Synapses drive local mitochondrial ATP synthesis to fuel plasticity

Background: Our brains are constantly forming and stabilizing memories, a process that relies on synaptic plasticity. Supplying ATP to these synapses, which can be located hundreds (or even thousands) of microns away from the neuronal cell body, poses a significant challenge. Therefore, synapses require an immediate and local energy supply, typically provided by mitochondria stabilized near dendritic spines. While it is known that mitochondrial dysfunction is associated with various neurological disorders, the fundamental mechanisms underlying dendritic spine energetics and how synapses signal their energy needs to local mitochondria have remained largely unknown. The primary obstacle in understanding spine energetics has been the lack of sensitive tools to measure ATP within small compartments like spines and mitochondria, coupled with the need to combine these measurements with single-spine stimulation and plasticity induction.

This preprint addresses these challenges by employing newly engineered spine- and mitochondria-targeted ATP reporters, along with two-photon glutamate uncaging, to study ATP dynamics upon synaptic plasticity induction at high resolution.

Key Findings of the preprint:

Spn-ATP: A quantitative reporter of spine ATP

The researchers begin with the introduction of a new quantitative spine ATP reporter called Spn-ATP. This reporter consists of luciferase (luc) fused to the postsynaptic protein Homer2 and the pH-sensitive fluorescent protein mOrange2 (mOrg) for ratiometric measurements (Fig.2A). Due to the low light emission of luminescence reporters, the authors custom-built an inverted spinning disk confocal microscope with specialized cameras for sensitive luminescence imaging and simultaneous two-photon manipulation. Using this setup and a calibrated Spn-ATP reporter, they measured steady-state ATP_spine concentrations at 0.41 ± 0.01 mM in areas 50–100 µm from the soma. They also found that steady-state ATP_spine decreases with increasing distance from the soma (Fig.2B).

Fig. 2 from the preprint: Synaptic plasticity induction increases ATP_spine.

(A) Illustration showing Spn-ATP for ATP_spine measurements following synaptic plasticity induction (lightning bolt). (B) Representative images showing spine calcium (white, GCaMP6s) and Spn-ATP luminescence (orange, luc) increase upon synaptic plasticity induction (white and black asterisk). The intensity scale is in arbitrary units. Image made available under a CC-BY-NC-ND 4.0 International license.

Synaptic plasticity induction increases ATP_spine and drives local ATP_mito synthesis

Contrary to expectations of ATP depletion, synaptic plasticity induction via two-photon glutamate uncaging at individual spines led to an instant (~1 min) and sustained (~11+ min) increase in ATPspine of about 1.5-fold. This increase was primarily fueled by mitochondrial oxidative phosphorylation initially, while both oxidative phosphorylation and glycolysis contributed to the sustained increase. Inhibition of these pathways affected the ATP_spine increase and impaired spine structural plasticity. This instant and sustained energy supply is considered essential for synaptic plasticity. The increase in ATP_spine was spatially restricted to spines within 10 µm of the stimulated spine, pointing to a local ATP source, likely mitochondria.

To investigate the mitochondrial role, the authors developed Mito-ATP (Fig.3A), a reporter targeting luciferase to the mitochondrial matrix for ratiometric ATP measurements. Using Mito-ATP, they observed a rapid rise in ATP_mito within the first two minutes of synaptic plasticity induction, which was sustained for at least 12 minutes. This increase was dependent on mitochondrial oxidative phosphorylation and, importantly, was spatially restricted to a ~10 µm compartment within mitochondria near the stimulated spine.

 

Fig. 3 from the preprint. (A) Illustration showing the design of the mitochondrial ATP sensor, Mito-ATP, for ATP_mito measurements following synaptic plasticity induction (lightning bolt). Luciferase (luc) is conjugated to the fluorescent protein mCherry2 (mCh) and targeted into the mitochondrial matrix. (B) Representative images showing spine calcium influx in the plasticity-induced spine (white and black asterisk) measured using GCaMP6s fluorescence. In response to spine plasticity induction, there is a spatially restricted increase in mitochondrial luc luminescence at the base of the plasticity-induced spine without a change in the mCh fluorescence. Scale bar 10 µm. Image made available under a CC-BY-NC-ND 4.0 International license.

ER plays a regulatory, but not obligatory, role in mitochondrial calcium influx

Having established a method for quantifying steady-state ATP levels at the spine and observing their spatial gradient, the researchers next sought to investigate how synaptic activity and plasticity events dynamically alter these local ATP concentrations. Synaptic plasticity induction increases local mitochondrial calcium influx. This study explored the link between synaptic inputs and mitochondrial ATP synthesis, hypothesizing that calcium, known to activate Krebs cycle enzymes, could be the signal.

To test the hypothesis this, they used special tools to track calcium levels in both the cytoplasm and the mitochondria of neurons. When they stimulated a single synapse with via glutamate uncaging they observed a rapid and simultaneous increase in calcium within the nearby mitochondria.

This mitochondrial calcium influx had specific requirements: it depended on the activation of NMDAR receptors at the synapse and only occurred if the calcium concentration in the dendrite, right at the base of the stimulated spine, reached a certain threshold. Interestingly, this calcium uptake by mitochondria was very localized, happening only within a small region (around 20 micrometers) of the stimulated synapse. This result suggests that calcium acts as a fast, local, and sustained signal that tells the mitochondria to produce more ATP specifically for that active spine (Fig.7).

The authors also looked at the role of the endoplasmic reticulum (ER), another cellular compartment that stores calcium. By interfering with the ER’s ability to take up or release calcium, they found only a minor impact on the mitochondrial calcium uptake following synaptic stimulation. This observation indicates that while the ER might have some influence, it’s not essential for this process of mitochondrial calcium entry triggered by synaptic activity. This was further confirmed by the fact that their standard stimulation method didn’t effectively release calcium from the ER, and because many spines don’t have much ER present within them.

Fig. 7 from the preprint. Illustration showing the significance of synaptic calcium signal input (green sphere) driving local mitochondrial calcium signaling (green sphere, green cristae) enabled by local mitochondrial stabilization (VAP, actin tethers), resulting in spatially restricted mitochondrial ATP production (yellow cristae) and its distribution to plasticity-induced spines (yellow spines) measured using mitochondrial- (Mito-ATP) and spine-targeted (Spn-ATP) ATP reporters, respectively. Image made available under a CC-BY-NC-ND 4.0 International license.

Why I think this work is important: This preprint presents a significant advance in our understanding of the energetic groundings of synaptic plasticity by allowing for a direct and quantitative measurement of ATP dynamics during synaptic activity and plasticity. Furthermore, the work highlights the critical role of mitochondrial stability in ensuring a reliable local energy supply for sustained synaptic function, with clear implications for neurological disorders. The comparison of steady-state ATP levels and energy dynamics between dendritic spines and presynaptic terminals provides valuable insights into the distinct energetic requirements of these neuronal compartments.

 

doi: https://doi.org/10.1242/prelights.40383

Read preprint (No Ratings Yet)