Molecular Human Reproduction Vol.8, No.9 pp. 823–832, 2002
A 60–66 kDa protein with gonadotrophin surge attenuating
factor bioactivity is produced by human ovarian granulosa
cells
Paul A.Fowler1,6, Tarja Sorsa-Leslie1,2, Phillip Cash3, Bryan Dunbar2, William Melvin2,
Yvonne Wilson4, Helen D.Mason5 and William Harris2
Departments of 1Obstetrics and Gynaecology, 2Molecular and Cell Biology and 3Medical Microbiology, University of Aberdeen,
Aberdeen, AB25 2ZD, 4The Assisted Conception St James’s University Hospital, Leeds and 5Departments of Obstetrics and
Gynaecology and of Physiology, St George’s Hospital Medical School, London, SW17 0RE, UK
6To
whom correspondence should be addressed. E-mail: [email protected]
We aimed to confirm the ovarian site of gonadotrophin surge-attenuating factor (GnSAF) production and produce granulosa/
luteal cell-conditioned medium (G/LCM) containing GnSAF for purification studies. Blue dye affinity chromatography followed
by pseudochromatofocusing of G/LCM yielded bioactive fractions at pH 5.74 and 5.77. The former had a major 60–66 kDa
band with an internal amino acid sequence of EPQVYVHAP following tryptic digestion. A rat polyclonal antiserum (rPAb)
raised against this band completely blocked in-vitro GnSAF bioactivity in human follicular fluid, serum and G/LCM. GnSAF
bioactivity was localized to a 64 kDa band of serum-free G/LCM and following 2D gel electrophoresis, one of the spots
recognized by Western blotting with the GnSAF rPAb had an N-terminal amino acid sequence of NH-XVPQGNAXXN. Neither
amino acid sequence had significant homology with proteins in the human genome database. When ovarian tissues from
spontaneously cycling women were cultured under serum-free conditions, neither theca- nor stroma-conditioned media contained
GnSAF bioactivity. However, granulosa cell-conditioned medium significantly reduced GnRH-induced LH secretion, an effect
that was reversed by incubation with the GnSAF rPAb. In conclusion, we have confirmed that human granulosa cells produce
GnSAF within the ovary and have two candidate amino acid sequences for GnSAF. We have also demonstrated that serumfree granulosa cell culture constitutes the method of choice for the characterization of GnSAF since recovery of bioactivity is
superior in the presence of fewer serum proteins.
Key words: FSH/GnRH/granulosa cells/LH/pituitary
Introduction
Treatment with FSH causes reduced GnRH-induced LH secretion in
women, monkeys, cows, pigs and rats (Littman and Hodgen, 1984;
Danforth et al., 1987; Koppenaal et al., 1993; Messinis et al., 1993;
Fowler and Price, 1997), resulting in endogenous LH surges being
reduced or abolished (Sopelak and Hodgen, 1984; Messinis and
Templeton, 1986). This effect has been attributed to the ovarian
hormone GnSAF (also termed GnSAF, GnSIF, attenuin) (Fowler and
Templeton, 1996) which reduces GnRH self-priming, an important
component of the LH surge, both in vivo (Messinis and Templeton,
1991) and in vitro (Koppenaal et al., 1992). The biological effect of
GnSAF is not due to steroids or inhibin because as early as 8 h after
a single FSH injection, pituitary responsiveness to GnRH is reduced
in women whereas circulating estradiol and inhibin do not rise
significantly until 24 h after the FSH injection (Messinis et al., 1991,
1993, 1994a; Burger et al., 1998). This is supported by in-vitro data
demonstrating that inhibin antibodies do not bind to GnSAF (Byrne
et al., 1995). Since the treatment of post-menopausal women with
exogenous FSH has no effect on pituitary responsiveness to GnRH,
a direct pituitary effect of FSH is not responsible for the GnSAF
effect of reduced GnRH-induced LH secretion (Messinis et al.,
© European Society of Human Reproduction and Embryology
1994b). Furthermore, GnSAF bioactivity is present in both follicular
fluid (FF) (Fowler et al., 1995) and serum (Fowler et al., 1994a) and
is produced primarily by small follicles (Fowler et al., 1994b, 2001),
supporting the concept that this hormone has a role in ‘clamping’
pituitary responsiveness to GnRH during the early to mid-follicular
phase (de Koning et al., 2001).
Although three putative GnSAF amino acid sequences (Tio et al.,
1994; Danforth and Cheng, 1995; Pappa et al., 1999) have been
published, they differ, were derived from proteins between 12–69
kDa in mass, and have not been confirmed as GnSAF. The purification
of GnSAF has been fraught with problems, including low concentrations in biological fluids despite high bioactivity, co-elution with
serum albumin and the large number of proteins in FF and serum
(Fowler and Templeton, 1996). These difficulties have hampered
advances in the field.
The current study was devised specifically to answer a connected
series of questions: (i) is GnSAF definitely produced by the granulosa
cell? (ii) Are cultured granulosa cells a potential source of GnSAF
for characterization studies? (iii) Can protein contamination of GnSAF
preparations be reduced using in-vitro culture techniques?
In this paper we present evidence that the human granulosa cell
823
P.A.Fowler et al.
produces the bioactive GnSAF molecule which has a molecular mass
of 64 kDa and an isoelectric focusing point (pI) of 5.7–5.8. We also
present two further candidate GnSAF amino acid sequences.
Materials and methods
Source of material
Human follicular fluid and IVF granulosa/luteal cells
All protocols employing human subjects were given joint ethical committee
approval at Aberdeen and patients all gave informed consent. On three
occasions over the time course of this long study human (h)FF was aspirated
from follicles 艋18 mm in diameter from 40 women undergoing IVF in
Aberdeen, and pooled and desalted as previously described (Fowler et al.,
2001). Subsequently, 500 µl aliquots of these hFF pools were stored at –20°C
and used as a GnSAF bioactivity quality control (QC), producing a 40–60%
reduction in GnRH-induced LH at 50 µl/well, in all bioassays performed as
part of the present study as previously described (Fowler et al., 2001).
Granulosa/luteal cells were recovered from hFF obtained from women
undergoing IVF as follows. The aspirated hFF in the collection tubes was
allowed to settle for 1 h at 8°C and then, under sterile conditions, all but the
lower 1 ml of hFF was aspirated from the tubes and discarded. The remaining
hFF was then pooled and purified by Histopaque (Sigma-Aldrich Co. Ltd,
Poole, Dorset, UK) sedimentation (3 ml Histopaque:12 ml hFF, at 850 g for
20 min at 8°C). The cells were washed with 20 mmol/l phosphate-buffered
saline (PBS), centrifuged at 850 g for 20 min at 8°C and the pellet was mixed
with culture medium. For the first purification attempt, the cells were cultured
in serum-free DMEM [supplemented with 0.1% bovine serum albumin (BSA),
2% penstrep (Sigma-Aldrich), 0.5% gentamycin (Sigma-Aldrich), 1.2 g sodium
bicarbonate/500 ml and 10 IU FSH/ml (Metrodin; Serono)] at 100 000 cells/
500 µl medium in 24-well culture dishes for 72 h. Once 460 ml of this
conditioned medium (G/LCM) had been accumulated, it was tested for GnSAF
bioactivity and used as indicated below. For the second purification attempt,
a further 2000 ml of G/LCM, obtained by culturing granulosa/luteal cells
in serum-free and BSA-free M199 [supplemented with 2% penstrep, 1%
gentamycin, 2% amphoterecin (all Sigma-Aldrich) and 10 IU FSH/ml], was
also tested for GnSAF bioactivity and then used as indicated below. In neither
case was inhibin removed because very high concentrations are required to
significantly affect GnRH-induced LH secretion from rat pituitary cells (Fowler
et al., 1994b). Nevertheless, inhibin bioactivity was monitored during the
purification process to confirm that GnSAF bioactivity was not due to inhibin
contamination.
All G/LCM was concentrated 50-fold using a tangential-flow concentration
system (Minitan S™; Millipore Corp, Bedford, MA, USA) with a 10 kDa
MWcut-off polysulphone filter. The concentrated conditioned medium was
then desalted into sterile distilled water.
Spontaneous cycle human granulosa, theca and stroma
Granulosa cells and thecal and stromal tissue were obtained from spontaneously
cycling women. The ovaries were obtained from women undergoing oophorectomy for non-ovarian gynaecological pathology. Approval for the current
study was granted by the local ethics committees of the hospitals concerned
and each woman gave informed consent. Samples were only collected from
women with a history of regular cycles and surgery was performed at random
stages of the cycle. Individual follicles were microscopically dissected intact
from the surrounding stroma. Follicles were then incised and the granulosa
cells flushed from the surface and cultured. The remaining theca ‘shell’ was
minced into small pieces, washed twice and 2–3 mg pieces were cultured.
Stroma dissected from the ovaries was chopped, washed and 5–6 mg were
cultured. All cultures were in serum-free M199, without any added androgen,
as previously reported (Mason et al., 1990). After 48 h, the granulosa cellconditioned medium (GCM), theca-conditioned medium (TCM) and
stroma-conditioned medium (SCM) were sent to Aberdeen, desalted and
stored at –20°C.
Purification methods used
Dyematrex Blue chromatography
The G/LCM batches were mixed at a ratio of 1 ml medium:5 ml Dyematrex
Blue A (Amicon Ltd, Stonehouse, Glos., UK), which has a high affinity for
824
serum albumin, for 48 h at 7°C in 20 mmol/l Tris–HCl, pH 7.5. The G/LCM
proteins that did not bind to the blue dye were recovered by centrifugation
(400 g) while the bound proteins were eluted using 20 mmol/l Tris–HCl ⫹
1.5 mol/l KCl, pH 7.5. Unbound and bound fractions were concentrated to
1 ml using Macrosep 10 kDa MWCO filters at 5000 g (Flowgen Instruments
Ltd, Sittingbourne, Kent, UK) and then desalted.
Pseudochromatofocusing (PCF) of G/LCM
Blue-dye treated G/LCM was further purified by pseudochromatofocusing
(PCF) using 3.3 g of Bakerbond CBX media (JT Baker Inc., Phillipsburg, NJ,
USA) packed into an HR 10/10 column (Amersham) as previously described
(Fowler and Price, 1997). The serum protein-depleted G/LCM samples were
equilibrated in MES buffer (2, N-morpholino, ethanesulphonic acid; SigmaAldrich), pH 5.0 and pseudochromatofocused at 200 ml/h. A pH gradient
from 4.8–6.8 was generated by sequentially passing 15 ml of 75% MES ⫹
25% sodium acetate [500 mmol/l sodium acetate (NaAc), pH 8], 50% MES
⫹ 50% NaAc and 25% MES ⫹ 75% NaAc down the column. Each run was
terminated with 140 ml of 100% NaAc buffer. The pH of each of the fractions
was determined prior to concentration to 1 ml through Macrosep 10 kDa
MWCO filters at 5000 g (Flowgen Instruments). Before bioassay, each
concentrated fraction was passed through a 0.2 µm filter and double-desalted.
The chromatographic fractions were bioassayed at 1, 5 and 25 µl/well in two
separate rat pituitary cell cultures.
2D gel electrophoresis
Proteins in concentrated, bioactive PCF fractions of blue dye-processed
G/LCM were visualized by 1D sodium dodecyl sulphate (SDS)–polyacrylamide gel electrophoresis (PAGE) or dried down using a vacuum centrifuge
and resuspended at a ratio of 250 µl of unprocessed G/LCM per 100 µl of
lysis buffer [0.01 mol/l Tris–HCl, pH 7.4, 1 mmol/l EDTA, 8 mol/l urea,
0.05 mol/l dithiothreitol (DTT), 10% (v/v) glycerol, 5% (v/v) NP40, 6%
(w/v) pH 3–10 Resolyte (Merck)] (Cash, 1989). Following centrifugation at
11 000 g, the supernatant was stored at –70°C.
Soluble proteins were analysed by 2D gel electrophoresis gels (2D gels)
using a small format gel system (Cash et al., 1995, 1997) with immobilized
pH gradient (IPG) gels for the first dimension separation. Briefly, 25 µl of
the bioactive fraction following PCF of BSA-free, blue dye-treated G/LCM
(fraction #10) in lysis buffer, prepared as described above, was adjusted to a
final volume of 125 µl in IPG re-swelling buffer [7 mol/l urea, 2 mol/l
thiourea, 4% (w/v) CHAPS, 0.3% (w/v) DTT, 2% (w/v) pH 4–8 Resolyte
(Merck)]. The loading volumes for the G/LCM fractions were estimated by
analysis of the soluble protein samples by SDS–PAGE (Cash, 1989). The
proteins were separated in the first dimension using 7 cm, pH 4–7 IPG gel
strips (Amersham). The dehydrated IPG strips were rehydrated overnight in
the IPG re-swelling buffer containing the conditioned media (Rabilloud et al.,
1994). Following their rehydration the IPG gel strips were electrophoresed on
a Multiphor II apparatus (Amersham). The IPG gels were then equilibrated
in buffer containing 50 mmol/l Tris–HCl, pH 8.8, 6 mol/l urea, 30% (v/v)
glycerol, 2% (w/v) SDS and 1% (w/v) DTT for 30 min followed by a second
equilibration for 30 min in the same buffer plus 2.5% (w/v) iodoacetamide.
The equilibrated strips were overlayed onto 7⫻8 cm 10–15% gradient
polyacrylamide slab gels and processed as described previously (Cash et al.,
1999). Proteins were located by staining with colloidal Coomassie Brilliant
Blue G250 (Cash et al., 1997). Protein spots of interest were cut from the
slab gels using a sterile scalpel and placed into 60 µl of sterile MilliQ water
ready for Edman sequencing.
Localization of GnSAF bioactivity in 1D and 2D gels
In order to estimate the molecular mass variants of GnSAF bioactivity,
BSA-free, blue dye-treated G/LCM was boiled 1:1 with treatment buffer
(0.0625 mol/l Tris, 2% SDS, 190% glycerol, 5% mercaptoethanol) for 10 min
and 10 µl volumes were run in eight lanes of 1D reducing SDS–PAGE gels
with prestained MW markers (BioRad Laboratories Ltd, Hemel Hempstead,
Herts, UK) in the outside lanes. One lane of MW marker and the adjoining
lane of G/LCM in each gel were stained with Coomassie Blue and aligned
with the remaining seven lanes of G/LCM and the single MW marker lane.
Two separate gels were used for each bioassay and were run as follows: 1D
SDS–PAGE according to the standard method (Laemmli, 1970). Briefly, 7.5%
separating gels, topped with 4% stacking gels, were used in a minigel apparatus
(Hoefer MightySmall; Amersham) and run at 15 mA for 60 min in running
buffer (0.025 mol/l Tris–HCl, 0.192 mol/l glycine, 0.1% SDS, pH 8.3) and
Production of human GnSAF
Figure 1. Dose-dependent GnSAF bioactivity in granulosa/luteal cell-conditioned medium (G/LCM): (a and b) bioactivity in raw medium while (c and d)
bioactivity in medium treated with Dyematrex Blue A to deplete serum proteins/albumin was limited to the unbound proteins (closed circles) rather than
bound proteins (open triangles). Data represent the mean of quadruplicate determinations from two different rat pituitary cell culture bioassays. The shaded
horizontal bar indicates the mean ⫾ SE range for control GnRH-induced LH secretion. Data are shown as means ⫾ SE.
visualized by Coomassie Blue staining. The main protein bands and strips of
gel with no visible bands (controls), were cut out with alcohol-washed
instruments on glass and washed in a rotary mixer under sterile conditions as
follows: (i) four 10 min washes in distilled water ⫹ 2.5% Triton-X, (ii) four
2 min washes with distilled water, and (iii) two 2 min washes in 20 mmol/l
PBS. After 2 min air drying the slices were incubated in culture medium for
48 h at 7°C. The gel slices were then discarded and debris removed from the
media by 0.8 and 0.2 µm filtration. The gel-conditioned media samples were
double-desalted into sterile distilled water, as described above, concentrated
to 250 µl and bioassayed at 1, 5 and 25 µl/well in two separate rat pituitary
cell cultures (25 µl being equivalent to 7 µl of starting material).
In order to locate GnSAF bioactivity in 2D gels, 500 µl of serum-free
G/LCM was dried down, resuspended in 500 µl of lysis buffer and divided
between five mini-2D gels, as described above. One gel was stained with
Coomassie Brilliant Blue while the remaining four were aligned in a stack
and sliced into nine equal squares which were then treated as described above,
concentrated to 400 µl and bioassayed at 1, 5 and 25 µl/well in three separate
rat pituitary cell cultures (25 µl being equivalent to 25 µl of starting material).
Edman sequencing
In the case of the BSA-containing G/LCM, the bioactive PCF fraction (fraction
#11) was concentrated to 50 µl using a vacuum centrifuge and applied to a
1D gel as previously described (Dunbar and Wilson, 1994). The major protein
band between 60 and 70 kDa was subjected to Edman sequencing as previously
described (Christie et al., 1993; Votyakova et al., 1999). Since the N-terminus
was blocked, the gel was repeated and the band subjected to tryptic digestion
using established protocols (Lui et al., 1996) and the Edman sequencing
repeated. In the case of the BSA-free G/LCM proteins separated on 2D gels,
the three excised protein spots identified by their position on the gel and
Western blotting were used subsequently for Edman sequencing (Christie
et al., 1993; Votyakova et al., 1999). The results of the Edman sequencing
were subjected to online searches (http://www.ncbi.nlm.nih.gov/blast/) of the
Human Genome, nr and Swissprot protein databases.
Antibody methods
GnSAF polyclonal antiserum production
BSA-containing G/LCM that was not used for tryptic digestion and Edman
sequencing of PCF fraction #11 was run down eight lanes of 10% 1D SDS–
PAGE gel (10 µl/lane) and the major 60–70 kDa band excised from each
lane. Four of these bands were homogenized in 100 µl Freund’s Complete
adjuvant while the other four were homogenized in 10 µl of Freund’s
Incomplete adjuvant (Sigma-Aldrich). The Freund’s Complete adjuvant mix
was used to immunise two Sprague-Dawley adult male rats (⬎500 g body
mass). The rats were given booster injections at 21 and 51 days with the
Freund’s Incomplete adjuvant and exsanguinated at 61 days. Aliquots of
100 µl aliquots of serum, previously collected from three post-menopausal
women (PMWS) as part of a different study, were used to immunise an
additional male rat in order to provide control antiserum free of antibodies
recognizing ovarian proteins.
Western blots
Gels were transferred onto 0.45 µm nitrocellulose or PVDF at 200 V for
75 min in transfer buffer (0.25 mol/l Tris–HCl, 19.2 mmol/l glycine, 0.01%
SDS, 2% methanol), washed in blocking buffer containing 2% non-fat
powdered milk in Tris-buffered saline (TBS) for 45 min then washed for
3⫻10 min in TBS containing 3 g Tris, 8 g NaCl, 0.2 g KCl/l, pH 8, prior to
incubation with primary antibody at room temperature for 90 min. The
membranes were washed (3⫻10 min, TBS-Tween 20) and incubated with
alkaline phosphatase-conjugated IgG diluted in TBS for 90 min. Recognized
proteins were directly visualized with 5-bromo-4-chloro-3-indolylphosphate
with nitroblue dimethyl formamide (Sigma-Aldrich). 1D SDS PAGE gels of
G/LCM transferred onto nitrocellulose were probed with anti-GnSIF kindly
provided by Dr D.R.Danforth and anti-GnSAF and anti-PMWS PAbs (see
above), while 2D gels were transferred onto nitrocellulose prior to probing
with anti-GnSAF and anti-PMWS PAbs.
Bioactivity blocking effect of GnSAF antiserum
In order to determine whether the antisera blocked in-vitro GnSAF bioactivity,
both the anti-GnSAF and anti-PMWS PAbs were mixed gently 1:1 in 100 µl
825
P.A.Fowler et al.
Figure 2. Pseudochromatofocusing resolution of GnSAF bioactivity detected in blue dye-treated, BSA-containing, granulosa/luteal cell-conditioned medium
(G/LCM). (a) The thin line shows the protein elution profiles at 280 nm while the thick lines show the pH of the fractions eluted from the column and the
circles bars (means ⫾ SE) show the effects of the eluted fractions of GnRH-induced LH secretion from rat pituitary cell culture bioassay. GnSAF bioactivity
was eluted in fraction #11 at pH 5.74, (b) presence of GnSAF bioactivity (suppression of GnRH-induced LH secretion, shown by closed circles), but not
inhibin bioactivity (suppression of basal FSH secretion, shown by open circles) in fraction #11. Data represent the mean of quadruplicate determinations from
three different rat pituitary cell culture bioassays. The shaded horizontal bar indicates the mean ⫾ SE range for control GnRH-induced LH secretion.
aliquots with hFF, pre-hCG IVF serum, G/LCM, GCM, TCM and SCM, for
1 h at 37°C in 24-well dishes on an orbital shaker. Controls were PAbs and
hFF, pre-hCG IVF serum, BSA-free G/LCM, GCM, TCM and SCM incubated
separately with culture medium only. The incubation experiment was repeated
twice. The samples were then applied to two separate bioassays at 1, 5 and
25 µl/well.
Bioassay, hormone and statistical methods
GnSAF bioassay
Adult female Sprague-Dawley rats (10–14 weeks old) were maintained under
a constant 12 h light:12 h dark, 22°C environment with ad libitum access to
food and water. For each cell culture 15 rats, selected at random during the
estrous cycle, were killed by stunning and cervical dislocation. Dispersion
and culture of the pituitary cells in serum-free defined medium (SFDM) was
carried out as previously described (Fowler et al., 1994b, 2001).
Experiments were carried out in quadruplicate wells: 200 µl of fresh SFDM
was added, together with the treatments made up to 25 µl with SFDM. All
the culture plates contained at least 12 control wells receiving SFDM only.
After 24 h incubation with the test substances, the medium was collected and
stored at –20°C for subsequent measurement of basal FSH as an index of
inhibin bioactivity. The wells were then treated with 0.1 µmol/l GnRH
(Fertagyl; Intervet UK Ltd, Cambridge, UK) in 50 µl of SFDM. In all dishes,
eight wells previously exposed to SFDM received GnRH alone while four
wells previously exposed to SFDM received 50 µl of SFDM instead of the
50 µl of GnRH challenge. These acted as controls for the magnitude of the
GnRH response. Cultures were terminated after 4 h incubation by collecting
the media which was stored at –20°C for subsequent measurement of GnRHinduced LH as an index of GnSAF bioactivity. The QC hFF preparations
were added to each bioassay at 1, 5 and 25 µl/well, in at least four wells/
dose/separate culture, to act as a GnSAF quality control. Bioassays in which
the QC hFF caused ⬍30% suppression of GnRH-induced LH secretion, or in
826
which the control GnRH response constituted ⬍50% increase in LH, were
repeated and the data discarded.
Hormone assays
Concentrations of gonadotrophins in cell-conditioned media from rat anterior
pituitary cell cultures were determined using: (i) a homologous rat radioimmunoassay for FSH with sensitivity and intra-assay and inter-assay coefficient of variation values of 0.6 ng FSH/ml (NIDDK-rFSH-RP-2) using
NIDDK-anti-rFSH-S11 and 7.1 and 11.2% respectively; (ii) a homologous rat
time-resolved fluoroimmunoassay for LH with sensitivity and intra-assay and
inter-assay coefficient of variation values of 0.2 ng LH/ml (NIDDK-rLHRP3) using NIDDK-anti-rLH-S11 and 5.4 and 7.9% respectively.
Statistical analysis
The in-vitro pituitary cell responses are expressed as percentages of the
relevant control gonadotrophin concentrations secreted from blank control
wells on the same culture dishes. These controls were either wells exposed
to SFDM alone (basal secretion) or wells exposed to SFDM ⫹ 0.1 µmol/l
GnRH. The differences between treatment groups and dose–responses were
assessed using two-way analysis of variance. Differences between treatments
and controls were tested by Dunnet’s post-hoc test and between treatments
by the Bonferroni–Dunn post-hoc test. The analyses were performed using
the Statview 5 program (Abacus Concepts Inc., Berkley, CA, USA). All
results are presented as means ⫾ SE.
Sequence of the study
d Production, Dyematrex Blue A and PCF purification of BSA-containing
d
d
d
d
G/LCM.
Identification of single major band of interest using anti-GnSIF.
Tryptic digestion and production of internal amino acid sequence.
Production and testing of anti-GnSAF antiserum raised in male rats.
Production, Dyematrex Blue A and PCF purification of serum/BSA-free
G/LCM.
Production of human GnSAF
Figure 3. Pseudochromatofocusing resolution of GnSAF bioactivity detected in blue dye treated, BSA-free, granulosa/luteal cell-conditioned medium
(G/LCM). (a) The thin line shows the protein elution profiles at 280 nm while the thick lines show the pH of the fractions eluted from the column and the
circles (means ⫾ SE) show the effects of the eluted fractions of GnRH-induced LH secretion from rat pituitary cell culture bioassay. GnSAF bioactivity was
eluted in fraction #10 at pH 5.77, (b) presence of GnSAF bioactivity (suppression of GnRH-induced LH secretion, shown by closed circles), but not inhibin
bioactivity (suppression of basal FSH secretion, shown by open circles) in fraction #10. Data represent the mean of quadruplicate determinations from three
different rat pituitary cell culture bioassays. The shaded horizontal bar indicates the mean ⫾ SE range for control GnRH-induced LH secretion.
Results
Production of GnSAF from granulosa/luteal cells and
depletion of serum proteins
Initial cultures in medium containing 0.1% BSA produced conditioned
medium which contained significant GnSAF bioactivity, suppressing
GnRH-induced LH to 58 ⫾ 8% of control at a dose of 25 µl/well
(P ⬍ 0.001, Figure 1a). Subsequently, BSA-free M199 was also
found to be suitable for the production of GnSAF bioactivity with a
dose of 25 µl/well, reducing GnRH-induced LH secretion to 38 ⫾
3% of control (P ⬍ 0.001, Figure 1b). With both types of culture
medium Dyematrex Blue A affinity chromatography depleted the
serum-protein/albumin content of the G/LCM, with GnSAF activity
remaining in the unbound fraction, reducing GnRH-induced LH
secretion to as little as 22 ⫾ 4% of control (Figure 1c,d).
Partial purification of GnSAF secreted by granulosa/luteal
cells
Figure 4. 1D SDS–PAGE protein profiles for BSA-containing granulosa/
luteal cell-conditioned medium (G/LCM) before (lane 1) and after (lane 2)
blue dye treatment and pseudochromatofocusing (PCF). Recognition of
proteins recognized by Western blot, using anti-porcine GnSAF, is shown in
lane 3 (before PCF) and lane 4 (after PCF).
d Identification of three main GnSAF candidate spots on 2D gel by localization
of GnSAF MW in 1D gels and bioactivity in 2D gels, followed by Edman
sequencing and production of N-terminal amino acid sequence.
d Confirmation that the granulosa cells from unstimulated ovaries are the
site of GnSAF production.
Recovery of bioactivity from both BSA-containing and BSA-free
G/LCM following CBX pseudochromatofocusing was similar (Figures
2 and 3). GnSAF bioactivity was recovered at pH 5.74 in fraction
#11 from medium with BSA and at pH 5.77 in fraction #10 from
medium without BSA. This activity was not due to inhibin since
fractions of G/LCM both with (fraction #11) and without BSA
(fraction #10) caused significant suppression of GnRH-induced LH
secretion (to 58 ⫾ 8 and 60 ⫾ 12% of control at 25 µl/well; P ⬍ 0.01
respectively) but not of basal FSH secretion (Figures 2b and 3b).
The major 60–70 kDa band (Figure 4, lane 2) resulting from PCF
827
P.A.Fowler et al.
Figure 5. Dose-dependent GnSAF bioactivity in (a) human follicular fluid, (b) serum from women collected prior to oocyte recovery and (c) BSA-free
granulosa/luteal cell-conditioned medium (G/LCM) is blocked by incubation with a rat anti-GnSAF polyclonal antiserum (open circles), but not following
incubation with a rat polyclonal antiserum raised against serum from post-menopausal women. Data represent the mean of quadruplicate determinations from
two different rat pituitary cell culture bioassays. The shaded horizontal bar indicates the mean ⫾ SE range for control GnRH-induced LH secretion. Data are
shown as means ⫾ SE.
of BSA-containing G/LCM stained well with anti-porcine GnSIF
antiserum (Figure 4, lane 4) although this antiserum also stained
other bands in the crude BSA-containing G/LCM.
Activity of the polyclonal antiserum raised against G/LCM
GnSAF
The rat antiserum raised against the 60–70 kDa band of partially
purified BSA-containing G/LCM (Figure 4, lane 2) was investigated
for its ability to block GnSAF bioactivity in vitro (Figure 5). GnSAF
bioactivity in IVF hFF, IVF pre-hCG serum and BSA-free G/LCM,
which reduced GnRH-induced LH secretion to 58 ⫾ 9, 57 ⫾ 10 and
26 ⫾ 16% of control (P ⬍ 0.001) respectively at 25 µl/well following
incubation with anti-PMWS (which was used as a control antiserum),
had no suppressive effects on GnRH-induced LH secretion following
incubation with anti-GnSAF (GnRH-induced LH secretion was
100 ⫾ 11, 105 ⫾ 6 and 133 ⫾ 9% of control respectively; not
significant).
Localization of GnSAF bioactivity in 1D and 2D gels
Protein-containing and blank 1D SDS–PAGE gel slices from bluedye treated BSA-free G/LCM had very distinct effects in the GnSAF
bioassay. The blank gel slices had little effect while the major band
at ~64 kDa significantly suppressed GnRH-induced LH secretion to
14 ⫾ 5% of control at 5 µl/well (P ⬍ 0.001, Figure 6). The 83 and
12–25 kDa bands only reduced GnRH-induced LH secretion to 43
⫾ 16% (P ⬍ 0.001) and 80 ⫾ 14% (not significant) of control
respectively at 5 µl/well. The cutting of nine squares from 2D gels
is shown in Figure 7a. Of these squares, significant bioactivity was
detected in Z1 and primarily in Z4 (GnRH-induced LH secretion
reduced to 64 ⫾ 6%, P ⬍ 0.05, and to 43 ⫾ 6%, P ⬍ 0.001, of
control respectively; Figure 7b).
Edman sequencing
The major 60–70 kDa band (Figure 4, lane 2) resulting from PCF of
BSA-containing G/LCM was subjected to Edman sequencing but was
found to be N-terminally blocked. The remaining sample was run on
a 1D gel and the 60–70 kDa band was subjected to tryptic digestion
and then yielded an internal amino acid sequence EPQVYVHAP by
Edman sequencing.
The bioactive PCF fraction of BSA-free G/LCM was subjected to
828
2D gel electrophoresis and Western blotted with our crude antiGnSAF polyclonal antiserum (Figure 8). The remaining serum proteins
were visible in the top right-hand quarter of the gel. However,
although the polyclonal antiserum caused very heavy streaking of 1D
gels (not shown), it stained a series of proteins at ~60–66 kDa, pH
4.5–5.0. This was the molecular mass range expected for GnSAF,
but a lower than expected pI, based both on data presented here and
in the literature. However, the three spots were within the region
from which GnSAF bioactivity was predominantly recovered
(Figure 7b). Of the three proteins spots in this region, only spot #2
contained sufficient protein for Edman sequencing (117 fmol). Edman
sequencing produced the N-terminal sequence NH-xVPQGNAGN.
Online database searching (http://www.ncbi.nlm.nih.gov/blast/) of
the Human Genome Protein database yielded no significant sequence
homology with either the N-terminal or internal sequences. The nr
and Swissprot databases yielded some sequence homology. The
internal amino acid sequence (EPQVYVHAP) had the following
sequence homologies: (i) 6 aa with human ornithine decarboxylase
antizyme inhibitor, MW 49 536, pI 4.63; (ii) 5 aa with human
replication control protein 1, MW 97 935, pI 9.34; and (iii) 5 aa with
human IG γ-1 chain c, MW 36 106, pI 8.46. The N-terminal amino
acid sequence (NH-xVPQGNAGN) had the following sequence
homologies: (i) 4 aa with human ubiquitously transcribed Y chromosome tetratricopeptide repeat protein, MW 149 578, pI 7.91; (ii) 4 aa
with human M-phase phosphoprotein 9, MW 24 257, pI 6.99; and
(iii) 4 aa with human mothers against decapentaplegic homologue 7
(SMAD7), MW 46 426, pI 8.63. None of the proteins identified by
on-line searches matched the proposed MW of pI values for GnSAF
bioactivity.
Localization of GnSAF production to granulosa cells from
unstimulated human ovaries
The granulosa, theca and stroma-conditioned BSA-free M199 produced very different results in the rat pituitary GnSAF bioassay. Only
GCM suppressed GnRH-induced LH secretion (Figure 9a), reducing
LH to 67 ⫾ 8% of control (P ⬍ 0.01) in the presence of anti-PMWS.
In contrast, SCM stimulated GnRH-induced LH secretion to 270 ⫾
130% of control (P ⬍ 0.05, at 25 µl/well) and TCM had little effect.
While the anti-GnSAF polyclonal completely blocked in-vitro GnSAF
bioactivity in GCM (P ⬍ 0.001), it had little effect on the response
of the gonadotroph to either TCM or SCM (Figure 9b).
Production of human GnSAF
Figure 6. Localization of GnSAF bioactivity in blue dye treated, BSA-free, granulosa/luteal cell-conditioned medium (G/LCM). (a) GnSAF dose–response
curve and (b) detection of GnSAF in the 64 kDa fraction of a reducing 1D SDS–PAGE gel at 1 and 5 µl/well doses. Data represent the mean of
quadruplicate determinations from two different rat pituitary cell culture bioassays. The shaded horizontal bar indicates the mean ⫾ SE range for control
GnRH-induced LH secretion.
Figure 7. Localization of GnSAF bioactivity in BSA-free, granulosa/luteal cell-conditioned medium (G/LCM). (a) Nine squares, Z1-Z9, cut from four
replicate 2D gels, (b) detection of GnSAF primarily in the Z4 gels square. Data represent the mean of quadruplicate determinations from three different rat
pituitary cell culture bioassays. The shaded horizontal bar indicates the mean ⫾ SE range for control GnRH-induced LH secretion.
829
P.A.Fowler et al.
Discussion
Figure 8. 2D gel of blue dye treated, BSA-free, granulosa/luteal cellconditioned medium (G/LCM) after pseudochromatofocusing (PCF)
showing (a) Coomassie Blue staining of proteins at 60–66 kDa and 4.5–5
pH (labelled 1–3) used for Edman sequencing and (b) proteins recognized
by Western blot using the rat anti-GnSAF polyclonal raised against the
60–70 kDa protein band shown in Figure 4.
This study demonstrates for the first time that human GnSAF is
produced specifically by the granulosa cell from unstimulated ovaries
and that the bioactivity is similar to that produced by IVF granulosa/
luteal cells. The GnSAF bioactivity was associated with proteins of
~64 kDa and a pI of 5.7–5.8 pH.
Our molecular mass and pI findings are in agreement with Danforth’s results (Danforth and Cheng, 1995; Mroueh et al., 1996) but
not those of Tio or Pappa (Tio et al., 1994; Pappa et al., 1999). It is
highly unfortunate that our anti-GnSAF polyclonal antiserum,
although an effective inhibitor of GnSAF bioactivity, did not prove
suitable for further characterization of the elusive GnSAF molecule.
Surprisingly, this antiserum recognized proteins at a lower pI range
in 2D gels than observed using chromatographic techniques. The
reasons for this will require further investigation. With regard to the
findings of Pappa et al. in their study a heat-treatment step was
included, and it may be that this results in cleavage of the GnSAF
molecule without destroying the bioactive moiety (Pappa et al., 1999).
It is interesting to note that we observed a minor reduction in GnRHinduced LH secretion from rat gonadotroph incubated with a 17 kDa
fraction of G/LCM, possibly indicative of limited GnSAF bioactivity.
However, it seems unlikely that the prolonged heat-treatment used
by Pappa et al. would fragment the GnSAF molecule to a greater
extent than the brief boiling the GnSAF received prior to 1D SDS–
PAGE in our hands. The robust nature of GnSAF bioactivity (Fowler
and Templeton, 1996) has unfortunately not assisted significantly in
its purification.
In addition, we also observed some GnSAF bioactivity at 83 kDa.
Over time, estimates of GnSAF molecular mass have varied from
⬍10 to ⬎100 kDa (Fowler and Templeton, 1996). While our
observation of a secondary GnSAF bioactivity peak at 83 kDa is
unlikely to be due to GnSAF binding with other proteins since
reducing conditions were employed, it is possible that GnSAF
undergoes a variety of post-transcriptional modification, and may
have a minor form at that mass. However, it is equally possible that
there is some other protein at this mass that can reduce GnRH-
Figure 9. Dose-dependent GnSAF bioactivity in cell/tissue-conditioned medium following incubation with (a) a rat polyclonal antiserum raised against serum
from post-menopausal women or (b) a rat anti-GnSAF polyclonal antiserum. GnSAF bioactivity was present in granulosa cell-conditioned medium (GCM:
closed circles) and blocked by the anti-GnSAF antiserum. There was no GnSAF bioactivity and no effects of the anti-GnSAF antiserum in media conditioned
by either stroma (SCM: inverted triangles) or theca (TCM: triangles). Data represent the mean of quadruplicate determinations from two different rat pituitary
cell culture bioassays. The shaded horizontal bar indicates the mean ⫾ SE range for control GnRH-induced LH secretion. Data are shown as means ⫾ SE.
830
Production of human GnSAF
induced LH secretion from rat pituitary cells. A final possibility is
that GnSAF was present at a sufficiently low abundance to not show
up following Coomassie Blue staining and was at the border of the
64 and 83 kDa bands, so that when the gels were sliced, some GnSAF
material was allocated to the 83 kDa band instead of the 64 kDa band.
Our results are disparate from those of Tio et al. who also used
conditioned medium, although from the Sertoli rather than granulosa/
luteal cells. Even in their purified material, Tio et al. observed some
inhibin-like bioactivity, since it caused a reduction in basal FSH
secretion from rat pituitary cultures (Tio et al., 1994). This is in
marked contrast to our findings and those of Danforth and Pappa
et al. (Danforth and Cheng, 1995; Mroueh et al., 1996; Pappa et al.,
1999). Combined with the differing molecular masses of these
proteins, the fact that Tio et al. derived their material from the male
rather than the female and continued to observe inhibin-like bioactivity
suggests that the bioactivity may not be of the same entity (Tio
et al., 1994).
The internal and N-terminal amino acid sequences may be part of
the same protein, but neither matches the other published putative
GnSAF sequences (Tio et al., 1994; Danforth and Cheng, 1995;
Pappa et al., 1999). Furthermore, there are no common proteins yet
identified which contain combinations of the five candidate GnSAF
amino acid sequences. However, it is possible that when GnSAF is
finally convincingly sequenced, one or more of these putative
sequences may prove to have been part of the bioactive molecule.
It is not possible at this stage to determine whether the sequences
derived in the current study represent parts of the GnSAF sequence
or of co-purified proteins. This criticism is true of the other published
sequences and can only be addressed by rigorous verification. Such
verification could include immunopurification of homogenous protein
from bioactive preparations (such as hFF) using antibodies against
synthetic peptides based on the amino acid sequences. The production
of antibodies which block or bind GnSAF bioactivity may be due to
recognition of GnSAF-binding or chaperone proteins rather than
recognition of the bioactive molecule itself. The screening of human
ovarian cDNA libraries with antibodies that recognize GnSAF bioactivity and the subsequent expression of a full length gene product
which could then be tested for GnSAF bioactivity is an attractive
alternative verification technique.
The lack of homology between the internal and N-terminal amino
acid sequences presented in this paper with the Human Genome
database suggests that these sequences may be from a novel protein
or proteins. The sequence homologies with proteins in the nr and
Swissprot databases were comparatively poor even though we had
only managed to obtain short sequences. In the case of the N-terminal
sequence, the three closest matches from the databases were: Y
chromosome tetratricopeptide repeat which is involved in X chromosome inactivation (Greenfield et al., 1998), M-phase phosphoprotein 9
which has a role in cell division (Matsimoto-Taniura et al., 1996)
and SMAD7 which transduces transforming growth factor-β signals
from the cell membrane to the nucleus (Wu et al., 2001). In contrast,
the three best matches with the internal amino acid sequence have
roles in cell growth (ornithine decarboxylase antizyme inhibitor)
(Nilsson et al., 2000), initiating cellular cDNA replication (replication
control protein 1) (Gavin and Stillman, 1995) and antigenic responses
(IgG). None of these functions would suggest a likely identity as
GnSAF and, in addition, the molecular masses and pIs of these
proteins do not match the values published by Danforth and those
presented in this study (Danforth and Cheng, 1995; Mroueh et al.,
1996). This supports our suggestion that the two sequences presented
above relate to novel proteins.
One interesting finding from our experiments was that SCM and,
to a lesser extent, TCM markedly stimulated GnRH-induced LH
secretion, the opposite effect to GCM. The fact that anti-GnSAF
antiserum did not significantly inhibit this stimulatory effect suggests
that it has no direct relationship with GnSAF. However, the cause of
the stimulatory effect of GnRH-induced LH secretion remains to be
established. It would also have been interesting to culture granulosa
cells from different sizes of spontaneous cycle follicles separately.
Unfortunately, the time involved in collecting sufficient material
precluded this particular experiment from the present study. Recovery
of GnSAF bioactivity from the spontaneous cycle GCM was superior
to the yield from IVF-derived granulosa/luteal cells. This hampered
the production of sufficient protein for repeat runs of chromatographic
separation. However, the scarcity of spontaneous cycle follicles meant
that we were unable to collect sufficient material from this source
for the GnSAF isolation experiments.
In terms of the physiological role of GnSAF, we have already
shown that in both stimulated and spontaneous cycles, GnSAF
concentration is considerably higher in hFF from small rather than
large follicles (Fowler et al., 1994b, 2001). This observation matches
our data suggesting that GnSAF has higher circulating titres during
the first half of the human follicular phase (Martinez et al., 2002).
Combined with the finding that the granulosa cell is the site of
specific GnSAF bioactivity production, these data support the concept
that GnSAF has a role in maintaining reduced pituitary responsiveness
to GnRH during the follicular phase, interacting with ovarian steroid
and signalling systems at the level of the pituitary gland (Byrne et al.,
1996; Fowler and Templeton, 1996; Tijssen et al., 1997; de Koning
et al., 2001).
In conclusion, we have demonstrated that BSA-free medium
conditioned by human granulosa/luteal cells is a suitable source of
GnSAF bioactivity with far fewer contaminating proteins, especially
serum albumin, than follicular fluid or serum. This conditioned
medium clearly contains GnSAF bioactivity with a molecular mass
between 60 and 70 kDa and a pI between 5.7 and 5.8. The two amino
acid sequences in this paper remain candidate sequences for GnSAF
and require further validation.
Acknowledgements
We are grateful to Dr D.R.Danforth, Department of Obstetrics and Gynaecology, Ohio State University, Columbus, Ohio, USA, for supplying some of
his GnSIF antiserum. We thank Mrs M.Fraser, P.Cunningham and E.Argo for
their expert technical assistance. We thank the staff at the Biological Services
Unit (University of Aberdeen) for maintaining the rats used in this study and
Dr A.F.Parlow at NIDDK’s National Hormone and Pituitary Program (Torrance,
California, USA) and SAPU (Carluke Hospital, Scotland) for hormone assay
materials. We are grateful to the BBSRC, MRC and Wellcome Trust for their
financial support.
References
Burger, H.G., Groome, N.P. and Robertson, D.M. (1998) Both inhibin A and
B respond to exogenous follicle-stimulating hormone in the follicular phase
of the human menstrual cycle. J. Clin. Endocrinol. Metab., 83, 4167–4169.
Byrne, B., Fowler, P.A., Fraser, M., Culler, M. and Templeton, A. (1995)
GnSAF bioactivity in serum from superovulated women is not blocked by
inhibin antibody. Biol. Reprod., 52, 88–95.
Byrne, B., Fowler, P.A. and Templeton, A. (1996) Role of progesterone and
non-steroidal ovarian factors in regulating gonadotrophin-releasing hormone
self-priming in vitro. J. Clin. Endocrinol. Metab., 81, 1454–1459.
Cash, P. (1989) Analysis of virus protein heterogeneity among group B
coxsackie viruses using a ‘mini’ two-dimensional gel electrophoresis
system. Electrophoresis, 10, 793–800.
Cash, P., Argo, E. and Bruce, K.D. (1995) Characterisation of Haemophilus
influenzae proteins by 2-dimensional gel electrophoresis. Electrophoresis,
16, 135–148.
831
P.A.Fowler et al.
Cash, P., Argo, E., Langford, P.R. and Kroll, J.S. (1997) Development of
an Haemophilus two-dimensional protein database. Electrophoresis, 18,
1472–1482.
Cash, P., Argo, E., Ford, L., Lawrie, L. and McKenzie, H. (1999) A
proteomic analysis of erythromycin resistance in Streptococcus pneumoniae.
Electrophoresis, 20, 2259–2268.
Christie, J.F., Dunbar, B. and Kennedy, M.W. (1993) The ABA-1 allergen of
the nematode Ascaris suum: epitope stability, mass spectrometry, and Nterminal sequence comparison with its homologue in Toxocara canis. Clin.
Exp. Immunol., 92, 125–132.
Danforth, D.R. and Cheng, C.Y. (1995) Purification of a candidate
gonadotrophin surge inhibiting factor from porcine follicular fluid.
Endocrinology, 136, 1658–1665.
Danforth, D.R., Sinosich, M.J., Anderson, T.L., Cheng, C.Y., Bardin, C.W.
and Hodgen, G.D. (1987) Identification of gonadotrophin surge-inhibiting
factor (GnSIF) in follicular fluid and its differentiation from inhibin. Biol.
Reprod., 37, 1075–1082.
De Koning, J., Lambalk, C.B., Helmerhorst, F.M. and Helder, M.N. (2001) Is
GnRH self-priming an obligatory feature of the reproductive cycle? Hum.
Reprod., 16, 209–214.
Dunbar, B. and Wilson, S.B. (1994) A buffer exchange procedure giving
enhanced resolution to polyacrylamide gels pre-run for protein sequencing.
Ann. Biochem., 216, 227–228.
Fowler, P.A. and Templeton, A. (1996) The nature and function of putative
gonadotrophin surge-attenuating/inhibiting factor (GnSAF/IF). Endocr. Rev.,
17, 103–120.
Fowler, P.A. and Price, C. (1997) Follicle-stimulating hormone stimulates
circulating gonadotrophin surge-attenuating/inhibiting factor bioactivity in
cows. Biol. Reprod., 57, 278–285.
Fowler, P.A., Cunningham, P., Fraser, M., McGregor, F., Byrne, B., Pappa,
A., Messinis, I.E. and Templeton, A. (1994a) Circulating gonadotrophin
surge-attenuating factor from superovulated women suppresses in-vitro
gonadotrophin releasing-hormone self-priming. J. Endocrinol., 143,
45–54.
Fowler, P.A., Fraser, M., Cunningham, P., Knight, P.G., Byrne, B., McLaughlin,
E., Wardle, P.G., Hull, M.G.R. and Templeton, A. (1994b) Higher
gonadotrophin surge-attenuating factor (GnSAF) bioactivity is found
in small follicles from superovulated women. J. Endocrinol., 143,
33–44.
Fowler, P.A., Fahy, U., Culler, M.D., Knight, P.G., Wardle, P.G., McLaughlin,
E.A., Cunningham, P., Fraser, M., Hull, M.G.R. and Templeton, A. (1995)
GnSAF bioactivity is present in follicular fluid from naturally cycling
women. Hum. Reprod., 10, 68–74.
Fowler, P.A, Sorsa, T., Harris, W.T., Knight, P.G. and Mason, H.D. (2001)
Relationship between follicle size and gonadotrophin surge-attenuating
factor (GnSAF) bioactivity during spontaneous cycles in women. Hum.
Reprod., 16, 1353–1358.
Gavin, K.A. and Stillman, B. (1995) Conserved initiator proteins in eukaryotes.
Science, 270, 1667–1671.
Greenfield, A., Carrel, L., Pennisi, D., Philippe, C., Quaderi, N., Siggers, P.,
Steiner, K., Tam, P.P.L., Monaco, A.P., Willard, H.F. and Koopman, P.
(1998) The UTX gene escapes X inactivation in mice and humans. Hum.
Mol. Genet., 7, 737–742.
Koppenaal, D.W., Tijssen, A.M.I. and de Koning, J. (1992) The effect
of gonadotrophin surge-inhibiting factor on the self-priming action of
gonadotrophin-releasing hormone in female rats in vitro. J. Endocrinol.,
134, 427–436.
Koppenaal, D.W., van Dieten, J.A.M.J., Tijssen, A.M.I. and de Koning, J.
(1993) Induction of the gonadotrophin surge-inhibiting factor by FSH and
its elimination: a sex difference in the efficacy of the priming effect of
gonadotrophin-releasing hormone on the rat pituitary gland. J. Endocrinol.,
138, 191–201.
Laemmli, U.K. (1970) Cleavage of structural proteins during assembly of the
head of bacteriophage T4. Nature, 22, 680.
Littman, B.A. and Hodgen, G.D. (1984) Human menopausal gonadotrophin
stimulation in monkeys: blockade of the luteinising hormone surge by a
highly transient ovarian factor. Fertil. Steril., 41, 440–447.
832
Lui, M., Tempst, P. and Erdjument-Bromage, H. (1996) Methodological
analysis of protein–nitrocellulose interactions to design a refined digestion
protocol. Ann. Biochem., 241, 156–166.
Martinez, F., Knight, P.G., Groome, N.P., Barri, P.N., Coroleu, V., Tur, R. and
Fowler, P.A. (2002) Circulating gonadotrophin surge-attenuating factor
(GnSAF), inhibin-A and inhibin-B are reduced in women with previous
spontaneous luteinisation. Hum. Reprod., 17, 634–640.
Mason, H.D., Morgan, R., Winston, R.M.L., Beard, R.W., Reed, M.J. and
Franks, S. (1990) Inhibition of oestradiol production by epidermal growth
factor in granulosa cells of normal and polycystic ovaries. J. Clin.
Endocrinol. Metab., 33, 511–517.
Matsimoto-Taniura, N., Pirollet, F., Monroe, R., Gerace, L. and Westendorf,
J.M. (1996) Identification of novel M phase phosphoproteins by expression
cloning. Mol. Cell. Biol., 7, 1455–1469.
Messinis, I.E. and Templeton, A. (1986) The effect of pulsatile follicle
stimulating hormone on endogenous luteinizing hormone surge in women.
Clin. Endocrinol., 25, 633–640.
Messinis, I.E. and Templeton, A. (1991) Attenuation of gonadotrophin release
and reserve in superovulated women by gonadotrophin surge attenuating
factor. Clin. Endocrinol., 34, 259–263.
Messinis, I.E., Hirsch, P. and Templeton, A.A. (1991) Follicle stimulating
hormone stimulates the production of gonadotrophin surge attenuating
factor (GnSAF) in vivo. Clin. Endocrinol., 35, 403–407.
Messinis, I.E., Lolis, D., Papadopoulos, L., Tsahalina, T., Papanikolaou, N.,
Seferiadis, K. and Templeton, A.A. (1993) Effect of varying concentrations
of follicle stimulating hormone on the production of gonadotrophin surge
attenuating factor (GnSAF) in women. Clin. Endocrinol., 39, 45–50.
Messinis, I.E., Lolis, D., Zikopoulos, K., Tsahalina, E., Seferiadis, K. and
Templeton, A.A. (1994a) Effect of an increase in FSH on the production
of gonadotrophin surge-attenuating factor in women. J. Reprod. Fertil.,
101, 689–695.
Messinis, I.E., Lolis, D., Papastergiopoulou, L., Milingos, S., Tsahalina, E.,
Seferiadis, K. and Templeton, A.A. (1994b) Effect of follicle stimulating
hormone treatment on the pituitary response to luteinizing hormonereleasing hormone in post-menopausal women. Hum. Reprod., 9, 241–244.
Mroueh, J.M., Arbogast, L.K., Fowler, P.A., Templeton, A., Friedman, C.I.
and Danforth, D.R. (1996) Identification of gonadotrophin surge-inhibiting
factor (GnSIF)/attenuin in human follicular fluid. Hum. Reprod., 11,
101–107.
Nilsson, J., Grahn, B. and Heby, O. (2000) Antizyme inhibitor is rapidly
induced in growth-stimulated mouse fibroblasts and releases ornithine
decarboxylase from antizyme suppression. Biochem. J., 346, 699–704.
Pappa, A., Seferiadis, K., Fotsis, T., Shevchenko, A., Marselos, M., Tsolas,
O. and Messinis, I.E. (1999) Identification of a candidate gonadotrophin
surge attenuating factor from human follicular fluid. Hum. Reprod., 14,
1449–1456.
Rabilloud, T., Valette, C. and Lawrence, J.J. (1994) Sample application by ingel rehydration improves the resolution of two-dimensional electrophoresis
with immobilized pH gradients in the first dimension. Electrophoresis, 15,
1552–1558.
Sopelak, V.M. and Hodgen, G.D. (1984) Blockade of the estrogen-induced
luteinizing hormone surge in monkeys: a nonsteroidal, antigenic factor in
porcine follicular fluid. Fertil. Steril., 41, 108–113.
Tijssen, A.M.I., Helder, M.N., Chu, Z.W. and de Koning, J. (1997) Intracellular
antagonistic interaction between GnRH and gonadotrophin surge-inhibiting/
attenuating factor bioactivity downstream of second messengers involved
in the self-priming process. J. Reprod. Fertil., 111, 235–242.
Tio, S., Koppenaal, D., Bardin, C.W. and Cheng, C.Y. (1994) Purification of
gonadotrophin surge-inhibiting factor from Sertoli cell-enriched culture
medium. Biochem. Biophys Res. Commun., 199, 1229–1236.
Votyakova, T.V., Wallace, H.M., Dunbar, B. and Wilson, S.B. (1999) The
covalent attachment of polyamines to proteins in plant mitochondria. Eur.
J. Biochem., 260, 250–257.
Wu, D., Luo, S., Wang, Y., Zhuang, L., Chen, Y. and Peng, C. (2001) Smads
in human trophoblast cells: expression, regulation and role in TGF-βinduced transcriptional activity. Mol. Cell. Endocrinol., 175, 111–121.
Submitted on January 11, 2002; resubmitted on April 8, 2002; accepted on
April 23, 2002