ATG3 - 3D Mechanism Simulator

Loading structural data… (9E8P)

References & Databases

START HERE — how to read E1~LC3 + ATG3 → ATG3~LC3 An E2 transthiolation cartoon on the human ATG3 frame that actually exists: 9E8P (2.70 Å) human ATG3 with crystal GABARAP conjugated at residue 264. Occupancy is C264K isopeptide (Lys264 NZ—Gly116 C = 1.33 Å), not WT Cys264 thioester. SEQADV writes LYS 264 ← UNP CYS 264. UniProt ACT_SITE is Cys264. Teaching cargo is LC3; this fold is GABARAP. Schematic E1~LC3 thioester is [schematic]. Lys264 ★ is the compiled Cys analog on this C264K mesh. Blue is the nucleophile, not hydride. Measured Å are deposited 9E8P pairs only. 4NAW is an ATG3 peptide fragment, not this fold. ATG12 is not this page. ATG4 next. Gap 86–193 is named, not drawn. Why no PE: 9E8P is a C264K isopeptide conjugate and PE (phosphatidylethanolamine) is a membrane lipid, not a ligand of this soluble E2 — no PE coordinates exist on this frame; LC3→PE lipidation (ATG7/ATG3) is a separate reaction, so this page names PE but cannot draw it. Not a pure A-page (covalent E2~cargo).
Crystal structures PDB 9E8P — human GABARAP–ATG3 conjugate, 2.70 Å (Ohashi 2025). Coordinates are public. Residue 264 is LYS (C264K). 4NAW is an ATG3 peptide fragment on an ATG12~ATG5–ATG16N complex — not the E2 fold, not this mesh. Human ATG3 UniProt Q9NT62.
Mechanism — E2 transthiolation ATG3 is the autophagy E2: E1~LC3 + ATG3-SH → ATG3~LC3 + ATG7. Compiled Cys264 thiol attacks the E1 thioester carbonyl. This deposit captures a C264K isopeptide stand-in, not the WT thioester. Ohashi et al., Biochemistry 2025, PMID:40628661. EC 2.3.2.- (UBA-like E2). ATG12 cargo is a different path and is not this page. Source: Ohashi 2025 · UniProt Q9NT62
0 — Setup: pocket shown empty, Lys264 analog present 9E8P holds C264K-conjugated GABARAP. The pocket is shown empty so ordered transfer can start. Opaque, run:null (no autoplay). Lys264 ★ is the compiled Cys analog (UniProt Cys264). Teaching cargo is LC3; this fold is GABARAP.Source: PDB 9E8P (2.70 Å)
① — E1~LC3 thioester binds Schematic E1~LC3 thioester docks as the ATG7 product — [schematic], not deposited. Teaching cargo is LC3; this crystal cargo fold is GABARAP. Carbonyl faces Lys264 NZ. ATG12 is not this page.Source: compiled E1→E2 handoff · ATG7 page
② — Cys analog attacks E1 thioester → ATG3~cargo (hypothetical, lit.-based) Blue nucleophile rides from Lys264 NZ (Cys analog) to the schematic E1~LC3 carbonyl. Thioester fades in place to crystal GABARAP Gly116 — the deposited C264K isopeptide C-terminus. This is not WT Cys thioester occupancy. Slow-motion on purpose. Blue = nucleophile, not hydride.Source: compiled E2 transthiolation · UniProt Q9NT62 Cys264
③ — Conjugate named — C264K isopeptide Crystal GABARAP Gly116 stays as the deposited conjugate C-terminus. Lys264 NZ ⋯ Gly116 C = 1.33 Å. Arg265 NH1 ⋯ Gly116 O = 2.32 Å. This is C264K, not WT Cys264 thioester. Schematic E1~LC3 is gone.Source: 9E8P LINK Lys264 NZ—Gly116 C
④ — ATG3~cargo stays — ATG4 next Crystal GABARAP Gly116 stays conjugated at Lys264. Teaching identity is ATG3~LC3 (LC3 vs GABARAP hedge). E1 has left. This page does not lipidated LC3-II. ATG4 next (pro-LC3 cut / delipidation). 4NAW is not this fold.Source: autophagy E2 bookkeeping · Ohashi 2025

How to read this page — a first-pass primer First pass: Setup (pocket empty, Lys264 ★, C264K mesh) → ① schematic E1~LC3 docks → ② blue Cys analog → carbonyl, fade to crystal Gly116 → ③ conjugate named (1.33 Å) → ④ conjugate stays; ATG4 next. Second pass: 1.33 / 2.32 / 2.25 Å are crystal facts on 9E8P; E1~LC3 is [schematic]. Blue = nucleophile. Third pass: open “Disagreements” for C264K ≠ WT Cys, LC3 vs GABARAP, 4NAW fragment, and ATG12 not this page.
Chemistry primer — E2 transthiolation ATG3 takes LC3 from ATG7 by attacking the E1 thioester. Compiled Cys264-SH attacks that carbonyl; ATG7 leaves and the product is ATG3~LC3, the species that ATG12~ATG5 later presents to PE. This deposit uses C264K so the conjugate is an isopeptide, not a thioester — named. Gold is unused on this page (no catalytic H⁺ sphere). Blue = nucleophile (Cys analog → carbonyl), not hydride. This is not ATG7 (E1) and not ATG12 cargo. PE (phosphatidylethanolamine) — the membrane-lipid anchor of LC3-II — is absent from 9E8P, so it is named, never drawn.
Words used on screen — C264K · LC3 vs GABARAP · E2 C264K — residue 264 is LYS on this deposit; UniProt ACT_SITE is Cys264. The LINK is Lys NZ—Gly C, not Cys SG—Gly C. LC3 vs GABARAP — teaching cargo is LC3; this fold is GABARAP. E2 — ubiquitin-like conjugating enzyme (transthiolation from E1). Isopeptide occupancy — deposited Lys264—Gly116, not WT thioester.

Legend & Fidelity

Mechanism fidelity — measured vs. inferred The measured layer: 9E8P (2.70 Å) human ATG3 with GABARAP; Lys264 NZ ⋯ Gly116 C = 1.33 Å; Lys264 NZ ⋯ Gly116 O = 2.25 Å; Lys264 CE ⋯ Gly116 C = 2.47 Å; Arg265 NH1 ⋯ Gly116 O = 2.32 Å; Arg265 NH1 ⋯ Gly116 C = 3.34 Å; Glu112 OE1 ⋯ Lys264 NZ = 3.59 Å. The inferred layer ([schematic]): E1~LC3 thioester, transferring nucleophile sphere, product identity as ATG3~LC3. Binding order is compiled staging, not a kinetic fit. The protein mesh is static.
Disagreements on record (a) C264K ≠ WT Cys264. Residue 264 is LYS; the conjugate is an isopeptide, not a thioester. (b) LC3 vs GABARAP. Teaching cargo is LC3; this fold is GABARAP. (c) Isopeptide occupancy ≠ E2 thioester. Crystal holds Lys264—Gly116, not Cys-SG~cargo. (d) 4NAW is an ATG3 peptide fragment, not this E2 fold. (e) ATG12 is a different cargo and is not this page. (f) Gap 86–193 is missing on this deposit, named not drawn. (g) EC 2.3.2.- is a class, not a fully assigned number. (h) no protein-domain swing.
Regulation — ATG12~ATG5 E3-like · starvation This mesh is ATG3 holding GABARAP/LC3 analog. The ATG12~ATG5·ATG16L1 conjugate is the E3-like scaffold that presents ATG3~LC3 to PE — that conjugate lives on the ATG5 occupancy page, not here. Starvation / mTOR / ULK1 gating of autophagy is drawer, not motion. Flexible region 86–193 is missing on 9E8P.Source: UniProt Q9NT62 · Ohashi 2025
Kinetics scope (no measured timecourse on this page) One E1~LC3 + ATG3 → ATG3~LC3 event is drawn, not a steady-state run. No measured k_cat/K_m is re-asserted here — no measured k_cat/K_m on this page [schematic]. Pair with a kinetics chapter.Source: Ohashi 2025 — no k_cat/K_m re-asserted
Sequence Human ATG3 UniProt Q9NT62 · NCBI Gene 64422. This screen uses PDB numbering (Lys264 on 9E8P = UniProt Cys264 engineered to Lys). Chain A Cα 27–310, gap 86–193. GABARAP chain C Cα 2–115 offset +1000. Gly116 is the ligand, not on the tube. GABARAP UniProt O95166.
Simulation map — what you see is what is written Setup = 9E8P pocket empty (opaque, no autoplay). ① schematic E1~LC3 moves in. ② slow-mo: blue Lys264 NZ → carbonyl; thioester fades to crystal Gly116. ③ conjugate named (1.33 Å). ④ Gly116 stays; ATG4 next. Only these events move.
Sources — bibliography of this page S1 — PDB 9E8P: human GABARAP–ATG3 conjugate, 2.70 Å — public coordinates; measured NZ⋯C = 1.33 Å · NH1⋯O = 2.32 Å S2 — Ohashi et al., Biochemistry 2025, PMID:40628661 — GABARAP–ATG3 backside interaction; C264K conjugate S3 — Sequence: UniProt Q9NT62 (ATG3, ACT_SITE Cys264) · O95166 (GABARAP) · NCBI Gene 64422 S4 — Kinetics: no k_cat/K_m re-asserted on this page

› Backbones and ligand poses are taken directly from the listed PDB entries, superposed into one common frame. Where no crystalline state exists for a step schematically and labeled as such in the narration. PE (the membrane-lipid anchor of LC3-II) has no coordinates on 9E8P or any ATG page, so it is named, never drawn.

ATG3GABARAPONnucleophile (blue)