Active substances of the future
Proteins are the molecular machines that keep life running. Thanks to artificial intelligence, their spatial structure can now be predicted from the order of their building blocks. This makes it possible to construct synthetic proteins that perform better than their natural counterparts.
Late Friday evening in April 2020, Alexandra "Lexi" Walls sat alone in her lab. There, at the University of Washington, she waited nervously for the outcome of what was probably the most important experiment of her life. The young structural biologist and expert on corona viruses had worked day and night for the past three months to develop a new vaccine against Sars-CoV-2. Their methodological approach not only had the potential to contain Covid-19, but also to transform vaccine science as a whole. The prospect of defeating all kinds of infectious diseases – from influenza to HIV – beckoned. In contrast to the vaccines used up until then, the one developed by Walls was not based on natural components. It consisted of artificial protein molecules designed on a computer and marked a major leap forward in biotechnology.
Proteins are intricate nanomachines that perform myriad biological functions in living things by constantly interacting with each other. They ensure that the organism digests food, fights off invaders, repairs cell and tissue damage, perceives its environment, processes stimuli and signals, contracts muscles, generates brain activity and reproduces. Proteins are elongated chains made up of building blocks called amino acids. As a rule, however, they are not thread-like, but twist and fold into complicated three-dimensional structures. The form they take depends on the type and order of their amino acids - because, depending on the type, they have different attraction or repulsion forces. The interactions between the various amino acids in a protein are usually so convoluted that for decades researchers have been unable to understand the principles by which such a molecule is given its final form. Not to mention that proteins are tiny: an average body cell contains 42 million of them. Many experts therefore assumed that we would never elucidate the protein folding mechanism.
But recent advances in the field of artificial intelligence (AI) are shedding light on the relationship between amino acid sequence and spatial architecture of proteins…
