Institut für Anatomie


Mikroskopische Anatomie und Strukturbiologie

Research interest

The Department of Microscopic Anatomy and Structural Biology led by Benoît Zuber develops and applies advanced methods to explore the fine structure of organs, cells, and proteins. To this end we mainly use cryo-electron microscopy and serial block face scanning electron microscopy

Time-resolved structural study of exocytosis

Isolated neuronal synapse

In our brain, information travels along the axon of neurons in the form of electrical signals. When the signal reaches the end of the axon, chemical messagers (neurotransmitters) are released and are detected by the neigbouring neuron, which in turn processes the signal. This happens at a specialized and asymetric interneuronal contact site called the synapse. Neurotransmitters are contained in membrane-bounded synaptic vesicles. Typically a synapse contains hundreds of vesicles. When an electrical signal arrives to the synapse, it triggers the entry of calcium in the presynaptic neuron, which in turn triggers the fusion of one or a few vesicles with the plasma membrane, resulting in the release of neurotransmitters in the extracellular space between the two neurons.

A number of questions about the mechanism of fusion remain open. We approach them with an original method where we analyze by cryo-electron tomography (LINK) the 3-dimensional ultrastructure of native isolated synapses (synaptosomes) that were fixed by vitrification milliseconds after exocytosis has been triggered. We focus in particular on the structure of membranes, as well as on the modification occurring to the dense network of filaments interconnecting the vesicles.

Collaborators: Vladan Lucic (MPI Biochemistry, Martinsried), Harvey McMahon (MRC LMB, Cambridge), Henning Stahlberg (University of Basel).

Membrane patches isolated from neuroendocrine cells

Neuroendocrine cells are another type of cells exhibiting calcium-triggered exocytosis. In this case large dense core vesicles filled with hormones fuse with the plasma membrane and release their content in the blood stream. Although the proteins involved in neuroendocrine secretion are in part identical and in part homolog to the neuronal proteins, the kinetics of neuroendocine secretion are slower those of neuronal secretion. In order to study the ultrastructural changes taking place in neuroendocrine secretion we have developed a system where the plasma membrane and associated structures, such as docked vesicles and the actin cortex, are isolated directly on an electron microscopy grid. The method, termed iMEM (isolation of Membrane patches for cryo-Electron Microscopy), allows a quick preparation of thin specimens directly observable by cryo-EM or cryo-ET. Furthermore, since the cytoplasmic side of the plasma membrane is exposed, physiological events can be easily triggered chemically, and intracellular structures can be labelled with gold-coupled labels.

Reconstituted SNARE fusion

The SNARE proteins mediate exocytosis membrane fusion. A proteoliposome in vitro reconstitution system initiated by J. Rothman’s lab has allowed to biochemically dissect how a population of proteoliposome mimicking synaptic vesicles fuse with a population of proteoliposome mimicking the plasma membrane. We analyse structural changes occuring at the level of the membranes and of the proteins by cryo-electron tomography of this system in order to better understand the function of the fusion machinery.

Collaborators: Harvey McMahon (MRC LMB, Cambridge)

Pore-forming toxin aerolysin

Many pathogenic bacteria produce protein toxins. The largest class of toxins consists of pore-forming toxins. They are secreted by bacteria in a water-soluble form and after binding to a receptor on target cells, typically eukaryotic cells, they insert and form a pore in their membrane. The cells become leaky, which may eventually lead to cell death. Production of pore-forming proteins is not only found in bacteria but also in other kingdoms of life, where they may assume different roles, e.g. in the defense against pathogens by the immune system. Fresh water bacteria Aeromonas hydrophila produce toxin aerolysin, which is the archetype of a large family of pore-forming proteins present in all kingdoms of life. Secreted water-soluble aerolysin binds to glycosylphosphatidylinositol (GPI) anchored membrane proteins, which are found in all eukaryotes. There, aerolysin C-terminal part gets cleaved off by host cell proteases, it forms a homoheptameric complex. The complex then rapidly forms a β-barrel pore in the plasma membrane. We have used single particle cryo-electron microscopy to obtain the atomic structure of aerolysin variants (mutants) blocked at different stages of the pore formation process Iacovache et al. (2016) Nature communications 7:12062. Our structures revealed that a loop refolds to form of long β-barrel, that a major collapse of the protein leads to membrane insertion of the barrel and, finally, that the cytoplasmic tip of the β-barrel folds outwards to form a rivet within the hydrophobic part of the membrane and thereby firmly anchor the pore in the membrane. Our aim is now to further dissect the mechanism of pore formation to understand in greater details how a water-soluble protein can insert in a membrane bilayer.

Collaborators: Gisou van der Goot (EPFL), Matteo Dal Perraro (EPFL), Nuria Cirauqui (EPFL/Federal University of Rio de Janeiro)


Cryo-electron microscopy

«Water is a beautiful substance. Its action is everywhere around us and in us. It has moulded the earth during geological times and it was the cradle of life. Even now, it is by far the most abundant constituent of living matter. Our body contains 65 % water and the most 'intelligent' part of it, our brain, is all water, except for a fifth of its mass. At the molecular level also, life works in, and with water. The liquid medium participates directly in many reactions and it provides the necessary mobility for the dynamics which is the essence of life. However, water has long been neglected in molecular biology.» (Dubochet J, et al. (1988). Q. Rev. Biophys. 21:129-228 ). This is partly due to the fact that electron microscopes work under high vacuum and that at room temperature water evaporates under high vacuum. A biological specimen inserted without special care in the microscope dramatically collapses and is of little use for the study of fine structures, which the subnanometer resolution of transmission electron microscopes would in principle enable. Since the early days of electron microscopy in 1940s, biologists have developed tricks to alleviate this problem. We can distinguish two fundamental approaches. The first and earlier one consists of chemically fixing the specimen before staining it with heavy metals, gently dehydrating it, and embedding it in a resin. This processing preserves relatively well the original structure of the specimen and since the sample is now totally dry it can be introduced in the microscope without it collapsing. Yet this solution is not perfect: not every molecule are efficiently fixed, and when their natural solvent, water, is removed they either aggregate or get washed away. Furthermore, the necessary use of heavy-metal stains limits achievable resolution since the microscope does not image biological molecules but stain molecules bound to them.

In response to these problems, an alternative approach termed cryo-electron microscopy, or electron cryomicroscopy (cryo-EM), was developed in the early 1980s with the aim of keeping the specimen fully hydrated. Here the specimen is frozen very rapidly in order for the water not crystallize, a phenomenon termed vitrification. The specimen is then transferred to the microscope and observed at -180°C, a temperature where water does not evaporate even under high vacuum. No staining is necessary. The specimen is in principle preserved in its native state, and resolution is only limited by the relatively low signal-to-noise ratio inherent to cryo-EM images. For well-ordered structures, a single image can provide a resolution of about 1nm. If multiple copies of the same structures can be averaged, near-atomic (<5Å), or even atomic resolution (<3Å) can be achieved. 3-dimensional information can be obtained by single particle analysis if the specimen consists of purified biological molecules, or by electron tomography for non-pure samples, such as tissues, cells or organelles.


The majority of eukaryotic cells and several bacteria are too thick to be imaged by cryo-EM directly after vitrification. They must be thinned down. One approach to do that consists of cutting series of ultrathin sections (~60nm thick) with a diamond knife, at -150°C, in order for water to remain vitreous. CEMOVIS (cryo-electron microscopy of vitreous sections) has had the reputation of being an art restricted to a handful of highly skilled and trained scientists. We have introduced a set of tools that greatly facilitate the use of CEMOVIS (Studer et al. (2014) J. Struct. Biol. 185:125-128). This makes it possible to generate in a given amount of time volumes of observable specimen that are serveral orders of magnitude larger than what can be obtained with the alternative method, cryo-focused ion beam milling. Specimens prepared by CEMOVIS however currently suffer from compression along the cutting direction. It is to be expected that knives with better gliding properties could significantly reduce this effect.

Serial block face scanning electron microscopy (SBFSEM)

SBFSEM was pioneered in the mid 2000’s by Winfried Denk and colleagues. This method makes it possible to obtain the 3-dimensional ultrastructure of “very large” samples (very large, at the electron microscopic scale). Resin-embedded samples are mounted on an ultramicrotome, which is installed within the specimen chamber of a scanning electron microscope. The microtome cuts sections as thin as 15 nm; usually though section thickness is set to 50 nm. After each cut, the section is removed and the surface of the block is imaged at low voltage. The resulting images resemble transmission electron microscopy images, although their resolution is limited to around 20-30 nm. We apply SBFSEM in a number of collaborative projects in various fields such as neurobiology, vascular biology, and parasitology.


With this approach we study several aspects of synapse biology such as postsynaptic receptors organisation and synaptic transmission. People interested to join our group are welcome to contact Benoît Zuber (