“Look, he’s on his way to the microscope!” Sjors Scheres (1975) points to a colleague who walks past at a fast pace with a steaming bowl. Liquid nitrogen is in it, more than 180 degrees below zero, and in there again is a gold-colored piece of confetti. At least that’s how it seems, but it is a grid – a metal plate of two millimeters in diameter, made of copper or gold, with a few drops of purified protein on it. Just now that protein has a flash freeze-undergoing treatment with another ice-cold substance: liquid ethane. “This ensures that the protein freezes so quickly that no crystals form and it is extra clearly visible under the microscope.”
That microscope is no ordinary microscope, Scheres explains, as he rapidly descends the stairs of the MRC Laboratory of Molecular Biology in Cambridge – he walks as fast as he talks. It is enthusiasm, passion for the profession. For about twenty years, Scheres has been one of the leading scientists in the field of cryo-electron microscopy, or cryo-EM for short. A technique in which electrons are shot through a frozen molecule (for example a protein) in order to visualize details down to the atomic level.
As a structural biologist, Scheres revolutionized his field in 2012 with the computer algorithm Relion that he developed. Trade magazine Nature called him one of the ‘ten people who mattered this year‘, he won several prizes and was made a fellow of the prestigious Royal Society. Because structural biologists worldwide are now working with Relion: thanks to the algorithm, good 3D images of proteins can be made for the first time, based on cryo-EM images. And that is crucial for research into Alzheimer’s disease, among other things. Scheres rushes past posters with close-ups of tau proteins, which play a crucial role in the brains of Alzheimer’s patients. “They were made with” – he triumphantly swings open a door – “this electron microscope!”
If I’m very honest, I mainly see a black and white box of blocks…
“Yes, the real magic happens inside… It’s in operation now so we can’t look into it. But imagine it this way: a robot arm puts the frozen grid in place and then all kinds of electrons shoot through it. We can then see these protein images in 3D down to the atomic scale on the computer. The microscope has been specially placed here on the ground floor, on a thick concrete slab – if you work at atomic resolution, you do not want unwanted vibrations to occur.
“Here, for example, you see a beautiful filament of a tau protein. An elongated fiber, which looks innocent enough in itself. But unwanted accumulations of these filaments occur in the brains of Alzheimer’s patients. They appear to differ in shape from tau proteins in the brains of healthy people, and in cross-section you see that they have a characteristic C-shape: the ‘Alzheimer’s structure’.”
And thanks to this microscope and your algorithm, we can study that structure in 3D?
“The first images of these accumulations in the Alzheimer’s brain were made almost a century ago, using polarized light. Later, these individual filaments were studied in more detail with X-rays and ‘normal’ electron microscopy. Cryo-EM initially produced far too low resolutions. It was not without reason that the method became blobology named. But around 2013, a resolution revolution took place. On the one hand, the electron detectors became more accurate, and on the other hand, Relion allowed us to make good 3D images of proteins for the first time. Including the tau proteins from the brains of deceased Alzheimer’s patients.”
What is the function of tau proteins?
“It appears that in the brains of healthy people they ensure the stabilization of microtubules, which are transport tubes in cells. In principle, our genes determine the order of amino acids, which in turn determine the 3D structure and function of the final proteins. But sometimes something goes wrong. Then some proteins fold into a different shape, and it seems that a snowball effect occurs in which they encourage other proteins to do the same. Eventually they accumulate.”
So Alzheimer’s isn’t the only disease where things go wrong?
“Remarkably, you also see C-shaped filaments of tau proteins in chronic traumatic encephalopathy, a disease in which nerve cells die due to repeated traumatic brain injuries, for example in professional rugby players or boxers. But the C’s are different in shape: less closed than in Alzheimer’s. You also have neurodegenerative diseases in which another protein causes problems. Parkinson’s, for example. Proteins can take different forms in different diseases.”
The MRC Laboratory is built in an X-shape – “a nod to the shape of chromosomes”. It is an internationally renowned research institute in the field of molecular biology; It is not without reason that vaccine developer AstraZeneca has moved to Cambridge. “They are now across the street.” Scheres goes up the stairs to the top floor. “From the restaurant there you have a beautiful view over Cambridge.”
Scheres grew up in North Limburg. “My father was the village vet, and I wanted to become that too. But when I was selected for veterinary medicine, I decided to study chemistry.” During his PhD, he wanted to unravel the structure of a protein, but laboratory research – purifying proteins – did not suit him. “Too much fiddling. Then I started developing software to help unravel those structures.”
Initially you mainly worked with X-rays.
“Yes, but it had already been automated to such an extent that little seemed possible in the field of software improvement. That’s when I started focusing on cryo-EM. That method was still in its infancy.”
But now…
“…the Sjors of twenty years ago would also find this not challenging enough, haha. Yes, we have come a long way with cryo-EM. But there are still challenges. We can already make images of rigid protein molecules down to atomic resolution, 0.1 nanometer. Only: most proteins are very mobile. That is why we now want to use artificial intelligence to ensure that these molecules can also be imaged down to the atomic level. At the same time, we want to unravel the protein structures of neurodegenerative diseases in even greater detail. If we can recreate these structures, this will hopefully lead to better insight into the molecular mechanisms that underlie them. And who knows, this may one day lead to better diagnostics and treatment.”