Researchers at the University of Pennsylvania are investigating a counterintuitive process called allostery that occurs in proteins by studying an analogous process in a macroscopic mechanical network.

Their research, published in the Proceedings of the National Academy of Sciences, could lead to a clearer understanding of why this phenomenon is so common in proteins.

​​​​​​​The research was led by Jason Rocks, a graduate student at Penn, and is part of a longtime collaboration between Andrea Liu, the Hepburn Professor of Physics in the School of Arts & Sciences at Penn, and Sidney Nagel, a professor of physics at the University of Chicago.

In allostery, a small regulatory molecule binding on one side of a protein will alter the ability of another kind of molecule, called the substrate, to bind somewhere else on the protein. This process is a way to turn on and off the ability of the protein to bind the substrate, and it plays a role in regulating the protein and the function it fills.

​​​​​​​“You might have a protein that's an enzyme,” said Liu. “And then you have some substrate that is chemically changed by this enzyme. When the substrate comes and binds to the protein, the chemical reaction occurs and the molecule goes off as another species.”

​​​​​​​But many of these enzymes are regulated by other small molecules. One might think that these regulatory molecules would simply plug into the binding pocket used by the substrate in order to prevent the substrate from binding.

“That seems like a pretty easy way to go about doing it,” said Rocks, “but it turns out a lot of proteins don't do anything even remotely similar to that.”

Instead, the regulatory molecule binds elsewhere on the protein.

The most canonical example of this, Rocks said, is hemoglobin in blood. Hemoglobin has four binding pockets for oxygen molecules. If an oxygen molecule binds to one of them, then it’s much easier for oxygen molecules to bind to the remaining pockets. This allows for more precise control over what the protein actually does.

​​​​​​​“It's a long-distance effect,” Liu said. “You're changing something here, and that's somehow changing the ability of this thing to bind over here.”

Rather than studying this process directly in proteins, the researchers investigated a mechanical network to see if they could get it to exhibit allostery properties.

“Our way of doing this is just the simplest thing you could possibly do,” Rocks said. “You start with some random network, and then all you do is you see what happens if you remove each of the bonds one at a time. We remove the bond that gets you closest to what your desired response is, and we just keep doing that until we've finally achieved the response that’s as big as we want.”

Their collaborators at the University of Chicago then created allosteric networks using laser cutting and 3-D printing.

“We took a bunch of spheres and smooshed them together,” Rocks said. “Once the whole thing was rigid, we converted it into a spring network, putting a node in the center of each of the spheres and drawing a bond between each pair of overlapping spheres.”

This is not so different from how some researchers describe folded proteins, he said.

“You can imagine the atoms each being kind of a sphere, and they're all smooshed together. So it's somewhat physical; all the bonds are going to be very local. For some reason networks with this very local bond structure are very close to almost any property that you'd want to achieve.”

The researchers found that, starting with the same initial network but removing slightly different bonds, they could tune the networks to have opposite responses.

“Not only is it ridiculously easy to tune these responses,” Rocks said, “but we actually have a lot of control over exactly what these responses are.”

The researchers also found that the number of changes they needed to make to tune the response was incredibly small. In networks with about 400 bonds on average, they found they only needed to remove five of them in order to get the desired response.

“These networks are incredibly adaptable,” Liu said. “They also appear to be pretty robust. So few bonds are involved in this that for a reasonably large system that, once I've tuned in this function, if I start removing bonds randomly I don't immediately lose it.”

This robustness is important because it means that some small mutation or change occurring somewhere in the protein would be unlikely to destroy the function.

Although they weren’t looking at proteins directly, the researchers believe that this research may give a good hint at what to expect in more complicated systems. The reason allostery is so common in proteins may be because it is so easy to achieve in mechanical networks.

“It seems like such a bizarre thing for nature to choose for proteins to behave in this way but it's not so surprising if it's just so easy to do,” Rocks said. “It seems that just through random mutations that eventually a protein would reach a property like this.”

This work has a wide range of potential applications. It could allow for the building of a new class of mechanical metamaterials, materials whose structure has been designed so that they have properties that one wouldn’t normally expect.

Many biomedical researchers are interested in protein allostery because it might provide a route to new drugs.

“A lot of drugs are made so they just plug right into the binding pocket. But, if you really understand how allostery , then you might be able to bind to some other part of the protein,” Rocks said. “You have a lot more options in making your drugs if you really understand how these allostery properties work.”

Although this is very “pie in the sky,” Liu said, someday researchers may even be able to use similar ideas to introduce new functions into proteins.

Currently, the researchers are investigating whether they can apply the same ideas to different kinds of networks, such as flow networks, with Penn colleagues Henrik Ronellenfitsch and Eleni Katifori.

One example of a flow network is blood vessels in the brain. The researchers want to see if they can direct flow to certain parts by making slight changes in the network.

“We haven't really explored a lot of the possibilities when it comes to tuning these things,” Rocks said. “We have no idea how big the systems can get, how many bonds you need for this to work and how many different functions you can tune at the same time. We're just trying to really explore the limits of the tunability of these networks and figure out how robust the solutions are once we find them.”

The research was supported by the federal Department of Energy, National Science Foundation, Simons Foundation and National Institute of Standards and Technology.