Evidence of endogenously produced hydrogen sulfide (H2S) and persulfidation in male reproduction

Persulfidation contributes to a group of redox post-translational modifications (PTMs), which arise exclusively on the sulfhydryl group of cysteine as a result of hydrogen sulfide (H2S) action. Redox-active molecules, including H2S, contribute to sperm development; therefore, redox PTMs represent an extremely important signalling pathway in sperm life. In this path, persulfidation prevents protein damage caused by irreversible cysteine hyperoxidation and thus maintains this signalling pathway. In our study, we detected both H2S and its production by all H2S-releasing enzymes (cystathionine γ-lyase (CTH), cystathionine β-synthase (CBS), and 3-mercaptopyruvate sulfurtransferase (MPST)) in male reproduction, including spermatozoa. We provided evidence that sperm H2S leads to persulfidation of proteins, such as glyceraldehyde-3-phosphate dehydrogenase, tubulin, and anchor protein A-kinase. Overall, this study suggests that persulfidation, as a part of the redox signalling pathway, is tightly regulated by enzymatic H2S production and is required for sperm viability.

www.nature.com/scientificreports/ role in H 2 S metabolism. Although H 2 S-releasing enzymes were detected in mouse testes 11 and CBS and CTH in human spermatozoa 12 , complete knowledge about their distribution through mammalian spermatozoa is lacking. Taken together, the results of previous studies indicate that endogenous production of H 2 S is associated with antioxidant defence and antiapoptotic and antiaging events reported in many tissues 13 . Unfortunately, studies on testicular tissue and spermatozoa are limited by the artificial supply of H 2 S rather than real H 2 S production 11,12,14,15 . Therefore, recent findings are unclear, and there are different conclusions depending on the donor concentration used. Although endogenous H 2 S production has been overlooked in male reproduction, these studies suggest that it has potential in reproductive physiology and deserves further attention. Our study provides the first evidence of physiological H 2 S production in spermatozoa and physiological contributions in the form of persulfidation in sperm physiology.

Results
Redox PTMs of cysteine do not drive maturation of male reproduction. In this experiment, we focused on persulfidation in the broad context of other redox PTMs (-SNO and -SOH), in which cross regulations with persulfidation have been reported 4,5 . Due to the physiological contribution of cysteine PTMs in sperm maturation, we assumed that redox PTMs control the onset of spermatogenesis and thus drive sexual maturity in males. Therefore, we compared testicular lysates from mouse males before puberty onset, 21-day-old (young) and fully matured, 12-14-week-old males (adult) by Western blot (WB) detection of cysteine modifications: -SNO, -SOH, and -(S) n H (Fig. 1). We did not find any differences between the young and adult groups in (i) protein distribution (Fig. 1a-d), (ii) individual band intensity ( Fig. 1e-g) or (iii) the total protein intensity ( Fig. 1h-j) in any of the following parameters, suggesting that redox PTMs do not drive male reproductive maturation. The WBs of each PTM showed a specific ladder of bands ( Fig. 1a-d). There was no band detected concurrently for -SNO/-SOH and -(S) n H, which suggests that there are different abundant proteins undergoing Persulfidation is abundant in testes compared to other tissues. Persulfidation (-(S) n H) plays a plausible role in ageing, apoptosis and stress defence in many tissues, but information about this PTM in male reproduction is still missing. To shed light on this issue, we performed quantitative and qualitative analyses of -(S) n H in the testis. Because no specific maturity-dependent protein pattern was observed, the index of -(S) n H was compared in different kinds of tissues. We selectively labelled -(S) n H, accordingly with 16 with slight modifications, while free -SH groups were blocked by MMTS and subsequently the -(S) n H groups were alkylated by IAM-PEG-biotin (Fig. 2a). Biotin was detected by streptavidin conjugated with horseradish peroxidase via WB detection (Fig. 2b). We observed that -(S) n H ranged from 40 to 150 kDa in the testes of adult mice. To validate the specificity of the method used, we prepared three specifically treated groups to detect: (S) n H + SH (no MMTS treatment), -(S) n H only (MMTS-treated), and naturally biotinylated proteins (nonalkylated control). The detected persulfidated proteins were then identified using pulldown assays and nano-LC-MS (Fig. 2c,d). We compared persulfidated proteins from the testis with those of the brain and liver, in which -(S) n H was previously widely described (Fig. 2c). We found proteins that were conservatively persulfidated across the tissues, but we found 68 proteins that were persulfidated only in the testis. Figure 2d represents persulfidated proteins specifically found in testes in the size range 55-75 kDa (bands in white rectangle marked with * in Fig. 2b). These findings suggest that -(S) n H targets proteins specifically in the testis, although these are widely expressed proteins.
H2S-releasing enzymes are present in germ cells independent of their maturation stages in mouse testes. We found that persulfidation (-(S) n H) is relatively abundant in the testes compared to the frequently studied liver and brain. -(S) n H is a well-known result of H 2 S action, and it is released enzymatically inside the cell. Therefore, we consider monitoring H 2 S production to be essential (Fig. 3). CBS, CTH and MPST have been previously detected in mouse testes 11 , but their distribution across developmental stages of germ cells has not been determined. Therefore, immunofluorescence detection of CBS, CTH, and MPST in testis sections was performed, depending on the developmental stages of germ cells within the seminiferous epithelial cycle. The results showed a strong dependency of enzymes on the cytoplasm of developing germ cells, regardless of the cell type and phase of spermiogenesis, distinguished in the Golgi, cap, and acrosomal stage (Fig. 3a). Enzyme independence of germ cell maturation was confirmed by WB performed on prepubertal and adult mice (Fig. 3b,c). The observation that H 2 S is not apparently associated with maturity level supports the versatility of H 2 S action in a cell. To elucidate H 2 S enzymatic production in testicular tissue, we performed colorimetric H 2 S detection (Fig. 3d). After the addition of pyridoxal-5′-phosphate (PxP), a cofactor of CBS and CTH, and l-cysteine, the substrate of enzymes, into the testis lysate, the production of H 2 S increased. To the best of our knowledge, we are the first to describe the relationship among H 2 S appearance, the enzymes responsible for its production, and Enzymatic production of H 2 S leads to persulfidation of protein cysteine in mouse spermatozoa. The aim of this experiment was to examine H 2 S-releasing enzymes in mouse spermatozoa during their passage through the caput into the cauda epididymis. For analysis of H 2 S production, we compared the pattern of H 2 S-releasing enzyme subcellular distribution with H 2 S fluorescence visualization and protein persulfidation (-(S) n H) in fully differentiated spermatozoa. First, we detected CBS, CTH, and MPST via WBs in mouse spermatozoa from the caput epididymis (Sp caput ) and cauda epididymis (Sp cauda ) (Fig. 4a). Caput spermatozoa showed a strong signal of all H 2 S-releasing enzymes, whereas caudal spermatozoa showed either decreased (CBS), weak (MPST) or almost no signal (CTH) detected by WBs. To enhance the observation of H 2 S-releasing enzymes in caudal spermatozoa, we performed immunocytochemistry of single sperm cells ( Fig. 4b-d). The signal of all enzymes along the entire length of the flagella was observed in caudal spermatozoa. Their enzymatic action was proved by H 2 S labelling by specific Sulfane Sulfur Probe 4 (SSP4) (Fig. 4e). Similar to H 2 S-releasing enzymes, the signal was emitted in the entire length of the flagella with the highest intensity in the midpiece. The observation of H 2 S production corresponding to H 2 S-releasing enzyme locations strongly supports the occurrence of H 2 S enzymatic production. Finally, we detected -(S) n H and found it exclusively in the midpiece (Fig. 4f). Although -(S) n H showed a slightly different pattern than H 2 S-releasing enzymes, it perfectly followed the site of the highest occurrence of H 2 S production. Therefore, we identified the midpiece as the location of H 2 S enzymatic activity, H 2 S production, and biochemical action.

Sperm H 2 S-releasing enzymes produce H 2 S in spermatozoa across mammalian species.
Based on previous findings of H 2 S production and the expression of H 2 S-releasing enzymes in mouse testes and spermatozoa, we suggest that H 2 S is enzymatically produced in spermatozoa across mammalian species. Therefore, we detected CBS, CTH, and MPST in human spermatozoa and in mouse and pig spermatozoa, the most common mammalian models. Enzyme detection was followed by elucidation of the release of SSP4-labelled H 2 S. Based on the known species-specific molecular weight of the individual enzymes based on the UniPr otKB Database (Fig. 5a, see the phylogenetic trees in Supplementary Fig. S1), we identified all enzymes via WBs (  www.nature.com/scientificreports/ enzymes responsible for most H 2 S production in a species-dependent manner. While H 2 S production colocalized with CTH in human spermatozoa, in boar spermatozoa, it colocalized instead with MPST. Sequential disappearance of H 2 S-releasing enzymes through spermatozoa maturation was shown by the evoked capacitation ( Fig. 5e) and zona pellucida-binding assays of boar spermatozoa undergoing the acrosomal reaction, the last step of sperm maturation (Fig. 5f). This result complements our previous finding that H 2 S-releasing enzymes gradually decrease from spermatozoa during their maturation in the epididymis (Fig. 4a). Based on these observations, we conclude that the presence of H 2 S-releasing enzymes is partially lost from the cytoplasmic membrane during remodelling, which accompanies sperm maturation. Therefore, these enzymes do not appear to be involved in the sperm fertilization of eggs.
Distribution and identification of persulfidates in human spermatozoa. In accordance with the aforementioned presence of H 2 S-releasing enzymes and H 2 S, we investigated the effects of H 2 S on persulfidation (-(S) n H) in the sperm protein of three normozoospermic donors using the approach described above. Concurrently, all three samples were subjected to flow cytometry analysis to obtain an overview of persulfidation occurrence across the entire sperm population. We observed -SH and -(S) n H with regard to plasma membrane integrity (PMI) using flow cytometry (Fig. 6a). In the control groups, spermatozoa were separated into three subpopulations according to susceptibility to PMI and 6-IAF staining as follows: the 1st quadrant (Q1) was live spermatozoa highly positive for 6-IAF, the 2nd quadrant (Q2) was dead spermatozoa highly positive for 6-IAF and the 4th quadrant (Q4) was live spermatozoa slightly positive for 6-IAF. When free thiols (i.e., -SH) were specifically blocked in the MMTS experimental group, most 6-IAF signals disappeared, and spermatozoa moved to the less intense 6-IAF quadrants Q3 (death) and Q4 (live). This observation was supported by the detection of -SH and -(S) n H in situ, occurring in whole sperm or exclusively in the midpiece in the control (Fig. 6b). Following free thiol blocking, the sperm head-rich signal disappeared. For both flow cytometry and in situ detection, cell death was noticeably accompanied by substantial thiol exposure through the spermatozoon, while persulfidation was stably localized in the midpiece independent of live or dead sperm status. This observation was supported by the analysis of three independent semen donors, while the analysis was performed on the subpopulation of live spermatozoa (i.e., Q1 and Q4 quartiles). There was a small population of spermatozoa (4.6-5.1%) showing high signal intensity belonging to -SH and -(S) n H in the control group. When -SH was blocked, only -(S) n H remained, and significantly weaker signals were detected in the sperm population in the MMTS group (Fig. 6c). Using the biotin-switched detection of -(S) n H in sperm lysate, we found that the abundance of persulfidated proteins did not show any capacitation-dependent difference (Fig. 6d), similar to our   Fig. S3). Concurrently, these persulfide-labelled samples were subjected to pulldown assays, followed by nano-LC-MS peptide detection. We identified 37 persulfidated proteins with 99% confidence, in most cases being in a donor-specific pattern (Fig. 6f). Five proteins were found to match at least in two donors, marked in bold in the table containing all characterized persulfidated sperm proteins (Fig. 6e). Altogether, nano-LC-MS findings of midpiece-occurring proteins are in accordance with the H 2 S-releasing enzyme distribution, H 2 S labelling and in situ detection of persulfidation, underlining the spatiotemporal requirement of H 2 S activity in target protein modulation.

Discussion
The sulfhydryl group (-SH) of cysteine provides a unique signalling pathway in which many redox molecules are involved. These molecules, such as nitric oxide (NO), hydrogen peroxide (H 2 O 2 ) or hydrogen sulfide (H 2 S), oxidize or reduce -SH to form various redox post-translational modifications (PTMs) that control protein activity. These redox PTMs are unstable and continuously replacing each other on cysteine and together creating sophisticated signalling pathways. Persulfidation (-(S) n H) plays a central role in this pathway, as it replaces S-nitrosylation (-SNO) and S-sulfenylation (-SOH) on cysteine and thus regulates not only protein activities but also prevents their inactivation by hyperoxidation. There are many proteins that are regulated by redox PTMs in male reproduction 2,10 but -(S) n H has not been investigated. In this study, we identified the existence of -(S) n H and its relation to endogenous H 2 S production and H 2 S-releasing enzyme localization in male reproduction with a focus on spermatozoa.
To the best of our knowledge, our study is the first to identify the protein -(S) n H in mouse testes and human spermatozoa. Subsequently, we compared persulfidated (-(S) n H) proteins found in mouse testes with -(S) n H from previously widely studied tissues, the brain and liver (Fig. 2c). Surprisingly, the largest amount of -(S) n H www.nature.com/scientificreports/ was detected in the testis, and 68 of the proteins were persulfidated uniquely in the testis (Fig. 2d). Although these proteins are not exclusive to the testes, they are apparently persulfidated in the testis only. It is well known that -(S) n H is the result of H 2 S action, and partially due to this effect, artificial supplementation of H 2 S has antiaging and antioxidant effects in many tissues, including spermatozoa 11,12,14 . There are several publications addressing H 2 S donor efficacy, although supplementation with exogenous H 2 S has possible toxic effects 15 . In contrast to these publications, there is no evidence about physiological endogenous H 2 S production and its consequences for male reproduction. Although all H 2 S-releasing enzymes have been previously detected in mouse testis 11 and CBS and CTH in human spermatozoa 12 and rat epididymis 17 , we immunodetected all three responsible enzymes in mouse testicular cross-sections and in spermatozoa of three mammalian species. Therefore, we claimed that enzymes are distributed in the cytoplasm regardless of the cell type or germ cell maturation stage. Although there was a steady distribution of enzymes across the seminiferous epithelium cycle, our experiments showed that spermatozoa from caput showed a higher intensity of H 2 S-releasing enzymes than more mature caudal spermatozoa (Fig. 4a). Spermatozoa obviously lose their H 2 S-releasing enzymes during passage through the epididymis. An interesting finding was made by a study describing the importance of H 2 S production for sperm quiescence in the rat epididymis 17 . In contrast to that in spermatozoa, the expression of CBS and CTH was increased towards the cauda epididymis 17 . Based on recent knowledge, the epididymal epithelium appears to compensate for sperm H 2 S-releasing enzyme loss during spermatozoa passage through the epididymis by increasing the self-production of enzymes. Nevertheless, the loss of H 2 S-releasing enzymes continues beyond the epididymis and is found in further steps of sperm maturation, capacitation and sperm-zona pellucida binding (Fig. 5d,e,f). The highlights of our study were the detection of enzymatic production of H 2 S and its consequences in the form of -(S) n H in mammalian spermatozoa. We detected H 2 S in the sperm flagella of mice (Fig. 4e), humans (Fig. 5c), and boar (Fig. 5d). Moreover, we immunodetected all H 2 S-releasing enzymes in all models used in sperm flagellum; therefore, we indirectly related endogenous H 2 S production to its enzymes. Subsequently, we followed -(S) n H as a major result of H 2 S action. To the best of our knowledge, for the first time, we described and characterized proteins that underwent -(S) n H of cysteine in spermatozoa. Protein persulfidation (-(S) n H) was strictly located in the sperm midpiece (Fig. 4f), which highly corresponds to H 2 S occurrence and the location of its enzymes. These observations are consistent with H 2 S properties; although H 2 S diffuses well across membranes, its short half-life, which lasts a few seconds, a maximum of minutes 18 , does not allow it to sufficiently affect proteins over long distances. To further elucidate the role of -(S) n H in sperm physiology, we detected -(S) n H depending on the live/dead status of human spermatozoa (Fig. 6). Surprisingly, live and dead spermatozoa did not differ from each other in terms of -(S) n H (Fig. 6a,b). Because cell death is accompanied by a decrease in pH, H 2 S could be released from pH unstable iron-sulfur complexes located in mitochondria, thereby maintaining the -(S) n H of nearby proteins even after cell death. To label -SH and -(S) n H, we used the affinity of iodoacetamidofluorescein (IAF) to the -SH group. IAF was previously used in a study that addresses the quality of cryopreserved bull spermatozoa depending on -SH content 19 . Based on IAF staining, the researchers distinguished several patterns, whose distribution was dependent on sperm viability. Spermatozoa labelled strictly in the midpiece were associated with higher viability than spermatozoa labelled along its entire length, as was the case in our study. Interestingly, the pattern associated with viable spermatozoa is strikingly similar to the pattern of -(S) n H. It is possible that viable spermatozoa contain -(S) n H in their midpiece instead of free -SH, as was previously suggested 19 . If so, -(S) n H located specifically in the mitochondrial sheath may play an important role in sperm metabolism and redox defence. We supported this statement by identifying -(S) n H using mass spectrometry. In most cases, the identified proteins were associated with mitochondrial metabolism and flagellar movement (Fig. 6e). Some of these proteins have been reported to undergo -(S) n H, including glyceraldehyde-3-phosphate dehydrogenase, tubulin 16 and l-lactate dehydrogenase 4 , but we were the first to observe that these proteins were persulfidated in human spermatozoa. Some persulfidation targets were previously discovered as S-nitrosylated, including A-kinase anchor protein, heat shock protein and semenogelin 10 , which supports the finding that S-nitrosylation serves as a -(S) n H precursor 4,5,20 . All these results prove that spermatozoa contain many proteins containing reactive cysteine through which proteins can be easily turned on and off by redox PTMs.
Spermatozoa are completely dependent on previously produced proteins once they leave the male reproductive tract; therefore, they are vulnerable to oxidative stress. Persulfidation could play a key role in this context because it prevents cysteine hyperoxidation and thus stops redox signalling pathway disruption and protein damage. The reducing abilities of H 2 S could be essential during sperm capacitation. As is known, capacitation is enhanced by reactive oxygen and nitrogen species, but overproduction of these reactive species leads to oxidative stress and cell death 1,21,22 . However, H 2 S, through persulfidation, could contribute to the maintenance of redox balance, and thus prevent premature capacitation, which is an often problem of sperm manipulation in vitro conditions (e.g. cell maintenance in vitro, cryopreservation). In this study, we provided evidence for the enzymatic production of H 2 S not only in the testis but also in spermatozoa. We detected CBS, CTH and MPST in mammalian spermatozoa and thus indirectly linked H 2 S with its enzymes. We visualized H 2 S and therefore were able to localize it to the sperm flagella, where it affects nearby proteins by persulfidation. We identified some persulfidated proteins seemingly crucial for sperm viability, and we outlined the impact of endogenous H 2 S production on male reproduction. We proved the existence of H 2 S-releasing enzymes, H 2 S, and persulfidation and considered the link between them in spermatozoa. Obviously, other sophisticated models of in vitro pharmacological treatment of sperm and/or targeted silencing of all H 2 S-releasing enzymes in somatic cells are needed for the achievement of experimental data leading to a comprehensive acknowledgement of H 2 S in the physiology of reproduction. Moreover, there is no doubt that H 2 S is an important signalling molecule that purposeful modulation deserves a knowledge transfer to different medical disciplines.

Methods
All chemicals were purchased from Sigma Aldrich unless specified otherwise. Peanut agglutinin from Arachis hypogaea (PNA) conjugated with Alexa Fluor™ 488 was purchased from Thermo Fischer Scientific (MA, USA, #L21409). The primary polyclonal antibodies anti-cystathionine β-synthase (anti-CBS), anti-cystathionine γ-lyase (anti-CSE) and anti-3-mercaptopyruvate sulfurtransferase (anti-3MPST) as well as the secondary antibody goat anti-rabbit-Alexa Fluor ® 647 (# ab150079) were purchased from Abcam (Cambridge, UK). Briefly, ejaculates were divided into the noncapacitated and capacitated groups. Spermatozoa were allowed to swim up from ejaculates into HTF-HEPES medium, which was placed over the ejaculate, for 2.5 h in a 37 °C water bath. In the case of the capacitated group, HTF-HEPES medium was enriched with 0.3% bovine serum albumin (BSA). Thereafter, all samples were processed according to the purpose stated below.

Animals, samples, and ethical statements.
Porcine zona pellucida-binding assay. Pig oocytes were obtained from ovaries of 6-to 8-month-old noncycling gilts (a crossbreed of Landrace × Large White), yielded at the slaughterhouse (Jatky Český Brod a.s., Český Brod, Czech Republic). First, cumulus-oocyte complexes were collected from ovarian follicles with a diameter of 2-5 mm by aspiration with a 20-gauge needle and handled in TL-HEPES-medium supplemented with 0.1 mg/ml polyvinyl alcohol (PVA). Immature oocytes were matured in vitro in modified tissue culture medium (mTCM; Gibco, Life Technologies, UK), as described earlier 26 . After 44 h of culture, cumulus cells were removed with 0.1% hyaluronidase, and matured oocytes with extruded polar bodies were selected for the binding assay. Spermatozoa stored in BTS medium were washed and resuspended in modified Tris-buffered medium (mTBM; Abeydeera et al., 1998) at a concentration of 1 mil/ml. Subsequently, 100,000 spermatozoa were added to oocyte-free zonas and coincubated in 0.5 ml of mTBM at 39 °C and CO 2 for 30 min. Thereafter, zona pellucida-bound spermatozoa were washed in PBS supplemented with PVA, fixed in 4% paraformaldehyde (PFA) enriched with 0.1% Triton TX-100 and 1 mM DTT for 15 min at 37 °C, washed and stored in PBS with sodium azide at 4 °C for immunocytochemistry.
Immunocytochemistry. Mouse, boar, and human spermatozoa were fixed and stored as described above.