(Received for publication, May 10, 1996, and in revised form, October 24, 1996)
From the Laboratory of Host Defenses, NIAID, National Institutes of Health, Bethesda, Maryland 20892
The eosinophil-derived neurotoxin (EDN) and
eosinophil cationic protein (ECP) are both small, cationic ribonuclease
toxins that are stored in and secreted by activated human eosinophilic leukocytes. We have previously shown that optimal expression of the EDN
gene is dependent on an interaction between an intronic enhancer
element or elements and the 5 promoter region. Here we present
evidence demonstrating that the gene encoding ECP is regulated in an
analogous fashion and that an intronic enhancer element functioning in
both genes is a consensus binding sequence for the transcription factor
NFAT-1. Our initial results demonstrate that one or more nuclear
proteins isolated from human promyelocytic leukemia (HL-60) cells bind
specifically at this consensus site (5
-GGAGAA-3
) within the intron of
the EDN gene and that disruption of this sequence reduced the
characteristic 20-30-fold increase in reporter gene activity observed
with the tandem EDN promoter/exon 1/intron construct to background
levels. The NFAT-1 consensus site in the ECP gene differs from that
found in the EDN gene by a single nucleotide (5
-GGAGAG-3
); the
conversion of the 3
G to an A resulted in a further enhancement of the
reporter gene activity supported by the ECP promoter/exon 1/intron
construct. Interestingly, no "supershift" was observed in gel shift
assays performed in the presence of anti-NFAT-1 antiserum, suggesting that a nuclear protein other than NFAT-1 may be acting at this consensus site.
Eosinophil-derived neurotoxin
(EDN1/RNS2) and eosinophil
cationic protein (ECP/RNS3) are members of the mammalian
ribonuclease (RNase A) superfamily and are two of four proteins found
in the large specific granules of human eosinophilic leukocytes (1, 2).
The genes encoding EDN and ECP are 90% homologous to one another and
both include two exons; each gene contains a noncoding exon 1, separated by a single intron from the coding sequence in exon 2. This
gene structure is characteristic of the ribonuclease gene family
(3-6). In our previous work, we have demonstrated that optimal
expression of the EDN gene is dependent on interaction between the 5
promoter region and the single intron (7). Further analysis of the EDN
gene demonstrated that a significant portion of this intron-mediated
enhancer activity resided within the first 60 base pairs of the intron,
which includes consensus binding sites for both AP-1 and NFAT-1
transcription factors (7).
The NFAT-1 sequence is a consensus binding site for nuclear factor of
activated T cells, preexisting (formerly known as NF-ATp), which was
originally described as a cyclosporin-sensitive T lymphocyte-specific transcription factor (8-11) involved in the regulation of gene expression of several cytokine genes, including murine interleukin-2 (8), interleukin-3 (12), interleukin-4 (13-16), interleukin-5 (17),
gp39 (18), granulocyte-macrophage colony-stimulating factor (19, 20),
and tumor necrosis factor (21). NFAT-1 exists as a cytosolic
protein that is dephosphorylated and translocated to the nucleus
following T lymphocyte activation by a calcium/calcineurin dependent
mechanism, where it interacts cooperatively with Fos-Jun dimers (11,
22, 23) to regulate gene expression. NFAT-1 is also involved in the
regulation of expression of cytokine genes in B lymphocytes (24-26)
and in mast cells (27) and of granulocyte-macrophage colony-stimulating
factor in endothelial cells (12). NFAT-1 immunoreactivity has been
observed in murine neuronal cell lines (28), suggesting an even wider
role for this factor as a regulator of gene expression.
Here, we find that optimal expression of the ECP gene also depends on the presence of an intronic enhancer element, and we present evidence demonstrating a role for an NFAT-1 consensus sequence and its binding protein in the regulated expression of both EDN and ECP.
All cell lines used in this investigation were
obtained from the American Type Culture Collection (Rockville, MD). The
HL-60 (human promyelocytic leukemia) and K-562 (human chronic
myelogenous leukemia) cell lines were cultured in RPMI 1640 (Biofluids,
Inc., Rockville, MD) supplemented with 10% heat-inactivated fetal
bovine serum (Hyclone Laboratories, Inc., Gaithersburg, MD), 2 mM L-glutamine (Quality Biologicals, Inc.,
Gaithersburg, MD), and 100 units/ml penicillin plus 100 µg/ml
streptomycin (Quality Biologicals). The HL-60 clone 15 promyelocytic
leukemia cells were grown in RPMI 1640 with fetal bovine serum,
L-glutamine, penicillin/streptomycin supplemented with 25 mM HEPPSO
(N-[2-hydroxyethyl]piperazine-N-[2-hydroxy-propanesulfonic acid] (Sigma), and maintained at pH 7.6. All cells
were grown at 37 °C with 5% carbon dioxide in a humidified
incubator.
The reporter gene used in these
studies was chloramphenicol acetyltransferase (CAT) as found in the
pCAT-basic plasmid vector (Promega, Madison, WI). Preparation of the
EDN reporter gene constructs were as described previously (7); the
construct EDN3-CAT described here is identical to that described as
EDN-PrExInCAT in the aforementioned publication. The ECP reporter gene
constructs (see Fig. 2B) were prepared as described (7) from
a bacteriophage with an ECP-encoding human genomic DNA
insert;2 the transcriptional start site of
the ECP gene was determined to be at the analogous site to the start
site in the gene encoding EDN.3
Sequence Mutations
The NFAT-1 consensus sequence
(nucleotides 126-131) found in the promoter of EDN was altered by
overlapping mutagenesis (29) of EDN3-CAT, creating EDN3M-CAT (see Fig.
3A) using the following oligonucleotide primer pairs: A and
D, and B and C. Primer A: 5-CTG CAG GCA GCA TAT AGT TTT CAT-3
,
nucleotides
312 to
288; primer B, 5
-CTG TAA GAA AAG AAG AGA AGT
AAC-3
, nucleotides 172-149; primer C: 5
-GGG GCA GCA ACT GAG TTC TCC
GAG AGC TGA CGT TAG, nucleotides 111-146; and primer D: 5
-CTA ACG TCA
GCT CTC GGA GAA CTC AGT TGC TGC CCC, nucleotides 146-111.
The NFAT-1 consensus sequence found in the promoter of EDN was
introduced into the homologous site in the ECP promoter (nucleotides 126-131) also by overlapping mutagenesis (29) of ECP3-CAT, creating ECP3A-CAT using the following oligonucleotide primers (E and F) paired
with primers A and D, respectively, as described above. Primer E:
5-AGG GCA GCA CCT GAG GGA GAA GTG AGC TGA AGT TAG, nucleotides
111-146; and primer F: 5
-CTA ACT TCA GCT CAC TTC TCC CTC AGG TGC TGC
CCT, nucleotides 146-111.
Cells from the human cell lines indicated were
grown to a density of 0.5 1 × 106 to 1 × 106 cells/ml, harvested by centrifugation, washed, and
resuspended at 30 × 106 cells/ml in growth medium as
described above. Ten µg of uncut reporter gene construct along with
10 µg of uncut pCMV (cytomeglovirus)--galactosidase plasmid
(Promega) were added to a 0.5-ml cell suspension in an electroporation
cuvette (0.4-cm gap) (Bio-Rad). Electroporation, harvest, CAT and
-galactosidase assays were as described previously (7).
Nuclear protein fractions
were obtained from cells of the HL-60 human promyelocytic leukemia cell
line for use in electrophoretic mobility shift assays as described
below. Cells (1 × 108 in log phase growth) were
harvested (2000 rpm for 5 min) and washed once in sterile, ice-cold
phosphate-buffered saline without calcium or magnesium. The washed
cells were resuspended in 1.25 ml of cold buffer containing 10 mM HEPES, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, 0.3 M sucrose, 0.1 mM EGTA, 0.5 mM
phenylmethylsulfonyl fluoride (Sigma), and protease
inhibitors (Boehringer Mannheim) as follows: 50 µg/ml antipain
dihydrochloride, 0.7 µg/ml pepstatin, 330 pg/ml phosphoramidon, 2 µg/ml aprotinin, 60 µg/ml chymostatin, 0.5 µg/ml leupeptin, 40 µg/ml bestatin, 10 µg/ml E-64, and 1 mg/ml Pefabloc. Resuspended
cells were harvested and resuspended in 0.5 ml of the same buffer and
lysed in with 15 strokes with a B type pestle in a Dounce homogenizer.
The supernatant containing the cytosolic contents from the lysed cells
was removed, and the pelleted nuclei were resuspended in 0.3 ml of
ice-cold buffer as described above, with the addition of 25% glycerol
and 0.2 mM EDTA. The resuspended nuclei were shaken (gentle
agitation) for 30 min at 4 °C, and the resulting supernatant,
harvested after centrifugation (12,000 rpm for 15 min at 4 °C), was
dialyzed in a 10-kDa pore size membrane (Spectrum Laboratories,
Houston, TX) at 4 °C against more than 100 volumes of buffer
containing 20 mM HEPES, pH 7.9, 20% glycerol, 10 mM KCl, 0.5 mM dithiothreitol, 0.2 mM EDTA, and 0.5 mM phenylmethylsulfonyl
fluoride for 3-5 h. The dialyzed supernatant was clarified by
centrifugation, and protein concentration was measured using the BCA
protein assay reagent system (Pierce). Nuclear extracts were aliquoted
and stored at 80 °C.
Complementary
oligonucleotides E and F as described above were hybridized by heating
to 65 °C, followed by gradual cooling (15 min) to 37 °C. The 5
ends of each strand were phosphorylated with 50 µCi
[
-32P]ATP (6000 Ci/mmol; DuPont NEN) and T4
polynucleotide kinase (Boehringer Mannheim) purified by
phenol-chloroform extraction and redissolved in 10 mM Tris,
pH 7.5, with 1 mM EDTA. One µl of the radiolabeled
oligonucleotide probe was added to 17 µl of gel shift incubation
buffer (Stratagene, La Jolla, CA) and 5 µl of a 1 µg/µl
preparation of nuclear proteins described above (or distilled
H2O as a negative control). Competing unlabeled
oligonucleotide was added in 100-fold excess prior to the addition of
the labeled probe. After a 30-min incubation at room temperature, 2 µl of 0.1% bromphenol blue dye was added, and 5 µl of each sample
were evaluated by electrophoresis in a 6% polyacrylamide DNA
retardation gel prepared with 0.5 × Tris-borate-EDTA (TBE)
(Novex, San Diego, CA) with 0.25 × TBE running buffer at a
constant 100V, 6-15 mA at room temperature.
Supershift analysis was performed using a polyclonal anti-NFAT-1 antiserum described by McCaffrey et al. (Ref. 10; Upstate Biotechnology, Inc., Lake Placid, NY). EMSA was performed as described above; 1 µl of a 1:1000 dilution of the NFAT-1 antibody was added 10 min after addition of the probe to the nuclear extract and then allowed to incubate an additional 30 min. Positive control included the NFAT-1 recognition sequence described in Ref. 10 incubated with the same dilution of antiserum and nuclear extract from Jurkat (human T lymphoma) cells, also as described (10).
Shown in Fig.
1A is a comparison of the genomic sequences
of EDN and ECP; exons 1 and 2 are demarcated by brackets. Also
indicated are the sequences of probes and consensus sites that are
described in the text. Fig. 1B diagrams the regions of ECP
gene used to create the reporter gene constructs. Segments including
the 5 promoter region (ECP1), promoter with exon 1 (ECP2) and promoter with exon 1 and intron (ECP3) were inserted 5
to the CAT reporter gene. Various human hematopoietic cell lines were transfected with
these constructs, and the ability of each to support reporter gene
activity was measured. As shown in Fig. 1C, the ECP promoter (ECP1-CAT) and promoter with noncoding exon 1 (ECP2-CAT) supported only
slightly more reporter gene activity than that supported by
(promoterless) CAT alone in the HL-60 promyelocytic leukemia cell line
(1.2-fold and 1.1-fold, respectively). Addition of the intron
(ECP3-CAT) led to a 9-fold increase in reporter gene activity. Similar
results were obtained in two additional cell lines tested; whereas
ECP1-CAT or ECP2-CAT supported relatively small increases in activity,
the activity of ECP3-CAT was measured at 8-fold and 18-fold over CAT in
the HL-60 clone 15 and K-562 cell lines, respectively. These results
suggest that optimal expression of the ECP gene depends on a sequence
element (or elements) present in the single intron, analogous to,
albeit less dramatic than, the results obtained previously for the gene
encoding EDN (Ref. 7; Fig. 1C).
A, sequence comparison of EDN and ECP
genes. Features indicated include consensus CAAT and TATA box sequences
within the 5 promoters (overlined sections a and
b, respectively), exons 1 and 2, and the single intron.
Consensus transcription factor binding sites (AP-1, NFAT-1, and PU.1) are overlined and labeled. Probes used
in the EMSA (Fig. 2) are shown as overlined sections c and d. B, schematic of the genomic structure of the
ECP gene. Areas of ECP sequence denoted by the filled bars
below were introduced 5
to the CAT reporter gene; ECP1, 5
promoter
region; ECP2, 5
promoter region with exon 1; ECP3, 5
promoter region
with exon 1 and intron. C, relative CAT activity of the ECP
reporter gene constructs described in B transfected into
cells of the HL-60, clone 15, and K-562 human hematopoietic cell lines.
The EDN3-CAT construct contains the promoter/exon 1/intron of the EDN
gene, as described in Ref. 7. The lane marked CAT indicates
the relative activity of the promoterless reporter gene construct,
pCAT-basic. The analysis of variance test indicates significant
differences between the values obtained for ECP3-CAT and those obtained
for EDN3-CAT in all three cell lines at p < 0.05 (n = 2 for HL-60 and clone 15; n = 3 for K-562).
EMSA/NFAT-1 Consensus Binding Sequence within the Intron of the EDN Gene
We have shown previously that a significant portion of the intron-mediated enhancer activity of EDN resides in the first 60 base pairs (nucleotides 68-133) of the intron (7). In order to assess the presence of specific binding sites, EMSA or gel shift analysis was performed with nuclear proteins isolated from the promyelocytic leukemia HL-60 cells. Nuclear proteins were incubated with radiolabeled double-stranded oligonucleotide probes c and d corresponding to nucleotides 68-103 and 104-137, respectively (Fig. 1A). As shown in Fig. 2A, two distinct bands migrating more slowly than the unbound probe (probe d) were detected after incubation with the HL-60 nuclear proteins. These bands could not be detected in incubations that included a 100-fold excess of unlabeled probe, suggesting recognition of a specific sequence element or elements by one or more nuclear proteins in this extract. Additionally, these bands could not be detected in incubations that included a double-stranded oligonucleotide that encoded the NFAT-1 consensus site (nucleotides 126-131). No competition was seen in the presence of a random 34-base pair double-stranded oligonucleotide or from oligonucleotide probes corresponding to any of the other sequences within the original probe (data not shown), suggesting the presence of a specific nuclear protein or proteins binding to the NFAT-1 consensus site (nucleotides 126-131). No specific binding was observed in gel shift experiments in which nuclear protein fractions were incubated with the double-stranded oligonucleotide probe c (nucleotides 68-103; data not shown).
In order to determine whether the protein binding to the consensus site could be identified as immunoreactive NFAT-1, supershift analysis was performed using a polyclonal antiserum raised against purified NFAT-1 protein (10). Although a supershift was observed using the anti-NFAT-1 antiserum in conjunction with the consensus site probe and nuclear extract described by McCaffrey et al. (10), no supershift was observed using the anti-NFAT-1 antiserum with the HL-60 nuclear proteins and probe d (Fig. 2B), suggesting the possibility that another protein, perhaps another member of the NFAT family of transcription factors, binds specifically to this intron-based consensus sequence.
Reporter Gene Activity/Mutagenesis of the EDN and ECP Intronic NFAT-1 Consensus SequencesIn order to determine the functional
significance of this consensus site, each nucleotide was altered
(5-GGAGAA-3
to 5
-TTCTCC-3
), creating EDN3M-CAT (Fig.
3A). The HL-60, K562, and clone 15 cell lines
were transfected with this construct, and the reporter gene activity
was compared directly to that supported by EDN3-CAT and CAT alone. We
have previously shown that EDN3-CAT supported reporter gene activity
20-30-fold over that of the promoterless CAT construct (Ref. 7; Fig.
1C). As shown in Fig. 3B, the reporter gene
activity of EDN3M-CAT was reduced to that observed for promoterless CAT alone in both the HL-60 and K562 cells. In the clone 15 cell line, disruption of the NFAT-1 site did not totally abolish reporter gene
activity but reduced it to 50% of that observed for the EDN3-CAT construct with the consensus site intact.
Interestingly, the NFAT-1 consensus site in the analogous site in the
ECP gene differs by a single nucleotide, with a G, rather than an A, at
the 3 end (see Fig. 1A). To evaluate the role played by
this single base pair change in this apparently crucial site, a 3
G to
A conversion was introduced into the consensus site encoded by the ECP
intron, creating ECP3A-CAT (Fig. 3A). The HL-60, K562, and
clone 15 cell lines were transfected with this construct, and the
reporter gene activity was compared directly to that supported by
ECP3-CAT and CAT alone. We have previously shown that ECP3-CAT supported reporter gene activity 9-fold over that of the promoterless CAT (Fig. 1C); as shown in Fig. 3C, the reporter
gene activity of the altered construct was significantly greater than
that of ECP3-CAT, from 15- and 17-fold over CAT in the HL-60 and K562 cell lines, respectively, to 28-fold over CAT in clone 15, similar to
the levels of enhancement observed for EDN3-CAT in this cell line.
In our previous work we showed that that the presence of the
single intron was necessary for optimal expression of the gene encoding
the eosinophil ribonuclease EDN (7). In this work, we have shown that
the gene encoding the eosinophil ribonuclease ECP is regulated in an
analogous fashion and that a consensus binding sequence for the
transcription factor NFAT-1 (8-11) found in the introns of both EDN
and ECP genes play a crucial role in enhancing their expression.
Although originally described as specific to T lymphocytes, NFAT-1 has
been found to participate in transcriptional regulation in mast cells,
endothelial cells, and B lymphocytes (12, 24-27). This work suggests
the possibility of a role for NFAT-1 and/or related proteins (30, 31)
in the transcriptional control of genes expressed in human
granulocytes. Interestingly, whereas the NFAT-1 sequence accounts for
the entire intron-mediated enhancer activity found in both the
promyelocytic HL-60 cells and the chronic myelogenous leukemia K-562
cells, this is not the case in the eosinophilic variant of HL-60, clone
15; in the clone 15 cells, elimination of the NFAT-1 consensus in the
EDN intron (EDN3M-CAT) reduced the intron-mediated enhancer activity by
only 50%. Similarly, the reporter gene activity of the ECP3A-CAT construct, with the 3 G to A conversion of the consensus site in the
ECP intron, showed the most dramatic enhancement in the clone 15 cells.
The results obtained with the EDN3M-CAT construct suggest the presence
of yet another (perhaps eosinophil-specific) active intronic enhancer
element participating in the promoter/intron interaction in the clone
15 cell line; the results obtained with the ECP3A-CAT construct suggest
a cooperative interaction between this putative site and the 3
A
version of the NFAT-1 consensus sequence.
The NFAT-1 consensus sequence situated in the EDN intron, 5-GGAGAA-3
,
can also be found in the regulatory regions of genes encoding other
granulocyte proteins, including eosinophil peroxidase (32), lactoferrin
(33), Charcot-Leyden crystal protein (34), and major basic protein
(35); this consensus sequence may have a wider role in the regulation
of gene expression in the granulocyte lineages than has been previously
appreciated. Particularly interesting is the potential relationship
between this consensus sequence and the rare genetic disease,
neutrophil-specific granule deficiency, which is a disorder of
transcription of granule proteins affecting both the neutrophilic and
eosinophilic granulocyte lineages (36, 37). Although NFAT-1 is
expressed in a variety of tissues, the results of the supershift
analysis suggest a different identity for this protein; it remains to
be seen whether this consensus sequence binding protein is a variant of
NFAT-1, a member of the NFAT family of transcription factors (34, 35),
or something completely unique. Future studies will seek to identify
this protein, to determine its role in the regulation of EDN and ECP
gene expression, and to assess its involvement in the physiology and
pathophysiology of the human granulocyte lineages.