ABOVE: Researchers identified a protein that spatially stabilizes mitochondria (shown in red) in neuronal dendritic spines (shown in cyan). Ojasee Bapat, Max Planck Florida Institute for Neuroscience.

During her graduate studies at Weill Cornell Medical College, Vidhya Rangaraju, now a neuroscientist at the Max Planck Florida Institute for Neuroscience, became fascinated with the brain. An engineer by training, she began her journey into neuroscience by building tools to track ATP in neurons. One question that has intrigued her ever since is how brain cells meet their energy needs, particularly at their synapses, which are often located far from the cell body and undergo intense remodeling to support neuronal activity. 

In a recent study, Rangaraju and her team described how the vesicle-associated membrane protein-associated protein (VAP) stabilizes mitochondria in dendrites, the branch-like structures neurons use to receive information from other neurons. They also showed that knocking out the VAP gene impaired the plasticity of dendritic spines. These findings, published in Nature Communications, suggest a role for VAP in mitochondria spatial stability and synaptic plasticity in dendrites.1   

“A lot more studies have happened in axons than in dendrites. It is nice to see somebody studying dendritic mitochondria because they're surely every bit as important as the axonal ones,” said Thomas Schwarz, a neuroscientist at Boston Children's Hospital who was not involved in the research. 

Mitochondria are the main suppliers of energy to neurons, but they look and behave differently depending on where they are situated in the brain cell. Mitochondria located in the axon, a neuron’s output structure, are discrete organelles that are either stationary or motile.2 In contrast, mitochondria are jammed together and often overlap in dendrites. “That makes it so much harder to see them—to see their moves and to see them as discrete organelles,” Schwarz explained.

The lack of knowledge about mitochondrial dynamics in dendrites motivated Rangaraju to investigate them more closely. As a postdoctoral researcher in the laboratory of neurobiologist Erin Schuman at the Max Planck Institute for Brain Research, Rangaraju and her colleagues found that, in dendrites, the organelles form long structures called mitochondrial compartments that often consist of stacked mitochondrial filaments.3 The team also showed that these compartments tether to the cytoskeleton and stay in the same place for up to two hours, providing energy for local protein synthesis and synaptic plasticity.

Although Rangaraju had hints that one or more proteins anchored the cellular powerhouses to the cytoskeleton in the dendrites, the identity of those molecular anchors was unclear. To unveil these mysterious proteins, Rangaraju and her team used hippocampal cell cultures expressing an engineered version of the ascorbate peroxidase apurinic/apyrimidinic endodeoxyribonuclease 2 (APEX2) on the outer membranes of mitochondria. By adding specific substrates to their neuronal cultures, the researchers induced the biotinylation of endogenous proteins that were very close to APEX2. The team then performed liquid chromatography coupled with mass spectrometry on these neuronal cultures and identified 129 candidate proteins. 

According to Xinnan Wang, a cell biologist at Stanford University who was not involved in the research, combining an unbiased approach such as proteomics with the proximity labeling strategy based on the APEX2 enzyme is an efficient way to find a protein’s endogenous binding partners. “It’s very elegant,” she said.

Rangaraju’s previous work suggested that disrupting the cytoskeletal architecture by actin depolymerization affected the stability of dendritic mitochondrial compartments.3 Therefore, her team decided to filter their proteomic dataset using a repository for protein interactions (BioGRID) and found that 13 out of the 129 candidates were proteins known to exclusively interact with actin. 

The researchers then narrowed their list to eight candidates with tethering roles for functional screening. To determine whether any of these proteins tether mitochondria to actin, the team knocked down each candidate gene individually and visualized changes in the mitochondria-actin interaction using a GFP-labeled probe. “It turned out all eight candidates showed an effect and were important for mitochondria-actin interaction,” said Rangaraju.    

Although all the proteins helped attach the organelle to actin in the cytoskeleton, the researchers did not know whether they were all essential for keeping mitochondria stationary over time. To address this question, the team used a mitochondrial matrix-targeted photoactivatable green fluorescent protein (mito-PAGFP). By shining light into specific dendritic segments, the team restricted the expression of mito-PAGFP fluorescence to only mitochondria in those regions, providing a way to visualize whether the photoactivated mitochondria stayed in the same place or moved away over the course of 60 minutes. Mitochondria lacking the candidate protein VAP showed a reduction in length after photoactivation, suggesting destabilization of the dendritic mitochondrial compartment.

The team next explored the functional relevance of these findings by assessing how VAP affected synaptic plasticity, the ability of synapses to strengthen or weaken over time. They measured changes in dendritic spine head size after stimulating the neurons since increases in the head size of these tiny protrusions often correlate with synaptic plasticity and brain processes, such as memory formation. The absence of VAP reduced the expected increase in spine head size after neuronal stimulation. Moreover, the dendritic spines of neurons lacking VAP did not maintain their sizes, but shrank over time. 

“It's really interesting that VAP was important for sustaining the anatomical change,” noted Schwarz.

While Wang found the results compelling and convincing, she believes that an important next step is to evaluate whether the same mechanism can be detected in an in vivo setting. Additionally, because VAP is implicated in neurological diseases such as amyotrophic lateral sclerosis (ALS), she is curious to see if the same mechanism might be important in a disease context.4,5 Rangaraju plans to examine the role of VAP in ALS-related motor learning deficits in animal models and to obtain samples from patients with ALS to see if defects in mitochondrial stabilization and synaptic plasticity appear in patient-derived neurons.

For Wang, the study findings also raise some interesting questions about the similarities and differences in the mechanisms that regulate mitochondrial dynamics in dendrites and axons. “That really opens up a whole field about mitochondria compartmentalization,” she said.