The billions of synaptic connections in the brain are constantly remodeled in response to changes in neural activity, a phenomenon known as synaptic plasticity. These alterations in synapse form and function allow neurons to store and relay information, and therefore provide a cellular basis for learning and memory.
Yukiko Goda began studying how neurons alter the strength of their synapses as a postdoc with Chuck Stevens at the Salk Institute in San Diego (1). This marked a major change in direction from her graduate studies in cell biology, where Goda took a biochemical approach to understanding late endosome to Golgi trafficking with Suzanne Pfeffer at Stanford (2). In her own lab—initially at UC San Diego and now at University College London—Goda has explored the cell biology of synaptic plasticity, studying how the actin cytoskeleton controls synaptic vesicle release and synapse morphology upon neuronal activation (3, 4). More recently, Goda has investigated how cell adhesion proteins such as b-catenin (5) and b3-integrin (6) modulate synaptic signaling by changing the activity or abundance of postsynaptic neurotransmitter receptors.
In a recent interview, Goda discussed some of her earliest scientific memories and what she has learned about the cellular mechanisms of synaptic plasticity.
CONNECTING TO NEUROSCIENCE
Where did you grow up?
I was born near Osaka, but I moved back and forth between the West and Japan because of my father's business—he worked for a trading company. So I've always straddled different cultures. When I was in high school, my father was briefly transferred to Toronto. By the time he transferred back, I'd been accepted into the University of Toronto with a scholarship. I didn't want to go through the Japanese university entrance exam, so my parents let me stay behind on my own.
“There are a billion synapses in just a cubic millimeter of cortical tissue.”
How did you first become interested in science?
It was a little bit by accident because I'd always been good at Japanese and arts subjects. But when I moved to Toronto, relearning English was really hard: it took me hours just to write a short essay. On the other hand, math and science topics were really easy. Especially math—I had already learned calculus in Japan, so I didn't have to do any homework. Maybe it was laziness that led me into science!
I chose mostly science classes at university. One day after class, my chemistry professor took me up to his lab and it looked really cool. I got a summer scholarship to work in an organic chemistry lab, and the following summer I applied for a biology studentship. I ended up in Jack Greenblatt's lab, working on bacteriophage transcription. I really had fun there, planning experiments and seeing their outcomes.
I developed an interest in protein trafficking around the time I was applying to grad school. Jim Rothman had written some articles comparing protein sorting at the Golgi to distillation plates. Having some chemistry background, that really attracted me. Suzanne Pfeffer, who used to be in Jim's lab, had just started her own group when I arrived at Stanford. It was terrific—I could choose any project I wanted and Suzanne was very enthusiastic and supportive.
Why did you move into neurobiology for your postdoc?
I took a Cold Spring Harbor course on developmental neurobiology. One of the organizers told me about synaptic plasticity: how it might be the way that neurons encode information, and that Chuck Stevens' group at the Salk could reproduce the process in vitro.
I thought that was cool and, with my biochemistry training, thought that you could use cultured neurons to look at the molecular basis of synaptic plasticity. I didn't dare apply to Chuck Stevens' lab though, because he was an electrophysiologist who did completely different techniques from me.
I applied to a more molecular-based neurobiology lab at the Salk, but Chuck happened to be there on the day I interviewed and we had a really nice, long chat. He liked some of my ideas and said I could join his lab. I think my ignorance of neurobiology helped because Chuck always likes to propose tough projects. I set out to do some technically difficult experiments, without appreciating how tricky they were—Chuck was probably amused—but I became very proficient at patch clamping.
I was used to biochemistry experiments, where you're basically handling dead tissue. But when you perform electrical recordings in neurons, the neuron is really alive and sending out signals that you record in real time. So it was very exciting. And neurons are very beautiful, at least compared to bands on a gel.
Most of your studies use cultured hippocampal neurons. What are the advantages of that system?
If you're interested in how synapses work, dissociated cultures are amazing. Many properties of synaptic plasticity are preserved in cultured neurons but you can manipulate the cells and map out their exact connectivity so that you know which presynapse partners with which postsynapse. You can't do that in vivo—there are a billion synapses in just a cubic millimeter of cortical tissue. So as long as you aren't addressing higher order circuit-level questions about how information is relayed and interpreted, I think cultured neurons are really useful. Slice cultures are sort of in between: they preserve a lot of the circuitry, but at the same time you can do molecular manipulations and image the connections more readily. We're using slice cultures more and more in our lab.
Recently, your lab has focused on a type of plasticity called synaptic homeostasis. What is it exactly?
When neuronal networks experience different levels of activity, synapses behave sort of like thermostats. When the network is relatively inactive, synapses turn up their sensitivity by increasing the release probability of synaptic vesicles from the presynapse or changing the abundance of postsynaptic receptors. If network activity is too high, synapses turn themselves down to become less sensitive. So synaptic homeostasis works in the opposite direction to other types of plasticity like long-term potentiation, where active synapses become stronger still.
In the past, you've investigated how the actin cytoskeleton modulates synaptic activity…
Yes, although we're moving away from that now. We studied how electrical signals cause actin remodeling, and how changes in actin could, in turn, change the efficacy of neurotransmitter release, or the way in which postsynaptic receptors are recruited and stabilized. But actin is really hard to study because it's everywhere. I thought that, rather than looking at actin directly, we should look at something a bit more upstream. Adhesion proteins mechanically link the pre- and postsynaptic sites and almost all of them connect to the actin cytoskeleton. Some of them could signal to control synaptic plasticity.
What are you learning about the function of adhesion proteins?
They're not simply glue—they're very sophisticated in reading out different signals. Their functions might change as the synapse matures. Initially they specify adhesion and connectivity, but then they might swap partners and participate in really fine-tuning synaptic signaling. For integrins, there seems to be a division of labor where β3-integrin is dedicated to modulating postsynaptic neurons whereas other integrins control presynaptic release.
A number of adhesion proteins are turning up in neurological diseases like autism. These are non-lethal diseases with only very slight changes in neural circuit function. So adhesion proteins probably play a subtle but important role in modulating synaptic strength.
What is your lab working on now?
We're interested in the relationship between synaptic homeostasis and other types of plasticity. How does a synapse distinguish signals that induce long-term potentiation from signals that require a homeostatic response? And how do adhesion proteins help discriminate between these different signals? We're also interested in the role of adhesion proteins in trans-synaptic communication. Does the synapse use homophilic adhesion proteins like cadherins to coordinate the pre- and postsynaptic sites, or could they have separate roles on each side of the synapse?
“Adhesion proteins aren't simply glue: they're very sophisticated.”
In addition, we're becoming interested in how neighboring synapses talk to each other. We found several years ago that synaptic vesicles are shared between neighboring presynaptic boutons along the axon. So what specifies the identity of individual synapses? We find that the postsynaptic site might locally specify how the presynaptic site behaves. But a synapse could also be influenced by its neighbor to some degree, and we're looking into how that would impact synaptic plasticity.
A neuron receives several thousand synaptic inputs, so if you make just one synapse stronger or weaker, it doesn't really matter for that particular neuron. Groups of synapses must act collectively as a functional unit. Thanks to improvements in imaging technologies, we can begin to address how individual synapses behave within a population of neighboring synapses, especially using cultured neurons where we know the exact identity of each connection.
What do you like to do outside of the lab?
I have lots of other interests. I like to cook and I like to dance—both classical ballet and modern. I've always done other things in parallel to research and I think that has helped sustain my scientific output. I can't do only one thing at once.
Text and Interview by Ben Short