Tài liệu Báo cáo khoa học: Gene transcription of fgl2 in endothelial cells is controlled by Ets-1 and Oct-1 and requires the presence of both Sp1 and Sp3 pdf

Gene transcription of fgl2 in endothelial cells is controlled by Ets-1
and Oct-1 and requires the presence of both Sp1 and Sp3
Mingfeng Liu
1
, Julian L. Leibowitz
2
, David A. Clark
1,4
, Michael Mendicino
1
, Qin Ning
1
, Jin Wen Ding
1
,
Cheryl D’Abreo
3
, Laisum Fung
1
, Philip A. Marsden
3
and Gary A. Levy
1
1
Multi Organ Transplant Program, Toronto General Hospital and The University of Toronto, Canada;
2
Department of Pathology and
Laboratory Medicine, Texas A & M University System College of Medicine, USA;
3
Renal Division and Department of Medicine,
St. Michael’s Hospital and University of Toronto, Canada;

4
McMaster University, Ontario, Canada
The immune coagulant fgl2/fibroleukin has been previously
shown to play a pivotal role in the pathogenesis of murine
and human fulminant hepatitis and fetal loss syndrome.
Constitutive expression of fgl2 transcripts at low levels are
seen in cytotoxic T cells, endothelial, intestinal and tropho-
blast cells, while specific factors (such as virus and cytokines)
are required to induce high levels of fgl2 expression in other
cell types including monocytes/macrophages. To address
the transcriptional mechanisms that regulate constitutive
expression of fgl2, murine genomic clones were characterized
and the transcription start site was defined by 5¢-RACE and
primer extension. A comprehensive assessment of basal
fgl2 promoter activity in murine vascular endothelial cells
defined a minimal 119 bp region responsible for constitutive
fgl2 transcription. A complex positive regulatory domain
(PRD) spanning a 39-bp sequence from )87 to )49 (relative
to the transcription start site) was identified. Electrophoretic
mobility shift assay studies in vascular endothelial cells
revealed that the nucleoprotein complexes that form on this
positive regulatory domain (PRD) contain Sp1/Sp3 family
members, Oct-1, and Ets-1. Heterologous expression studies
in Drosophila Schneider cells confirmed that the constitutive
expression of this gene is controlled by Ets-1 and requires the
presence both of the Sp1 and Sp3 transcription factors. The
presence of this complex multicomponent PRD in the fgl2
proximal promoter is consistent with the observation that,
in vivo, fgl2 expression is tightly regulated. Moreover, viral
induced fgl2 expression also requires the presence of this

PRD. These results clearly demonstrate that multiple cis
DNA elements in a clustered region work cooperatively to
regulate constitutive fgl2 expression and interact with indu-
cible elements to regulate viral-induced fgl2 expression in
endothelial cells.
Keywords: fibroleukin; fgl2; hepatitis; immune coagulant;
transcription regulation.
Activation of the coagulation system represents an import-
ant facet of immune and inflammatory reactions, account-
ing for the fibrin deposition that is commonly observed in
these reactions [1]. Cellular procoagulants and the soluble
factors of the coagulation cascade are important partici-
pants in a number of human diseases including allograft
rejection [2,3], glomerulonephritis [4,5], septic shock [6,7],
and bacterial and viral infections [8]. For example, tissue
factor (TF) [9], the transmembrane receptor of factor VII, is
the major procoagulant of the classical extrinsic pathway of
coagulation resulting in thrombin generation, and subse-
quent fibrin deposition. Furthermore, TF has important
roles in the regulation of angiogenesis in cancer, embryonic
blood vessel development, and intracellular signaling
[10–12]. Thrombin is involved in many biological processes.
In addition to its function in blood coagulation and wound
healing, thrombin has diverse functions in many different
cell types. For instance, thrombin induces proliferation of
fibroblasts and smooth muscle cells, and neurite retraction
and synapse reduction in neurons [12]. Moreover, thrombin
is a chemotactic agent for inflammatory cells such as
macrophages and neutrophils [12].
fgl2/fibroleukin was originally described as a fibrinogen-

like protein, constitutively expressed in cytotoxic T cells
[13]. This gene was subsequently determined to encode an
immune coagulant and was localized to the proximal
region of mouse chromosome 5 [14,15]. Fgl2 is a member
of the fibrinogen family, which includes tenascin, cyto-
toxin, and fibrinogen [16–18]. Fibrinogen-like protein
family members all contain fibrinogen-related domains
(FRED), known to represent key regions for protein–
protein interaction. As a procoagulant, fgl2 is involved in
cleaving prothrombin to thrombin endogenously and
when functionally expressed in a heterologous system
[14,19]. Thus, an important biological function of fgl2
might be to regulate the production of thrombin via a
pathway independent of TF, especially in settings where
TF-independent procoagulant activity has been documen-
ted. For instance, fgl2 plays a critical role in the
pathogenesis of both human and mouse fulminant viral
Correspondence to G. A. Levy, Toronto General Hospital, 621
University Ave., NU-10–116, Toronto, Ontario, Canada M5G 2C4.
Fax: 416 340 3378, Tel.: 416 340 5166,
E-mail:
Abbreviations: EMSA, electrophoretic mobility shift assays; fgl2,
fibrinogen-like protein; LUC, luciferase; MHV-3, murine hepatitis
virus strain 3; NE, nuclear extracts; PRD, positive regulatory domain;
Stat3, signal transducer and transactivator 3.
(Received 6 December 2002, revised 4 March 2003,
accepted 27 March 2003)
Eur. J. Biochem. 270, 2274–2286 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03595.x
hepatitis [20]. We have previously demonstrated the
pivotal role of fgl2 in a murine model for murine hepatitis

virus strain 3 (MHV-3) induced fulminant hepatic failure
[21,22]. Administration of neutralizing antibodies against
fgl2 conferred protection against MHV-3 induced fulmi-
nant hepatic failure in susceptible BALB/cJ mice [23].
Humanfgl2wasclonedandshowntobeexpressedin
endothelial cells and macrophages of livers from patients
with hepatitis [20,24]. Expression of fgl2 during both
mouse and human fulminant hepatitis correlates with
fibrin deposition, a typical feature of fulminant hepatic
failure [20,21,25,26].
As a multifunctional protein, fgl2 has a transcription
regulation mechanism that seems to reflect its function in
different cells and developmental stages. Fgl2 mRNA is
constitutively expressed in some cell types [13,20,26],
but transcription can also be robustly induced by viruses
and cytokines, such as interferon-c (IFN-c) [26–28]. Con-
stitutive activity of the fgl2 promoter suggests that fgl2
functions as a matrix/adhesion protein, possibly necessary
for development of a normal fetus and in regulating
immune responses within the intestine. Induction of fgl2
in macrophages in response to viruses results in the
production of a potent coagulant activity with the charac-
teristics of a prothrombinase. This combination of consti-
tutive expression and induction may permit a rapid response
to inflammation. However, the mechanisms regulating fgl2
expression, either constitutive or induced, remain unclear. A
comprehensive understanding of the general transcription
machinery on the native fgl2 promoter is necessary for
developing further insight into the TF-independent coagu-
lation pathway in diseased states. We report here the

structural and functional characterization of the 5¢-flanking
region of the mouse fgl2 gene in vascular endothelial cells.
Through a series of studies focusing on constitutive
expression in endothelial cells, we identified that Ets-1
involves in the fgl2 constitutive expression within a positive
regulatory domain formed by Sp1, Sp3, Ets-1, and Oct-1. In
addition, we demonstrated that this positive regulatory
domain is also required for MHV-3 nucleocapsid protein-
induced fgl2 expression.
Materials and methods
Mice
Female BALB/cJ mice, 6–8 weeks of age were purchased
from Jackson Laboratories (Bar Harbor, ME), and were
housed in the animal facility at the Toronto Hospital
Research Institute, University of Toronto.
Cells
Peritoneal macrophages used for primer extension and
5¢-RACE were harvested from BALB/cJ mice 5 days after
intraperitoneal administration of 1.5 mL of 5% thioglycol-
late (Difco Laboratories, Detroit, Michigan) as previously
described [29]. Macrophages were greater than 95% pure as
determined by morphology and viability exceeded 98% by
trypan blue exclusion. MHV-3 was grown and titrated as
described previously [8]. The murine SVE-10 endothelial cell
line and Schneider’s Drosophila line 2 (SL2) were obtained
from American Type Culture Collection (ATCC Rockport,
MD). Cells were maintained in DMEM (Dulbecco’s
modified Eagle’s medium) (SVE-10, ATCC) or Schneider’s
Drosophila medium (Invitrogen) supplemented with 10%
fetal bovine serum and maintained as described [30].

Sequencing
Plasmid DNA fragments were sequenced by cycle sequen-
cing on an automated DNA sequencer (Model 377, Applied
Biosystems, Foster City, CA) using dideoxy dye terminator
chemistry and primer-directed strategies. Sequences were
determined on both DNA strands.
Primer extension analysis
A 30 base oligonucleotide (5¢-CCTCCACCGCTCGGCA
GGCAGCGAGGACGG-3¢) complementary to nucleo-
tides 1363–1392 (GenBank Accession AF025817) at the
5¢ end of the mouse fgl2 coding sequence was purified by
polyacrylamide gel electrophoresis and used for the primer
extension reaction. End labelling and primer extension
reactions were performed as described previously [31].
Labelled oligonucleotide (100 000 c.p.m.) was hybridized
with 50 lg of total RNA from MHV-3 infected (6 h)
BALB/cJ mouse peritoneal macrophages. Nucleic acids
were recovered by ethanol precipitation and reverse tran-
scribed with 100 U of MMLV reverse transcriptase
(MMLV-RT). The primer extension products were extrac-
ted with phenol/chloroform, precipitated with ethanol and
analyzed in 8% sequencing gel. The same primer was used
to create a dideoxynucleotide chain termination-sequencing
ladder from a double stranded genomic DNA fragment
loaded adjacent to the primer extension sample.
5¢-RACE analysis
5¢-RACE analysis was performed as previously described
with minor modifications [31,32]. Briefly, 0.5 lgoftotal
RNA extracted from MHV-3 treated (at 10 pfu for 24 h)
Balb/cJ liver and peritoneal macrophages infected MHV-3

for 6 h were heated to 95 °C for 5 min, cooled on ice and
followed by mixing with one picomole of an antisense
primer (5¢-ATCTCGATGGTCGTCAGCC-3¢), which
corresponds to nucleotides 1627–1646 (GenBank Acces-
sion AF025817) of the BALB/cJ fgl2 gene. Reverse
transcription was carried out for 2 h at 42 °C with 1 unit
of MMLV-RT. The sample was heated to 95 °Cfor
5 min to inactivate the MMLV-RT and digested with
1 unit of RNase at 37 °C for 20 min. First round PCR
was performed in a total volume of 100 lL containing
the following: 50 lLofH
2
O, 2.5 pmole PCR 4 (sense
primer: 5¢-GACTCGAGTCGACGAATTCAAT-3¢),
25 pmole PCR 5 (5¢-GACTCGAGTCGACGAATT
CAA-3¢), 25-pmole of gene specific antisense primer
PCA2 (5¢-TGCCACTGCTTCCTTGAGG-3¢), 10 lLof
10 · PCR buffer, 10 lL25m
M
MgCl
2
,2lLof10m
M
dNTP, 20 lL tailed first strand cDNA and 0.5 units of
Taq polymerase. The second PCR amplification was
performed using the same conditions except that a
different antisense primer was used (PCA3, 5¢-AGCAC
CTCCTCCATGCTGC-3¢) with an annealing temperature
Ó FEBS 2003 fgl2 expression regulation in endothelial cells (Eur. J. Biochem. 270) 2275
of 57 °C. The PCR product was recovered and digested

with restriction endonucleases SacII and EcoRI. The
fragment was subcloned into the pBluescript plasmid
(Strategen). The transformants were subjected to DNA
sequence analysis.
RT-PCR detection of fgl2
Expression of fgl2 mRNA was assessed using RT-PCR.
Total RNA from BALB/cJ macrophages, macrophages
infected with MHV-3, SVE-10 murine endothelial cells, and
murine small intestine RNA (Clontech, Palo Alto, CA) were
used to synthesize the first strain cDNA. PCR was then
performed in 50-lL reactions using 1-lL portions of cDNA
and the primers fgl2318 (5¢-TGCCCACGCTGACCATC
CA-3¢) corresponding to nucleotides 318–336 of BALB/cJ
fgl2 cDNA (M15761) and fgl21224 (5¢-GAGACAAC
GATCGGTACCCCT-3¢) corresponding to nucleotides
1224–1244 of BALB/cJ fgl2 cDNA. (M16238), which
yield a 906-base pair band in 1% agarose DNA gel after
30 cycles of PCR reaction. Amplification products were not
obtained when reverse transcriptase was omitted (data not
shown). RT-PCR for b-actin was set upas an internal control
to ensure equal loading and first strand synthesis with
forward primer, 5¢-ATGTTTGAGACCTTCAACAC-3¢,
and reverse primer, 5¢-CACGTCACACTTCATGAT
GGA-3¢.
DNA constructs
AsingleSalI site was introduced into pM166 after the
initiating ATG by site directed mutagenesis. The restriction
fragment extending 3.5 kb upstream from the introduced
SalI site was excised by digestion with SalIandSmaI
(utilizing the SmaI contained in the pBluescript vector) and

subcloned into the luciferase reporter vector pGL2-Basic
using the SmaIandXhoI sites in this vector to construct the
plasmid pfgl2()3500/+9)Luc. A 1.3-kb fragment was also
released from pM166 by digestion with SalIandEcoRV
and subcloned into pGL2-Basic using the SmaIandXhoI
sites in this vector to construct a shorter reporter
pfgl2()1320/+9)Luc.
5¢-Truncation of the fgl2 promoter in pGL2-Basic was
created by PCR. 5¢-deleted promoter fragments were
amplified from a pM166 template using a common
antisense primer (5¢-GCCACAACCAACCAGGAAG-3¢,
positions 1335–1353 in GenBank Accession AF025817) and
a series of sense primers at varying distances upstream.
Upstream sense primers were 5¢-TCTTGGGAAATCTGG
TTAGAG-3¢ for fragment pfgl2()985/+9), 5¢-GGTCAGT
ATGCACAAGTGAG-3¢ for fragment pfgl2()723/+9),
5¢-GAGCTGAGTGATGGGGAAGGA-3¢forpfgl2()681/
+9), 5¢-CCACTGACGATTACATAGCC-3¢ for fragment
pfgl2()612/+9), 5¢-GGACCTTTGTTCTGATTAGGG
GC-3¢ for fragment pfgl2()498/+9), 5¢-CGCAGACATT
TAGACGTTCC-3¢ for pfgl2()360/+9), 5¢-GGGCACTG
GTATTACAACTGT-3¢ for pfgl2()294/+9), 5¢-CTCCT
CCTGTGTGGCGTCTGA-3¢ for fragment pfgl2()119/
+9), and 5¢-GAACGCCTGAGTCAGGCGGCGG-3¢
for fragment pfgl2()58/+9). The PCR products represent-
ing these different promoter fragments were subcloned into
the plasmid PCR2.1 (Invitrogen). These fragments were
then excised from PCR2.1 by digestion with HindIII and
SalI and transferred into the pGL2-Basic utilizing the
HindIII and XhoIsites.A3¢-deleted promoter fragment was

amplified from a pM166 template using a common sense
primer 5¢-GAATAAGGAGGGCAGGGTGAA-3¢ (posi-
tions )1320 to )1302 in GenBank accession no. AF025817),
and the antisense primers 5¢-TAGTGGGGAAAGA
GTTGGAACG-3¢ for fragment )1320/)274. The PCR
products representing this promoter fragment was sub-
cloned into PCR2.1 and transferred into pGL2-Basic using
the KpnIandXhoI sites in the vector and KpnIandSalIsites
in the PCR2.1 clones of the promoter fragments. All
promoter constructs were sequenced to confirm their
orientation and to verify their sequences.
pCR3.1A59 MHV nucleocapsid protein (N protein)
expression vector was reported previously [33]. The N gene
fragments were subcloned into the expression vector
pCR3.1 (Invitrogen), under the control of the cytomegalo-
virus promoter and bovine growth hormone 3¢-processing
signals.
Drosophila Eukaryotic Expression Constructs Expression
cassettes for Sp1, Sp3 variants, and Ets-1 were based upon
pPacUO, a transient episomal vector which contains the
2.6-kb Drosophila actin 5C promoter, a 0.7-kb 5¢-UTR
Ultrabithorax (Ubx) internal ribosome entry site, the first
eight codons of the Ubx open reading frame and 1.1 kb of
3¢-UTR from the actin 5c gene. These constructs have been
reported previously [30].
Linker scanning analysis
Seven site-directed mutants were created within a 70-bp
region ()119 to )41) of the fgl2 promoter in the
plasmid pfgl2()1320/+9)Luc using primers listed in
Table 1. Each mutant contains the sequence GGTACC,

a KpnI restriction site. These mutations were introduced
using paired primers according to the manufacturer’s
instructions (Invitrogen). All mutants were verified by
sequencing.
Transient transfection and luciferase assay
Transient transfections were carried out using Lipofect-
amine according to the manufacturer’s instruction (Cana-
dian Life Technology, Burlington, Canada). Cells were
plated at 5 · 10
5
per well in six-well plates 18 h before
transfection. Transfection conditions were optimized using
the SV40 promoter/enhancer luciferase control plasmid,
pGL2-Control, to confirm that increasing amounts of
templates resulted in proportional increases in reporter
activity. Endothelial cells were cotransfected with 1 lgof
reporter construct and 0.5 lgofpRSV-b-gal DNA to
control for transfection efficiency as previously described
[30]. For Drosophila Schneider studies, cells were cotrans-
fected with 1 lg of experimental luciferase construct, the
indicated amount of expression plasmids, and 0.5 lgpRSV-
b-gal. The total amount of DNA transfected was kept
constant (2 lg) with the addition of pPacUO. Each
transfection experiment was performed in triplicate and
repeated a minimum of three times. Luciferase and
b-galactosidase assays were carried out 48 h post transfec-
tion as described previously [30].
2276 M. Liu et al.(Eur. J. Biochem. 270) Ó FEBS 2003
Electrophoretic mobility shift assays (EMSA)
Nuclear extracts from endothelial SVE-10 cells were collec-

ted as described [34]. For EMSA, double stranded oligo-
nucleotide probes (Table 2) were 5¢-end-labelled with
[c-P
32
]ATP (Amersham) using T4 polynucleotide kinase.
ForeachEMSAreaction2–5lg of nuclear extracts were
incubatedfor15minonicein20lL of binding buffer.
1 · 10
4
d.p.m. of probe was added to each reaction and the
mixtures were incubated at room temperature for 30 min.
For supershifts, 2 lL of antibodies were incubated with
nuclear extracts for 45 min before adding DNA probe.
Anti-Sp1, Sp3, Stat3, Oct-1, Oct-2, Ets1/2, and PU.1 Ig are
from Santa Cruz Biotechnology Inc. (Santa Cruz, CA,
USA). For competition, 100 · cold oligo were added to the
reaction. Consensus double stranded oligos Oct-1, 5¢-TGT
CGAATGCAAATCACTAGAA-3¢; Sp1, 5¢-ATTCGATC
GGGGCGGGGCGAGC-3¢; Ets/Pea3, 5¢-GATCTCGAG
CAGGAAGTTCGA-3¢; Ets (PU.1), 5¢-GGGCTGCTTG
AGGAAGTATAAGAAT-3¢;Stat3,5¢-GATCCTTCTG
GGAATTCCTAGATC-3¢; and C/EBP, 5¢-TGCAGATT
GGGCAATCTGCA-3¢ are from Santa Cruz Biotechno-
logy. The binding reactions were size-fractionated on a
nondenaturing, 5% acrylamide gel, run at 150 V at room
temperature for 2 h in 1 · Tris/borate/EDTA buffer.
Results
Mapping the transcription start site of
fgl2
gene

To map the transcription start site of the fgl2 gene and
exclude any 5¢-mRNA sequence diversity we employed
5¢-RACE. Clones representing the 5¢-end of the fgl2
mRNAs were isolated by 5¢-RACE as described in Mate-
rials and methods. The results are summarized in Fig. 1A.
Twenty-six of the 33 cDNA clones initiated with an A at
position 1320, 34 nucleotides downstream of a putative
TATTAAA box, which appears to represent the major
transcription initiation site. For the remainder of this report
we will refer to the transcriptional start site at nucleotide
1320 in AF025817 as position +1. Although there is a
second potential upstream TATA box at position )314 with
respect to transcription initiation, the usage of this TATA
box was not observed in our studies. Determination of the
5¢-end of mRNAs isolated from liver tissue and cloned by
5¢-RACE showed similar results (data not shown). Thus, we
observed no differences in the fgl2 transcription start site
using both in vivo and in vitro RNA sources. To complement
these findings, we performed primer extension analysis. As
shown in Fig. 1B, compared to uninfected macrophages,
major fgl2 transcripts only appeared in mRNA from
Table 2. Oligonucleotides used in EMSA analysis. Numbers after fgl2 in the name column are the 5¢-ending and 3¢-ending nucleotide location of
each oligo used. Underlined are the potential cis-DNA elements identified. m after fgl2 stands for mutant. Mutated sequences are the same as the
mutated sequences in Table 1 used in the luciferase assay. Ets core GGAA is italicized to indicate its overlapping with consensus Stat3 binding site
underlined.
Name Actual mouse fgl2 sequences (5¢ to 3¢-sense strand)
fgl2 )82/)55
GCG CCC GCC CTT TTC TGG GAA CTC AGA A
fgl2 )98/)69 AGA CTG TGA TGC AAA TGC GCC CGC CCT TTT
fgl2 )57/)32 AGA ACG CCT GAG TCA G GCG GCG GTG GC

fgl2 )85/)66 AAT GCG CCC GCC CTT TTC TG
fgl2 m)85/)66 AAT GCG CCA GGT ACC TTC TG
fgl2 )97/)77 GAC TGT GAT GCA AAT GCG CCC
fgl2 m)97/)77 GAC TGT GAT GCG GTA CCT CCC
fgl2 )76/)56 GCC CTT TTC TGG GAA CTC AGA
fgl2 m)76/)56 GCC CTT TTG AGG TAC CTC AGA
Table 1. Primer pairs used to construct linker-scanning fgl2 promoter mutants. Underlined are actually mutated sequences.
Name Primer sequences
Mu-S)108/)99
CTC CTC CTG TAA GGT ACC ACA GAC TGT GAT GC
Mu-AS)108/)99 GCA TCA CAG TCT GTG GTA CCT TAC AGG AGG AG
Mu-S)97/)91 GTG GCG TCT GAG GTA CCA ATG CAA ATG CGC
Mu-AS)97/)91 GCG CAT TTG CAT TGG TAC CTC AGA CGC CAC
Mu-S)87/)80 GAG ACT GTG ATG CGG TAC CTC CCG CCC TTT C
Mu-AS)87/)80 GAA AAG GGC GGG AGG TAC CGC ATC ACA GTC TC
Mu-S)77/)71 GCA AAT GCG CCA GGT ACC TTC TGG GAA CTC
Mu-AS)77/)71 GAG TTC CCA GAA GGT ACC TGG CGC ATT TGC
Mu-S)68/)59 CCG CCC TTT TGA GGT ACC AGA GAA CGC CTG
Mu-AS)68/)59 CAG GCG TTC TCT GGT ACC TCA AAA GGG CGG
Mu-S)57/)50 CTG GGA ACT CAT GGT ACC ACA GTC AGG CGG
Mu-AS)57/)50 CCG CCT GAC TGT GGT ACC ATG AGT TCC CAG
Mu-S)50/)44 CTC AGA ACG CCA GGT ACC CGC GGC GGT GGC
Mu-AS)50/)44 GCC ACC GCC GCG GGT ACC TGG CGT TCT GAG
Ó FEBS 2003 fgl2 expression regulation in endothelial cells (Eur. J. Biochem. 270) 2277
MHV-3 infected macrophages and transcription was shown
to initiate at a site consistent with our 5¢-RACE results
(lane T, corresponding to a nucleotide A at that position).
Mapping the region essential for fgl2 transcription
in endothelial cells
Recently, BALB/cJ mouse endothelial cells have been

shown to express fgl2 by in situ hybridization [26]. Using
RT-PCR, we confirmed that the mouse endothelial cell line,
SVE-10, constitutively expresses fgl2 (Fig. 2A). To identify
functionally important DNA regions necessary for the basal
level of fgl2 transcription in endothelial cells, we constructed
a series of plasmids containing a luciferase reporter under
the control of successively deleted fgl2 promoters (Fig. 2B).
These constructs contain fgl2 promoter all the way to the
beginning of ATG start codon to avoid the discrepancy of
minor transcription start variance between macrophages
and endothelial cells. Endothelial cells transiently transfected
withthelongestfragmentofthe5¢-fgl2 flanking region
()3500/+9) resulted in luciferase activity that averaged
10% of that observed in cells transfected with constructs
under the control of the strong-heterologous viral SV40
promoter/enhancer (not shown). As shown in Fig. 2, serial
5¢-deletion with regions spanning )3.5 bp to )119 bp
resulted in only minor changes in fgl2 promoter activity,
implying that elements in this region are not important for
constitutive activity of the fgl2 promoter. However, further
deleting regions from )119 bp to )58 bp significantly
reduced fgl2 promoter activity to approximately 20%
of that observed with the longer constructs. To further
confirm the 5¢-truncation results, we assessed the activity of
pfgl2()1320/)274)LUC relative to pfgl2()1320/+9)LUC
in SVE-10 cells. Results indicated that deletion of regions
spanning )274/+9 abrogated fgl2 promoter activity
(Fig. 2D), suggesting that this region is necessary for fgl2
transcription. A detailed sequence inspection of the region
spanning )120 nucleotide to )40 nucleotide indicated many

potential cis-acting elements, including Sp1, Ets, Stat3,
Octamer [35] binding sites and AP1 sites (Fig. 3) and these
motifs might be critical for fgl2 expression. Together, the
5¢-deletion and 3¢-deletion analyses indicate that sequences
from )119 through +9 contain sequences that are sufficient
for constitutive fgl2 transcription.
Linker-scanning and EMSA analyses
To gain further insight into the regulation of the fgl2
promoter, the region between )119 to )41 bp, just
upstream of the putative TATA box (Fig. 4A), was
subjected to systematic site-directed mutagenesis [30]. Seven
constructs were created in pfgl2()1320/+9)LUC so that
each construct contained an 8–10 bp mutation (Table 1).
These constructs were transfected into SVE-10 cells and
luciferase activity was assessed. Mutation of position )108
to )99 and )97 to )91 had no significant impact on fgl2
promoter activity. Mutation of positions )87 bp to )80
(containing an Octamer motif 5¢-ATGCAAAT-3¢)
Fig. 1. Mapping of the transcription start site
of fgl2 mRNA. (A) 5¢-RACE of fgl2 mRNA
from macrophages. Stars indicate the 5¢-ter-
minus of 33 clones isolated after 5¢-RACE. (B)
Primer extension analysis of fgl2 mRNA. The
left side of the figure is the sequence ladder
using the same primer and the pM166 clone as
template. Arrow indicates the nucleotide that
is matched to the band present in the primer
extension reaction on the right. The sequence
of the primer is 5¢-CCTCCACCGCTCGG
CAGGCAGCGAGGACGG-3¢. The lanes

on the right correspond to primer extension
reactions performed with RNA extracted
from uninfected macrophages (M); MHV-3
infected macrophages (M + MHV-3).
2278 M. Liu et al.(Eur. J. Biochem. 270) Ó FEBS 2003
significantly reduced promoter activity. The mutation at
positions )77 to )71 (containing a Sp1 binding site:
5¢-GCCCGCCC-3¢) reduced the promoter activity to 20%
of the parental wild-type promoter. The mutations at
positions )68 to )59 (containing potential Ets family
binding site, and a Stat3-like motif: 5¢-TTCTGGGAACT-3¢)
and from )57 to )49 (which is overlapping the Ets/Stat3
binding sites) similarly decreased promoter activity to 40%
and 50% of that observed with the wild-type promoter.
However, mutating the AP1 site (from )50 to )44:
5¢-TGAGTCAG-3¢) did not have any impact on fgl2
promoter activity in vascular endothelial cells (Fig. 4B).
Thus, these results suggest that a 39-bp positive regulatory
region spanning nucleotides )87 to )49 contains cis-acting
elements necessary for fgl2 constitutive promoter activity in
SVE-10 endothelial cells.
EMSAs were used to examine the nature of the
functionally important nucleoprotein complexes that form
at cis-regulating elements important for constitutive fgl2
promoter activity in SVE-10 endothelial cells. Using
EMSA probes spanning fgl2–82/)55 and )98/)69 we
detected protein–DNA complexes with endothelial cell
nuclear extracts (not shown). These DNA-protein com-
plexes were specific as they were eliminated by competing
unlabelled oligonucleotide containing these sequences but

not with mutant oligonucleotides (fgl2 m-97/)77, m-85/
)66, and m-76/)56, Table 2). As both probes contain the
Sp1 motif, we predict that DNA-protein complexes could
be formed with Sp1 and Sp3 transcription factors as seen
Fig. 2. Functional analysis of fgl2 promoter. (A) Analysis of fgl2 RNA by semiquantitative RT-PCR. RT-PCR was performed with 4 lgofeach
RNA sample. M, macrophages; M + MHV-3, macrophages infected with MHV-3; endothelial cells, RNA from murine endothelial SVE-10 cells;
intestine, total mouse intestinal RNA. (B) Fgl2 promoter-luciferase constructs. The numbers give the 5¢-ending and 3¢-ending nucleotide of each
construct. (C) Relative luciferase activity of 5¢-truncations of fgl2 promoter constructs in endothelial cells. Activity of pGl2()1320/+9)Luc is set at
100%. (D) Luciferase activity of 3¢-truncation of fgl2 promoter construct in endothelial cells. The data is the average of at least three separate
experiments each performed in triplicate. The data was corrected for variations in transfection efficiency by normalization to the activity of a
cotransfected plasmid expressing b-galactosidase.
Fig. 3. Schematic drawing representing the proximal promoter of the
mouse fgl2 gene. Positive regulatory region (PRD) and functional cis-
regulatory DNA elements are shown. Identified cis elements are
underlined. Boxes indicate mutated regions. Probes used in EMSA are
also shown as underline with specific numbers.
Ó FEBS 2003 fgl2 expression regulation in endothelial cells (Eur. J. Biochem. 270) 2279
in the case of eNOS in vascular endothelial cells [30]. To
test this hypothesis, we performed an EMSA experiment
with a probe containing the fgl2 Sp1 binding site (probe
)85/)66, Table 2 and Fig. 3). As shown in Fig. 5, at least
four nucleoprotein complexes were seen using endothelial
cell nuclear extracts (lane 2). These complexes were
competed away by the same unlabelled probe (lane 3)
but not by an unrelated nonspecific probe (lane 4). The
addition of anti-Sp1 IgG
1
shifted the slowest migrating
complex to a higher position in the gel (lane 5) while anti-
Sp3 antibody detected two forms of Sp3 complexes, the

second nucleoprotein complex (lane 6) and an additional
Sp3 complex, which is in close proximity to the slowest
moving Sp1 and did not clearly separate in the gel
utilized. The same Sp1/Sp3 pattern has been demonstrated
for the eNOS promoter in endothelial cells [30]. Consis-
tent with our linker-scanning analysis, which suggested
that AP1 is not functioning in the constitutive expression
of fgl2, no DNA–protein complex was observed with a
probe )57/)32 (Fig. 3) which contains the putative AP1
binding site (data not shown).
To test whether the Octamer motif, which corresponds to
the mutant 3 (Fig. 4), is functional in the fgl2 promoter, a
cross competition EMSA was performed. As seen in
Fig. 6A, protein–DNA complexes formed by consensus
Oct-1 motif were specifically blocked by both excess
unlabelled consensus Oct-1 oligo or fgl2 oligo)97/)77 as
shown in Table 2 and Fig. 3, which contains an identical
Oct-1 motif ATGCAAAT (lanes 3 and 4). Competition was
not observed with an excess of unlabelled mutant fgl2–97/
)77 oligonucleotide (Table 2). Consistent with this result,
unlabelled consensus Oct-1 oligonucleotide, or fgl2 probe
()97/)77), were able to inhibit formation of the nucleopro-
tein A formed by labelled fgl2–97/)77 probe, but not by the
mutated fgl2–97/)77 oligo (lane 8–10). These results suggest
that complex A is formed by DNA binding proteins
functionally related to members of the Octamer family. In
addition, similar to that of the consensus Oct-1 probe, the
addition of an antiOct-1 antibody significantly reduced the
intensity of complex A of probe )97/)77 (Fig. 6B, lane 5
and 7). Thus, we infer that complex A represents the

interaction of Oct-1 with the fgl2 Octamer DNA motif. The
Fig. 4. Linker-scanning analysis of important cis elements in fgl2 pro-
moter. (A) Linker scanning mutations of fgl2 promoter. All mutant
constructs were made using pfgl2()1320/+9)Luc as template. The
numbered boxes indicate the specific nucleotides mutated which were
also shown in Table 1 and Fig. 5. (B) Luciferase activity of wild-type
and mutated fgl2 promoter in SVE-10 endothelial cells. Luciferase
activity was analyzed as described in Fig. 2.
Fig. 5. EMSA analysis of cis elements overlapping the 38 bp identified
in Fig. 4 of the fgl2 promoter. Nuclear extracts (NE) from SVE-10
endothelial cells were incubated with probes shown in Table 2 and
Fig. 5. Arrows indicate the bands that are interacting with these
probes. All the lanes have labelled probes. The adding of extracts and
cold probe are indicated in the top of the panel. Supershift, ss-Sp1
and ss-Sp3, were shown in presence of anti-Sp1 and anti-Sp3 Ig. Cold
and mutated competitors are indicated at the top of the figure.
2280 M. Liu et al.(Eur. J. Biochem. 270) Ó FEBS 2003
nucleocomplex B and C in panel A could represent binding
of the Sp1 protein family, given that the probe contains a
half site of a Sp1 motif (Fig. 3). Consistent with this view, a
consensus Sp1 oligonucleotide blocked the formation of
these complexes (Fig. 6A, lane 12). To assess this possibility,
we performed a supershift assay using antibodies against
Sp1 and Sp3. We did not see a supershift or blocking of the
complexes B and C (data not shown). Thus, it can be
concluded that these complexes are not formed by the
binding of Sp1 or Sp3. However, we cannot rule out a Sp1
related transcription factor or an unknown factor at this
time. Future studies are required to address this issue.
We next examined whether the motifs located between

)68/)57 (Table 2 and Fig. 3), corresponding to mutant 5
and 6, are functional. As shown in Fig. 7A, two DNA-
protein complexes were detected. Both complexes were
competed away by a cold fgl2 probe (lane 3) but only
partially affected by a mutant fgl2 probe (Table 2) (lane 4).
The slower moving complex was blocked by the addition of
two Ets consensus oligonucleotides (Ets1/Pea3 and PU.1,
lanes 5 and 6). We observed that formation of the faster
moving complex was blocked by a Stat3 consensus oligo-
nucleotide (lane 7). Unlabelled fgl2 probe )76/)57, but not
a mutant fgl2–68/)57 probe, blocked the binding to a
labelled Stat3 consensus oligonucleotide and provided
further evidence for the involvement of Stat3 cis DNA
elements in fgl2 expression (Fig. 7B). Importantly, we were
unable to detect either a supershift or shift abrogation using
antibodies directed against Ets (Ets1/Pea3, PU.1, data not
shown). This was not unexpected as it has been reported
previously that Ets antibodies rarely work in EMSA [30].
We were also unable to detect Stat3 supershift with
antibody against Stat3 (data not shown). An EMSA probe
spanning )57/)32 (Table 2 and Fig. 3) did not demonstrate
nucleoprotein complex formation in EMSA (data not
shown). This may suggest that the effect of mutant 6 in
Fig. 4B could represent the edge effect of mutation in this
area that affected the binding of Ets/Stat3 to its motif.
Ets-1 is critical in regulating fgl2 expression
and requires the presence of both Sp1 and Sp3
Collectively the linker-scanning and EMSA analysis suggest
that Oct 1, Sp1/Sp3, Ets family members, and a Stat3-like
protein may form a multicomponent nucleoprotein com-

plex upon the 39 bp PRD region and functionally contri-
bute to the constitutive expression of fgl2, especially in
vascular endothelial cells. To further evaluate this hypo-
thesis, we employed the Drosophila Schneider cell line
(SL2), which is deficient in constitutive Sp1, Sp3, and Ets-1
transcription factors [36,37]. As shown in Fig. 8A, cotrans-
fection of the pfgl2LUC()1320/+9) promoter, along with
increasing amounts of Sp1 expression cassette resulted in a
concentration-dependent increase in functional fgl2 pro-
moter activity (22-fold maximum effect). Increasing
amounts of a Sp3 expression cassette also increased fgl2
promoter activity in a concentration-dependent fashion,
though to a lesser extent (Figs 8B, fivefold increase).
However, cotransfection of increasing amounts of an Ets-1
expression cassette alone had no important effect on fgl2
promoter activity (Fig. 8C). To examine whether Sp1, Sp3,
and Ets-1 exert an interactive effect on fgl2 promoter
activity, we cotransfected combinations of the varied
transcription factors. As shown in Fig. 9A, cotransfection
Fig. 6. EMSA analysis of the function of Octamer motif in the fgl2 promoter. Lanes 2–5 of panel A and lanes 2–5 of panel B, NE were incubated with
labelled consensus Oct-1. Lanes 6–12 of panel A and lanes 6–8 of panel B, NE were incubated with
32
P-labelled fgl2 fragment )97/)77. All lanes
contained hot probes. Addition of NE, cold probe competitors, and antibodies are indicated at the top of the panel. (A) Cross competition with
consensus Oct-1 and fgl2–97/)77. (B) Antibody analysis of Oct-1 binding in the presence of antiOct-1 and antiOct-2 antibody. Arrows indicate the
specific Oct-1 DNA-protein complexes. All probes are shown in Table 2 and Fig. 5.
Ó FEBS 2003 fgl2 expression regulation in endothelial cells (Eur. J. Biochem. 270) 2281
of increasing amounts of Sp3 along with 50 ng of Sp1 (half-
maximum amount) demonstrated increased activity of fgl2
promoter compared to Sp1 or Sp3 alone (Fig. 9A). In

contrast, increasing amounts of Ets-1 along with 50 ng of
Sp1 or a threshold amount of Sp1 (10 ng) did not increase
fgl2 promoter activity (data not shown). To examine the
possibility that Ets-1 functions with Sp3, we performed a
cotransfection assay using Ets-1 and Sp3 expression vectors
in combination with an fgl2 promoter construct in SL2
cells. Similar to that of Sp1, Ets-1 had no incremental effect
on fgl2 promoter activity when used with either 50 ng or a
threshold amount of Sp3 (10 ng) (Fig. 9A). However, when
both threshold amounts (10 ng) of Sp1 and Sp3 were added
together, with increasing quantities of Ets-1, a cooperative
positive functional interaction was evident (Fig. 9B). To
define the fgl2 cis-element implicated in the responses of the
fgl2 promoter to these transcription factors, mutations
corresponding to mutant 4 ()77/)71) and 5 ()68/)59) were
employed. As shown in Fig. 9B, mutation of the Sp1/Sp3
binding site (mut4) deceased luciferase activity of the Sp1/
Sp3 interaction or the Sp1/Sp3/Ets-1 interaction by
approximately 90% and 80%, respectively. Mutation of
the Ets-1 binding site (mut5) also blocked the effect of
Ets-1 in the presence of Sp1/Sp3.
Effect of the fgl2 positive regulatory region
on induced expression of fgl2
Previously, our laboratory has provided evidence that fgl2
expression is responsible for the pathogenesis of MHV-3
induced fulminant hepatitis through its procoagulant
activity [14,26]. We have also recently demonstrated that
Fig. 7. EMSA analysis of the function of Ets/Stat3 cis elements. (A) Cross competition of consensus Ets/PEA3, Ets/PU1, and Stat3 cis elements
with corresponding transcription factors. NEs from endothelial cells were incubated with labelled fgl2 probe )76/)57 (Table 2 and Fig. 5). The
adding of NE and cold probes are indicated at the top of the panel. Arrows indicate specific Ets, Stat3 DNA-protein complexes. (B) Cross

competition of Stat3 DNA cis element with its binding protein. Consensus Stat3 was labelled and competes with cold fgl2 oligonucleotide
)76/)57 (Fig. 5).
Fig. 8. Sp1, Sp3, and Ets-1 transactivate mouse fgl2 promoter/reporter
luciferase constructs in Drosophila Schneider cells. Assay of pro-
moter activity of pfgl2LUC()1320/+9) promoter/reporter luciferase
construct cotransfection with increasing amounts of pPacUSp1
(10–250 ng, panel A), pPacUSp3 (10–250 ng, panel B),and pPacUEts-1
(25–250 ng, panel C). Luciferase activity was analyzed as described in
Fig. 2.
2282 M. Liu et al.(Eur. J. Biochem. 270) Ó FEBS 2003
the nucleocapsid (N) protein of MHV-3 can induce
transcription of fgl2 and implicated regions of the fgl2
promoter located )372 to )306 bp upstream of the
transcription initiation site [33]. Therefore, we undertook
studies to determine the contribution of the PRD at
nucleotide )87 to nucleotide )49 that is implicated in the
constitutive expression of fgl2 to the induced transcription of
fgl2 in response to MHV-3. We examined whether a
mutation of the PRD would abolish the induction of fgl2
by MHV-3 nucleocapsid protein. As shown in Fig. 10, we
cotransfected plasmid pfgl2()1320/+9) with a constitu-
tively active nucleocapsid protein expression plasmid
(pCR3.1A59) into endothelial cells. Co-transfection of the
N gene expressed by pCR3.1A59 induced a 5.5-fold activa-
tion of the fgl2 promoter relative to that observed with
cotransfection with empty vector (pCR3.1). Mutation of the
Sp1/Sp3 (mut4) site completely abolished this activation.
Discussion
Evidence indicates that inducible expression of fgl2 correlates
with the vascular thrombosis of the liver seen in fulminant

Fig. 10. The effect of PRD on nucleocapsid protein induced fgl2
expression. Wild-type pfgl2()1320/+9)Luc (which was set as 1) and
mutated fgl2 promoter-luciferase construct pfgl2(mut4)Luc were sep-
arately cotransfected with pCR3.1A59 into SVE-10 endothelial cells.
The results were expressed as fold increase compared to cotransfection
with pCR3.1 empty vector.
Fig. 9. Cooperative activity of Sp1, Sp3, and Ets-1 on fgl2 promoter/reporter luciferase constructs in Drosophila Schneider cells. (A) Assay of fgl2
promoter activity upon cotransfection increasing amount of Sp3 with half-maximal amounts of pPacUSp1 (50 ng, left panel) or increasing amount
of Ets-1 with 50 ng of either Sp1 or Sp3. (B) Assay of fgl2 promoter activity upon cotransfection of thresholds amount of both Sp1 and Sp3 (10 ng,
right panel) and assay of promoter activity of wild-type (pfgl2 ()1320/+9) and linker-scanning mutant 4 and 5 following cotransfection with 10 ng
of pPacUSp1, pPacUSp3, and pPacUEts-1. Luciferase activity was analyzed as described in Fig. 2.
Ó FEBS 2003 fgl2 expression regulation in endothelial cells (Eur. J. Biochem. 270) 2283
viral hepatitis [14,21,26]. The function of constitutively
expressed fgl2 at low levels in cells or tissues including
cytotoxic T cells, trophoblasts, intestinal epithelial cells and
endothelial cells remains to be defined. We hypothesize that a
variety of tissues may express basal levels of a TF/factor VII-
independent procoagulant. Indeed, constitutive expression
of TF in some cells and tissues is known to play this role, such
as in the CNS [9]. The biological importance of constitutive
fgl2 expression is exemplified by the observation that the
fgl2
–/–
mouse has a developmental defect with measured fetal
loss of fgl2
–/–
embryos due to hemorrhage at the fetomater-
nal interface (manuscript in preparation). Constitutive fgl2
promoter activity was also seen in a small percentage of the
primary endothelial cells isolated from these fgl2

–/–
mice, in
which the fgl2 coding region is replaced with b-galactosidase
(unpublished data). The constitutively expressed fgl2 might
reflect functional contributions of fgl2 as a matrix/adhesion
protein, possibly necessary for development of a normal
fetus.
In this study we have functionally characterized the
5¢-flanking region of the mouse fgl2 gene. We mapped the
5¢-terminal region of the fgl2 mRNA using primer extension
analysis and 5¢-RACE. Sequence analysis revealed a TATA
box (TATTAAA) located 33 bp upstream of the major
transcription start site. The results of transient transfection
with a 3¢-deletion construct that removed this major
transcription start site and TATA box ()1320/)274)
provided functional evidence for this conclusion.
When we performed functional analysis of the 5¢-flanking
region of the fgl2 gene, we demonstrated in murine
endothelial cells that a region 119 bp upstream from the
transcriptional start site and 9 bp downstream are sufficient
for the basal expression of the fgl2 gene. This finding is
consistent with the observation that fgl2 is constitutively
expressed in cultured endothelial cells, as well as in the
primary endothelial cells in low level. Although the maximal
activity we observed with the fgl2 promoter is approxi-
mately 10% of the strong SV-40 promoter/enhancer, this
is not an unusual finding. For example, the promoter of the
constitutively expressed human endothelial NOS promoter
evidences approximately 10% of the activity of the SV-40
promoter/enhancer [30].

A more comprehensive mutational analysis of this fgl2
promoter segment revealed tightly clustered cis-regulatory
DNA elements spanning 39 bp in the proximal region of the
fgl2 promoter. As shown in Fig. 4, mutating nucleotides
from )87 to )80 (containing Oct-1), )77 to )71 (containing
Sp1), )68 to )57 (containing Ets/Stat3) all significantly
reduced fgl2 promoter activity. The linker-scanning mutant
)77 to )71, which contained the Sp1 binding site, had the
greatest effect on fgl2 promoter activity, reducing it to the
level observed with the deletion construct pfgl2()51/
+9)LUC. As mutation of any of the above fragments
significantly affected basal expression of fgl2, it is possible
that the nucleoprotein complexes that form upon this PRD
cis region interact with the core transcription machinery
that forms upon the TATA box to direct basal expression of
fgl2 [38]. The presence of a multicomponent PRD in the fgl2
proximal promoter suggests the complexity of fgl2 as a
proinflammatory gene.
As well as the functional Sp1/Sp3 binding site, we
identified that the octamer site (ATGCAAAT) in the fgl2
promoter is functional and interacts with the Oct-1
transcription factor (Fig. 6). It has been shown that POU
homeodomain family proteins bind to this consensus motif
and regulate gene transcription in diverse cells through
interaction with a variety of partner DNA binding proteins,
including Sp1 family members and Ets family members, as
observed in the current studies [35,39]. In the endothelial
cell, Oct-1 has been demonstrated to positively or negatively
regulate a variety of genes. For example, the expression of
endothelial cell-specific TIE2 gene is dependent on Oct-1

and an endothelial cell-restricted cofactor [40]. We also
identified that the Ets/Stat3 motifs are functional and
contribute to fgl2 constitutive expression in vascular endo-
thelial cells (Fig. 7). Ets proteins have been shown to be
promiscuous in their binding to the same GGAA core
sequence [41]. Consistent with reports from others, we did
not detect supershifts with antiEts1/2 and antiPU.1 anti-
body. However, we did demonstrate that both Ets1/Pea3
consensus and PU.1 consensus oligonucleotides were able to
inhibit the formation of nucleoprotein complexes with the
fgl2 GGAA core.
It is known that Ets family members generally serve as
coactivators for the transcriptional activator Sp1 and
synergistically regulate gene expression [42]. The presence
of an Ets site immediately proximal to the Sp1 site in the fgl2
promoter may explain the constitutive expression of fgl2.
With our EMSA results (Fig. 7) suggesting the involvement
of Ets family members in the regulation of fgl2 promoter,
we found that Ets-1, a major Ets family member in
endothelial cells, participated in activation of the fgl2
promoter. Our results suggested that Ets-1 can cooperate
with Sp1/Sp3 in a positive manner in Drosophila cells that
areknowntobedeficientinconstitutivelyexpressedEtsand
Sp1 protein family members (Figs 8 and 9). Removing any
one of these factors from the system or mutating any of
these two activator recognition sites found in the PRD
resulted in a marked decrease in functional promoter
activity.
Another motif, TTCTGGGAACT that overlaps the Ets
binding site, differed by one nucleotide from the consensus

Stat3 binding site (TTCTGGGAATT) [43] and showed
specific DNA–protein interactions in endothelial cells.
Although Stat3 is commonly implicated in activated gene
expression, constitutively activated Stat3 has been shown to
be present in many cell types, including endothelial cells [43].
Importantly, Stat3 is known to regulate constitutive gene
expression through cooperative interactions with other
transcriptional coactivators [43,44]. As we did not detect
supershifts with anti-Stat3 antibody, either with fgl2
sequence or a consensus Stat3 oligo, this issue will require
further study, such as by utilizing the dominant nega-
tive Stat3 expression system to test its function in fgl2
expression.
Among the identified DNA cis-elements, the Sp1 motif
plays an important role in the regulation of many consti-
tutively expressed genes [45]. Sp1 might be initially recruited
to the fgl2 promoter and then facilitate the further
recruitment and binding of other positive trans-regulators,
such as Oct-1, Ets-1, and Stat3-like protein (Fig. 5). In other
systems Sp1 has been shown to physically interact and
recruit trans-acting factors, such as AP1, GATA, NF-jB,
and Ets-1 [30]. For example, in the promoter of parathyroid
2284 M. Liu et al.(Eur. J. Biochem. 270) Ó FEBS 2003
hormone-related protein (PTHrP), Sp1 and Ets-1 proteins
cooperate to transactivate the expression of PTHrP [46].
Consistent with previous findings, our EMSA also detec-
ted two forms of Sp3 protein–DNA complexes with a fgl2
probe (Fig. 5). Sp3 is known to act either positively or
negatively to regulate gene expression [37,45] and to have
at least three human variants with molecular masses of

115, 80 and 78 kDa that are abundantly expressed in a
broad number of mammalian cells. In EMSA, the two
shorter forms comigrate as a fast protein–DNA complex,
while the longer form migrates as a slower protein–
DNA complex [47]. In addition, Sp1 has also been
established to be responsible for induced expression of
IL-10 [48]. Our data from endothelial cells and the
Drosophila system, supports the contention that both Sp1
and Sp3 positively affect fgl2 constitutive expression.
Importantly, the presence of both Sp1 and Sp3 is required
for Ets-1 function.
Taken together, our data suggest that DNA regions
spanning 39 bp, from )87 to )49 of the fgl2 promoter,
represent functional cis-DNA regulatory elements that
interact with Oct-1, Ets-1, and Sp1/Sp3. As fgl2 expression
is tightly controlled [49], the interaction of transcription
factors Sp1/Sp3 with the fgl2 promoter form a core to
recruit both Oct-1 and Ets-1 to form a multicomponent
PRD that is sufficient to confer basal level expression of fgl2
in endothelial cells. In addition, this PRD is required to
interact with inducible elements upon activation to enhance
fgl2 expression. This mechanism may explain how fgl2 can
be both constitutively transcribed and rapidly responsive to
proinflammatory stimuli. Although the observations pre-
sented in this study suggest that this PRD is necessary for
the inducible expression of fgl2, we plan to undertake
additional studies aimed at more precisely elucidating
specific mechanisms regulating induction of fgl2 promoter
activity in response to proinflammatory stimuli such as
IFN-c or other viral factors.

Acknowledgements
This work was supported in part by grant 13298, MOP 37780 and
MOP 37778 from the Canadian Institute of Health Research (CIHR)
and grant AI43368 from the National Institutes of Health.
References
1. Cosgriff, T.M. (1989) Viruses and hemostasis. Rev. Infect. Dis.
Suppl. 4, S692.
2. Halloran, P., Aprile, M., Haddad, G. & Robinette, M. (1982) The
significance of elevated procoagulant activity in the monocytes of
renal transplant recipients. Transplant. Proc. 14, 669–672.
3. Silverman, R., Cohen, Z., Levy, G., Craig, M., Cullen, J. &
Langer, B. (1987) Immune responses in small intestinal trans-
plantation in the rat: correlation of histopathology and monocyte
procoagulant activity. Surgery 102, 395–401.
4. Tipping, P.G., Lowe, M.G. & Holdsworth, S.R. (1988) Glo-
merular macrophages express augmented procoagulant activity in
experimental fibrin-related glomerulonephritis in rabbits. J. Clin.
Invest. 82, 1253–1259.
5. Cole, E.H., Schulman, J., Urowitz, M., Keystone, E., Williams, C.
& Levy, G.A. (1985) Monocyte procoagulant activity in glomer-
ulonephritis associated with systemic lupus erythematosus. J. Clin.
Invest. 75, 861–868.
6. Kucey, D.S., Cheung, P.Y. & Rotstein, O.D. (1992) Platelet-
activating factor modulates endotoxin-induced macrophage pro-
coagulant activity by a protein kinase C-dependent mechanism.
Infect. Immun. 60, 944–950.
7. Sullivan, B.J., Swallow, C.J., Girotti, M.J. & Rotstein, O.D. (1991)
Bacterial translocation induces procoagulant activity in tissue
macrophages. A potential mechanism for end-organ dysfunction.
Arch. Surg. 126, 586–590.

8. Levy, G.A., Leibowitz, J.L. & Edgington, T.S. (1981) Induction of
monocyte procoagulant activity by murine hepatitis virus type 3
parallels disease susceptibility in mice. J. Exp. Med. 154, 1150–
1163.
9. Konigsberg,W.,Kirchhofer,D.,Riederer,M.A.&Nemerson,Y.
(2001) The TF: VIIa complex: clinical significance, structure-
function relationships and its role in signaling and metastasis.
Thromb. Haemost. 86, 757–771.
10. Carmeliet,P.,Mackman,N.,Moons,L.,Luther,T.,VanGres-
sens, P.V.I., Demunck, H., Kasper, M., Breier, G., Evrard, P.,
Muller, M., Risau, W., Edgington, T. & Collen, D. (1996) Role of
tissue factor in embryonic blood vessel development. Nature 383,
73–75.
11. Huang, X., Molema, G., King, S., Watkins, L., Edgington, T.S. &
Thorpe, P.E. (1997) Tumor infarction in mice by antibody-direc-
ted targeting of tissue factor to tumor vasculature. Science 275,
547–550.
12. Ryu,J.,Pyo,H.,Jou,I.&Joe,E.(2000)ThrombininducesNO
release from cultured rat microglia via protein kinase C, mitogen-
activated protein kinase, and NF-kappa B. J. Biol. Chem. 275,
29955–29959.
13. Koyama, T., Hall, L.R., Haser, W.G., Tonegawa, S. & Saito, H.
(1987) Structure of a cytotoxic T-lymphocyte-specific gene shows a
strong homology to fibrinogen beta and gamma chains. Proc. Natl
Acad. Sci. USA 84, 1609–1613.
14. Parr, R.L., Fung, L., Reneker, J., Myers-Mason, N., Leibowitz,
J.L. & Levy, G. (1995) Association of mouse fibrinogen-like
protein with murine hepatitis virus-induced prothrombinase
activity. J. Virol. 69, 5033–5038.
15. Qureshi, S.T., Clermont, S., Leibowitz, J., Fung, L.S., Levy, G. &

Malo, D. (1995) Mouse hepatitis virus-3 induced prothrom-
binase (Fgl2) maps to proximal chromosome 5. Genomics 29,
307–309.
16. Chiquet-Ehrismann, R. (1995) Tenascins, a growing family of
extracellular matrix proteins. Experientia 51, 853–862.
17.Ruegg,C.E.,Gandolfi,A.J.,Nagle,R.B.,Krumdieck,C.L.&
Brendel, K. (1987) Preparation of positional renal slices for study
of cell-specific toxicity. J. Pharmacol. Meth. 17, 111–123.
18. Xu, X. & Doolittle, R.F. (1990) Presence of a vertebrate fibrino-
gen-like sequence in an echinoderm. Proc. Natl Acad. Sci. USA 87,
2097–2101.
19. Yuwaraj, S., Ding, J., Liu, M., Marsden, P.A. & Levy, G.A. (2001)
Genomic characterization, localization, and functional expression
of FGL2, the human gene encoding fibroleukin: a novel human
procoagulant. Genomics 71, 330–338.
20. Levy, G.A., Liu, M., Ding, J., Yuwaraj, S., Leibowitz, J., Mars-
den, P.A., Ning, Q., Kovalinka, A. & Phillips, M.J. (2000)
Molecular and functional analysis of the human prothrombinase
gene (HFGL2) and its role in viral hepatitis. Am.J.Pathol.156,
1217–1225.
21. Fingerote, R.J., Abecassis, M., Phillips, M.J., Rao, Y.S., Cole,
E.H.,Leibowitz,J.&Levy,G.A.(1996)Lossofresistanceto
murine hepatitis virus strain 3 infection after treatment with cor-
ticosteroids is associated with induction of macrophage pro-
coagulant activity. J. Virol. 70, 4275–4282.
22. Pope, M., Rotstein, O., Cole, E., Sinclair, S., Parr, R., Cruz, B.,
Fingerote, R., Chung, S., Gorczynski, R. & Fung, L. (1995)
Pattern of disease after murine hepatitis virus strain 3 infection
Ó FEBS 2003 fgl2 expression regulation in endothelial cells (Eur. J. Biochem. 270) 2285
correlates with macrophage activation and not viral replication.

J. Virol. 69, 5252–5260.
23. Fung, L.S., Neil, G., Leibowitz, J., Cole, E.H., Chung, S., Crow,
A. & Levy, G.A. (1991) Monoclonal antibody analysis of a unique
macrophage procoagulant activity induced by murine hepatitis
virus strain 3 infection. J. Biol. Chem. 266, 1789–1795.
24. Ruegg, C. & Pytela, R. (1995) Sequence of a human transcript
expressed in T-lymphocytes and encoding a fibrinogen-like
protein. Gene 160, 257–262.
25. Schwartz, B.S., Levy, G.A., Fair, D.S. & Edgington, T.S. (1982)
Murine lymphoid procoagulant activity induced by bacterial
lipopolysaccharide and immune complexes is a monocyte pro-
thrombinase. J. Exp. Med. 155, 1464–1479.
26. Ding, J.W., Ning, Q., Liu, M.F., Lai, A., Leibowitz, J., Peltekian,
K.M.,Cole,E.H.,Fung,L.S.,Holloway,C.,Marsden,P.A.,Yeger,
H., Phillips, M.J. & Levy, G.A. (1997) Fulminant hepatic failure in
murine hepatitis virus strain 3 infection: tissue-specific expression
of a novel fgl2 prothrombinase. J. Virol. 71, 9223–9230.
27. Clark, D.A., Ding, J.W., Chaouat, G., Coulam, C.B., August, C.
& Levy, G.A. (1999) The emerging role of immunoregulation of
fibrinogen-related procoagulant Fgl2 in the success or sponta-
neous abortion of early pregnancy in mice and humans. Am.
J. Reprod. Immunol. 42, 37–43.
28. Lafuse, W.P., Castle, L., Brown, D. & Zwilling, B.S. (1995) The
cytotoxic T lymphocyte gene FIBLP with homology to fibrinogen
beta and gamma subunits is also induced in mouse macrophages
by IFN-gamma. Cell Immunol. 163, 187–190.
29.Li,C.,Fung,L.S.,Chung,S.,Crow,A.,Myers-Mason,N.,
Phillips, M.J., Leibowitz, J.L., Cole, E., Ottaway, C.A. & Levy, G.
(1992) Monoclonal antiprothrombinase (3D4.3) prevents morta-
lity from murine hepatitis virus (MHV-3) infection. J. Exp. Med.

176, 689–697.
30. Karantzoulis-Fegaras, F., Antoniou, H., Lai, S.L., Kulkarni, G.,
D’Abreo, C., Wong, G.K., Miller, T.L., Chan, Y., Atkins, J.,
Wang, Y. & Marsden, P.A. (1999) Characterization of the human
endothelial nitric-oxide synthase promoter. J. Biol. Chem. 274,
3076–3093.
31. Marsden, P.A., Heng, H.H., Scherer, S.W., Stewart, R.J., Hall,
A.V., Shi, X.M., Tsui, L.C. & Schappert, K.T. (1993) Structure and
chromosomal localization of the human constitutive endothelial
nitric oxide synthase gene. J. Biol. Chem. 268, 17478–17488.
32. Hall, A.V., Antoniou, H., Wang, Y., Cheung, A.H., Arbus, A.M.,
Olson,S.L.,Lu,W.C.,Kau,C.L.&Marsden,P.A.(1994)
Structural organization of the human neuronal nitric oxide syn-
thasegene(NOS1).J. Biol. Chem. 269, 33082–33090.
33. Ning,Q.,Liu,M.,Kongkham,P.,Lai,M.M.,Marsden,P.A.,
Tseng, J., Pereira, B., Belyavskyi, M., Leibowitz, J., Phillips, M.J.
& Levy, G. (1999) The nucleocapsid protein of murine hepatitis
virus type 3 induces transcription of the novel fgl2 prothrombinase
gene. J. Biol. Chem. 274, 9930–9936.
34. Fan, J., Marshall, J.C., Jimenez, M., Shek, P.N., Zagorski, J. &
Rotstein, O.D. (1998) Hemorrhagic shock primes for increased
expression of cytokine-induced neutrophil chemoattractant in the
lung: role in pulmonary inflammation following lipopolysaccha-
ride. J. Immunol. 161, 440–447.
35. Clark, M.E. & Mellon, P.L. (1995) The POU homeodomain
transcription factor Oct-1 is essential for activity of the gonado-
tropin-releasing hormone neuron-specific enhancer. Mol. Cell
Biol. 15, 6169–6177.
36. Courey, A.J. & Tjian, R. (1988) Analysis of Sp1 in vivo reveals
multiple transcriptional domains, including a novel glutamine-rich

activation motif. Cell 55, 887–898.
37. Hagen, G., Muller, S., Beato, M. & Suske, G. (1994) Sp1-mediated
transcriptional activation is repressed by Sp3. EMBO J. 13, 3843–
3851.
38. Thanos, D. & Maniatis, T. (1995) Virus induction of human IFN
beta gene expression requires the assembly of an enhanceosome.
Cell 83, 1091–1100.
39. Belsham, D.D. & Mellon, P.L. (2000) Transcription factors Oct-1
and C/EBPbeta (CCAAT/enhancer-binding protein-beta) are
involved in the glutamate/nitric oxide/cyclic-guanosine 5¢-mono-
phosphate-mediated repression of mediated repression of gona-
dotropin-releasing hormone gene expression. Mol. Endocrinol. 14,
212–228.
40. Fadel, B.M., Boutet, S.C. & Quertermous, T. (1999) Octamer-
dependent in vivo expression of the endothelial cell-specific TIE2
gene. J. Biol. Chem. 274, 20376–20383.
41. Cieslik, K., Zembowicz, A., Tang, J.L. & Wu, K.K. (1998)
Transcriptional regulation of endothelial nitric-oxide synthase by
lysophosphatidylcholine. J. Biol. Chem. 273, 14885–14890.
42. Gegonne, A., Bosselut, R., Bailly, R.A. & Ghysdael, J. (1993)
Synergistic activation of the HTLV1 LTR Ets-responsive
region by transcription factors Ets1 and Sp1. EMBO J. 12, 1169–
1178.
43. Schaefer, T.S., Sanders, L.K. & Nathans, D. (1995) Cooperative
transcriptional activity of Jun and Stat3 beta, a short form of
Stat3. Proc. Natl Acad. Sci. USA 92, 9097–9101.
44. Bromberg, J. & Darnell, J.E. Jr (2000) The role of STATs in
transcriptional control and their impact on cellular function.
Oncogene 19, 2468–2473.
45. Tone, M., Powell, M.J., Tone, Y., Thompson, S.A. & Waldmann,

H. (2000) IL-10 gene expression is controlled by the transcription
factors Sp1 and Sp3. J. Immunol. 165, 286–291.
46. Schoenherr, C.J. & Anderson, D.J. (1995) The neuron-restrictive
silencer factor (NRSF): a coordinate repressor of multiple neuron-
specific genes. Science 267, 1360–1363.
47.Kennett,S.B.,Udvadia,A.J.&Horowitz,J.M.(1997)
Sp3 encodes multiple proteins that differ in their capacity to
stimulate or repress transcription. Nucleic Acids Res. 25, 3110–
3117.
48. Brightbill, H.D., Plevy, S.E., Modlin, R.L. & Smale, S.T. (2000) A
prominent role for Sp1 during lipopolysaccharide-mediated
induction of the IL-10 promoter in macrophages. J. Immunol. 164,
1940–1951.
49. Marazzi, S., Blum, S., Hartmann, R., Gundersen, D., Schreyer,
M., von Argraves, S.F.V., Pytela, R. & Ruegg, C. (1998) Char-
acterization of human fibroleukin, a fibrinogen-like protein
secreted by T lymphocytes. J. Immunol. 161, 138–147.
2286 M. Liu et al.(Eur. J. Biochem. 270) Ó FEBS 2003