(Received for publication, August 9, 1996, and in revised form, October 17, 1996)
From the Departments of Molecular Cardiology and
§ Cancer Biology, Research Institute, Cleveland Clinic
Foundation, Cleveland, Ohio 44195 and the ¶ Division of
Cardiology, University of Cincinnati Medical Center,
Cincinnati, Ohio 45267
Primate -chymases are mast cell neutral
proteases that are involved in regulating several regulatory peptides
including angiotensin II. Because of significant substrate specificity
differences among the chymase group of enzymes, animal models that
overexpress primate chymases are crucial for delineating the in
vivo function of these enzymes. Activation of
-prochymase
requires processing enzymes and proteoglycans found in mast cell
secretory granules. Thus, the development of models overexpressing
active primate chymase requires a mast cell-specific promoter. We show
that the 571-base pair (bp) 5
-upstream sequence of the baboon chymase
gene, which encodes an
-chymase, coupled to the prokaryotic
lacZ gene allows the targeting of
-galactosidase to mast
cells in transgenic mice. Tissue expression of the transgene is similar
to the expression of the endogenous mouse
-chymase mouse mast cell
protease-5. A mouse mast cell line that endogenously expresses mouse
mast cell protease-5 (JKras mast cells) also selectively
supports the expression of this transgene. In vitro
transcription studies in JKras mast cells shows the
critical role of a GATA cis-regulatory motif in baboon
chymase promoter, located ~430-bp upstream of the transcription start
site. These results suggest that the 571-bp domain of the baboon
chymase promoter contains most, if not all, of the mast cell-specific
region of the promoter. We describe here for the first time a promoter
that directs expression of transgenes specifically to mouse mast cells.
This promoter should be generally applicable for dominant expression of
mast cell regulatory proteins.
Mast cells originate from hematopoietic stem cells, but mature
mast cells reside in mucosal and connective tissue (1). Through the
release of proteases, cytokines, and other chemical mediators, mast
cells participate in a variety of immunological, inflammatory, and
cardiovascular responses. Mast cell proteases include chymases,
tryptases, and carboxypeptidase A (2). Our recent studies and those of
other investigators suggest that chymases can activate and inactivate
regulatory peptides and thus can produce complex effects on the
cardiovascular system. For example, some chymases, e.g.
human, baboon, and dog chymases, form angiotensin II efficiently (3,
4), whereas other chymases, e.g. rat chymase-1, degrade
angiotensin II (5, 6). Angiotensin II is a potent vasoconstrictor
peptide (7) and a myocyte growth factor (8). Recent phylogenetic
studies indicate that mammalian chymases occur in two distinct
isoenzyme groups, and
(6). A single chymase gene
(chm gene) encoding an
-chymase has been identified in
all mammals studied, including human, baboon, dog, rat, gerbil, and
mouse (6, 9-12). In contrast,
-chymases are species-specific. Mice,
rat, and gerbil contain four, two, and one
-chymase-encoding
chm genes, respectively (6, 11).
-Chymases have not been
demonstrated to be present in humans and baboons. Substrate specificity
studies indicate that
- and
-chymases differ in their
specificities and that there are more subtle substrate specificity
differences between some
-chymases (3-6, 13).
Although the overall function of mast cells can be explored in mast
cell-deficient mice, the in vivo role of -chymases cannot be delineated because highly specific inhibitors of these proteases are
not currently available. Also, subtle differences in substrate specificity between mammalian
-chymases make questionable the use of
rodent models for the study of primate
-chymase function. One
approach toward understanding the in vivo role of primate
-chymases is to create transgenic models that overexpress
functionally active forms of these enzymes. Our recent studies show
that the chymase propeptide sequence is important in the correct
folding of this protease within the cell (14) and activation of
prochymase requires heparin and dipeptidylpeptidase I, which occur
abundantly in mast cells (15). Therefore, to assess the effect of
dominant expression, it is necessary to target the zymogen to mast
cells to ensure processing and proper tissue localization of chymase. The unavailability of a promoter that selectively targets the mast cell
in vivo has hampered investigation of the in vivo
role of these mast cell proteases.
We recently cloned the baboon chm gene, including its
~2.5-kb1 5-upstream sequence. The
5
-upstream sequence of this gene is highly homologous to the
corresponding region of the human chm gene described by
Caughey et al. (16). In this report we show that the 571-bp
5
-upstream sequence of the baboon chm gene coupled to the
prokaryotic lacZ gene allows the targeting of the
-galactosidase reporter to mast cells in transgenic mice. We also
show that tissue expression of the lacZ gene directed by the
baboon chm 571-bp 5
-upstream region is similar to the
endogenous pattern of expression of the mouse
-chymase mmcp5 and
that this promoter is highly active in JKras mast cells, a
mouse mast cell line that endogenously expresses mmcp5. Furthermore,
in vitro transcriptional studies in this mouse mast cell
line demonstrate the critical role of a GATA cis-regulatory
motif in baboon chm promoter-directed expression of the
reporter gene. These studies provide a novel tool for the specific
expression of genes in mast cells that is likely to be crucial in
uncovering the in vivo role of primate chymases as well as
other mast cell proteins and provide insight into cis-trans interactions that may be important in baboon chm gene
expression in mast cells.
Southern blot analysis indicated
the presence of a single 7-kb EcoRI-digested band from
baboon genomic DNA that hybridized to a 32P-labeled human
chymase cDNA probe. EcoRI-digested 6-8-kb fragments of
baboon genomic DNA were cloned into a phage gt10 vector system (Stratagene, La Jolla, CA) as per manufacturer's protocol. The 7-kb
insert, which strongly hybridized to the human chymase cDNA probe,
was subcloned into a pBluescript KS+ phagemid (Stratagene).
The nucleotide sequence of the overlapping fragments was determined in
both directions.
The 5-end of baboon heart-derived chymase mRNA was defined by
primer extension analysis (17). Poly(A)+ RNA, prepared from
the baboon left cardiac ventricle, was incubated for 1 h at
42 °C in a reaction buffer containing Moloney murine virus reverse
transcriptase (Boehringer Mannheim) and a 30-bp 32P-labeled
oligonucleotide derived from the 3
-end of the first exon of the baboon
chm gene. After phenol/chloroform extraction and ethanol
precipitation, the sample was electrophoresed on a 6% polyacrylamide
gel, and the size of the prominent radiolabeled DNA fragment extended
onto the primer was determined.
The
lacZ gene was released from a pSV--galactosidase plasmid
by SalI and BamHI digestion (the SalI
site is created by PCR and located just before the ATG translation
start site). Two different transgenes were constructed by ligation of
2,426- or 571-bp fragments of the baboon chm gene's
5
-upstream region with the lacZ gene. The linearized
transgene constructs were microinjected (18); ~2 pl of the DNA
solution (~1000 copies) was microinjected into one pronucleus of each
mouse (FVB) F2 embryo. Injections were made by the Transgenic Core
Facility of the Cleveland Clinic Foundation. Two-cell stage embryos
were transferred to the oviducts of 0.5-day post-coitus pseudopregnant
Swiss FBR females. The offspring were screened for transgene
integration by Southern blot analysis of tail DNA using a
32P-labeled 3.34-kb prokaryotic lacZ gene as
probe.
A previously described histochemical
staining procedure was used to assess -galactosidase activity (18).
Three-week-old transgenic mice as well as their nontransgenic
littermates were sacrificed. Tissue samples from each mouse were fixed
for 40-60 min in 2% formaldehyde, 0.2% gluteradehyde, 0.02% Nonidet
P-40, and 0.01% sodium deoxycholate in phosphate-buffered saline,
embedded in the embedding medium OCT and frozen on dry ice. Serial
6-µm tissue sections were cut with a cryostat, loaded on
charge-coated slides (Fisher), and refixed in the same solution for 10 min. After three washes in phosphate-buffered saline, sections were submerged in 0.1 M phosphate buffer, pH 7.3, containing 1 mg/ml 4-chloro-5-bromo-3-indolyl-
-D-galactoside
(Boehringer Mannheim), 2 mM MgCl2, 10 mM K3Fe(CN)6, and 10 mM
K4Fe(CN)6. Incubation was carried out for 3-4
h at 37 °C. After
4-chloro-5-bromo-3-indolyl-
-D-galactoside staining,
sections were dehydrated and mounted in Permount (Fisher). Adjacent
slides were stained with hematoxylin and eosin or with 1% toluidine
blue (T-blue).
Total RNA was isolated from 1 × 108 cells using an RNAzol kit (TEL-TEST, Friendswood, TX). 20-µg RNA samples were size-fractionated on a 1% formaldehyde-agarose gel and transferred to a nitrocellulose filter (19). After UV cross-linking, the blot was hybridized with 32P-labeled human chymase cDNA at high stringency.
Cell Lines and Tissue CultureMouse mastocytoma P815 cells, mouse erythroid leukemia (MEL) cells, and NIH3T3 (3T3) fibroblasts were obtained from the American Type Culture Collection. JKras mast cells (20) were obtained from Dr. James N. Ihle (St. Jude Children's Research Hospital, Memphis, TN). P815 mast cells and MEL cells were maintained in RPMI 1640 medium containing 10% fetal bovine serum. JKras mast cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum and murine interleukin-3 (20 units/ml). 3T3 fibroblasts were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. Cells were kept at 37 °C in an atmosphere of 5% CO2 and split (~1:4) every 5 days.
Construction of Recombinant PlasmidsBaboon chm
promoter-CAT chimeras were made by fusing varying lengths of the baboon
chm gene's 5-upstream sequence to the coding sequences for
CAT. These baboon chm promoters of varying lengths were
generated by PCR amplification. For each construct two primers were
used; each contained either a terminal PstI site or a
SalI site to facilitate cloning into the PstI and
SalI sites immediately upstream of the CAT reporter gene in
the plasmid pSV40-CAT enhancer (Promega, Madison, WI). The number
following the
in Fig. 1 indicates the end of the 5
-upstream
sequence with respect to the transcription-initiation site (designated
+1) was determined by primer extension analysis. In addition to the
specified amounts of the 5
-upstream sequence, these constructs
contained the transcription-initiation site and 30 bp of the
5
-upstream region between the transcription-initiation site and the
putative translation start site of the baboon chm gene.
Transient Cell Transfection and CAT Activity Assays
Transient transfection of 3T3 fibroblasts was performed
using the calcium/phosphate co-precipitation method (19).
JKras mast cells and P815 mast cells were transiently
transfected by the DEAE-dextran method (19). As an internal control for
these transfections, the pSV-lacZ plasmid was co-transfected
with the test plasmid (1:5). The cells were harvested and ruptured by
three freeze-thaw cycles. One-half of the extract was heated for 10 min
at 65 °C to destroy endogenous deacetylating activity. After removal
of the cellular debris, the heat-treated extract was assayed for CAT
activity. The remaining extract was assayed for -galactosidase activity. Based on
-galactosidase activity, CAT activity in each sample was adjusted for between plate and between experiment
variability in transfection efficiency. Transient transfection of COS-1
cells was performed using the DEAE-dextran method (19). The
pSV-lacZ plasmid was co-transfected with the test plasmid
(1:5) as an internal control. Human GATA-2 cDNA plasmid, a gift
from Dr. S. H. Orkin (Harvard Medical School, Boston), was
co-transfected with the
571-bchm-CAT construct to
determine transcriptional activation of this promoter construct by
human GATA-2. CAT activity in each sample was adjusted for between
plate and between experiment variability in transfection
efficiency.
JKras mast cell
nuclear extracts were prepared as described previously (21). Nuclear
extracts were dialyzed overnight at 4 °C against 20 mM
HEPES buffer, pH 7.9, containing 100 mM KCl, 1 mM EDTA, 1 mM EGTA, 0.5 mM
dithiothreitol, 1 µg/ml aprotinin, 0.5 mM
phenylmethylsulfonyl fluoride, and 25% glycerol, and stored at
70 °C. In vitro transcription reactions were performed
with 2 nM supercoiled template DNA and 10-15 µl of
nuclear extract (10-15 mg/ml) in a 50-µl reaction mixture containing
12 mM HEPES buffer, pH 7.9, containing 50 mM
KCl, 25 mM (NH4)2SO4,
7.5 mM MgCl2, 0.1 mM EDTA, 0.3 mM dithiothreitol, 2 units of RNasin, 12% glycerol, 600 µM ATP, 600 µM UTP, 600 µM
GTP, 25 µM CTP, and 10 µCi of
[
-32P]CTP (Amersham Corp.). Transcription reactions
were allowed to proceed for 60 min at 30 °C and terminated by the
addition of 0.2% SDS, 10 mM EDTA, and 100 µg of tRNA
carrier/ml. RNA in the transcription mixture was extracted with 200 µl of phenol/chloroform/isoamyl alcohol (25:24:1), ethanol
precipitated, and dissolved in diethyl pyrocarbonate-treated Tris/HCl
buffer, pH 8.0, containing 1 mM EDTA. The RNA samples were
separated on a 1% formaldehyde-agarose gel in 1 × MOPS buffer,
transferred to a nitrocellulose filter, and subjected to
autoradiography for 16 h.
Electrophoretic
mobility shift assays were performed as described (22). Binding
reactions were performed with 20 µg of JKras mast cell
nuclear extract and ~0.5 ng of 32P-labeled
oligonucleotide corresponding to a fragment (417 to
441 bp) of the
baboon chm promoter (100,000 cpm) in 20 mM HEPES buffer, pH 7.9, containing 50 mM KCl, 1 mM
EDTA, 1 mM dithiothreitol, 5% glycerol, 100 mg/ml bovine
serum albumin, and 1 µg/ml poly d(I-C). To study the effect of
competitor oligonucleotides on the binding of the
32P-labeled oligonucleotide to JKras mast cell
nuclear trans-acting factors, a 10-, 30-, or 60-fold excess
of competitor oligonucleotide was added to the reaction mixture 10 min
before the addition of the radiolabeled probe. 32P-Labeled
DNA-protein complexes were resolved from unbound radiolabeled DNA on
6% polyacrylamide gels run in 22.5 mM Tris borate buffer, pH 8.3. The gels were dried under vacuum and autoradiographed for ~16
h.
Single-stranded cDNAs
were prepared from total RNA from JKras mast cells, MEL
cells, P815 cells, 3T3 fibroblasts, and mouse spleen and from tissues
from the transgenic mice and their nontransgenic littermates using
oligo(dT) primers and Moloney murine virus reverse transcriptase. The
resulting single-stranded cDNAs were amplified using a 0.5 µM concentration of each primer. For mmcp5,
5-TTCTGAGAACTACCTGTCGGCC-3
was used as the sense strand, and
5
-GAATTGTGTCGAAAACTGGTGAAGTG-3
was used as the antisense strand. For
mouse GATA-1, 5
-CAACAGTATGGAGGGAATTCCTGGGGGCT-3
was used as the sense
strand, and 5
-GGCCGGTTCTGACCATTCATTTTGTGATA-3
was used as the
antisense strand. For mouse GATA-2,
5
-GACGGAGAGCATGAAGATGG-3
was used as the sense strand, and
5
-AGGCATTGCACAGGTAGTGG-3
was used as the antisense strand; these
primers were based on the sequence of human GATA-2. For mouse GATA-3,
5
-CCTTGGAACCTCAGCCCTTTTTCCAAGACCTCCA-3
was used as the sense strand,
and 5
-GGCCGGTTCTGACCATTCATTTTGTGATA-3
was used as the antisense
strand. For lacZ, 5
-CCAGCGTTTCAACGCGCTGTATGGCG-3
was
used as the sense strand, and 5
-AGACGTCACGGAAAATGCCGCTCATC-3
was used
as the antisense strand. For GAPDH, 5
-CAAAGTTGTCATGGATGACCTTGGC-3
was
used as the sense strand, and 5
-GTCTTCACCACCATGGAGAAGGCTG-3
was used
as the antisense strand.
Autoradiographic data were analyzed using a Molecular Dynamics PhosphorImager. RT-PCR agarose gel bands were quantitated using a densitometric image analyzer (densitometer) driven by a Macintosh computer. Quantitative data are presented as the means ± S.E., and differences between means were evaluated using a Student's t test, with p < 0.05 considered statistically significant.
The nucleotide sequence and
organization of the baboon chm gene is shown in Fig.
1. Based on a comparison with the baboon chymase
cDNA structure, this gene has five coding blocks. The first coding
block is 58 bp in length; it encodes, in an open reading frame, the
first 19 amino acids (21 to
3) of the preproenzyme. The second,
third, fourth, and fifth coding blocks of the baboon chm
gene encode amino acids
2 to 49, 50 to 94, 95 to 179, and 180 to 226. As assessed by primer extension analysis, the primary transcription-initiation site of the baboon chm gene is 30 bp upstream of the translation-initiation site (data not shown). The
translation-initiation ATG codon of the baboon chm gene was predicted based on the structure of the baboon chymase cDNA. The 5
-upstream region of the baboon chm gene contains a CAAT
and a TATA box,
108 to
105 bp and
29 to
25 bp upstream,
respectively, of the transcription-initiation site. A consensus
polyadenylation motif, AATAAA, which is followed by a cleavage site
motif, consisting of a CA sequence located 15 bp downstream along with
a GT cluster, is found in the 3
-untranslated region. The structure of
the baboon chm gene's 5
-upstream region (
2,426 to
1
bp) is ~92% identical to that of the human chm gene.
Baboon chymase is 97% identical to human chymase (9) and 76%
identical to its mouse homolog mmcp5 (12).
To identify the baboon chm promoter, two
constructions were made. These constructs consist of 571- and 2,426-bp
5-upstream sequences of the baboon chm gene linked to the
prokaryotic lacZ gene, i.e.
571-bchm-lacZ and
2,426-bchm-lacZ,
respectively.
571-bchm-lacZ and
2,426-bchm-lacZ constructions were used to generate
transgenic founders. Germline transmission occurred with
571-bchm-lacZ (two female and one male founder) and
2,426-bchm-lacZ (1 male founder) constructs. A
571-bchm-lacZ male founder and a
2,426-bchm-lacZ male founder were used to generate
F1 transgenic heterozygous mice. Subsequent tissue and cell
expression of the lacZ gene was determined in F1
heterozygous transgenic mice.
Tissue expression of -galactosidase mRNA in transgenic
heterozygous mice and mmcp5 expression in nontransgenic littermates was
quantified using RT-PCR. Ratios of mmcp5/GAPDH in tissues from
nontransgenic mice and lacZ/GAPDH mRNA in tissues from
571-bchm-lacZ and
2,426-bchm-lacZ
transgenic mice are summarized in Fig. 2. The relative
distribution of GAPDH-normalized lacZ expression in the
stomach, intestine, spleen, heart, lung, and brain tissues was similar
between the
571-bchm-lacZ and
2,426-bchm-lacZ constructs and also similar to that of
mmcp5. To identify the cell types in transgenic mice that express the
lacZ gene, we examined T-blue-staining cells (T-blue
produces a deep purple stain in mast cells) and
-galactosidase-containing cells in serial sections of the stomach from the
571-bchm-lacZ and
2,426-bchm-lacZ
transgenic mice. Fig. 3 indicates the selective
expression of
-galactosidase in T-blue staining cells in the
submucosa of the stomach from a
571-bchm-lacZ transgenic
mouse. Similar results were observed in the gastric submocosa and small
intestine of other
571-bchm-lacZ (n = 2) and
2,426-bchm-lacZ (n = 2) transgenic
mice (data not shown). Fig. 4 shows representative cells
in the
571-bchm-lacZ transgenic left cardiac ventricle
staining for T-blue or
-galactosidase. To determine if
571-bchm promoter-driven lacZ expression
occurs in circulating mast cell progenitors, leukocytes from
571-bchm-lacZ transgenic mice were examined for
-galactosidase activity.
-Galactosidase activity was not observed
in several leukocyte samples (data not shown).
Photomicrographs showing the expression of the
571bchm-lacZ construct in the gastric submucosa of an
F1 heterozygous transgenic mouse. Adjacent sections of
the stomach of a transgenic mouse harboring the
571-bchm-lacZ construct were stained with hematoxylin and
eosin (a), T-blue (b), and
-galactosidase
(c). In b, purple-stained cells in the gastric
submucosa indicated by arrows represent mast cells. In
c,
-galactosidase activity is found in cells located in
the same position as the purple-stained mast cells shown in
b. d and e show magnified views of
stained cells in b and c, designated
by-arrows with asterisks, respectively.
Baboon chm Gene 5
We examined several murine
mast cell lines to see which ones endogenously express the mouse
-chymase mmcp5. Using a human chymase (an
-chymase-like mmcp5
(6)) cDNA probe, Northern blot analysis indicated chymase-like
mRNA in mouse JKras mast cells but not in mouse
mastocytoma P815 cells or mouse erythroid leukemia MEL cells (Fig.
5A). Using mouse chymase subtype selective PCR primers, Fig. 5B shows that JKras mast cells
contain mmcp5 mRNA. Using mouse
-chymase subtype-selective PCR
primers showed the absence of
-chymase mRNA in JKras
mast cells (data not shown). JKras mast cells are
interleukin-3-dependent Harvey sarcoma virus-infected immune mast cells from the mouse spleen (20).
Fig. 5 shows that mmcp5 is endogenously expressed in JKras
mast cells but not P815 mast cells. To examine whether this
discrimination in the expression of an -chymase-encoding
chm gene in these two mast cell lines is mimicked by the
transcriptional activity of the
571 baboon chm
promoter-CAT construction,
571-directed expression of the CAT gene
was determined in JKras mast cells and P815 mast cells and
compared with its transcriptional activity in 3T3 fibroblasts. Table
I shows CAT activity in JKras mast cells, P185 mast
cells, and 3T3 fibroblasts after transient transfection with the baboon chm-CAT construction
571. The results of the transfection
assays were expressed as a ratio of activity of
571 to that of the
nonspecific control plasmid (SV40-CAT control). The relative CAT
activity directed by
571 in JKras mast cells was 25-fold higher than
in P815 mast cells and 51-fold higher than in mouse 3T3
fibroblasts.
|
We studied the
in vitro regulation of the baboon chm promoter in
JKras mast cells. To map the location of
cis-acting elements within the baboon chm
promoter, transcription studies were carried out in JKras
mast cell nuclear extracts with deletion mutants of the 571-bp baboon
chm promoter-CAT construct (571-bchm-CAT). The
results are summarized in Fig. 6. The transcriptional
activity of the
453-bchm-CAT construct was 11-fold
greater than of the
426-bchm-CAT construct but similar to
that of
571- and
506-bchm-CAT constructs, indicating
the presence of a positive regulatory element between
426 and
453
bp of the baboon chm gene's 5
-upstream region. Motif
analysis indicates that this region contains a GATA-binding element,
situated at
432/
427 bp upstream of the transcription-initiation site. To determine if the putative
432/
427 GATA-binding site is a
necessary for transcription, we examined transcriptional activities of
the
435-bchm-CAT construct and its mutant
[A
430
C]-
435-bchm-CAT containing a
defective GATA-binding site (Fig. 7). Transcriptional
activity of [A
430
C]-
435-bchm-CAT in
JKras mast cell nuclear extracts was ~12-fold lower than
that of
435-bchm-CAT, indicating the importance of this
single GATA cis-regulatory site in the transcriptional
activation of the baboon chm gene.
Previous studies have indicated synergistic interactions between GATA
and other transcription factors in gene regulation (23, 24). To
identify putative cis-acting elements in the
571-bchm-CAT construct that could synergize with the GATA
cis-regulatory element, transcriptional activity of
additional
571-bchm-CAT mutants was examined. Deletion of
the 139-bp sequence in the
571-bchm-CAT construct
5
-upstream of the GATA site (i.e.
432-bchm-CAT) did not reduce the transcriptional activity
of the mutant promoter construct in JKras mast cell nuclear
extract (construct 2, Fig. 8). However,
replacement of the 41-bp sequence immediately 3
to the GATA site in
the
432-bchm-CAT construct with an indifferent sequence
produced a ~50% decrease (p < 0.05) in
transcription (construct 3, Fig. 8). This indifferent
sequence, shown in Fig. 8, was 41% homologous to the wild type.
Deletion of a 333-bp sequence,
386 to
53 bp upstream of the
transcription-initiation site, in the
432-bchm-CAT
construct did not affect transcriptional activity (construct
4, Fig. 8).
GATA-2 Is Present in JKras Mast Cells and Supports Transcription of the Baboon chm Promoter
Electrophoretic mobility shift analysis
using an end-labeled DNA fragment comprising the sequence 441 to
417 bp of the baboon chm promoter containing a
GATA-binding element indicated the presence of a
trans-acting factor in JKras mast cell nuclear
extracts (Fig. 9). To characterize the specificity of
this binding, competition assays were performed using mutated forms of
the
441/
417 baboon chm gene DNA fragment. Mutation of
the GATA sequence to G
TA or
ATA produced a
loss of binding to the JKras mast cell
trans-acting factor (Fig. 9). The sequence specificity
illustrated by these experiments is similar to that previously shown
for GATA-1, GATA-2, and GATA-3 binding proteins (25).
RT-PCR analysis with specific primers for GATA-1, GATA-2, and GATA-3
was used to examine expression of these transcription factors in
JKras mast cells. Fig. 10 shows that GATA-2
but not GATA-1 or GATA-3 mRNA is present in JKras mast
cells and P815 mast cells. The presence of GATA-1 but not GATA-2 or
GATA-3 mRNA in MEL cells and GATA-2 mRNA in P815 cells has
previously been shown. These observations of GATA-binding protein
subtypes in MEL and P815 cells indicates the specificity of these PCR
primers (Fig. 10). Because GATA-3 mRNA could not be identified in
these cells as a positive control, RT-PCR analysis was also performed with total RNA from mouse spleen, which is known to contain lymphoid cells expressing GATA-3. The absence of GATA-1, GATA-2, and
GATA-3 mRNA in 3T3 fibroblasts also indicates the specificity of
the PCR primers.
To determine if ectopically expressed GATA-2 can support
571-bchm-CAT expression in a heterologous cell, we
determined transcriptional activity of the
571-bchm-CAT
construct in COS-1 cells transfected with a human GATA-2 cDNA
construct (26). Fig. 11 shows an ~5-fold increase in
transcription of the
571-bchm-CAT construct in COS-1 cells transfected with human GATA-2 cDNA compared with mock
transfected COS-1 cells.
Chymase occurs in numerous human (10) and baboon (27) tissues
including the alimentary tract, heart, lungs, blood vessels, and
reproductive tissues is implicated in diverse functions and represents
a differentiation-specific marker that identifies a population of
mature tissue resident mast cells (28). Using a transgenic approach, we
show that the 571-bp 5-upstream region of the baboon chm
gene is a promoter that allows restricted mast cell expression of a
reporter gene. These findings, as well as our identification of a
critical cis- and trans-acting positive regulatory element within this promoter, shed light on how
chm genes are transcriptionally regulated.
Introduction in mice of constructs containing 0.571- or 2.426-kb
fragments of the 5-upstream region of the baboon chm gene coupled to the prokaryotic lacZ gene, i.e.
571-bchm-lacZ or
2,426-bchm-lacZ, respectively, allows the restricted expression of
-galactosidase in
T-blue staining mast cells resident in the gastric submocosa of
transgenic mice. Cells containing
-galactosidase with mast cell-like
granules were also detected in the heart and intestine of mice
containing the
571-bchm-lacZ transgene; however, because the
-galactosidase-containing cells from the heart did not cluster like mast cells of the gastric submucosa, it was not possible to
confirm from T-blue staining of adjacent sections whether these cells
were mast cells. Circulating progenitor mast cells must normally be
resident in tissue before ultrastructural characteristics and protease
expression patterns associated with fully differentiated mast cells are
observed (1). We did not observe any cells with
-galactosidase
activity in circulating leukocyte samples from mice harboring the
571-bchm-lacZ transgene. The relative lack of expression
of the lacZ gene in brain tissue is also relevant to
tissue-specific expression of the chm gene. Blood cells
cannot readily penetrate the blood-brain barrier, and thus progenitor mast cells cannot get into brain tissue. These observations illustrate the high fidelity of the 571-bp 5
-upstream region of the baboon chm gene in conferring development-specific in
vivo expression of the lacZ gene in mast cells.
Previously, our in situ hybridization and electron
microscope immunohistochemical studies showed the presence of human
chymase and its mRNA in some vascular endothelial cells (10). In
all tissues studied from mice expressing the
571-bchm-lacZ or
2,426-bchm-lacZ
transgenes,
-galactosidase activity was not detected in endothelial
cells. This finding suggests that either mouse endothelial cells do not
contain trans-activating factors necessary for chymase
expression or cis-regulatory sequences other than those
present in the ~2.5-kb 5
-upstream fragment of the baboon
chm gene are required for chymase expression in endothelial cells.
Distribution of -galactosidase mRNA in tissues from mice
containing the
571-bchm-lacZ or
2,426-bchm-lacZ transgenes was similar to the
distribution of mmcp5 mRNA (mmcp5 is the mouse
-chymase). Given
that the same observations were made in two independently generated
groups of transgenic mice, we believe that transcriptional specificity
in transgene expression is likely to be due to baboon chm
regulatory sequences rather than to integration of the constructs near
a strong endogenous mast cell regulatory domain. These studies and the
observation that
571-bchm-driven CAT expression is high
in a mouse mast cell line that endogenously expresses the mmcp5 gene
(JKras mast cells) but low in a mouse mast cell line that
does not (P815 mast cells) implies that this baboon promoter could show
in vivo transcriptional specificity between mmcp5-expressing
connective tissue mast cells and mucosal mast cells that do not express
mmcp5. Additional in vivo studies are necessary to prove
that
571-bchm-driven reporter expression is limited to
mmcp5-expressing mast cells in vivo; however, the use of
this promoter for generating transgenic mouse models with dominant mast
cell expression of active primate chymases is also likely to be
dependent on proteoglycan composition of the mast cells. Mmcp5 is found
in all populations of mouse mast cells containing heparin sulfate
proteoglycans but not in chondroitin sulfate-containing mucosal mast
cells (29). Because heparin sulfate but not chondroitin sulfate is an
important cofactor in
-prochymase activation (15),
571-bchm-driven zymogen expression in vivo is
likely to lead to
-prochymase activation in connective tissue mast
cells but not in mucosal mast cells.
Transcriptional regulation of a gene is achieved through interactions
of cis-regulatory regions uniquely present in a gene and
cell-specific trans-activating factors with components of the basal transcriptional complex assembled at the initiation site. We
used the 571-bchm-CAT promoter construct to identify cis-regulatory regions of the baboon chm promoter
critical for transcriptional activation in JKras mast cells.
Deletional and site-specific mutational analyses revealed a strong
cis-acting positive regulatory GATA element between
453 to
426 bp upstream of the transcription-initiation site. A GATA-binding
site is also present ~430 and ~100 bp upstream of the
transcription-initiation site of the human chm (16, 30) and
mmcp5 (12) genes, respectively, and therefore, GATA-binding proteins
could be involved in transcriptional activation of these
-chymase
encoding chm genes in mast cells.
We found GATA-2 but not GATA-1 or GATA-3 DNA-binding proteins in
JKras mast cells. The further observation that CAT activity in COS-1 cells increases by ~5-fold after transfection with human GATA-2 cDNA shows the importance of this transcription factor in
trans-activation of the baboon chm promoter.
GATA-1, GATA-2, and GATA-3 transcription factors have markedly
differing patterns of cellular expression, but they recognize the same
cis-regulatory motif (A/T)GATA(A/G) (31). In
cell-dependent expression of genes requiring a GATA
cis-trans-interaction, a high level of transcriptional activation can be achieved through synergistic interactions with other
transcription factors. For example, GATA-1 and Krüppel family
factors cooperate from a distance in directing expression of the globin
gene in erythroid cells (23). A similar cooperative interaction has
been described between GATA-2 and AP1 in trans-activation of
the endothelin-1 gene (24). In these examples, synergistic interactions
between GATA and Sp1, AP1, or EKLF produce a substantial effect,
i.e. a >10-fold increase in transcription over a simple additive effect. Using motif analysis, we were unable to locate cis-regulatory motifs for AP1, Sp1, or EKLF in the 571-bp
baboon chm promoter. Furthermore, deletion or replacement of
DNA sequences 5 or 3
to the GATA cis-regulatory element
produced small decreases (<50%) in transcriptional activation of the
571-bchm-CAT construct in JKras mast cell
nuclear extracts. These data indicate an absence of strong synergistic
interactions between GATA-2 and other trans-acting factors
in the transcriptional activation of the baboon chm promoter in JKras mast cells.
In summary, we show that the 571-bp 5-upstream sequence of the baboon
chm gene can be used to selectively target mouse mast cells
in vivo and in vitro and thus provides an
important tool in the generation of transgenic mice where proteins need
to be targeted to mast cells. We also show that a GATA
cis-trans-interaction is both necessary and sufficient for
transcriptional activation of the baboon chm promoter
in vitro in JKras mast cells. However, given that
GATA-2 is found in differentiated mast cells but also has an important
role early in hematopoietic cell development (31-33), it is clear that
engaging the GATA cis-regulatory element must be one of
multiple cis-trans-interactions necessary for
transcriptional specificity. The finding that the P815 mast cell
contains abundant GATA-2 but does not support a high level expression
of CAT with the
571-bchm-CAT promoter construct suggests
that suppression of dominant negative interactions as well as enhanced
GATA-2 expression could coordinate chm gene expression as
progenitor mast cells differentiate into mature tissue resident mast
cells. The availability of a mast cell-specific promoter now makes it
possible to study how chm gene expression is suppressed in
primate mast cell progenitors as well as in other GATA transcription
factor-expressing hematopoietic cells.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U38463[GenBank] and U38521[GenBank].
We thank Dennis Wilk and Xiaopu Liu for excellent technical assistance and Christine Kassuba for manuscript editing.