The enzyme doesn't come back for two weeks
IUPAC: 2-phenylethylhydrazine · brand Nardil · sulfate salt · MW 136.20 g/mol (free base) · CAS 51-71-8 · ATC N06AF03 · first approval 1961
Phenelzine (Nardil). A hydrazine irreversible, non-selective monoamine oxidase inhibitor — the pharmacological opposite of a reversible reuptake blocker. Where an SSRI politely occludes a transporter and washes out in a day, phenelzine is oxidised by its own target and then welds itself covalently onto the flavin cofactor, permanently killing the enzyme. Recovery requires the cell to transcribe and translate brand-new MAO protein — roughly a two-week job. This is the first-generation antidepressant that made the "cheese reaction" a household phrase.
Monoamine oxidase (MAO) is a flavin-dependent amine oxidase anchored to the outer mitochondrial membrane. Two isoforms — MAO-A (preferring serotonin and noradrenaline) and MAO-B (preferring phenylethylamine and, in gut/liver, dietary tyramine) — use a covalently bound FAD cofactor to oxidatively deaminate monoamines. Phenelzine is not a passive blocker: it is a mechanism-based (suicide) inhibitor. The enzyme grabs phenelzine's terminal hydrazine as if it were a normal amine substrate, oxidises it, and in doing so generates a reactive species that forms a covalent bond with the N5 of the flavin ring.
The result is a permanently inactivated enzyme. Unlike moclobemide or harmine (reversible, competitive MAO-A inhibitors that dissociate), phenelzine's inhibition does not wash out with the drug. Plasma phenelzine is cleared within a day, but the enzyme it destroyed stays dead. Restoration of monoamine-oxidising capacity depends entirely on de novo synthesis of new MAO protein, which is why the pharmacodynamic footprint (and the danger window) outlasts the drug's half-life roughly fifteen-fold.
MAO-A oxidatively deaminates 5-HT, noradrenaline and dopamine. Covalent inhibition (IC50 ~30 nM, rat brain) raises intraneuronal and synaptic monoamines — the antidepressant/antipanic mechanism. This isoform's loss is also what makes dietary tyramine dangerous.
MAO-B (IC50 ~76 nM) handles phenylethylamine, benzylamine and a large share of gut tyramine metabolism. Non-selective blockade means both the neuronal and the gut/hepatic tyramine firewalls are down simultaneously.
MAO oxidises the –NH–NH2 hydrazine, generating a diazene/aryl-radical intermediate that alkylates flavin N5. The drug is destroyed and the enzyme with it — a one-shot, stoichiometric kill visualised in the covalent-inhibitor MAO-B structures.
A phenelzine metabolite, β-phenylethylidenehydrazine (PEH), inhibits GABA-transaminase and elevates whole-brain GABA. This GABAergic secondary effect is thought to underlie phenelzine's anxiolytic edge over other MAOIs.
Phenelzine also potently inactivates the copper-containing semicarbazide-sensitive amine oxidase VAP-1/AOC3 (IC50 ~20 nM, human) — another hydrazine-sensitive target, relevant to its vascular/off-target profile.
Because LSD1 is itself a FAD-dependent amine oxidase, phenelzine covalently inhibits it too — captured directly in PDB 6NR5 (human LSD1 + phenelzine). A real epigenetic off-target, now studied for its own sake.
Phenelzine's defining PK feature is a total disconnect between plasma half-life and pharmacodynamic duration. The parent compound is well absorbed orally and eliminated with a plasma T½ of only ~11.6 hours, yet the enzyme inhibition it produces lasts roughly two weeks. Standard steady-state PK reasoning does not apply: you are not dosing to maintain a plasma level, you are titrating cumulative, irreversible enzyme destruction.
Metabolism is a branch point that matters clinically. Phenelzine is partly oxidised (in part by MAO itself, an autocatalytic quirk) to phenylacetic acid and phenylethylamine-derived products, and partly N-acetylated by NAT2. A minor oxidation product, β-phenylethylidenehydrazine (PEH), is pharmacologically active as a GABA-transaminase inhibitor. NAT2 acetylator status shifts how much parent drug survives to inhibit MAO versus being shunted to acetylated metabolites.
The practical consequence: stopping phenelzine is not the same as clearing phenelzine. Even after the last dose has been eliminated, MAO remains inhibited until enough new enzyme is synthesised. Every guideline built around this drug — the tyramine diet, the serotonergic washout, the surgical/anaesthetic precautions — is timed to enzyme resynthesis (~14 days), not to plasma clearance.
Reversibility is the single most important pharmacological fact about phenelzine. A reversible inhibitor (moclobemide, harmine) sits in the active site in equilibrium with free drug — remove the drug, and the enzyme is fine within hours. Phenelzine forms a covalent bond. There is no equilibrium to shift. The only route back to functional enzyme is ribosomal synthesis of fresh MAO protein, governed by the enzyme's turnover rate, which in humans takes on the order of 10–14 days to restore meaningful capacity.
This is why every serious phenelzine interaction carries a two-week tail. Switching from phenelzine to an SSRI, SNRI, triptan, tramadol, or another serotonergic requires a 14-day drug-free interval — not because phenelzine is still in the blood, but because MAO is still offline. Conversely, when switching to phenelzine from a long-half-life serotonergic like fluoxetine, the washout runs the other way and stretches to five weeks (fluoxetine/norfluoxetine clearance). Emergency clinicians treat any patient who has taken phenelzine within the last two weeks as fully MAOI-active.
Clinically, phenelzine is a broad-spectrum antidepressant with particular strength in atypical depression and panic/social anxiety. Its efficacy signature differs from selective agents precisely because it removes the shared clearance enzyme for all the monoamines at once, and layers a GABAergic effect on top.
Loss of MAO-A raises intraneuronal and synaptic 5-HT across cortical and limbic circuits, driving the antidepressant and antipanic effect. This same 5-HT surplus is the substrate for serotonin syndrome when a second serotonergic agent is added — the therapeutic mechanism and the lethal-interaction mechanism are one and the same.
Elevated noradrenaline improves energy and motivation but sets up the cardiovascular hazard. With MAO gone, presynaptic terminals accumulate large releasable NA stores — the loaded gun that dietary tyramine or a sympathomimetic later fires as a hypertensive crisis. Paradoxically, chronic MAOI use can also cause orthostatic hypotension via accumulation of the false transmitter octopamine.
Reduced dopamine catabolism (both isoforms contribute) raises dopaminergic tone, contributing to relief of anhedonia and psychomotor retardation in atypical depression — an effect selective serotonergic drugs achieve poorly.
Uniquely among MAOIs, phenelzine (via its PEH metabolite) inhibits GABA-transaminase and raises whole-brain GABA. This enhanced inhibitory tone is a plausible basis for its robust anti-anxiety and anti-panic profile, distinguishing it mechanistically from tranylcypromine and from the reversible MAO-A inhibitors.
A reversible inhibitor collapses the conformational entropy of a binding pocket only while it is bound; release the ligand and the
pocket's Shannon entropy H_pocket recovers. Phenelzine represents the thermodynamic extreme of the entropy-collapse concept:
the covalent flavin adduct converts a reversible binding equilibrium into an irreversible chemical bond. In FlexAID∆S terms, the
final state is not a low-ΔS_conf complex that can re-expand — it is a permanently frozen microstate with the flavin N5
locked, entropy of the catalytic machinery driven toward zero and unable to recover without proteolysis and resynthesis.
Docking phenelzine as a non-covalent ligand into MAO would badly misprice it. The tiny 2-phenylethylhydrazine scaffold (MW 136, 3 rotatable
bonds) has poor shape complementarity to the hydrophobic MAO substrate cavity — a naïve ΔG_bind from a rigid-receptor score would look
weak. The pharmacology is not equilibrium affinity; it is kinact/KI, the rate of covalent inactivation. FlexAID∆S
flags this as a case where the entropic/enthalpic bookkeeping must include a bond-forming reaction coordinate, not merely a docking pose —
exactly the regime where equilibrium ΔS modelling hands off to covalent-mechanism modelling.
The visualised MAO-B covalent complexes (deprenyl, rasagiline analogues) make the geometry explicit: the inhibitor sits directly over the re-face of the flavin, poised for N5 attack. Phenelzine reaches the same endpoint by a different chemical route — hydrazine oxidation rather than propargylamine Michael addition — but the entropy fingerprint of the product is identical: a dead, rigid, single-microstate enzyme.
Phenelzine itself is a manageable antidepressant. What kills people is what they take with it — or within two weeks of stopping it. There are two independent lethal pathways: serotonergic combinations → serotonin syndrome, and tyramine / sympathomimetics → hypertensive crisis. Both are avoidable with information. Non-moralising, clinical, specific.
| Agent | Mechanism | Outcome | Risk |
|---|---|---|---|
| SSRIs / SNRIs | MAO block + reuptake block → runaway synaptic 5-HT | Serotonin syndrome (hyperthermia, clonus, rigidity, autonomic instability) | Fatal |
| MDMA / MDA | Massive 5-HT release into a synapse with no clearance enzyme | Fulminant serotonin syndrome + hyperthermia | Fatal |
| Tramadol · meperidine | Opioids with intrinsic serotonin-reuptake inhibition; also ↓ seizure threshold | Serotonin syndrome, seizures | Fatal |
| DXM (dextromethorphan) | Serotonin reuptake inhibition; common in OTC cough syrup | Serotonin syndrome | Fatal |
| Triptans | 5-HT1 agonism added to serotonergic overload | Serotonin syndrome | Avoid |
| Other MAOIs / linezolid | Additive irreversible MAO blockade | Serotonin syndrome + hypertensive crisis | Fatal |
| Tyramine-rich food | Gut/hepatic MAO firewall gone → tyramine floods circulation, displaces stored NA | Hypertensive crisis (severe headache, stroke, MI) | Fatal |
| Pseudoephedrine · decongestants | Indirect sympathomimetic releases the enlarged NA stores | Hypertensive crisis | Fatal |
| Amphetamine · cocaine | Sympathomimetic NA release / reuptake block on top of MAOI | Hypertensive crisis, hyperthermia | Fatal |
DPK + FAD),
2.2 Å (De Colibus, Binda et al., PNAS 2005). No phenelzine–MAO co-crystal exists; deprenyl is shown as an honest
structural analog of the same irreversible covalent flavin-adduct mechanism class. A real phenelzine–FAD covalent adduct is captured in the
related flavoenzyme LSD1 (PDB 6NR5). Rotate · scroll to zoom · right-drag to translate.
| Target | Potency | Rel. | Mechanism |
|---|---|---|---|
|
MAO-A
Amine oxidase [flavin] A (MAOA)
|
IC50 = 30 nM
pChEMBL 7.52 · rat brain
|
Irreversible | |
|
MAO-B
Amine oxidase [flavin] B (MAOB)
|
IC50 = 76 nM
pChEMBL 7.12 · rat brain
|
Irreversible | |
|
AOC3 / VAP-1
Semicarbazide-sensitive amine oxidase
|
IC50 = 20 nM
pChEMBL 7.70 · human
|
Covalent | |
|
GABA-T
GABA transaminase (via PEH metabolite)
|
Inhibitor
brain GABA ↑ (secondary)
|
Indirect | |
|
LSD1 / KDM1A
Lysine-specific demethylase 1
|
Covalent
FAD adduct · PDB 6NR5
|
Off-target |