Ca2+ efflux via plasma membrane Ca2+-ATPase mediates chemotaxis in ascidian sperm

When a spermatozoon shows chemotactic behavior, transient [Ca2+]i increases in the spermatozoon are induced by an attractant gradient. The [Ca2+]i increase triggers a series of stereotypic responses of flagellar waveforms that comprise turning and straight-swimming. However, the molecular mechanism of [Ca2+]i modulation controlled by the attractants is not well defined. Here, we examined receptive mechanisms for the sperm attractant, SAAF, in the ascidian, Ciona intestinalis, and identified a plasma membrane Ca2+-ATPase (PMCA) as a SAAF-binding protein. PMCA is localized in sperm flagella membranes and seems to interact with SAAF through basic amino acids located in the second and third extracellular loops. ATPase activity of PMCA was enhanced by SAAF, and PMCA inhibitors, 5(6)-Carboxyeosin diacetate and Caloxin 2A1, inhibited chemotactic behavior of the sperm. Furthermore, Caloxin 2A1 seemed to inhibit efflux of [Ca2+]i in the sperm, and SAAF seemed to competitively reduce the effect of Caloxin 2A1. On the other hand, chemotactic behavior of the sperm was disordered not only at low-Ca2+, but also at high-Ca2+ conditions. Thus, PMCA is a potent candidate for the SAAF receptor, and direct control of Ca2+ efflux via PMCA is a fundamental mechanism to mediate chemotactic behavior in the ascidian spermatozoa.

for chemotactic response of the sperm, is mediated by the Ca 2+ channels 23 . Moreover, we have revealed that transient [Ca 2+ ] i increases in the spermatozoon (Ca 2+ bursts) and is induced in a SAAF gradient, triggering a series of stereotypic responses of flagellar waveforms that comprise turning and straight-swimming 24 .
In contrast to the knowledge on sperm attractants and the participation of Ca 2+ , the sperm receptors for the attractants are almost unknown, and only the attractant receptors of the sea urchin Arbacia punctulata and the starfish Asterias amurensis have been identified as guanylyl cyclases [26][27][28] . Therefore, the molecular mechanisms for the [Ca 2+ ] i modulation by attractants still remain unknown. In the case of the sea urchin and starfish, cGMP produced by the receptor guanylyl cyclase opens the K + -selective cGMP-gated channel, resulting in hyperpolarization 29 . This seems to activate the Na + /H + exchanger and leads to alkalization, resulting in Ca 2+ increase via the alkalization-gated channel CatSper 30 . The involvement of cGMP in regulating sperm function is not known in species other than the echinoderm. In the mammalian sperm, CatSper seems to act as a chemoattractant receptor 31,32 . In the ascidian, cGMP does not seem to be involved in chemotactic behavior 3 , and the SAAF receptor may not be the guanylyl cyclase. Thus, identification of the SAAF receptor is required to understand Ca 2+ signaling and the molecular mechanisms of ascidian sperm chemotaxis. Furthermore, in the case of any other species, Ca 2+ influx and [Ca 2+ ] i increases in the sperm cell are focused in chemotactic behavior, and Ca 2+ efflux and [Ca 2+ ] i decreases are scarcely examined; despite the need for prompt [Ca 2+ ] i decrease, it has been observed only in sperm activation of the sea urchin 33 .
In this study, we attempted to identify the SAAF receptor on the sperm of the ascidian, C. intestinalis. We show here that the PMCA located on the sperm tail binds to the SAAF and mediates chemotactic behavior in the ascidian. Furthermore, modification of the extracellular Ca 2+ concentration ([Ca 2+ ] ex ) disrupted the Ca 2+ bursts and the sperm chemotactic behavior. Thus, PMCA is a potential candidate for the SAAF receptor on the sperm. Regulation of the flagellar waveform in chemotactic behavior requires precise control of the Ca 2+ efflux.

Results
PMCA localized on sperm flagellar membranes acts as a target for SAAF. In order to identify a receptor for SAAF on the sperm of the ascidian, C. intestinalis, we attempted to discover SAAF-binding proteins. The SAAF-binding proteins were purified from the sperm membrane fraction by the pull-down assay using a resin conjugated with bio-SAAF, which is the biotinylated derivative of the C(4)-hydroxy group (Supplemental Fig. S1A,B). These were identified by the peptide-mass-fingerprint (PMF) method using MALDI-TOF/MS with the genome database of C. intestinalis (Ghost: http://ghost.zool.kyoto-u.ac.jp/cgi-bin/gb2/gbrowse/kh/) 34 . The most abundant SAAF-binding protein, the 370-kDa protein (Fig. 1A), was a product of the predicted gene model KH.C8.156, which is similar to the human PMCA3 (plasma membrane Ca 2+ /ATPase 3; ATP2B3) (Supplemental Fig. S1C). Another SAAF-binding protein, the 330-kDa protein was also identified as a product of KH.C8.156 (Supplemental Fig. S1), but the other proteins could not be identified by the PMF method. Thus, we concluded that PMCA is a potent candidate for the SAAF receptor. After searching through the genome database, we only found one PMCA gene (Atp2b) in C. intestinalis (KH.C8.156), which seems to diverge from the common ancestral gene of Atp2b1-4 (Fig. 1B). The SAAF-binding proteins were analyzed by a western blot assay with an anti-pan PMCA antibody (mAb 5F10). PMCA was detected as a 130 kDa band (Fig. 1A, arrow), and, as high molecular weight aggregates in a pull-down fraction by SAAF (Fig. 1A, asterisk), one of which was the 370 kDa protein (Fig. 1A, arrow head). We cloned and sequenced mRNAs of Atp2b from the testis cDNA and finally found two splice variants (Fig. 1C, Supplemental Fig. S2). There were differences in usage of the 6 th and 21 st exons, and, the main difference in the two transcripts was the C-terminus region after the CaM binding site (Fig. 1C). Unexpectedly, both variants (Atp2b-var.a: 133 kDa, Atp2b-var.b: 128 kDa) were different from the predicted product of KH.C8.156 (107 kDa) in the genome database (Fig. 1C). We checked the sequence in the genome database and found that there is a gap in the sequence between exon 17 and 19, resulting in an error of prediction in the product of KH.C8.156 (Supplemental Fig. S3). RT-PCR analysis showed that one of the splice variants (Atp2b-var.a) was expressed ubiquitously, and the other one (Atp2b-var.b) was highly expressed in the testis and weakly expressed in the stomach, intestine, and heart (Fig. 1D). Furthermore, western blot analysis with variant-specific anti-PMCA antibodies showed that only Atp2b-var.b was present in the sperm plasma membrane (Fig. 1E,F). Immunostaining experiments using the mAb 5F10 revealed that PMCA was present in the sperm flagella (Fig. 1G). Thus, we concluded that Atp2b-var.b is the dominant PMCA isoform in the sperm of C. intestinalis and is defined as the sperm PMCA. SAAF interacts with PMCA and enhances ATPase activity. In order to examine the molecular interaction between PMCA and SAAF, the membrane fraction of Sf9 cells expressing recombinant Atp2b-var.b was subjected to an interaction assay with SAAF using a quartz crystal microbalance (QCM) method. PMCA interacts with SAAF and the dissociation constant (K D ) of the interaction between SAAF and Atp2b-var.b was calculated as 240 ± 50 nM (Fig. 2). Next, we attempted to identify the SAAF-binding region in PMCA. The deduced amino acid sequence of Atp2b-var.b was subjected to the TMpred program (http://embnet.vital-it.ch/software/ TMPRED_form.html) 35 . Using this program, the membrane-spanning regions and the orientation of the protein were predicted (Fig. 3A, Supplemental Fig. S2). Because SAAF is a sulfate-conjugated polyhydroxysterol, it is highly hydrophilic and negatively charged. The interaction between SAAF and PMCA was proposed to be driven by electrostatic interactions between complementarily charged residues, the positively charged amino acids (basic amino acids) were replaced by neutral amino acids around the extracellular loops (ExLoops) of Atp2b-var.b (Fig. 3A,B). When these mutant-recombinant proteins were compared to the wild-type recombinant proteins, replacing W395A, R396A, and K409A around ExLoop2 (Atp2b-var.b-mut1) and R896A, K874A, and W880A around ExLoop3 (Atp2b-var.b-mut2) reduced the interaction with SAAF (Fig. 3C). On the other hand, replacing R951A, H956A, and H962A around ExLoop4 (Atp2b-var.b-mut3) did not affect the interaction between SAAF and PMCA (Fig. 3C). Thus, PMCA actually binds SAAF directly, and ExLoop2 and ExLoop3 are potent binding sites for SAAF. Because PMCA is a P-type primary ion transporter from the ATPase family, we attempted to assay PMCA activity by measuring its ATPase activity. In the assay, membrane fractions of the sperm and of the Atp2b-var.b-expressing Sf9 cells were pre-incubated, and then the ATPase activity was measured with and without SAAF and bio-SAAF. Both membranes had activity of an ATPase. However, SAAF increased the activity only in the membrane of Atp2b-var.b-expressing Sf9 cells, whereas bio-SAAF increased the activity of both membrane fractions (Fig. 4). Significant increase by bio-SAAF was observed in the membrane of Atp2b-var.b-expressing Sf9 cells (Fig. 4B).

PMCA mediates sperm chemotaxis towards SAAF. SAAF interacts with PMCA and affects its
ATPase activity, which could suggest that SAAF modulates Ca 2+ efflux via PMCA. Thus, we examined the role of PMCA in sperm chemotaxis. When the SAAF capillary was inserted into a sperm solution pre-treated with 5(6)-Carboxyeosin diacetate (CEDA), an inhibitor of PMCA via ATPase activity, most of the sperm lost their chemotactic behavior toward the capillary tip (Fig. 5A). Quantitative analysis of sperm chemotaxis using the linear equation chemotaxis index (LECI) 19 showed that pre-treatment with 10 µM CEDA partially blocked chemotaxis. The LECI was reduced to one-third that of the control sperm ( Fig. 5B). At 100 µM CEDA, sperm chemotaxis was completely blocked and the LECI became almost zero (Fig. 5B). Another PMCA inhibitor Caloxin 2A1, which selectively binds to ExLoop2 of human PMCA 36,37 also blocked chemotactic response, although, most of sperm was attached to the glass slide (Fig. 6). These results suggest that PMCA contributes to the regulatory mechanism of sperm chemotaxis.
Caloxin 2A1 inhibited the Ca 2+ pump of the ascidian PMCA in the absence of SAAF, and Caloxin 2A1 increased the [Ca 2+ ] i baseline in the sperm head in a dose dependent manner (Fig. 7A). If the sperm showed a chemotactic response toward SAAF, [Ca 2+ ] i during the Ca 2+ burst should have increased in the presence of Caloxin 2A1 (Fig. 7B,C, and Supplemental Fig. S8). These results suggest that PMCA really reduces the [Ca 2+ ] i in the sperm. On the other hand, even in the presence of Caloxin 2A1, the baseline [Ca 2+ ] i of some spermatozoa showed a response to SAAF, which was similar to that in the control (Fig. 7C). Even though number of spermatozoa showing higher intensity of [Ca 2+ ] i baseline (>1.5) was increased by Caloxin 2A1, 70% (1 mM Caloxin) and 50% (2 mM Caloxin) of the sperm still showed the same [Ca 2+ ] i baseline as the control (<1.5) (Supplemental Fig. S8). These results suggest that Caloxin 2A1 inhibits PMCA, and that SAAF activates PMCA competitively with Caloxin 2A1, resulting in maintenance of the [Ca 2+ ] i baseline.

Modification of [Ca 2+
] ex disrupts Ca 2+ bursts. Principal role of PMCA in many mammalian cells is maintenance of Ca 2+ homeostasis 14 . In the ascidian sperm, inhibition of PMCA resulted in an increase in the [Ca 2+ ] i level (Fig. 7). Chemotactic behavior of the spermatozoa as is mediated by Ca 2+ bursts in the sperm, and the Ca 2+ bursts require extracellular Ca 2+ 24 . Therefore, we examined the effects of the [Ca 2+ ] ex on chemotactic behavior. When the [Ca 2+ ] ex was lower than 10 mM, the chemotactic response of sperm was concentration-dependent ( Fig. 8A,B). At low [Ca 2+ ] ex (0.1 mM), changes in the [Ca 2+ ] i could not be observed. Interestingly, the chemotactic behavior also decreased when the concentration of [Ca 2+ ] ex was higher than 10 mM (Fig. 8A,B). Spermatozoa still demonstrated a chemotactic behavior like 'turning' at high [Ca 2+ ] ex (100 mM), but the timing of the response was disrupted (Fig. 8C,D). The 'turning' response is always observed at the initiation point of the Ca 2+ bursts 24 . In normal conditions (10 mM [Ca 2+ ] ex ), the Ca 2+ bursts occur at the Distal Phase 24 (Supplemental Fig. S9B). However, in the higher [Ca 2+ ] ex condition, initiation points of the Ca 2+ bursts become diverse. The Ca 2+ bursts are observed even in the Proximal Phase (Supplemental Fig. S9C). Duration of the Ca 2+ bursts and the [Ca 2+ ] i peak of the flagellum at high [Ca 2+ ] ex increased as compared to those in the 10 mM [Ca 2+ ] ex condition (Supplemental Table S1). Moreover, a decreasing rate of [Ca 2+ ] i in the flagella was suppressed at high [Ca 2+ ] ex conditions (Supplemental Table S1). Thus, high [Ca 2+ ] ex disrupts timing of the Ca 2+ bursts and potentiates them, resulting in a disruption of the chemotactic response of the sperm flagella. Pattern of the sperm trajectories in the high [Ca 2+ ] ex condition ( Fig. 8A) does not seem the same as that in the presence of the PMCA inhibitors (Figs 5 and 6). These results support the finding that the [Ca 2+ ] i increase during the chemotactic response is a passive phenomenon mediated by some Ca 2+ channels.

Discussion
In this study, we revealed that the sperm attractant seems to modulate Ca 2+ efflux via PMCA on the sperm plasma membrane in the ascidian, C. intestinalis, resulting in a chemotactic response. Thus, PMCA seems to act as a receptor for SAAF. There are four PMCA genes (Atp2B1 -4) with many splice variants in mammalian and probably in most vertebrate species (see Fig. 1B). Furthermore, PMCA is known as a housekeeping protein for dehydrogenase (Gapdh). The images were gathered from different gels for juxtaposing. Full-length gels are presented in Supplemental Fig. S4. (E,F) Expression of the protein Atp2b-var.a (E) and Atp2b-var.b (F) in the sperm membrane. Two columns on the left show immunostaining with anti-pan PMCA (5F10). Two columns on the right show immunostaining with anti-Atp2b-var.a (E) or with anti-Atp2b-var.b (F) antibodies. In the sperm, only Atp2b-var.b was expressed. Predicted molecular weights from deduced amino acid sequences of Atp2b-var.a and Atp2b-var.b are 133 kDa and 128 kDa, respectively. The images of (f) were gathered from different gels for juxtaposing. Full-length gels are presented in Supplemental Figs S5 and S6. Specificity of anti Atp2b-var.b antibody was shown in Supplemental Fig. S7. (G) Indirect immunofluorescence assay with the mAb 5F10 and Anti Atp2b-var.b antibody showed that PMCA was selectively localized at the sperm tail (left panels). Scale bar = 10 µm. We show here that there is only one PMCA gene in the ascidian, C. intestinalis, and that one of the splice variants, Atp2b-var.b, is specifically expressed on the sperm flagella. Even in mammals, some splice variants of Atp2B2 and Atp2B3 show restricted tissue distribution and seem to be involved in tissue-specific functions 39 . Moreover, PMCA4-deficient male mice are infertile due to a deficiency in sperm motility 17 . Thus, PMCA may be an important signaling molecule in the sperm of many animals.
In this study, we show that SAAF binds to PMCA and accelerates its ATPase activity. Furthermore, changing several amino acids around the extracellular loops 2 and 3 of PMCA result in a loss of SAAF binding. Therefore, the target molecule for SAAF is PMCA, and SAAF seems to induce Ca 2+ efflux. Interestingly, the ATPase activity of the recombinant PMCA was increased by both SAAF and bio-SAAF, but that of the sperm membrane was activated only by bio-SAAF. Previously we showed that bio-SAAF seemed to have a higher affinity to the sperm than SAAF 40 . Thus, bio-SAAF may bind and activate PMCA constitutively, and that is one reason why bio-SAAF worked as a bait for PMCA in this study. Because of its high affinity, bio-SAAF has no ability for sperm attraction 40 . Since sperm showing chemotactic response should be finely and locally controlled under various SAAF concentrations, the affinity between SAAF and PMCA should not be so high to sense the right concentration. We discuss the affinity between SAAF and PMCA below.
We showed here that two PMCA inhibitors, CEDA and Caloxin 2A1 suppress chemotactic behavior of the sperm, even though their effects differ. CEDA showed stronger inhibitory effects on the chemotactic response than Caloxin 2A1. Generally, PMCA has 10 transmembrane regions and 5 extracellular loops 13 . CEDA probably interacts with the ATP binding region at the intracellular loop between transmembrane 4 and 5, which is identical between the human PMCAs and the Ciona PMCA (see Supplemental Fig. S2), and inhibits ATPase activity 37,41 . Thus, effects of CEDA may inhibit other ATPases such as Na + /K + pump. On the other hand, Caloxins were developed as the selective peptide inhibitors for human PMCAs as they affect the PMCA activity extracellularly 36 . Caloxin 2A1 selectively binds to ExLoop2 of human PMCA 36,37 which is conserved between the human PMCAs and the Ciona PMCA (see Fig. 3A). As shown in the results, SAAF also binds to ExLoop2; thus, Caloxin 2A1 probably interacts with PMCA and competitively with SAAF. Since affinity of Caloxin 2A1 towards PMCA is relatively low (inhibition constant is 529 µM) 36,37 , Caloxin 2A1 could not inhibit SAAF completely. SAAF kept the [Ca 2+ ] i baseline of the sperm same as that of the control even in the presence of Caloxin 2A1 (Supplemental Fig. S8). Interestingly, the pattern of sperm trajectories in the presence of both PMCA inhibitors (Figs 5 and 6) does not seem to be the same as that in the high Ca 2+ conditions (Fig. 8A). These results show that PMCA is involved in chemotactic behavior of the Ciona sperm.
The receptor of sperm attractants has been described in echinoderm species. Specifically, the receptor for sperm attractants in the sea urchin is a transmembrane-type guanylyl cyclase 26 or a guanylyl cyclase-associated protein 42 . Furthermore, in the starfish, Asterias amurensis, it is also a guanylyl cyclase that acts like a receptor 27,28 . In these species, a guanylyl cyclase is activated by the sperm attractant and synthesizes cGMP, resulting in hyperpolarization 29,43,44 and alkalization of cytosol 30,45 . Finally, alkalization may open the sperm-specific Ca 2+ channel CatSper, resulting in a [Ca 2+ ] i increase 30 . In human sperm, progesterone, one of the potent candidates of the mammalian sperm attractant, seems to bind to the orphan enzyme α/β hydrolase domain-containing protein 2 (ABHD2) and mediates activity of CatSper by depletion of 2-arachidonoylglycerol 46 . Thus, the Ca 2+ channel CatSper seems to be linked with the receptor of the sperm attractants, and the [Ca 2+ ] i increase which is induced by the attractants controls chemotactic behavior of the sperm in these species. On the other hand, the present study shows that PMCA acts as the attractant receptor, and the Ca 2+ efflux induced by the attractant, mediates sperm chemotaxis in the ascidian. These models are contradictory, but our ascidian model fits our previous results. Addition of SAAF does not induce Ca 2+ increase, and Ciona sperm seems to sense decreases in SAAF concentration, thus resulting in Ca 2+ bursts 24,47 .
In the present study, we show that normal Ca 2+ bursts and chemotactic responses were observed only in the presence of 10 mM Ca 2+ . At low [Ca 2+ ] ex , no [Ca 2+ ] i change was observed and the chemotactic response was reduced. At high [Ca 2+ ] ex , the [Ca 2+ ] i increases were potentiated and the chemotactic response was disrupted (Fig. 8). This suggests that the [Ca 2+ ] i increase is a passive phenomenon, and that the ascidian tunes its mechanism for controlling sperm flagellar beating in to the seawater condition, which contains 10 mM Ca 2+ .
To examine the molecular mechanisms of sperm attraction, evaluating the binding affinity between the sperm attractant and its receptor is important. In the sea urchin, the binding affinity of the sperm attractant and its receptor seems to be very high. The K D was 0.19-15 pM in Lytechinus pictus 48 , and 90 ± 84 pM in Arbacia punctulata 49 . These K D values are almost similar to those of an antibody and its target molecule, that is, the sperm attractant of the sea urchin irreversibly binds to its receptor. In this case, the attractant may affect sperm movement when it binds to the sperm. In fact, guanylyl cyclase, the sperm attractant receptor of the sea urchin, A. punctulata, is highly dense on the sperm flagellum 49 , and it activates cGMP production and Ca 2+ fluctuations when it binds to the attractant 50 . On the other hand, in the ascidian, C. intestinalis, the K D value of interaction between the sperm attractant SAAF and PMCA was calculated in the present study and was found to be 240 ± 50 nM. This value is similar to the one calculated for calcium and its chelator or sensor, calmodulin (500 to 5000 nM) 51 , BAPTA (160 to 700 nM) 52 , fluo-4 (345 nM) 53 , etc. We previously estimated that the SAAF gradient around the capillary containing 1 µM SAAF ranges from 10 to 200 nM 24 . That is, the affinity between SAAF and PMCA is suitable for sensing the concentration of the sperm attractant SAAF. Probably, echinoderms and ascidians are using a completely different basic molecular design for sperm chemotaxis even though the regulation of flagellar beating by [Ca 2+ ] i is the same. These different results may provide new insights into the specificity and diversity of fertilization mechanisms.
We previously demonstrated that NCX, another molecule responsible for Ca 2+ efflux on the plasma membrane, exists on the ascidian sperm flagella and contributes towards the control of transient [Ca 2+ ] i in chemotactic behavior. Inhibition of NCX decreased the alteration of swimming-path curvature in the chemotactic behavior and diminished turning and straight-swimming 11  in the ascidian sperm. On the other hand, the capacitative Ca 2+ entry, which is mediated by the store-operated Ca 2+ channel, seems to mediate flagellar movements to establish the chemotactic behavior of the ascidian sperm 23 . The capacitative Ca 2+ entry is induced by depletion of internal Ca 2+ stores and lasts for a few minutes.
Here, we propose a new working hypothesis showing the molecular mechanism of sperm chemotaxis in the ascidian C. intestinalis (Fig. 9) ] i reaches low levels (refractory phase). It is still unknown whether Orai1 and Stim1, which compose the store-operated Ca 2+ channels, are involved in the sperm chemotaxis system. On the other hand, CatSper, which is the Ca 2+ channel having an important role in mammalian sperm, also exists in the ascidian 54 . Function of CatSper in the ascidian sperm is still unknown. CatSper might act as the Ca 2+ channel in the hypothesis instead of the store-operated Ca 2+ channel. Furthermore, difference of amino-acid sequences between the two splice variants of PMCA was only found in the C-terminus region (see Supplemental Fig. S2). The C-terminus tail of the mammalian PMCA contains a PDZ-binding domain and it interacts with many other proteins 13 . Because none of the ascidian PMCA variants contain a PDZ-domain, the C-terminus region of the ascidian PMCA may have a role in a novel signaling cascade. Further studies on the roles of PMCA and Ca 2+ channels will provide new insights into the mechanisms and functions of [Ca 2+ ] i changes.

Preparation of the SAAF-resin.
Bio-SAAF was immobilized on streptavidin-coated beads by using the affinity of biotin for streptavidin (TetraLink ™ Tetrameric Avidin Resin; Promega Japan, Tokyo, Japan). Binding capacity of the beads was 30 nmol of biotin/mL. In general, 150 nmol of bio-SAAF was added to 10 mL of 50% slurry of streptavidin-coated beads. The beads were previously obtained by washing with 10 mL of PBS three times and incubating for 30 min at room temperature on a rotary mixer at low speed. The unbound material was carefully removed, and the beads were washed with 10 mL of PBS three times and immediately used or stored at 4 °C. For simplicity, the SAAF bound to the beads is referred to as the SAAF-resin and the beads bound with biotin as the control resin.       were ASEENSTENNQTQNSVA and ARHHSSHHAHLEPV, respectively. For detecting the anti-Atp2b-var.a and Atp2b-var.b antibodies, HRP-conjugated anti-rabbit IgG (Bethyl Laboratories; Montgomery, TX, USA), and ECL Prime (GE) were used. Luminescence was detected using an X-ray film (Fujifilm, Tokyo Japan) and images were acquired using a scanner GT-X770 (Epson, Suwa, Japan) (Fig. 1A,E) or imaged by a gel imager (EzCapture II; ATTO) (Figs 1F and 3). Acquired images were processed by Adobe Photoshop and Adobe Illustrator (Adobe Systems). Specificity of the anti-Atp2b-var.b was shown in Supplemental Fig. S7. For immunostaining, the sperm suspension was put onto a coverslip and fixed with 4% paraformaldehyde for 10 min, after which it was washed twice with ASW for 5 min. The fixed sperm was permeabilized with 0.1% NP-40 for 15 min, blocked with 1% BSA for 1 h and incubated with 5 µg/mL 5F10 or x1/500 anti-Atp2b-var.b antibody with 1% BSA in PBS for 45 min. The sperm was washed 3 times with PB and incubated with x1/2000 Alexa 488-conjugated anti-mouse IgG(H + L) (ThermoThermo Fisher Scientific) or x1/2000 DyLight550-conjugated anti-rabbit IgG (Abcam) for 30 min, respectively. After washing two times with PBS, the sperm was observed using a fluorescent microscope (IX-71; Olympus). Images were recorded on a PC connected to a digital CCD camera (Retiga Exi; QImaging, Surrey, Canada) using an imaging application (TI workbench) 57 . Acquired data was processed by Adobe Photoshop and Adobe Illustrator (Adobe Systems).

Molecular interaction assay.
Interaction between SAAF and PMCA was investigated by a highly sensitive 30-MHz quartz crystal microbalance (QCM; NAPiCOS; Nihon Dempa Kogyo, Tokyo, Japan). The first channel of a sensor chip was coated with PMCA-expressing Sf9 membranes (5 mg/mL of proteins) suspended in PBS (pH 7.4) containing 0.1% of NP-40. The second channel of the same sensor chip was coated with Sf9 membranes (5 mg/mL of proteins), as a reference. The sensor chip was washed three times with PBS, placed into the chamber, perfused with PBS until the frequency was stabilized. Then, the sensor chip was perfused with SAAF (125 µM) dissolved in PBS, and the change in frequency was recorded. In order to examine SAAF and mutant PMCA interactions, the first channel of a sensor chip was coated with mutant PMCA-expressing Sf9 membranes. The second channel on the same sensor chip was coated with PMCA-expressing Sf9 membranes, as a reference. After washing, perfusion, and blocking the sensor chip, it was perfused with SAAF, and the change in frequency was recorded. All experiments were performed at 20 °C, with a flow rate of 50 μL/min. To evaluate the binding affinity, changes in the frequency of cumulative perfusion were analyzed. Affinities were evaluated by the Michaelis-Menten equation, and fitted curves were obtained by calculating the average of three experimental values. The dissociation constant (K D ) was calculated with NAPiCOS Analysis software (Nihon Dempa Kogyo).
Assay for ATPase activity of PMCA. Enzyme activity of PMCA from Ciona sperm and Sf9 membranes was determined by measuring the inorganic phosphate produced using a coupled enzyme assay kit (EnzChek ® Phosphate Assay Kit, Thermo Fisher Scientific) according to the manufacturers' instructions. In brief, membranes were incubated for 5 min at 25 °C in a reaction mixture containing 50 mM Tris-HCl, pH 7.4, 1 mM MgCl 2 , 0.2 mM Figure 9. Working hypothesis of the function of SAAF on the sperm plasma membrane. When SAAF concentration increases, PMCA may be continuously activated by SAAF, and Ca 2+ may be kept at low levels, though Ca 2+ may enter through Ca 2+ channel (Ascending Phase). When SAAF concentration decreases, SAAF may detach from PMCA and become inactivated, resulting in [Ca 2+ ] i increase (descending phase). Intracellular Ca 2+ is excreted by NCX and PMCA and finally [Ca 2+ ] i reaches low levels (refractory phase). Pair of yellow ovals show Ca 2+ channel, Red oval: PMCA, Green oval: NCX, Red rectangle: SAAF. Blue arrows indicate flow of Ca 2+ .
2-Amino-6-mercapto-7-methylpurine riboside (MESG) as a substrate, 1 U/mL purine nucleoside phosphorylase (PNP) as a converting enzyme, and 100 µM CaCl 2 . The enzymatic reaction was initiated by addition of 1 mM ATP. ATPase activity was calculated based on the difference in ATP hydrolysis between samples incubated in the presence and absence of 10 µM SAAF or bio-SAAF. Enzymatic conversion of MESG by PNP was measured for 5 min at room temperature at a wavelength of 360 nm in a recording spectrophotometer (Ultrospec2100pro, Amersham Biosciences).
Analysis of sperm chemotaxis. Sperm chemotaxis was examined as described previously 24 . Briefly, semen was diluted 10 4 to 10 5 times in the medium for experiment (e.g. ASW containing CEDA or Caloxin 2A1; high and low Ca 2+ -ASW). Theophylline (Sigma-Aldrich Japan, Tokyo, Japan) was added to the suspension in a final concentration of 1 mM for the induction of motility activation which is mediated by an increase in intracellular cAMP 3 , except in a series of Caloxin 2A1 experiments. Series of Caloxin 2A1 experiments including control were performed with the sperm suspension not containing theophylline, since the Caloxin 2A1-treated theophylline-activated sperm completely adhered to the glass slide and there was no free-swimming sperm. The activated-sperm suspension was placed in the observation chamber, and sperm movement around the micropipette tip containing SAAF was recorded. The position of the sperm head was analyzed with Bohboh software (BohbohSoft, Tokyo, Japan) 58 . The parameters of chemotactic activity (trajectory, the distance between the capillary tip and the sperm, and LECI) were calculated as described previously 19 .

Imaging analysis of [Ca 2+
] i in the flagella of swimming sperm. For Ca 2+ imaging, Fluo-8H AM ( Fig. 7) or Fluo-4 AM (Fig. 8) was used as a fluorescent probe. The dye-loaded sperm was prepared as described previously 24 . Fluorescent images of the sperm were observed by a microscope (IX71, Olympus, Tokyo, Japan), and captured on a PC connected to a digital CCD camera (ImagEM, C9100-13; Hamamatsu Photonics, Hamamatsu, Japan) at 32.5 frames/sec using Aquacosmos (Hamamatsu Photonics), or a digital CCD camera (Retiga Exi; QImaging) at 50 frames/sec using an imaging application (TI workbench 57 ), as described previously 24,59 . For fluorescence illumination, a stroboscopic lighting system with a power LED was used as described 24 . Fluorescent signal intensity and sperm flagellar bending were also analyzed using the Bohboh software 58 .
Statistical analysis. All experiments were repeated at least three times with different specimens. Data is expressed as the mean ± SD. Statistical significance against control (Figs 4B, 5B, 6B and 7), or the normal SW (Fig. 8B) was calculated using the Student's t-test; P < 0.05 was considered significant.