The long-tail opioid
IUPAC: (RS)-6-(dimethylamino)-4,4-diphenylheptan-3-one · C21H27NO · MW 309.45 g/mol · CAS 76-99-3 (racemate) · ChEMBL651
Methadone (racemic). A synthetic diphenylheptane opioid — no morphinan ring, built instead on an open diphenyl-ketone chain. It is two drugs in one molecule: R-methadone (levomethadone) carries almost all the µ-opioid agonism, while S-methadone (dextromethadone) is a weak opioid but the more potent NMDA-receptor antagonist and hERG/QT blocker. Long-acting, orally reliable, and the workhorse of opioid agonist therapy. Street/context names: dolo, done, meth (do not confuse with methamphetamine), physeptone, the green liquid, juice, biscuits.
Methadone's primary action is full agonism at the µ-opioid receptor (MOR / OPRM1), binding the orthosteric pocket with Ki = 13 nM (human MOR, [³H]DAMGO displacement; ChEMBL, J. Med. Chem. 2021). Like morphine and fentanyl it activates Gi/o, closing voltage-gated Ca²⁺ channels, opening GIRK K⁺ channels, and inhibiting adenylyl cyclase — hyperpolarising neurons and blunting nociceptive transmission. Unlike fentanyl, methadone is a largely unbiased/balanced agonist (no strong G-protein bias), and its distinguishing feature is not the opioid arm at all — it is the second target.
Methadone is also a non-competitive (open-channel) antagonist of the NMDA glutamate receptor, binding the intra-channel MK-801/PCP site. This NMDA block is carried predominantly by the S-(+)-enantiomer (dextromethadone) and sits in the low-micromolar range (Ki ≈ 3–8 µM at the [³H]MK-801 site; Gorman et al. 1997, Neurosci. Lett. 223 — literature value, not a ChEMBL binding record). It is pharmacologically real but weak relative to the nanomolar opioid affinity, so at analgesic plasma levels the NMDA contribution is modest — most relevant to blunting opioid tolerance, attenuating wind-up/hyperalgesia, and contributing to methadone's efficacy in neuropathic and opioid-rotation contexts.
Binds the MOR orthosteric site (Ki 13 nM, ChEMBL). Full efficacy Gi/o coupling → ↓cAMP, GIRK activation, Ca²⁺-channel inhibition → analgesia, euphoria, and — the fatal endpoint — brainstem respiratory depression at the pre-Bötzinger complex.
Non-competitive antagonist at the NMDA-receptor MK-801/PCP site (Ki ≈ 3–8 µM, literature). Dampens glutamatergic wind-up and NMDA-driven opioid tolerance — a proposed reason methadone retains potency after tolerance to other µ-agonists.
R-(−)-methadone (levomethadone) holds ~8–50× the µ-opioid potency of the S-enantiomer and drives analgesia. S-(+)-methadone (dextromethadone) is the stronger NMDA blocker and the stronger hERG blocker — the cardiac liability rides largely on the S-isomer.
Methadone blocks the hERG (Kv11.1) cardiac potassium channel, prolonging ventricular repolarisation (QTc). This is an off-target liability distinct from the opioid and NMDA actions — the mechanistic root of torsades de pointes (see §03).
Methadone has weak reuptake-inhibitory activity at SERT/NET reported in some assays, but affinity is far lower than at MOR. Dopamine-receptor binding is negligible/weak (D2 Ki > 100 µM; D3 Ki ≈ 2.1 µM; ChEMBL) — reward is opioid-mediated, not direct dopaminergic.
Unlike morphine (M6G) or codeine (→morphine), methadone's major metabolite EDDP is pharmacologically inactive. All opioid effect comes from the parent drug — so effect duration is governed entirely by the parent's long, variable half-life.
Methadone's defining danger is not potency — it is time. The elimination half-life is long and extraordinarily variable: roughly 15–60 hours, and it can stretch past 120 h in some individuals. Critically, the analgesic duration (4–8 h) is far shorter than the elimination half-life. A patient re-doses for pain relief while the drug is still accumulating in deep tissue compartments (Vd ≈ 4 L/kg, highly lipophilic, ~85–90% protein-bound to α₁-acid glycoprotein). Plasma levels keep climbing for days after a dose change even though the felt effect has worn off.
Because half-life > duration of effect, methadone reaches steady state only after 5–10 days (5 half-lives). Deaths cluster in the induction phase (first 1–2 weeks) and after any upward dose adjustment: the patient feels under-medicated, takes more, and the trough plasma level silently ratchets up until it crosses the respiratory-depression threshold — often during sleep, 2–4 days in. "Start low, go slow" is not caution theatre; it is the physics of a drug whose feedback signal lags its real concentration by days.
Metabolism: Methadone is N-demethylated — chiefly by CYP2B6 and CYP3A4, with contributions from CYP2D6, CYP2C19 and CYP2C9 — to an unstable intermediate that spontaneously cyclises to EDDP (2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine), then to EMDP. Both are pharmacologically inactive. EDDP is the urine-screen marker that distinguishes genuine methadone dosing from spiked samples (parent-only urine suggests tampering).
The interaction minefield: because clearance rides on CYP3A4 and CYP2B6, methadone plasma levels swing wildly with co-medications. CYP inducers (rifampicin, carbamazepine, phenytoin, efavirenz, St John's Wort) can crash levels and precipitate withdrawal or — on stopping the inducer — cause rebound toxicity. CYP3A4 inhibitors (many azole antifungals, macrolides, some protease inhibitors, grapefruit) raise levels and stack respiratory-depression and QT risk. There is no safe "eyeball" dose adjustment; this is a therapeutic-drug-monitoring drug.
Racemic methadone means each dose is really two drugs with different clearances. CYP2B6 is polymorphic: slow-metabolisers (e.g. CYP2B6 *6/*6) accumulate the S-enantiomer preferentially — and the S-enantiomer is the one that blocks hERG. So the same milligram dose produces a higher QT liability in genetically slow CYP2B6 metabolisers, an under-appreciated driver of the torsades signal (see §03).
Methadone blocks the hERG (Kv11.1) channel that carries the rapid delayed-rectifier potassium current IKr — the current responsible for phase-3 repolarisation of the ventricular action potential. Blocking it lengthens the QT interval in a dose-dependent way. The S-enantiomer is the more potent hERG blocker, tying the cardiac risk back to the same isomer that carries the NMDA action and that accumulates in CYP2B6 slow-metabolisers.
Prolonged QTc is not itself the danger — it is the substrate. When repolarisation is delayed enough, early after-depolarisations can trigger torsades de pointes, a polymorphic ventricular tachycardia that can degenerate into ventricular fibrillation and sudden cardiac death. Risk climbs steeply above QTc ~500 ms, and rises with dose (especially >100–120 mg/day), with hypokalaemia and hypomagnesaemia, with structural heart disease, and — critically — with stacking other QT-prolonging drugs.
Practical monitoring: guideline practice is a baseline ECG, a repeat within ~30 days and annually, and again after dose increases or when adding an interacting drug — with dose review if QTc > 450–500 ms. This is why methadone induction belongs in a structured program, not a one-line script.
Set against that risk profile is one of the best-evidenced interventions in all of addiction medicine. Methadone maintenance / opioid agonist therapy (OAT) exploits the same long half-life that makes induction dangerous: once at steady state, a single daily oral dose holds plasma levels above the withdrawal threshold and below the euphoria/intoxication threshold for a full 24 hours. It occupies MOR, blunts craving, blocks the reinforcing "rush" of injected short-acting opioids by cross-tolerance, and removes the need to inject illicit, unpredictably-dosed supply.
Methadone (and buprenorphine) maintenance roughly halves all-cause and overdose mortality in people with opioid use disorder versus no treatment, and cuts injection-driven HIV and hepatitis C transmission, criminal-justice involvement, and overdose. Oral, once-daily, and long-acting is exactly what you want for a stabilising agent. The drug that is dangerous when titrated carelessly is protective when titrated properly — the pharmacology didn't change, the structure around it did. That is the whole thesis of harm reduction: meet the physiology where it is.
The naloxone caveat — read this twice. If methadone causes respiratory depression, naloxone works, but it does not last as long as methadone does. Naloxone's duration is roughly 30–90 minutes; methadone's effect lasts many hours to more than a day. After a naloxone bolus wears off, the still-abundant methadone re-binds MOR and the patient re-narcotizes — respiratory depression returns, often after the bystander or even the ED thinks the crisis is over.
Methadone is the antithesis of a rigid pharmacophore. Its open diphenyl-heptanone chain has
seven rotatable bonds (ChEMBL rtb = 7) versus the caged bicyclics of morphinan opioids. In the
free state this floppy chain populates a broad ensemble of conformers — a high ligand conformational
Shannon entropy H_lig. To bind MOR it must fold into a single receptor-competent pose,
collapsing that ensemble.
In FlexAID∆S terms this is a substantial ligand conformational entropy penalty
(−TΔS_conf > 0): methadone pays an entropic tax on binding that a pre-organised ligand
like fentanyl largely avoids. That the observed affinity is still nanomolar (Ki 13 nM,
ΔG ≈ −RT·ln Ka ≈ −10.8 kcal/mol at 310 K) implies the enthalpic contact
term must be correspondingly favourable — the two phenyl rings and the protonated dimethylamino
nitrogen make strong, complementary contacts (aromatic stacking + the conserved Asp3.32 salt bridge
common to aminergic/opioid orthosteric sites) that overpay the entropy debt.
The receptor side shows the familiar pocket entropy collapse: the apo MOR orthosteric
site samples multiple side-chain rotamers (high H_pocket) that funnel into one arrangement
on agonism. Methadone's story is therefore a two-sided entropy ledger — flexible ligand and
flexible pocket both losing configurational entropy — which is precisely the decomposition FlexAID∆S is
built to resolve: separating the entropy the ligand pays from the entropy the protein pays, rather than
hiding both inside a single docking score.
Methadone is safe when respected and lethal when rushed. The killers are the accumulating half-life, the QT interval, and depressant combinations. Every one of these is manageable with information.
If you use, don't use alone, and don't stack depressants on a long-acting opioid. Methadone maintenance is a proven way to stay alive — engaging with a program is harm reduction, not surrender. Resources: TripSit · TripSit Combo · DanceSafe.
| Target | Affinity | Rel. | Mechanism |
|---|---|---|---|
|
MOR
µ-opioid receptor (OPRM1)
|
Ki = 13 nM
pChEMBL 7.89 · primary
|
Full agonist | |
|
NMDA-R
GluN MK-801 / PCP site
|
Ki ≈ 3–8 µM
lit. (S-isomer)
|
Open-channel antag. | |
|
D3
Dopamine D3 receptor (DRD3)
|
Ki = 2,110 nM
weak
|
Weak binder | |
|
D4
Dopamine D4 receptor (DRD4)
|
Ki = 26,200 nM
negligible
|
Weak binder | |
|
CYP2B6
Cytochrome P450 2B6
|
Ki = 10 µM
metabolism route
|
Substrate/inhib. |