Impact
Released September 21, 2022
Alphafold’s forecast paves the way for new treatments that could affect more than 10 million people worldwide
It was a source of hard-earning satisfaction after often felt like a difficult battle. David Komander and his colleagues have finally announced the long-standing structure of Pink1. Mutations in the gene encoding this protein cause neurodegenerative diseases with a wide range of progressive symptoms, particularly early-onset Parkinson’s disease, with body tremor and difficulty in moving. However, when other science teams exposed their own structures of the same protein, something became clear.
“The other two structures that came out were very different from those that were done by our group,” says Zhong Yan Gan, a doctoral student at Komander’s Lab, co-directed by Associate Professor Grant Dewson at Wehi (Walter and Eliza Hall Institute of Medical Research) in Melbourne, Australia. They were strange and had unique features that didn’t seem to exist to others. The interests were high. Understanding Pink 1 can help you unlock new treatments that address the basic causes of Parkinson’s disease. This affects more than 10 million people around the world.
Komander’s team was confident in their findings, but the contrasting results raised some major questions. And in the highly competitive field of research, they knew that they weren’t alone in hunting answers. “Not only have these very difficult nuts cracked, but after they crack, you suddenly open up all of this whole area doing something very similar,” says Commander.
The team ultimately unraveled the mystery, but needed more years of research, a random discovery, and help from the deep mind’s protein structure prediction system, Alphafold.
Symptoms of Parkinson’s disease occur when someone’s brain does not produce enough chemical dopamine. Most people who have obtained Parkinsons don’t know a specific cause, but about 10% of patients can refer to a specific genetic mutation. In these cases, Parkinson’s disease tends to develop early and affects people before they reach the age of 50.
One of those genetic variations lies in the gene encoding the Pink1 protein. Pink1 plays an important role in the failure and removal of mitochondria, often referred to as intracellular power plants. “As you get older, mitochondria can get older and get damaged,” says Gunn. “Pink1 is part of the body’s mechanism that recycles old mitochondria and gives way to new ones.”
When this mechanism is alleviated, damaged mitochondria accumulate, leading to the loss of dopamine-producing neurons, which ultimately leads to Parkinson’s disease. Therefore, one way to find a better way to treat this condition is to better understand Pink1 and its role.
When researchers discovered that Pink1 could cause Parkinson’s disease in 2004, finding its structure became an important goal, but it wasn’t approaching because human Pink1 was too unstable for it to be produced in a lab. Scientists have been pushed to cast the net more widely, and found that insect versions such as human lice are stable enough to produce and study in the lab.
That’s back to the beginning of our story. The Komander’s team unveiled the Pink1 structure in 2017. However, when other researchers published different structures of the same protein from different insects (wheat flour), they knew that there was only part of the story. That wasn’t totally surprising. After all, proteins are dynamic molecules. “They’re like machines and can take different shapes,” Gunn says. What if the exposed structure is just one of those shapes?
“
We had these new structures, and at the time we were the only people on earth who knew what Pink1 would look like during activation.
David Commander, Biochemist
Gan took on the ambitious task of understanding what Pink1 looks like as a PhD project in every step of the Activation Process. During this work, he found something strange. “We would normally ignore it as just agglomerated together, like the white kind of scrambled egg,” Commander says.
However, Cancer believes this mass is worthy of further investigation, and with the help of Dr. Alisa Glucova, decided to probe the molecules at the atomic scale using a cryoelectron microscope (cryo-em). “I remember telling Zhong, ‘Yeah, you can try it, but it never works,'” Komander admits.
The gun’s tenacity was rewarded with spades. What he discovered was the very molecule that researchers were looking for: Pink1. But why is it so big? It turns out that Pink1 likes the company. Instead of a single protein, it was grouped into pairs of molecules known as dimers and placed itself in even larger layers. “The six dimers of Pink1 were assembled into a large bagel-shaped structure,” says Gan.
This accidental discovery allowed him to use cryo-em. This did not work with as small molecules as a single Pink1 to solve the physical structure of proteins. The team had the answer.
The previously published structure of Pink1 was not incorrect. These were different forms in which proteins were collected at different stages of the activation process. But there was a catch. All of this experimental work was carried out using insect-derived Pink1. To understand the implications of their findings for humans with Parkinson’s disease, we need to investigate whether their findings have been extended to human versions of proteins.
Komander and his team turned their eyes to Alphafold. “We had these new structures, but back then we were the only people on the planet who knew what Pink1 would look like during activation,” Komander says. So they used Alphafold to invoke predictions of the structure of human-supplied Pink1, which were then on the screen. How accurate AlphaFold’s predictions were “completely shocking,” he says.
When Gan then placed two protein sequences in Alphafold to predict the structure of human Pink1 dimers, the results were largely indistinguishable from experimental studies with insect proteins. “That dimer was essentially showing exactly how these two proteins interact, so we can work together to form some of these complexes that we saw,” Commander says.
“
Rather than dealing with the fact that it is broken, we can start thinking about what kind of drugs we need to develop to fix proteins.
David Commander
This close consistency between some experimental results and the predicted structure of Alphafold gave the team confidence that AI systems could provide meaningful knowledge beyond empirical work. They continued to model the effects of certain mutations on the formation of dimers using Alphafold, investigating how those mutations could lead to Parkinsonson and how their suspicions were confirmed.
“We were able to generate some real insights right away for people who have these specific mutations,” Commander says. These insights could ultimately lead to new treatments. “We can start thinking about what kind of drugs we need to develop to fix proteins, rather than dealing with the fact that they’re broken,” Commander says.
They submitted their findings on the mechanism of activation of Pink 1 to Nature in August 2021, and this paper was accepted in early December 2021. Researchers at Trempe Lab in Montreal, Canada reached a similar conclusion, and when the team’s paper was published in December 2021, Wehi’s authors had to quickly track the final revision. “We were told to complete the paper three days before Christmas so that it can be released in 2021,” Commander says. “It was a brutal timeline.”
Ultimately, these well-known papers came out within weeks of each other. Both provide important insights into the molecular basis of Parkinson’s disease.
Of course, there are many questions left for researchers in this field, and Alphafold is free to help them reach some of the answers. For example, Sylvie Callegari, a senior postdoctoral researcher at Komander’s Lab, used Alphafold to find a large protein structure called VPS13C by stitching together smaller pieces of protein.
“Now we can ask a variety of questions,” she says. Instead of “What does it look like?” you can ask, “How does it work?”, “How does this protein mutation cause illness?”
One of the many goals of Alphafold is to accelerate medical research, which is also applied to the genetic sequences of early-onset Alzheimer’s disease people at Wehi, allowing researchers to investigate the causes of individual cases. “Alphafold can do that based on a fantastical, correct human model,” says Komander. “It’s very powerful.”
