Báo cáo khoa học: The Runx3 distal transcript encodes an additional transcriptional activation domain doc

The Runx3 distal transcript encodes an additional
transcriptional activation domain
David D. Chung
1
, Kazuho Honda
2,
*, Lorraine Cafuir
2
, Marcia McDuffie
2,3
and David Wotton
1
1 Center for Cell Signaling and Department of Biochemistry, and Molecular Genetics, University of Virginia School of Medicine,
Charlottesville, VA, USA
2 Department of Microbiology, University of Virginia School of Medicine, Charlottesville, VA, USA
3 Department of Medicine, University of Virginia School of Medicine, Charlottesville, VA, USA
The transcriptional regulator Runx3 is expressed in
several hematopoietic cell lineages at different stages of
development, including CD8
+
thymocytes and periph-
eral CD8
+
T cells. Two independent lines of Runx3
mutant mice, in which Runx3 targets different molecu-
lar domains, have been established via induced muta-
tions [1,2]. Hematopoietic precursors from both lines
produce decreased numbers of peripheral CD8
+
T cells that aberrantly coexpress CD4 [3,4]. These cells
fail to expand into functional cytotoxic CD8

+
T-cell
populations in vivo or in vitro. Mutational analysis of
the CD4 silencer has shown that transcriptional repres-
sion in CD8
+
T cells requires binding of Runx3 to
runt-specific binding sites in the first intron of Cd4.
Keywords
cell differentiation; runt domain; Runx3;
transcription
Correspondence
M. McDuffie, University of Virginia School
of Medicine, Aurbach Medical Research
Building, Box 801390 (FedEx: room 1253),
Charlottesville, VA 22908, USA
Fax: +1 434 243 9143,
Tel: +1 434 924 1707
E-mail:
D. Wotton, Center for Cell Signaling,
University of Virginia, Room 7008,
Hospital West, Hsc 800577, Charlottesville,
VA 22908, USA
Fax: +1 434 924 1236,
Tel: +1 434 243 6752
E-mail:
*Present address
Department of Pathology, Tokyo Women’s
Medical University, Tokyo, Japan
(Received 5 February 2007, revised 5 April

2007, accepted 9 May 2007)
doi:10.1111/j.1742-4658.2007.05875.x
The runt family transcriptional regulator, Runx3, is upregulated during the
differentiation of CD8 single-positive thymocytes and is expressed in peri-
pheral CD8
+
T cells. Mice carrying targeted deletions in Runx3 have
severe defects in the development and activation of CD8
+
T cells, resulting
in decreased CD8
+
T-cell numbers, aberrant coexpression of CD4, and
failure to expand CD8
+
effector cells after activation in vivo or in vitro.
Expression of each of the three vertebrate runt family members, including
Runx3, is controlled by two promoters that generate proteins with alter-
native N-terminal sequences. The longer N-terminal region of Runx3,
expressed from the distal promoter, is highly conserved among family
members and across species. We show that transcripts from the distal
Runx3 promoter are selectively expressed in mature CD8
+
T cells and are
upregulated upon activation. We show that the N-terminal region encoded
by these transcripts carries an independent transcriptional activation
domain. This domain can activate transcription in isolation, and contri-
butes to the increased transcriptional activity observed with this isoform as
compared to those expressed from the ancestral, proximal promoter.
Together, these data suggest an important role for the additional N-ter-

minal Runx3 activation domain in CD8
+
T-cell function.
Abbreviations
eYFP, enhanced yellow fluorescent protein; GBD, Gal4 DNA-binding domain; HDAC5, class II histone deacetylase; Ig-Ca, IgG a-chain
constant region; TGF, transforming growth factor; YFP, yellow fluorescent protein.
FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS 3429
However, the mechanism(s) by which Runx3 selectively
promotes CD8 single-positive development in the thy-
mus and the expansion of CD8
+
effector cells in the
periphery is not yet clear.
Runx3 is one of a large family of transcriptional reg-
ulators, closely related in both structure and function
to the Drosophila runt protein [5]. Found in all animals
from worms to humans, this gene family was originally
defined by the presence of the DNA-binding runt
domain, which is essentially invariant among all family
members throughout evolution and results in similar
target-binding motifs for all family members [6]. In
addition to DNA-binding affinity, nuclear import or
retention and protein half-life are also regulated
through this domain [7–11]. Vertebrates have three
runt family genes, Runx1–3 (also referred to as
CBFa1–3 or PEBP2aA–C [12]). Runx3 has been
shown to recruit other transcription factors to sites of
transcriptional regulation, integrating signaling path-
ways critical for hematopoietic cell development and
activation. Documented interactions include those with

transforming growth factor (TGF)-b-activated Smads
and several Ets family members, as well as repressor
complexes from the mSin3a and Groucho ⁄ TLE famil-
ies [13–17]. Recruitment of TLE repressor complexes
has been shown to be essential for silencing of Cd4 by
Runx3 in CD8
+
T cells [13].
In vertebrates, all runt family genes contain two
highly conserved alternative promoters, proximal (P)
and distal (D). The 5¢-coding sequence of invertebrate
runt family members is homologous to the sequence
controlled by the vertebrate proximal promoter, sug-
gesting that acquisition of the distal promoter was
associated with the evolution of more complex require-
ments for control of this essential developmental regu-
lator early in vertebrate evolution [6]. To date, the
activity of runt family members has been characterized
primarily using transcripts generated from the prox-
imal (ancestral) promoter, with attention focused on
C-terminal functional domains. Previous reports have
suggested that transcripts from the two alternative pro-
moters are differentially expressed [18,19]. However,
no studies specifically testing for functional differences
between Runx3 isoforms have been published, and the
function of the longer N-terminal domain in Runx3
has not been extensively analyzed.
Here we show that Runx3 transcripts from the distal
promoter are largely restricted to CD8
+

T cells in the
periphery, clearly distinguishing them from CD4
+
T cells. Additionally, splenocytes from C57BL ⁄ 6 (B6)
mice selectively upregulate the transcript from the
distal promoter after culture with activating antibody
to CD3. In contrast, differences in expression of total
Runx3 transcripts in CD4
+
and CD8
+
T cells are
small, suggesting that transcripts from the proximal
promoter are unlikely to explain the marked differ-
ences in dependence on Runx3 activity in the two sub-
sets. Analysis of the function of the longer Runx3
isoform, expressed from the distal promoter, demon-
strates that the N-terminal region contains an
independent transcriptional activation domain that
contributes to increased transcriptional activity by this
Runx3 isoform. We suggest that the N-terminal activa-
tion domain, encoded by transcripts from the Runx3
distal promoter, is required for normal function of per-
ipheral T cells, particularly those of the CD8
+
subset.
Results
Differential expression of Runx3 transcripts
The highly homologous mouse and human RUNX3
genes are transcribed from two alternative promoters

(human homolog shown in Fig. 1A). Transcripts
encode distinct five amino acid or 19 amino acid
N-terminal sequences from the proximal (P) and distal
(D) promoters, respectively (Fig. 1B). A high level of
sequence conservation is present within the N-termini
of proteins generated from each of the two promoters
across vertebrate species, suggesting a critical func-
tional distinction between the two isoforms (Fig. 1C).
Data from two independent targeted mutant mouse
strains show that normal Runx3 activity is absolutely
required for the normal development and clonal
expansion of CD8
+
T cells but has no detectable
impact on CD4
+
T-cell development or proliferation.
Although no disruption of CD4
+
T-cell development
or activation was noted in mice carrying deletion
mutations of Runx3, a recent report [20] showed that
Runx3 is upregulated by the transcription factor T-bet
during the functional maturation of T
H
1 cells, and
subsequently cooperates with T-bet in the repression of
interleukin-4 expression and, to a lesser degree, in
the upregulation of interferon-c characteristic of this
subset.

Using primers binding to a sequence in the runt
domain conserved among all runt family members
(exon 4) and a Runx3-specific sequence in the C-termi-
nus (exon 6), we compared expression levels for all
full-length Runx3 transcripts in positively selected
CD4
+
and CD8
+
splenocytes. Both quantitative den-
sitometry (data not shown) and real-time quantitative
PCR (Fig. 2A) revealed only a modest overall differ-
ence in expression of Runx3 between the CD4
+
and
CD8
+
subsets (< 1.6-fold), which seemed unlikely to
explain a selective effect of Runx3 deficiency on CD8
+
Transcriptional activation by Runx3 D. D. Chung et al.
3430 FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS
survival and activation. We were therefore interested
to test whether there were differences in expression
from the two promoters between the CD4
+
and
CD8
+
T-cell subsets.

No primers suitable for real-time quantitative PCR
specific for transcripts from either the proximal or
distal promoter were identified, probably because of
the extremely high G ⁄ C content of these exons. How-
ever, we were able to amplify a 433 bp fragment from
the 5¢-UTR of the distal transcript that could be used
for quantitation through densitometry of PCR prod-
ucts on serial dilutions of template. Using this method,
we found that distal transcripts are more highly
expressed in the CD8
+
subset, suggesting that differen-
tial effects of Runx3 deletion on CD4
+
and CD8
+
T cells may result selectively from loss of the Runx3
isoform generated from this promoter (Fig. 2A).
With culture conditions that produced optimal B6
splenocyte proliferation at 72 h in response to plate-
bound anti-CD3, CD8
+
T cells reproducibly expanded
more efficiently than CD4
+
T cells (Fig. 2B). Forty
hours after initiation of the cultures, when viable
CD4
+
and CD8

+
T-cell numbers had not changed sig-
nificantly from the input numbers (data not shown),
we detected only small increases in total Runx3 tran-
scripts, which seemed unlikely to explain the require-
ment for Runx3 in the clonal expansion or functional
maturation of CD8
+
T cells (Fig. 2C). However,
quantification of distal transcripts alone again showed
a dramatic increase in expression levels after anti-CD3
stimulation. Thus, it appears that Runx3 distal tran-
scripts are more highly expressed in CD8
+
T cells
and are specifically further upregulated on T-cell
stimulation.
Increased transcriptional activity from protein
expressed from the distal promoter
As little was known about functional differences
between the two isoforms of Runx3, we compared
A
B
C
Fig. 1. Structure of the RUNX3 gene. (A) The RUNX3 gene consists of seven exons. Expression of alternative transcripts is regulated by
TATA-less promoters upstream of exons 1 and 3 (arrows; P, proximal promoter; D, distal promoter). The two transcripts are produced from
translational start sites in exon 3 (transcript 1) and exon 2 (transcript 2), resulting in protein-coding sequences (marked in gray) differing only
at the N-terminus. (B) The DNA and corresponding amino acid sequences of the N-termini of the two alternative transcripts are shown. The
splice site between exons 2 and 3 in transcript 2 is underlined. (C) An alignment of vertebrate P2-RUNX3 N-terminal sequences is shown.
Amino acids that are identical or similar in at least two sequences are shaded black and gray, respectively. The start of the runt domain is

indicated. The region unique to P2-RUNX3, and the C-terminal ends of the two GBD fusions (encoding amino acids 1–31 or 1–66 of
P2-RUNX3) are shown.
D. D. Chung et al. Transcriptional activation by Runx3
FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS 3431
their ability to activate transcription of known runt
family target promoters. Using clones for both tran-
scripts from the highly homologous human RUNX3
gene (Fig. 1C [19]), we first tested the ability of full-
length P-RUNX3 and D-RUNX3 to activate the
p6OSE2 reporter construct. This synthetic promoter,
consisting of a tandem array of six runt domain-bind-
ing sites derived from the osteocalcin gene promoter, is
a sensitive reporter for all runt family members. The
activity of p6OSE2 in the presence of Flag–D-Runx3
was consistently two-fold higher than with Flag–P-
Runx3 (Fig. 3A). Higher transcriptional activity of
D-RUNX3 was also observed using both Myc epitope-
tagged and untagged expression constructs (Fig. 3B,C),
ruling out possible differential effects of the N-terminal
epitope tags. Both isoforms were expressed at similar
levels [a P-RUNX3 ⁄ D-RUNX3 ratio of 1.09 : 1 (Myc)
and 1.12 : 1 (Flag)], suggesting that this difference in
transcriptional activity was not due to differences in
expression levels (Fig. 3F).
Although DNA binding by all runt family members
is similar in vitro, the osteocalcin promoter is a natural
target for the runt family member Runx2 in vivo.To
analyze transcriptional activation by P-RUNX3 and
D-RUNX3 using a known RUNX3 target, we used a
reporter driving luciferase expression by the IgG

a-chain constant region (Ig-Ca) promoter. This pro-
moter is regulated by Runx3 together with TGF-b-
activated Smad complexes in vivo [21,22]. Although
both RUNX3 isoforms enhanced the response to
increasing levels of added TGF-b, D-RUNX3
increased activation of the Ig-Ca-luciferase reporter
more effectively than P-RUNX3 (Fig. 3D), confirming
a general increase in the activity of D-RUNX3 as com-
pared to P-RUNX3. With this reporter, the difference
between P-RUNX3 and D-RUNX3 was less dramatic
than with p6OSE2, possibly because p6OSE2 is
entirely dependent on RUNX3 for activation.
Interactions between p300 ⁄ CBP and a C-terminal
domain in RUNX3 have been shown to regulate
nuclear localization and transcriptional activity of
RUNX3 via acetylation of residues in and flanking the
runt domain [10]. Overexpression of the adenovirus
E1a protein blocks numerous transcriptional responses
through its dose-dependent inhibition of p300 ⁄ CBP
activity [22]. We used this titratable inhibition to deter-
mine whether the two RUNX3 isoforms differed in
A
B
C
Fig. 2. Selective expression of distal transcripts of Runx3 in
mature, peripheral CD8
+
T cells and activated splenocytes. Real-
time quantitative PCR (all transcripts) or densitometry of PCR prod-
ucts (distal transcripts) was performed on RNA isolated from the

relevant tissues in order to determine the relative representation of
transcripts from the two promoters. (A) RNA from bead-purified
splenic CD4
+
and CD8
+
T cells was compared in two separate
experiments on RNA from two mice (Expt 1) or RNA from a single
mouse (Expt 2). A representative gel used for densitometry is
shown below. For each test sample: lane 1 shows a control tem-
plate with no reverse transcriptase; lanes 2–4 show the results of
RT-PCR on decreasing concentrations of template. (B) B6 spleno-
cytes from three individual 4-month-old female mice were cultured
for 72 h, with or without anti-CD3. Cells were counted and ana-
lyzed by flow cytometry. Absolute numbers were calculated from
viable total cell counts using the frequency of cells in each subset
after exclusion of dead cells via 7-amino-actinomycin D staining (SD
in parentheses). (C) Runx3 expression was quantitated in RNA
pooled from the splenocytes of two individual mice in each of
two experiments after 40 h of culture, with or without anti-CD3.
Expression was normalized by levels of 18S RNA to ensure
comparisons based on equal amounts of input total RNA and is
shown relative to that in CD4
+
T cells (A) or unstimulated, cultured
splenocytes (C).
Transcriptional activation by Runx3 D. D. Chung et al.
3432 FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS
their dependence on p300 by coexpressing P-RUNX3
or D-RUNX3 with increasing amounts of E1a. As

shown in Fig. 3E, transcriptional activation by both
P-RUNX3 and D-RUNX3 was inhibited equally well
by coexpressed E1a, suggesting that both depend on
p300 ⁄ CBP for their ability to activate transcription.
To rule out a general alteration in RUNX3 activity
caused by the presence of alternative N-terminal
regions, we tested a selection of known RUNX3 func-
tions. As shown in Fig. 4A, when expressed as fusions
to enhanced yellow fluorescent protein (eYFP), both
RUNX3 isoforms were localized predominantly to the
AB C
D
E
F
Fig. 3. Transcriptional activation by P-RUNX3 and D-RUNX3. HepG2 cells were transfected with the p6OSE2 luciferase reporter, together
with Flag epitope-tagged P-RUNX3 and D-RUNX3 (A), Myc epitope-tagged P-RUNX3 and D-RUNX3 (B), or untagged expression vectors enco-
ding P-RUNX3 and D-RUNX3 (C). (D) HepG2 cells were transfected with the Ig-Ca luciferase reporter, and either P-RUNX3 or D-RUNX3, as
indicated; 18 h before analysis, cells were treated with the indicated concentration of TGF-b. (E) HepG2 cells were transfected with the
p6OSE2 luciferase reporter and untagged P-RUNX3 and D-RUNX3 with increasing amounts of an E1a expression plasmid. Luciferase activity
in all panels is shown in arbitrary units, as the mean ± SD of duplicate transfections. (F) Relative expression of transfected Myc- or Flag-
tagged P-RUNX3 and D-RUNX3 was assayed by western blotting with an antibody against the Myc or Flag epitopes.
D. D. Chung et al. Transcriptional activation by Runx3
FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS 3433
nucleus, with no apparent differences in subnuclear
localization. Next, we analyzed the interaction of
RUNX3 with two known partners by coimmunoprecip-
itation from transfected cells. COS1 cells were trans-
fected with expression plasmids encoding Smad3 and
Flag-tagged RUNX3, and immunocomplexes were iso-
lated on anti-Flag agarose. Despite the difference in the

activation of the Ig-Ca reporter, we found no significant
differences in the ability of the two isoforms to interact
with the TGF-b-responsive Smad3, which interacts with
both p300 and RUNX3 to initiate class switching to
IgA production in B lymphocytes (Fig. 4B). Activation
and stabilization of Runx3 by p300 ⁄ CBP can be
reversed through Runx-dependent recruitment of the
class II histone deacetylase, HDAC5, to transcriptional
complexes containing runt family proteins [10]. There-
fore, we tested for interaction of HDAC5 with Runx3
using COS1 cells transfected with Myc-tagged
P-RUNX3 or D-RUNX3, with or without Flag-tagged
HDAC5. Flag immunocomplexes were analyzed for
coprecipitating RUNX3. As with the Smad3 interac-
tion, we observed no significant difference in the ability
of the RUNX3 isoforms to interact with HDAC5, sug-
gesting that the potential for destabilization of RUNX3
binding through deacetylation, and the potential for
transcriptional repression through HDAC5-mediated
histone deacetylation, is similar for both isoforms
(Fig. 4C). Taken together, these results imply that the
enhanced transcriptional activation by D-RUNX3 does
not simply result from altered interactions with known
binding partners or protein localization, and suggest the
possibility that the longer N-terminal region present
in D-RUNX3 contributes directly to transcriptional
activation.
A novel activation domain in D-RUNX3
To analyze the transcriptional activation potential of
the RUNX3 isoforms, independent of the runt domain

binding to DNA, we created a series of RUNX3
fusions to the Gal4 DNA-binding domain (GBD), and
tested them using a heterologous promoter. Full-length
P-RUNX3 and D-RUNX3 constructs, fused to the
GBD, were cotransfected into HepG2 cells together
with the (Gal)
5
-TATA-luc reporter, which contains five
Gal4 operators upstream of a minimal TATA element.
As seen with the Runx binding site reporters,
D-RUNX3 was significantly more active than
P-RUNX3, particularly at low levels of the transfected
GBD fusion construct (Fig. 5A). Thus, at the lowest
level of transfected plasmid, GBD–D-RUNX3
increased activation more than two-fold, whereas
GBD–P-RUNX3 increased activity by only 30%. This
suggests that transcripts from the distal promoter
encode a transcriptional activation function independ-
ent of DNA binding by the runt domain, or recruit-
ment of additional transcriptional activators by the
rest of the protein. In confirmation of this, GBD
fusion constructs that excluded the runt domain and
everything C-terminal to it showed similar differences
in reporter activation (Fig. 5B). The N-terminal region
of D-RUNX3 encompassing amino acids 1–66 activa-
ted expression of the (Gal)
5
-TATA-luc reporter up to
eight-fold, whereas the comparable P construct (amino
acids 1–52; Figs 1C and 5D) performed little better

than the GBD alone. Further analysis of the
D-RUNX3 N-terminal domain demonstrated that
either the N-terminal 19 or 31 amino acids were suffi-
cient to activate transcription when targeted via the
ABC
Fig. 4. RUNX3 interactions with Smad3 and HDAC5. (A) RUNX3 proteins are nuclear. COS1 cells were transfected with YFP fused P-RUNX3
and D-RUNX3 expression constructs. Hoechst and YFP images are shown. (B) COS1 cells were transfected with SMAD3 together with Flag
tagged P-RUNX3 or D-RUNX3 expression constructs. Protein complexes were collected on anti-Flag agarose, and coprecipitating SMAD3
was visualized by western blot. (C) COS1 cells were cotransfected with Flag-tagged HDAC5 Myc-tagged P-RUNX3 or D-RUNX3, as indica-
ted, and the presence of RUNX3 proteins in Flag immunocomplexes detected by Myc western blot.
Transcriptional activation by Runx3 D. D. Chung et al.
3434 FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS
GBD (Fig. 5B). These constructs contain only the
most highly conserved region of the D-RUNX3 N-ter-
minus (Figs 1C and 5D). Importantly, all of the trun-
cated GBD fusion constructs were expressed at very
similar levels (Fig. 5D). Thus, the extreme N-terminus,
contained within amino acids 1–19 of D-RUNX3,
constitutes an independent transcriptional activation
domain that contributes to overall transcriptional acti-
vation by this RUNX3 isoform.
Discussion
We have shown that transcripts from the distal Runx3
promoter are clearly more highly expressed in the
CD8
+
T-cell subset than in the CD4
+
subset in mice.
In contrast, little difference in total Runx3 transcripts

is observed between these two cell types. As Runx3
has been shown to be absolutely required for the nor-
mal phenotype and function of CD8
+
T cells, our
results suggest that this subset may specifically require
the longer D-Runx3 isoform. Transcripts from the
distal promoter are also selectively upregulated in
splenocytes stimulated with anti-CD3, further support-
ing a critical role for this Runx3 isoform in T-cell
activation.
Functional comparison of the proteins encoded by
each transcript demonstrates that both P and D iso-
forms of highly homologous human RUNX3 isoforms
A
CD
B
Fig. 5. The N-terminus of D-RUNX3 contains an activation domain. HepG2 cells were transfected with the (Gal)
5
-TATA-luc reporter, together
with the indicated fusions between the GBD and P-RUNX3 and D-RUNX3. Luciferase activity was assayed and is presented as in Fig. 3.
(A) Increasing amounts of GBD and GBD–P-RUNX3 or GBD–D-RUNX3 fusions, encoding full length P-RUNX3 or D-RUNX3, were transfected.
(B) GBD alone or fusions to the N-terminal 52 amino acids of P-RUNX3 or the N-terminal 66 amino acids of D-RUNX3, and two fusions to
the N-terminal 31 or 19 amino acids of D-RUNX3, were assayed as in (A). (C) The relative expression of transfected GBD, GBD–P(1–52),
GBD–D(1–66), GBD–D(1–31) and GBD–D(1–19) fusions was assayed by western blot with an antibody against GBD. (D) The GBD fusions
used in this figure are shown schematically.
D. D. Chung et al. Transcriptional activation by Runx3
FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS 3435
interact similarly with Smad3 and HDAC5, and that
transcriptional activation by both is dependent on the

CBP ⁄ p300 complex. However, we show that the D iso-
form, generated from transcripts of the distal RUNX3
promoter, is consistently more active in transcriptional
reporter assays. This is true whether it is targeted to
DNA via its own runt domain or by fusion to the
heterologous Gal4 DNA-binding domain, suggesting
that differences in transcriptional activity are not due
to differences in DNA binding between the two iso-
forms. Deletion analyses revealed that D-RUNX3
contains a previously uncharacterized transcriptional
activation domain within the first 19 amino acids. This
domain can act in isolation, when targeted to a pro-
moter via a heterologous DNA-binding domain,
suggesting that it is a functionally independent tran-
scriptional activation domain. This domain clearly
functions to enhance the activity of the more
C-terminal activation domain in Runx3. However, it is
also possible that it may have distinct functions at
specific target promoters.
The N-terminal sequence encoded by distal Runx3
transcripts is highly conserved in runt family members
from vertebrate species, suggesting that the alternative
N-termini of vertebrate Runx proteins control critical
functions. In support of this idea, the D isoform of the
family member Runx1 has been shown to bind with
higher affinity than the P isoform to runt domain-
binding elements from both myeloperoxidase and
T-cell receptor b-chain promoters [23]. Moreover, the
two isoforms were shown to have differential effects
on precursor expansion and myeloid differentiation

during induced maturation of the promyelocytic
32Dcl.3 cell line. Granulocyte colony-stimulating
factor-mediated maturation occured rapidly in 32Dcl.3
cells transfected with control vector or D-Runx1,
whereas P-Runx1 promoted ongoing expansion prior
to terminal differentiation and thus produced a seven-
fold increase in the final number of mature granulo-
cytes. Interpretation of these results is complicated by
subsequent studies that showed an inverse relationship
between DNA-binding affinity and transcriptional acti-
vation for Runx1 [24]. However, the clear functional
difference between isoforms of Runx1 supports the
hypothesis that selective expression of D-Runx3 in
CD8
+
T cells plays a critical role in the normal func-
tion of this subset.
Interestingly, deletion analyses of Runx2 demonstra-
ted that the 19 amino acid region in the longer isoform
of this runt family member was required for full tran-
scriptional activation in a reporter assay. In contrast
to our results, however, the isolated N-terminus of the
longer Runx2, or the entire region N-terminal to the
runt domain, was only able to very weakly activate
transcription of a heterologous DNA-binding element
[25]. Differential activation of Runx2 target genes by
alternative Runx2 isoforms was also confirmed in two
later studies, although the differences were promoter
specific [26,27]. Additionally, functional differences
were seen in transgenic rescue models that tested the

recovery of bone formation by each isoform in Runx2
null mice [28]. More recently, the N-terminal region of
the shorter Runx1 isoform has also been shown to
contribute to transcriptional activation [24]. Thus it
appears that for all three vertebrate Runx proteins, the
N-termini play a role in their ability to activate
transcription.
Taken together, the results of analysis of the func-
tional differences between long and short isoforms of
the vertebrate runt family members clearly suggest dif-
ferences in overall transcriptional activity. In addition, it
is likely that some specific target genes are more
sensitive to these differences. We hypothesize that
increased expression of D-Runx3 in CD8
+
T cells may
result simply in an overall increase in the expression of
Runx3 target genes, with an increased sensitivity to
signals such as TGF-b. However, it is also possible that
the control of CD8
+
T-cell development and activation
may require the activation of a unique set of target genes
controlled by the activation domain in D-Runx3. In
either case, specific identification of D-Runx3-dependent
transcriptional targets is likely to provide a productive
strategy for uncovering the function of this isoform,
which is specifically required for normal CD8
+
T-cell

development and activation.
Experimental procedures
Mice
C57BL ⁄ 6 mice were bred and maintained in the vivarium
at the University of Virginia, using founders obtained
from The Jackson Laboratory (Bar Harbor, ME). All
protocols using mice were reviewed and approved by the
Animal Care and Use Committee of the University of
Virginia.
Splenocyte culture and flow cytometry
Splenocytes were processed for flow cytometry after culture
for up to 72 h in modified Dulbecco’s medium containing
10% heat-inactivated fetal bovine serum, with or without
with plate-bound purified anti-CD3 (145-2C11). Samples
stained with allophycocyanin-labeled GK1.5 (CD4) and
fluorescein isothiocyanate-labeled 53-6.72 (CD8) (BD Bio-
sciences ⁄ Pharmingen, San Jose, CA) were analyzed using a
Transcriptional activation by Runx3 D. D. Chung et al.
3436 FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS
FACScan cytometer equipped with cellquest software
(BD Biosciences ⁄ Pharmingen).
Quantitation of RUNX3 transcripts
RNA was isolated from cells after 40 h in culture using
a QIAshredder Mini Spin Column and RNeasy Mini Kit
(Qiagen, Valencia, CA). cDNA was synthesized using the
First-Strand cDNA Synthesis Kit (GE Healthcare Life
Sciences, Piscataway, NJ). Real-time quantitative PCR for
total Runx3 transcript levels was performed in triplicate
using the QuantiTect SYBR Green RT-PCR Kit (Qiagen) in
an iCycler (Bio-Rad, Hercules, CA), with the following prim-

ers: CAGGTTCAACGACCTTCGAT (exon 4) and AGGC
CTTGGTCTGGTCTTCT (exon 6). Quantitative determin-
ation of mRNA levels from the distal promoter was per-
formed using the primers GGTGAGCCTCGTTCATTCAT
(sense) and GGTCAGACCCACTTGGTTGG (antisense) to
generate a single 433 bp product from the 5¢-UTR of the
distal promoter, followed by electrophoresis through 1%
agarose and visualization using ethidium bromide. Densito-
metry of PCR products was performed using genesnap
software (SynGene, Frederick, MD). For both methods,
Runx3 transcript levels were standardized to expression of
ribosomal 18S RNA using TaqMan Ribosomal RNA
Control Reagents (Applied Biosystems, Foster City, CA).
Plasmids
An expressed sequence tag clone (IMAGE: 3615873) for
human D-Runx3 was obtained from the American Type
Culture Collection. 6Myc-D-Runx3 was generated by PCR
from 6Myc-P-Runx3 (gift of Y. Ito, Kyoto University,
Japan). Untagged human SMAD3 and P-Runx3 and D-
Runx3 were expressed from pCMV5. Flag-tagged P-Runx3
and D-Runx3 constructs were generated in a modified
pCMV5. GBD fusions were created within pM (Clontech,
Mountain View, CA). Yellow fluorescent protein (YFP)
fusions were created within a modified pCS2 vector,
containing an N-terminal enhanced eYFP tag (BD
Biosciences ⁄ Pharmingen). p6OSE2-luc was a kind gift
from R. Derynck (UCSF, CA). pBJ5-Flag-HDAC5 was a
gift from S. Schreiber (Harvard, MA).
Luciferase assays
HepG2 cells were transfected with firefly luciferase report-

ers, a phCMVRLuc control (Promega, Madison, WI) and
Runx3 expression constructs using Exgen 500 (MBI Fer-
mentas, Hanover, MD). HepG2 cells were chosen specific-
ally because they express negligible levels of runt family
members. After 48 h, promoter activity was assayed with a
luciferase assay kit (Promega), using a Berthold (Oak
Ridge, TN) LB953 luminometer. Results were standardized
using renilla luciferase activity, assayed with 0.09 lm colen-
terazine (Biosynth, Naperville, IL).
Immunoprecipitation and western blotting
COS1 cells were maintained in DMEM with 10% bovine
growth serum (Hyclone, Logan, UT) and were transfected
using LipofectAmine (Invitrogen, Carlsbad, CA). Thirty-six
hours after transfection, cells were lysed by sonication in
75 mm NaCl, 50 mm Hepes (pH 7.8), 20% glycerol, 0.1%
Tween020, 0.5% Nonidet-P40 with protease and phospha-
tase inhibitors. Immunocomplexes were precipitated with
Flag M2–agarose (Sigma, St Louis, MO). Following
SDS ⁄ PAGE, proteins were electroblotted to Immobilon-P
(Millipore, Billerica, MA) and incubated with antisera spe-
cific for Flag (Sigma), Smad2 ⁄ 3 (Chemicon, Temecula,
CA), and Myc (9E10; University of Virginia, Lymphocyte
Culture Center). RUNX3 levels were quantified using Alexa
Fluor 680 anti-(mouse IgG) (1 : 2000) as secondary anti-
bodies. Membranes were scanned and analyzed using odys-
sey software (LI-COR).
Fluorescence microscopy
COS1 cells were split onto four-well chamber slides (Nunc,
Rochester, NY) and transfected with eYFP-tagged fusion
proteins using Fugene 6 (Roche, Indianapolis, IN). After

22–26 h, cells were stained with Hoechst 33342 and imaged
with a Zeiss (Thornwood, NY) Axiovert 135T inverted
fluorescence microscope on a heated stage with YFP and
4’,6-diamidino-2-phenylindole filter sets (Omega Opticals,
Brattleboro, VT). Images were visualized, captured, and
converted to 8-bit.tif files with a Zeiss 32·⁄0.40 objective
and a Hamamatsu (Bridgewater, NJ) Orca II cooled
charge-coupled device camera controlled by openlab soft-
ware (Improvision, Lexington, MA).
Acknowledgements
This work was supported by the University of Virginia
SCOR in Systemic Lupus Erythematosus (P50
AR45222). The authors would like to thank R. Dery-
nck, Y. Ito and S. Schreiber for generously providing
plasmids.
References
1 Li QL, Ito K, Sakakura C, Fukamachi H, Inoue K, Chi
XZ, Lee KY, Nomura S, Lee CW, Han SB et al. (2002)
Causal relationship between the loss of RUNX3 expres-
sion and gastric cancer. Cell 109, 113–124.
2 Levanon D, Bettoun D, Harris-Cerruti C, Woolf E,
Negreanu V, Eilam R, Bernstein Y, Goldenberg D,
D. D. Chung et al. Transcriptional activation by Runx3
FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS 3437
Xiao C, Fliegauf M et al. (2002) The Runx3 transcrip-
tion factor regulates development and survival of
TrkC dorsal root ganglia neurons. EMBO J 21,
3454–3463.
3 Taniuchi I, Osato M, Egawa T, Sunshine MJ, Bae SC,
Komori T, Ito Y & Littman DR (2002) Differential

requirements for Runx proteins in CD4 repression and
epigenetic silencing during T lymphocyte development.
Cell 111, 621–633.
4 Woolf E, Xiao C, Fainaru O, Lotem J, Rosen D,
Negreanu V, Bernstein Y, Goldenberg D, Brenner O,
Berke G et al. (2003) Runx3 and Runx1 are required
for CD8 T cell development during thymopoiesis. Proc
Natl Acad Sci USA 100 , 7731–7736.
5 Kagoshima H, Shigesada K, Satake M, Ito Y,
Miyoshi H, Ohki M, Pepling M & Gergen P (1993)
The Runt domain identifies a new family of hetero-
meric transcriptional regulators. Trends Genet 9,
338–341.
6 Rennert J, Coffman JA, Mushegian AR & Robertson
AJ (2003) The evolution of Runx genes I. A comparat-
ive study of sequences from phylogenetically diverse
model organisms. BMC Evol Biol 3,4.
7 Tang YY, Shi J, Zhang L, Davis A, Bravo J,
Warren AJ, Speck NA & Bushweller JH (2000)
Energetic and functional contribution of residues in
the core binding factor beta (CBFbeta) subunit to
heterodimerization with CBFalpha. J Biol Chem 275,
39579–39588.
8 Zhang L, Li Z, Yan J, Pradhan P, Corpora T, Cheney
MD, Bravo J, Warren AJ, Bushweller JH & Speck NA
(2003) Mutagenesis of the Runt domain defines two
energetic hot spots for heterodimerization with the core
binding factor beta subunit. J Biol Chem 278, 33097–
33104.
9 Huang G, Shigesada K, Ito K, Wee HJ, Yokomizo T &

Ito Y (2001) Dimerization with PEBP2beta protects
RUNX1 ⁄ AML1 from ubiquitin-proteasome-mediated
degradation. EMBO J 20, 723–733.
10 Jin YH, Jeon EJ, Li QL, Lee YH, Choi JK, Kim WJ,
Lee KY & Bae SC (2004) Transforming growth factor-
beta stimulates p300-dependent RUNX3 acetylation,
which inhibits ubiquitination-mediated degradation.
J Biol Chem 279, 29409–29417.
11 Chakraborty S, Sinha KK, Senyuk V & Nucifora G
(2003) SUV39H1 interacts with AML1 and abrogates
AML1 transactivity. AML1 is methylated in vivo.
Oncogene 22, 5229–5237.
12 van Wijnen AJ, Stein GS, Gergen JP, Groner Y, Hie-
bert SW, Ito Y, Liu P, Neil JC, Ohki M & Speck N
(2004) Nomenclature for Runt-related (RUNX) pro-
teins. Oncogene 23, 4209–4210.
13 Yarmus M, Woolf E, Bernstein Y, Fainaru O, Negrean-
u V, Levanon D & Groner Y (2006) Groucho ⁄ transduc-
in-like Enhancer-of-split (TLE)-dependent and
-independent transcriptional regulation by Runx3. Proc
Natl Acad Sci USA 103, 7384–7389.
14 Erman B, Cortes M, Nikolajczyk BS, Speck NA & Sen
R (1998) ETS-core binding factor: a common composite
motif in antigen receptor gene enhancers. Mol Cell Biol
18, 1322–1330.
15 Miyazono K, Maeda S & Imamura T (2004) Coordinate
regulation of cell growth and differentiation by TGF-
beta superfamily and Runx proteins. Oncogene 23,
4232–4237.
16 Lutterbach B, Westendorf JJ, Linggi B, Isaac S, Seto E

& Hiebert SW (2000) A mechanism of repression by
acute myeloid leukemia-1, the target of multiple chro-
mosomal translocations in acute leukemia. J Biol Chem
275, 651–656.
17 Javed A, Guo B, Hiebert S, Choi JY, Green J, Zhao
SC, Osborne MA, Stifani S, Stein JL, Lian JB et al.
(2000) Groucho ⁄
TLE ⁄ R-esp proteins associate
with the nuclear matrix and repress RUNX
(CBF(alpha) ⁄ AML ⁄ PEBP2(alpha) dependent activation
of tissue-specific gene transcription. J Cell Sci 113(12),
2221–2231.
18 Rini D & Calabi F (2001) Identification and comparat-
ive analysis of a second runx3 promoter. Gene 273,
13–22.
19 Bangsow C, Rubins N, Glusman G, Bernstein Y,
Negreanu V, Goldenberg D, Lotem J, Ben-Asher E,
Lancet D, Levanon D et al. (2001) The RUNX3 gene )
sequence, structure and regulated expression. Gene 279,
221–232.
20 Djuretic IM, Levanon D, Negreanu V, Groner Y, Rao
A & Ansel KM (2007) Transcription factors T-bet and
Runx3 cooperate to activate Ifng and silence Il4 in T
helper type 1 cells. Nat Immunol 8, 145–153.
21 Hanai J, Chen LF, Kanno T, Ohtani-Fujita N, Kim
WY, Guo WH, Imamura T, Ishidou Y, Fukuchi M, Shi
MJ et al. (1999) Interaction and functional cooperation
of PEBP2 ⁄ CBF with Smads. Synergistic induction of
the immunoglobulin germline Calpha promoter. J Biol
Chem 274, 31577–31582.

22 Park SR, Seo GY, Choi AJ, Stavnezer J & Kim PH
(2005) Analysis of transforming growth factor-beta1-
induced Ig germ-line gamma2b transcription and its
implication for IgA isotype switching. Eur J Immunol
35, 946–956.
23 Telfer JC & Rothenberg EV (2001) Expression and
function of a stem cell promoter for the murine
CBFalpha2 gene: distinct roles and regulation in
natural killer and T cell development. Dev Biol 229,
363–382.
24 Liu H, Carlsson L & Grundstrom T (2006) Identi-
fication of an N-terminal transactivation domain of
Runx1 that separates molecular function from global
differentiation function. J Biol Chem 281, 25659–
25669.
Transcriptional activation by Runx3 D. D. Chung et al.
3438 FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS
25 Thirunavukkarasu K, Mahajan M, McLarren KW,
Stifani S & Karsenty G (1998) Two domains unique to
osteoblast-specific transcription factor Osf2 ⁄ Cbfa1
contribute to its transactivation function and its
inability to heterodimerize with Cbfbeta. Mol Cell Biol
18, 4197–4208.
26 Xiao ZS, Hinson TK & Quarles LD (1999) Cbfa1
isoform overexpression upregulates osteocalcin gene
expression in non-osteoblastic and pre-osteoblastic cells.
J Cell Biochem 74, 596–605.
27 Harada H, Tagashira S, Fujiwara M, Ogawa S,
Katsumata T, Yamaguchi A, Komori T & Nakatsuka
M (1999) Cbfa1 isoforms exert functional differences

in osteoblast differentiation. J Biol Chem 274,
6972–6978.
28 Kanatani N, Fujita T, Fukuyama R, Liu W, Yoshida
CA, Moriishi T, Yamana K, Miyazaki T, Toyosawa S
& Komori T (2006) Cbf beta regulates Runx2 function
isoform-dependently in postnatal bone development.
Dev Biol 296, 48–61.
D. D. Chung et al. Transcriptional activation by Runx3
FEBS Journal 274 (2007) 3429–3439 ª 2007 The Authors Journal compilation ª 2007 FEBS 3439