The biology and dynamics of mammalian cortical granules

Mammalian fertilization is a sequence of coordinated events involving multiple steps of mutual recognitions between haploid male and female gametes. In mammals, fertilizable oocytes are ovulated along with the first polar body, as well as the extracellular matrix (ECM) zona pellucid (ZP) which is composed of three glycoproteins, ZP1, ZP2, and ZP3 [1, 2] and the cumulus oophorus which is made of several layers of ovarian follicular granulose cells embedded in hyaluronic acid-containing ECM (Figure 1). Passage of a sperm through these semi-permeable investments surrounding an oocyte is absolutely essential in the initiation of fertilization. First, sperm need to pass through the cumulus oophorus and they do so by disrupting the macromolecular structure of the matrix [3] and by sperm hyaluronidase activity [4]. Next, formation of a pathway through the ZP is triggered by the recognition of a cognate receptor on sperm and the ZP3 glycoprotein of the oocyte by protein-carbohydrate recognition [5, 6], followed by the induction of proteolytic enzyme release from the sperm head termed acrosome reaction. Sperm penetration through the ZP is subsequently facilitated by the binding between acrosome reacted sperm and ZP2 glycoproteins [7]. Most studies of mammalian fertilization support the model described above regarding gamete recognition and interaction; however, a recent report has revealed that ZP3 glycoprotein is not essential and sufficient to induce sperm acrosome reaction while binding to ZP2 is account for sperm-egg recognition during fertilization in mice [8]. Whether this model for gamete recognition is adequate for all mammals still needs to be determined experimentally.

Prior to fertilization, a sperm needs to swim through the perivitelline space (PVS) and fuses with the oocyte through a fertilin/integrin- mediated adhesion process [9‐15]. Following gamete fusion, the fertilized oocyte resumes meiosis and protrudes the second polar body, which marks the end of fertilization process. However, a successful fertilization is not guaranteed unless preventative mechanism(s) inhibiting fertilization by more than one sperm, a condition known as polyspermy, is established immediately following gamete fusion since polyploidy is fatal to an embryo. Physiological polyspermy or penetration of the oocyte by more than a single sperm occurs in numerous species including insects, reptiles, and birds, while mammalian polyspermic fertilization is considered abnormal, which leads to developmental failure of the zygote and spontaneous abortion in humans. A recent study of porcine oocytes reveals that pig is a species with a high incidence of polyspermy under physiological conditions or in vitro fertilization [16]. Interestingly, porcine oocyte cytoplasm has the capability to remove the accessory sperm [17] and some poly-pronuclear pig eggs can develop to term if accessory sperm do not interrupt the embryo genome [18], demonstrating that this species depends on unique mechanisms to deal with the occurrence of polyspermy. Nevertheless, a defense against polyspermic fertilization needs to be established rapidly after fertilization under most circumstances in mammals.

In mammals, fertilization occurs internally (e.g. cervix, uterotubal junction) and female reproductive tract limits the arrival of sperm to the fertilization site [19]. In mice, 50 million sperms are ejaculated and only 100 to 200 sperm actually reach the site of ovulated oocytes [20], supporting by the observation that in vitro fertilization has higher incidence of polyspermy in animals and humans [16, 21, 22] and the frequency of polyspermic fertilization is directly related to the sperm concentration [23]. However, this mechanism is not completely dependable, and a more stable and long lasting system against polyspermic fertilization is required. The oocyte plasma membrane block to polyspermy has been described in several mammalian species such as humans, mice, and rabbits [24‐26]. In some invertebrates, a depolarization of the plasma membrane potential prevents sperm from fusing with the oocytes immediately after fertilization [27, 28]; however, such an electrical fast block at the level of plasma membrane is not present in mammals. Interestingly, a prior study of in vitro fertilization of bovine oocytes has suggested that the cumulus oophorus also plays a role in blocking polyspermy by engulfing numerous sperm.

In mammals, the primary mechanism in preventing polyspermic fertilization involves biochemical modifications of the surrounding of fertilized oocytes as well as the gamete itself; depending on the species, polyspermy block resides at the zona pellucida, the oolemma, or the perivitelline space [29, 30]. Generally, the establishment of polyspermy block is initiated and achieved by molecules of cortical granule origin following gamete fusion. Cortical granules are membrane bound secretory organelles located in the cortex of unfertilized oocytes in many invertebrates and vertebrates [31‐44]. Mammalian cortical granules, which were first described by C.R. Austin in hamster oocytes using phase contrast microscopy, are derived from Golgi complexes during oocyte growth [41, 45]. Following fertilization, cortical granules undergo exocytosis to release their contents into the perivitelline space (Figure 1). This secretory process is calcium dependent and is known as the cortical reaction [46‐49]. The released cortical granule proteins are responsible for blocking polyspermy by modifying the oocytes' extracellular matrices, such as the zona pellucida in mammals, the vitelline envelope in echinoderms, and the fertilization layers in amphibians [35, 42, 44, 50‐55]. Although other biochemical and structural changes in ZP following cortical granule exocytosis are inferred, only the proteolytic cleavage of ZP2 to a form called ZP2f has been experimentally described [56]. This modification of ZP, termed zona reaction or zona hardening, exerts polyspermic block by inhibiting additional sperm to bind ZP and hamper bound sperm to penetrate this ECM [57]. Furthermore, the accumulation of cortical granule components on oolemma and PVS, as well as the fusion of cortical granule membrane with oolemma may play roles in the establishment of a plasma membrane and PVS block to polyspermy by hindering gamete interaction or modifying an incoming sperm [29, 44].

Mammalian cortical granules range in size from 0.2 um to 0.6 um in diameter and appear morphologically similar to each other at the ultrastructural level [41, 58‐60]. Cortical granules are distinct and different from most other regulatory secretory organelles in that they are not renewed once released. These granules are only synthesized in female germ cells and transform an egg upon sperm entry; therefore, this unique organelle has inherent interest for our understanding of the fertilization process. The following review will focus on the biology and dynamics of cortical granules in mammals; topics such as the formation, distribution, pre-fertilization release of cortical granules, molecular mechanisms that regulate granule secretion, granule contents (including biological functions if available), as well as granule population heterogeneity, will be discussed.

Mammalian cortical granules first appear in the early stages of oocyte growth; however, the exact window of time in which cortical granules are synthesized is different from species to species [41]. In rat and mouse, cortical granules are first seen in the unilaminar follicle [46, 61‐63], whereas in the human, monkey, hamsters, and rabbit, they first appear in multilayered follicles [46, 64‐68]. During the early stage of follicular growth, Golgi complexes undergo hypertrophy and proliferation, and the formation of cortical granules from Golgi complexes occurs at this stage [41, 46, 69]. At first, small vesicles are formed from hypertrophied Golgi complexes that migrate toward the subcortical region of the oocytes. These vesicles then coalesce to form mature cortical granules that eventually separate from Golgi complexes [41]. Subcellular structures containing both granules and Golgi complex-like organelles have been observed in growing mouse oocytes [70], and these structures likely represent the sites of cortical granule synthesis. The production of cortical granule in mammalian oocytes is a continuous process, and newly synthesized granules are translocated to the cortex up until the time of ovulation [41, 46, 61, 65, 70‐75]. Several lines of evidence show that the granule migration is a cytoskeleton-dependent process and microfilaments are required for this cortical translocation in both non-mammalian and mammalian animal models, such as sea urchin, porcine, human, and mouse [76‐80]. In mammalian oocytes, the migration of cortical granules is an important step in cytoplasmic maturation and has been used routinely as a criterion in assessing the maturity and organelle organization of developing oocytes [81].

Translocated cortical granules are evenly distributed in the cortex of unfertilized oocytes; however, some granules undergo complex changes in distribution prior to ovulation [82‐84], in results of areas devoid of cortical granules (Figure 2). This meiotic spindle-associated granule-free area was first noted 30 years ago in mouse oocytes, although the extent of this domain apparently was not fully appreciated [46, 71]. Recent studies have reported the development of two distinct cortical granule free domains (CGFDs) over the metaphase I and metaphase II spindles in hamster and mouse oocytes, respectively [60, 70, 82, 84‐88]. Interestingly, CGFD appears to be unique to rodents since feline, equine, bovine, porcine, and human oocytes lack such a domain [75, 89‐93]. Several lines of evidence show that redistribution and/or exocytosis of cortical granules during meiotic maturation are likely mechanisms for the formation of CGFDs, supported by the observations of an increase in granule density in the periphery of the CGFD and a decrease in the total number of cortical granules during meiotic maturation [82, 83, 85]. In mice, redistribution of granules appear to be the main factor contributing to development of the first CGFD over metaphase I spindle since cortical granule exocytosis is not evident this time [88] and the formation of this domain is not inhibited in the presence of BAPTA, a Ca²⁺ chelator that prevents cortical granule exocytosis [87]. However, a study on hamster oocytes reported that small extend of cortical granule release was observed during metaphase I, and both granule migration and exocytosis are claimed to be required for generating first CGFD [82]. The exact mechanisms responsible for the first CGFD formation in rodents may be different from species to species and will need to be determined experimentally. The second CGFD formed over the spindle in metaphase II oocytes is thought largely due to redistribution of cortical granules, supported by the observation of increased granule density at the edge of the granule-containing domain [87, 88]. Studies in mice also reveal that this cortical granule migration during metaphase I and II transition is chromatin-mediated [87]; however, disruption of the spindle with microtubule inhibitors does not prevent formation of a CGFD over the chromatin [80]. While the formation of both cortical granule free domains appears to occur by granule redistribution, exocytosis has been detected during the metaphase I to metaphase II transition at both light and electron microscopy levels [82, 86, 88]. Interestingly, the pre-fertilization exocytotic event documented during this meiotic transition period does not participate in the second CGFD development and only contributes to the enlargement of the granule-free area since calcium chelator does not abolish the domain formation [87]. Taken together, present findings demonstrate that granule redistribution is the primary mechanism responsible for cortical granule free domain formation in mammals. Changes in cortical granule distribution and formation of CGFDs in mouse oocytes are summarized in Figure 3. Prior study of mouse oocytes reveals that the pre-fertilization exocytosis of cortical granules during metaphase I and II transition involves a specific population of granules [88]. In addition, this exocytotic event takes place during polar body extrusion and only in the cleavage furrow. The finding is supported by the observation that the mean total numbers of cortical granules per oocyte decreases after polar body extrusion (from ~7400 to 4100) [85]. Moreover, an electron microscopic study of cortical granules revealed that granules undergoing exocytosis prior to gamete fusion were predominantly confined to an ultrastructurally "light" population [60]. Aside to the pre-fertilization release aforementioned, small populations of cortical granules have been described to undergo exocytosis at different time prior to fertilization. One wave of cortical granules release occurs during metaphase I and has been reported only once in thin sections of hamster oocytes [82]. Another wave of the release, which has only been observed ultrastructurally, takes place after addition of sperm but before sperm penetration when oocytes are inseminated in vitro [60]. However, comparable exocytosis of granules in other mammals has not been described.

Based on the preceding review, we know that pre-fertilization exocytosis of cortical granules, as well as formation of cortical granule free domains, that take place at a specific time and place are well documented in some mammals. Although the exact biological significance and functions of CGFDs and pre-fertilization release of cortical granules are elusive, several hypotheses are available. First, the formation of CGFDs ensures that the first and second polar bodies to be extruded with few cortical granules in them; therefore allowing maximum number of cortical granules to exert their biological function(s) at the right window of time. Second, the pre-fertilization release of cortical granules during the metaphase I and II transition may be involved in modification of zona pellucida, perivitelline space, and/or oolemma directly above the site of the release, thereby minimizing the likelihood for sperm to bind and penetrate oocytes through this area where maternal chromatin is located. Establishment of this local block to sperm entry will reduce the possibility of damaging maternal genetic materials when paternal chromatin undergoes decondensation following gamete fusion. This may be particular importance in rodents in which the oocyte surface area is typically 25% that of larger mammals and therefore the risk is much higher for sperm entry in this region. This hypothesis is supported by findings showing that sperm-oocyte fusion occurs less frequently in the CGFD [60, 94]. Lastly, the limited cortical granule exocytosis occurs during oocyte maturation and prior to fertilization may condition the zona by altering the physico-chemical properties slightly in such a way that the zona can be penetrated only by a very "strong" sperm [82, 86]. This postulation is supported by finding from an early report on the loss of cortical granules and the development of CGFD in oocytes matured in vitro cultures are accompanied by small degree of ZP2 to ZP2f conversion [83]. In conclusion, most but not all mammalian cortical granules under exocytosis following gamete membrane fusion [41]. Despite the general function of cortical granules in establishing block to polyspermy, it is highly probable that CGFDs and pre-fertilization release perform unique biological function(s) since they occur in a specific location at a specific time. However, the exact physiological functioning of these events still needs to be determined experimentally.

The total number of mammalian cortical granule proteins is not known; however, it has been estimated to be between four and fourteen [56, 119, 120]. The amount of cortical granule material available in mammalian oocytes (estimated to be picogram quantities/oocyte) is scanty, which is a major factor that hampers studies on granule constituents. The lack of specific probes to mammalian cortical granules makes the characterization of cortical granules even more difficult. Nevertheless, several molecules have been inferred or demonstrated to be of cortical granule origin. Components identified and characterized in mammalian cortical granules will be described in the following section.

Different types of cortical granules have been described in unfertilized oocytes of some species such as fish, amphibians, crustaceans, echinoderms, and mammals [34, 37‐40, 58, 60, 74, 164‐167]. In the next section, works focused on granule population in non-mammalian species will be presented first, followed by studies in mammals, to provide a more fundamental understanding and broader perspective on cortical granule heterogeneity.

In the Zebrafish, Brachydanio rerio, two types of cortical granules, the dark type (electron-dense) and the light type (electron-lucent), have been identified at the electron microscopy level [38]. Differential cytochemical staining with either Alcian blue or toluidine blue also confirms the presence of these two types of cortical granules in unfertilized oocytes of Zebrafish [38]. The killifish, Fundulus heteroclitus, has two populations of cortical granules; at the ultrastructural level, one population contains filamentous and/or granular material, while the other does not [39]. In the frog, Xenopus laevis, two types of cortical granules are differentially distributed in unfertilized oocytes [34]. One type has a homogeneous matrix of moderate electron density and is located in the animal hemisphere, while the other type has a loose flocculent material and is located in the vegetal hemisphere [34]. In the horseshoe crab, Limulus polyphemus L., two types of cortical granules have been described, one appears to be electron-translucent and the other to be electron-dense [37]. In lobsters, Homarus americanus and H gammarus, four types of cortical granules, that are distinguishable based on their ultrastructural appearance, are present in the cortex of ovarian oocytes and are referred to as high-density, low-density, moderately dense, and ring cortical granules. The contents of each type of granule are released in sequence from the cortex of the oocyte into the perivitelline space following fertilization [40]. In the sea urchin, Strongylocentrotus purpuratus, the presence of biochemically distinct populations of cortical granules has been shown based on the immunocytochemical observations. One population of cortical granules contains an antigen recognized by two specific antibodies, 1G8 and B2C2, while the other population does not [164]. In another species of the sea urchin, Lytechinus variagatus, four types of granules have been identified immunocytochemically in the unfertilized oocytes [165]. These granules ultimately migrate to the cortex and sequentially release their contents into the perivitelline space upon fertilization.

Mammalian cortical granules are generally similar in morphology [46, 58, 67, 168, 169] although dark and light granules have been reported based on their ultrastructural appearance [58, 60, 74, 166, 167]. Unlike some species such as fish, amphibians, crustaceans, echinoderms, heterogeneity of cortical granules in mammals is well documented but poorly understood. In humans, two populations of cortical granules of which one has a homogeneous electron-dense core and the other has granular or fluffy contents are present in unfertilized oocytes [166, 167]. In pig, hamster, and mouse oocytes, dark and light forms of cortical granules have also been described based on their electron density [58, 60, 74]. It is not known if this difference in electron density represents different stages in granule maturation, different types of granules, or different stages in exocytosis [58, 60, 170]. Not till recently, biochemically distinct populations of cortical granules are demonstrated cytochemically in unfertilized mouse oocytes. By taking advantage of two probes specific to mammalian cortical granules, lectin LCA (mannosyl glycoconjugate-binding) and the ABL₂ antibody (PAD protein-binding), two populations of cortical granules, of which one contains PAD and the other lacks of this molecule, are shown to be present in unfertilized oocytes [88]. However, whether the heterogeneity in granule populations corresponds to different biological functioning is still a question needs to be answered.

To date, only several partially characterized molecules have been reported to be of cortical granule origin in mammals. Cortical granules of invertebrates (e.g. sea urchins and starfish) and some vertebrates (e.g. fish and frogs) are better studied as these model animals served rich source for experimentation, due to the ease in oocyte isolation and manipulation. Our understanding of mammalian cortical granule biochemistry and biogenesis has been hampered for several reasons. First, characterization of granule components is a laborious task due to the paucity of mammalian cortical granule material available per oocyte and low number of oocytes retrievable per animal. It is estimated that each oocyte (i.e. in mice) contains ~20 ng of total proteins and only 2.5% to 5% of these proteins are from cortical granules [119]. Second, the lack of specific probes for mammalian cortical granules makes analytical tests difficult to perform. Studies of cortical granules in humans can be even more troublesome since experimental samples are mostly obtained from volunteer donors, who undergo reproductive assistant treatments, and the number and quality of oocytes collected cannot always be optimal. Although non-mammalian animals serve as excellent models and provide research basis for mammalian species, one should keep in mind that direct extrapolation of information between mammalian and other animal modes requires caution because cellular and molecular differences among different species can affect the experimental outcome.

From the preceding review, we know that mammalian cortical granules are secretory organelles, which are stimulated to undergo calcium dependent and SNARE protein-mediated exocytosis upon gamete fusion, in result of a block to polyspermy. Evolved and extended from the historical viewpoint, mammalian cortical granules are more complex and have more functions than previously realized. Biochemical heterogeneity and complex redistribution of granule in the cortex of unfertilized oocytes prior to ovulation are documented. A small population of granules is evident to exocytose prior to sperm penetration; a number of granule components are reported to function following fertilization in regulating embryonic cleavage and preimplantation development. In addition to providing biological basis to better understand reproduction process, new findings of mammalian cortical granules have made scientists to start apprehending their clinical relevance. Biological parameters of cortical granules, such as granule synthesis and translation to the cortex, are used routinely as criterion to determine the developmental stage of human oocyte grown in vivo or in vitro. Timely release of these granules is essential for a successful fertilization to occur, especially for individuals who are reproductively challenged and require artificial assistance. Interestingly, low fertilization rate was observed following cryopreservation of mouse oocytes and this low rate was attributed to cryopreservant-induced premature cortical granule discharge [171, 172]. Previous study of the cortical granules in unfertilized human oocytes from failed IVF trials also showed a significant decrease in number of granules, suggesting that premature release of cortical granules and early establishment of block to polyspermy prior to insemination may be the cause for some unsuccessful cases of in vitro fertilization [173].

Cortical granules are first identified in sea urchins by Harvey in 1910 [174], inspired by Derbès' observation of the fertilization envelope 150 years ago. Not till four decades later, mammalian cortical granules were documented in hamster oocytes by Austin [45]. Fifty five years of time is full of excitement as we have witnessed many technological breakthroughs in reproductive biology and fascinating scientific discoveries of mammalian cortical granules. However, this unique organelle is far from being fully appreciated and a broader perspective on its biochemistry and functional role(s) should be developed and presented. On the quest to find answers to these questions we encounter in studies of mammalian cortical granules, continuous efforts made by scientists in the field of reproductive biology should led us a step closer to better understanding the nature of cortical granules and embracing the beauty of this organelle.