European Journal of Endocrinology (2002) 146 583–589 ISSN 0804-4643
Aromatase and breast cancer: W39R, an inactive protein
C Nativelle-Serpentini, S Lambard, G E Seralini and P Sourdaine
´ ´ ´
Institut de Biochemie et Biologie Appliquee, EA 2608, Laboratoire de Biochimie et de Biologie Moleculaire, Universite de Caen, Esplanade de la Paix,
14032 Caen Cedex, France
(Correspondence should be addressed to G E Seralini; Email: email@example.com)
Background: Aromatase (CYP19) catalyzes the conversion of androgens into estrogens. It is in
particular involved in development, reproduction and breast cancer. One of its polymorphisms,
W39R localized in the N-terminal region of CYP19, signiﬁcantly decreases breast cancer risk
among Japanese women and was chosen for this study. In this work, we studied the structure –
function relationships between W39R polymorphism and CYP19 enzyme activity.
Objective: To examine the kinetic properties of the mutant W39R recombinant protein in transfected
human cells devoid of steroidogenic activity.
Methods: Expression vectors for the wild-type or the mutated R39 aromatase were transiently trans-
fected into E293 human embryonal kidney cells. The conversions of androstenedione to estrone and
of testosterone and nortestosterone to 17b-estradiol were assayed by RIA. Expression of recombinant
cDNAs was analyzed by semi-quantitative RT-PCR and immunoblotting.
Results: W39R recombinant protein was devoid of aromatase activity whatever the substrate used.
This absence of activity was not due to the lack of expression of the recombinant enzyme since the
mRNA and protein were detected.
Conclusion: Our present in vitro study shows that the R39 mutant is unable to synthesize estrogens.
This work provides a novel observation, being consistent with the fact that Japanese women with the
variant allele arg have signiﬁcantly lower risk of developing a breast tumor.
European Journal of Endocrinology 146 583–589
Introduction cancer (8). However, recently, Miyoshi et al. (9) have
identiﬁed two new polymorphisms. The ﬁrst one, also
The human aromatase (CYP19), in concert with the described by Kristensen et al. (10), is a tetranucleotide
ubiquitous ﬂavoprotein NADPH-cytochrome P450 repeat in intron 4 which could be associated with
reductase, is a key enzymatic complex that catalyzes breast cancer susceptibility. The second polymorphism,
the conversion of androgens into estrogens. It is now chosen for our study, is localized in the N-terminal
well established that estrogens are involved in various region of the CYP19 gene among Japanese women.
physiological functions but that they also promote the Miyoshi et al. (9) observed that this single homozygous
growth of breast cancer cells (1, 2). This enzymatic or heterozygous amino acid substitution from trypto-
complex belongs to the class of mammalian endo- phan to arginine at codon 39 (W39R), signiﬁcantly
plasmic reticulum cytochrome P450, anchored with decreased the breast cancer risk. The same authors
the N-terminal membrane domain extending from speculated that this polymorphism may reduce the
A20 to W39 (3, 4) (Throughout, the conventional activity of CYP19 enzyme, resulting in the lower pro-
single letter abbreviations for amino acids are used: duction of estrogens. In this work, we tested this
A, alanine; C, cysteine; E, glutamic acid; Q, glutamine; hypothesis, i.e. the relationship between W39R poly-
R, arginine; W, tryptophan.) Moreover, Amarneh and morphism and CYP19 enzyme activity. Therefore in
Simpson (5) have reported that the deletion of the the present work, the W39R mutant was generated
ﬁrst 53 residues of the human aromatase protein in order to determine its potential impact on the
abolishes its activity. In the last decade, several point structure – function relationships of aromatase. A
mutations of CYP19 have been associated with a loss mammalian cell expression plasmid, pCMV-HA, was
of the aromatase function (for review see (6)). On the mutated or not, and transiently expressed in E293
other hand, one polymorphism, R264C, in the coding human embryonic kidney cells in order to determine
area of the gene was observed (7) but without affecting the kinetic properties of the recombinant proteins.
activity and without being correlated with breast E293 cells were used because they are an appropriate
q 2002 Society of the European Journal of Endocrinology Online version via http://www.eje.org
584 C Nativelle-Serpentini and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2002) 146
cell line to study recombinant steroidogenic enzyme E293 cell culture and transient transfection
(11) since they are naturally devoid of aromatase, assay
5a-reductase and 17b-dehydrogenase enzyme activi-
The stable reductase-transfected human embryonal
ties (12). In order to investigate if this polymorphism
kidney cell line E293 (a gift from Dr Van Luu-The,
at codon 39 affects the expression of aromatase or its
CHUL, Quebec, Canada) was maintained at 37 8C (5%
cellular localization, we have performed a semi-quanti-
CO2, 95% air) in EMEM phenol red-free medium
tative RT-PCR analysis of mRNA from transfected E293
containing 2 mmol/l glutamine, 100 U/ml antibiotics
cells and a CYP19 protein analysis in microsomal frac-
(penicillin, streptomycin and Fungizone), 10% heat-
tions by Western blot. According to the aromatase
inactivated fetal calf serum (supreme serum), and 1%
sequence, one possible structural modiﬁcation induced
non-essential amino acids (BioWhittaker, Gagny,
by R39 could result from its ionic interaction with the
France). Forty-eight hours prior to transfection, cells
i+3 residue, E42, leading to a modiﬁcation of the hinge
were grown up to 50% conﬂuence on 24-well cell
region (13). In order to check this hypothesis, we have
culture plates, washed with medium and supplemented
performed the double mutation W39R-E42A. In such a
with 500 ml EMEM-free serum. Cell transfection was as
protein, alanine will be unable to form an electrostatic
reported by Boussif et al. (15). Brieﬂy, 2 mg plasmid and
interaction with R39 and, thus, could restore the
54 nmol polyethylenimine (PEI) aqueous solution (pH
7) were diluted separately with 50 ml NaCl
In this study, we have shown that the W39R recom-
(150 mmol/l), incubated 10 min at room temperature
binant protein is devoid of aromatase activity whatever
and mixed together. After a further 10 min incubation,
the substrate used and that the double mutation
PEI – DNA mix was added to each well. Then, after 3-4 h
W39R-E42A is unable to restore the activity.
incubation at 37 8C, cells were supplemented with
500 ml EMEM-10% serum. Aromatase activity ‘in cell’
was evaluated 48 h post-transfection. Transfection efﬁ-
ciency was evaluated by ELISA of the wild-type and
Materials and methods several mutated proteins without showing differences
in recombinant aromatase expressions (16).
pCMV-human aromatase cDNA construction
Aromatase activity ‘in cell’
A 2920 bp fragment of human aromatase cDNA
EcoR1-EcoR1 was subcloned into the pCMV EcoRI The conversion of androstenedione to estrone and of
site (14). Orientation was then checked by sequencing. testosterone and nortestosterone to 17b-estradiol was
The pCMV-HA was puriﬁed from the transformed measured in E293 cell supernatants by RIA according
JM109 bacterial strain by using the Qiagen Plasmid to Auvray et al. (14). Brieﬂy, transfected cells were
Mega kit (Qiagen, Courtaboeuf, France). The length, washed with EMEM and incubated with increasing
the concentration and the purity of the plasmid-cDNA concentrations of androstenedione (0 –800 nmol/l) or
construction were veriﬁed by 1% agarose electrophor- 200 nmol/l testosterone or 20 nmol/l nortestosterone
esis and ethidium bromide staining. for 45 min at 37 8C under a 5% CO2 atmosphere. After
incubation, cells were placed for 5 min on ice and
medium was extracted by 10 volumes diethyl ether.
Anti-estrone and anti-estradiol rabbit antibodies (Ref
PF1247 and Ref 1287, P.A.R.I.S., Compiegne, France)
Site-directed mutagenesis were prepared according to the manufacturer’s instruc-
The introduction of speciﬁc base changes in the aroma- tion. The extraction efﬁciency was 80^5% and the
tase cDNA was accomplished by using the Quickchange sensitivity of this RIA was 10 pg/ml. Intra- and inter-
site directed mutagenesis kit from Stratagene (France). assay coefﬁcients of variations were approximately 3
Oligonucleotide primers used for the mutagenesis and 7% respectively. Results are the mean of at least
experiments were synthesized at Eurobio (Les Ulis, triplicate experiments^S.D.
France). Sequences of oligonucleotides were 50 -CCT
TTT TCT CTT GGT GAG GAA TTA TG-30 and 50 -CAT
RNA extraction and semi-quantitative RT-PCR
AAT TCC TCA CCA AGA GAA AAA GG-30 for W39R
and 50 -CTT GGT GAG GAA TTA TGC GGG CAC ATC Total cellular RNA was isolated from E293 cells, alone
C-30 and 50 -GGA TGT GCC CGC ATA ATT CCT CAC or transfected with pCMV-HA (used as positive control)
CAA G-30 for W39R-E42A. All mutated cDNAs were or pCMV-R39 and pCMV-R39-A42, using the TRIzol
sequenced by using the ABI Prism BigDye Terminator reagent (Gibco BRL, Cergy Pontoise, France). Total
Cycle Sequencing Ready Reaction from Biosystems RNA (3 mg) was treated for 10 min at 37 8C with 2
(Warrington, UK) to conﬁrm that there were no other units RNAse-free DNAse (Promega, Charbonnieres, `
mutations except those designed. France) and reverse transcribed with 200 IU M-MLV
EUROPEAN JOURNAL OF ENDOCRINOLOGY (2002) 146 Aromatase and breast cancer 585
RT (Promega) and 0.2 mg oligo dT primer (12 – 18 Subcellular fractionation and Western
bases), in RT buffer (Promega) containing 500 mmol/l analysis
dNTPs and 24 IU Rnasin for 1 h at 37 8C, in a total
Transfected cells from conﬂuent T75 tissue culture
volume of 10 ml. PCR ampliﬁcation of 5 ml of reversed
ﬂasks were scraped into 20 mmol/l Tris –HCl (pH
transcribed sample was carried out in a ﬁnal volume
7.5), 20% glycerol, 15 mmol/l b-mercapto-ethanol,
of 50 ml PCR buffer supplemented with 200 mmol/l
50 mmol/l NaCl, 5 mmol/l EDTA, 1 mmol/l dithio-
dNTP, 50 pmol of each primer, 1.5 mmol/l MgCl2 and
threitol containing 4 mg/ml polymethylsulfonyl
1.5 IU Taq DNA polymerase from Promega. Ampliﬁca-
ﬂuoride, and a mixture of antiproteases (benzamidine,
tion was performed on a Stratagene Robocycler
aprotinin, leupeptin, pepstatin A and antipain at
through 30 PCR cycles with the proﬁle: 94 8C (1 min),
1 mg/ml each). Then cell lysates were centrifuged at
60 8C (1 min) and 72 8C (2 min) and an extra cycle of
20 000 g for 15 min at 4 8C and the supernatant was
72 8C (5 min). The sequences of sense and antisense pri-
transferred to a fresh tube and centrifuged at
mers were 50 -AAA GGA AAT CCA GAC TGTT ATT GGT
100 000 g for 1 h at 4 8C to obtain the microsomal
GAG-30 (exon IX, 1116 – 1143) and 50 -GTA TCT TCT
fraction. Each fraction was mixed with an equal
GTA TGC TCT CAA CAC ACT GTG-30 (exon X, 1511 –
volume of twofold concentrated Laemmli sample
1540) respectively. The resulting PCR product of
buffer (17) and denatured by heating in boiling water
424 bp was resolved on a 1.5% agarose gel and visual-
for 2 min. Proteins (100 mg) were resolved on a 10%
ized by ethidium bromide staining.
SDS-PAGE using Tris base (0.25 mol/l), glycine
The internal control (GAPDH) was ampliﬁed for 23
(1.92 mol/l), SDS (1%) running buffer and then electro-
cycles, in the conditions described above. The sequences
transferred onto PVDF membrane (Appligen, Illkirch,
of sense and antisense primers used to detect the PCR
France). Proteins were detected by immunoblotting
identiﬁcation of the GAPDH mRNAs were 50 -TGA
with the P450 aromatase antiserum (1:1000) pro-
ACG GGA AGC TCA CTG GCA TGG CTT T-30
duced in our laboratory (18), followed by an incubation
(731– 758) and 50 -GTG TGG TGG GGG ACT GAG TGT
with horseradish peroxidase-linked anti-rabbit IgG at
GGC AGG GAC-30 (1132 –1162) respectively. The result-
1:3000 (BioRad, Ivry sur Seine, France) and addition
ing PCR product of 431 bp was resolved on a 1.5%
of Sigma Fast 3,30 -diaminobenzidine (Tablet Sets;
agarose gel and visualized by ethidium bromide staining.
Sigma, Saint Quentin Fallavier, France).
Twelve microliter aliquots of the PCR reaction were
size separated on 1.5% agarose equilibrated in
0.04 mol/l Tris –acetate, 0.001 mol/l EDTA. Gels were Results
stained with ethidium bromide (1 mg/ml), photographed
using Polaroid ﬁlm under UV light and analyzed using an The aim of our study was to examine the enzyme
AGFA Snap-Scan 1200p Scanner, Adobe Photoshop kinetic properties of the mutants W39R and
(version 4) software and the NIH image computer pro- W39R-E42A in transfected E293 cells and to determine
gram (http://rsb.info.nih.gov/nih-image). the importance of these residues in protein activity.
Figure 1 Aromatase activity and mutations. (A)
Enzymatic activity of pCMV-HA (upper curve), and
of R39 and R39-A42 (lower two curves)
transfected in E293 cells, incubated for 45 min with
increasing doses of androstenedione
(0– 400 nmol/l). Results are expressed as pg
estrone (E1) formed/min per mg protein and are
the mean of triplicate values^S.E.M. Insert:
apparent Km value determination of the wild-type
enzyme by a Lineweaver– Burk plot (results are
representative of one from three experiments
showing similar proﬁles). (B, C) Enzymatic activity
of pCMV, pCMV-HA, R39 and R39-A42
transfected E293 cells, incubated for 45 min with
200 nmol/l testosterone (B) or nortestosterone (C).
Results are expressed as pg estradiol (E2)
formed/min per mg protein and are the mean of
586 C Nativelle-Serpentini and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2002) 146
Cells were transfected with either the wild-type plasmid for R39-A42 in comparison with mRNA coding for
(pCMV-HA) or with the mutated plasmids (W39R and the wild-type enzyme. The Western blot analysis of
W39R-E42A) or with pCMV plasmid, used as control. the wild-type protein and the recombinant R39 and
Forty-eight hours after transfection, cells were incu- R39-A42 proteins on subcellular fractions, illustrated
bated for 45 min with increasing concentrations of in Fig. 3, reveals that the proteins are seen only in
androstenedione (0-800 nmol/l) or with 200 nmol/l the microsomal fraction and not in the cytosol.
testosterone or nortestosterone. As illustrated in Fig.
1, the mutation of W39R and the double mutation
induced a total activity loss with androstenedione Discussion
(Fig. 1A). With 200 nmol/l testosterone as substrate,
both mutations decreased strongly the aromatase Recent studies of 204 Japanese women have shown
activity (activities below 6 and 3% relative to the that the single amino acid substitution in aromatase
wild-type for R39 and R39-A42 respectively) (Fig. 1B). of tryptophan by arginine at codon 39 was signiﬁcantly
In the same manner, a lower activity with 200 nmol/l associated with a decrease in breast cancer risk (9).
nortestosterone was observed for the R39 and Those authors speculated that this polymorphism
R39-A42 (41 and 27% of the wild-type respectively) may reduce the activity of the CYP19 enzyme, conse-
(Fig. 1C). Moreover, Fig. 1A shows a typical result of quently resulting in a lower production of estrogens.
kinetic analysis of the wild-type aromatase. From We tested this hypothesis.
these results, the apparent Km value of the wild-type In our study, aromatase activity was measured ‘in
aromatase expressed in E293 was 76:6^14:1 nmol=l cell’ and allowed us to determine an apparent Km
for androstenedione. To study the effects of the point value for the wild-type protein for androstenedione
mutation T/A on the CYP19 expression, aromatase (76 nmol/l) in the same range as values previously
mRNA from the E293 transfected cells was analyzed reported by others (19). However, different estimates
by RT-PCR. As illustrated in Fig. 2A, the corresponding of apparent Km and Vm values can be obtained from
424 bp RT-PCR product of aromatase mRNA was microsomal and intact cell methods. According to
absent in untransfected E293 cells and present in Zhou et al. (20), similar Km values were obtained with
E293 cells transfected with pCMV-HA, pCMV-HA- both approaches but Vm values were lower by using
W39R and pCMV-HA-W39R-E42A, and in granulosa microsomes. The authors suggested instability of the
cells used as a positive control. As illustrated in microsomal preparation. Moreover, Kadohama et al.
Fig. 2C, the semi-quantitative RT-PCR analysis of (21) have suggested that the cytochrome P450 reduc-
CYP19 mRNA levels, by using GAPDH mRNA as tase/aromatase ratio may be involved in differences in
standard (Fig. 2B), showed a 34% decrease in mRNA the two protocols. In such a way, we have used stable
coding for R39 and a 56% decrease in mRNA coding reductase transfected E293 cells. Our results show
Figure 2 (A) RT-PCR ampliﬁcation of aromatase mRNA
followed by each negative control without reverse
transcriptase from granulosa cells (lanes 1 and 2),
untransfected cells (3 and 4), transfected cells with R39 (5
and 6), with R39-A42 (7 and 8) and with pCMV-HA (9 and
10). The size and position of the expected 424 bp transcript
is indicated on the right. (B) RT-PCR ampliﬁcation of
GAPDH mRNA and negative controls. The size and position
of the expected 431 bp transcript is indicated on the right.
(C) Semi-quantiﬁcation by RT-PCR of P450 aromatase
mRNA levels in E293 cells transfected with 2 mg, pCMV-HA
(W39), R39 and R39-A42; the respective P450
aromatase/GAPDH ratios are shown. Each bar is the mean
of three separate values^S.E.M. *P , 0.01, **P , 0.05.
EUROPEAN JOURNAL OF ENDOCRINOLOGY (2002) 146 Aromatase and breast cancer 587
that mutation of residue 39 abolishes or reduces con- and, thus, could restore the aromatase activity. In the
siderably the protein activity, whatever the substrate present study, the double mutant R39-A42 was
tested, and consequently suggest that this position devoid of activity whatever the substrate tested. This
appears to be important for aromatase activity. Slight lack of activity may result from a lower expression
differences according to the substrates is in agreement rate of the double mutant, since corresponding
with our previous results. According to multiple mRNA levels were 56% lower than these of the wild-
sequence analysis, W39 is well conserved in most type. Another hypothesis to explain this lack of activity
P450 aromatase enzymes (human (22), horse (23), may be that the mutation W39R could prevent the
pig (24) and rabbit (25)) except in rat (26), as well as interaction between the cytochrome P450 aromatase
in other P450s involved in steroidogenesis (human and the reductase. For instance, Lee-Robichaud et al.
P450scc (27), steroid 17a-hydroxylase/lyase (34) have described the importance of arginine residues
(CYP17A) (28)) or other P450s such as P450 2E1 which are involved in protein – protein interaction
(29) and P450 2C2 (30). Tryptophan is a hydrophobic between cytochrome b5 and human CYP17 and conse-
amino acid, in contrast to arginine, which is a posi- quently affect the lyase activity of the enzyme required
tively charged residue, and consequently could be for androgen formation. Moreover, Muller-Enoch &
involved in an ionic interaction with E42 according Gruler (35) have recently reported that the N-terminal
to structural models (31). Moreover, several ionic binding domain of cytochrome P450 enzymes deter-
interactions between arginine and glutamic acid have mines the complexation process of the binary P450
been described in P450 cytochromes which could reductase system. We also hypothesized that R39 may
favor secondary structures of the protein. Carani et al. form a salt-bridge with another glutamic acid residue
(32) have identiﬁed a homozygous mutation, R365Q located far away in the sequence, as has been described
on the CYP19 gene, which is associated with a very in CYP19 between R9 and E296 (36), and conse-
low activity of the enzyme. According to the computer quently could prevent the interaction between P450
model of Graham-Lorence et al. (31), R365 may form a aromatase and NADPH reductase. Further studies will
salt-bridge with E362 and may play a major role in the be necessary to explore this hypothesis.
stabilization of the general structure of the aromatase Genetic polymorphism of the gene CYP19 may be
protein. Moreover, glutamic acid is likely to play a involved in mechanisms other than in the protein
role in aromatase structure and function by affecting structure, such as mRNA stability, modulation of tran-
enzyme activity and modulating substrate binding scription or post-translational regulation of expression.
afﬁnities (33). We speculated that this W39R mutation Recently, Kristensen et al. (37) have observed that the
could destabilize the structure of aromatase by forming C/T polymorphism in exon 10 of the CYP19 gene is
a salt-bridge with E42, leading to its inactivity. Thus, to associated with variations in mRNA levels. They
check this possibility, we analyzed the double-mutant reported that, according to the computer model of
pCMV-HA-R39-A42. In such a case, alanine will be Walter et al. (38), the C/T change may inﬂuence
unable to form an electrostatic interaction with R39 mRNA stability. In addition, it was shown that the
introduction of a polar residue in the hydrophobic
anchor domain of cytochromes P450 2E1 and 2B4
changes their cellular location from membrane to cyto-
sol (39). Because a point mutation could modify the
expression of the protein (37), or its cell trafﬁcking
(39), we have further analyzed this mutant at the
RNA and subcellular levels. To study the effects of the
single nucleotide mutation T/A on CYP19 gene
expression, aromatase mRNA from the E293 trans-
fected cells were analyzed by RT-PCR. The results
show that the corresponding product of aromatase
mRNA was detected in E293 cells transfected with
R39. However, the semi-quantitative RT-PCR analy-
sis of mRNA levels have shown that they were sig-
niﬁcantly affected by the mutation. In fact, a 34%
decrease in HA-R39 mRNA levels was observed in com-
Figure 3 Detection of aromatase proteins. E293 cells used for parison with HA-W39 mRNA levels. The new mRNA
enzymatic assays were harvested and centrifuged; 100 mg structures may affect their half-lives but this hypothesis
proteins from microsomal (M) and cytosolic (C) fractions were needs further investigation.
resolved on SDS 10% polyacrylamide gels. The proteins were In order to determine whether the enzyme localiza-
electroblotted onto a PVDF membrane and probed with rabbit
anti-P450 aromatase antiserum. The size and position of the tion was modiﬁed by the W39R mutation, an immuno-
expected 53 kDa aromatase protein is shown on the right and the blot analysis was performed on subcellular fractions. As
molecular mass markers (MW, kDa) are on the left. illustrated in Fig. 3, aromatase was detected only in the
588 C Nativelle-Serpentini and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2002) 146
microsomal fraction and not in the cytosol as was increase the risk of breast cancer. Pharmacogenetics 1998 8
observed for P450 2E1 and 2B4 (39). However, the 43 –48.
11 Dufort I, Soucy P, Labrie F & Luu-The V. Molecular cloning of
mutants are completely inactive whatever the human type 3 3 alpha-hydroxysteroid dehydrogenase that differs
substrate used. This inactivity may partly result from from 20 alpha-hydroxysteroid dehydrogenase by seven amino
a decrease in the protein expression but also from a acids. Biochemical and Biophysical Research Communications 1996
structural modiﬁcation of the enzyme. 228 474 –479.
12 Corbin CJ, Trant JM, Walters KW & Conley AJ. Changes in
In conclusion, our present study provides a novel testosterone metabolism associated with the evolution of
observation, being consistent with the fact that placental and gonadal isozymes of porcine aromatase cytochrome
Japanese women with the variant allele arg have a P450. Endocrinology 1999 140 5202–5220.
signiﬁcantly lower risk than the non-carrier of 13 Yamazaki S, Sato K, Suhara K, Sakaguchi M, Mihara K & Omura
developing a breast tumor, as hypothesized by Miyoshi T. Importance of the proline-rich region following signal-anchor
sequence in the formation of correct conformation of
et al. (9). However, the lack of activity of the mutated microsomal cytochrome P-450s. Journal of Biochemistry 1993
W39R protein raises the question of the consequences 114 652 –657.
of such an homozygous mutation in the normal post- 14 Auvray P, Sourdaine P & Seralini GE. PAAn-1b and PAAn-E: two
natal female physiology. In this context, further in phosphorothioate antisense oligodeoxynucleotides inhibit human
aromatase gene expression. Biochemical and Biophysical Research
vivo studies concerning their aromatase status (activity, Communications 1998 253 1– 9.
expression) and estrogen levels would be interesting. 15 Boussif O, Lezoualc’h F, Zanta MA, Mergny MD, Scherman D,
Demeneix B et al. A versatile vector for gene and oligonucleotide
transfer into cells in culture and in vivo: polyethylenimine.
PNAS 1995 92 7297–7301.
Acknowledgements 16 ´
Auvray P, Nativelle C, Bureau R, Dallemagne P, Seralini GE &
Sourdaine P. Study of substrate speciﬁcity of human aromatase by
This work was supported by Agro-biological industries site directed mutagenesis. European Journal of Biochemistry (In Press).
(AGROBIO) Committee, Fonds FEDER (European 17 Laemmli UK. Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 1970 227
Funds for Development and Research) and the Ligue 680– 685.
Nationale Contre le Cancer (Comite de la Manche et 18 Almadhidi J, Seralini GE, Fresnel J, Silberzahn P & Gaillard JL.
du Calvados). Immunohistochemical localization of cytochrome P450 aroma-
tase in equine gonads. Journal of Histochemistry and Cytochemistry
1995 43 571 –577.
19 Zhou DJ, Korzekwa KR, Poulos T & Chen SA. A site-directed
References mutagenesis study of human placental aromatase. Journal of
Biological Chemistry 1992 267 762 –768.
1 Miller WR, Mullen P, Sourdaine P, Watson C, Dixon JM & Telford J. 20 Zhou D, Cam LL, Laughton CA, Korzekwa KR & Chen S.
Regulation of aromatase activity within the breast. Journal of Mutagenesis study at a postulated hydrophobic region near the
Steroid Biochemistry and Molecular Biology 1997 61 193 –202. active site of aromatase cytochrome P450. Journal of Biological
2 Blankenstein MA, van de Ven J, Maitimu-Smeele I, Donker GH, de Chemistry 1994 269 19501–19508.
Jong PC, Daroszewski J et al. Intratumoral levels of estrogens in 21 Kadohama N, Zhou D, Chen S & Osawa Y. Catalytic efﬁciency of
breast cancer. Journal of Steroid Biochemistry and Molecular Biology expressed aromatase following site-directed mutagenesis.
1999 69 293– 297. Biochimica et Biophysica Acta 1993 1163 195–200.
3 Haugen DA, Armes LG, Yasunobu KT & Coon MJ. Amino- 22 Corbin CJ, Graham-Lorence S, McPhaul M, Mason JI, Mendelson
terminal sequence of phenobarbital-inducible cytochrome CR & Simpson ER. Isolation of a full-length cDNA insert
P-450 from rabbit liver microsomes: similarity to hydrophobic encoding human aromatase system cytochrome P-450 and its
amino-terminal segments of preproteins. Biochemical and expression in nonsteroidogenic cells. PNAS 1988 85
Biophysical Research Communications 1977 77 967–973. 8948–8952.
4 Omura T. Forty years of cytochrome P450. Biochemical and 23 Tomilin A, Auvray P, Moslemi S, Sourdaine P, Drosdowsky MA &
Biophysical Research Communications 1999 266 690–698. ´
Seralini GE. Molecular cloning and characterisation of equine
5 Amarneh B & Simpson ER. Expression of a recombinant cytochrome P450 aromatase. Fourth International Aromatase Con-
derivative of human aromatase P450 in insect cells utilizing ference, Tahoe City, California 1996, Abstract A22.
the baculovirus vector system. Molecular and Cellular Endocrinology 24 Choi I, Simmen RC & Simmen FA. Molecular cloning of
1995 109 R1–R5. cytochrome P450 aromatase complementary deoxyribonucleic
6 Faustini-Fustini M, Rochira V & Carani C. Oestrogen deﬁciency in acid from periimplantation porcine and equine blastocysts
men: where are we today? European Journal of Endocrinology 1999 identiﬁes multiple novel 50 -untranslated exons expressed in
140 111 –129. embryos, endometrium, and placenta. Endocrinology 1996 137
7 Sourdaine P, Parker MG, Telford J & Miller WR. Analysis of the 1457–1467.
aromatase cytochrome P450 gene in human breast. Journal of 25 Delarue B, Mittre H, Feral C, Benhaim A & Leymarie P. Rapid
Molecular Endocrinology 1994 13 331– 337. sequencing of rabbit aromatase cDNA using RACE PCR.
8 Watanabe J, Harada N, Suemasu K, Higashi Y, Gotoh O & ´
Compte-Rendu de l’Academie des Sciences III 1996 319 663–670.
Kawajiri K. Arginine-cysteine polymorphism at codon 264 of 26 Hickey GJ, Krasnow JS, Beattie WG & Richards JS. Aromatase
the human CYP19 gene does not affect aromatase activity. cytochrome P450 in rat ovarian granulosa cells before and
Pharmacogenetics 1997 7 419 –424. after luteinization: adenosine 30 ,50 -monophosphate-dependent
9 Miyoshi Y, Iwao K, Ikeda N, Egawa C & Noguchi S. Breast cancer and independent regulation. Cloning and sequencing of rat
risk associated with polymorphism in CYP19 in Japanese women. aromatase cDNA and 50 genomic DNA. Molecular Endocrinology
International Journal of Cancer 2000 89 325 –328. 1990 4 3– 12.
10 Kristensen VN, Andersen TI, Lindblom A, Erikstein B, Magnus P 27 Chung BC, Matteson KJ, Voutilainen R, Mohandas TK & Miller
& Borresen-Dale AL. A rare CYP19 (aromatase) variant may WL. Human cholesterol side-chain cleavage enzyme, P450scc:
EUROPEAN JOURNAL OF ENDOCRINOLOGY (2002) 146 Aromatase and breast cancer 589
cDNA cloning, assignment of the gene to chromosome 15, and and Arg358 of CYP17 with cytochrome b5. Biochemical Journal
expression in the placenta. PNAS 1986 83 8962–8966. 1998 332 293–296.
28 Matteson KJ, Picado-Leonard J, Chung BC, Mohandas TK & Miller 35 Muller-Enoch D & Gruler H. Complexation of membrane-bound
WL. Assignment of the gene for adrenal P450c17 (steroid 17 enzyme systems. Zeitschrift fur Naturforschung (C) 2000 55
alpha-hydroxylase/17,20 lyase) to human chromosome. Journal 747–752.
of Clinical Endocrinology and Metabolism 1986 63 789–791. 36 Yano J, Koo L, Schuller D, Li H, Ortiz de Montellano P & Poulos T.
29 Song BJ, Gelboin HV, Park SS, Yang CS & Gonzalez FJ. Crystal structure of a thermophilic cytochrome P450 from the
Complementary DNA and protein sequences of ethanol-inducible Archaeon Sulfolobus solfactaricus. Journal of Biological Chemistry
rat and human cytochrome P-450s. Transcriptional and post- 2000 275 31086–31092.
transcriptional regulation of the rat enzyme. Journal of Biological 37 Kristensen VN, Harada N, Yoshimura N, Haraldsen E, Lonning
Chemistry 1986 261 16689–16697. PE, Erikstein B et al. Genetic variants of CYP19 (aromatase) and
30 Okino ST, Quattrochi LC, Pendurthi UR, McBride OW & Tukey breast cancer risk. Oncogene 2000 19 1329–1333.
RH. Characterization of multiple human cytochrome P-450 1 38 Walter JM, Cain RB & Higgins IJ. Cell-free benzo[a]pyrene
cDNAs. The chromosomal localization of the gene and evidence hydroxylase activity in marine zooplankton. Journal of the
for alternate RNA splicing. Journal of Biological Chemistry 1987 Marine Biological Association of the United Kingdom 1979 59
262 16072–16079. 553–563.
31 Graham-Lorence S, Amarneh B, White RE, Peterson JA & 39 Pernecky SJ, Larson JR, Philpot RM & Coon MJ. Expression of
Simpson ER. A three-dimensional model of aromatase truncated forms of liver microsomal P450 cytochromes 2B4
cytochrome P450. Protein Science 1995 4 1065– 1080. and 2E1 in Escherichia coli: inﬂuence of NH2-terminal region on
32 Carani C, Qin K, Simoni M, Faustini-Fustini M, Serpente S, Boyd J localization in cytosol and membranes. PNAS 1993 90
et al. Effect of testosterone and estradiol in a man with aromatase 2651–2655.
deﬁciency. New England Journal of Medicine 1997 337 91–95.
33 Chen S & Zhou D. Functional domains of aromatase cytochrome
P450 inferred from comparative analyses of amino acid
sequences and substantiated by site-directed mutagenesis experi-
ments. Journal of Biological Chemistry 1992 267 22587–22594.
34 Lee-Robichaud P, Akhtar ME & Akhtar M. Control of androgen Received 1 October 2001
biosynthesis in the human through the interaction of Arg347 Accepted 6 December 2001