Molecular Human Reproduction, Vol. 9, No. 11, 645-661,
November 2003
© 2003 European Society of Human Reproduction and Embryology
Article |
Multiple forms of redox activity in populations of human spermatozoa
Submitted on June 23, 2003; accepted on July 14, 2003
ARC Centre of Excellence in Biotechnology and Development and Discipline of Biological Sciences, School of Environmental and Life Sciences, Faculty of Science and Information Technology, University of Newcastle, NSW 2308, Australia
1 To whom correspondence should be addressed at: Discipline of Biological Sciences, School of Environmental and Life Sciences, University of Newcastle, University Drive, Callaghan, NSW 2308, Australia. e-mail: jaitken{at}mail.newcastle.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
In this study we have examined the biochemical attributes of the redox systems that regulate human sperm function using 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulphophenyl)-2H-tetrazolium, monosodium salt (WST-1), lucigenin and luminol-peroxidase as probes. WST-1 was readily reduced by human spermatozoa in the presence of an intermediate electron acceptor (IEA) or NAD(P)H. The IEA-mediated activity resembled a previously described trans-membrane NADH oxidase in being inhibited by capsaicin, superoxide dismutase (SOD) and N-ethyl maleimide, but differed in its sensitivity to p-chloromercuriphenylsulphonic acid (pCMBS). The NAD(P)H-induced WST-1 reduction resembled the superficial oxidase described previously, in its sensitivity to pCMBS, but differed in its suppression by capsaicin. Lucigenin was reduced by human spermatozoa in a manner that could be inhibited by SOD and stimulated by NAD(P)H or 12-myristate, 13-acetate phorbol ester. A23187 also stimulated human spermatozoa via a diphenylene iodonium-sensitive pathway detectable with luminol-peroxidase but not lucigenin. Defective sperm populations recovered from the low-density region of Percoll gradients were characterized by high levels of redox activity that was only discernable with lucigenin. We conclude that human spermatozoa possess multiple plasma membrane redox systems that are involved to varying extents in the physiological control and pathological disruption of sperm function. Their distinct pharmacological profiles should significantly assist attempts to resolve and characterize these systems.
Key words: human spermatozoa/lucigenin/luminol/Percoll fractionation/redox activity
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The capacitation of mammalian spermatozoa is a redox-regulated event initiated by low-level reactive oxygen species (ROS) generation and mediated by high levels of phosphotyrosine expression (Bize et al., 1991; De Lamirande and Gagnon, 1993; Aitken et al., 1995, 1998; Visconti et al., 1995; Leclerc et al., 1997). The linkage between redox activity on the one hand and tyrosine phosphorylation on the other, appears to involve an increase in intracellular cAMP levels that, through a novel PKA-mediated signal transduction cascade, activates tyrosine kinase activity (Visconti et al., 1995; Zhang and Zheng, 1996; Aitken et al., 1998; Lewis and Aitken, 2001).
The physiological significance of this redox activity is emphasized by the suppressive effects of ROS scavenging enzymes such as catalase and superoxide dismutase (SOD) on the capacitation status of mammalian spermatozoa (Bize et al., 1991; De Lamirande et al., 1993; Griveau et al., 1994, 1995; Aitken et al., 1998). In addition, disruption of the plasma membrane redox activity human spermatozoa with agents such as the flavoprotein inhibitor diphenylene iodonium (DPI), is known to suppress the fertilizing capacity of these cells (Aitken et al., 1998). Conversely, the artificial stimulation of redox activity by the addition of NAD(P)H to the culture medium, has been shown to stimulate capacitation in both rat (Lewis and Aitken, 2001) and human spermatozoa (Aitken et al., 1995, 1996, 1997; De Lamirande et al., 1998).
The redox activity intrinsic to human spermatozoa is not just of significance in the physiological programming of spermatozoa for fertilization, it is also involved in the aetiology of sperm pathology (Aitken and Fisher, 1994; Sharma and Agarwal, 1996). The susceptibility of these cells to oxidative stress has been appreciated since MacLeod (1943) demonstrated that the loss of sperm motility in oxygenated medium could be reversed by the addition of catalase. This vulnerability to ROS stems from the high unsaturated fatty acid content of the sperm plasma membrane and the consequent susceptibility of this structure to lipid peroxidation (Jones et al., 1979; Aitken et al., 1993b). This tendency to undergo peroxidative damage is also compounded by the relative lack of antioxidant protection offered by these cells as a result of the restricted volume and limited distribution of cytoplasm in which to house the antioxidant enzymes that protect most cells types from oxidative stress (Jones et al., 1979; Aitken and Baker, 2002). As a consequence, direct exposure of human spermatozoa to high levels of ROS has been shown to suppress sperm function via mechanisms that involve the induction of lipid peroxidation (Aitken et al., 1989a, 1993a,b; Griveau et al., 1995). The fact that catalase can protect spermatozoa from oxidative damage suggests that hydrogen peroxide is a particularly damaging oxidant as far as spermatozoa are concerned, suppressing the functional competence of these cells and disrupting their genomic integrity (Aitken et al., 1993a, 1998; Oehninger et al., 1995; Kemal Duru et al., 2000).
The importance of ROS in the aetiology of defective human sperm function is further supported by the high levels of chemiluminescence detected in sperm preparations from such patients using luminol or lucigenin as the probe (Aitken and Clarkson, 1987; Aitken et al., 1989a,b; McKinney et al., 1996; Gil-Guzman et al., 2001; Kobayashi et al., 2001; Ollero et al., 2001). Moreover, the subpopulation of human spermatozoa generating high levels of chemiluminescence in infertile patients have been isolated by virtue of their tendency to migrate to the low-density region of discontinuous Percoll gradients (Aitken and Clarkson, 1988; Aitken et al., 1989a; Gil-Guzman et al., 2001; Ollero et al., 2001). The low density of these redox-active spermatozoa appears to be due to the presence of excess residual cytoplasm, one consequence of which may be to enhance the supply of substrate to the redox-active systems in these cells through the generation of NAD(P)H (Gomez et al., 1996; Ollero et al., 2001; Baker et al., 2002).
In view of the physiological and pathological significance of redox activity in the control of human sperm function, it will be important to characterize the enzyme systems involved at a biochemical level. An NADPH oxidoreductase activity has been described in these cells (Aitken et al., 1997) and an NADPH oxidase (NOX5), distantly related to the catalytic subunit of the neutrophil oxidase, gp91phox, has been detected in human testicular germ cells (Banfi et al., 2001). Whether this is the only redox system present in human spermatozoa and whether it is the oxidase responsible for oxidative stress in the aetiology of male infertility has not been resolved. In this study, we have used a variety of redox-sensitive probes to examine the nature of the redox systems present in human spermatozoa and determine the attributes of the pathways up-regulated in populations of defective human spermatozoa.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Chemicals
Unless otherwise stated, all chemicals were purchased from Sigma Chemical Company, St Louis, MO, USA.
Spermatozoa
The study population comprised a population of 50 unselected donors (World Health Organization, 1999) who had been counselled to exclude individuals exhibiting a high risk for sexually transmitted diseases such as HIV, which might have influenced the quality or cellular composition of their semen. Institutional and State Government ethical approval was secured for the use of human semen samples for the purposes of this research. The semen samples were produced by masturbation and collected into sterile containers for immediate transportation to the laboratory. After allowing at least 30 min for liquefaction to occur, the spermatozoa were fractionated on a discontinuous two-step Percoll gradient. For this procedure, an isotonic solution was prepared by adding to 90 ml of Percoll (Pharmacia LKB, Uppsala, Sweden), 10 ml of 10x Ham’s F10 (ICN Biochemicals, Seven Hills, NSW, Australia) supplemented with 100 mg of polyvinyl alcohol (PVA), 3 mg of sodium pyruvate, 0.37 ml of a 60% sodium lactate syrup and 200 mg of sodium hydrogen carbonate to give an isotonic preparation that was designated 100% Percoll (Lessley and Garner, 1983). This solution was diluted 1:1 with HEPES-buffered (20 mmol/l) Biggers–Whitten–Whittingham (BWW) medium (Biggers et al., 1971), supplemented with 1 mg/ml PVA and discontinuous gradients were created by layering this low-density Percoll preparation above 3 ml of isotonic Percoll. Liquefied semen was then pipetted onto the gradient and centrifuged for 20 min at 500 g. Spermatozoa were recovered from the base of the gradient and the low-density/high-density Percoll interface, washed with 7 ml of BWW medium, centrifuged (5 min at 500 g) and finally resuspended at 20 x 106/ml in either Hank’s balanced salt solution supplemented with 1 mg/ml PVA for the experiments involving luminol (5-amino-2,3-dihydro-1,4-phthalazinedione) and horse radish peroxidase (HRP) or HEPES-buffered BWW medium supplemented with 1 mg/ml PVA for all other treatments. Under these conditions the spermatozoa readily capacitated over a 180 min incubation period, as indicated by the acquisition of high levels of tyrosine phosphorylation (data not shown).
All of the preparations used in this study were polymorph-free in that they had been screened with a formyl methionyl leucyl phenylalanine (fMLP) provocation test (Krausz et al., 1992). If a positive signal was obtained, the sperm suspension was incubated with magnetic beads coated with an anti-CD45 monoclonal antibody (Dynabead M-450, CD45 pan leukocyte, product 11120; Dynal, Oslo, Norway). Following this treatment, the samples were tested again with the fMLP provocation test and only if a negative response was obtained were they carried forward for analysis.
Luminometry
Lucigenin (bis-N-methylacridinium nitrate)- and luminol-dependent chemiluminescence was recorded on a Berthold 953 luminometer (Berthold Detection Systems GmbH, Crown Scientific Pty Ltd, Moorebank, Australia) at 37°C using 400 µl aliquots of spermatozoa (10 x 106/ml). Media blanks were run for every treatment in order to ensure that the signals recorded were not due to the spontaneous reduction of the probe. The values obtained in these media-only control incubations were subtracted from those obtained in the presence of spermatozoa. When lucigenin was used as the probe, 400 µl of sperm suspension was placed in a sample tube and mixed with 4 µl of a 25 mmol/l lucigenin stock solution dissolved in dimethylsulphoxide. NAD(P)H was used at a final concentration of 2.5 mmol/l while 12-myristate, 13-acetate phorbol ester (PMA) was employed at a final concentration of 100 nmol/l.
For the luminol assay, 400 µl of spermatozoa, at a concentration of 10 x 106/ml, were placed in a 3 ml sample tube and 4 µl of a 25 mmol/l luminol stock solution in dimethyl sulphoxide, was then added to give a final probe concentration of 250 µmol/l. The luminol-dependent chemiluminescence was enhanced by the addition of 11.52 U/ml HRP. The luminometer results were recorded as continuous traces and as integrated photon counts over a fixed period of time (2 min).
2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulphophenyl)-2H-tetrazolium, monosodium salt (WST-1) reduction
Human spermatozoa were prepared at a concentration of 20 x 106 /ml in BWW supplemented with 1 mg/ml PVA. Aliquots (100 µl) of these sperm suspensions were then added to the wells of a microtitre plate, PVA-supplemented BWW serving as the cell-free control. Redox activity was monitored following the addition of 10 µl of a 4.5 mmol/l solution of WST-1 (Dojindo, Kumamoto, Japan) and the reduced product measured at 0, 5, 10, 15, 30, 60, 120 and 180 min using a Biorad Ultramark microtitre plate reader (BioRad, Regents Park, NSW, Australia) with the filters set at 415 nm and, as a reference wavelength, 550 nm. The additions were made in a volume of 10 µl. Media blanks were run for every treatment in order to ensure that the signals recorded were not due to the spontaneous reduction of the probe. The values obtained in these media-only control incubations were subtracted from those obtained in the presence of spermatozoa so that only cell-dependent probe reduction was recorded. For certain experiments, a commercially available WST-1 preparation was used (Cell Proliferation Reagent WST-1; Roche Diagnostics, Mannheim, Germany) that incorporated an electron coupling reagent in the form of 1-methoxy phenazine methosulphate (PMS), to facilitate reduction of the probe (Berridge and Tan, 2000a,b).
Impact of CHAPS
In order to gain an insight into the site of redox activity in human spermatozoa, these cells were treated with a mild detergent, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulphonate (CHAPS; Research Organics, Cleveland, OH). For this analysis, human spermatozoa were Percoll-purified as described above and diluted to 10 x 106/ml in BWW medium supplemented with 1 mg/ml PVA. An aliquot of these cells was then treated with 0.05% CHAPS (400 µl of 0.5% CHAPS in PBS added to 4 ml of sperm suspension), the controls being treated with PBS alone. The cells were then agitated on an Eppendorf Thermomixer (Eppendorf South Pacific Pty, North Ryde, Australia) set to 37°C for 30 min. At the end of this period the cells were centrifuged at 16 000 g on a Heraeus Biofuge (Heraeus Ltd, Hanau, Germany) for 15 min and the supernatant removed. The cell pellet was then resuspended in BWW supplemented with PVA and centrifuged again at 16 000 g for 15 min. On this occasion, the supernatant was discarded and the pellet diluted in BWW–PVA to 10 x 106 cells/ml.
Inhibitors
The following inhibitors were used to determine the pharmacological profile of the redox activity detected with the various redox detection systems: the flavoprotein inhibitor, DPI (100 µmol/l); catalase (3000 IU) to remove hydrogen peroxide; potential inhibitors of NADPH oxidase activity, resiniferatoxin (RFT; 10 µmol/l) and capsaicin (CAP; 100 µmol/l); the thiol reactive agents p-chloromercuriphenylsulphonic acid (pCMBS; 100 µmol/l) and N-ethyl maleimide (NEM; 100 µmol/l); SOD (300 IU) to remove superoxide anion; retinoic acid (RA; 20 µmol/l) a potential inhibitor of NADH oxidase activity (Dai et al., 1997) and the mitochondrial electron transport inhibitor, rotenone (200 µmol/l). Unless otherwise indicated, inhibitors were added 10 min before the addition of agonist [PMA, NAD(P)H or A23187]. The inhibitors were made up as stock solutions in DMSO, with the exception of RA (ethanol), catalase (PBS) and SOD (PBS).
NAD(P)H oxidation
Percoll-purified human spermatozoa were suspended at 4 x 106/ml in Hank’s balanced salt solution and treated with 20 µmol/l NAD(P)H. Oxidation of the pyridine dinucleotides was followed at 37°C using an excitation wavelength of 340 nm and an emission wavelength of 460 nm. Measurements were taken at 12 s intervals for a period of 10 min in a Shimadzu spectrofluorimeter (RF-5301PC; Shimadzu, Torrance, CA, USA).
Statistics
All experiments were repeated at least three times on independent samples and the results analysed by ANOVA using the SuperANOVA program (Abacus Concepts Inc., CA, USA) on a MacIntosh G4 computer; post hoc comparison of group means was by Fisher’s PLSD test. The chemiluminescence data were square-root-transformed prior to analysis to improve the normality of the data distribution. Paired comparisons were conducted using the paired t-test on the Statview program (SAS Institute, NC, USA). The significance level was set at P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
WST-1 reduction
Incubation of purified human sperm populations, free of detectable leukocyte contamination, in the presence of the electron acceptor, WST-1, resulted in a slight reduction of the probe that was statistically significant only after 180 min (P < 0.05; Figure 1A). An important feature of this spontaneous reduction of WST-1 by human spermatozoa was its inconsistency; some samples being redox active in this regard while others exhibited no detectable activity. Inclusion of an intermediate electron acceptor (IEA) consistently and significantly enhanced the levels of detectable WST-1 reduction (Figure 1B; P < 0.001), possibly by sensitizing the system for trans-plasma membrane electron flux as suggested by Berridge and Tan (2000a,b). This enhanced redox activity was significantly suppressed by SOD, indicating a role for the superoxide anion in signal generation, but was actually enhanced by catalase (Figure 2; P < 0.001). Thiol involvement in WST-1/PMS reduction, was suggested by the significant inhibitory effects observed with pCMBS and NEM. Known inhibitors of plasma membrane electron transport systems, CAP and RA, also significantly inhibited WST-1 reduction in the presence of PMS (P < 0.001). However, the flavoprotein inhibitor, DPI, could not significantly inhibit reduction of this probe in the presence of an intermediate electron coupling reagent, in contrast to the NAD(P)H-induced WST-1 reduction described below (Figure 3). Furthermore, WST-1 reduction in the presence of PMS did not appear to involve electron leakage from the sperm mitochondria because rotenone had no significant effect on the enhanced signal generated with this system (Figure 2).
|
|
|
The addition of an electron source to the sperm suspensions in the form of NAD(P)H enhanced the reduction of the probe in a manner that was significantly dose-dependent (P < 0.001), the optimal doses being between 0.25 and 2.0 mmol/l (Figure 3). The level of WST-1 reduction was significantly higher with NADH as the electron donor, rather than NADPH (P < 0.001). When WST-1 reduction was enhanced by NADPH addition, catalase, but not SOD, could inhibit the signal generated (Figure 4A). In addition, WST-1 reduction in the presence of NADPH was extremely sensitive to inhibition by pCMBS and, to a lesser extent, NEM, suggesting the involvement of thiols, particularly surface thiols, in the reduction process (Figure 4B). DPI was also active in the suppression of NADPH-induced WST-1 reduction, indicating a flavoprotein involvement in the reduction process. RFT had a mild inhibitory effect on WST-1 reduction under these conditions whereas CAP, a putative inhibitor of trans-plasma membrane redox activity (Berridge and Tan, 2000a,b) and the t-NOX surface oxidase (Chueh et al., 2002), had no impact on NADPH-induced WST-1 reduction (Figure 4B).
|
NADH also stimulated an increase in trans-membrane electron flux as measured by WST-1 reduction (Figure 5). However, the chemistry of this NADH-induced reduction was quite distinct from the NADPH-activated process in that it was clearly inhibitable by SOD (P < 0.01), indicating a role, albeit indirect, for superoxide anion in the reduction process. It was also much less susceptible to DPI inhibition (P < 0.05) and, in complete contrast to NADPH-induced WST-1 reduction, could be significantly suppressed by CAP (P < 0.01). Although NADH-dependent WST-1 reduction was, like the NADPH response, inhibitable by pCMBS, it differed from the latter in being resistant to NEM suppression (Figure 5).
|
Lucigenin mediated chemiluminescence
Purified populations of human spermatozoa isolated from the high-density region of Percoll gradients, and free of detectable leukocyte contamination, did not give significant spontaneous signals with luminol or isoluminol; however, a low level of activity was detectable with lucigenin (Figure 6). This basal lucigenin signal could be silenced by the addition of 300 IU SOD, but not any of the other inhibitors assessed (NEM, RFT, CAP, rotenone, catalase, DPI and pCMBS; data not shown). The addition of NADPH significantly enhanced this lucigenin-dependent chemiluminescence, but had no impact on the signals obtained with luminol or isoluminol (Figure 6A,C,D). The response to NADPH contained a component, representing
64% of the total signal, that was SOD inhibitable (designated ‘
’ in Figure 6A). The remaining 36% of the NADPH response could not be suppressed by extracellular SOD (‘ß’ in Figure 6A) possibly as a result of the inability of this extracellular enzyme to access the intracellular sites at which lucigenin was being reduced. As revealed in Figure 6, exactly the same level of SOD-resistant signal intensity was observed, regardless of whether this enzyme was added before or after the NADPH.
|
In addition to the suppressive effect of SOD, this chemiluminescent response to NADPH could be suppressed by DPI (P < 0.01) and pCMBS (P < 0.05; Figure 7A) emphasizing the importance of flavoproteins and surface thiols in the origins of this NADPH-dependent lucigenin signal. NADH could also stimulate a lucigenin-dependent chemiluminescent response, however, this activity was not significantly suppressed by any of the inhibitors assessed (DPI, CAP, NEM, pCMBS, RFT and catalase). SOD did inhibit this NADH-dependent lucigenin signal, but the level of suppression did not reach statistical significance on post hoc testing (Figure 7B). In contrast, rotenone significantly enhanced this NADH-induced response (P < 0.01; Figure 7B), possibly by suppressing the utilization of intracellular NADH by the mitochondrial electron transport chain.
|
Luminol peroxidase
Banfi et al. (2001) recently reported the existence of an NADPH oxidase in the male germ line (NOX5) that could be monitored as a calcium-dependent chemiluminescent signal generated in the presence of a detection system comprising luminol and HRP. Human spermatozoa recovered from the high-density region of Percoll gradients gave low-level luminol-peroxidase signals that were not significantly suppressed by any of the inhibitors tested including catalase, DPI, CAP, SOD, NEM, pCMBS, RFT and rotenone (data not shown). However, following addition of A23187, a chemiluminescent response was observed (Figure 8A) that was sensitive to inhibition by catalase, DPI, CAP and SOD (Figure 8B). Thiols were clearly not involved in this response, since it could not be inhibited with pCMBS or NEM. Moreover, neither RFT nor rotenone could inhibit this A23187-dependent increase in chemiluminescence (Figure 8B). Interestingly, signals were not observed with human spermatozoa when this ionophore was used in conjunction with lucigenin as the chemiluminescent probe (data not shown).
|
Location of redox activity
In order to gain some insight into the location of the electron transfer events recorded in this study, cells were extracted with CHAPS and redox activity compared in the extractable and non-detergent-soluble fractions. In terms of WST-1 reduction, significant activity was observed in the cells as well as the supernatants prepared in the presence and absence of CHAPS exposure (Figure 9). The CHAPS-soluble fraction exhibited the greatest activity with both NADH and NADPH as electron donors but significant WST-1 reduction was also found in the cell pellet and supernatant of control cell suspensions treated with PBS alone. These results suggest that human spermatozoa possess two types of WST-1 reductase activity. One that is superficially located and can be readily removed from the intact cell by a simple washing procedure. This activity appears in the supernatant of washed cells that have not been exposed to CHAPS. A second activity remains bound to the cell after such simple washing procedures and is presumably associated with the plasma membrane since it can be released by treatment with CHAPS leaving the residual cell bodies with little or no detectable activity (Figure 9A).
|
When lucigenin was used as a probe a different pattern of activity was observed. In this case, little (with NADH) or no (with NADPH) activity could be released from the cells by a simple washing procedure and all detectable lucigenin reductase activity was destroyed on treatment with CHAPS in both the cells and supernatants (Figure 9A,B).
Redox activity and sperm quality
The data presented above demonstrates that human spermatozoa are redox-active cells that contain multiple pathways for regulating electron flux in the vicinity of the plasma membrane. Since high levels of redox activity have been associated with abnormal sperm function (Aitken and Krausz, 2001), it was of clinical interest to determine which of these redox systems were elevated in populations of defective spermatozoa.
In keeping with previous findings (Aitken and Clarkson, 1988; Gomez et al., 1996), low-density Percoll fractions free of detectable leukocyte contamination exhibited significantly higher spontaneous lucigenin signals that the corresponding 100% Percoll samples (2.13 ± 0.24 
c.p.m. versus 1.05 ± 0.10 c.p.m.; P < 0.001). These spontaneous lucigenin signals generated by low-density Percoll fractions were suppressed by the superoxide scavenger SOD and the flavoprotein inhibitor DPI (P < 0.001; Figure 10A). The possible involvement of thiol groups in this spontaneous chemiluminescence was suggested by the inhibitory effects of pCMBS (P < 0.01) and NEM (P < 0.001; Figure 10A). The trans-plasma membrane redox system inhibitor, CAP, also had a significant effect on this signal (P <0.05), whereas RFT and catalase had no effect (Figure 10A). The high level of redox activity associated with these defective low-density fractions was not due to electron leakage from the sperm mitochondria because rotenone had no impact whatsoever on the lucigenin-dependent chemiluminescence associated with such samples (Figure 10A).
|
The addition of PMA (100 nmol/l) to these low-density Percoll fractions elicited a lucigenin-dependent chemiluminescence response that was significantly higher than in the corresponding high-density Percoll fractions (Figure 11D). These high-level PMA-dependent responses were susceptible to inhibition by SOD (P < 0.01), DPI (P < 0.001) and NEM (P < 0.01) but not to any of the other potential inhibitors assessed including pCMBS, CAP, RFT, rotenone and catalase (Figure 10B). The impact of PMA addition is clear when the suppressive effects of CAP and pCMBS are compared; both of these reagents were effective inhibitors of the spontaneous lucigenin signals generated by low-density Percoll fractions, but neither was active when these cell populations were stimulated with PMA (Figure 10A,B).
|
The addition of NADPH or NADH also elicited redox activity in human spermatozoa that was significantly more intense in the low-density Percoll sperm populations than in the high-density fractions (P < 0.001; Figures 11A,C and 12A) and highly correlated with the PMA-elicited lucigenin-dependent chemiluminescence (r = 0.680; n = 72; P < 0.001). Although the inhibitor profiles presented above suggested that NADH and NADPH induced lucigenin responses via different biochemical mechanisms, the responses of individual sperm preparations to the two nucleotides were highly correlated (r = 0.840; P < 0.001; Figure 12B), suggesting a common mechanism for the variation in redox activity, such as the retention of excess residual cytoplasm. Notwithstanding this high overall level of correlation, individual samples occasionally departed from this general trend (circled points in Figure 12B) suggesting that occasionally factors predispose a given sample towards a higher lucigenin response with one dinucleotide rather than the other, particularly in the case of NADH.
|
Fluorometric analyses demonstrated significant rates of spontaneous NADH oxidation in these low-density Percoll fractions, that were significantly correlated with the NADH-elicited, lucigenin-dependent chemiluminescent responses given by the same samples (r = 0.680; n = 14; P < 0.001). In contrast, no significant NADPH oxidation could be detected in these samples at the concentration (50 µmol/l) examined. Unfortunately, it was not technically possible to assess the oxidation of higher concentrations of NADPH in the range used for stimulating high levels of redox activity (0.125–2.0 mmol/l).
In contrast to the above results with PMA and NAD(P)H, addition of A23187 to low-density Percoll sperm populations elicited a chemiluminescent response that, though higher than in the high-density fraction, did not reach statistical significance (Figure 11B). Comparison of the relative intensities of the A23187/luminol-peroxidase and NADH/lucigenin signals achieved in individual samples, demonstrated that while most samples show some level of responsiveness to A23187, the response to NADH was highly variable (Figure 8A,C).
In order to determine whether the redox activity detected with WST-1 would also detect the elevated redox activities associated with low-density Percoll fractions, a comparative analysis of the signals generated by WST-1 and lucigenin was undertaken. The results of this study revealed that whereas NADH/lucigenin was an extremely effective system for detecting the enhanced redox activity associated with these defective sperm populations, WST-1 could not discriminate in this context, regardless of whether probe reduction was promoted with PMS or NAD(P)H (Figure 13A). Regression analysis also revealed that while the signals obtained with NADH and NADPH using WST-1 as the probe were highly correlated (r = 0.94; P < 0.001; Figure 13B) there was no significant correlation between WST-1/NADH and lucigenin/NADH (Figure 13C). Evidently, these probes are measuring quite different aspects of the redox activity associated with human spermatozoa.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Electron transfer reactions at the level of the plasma membrane play an extremely important role in the regulation of cell biochemistry (Baker and Lawen, 2000). The results obtained in this study indicate that human spermatozoa possess highly active redox systems capable of reducing membrane impermeant probes such as WST-1 and lucigenin. Moreover, there appear to be several distinct redox pathways operating in these cells, in terms of their location and pharmacological profile. Thus, a significant component of the WST-1 activity could be removed from the spermatozoa by an additional washing step, suggesting that the activity is absorbed onto the sperm surface in a manner consistent with the cell-surface NADH oxidase activity described previously by Berridge and Tan (2000a,b). If such superficial oxidoreductases are active in vivo, they must use alternative electron donors, since it is extremely unlikely that significant quantities of NAD(P)H exist in the extracellular space. Such activity apart, most of the sperm’s capacity for probe reduction is resistant to washing but can be completely removed from the cells by a brief exposure to CHAPS. Such findings support the notion that these redox systems are associated with the plasma membrane (Vernet et al., 2001). In keeping with studies conducted in many cell types (e.g. Meier et al., 1991; Souza et al., 2001; Didion and Faraci, 2002) we have used high concentrations of extracellular NAD(P)H as a means of driving these redox systems. The assumption is that under these circumstances, sufficient co-enzyme crosses the sperm plasma membrane to fuel the latter’s electron transport systems from the cytoplasmic surface. Once the molecular identities of sperm oxidoreductases are known, we shall be in a position to conduct ultrastructural immunolocalization studies, to confirm the subcellular localization of these enzymes.
The redox activity detected in human spermatozoa is certainly biologically relevant because the capacitation of these cells can be enhanced by the addition of exogenous NADPH and suppressed by redox inhibitors, such as DPI (Aitken et al., 1995, 1998; Leclerc et al., 1997). In addition, sperm pathology has been clearly linked with abnormally high levels of redox activity on the part of the spermatozoa, at least in part, associated with the retention of excess residual cytoplasm in the midpiece of the cell (Aitken and Clarkson, 1987; Sharma and Agarwal, 1996; Gil-Guzman et al., 2001; Ollero et al., 2001).
It has frequently been stated that these redox-regulated activities are due to the formation of ROS. The evidence for this assertion is certainly strong. Thus, both SOD and catalase have been shown to suppress the capacitation of mammalian spermatozoa under different experimental conditions, while direct exposure of mammalian spermatozoa to low levels of ROS has been shown to stimulate various aspects of sperm function, particularly capacitation (Bize et al., 1991; De Lamirande and Gagnon, 1993; Griveau et al., 1994; Aitken et al., 1995, 1998; De Lamirande et al., 1998). In addition, there is abundant evidence linking defective sperm function with oxidative stress. Hence, the pathologically low levels of motility and fertilizing potential observed in the spermatozoa of infertile patients is clearly associated with high levels of lipid peroxidation in these cells (Jones et al., 1979; Aitken et al., 1989a; Sharma and Agarwal, 1996). The DNA fragmentation recorded in the spermatozoa of infertile patients and heavy smokers has also been correlated with the presence of 8-hydroxyguanine, a marker of oxidative stress (Fraga et al., 1996; Kodama et al., 1997). Moreover, defective sperm function is exacerbated by high oxygen tensions in the medium (Whittington and Ford, 1998) and, in certain studies, with abnormally low levels of antioxidant activity in seminal plasma (Lewis et al., 1995). Direct exposure of mammalian spermatozoa to excessive ROS, particularly hydrogen peroxide, has also been shown to suppress sperm function or disrupt DNA integrity in several independent studies (Bize et al., 1991; Aitken et al., 1993a; De Lamirande et al., 1993; Oehninger et al., 1995; Lopes et al., 1998; Armstrong et al., 1999). Although oxidative stress can arise through antioxidant depletion and leukocyte contamination as well as through the excessive generation of ROS by the spermatozoa themselves, the latter appears to be the dominant mechanism by which oxidative stress is created in the human ejaculate (Aitken et al., 1992b; Gomez et al., 1996; Sharma and Agarwal, 1996; Gil-Guzman et al., 2001; Kobayashi et al., 2001; Ollero et al., 2001).
Despite such experimental and circumstantial evidence, the direct demonstration of ROS production by these cells has been more difficult than expected. Several (Kumar et al., 1991; Zhang and Zheng, 1996), but not all (Richer and Ford, 2001), researchers have detected a clear superoxide anion signal in populations of purified mammalian spermatozoa using electron spin resonance techniques. Moreover, previous studies using acetylated oxidized cytochrome c and peroxidase-catalysed acetylated reduced cytochrome c (in the presence or absence of SOD and catalase respectively) have clearly indicated that mammalian spermatozoa produce both superoxide anion and hydrogen peroxide; the latter coming almost entirely from the dismutation of superoxide anion (Holland et al., 1982; Alvarez and Storey, 1984, 1985; Alvarez et al., 1987).
Although the chemiluminescence signals detected with luminol and lucigenin are frequently held to reflect ROS production by these cells, this is not necessarily the case. As recently pointed out by Richer and Ford (2001), lucigenin can undergo a one-electron reduction to create a radical species that may give up its electron to oxygen, artificially creating the superoxide anion. The latter then promotes an oxygenation reaction with lucigenin, creating an unstable dioxetane that decomposes to an excited-state acridone with the generation of light. Such chemistry would lead to an SOD-sensitive chemiluminescent signal that was reflective of redox activity but not primary superoxide anion generation. Since it is not possible to determine the proportion of a SOD-sensitive signal that represents primary superoxide anion generation or secondary production due to redox cycling of the probe, it is preferable to consider the read-out of lucigenin-dependent chemiluminescence reactions as being indicative of redox activity, rather than ROS production, and the enzymes responsible as oxidoreductases rather than oxidases (Vernet et al., 2001). Moreover, since lucigenin is present as a divalent cation at neutral pH it will be relatively membrane impermeant and be particularly sensitive to redox activity at the cell surface.
Recently, WST-1 has been introduced as a novel probe for the measurement of redox activity at the level of the plasma membrane (Berridge and Tan, 2000a,b). This sulphonated tetrazolium salt is thought to be membrane impermeant, exhibits low levels of background absorbance and can undergo two electron reduction to generate a soluble formazan that is readily amenable to spectrophotometric quantification. WST-1 has been used to detect superoxide anion by activated leukocytes (Tan and Berridge, 2000) and to obtain evidence for two distinct electron transport systems in the plasma membrane of mammalian cells: (1) a trans-membrane system capable of supporting the flux of 200 million electrons/s/cell from NADH to the outer surface of the plasma membrane and (2) a surface oxidase activated by the addition of NAD(P)H to the incubation medium (Berridge and Tan, 2000a,b).
In order to determine whether the human spermatozoon possesses similar electron transport systems in its plasma membrane, the reduction of WST-1 in the presence of these cells was analysed. In the absence of either an IEA or NAD(P)H, human spermatozoa were capable of stimulating an extremely low level of WST-1 reduction over a 3 h incubation period in vitro. Such spontaneous reduction of WST-1 in the absence of an IEA has previously been observed by Tan and Berridge (2000) in activated neutrophils and is thought to be indicative of the primary production of superoxide anion by these cells. Notwithstanding the low intensity of these spontaneous WST-1 signals, such data would support the chemiluminescence data generated using lucigenin or MCLA as the probe (Aitken et al., 1992a; De Lamirande and Gagnon, 1993, 1995; De Lamirande et al., 1998) in suggesting that populations of capacitating human spermatozoa do spontaneously produce low levels of superoxide anion.
This WST-1 signal was significantly enhanced by the inclusion of an IEA in the form of PMS. Such PMS-facilitated, WST-1 reduction has been observed in a variety of disparate cell types, particularly rapidly proliferating cell lines such as HeLa cervical carcinoma or Jurkat T-lymphoblastic cells. To a lesser extent, resting cells including non-activated neutrophils also reduce WST-1 in the presence of PMS (Berridge and Tan, 2000a,b). As in other cell types, PMS enhanced WST-1 reduction by spermatozoa was profoundly inhibited by SOD. Although this could be supportive of spontaneous superoxide anion generation in populations of capacitating human spermatozoa (De Lamirande and Gagnon, 1993), it is also possible that the production of this reactive oxygen intermediate is a consequence, rather than a cause, of WST-1 reduction. A probable sequence of reactions leading to WST-1 reduction are indicated below, based on the work of Picker and Fridovich (1984) and Berridge and Tan (2000a,b):
This scheme envisages that there is an electron transport system in the sperm plasma membrane that is capable of performing a two-electron reduction of PMS to form PMSH2 as indicted in reaction (a). This could be a trans-membrane electron transport system of the type suggested by Berridge and Tan (2000a,b), that derives its electrons from cytoplasmic NAD(P)H and then shuttles them to the exterior surface of the cell using ubiquinone as an intermediate electron carrier. Once reduced, PMSH2 will then react with WST-1 to form a WST-I radical (WST-1H·) that, by dismutation, will generate the fully reduced soluble formazan product (WST-1H2) detected in the spectrophotometric assay, as depicted in reactions (c) and (e). If ground state oxygen is available, the WST-1H· radical will give up its electron to generate superoxide anion as indicated in reaction (d). If SOD is present, the superoxide will be rapidly removed, driving reaction (d) to the right. As a consequence, WST-1H· availability will be limited and reaction (e) suppressed, leading to the inhibition of formazan formation. Thus, while SOD can inhibit WST-1 reduction under these conditions, superoxide anion may play no direct part in the reduction of the probe. Using a similar logic, WST-1 reduction in the absence of PMS may simply reflect the presence of plasma membrane oxidoreductases capable of performing a one-electron reduction of the probe:
The WST-1 radical would then go on to form the reduced formazan product as a result of dismutation, as indicated in reaction (e). Although SOD may well be suppressive under such circumstances, this is only the case when reducing the availability of WST-1H· as in (d) and is not necessarily evidence for primary superoxide anion production.
The trans-membrane NADH oxidase detected by Berridge and Tan (2000a,b) using PMS/WST-1 and the activity detected in human spermatozoa share similarities, not just because of their susceptibility to SOD but also because of the inhibitory effects of CAP, the potent vanilloid inhibitor, as well as NEM, a membrane permeant thiol-reactive agent. However, the profound inhibitory effect of the membrane impermeant thiol-blocking agent, pCMBS, and the high sensitivity to RA, contrasts with the trans-membrane NADH oxidase described by Berridge and Tan (2000a,b). Clearly the WST-1/PMS oxidoreductase activity observed in human spermatozoa is distinct from that observed in other cell types.
WST-1 reduction was also enhanced by the addition of either NADH or NADPH to suspensions of human spermatozoa, the former being the preferred substrate. In the presence of NADPH, the reduction of WST-1 was highly sensitive to pCMBS (Table I), in a similar fashion to the surface oxidase described by Berridge and Tan (2000a,b) in a variety of cell types (Jurkat, HeLa, 143B and 32Dc123 cells). However, the sperm NADPH oxidoreductase was also sensitive to NEM, a reagent that actually stimulated the surface oxidase described by Berridge and Tan (2000a,b) (Table I). The inhibitory effect of DPI on NADPH-induced WST-1 reduction suggests the involvement of a flavoprotein, while the lack of any inhibitory effect of SOD would be consistent with a direct two-electron reduction of the probe by a diaphorase-type enzyme:
|
Spermatozoa are known to possess a diaphorase-like activity but it is usually thought of as an enzymatic activity located in the cytosol (Caldwell et al., 1976; Gomez et al., 1996; Zini et al., 1998). Moreover, the fact that NADPH oxidation by human spermatozoa is insensitive to dicoumarol (Aitken et al., 1997) suggests that this two-electron, WST-1 reductase is not classical DT-diaphorase.
A completely different biochemical pathway was activated when NADH was used to provide a source of electrons for WST-1 reduction. In this case, SOD exerted a powerful inhibitory effect, as did pCMBS and CAP. Thus, in the presence of NADH, we can propose that a plasma membrane oxidoreductase effected the one-electron reduction of WST-1 to generate a radical species that then produced fully reduced product (WST-1H2) by dismutation (reaction e). By driving reaction (d) to the right, SOD would have suppressed the availability of WST-1H· radicals and thus inhibited the formation of soluble formazan. This NADH oxidoreductase activity was similar to the surface NADH oxidase described by Berridge and Tan (2000a,b) in being suppressed by pCMBS and SOD but differed in its sensitivity to CAP inhibition and lack of NEM stimulation. Thus, while cell-surface NADH oxidase activity has been observed on all cell types examined, the activity recorded in the presence of human spermatozoa seems to possess attributes that are unique to these cells (Table I).
The spontaneous lucigenin, luminol and isoluminol signals generated by purified human sperm suspensions from the high-density region of Percoll gradients were extremely low, as befits such functionally competent cells (Aitken and Clarkson, 1988; Ollero et al., 2001). On addition of NAD(P)H, only the lucigenin probe generated a signal. This signal was partially inhibited by SOD suggesting that superoxide anion was involved in the oxygenation of the probe. Whatever the chemistry of this chemiluminescent signal, it was quite distinct from the oxidoreductases observed with WST-1. Thus, in complete contrast to the WST-1/NADPH response, the NADPH/lucigenin signal could be inhibited by SOD but not NEM. Similarly, the NADH/lucigenin signal could not be inhibited with pCMBS or CAP, both of which had been such successful inhibitors of the WST-1/NADH response. Furthermore, only the lucigenin signals could differentiate between the high- and low-density Percoll fractions; the WST-1 signals did not differentiate between these two populations of cells, regardless of which promoter was used (PMS, NADH or NADPH).
The high levels of redox activity associated with the poor quality spermatozoa recovered in the low-density Percoll fractions has been reported by several independent laboratories (Aitken and Clarkson, 1988; Gomez et al., 1996; Sharma and Agarwal, 1996; Gil-Guzman et al., 2001; Ollero et al., 2001). In this study we were able to demonstrate that the high spontaneous lucigenin signals generated by low-density sperm populations could be powerfully inhibited by thiol reactive agents (pCMBS, NEM) and DPI, suggesting the involvement of cysteine-rich flavoproteins in the origins of this activity. At the doses used, DPI is also a potential inhibitor of mitochondrial ROS production (Li and Trush, 1998). However, the fact that rotenone had no impact on the signals recorded in this study, indicates that mitochondria are not contributing significantly to the redox status of human spermatozoa. The addition of the protein kinase C (PKC) activator, PMA, could also augment this signal via pathways that were inhibitable by DPI and NEM as well as SOD. The signals generated following the addition of PMA or NAD(P)H were not only highly correlated but, like the spontaneous lucigenin chemiluminescence, significantly elevated in the low-density Percoll fractions, suggesting a common causative mechanism. The ability of PMA to induce redox activity involves activation of PKC since this action can be blocked by staurosporine and cannot be induced by phorbol esters that are not activators of PKC, such as the 4-phorbol-12,13-didecanoate (unpublished observations). Whether defective sperm populations possess elevated PKC activity is not yet known. However, these defective sperm populations do possess a high proportion of cells that have retained excess residual cytoplasm (Gomez et al., 1996; Gil-Guzman et al., 2001; Ollero et al., 2001). Possibly, the additional cytosol and plasma membrane associated with such cytoplasmic remnants accounts for the elevated lucigenin reductase activities, by virtue of the increased availability of electron donors [NAD(P)H], reductases and activators, such as PKC. Why WST-1 reduction should not be elevated in the low-density sperm fractions for the same reason is difficult to discern. Both lucigenin and WST-1 are relatively membrane impermeant and both can potentially be activated by a one-electron reduction, yet only lucigenin can discriminate. Differences in the chemical structures of the probes and their physical properties must result in the targeting of separate biochemical pathways.
Banfi et al. (2001) have recently reported the presence of an NADPH oxidase (NOX5) in the male germ line that is sensitive to calcium, can serve as a proton transporter and generates signals that were detectable with luminol-peroxidase. In keeping with these observations we found that the induction of a calcium signal with A23187 in the presence of luminol-peroxidase consistently generated a chemiluminescent response in populations of human spermatozoa. Although SOD inhibited this response, superoxide anion production is probably a consequence rather than a cause of the observed chemiluminescence (Aitken et al., 1992a). Thus, the combination of hydrogen peroxide and HRP leads to a one-electron oxidation of the probe to create a luminol radical. The latter then interacts with ground state oxygen to create a superoxide anion that, in turn, reacts with luminol radicals to generate a luminol endoperoxide. This molecule then decomposes with the omission of light.
In common with PMS/WST-1 and NADH/WST-1, this response to A23187 was suppressed by CAP. In addition, DPI significantly inhibited this signal in keeping with the involvement of NOX5, a flavoprotein, in the genesis of the chemiluminescent response. One of the distinguishing characteristics of this calcium-dependent response was its resistance to pCMBS, in complete contrast to the redox activity detected by WST-1. Another interesting aspect was that although the A23187-induced, luminol-peroxidase signals were higher in the low- compared with the high-density Percoll fractions, this probe did not have the same ability as lucigenin to discriminate the dysfunctional sperm populations. Although both luminol- and lucigenin-dependent chemiluminescence ultimately depend upon oxygenation of the probe, the major difference is that creation of a reactive radical involves a one-electron reduction in the case of lucigenin and a one-electron oxidation in the case of luminol. Either superoxide anion generation or lucigenin reductase activity must be selectively elevated in the low-density Percoll fractions in order to account for this difference in probe sensitivity.
In conclusion, this study has demonstrated that human spermatozoa are redox-active cells that possess a variety of different electron transport systems associated with the plasma membrane. Some of these activities are highly active in purified suspensions of functional spermatozoa from the high-density region of Percoll gradients. This applies particularly to the A23187-induced luminol-peroxidase signals and the WST-1 responses, both of which may reflect redox activities involved in the physiological regulation of sperm function. The lucigenin-dependent activities, on the other hand, were elevated in the defective sperm populations isolated in the low-density fraction of discontinuous gradients. Further elucidation of the biochemical pathways responsible for these various forms of redox activity should contribute significantly to our understanding of the way in which cellular redox activity is involved in the physiological regulation and pathological disruption of human sperm function.
|
Acknowledgements |
|---|
The authors gratefully acknowledge the support of the Ernst Schering Trust and Rockefeller Foundation in the support of this research programme through the AMPPA network. We are also grateful to Jodie Powell for her help with the recruitment and management of the semen donor panel.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Aitken, R.J. and Baker, M.A. (2002) Reactive oxygen species generation by human spermatozoa: a continuing enigma. Int. J. Androl., 25, 191–194.[ISI][Medline]
Aitken, R.J. and Clarkson, J.S. (1987) Cellular basis of defective sperm function and its association with the genesis of reactive oxygen species by human spermatozoa. J. Reprod. Fertil., 81, 459–469.[Abstract]
Aitken, R.J. and Clarkson, J.S. (1988) Significance of reactive oxygen species and anti-oxidants in defining the efficacy of sperm preparation techniques. J. Androl., 9, 367–376.
Aitken, R.J. and Fisher, H. (1994) Reactive oxygen species generation and human spermatozoa: the balance of benefit and risk. Bioessays, 16, 259–267.[CrossRef][ISI][Medline]
Aitken, R.J. and Krausz, C.G. (2001) Oxidative stress, DNA damage and the Y chromosome. Reproduction, 122, 497–506.[Abstract]
Aitken, R.J., Clarkson, J.S. and Fishel, S. (1989a) Generation of reactive oxygen species, lipid peroxidation and human sperm function. Biol. Reprod., 41, 183–187.[Abstract]
Aitken, R.J., Clarkson, J.S., Hargreave, T.B., Irvine, D.S. and Wu, F.C.W. (1989b) Analysis of the relationship between defective sperm function and the generation of reactive oxygen species in cases of oligozoospermia. J. Androl., 10, 214–220.
Aitken, R.J., Buckingham, D.W. and West, K.M. (1992a) Reactive oxygen species and human spermatozoa; analysis of the cellular mechanisms involved in luminol- and lucigenin-dependent chemiluminescence. J. Cell. Physiol., 151, 466–477.[CrossRef][ISI][Medline]
Aitken, R.J., Buckingham, D., West, K., Wu, F.C., Zikopoulos, K. and Richardson, D.W. (1992b) Differential contribution of leucocytes and spermatozoa to the high levels of reactive oxygen species recorded in the ejaculates of oligozoospermic patients. J. Reprod. Fertil., 94, 451–462.[Abstract]
Aitken, R.J., Buckingham, D. and Harkiss, D. (1993a) Use of a xanthine oxidase oxidant generating system to investigate the cytotoxic effects of reactive oxygen species on human spermatozoa. J. Reprod. Fertil., 97, 441–450.[Abstract]
Aitken, R.J., Harkiss. D. and Buckingham, D.W. (1993b) Analysis of lipid peroxidation mechanisms in human spermatozoa. Mol. Reprod. Dev. 35, 302–315.[CrossRef][ISI][Medline]
Aitken, R.J., Paterson, M., Fisher, H., Buckingham, D.W. and van Duin, M. (1995) Redox regulation of tyrosine phosphorylation in human spermatozoa and its role in the control of human sperm function. J. Cell Sci., 108, 2017–2025.[Abstract]
Aitken, R.J., Buckingham, D.W., Harkiss, D., Fisher, H., Paterson, M. and Irvine, D.S. (1996) The extragenomic action of progesterone on human spermatozoa is influenced by redox regulated changes in tyrosine phosphorylation during capacitation. Mol. Cell. Endocrinol., 117, 83–93.[CrossRef][ISI][Medline]
Aitken, R.J., Fisher, H.M., Fulton, N., Gomez, E., Knox, W., Lewis, B. and Irvine, S. (1997) Reactive oxygen species generation by human spermatozoa is induced by exogenous NADPH and inhibited by the flavoprotein inhibitors diphenylene iodonium and quinacrine Mol. Reprod. Dev., 47, 468–482.