Quantum Physics In Proteins

Artificial intelligence provides unmatched insights into the function of biomolecules

A new analysis method provides previously unattainable insights into the extremely fast dynamics of biomolecules. The development team led by Abbas Ourmazd from the University of Wisconsin Milwaukee and Robin Santra from DESY presents the clever combination of quantum physics and molecular biology in the journal "Nature". The researchers have used it to track how the photoactive yellow protein (PYP) changes its structure in less than a trillionth of a second after it has been stimulated by light.

"In order to understand biochemical processes in nature, such as photosynthesis in certain bacteria, it is important to know the detailed process," explains Santra, explaining the motivation. “When photoactive proteins are struck by light, they change their spatial structure, and this structural change determines the role that a protein assumes in nature.” So far, however, it has hardly been possible to follow the exact course of such structural changes: they can only be Determine and theoretically interpret the initial and final state of a molecule before and after a reaction. “But we don't know exactly how the change in energy and shape in between takes place,” says Santra. "It's like you can see someone clasped their hands, but you can't see them bend their fingers for it."

While the hand is big enough and the movement is slow enough that we can observe it with our eyes, in the realm of molecules it is not so easy. The energy state of a molecule can be determined very precisely with the help of spectroscopy. And to analyze the shape of molecules, researchers use bright X-ray light like from an X-ray laser. Thanks to its very short wavelength, it can decipher very small spatial structures, such as the positions of atoms in a molecule. However, this does not create an image as in a photo, but a characteristic scatter pattern of the X-rays from which the spatial structure can be calculated.

Short, bright X-ray flashes
Since the movement on the molecular level is extremely fast, the scientists have to use extremely short X-ray flashes, otherwise the image will be smeared. It was only with the invention of the X-ray laser that it was possible to produce sufficiently bright and short X-ray flashes to capture these dynamics. Since molecular dynamics takes place in the field of quantum physics, where the laws of physics familiar from everyday life no longer apply, the measurements cannot be understood without a quantum physical analysis.

There is one special feature of the photoactive proteins to be considered: Due to the irradiated light, their electron shell changes into an excited quantum state, which causes an initial change in shape of the molecule. This change in shape can in turn lead to an overlap of the excited and the ground quantum state. The result is a quantum leap from the excited back to the ground state, whereby the changed shape of the molecule initially remains. The funnel-shaped intersection of the quantum states is scientifically called a conical intersection and opens a path to a new spatial structure of the protein in the quantum mechanical ground state.

The team led by Santra and Ourmazd has now succeeded for the first time in deciphering the structural dynamics of a photoactive protein at such a funnel-shaped intersection. To do this, they resorted to the help of machine learning. For such a description of the dynamics would actually have to consider all the movement possibilities of all the particles involved. However, this quickly leads to confusing, no longer solvable calculations.

Around 6000 dimensions
"The photoactive yellow protein that we examined consists of around 2,000 atoms," reports Santra, who is a senior scientist at DESY and a professor of physics at the University of Hamburg. “Since every atom can basically move in all three spatial dimensions, there are a total of 6000 movement options. This leads to a quantum mechanical calculation with 6000 dimensions - and that cannot be solved even with today's most powerful computers. "

Using computer analysis based on machine learning, however, collective movement patterns of the atoms in the complex molecule could be identified. "It's like folding your hands: We don't look at each atom individually, but at their concerted movement," explains Santra. In contrast to a hand, in which the collective possibilities of movement are obvious, these are not so easy to recognize in the case of the atoms of a molecule. In this way, however, the computer was able to reduce the approximately 6000 dimensions to four. With the demonstration of this new method, Santra's team can also for the first time characterize a conical intersection of quantum states in a complex molecule made up of thousands of atoms.

The detailed calculation shows how this conical funnel forms in the four-dimensional space and how the photoactive yellow protein changes its structure after being excited with light and thereby falls back into its original quantum state. The scientists can now describe this process in steps of a few dozen femtoseconds (quadrillionths of a second) and thus contribute to an understanding of the photoactive processes. "Quantum physics provides new insights into a biological system, and biology provides new ideas for quantum mechanical methodology," says Santra, who is also a member of the Hamburg Cluster of Excellence "CUI: Advanced Imaging of Matter". "Both areas fertilize each other."