Molecular Cell, Vol. 17, 453–462, February 4, 2005, Copyright ©2005 by Elsevier Inc. DOI 10.1016/j.molcel.2004.12.028
Proximity among Distant Regulatory Elements
at the ?-Globin Locus Requires GATA-1 and FOG-1
globin genes (Bender et al., 2000; Epner et al., 1998;
Reik et al., 1998; Schubeler et al., 2001). Several models
exist to explain how the LCR can enhance the rate of
1999; Engel and Tanimoto, 2000). Among these, looping
models propose that physical interaction between the
LCR and a promoter is a prerequisite for activated tran-
scription (Choi and Engel, 1988; Foley and Engel, 1992).
In addition, primary transcript fluorescence in situ hy-
bridization experiments demonstrated that only one hu-
man globin gene is active at a given allele at any time,
which favors a model where the LCR interacts with one
?-globin promoter at the expense of another (Gribnau
et al., 1998). Two reports provided independent and
direct evidence that the LCR is in proximity with the
actively transcribing globin gene to form what has been
termed an active chromatin hub (ACH) (Carter et al.,
2002; Tolhuis et al., 2002). This structure is detectable
require transcription (Palstra et al., 2003). Furthermore,
developmental switching of globin gene expression is
reflected in chromatin fiber alterations in which active
ones (Palstra et al., 2003). Although several regulatory
elements cluster in three-dimensional (3D) space, dele-
tions of different combinations of elements in the locus
revealed an essential role for the combined presence
of HS3 and the adult ? promoter for ACH formation and
chromatin organization (Patrinos et al., 2004).
The term “loop” is used here to imply a chromatin
conformation where two distal DNA fragments located
in cis along the chromatin fiber are physically close to
one another, but not to intervening DNA sequences. Sev-
eral mechanisms have been proposed to explain looped
patterns of proximity. For example, loop formation might
torsbound atdistal regulatorysites withina locus.Alter-
natively, colocalization of sequences within shared nu-
clear compartments may account for the observed
proximity of specific sequences in the same locus (Ko-
sak and Groudine, 2004; Osborne et al., 2004). In either
case, the molecular basis and functional significance of
such long-range interactions are only poorly under-
GATA-1 and its cofactor FOG-1 are essential tran-
scription factors for development of the erythroid cell
lineage (Fujiwara et al., 1996; Pevny et al., 1991; Tsang
at all erythroid-specific promoters and enhancers.
GATA-1 function requires direct interaction with FOG-1
as revealed through synthetic and naturally occurring
point mutations of GATA-1 that diminish FOG-1 binding
(Crispino et al., 1999). For example, a point mutation
(V205M) in the N-terminal zinc finger of GATA-1 disrupts
FOG-1 binding, resulting in severe dyserythropoietic
FOG-1 can augment or inhibit GATA-1 activity in differ-
ent promoter contexts, a mechanism for coactivation at
the ?-globin locus remains to be determined. Although
GATA-1(V205M) binds to naked DNA in vitro with the
Christopher R. Vakoc,1,3Danielle L. Letting,1,3
Nele Gheldof,2Tomoyuki Sawado,4M.A. Bender,5
Mark Groudine,4Mitchell J. Weiss,1,3Job Dekker,2
and Gerd A. Blobel1,3,*
1Division of Hematology
The Children’s Hospital of Philadelphia
Philadelphia, Pennsylvania 19104
2Program in Gene Function and Expression
Department of Biochemistry
and Molecular Pharmacology
University of Massachusetts Medical School
Worcester, Massachusetts 01605
3University of Pennsylvania School of Medicine
Philadelphia, Pennsylvania 19104
4Division of Basic Sciences
Fred Hutchinson Cancer Research Center
Department of Radiation Oncology
University of Washington School of Medicine
Seattle, Washington 98195
5Division of Clinical Research
Fred Hutchinson Cancer Research Center
Department of Pediatrics
University of Washington School of Medicine
Seattle, Washington 98195
Recent evidence suggests that long-range enhancers
and gene promoters are in close proximity, which
might reflect the formation of chromatin loops. Here,
we examined the mechanism for DNA looping at the
?-globin locus. By using chromosome conformation
capture (3C), we show that the hematopoietic tran-
scription factor GATA-1 and its cofactor FOG-1 are
required for the physical interaction between the
?-globin locus control region (LCR) and the ?-major
globin promoter. Kinetic studies reveal that GATA-1-
induced loop formation correlates with the onset of
?-globin transcription and occurs independently of
new protein synthesis. GATA-1 occupies the ?-major
globin promoter normally in fetal liver erythroblasts
from mice lacking the LCR, suggesting that GATA-1
binding to the promoter and LCR are independent
events that occur prior to loop formation. Together,
these data demonstrate that GATA-1 and FOG-1 are
providing general insights into long-range enhancer
The murine ?-globin locus contains four ?-like globin
genes (?y, ?h1, ?maj, and ?min) and an upstream LCR
consisting ofsix DNase I hypersensitivesites (HSs). Tar-
geted deletions in mice revealed an absolute require-
ment of the LCR for high-level transcription of all ?-like
same affinity as wild-type (wt) GATA-1 (Crispino et al.,
1999),chromatin immunoprecipitation(ChIP) studiesre-
vealed impaired binding to a subset of target sites in
the context of cellular chromatin (Letting et al., 2004).
Moreover, target gene occupancy by GATA-1 is dimin-
ished in hematopoietic cells that lack FOG-1 (Pal et al.,
2004). These studies suggest that FOG-1 functions in
part by facilitating GATA-1 binding to chromatin tar-
gets in vivo.
Here, we examined whether GATA-1 and FOG-1 are
required for the physical interaction between the LCR
and the ?-globin promoter by using the 3C assay. We
show that restoration of GATA-1 activity in GATA-1 null
G1E cells stimulated proximity between the LCR and
intervening embryonic genes, thus forming a looped
structure. In addition, we demonstrate that FOG-1 bind-
ing to GATA-1 is essential to induce full loop formation.
that GATA-1 directly occupies the promoter indepen-
dently of LCR proximity. Together, these findings sup-
port a model where GATA-1 and FOG-1 anchor a loop
between distant regulatory elements to activate tran-
BglII restriction enzyme digestion was used for the
3C assay based on previous reports of equal nuclear
digestion among different tissues (Tolhuis et al., 2002).
BglII restriction fragments are appropriately spaced
within the LCRand among each ?-globingene such that
they can resolve a looped structure (Tolhuis et al., 2002)
(Figure 1A). Within the LCR, BglII fragments used in this
study are in the HS2 core element or in a region located
between HS2 and HS3. Because the extent of nuclear
of ligation product formation, the efficiency of nuclear
BglII digestion was determined by Southern blotting.
Using a probe within the 5.4 kb BglII fragment encom-
passing the ?-major gene showed that the ratio of liber-
ated fragment to incompletely digested higher-molecu-
lar weight species was equal before and after GATA-1
activation (Supplemental Figure S1 available online at
DC1/). Thus, there is no detectable difference in diges-
tion efficiency at the site of actively induced ?-major
transcription upon G1E differentiation.
To assay the effect of GATA-1 on LCR proximity to
each globin gene, a primer within a BglII fragment con-
taining HS2 was used in pair-wise combination with a
primer within a BglII fragment in the vicinity of each
?-globin gene (Figure 1A). All PCR products were identi-
fied to be the expected ligation-dependent products by
tion. PCR products were quantified by using32P-labeled
?-dCTP in the PCR reaction followed by gel electropho-
resis and phosphorimager analysis. All PCR reactions
were performed in triplicate and averaged. To allow for
comparison between different PCR products, signals
consisting of defined molar amounts of each PCR prod-
uct analyzed in parallel. Signals were also normalized
to a control interaction between two BglII fragments at
the GAPDH locus to account for variations of template
amount or quality. GAPDH expression does not change
upon G1E differentiation (Supplemental Figure S2).
Analysis of GATA-1-ER-expressing G1E cells by 3C
showed that in the absence of tamoxifen the relative
proximity of each ?-like globin gene with the LCR de-
creased in a distance-dependent manner (Figure 1B,
white bars). Such a pattern is expected for random colli-
sions of fixed points along a flexible linear fiber and may
reflect the nontranscribing chromatin conformation of
for 20 hr (black bars), the HS2-?-major ligation product
increased 3.3-fold (n ? 4), a magnitude comparable to
that observed previously comparing expressing and
nonexpressing primary tissues (Tolhuis et al., 2002).
Proximity of the LCR with the ?y and ?h1 embryonic
consistent with low-level transcription of these genes
in the definitive stage G1E cells. A moderate increase
in proximity was observed between HS2 and ?-minor,
which likely reflects the LCR-dependent activity of this
gene. Based on the observation of induced proximity of
ing ?y and ?h1 globin genes, we interpret the locus as
transitioning from a linear to a looped conformation
Results and Discussion
GATA-1 Regulates Enhancer-Promoter Proximity
at the ?-Globin Locus
digestion, nuclear ligation, and PCR-based quantifica-
tion of ligated DNA products to determine proximity
between different regions along the chromosome (Cul-
len et al., 1993; Dekker et al., 2002). Two restriction
fragments that are in proximity in the nucleus display
higher crosslinking frequency and therefore higher liga-
tion frequency than two fragments further apart from
one another. To determine a role for GATA-1 in the for-
mation of a looped ?-globin locus, 3C was applied
to the GATA-1 null murine cell line G1E.
G1E cells are immortalized erythroid precursor cells
derived from in vitro-differentiated GATA-1?/?ES cells
(Weiss et al., 1997). Stable expression in these cells
of GATA-1 fused to the ligand binding domain of the
estrogen receptor (GATA-1-ER) permits ligand-induc-
ible activation of GATA-1, leading to synchronous cell
maturation, globin gene transcription, and cell cycle ar-
rest. By several criteria, including gene profiling and
expressionresembles the transition from the BFU-Estage
et al., 2004). GATA-1-ER binds to GATA elements in a
manner indistinguishable from endogenous GATA-1 in
primary erythroid cells (Johnson et al., 2002) and acti-
The G1E system has thus proven valuable for studying
GATA-1 function, including defining its role in chromatin
related processes (Kiekhaefer et al., 2002; Letting et al.,
2003), identification of new target genes (Gregory et
al., 1999; Kihm et al., 2002; Shirihai et al., 2000), and
structure-function analyses (Weiss et al., 1997). More-
over, the synchrony of GATA-1-induced differentiation
in this system is ideally suited for kinetic studies of
GATA-1 in its natural environment.
Chromatin Loop Formation by GATA-1 and FOG-1
Figure 1. GATA-1 Is Necessary to Transition
the ?-Globin Locus into a Looped Conforma-
tion as Determined by 3C Analysis
(A) Representation of the murine ?-globin lo-
cus and BglII fragments used for this study.
(B) Graph of the relative proximity of different
regions across the ?-globin locus with HS2 of
the LCR before and after GATA-1 activation.
Data shown are the average of four indepen-
dent experiments with PCR reactions from
each experiment performed in triplicate and
averaged. Each signal was normalized to
control templates and a GAPDH interaction.
Error bars denote SD. Gel shows representa-
tive duplicate PCR products of each 3C inter-
(C) Graph of the relative proximity of an inter-
HS2/3 fragment with the ?-major gene before
and after GATA-1 activation.
Figure 2. Loop Formation Correlates with
(A and B) Occupancy of GATA-1 at HS3 (A)
and HS2 (B) measured by ChIP using a
GATA-1 (N6) antibody or isotype-matched
treatment. Results are the average of two in-
(C and D) HS2-?-major (C) or HS2-?h1 (D)
proximity assayed by 3C at indicated time
points. Results shown are the average of
three independent experiments.
(E) TotalRNA pol II(left and center)and phos-
phoserine 5 pol II (right) measured by ChIP
at 0, 4, 8, and 12 hr at the promoter (left) or
Due to low sensitivity of the phosphoserine
5 pol II ChIP, isotype control signals were
subtracted from IP signals to show associa-
tion. For the promoter ChIP, a representative
experiment is shown.
Chromatin Loop Formation by GATA-1 and FOG-1
of tamoxifen to activate GATA-1-ER yielded identical
results (data not shown). Because estrogen and tamoxi-
fen have opposing effects on the ligand binding domain
of the ER, this suggests that the observed effects are
similar results were obtained in two independent G1E
clones expressing GATA-1-ER (data not shown).
HS3 is a major site of GATA-1 occupancy at the locus
and has been implicated in regulating the proximity of
the LCR and active genes in the locus (Patrinos et al.,
2004). Although a BglII fragment that selectively encom-
passes HS3 is lacking, one is present between HS2 and
HS. When a primer within this fragment (denoted HS2/3)
was paired with the ?-major primer, GATA-1 induced
proximity between these regions increased by 3-fold
(Figure 1C), similar to what was found for HS2-?-major.
Thus, analysis of independent restriction fragments
within the LCR confirmed GATA-1 induced LCR-pro-
moter proximity. Importantly, DNase I hypersensitivity
within the assayed restriction fragment does not appear
supports that the 3C assay measures distance between
chromosomal regions rather than variation in nuclease
dent clone that expressed GATA-1-ER at somewhat
lower levels but displayed similar proximity between
LCR and ?-major promoter after tamoxifen treatment
(data not shown).
Loop Formation by GATA-1 Occurs in the Absence
of New Protein Synthesis
To rule out that the effects of GATA-1 on LCR-promoter
formed 3C analysis on cells treated with tamoxifen and
the translation inhibitor cyclohexamide. Because G1E
cells undergo apoptosis in the presence of this drug
(data not shown), GATA-1-ER-expressing G1E cells
were generated that overexpress the antiapoptotic fac-
tor Bcl-XL. G1E Bcl-XLcells activate ?-globin synthesis
normally upon tamoxifen treatment (Welch et al., 2004).
Treatment of Bcl-XL-containing cells with cyclohexa-
by [35S]-methionine uptake (Figure 3A), and cells re-
mained viable for over 12 hr. As shown in Figure 3B,
GATA-1 augments HS2-?-major proximity in the Bcl-
XL-overexpressing G1E cells to levels similar to those
observed in previous experiments. Notably, cyclohexa-
mide treatment did not significantly inhibit the induction
of HS2-?-major proximity. As expected, GATA-1 occu-
pancy as determined by ChIP assay was unaffected by
the presence of cyclohexamide (data not shown). These
findings show that LCR-?-major proximity can be in-
duced directly by GATA-1 and does not require synthe-
sis of downstream target genes.
GATA-1 Occupancy Correlates with Loop
Formation In Vivo
scription factors such as p45/NF-E2 and EKLF, which
are known to act at the ?-globin locus. Therefore,
GATA-1 might induce loop formation indirectly through
the effects of GATA-1 are direct, we first compared the
kinetics of LCR-promoter proximity with that of GATA-1
occupancy as measured by ChIP. GATA-1-ER-express-
ing G1E cells were treated with tamoxifen for 0, 4, 8,
12, 16, or 20 hr. Consistent with our previous findings,
occupancyof GATA-1atHS3 oftheLCR wasdetectable
at 8 to 12 hr (Figure 2A). A similar pattern was observed
at HS2 (Figure 2B). 3C analysis over the same time
course revealed increased HS2-?-major proximity after
4 hr of tamoxifen treatment that was maintained at later
time points (Figure 2C). In contrast, HS2-?h1 proximity
was unchanged (Figure 2D). Hence, loop formation cor-
relates with GATA-1 occupancy at the LCR. Moreover,
induction of ?-major transcription begins approximately
7–12 hr after GATA-1 activation (Letting et al., 2003;
lates with or may even precede ?-major transcription.
This result is consistent with loop formation being a
prerequisite for transcription rather than a consequence
thereof. Although there is generally a good correlation
between GATA-1 occupancy at HS3 and LCR-promoter
proximity, there appears to be a further increase in
GATA-1 occupancy at HS2 at 20 hr that does not result
in a further increase in proximity between HS2 and the
?-globin promoter. This suggests that GATA-1 can trig-
ger proximity when present at the LCR at submaximal
amounts. Therefore, a further 2-fold increase at one HS
might not result in a measurable increase in proximity.
This viewis supportedby our 3Canalysis ofan indepen-
Onset of Loop Formation Correlates with Elongating
RNA Polymerase II at the ?-Major Gene
The LCR is required for the efficient transition from pre-
initiation complex formation to the elongating form of
RNA polymerase II (pol II) (Sawado et al., 2003). This is
reflected in the diminished presence at the 3? end of
the ?-major coding sequence of total pol II and pol II
phosphorylated at serine 5 of its C-terminal domain in
the absence of the LCR (Sawado et al., 2003). To assess
the temporal relationship between GATA-1-induced
loop formation and pol II recruitment/elongation, time
courseChIP experimentswerecarried outby usinganti-
bodies against total pol II or phosphoserine 5 pol II.
PCR primers that amplify either the ?-major promoter
to assess pol II recruitment or within intron 2 of the
?-major gene to allow measurement of elongating pol
II were used. The results show that an increase in elon-
gating pol II at the ?-major gene occurs at 4 hr and
peaks at 8 hr of tamoxifen treatment (Figure 2E, middle).
This correlates well with the kinetics of loop formation
(Figure 2C). Although serine 5 phosphorylation of RNA
pol II generally displays a similar pattern, there might
be a delay relative to total pol II association (Figure
2E, right). A more detailed time course could address
whether pol II association and phosphorylation might
be ordered processes. A similar kinetic of pol II recruit-
ment to the promoter was observed (Figure 2E, left).
Collectively, these findings demonstrate that the kinet-
ics of LCR-promoter proximitycorrelate with two known
downstream functions of the LCR in recruiting pol II to
the promoter and regulating the transition from initiation
Figure 3. GATA-1 Induces Loop Formation in the Absence of New
(A) G1E Bcl-XLcell lines were treated with estradiol,35S methionine,
and in the presence (?) or absence (?) of cyclohexamide (CHX).
Whole-cell extracts were assayed for35S incorporation to measure
new protein synthesis.
(B) 3C analysis of HS2-?-major proximity in G1E Bcl-XLcell lines in
the presence or absence of CHX. Results are the average of two
FOG-1 Binding Is Required for GATA-1-Induced
Changes in Locus Conformation
FOG proteins are a conserved family of multitype zinc
finger transcription factors known to bind and regulate
GATA factor function. Because they do not have any
DNA binding activity of their own, FOG proteins are
recruited to chromatin by GATA factors in a manner
resembling a coactivator/corepressor (Anguita et al.,
2004; Pal et al., 2004). GATA-1 bearing the V205M point
mutation that impairs binding to FOG-1 is severely com-
promised in inducing ?-globin gene expression (Nichols
et al., 2000) (Figure 4A). This suggests that FOG-1 func-
tions asa GATA-1coactivator atthe ?-globinlocus. One
potential caveat of this interpretation is that the V205M
teins, thereby compromising GATA-1 function. How-
ever, a compensatory mutation in FOG-1 (S706R) that
restores binding to GATA-1(V205M) rescues erythroid
maturation, arguing against this possibility (Crispino et
al., 1999). Although FOG-1 is required for globin gene
activation by GATA-1, the mechanism by which FOG-1
assists GATA-1 during this process remains unclear. To
to loop the locus into an active conformation, we per-
Figure 4. FOG-1 Is Required for GATA-1-Induced Loop Formation
(A) qRT-PCR analysis of ?-major mRNA in G1E cells expressing
GATA-1-ER or GATA-1(V205M)-ER before or after 20 hr of tamoxi-
conditions. Results are the average of four independent experi-
formed 3C in G1E cells expressing GATA-1(V205M)-ER
at levels similar to GATA-1-ER cells (Letting et al., 2004).
the increase in HS2-?-major proximity was substantially
impaired for GATA-1(V205M)-ER (1.8-fold) when com-
pared to GATA-1-ER cells (3.8-fold). These results were
verified in a second independently derived clone of
GATA-1(V205M)-ER expressing G1E cells (data not
shown). These results indicate that FOG-1 binding to
Chromatin Loop Formation by GATA-1 and FOG-1
GATA-1 is required to efficiently loop the ?-globin locus
into an active conformation. The residual increase in
LCR-promoter proximity in the GATA-1(V205M)-ER-
GATA-1 Occupies the ?-Major Promoter
Independently of the LCR
Two mechanisms might account for a FOG-1 require-
ment in loop formation. FOG-1 might act by facilitating
promoter independently of the LCR, which would pro-
vide one essential “anchor” for subsequent loop forma-
tion. In an alternative model, FOG-1 might be required
to deliver LCR bound GATA-1 to the promoter via a
FOG-1-facilitated looping mechanism. The latter model
was suggested by several observations. First, GATA-1
(Letting et al., 2004). Second, the lack of spatial conser-
vation of GATA sites at the promoter (Hardison et al.,
1997) raised the possibility that the detection of GATA-1
at the promoter by ChIP could reflect LCR proximity
transfer would be similar to the previously proposed
RNA polymerase transfer model (Johnson et al., 2001).
To distinguish between these two mechanisms, we ex-
amined GATA-1 occupancy at the ?-major promoter in
day 14.5 postcoitum fetal liver erythroid cells from mice
that are homozygous for a deletion spanning the entire
murine LCR, including HS1 through HS6 (Bender et al.,
2000). As shown in Figure 5A, GATA-1 occupancy at the
?-major promoter is similar in the presence or absence
of the LCR. These results were verified by using an
for possible generalized variation in GATA-1 occupancy
between wt and mutant mouse strains, GATA-1 occu-
pancy was examined at the FOG-1 locus (Figure 5B)
and EKLF promoter (Supplemental Figure S4) and found
to be the same. Finally, ChIP experiments with antibod-
ies against the p45 subunit of the hematopoietic tran-
scription factor NF-E2 displayed the same amount of
p45 binding at the ?-major promoter at wt and ?LCR
alleles, consistent with previous results (Sawado et al.,
2003). This finding suggests that loop formation as de-
tected by the 3C assay is not a prerequisite for GATA-1
recruitment to the ?-major promoter. Conversely,
GATA-1 binding to the LCR is unlikely to require the
presence of a promoter because GATA-1 is functional
at a stably integrated DNA fragment containing HS4 of
and is required for the formation of a DNase HS (Stama-
cate that transcription factor complexes assemble at
the LCR and the active promoters independently of
Based on these observations, we propose a model
for loop formation where direct DNA binding of GATA-1
to an enhancer and a promoter separated by over 40
kb can lead to juxtaposition of these elements within
the nucleus. FOG-1 likely acts at a step prior to loop
formation by facilitating sequence-specific GATA-1 oc-
cupancy at the promoter in the context of chromatin
(Letting et al., 2004; Pal et al., 2004). The mechanism
Figure 5. Normal GATA-1 Occupancy at the ?-Major Promoter in
the Absence of the LCR
ChIP experiments were performed in wt-LCR or ?LCR homozygous
mice by using GATA-1 (N6) or p45 antibody.
(A) ?-major promoter.
(B) FOG-1 locus (PreCRM1 region as described in Welch et al.
). Result shown is a representative experiment.
by which FOG-1 exerts this function and why the FOG-1
requirement varies among GATA sites remain important
questions in the future. The observation that four out of
nine FOG-1 zinc fingers can bind to GATA-1 raises the
possibility thatFOG-1 might bind simultaneouslyto pro-
moter and LCR bound GATA-1, forming a bridge that
stabilizes a chromatin loop. However, previous studies
indicate that FOG-1 with a single intact GATA-1 binding
zinc finger is sufficient for erythroid differentiation (Can-
tor et al., 2002). Therefore, to simultaneously contact
multiple GATA-1 molecules, FOG-1 would have to form
dimers or multimers for which there is currently no evi-
Together, our results show that GATA-1 is critical for
the configuration of the ?-globin locus. However, addi-
tional erythroid transcription factors such as NF-E2 and
cient because both are expressed in G1E cells in the
absence of GATA-1 (Johnson et al., 2002; Letting et
al., 2003). Recent evidence indicates that EKLF is also
required to achieve maximum proximity among regula-
tory elements at the ?-globin locus in vivo (Drissen et
al., 2004). The mechanism by which GATA-1 regulates
long-distance interactions likely involves formation of a
higher-order protein complex that stabilizes proximity
between enhancers and promoters (Sieweke and Graf,
1998).A physicalinteractionbetweenEKLF andGATA-1
(Merika and Orkin, 1995) might be a critical component
of such a protein bridge. We speculate that bringing
the LCR into proximity with the ?-major promoter by
that increase preinitiation complex recruitment and en-
hance the transition from transcription initiation to elon-
gation (Sawado et al., 2003).
The detection by ChIP of similar amounts of GATA-1
at the promoter in the presence and absence of the
LCR was surprising because LCR proximity would be
expected to increase GATA-1 concentration at the pro-
moter. This suggests that the ChIP assay preferentially
detects GATA-1 molecules that are directly associated
with their cognate elements rather than the fraction of
GATA-1 that is placed in promoter proximity indirectly
via looping. This emphasizes the utility of applying ChIP
and 3C assays in parallel to relate sequence-specific
occupancy oftranscription factorsto higher-orderchro-
matin structure at an endogenous locus.
Although the presence of substantial amounts of
GATA-1 bound to the ?-major promoter even in the ab-
sence of the LCR suggests that GATA-1 binds directly
to the ?-major promoter, it is noteworthy that GATA
sites at the ?-major globin promoter are not conserved
with regard to their position and in relation to other
elements (Hardison et al., 1997). This had originally led
to the speculation that GATA-1 that is detected at the
promoter by ChIP might reflect indirect association with
the promoter via looping. However, our results show
that this is not the case, but rather, that GATA-1 binds
to the promoter in the absence of the LCR either directly
can bind to syntenic but spatially nonconserved GATA
sites in the upstream enhancer of the GATA-1 gene
(Valverde-Garduno et al., 2004). Thus, sequence com-
parison between species is not always a reliable pre-
dictor of important regulatory sites.
It has been recently reported that transcribed genes
can colocalize within the nucleus to shared “transcrip-
tion factories” (Osborne et al., 2004). Therefore, one
mechanism for proximity among DNA elements is colo-
calization within a shared subnuclear compartment (Os-
borne et al., 2004). Alternatively, proximity among DNA
elements could occur by local folding of chromatin into
looped structures independently of changes in nuclear
location or transcription status. The latter mechanism
is supported by evidence of locus conformational
changes in the absence of transcription (Palstra et al.,
2003;Patrinos etal.,2004;Spilianakis andFlavell,2004).
Thus, a gene-specific 3D structure that is formed by
erythroid transcription factors might precede recruit-
ment to a transcription factory. However, proximity
among regulatory regions in the absence of ongoing
transcription could reflect colocalization to a distinct
subnuclear compartment where genes are poised for
activation (Ragoczy et al., 2003). The 3C assay likely
of both shared nuclear position and locus folding ef-
induced proximity among bound elements detected by
3C is in part due to localization in a shared nuclear com-
Our findings are reminiscent of earlier reports with a
similar nuclear ligation assay showing that the estrogen
receptor can induce proximity of an enhancer with a
promoter at the rat prolactin locus (Cullen et al., 1993;
Gothard et al., 1996). Notably, GATA-3, STAT6, and
formation at the IL4/IL5/IL13 locus (Spilianakis and Fla-
vell, 2004); however, a direct mechanism of action re-
mains to be established. How might diverse families
of transcription factors accomplish the general task of
anchoring enhancer-promoter loops? The chromatin fi-
ber may require structural transitions as reflected in
DNase I hypersensitivity to become adequately flexible
to allow collisions between distant enhancers and pro-
moters. Subsequently, a collision may become stabi-
lized by the physical interaction between transcription
factors bound at each element to establish a chroma-
G1E cells were cultured as described previously (Weiss et al., 1997).
Cells were treated with 1 ?M tamoxifen or 0.5 mM cyclohexamide
where indicated. G1E cells overexpressing Bcl-XLhave been de-
scribed (Welch et al., 2004). To measure translation inhibition by
cyclohexamide, cells were grown in methionine-free DMEM with
10% dialyzed FCS and Epo/Kit ligand, nonessential amino acids
washed two times with PBS, and lysed in SDS-laemmli buffer. Ex-
tract was run on 12% SDS PAGE gel and exposed o/n to film.
35S methionine. Cells were incubated 3 hr at 37?C,
The 3C assay was performed essentially as described previously
(Tolhuis et al., 2002) with small modifications. 1 ? 107cells were
crosslinked in 4.5 ml of 1 ? PBS with 1% formaldehyde at room
temperature for 10 min. The crosslinking reaction was stopped by
adding glycine to a final concentration of 0.125 M. Cells were centri-
fuged at 500 ? g in a Beckman GS-6R centrifuge, resuspended in
1 ml of 4?C cell lysis buffer (10 mM Tris [pH 8.0], 10 mM NaCl, 0.2%
NP40, and protease inhibitors), and incubated on ice for 10 min.
Samples were kept on ice from this point forward. Cell lysis was
completed with ten strokes of a Dounce homogenizer (pestle A).
Nuclei were washed with 0.5 ml of the appropriate 1? restriction
enzyme buffer. Nuclei were then resuspended in 762 ?l of the same
1? restriction enzyme buffer. SDS was added to a final concentra-
tion of 0.1%, and nuclei were incubated at 37?C for 15 min. Triton
X-100 was then added to the final concentration of 1% to sequester
SDS. Digestion was performed with 800 U of restriction enzyme at
37?C overnight. The reaction was stopped by adding SDS to a final
concentration of 2% and incubating at 65?C for 30 min. Crosslinks
were reversed overnight at 65?C for half of each sample for control
Southern blots to measure digestion efficiency. The remaining sam-
ple was diluted into an 8 ml ligation reaction buffer containing 1%
Triton X-100, 50 mM Tris 7.5, 10 mM MgCl2, 10 mM DTT, 0.1 mg/
ml BSA, 1 mM ATP, and 4000 U of T4 DNA Ligase (NEB). Ligations
were incubated at 16?C for 2 hr. Reactions were stopped by adding
EDTA to a concentration of 10 mM. Samples were treated with
100 ?g Proteinase K and incubated at 65?C to reverse crosslinks.
Samples were extracted with phenol:chloroform:IAA four times,
redissolved in deionized water and treated with 0.5 ?g RNase A.
Quantitative PCR reactions were performed by using dilutions of
Chromatin Loop Formation by GATA-1 and FOG-1
experimental templates into a linear range of signal for each individ-
ual primer pair in each individual experiment. Final quantification of
results was performed by performing the PCR reaction in the pres-
ence of32P-labeled ?-dCTP and exposing dried acrylamide gels to
a phosphor screen. Band intensities were quantified with Im-
ofdefined molaramounts ofeachgel-purified PCRproduct. Aquan-
titative PCR reaction was performed with control templates in paral-
lel with experimental templates. In addition, all PCR reactions for
interactions at the ?-globin locus from an experimental template
were normalized to a control interaction at the GAPDH locus to
equalize for small differences in template amount or quality.
BglII Primersused in thisstudyare asfollows: HS2,5?-ATGACTCAG
CACTGCTGTGCTCAAGCC-3?; HS2/3, 5?-ACATGAGGCTACTCTATTG
TCAGACTGTGC-3?; ?y, 5?-AATGTAAGAACAAGCCTCCATTTGTC
AAG-3?; ?h1, 5?-TCAGGAATGTTCCCAACTTTCACTCAATTCCCC-3?;
?-major, 5?-GGTGGAAGGGGGTATTATGAACATTCGG-3?; ?-minor,
5?-CTGCTCTTTCTTCTTCTTTACTTTACTCTCC-3?; 3 prime distal,
5?-TAGCTGTGGAGAGCAGGAGGTCTGCTAATGCC-3?; and GAPDH,
Basic Research Fellow and a Leukemia & Lymphoma Society Spe-
Received: September 28, 2004
Revised: November 30, 2004
Accepted: December 20, 2004
Published: February 3, 2005
Anguita, E., Hughes, J., Heyworth, C., Blobel, G.A., Wood, W.G.,
and Higgs, D.R. (2004). Globin gene activation during haemopoiesis
is driven by protein complexes nucleated by GATA-1 and GATA-2.
EMBO J. 23, 2841–2852.
Bender, M.A., Bulger, M., Close, J., and Groudine, M. (2000). Beta-
globin gene switching and DNase I sensitivity of the endogenous
beta-globin locus in mice do not require the locus control region.
Mol. Cell 5, 387–393.
model forlong-distance geneactivation. Genes Dev.13, 2465–2477.
Cantor, A.B., Katz, S.G., and Orkin, S.H. (2002). Distinct domains of
the GATA-1 cofactor FOG-1 differentially influence erythroid versus
megakaryocytic maturation. Mol. Cell. Biol. 22, 4268–4279.
Carter, D., Chakalova, L., Osborne, C.S., Dai, Y.F., and Fraser, P.
(2002). Long-range chromatin regulatory interactions in vivo. Nat.
Genet. 32, 623–626.
globin gene switching. Cell 55, 17–26.
Crispino, J.D., Lodish, M.B., MacKay, J.P., and Orkin, S.H. (1999).
interaction in differentiation: the GATA-1:FOG complex. Mol. Cell
Cullen, K.E., Kladde, M.P., and Seyfred, M.A. (1993). Interaction
ence 261, 203–206.
Dekker, J., Rippe, K., Dekker, M., and Kleckner, N. (2002). Capturing
chromosome conformation. Science 295, 1306–1311.
Drissen, R., Palstra, R.J., Gillemans, N., Splinter, E., Grosveld, F.,
Philipsen, S., and de Laat, W. (2004). The active spatial organization
of the beta-globin locus requires the transcription factor EKLF.
Genes Dev. 18, 2485–2490.
activity: new insights into beta-globin locus regulation. Cell 100,
Epner, E., Reik, A., Cimbora, D., Telling, A., Bender, M.A., Fiering,
S., Enver, T., Martin, D.I., Kennedy, M., Keller, G., and Groudine, M.
(1998). The beta-globin LCR is not necessary for an open chromatin
structure or developmentally regulated transcription of the native
mouse beta-globin locus. Mol. Cell 2, 447–455.
Foley, K.P., and Engel, J.D. (1992). Individual stage selector element
mutations lead to reciprocal changes in beta- vs. epsilon-globin
gene transcription: genetic confirmation of promoter competition
during globin gene switching. Genes Dev. 6, 730–744.
nick, E.H. (1999). Requirement of an E1A-sensitive coactivator for
long-range transactivation by the beta-globin locus control region.
J. Biol. Chem. 274, 26850–26859.
Fujiwara, Y., Browne, C.P., Cunniff, K., Goff, S.C., and Orkin, S.H.
(1996). Arrested development of embryonic red cell precursors in
mouse embryos lacking transcription factor GATA-1. Proc. Natl.
Acad. Sci. USA 93, 12355–12358.
Gothard, L.Q., Hibbard, J.C., and Seyfred, M.A. (1996). Estrogen-
mediated induction of rat prolactin gene transcription requires the
formation of a chromatin loop between the distal enhancer and
proximal promoter regions. Mol. Endocrinol. 10, 185–195.
Gregory, T., Yu, C., Ma, A., Orkin, S.H., Blobel, G.A., and Weiss,
M.J. (1999). GATA-1 and erythropoietin cooperate to promote ery-
throid cell survival by regulating bcl-xL expression. Blood 94, 87–96.
Gribnau, J., de Boer, E., Trimborn, T., Wijgerde, M., Milot, E., Gros-
ChIP was performed as described previously (Forsberg et al., 1999).
Antibodies used are GATA-1 N6 or C20, FOG-1 M20, p45 NF-E2
C19, and RNA pol II N20 (all from Santa Cruz). Ser5 pol II H14
antibody was purchased from Covance. For mouse experiments,
wt and ?LCR homozygous mouse day 14.5 fetal livers were used
as a source of chromatin. ?LCR homozygous mice are not viable
but are rescued by a YAC expressing the human ?-globin gene
(Bender et al., 2000). For phosphoserine RNA pol II ChIPs, 10 mM
?-glycerophosphate and 1 mM Na Vanadate were included in all
buffers to inhibit phosphatase activity. Results were quantified by
using real-time PCR with SYBR Green dye on an ABI Prism 7000
system. A standard curve was generated for each primer pair by a
dilution series of the input sample. All PCR signals from IP samples
were referenced to their respective input standard curve to normal-
ize for differences in cell number and for primer efficiency. Primers
used in this study are as follows: HS3 F, 5?-CTAGGGACTGAGAGA
GGCTGCTT-3?; HS3 R, 5?-ATGGGACCTCTGATAGACACATCTT-3?;
HS2 F, 5?-GGGTGTGTGGCCAGATGTTT-3?; HS2 R, 5?-CACCTTCC
CTGTGGACTTCCT-3?; ?-major promoter F, 5?-CAGGGAGAAATAT
GCT-3?; ?-major promoter R, 5?-GTGAGCAGATTGGCCCTTACC-3?;
?-major intron F, 5?-CTTCTCTCTCTCCTCTCTCTTTCTCTAATC-3?;
?-major intron R, 5?-AATGAACTGAGGGAAAGGAAAGG-3?; ?-major
IVR5 F, 5?-GTATGCTCAATTCAAATGTACCTTATTTTAA-3?; ?-major
IVR5 R, 5?-TTACCTCTTTATTTCACTTTTACACATAGCTAA-3?; EKLF
promoter F, 5?-TCTGCTCAAGGAGGAACAGAGCTA-3?; EKLF pro-
moter R, 5?-GGCTCCCTTTCAGGCATTATCAGA-3?; FOG-1 locus F,
5?-TGCAAGTCCCATCCTGATAAGA-3?; and FOG-1 locus R, 5?-GCA
RNA was extracted with Trizol reagent (Invitrogen). Reverse tran-
scription reactions were performed with Superscript II (Invitrogen).
Results were quantified by using real-time PCR with SYBR Green
dye on an ABI Prism 7000 system. ?-major primers were 5?-AAC
GATGGCCTGAATCACTTG-3? and 5?-AGCCTGAAGTTCTCAGGAT
CCA-3?. GAPDH primers were 5?-GATGCCCCCATGTTTGTGAT-3?
We thank Tom Kadesch, Steve Liebhaber, and Ben Olenchock for
G1E microarray database information. G.A.B. is supported by Na-
is supported by NIH training grant T32HL0743926. D.L.L. is sup-
ported by NIH training grant T32GM008216. J.D. and N.G. are sup-
ported by grants from the Worcester Foundation and by NIH grant
HG03143. M.G., T.S., and M.A.B. are supported by NIH grants
DK44746 and HL65440. T.S. is an American Society of Hematology
veld, F., and Fraser, P. (1998). Chromatin interaction mechanism of
transcriptional control in vivo. EMBO J. 17, 6020–6027.
Halachmi, S., Marden, E., Martin, G., MacKay, H., Abbondanza, C.,
Hardison, R., Slightom, J.L., Gumucio, D.L., Goodman, M., Stoja-
novic, N., and Miller, W. (1997). Locus control regions of mammalian
beta-globin gene clusters: combining phylogenetic analyses and
experimental results to gain functional insights. Gene 205, 73–94.
Johnson, K.D., Christensen, H.M., Zhao, B., and Bresnick, E.H.
(2001). Distinct mechanisms control RNA polymerase II recruitment
to a tissue-specific locus control region and a downstream pro-
moter. Mol. Cell 8, 465–471.
Johnson, K.D., Grass, J.A., Boyer, M.E., Kiekhaefer, C.M., Blobel,
G.A., Weiss, M.J., and Bresnick, E.H. (2002). Cooperative activities
of hematopoietic regulators recruit RNA polymerase II to a tissue-
specific chromatin domain. Proc. Natl. Acad. Sci. USA 99, 11760–
Kiekhaefer, C.M., Grass, J.A., Johnson, K.D., Boyer, M.E., and Bres-
nick, E.H. (2002). Hematopoietic-specific activators establish an
overlapping pattern of histone acetylation and methylation within a
mammalian chromatin domain. Proc. Natl. Acad. Sci. USA 99,
Kihm, A.J., Kong, Y., Hong, W., Russell, J.E., Rouda, S., Adachi, K.,
Simon, M.C., Blobel, G.A., and Weiss, M.J. (2002). An abundant
erythroid protein that stabilizes free alpha-haemoglobin. Nature
Kosak, S.T., and Groudine, M. (2004). Form follows function: the
genomic organization of cellular differentiation. Genes Dev. 18,
Letting, D.L., Rakowski, C., Weiss, M.J., and Blobel, G.A. (2003).
Formation of a tissue-specific histone acetylation pattern by the
Letting, D.L., Chen, Y.Y., Rakowski, C., Reedy, S., and Blobel, G.A.
(2004). Context-dependent regulation of GATA-1 by friend of
GATA-1. Proc. Natl. Acad. Sci. USA 101, 476–481.
Merika, M., and Orkin, S.H. (1995). Functional synergy and physical
interactions of the erythroid transcription factor GATA-1 with the
Kruppel family proteins Sp1 and EKLF. Mol. Cell. Biol. 15, 2437–
Nichols, K.E., Crispino, J.D., Poncz, M., White, J.G., Orkin, S.H.,
Maris, J.M., and Weiss,M.J. (2000). Familial dyserythropoietic anae-
mia and thrombocytopenia due to an inherited mutation in GATA1.
Nat. Genet. 24, 266–270.
Osborne, C.S., Chakalova, L., Brown, K.E., Carter, D., Horton, A.,
Debrand, E., Goyenechea, B., Mitchell, J.A., Lopes, S., Reik, W., and
Fraser, P. (2004). Active genes dynamically colocalize to shared
sites of ongoing transcription. Nat. Genet. 36, 1065–1071.
Pal, S., Cantor, A.B., Johnson, K.D., Moran, T.B., Boyer, M.E., Orkin,
S.H., and Bresnick, E.H. (2004). Coregulator-dependent facilitation
of chromatin occupancy by GATA-1. Proc. Natl. Acad. Sci. USA
Palstra, R.J., Tolhuis, B., Splinter, E., Nijmeijer, R., Grosveld, F.,
and de Laat, W. (2003). The beta-globin nuclear compartment in
development and erythroid differentiation. Nat. Genet. 35, 190–194.
Patrinos, G.P., de Krom, M., de Boer, E., Langeveld, A., Imam, A.M.,
Strouboulis, J., de Laat, W., and Grosveld, F.G. (2004). Multiple
interactions between regulatory regions are required to stabilize an
active chromatin hub. Genes Dev. 18, 1495–1509.
Pevny, L., Simon, M.C., Robertson, E., Klein, W.H., Tsai, S.F., D’A-
gati, V., Orkin, S.H., and Costantini, F. (1991). Erythroid differentia-
tion in chimaeric mice blocked by a targeted mutation in the gene
for transcription factor GATA-1. Nature 349, 257–260.
Ragoczy, T., Telling, A., Sawado, T., Groudine, M., and Kosak, S.T.
(2003). A genetic analysis of chromosome territory looping: diverse
roles for distal regulatory elements. Chromosome Res. 11, 513–525.
Reik, A., Telling, A., Zitnik, G., Cimbora, D., Epner, E., and Groudine,
M. (1998). The locus control region is necessary for gene expression
in the human beta-globin locus but not the maintenance of an open
chromatin structure in erythroid cells. Mol. Cell. Biol. 18, 5992–6000.
Sawado, T., Halow, J., Bender, M.A., and Groudine, M. (2003). The
beta -globin locus control region (LCR) functions primarily by en-
hancing the transition from transcription initiation to elongation.
Genes Dev. 17, 1009–1018.
Schubeler, D., Groudine, M., and Bender, M.A. (2001). The murine
beta-globin locus control region regulates the rate of transcription
but not the hyperacetylation of histones at the active genes. Proc.
Natl. Acad. Sci. USA 98, 11432–11437.
Shang, Y., Hu, X., DiRenzo, J., Lazar, M.A., and Brown, M. (2000).
Cofactor dynamics and sufficiency in estrogen receptor-regulated
transcription. Cell 103, 843–852.
ABC-me: a novel mitochondrial transporter induced by GATA-1 dur-
ing erythroid differentiation. EMBO J. 19, 2492–2502.
blood cell differentiation. Curr. Opin. Genet. Dev. 8, 545–551.
Spilianakis, C.G., and Flavell, R.A. (2004). Long-range intrachromo-
somal interactions in the T helper type 2 cytokine locus. Nat. Immu-
nol. 5, 1017–1027.
Stamatoyannopoulos, J.A., Goodwin, A., Joyce, T., and Lowrey,
C.H. (1995). NF-E2 and GATA binding motifs are required for the
formation of DNase I hypersensitive site 4 of the human beta-globin
locus control region. EMBO J. 14, 106–116.
Tolhuis, B., Palstra, R.J., Splinter, E., Grosveld, F., and de Laat, W.
(2002). Looping and interaction between hypersensitive sites in the
active beta-globin locus. Mol. Cell 10, 1453–1465.
Tsang, A.P., Fujiwara, Y., Hom, D.B., and Orkin, S.H. (1998). Failure
GATA-1 transcriptional cofactor FOG. Genes Dev. 12, 1176–1188.
Valverde-Garduno, V., Guyot, B., Anguita, E., Hamlett, I., Porcher,
C., and Vyas, P. (2004). Differences in the chromatin structure and
cations for cis-element identification. Blood 104, 3106–3116.
Weiss, M.J., Yu, C., and Orkin, S.H. (1997). Erythroid-cell-specific
properties of transcription factor GATA-1 revealed by phenotypic
rescue of a gene-targeted cell line. Mol. Cell. Biol. 17, 1642–1651.
Welch, J.J., Watts, J.A., Vakoc, C.R., Yao, Y., Wang, H., Hardison,
R.C., Blobel, G.A., Chodosh, L.A., and Weiss, M.J. (2004). Global
regulation of erythroid gene expression by transcription factor
GATA-1. Blood 104, 3136–3147.