Olfaction: How Smell Works

Why olfaction matters

• It is the oldest sense. Chemoreception predates vision and hearing by hundreds of millions of years; even bacteria run chemotaxis toward nutrients. In vertebrates the olfactory system wires straight into the limbic system, which is why smell is bound so tightly to emotion, appetite, and memory rather than to cool, deliberate analysis.
• It uses the biggest gene family we have. Roughly 400 intact odorant-receptor genes (plus several hundred pseudogenes) make up about 2% of the human protein-coding genome. No other single job in the body commands so many genes — a measure of how much evolutionary investment went into distinguishing molecules in the air.
• It anchors flavor. What we call taste is mostly smell. The tongue reports only five qualities (sweet, salty, sour, bitter, umami); the richness of coffee, wine, or strawberry arrives through retronasal olfaction as volatiles rise from the mouth to the epithelium. Pinch your nose and a jellybean loses almost all its identity.
• Its loss is an early warning. Reduced smell (hyposmia) is one of the earliest signs of Parkinson's and Alzheimer's disease, often appearing years before motor or memory symptoms, because the olfactory bulb and entorhinal cortex are among the first regions to accumulate pathology.
• It made anosmia a household word. Sudden smell loss became a hallmark symptom of COVID-19, affecting a large fraction of cases. The virus attacks the supporting sustentacular cells of the epithelium rather than the neurons directly, which is why smell usually returns as the tissue heals — sometimes with distortions (parosmia) during the imperfect rewiring.
• It is a survival alarm. The nose detects hydrogen sulfide, smoke, and spoilage at concentrations far below any harmful dose. Natural gas is odorless, so utilities deliberately add tert-butyl mercaptan (thiol) so a leak is smelled at parts-per-billion — an engineering decision that leans entirely on human olfactory sensitivity.

How olfaction works, step by step

Smell begins when you inhale. Volatile molecules — small, hydrophobic, typically under 300 daltons — ride the airstream to the olfactory epithelium, a postage-stamp of tissue (a few square centimeters) high in the nasal cavity. Odorants dissolve into the overlying mucus, sometimes ferried by odorant-binding proteins, and reach the cilia of the olfactory receptor neurons, where the receptor proteins sit. Each mature neuron expresses exactly one type of odorant receptor out of the ~400 available — the "one neuron, one receptor" rule — enforced by a remarkable feedback mechanism that shuts off all other receptor genes once one is chosen.

Binding an odorant triggers a classic GPCR cascade. The receptor changes shape and activates the olfactory-specific G protein G_olf. Its α subunit stimulates adenylyl cyclase III (AC3), which converts ATP into the second messenger cyclic AMP (cAMP). cAMP binds directly to cyclic-nucleotide-gated (CNG) channels, opening them so Na⁺ and Ca²⁺ pour into the cilium. The rising Ca²⁺ then opens calcium-activated chloride channels (anoctamin-2, ANO2); because olfactory neurons accumulate chloride internally, Cl⁻ flows out, amplifying the depolarization. A few bound molecules thus become a large receptor potential — the cascade's gain is enormous.

When the receptor potential crosses threshold, the neuron fires action potentials down a thin unmyelinated axon that pierces the cribriform plate of the ethmoid bone and enters the olfactory bulb. Here the code is sorted spatially: every neuron expressing the same receptor — scattered across the epithelium — converges its axons onto the same one or two glomeruli, spherical knots of neuropil about 100–200 µm across. A single glomerulus becomes a labeled channel for one receptor type. An odorant that activates a particular combination of receptors therefore lights up a particular spatial pattern of glomeruli — a chemical image, or "odor map," on the surface of the bulb.

Inside each glomerulus, receptor axons synapse onto the dendrites of mitral and tufted cells, the bulb's output neurons. Lateral inhibition from periglomerular and granule cells sharpens the contrast between active and inactive channels, much as it does in the retina. The mitral and tufted axons bundle into the lateral olfactory tract and — this is the anatomical surprise — project directly to the cortex: the primary olfactory (piriform) cortex, the amygdala, and the entorhinal cortex, without the obligatory thalamic relay that every other sense must pass through. Only later does a secondary pathway route olfactory information through the mediodorsal thalamus to the orbitofrontal cortex for conscious identification. Piriform cortex reads the glomerular pattern as a whole, and that ensemble — not any single cell — is what you experience as "coffee" or "smoke."

The combinatorial code: 400 receptors, a trillion smells

The genius of olfaction is that it does not need a dedicated receptor for every smell. In 1999 Bettina Malnic, Junzo Hirono, Takaaki Sato, and Linda Buck showed the underlying logic directly: each odorant is recognized by a combination of receptors, and each receptor recognizes a range of odorants that share chemical features. A molecule might strongly activate receptor A, weakly activate B, ignore C, and mildly touch D. A second molecule that differs by a single carbon shifts that fingerprint, and the brain reads the shift as a different smell.

This is why olfaction behaves like a barcode reader rather than a keyboard of labeled buttons. With ~400 receptor types each responding in graded fashion, the number of distinguishable activation patterns is astronomically large — the basis for the 2014 estimate of more than a trillion discriminable stimuli. It also explains olfaction's notorious sensitivity to structure: (R)- and (S)-carvone, mirror-image enantiomers of the same molecule, smell of spearmint and caraway respectively because they dock differently into chiral receptor pockets. Move one functional group and rose can become rot.

Common misconceptions

• "Humans have a terrible sense of smell." This is a nineteenth-century myth. A 2017 review by John McGann in Science argued the "microsmatic human" idea traces to the anatomist Paul Broca's flawed reasoning, not data. Humans have a smaller relative bulb than dogs but a comparably rich receptor repertoire and outperform dogs on some odors; the perceptual limits are largely in labeling, not detection.
• "There is a map of tastes and smells with fixed spots." The old "tongue map" is fiction, and olfaction has no fixed one-smell-one-place code either. The glomerular "map" is a pattern of relative activity across many channels, not a single labeled line, and the same neuron participates in coding many different odors.
• "We smell with our nostrils / the whole nasal cavity." Almost all inhaled air bypasses the sensory sheet. The olfactory epithelium is a small patch high in the cavity; only a fraction of each breath is deflected up to it, which is why sniffing — which increases turbulent upward flow — sharpens weak smells.
• "Smell goes through the thalamus like the other senses." It does not, at least not obligatorily. Primary olfactory cortex receives bulb output directly. The thalamic (mediodorsal) route exists but is secondary, feeding orbitofrontal cortex for conscious discrimination — a genuine anatomical exception, not a simplification.
• "Pheromones drive human behavior through a vomeronasal organ." Rodents have a functional vomeronasal organ and V1R/V2R receptors, but the human vomeronasal organ is vestigial and the associated TRPC2 gene is a pseudogene. Claimed human pheromones like androstadienone have not held up in rigorous replication.
• "Odorant receptors exist only in the nose." Odorant-receptor genes are expressed ectopically in testis (OR1D2 in sperm), kidney, gut, and skin, where they sense metabolites and modulate physiology — an active frontier well beyond smell itself.

Olfaction vs vision vs taste: sensing compared

Feature	Olfaction (smell)	Vision (phototransduction)	Gustation (taste)
Stimulus	Volatile odorant molecules	Photons (light)	Dissolved tastants
Receptor class	GPCR (odorant receptors)	GPCR (opsins/rhodopsin)	GPCR (T1R/T2R) + ion channels
Number of receptor types	~400 in humans	4 (rod + 3 cone opsins)	~25 bitter, few sweet/umami
Second messenger	cAMP ↑ (Golf → AC3)	cGMP ↓ (transducin → PDE)	IP₃/Ca²⁺ (gustducin)
Response direction	Depolarizing (channels open)	Hyperpolarizing (channels close)	Depolarizing
Coding scheme	Combinatorial across receptors	Opponent + trichromatic	Labeled-line by quality
Thalamic relay	Bypassed (direct to cortex)	LGN of thalamus (obligatory)	VPM of thalamus (obligatory)
Neuron turnover	Continuous (adult neurogenesis)	None (photoreceptors permanent)	Taste cells renew ~10 days

The olfactory transduction cascade at a glance

Step	Molecular player	What happens
1. Capture	Odorant + mucus / OBP	Volatile dissolves, reaches cilia
2. Binding	Odorant receptor (GPCR)	Receptor changes conformation
3. G-protein	Golf (α subunit)	Activates adenylyl cyclase III
4. Second messenger	cAMP	Synthesized from ATP by AC3
5. Channel opening	CNG channel	Na⁺ / Ca²⁺ influx, depolarization
6. Amplification	ANO2 (Ca²⁺-gated Cl⁻)	Cl⁻ efflux boosts the receptor potential
7. Firing	Voltage-gated Na⁺ channels	Action potentials to the bulb
8. Convergence	Glomerulus (bulb)	Like axons pool onto one channel
9. Output	Mitral / tufted cells	Direct projection to cortex

Famous experiments and history

• Buck & Axel (1991). Working in Richard Axel's lab, Linda Buck used degenerate PCR against conserved GPCR sequences to hunt for a large receptor family expressed only in the nose. She found a multigene family of about 1,000 genes in rat — the largest in the genome — reported in Cell 65: 175–187. The paper opened the molecular era of smell and, thirteen years later, won them the 2004 Nobel Prize in Physiology or Medicine.
• One neuron, one receptor. Follow-up in-situ hybridization and single-cell studies (Ressler, Sullivan, Buck; Vassar, Chao, Sitcheran, Axel, early-to-mid 1990s) established that each neuron transcribes a single receptor allele and that neurons expressing the same receptor project to shared glomeruli — mapping the epithelial code onto a spatial code in the bulb.
• Malnic & Buck (1999). By recording calcium responses of individual mouse neurons to panels of odorants and then identifying which receptor each neuron expressed, Malnic, Hirono, Sato, and Buck proved the combinatorial code directly: each odorant is read by a unique combination of receptors, published in Cell 96: 713–723.
• Mombaerts glomerular mapping (1996). Peter Mombaerts and colleagues genetically tagged neurons expressing a single receptor and watched their axons converge with pinpoint accuracy onto specific glomeruli in the mouse bulb — the elegant "wiring" demonstration behind the odor map.
• The trillion-odor estimate (2014). Caroline Bushdid, Marcelo Magnasco, Leslie Vosshall, and Andreas Keller had subjects discriminate complex odorant mixtures and extrapolated a lower bound above one trillion discriminable stimuli (Science 343: 1370–1372), overturning the century-old "10,000 smells" figure — though the exact number remains debated on statistical grounds.