Bicarbonate stimulated phospholipid scrambling induces cholesterol redistribution and enables cholesterol depletion in the sperm plasma membrane

Mammalian sperm cells released into the lumen of the seminiferous tubules in the testis are immotile and incompetent to interact with the oocyte and its extracellular vestment, the zona pellucida (Toshimori, 1998). The sperm cell requires two physiological maturation phases in order to acquire optimal fertilization properties: (1) epididymal maturation, a process in which severe surface protein and lipid modifications take place (Nikolopoulou et al., 1985; Haidl and Opper, 1997; Toshimori, 1998), which results in the generation of sperm cell motility (Amann et al., 1982); (2) further surface modifications and activation in the female genital tract (especially in the lumen of the oviduct) (Yanagimachi, 1994). In vitro activation of sperm triggers diverse signaling pathways such as cAMP-dependent protein kinase (PKA) and induced protein tyrosine phosphorylation (Visconti et al., 1995a; Visconti et al., 1995b) and leads ultimately to the generation of sperm cells with high binding affinity for the zona pellucida. The sperm activation processes are collectively termed capacitation (Yanagimachi, 1994).

The amount and distribution of cholesterol in the sperm plasma membrane alter upon capacitation and these cholesterol alterations are believed to play a role in modulating signaling pathways in sperm cells (Visconti et al., 1999a; Visconti et al., 1999b). One of the key events in sperm capacitation is the activation of adenylate cyclase by high levels of bicarbonate that are present in in vitro fertilization media, and proposed to be locally enriched in upper parts of the female genital tract (i.e. in the lumen of the oviduct), but virtually absent in epididymal and seminal plasma (Harrison, 1996). Increased cAMP levels activate cAMP-dependent PKAs and indirectly induce protein tyrosine phosphorylation by a yet unknown signaling pathway. Bicarbonate also induces PKA-dependent changes in the lipid architecture of the sperm plasma membrane (Harrison and Miller, 2000), due to phospholipid scrambling (Gadella and Harrison, 2000). These membrane lipid changes can be monitored by merocyanine-540 (M540), which has been used as a probe to monitor bicarbonate activation in individual cells using flow cytometry, and only a subpopulation of sperm cells appeared to be activated (Harrison et al., 1996; Flesch et al., 1999). The role of cholesterol in activation of sperm cells has recently been studied using cyclodextrin, an agent that extracts cholesterol from membranes. It has been demonstrated that sperm cells incubated in the absence of bicarbonate but with cyclodextrin had markedly activated PKA (Visconti et al., 1999a) and enhanced tyrosine phosphorylation levels (Visconti et al., 1999a; Cross et al., 1999). Moreover, cholesterol efflux has been demonstrated during in vitro capacitation by albumin (Langlais et al., 1988; Visconti et al., 1999b) as well as by lipoproteins that originate from oviductal and follicular fluids (under fertilization conditions in vivo the sperm will encounter both components in the oviduct) (Ehrenwald et al., 1990). Ravnik et al. supported this by demonstrating that an active lipid transfer protein I, which is present in the oviduct, supports sperm capacitation but not the acrosome reaction (Ravnik et al., 1995). The bicarbonate- and albumin-mediated lipid changes seem to be related to the initiation of the acrosome reaction (a Ca²⁺-dependent exocytotic multiple fusion event between the apical sperm plasma membrane and the underlying outer acrosomal membrane) (Zarintash and Cross, 1996).

Filipin complexes cholesterol into clusters that can be visualized using freeze fracture techniques and electron microscopy at the level of the sperm surface. Owing to its intrinsic UV fluorescent properties, the lateral distribution can also be followed using fluorescence microscopy. In this study we analyzed the effects of the capacitation factors, bicarbonate, calcium and albumin, on the lateral organization of cholesterol in sperm cells using filipin as a microscopical marker for cholesterol. We also tested their inducing effects on reduction of cellular cholesterol levels by determining the molecular composition of lipid extracts from washed sperm cell pellets after various incubations. Sperm suspensions were also activated by bicarbonate, and non-responding cells were sorted from responding cells that acquired high M540 fluorescence. The lipid composition and filipin labeling was determined in both cell subpopulations.

We report that the cholesterol organization in ejaculated sperm cells is heterogeneous due to variations in the extent of epididymal maturation between individual cells. Sperm cells with low levels of cholesterol are activated by bicarbonate (monitored by M540 fluorescence), which causes a lateral redistribution of cholesterol in these cells. The redistributed cholesterol then becomes available for extraction by albumin. A model is presented to explain this biphasic change in cholesterol organization for sperm cell signaling and its importance for the eventual fertilization of the oocyte is discussed.

Merocyanine-540 (M540), propidium iodide and Yo-Pro 1 were purchased from Molecular Probes Inc. (Eugene, OR), PNA-FITC from EY-laboratories Inc. (San Mateo, CA) and Filipin from Sigma (St Louis, MO). All organic solvents used for lipid analyses were obtained from Labscan (Dublin, Ireland) and were of HPLC grade. Lipid standards were obtained from Avanti Polar Lipids (Alabaster, AL).

Semen was obtained from the Cooperative Center for Artificial Insemination in Pigs ‘Utrecht en de Hollanden’ (Bunnik, The Netherlands). Freshly ejaculated semen was filtered through gauze to remove gelatinous material and diluted, washed and stored in Beltsville Thawing Solution as described previously (Gadella et al., 1999a). All buffers and other solutions used were iso-osmotic (285-300 mOsm) and kept at room temperature unless stated otherwise. Sperm cells were washed through a discontinuous Percoll^® (Pharmacia, Uppsala, Sweden) density gradient as described before (Harrison et al., 1993; Flesch et al., 1998).

The investigations centered on sperm behavior during incubation in one of two media: (1) The ‘control’ medium, Hepes buffered Tyrode’s (HBT: 120 mM NaCl, 21.7 mM lactate, 20 mM Hepes, 5 mM glucose, 3.1 mM KCl, 2.0 mM CaCl₂, 1.0 mM pyruvate, 0.4 mM MgSO₄, 0.3 mM NaH₂PO₄ and 100 μg/ml kanamycin; 300 mOsm/kg, pH 7.4); and (2) the ‘capacitating’ medium, HBT-Bic, that is, HBT containing 15 mM NaHCO₃ in equilibrium with 5% CO₂ in humidified atmosphere (the bicarbonate replaced a molar equivalent of NaCl so that osmolality was maintained). HBT, though relatively physiological, does not induce capacitative changes, whereas HBT-Bic induces capacitative changes (Gadella and Harrison, 2000). The two media were supplemented with one, or a combination of the following components: (1) 2 mM CaCl₂ or 1 mM EGTA; (2) 0.3% (w/v) BSA (defatted fraction V; Boehringer Mannheim, Almere, The Netherlands), or a mixture of 0.5 mg/ml polyvinyl alcohol and 0.5 mg/ml polyvinylpyrrolidone. Sperm suspensions were incubated at 38.5°C in a cell incubator with humidified air containing 5% CO₂. Sperm suspensions in HBT were placed in air tight sealed tubes during incubation, whereas suspensions in HBT-Bic were placed in the opened tubes in the cell incubator.

For flow cytometric purposes, sperm cells were also capacitated in HBT-Bic containing 2.7 μM M540 (a reporter probe for phospholipid scrambling) (Gadella and Harrison, 2000), and 25 nM Yo-Pro 1 (a membrane impermeable nucleic acid stain) (Harrison et al., 1996) and 0.5 mg/ml polyvinyl alcohol and 0.5 mg/ml polyvinylpyrrolidone. In vitro capacitation was performed in airtight sealed 5 ml flow cytometer tubes (Becton Dickinson, San Jose, CA) containing 3 ml medium, which were flushed with air containing 5% carbon dioxide before closing. Capacitation was performed for approximately half an hour in a shaking waterbath at 38.5°C before flow cytometric analysis and sorting. From that time point, built up M540-positive and -negative cell subpopulations were stable, although the amount of M540 staining in positive cells increased somewhat during further incubation.

Sperm cell sorting and analysis were performed on a FACS Vantage SE (Becton Dickinson). The system was triggered on the forward light scatter signal (FSC). Yo-Pro 1 and M540 were both excited by an argon ion laser (Coherent Innova, Palo Alto, CA) with 200 mW laser power at 488 nm. Yo-Pro 1 was measured through a 500 nm long pass filter. M540 emission was deflected with a 560 nm short pass dichroic mirror in the emission pathway and measured through a 575±26 nm band pass filter. Sperm cells were analyzed at a rate between 8000 and 10,000 per second. For each file, 10,000 events were stored in the computer for further analysis with Cell-Quest software (Becton Dickinson) or WinMDI 2.8 (http://facs.scripps.edu/). FSC and sideward light scatter (SSC) were recorded and only sperm cell-specific events, which appeared in a typically L-shape scatter profile (Harrison et al., 1996), were positively gated for further analysis. During sorting the sample-input tube on the FACS Vantage SE was kept at 38.5°C and 5% CO₂ to maintain constant incubation conditions during the complete sorting procedure using a controlled temperature bath/circulator. Sperm cells were run through the machine using PBS as a sheath fluid. Two subpopulations were sorted: (1) sperm cell events that were not stained with Yo-Pro 1 (viable) showing low M540 fluorescence (non-responding cell subpopulation); and (2) viable cells with high M540 fluorescence (responding cells). Sorted cells were collected in precooled 50 ml tubes that were placed in a tube holder which was kept at −20°C. Sorted sperm cell events were also collected immediately on microscopic slides and subsequently examined with a spectral confocal microscope as stated below. Alternatively, sorted sperm cells were collected at room temperature in tubes that were half filled with Karnovsky fixative (2.5% glutaraldehyde, 2% paraformaldehyde, 80 mM Na-cacodylate, 500 μM MgCl₂, 250 μM CaCl₂, pH 7.4) until tubes were three-quarters full and further processed to visualize the surface cholesterol organization with filipin. In order to analyze the efficiency of cell sorting, sperm cells were collected in flow cytometer tubes and rerun within 10 minutes.

The acrosomal status was checked routinely by staining the same incubated sperm samples that were used for sperm sorting as described above with 5 μg/ml fluorescein-conjugated peanut agglutinin (PNA-FITC; as a marker probe for acrosomal leakage) (Flesch et al., 1998) and 1 μm propidium iodide (as marker probe for cell deterioration) (Ashworth et al., 1995), and subsequent analysis on a FACScan (Becton Dickinson) as described before (Szasz et al., 2000). Alternatively 10 μl of the labeled sperm suspension was used to make microscopic slides (Szasz et al., 2000) and 200 cells were counted in triplicate for each treatment on damage of the acrosome or the plasma membrane.

Sperm cells were incubated in various media for 2 hours at 38.5°C in a humidified air of 5% CO₂ and stained with 2.7 μM M540 and 25 nM Yo-Pro 1 for 10 minutes. Aliquots of 250 μl sperm suspension (containing approximately 1 million sperm cells) were placed into a life chamber (37°C, 5% CO₂ in humidified atmosphere) and placed on an epifluorescence microscope (Leica DMRE, Leica GmbH, Germany) equipped with a Hg lamp (100 mW) and a filter block (480 nm excitation filter, 500 nm dichroic mirror and a 520 nm long pass emission filter) in order to simultaneously assess M540 fluorescence (red) and Yo-Pro 1 fluorescence (green). From each sperm sample, 200 cells were counted in triplicate. For visualization (sorted), sperm samples were placed under a spectral confocal microscope (Leica TCS SP) and excited with the 488 nm Argon laser line. Yo-Pro 1 fluorescence was detected with photomultiplier tube 1 (emission selected in the wavelength range of 500-550 nm) and M540 fluorescence with photomultiplier tube 2 (580-620 nm). Single scans were made to capture labeling patterns of hyperactivated sperm cells.

Sperm cell suspensions that were incubated for 2 hours in HBT or HBT-Bic (either in the absence or presence of 0.3% (w/v) BSA were further subjected to lipid extraction as described (Bligh and Dyer, 1959). All sperm suspensions were washed through a 30% Percoll cushion prior to lipid extraction because this procedure was required to separate BSA from the sperm cells (10 minutes 700 g). Alternatively, sperm cells that were sorted for low and high M540 fluorescence and collected in tubes at −20°C were centrifuged (285,000 g, 70 minutes, 2°C), supernatant was discarded and the lipids from the resulting cell pellets were extracted. The composition of lipid classes of total sperm populations was detected by high performance liquid chromatography (HPLC) that consisted of an LKB low pressure mixer, a model 2248 pump (Pharmacia, Uppsala, Sweden), and a Rheodyne injector. Lipid classes of sorted sperm cell populations were separated on a Lichrospher DIOL-100 column (250×3.2 mm, 5 μm particle size) obtained from Alltech Applied Sciences (Breda, The Netherlands). Elution was performed at 40°C using an adapted method (Silversand and Haux, 1997). In brief, lipid classes were eluted with a ternary gradient using the solvents hexane/acetone 99/1 v/v (A), hexane/2-propanol/acetone 82/17/1 v/v/v (B), and 2-propanol/water/acetone 85/14/1 v/v/v (C). The gradient was developed as follows: (time in minutes, %A,%B,%C); (0, 90, 10, 0); (10, 57, 43, 0); (11, 20, 70, 10); (15, 0, 80, 20); (38, 0, 60, 40); (40, 0, 60, 40); (45, 0, 100, 0); (49, 90, 10, 0); (55, 90, 10, 0). Lipids were detected using a Varex MKIII light scattering detector obtained from Alltech (Deerfield, IL). The detector was calibrated with lipid standards at a drift tube temperature of 90°C and a gas flow of 1.8 l/min (Brouwers et al., 1998). Lipid classes and molecular species composition were determined by online electrospray ionization mass spectrometry on a Sciex API-365+ triple quadrupole mass spectrometer (PE Biosystems, Nieuwerkerk a/d IJssel, The Netherlands).

The apical plasma membranes were isolated from sperm suspensions after a 2 hour incubation in either HBT, HBT-bic or HBT-Bic containing 0.3% (w/v) BSA as described (Flesch et al., 1999; Flesch et al., 2001). The phospholipid classes and cholesterol content were detected using the HPLC method described above.

Sperm suspensions were immediately fixed after capacitation incubations by diluting 1:1 with 4% glutaraldehyde, 150 mM Na-cacodylate buffer pH 7.4. Sperm samples were fixed for 30 minutes under gentle shaking. Before staining, cell suspensions were washed twice (500 g, 15 minutes) in 0.15 M Na-cacodylate buffer (pH 7.4). After washing, sperm cells were resuspended in 0.15 M Na-cacodylate buffer containing 25 μM filipin (an antibiotic that can aggregate unesterified sterols into complexes) (Friend, 1982). Filipin was dissolved in the buffer from a 10 mM DMSO stock solution. Blanc samples were treated with similar DMSO concentrations without filipin. Tubes were wrapped in aluminum foil to keep the fluid in the dark and labeling was performed for 30 minutes under gentle shaking. Subsequently, tubes were centrifuged (500 g, 15 minutes) and washed in 0.15 M Na-cacodylate buffer. Cell pellets were mixed with 30 vol% EM-grade glycerol and 70 vol% 0.15 M Na-cacodylate buffer for 1 hour under gentle shaking. Ultrastructural localization of filipin labeling was evaluated by electron microscopy. For this purpose 1 μl aliquots of sperm suspensions were pipetted on golden heads and were frozen in a mixture of liquid and solid nitrogen. The cells were stored in liquid nitrogen until further processing. Cells were freeze fractured in a Balzers BAF-300 device at −105°C and a vacuum of 10⁻⁷ Torr. A coat of 200 nm platinum/carbon was evaporated under an angle of 45° followed by a second coat of carbon at 90°. The cells were digested overnight in household bleach and the replicas were examined in a Philipis CM 10 electron microscope (Philips, Eindhoven, The Netherlands). The freeze fracture procedure was further performed as described previously (De Leeuw et al., 1990).

The fluorescent properties of filipin immobilized to the sperm cells were analyzed in an LS50 luminescence spectrophotometer (Perkin Elmer Ltd, Beaconsfield, UK). Emission scans were made at 357 nm excitation, in the range of 400-600 nm. Excitation scans were made at 480 nm, in the range of 275-400 nm (in all cases, 5 nm slit width settings were used). Filipin fluorescence was also observed on a fluorescence microscope (Leica DMRE) equipped with a Hg lamp (100 mW) and a UV filter block (340-380 nm excitation filter, 400 nm dichroic mirror and a 425 nm long pass emission filter). Specimens were mounted and coverslips were sealed with nail polish for fluorescence microscopical inspection. Fluorescence patterns of filipin cholesterol complexes were observed in three different sets of boar sperm samples that were incubated for 2 hours at 38.5°C in humidified air: (1) in HBT with and without 0.3% (w/v) BSA; (2) in HBT-Bic with and without 0.3% (w/v) BSA (BSA replaced for 0.5% (w/v) PVP in combination with 0.5% (w/v) PVA; and (3) in HBT-Bic, subsequently sorted for low and high M540 fluorescence, and collected in fixative. From each sample, 200 cells were counted in triplicate.

Sperm suspensions from incubated specimens (total cell population as well as the low and high M540 sorted subpopulations) were diluted to a concentration of 1 million cells per ml, fixed as described above, and 200 cells were counted in triplicate from each preparation for three cell morphology types: (1) normal well-matured sperm cell without cytoplasmic remnants; (2) normal but poorly matured sperm cell containing visible cytoplasmic droplets; (3) deteriorated sperm cell or sperm cell exhibiting abnormal morphology. Sperm morphology was scored under an Olympus 209376 Phase-Contrast microscope (100× objective, 10× ocular; Olympus, Tokyo, Japan). Sperm suspensions were routinely assessed for acrosomal integrity as described before (Flesch et al., 1999).

Ejaculates of three different boars were examined in triplicate after incubation at 38.5°C for 2 hours. The effect of medium composition on lipid composition, morphology, acrosomal status and capacitation was analyzed using ANOVA in combination with Bonferroni’s multiple comparison test. Lipid compositions of sorted sperm cells were statistically compared using a paired Student t-test.

Lipids were extracted from sperm samples after a 2 hour incubation period in HBT and HBT-Bic media. The lipid composition (including the cholesterol concentration) remained unaltered in HBT-Bic-treated sperm cells when compared with the HBT-treated sperm cells (Fig. 1a). Addition of 0.3% (w/v) albumin to HBT-Bic media resulted in a marked decrease in cellular levels of cholesterol and this effect was calcium independent (Fig. 1a). The albumin-mediated extraction of cholesterol was dependent on bicarbonate because there was no drop in cellular cholesterol levels in HBT-treated cells in the absence of bicarbonate (Fig. 1a). Albumin specifically extracted cholesterol from the bicarbonate-activated sperm cells since no changes in other lipid classes were detected (Fig. 1b). In a parallel experiment, sperm cells were first incubated for 2 hours in HBT, or HBT-Bic in the absence or presence of albumin and further processed to isolate plasma membranes as described before (Flesch et al., 1998; Flesch et al., 1999). The results clearly demonstrate that the albumin/bicarbonate-mediated extraction of cholesterol was mediated at the level of the sperm plasma membrane and there was no effect on other lipid classes (Fig. 1c). The relative proportion of the other lipid classes increased somewhat due to the lower amount of cholesterol.

Sperm cells were incubated in HBT-Bic in the presence of M540 (to detect bicarbonate-mediated scrambling of phospholipids) (Gadella and Harrison, 2000) and Yo-Pro 1 (to distinguish viable from deteriorated cells) (Harrison et al., 1996) in a flow cytometer tube (38.5°C) at the sample-input of the FACS Vantage SE flow cytometer. Sperm cells were continuously run through the flow cytometer. The subpopulation of viable sperm-specific events that acquired high M540 fluorescence and the subpopulation that remained low fluorescent for M540 were sorted into two collection tubes (Fig. 2a). Rerun of these two viable sperm cell subpopulations through the flow cytometer within a time period of 10 minutes revealed that the cells remained viable and did not change M540 fluorescence characteristics (Fig. 2b,c). The efficiency of sorting was >96%. Labeling patterns of M540 were established within 15 minutes and these labeling patterns did not change during the course of the sorting experiments. However, a steady build-up of M540-responding cells and a decrease in non-responding cells appeared after longer incubations.

The viable sperm events sorted for low and high M540 fluorescence were immediately analyzed under a confocal microscope (Fig. 2d-f). Unsorted cells were either low or high fluorescent for M540 (for the sperm sample depicted in Fig. 2d, approx. 54% of the viable cells had high M540 fluorescence), whereas >98% of the cells sorted for low M540 fluorescence had the M540 labeling pattern shown in Fig. 2e, and >98% of the cells sorted for high M540 fluorescence had the M540 labeling pattern shown in Fig. 2f. During the sorting experiment the sperm cells remained acrosome intact and viable, as was detected by staining the cells with PNA-FITC in combination with propidium iodide (the incidence of damaged acrosomes and/or cell deterioration was <10%).

Lipids were extracted from viable sperm cell subpopulations with either low or high M540 fluorescence. Total lipid extracts were analyzed after a single HPLC run using a light scattering detector (Fig. 3a). The lipid composition of the sperm cell membranes is given in Table 1. Although appreciable differences in lipid composition were observed between boars, a paired t-test revealed no significant difference in the composition of lipid classes between cells with high and low membrane fluidity from the same boar (P>0.05). However, the difference in cholesterol levels was highly significant (P<0.01). To investigate whether the observed differences in membrane fluidity were related to differences in the fatty radyl moieties of the phospholipids, online electrospray mass spectrometry was performed on the lipid classes as they eluted from the column (depicted for PC species in Fig. 3b). However, no differences in the molecular species composition were observed between cells with high and low M540 fluorescence (Fig. 3b).

Sperm cells were labeled with filipin in order to reveal the lateral organization of cholesterol and other free sterols at the sperm surface (Friend, 1982). Filipin aggregates sterols into complexes that can be distinguished on freeze fracture replicas of the sperm plasma membrane (Fig. 4a,b), whereas unlabeled cells were devoid of such complexes (Fig. 4c; the small dots represent trans-membrane proteins). Two different types of filipin labeling were observed (Fig. 4a,b; labeled type A and B, respectively). However, due to the fact that the fracture plane only rarely runs through extended areas of the sperm head plasma membrane, this very laborious technique cannot be used to assess the relative frequencies of cells with one of each filipin labeling pattern. Filipin formed complexes only with sterols at the sperm surface (Fig. 5).

Filipin contains a set of five conjugated trans double bonds (-CH=CH-; Fig. 6, inset) and therefore can be used as a UV fluorescent probe. The excitation and emission properties of filipin immobilized to sperm cells are depicted in Fig. 6. The excitation peak of 357 nm in combination with the emission peak at 480 nm makes this probe suitable for UV-fluorescence detection. In fact UV-fluorescence detection of filipin sterol complexes on sperm cells revealed surface labeling patterns similar to those detected with electron microscopy on freeze fracture replicas (compare Fig. 4a, pattern A with Fig. 7A; and Fig. 4b, pattern B with Fig. 7B). When sperm samples were treated with HBT-Bic in combination with 0.3% (w/v) albumin, a subpopulation of cells showed negative filipin labeling (Fig. 7C, pattern C). The filipin pattern of Fig. 7C is difficult to examine using electron microscopy on freeze fracture replicas because albumin further destabilizes the sperm plasma membrane, which hampers fracture planes to run through extended plasma membrane areas of the sperm head. The relative frequencies of the three types of filipin labeling were scored by UV-fluorescence microscopy for six individual sperm ejaculates. The transition from filipin pattern A to B reflects the cholesterol redistribution (Fig. 8). Clearly, bicarbonate was necessary to induce cholesterol redistribution, whereas the inclusion of BSA did not affect cholesterol redistribution. However, both bicarbonate and BSA were necessary to induce cholesterol efflux as reflected by pattern C (Fig. 8).

Phospholipid scrambling was followed by staining the same sperm samples with M540 (Gadella and Harrison, 2000) and relative frequencies of low and high fluorescent M540 cells were scored. Sperm suspensions incubated in HBT almost exclusively expressed low M540 fluorescence, whereas a variable but considerable shift to high M540 fluorescence was observed in sperm suspensions treated with HBT-Bic (Fig. 8). It should be noted that inclusion of 0.3% (w/v) albumin did not change this distribution in labeling pattern frequencies and therefore did not affect phospholipid scrambling (Flesch et al., 1999). Moreover, viable sperm subpopulations that were sorted for low and high M540 fluorescence showed the corresponding filipin labeling: >95% of the low M540 fluorescent sorted cells showed filipin labeling pattern A, whereas >90% of the high M540 fluorescent sorted cells showed filipin labeling pattern B (or in the presence of BSA, pattern C; Fig. 8, Fig. 9). Collectively, these data indicate that the membrane architectural changes picked up by M540 probe allow albumin-mediated cholesterol efflux.

Cell morphology of the same sperm samples was scored under a phase-contrast microscope. The sperm morphology (cytoplasmic droplet content and acrosome morphology) did not change under the different HBT and HBT-Bic incubations tested (data not shown). It was noted that the highly variable M540 and filipin response correlated very well with the degree of epididymal maturation of the sperm cells (Fig. 9). The proportion of sperm cells that were devoid of cytoplasmic droplets correlated very well with the proportion of sperm cells that responded to bicarbonate (both with respect to the filipin response and to the M540 response; Fig. 9). The presence of a large population of sperm cells with cytoplasmic droplets is considered to be an indication of sub-optimal epididymal maturation (Briz et al., 1995). Therefore, variations in bicarbonate-induced effects in the ejaculated sperm samples may well relate to differences in epididymal maturation efficiencies.

In this paper, the dynamics of cholesterol organization in the plasma membrane of bicarbonate-activated sperm cells was assessed. Filipin labeling studies revealed that bicarbonate induced a redistribution of cholesterol (and other free sterols present only in trace amounts) in the sperm head from a more general distribution (Fig. 7A) to the apical site (Fig. 7B). Furthermore, cholesterol was extracted from sperm cells incubated with bicarbonate if an acceptor (e.g. BSA) was present (Fig. 7C). Bicarbonate plays a major role in the activation of sperm cells (Boatman and Robbins, 1991; Suzuki et al., 1994; Shi and Roldan, 1995; Visconti et al., 1995a; Visconti et al., 1995b). In pigs, bicarbonate brings about modifications in sperm surface coating (Ashworth et al., 1995), a PKA-mediated scrambling of phospholipids (Gadella and Harrison, 2000), induction of protein tyrosine phosphorylation (Flesch et al., 1999) and an increase in the ability of viable acrosome-intact cells to bind to zona pellucida components (Harkema et al., 1998). Of particular interest has been the observation that bicarbonate-induced scrambling of phospholipids in boar sperm cells can be monitored with M540 (Gadella and Harrison, 2000). In other cell types, an increase in M540 has also been interpreted as indicative for a loss of transverse phospholipid asymmetry in the plasma membrane (Williamson and Schlegel, 1994; Verhoven et al., 1995). In this paper, the effect of bicarbonate on the organization of cholesterol was measured because several studies have indicated that the lateral distribution of cholesterol (Friend, 1982) and lipoprotein cholesterol removal from the plasma membrane (Ravnik et al., 1995; Visconti et al., 1999b) are important events in sperm capacitation and modulate cell signaling molecules (Visconti et al., 1999a; Visconti et al., 1999b).

Percoll-washed sperm suspensions were incubated with HBT-Bic and stained with M540, and viable cells were sorted for high and low M540 fluorescence (Fig. 3). The phospholipid composition of the high and low M540 sperm subpopulations was nearly identical in the two sperm subpopulations, whereas the high M540 subpopulation had 30% reduced cholesterol levels when compared with the low M540 subpopulation (Table 1). In the absence of albumin, bicarbonate did not affect the lipid composition and cholesterol levels in complete sperm suspensions (Fig. 2). Therefore, it can be concluded that bicarbonate did not induce a reduction in cholesterol in the high M540 cells (the sorting experiments were performed in the absence of albumin). This implies that individual ejaculated sperm cells contain variable amounts of cholesterol and that the cells with low cholesterol levels were primed by bicarbonate (as detected with M540). It should be noted that the high and low cholesterol-containing cells (i.e. the M540 responding and non-responding cells, respectively) contained an identical phospholipid composition and unsaturation degree in fatty acids attached to phospholipids (Fig. 3). Therefore, the bicarbonate-induced M540 response depended only on the intrinsic level of cholesterol in each individual cell at ejaculation. With this respect it is interesting to note that the M540 response reflects the scrambling of phospholipids in the sperm plasma membrane (Gadella and Harrison, 2000). Bicarbonate appears to stimulate directly a sperm-specific adenylate cyclase (Garty and Salomon, 1987, Chen et al., 2000), which results in increased cAMP, which activates PKA to initiate one or more as yet unidentified protein phosphorylation cascades resulting in the phosphorylation of protein tyrosine residues (Visconti and Kopf, 1998; Flesch et al., 1999). Our results imply that only low cholesterol levels in sperm cells allow all these bicarbonate-triggered events. The variety of cholesterol levels is probably a result of differential efficiencies in epididymal maturation. (1) Severe modifications in the lipid composition take place during this process, including a decrease in cholesterol (Rana et al., 1991; Haidl and Opper, 1997). (2) Sperm cells with uncomplete matured morphology (i.e. with cytoplasmic remnants) (Briz et al., 1995) never acquired high M540 fluorescence after incubation in HBT-Bic. In fact, the relative number of incomplete matured sperm cells (as scored on cytoplasmic remnants) correlates very well with the amount of cells that did not show high M540 fluorescence after incubation in HBT-Bic (Fig. 9). Furthermore, the sorted high M540 fluorescent sperm cell subpopulation did not contain cells with incompletely matured morphology. Taken together, these data indicate that differences in epididymal maturation are responsible for the variations in cholesterol levels in individual ejaculated sperm cells. Only the cells with relatively low cholesterol levels are primed by bicarbonate (resulting in high M540 fluorescence). It is most likely that the boar variation in M540 response to bicarbonate (Harrison et al., 1996) is due to boar to boar variation in epididymal sperm maturation.

Cholesterol is distributed non-randomly in and between biological membranes and can form lateral domains in the plasma membrane in a variety of cell types (Schroeder et al., 1996) including sperm cells (Friend, 1982). It is believed that the dynamics in these cholesterol domains play a role in receptor effector coupling, membrane ion transport and membrane-mediated cell signaling (Sweet and Schroeder, 1988). In sperm cells, cholesterol seems to modulate PKA activity and protein tyrosine phosphorylation (Visconti et al., 1999b). Obviously, it was of interest to detect the membrane organization of cholesterol in capacitating sperm cells. For this purpose, sperm cells were labeled with filipin, a polyene antibiotic that forms 25-30 nm complexes with 3-β-hydroxysterols (including cholesterol), which are visible in freeze-fractured membranes (Verkleij et al., 1973) (Fig. 5). The lateral distribution of filipin-sterol complexes on sperm cells could be alternatively visualized by UV-fluorescence due to the intrinsic fluorescent properties of filipin (Fig. 6). Both visualization techniques gave similar labeling results (compare Fig. 4a,b with Fig. 7A,B). Sperm cells incubated in HBT showed predominantly labeling pattern A, whereas incubation in HBT-Bic caused a shift of a subpopulation to labeling pattern B (Fig. 8). Interestingly, sperm cells incubated in HBT-Bic that were sorted for high M540 fluorescence showed almost exclusively filipin labeling pattern B, whereas cells that were sorted for low M540 fluorescence showed labeling pattern A. This indicates that the bicarbonate-mediated M540 response that occurred only in sperm cells with relatively low cholesterol levels (Fig. 2, Fig. 3), induced a reordering of cholesterol in these cells (in the absence of bicarbonate only a few cells appeared with labeling type B). The M540 response closely relates to the induction of phospholipid scrambling (Gadella and Harrison, 2000) as well as the shift and/or depletion response detected with filipin (Fig. 9). Therefore, topological redistribution of cholesterol probably relates to lateral (and perhaps transbilayer) membrane cholesterol redistributions. It should be noted that cholesterol redistributions may occur after fixation. Therefore, the lateral distribution of filipin may not reflect the natural cholesterol organization in living cells. However, one can conclude from the labeling patterns of sperm cells incubated in HBT compared with HBT-Bic, that bicarbonate induces rearrangement in cholesterol ordering.

The changes in cholesterol organization upon bicarbonate activation of sperm cells enable albumin-mediated cholesterol extraction. Sperm cells incubated in HBT supplemented with albumin showed no alterations in cholesterol distribution (Fig. 8) nor a reduction in cholesterol levels or other modifications in lipid composition (Fig. 1). However, when sperm cells were stimulated in HBT-Bic supplemented with albumin, a marked reduction in cholesterol levels was noted (Fig. 8). Moreover, a subpopulation of sperm cells with filipin labeling pattern B shifted to very low filipin labeling (pattern C; Fig. 7C). These data collectively indicate that cholesterol is extracted from the bicarbonate-stimulated sperm subpopulation only. Albumin mediated the cholesterol efflux at the cell surface (Fig. 3c). The importance of lipoprotein-mediated cholesterol extraction in mammalian fertilization has been widely studied since the mid-1980s (Cross, 1998). In fact, incubation media used to capacitate sperm cells in vitro (IVF media) or in vivo (oviduct fluid) must contain cholesterol acceptor proteins for optimal fertility results (Ehrenwald et al., 1990; Yanagimachi, 1994; Ravnik et al., 1995). More recently, the modulating effect of cholesterol efflux in signal transduction in rodent sperm has been reported (Visconti et al., 1999a; Visconti et al., 1999b).

In summary, bicarbonate was shown to have a biphasic effect on cholesterol organization in sperm cells: (1) filipin labeling pattern changed from a homogeneous to an apical distribution pattern in the subpopulation of sperm cells with low cholesterol levels; and (2) only the bicarbonate-responsive cells were susceptible to lipoprotein-mediated cholesterol extraction, leading to sperm cells with very low filipin staining. It should be noted that rodent sperm cells are collected from the cauda epididymides, a site where sperm cells are stored and undergo final maturation changes (Briz et al., 1995). During collection of rodent ejaculates, sperm cells are mixed with the spermicidal coagulation plug (Eddy and O’Brien, 1994). This makes rodents a less suitable model system for monitoring the biphasic effects of bicarbonate on cholesterol organization and cell signaling; non-rodents are more favourable species for such studies. For instance, ejaculated porcine sperm can be stored for up to a week in dilution buffers and can easily be worked up for capacitation in vitro (Harrison, 1996). The bicarbonate-mediated changes in cholesterol topology coincide with a partial scrambling in phospholipid asymmetry (this study) (Gadella and Harrison, 2000), which is driven by the activation of PKA (Harrison and Miller, 2000). Interestingly, there is considerable evidence that maximal activated adenylate cyclase activity may require an optimal transbilayer fluidity gradient (Sweet and Schroeder, 1988), which may explain why only sperm cells with low cholesterol (i.e. with higher membrane fluidity) were activated by bicarbonate (which directly activates adenylate cyclase) (Okamura et al., 1985; Garty and Salomon, 1987) (Chen et al., 2000). The data indicate that the membrane changes detected by M540 (i.e. phospholipid scrambling and cholesterol redistribution) are required for albumin-mediated cholesterol extraction. The lateral concentration of cholesterol into the apical plasma membrane of the sperm head may indicate that a membrane raft is formed at this surface area (Simons and Toomre, 2000; Smart et al., 1999). Based on this possibility, we have proposed a model in which the biphasic behaviour of cholesterol (lateral concentration followed by a cholesterol efflux) is explained in this raft theory (Fig. 10). The presence of caveolin-1 in the sperm head (B.M.G. et al., unpublished) as well as the relocalisation of PH-20 (also called 2B1), a GPI-anchored plasma membrane protein, to the apical surface area of sperm head (Seaton et al., 2000) may indeed indicate that an apical membrane raft is formed in the sperm cell. In Fig. 10 it is proposed that the bicarbonate-induced phospholipid scrambling (which is exclusive for the apical head plasma membrane of capacitating intact sperm cells) (Gadella et al., 1999b) is required for this raft formation (Fig. 10B). Within the formed apical membrane, raft scavenger receptors may enable the transport of cholesterol out of the membrane lipid bilayer (in analogy with membrane rafts from other cell types) (Fielding and Fielding, 2000). This will result in effective depletion of cellular cholesterol levels only if an extracellular cholesterol acceptor is present (Fig. 10C). Activation of sphingomyelinase (Gadella and Harrison, 2000) may contribute to the efflux of cholesterol (Cross, 2000).

The role of cholesterol as a cell signaling modulator appears to be an important general cell biological phenomenon (Schroeder et al., 1996; Tepper et al., 2000). With respect to sperm physiology, it is reported that boar sperm cells incubated in bicarbonate-depleted capacitation medium (Flesch et al., 2001), and cytoplasmic droplet containing bovine sperm incubated in bicarbonate-enriched capacitation medium (Thundathil et al., 2001) failed to bind to the zona pellucida. Therefore, the described changes in cholesterol organization are probably required for final, functionally prepared sperm cells to interact with the zona pellucida (by allowing tyrosine phosphorylation of the zona receptors) (Flesch et al., 2001) and enable one of them to fertilize the female gamete.

We thank G. Arkestein and Ing C. van Ballegooijen for expert technical support. B.M.G. is a senior academy fellow of the Royal Dutch Academy of Sciences and Arts (KNAW). F.M.F. was a PhD student financed by the Graduate School Animal Health. P. de Ley and C. Eijndhoven are thanked for artwork.