Sperm selection and competition in pigs may be mediated by the differential motility activation and suppression of sperm subpopulations within the oviduct

Sperm motility is widely recognised as a significant and general determinant of male reproductive success. However, species differences in the events preceding fertilisation have led to the evolution of sophisticated motility control mechanisms that influence the relative reproductive success of males. In freshwater fishes, sperm motility is activated by sudden exposure to water and spermatozoa must reach and fertilise eggs within a very short activation time. In this situation the ability to produce faster spermatozoa provides a fertility advantage, as demonstrated by Gage et al. in studies of Atlantic salmon (Gage et al.,2004). In similar vein, Creech et al. showed that nitric oxide produced near the micropyle of fathead minnow oocytes stimulates sperm velocity, providing a further advantage to those spermatozoa with potential for faster motility (Creech et al.,1998). It is also becoming clear that modification of flagellar beating patterns in response to external signalling factors is the basis of sperm chemotaxis across a wide range of species, from species with external fertilisation such as sea urchins and siphonophores to those whose fertilisation occurs internally (Ralt et al., 1994; Sun et al.,2005; Xiang et al.,2004; Xiang et al.,2005).

In species with internal fertilisation, the significance of sperm motility is made more complex because the spermatozoa encounter different environments within the male and female reproductive tract, and are exposed to factors with cell signalling capabilities. Moreover, the initial stages of sperm transport are mediated by the female tract itself. Sperm transport from the vagina,through the cervical canal and towards the oviduct, is mediated by a combination of intrinsic sperm motility and peristaltic movements of the female reproductive tract. However, the entrance to the oviduct, the uterotubal junction, acts as a selective barrier to sperm transport. Several studies have indicated that in order to reach the oviductal environment,spermatozoa must be physically intact [i.e. possess intact plasma membrane and acrosome (Ellington et al.,1999; Esponda and Moreno,1998)], functionally competent at the molecular level(Nakanishi et al., 2004) and appropriately motile (Herrmann et al.,1999; Olds-Clarke,1996).

Heterospermic insemination studies using mixed semen from two or more males have consistently demonstrated that conception rates are skewed in favour of certain individuals [cattle (Beatty et al.,1969; Beatty et al.,1976; Stewart et al.,1974), pigs (Berger et al.,1996; Stahlberg et al.,2000), rabbits (Vicente et al., 2004)] even though equivalent numbers of live sperm have been used. This situation may mimic the processes of sperm competition, a phenomenon that occurs among many species, including mammals(Birkhead, 1995; Parker, 1970). Physiological explanations for this paternity skewing effect still remain unsatisfactory when interpreted in terms of standard sperm function tests such as motility and viability.

In her study of heterospermic insemination and skewed paternity in pigs,Berger et al. found a correlation between dominance in conception and sperm-egg binding ability (Berger et al.,1996). This implies that ability to reach the eggs should be correlated with ability to traverse the uterotubal junction and to negotiate the oviduct; it is not yet clear whether these abilities are actually correlated. Detailed studies of sperm motility in mice have, nevertheless,shown that the uterotubal junction effectively blocks the transit of spermatozoa with defective flagellar function(Herrmann et al., 1999; Olds-Clarke, 1986).

These findings emphasise the importance of intrinsic sperm motility in populating the oviductal reservoir. Upon ejaculation, mammalian spermatozoa are briefly exposed to seminal plasma but then enter the uterine environment prior to reaching the oviduct. Once within the oviductal isthmus, significant modification of motility occurs as this region is involved in storing spermatozoa for periods of varying duration prior to fertilisation(Birkhead and Moller, 1993). Observations of spermatozoa taken from the isthmic reservoir of several species have consistently shown that their motility is suppressed (e.g. Bedford and Breed, 1994; Burkman et al., 1984; Overstreet and Cooper, 1975; Overstreet et al., 1980) and motility modulation has also been observed in spermatozoa exposed to oviductal fluid (McNutt et al., 1994). The sperm reservoir within the oviductal isthmus therefore contains a selected and highly fertile subpopulation of the original ejaculate(Ellington et al., 1999; Gualtieri and Talevi, 2003; Suarez, 1998) and it is unsurprising that measures of motility based on whole ejaculates show little correlation with fertility under competitive conditions such as heterospermic insemination.

Previous experiments have shown that boar spermatozoa are highly sensitive to the presence of bicarbonate, responding with rapid motility activation(Harrison et al., 1996a; Harrison and Miller, 2000; Holt and Harrison, 2002; Okamura et al., 1985; Tajima et al., 1987) and modification of membrane lipid architecture(Harrison, 2004). When semen samples are washed using a bicarbonate-free Percoll density gradient the viability of the resultant population is typically 80–90%; however, the cells are often quiescent. Adding >5 mmol l^–1bicarbonate/CO₂ to the sperm suspension without changing external pH, rapidly (<2 min) stimulates the motility of some, but not all,spermatozoa. This mimics the situation within the porcine female reproductive tract, where entry to the oviduct is accompanied by an elevation of bicarbonate concentration from <5 mmol l^–1 in the uterine horns to approximately 35 mmol l^–1 in oviductal fluid(Rodriguez-Martinez et al.,2005) (J. M. Vazquez, personal communication). Here we investigate the relative proportions of such bicarbonate-sensitive spermatozoa in different ejaculates and examine the possibility that these might represent the functionally privileged population selected for entry into the oviduct. However, if sperm motility suppression is induced in the oviduct, the model would be incomplete without also investigating the interaction with oviductal factors. We therefore tested the effects of a soluble fraction of oviductal epithelial cell apical plasma membrane proteins (APM), which is known to enhance sperm survival in vitro(Elliott et al., 2001; Fazeli et al., 2000; Fazeli et al., 2003; Holt et al., 2005), on the degree of motility stimulation induced by bicarbonate/CO₂.

To examine these effects in detail we analysed trajectory data from individual spermatozoa using pattern analysis(Abaigar et al., 1999); this analysis allows us to examine and compare the behaviours of sperm subpopulations within ejaculates and to see how the population responses vary between individual boars. In essence, the study confirms the existence of a bicarbonate-sensitive sperm subpopulation within each ejaculate and shows that the oviductal protein fraction suppresses sperm activation in a dose-dependent manner. Interestingly, the oviductal proteins also modify the nature of sperm motion, higher doses leading to increasingly linear trajectories. We also confirmed that the oviductal proteins do not suppress motility by blocking bicarbonate/CO₂ transport; on the contrary it was found that they caused enhanced intracellular alkalinisation, a process associated with the onset of capacitation (Zeng et al.,1996).

Sperm-rich fractions of semen were collected from fertile boars (Sus scrofa Linnaeus 1758) from the colony kept by JSR Genetics Limited at Thorpe Willoughby (Selby, Yorkshire, UK) for commercial artificial insemination; various breeds were represented. The semen samples were supplied(via Royal Mail Special Delivery, arriving the day after semen collection)already diluted in Beltsville thawing solution (BTS) extender, and were stored for no more than one more day.

Prior to experimentation, sperm were isolated from the diluted semen by sedimentation through a two-step Percoll gradient. 7-ml aliquants of diluted semen were layered over 2 ml 35% Percoll on 2 ml 70% Percoll; the Percoll suspensions were iso-osmotic, prepared according to the method of Vincent and Nadeau (Vincent and Nadeau,1984) in a saline medium: 137 mmol l^–1 NaCl, 10 mmol l^–1 glucose, 2.5 mmol l^–1 KOH, 20 mmol l^–1 Hepes, pH 7.4 at 20°C. Centrifugation was performed for 15 min at 200 g_max followed by 900 g_max. After centrifugation, the supernatant layers were removed by aspiration to leave about 0.4 ml of the 70% Percoll, in which the loose sperm pellets were resuspended. These preparations (final concentration approximately 4×10⁸ sperm ml^–1, viability >90%) were kept at ambient temperature protected from draughts and used within 2 h of washing.

A basal Tyrode's-based incubation medium (M solution)(Harrison et al., 1996b)consisted of 116 mmol l^–1 NaCl, 3.1 mmol l^–1KCl, 0.4 mmol l^–1 MgSO₄, 0.3 mmol l^–1 NaH₂PO₄, 5 mmol l^–1 glucose, 21.7 mmol l^–1 sodium lactate, 1 mmol l^–1 sodium pyruvate, 1 mmol l^–1 EGTA,20 mmol l^–1 Hepes (adjusted with NaOH to pH 7.6 at 20°C),3 mg BSA ml^–1, 100 μg kanamycin ml^–1, and 20 μg phenol red ml^–1; the final pH at 38°C was 7.4 and the osmolality 300 mOsmol kg^–1. Bicarbonate/CO₂ was added in the form of suitable aliquants of a 300 mmol l^–1 aqueous solution of NaHCO₃ saturated with 100% CO₂ (a ratio of bicarbonate:CO₂ yielding,after dilution, pH 7.4 at 38°C; addition of this stock did not disturb the pH of the medium). Control incubations received similar aliquants of 300 mmol l^–1 NaCl unless stated otherwise. To prevent loss of CO₂ during subsequent incubation, the bicarbonate-containing suspensions were maintained under a CO₂-containing atmosphere. 4,4′-diisothiocyanatostilbene-2,2′-disulphonic acid (DIDS) was prepared as a 1 mmol l^–1 stock solution in M solution and added as appropriate to the experimental tubes (final concentrations 100μmol l^–1 or 600 μmol l^–1) prior to the addition of spermatozoa.

Measurements of intracellular pH (pH_i) were undertaken using the H⁺ -sensitive fluoroprobe 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein,acetoxymethyl ester (BCECF) following the procedures described by Browning and Wilkins (Browning and Wilkins,2004) and Santi et al. (Santi et al., 1998). Fluorescence ratios were measured (using a Gemini Spectramax microplate reader; Molecular Devices Ltd, Wokingham, UK; excitation wavelengths 490 nm and 440 nm; emission 535 nm) in triplicate, from 100 μl cell suspensions in 96-well plates. Calibration of the procedure was undertaken using a nigericin-high K⁺ technique over the pH range 6.2–8.0. A linear calibration curve (pH vs R_BCECF)was constructed for each experiment; R² for acceptable calibration curves always exceeded 0.995.

Oviducts were obtained from a local abattoir and placed in ice within 20 min of slaughter. These were transported to the laboratory in ice cold PBS and separated from the remainder of the reproductive tissue, washed five times in PBS and kept on ice for the duration of the procedure. As previous studies in this laboratory have shown that apical plasma membrane preparations (APM) and solubilised APM (sAPM) from isthmus and ampulla are equally effective in prolonging sperm viability in vitro(Fazeli et al., 2003), the two regions of oviductal tissue were pooled for this study. The methods used for preparing APM and sAPM have been described previously(Holt et al., 2005).

Because each video recording of motility took about 3 min, and needed to be initiated at set intervals after the initiation of treatments, incubations were carried out one at a time. Six or eight incubations were normally carried out on a single day, from which direct comparisons (either between different ejaculates or between treatments) could be made.

Incubations were carried out as follows. M solution (1 ml), containing sAPM or DIDS if appropriate, was prewarmed to 38°C in a capped 15-ml polystyrene tube (Sterilin, Stone, Staffs, UK). An aliquant (3–5 μl)of washed sperm (final concentration approximately 2×10⁶cells ml^–1) was added, and the suspension incubated at 38°C for 10 min. A 60 μl sample was then removed for motility analysis. Next, half the remaining suspension was transferred to an empty prewarmed tube and to it was added the control treatment (NaCl as control for bicarbonate addition). 12 min after the initiation of incubation, the bicarbonate/CO₂ mixture (`activator') was added to the rest of the sperm suspension (in the first tube), achieving a final bicarbonate concentration of 15 mmol l<sup>–1</sup>. Incubation of both tubes was continued, and further 60 μl samples were taken for motility analysis: from the `experimental' tube at 5-min intervals up to a maximum of 32 min after addition of activator. A final sample was taken from the `control' tube 5 min after the last activated sample.

Sperm motility was recorded by videomicroscopy as described by Holt and Harrison (Holt and Harrison,2002). 60 μl samples were taken from the sperm suspensions and placed on electropositive glass slides; microscopy was performed using an Olympus BH-2 microscope with a ×10 negative–high phase contrast objective. Sperm video sequences were recorded on CDs using a VCD recorder(VDR-3000; Datavideo UK Ltd, Manchester, UK).

Two hundred individual sperm trajectories were analysed quantitatively for each of the treatment/time combinations, using a Hobson Sperm Tracker (Hobson Tracking Systems, Sheffield, UK) operating at 50 Hz within an IBM-compatible Pentium computer. The `search radius' used was 5.9 μm, and the `minimum track points' setting was 50 frames.

The measured descriptors of sperm motion are summarised in Table 1. Details regarding the use of the Hobson Sperm Tracker and discussion of these parameters may be found in earlier publications (Abaigar et al., 1999; Davis and Siemers,1995).

Detailed exploration of individual sperm trajectories was undertaken using purpose-written software (GET XY) supplied by Prof. G. S. Hobson (Hobson Tracking Systems, Sheffield, UK). This software provides series of x,y coordinates of sequential sperm head positions at a resolution of 50 Hz from video recordings; these can be used to reconstruct and plot sperm trajectories using auxiliary software such as Microsoft Excel or Statistica for Windows(Statsoft UK, Letchworth, UK).

Intracellular pH responses to the presence of sAPM or 100 μmol l^–1 DIDS (with and without bicarbonate addition) were tested directly in parallel with the sperm motility responses; this was achieved using split semen samples (N=9 boars). pH_i response profiles were recorded at 5-min intervals up to 60 min.

sAPM proteins or bovine serum albumin were coupled to polymeric beads with derivatised (formyl groups), chemically active surface properties (Toyopearl AF-formyl 650M; Tosoh Corporation, Japan) using a carbodiimide amine coupling procedure following the manufacturer's protocol. Protein solutions (sAPM and bovine serum albumin; 0.125–2.00 mg ml^–1) were mixed with resin and NaCNBH₃ and agitated gently overnight at room temperature. Formyl groups were blocked using further NaCNBH₃treatment for 1 h at 25°C. Efficiency of protein coupling was determined using a fluorescamine-based protein assay suitable for a fluorescence microplate reader (Lorenzen and Kennedy,1993).

Percoll-washed semen samples were diluted to a concentration of 100–120×10⁶ ml^–1 in modified Tyrode's solution (TLP) containing 4 mg ml^–1 BSA. 0.5 ml of sperm suspension was applied to small columns of protein-coupled resin beads (0.25 ml packed volume prepared in 1 ml disposable syringes). The sperm samples were washed through the columns with a series of 16×0.5 ml aliquants of M medium. The columns were then washed with 0.4 ml of fixative (4% v/v glutaraldehyde, 4% w/v formaldehyde in 100 mmol l^–1 phosphate buffer). Finally, the column materials were extruded and aliquants of resin beads were applied to glass coverslips that had been precoated with 0.5%poly-l-lysine. The preparations were subsequently dehydrated and sputter-coated with gold for scanning electron microscopy.

Sperm trajectory data from the Hobson Sperm Tracker were downloaded into Statistica for Windows (Statsoft UK, Letchworth, UK) for analysis. For the preliminary analysis of summary statistics, mean parameter values were derived from each set of 200 sperm trajectories. These represented replicate values and were used in initial analyses of variance (ANOVAs) after log or arcsine transformation (Zar, 1984). Planned comparisons within the experimental designs were analysed using contrast analysis. More detailed analyses were also performed using multivariate pattern analysis software PATN(Abaigar et al., 1999).

Multivariate analyses of sperm motion parameters were carried out using the computer program PATN (Belbin,1993). The program uses a series of procedures to analyse and compare the motility parameter values associated with each spermatozoon so as to identify sub-groups within the sperm population (`patterns'). The identification of the sub-groups and their hierarchical classification is carried out by the program independently of the investigator, who is simply required to judge to what degree subgroups may be combined to yield a sufficiently small number of groups to allow practical interpretation. In all of the experiments described here the PATN software identified three sperm subpopulations.

A more complete description and illustration of the use of PATN analysis to identify subpopulations within boar sperm samples is given by Abaigar et al.(Abaigar et al., 1999). However, a few key points are worth mentioning. The PATN analysis was performed using data from all individual spermatozoa within a single experiment and the data need not be normally distributed. Occasional trajectories where AREA>600 μm² (representing groups of spermatozoa) were excluded from the data; any zero values for straight line velocity (μm s^–1; VSL), beat cross frequency (Hz; BCF) and amplitude of lateral head displacement (ALH) were transformed to 0.1. Data sets for analysis were prepared by merging raw data files from every measured sperm sample. As each of these contains data from approximately 200 spermatozoa, the combined data sets contain many thousands of cases (sets of values for each sperm trajectory). Unlike many conventional cluster analysis software packages, PATN is not limited by the computer memory and such large amounts of data can be analysed on a conventional PC. Upon completion of the PATN analysis, each individual spermatozoon was categorised as belonging to one of the small number of groups, or subpopulations, described above. In this study the groups were distinguished on the basis of multivariate combinations of motion descriptors, and qualitative interpretation of the group structure was therefore based on the descriptive interpretation of the sperm motion behaviour that each group represents. Multivariate group centroids were calculated to assist with this interpretation. In some instances, treatment statistics relating to a single subpopulation are presented; this is possible where the subpopulations continue to show significant within-group variance after the PATN analysis.

Once the sub-populations had been identified, the relative frequencies of spermatozoa within each experimental sample, and belonging to each group, were compared by ANOVA using Statistica for Windows (Statsoft UK, Letchworth, UK). Replicated experiments were evaluated by combining frequency data(percentages) across replicates, then using ANOVAs for further analysis. Data expressed as percentages were subjected to arcsine transformation prior to ANOVA.

Summary tables showing PATN group motility characteristics derived for each experiment are shown in Tables 2, 3, 4.

The first two experiments (experiments 1 and 2) were aimed at characterising the modulatory effects of oviductal proteins on the bicarbonate-induced motility response. In the first experiment the effects of a fixed concentration of sAPM [200 μg ml^–1; the concentration that promotes prolonged sperm survival in vitro(Elliott et al., 2001; Fazeli et al., 2003)] on bicarbonate-stimulated sperm motility was tested. Spermatozoa were pre-incubated in the presence or absence of sAPM for 10 min before adding bicarbonate/CO₂ (or NaCl as a control). Sperm motility was observed before bicarbonate/CO₂ or NaCl addition, at 2 min after bicarbonate/CO₂ addition and at 5 min intervals up to 17 min. A final observation of NaCl-treated control spermatozoa was performed at 22 min. The same approach was used in the second experiment, except that the dose–response effect of sAPM was examined (200, 100, 50, 25 and 0 μg ml^–1).

Experiment 1: data from 12,012 individual spermatozoa were obtained from eight replicates (boars); individual sperm motion parameters were analysed but the whole data set was also subjected to PATN analysis which classified them into three groups (Table 2). Group 1 was identified as a fast-linear subpopulation, whereas the other two groups were slow and non-linear.

Examination of the activation effects of bicarbonate/CO₂, and the attenuating effects of sAPM on individual spermatozoa, revealed that these were more complex than is apparent from simple statistical comparisons. Fig. 1A,C shows that prior to bicarbonate/CO₂ stimulation most spermatozoa move slowly along complex and nonlinear tracks (tracks 1 and 2 in Fig. 1E). The addition of bicarbonate/CO₂ in the presence or absence of sAPM stimulated a proportion of spermatozoa to higher velocity and track linearity(Fig. 1B,D; a cloud of new points is visible in the top right corner of B). Nevertheless, many spermatozoa failed to respond to bicarbonate/CO₂ activation,remaining in the bottom left hand corner of the scatterplots.

Fig. 2A,B illustrate the sperm velocity (VSL) and flagellar beat cross frequency (BCF) responses that occurred when Percoll-washed spermatozoa were exposed to 15 mmol l^–1 bicarbonate/CO₂. VSL and BCF both increased significantly (P<0.01) upon bicarbonate/CO₂ addition,reflecting the rapid and visible rise in sperm activity and straightening of trajectories. Spermatozoa activated in the absence of sAPM showed forward progression, but many lateral excursions of the sperm head were typical (track 3 in Fig. 1E).

After the initial rapid increase, VSL remained constant throughout the 17 min incubation period; VSL and BCF of the untreated control sample (addition of NaCl only and measured at 22 min) remained unchanged throughout incubation,emphasising the strict requirement for bicarbonate/CO₂ to induce the activation effect. Exposure to 200 μg ml^–1 sAPM for 10 min prior to bicarbonate/CO₂ addition significantly reduced, but did not abolish, the stimulation effect on both mean VSL (P<0.01)and mean BCF (P=0.015).

Inspection of the scatterplots suggested that sAPM reduced the proportion of spermatozoa undergoing bicarbonate-induced activation (compare Fig. 1B and D). This was confirmed when the relative frequencies of fast-linear spermatozoa (determined by PATN analysis) were compared (Fig. 2C,D).

Frequency variations of fast-linear spermatozoa along the experimental time course showed that, when sAPM was absent, bicarbonate/CO₂ addition caused rapid recruitment into the fast-linear group(Fig. 2C). This frequency reached a maximum at 12 min post-bicarbonate/CO₂ addition, followed by a significant decline. However, pre-exposure to sAPM induced a considerable change in this profile. About 40% of spermatozoa were classified as group 1(fast-linear) at time 0 min; however, bicarbonate/CO₂ addition failed to induce a statistically significant (P=0.248) initial rise in group 1 frequency. There was nevertheless an increase that was immediately followed by a decline that could be seen consistently in each of the eight replicates (Fig. 2D). In addition to these statistical effects the sAPM induced qualitative increases in track linearity in typical sperm trajectories(Fig. 1F). Comparing these tracks with the control, non sAPM-treated spermatozoon(Fig. 1E, track 3) illustrates this increased linearity of forward progression, where fewer lateral excursions result in the convergence of VSL and curvilinear velocity (μm s^–1; VCL) and a ratio that approaches unity.

As with many animal experiments, replicate (individual boar) differences were significant within the analyses (F_7,76=3.62; P=0.002). More specific investigation of these differences revealed that they were systematic. In the absence of sAPM, there was a negative correlation (r²=–0.653; P=0.11) between the relative sizes of the fast-linear sperm subpopulations before and after bicarbonate/CO₂ addition (times 0 and 7 min); however, this correlation became positive but weaker in the presence of sAPM(r²=0.432; P=0.285; compare Fig. 3A and B). Not only was this reversal of slope observed, but when the individual boars were ranked by size of the post-activation cohort of active spermatozoa (sAPM absent), their respective rankings were changed by the presence of sAPM(Fig. 3C). Notably, the boars occupying the first and last rank positions reversed their ranks in the presence and absence of sAPM.

In experiment 2, the frequency of fast-linear spermatozoa within ejaculates was negatively correlated with sAPM concentration(r²=0.74; P<0.001; Fig. 4A); the inhibitory effect was most noticeable with bicarbonate/CO₂-stimulated samples although it was even weakly apparent without stimulation. The overall stimulatory effect of bicarbonate/CO₂ nevertheless remained statistically significant (F_1,141=38.3, P<0.0001). Data were obtained from 28,273 individual spermatozoa(6 boars) and subjected to PATN analysis(Table 3).

Testing the dose–response relationship (0, 25, 50, 100, 200 μg ml^–1) and the suppression of bicarbonate/CO₂activation with an orthogonal contrast vector (–2, –1, 0, +1, +2)confirmed that the dose response was directly proportional to the log_{(base 2)} of the sAPM concentration(F_1,141=43.52, P<0.0001). It is important to point out that this result represents quantal responses, i.e. fewer spermatozoa became activated in the presence of increasing doses of sAPM, but it is noteworthy that almost 40% of the available spermatozoa still became activated at the highest dose of sAPM (Fig. 4A). Most motion parameters of these progressive spermatozoa (e.g. VCL, VSL and BCF) changed relatively little in response to increasing sAPM concentration, but linearity (LIN) showed a positive dose–response relationship (Fig. 4B).

These results support the hypothesis that oviductal proteins act upon a particular subset of spermatozoa, and moreover, suggest that individual spermatozoa exhibit varying sensitivities to sAPM. The results of the previous experiment were confirmed, in that even at the highest concentration tested(200 μg ml^–1) approximately 30–40% of spermatozoa remained unaffected by the proteins. The strict dose relationship is consistent with the view that sAPM acts via a receptor–ligand mechanism in which individual spermatozoa are inactivated when they interact with different numbers of signalling molecules.

In searching for alternative explanations for the inhibitory effects of sAPM on spermatozoa we considered the possibility that some of the proteins in the sAPM fraction might be preventing the uptake of bicarbonate/CO₂through anion channels. As spermatozoa are known to possess bicarbonate-exchange proteins (Darszon et al., 2001; Demarco et al.,2003; Holappa et al.,1999; Parkkila et al.,1993) that exchange with chloride or sodium, we used the specific bicarbonate transport inhibitor,4,4′-diisothiocyanatostilbene-2,2′-disulphonic acid (DIDS)(Zeng et al., 1996), to compare the effects of bicarbonate transport inhibition with the oviductal protein interactions.

In experiment 3 sperm motility responses to 50 μg ml^–1sAPM were compared with those to 100 μmol l^–1 and 600μmol l^–1 DIDS, both in terms of motility responses and effects on intracellular pH_i. Motility effects were tested up to 32 min after bicarbonate/CO₂ addition, and a final control (NaCl treated) was examined at 35 min. Intracellular pH responses to the presence of sAPM or 100 μmol l^–1 DIDS (with and without bicarbonate/CO₂ addition) were tested directly in parallel with the sperm motility responses; this was achieved using split semen samples.

Data were obtained from 39 989 spermatozoa and subjected to PATN analysis using the parameters VCL, average path velocity (μm s^–1;VAP), straightness (%; STR) and BCF. It became evident during the analysis that 600 μmol l^–1 DIDS produced anomalous results (i.e. a considerable degree of sperm motility stimulation together with unstable and highly variable pH_i effects) and these results are not reported here. However, PATN analysis for the classification of sperm subpopulations was nevertheless performed on the dataset that included these treatments. All sperm motion data was classified into three groups distinguished on the basis of their velocity and linearity (Table 4).

After bicarbonate/CO₂ addition, both 100 μmol l^–1 DIDS and 50 μg ml^-1 sAPM reduced the extent of bicarbonate-induced sperm stimulation(Fig. 5A; control vssAPM and DIDS; F_1,161=6.12; P=0.014). However, in contrast to sAPM, 100 μmol l^–1 DIDS also significantly reduced the BCF of the activated sperm group(Fig. 5B,C; contrast within ANOVA on the fast-linear sperm subpopulation; F_1,24536=18.9; P<0.001). Although 100 μmol l^–1 DIDS and sAPM both suppressed sperm activation, the pH_i measurements confirmed that sAPM did not act through the inhibition of bicarbonate uptake. Bicarbonate/CO₂ addition initiated a linear rise in pH_i in all treatments; but whereas sAPM enhanced the rate of pH_i elevation with respect to the control treatment (F_1,356=52.45; P<0.001; Fig. 6B vs C) the DIDS caused significant inhibition (F_1,252=93.32; P<0.001) (Fig. 6A vsB). After 60 min incubation in the presence of bicarbonate/CO₂ the sAPM-treated samples attained pH_ivalues around 8.5, while the DIDS-treated samples only reached about pH 7.6.

To see whether spermatozoa interacted with oviductal proteins on a solid surface in a manner that was similar to the situation within the oviduct, sAPM proteins were covalently bound to the surface of chemically reactive polymeric beads. Qualitative analyses of the interactions revealed that whereas a significant proportion of the sperm population failed to bind(Fig. 7A), 10–20 spermatozoa did indeed become bound to each bead. The binding interaction consistently occurred via the anterior acrosomal region of the sperm heads (Fig. 7B,C) whereas their flagellae protruded away from the bead surface. Control beads, prepared using BSA, bound spermatozoa indiscriminately via their heads and tails(not shown).

The bicarbonate-induced activation of boar sperm motility provided an excellent test system for evaluating both the differential immediate responses of sperm subpopulations and the way in which the presence of sAPM modulates these effects. Boar spermatozoa exhibit considerable sensitivity to their bicarbonate/CO₂ environment and thus maximal stimulation might be expected within the oviduct where the concentration is more than twofold higher than that used in the present series of experiments. The formation of a sperm reservoir in the oviductal isthmus, where many spermatozoa become attached to oviductal epithelial cells suggests, however, that in some spermatozoa the motility activation response is suppressed by additional signal transduction pathways. However, some spermatozoa do not bind to the oviductal epithelium and may reach the site of fertilisation relatively quickly (<1 h) when insemination and ovulation are closely synchronised. Such basic observations suggest that two functionally distinct sperm subpopulations have the ability to reach and enter the oviduct. Entry to the oviduct itself is a highly selective process in its own right, being dependent on the intrinsic motility of each individual spermatozoon(Olds-Clarke, 1996). The present data implies that this response is both sperm specific and boar specific, thereby providing considerable opportunity for the operation of selective mechanisms.

The sAPM protein fraction used in these experiments has previously been evaluated for its ability to maintain sperm viability over an extended period at body temperature in vitro(Elliott et al., 2001; Fazeli et al., 2003). sAPM consists of a subset of peripheral membrane proteins that can be removed in soluble form from the apical plasma membranes of oviductal epithelial cells by treatment with 1 mol l^–1 NaCl(Holt et al., 2005). Prior to these observations, Smith and Nothnick(Smith and Nothnick, 1997) had demonstrated that apical plasma membrane fractions from the rabbit oviductal epithelium possess the ability to enhance rabbit sperm survival in vitro. These observations lend support to the view that the sAPM effects seen here are analogous to in vivo interactions between spermatozoa and the oviductal epithelial cells.

Before bicarbonate/CO₂ stimulation, but in the presence of sAPM,approximately 40% of the spermatozoa were classified into the fast-linear group; this was only about 10% lower than the corresponding value measured in the absence of sAPM. The sAPM therefore initially affected only about 10% of the total sperm population, which implies that the 40% of unsuppressed spermatozoa are responsive to neither bicarbonate/CO₂ stimulation nor suppression by oviductal protein(s). The origin of such diversity is unclear but is partly attributable to differences in sperm maturation status as a result of mixing within the epididymis(Flesch et al., 2001). Alternatively, signalling responses within individual spermatozoa may be dependent on its intrinsic genetic characteristics. Such variation, at least in mice carrying t-locus mutations, is known to depend on the regulation of the intracellular sperm-specific motility kinase (Smok)(Herrmann et al., 1999) which,if over-expressed and not corrected by other regulatory mechanisms, causes flagellar beat patterns that are incompatible with entry to the oviduct and fertilisation. As the t-locus mutations are also sex linked, this inappropriate motility is responsible for sex ratio skewing of litters(Olds-Clarke, 1991).

The presence of sAPM suppressed further recruitment into the fast-linear group, despite the initial transient increase. If this occurs within the oviductal environment it would allow both oviductal reservoir formation and the continued progression of an unaffected subpopulation. The observation that spermatozoa became attached to sAPM-coated beads provides further support for this model. The sperm-bead interaction bears striking resemblance to published micrographs of spermatozoa bound to oviductal cells in vitro and in vivo (Green et al.,2001; Hunter,1981). Although it may seem surprising that flagellar activity is suppressed by a binding mechanism that principally involves the sperm head,recent studies of flagellar wave modulation have suggested that the calcium storage sites at the sperm neck region are crucially important in determining the directionality of sperm progression(Ho and Suarez, 2003). In this context it is of interest that the unsuppressed spermatozoa responded to sAPM by increasing their track linearity. This may mean that, within the oviduct,some spermatozoa are assisted to reach the site of fertilisation as rapidly as possible.

The dose-responsive nature of the sperm–sAPM interaction supports the belief that the effects are mediated via receptor-ligand interactions. As the analysis was based on population frequencies, the results show that increasing concentrations of sAPM lead to the suppression of ever more individual spermatozoa. At the same time the spermatozoa that remain active appear to exhibit an increasingly linear pattern of movement. This is consistent with in vivo observations of boar spermatozoa in the oviduct(Suarez et al., 1992) whose flagellae demonstrate higher radius of curvature, i.e. are straighter, and whose movements were more linear than equivalent free-swimming spermatozoa. This complex subpopulation response differs from responses obtained by exposing spermatozoa to glycolytic or mitochondrial inhibitors, which typically cause global reductions in sperm activity. In the oviductal environment such a global response would be inappropriate because it would synchronously prevent the activity of the entire sperm population.

Although the stimulatory pathway is now well documented(Chen et al., 2000; Holt and Harrison, 2002; Okamura et al., 1985; Tajima et al., 1987), the existence of an opposing pathway mediated by oviductal components has yet to be elucidated. The observation that sAPM does not cause motility suppression by blocking the uptake of the bicarbonate anion supports the belief that spermatozoa are controlled by a specific sperm-oviduct signal transduction pathway. In fact, the present data showed enhancement of intracellular pH elevation. Such an effect would be consistent with the hypothesis that when spermatozoa become bound to the oviduct epithelium they undergo accelerated capacitation (Fazeli et al.,1999), a process that is accompanied by increased pH_i(Murphy and Yanagimachi, 1984; Zeng et al., 1996). The existence of sperm–oviduct signal transduction pathways that cause de novo gene transcription in the oviduct upon the arrival of spermatozoa have been confirmed recently(Fazeli et al., 2004).

These results may have profound implications for the operation of sperm selection, or self-selection, and sperm competition mechanisms in the oviduct. They show that the stimulatory effects of bicarbonate are exerted differentially at both the individual sperm and individual boar levels. More importantly, the oviductal proteins also differentially suppressed the activation of individual spermatozoa from the different boars, and transformed the way in which boars were ranked after sperm activation. In fact, the data showed that in some boars the proportion of fast-linear sperm in a sperm sample was actually diminished by the addition of bicarbonate/CO₂when oviductal proteins were present. If these boar-specific effects are a reflection of the processes that take place in vivo, the skewed results obtained from heterospermic insemination experiments(Berger et al., 1996; Popwell and Flowers, 2004; Stahlberg et al., 2000) might readily be explained. The presence of a universally conserved bicarbonate receptor in spermatozoa of many species(Chen et al., 2000; Livera et al., 2005; Pastor-Soler et al., 2003) and specific demonstrations that bicarbonate activates motility in human(Luconi et al., 2004; Luconi et al., 2005) and mouse(Wennemuth et al., 2003)spermatozoa, imply, moreover, that the oviductal effects may be of universal significance for sperm selection and sperm competition.

Our data suggest that if spermatozoa successfully enter the oviduct two outcomes are possible; they either bind to the epithelial surface and form the isthmic reservoir, or progress immediately along the oviduct towards the ampulla. At present it is impossible to predict, in the absence of oviductal proteins, the relative extent to which these outcomes are favoured. However,if spermatozoa from different boars bind unequally to the oviductal epithelium, then the relative conception rates are likely to depend largely on the synchrony of insemination with ovulation, and the relative dynamics of capacitation and sperm release. Spermatozoa that do not bind to the oviduct would have a greater chance of fertilisation if oocytes are present around the time of insemination, while spermatozoa that were stored, and then released some hours later, would be able to fertilise oocytes if ovulation were delayed until some time after insemination.

We therefore propose a working hypothesis as a basis for future experiments in sperm selection and sperm competition. Our data confirm that boar spermatozoa are exquisitely sensitive to the presence of bicarbonate/CO₂; this is consistent with the molecular identification of soluble adenylyl cyclase as a bicarbonate sensor in spermatozoa (Chen et al.,2000). A crucial aspect of our hypothesis is the observation that exposure to bicarbonate affects individual spermatozoa differentially. Some spermatozoa are actively progressive in the absence of bicarbonate while others are quiescent until activated. Together, these two groups typically make up about 50–80% of all spermatozoa in an ejaculate, the remainder apparently being rather insensitive to their environment. We propose that the inter-boar differences in relative proportions of these sperm subpopulations are likely to be responsible for between-boar differential fertility. More specifically, we propose that differences in the relative numbers of spermatozoa physically enabled to enter the oviduct would differ between boars, even if the inseminate contained equal numbers of spermatozoa from different boars.

 
• ALH

  amplitude of lateral head displacement

 
• ANOVA

  analysis of variance

 
• APM

  apical plasma membrane preparation

 
• BCECF

  2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein,acetoxymethyl ester

 
• BCF

  beat cross frequency (Hz)

 
• BTS

  Beltsville thawing solution

 
• DIDS

  4,4′-diisothiocyanatostilbene-2,2′-disulphonic acid

 
• EGTA

  ethylene glycol bis(2-aminoethyl ether)-N,N,N′N′-tetraacetic acid

 
• LIN

  linearity (%)

 
• PATN

  name of pattern analysis software

 
• STR

  straightness (%)

 
• TLP

  modified Tyrode's solution

 
• sAPM

  solubilised apical plasma membrane protein fraction

 
• VAP

  average path velocity (μm s^–1)

 
• VCL

  curvilinear velocity (μm s^–1)

 
• VSL

  straight line velocity (μm s^–1)

We thank JSR-Newsham Ltd (Thorpe-Willoughby, Yorks, UK) and PIC Ltd(Fyfield Wick, Oxfordshire, UK) for supplying the boar semen used in these experiments. We also thank the BBSRC, Sygen Plc and Genus Breeding Ltd for grants in support of this work and the Royal National Orthopaedic Hospital,Stanmore, for use of their electron microscope.