Developmental Cell, Vol. 3, 499–510, October, 2002, Copyright 2002 by Cell Press
Plzf Mediates Transcriptional Repression
of HoxD Gene Expression
through Chromatin Remodeling
SMRT, N-CoR, Sin-3, and, in turn, class I and II histone
deacetylases to the transcriptional complex (Hong et
al., 1997; He et al., 1998; Grignani et al., 1998; Lin et al.,
1998; David et al., 1998; Lemercier et al., 2002).
We have generated mice with a null mutation in Plzf,
which show striking patterning defects in both the limb
and axial skeleton, including homeotic transformations
of anterior skeletal elements in the developing limb into
posterior structures (Barna et al., 2000). These transfor-
mations are accompanied by the anteriorization and ec-
topic expression of each member of the 5? AbdB HoxD
gene complex in the developing hindlimb. These results
gene expression; however, the molecular mechanisms
by which Plzf would restrict the posterior boundaries of
expression of the HoxD genecomplex are still unknown.
Several models have been proposed to account for reg-
ulation of Hox gene expression which illustrate (1) the
importance of cis elements within these genes which
would regulate their transcription in space and time and
vent posterior Hox genes from being activated at an
earlier stage through a repressive chromatin configura-
tion (Dolle et al., 1989; van der Hoeven et al., 1996;
been , so far, implicated in this process.
activation of the AbdB HoxD gene complex through
binding to cis elements within Hox genes and recruit-
ment of histone deacetylases as well as Polycomb pro-
teins, in turn favoring the transition from a euchromatic
to a heterochromatic chromatin state. These results will
be discussed in the context of previous models of Hox
Maria Barna,1Taha Merghoub,1Jose ´ A. Costoya,1
Davide Ruggero,1Matthew Branford,1
Anna Bergia,2Bruno Samori,2
and Pier Paolo Pandolfi1,3
1Molecular Biology Program
Department of Pathology
Memorial Sloan-Kettering Cancer Center
New York, New York 10021
2Department of Biochemistry
University of Bologna and
The molecular mechanisms that regulate coordinated
and colinear activation of Hox gene expression in
space and time remain poorly understood. Here we
demonstrate that Plzf regulates the spatial expression
of the AbdB HoxD gene complex by binding to regula-
sion and can recruit histone deacetylases to these
sites. We show by scanning forced microscopy that
Plzf, via homodimerization, can form DNA loops and
bridge distant Plzf binding sites located within HoxD
gene regulatory elements. Furthermore, we demon-
teins on DNA. We propose a model by which the bal-
ance between activating morphogenic signals and
transcriptional repressors such as Plzf establishes
proper Hox gene expression boundaries in the limb
The spatial and temporal order of Hox gene activation
is colinear with the physical position of the genes along
their respective clusters (Gaunt et al., 1989; Dolle et al.,
1989; Izpisua-Belmonte et al., 1991). In vertebrates, the
transcriptional mechanisms involved in Hox gene regu-
lation are largely unknown. The PLZF (promyelocytic
leukemia zinc finger) gene was identified by virtue of its
involvement in chromosomal translocations associated
with acute promyelocytic leukemia (APL) (Chen et al.,
1993). PLZF is a nuclear protein (Reid et al., 1995) con-
taining at the C terminus nine Kru ¨ppel-type zinc-finger
domains, which recognize specific DNA sequences (Li
et al., 1997; Sitterlin et al., 1997). At the N terminus,
association and transcriptional repression when fused
to a heterologous DNA binding region (Bardwell and
Treisman, 1994; Dong et al., 1996). PLZF functions as
a transcriptional repressor through its ability to recruit,
via the BTB/POZ domain, nuclear corepressors such as
Selective Deregulation of Spatial but Not
Temporal Expression of the Abdb HoxD
Complex in Plzf?/?Mice
Throughout normal limb development, 5? HoxD genes
display restricted patterns of expression which corre-
late with both the temporal and spatial colinearity of
the complex. Whole-mount in situ hybridization of
Plzf?/?embryos did not reveal any premature temporal
activation of 5? HoxD genes, as Hoxd11 transcripts were
not expressed at an earlier time than wild-type embryos
(Figure 1A). However, when 5? HoxD transcripts are first
expressed in spatially restricted posterior domains of
the limb, their expression was anteriorized in Plzf?/?
embryos, showing a uniform expression of transcripts
along the entire extent of the hindlimb bud (Figure 1B;
and data not shown). As previously reported (Barna et
al., 2000), the Plzf?/?phenotype is largely restricted to
the hindlimb, which may be due to redundancy of Plzf
function in the forelimb because of the presence of a
Figure 1. Aberrant HoxD Gene Expression in
Plzf?/?Mice and Its Rescue in Plzf?/?Limb
(A) Expression of Hoxd11 in Plzf?/?and wild-
type embryos at 9.0 and 9.5 dpc.
ing hindlimb of Plzf?/?and wild-type controls
at 10.5 dpc.
(C) Northern blot analysis of Hoxd13 and
Hoxb2 in Plzf?/?and wild-type limb cells as
wellas inPlzf?/?limbcells followingtransfec-
tion of a Plzf expression vector or an empty
Plzf homolog. Similarly, these changes in Hox gene ex-
pression are limited to the hindlimb. Anteriorization of
HoxD expression occurred prior to any visible morpho-
logical defects in Plzf?/?embryos and is consistent with
the misexpression of 5?HoxD genes at later stages of
Plzf inactivation therefore results in loss of spatial colin-
ear expression of 5?HoxD genes, while their temporal
sequential activation does not appear to be affected.
1997; Sitterlin et al., 1997), as cis regulatory regions of
this gene have been previously identified (Gerard et al.,
1993). We identified five putative Plzf binding sites. Two
of the binding sites were located in the promoter of
highly conserved regulatory regions, known as region
VI and IX, present in the 3? UTR of the gene (Figure 2A).
We employedlimb extracts from Plzf?/?and wild-type
embryos to perform gel shift retardation assays utilizing
labeled oligonucleotide probes spanning the Plzf bind-
ing sites present within the Hoxd11 gene (Plzf binding
sites 1–5). A shift is present only in wild-type extracts,
but not in extracts from Plzf?/?mutants, is competed
by a cold oligonucleotide, and is absent with a mutant
oligonucleotide in all cases (Figure 2B). As a control,
of shifting a SP3 oligonucleotide (data not shown). The
presence of Plzf in the retarded DNA protein complex
was confirmed by Western blot analysis of the proteins
eluted from the region of the gel encompassing the
shifted band observed utilizing wild-type extracts (Fig-
ure 2C). We confirmed that this interaction is direct by
the ability of GST-Plzf to bind to an oligo spanning
Hoxd11 RRIX (Figure 2E).
To assess whether Plzf binds to Hox regulatory ele-
ments in vivo, we performed DNA immunoprecipitation
limb cells. We specifically detected Hoxd11 RRIX in the
DNA that was bound by Plzf in vivo (Figure 2D).
Plzf possesses an evolutionarily conserved BTB/POZ
domain that has been shown to mediate homodimeriza-
tion (Minucci et al., 2000). In order to assess if Plzf
the HoxD locus, we utilized scanning force microscopy
(SFM) to visualize stretches of Hoxd11 genomic DNA in
the presence of GST-Plzf. To this end, purified GST-
Plzf was incubated with different DNA fragments within
Hoxd11 (Figures 2A and 2G). The samples were depos-
ited on freshly cleaved mica and imaged by SFM. The
position of GST-Plzf along the DNA was determined by
Elevated HoxD Gene Expression Is Rescued by
the Reintroduction of Plzf in Limb Cultures
Derived from Plzf?/?Embryos
As alterations in limb development may result in the
ectopic activation of 5?HoxD genes, whose elevated
expression are a response to these morphological
changes, we sought to test whether the anteriorization
of HoxD transcripts in Plzf?/?embryos was a direct or
indirect consequence of Plzf activity. To this end, we
utilized low-density primary limb cells from wild-type
and Plzf?/?embryos to assay the level of expression
of one of the 5?HoxD genes that showed anteriorized
expression in the hindlimb bud, Hoxd13. Plzf?/?limb
cultures showed a marked increase in Hoxd13 tran-
scripts with respect to wild-type cultures, by Northern
blot analysis (Figure 1C; and data not shown). The rein-
troduction of Plzf in Plzf?/?limb cultures (Experimental
Procedures) reduced the level of Hoxd13 transcripts to
while it had no effect on the expression of Hoxb2, a Hox
therefore suggest that Plzf directly modulates the expres-
sion of 5?HoxD genes in primary limb cells.
Plzf Binds to Multiple Sites within the 5?HoxD
Gene Cluster and Mediates Long-Range
To understand the mechanisms by which Plzf mediates
for the presence of putative Plzf binding sites (Li et al.,
Role of Plzf in Hox Gene Regulation
Figure 2. Plzf Binds to Multiple Sites within Regulatory Regions of the Hoxd11 Gene and Mediates Long-Range Interactions through DNA
(A) Schematic representation of the Hoxd11 gene, which contains five (1–5) Plzf binding sites (indicated by the star). Evolutionarily conserved
regulatory regions within the gene are boxed and numbered.
(B) Electromobility shift analysis with oligonucleotides spanning each of the Plzf binding sites within Hoxd11 incubated with Plzf?/?and wild-
type limb extracts.
(C) Western blot analysis of the shifted complex observed in (B) utilizing wild-type limb extracts and an oligonucleotide spanning the Plzf
binding site within Hoxd11 regulatory region IX. Lane 1, COS-1 cells transfected with a Plzf expression vector. Lane 2, proteins from shifted
band in (B).
(D) ChIP assays of Hoxd11 RRIX in mouse limb cells utilizing a Plzf-specific Ab or a preimmune sera control (CNT).
(E) Electromobility shift analysis spanning the Plzf binding site within Hoxd11 regulatory region IX, utilizing GST-Plzf.
(F) Electromobility shift analysis with oligonucleotides spanning the Plzf binding site within Hoxd13 (Experimental Procedures) incubated with
Cos cell extracts and Cos cell extracts transfected with Plzf.
(G) SFM images of looped complexes of Hoxd11 DNA fragments mediated by GST-Plzf. Three representative examples are shown. A schematic
representation of the corresponding looped structures between different Plzf binding sites within Hoxd11 is illustrated.
(H) The x axis (ticks represent 25 nm2intervals) shows statistical analysis of the area protein bound on DNA at loop junctions (light yellow
bars); at single DNA binding sites (bright yellow bars); and on the surface of the mica substrate (red bars). The y axis shows the frequency
in percent at which GST-Plzf and Hoxd11 complexes are observed. The arrows indicate the highest frequency (percent of complexes observed)
of GST-Plzf bound to mica, at single binding sites, and at loop junctions.
measuring the DNA contour length from the center of
the protein to each DNA end (data not shown). Surpris-
ingly, we consistently observed complexes between
GST-Plzf and Hoxd11 genomic DNA fragments which
resulted in looped structures. In fact, 7%–10% of all
tive examples of looped complexes are shown (Figure
2G). Loop formation occurred between different Plzf
binding sites within the Hoxd11 locus, schematically
represented in Figure 2G.
To assess whether the DNA looping was mediated by
homodimerization of the Plzf protein, we measured the
size of the protein molecules either deposited on mica,
bound at single DNA sites, or at DNA loop junctions
(Nettikadan et al., 1996). The area of protein molecules
deposited on mica (n ? 1050) is centered on the value of
30 nm2that matches the expected value of a monomeric
molecules on single sites along the DNA chain showed
two main peaks: one at about 30 nm2and the other at
about double this value. The distribution of the dimen-
tions is shifted toward higher values and is centered
between 60 and 90 nm2(Figure 2H). These results
strongly suggest that Plzf tends to bind DNA as a dimer
or even as a trimer, mostly when DNA loops are formed.
Thus,Plzf iscapable ofbinding tomultiple siteswithin
Hoxd11 regulatory elements. Furthermore, sequence
analysis of other 5? HoxD genes identified Plzf binding
sites in all cases and oligos containing these sites that
were all bound by Plzf in gel shift analysis (see Experi-
mental Procedures for a detailed description of these
sites and Figure 2F). Moreover, SFM analysis of Plzf
bound to Hoxd11 genomic DNA provides evidence for
long-range interactions between distant Plzf binding
sites located within HoxD regulatory elements mediated
by Plzf di- or trimerization.
Plzf Directly Mediates Transcriptional Repression
of HoxD Gene Expression
to Hox gene regulatory elements, we generated two
luciferase reporter constructs corresponding to the
Hoxd11 promoter (d11 promoter) and regulatory region
IX (RRIX) in which we demonstrated the presence of
sequences bound by Plzf (see above). While Plzf?/?em-
bryos showed ectopic expression of Hoxd11 in more
anterior regions of the hindlimb, accompanied by ho-
meotic transformations, the forelimbs of these mice
were relatively unaffected (Barna et al., 2000). We there-
fore also tested whether the transcriptional repressive
abilities of Plzf differed in the forelimb with respect to
the hindlimb. Luciferase activity was assayed following
transfection of these constructs in low-density primary
limb cultures derived from Plzf?/?and wild-type em-
bryos. Strikingly, we observed a 3- to 4-fold increase in
the basal activity of the d11 promoter reporter and a
45-fold increase in basal activity of the RRIX reporter
in Plzf?/?hindlimb cells with respect to wild-type cells
(Figure 3A). However, we did not observe any statisti-
cally significant difference in the basal activity of the
RRIX reporter or the d11 promoter reporter between
wild-type and Plzf?/?forelimb cells (Figure 3A). Further-
more, the basal activity of both reporter constructs was
markedly higher in wild-type forelimbs in comparison to
wild-type hindlimbs, suggesting that Plzf is unable to
mediate transcriptional repression of these Hoxd11 re-
porter constructs in the context of the forelimb (Figure
3A). These results are consistent with the hindlimb-spe-
cific phenotype in Plzf?/?mice.
We next monitored whether we could rescue the ele-
vated luciferase activity following cotransfection of Plzf
cells. The RRIX reporter as well as the d11 promoter
reporter was repressed in a dose-dependent manner,
by increasing concentrations of Plzf. The repression
conferred by Plzf on the Hoxd11 reporter constructs
was almost 80% at its highest dose (Figure 3B). Thus,
Plzf is essential for the transcriptional repression of
Hoxd11 reporter constructs in the hindlimb.
Figure 3. PlzfMediatesTranscriptionalRepressionofHoxDRegula-
(A) Relative luciferase activity of Hoxd11 regulatory regions fused to
the SV40 minimal promoter corresponding to the Hoxd11 promoter
(pGL3-d11Prom.) and Regulatory Region IX (pGL3-RRIX) in Plzf?/?
and wild-type hindlimb and forelimb low density cultures.
(B) Rescue of elevated pGL3-d11Prom. and pGL3-RRIX luciferase
levels in Plzf?/?hindlimb cells following the reintroduction of Plzf in
a dose-dependent manner (Experimental Procedures).
(C) Plzf?/?limb cellswere transfectedwith thepGL3-RRIX luciferase
reporter, and Plzf in the presence or absence of TSA (Experimental
(D) ChIP assay of Hoxd11 RRIX in limb cells utilizing an anti-acet-
ylated Histone H3 antibody, preimmune sera control (CNT), or no
antibody (No Ab.). The PCR products from Plzf?/?and wild-type
samples were run simultaneously on the same gel.
Plzf Mediates Transcriptional Repression of HoxD
Reporter Constructs and Can Recruit
PLZF forms complexes with nuclear corepressors such
as SMRT, N-CoR, and Sin-3, thus recruiting histone de-
acetylases (HDACs) to the transcription complex, re-
sulting in nucleosome assembly and transcriptional
repression (see Introduction). To determine if HDAC-
associated enzymatic activity is required for the ability
of Plzf to repress the Hoxd11 reporter constructs, we
examined the effect of Trichostatin-A (TSA), a specific
TSA significantly reverted the transcriptional repressive
activity of Plzf (Figure 3C).
We next determined whether the chromatin state of
HoxD regulatory regions that were bound by Plzf were
altered by the absence of Plzf in vivo. To this end we
Role of Plzf in Hox Gene Regulation
Figure 4. The Transcriptional Repressive Activity of Plzf on HoxD Gene Expression in Anterior versus Posterior Micromass Cultures Is Antago-
nized by Posteriorizing Signals
(A) Anterior and posterior micromass limb cultures were established from 10.5 dpc wild-type embryos (Experimental Procedures).
(B) Transcriptional repression of Hoxd11 reporter gene constructs by Plzf in anterior versus posterior micromass cultures represented as
percent repression of the basal activity of Hoxd11 luciferase reporter. The activity of the reporter gene without cotransfection of Plzf is shown
(C) Electromobility shift analysis spanning the Plzf binding site within Hoxd11 regulatory region IX (RRIX) and the Hoxd11 promoter (Prom 1),
utilizing cell extracts obtained from anterior and posterior regions of the limb bud from Plzf?/?and wild-type embryos. The arrow indicates
the shifted band containing Plzf complexes.
(D) Immunoprecipitation of HDAC-1 in anterior and posterior regions of the limb bud with a Plzf antibody.
(E) Schematic representation of the location of the Plzf binding site (Plzf bs) and RARE within Hoxd11 regulatory region IX (RRIX). Cos-1 cells
were transfected with the pGL3-RRIX luciferase reporter, Plzf, or RAR? in the presence of RA (Experimental Procedures). Similar results were
obtained in limb cells (Figure 6G).
and wild-type hindlimb cells. We observed a marked
increase of acetylated histones on HoxD regulatory re-
gions in the absence of Plzf (Figure 3D). These results
therefore implicate HDAC-mediated transcription re-
posterior regions of the limb, we utilized a limb micro-
rior characteristics of limb bud cells in ex vivo cultures.
(Vogel and Tickle, 1993) (see Experimental Procedures).
Plzf showed a dose-dependent transcriptional repres-
sive ability in anterior limb micromass cultures (cells
corresponding to regions of the limb bud where HoxD
genes are excluded at 10.5 dpc). In contrast, even at
the highest dose of Plzf, Hoxd11 reporter expression
was not repressed in posterior limb micromass cultures
(cells corresponding to regions of the limb bud where
HoxD genes are expressed at 10.5 dpc) (Figures 4A
We next sought to determine the molecular basis for
this differential repressive ability of Plzf in anterior ver-
sus posterior cells. In gel shift experiments, Plzf was
bound to Hoxd11 regulatory elements and was able to
recruit HDAC in both anterior and posterior regions of
the limb (Figures 4C and 4D). In the limb bud, Shh or
Plzf Differentially Mediates Transcriptional
Repression in Anterior versus Posterior Limb
Micromass Cultures: Balance of trans-Acting
and trans-Repressing Factors
mation, and at 10.5 dpc, Plzf is present throughout the
limb (Barna et al., 2000) (data not shown). Plzf is there-
fore not restricted in expression to regions of the limb
where the HoxD genes are not expressed. In order to
test whether Plzf functionally possessed differential
transcriptional repressive properties in anterior versus
Figure 5. Plzf Tethers the Polycomb Protein
Bmi-1 to Hoxd11 Regulatory Regions, Pre-
venting Transcriptional Activity by trans-Act-
(A) Immunofluoresence of Cos-1 transfected
cells and primary limb cells with Plzf and
(B) Cos cells were transfected with Plzf and/
or Bmi-1, and coimmunoprecipitation experi-
ments were performed with anti-Bmi-1 and
anti-Plzf antibodies, or pre-immune sera.
(C–D) In-vitro pull-down assays performed
(G) alone with in vitro translated
proteins or total body embryo cell extracts.
(E) A biotinylated Hoxd11 RRIX oligo was in-
Plzf, and subsequent Western blot analysis
was performed with a Bmi-1 Ab.
fected with the Hoxd11 RRIX reporter con-
struct and Bmi-1.
Hoxd11 RRIX reporter construct and Bmi-1,
Plzf, and/or RAR? in the presence of RA (Ex-
RA has been shown to induce HoxD gene expression
when ectopically placed in anterior regions of the limb
et al., 1994; Johnson and Tabin, 1997). We tested
whether the presence of either RA or Shh would alter
the ability of Plzf to mediate transcription repression of
Hoxd11 regulatory elements. We also took advantage
of the fact that a nuclear hormone receptor, retinoic
acid receptor ? (RAR?)-responsive element (RARE) was
present within the RRIX reporter (Figure 4E). This RARE
has previously been shown to be important for correct
Hoxd11 expression in vivo (Gerard et al., 1996). Addition
of Shh or transfection of RAR? in the presence of RA
ity of Plzf (Figures 4E and 5G; and data not shown).
Therefore, posteriorizing signals derived from the de-
repressive ability of Plzf. This data provide a possible
mechanism for how Plzf differentially mediates tran-
scriptional repression of HoxD genes in anterior versus
posterior regions of the limb bud.
not bind to DNA directly. Therefore, how PcG proteins
would be tethered on DNA in order to mediate Hox gene
silencing remains unknown. Mammalian PcG proteins
Ring-1, Bmi-1, and hPc2 form a protein complex that is
localized to discrete nuclear structures known as PcG
bodies associated with heterochromatin (Saurin et al.,
1998). PLZF also localizes to discrete nuclear foci of
unknown functional significance (Reid et al., 1995). Due
to the ability of Plzf to directly mediate transcriptional
repression of Hox gene expression, we tested whether
Plzf would colocalize in the nucleus with PcG bodies.
nuclear foci both in transfected cells as well as primary
limb cultures, suggesting that Plzf is present within the
(Figure 5A). We next tested whether Plzf would physi-
cally interact with one of the mammalian PcG proteins,
Bmi-1, in vivo. Using coimmunoprecipitation assays, we
determined that Plzf associated with Bmi-1 (Figure 5B).
Plzf and Bmi-1 interaction was direct as assessed by
Furthermore, GST-Bmi-1was able to pulldown endoge-
determined whether Plzf and Bmi-1 would physically
associate on DNA, utililizing an oligo affinity binding
assay (OABA). Bmi-1 was only associated with RRIX
when Plzf was present (Figure 5E).
ciation of Plzf and Bmi-1, we performed at first trans-
repression assays on the Hoxd11 RRIX reporter in
Plzf?/?and wild-type limb cells. Strikingly, Bmi-1 was
Plzf Directly Tethers Polycomb on DNA which
Antagonizes Posteriorizing Signals in the Limb
Polycomb (PcG) family proteins are transcriptional re-
pressors of Hox genes; and they act to maintain correct
Hox gene expression boundaries through an epigenetic
ture (Pirrotta, 1998, Paro, 1990). Most PcG proteins do
Role of Plzf in Hox Gene Regulation
Figure 6. The Effect of Mutations in Plzf
Binding Sites within a Hoxd11/lacZ Reporter
Construct In Vivo
(A) Schematic representation of the Hoxd11/
lacZ reporter constructs utilized for the gen-
eration of transgenic mice.
(B) A stable transgenic line carrying the
Hoxd11/lacZ transgene was generated and
?-gal expression was monitored in 10.5 dpc
wild-type and Plzf?/?embryos.
(C) Mutations in all five Plzf binding sites
within the Hoxd11/lacZ transgene (Hoxd11/
lacZ mut) and a single mutation in the Plzf
binding site corresponding to Hoxd11 Regu-
latory Region IX (Hoxd11/lacZ mut bs5) were
generated and ?-gal expression was moni-
tored in embryos which expressed the trans-
gene. Arrow indicates ?-gal-positive cells
within the anterior margins of the limb bud.
able to repress the basal activity of Hoxd11 RRIX re-
porter in wild-type, but not in Plzf?/?limb cells (Figure
5F). We next tested whether the association between
Plzf and Bmi-1 would affect the function of other trans-
acting factors in the limb. Posteriorizing signals in the
limb such as retinoic acid (RA) have been shown to be
important in affecting the transcriptional state of Hox
genes in anterior tissues (Johnson and Tabin, 1997). We
took advantage of the fact that a RARE binding site for
factor, RAR?, was present in close vicinity to the Plzf
binding site (see previous paragraphs and Figure 4E).
In primary limb cells, RAR? acted as a potent transcrip-
tional activator of the HoxD11 RRIX reporter in the pres-
ence of RA (Figure 5G). Plzf was not able to mediate
transcriptional repression of the HoxD11 RRIX reporter,
in the presence of RAR?, although it was able to sup-
press expression to basal level of activity (Figure 5G).
Strikingly, cotransfection of Plzf and RAR? in the pres-
ence of Bmi-1 completely restored the transrepressive
potential of Plzf (Figure 5G).
Thus, Plzf directly interacts with Polycomb family
members, and this may serve to tether their chromatin
remodeling potential to specific cis-acting elements
within the Hox locus. Furthermore, the interaction be-
tween Plzf and Polycomb members such as Bmi-1 may
antagonize posterior activating signals in the limb bud
such as RA, which act on target nuclear hormone re-
Regulation of Hoxd11 Gene Expression
by Plzf In Vivo
Regulation, in vivo, of the Hoxd11 gene can be faithfully
studied in transgenic mice using a 11 kb fragment of
Hoxd11 genomic DNA, known as Hoxd11/lacZ which is
capable of reproducing the spatially restricted expres-
sion pattern of Hoxd11 in the hindlimb (Gerard et al.,
1993). First, we generated a stable line of Hoxd11/lacZ
transgenic mice to examine the expression of this con-
struct in Plzf?/?embryos. While wild-type embryos
showed a restricted expression of ?-gal to the posterior
mesoderm of the developing hindlimb, Plzf?/?embryos
displayed ?-gal expression throughout the entire ante-
rior-posterior extent of the limb bud (Figures 6A and 6B)
that mimicked the anteriorization of Hoxd11 transcripts
in Plzf?/?embryos (see Figure 1B).
Mutations in all of the Plzf binding sites within the
Hoxd11/lacZ reporter (Figure 6A; and Experimental Pro-
cedures) resulted in anteriorization of ?-gal expression
within the wild-type limb bud (Figure 6C) and fully repro-
duced the expression of the Hoxd11/lacZ reporter con-
struct in a Plzf?/?background (Figure 6B). Interestingly,
a single mutation in the Plzf binding site corresponding
to regulatory region IX of Hoxd11 (see above para-
graphs) also resulted in a partial anteriorization of ?-gal
expression which was never as extensive as when all
five Plzf binding sites were mutated (Figure 6C).
Plzf therefore directly acts in vivo to regulate the spa-
tial expression of Hoxd11 in the limb bud by binding to
regulatoryelements requiredforrestricted geneexpres-
sion boundaries. When these Plzf binding sites are mu-
anterior regions of the developing limb.
through the limb bud. Furthermore, we have demon-
strated that the reintroduction of Plzf in limb cultures
derived from Plzf?/?embryos represses HoxD gene ex-
pression. We have demonstrated the importance of Plzf
to bind to Hox gene regulatory elements in vivo within
lated expression of the transgene. Moreover, given the
presence of multiple Plzf binding sites bound by Plzf
within the HoxD locus and that all 5? HoxD genes are
anteriorized in Plzf?/?embryos, it is likely that Plzf can
regulate the expression of the entire 5? HoxD complex
by directly binding to regulatory elements within each
gene member of the complex. Furthermore, we have
demonstrated that the transcriptional activity of Plzf is,
in turn resulting in chromatin remodeling from an open
to a closed heterochromatic status via histone deacety-
lation. This is of particular relevance as several models
of Hox gene regulation have predicted the importance
of transitionsof thechromatin configurationin Hoxgene
regulation; however, the mechanism was largely un-
We have visualized DNA loop formation mediated by
Plzf molecules bound at distantly located sites. There-
fore, in addition to the ability of Plzf to remodel chroma-
tin, its homodimerization status could mediate long-
range interactions of DNA elements present within the
HoxD locus. Long-range interactions between regula-
tory regions containing different Plzf binding sites may
sion of distant Hox genes within the 5? HoxD locus.
Plzf in the Control of Spatial Expression
of the HoxD Gene Complex
Tremendous efforts have been made at understanding
rate expression of vertebrate Hox genes. Taken to-
gether, several experimental approaches have high-
lighted theimportance ofcis-acting regulatoryelements
as well as a higher order or global mechanism acting
reporter gene constructs have demonstrated the ability
of regulatory regions within Hox genes to reproduce
important patterns of expression of the endogenous
genes (e.g., Behringer et al., 1993; Gerard et al., 1993;
Beckers et al., 1996). On the other hand, temporal and
spatial colinearity of Hox gene expression may also be
regulated by long-range interactions with a “locus con-
trol region” functionally related to the Globin gene com-
plex (Hanscombe et al., 1991), or by the progressive
configuration (closed to open) (Duboule 1992; van der
Hoeven et al., 1996; Kondo et al., 1998). In fact, large
deletions near the 5? end of the HoxD complex result in
deregulation in the temporal colinearity of Hox gene
to a higher order silencing mechanism that would pre-
and Duboule 1999). It is intuitive that, if concomitantly
active, local regulatory influences would be subordinate
to this more general regulatory system. One question
tion of specific trans-acting factors which would inte-
grate both local and/or global regulatory influences in
order to mediate the correct expression of Hox genes
either in time or in space: Plzf can indeed serve this
Plzf inactivation results in a loss of the restricted spa-
tial expression of 5? HoxD genes, in that all the 5? HoxD
genes of the AbdB complex are ectopically expressed
Balance between trans-Actin and trans-Repressing
Signals in Regulation of Hox Gene
Expression in the Limb
We demonstrate that the ability of Plzf to mediate tran-
scriptional repression of Hox genes is severely compro-
mised in posterior regions of the limb (where HoxD
genes are expressed at this stage). Shh or RA has been
shown to induce HoxD genes in a sequential manner
when ectopically placed in anterior regions of the limb
where HoxD genes are not normally expressed (Helms
et al., 1994; Johnson and Tabin, 1997). This suggests
of HoxD genes in posterior regions of the limb, although
the mechanisms of this activation remain unclear. Here
we show that posteriorizing factors such as Shh or RA
can overcome Plzf trans-repressive ability. Therefore, a
rizing factors may ultimately regulate the chromatin
state of the 5? HoxD locus (Figure 7). This is supported
by the fact that RAR? in its liganded state remodels
ases (HATs) (reviewed in McEwan, 2000). This activity
could be counterbalanced by the ability of Plzf to recruit
HDAC, thus favoring a more heterochromatic chromatin
Plzf Recruitment of Polycomb: Tilting
the Balance between trans-Acting
and trans-Repressing Signals
We show that Plzf directly interacts with PcG members
such as Bmi-1 and colocalizes with PcG bodies. Bmi-1
Role of Plzf in Hox Gene Regulation
eling. Like the gap proteins, Plzf specifically binds DNA,
thus directly repressing Hox gene expression. However,
unlike gap, Plzf expression is maintained throughout
tion between PLZF and PcG in vertebrate Hox gene
regulation and demonstrated that recruitment of PcG
may play an important role in preventing activation of
Hox genes by transcriptional activators such as RAR?.
Therefore, the Plzf and PcG complex might tilt the bal-
ance between trans-acting and trans-repressing fac-
tors, favoring maintenance of the transcriptional re-
pressed state (Figure 7). In agreement with this notion,
PcG have been already implicated in limiting the acces-
sibility of trans-acting factors to DNA (Zink and Paro,
1995). Furthermore, M33 a PcG member that belongs
to the same complex as Bmi-1 has been implicated in
defining access to RAREs localized in regulatory ele-
ments present within Hox genes, including Hoxd11, as
M33-deficient mice show altered RA sensitivity (Core et
al., 1997; Bel-Vialar et al., 2000).
Although our data underscore the importance of Plzf/
PcG interactions in the limb, the homeotic transforma-
tions in axial skeletal structures observed in both Plzf
and Bmi-1 null mice (van der Lugt et al., 1994, 1996;
Alkema et al., 1995; Barna et al., 2000) suggest that this
expression patterns in other embryonic structures.
Figure 7. Model for the Role of Plzf in Mediating Transcriptional
Repression of the HoxD Gene Locus
Plzf represses the expression of genes belong to the HoxD cluster
in the limb by directly binding to regulatory elements and can recruit
histone deacetylases to these sites. Plzf may therefore regulate the
remodeling, changing the chromatin conformation from open to
closed. In addition, Plzf can mediate long-distance interactions be-
ization. The transcriptional repressive ability of Plzf is antagonized
in posterior regions of the limb by posteriorizing signals such as RA
the trans-acting activity of these morphogens, thereby establishing
ing the spatial expression of HoxD gene expression in the limb bud.
Implications for Aberrant HOX Gene Expression
and Polycomb Function in Leukemogenesis
Deregulated function and/or expression of specific HOX
genes have being directly implicated in leukemogenesis
(reviewed in Look, 1997). In APL, as a consequence
of translocations between chromosome 11 and 17, the
PLZF gene fuses to the RAR? gene resulting in two
fusion genes which are coexpressed in the leukemic
blast: PLZF-RAR? and RAR?-PLZF (reviewed in Rego
and Pandolfi, 2001). Both proteins can act as dominant-
negative PLZF mutants (He et al., 2000). It will be there-
fore important to determine the expression pattern of
potential target HOX genes whose correct expression
of PLZF function and the consequent aberrant expres-
sion of HOX genes could in turn participate in leukemo-
In APL, PLZF-RAR? can also act as a potent and RA-
insensitive transcriptional repressor of RAR? through
aberrant nuclear corepressor/HDAC associations, thus
rendering the APL blasts unresponsive to the differenti-
ating effects of RA (Grignani et al., 1998; He et al., 1998;
Lin et al., 1998). Surprisingly, however, PLZF-RAR? can
exert this role, at least in part, epigenetically since RA
can readily induce its physical degradation (Rego et al.,
2000). The fact that Bmi-1 and Plzf physically interact
may provide a mechanism underlying this aberrant epi-
genetic repressive ability by PLZF-RAR?. The recruit-
ment of BMI-1 to the leukemogenic transcription com-
plex may prevent proper transcriptional activation of
RAR? target genes even in the absence of the fusion
hasbeen previouslyshowntoregulate correctHoxgene
expression boundaries in vertebrates (van der Lugt et
al., 1994; Alkema et al., 1995). The mechanism by which
PcG members such as Bmi-1 would act as a trans-
repressive factor of target Hox gene expression re-
mained largely unknown. Our data indicate that Plzf
directly binds Hox regulatory elements and can act to
tether PcG members on DNA. Plzf directly interacts with
Bmi-1, and it may interact with other PcG members
During Drosophila embryogenesis the transcriptional
activity of gap and pair-rule transcription factors plays
pattern of Hox genes. However, due to the transient
of embryonic development, the maintenance of expres-
sion boundaries of homeotic genes requires the activity
of two antagonistic group of genes, PcG and the
trithorax group (TrxG) (Kennison, 1993; Simon, 1995;
transcriptional repression of homeotic genes through an
genomic DNA carrying a mutation in this binding site into the
Whole-Mount In Situ Hybridization
Whole-mount in situ hybridizations were carried out as previously
described (Barna et al., 2000).
Electrophoretic Mobility Shift Assay
Hindlimbs were removed from embryos, trypsinized, and homoge-
nized in 400 mM NaCL, 10 mM HEPES (pH 7.9), 0.1 mM EGTA, 0.5
mM DTT, 5% Glycerol, 0.5 mM PMSF, and 1% Triton. Lysates were
ultracentrifuged at 35,000 ? g for 30 min at 4?C. The supernatant
was collected and utilized for electrophoretic mobility shift assay
(EMSA). The following oligonucleotides were employed in EMSA
reactions spanning Plzf binding sites within Hoxd11: RRIX (5? CTT
CCAAAATGTCAAGGTCATCACCTTTAACCTCT 3?), RRVI (5? GGG
AACATGGTAAATGTAAACATCCCTTTC 3?), Intron (5? GGGCGTG
AACACATGTCCACGCCGCACTCT3?), Promoter 1 (5? CCAACACAG
TGAAAGCTCCAAGAGACTTGA 3?), and Promoter 2 (5? GCAGA
GAAA TATGTAAATCAGGGCTCCCTG 3?). Binding reaction mixes
for gel retardation assays were carried out as previously described
(Zhong et al., 1999). The shifted band observed utilizing RRIX oligos
was cut out of the gel and analyzed by SDS-PAGE, followed by
Western blot analysis utilizing an anti-Plzf antibody (Barna et al.,
2000). Plzf consensus binding sites (underlined) shifted in EMSA
in all the members of the Abdb Hox gene cluster: Hoxd13 (5? AAGAC
ATCTGGTTCCAGAAC3?); HoxD12 (5? CCGGGCGTGAACACATGTC
CACGCCCGCAC 3?); HoxD10(5?ACAACAAAAGAGCTAAAAGGAGA
HoxD9(5?GGAGCCCTATTCTATGTAAATGTCCCTCAT 3?). (Zappa-
vigna et al., 1991; Renucci et al., 1992; Gerard et al., 1996).
Northern Blot Analysis
Total RNA was prepared from limb cultures using Trizol reagent
(GIBCO-BRL). For Northern blot analysis, denatured total RNA (10
?g) was hybridized with the mouse Hoxd13, Hoxb2, and GAPDH
fragments as a probe.
Cell Culture and Transfection
Hindlimb buds were dissected from embryos and the ectoderm
removed after soaking the limbs in 2% trypsin (GIBCO, 1:250) in
calcium and magnesium free Hanks buffered salt solution (GIBCO),
pH 7.4, for 20–30 min. The mesenchyme was disaggregated and
the cells centrifuged. Low-density monolayer layer cultures were
obtained by resuspending cells in CMRL medium (GIBCO) with fetal
bovine serum (FBS) and seeding at a concentration of 1,000,000
primary cells in a 35 mm dish. High-density micromass limb bud
cultures of 10.5 dpc mouse embryos were prepared as described
to posterior limb bud cells at a concentration of 1 ?g/ml to maintain
polarizing activity (Vogel and Tickle 1993). Primary limb bud cells
were transfected using lipofectamine plus reagent (GIBCO). The
luciferase reporter plasmids (400 ng) were cotransfected with in-
creasing concentrations of the Plzf expression vector or an empty
vector (72, 180, and 360 ng). The RRIX reporter plasmid (400 ng)
was cotransfected with Bmi-1 (424 ng), RAR? (424 ng), Shh (480
ng). For cells transfected with a RAR? expression plasmid, RA was
supplemented in the media at a concentration of 30 ng/ml. In cases
where multiple expression plasmid were used, the plasmids were
cotransfected at a 1:1 molar ratio. The TK-?-galactosidase expres-
sion vector or TK-Renilla (25 ng) was also cotransfected in order to
normalize for transfection efficiency. Forty-eight hours later, cells
were harvested and prepared for FACS sorting or determination of
luciferase activity with a luminometer according to the manufactur-
er’s instructions (Promega). For the TSA experiments, limb cells
were treated with the drug at a concentration of 200 nM for 48 hr,
and the medium containing the drug was replaced every 12 hr.
DNA Immunoprecipitation Assay
Single cell suspensions of limbs derived from Plzf?/?and wild-type
embryos were prepared. Chromatin immunoprecipitation assays
were performed using the chromatin immunoprecipitation assay kit
(Upstate Biotech) according to the manufacturer’s directions. PLZF
Ab or rabbit preimmune sera were utilized. Alternatively, an anti-
acetylated Histone H3 antibody was employed (Upstate Biotech).
ThePCRreaction wascarriedoutwithprimers spanningtheHoxd11
regulatory region IX: 5? AAGATGCACAGCAGCTCATG 3? and 5? GTC
TGGATGTATGAGCCTG 3?. ThePCR product was runon an agarose
gel and subjected to Southern blot analysis with the internal oligo:
5? GAATAATTAGGCGCCTTAAAGT 3?.
A genomic fragment of 480 bp corresponding to the regulatory
element located in region IX of the mouse Hoxd11 gene (Gerard et
al., 1993) was amplified by polymerase chain reaction using two
specific primers (5? AAGATGCACAGCAGCTCATG 3? and 5? ACTG
CAGCTCTTCATTACAG 3?). This fragment was subcloned into the
pCR 2.1 and cloned into pGL3-Promoter vector (Promega) digested
with Xho1and Sac1.This gavethe pGL3-RRIXconstruct. Agenomic
fragment of 1.5 kb corresponding to the promoter region of the
mouse HoxD-11 gene was released from the pGemE/ElacZpA con-
struct (Gerard et al., 1993) using PstI and SalI and cloned into pSP72
vector (Promega) digested with the same restriction enzyme sites.
The cloned fragment was subsequently excised with HindIII and
same restriction enzymes. This gave the pGL3-Promoter construct.
The Hoxd11/lacZ mut construct was generated by introducing
mutations in all five Plzf binding sites within the pGemE/ElacZpA
Mutagenesis Kit (Stratagene) following the manufacturer’s sug-
gested protocol employing oligonucleotides containing the desired
mutation (underlined) in Plzf binding site 1 (5? GGGCTCCAACACAG
TGAAAGGGGGAAGAGACTTGAACACAGAAG3?); Plzf binding site 2
3?); Plzf binding site 3 (5? GTGTCCGGGCGTGAACACAGGGGCAC
GCCGCACTCTACTGTGC3?); Plzf binding site 4 (5?GGGGAACATGG
TAAAGGGGAACATCCCTTTCCAATTTTACTGCC3?), and Plzf bind-
ing site 5 (5?GTATGCCTTTGAACTTCCAAAAGGGGAAGGTCATCAC
CTTTAACCTCTC3?). The insertion of desired mutations was verified
tion in Plzf binding site 5 was created by recloning fragments of
Scanning Force Microscopy
Hoxd11 genomic DNA fragments were incubated with GST, GST-
PLZF, and PLZF proteins for half an hour on ice in binding buffer
(400 mM KCL, 200 mM HEPES, pH 7.5, 10 mM MgCl2, 1 mM EGTA,
5 mM DTT). The reaction mixtures were diluted 10-fold with 10
mM MgCl2just before the deposition on freshly cleaved mica. The
samples were prepared with three different protein to DNA molar
ratios, 2, 4, and 8, such that the initial concentration of DNA was
40 nM and that of the protein ranged from 80 to 320 nM. After a
deposition time of about 3 min, the surface was rinsed with water
and dried with a gentle flux of nitrogen. Imaging was performed in
tapping mode with PointProbe noncontact silicon probes (Nano-
Sensors, Germany) on a NanoScope IIIa scanning force microscope
system equipped with a multimode head and a type E piezoelectric
scanner (Digital Instruments). Images were recorded with a 10 ? 15
mm linear scanning speed at a sampling density of 4 ? 9 nm2
per pixel. Raw SFM images were processed only for background
removal (flattening) using the microscope manufacturer’s image-
processing software. DNA molecule lengths were measured from
the SFM images using ALEX, a software package written for probe
microscopy image processing (Rivetti et al., 1996).
GST Pull-Down Assays
In vitro pull-down analysis was performed using GST-fused PLZF
and Bmi-1 proteins and GST alone (from pGEX5T; Pharmacia) as
used in the protein binding experiments using either total body
ern blot analysis was carried out for the GST pull-down assay on
Role of Plzf in Hox Gene Regulation
total body embryo extracts with a Plzf monoclonal antibody (Onco-
gene) at 2 ?g/ml concentration.
Behringer, R.R., Crotty, D.A., Tennyson, V.M., Brinster, R.L., Pal-
box of the Hox-1.4 gene direct tissue-specific expression of lacZ
during mouse development. Development 117, 823–833.
(Zhong et al., 1999). Anti-PLZF (Barna et al., 2000) was used at a
dilution of 1:200 and anti-Bmi-1 (kind gift of M. van Lohuizen) at
Bel-Vialar, S., Core, N., Terranova, R., Goudot, V., Boned, A., and
Djabali, M. (2000). Altered retinoic acid sensitivity and temporal ex-
pression of Hox genes in polycomb-M33-deficient mice. Dev. Biol.
Chen, Z., Brand, N.J., Chen, A., Chen, S.J., Tong, J.H., Wang, Z.Y.,
Waxman, S., and Zelent, A. (1993). Fusion between a novel Kruppel-
a variantt (11;17)translocation associated withacute promyelocytic
leukaemia. EMBO J. 12, 1161–1167.
Oligo Affinity Binding Assay
Theoligo affinitybindingassay (OABA)wasperformed aspreviously
described (Zhong et al., 1999). Biotinylated RRIX oligo and mutated
RRIX (100 ng; sequences as in EMSA assays; GeneLink) and Cos-1
cell lysates transfected with Bmi-1 alone or PLZF and Bmi-1 (70 ?g)
Core, N., Bel, S., Gaunt, S.J., Aurrand-Louis, M., Pearce, J., Fisher,
A., and Djabali, M. (1997). Altered cellular proliferation and meso-
derm patterning in Polycomb-M33-deficient mice. Development
Production of Transgenic Mice
For injection, the pGemE/ElacZpA (Gerard et al., 1993), Hoxd11/
lacZ mut, and Hoxd11/lacZ bs5 mut constructs were excised with
NsiI, and a fragment of 10.7 kb was injected in all cases. A Plzf?/?
transgenic line carrying the Hoxd11/lacZ transgene was obtained
injecting this construct into Plzf?/?eggs. In addition, all the con-
structs were tested in transient assays using F1 eggs. Eleven em-
transient assays, and all embryos showed a posteriorly restricted
expression pattern in the hindlimb as previously described (Gerard
et al., 1993). Four embryos expressing the Hoxd11/lacZ mut trans-
gene were obtained, and two embryos showed an anteriorization
of ?-gal expression throughout the hindlimb that was never ob-
served in the Hoxd11/lacZ transgene. Eight embryos expressing the
Hoxd11/lacZ bs5 mut transgene were obtained, and three embryos
showed a partial anteriorization of ?-gal expression in the hindlimb
that was never observed in the Hoxd11/lacZ transgene.
David, G., Alland, L., Hong, S.H., Wong, C.W., DePinho, R.A., and
Dejean, A. (1998). Histone deacetylase associated with mSin3A me-
diates repression by the acute promyelocytic leukemia-associated
PLZF protein. Oncogene 16, 2549–2556.
Dolle, P., Izpisua-Belmonte, J.C., Falkenstein, H., Renucci, A., and
plex homoeobox-containing genes during limb pattern formation.
Nature 342, 767–772.
Dong, S., Zhu, J., Reid, A., Strutt, P., Guidez, F., Zhong, H.J., Wang,
Z.Y., Licht, J., Waxman, S., Chomienne, C., et al. (1996). Amino-
terminal protein-protein interaction motif (POZ-domain) is responsi-
ble for activities of the promyelocytic leukemia zinc finger-retinoic
acid receptor-alpha fusion protein. Proc. Natl. Acad. Sci. USA 93,
Duboule, D. (1992). The vertebrate limb: a model system to study
the Hox/HOM gene network during development and evolution. Bi-
oessays 14, 375–384.
hyde, 5 mM EDTA, 2 mM MgCl2, 1? PBS) for 2 hr. The embryos
were washed in detergent rinse (2 mM MgCl2, 0.2% NP-40, 1? PBS)
two times for 15 min on an orbital shaker. The embryos were next
incubated in staining solution (5 mM potassium ferricyanide, 5 mM
ferrocyanide, 2 mM MgCl2, 2 mg/ml X-gal [dissolved in DMSO], 1?
PBS) overnight at 30?C.
Gaul, U.,and Jackle, H.(1990). Role ofgap genes inearly Drosophila
development. Adv. Genet. 27, 239–275.
Gaunt, S.J., Krumlauf, R., and Duboule, D. (1989). Mouse homeo-
genes within a subfamily, Hox-1.4, -2.6 and -5.1, display similar
anteroposterior domains of expression in the embryo, but show
stage- and tissue-dependent differences in their regulation. Devel-
opment 107, 131–141.
of regulatory elements involved in the activation of the Hoxd-11
gene during late gastrulation. EMBO J. 12, 3539–3550.
We thank Lee Niswander foruseful discussions and for critical read-
ing of the manuscript; Denis Duboule and Mathew Gerard for the
Hoxd11 genomic region, the Hoxd11-LacZ expression vector, and
for suggestions; Maarten van Lohuizen for advice and the Bmi-1
antibody; Willy Mark, Joanne Ingenito, and the whole staff of the
transgenic facility for injections of multiple constructs. P.P.P. is
a Scholar of the Leukemia and Lymphoma Society (LLS). T.M. is
supported by a fellowship from LLS. J.A.C. is supported by a FPI
fellowship of the Spanish Ministry for Education Culture and Sports.
This work was supported by the NIH (CA-71692 awarded to P.P.P.)
and MURST (Biotechnology Program and PRIN to B.S.).
Gerard, M., Chen, J., Gronemeyer, H., Chambon, P., Duboule, D.,
and Zakany, J. (1996). In vivo targeted mutagenesis of a regulatory
element required for positoning the HoxD-11 and HoxD-10 expres-
sion boundaries. Genes Dev. 10, 2326–2334.
Grignani, F., De Matteis, S., Nervi, C., Tomassoni, L., Gelmetti, V.,
Cioce, M., Fanelli, M., Ruthardt, M., Ferrara, F.F., Zamir, I., et al.
(1998). Fusion proteins of the retinoic acid receptor-alpha recruit
histone deacetylase in promyelocytic leukaemia. Nature 391, 815–8.
Hanscombe, O., Whyatt, D., Fraser, P., Yannoutsos, N., Greaves,
D., Dillon, N., and Grosveld, F. (1991). Importance of globin gene
order for correct developmental expression. Genes Dev. 5, 1387–
Received: July 8, 2002
Revised: August 8, 2002
He, L.Z., Guidez, F., Tribioli, C., Peruzzi, D., Ruthardt, M., Zelent,
A., and Pandolfi, P.P. (1998). Distinct interactions of PML-RARalpha
and PLZF-RARalpha with co-repressors determine differential re-
sponses to RA in APL. Nat. Genet. 18, 126–135.
Alkema, M.J., van der Lugt, N.M., Bobeldijk, R.C., Berns, A., and
van Lohuizen, M. (1995). Transformation of axial skeleton due to
overexpression of bmi-1 in transgenic mice. Nature 374, 724–727.
Bardwell, V.J., and Treisman, R. (1994). The POZ domain: a con-
served protein-protein interaction motif. Genes Dev. 8, 1664–1677.
Barna, M., Hawe, N., Niswander, L., and Pandolfi, P.P. (2000). Plzf
Beckers, J., Gerard, M., and Duboule, D. (1996). Transgenic analysis
of a potential Hoxd-11 limb regulatory element present in tetrapods
and fish. Dev. Biol. 180, 543–553.
He, L., Bhaumik, M., Tribioli, C., Rego, E.M., Ivins, S., Zelent, A.,
and Pandolfi, P.P. (2000). Two critical hits for promyelocytic leuke-
mia. Mol. Cell 6, 1131–1141.
Helms, J., Thaller, C., and Eichele, G. (1994). Relationship between
retinoic acid andsonic hedgehog, two polarizing signalsin the chick
wing bud. Development 120, 3267–3274.
Hong, S.H., David, G., Wong, C.W., Dejean, A., and Privalsky, M.L.
(1997). SMRT corepressor interacts with PLZF and with the PML-
retinoic acid receptor alpha (RARalpha) and PLZF-RARalpha onco-
proteins associated with acute promyelocytic leukemia. Proc. Natl.
Acad. Sci. USA 94, 9028–9033.
Izpisua-Belmonte, J.C., Falkenstein, H., Dolle, P., Renucci, A., and
Duboule, D. (1991). Murine genes related to the Drosophila AbdB
homeotic genes are sequentially expressed during development of
the posterior part of the body. EMBO J. 10, 2279–2289.
Johnson, R.L., and Tabin, C.J. (1997). Molecular models for verte-
brate limb development. Cell 90, 979–990.
Kennison, J.A. (1993). Transcriptional activation of Drosophila ho-
meotic genes from distant regulatory elements. Trends Genet. 9,
Kondo, T., and Duboule, D. (1999). Breaking colinearity in the mouse
HoxD complex. Cell 97, 407–417.
Kondo, T., Zakany, J., and Duboule, D. (1998). Control of colinearity
in AbdB genes of the mouse HoxD complex. Mol. Cell 1, 289–300.
Lemercier, C., Brocard, M.P., Puvion-Dutilleul, F., Kao, H.Y., Albagli,
O., and Khochbin, S. (2002). Class II histone deacetylases are di-
rectly recruited by BCL6 transcriptional repressor. J. Biol. Chem.
Li, J.Y., English, M.A., Ball, H.J., Yeyati, P.L., Waxman, S., and Licht,
J.D. (1997). Sequence-specific DNA binding and transcriptional reg-
ulation by the promyelocytic leukemia zinc finger protein. J. Biol.
Chem. 272, 22447–22455.
Lin, R.J., Nagy, L., Inoue, S., Shao, W., Miller, W.H., Jr., and Evans,
R.M. (1998). Role of the histone deacetylase complex in acute pro-
myelocytic leukaemia. Nature 391, 811–814.
Look, A.T. (1997). Oncogenic transcription factors in the human
acute leukemias. Science 278, 1059–1064.
McEwan, I.J. (2000). Gene regulation through chromatin remodeling
by members of the nuclear receptor superfamily. Biochem. Soc.
Trans. 28, 369–373.
Minucci, S., Maccarana, M., Cioce, M., De Luca, P., Gelmetti, V.,
Segalla, S., Di Croce, L., Giavara, S., Matteucci, C., Gobbi, A., et al.
(2000). Oligomerization of RAR and AML1 transcription factors as
a novel mechanism of oncogenic activation. Mol. Cell 5, 811–20.
Nettikadan, S., Tokumasu, F., and Takeyasu, K. (1996). Quantitative
analysis of the transcription factor AP2 binding to DNA by atomic
force microscopy. Biochem. Biophys. Res. Commun. 226, 645–649.
Paro, R. (1990). Imprinting a determined state into the chromatin of
Drosophila. Trends Genet. 6, 416–421.
Pirrotta, V. (1998). Polycombing the genome: PcG, trxG, and chro-
matin silencing. Cell 93, 333–336.
Rego, E.M., He, L.-Z., Warrell, R.P., Jr., Wang, Z.-G., and Pandolfi,
P.P. (2000). RA and As2O3treatment in transgenic models of APL
unravel the distinct nature of the leukemogenic process induced by
USA 97, 10173–10178.
Rego, E.M., and Pandolfi, P.P. (2001). Analysis of the molecular
Hematol. 38, 54–70.
Reid, A., Gould, A., Brand, N., Cook, M., Strutt, P., Li, J., Licht,
J.D., Waxman, S., Krumlauf, R., and Zelent, A. (1995). Leukemia
translocation gene, PLZF, is expressed with a speckled nuclear
pattern in early hematopoietic progenitors. Blood 86, 4544–4552.
Renucci, A., Zappavigna, V., Zakany, J., Izpisua-Belmonte, J.C.,
Burki, K., and Duboule, D. (1992). Comparison of mouse and human
HOX-4 complexes defines conserved sequences involved in the
regulation of Hox-4.4. EMBO J. 11, 1459–1468.
Rivetti, C., Guthold, M., and Bustamante, C. (1996). Scanning force
trapping studied by statistical polymer chain analysis. J. Mol. Biol.
Saurin, A.J., Shiels, C., Williamson, J., Satijn, D.P., Otte, A.P., Sheer,
D., and Freemont, P.S. (1998). The human polycomb group complex
associates with pericentromeric heterochromatin to form a novel
nuclear domain. J. Cell Biol. 142, 887–898.
Simon, J. (1995). Locking in stable states of gene expression: tran-
scriptional control during Drosophila development. Curr. Opin. Cell
Biol. 7, 376–385.
Sitterlin, D., Tiollais, P., and Transy, C. (1997). The RAR alpha-PLZF
chimera associated with Acute Promyelocytic Leukemia has re-
tained a sequence-specific DNA-binding domain. Oncogene 14,
van der Hoeven, F., Zakany, J., and Duboule, D. (1996). Gene trans-
positions in the HoxD complex reveal a hierarchy of regulatory con-
trols. Cell 85, 1025–1035.
van der Lugt, N.M., Domen, J., Linders, K., van Roon, M., Robanus-
Maandag, E., te Riele, H., van der Valk, M., Deschamps, J., Sofro-
niew, M., van Lohuizen, M., et al. (1994). Posterior transformation,
neurological abnormalities, and severe hematopoietic defects in
mice with a targeted deletion of the bmi-1 proto-oncogene. Genes
Dev. 8, 757–769.
The Polycomb-group homolog Bmi-1 is a regulator of murine Hox
gene expression. Mech. Dev. 58, 153–164.
Vogel, A., and Tickle, C. (1993). FGF-4 maintains polarizing activity
of posterior limb bud cells in vivo and in vitro. Development 119,
Yoshida, M., Kijima, M., Akita, M., and Beppu, T. (1990). Potent and
specific inhibition of mammalian histone deacetylase both in vivo
and in vitro by trichostatin A. J. Biol. Chem. 265, 17174–17179.
Zappavigna, V., Renucci, A., Izpisua-Belmonte, J.C., Urier, G.,
Peschle, C., and Duboule, D. (1991). HOX4 genes encode transcrip-
tion factors with potential auto- and cross-regulatory capacities.
EMBO J. 10, 4177–4187.
Zhong, S., Delva, L., Cenciarelli, C., Gandini, D., Zhang, H., Fagioli,
M., Kalantry, S., and Pandolfi, P.P. (1999). A RA-dependent tumor-
growth suppressive transcription complex is the target of the PML-
RAR? and T18 oncoproteins. Nat. Genet. 23, 287–295.
Zink, D., and Paro, R. (1995). Drosophila Polycomb-group regulated
chromatin inhibits the accessibility of a trans-activator to its target
DNA. EMBO J. 14, 5660–5671.