The seminal acrosin‐inhibitor ClTI1/SPINK2 is a fertility‐associated marker in the chicken

Abstract The seminal plasma is a very complex fluid, which surrounds sperm in semen. It contains numerous proteins including proteases and protease inhibitors that regulate proteolytic processes associated with protein activation and degradation. We previously identified a seminal protein, chicken liver trypsin inhibitor 1 (ClTI‐1) over expressed in semen of roosters with high fertility, suggesting a role in male fertility. In the present study, we showed that ClTI‐1 gene is actually SPINK2. Using normal healthy adult roosters, we showed that SPINK2 amount in seminal plasma was positively correlated with male fertility in chicken lines with highly contrasted genetic backgrounds (broiler and layer lines). Using affinity chromatography combined to mass spectrometry analysis and kinetic assays, we demonstrated for the first time that two chicken acrosin isoforms (acrosin and acrosin‐like proteins) are the physiological serine protease targets of SPINK2 inhibitor. SPINK2 transcript was overexpressed all along the male tract, and the protein was present in the lumen as expected for secreted proteins. Altogether, these data emphasize the role of seminal SPINK2 Kazal‐type inhibitor as an important actor of fertility in birds through its inhibitory action on acrosin isoforms proteins.

Bird reproduction differs from mammals by many aspects. It combines a highly specialized internal fertilization system with specific storage of sperm in the female tract (sperm storage tubules [SST]), physiological polyspermy, specific relationships between sperm and oocyte perivitelline layer, oviparity with telolecithal egg but very rapid divisions of the zygote in the female tract before oviposition. All adaptations making easier the "adaptive freedom" of flying animals.
Moreover, bird sperm biology faces a spermatogenesis producing a high number of sperm (billions in the chicken) in a short time (12 days in the chicken), a lack of accessory glands in the male tract, and a post gonadic process in the female tract that is much longer (3 weeks in the chicken, 2-3 months in the turkey) where sperm evolve in highly different environments depending on the stage of albumen/shell secretions (reviewed by Blesbois, 2018). In birds, acrosin is assumed to be involved in the interaction of sperm with the inner perivitelline layer that surrounds the oocyte at the earliest step of the fertilization process (Etches, Clark, Toner, Liu, & Gibbins, 1996;Lemoine, Grasseau, Brillard, & Blesbois, 2008;Slowinska et al., 2010;Slowinska, Liszewska, Dietrich, & Ciereszko, 2012).
In semen, SPINK2 levels are reduced in specific pathological conditions (men azoospermia, specific reduced SPINK2 mutant mice (Kherraf et al., 2017;Lee et al., 2011;Rockett, Patrizio, Schmid, Hecht, & Dix, 2004). SPINK2 protein was recently identified as essential for acrosome biogenesis during mouse spermatogenesis (Kherraf et al., 2017). However, postgonadic SPINK2 biochemical and physiological functions in nonpathological conditions have been poorly explored and make part of the aim of the present study using the chicken model. Using an experimental line of chicken, we recently identified a seminal protein, chicken liver trypsin inhibitor 1 (ClTI-1) that was present with variable abundance in semen of healthy roosters with contrasted fertility (Labas et al., 2015). This protein was previously purified from chicken liver and seminal plasma using a chromatography based on its trypsin inhibitor property (Kubiak, Jakimowicz, & Polanowski, 2009;Lessley & Brown, 1978) but its function as well as its target proteases remain unknown.
In the present work, we combined several complementary approaches to better decipher the post gonadic biochemical and physiological roles of the seminal ClTI-1/SPINK2 protein in male fertility of physiologically normal, but with contrasted fertility, breeder males, and identified the physiological proteases that are actually inhibited by ClTI-1/SPINK2.

| Animals and sampling
Adult birds were housed at the INRA experimental unit UE-PEAT at Nouzilly (France). Their breeding followed the European welfare and the French Direction of Veterinary Services regulations (agreement number: C37-175-1). Thirty-week-old males were obtained from commercial pedigree stocks (Hubbard, Quintin, France; Novogen, Loue, France) and housed in individual battery cages under 14L/10D photoperiod and fed with a standard diet of 12.5 MJ/day. Females used for artificial insemination (AI) were 41-week-old ISABROWN hens (ISA, Ploufragan, France) housed in five hens rooms under a 14L/10D photoperiod and fed a standard diet of 12.5 MJ/day, supplemented with calcium. Semen was routinely collected twice a week by massage (Burrows & Quinn, 1937). Sperm concentrations were immediately determined by light absorption of semen with a photometer (Accucell Photometer, IMV Technologies, L'Aigle, France) at a wavelength of 530 nm (Brillard & McDaniel, 1985). Individual ejaculates were then diluted 1:1 in Beltsville poultry semen extender (BPSE; Brillard & McDaniel, 1985) and then used directly to analyze semen or inseminated females. Semen was centrifuged at 600g for 10 min at 20°C. The supernatant corresponding to seminal plasma was centrifuged twice again at 10,000g for 10 min at 4°C to remove insoluble cellular debris. Pure seminal plasma was stored at −20°C until further use.
Male reproductive tract tissues were obtained from 40-week-old animals. For reverse transcription polymerase chain reaction (RT-PCR), tissues were immediately frozen in liquid nitrogen, and for immunohistochemical study, the tissues were fixed in 4% paraformaldehyde (PAF) solution. THÉLIE ET AL.

| 763
2.2 | In vitro semen quality evaluation Mass motility was measured as a subjective evaluation of the speed of the movement of spermatozoa in 10 μl of semen (scale 0-8), as previously described . In this motility scale, the "0" value corresponds to a total lack of movement and 8 represents whirlwinds covering 30-60% of the area. A mean of five repetitions was made per male.
Measurements of motility were evaluated by the computerassisted sperm analysis system with an HTM-IVOS (Hamilton-Thorn Motility Analyzer; IVOS, IMV Technologies), as previously described (Nguyen et al., 2014).
Sperm viability was assessed with SYBR-14/propidium iodide fluorescent dyes (Chalah & Brillard, 1998). Sperm was diluted in Lake 7.1 buffer down to 20 × 10 6 cells/ml and 5 μl SYBR-14 was added before incubation for 10 min at 4°C in darkness. Afterward, 2 μl of propidium iodide was added and the incubation was further processed in the darkness for 5 min at 4°C. After incubation, sperm viability was assessed by flux cytometry measurements using a EasyCyte Guava system (Millipore, Molsheim, France). This test is the reference test to define fertility, and allow obtaining values to establish two cohorts of animals, namely fertile and subfertile animals. Animals were, therefore, considered fertile when rates were above 70% and subfertile when rates were below 70%. For DNA synthesis, 500 ng of total RNA were denatured in the presence of a mix of oligodT (25 ng) and random hexamers (12.5 ng)

| RNA isolation and PCR
for 5 min at 65°C. RT was performed at 42°C for 50 min using SuperScript II Reverse Transcriptase (Life Technologies) in the presence of dNTP (0.5 mM) and RNAsine (RNAse inhibitor, 2 U). All products were from Promega (Charbonnières-les-Bains, France).
The thermal cycling conditions were 95°C for 2 min, followed by 30 cycles of 95°C for 30 s, 60°C for 30 s and 72°C for 20 s and a final elongation at 72°C for 10 min. Samples were then submitted to agarose electrophoresis.

| In situ hybridization
Fragments of testis, epididymis, and ductus deferens were fixed in PAF 4%/phosphate-buffered saline for 24 hr at 4°C, dehydrated, embedded in paraffin and serially sectioned (10 μm). Sections were deparaffinized in xylene, rehydrated gradually and washed in water.
Sections were treated (deproteination, acetylation) and dehydrated again. Hybridization was performed at 65°C for 12 hr in the presence of 1 μg of antisense or sense probes. After posthybridization washes, digoxigenin (DIG) activity was detected using anti-DIG-AP antibody (12 hr at 4°C) and revealed with NBT-BCIP. SPINK2 PCR fragment was subcloned in the pCS2 + vector. Sense and antisense probes were generated using DIG RNA labeling kit (LifeScience, Saint-Quentin Fallavier, France).

| Immunohistochemical detection of SPINK2 in chicken tissues
Ten microns sections were prepared as described above. To achieve antigen staining, the slides were immersed in citrate buffer (10 mM, pH 6.0) and heated in a microwave oven (6 min, 440 W). Nonspecific binding was prevented with TBS-bovine serum albumin 0.1% buffer incubation before incubation overnight at 4°C in a humidified chamber in the presence of SPINK2 antibody (1:200). Subsequently, sections were incubated with anti-rabbit-AP antibody (1:200) for 1 hr. The staining was revealed using NBT/BCIP for one to 2 hr at 37°C. Negative controls included sections that were incubated without the primary antibody and with the secondary antibody only.

| Immunofluorescence detection of SPINK2 in chicken sperm
Fresh spermatozoa were fixed for 5 min 4% PAF solution and labeled as previously described (Nguyen et al., 2014) with anti-SPINK2 antibody (1:200) followed by anti-rabbit secondary antibody conjugated with Alexa Fluor® 584 (Thermo Fisher Scientific, Courtaboeuf, France). Negative controls included sperm that were incubated without the primary antibody and with the secondary antibody only.
Images obtained were analyzed and quantified using Fiji (Schindelin et al., 2012) to detect differences of labeling between roosters.
2.9 | Purification of SPINK2 from chicken seminal plasma and identification by MALDI-TOF mass spectrometry

| Purification and identification of serine proteases from sperm extracts
Pooled semen (7 ml) was centrifuged at 600g for 10 min at 20°C. The supernatant was discarded and sperm was centrifuged again at 12,000g for 10 min at 4°C. Sperm pellet was resuspended in 6 M urea, placed in an ice bath and sonicated for 15 s, set at 40% relative output and centrifuged at 14,000g for 15 min at 4°C.

| Statistical analysis
Data are represented as means ± SEM. Statistical analysis for multiple comparisons was performed using nonparametric Kruskal-Wallis test follow by Tukey's least significant difference (LSD).
Percentages were transformed to arcsine square-root before analysis. The data were analyzed using R (http://cran.r-project.org; R version 3.5.1). The level of significance was set at a p < 0.05.

| Characterization of SPINK2 sequence
In our previous proteomic investigation (Labas et al., 2015), ClTI-1 protein was identified as a 55 amino-acid proteoform with several posttranslational modifications including a proteolytic event leading to the excision of the N-terminal methionine, and three disulfide bridges.
Using a protein and genomic sequence alignment tool, we found that this proteoform matched with 100% of sequence coverage and an E-

| Expression profile of SPINK2 transcript and protein in chicken male
By RT-PCR using specific primers, we observed that the signal was much stronger in the testis and epididymis in contrast to the liver and heart where SPINK2 mRNA was barely detectable (Figure 1a). Using in situ hybridization, we detected SPINK2 transcripts in the whole epithelium of seminiferous tubules of testis, but neither in the interstitium, nor in the spermatozoa inside the lumen (Figure 2A-a,b).
SPINK2 mRNA was also detected in epididymal and ductus deferens epithelium (Figure 2A-c-f). No signal was detected in the lumen.
By WB, we detected an expected 5-10 kDa band in chicken seminal plasma, sperm extracts, and testis ( Figure 1b). In the testis and the epididymis, the protein was detected in the lumen with a strong signal ( Figure 2B), suggesting that the protein was secreted MALDI-TOF spectra confirmed that SPINK2 was present in the F2 fraction at m/z 3,013 and 6,025 ± 0.05% corresponding to di-and mono-charged forms, respectively.
The inhibitory activity of purified SPINK2 was first assessed against trypsin to verify that the purification process did not affect SPINK2 inhibitory activity. One hundred nanomolar of SPINK2 inhibited 63% (2 nM) of trypsin activity (data not shown). The inhibitory activity of SPINK2 was further assessed on sperm extracts.

| Physiological target proteases of SPINK2 protein
We first performed a benzamidine chromatography to isolate serine proteases composing the sperm extract (Lessley & Brown, 1978;Thurston, Korn, Froman, & Bodine, 1993). Results are displayed in It is noteworthy that zymography allows to detect the activity of both latent and active forms of proteases, as published previously (Toth, Sohail, & Fridman, 2012). The serine protease activity corresponding to a 25 kDa band, was completely inhibited in a dose-dependent manner by SPINK2 (Figure 6, lines 16 and 160, asterisk). The corresponding 25 kDa band was further analyzed by mass spectrometry for protein identification. Data showed that this 25 kDa band contains both acrosin and acrosin like proteins, as the unique serine proteases found in this sample (Figure 5a, asterisk). The analysis of mass spectrometry results also suggests that acrosin and acrosin-like are in their active form: no peptides matching with propeptide sequence could be identified on the 16 and 38 peptides identified for acrosin and acrosin-like, respectively.
To conclude, acrosin and acrosin-like proteins are (a) the only proteases isolated by the affinity chromatography with benzamidine, 3.5 | Correlations between seminal SPINK2 protein amount, amidase activity, and male fertility In our previous study (Labas et al., 2015), In parallel, we evaluated the acrosin amidolytic activity in seminal plasma of broiler and layer males (Figure 7a,b; acrosin amidolytic activity) and we compared with the SPINK2 protein amount (Figure 7a,b; SPINK2 WB quantification). We observed that the acrosin activity is lower in highly fertile males where the SPINK2 protein amount is high.
This negative correlation corroborates the above data and strengthens the conclusion that acrosin is a SPINK2 target in the seminal plasma.

| DISCUSSION
In the present study, we characterized ClTI-1, one of the seminal plasma proteins identified in our previous study as potential markers of chicken fertility (Labas et al., 2015) and investigated its postgonadic function for the first time.
Sequences of ClTI-1 were also referenced in protein databases of other avian species including turkeys, quails, doves, and ducks and of other non-bird species such as mouse, zebrafish, Xenopus, and Drosophila. Most of these protein sequences were predicted consequently to automatic chicken genome annotation in databanks.
Previous in vitro studies showed that Kazal-type protease inhibitors possess trypsin inhibition properties (Lee et al., 2011;Lin et al., 2008;Slowinska et al., 2014;Zalazar et al., 2012). In our study, we showed for the first time using sperm of physiologically normal and healthy animals, that SPINK2 displays trypsin and trypsin-like inhibition activity on sperm proteases. High inhibition efficiency of SPINK2 suggested that physiological function of SPINK2 is associated with semen protection against proteolytic activities of trypsinlike protease(s). Using benzamidine chromatography combined to amidase activity measurement, inhibition assay, and mass spectrometry analysis, we identified only two proteases, acrosin and acrosinlike protein.
Using this benzamidine affinity chromatography, we were able to purify 27 proteins from seminal plasma but only two are proteases (acrosin and acrosin-like;  Figure 3; the hydrolytic potential of serine proteases contained in this fraction [thus acrosin and acrosin-like] is inhibited in a dose-dependent manner by SPINK2). In parallel, the zymography performed with purified SPINK2, showed the disappearance of some bands, which after proteomic analyses F I G U R E 7 SPINK2 WB quantification, acrosin amidolytic activity and fertility rate in different chicken lines. (a,b) SPINK2 amount WB quantification, acrosin evaluation amidolytic activity, and fertility rate of good (gray histograms) and bad (black histograms) male breeders from broiler (a) and layer (b) lines. Values represent means ± SEM. Asterisks indicate significant differences between good and bad males (p < 0.05).
(c) Western blot profiles of SPINK2 in seminal plasma of broiler breeder males (meat line) with contrasted fertility (high = good; low = bad). The loading control (SyproRuby counterstained membrane) is presented in Figure S2.  (Baba, Michikawa, Kashiwabara, & Arai, 1989). Acrosin activity is known to be regulated by Kazal inhibitors (Laskowski & Kato, 1980;Slowinska et al., 2010). In turkeys, two forms of acrosin have been described, acrosin and acrosin-like Slowinska et al., 2010). In the present study, we identified for the first time two homologs of acrosin in the chicken species (38 and 39 kDa as proprotein forms of acrosin and 27 kDa as active protease). Both acrosin (GeneID 426864) and acrosin-like (GeneID 769176) proteins are localized within a 13 kb locus on chromosome 1 in G. gallus genome. We found the turkey acrosin-like specific peptide (SLQEYVEPYRVLQEAKVQ-LIDL) in the chicken acrosin sequence. Recently by coexpression of acrosin and SPINK2 proteins in human cell line HEK293, it was demonstrated that SPINK2 prevents cell proliferation arrest induced by proacrosin (Kherraf et al., 2017). We thus conclude that, in the chicken, like in the mouse, SPINK2 regulates the activity of acrosin to prevent uncontrolled/deleterious sperm proteolysis and too early acrosome reactions. However, it is necessary to highlight that acrosin from mammalian and avian species may have diverging functions since the sequence identity between these species varies from 36% to 43%, and since most avian acrosins lack the conventional prolinerich propeptide that is required for activation of mammalian acrosins ( Figure S4).
The chicken is known to be a very robust model to study candidate fertility marker proteins since the fertility evaluation is noninvasive and highly powerful in this species (availability of males with contrasted fertility) allowing us to complete the functional study by fertility tests (Soler et al., 2016). We showed positive correlations between fertility and SPINK2 amount in the seminal plasma of highly different genetic lines (meat and egg lines), confirming that SPINK2 function is important at the physiological level, regardless of the genetic background. We also revealed that the amidase activity (reflecting enzymatic activity of acrosin) is conversely correlated to the fertility rate and the SPINK2 amount. Our results demonstrated that SPINK2 amount in sperm should be a good candidate marker to predict male fertility.
Interestingly, in our previous proteomic study, acrosin, the main target of SPINK2 in sperm (as demonstrated in the present study), was also detected as differentially expressed in Label Rouge roosters' semen with contrasted fertility (overexpressed in males with low fertility rate; Labas et al., 2015).
In mouse, Spink2 genomic invalidation demonstrated the importance of the protein at the gonadic level. SPINK2 allowed spermatid differentiation and acrosome formation during spermatogenesis and was still present on the acrosome of mature sperm of mouse and human to protect it from proteolysis (Kherraf et al., 2017). In the present study, we demonstrated a role of SPINK2 at the postgonadic level. We showed that ClTI-1/SPINK2 is preferentially expressed in the male reproductive tract (at both the mRNA and protein levels; Figures 1 and 2) but not in the liver and the heart, and is recovered in the seminal plasma where it interacts with sperm proteins. Our results suggest that, in avian species, diffuse secretory cells of the testis, efferent ducts, epididymis, and deferent ducts secrete SPINK2 allowing its accumulation in the seminal plasma.
SPINK2 was previously suspected to be a secreted protein of mammals' male tract since it was found in mice testis, epididymal, and seminal vesicles fluids. Bird sperm develop adaptations to avoid too early acrosome reactions. Consequently, they need very high antiprotease protections before reaching the site of fertilization.
To conclude, the ClTI-1/SPINK2 protein, is a single domain Kazaltype serine protease inhibitor, that targets the chicken acrosin and F I G U R E 8 Correlation between SPINK2 protein amount and fertility rate. Positive correlations for males from three genetic lines (■ broilers and • layers) between fertility rate and SPINK2 relative abundance acrosin-like proteins in seminal plasma. The role of seminal SPINK2 acrosin inhibitor in fertility is likely to prevent too early acrosin activation that would be deleterious for sperm cells. Seminal SPINK2 is present in higher amount in highly fertile males in contrast to acrosin that is over-abundant in low fertile males. From these results, we strongly suggest that high seminal SPINK2 induce efficient inhibition of acrosin activity which is important for fertility. Thus, SPINK2 seems to be a relevant candidate fertility marker in chicken species. Whether SPINK2 has the same function and is a fertility marker also in other species is a crucial question that should be addressed in the future.

ACKNOWLEDGMENTS
We thank the staff of the experimental unit PEAT for animal breeding and semen collection, Isabelle Grasseau for quality semen evaluation and Maryse Mills for zymography experiment help. We thank Valérie Labas, Lucie Combes-Soia, and the PAIB platform for mass spectrometry analysis. We thank the SYSAAF for its kind support. This study was supported by grants from the French