Sperm membrane proteins DCST1 and DCST2 are required for the sperm-egg fusion process in mice and fish

The process of sperm-egg fusion is critical for successful fertilization, yet the underlying mechanisms that regulate these steps have remained unclear in vertebrates. Here, we show that both mouse and zebrafish DCST1 and DCST2 are necessary in sperm to fertilize the egg, similar to their orthologs SPE-42 and SPE-49 in C. elegans and Sneaky in D. melanogaster. Mouse Dcst1 and Dcst2 single knockout (KO) spermatozoa are able to undergo the acrosome reaction and show normal relocalization of IZUMO1, an essential factor for sperm-egg fusion, to the equatorial segment. While both single KO spermatozoa can bind to the oolemma, they rarely fuse with oocytes, resulting in male sterility. Similar to mice, zebrafish dcst1 KO males are subfertile and dcst2 and dcst1/2 double KO males are sterile. Zebrafish dcst1/2 KO spermatozoa are motile and can approach the egg, but rarely bind to the oolemma. These data demonstrate that DCST1/2 are essential for male fertility in two vertebrate species, highlighting their crucial role as conserved factors in fertilization.

DCSTAMP and OCSTAMP proteins represent an interesting group of proteins to study in the context of cell-cell fusion, since they have been shown to play a role in osteoclast and foreign body giant cell (FBGC) fusion [13][14][15] . They belong to the class of DC-STAMPlike domain-containing proteins and are multi-pass transmembrane proteins with an intracellular C-terminus containing a non-canonical RING finger domain 13,16,17 . DCSTAMP was shown to localize to the plasma membrane and endoplasmic reticulum (ER) membrane in dendritic cells and osteoclasts [16][17][18][19] . These cell types in Dcstamp KO mice show no apparent defect in differentiation into the osteoclast lineage and cytoskeletal structure, yet osteoclasts and FBGCs are unable to fuse to form terminally differentiated multinucleated cells 14 . Even though OCSTAMP is widely expressed in mouse tissues 20 , the only reported defect in Ocstamp KO mice is the inability to form multinucleated osteoclasts and FBGCs 13,15 . The fusion defect is not due to a change in the expression levels of osteoclast markers, including Dcstamp 13,15 . These results established an essential role for DC-STAMP-like domain-containing proteins in cell-cell fusion.
DC-STAMP-like domain-containing proteins, namely testis-enriched Sneaky, SPE-42, and SPE-49, are necessary for male fertility in Drosophila 21,22 and C. elegans 23-25 , respectively. Specifically, sneaky-disrupted fly spermatozoa can enter the egg, but fail to break down the sperm plasma membrane; the male pronucleus thus does not form and embryonic mitotic divisions do not occur 22 . Spe-42 and spe-49 mutant C. elegans spermatozoa can migrate into the spermatheca, the site of fertilization in worms, but these mutants are nearly or completely sterile, respectively, suggesting that SPE-42 and SPE-49 are involved in the ability of spermatozoa to fertilize eggs [23][24][25] . Sneaky, SPE-42 and SPE-49 have homologs in vertebrates called DCST1 and DCST2, but the roles of these proteins have remained undetermined. Here, we analyzed the physiological function of Dcst1 and Dcst2 and their effect on sperm fertility using genetically modified mice and zebrafish.

DCST1 and DCST2 are required for male fertility in mice.
RT-PCR analysis with multiple mouse tissues showed that Dcst1 and Dcst2 mRNAs are abundantly expressed in mouse testis ( Figure 1A). Using published single-cell RNAsequencing data 28 , we found that Dcst1 and Dcst2 mRNAs peak in mid-round spermatids, indicating that the expression patterns of Dcst1 and Dcst2 are similar to that of the other sperm-egg fusion-related genes Izumo1, Fimp, Sof1, Tmem95, and Spaca6 ( Figure 1B).
Using CRISPR/Cas9-mediated mutagenesis, we generated Dcst2 mutant mice lacking 7,223 bp (Dcst2 del/del ), which resulted in the deletion of almost all of the Dcst2 open reading frame (ORF) (Figure S2A-C). Of note, the expression level of Dcst1 mRNA in Dcst2 del/del testis decreased (Figure S2D), suggesting that the deleted region is required for Dcst1 expression in the testis. As shown in Figure S2A, Dcst1 and Dcst2 are tandemly arranged such that parts of their 5' genomic regions overlap. To assess the role of each gene, we generated Dcst1 indel mice (Dcst1 d1/d1 ) and Dcst2 indel mice (Dcst2 d25/d25 ) ( Figure S3A and B). RNA isolation from mutant testes followed by cDNA sequencing revealed that Dcst1 d1/d1 has a 1-bp deletion in exon 1, and Dcst2 d25/d25 has a 25-bp deletion in exon 4 (Figure S3C and D). Both deletions result in frameshift mutations leading to premature stop codons.  Figure 1C). Unexpectedly, the indel mutations in Dcst1 d1/d1 and Dcst2 d25/d25 decreased the expression level of Dcst1 mRNA in Dcst2 d25/d25 testis and Dcst2 mRNA in Dcst1 d1/d1 testis, respectively (Figure S3C). To evaluate the influence of the decreased expression level of Dcst1 and Dcst2 mRNAs on male fertility, we obtained double heterozygous (Dcst1 d1/wt and Dcst2 del/wt ) (dHZ) males through intercrossing. DHZ males showed a decreased expression level of both Dcst1 and Dcst2 mRNA in the testis, but 4 their fertility was comparable to that of the control (Figure S3E), indicating that the expression levels of Dcst1 mRNA from the Dcst2 d25 allele and Dcst2 mRNA from the Dcst1 d1 allele are decreased but still sufficient to maintain male fertility. This data reconfirms that DCST2 is indispensable for male fertility. Hereafter, we used Dcst1 d1/d1 and Dcst2 d25/d25 male mice for all experiments unless otherwise specified.
To confirm that the Dcst1 and Dcst2 disruptions are responsible for male sterility, we generated transgenic mice in which a testis-specific Calmegin (Clgn) promoter expresses mouse DCST1 and DCST2 with an HA tag at the C-terminus (Figure S6A and B). When Dcst1 d1/d1 males with the Dcst1-3xHA transgene and Dcst2 d25/d25 males with the Dcst2-3xHA transgene were mated with wild-type females, the females delivered normal numbers of offspring [pups/plug: 5.7 ± 0.5 (Dcst1 d1/d1 ; Tg, 25 plugs), 7.6 ± 2.7 (Dcst2 d25/d25 ; Tg, 15 plugs)] ( Figure 3A). We could detect HA-tagged DCST1 and HAtagged DCST2 in TGCs and spermatozoa at the expected sizes for the full-length proteins ( Figure 3B, arrowheads), though both proteins appear to be subject to post-translational processing or protein degradation.
To reveal the localization of DCST1 and DCST2 in spermatozoa, we performed immunocytochemistry with an antibody detecting the HA epitope and peanut agglutinin (PNA) as a marker for the sperm acrosome reaction. As shown in Figure 3C, PNA in the anterior acrosome was translocated to the equatorial segment after the acrosome reaction as shown previously 29 . While HA-tagged DCST1 could rarely be observed in spermatozoa, HA-tagged DCST2 was detected within the anterior acrosome of acrosomeintact spermatozoa, and then translocated to the equatorial segment after the acrosome reaction ( Figure 3C), mirroring the relocalization of IZUMO1 upon the acrosome reaction 30 . Fluorescence in the sperm tail was observed in both control and Dcst2-HA Tg spermatozoa, indicating that this signal in the tail was non-specific ( Figure 3C).
Taking advantage of the HA tag, we performed co-immunoprecipitation (co-IP). While HA-tagged DCST1 was detected only in TGCs, HA-tagged DCST2 was detected in both TGCs and spermatozoa ( Figure 3D). We could not detect IZUMO1 in these IP samples (Figure 3D), suggesting that DCST1 and DCST2 do not form a complex with IZUMO1. However, co-expression of Dcst1-3xFLAG and Dcst2-3xHA in HEK293T cells revealed the presence of a DCST1/DCST2 complexes (Figure 3E), which is in line with proposed complex formation between OCSTAMP and DCSTAMP during osteoclast fusion 31-33 .
HEK293T cells expressing DCST1/2 and IZUMO1 bind to, but do not fuse with, ZPfree eggs. To assess whether DCST1 and DCST2 are sufficient for inducing sperm-egg fusion, we overexpressed Dcst1-3xFLAG, Dcst2-3xHA, and Izumo1-1D4 in HEK293T cells ( Figure 4A). HEK293T cells overexpressing IZUMO1 could bind to, but not fuse with, ZP-free eggs (Figure 4B), which was consistent with previous reports 4,9 . In contrast, HEK293T cells overexpressing only DCST1 and DCST2 failed to bind to ZP-free eggs (Figure 4B and C). Co-expression of IZUMO1 and DCST1/2 allowed the cells to bind to ZP-free eggs [4.24 ± 2.41 cells/eggs (IZUMO1), 2.01 ± 1.93 cells/eggs (DCST1/2 and IZUMO1)], but did not facilitate fusion with the oolemma (Figure 4B and C). Thus, though DCST1 and DCST2 appear to have a role in the sperm-egg fusion process, they are not sufficient to induce fusion, even in conjunction with IZUMO1.
Sperm-expressed Dcst1/2 are also required for fertilization in zebrafish.
To assess to what extent our findings in mice could be expanded among vertebrate species, we asked what the roles of DCST1/2 are in an evolutionarily distant vertebrate species, the zebrafish. The orthologous zebrafish genes dcst1 and dcst2 are expressed specifically in testis and arranged similarly to mouse Dcst1/2 (Figure S7A and B). We therefore generated three independent KO fish lines, dcst1 -/-, dcst2 -/-, and dcst1/2 -/-, by CRISPR/Cas9-mediated mutagenesis (Figure S7B and C). 6 Lack of zebrafish Dcst2 alone or in combination with Dcst1 caused complete sterility in males, whereas lack of Dcst1 alone led to severe subfertility [5.5 ± 3.6% fertilization rate (dcst1 -/-, 16 clutches)] ( Figure 5A). The fertility of heterozygous males and KO females, however, was comparable to wild-type control males. (Figure 5A). Thus, similar to mice, Dcst1/2 are essential for male fertility in zebrafish. For further phenotypic analyses we decided to focus on the dcst2 -/mutant (unless stated otherwise), since loss of Dcst2 on its own is sufficient to cause complete sterility.
To understand what causes the fertility defect, we first determined whether spermatozoa were produced in mutant males. Dcst1 -/-, dcst2 -/-, and dcst1/2 -/males showed normal mating behavior and produced morphologically normal spermatozoa, indicating that zebrafish Dcst1/2 are not crucial for spermatogenesis ( Figure 5B). To examine where Dcst2 is localized in wild-type zebrafish spermatozoa, which lacks an acrosome, we produced antibodies against the C-terminal RING finger domain of zebrafish Dcst2. Dcst2 antibodies could detect zebrafish Dcst2 protein as determined by western blotting of wild-type and dcst2 -/sperm lysates ( Figure S7D) and immunofluorescence staining of zebrafish embryos overexpressing dcst2(RING)-superfolder GFP (sfGFP) mRNA ( Figure S7E). Interestingly, immunofluorescence against Dcst2 strongly stained wildtype spermatozoa at the periphery of the head in punctae and occasionally the mid-piece ( Figure 5C). Weaker staining of the tail region was also detected in dcst2 KO spermatozoa, suggesting that this signal was unrelated to Dcst2. Thus, Dcst2 localizes to the periphery of the sperm head.
When added to wild-type eggs, dcst2 KO spermatozoa were able to reach and enter the micropyle, the funnel-shaped site of sperm entry, similar to wild-type spermatozoa ( Figure 5D; Movie S3 and S4). We therefore conclude that Dcst2 is neither required for overall sperm motility nor for spermatozoa to approach and enter the micropyle. However, in contrast to wild-type spermatozoa which remained attached to the egg (Movie S3), most of the entering mutant spermatozoa subsequently detached and drifted away from the micropyle (Movie S4), suggesting that spermatozoa lacking Dcst2 are defective in stable binding to the oolemma. We previously established an assay to assess sperm-egg binding during zebrafish fertilization 34 . Building on this assay, we used live imaging of spermatozoa and eggs to quantify the number of wild-type sperm adhered to the oolemma within a physiologically relevant time frame [1.97 ± 0.97 spermatozoa/100 µm (12 eggs)] (Figure 5E-F and Movie S5). Performing this assay with dcst2 KO spermatozoa revealed that these spermatozoa were unable to adhere stably to wild-type eggs [0.05 ± 0.1 spermatozoa/100 µm (9 eggs)] (Figure 5E-F and Movie S5). We therefore conclude that zebrafish Dcst2 is required for stable binding of spermatozoa to the oolemma.

Discussion
Here, we demonstrate that the testis-enriched proteins DCST1/2 are necessary for male fertility in mice and fish. These results agree with the recently reported independent findings in mice 35 and the known requirements of Drosophila Sneaky and C. elegans SPE-42/49 in sperm for successful reproduction 21-25 , indicating that the physiological function of DCST1/2 is widely conserved among bilaterians. The conservation of DCST1/2 in sequence and function is remarkable for the otherwise rapidly evolving group of sperm-egg interacting proteins 36,37 . Contrary to vertebrates, which have both DCST1/2 and the related DCSTAMP/OCSTAMP proteins (Figure S1A), invertebrates have solely DCST1/2.
Having established the essential role of DCST1/2 for male fertility, questions arise concerning the molecular processes in which they are involved and how they contribute to achieve fertilization. We detected mouse DCST2 at the equatorial segment of acrosome-reacted spermatozoa and zebrafish Dcst2 at the periphery of the sperm head. Given that DCST1/2 are TM proteins, these localization patterns suggest that part of the proteins are exposed on the sperm surface. Given our result that DCST1/2 form a complex, we speculate that a DCST1/2 complex in the sperm membrane either helps organize the presentation of other fusion-related sperm proteins or directly interacts with other binding-and/or fusion-relevant molecules on the oolemma. To test this hypothesis, we overexpressed DCST1 and DCST2 in IZUMO1-expressing HEK293T cells. These cells could bind to, but not fuse with, ZP-free eggs ( Figure 4B-C), suggesting that DCST1/2 are not sufficient to induce fusion in a heterologous system. This observed lack of fusion could be due to the absence of other sperm-oocyte fusion-related factors (FIMP, SOF1, TMEM95, and SPACA6). Future investigation is needed to uncover the mechanism by which DCST1/2 act during fertilization.
The function of DCST1/2 in the sperm-egg fusion process differs between mice and fish: mouse DCST1/2 are required for the fusion process after sperm-egg binding (Figure 2), while zebrafish Dcst1/2 are required for sperm-egg binding (Figure 5). Given the diversity of the fertilization process across the animal kingdom, it may be that while DCST proteins are highly conserved, they may have evolved different roles to fit into the specific context of fertilization for a given species or animal group. Application of CRISPR/Cas9-mediated KO technology on a larger set of animals with divergent modes of fertilization will shed light on both the conservation and possible species-specific diversification in DCST1/2's function in fertilization.

Materials and Methods
Animals. B6D2F1, C57BL/6J, and ICR mice were purchased from Japan SLC and CLEA Japan. Mice were acclimated to 12-h light/12-h dark cycle. All animal experiments were approved by the Animal Care and Use Committee of the Research Institute for Microbial Diseases, Osaka University, Japan (#Biken-AP-H30-01). Zebrafish (Danio rerio) were raised according to standard protocols (28°C water temperature; 14/10-hour light/dark cycle). TLAB zebrafish served as wild-type zebrafish for all experiments, and were generated by crossing zebrafish AB stocks with natural variant TL (Tüpfel longfin) stocks. Dcst1 -/-, dcst2 -/-, and dcst1/2 -/mutant zebrafish were generated as part of this study as described in detail below. All fish experiments were conducted according to Austrian and European guidelines for animal research and approved by the local Austrian authorities (animal protocol GZ: 342445/2016/12).

RT-PCR for mouse multi-tissue expression analyses.
Total RNA was reverse-transcribed into cDNA using a SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen). PCR was conducted with primer sets (Table  S1) and KOD-Fx neo (TOYOBO). The PCR conditions were initial denaturation at 94°C for 3 minutes, denaturing at 94°C for 30 seconds, annealing at 65°C for 30 seconds, and elongation at 72°C for 30 seconds for 30 or 35 cycles in total, followed by 72°C for 2 minutes.

Mouse sperm motility and in vitro fertilization.
Cauda epididymal spermatozoa were squeezed out and dispersed in PBS (for sperm morphology) and TYH (for sperm motility and IVF) 39 . After incubation of 10 and 120 minutes in TYH, sperm motility patterns were examined using the CEROS II sperm analysis system 40-42 . IVF was conducted as described previously 43 . Protein extracts from the remaining sperm suspension in PBS and TYH drops were used for co-IP experiments.

Mouse sperm fusion assay.
The fusion assay was performed as described previously 9 . To visualize IZUMO1 distribution in spermatozoa, spermatozoa after incubation of 2.5 hours in TYH drops were then incubated with the IZUMO1 monoclonal antibody (KS64-125, 1:100) for 30 minutes. Then, the spermatozoa were incubated with ZP-free eggs in TYH drops with the mixture of IZUMO1 monoclonal antibody (KS64-125, 1:100) and goat anti-rat IgG Alexa Fluor 488 (1:200) for 30 minutes. Then, the eggs were gently washed with a 1:1 mixture of TYH and FHM medium three times, and then fixed with 0.2% PFA. After washing again, IZUMO1 localization was observed under a fluorescence microscope (BZ-X700, Keyence).

HEK293T-oocyte binding assay.
Mouse Dcst1 ORF-3xFLAG, mouse Dcst2 ORF-3xHA, mouse Izumo1 ORF-1D4 with a Kozak sequence (gccgcc) and a rabbit polyadenylation [poly (A)] signal were inserted under the CAG promoter. These plasmids (0.67 μg/each, total 2 μg) were transfected into HEK293T cells using the calcium phosphate-DNA coprecipitation method 47 . After 2 days of transfection, these cells were resuspended in PBS containing 10 mM (ethylenedinitrilo)tetraacetic acid. After centrifugation, the cells were washed with PBS, and then incubated with ZP-free eggs. After 30 minutes and then more than 6 hours of incubation, the attached and fused cell numbers were counted under a fluorescence microscope (BZ-X700, Keyence) and an inverted microscope with relief phase contrast (IX73, Olympus). Proteins were extracted from the remaining HEK293T cells with a lysis buffer containing Triton-X 100 [50 mM NaCl, 10 mM TrisꞏHCl, 1% (vol/vol) Triton-X 100 (Sigma Aldrich), pH 7.5] containing 1% (vol/vol) protease inhibitor mixture, and then used for western blotting and co-IP.

Immunocytochemistry.
After 3-hour incubation of mouse spermatozoa in TYH drops, the spermatozoa were washed with PBS. The spermatozoa suspended with PBS were smeared on a slide glass, and then dried on a hotplate. The samples were fixed with 1% PFA, followed by permeabilization with Triton-X 100. The spermatozoa were blocked with 10% goat serum (Gibco) for 1 hour, and then incubated with a mouse monoclonal antibody against HA tag (1:100) for 3 hours or overnight. After washing with PBS containing 0.05% (vol/vol) Tween 20, the samples were subjected to the mixture of a goat anti-mouse IgG Alexa Fluor 546 (1:300) and Alexa Fluor 488-conjugated Lectin PNA (1:2,000) for 1 hour. After washing again, the samples were sealed with Immu-Mount (Thermo Fisher Scientific) and then observed under a phase contrast microscope (BX-50, Olympus) with fluorescence equipment. Zebrafish spermatozoa were fixed with 3.7% formaldehyde immediately after collection at 4°C for 20 minutes. Spermatozoa were spun onto a SuperFrost Ultra Plus slide (Fisher Scientific) with a CytoSpin 4 (Thermo Fisher Scientific) at 1,000 rpm for 5 minutes followed by permeabilization with ice-cold methanol for 5 minutes. The slide was washed with PBS before blocking with 10% normal goat serum (Invitrogen) and 40 µg/mL BSA in PBST for 1 hour and then incubated with the mouse anti-zebrafish-Dcst2 antibody (1:25) overnight at 4°C. After washing with PBST, the slide was incubated with goat antimouse IgG Alexa Fluor 488 (1:100) for 2 hours before washing with PBST once more. After mounting using VECTASHIELD Antifade with DAPI (Vector Laboratories), spermatozoa were imaged with an Axio Imager.Z2 microscope (Zeiss) with an oil immersion 63x/1.4 Plan-Apochromat DIC objective.

Fertility assessment of adult zebrafish.
The evening prior to mating, the fish assessed for fertility and a TLAB wild-type fish of the opposite sex were separated in breeding cages. The next morning, the fish were allowed to mate. Eggs were collected and kept at 28°C in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4, 10 -5 % Methylene Blue). The rate of fertilization was assessed approximately 3 hours post-laying. By this time, fertilized embryos have developed to ~1000-cell stage embryos, while unfertilized eggs resemble one-cell stage embryos. Direct comparisons were made between siblings of different genotypes (wild-type, heterozygous mutant, homozygous mutant).

Collection of zebrafish eggs and spermatozoa.
Un-activated zebrafish eggs and spermatozoa were collected following standard procedures 48 . The evening prior to sperm collection, male and female zebrafish were separated in breeding cages (one male and one female per cage). To collect mature, un-activated eggs, female zebrafish were anesthetized using 0.1% w/v tricaine (25x stock solution in dH2O, buffered to pH 7.0-7.5 with 1 M Tris pH 9.0). After being gently dried on a paper towel, the female was transferred to a dry petri dish, and eggs were carefully expelled from the female by applying mild pressure on the fish belly with a finger and stroking from anterior to posterior. The eggs were separated from the female using a small paintbrush, and the female was transferred back to the breeding cage filled with fish water for recovery. To collect wild-type or mutant spermatozoa, male zebrafish were anesthetized using 0.1% tricaine. After being gently dried on a paper towel, the male fish was placed belly-up in a slit in a damp sponge under a stereomicroscope with a light source from above. Spermatozoa were collected into a glass capillary by mild suction while gentle pressure was applied to the fish's belly. Spermatozoa were stored in ice-cold Hank's saline (0.137 M NaCl, 5.4 mM KCl, 0.25 mM Na2HPO4, 1.3 mM CaCl2, 1 mM MgSO4, and 4.2 mM NaHCO3). The male was transferred back to the breeding cage containing fish water for recovery. For western blot analysis, spermatozoa from 3 males was sedimented at 800 x g for 5 minutes. The supernatant was carefully replaced with 25 µL RIPA buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM MgCl2, 1% NP-40, 0.5% sodium deoxycholate, 1X complete protease inhibitor (Roche)] including 1% SDS and 1 U/µL benzonase (Merck). After 10 minutes of incubation at RT, the lysate was mixed and sonicated 3 times for 15 seconds of 0.5-second pulses at 80% amplitude (UP100H, Hielscher) interspersed by cooling on ice.
Zebrafish sperm approach and binding assays. Imaging of zebrafish sperm approach Spermatozoa were squeezed from 2-4 wild-type and mutant male fish and kept in 150 µL Hank's saline containing 0.5 µM MitoTracker Deep Red FM (Molecular Probes) for >10 minutes on ice. Un-activated, mature eggs were obtained by squeezing a wild-type female. To prevent activation, eggs were kept in sorting medium (Leibovitz's medium, 0.5 % BSA, pH 9.0) at RT. The eggs were kept in place using a petri dish with cone-shaped agarose molds (1.5% agarose in sorting medium) filled with sorting medium. Imaging was performed with a LSM800 Examiner Z1 upright system (Zeiss) with a 20x/1.0 Plan-Apochromat water dipping objective. Before sperm addition, sorting media was removed and 1 mL of E3 medium was carefully added close to the egg. 5-10 µL of stained spermatozoa was added as close to the egg as possible during image acquisition. The resulting time-lapse movies were analyzed using FIJI.
Imaging and analysis of zebrafish sperm-egg binding Spermatozoa was squeezed from 2-4 wild-type and mutant male fish and kept in 100 µL Hank's saline + 0.5 µM MitoTracker Deep Red FM on ice. Un-activated, mature eggs were squeezed from a wild-type female fish and activated by addition of E3 medium. After 10 minutes, 1-2 eggs were manually dechorionated using forceps and transferred to a cone-shaped imaging dish with E3 medium. After focusing on the egg plasma membrane, the objective was briefly lifted to add 2-10 µL of stained spermatozoa (approximately 200,000-250,000 spermatozoa). Imaging was performed with a LSM800 Examiner Z1 upright system (Zeiss) using a 10x/0.3 Achroplan water dipping objective. Images were acquired until spermatozoa were no longer motile (5 minutes). To analyze sperm-egg binding, stably-bound spermatozoa were counted. Spermatozoa were counted as bound when they remained in the same position for at least 1 minute following a 90second activation and approach time window. Data was plotted as the number of spermatozoa bound per 100 µm of egg membrane for one minute.

Statistical analyses.
All values are shown as the mean ± SD of at least three independent experiments. Statistical analyses were performed using the two-tailed Student's t-test, Mann-Whitney U-test, and Steel-Dwass test after examining the normal distribution and variance ( Figure 2B and E, Figure 4C, Figure S4B, and Figure S5B). For zebrafish data, statistical analyses were performed in GraphPad Prism 7 software.

Data availability statement
RNA-seq data reported here (zebrafish adult tissues) were deposited at the Gene Expression Omnibus (GEO) and are available under GEO acquisition number GSE171906. The authors declare that the data that support the findings of this study are available from the corresponding authors upon request.

C) Localization of DCST2 in spermatozoa.
The HA-tagged DCST2 was localized in the anterior acrosome before the acrosome reaction, and then translocated to the equatorial segment in acrosome-reacted spermatozoa (arrows). PNA was used as a marker for the acrosome reaction. The fluorescence in the sperm tail was non-specific.

D) Co-IP and western blotting of the interaction between IZUMO1 and DCST1/2.
The TGC and sperm lysates from Ctrl, Dcst1;Tg, and Dcst2;Tg males were incubated with anti-HA tag antibody-conjugated magnetic beads, and then the eluted protein complex was subjected to western blotting. The HA-tagged DCST1 was detected only in the IP product from TGC, and the HA-tagged DCST2 was detected in the IP-product from TGC and spermatozoa. IZUMO1 was not detected in the co-IP products. Red and blue triangle marks show the expected molecular size of DCST1 (about 80 kDa) and DCST2 (about 77 kDa), respectively. E) Interaction between DCST1 and DCST2 in HEK293T cells. The protein lysate collected from HEK cells overexpressing Dcst1-3xFLAG and Dcst2-3xHA was incubated with anti-HA tag antibody-conjugated magnetic beads. The FLAG-tagged DCST1 was detected in the eluted protein complex. ADAM1B, a sperm protein that localizes to the sperm surface and is not involved in sperm-egg fusion, was used for negative control.

11
Branches supported by ultrafast bootstrap values (>=95%) were marked with a blue dot.

12
Sequence accessions were added next to the species names. for dcst2 (Table S4)