| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Shionogi Institute for Medical Science, Mishima, Settsu-shi, Osaka, Japan; and
Department of Bacteriology, Kinki University School of Medicine, Ohno-Higashi, Osaka-Sayama, Osaka, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-type T cells, and NK cells. In humans, there are two highly
homologous genes encoding SCM-1
and SCM-1ß. Here we examined the
regulatory mechanism of the SCM-1 genes. The luciferase reporter gene
under the control of the 5' flanking region of 0.7 kb was strongly
induced upon activation with anti-CD3 or PHA plus PMA only in
SCM-1-producer T cell lines through a cyclosporin A-sensitive
mechanism. An element termed E1 located at -108 to -95 nt relative to
the major transcription start site was found to be critical for the
promoter activity. In electrophoretic mobility shift assays using the
E1 oligonucleotide as probe, nuclear extracts from unstimulated T and B
cell lines formed a constitutive complex termed complex I, while
nuclear extracts from stimulated SCM-1-producer T cell lines formed a
higher mobility complex termed complex II with a concomitant decrease
in complex I. The shift from complex I to complex II seen only in
SCM-1-producer T cell lines upon activation was completely suppressed
by cyclosporin A. Both complexes were critically dependent on the NF-AT
core sequence TTTCC in the E1 element and were partially supershifted
by anti-NF-ATp. One-hybrid assays in yeast isolated NF-ATp as an E1
binding protein, and transfection of NF-ATp into T and B cell lines
strongly enhanced the activation-dependent SCM-1 promoter activity.
Collectively, a unique mechanism involving NF-ATp appears to regulate
the cell type-specific and activation-dependent expression of the SCM-1
genes. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
ß-type thymocytes (1, 3, 4).
Furthermore, epidermal dendritic -type T cells and intestinal
intraepithelial -type T cells were found to produce copious
amounts of this cytokine upon activation (7).
IL-2-activated murine NK cells and human NK clones were also shown to
express its mRNA (6). Thus, SCM-1/lymphotactin/ATAC is a
unique cytokine whose expression is highly selective for certain
subsets of T cells and NK cells.
In our previous study we isolated genomic clones hybridizing with the
SCM-1 cDNA. There are two highly homologous SCM-1 genes, termed
SCYC1 and SCYC2, in the human genome
(8). These two genes are closely mapped to human
chromosome 1q23, similarly inducible in peripheral blood T cells and
Jurkat T cell line upon activation, and encode SCM-1 and
SCM-1ß with only two amino acid differences (8). We have
also identified a specific functional receptor for SCM-1 and
SCM-1ß, now termed XCR1, which is a seven-transmembrane G
protein-coupled receptor selectively expressed in tissues such as
placenta, spleen, and thymus (9).
In the present study we have investigated the molecular mechanism of SCM-1 gene expression in particular subsets of T cells. Because cyclosporin A (CsA) was shown to suppress the induction of this protein in activated CD8+ T cells (3), the NF-AT family transcription factors are likely to be involved in its expression (10). We have found that an NF-AT-like cis-acting element in the 5' flanking regions of the SCM-1 genes termed E1 is essential and sufficient for activation-dependent, cell type-specific, and CsA-sensitive gene expression. We have further demonstrated that E1 is a site of constitutive and activation-inducible DNA-protein interactions involving NF-AT family proteins.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Human T cell lines, Jurkat, PEER, and HPB-ALL, and a human B cell line BALL-1 were described previously (11). These cell lines and HeLa-S3 were maintained in RPMI 1640, 10% heat-inactivated FCS, and kanamycin (30 µg/ml). PBMC were isolated from heparinized venous blood from healthy donors by Ficoll-Hypaque density gradient centrifugation and cultured in RPMI 1640, 10% FCS, PHA (Life Technologies, Grand Island, NY; 1/100), and 100 U/ml of rIL-2 (Shionogi, Osaka, Japan). After 2 days, lymphocytes were harvested, washed twice in PBS, and resuspended in RPMI 1640, 10% FCS, and 100 U/ml of IL-2. After overnight culture, CD4+ and CD8+ T cells were negatively selected using magnetic beads (Dynal, Great Neck, NY) after incubation with anti-CD8 or anti-CD4 (Dako, Carpenteria, CA) for 30 min on ice, respectively. After that, whole and selected populations were expanded in RPMI 1640, 10% FCS, and 100 U/ml of IL-2 for 1 wk and used for transfection experiments. Flow cytometric analysis routinely demonstrated that CD4+ and CD8+ T cell populations were >80 and >90% pure, respectively.
Plasmid construction
The promoterless luciferase plasmid pGV-B was purchased from
Toyo Ink (Tokyo, Japan). To generate a series of 5' deletions, DNA
fragments were amplified from the SCM-1 genomic template by PCR
using a 3' oligonucleotide tailed with an XhoI site
(5'-CTCGAGCTGTGCAAGGAGA-3') and 5' oligonucleotides of 20 bp tailed
with a MluI site and beginning at -669, -497, -279,
-127, -117, -108, -98, -83, -62, -42, and -32 relative to the
major transcriptional start site (see Fig. 1
). PCR products were digested with
MluI and XhoI and cloned into the MluI
and XhoI sites of the reporter plasmid pGV-B. Base
substitutions were also introduced using PCR. An N-terminally truncated
464-aa form of NF-ATp (12), which contains the domains
required for DNA binding and for formation of a transcriptionally
active complex with Fos and Jun, was subcloned into the NotI
site of pEBV-His (Invitrogen, San Diego, CA). To increase fidelity, all
PCR amplifications were performed using Pfu polymerase (Stratagene, La
Jolla, CA) for only 15 cycles with a large amount of template (100
ng/reaction). All PCR-derived sequences were checked by DNA
sequencing.
|
pCMV-ß-gal, a CMV promoter-driven ß-galactosidase (ß-gal)
construct, was purchased from Promega (Madison, WI). It was used for
normalization of transfection efficiency. For transfection of each cell
line, 
1 x 107 cells were mixed with 30
µg of a test plasmid and 5 µg of pCMV-ß-gal in 200 µl of PBS
and transfected by electroporation at 250 V/500 µF. For transfection
of primary T lymphocytes, we followed the procedure described
previously (13). In brief, about 2 x
107 cells were mixed with 100 µg of a test
plasmid and 10 µg of pCMV-ß-gal in 0.8 ml of RPMI 1640, 10% FCS,
and 100 U/ml of IL-2 and transfected by electroporation at 350 V/960
µF. Transfected cell lines were cultured in RPMI 1640 and 20% FCS
for 18 h. Transfected lymphocytes were cultured in RPMI 1640, 20%
FCS, and 100 U/ml of IL-2 for 18 h. After that, cell cultures were
divided and further treated without or with PHA (1/100) and PMA (30
ng/ml; Sigma, St. Louis, MO) or with anti-CD3 (OKT3) that was
immobilized on plastic dishes by incubation at concentration of 1
µg/ml overnight. After additional 24-h culture, cells were washed
twice in PBS and resuspended in a lysis buffer for luciferase assay
(Promega). After 10 min, cell extracts were obtained by centrifugation.
After normalization of transfection efficiency by ß-gal assay,
luciferase assay was performed using the PicaGene kit (Toyo Inki) as
described previously (14). Light units were measured on a
luminometer (Lumat LB9501, Berthold, Wildbad, Germany).
Electrophoretic mobility shift assay
The following double-stranded oligonucleotides were synthesized
using a DNA synthesizer (Cyclon Plus DNA Synthesizer,
MilliGene/Biosearch, Bedford, MA): E1,
5'-ATTGCAAAGACTTTCCATGATCCTA-3'; muE1, 5'-ATTGCAAAGACTCTAGATGATCCTA-3';
NF-ATIL-2,
5'-GGGGAGGAAAAACTGTTTCATACAGAAGGCGT-3';
AP-1IL-2, 5'-TTCCAAAGAGTCATCAG-3'; and a
series of mutated E1 oligonucleotides with 2-bp substitutions shown in
Fig. 5E. The following oligonucleotides were purchased
from Promega: Oct-1, 5'-TGTCGAATGCAAATCACTAGAA-3'; and SP-1,
5'-ATTCGATCGGGGCGGGGCGAGC-3'. EMSA was conducted as
described previously (15). In brief, cells were stimulated
without or with PHA (1/100) and PMA (30 ng/ml) for 4 h in the
absence or the presence of CsA (100 ng/ml; Sandoz Research Institute,
Hanover, NJ). Nuclear extracts were prepared as described previously
(16). Double-stranded oligonucleotides were end labeled
using [-32P]ATP and T4 polynucleotide kinase
(Pharmacia LKB Biotechnology, Piscataway, NJ). Polyclonal
anti-NF-ATp was purchased from Upstate Biotechnology (Lake Placid,
NY). Polyclonal anti-NF-ATc and anti-NF-ATx were purchased from
Santa Cruz Biotechnology (Santa Cruz, CA). Nuclear extracts (3
µg/reaction) were incubated without or with competitors or
anti-NF-ATp (0.5 µg) for 15 min at room temperature and mixed
with 32P-labeled probes (3 x
104 cpm/ng, 1 ng/reaction) for another 20 min at
room temperature in 20 µl of a solution consisting of 10 mM Tris-HCl
(pH 7.5), 50 mM NaCl, 0.5 mM DTT, 0.1 mM EDTA, 10% glycerol, 1%
Ficoll, and 2 µg/ml poly(dI-dC). Bound complexes were separated from
free probes by electrophoresis on a 5% polyacrylamide gel. Gels were
dried, and complexes were visualized by autoradiography.
|
The yeast one-hybrid assay (17) was conducted by using Matchmaker One-Hybrid System (Clontech, Palo Alto, CA). Briefly, the concatemer consisting of six tandem repeats of the E1 element was linked to a low activity promoter directing expression of HIS3. The reporter construct was integrated into the recipient yeast genome. This yeast strain was transformed with a Jurkat cDNA library expressing fusion proteins with the target-independent Gal4 activation domain and plated on medium lacking histidine. HIS+ colonies were isolated, plasmids were prepared, and inserts were sequenced. The same reporter system directing expression of lacZ was used to test positive clones (17).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previously, we have described the presence of two highly
homologous SCM-1 genes encoding SCM-1 and SCM-1ß in the human
genome (8). These two genes are similarly inducible upon
cell activation (8). As shown in Fig. 1
, the 0.7-kb 5'
flanking sequences of the SCM-1 and SCM-1ß genes are 97%
identical. Computer analysis reveals a number of potential regulatory
elements. We first tested whether the 5' flanking sequences of the
SCM-1 and SCM-1ß genes were similarly inducible by cell
activation. We generated reporter genes by introducing the 0.7-kb 5'
flanking regions of SCM-1 and SCM-1ß in the upstream of the
promoterless luciferase gene pGV-B. We transfected these reporter
plasmids into Jurkat, a human CD4+ T cell line
capable of expressing SCM-1 (8). At 18 h
posttransfection, cells were stimulated with immobilized anti-CD3
or PHA and PMA, and cell extracts were prepared for luciferase assay
after additional 24-h culture. As shown in Fig. 2A, the 0.7-kb 5' flanking
regions of the SCM-1 and SCM-1ß genes showed very low levels of
basal expression, but were strongly induced upon stimulation with
anti-CD3 or PHA and PMA. Thus, the 0.7-kb 5' flanking regions of
SCM-1 genes contained essential regulatory elements for the
activation-dependent gene expression. Because these 0.7-kb 5'
franking regions are 97% identical at the DNA level (Fig. 1), our
subsequent experiments were conducted using the SCM-1 gene
promoter.
|
B we transfected the reporter plasmid with the
0.7-kb 5' flanking region of the SCM-1 gene (-669/luc) into
total T cells, CD4+ T cells and
CD8+ T cells. Upon treatment with PHA and PMA,
the reporter gene was strongly induced in primary human T cells.
Furthermore, a higher level of induction was consistently seen in
CD8+ T cells (90% purity) than in
CD4+ (80% purity). Thus, the 0.7-kb 5'
flanking region of the SCM-1 gene was capable of regulating
activation-dependent gene expression in primary human T lymphocytes
with preference for CD8+ T cells. Expression of
the reporter gene in the CD4+ T cell populations
might be due in part to the relatively low purity of
CD4+ T cells (80%) and/or to human
equivalents of the murine NK1.1+
CD4+ T cells that were mentioned to express
lymphotactin/SCM-1 (4).
In Fig. 2C, we further tested the cell type-specific
promoter activity by using five human cell lines. The CMV-ß-gal
plasmid was used to normalize transfection efficiency by ß-gal
activity. Consistent with the induction of the endogenous SCM-1 genes
(8), the treatment with PHA and PMA strongly activated the
promoter in Jurkat (a CD4 single-positive T cell line) and PEER (a
double-negative T cell line), but not in HPB-ALL (a
double-positive T cell line), BALL-1 (a B cell line), or HeLa (an
epithelial cell line) (11). Thus, the 0.7-kb 5'
flanking sequence of the SCM-1 gene is functional only in certain T
cell lines that are capable of expressing the endogenous SCM-1 genes
upon activation (8). We also tested the effect of CsA on
the stimulation-dependent activation of the promoter in Jurkat. As
shown in Fig. 2D, CsA completely abrogated the activation of
the promoter by PHA and PMA with an IC50 of about
1 ng/ml. Collectively, these results clearly demonstrated that the
0.7-kb 5' flanking regions of the SCM-1 genes contain major
cis-acting elements that regulate the activation-dependent,
cell type-specific, and CsA-sensitive expression of the SCM-1
genes.
Deletion analysis of the 5' flanking region
To determine the cis-acting regulatory elements in the
5' flanking region of the SCM-1 gene, we transfected a series of
luciferase reporter plasmids with progressively deleted 5' flanking
regions into Jurkat cells. As shown in Fig. 3, all the plasmids showed a very low
level of basal promoter activity. Upon PHA and PMA stimulation,
-669/luc and -497/luc induced luciferase activity by about 30-fold;
-279/luc, -117/luc and -108/luc induced luciferase activity by about
15-fold; -98/luc, -83/luc, and -62/luc induced luciferase activity
by about 3-fold; -42/luc and -32/luc induced luciferase activity by
<2-fold. From these results, at least three regions (-497 to -279,
-108 to -98, and -62 to -42) appeared to contribute to the full
inducibility of the SCM-1 promoter upon treatment with PHA and PMA,
and the region between -108 and -98 contained a major
cis-acting element. A close inspection of the sequence in
and around this region revealed a sequence 5'-GACTTTCCAT-3' (-101 to
-92) that contained an inverted core sequence of the NF-AT site
(TTTCC) (10). Because NF-AT is known to be involved in the
expression of various cytokine genes in T cells by activation-dependent
and CsA-sensitive mechanisms (10), our subsequent study
was focused on this element. For the sake of convenience, we designated
the sequence between -108 and -83 as element 1 (E1; Fig. 1).
|
We first examined the effect of mutations in the NF-AT core
sequence of E1 on the promoter activity of the -669-bp 5' flanking
region. The luciferase reporter plasmids with the wild-type E1 or a
mutated E1 with three base substitutions in the NF-AT core sequence
(Fig. 4A) were transfected
into Jurkat and PEER. As shown in Fig. 4B, the mutations at
this site effectively abrogated the stimulation-dependent induction of
the reporter driven by the -669-bp 5' flanking region in both cell
lines. To further test whether E1 was an activation-inducible and
CsA-sensitive regulatory element, we constructed an artificial reporter
plasmid by inserting six tandem repeats of E1 in the upstream of the
-42/luc (Fig. 3). As shown in Fig. 4C, the reporter driven
by the tandem repeats of E1 was indeed induced by about 10-fold in
transfected Jurkat by PHA and PMA. Furthermore, this induction was
fully abrogated by CsA. Taken together, E1 containing the NF-AT core
sequence is a critical regulatory element for the activation-dependent
and CsA-sensitive expression of the SCM-1 promoter.
|
To demonstrate specific interactions of nuclear proteins with E1,
EMSA was conducted using the E1 probe (Fig. 5A). Unexpectedly, even the
nuclear extract from unstimulated Jurkat was capable of forming a
DNA-binding complex, termed complex I. On the other hand, the nuclear
extract from stimulated Jurkat formed a higher mobility complex, termed
complex II, with concomitant decrease in complex I. Both complexes were
blocked by the wild-type E1 site probe (wt), but not by a mutated E1
site probe with three base substitutions in the NF-AT core sequence
(mu, see Fig. 4A). Moreover, the mutated E1 site probe
(mu-E1) failed to form any complexes with the nuclear extracts from
either unstimulated or stimulated Jurkat. These results indicated that
the NF-AT core sequence of the E1 site, which is critical for the
enhancer activity of the E1 site (Fig. 4), is also critical for the
formation of both complex I and complex II.
EMSA was also conducted to test the effect of CsA. As shown in Fig. 5B, the nuclear extract of Jurkat cells stimulated with PHA
and PMA in the presence of CsA only formed complex I and failed to form
complex II. Thus, the shift from complex I to complex II observed upon
stimulation with PHA and PMA was effectively blocked by CsA treatment.
This further supported the idea that the shift from complex I to
complex II was required for the transcriptional activation.
By cold competition experiments, we next analyzed the nature of the
nuclear proteins binding to E1 in complex I and II. As shown in Fig. 5C, both complexes were clearly inhibited by the E1 probe at
10-fold excess and by the IL-2 NF-AT site probe at 10- to 30-fold
excess (15). On the other hand, the IL-2 AP-1 site probe
(15) hardly affected complex I even at a 100-fold excess.
This probe, however, partially inhibited complex II at a 30- to
100-fold excess. For an internal control, EMSA was also conducted using
the IL-2 NF-AT site probe (15). As expected, a complex was
formed only with the extract from stimulated Jurkat. Furthermore, this
complex was inhibited by the IL-2 NF-AT site probe (and the E1 site
probe) as well as by the IL-2 AP-1 site probe (10, 15)
(Fig. 5D). Collectively, these results suggested that both
complexes contained proteins related to NF-AT. Furthermore, the complex
II might contain a cofactor related to the AP-1 family. However, there
is no consensus AP-1 site as defined by T(T/G)ANT(A/C)A
(18) in or around the E1 site (Fig. 1).
To further define the critical nucleotides in the E1 site that were
involved in the formation of constitutive complex I and inducible
complex II, we next conducted EMSA in the presence of mutated E1 site
probes with scanning-type 2-bp substitutions as cold competitors. As
shown in Fig. 5E, both constitutive complex I and
activation-inducible complex II showed essentially identical inhibition
patterns, indicating that the same nucleotide sequence in the E1 site
was involved in the formation of both complexes. Furthermore, the lack
of competition by M6, M7, and M8 revealed that TTTCCA corresponding
exactly to the NF-AT core sequence was essential for the formation of
both complexes (10).
Cell type-specific formation of complex II
By using the E1 probe, EMSA was further conducted with nuclear
extracts from Jurkat, PEER, BALL-1, and HPB-ALL (Fig. 6, upper panel). Both Jurkat
and another SCM-1 producer T cell line, PEER (8), formed
the constitutive complex I before activation and the inducible complex
II with a concomitant decrease or disappearance of complex I upon
activation with PHA and PMA. SCM-1-nonproducer cell lines BALL-1 and
HPB-ALL (8) also formed complex I, but failed to form
complex II even after activation with PHA and PMA. Thus, the formation
of complex II was seen only with the nuclear extracts from SCM-1
producer cell lines upon stimulation with PHA and PMA (Fig. 2). As
internal controls, the same nuclear extracts were probed with the IL-2
NF-AT site probe (15). In this case, the DNA-binding
complex was seen only with extracts from stimulated Jurkat and HPB-ALL
(Fig. 6, lower panel).
|
To test the presence of the NF-AT family proteins in complexes I
and II, the effect of polyclonal anti-NF-ATp was examined in EMSA.
As shown in Fig. 7A, complex I
formed with the nuclear extract from unstimulated Jurkat was partially
supershifted by anti-NF-ATp, while both complex I and complex II
formed with the nuclear extract from activated Jurkat were reduced by
anti-NF-ATp (panel E1). A faint supershift band
was also seen. The discrepancy of the effect of anti-NF-ATp on
complex I formed by unstimulated Jurkat and that formed by stimulated
Jurkat was unexpected and might indicate that they were not identical.
As internal controls, the same Ab was shown to partially supershift the
IL-2 NF-AT nuclear complex (panel NFAT), but to have
no effect on the SP-1 nuclear complex (panel SP-1).
We further examined the effects of various anti-NF-AT Abs in EMSA
using the nuclear extracts from PEER. As shown in Fig. 7B,
anti-NF-ATp again partially supershifted complex I formed by the
nuclear extract of unstimulated PEER. It also slightly reduced and
faintly supershifted the complex II formed by the nuclear extract from
stimulated PEER even though the effect was less dramatic than that seen
with Jurkat. No such effects were seen with anti-NF-ATc and
anti-NF-ATx. Collectively, these results support that both
complexes I and II contain NF-ATp or some proteins immunologically
cross-reactive with anti-NF-ATp.
|
To identify proteins binding to E1, we conducted one-hybrid
screening in yeast (17). The reporter construct consisted
of six tandem repeats of E1 site linked to a low activity promoter
directing His3 gene expression. A Jurkat cDNA library of
2 x 106 clones expressing a translational
fusion with the Gal4 trans-activating domain was transformed
into a his- yeast strain carrying the
HIS3 reporter plasmid. About 2 million transformants were
plated on the medium lacking histidine, and 20
HIS+ colonies were isolated. By partial
sequencing, we found that 10 of 20 positive clones were NF-ATp. We
sequenced one NF-ATp clone and found that it encoded an N-terminally
truncated NF-ATp starting from the amino acid position at 401 and
consisting of 521 aa fused in-frame to the transcriptional activation
domain of Gal4. To measure its trans-activation activity, we
conducted ß-gal reporter assay in yeast using a reporter construct
consisting of six tandem repeats of E1 site linked to a low activity
promoter directing lacZ gene expression (17).
As shown in Fig. 8, the NF-ATp-Gal4
fusion protein was capable of inducing the reporter gene by about
20-fold. Thus, NF-ATp is capable of directly binding to the E1 site in
yeast.
|
To further examine the effect of NF-ATp on the SCM-1 promoter, we
cotransfected -669/luc with wild-type E1 site (wt) or mutated E1 site
(µ; see Fig. 4) and either a control vector or a vector expressing an
N-terminally truncated 464-aa functional NF-ATp (12) into
Jurkat and BALL-1. As shown in Fig. 9,
transfection of functional NF-ATp into Jurkat significantly augmented
the activity of the SCM-1 promoter with the wild-type E1 site, but not
that with the mutated E1 site upon stimulation with PHA and PMA.
Furthermore, transfection of NF-ATp into the SCM-1-nonproducer BALL-1
strongly induced the activity of the wild-type promoter, but not that
of the mutated one upon stimulation with PHA and PMA. The latter
observation may support the lack of expression of functional NF-ATp in
BALL-1 (19). Thus, an N-terminally truncated functional
NF-ATp (12) is capable of trans-activating the
SCM-1 promoter through the E1 site upon cell activation.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
120 kDa), NF-AT2/NF-ATc (80 kDa), NF-AT3 (100
kDa), and NF-AT4/NF-ATx (120 kDa) (12, 20, 21, 22, 23). They
share a highly homologous DNA-binding domain that shows a weak sequence
similarity to the DNA-binding domain of the Rel family proteins
(24, 25). Activation of T cells leads to translocation of
cytoplasmic NF-AT into the nucleus through dephosphorylation of NF-AT
by Ca2+/calmodulin-dependent phosphatase
calcineurin, which is the target of the immunosuppressive drugs CsA and
FK506 (26). Within nuclei, translocated NF-AT and a de
novo synthesized cofactor cooperatively bind to a composite site
(10). Thus, most NF-AT binding sites in various cytokine
promoters are accompanied by the binding sites for auxiliary
transcription factors such as AP-1 and Oct (10). For
example, NF-AT and AP-1 cooperatively bind to multiple composite sites
in the promoter of IL-2 for full induction in activated T cells
(27). The distal NF-AT site in the IL-4 promoter also
consists of a composite binding site for NF-AT and AP-1
(28). Recently, the proto-oncogene c-maf was
found to be expressed highly selectively in Th2-type T cells and to act
in synergy with NF-AT to trans-activate the IL-4 promoter by
binding adjacent to the proximal NF-AT site (29).
Furthermore, a nuclear factor designated NIP45 has been shown to
interact with the Rel homology domain of NF-ATp and to act
in synergy with NF-ATp and c-Maf in the expression of IL-4
(30). Thus, it is likely that cell type-specific cofactors
assembling at a given NF-AT site determine the cell type-specific
expression of various cytokines.
SCM-1/lymphotactin/ATAC is a unique cytokine whose expression is highly
selective for activated CD8+ T cells,
double-negative thymocytes, intradermal +
dendritic T cells, intestinal intraepithelial
+ T cells, and NK cells
(1, 2, 3, 4, 5, 6, 7, 8). Like IL-2 and IL-4, which serve as models to
examine differential transcriptional regulation between Th1 and Th2
(31, 32), SCM-1 may provide a unique model for
activation-dependent gene expression in particular subsets of T cells
and NK cells. In the present study we have shown that the SCM-1
promoter has an element termed E1 between -108 and -90 relative to
the transcription initiation site (Fig. 1). E1 is essential and
sufficient for the activation-dependent and CsA-sensitive expression of
the SCM-1 genes (
Figs. 2–4). Importantly, the E1 site contains the
NF-AT core sequence (TTTCC) (10) that is essential for its
enhancer activity (Fig. 4). Unlike most other NF-AT sites
(10), however, the E1 site probe formed a constitutive
complex, termed complex I, with nuclear extracts from unstimulated
SCM-1 producer T cell lines and even from SCM-1 nonproducer cell lines
(Fig. 5). With nuclear extracts from stimulated SCM-1 producer T cell
lines, however, the E1 probe formed a higher mobility complex, termed
complex II, with a concomitant decrease in complex I (Figs. 5 and 6).
Furthermore, the shift from complex I to complex II upon cell
activation was effectively suppressed by treatment with CsA (Fig. 5).
Both complexes were critically dependent on the NF-AT core sequence
(Fig. 5) and were partially supershifted by anti-NF-ATp, but not by
anti-NF-ATc or anti-NF-ATx (Fig. 7). Furthermore, NF-ATp was
found to be a protein capable of binding to the E1 site bait in the
yeast one-hybrid assay (Fig. 8) (17). Exogenously
transfected N-terminally truncated functional NF-ATp (12)
was further shown to be capable of trans-activating the
SCM-1 promoter in the E1-dependent manner upon stimulation with PHA and
PMA (Fig. 9). Collectively, these results support the following
conclusions. Transcription from the SCM-1 promoter takes place through
a shift from the preexisting complex I to the activation-inducible
complex II. The NF-AT core sequence (TTTCC) in E1 is critical for the
formation of both complexes. NF-ATp is capable of binding to the NF-AT
core sequence of E1. NF-ATp appears to be associated with both
complexes, but induces transcription of the SCM-1 promoter only upon
stimulation with PHA and PMA.
The SCM-1 E1 site thus presents notable differences from the classical
IL-2 NF-AT sites (10). In the case of IL-2 NF-AT sites,
NF-AT proteins bind only after stimulation-induced nuclear
translocation (10). In the case of E1, however, not only
the inducible complex II but also the pre-existing complex I seem to
contain NF-ATp or proteins immunologically reactive with
anti-NF-ATp. Second, the IL-2 AP-1 site competitor only weakly
inhibited the activation-inducible complex II of E1 (Fig. 5). This
contrasts to the IL-2 NF-AT sites, because the IL-2 AP-1 site
competitor effectively inhibits the complex between the IL-2 NF-AT
probe and nuclear extracts from stimulated T cells (10, 15) (Fig. 5). In fact, no potential AP-1 site with a consensus
sequence of T(T/G)ANT(A/C)A (18) is present adjacent to or
within the E1 element (Fig. 1). Like SCM-1 E1, the
activation-responsive element (ARE) located between -88 and -60 of
the murine IL-4 gene was bound by proteins from both unstimulated and
stimulated nuclear extracts of EL-4 T cells (33).
Furthermore, both constitutive and inducible complexes of the IL-4 ARE
contained proteins immunoreactive with anti-NF-ATp and
anti-NF-ATc (33). In contrast to the SCM-1 E1 site,
however, the AP-1 family members were clearly associated with the
activation-induced complex of the IL-4 ARE (34). The same
ARE was also critical for activation-dependent expression of the IL-4
gene in murine mast cells and was also shown to be bound by nuclear
proteins from both unstimulated and stimulated mast cells
(35). Both constitutive and inducible complexes contained
proteins immunoreactive with anti-NF-ATp (35). In
contrast to T cells, however, the AP-1 family members were not
associated with the activation-inducible complex of mast cells
(35). The observation of mast cells is thus rather similar
to our present findings. In contrast to the SCM-1 E1 site, however, the
constitutive complex of ARE formed by nuclear extracts from
unstimulated mast cells had an electrophoretic mobility much faster
than the activation-induced complex formed by nuclear extracts from
stimulated mast cells (35).
Collectively, some NF-AT proteins appear to be constitutively present in the nuclei of certain types of cells and capable of binding to elements such as SCM-1 E1 or IL-4 ARE even without cell activation. In fact, a protein with a molecular mass of about 41 kDa and reactive with anti-NF-ATp was shown to be present in the nuclei of unstimulated murine mast cells (35). It remains to be seen whether this protein is encoded by an unknown gene belonging to the NF-AT family or produced by an alternative splicing or a posttranslational modification of NF-ATp. Recently, the carboxyl-terminal end of NF-AT4/x has been shown to be required for its maximum trans-activation activity, and notably, the same sequence is well conserved in other NF-AT family proteins (36). Furthermore, multiple isoforms of NF-AT proteins with distinct carboxyl-terminal regions and with different trans-activation activities have been shown to be generated by alternative splicing (21, 36, 37). Thus, some isoforms of NF-ATp may be capable of trans-locating into the nucleus without cell activation and may even behave as a negative factor. Upon cell activation, they may be replaced by transcriptionally active NF-ATp translocated from the cytoplasm through dephosphorylation by calcineurin. Transcriptional activation may also require some de novo synthesized cofactors specific for each cell type and to be associated with NF-ATp. The E1 element of the SCM-1 genes may be useful to explore this new aspect of transcriptional regulation by the NF-AT system.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Osamu Yoshie, Department of Bacteriology, Kinki University School of Medicine, 377-2 Ohno-Higashi, Osaka-Sayama, Osaka 589-8511, Japan. E-mail address:
3 Abbreviations used in this paper: SCM-1, single C motif-1; ATAC, activation-induced, T cell-derived and chemokine-related molecule; CsA, cyclosporin A; XCR1, XC chemokine receptor 1; E1, element 1; ARE, activation-responsive element; ß-gal, ß-galactosidase.
Received for publication October 13, 1998. Accepted for publication July 7, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
type chemokine, SCM-1. FEBS Lett. 360:155.[Medline]
T cells. J. Immunol. 157:985.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  B. O. Kim, Y. Liu, B. Y. Zhou, and J. J. He Induction of C Chemokine XCL1 (Lymphotactin/Single C Motif-1{alpha}/Activation-Induced, T Cell-Derived and Chemokine-Related Cytokine) Expression by HIV-1 Tat Protein J. Immunol., February 1, 2004; 172(3): 1888 - 1895. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Gebert, W. Fischer, E. Weiss, R. Hoffmann, and R. Haas Helicobacter pylori Vacuolating Cytotoxin Inhibits T Lymphocyte Activation Science, August 22, 2003; 301(5636): 1099 - 1102. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. D. Cristillo, M. J. Macri, and B. E. Bierer Differential chemokine expression profiles in human peripheral blood T lymphocytes: dependence on T-cell coreceptor and calcineurin signaling Blood, January 1, 2003; 101(1): 216 - 225. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Nagai, S.-i. Hashimoto, T. Yamashita, N. Toyoda, T. Satoh, T. Suzuki, and K. Matsushima Comprehensive gene expression profile of human activated Th1- and Th2-polarized cells Int. Immunol., March 1, 2001; 13(3): 367 - 376. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
