When chemist and computational biologist Rommie Amaro was first profiled by The Scientist back in 2012, she had just won a Presidential Early Career Award, which recognizes outstanding scientists and engineers who are launching their independent careers and show exceptional leadership potential. She was hard at work developing computational models for disease targets from cancer to influenza and forming strategic collaborations to translate her discoveries. “When we work together with experimentalists is when the projects really just take off,” Amaro said at the time.
Now, 14 years later, Amaro is still a protein explorer; her laboratory at the University of California, San Diego, leverages the power of computational modeling to solve problems in drug discovery and explore molecular-level biophysics. For The Scientist’s 40th Anniversary, we spoke with Amaro again to learn more about her discoveries during the COVID-19 pandemic, how she works with massive datasets, and why it’s important to understand how proteins wiggle. “We're basically solving Newton's equation of motion for all the atoms in these complex systems to predict how these proteins wiggle and jiggle,” she said.
Receptor Proteins Jiggle Differently on a Crowded Dancefloor
Much of Amaro’s work has focused on creating detailed, three-dimensional, in situ simulations of proteins, including ones from viruses. Unlike in vitro protein experiments, such as X-ray crystallography, which look at single molecules, she explores a protein’s role within complex environments. “I had this idea of trying to make these very large-scale, highly detailed simulations, [and] trying to understand, for example, instead of just modeling a single protein by itself, model the protein inside the whole viral envelope,” Amaro said.
This approach led her to some key discoveries about how proteins behave within biological systems. “One of the most exciting [discoveries] that we found recently is that there is a difference in how the receptors move,” she explained. “If they're just alone in solution, they have a certain movement…If things get really crowded [in a cell], I intuitively thought they would become more rigid or less able to move.”
Instead, the simulations showed that in crowded conditions mechanical motion redistributes along particular ‘joints’ of the receptors, making them jiggle differently. “What we see is that if it gets more crowded, [the receptor] actually might bend more, or some domains actually might open up more,” Amaro said. “It’s like when you're alone in your living room, and you're having like a dance party on a Friday night by yourself. You dance one way, and then when you're in a very crowded dance club on the dance floor, you're just going to interact a little bit differently.”
Amaro is also fascinated by how receptor proteins interact with their ligands. In a recent project, she explored the sensitivity of receptors to hallucinogenic molecules, like lysergic acid diethylamide (LSD). “[We] can make the same molecule of LSD with one hydrogen moved to a different place on the molecule, and it still activates the receptor, but you don't get the hallucinogenic effects,” she said. “These receptors can be so sensitive.” Because LSD is one of the few molecules that can cause the growth of synapses, the research could be applied in experimental fields like brain remodeling, retaining the drug’s therapeutic properties while eliminating the hallucinogenic effects.
COVID-19, Massive Data, and Going Further with AI
Following on from her earlier work in influenza, Amaro and her team pivoted during the COVID-19 pandemic and intensively studied the SARS-CoV-2 virus, exploring the role of the spike protein in infection. Here too, her unique computational approach of creating complex 3D simulations of proteins in their native environments led to a major discovery about how the virus gains entry into human cells.1 Scientists had long known that viral surface proteins are coated in sugar molecules called glycans, which shield the virus from the human immune system, but they thought that was the extent of their involvement. “We were able to show that these glycans did more than that—they actually participated in the viral infection mechanism itself,” Amaro explained. “That was cool because it basically opened a new sort of paradigm for glycans.”
However, working at the bleeding edge of computational biology and creating such enormous datasets is not without its challenges. “It's hard to be sometimes very ahead of the field,” Amaro said. “These billion-atom simulations that we were able to get up and running just created enormous data sets, bigger by [multiple] orders of magnitude. So, the data processing and actually how we begin [to] think about molecules in their real in vivo context, it's a real gear change…That's something we are still working on.”
Amaro said that the rise of artificial intelligence and the resulting increase in the number of researchers vying for computing power has caused some career introspection. “Mathematical or analytical problems with closed-form solutions, which used to take a mathematically inclined graduate student maybe like four to six years to solve, one can now do it in [about] a weekend,” she said. “We've reached the point now where it has matured to a completely disruptive technology, and I think for many scientists, it really makes you pause and think about your role in the whole process and how you can effectively use these tools to really go deeper and farther than you imagined possible even a year ago.”
- Casalino L, et al. Beyond shielding: The roles of glycans in the SARS-CoV‑2 spike protein. ACS Cent Sci. 2020;6(10):1722-1734.


















