and sequenced on
SOLiD System library preparation protocol using
SOLiD System 4 (ABI, Carlsbad, CA, USA) instrument
reagents from New England Biolabs (NEB). P1 adaptors
by paired-end chemistry (35/25bp) using a quarter of the
had unique 6-bp barcodes at their 30-ends to distinguish
slide.
nucleosomes from each of the six cell types analysed.
Ligationproductswereseparatedby5%nativepolyacryl-
Sequence alignments
amide gel, and they were extracted by the crush and soak
method described earlier in the text. DNA was further After sequencing, the colourspace reads were grouped by
purified and concentrated by QIAquick PCR purification barcode and aligned to the human reference genome
kitandthennicktranslatedbyDNApolymeraseI(NEB) (GRCh37/hg19) using BioScope version 1.3.1 from ABI
as described in the SOLiD System library preparation (Carlsbad, CA, USA). After alignment, the BAM files
protocol. The reaction was stopped with buffer PB and were processed, and a text file was generated, which
purified by QIAquick PCR purification kit. contains, for each read, the genomic coordinates, strand,
sequence, alignment length and the number of locations
the read has aligned to in the genome. Only the uniquely
BAC DNA isolation and enrichment of targets
mapped reads were kept for the subsequent analyses. To
by BAC hybridization
construct the nucleosome occupancy score, we further
BAC123s has a 250.3-kb insert containing the whole- filtered out the reads of length between 137 and 157bp.
coding region of the CFTR gene, also 40.1kb of DNA 50 Let R be the number of sequencing reads centred at
j
and 25kb 30 to the gene (13). BAC DNA was isolated genomic location j. We defined the centre weighted nu-
using the PureLink HiPure Midi Kit (Life Technologies) cleosome occupancy score at location i as follows:
according to the manufacturer’s protocol and verified by
X
digestion withNotIandNruI followedbypulsed-field gel oi ¼ RjGði(cid:5)j;20Þ, ð1Þ
electrophoresis (13). Enrichment of SOLiD nucleosome jj(cid:5)ij(cid:6)60
libraries (2mg) was done essentially as described previ-
whereGisaGaussiandensitywithstandarddeviation20.
ously (22). In summary, (cid:4)150ng of BAC DNA was
This Gaussian weight function covers ±60bp of the pro-
labelled with biotin-dUTP (Enzo Life Sciences) and
[a-32P]dCTPusinganicktranslationkit(Roche),accord- jected nucleosome centre and decays towards the two
ingtothemanufacturer’s protocolat15(cid:2)Cfor1h.[a-32P] ends. Compared with uniform weight function, the
Gaussian function is effective in avoiding spikes in the
was used to indirectly determine the incorporation of
reads occupancy curve frequently observed in the linker
biotin-dUTP. The biotinylated products were purified
region because of overlap of reads arising from two
through P-30 columns (Bio-Rad) and lyophilized.
neighbouring nucleosomes, and thus provides a sharper
Labelled BAC was dissolved in Cot-1 DNA (2mg/ml)
boundary of closely positioned adjacent nucleosomes.
(Invitrogen). The solution was overlaid with mineral oil,
denatured by incubation at 95(cid:2)C for 5min and incubated
at 65(cid:2)C for 15min. Five microlitres of 2(cid:3)hybridization Bioinformatic identification of overrepresented
buffer (22) was added and incubated at 65(cid:2)C for 6h. transcription factor binding sites in nucleosome
depleted regions
TocapturenucleosomeDNA,1–2mgofadapter-ligated
genomic DNA in 5ml of water was overlaid with mineral The paired-end SOLiD reads that mapped to the CFTR
oil,5mlof2(cid:3)hybridizationbufferwasaddedandthenthe locusonGenomeReferenceConsortiumHumanBuild37
entire sample was transferred to the tube containing the (GRCh37) were compared across different cell types by
Cot-1blockedBACDNA.Thehybridizationreactionwas using the edgeR Bioconductor package (23). The edgeR
incubated at 65(cid:2)C for a further 72–80h. In all, 150ml of package was initially designed for examining differential
streptavidin-coated beads (Invitrogen) were washed twice expression of replicated count data, but it is equally ap-
with200mlstreptavidinbead-bindingbuffer(10mMTris– plicable to analysis of high-throughput sequencing of
HCl,pH7.5,1mMEDTA,pH8,and1MNaCl)andthen nucleosome-depleted regions. First, discrete genomic
resuspended in 150ml of the same buffer. Next, the whole intervals were defined over which to compare the
hybridization reactionwasaddedtothestreptavidinbead number of overlapping sequence reads. These ((cid:4)25 k)
suspensionandboundatroomtemperaturefor30minon intervals consisted of a 100-bp window starting at the be-
a rotator. Streptavidin beads were removed from the ginningoftheCFTRlocusandslidingforwards10bpata
binding buffer and washed once, at 25(cid:2)C for 15min in timeuntilreachingtheendofthelocus(chr7:117082000–
1ml of 1(cid:3)SSC with 0.1% SDS, and then three times, 117332130). Next, the number of sequencing reads that
each at 65(cid:2)C for 15min in 1ml 0.1(cid:3)SSC with 0.1% overlapped(atthemidpointofthesequence)witheachof
SDS. The nucleosome DNA hybridized to the labelled theseintervals,ineachcelltype,wascomputedwithaperl
BAC DNA was eluted by addition of 100ml of program. Finally, edgeR was run on the resultant
0.1MNaOH at 25(cid:2)C for 10min. Eluted DNA solution sequence counts for each of these intervals, comparing
was neutralized by addition of 100ml of 1MTris–HCl each cell type with all other cell types to determine
(pH 7.5) and desalted by using a P-30 column. which intervals were consistently nucleosome depleted,
2860 NucleicAcidsResearch,2013,Vol.41,No.5
or were depleted in one cell type (e.g. 16HBE14o-) an apparent depletion of nucleosomes in the BAC
compared with another cell type (e.g. skin fibroblasts). hybridization-generated data. Nucleosome positioning
The edgeR program output all intervals that were repre- data from the ENCODE consortium (Stanford/BYU)
sented at a statistically different level between cell types, show a comparable depletion of signal in this area, which
and different P-value cut-offs were then used to select the is likely because of the >70% GC content of this region.
mostsignificantdifferencesbetweencelltypes.Next,these The coincidence of nucleosome positioning between the
regions were analysed with the Clover program for iden- two methods was high in the core of the intron 11 DHS
tificationoffunctionalsequencemotifsinDNAsequences enhancersequence(SupplementaryFigureS2).
(24). The CFTR locus was used as the background
sequence for these analyses.
Specific loss of nucleosomes at well-characterized binding
sites for known trans-acting factors
Chromatin immunoprecipitation
Enhancer elements
ChIPexperimentswerecarriedoutbystandardprotocols. We previously identified several enhancer elements
The following antibodies were used: GR (SantaCruz associated with DHS within introns of the CFTR gene
sc1003); FOXA1 (Abcam, ab5089) and BAF155 that interact with the promoter in a cell-type–specific
(sc10756). Rabbit and goat IgG were from Santa Cruz manner.Insomecases,wealsocharacterizedcriticaltran-
(sc-2027 and sc-2028). All primers used for qPCR are
scription factors that bind to the elements and contribute
listed in Supplementary Table S1.
to enhancer function. A weak intestinal-specific enhancer
inthefirstintronofthegeneat185+10kb(185isthelast
codingbaseinexon1)(25)wasshowntobindhepatocyte
RESULTS
nuclear factor 1 (HNF1) both in vitro and in vivo (15,26).
Statistics of SOLiD DNA sequence analysis Inspection of the nucleosome positioning data for this
region reveals specific loss/repositioning of nucleosomes
Data summarizing the sequence analysis for the six cell
within the core region of this enhancer element in Caco2
types (including two biological replicas of the primary
intestinal carcinoma cells (Figure 1A). We next analysed
NHBE cells) are shown in Table 1. Total unique
the precise region (chr7: 117130188–117130198) of the
sequence tags mapping to the hg19 human genome build
conserved HNF1 binding site that we showed to be
ranged from 1412498 for primary epididymis cells to
critical for the activity of the enhancer (26) and found it
11411833 for 16HBE14o- cells, with the range likely re-
to coincide with the nucleosome depleted region.
flecting absolute amounts of barcoded, purified
Enrichment of HNF1 binding at this element in Caco2
mononucleosomal DNA in the complex mixture that
cells was confirmed by ChIP (Figure 1B). This region
was sequenced. The unique tags within the CFTR BAC
was also enriched for the enhancer signature protein
ranged from 1040691 to 6936596, respectively, in the
p300 in ChIP experiments (15).
same samples, demonstrating percentage enrichment
Astrongercell-type–selectiveenhancerexistsinintron11
efficiencies of 74–61%. The lowest enrichment efficiency
of the CFTR gene at 1811+0.8kb, which is also enriched
was 34% in NHBE2, whereas fibroblasts, NHBE1 and
for p300 and was shown by chromosome conformation
Caco2 all showed enrichments of between 52 and 56%.
capture(3C)tointeractdirectlywiththeCFTRpromoter
Sequence coverage by nucleosome was calculated accord-
(15,27). ChIP-seq data from the ENCODE consortium
ing to the formula (number of mapped reads)(cid:3)(147)/
identify binding sites for a number of transcription
(BACsize)(7)andrangedfrom612to4079,thusconfirm-
factors, including FOXA1 and C/EBPb in this core of
ing adequate coverage of nucleosome positions across the
this enhancer element (Figure 2A). This is consistent with
BAC for each cell type.
thedepletion/relocationofatleasttwonucleosomesinthis
region in Caco2 cells in which the intron 11 element is
Comparative analysis of nucleosome positions using
highlyactive(15).Incontrast,well-positionednucleosomes
different methodologies
are seen at this site in skin fibroblast and primary airway
We previously developed a method of detecting nucleo- (NHBE) cells in which the element is inactive. Moreover,
some positioning by tiled qPCR of purified usingChIPwithanantibodyspecificforFOXA1,weshow
mononucleosomal DNA (17). This method was applied no enrichment of this factor in the DHS11 region
to analysis of the CFTR promoter and the intron 11 (Figure 2B), despite the expression of FOXA1 protein in
(1811+0.1kb) DHS region (where 1811 is the last coding these cells (Figure 2C). In contrast, there is strong enrich-
base in exon 11), which contains a potent enhancer of ment of FOXA1 at the DHS 11 enhancer in Caco2 cells
CFTR promoter activity (15). It was of interest to (Figure 2B), consistent with our recent data showing
compare the nucleosome positioning data generated by binding of both FOXA1 and FOXA2 (which recognize
thismethodandtheBAChybridizationmethoddescribed thesameDNAsequencemotif)tothisregion(28).
here. There was generally good overlap between nucleo-
some mapping generated by the two methods across the Insulator elements
2-kbregionanalysed(SupplementaryFigureS1),withthe Enhancer-blocking insulator elements are important for
exceptionofthe500-bpregionimmediately50 tothetran- establishing chromatin domains and transcriptional hubs
scription start site. Where we previously noted three well- genome wide and are often located at the functional
positioned nucleosomes using the qPCR method, there is boundaries of individual loci. CCCTC binding factor
NucleicAcidsResearch,2013,Vol.41,No.5 2861
Table 1. Statistics of SOLiD paired-end DNA sequence analysis
Cell type/barcode Fibroblast NHBE1 NHBE2 16HBE14o- Caco2 Epididymis
AGCTTA GTCATC GCATGT AAGTAA GTGCCT TAAAGT
Unique tags within CFTR BAC 5635722 4695178 1054913 6936596 3121914 1040691
Unique tags at NON-target region 4491093 4282556 2051985 4475238 2832115 371808
Unique tags in hg19 10126815 8977734 3106898 11411833 5954029 1412498
Non-unique tags in hg19 639090 262297 196299 823494 361968 76100
Total number of tags in hg19 10765905 9502327 3303197 12235327 6315997 1488598
Enrichment efficiency (%) 56 52 34 61 52 74
Unique pairs in hg19 (%) 94 94 94 93 94 95
Coverage per nucleosome 3314 2761 620 4079 1836 612
A
Scale 1 kb
chr7: 117129000 117129500 117130000 117130500 117131000 117131500
185+10kb DHS (intron1)
Fibroblast
NHBE1
s
d
ea NHBE2
r
e
c
n
ue 16HBE14o
q
e
S
Caco2
Digital DNaseI Hypersensitivity Clusters from ENCODE
Scale 50 bases hg19
chr7: 117,130,160 117,130,170 117,130,180 117,130,190 117,130,200 117,130,210 117,130,220 117,130,230 117,130,240
--->TTGGAATCAGACAGACCTGGCTGGAATCCTAACTCTGTCACTTATTAACAATGTGATCTTAGGCAATTTACTTAATCTCTCTGAACCTCAGCTACTCTCGT
HNF1
B HNF1
CFTR 23 8 10 11 18 20 24
Caco2
-20.9 PkrbomoteIrnt D1HS 1 HS 1D0(HaSb )10(DcH)S 11 +6.8+ 1k5b.6 kb
D
Figure 1. Nucleosome positionsacrossCFTRintron 1(185+10kb) DHSidentifythe locationofacritical HNF1 bindingsite.(A)Sequencereads
showingnucleosomepositionsacrossCFTRintron1(185+10kb)DHS(25)infibroblasts,NHBE,16HBE14o-andCaco2cells.Thelocationofthe
DHS is marked above the reads, and digital DNase hypersensitivity data from ENCODE confirm its location below the reads. Asterisk denotes
depletionofanucleosomeattheDHScorethatcorrespondstotheHNF1bindingsiteinCaco2cells.Belowthereads,single-baseresolutionofthe
region marked by asterisk in (A) shows the HNF1 binding site (26). (B) Enrichment of HNF1 assayed by ChIP is greatest at the intron 1 DHS,
although it is also seen at other elements across the CFTR locus in Caco2 cells.
2862 NucleicAcidsResearch,2013,Vol.41,No.5
A
1 kb
Scale
chr7 : 117227500 117228000 117228500 117229000 117229500 117230000 117230500 117231000
CFTR
1811+0.8kb DHS (intron11)
Fibroblast
NHBE1
s
d
a NHBE2
e
r
e
c
n
e 16HBE14o
u
q
e
S
Caco2
Digital DNaseI Hypersensitivity Clusters from ENCODE
Transcription Factor ChIP-seq from ENCODE
FOXA1
CEBPB
p300
B FOXA1 C kDa Caco2 16HBE14oN-HBE
CFTR 23 8 10 11 18 20 24
FOXA1
50
NHBE β-tubulin
50
-20.9 kbP-r4o kmboteIrnt D1HS 1 Int 8DHS 1D0(HaSb )10(DcH)S 11Int 17a Int 23
Caco2
Figure 2. Nucleosome positions across CFTR intron 11 (1811+0.8kb) DHS show the FOXA binding site that contributes to enhancer activity of
thiselement.(A)SequencereadsshowingnucleosomepositionsacrossCFTRintron11(1811+0.8kb)DHS(15)infibroblasts,NHBE,16HBE14o-
andCaco2cells.ThelocationoftheDHSismarkedabovethereadsanddigitalDNasehypersensitivitydataandtranscriptionfactorChIP-seqdata
fromENCODEareshownbelowthereads.AsteriskdenotesdepletionofnucleosomesattheDHScorethatcorrespondstotheFOXAbindingsite
inCaco2cells.(B)ChIPwithanantibodyspecificforFOXA1showsnoenrichmentinNHBEcellswheretheDHS11enhancerisinactive,despite
the presence of low levels of FOXA1 protein and strong enrichment of DHS 11 and other cis-acting regulatory elements in CFTR in Caco2 cells.
(C)RelativeexpressionofFOXA1proteininCaco2,16HBE14o-andNHBEcells.Westernblotofwhole-celllysatesprobedwiththesameantibody
used in ChIP experiments or anti b-tubulin and ImageJ quantitation of a scan of the same blot, relative to the b-tubulin signal.
NucleicAcidsResearch,2013,Vol.41,No.5 2863
(CTCF) is frequently associated with these insulator withthenucleosomepositioningdataatthesesites(Figure
elements where it interacts with cohesin complex compo- 3A and B). The CTCF binding site at the core of the
nents (29,30). We previously described CTCF-binding, (cid:5)20.9-kb DHS is flanked by two well-positioned nucleo-
enhancer-blocking insulator elements associated with somes, with the 50-peak being more evident in all cases
DHS at (cid:5)20.9kb with respect to the CFTR translational than the 30-peak (Figure 3A). Moreover, these flanking
start site and at a DHS at+6.8-kb downstream from the nucleosomes seem to have been displaced to increase the
last coding base (31–33). The (cid:5)20.9-kb DHS is seen in nucleosome-freeregioncoincidingwiththeCTCFbinding
most epithelial cell types examined and also in some site.Theusualperiodicityof(cid:4)150bpbetweennucleosome
other cell types. In contrast, the +6.8-kb DHS is only peaks is increased to 270bp for nucleosomes flanking the
evident in human lung tissue and epididymis epithelial CTCF binding site. At the+6.8-kb insulator element, the
cells, although low levels of CTCF binding are also seen nucleosomes flankingthe CTCF binding site are also well
inCaco2cells.Theseobservationsaremarkedlyconsistent positioned, likely with a loss/repositioning of at least one
A
2 kb
Scale
chr7: 117098000 117099000 117100000 117101000 117102000 117103000
-20.9kb
Fibroblasts
NHBE1
s
d
a NHBE2
e
r
e
c
n
e 16HBE14o
u
q
e
S
Caco2
Digital DNaseI Hypersensitivity Clusters from ENCODE
Transcription Factor ChIP-seq from ENCODE
CTCF
Rad21
SMC3
B
Scale 500 bases
chr7117313600 117313800 117314000 117314200 117314400 117314600 117314800 117315000
+6.8kb DHS
Fibroblasts
NHBE1
s
d
a
e
r NHBE2
e
c
n
e
u
q 16HBE14o-
e
S
Caco2
Epididymis
Digital DNaseI Hypersensitivity Clusters from ENCODE
Transcription Factor ChIP-seq from ENCODE
CTCF
Rad21
SMC3
Figure 3. The(cid:5)20.9and+6.8kbDHSinCFTRthatencompassenhancer-blockinginsulatorsshownucleosomepositioningbyCTCFbinding.Sequence
readsshowingnucleosomepositionsacross(A)CFTR(cid:5)20.9-kbDHS(15)and(B)+6.8-kbDHSinfibroblasts,NHBE,16HBE14o-,Caco2andprimary
epididymiscells.ThelocationoftheDHSismarkedabovethereadsanddigitalDNasehypersensitivitydata,andtranscriptionfactorChIP-seqdatafrom
ENCODEareshownbelowthereads.AsteriskdenotesnucleosomedepletionattheDHScorethatcorrespondstotheCTCFsite.
2864 NucleicAcidsResearch,2013,Vol.41,No.5
nucleosome, as the distance between the peaks is (cid:4)300bp (cid:5)35-kb DHS in Caco2 cells reveals some displacement of
(Figure3B).Alsoofnoteisthedifferenceinprofileofthe nucleosomes (marked by arrows in Figure 4A) that may
sequence peaks between the flanking nucleosomes in dif- correspond to the activity of BAF155 at this site.
ferentcelltypes.ChIPdataforCTCFbindingintheDHS
+6.8-kbregionshowed15-foldenrichmentinprimaryepi- Identification of transcription factor binding sites that are
didymis cells in which the DHS is evident in comparison overrepresented in regions where a specific nucleosome is
with 6-fold in Caco2 cells that lack the DHS (33). lost in a cell-specific manner
Consistent with this observation, complete loss of nucleo-
Next,weinvestigatedwhetherwecouldidentifytranscrip-
somes between the flanking peaks is seen in epididymal
tion factors de novo that regulated CFTR expression by
cells, whereas there is low occupancy of this region in
binding to nucleosome-depleted regions, either by them-
Caco2 cells (Figure 3B). The +6.8-kb DHS is also
selves displacing nucleosome(s) or by occupying regions
evident in adult human lung tissue (34); therefore, it is
that were made accessible by the interaction of other
of interest that one culture of primary human airway epi-
factors. To achieve this, nucleosome occupancy across
thelial cells (NHBE2) shows a profile similar to the
the CFTR locus in skin fibroblasts (where the gene is
primary epididymis, whereas the other (NHBE1) is more
inactive and shows consistent nucleosome positioning)
similar to Caco2.
was compared with that of 16HBE14o- cells (where the
gene is highly expressed) to identify nucleosome-depleted
Other elements
regions that were restricted to the latter cell type. The
We examined nucleosome positioning across multiple
sequence of these regions was then evaluated using the
other cell-type–specific DHS in the CFTR locus. One
CLOVER algorithm to look for overrepresented tran-
such element is associated with a DHS at 21.5kb with
scription factor binding sites (Supplementary Table S2),
respect to last coding base of the gene and is only
usingthewholeCFTRBACasabackgroundsequencefor
evident in primary airway epithelial cells (15,16). Again,
comparison. This analysis revealed half-sites for multiple
specificdepletionofnucleosomesisseeninthesecelltypes;
nuclear hormone receptors including the glucocorticoid
however, in this case, at least three nucleosomes are
receptor (GR) and for hepatocyte nuclear factor 4
affected, two lie adjacent to each other, whereas the
(HNF4). To determine whether these sites were of func-
third is separated from them by two well-positioned nu-
tional relevance in CFTR expression, we treated
cleosomes (Supplementary Figure S3). The underlying
16HBE14o- cells with dexamethasone (100nM) for 3h
trans-acting factors are currently being investigated.
and then performed ChIP with an antibody specific for
GR. As the GR can bind to chromatin as a homodimer
Identification of predicted binding sites for novel factors
or in combination withother hormone receptors, changes
that bind in nucleosome depleted regions in vivo
in GR occupancy at predicted nucleosome-free GRE
We also evaluated the nucleosome positioning at a homodimer and GRE half binding sites in CFTR were
cell-type–specific DHS at (cid:5)35kb with respect to the then analysed (Figure 5, GRE half-sites and GRE
CFTR translational start site that is particularly evident homodimers sites in nucleosome-depleted regions are
in 16HBE14o- cells and primary epididymis cells and is shown above the ChIP graph, and the motifs used for
associated with enhancer activity in airway epithelial cells these predictions are shown in Supplementary Table S3).
(16).Todate,wehavenotidentified thecriticaltranscrip- MultiplesitesshowedenrichmentofGRafterdexametha-
tion factors that interact with this region. However, the sone induction, including in intron 1 chr7: 117143050–
availability of nucleosome positioning data across this 117143219, intron 10 at chr7: 117207782–117208258
region identified BAF155 (a component of the SWI/ and chr7: 117207992–117208126, intron 21 at chr7:
SNF-related,matrixassociated,actin-dependentregulator 117297508–117297802 and 30 to the gene at+16kb (at
of chromatin structure, SMARC) as a strong candidate chr7: 117323222–117323391) and +24kb (at chr7:
for interaction with this element. Selective depletion of 117331607–117331776) with respect to the end of the
at least three nucleosomes is seen in 16HBE14o- cells at coding sequence (Figure 5). In contrast, no changes in
the core of the (cid:5)35-kb DHS (Figure 4A), a region that GR occupancy were seen at any site after dexamethasone
correspondstoChIP-seqpeaksforC/EBPb(weakenrich- induction of skin fibroblasts. Moreover, using a RT–
ment) and BAF155 in ENCODE data. We were unable qPCR assay for CFTR mRNA, dexamethasone was
shown to decrease CFTR expression under the same ex-
to confirm C/EBPb binding to this site by ChIP in
perimental conditions (Figure 5 inset panel), suggesting
16HBE14o- cells, although this transcription factor is ex-
the observed GR occupancy changes were functionally
pressedatahighlevelinthesecells(microarrayexpression
relevant.
data not shown). However, the predicted BAF155 occu-
pancy at the (cid:5)35-kb DHS was confirmed by ChIP in
16HBE14o-cells (arrowed in Figure 4B). BAF155
DISCUSSION
protein is expressed in 16HBE14o- and Caco2 cells but
not in fibroblasts (Figure 4C). Consistent with these The dynamic competition between nucleosome position-
data, a low level of BAF155 enrichment is seen at ingon thechromatin fibreand theability oftranscription
multiple DHS across the locus by ChIP in Caco2 cells factors to bind their recognition sequences and modulate
but not in fibroblasts (Figure 4B). Moreover, inspection gene expression patterns is complex (6,7). Although some
of the nucleosome positioning profile at the core of the proteins, such as pioneer factors, have the capability to
NucleicAcidsResearch,2013,Vol.41,No.5 2865
A
2 kb
Scale
chr7: 117083500 117084500 117085500 117086500 117087500
-35kb DHS
Fibroblast
NHBE1
s
d
a NHBE2
e
e r
c
n
e 16HBE14o
u
q
e
S
Caco2
Digital DNaseI Hypersensitivity Clusters from ENCODE
Transcription Factor ChIP-seq from ENCODE
CEBPB H
BAF155
BAF155
B CFTR 23 8 10 11 18 20 24 C 16HBE14oC-aco2 NHBE Fibroblasts
kDa
16HBE14o- 140 Baf155
50 β-tubulin
-44-35 -20.9 -4 -P2rom DHS10(c)DHS11 DHS19 DHS23+15.6+21.5+36.6
Caco2
Fibroblast
-50kb 0 +50kb +100kb +150kb +200kb
Position Relative to CFTR Translational Start Point (10^4)
Figure 4. Nucleosomedepletionatthe(cid:5)35-kbDHSin16HBE14o-cellsidentifiesaroleforBAF155inCFTRregulationinairwayepithelialcells.
(A) Sequence reads showing nucleosome positions across CFTR (cid:5)35-kb DHS (15,16) in fibroblasts, NHBE, 16HBE14o- and Caco2 cells. The
location of the DHS is marked above the reads, and digital DNase hypersensitivity data and transcription factor ChIP-seq from ENCODE are
shown below the reads. Asterisk denotes depletion of several nucleosomes at the DHS core that correspond to the BAF155 binding site in
16HBE14o- cells. (B) ChIP with an antibody specific for BAF155 shows enrichment of (cid:5)35kb DHS (arrow) and other cis-acting regulatory
elements in CFTR in 16HBE14o- and Caco2 cells but not in fibroblasts. (C) Relative expression of BAF155 protein in Caco2, 16HBE14o-,
NHBE cells and fibroblasts. Western blot of whole-cell lysates probed with the same antibody used in ChIP experiments or anti b-tubulin and
ImageJ quantitation of a scan of the same blot, relative to the b-tubulin signal.
interact with and disrupt compacted heterochromatin by 250-kb locus to identify cell-type–specific transcription
recruiting chromatin-remodelling complexes (35), others factorbinding sites.Applying anovelmethod tomapnu-
canonlybindtoopenchromatin.Moreover,alltranscrip- cleosomes inprimary human epithelial cellsand epithelial
tion factors must overcome the energetically favourable cell lines, we generated precise nucleosome occupancy
nucleosome core particle by destabilizing its interaction data across the CFTR locus. Purified mononucleosomal
with the DNA strand (8,9). Despite the power of DNA from each cell type was hybridized to a 250-kb
multiple search engines to predict transcription factor BAC encompassing the locus, purified, amplified and
binding sites in silico based on DNA sequence alone, sequenced.Sequenceswerealignedwithhg19,andthenu-
these bioinformatic methods have their limitations. cleosome positions were then compared between the
Hence,weexaminedthepotentialofusingexperimentally diverse cell types, enabling specific differences in nucleo-
determined nucleosome positioning information across a someoccupancytobecorrelatedwithtranscriptionfactor
2866 NucleicAcidsResearch,2013,Vol.41,No.5
-50kb 0 +50kb +100kb +150kb +200kb
CFTR
nucleosome depleted in 16HBE14o- (1e-100) GRE half sites
nucleosome depleted in 16HBE14o- (1e-100) GRE homodimer sites
4.0 on
3.0 Rexpressi11..05 * GR DEteOxH
FT 0.5
C
2.0 Relative 0 DEX EtOH
)G
gI 1.0
o
t
e
v
ita 0.0
leR(tn 16HBE14o- -2Prom i1 i10i i10iii10iii i21 +16+24
e
m
h 4.0
c
irn
E
d 3.0
lo
F
2.0
1.0
0.0
Fibroblast
Figure 5. Dexamethasone enhances occupancy of GR binding sites in 16HBE14o- selective nucleosome-free regions and suppresses CFTR expres-
sion.ChIPwithanantibodyspecificforGRshowsdexamethasone-activatedoccupancyofGRbindingsitesin16HBE14o-cellsatCFTRintrons1,
10 and 21 and 30 to the gene at+16 and+24kb. Fibroblasts show no change in GR occupancy. Black bars: EtOH, open bars: 100nM dexa-
methasone in EtOH. Clover predictions for GRE half sites and homodimer sites (Supplementary Table S3) in 16HBE14o- selective
nucleosome-depleted regions are shown above the graph. Inset: Dexamethasone significantly reduces CFTR mRNA as measured by qRT–PCR.
*P<0.01, t-test.
interactions. Moreover, bioinformatics approaches to sequence-dependent bias. In the context of our experi-
revealTFbindingmotifsthatwereoverrepresentedinnu- ments, thisisnotsignificant,aswearecomparingnucleo-
cleosome-depleted regions identified the involvement of some occupancy/depletion in different cell types which all
theGRinCFTRregulation, amechanismthatwasprevi- have essentially the same genomic sequence. Moreover,
ously unknown. recent data showed that chemical mapping produces
The application of this method for nucleosome pos- nucleosome-positioning data that are similar to those
itioning for defined regions of the human genome has derived from MNase I (37). Second, the deep sequencing
some advantages over the ENCODE genome-wide data approach to nucleosome positioning may reveal appar-
that are available on the UCSC genome browser (http:// ently depleted regions that are artifacts of high GC
genome.ucsc.edu). Specifically the ability to map nucleo- content. An example is the presence of a gap of (cid:4)500bp
somes in chromatin derived from primary cell types for immediately 50 to the CFTR transcription start site when
which limited material is available and the ability to compared with our previous PCR tiling approach to map
generate greater sequence coverage per nucleosome. nucleosomes. However, this gap is also present in the nu-
Moreover, by combining multiple BACs, it should be cleosome positioning data on the ENCODE browser and
possible to generate a high-resolution map of nucleosome corresponds to a region of high GC content, which are
positioning over megabase regions of genomic DNA in difficult to sequence. The high GC content could also
addition to the 250–300kb regions cloned in single potentially influence the kinetics of BAC hybridization.
BACs. Two potential technical limitations of the method The identification of novel factors involved in regula-
should be noted. First, although the use of MNase I to tion of CFTR expression by analysing cell-type–selective
mapnucleosomepositionsisgenerallywellaccepted,there nucleosome-free regions warrants further discussion.
is some evidence (36) that MNase I has a First, we show enrichment of BAF155 in vivo at the