Isocitrate Dehydrogenase (NAD⁺) - 3D Mechanism Simulator

Loading structural data… (8GRU)

References & Databases

START HERE — how to read the isocitrate → α-KG picture IDH3 does oxidative decarboxylation in three textbook moves: isocitrate hands a hydride to NAD⁺ (making oxalosuccinate + NADH), oxalosuccinate loses CO₂ (the β-carboxyl), and the leftover enol is protonated to α-ketoglutarate. The rare luck of this page: the crystal 8GRU holds all three players at once — isocitrate, metal, NAD⁺ — in the closed active site, so the landing seats are measured, not posed. Two honest asterisks travel with every number: the metal on screen is Ca²⁺, the crystallization stand-in (the working metals Mn²⁺/Mg²⁺ never crystallized here), and the protein carries a Q139A clasp mutation far from the active site that made this complex trappable. Measured numbers (3.29 / 3.11 / 3.09 / 2.50 / 2.57 Å) reproduce from the deposited 8GRU coordinates; the intermediates relabel the same heavy-atom skeleton because H atoms are not drawn.
Crystal structure PDB 8GRU — human IDH3 α-Q139Aβ heterodimer with ICT + Ca²⁺ + NAD⁺ at the α active site (plus NAD⁺ at the β pseudo-allosteric site, text-only here), 2.85 Å, Chen X, Sun P, Liu Y et al., J Biol Chem 2022, PMID:36375638 — the first pseudo-Michaelis complex of HsIDH3 (the paper's own claim — not of IDHs in general). Contrast frame: WT αβ + NAD⁺ only, PDB 6KDY (Sun et al. 2019, PMID:31515270), whose metal site is distorted and binds no substrate — which is why 8GRU, not 6KDY, carries this page.
Mechanism — Lys184B–Asp230A–Tyr126A triad Lys184B ★ (β subunit, deprotonated by Asp230A ★) is the general base for the isocitrate α-OH; the C2 hydride goes to NAD⁺ C4 (re-face) giving oxalosuccinate; the C6 β-carboxyl leaves as CO₂; Tyr126A ★ protonates C3 to α-KG and Asp230A restores Tyr126 (Chen 2022 Fig.7 — by comparison with the HsIDH1 pseudo-Michaelis (6PAY) and the human NADP-IDH mechanism; E. coli QM/MM homolog Neves 2016; induced-fit trilogy Gonçalves 2012). The metal chelates the α-carboxyl + α-OH through both chemical steps. This triad replaces the old His263/Asp317 labels previously shown on this page — those residues are pocket-lining, not catalytic, and the Y126F mutant specifically kills catalysis (Dange & Colman 2010). Source: Chen et al. 2022 (PMID:36375638) — triad + pseudo-Michaelis · Gonçalves et al. 2012 (PMID:22891681) — 3-step induced-fit trilogy · Neves et al. 2016 — E. coli NADP-IDH QM/MM (homolog); ≈16.5 kcal·mol⁻¹ barrier, rate-limiting assignment contested — see ③ card · Dange & Colman 2010 (PMID:20435888) — αY126 catalytic acid · M-CSA entry 7 (homolog EcoIDH chemistry, disclosed as homolog)
0 — Setup: 8GRU pseudo-Michaelis The deposited closed enzyme holds isocitrate + Ca²⁺ + NAD⁺ exactly as drawn — opaque, run:null. Read the two asterisks before anything else: Ca²⁺ stands in for Mn²⁺/Mg²⁺ (Ca²⁺ inhibits the solution reaction), and the α-Q139A mutation lives in the distant clasp domain, far from the active-site chemistry, where it mimics the activator-bound interface. The β-subunit NAD⁺ at the pseudo-allosteric site is not drawn (text-only).Source: PDB 8GRU (2.85 Å), Chen 2022
① — Isocitrate + metal dock NAD⁺ waits in its 8GRU seat while isocitrate + metal move onto measured seats: the α-carboxyl onto the Arg88/98/119 clamp (2.50 Å salt bridge to Arg119 shown on step ③), the γ-carboxyl to Thr74/Ser82/Asn84, the α-OH onto the metal + Lys184B. No schematic posing anywhere in this step — the seats are the crystal's.Source: measured seats, PDB 8GRU (Chen 2022 Fig.6D contacts)
② — Hydride transfer → oxalosuccinate + NADH Proposed base coupling (Chen 2022); established hydride flight. (1) Lys184B ★ takes the α-OH proton (gold ball + arrow O7 → NZ, 3.11 Å seat — H⁺); (2) the hydride flies C2 → NAD⁺ C4 (blue ball + arrow, 3.29 Å, re-face — H⁻ carrying 2 e⁻, never gold) — oxalosuccinate + NADH. Because H atoms are not drawn, oxalosuccinate is the same heavy-atom mesh relabeled over the same seat (honest swap, stated on screen). Slow-motion on purpose so the two throws read in order.Source: Chen 2022 Fig.7 (Lys184B–Asp230A activation) · hydride re-face stereochemistry per ternary-complex structures (Gonçalves 2012)
③ — Decarboxylation → enol + CO₂ The β-carboxyl (C6) leaves as CO₂ — born at its deposited seat, out on its own vector — proposed as the rate-limiting step in one E. coli QM/MM study (≈16.5 kcal·mol⁻¹, Neves 2016), contested by KIE work pointing at the hydride/protonation steps. What stays is the enol on the way to α-KG. The frame contacts shown (Arg119…α-carboxyl 2.50 Å, Asp230…metal 2.57 Å) reproduce from 8GRU.Source: Neves 2016 (E. coli homolog; barrier figure cited with hedge) · measured contacts PDB 8GRU (see tools/manifests/idh3.json geomClaims)
④ — Protonation → α-ketoglutarate Tyr126A ★ protonates C3 (gold ball OH → C3); the enol collapses to α-ketoglutarate. Asp230A ★ restores Tyr126 through the relay (text-only — relay waters not drawn). Note the honesty bound: in the Michaelis frame Tyr126-OH sits 6.35 Å from C3 — the protonation happens after decarboxylation reshapes the pocket, so no distance is claimed here, only the residue assignment (Y126F kills catalysis).Source: Dange & Colman 2010 (αY126 catalytic acid) · Chen 2022 Fig.7 (Asp230→Tyr126 restoration)
⑤ — Products released — metal stays home [schematic exits] α-KG toward OGDH, CO₂ out, NADH toward the respiratory chain — each on its own vector (Rule 4). The metal stays home for the next turn; NAD⁺ repriming by exchange is text-only. Net: isocitrate + NAD⁺ → α-KG + CO₂ + NADH.Source: Rule 4 display convention — separate product exits; no external geometry claimed

Legend & Fidelity

Mechanism fidelity — measured vs. inferred The measured layer: the 8GRU α-Q139Aβ pseudo-Michaelis frame (2.85 Å) with ICT + Ca²⁺ + NAD⁺ as deposited seats, the 6KDY WT αβ + NAD⁺ contrast frame, and the five contacts reproduced from 8GRU into the manifest geomClaims (hydride seat 3.29 Å · Lys184B–α-OH 3.11 Å · 3.09 Å · Arg119–α-carboxyl salt bridge 2.50 Å · Asp230–metal 2.57 Å). The inferred layer (all [schematic]): every state past the Michaelis frame — the CO₂ birth vector and product exits (step ⑤ is tagged [schematic] outright), the enol→α-KG tautomerization, and the C3-protonation geometry, which happens only after decarboxylation reshapes the pocket (Tyr126…C3 sits 6.35 Å in the Michaelis frame, so no distance is claimed — an honesty bound, not a number). The Ca²⁺ screen metal is a stated stand-in for Mn²⁺/Mg²⁺ and the Q139A frame is a constitutively-active mutant, both disclosed rather than disguised. Where a transition cannot be attached to a crystalline state, the drawerNote carries the schematic framing.
Why is the "exits / protonation" "[schematic]"? 8GRU is pseudo-Michaelis: enzyme, substrate, cofactor and (stand-in) metal are all bound, but no chemistry is caught mid-flight. Decarboxylation is the first step that moves the scaffold, so the carbon skeleton, the CO₂ route and the α-KG protonation after it have no deposited coordinates at all — they are rebuilt on the same seats and tagged [schematic], with the incident Tyr126–C3 honesty bound stated instead of a fabricated distance. Everything drawn before that event travels on measured seats, which is why this page's schematic tag starts at the chemistry, not at the binding.
Catalytic residues — Lys184B · Asp230A · Tyr126A (+ clamp + metal ligands) Lys184B ★ = general base (α-OH deprotonation). Asp230A ★ = triad hub (activates Lys184B, ligates the metal at 2.57 Å, restores Tyr126). Tyr126A ★ = general acid (C3 protonation). Asp234A + Asp217B = metal ligands (deposited chelation geometry). Arg88/98/119 = carboxylate clamp (2.50 Å salt bridge measured). Thr74/Ser82/Asn84 = γ-carboxyl pocket. Asp268A = NAD⁺-vs-NADP⁺ specificity (ribose 2-OH recognition, Sun 2019 — text-only, no distance claimed on the 8GRU NAD pose). Active site spans both subunits: the α subunit alone cannot catalyze (Chen 2022).
Disagreements on record (a) Base identity. Lys184B–Asp230A coupling is Chen 2022's proposal from the 8GRU pseudo-Michaelis + HsIDH1 (6PAY) comparison and the human NADP-IDH mechanism; the older literature routes the base through Asp or water relays — drawn as proposal (step ② badge). (b) Ca²⁺ paradox. The screen metal is Ca²⁺, which inhibits the solution reaction; the working metals Mn²⁺/Mg²⁺ have no IDH3 co-structure — every metal contact is labeled stand-in. (c) Q139A. The frame is a constitutively-active mutant; WT needs CIT+ADP to reach this closed state — the regulation card tells that half. (d) Enol vs direct. Decarboxylation via an enol that tautomerizes follows Gonçalves/Neves; drawn as the consensus, flagged where the proton lands.
Words used on screen — oxidative decarboxylation · re-face · pseudo-Michaelis Oxidative decarboxylation — oxidation (hydride out) plus CO₂ loss in one enzyme visit; the oxidation pays for the decarboxylation. Re-face — the hydride lands on one specific face of the flat NAD⁺ nicotinamide ring (the classic stereochemistry, proven by isotope labeling). Pseudo-Michaelis — enzyme + substrate + cofactor + (stand-in) metal all bound, but no chemistry caught mid-flight; the closest crystallography gets to the starting line. Oxalosuccinate — the β-keto-acid intermediate; too unstable to isolate — it decarboxylates spontaneously and never leaves the active site (Chen 2022: 'unstable… quickly decarboxylated'). Gold ball + gold arrow = proton (H⁺); blue ball + blue arrow = hydride (H⁻, 2 e⁻) — the two flying H's never share a color, so step ② reads apart with the sound off.
Kinetics scope (no measured timecourse on this page) One isocitrate → α-KG event is drawn, not a steady-state run. IDH3 is the rate-limiting step of the TCA cycle, and its holoenzyme answers to S0.5,ICT ≈ 3.5 mM at rest → ≈ 0.4 mM with CIT + ADP (≈8-fold activation); the isolated αγ heterodimer activates ≈21-fold while αβ does not activate at all (Sun et al. 2020, rounded, attributed — see Sources). One E. coli QM/MM study puts decarboxylation highest on its computed hill (≈16.5 kcal·mol⁻¹, Neves 2016) — KIE studies contest the assignment, so no rate is drawn on screen.Source: Sun et al. 2020 (S0.5 values) · Neves et al. 2016 (barrier figure, hedged)
Regulation — CIT + ADP on, ATP + NADH off, metal required The γ subunit allosteric site (Arg97G/Tyr135G/Arg272G) binds citrate + ADP → closed active site, ≈8-fold S0.5 improvement (holoenzyme). The β pseudo-allosteric site cannot bind activators in WT (explains αβ's silence). ATP competes with ADP; NADH competes with NAD⁺ at the active site (Sun 2019). Mn²⁺/Mg²⁺ required, Ca²⁺ inhibits. TCA-cycle logic: high energy charge (ATP, NADH) throttles the committing step; ADP + citrate (substrate surplus) opens it.Source: Sun et al. 2020 (DOI 10.1038/s41421-020-00220-7) · Sun et al. 2019 (NADH competition, PMID:31515270)
Simulation map — what you see is what is written Setup = the 8GRU pseudo-Michaelis (ICT + Ca²⁺ + NAD⁺, opaque, run:null) with the Q139A/Ca²⁺ asterisks. ① ICT + metal move onto deposited seats; NAD⁺ label stays. ② STAR slow-mo: Lys184B gold arrow takes the α-OH proton (gold ball O7→NZ), hydride blue ball + blue arrow flies C2→C4N (3.29 Å seat); three measured contacts label in turn (3.29 / 3.11 / 3.09 Å) and clear; ICT fades as oxalosuccinate (same atoms) fades in; NAD⁺ relabels NADH. ③ two frame contacts label (2.50 / 2.57 Å) and clear; CO₂ is born at the C6 seat and flies out; the enol/α-KG skeleton stays. ④ STAR: Tyr126A gold arrow + gold ball OH→C3; α-KG final. ⑤ α-KG + CO₂ + NADH release on three vectors; metal stays. Only these events move; everything else is a static frame of 8GRU.
Sequence Human IDH3 α UniProt P50213 · β UniProt O43837 · γ UniProt P51553 · NCBI Gene 3419 (IDH3A). Screen shows 8GRU PDB-author numbering (α Tyr126/Arg88/Asp230 · β Lys184/Asp217 — displayed as deposited; B-suffixed keys are β-subunit).
Sources — bibliography of this page S1 — PDB 8GRU (2.85 Å): HsIDH3 α-Q139Aβ + ICT + Ca²⁺ + NAD⁺; Chen et al. 2022, J Biol Chem 298:102695, PMID:36375638 S2 — PDB 6KDY: WT αβ + NAD⁺ (contrast frame); Sun et al. 2019, JBC 294:16214–16227, PMID:31515270 S3 — Sun, Liu, Ma & Ding 2020, Cell Discov 6, DOI 10.1038/s41421-020-00220-7 (holoenzyme kinetics S0.5 + CIT/ADP activation; allosteric site residues) S4 — Gonçalves et al. 2012, Biochemistry, PMID:22891681 (induced fit; dehydrogenation→decarboxylation→tautomerization trilogy) S5 — Neves, Fernandes & Ramos 2016, ACS Catal 6:357–368, DOI 10.1021/acscatal.5b01928 (QM/MM of E. coli NADP-IDH — homolog, not human; the ≈16.5 barrier figure is quoted-with-hedge on the ③ card) S6 — Dange & Colman 2010, J Biol Chem 285:20520–20525, PMID:20435888 (αY126 catalytic acid; βY137 NAD/ADP role; γY135 required) S7 — UniProt P50213 (α) · O43837 (β) · P51553 (γ) · NCBI Gene 3419 · BRENDA EC 1.1.1.41 · M-CSA entry 7 (homolog mechanism, disclosed) S8 — Repo display conventions: Rule 4 (separate product exits) · Rule 34 geometry (5 geomClaims reproduced in tools/manifests/idh3.json) · [schematic]/same-seat relabel = stated on screen, no deposition claimed

› 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 (e.g., product exits, proton-relay waters), the pose is drawn schematically and labeled as such in the narration.

α-subunitβ-subunitONCa