Title:	Cryo-EM of native membrane protein complexes underlying fertilization
DNr:	Berzelius-2025-276
Project Type:	LiU Berzelius
Principal Investigator:	Luca Jovine <luca.jovine@ki.se>
Affiliation:	Karolinska Institutet
Duration:	2025-08-29 – 2026-03-01
Classification:	10601
Homepage:	http://jovinelab.org
Keywords:

Abstract

Recognition and fusion between egg and sperm at fertilization is a fundamental biological process that allows the transmission of genetic information between generations. Despite the crucial importance of this event, our knowledge of the molecular basis of gamete interaction remains very limited. In organisms ranging from worm to human, more than 10 sperm proteins have been identified that are essential for fertility and thought to assemble into a so-called "fertilization synapse" complex (FSC) which first recognizes the plasma membrane of the egg and then fuses with it. Notably, some of these molecules have been conserved throughout evolution, and structures of a few non-conserved FSC components have been determined in isolation. However, there is no structure of any of the conserved subunits. Most importantly, the complexity of the full assembly has so far precluded its preparative reconstitution from recombinant proteins, as well as its definitive biochemical characterization (both in terms of exact composition and stoichiometry) and 3D structure determination. To overcome this impasse, we managed to isolate native secretion vesicles from sperm that contain the FSC and recently collected a cryo-EM dataset from this material. Due to the major biological importance of the FSC and the implications of its structure for understanding mutations associated with infertility and the possible development of non-hormonal contraceptives, this is a very exciting development. However, we expect that processing of this data will be highly complex due to the native nature of the vesicles, which probably contain many other membrane proteins in addition to the FSC, and the lack of any information (beyond AlphaFold models of some possible subcomplexes) on the expected 3D architecture of the latter. Reconstructing the structure of the FSC will thus require a large number of trials with variable parameters, which will necessarily depend on far greater computational resources than those available via the local workstations of our group. We are therefore applying for Berzelius machine time in order to efficiently attack this problem and maximize the chance that our dataset will yield the first structure of a FSC.