A crystallographic view of egg-sperm interaction

Abstract: Animal reproduction is a heavily studied biological event, however our knowledge regarding the molecular basis of the recognition between gametes – egg and sperm – still remains very incomplete. In particular, it is currently unclear whether protein epitopes important for fertilization have been conserved among different species during evolution, and why gamete interaction proteins generally are subjected to strong positive Darwinian selection. Furthermore, there has not been a reported X-ray crystal structure of an egg-sperm protein complex, in any organism, prior to sperm penetration. The aim of this thesis is to shed light on these important biological questions, by applying structural biology techniques to a selection of key reproductive proteins ranging from invertebrate to vertebrate. In PAPER I, it was hypothesized that, sperm binding region in yeast, invertebrate and vertebrate adopts the same three dimensional fold as egg zona pellucida (ZP) protein ZP3, which has been suggested to be involved in mediating binding between gametes in humans. This also led us to the hypothesis that the N-terminal region of ZP2 might contain three ZP-N domains. Based on this theory, despite the lack of sequence similarity, we suggest that the 22 tandem repeats observed in the invertebrate egg coat protein vitelline envelope receptor for lysin (VERL), as well as a protein domain in yeast, essential for mating, also adopts the same ZP-N fold. This suggests that, even though being separated by 0.6-1 billion years of evolution, these reproductive proteins all share a common structural fold important for egg-sperm recognition. In PAPER II, we facilitate protein expression by using an approach to obtain milligram recombinant protein transiently expressed in mammalian cells, either as a rigid or cleavable fusion with a mammalianized version of bacterial maltose-binding protein (mMBP). We show that mMBP can increase protein expression up to 200-fold and also assist crystallization of target proteins, which did not crystallize by themselves. Furthermore, we show that mMBP can also be used during structure determination as a search model for molecular replacement, highlighting the diversity of mMBP, not only for increasing protein expressing but also guiding crystallography. In PAPER III, we show that invertebrate egg coat protein VERL, consisting of 22 tandem repeats, previously suggested to be ZP-N domains (paper I), shares a similar 3D structure as mammalian egg coat protein ZP2. By solving the crystal structures of VERL repeats and mammalian egg coat protein ZP2 ZP-N1, we show that these reproductive proteins share a common 3D fold, proving that this specific immunoglobulin-like fold has been conserved during evolution. Furthermore, we identified the binding region between VERL and lysin, while at the same time suggesting why VERL repeats are under positive selection. We also solved the crystal structure of the VERL/lysin complex. This structure is not only the first structure of an egg-sperm protein complex on the egg coat, but also further elucidates the role of positive Darwinian selection on these reproductive proteins. Finally, our structures also suggest a very simple mechanism of how lysin-VERL binding creates a hole by splaying apart the VERL filaments allowing the sperm to enter the egg and fuse with the plasma membrane. In PAPER IV, the crystal structure of the mammalian egg protein Juno, known previously as folate receptor (FR) 4, was solved. Previous data showed that Juno is crucial for fertilization, since female mice lacking Juno are infertile. Although our structure shows a similar fold as FRα and FRß, Juno is not able to bind vitamin B9/folic acid. One of the reasons why Juno is not able to bind vitamin B9 might be due to high flexibility within the folate binding region. Even though Juno has lost the ability to bind vitamin B9 it has gained the ability to bind sperm, via sperm protein Izumo1.