Chemicals, unless otherwise indicated, were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, United States). Rotenone and antimycin A were from Cayman Chemical (Cat #13,995 and 19,433, respectively; Ann Arbor, MI, United States). The antibodies used were: anti-phospho-protein kinase A substrates (pPKAs) (Cat #9624s; Cell Signaling Technology, Danvers, MA, United States), anti-phosphotyrosine (pY) (clone 4G10; Millipore Corporation, Temecula, United States), anti-β-tubulin (Cat #T4026; Sigma-Aldrich), anti-rabbit and anti-mouse conjugated with horseradish peroxidase (HRP) (Sigma-Aldrich and Vector Laboratories, Inc., Burlingame, CA, United States, respectively), and anti-cAMP antibody (gently provided by Dr. Parlow, National Hormone and Peptide Program, Harbor-UCLA Medical Center, CA, United States). 20-O-monosuccinyladenosine-30,50-cyclic monophosphate tyrosylmethyl ester (TME-cAMP) was radiolabeled with Na¹²⁵I by the method of chloramine-T 90 (specific activity 600 Ci/mmol) as described (Piroli et al., 1992). Fluo-4 AM and pluronic acid were purchased from Invitrogen, Thermo Fisher Scientific (Waltham, MA, United States); while propidium iodide (PI) was from Santa Cruz Biotechnology (Dallas, TX, United States). Biggers-Whitten-Whittingham medium (BWW (World Health Organization, 2010)) was used throughout the study. It consisted of (in mM): 94.3 NaCl, 4.78 KCl, 1.2 MgSO₄, 1.7 CaCl₂, 1 KH₂PO₄, 5.5 glucose, 0.27 Na-pyruvate and 25 Na-lactate. This medium was called “NUTR”. Medium devoid of glucose, pyruvate and lactate was used for starving condition (“STRV”). For SER experiments, 2x NUTR medium was added to starved sperm, which contained the same components of NUTR but with double amount of the energy sources (11 mM glucose, 0.54 mM Na-pyruvate and 50 mM Na-lactate). One hour before use, media were supplemented with penicillin G, 25 mM NaHCO₃ and 0.5% bovine serum albumin (BSA; Cat #A7960; Sigma-Aldrich) and placed at 37°C in a 5% (v/v) CO₂ atmosphere. All media were used at pH = 7.4.

Semen samples used in the study were provided by normozoospermic volunteers, and obtained under donors’ written consent. Protocols were reviewed and approved by the Ethics Committee of the Instituto de Biología y Medicina Experimental, Buenos Aires (Ref: CE 001/April 2019).

Samples were obtained by masturbation after approximately 48 h of abstinence and subjected to routine analysis following WHO guidelines (World Health Organization, 2010). Only samples that fulfilled WHO normality criteria for semen volume, total sperm number, motility and vitality were included in the study. After liquefaction, samples were divided into two aliquots and motile sperm were selected by the swim-up procedure (World Health Organization, 2010) using NUTR or STRV medium (as stated in each case). Briefly, 0.5–1 ml of semen were placed in a 15-ml sterile conical tube and 1 ml of medium was gently layered over it. Tubes were incubated at 45° angle, at 37°C in a 5% (v/v) CO₂ atmosphere. After 1 h, the upper layers containing highly motile sperm were recovered and 5 ml of the corresponding medium were added. Samples were centrifuged at 350 × g for 7 min, the supernatants were discarded, fresh media were added and the centrifugation step was repeated in the respective media. The sperm pellets were resuspended in the corresponding medium, sperm concentration was determined with a Neubauer hemocytometer and adjusted to 7 × 10⁶ cells/mL.

When indicated, motile sperm were selected by density gradient centrifugation using Percoll. This product was chosen because it lacks glucose. Isotonic Percoll was prepared with 9 volumes of Percoll and 1 volume of 10x concentrated PBS, supplemented with 0.3% BSA. Eighty and 40% (v/v) Percoll solutions were prepared using NUTR or STRV medium as required. The discontinuous gradients were done in a 15-ml sterile conical tube by layering 1 ml of 40% Percoll over 1 ml of 80% Percoll. Aliquots of 0.5–1 ml of semen were gently placed on top of each gradient and they were centrifuged at 300 × g for 15 min. Sperm pellets were placed in a new tube containing 7 ml of the corresponding medium and samples were centrifuged at 350 × g for 7 min. The washing procedure was repeated using 2 ml of medium. Sperm pellets were resuspended in NUTR or STRV medium and sperm concentration was determined.

To promote capacitation sperm were incubated for the indicated periods of time in 150 µL of the corresponding medium at 37°C in a 5% (v/v) CO₂ atmosphere. For SER treatment, sperm were first incubated in STRV medium and at the time of recovery, 150 µL of 2x NUTR were added. As a control, sperm incubated in NUTR were diluted at the same time point with 150 µL of NUTR. Replicate experiments were run with different donors, unless indicated.

Sperm motility parameters were evaluated using the Sperm Class Analyzer® system (SCA v.6.2.0.1.; Microptic SL, Barcelona, Spain) that acquires 60 frames per second (s). Sperm suspensions (15 µL) were placed onto a slide, using a 22 × 22 mm² coverslide (preparation depth of 31 µm), and temperature was maintained at constant 37°C using a temperature-controlled stage. At least 5 microscopic fields and at least 300 sperm were analyzed. The following parameters were assessed: curvilinear velocity (VCL, µm/s), straight line velocity (VSL, µm/s), average path velocity (VAP, µm/s), linearity (LIN: VSL/VCL × 100, %), straightness (STR: VSL/VAP × 100, %), wobble (WOB: VAP/VCL × 100, a measure of sperm head side to side movement, %), amplitude of lateral head displacement (ALH, µm, which measures the magnitude of the lateral displacement of the sperm head about the average path) and beat cross frequency (BCF, Hz). Sperm motility was classified as follows: rapid progressive (VCL ≥ 35 μm/s; STR ≥ 80%), medium progressive (VCL ≥ 15 μm/s; STR ≥ 80%), in situ (VCL<15 μm/s; VAP ≥ 5 μm/s) and immotile (VAP < 5 μm/s). Percentages of total (rapid progressive + medium progressive + in situ) and progressive (rapid + medium progressive) motility were recorded. Drifting was set in 25 μm/s. Sperm were considered hyperactivated when presenting VCL ≥ 150 µm/s, LIN < 50% and ALH ≥ 3.5 µm [modified from (Mortimer et al., 1998)].

ATP levels were determined using a commercial kit (Cat #700410; Cayman Chemical, Ann Arbor, MI, United States). Motile sperm (4 × 10⁶) were incubated for the indicated periods of time under capacitating conditions. The medium was removed by centrifugation (400 × g for 5 min), and the sperm pellet was washed with PBS at room temperature. A second wash was performed with PBS at 4°C, followed by centrifugation at 700 × g for 4 min. Sperm were resuspended in 40 µL of ATP Detection Sample buffer (1x) at 4°C, homogenized by repeated pipetting and stored at −20°C until use. On the day of measurement, samples were thawed, diluted 1:50 with ATP Detection Sample buffer (1x) and maintained on ice. ATP detection standards were prepared and the assay was run following the manufacturer’s instructions. Luminiscence signal was recorded using a GloMax® 96 Microplate Luminometer (Promega; Madison, WI, United States). Results were expressed as pmol/10⁶ sperm and normalized against the control condition.

Sperm intracellular cAMP levels were determined by radioimmunoassay (RIA). Sperm were centrifuged at 700 × g for 4 min, the supernatant was discarded and 1 ml of PBS at 37°C was added. The centrifugation step was repeated and the sperm pellet was resuspended in 300 μL of absolute ethanol, followed by vortexing for 1 min. The sample was centrifuged at 15,000 × g for 4 min, the supernatant was recovered and the ethanol was subjected to evaporation by heating at 55–60°C for 4 h. An aliquot of 220 μL of 50 mM Na-acetate (pH 6.2) was added, it was vortexed for 1 min and the sample was stored at −20°C until cAMP determination as described (Del Punta et al., 1996; Abiuso et al., 2014). cAMP standards of 0–5,000 fmol in 100 μL of Na-acetate buffer were used, and duplicate determinations of the standards and samples (extracts corresponding to 10 × 10⁶ sperm) were done. A 10-μL aliquot of acetic anhydride and triethylamine (1:2 v/v) was added and incubated at room temperature for 10 min. Then, TME-cAMP labeled with I¹²⁵ (25,000 cpm) and the anti-cAMP antibody (1:17,500), both diluted in 100 μL of acetate buffer, were added. After overnight incubation at 4°C, the antigen-antibody complexes were precipitated with 50 μL of 2% BSA and 2 ml of cold ethanol (95%), and samples were centrifuged at 700 × g during 12 min. Supernatant was aspirated and the pellet radioactivity was determined using a Packard Auto-Gamma (Packard Instrument Co., Downers Grove, IL, United States). The inter and intra-assay coefficients of variability were 4.8 and 3.2, respectively. Results were expressed as fmol/10⁶ sperm and normalized against the control condition.

Sperm [Ca²⁺]_i levels were assessed by flow cytometry using Fluo-4 AM. After incubation in the appropriate medium, samples were centrifuged at 400 × g for 5 min at room temperature and resuspended in 250 μL of non-capacitating NUTR or STRV medium (devoid of NaHCO₃ and BSA), containing 1 μM Fluo-4 AM and 0.02% pluronic acid for 15 min at 37°C in the darkness. Samples were centrifuged again and resuspended in 500 μL of non-capacitating NUTR or STRV medium. Before collecting data, 2 μg/ml of PI was added to monitor viability. When indicated, 100x aliquots of energy sources were added to starved sperm and registered after 1 min. In this case, an equal amount of NUTR medium was added as control. Data were recorded as individual cellular events using a MACSQuant Analyzer 16 cytometer (Miltenyi Biotec Inc.). Forward-scatter area (FSC-A) and side-scatter area (SSC-A) data were collected from 20,000 events per sample in order to define sperm population as previously described (Escoffier et al., 2012). Doublet exclusion was performed analyzing two-dimensional dot plot FSC-A vs. forward-scatter height (FSC-H). Doublets exhibit a higher signal width or area to height ratio compared to single cells (singlets); events deviating from the diagonal are doublets. Positive cells for Fluo-4 AM and PI were collected using the filters for Fluorescein isothiocyanate (FITC; 530/30) and Peridinin chlorophyll protein complex (PerCP; 670LP), respectively. The two indicators had minimal emission overlap, but still compensation was done. Data were analyzed using FlowJo software (V10.0.7). Results were expressed as median fluorescence and normalized against the control condition.

Motile sperm were incubated in the corresponding medium for the indicated periods of time. Reactions were stopped by adding 500 µL of PBS and cells were centrifuged for 5 min at 600 × g. Sperm pellets were resuspended in Laemmli sample buffer without 2-mercaptoethanol, heated for 5 min at 100°C and centrifuged at 15,000 × g for 5 min. The supernatants were recovered and stored at −20°C until used. Samples were supplemented with 5% 2-mercaptoethanol, boiled for 5 min, and subjected to SDS-PAGE in 10% polyacrylamide gels and Western immunoblotting. To analyze pPKAs, membranes were developed with anti-pPKAs (1:3,000) and anti-rabbit IgG conjugated with HRP. To evaluate pY, primary antibody (1:5,000) and anti-mouse conjugated with HRP were used. As loading control, membranes were developed with anti-β-tubulin (1:5,000). The reactive bands were detected by enhanced chemiluminiscence (ECL) using standard procedures. The same membranes were developed with each of the 3 primary antibodies, followed by stripping with 0.2 N NaOH for 2 min. In some cases, pPKAs was first developed; in other cases, pY was detected in the first place. Western immunoblot images were analyzed with ImageJ 1.48 k (National Institute of Health, United States), following the specifications of ImageJ User Guide, IJ 1.46r. The optical densities of all bands, from 250 to 25 kDa, were quantified and expressed as relative to the β-tubulin band. Results from the NUTR condition were considered 100%.

Statistical analyses were done using the GraphPad Prism program (version 6.01 for Windows; GraphPad Software, San Diego, CA, United States). Data were expressed as mean ± standard error of the mean (SEM). To assume normal distribution, percentages were converted to ratios and subjected to the arcsine square root transformation. Results were compared by one or two-way analysis of variance (ANOVA) and multiple comparison tests, or Student’s t test, as indicated. A p value of < 0.05 was considered statistically significant.

To determine the effect of starvation on human sperm motility, we split each semen sample into two aliquots and performed the swim-up procedure using BWW lacking glucose, pyruvate and lactate (STRV) or medium with nutrients (NUTR), as control. After swim-up, motile sperm were washed twice and resuspended in STRV or NUTR media, respectively. Motility parameters were then evaluated by CASA either immediately after the swim-up (0 h) or after 3-h incubation (Figure 1A). At both time points, sperm incubated in NUTR medium showed high percentages of total and progressive motility (Figures 1B,C). However, contrary to our observations with mouse sperm (Navarrete et al., 2019), human sperm continued to be motile when incubated for 3 h in the STRV condition (Figure 1B). In the absence of exogenous substrates, the population of sperm showing progressive motility was significantly reduced (Figure 1C) and the average VCL was diminished compared to those of the NUTR medium (Figure 1D). In addition to VCL, sperm incubated in the STRV condition also displayed lower average values on VSL, VAP, WOB, ALH and BCF (Table 1). Moreover, in the absence of energy metabolites, sperm did not undergo hyperactivation (Figure 1E). Similar results were obtained when motile sperm were selected by density gradient centrifugation using Percoll (Supplementary Figure S1 and Supplementary Table S1).

Overall, these data indicate that, contrary to mouse sperm, human sperm remain motile in media devoid of energy nutrients suggesting that human sperm can obtain sufficient energy to continue moving, possibly from internal sources. To further evaluate this possibility, in another set of experiments, sperm were incubated for up to 48 h in STRV or NUTR media and motility parameters were evaluated at different time points (Figure 2A). As expected, in both conditions there was a decline in total motility with increasing incubation times (Figure 2B). However, even after 48-h incubation, a fraction of the sperm population continued to move despite the absence of exogenous metabolites. Other parameters such as the percentage of progressive motility also decreased with time (Figure 2C). Noteworthy, VCL was maintained in the STRV condition during the entire period of the experiment (Figure 2D) and no hyperactivation was observed in these cells at any of the time points analyzed (Figure 2E). Significantly reduced values were obtained for most of the kinematic parameters evaluated when sperm were incubated in STRV (Supplementary Figure S2).

To determine the contribution of OXPHOS in the maintenance of human sperm motility in the absence of energy nutrients, sperm incubated either in NUTR or STRV medium were exposed the last 15 min to carbonyl cyanide p-chlorophenylhydrazone (CCCP, a mitochondrial uncoupler), rotenone (ROT, an inhibitor of complex I of mitochondrial electron transport chain) or antimycin A (AA, an inhibitor of complex III of mitochondrial electron transport chain) and motility parameters were assessed (Figure 3A). In NUTR medium, no significant decline in any of the motility parameters was observed when sperm were exposed to either CCCP, ROT or AA (Figures 3B–D). However, cells incubated in starvation stopped moving when OXPHOS was inhibited (Figures 3B–D). These results indicate that in nutrients containing medium, glycolysis is sufficient to maintain sperm motility and hyperativation. However, in the absence of exogenous energy substrates, sperm motility is maintained exclusively by OXPHOS-mediated ATP production.

Based on the evidence in the murine and bovine models (Navarrete et al., 2019), we aimed to determine the effect of SER treatment on human sperm motility. After swim-up, sperm were incubated for different periods of time (1.5, 3 or 20 h) in STRV medium. Then, energy substrates were added, and motility was recorded at 15, 30, 45, 60 and 180 min (Figure 4A). Sperm incubated for 1.5 and 3 h under energy restriction were able to recover all motility parameters after addition of energy substrates, at least to similar levels than those consistently incubated in the presence of nutrients (Figures 4B,C). Interestingly, addition of exogenous nutrients to sperm starved for 20 h not only allowed the recovery of total and progressive motility and VCL, but also led to significantly higher hyperactivation percentages in comparison to the control (Figure 4D). The hyperactivation values obtained in SER treatment after 20-h incubation were similar to those of cells incubated for 3 h in NUTR medium (Figure 4C).

Sperm capacitation is associated with the acquisition of hyperactivated motility, which has high ATP demands (Edwards et al., 2007; Hereng et al., 2011). Moreover, during capacitation there is an activation of the cAMP/PKA pathway that leads to increased phosphorylation of PKA substrates (pPKAs) and tyrosine residues (pY) (Gervasi and Visconti, 2016; Puga Molina et al., 2018). First, we analyzed the effect of starvation on several capacitation-associated parameters (Figure 5A). As observed in previous experiments, sperm incubated for 4 h under starvation did not develop hyperactivated motility (Figure 5B and Supplementary Figure S3). These cells showed lower levels of ATP and cAMP in comparison to sperm incubated in the presence of energy substrates (Figures 5C,D). Similar to our findings with mouse sperm (Sánchez-Cárdenas et al., 2021), starvation increased [Ca²⁺]_i (Figure 5E). Regarding protein phosphorylation, while no significant differences were obtained for pPKAs, pY was significantly reduced in these cells (Figure 5F).

Next, starved sperm were subjected to recovery in NUTR medium for 1 min and sperm parameters were analyzed (Figure 6A). One-min incubation in the presence of exogenous nutrients led to the development of sperm hyperactivation (Figure 6B and Supplementary Figure S3). The 1-min recovery treatment also resulted in an increase in both ATP and cAMP concentrations (Figures 6C,D) and a decrease in [Ca²⁺]i (Figure 6E). Under these conditions, no improvement in pY levels were observed (Figure 6F).