An AI that builds glowing enzymes-breakinglatest.news-Breaking Latest News

It has been almost three years since the Google subsidiary DeepMind announced a discovery: the AlphaFold neural network had succeeded in determining the three-dimensional structure of almost all proteins known to date. The discovery was not just a breakthrough in the field of protein folding. She has also demonstrated the potential of machine learning and artificial intelligence for medicine.

Researchers at the University of Washington made use of this knowledge in a recent study published in the journal Nature. Instead of calculating the structure of existing proteins from the sequence of bases, she had an AI create new proteins. Namely so-called luciferases: These are enzymes that cause bioluminescence through a chemical reaction. Among other things, they make fireflies glow.

“We were able to develop very efficient enzymes from scratch on the computer instead of relying on enzymes found in nature. This breakthrough means that, in principle, tailor-made enzymes could be developed for almost any chemical reaction,” says molecular biologist Andy Hsien-Wei Yeh , who led the study.

The new enzymes could be used, among other things, in imaging diagnostics – for example to examine processes in cells or in medicine to make biochemical reactions visible that are difficult or impossible to do with traditional methods. Diagnostics using bioluminescence is a comparatively young technology, but it is said to have great potential. In other areas, too, for example in recycling, people are looking for new enzymes that can perform specific tasks.

Finding new enzymes from scratch has been difficult until now. In order for them to become active, they not only need the right nutrient medium, the right substrate, but also perfect conditions in terms of temperature and chemical substances. Above all, however, enzymes only work as desired if they can dock to the right proteins, which in turn depends on the 3D structure of the proteins.

For their study, the researchers used diphenylterazine (DTZ), a substance used in organisms to generate light, as the target substrate. The next step was to find enzymes that can dock onto DTZs in order to trigger the chemical reaction that causes bioluminescence. This is where artificial intelligence came into play: Using trRosetta, a neural network designed to predict protein structures, the researchers managed to “hallucinate” around 1,600 protein scaffolds that could potentially react with DTZ.

To test the results, the team painstakingly encoded the proposed scaffolds in DNA and then injected them into bacteria to see if the enzymes proposed by the AI ​​actually produce light when combined with DTZ.

In the end, the “winner” was a new luciferase, which the experts dubbed LuxSit (Latin for “Let there be light”). A subsequently further optimized version, LuxSit-i, made the cells glow so intensely that the light could be seen with the naked eye. It was also very stable and retained its structure even at 95 degrees Celsius.

‘The development of highly active and specific biocatalysts from scratch with broad applications in biomedicine is an important milestone for computational enzyme design,’ says the study in Nature, which has already undergone a peer-review process. With the help of neural networks, it is possible to develop specific enzymes that can perform a variety of different tasks and make cell processes visible by emitting different types of light.

Technology Review” class=”a-u-hide-from-tablet” height=”693″ high-dpi-quality=”100″ quality=”100″ src=”https://www.heise.de/Magazin-Banner/tr_mobil.jpg” style=”aspect-ratio: 1200 / 693;” width=”1200″>

More from MIT Technology Review

Technology Review” class=”a-u-show-from-tablet” height=”500″ high-dpi-quality=”100″ quality=”100″ src=”https://www.heise.de/Magazin-Banner/tr_desktop.jpg” style=”aspect-ratio: 1830 / 500;” width=”1830″>More from MIT Technology Review

More from MIT Technology Review

(bsc)

To home page


Read more here: Source link