Caryophyllene: the terpene that met a cannabinoid receptor

Can a flavour molecule act on a receptor?

Before 2008, beta-caryophyllene sat in a quiet drawer. It was classified as a sesquiterpene, and the people paying attention were flavour chemists, perfumers and essential-oil analysts. A well characterised molecule with no particular link to cannabinoid pharmacology.

Then came a screening result. In the work of Gertsch and colleagues from 2008, beta-caryophyllene was reported to bind one of the two known cannabinoid receptors directly, and the authors attached a label to it: dietary cannabinoid [1]. The reasoning behind that phrase was plain. The molecule is a common constituent of food plants, so it turns up in ordinary meals and not only in cannabis, and the phrase went straight into the title of the paper [1].

Which receptor is the whole story. Screening against both cannabinoid receptors found nanomolar affinity at CB2 and full agonist behaviour there, alongside no measurable affinity at CB1 [1]. Pharmacology has a word for that pattern. Selectivity: activity at one receptor subtype and effectively none at the other. It is a narrow claim, and the narrowness is what makes it interesting.

• Status before 2008: a sesquiterpene studied mainly by flavour chemists, perfumers and essential-oil analysts.
• CB2: nanomolar affinity and full agonist behaviour in the 2008 screening by Gertsch and colleagues [1].
• CB1: no measurable affinity in the same screening [1].
• Selectivity: the term for activity at one receptor subtype and effectively none at the other.
• Oral administration: in preclinical models, the 2008 study reported receptor-dependent responses, which were absent in CB2-null models [1].
• The label: dietary cannabinoid, justified by the molecule being a common constituent of food plants and used in the study title [1].

Two rings and a moving boiling point

The "beta" in beta-caryophyllene is functional, not decoration. It identifies the bicyclic molecule that the 2008 study worked with, meaning the structure built around two joined rings, and it separates that molecule from its close relatives [1]. If a source drops the prefix, you cannot be certain which of those relatives is on the page. In a field where one receptor result belongs to one structure, that matters.

The physical behaviour is easier to pin down. Beta-caryophyllene boils at roughly 254 to 262 °C at atmospheric pressure. Under reduced pressure the figure drops considerably, which is why distillation values quoted in the literature vary so widely. Two sources can both be correct and still disagree by a wide margin, because they are describing different pressures. Anyone comparing numbers needs the pressure alongside the temperature, or the comparison means nothing.

This is also why the molecule had a professional audience long before cannabinoid receptors entered the conversation. Flavour and fragrance work depends on knowing how a compound behaves when heat is applied, how it survives a distillation step, and what it contributes to a finished profile. Those questions were being answered while the receptor question had not yet been asked.

• The prefix: beta marks the bicyclic structure used in the 2008 receptor work, not a stylistic variant of the name [1].
• Boiling point: approximately 254 to 262 °C at atmospheric pressure.
• Under vacuum: much lower, which explains the spread of distillation figures across published sources.
• Reading a number: a boiling point without a stated pressure is not comparable to one that has it.
• First audience: flavour chemists, perfumers and essential-oil analysts, working on the molecule long before 2008.
• Occurrence: food plants, the basis on which the 2008 authors chose the word dietary [1].

Nine steps from screening to claim

1. Start with the receptors. Two cannabinoid receptors were known, CB1 and CB2, and the screening reported by Gertsch and colleagues in 2008 covered both of them rather than one [1].
2. Measure binding. At CB2, the 2008 study found affinity in the nanomolar range, which is the region where a molecule is considered to bind a receptor properly rather than incidentally [1].
3. Establish the mode. Binding alone says little, so the same study characterised the behaviour at CB2 as that of a full agonist, meaning the receptor is switched on rather than partly occupied [1].
4. Check the second receptor. Against CB1, the 2008 screening found no measurable affinity at all, which is a result in its own right and not a gap in the data [1].
5. Name the pattern. One subtype active, the other effectively untouched, is what pharmacologists call selectivity, and it is the shape of the whole finding [1].
6. Move beyond the plate. Oral beta-caryophyllene was then examined in preclinical models, where the 2008 study reported responses that depended on the receptor being present [1].
7. Remove the receptor. Those same responses were absent in CB2-null models, and that absence is what ties the observations to CB2 instead of to some unrelated mechanism [1].
8. Justify the label. Because the molecule is a common constituent of food plants, the authors described it as a dietary cannabinoid and put the phrase in the title of the 2008 paper [1].
9. Stop at the limit. The study measured no CB1 activity, so any question about intoxication is a separate question that this dataset does not answer [1].

One receptor engaged, one untouched

Two points are worth keeping straight, because they are the ones that get muddled fastest.

The first concerns CB1. CB1 is the receptor central to THC intoxication, and the 2008 study by Gertsch and colleagues found no measurable beta-caryophyllene activity there [1]. Activity at CB2 does not imply an effect comparable to activity at CB1. Two receptors, two pharmacologies, and in this case one of them engaged [1]. That distinction is not a technicality. It is the difference between a receptor result and an experience, and the study only speaks to the first.

The second concerns the word cannabinoid itself. Many readers hear it and picture something shaped like THC. The definition is broader than that. It covers the body's own signalling molecules, anandamide and 2-arachidonoylglycerol, better known as 2-AG, and neither of those is structurally THC-like. Cannabinoid describes what a molecule does at these receptors, not what family tree it comes from. Under that definition, and on the basis of the 2008 report, beta-caryophyllene qualifies at CB2 [1].

So the terse answer to the question people actually ask is this: a terpene that behaves as a cannabinoid at one receptor, according to the study that reported it [1]. Not a THC relative. Not an intoxicant. A selective CB2 agonist in the dataset that carries the claim [1].

It is also worth noting what the 2008 paper did not do. It did not measure the compound in a finished extract, and it did not quantify what any given amount does in people. Screening data and human data are different currencies. We keep them in separate columns, and we say so plainly rather than letting a receptor result stand in for something it never measured.

Where the 2008 finding landed

The 2011 synergy proposal

The receptor result arrived in the middle of an open argument. Do terpenes contribute anything to the activity of a whole-plant extract, or do the cannabinoids do the work on their own? In 2011, Russo published a review proposing phytocannabinoid-terpenoid synergy, compiling the individual terpene data available at the time and setting out where plausible interaction points might sit [2]. That proposal is the origin of the entourage effect discussion, the idea that the components of a hemp extract may act together rather than in isolation [2]. Within that argument, the beta-caryophyllene work is the one hard receptor measurement anyone can point to, because it names a receptor, an affinity range and a mode of action [1]. The rest of the picture remains a proposal in need of testing [2].

Reporting practice since 2014

Our approach here has not changed since 2014. Report what a study found, name the study, and stop where the evidence stops. One of the first CBD brands in the world does not need to stretch a screening result into a promise, and Swiss-minded work means writing the limit down next to the finding.

Which is why this page gives you a receptor, a pressure-dependent boiling point and two citations rather than a conclusion about how anyone feels. The 2008 data is specific [1]. The 2011 framework is a hypothesis [2]. Both are worth knowing, and neither is worth overstating.