Molecular interplay between sperm and oocyte: a narrative review

Abstract

BACKGROUND

Fertilization ensures the transmission of genetic material across generations through a series of precisely coordinated physiological and molecular events. To fertilize an oocyte, a spermatozoon must pass through the cumulus cell layer, penetrate the zona pellucida (ZP), and ultimately adhere to and fuse with the oolemma (the oocyte plasma membrane). Upon fusion, the oocyte initiates mechanisms to block additional sperm entry at both the ZP and oolemma. These processes are highly dynamic in space and time, posing substantial technical barriers to their mechanistic dissection. Nonetheless, recent in silico, in vitro, and in vivo studies have begun to elucidate how intricate networks of intracellular signaling cascades and extracellular protein–protein interactions orchestrate successful fertilization in vertebrates. However, the extent to which these findings accurately reflect the biology of human sperm–oocyte interactions remains obscure, owing to ethical constraints on human gamete experimentation and the limited availability of patients harboring pathogenic variants in fertilization-related genes.

OBJECTIVE AND RATIONALE

This narrative review synthesizes current knowledge of the molecular determinants governing mammalian sperm–oocyte interactions, summarizes relevant genetic anomalies identified in infertile patients, and discusses emerging experimental approaches for the direct investigation of human fertilization. We also explore how recent mechanistic insights and technological innovations may inform the diagnosis and treatment of fertilization disorders and guide the development of novel contraceptive strategies.

SEARCH METHODS

We searched PubMed, Google Scholar, and Scopus to identify research and review articles published in English. Studies limited to non-mammalian species and non-peer-reviewed preprints were excluded. Searches used terms related to fertilization, sperm–oocyte interactions, ZP, cumulus cells, polyspermy block, and human infertility, alone or in combination. Additional searches targeted key proteins and emerging technologies relevant to mammalian fertilization, clinical diagnostics, and contraceptive development.

OUTCOMES

Our mechanistic understanding of mammalian gamete interactions has predominantly stemmed from in vitro and in vivo animal studies, which have revealed key molecular processes, such as sperm hyaluronidase-mediated cumulus matrix dispersal, translocation of sperm acrosomal membrane proteins to enable oolemma interaction, and ZP glycoprotein cleavage underlying the polyspermy block. While studies in model species remain indispensable, translating this knowledge to human biology requires meticulous validation. The integration of interdisciplinary approaches, such as humanized mouse models, artificial human oocytes, xenospecies fertilization assays, antibody inhibition studies, and high-throughput interactome screening, offers promising avenues to clarify interspecies discrepancies and generate insights more directly relevant to human gamete interactions.

WIDER IMPLICATIONS

Advances in the mechanistic dissection of sperm–oocyte interactions are anticipated to support the development of diagnostic tools and therapeutic interventions for infertility caused by defective fertilization. In parallel, these discoveries may enable the rational design of safe, reversible contraceptives that selectively block gamete interactions without compromising other physiological processes.

REGISTRATION NUMBER

N/A.

Graphical Abstract

Progressive stages and mechanistic studies of mammalian sperm–oocyte interactions. ZP, zona pellucida.

Introduction

The exploration of human reproductive biology can be traced back to ancient Greece, where Aristotle proposed that during sexual reproduction, the male contributes the ‘form’, the blueprint for the offspring, while the female provides the ‘matter’, the raw materials for a new individual. This theory dominated scientific thought for nearly two millennia, until the microscopic identification of mammalian sperm and oocytes by Antonie van Leeuwenhoek and Karl Ernst von Baer, respectively (reviewed by Lopata, 2009; Cobb, 2012; Andrade-Rocha, 2017). However, the true nature of fertilization remained poorly understood until the late 19th century, when Oscar Hertwig observed the entry of a sperm into an oocyte in sea urchins, providing the first definitive evidence that male and female gametes contribute equally to the formation of a progeny organism (reviewed in Briggs and Wessel, 2006). In the mid-20th century, advances in cell culture and microscopy enabled more sophisticated investigations into mammalian fertilization, ultimately leading to the development of IVF: one of the most important ARTs to date (Chang, 1959; Yanagimachi and Chang, 1963; Edwards et al., 1969; Steptoe and Edwards, 1978). These breakthroughs collectively opened a new era of reproductive medicine and bioengineering, laying the foundation for contemporary infertility treatments and mammalian genome manipulation.

Building upon these conceptual and technological developments, researchers have since elucidated the cellular choreography of fertilization, and have begun to uncover the molecular mechanisms largely through studies in model vertebrates (e.g. ungulates, rodents, and fish). During natural conception, sperm deposited in the vagina travel distances thousands of times their own length, traversing the cervix and uterus to reach the Fallopian tubes, where fertilization takes place (Fig. 1). To successfully fertilize an oocyte, a spermatozoon must penetrate the cumulus cell layer and the zona pellucida (ZP) surrounding the oocyte and then bind to and fuse with the oocyte plasma membrane (hereafter referred to as the oolemma; Fig. 1). This multistep process demands a precisely timed, spatially coordinated, and molecularly stringent interplay between sperm and oocyte, rendering fertilization one of the most intricate and underexplored forms of cell–cell interaction in biology.

Figure 1.

Sperm–oocyte interactions. Successful natural conception requires synergized sperm deposition (1) and ovulation (2). Following ejaculation, sperm ascend through the cervix and traverse the uterotubal junction to reach the Fallopian tubes, where they undergo capacitation (3). As sperm approach the oocyte, they interact with and pass through the cumulus cell layer (4), bind to and penetrate the zona pellucida (ZP) to enter the perivitelline space (5), and ultimately fuse with the oolemma (6), thereby triggering mechanisms that block polyspermy (7). This narrative review mainly focuses on steps 4–7, during which an intricate cascade of molecular interactions takes place between sperm and oocyte.

Sperm are maintained in a functionally quiescent state prior to ejaculation and must undergo ‘capacitation’, a series of biochemical and physiological changes that confer fertilization competence, during their journey through the female reproductive tract (Austin, 1951, 1952; Chang, 1951; Austin and Bishop, 1958). After insemination into the vagina, semen rapidly liquefies, usually within 15–30 min in humans, to transform from a coagulated mass into a watery fluid that allows sperm to swim out of seminal plasma rich in decapacitation factors such as CRISP1, NYD-SP27 (an isoform of PLCζ1), SERPINE2, semenogelins (SEMG1 and SEMG2), and SPINK3 (de Lamirande et al., 2001; Roberts et al., 2003; Nixon et al., 2006; Da Ros et al., 2008; Bi et al., 2009; Lu et al., 2011; Yamasaki et al., 2017; Weigel Muñoz et al., 2019). As sperm ascend through the uterus toward the Fallopian tubes, they encounter luminal components, including albumin and high-density lipoproteins, that gradually strip residual decapacitation factors (e.g. cholesterol, SERPINE2, SPINK3, and SPINKL) from the sperm plasma membrane (Visconti et al., 1999; Lin et al., 2008; Lu et al., 2011; Zalazar et al., 2020). Concurrently, an elevated concentration of bicarbonate (${\text{HCO}}_3^{-}$) in the female tract activates soluble adenylyl cyclase, leading to a rise in cAMP levels (Chen et al., 2000). This, in turn, stimulates protein kinase A activity, initiating a cascade of phosphorylation events, especially on tyrosine residues (Visconti et al., 1995). The resulting proteomic changes culminate in sperm hyperactivation, characterized by high-amplitude, asymmetric flagellar beating, and a non-linear swimming trajectory, in contrast to the symmetric, progressive motion of non-capacitated sperm (Suarez, 2008).

During capacitation, the depletion of cholesterol destabilizes the sperm plasma membrane, making sperm more susceptible to eliciting the acrosome reaction (AR), in which the acrosome, a membranous organelle encapsulating the anterior portion of the sperm nucleus, fuses with the overlying plasma membrane at multiple points, releasing hydrolytic enzymes and exposing acrosomal membrane proteins critical for subsequent interactions with the oocyte (Cross, 1996; Khorasani et al., 2000; Cohen et al., 2016). In human sperm, the AR can be induced in vitro by various physiological stimuli, including steroids (e.g. progesterone), prostaglandins, and ZP glycoproteins (e.g. ZP3 and ZP4), as well as pharmacological agonists such as calcium ionophores (e.g. A23187 and ionomycin) (Aitken et al., 1993; Schaefer et al., 1998; Shimizu et al., 1998; Bronson et al., 1999; Chiu et al., 2008a; Sánchez-Cárdenas et al., 2014). Among these, progesterone has emerged as a potent inducer by activating a sperm-specific calcium channel, CatSper, which mediates a rapid calcium influx required for the AR (Uhler et al., 1992; Turner et al., 1994; Smith et al., 2012). Nevertheless, the precise timing and location of the AR in vivo, whether it occurs before sperm approach the oocyte, during transit through the cumulus matrix, or upon ZP binding, is not completely defined.

The temporal regulation of capacitation and the AR is believed to be synchronized with ovulation, although the exact dynamics remain poorly understood. Upon arrival at the isthmic region of the Fallopian tube, non-capacitated, acrosome-intact sperm bind to the local epithelial cells, forming a transient reservoir that preserves sperm viability for up to 5 days (Aitken and Nixon, 2013). Sperm are released from this reservoir upon acquiring hyperactivated motility and are subsequently guided by rheotaxis, thermotaxis, and chemotaxis to the oocyte, which is ovulated at the metaphase II stage of meiosis, a state of full maturation for fertilization (Giojalas and Guidobaldi, 2020; Eisenbach, 2025). The orchestrated release, activation, and redirection of sperm toward the oocyte exemplify the exquisite spatiotemporal control that underlies successful mammalian fertilization.

Mammalian gametes share fundamental cellular architectures and conserved fertilization processes. Therefore, in this narrative review, we focus on the molecular mechanisms surrounding mammalian sperm–oocyte interactions, emerging methodologies for directly dissecting these processes in humans, and recent discoveries in basic science which inform advanced infertility diagnosis and treatment, as well as the rational design of novel reversible contraceptive strategies.

Methods

We conducted comprehensive literature searches using PubMed, Google Scholar, and Scopus to identify research and review articles published in English between May 1946 and February 2026. Studies focusing exclusively on non-mammalian species and non-peer-reviewed preprints were excluded. Search terms, such as ‘fertilization’, ‘sperm’, ‘oocytes’, ‘sperm–oocyte interactions’, ‘cumulus cells’, ‘zona pellucida’, ‘polyspermy block’, ‘oolemma’, ‘acrosome’, and ‘human infertility’, were used either individually or in combination. Additional searches targeted specific proteins implicated in mammalian sperm–oocyte interactions, as well as emerging technologies with potential to advance mechanistic studies, clinical diagnostics, or contraceptive development. Titles and abstracts of retrieved articles were screened for relevance, and shortlisted articles underwent full-text assessment. Reference lists of included publications were also examined to identify additional related studies not captured by the initial keyword searches. Finally, all citations were cross-checked against the Crossref Retraction Watch Database to ensure that retracted articles or those subject to expressions of concern were excluded.

Molecular basis of sperm–cumulus interaction

The female reproductive tract constitutes a rigorous selection arena for sperm. Among the hundreds of millions of sperm ejaculated during intercourse, only a small fraction, typically no more than a few hundred, can successfully reach the vicinity of the oocyte in the Fallopian tubes (Suarez and Pacey, 2006; Suarez and Wu, 2017). These sperm must then overcome three additional barriers, namely the cumulus cell layer, the ZP, and the oolemma, to fuse with the oocyte.

Formation and structure of the cumulus oophorus

The outermost barrier encountered by sperm is the cumulus cell layer, or cumulus oophorus, a dense cluster of several thousand somatic cells derived from follicular granulosa cells. During folliculogenesis, granulosa cells encase the oocyte, providing structural support and creating a specialized microenvironment optimal for oocyte growth and maturation (Nagyová et al., 2021; Turathum et al., 2021; Wang, 2025). Granulosa cells emanate thin cytoplasmic extensions called transzonal projections, which traverse the ZP and establish contacts with the oolemma through gap junctions, adherens junctions, and tight junctions (Mora et al., 2012; Xie et al., 2023). These structures facilitate bidirectional exchange of metabolic substrates and signaling molecules, thereby coordinating the development of both somatic and germ cells (Li and Albertini, 2013; Clarke, 2022).

In response to FSH secreted by the anterior pituitary gland during the follicular phase of the menstrual cycle, granulosa cells proliferate and differentiate, forming a fluid-filled antrum (Barbieri, 2014; Sugiura et al., 2023). As the follicle matures, granulosa cells segregate into two spatially distinct subpopulations: mural granulosa (MG) cells, which line the follicle wall and are separated from the adjacent theca cell layers by a basement membrane, and cumulus granulosa (CG) cells, which remain tightly associated with the oocyte (Khamsi and Roberge, 2001; Shirafuta et al., 2024). Despite their anatomical differences, MG and CG cells function in a coordinated manner to regulate meiotic arrest and resumption of the growing oocyte, thereby synchronizing oocyte maturation with ovulation (Jaffe and Egbert, 2017; Gao et al., 2022). By the late follicular phase, elevated estradiol levels produced by MG cells trigger a surge of LH from the anterior pituitary (Arroyo et al., 2020). LH, in turn, upregulates HAS2 expression in CG cells, promoting the synthesis of high-molecular-weight hyaluronic acid (or hyaluronan) (Salustri et al., 1999; Zhuo and Kimata, 2001; Yung et al., 2019). Secreted hyaluronan binds extracellularly to its principal receptor CD44 on CG cells, while TNFAIP6 (or TSG6), PTX3, inter-α-trypsin inhibitor (IαI), and Versican crosslink hyaluronan chains into larger complexes, aggregating CG cells into a mucoelastic extracellular matrix (Ohta et al., 1999; Russell et al., 2003; Salustri et al., 2004; Di Giacomo et al., 2016). Concomitantly, LH, together with signaling molecules from both granulosa cells and the oocyte (e.g. EGF-like growth factors, GDF9 and BMP15), induces prominent volumetric expansion of CG cells (Kawashima et al., 2012). The oocyte, surrounded by multiple layers of CG cells, is then dislodged from the follicle wall, forming a cumulus–oocyte complex (COC) floating freely within the follicular fluid. Finally, extracellular matrix remodeling by proteolytic enzymes (e.g. matrix metalloproteinases and plasminogen activator) within the follicle wall, together with progesterone-dependent follicle contraction (as recently demonstrated in mice), results in follicle rupture and release of the COC into the Fallopian tubes (McIntush and Smith, 1998; Thomas et al., 2024).

Given the intimate crosstalk between cumulus cells and the oocyte, correlations have been observed between the transcriptome, proteome, metabolome, structure, and functionality of cumulus cells, and the quality and fertilization potential of oocytes (Assidi et al., 2011; Ishikawa-Yamauchi et al., 2024; Massoud et al., 2024; Morimoto et al., 2024), making cumulus cells valuable biomarkers for the clinical assessment of female fertility. Notably, transplantation of aged mouse oocytes into follicles from young mice can rejuvenate oocyte function (Wang et al., 2024b), suggesting a promising cell-based therapeutic strategy for restoring oocyte quality, which complements existing pharmacological interventions and dietary supplements (Wang, 2025).

COC-mediated sperm chemoattraction and activation

After passing through the uterotubal junction, sperm are guided over long distances along the oviduct toward the site of fertilization via two major mechanisms: rheotaxis, in which sperm swim against the direction of oviductal fluid flow, and thermotaxis, whereby sperm migrate up a temperature gradient from the oviductal isthmus toward the ampulla (reviewed by Giojalas and Guidobaldi, 2020; Eisenbach, 2025). Once sperm arrive at the ampulla, they are more precisely navigated toward the oocyte and undergo final functional maturation in response to chemical cues secreted by the COC (Fig. 2A).

Figure 2.

Sperm–cumulus interaction. (A) In the Fallopian tubes, sperm are guided toward the oocyte by chemoattractants released from the cumulus–oocyte complex (COC), such as DEFB119, CCL20, and progesterone. DEFB119 is produced by both the COC and the oviductal epithelium, whereas CCL20 is primarily derived from the COC. Progesterone, in contrast, is predominantly secreted by cumulus cells. SPAM1 is a representative sperm acrosomal protein that dissociates cumulus cells by hydrolyzing hyaluronan in the cumulus matrix. (B) The COC- and oviduct-derived molecular cues are able to trigger sperm hyperactivation, chemotaxis, and the acrosome reaction in vitro by activating CatSper-mediated calcium signaling. Δ2AG, depletion of 2-arachidonoylglycerol; Ca²⁺, calcium ions; [Ca²⁺]_i, intracellular calcium concentration; P4, progesterone. (C) The cumulus matrix is composed of long hyaluronan chains crosslinked by hyaluronan-binding proteins such as TNFAIP6 and PTX3. Sperm penetrate the cumulus oophorus by digesting its hyaluronan-rich extracellular matrix with sperm-borne hyaluronidases such as SPAM1. GlcA, D-glucuronic acid; GlcNAc, N-acetyl-D-glucosamine.

The chemotactic behavior of human spermatozoa was first demonstrated in vitro, where sperm altered their swimming trajectories upon exposure to human follicular fluid (Ralt et al., 1991; Jeon et al., 2001). However, follicular fluid is unlikely the principal chemoattractant in vivo, as only a minute volume is released with the COC at ovulation, and effective chemoattraction requires a sustained and spatially stable chemical gradient surrounding the COC over the 12- to 24-h window of human fertilization (Eisenbach, 2025). Instead, culture media conditioned by COCs, isolated cumulus cells, or cumulus-free oocytes have all been shown to induce chemotactic responses in human sperm, suggesting that chemoattractants are produced by both cumulus cells and the oocyte (Sun et al., 2005).

Progesterone, continuously synthesized and secreted by cumulus cells following ovulation, is a well-characterized chemoattractant for human sperm (Chian et al., 1999; Oren-Benaroya et al., 2008). Cumulus cells are densely packed near the ZP and more loosely arranged at the periphery. This unique radial architecture, coupled with the highly viscous extracellular matrix, facilitates the formation of a stable progesterone gradient within the cumulus oophorus (Villanueva-Díaz et al., 1995; Wang et al., 2001; Teves et al., 2006). In vitro, human sperm exhibit prominent chemoattraction toward picomolar concentrations of progesterone (1–100 pmol/l), displaying a uniformly oriented swimming trajectory toward the source. Notably, depletion of progesterone from cumulus cell-conditioned media using anti-progesterone antibody-conjugated magnetic beads completely abolishes sperm chemotaxis, demonstrating progesterone as the dominant chemoattractant derived from cumulus cells (Oren-Benaroya et al., 2008). Such picomolar progesterone gradients have also been exploited for selecting functional spermatozoa from subfertile semen samples (Gatica et al., 2013).

The precise molecular mechanisms by which sperm detect such low concentrations of progesterone remain incompletely understood (Eisenbach, 2025). Unlike its canonical genomic actions in somatic cells, where progesterone binds nuclear receptors to initiate transcriptional regulation, it functions in sperm via a non-genomic mechanism (Calogero et al., 2000), as transcription is largely inactive in sperm due to the inaccessibility of highly condensed chromatin. In mice, only acrosome-intact spermatozoa, but not acrosome-reacted ones, respond to progesterone, indicating that chemotactic signaling occurs prior to the AR and is likely modulated by receptors on the sperm plasma membrane (Guidobaldi et al., 2017b). The sperm-specific cation channel, CatSper, localized to the principal piece of the flagellum, has been proposed as a key component of this mechanism (Publicover and Barratt, 2011; Smith et al., 2012; Rahban and Nef, 2020). Progesterone does not directly bind CatSper; instead, it interacts with α/β hydrolase domain-containing protein 2 (ABHD2), which hydrolyzes an endogenous CatSper inhibitor, 2-arachidonoylglycerol, at the sperm plasma membrane (Miller et al., 2016). The resulting disinhibition of CatSper leads to an intracellular calcium rise, essential for hyperactivation and the AR (Gomez-Olivieri et al., 2025). Thus, the progesterone–ABHD2–CatSper axis represents an important molecular pathway by which cumulus cell-derived steroid triggers sperm functional activation at fertilization (Fig. 2A and B). However, whether this signaling cascade also mediates directional steering of spermatozoa during chemotaxis remains an open question (Rahban and Nef, 2020).

More recently, the human β-defensin DEFB119 (orthologous to mouse DEFB19) has been identified as another physiological sperm chemoattractant produced by both the oviductal epithelium and the COC (Li et al., 2022). This β-defensin forms an ascending concentration gradient from the uterus to the oviduct and elicits sperm chemotactic responses both in vitro and in vivo, supporting a multistage chemotactic model where DEFB19/119 mediates long-range navigation of sperm along the female tract and synergizes with progesterone for short-range attraction near the oocyte. Female mice lacking DEFB19 are subfertile; they exhibit reduced sperm numbers in their oviducts after pairing with wild-type males. Intriguingly, DEFB19/119 also activates CatSper to induce calcium influx through an unknown mechanism (Li et al., 2022).

The COC-derived chemokine CCL20 and its receptor CCR6, a G-protein coupled receptor on the sperm surface, constitute another ligand–receptor pair implicated in human sperm chemotaxis (Caballero-Campo et al., 2014; Duan et al., 2020). Exposure of capacitated human sperm to recombinant CCL20 induces directed swimming, which can be abrogated by neutralizing antibodies against CCL20 (Duan et al., 2020). Interestingly, DEFB1, a β-defensin produced by both the male genital epithelium and the uterine epithelium, also binds CCR6 (Diao et al., 2014), suggesting that CCR6 is a shared receptor for both female- and male-derived ligands. CCR6 co-localizes with CatSper in sperm flagella, and ligand binding to CCR6 triggers CatSper-dependent calcium influx (Diao et al., 2017), indicating its potential role as a proximal sensor that transduces extracellular chemotactic cues to the CatSper channel.

It has also been proposed that female-derived odorants and their olfactory receptors in sperm contribute to chemotaxis. hOR17-4, a G-protein-coupled olfactory receptor first discovered in human testicular tissue, mediates responses to bourgeonal, a synthetic odorant mimicking physiological chemoattractants (Spehr et al., 2003). Bourgeonal stimulates chemotactic responses and calcium influx in human sperm; its interaction with hOR17-4 activates membranous adenylyl cyclase, which converts ATP to cAMP that subsequently leads to CatSper activation (Spehr et al., 2003, 2004). The bourgeonal-mediated sperm chemotaxis and calcium entry can be blocked by hOR17-4 antagonists or adenylyl cyclase inhibitors (Spehr et al., 2004). In concert with these findings, men with unexplained infertility display both reduced nasal sensitivity and impaired sperm chemoattraction to bourgeonal (Ottaviano et al., 2013), supporting a physiological role for olfactory signaling in sperm navigation. Recent studies have uncovered several pairs of follicle fluid-derived odorants and sperm olfactory receptors, including androstenone–OR4D1, furaneol–OR7A5, methional–OR2H1, and propionic acid–OR51E2 (Hartmann et al., 2013; Flegel et al., 2015; Teveroni et al., 2022; reviewed by Elango and Kekäläinen, 2025); however, their involvement in COC-mediated human sperm chemotaxis awaits confirmation.

To date, how exactly COCs mediate sperm chemotaxis in vivo remains elusive. Nonetheless, many known chemoattractants converge on CatSper-dependent calcium signaling, a pathway essential for both sperm hyperactivation and the AR. This suggests that the COC-derived signaling molecules are, at a minimum, pivotal for final functional maturation in sperm and may represent the earliest molecular interactions between sperm and oocyte.

Sperm interaction with the cumulus oophorus

Once sperm approach the COC, they must make their way through the cumulus oophorus by digesting its viscous, hyaluronan-rich extracellular matrix. Hyaluronan is a linear glycosaminoglycan composed of repeating disaccharide units of D-glucuronic acid and N-acetyl-D-glucosamine linked by β-1,4-glycosidic bonds. Hyaluronidases cleave these bonds, loosening the cumulus matrix and enabling sperm progression (Fig. 2C; Kobayashi et al., 2020).

In both humans and mice, hyaluronidases are encoded by two major gene clusters: HYAL1, HYAL2, and HYAL3 on human chromosome 3 (orthologous to mouse Hyal1, Hyal2, and Hyal3 on chromosome 9) and SPAM1, HYAL4, and HYAL6P on human chromosome 7 (corresponding to mouse Spam1, Hyal4, Hyal5, and Hyal6 on chromosome 6). Humans lack a functional ortholog of mouse Hyal5, and human HYAL6P (also known as HYAL1P) is annotated as a pseudogene. Tissue expression profiles from the Human Protein Atlas (Uhlén et al., 2005, 2015) and the mouse ENCODE database (Dunham et al., 2012) indicate that HYAL3, HYAL4, and SPAM1 are predominantly expressed in the testis of both species, whereas HYAL1 and HYAL2 show weak expression in male reproductive tissues (also reviewed by Kobayashi et al., 2020; Seol et al., 2022); mouse Hyal5 and Hyal6 are testis-specific. In humans, functional dissection of individual hyaluronidases in cumulus dispersal remains incomplete, largely owing to limited availability of human gametes and effective analytical methodologies; our current understanding mainly originates from IVF and mutagenesis studies in mice.

SPAM1, also known as PH-20 or HYAL7, is a glycosylphosphatidylinositol (GPI)-anchored protein localized to the sperm head surface before and after the AR in humans, mice, and other mammals (Lin et al., 1994; Overstreet et al., 1995; SABEUR et al., 1997). It has a soluble isoform that resides in the acrosomal matrix and is released during the AR (Hunnicutt et al., 1996). Mouse sperm can also acquire SPAM1 from epididymal and oviductal fluids (Chen et al., 2006; Griffiths et al., 2008), although the physiological significance of exogenous SPAM1 uptake remains inconclusive. SPAM1 was initially regarded as the primary hyaluronidase mediating cumulus dispersal; preincubation of acrosome-intact mouse sperm with anti-SPAM1 antibodies prevents their passage through the cumulus oophorus (Lin et al., 1994), and Spam1 knockout sperm exhibit delayed cumulus dispersal and fertilization in vitro. Nevertheless, these knockout sperm can eventually traverse the cumulus cell layer and fertilize oocytes, and mutant males produce normal litter sizes (Baba et al., 2002).

HYAL5 is another GPI-anchored hyaluronidase, detected on both the sperm plasma membrane and acrosomal membrane (Kim et al., 2005). Disruption of Hyal5 alone in mice neither affects the expression of Spam1 nor impairs male fecundity; knockout sperm retain the ability to dissociate the cumulus cells and fertilize oocytes (Kim et al., 2005). Strikingly, double knockout mice lacking both Spam1 and Hyal5 are severely subfertile, with sperm exhibiting profound defects in cumulus dispersal and fertilizing ability (Park et al., 2019; Seol et al., 2022). These results highlight the cooperative and overlapping functions of sperm hyaluronidases in digesting the cumulus matrix (Zhang et al., 2005; Miller et al., 2007; Reitinger et al., 2007).

Although the expression of HYAL4 is biased toward the testis, its transcriptional levels are low in both humans and mice, and its involvement in cumulus matrix remodeling remains uncertain. Male mice lacking Hyal3 or Hyal6 are fertile; Hyal6-null sperm disperse the cumulus cells and fertilize oocytes normally, whereas Hyal3 knockout sperm show delayed cumulus penetration and reduced AR rates in vitro (Reese et al., 2010; Bang et al., 2022).

Sperm hyaluronidases exhibit varied enzyme activities in a pH-dependent manner. Enzymes active at neutral pH (e.g. HYAL5) are postulated to facilitate cumulus dispersal in acrosome-intact sperm (Kim et al., 2005), while those with optimal activity under acidic conditions (e.g. soluble SPAM1) may function during or after the AR (Evans et al., 2003), coinciding with local acidification inside the acrosome. Collectively, studies on sperm–cumulus interaction to date suggest that cumulus penetration is a multistep process coordinated by various sperm-borne hyaluronidases operating at different times and within distinct chemical microenvironments. Moreover, mouse sperm lacking arylsulfatase A (AS-A), an enzyme that cleaves sulfatides, or acrosin, a trypsin-like serine protease, also show delayed cumulus dispersal and fertilization (Baba et al., 1994; Wu et al., 2007; Isotani et al., 2017), implying that sperm deploy redundant enzymatic mechanisms to traverse the cumulus oophorus, in addition to hyaluronan–hyaluronidase hydrolysis.

On the oocyte side, loss of TNFAIP6 disrupts cumulus matrix formation, leading to impaired ovulation, reduced oocyte fertilizing ability, and female sterility (Fülöp et al., 2003). Depletion of PTX3, a TNFAIP6-interacting pentraxin produced by cumulus cells during cumulus expansion, causes female infertility in mice. Ptx3 knockout oocytes exhibit impaired cumulus matrix stability; they can be fertilized in vitro but not in vivo (Salustri et al., 2004). It was hypothesized that the failure of maintaining an intact cumulus matrix results in defective sperm chemoattraction, thereby leading to impaired in vivo fertilization. Female mice lacking the hyaluronan-interacting protein IαI are severely subfertile; their oocytes are completely devoid of cumulus matrix and rarely fertilized by sperm (Zhuo et al., 2001). Female mice deficient in PTGER2 are severely subfertile; their cumulus cells overexpress chemokines such as Ccl2, Ccl7, and Ccl9, resulting in excessive integrin engagement and increased resistance to hyaluronidases within the cumulus matrix, thereby preventing sperm penetration (Tamba et al., 2008). Moreover, impaired cumulus expansion, reduced hyaluronan, defective cumulus matrix integrity, and downregulation of hyaluronan-associated genes such as Ptx3 and Tnfaip6 have been observed in the oocytes ovulated by aged female mice, providing a linkage between reduced cumulus quality and age-related female infertility (Babayev et al., 2023). Hence, in addition to a collection of sperm hyaluronidases and auxiliary enzymes, a properly assembled and maintained cumulus oophorus is another prerequisite for successful sperm penetration and fertilization.

Defective sperm–cumulus interactions in humans

Determined by the diameters of halos formed around the sperm head upon exposure to hyaluronan, acrosomal hyaluronidase activity was found to positively correlate with sperm concentration, motility, and morphology, as well as the fertilization rate (Savion et al., 1986; Hirayama et al., 1989; Abdul-Aziz et al., 1995). Genetic mutations in HYAL1, HYAL2, and HYAL3 have been reported in humans, leading to mucopolysaccharidosis type IX, multiple developmental defects (e.g. cleft lip or palate, heart anomalies, flattened nasal bridge, and hypertelorism), and male infertility, respectively (Triggs-Raine et al., 1999; Muggenthaler et al., 2017; Fasham et al., 2022; Aslam et al., 2024; reviewed by Fink and Triggs-Raine, 2024). Compared with healthy males with an average sperm concentration of 120 million/ml, sperm concentrations in male patients carrying HYAL3 mutations are reduced to 10 million/ml or less, indicating oligospermia. Nevertheless, whether the reduced sperm count in these patients is caused by the genetic mutations or whether their sperm can efficiently disperse cumulus cells was not reported (Aslam et al., 2024).

To date, there are few reports on infertile patients with compromised sperm–cumulus interactions. Such defects, derived from impaired sperm chemotaxis or cumulus penetration or aberrant cumulus matrix formation or stability, can be readily overlooked by clinicians due to a lack of standardized diagnostic tools or strategies in vitro. However, it is anticipated that these can be bypassed through IVF, ZP drilling, subzonal injection (SUZI), or ICSI.

Molecular basis of sperm–ZP interaction

After passing through the cumulus oophorus, sperm encounter the second physical barrier, the ZP, an extracellular glycoprotein coat surrounding the oocyte (Fig. 3).

Figure 3.

Sperm–ZP interaction and polyspermy block. To fertilize an oocyte, a spermatozoon needs to interact with and penetrate the zona pellucida (ZP) to reach the oolemma (1), followed by binding to and fusion with the oolemma (2). Sperm–oolemma fusion initiates the cortical reaction (3) and cleanses the oocyte surface of sperm receptors, thereby establishing a plasma membrane (PM) block to polyspermy (4). Ovastacin secreted from the cortical granules cleaves the ZP2 protein in the zona pellucida (ZP), which, in turn, stimulates ZP structural remodeling and a ZP block to polyspermy (5). The zinc ions (Zn²⁺) released during the cortical reaction are postulated to mediate sperm chemorepellent, which also contributes to polyspermy block (6).

Biogenesis and architecture of the ZP

In the ovarian reserve, non-growing oocytes within primordial follicles remain arrested at the diplotene stage of the meiotic prophase I, a prolonged dormant state (Telfer and McLaughlin, 2007). At this stage, oocytes lack a ZP and are directly encapsulated by a monolayer of squamous pre-granulosa cells (Clarke, 2021). Upon follicular activation, oocytes undergo drastic growth and enlargement, during which de novo synthesis, intracellular processing, and secretion of ZP glycoproteins take place (Wassarman and Litscher, 2022).

Humans possess four ZP proteins, ZP1, ZP2, ZP3, and ZP4, whereas mice only have three, ZP1–ZP3, with Zp4 being a pseudogene in the Mus subgenus (e.g. Mus musculus and Mus caroli). Other murine species, such as Mus mattheyi, Mus pahari, and Mastomys coucha, retain a functional ZP4 and produce all four proteins (Izquierdo-Rico et al., 2020). All ZP proteins are synthesized as type-I single-pass transmembrane precursors, each of which contains an N-terminal signal peptide, a central ZP domain composed of immunoglobulin (Ig)-like ZP-N and ZP-C subdomains connected by a short interdomain linker, a consensus furin cleavage site upstream of an external hydrophobic patch (EHP), a C-terminal transmembrane-like domain, and a short cytoplasmic tail (Litscher and Wassarman, 2020a,b; Wassarman and Litscher, 2021). ZP1 and ZP4 each carry an extra ZP-N subdomain (ZP-N1) immediately upstream of the central ZP domain, while ZP2 harbors three tandem ZP-N repeats (ZP-N1, ZP-N2, and ZP-N3) in this region; in contrast, ZP3 lacks auxiliary ZP-N repeats (Monné et al., 2008). Further, ZP1 and ZP4 bear a conserved trefoil motif, a three-loop structure stabilized by three intramolecular disulfide bonds, positioned between their ZP-N1 and ZP domains (Thim, 1997).

During biosynthesis, ZP precursors remain tethered to the membrane as they undergo extensive N- and O-linked glycosylation and traffic from the endoplasmic reticulum (ER) through the Golgi apparatus to the oolemma (Wassarman et al., 2004). Prior to secretion, immature ZP proteins are maintained in a polymerization-incompetent state by the internal hydrophobic patch and EHP flanking the ZP-C subdomain. Proteolytic cleavage at the conserved furin site by an unidentified proprotein convertase dissociates the C-terminal EHP and transmembrane segment, triggering conformational rearrangements that expose the polymerization interface within the ZP domain (Jovine et al., 2002; Hoodbhoy et al., 2006; Litscher and Wassarman, 2020b). Once secreted, mature ZP proteins assemble via head-to-tail interactions to generate long, unbranched fibrils that are subsequently organized into a higher-order extracellular matrix. Specifically, ZP2–ZP3 heterodimers polymerize into filaments that are crosslinked by ZP1 and ZP4, collectively forming a dense 3-dimensional mesh with a thickness of 10–31 µm (Bertrand et al., 1995; Nishimura et al., 2019; Wassarman and Litscher, 2021). This unique supramolecular lattice confers both mechanical rigidity and elasticity, allowing the ZP to protect the growing oocyte, mature oocyte, and preimplantation embryo, while permitting small-molecule diffusion, sperm interaction, and blastocyst hatching.

Genetic studies in mice have clarified the individual functions of ZP proteins. Female mice lacking Zp1 ovulate oocytes with a fragile, loosely organized ZP, leading to reduced fecundity, likely due to premature embryo hatching (Rankin et al., 1999). Structural analysis indicates that ZP1 homodimerization via its ZP-N1 domain is essential for crosslinking ZP2 and ZP3 filaments into a stable matrix (Nishimura et al., 2019). Corroboratively, mice lacking all endogenous ZP subunits but expressing only human ZP2 and ZP3 (without ZP1) fail to form a functional ZP (Yauger et al., 2011). Zp2 knockout females form a thin, unstable ZP during early folliculogenesis, which is lost in pre-ovulatory follicles, leading to impaired follicle maturation and reduced ovulation (Rankin et al., 2001). In contrast, Zp3 knockout oocytes are completely devoid of the ZP throughout oogenesis, and their cumulus cells do not assemble into a stable oophorus (Liu et al., 1996; Rankin et al., 1996). The function of ZP4 has been interrogated in rats and rabbits by gene disruption. While ZP4 is dispensable for ZP morphology and female fertility in rats, it plays an important role in rabbits; the knockout females are severely subfertile, with oocytes surrounded by thinner ZPs and aberrant development of preimplantation embryos (Lamas-Toranzo et al., 2019; Zeng et al., 2021).

Sperm interaction with the ZP

The sperm–ZP interaction has been described in three sequential steps: primary binding of acrosome-intact sperm to the ZP, secondary binding of acrosome-reacted sperm to the ZP, and penetration through the ZP aided by acrosomal enzymes (Tumova et al., 2021). This model was based on an early assumption that the AR was triggered exclusively by ZP glycoproteins. However, subsequent discoveries challenged this theory. Progesterone secreted by cumulus cells can induce the AR before ZP binding, and acrosome-reacted sperm are frequently observed within the cumulus matrix in both humans and mice. These findings suggest that the AR is a spatiotemporally heterogeneous event, which may occur before reaching the COC, within the cumulus, or upon ZP interaction, depending on the physiological cues. Nevertheless, ZP proteins remain potent AR inducers and likely serve as the final checkpoint for sperm activation at fertilization.

Beyond different structural roles, individual ZP proteins exhibit distinct functions in sperm binding. Human sperm–ZP interaction has been studied using fluorophore-conjugated recombinant proteins or humanized mouse models. Generally, recombinant ZP1, ZP3, and ZP4 predominantly bind to the sperm acrosomal cap and equatorial segment before the AR and mainly bind to the equatorial segment after the AR (Chakravarty et al., 2005; Bansal et al., 2009; Ganguly et al., 2010). In contrast, recombinant ZP2 does not bind acrosome-intact but acrosome-reacted sperm at their equatorial segment and midpiece (Tsubamoto et al., 1999; Chakravarty et al., 2008b). However, human ZP2 and ZP3 directly purified from oocytes display distinct sperm binding patterns. Endogenous ZP2 binds both acrosome-intact and acrosome-reacted sperm at the equatorial segment, post-acrosomal region, and midpiece (Kerr et al., 2002; Chiu et al., 2008b). Similarly, native ZP3 binds broadly to the acrosomal cap, equatorial segment, and midpiece before the AR and binds only to the midpiece after the AR (Chiu et al., 2008a). Such discrepancies likely reflect differences in folding, glycosylation, or other post-translational modifications between native and recombinant proteins, as well as technical complications associated with antibody-based visualization of sperm-bound native ZP proteins.

To better understand how ZP proteins mediate sperm binding, human ZP1, ZP2, and ZP3 have been individually expressed in mice lacking the corresponding murine genes, while human ZP4 has been expressed in wild-type mice. Among these, only transgenic expression of human ZP2, but not ZP1, ZP3, or ZP4, enables human sperm binding and penetration of the ZP (Baibakov et al., 2012). This interaction can be competitively inhibited by agarose beads coated with the ZP2-N1 domain, demonstrating this region as a putative sperm binding interface (Baibakov et al., 2012; Avella et al., 2016). Intriguingly, although recombinant ZP1, ZP3, and ZP4 can bind sperm in vitro, they do not support sperm–ZP interaction when transgenically expressed in mice, possibly because of structural masking of their binding domains within the ZP lattice or competitive interactions with other ZP proteins.

Glycosylation is a pivotal determinant of ZP protein function, influencing both sperm receptivity and induction of the AR. The sialyl-Lewis X sequence, a highly enriched epitope on the N-glycans of human ZP proteins, has a direct role in sperm binding. Multivalent sialyl-Lewis X-bearing glycoconjugates bind the head of capacitated sperm and competitively suppress sperm–ZP interaction (Pang et al., 2011). Consistent with the functional requirement for glycosylation, baculovirus-expressed recombinant ZP1, ZP3, and ZP4, which undergo N-linked glycosylation, can stimulate the AR, whereas their non-glycosylated counterparts purified from E. coli fail to do so (Gupta, 2021). ZP-mediated AR induction is proposed to involve multiple partially redundant signaling pathways. In the classic model, ZP3 binding to its putative sperm receptor β-1,4-galactosyltransferase (GalTase) activates a pertussis toxin-sensitive heterotrimeric guanine nucleotide-binding regulatory protein (G protein) (Ward et al., 1992; Gong et al., 1995; Lu and Shur, 1997; Shi et al., 2001; Schuffner et al., 2002; Ni et al., 2007). Upon receptor activation, the GDP to GTP exchange on Gα_i leads to the release of Gα_i–GTP from the Gβγ subunit complex. Given that Gα_i canonically suppresses adenylyl cyclase and reduces cAMP levels (Federman et al., 1992), signaling mediated by the released Gβγ subunits is more likely contributing to the activation of phospholipase C (PLC) isoforms (Katz et al., 1992), such as PLCβ1 and PLCδ4 in sperm. Activated PLC hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP₂) to generate diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP₃). IP₃ promotes calcium release from intracellular stores, whereas DAG activates protein kinase C and may modulate calcium channel activity (Walensky and Snyder, 1995; Spungin and Breitbart, 1996). The resulting elevation of intracellular calcium, through both internal release and extracellular influx via voltage-dependent and other calcium channels, serves as a central trigger for AR initiation (Gupta and Bhandari, 2011; Simons and Fauci, 2018). Notably, PLCδ4-deficient mice exhibit an impaired ZP-induced AR, indicating that PLC-mediated PIP₂ hydrolysis is necessary for this process (Fukami et al., 2001, 2003). Pharmacological assays further indicate the involvement of alternative signaling pathways in reinforcing the ZP-induced AR, which can be inhibited by extracellular calcium chelator EGTA (Ni et al., 2007), the 20S proteasome inhibitor MG132 (Chakravarty et al., 2008a), the tyrosine kinase inhibitor tyrphostin A48, or the muscarinic receptor antagonist 3-quinuclidinyl benzilate (Bailey and Storey, 1994). Importantly, some of these inhibitory effects are insensitive to pertussis toxin, indicating the existence of G_i-independent or non-canonical mechanisms that converge on calcium mobilization and acrosome exocytosis. Despite the extensive in vitro evidence that ZP glycoproteins can induce the AR, the physiological timing of this event in vivo remains debated (Hirohashi and Yanagimachi, 2018). Live imaging in mice has indicated that many sperm undergo the AR before contacting the ZP; however, it is unknown whether this phenomenon is conserved in human fertilization, as direct observation of this process in humans is ethically and technically challenging (Jin et al., 2011; Hirohashi and Yanagimachi, 2018).

Numerous sperm proteins have been proposed as candidate ZP receptors, such as acrosin, AS-A, C1orf56, CRISP1, GalTase, MFGE8, SPACA2, SPAM1, ZAN, ZRK, ZP3R, and ZPBP1 (reviewed in Tumova et al., 2021). Acrosin, a fucose-binding serine protease, is stored in the acrosome as an inactive zymogen, proacrosin, and activated and released during the AR. Although acrosin exhibits proteolytic activity, it binds non-enzymatically to the mannosyl and fucosyl residues and sulfated glycans on ZP proteins (Töpfer-Petersen and Henschen, 1987; Francavilla et al., 1994; Howes and Jones, 2002; Vazquez-Levin et al., 2007). While an anti-acrosin antibody, AcrC5F10, inhibits proacrosin activation and ZP-induced AR without affecting sperm–ZP binding (Veaute et al., 2010), inhibition of acrosin activity by a soybean trypsin inhibitor abolishes ZP penetration but not binding (Liu and Baker, 1993). Ablation of acrosin in mice and rats does not impair ZP interaction but delays cumulus penetration (Baba et al., 1994; Isotani et al., 2017). In contrast, disruption of acrosin in hamsters and humans results in male-specific sterility; the mutant sperm can bind to but not penetrate the ZP (Hirose et al., 2020; Hua et al., 2023).

The sperm hyaluronidase SPAM1 shows limited involvement in ZP interaction, based on mouse mutagenesis analyses (Park et al., 2019). However, a recent structural analysis indicates that SPAM1 contains an epidermal growth factor (EGF)-like domain potentially functioning as a ZP binding interface (Im et al., 2025). SPAM1 has also been postulated to mediate ZP recognition by complexing with AS-A and HSPA2 (Redgrove et al., 2012). In this model, the complex initially positions SPAM1 on the sperm surface to disperse the cumulus matrix. As sperm traverse the cumulus, the protein complex is reoriented to expose AS-A, which directly binds ZP3 (Carmona et al., 2002; Tantibhedhyangkul et al., 2002; Redgrove et al., 2012). Whether this HSPA2–SPAM1–AS-A complex exists and functions at human sperm–ZP recognition requires confirmation.

GalTase binds O-linked oligosaccharides of ZP3 (Gong et al., 1995). However, depletion of GalTase in mice does not impair fertility, with sperm capable of ZP binding and penetration albeit with reduced AR efficiency (Lu and Shur, 1997). More recently, sperm C1orf56 has been identified as a candidate ZP receptor that recognizes the sialyl-Lewis X motifs on ZP glycans in humans. An antibody against C1orf56 blocks both sperm–ZP binding and ZP-stimulated AR in vitro (Wang et al., 2021c).

In mice, sperm–ZP binding is regulated by a network of testis-enriched proteins, such as ADAM3, CALR3, CLGN, PDILT, TEX101, as well as CST, LYPD, SPACA4, and PATE family proteins (Yamaguchi et al., 2009; Ikawa et al., 2011; Tokuhiro et al., 2012; Fujihara et al., 2019, 2021). Ablation of these proteins disrupts ADAM3 maturation and stability, leading to defective sperm migration through the uterotubal junction in vivo and failure in binding and penetrating the ZP of cumulus-free oocytes in vitro. Despite the central role of Adam3 in mouse fertilization, it is annotated as a pseudogene, ADAM3A, in humans, highlighting remarkable species divergence in the molecular basis of sperm–ZP interaction. Recently, biallelic mutations in CALR3 have been identified in humans, causing fertilization failure associated with abnormal acrosome morphology and compromised sperm–ZP binding (Gao et al., 2024).

Following sperm binding to the ZP, penetration is achieved through the combined action of hyperactivated motility and enzymatic digestion by acrosomal hydrolases, including acrosin and hyaluronidases. In addition to these enzymes, inhibition assays have shown that other proteases, such as the matrix metalloproteinase MMP2 and the testis serine protease TESP5, also contribute to ZP degradation (Honda et al., 2002; Ferrer et al., 2012; Saldívar-Hernández et al., 2015). Moreover, sperm-associated 26S proteasomes have emerged as a potential contributor of ZP penetration. In vitro assays demonstrate that isolated sperm proteasomes can degrade ZP proteins, and inhibition of proteasomal activity reduces the efficiency of fertilization and polyspermy block, suggesting a role in sperm-mediated ZP remodeling (Zimmerman et al., 2011). Consistently, the sperm 26S proteasome digests ZP3 in vitro; this degradation is blocked by proteasome inhibitors and is enhanced in the presence of ubiquitin-aldehyde (Zimmerman et al., 2011), supporting a model where ubiquitination of ZP glycoproteins targets them for proteasomal processing.

Defective sperm–ZP interactions in humans

Female infertility caused by mutations in ZP proteins has been reported in numerous studies. Homozygous and compound heterozygous mutations in ZP1 lead to female infertility caused by empty follicle syndrome or defective oocyte maturation (Huang et al., 2014; Dai et al., 2019a; Sun et al., 2019, 2023; Zhou et al., 2019; Liu et al., 2020; Luo et al., 2020; Okutman et al., 2020; Xu et al., 2020; Wang et al., 2021b, 2025; Wu et al., 2021; Zou et al., 2022; Pujalte et al., 2023; He et al., 2025). Among these, a homozygous frameshift mutation in ZP1 abolishes ZP formation, owing to cytoplasmic retention of truncated ZP1 together with ZP3 (Huang et al., 2014). Heterozygous mutations in the N-terminus or the trefoil domain of ZP1 impair ZP assembly and female fertility (Cao et al., 2020, 2024).

Likewise, homozygous and compound heterozygous mutations in ZP2 lead to female infertility characterized by empty follicle syndrome, abnormal ZP morphology, or defective oocyte maturation (Dai et al., 2019b; Zhou et al., 2019; Luo et al., 2020; Sun et al., 2021, 2023; Jia et al., 2022; Zeng et al., 2023b; Yang et al., 2025a). A heterozygous mutation in the consensus furin cleavage site of ZP2 causes empty follicle syndrome by preventing ZP2 maturation and extracellular trafficking (Shen et al., 2022). In addition, a heterozygous mutation in the ZP-C subdomain decreases ZP2 expression and secretion, abolishing ZP formation and impairing folliculogenesis (Liu et al., 2023). Other heterozygous mutations located in the short linker between ZP-N and ZP-C or within the ZP-C domain disrupt ZP2–ZP3 interaction, resulting in abnormal ZP morphology and female infertility (Hou et al., 2022).

Mutations in ZP3, the core component of the ZP, also result in female infertility. Compound heterozygous mutations cause empty follicle syndrome in female patients (Chen et al., 2021; Jia et al., 2022; Kong et al., 2023). Heterozygous mutations in ZP3 can exert dominant-negative effects, leading to a wide spectrum of phenotypes, such as empty follicle syndrome, abnormal ZP morphology, defective oocyte maturation, or ovulatory dysfunction (Chen et al., 2017; Zhou et al., 2019, 2024; Cao et al., 2020; Zhang et al., 2021; Sun et al., 2023; Liu et al., 2025; Sa et al., 2025; Wang et al., 2025).

Reports on ZP4 mutations are relatively scarce. However, heterozygous mutations in the N-terminus and the ZP domain have been associated with thin and disorganized ZP structures, leading to infertility (Wei et al., 2022). Taken together, mutations in any of the four ZP genes compromise ZP formation, leading to fertilization failure either in vivo or in vitro. In most cases, infertility caused by ZP abnormalities can be bypassed by ICSI. However, in patients with empty follicle syndrome or a complete absence of the ZP, treatment options remain extremely limited, and oocyte donation often becomes the only viable alternative.

On the male side, low sperm acrosin activity has been associated with deficiencies in acrosomal enzymes, including SPAM1, as well as reduced glutathione, which protects sperm from oxidative stress and prevents premature acrosome exocytosis (Lin et al., 2023). In patients receiving IVF or ICSI treatments, low sperm acrosin activity correlates with reduced cumulative live birth rates (Zhao et al., 2024). Strikingly, male patients carrying a biallelic W56X mutation in acrosin are sterile; their sperm can bind to but not penetrate the ZP. These individuals exhibit normal semen parameters and normal gross morphology of sperm, but ultrastructural analyses reveal detachment of the acrosome from the nucleus (Hua et al., 2023). Similarly, a homozygous frameshift and two compound heterozygous mutations in CALR3 cause around 80% of sperm with abnormal head morphology, characterized by the acrosome dislodging from the sperm nucleus (Gao et al., 2024). Despite normal semen parameters, the mutant sperm fail to bind to or penetrate the ZP. Importantly, sperm carrying mutations in acrosin or CALR3 can achieve fertilization success with the assistance of SUZI or ICSI (Hua et al., 2023; Gao et al., 2024), demonstrating that they retain the ability of membrane fusion and oocyte activation.

Furthermore, ZP-binding ability is increasingly recognized as a marker of human sperm quality. Human ZP preferentially binds sperm with higher fertilization potential and genomic integrity, although the underlying molecular mechanisms remain poorly defined (Leung et al., 2023). A deep learning algorithm has recently been developed to identify human sperm with ZP-binding capability (Leung et al., 2025), facilitating the identification of males at high risk of unexpected IVF failure in clinical settings.

Molecular basis of sperm–oolemma interaction

After traversing the ZP and entering the perivitelline space, an acrosome-reacted sperm binds to and fuses with the oolemma, completing the union of the male and female gametes (Fig. 3). Compared with studies on sperm–cumulus and sperm–ZP interactions, major progress in deciphering the molecular mechanisms of sperm–oolemma interaction has been achieved only recently, largely enabled by CRISPR/Cas9-based genome editing in mice and zebrafish. Although the full mechanism is not yet consolidated, current evidence indicates that sperm–oolemma interaction involves an array of sperm ligands and oolemma receptors. Their interplay bridges the two membranes and establishes a ‘fertilization synapse’, a specialized microenvironment that promotes membrane fusion.

Roles of IZUMO1 and JUNO in sperm–oolemma interaction

To date, IZUMO1 on sperm and JUNO (also known as IZUMO1R or FOLR4) on oocytes constitute the only validated ligand–receptor pair between mammalian gametes (Inoue et al., 2005; Bianchi et al., 2014). The story of IZUMO1 began nearly half a century ago with the discovery of OBF13 (Okabe, Butsu-Metsu, Friday the 13th), a mouse anti-sperm antibody identified by Masaru Okabe (Okabe et al., 1986). This antibody blocked fertilization in vitro and was later shown by mass spectrometry to target IZUMO1, a type-I single-pass transmembrane protein localized to the acrosomal membrane (Okabe et al., 1986, 1987; Kawai et al., 1989; Inoue et al., 2005; Satouh et al., 2012). Mice lacking Izumo1 exhibit male-specific sterility despite producing morphologically intact sperm with normal motility and the ability to elicit the AR and penetrate the ZP. However, their sperm fail to fuse with the oolemma, which prevents ZP hardening and the establishment of polyspermy block. Consequently, multiple sperm enter the perivitelline space, yet none successfully fertilize the oocyte (Inoue et al., 2005). This phenotype is also shared among knockout mice lacking other male or female factors crucial for sperm–oolemma interaction [e.g. Fimp, Spaca6, Tmem95, Sof1, Dcst1, and Dcst2 knockouts (Fujihara et al., 2020; Noda et al., 2020, 2022), and Cd9 and Juno knockouts (Bianchi et al., 2014; Stickler and Avella, 2024)].

A decade after the discovery of IZUMO1, its oocyte receptor JUNO, a GPI-anchored folate receptor-like protein, was identified through an in vitro interactome screen using pentamerized IZUMO1 to probe cultured cells randomly expressing oocyte surface proteins (Bianchi et al., 2014). The essential role of JUNO was also validated by mouse mutagenesis; Juno knockout females are infertile because their oocytes are unable to fuse with wild-type sperm. The importance of the IZUMO1–JUNO interaction has been demonstrated in human fertilization. Antibodies against JUNO block human IVF, whereas anti-IZUMO1 antibodies prevent human sperm from fertilizing ZP-free golden hamster oocytes, which are known to fuse with sperm from a wide range of mammalian species, such as mice, dogs, and pigs (Inoue et al., 2005; Bianchi and Wright, 2015; Jean et al., 2019; Tang et al., 2022).

X-ray crystal structures of IZUMO1, JUNO, and their co-structure have revealed the detailed functions of the conserved domains and the molecular basis of their recognition (Aydin et al., 2016; Han et al., 2016; Kato et al., 2016; Nishimura et al., 2016; Ohto et al., 2016; Lu et al., 2025). The ectodomain of IZUMO1 adopts a boomerang-shaped structure, consisting of an N-terminal four-helix bundle (4HB) and a C-terminal Ig-like domain linked by a β-hairpin hinge (Aydin et al., 2016; Ohto et al., 2016). JUNO harbors eight α-helices and four β-strands that form a globular architecture stabilized by eight disulfide bonds. Unlike other folate receptors, JUNO has lost the ability to accommodate folic acid due to the absence of six key folate-binding residues (Kato et al., 2016; Ohto et al., 2016). IZUMO1–JUNO binding involves IZUMO1’s α2–α3 loop of the 4HB, the β-hairpin hinge, and β8–β9 loop of the Ig-like domain engaging a conserved surface on JUNO opposite to its non-functional folate-binding pocket. A single tryptophan (W148) within the central β-hairpin of IZUMO1 is one of the key residues mediating JUNO binding; an alanine substitution completely abolishes interaction with JUNO (Aydin et al., 2016; Kato et al., 2016; Ohto et al., 2016).

The initial study reported that Izumo1 knockout mouse sperm bound to the oolemma as efficiently as wild-type sperm, giving the impression that IZUMO1 was dispensable for gamete binding but essential for fusion (Inoue et al., 2005). However, a subsequent investigation revealed that most sperm bound to ZP-free oocytes were acrosome-intact. Such binding is physiologically irrelevant, as it reflects the inherent stickiness of the mouse sperm surface rather than specific extracellular protein–protein interactions (Matsumura et al., 2021). To distinguish acrosome-intact from acrosome-reacted sperm, live mouse sperm were probed with a fluorophore-conjugated anti-IZUMO1 antibody prior to IVF (Matsumura et al., 2021; Deneke et al., 2024; Nishio et al., 2024). These refined assays indicated that Izumo1-deficient mouse sperm show reduced binding to ZP-free oocytes. Similarly, Izumo1 knockout rat sperm fail to bind the rat oolemma (Matsumura et al., 2021). In concert with these findings, HEK293T cells transiently expressing IZUMO1 adhere strongly to but do not fuse with ZP-free mouse oocytes. Likewise, binding, but not fusion, was detected between cells overexpressing IZUMO1 and JUNO, suggesting that this ligand–receptor pair contributes to sperm–oolemma binding but is insufficient to drive membrane fusion (Bianchi et al., 2014; Noda et al., 2020).

Contrary to these observations, recent studies have uncovered that IZUMO1 exhibits fusogenic activity in vitro (Brukman et al., 2023, 2024). When ectopically expressed in Baby Hamster Kidney (BHK) cells, IZUMO1 induced both unilateral cell–cell fusion (i.e. fusogen on one membrane) at a rate of ∼6% and bilateral fusion (i.e. fusogen on both membranes) of ∼8%, compared to ∼1.5% in cells transfected with empty vectors. Similarly, IZUMO1 also induces fusion, but to a lesser extent, in HEK293T cells, with fusion rates increasing from ∼1% in blank controls to ∼2.5% and 3% for IZUMO1-mediated unilateral and bilateral fusion, respectively. Interestingly, the W148A mutation in IZUMO1 ablates its interaction with JUNO but does not impair IZUMO1-mediated cell–cell fusion, while triple mutations within the 4HB (F28A, W88A, and W113A) do not affect JUNO binding but inhibits cell–cell fusion (Brukman et al., 2023). These observations indicate that the β-hairpin hinge of IZUMO1 is critical for JUNO binding but not for its fusogenic activity. Instead, the 4HB region may serve as the functional domain that drives gamete fusion. Further, sperm were found capable of fusing with JUNO-expressing BHK cells and triggering fusion between cells overexpressing JUNO (Brukman et al., 2023, 2024), suggesting that additional factors on sperm are required to complete the fusion process.

Other factors underpinning sperm–oolemma interaction

Since 2020, multiple additional sperm proteins have been identified as essential for sperm–oolemma interaction through gene disruption in mice and zebrafish. These include SPACA6 (Lorenzetti et al., 2014; Barbaux et al., 2020; Noda et al., 2020; Binner et al., 2021), TMEM95 (Pausch et al., 2014; Fernandez-Fuertes et al., 2017; Lamas-Toranzo et al., 2020; Noda et al., 2020), TMEM81 (Deneke et al., 2024), FIMP (Fujihara et al., 2020), DCST1/2 (Inoue et al., 2021; Noda et al., 2022), SOF1 (or LLCFC1; Noda et al., 2020), and FREY1 (Contreras et al., 2022; Hao et al., 2022; Lu et al., 2023). Among these proteins, SPACA6, TMEM95, and TMEM81 are type-I transmembrane proteins, sharing structural features with IZUMO1.

Mouse Fimp encodes both a membrane-bound and a soluble isoform lacking a signal peptide. Transgenic expression of a mCherry-tagged membrane-type FIMP fully restores the subfertility of Fimp knockout males, indicating that the soluble form is dispensable (Fujihara et al., 2020). FIMP localizes to the equatorial segment of acrosome-intact sperm but is lost in more than half of the sperm after the AR (Fujihara et al., 2020). Its precise molecular role in sperm–oolemma interaction remains to be elucidated. SOF1 is a putative secreted protein with unknown function (Noda et al., 2020). FREY1, a type-II transmembrane protein localized to the ER, interacts with and stabilizes subunits of a testis-expressing oligosaccharyltransferase complex, which mediates N-linked glycosylation of acrosomal membrane proteins. Loss of FREY1 downregulates acrosomal membrane proteins such as CD46, EQTN, IZUMO1, SPACA1, and SPACA6 and impairs the sperm–oolemma interaction (Lu et al., 2023).

IZUMO1, SPACA6, and TMEM95 have been categorized into an IST superfamily, featured by an N-terminal 4HB with a triangular face and a pair of disulfide-bonded CXXC motifs (Vance et al., 2022). SPACA6 shares structural similarities to IZUMO1, with an N-terminal distorted 4HB and a C-terminal Ig-like domain linked by a β-hairpin. By contrast, TMEM95 possesses only three helices (α1, α3, and α4) in its N-terminus and lacks an Ig-like domain (Tang et al., 2022). In mice, both SPACA6 and TMEM95, like IZUMO1, are localized to the acrosomal membrane and translocated to the equatorial segment before and after the AR, respectively (Fig. 4A), with an exception that IZUMO1 occasionally redistributes to the entire head surface of acrosome-reacted sperm (Satouh et al., 2012; Noda et al., 2020). Remarkably, SPACA6 is absent in mutant sperm lacking IZUMO1, TMEM95, SOF1, FIMP, DCST1, DCST2, or TMEM81 (Inoue et al., 2021; Lu et al., 2023), suggesting that SPACA6 forms a complex with one or more of these factors on the sperm acrosomal membrane to retain its stability.

Figure 4.

Sperm–oolemma interaction. (A) The acrosomal membrane proteins are localized to the outer (OAM) and inner acrosomal membranes (IAM), both the acrosomal (AM) and plasma membranes (PM), and equatorial segment, respectively, before, during, and after the acrosome reaction. (B) A suite of proteins has been validated as essential for sperm–oolemma interaction through mouse knockout studies. Notably, TMEM95 has a putative GPI-anchored receptor on the oolemma. (C) AlphaFold-Multimer predicts a trimeric sperm complex of IZUMO1, SPACA6, and TMEM81 interacting with the oocyte JUNO (adapted from Deneke et al., 2024).

Of the IST family proteins, the physiological function of TMEM95 has been most extensively studied in humans. TMEM95 features an evolutionarily conserved, positively charged surface within the α3–α4 loop of its N-terminal helix bundle (Tang et al., 2022). Antibodies targeting this interface inhibit human sperm fusion with ZP-free golden hamster oocytes but do not affect sperm binding. In contrast, IZUMO1 antibodies against the JUNO-interacting surface completely abolish sperm binding (Tang et al., 2022). Although human TMEM95 does not interact with JUNO, it strongly binds the hamster oolemma, demonstrating the presence of an unknown oocyte receptor. This binding can be abolished by phosphoinositide PLC, suggesting that the receptor might be GPI-anchored. It has been postulated that TMEM95 binds its putative receptor to reinforce membrane binding or promote fusion downstream of IZUMO1–JUNO interaction (Tang et al., 2022). Mutations in bovine TMEM95 impair male fertility (Pausch et al., 2014; Fernandez-Fuertes et al., 2017; Zhang et al., 2019). Furthermore, AlphaFold predicts that TMEM95 forms a heterodimer with FIMP, a putative fusion-inducing complex that awaits experimental validation (Elofsson et al., 2024).

Recent AlphaFold-Multimer predictions suggest that IZUMO1 and SPACA6 interact with TMEM81 to form a trimeric complex conserved across vertebrates such as zebrafish, humans, and mice (Deneke et al., 2024; Elofsson et al., 2024). This trimer does not interfere with the IZUMO1–JUNO interaction, and AlphaFold further predicts a tetrameric assembly involving JUNO with high confidence (Fig. 4B and C). The presence of this trimer on the sperm surface has been experimentally validated in zebrafish by co-immunoprecipitation–mass spectrometry; cell-based assays further indicate that the trimer interacts with Bouncer, the sole known key fertilization factor on zebrafish oocytes (Deneke et al., 2024). The trimer was also detected in the giant plasma membrane vesicles released by FreeStyle 293-F cells expressing full-length human IZUMO1, SPACA6, and TMEM81, providing indirect evidence for trimerization in mammals (Deneke et al., 2024). AlphaFold predicts that TMEM81 harbors a β-hairpin and an Ig-like domain, similar to IZUMO1 and SPACA6, but lacks an N-terminal 4HB. Mutations at predicted IZUMO1–SPACA6–TMEM81 interfaces destabilize this trimeric complex, thereby abolishing the fertilizing ability of zebrafish sperm (Deneke et al., 2024); however, structural and biochemical analyses are warranted to consolidate these protein interactions in humans and mice.

Unlike the vertebrate-restricted IZUMO1, SPACA6, TMEM95, and TMEM81, DCST1 and DCST2 exhibit broader conservation across both vertebrates and invertebrates, including fruit flies and nematodes (Inoue et al., 2021; Noda et al., 2022). These multi-pass transmembrane proteins contain a C-terminal RING-finger domain, resembling invertebrate orthologs such as Sneaky in Drosophila, and SPE-42 and SPE-49 in Caenorhabditis elegans (Kroft et al., 2005; Wilson et al., 2006, 2011, 2018). In mammals, DCST1 and DCST2 share homology with DC-STAMP and OC-STAMP, which regulate cell–cell fusion in osteoclasts and foreign body giant cells (Ishii et al., 2018). Oc-stamp knockout mice fail to form multinucleated osteoclasts or giant cells (Miyamoto et al., 2012; Witwicka et al., 2015); their fusogenic activity is pending evaluation. DCST1 and DCST2 interact and are interdependent; depletion of one results in the loss of the other (Noda et al., 2022). However, overexpression of the two proteins in cultured cells does not induce fusion with ZP-free oocytes (Inoue et al., 2021; Noda et al., 2022).

On the oolemma, the tetraspanins CD9 and its paralog CD81 are important mediators of sperm–oolemma interaction, in addition to JUNO (Kaji et al., 2000; Le Naour et al., 2000; Miyado et al., 2000; Rubinstein et al., 2006). Female mice lacking Cd9 or Cd81 exhibit severe or mild subfertility, respectively, while double knockouts completely abolish female fertility (Rubinstein et al., 2006). Cd9-deficient oocytes show short and sparse microvilli (Runge et al., 2007), whereas deletion of Cd81 does not affect microvillar morphology (Benammar et al., 2017). Notably, overexpression of CD81 in Cd9 knockout oocytes restores fertilization rates to up to ∼50%, demonstrating partial functional redundancy between the two proteins (Kaji et al., 2002). Both CD9 and CD81 are secreted in extracellular vesicles from oocytes but differ in localization; CD9 is immunodetected at both the oolemma and the perivitelline space, while CD81 is predominantly distributed beneath the ZP (Ohnami et al., 2012). Interestingly, co-incubating Cd9 knockout oocytes with wild-type oocytes or CD9-containing exosomes restores their fertilizing ability, presumably through exosome incorporation into sperm and/or oocyte (Miyado et al., 2008).

Structurally, CD9 and CD81 contain four transmembrane helices and a small and a large extracellular loop (LEL), arranging into a cone-shaped tertiary structure with a central intramembrane cavity (Zimmerman et al., 2016; Umeda et al., 2020). Transmembrane proteins with such conical architectures naturally create local membrane curvature (McMahon and Gallop, 2005). Recently, CD9 has been demonstrated as a membrane curvature sensor, which is enriched in highly curved membrane tubes mechanically pulled from giant plasma membrane vesicles (Dharan et al., 2022). Together with the disrupted microvillar architecture observed in Cd9 knockout oocytes, these findings underscore the crucial role of CD9 in shaping membrane curvature at the oocyte surface, which ultimately promotes the formation of properly extended microvilli. Mutations within the LEL of CD9 impair sperm–oolemma interaction (Umeda et al., 2020), suggesting that CD9 interacts with unidentified cofactors in cis or receptors in trans to facilitate sperm interaction. Deep learning-based modeling predicts a higher-order complex involving interactions between the IZUMO1–SPACA6–TMEM81 trimer with both JUNO and CD9. In this model, the 4HB and hinge region of SPACA6 bind the 4HB of IZUMO1 on the opposite side of the JUNO-interacting surface, constituting a concave pocket that engages the LEL of CD9 (Elofsson et al., 2024). Co-localization and co-immunoprecipitation analyses in HEK293T cells overexpressing these factors suggest a potential association between JUNO and CD9 (Frolikova et al., 2023). However, the existence of such a higher-order complex at the sperm–oocyte interface remains speculative and requires experimental validation. Beyond its potential role as a complex subunit, CD9 regulates molecular compartmentalization at the oocyte surface, excluding GPI-anchored proteins such as JUNO and CD55 from the cortical cap (Inoue et al., 2020). CD9 also activates the cytosolic protein kinase PTK2B, which remodels actin at the sperm binding site to promote fusion (Wang et al., 2021a).

MAIA (or FCRL3), a protein expressed on the human oocyte surface, has been debated regarding its role as a secondary oocyte receptor for IZUMO1 following JUNO binding. In mice, disruption of a putative MAIA ortholog, Fcrl5, does not interfere with female fertility or sperm–oolemma interaction (Vondrakova et al., 2022). Cell-based in vitro assays indicated that MAIA interacted with IZUMO1 in trans and JUNO in cis, and co-expression of MAIA and JUNO in HEK293T cells significantly enhanced their fusion with human sperm (Vondrakova et al., 2022). Nevertheless, these findings have been challenged by a subsequent study, which detected no direct interaction between MAIA with either IZUMO1 or JUNO, using avidity-based extracellular interaction screening (AVEXIS) or binding assays between recombinant proteins and intact cell membranes (Bianchi et al., 2024). Furthermore, MAIA failed to induce fusion between JUNO- and IZUMO1-expressing cells in a split-GFP complementation assay (Bianchi et al., 2024). Therefore, the contribution of MAIA to mammalian sperm–oolemma interaction remains inconclusive and warrants further investigation.

Historical observations of sperm–oolemma fusion in humans, hamsters, and mice by transmission electron microscopy indicate that gamete membrane fusion can be subdivided into two steps: sperm membrane at the equatorial region first fuses with the microvilli on the oolemma, and after the primary fusion, a wave of cytoplasm is excited at the oocyte surface to cover the sperm head, which ultimately engulfs and internalizes the entire sperm (Yanagimachi and Noda, 1970; Thompson et al., 1974; Sathananthan and Chen, 1986). These observations imply the potential involvement of phagocytosis in final sperm internalization following initial membrane fusion. Indeed, a previous study discovered that phosphatidylserine (PtdSer) on the sperm surface serves as a phagocytotic signal by interacting with its oocyte receptors during fertilization. Blocking either PtdSer or its oocyte receptors significantly reduces sperm–oolemma fusion. However, individual disruptions of PtdSer receptors pose minor or no effects on fertilization success, indicating that PtdSer-mediated sperm internalization may be redundantly regulated by a panel of oocyte receptors (Rival et al., 2019).

A more recent work proposes a non-canonical phagocytosis model in which the oocyte actively engulfs the spermatozoon as a prerequisite for fertilization (Inoue et al., 2025). This process unfolds in two sequential stages, termed the oocyte tentacle stage and the SEAL stage (sperm engulfment activated by IZUMO1–JUNO linkage and gamete fusion-related factors). In the oocyte tentacle stage, the IZUMO1–JUNO interaction establishes membrane binding between sperm and oocyte, drives JUNO accumulation at the sperm binding site, and triggers elongation of microvilli to form lamellipodium-like protrusions that extend to the sperm head, a process likely facilitated by CD9. In the subsequent SEAL stage, the oolemma envelops the entire sperm head in a phagocytic-like manner, which requires sperm proteins DCST1/2, FIMP, SPACA6, TMEM81, and TMEM95, and culminates in JUNO depletion from the phagocytosis site (Inoue et al., 2025).

Defective sperm–oolemma interactions in humans

Numerous mouse mutagenesis studies suggest that infertile patients with defects in sperm–oolemma interaction may nevertheless produce morphologically normal gametes, rendering such anomalies virtually undetectable by routine semen analysis. To date, no human patients carrying mutations in the key factors mediating sperm–oolemma interaction have been definitively identified (Hayasaka et al., 2007; Granados-Gonzalez et al., 2008). However, two mutational analyses of JUNO have been conducted on women with recurrent fertilization failure, polyspermy during IVF, or infertility with unknown etiology (Takaiso et al., 2016; Yu et al., 2018). Although the studies suggest a potential link between JUNO variants and impaired fertilization, they did not directly assess the fertilizing ability of oocytes from the female patients. More recently, an H177Q variant of JUNO was identified in ∼9% of a female idiopathic infertility cohort (Boult et al., 2025). This study demonstrated that the IZUMO1–JUNO interaction adopts a catch-bond strategy, whereby mechanical force generated by sperm flagellar beating increases the bond lifetime between the two proteins. The H177Q substitution compromises the mechanostability of the IZUMO1–JUNO complex, potentially destabilizing sperm–oolemma binding and contributing to infertility (Boult et al., 2025). IZUMO1 protein levels were reported to be significantly reduced in sperm from patients experiencing total fertilization failure, compared to healthy controls (Enoiu et al., 2022). Given the sensitivity of SPACA6 stability to the integrity of the putative sperm fusion complex, future studies could explore the use of anti-SPACA6 antibodies as a potential diagnostic tool to assess sperm competence for interaction with the oolemma.

Molecular basis of the polyspermy block

In humans, oocytes fertilized by multiple sperm undergo abnormal embryogenesis that results in spontaneous abortion (Sevenster et al., 1987). In some cases, however, polyspermic conceptuses develop into gestational trophoblastic diseases such as a hydatidiform mole, a pathological condition that severely threatens maternal health (Vejerslev et al., 1987; Nikiforaki et al., 2014). To prevent such detrimental outcomes, the fertilized oocyte rapidly triggers biochemical and structural changes at both the ZP and the oolemma, leading to the blockage of additional sperm entry (Evans, 2020; Fahrenkamp et al., 2020). These post-fertilization events are referred to as the ZP block to polyspermy (ZPBP) and the plasma membrane block to polyspermy (PMBP).

ZP block to polyspermy

Following gamete fusion, the sperm delivers a sperm-specific phospholipase PLCζ1 into the cytoplasm of the oocyte. Similar to other PLC enzymes, PLCζ1 hydrolyzes PIP₂ into DAG and IP₃. Binding of IP₃ to its receptors on the ER triggers the release of intracellular calcium stores, generating oscillatory calcium waves for oocyte activation (Saunders et al., 2002; Kouchi et al., 2004; Yoon et al., 2012; Nomikos et al., 2013; Nozawa et al., 2018).

Meanwhile, this calcium surge drives the exocytosis of cortical granules, Golgi-derived secretory organelles distributed beneath the oocyte cortex (Fig. 3). Upon fusion with the oolemma via SNARE-dependent mechanisms, cortical granules release a cocktail of molecules, including the metalloprotease ovastacin (or ASTL), glycosidases, proteases, lectins, and inorganic zinc, into the perivitelline space (Burkart et al., 2012; de Paola et al., 2015; Duncan et al., 2016; Körschgen et al., 2017; Que et al., 2017; Tokuhiro and Dean, 2018; Fahrenkamp et al., 2020). These factors cooperatively mediate remodeling of the ZP, initiating biochemical processes including ovastacin-induced ZP2 cleavage at the N-terminus (Burkart et al., 2012; Kang et al., 2023; Nishio et al., 2024), ZP3 deglycosylation by glycosidases and proteases (Miller et al., 1993), lectin-mediated ZP glycan crosslinking (Avilés et al., 1996), and ZP uptake of zinc ions released from the oocyte (Tokuhiro and Dean, 2018). As a result, the ZP becomes hardened, exhibiting decreased sperm receptivity and enhanced resistance to proteolytic digestion, thereby effectively preventing further sperm binding and penetration (Nishio et al., 2024). A recent study has provided structural insights into how ZP hardening is achieved. Post-fertilization cleavage of human ZP2 leads to oligomerization of its N-terminal domain, which extensively crosslinks ZP filaments, rigidifying the ZP and rendering it physically impenetrable by sperm (Nishio et al., 2024).

Plasma membrane block to polyspermy

Fertilized oocytes also establish a PMBP that prevents additional sperm fusion with the oolemma (Fig. 3; Sengoku et al., 1995). After fertilization, JUNO is rapidly shed from the oolemma in the form of extracellular vesicles (Bianchi et al., 2014). This process eliminates functional sperm receptors on the oolemma, while the released JUNO-containing vesicles may further bind to acrosome-reacted sperm in the perivitelline space, thereby neutralizing their fusion ability (Bianchi et al., 2014; Bianchi and Wright, 2016). Similarly, the abundance of CD9 is significantly reduced on the surface of fertilized oocytes (Żyłkiewicz et al., 2010; Ravaux et al., 2018). A recent work in the mouse model suggests that these shed sperm receptors, including JUNO and CD9, bind excess sperm in the perivitelline space to reinforce the PMBP (Dubois et al., 2026). Moreover, while unfertilized oocytes exhibit uniformly distributed thin microvilli, fertilized oocytes display thicker and irregularly distributed microvilli, suggesting a fertilization-induced remodeling of the oolemma architecture (Żyłkiewicz et al., 2010). Such structural reorganization presumably impedes the oocyte’s competence to support additional fusion events. It has also been proposed that fertilization-triggered cytosolic calcium increase has a role in the PMBP. Suppression of calcium transients compromises the establishment of the PMBP, indicating that calcium signaling not only drives the cortical reaction but also reduces sperm receptivity at the oolemma (Gardner et al., 2007; Evans, 2020). Despite these insights, the PMBP remains underexplored compared to the ZPBP. Current evidence supports a multilayered model in which receptor shedding, membrane remodeling, and calcium-dependent signaling act synergistically to prevent polyspermy.

Defective polyspermy blocks in humans

Clinical evidence of defective polyspermy blocks has primarily emerged in infertile men and women carrying mutations in PLCζ1 and ASTL, respectively. In men, spermatozoa harboring PLCζ1 mutations fail to evoke calcium oscillations after fertilization. As a result, the exocytosis of cortical granules does not occur, the ZPBP is not established, and oocyte activation is impaired, resulting in polyspermy and delayed pronuclear formation. Remarkably, these patients were able to achieve pregnancies through ICSI followed by artificial oocyte activation using the calcium ionophore A23187 (Peng et al., 2023; Che et al., 2024; Tong et al., 2024). In women, biallelic variants in ASTL have been associated with fertilization failure characterized by polyspermy and, in some cases, defective oocyte maturation. Loss of ASTL compromises the ZPBP, permitting excessive sperm entry. Nevertheless, these female patients were able to conceive using ICSI, which ensures monospermic fertilization in vitro (Wu et al., 2023; Zeng et al., 2023a).

Emerging approaches for studying human sperm–oocyte interactions

Human gametes are highly regulated biological materials in research laboratories, subject to strict legal frameworks and ethical guidelines. These safeguards, while essential for preserving human dignity and reproductive rights, create substantial challenges for direct investigations of human reproductive biology, especially for elucidating the cellular and molecular mechanisms underlying sperm–oocyte interactions.

At fertilization, mature gametes exhibit minimal transcriptional and translational activities. As a consequence, sperm–oocyte interactions rely almost entirely on a preassembled repertoire of proteins synthesized earlier during gametogenesis, leaving little opportunity for in vitro manipulation. In oocytes, however, functional perturbation is partially feasible through approaches such as mRNA microinjection or electroporation (gain-of-function) or ubiquitin-mediated targeted protein degradation (loss-of-function). These approaches, however, remain constrained by the temporal expression window and subcellular localization of the target protein (Kato et al., 2016; Clift et al., 2017, 2018; Nozawa et al., 2018; Dai et al., 2019c; Satouh et al., 2024). Earlier germ cells are more amenable to genome editing; yet, there is still no reliable method for efficiently and stably producing fully competent gametes in culture dishes (Saitou and Hayashi, 2021; Malo et al., 2024).

Genetically modified mouse models

To circumvent these limitations, reproductive biologists have turned extensively to genetically modified mouse models, a strategy that has drastically expanded with the advent of CRISPR/Cas9-based genome editing (Young et al., 2015; Scott and Gruzdev, 2019). This technology has significantly reduced the time, cost, and labor for generating genome-engineered mice while improving editing precision (Lu et al., 2019; Imai et al., 2022). Over the past decade, such models have been instrumental for identifying novel factors involved in distinct stages of sperm–oocyte interactions (Fujihara et al., 2020; Noda et al., 2020, 2022; Lu et al., 2023). Moreover, transgenesis, via either random genomic integration or CRISPR/Cas9-mediated homology recombination, has enabled targeted insertion, deletion, and substitution of specific DNA sequences, facilitating detailed mechanistic dissection of molecular events at fertilization. Applications include mutagenesis to probe the function of protein domains or residues, epitope or fluorescent tagging for localization analysis and affinity purification, fusion with biotin ligases for proximity labeling, degron tagging for inducible and selective protein degradation, and integration of synthetic circuits to modulate or monitor protein activity (Satouh et al., 2012; Shimada et al., 2021; Oura et al., 2022; Kaneda et al., 2025).

Reproductive disorders are often multifactorial, and genetic variants identified by whole-exome sequencing are not necessarily the definitive cause of infertility. To address this complexity, researchers have generated mouse models carrying patient-derived mutations linked to male infertility phenotypes such as oligozoospermia, asthenozoospermia, teratozoospermia, or oligoasthenoteratozoospermia, as well as female infertility conditions such as aberrant oocyte formation, maturation, or activation (Jiao et al., 2021; Liu et al., 2021; Maddirevula et al., 2022; Zhang et al., 2022; Muroňová et al., 2024; Ye et al., 2024).

Humanized mouse models

Despite their transformative value in reproductive science, mouse models exhibit inherent translational limitations. Discrepancies between human and mouse fertilization, arising from reproductive isolation, restrict the direct translation of mouse discoveries to humans (Gharib and Robinson-Rechavi, 2011). This underlines a need for alternative experimental approaches to bridge the translational gap.

Humanized mouse models are widely employed in biomedical research to assess the functions of human genes, or the biological consequences of patient-derived mutations within an in vivo physiological context (Zhu et al., 2019). In the field of fertilization, humanized mice have been intensively utilized to investigate the molecular mechanisms underlying ZP maturation, sperm–ZP interaction, and species-specific ZP recognition. Mice have been engineered to express full-length human ZP proteins or mouse ZP proteins modified to incorporate functional domains of the human orthologs. A particularly notable humanized model, in which mouse ZP1–3 proteins were knocked out and replaced with a complete set of human ZP proteins, demonstrated that mouse oocytes encapsulated by human ZP can be recognized and penetrated by human spermatozoa (Baibakov et al., 2012).

However, such transgenic mice show female subfertility due to species incompatibility at sperm–ZP recognition (Baibakov et al., 2012), making these lines difficult to maintain and limiting their utility as routine in vitro platforms for studying human sperm–ZP interaction. Similar challenges are anticipated for mouse models expressing human proteins involved in sperm–oolemma interaction, where fertilization efficiency is likely to be compromised due to disrupted endogenous molecular machinery. To enable modeling of human sperm–oolemma interaction while preserving fertility in mice, we envisage leveraging protein engineering, whereby ligands are rationally redesigned to accommodate interactions with both human and mouse receptors (Cao et al., 2022). For example, introduction of an L141Y substitution into mouse JUNO, mimicking the corresponding residue in hamster JUNO, results in significantly enhanced affinity for mouse IZUMO1 (Lu et al., 2025). This finding raises the possibility of developing ‘hamsterized’ mouse models with reduced species incompatibility at the sperm–oolemma interface, providing a promising framework for dissecting the molecular principles underlying human gamete fusion, especially when combined with mutagenesis and inhibitory assays.

Protein engineering and inhibition assays

Viral infection is orchestrated by sequential, spatiotemporal dynamic protein–protein interactions from initial attachment to final membrane fusion, conceptually resembling sperm–oolemma interaction during fertilization (Clark, 2018). Protein engineering, the deliberate modification or production of proteins with tailored properties, has emerged as a central toolkit for dissecting and perturbing virus–host interactions. Engineered viral fusogens and receptors have enabled precise mapping of receptor binding interfaces, identification of fusogenic motifs, and resolving of structural dynamics at near-atomic resolution (Shah et al., 2022). These insights have directly informed the rational design of vaccines, monoclonal antibodies, and small-molecule inhibitors against viral pathogens (Rey and Lok, 2018; Bhutkar et al., 2022). Engineered viruses and viral fusogens are further repurposed for therapeutic applications such as targeted drug delivery and gene therapy (Mandalawatta et al., 2024; Wang et al., 2024a), highlighting the robustness and versatility of protein engineering in both basic research and biomedical innovation.

Inspired by these advances, we envision harnessing protein engineering to unravel the molecular interplay between sperm and oocyte through the following complementary strategies: (i) protein mutagenesis, such as site-directed mutagenesis, truncation, domain swapping, and alanine scanning, to pinpoint binding interfaces, functional motifs, and the role of post-translational modifications (Arming et al., 1997; Chen et al., 1998; Takahashi et al., 2001); (ii) recombinant protein expression to generate soluble ectodomains, peptides, and inhibitory antibodies for structural and binding analyses (Aydin et al., 2016; Tang et al., 2022; Vance et al., 2022); (iii) directed protein evolution to develop high-affinity ligands and antibodies capable of modulating or disrupting sperm–oocyte interactions (Lu et al., 2025); (iv) terminal fusion with epitope tags, fluorescent proteins, or biotin ligases for visualizing protein localization, monitoring dynamics in live cells, and profiling interactomes (Satouh et al., 2012; Kaneda et al., 2025); and (v) targeted protein degradation to selectively eliminate fertilization factors for direct investigation of their physiological functions (Clift et al., 2018). These protein engineering approaches, in conjunction with inhibition assays employing recombinant proteins, transfected cultured cells, gametes, or their combinations (Tang et al., 2022; Lu et al., 2025), are expected to provide deeper insights into the molecular basis of sperm–oocyte interactions.

Xenospecies IVF assays

The human sperm penetration assay, commonly known as the hamster test, is a conventional method for evaluating the fertilizing potential of human spermatozoa in clinical settings by leveraging the nature that ZP-free golden hamster (Mesocricetus auratus) oocytes can be fertilized by human sperm. After insemination, the number of internalized sperm heads is assessed by gently pressing the oocytes with a coverslip and observing them under a phase contrast microscope. This technique was initially included in the World Health Organization (WHO) laboratory manual for the examination and processing of human semen, but it was removed from the latest 6th edition published in 2021 due to its obsolete nature (WHO, 2021). Indeed, several reports have pointed out the insufficient reliability, poor consistency, and limited clinical relevance of the hamster test to predict the outcomes of natural conception or IVF (Jane Rogers, 1985; Mao and Grimes, 1988). These drawbacks likely originate from the following factors: (i) human sperm are extremely heterogeneous even within a single ejaculate; (ii) semen quality is readily influenced by numerous external elements such as the donor’s health status and duration of sexual abstinence; (iii) sperm fertilizing ability does not necessarily correlate with genome integrity; and (iv) female factors also play an important role in the success of fertilization and embryogenesis.

Despite the concerns regarding its clinical significance, we have demonstrated that the hamster test remains a robust experimental tool for studying human sperm–oolemma interaction in vitro (Tang et al., 2022; Lu et al., 2025). To derive reproducible outcomes between technical replicates, we purify human sperm from fresh ejaculates by Percoll gradients, isolate the fraction with 100% motility, and incubate sperm in Biggers, Whitten and Whittingham medium containing bicarbonate, fetal bovine serum, and calcium ionophore A23187 to stimulate capacitation and the AR, before inseminating ZP-free hamster oocytes. Coupling the optimized hamster test with inhibitory recombinant antibodies, we have gained initial mechanistic insight into the role of human TMEM95 in sperm–oolemma binding and fusion (Tang et al., 2022). Building upon this work, we propose that xenospecies fertilization assays, in combination with protein engineering technologies, can serve as a robust platform to mechanistically dissect the function of individual human sperm proteins during fertilization. Nevertheless, findings derived from these assays should be interpreted with caution, as the hamster oolemma may express a repertoire of surface proteins that are absent from the human oocyte or exhibit distinct species-specific functions.

In addition to the hamster test, several other xenospecies and intraspecies IVF assays have emerged as functional tools to interrogate the causes of ICSI failure in clinical settings. These include mouse and bovine oocyte activation tests (MOAT and BOAT), in which a human spermatozoon is introduced into a mouse or bovine oocyte by ICSI to evaluate pronuclear formation (Vanden Meerschaut et al., 2013; Durban et al., 2015), as well as mouse and human oocyte calcium analyses (MOCA and HOCA), where a human spermatozoon is injected into a mouse or an in vitro matured human oocyte to assess sperm-induced oocyte activation by monitoring intracellular calcium oscillations, the hallmark of successful oocyte activation (Cardona Barberán et al., 2024; Parrella et al., 2024). Beyond their roles in supporting the diagnosis of fertilization failure associated with compromised oocyte activation, these xenospecies and intraspecies fertilization assays could be further developed into robust platforms for the mechanistic dissection of post-fertilization signaling events underlying oocyte activation and pronuclear formation in humans.

High-throughput interactome screening

The interactions between sperm and oocyte are mediated by a complex network of ligand–receptor pairs. For many years, these protein–protein interactions were investigated using low-throughput genetics and structural biology approaches. The recent advent of AlphaFold2 and its derivatives, such as AlphaFold-Multimer and AlphaMissense, represents a revolutionary shift in protein biology (Jumper et al., 2021; Cheng et al., 2023; Abramson et al., 2024; Omidi et al., 2024). These artificial intelligence-powered platforms permit high-confidence structural predictions of both individual and multimeric proteins directly from their amino acid sequences, bypassing long-lasting technical difficulties related to protein expression, purification, and crystallization.

By incorporating AlphaFold2-based structural modeling into the experimental pipeline, researchers have derived novel mechanistic insights into sperm–oocyte interactions that were previously inaccessible (Deneke et al., 2024; Elofsson et al., 2024; Nishio et al., 2024). Moving forward, we envision utilizing AlphaFold to construct comprehensive, high-resolution atlases of cis and trans protein–protein interactions within and between sperm and oocyte. Mapping these interactome networks with the respective stages of sperm–oocyte interactions will advance protein function prediction, facilitate rational interpretation of infertility-related genetic variants, and enhance genotype–phenotype correlations.

Despite the advances of in silico complex modeling, experimental validation remains indispensable. Membrane proteins are notoriously difficult to solubilize, and their extracellular interactions are often weak and transient, complicating detection through conventional affinity-based methods. To mitigate these challenges, several in vitro interactome screening systems have been developed. A representative example is AVEXIS, which enabled the landmark discovery of JUNO by using a pentamerized IZUMO1 ectodomain to enhance avidity (Bushell et al., 2008; Bianchi and Wright, 2016). However, AVEXIS requires the expression and purification of oligomerized recombinant proteins, often constituting a major bottleneck. To enable more sensitive screening, synthetic molecular circuits can be incorporated as downstream reporters of protein–protein interactions. Examples include the synthetic Notch receptor (Morsut et al., 2016), binding-induced inhibition of Gaussia luciferase (Galaway and Wright, 2020), as well as other binding-induced transcriptional activation systems (Yang et al., 2025b). In addition, bait proteins can be displayed on lentiviruses or virus-like particles, providing an alternative to recombinant protein expression and purification (Cao et al., 2021).

Artificial human gametes

Producing functional gametes in vitro is one of the most rapidly progressing yet technically challenging areas in stem cell biology. Artificial gametes are highly demanded in reproductive biology, as they provide ethically permissible and scientifically valid systems to study gametogenesis, fertilization, and reproductive disorders (Villalba, 2024; Kaltsas et al., 2025).

Previous works have demonstrated that mouse pluripotent stem cells can be induced into primordial germ cell-like cells, which, upon co-culturing with fetal ovarian somatic cells, differentiate into mature oocytes capable of producing viable offspring, albeit at low efficiency (Hayashi et al., 2012; Hikabe et al., 2016; Yoshino et al., 2021). In the male germline, organotypic culture of mouse and rat seminiferous tubules remains the only method that supports the production of fully functional spermatozoa in vitro, again with limited efficiency and under complex culture conditions (Sato et al., 2011; Matsumura et al., 2023; Kamoshita et al., 2025). With optimized and simplified culture protocols, we anticipate that artificial gametes will have higher fertilization competence, establishing a renewable and customizable platform for high-throughput interactome mapping, protein functional analyses, and screening of small molecules and antibodies targeting sperm–oocyte interactions.

Advanced imaging technologies

A broad range of imaging technologies has been applied to observe mammalian gametes and their interactions across multiple spatial and temporal scales, ranging from in vivo dynamics to molecular architecture. At the organismal and tissue levels, live imaging approaches using standard light and fluorescent microscopy and optical coherence tomography have enabled the visualization of dynamic processes such as sperm migration within the female reproductive tract (Fujihara et al., 2019; Umezu and Larina, 2023) and sperm traversal through the cumulus matrix and the ZP (Jin et al., 2011; Morohoshi et al., 2023). At the cellular level, scanning electron microscopy and spinning disk or laser scanning confocal microscopy have allowed real-time or static observations of gamete morphology, behavior, interactions, and surface protein dynamics, including the 3-dimensional structure of the ZP (Ishikawa-Yamauchi et al., 2024), acrosome exocytosis (Lu et al., 2023; Morohoshi et al., 2023), sperm–oocyte binding and fusion (Satouh et al., 2012; Inoue et al., 2025), pronuclear formation, and oocyte activation (Nozawa et al., 2018). At the subcellular and ultrastructural levels, transmission electron microscopy and super-resolution microscopy (e.g. stochastic optical reconstruction microscopy) have been widely used to examine detailed gamete structures, such as the cumulus oophorus and the ZP (Baena and Terasaki, 2019; Ishikawa-Yamauchi et al., 2024), oocyte microvillar (Żyłkiewicz et al., 2010; Benammar et al., 2017), sperm acrosome (Romarowski et al., 2018; Morohoshi et al., 2023; Kaneda et al., 2025), and the sperm–oolemma interface (Yanagimachi and Noda, 1970; Thompson et al., 1974; Sathananthan and Chen, 1986). Coupled with fluorescent protein tagging or immunofluorescence, these microscopic techniques have further enabled the precise localization of fertilization factors in both living and fixed cells.

Recent advances in cryogenic electron microscopy (cryo-EM) and cryogenic electron tomography (cryo-ET) further provide powerful means to resolve the structures of fertilization-related proteins, as well as their complexes, at unprecedented resolution and under near-native conditions. While cryo-EM can reveal near-atomic details of individual proteins and macromolecular assemblies without the need for crystallization, cryo-ET can reconstruct 3-dimensional tomograms of vitrified cells, thereby visualizing the spatial organization of protein complexes in situ. Indeed, recent cryo-EM and cryo-ET studies have already elucidated the architecture of protein monomers or multimers in gametes (Im et al., 2025; Shao et al., 2025), as well as large protein complexes and subcellular structures in the sperm manchette (Judernatz et al., 2025), central apparatus (Zhu et al., 2025), and flagellum (Lin et al., 2021; Gui et al., 2022; Zhao et al., 2022; Chen et al., 2023; Zhou et al., 2023), and in the oocyte cytoplasm (Jentoft et al., 2023). We expect that cryo-EM/ET will soon be exploited to resolve protein complexes at the gamete surface and to capture their higher-order assemblies and conformational changes during fertilization.

Development of novel contraceptive approaches

Developing safe, effective, and reversible contraceptives with minimal side effects has been a long-standing objective in reproductive medicine. While several hormonal contraceptives are commercially available, they are often associated with systemic side effects. However, to date, FDA-approved non-hormonal contraceptives exist only for women, whereas male options remain largely confined to preclinical stages, despite the identification of numerous promising sperm-specific targets over the last decade (Garcia and Matzuk, 2024; Mannowetz et al., 2025). Current efforts in male contraceptive development predominantly target spermiogenesis (Chang et al., 2021; Yu et al., 2021; Ku et al., 2024), sperm motility, hyperactivation, and the AR (O’Rand et al., 2018; Rennhack et al., 2018; Balbach et al., 2023; Lyon et al., 2023; Garcia and Matzuk, 2024).

Contraceptives targeting sperm–cumulus interaction

Contrary to chemoattractants, chemorepellents are physiological or pharmacological substances that drive sperm away from the source. In the Fallopian tubes, the fertilized oocyte releases zinc ions, which induce chemorepulsive behavior in surrounding sperm, providing an additional mechanism of polyspermy block (Fig. 3; Guidobaldi et al., 2017a; Eisenbach, 2025).

Synthetic progesterone receptor ligands (sPRLs) are chemicals that function on progesterone receptors to trigger either agonist or antagonist effects on progesterone signaling. Ulipristal acetate (UPA), an sPRL approved for emergency contraception, has been shown to induce repulsive motility patterns in human sperm at a concentration of 100 pg/ml (Guidobaldi et al., 2017a), suggesting that the contraceptive action of UPA may involve perturbing cumulus-mediated sperm chemotaxis. However, sPRLs are classified as hormonal contraceptives that can potentially modulate the endocrine system by interacting with hormone receptors. This emphasizes the need to identify non-hormonal chemorepellents, akin to zinc ions, that disrupt sperm chemotaxis without adverse hormonal effects.

The antimicrobial compound polyphenylene carboxymethylene (PPCM) has been formulated as a topical vaginal gel for both non-hormonal female contraception and sexually transmitted infection prevention and is close to human clinical trials (Weitzel et al., 2020). In preclinical tests, PPCM induces dose-dependent sperm inactivation characterized by premature acrosome loss (North et al., 2022). Further, PPCM inhibits hyaluronidase and acrosin (Weitzel et al., 2020; North et al., 2022), thereby preventing sperm from traversing the cumulus cell layer and the ZP.

Contraceptives targeting sperm–ZP interaction

The monoclonal antibody IE-3 is known to inhibit fertilization by targeting the N-terminal ZP-N1 domain of mouse ZP2, a region implicated in mediating both sperm–ZP binding and polyspermy block (Sun et al., 1999; Nishio et al., 2024). The structural basis of IE-3 recognition of ZP2 has recently been elucidated (Dioguardi et al., 2025), providing valuable insights into the rational design of ZP-targeting immunocontraceptives.

Contraceptive vaccines targeting ZP glycoproteins have been studied extensively, particularly in wildlife population control. Three vaccines, SpayVac, ZonaStat-H, and PZP-22, are in practical use for regulating fertility in wild animals such as wild horses, white-tailed deer, and marsupials (Gupta et al., 2011; Gupta, 2023). In primate models, recombinant ZP3 immunization efficiently suppressed fertility (Paterson et al., 1996; Kerr et al., 1998). Despite the promising efficacy of ZP-based contraceptives, their application in human contraception has been limited by potential ovarian pathology (e.g. oophoritis and accelerated primordial follicle depletion), highlighting a need for identifying highly specific ZP epitopes capable of stimulating a contraceptive immune response without compromising ovarian function (Gupta et al., 2011).

Contraceptives targeting sperm–oolemma interaction

The final step of fertilization, sperm–oolemma interaction, has also emerged as a promising target for contraceptive development. A chimeric recombinant protein carrying epitopes from IZUMO1, SPACA3, and SPAM1 elicited antibodies that significantly reduced fertility in mice (Mortazavi et al., 2021). Two small molecules against JUNO inhibited mouse IVF and impaired hamster oocyte penetration by human sperm. Interestingly, despite these molecules being predicted to target the IZUMO1-binding interface, they do not interfere with the IZUMO1–JUNO interaction in vitro (Stepanenko et al., 2022). In addition, the naturally occurring anti-IZUMO1 antibody, OBF13, has been shown to block fertilization in mice with a well-characterized inhibition mechanism (Lu et al., 2025). These structural insights directly inform how antibodies targeting the gamete fusion machinery could be engineered into highly specific non-hormonal contraceptives.

Concluding remarks and future perspectives

Over the past several decades, research on sperm–oocyte interactions has progressed drastically, evolving from descriptive observations to a mechanistic understanding of the molecular networks governing fertilization. Nevertheless, despite this remarkable progress, many fundamental questions remain unanswered. For instance, what are the principal sperm receptors of chemoattractants and ZP glycoproteins? What mechanisms potentiate species-specific recognition between gametes? Apart from IZUMO1–JUNO, what additional ligand–receptor pairs mediate sperm–oolemma adhesion and fusion? What is the bona fide fusogen that drives membrane fusion at fertilization?

Addressing these mysteries holds enormous promises beyond basic biology. A comprehensive understanding of the cellular and molecular basis of gamete recognition and fusion will lay the foundation for next-generation reproductive engineering and medicine. Potential applications may include the following: on-chip chemotaxis for high-quality sperm selection; fusogen-mediated, gamete-specific gene and drug delivery; engineered gamete surfaceomes for xenospecies fertilization; IVF and SUZI with immotile or immature sperm; and precision diagnostics for fertilization failure with novel molecular biomarkers.

Such capabilities to manipulate gamete interactions in vitro would revolutionize ARTs, thereby offering safe, efficient, and personalized treatments for infertile patients. By continuing to decode the molecular dialogue between sperm and oocyte, reproductive science is poised to enter a new era, in which mechanistic discoveries at the bench translate into diagnostic innovations and therapeutic strategies in the clinic. Clinical insights, in turn, will feed back into basic science, guiding refined research directions and fostering the development of more advanced reproductive engineering technologies.

Data availability

No new data were generated or analyzed in support of this research.

Authors’ roles

Y.L. and S.T. conceptualized, designed, and supervised the work. All authors reviewed the literature and wrote the manuscript.

Funding

This work was supported by: the World Premier International Research Center Initiative (WPI); the Japan Society for the Promotion of Science (JSPS) Doctoral Course Research Fellowship (to H.-Y.C.); the Gruber Science Fellowship at Yale from the Gruber Foundation (to T.G.); the National Science Foundation Graduate Research Fellowship Program DGE-2139841 (to T.G.); the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health grant R00HD104924 (to S.T.); the David Sokal Innovation Award from the Male Contraceptive Initiative 2024-303 (to S.T.); the JSPS Grant-in-Aid for Scientific Research (B) JP24K02033 (to Y.L.); the Life Science Research Grant from the Takeda Science Foundation 2024032851 (to Y.L.), the Research Encouragement Grant from the Asahi Glass Foundation (to Y.L.), the PRIMe Joint Research Grant 1301800004 (to Y.L.), and the Research Grant from the Mochida Memorial Foundation for Medical and Pharmaceutical Research (to Y.L.).

Conflict of interest

The authors declare no conflict of interest.