Researchers at the Nano Life Science Institute (WPI-NanoLSI) at Kanazawa University, Kyoto University, SOKENDAI, and the National Institute for Physiological Sciences have revealed how CaMKIIα—a key brain protein involved in learning and memory—organizes itself into chain-like structures. Their study, published in Science Advances, shows how activation causes these protein structures to grow and how a mutation associated with neurodevelopmental disorders changes their organization, findings that could deepen our understanding of memory formation and neurological disease.
As new memories form the connections between brain cells can become stronger. A protein called CaMKIIα plays an important part in this process and is found in especially large amounts at the points where brain cells communicate. Scientists have long known that more of this protein gathers at active connections, but they have not been able to see clearly how the individual protein molecules come together.
In the new study, a team led by Mikihiro Shibata, used high-speed atomic force microscopy, a technique that can visualize individual molecules at very high resolution, to observe how CaMKIIα molecules move and interact in real time. The researchers discovered that, under crowded conditions similar to those inside brain-cell connections, the molecules join together to form chain-like structures.
From freely moving molecules to chain-like clusters
CaMKIIα normally assembles as a ring-shaped complex called a holoenzyme, typically containing 12 protein subunits. At low concentrations, where the holoenzymes could diffuse freely, more than 95% appeared as individual particles, whether inactive or activated. Stable clusters did not form.
The picture changed when the researchers recreated two important features of the postsynaptic environment: high molecular density and restricted movement. Under these spatially confined conditions, CaMKIIα holoenzymes contacted one another and assembled into stable chains containing several holoenzymes. The contact points observed by HS-AFM indicated that the interactions occurred through the proteins' kinase domains—the regions responsible for catalytic activity.
The clusters began to form at densities below those estimated within the postsynaptic density, the protein-rich signaling region of a synapse. This suggests that the crowded and constrained environment of a dendritic spine could strongly favor higher-order CaMKIIα organization.
Activation makes the molecular chains grow
When calcium levels rise inside a brain cell, another molecule called calmodulin switches on CaMKIIα. This causes the working parts of the protein, known as kinase domains, to move outward. High-speed atomic force microscopy showed that, as the protein opened up, the CaMKIIα molecules formed larger chain-like structures. The distance between neighboring molecules increased by about four nanometers, confirming that the activated protein had adopted a more extended shape.
The activated protein can also add a small chemical tag, called a phosphate group, to itself. This process, known as autophosphorylation, helps CaMKIIα remain active after calcium levels fall. The researchers found that adding this tag at a particular site helped the larger chains remain stable. Computer simulations supported the observations, showing that the proteins are more likely to form larger groups when they open up and when their movement is restricted.
"Our observations connect the structural changes of individual CaMKIIα holoenzymes with their collective organization at a larger scale," says Shibata. "The results suggest that activation does more than switch on kinase activity: it also changes how CaMKIIα molecules assemble with one another."
A possible bridge between molecular activation and synaptic organization
The long-lasting strengthening of connections between brain cells is known as long-term potentiation, or LTP. Scientists consider LTP to be one of the main processes through which the brain forms memories.
The researchers propose that, during LTP, activated CaMKIIα molecules attach to receptors at the receiving side of a connection between brain cells. Once held in place, these molecules may provide starting points around which other CaMKIIα molecules gather. The proteins can then join together to form chains and larger groups. This could explain how brain-cell connections accumulate the large amounts of CaMKIIα needed for effective communication, even though only some of the protein molecules can attach directly to receptors.
The study suggests that the contacts between individual CaMKIIα molecules are weak. However, when many proteins are crowded together and cannot move freely, a large number of weak contacts can combine to produce stable structures. This may help explain how small changes in the shape of individual CaMKIIα molecules lead to the formation of much larger protein groups at connections between brain cells.
A disorder-associated variant clusters abnormally
The team also studied a changed form of CaMKIIα known as P212L. The name indicates that one of the protein's building blocks has been replaced by another at position 212. This change results from a mutation in CAMK2A, the gene that provides the instructions for making CaMKIIα. It is described as a de novo mutation because it arises spontaneously rather than being inherited from a parent. The mutation has been associated with intellectual disability and other disorders affecting brain development.
Previous research found that P212L is activated more easily than the normal protein. In mice, it can also produce an unusually strong form of long-term potentiation (LTP), the lasting strengthening of connections between brain cells.
In the present experiments, P212L formed much larger groups than the normal protein, even while it was in its resting, non-activated state. The researchers suggest that the mutation makes it harder for the protein to remain in its normally folded, switched-off shape. As a result, it is more likely to open up and join with other CaMKIIα molecules.
"The abnormal basal clustering of the P212L variant offers a possible molecular link between altered CaMKIIα organization and excessive synaptic responsiveness," explains Shibata. "It will now be important to test whether comparable clusters form in living synapses and how changes in cluster size affect neuronal function."
What the study does—and does not—show
The study directly demonstrates CaMKIIα self-organization in a purified experimental system and identifies conditions that control cluster size. It does not yet demonstrate the same chain-like structures inside living neurons. The HS-AFM observations were made in two dimensions on a mica surface, where non-specific adsorption may influence molecular behavior. The authors therefore plan to develop improved substrates and imaging conditions and to investigate whether similar reversible cluster dynamics occur in intact synaptic environments.
Despite these limitations, the work provides a new framework for understanding how the local concentration, activation state and movement of CaMKIIα may work together to organize synaptic signaling. The findings are expected to inform future models of learning and memory and studies of neurodevelopmental disorders linked to CAMK2A mutations.
Key concepts and methods
CaMKIIα (calcium/calmodulin-dependent protein kinase II alpha) – A highly abundant neuronal enzyme essential for synaptic plasticity, learning and memory.
Holoenzyme – The complete CaMKIIα complex, typically formed from 12 subunits arranged around a central hub.
Long-term potentiation (LTP) – A persistent strengthening of communication between neurons and a widely studied cellular mechanism underlying learning and memory.
High-speed atomic force microscopy (HS-AFM) – An imaging technique that records the shape and movement of individual biomolecules in solution at nanometer resolution.
Mesoscale organization – Structures larger than individual molecules but smaller than whole cellular compartments; here, approximately 5–500 nanometers.
Reference
CaMKIIα holoenzymes self-organize into chain-like mesoscale clusters. Taisei Suzuki, Takashi Sumikama, Keisuke Matsushima, Kodai Hasegawa, Ayumi Sumino, Kenichi Umeda, Noriyuki Kodera, Tamoghna Das, Carsten Beta, Hideji Murakoshi, and Mikihiro Shibata. Science Advances. [PUBLICATION DATE, VOLUME, ARTICLE NUMBER AND DOI TO BE CONFIRMED].
Funding
This research was supported by the World Premier International Research Center Initiative (WPI), MEXT, Japan; JSPS KAKENHI (JP24K21942, JP25H00972, JP22H04926, JP23H04244 and JP24H01298); the Mochida Memorial Foundation for Medical and Pharmaceutical Research; the Uehara Memorial Foundation; the Naito Foundation; the Takeda Science Foundation; JST CREST (JPMJCR1762); JST SPRING (JPMJSP2135); and JST ERATO (JPMJER2403).
