Caption: Wohlever’s research shows that ALS-causing mutations will disrupt the ability of the STI1 domain to bind to the transmembrane domain or placeholder sequences. (STI1 domain, in blue, and a membrane protein (substrate), in magenta. ALS causing mutations are shown in yellow.)
By Phoebe Ingraham Renda
When oil separates from water in a bottle of salad dressing—a process called liquid-liquid phase separation—you probably give it little thought. Inside human cells, a similar process occurs constantly and is one of the key strategies cells use to compartmentalize thousands of different reactions. But when that process is interrupted, causing the liquid-like droplets to turn rigid, like if the salad dressing were tossed in the freezer, the consequences can be detrimental.
“This process is associated with a lot of neurodegenerative diseases where we go from this liquid-like mixed state to a gel state, to a solid-like fibrous aggregate—and generally the more solid you go, the more that’s associated with disease,” says Matthew Wohlever, assistant professor of cell biology, School of Medicine, University of Pittsburgh. His research team studies how disruptions in liquid-liquid phase separation contribute to amyotrophic lateral sclerosis (ALS), a neurodegenerative disease affecting nerve cells in the brain and spinal cord. In ALS, proteins called ubiquilins are known to form a dense aggregate state in neurons due to dysregulated phase separation—a hallmark of the disease.
In a study published March 20 in The EMBO Journal, Wohlever (pictured left), corresponding author, and his School of Medicine research team—including Joan Onwunma, a former master’s student at University of Toledo; Saeed Binsabaan, a former postdoctoral researcher; and Shawn Allen, a molecular biophysics and structural biology PhD student—revealed, for the first time, how mutations linked to ALS interrupt ubiquilin phase separation.
The team’s X-ray crystallography work resolved the first crystal structure of any ubiquilin STI1 domain bound to a transmembrane domain. The structure showed that the STI1 domain forms a hot dog bun-like shape, binding its substrate within the groove in the same way a bun holds a hot dog. However, that groove is very hydrophobic, meaning that if no substrate is present, that protein domain reacts unfavorably with water, which would drastically impede its ability to phase separate.
To combat this biological challenge, the protein relies on four internal placeholder sequences that work to take the place of the substrate in the STI1 hydrophobic groove. The STI1-placeholder interaction allows ubiquilins to dynamically self-associate and form droplets, a key aspect of liquid-liquid phase separation.
The identification of these placeholder sequences, Wohlever says, first arose from an “aha” moment when sharing data with Carlos Castañeda, associate professor of biology and of chemistry, College of Arts and Sciences, Syracuse University, who was studying the yeast version of this domain using nuclear magnetic resonance. “When we put our two pieces of data together, it was like two missing pieces of a jigsaw puzzle.”
In complementary projects, published back-to-back in The EMBO Journal, Castañeda’s team tested and validated the placeholder model in yeast, while Wohlever’s team translated and validated the model with human ubiquilins and began to explore the model’s relevance to ALS. To gain a structural explanation for how each known ALS-causing ubiquilin mutation affects the domain’s structure and function, the team mapped each mutation to the protein’s three-dimensional structure. That mapping revealed that all the ALS-causing ubiquilin mutations occurred in either the STI1 domain or the placeholder sequences.
“This model explains for the first time how ALS-causing mutations lead to differences in phase separation and the altered phase separation seen with ALS-causing mutations,” says Wohlever.
While researchers don’t fully understand the biological importance of liquid-liquid phase separation dynamics, Wohlever’s team is working on testing a few hypotheses to begin answering that question.
“Now that we have a structure and we know what something looks like, we can understand how to fix it when it’s broken,” says Wohlever. “We now understand for the first time how these mutations disrupt phase separation and we can now start to develop interventional strategies.”
Each mutation, Wohlever predicts, could have different effects or work through different biological mechanisms. Identifying and validating the biological effects of each mutation, he says, could lead to more personalized medicine based on which mutation a patient has: “For the first time we have the resolution to really try to understand that.”