| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Special Articles |
Medical Research Council Group in Molecular Endocrinology, Centre Hospitalier Université Laval Research Center and Laval University (A.-M.M., M.L.R, M.D., J.S.), Québec, Canada G1V 4G2; Laboratoire de Biochimie Endocrinienne (V.T., F.M., M.G.F., Y.M.), Institut National de la Sante et de la Recherche Medicale U329, Université de Lyon and Hôpital Debrousse, 69322 Lyon Cedex 05, France; Service d’Explorations Fonctionnelles Endocriniennes (S.C., M.C.R.-D.), Hôpital Armand Trousseau, Paris, France; Service d’Endocrinologie Pédiatrique (J.-L.C.), Hôpital Saint-Vincent de Paul, Paris, France; Department of Paediatrics (W.G.S., M.P.), Division of Paediatric Endocrinology, Christian Albrechts-University of Kiel, Germany D-23946
Address correspondence and requests for reprints to: Dr. Jacques Simard, Laboratory of Hereditary Cancers, Centre Hospitalier Université Laval Research Center, 2705 Laurier Boulevard, Québec City, Québec, Canada, G1V 4G2. E-mail: jacques.simard{at}crchul.ulaval.ca
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsMethodsResultsReferences |
|---|
5-4 isomerase
(3ßHSD) deficiency is a form of congenital adrenal
hyperplasia that impairs steroidogenesis in both the adrenals and
gonads resulting from mutations in the HSD3B2 gene and
causing various degrees of salt-wasting in both sexes and incomplete
masculinization of the external genitalia in genetic males. To identify
the molecular lesion(s) in the HSD3B2 gene in the 11
patients from the seven new families suffering from classical 3ßHSD
deficiency, the complete nucleotide sequence of the whole coding region
and exon-intron splicing boundaries of this gene was determined by
direct sequencing. Five of these families were referred to Morel’s
molecular diagnostics laboratory in France, whereas the two other
families were investigated by Peter’s group in Germany. Functional
characterization studies were performed by Simard’s group in Canada.
Following transient expression in 293 cells of each of the mutant
recombinant proteins generated by site-directed mutagenesis, the effect
of the 25 mutations on enzyme activity was assessed by incubating
intact cells in culture with 10 nM
[14C]-DHEA as substrate. The stability of
the mutant proteins has been investigated using a combination of
Northern and Western blot analyses, as well as an in
vitro transcription/translation assay using rabbit reticulocyte
lysates. The present report describes the identification of 8
mutations, in seven new families with individuals suffering from
classical 3ßHSD deficiency, thus increasing the number of known
HSD3B2 mutations involved in this autosomal recessive
disorder to 31 (1 splicing, 1 in-frame deletion, 3 nonsense, 4
frameshift and 22 missense mutations). In addition to the mutations
reported here in these new families, we have also investigated for the
first time the functional significance of previously reported missense
mutations and or sequence variants namely, A82T, A167V, L173R, L205P,
S213G and K216E, P222H, T259M, and T259R, which have not previously
been functionally characterized. Furthermore, their effects have been
compared with those of the 10 previously reported mutant enzymes to
provide a more consistent and comprehensive study. The present results
are in accordance with the prediction that no functional 3ßHSD type 2
isoenzyme is expressed in the adrenals and gonads of the patients
suffering from a severe salt-wasting form of CAH due to classical
3ßHSD deficiency. Whereas the nonsalt-losing form also results from
missense mutation(s) in the HSD3B2 gene, which cause an
incomplete loss in enzyme activity, thus leaving sufficient enzymatic
activity to prevent salt wasting. The functional data described in the
present study concerning the sequence variants A167V, S213G, K216E and
L236S, which were detected with premature pubarche or hyperandrogenic
adolescent girls suspected to be affected from nonclassical 3ßHSD
deficiency, coupled with the previous studies reporting that no
mutations were found in both HSD3B1 and/or
HSD3B2 genes in such patients strongly support the
conclusion that this disorder does not result from a mutant 3ßHSD
isoenzyme. The present study provides biochemical evidence supporting
the involvement of a new molecular mechanism in classical 3ßHSD
deficiency involving protein instability and further illustrates the
complexity of the genotype-phenotype relationships of this disease, in
addition to providing further valuable information concerning the
structure-function relationships of the 3ßHSD superfamily. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsMethodsResultsReferences |
|---|
5-4-isomerase
(3ßHSD) (1, 2, 3, 4, 5) catalyzes the 3ß-hydroxysteroid dehydrogenation and
5 to 4-isomerization
of the 5-steroid precursors pregnenolone
(PREG), 17-hydroxypregnenolone (17OH-PREG), dehydroepiandrosterone
(DHEA), and androst-5-ene-3ß, 17ß-diol
(5-diol) into the respective
4-ketosteroids, namely progesterone (PROG),
17
-hydroxyprogesterone (17OH-PROG),
4-androstenedione
(4-DIONE), and testosterone (T) (6). This
enzymatic activity is, thus, required for the biosynthesis of
glucocorticoids, mineralocorticoids, PROG, androgens, and estrogens. In
the human there are two 3ßHSD isoenzymes, chronologically designated
type 1 and 2, that are 93.5% homologous and are encoded by two genes
on chromosome 1p13.1 (1, 6, 7, 8, 9, 10, 11, 12). The type 1 (HSD3B1) gene is
the almost exclusive 3ßHSD expressed in the placenta and peripheral
tissues including the mammary gland, prostate, and the skin, whereas
the type 2 (HSD3B2) gene is the predominant 3ßHSD
expressed in the human adrenal gland, ovary, and testis (6, 8, 12, 13, 14).
It has been recently reported that as children mature there is a
decrease in the expression of 3ßHSD in the adrenal reticularis that
may contribute to the increased production of DHEA and
DHEAS seen during adrenarche (15). Classical 3ßHSD deficiency results from mutations in the HSD3B2 gene, whereas the HSD3B1 gene is normal in these patients, and is responsible for a severe form of congenital adrenal hyperplasia (CAH). Since the first reports by Bongiovanni (16, 17), many patients of both sexes have been described, and the heterogeneity of the clinical presentation demonstrated. In contrast to the two most frequent causes of CAH, 21-hydroxylase and 11ß-hydroxylase deficiencies, which are adrenal defects, the severe form of 3ßHSD deficiency impairs steroidogenesis in both the adrenals and the gonads, resulting in decreased secretion by these tissues of not only cortisol and aldosterone, but also of PROG, androgens, and estrogens (18–24 and references therein).
Newborns affected by CAH due to classic 3ßHSD deficiency exhibit
varying degrees of salt wasting associated with male
pseudohermaphroditism (18–25 and references therein). The expected
severe inhibition of T biosynthesis by the fetal testis resulting in a
marked decrease in 3ßHSD activity provides an explanation for the
incomplete masculinization of the external genitalia seen in the male
patients studied. Furthermore, males affected with
pseudohermaphroditism and complete or partial 3ßHSD deficiency have
intact Wolffian duct structures, including vas deferens. This is also
the case in 17ßHSD type 3 deficiency as well as 5-reductase type 2
deficiency, which is consistent with the hypothesis that a principal
effect of 3ßHSD deficiency is to reduce the formation of
dihydrotestosterone below the level required for the development of
external genitalia (22, 25). On the other hand, complete or partial
inhibition of 3ßHSD activity in the adrenals and ovaries was not
accompanied by a noticeable alteration in the differentiation of the
external genitalia of female patients, as indicated by the absence of
ambiguity of external genitalia (18–25 and references therein).
The salt-losing form of classic 3ßHSD deficiency is usually diagnosed during the first few months of life due to insufficient biosynthesis of aldosterone and consequent salt loss that may be fatal if not diagnosed and treated early (16, 25, 26, 27, 28, 29, 30). In contrast, the nonsalt-losing form of 3ßHSD deficiency may be diagnosed either at a young age in the presence of indicating factors, such as a family history of death during early infancy (31), perineal hypospadias in male newborns (32, 33), or failure to gain weight (29), or the diagnosis may be made at a later age (30, 34, 35, 36). Due to the fact that sexual differentiation is normal in female newborns affected by nonsalt-losing 3ßHSD deficiency, the proper diagnosis is delayed until adrenarche (35) or puberty (36).
An elevated ratio of 5- to
4-steroids is considered the best biological
parameter for the diagnosis of 3ßHSD deficiency (37, 38). It is well
recognized, however, that levels of 17OH-PROG and
4-DIONE plasma and other
4-steroids are frequently elevated in
3ßHSD-deficient patients (17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 34, 35, 36, 37, 38). Such observations
are consistent with a functional 3ßHSD type 1 that is expressed in
peripheral tissues and is responsible for the extra-adrenal and
extragonadal conversion of 5-hydroxysteroid
precursors into the corresponding
4-3-ketosteroids. This peripheral 3ßHSD
activity could well explain why some patients were misdiagnosed as
having 21-hydroxylase deficiency, in view of elevated 17OH-PROG and
mild virilization of girls at birth (21).
The significant peripheral conversion of
5-hydroxysteroids in 3ßHSD-deficient
patients is in agreement with the finding that a large proportion of
androgens (about 40% in men) and the majority of estrogens formed in
children and in anovulatory or postmenopausal women arise from
extragonadal steroidogenesis (for review see Ref. 39). However,
although very low levels of type 1 3ßHSD messenger RNA (mRNA) can be
detected in normal gonads by sensitive ribonuclease protection assay,
4 steroids can originate from gonadal 3ßHSD
type 1 activity, which possesses a roughly 10-fold higher affinity than
the type 2 isoenzyme and which could be stimulated after an increase in
LH secretion, resulting from low-circulating androgen levels at puberty
(8, 18, 40). Indeed, a male affected with proven severe 3ßHSD
deficiency has fathered children (18, 26).
The nonclassical form of 3ßHSD deficiency, also referred to as
attenuated or late-onset deficiency has been described in older females
with hyperandrogenism beginning at adulthood and children with
premature pubarche presumed to have nonclassical 3ßHSD deficiency
(41, 42, 43, 44). No mutations were found in both HSD3B1 and/or
HSD3B2 genes in these patients (45, 46, 47), and on
reexamination some patients no longer showed an elevated
5/4 ratio (45).
Moreover, Morel’s group also found no mutation in both
HSD3B1 and HSD3B2 genes in 20 girls having a
17OH-PREG peak after ACTH stimulation between 30 and 90 nmol/L (21, 48, 49). It has been concluded that it is difficult, if not impossible, to
provide any kind of accurate statement regarding the clinical features,
pathophysiology, or diagnosis (50).
To date, 24 mutations (1 splicing, 3 nonsense, 3 frameshift, and 17 missense mutations) in the HSD3B2 gene were detected from approximately the same number of families with individuals suffering from classical 3ßHSD deficiency (6, 21, 23, 24, and references therein; Alos, N., Moisan, A.M., Ward, L., et al., submitted for publication). In addition to providing a molecular explanation for the heterogeneous clinical presentations, the functional characterization of these mutant enzymes also generated valuable information concerning the structure-function relationships of the 3ßHSD superfamily (6, 21). In this regard, the goal of the present study was 2-fold. First, we now report the identification of eight mutations in the HSD3B2 gene in 11 patients suffering from classical 3ßHSD deficiency originating from seven new families. Five of these families were referred to Morel’s molecular diagnostics laboratory in France (49), whereas the other two families were investigated by Peter’s group in Germany. Once the mutations had been identified by these two European laboratories, experiments were designed to assess the effect of these mutations on the expression and activity of the 3ßHSD type 2 to gain a better understanding of not only the relation between the molecular defect and the phenotypic manifestation of classical 3ßHSD deficiency, but also on the structure-function relationships of this isoenzyme. Second, in addition to these mutations reported herein, we have also studied the functional significance of recently reported missense mutations and or sequence variants—namely, A82T (51), A167V (52), L173R (22), L205P (53), S213G and K216E (54), P222H (55), and T259R (56)—that have not previously been functionally characterized. To perform a consistent and more comprehensive study into the effects of missense mutations in the HSD3B2 gene, we have also reassessed using the current experimental procedures the activity and expression of previously reported mutant enzymes, including A10E (Alos, N., Moisan, A.M., Ward, L., et al., submitted for publication), G15D (57), N100S (58), L108W (61), G129R (60), E142K (19), P186L (59), A245P (19), Y253N (19), and Y254D (61). The present study, therefore, provides evidence supporting the involvement of a new molecular mechanism in classical 3ßHSD deficiency and further illustrates the complexity of the genotype-phenotype relationship of this disease.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsMethodsResultsReferences |
|---|
Our study examined seven unrelated families, four of which were
affected by the severe salt-losing form of classical 3ßHSD
deficiency, whereas three other families were found to be affected by
the nonsalt-wasting form of classical 3ßHSD deficiency. As indicated
in Table 1
, three
of the families are French (families 9, 14, and 21), two families
originated from Sri-Lanka with one now living in France (family 11),
and the other family lives in Germany (family 12). One family
originated from Algeria (family 7), whereas the last family was from
Egypt (family 16). Families 7, 9, 11, 14, and 21 were referred to Dr.
Y. Morel’s laboratory for molecular diagnosis, whereas families 12 and
16 were referred to Dr. M. Peter’s laboratory.
|
|
Family 7: The two Algerian siblings have been followed by Dr. D. Barama (Algeria). Patient 8 was noted to have perineal hypospadias with micropenis and palpable testes within the scrotum at birth. The karyotype of this individual was established to be 46XY. The index case underwent a salt-wasting crisis and was started on substitutive glucocorticoid and mineralocorticoid therapy. At 16 months of age, while undergoing treatment, the patient had a bad compliance, 17OH-zPREG levels were found to be severely elevated at a value of 461.6 nmol/L, whereas 17OH-PROG levels were 3.9 nmol/L. Patient 9, a 46XX individual, was found to have mild clitoromegaly at 1 month of age. This individual underwent a salt-wasting crisis and was first misdiagnosed to be suffering from the salt-wasting form of 21-hydroxylase deficiency. At 9 months of age, while undergoing substitutive glucocorticoid and mineralocorticoid therapy, 17OH-PREG levels were still found to be elevated 195.18 nmol/L, whereas 17OH-PROG levels were 5.59 nmol/L.
Family 9: The affected individuals from the French family 9 have been followed by Dr. J. L. Chaussain (Paris, France). Patient 11, a 46XY individual is the eldest of two siblings. His younger sister had normal external genitalia at birth (patient 12), while at birth the index case had perineal hypospadias with palpable testes within the bifid scrotum. At 2 days of age, 17OH-PREG levels were found to be severely elevated at a value of 712.66 nmol/L, whereas 17OH-PROG levels were 105.87 nmol/L, indicative of the blockade of adrenal and gonadal 3ßHSD activity. Sodium and potassium levels were found to be normal until day 10 when a salt-losing crisis occurred (Na+: 132 mmol/L; K+: 5.5 mmol/L; renin levels were also increased to 148 pg/mL). The individual responded well to substitutive glucocorticoid and mineralocorticoid therapy, and by day 81 17OH-PREG levels had fallen to 390.91 nmol/L and 17OH-PROG levels were 75.625 nmol/L. The patient is currently continuing with this therapeutic regime.
Family 11: Patient 15 was first diagnosed by Dr. S. Cabrol (Paris,
France) and was the second child from consanguinous parents originating
from Sri-Lanka. The first child, a girl, died at the age of
11/2 months in Sri-Lanka.
Patient 15 was established as having a 46XY karyotype, and at birth
perineal hypospadias with micropenis with palpable testes within the
scrotum was noted. At 12 days of age, he underwent a salt-wasting
crisis with low natremia (127 mmol/L) and high kalaemia (6.7 mmol/L)
and was diagnosed to be suffering from a classical form of 3ßHSD
deficiency. Substitutive therapy was started on day 10. The levels of
17OH-PREG and 17OH-PROG were found at day 16 to be 75.17 nmol/L and
114.95 nmol/L, respectively. At 2 months of age, the patient underwent
an ACTH-stimulation test with the following results (all expressed as
nmol/L): 17OH-PREG (basal = 48.11; 60 min = 204.48);
17OH-PROG (basal = <0.302; 60 min = 5.75); DHEA
(basal = 2.08; 60 min = 6.24); and 4-DIONE
(basal = <0.35; 60 min = 6.64); 17OH-PREG/17OH-PROG ratio
(basal = >159; 60 min = 35.56);
DHEA/4-DIONE ratio (basal = >5.94; 60
min = 0.93).
Family 12: The index case of this family, also originating from Sri-Lanka, was clinically diagnosed by Dr. Peter Beyer (Kinderklinik, Germany), and diagnosis was confirmed by Drs. M. Peter and M.G. Sippell (Kiel, Germany). Patient 16, was born in January 1998 and at birth was noted to have perineal hypospadias. Karyotype analysis determined this individual to be 46XY. Although the consanguinuity of the parents remains unknown, they were both found to be heterozygous for the identified mutation. During the 1st week of life the patient underwent a salt-wasting crisis (Na+: 128 mmol/L; K+ 7.5 mmol/L), and based on first hormonal analysis the exact cause of this salt-wasting form of CAH was difficult to determine. Subsequent genetic analysis confirmed 3ßHSD deficiency, and at 9 months of age the patient was successfully treated with substitutive therapy.
Cases suffering from the nonsalt-wasting form of classical 3ßHSD deficiency
Family 14: For both patients 18 and 19, the clinical data has
previously been reported as patients 4 and 5, respectively, in Gendrel
et al. (32); patient 18 also corresponds to patient 4 in
Chaussain et al. (62). Both patients 18 and 19 have a 46XY
karyotype and were noted to have perineal hypospadias with palpable
testes within the scrotum at birth. Patient 18 had at 6 days of age
diarrhea with low plasma sodium levels (130 mmol/L), and moderate
salt-loss occurred (urinary Na+ 17 mmol/24h,
under a sub-normal sodium diet 10–15 mmol/day). At day 12,
DHEA levels were recorded as 58.94 nmol/L;
4-DIONE: 5.23 nmol/L;
DHEA/4-DIONE ratio as 11.3, no
neonatal peak of testosterone was recorded. The patient failed to
thrive and was started on substituve therapy. Patient 19 was the eldest
of the two siblings. At 8 days of age diarrhea with normal plasma
sodium, but with subclinical salt-loss, occurred (urinary
Na+: 8 mmol/day, under a normal sodium diet 5.5
mmol/day). Normal growth without any treatment occurred until 3 yr, 10
months of age when the youngest sibling (patient 18) was diagnosed to
be suffering from 3ßHSD deficiency. At that point in time,
DHEA levels were found to be 40.22 nmol/L;
4-DIONE 1.57 nmol/L, ratio
DHEA/4-DIONE 25.6; renin >13.20
µg/L.h). In addition, there was no evidence of an increase in T
levels after an hCG test.
Family 16: The two Egyptian 46XY index cases were clinically diagnosed
by Dr. I. Ghaley (Diabetes, Cairo University, Cairo, Egypt), and
diagnosis was confirmed by Drs. M. Peter and M. G. Sippell (Kiel,
Germany). Their parents are first-degree cousins. Patient 22 was born
in December 1993, and at birth scrotal hypospadias with palpable testes
within the labialscrotal folds was noted with a phallus of 2.8 cm. No
salt-wasting was observed within this individual. At 6 months, of age
DHEA levels were found to be 5.88 nmol/L;
4-DIONE 0.7 nmol/L;
DHEA/4-DIONE ratio 8.4, and at 40
months of age this patient underwent an ACTH-stimulation test with the
following results (all expressed as nmol/L): 17OH-PROG (basal =
12.3; 60 min = 31.4); DHEA (basal = 8.65; 60
min = 10.38); and 4-DIONE (basal = 0.7; 60
min = 1.05) DHEA/4-DIONE
ratio (basal = 12.35; 60 min = 9.88). Patient 23 is the
younger brother of patient 22 and was born in November 1995. At birth,
the same scrotal hypospadias with palpable testes within the
labialscrotal folds was noted with a phallus of 2 cm. At 4 months of
age, 17OH-PROG levels were found to be 13.3 nmol/L, DHEA
30.45 nmol/L; 4-DIONE 2.8 nmol/L;
DHEA/4-DIONE ratio 10.86.
Thereafter, at 16 months of age, this patient underwent an
ACTH-stimulation test with the following results (all expressed as
nmol/L): 17OH-PREG (basal = 9.42; 60 min = 117.9); 17OH-PROG
(basal = 1.87; 60 min = 14.76); 17OH-PREG/17OH-PROG ratio
(basal = 5.03; 60 min = 7.98).
Family 21: Patient 32 was diagnosed by Dr. M-C. Raux-Demay (Paris,
France). This 46XY individual, was noted at birth to have perineal
hypospadias with micropenis, with no evidence of palpable testes. At 3
days of age basal steroid levels were as follows: DHEA
93.61 nmol/L; 4-DIONE 2.79 nmol/L;
DHEA/4-DIONE ratio 33.55. At day
13, an ACTH-stimulation test was done with the following results (all
expressed as nmol/L): 17OH-PREG (basal = 10.5; 60 min =
312.73); DHEA (basal = 13.17; 60 min = 135.21);
and 4-DIONE (basal = 3.45; 60 min = 8.38);
DHEA/4-DIONE ratio (basal =
3.81; 60 min = 16.13). There was no salt-wasting within
this individual, and growth was normal.
|
Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsMethodsResultsReferences |
|---|
Selective amplification of the HSD3B2 gene fragments was performed as described previously (18, 19). Three different primer pairs for HSD3B2 were used for amplification of the coding region and the exon-intron splicing junction boundaries. The primers used for PCR amplifications are the same as those used previously (18).
Direct sequencing of PCR products
Direct sequencing of PCR products was performed as described previously (63). Briefly, PCR products were purified with microspin S400-HR columns (Pharmacia Biotech, Uppsala, Sweden) to remove salt, residual primers, and unincorporated deoxynucleotide triphosphates. Approximately 80–100 ng PCR products were directly sequenced using the AmpliTaq FS dye terminators kit (PE Applied Biosystems, Foster City, CA). Each exon was sequenced on both strands. After 25 cycles in 9600 GeneAmp (PE Applied Biosystems; (30 sec at 95 C and 4 min, 30 sec at 60 C), the reaction products were purified on sephadex G50 microspin columns, dried under vacuum, and dissolved in 4 µl of a (5:1) formamide:EDTA mix. Electrophoresis was performed with a 7% acrylamide/bisacrylamide 19/1 sequencing gel during 10 h with a 373A model automatic sequencer, and the data were analyzed using Sequed software (Applied Biosystem).
The above methodology was used for all the patients, except patients 16, 22 and 23. For these two patients the nucleotide sequence of both strands of the PCR products were directly determined by thermo-cycle sequencing using the Thermo Sequenase radiolabelled terminator cycle sequencing kit, following the manufacturers instructions (USB Corporation, Cleveland, Ohio), as described previously (64). The primers used were the same as those used previously (18).
Site-directed mutagenesis
The oligonucleotide sequences for each mutation were designed such that the desired mutation was in the middle of the primer with 15 bases of correct sequence on either side. Site-directed mutagenesis was performed using the QuickChange Site-directed mutagenesis kit form Stratagene Cloning Systems (La Jolla, CA), according to the supplier’s protocol. The correct sequence for each mutation was confirmed by manual sequencing using the dideoxy nucleotide chain termination method (65) using a T7 sequencing kit from Amersham Pharmacia Biotech, Inc. (Picastaway, NJ).
Transcription/translation
Transcription/translation was performed using the TNT Quick coupled transcription/translation system from Promega Corp. (Madison, WI), according to the manufacturer’s instructions. Briefly, 0.5 µg DNA was added to a 16-µl master mix and 1 µCi [35S]-methionine; this was then mixed and incubated at 30 C for 90 min, after which the samples were placed on ice. Translation was then assessed by separation on a 12% SDS-PAGE gel, the gel was dried using a gel dryer, followed by exposure to Hyperfilm-MP x-ray film overnight.
Cell culture and transfection
All media and supplements for cell culture were purchased from Life Technologies, Inc. (Grand Island, NY), except FCS, which was purchased from HyClone Laboratories, Inc. (Logan, UT). Human embryonic kidney 293 cells were obtained from the American Type Culture Collection (Manassas, VA) and were cultured in DMEM/low glucose supplemented with 10% FCS, 1% glutamine, 100 IU/mL penicillin, and 50 µg/mL streptomycin until confluence. The cells were passaged and plated in 6-well plates at a density of 450,000 cells per well. The cells were allowed to settle overnight, after which the medium was changed to DMEM without any supplements just before transfection. Transient transfection was performed using ExGen 500 cationic polymer transfection reagent (MBI Fermentas Inc, Ontario, Canada), according to the supplier’s protocol. Transfection efficiency was monitored by cotransfecting with ß-galactosidase, the activity of which was assessed using the chemiluminescent reporter gene assay system from Tropix Inc. (Bedford, MA). Cells were also transfected with the pCDNA3 vector to act as a negative control. One day after transfection the cells were incubated with substrate to assess enzyme activity, as detailed below.
Northern blot analysis
Total RNA was prepared from the transfected cells using TRI
Reagent and a modified single-step method based on that of Chomczynski
and Sacchi (66). For Northern analysis, 3 µg RNA was loaded per lane
and subjected to electrophoresis in a 1% agarose gel containing 2%
formaldehyde in 1X [N-morpholino]propanesulfonic acid buffer. The gel
was then transferred by capillary action to GeneScreen Plus
hybridization transfer membrane (NEN Life Science Products, Inc., Boston, MA), and RNA was immobilized by
ultraviolet-cross-linking. Hybridization was performed using an
-[32P]dCTP-labeled 448-bp fragment of
3ßHSD type 2 generated by PCR, in 50% formamide-containing buffer at
42 C, according to the membrane suppliers protocol. The membranes were
washed 2x standard saline-citrate (SSC) (1x SSC = 150
mM sodium chloride, 15 mM tri-sodium citrate)
at room temperature for 10 min, and once in 2x SSC/1% SDS at room
temperature for 10 min, followed by 2x SSC/1% SDS at 52 C for 10 min.
Control hybridization was performed by cohybridization with a human
glyceraldehyde-3-phosphate dehydrogenase (GAPDH) complementary
DNA (cDNA) HindIII-XbaI fragment of 548 bp.
Membranes were exposed to Hyperfilm-MP x-ray film at -80 C for 4–16
h.
Western analysis
Western analysis of proteins, prepared from transiently transfected 293 cells, was performed by SDS-PAGE on discontinuous acrylamide gels. Samples were prepared for loading by denaturing at 95 C in 2% SDS, 10% glycerol, 62.5 mmol/L Tris (pH 6.8), and 0.1% dithiothreitol and electrophoresed at 200 volts through 4% stacking and 12% resolving gels using the Mini-Protean II Western apparatus (Bio-Rad Laboratories, Inc., Richmond, CA). Total protein (25 µg) was loaded per lane, and prestained molecular weight markers (Bio-Rad Laboratories, Inc.) were run in parallel lanes. After electrophoresis, proteins were transferred to Hybond-ECL nitrocellulose membrane (Amersham Pharmacia Biotech Inc.), blocked for nonspecific binding with 5% nonfat milk-PBS-0.1% Tween 20, [PBS: 0.05 mol/L (pH7.6)] 16 h at 4 C, then washed briefly in PBS-Tween 20 solution. Membranes were incubated with a polyclonal antibody directed against human placental type 1 3ßHSD at a dilution of 1:2000 in blocking solution (5% nonfat milk-PBS-0.1% Tween 20) for 2 h at room temperature, washed with PBS-0.1% Tween 20, and incubated with donkey antirabbit IgG peroxidase-conjugated secondary antibody (Amersham Pharmacia Biotech Inc.) at a dilution of 1:10,000 in blocking solution for 1 h at room temperature. Membranes were washed and proteins were visualized using the Renaissance plus detection kit (NEN Life Science), followed by exposure of the membranes to x-ray film for 1–10 min.
Assay of 3ßHSD type 2 enzymatic activity
To determine 3ßHSD type 2 enzyme activity in intact 293 cells transiently transfected with the mutant cDNAs, the cells were incubated for 0.5–6 h with 10 nM [4-14C(N)-DHEA (55.2 mCi/mmoL), as described previously (57). Steroids were extracted from the media by the addition of 4 volumes of diethyl ether, and the incubation mixture was chilled in a dry-ice/ethanol bath. Steroids were separated by thin-layer chromatography using a mobile phase of toluene:acetone (4:1) and analyzed using a Phosphorimager imaging system (Molecular Dynamics, Inc., Sunnyvale, CA). All results are expressed as the mean ± SE of at least two separate transfection experiments performed in triplicate.
Establishment of apparent Km and Vmax
The transiently transfected cells were incubated with 10 nM up to 50 µM DHEA, including 10 nM [4-14C(N)]-DHEA and various concentrations of unlabeled DHEA for 15 min (wild-type, A167V, L236S), 30 min (L173R, K216E, S213G), 1 h (A10V, A245P, G294V), 2 h (G129R), or 3 h (A82T) to establish the apparent Km and Vmax for each mutant compared to the wild-type. Steroid measurement was performed as described in the previous section. Under the conditions used, first order kinetics were always maintained. All results are expressed as the mean ± SE of at least three separate experiments performed in triplicate. The apparent Km and Vmax for each mutant protein were calculated using the ENZFITTER software (Biosoft, Cambridge, UK), as described previously (67).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsMethodsResultsReferences |
|---|
To identify the molecular lesion(s) in the HSD3B2 gene in patients from the seven new families suffering from classical 3ßHSD deficiency, the complete nucleotide sequence of the whole coding region and exon-intron splicing boundaries of this gene was determined by direct sequencing, as described in the previous section. The molecular diagnostic testing for patients from families 7, 9, 11, 14, and 21 were performed in Dr. Y. Morel’s laboratory (Lyon, France), whereas those from families 12 and 16 were tested in Dr. M. Peter’s laboratory (Kiel, Germany).
Molecular testing for the Algerian family 7 revealed in the two salt-wasting patients 8 and 9, the presence in exon IV in the HSD3B2 gene of a homozygous C to A transversion converting codon 222 encoding a Pro residue (CCA) to CAA encoding a Gln residue.
Elucidation of the sequence of exon IV in the HSD3B2 gene from two other salt-wasting patients 11 and 12 from the French family 9 indicated the presence of a heterozygous C to T transition converting codon 259 (ACG) encoding Thr to ATG encoding Met. The frameshift mutation, 867delG was identified in the other allele. This mutation resulting from a deletion of a G in codon 290 would lead to a predicted truncated protein of 299 amino acids (including the first Met).
Independent molecular testing by the two European laboratories of the two male pseudohermaphrodite salt-wasting patients 15 and 16, from families 11 and 12, which originate from Sri-Lanka, revealed the presence of a 27-bp deletion in exon IV deleting the terminal base pair of codon 229 and all of codons 230–237, in addition to the first two base pairs of codon 238. This deletion did not alter the reading frame, but deleted the amino acid residues Ala-His-Leu-Ala-Leu-Arg-Ala. Both the individuals identified to be carrying this mutation were homozygous, whereas their parents were found to be heterozygous.
Following direct sequencing of exon IV in the HSD3B2 gene of the two male nonsalt-wasting pseudohermaphrodite patients 18 and 19 of the French family 14, the presence of a C to T transition was identified in the maternal allele converting codon 155 (CCG) encoding a Pro to CTG encoding a Leu. However, a G to T transversion was identified in the paternal allele converting codon 294 (GGC) encoding a Gly residue to GTC encoding a Val.
Direct sequencing of exon II of the HSD3B2 gene in the two Egyptian pseudohermaprodite patients 22 and 23 from family 16, who are suffering from a nonsalt-wasting form of 3ßHSD deficiency, revealed the presence of a homozygous C to T transition converting codon 10 (GCA) encoding an Ala residue to GTA encoding a Val residue. In addition, this mutation was found in the heterozygous state in both parents.
Finally, the presence of a heterozygous T to C transition converting codon 236 encoding a Leu (TTG) to TCG, encoding a Ser residue, was found in the HSD3B2 gene of the nonsalt-wasting male pseudohermaphrodite patient 32 from the French family 21. We have also identified in his other allele the frameshift mutation 867delG, which was previoulsy observed in the other French family 9.
Effect on 3ßHSD activity of the in-frame deletion 687del27 and of all missense mutations found in patients suffering of the severe salt-wasting form of 3ßHSD deficiency.
Site-directed mutagenesis for each of the mutants was performed as detailed in Methods, and sequence analysis was performed to confirm that the correct substitution or deletion had been achieved for each mutant cDNA construct. Following transient transfection of each mutant construct in 293 cells, the effect of the various mutations on enzyme activity was assessed by incubating intact cells in culture with 10 nM [14C]-DHEA as substrate for the indicated time periods. Such analyses were performed at least in two independent transfection experiments done in triplicate.
For the first time, we have characterized the effect of mutations
L205P, P222Q, T259M, T259R, and 687del27 on 3ßHSD activity in
comparison to the effect on all other reported missense mutations found
in patients suffering from the severe salt-wasting form of 3ßHSD
deficiency (Table 1
). As illustrated in Fig. 1, no significant transformation of
[14C]-DHEA was observed in intact
cells transfected with plasmid constructs expressing mutated
recombinant A10E, G15D, L108W, E142K, P186L, L205P, P222Q, Y253N,
T259M, T259R, and 687del27 proteins. Northern blot analyses
demonstrated that both wild-type and mutant transcripts were expressed
at equal levels in transfected 293 cells, whereas no significant
endogenous type II 3ßHSD mRNA (1.7 kb) was detected in mock
pCDNA3 transfected cells, thus confirming the
efficiency of 3ßHSD transcription from the adenoviral promotor of the
vector. Cohybridization to GAPDH was performed to determine equal
loading of RNA for each sample.
|
, which is in consistent with the severity of the
disorder in these index cases. Effect on 3ßHSD activity of all missense mutations found in patients suffering from the nonsalt-wasting form of 3ßHSD deficiency
Following the same experimental procedures described above, we
next characterized for the first time the functional significance of
the missense mutations A10V, A82T, P155L, L173R, L236S, P155L, P222H,
and G294V (Table 1; Fig. 1). To perform a consistent and more
comprehensive study, we have also reassessed, using the current
experimental procedure, the activity of previously reported mutant
enzymes detected in patients affected by the nonsalt-wasting form of
the disease (19, 58, 60), including N100S, G129R, A245P, and Y254D.
As illustrated in Fig. 1, no significant conversion of
[14C]-DHEA was obtained in intact
293 cells transfected with plasmid constructs expressing mutated
recombinant P155L, P222H or Y254D proteins. It can also be seen that in
cells transfected with pCDNA3-G294V or
pCDNA3-G129R, the apparent activity after a 1-h
incubation (percent conversion of
[14C]-DHEA in cells transfected
with each vector expressing the mutant protein/% conversion of
[14C]-DHEA in cells tranfected
with pCDNA3-3ßHSD type 2 x 100) were
20.5% and 11.7%, respectively. The in vitro results
obtained in intact cells suggest that the combined residual activity
catalyzed by these mutant type 2 enzymes approximates to 10.25%
(P155L/G294V) and 5.85% (P222H/G129R). However, as described below,
another mechanism(s) might affect such residual activity.
In 293 cells tranfected with pCDNA3-A10V,
pCDNA3-A82T, pCDNA3-N100S,
pCDNA3-L173R, or
pCDNA3-A245P, the apparent residual activities
after a 1-h incubation were 29.1%, 7.6%, 2.8%, 52.8%, and 35.4%,
respectively. All these mutant alleles were detected in the homozygous
state. Surprisingly, the activity of the mutant recombinant L236S
protein, which was found in the compound heterozygous patient 32
(L236S/867delG), was almost superimposable with that of the wild type
3ßHSD type 2 isoenzyme. Unexpectedly, no significant activity up to a
4-h incubation was observed with the T259M mutant protein, which was
identified in the homozygote nonsalt-losing 46XX individuals (patients
35 and 36), although low, but not significant, activity was observed
after 6-h incubation (Fig. 1).
Effect on 3ßHSD activity of sequence variants found in carriers with the nonclassic form of 3ßHSD deficiency
Several sequence variants and/or mutations (A167V, S213G, K216E, and L236S,) were detected with premature pubarche or hyperandrogenic adolescent girls suspected to be affected from nonclassical (late-onset) 3ßHSD deficiency (52, 54). To gain further knowledge concerning the molecular basis of such a disorder we next investigated the apparent activity of these mutant enzymes. In 293 cells transfected with pCDNA3-A167V, pCDNA3-S213G, pCDNA3-L236S, or pCDNA3-A245P, the apparent residual activities after a 1-h incubation were 81.45%, 58.4%, 58.95%, and 100%, respectively. Knowing that all previously reported heterozygous carriers bearing a deleterious mutation in the HSD3B2 gene were typically asymptomatic, these results provide additional molecular proof to support the conclusion that other genetic or environmental/hormonal influences may contribute to the expression of the observed symptoms (45, 48, 49, 50, 52, 68).
Kinetic analysis of mutant proteins
After the initial assessment of the effect of the various
mutations on enzyme activity, the mutations that were found to retain
activity were further analyzed to gain more information about the
apparent Km and Vmax for
each mutant enzyme. The transiently transfected cells were incubated
with various concentrations of DHEA ranging from 10
nM-50 µM (using 10 nM
[14C]-DHEA), under conditions
ensuring that first order kinetics were maintained. These analyses were
performed during at least three independent tranfection experiments, in
triplicate. The results, as shown in Fig. 2, demonstrate that certain mutant
enzymes [i.e. A10V (3.80 ± 0.27
µM), A167V (0.96 ± 0.10
µM), L173R (1.25 ± 0.04
µM), S213G (2.50 ± 0.26
µM), K216E (6.15 ± 1.53
µM), and L236S (2.73 ± 0.48
µM)] have only a small, if any, alteration in
their apparent Km compared to that of the
wild-type enzyme (1.91 ± 0.30 µM). On the
other hand, the expressed mutant A82T, G129R, and A245P proteins had a
much lower apparent affinity for DHEA with
Km values of 28.46 ± 4.18
µM and 19.58 ± 3.80
µM, respectively. The present results are
consistent with those obtained previously using cell homogenates
instead of intact cells (G129R, Km = 14 ± 2
vs wild type Km = 2.1 ± 0.2
µM; 61). However, there are variations in the
Vmax values for certain mutant proteins, for
example the Vmax for mutant L173R is
significantly reduced (11.47 ± 2.42 vs 127.01 ± 26.26 for
the wild-type), whereas A10V (40.01 ± 17.82), A167V (44.78
± 21.18) and G294V (39.17 ± 14.43) are approximately one third
that for the wild-type enzyme, which is important to note.
|
In the past few years, we have frequently observed that it is
difficult, if not impossible, to detect certain transiently expressed
mutant recombinant proteins by Western blot analysis. The next study
was designed to gain further biochemical data concerning this
phenomena. We thus compared the levels of expression and stability of
25 mutant recombinant 3ßHSD type 2 proteins. As illustrated in Fig. 3A, all transcripts were expressed in
transfected 293 cells following transient expression with the indicated
expression vector constructs, as revealed by Northern blot analysis.
The cells were also transfected with the pCDNA3
vector alone to show no endogenous expression of 3ßHSD type 2 mRNA.
The transfection efficiency has also been confirmed by using human
GAPDH as a control. In parallel, an in vitro
transcription/translation (TNT) rabbit reticulocyte lysate assay using
the mutant cDNA constructs was performed to show that each
pCDNA3 construct is adequately translated into a
[35S]-labeled-42-kDa protein, indicative of the
normal expression levels of mutant recombinant 3ßHSD type 2 proteins.
To determine whether all mutant proteins are, indeed, recognized by our
polyclonal antibodies, Western blot analysis of the corresponding
samples from the TNT assay were hybridized with an antihuman 3ßHSD
type 1 antibody. The data illustrated in Fig. 3B support the notion
that there is no detectable change on the polyclonal antibody binding
site as a consequence of these mutations. We have then performed
Western blot analysis using the homogenates purified from the same
cells transfected with the indicated expression vector constructs,
which have been used for RNA blot analysis illustrated in Fig. 3A. A
42-kDa band corresponding to the 3ßHSD type 2 protein was detectable
in several but not all homogenate preparations from 293 transfected
cells expressing the indicated wild type or mutant recombinant
proteins. The nonspecific band observed may also be used as an internal
control for loading. Such a difference in immunoblot signal levels has
been observed in several independent experiments.
|
The present study describes the identification of eight mutations,
in seven families with individuals suffering from classical 3ßHSD
deficiency, thus increasing the number of known HSD3B2
mutations involved in this autosomal recessive disorder to 31 (1
splicing, 1 in-frame deletion, 3 nonsense, 4 frameshift, and 22
missense mutations). In addition to providing further molecular
explanations for the heterogeneous clinical presentations, the
functional characterization of these mutant enzymes also generated
valuable information concerning the structure-function relationships of
the 3ßHSD superfamily. Indeed, in addition to these mutations
reported herein, we have also studied for the first time the functional
significance of previously reported missense mutations and or sequence
variants—namely, A82T (51), A167V (52), L173R (22), L205P (53), S213G
and K216E (54), P222H (55) T259M (55), and T259R (56)—that have not
previously been functionally characterized. Furthermore, their effects
have been compared to those of previously reported mutant enzymes,
including A10E (Alos, N., Moisan, A.M., Ward, L., et al.,
submitted for publication), G15D (57), N100S (58), L108W (59), G129R
(60), E142K (19), P186L (59), A245P (19), Y253N (19), and Y254D (61) to
provide a more consistent and comprehensive study with the same
experimental procedures using transfected intact 293 cells (Table 1).
Finally, the present study, therefore, provides evidence supporting the
involvement of a new molecular mechanism in classical 3ßHSD
deficiency involving protein instability and further illustrates the
complexity of the genotype-phenotype relationships of this disease.
Our present data further support the fact that it is more appropriate to assess the enzymatic activity of transiently expressed mutant proteins using intact 293 cells rather than homogenates from cells, due to the fact that the addition of exogenous cofactor can drive a reaction that may not occur in vivo. This conclusion is also consistent with our previous report showing that the homozygous mutation G15D found in a patient suffering from a severe salt-wasting form completely abolished the activity of the transiently expressed mutant protein in intact cells, whereas a significant residual activity was observed when using cell homogenates (57). The use of an excess of exogenous cofactor in studies using cell homogenates was responsible, at least in part, for this apparent discrepancy because this mutation, which is located in the NAD-binding domain, markedly decreases the affinity for the cofactor. The mutations N100S, L108W, and P186L also severely affected the affinity for the cofactor (57, 58). The advantage in using intact cells to assess the enzymatic characteristics was also indicated by several studies describing the properties of numerous hydroxysteroid dehydrogenases. For example, the type 1 isozyme of 11ß-hydroxysteroid dehydrogenase was always thought to act as both a dehydrogenase and an oxoreductase, and it was not until assessment of enzyme activity in intact cells (79 and references therein) that it became apparent that this isozyme behaves as a reductase in vivo, but in vitro can also act as a dehydrogenase due to the addition of exogenous cofactor (79 and references therein). A similar conclusion has been proposed concerning members of the 17ßHSD family (80). Furthermore, it was recently shown that the activity of the type 5 isozyme of 17ßHSD was unstable when assessed in homogenates of transfected cells as compared to assessment when using intact cells (81). However, we have previously observed in cells expressing the mutant protein A245P that no activity was obtained in cell homogenates in the absence of glycerol, whereas significant activity could be measured in intact cells (19).
Although the exact molecular and cellular explanation for the apparent
instability of various mutant recombinant proteins in intact
transfected 293 cells remains to be elucidated, our results illustrate
that it might be difficult, if not impossible, to rigorously measure
the levels of expression of some of these mutant proteins to obtain an
accurate estimate of their Vmax value. We are,
thus, suggesting that the various degrees of protein instability may
explain, at least in part, not only the observed decrease in the
Vmax values for several mutant proteins and more
specifically for those with the L173R or G294V substitution, but also
the absence of activity observed in 293 cells transiently expressing
mutant recombinant proteins A10E, G15D, L108W, P186L, A245P, Y253N,
T259M, and T259R. For example, although the overall efficiency of the
N100S protein found in the homozygous nonsalt-loser patient 17 (58) is
closely similar (
0.1%) to that of L108W and P186L found in the
compound heterozygous salt-loser patient 4 (59) The present study
illustrates the inherent limitations of such experimental approaches
that should be used in conjunction to further evaluate the impact of a
mutation on in vivo enzymatic activity. It should be noted
that, although the values obtained in the assessment of apparent
Km and Vmax in this study
are consistent with those of previously published mutations (60),
throughout these studies we have assessed enzyme activity for all the
mutants in intact cells to most resemble the situation in the cells of
the patient. In fact, the present results, thus, suggest that a single
amino acid substitution can alter the phenotypic expression of a
protein by altering its stability. In agreement with this finding, it
has previously been shown that a single amino acid mutation in
Cytochrome P450 (P4502C13), namely Ser180 to Cys,
located in a highly conserved region in the P4502C subfamily,
determines a polymorphism by altering protein stability (82).
The present study is also in agreement with the prediction that no
functional 3ßHSD type 2 isoenzyme is expressed in the adrenals and
gonads of the patients suffering from a severe salt-wasting form of CAH
due to classical 3ßHSD deficiency, as summarized in Table 1. Note
that the 10 patients suffering from a severe salt-wasting form of
3ßHSD deficiency, who bear homozygous [W171X (18), R259X (40, 56),
Y308X (56), 820delAA (77)] or compound hetrozygous [W171X/558insC
(18), 820delAA/952delC (78)] mutations, were not included in Table 1.
It has been predicted that these nonsense and frameshift mutations lead
to nonfunctional truncated proteins. Taken together, all these results
are in perfect agreement with the severity of this form of CAH. On the
other hand, the observed peripheral conversion of
5-hydroxysteroids in these patients is
consistent with the now recognized important biosynthesis of sex
steroids in peripheral tissues (39). However, although very low levels
of 3ßHSD type 1 transcripts can be detected in normal gonads,
4 steroids can originate from gonadal 3ßHSD
type 1 activity, which possesses a roughly 10-fold higher affinity than
the type 2 isoenzyme and which could be stimulated after an increase in
LH secretion, resulting from low-circulating androgen levels at puberty
(8, 18, 40).
In addition, the present study demonstrates that the nonsalt-losing
form of classic 3ßHSD deficiency also results from missense
mutation(s) in the HSD3B2 gene, which causes an incomplete
loss of enzymatic activity, thus leaving sufficient enzymatic activity
to prevent salt wasting (19, 58, 60). On the other hand, it is worth
noting that the kinetic properties of the N100S protein, found in a
nonsalt-losing patient (58) are quite similar to those of L108W and
P186L mutant proteins, which were detected in a severe salt-losing form
of classical 3ß-HSD deficiency (59) when kinetic properties were
measured using cell homogenates. The functional characterization of
these mutant proteins in cell homogenates cannot provide an explanation
for the heterogeneity responsible for the severe salt-losing form
(L108W:P186L) down to the clinically inapparent form of salt loss
(N100S) of classic 3ßHSD deficiency; nevertheless, the hormonal
profile of the N100S mutation suggests that salt loss was compensated
for by a limited capacity of aldosterone biosynthesis at the price of
high renin synthesis (Table 1). In addition, the present data
suggesting the instability of L108W and P186L, but not of N100S, may
well explain, at least in part, the observed difference in the levels
of activity using intact cell assays which is in accordance with the
phenotypic differences observed in these two patients.
It is also of interest to mention that the striking phenotypic differences observed between the homozygous salt-losing patients 1 and 2, bearing the A10E mutation (Alos, N., Moisan, A.M., Ward, L., et al., submitted for publication), and the nonsalt-losing patients 22 and 23, bearing the A10V mutation, is in accordance with their respective enzymatic properties, as determined in intact cells. Such a difference may be the result of both the observed apparent instability of the A10E protein coupled with the fact that Glu is a negatively charged residue, whereas Val, like Ala, is a nonpolar residue. It should also be noted that this Ala is located in the highly conserved putative NAD-binding domain (21, 57, 83, 84).
In general, the present functional data are in close agreement with the
severity of the disease in patients suffering form the nonsalt-wasting
form of 3ßHSD deficiency. In this respect, it can be estimated that
the 3ßHSD activity catalyzed by the mutant type 2 proteins in the
compound heterozygotes P155L/G294V, P222H/G129R and G129R/6651(G to A),
as well as the homozygotes A10V, A82T, and A245P will be 10.2, 5.8,
5.8, 29, 7.6, and 35.4%, respectively, as measured in intact cells. On
the other hand, knowing that all heterozygote carriers of a deleterious
mutation in the HSD3B2 gene are asymptomatic, it was
unexpected to observe such a relatively high activity of L173R,
i.e. 52.8% (Fig. 1C), but the apparent instability of L173R
would most likely be involved in further reducing the activity
catalyzed by the mutant 3ßHSD type 2 protein in the cells of the
adrenals and gonads in patients 30 and 31. It is also possible that the
apparent instability of A245P and G294V will play a role in further
decreasing the activity in patients from families 14 and 22, as also
suggested by their Vmax/Km
values.
Although the results described above illustrate very well the almost perfect genotype-phenotype relationships, there are several examples supporting the notion that there are exceptions to the rule. First, the homozygous A82T mutation that was previously reported in four Brazilian patients 24, 25, 26 and 27 (36, 51). In family 18, it associated with precocious puberty, whereas in the unrelated family 17, it was, indeed, associated with male pseudohermaphroditism, but had no effect in the homozygous female relative. However, it has recently been demonstrated that although the homozygous A10E patient, patient 2, presented with the typical phenotype of ambiguous genitalia at birth with normal masculinization at puberty, the female patient (patient 1) also harboring this homozygous mutation, presented with spontaneous feminization and menarche (Alos, N., Moisan, A.M., Ward, L., et al., submitted for publication), in contrast to the other 46XX patient with severe 3ßHSD deficiency (due to a homozygous W171X mutation), for whom follow-up at pubertal age has been reported, who was hypogonadal (18, 85). Another astonishing example is the observation that the Brazilian patients 35 and 36 bearing the homozygous mutation T259M suffer from a nonsalt-wasting form of the disease, whereas the compound heterozygous French patients 11 and 12 with T259M/867delG are affected by the severe salt-wasting form. It is, thus, impossible to correlate the phenotype of patients 35 and 36 with the present data showing that cells expressing mutant T259M proteins have no 3ßHSD activity, and in comparison with the T259R mutation, we provide evidence suggesting the instability of the mutant T259M protein.
It was also unexpected to observe that the L236S mutation, which was found in the compound heterozygous nonsalt-losing patient 32 (L236S/867delG), possesses the same enzyme activity as the wild-type enzyme, and, furthermore, there is no evidence that this mutation affects the stability of the protein. Although some genetic alterations affecting the synthesis of a multimeric protein may have a dominant effect (86), such evidence remains to be demonstrated for the 3ßHSD family. Nevertheless, this hypothesis is, thus, difficult to reconcile with the well established fact that heterozygous carriers are asymptomatic (6, 20, 21, 22). On the other hand, we cannot rule out that the L236S mutation could be in linkage disequilibrium with another deleterious mutation affecting the expression or the splicing of this gene.
The present study also demonstrates that not only the L236S mutation
but also the heterozygous A167V sequence variant leads to proteins that
have similar activity to the native enzyme, whereas mutant S213G and
K216E proteins had only minor impact on the activity. It should also be
noted that the mother of patient 37 was also a heterozygous carrier for
this variant, but did not have any symptoms of hyperandrogenism (52).
The present functional data concerning these sequence variants coupled
with the previous studies reporting that no mutations were found in
both HSD3B1 and/or HSD3B2 genes in the patients
affected by premature pubarche or hyperandrogenism (45, 46, 47, 48, 49, 68),
strongly support the conclusion that this disorder does not result from
a mutant 3ßHSD isoenzyme. We cannot refute the possibility that
inherited mutation(s) could be located farther upstream in the putative
promoter region of the HSD3B2 gene, leading to an aberrant
level of expression of a normal type II 3ßHSD protein. However, the
latter hypothesis is markedly weakened by the fact that all patients
come from unrelated pedigrees and diverse ethnic origins. On the other
hand, because 3ßHSD gene expression and activity are under complex
multiple hormonal regulation (6), it cannot be ruled out that at least
some forms of NC3ßHSD deficiency result from a genetic or acquired
origin acting indirectly on these modulatory parameters. There is also
the possibility of the implication of a steroidogenic enzyme different
from known 3ßHSD isoenzymes. Does it possibly involve dysregulation
of 17-hydroxylase and 17,20-lyase activities, as has been suggested
(87)? If so, is this dysregulation of genetic origin? All these
hypotheses remain to be further studied to gain more understanding of
this puzzling but frequent disease.
Finally, the functional characterization of these mutant enzymes also
generated valuable information concerning the structure-function
relationships of the 3ßHSD superfamily. Indeed, as indicated in Table 1, of special interest to note is that the amino acid residues that are
the sites of the missense mutations are generally in highly conserved
regions in members of the vertebrate 3ßHSD isoenzymes characterized
thus far in the human, macaque, bovine, rat, mouse, hamster, chicken,
and rainbow trout (for a review, see Ref. 21). This finding strongly
suggests the crucial role of these residues for the catalytic activity
of these enzymes. However, for example, although amino acid
Pro186 is also well conserved in the vertebrate
3ßHSD family, it is not conserved in all members; namely rat type 3,
mouse types 4 and 5 and hamster type 3, which are specific
3-ketoreductases responsible for the conversion of 3-keto-saturated
steroids using NADPH as cofactor, which do not share this amino acid
residue at this position (6, 21). It is also of interest to mention
that mutations A10E, A10V and G15D change an amino acid in the highly
conserved Gly-X-X-Gly-X-X-Gly region found in all members of the
3ßHSD superfamily (89), which is similar to the common
Gly-X-Gly-X-X-Gly conserved sequence present in most NAD(H)-binding
enzymes (83, 84). Moreover, mutations A82T and G294V create a
substitution in each of the two predicted membrane-spanning domains (6, 88). Furthermore, mutation P155L is located in the first of the two the
characteristic Y-X-X-X-K sequences located in the region from
Tyr154 to Lys158 and
Tyr269 to Lys273, which is
found in the active site of short-chain alcohol dehydrogenases (83, 90). Affinity labeling of purified human type I 3ßHSD identified two
tryptic peptides, comprising amino acids Asn176
to Arg186 and Gly251 to
Lys274 that should contain residues involved in
the putative substrate-binding domain (91). Thus, the exact role of the
first YXXXK motif in the 3ßHSD family remains to be confirmed.
Finally, recent findings have shown that His261
is a critical amino acid residue for 3ßHSD activity and
Tyr253 or Tyr254
participates in the isomerase activity of the human 3ßHSD type 1
enzyme (92, in addition to evidence that Tyr253
functions as the general acid (proton donor) in the isomerase reaction
(93). Consequently mutations located within this area will inevitably
have a major effect on enzyme activity, as exemplified in the case of
mutations Y253N, Y254D, T259M, and T259R, which completely abolish
enzyme activity.
In summary, these studies have provided further insight into the molecular basis of 3ßHSD deficiency and have highlighted the fact that mutations in the HSD3B2 gene can result in a wide spectrum of molecular repercussions that are associated with the different phenotypic manifestions of classical 3ßHSD deficiency. The present results have also demonstrated the fact that the genotype correlates very well with the phenotype in most cases studied, with the exception of patients 35 and 36 bearing the homozygous mutation T259M, whereas additional studies will be needed to elucidate the real impact of the locus harboring the L236S mutation. Moreover, the present study further demonstrates the importance of the measurement of 17OH-PREG levels, which always exceeded 100 nmol/L in patients suffering from classical 3ßHSD deficiency. The functional characterization of all the missense mutations known to be involved in this disease also provides valuable information concerning the structure-function relationships of the 3ßHSD superfamily. Finally, the present studies have highlighted the fact that various mutations seem to have a drastic effect on the stability of the protein, thus providing the first molecular evidence of a new mechanism involved in classical 3ßHSD deficiency.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Contributed equally to this work and should be considered as equal
first author.
Received October 13, 1999.
Accepted October 15, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsMethodsResultsReferences |
|---|
5-4-isomerase in
adrenal cortex mitochondria and microsomes. Endocrinology. 134:751–759.[Abstract]
5-4-isomerase gene family.
J Endocrinol. 150:S189–S207.
5-ene isomerase in
human adrenals and gonads. Mol Endocrinol. 5:1147–115.[CrossRef][Medline]
5-4-isomerase gene and
its expression in mammalian cells. J Biol Chem. 265:20469–20475.
(Additions and corrections, J Biol Chem. 267:3551).5-4-isomerase (3ßHSD)
gene: adrenal and gonadal specificity. DNA Cell Biol. 10:701–711.[Medline]
5-4-isomerase close to D1S514 and the centromeric
D1Z5 locus. Cytogenet Cell Genet. 69:59–62.[Medline]
5-4-isomerase
type 1 gene transcription in human breast cancer cell lines and in
normal mammary epithelial cells by interleukin-4 and interleukin-13. Mol Endocrinol. 13:66–81.
