Characterization of promoter 3 of the human thromboxane
A
2
receptor gene
A functional AP-1 and octamer motif are required for basal
promoter activity
Adrian T. Coyle and B. Therese Kinsella
Department of Biochemistry, Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin, Ireland
The prostanoid thromboxane (TX)A
2
induces activa-
tion and aggregation of platelets, constriction of vascu-
lar (V) and bronchial smooth muscle (SM) and of
renal mesangial cells [1–4], and may induce other
diverse cellular responses including mitogenic and ⁄ or
hypertrophic growth of VSM [5,6], inhibition of angio-
genesis ⁄ neo-vascularization [7] and apoptosis of
CD4 ⁄ CD8
++⁄ –
thymocytes [8]. Alterations in the level
of this potent autocoid, or of its specific synthase or
its receptor (TP) are widely implicated in a variety of
vascular diseases including thrombosis, unstable angina,
asthma, systemic and pregnancy-induced hyperten-
sion, and glomerulonephritis [9–12]. Moreover, mice
deficient in the TXA
2
receptor (TP
– ⁄ –
) display
increased bleeding and altered hemodynamic proper-

core Prm3 activity in both cell types. Furthermore, three distinct regulatory
regions comprising of an upstream repressor sequence, located between
)404 to )320, and two positive regulatory regions required for efficient
basal gene expression, located between )154 to )106 and )50 to +1, were
identified within the core Prm3. Deletion and site-directed mutagenesis of
consensus Oct-1 ⁄ 2 and AP-1 elements within the latter )154 to )106 and
)50 to +1 regions, respectively, substantially reduced Prm3 activity while
mutation of both elements abolished Prm3 activity. Electromobility shift
and supershift assays confirmed the specificity of nuclear factor binding to
the latter Oct-1 ⁄ 2 and AP-1 elements. Moreover, herein it was established
that the core AP-1 element mediates phorbol myristic acid-induction of
Prm3 activity hence providing a mechanistic explanation of phorbol ester
up-regulation of TPb mRNA expression.
Abbreviations
AP-1, activator protein-1; EMSA, electromobility shift assay; FBS, fetal bovine serum; HEK, human embryonic kidney; HEL, human
erythroleukemia; I, intron; NT, nucleotide; PMA, phorbol myristic acid; Prm, promoter; RLU, relative luciferase units; TP, thromboxane
receptor; TI, transcription initiation; TXA2, thromboxane A2; UAS, upstream activation sequence; URS, upstream repressor sequence;
UTR, untranslated region.
1036 FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS
receptor in the dynamic regulation of haemostasis
[13,14].
As a member of the G protein coupled receptor
(GPCR) superfamily, the TXA
2
receptor or TP is pri-
marily coupled to Gq-dependent activation of phos-
pholipase (PLC) Cb isoforms [1,3]. In humans, but not
in nonprimates, TXA
2
signals through two TP isoforms

of TPa within its unique C-tail domain [22,23].
Hence, whilst the biological significance for the
existence of two TP receptors in humans is indeed
unclear, there is mounting evidence that they undergo
differential signaling and regulation, strengthening the
viewpoint that TPa and TPb may have distinct physio-
logic ⁄ pathophysiologic roles. Consistent with this, TPa
and TPb are also subject to differential expression and
gene regulation [24,25]. Whilst TPa and TPb mRNAs
are coexpressed in a range of cell ⁄ tissue types of rele-
vance to TXA
2
biology, there are extensive differences
in the relative levels of expression of TPa:TPb mRNA
in several tissues [24]. Moreover, recent studies have
confirmed that TPa and TPb expression are under the
genetic control of distinct promoters within the single
human TP gene located on chromosome 19 [16,25].
Whilst the originally identified promoter (Prm) 1
directs TPa expression, a novel promoter (Prm3) was
identified within the human TP gene that exclusively
directs TPb expression [25]. Similar to that of the pre-
viously characterized Prm1 and Prm2, Prm3 lacks a
consensus TATA box or initiator element and, hence,
the transcription factor elements directing basal Prm3-
activity remain to be identified [25].
The aim of the current study was to define the core
promoter and to identify the cis -acting elements regu-
lating basal Prm3 activity with the view to determining
the key factors that regulate TPb expression in human

2
receptor (TP) gene that
directs expression of TPb in HEL92.1.7 and HEK293
cells [25]. A schematic of the human TP gene highlight-
ing the positions of the previously identified Prm1,
Prm2 [16,25] and the novel Prm3 [25] relative to its
translational start site (ATG, designated +1) is
presented in Fig. 1. In order to gain further insights
into the modes of regulation of TPb, the aim of the
current study was to map the minimal transcriptional
unit and to identify the key regulatory elements within
Prm3 directing basal gene expression.
The recombinant pGL3Basic encoding Prm3 direc-
ted 3.65 ± 0.23 RLU and 3.0 ± 0.25 RLU of luci-
ferase activity in HEL (Fig. 1A) and HEK293
(Fig. 1C) cells, respectively, whilst the empty pGL3
Basic vector directed minimal activity in either cell
type (Fig. 1A,C). Progressive 5¢ deletion of Prm3
sequences in pGL3Basic to generate )975 and )404
subfragments did not significantly affect luciferase
A. T. Coyle and B. T. Kinsella Thromboxane A2 receptor gene expression
FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS 1037
expression in either HEL (Fig. 1A) or HEK293
(Fig. 1C) cells. Similarly, whilst the overall levels of
Prm3-luciferase activity directed by the pGL3Enhancer
plasmids, containing an SV40 enhancer element down-
stream of the luciferase gene, were generally two- to
three-fold higher than by the equivalent pGL3Basic
plasmids in both HEL (7.57 ± 0.53 RLU; Fig. 1B)
and HEK (9.60 ± 0.41 RLU; Fig. 1D) cells, there was

+1 Luc
-975
pGL3Basic
+1
Luc
-50
+1 Luc
-106
+1 Luc
-154
+1 Luc
-320
+1 Luc
-404
+1 Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1 Prm3
+786
8
01234567
0246 12810
0 5 10 15 20 25
Luciferase Activity (RLU)

+1
Luc
+1 Luc
-975
+1 Luc
-975
pGL3Basic
+1
Luc
pGL3Basic
+1
Luc
-50
+1 Luc
-50
+1 Luc
-106
+1 Luc
-106
+1 Luc
-154
+1 Luc
-154
+1 Luc
-320
+1 Luc
-320
+1 Luc
-404
+1 Luc

+1
Luc
-50
+1 Luc
-106
+1
Luc
-154
Luc
-320
+1 Luc
-404
+1 Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1
Prm2 Prm3
+786
-1394
+1
Luc
+1 Luc
-975
pGL3Basic

+1 Luc
-50
+1 Luc
-106
+1
Luc
-106
+1
Luc
-154
Luc
-154
Luc
-320
+1 Luc
-320
+1 Luc
-404
+1 Luc
-404
+1 Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1

+1
Luc
-320
+1 Luc
-404
Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1
Prm2 Prm3
+786
+1
-1394
+1
Luc
+1
Luc
-975
pGL3Enhancer
+1
Luc
-50
+1
Luc

-975
+1
Luc
-975
pGL3Enhancer
+1
Luc
pGL3Enhancer
+1
Luc
-50
+1
Luc
-50
+1
Luc
-106
+1
Luc
-106
+1
Luc
-154
+1
Luc
-154
+1
Luc
-320
+1 Luc

Luc
+1 Luc
-975
pGL3Enhancer
+1
Luc
-50
+1 Luc
-106
Luc
-154
+1 Luc
-320
+1 Luc
-404
+1 Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1
Prm2
Prm3
+786
-1394
+1

Luc
+1 Luc
-975
pGL3Enhancer
+1
Luc
-50
+1 Luc
-106
Luc
-154
+1 Luc
-320
+1 Luc
-404
+1 Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1
Prm2
Prm3
+786
-1394
+1

-404
+1 Luc
-404
+1 Luc
-8500
E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1
Prm2
Prm3
+786
-8500
E1E1
E2
E1b
-1979 +1
-3308
-5895
-1394
Prm1
Prm2
Prm3
+786
Fig. 1. Effect of 5¢-deletion mutagenesis on Prm3-directed luciferase expression. (A–D) A schematic figure of the human TP genomic region
spanning nucleotides )8500 to +786 encoding promoter (Prm) 1, Prm2 and Prm3, in addition to exon (E) 1, E1b and E2, which are illustrated

recombinant plasmids was not substantially different
from that of the corresponding empty pGL3Basic
(compare 0.35 ± 0.08 RLU vs. 0.11 ± 0.03 RLU)
or pGL3Enhancer (compare 0.84 ± 0.11 RLU vs.
0.32 ± 0.01 RLU) vectors. Similar data were observed
in HEK293 cells (data not shown). Moreover, the
possible requirement for regulatory DNA sequences 3¢
of the +1 translational start site was investigated by
comparing luciferase activity of the previously charac-
terized )404 to +1 fragment to that of a )404 to
+119 fragment, containing an additional 119 bp of TP
genomic sequence downstream of the translational
start site (Fig. 2A,B). However, the level of )404 to
+119 directed luciferase activity was not significantly
different from that of the Prm3-directed luciferase
activity (e.g. )404 to +1) expressed in HEL cells irres-
pective of whether recombinant pGL3Basic (Fig. 2A)
or pGL3Enhancer (Fig. 2B) based-vectors were used.
Similar data were observed in HEK293 cells (data not
shown). In summary, we have identified three regula-
tory regions within Prm3 that contribute to basal pro-
moter activity, one that negatively ()404 to )320)
regulates the action of Prm3 while two of which posi-
tively ()154 to )106, )50 to +1) regulate basal Prm3
activity. Moreover, we have confirmed that nucleotides
)118 to +1 are essential for the core Prm3.
Identification of a functional Oct-1 ⁄ 2 element
within promoter 3
In order to further localize and identify the positive
regulatory element(s) positioned between )154 to )106

-404
+ 119
-1394
E2
+1
+786
*
*
*
*
012345
Luciferase Activity (RLU)
0246810
Luciferase Activity (RLU)
+1
Luc
+1
Luc
-118
Luc
-118
Luc
+119
Luc
+119
Luc
-118
-404
+ 119
-1394

-404
+1 19
-1394
E2
+1
+786
*
*
*
*
+1
Luc
-118
Luc
+119
Luc
-118
-404
+1 19
-1394
E2
+1
+786
*
*
*
*
+1
Luc
+1

+1
+786
*
*
*
*
A
B
Fig. 2. Localization of the core Prm3 by
5¢- and 3¢-deletion analysis. (A and B) The
TP genomic region spanning nucleotides
)1394 to +786, and encoding Prm3 ()1394
to +1) in addition to exon (E) 2, is illustrated
above each panel. Recombinant pGL3Basic
(A) or pGL3Enhancer (B) plasmids encoding
Prm3a ()404 to +1), Prm3f ()404 to )118)
and Prm3e ()404 to +119) were cotransfect-
ed with pRL-TK into HEL92.1.7 cells. Firefly
and renilla luciferase activity was assayed
48 h post-transfection; mean firefly relative
to renilla luciferase activity are expressed in
arbitrary relative luciferase units (RLU ±
SEM; n ¼ 5). The asterisks (*) indicate that
the level of Prm3f-directed luciferase activity
was significantly reduced relative to
Prm3a-directed luciferase expression, where
****P £ 0.0001.
A. T. Coyle and B. T. Kinsella Thromboxane A2 receptor gene expression
FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS 1039
deletion of a 13 bp gene segment between )119 and

factors capable of binding to the latter Oct-1 ⁄ 2 site
centered at )105, electromobility shift assays (EMSAs)
were carried out using a radiolabeled double-stranded
DNA probe spanning nucleotides )115 to )92 (Oct-
1 ⁄ 2
WT
; Kin195) and nuclear extract prepared from
HEL cells. Incubation of the radiolabeled Oct-1 ⁄ 2
WT
probe with HEL nuclear extract resulted in the appear-
ance of a single-labeled DNA–protein band (Fig. 4A,
lane 2) that was efficiently inhibited by an excess of
the corresponding nonlabeled double-stranded Oct-
1 ⁄ 2
WT
oligonucleotide (Fig. 4A, lane 3) or by a dou-
ble-stranded oligonucleotide containing a recognized
consensus Oct-1 ⁄ 2 (Fig. 4A, lane 5). The specificity of
nuclear factor binding to the latter Oct-1 ⁄ 2 site was
also verified by the failure of excess double-stranded
-154 +1
Luc
-140 +1
Luc
-119 +1
Luc
-106 +1 Luc
AP-1
(-27)
Oct1

Oct1
(-123)
Oct1 /2
(-105)
-1394
+1
AP-1
(-27)
Oct1
(-123)
Oct1 /2
(-105)
-1394-1394
+1+1
*
*
*
*
*
*
*
*
+1
Luc
-404
-404
+1
Luc
+1 Luc
-404

+1 Luc
-404
+1
Luc
-320
+1 Luc
-320
+1
Luc
-320
AP-1
(-27)
Oct1
(-123)
Oct1 /2
(-105)
-1394
+1
*
*
*
*
*
*
*
*
*
*
*
*

Oct1
(-123)
Oct1 /2
(-105)
-1394
+1
+1
Luc
-404
+1
Luc
+1
Luc
-404
-404
+1
Luc
-404
+1
Luc
+1
Luc
+1
Luc
+1 Luc
-404
+1 Luc+1 Luc
-404
+1
Luc

Oct1 /2
(-105)
-1394
AP-1
(-27)
Oct1
(-123)
Oct1 /2
(-105)
-1394-1394
A
B
+1+1
Fig. 3. Identification of a functional Oct-1 ⁄ 2
site within Prm3. (A and B) Scheme of the
TP genomic region spanning Prm3 ()1394 to
+1) in addition to the relative positions of
putative Oct-1, Oct-1 ⁄ 2 and AP-1 elements
is illustrated above each panel. Recombinant
pGL3Basic plasmids encoding Prm3aa ()154
to +1), Prm3ax () 140 to +1), Prm3ac ()119
to +1) and Prm3aab ()330 to +1) (A) or
Prm3a ()404 to +1) or Prm3ab ()320 to +1)
and their site-directed variants Prm3a
Oct-1
*
and Prm3ab
Oct-1
*, Prm3a
Oct-1 ⁄ 2

WT
DNA probe
(Kin195 and its complement corresponding to nucleotides )115 to )92 of Prm3) was used in EMSAs (A) or in supershift assays (B) using
nuclear extracts from HEL92.1.7 cells. (A)
32
P-labeled Oct-1 ⁄ 2
WT
probe was incubated: without nuclear extract (lane 1); with nuclear extract
alone (lane 2); with nuclear extract in the presence of excess nonlabeled double-stranded specific competitor Oct-1 ⁄ 2
WT
oligonucleotide
(Kin195 and its complement, lane 3); with nuclear extract in the presence of excess nonlabeled double-stranded noncompetitor Oct-1 ⁄ 2*
oligonucleotide (Kin193 and its complement, lane 4); with nuclear extract in the presence of excess nonlabeled double-stranded consensus
Oct-1 ⁄ 2 oligonucleotide (Kin340 and its complement, lane 5); with nuclear extract in the presence of excess nonlabeled double-stranded
Ap-1 noncompetitor oligonucleotide (Kin189 and its complement, lane 6). (B)
32
P-labeled Oct-1 ⁄ 2
WT
probe was incubated without nuclear
extract (lane 1); with nuclear extract alone (lane 2); with nuclear extract preincubated for 30 min with anti-(Oct-1) IgG (sc-232x; lane 3); with
nuclear extract preincubated for 30 min with anti-(Oct-1) IgG in the presence of excess nonlabeled double-stranded consensus Oct-1 ⁄ 2
WT
oligonucleotide (Kin340 and its complement, lane 4); with nuclear extract preincubated for 30 min with anti-(Oct-2) IgG (sc-233x; lane 5); with
nuclear extract preincubated for 30 min with anti-(Oct-2) IgG in the presence of excess nonlabeled double-stranded consensus Oct-1 ⁄ 2*
oligonucleotide (Kin340 and its complement, lane 6). The arrow indicates the supershifted transcription factor: DNA complex detected in the
presence of the anti-(Oct-2) IgG (lane 5). DNA–protein complexes were subject to PAGE followed by autoradiography, as outlined. (C and D)
Western blot analysis of Oct-1 (C) and Oct-2 (D) expression in whole cell protein (60 lgÆ lane
)1
) prepared from HEL (C and D; lane 1) and
HEK293 (C and D; lane 2) cells. The positions of the molecular size markers (kDa) are indicated to the left and right of the (C) and (D),

ical for efficient basal Prm3-directed gene expression
and have confirmed the ability of both Oct-1 and
Oct-2 to bind and regulate Prm3-directed gene
expression.
Identification of a functional AP-1 element within
promoter 3
To further investigate the positive regulatory element(s)
located between )50 to +1 of Prm3 that directs low,
though significant, luciferase activity in both HEL
and HEK293 cells (Fig. 1), matinspector
TM
analysis
[26] of Prm3 revealed the presence of a high consen-
sus AP-1 element centered at )27 (Fig. 4) located
some 15 bp 5¢ of the previously identified transcrip-
tion initiation site within the TPb mRNA [25]. Hence,
to ascertain the functional role of this AP-1 site in
mediating basal Prm3 activity, its consensus core
sequence was disrupted by site directed mutagenesis
(GTGACT to GATCCT) in a range of 5¢-deletion
subfragments and the ability of the mutated AP-1
(AP-1*) relative to the AP-1
WT
Prm3 subfragments to
direct luciferase activity in HEL (Fig. 5A,B) and
HEK293 (Fig. 5C,D) cells was investigated. Following
transfection into HEL cells, in general the 5¢-deletion
fragments containing the mutated AP-1* site yielded
approximately 2.5-fold reductions in luciferase activity
relative to that of the corresponding subfragments

centered at )27, EMSAs were carried out using a
radiolabeled double-stranded oligonucleotide probe
spanning nucleotides )32 to )10 (AP-1
WT
) of Prm3
and nuclear extracts prepared from HEL92.1.7 cells.
Incubation of the radiolabeled AP-1
WT
probe (Kin189;
Fig. 6) with HEL nuclear extract resulted in the forma-
tion of a single radiolabeled nuclear factor–DNA com-
plex (Fig. 6, lane 2) that was efficiently competed by
an excess of the corresponding nonlabeled double-
stranded AP-1
WT
oligonucleotide (Fig. 6, lane 3) or
by a double-stranded oligonucleotide containing a
recognized consensus AP-1 sequence (Fig. 6, lane 5).
The specificity of nuclear factor binding to the radio-
labeled AP-1
WT
probe was further confirmed by the
failure of both a double-stranded oligonucleotide span-
ning nucleotides )32 to )10 but containing a mutated
AP-1* site (Kin162; Fig. 6, lane 4) and a double-stran-
ded oligonucleotide based on the previously identified
Oct-1 ⁄ 2 (Oct-1 ⁄ 2
WT
, Kin195, Fig. 6, lane 6) to inter-
fere with nuclear factor: DNA complex formation.

-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc
+1
-404
Luc
+1
Luc
+1
-320
-154
Luc
+1
Luc
+1
-50
Luc
+1
-50
***

-320
-154
Luc
+1
Luc
+1
-50
Luc
+1
-50
01234567
Luciferase Activity (RLU)
+1
-1394
Luc
+1
-1394
Luc
-1394
Luc
-404
Luc
+1
-404
Luc
+1
-320
Luc
-320
LucLuc

+1
-320
Luc
+1
-320
-154
Luc
+1-154
Luc
+1
Luc
+1
-50
Luc
+1
-50
Luc
+1
-50
Luc
+1
-50
***
****
**** ****
***
***
****
**** ****
***

Luc
+1
Luc
+1
-50
Luc
+1
-50
***
**
***
***
***
01234567
Luciferase Activity (RLU)
+1
-1394
Luc
-404
Luc
+1
+1
-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1

-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc
+1
-404
Luc
+1
Luc
+1
-320
-154
Luc
+1
Luc
+1
-50
Luc
+1
-50
01234567

E2
+786
-1394
Promoter
AB
DC
3
AP-1
+1
E2
+786+1
E2E2
+786
-1394
Luc
+1-1394
Luc
+1
-404
Luc
+1-404
Luc
+1
Luc
+1
-320
Luc
+1
-320
-154

-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc
+1
-404
Luc
+1
Luc
+1
-320
-154
Luc
+1
Luc+1
-50
Luc
+1
-50
*** ***
****

-320
-154
Luc
+1
Luc+1
-50
Luc
+1
-50
024681012
Luciferase Activity (RLU)
+1
-1394
Luc
-404
Luc
+1
+1
-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394

Luc
+1
+1
-320
Luc
+1
-320
Luc
-154
Luc
+1
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Promoter 3
AP-1
+1
E2
+786+1
E2E2
+786
-1394
Luc

*** ***
****
****
**
0 5 10 15 20
Luciferas e Activit
y
(RLU)
+1
-1394
Luc
-404
Luc
+1
+1
-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc
+1
-404

Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc
+1
-404
Luc
+1
Luc
+1
-320
-154
Luc
+1
Luc+1
-50
Luc
+1
-50
0 5 10 15 20
Luciferas e Activit
y

+1
Luc+1
-50
Luc
+1
-50
0 5 10 15 20
Luciferas e Activit
y
(RLU)
+1
-1394
Luc
-404
Luc
+1
+1
-320
Luc
-154
Luc
+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc

+1
-1394
Promoter 3
AP-1
+1
E2
+786
-1394
Luc
+1
-404
Luc
+1
Luc
+1
-320
-154
Luc
+1
Luc+1
-50
Luc
+1
-50
0 5 10 15 20
Luciferas e Activit
y
(RLU)
+1
-1394

-1394
Promoter 3
AP-1
+1
E2
+786+1
E2E2
+786
-1394
Luc
+1-1394
Luc
+1
-404
Luc
+1-404
Luc
+1
Luc
+1
-320
Luc
+1
-320
-154
Luc
+1
-154
Luc
+1

renilla luciferase activity was assayed 48 h post-transfection; results are presented as mean firefly relative to renilla luciferase activity,
expressed in arbitrary relative luciferase units (RLU ± SEM; n ¼ 5). The asterisks (*) indicate that mutation of the AP-1 element significantly
reduced Prm3-directed luciferase activity in HEL and HEK293 cells, where *, **, ***, **** indicate P £ 0.05, P £ 0.02, P £ 0.001,
P £ 0.0001, respectively.
A. T. Coyle and B. T. Kinsella Thromboxane A2 receptor gene expression
FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS 1043
(Fig. 7B) plasmids containing these mutations resulted
in a complete loss in Prm3-directed luciferase expres-
sion. More specifically, there was a 10- to 16-fold
reduction in luciferase activity directed by the Prm3
subfragment containing the mutated AP-1*, Oct-1 ⁄ 2*
sites relative to the corresponding subfragments con-
taining the wild-type sequences in either pGL3Basic
(Fig. 7A) or pGL3Enhancer (Fig. 7B). In fact, the
level of luciferase activity directed by the AP-1*,
Oct-1 ⁄ 2* subfragment in either pGL3Basic or
pGL3 Enhancer was not substantially greater than
that level found in cells transfected with the equival-
ent promoter-less empty vectors (Fig. 7A,B). Similar
data were generated in HEK293 cells (data not
shown). These data strongly indicate that the Oct-1 ⁄ 2
and AP-1 elements independently regulate Prm3 activ-
ity and that disruption of both sites obliterates basal
Prm3 activity.
Investigation of the role of the AP-1 site at -27
in phorbol myristic acid (PMA) induction of
promoter 3
Previous studies have shown that TPb mRNA and
Prm3-directed luciferase activity in HEL92.1.7 cells is
up-regulated in response to phorbol myristic acid

mediates PMA-induction of Prm3 expression. In
addition, we have identified the presence of a negat-
ive regulatory region between )404 and )320
upstream of the core promoter that acts as an
upstream repressor sequence.
Fig. 6. Demonstration of nuclear factor binding to a putative AP-1
element within Prm3 by electromobility shift assay. A
32
P-labeled
double-stranded AP-1
WT
DNA probe (Kin189 and its complement)
was used in electromobility shift assays (EMSAs) using nuclear
extracts prepared from HEL92.1.7 cells.
32
P-labeled AP-1
WT
probe
was incubated: without nuclear extract (lane 1); with nuclear extract
(lane 2); with nuclear extract in the presence of excess of nonlabe-
led specific double-stranded competitor AP-1
WT
oligonucleotide
(Kin189 and its complement; lane 3); with nuclear extract in the
presence of excess nonlabeled double-stranded AP-1* noncompeti-
tor oligonucleotide (Kin162 and its complement where the putative
AP-1 element centered at )27 was mutated, lane 4); with nuclear
extract in the presence of excess nonlabeled consensus double-
stranded AP-1 oligonucleotide (Kin338 and its complement, lane 5);
with nuclear extract in the presence of excess nonlabeled double-

effects of TXA
2
can be modulated in an isoform
and ⁄ or cell ⁄ tissue specific manner. The overall aim of
the current study was to carry out a detailed functional
characterization of Prm3, identifying the cis-acting ele-
ments regulating basal Prm3 activity with a view to
defining the key factors that direct TPb expression
under normal cellular conditions.
Similar to that of the previously characterized Prm1,
Prm3 belongs to the class of TATA-less promoters
[16,25]. In TATA-less promoters, assembly of the pre-
initiation complex relies on binding of multiple general
transcription factors, such as SP-1, in proximity to the
transcription initiation site [28]. Herein, successive
5¢-deletion of Prm3 to either 106 bp ()106 to +1) or
50 bp ()50 to +1) yielded a subfragment that retained
a significant, albeit reduced, ability to direct reporter
gene expression in both HEL and HEK293 cells whilst
deletion of the 3¢-terminal 118 bp of Prm3 ()118
to +1) led to a complete loss of promoter activity in
both cell types. Collectively these data established that
the critical core element(s) are located within the )118
to +1 region of Prm3.
Upstream activation sequences (UAS) and upstream
repressor sequences (URS) are gene-specific sequences
controlling the rate of transcription initiation [29].
Negative regulatory elements in particular have been
identified in a number of TATA-less promoters [30,31].
Consistent with this, successive 5¢-deletion of nucleo-

-404
-320
+1
+1
Luc
-320
+1
Luc
+1
Luc
-320
+1 Luc
-320
+1 Luc+1 Luc
-320
pGL3Basic
+1
Luc
pGL3Basic
+1
Luc
02
Luciferase Activity (RLU)
*
*
*
*
02
Luciferase Activity (RLU)
*

-320
-320
A
B
+1+1
024681012
Lucife rase Activity (RLU)
+1 Luc
-320
+1 Luc
-320
pGL3Enhancer
+1
Luc
*
*
*
*
AP-1
(-27)
Oct 1
(-123)
Oct 1/2
(-105)
-1394
-404
-320
+1
024681012
Lucife rase Activity (RLU)

*
*
*
*
024681012
Lucife rase Activity (RLU)
+1 Luc
-320
+1 Luc+1 Luc
-320
+1 Luc
-320
+1 Luc+1 Luc
-320
pGL3Enhancer
+1
Luc
pGL3Enhancer
+1
Luc
*
*
*
*
*
*
*
*
AP-1
(-27)

expression. (A and B) The TP genomic
region spanning Prm3 ()1394 to +1) in addi-
tion to the relative positions of putative Oct-
1, Oct-1 ⁄ 2 and AP-1 elements are illustrated
above each panel. Recombinant pGL3Basic
(A) or pGL3Enhancer (B) plasmids encoding
Prm3ab ()320 to +1) or its respective site-
directed variant Prm3ab
Oct1 ⁄ 2
*
,AP)1
*, where
both the Oct-1 ⁄ 2 and AP-1 elements cen-
tered at )105 and )27 of Prm3, respect-
ively, was disrupted by site-directed
mutagenesis were cotransfected with
pRL-TK into HEL92.1.7. Firefly and renilla
luciferase activity was assayed 48 h post-
transfection; results are presented as mean
firefly relative to renilla luciferase activity,
expressed in arbitrary relative luciferase
units (RLU ± SEM; n ¼ 5). The asterisks (*)
indicate that either deletion or site-directed
mutagenesis of Prm3 sequences signifi-
cantly reduced luciferase expression in HEL
cells, where **** indicates P £ 0.0001.
A. T. Coyle and B. T. Kinsella Thromboxane A2 receptor gene expression
FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS 1045
The octamer sequence element (consensus 5¢-ATGC
AAAT-3¢) present in promoters of immunoglobulin

-404
Luc
+1
-1394
A
B
AP-1
+1
E2
+786
-1394
Luc
+1
Luc
+1
-404
1234567
Fig. 8. Effect of PMA on Prm3-directed luciferase expression and nuclear factor binding. (A) The TP genomic region spanning nucleotides
)1394 to +786 encoding Prm3 ()1394 to +1) in addition to an AP-1 element and exon (E) 2 are illustrated above the panel. Recombinant
pGL3Basic plasmids encoding Prm3 ()1394 to +1), Prm3a ()404 to +1) or their respective site-directed variants Prm3
AP)1
* Prm3a
AP)1
*,
where the AP-1 element centered at )27 of Prm3 was mutated, were cotransfected with pRL-TK into HEL92.1.7 cells. Thirty-six hours post-
transfection, cells were incubated with either 100 n
M PMA or the vehicle (0.1% dimethylsulfoxide) for 16 h. Thereafter, firefly and renilla
luciferase activity was assayed; results are presented as mean firefly relative to renilla luciferase activity, expressed in arbitrary relative
luciferase units (RLU ± SEM; n ¼ 4). The asterisks (*) indicate that luciferase expression in HEL cells was significantly altered in PMA-
treated cells relative to vehicle treated cells, where **, *** indicates P £ 0.02, P £ 0.001, respectively. (B) A

whereby an A ⁄ G mutation severely affected Oct-1 or
Oct-2 binding in vitro [46].
Furthermore, EMSAs confirmed the presence of
transcription factors capable of binding to a double-
stranded DNA probe ()115 to )92; Oct-1 ⁄ 2
WT
) span-
ning the Oct-1 ⁄ 2 site at )105 in nuclear extracts from
HEL92.1.7 cells. Nuclear factor ⁄ DNA complex forma-
tion was efficiently competed by an excess of the
corresponding nonlabeled double-stranded Oct-1 ⁄ 2
WT
oligonucleotide but was not competed by the equival-
ent double-stranded oligonucleotide harboring the
A ⁄ G mutation within the core Oct-1 ⁄ 2* site. Western
blot analysis confirmed abundant expression of Oct-2,
but not Oct-1, in HEL cells and supershift assays
employing anti-(Oct-2) IgG further confirmed Oct-2
binding to the Oct-1 ⁄ 2 site of Prm3. Owing to the
absence of Oct-1 expression in HEL cells, these data
did not rule out the possibility that Oct-1 may regulate
Prm3-directed gene expression in other cell types, such
as in HEK293 cells where it is abundantly expressed.
Consistent with this, heterologous over-expression of
both Oct-1 and Oct-2 significantly increased Prm3-
directed luciferase activity in HEK293 and HEL cells.
Hence, it is evident that Oct-2 can function as a trans-
acting element capable of regulating Prm3 and TPb
expression in megakaryocytic HEL92.1.7 cells. In addi-
tion Oct-1 can also regulate Prm3 activity in other

oligonucleotide but not by the corresponding double-
stranded oligonucleotide containing the mutated
AP-1* element. Moreover, EMSAs employing a pan-
Jun antibody confirmed binding of Jun protein to the
AP-1 element that was specifically competed by
the nonlabeled double-stranded AP-1
wt
but not by the
AP-1* oligonucleotide (data not shown). The com-
bined contribution of the AP-1 and Oct-1 ⁄ 2 cis-acting
elements were confirmed whereby mutation of both
the AP-1* and Oct-1 ⁄ 2* elements yielded 10- to
16-fold reductions in luciferase expression, effectively
abolishing Prm3 activity in both HEL and HEK293
cells. Consistent with our observations, both AP-1
and Oct members can bind and direct transcription
from a number of TATA-less promoters [48–51]. In
addition, although both the AP-1 complex and certain
Oct family members are ubiquitously expressed, they
can also regulate transcription in a cell type-specific
manner through the recruitment of cell specific tran-
scription factors and coregulators [52].
The AP-1 binding sequence was originally identified
as a tetradecanoyl phorbol myristic acid (TPA) ⁄ phor-
bol myristate acetate (PMA) response element (TRE)
and treatment of cultured cells with PMA results in a
strong increase in AP-1 binding to TREs [53]. Prm1-
and Prm3- reporter gene expression and TPa and TPb
mRNA expression are up-regulated by PMA in HEL
cells [25]. While the site of action of PMA within

core promoter elements that are critical for basal Prm3
activity and the AP-1 element mediates PMA-induc-
tion of Prm3 activity. In addition, we have identified a
negative URS between )404 and )320. Collectively,
these data provide valuable insights into the factors
determining both the basal and PMA up-regulation of
TPb expression and may provide a platform to deter-
mine the relative contributions of differentially regula-
ted expression of TPa and TPb to haemostasis and
possibly to vascular disease.
Experimental procedures
Materials
pGL3Basic, pGL3Enhancer, pRL-Thymidine Kinase
(pRL-TK) and Dual LuciferaseÒ Reporter Assay System
were obtained from Promega Corporation (Madison, WI,
USA). DMRIE-CÒ was from Invitrogen Life Techno-
logies (CA, USA). EffecteneÒ was from Qiagen Ltd
(Crawley, UK). [
32
P]ATP[cP] (6000 Ci Æmmol
)1
at 10 mCiÆ
mL
)1
) was from Valeant Pharmaceuticals (ICN; Costa
Mesa, USA). All other reagents were molecular biology
grade. Anti-(Oct-1) (sc-232x), anti-(Oct-2) (sc-233x), anti-(c-
jun) (sc-44x) Igs were obtained from Santa Cruz Biotech-
nology.
Construction of luciferase-based genetic reporter

(3) Prm3ab, pGL3b:Prm3ab & pGL3e:Prm3ab. (Primer
Kin146; 5¢-dGAGA
GGTACCTGGGAGGCTGAGATGG-
3¢, corresponding to NTs )320–304).
(4) Prm3aa, pGL3b:Prm3aa & pGL3e:Prm3aa. (Primer
Kin145; 5¢-dGAGA
GGTACCTAGGAGTTCACCAGA
GC-3¢, corresponding to NTs )154 to )137).
(5) Prm3ax; pGL3b:Prm3ax & pGL3e:Prm3ax. (Primer
Kin177; 5¢-dGAGA
GGTACCAGCTACTTACACTGAAA
TGCAG-3¢, corresponding to NTs )140 to )118).
(6) Prm3ac; pGL3b:Prm3ac & pGL3e:Prm3ac. (Primer
Kin188; 5¢-dGAGA
GGTACCGAATTAATCACAAGCAA
ATCTTCTC-3¢, corresponding to NTs )119 to )94).
(7) Prm3aab; pGL3b:Prm3aab & pGL3e:Prm3aab. (Primer
Kin160; 5¢-dGAGA
GGTACCGCAAATCTTCTCTCGCC
TCC-3¢, corresponding to NTs )106 to )86).
(8) Prm3aaa; pGL3b:Prm3aaa & pGL3e:Prm3aaa. (Primer
Kin161, 5¢-dGAGA
GGTACCGCAGCATCGGCCTGATG
GG-3¢, corresponding to NTs )50 to )31).
The gene fragment Prm3e ()404 to +119 of Prm3) was
amplified by PCR using pWE15:TXR [55] as template and
primers Kin143 (5¢-dGAGA
GGTACCCTCACGCCTGTA
ATCCCAG-3¢, corresponding to NTs )404 to )385) and
Kin245 (5¢-dAGAG

AP)1
*, pGL3e:
Prm3
AP)1
*, pGL3b: Prm3a
AP)1
*, pGL3e:Prm3a
AP)1
*,
pGL3b:Prm3ab
AP)1
*, pGL3e: Prm3ab
AP)1
*, pGL3b:
Prm3aa
AP)1
*, pGL3e:Prm3aa
AP)1
*, pGL3b:Prm3aaa
AP)1
*
and pGL3e:Prm3aaa
AP)1
*.
Mutation of the Oct-1 element with the sequence aaA
TGCa to aaTTCCa (core bases shown in uppercase letters)
centered at )123 within Prm3 was performed using the
mutator primers Kin175 (5¢-CACCAGAGCTACTTACA
CTGAATTCCAGAATAATCACAAGCAAATC-3¢; sense
primer) vs. its complement generating pGL3b:Prm3a

Prm3ab
Oct1 ⁄ 2
*
,AP)1
* and pGL3e:Prm3ab Oct1
⁄ 2
*
,AP)1
*. In
each case, mutated bases within the mutator primers are
highlighted in bold type.
Cell culture
All mammalian cells were grown at 37 °C in a humid envi-
ronment with 5% CO
2
. Human erythroleukemic (HEL)
92.1.7 cells and human embryonic kidney (HEK) 293 cells
were cultured in RPMI 1640, 10% fetal bovine serum and
in Eagle’s minimal essential medium (MEM), 10% fetal
bovine serum, respectively.
Assay of luciferase activity
HEK293 cells were plated in MEM, 10% FBS in six well
dishes at 1 · 10
5
cells per well. At 70–80% confluence, cells
were cotransfected with control ⁄ recombinant pGL3Basic or
pGL3Enhancer vectors (0.4 lg per well), encoding firefly
luciferase, along with pRL TK (50 ng per well), encoding
renilla luciferase, using effectene as recommended by the
supplier. To investigate the effect of Oct-1 or Oct-2 on

incubator, after which 2 mL of
RPMI 1640 containing 15% fetal bovine serum was
added. Forty-eight hours after transfection, cells were
washed in ice-cold NaCl ⁄ P
i
and harvested at 1200 g for
5 min at 4 °C. Where relevant, 36 h prior post-transfection
HEL cells with incubated with PMA (100 nm) or, as con-
trols, with an equivalent volume of the vehicle (0.1%
dimethylsulfoxide) and cells were further incubated for
16 h prior to harvesting. Cell pellets were resuspended in
reporter lysis buffer (100 lL) and were lysed by repeated
trituration. Cell lysates were prepared by centrifugation at
14 000 g for 1 min at room temperature. To investigate
the effect of Oct-1 or Oct-2 on Prm3-directed luciferase
expression, pcDNA3:HaOct-1 or pcDNA3:HaOct-2 plas-
mids [46] were transiently over-expressed in HEL cells
along with recombinant pGL3Basic vectors encoding Prm3
or its subfragments. Briefly, recombinant pGL3b:Prm3ab
(1.0 lg per well) plus pRL TK (200 ng per well) plasmids
were transiently cotransfected along with either pcDNA3:
HaOct-1 (1.0 lg per well), pcDNA3:HaOct-2 (1.0 lg per
well) or, as a negative control, pcDNA3 (1.0 lg per well)
using DMRIE-C. Cells were harvested 48 h post-transfec-
A. T. Coyle and B. T. Kinsella Thromboxane A2 receptor gene expression
FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS 1049
tion and cells lysed and prepared for luciferase assays as
described above.
HEK293 and HEL cell supernatants were assayed for
both firefly and renilla luciferase activity using the Dual

tol was excluded from the various isolation buffers.
Electrophoretic mobility shift and supershift
assays
Oligonucleotides corresponding to the sense and antisense
strands of each probe (0.35 lm of each) were annealed in
1· T4 polynucleotide kinase (PNK) buffer (70 mm
Tris ⁄ HCl, pH 7.6.,10 mm MgCl
2
,5mm dithiothreitol;
8 lL) by heating at 95 °C for 2 min followed by slow cool-
ing to room temperature for 30 min. The resulting double-
stranded probes were then radiolabeled in 10 l L reactions
containing 0.35 lm double-stranded oligonucleotide, 1 lL
[
32
P]ATP[cP] (6000 CiÆmmol
)1
at 10 mCiÆmL
)1
; ICN) and
1 lL T4 PNK (10 UÆlL
)1
)at37°C for 30 min. Following
labeling, the reactions were diluted 1 : 10 with 10 mm
Tris ⁄ HCl, pH 8.0.,1 mm EDTA (TE, pH 8.0) buffer to
achieve a final concentration of
32
P-radiolabeled labeled of
0.035 lm. The unincorporated [
32

are in bold italics); (c) AP-1
WT
(Kin189; 5¢-dGGTGGTGAC
TGATCCCTCAGGGC-3¢; corresponding to NTs )32 to
)10 of Prm3); (d) mutated AP-1* (Kin162; 5¢-dCGGCCT
GATGGGGTGGATCCTGATCCCTCAGGGC-3¢; corres-
ponding to NTs )46 to )7 of Prm3 where bases mutated
from the wild-type Prm3 sequence are in bold italics); (e) SP-
1 consensus site (Promega) with the sequence: 5¢-dATTCG
ATCGGGGCGGGGCGAG-3¢; (f) AP-1 consensus site
(Promega; Kin338; 5¢ dCGCTTGATGAGTCAGCCGGAA-
3¢); (g) octamer (Oct) consensus site (Promega; Kin340;
5¢-dTGTCAGATGCAAATCACTAGAA-3¢). Note, only
forward oligonucleotides are given and sequences of the
corresponding complementary strands are omitted.
For electrophoretic mobility supershift assays, nuclear
extracts (4 lg total protein) was preincubated with 2 lL for
either anti-(Oct-1) (sc-232x; 2 mgÆmL
)1
stock concentra-
tion), anti-(Oct-2) (sc-233x; 2 mgÆmL
)1
stock concentra-
tion), anti-(c-jun) (sc-44x; 2 mgÆmL
)1
stock concentration)
for 30 min at 4 °C. Thereafter, nuclear extract ⁄ antibody
mixtures were incubated for 15 min at room temperature
with ⁄ without a 57-fold molar excess of unlabeled double-
stranded competitor ⁄ noncompetitor oligonucleotide (2 lm)

come Trust, The Health Research Board and Enter-
prise Ireland. We are grateful to Dr Harinder Singh,
Department Molecular Genetics and Cell Biology,
University of Chicago for the gifts of pcDNA3:HaOct1
and pcDNA3:HaOct2.
References
1 Narumiya S, Sugimoto Y & Ushikubi F (1999) Prosta-
noid receptors: structures, properties, and functions.
Physiol Rev 79, 1193–1226.
2 Breyer RM, Bagdassarian CK, Myers SA & Breyer MD
(2001) Prostanoid receptors: subtypes and signalling.
Annu Rev Pharmacol Toxicol 41, 661–690.
3 Kinsella BT (2001) Thromboxane A2 signalling in
humans: a ‘Tail’ of two receptors. Biochem Soc Trans
29, 641–654.
4 Lianos EA & Bresnahan BA (1999) Effect of thrombox-
ane A2 inhibition and antagonism on prostaglandin and
leukotriene synthesis in glomerular immune injury.
J Lab Clin Med 134, 478–482.
5 Dorn GW, 2nd Becker MW & Davis MG (1992) Disso-
ciation of the contractile and hypertrophic effects of
vasoconstrictor prostanoids in vascular smooth muscle.
J Biol Chem 267, 24897–24905.
6 Ali S, Davis MG, Becker MW, Dorn GW & 2nd.
(1993) Thromboxane A2 stimulates vascular smooth
muscle hypertrophy by up-regulating the synthesis and
release of endogenous basic fibroblast growth factor.
J Biol Chem 268, 17397–17403.
7 Ashton AW, Yokota R, John G, Zhao S, Suadicani SO,
Spray DC & Ware JA (1999) Inhibition of endothelial

TM, Grosser T, Lawson JA & FitzGerald GA (2002)
Role of prostacyclin in the cardiovascular response to
thromboxane A2. Science 296, 539–541.
15 Hirata M, Hayashi Y, Ushikubi F, Yokota Y,
Kageyama R, Nakanishi S & Narumiya S (1991)
Cloning and expression of cDNA for a human
thromboxane A2 receptor. Nature 349, 617–620.
16 Nusing RM, Hirata M, Kakizuka A, Eki T, Ozawa K
& Narumiya S (1993) Characterization and chromoso-
mal mapping of the human thromboxane A2 receptor
gene. J Biol Chem 268, 25253–25259.
17 Raychowdhury MK, Yukawa M, Collins LJ, McGrail
SH, Kent KC & Ware JA (1994) Alternative splicing
produces a divergent cytoplasmic tail in the human
endothelial thromboxane A2 receptor. J Biol Chem 269,
19256–19261.
18 Hirata T, Ushikubi F, Kakizuka A, Okuma M &
Narumiya S (1996) Two thromboxane A2 receptor
isoforms in human platelets: opposite coupling to
adenylyl cyclase with different sensitivity to Arg60 to
Leu mutation. J Clin Invest 97, 949–956.
19 Vezza R, Habib A & FitzGerald GA (1999) Differential
signaling by the thromboxane receptor isoforms via the
novel GTP-binding protein, Gh. J Biol Chem 274,
12774–12779.
20 Parent JL, Labrecque P, Orsini MJ & Benovic JL
(1999) Internalization of the TXA2 receptor alpha and
beta isoforms: role of the differentially spliced COOH
terminus in agonist-promoted receptor internalization.
J Biol Chem 274, 8941–8948.

mammalian Jun proteins: redundancy and specificity.
Oncogene 20, 2378–2389.
28 Zhang Y & Dufau ML (2003) Dual mechanisms of
regulation of transcription of luteinizing hormone
receptor gene by nuclear orphan receptors and histone
deacetylase complexes. J Steroid Biochem Mol Biol 85,
401–414.
29 Struhl K (1995) Yeast transcriptional regulatory mecha-
nisms. Annu Rev Genet 29, 651–674.
30 Hata K & Mizuguchi J (2004) Genomic organization
and characterization of the promoter for the E2A gene.
Gene 325, 53–61.
31 van Rietschoten JG, Smits HH, van de Wetering D,
Westland R, Verweij CL, den Hartog MT & Wierenga
EA (2001) Silencer activity of NFATc2 in the interleu-
kin-12 receptor beta 2 proximal promoter in human T
helper cells. J Biol Chem 276, 34509–34516.
32 Zhao FQ, Adachi K & Oka T (2002) Involvement of
Oct-1 in transcriptional regulation of beta-casein gene
expression in mouse mammary gland. Biochim Biophys
Acta 1577, 27–37.
33 Strom AC, Forsberg M, Lillhager P & Westin G (1996)
The transcription factors Sp1 and Oct-1 interact physi-
cally to regulate human U2 snRNA gene expression.
Nucleic Acids Res 24, 1981–1986.
34 Das G & Herr W (1993) Enhanced activation of the
human histone H2B promoter by an Oct-1 variant
generated by alternative splicing. J Biol Chem 268,
25026–25032.
35 Verrijzer CP & Van der Vliet PC (1993) POU domain

malian brain development. Nature 340, 35–41.
43 Cockerill PN & Klinken PS (1990) Octamer-binding
proteins in diverse hemopoietic cells. Mol Cell Biol 10 ,
1293–1296.
44 Fujimori K, Kanaoka Y, Sakaguchi Y & Urade Y
(2000) Transcriptional activation of the human hemato-
poietic prostaglandin D synthase gene in megakaryo-
blastic cells. Roles of the oct-1 element in the
5¢-flanking region and the AP-2 element in the
untranslated exon 1. J Biol Chem 275, 40511–40516.
45 Groenen MA, Dijkhof RJ, van der Poel JJ, van Digge-
len R & Verstege E (1992) Multiple octamer binding
sites in the promoter region of the bovine alpha
s2-casein gene. Nucleic Acids Res 20, 4311–4318.
46 Shah PC, Bertolino E & Singh H (1997) Using altered
specificity Oct-1 and Oct-2 mutants to analyze the regu-
lation of immunoglobulin gene transcription. EMBO
J 16, 7105–7117.
47 Becker JC, Nikroo A, Brabletz T & Reisfeld RA
(1995) DNA loops induced by cooperative binding of
transcriptional activator proteins and preinitiation com-
plexes. Proc Natl Acad Sci USA 92, 9727–9731.
Thromboxane A2 receptor gene expression A. T. Coyle and B. T. Kinsella
1052 FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS
48 Wang F, Gao JX, Mimura J, Kobayashi A, Sogawa K
& Fujii-Kuriyama Y (1998) Structure and expression of
the mouse AhR nuclear translocator (mArnt) gene.
J Biol Chem 273, 24867–24873.
49 Wolf I, Pevzner V, Kaiser E, Bernhardt G, Claudio E,
Siebenlist U, Forster R & Lipp M (1998) Downstream

coids increase retinoid-X receptor alpha (RXRalpha)
expression and enhance thyroid hormone action in
primary cultured rat hepatocytes. J Mol Endocrinol
22, 81–90.
A. T. Coyle and B. T. Kinsella Thromboxane A2 receptor gene expression
FEBS Journal 272 (2005) 1036–1053 ª 2005 FEBS 1053