The recombinant subdomain IIIB of human serum albumin displays activity of gonadotrophin surge-attenuating factor

Sotiria Tavoulari1, Stathis Frillingos1,3, Panayiota Karatza1, Ioannis E. Messinis2 and Konstantin Seferiadis1

1 Laboratory of Biological Chemistry, University of Ioannina Medical School, Ioannina 45110 and 2 Department of Obstetrics and Gynecology, University of Thessalia Medical School, Larissa 41222, Greece

3 To whom correspondence should be addressed. e-mail: efriligo{at}cc.uoi.gr


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: Gonadotrophin surge-attenuating factor (GnSAF) is an as yet unidentified ovarian factor that acts on the pituitary to attenuate the pre-ovulatory LH surge. In a previous study, GnSAF bioactivity was proposed to derive, at least in part, from a C-terminal domain (95peptide) of human serum albumin (HSA). METHODS AND RESULTS: We employ here the expression–secretion system of Pichia pastoris to produce and assay selected recombinant polypeptides of HSA for GnSAF activity. We show that the C-terminal 95peptide of HSA (residues 490–585; subdomain IIIB) can be expressed from P.pastoris in secreted form and supernatants from clones expressing this polypeptide reduce the GnRH-induced LH secretion of primary rat pituitary cultures by 50–82%. When expressed in the same system, HSA domain III (residues 381–585) or full-length HSA (residues 1–585) are inactive. The bioactive subdomain IIIB is also separable from either domain III or full-length HSA on Blue Sepharose chromatography. CONCLUSIONS: Taken together, the findings highlight the putative importance of HSA subdomain IIIB as a GnSAF-bioactive entity and introduce a unique experimental tool to engineer this molecule for structure–function analysis.

Key words: albumin/GnRH/GnSAF/LH/Pichia pastoris


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Several lines of functional evidence have been presented in the literature that the ovary produces a non-steroidal factor, named gonadotrophin surge-attenuating factor (GnSAF) (Messinis and Templeton, 1989Go) or gonadotrophin surge-inhibiting factor (GnSIF) (Sopelak and Hodgen, 1984Go), that acts on the pituitary gonadotrophs. This factor plays an important role in maintaining a low LH concentration during the follicular phase of the ovarian cycle by reducing the pituitary cell responsiveness to GnRH without affecting basal gonadotrophin secretion (Messinis and Templeton, 1991Go; Koppenaal et al., 1992Go; Tijssen et al., 1997Go). GnSAF is distinct from inhibin (Ying, 1988Go), a gonadal protein factor which acts primarily by suppressing the basal secretion of FSH (Fowler et al., 1990Go; Byrne et al., 1995Go; Mroueh et al., 1996Go; Pappa et al., 1999aGo,b). GnSAF bioactivity has been demonstrated, by monitoring changes in the GnRH-induced LH secretion levels of rat pituitary primary cell cultures, in the ovarian follicular fluid (FF) of human (Busbridge et al., 1990Go; Fowler et al., 1990Go; Knight et al., 1990Go; Mroueh et al., 1996Go) and other mammalian species (Schenken and Hodgen, 1986Go; Danforth et al., 1987Go; Busbridge et al., 1988Go; Danforth and Cheng, 1995Go), as well as in culture supernatants from human ovarian granulosa cells (Fowler et al., 2002Go).

Several attempts have been made to purify and identify the bioactive GnSAF molecule(s), and five putative GnSAF/IF sequences have been proposed (Tio et al., 1994Go; Danforth and Cheng, 1995Go; Pappa et al., 1999aGo; Fowler et al., 2002Go). The five sequences have no apparent consensus and refer to proteins ranging in size from 12 to 69 kDa. Among them, only one displays substantial homology to a known portion of the human genome (Pappa et al., 1999aGo); this sequence has been identified as a 12.5 kDa protein corresponding to the C-terminal 95peptide of human serum albumin (HSA), raising the exciting possibility that a bioactive C-terminal domain of HSA may participate in the regulation of LH secretion (Pappa et al., 1999aGo). In the progress of the present study, we have employed in silico analysis and found out that this 95peptide corresponds to HSA subdomain IIIB (He and Carter, 1992Go); therefore, we refer to this sequence with the alternative name subdomain IIIB or DIIIB throughout this manuscript.

To date, none of the five putative GnSAF/IF sequences has been confirmed as pertinent to a genuine GnSAF molecule, and the possibility that other, unrelated co-purifying proteins may contribute to the above findings cannot be ruled out in each case. In general, the relevant types of isolation studies have major drawbacks related to the low concentration of GnSAF in biological fluids, high bioactivity and a background of large numbers and abundance of other proteins. To overcome such problems and provide more clear evidence, we have produced the C-terminal 95peptide of HSA in recombinant form using the heterologous expression–secretion system of the methylotrophic yeast Pichia pastoris and subjected it to GnSAF bioassay. In parallel experiments, we have analysed the full-length HSA molecule or other domains of HSA, engineered and produced in recombinant form in the same system. Pichia pastoris was selected because it is a low-complexity expression–secretion system that has also been employed in the past to study domains of HSA (Dockal et al., 1999Go).

Our data show that, consistent with the proposal of Pappa et al. (1999aGo), the C-terminal polypeptide of HSA (subdomain IIIB) possesses specific GnSAF bioactivity, that is not reproducible with larger domains of HSA or intact HSA itself.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Chemicals
Oligodeoxynucleotides were synthesized by BioSpring GmbH (Frankfurt, Germany). Restriction endonucleases were purchased from New England Biolabs (Beverly, MA), Pyrococcus furiosus (Pfu) DNA polymerase from Stratagene (La Jolla, CA) and high fidelity Taq polymerase from Roche Molecular Biochemicals (Manheim, Germany). Rat LH (rLH-RP-3, rLH-I-10), rat FSH (rFSH-RP-2, rFSH-I-9) and antisera (anti-rLH-S11, anti-rFSH-S-11) were provided by Dr Parlow of the National Institute of Diabetes and Digestive and Kidney Diseases through the National Hormones and Pituitary Program (Baltimore, MD). GnRH and HSA (catalogue number A-3782) were obtained from Sigma Chemical Co. (St Louis, MO). Anti-myc monoclonal antibody 9E10 was purified from the hybridoma using standard techniques and obtained from Dr Carol Murphy. Anti-HSA polyclonal antibody was purchased from Abcam (Cambridge, UK). Anti-mouse and anti-rabbit IgG–horseradish peroxidase (HRP) conjugates were from Amersham Pharmacia Biotech (London, UK). Blue Sepharose 6 Fast Flow was also from Amersham Pharmacia Biotech.

Bacterial and yeast strains and plasmids
Escherichia coli TOP10F' [F'{proAB, lacIq, lacZ{Delta}M15, Tn10(TetR)}mcrA,{Delta}(mrr-hsdRMS-mcrBC), {phi}80lacZÄM15, {Delta}lac X74, deoR, recA1, araD139, {Delta}(ara-leu)7697, galU, galK, rpsL(StrR), endA1, nupG{lambda}] was used for initial propagation and verification of the recombinant plasmids. Pichia pastoris GS115 [his4] (Invitrogen, Groningen, The Netherlands) was used as a host strain for the heterologous expression of HSA domains. Strain GS115/albumin (Invitrogen), having the full-length cDNA of HSA incorporated into the genome under control of a methanol-inducible promoter (AOX1 locus), was used as a control for secreted expression and production of recombinant HSA. Novel GS115 strains harbouring specified domains of HSA were engineered in the course of this study. Plasmid pPICZaA (Invitrogen) was used for cloning and for expression–secretion of the selected recombinant sequences in the P.pastoris system, in each case. Selection of GS115 harbouring pPICZaA-based episomes was carried out on growth media containing zeocin (0.1 mg/ml).

Design and construction of recombinant P.pastoris strains
DNA encoding selected HSA domains was amplified by PCR using Pfu polymerase and chromosomal DNA of GS115/albumin as a template. The sense and antisense primers used had been designed to carry non-annealing overhangs at their 5' end containing an EcoRI and a NotI restriction site, respectively (see Table I). The pair of primers DIIIB-(EcoRI)-sense and (NotI)-antisense or (NotI)-His6-antisense was used to amplify the sequence of HSA codons 490–585 (corresponding to subdomain IIIB); the pair of primers DIII-(EcoRI)-sense and (NotI)-antisense or (NotI)-His6-antisense was used to amplify the sequence of HSA codons 381–585 (domain III). In both cases, the (NotI)-His6-antisense primer was employed to engineer a C-terminal sequence tag of ~2.5 kDa containing the c-myc epitope (Glu-Gln-Lys-Leu-Ile-Ser-Glu-Glu-Asp-Leu-Asn) and a His6 tail, for optimal detection and/or purification of the recombinant polypeptides. Two additional constructs were made by using primers DI-(EcoRI)-sense and DI-(NotI)-antisense to amplify the sequence of HSA codons 1–197 (domain I), or primers DII-(EcoRI)-sense and DII-(NotI)-antisense to amplify the sequence of HSA codons 189–385 (domain II). Due to the engineered EcoRI site (Table I), a short extension by two amino acid residues (Glu-Phe) is introduced at the N-terminus of the processed recombinant polypeptides, in all cases. The PCR products were digested with EcoRI and NotI and ligated to a similarly treated vector pPICZaA (Invitrogen); this subcloning step places the desired HSA sequences at the AOX1 locus to allow efficient incorporation into the P.pastoris genome and methanol-inducible expression/secretion of the products.


View this table:
[in this window]
[in a new window]
 
Table I. Sequences of the primers used for cloning of HSA domains
 
Initial propagation of the recombinant plasmids for large-scale DNA preparation was carried out in E.coli TOP10F' transformed according to Inoue et al. (1990Go). The DNA sequence was verified in an automated DNA sequencer (MWG-Biotech, Ebersberg, Germany). The recombinant plasmid DNA (3 µg) was linearized at the PmeI restriction site of the AOX1 locus and transferred into P.pastoris GS115 with chemical transformation following the manufacturer’s protocol (Invitrogen). Cells were plated on YPDS-zeocin selection media [1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose, 1 mol/l sorbitol, 0.1 mg/ml zeocin] and incubated for 2–4 days at 30°C. Incorporation of the corresponding HSA sequence into the P.pastoris genome was confirmed by direct PCR screening (Linder et al., 1996Go).

Expression of recombinant peptides
All incubations were carried out at 30°C on an orbital shaker at 250 r.p.m. Recombinant P.pastoris clones were grown on buffered glycerol-complex medium [BMGY: 1% (w/v) yeast extract, 2% (w/v) peptone, 0.1 mol/l KPi pH 6.0, 1.34% (w/v) yeast nitrogen base, 4 x 10–5% (w/v) biotin, 1% (v/v) glycerol] for 20 h, normalized to an OD600 of 1.0 and transferred to buffered methanol complex medium [BMMY: 1% (w/v) yeast extract, 2% (w/v) peptone, 0.1 mol/l KPi pH 6.0, 1.34% (w/v) yeast nitrogen base, 4 x 10–5% (w/v) biotin, 0.5% (v/v) methanol] for induction of recombinant coding sequences through the AOX1 promoter. Growth on BMMY was maintained for an additional 72 h. Subsequently, culture supernatants were collected, filter sterilized, snap-frozen in liquid nitrogen and stored at –80°C until use. Pelleted cells were also harvested for total protein determination.

Immunoblot analysis
Culture supernatants from clones expressing recombinant domains of HSA were analysed by 15% or 18% SDS–PAGE (Laemmli, 1970Go). Visualization of protein bands was performed with both Coomassie brilliant blue and silver staining. Proteins were electroblotted to polyvinylidene difluoride membranes (BioTrace-PVDF; Pall Corporation, Ann Arbor, MI) and probed with monoclonal anti-myc 9E10 antibody (at a dilution of 1:2000) or polyclonal anti-HSA antibody (at a dilution of 1:15 000). Immunoreactive bands were detected with enhanced chemiluminescence (Amersham Pharmacia Biotech) following incubation with anti-mouse IgG– or anti-rabbit IgG–HRP conjugate, respectively (at a dilution of 1:5000 in either case).

Rat pituitary cell cultures
Pituitaries were excised from 2-month-old female Wistar rats that had been sacrificed by stunning and cervical dislocation. The anterior pituitary tissue was separated and used to prepare primary cell cultures, as described (Pappa et al., 1999aGo,b) with minor modifications. Anterior pituitary cells (200 000 per well) were spread on 24-well tissue culture plates. Cells were cultured for 48 h in SFDM (Dulbecco’s modified Eagle’s medium nutrient mixture F-12 Ham supplemented with10 µg/ml insulin, 5 µg/ml transferrin, 100 U/ml penicillin, 100 mg/ml streptomycin, 10% (v/v) fetal calf serum (FCS) and 2.5 mmol/l L-glutamine) and then subjected to GnSAF bioassay.

GnSAF bioassay
Primary pituitary cells pre-incubated for 48 h in SFDM were transferred to SFDM without FCS (SFDMminus) (0.5 ml per well) for another 48 h. During this second period, cells were incubated in triplicate in the presence of test substances (25 µl per well, unless otherwise indicated); at least six control wells receiving SFDMminus without other additions were included in all culture plates. Subsequently, cells were washed with medium and incubated with SFDMminus containing GnRH (0.1 µmol/l) together with test substances (25 µl per well, unless otherwise indicated), for another 4 h. During this final period, three of the control wells were incubated with GnRH alone to measure optimal GnRH-induced LH levels and three of them received SFDMminus without additions to measure basal LH secretion. Cell-conditioned media were collected and stored at –20°C until assay. Concentrations of LH in the culture media were determined with competitive enzyme-linked immunosorbent assay (ELISA) as described (Pappa et al., 1999bGo) using hormone preparations NIDDK-rLH-RP-3 and rLH-I-10 and antiserum NIDDK-anti-rLH-S-11. The useful range of the assay was from 0.5 to 50 ng/ml (NIDDK-rLH-RP-3). Intra- and inter-assay coefficients of variation were, for basal secretion, 11 and 8%, respectively, and for GnRH-induced LH secretion, 11 and 16%, respectively.

GnSAF bioactivity is calculated according to the formula [1 – A/B] (%) where A = LH(ng/ml)[GnRH + SAMPLE] – LH(ng/ml)BASAL, B = LH(ng/ml)GnRH – LH(ng/ml)BASAL, LH(ng/ml)BASAL = the basal secretion level of LH (3.0 ± 0.4 ng LH per well, n = 12), LH(ng/ml)GnRH = the GnRH-induced LH secretion (46.5 ± 8.5 ng LH per well, n = 12) and LH(ng/ml)[GnRH + SAMPLE] = the GnRH-induced LH secretion attained in the presence of the sample tested.

In each calculation, the values LH(ng/ml)BASAL, LH(ng/ml)GnRH and LH(ng/ml)[GnRH + SAMPLE] are determined from wells of the same culture dish.

Bioactivity-blocking effect
To determine whether anti-HSA polyclonal antibody blocks GnSAF activity of IIIB polypeptide, supernatants from GS115/DIIIB or control supernatants (25 µl) were pre-incubated on an orbital shaker with anti-HSA (Abcam) (10 µl; at the desired dilution) in SFDMminus (final volume, 0.5 ml) at 37°C for 5 min. These samples subsequently were subjected to GnSAF bioassay, as described above.

Effects on basal LH and basal FSH levels
To assess the effect of P.pastoris supernatants on basal LH and basal FSH secretion, pituitary cell-conditioned media collected after 48 h of incubation in the absence or presence of test substances were stored at –20°C and used to determine concentrations of LH and FSH. Concentrations of FSH were determined with competitive ELISA (Pappa et al., 1999bGo) using hormone preparations NIDDK-rFSH-RP-2 and rFSH-I-9 and antiserum NIDDK-anti-rFSH-S-11. The useful range of the assay was from 1.25 to 40 ng/ml (NIDDK-rFSH-RP-2). The basal secretion level of FSH after 48 h was 10.9 ± 3.8 ng FSH per well (n = 6). Concentrations of LH were determined with competitive ELISA, as described above. The basal secretion level of LH after 48 h was 13.7 ± 1.3 ng LH per well (n = 6).

Blue Sepharose chromatography
Pichia pastoris culture supernatants were harvested by centrifugation, filtered and dialysed against 100 vols of binding buffer (20 mmol/l NaPi pH 7.5, 0.15 mol/l NaCl) with three changes. Dialysed samples (1 ml) were incubated with Blue Sepharose 6 Fast Flow (gel volume, 50 µl) pre-equilibrated in binding buffer. After washing with 30 vols of binding buffer, elution was performed twice with a 100 µl volume of 20 mmol/l NaPi pH 7.5, 2 mol/l NaCl. All of the above steps were performed at 4°C. Samples subjected to chromatography, flow-through fractions and eluting protein fractions were analysed by SDS–PAGE (18%) followed by silver staining.

Protein determinations
Protein concentrations were determined with the bicinchoninic acid (BCA) method (Pierce, Rockford, IL) using bovine serum albumin as a standard. Assessment of total intracellular protein in cell lysates of P.pastoris (yielding 0.57–0.65 µg of extracted protein per ml of cell culture extract, under the conditions employed) was used to normalize the volumes of supernatants analysed by SDS–PAGE and assayed for GnSAF activity. The P.pastoris lysates were prepared using glass beads (0.5 mm diameter) and a Mini-BeadBeater cell disrupter (Biospec Products, Bartlesvill, OK).

Bionformatics
To estimate domain boundaries, the full-length HSA sequence was subjected to in silico analysis using CATH (Class-Architecture-Topology-Homologous superfamily) (Pearl et al., 2000Go).

Statistical analysis
Differences between treatments and controls were tested by Student’s t-test. Results are presented as means ± SEM.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Production of the sequence 490–585 of HSA in recombinant form
It has been reported recently (Pappa et al., 1999aGo) that a C-terminal 95peptide of HSA (residues 490–585) can be isolated from the FF of superovulated women as a GnSAF-bioactive entity. To analyse this finding further and establish whether GnSAF bioactivity of this HSA peptide is specific, we undertook to produce it in recombinant form from the expression–secretion system of P.pastoris. The cDNA sequence corresponding to residues 490–585 of HSA (subdomain IIIB) was mobilized from GS115/albumin and subcloned in vector pPICZaA. Two versions were engineered, one with a stop codon introduced immediately downstream of the HSA coding sequence (DIIIB) and one tagged at the C-terminus with the myc epitope and a His6 tail (DIIIB-myc-His6). After verification of the sequence, recombinant DNA was used to transform P.pastoris GS115, several zeocin-resistant clones were selected and incorporation of the recombinant sequence into the yeast genome was verified with direct PCR (data not shown).

The transformed P.pastoris clones were grown on the appropriate medium for induction with methanol. After induction of expression, supernatants were analysed by SDS–PAGE (15 or 18%) followed by silver staining and immunoblotting. Silver staining of the electrophoresed proteins revealed the presence of major protein bands at the approximate sizes expected for DIIIB (12 kDa) and DIIIB-myc-His6 (14 kDa), respectively (Figure 1A). Clearly, these bands comprise the vast majority of total protein in the medium, while they are absent from equivalent electropherograms of P.pastoris transformed with empty vector alone (Figure 1A). Immunoreactivity of these protein bands with anti-HSA antibody (for either DIIIB or DIIIB-myc-His6 versions) (Figure 1B) and/or anti-myc antibody 9E10 (not shown) identifies the proteins as corresponding to the designed recombinant polypeptides of HSA (DIIIB and DIIIB-myc-His6, respectively). The heterologous expression of these polypeptides is not accompanied by over-representation of other bands corresponding to native secreted proteins of P.pastoris, as evidenced from the silver staining picture (Figure 1A).



View larger version (56K):
[in this window]
[in a new window]
 
Figure 1. Production of recombinant polypeptide 490–585 of HSA (subdomain IIIB) in P.pastoris GS115. Supernatants (0.1 ml) from representative induced clones of GS115/DIIIB (lane 1), GS115/DIIIB-myc-6His (lanes 2, 4 and 5) or GS115/empty vector (lanes 3 and 6) were subjected to SDS–PAGE (18%) and analysed by silver staining (A) and anti-HSA immunoblotting (B). Pre-stained molecular weight standards (Bio-Rad, low range) and unstained Precision protein standards (Bio-Rad, broad range) are shown on the right of (A) (upper and lower part, respectively).

 
The recombinant polypeptide 490–585 of HSA displays GnSAF bioactivity
Supernatants from P.pastoris clones were assayed for GnSAF bioactivity on primary cultures of rat pituitary cells. Aliquots of 25 µl, corresponding to an intracellular protein content of 0.57–0.65 µg/ml culture, were assayed in all cases, including both GS115/DIIIB or GS115/DIIIB-myc-His6 and GS115 control clones. Supernatants from five individual clones expressing recombinant polypeptide 490–585 of HSA (DIIIB-1, DIIIB-2, DIIIB-3, DIIIB-myc-His6-1 and DIIIB-myc-His6-2) were tested; all of them were found to reduce GnRH-induced LH secretion to a highly significant extent, independently of the presence or absence of the C-terminal tag (Figure 2 and Table II). Maximal inhibition was observed with clones GS115/DIIIB-1 and GS115/DIIIB-myc-His6-1 which reduced GnRH-induced LH secretion to 23 ± 8% (n = 6; P < 0.0001) and 30 ± 9% (n = 3; P < 0.01), corresponding to GnSAF activities of ~82 and 77%, respectively. The remaining three clones reduced GnRH-induced LH secretion to levels ranging from 51 to 34% (with respective GnSAF activities ranging from 50 to 68%). The differences in activities observed with different clones might be related to different levels of expression of DIIIB in the clone supernatants. In all cases, control supernatants from GS115 transformed with vector pPICZaA alone were assayed in parallel and found to contain negligible attenuation activity (Figure 2 and Table II).



View larger version (27K):
[in this window]
[in a new window]
 
Figure 2. GnSAF activity of supernatants of P.pastoris expressing the recombinant polypeptide 490–585 of HSA (subdomain IIIB) in its untagged (A) or tagged version (B). Culture supernatants (25 µl/well) of induced GS115/empty vector (A and B), GS115/DIIIB-1, GS115/DIIIB-2, GS115/DIIIB-3 (A), GS115/DIIIB-myc-His6-1 or GS115/DIIIB-myc-His6-2 (B), as indicated, were filter sterilized and assayed for GnSAF bioactivity, as described in Materials and methods. The effects on the GnRH-induced LH secretion of rat pituitary primary cells are presented as percentages of secreted LH relative to LH secreted from control cells incubated with 0.1 µmol/l of GnRH alone, as measured with competitive ELISA. Each bar represents mean values ± SEM of triplicate determinations. Values significantly lower than control (P < 0.01) are denoted by an asterisk. Cross-hatched horizontal bars denote the mean ± SEM range for control GnRH-induced LH secretion. Interrupted lines indicate basal LH secretion elevated by an increment of three SEM values.

 

View this table:
[in this window]
[in a new window]
 
Table II. GnSAF activities of recombinant P.pastoris clones
 
Polypeptide 490–585 of HSA does not affect basal LH or basal FSH secretion
Dose–response analysis of GS115/DIIIB supernatants in the range of 2.5–25 µl (Figure 3A) shows that the attenuation activity on the GnRH-induced LH secretion is evident from doses of at least 15 µl per well. In the same range (5–25 µl), GS115/DIIIB supernatants do not affect the basal LH secretion levels (Figure 3C) or the basal FSH secretion levels (Figure 3D).



View larger version (29K):
[in this window]
[in a new window]
 
Figure 3. Analysis of the GnSAF activity of HSA DIIIB. (A) Dose–response curve of the effect on GnRH-induced LH secretion. (B) Blockade of the activity by anti-HSA antibodies. (C) Lack of effect on basal LH secretion. (D) Lack of effect on basal FSH secretion. Supernatants from induced GS115/DIIIB were assayed at doses of 2.5–25 µl per well, as indicated (A, C and D). Data are presented as mean ± SEM of triplicate determinations. The shaded horizontal bars indicate the mean ± SEM range for control GnRH-induced LH (A), basal LH (C) or basal FSH secretion (D). Supernatants from induced GS115/DIIIB (25 µl per well) were also assayed for their effect on GnRH-induced LH secretion after pre-incubation in the absence or presence of anti-HSA polyclonal antibody (1 µg per well), as described in Materials and methods (B). Data are presented as mean ± SEM of triplicate determinations.

 
GnSAF activity of the HSA polypeptide 490–585 is blocked by anti-HSA antibody
Since the recombinant polypeptide 490–585 of HSA (DIIIB) immunoreacts with anti-HSA polyclonal antibodies on western blot analysis (Figure 1), we tested whether the GnSAF bioactivity of the corresponding DIIIB preparations can be blocked by these antibodies. Supernatants from induced P.pastoris GS115/DIIIB (25 µl per well) were subjected to GnSAF bioassay after pre-incubation with or without anti-HSA, and the effect on GnRH-induced LH secretion was found to be completely abrogated in the presence of anti-HSA antibodies (1 µg per well); the dose of anti-HSA applied was approximately equal, on a molar basis, to the estimated content of IIIB polypeptide in the GS115/DIIIB supernatant (Figure 3B).

Intact HSA does not display GnSAF bioactivity
Supernatants from induced P.pastoris GS115/albumin were analysed by SDS–PAGE and Coomassie brilliant blue staining (Figure 4A) and assayed for GnSAF activity (Figure 4B). Clearly, under the conditions employed, culture supernatants from GS115/albumin appear to contain an overexpressed protein corresponding to recombinant HSA (Figure 4A) but fail to display any detectable GnSAF bioactivity, indistinguishable from supernatants of GS115 transformed with empty vector (Figure 4B and Table II). To reinforce the significance of these findings further, pure HSA (Sigma A3782) was also assayed in the same system and found to display negligible GnSAF bioactivity at doses of 0.125, 0.25, 1 or 2 µg (Figure 4C); the doses of full-length HSA applied (0.1–2.0 µg, or 2–30 pmol) were equivalent to or higher, on a molar basis, than doses of HSA DIIIB estimated to be active on the GnSAF bioassay (see Pappa et al., 1999aGo).



View larger version (34K):
[in this window]
[in a new window]
 
Figure 4. Full-length HSA does not display GnSAF activity. Supernatants from induced P.pastoris GS115/empty vector (lane 1, bar 1) and GS115/albumin (lane 2, bar 2) were analysed by SDS–PAGE (15%) followed by Coomassie staining (0.1 ml/lane) (A) and subjected to GnSAF bioassay (25 µl/well) (B). In (C), pure HSA (Sigma A3782) was assayed in the GnSAF bioassay at four different doses (0.125, 0.25, 1.0 and 2.0 µg). Pre-stained molecular weight standards (A) are as in the legend to Figure 1. Values in (B) and (C) represent means ± SEM of triplicate determinations. The effects on the GnRH-induced LH secretion of rat pituitary primary cells are presented as percentages of secreted LH relative to LH secreted from control cells incubated with 0.1 µmol/l of GnRH alone, as measured with competitive ELISA. The cross-hatched horizontal bar in (B) and shaded horizontal bar in (C) indicate the mean ± SEM range for control GnRH-induced LH secretion, in each case.

 
Domains I, II or III of HSA do not display GnSAF bioactivity
The cDNA sequence corresponding to residues 381–585 of HSA (domain III; Dockal et al., 1999Go) was mobilized from GS115/albumin and used to produce two recombinant versions, GS115/DIII and GS115/DIII-myc-His6 (see Materials and methods). Similarly, the cDNA sequence of residues 1–197 (HSA domain I) or residues 189–385 (HSA domain II) was used to produce recombinant GS115/DI or GS115/DII, respectively. Culture supernatants of induced GS115/DI-1, GS115/DII-1, GS115/DIII-1 or GS115/DIII-myc-His6-1 clones were analysed by SDS–PAGE and immunoblotting and shown to contain a major protein band corresponding to overexpressed DI, DII, DIII or DIII-myc-His6, respectively (Figure 5A and B). No significant GnSAF bioactivity was found for these supernatants (25 µl) when assayed in parallel with clones of GS115/DIIIB and GS115/DIIIB-myc-His6 as positive controls (Table II); moreover, higher doses of the supernatants (30 and/or 35 µl) display marginal (GS115/DIII-1, GS115/DII-1) or undetectable bioactivity (GS115/DI-1, GS115/DIII-myc-His6-1) (Figure 5C and D).



View larger version (45K):
[in this window]
[in a new window]
 
Figure 5. Recombinant HSA domains I, II or III do not display GnSAF bioactivity. Culture supernatants (0.1 ml) of induced P.pastoris GS115/empty vector (lane 1) GS115/DIII-1 (lane 2), GS115/DIII-myc-His6-1 (lane 3), GS115/DI-1 (lane 4) and GS115/DII-1 (lane 5) were subjected to 18% (lanes 1–3) or 15% SDS–PAGE (lanes 4 and 5) and analysed by silver staining (A and B, left) and anti-HSA immunoblotting (A and B, right). Supernatants from induced GS115/DIII-1 (closed symbols) and GS115/DIII-myc-His6-1 (open symbols) (C), or induced GS115/DI-1 (closed symbols) and GS115/DII-1 (open symbols) (D) were assayed for GnSAF activity at doses of 10–35 µl/well, as indicated. Data are presented as the mean ± SEM of triplicate determinations. The shaded horizontal bars indicate the mean ± SEM range for control GnRH-induced LH secretion.

 
Chromatographic separation of the bioactive HSA subdomain IIIB (residues 490–585) from full-length HSA and HSA domain III
The finding that a C-terminal subdomain of HSA (DIIIB) appears to possess GnSAF activity, while larger entities that contain DIIIB (full-length HSA or HSA domain III) do not, raises the question of how to distinguish between DIIIB and its non-bioactive counterparts in human biological fluids and/or relevant experimental systems. In this respect, polyclonal anti-HSA antibodies are shown here to recognize DIIIB, DIII and intact HSA (Figures 1 and 5, and data not shown) while antisera with GnSAF/IF-blocking activity (Fowler et al., 2002Go) have been found to cross-react with HSA. On the other hand, GnSAF bioactivity of human ovarian cell cultures is found to separate completely from HSA by Dyematrex Blue A affinity chromatography (Fowler et al., 2002Go). These observations prompted us to apply Blue Sepharose chromatography in an attempt to compare the binding properties of DIIIB with those of DIII and intact HSA. Full-length HSA (Peters, 1996Go) as well as the three major HSA domains (DI, DII and DIII) (Dockal et al., 1999Go) are known to contain Cibacron blue-binding sites and bind to the resin. Consistently, as shown in Figure 6, the bulk of either pure HSA (Sigma A3782) or DIII contained in supernatants from clone GS115/DIII-1 binds to Blue Sepharose 6 Fast Flow and elutes from the column with an NaCl concentration of 2 mol/l. In contrast, HSA DIIIB contained in supernatants from clone GS115/DIIIB-1 does not bind to the resin and is recovered stoichiometrically in the flow-through fractions (Figure 6).



View larger version (82K):
[in this window]
[in a new window]
 
Figure 6. Comparison of HSA domain III, HSA subdomain IIIB and full-length HSA on Blue Sepharose chromatography. Dialysed culture supernatants (1 ml) of induced P.pastoris GS115/DIII-1 (lanes 1–4) or GS115/DIIIB-1 (lanes 5–8) and pure intact HSA (Sigma A3782) (0.1 mg/ml) (lanes 9–12) were subjected to Blue Sepharose chromatography, as described in Materials and methods. Proteins were eluted with 20 mmol/l NaPi pH 7.5, 2 mol/l NaCl in two sequential steps (0.1 ml/elution step). Aliquots corresponding to 10% of dialysed samples (lanes 1, 5 and 9), 10% of flow-through fractions (lanes 2, 6 and 10) and 50% of protein fractions eluting at the first (lanes 3, 7 and 11) and the second elution step (lanes 4, 8 and 12) were subjected to SDS–PAGE (18%) and silver staining.

 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
In the present study, we demonstrate that a C-terminal polypeptide sequence of HSA (residues 490–585) displays properties of GnSAF when expressed independently from an heterologous expression–secretion system. Our data are consistent with in silico analysis predicting the behaviour of this sequence as an independent domain, with junctions approximating those of HSA subdomain IIIB (sequence 495–572; He and Carter, 1992Go). More importantly, the data confirm and extend previous evidence that this polypeptide is a candidate GnSAF molecule isolatable from human FF of superovulated women (Pappa et al., 1999aGo).

Interestingly, HSA belongs to a multi-gene protein family that also includes {alpha}-fetoprotein and vitamin D-binding protein (Gibbs and Dugaiczyk, 1987Go); the molecule displays binding capacity for several ligands, including long chain fatty acids, drugs and peptide hormones (Peters, 1996Go). Domain analysis of the sequence of HSA in silico predicts three major structurally similar domains, with approximate sequences of codons 1–197 (domain I), 189–385 (domain II) and 381–585 (domain III), each of which is composed of two subdomains (IA and IB, IIA and IIB, IIIA and IIIB, respectively); this characteristic domain architecture of HSA has been emphasized in the past (Gibbs and Dugaiczyk, 1987Go; He and Carter, 1992Go; Peters, 1996Go), while clear experimental evidence that domains I, II and III are expressible as structurally independent proteins in the P.pastoris system has been provided by Dockal et al. (1999Go). In our present study, subdomain IIIB is also shown to be expressible in secreted form in the same system, providing an important experimental tool to study the mechanism of this putative GnSAF biomolecule.

The observed GnSAF bioactivity is not attributable to endogenous secreted factors of P.pastoris. It is also not derived from intact HSA (residues 1–585) or larger HSA domains such as domain III (residues 381–585) that are inactive when used as the recombinant expressed entities. Similarly, the two other major HSA domains, I and II, are inactive. Taken together, these findings indicate that the C-terminal polypeptide corresponding to subdomain IIIB of HSA possesses inherent GnSAF-like activity that is cryptic in the structural context of either full-length albumin or domain III and released only when the active moiety is separated from HSA with an as yet unidentified mechanism. This might be due to a specific proteolytic cleavage event, as postulated by Pappa et al. (1999aGo), or even to alternative splicing of the HSA transcript. In any case, the results are reminiscent of other examples where specific polypeptide portions exert different biological effects from their corresponding parental proteins, such as the angiogenesis inhibitors endostatin, a 20 kDa C-terminal portion of collagen XVII (O’Reilly et al., 1997Go), and angiostatin, a 38 kDa internal sequence of plasminogen (O’Reilly et al., 1994Go).

Although it remains unclear whether the endogenous human GnSAF bioactivity derives from one molecular entity or a number of different contributing molecules, as discussed below, the observation that a C-terminal subdomain of HSA can act as a GnSAF raises very important questions (i) on the in vivo mechanism of production of this putative GnSAF moiety in the ovarian granulosa cells and/or the FF; and (ii) on the mechanism of attenuation of the GnRH-induced LH surge in the pituitary gonadotrophs. The availability of recombinant GnSAF-like molecules from heterologous, simple biological systems such as the one presented here will greatly facilitate studies on the molecular details of the GnSAF mechanism. In this respect, studies have already been initiated to examine the effect of the relevant HSA domains on the clonal gonadotrophic cell line L{beta}T2, an in vitro model for GnRH-mediated signalling leading to pulsatile LH secretion (Turgeon et al., 1996Go; Liu et al., 2002Go; Kakar et al., 2003Go).

Much research has been carried out in the past decade to identify molecule(s) responsible for GnSAF bioactivity. Purification, sequencing and structure–functional analysis of such molecule(s) would be of major importance in the field of human fertility and provide crucial improvements in the application of IVF programmes (Messinis and Templeton, 1989Go, 1990a,Gob, 1991). Originally, Fowler et al. (1992Go) analysed human FF and recovered GnSAF/IF bioactivity in ultrafiltration fractions of 10–30 kDa without proceeding with further purification. Since then, four systematic identification trials have been described with varying results and no apparent consensus, yielding proteins between 12 and 69 kDa in size (Tio et al., 1994Go; Danforth and Cheng, 1995Go; Pappa et al., 1999aGo; Fowler et al., 2002Go). Tio et al. (1994Go) reported the isolation from rat Sertoli cell-conditioned medium of a 37 kDa protein with a partial N-terminal sequence of NH2-SDXXPQL; this molecule displayed GnSAF bioactivity but was not completely devoid of inhibin-like activity since it inhibited basal FSH secretion to some extent. Danforth and Cheng (1995Go) isolated a candidate GnSAF/IF of 69 kDa with a partial N-terminal sequence of NH2-SKPLAE from porcine FF. By using antibodies raised against the 69 kDa porcine protein, Mroueh et al. (1996Go) identified two putative GnSAF/IF-like homologues in human FF with molecular sizes of 63 and 59 kDa. More recently, Fowler et al. (2002Go) showed that GnSAF/IF activity associated with proteins of 60–66 kDa and a pI of 5.7–5.8 is present in culture supernatants from human granulosa/luteal cells of superovulated ovaries as well as from granulosa cells of unstimulated ovaries, and provided two candidate GnSAF/IF sequences (EPQVYVHAP and NH2-XVPQGNAGN).

None of the sequences of the aforementioned studies shared any significant homology with known proteins in computer-assisted database searches; furthermore, none has been confirmed as GnSAF thus far, and it remains unclear whether they represent different parts of one novel molecule, partial sequences of different bioactive molecules, or even unrelated co-purifying proteins or mixtures thereof. In a separate study, Pappa et al. (1999aGo) isolated from human FF a candidate GnSAF molecule of 12.5 kDa with a sequence corresponding to the C-terminal 95peptide of HSA, on the basis of mass spectrometric analysis. Although this work is also not exempt from criticism concerning putative co-purification contaminants, it is unique because (i) it yields a candidate sequence with homology to a known DNA portion of the human genome; (ii) it agrees with the original fractionation results of Fowler et al. (1992Go); and (iii) it is strongly supported by our present finding that the relevant polypeptide still behaves as a GnSAF when produced in recombinant form from the heterologous expression–secretion system of the yeast P.pastoris, and anti-HSA antibodies can block the GnSAF-like activity associated with this polypeptide in the P.pastoris supernatants.

GnSAF is different from inhibin (Ying, 1988Go; Fowler et al., 1990Go). Although production of GnSAF, as well as inhibin, is considered to be stimulated by FSH in both rats and humans (Koppenaal et al., 1992Go; Messinis et al., 1993Go), it has been questioned if GnSAF is regulated by FSH in the rat ovarian cycle (Tio et al., 1998Go). In any case, unequivocal distinction between GnSAF and inhibin can be made on the basis that the latter suppresses the basal secretion level of FSH, an activity that is not observed with GnSAF itself. As shown here, the bioactivity of the C-terminal 95peptide of HSA (subdomain IIIB) is free from contributions of inhibin-like activity, indicating that the corresponding moiety represents a genuine GnSAF-like molecule.

In conclusion, using the expression–secretion system of P.pastoris, we have confirmed that subdomain IIIB of HSA possesses GnSAF-like activity. Taken together with the results of Pappa et al. (1999aGo), and in the light of the other findings on candidate GnSAF molecules (Mroueh et al., 1996Go; Fowler et al., 2002Go), the data favor the interpretation that GnSAF may involve the coordinated action of more than one polypeptide species in the regulation of GnRH-induced LH secretion from pituitary gonadotrophs.


    Acknowledgements
 
We thank Dr Anastasia Politou for valuable help with using structural domain prediction programs in silico, and Dr Carol Murphy for kindly providing monoclonal antibody 9E10. This work was supported in part from the Greek Ministry for Research and Technology program PENED-99 (code no. 1438).


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Busbridge NJ, Buckley DM, Cornish M and Whitehead SA (1988) Effects of ovarian hyperstimulation and isolated preovulatory follicles on LH responses to GnRH in rats. J Reprod Fertil 82,329–336.[Abstract]

Busbridge NJ, Chamberlain GVP, Griffiths A and Whitehead SA (1990) Non-steroidal follicular factors attenuate the self-priming action of gonadotropin-releasing hormone on the pituitary gonadotroph. Neuroendocrinology 51,493–499.[ISI][Medline]

Byrne B, Fowler PA, Fraser M, Culler MD and Templeton A (1995) Gonadotropin surge-attenuating factor bioactivity in serum from superovulated women is not blocked by inhibin antibody. Biol Reprod 52,88–95.[Abstract]

Danforth DR and Cheng CY (1995) Purification of a candidate gonadotropin surge inhibiting factor from porcine follicular fluid. Endocrinology 136,1658–1665.[Abstract]

Danforth DR, Sinosich MJ, Anderson TL, Cheng CY, Bardin CW and Hodgen GD (1987) Identification of gonadotropin surge-inhibiting factor (GnSIF) in follicular fluid and its differentiation from inhibin. Biol Reprod 37,1075–1082.[Abstract]

Dockal M, Carter DC and Ruker F (1999) The three recombinant domains of human serum albumin. Structural characterization and ligand binding properties. J Biol Chem 274,29303–29310.[Abstract/Free Full Text]

Fowler PA, Messinis IE and Templeton AA (1990) Inhibition of LHRH-induced LH and FSH release by gonadotrophin surge-attenuating factor (GnSAF) from human follicular fluid. J Reprod Fertil 90,587–594.[Abstract]

Fowler PA, Townsend C, Messinis IE, Cunningham P and Templeton A (1992) Gonadotrophin surge attenuating factor attenuates in vitro LH secretion induced by gonadotrophin releasing hormone from cultured ovine pituitary cells only during the breeding season. J Endocrinol 135,221–227.[Abstract]

Fowler PA, Sorsa-Leslie T, Cash P, Dunbar B, Melvin W, Wilson Y, Mason HD and Harris W (2002) A 60–66 kDa protein with gonadotrophin surge attenuating factor bioactivity is produced by human ovarian granulosa cells. Mol Hum Reprod 8,823–832.[Abstract/Free Full Text]

Gibbs PE and Dugaiczyk A (1987) Origin of structural domains of the serum-albumin gene family and a predicted structure of the gene for vitamin D-binding protein. Mol Biol Evol 4,364–379.[Abstract]

He XM and Carter DC (1992) Atomic structure and chemistry of human serum albumin. Nature 358,209–215.[CrossRef][ISI][Medline]

Inoue H, Nojima H and Okayama H (1990) High efficiency transformation of Escherichia coli with plasmids. Gene 96,23–28.[CrossRef][ISI][Medline]

Kakar SS, Winters SJ, Zacharias W, Miller DM and Flynn S (2003) Identification of distinct gene expression profiles associated with treatment of LbetaT2 cells with gonadotropin-releasing hormone agonist using microarray analysis. Gene 308,67–77.[CrossRef][ISI][Medline]

Knight PG, Lacey M, Peter JLT and Whitehead SA (1990) Demonstration of a nonsteroidal factor in human follicular fluid that attenuates the self-priming action of gonadotropin-releasing hormone on pituitary gonadotropes. Biol Reprod 42,613–618.[Abstract]

Koppenaal DW, Tijssen AMI 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[Abstract]

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227,680–685.[ISI][Medline]

Linder S, Schliwa M and Kube-Granderath E (1996) Direct PCR screening of Pichia pastoris clones. Biotechniques 20,980–982.[ISI][Medline]

Liu F, Austin DA, Mellon PL, Olefsky JM and Webster NJ (2002) GnRH activates ERK1/2 leading to the induction of c-fos and LHbeta protein expression in LbetaT2 cells. Mol Endocrinol 16,419–434.[Abstract/Free Full Text]

Messinis IE and Templeton AA (1989) Pituitary response to exogenous LHRH in superovulated women. J Reprod Fertil 87, 633–639.[Abstract]

Messinis IE and Templeton AA (1990a) Superovulation induction in women suppresses luteinizing hormone secretion at the pituitary level. Clin Endocrinol 32,107–114.[ISI][Medline]

Messinis IE and Templeton AA (1990b) In-vivo bioactivity of gonadotrophin surge attenuating factor (GnSAF). Clin Endocrinol 33,213–218.[ISI][Medline]

Messinis IE and Templeton AA (1991) Attenuation of gonadotrophin release and reserve in superovulated women by gonadotrophin surge attenuating factor (GnSAF). Clin Endocrinol 34,259–263.[ISI][Medline]

Messinis IE, Lolis D, Papadopoulos L, Tsahalina T, Papanikolaou N, Seferiadis K and Templeton AA (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.[ISI][Medline]

Mroueh JM, Arbogast LK, Fowler P, Templeton AA, Friedman CI and Danforth DR (1996) Identification of gonadotrophin surge-inhibiting factor (GnSIF)/attenuin in human follicular fluid. Hum Reprod 11,490–496.[Abstract]

O’Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Moses M, Lane WS, Cao Y, Sage EH and Folkman J (1994) Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79,315–328.[ISI][Medline]

O’Reilly MS, Boehm T, Shing Y, Fukai N, Vasios G, Lane WS, Flynn E, Birkhead JR, Olsen BR and Folkman J (1997) Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88,277–285.[ISI][Medline]

Pappa A, Seferiadis K, Fotsis T, Shevchenko A, Marselos M, Tsolas O and Messinis IE (1999a) Purification of a candidate gonadotrophin surge attenuating factor from human follicular fluid. Hum Reprod 14,1449–1456.[Abstract/Free Full Text]

Pappa A, Seferiadis K, Marselos M, Tsolas O and Messinis IE (1999b) Development and application of competitive ELISA assays for rat LH and FSH. Theriogenology 51,911–926.[CrossRef][ISI][Medline]

Pearl FM, Lee D, Bray JE, Sillitoe I, Todd AE, Harrison AP, Thornton JM and Orengo CA (2000) Assigning genomic sequences to CATH. Nucleic Acids Res 28,277–282.[Abstract/Free Full Text]

Peters T. Jr (1996) All About Albumin: Biochemistry, Genetics and Medical Applications. Academic Press, Inc., Orlando, FL.

Schenken RS and Hodgen GD (1986) Follicle-stimulating hormone blocks estrogen-positive feedback during the early follicular phase in monkeys. Fertil Steril 45,556–560.[ISI][Medline]

Sopelak VM and Hodgen GD (1984) Blockade of the estrogen-induced luteinizing hormone surge in monkeys: a nonsteroidal, antigenic factor in porcine follicular fluid. Fertil Steril 41,108–113.[ISI][Medline]

Tijssen AM, Helder MN, Chu ZW 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.[Abstract]

Tio S, Koppenaal D, Bardin CW and Cheng CY (1994) Purification of gonadotropin surge-inhibiting factor from Sertoli cell-enriched culture medium. Biochem Biophys Res Commun 199,1229–1236.[CrossRef][ISI][Medline]

Tio S, van Dieten JA and de Koning J (1998) Immunoneutralization of follicle stimulating hormone does not affect gonadotrophin surge-inhibiting factor/attenuating factor bioactivity during the rat ovarian cycle. Hum Reprod 13, 2731–2737.[Abstract/Free Full Text]

Turgeon JL, Kimura Y, Waring DW and Mellon PL (1996) Steroid and pulsatile gonadotropin releasing hormone (GnRH) regulation of luteinizing hormone and GnRH receptor in a novel gonadotrope cell line. Mol Endocrinol 10,439–450.[Abstract]

Ying SY (1988) Inhibins, activins, and follistatins: gonadal proteins modulating the secretion of follicle-stimulating hormone. Endocr Rev 9,267–293.[Abstract]

Submitted on November 25, 2003; accepted on January 9, 2004.