(Received for publication, February 19, 1997)
From the The expression of the prolactin receptor is under
the control of two putative tissue-specific (PI, gonads; PII, liver)
and one common (PIII) promoters (Hu, Z. Z., Zhuang, L., and Dufau, M. L. (1996) J. Biol. Chem. 271, 10242-10246). The three
promoter regions were co-localized to the rat chromosomal locus 2ql6,
in the order 5 The functional diversity of prolactin, involving lactation and
reproduction, growth and metabolism, immune regulation, behavior, and
homeostasis, has been well documented in the past several decades (2,
3). However, the mechanisms underlying the regulation of prolactin
receptors present in specific target tissues have only recently been
investigated. Prolactin receptors are widely expressed, and multiple
mRNA transcripts corresponding to the long and short forms of the
receptor are present with various proportions in different tissues (4,
5). The diverse actions of prolactin could be manifested by the
expression of different receptor forms and signal transduction pathways
and also by differential control of gene transcription and mRNA
regulation of PRLR1 in target tissues.
Recently, the complexity of the mechanism by which the PRLR gene is
controlled was revealed by the demonstration of multiple and
tissue-specific promoters of the rat prolactin receptor gene (1). Three
putative promoters PI, PII, and PIII that initiate the transcription of
alternative first exons E11, E12, and
E13, respectively, were identified in the rat gonads and
the liver. The three alternate first exons are alternatively spliced to
a common noncoding exon 2 that precedes the third exon containing the
translation initiation codon of the prolactin receptor. E11 is expressed in ovaries and in Leydig cells, E12 in the
liver, and E13 in all three tissues, indicating that PI is
gonad-specific, PII is liver-specific, and PIII is a common promoter
for the prolactin receptor in these tissues and is the sole promoter
utilized in the rat mammary gland (our current study). The various
prolactin receptor forms were found not to be dependent on the
utilization of individual promoters, which may result from a
post-transcriptionally regulated process. However, these accounted for
the 5 In the present study, we have determined the co-localization and the
relative orientation of the three 5 Adult male, prepubertal female, and lactating
Harlan Sprague Dawley rats (Charles River, Wilmington, MA) were housed
in pathogen-free, temperature- and light-controlled conditions
(20 °C; alternating light-dark cycle with 14 h of light and
10 h of darkness; lights on at 0600 h). The animals were
given free access to tap water and standard Purina lab chow
(Ralston-Purina, St. Louis, MO). All animals were killed between 1000 and 1100 h by asphyxiation with CO2. All animal
studies were approved by the NICHD Animal Care and Use Committee
(protocols 94-040 and 94-041). 40-day-old adult male rats were used for
preparation of Leydig cells. 24-day-old immature female rats were
injected subcutaneously daily for 3 days with 1.5 mg/day
17 Cultures of mouse Leydig tumor cells (MLTC-1), a stable
steroidogenic cell line that expresses prolactin and luteinizing
hormone receptors (kindly provided by Dr. R. V. Rebois, National
Institutes of Health, Bethesda, MD), were maintained in RPMI 1640 (BioWhittaker, Walkersville, MD) supplemented with 10% fetal bovine
serum and 1 × antibiotic/antimycotic mixture (Life Technologies,
Inc.), and the human hepatoma cell line (HepG2, American Type Culture Collection, Rockville, MD), which expresses prolactin receptors, was
maintained in minimal essential medium (Life Technologies, Inc.)
supplemented with 10% fetal bovine serum, non-essential amino acids,
and sodium pyruvate.
Rat ovarian granulosa cells were prepared from estrogen-primed rat
ovaries as described previously (10) with some modifications. Briefly,
the ovaries were removed and cleared of the surrounding fat tissues.
These were washed in Medium 199 (M199, BioWhittaker) containing 0.2%
bovine serum albumin and 10 mM HEPES, pH 7.2 (Medium A),
and punctured 20 times with a 23-gauge needle in Medium B (Medium A
with additional 6.8 mM EGTA) followed by incubation in a
CO2 incubator at 37 °C for 10 min and centrifugation at
700 rpm for 10 min. The ovaries were then incubated with Medium C (Medium A with additional 1.8 mM EGTA and 0.5 M
sucrose) for 6 min in CO2 incubator at 37 °C and
centrifuged as above. To disperse the granulosa cells, the ovaries were
pressed against a 40-mesh metal sieve with a spatula, and the dispersed
cells were collected. The cells were treated with trypsin (Sigma) (20 mg/ml) for 1 min at 37 °C followed by addition of soybean trypsin
inhibitor (Sigma) (300 µg/ml) and DNase I (Sigma) (100 µg/ml) for 4 min at 37 °C. The enzymes were removed by centrifugation, and the
cells were washed twice with Medium A. Cells were suspended in medium
Dulbecco's modified Eagle's medium/F12 (Life Technologies, Inc.)
supplemented with 1% fetal bovine serum, and with 15 ng/ml ovine
follicle stimulating hormone (ovine FSH-20, National Pituitary Program,
NIDDK), 10 ng/ml testosterone (Sigma) and plated at a density of
0.5 × 106 cells/cm2 in 24-well plates and
cultured at a CO2 incubator for 3-5 days.
Rat Leydig cells were prepared from adult male rat testes by
collagenase digestion and purified by centrifugal elutriation (11). The
purified Leydig cells were suspended in M199 containing 0.1% bovine
serum albumin and 1 × antibiotic/antimycotic and plated at the
density of 0.5 × 106 cells/cm2 and
cultured for up to 20 h.
Lambda phage DNA with genomic PRLR
DNA inserts of 15 and 14.5 kb were used for fluorescence in
situ hybridization analysis. The probes were labeled with biotin
(lambda phage clone For mapping of the plasmid DNA isolated from a P1 genomic library
(Genome Systems, Inc., St. Louis, MO) containing the 5 DNA fragments of the 5 For RT-PCR analyses
of RNA from MLTC, first strand cDNA was synthesized with random
primers using Superscript reverse transcriptase (Life Technologies,
Inc.) at 42 °C for 30 min. Primers used for PCR amplification of
exons E11 and E13 were
5 Nuclear proteins from MLTC,
rat granulosa cells, and rat Leydig cells were extracted as described
(14) with some modifications. Briefly, cells were suspended in buffer A
containing 10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl, 0.1 mM EDTA, 0.5 mM DTT,
0.4 mM Pefabloc SC (Boehringer Mannheim), 2 µg/ml
leupeptin (Sigma), and 2 µg/ml pepstatin A (Sigma) with additional
0.3 M sucrose and 2% Nonidet P-40 and homogenized 20 strokes with type B pestle. The nuclei were recovered by centrifugation
of the cell homogenate through a 1.5 M sucrose cushion
contained in buffer A and lysed in buffer B (10 mM HEPES,
pH 7.9, 420 mM KCl, 1.5 mM MgCl, 0.1 mM EDTA, 10% glycerol, 0.5 mM DTT, 0.4 mM pefabloc SC, 2 µg/ml leupeptin, and 2 µg/ml
pepstatin A). The nuclear lysate was dialyzed against buffer D (20 mM HEPES, pH 7.9, 100 mM KCl, 0.1 mM EDTA, 20% glycerol, 0.5 mM DTT, 0.4 mM pefabloc SC, 2 µg/ml leupeptin and pepstatin A)
overnight at 4 °C. The concentration of the nuclear proteins was
measured by the protein assay kit (Bio-Rad).
Double-stranded DNA fragment
corresponding to the region of PI at Synthesized
oligonucleotides were used for gel mobility shift assays unless
otherwise indicated. The oligomers were annealed and 5 MLTC cells were
transfected at 50-70% confluency ~48 h after plating using
N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsulfate (Boehringer Mannheim) as described previously (16). Cells
were harvested 40-60 h after transfection, and whole cell lysates were
assayed for luciferase activity (17). Since the primary rat gonadal
cells were not amenable for transient expression of PRLR-PI/LUC
reporter gene constructs by conventional liposome-mediated transfection
in our experiments, the recombinant adenovirus infection method was
applied in this study. Adenovirus reporter gene constructs (PI/LUC/Adv)
were added at multiplicity of infection of 100 to the granulosa or
Leydig cell culture and incubated for 20 h. Adenovirus construct
containing the promoterless luciferase gene served as background
control and a construct containing the SV40 promoter-luciferase gene as
positive control. The granulosa cells were infected after 3 days of
culture, and the infection was allowed to continue for 20 h before
termination. The Leydig cells were infected 2 h after plating, and
human chorionic gonadotropin was added to a final concentration of 20 ng/ml 6 h before termination. The infection was also allowed to
continue for 20 h. Whole cell lysates were used for measurement of
luciferase activity.
The fluorescence in situ hybridization
method was used to establish the chromosomal location of the rat PRLR
gene and to determine whether the three PRLR gene promoter regions that
direct transcription of alternative first exons reside at the same gene
locus. The genomic probe (15 kb) containing the 5
Due to the lack of appropriate rat gonadal stable cell
lines, we have utilized a mouse Leydig tumor cell line (MLTC-1) (18) for investigation of the transcriptional control of the PRLR gene promoters. For these studies, it was necessary to initially
characterize the endogenous expression of PRLR alternative exons in
these cells. Northern blot analysis showed a major mRNA transcript
of 9.5 kb comparable to the 9.7 kb of the rat species using 5
The promoter utilization of the PRLR gene in rat mammary gland was also
examined. Northern blot analyses showed that the major mRNA
transcript in this tissue is the 1.8-kb species (Fig. 2C, left), which is also seen in the Leydig cells
(right) and the liver (1), and corresponds to the short form
of the receptor (1, 4). 5 A
luciferase reporter gene construct containing the 5
The PRLR PI/LUC constructs
with serial deletion of the 5
DNase I footprinting analysis using nuclear extracts from MLTC was
employed to examine the specific nuclear protein binding activities
within the promoter domain. DNA probes corresponding to the region of
Electrophoretic mobility shift assay of MLTC nuclear extract using
oligonucleotide probes (probes A and B, Fig. 6)
containing the footprint sequence showed a major protein complex (Fig.
6, lanes 1, 10, and 15) that was specifically
competed by unlabeled wild-type sequence even at a molar excess of 10 (lanes 2-4, 16, and 17) and was either
supershifted (lanes 12 and 22) or inhibited (lanes 13, 14, and 23) by specific SF-1
polyclonal antibodies but not by normal rabbit serum (NRS,
lanes 11 and 21). Mutation of the SF-1 element
(m1, m2, and m3, Fig. 6, lanes 5-7,
18-20, and below) did not compete for the SF-1
protein binding, indicating that CCAAGGTCA ( The trans-factors for promoter I present in the nuclei of the rat
gonadal cells were also evaluated using EMSA. A single protein complex
from granulosa cell nuclear extracts was shown to bind CCAAGGTCA (Fig. 7, lanes 1-10) as
was found in MLTC cells. The retarded band was supershifted (lane
12) or inhibited (lane 14) by SF-1 antibodies but not
by normal rabbit serum (lane 11). Similar results were
obtained with rat Leydig cell nuclear protein extracts (lanes
15-23). However, the SF-1 binding complex was resolved as a
doublet, and these bands may represent different states of the SF-1
protein.
Mutation of the SF-1 binding site not only
abolished the SF-1 protein binding activity (Figs. 6 and 7) but also
markedly decreased the promoter activity (by 8-fold) in MLTC (Fig.
8, construct 2). Similarly, deletion of SF-1
element between
The promoter domain
Previous studies have identified three putative PRLR gene
promoters (PI, PII, and PIII) that are alternatively utilized in a
tissue-specific manner and suggested that the expression of PRLR in
different tissues may be subject to differential regulation by
different regulators at various levels (1). In this study, we have
demonstrated that the three alternative promoters are co-localized in
the rat chromosome 2q16 in the order of 5 Our current study has predominantly focused on the elucidation of the
mechanism controlling the gonad-specific expression of E11
by promoter I. The mouse Leydig tumor cell line (MLTC) has been used
for characterization of cis-elements and transcription factor(s)
required for promoter I activity. Analyses of endogenous PRLR mRNA
transcripts in MLTC indicated that the mouse gonadal cell line
expresses the counterparts of the rat first exons E11 and
E13 and therefore possess promoters I and III activities. The expression of the PI( From deletion analyses, we have identified a 152-bp region ( To our knowledge, this is the first report of a receptor gene that
requires SF-1 as the major transcriptional activator and for the
tissue-specific utilization of its alternative promoters. Although the
gonadotropin releasing hormone receptor gene was recently shown to
require a functional SF-1 element for its specific expression in
Additional factors may be required for full activation of this
promoter, since mutation or deletion of SF-1 motif ( Our studies have demonstrated that the SF-1 protein binds to its
cognate regulatory element within the PI promoter domain and exerts a
dominant effect on the activation of this promoter. In addition, a
CCAAT element may have minor contribution to the basal promoter
activity, and a TATA-like region may provide a constitutive inhibitory
site for this promoter. These important functional gene structures are
amenable to multifunctional regulation during the development of
testicular Leydig cells from fetal to adult life and at the
different stages of the ovarian cycle.
We thank Drs. Joseph L. Alcorn and Carole A. Mendelson (University of Texas Southwestern Medical Center, Dallas, TX)
for providing plasmids pAC and pJM17 (13) and Dr. Ken-ichirou Morohashi
(Kyushu University, Kyushu, Japan) for providing Ad4BP antibody (15).
Section on Molecular Endocrinology,
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
-PIII-PI-PII-3
. To investigate the mechanisms of gonad-specific utilization of PI, the promoter domain, regulatory cis-elements, and trans-factors were identified in gonadal cells. The
promoter domain localized to the 152-base pair 5
of the
transcriptional start site at
549 is highly active in gonadal cells
but has minimal activity in hepatoma cells. It contains a steroidogenic
factor 1 (SF-1) element (
668) that binds the SF-1 protein of nuclear extracts from gonadal cells and is essential for promoter activation. A
CCAAT box (
623) contributes minimally to basal activity in the
absence of the SF-1 element, and two adjacent TATA-like sequences act
as inhibitory elements. Thus, PI belongs to a class of
TATA-less/non-initiator gene promoters. These findings demonstrate an
essential role for SF-1 in transcriptional activation of promoter I of
the prolactin receptor gene, which may explain the tissue-specific
expression of PI in the gonads but not in the liver and the mammary
gland.
-untranslated region heterogeneity of the first exons of the PRLR
mRNA transcripts (E11, E12, and
E13) (1). The heterogeneity of the 5
-untranslated region
of the PRLR transcripts was also reported recently in the gonads and
the liver by others (6, 7). Sequence analyses of 5
-flanking regions
showed a lack of consensus TATA box sequences positioned within the
expected distance from the transcriptional start site (TSS) in all
three putative promoter regions (1). However, potentially functional
non-canonical TATA box sequences are present in these regions, and in
PI two adjacent TATA-like sequences reside 10 and 23 bp 5
from the
TSS. In addition, consensus sequences for several transcription factors
are also present in these promoters, specifically SF-1, C/EBP, and CAAT
box in PI; GATA-1, ERE and AP-1 in PII; and GAGA, GATA-1 and AABS in
PIII. Therefore, it is assumed that different transcription factors and
other cellular and extracellular regulators may be involved in the
control and regulation of the PRLR gene in diverse target tissues.
-untranslated region/promoter regions in the rat chromosome, and we have investigated the underlying mechanisms of the tissue-specific promoter control of the PRLR gene in
the gonads by the characterization of the gonad-specific promoter
domain PI and its relevant cis-elements and trans-factors. These
studies have demonstrated that steroidogenic factor-1 (SF-1) (8), also
known as Ad4BP (9), is an essential transcriptional activator of PI in
testicular and ovarian cells. Such regulation by SF-1 appears to
determine the specific gonadal utilization of promoter I.
Animals
-estradiol (dissolved in propylene glycol) (Sigma). On the 4th
day, the animals were sacrificed for preparation of ovarian granulosa
cells. The mammary gland of lactating female rats were resected and
freed of surrounding adipose tissue and used for mRNA
preparation.
-Flanking/Promoter Regions
11-1, 15 kb, containing PRLR PI/PII regions) or
spectrum orange (lambda phage clone
3a, 14.5 kb, containing PIII
region) by nick translation (Life Technologies, Inc.) and hybridized to
rat metaphase chromosomes as described previously (12). Briefly, about
200 ng of each probe was used in 10 µl of hybridization mixture
containing 55% formamide, 2 × SSC, and 1 µg of human
CotI DNA (Life Technologies, Inc.) that was denatured at
75 °C for 5 min. Slides with rat metaphase chromosomal spreads were
denatured in 70% formamide, 2 × SSC at 72 °C for 2 min, and
hybridized with the specified probes at 37 °C in a moist chamber
overnight. Slides were then washed three times in 50% formamide,
2 × SSC at 45 °C for 3 min each. The hybridization signal of
the probe was detected by two layers of fluorescein isothiocyanate-conjugated avidin (Vector Laboratories, Inc, Burlingame, CA) and amplified with one layer of anti-avidin antibody (Vector). Slides were counterstained with 0.5 µg/ml
4,6-diamidino-2-phenylindole in an antifade solution (1 mg/ml
p-phenylenediamine dihydrochloride, 10% phosphate-buffered
saline (v/v), 90% glycerol (v/v), 4.2% sodium carbonate (w/v)) and
was examined with a Zeiss Axiophot microscope equipped with a dual
bandpass filter.
-flanking regions of the rat PRLR gene, the plasmid was digested with
BamHI or in combination with NotI followed by
Southern blot hybridization with oligonucleotide probes derived from
E11, E12, and E13.
-flanking region containing promoter
I were either generated by restriction enzyme digestion
(KpnI/XbaI,
1566/
124) or by PCR
amplification. For generation of plasmid constructs, the 5
-flanking
DNA fragments of PI were ligated 5
to the luciferase gene (LUC) of the
linearized plasmid pGL2 (Promega, Madison, WI). The pGL2 constructs
were numbered relative to the translation initiation codon
(PI(
#/
#)/LUC). The recombinant adenovirus (Adv) luciferase reporter
constructs containing the putative promoter region (5
-flanking region
of the PRLR) 5
to the luciferase gene [PI(
#/
#)/LUC/Adv] were
prepared as described previously (13). Briefly, the 5
-flanking region
of PRLR-luciferase (5
PRLR/LUC) minigenes were excised from pGL2
plasmid constructs using SmaI/BamHI and inserted
to the EcoRV/BamHI site of pAC plasmid, which
contains partial sequences of the adenovirus 5 genome. The resulting
pAC/PI/LUC constructs were co-transfected with plasmid pJM17, which
contains the remainder of the adenovirus genome into human embryonic
293 cells. In vivo homologous recombination of the plasmids
yields recombinant viral genome (PI/LUC/Adv) and subsequent generation
of infectious viral particles. Viral plaques were isolated and
propagated to a titer of 109 ml
1 in the 293 cells and were used for infection of primary cultures of rat gonadal
cells. Structure of the fusion genes was verified by PCR amplification
of the insert and the subsequent DNA sequence analysis.
-RACE PCR, and Northern Blot
-GTGGCCAGAGCCATGGACAG-3
(forward) and
5
-AAACTCTTTCCTCGGAGGTCACTAG-3
(reverse),
5
-TCTCAGAGACACGCGGCTG-3
(forward) and
5
-TTCTGCTGGAGAGAAAAGTCTG-3
(reverse), respectively. PCR fragments
were resolved on 1.5% agarose gel. 5
-RACE PCR analyses of RNA from
luciferase reporter plasmid PI(
1566/
124)/LUC transfected in MLTC
were performed as described previously (1). First strand cDNA was
synthesized using primer GS1 (5
-AGCGGTTCCATCCTCTAG-3
) (+7 to +24 of
the luciferase coding region) and 3
-end-tailed with dCTP using
terminal deoxynucleotidyltransferase followed by PCR with primer GS2
(5
-CTTTATGTTTTTGGCGTCTTCCA-3
) (+44 to +66 of the LUC coding region)
and dG-adaptor primer
(5
-GCGAATTCTCGAGATCTGGGIIGGGIIGGGIIG-3
), where I represents inosine. The PCR products were resolved
on 1.5% agarose gel and subjected to Southern blot analyses using nested oligonucleotide probe GS3 (5
-TCTACCTAACCCGCCCACTGGTT-3
) within
E11. Primers used for 5
-RACE PCR analyses of rat mammary gland PRLR mRNA were same as described previously (1).
Poly(A)+ RNAs from the rat ovary, Leydig cells, mammary
gland, and MLTC were prepared and analyzed by Northern blot as
described previously (1, 4).
827 to
440 was used for DNase
I footprinting analysis. One strand end-labeled DNA probes were
generated by first labeling the 5
end with [
-32P]ATP
(3000 mCi/mmol, DuPont NEN) of either forward primer
(5
-TGTCTGCCTCATGAGAATAC-3
) or reverse primer
(5
-TCTACCTAACCCGCCCACTGGTT-3
) followed by PCR amplification of the
DNA fragment with one labeled and one unlabeled primer. For each
footprinting reaction, 5 × 104 cpm of the probe was
added to 10-20 µg of nuclear protein in 50 µl of mixture
containing 20 mM HEPES, pH 7.5, 50 mM KCl, 0.5 mM EDTA, 1 mM DTT, 5% glycerol and incubated
at room temperature for 15 min. DNase I digestion was performed by
adding 1 × DNase I buffer (25 mM NaCl, 10 mM HEPES, 5 mM MgCl, and 1 mM
CaCl2) containing 1 unit of DNase I (Promega) and
incubation for 1 min at 22 °C. The reaction was terminated by adding
10 µl of stop buffer (200 mM NaCl, 30 mM
EDTA, 1% SDS, and 100 µg/ml yeast RNA) and 110 µl of
phenol/chloroform. The digested DNA fragments were recovered by ethanol
precipitation and resolved on 6% polyacrylamide-urea gel
electrophoresis.
-end-labeled
with [
-32P]ATP (3000 mCi/mmol, DuPont NEN). 1-3 µg
of nuclear protein was added to 20 µl of reaction containing 12 mM HEPES, pH 7.6, 60 mM KCl, 4 mM
Tris-HCl, 5% glycerol, 1 mM EDTA, 1 mM DTT,
and 25 µg/ml polydeoxyinosinic deoxycytidylic acid on ice for 15-30
min followed by addition of 5 × 104 cpm of the probe
for additional 15 min. For competition assay, unlabeled DNA sequences
were added to the reaction 15 min prior to the addition of the probe.
For supershift assay, SF-1 antibody 1 (a gift from Dr. Ken-ichirou
Morohashi, Kyushu University, Japan) (15) or antibody 2 (8) (purchased
from Upstate Biotechnology, Inc, Lake Placid, NY) was incubated with
nuclear proteins for 30 min prior to the addition of the probe.
DNA-protein complexes were resolved on 5% native polyacrylamide gel
electrophoresis.
Chromosomal Localization of PRLR Promoters and Their Order in the
Rat Genome
-flanking region
corresponding to both PI and PII promoter regions has located these
regions at the chromosomal locus 2q16 (Fig. 1,
above). The same finding was obtained using the genomic
probe containing coding exons 4 and 5 of the rat PRLR (not shown). The
PIII promoter region was also located at the identical locus on the
same metaphase chromosome when probed with a genomic fragment (14.5 kb)
containing the PIII region (Fig. 1, above). The order of the
three promoters in the gene was determined by mapping overlapping
genomic clones (Fig. 1, middle). The size of the DNA insert
cloned from the rat P1 genomic library is ~96 kb as determined by
restriction digestion and agarose gel electrophoresis (Fig. 1,
middle left). A 55-kb BamHI fragment was
hybridized by two oligomer probes derived from E11 (Fig. 1,
middle right, lanes 1-3) but not from
E12 (not shown) or E13 (lane 4).
Lanes 2 and 3 show a 35-kb positive fragment generated from the 55-kb fragment with NotI digestion, which
cleaves at a unique site of the 16-kb vector. Furthermore, a
BamHI-digested fragment of 8 kb was specifically hybridized
by the oligomer probe derived from E13 (lane 4).
However, no hybridization of this genomic DNA was revealed by the
oligomer probes derived from E12 (not shown). These results
indicated that this 96-kb genomic fragment of the rat PRLR gene
contained both regions corresponding to exons E11 and
E13 as well as their putative promoter regions PI and PIII
but not PII and exon E12. Since previously we have
established that PI region is located 10 kb 5
of PII region (1), we
could derive that the order of these 5
-flanking/promoter regions of the PRLR gene is 5
-PIII-PI-PII-3
, which spans over 20 kb in the
genome.
Fig. 1.
Chromosomal localization and genomic order of
rat PRLR 5-flanking/promoter regions. Above, genomic probes
containing 5
-flanking/promoter regions, PI and PII (15 kb), and PIII
region (14.5 kb) were labeled with biotin and spectrum orange,
respectively, by nick-translation and hybridized to the metaphase
chromosomes of the rat fibroblast. The signals in green
(PI/PII) and red (PIII) were co-localized to the long arm of
chromosome 2 at the 2q16 locus as indicated by the G-banding
method (left). Middle: left, schematic
representation of the plasmid DNA insert restriction map by
BamHI. This clone contains a PRLR gene insert of 96 kb with
at least nine BamHI sites (indicated by
cross-lines). The NotI site is unique to the
vector (16 kb, closed bar). The positions of E11
and E13 in the plasmid were determined by Southern
hybridization of the digested plasmid DNA by oligonucleotide probes
derived from E11, E12, and E13
separately. Right, Southern blot of digestion of plasmid DNA
with BamHI (lane 1) and with BamHI
plus NotI (lanes 2-4). Lanes 1 and
2 were hybridized with sequence 5
-TCCTCTCTCACCAGGCGAAAC-3
(E11), lane 3 with
5
-ACAGTACTGGGAGAATGGCTCTAAG-3
(E11), and lane
4 with 5
-GCCCCGTGTAAAATC CAGAACGCAG-3
(E13). No
hybridization signal was obtained with probe (5
-TCCC
TAGAATCCAGCTCGCTTGAC-3
) from E12 (not shown). *,
NotI shows partial digestion, which yields a 35-kb fragment
from the 55-kb BamHI fragment. Below, the
schematic drawing shows the order of the three first exons and
corresponding 5
-flanking regions of the rat PRLR gene.
[View Larger Version of this Image (39K GIF file)]
-coding
region of the rat PRLR as the probe (Fig. 2A,
left). In addition, minor transcripts of 4.2, 2.4, and 1.4 kb were
also revealed in MLTC. Based on the high similarity between the rat and
the mouse PRLR cDNAs, primers derived from the rat E11
and E13 were used for RT-PCR analyses of mRNA from rat
ovaries and MLTC cells. PCR products of identical sizes for
E11 (325 bp) (Fig. 2B, lanes 2-4) and for E13 (229 bp) (lanes 5-7) were observed for the
rat ovary and MLTC (mouse), indicating analogous expression of
alternative first exons E11 and E13 in both
species. Furthermore, recent genomic cloning of the mouse
E13 and its 5
-flanking region showed 94% sequence
similarity between the rat and the mouse within the noncoding exon
E13 and the proximal 5
-flanking region (not shown). These results indicate that as in rat gonads, the PRLR expression is also
under the control of putative promoters PI and PIII in MLTC cells.
Fig. 2.
Endogenous expression of mouse counterparts
of the rat PRLR exons E11 and E13 in MLTC.
A, Northern blot analyses of PRLR mRNAs from MLTC cells
(left) and rat ovary (right). The
32P-labeled probe of 0.4 kb used contained the common
coding sequences for the long and short form of the rat PRLR (4).
Left lane shows major transcript at 9.5 kb and minor species
at 4.4, 2.4, and 1.4 kb. This lane is to be compared with the
right lane showing the 9.7, 2.1, and 1.8-kb species
previously demonstrated in rat ovaries (4). B, RT-PCR
analyses of RNAs from MLTC (lanes 3, 4, 6, and 7)
and rat ovaries (lanes 2 and 5). DNA standards
are shown in lanes 1 and 8. Primers were derived
from E11 and E13 as indicated below
the gel. The primers used were 5-GTGGCCAGAGCCATGGACAG-3
(upper
strand) and 5
-AAACTCTTTCCTCGGAGGTCACTAG-3
(lower
strand) for E11, 5
-TCTCAGAGACACGCGGCTG-3
(upper strand) and 5
-TTCTGCTGGAGAGAAAAGTCTG-3
(lower strand) for E13. The expected band size
is indicated by the arrow in base pairs (bp).
C, Northern blot analyses of PRLR mRNA from rat mammary
gland (left) and Leydig cells (right). The probe
used was the same as in A. The 1.8 kb is the major species in mammary tissue, whereas the 9.7 and 1.8 kb are the major species in
rat Leydig cells (1). D, 5
-RACE PCR analyses of PRLR
mRNA from the rat mammary gland. Primers at different locations
were selected to generate different sizes of PCR products. PCR products of identical sizes as from those of the rat gonads and liver were obtained and hybridized by oligomer probes derived from E13
(lanes 1 and 2) but not from E11
(lanes 3 and 4) or E12 (lanes
5 and 6).
[View Larger Version of this Image (37K GIF file)]
-RACE PCR analyses of mRNA from the rat
mammary gland showed that only E13 sequences were present
in this tissue by sequencing the PCR products (not shown) and by
Southern blot analysis of the PCR products (Fig. 2D).
Neither E11 nor E12 sequences were found to be
present in the mammary tissue by Southern hybridization by
oligonucleotide probes derived from E11 (lane 3 and 4) or E12 (lane 5 and
6). This result indicated that PIII is the sole promoter utilized in the rat mammary gland and further demonstrated that the
usage of PI is restricted to the gonads among the tissues examined.
-flanking region
with adjacent E11 sequence [PI(
1566/
124)/LUC] was examined for its promoter activity in the steroidogenic MLTC and the
non-steroidogenic cell line HepG2. This genomic fragment exhibited strong promoter activity, which was 43-fold over that of the
promoterless vector and about 60% that of the SV40 promoter in MLTC.
In contrast, only minimal induction was observed in HepG2 cells, where
the SV40 promoter displayed activities comparable to those observed in
MLTC cells (Fig. 3, left). This finding
indicated the specificity of activation of promoter I in MLTC and is
consistent with the pattern of promoter utilization in vivo,
where PI activity was found only in gonads but not in the liver (1) and
the mammary gland (Fig. 2D). The transcription initiation
site from the PI(
1566/
124)/LUC construct transfected in MLTC was
determined by the 5
-RACE PCR method. An extended product of ~540 bp
was revealed on Southern hybridization of the PCR products by a nested
probe (GS3, Fig. 3, right, above), and the
transcriptional initiation site derived from this experiment was
consistent with that reported for the rat ovary (
549 relative to
translation initiation codon at +1) (1) (Fig. 3, right,
below). These results indicate that MLTC is an appropriate cell
line for characterization of the regulatory components of this
gonad-specific promoter (PI), since the PI/LUC construct is highly
active and faithfully transcribed in MLTC.
Fig. 3.
Promoter I activity in MLTC and HepG2 cells
and determination of the transcription start site for the expressed
PRLR-PI/LUC gene construct in MLTC cells. Left, the
luciferase reporter construct PI(1566/
124)/LUC containing the
5
-flanking region and partial exon 1 (E11) sequences was
transiently expressed in MLTC and HepG2 cells. The luciferase
activities are expressed in fold induction over that of the basic
vector (lacking PRLR gene sequence). Right, 5
-RACE PCR
analysis of reporter gene mRNA transcripts from
PI(
1566/
124)/LUC transfected in MLTC. The two gene-specific primers
(GS1 and GS2) used were derived from regions of
the luciferase gene to avoid interference from the endogenous mouse
PRLR mRNA. Specific extended products of ~540 bp were generated from both total RNA (lane 1) and poly(A+) RNA
(lane 2). The scheme for the analysis is indicated
below the Southern blot. The sequences for GS1 and GS2 were
5
-AGCGGTTCCATCCTCTAG-3
and 5
-CTTTATGTTTTTGGCGTCTTCCA-3
located at
+24 and +66 of the luciferase coding region, respectively. The
wavy line is a portion of the polylinker sequence from the
vector (55 bp). CC in the lower right indicates
dC-tailing of the first strand cDNA synthesized with GS1. The
sequence of dG adaptor primer is
5
-GCGAATTCTCGAGATCTGGGIIGGGIIGGGIIG-3
. The calculated size from the
reporter construct is 538 bp.
[View Larger Version of this Image (42K GIF file)]
-flanking region were expressed in MLTC
to localize the minimal promoter domain. Deletion of the 5
-flanking
sequences beyond
700 had no significant effect on luciferase activity
(Fig. 4). The region between
700 and
549 was
required for promoter activity, whereas deletion of the downstream exon
1 sequences (
124 to
549) had no significant effect on promoter
activity. Thus, the 152-bp region between
700 and
549 retained full
promoter activity in transfected MLTC cells. In contrast, the
PI(
1566/
124)/LUC construct (Fig. 3) and the deletion mutants of PI
including the minimal promoter domain
700/
549 (not shown) exerted
only minor promoter activities in HepG2 cells, further indicating the
tissue specificity of promoter I for expression in gonadal cells. The
regulatory cis-elements and transcriptional activators required for
this promoter were further analyzed by the DNA-protein binding analyses
and site-directed mutagenesis of regulatory elements within the
promoter domain.
Fig. 4.
Basal activities of deletion constructs of
the 5-flanking region PI-exon E11 in MLTC.
Left, schematic representation of deletion constructs
expressed in MLTC cells. Luciferase reporter gene constructs with
progressive 5
deletions of the 5
-flanking region as well as the 3
deletions of the adjacent first exon E11 are shown. Plasmid
constructs are identified by nucleotide position calculated from the
first base of translation initiation codon (ATG, +1). The transcription
initiation site is indicated by arrowhead
(
549). Right, the activities of the constructs
were expressed as fold induction of luciferase activity over that of the basic vector. Results are expressed as mean ± S.E. of four to
six independent experiments in triplicate.
[View Larger Version of this Image (32K GIF file)]
440 to
827 labeled at 5
ends of either upper or lower strands were
used in the DNase I footprinting assay (Fig. 5). Regions
at
679 to
661 and
681 to
663 were protected for upper (Fig. 5,
lanes 3-5) and lower strands (lanes 13-15),
respectively, indicating a single footprint with asymmetrical binding
of the protein complex to the double helixes of this region (Fig. 5, middle). Within this footprint resides the consensus binding
element for the steroidogenic factor 1 CCAAGGTCA.
Preincubation with unlabeled DNA sequences containing the SF-1 element
(see also Fig. 6) can effectively prevent formation of
the footprint (Fig. 5, lanes 6, 7, 16, and 17),
whereas the mutated SF-1 sequence (m1, see also Fig. 6) had
no effect (lanes 8 and 18). Although the region
628/
648 downstream of the putative SF-1 element contains a
consensus C/EBP binding site (TTGTGTAA), sequences containing either
wild type (lanes 9 and 19) or mutation of the
C/EBP binding site (lanes 10 and 20) did not
affect the footprinting pattern, indicating that the nuclear protein
complex binds specifically to a single region containing the SF-1
element.
Fig. 5.
DNase I footprinting analysis of promoter I
protein binding domain using MLTC nuclear extracts. Both upper
strand end-labeled probe (left) and lower strand end-labeled
probe (right) were used in the assay. Probes were incubated
with increasing concentrations of nuclear proteins (lanes
3-5 and 13-15), or bovine serum albumin (BSA) (20 µg, lanes 2 and 12) as
control. The probes were also incubated with nuclear proteins in the
presence of unlabeled DNA competitors as indicated (lanes
6-10 and 16-20). The molecular ladders are from
sequencing reactions using primers derived from the labeled ends of
either probes (see "Materials and Methods"). The protected region
on either strand was indicated by shaded boxes. The
sequences protected on both strands and their positions in relation to
the transcription initiation site (549) are indicated (between
gels).
[View Larger Version of this Image (126K GIF file)]
Fig. 6.
Electrophoresis mobility shift assay (EMSA)
of binding of DNA footprinting sequences by nuclear proteins from MLTC
cells. Double-stranded DNA probes A (689/
650) and B
(
689/
628) containing sequences corresponding to the footprint
(sequences in bold, below) were used for the EMSA. Probe A
contains the consensus SF-1 element (underlined sequences
shown below) and probe B (DNA fragment generated by PCR)
contains additional consensus sequence for C/EBP element (TTGTGTAA).
Wild-type and mutated sequences of the probe were shown
below. Probes were incubated with the nuclear protein in the
absence (lanes 1, 10, and 15) or presence of
excess molar of unlabeled DNA competitors, wild-type DNA (lanes
2-4, 16, 17, and 24), or mutated DNA (lanes
5-9 and 18-20), or in the presence of normal rabbit
serum (NRS, lanes 11 and 21) or SF-1 antibodies (Ab1, lanes 12 and 22; Ab2,
lanes 13, 14, and 23). The specific DNA-protein
complex (SF-1) and free probes are indicated by
arrows. SF-1 is defined as the protein from ovarian
granulosa and Leydig cells that is responsible for DNA-specific
retardation of the designated band.
[View Larger Version of this Image (99K GIF file)]
676 to
668)
is required for the SF-1 binding, conforming to the sequence
requirement for SF-1 in other genes (19). It was noted that a sequence
(AACAGGCCA,
682 to
674) resembling the SF-1 consensus
sequence is overlapping (underlined) at 5
with the SF-1 binding site
(in bold) (CGAAACAGGCCAAGGTCAAAC). To more
accurately define the SF-1 binding site, the adjacent nucleotides were
mutated and the DNA binding activity was evaluated. When the upstream
AGG (
677) was mutated to cta (Fig. 6,
m4), or cat, or tta (not shown), the
SF1 binding was completely competed (Fig. 6, lane 8) by
unlabeled mutant sequences. Furthermore, mutation of nucleotides
AAC at
665 (m5) and at further upstream sites (
680, not
shown) to ctg also competed for the binding, indicating that
these flanking nucleotides are not required for SF-1 binding activity.
These results demonstrate that the SF-1 binding site resides at
CCAAGGTCA (
676/
668) but not at AACAGGCCA
(
682/
674). Moreover, the longer probe B (62 bp) containing the
consensus sequence for C/EBP (TTGTGTAA) in addition to the SF-1 element (Fig. 6, lower right) was shown to bind to the same protein
complex(es) as the shorter probe A (30 bp) lacking the C/EBP sequence
(lanes 15-23). This DNA-protein complex was not competed by
the sequence
648/
628 (lane 24), further indicating that
the consensus C/EBP site did not have significant binding activity at
the given condition (Fig. 6). These findings are consistent with those
derived from footprinting analysis that revealed a single protected
region corresponding to the SF-1 element.
Fig. 7.
EMSA of binding of DNA containing SF-1
binding sequence by nuclear proteins from rat ovarian granulosa and
testicular Leydig cells. Probe A (689/
650) was incubated with
the ovarian (lane 1-14) or testicular nuclear extracts
(lanes 15-23) in the absence (lanes 1, 10, and
15) or the presence of designated molar excess of unlabeled
competitors (lanes 2-4 and 16, wild type; lanes 5-9 and 17-19, SF-1 mutated). The
specificity of SF-1 binding is confirmed by the use of two SF-1
antibodies Ab1 and Ab2 (lanes 12-14 and 21-23)
to be compared with the negative control of normal rabbit serum
(NRS, lanes 11 and 20). The two
antibodies used are presumed to act as follows. Ab1 supershifts the
SF-1·DNA complex after binding to the exposed surface of SF-1. In
contrast, Ab2 inhibits the binding of SF-1, probably by interference or
blockade of the SF-1 protein site that is required for interaction with the SF-1 DNA element. Specific DNA-protein complexes are indicated by
arrowhead. Free DNA probes are also indicated
(Free).
[View Larger Version of this Image (86K GIF file)]
660 and
700 decreased the promoter activity by
10-fold (construct 5). This indicates a crucial
role for SF-1 in the transcriptional activation of promoter I in MLTC.
A further reduction to basic level in promoter activity was observed
when the deletion was extended to the position
627 (construct 6), which disrupted the consensus
CCAAT box sequences (CCAATTA,
629/
623), suggesting that a sequence
element downstream of
660 may also contribute to the basal promoter
activity. Mutation of the CCAAT box caused a small but significant
reduction (30%) in promoter activity (Fig. 8, construct 3).
However, the activity attributable to the CCAAT element was only 10%
in the absence of SF-1 sequence (mutation or deletion). Therefore, the
CCAAT box element may participate minimally in the basal
transcriptional activation of PI. In contrast, mutation of the two
adjacent TATA-like sequences positioned at 10 and 23 bp 5
from the
transcriptional initiation site caused marked activation of promoter I
(by 113%).
Fig. 8.
Functional analysis in MLTC of PI promoter
domain by mutation of putative regulatory cis-elements. The
minimal promoter region of PI (700/
549) (construct 1)
and its mutants (constructs 2-4) as well as deletion
mutants (constructs 5 and 6) were transiently expressed in MLTC. The relative promoter activities are indicated as
percentages of the wild-type activity. Data were collected from 3 to 6 independent transfections of triplicate wells and expressed as
mean ± S.E. Arrows indicate the transcriptional
initiation site. Numbers indicate the nucleotide position
relative to the translation initiation site.
[View Larger Version of this Image (20K GIF file)]
700/
549 was also evaluated for its
transcriptional activation in primary cultures of rat granulosa cells and Leydig cells using recombinant adenovirus-reporter gene infection method. In granulosa cells, the wild-type PI promoter domain
(
700/
549) showed 7.6-fold induction of luciferase activity over
that of the promoterless adenovirus-reporter gene construct
(Basic) (Fig. 9, left), an
induction that is about 28% that observed for the control SV40
promoter (SV40). Mutation of the SF-1 element reduced the
activity to 25% wild type (SF1X), indicating that this
factor is also a major transcriptional activator of this promoter in the granulosa cells. In rat Leydig cell, PI promoter domain caused a
significant induction (5.9-fold) over the basic construct, and the SF-1
mutated construct displayed reduced promoter activity to 33% wild type
(Fig. 9, right) but significantly higher than that of the
basic. Thus, the findings in the Leydig cells are consistent with those
observed in MLTC and the rat granulosa cells.
Fig. 9.
Transient expression of adenovirus constructs
of promoter I and its mutants in rat granulosa and Leydig cells.
Rat granulosa cells (left) and Leydig cells
(right) were transiently infected with the adenovirus
constructs containing the minimal promoter domain of
PI(700/
549)/LUC/Adv (wild type) or its SF-1 mutant adenovirus
construct (SF1X). Promoterless (Basic) and SV40 promoter-driven (SV40) adenovirus constructs were also used
for infection of both cell types. The promoter activities are indicated as percentages of the wild-type activity (100%). The data are expressed as mean ± S.E. of three independent infections in
quadruplicate.
[View Larger Version of this Image (31K GIF file)]
-PIII-PI-PII-3
, indicating
that multiple PRLR mRNA species are transcribed from the same PRLR
gene through usage of alternative promoters.
1564/
124)/LUC construct in MLTC and HepG2
cells has not only demonstrated faithful transcription initiation from
this transfected construct but also revealed preferential activation of
PI in MLTC. This is in agreement with the in vivo expression
pattern (1), where PI is utilized in the rat gonads but not in the rat
liver.
700 to
549) as the minimal promoter I domain of the PRLR gene (Fig. 4),
whereas sequences upstream of
700 do not influence the
transcriptional activation of this promoter in transfected MLTC cells.
We demonstrated that steroidogenic factor-1 binds with high affinity to
the sequence CCAAGGTCA (
676/
668) positioned 119 bp
upstream of the transcription initiation site, which is essential for
the activity of this promoter in MLTC cultures as well as in primary
cultures of rat ovarian granulosa and Leydig cells. The position of the
SF-1 element in this promoter domain is consistent with its location in
SF-1-dependent steroidogenic P450 genes (9, 19). SF-1 is a
zinc finger DNA binding protein and is a member of the orphan nuclear
receptor superfamily (9, 20). SF-1 was first identified as a
transcriptional activator for steroidogenic P450 genes (9, 19) and
subsequently was found to be important for a number of other genes
including inhibin, müllerian inhibiting substance, oxytocin, and
gonadotropin
and
subunits (21-25). The targeted disruption of
the Ftz/F1 gene in the mouse has demonstrated that SF-1 is
not only essential for steroidogenesis but is also crucial for gonadal
and adrenal development and sexual differentiation (8). The absence of an SF-1 binding element in the promoter region of PII and PIII (1) may
explain the lack of expression of the liver-specific promoter II in the
gonads and suggests that a different transcriptional mechanism is
involved in the activation of promoter PIII. Conversely, SF-1 protein
is not expressed in the liver or mammary gland, which may explain the
lack of PI activity in both tissues. In contrast, hepatocyte nuclear
factor 4, an abundant transcription factor in the liver, has been
reported to be a transcriptional activator of the promoter II that is
functional in the liver (6). In our study, only PIII but neither PI nor
PII activities was found in the rat mammary gland as revealed by
Northern blot and 5
-RACE PCR analyses of mRNA transcripts,
consistent with the notion that PI and PII are utilized specifically in
the gonads and liver, respectively.
T3-1 cells, mutation of an upstream element in the presence of the
downstream SF-1 element caused 63% reduction in promoter activity,
indicating that SF-1 is not a major transactivator but a co-activator
of the gonadotropin releasing hormone receptor gene promoter (26).
Although the rat luteinizing hormone receptor gene contains an SF-1
consensus sequence located 5
adjacent to its promoter domain, it was
not found to be functional in the rat.2
Furthermore, different mechanisms are involved in the promoter activation of all other pituitary hormone receptors cloned to date.
Therefore, the gene for the PRLR can be viewed as one of a class of
genes regulated by SF-1, unique among those of the receptors for
pituitary trophic hormones. This may imply important functional
correlation between PRLR and other SF-1 regulated gonadal-specific genes in the regulation of reproductive functions.
676/
668) greatly (70-90%) but not completely abolished the promoter activity. Although our current study using DNA-protein binding analyses has not
identified the existence of an additional nuclear protein binding site
within the promoter region that may contribute to the basal promoter
activity, this could not be ruled out. It is probable that the presence
of low affinity or unstable nuclear protein binding may not be readily
detected by these approaches. Among other potential functional elements
present within this promoter, the consensus C/EBP element (
643/
636)
did not bind to C/EBP or participate in the activation of the PI
promoter (not shown). The CCAAT element and the two adjacent TATA-like
sequences present within this promoter located 75 and 10 base pairs,
respectively, from the TSS, appeared to affect the basal promoter
activation (Fig. 8). The disruption of the CCAAT box sequence caused a
minor reduction (30%) in the promoter activity. However, in the
absence of a functional SF-1 element (mutation or deletion), the region containing the CCAAT box showed only ~10% of the wild-type activity. This activity was comparable to that observed in HepG2 cells
transfected with PI(
1566/
124)/LUC. However, the minimal activity
presumably induced by the CCAAT element in transfected HepG2 cells was
not effective for transcriptional activation in vivo since
expression of the exon E11 was not observed in liver or
mammary gland. The pentanucleotide CCAAT sequence is commonly found
60-80 bp upstream of the TSS (27). A number of nuclear proteins that
can specifically bind to the CCAAT element have been identified,
including human CCAAT-binding proteins CP1, CP2, and NF-1 (28) and the
family of CCAAT/enhancer-binding protein (C/EBP
,
,
,
,
, and CHOP10) (29). However, the possibility that member(s) of this
large family of proteins may interact with the CCAAT element in PI
promoter and affect its basal activity remains speculative. In
addition, the TATA-like elements and presumably their associated
proteins may exert constitutively down-regulatory influence(s) in
promoter I function, conceivably through interaction with SF-1 or the
preinitiation complex, whose formation should be presumably independent
of the TATA boxes in this promoter. Therefore, the PRLR promoter I
belongs to the class of TATA-less/non-initiator, GC box-less gene
promoters.
*
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
¶
To whom correspondence should be addressed: Bldg. 49, Rm. 6A
36, 49 Convent Dr. MSC 4510, National Institutes of Health, Bethesda, MD 20892-4510. Tel.: 301-496-2021; Fax: 301-480-8010.
1
The abbreviations used are: PRLR,
prolactin receptor; SF-1, steroidogenic factor-1; 5-RACE, rapid
amplification of cDNA 5
ends; RT-PCR, reverse
transcriptase-polymerase chain reaction; bp, base pairs; kb,
kilobase(s); DTT, dithiothreitol; LUC, luciferase; MLTC, mouse Leydig
tumor cells; EMSA, electrophoresis mobility shift assay; C/EBP,
CCAAT/enhancer-binding protein; TSS, transcriptional start site; Adv,
adenovirus.
2
C.-H. Tsai-Morris and M. L. Dufau, unpublished
observation.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.