| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Molecular Physiological Chemistry (S.K.H., H.T., A.W., M.O.), and Basic Laboratory Science (Y.N.), Osaka University Medical School, Suita, Osaka, 565 Japan; and Department of Anatomy (O.H.), Nara Medical University, Kashihara, Nara 634, Japan
Address all correspondence and requests for reprints to: Mitsuhiro Okamoto, Department of Molecular Physiological Chemistry, Osaka University, Medical School, 2–2 Yamadaoka, Suita, Osaka, 565, Japan. E-mail: mokamoto{at}mr-mbio.med.osaka-u.ac.jp
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The final steps of the production of both glucocorticoid and mineralocorticoid are catalyzed by cytochrome P450(11ß), i.e. steroid 11ß-hydroxylase [P450(11ß)] family enzymes. In the adrenal cortex of ungulates [such as cattle (2, 3), pigs (4), and sheep (5)] and of amphibians [such as bullfrogs (6)], a single P450(11ß) enzyme produces aldosterone and cortisol in the ZG and in the ZF/ZR, respectively, in different catalytic manners. In contrast, two isoforms are present in the adrenal gland of rodents [such as rats (7, 8), mice (9), and hamsters (10, 11)] and of humans (12, 13). One, cytochrome P450(aldo), i.e. aldosterone synthase [P450(aldo)], is responsible for the production of aldosterone in the ZG and the other, P450(11ß), is liable for the production of glucocorticoids in the ZF/ZR. The expressions of these P450(aldo)/P450(11ß) genes are controlled by different signaling pathways in the respective zones. The molecular mechanisms underlying the specifically differentiated regulations of the steroidogenesis are poorly understood.
Electron-microscopic observation of rat, mouse, and dog adrenal cortex suggested that the ZG could be further divided into two layers. The cells in the first layer, adjacent to the capsule, have elongated mitochondria, whereas those in the second layer, sometimes called the zona intermedia (ZI), have small and polymorphic (round, annular, or elongated) mitochondria (14). In the normal rat, these mitochondria have tubulovesicular, plate-like, and/or straight tubular cristae, each mitochondrion containing one or more types. The mitochondrial features of the ZG cells have been reported to reflect the degree of aldosteronogenesis, taking place in different types of mitochondria (15).
Immunohistochemical studies, performed by using the antibody specific to P450(aldo) or that specific to P450(11ß), have demonstrated the presence of two types of cell in the rat ZG, the cells expressing P450(aldo) and those expressing neither P450(aldo) nor P450(11ß) (16). The former seemed to be identical to those having elongated mitochondria. Stimulation of glomerulosa function by Na+-depleted diet seemed to coincide with an increase in the number of these cells. Thus, they are considered as those in a steroidogenically active stage. The latter cells, apparently negative for steroidogenic activity, seemed to correspond with those of the ZI. Because most proliferating cells were present in this zone, these cells have been considered undifferentiated stem cells.
When rats are subjected to bilateral adrenal enucleation surgery, the cells in residual capsule regenerate within a month, and they form a glomerulosa-like outer zone and a fasciculata/reticularis-like inner zone (17). This suggests that cells attached to the adrenal capsule, mainly composed of the ZG cells, have potential to act as stem cells of the cortical differentiation, being capable of converting into all three zones. The investigation carried out by Engeland et al. (18) suggested that these cells, at the early stage of adrenal regeneration (within 1–3 days after the surgery), were P450(aldo)-negative, and their mitochondria did not look like those of aldosteronogenic cells. It seems that most of the capsule-attached cells were converted immediately after the surgery to those negative for steroidogenic activity, like the ZI cells. However, most cells that were positive for proliferating cell markers, scattered in the regenerating adrenal, were positive for P450(11ß), a marker for the ZF/ZR cells. These observations suggest that the proliferating cells in the regenerating adrenal cortex may be differentiation-initiated ZF/ZR-like cells, rather than the stem cells.
As the first step to elucidate the mechanism underlying the zonal differentiation of adrenal cortex, we have attempted to isolate proteins specifically expressed in the rat ZG. Developing a new simple complementary DNA (cDNA) subtraction method, we successfully isolated three distinct cDNA fragments. Here, we report the characterization of a protein encoded by one of them, ZG specific clone (ZOG), a rat homolog of preadipocyte factor-1 (Pref-1). Immunohistochemical studies demonstrated that most cells in ZG, including the P450(aldo)/P450(11ß) double-negative ZI cells, expressed Pref-1 protein. Moreover, we found that there was a cell layer between ZG and ZF that was Pref-1-, P450(aldo)-, and P450(11ß)-negative.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All procedures were followed, according to the guidelines for animal care of the Medical School of Osaka University.
High-Na+ and high-K+ diet treatment
One group of rats (5 weeks old, male) was given an
Na+-enriched diet (containing 8% NaCl, Oriental Co.,
Chiba, Japan) with tap water, and the other group was given a
low-Na+ diet (containing 0.3% NaCl) with
K+-loaded water (150 mM KCl) for 7 days.
Dexamethasone and ACTH treatment
Dexamethasone phosphate (DX) was purchased from Banyu Seiyaku
Co., Tokyo, Japan, and ACTH was from Daiichi Seiyaku Co., Tokyo, Japan.
Male 5-week-old rats were separated into three groups. Those in the
first group were injected ip with 0.1 mg DX, dissolved in 0.6% NaCl,
daily for 7 days. The second group was injected ip with 0.1 mg ACTH for
7 days, and the third was also injected ip with vehicle for 7 days.
Adrenal regeneration experiment
Male rats (6 weeks old) were separated into two groups; one
group was used as control, and the other group was used for adrenal
regeneration. The rats were subjected to anesthesia (Nembutal, Dainabot
Co.), and the adrenal glands were bilaterally enucleated. Enucleation
was performed by splitting the capsule and extruding the inner cortical
and medullary tissue. After surgery, rats were given laboratory chow
and tap water under the standard maintenance conditions. At 3, 7, and
30 days after the surgery, their adrenal glands were isolated and
separated for Northern blot analysis and for immunohistochemistry.
mRNA purification and double-strand cDNA synthesis
All adrenal tissues or cells were frozen in liquid nitrogen
until use. Total cellular RNAs were isolated using the guanidine
thiocyanate method (19). About 10 µg polyA-rich RNA was purified from
0.5 mg of total RNA fraction, using an mRNA isolation kit (Takara,
Kyoto, Japan), and was used for cDNA synthesis employing an oligo
(deoxythymidine) primer and avian myeloblastosis virus reverse
transcriptase (Amersham, Arlington Heights, IL). Then the second-strand
synthesis was conducted by means of ribonuclease H and DNA polymerase I
(Amersham).
PCR-coupled cDNA subtraction and isolation of three glomerulosa
specific clones
The double-stranded cDNA populations from K+-induced
capsule and Na+-induced decapsulated portions were prepared
from 1 µg mRNA and were designated Tester (containing target cDNA)
and Driver, respectively. Tester and Driver were then digested with
Sau 3A1 (a four-base cutter restriction endonuclease). One
microgram of Tester was then ligated to 0.5 nmol of 24 mer (24 mer T,
5'-AGTGACTACTGCAGACCGTGAAAG-3') and of 15 mer (15 mer T,
5'-GATCCTTTCACGGTC-3') dephosphorylated oligonucleotides (adaptor set
T) in 60 µl reaction volume with Escherichia coli T4 DNA
Ligase (Takara). After ligation, Tester was ethanol precipitated and
resuspended in 50 µl distilled water. Driver was ligated to 0.5 nmol
of 24 mer (24 mer D, 5'-ACCGAAGCTTGTAGACTGTCTACG-3') and of 15 mer (15
mer D, 5'-GATCCGTAGACAGTC-3') dephosphorylated oligonucleotides
(adaptor set D) and treated similarly.
Ten microliters of adaptor-ligated Tester DNA was amplified by PCR in
100 µl reaction volume containing 10 µl PCR reaction buffer, 8 µl
2.5 mM deoxynucleotide triphosphate mixture, and 1 µl
(100 pmol) of the 24 mer T. After heating to 72 C for 3 min in a
thermal cycler (Perkin Elmer Cetus, Foster, CA) and adding 15 U
Taq DNA polymerase (Takara), the mixture was overlaid with
mineral oil and incubated for 5 min at 72 C to fill in 5'-protruding
ends. The PCR was carried out in 20 cycles (94 C, 45 sec; 61 C, 1 min;
72 C, 3 min) with the last cycle followed by an extension at 72 C for
10 min. The amplified Tester (0.5 µg) was mixed with 2.5 µg Driver
(in a ratio of 1:5), and the DNA was phenol/chloroform-extracted and
ethanol-precipitated. The DNA was then resuspended in 4 µl EEx3
buffer [10 mM N-(2-hydroxyethyl)
piperazine-N'-(3-propanesulfonic acid) (HEPPS, Sigma, St.
Louis, MO), pH 8.0 at 20 C, 1 mM EDTA (EE)], and 1 µl of
5 M NaCl solution was added. The solution was overlaid with
mineral oil and heat denatured at 95 C for 5 min. The DNA was
hybridized (for 20 h) at 68 C. At the end of hybridization, DNA
was ethanol-precipitated and resuspended in distilled water. There are
three types of hybridized cDNA populations at this step: 1)
Tester/Tester; 2) Tester/Driver; and 3) Driver/Driver. Target clones of
interest should accumulate in the Tester/Tester self-hybrid
populations. The ends of hybridized cDNAs of Tester/Driver (produced
from cDNAs expressed nonspecifically) could not make double strand
because of the difference in sequences between 24 mer oligonucleotide
adaptor primers. To remove these single-strand adaptor portions,
hybridized products were treated with 15 U mung bean nuclease (Takara)
at 37 C for 30 min in 20 µl reaction volume. The digested sample was
diluted (1:4) in 50 mM Tris-HCl (pH 8.9), inactivated by
heating at 95 C for 5 min, and chilled on ice. The target cDNAs
(Tester/Tester hybridized population) were specifically amplified by
PCR in 20 cycles with 24 mer oligonucleotide (24 mer T) under the same
conditions as described above. The amplified product was
phenol/chloroform-extracted, ethanol-precipitated, and resuspended in
TE (10 mM Tris-HCl, pH 8.0, and 1 mM EDTA),
giving the first subtracted product. A sample (0.5 µg) from the first
subtracted product was mixed with 2.5 µg Driver amplicon, and the
subtraction was conducted again as mentioned above. After the third
cycles of the hybridization-amplification procedures, products were
phenol/chloroform extracted, ethanol precipitated and resuspended in 20
µl distilled water, and digested with Sau 3A 1 (20 U) in a
40-µl reaction volume. The digested products were then
ethanol-precipitated, resuspended in water, and separated in 1%
agarose gel. The products having 100–400 bp length were recovered from
the gel and purified by a spin column (Qiagen, Chatsworth, CA). The
purified DNA fragments were cloned into the BamHI site of
pUC 18 for sequencing. We successfully isolated three independent cDNA
clones that were specifically expressed in the ZG. Because one clone,
ZOG, showed the strongest hybridization signal in the Northern blot
(Fig. 1
), this clone was subjected to
further investigation.
|
MosElox cloning vector (Amersham). For screening, 4 x
105 phages were plated and replicas were made from the
resultant plaques (19). Positively hybridizable plaques were screened
with [32P]-labeled cDNA fragment isolated from
subtraction hybridization. Seven independently isolated clones had
almost the same length of cDNA insert (1.6 kb). The longest
EcoRI-digested cDNA fragment was subcloned into pBluescript
KS (-) (Stratagene, La Jolla, CA). The recombinant plasmid was named
pBlue-ZOG and was used for sequencing.
Northern blot analysis
We used the XhoI-EcoRV cDNA fragment of
pBlue-ZOG and the XhoI-BamHI fragment of
pcP-45011ß-62 (21), which hybridized with both P450(11ß) and
P450(aldo) messages, for Northern blot analyses. To prepare the cDNA
probe for rat cytochrome P450 cholesterol side chain cleavage
[P450(SCC)], we amplified the cDNA fragment by RT-PCR using
5'-GGGAATTCTCCTGCGAGGGTCCTA-3' and 5'-GGAAGCTTGAGAGGCTGGAAGTTG-3'
oligonucleotides from the total RNAs extracted from rat adrenal grand.
The product was digested with EcoRI and HindIII
and was cloned into the EcoRI and HindIII sites
of pET28a plasmid vector (Novagen, Madison, WI). The resultant plasmid
was named pETrSCC. For the control cDNA probe, we used the rat
glyceraldehyde-3-phosphate-dehydrogenase (G3PDH) cDNA, which was also
amplified by RT-PCR with the primer set commercially supplied from
CLONTECH, and cloned into the pMOS-T vector (Amersham). Steroidogenic
factor-1 (Ad4BP/SF-1) cDNA was kindly given by Dr. K. Morohashi at the
National Institute of Basic Biology, Okazaki, Japan. Total RNAs (10
µg) were size-fractionated on 1% agarose gel in 2.2 M
formaldehyde, 20 mM
3-[N-morpholino]propanesulfonic acid buffer (pH 7.0), 1
mM EDTA, stained with ethidium bromide, and transferred to
nylon membrane (Amersham). Filters were hybridized at 42 C for 16
h under identical conditions, with the addition of 1 x
107 cpm/ml [32P]-labeled random primed cDNA
probes. After hybridization, filters were washed and exposed to x-ray
film (RX, Fuji Film Co., Ltd., Tokyo, Japan), at -80 C, with an
intensifying screen.
Competitive RT-PCR analysis
The size and the sequence of mRNA for rat P450(11ß) and
P450(aldo) are so similar that the signals for individual messages
cannot be distinguished from each other in Northern blot analysis. To
estimate the ratio of P450(aldo) mRNA to P450(11ß) mRNA, we used
competitive RT-PCR analysis (22). Total RNAs were prepared from adrenal
glands of normal, Na+-, and K+-induced rats, as
follows. One microgram of total RNAs was reverse-transcribed by 200 U
Super Script RT (GIBCO BRL, Rockville, MD) in the presence of 150 ng
random hexanucleotides in a 20-µl reaction mixture, according to the
protocol recommended by the manufacturer. The products were heated
at 70 C for 5 min and were used for PCR. An aliquot of the reaction
mixture (4 µl) was added to a 45-µl mixture containing 50 pmol
P450(11ß)/P450(aldo) common forward (5'-CGTGTCACGTCCCCCTCG-3')
and reverse (5'-CATACTCAGAGATGACAT-3') primers, 4 µl of 2.5
mM deoxynucleotide triphosphate mixture, 5 µl of 10x
buffer, and 1 µl (5 U) Taq DNA polymerase (Takara). The
reaction was subjected to 30 cycles of PCR, at 94 C for 40 sec, at 58 C
for 1 min, and at 72 C for 1 min, and with the final extension reaction
for 5 min. The amplified DNA fragments were fractionated on 1% agarose
gel, stained with ethidium bromide, transferred to nylon membrane
(Amersham), and subjected to Southern blot analysis. Filters were
hybridized at 42 C for 16 h with [32P]-labeled
P450(11ß) specific oligonucleotide (5'-TAAACATTCAGTCCAATA-3'), or
P450(aldo) specific oligonucleotide (5'-TGGATGTCCAGCAAAGTC-3').
Hybridized filters were washed and exposed to x-ray film (RX, Fuji Film
Co., Ltd., Tokyo, Japan), at -80 C, with an intensifying screen.
Preparation of antibody
The peptide of 139 amino acid residues from the C-terminal
side of the sixth epidermal growth factor (EGF)-like repeat was
expressed in Escherichia coli and was used as an antigen to
produce anti-Pref-1 antibody. To construct the expression plasmid, we
used Pref-1 cDNA, which was amplified by the specific forward primer
linked to EcoRI adaptor
(5'-GGGAATTCGCGAAGAAGCGCGGGAC-3',
nucleotide position, 868–885) and the reverse primer linked to
HindIII adaptor (5'-GGCAAGCTTAGATATCCTCATCACC-3', nucleotide
position, 1263–1290). The reaction was subjected to 20 cycles of PCR
at 94 C for 40 sec, at 60 C for 1 min, and at 72 C for 1 min, with the
final extension reaction being for 5 min. The amplified DNA fragment
was digested with EcoRI and HindIII and ligated
into EcoRI and HindIII sites of pET28a (Novagen).
The positive clones were selected, and one of them was named pET-ZOGc.
Escherichia coli, BL-21(DE3) (Novagen), was transfected with
pET-ZOGc plasmid, isopropyl-ß-D-thiogalactopyranoside
(final concentration, 1 mM) was added at the logarithmic
growth phase (OD600 = 0.3–0.5) and was incubated for
3 h. The harvested Escherichia coli was suspended in
homogenizing buffer [10 mM Tris (pH 8.0), 1 mM
EDTA, 100 mM NaCl, and 1% Triton-X100] and was sonicated.
The expressed Pref-1 protein was precipitated as an inclusion body by
centrifugation at 30,000 x g for 1 h. The pellet
was resuspended in homogenizing buffer without Triton-X100 and was
subjected to SDS-PAGE analysis. The separated proteins were stained
briefly with brilliant blue R (Katayama, Co. Ltd., Osaka, Japan), and
the region of the gel containing the target protein was recovered, cut
into small pieces, and put into dialysis bags (Nakarai Chemicals, Ltd,
Japan) with 3 ml SDS-PAGE running buffer [25 mM Tris, 150
mM glycine (pH 8.3), and 0.1% SDS]. The electrophoresis
was used to elute the recombinant protein from the gel. Four hundred
micrograms of the purified protein was well emulsified with an equal
volume of complete adjuvant and injected into two rabbits (2 kg each).
The same animals were further injected with 200 µg of the purified
protein emulsified with incomplete adjuvant. With an interval of 2
weeks, blood was collected from the animals, and serum (containing
anti-Pref-1 antibody) was separated. The anti-Pref-1 IgG was purified
using a protein A agarose column (Amersham), according to the
manufacturer’s recommendation, was precipitated with 30% ammonium
sulfate, and was dialyzed in PBS (pH 7.4). Three milligrams of
anti-Pref-1 IgG was labeled with biotin, using a protein-labeling kit
(Boehringer Mannheim), and was concentrated to 0.3 mg/ml.
To prepare an antigen for making anti-P450(11ß)/P450(aldo) antibody, we expressed recombinant P450(11ß) protein in Escherichia coli transformed with pTrc11ß expression vector (Nonaka Y., T. Fujii, N. Kagawa, M. R. Waterman, and M. Okamoto, unpublished data), and purified the expressed protein. The antigen for rat P450(SCC) was also prepared, using an Escherichia coli-expression system with pETrSCC expression vector. Anti-Ad4BP/SF-1 and anti-inner zone antigen (anti-IZA) IgG(s) were generously given by Dr. K. Morohashi and Dr. G. P. Vinson, respectively.
Immunohistochemical analysis
Twelve adrenal glands at each embryonic stage, 17.5 days
postcoitum (d.p.c.) and 20.5 d.p.c., and 12 control adrenals from
6 rats (8 weeks old), were fixed in 2% paraformaldehyde-PBS (pH7.4)
for 2 h. Serial frozen sections (6 µm) were prepared on glass
slides and stained immunohistochemically (23). For double
immunostainings, the sections were incubated with the antiserum to
P450(11ß/aldo), diluted 1:1,000 for 12 h, followed by Cy3
(yellow)-labeled goat antirabbit IgG (1:300) for 2 h. After
washing in PBS, the sections were incubated with 50% normal rabbit
serum-PBS for 1 h, the biotin labeled IgG to Pref-1 (0.015 mg/ml)
for 2 h, and then fluorescein isothiocyanate (green)-labeled
streptavidin (1:50) for 2 h. Photographs were taken using a
fluorescent microscope (Carl Zeiss, Oberkochen, Germany).
Preparation of Pref-1 protein-expressing Y-1 transformants
To prepare the expression vectors of Pref-1 protein, we first
digested the plasmid pBlue-ZOG with HindIII and
Nru I. As a result of this double digestion, a short
fragment was removed from the 5' end of Pref-1 cDNA. This digested
plasmid was fractionated on agarose gel, recovered, and purified. Two
designed oligonucleotides for an adaptor (5'-AGCTTCATATGATCG-3' and
5'-CGATCATATGA-3') were annealed and ligated to the
HindIII/Nru I digested plasmid at 16 C overnight.
Plasmids containing the correct sequence were further digested with
HindIII and XbaI and recloned into pRc/RSV. The
resultant plasmid, named pRc-rZOG, was used for preparation of Pref-1
protein-expressing stable Y-1 transformant.
To isolate stable transformant, mouse Y-1 cell line was plated at 2 x 106 cells per 10-mm dish in DMEM containing 10% FCS (GIBCO BRL). Before the confluent stage of growth, cells were trypsinized, collected, and transfected with plasmid DNA, which was digested with Xnm I (20 µg per dish) via the electrotransfection method (3). The cells were selected for 2 weeks in DMEM containing 400 µg/ml genetecine (neomycin) (GIBCO BRL). The genetecine resistant cells were then collected from the culture plates by trypsinization and were analyzed for the expression of Pref-1 protein by immunoblotting (19). The plated genetecine resistant cells were also fixed in PBS containing 4% paraformaldehyde (Merck, Darmstadt, Germany) for 15 min at room temperature and were immunostained with rabbit polyclonal anti-Pref-1 antibody, as mentioned above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, and its cloning are described under Materials and Methods. As an
attempt to isolate ZG-specific cDNAs, we employed PCR-coupled
subtraction hybridization technique, a modified method of genomic
representational difference analysis (24). The original
representational difference analysis was developed as a method to
differentially identify the specific regions in genomic DNA, such as
large deletion or insertion (24, 25). This method was successfully
adapted to isolate ZG-specific genes. But our method had the
disadvantage of most of the isolated clones being expressed little in
the tissues. Because several mRNAs in the ZG, such as those for
P450(aldo) (7) and angiotensin II receptor (AT1B) (26), are highly
similar in sequence to those in the ZF, such as those for P450(11ß)
(21) and angiotensin II receptor (AT1A) (27), the isolation of these
clones failed. For differential subtraction, we mixed the Driver cDNAs
and the Tester cDNAs at a ratio of 5:1. Because the capsular portion of
the adrenal gland used for preparation of the Tester cDNAs may
inadvertently contain a few fasciculata cells, ratios of the Driver to
the Tester higher than 5:1 were not suitable for isolation of
ZG-specific clones. Hence, a preliminary test performed in the ratio
50:1 could not result in isolation of any clone (data not shown).
The full-length cDNA isolated from the capsular portion, shown in Fig. 2A, was composed of 1,560 nucleotides,
with a polyadenylation signal sequence (AATAAT) near the poly (A) tail
(28). The protein encoded by the DNA had 383 amino acid residues with a
molecular mass of approximately 41 kDa, having a feature of a single
membrane-spanning protein with a putative signal peptide, six EGF-like
repeat motifs, a transmembrane domain, and a short cytosolic stretch.
Three potential N-linked glycosylation sites were present in its
extracellular domain. Database search revealed that this protein was
highly homologous to four previously reported proteins, such as
pheochromocytoma-specific protein (pG2) (29), human fetal antigen 1
(FA-1) purified from the amniotic fluid (29, 30), small cell lung
carcinoma-specific protein (
-like protein, dlk) (31), and mouse
Pref-1 (32). These independently identified proteins are a little
different from one another in their sequences but are thought to be
produced from a single gene by means of alternative splicing and
posttranslational modification. Recently, a clone for the rat homolog
of Pref-1 was isolated by using a PCR amplification by Carlsson
et al. (33). Our clone differed from the published rat
Pref-1 in six nucleotides with four amino acid substitutions. Because
these amino acid substitutions found in our clone, excepting Ala 323,
are conserved in the sequences of mouse and human homologs (Fig. 2, A
and B), we concluded that our sequence may be correct. The reported
sequence may contain the artificial mutations caused by the
PCR-mediated cloning.
|
), indicated the absence of
Pref-1 expression in the fasciculata-derived Y-1 cell line.
|
).
The smearlike appearance of Pref-1 protein may be explained by the
posttranslational glycosylations. When the cell homogenates were stored
at -20 C for 24 h, most immunopositive bands disappeared,
suggesting that the expressed Pref-1 protein was rapidly degraded by
proteolysis. Y-1 cells that stably expressed Pref-1 protein were
selected and stained with anti-Pref-1 protein antibody. Most cells were
strongly positive in the cytosol (Fig. 3C). At higher magnification,
the level of immunostain near the cell’s nucleus seemed to be greater
than that near the plasma membrane (not shown). Immunohistochemistry of
Y-1 cell transformants showed that Pref-1 protein existed in the
cytoplasm, rather than at the plasma membrane. Because the antibody was
raised against the C-terminal peptide fragment, it was reasonable that
the antibody could not recognize an extracellular domain of Pref-1
protein. However, it should be pointed out that immunohistochemical
study performed using the other antibody, which was raised against a
peptide including the EGF-repeats, also provided similar staining in
human adrenal ZG cells. When the full-length Pref-1 protein was
expressed in COS cells (a cell line lacking intrinsic Pref-1), a 25-kDa
cell-associated protein was detected by anti-Pref-1 antibody. This
protein was identified as the residual protein after cleavage and
release of some region of the extracellular domain (34). The picture
showed that control Y-1 cells were immunonegative (Fig. 3D). These
results established the reliability of anti-Pref-1 protein
antibody.
Immunohistochemistry of adult adrenal glands
The study was undertaken to identify Pref-1 protein-expressing
cells in adult adrenals. Strong signals for Pref-1 protein were found
in ZG and medulla (Fig. 4A). The
medullary expression of Pref-1 protein well agrees with the findings of
pG2 mRNA expression in human pheochromocytoma (35) and dlk mRNA
expression in rat PC12 cells (31). Because human pheochromocytoma,
tumor derived from mature adrenal chromaffin cells, expresses Pref-1
mRNA (not shown), the immunoreactivity in the rat adrenal medulla may
be specific. As expected, an antibody that recognized both P450(11ß)
and P450(aldo) stained the cells in ZF/ZR, as well as those in the
outermost part of ZG (Fig. 4B).
|
; C, D, and E). The antibody against Pref-1 protein stained most cells
in the ZG, as well as those in the medulla (Fig. 4, C and E). In
contrast the anti-P450(11ß)/(aldo) antibody stained only the
outermost layer of the ZG and most cells in the ZF (Fig. 4, C and D).
This suggests that the cells containing aldosterone synthase P450,
P450(aldo), were present in a few-cells-thick layer of ZG underneath
the capsule, whereas the cells containing 11ß-hydroxylase,
P450(11ß), were distributed over ZF. It should be noted that the
cells expressing P450(aldo) also contained Pref-1 protein, but not all
the cells expressing Pref-1 protein contained P450(aldo). Thus, the
width of the cell layer expressing Pref-1 protein was wider than that
expressing P450(aldo). These observations were confirmed in all adrenal
glands examined.
To our surprise, a cell layer free of Pref-1 protein, P450(aldo),
or P450(11ß) was found between the ZG and the ZF (indicated by a
star in Fig. 4, C and D). This finding suggests that a zone
recently reported by Mitani et al. (16) that was
immunonegative for P450(aldo) and P450(11ß) was, in fact, composed of
two layers (one devoid of both Pref-1 protein and P450s, and the other
lacking P450s but containing Pref-1 protein).
Immunohistochemical study of embryonic adrenal glands
To examine whether Pref-1 protein was expressed during embryonic
development, we double-stained the adrenal tissues on embryonic
days 17.5 d.p.c. (Fig. 5, A and D)
and 20.5 d.p.c. (Fig. 5, B and E). The green stains
were the immunoproducts with biotin-labeled anti-Pref-1 protein
antibody, and the yellow stains those with
anti-P450(11ß)/(aldo) antibody. On the embryonic stage 17.5
d.p.c., Pref-1 protein had already been expressed in cells that were
distributed at the periphery of the adrenal gland. There were no
P450(aldo)-positive cell layer at the outermost part of the adrenal
gland. There were a few cells, scattered inside the cortical tissue,
that expressed both Pref-1 protein and P450(11ß) (shown by
arrowheads in Fig. 5D). Several aggregating cells,
expressing only Pref-1, were found inside the adrenal (shown by
asterisks in Fig. 5A). These latter cells, at the innermost
part of adrenal tissue, were probably of the medullary origin,
because they were stained with anti-tyrosine hydroxylase antibody (data
not shown). The cell layer devoid of both Pref-1 protein and P450s was
not found on the day 17.5 d.p.c. (Fig. 5, compare D with F).
|
, B and E) were very similar to those of the adult
adrenal cortex (Fig. 5, C and F), with the exception that the outermost
cells expressing both Pref-1 and P450(aldo) were fewer in the embryo
than in the adult. The cell layer devoid of both Pref-1 protein and
P450s was not clearly found on day 20.5 d.p.c. (Fig. 5, B and
E).
Effects of a high-Na+ or high-K+ diet on
the expression of Pref-1 message
Northern blot analysis, using total RNAs prepared from adrenal of
rats given a normal, high-Na+, or high-K+ diet,
is shown in Fig. 6. The message level of
Pref-1 was fairly constant among the three groups of animals. By
performing Southern hybridization of RT-PCR products, we could
distinguish the message for P450(aldo) from that for P450(11ß). As
expected, the level of P450(aldo) message of Na+-loaded
animals was significantly lower, and that of K+-loaded
animals was higher, than that of control animals. In contrast, the
level of P450(11ß) message was apparently constant among the three
groups. The significant induction of P450(aldo) in
Na+-restricted and K+-loaded rats has been
previously reported (36, 37). These results suggest that Pref-1 protein
may not directly regulate the expression of P450(aldo).
|
). The P450(aldo) messages
could not be detected in the ZG of DX-treated animals, whereas that in
the ACTH-treated animals was less than that in the control. These
results of control hybridization agree well with a previous report
indicating that the level of P450(aldo) message, detected by in
situ hybridization analysis, in adrenals of chronic stress rats
was lower than that of normal rats under the influence of plasma ACTH
(38). The level of Pref-1 message in DX-treated animals was a little
higher than that in the control, whereas the ACTH-treatment little
affected the level. The message for side-chain cleavage P450
[P450(SCC)] significantly decreased after the DX-treatment, but that
seemed to be little affected by the ACTH-treatment. These results
suggest that the level of Pref-1 message in ZG was fairly constant
under the pituitary hormonal influence.
|
When the major portion of cortex and the entire medulla are
surgically excised from rat adrenal gland, cells attached to the
remaining capsule, mostly ZG cells, begin to proliferate and
differentiate into ZF/ZR-like cells, an original mass of the gland
being recovered within a month. To gain an insight into the
physiological role of Pref-1 during the adrenal regeneration, we
examined the expression of its message. On the day 3, the messages for
P450(SCC) and Ad4BP/SF-1 decreased to one third of the ZG fraction (day
zero) and gradually recovered during the course of regeneration (see
Fig. 8A). The level of the Pref-1 message
in the capsule taken on day 3 after the surgery was quite low (less
than one tenth) compared with that taken before the surgery, and it
reached the control total adrenal level on day 7. Taking into account
the contaminating RNAs derived from fibrin clot in the regenerating
adrenals, we normalized the levels of Pref-1 message by those of
Ad4BP/SF-1 message (Fig. 8B). Although the relative amount of Pref-1
message in the ZG fraction on day zero decreased on day 3 after the
surgery, the specific message level at this stage was not significantly
lower, compared with those of the control total adenal and the
regenerated adrenal (day 7 and day 30). These data suggest that the
expression of Pref-1 message may be suppressed but maintained at the
level of the total normal adrenal cortex in a homeostatic manner during
the regeneration.
|
). Compared with the
sham-operated adrenal ZG (Fig. 9A), the regenerating adrenal (Fig. 9B)
contained a very low amount of Pref-1 protein. Only a few cells were
positively stained with the antibody. When the adrenal sections were
treated with anti-Ad4BP/SF-1 antibody (Fig. 9, C and D), the
immunopositive cells, exclusively stained in their nuclei, were
distributed over all the adrenocortical zones. It should be noted that
the fluorescence intensity was equal, irrespective of the cell’s
location. As expected, anti-P450(SCC) antibody exclusively stained the
cytoplasm of all the cortical cells in the normal adrenal (Fig. 9E),
even though the intensity in ZG seemed to be lower than that of ZF. In
the regenerating adrenal of day 3 (Fig. 9F), all the cortical cells
were immunopositive for P450(SCC). Next, the adrenals were treated with
monoclonal antibody against IZA, which had been reported to
specifically recognize the ZF/ZR cells (39). The photomicrograph of
Fig. 9G confirmed that the antibody stained the cells in the ZF but not
those in the ZG. In the regenerating adrenal of day 3 (Fig. 9H),
several cells were stained with this antibody, indicating the presence
of ZF-like cells at the early phase of adrenal regeneration. Moreover,
some of the cells in the regenerating adrenal were found to be
immunopositive for both Pref-1 protein and IZA (not shown), suggesting
that cells intermediate between the ZG and ZF/ZR cells were present at
the early phase of regeneration.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
With the characteristic spacing between cysteine residues and the presence of a proline residue preceding the second cysteine residue in the third EGF-like repeat, the Pref-1/dlk/FA-1 protein was strikingly similar to those of Drosophila homeotic gene products, Delta (40), Serrate (41) and Notch (42), and of C. elegans Lin-12 (43) and Glp-1 (44). Delta and Notch, interacting with each other via their EGF-like domains, are implicated in cell-fate determination during the animals’ development. Recently, mammalian homologs of Delta (45), Serrate (46, 47), and Notch (48, 49) were also isolated; and it was experimentally shown that overexpression of these molecules suppressed muscle cell differentiation in culture medium (46) and involved in mammalian neurogenesis (50). The mutations that alter the amino acids in EGF-like repeats of Notch3 caused not only abnormal differentiation but also produced a disease called cadasil, a disorder characterized by stroke and dementia (51, 52). Furthermore, mouse mammary tumor virus late promoter, fused to the Notch-related mouse gene int-3, could produce mammary tumors (53). Taking the structural similarity between these proteins and Pref-1 into consideration, the latter may have a role in the cell-fate determination during the organogenesis of adrenal gland and/or the adrenocortical zonal differentiation.
Immunohistochemical evidence confirmed that the Pref-1 protein-expressing glomerulosa zone in adult rats was actually composed of two layers of cells; the first layer underneath the capsule contained Pref-1 protein and P450(aldo), and the second layer beneath the first contained Pref-1 protein but not P450(aldo). This suggests that the functionally active aldosteronogenic cells are, in fact, present at the outermost part of the ZG, and they may differentiate from the cells underneath that possess only Pref-1 protein. Mitani and co-workers (16) have shown a cell layer between the ZG and ZF that could be stained with neither P450(aldo) nor P450(11ß) antibodies. Although the physiological function of the double negative cell layer remains to be clarified, the unstained zone was shown to contain replicating cells. Based on these findings, the investigators proposed that the unstained zone was in an undifferentiated stage, in terms of steroidogenic activity. They also proposed that this zone was the stem cell layer of rat adrenal cortex. We could divide the P450(11ß)/(aldo) double negative cell layer into an additional two sublayers: one was Pref-1 positive and the other, Pref-1/P450(11ß)/(aldo) triple negative. It is conceivable that the latter contains either very primitive cells or differentiation-initiated cells from the Pref-1-expressing ZG cells could differentiate into the P450(11ß)-expressing ZF.
Immunohistochemical study of the embryonic adrenal glands proposed
further information. First, the outermost part of the glomerulosa zone
on embryonic day 17.5 d.p.c. did not possess the aldosteronogenic
cells, whereas the cells expressing Pref-1 protein were already
present. That the aldosteronogenic cells were clearly found at
embryonic day 20.5 d.p.c. strongly suggests that these cells came
from the Pref-1-expressing cells during the zonal differentiation of
adrenal gland. Previous study also showed the absence of
aldosteronogenic cells in an embryonic stage between 16.5 and 17.5
d.p.c. (54). Second, several cells possessing both Pref-1 protein and
P450(11ß) were scattered at the ZG/ZF border in the cortical tissue
at 17.5 d.p.c. (indicated by an arrow in Fig. 5D), and
these cells were no longer found on day 20.5 d.p.c. (Fig. 5E),
suggesting the differentiation of a part of the Pref-1-expressing cells
into the functionally active P450(11ß)-expressing ZF cells via an
intermediate cell types that expressed both Pref-1 and P450(11ß).
Observations with regenerating adrenal experiments presented in this
paper also support the above hypothesis.
The expression of Pref-1 message was suppressed at an early stage of adrenal regeneration after the enucleation. The level of the message seemed to be restored to the normal level within a week. Only a few cells in the regenerating adrenal, on day 3 after the surgery, contained Pref-1 protein. In contrast, IZA (the ZF/ZR-specific antigen) was present in most cells at this stage. Engeland et al. reported that the cells expressing P450(aldo) protein also disappeared in the early stage of adrenal regeneration and that differentiation occurred from a ZG cell to an intermediate or ZF cell phenotype (55). It seems that most of the aldosteronogenic cells were converted immediately after the surgery to those negative for steroidogenic activity, like the ZI cells. However, the most signals for proliferating cell markers were not detected in these ZI-like cells but were detected in ZF/ZR-like cells expressing P450(11ß) (18). These observations suggest that the proliferating cells in the regenerating adrenal cortex may be differentiation-initiated ZF/ZR-like cells, rather than the stem cells. Our studies indicate that the decrease of ZG-like cells, demonstrated by the repression of Pref-1 protein expression, and the appearance of ZF/ZR-like cells occur at the early phase of adrenal regeneration. It should be mentioned that several cells in the regenerating adrenal gland on day 3 contained both Pref-1 protein and IZA (data not shown), suggesting that they seemed to be in a stage intermediate between the ZG and ZF cells, as shown in an embryonic stage 17.5 d.p.c. Taken together, these findings indicate that the expression of Pref-1 gene is down-regulated at the early phase of adrenocortical regeneration. Similarly, Pref-1 gene has been reported to be down-regulated during the adipocyte differentiation (32).
These findings suggest that the Pref-1-expressing cells are at a rather primitive stage in the adrenocortical development. Further studies are required, to clarify the mechanisms of Pref-1 expression and adrenocortical differentiation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 These authors contributed equally to this paper.
Received November 12, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
