Glycodelin-S in human seminal plasma reduces cholesterol efflux and inhibits capacitation of spermatozoa.

Tight control of sperm capacitation is important for successful fertilization. Glycodelin-S is one of the most abundant glycoproteins in the human seminal plasma. However, its function is unclear. We investigated the role of glycodelin-S on capacitation of human spermatozoa. Binding kinetics experiments demonstrated the presence of two saturable and reversible binding sites of glycodelin-S on human spermatozoa. Differently glycosylated other isoforms of glycodelin, glycodelin-A and -F, did not compete with glycodelin-S for these binding sites, suggesting that the glycodelin-S binding sites are different from those of the other isoforms. Indirect immunofluorescent staining revealed specific binding of glycodelin-S around the sperm head. This immunoreactivity was greatly reduced in spermatozoa that had migrated through the cervical mucus surrogates. Glycodelin-S at physiological concentrations significantly reduced the bovine serum albumin and cyclodextrin-induced cholesterol efflux and down-regulated the adenylyl cyclase/protein kinase A/tyrosine kinase signaling pathway, resulting in suppression of capacitation. Deglycosylation abolished glycodelin-S binding and the effect of glycodelin-S on bovine serum albumin-induced capacitation. This indicates that the carbohydrate moiety of glycodelin-S is critical for the function of the molecule. It is concluded that glycodelin-S in seminal plasma maintains the uncapacitated state of human spermatozoa.

Freshly ejaculated mammalian spermatozoa are not immediately capable of fertilizing an oocyte. They acquire their fertilizing capacity by an ill-defined process termed capacitation that normally occurs during migration of the spermatozoa in the female genital tract (1). After capacitation, the spermatozoa are capable of undergoing acrosome reaction upon stimulation by the zona pellucida (ZP) 1 proteins (2) and progesterone (3), thereby releasing acrosomal enzymes for penetration through the ZP. Capacitated human spermatozoa remain responsive to ZP-induced acrosome reaction in vitro for 50 -240 min only (4). Thus a tight control in the timing of capacitation is important to ensure the occurrence of appropriate sequence of events in the fertilization process. The male sex accessory glands secrete complex mixture of proteins, glycoproteins, peptides, glycopeptides, and prostaglandins into seminal plasma (1). The seminal plasma is capable of maintaining spermatozoa in an uncapacitated state. This is essential to ensure that capacitation would not start before migration of the spermatozoa to the fertilization site. In fact, seminal plasma can "decapacitate" capacitated spermatozoa (5).
GdA is the most extensively studied glycodelin isoform. Its abundance in endometrial glands and its immunosuppressive activity suggest that it may play a role in the fetomaternal defense system (9,10). GdA is the first endogenous glycoprotein found to inhibit spermatozoa-ZP binding (11). Its absence in the endometrium in the periovulatory period is related to the presence of a fertilization window (6). GdF also inhibits spermatozoa-ZP binding (8). GdF, but not the other glycodelin isoforms, suppresses progesterone-induced acrosome reaction (8,12), suggesting that it may protect spermatozoa from premature acrosome reaction before binding to the ZP. Glycodelin-S does not adversely affect spermatozoa-zona binding (6), and its action on spermatozoa is not clear.
GdS is one of the most abundant glycoproteins in the seminal plasma with a concentration about 1700 nM (13). Unlike GdA and GdF, GdS does not affect spermatozoa-ZP binding (14). GdS has unusual fucose-rich glycans, and its major complex type glycan structures are bi-antennary with Lewis x and Lewis y antennae. The glycosylation of GdS is unusual among secreted human glycoproteins, because it has no sialylated glycans, and its Asn-28 contains only high mannose structures, whereas Asn-63 contains only complex type glycans (14). The immunosuppressive properties of the Lewis x and Lewis y epitopes in GdS may contribute to the low immunogenicity of semen in women despite frequent exposure.
In this report, we studied the role of GdS in regulating capacitation of spermatozoa in view of three objectives. The first objective was to investigate the binding of GdS to spermatozoa, a prerequisite for the molecule to exert its effect on spermatozoa. The second objective was to determine the relationship between GdS and capacitation. The third was to study the mechanism of action of GdS on capacitation. In particular, we hypothesize that GdS suppresses capacitation by reducing cholesterol efflux from spermatozoa. We also address the ability of GdS to suppress bovine serum albumin (BSA)-induced or cyclodextrin-induced cholesterol efflux and the effect of GdS on enzymes involved in the intracellular signaling pathways leading to capacitation.

MATERIALS AND METHODS
Semen Samples-The Ethics Committee of the University of Hong Kong approved the research protocol. Only samples with normal semen parameters according to World Health Organization criteria (15) were used. Spermatozoa were processed by using Percoll (Amersham Biosciences) density gradient centrifugation as described previously (8). The processed spermatozoa were then resuspended in Earle's balanced salt solution supplemented with 0.3% BSA, 0.3 mM sodium pyruvate, 0.16 mM penicillin-G, 0.05 mM streptomycin sulfate, and 14 mM sodium bicarbonate (EBSS/0.3%BSA) to a concentration of 2 ϫ 10 6 spermatozoa/ml.
Purification of Glycodelin Isoforms-Glycodelin-A (GdA), -S (GdS), and -F (GdF) were purified from amniotic fluid, seminal plasma, and follicular fluid respectively, using a monoclonal anti-glycodelin antibody (clone F43-7F9) Sepharose column as described (12). The bound glycodelin was eluted with 0.1% trifluoroacetic acid and dialyzed against 100 mM sodium phosphate buffer (pH 7.2). GdS was further purified by anion exchange chromatography as described (12). GdF was purified from human follicular fluid collected during oocyte retrieval from women undergoing assisted reproduction treatment in Queen Mary Hospital, Hong Kong. The follicular fluid was passed successively through Hi-Trap blue, protein-G, ConA-Sepharose columns (Amersham Biosciences), Amicon-10 concentrator (Amicon Inc., Beverly, CA), Mono-Q, and Superose columns. The concentrations of the purified glycodelin isoforms were determined by the Bio-Rad Protein Assay kit.
Deglycosylated GdS was prepared using the N-glycosidase F deglycosylation kit (Bio-Rad) as described (12). The deglycosylated protein was obtained after three successive protein precipitations, redissolved in 20 l of PBS, and further purified by gel filtration chromatography in a SMART system (Amersham Biosciences). The configuration of the glycodelin protein core is unlikely to be affected by the procedure, because deglycosylated glycodelin prepared as described possesses the same immunosuppressive activity as purified GdA (16) and recombinant glycodelin from Escherichia coli without post-translational modifications (17). Moreover, thermodynamic evidence is available showing that the native folding of glycodelin is not influenced by glycosylation (18). The purity of the deglycosylated protein was analyzed by SDS-PAGE, and its concentration was determined.
Radioactively labeled GdS was prepared by mixing 50 g of GdS in 0.02 ml of 0.05 M PBS (pH 7.4) with 2 mCi of sodium 125 I (20 l, Amersham Biosciences) and freshly prepared chloramine T (100 g in 0.02 ml of 0.05 M PBS, pH 7.4) as described (12). The first radioactive peak containing iodinated glycodelin was collected.
Cervical Mucus Penetration-Human cervical mucus was not used in this study for three reasons. First, cervical mucus contains glycodelin of unknown isoform (19). Second, the mucus is not homogeneous, because it may be contaminated with vaginal fluid. Third, its viscosity fluctuates widely depending on the day of menstrual cycle, making interpretation of data difficult. Two commercially available substitutes of cervical mucus were used. They were methylcellulose with a viscosity of 4000 centipoises (MC4000, Aldrich) and hyaluronic acid from rooster comb with molecular mass range of 1-4 ϫ 10 6 Da (Sigma). Previously, MC4000 and hyaluronic acid at concentrations of 10 and 6 mg/ml, respectively, have been used successfully as cervical mucus surrogate (20,21). The substitutes were dissolved in EBSS/0.3% BSA.
To perform the mucus penetration test, a capillary model similar to that described for studying the cumulus oophorus was used (22). Briefly, a sterile glass capillary (Microcaps, Drummund, PA) was successively filled with EBSS/0.3% BSA to a length of 5 cm (medium column), and the mucus was substituted to form a mucus column of 5 cm in length. The end of the capillary with the mucus column was dipped into a 100-l droplet of liquefied semen overlaid with mineral oil. Another capillary containing only EBSS/0.3% BSA served as a control. Spermatozoa were allowed to migrate through the mucus col-umn into the medium column for 30 min at 37°C under 5% CO 2 in air before the capillary was cut at position of 1 cm above the interface between the mucus column and the medium column. Spermatozoa in the medium column were collected. Spermatozoa that had swum to the corresponding level in the control capillary were used as controls.
Determination of Acrosomal Reaction and Capacitation-fluorescein isothiocyanate-labeled Pisum sativum agglutinin (Sigma) and Hoechst staining techniques were used to evaluate the acrosomal status of the spermatozoa as described (23). The fluorescence patterns of 300 spermatozoa in randomly selected fields were determined under a fluorescence microscope (Zeiss, Oberkochen, Germany) with 1000ϫ magnification. A digital camera (COOLSNAP, Photometrics, AZ) and its associated software (Photometrics) were used to capture the image at room temperature. The filter set used for Hoechst staining consisted of an excitation filter G365, a chromatic beam splitter FT395, and a barrier filter LP420, whereas that for fluorescein isothiocyanate-labeled P. sativum agglutinin consisted of an excitation filter BP 450 -490, a chromatic beam splitter FT510, and a barrier filter LP520. Spermatozoa without staining or those with fluorescein isothiocyanatelabeled P. sativum agglutinin staining confined to the equatorial segment only were considered as acrosome-reacted spermatozoa.
Sperm capacitation was assayed by the chlortetracycline staining (CTC) method as described before (24). The capacitation status and acrosomal status of 200 spermatozoa were evaluated under a fluorescence microscope (Zeiss) at ϫ630 magnification with a filter set consisting of an excitation filter BP 450 -490, a chromatic beam splitter FT510, and a barrier filter LP520. Five CTC staining patterns of the sperm head were identified (25). CTC4 pattern (uniform head fluorescence) was the main capacitated pattern, and CTC5 pattern (decrease in or loss of uniform head fluorescence) was the acrosome-reacted pattern. The incidence of uncapacitated patterns (CTC1-3) decreased, whereas that of the CTC4 and CTC5 increased with the duration of capacitation of the spermatozoa (24).
Quantification of Cholesterol-Sperm cholesterol was extracted by the method of Folch and coworkers (26) adapted for spermatozoa. Briefly, sperm suspension was washed thrice with PBS before vortexing with 8 ml of chloroform:methanol (2:1, v/v). Distilled water (1.5 ml) was then added. The mixture was vortexed again and allowed to stand at room temperature for 1 h before centrifugation at 500 ϫ g for 10 min at 4°C. The upper layer was resuspended in 8 ml of chloroform:methanol: water (86:16:1, v/v) and centrifuged at 500 ϫ g for 10 min at 4°C, after which the chloroform extracts were pooled. The chloroform in the extract was removed in a rotary evaporator (Virtis, Gardiner, NY). The cholesterol content in the dried extract was then determined by a cholesterol quantification kit (BioVision) according to the manufacturer's instructions. Sperm cholesterol content was expressed as the amount of cholesterol per million of spermatozoa.
Determination of cAMP-Intracellular cAMP was extracted as described by Calogero and coworkers (27). Briefly, spermatozoa that had been incubated for various durations were collected by centrifugation at 3,000 ϫ g for 5 min, and resuspended in 0.5 ml of ice-cold 90% ethanol. After 30 min at Ϫ20°C and 30 min at 4°C, the samples were centrifuged at 19,000 ϫ g. The supernatants were collected, and the ethanol was evaporated in a rotary evaporator. The dried pellets were stored at Ϫ20°C until used. Intracellular cAMP was then determined using a non-radioactive cAMP enzyme-linked immunosorbent assay Kit (R&D, Minneapolis, MN) according to the manufacturer's instructions.
Determination of Protein Kinase Activity-Spermatozoa were washed thrice with PBS before solubilization by sonication in 100 l of homogenizing buffer (20 mM PBS, pH 7.4, containing 2 mM EDTA, 1 mM dithiothreitol, 0.1 mM phenylmethanesulfonyl fluoride, 0.1 mM vanadate, 1 mM MgCl 2 , 100 mM NaCl, and 0.05% Triton X-100) for 15 min at 4°C as described (28). The supernatant was obtained after centrifugation at 15,800 ϫ g for 30 min at 4°C. The kinase activity in the supernatant was determined at room temperature. Enzyme-linked immunosorbent assay-based tyrosine kinase (Molecular Probes) and protein kinase A (Calbiochem) assay kits were used according to the manufacturers ' instructions. Standard curves were run along with the test samples in each experiment. The protein content in the extracts was determined using commercial kit (Bio-Rad). One unit of protein kinase A and tyrosine kinase activity was defined as the amount of enzyme required to catalyze the transfer of 1 pmol of phosphate to the substrates, RFARKGSLRQKNV and TSTEPQYQPGENL, respectively, in 1 min at 30°C.
Detection of Tyrosine Phosphorylation-Spermatozoa (2 ϫ 10 6 ) were concentrated by centrifugation at 20,000 ϫ g for 2 min at room temperature and washed once in 0.5 ml of PBS at room temperature. The sperm pellet was then resuspended in SDS sample buffer without mercaptoethanol and boiled for 5 min. After a further centrifugation at 20,000 ϫ g for 2 min, the supernatant was collected. 2-Mercaptoethanol was added to a final concentration of 5%. The sample was boiled for 3 min before subjected to SDS-PAGE using 10% gels in a Mini-PROTEIN 3 Electrophoresis System (Bio-Rad). The gel was blotted on a polyvinylidene fluoride membrane (Millipore, Bedford, MA) for Western blot analysis using the Chromogenic Western Blot Kit (Invitrogen) according to the manufacturer's instructions. Monoclonal primary antibody, clone 16F4 (Calbiochem) specific for tyrosine phosphorylation was used. In addition, anti-tubulin antibody (Sigma) was used to reveal the equal sample loading of each lane.
Equilibrium Binding of 125 I-Glycodelin-S to Spermatozoa-Four semen samples were used (n ϭ 4). Spermatozoa (200,000 in 100 l of EBSS/0.3% BSA) were incubated with different concentrations (0.16 -4700 nM) of 125 I-GdS at 37°C for 3 h. The binding was terminated by the addition of 1.5 ml of ice-cold PBS followed by centrifugation at 300 g for 3 min. The spermatozoa were further washed with fresh EBSS/0.3% BSA. The radioactivity associated with the spermatozoa was counted with a gamma counter (Model 5500B, Beckman). Specific binding of GdS was determined by subtracting the radioactivity bound on the spermatozoa in the presence of a 100-fold concentration of unlabeled GdS from that in the absence of unlabeled protein (12). The determinations of total binding and nonspecific binding were done in triplicate.
Specificity of Glycodelin-S Binding to Spermatozoa-Competition binding analysis was used to compare the affinity of GdS binding sites for glycodelins and other lipocalins (12). The binding of 2000 nM of 125 I-GdS to 200,000 spermatozoa (n ϭ 4) in 100 l of EBSS/0.3% BSA was determined in the presence of an increasing concentration (200, 2000, 5000, 0.2 ϫ 10 5 , and 1 ϫ 10 5 nM) of unlabeled GdS, deglycosylated GdS, GdA, GdF, bovine ␤-lactoglobulin A (Sigma), and human retinolbinding protein (Sigma) or buffer alone at 37°C for 360 min. The cell-bound radioactivity was determined as described above. Each individual experiment was repeated thrice.
Immunofluorescence Staining of Glycodelin-S on Human Spermatozoa-Liquefied semen was divided into three aliquots. Motile spermatozoa in one aliquot were collected by the swim-up technique (15). Briefly, the semen was gently overlaid with EBSS/0.3% BSA and incubated for 1 h at 37°C. Spermatozoa that swam into the medium were collected. Spermatozoa in the second aliquot were collected after cervical mucus penetration as described above. Spermatozoa in the last aliquot were processed by Percoll gradient centrifugation, followed by incubation with 1700 nM GdS or deglycosylated GdS for 60 min at 37°C under 5% CO 2 in air. The concentration of GdS used was physiological in human seminal plasma (13). Monoclonal anti-glycodelin antibody (clone F43-7F9) recognizing the protein core of glycodelins was used to determine glycodelin immunoreactivity in the spermatozoa from the three aliquots as described (8). The staining was observed with an excitation filter BP 450 -490, chromatic beam splitter FT510, and barrier filter LP520. Percoll processed spermatozoa without prior GdS treatment was used as the control.
Effect of Glycodelin-S on Albumin and Methyl-␤-cyclodextrin-induced Capacitation-Albumin acts as a cholesterol acceptor in the fluid of female reproductive tract, leading to sperm capacitation. Therefore, it has been widely used in media for capacitation of spermatozoa. To study the effect of GdS on BSA-induced capacitation, Percoll processed spermatozoa (n ϭ 5) were incubated with 0.3, 3, 30, 300, 900, 1500, 2000, and 3000 nM of GdS, deglycosylated GdS, GdA, or EBSS/0.3% BSA (control) at 37°C under 5% CO 2 in air for 60 min. Spermatozoa were then capacitated in EBSS containing 3% BSA (EBSS/3% BSA) for 3 h under the same conditions before washing with EBSS/0.3% BSA. The capacitation status and cholesterol content of spermatozoa from each group was then evaluated by the methods described above.
The effect of GdS on cyclodextrin-induced capacitation was also investigated. ␤-Cyclodextrins are cyclic heptasaccharides that bind cholesterol and induce cholesterol efflux from cells. ␤-Cyclodextrins stimulate sperm cholesterol efflux, capacitation, and ZP/progesteroneinduced acrosome reaction (29 -33). The procedure in this experiment was similar to that described for BSA except that 3 mM methyl-␤cyclodextrin in EBSS/0.3% BSA, pH 7.4 (M␤CD), was used instead of EBSS/3% BSA, and the capacitation time was 30 min. The M␤CD concentration used decreased sperm cholesterol efficiently and had a minimal effect on sperm viability (30). In this study, the percentages of viable spermatozoa after incubation in M␤CD ranged from 92% to 99% as determined by the trypan blue exclusion test.
To determine whether GdS could reverse the capacitation process, Percoll-treated spermatozoa (n ϭ 5) were first capacitated in EBSS/3% BSA or M␤CD at 37°C under 5% CO 2 in air for 3 h. The treated spermatozoa were washed and incubated in 30, 300, 1500, and 3000 nM of GdS for 60 min in the same conditions. The capacitation status of the spermatozoa was then evaluated as described above.
Effect of Glycodelin-S on Zona Pellucida-induced Acrosome Reaction-To confirm the result of the CTC staining, the zona pellucidainduced capacitation assay was performed. The assay was based on the observation that only capacitated spermatozoa are able to undergo acrosomal exocytosis upon ZP stimulation (1,2). In this assay, human ZP was separated from the oocytes using micropipette under a microscope and heat-solubilized at 70°C for 90 min in sodium carbonate solution, pH 9 (34). The amount of zona pellucida used is expressed as the number of zona pellucida solubilized per microliter of the final incubation medium (ZP/ml). Percoll processed spermatozoa (n ϭ 5) were incubated with 0, 30, 300, 900, 1500, 2000, and 3000 nM GdS, deglycosylated GdS, GdA, or EBSS/0.3% BSA (control) at 37°C under 5% CO 2 in air for 60 min. They were then capacitated in EBSS/3% BSA for 3 h or M␤CD for 30 min, washed with EBSS/0.3% BSA, and incubated for 30 min in EBSS/0.3% BSA containing solubilized ZP (5 ZP/ml). The percentage of acrosome-reacted spermatozoa was then determined by P. sativum agglutinin-fluorescein isothiocyanate staining.
Investigation of Cholesterol Distribution by Filipin-To confirm the effect of GdS on sperm cholesterol content, in situ double staining of GdS and cholesterol with monoclonal anti-glycodelin antibody and filipin, respectively, was performed. Filipin is an antibiotic that forms complexes with non-esterified membrane cholesterol. Percoll-processed spermatozoa were incubated with 2000 nM GdS, deglycosylated GdS, GdA, or EBSS/0.3% BSA (control) at 37°C under 5% CO 2 in air for 60 min. They were then capacitated in EBSS/3% BSA for 3 h or M␤CD for 30 min, washed with fresh EBSS/0.3% BSA, and smeared on glass slides before fixation in 2% formaldehyde in PBS for 30 min at room temperature and washed thrice in PBS containing 1% BSA. The spermatozoa were then stained with 25 M filipin in PBS in the dark for 30 min. Identical samples treated similarly without filipin were used as controls. After filipin staining, the spermatozoa were washed thrice with PBS before immunostaining for glycodelin as described above. Filipin and glycodelin fluorescence was observed under a fluorescence microscope (Zeiss) with 1000ϫ magnification. The filters used for filipin staining were excitation filter G365, chromatic beam splitter FT395, and barrier filter LP420.

Effects of Glycodelin-S on Albumin and M␤CD-induced Changes in cAMP Concentration and Protein
Kinase Activities-It has been shown that cholesterol efflux promotes protein kinase A activities, protein tyrosine phosphorylation, and capacitation (31)(32)(33)35). cAMP is also known to be important in capacitation (35). Its analogues enhance capacitation and overcome the suppressive effect of cholesterol on BSAinduced capacitation (31)(32)(33).
To evaluate the potential site of action of GdS in these signaling pathways, Percoll-processed spermatozoa (n ϭ 5) were incubated with 2000 nM GdS, deglycosylated GdS, GdA, or EBSS/0.3% BSA (control) at 37°C under 5% CO 2 in air for 60 min. They were capacitated in EBSS/3% BSA for 3 h or M␤CD for 30 min, with or without 1 mM Sp-cAMPS, a membrane-permeable cAMP analogue resistant to cyclic nucleotide phosphodiesterases that activates protein kinase A. The intracellular cAMP level, protein kinases activities, and protein tyrosine phosphorylation pattern of the treated spermatozoa were evaluated as described above.
Data Analysis-All the data were expressed as mean Ϯ S.E. The data were analyzed by using statistical programs (SigmaPlot 8.02, Ligand Binding Analysis Module & SigmaStat 2.03, Jandel Scientific, San Rafael, CA). For all experiments, the non-parametric analysis of variance on rank test for multiple comparisons was used. The parametric Student's t test or the non-parametric Mann Whitney U test were used where appropriate as the post-test. A probability value of Ͻ0.05 was considered to be statistically significant.

Equilibrium Binding of 125 I-Glycodelin-S to Spermatozoa-
Specific binding of 125 I-GdS to spermatozoa increased with the concentration of GdS used up to 2000 nM, after which no further increase was observed (Fig. 1). This result indicated that the binding was saturable. Analysis of the saturation data revealed a curvilinear plot. Scatchard plots best fitted by nonlinear regression analysis (R 2 Ͼ 0.98) suggested the presence of two specific binding sites (Fig. 1A). The low affinity binding sites (K D 1413 Ϯ 316 nM; B max 17 Ϯ 3 pmol/2 ϫ 10 6 spermatozoa) were more abundant than the high affinity binding sites (K D 104 Ϯ 20 nM; B max 8 Ϯ 1 pmol/2 ϫ 10 6 spermatozoa). Hill equation analysis of the binding data yielded a Hill coefficient of less than unity (0.75 Ϯ 0.07) (Fig. 1B), further suggesting binding heterogeneity.
Binding Kinetics- Fig. 2 shows the binding kinetics of 125 I-GdS to spermatozoa at 37°C. Specific binding increased rapidly for the first 10 min, reaching equilibrium after about 30 min. These association data ( Fig. 2A) were best fitted by a double-exponential equation (R 2 Ͼ 0.97), indicating the presence of two populations of binding sites with different observed association rate constants (K obs1 and K obs2 ). The major population (69.4% of the total) had a slow observed association rate constant of 0.11 Ϯ 0.05 per minute (K obs1 ) compared with the minor population (30.6% of the total) with an observed association rate constant of 0.06 Ϯ 0.08 per minute (K obs2 ).
The dissociation kinetics at 37°C was also best described by two-exponential functions (R 2 Ͼ 0.98) (Fig. 2B), supporting the presence of two binding site populations. These binding sites had different dissociation constants (K off1 and K off2 ). The larger population (65.7% of the total) had a dissociation rate constant of 0.028 Ϯ 0.002 min Ϫ1 (K off1 ) corresponding to a dissociation half-life of 25 min (obtained from natural log 2/K off ). The smaller population (34.3% of the total) had a dissociation rate constant of 0.007 Ϯ 0.0006 min Ϫ1 (K off2 ) and a half-life of 99 min.
The kinetic data indicated that the binding of GdS was reversible. The true association rate constant (K on ) was calculated from the equation, (K obs Ϫ K off )/L, where L is the concen-tration of 125 I-GdS used in the binding kinetic experiment. The K D values of the two binding sites derived from the rate constants according to the relationship: K D ϭ K off /K on were 136 and 1700 nM. These values closely agreed with that obtained in the equilibrium binding study.
Specificity of Glycodelin-S Binding to Human Spermatozoa- Fig. 3 shows the results of competitive binding to human spermatozoa between 125 I-GdS and different lipocalin family members. As expected, unlabeled GdS inhibited the binding of 125 I-GdS in a dose-dependent manner with a half-maximal inhibition (IC 50 ) of 2300 nM. Glycodelin-F, glycodelin-A, and deglycosylated GdS inhibited the binding at high concentrations (IC 50 Ͼ 2.5 ϫ 10 5 nM), whereas the other lipocalin proteins tested did not affect the binding of 125 I-GdS to spermatozoa at all (IC 50 Ͼ 20 ϫ 10 5 nM).
Localization of Glycodelin Immunoreactivity on Human Spermatozoa-Glycodelin immunoreactivity was localized to the whole head of swim-up-processed spermatozoa (Fig. 4A). A vast majority (85-90%) of these spermatozoa possessed glycodelin immunoreactivity. The immunoreactivity of the sperm head disappeared after Percoll processing (Fig. 4G). It was greatly reduced after passing through the cervical mucus surrogates (Fig. 4, B and C), but regained after incubation with GdS (Fig. 4E). Percoll-processed spermatozoa incubated with deglycosylated GdS (Fig. 4F) did not possess glycodelin immunoreactivity.
Effect of Glycodelin-S on Acrosomal Status of Human Spermatozoa-Glycodelin-S, GdA, or deglycosylated GdS at concentrations of 0.3-3000 nM did not affect spontaneous acrosome reaction as determined by P. sativum agglutinin-fluorescein isothiocyanate and CTC staining (data not shown). GdS in all the concentrations tested did not affect sperm capacitation, whereas it inhibited 3% BSA-and M␤CD-induced capacitation of Percoll-processed spermatozoa at concentrations Ͼ1500 nM and Ͼ2000 nM, respectively (Fig. 5A). Using the CTC staining, the percentages of 3% BSA-and M␤CD-induced capacitated spermatozoa were 37 Ϯ 2% and 49 Ϯ 4%, respectively. These values decreased with the addition of GdS. The corresponding values at GdS concentration of 3000 nM were 14 Ϯ 3% and 29 Ϯ 3%. A similar observation was found using the ZP-induced capacitation assay (Fig. 5B). In the absence of GdS, spermatozoa treated with 3% BSA or M␤CD had a significantly higher percentage of ZP-induced acrosome reacted spermatozoa than those incubated in EBSS/0.3% BSA. The stimulatory activities of 3% BSA and M␤CD were significantly inhibited by GdS at concentrations Ͼ1500 nM and Ͼ2000 nM, respectively (Fig. 5B). Deglycosylated GdS and GdA had no effects on 3% BSA-and M␤CD-induced capacitation (data not shown). Although GdS inhibited 3% BSA-and M␤CD-induced capacitation, it could not revert the capacitation status of spermatozoa that had been capacitated in 3% BSA or M␤CD as determined by CTC staining or ZP-induced acrosome reaction (data not shown).
Effect of Glycodelin-S on Sperm Cholesterol Content-There was a significant decrease in cholesterol content in spermatozoa treated with 3% BSA or M␤CD compared with those incu-bated in EBSS/0.3% BSA (Fig. 6). M␤CD was more efficient than 3% BSA in enhancing cholesterol efflux. GdS alone had no effect on sperm cholesterol content. However, GdS at concentrations Ͼ900 nM significantly inhibited 3% BSA-induced cholesterol loss as compared with the control. The cholesterol content of spermatozoa that had been previously treated with 3000 nM GdS remained high (329 Ϯ 37 pmol/10 6 spermatozoa) after treatment with 3% BSA, and it was significantly higher than that without prior GdS treatment (197 Ϯ 24 pmol/10 6 spermatozoa). GdS at concentrations Ͼ2000 nM also significantly reduced M␤CD-induced cholesterol loss; the cholesterol concentrations of spermatozoa with or without 3000 nM GdS were 285 Ϯ 19 and 142 Ϯ 18 pmol/10 6 spermatozoa, respectively (Fig. 6).
The same conclusion was obtained from the filipin staining. In Percoll-processed but non-capacitated spermatozoa, the intensity of filipin staining was uniformly high in the acrosome and equatorial regions and weak in the postacrosomal region and the tail (Fig. 7A). These data are consistent with a previous report (36). Capacitation in 3% BSA reduced the intensity of staining on the sperm head. In addition, the fluorescent signal became aggregated and gave a heterogeneous appearance (Fig.  7B). Compared with 3% BSA, the cholesterol removal effect of M␤CD was higher resulting in almost complete disappearance of the signal (Fig. 7D). The binding of GdS on spermatozoa suppressed the effect of 3% BSA and M␤CD as demonstrated by the strong filipin staining after capacitation treatment (Fig.  7, C and E). Deglycosylated GdS and GdA had no effect on sperm cholesterol content (data not shown).
Effects of Glycodelin-S on Intracellular cAMP Concentration- Fig. 8 shows the effect of GdS on intracellular levels of cAMP in capacitated and non-capacitated spermatozoa. GdS alone had no effect on the cAMP concentration as demonstrated by similar cAMP concentrations in spermatozoa with or without prior GdS treatment. On the other hand, both 3% BSA and M␤CD induced a significant increase in cAMP concentration, the effect of M␤CD being more rapid. The cAMP level increased from 35 Ϯ 4 fmol/10 6 spermatozoa to 52 Ϯ 2 fmol/10 6 spermatozoa after 180 min of 3% BSA treatment, and from 39 Ϯ 3 fmol/10 6 spermatozoa to 76 Ϯ 2 fmol/10 6 spermatozoa after 30 min of M␤CD treatment. Incubation with 2000 nM GdS significantly reduced the 3% BSA (36 Ϯ 1.8 fmol/10 6 spermatozoa after 180 min)-and M␤CD (61 Ϯ 2.4 fmol/10 6 spermatozoa after 30 min)-induced up-regulation of cAMP level. Neither deglycosylated GdS nor GdA had any effect on BSA-or M␤CDmediated increase in protein kinase activities and tyrosine phosphorylation (data not shown).
The results of tyrosine phosphorylation also demonstrated that 2000 nM GdS inhibited the capacitation-associated increase in protein tyrosine phosphorylation mediated by 3% BSA (Fig. 10, lanes 5 and 6) and M␤CD (Fig. 10, lanes 9 and  10). On the other hand, the protein kinase activities (Fig. 9, A  and B) and tyrosine phosphorylation (Fig. 10, lanes 1 and 2) of  The effects of GdS on BSA and M␤CD-induced kinase activities were not due to abrogation of BSA and M␤CD function, because Sp-cAMPS could overcome the effect of GdS. The addition of Sp-cAMPS supported protein kinase activity (Fig. 9, A and B) and protein tyrosine phosphorylation (Fig. 10, lanes 8  and 12) to levels similar to those seen with 3% BSA (Fig. 10,  lane 7) or M␤CD (Fig. 10, lane 11) incubation with or without GdS. This demonstrates that the effect of GdS is directed upstream of protein kinase A and tyrosine kinase activation in the signal transduction cascade. Again, neither deglycosylated GdS nor GdA had any effect on the BSA-or M␤CD-mediated increase in protein kinase activity and tyrosine phosphorylation (data not shown).

DISCUSSION
This is the first report on characterization of the binding of GdStohumanspermatozoa.Thebindingistime-andconcentrationdependent. There are two binding sites for GdS; a low and a high affinity binding site. Three observations indicate that, on human spermatozoa, these binding sites are different from those of GdA and GdF (8,12). First, the affinity of GdS to spermatozoa is much lower than that of GdF. Even the K D of the high affinity binding site of GdS (about 100 nM) is four times higher than that of the low affinity binding site of GdF (about 25 nM). The low binding affinity of GdS explains why it is readily removed from the sperm surface and has a short half-life of dissociation. Second, the binding sites of GdA and GdF are confined to the acrosomal region (8), whereas the binding sites of GdS are spread all over the sperm head. Third, in the competition binding assay, GdA and GdF cannot significantly compete with GdS for its binding sites. Likewise, the other lipocalin family members are ineffective in competing with GdS. Thus, the binding of GdS to spermatozoa is specific. This is in line with the suggestion that lipocalins do not promiscuously bind to human spermatozoa (12).
The ejaculated spermatozoa have to traverse through the cervical mucus to enter the upper female genital tract for fertilization. The cervical mucus has several proposed functions, including regulation of sperm transport, selection of spermatozoa according to their motility and morphology, protection of the endocervical epithelium against bacterial invasion and fluid loss (37), and initiation of capacitation (38). However, the molecular mechanism of action of these functions is poorly understood.
The present study provides a molecular basis for the initiation of capacitation by cervical mucus, i.e. removal of seminal plasma factors that maintain spermatozoa in an uncapacitated state (uncapacitation factors). The glycodelin immunoreactivity of spermatozoa became greatly reduced after cervical mucus penetration. This result is consistent with the observation that interaction between spermatozoa and cervical mucus results in loss of decapacitation factors (39). Data derived from the present study suggest that GdS is an uncapacitation factor.
Although the swim-up process retains glycodelin immunoreactivity on ejaculated human spermatozoa, Percoll gradient centrifugation removes the immunoreactivity. The main difference between the two sperm-processing procedures is that spermatozoa are forced through a colloid solution by centrifugal force in the latter, but not in the former. The passage of spermatozoa through the Percoll gradient could be similar to the penetration through cervical mucus. A previous study also showed that Percoll processing reinforced shedding of loosely attached seminal plasma components from spermatozoa (40). Transmission electron microscopy study suggested that Percoll processing removed a coating envelope from the head and tail regions of spermatozoa, which lead to higher zona-free hamster egg-penetration ability of Percoll processed spermatozoa, com- pared with those prepared by the swim-up technique (41).
Capacitation is generally assessed by the ability of the spermatozoa to undergo acrosome reaction in response to physiological inducers, such as ZP or progesterone (1)(2)(3). Capacitation starts in the female reproductive tract, and acrosome reaction marks completion of the process. Some components of the seminal plasma are adsorbed on the surface of ejaculated spermatozoa (43). The fact that most of these seminal plasma components are shed from spermatozoa during their ascent to the oviduct is consistent with the proposal that the removal of the seminal plasma components is necessary for capacitation (44). Spermatozoa are exposed to uncapacitation factors in seminal plasma, but only after ejaculation. A rapid binding of these factors to spermatozoa is important to ensure that ejaculated spermatozoa will remain uncapacitated in the vagina. On the other hand, it is also essential that these factors can be removed from spermatozoa during their passage through the cervical mucus to allow initiation of capacitation. The rapid association to and dissociation of GdS from spermatozoa is compatible with its role as an uncapacitation factor. Of note, cervical mucus contains measurable level of glycodelin (19), but its glycosylation pattern and physiological role are unknown.
The present study demonstrates for the first time that GdS neither affects the acrosome reaction nor decapacitates capacitated spermatozoa, but it inhibits BSA-induced capacitation, shown by both CTC-and ZP-induced capacitation assay. Because GdA has no similar activity, this biological activity of GdS is isoform-specific. The concentration of GdS (Ͼ1500 nM) required to elicit the biological activity is compatible with the concentration of GdS in seminal plasma (1700 nM) (13). This suggests that such activity may have physiological relevance. These observations strongly indicate that one of the functions of GdS in seminal plasma is to maintain the uncapacitated state in human spermatozoa.
Coincidentally, recombinant glycodelin from yeast, Pichia pastoris, at a concentration of 1500 nM inhibits capacitation of human and hamster spermatozoa (45). However, the degree of similarity of the glycans of this recombinant glycodelin with GdS is unknown. Glycosylation is important for the biological activity of other glycodelin isoforms (6,8,12,45), and like in the case of the other glycodelin isoforms, the biological activity of GdS is also glycosylation-dependent. Interestingly, carbohydrate residues are also involved in the activity of decapacitation factors in the human spermatozoa (46,47).
It is interesting to note that albumin is present in high concentration in the uterine fluid and follicular fluid, but is virtually absent from the seminal plasma (48). In the follicular fluid, albumin acts as a sterol acceptor to induce cholesterol efflux from the sperm plasma membrane leading to capacitation (33,49). The disappearance of glycodelin immunoreactivity after cervical mucus penetration and the ability of GdS to suppress BSA-induced capacitation are consistent with a regulatory role of GdS on capacitation in vivo, such that the removal of GdS during cervical mucus penetration allows albumin in the uterine fluid to initiate capacitation.
The ability of GdS in inhibiting capacitation is related to suppression of BSA-and M␤CD-induced capacitation. Because BSA might be contaminated by traces of other serum components, M␤CD was used to confirm that the effect of GdS on capacitation was via modulating cholesterol efflux. M␤CD is a non-physiological cholesterol acceptor that promotes cholesterol efflux, tyrosine phosphorylation, and capacitation of spermatozoa (29 -33). M␤CD is more effective than 3% BSA in inducing capacitation. This may be due to the smaller size of M␤CD, which allows a more rapid kinetics of cholesterol removal from cells.
Cholesterol is a plasma membrane lipid component that is important in regulating membrane fluidity and permeability, and the mobility of integral proteins and functional receptors in the membranes. Loss of cholesterol from the plasma membrane initiates capacitation (1,33,49). It begins when the spermatozoa leave the seminal plasma and is associated with a decrease in the cholesterol/phospholipid ratio and an increase in fluidity and permeability of the sperm membrane (1,49). Incubation with exogenous cholesterol reduces cholesterol loss and suppresses progesterone-, calcium ionophore-, and ZP-induced acrosome reaction (32,33,49) and inhibits the fertilizing ability of spermatozoa (50). It has been suggested that cholesterol in the seminal plasma inhibits capacitation (49). This is supported by the observations that cholesterol acceptors such as serum albumin and cyclodextrins stimulate sperm capacitation (30 -33, 49, 51, 52). Pretreatment of cholesterol acceptors with excessive cholesterol blocks the capacitation promoting activity of the acceptors (30,31). Given this background, results of the present study demonstrate that GdS in seminal plasma inhibits capacitation by reducing cholesterol loss.
The present study expands upon previous reports (30 -32, 35) suggesting that capacitation of human spermatozoa may involve the loss of cholesterol and the subsequent increase in cAMP concentration, leading to the activation of protein kinase A and tyrosine kinase and to protein tyrosine phosphorylation. Here, we provide direct evidences on the existence of such signaling pathway by determining the change in intracellular protein kinase activities and cAMP concentration of human spermatozoa after induction of cholesterol efflux. This regulation of protein tyrosine phosphorylation by cAMP and protein kinase A pathway is unique to spermatozoa. It is in contrast to somatic cells, in which the activation of protein tyrosine phosphorylation is mediated through plasma membrane receptors with intrinsic tyrosine kinase activity or associated with tyrosine kinases. This is the first report demonstrating the ability of GdS to suppress 3% BSA-and M␤CD-induced capacitation-associated increases in cholesterol efflux, cAMP concentration, protein kinase activities, and protein tyrosine phosphorylation. This effect was isoform-specific, as GdA was unable to produce such effect even at high concentrations. GdS alone had no effect on these intracellular signaling events, suggesting that the activity of GdS on 3% BSA-and M␤CD-induced changes in protein kinase activities and protein tyrosine phosphorylation is not due to a direct action of GdS on these molecules. The observation that exogenous cAMP agonists, Sp-CAMPS, overcome such activity of GdS further indicates that the effect of GdS is directed upstream of protein kinase activation in the signal transduction cascade.
How GdS prevents cholesterol loss is unknown. The plasma membrane of sperm head has a glycocalyx of 100 -150 Å in thickness with glycoproteins anchoring to the lipid bilayer of the membrane. The glycocalyx retains a layer of immobile water around the sperm head. It has been proposed that cholesterol efflux from membranes involves the entrance of the cholesterol acceptor into the immobile water layer and subsequent binding of the acceptor with cholesterol (53). Therefore, one possibility of GdS action is that it covers the glycocalyx and prevents entrance of cholesterol acceptors, thus reducing removal of cholesterol. As glycodelin has significant homology with another lipocalin member, ␤-lactoglobulin, which binds cholesterol (54), it is also possible that sperm-bound glycodelin-S may bind to cholesterol directly, thereby reducing cholesterol efflux from the sperm plasma membrane. Another possibility is that the binding of GdS on sperm cells causes cholesterol redistribution on the sperm membrane, thereby reducing accessibility of cholesterol to its acceptors. A similar condition exists in bovine seminal plasma. It contains a glycoprotein, PDC-109, that immobilizes membrane lipid, including cholesterol and retards capacitation upon binding to spermatozoa (55). Interestingly, albumin-mediated cholesterol depletion only occurs after bicarbonate induced membrane lipid scrambling (40,56). The effect of GdS on bicarbonate-induced capacitation is under investigation in our laboratory.
Taken together, the present results and previous observations show that tissue-specific glycosylation of glycodelin modulates sperm function for fertilization. GdS in the seminal plasma acts as a natural uncapacitation factor by modulating sperm cholesterol efflux. Based on current data from this and previous reports, the sequence of events appears as follows: GdS is removed from spermatozoa as they migrate across the cervical mucus, initiating capacitation. Subsequently, follicular fluid-derived GdF in oviductal fluid attaches onto the acrosome region of the sperm head and prevents premature acrosome reaction (8,12). GdF is then removed by the cumulus/corona cells (42), whereby progesterone-induced acrosome reaction is restored, and sperm-egg binding capacity is induced initiating the fertilization process.