Buffalo Sperm Surface Proteome Pro ling Reveals an Intricate Relationship Between Innate Immunity and Reproduction

Vipul Batra National Dairy Research Institute Vanya Bhushan National Dairy Research Institute Syed Ali National Dairy Research Institute Parul Sarwalia National Dairy Research Institute Subhash Solanki National Dairy Research Institute Arumugam Kumaresan National Dairy Research Institute Rakesh Kumar National Dairy Research Institute Tirtha Datta (  tirthadatta@gmail.com ) National Dairy Research Institute


Introduction
The voyage of the spermatozoa in the female reproductive tract (FRT) entails surmounting of numerous impediments including the physical, thermal, chemical and immunological barriers. These include the vaginal acidic pH, the mucus in the cervix, the leukocytes and anti-sperm antibodies of the immune system especially in the uterus, and the narrow utero-tubal junction in the oviduct (1, 2; 3, 4 and 5). To overcome these obstructions the spermatozoa must acquire surface properties primarily customized for this strenuous voyage. The process of sperm surface remodelling (SSR), which occurs during the epididymal transit of the spermatozoa tailors the sperm surface which assists them in survival and fertilization in the FRT (6, 7, and 8). The biomolecular constitution of the mammalian testicular spermatozoa changes continuously and progressively in the luminal uid of the various epididymal regions due to the secretory and re-absorptive actions of the epithelial cells that line this organ (9, 10, 11 and 12). The remodelling events include a) enzymatic cleavage of the membraneassociated proteins b) variations in the composition of membrane-lipids c) re-organization of the glycoconjugates (GCs) associated with the sperm glycocalyx d) removal or addition of (glyco)proteins (9, 13, and 14). A blend of distinct secretagogues is known to be added onto the sperm-surface in these three epididymal regions viz. caput, corpus and cauda. A majority of these secretagogues include the immune-related (glyco)proteins often implicated in sperm survival and fertility. The epididymal secreted proteins involved in sperm maturation could be adhered on their sperm plasma membrane either by low binding a nity or they could be transmembrane. Many of these components bind transiently, for example, molecules acquired in the distal epididymal regions which are obligatory for traversing the array of mucosal uids and extracellular matrices in the FRT (15,16). The adhered proteins in the peripheral sperm environment bind the sperm-surface most likely by the electrostatic/hydrophobic interactions. These proteins change the sperm-surface characteristics as they interact with the transiting spermatozoa. The epididymal secretome usually involves the binding of hydrophobic proteins, the glycans-modifying enzymes such as glycosidases and glycosyltransferases (17, 18 and 19), proteases and protease inhibitors (20), proteins involved in immunological protection (3) and the ones that protect the sperm from oxidative injuries (21). Besides, many membranous vesicles rich in cholesterol, sphingolipids and Ca + 2 known as epididymosomes exist in the epididymal lumen. These extracellular vesicles mainly carry the GPI linked proteins, many of which are inserted in the sperm plasma membrane (22, 23 and 24).
As mentioned earlier, many of the added (glyco) proteins on sperm-surface belong to defence family and their glycosylation patterns are critical for either stabilizing the sperm-membrane during the immune attack by immune cells or assisting in immune-evasion in the FRT (25, 26, 27 and 28). The rendering of a highly glycosylated surface coat on the spermatozoa after epididymal transit, not only acts as a barrier between the spermatozoa and female immune system, but also assists in them in cervical mucus penetration (CMP), oviductal epithelial cell (OEC) binding, identi cation of the zona pellucida and oolemma, and juxtaposition of the sperm and the oocyte plasma membrane (3, 19, 29, 30 and 31). The sperm-surface proteins and their associated glycans also play a key role in the acquisition of motility and fertilizing ability in the epididymis, their protection, selection and secondary maturation in the FRT (19 and 32).
The buffalo was considered as a model for this study due to its economic importance in the agriculture-based economies. It has been reported that more people depend on buffalo than on any other domestic animal (33).
Although it is a premier dairy animal with superior milk-producing ability, idiopathic male infertility is a common reproductive limitation in buffalo. A sizeable number of high genetic merit bull calves originally selected for AI programs are discarded because their semen ends up yielding dismal conception rates (CRs) between 30 to 50%, re ecting poor fertilizing ability (34, 35 and 36). The factors that contribute to male fertility are relatively poorly understood, especially in bovine species (37). The prediction of fertility assessment currently relies on analyses of sperm functional parameters apart from the physical examination of the bulls, nonetheless, the correlation between these parameters and CR are often inconsistent (38). Therefore, a better understanding of the novel factors which regulate fertility e.g. sperm-surface proteins is required to gain insights into the factors behind idiopathic male infertility.
The objective of this study was the identi cation and in silico characterization of the post-testicular maturation antigens and peripheral proteins that interact with the buffalo sperm plasma membrane either through noncovalent (ionic) interactions or through a GPI-anchor. We also sought to determine the existence of the epididymal expressed BDs on buffalo spermatozoa and to predict their reproductive functional signi cance through immuno uorescence and in vitro fertilization experiments.

Results
Only the ejaculates that were milky or creamy in colour, homogenous in consistency i.e. free from akes/clumps with a minimum sperm concentration of 600 x 10 6 /ml were considered for swim-up and further downstream experiments. The average motility and viability of seven representative sample ejaculates after processing was 81.84 ± 1.20 and 85.85 ± 1.16 respectively. The samples were diluted according to the experiments, as mentioned, wherever required.
Hundreds of proteins are bound on buffalo sperm surface either through electrostatic interactions or by a GPI anchor The extracted sperm-surface proteins indicated enough diversity among the types of protein removed using the seven treatment groups viz. 2X-30, 2X-60, 4X-30, 4X-60, and 1U/mL, 1.5U/mL and 2U/mL representing the elevated salt extractions (2X/4X-DPBS for 30/60 min) and PI-PLC extractions, respectively (Supplementary Figs. 1 and 2). All the elevated salt and PI-PLC treatments' extracted sperm-surface proteins produced > 20000 PEP-XML spectra. The iprophet tool correctly identi ed more than 300 proteins in all of the treatments at p > = 0.99 where p indicates the probability that the spectra have been correctly matched to its analogous peptide (Table 1). A total of 317, 391, 395 and 432 proteins were identi ed in 2X-30, 4X-30, 2X-60 and 4X-60 (DPBS) treatments respectively. On the other hand, 385, 353 and 364 proteins were identi ed in the 1U, 1.5U and 2U/mL PI-PLC treatments, respectively. At p ≥ 0.99 zero proteins were found to be incorrectly identi ed. Many proteins were found to be unique to each sub-group of either treatment (elevated salt or PI-PLC) demonstrating that the individual combinations of incubation time and salt/enzyme concentration exerted disparate effects on disrupting the non-covalent/GPI interactions among the proteins of sperm surface. Moreover, nearly 30% of the proteins were observed as common between any two treatment subgroups (Supplementary Figs. 1 and 2). Overall, we report a total of 352 buffalo sperm-surface proteins that were identi ed in the protein fractions The BDs Spag-11D and BD-129 had the highest pI while the Acrosin inhibitor 1 had the lowest pI (4.25). Only three (5%) proteins viz. Sperm acrosome membrane-associated protein1, Angiotensin-converting enzyme and an uncharacterized protein (F1MD73) were predicted to contain a transmembrane segment. A high level of PTMs, especially glycosylation appears to modify the analyzed proteins. More than 80% of the analyzed proteins were Proteins involved in the immune response and reproductive processes adorn the buffalo sperm surface GO analysis was performed on the identi ed 119 extracellular (EC) sperm-surface proteins ( Supplementary   Fig. 2) wherein the annotation terms for Biological Process, Molecular Function and Cellular Component were determined. The 119 buffalo sperm-surface proteins were successfully mapped to 63 entries in the background dataset. The singular enrichment analysis (SEA) for Biological Process terms' identi ed reproductive processes, sexual reproduction, immune response and response to biotic/abiotic stimulus terms as the major GO annotations (Table 2) in the input list vis-à-vis the background reference dataset, the Bovine genome locus (Bovine Genome Database): GLEAN_03528. The scatter plot analysis (SPA) using SimRel for Biological Process similarly indicated semantic similarities between reproductive process functions, immune response and response to biotic/abiotic stimulus terms ( Fig. 1) as observed by their closeness in the displayed two-dimensional space.
The SEA for Molecular Function indicated that the majority of proteins were ( Fig. 1) involved in catalytic and binding (carbohydrate or protein) functions. The SPA for Molecular Function (Fig. 1) also identi ed protein binding and catalytic activity as the major GO terms with the highest uniqueness index values and the least dispensability scores. Most of the proteins were found to be extracellular, vesicular or part of the plasma membrane as indicated by the SEA and SPA for the Cellular Component terms (  Overall, both the treatments reduced the availability of respective cognate glycans for most lectins except the PNA after salt treatment. Contrarily, the PI-PLC treatment led to increased exposure of α-2, 3 linked sialic acid and asialylated galactosyl (β-1, 3) N-acetylgalactosamine. Furthermore, both the treatments were signi cantly different from each other vis-à-vis the MFI produced upon lectin binding on the surface of the buffalo spermatozoa. Differential spatial distribution of BuBD-129 and 126 The peptides GRCKEYCNMDEKELDK for BuBD-129 and NKTGNCRSTCRNGEK for BuBD-126 were predicted to be highly antigenic and were thus adjudged as the best B-cell epitopes. This is because they were predicted to be preferentially present in turns and loops and had a comparatively higher probability for being found on the surface ( Supplementary Fig. 4). Initially, the crude concentration of the isolated IgGs assayed by measuring the Blocking BuBD-129 on sperm surface hinders cleavage, Morula and Blastocyst formation rates The addition of anti-BuBD-129 antibody in the fertilization medium appeared to hamper the fertilization in a dosedependent manner (Fig. 4). The percentage of cleaved oocytes decreased in the 1:15000 dilution group compared to the control group which further dropped signi cantly (P < 0.05) in the 1:10000 and 1:5000 (P < 0.00001) dilution. Both the 1:10000 and 1:15000 differed signi cantly (p < 0.001) from the 1:5000 dilution and the control group for the number of cleaved oocytes. The subsequent stages of embryo development e.g. the morula formation also exhibited a similar trend. The percentage of morula formed decreased in the 1:15000 dilution but declined signi cantly (P < 0.05) in 1:10000 dilution which further reduced (P < 0.00001) in the 1:5000 dilution group. As expected, the blastocyst formation rate was highest in control which declined (p < 0.01) on the addition of anti-BuBD-129 in 1:15000 and 1:10000 dilution groups (Fig. 4). No blastocyst was formed in the 1:5000 dilution group.

Discussion
The present study was designed to identify the proteins associated with peripheral coats on the buffalo spermatozoa acquired during their transit through the epididymis and other ducts before ejaculation. The overrepresented immune-related glycoproteins/glycoconjugates of sperm surface are known to regulate male fertility e.g. by assisting in immune-evasion (28 and 29). We sought to speci cally remove i) the proteins bound through electrostatic interactions (by elevated NaCl) ii) the GPI-anchored proteins (by PI-PLC). We also wanted to determine if the previously detected epididymal transcripts of buffalo beta-defensins (BDs) are translated and their gene products are subsequently applied to the buffalo sperm surface. Shotgun proteomic pro ling revealed that the majority of the extracellular proteins is involved in immune response or reproductive processes ( Table 2).
Eight BDs including two BDs, implicated in male fertility viz. the heavily O-glycosylated BuBD-129 and BuBD-126 were also identi ed along with other sperm-surface proteins (Supplementary sheet-Results). The in silico prediction of glycosylation of sperm surface was validated by ow cytometry using six lectins (Fig. 2). The presence of BuBD-129 and 126 was con rmed by immuno uorescence which revealed a differential immunolocalization pattern of these BDs (Fig. 3). Besides, the blocking of BuBD-129 with antibodies was found to hamper the fertilization of buffalo oocytes which subsequently affected embryogenesis (Fig. 4).
The sperm surface plays a crucial role in biomolecular interactions, intracellular communication and gamete recognition. However, the testicular spermatozoa still require distinct post-gonadal modi cations to become competent to traverse the FRT and fertilize the oocyte (19 and 40). The concluding stages of the sperm differentiation, including the tailoring of its surface, occur outside the gonads and do not appear to be regulated or mediated by the germline genome. Therefore, subtle interactions between the sperm and the luminal milieu of the epididymis modify the surface of spermatozoa in a series of sequential biochemical modi cations which includes removal or addition of (glyco)proteins and changes in the glycoconjugates (GCs) associated with these proteins (9, 13, 19 and 41). Broadly, two distinctive and separate populations of (glyco)proteins have been described on mammalian spermatozoa which are differentiated based on their interactions with sperm-surface. One of them is adsorbed onto the sperm surface by either electrostatic or hydrophobic interactions and isn't integrated into the sperm plasma membrane (3, 19, 42 and 43). To elucidate such a sperm surface antigen, we sought to remove the non-covalently linked epididymal proteins from the buffalo spermatozoa surface using an elevated NaCl concentration (DPBS treatment). It had been documented more than a decade ago that a population of non-covalently bound sperm surface (glyco)proteins could be released by exposing the macaque differs from the DEFB-126 binding pattern observed in monkeys (79) or rodents (80). The BuBD-126 was found to localize to the post-acrosomal region and the tail rather than environing the whole sperm surface. These speciesspeci c differences could be ascribed to the high variability in the distribution pattern of glycans and the differential sperm associated glycan topography(SpAGT) amongst various species (28).
The blocking of BuBD-129 drastically reduced the cleavage as well as the blastocyst formation rates during IVF. It is well established that the antibodies directed against the sperm-speci c antigens are a major cause of immunological infertility since they perturb the normal fertilization process (3, 50 and 81). How the anti-sperm antibodies target the sperm is not clearly understood. One facet of this obscurity may be unravelled by our demonstration that blocking the BuBD-129 precluded a successful fertilization event which con rmed their presence on buffalo sperm even after capacitation. Likewise, the ortholog of primate DEFB-126 in mice has been demonstrated to incorporate in the oolemma during gamete fusion which subsequently oated and extended out from the fused spermatozoa (90 and 91). Besides, its ortholog in cattle has been reported to be retained on the sperm surface, after induction of in vitro capacitation (51). This suggests that the BBD-126, like its ortholog defb22, remains associated with the spermatozoa during fertilization indicating an additional role in fertilization (80). The antibodies to heavily glycosylated BuBD-129 interfered with normal fertilization event apparently by inhibiting the recognition of oocytes by spermatozoa. It has been proposed that these features may be true for all the BDs with high levels of O-glycosylation (84). Similar to BuBD-129, antibodies against another member of the CAP superfamily, CRISP1 have been demonstrated to obstruct fertilization by interfering in the sperm-oocytes fusion process, thus reducing fertility. The cysteine-rich defensin-like peptides appear to be integral to the reproductive success of organism ranging from invertebrates, plants to higher primates (85, 86 and 87). It is equally likely that the blocking of BuBD-129 changes the attributes of sperm surface features which not only in uence fertility but also the developmental potential of the subsequent embryos. The BuBD-129 probably imparts surface properties that are essential for fertilization by the buffalo sperm due to its glycocalyx and uniform distribution on entire buffalo sperm (88). These results suggest multiple and putatively epistatic roles of the BD genes in immune response and reproductive physiology of buffalo, e.g.in the process of sperm-oocytes interaction.

Conclusion
The buffalo sperm surface is heavily glycosylated and many glycoproteins are applied as peripheral coats Data processing The generated raw les for the seven samples were analyzed using Trans-Proteomic Pipeline TPP v5.1 (Syzygy) rev. 0 (90), against a database generated from UniProt knowledgebase (Bos taurus, Bubalus bubalis and betadefensin, downloaded January 3, 2019; www.uniprot.org) using the Comet search engine (91). The precursor and fragment mass tolerances were set at 10 ppm and 0.5 Da, respectively. The enzyme speci city was set to trypsin/Panda maximum of two missed cleavages were allowed. Carbamidomethyl on cysteine (C) was considered as xed modi cation while oxidation of methionine and N-terminal acetylation were considered as variable modi cations for the database search. The peptide spectrum match and protein false discovery rates (FDR) were set to 0.01 to increase the con dence and remove the false-positive identi cations. Only the proteins with iProphet probability greater than 0.99 were considered for further analysis (92).

Singular Enrichment and Scatter Plot analyses of the identi ed proteins
The gene-ontology and singular enrichment analysis for identi ed proteins in all of the seven treatments were performed using agriGO (93)

Immunocytochemistry
The swim-up fraction of spermatozoa (40 x 10 6 cells/ml) was re-suspended in NCM and was added onto poly-Llysine coated slides. The NCM was removed after 15 min since the spermatozoa adhered to the slide surface by that time. Subsequently, the cells were washed twice in PBS and xed in 2% paraformaldehyde and 0.1% glutaraldehyde for 20 min at RT. The movement of the GPI-linked proteins is prevented by using low glutaraldehyde concentration. The spermatozoa were then washed with PBS thrice, and the slide surface was then blocked with blocking buffer (1% BSA in PBST-0.1 % Tween 20 in 1X-PBS) for 1 hour at room temperature.
The cells were later incubated with the primary polyclonal antibody (1:1000 dilutions) against BuBD-129 and 126 overnight at 4°C. The sperm were washed with PBST thrice and were then incubated with FITC conjugated goat anti-rabbit IgG secondary antibody (1:5000 dilutions; Sigma-Aldrich) in dark for 1 hour at RT followed by nal washings with PBST (3x). After the nal washing, a coverslip was mounted onto a glass slide onto which one drop of mounting medium, Dabco® 33-LV was placed. The cells were then observed under a BX-51 Olympus uorescence microscope. The loss of the BuBD-129 and 126 from the sperm-surface after incubation of spermatozoa in 2X-DPBS or exposing them to 2U/ml of PI-PLC was similarly monitored by immunocytochemistry (ICC) after the stipulated times of incubation.

IVF study
The IVF was performed using the procedure described by Verma et al. (97) in the control and three treatment groups. The control group was without anti-BuBD-129 and the three treatment groups comprised of three different concentrations of anti-BuBD-129 (0.5mg/ml) antibody in the fertilization medium drops viz. the 1:15000, 1:10000 and 1:5000 dilution groups. A total of four biological replicates were used for the IVF experiments. The data were analyzed on GraphPad prism7 (for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com) to compare the observed differences in cleavage and blastocyst rates among the control and treatment groups. Ethics approval and consent to participate: The animal study was reviewed and approved by Institutional Animal Ethics Committee (IAEC), National Dairy Research Institute (NDRI). All experiments were performed in accordance with guidelines and regulations laid by IAEC-NDRI. The study was carried out in compliance with the ARRIVE guidelines.

Consent for publication:
Not Applicable Study was designed by VB 1 , RK and TKD. Analysis of proteomics data was done by VB 1 , VB2 and SAA. The IVF was performed by PS. The images were generated by AK and SS. The manuscript was written by VB 1 . All the authors read and approved the nal manuscript Figures Figure 1 The scatter plot analysis (SPA) results for Biological Process, Molecular Function and Cellular Component GO terms of buffalo sperm surface proteins The overall research methodology followed for BuBD identi cation on the buffalo sperm surface by LC-MS/MS.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.