| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Biology, Boston University, Boston, Massachusetts 02215
Address all correspondence and requests for reprints to: Gloria V. Callard, Department of Biology, Boston University, 5 Cummington Street, Boston, Massachusetts 02215. E-mail: gvc{at}bio.bu.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
14:1), the ratios are reversed in retina (B/A, 1:25).
Further studies are required to resolve the evolutionary and functional
implications of multiple CYPl9 genes and P450arom isozymes in goldfish,
their differential expression in brain and ovary, and whether
observations can be generalized to other vertebrates. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The ability of the gonads and brain to aromatize androgen is an evolutionarily conserved characteristic of vertebrates (3, 4). However, teleost fish, including the goldfish Carassius auratus, are unique among vertebrates in having exceptionally high levels of brain aromatase activity: e.g. 100- to 1000-fold higher than mammalian brain but similar to mammalian ovary (5, 6). Enzyme activity is correspondingly high in teleost pituitary and retina, but ovarian aromatase is less than one-tenth that of brain (6, 7). To address the molecular mechanism of high constitutive neural aromatase expression, a full-length 3-kb P450arom cDNA was isolated from a goldfish brain library and used to show that high accumulated levels of P450arom messenger RNA (mRNA) in brain correspond to high enzyme levels (8). Paradoxically, the brain-derived cDNA failed to hybridize with ovarian mRNA under any condition. This observation led us to postulate the existence of different brain and ovarian mRNA variants. Although evidence from chickens (9) and medaka fish (10) supports studies in humans indicating that the CYPl9 gene exists as a single copy in the haploid genome (11, 12), multiple CYPl9 loci encoding different P450arom isoforms in placenta, ovary, and blastocysts were identified in pigs (13, 14, 15).
Here we report the isolation of a second P450arom cDNA from goldfish ovaries, and present evidence for multiple CYPl9 gene loci encoding the brain- and ovarian-derived P450arom forms. We show that both proteins are able to catalyze estrogen production from radiolabeled androgen, and are differentially expressed in neural and gonadal tissues. Preliminary findings have been reported elsewhere (16).
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Primers 1–11 were used with
the ovarian P450arom cDNA, whereas 12–16 are complementary to brain
P45Oarom cDNA:
|
Primer 2: nucleotides 1395–1416: 5' CACCATNGCDATRWRYTTNCC 3'.
Primer 3: nucleotides 1316–1337: 5' TGGAAGTTGTCTAGACTGAAC 3'.
Primer 4: nucleotides 1303–1324: 5' GACTGAACTCATTCGGCTTGG 3'.
Primer 5: nucleotides 338–359: 5' RGTBTGGATCWVYGGAGARGA 3'.
Primer 6: nucleotides 398–423: 5' TGAGGTGTACAGAGATTTCCTCAAC 3'.
Primer 7: nucleotides 1301–1329: 5' CCCCAAGCCGAATGAGTTCAGTCTAGAC 3'.
Primer 8: nucleotides 17–43: 5' TCTTATGGCACGTGAACTTCTCCAGC 3'.
Primer 9: nucleotides 1848–1874: 5' CAGTAGAAGACACCAGTAATTCATAG 3'.
Primer 10: nucleotides 1220–1241: 5' TGATGTCATCGAAGGCTACAA 3'.
Primer 11: nucleotides 1274–1299: 5' GAATTCCGATCTGTGCATCCGACCC 3'.
Primer 12: nucleotides 1241–1262: 5' ACGACGAGATCGATGGCTACC 3'.
Primer 13: nucleotides 1362–1383: 5' AAGTAACGACTGGGAACGGTG 3'.
Primer 14: nucleotides 1338–1359: 5' TCGAAGTTCTCCAAGTTGAAT 3'.
Primer 15: nucleotides 1206–1227: 5' GGTTGTGGACTTCATCATGAG 3'.
Primer 16: nucleotides 1338–1359: 5' TCGAAGTTCTCCAAGTTGAAT 3'.
RT-PCR cloning of goldfish aromatase cDNA from ovaries
cDNA was synthesized from 5 µg total RNA with oligo(dT)
primers using SuperScript II reverse transcriptase (Gibco-BRL,
Gaithersburg, MD) according to the manufacturer’s instructions. An
aliquot (15%) of the first-strand reaction was amplified with
degenerate primers 1 and 2. PCR was performed in 50-µl final vol
containing 5 µl 10 x reaction buffer, 2 mM
MgCl2, 200 µM deoxynucleotide triphosphate, 2
µM of each primer, and 2.5 U Taq DNA
Polymerase (Promega, Madison, WI). The following PCR conditions were
used for the first 5 cycles: 94 C for 1 min, 37 C for 2 min, and 72 C
for 3 min; during the remaining 30 cycles, the annealing temperature
was increased from 37 C to 50 C. A 240-bp fragment was purified from
2% agarose with GLASSMILK (GENECLEAN III kit; BIO 101, Vista, CA), T-A
cloned into pCR II vector (Invitrogen, Carlsbad, CA) and sequenced.
Subsequently, the 1-kb PCR fragment was amplified using the RT-reaction
primed with sequence-specific primer 3. The PCR was conducted as
indicated above but with 0.2 µM of primer 4, 2
µM of degenerate primer 5, and Taq-mix
consisting of 2.5 U Taq DNA polymerase and 0.25 U Pfu DNA
Polymerase (Stratagene, La Jolla, CA). The annealing temperature during
the first 5 cycles was 43 C; otherwise PCR conditions were as described
above. After purification, the 1-kb PCR product was T-A cloned into pCR
II vector and sequenced. The remaining cDNA was obtained with the
Marathon Kit (Clonetech, Palo Alto, CA) according to the
manufacturer’s protocol. Primers 6 and 7 were used for 5'-and 3'-rapid
amplification of cDNA ends (RACE), respectively. The amplification of
double-stranded cDNA was done with Advantage KlenTaq polymerase mix
containing a proofreading polymerase (Clonetech). 5'-RACE was carried
out for 30 cycles under the following conditions: 94 C for 30 sec, 60 C
for 45 sec, and 72 C for 2 min. For 3'-RACE, a touch-down program was
used (5 cycles: 94 C for 30 sec and 72 C for 2 min; 5 cycles: 94 C for
30 sec and 70 C for 2 min; and 25 cycles: 94 C for 20 sec and 68 C for
2 min). The full-length cDNA was generated by PCR using Advantage
KlenTaq polymerase mix and primers 8 and 9, derived from the 5' and 3'
ends of the cDNA. A PCR product of approximately 2 kb was silica
purified (Geno-Bind; Clonetech) from a 1% agarose gel and T-A cloned
into pGEM-T vector (Promega). The nucleotide sequences of three
independent cDNAs derived from four separate amplification reactions,
pooled together, were determined.
Sequencing and computer analysis
Sequencing of ovarian aromatase cDNA was done on double-stranded
DNA with Sequenase (U.S. Biochemical Corp., Cleveland, OH) according to
the manufacturer’s recommendations. Complete sequence was obtained
using Sp6 and T7 primers (Promega) complementary to the vector and
sequential internal primers. The nucleotide and amino acid sequences
were analyzed using the WI Package Version 9.0, Genetics Computer Group
(GCG), Madison, WI. Phylogenetic analysis was done by multiple
alignments of deduced amino acid sequences using the Neighbor-Joining
method within the Clustal W program (17).
DNA isolation and Southern analysis
Genomic DNA was isolated from brains of two individual goldfish
as previously described (8). Goldfish were purchased from Grassyfork
Fisheries (Martinsville, IN). Approximately 10 µg DNA of each fish
was digested with EcoRI, HindIII,
PstI, or XbaI (New England Biolabs, Boston, MA)
restriction enzymes, electophoresed on a 0.8% agarose gel, and
transferred to a Magnacharge nylon membrane (MSI, Westborough, MA). The
filter was hybridized at 42 C with an ovary-derived P450arom cDNA probe
(nucleotides: 1173–1322). The probe was generated by cutting the
240-bp PCR product (see above) with XbaI enzyme, and labeled
by random priming to a specific activity 1 x 109
cpm/µg (18). The most stringent wash was performed at 60 C in
0.1 x SSC/0.1% SDS. After autoradiography, the membrane was
reprobed with a brain-derived cDNA fragment, generated by
EcoRI restriction enzyme digestion of the RT-PCR product
amplified with primers 1 and 2. The amplified product was identical to
the corresponding sequence of the reported goldfish P450arom cDNA
(nucleotides: 1195–1336) (8).
RNA isolation and Northern analysis
Different tissues of 50 fish (mixed males and females) were
pooled by type and total RNA was isolated using Tri Reagent (Molecular
Research Center, Cincinnati, OH). Poly(A) RNA was purified from total
forebrain and ovarian RNA with PolyATract mRNA isolation system
(Promega), then 5 µg/lane poly(A) RNA was fractionated on a 1%
formaldehyde agarose gel and transferred to Magnacharge nylon membrane
(MSI, Westborough, MA). The RNA blot was prehybridized in 50%
formamide, 5 x SSPE, 2 x Denhardt’s solution, 100 µg
yeast transfer RNA, and 0.1% SDS at 42 C. The ovarian aromatase
complementary RNA probe was derived from the 240-bp PCR product cloned
into pCR II vector (Invitrogen) (see above). The plasmid was linearized
with HindIII and labeled with [32P]uridine
triphosphate using the MAXIscript In Vitro Transcription kit
(Ambion, Austin, TX) to a specific activity of approximately 1 x
109 cpm/µg. Hybridization was performed in the
prehybridization buffer at 42 C with 4 x 106 cpm/ml
riboprobe. The final stringent wash was at 68 C in 0.1 x
SSPE/0.01% SDS. A 1-kb goldfish actin cDNA (derived by RT-PCR with
degenerate primers, our unpublished data) was used as probe to
standardize for RNA loading.
RT-PCR
Total RNA isolated from forebrain, mid/hindbrain, pituitary,
ovary, and retina was DNase I-treated (Promega) according to the
standard procedure (18). Each RNA (5 µg) was reverse transcribed as
indicated above. PCR was performed with 10% of the RT reaction,
diluted as follows: forebrain and ovary, 1:1, 1:3, 1:9, 1:27, 1:81, and
1:243, and retina, 1:1, 1:2, 1:4, 1:8, 1:16, and 1:32. The PCR mix
consisted of the same components as for amplification with primers 1
and 2, except two sets of primer pairs (10 and 3 plus 12 and 13) were
added at a concentration of 0.2 µM. The reaction mixture
was denatured for 3 min at 94 C. Thirty cycles of amplification were
conducted with 30 sec denaturation at 94 C, 30 sec annealing at 60 C,
and 1 min extension at 72 C. The reaction was incubated for an
additional 10 min at 72 C to allow completion of extension. One half of
each PCR reaction was analyzed on an 8% polyacrylamide gel. After
staining with ethidium bromide (EtBr), gel image acquisition was
performed with a Videk CCD camera (Kodak, Canandaigua, NY) and Bio
Image’s Low Light Scanning software (Bio Image, Ann Arbor, MI). A
600 ± 70-nm band-pass filter was used to reduce background
fluorescence. Authenticity of PCR products was confirmed by Southern
analysis of aliquots of forebrain and ovarian RT-PCR reactions, after
separation on a 3% Nu-sieve GTG/1% agarose gel (FMC Bioproducts,
Rockland, ME). The DNA blot was probed at 55 C with an oligonucleotide
11, labeled with [
-32P]ATP using T4 polynucleotide
kinase (Promega) (18). The membrane was washed for 5 min in 6 x
SSC/0.1% SDS, autoradiographed, and then reprobed with an
oligonucleotide 14 under the same conditions.
PCR of genomic DNA
Genomic DNA (200 ng) from two goldfish (different from
individuals used for Southern analysis) was subjected to PCR with
either brain (15 and 16) or ovarian (10 and 3) primer pairs (see legend
of Fig. 5C). The PCR conditions were as described for RT-PCR, except
that the final volume of PCR reaction was 100 µl. An aliquot (20%)
of the reaction mixture was digested with the restriction enzymes
EcoRI or ClaI to generate distinct size fragments
for each amplified product. Undigested and digested products were
resolved on an 8% polyacrylamide gel, EtBr-stained and photographed
with FOTO/Phoresis UV documentation system (Fotodyne, Hartland,
WI).
|
Transient transfection
COS 7 African green monkey kidney cells were grown in Temin’s
modified Eagle’s medium (TMEM) supplemented with 10% FBS (Sigma
Chemical Co., St. Louis, MO). Transfections were done in 60-mm dishes
(two dishes per construct, 65% confluent cells) with 10 µg/ml
plasmid DNA construct (see above) and 30 µg/ml polybrene (Sigma) by
the dimethylsulfoxide-polybrene method as described by Kawai and
Nishizawa (20). Mock-transfected cells were treated similarly only
without addition of DNA constructs.
Aromatase assays
Aromatase activity was measured by [3H]androgen to
[3H]estrogen conversion. Substrates were
[7-3H]androstenedione (-A; specific activity, 22.5
Ci/mmol) and [1
,2-3H]testosterone (-T; specific
activity, 53.5 Ci/mmol; New England Nuclear, Boston, MA), repurified
just before use by thin-layer chromatography on silica gel plates
(ether/hexane, 3:1 vol/vol) and adjusted to 10 Ci/mmol specific
activity with respective radioinert androgen. Sixty hours after cell
transfection, medium was replaced with fresh medium containing
substrate (50 nM final concentration) and aliquots taken at
2, 6, and 24 h of incubation. Cells were harvested after 24 h
for determination of protein. Controls were mock-transfected cells and
no-cell cultures. [3H]Estrogen products were analyzed by
procedures previously described and validated for goldfish brain (6).
In brief, media were extracted with diethyl ether and steroids present
in the ether residues were separated by thin-layer chromatography in
the system above. Silica from areas corresponding to estrone
(E1) and estradiol (E2) standards was scraped
directly into vials for determination of radioactivity. To verify
authenticity of estrogen products, methanolic eluates of E1
and E2 areas of representative samples were evaporated to
dryness and partitioned twice between 0.05N NaOH and carbon
tetrachloride. Results were expressed as counts per minute product
formed per milligrams cell protein.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
240-bp fragment was
amplified with a degenerate pair of primers 1 and 2 (see Fig. 1
and
Materials and Methods) designed to sequences in the highly
conserved Ozols peptide and heme-binding regions. The isolated cDNA
fragment was 75% homologous to the corresponding region of the
goldfish P450arom derived from the brain cDNA library (8). In the
second step of cloning, a 1-kb RT-PCR fragment was obtained using
one degenerate and one sequence-specific primer (4 and 5). The sequence
of degenerate primer 5 was chosen from the conservative region A
described by Tanaka et al. (10). By means of 5'/3'-RACE and
primers to newly identified sequences (6 and 7), the remaining sequence
information was determined. The full-length cDNA was amplified with a
pair of primers (8 and 9) targeted to the 5'- and 3'-ends. The products
of four independent PCR reactions were combined and cloned into the
pGEM vector. Three P450arom cDNA clones were fully sequenced, and two
of them were identical (shown in Fig. 2).
The third P450arom clone, when compared with the other two, had several
nucleotide substitutions in the ORF: four synonymous (T293C, T380A,
G1004A, and C1514A) and three nonsynonymous (C222T, T381A, and A807G).
The latter resulted in the following amino acid substitutions: P68S,
S121T, and R263G. Four more substitutions were observed in the 3'
untranslated region (UTR): A1687C, A1773C, A1780T, and T1812A. Some
differences in the sequence among clones may be explained by allelic
polymorphism (because starting material was RNA pooled from ovaries of
25 fish), whereas others may be due to PCR errors despite the use of
a proofreading enzyme in the polymerase mix.
|
). The
ORF encoded a protein of the calculated mol wt of 58.7K. Similar to the
reported fish ovarian aromatases (21, 22, 23, 24), the goldfish P450arom had a
second potential initiation site 30 bp downstream from the first, but
neither of the two ATGs had a nucleotide context that perfectly matched
the proposed consensus sequence for initiation of translation (25).
This differs from the brain P450arom, which has a single ATG that
exactly corresponds to the initiation consensus sequence (8; Table 1).
|
) and other P450arom forms. The 3'UTR was
322 bp in length and contained two polyadenylation signals, 9 bp and 18
bp upstream from the poly(A) tail. It was the major feature accounting
for size differences between the cloned ovarian and brain-derived cDNAs
(1.9 vs. 3 kb; Table 1).
Figure 3 shows the deduced amino acid
sequence of the goldfish ovary-derived P450arom aligned with other
P450arom forms (8, 21, 22, 23, 24, 26, 27). The goldfish ovarian P450arom
shared 62/78% overall sequence identity/similarity with the goldfish
brain aromatase; 68–72/82–85% with other fish aromatases; and
53/71% and 52/73% with chicken and human aromatases, respectively.
The percent of identity/similarity was higher (64–97/79–100%) in the
regions of high homology, including the I-helix, an aromatase-specific
conserved region, and the heme-binding region, and was lowest at the
termini of the different P450arom forms. Cross-comparison of the
goldfish ovarian and brain sequences showed amino acid substitutions in
all nine putative coding exons, indicating that the two P450aroms are
more likely to be derived from different genes than from alternative
exon splicing. The ovarian P450arom had an extended N-terminus and a
shortened carboxy terminus relative to the brain form. Where
overlapping, there were a total of 192 amino acid differences, of which
30% were nonconservative substitutions. Many of the differences
were clustered at the N- terminus and the central part of molecule,
including regions of putative exons II, V, and VII, and the N-terminus
of exon IX. Based on mutational analysis and molecular modeling, amino
acids known to be essential for catalytic functions in the human
P450arom [I133 (28), E302, P308, D309, T310 (for review, see Ref.29),
and R435, C437 (30)] were identical in both goldfish forms. Many of
the substitutions between goldfish ovarian and brain P450aroms were
brain-specific in that they are conserved among all other fish ovarian
forms: F72W, S94C, I218V, H262Y, H265Y, Q270K, L353V, R423K, and S452C.
Others distinguish the brain P450arom from all nonneural aromatases,
including chicken and human, and were located in putative functional
domains: I-helix (E314D and N315D) and heme-binding domain (M467T)
(based on numbering of the ovarian sequence). Of these, only the first
was a conservative substitution. Also, a histidine residue at position
128 of the human P450arom, which may be involved in orientation of the
substrate in the active site (29), corresponded to a histidine in the
goldfish ovarian form but phenylalanine in the brain P450arom.
Consistent with the glycosylation site described at the N terminus of
human aromatase (31), a consensus N-glycosylation site (N-X-S/T) was
identified in the amino terminal region of the ovarian form at N30, but
the 5'-most glycosylation site in the brain form was at N42. Neither of
these sequences was present in other fish aromatases. Further analysis
with the Motifs program of the WI Package (GCG) indicated that the
goldfish ovarian P450arom had consensus sequences for phosphorylation
by cAMP-dependent protein kinase A at residue 409 (KRGT) and protein
kinase C (PKC)-dependent sites at 133, 219, and 383 (S/T-X-R/K). The
goldfish brain P450arom lacked the protein kinase A phosphorylation
consensus sequence, but had all three PKC sites in corresponding
positions (117, 198, and 361). The first two of these were conserved in
all P450arom forms compared in Fig. 3, and the third was present in
trout, chicken, and human, indicative of their possible functional
importance. Each of the two goldfish forms had additional uniquely
positioned PKC- and casein kinase-dependent phosphorylation
sequences.
|
). As expected, all
fish ovarian aromatases were clustered together. However, the goldfish
neural P450arom branched independently from the group comprising the
fish aromatases, including the goldfish ovarian counterpart, and the
calculated percent sequence divergence was indicative of a long
evolutionary history from the branch point.
|
). Genomic DNA was
digested with different restriction enzymes, electrophoresed on an
0.8% agarose gel, transferred to a nylon membrane, and hybridized with
the 32P-labeled probes, as described in Materials and
Methods. Each probe gave a unique hybridization pattern with four
different restriction digestions and with no evidence of
cross-hybridization at the stringency conditions used (compare Fig. 5, A and B). The brain-derived probe demonstrated a single band with each
of the four restriction enzymes tested (EcoRI, 1.7 kb;
HindIII, 3.9 kb; PstI, 2.8 kb; and
XbaI, 10 kb); thus, it is encoded by a single gene locus.
By contrast, the ovary-derived probe gave one to three bands. In all
but HindIII digestions, results were identical for two
individuals: EcoRI, 6.8, 2.4, and 1.65 kb (the latter band
was seen on longer exposure); HindIII, 10 and 5.8 kb (the
latter was not seen in the second individual); PstI, 10 and
3.4 kb; and XbaI, 5.6 and 3.8 kb. The third band obtained
with the EcoRI restriction digestion is due to an internal
EcoRI site in the ovarian probe. The identification of at
least two fragments with most of the restriction digestions tested
suggests that more than one CYP19 gene locus of the ovary-type sequence
is present in the goldfish genome. Most importantly, the results of
Southern analysis show the existence of two different gene loci in the
goldfish genome encoding separate brain- and ovary-derived P450arom.
This conclusion is supported by PCR analysis (Fig. 5C), in which
regions of CYP19 genes corresponding to brain- and ovary-derived cDNAs
were amplified from genomic DNA with type-specific primer pairs to give
different-sized products. Amplified DNA products of both types were
further analyzed by restriction digestion to give different-sized
cleavage products for the two sequences. Thus, Southern and genomic
PCR-restriction analyses both show that two P450arom mRNAs are
transcribed from separate and distinct gene loci.
Tissue distribution of two P450arom mRNA variants
Northern blot analysis of goldfish poly(A) RNA was done to
determine the size and number of ovary-derived aromatase mRNAs (Fig. 6). Two bands (3.0 and 1.9 kb) were
observed with ovarian poly(A) RNA, and the 1.9-kb transcript was the
most abundant. The size of the smaller transcript corresponds to the
size of the P450arom cDNA isolated by RT-PCR from ovaries. The presence
of the larger mRNA species may be explained by use of an alternative
polyadenylation signal located downstream of the 3'most end of the cDNA
shown in Fig. 2. Importantly, the 3.0-kb mRNA seen in goldfish ovaries
cannot be ascribed to cross-hybridization with the brain-type mRNA,
because forebrain poly(A) RNA failed to show a hybridization signal
with the ovarian probe, although use of the brain probe in the same
experiment (not shown) confirmed the presence of abundant brain-type
message (8).
|
illustrates representative
EtBr-stained gels of RT-PCR products. Analysis of integral intensity of
RT-PCR products by Bio Image’s Low Light Scanning software
demonstrated that in neural tissues, such as forebrain and
mid/hindbrain (not shown), the brain variant was at least 14 times more
abundant than the ovarian type, whereas in ovaries, only the ovary
variant was detected. In pituitary, the brain variant predominated, but
ratios were not determined (not shown). Interestingly, even though two
types of mRNA were found in retinal tissue, similar to forebrain, their
ratio was reversed (1:25). In Fig. 7B, results were validated by
Southern analysis of forebrain and ovary RT-PCR products hybridized
with internal oligonucleotides. Whereas the brain-type variant was much
more abundant than the ovary type in forebrain, no brain-type P450arom
cDNA variant was detected in ovaries.
|
shows that COS cells
transfected with both pcDNA3.1-BRAROM and -OVAROM constructs
transformed androgen to estrogen in a linear fashion over 24 h of
incubation. Both A and T were aromatized; however, as previously
reported, E1 was the major estrogen produced from both
substrates due to high constitutive expression of l7ß-hydroxysteroid
dehydrogenase in COS 7 cells (22). The radioactivity eluted from
E1 areas partitioned mainly into the phenolic fraction,
indicating authentic estrogen. When expressed on the basis of total
cellular protein (due to the lack of isoform-specific goldfish
antibodies), estrogen yields were greater from A than from T and from
cells expressing the brain construct compared with the ovarian
construct. Mock-transfected cells and no-cell cultures produced no
detectable E1. Results were similar in a second experiment
when cells were transfected with the PSGFLAG5-BRAROM and -OVAROM
constructs, except that differences in estrogen yields from the brain-
and ovarian-derived isozymes were negligible (from T) or reversed (from
A) (not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Alignment of the newly isolated P450aromA sequence with previously
reported fish aromatases shows that it is more closely related to
ovarian aromatases of other fish species than to the goldfish brain
counterpart. Nonetheless, a uniformly high percentage of identity was
observed among all fish aromatases when functionally important regions
(I-helix, heme-binding) and specific amino acid residues were compared,
and the overall percent of amino acid identity when compared with
chicken and human aromatases was similar to other fish (21, 22, 23, 24).
Interestingly, all fish ovarian P450aroms reported so far are in a
range of 518–524 amino acids, whereas goldfish brain, chicken, and
mammalian aromatases have 503–510 amino acids. The difference in
length is due mainly to the longer N-termini of ovary-derived fish
sequences (Fig. 3) as previously noted (10). This characteristic
separates the goldfish P450aromB from P450aromA and other teleostean
P450aroms. The phylogenetic tree depicted in Fig. 4 further illustrates
that the goldfish neural aromatase segregates from known ovarian
sequences to form a distinct branch. Compared with the two goldfish
isoforms, which have only 62% overall amino acid identity, the porcine
isoforms (ovarian, blastocyst, and placental types) are much more
homologous (87–93% identity) and virtually identical within certain
exons (for review, see Ref.15). Although a 700-bp P450arom fragment
has been isolated from quail brain (37), and 5' cDNA variants have been
identified in rodent and primate brain (38, 39), to our knowledge no
full-length P450arom cDNA has been characterized in the brain of any
species. Efforts to obtain additional brain P450arom sequences will
help to resolve evolutionary relationships within the CYPl9 family and
test the predictive value of a separate brain-type isoform in
goldfish.
Tetraploidization, or genomic duplication, occurred at several points during early vertebrate evolution (>500 million years ago), and is believed to be the genetic basis of subsequent rapid evolutionary change and the origin of multigene families (40). A more recent genome duplication event (15–20 million years ago) led to tetraploidy in present-day goldfish, which have multiple loci for many single-copy mammalian genes (41). However, tetraploidy per se cannot account for multiple CYPl9 genes in goldfish because corresponding P450aromA and P450aromB sequences of putative exon IX were identified by PCR cloning in the genome of the zebrafish, a closely related diploid species (16). The presence of multiple CYPl9 genes in pigs (13, 14, 15) and a CYPl9 pseudogene in the cow (42) signifies that CYPl9 gene duplication originated in a common ancestor of fish and mammals or, more likely, is the end result of two or more independent duplication events during the course of vertebrate evolution. The idea that CYPl9 gene duplication followed by diversification and/or convergence was a common occurrence at many points in evolution is further indicated by the presence of alternatively spliced 5'-exons and promoters in the CYPl9 genes of several mammals (1, 38, 39).
The two goldfish CYP19 genes are nonequivalent in their tissue-specific expression, indicating distinct promoters and regulatory mechanisms. RT-PCR analysis shows that expression of the brain-derived P450aromB is restricted to brain and retina, which have high levels of mRNA and enzyme activity, whereas P450aromA is the only form expressed in ovary, which has relatively low aromatase activity (6). Based on our initial results from ribonuclease protection assays (unpublished studies), transcript levels in ovary are low, reflecting low enzyme levels. Interestingly, the CYPl9A and CYPl9B genes have overlapping expression in neural tissues, but the B/A ratios differ regionally: brain, 14:1 and retina, 1:25. Tissue-specific and coexpression of different P450arom isoforms has been seen in pigs (43). An important question is whether the two isoforms expressed in goldfish brain are colocalized in the same cell populations or have unique distributions in different subsets of cells. Although human placental P450arom antibody immunoreacts with a single 56-kDa protein in goldfish brain membranes, and recognizes specific P450arom-labeled cells in sections of goldfish brain and retina (7, 44), further studies using variant-specific antibodies are required to determine whether earlier work may have underestimated the amount of immunolabeling per cell or failed to recognize the full population of P450arom-expressing cells. Recently, it has been shown that two P450arom mRNA variants are present in the rat brain, but only one is a full-length transcript that correlates neuroanatomically with enzyme levels (45, 46). Multiple brain forms, if present in mammals and birds, might also account for reported discrepancies between immunoreactivity and enzyme activity (for review, see Ref.47).
In addition to sequence analysis, formal proof that each of the two CYPl9 genes encodes an authentic P450arom isozyme is based on the ability of the recombinant proteins to catalyze the transformation of androgen to estrogen. Due to high constitutive expression of l7ß-hydroxysteroid dehydrogenase activity in COS cells, and resulting conversion of added T to A (22), functional differences between the two isozymes with respect to substrate preference or turnover cannot be tested in this cell line, nor can we compare product yields without antibodies to quantify aromatase-specific protein. Nonetheless, it is interesting that the goldfish P450aromB isoform differs from the ovarian P450aromA by having phenylalanine instead of histidine at a position corresponding to H128 of the human P450arom. Mutations at this position (H to A or Q) resulted in an approximately 30% increase in the Michaelis-Menten constant (Km) with A, but not T, as substrate, but maximun velocity (Vmax) was reduced to only 20% of wild-type with both substrates (for review, see Ref.29). Based on the three-dimensional model for aromatase, it has been suggested that this residue may affect the reaction indirectly by stabilizing the substrate in the active site (29). To explain a surplus of l9-hydroxylating activity relative to aromatase in rat brain membranes, and differences between brain and placental aromatases in their sensitivity to selective P450 inhibitors, Fishman and associates (48) speculated that aromatases of neural and nonneural tissues may differ. Further work is required to determine whether the goldfish isozymes differ in their basic catalytic properties, or whether sequence differences are indicative of regulation at pre- or posttranslational levels.
We conclude that CYPl9A and CYPl9B have had a long evolutionary history as separate functional genes, and that sequence divergence affected both transcribed and regulatory regions. Analysis of the gene promoter regions is in progress (49). Taken together, these observations and the differential expression of P450aromA and P450aromB in neural and gonadal tissues, signifies that retention of multiple gene loci and isozymes in the teleost lineage has had some adaptive advantage. Whether this is a cause or consequence of high neural estrogen biosynthesis remains unknown.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Recipient of an Endocrine Society Summer Fellowship.
Received October 7, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-reductase activities correlate significantly with
seasonal reproductive cycles in goldfish (Carassius
auratus). Endocrinology 122:1349–1356 homolog can
mobilize the human thyroid hormone receptor to transactivate a human
promoter. Biochemistry 36:9221–9231[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  D. Vizziano-Cantonnet, D. Baron, S. Mahe, C. Cauty, A. Fostier, and Y. Guiguen Estrogen treatment up-regulates female genes but does not suppress all early testicular markers during rainbow trout male-to-female gonadal transdifferentiation J. Mol. Endocrinol., November 1, 2008; 41(5): 277 - 288. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Zhang, W. Zhang, H. Yang, W. Zhou, C. Hu, and L. Zhang Two cytochrome P450 aromatase genes in the hermaphrodite ricefield eel Monopterus albus: mRNA expression during ovarian development and sex change J. Endocrinol., November 1, 2008; 199(2): 317 - 331. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Liu, F. Wu, B. Jiao, X. Zhang, C. Hu, B. Huang, L. Zhou, X. Huang, Z. Wang, Y. Zhang, et al. Molecular cloning of doublesex and mab-3-related transcription factor 1, forkhead transcription factor gene 2, and two types of cytochrome P450 aromatase in Southern catfish and their possible roles in sex differentiation J. Endocrinol., July 1, 2007; 194(1): 223 - 241. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F.-A. Weltzien, C. Pasqualini, M.-E. Sebert, B. Vidal, N. Le Belle, O. Kah, P. Vernier, and S. Dufour Androgen-Dependent Stimulation of Brain Dopaminergic Systems in the Female European Eel (Anguilla anguilla) Endocrinology, June 1, 2006; 147(6): 2964 - 2973. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T.-T. Wong, S. Ijiri, and Y. Zohar Molecular Biology of Ovarian Aromatase in Sex Reversal: Complementary DNA and 5'-Flanking Region Isolation and Differential Expression of Ovarian Aromatase in the Gilthead Seabream (Sparus aurata) Biol Reprod, May 1, 2006; 74(5): 857 - 864. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y Kazeto and J M Trant Molecular biology of channel catfish brain cytochrome P450 aromatase (CYP19A2): cloning, preovulatory induction of gene expression, hormonal gene regulation and analysis of promoter region J. Mol. Endocrinol., December 1, 2005; 35(3): 571 - 583. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P Black, J. Balthazart, M. Baillien, and M. S Grober Socially induced and rapid increases in aggression are inversely related to brain aromatase activity in a sex-changing fish, Lythrypnus dalli Proc R Soc B, November 22, 2005; 272(1579): 2435 - 2440. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Moncaut, G. Somoza, D. M Power, and A. V M Canario Five gonadotrophin-releasing hormone receptors in a teleost fish: isolation, tissue distribution and phylogenetic relationships J. Mol. Endocrinol., June 1, 2005; 34(3): 767 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Kuntz, A. Chesnel, S. Flament, and D. Chardard Cerebral and gonadal aromatase expressions are differently affected during sex differentiation of Pleurodeles waltl J. Mol. Endocrinol., December 1, 2004; 33(3): 717 - 727. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kagawa, K. Gen, K. Okuzawa, and H. Tanaka Effects of Luteinizing Hormone and Follicle-Stimulating Hormone and Insulin-Like Growth Factor-I on Aromatase Activity and P450 Aromatase Gene Expression in the Ovarian Follicles of Red Seabream, Pagrus major Biol Reprod, May 1, 2003; 68(5): 1562 - 1568. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. T. Ankley, M. D. Kahl, K. M. Jensen, M. W. Hornung, J. J. Korte, E. A. Makynen, and R. L. Leino Evaluation of the Aromatase Inhibitor Fadrozole in a Short-Term Reproduction Assay with the Fathead Minnow (Pimephales promelas) Toxicol. Sci., May 1, 2002; 67(1): 121 - 130. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Okubo, S. Nagata, R. Ko, H. Kataoka, Y. Yoshiura, H. Mitani, M. Kondo, K. Naruse, A. Shima, and K. Aida Identification and Characterization of Two Distinct GnRH Receptor Subtypes in a Teleost, the Medaka Oryzias latipes Endocrinology, November 1, 2001; 142(11): 4729 - 4739. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Gen, K. Okuzawa, N. Kumakura, S. Yamaguchi, and H. Kagawa Correlation Between Messenger RNA Expression of Cytochrome P450 Aromatase and Its Enzyme Activity During Oocyte Development in the Red Seabream (Pagrus major) Biol Reprod, October 1, 2001; 65(4): 1186 - 1194. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kishida and G. V. Callard Distinct Cytochrome P450 Aromatase Isoforms in Zebrafish (Danio rerio) Brain and Ovary Are Differentially Programmed and Estrogen Regulated during Early Development Endocrinology, February 1, 2001; 142(2): 740 - 750. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Okubo, S. C. Mok, and S. Chen Regulation of Aromatase Expression in Human Ovarian Surface Epithelial Cells J. Clin. Endocrinol. Metab., December 1, 2000; 85(12): 4889 - 4899. [Abstract] [Full Text]
|
|
|
|
|
|
R. Sampath Kumar, S. Ijiri, and J. M. Trant Changes in the Expression of Genes Encoding Steroidogenic Enzymes in the Channel Catfish (Ictalurus punctatus) Ovary Throughout a Reproductive Cycle Biol Reprod, December 1, 2000; 63(6): 1676 - 1682. [Abstract] [Full Text]
|
|
|
|
|
|
C. J. Corbin, J. M. Trant, K. W. Walters, and A. J. Conley Changes in Testosterone Metabolism Associated with the Evolution of Placental and Gonadal Isozymes of Porcine Aromatase Cytochrome P450 Endocrinology, November 1, 1999; 140(11): 5202 - 5210. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) and antisense (
) primers are shown and correspond to
oligonucleotides 1–9 listed in Materials and Methods.
Hatched bar indicates coding region.
DNA digested with HindIII) are indicated
between A and B. Genomic DNA (200 ng)
from two individuals was amplified with primers targeting sequences
within presumptive exon IX (10, 11) but specific for brain- (15/16) or
ovary-derived cDNAs (10/3)(C). Size of undigested PCR products
generated with brain- (lanes 1 and 3) and ovary-type primers (lanes 2
and 4) was 153 bp and 116 bp, respectively. Digestion of brain product
with EcoRI generated 131 bp and 22 bp fragments. The
latter fragment was not visible on gel. Fragments of 74 bp and 42 bp in
length were obtained with EcoRI cleavage of product
amplified with ovary-specific primers. ClaI site was
present in brain- but not in ovary-derived sequence, and generated 108
and 45 bp cleavage products.
) or no DNA construct (
) and incubated with [3H]A
(
, 
, 
) or -T (•,