Hypothesis\n\Neuronal C5aR1 activation, triggered by locally generated C5a, phosphorylates ULK1 at a distinct site that biases the autophagy machinery toward mitochondria‑rich synapses first, establishing a tri‑tiered elimination hierarchy: (1) synapses with high mitochondrial ROS, (2) complement‑tagged but metabolically stable synapses, (3) quiescent synapses. Disrupting this phosphorylation shifts the order, causing premature removal of low‑activity synapses and preserving damaged ones, which drives cognitive decline in aging.\n\n## Mechanistic Model\n\n- C5a generation: Astrocytes and microglia release C5a in response to Aβ oligomers; neuronal C5aR1 couples to Gαi, lowering cAMP and activating AMPK.[2]\n- AMPK‑ULK1 crosstalk: AMPK phosphorylates ULK1 at Ser555 (canonical) and, when C5aR1 is engaged, additionally at Ser757 via a PKA‑independent kinase (e.g., IKKβ). This dual phosphorylation creates a conformational bias that favors recruitment of the autophagy adaptor NIX/BNIP3L to mitochondrial outer membrane.[4,5]\n- Synaptic triage: Mitochondrial‑laden synapses exhibit elevated ROS, exposing cardiolipin on the outer membrane, which binds LC3 with high affinity when ULK1 is primed by C5aR1‑Ser757.[1,3] Consequently, these synapses are engulfed first. As mitochondrial stress resolves, C5aR1 signaling wanes, ULK1 reverts to basal Ser555 phosphorylation, and autophagy proceeds on complement‑opsonized synapses via CR3‑mediated microglial phagocytosis.[6]\n- Failure mode: Genetic ablation of the C5aR1‑Ser757 site (or pharmacological IKKβ inhibition) removes the mitochondrial bias, causing ULK1 to act on random synaptic cargo. In aged mice, this leads to early loss of low‑activity synapses (detected by reduced PSD‑95) while damaged, ROS‑high synapses persist, correlating with worsened memory performance.\n\n## Predictions and Experiments\n\n1. Phospho‑specific detection: Use a custom pSer757‑ULK1 antibody to show increased signal in hippocampal neurons of 12‑month‑old APP/PS1 mice compared with young controls; signal should co‑localize with mitochondrial markers (TOM20) and C5aR1.[2,4]\n2. Synapse‑specific autophagy flux: Express mito‑Keima and synaptic‑GFP‑LC3 reporters; predict higher mito‑Keima redox shift in synapses exhibiting C5aR1‑Ser757‑ULK1 activation, blocked by C5aR1 antagonist (PMX53) or IKKβ inhibitor (SC-514).[5]\n3. Behavioral rescue: Viral knock‑in of ULK1‑S757A (non‑phosphorylatable) in aged mice should accelerate synapse loss measured by EM and impair spatial learning in the Morris water maze; conversely, ULK1‑S757D phosphomimetic should preserve mitochondrial synapses and rescue cognition.\n4. Hierarchy mapping: Apply array‑tomography to quantify C3b/iC3b, TOM20, and PSD‑95 across individual synapses; expect a sequential decrease: first loss of PSD‑95 in TOM20‑high/C3‑low synapses, followed by TOM20‑low/C3‑high synapses in wild‑type but not in ULK1‑S757A mutants.\n\n## Potential Implications\n\nIf the C5aR1‑ULK1Ser757 axis dictates the synaptic autophagy hierarchy, then aging‑related cognitive decline may stem not from excessive autophagy per se but from a mistimed triage mechanism. Therapeutically, fine‑tuning C5aR1 signaling—rather than global autophagy inhibition—could restore the proper order of synapse preservation and clearance, offering a precision approach to mitigate neurodegenerative pathology.\n\nReferences (inline citations correspond to the numbered sources provided):\n[1] https://www.science.org/doi/10.1126/science.aad8373\n[2] https://pmc.ncbi.nlm.nih.gov/articles/PMC2230831/\n[3] https://pubmed.ncbi.nlm.nih.gov/34738335/\n[4] https://doi.org/10.1038/s41419-019-1464-x\n[5] https://pmc.ncbi.nlm.nih.gov/articles/PMC10330548/\n[6] https://doi.org/10.1038/s41467-020-15119-w