An interaction between two proteins — Izumo1, which is produced by sperm, and Juno, its receptor on eggs — enables human fertilization. However, the details of this interaction have been elusive. In two papers, Aydin et al.1 (page 562) and Ohto et al.2 (page 566) present the structures of Izumo1, Juno and the two proteins in complex, determined by X-ray crystallography at atomic-level resolution.
Following human copulation, motile sperm move towards eggs in the female's Fallopian tubes. The acidic environment of the female reproductive tract triggers an activation step, in which sperm become hypermobile and penetrate the outer protective layer of the egg. A second activation step occurs when or shortly before the sperm binds to the zona pellucida — the tough inner layer that surrounds the egg. During this step, the acrosome — an organelle at the tip of the sperm head — releases digestive enzymes that break down the zona pellucida. This acrosome reaction allows the sperm to bind to Juno on the egg membrane, following which the two cells' membranes fuse and the cells merge. In turn, the egg releases enzymes that crosslink glycoproteins of the zona pellucida to make it impenetrable, preventing fertilization by multiple sperm (polyspermy)3,4.
Izumo1, which is named after a Japanese marriage shrine, was first identified in 2005 by its binding to an antibody that blocked sperm–egg fusion5. The protein remains concealed intracellularly in the inner acrosomal membrane until the acrosome reaction occurs, when the inner membrane becomes part of the cell surface. Juno, named after the Roman goddess of love and marriage, was identified almost a decade later6 as a membrane-anchored protein that is required for female fertility, sperm–egg membrane fusion, and egg binding by Izumo1. One structure of mouse Juno has been published this year7, and another will soon be published in Nature Communications8. But structures of the extracellular domain of Izumo1, human Juno and the Juno–Izumo1 complex have remained unknown.
Juno was originally called folate receptor-δ, and shares close to 60% amino-acid identity with human folate receptors6 (receptors for folic acid and its derivatives). The structures of Juno from mice7,8 and the current studies reveal that the protein has an almost identical fold to that of folate receptors9,10: globular, stabilized by eight disulfide bonds (S–S) and with a deep, ligand-binding pocket. But several key amino-acid residues in Juno's ligand-binding pocket differ from those of the folate receptors, consistent with the fact that Juno cannot bind folates4.
Both groups find that the extracellular region of Izumo1 has two domains — a four-helix bundle at the protein's amino terminus and an immunoglobulin-like domain at the carboxy terminus. The two domains are connected by a hinge region consisting of a β-hairpin structure with loops at either end that are anchored to the two folded domains by disulfide bonds. The researchers show that Izumo1 and Juno form a high-affinity complex in a 1:1 ratio. A surface of Juno distant from the pocket binds the outside of the hinge and makes contacts with both Izumo1 domains (Fig. 1).
Ohto and colleagues crystallized structures of free and Juno-bound Izumo1 in the same elongated conformation. By contrast, Aydin et al. report that Izumo1 alone adopts a boomerang-shaped conformation, in which the hinge is almost 40° more closed than that of Juno-bound Izumo1. The authors validated the approximate shape using a technique known as small-angle X-ray scattering. This provides low-resolution structural information about the protein in solution, thereby avoiding potential conformational biases that can arise in X-ray crystallography owing to crystal packing. These data indicate that the boomerang-shaped conformation is probably the predominant conformation of Izumo1 in solution. Moreover, although Juno binds to the outer hinge surface, the region most strongly stabilized by this binding seems to be inside the hinge. This suggests that the hinge can adopt different positions in Izumo1 alone, but that Juno fixes the conformation of Izumo1 by simultaneously binding to both domains.
Although binding interfaces are typically the most evolutionarily conserved surfaces of proteins, the Izumo1–Juno interface is less conserved than the remainder of either protein. Both groups suggest that variation at the binding surfaces might contribute to species specificity during fertilization, because sperm–egg fusions retain some specificity even if the zona pellucida (the main block to cross-species fertilization) is removed11. Ohto and colleagues introduced genetic mutations into mouse Izumo1 that strongly reduced the affinity of the Izumo1–Juno interaction. Expression of wild-type Izumo1 in monkey kidney cells (which do not normally express Izumo1) enabled these cells to bind efficiently to mouse eggs that lacked the zona pellucida, whereas cells that expressed the mutant protein could not. These results clearly confirm the interface identified in these structures and its importance in mediating sperm–egg docking.
Why would a protein-binding receptor evolve from a folate receptor? It is tempting to speculate that an unidentified, non-folate ligand might bind the pocket of Juno to modulate the receptor's activity. Folate receptors are exquisitely pH-sensitive and release folic acid under acidic conditions10, and Ohto et al. demonstrated that slight acidification drastically decreased Juno's affinity for Izumo1. Together, ligand binding and pH changes could enable Juno to regulate Izumo1 binding at multiple levels.
Although the interaction between Izumo1 and Juno in sperm–egg recognition and adhesion has been structurally and biophysically characterized, the transition from initial binding to membrane fusion remains unclear. Izumo1 stays in the membrane following binding, whereas Juno is shed. This shedding might rapidly block polyspermy before the slow hardening of the zona pellucida is completed6. Previous work12 suggests that Izumo1 undergoes stable dimerization through a disulfide-exchange reaction, dissociating from Juno to enable recruitment of membrane-fusion machinery. Indeed, Ohto et al. provide evidence that the disulfide bonds in Izumo1 are easily broken — perhaps stabilization of Izumo1 following Juno binding could expose disulfides for exchange. Testing this hypothesis and determining how Izumo1–Juno binding triggers membrane fusion will require the identification of proteins that bind to Izumo1 after Juno shedding, and the reconstitution of events that follow initial binding in cells.