Reverse target screening identifies putative L-AAA targets in astrocyte-selective toxicity

L--aminoadipic acid (L-AAA) is widely used to ablate astrocytes, yet the molecular basis of its cell-type selectivity remains unclear. We combined structure-based reverse screening with an enantiomeric control D--aminoadipic acid (D-AAA), network propagation on a signed, directed interactome, and cell-type-resolved transcriptomics to nominate mechanistic targets of L-AAA. Two independent AlphaFold-enabled screens converged on a small set of proteins; integrating apoptosis connectivity and astrocyte enrichment prioritized pyruvate carboxylase (PC). Docking placed L-AAA in the pyruvate/acetyl-CoA pocket of PC and supported stereoselective engagement over D-AAA. These results motivate a testable model in which L-AAA limits PC-driven anaplerosis in astrocytes, constraining oxaloacetate/aspartate supply and redox buffering to lower the threshold for apoptosis. The workflow generalizes to polar metabolites that evade standard chemoproteomics and reframes how L-AAA-based astrocyte ablation should be interpreted. SignificanceL--aminoadipic acid is a staple tool for selective astrocyte ablation, but its molecular effectors have remained elusive. By unifying AlphaFold-enabled reverse screening with network-level apoptosis inference and single-cell expression filters, we nominate pyruvate carboxylase as a plausible mediator of L-AAA toxicity. The resulting metabolism-centric mechanism--partial inhibition of glial anaplerosis that depletes aspartate and erodes NADPH-glutathione buffering--offers concrete, falsifiable predictions and clarifies how L-AAA should be interpreted in vivo and in vitro. Our pipeline is broadly applicable to small, highly polar metabolites that are challenging to profile by conventional target-deconvolution approaches.