[5] are often localized far away from the soma, mitochondria are actively transported to these sites [6][7][8][9][10][11]. Also, the removal and degradation of mitochondria are tightly regulated [9,12,13], because dysfunctional mitochondria are a source of reactive oxygen species, which can damage the cell [14]. Deficits in mitochondrial trafficking have been proposed to contribute to the pathogenesis of Parkinson's disease, schizophrenia, amyotrophic lateral sclerosis, optic atrophy, and Alzheimer's disease [13,[15][16][17][18][19]. In neuronal cultures, about a third of mitochondria are motile, whereas the majority remains stationary for several days [8,20]. Activity-dependent mechanisms cause mitochondria to stop at synaptic sites [7,8,20,21], which affects synapse function and maintenance. Reducing mitochondrial content in dendrites decreases spine density [22,23], whereas increasing mitochondrial content or activity increases it [7]. These bidirectional interactions between synaptic activity and mitochondrial trafficking suggest that mitochondria may regulate synaptic plasticity. Here we investigated the dynamics of mitochondria in relation to axonal boutons of neocortical pyramidal neurons for the first time in vivo. We find that under these circumstances practically all mitochondria are stationary, both during development and in adulthood. In adult visual cortex, mitochondria are preferentially localized at putative boutons, where they remain for several days. Retinal-lesion-induced cortical plasticity increases turnover of putative boutons but leaves mitochondrial turnover unaffected. We conclude that in visual cortex in vivo, mitochondria are less dynamic than in vitro, and that structural plasticity does not affect mitochondrial dynamics.
RESULTS AND DISCUSSION
Few Motile Mitochondria in Axons of Pyramidal Neurons in Visual Cortex In VivoTo fluorescently label mitochondria and the neuronal structures in which they are localized, we performed in utero electroporation with two DNA constructs driving expression of mitomTurquoise2 and membrane-associated YFP. A cranial window was implanted once mice had reached the age of 8-10 weeks. This allowed us to visualize mitochondria in axonal arbors of layer 2/3 pyramidal neurons in V1 in vivo using two-photon microscopy (Figures 1A-1C; Movie S1). Two to three weeks after cranial window implantation, optical imaging of intrinsic imaging was performed to localize monocular V1. One week later, axonal branches in layer 1 and upper layer 2 were imaged using in vivo two-photon microscopy. We found that the density of mitochondria was 0.09 per mm ( Figure 1D). Surprisingly, the fraction of axonal mitochondria that were motile was only 0.83% ( Figure 1E), much lower than the 10%-30% previously observed in neuronal cultures [6][7][8][9][10]. We therefore asked whether the difference in mitochondrial motility was caused by the different experimental condition (in vivo versus in vitro) or by the difference in the age of the imaged neurons. We assessed mitochondrial motilit...