| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Hanson Center for Cancer Research, Institute of Medical and Veterinary Science, Adelaide, Australia
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Many previous studies have been devoted to studying the transcriptional
regulation of the IL-3 and GM-CSF genes. These closely linked genes,
together with their known regulatory elements, are located within a
30-kb segment of DNA (Refs. 10, 11, 12, 13, 14 and Fig. 1
A). We have used the approach of mapping DNase I
hypersensitive (DH)5
sites across the entire IL-3/GM-CSF locus to identify distal DNA
elements that govern the correct developmentally regulated and
differential expression of the IL-3 and GM-CSF genes
(10, 11, 12, 13, 14). These studies identified inducible DH sites 14
kb upstream of the IL-3 gene, and 3 kb upstream of the GM-CSF gene that
function as inducible enhancers. These enhancers are dependent upon a
combination of kinase and calcium signaling pathways, which in T cells
are linked to the TCR. Both enhancers are activated via arrays of NFAT
sites, and are repressed by the immunosuppressant cyclosporin A (CsA)
that blocks the calcium-dependent induction of NFAT
(10, 11, 12, 13, 14). Despite these similarities, the two enhancers
have distinct tissue-specific functions that largely mirror the
expression patterns of the IL-3 and GM-CSF genes. The -14-kb IL-3
enhancer functions exclusively in T cells, and this specificity is
mediated by a composite NFAT/Oct element that recruits NFAT and Oct
factors together with the NFAT cofactor NIP45 and the lymphoid-specific
Oct cofactor OCA-B (10). In contrast, the -3-kb GM-CSF
enhancer functions in a much wider range of NFAT-expressing cells such
as T cells, myeloid cells, and endothelial cells (12), and
is activated via composite NFAT/AP-1 elements (13). The
fact that the locus has evolved with two distinct enhancers also makes
it more likely that the IL-3 and GM-CSF genes are independently
regulated. Furthermore, it is already established that a 10-kb GM-CSF
transgene is correctly regulated in vivo in isolation from the IL-3
locus (12).
|
The aim of this study was to further characterize the tissue-specific distribution of DH sites upstream of the IL-3 gene, and to determine their role in its regulation. This investigation led to the identification of a previously unidentified inducible DH site at -4.5 kb that is closely linked to a constitutive DH site at -4.1 kb. The -4.5 DH site functioned as an inducible NFAT-dependent enhancer, and it cooperated with the -14-kb IL-3 enhancer to activate the IL-3 promoter. However, the -4.5-kb enhancer had a broader tissue-specific range than the -14-kb enhancer because it functioned in both T cells and mast cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
DH sites were assayed essentially as described previously
(11, 12, 13, 14, 22). In brief, cells were either unstimulated, or
stimulated for 4–6 h with 20 ng/ml PMA and 2 µM A23187 (PMA/I), and
isolated nuclei were digested for 3 min at 22°C with a range of 1–18
µg/ml DNase I. Optimally digested DNA samples were digested with
either BamHI, EcoRI, or SpeI, and
probed with 32P-labeled probes. DH sites upstream
of a BamHI site located 10.1 kb upstream of the IL-3 gene
were probed with a 0.7-kb EcoRI-BamHI fragment of
DNA (-10.1 to -10.8 kb) isolated from 
-66 (11, 14).
DH sites upstream of an internal IL-3 gene EcoRI site at +2
kb were probed with a mixture of a 0.56-kb
StuI/EcoRI fragment from the 3' end of the IL-3
gene, and a 0.42-kb SmaI/SacI fragment from the
5' end of the IL-3 gene isolated from J1–16 (4). This
probe design avoids a repeat element within the IL-3 gene.
SpeI-digested DNA was probed with a 0.5-kb
SpeI/BamHI fragment located 6.1–6.6 kb upstream
of the IL-3 gene and isolated from -66 (11). The DNA
samples analyzed in this study have been selected on the basis that
they are optimal for the detection of ubiquitous DH sites downstream of
the IL-3 gene (12).
DNA sequence homology
The human and mouse IL-3/GM-CSF locus sequences were compared using the Blast2 program (23) located on the National Center for Biotechnology Information server (http://www.ncbi.nlm.nih.gov/BLAST/). A region extending from 20 kb upstream of the IL-3 gene to the GM-CSF promoter (13 kb downstream of the human IL-3 promoter) was compared with the mouse GenBank file GI:7542829. The human sequence was compiled from GenBank files AC004511, AC009175, and L77036, and our own unpublished primary sequence data. This compiled sequence is almost identical with the newly released GenBank file AC034216 that corresponds to the completed sequence of the same region from a different allele. DNA sequences rated as highly conserved encompassed 200–360 bp of 80–83% homology. DNA sequences rated as moderately conserved encompassed either 50–120 bp of 80% homology, or 147–484 bp of 61–76% homology.
Plasmids
The pXP1 luciferase reporter gene plasmid was a gift from Dr. S.
Nordeen (University of Colorado, Denver, CO; Ref. 24).
pXPG is a modified version of pXP1 containing the
Luc+ gene from pGL3 (Promega, Madison, WI) and a
novel high copy plasmid origin of replication, and is designed to avoid
serious read-through transcription artifacts that appear to occur with
pGL3 and to replace the low copy pBR322 origin of replication in pXP1
(25). pXPG-GM55 (25) contains the minimum
GM-CSF promoter core region (-55 to +28) in pXPG. pXPG-herpes simplex
thymidine kinase (TK)229 was generated by cloning a 229-bp
HindII/PvuII fragment spanning the TK promoter
(obtained from the Promega pRL-TK vector) into pXPG. pXPG-TK229-SV40E
was generated by cloning a 175-bp BamHI fragment containing
SV40 enhancer sequences from SV40 positions 113–270 into pXPG-TK229.
pIL3H (referred to as pIL3 in Ref. 14) contains the -559
to +50 fragment of the IL-3 promoter in pXP1. pXPG-IL3H was generated
by subcloning a 611-bp HindIII IL-3 promoter fragment from
pIL3H into pXPG. pIL3H-S1.3 and pS1.3 were generated by subcloning a
1.3-kb SacI fragment from 3.3–4.6 kb upstream of the IL-3
locus from 66 (11) into the SacI sites of
pIL3H and pIC19H. pXPG-GM55-IL3E (referred to as pIL3E in Ref.
10) contains a 330-bp NruI/AccI fragment from
the human -14-kb IL-3 enhancer cloned into pXPG-GM55. pXPG-GM55-GME
(referred to as pGME in Ref. 10) contains a 425-bp
BamHI/MscI fragment from the human GM-CSF enhancer cloned
into pXPG-GM55. pXPG-GM55-SS245, pXPG-IL3H-SS245, and pXPG-TK229-SS245
were generated by cloning a 245-bp Ecl136II/StuI 5' fragment
of pS1.3 into pXPG-GM55, pXPG-IL3H, and pXPG-TK229, respectively.
pIL3H-B1.2 (referred to as pIL3B1.2 in Ref. 14),
pXPG-IL3H-B1.2, and pIL3H-S1.3-B1.2 were generated by cloning a 1.2-kb
BglII fragment spanning the human IL-3 enhancer into pIL3H,
pXPG-IL3H, and pIL3H-S1.3. pXPG-GM55-SE546 was generated by cloning a
546-bp StuI/EcoRV (SE546) fragment of pS1.3
encompassing the -4.1 DH site into pXPG-GM55.
PCR was used to generate 5' deletion fragments of the SS245 enhancer for cloning where no appropriate restriction enzymes were present. Primers used for this purpose include an introduced SmaI site. The 5' primers were: AGGAGGCCCGGGCTCTCCACC, TGTCACCCGGGCCACCAGCGG, and TCTGACCCGGGTGATGCCATGG. The 3' primer was TGGACCCGGTCTAATGGATCCC. The three PCR-generated fragments were digested with SmaI and StuI, and cloned into pXPG-GM55 to give to the SS285, SS146, and SS95 derivatives, respectively. The pXPG-GM55-MS197 plasmid was generated by cloning a 197-bp MscI/StuI fragment of the SS245 enhancer into pXPG-GM55.
Site-directed mutagenesis was performed on a plasmid carrying the SS245
enhancer using PCR and the following double-stranded oligonucleotides
(altered bases are depicted in lower case): 
AML1:
CTGAGCTGCCTTCTGTccTGACAAAACCAGGCGATG,
AAP-1:CCTTCTGTGGTGACcgAACCAGGCGATGTCATC,
BNFAT:CACTTGTGATGCCATccAAAGGGTGGCGGAGG,
Sp1:GCCATGGAAAGGGTttCaaAGGAGCTTGTCACAGTGACTGAGG, and
CNFAT:GGTACCCCAGAAAATTggACTAGGGGGTCGACAAGC. SS245 enhancer
fragments containing the appropriate mutations were isolated from the
resulting plasmids and cloned into pXPG-GM55.
pE°tk.neo is a selection plasmid used to create stably transfected G418-resistant cell lines (26).
Human cell lines and tissue culture
The Jurkat, CEM, and HSB2 T cell lines, the KG1a, U937, and K562 myeloid cell lines, and the Raji and Ball-1 B cell lines are all human leukemic cell lines that have been used by us in previous studies (10, 11, 12, 13, 14). HMC-1 is a mast cell line (27) provided by Dr. J. Butterfield (Mayo Clinic, Minneapolis, MN). KG1 is a CD34+ myeloblastic cell line, whereas KG1a is a subclone of KG1 that has acquired some T cell characteristics (28). HuT78 and Molt4 are leukemic T cell lines, HepG2 is a hepatic cancer cell line, 5637 is an epithelial bladder carcinoma cell line, and HeLa is a cervical carcinoma cell line. Lung fibroblasts were obtained as passaged embryonic lung fibroblasts from CSL (Melbourne, Australia). Passaged endothelial cells were provided by Dr. J. Gamble (Hanson Centre for Cancer Research, Adelaide, Australia) and were prepared from cells harvested by trypsinization of umbilical cords (HUVECs) as in our previous study (29). PBMCs were purified by centrifugation of peripheral blood over Lymphoprep, and collecting cells trapped at the interface. Peripheral blood monocytes were purified from the mononuclear cell fraction by centrifugal elutriation and were provided by the laboratory of Dr. A. Lopez (Hanson Centre for Cancer Research). Primary T cells were purified by isolating the mononuclear cell fraction of peripheral blood and depleting the B cells and myeloid cells by adherence to nylon wool. T lymphoblasts were prepared by activating PBMCs with 2 µg/ml PHA for 3 days followed by culture for 6 days in 5 ng/ml recombinant human IL-2.
With the exception of 5637 cells that were grown in 7% FCS, all cell lines were grown in RPMI 1640 supplemented with 10% heat-inactivated FCS, 25 mM HEPES (pH 7.3), 2 mM L-glutamine, 28 mM NaHCO3, 50 U/ml penicillin, and 50 mg/ml streptomycin. Human T cells were grown in media supplemented with 50 µM 2-ME.
Transient transfection and luciferase assay
A total of 4.5 x 106 cells were transfected with 5–10 µg of CsCl-purified reporter plasmid DNA by electroporation as described (11). Cells were cultured for 20–24 h, and then stimulated with a combination of 25 ng/ml PMA plus 1.2 µM calcium ionophore A23187 (PMA/I) in the presence or absence of 0.1 µM CsA and incubated for 9 h before harvesting, and assayed for luciferase activity as described (11).
Stable transfection assays
Jurkat cells were transfected with 10 µg of the FspI-linearized reporter plasmid and 1 µg of the PvuI-linearized pE°TKneo (26) as described above. Cells were allowed to recover for 2 days before commencing G418 selection. The cells were incubated with 600 µg/ml G418 for 20 or 21 days, after which time cells in the unelectroporated control flask were all dead. Cells were removed from G418 selection for at least 24 h before stimulating 1 x 106 cells in 10 ml with PMA/I. Cells were stimulated for 9 h, and then harvested and assayed for luciferase activity as described (11).
EMSA
Nuclear extracts were prepared from stimulated and nonstimulated Jurkat cells by a modification of the method of Lavery and Schibler (30). Jurkat cells were stimulated with PMA/I for 2 h. Cells were washed with PBS and resuspended in five times the pellet volume of a hypotonic buffer (solution 1: 10 mM HEPES (pH 8.0), 10 mM KCl, 0.75 mM spermidine, 0.15 mM spermine, 0.2 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5 mM PMSF, 10 µg/ml aprotinin, and 10 µg/ml leupeptin) and incubated on ice for 10 min to lyse the cells. Nuclei were spun down, resuspended with five times the pellet size of solution 1, homogenized with a Dounce glass homogenizer 5–10 times, and combined with 10% the volume of solution 2 (50 mM HEPES (pH 8.0), 10 mM KCl, 0.75 mM spermidine, 01.5 mM spermine, 0.2 mM EDTA, 70% sucrose, 1 mM DTT, 0.5 mM PMSF, 10 µg/ml aprotinin, 10 µg/ml leupeptin). The nuclei were pelleted by centrifugation and resuspended in nuclei suspension buffer (20 mM Tris-HCl (pH 7.9), 75 mM NaCl, 0.5 mM EDTA, 0.85 mM DTT, 0.125 mM PMSF, 50% glycerol) so that the nuclei concentration was at 4 x 106 nuclei/µl. Nine times the nuclei suspension volume of NaCl-urea-Nonidet P-40 buffer (1.1 M urea, 0.33 M NaCl, 1.1% Nonidet P-40, 27.5 mM HEPES (pH 7.6), 1.1 mM DTT) was then added drop wise. After all of the NaCl-urea-Nonidet P-40 buffer was added, samples were gently vortexed and left on ice for 15 min until lysis was complete. Nuclei lysates were then spun at 13,000 rpm in a microcentrifuge for 15 min, and the supernatants were transferred to new tubes and glycerol concentrations were increased to 10%. Nuclear extracts were then aliquoted and stored at -70°C.
Binding in EMSAs was performed for 20 min at room temperature in 20-µl reaction volumes that contained 10 mM HEPES (pH 8.0), 0.1 mM EDTA, 0.1 M NaCl, 2 µg of poly(dI-dC), 10% glycerol, 0.03% Nonidet P-40, 1 mM DTT, 1 mM EDTA), 2 mM EGTA, 0.2 ng of end-filled 32P-labeled probes, and 5 µg of nuclear extract. Where oligonucleotide duplex competitors were included, 20 ng of competitor was added in 1 µl. When Abs were used, the extracts were preincubated with Abs on ice for 20 min, and then all the other components were added. We used an Ab against AML1 provided by Dr. N. Speck (Dartmouth Medical School, Hanover, NH) and an NFATC2 Ab (R59) provided by Dr. A. Rao (Harvard Medical School, Boston, MA). EMSAs were performed on 15-cm long, 1.5-mm thick 4% polyacrylamide gels containing 25 mM Tris/Borate and 0.5 mM EDTA. The gels were then fixed with 0.1% hexadecyl trimethylammonium bromide (cetrimide) for 20 min and dried before autoradiography.
Recombinant AP-1 containing the DNA-binding domains of cFos and cJun, and recombinant NFAT containing the DNA-binding domain of NFATC2 were also used in EMSAs to confirm NFAT and AP-1 binding as in our previous studies (13). These assays incorporated 0.2 ng of NFAT and/or 2 ng of Fos plus 2 ng of Jun with 0.1 µg of poly(dI-dC).
Oligonucleotides used as probes in EMSAs
The following double-stranded oligonucleotides were used for
binding assays: probe A AML1/AP-1: GCTGCCTTCTGTGGTGACAAAACCAGGCG,
probe B NFAT/Sp1: TGCCATGGAAAGGGTGGCGGAGGAGCTTGT, probe B2
NFAT/Sp1/AP-1: CTTGTGATGCCATGGAAAGGGTGGCGGAGGAGCTTGTCACAGTG, probe C
NFAT: CGGGTACCCCAGAAAATTCCACTAGGCCTGTG, probe C2 NFAT:
CCAGGTTCTTGCCTGTGGCGGGTACCCCAGAAAATTCCACTAGGCCTGTG, probe D
NFAT/AP-1:
CCAGGCGATGTCATCAGGGCCACCAGCGGAAATACAACCGAGGTCATGGG, probe A
AML1/AP-1: GCTGCCTTCTGTccTGACAAAACCAGGC, probe A AML1/AP-1:
GCTGCCTTCTGTGGTGACcgAACCAGGC, probe B NFAT/Sp1:
TGCCATccAAAGGGTGGCGGAGGAGCTTGT, probe B NFAT/Sp1:p
TGCCATGGAAAGGGTttCaaAGGAGCTTGT, probe C NFAT:
CGGGTACCCCAGAAAATTggACTAGGCCTGTG, and probe C2 NFAT:
CCAGGTTCTTGCCTGTGGCGGGTACCCCAGAAAATTggACTAGGCCTGTG.
Oligonucleotides used as competitors in EMSAs
The following oligonucleotides were used as competitors in
EMSAs: GM170: GATCCTGGAGTGACTCAAGCCCCTGTTTCCTACAG and GM430:
GATCTCACACATCTTTCTCATGGAAAGATGA, derived from the GM-CSF enhancer
and containing GM170 AP-1/NFAT and GM420 NFAT binding sites,
respectively (11); TCR-AML1:
GATCCGGCAATGCATGTGGTTTCCAACCGTTAATG AML-1 site derived from the
TCR
enhancer (31); stromelysin AP-1:
GATCTGGATCACCCGCAGCTTGACTCATCCTTGCA, AP-1 site derived from the
stromelysin gene promoter (11); and Sp1:
GATTCGATCGGGGCGGGGCGAGC, an artificial Sp1 site.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The -14-kb IL-3 enhancer can activate the IL-3 promoter in T cell
lines, but its role in normal T cells and other IL-3-expressing cells
has not previously been investigated. To further define the
tissue-specific functions of this enhancer, we mapped DH sites in a
variety of cell types (Fig. 1
). As models
of T cells that efficiently express IL-3, we analyzed freshly isolated
peripheral blood T cells and the Jurkat and CEM T cell lines
(10). We also assayed Hut78 T cells and HMC-1 mast cells
that produce low levels of IL-3, and the T cell line HSB2 that does not
express IL-3 (expression data not shown). As a model of primitive
undifferentiated myeloid cells, we assayed the
CD34+ cell line KG1a that has not yet acquired
the capacity to express IL-3 (10). Cells were stimulated
with PMA/I to activate signaling pathways which in T cells are linked
to the TCR. DNase I analyses of these cells confirmed that the -14-kb
region exists as an inducible DH site in Jurkat and CEM T cells. In
contrast, the -14-kb DH site could not be induced in peripheral blood
T cells, HSB2 or Hut78 T cells, HMC-1 mast cells (please ignore the
blemish), or in KG1a cells (Fig. 1B). The previously
identified -24-kb DH site (14) was seen in this study to
be highly induced in HSB2 cells, and a novel constitutive DH site
existed at -11 kb in Hut78 cells that was not present in any other
cell types.
To determine whether enhancer function mirrored chromatin structure in
these cell types, we assayed the -14-kb IL-3 enhancer in the context
of a minimal promoter upstream of the luciferase gene in plasmids
transfected into CEM, HSB2, and HMC-1 cells (Fig. 1C). As
observed previously, the IL-3 enhancer was efficiently induced in CEM
cells to a level similar to that supported by the GM-CSF enhancer.
However, in contrast to the GM-CSF enhancer, the IL-3 enhancer
functioned poorly in HSB2 cells, and not at all in HMC-1 mast cells.
Because the -14-kb enhancer only functioned in a subset of
IL-3-expressing cells, the need arose to seek regulatory elements in
other regions of the IL-3 locus.
An array of tissue-specific DH sites exists upstream of the IL-3 gene
The human IL-3 gene is embedded within an extensive array of
constitutive DH sites (11). To define the tissue-specific
distribution and properties of these elements, we mapped DH sites
upstream of the IL-3 gene in a wide range of cell types (Fig. 2). The ability of each cell type to
express IL-3 is indicated below the DNase I assay (Fig. 2A).
Upstream DH sites were mapped from an EcoRI site located 2.0
kb downstream of the transcription start site (Fig. 2A).
Note that each lane in Fig. 2A represents just the midpoint
in a DNase I titration that has been previously optimized for the
detection of ubiquitous DH sites downstream of the IL-3 gene, as in a
previous study (12). We detected a very prominent
DH site at -4.1 kb in all T lineage cells, the mast cell line HMC-1,
and the primitive CD34+ myeloid cell lines KG1
and KG1a, but this site was absent in other more mature myeloid cells,
B cell lines, and all nonhemopoietic cell types. Five additional DH
sites were detected at -0.1, -1.0, -1.5, -3.1, and -5.5 kb.
Whereas the -0.1-kb DH site within the IL-3 promoter was present in
both T cells and KG1a cells, the four DH sites at -1.0, -1.5, -3.3,
and -5.5 kb were confined exclusively to T lineage cells. The -4.1-kb
and -1.5-kb sites were present in all T lineage cells, but the DH
sites in the promoter and at -1.0, -3.1, and -5.5 kb were only
detected in some of the T cell lines, and they varied considerably in
their intensities. In a more extensive series of analyses of primary T
cells and cultured T lymphoblasts, the promoter was occasionally
detected as a weak DH site (data not shown). Seeking inducible DH
sites, the same cell lines were also assayed in parallel after
stimulation with PMA/I, but no change in the patterns of sites was
detected by the EcoRI probe (data not shown).
|
B) that pre-exists in T cells as a weak
DH site and is further induced by PMA/I. The -4.5-kb DH site was most
strongly induced in T lymphoblasts cultured from human blood, and was
also present in freshly isolated peripheral blood T cells that had been
stimulated with PMA/I. This site was also induced to variable degrees
in the Jurkat, CEM, HSB2, and HuT78 T cell lines. At least in the case
of Jurkat and HSB2 cells, its induction was suppressed to the
pre-existing level in the presence of CsA, which typically functions in
the repression of cytokine genes by suppressing induction of NFAT. In
those lanes where the upper part of the gel is clearly resolved, it is
possible to see a parallel induction of the DH site in the
NFAT-dependent GM-CSF enhancer, and its suppression by CsA. In KG1a and
HMC-1 cells, DH sites are clearly visible extending from the -4.1-kb
site to the +10-kb GM-CSF enhancer, but little or no induction of the
-4.5-kb DH site is evident in these cells. However, it is possible to
detect weak induction of the -4.5-kb DH site in HMC-1 cells by
examining a darker reproduction of this image (Fig. 2B,
right inset), and examination of the fluorescently stained
DNA suggested that the HMC-1 sample was underrepresented in this
analysis. DH sites upstream of the IL-3 gene encompass conserved DNA elements
As a further aid in the identification of regulatory elements in
the IL-3 locus, we searched for DNA sequences that are conserved
between the mouse and human IL-3 genes (Fig. 2C). This
analysis extended from 20 kb upstream of the human and mouse IL-3 genes
to the GM-CSF promoter downstream of the IL-3 gene. This region
encompasses all of the known genes and regulatory elements, and most of
the known DH sites in the IL-3/GM-CSF locus (Fig. 2C,
top row). Conserved regions are depicted on the two lower
rows of Fig. 2C as either strongly (
) or moderately (
)
conserved. Significantly, the most extensive highly conserved elements
in this locus were not the coding regions, but the IL-3 -4.5-kb DH
site (221 bp 79% or 376 bp 70%), the GM-CSF enhancer at +10 kb (227
bp 80% or 417 bp 76%), and the GM-CSF promoter at +13 kb (361 bp
83%). Only relatively short regions of the IL-3 promoter (93 bp 84%)
and the IL-3 gene (176 bp 62%) were conserved, and the -14-kb IL-3
enhancer was not conserved at all. As a whole, the blocks of conserved
sequences upstream of the IL-3 gene exhibit a striking association with
the locations of DH sites. In addition to the above mentioned highly
conserved elements, the -1.0 and -1.5-kb DH sites span a stretch of
484 bp of 63% conserved DNA, the -3.1-kb DH site is flanked by two
blocks of DNA that contain 152 bp 66% and 301 bp of 61% conserved
sequence, the -4.1-kb DH site includes 49 bp of 81% conserved
sequence, and the -5.5-kb DH site is centered just upstream of an
overlapping block of 290 bp of 65% conserved sequence.
To further investigate inducible mechanisms of IL-3 gene regulation, we
focused on the inducible DH site at -4.5 kb. Shown in Fig. 3 is the 221-bp conserved sequence
encompassing the -4.5-kb site that is 79% homologous with a 218-bp
sequence located 4.7–4.9 kb upstream of the mouse IL-3 gene. The 3'
boundary of the DH site is defined by a StuI site (Fig. 2B, lane 2) that also effectively defines the 3'
border of the conserved DNA sequence (Fig. 3). Given that the DH site
itself spans 
200 bp, there is a striking relationship between the
location of the conserved domain and the DH site. Within the 221-bp
conserved sequence lie three potential NFAT sites that could account
for much of the inducible CsA-sensitive activity of the DH site. Based
on their similarity to known NFAT sites, at least two of the NFAT-like
elements are likely to be conserved in the mouse as NFAT binding sites
(Fig. 3, NFATB and NFATC).
Although there were no other perfect matches to transcription factor
consensus sequences, we identified several sites that are likely to
function as moderate to weak transcription factor binding sites. The
best of these matches included a conserved AP-1 site that overlapped an
AML-1 site, and an Sp1 site adjacent to the central
NFATB motif. Although the AML1 and Sp1 sites were
not conserved in the mouse, an ideal consensus AML1 site was present in
the mouse sequence just 21 bp further downstream than the AML1 site in
the human sequence. Four additional nonideal, partially conserved
consensus sequences are underlined in Fig. 3. These include a second
AP-1-like element (CgaTGTcaTCA) that coresides with the mouse AML1
site, a third AP-1-like element (cGAGGTCA) that exists as the more
ideal tGAGGTCA in the mouse sequence, a fourth AP-1-like element
(TGTCACA) downstream of the Sp1 site, and a GATA-like element
(AGATtA) that most closely matches the GATA3 consensus (Ref.
32 ; the nonconserved bases in these elements are depicted
in lowercase). In the human sequence, the GATA-like element exists as a
palindrome containing two overlapping complementary CAGATTa/tAA
sequences.
|
To test the functions of the -4.5 DH site, a 245-bp
SacI-StuI fragment (SS245; Fig. 4A) encompassing the DH site
and most of the conserved sequences was linked to a -559 to +50
fragment of the IL-3 promoter upstream of a luciferase reporter gene in
the plasmid pXPG-IL3H. The resulting plasmid (pXPG-IL3H-SS245) was
tested in parallel with pXPG-IL3H in transient transfection assays in
Jurkat and CEM cells (Fig. 4B). These cells were either
stimulated (+) with PMA/I or left unstimulated (0). To provide a direct
comparison with the -14-kb IL-3 enhancer, the plasmid pXPG-IL3H-NA330
was also assayed. The SS245 fragment increased the activity of the IL-3
promoter in Jurkat and CEM cells by an average of 2-fold, indicating
that the -4.5-kb region functioned as an inducible enhancer. The SS245
enhancer appeared to be less active than the -14-kb IL-3 enhancer,
which increased the activity of the IL-3 promoter by 3-fold.
|
C). The SS245 fragment
increased the activity of the minimal GM-CSF promoter on average by
40-fold, and was almost as active as the GM-CSF enhancer. The
inducible activity of the SS245 enhancer was abolished in the presence
of CsA. In contrast, the SE546 fragment did not exhibit any enhancer
activity in this assay system.
The SS245 fragment exhibited the classical properties associated with
enhancers in that it also activated the promoter when inserted in the
reverse orientation, and in stable transfection assays could function
at a distance when assayed in the context of the larger 1.3-kb
SacI fragment depicted in Fig. 4A (see Fig. 8).
|
To assess the tissue-specific inducibility of the -4.5-kb IL-3
enhancer, the SS245 fragment was cloned upstream of the constitutively
active TK promoter in a luciferase reporter gene plasmid (pXPG-TK229).
For a control, a 175-bp fragment of the SV40 enhancer, which is widely
active, was also cloned upstream of the TK229 promoter. These plasmids
were transiently transfected into Jurkat T cells, HMC-1 mast cells,
Raji B cells, KG1a myeloid cells, and K562 proerythroid cells and
stimulated with PMA/I (Fig. 5). The
-4.5-kb SS245 enhancer increased inducible promoter activity by 2- to
3-fold in both Jurkat T cells and HMC-1 mast cells, but was inactive in
cell types that do not express IL-3 (KG1a, K562, and Raji cells). The
SS245 enhancer did not significantly influence the constitutive
activity of the TK229 promoter, in contrast to the SV40 enhancer that
was active in all of the cell lines and functioned to increase both
inducible and constitutive promoter activity.
|
To investigate mechanisms of SS245 enhancer function, EMSAs were
performed on probes encompassing the transcription factor consensus
sequences highlighted in Fig. 3 using nuclear extracts of unstimulated
and PMA/I-stimulated Jurkat cells (Fig. 6, A–D). The sequences of the
four AML1/AP-1, NFAT/Sp1, NFAT, and NFAT/AP-1-like elements used as
probes (Fig. 6, A, B, C, and
D, respectively) are shown below these EMSAs, and their
locations depicted as bars above the map of the enhancer in Fig. 6.
Probe A weakly associated with an inducible AP-1-like complex and a
constitutive AML1-like complex. The AML1-like complex with probe A was
supershifted by an AML1 Ab. The probe A AML1 and AP-1 complexes were
also specifically competed by consensus binding sites for these factors
(data not shown). Probe B was a moderately strong binding site for
constitutively expressed Sp1-like factors and for inducible NFAT-like
factors. The probe B Sp1 and NFAT-like complexes were specifically
competed by the Sp1 or the GM430 NFAT-binding oligonucleotides,
respectively. Probe C functioned as a moderate affinity binding site
for inducible NFAT-like factors, and the NFAT-like complex was
supershifted by an NFATC2 Ab (Fig. 6) and specifically competed by the
GM430 NFAT-binding oligonucleotide (data not shown). Probe D, that
included loose matches to NFAT and AP-1/CRE-like elements, generated a
diffuse ladder of weakly inducible complexes that were specifically
inhibited in the presence of the GM170 oligonucleotide that contains
strong binding sites for both NFAT and AP-1, but not by the Sp1
oligonucleotide competitor. A comparison of the relative affinities of
the three NFAT-like elements suggested that probe C was the strongest
NFAT site, binding NFAT 40% as efficiently as the GM430
high-affinity NFAT site (data not shown).
|
were altered to the bases
shown underneath each probe in lowercase and data not shown). However,
note that the precise location of the Sp1 site in probe B is difficult
to predict because it encompasses two overlapping sequences, GTGGCGGAG
and GCGGAGGAG, which both resemble the Sp1 consensus sequence, and the
introduced mutation would strongly suppress binding to both. Also, note
that the first of these two potential Sp1 sites closely resembles
another Sp1-like element located just upstream of the 3' NFAT site C
(GTGGCGGGT, underlined in probe C2). However, probe C2 did not form any
Sp1-like complexes with nuclear extracts (data not shown), which also
makes it more likely that Sp1 contacts GCGGAGGAG rather than
GTGGCGGAG within probe B. Probe C also resembles an NF-
B site,
but it did not form any NF-B-like complexes, and its NFAT complex
was not competed efficiently by an NF-B site (29) from the
E-selectin gene (data not shown).
Purified recombinant proteins were also used to assay binding of NFAT
and AP-1-like complexes to regions A, B, C, and D of the enhancer (Fig. 6). This very sensitive approach uses relatively high concentrations of
factors and detects binding even at very low-affinity sites that are
not detected when nuclear extracts are used. Probes A, B2, C2, and D
were assayed in parallel with the stomelysin gene high-affinity AP-1
site for binding to dimers of the DNA-binding domains of cFos and cJun.
Probe A functioned as a medium-affinity AP-1 site, and probes B2 and D
each contained at least one and potentially two low-affinity AP-1
sites. Probe C2 supported a very low level of AP-1 binding that is
unlikely to be significant. EMSAs with the purified DNA-binding domain
NFATC2 confirmed that probes C2 and B2 encompassed at least one
medium-affinity NFAT site, and mutation of the GG core of NFAT site C
within probe C2 essentially abolished NFAT binding. Probe D contained a
low-affinity NFAT site, whereas probe A did not bind significant
amounts of NFAT.
Because NFAT and AP-1 are often found bound cooperatively to sites that
have the consensus sequence GGANNNNNTGANTCA (13), it
was important to search for any interactions of this nature within the
SS245 enhancer. However, an analysis of the sequence revealed that no
such composite consensus sequences exist within the SS245 enhancer.
Although probes B2 and C2 each contain sites for both NFAT and AP-1,
our previous study (13) suggests that they exist in
conformations that are unlikely to support cooperative binding. For
example, probe D has NFAT and AP-1 elements separated by an extra half
turn of the helix, and is in a configuration that resembles the GM-CSF
enhancer GM170 NFAT/AP-1 element to which NFAT and AP-1 bind quite
independently (13). In the case of probe C, we directly
tested the ability of NFAT to recruit AP-1 to cryptic AP-1 sites, but
only very low level AP-1 binding was detected, and the same level of
AP-1 binding occurred when the NFAT site was deleted (C-NFAT,
Fig. 6).
The SS245 enhancer requires three NFAT sites for function
To study the contribution of the identified binding sites to
enhancer function, we introduced each of the previously tested
transcription factor site mutations (Fig. 6) into the SS245 enhancer.
Plasmid constructs containing the mutated enhancer fragments upstream
of the GM55 promoter in pXPG-GM55 were tested in Jurkat T cells in
transient transfection assays. The effects of these mutations are
summarized below each site in the map in Fig. 6. Loss of the NFAT
sites in regions B and C reduced the activity of these plasmids to 27
and 18% of that supported by the intact enhancer. This residual
activity was not substantially greater than the activity obtained with
just the GM55 promoter, which was 10% of the GM55-SS245 activity. In
contrast, deletion of the AML1, AP-1, or Sp1 sites in regions A and B
had little significant impact on enhancer activity.
To further define DNA regions needed for enhancer function, we created
enhancer fragments that have 5' deletions, as shown in Fig. 7. To broaden the scope of this analysis,
the 5' deletion series commenced at a point 40 bp upstream of the
SacI site (SS285), and proceeded to a point 95 bp upstream
of the StuI site (SS95). The boundaries of the SS146 and
SS95 segments are indicated in Fig. 3. These enhancer fragments were
cloned upstream of the minimal promoter in pXPG-GM55 and transiently
transfected into Jurkat T cells (Fig. 7). Cells were either stimulated
with PMA/I or left unstimulated. Deletions of the 5' flanking sequences
and the AML1/AP-1 region did not significantly affect the SS245
enhancer function, as indicated by the activities of the SS285, MS197,
and SS146 plasmids that did not vary substantially from the SS245
plasmid. However, the deletion of an additional 51 bp resulted in an
almost complete loss of enhancer activity in the SS95 plasmid (12% of
the SS245 activity). This 51-bp region encompassed NFAT site D and
nonideal consensus AP-1 and GATA-like elements (Fig. 3). Based on the
above observations, we conclude that the essential core of the enhancer
is located within a 146-bp region upstream of the StuI site,
and that its activity relies on three distinct regions that each
encompass NFAT sites.
|
We also used a stable transfection assay to examine the
cooperativity between the IL-3 promoter, the -4.5/-4.1-kb region, and
the -14-kb enhancer. For this purpose, the IL-3 promoter (-559 to
+50) fragment was cloned in the pXP1 luciferase plasmid
(11) to give pIL3H. The 1.2 kb BglII fragment
(B1.2) encompassing the -14-kb IL-3 enhancer (11) and the
1.3-kb SacI (S1.3) fragment encompassing both the -4.5-kb
and -4.1-kb DH sites (see Fig. 4A) were cloned in the
BglII and SacI sites upstream of the IL-3
promoter in the pIL3H plasmid to generate pIL3H-B1.2 and pIL3H-S1.3,
respectively. In addition, the B1.2 fragment was cloned in the
BglII site of pIL3H-S1.3 to give pIL3H-S1.3-B1.2. These
constructs were stably integrated into the genome of Jurkat T cells in
two independent transfection assays. Luciferase activity was determined
after stimulation with PMA/I, activities were corrected for average
gene copy number from each transfection, and the activities from each
pair were averaged to create Fig. 8. The
luciferase activities increased stepwise with the addition of each of
the S1.3 and B1.2 elements to the IL-3 promoter. The S1.3 element
enhanced the activity of the IL-3 promoter by 3.4-fold over the
promoter alone. The addition of the B1.2 enhancer in
pIL3H-S1.3-B1.2 increased the activity by a further 2.2-fold.
However, it should be noted that this assay system tended to give
somewhat more variable results overall than the transient transfection
assay system, and was less reliable as a source of quantitative
information.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The combined approaches of studying chromatin structure and DNA
sequence conservation have proved to be invaluable in the search for
regulatory elements in the IL-3 locus. Although the -4.5-kb enhancer
was identified by studying chromatin structure, it could equally well
have been identified by seeking conserved DNA elements had this
information been available at the beginning of this study. The -4.5-kb
IL-3 enhancer represents the most highly conserved region of the IL-3
locus, and it could perhaps be compared with the CNS-1 enhancer between
the IL-4 and IL-13 genes that was originally identified solely on the
basis of sequence conservation (33). The degree of
homology between the human and mouse -4.5-kb enhancers (221 bp 79%)
is similar to that observed for other key enhancers that function in T
cells such as the GM-CSF enhancer (227 bp 80%), the IL-4/IL-13 CNS-1
enhancer (401 bp 84%; Ref. 33), the IL-2 proximal
enhancer/promoter region (580 bp 86%; Ref. 34), and the
human TCR
and TCR enhancers (240 bp 84% and 370 bp 70%;
Ref. 35). A common feature of many such elements is that
they can be identified both as DH sites and as isolated blocks of
conserved sequence distal to the coding regions. In contrast, the
-14-kb IL-3 enhancer is not in the least conserved and could not have
been identified in silico.
The roles of NFAT and CsA in enhancer function
The -4.5-kb IL-3 enhancer shared many properties with the -14-kb IL-3 enhancer and the GM-CSF enhancer. The -4.5-kb IL-3 enhancer was first identified as a tissue-specific inducible DH site that was induced in T cells by stimuli that mimic TCR signaling pathways and repressed by CsA. The basis for the CsA-sensitivity in each case appears to reside on the dependency of these enhancers upon NFAT. The -4.5-kb IL-3 enhancer relies on an array of three NFAT sites for its activity, and deletion of NFAT sites B or C or of a 51-bp region encompassing NFAT site D almost abolished its activity. CsA functions to block the calcineurin-dependent activation of NFAT (36), and it is likely that NFAT is required for both the chromatin remodeling and enhancer function of all three enhancers in the IL-3/GM-CSF locus. NFAT may in fact be the initiator of chromatin remodeling in these enhancers, because we have previously found that NFAT sites from the GM-CSF enhancer are sufficient to create inducible DH sites in transfected Jurkat cells (P. N. Cockerill, unpublished observations). Like many other cytokine genes, the IL-3 gene is suppressed by CsA (36), and the NFAT sites in the promoter (7) and the two upstream enhancers may play a major role in this effect.
Differential regulation of enhancers
One of the most significant findings of this study was the
discovery that the IL-3 gene is regulated by two distinct enhancers
that direct different activities. The -14-kb enhancer is not conserved
in the mouse genome and may be a recent acquisition in the human
genome, serving a highly specialized function. However, it is not
unprecedented to find that genes expressed in T cells are controlled by
more than one enhancer. The TCR
locus encompasses two enhancers that
have overlapping, but also unique, functions (37). The CD8
locus similarly contains more than one enhancer, and these are
distinguished by the developmental stage at which they become active in
the thymus (38).
Although they are both NFAT-dependent, the -14 and -4.5-kb enhancers are governed by distinct patterns of combinatorial regulation that direct very specific patterns of activity. The -14-kb enhancer does not function outside of the lymphoid lineage because it is absolutely dependent upon the lymphoid-specific Oct cofactor OCA-B (14). Unlike the T and B cell lines that we have analyzed (10), we were unable to detect significant OCA-B expression in HMC-1 mast cells (data not shown), and this probably accounts for the inactivity of the -14-kb enhancer in this cell type. In contrast, the -4.5-kb enhancer has a wider range of activity, and it appears to rely primarily on an array of three NFAT sites for its activity. However, it remains to be determined if these NFAT sites are sufficient to account for its functions. NFAT typically functions in strict cooperation with other classes of transcription factor, and further studies will be required to determine whether this is also the case in the -4.5-kb enhancer.
Although we did not identify any function for the AML1 or GATA consensus elements in our in vitro studies, it is nevertheless significant that the -4.5-kb enhancer encompasses the combination NFAT and AML1 binding sites, and a potential GATA site. These three families of proteins are widely expressed in the T cells and myeloid lineage cell types that express IL-3, and potentially direct a specialized pattern of combinatorial regulation that is appropriate for the IL-3 locus. A similar combination of regulatory elements exists within the IL-3 promoter. The presence of a potential GATA3 site in the -4.5-kb enhancer is interesting because GATA3 directs Th2 T cell differentiation (39), and Th2-specific elements such as the IL-5 promoter (40, 41) and the 3' IL-4 enhancer (42) encompass NFAT and GATA3 sites. Hence, the -4.5-kb enhancer may be up-regulated in parallel with IL-4 and IL-5 in Th2 T cells. Conversely, we found that the -14-kb enhancer was down-regulated in the cell line HSB2 that exhibits high level IL-5 expression (43). This presents the interesting possibility that the -4.5 and -14-kb enhancers may be differentially regulated in Th1 and Th2 T cells. If it could be established that the -14-kb enhancer was preferentially active in Th1 rather than Th2 T cells, then this might account for the need to have two distinct enhancers directing IL-3 gene transcription. However, it was not possible to resolve these issues in this in vitro study, and a more physiological system will be needed to identify the true in vivo functions of the -4.5 and -14-kb enhancers.
Other regulatory elements in the IL-3/GM-CSF locus
The IL-3/GM-CSF locus is now known to encompass three inducible enhancers, but there still exist many other elements in this locus for which the function is unknown. There exists an extensive array of constitutive DH sites spanning the IL-3 gene that potentially provide mechanisms controlling the developmental regulation of the IL-3 gene, or for insulating the GM-CSF gene from the IL-3 gene. The presence of the -4.1-kb DH site in KG1, KG1a, and HMC-1 cells and in all T lineage cells was particularly interesting. KG1 is a CD34+ cell line that does not make IL-3, but it may be representative of an early stage of myeloid development (28). Because the -4.1-kb site is the only DH site detected upstream of the IL-3 gene in KG1 cells, it may be the first of the IL-3 DH sites to appear during hemopoietic development. Hence, the constitutive -4.1-kb DH site may represent a developmental marker present in primitive myeloid or lymphoid precursor cells that develop into IL-3-producing cells. In support of this concept, the -4.1-kb DH site can also be detected as an isolated DH site upstream of the IL-3 gene in primitive CD34+ M1 category AML samples (P. N. Cockerill, unpublished observations). However, we were unable to adequately assess the function of the -4.1-kb DH site using in vitro approaches because it lacked classical enhancer activity, and it may have a role that can only be detected in vivo. The -4.1-kb DH site may function by increasing accessibility within the IL-3 locus, thereby promoting subsequent activation of the enhancer and promoter regions by inducible factors. The -1.5-kb DH site may similarly represent a developmental marker of a primed IL-3 locus because this exclusively T cell-specific site is invariably present as a constitutive DH site in all T lineage cells.
In conclusion, the IL-3 locus is likely to be activated in a stepwise process by a combination of developmentally regulated elements that may prime the locus at the level of chromatin structure, and inducible tissue-specific elements that respond to agents that activate the immune system. The role of the downstream elements in the regulation of the IL-3 gene remains to be determined. Our future studies will attempt to determine whether the IL-3 gene is regulated independently of the GM-CSF gene and enhancer, and whether the downstream cluster of ubiquitous DH sites defines one of the boundaries of the functional IL-3 domain within the genome.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Department of Immunology, Duke University Medical Center, Durham, NC 27710.
3 Current address: Division of Medicine, University of Tasmania, Hobart, Tasmania, Australia.
4 Address correspondence and reprint requests to Dr. Peter Cockerill at the current address: Molecular Medicine Unit, Department of Medicine, University of Leeds, Clinical Sciences Building, St. James University Hospital, Leeds LS9 7TF, U.K. E-mail address: medpnc{at}leeds.ac.uk
5 Abbreviations used in this paper: DH, DNase I hypersensitive; CsA, cyclosporin A; PMA/I, PMA plus calcium ionophore A23187; TK, herpes simplex thymidine kinase.
Received for publication March 18, 2002. Accepted for publication June 3, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene expression after stable integration. I. Role of the intronic MAR and enhancer in plasmacytoma cells. J. Biol. Chem. 264:21183. hydroxylase gene may involve the liver-enriched bZIP protein DBP. Genes Dev. 7:1871. locus in gene regulation and T cell development. Immunity 16:453.[Medline]
gene expression in subsets of thymocytes and peripheral T cells. Immunity 9:497.[Medline]
This article has been cited by other articles:
|
|
 |
|
  N. Soto-Nieves, I. Puga, B. T. Abe, S. Bandyopadhyay, I. Baine, A. Rao, and F. Macian Transcriptional complexes formed by NFAT dimers regulate the induction of T cell tolerance J. Exp. Med., April 13, 2009; 206(4): 867 - 876. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Bowers, F. Mirabella, F. J. Calero-Nieto, S. Valeaux, S. Hadjur, E. W. Baxter, M. Merkenschlager, and P. N. Cockerill A Conserved Insulator That Recruits CTCF and Cohesin Exists between the Closely Related but Divergently Regulated Interleukin-3 and Granulocyte-Macrophage Colony-Stimulating Factor Genes Mol. Cell. Biol., April 1, 2009; 29(7): 1682 - 1693. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Sly, J. Kalesnikoff, V. Lam, D. Wong, C. Song, S. Omeis, K. Chan, C. W. K. Lee, R. P. Siraganian, J. Rivera, et al. IgE-Induced Mast Cell Survival Requires the Prolonged Generation of Reactive Oxygen Species J. Immunol., September 15, 2008; 181(6): 3850 - 3860. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Mellstrom, M. Savignac, R. Gomez-Villafuertes, and J. R. Naranjo Ca2+-Operated Transcriptional Networks: Molecular Mechanisms and In Vivo Models Physiol Rev, April 1, 2008; 88(2): 421 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. G. Bert, B. V. Johnson, E. W. Baxter, and P. N. Cockerill A Modular Enhancer Is Differentially Regulated by GATA and NFAT Elements That Direct Different Tissue-Specific Patterns of Nucleosome Positioning and Inducible Chromatin Remodeling Mol. Cell. Biol., April 15, 2007; 27(8): 2870 - 2885. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. V. Johnson, A. G. Bert, G. R. Ryan, A. Condina, and P. N. Cockerill Granulocyte-Macrophage Colony-Stimulating Factor Enhancer Activation Requires Cooperation between NFAT and AP-1 Elements and Is Associated with Extensive Nucleosome Reorganization Mol. Cell. Biol., September 15, 2004; 24(18): 7914 - 7930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-N. Lee, J. Tuckerman, H. Nechushtan, G. Schutz, E. Razin, and P. Angel c-Fos as a Regulator of Degranulation and Cytokine Production in Fc{epsilon}RI-Activated Mast Cells J. Immunol., August 15, 2004; 173(4): 2571 - 2577. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Lee, H. Y. Chung, L. A. Ehrlich, D. F. Jelinek, N. S. Callander, G. D. Roodman, and S. J. Choi IL-3 expression by myeloma cells increases both osteoclast formation and growth of myeloma cells Blood, March 15, 2004; 103(6): 2308 - 2315. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G. Hogan, L. Chen, J. Nardone, and A. Rao Transcriptional regulation by calcium, calcineurin, and NFAT Genes & Dev., September 15, 2003; 17(18): 2205 - 2232. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
