Differences in ATP Generation Via Glycolysis and Oxidative Phosphorylation and Relationships with Sperm Motility in Mouse Species*

Background: Sperm from mouse species may produce ATP through glycolysis or respiration. Results: Sperm with high respiration/glycolysis ratio and high reliance on respiration produce more ATP and swim faster. Conclusion: The usage ratio of ATP production pathways defines sperm motility in mouse species. Significance: Sperm metabolism in mice has evolved ways to produce faster sperm. Mouse sperm produce enough ATP to sustain motility by anaerobic glycolysis and respiration. However, previous studies indicated that an active glycolytic pathway is required to achieve normal sperm function and identified glycolysis as the main source of ATP to fuel the motility of mouse sperm. All the available evidence has been gathered from the studies performed using the laboratory mouse. However, comparative studies of closely related mouse species have revealed a wide range of variation in sperm motility and ATP production and that the laboratory mouse has comparatively low values in these traits. In this study, we compared the relative reliance on the usage of glycolysis or oxidative phosphorylation as ATP sources for sperm motility between mouse species that exhibit significantly different sperm performance parameters. We found that the sperm of species with higher oxygen consumption/lactate excretion rate ratios were able to produce higher amounts of ATP, achieving higher swimming velocities. Additionally, we show that the species with higher respiration/glycolysis ratios have a higher degree of dependence upon active oxidative phosphorylation. Moreover, we characterize for the first time two mouse species in which sperm depend on functional oxidative phosphorylation to achieve normal performance. Finally, we discuss that sexual selection could promote adaptations in sperm energetic metabolism tending to increase the usage of a more efficient pathway for the generation of ATP (and faster sperm).

Mammalian spermatozoa move forward as a result of thrust generated by the flagellum, a cell component containing the axoneme whose microtubules are associated with large ATPases (dyneins). Sperm motility is directly dependent upon the availability of energy obtained through ATP hydrolysis (1)(2)(3) and accounts for about 70% of total ATP consumption (4). Because spermatozoa are transcriptionally inactive and have a small cytoplasmic volume, their ability to adjust their enzymatic load is very limited (5). In addition, sperm face extensive changes with regard to the availability of oxygen and exogenous metabolic substrates in the different regions of the female tract as they travel to the site of fertilization (2,6,7). Previous studies have hypothesized that mammalian sperm have overcome these limitations by evolving flexible and adaptable metabolic processes that are compartmentalized to specific regions of the cell and function in a localized manner (3,5,8,9).
Mammalian sperm rely mainly on two metabolic pathways to produce ATP, namely oxidative phosphorylation (OXPHOS) 4 and anaerobic glycolysis, which are localized to different regions of the cell. On the one hand, OXPHOS takes place in the mitochondria, which are tightly packed in the sperm mid-piece. Numerous studies have provided evidence supporting the role of OXPHOS as the main ATP provider for sperm motility in several mammalian species; in these species, mitochondrial membrane potential and oxygen consumption rate are positively associated with ATP content, proportion of motile sperm, and sperm velocity (1)(2)(3)10). On the other hand, glycolysis takes place in the principal piece, which occupies the major part of the flagellum. Several glycolytic enzymes have been identified mainly in the fibrous sheath of the principal piece in mammalian species (11). Because OXPHOS yields more ATP per mol of glucose than glycolysis, mitochondria are abundant in the mid-piece of mammalian sperm and may use a wider range of substrates (fatty acids, monocarboxylic acids, and amino acids); respiration has been historically regarded as the main source of ATP production for sperm motility, relegating glycolysis to a secondary role (12)(13)(14)(15). However, there is a wide range of variation among mammalian species with regard to the primary mechanism for ATP production. There are species whose sperm have high respiration rates and cannot support motility with glycolysis alone (boar and horse), species with both high respiration and glycolysis (bull and guinea pig), and others with sperm that rely mainly on glycolysis (human) (9).
In the mouse, numerous experiments on epididymal sperm revealed that both glycolysis and OXPHOS are able to sustain vigorous sperm motility in the presence of their specific substrates (5, 10, 16 -19). Furthermore, mouse sperm are not able to sustain ATP levels and progressive motility when treated with uncoupler agents (17,18) or respiratory inhibitors (10) in glucose-free media. However, two studies provided evidence that an active glycolytic pathway is essential for mouse sperm motility and male fertility. Using glyceraldehyde phosphate dehydrogenase knock-out mice (20) or chemical inhibition of glycolysis (21), these studies showed that ATP production and motility were arrested when glycolysis was halted, even if sperm mitochondria were fully functional. The conclusion that glycolysis was essential for mouse sperm was challenged (1-3, 10) on the basis that in such studies glycolysis was interrupted after its ATP-consuming phase but before its ATP synthesis phase, turning this process into an ATP consumer. Additional studies were carried out using knock-out mice for other glycolytic enzymes such as phosphoglycerate kinase 2 (PGK2) (22), enolase 4 (Eno4) (23), and lactate dehydrogenase C (LDHC) (18,24,25). PGK2 catalyzes steps before the ATP-producing phase of the glycolysis, whereas the others are involved in steps during (Eno4) and after (LDHC) the ATP-producing phase. Such studies found that sperm from all these knock-out mice have impaired sperm motility and that males experience fertility loss, while maintaining normal mitochondrial activity (at least in the LDHC knock-out). In addition, multiple isoforms of glycolytic enzymes in mouse sperm are germ cell-specific, having their expression restricted to spermatogenesis or being produced as sperm-specific splicing variants. These include glyceraldehyde-3-phosphate dehydrogenase-S (GAPDHS), PGK2, LDHC, aldolase A, Eno4, muscle-type phosphofructokinase, and hexokinase (23, 26 -32). Many of these enzymes are localized to the principal piece of mouse sperm in close proximity to the axoneme, both in the liposoluble fraction (hexokinase, PGK2, Eno4, and LDHC) and in association with the fibrous sheath (aldolase A, GAPDHS, pyruvate kinase, and lactate dehydrogenase A) (8,33,34). The unique features of these isozymes may respond to the specificity of the localization of the glycolytic pathway in the principal piece of the sperm flagellum (35). A fully active glycolytic pathway is required for multiple steps in the fertilization cascade, including capacitation-dependent tyrosine phosphorylation, hyperactivated motility, and oocyte penetration (5,17,18,34,36,37). Taken together, evidence suggests that glycolysis is the main source of ATP required to sustain the motility of mouse spermatozoa.
Importantly, knowledge of mouse sperm bioenergetics derives from studies of the laboratory mouse (Mus musculus) and, in particular, of a few commonly used mouse strains (CD1, C57BL, and 129SvEv). Recent comparative studies in several muroid rodents have revealed considerable differences in sperm traits and performance within this group of rodents (38 -41, 68). A strong association was found across species between the content of sperm ATP and both the proportion of motile sperm and sperm swimming velocity in freshly collected spermatozoa (40) and after a period of incubation (68). These results underscore the link between ATP and sperm performance and that sperm from different species may have different demands for efficient fertilization. Interestingly, M. musculus exhibited comparatively low ATP content, low percentage of motile sperm, and low sperm swimming velocity in relation to other mouse species. This may perhaps relate to the observation that this species under natural conditions, although under selection for fertilization efficiency, may experience a lower selective pressure for high performance derived from a malemale postcopulatory competitive scenario. Thus, in M. musculus, glycolysis, being a less efficient pathway for ATP generation, may suffice for energy production, whereas in other mouse species, sperm that are required to perform at a higher level may need to resort to additional and more efficient ATPsupplying pathways.
Therefore, the differences observed between mouse species in an earlier comparative study, and the fact that spermatozoa from M. musculus rely highly on glycolysis, prompted the following question. Could the differences in ATP content and sperm swimming velocity between mouse species be related to variation in the usage of different metabolic pathways to generate ATP? In other words, because OXPHOS typically represents a more efficient (per mol of glucose) generation of ATP in comparison with glycolysis, could the higher ATP content and swimming velocity of sperm from other mice result from a relatively higher reliance on OXPHOS for ATP and sperm motility? A prediction that follows from this question is that species with higher reliance on OXPHOS to generate ATP for sperm motility would suffer a more pronounced decrease in sperm ATP content and swimming performance when OXPHOS is inhibited. The aim of this study was therefore to compare the relative reliance on the usage of OXPHOS and glycolysis as the ATP source to sustain sperm motility between mouse species that exhibit significantly different sperm performance parameters.

Animals and Sperm Collection
This study focused on three mouse species as follows: the house mouse (M. musculus), from which the majority of laboratory mouse strains are derived; the western Mediterranean mouse (Mus spretus), and the steppe mouse (Mus spicilegus). These three species represent a closely related monophyletic group with an estimated divergence time of ϳ1.7 million years (42,43). However, comparative studies have revealed notable differences between these species regarding sperm numbers, percentage of motile sperm, swimming velocity, and ATP content (38 -41, 68).
Adult males (4 -6 months old) of the three species were acquired from the Institut des Sciences de l'Evolution, CNRS-Université Montpellier 2, France. These animals belong to the following wild-derived strains (which have been kept in captivity for only a few generations): M. musculus, strain MPB (from Bialowieza, Poland); M. spretus, strain SEB (from Barcelona, Spain), and M. spicilegus, strain ZRU (from Kalomoyevka, Ukraine). Animals were maintained under standard conditions (14 h of light and 10 h of darkness, 22-24°C), with food and water provided ad libitum. Each male to be used in this study was housed alone (i.e. in individual cages) for at least 2 weeks before sperm collection. Males were sacrificed by cervical dislocation and weighed immediately. Testes were then removed and weighed. All procedures followed Spanish Animal Protec-  (44). To remove substrates present in the epididymal fluid, the original sperm suspension was transferred to a tube and centrifuged for 2 min at 500 ϫ g, after which the supernatant was removed, and the sperm were resuspended in mT-H. Then, sperm concentration was estimated by using a modified Neubauer chamber and adjusted to ϳ20 ϫ 10 6 sperm ml Ϫ1 by adding mT-H. The resulting sperm suspension was transferred to plastic tubes, and sperm parameters were assessed immediately. In all procedures, large-bore pipette tips were used to minimize damage to sperm membranes. Five individuals of each species were used for each set of experiments as follows: set 1, oxygen consumption rate and extracellular acidification rate measurements; set 2, motility, velocity, and ATP content measurements; set 3, lactate excretion rate measurements.

Oxygen Consumption Rate (OCR) and Extracellular Acidification Rate (ECAR) Assessment
An XF24 extracellular flux analyzer (Seahorse Bioscience, North Billerica, MA) was used to measure the real time OCR (pmol of min Ϫ1 ) and ECAR (milli-pH min Ϫ1 , an estimate of the rate of lactate release to the medium) of mouse sperm. Because the ECAR measurement method precludes the use of a pH buffer in the medium, an unbuffered mT medium in which Hepes was replaced with NaCl and pH was adjusted to 7.4 at 37°C was used instead of mT-H for sperm collection, dilution, and incubation. After sperm concentration was adjusted, 2 ϫ 10 6 sperm cells were seeded on each of 20 wells of a XF24 plastic microplate that had been previously coated with concanavalin A to ensure sperm adhesion. Plates were coated by placing 20 l of a 0.5 mg/ml concanavalin A solution in each well and letting it dry. Four wells were left without cells to perform background corrections. After 1 min, the plate was centrifuged for 2 min at 1200 ϫ g. This centrifugation was repeated changing the plate orientation to ensure an even adhesion of cells to the bottom of the well. The supernatant was removed from each well, and 500 l of medium were added per well. Then, the plate was placed in the instrument. All the OCR and ECAR measurements were performed at 37°C under air. Taking account the time required for cell counting, seeding, and instrument equilibration, the first OCR and ECAR measurements were taken 44 min after collection of samples. Both OCR and ECAR were recorded for the next 21 min. Subsequently, 1 M carbonyl cyanide p-trifluoromethoxyphenylhydrazone ("FCCP" group) or mT medium ("control" group) were added to each well to measure maximum and spare OCR (because FCCP acts as an uncoupler of mitochondrial respiration). Measurements continued for 26 min, after which 1 M rotenone and 1 M antimycin A were added to all wells to quantify the mitochondria-independent OCR. OCR and ECAR were monitored for an additional 16 min, and the measurement run ended ϳ107 min after sperm collection. A schematic representation of the experimental design is shown in Fig. 1a. After the measurement run ended, sperm in each well were fixed by adding 100 l of 4% formaldehyde solution, they were detached from the bottom of the well (complete detachment was confirmed under the microscope) and counted using a modified Neubauer chamber. The OCR and ECAR values for each well were normalized by the number of sperm present and are reported as amol of O 2 min Ϫ1 sperm Ϫ1 and nano-pH min Ϫ1 sperm Ϫ1 . Basal mitochondrial OCR was calculated by subtracting the OCR values obtained after the addition of antimycin A and rotenone from the OCR values of the first 21 min of the experiment. Maximum mitochondrial OCR was calculated by subtracting the OCR values obtained after the addition of antimycin A and rotenone from the OCR values recorded in the FCCP group in the 26 min after FCCP addition. Spare OCR capacity was calculated by subtracting the basal OCR values from the maximum OCR values. Basal ECAR was calculated averaging the ECAR values obtained in the first 21 min of the experiment.

Sperm Motility and Swimming Velocity
Assessments of these parameters were adjusted to follow the timing of the OCR and ECAR measurements (see Fig. 1b). Sperm suspensions (20 ϫ 10 6 sperm ml Ϫ1 in mT-H) were incubated in plastic tubes at 37°C under air during 56 min (corresponding to the third OCR-ECAR measurement), and sperm parameters were recorded (hereafter, "basal" conditions). Subsequently, the suspensions were divided in two aliquots, incubated for another 35 min, and either 1 M antimycin A plus 1 M rotenone ("AA ϩ R" group), or an equivalent volume of mT-H (control group) were added, coinciding with the second addition of the OCR-ECAR experiment. After 16 min of incubation, sperm parameters were assessed. Sperm motility was evaluated by examining 10 l of the sperm suspension that was placed between a pre-warmed slide and a coverslip at ϫ100 magnification under phase contrast optics. The percentage of motile (MOT) sperm was estimated by at least two independent, experienced observers; estimations from the different observers were averaged and rounded to the nearest 5% value. Additionally, the quality (Q) of sperm movement was ranked in a scale from 1 to 5 (from least to most vigorous movement) and a sperm motility index (SMI) was calculated using the following equation: SMI ϭ (Q ϫ 20 ϩ MOT)/2.
To assess sperm velocity, an aliquot of sperm suspension was diluted to ϳ5 ϫ 10 6 sperm ml Ϫ1 , placed in a pre-warmed microscopy chamber with a depth of 20 m (Leja, Nieuw-Vennep, The Netherlands), and filmed at ϫ40 using a phase contrast microscope connected to a digital video camera (Basler A312fc, Vision Technologies, Glen Burnie, MD). A minimum of 150 sperm trajectories was assessed using a computer-assisted sperm analyzer (Sperm Class Analyzer version 4.0, Microptic, Barcelona, Spain), and the following velocity parameters were estimated for each trajectory: curvilinear velocity (VCL, m s Ϫ1 ), straight line velocity (VSL, m s Ϫ1 ), average path velocity (VAP, m s Ϫ1 ), linearity (LIN ϭ VSL/VCL), straightness (STR ϭ VSL/VAP), wobble (WOB ϭ VAP/VCL), amplitude of lateral head displacement (ALH, m), and beatcross frequency (BCF, Hz). The final value for these parameters was calculated as the mean of all the individual trajectories for each control or treated sample.

Sperm ATP Content
Samples for ATP quantification were obtained at the same time points described above (see Fig. 1b). Sperm ATP content was measured using a luciferase-based ATP bioluminescent assay kit (ATP Bioluminescence Assay Kit HS II, Roche Applied Science). A 100-l aliquot of diluted sperm suspension was mixed with 100 l of cell lysis reagent, vortexed, and incubated at room temperature for 5 min. The resulting cell lysate was centrifuged at 12,000 ϫ g for 2 min, and the supernatant was recovered and frozen in liquid N 2 . Bioluminescence was measured in triplicate in 96-well plates using a luminometer (Synergy HT, Biotek Instruments Inc.). In each well, 50 l of luciferase reagent was added to 50 l of sample (via auto-injection), and following a 1-s delay, light emission was measured over a 5-s integration period. Standard curves were constructed using solutions containing known concentrations of ATP diluted in mT-H and cell lysis reagent in a proportion equivalent to that of the samples. ATP content was expressed as amol of sperm Ϫ1 .

Sperm Lactate Excretion Rate
To examine whether the ECAR values obtained using the Seahorse XF24 extracellular flux analyzer mirrored values of lactate excretion or other acidifying compounds, the sperm basal lactate excretion rate (LER) was measured. At 40, 50, 60, and 70 min after swim out, an aliquot of sperm was centrifuged at 2500 ϫ g for 2 min, and the supernatant was frozen in liquid N 2 . Lactate levels in the supernatant (extracellular medium) were determined using a commercial kit (lactate assay kit K607, BioVision, Mountain View, CA) based on an enzymatic reaction by lactate oxidase and interaction of the product with a probe to produce color. After preparing the reaction mix according to the supplier's instructions, optical density was measured at ϭ 570 nm using a microplate reader (Varioskan Flash, Thermo Fisher Scientific Inc.). Standard curves were constructed using solutions containing known concentrations of lactate diluted in mT-H and cell lysis reagent in a proportion equivalent to that of the samples. LER was estimated through linear least squares regression using the four points provided by the measurements and was expressed as amol min Ϫ1 sperm Ϫ1 .

Data Analysis
Principal Component Analyses for Sperm Velocity Parameters-Because the sperm velocity parameters tend to be highly correlated (39), principal component analyses (PCA) were performed to obtain summarized variables to integrate the velocity information. The variables were divided in two groups termed sperm "velocity" (VCL, VSL, and VAP) and sperm "trajectory shape" (LIN, STR, WOB, ALH, and BCF), and one independent PCA was carried out for each group. The first principal component for the velocity group (VPC1) accounted for 97.5% of the variability in the three summarized variables (VCL, VSL, and VAP), whereas the second principal component (VPC2) only accounted for 2.5%. The values for each of the three sperm velocity descriptors were significantly correlated with VPC1, but they showed no significant correlation with VPC2 (data not shown). In the case of the trajectory shape group, the first principal component (TPC1) accounted for 88.3% of the variability, and the second principal component (TPC2) represented 8.7%. The values of all five variables in this group (LIN, STR, WOB, ALH, and BCF) were significantly correlated with TPC1, and only ALH and BCF showed a significant correlation with TPC2 (data not shown). Thus, considering the high percentage of variability absorbed by the PC1 in both variable groups, PC1 values for each treatment and species (hereafter referred to "overall sperm velocity" and "overall trajectory shape") were used as our integrated sperm velocity measures.
Length-adjusted Metabolic Measures-Because larger cells might contain greater quantities of ATP, differences in cell size should be taken into account when testing possible variations in sperm ATP concentration. We calculated the length-adjusted ATP concentration (amol m Ϫ1 ) as the ratio between the amount of ATP per sperm for each species and its total sperm length (40). Furthermore, because the two metabolic pathways evaluated in this study take place in two different regions of the sperm flagellum, midpiece (OXPHOS) and principal piece (glycolysis), and these regions vary in size between the species evaluated, we chose to proceed in the same way with OCR, ECAR, and LER values. Thus, we estimated length-adjusted OCR (amol O 2 min Ϫ1 m Ϫ1 ) using the midpiece length and the length-adjusted ECAR (nano-pH min Ϫ1 m Ϫ1 ) and LER (amol min Ϫ1 m Ϫ1 ) using the principal piece length. Total sperm length, midpiece length, and principal piece length was obtained for the three species (68).
Statistical Analyses-The effect of membrane decoupling and OXPHOS inhibition on OCR was analyzed with a twofactor repeated-measures ANOVA for each species, using treatment and time as factors (with three and nine levels, respectively). Differences between conditions were analyzed through a post hoc multiple comparisons test using the Bonferroni correction. Basal values for MOT, SMI, velocity variables, and their summarized parameters, OCR (basal, maximum, and spare), ECAR, and LER, were compared between species by means of a one-factor ANOVA using species as a factor. A post hoc test (45) was used to determine pairwise differences between species. To evaluate the influence of OXPHOS inhibition in sperm performance parameters, the control and AA ϩ R groups were compared for each species by means of paired t test. All variables were log 10 -transformed for statistical purposes, with the exception of percentages MOT, LIN, STR, and WOB, which were arcsine-transformed. The statistical analyses were performed using SPSS Statistics (SPSS version 21.0.0.1; SPSS, IBM Corp., Somers, NY) and InfoStat version 2011p (Grupo Infostat, Universidad Nacional de Córdoba, Córdoba, Argentina) with ␣ ϭ 0.05.

Interspecific Comparison of OCR, Lactate Production and
Sperm Performance Traits-The extracellular flux measurements revealed that sperm oxygen consumption maintained a relatively steady rate in control conditions for the three species throughout the duration of the experiment (Fig. 2, a-c). Moreover, the sperm of the three species responded to the addition of FCCP with a significant increase in OCR and showed a significant decrease in this parameter under the influence of antimycin A ϩ rotenone in both treatments (control and FCCP) (Fig. 2, a-c). In addition, ECAR values were also stable for the most part of the experiment, showing a slight increase after the addition of antimycin A ϩ rotenone in M. musculus and M. spretus (Fig. 2, d-f).
Although the overall pattern of stimulation and inhibition of OCR was similar for the three species, significant quantitative differences were observed. Basal OCR values, both absolute and length-adjusted, were significantly higher in M. spretus and M. spicilegus than in M. musculus ( Fig. 3a and Table 1). Moreover, after the addition of FCCP, the same significant differences were observed in the maximum OCR ( Fig. 3b and Table 1) and spare OCR capacity ( Fig. 3c and Table 1). The average percentage of OCR stimulation caused by the uncoupling of mitochondrial respiration revealed interspecific differences (M. musculus, 211%; M. spretus, 247%; and M. spicilegus, 278%); nonetheless, these differences did not reach statistical significance (ANOVA, F ϭ 1.47, p ϭ 0.2714). In contrast, basal LER ( Fig. 3d and Table 1) and ECAR (Table 1) values were significantly higher in M. musculus than in the other two species in both their absolute and length-adjusted measures. The remarkable similarity of the interspecific patterns shown by the ECAR and LER measurements (Fig. 4) suggests that most of the extracellular acidification recorded in our extracellular flux experiments do correspond to the excretion of lactate by sperm.
In addition, we calculated the OCR/ECAR ratio as a measure of the relative magnitude of OXPHOS versus glycolysis in each species. M. spretus and M. spicilegus presented a significantly higher OCR/ECAR ratio than M. musculus ( Fig. 5a (40,68) have reported a significant relation between differences in ATP content and sperm performance in muroid rodents. Our results confirmed these findings because M. spretus and M. spicilegus showed a higher percentage in motile sperm (Table 1) and a higher sperm motility index ( Fig. 5c and Table 1). Moreover, the three velocity parameters (VCL, VSL, and VAP) ( Table 1), along with their summarized scores (overall sperm velocity; VPC1) ( Fig. 5d and Table 1) showed higher values in M. spretus and M. spicilegus than in M. musculus. Finally, when the descriptors of sperm trajectory shape were analyzed, M. spretus and M. spicilegus showed sperm with more linear and progressive trajectories (higher LIN, STR, and WOB), and more frequent (higher BCF) but less pronounced (lower ALH) lateral displacement (Table 1). These differences were also reflected in their summa- rized scores for these variables (overall trajectory shape; TPC1), which were significantly higher in M. spretus and M. spicilegus than in M. musculus ( Table 1).
Effect of OXPHOS Inhibition in the Sperm of Three Species of the Genus Mus-The inhibition of mitochondrial respiration affected the sperm of the three species in different manner and intensity. Significant differences between the control and AA ϩ R groups are summarized for each species in Table 2. In the case of M. musculus sperm, no significant differences were registered between groups in any of the ATP values (absolute and length-adjusted) (Fig. 6a), motile sperm percentage, SMI (Fig.  6b), and the summarized velocity (Fig. 6c) and trajectory shape variables (Fig. 6d). When analyzing the individual velocity and trajectory shape variables, we only found a significant 6% decrease in the VCL as a result of the inhibition of OXPHOS (Table 2). However, M. spretus and M. spicilegus showed significant alterations in sperm motility and energy production as a response of OXPHOS inhibition. In both species, sperm ATP content (both absolute and length-adjusted) (Fig. 6a), motile sperm percentage, and SMI (Fig. 6b), decreased as a result of the incubation in the presence of 1 M antimycin A and 1 M rotenone. However, although M. spicilegus sperm decreased both their overall sperm velocity (Fig. 6c) and overall trajectory shape (Fig. 6d) values under the effect of OXPHOS inhibitors, M. spretus sperm only showed significant decreases in the first of the two variables (Fig. 6, c and d). According to these results, the separate analysis of the sperm velocity and trajectory-shape variables revealed a significant decrease in the eight variables  measured for M. spicilegus (Table 2) but only in the three variables that composed the overall sperm velocity score (VCL, VSL, and VAP) for M. spretus (Table 2). Remarkably, after the incubation under respiration inhibiting conditions, M. spicilegus sperm parameters were very similar to those of the M. musculus control group (length-adjusted ATP concentration F ϭ 2.85, p ϭ 0.1256; SMI F ϭ 2.39, p ϭ 0.1563; overall sperm velocity F ϭ 0.08, p ϭ 0.7887; overall trajectory shape, F ϭ 0.50, p ϭ 0.4960) (Fig. 6).

Differences in Sperm Energetic Metabolism Translate into Variations in Sperm ATP Production and Swimming
Performance-The results of our study revealed the existence of significant variance in the bioenergetics metabolism of three different species of the genus Mus. Our study has in fact two main premises as follows: (a) that the variability in sperm performance seen in species of the genus Mus is a consequence of differences in the usage ratio of two ATP generation pathways (OXPHOS and glycolysis), and (b) that these differences could be adaptations to different levels of sperm competition (because faster sperm are more competitive). In this sense, our analyses showed that the sperm of M. musculus have a lower basal and maximum respiration rate than M. spretus and M. spicilegus, but it excretes lactate to the extracellular medium at a higher rate. Furthermore, these divergences regarding the metabolic pathways usage ratios, translate into variations in sperm ATP content and sperm performance. Thus, in accordance with our hypothesis, the species with higher sperm respiratory activity also showed higher sperm ATP content, percentage of motile cells, sperm motility index, faster sperm velocity, and more linear and progressive trajectories. Previous studies comparing the sperm of hominid (46) and felid (47) species have also found differences in the level of sperm metabolic traits (mitochondrial membrane potential). More specifically, a recent study found significant divergences in the levels of glucose consumption and glycolytic enzyme activity between two strains of laboratory mice (18). However, in all these cases, the differences that were registered at basal metabolic levels did not appear to translate into divergences in sperm performance traits (18,47) or ATP content (18). To our knowledge, this study is the first to produce evidence relating variations in the usage of sperm metabolic pathways to differences in sperm performance among closely related mammalian species.
Our results indicate that a higher basal rate of OXPHOS could benefit sperm performance by increasing the amount of ATP available for flagellar motility. Moreover, this trend corresponded with a decrease in the rate at which lactate was excreted to the extracellular medium, which would intuitively lead to infer a reduced glycolytic activity in these cells. Yet, this would constitute a somewhat premature conclusion; although sperm function as a highly compartmentalized cell (48), interactions between components of glycolytic and respiratory pathways have been found in the mouse sperm (17). As an example, the pyruvate dehydrogenase complex (usually located in the mitochondrial matrix), has been detected in the principal piece of hamster sperm (49). Furthermore, the GLUT8 glucose transport facilitator has been found in the mouse sperm midpiece where it is required for the maintenance of mitochondrial membrane potential (50).
Regardless of whether lactate (51)(52)(53) or pyruvate (18,34,54) is the primary respiratory substrate entering mouse sperm mitochondria, the composition of the medium used in our experiments made active glycolysis a prerequisite for OXPHOS. Although the physiological (in vivo) relevance of the uptake by mitochondria of glycolytic products excreted by the principal piece is uncertain, the different regions of the mouse female reproductive tract (vagina, uterus, and oviduct) present relatively high lactate/glucose ratios around the time of ovulation (2, 7) that would enable sperm to import sufficient exogenous lactate to fuel OXPHOS (52,53,55,56). Furthermore, when sperm are suspended in seminal plasma, fructose is by far the main metabolic substrate (2). Under these conditions, the ability to incorporate sperm-excreted lactate would enable these cells to produce ATP via both metabolic pathways (OXPHOS and glycolysis). Thus, the onset of OXPHOS using sperm-derived lactate could indeed function to "jump-start" sperm metabolism upon sperm activation when sperm plus seminal plasma enter the vagina.
As a consequence of our experimental setup, the total amount of lactate present in the extracellular medium is the result of the balance between the lactate produced by the glycolytic activity of the cell and the lactate used as respiratory substrate (possibly after an LDH-catalyzed conversion to pyruvate), making the OCR/LER ratio (the amount of O 2 consumed per mol of lactate excreted) a rather useful measure of the relative usage of each metabolic pathway. Altogether, the comparison of the sperm metabolic output between the three species studied suggests that those sperm in which the products of glycolysis enter the OXPHOS pathway in a higher proportion are able to produce higher amounts of ATP, achieving higher swimming velocities.

Inhibition of OXPHOS Affects Sperm ATP Production and Swimming Performance in Species with High Respiratory
Rates-The inhibition of mitochondrial respiration elicited a decrease of sperm ATP content and sperm performance in the species with high respiration rates (M. spretus and M. spicilegus), while having no significant effect (beyond a 6% reduction in one velocity parameter) in M. musculus. Furthermore, although the sperm of both species with high respiration rates showed a significant reduction of sperm ATP content (Ϫ30% in M. spretus, Ϫ40% in M. spicilegus), their swimming performance was affected with different intensity by OXPHOS inhibition. In the case of M. spretus, sperm showed a moderate reduction of motility (SMI ϭ Ϫ10%) and the variables related to swimming velocity (VCL ϭ Ϫ7%, VSL ϭ Ϫ11%, and VAP ϭ Ϫ10%) when OXPHOS was inhibited. A more pronounced decrease in these variables was observed in M. spicilegus sperm (SMI ϭ Ϫ25%, VCL ϭ Ϫ19%, VSL ϭ Ϫ39%, and VAP ϭ Ϫ30%), in addition to significant variations of the sperm trajectory shape toward less linear trajectories (LIN ϭ Ϫ25%, STR ϭ Ϫ11%, and WOB ϭ Ϫ14%) with higher lateral displacements (ALH ϭ 20% and BCF ϭ Ϫ39%).
The relative importance of OXPHOS versus glycolysis in the production of ATP for mouse sperm motility is still a matter of debate (1)(2)(3)57). On the one hand, germ cell-specific variants of genes that encode for OXPHOS-related proteins, such as cytochrome c (58), succinyl-CoA transferase (59), and pyruvate dehydrogenase (60), have been identified in mouse sperm. Furthermore, male mice knock-out for testis-specific cytochrome c (16) and choline dehydrogenase (crucial for the maintenance of mitochondrial membrane potential) (61) have diminished motility, ATP levels, and fertility, and mtDNA mutator phenotype strain males are virtually infertile (62). In addition, several studies have shown that laboratory mouse sperm are able to rely solely on respiratory substrates to support motility in the absence of glucose (17,21,37). Notably, this ability is disrupted when glycolysis is inhibited and glucose is also present as a substrate (18 -21). On the other hand, glycolysis has been shown to be able to sustain the levels of ATP required for normal motility in mouse sperm, even after the inhibition of OXPHOS in the presence of oxidative substrates (17,18,20,22). The reliance of murine sperm on active glycolysis to achieve normal sperm motility has been recently explained in terms of metabolic compartmentalization (19); these authors observed that the inhibition of glycolysis, even in sperm supplied with respiratory substrates, promoted a decrease in the bending angle at the distal end of the flagellum. They proposed that, besides producing ATP when provided with glycolysable substrates, glycolytic enzymes located throughout the principal piece (PGK and GAPDH) would catalyze sequential rapid equilibrating reactions constituting an ATP-transferring mechanism from high ATP concentration zones (mitochondria) to low ATP concentration ones (distal flagellum), a process that would not be possible by passive ATP diffusion (63,64).
Our analyses on M. musculus sperm appear to confirm this previous evidence because its sperm appear to ignore the effects of OXPHOS inhibition. However, our main result revealed, for the first time, that in other species of the genus Mus (M. spretus and M. spicilegus) sperm depend on a functional respiratory pathway to achieve normal ATP production and motile performance. In addition, our results suggest that the species with more intense respiratory function and higher OCR/LER ratio (M. spicilegus) would have a higher degree of reliance upon OXPHOS. On a final note, the differences in the decrease of sperm performance under respiratory inhibition between M. spretus and M. spicilegus could be explained in terms of compensation between metabolic pathways. Of these two species, only M. spretus sperm showed a significant increase in lactate excretion rate after the addition of respiratory inhibitors. Thus, the sperm of M. spretus could be (at least partially) compensating for the decrease in ATP levels caused by respiratory inhibition by increasing their glycolytic rate.
Sperm Competition as a Possible Explanation for the Evolution of Sperm Metabolism-The evolution of differences in sperm metabolism between closely related mouse species may be regarded as an adaptive response to selective pressures imposed by sperm competition. Sperm competition is a phenomenon that arises from female promiscuity, in which case sperm from more than one male would compete to fertilize a given set of ova (65). In the case of muroid rodents, sperm competition has been found to promote a high number of adaptations in sperm traits tending to increase their competitiveness. Comparative studies on this group (which included numerous species of the genus Mus) have revealed that species with high levels of sperm competition have higher sperm numbers, percentages of acrosomal integrity, morphologically normal sperm, motile sperm (38), and higher sperm velocity (39, 40). In the particular case of sperm velocity, it has been shown to evolve through adaptations in regulatory processes (sequence of ion channels in the sperm tail (41)) and sperm structure (evolution of protamines and head morphology (39, 66)). More importantly, previous results from our group indicate that the increase in sperm swimming velocity in species with high sperm competition levels is mainly related to an elevation in sperm ATP production (40,68).
Sperm competition levels in mammals can be inferred by measuring the mass of the testes of a species in relation to its body mass (a reliable indicator of sperm competition level (67)). The inferred sperm competition levels for M. musculus, M. spretus, and M. spicilegus are respectively low (testes mass ϭ 0.63% of body mass), intermediate (1.73%), and high (2.82%), in comparison with other species of the genus Mus (38). Although a higher number of species may be required for a correlational analysis between sperm competition and sperm metabolic features, the trend observed in this study is clear; the high levels of sperm competition promote an increase in the ATP production capabilities of mouse sperm, and the species with higher sperm competition level present higher OXPHOS activity. Altogether, this evidence would suggest that sperm competition could promote adaptations in sperm energetic metabolism tending to increase the usage of OXPHOS in relation to glycolysis, because it would constitute a more efficient pathway for the generation of high amounts of ATP.