3D in situ imaging of female reproductive tract reveals molecular signatures of fertilizing spermatozoa in mice

Out of millions of ejaculated sperm, only a few reach the fertilization site in mammals. Flagellar Ca2+ signaling nanodomains, organized by multi-subunit CatSper calcium channel complexes, are pivotal for sperm migration in the female tract, implicating CatSper-dependent mechanisms in sperm selection. Here, using biochemical and pharmacological studies, we demonstrate that CatSper1 is an O-linked glycosylated protein, undergoing capacitation-induced processing dependent on Ca2+ and phosphorylation cascades. CatSper1 processing correlates with protein tyrosine phosphorylation (pY) development in sperm cells capacitated in vitro and in vivo. Using 3D in situ molecular imaging and ANN-based automatic detection of sperm distributed along the cleared female tract, we demonstrate that all spermatozoa past the UTJ possess intact CatSper1 signals. Together, we reveal that fertilizing mouse spermatozoa in situ are characterized by intact CatSper channel, lack of pY, and reacted acrosomes. These findings provide molecular insight into sperm selection for successful fertilization in the female reproductive tract.


27
In most mammals, millions or billions of spermatozoa are deposited into the cervix upon coitus. 28 Yet less than 100 spermatozoa are found at the fertilization site, called ampulla, and only 10-12 29 spermatozoa are observed around an oocyte (Kolle, 2015;Suarez, 2002). This implies the 30 presence of mechanisms to select sperm as they travel through the female reproductive tract and 31 to eliminate non-fertilizing, surplus spermatozoa once the egg is fertilized (Sakkas et al., 2015). 32 Recent ex vivo imaging studies combined with mouse genetics have shown that surface 33 molecules on the sperm plasma membranes such as ADAM family proteins are essential for the 34 sperm to pass through the utero-tubal junction (UTJ) (Fujihara et al., 2018). By contrast, whether 35 such selection and elimination within the oviduct requires specific molecular signatures and 36 cellular signaling of spermatozoa is not fully understood. 37 38 Mammalian sperm undergo capacitation, a physiological process to obtain the ability to fertilize 39 the egg, naturally inside the oviduct (Austin, 1951;Chang, 1951 suggests that in vitro sperm capacitation does not precisely reproduce the time-and space-44 dependent in vivo events in the oviduct. Protein tyrosine phosphorylation (pY), which has been 45 utilized as a hallmark of sperm capacitation over decades, showed different patterns in boar 46 sperm capacitated in vitro from ex vivo and in vivo (Luno et al., 2013). In mice, pY is not required 47 for sperm hyperactivation or fertility (Alvau et al., 2016;Tateno et al., 2013). Previous in vitro 48 studies that represent the population average at a given time may or may not have observed 49 molecular details of a small number of the most fertilizing sperm cells. 50 51 Capacitation involves extensive sperm remodeling that triggers cellular signaling cascades. 52 Cholesterol shedding and protein modifications occur within the plasma membrane (Visconti et 53 al., 1999;Vyklicka and Lishko, 2020). Cleavage and/or degradation of intracellular proteins by 54 individual proteases and ubiquitin-proteasome system (UPS) also participate in the capacitation 55 process (Honda et al., 2002;Kerns et al., 2016). Various capacitation-associated cellular signaling 56 pathways that include cAMP/PKA activation followed by pY increase and rise in intracellular pH 57 and calcium result in physiological outcomes such as acrosome reaction and motility changes 58 (Balbach et Figure 1B, D).

92
To better understand the processing of CatSper1, we first examined CatSper1 protein expression 93 in the testis and epididymis. Interestingly, the molecular weight of CatSper1 increases gradually 94 during sperm development and epidydimal maturation ( Figure 1A; upper), indicating that 95 CatSper1 undergoes post-translational modifications. We next examined the nature of the 96 modifications. Block of tyrosine phosphatases by sodium orthovanadate or addition of specific 97 protein phosphatases, PP1 or PTP, does not change the molecular weight of CatSper1 (Figure1-98 figure supplement 1A). In contrast, when sperm membrane was subjected to enzymatic 99 deglycosylation, O-glycosidase, but not PNGase F, shifts apparent molecular weight of CatSper1 100 to close to the CatSper1 band with the smallest molecular weight observed in testis ( Figure 1A, 101 B, Figure1-figure supplement 1B). These data suggest that CatSper1 in sperm is not a 102 phosphoprotein but an O-linked glycosylated protein.

104
CatSper1 resides in the subdomains of lipid rafts in mature sperm and processed during 105 capacitation 106 Sucrose density gradient centrifugation identifies CatSper1 in lipid raft subdomains in mature 107 sperm ( Figure 1C). Because cholesterol depletion destabilizes the plasma membrane during 108 sperm capacitation, one simple hypothesis is that the capacitation-associated changes in raft 109 stability and distribution (Nixon et al., 2007) render CatSper1 accessible to a protease activity. 110 Before inducing capacitation, CatSper1 is not processed in sperm cells, probably because the 111 CatSper1-targeting protease activity is normally not in the immediate vicinity to the CatSper 112 nanodomains in the flagellar membrane ( Figure 1-figure supplement 1C, E). Supporting this 113 notion, the protease activity readily cleaves CatSper1 by solubilizing the sperm membrane 114 fraction with Triton X-100 ( Figure 1-figure supplement 1C).

116
The CatSper1 N-terminus undergoes capacitation-associated degradation in vitro 117 We next investigated the location of CatSper1 cleavage and degradation using recombinant 118 CatSper1 proteins and sperm lysates. The CatSper1 antibody used in this study is raised against 119 the first N-terminal 150 amino acids of recombinant CatSper1 (Ren et al., 2001). C-terminal HA-120 tagged full-length (FL) or N-terminal deleted (ND) recombinant CatSper1 are expressed in HEK 121 293T cells for pull-down and detection by western blot (Figure 1E, F). Solubilized sperm lysates 122 degrade FL-CatSper1 and result in increased detection of cleaved CatSper1 by HA antibody 123 ( Figure 1G; upper). In contrast, protein levels of recombinant ND-CatSper1 are not affected by 124 incubation with sperm lysate ( Figure 1G; lower). These results demonstrate that the cytoplasmic 125 N-terminal domain of CatSper1 is the target region for proteolytic activity in sperm cells. How is 126 the CatSper proteolytic activity regulated? 127 128 CatSper1 degradation involves Ca 2+ and phosphorylation-dependent protease activity 129 At the molecular level, capacitation is initiated by HCO3uptake, which activates soluble adenylyl 130 cyclase (sAC), resulting in increased cAMP levels. HCO3also stimulates CatSper-mediated Ca 2+ 131 entry into sperm cells by raising intracellular pH (Figure 1-figure supplement 1E). We thus 132 examined whether the proteolytic activity requires cAMP/PKA and/or Ca 2+ signaling pathways. 133 Interestingly, adding a PKA inhibitor H89 or the St-Ht31 peptide, which abolishes PKA anchoring 134 to AKAP during sperm capacitation, accelerated CatSper1 degradation during sperm capacitation 135 ( Figure 1H, Figure 1-figure supplement 1D). Consistently, calyculin A, a serine/threonine protein 136 phosphatase inhibitor, suppresses the capacitation-associated CatSper1 degradation. These 137 data suggest that regulation of the proteolytic activity targeting CatSper1 involves protein 138 phosphorylation cascades (Figure 1-figure supplement 1E). Interestingly, adding Ca 2+ ionophore 139 A23178 to the sperm suspension was sufficient to induce CatSper1 processing even under non-140 capacitating conditions that do not support changes in PKA activity ( Figure 1I, J). Thus, a Ca 2+ 141 dependent protease that is indirectly regulated by protein phosphorylation such as calpain (Ono 142 et al., 2016) may process CatSper1. We observed that calpain inhibitors prevent CatSper1 from 143 capacitation-associated degradation ( Figure 1K). in the spermatozoa that passed the utero-tubal junction (UTJ) are arranged normally along the 168 tail, mostly protected from degradation, but in decreasing intensity and continuity more towards 169 UTJ ( Figure 2D, Figure 1-figure supplement 1E). In striking contrast, pY is not detected in the 170 spermatozoa from the ampulla but appears in the oviductal sperm increasingly towards UTJ 171 ( Figure 2E). Absence of EGFP reveals that spermatozoa from the ampulla are fully capacitated 172 and acrosome reacted (AR) but the those in the isthmus are undergoing AR ( Figure 2D Interestingly, we found that a few CatSper1 -/sperm cells that managed to arrive at the ampulla 178 are all not acrosome reacted ( Figure 2G), supporting the notion that CatSper-mediated Ca 2+ 179 signaling is required for sperm acrosome reaction (Stival et al., 2018). These results suggest that 180 escape of CatSper1 from the cleavage and subsequent degradation suppresses pY development, 181 enabling sperm to maintain hyperactivation capability, prime AR, and achieve the fertilization in 182 vivo. 183 184 3D in situ molecular imaging of gametes in the female reproductive tract 185 The physiological importance of tracing a small number of spermatozoa progressing to the 186 fertilization site prompted us to seek a method that enables direct molecular assessment of single 187 cells inside the intact female tract. We have adapted tissue clearing technologies to establish 188 three-dimensional (3D) in situ molecular imaging systems for fertilization studies (Figure 3, Figure  189 3 -figure supplement 1, Videos 1-6). We found that various tissue clearing methods (Chung et 190 al we anticipated that sperm cells that successfully reach the ampulla would be CatSper1-intact and 226 acrosome reacted. As expected, most sperm cells located in the ampulla exhibit linearly arranged 227 intact CatSper1 and reacted acrosomes ( Figure 4A, upper, Video 8). In the middle isthmus, both 228 CatSper1 and acrosome remain intact in most sperm cells, but mixed patterns are observed in 229 some cells ( Figure 4A, middle, Video 8). Interestingly, acrosome is largely intact in the sperm 230 clusters in the proximal isthmus close to UTJ whereas CatSper1 is barely detected ( Figure 4A, 231 lower, Video 8). This contrasts with the reduced but readily visible CatSper1 in the sperm from 232 the same region by microdissection ( Figure 2D). It is possible that the relatively longer tissue 233 processing time and subsequent labeling could have contributed to lower the signal to noise ratio 234 to a certain degree. Notably, 3D volume imaging of this mid isthmus regions reveals sperm cells 235 aligned in one direction towards the ampulla, providing unprecedented insight into sperm taxis in 236 the fertilization process (Video 8). Our qualitative but semi-quantitative analyses suggest that 237 CatSper1 is largely protected from degradation once in the oviduct; acrosome reaction initiates in 238 the mid-isthmus and is completed in the ampulla before interacting with the oocytes ( Figure 4B). 239 These results are consistent with our initial observations from microdissection and ex vivo imaging 240 studies ( Figure 2D, F), validating the information obtained by our in situ molecular imaging 241 platform. Taken together, we conclude that intact CatSper1, lack of pY, and reacted acrosome 242 are molecular and functional signatures of most fertilizing spermatozoa in the physiological 243 context.

245
Automatic detection of sperm in the voluminous female tract using artificial neural network 246 Processing 3D volumetric fluorescent data presents a significant challenge; analyses of sperm in 247 the female tract includes object identification in the voluminous specimen, object separation from 248 background noise, and object alignment in three dimensions. To address these logistics 249 problems, we took an advantage of the artificial neural network (ANN) approach for automatic 250 localization and signal isolation. We performed a proof-of-principle investigation utilizing CatSper1 251 distributions in sperm cells from our 3D in situ molecular imaging ( Figure 5). First, we manually cleavage and degradation, leading to a heterogeneous sperm population (Figures 1, 2). Building 286 on our observations of sperm cells from microdissection, ex vivo imaging, and CLARITY-based 287 in situ molecular imaging (Figures 2, 3, 4), we hypothesize that CatSper1 is a built-in countdown 288 timer for sperm death and elimination in the female tract; CatSper1 cleavage and degradation, 289 triggered in a time-and space-dependent manner along the female tract, signals to end sperm 290 motility, and ultimately sets sperm lifetime in vivo. With our newly developed automated ANN 291 method to obtain high-quality 3D fluorescent images of CatSper1 in the sperm cells from cleared 292 female tract samples, we further tested this idea by quantitatively analyzing the CatSper1 signals 293 in situ.

295
Our in situ imaging platform offers the typical resolution that a confocal microscopy can provide; 296 two separated CatSper1 arrangement along the sperm tail (Chung et al., 2017) are detected 297 without any computational processing ( Figure 3I, 4A). This encouraged us to develop an analytical 298 procedure to assess the status of CatSper1 quadrilateral and linear distributions. We isolated the 299 fluorescent signal from a proximal region of the principal piece close to the annulus where 300 CatSper1 signal is the most intense ( Figure 6B). To superpose the individual cross-sectional 301 images according to the expected 4 intensity peaks, we aligned randomly oriented transversal-302 projection images by placing the quadrant with the highest fluorescent intensity to upper right 303 corner ( Figure 6B, inset). The aligned images were then superposed ( Figure 6C) and used for 304 statistical purposes to represent quadrilateral arrangement of CatSper1 in individual sperm cells 305 ( Figure 6D). The individually processed images of sperm cells from the oviductal regions close to 306 UTJ, middle isthmus, and ampulla, regions were again superposed to create cumulative diagrams 307 and heat maps corresponding to these regions ( Figure 6E). They show quadrilateral distribution 308 of enriched CatSper1 signal more clearly from the sperm population in the ampulla compared to 309 the population in the oviduct close to UTJ ( Figure 6E, G, H).

311
To further quantify and statistically analyze our outputs, we divided the pre-processed images of 312 individual sperm cells on 80 round areas ( Figure 6 -figure supplement 1A) and calculated 313 fluorescent intensities among them. The quantified intensity from the 80 areas were plotted; the 314 observed 4 peaks (highest intensity) and valleys (lowest intensity) were used to calculate the delta 315 value among them to represent the quality of CatSper1 quadrilateral structure ( Figure 6F). Our 316 quantitative analysis ( Figure 6G, Figure 6 -figure supplement 1B) shows consistent results with 317 our previous semi-quantitative analysis by manual assignment of the CatSper1 patterns ( Figure  318 4). Together with the whole tissue image processing ( Figure 3E), the quantitative analysis clearly 319 visualizes that sperm populations located along the cleared oviduct have statistically different 320 CatSper1 quadrilateral intensity delta values ( Figure 6H). 321 322

324
CatSper1 as a molecular barcode for sperm maturation and transition in the female tract 325 Testicular spermatozoa undergo maturation and biochemical alterations in the intraluminal 326 environment of the epididymis (Cornwall, 2009). Glycan-modifying enzymes such as glycosidases 327 and glycosyltransferases are present in the epididymal luminal fluid (Tulsiani, 2003 The effect of CatSper1 truncation on channel activity and sperm motility remains to be determined 365 in future studies. CatSper1 truncation may be coordinated with molecular changes of other 366 CatSper subunits. For example, the protein level of CatSper2, but not CatSper3 or 4, also 367 decreases after capacitation when probed with the antibody recognizing its C-terminal domain cancer cells (Liu et al., 1994). We propose that capacitation-associated global pY development 405 represents degenerating sperm which might concomitantly lose motility. It is intriguing that 406 capacitation-associated reactive oxygen species (ROS) generation activates intrinsic apoptotic 407 cascade and compromises sperm motility (Koppers et al., 2011). Consistent with this idea, ROS 408 inactivates protein tyrosine phosphatase (Tonks, 2005) and enhances pY development in sperm 409 (Aitken et al., 1998 than those observed at the fertilization site in vivo (Suarez, 2006). Sperm capacitated in vitro do 421 not encounter the anatomically and spatially distinct environment of the female reproductive tract, 422 for example, missing their interaction with the oviductal epithelial cells. In vitro capacitation also 423 lacks secretory factors from the male and female reproductive tracts that can affect the surface 424 protein dynamics during the capacitation process (Flesch and Gadella, 2000). Mouse models that 425 typically use epidydimal sperm for in vitro studies do not contain secretions from male glands. The present study opens up new horizons to microscopically visualize and analyze molecular 448 events in single sperm cells that achieve fertilization. This will allow us to better understand 449 physiologically relevant cellular signaling pathways directly involved in fertilization. We also have 450 illustrated that the same approach of tissue-clearing based 3D in situ molecular imaging is 451 applicable to study gametogenesis in situ. Future areas for investigations as natural extensions 452 of the current study are gameto-maternal interaction, development, transport, and implantation of 453 early embryos and maternal-fetal communication. Developing gamete-specific antibodies and/or 454 knockout validated antibodies to probe molecular abundancy and dynamics in situ and post-455 processing tools for various parameters will be critical to this end. 456 457 EGFP mice were crossbred with CatSper1-null mice to generate Su9-DsRed;Acr-EGFP 463

Materials and Methods
CatSper1-null mice. WT C57BL/6 and B6D2F1 male and CD1 female mice were purchased from 464 Charles River Laboratories (Wilmington, MA) and Jackson laboratory (Bar Harbor, ME). Mice 465 were cared in accordance with guidelines approved by the Yale Animal Care and Use 466 Committees.

474
Antibodies and Reagents 475 In Recombinant protein expression 506 HEK293T cells were transfected with constructs encoding FL-CatSper1 or ND-CatSper1 to 507 express the recombinant proteins transiently. Polyethyleneimine was used for the transfection 508 following the manufacturer's instruction as previously. 509

Protein Preparation and Western blot. 511
Total Protein Extraction 512 Total proteins were extracted from sperm, testis, and cultured mammalian cells as previously 513 described (Hwang et al., 2019). In short, collected epididymal sperm cells were washed with PBS 514 and lysed in 2X LDS sampling buffer for 10 min at room temperature with agitation (RT). The 515 whole sperm lysates were centrifuged at 14,000 x g for 10 min at 4 °C. Testes were homogenized 516 in 0.32M sucrose and centrifuged at 1,000 x g for 10 min at 4 °C to remove cell debris and nuclei. 517 1% Triton X-100 in PBS containing protease inhibitor cocktail (cOmplete TM , EDTA-free, Roche) 518 was added to the cleared homogenates to make total testis lysate. The lysates were centrifuged 519 at 4 °C, 14,000 x g for 30 min and the supernatant was used for the downstream experiments. 520 Transfected HEK29T cells and COS-7 cells were washed and lysed with 1% Triton X-100 in PBS 521 with protease inhibitor cocktail (Roche) at 4 °C for 1 hr. Cell lysates were centrifuged at 14,000 x 522 g for 30 min. All the solubilized protein lyates from the sources described above were reduced by 523 adding dithiothreitol (DTT) to 50 mM and denature by heating at 75 °C for 5 min (testis and 524 cultured cells) or 10 min (sperm).

526
Discontinuous Sucrose Density Gradient Centrifugation 527 Discontinuous sucrose density gradient centrifugation was performed as previously described 528 (Kaneto et al., 2008). To isolate and solubilize membrane fraction without using a detergent, 529 cauda epididymal sperm cells washed and suspended in PBS (1.0 x 10 8 cells/ml) were sonicated 530 3 times for 1 sec each. Sonicated sperm cells were then centrifuged at 5,000 x g for 10 min at 4 531 °C and the solubilized membrane fraction in supernatant was collected. The solubilized 532 membrane fraction was pelleted by ultracentrifugation at 100,000 x g for 1 hr at 4 °C and 533 resuspended with PBS. The membrane suspension was mixed with equal volume of 80 % 534 sucrose in PBS. A discontinuous sucrose gradient was layered with the 40%, 30 %, and 5 % 535 sucrose solution from bottom to top in a tube discontinuously. The gradient was ultracentrifuged 536 at 200,000 x g for 20 hr at 2 °C. Proteins collected from each fraction were precipitated with 5 % 537 of trichloroacetic acid, ethanol washed, and dissolved in SDS sampling buffer. 538 539 Dephosphorylation of Sperm Membrane Proteins 540 Sperm membrane fractions from 1 x 10 6 sperm cells prepared as above were treated with protein 541 phosphatase 1, (PP1, 0.1 unit; NEB), protein tyrosine phosphatase (PTP, 5 unit; NEB), or sodium 542 orthovanadate (Na3VO4, 1 mM; NEB) to test dephosphorylation of CatSper1. The membrane 543 fractions were incubated with the phosphatases or Na3VO4 in a reaction buffer containing 20mM 544 HEPES, 0.1 mM EDTA and 0.1mM DTT at 30 °C for the indicated times. The isolated sperm 545 membrane was solubilized by adding Triton X-100 to final 0.1% in PBS (PBS-T) for the indicated 546 times at RT. 547 548 Enzymatic Deglycosylation 549 Glycosylation of CatSper1 from cauda sperm was tested using PNGase F (Sigma-Aldrich) and 550 O-glycosidase (NEB). Sperm cells were washed with 1x reaction buffer for each enzyme by 551 centrifugation at 800 x g for 3 min. Sperm pellets were re-suspended with each 1x reaction buffer 552 (20 mM or 50 mM sodium phosphate, pH7.5 for PNGase F and O-glycosidase, respectively) and 553 followed by sonication and centrifugation to collect sperm membrane fraction as described above. 554 Collected supernatants were incubated with denaturation buffer at 100 °C for 5 min to denature 555 glycoproteins before subject to enzymatic deglycosylation. Denatured protein samples were subjected to SDS-PAGE. Rabbit polyclonal CatSper1 (2 µg/ml), 572 CatSper3 (2 µg/ml), CatSperε (1.6 µg/ml), and CAIV (1:500) antibodies and monoclonal HA (clone 573 C29F4; 1:2,000), caveolin1 (clone 2297, 1:500), acetylated tubulin (clone 6-11B-1; 1:20,000), 574 phosphotyrosine (clone 4G10; 1:1,000), and ubiquitin (clone P4D1; 1:1,000) antibodies were used 575 for western blot. Anti-mouse IgG-HRP (1:10,000) and anti-rabbit IgG-HRP (1:10,000) were used 576 for secondary antibodies. 577 578 Sperm Migration assay 579 Sperm migration assay was performed as previously described (Chung et al., 2014). Briefly, 580 female mice were introduced to single-caged Su9-DsRed;Acr-EGFP males for 30 min and 581 checked for vaginal plug. Whole female reproductive tracts were collected 8 h post-coitus and 582 subjected to ex vivo imaging to examine spermatozoa expressing reporter genes in the tract 583 (Eclipse TE2000-U, Nikon).

585
Collection of in vivo Capacitated Sperm 586 Female reproductive tracts from timed-mated females to Su9-DsRed;Acr-EGFP or Su9-587 DsRed;Acr-EGFP CatSper1-null males were collected 8 h post coitus. Sperm cells were released 588 by micro-dissection of female reproductive tract followed by lumen flushing of each tubal segment 589 (cut into ~ 1-2 mm pieces). Each piece was placed in 50 µl of PBS on glass coverslips and the 590 intraluminal materials were fixed immediately by air-dry followed by 4% PFA in PBS. Ampulla and 591 uterine tissue close to UTJ were placed in 100 µl of PBS and vortexed briefly to release the sperm 592 within the tissues. Fixed sperm cells were subjected to immunostaining. 593 594 Sperm immunocytochemistry 595 Non-capacitated or in vitro capacitated sperm cells on glass coverslips were washed with PBS 596 and fixed with 4% paraformaldehyde (PFA) in PBS at RT for 10 minutes. Fixed samples were 597 permeabilization with PBS-T for 10 min and blocked with 10% normal goat serum in PBS for 1 hr 598 at RT. Blocked sperm cells were stained with primary antibodies, anti-CatSper1 (10 µg/ml) and 599 anti-phosphotyrosine (1:1,000), at 4 °C for overnight, followed by staining with secondary 600 antibodies for 1 hr at RT. Hoechst was used for counterstaining sperm head. Sperm cells were 601 mounted (Vectashield, Vector Laboratories) and imaged with confocal microscopes (Zeiss 602 LSM710 Elyra P1 and Olympus Fluoview 1000).

604
Tissue clearing and molecular labeling of the cleared tissues 605 CLARITY 606 All 3D volume images from the main figures (Figures 3 and 4) were taken from female tracts 607 subjected to CLARITY method (Chung et al., 2013) with slight modification by clearing tissue-608 hydrogel passively without involving electrophoresis. Timed-mated females (8 h post coitus) and 609 males were subjected to transcardiac perfusion using peristaltic pump. The mice were perfused 610 with each 20 ml of ice-cold PBS followed by freshly prepared hydrogel monomer solution (4% 611 acrylamide, 2% Bis-acrylamide, 0.25% Azo-inhibitor (VA-044, Wako), 4% PFA in PBS). The whole 612 female tract or testis-hydrogel were dissected from animals after perfusion and placed in 10 ml of 613 fresh hydrogel monomer solution for post-fixation. The collected tissues in monomer solution were 614 heated at 37 °C with degassing for 15 min, followed by incubation at 37 °C for 2-3 h for tissue 615 gelation. The gelated tissues were washed with clearing solution containing 200 mM boric acid 616 and 4% sodium dodecyl sulfate (pH8.5) three times for 24 h each by gentle rocking at 55 °C. 617 Cleared tissues were further washed with PBS-T for 24 h. The cleared female tracts were 618 subjected to dye-and/or immunolabeling: the cleared tissues were incubated with CatSper1 (7 619 µg/ml) or AcTub (1:100) antibodies in PBS-T for overnight at RT, followed by washing with PBS-620 T for 24 h. Washed samples were stained with the secondary antibodies (1:500) overnight.

621
Fluorescence dye conjugated PNA or WGA were used to detect sugar residues (1:1,000) and samples were put on imaging chamber filled with RIMS solution and imaged. All cleared tissues 625 were imaged with laser scanning microscope (Zeiss LSM710 Elyra P1). EC plan-Neofluar 10x/0.3, 626 LD LCI Plan-Apochromat 40x1.2, and Plan-Apochromat 63x/1.4 objectives were used for imaging. 627 Tile scanning and z-stacking for volume imaging were carried out with functions incorporated in 628 Zen black 2012 SP2 (Carl Zeiss) and Zen blue 2011 SP1 software (Carl Zeiss) was used for 3D 629 rendering. 630 X-CLARITY 631 Ovary, testis, and epididymis images (Figure 3 -figure supplement 1A, D-I) were taken from X-632 CLARITY method, following manufacturer's instruction (Logos biosystems). Animals 633 transcardially fixed with 4% PFA were post fixed in the fresh fixative for 4-6 h. The post-fixed 634 tissues were then immersed in a modified hydrogel solution (4% acrylamide, 0.25% Azo-inhibitor 635 (VA-044, Wako), 4% PFA in PBS) for 4-6 h. The samples were degassed and polymerized as 636 described in CLARITY method. The gelated tissues were washed with PBS and placed in 637 electrophoretic tissue clearing (ETC) chamber. Tissues in the ETC chamber were cleared by 638 clearing solution described above with active electrophoretic forcing of tissue for 6-8 hr. Cleared 639 tissues were washed and stained with β-actin and WGA.