The Side-Chain Homologous Series and its Structure-Activity Relationship

The series is the organising idea

Cannabis does not make one cannabinoid scaffold with one side chain. It makes a homologous series, because the polyketide arm of the biosynthesis will accept several different short acyl-CoA starter units. Whichever starter the enzyme picks up sets the alkyl chain length for everything downstream, so each starter generates a parallel copy of the entire cannabinoid family: a CBG, a THC, a CBD, a CBC, a CBN and so on, all carrying the same chain. The pentyl (C5) series dominates because hexanoyl-CoA is the abundant starter; the propyl (C3) series is the next most abundant; the others occur at trace to ultra-trace level. Understanding this is the difference between memorising a list of acronyms and being able to predict what should exist.

• C1, methyl: the orcinol series. Δ9-THC-C1, CBD-C1, cannabiorcol (CBN-C1). Built from a shorter starter, giving orcinol rather than olivetol as the resorcinol.
• C3, propyl: the varinic series, suffix -varin. THCV, CBDV, CBGV, CBCV, cannabivarin (CBNV), CBNDV, CBLV, and the corresponding acids. Built from butanoyl-CoA via divarinic acid.
• C4, butyl: THCB (tetrahydrocannabutol) and CBDB (cannabidibutol), isolated from the Italian FM2 medicinal variety in 2019 to 2020.
• C5, pentyl: the standard series. Δ9-THC, CBD, CBG, CBC, CBN and the rest, from olivetolic acid.
• C6, hexyl: THCH (tetrahydrocannabihexol) and CBDH (cannabidihexol), also first reported from FM2.
• C7, heptyl: the phorolic series, suffix -phorol. THCP and CBDP, reported in 2019; CBGP and CBCP are putative identifications.
• C8, octyl: Δ8-THC-C8, a classical octyl cannabinoid known from the Adams-era synthetic work and also catalogued in the JWH numbering as JWH-138. This is the compound behind the trade name THCJD.

Sources: Citti C 2019 · Linciano P 2020 · Linciano P 2020 · Hanuš LO 2016 · Radwan MM 2021 · Martin BR 1999

The naming convention, because the suffixes are not obvious

The trivial names encode the chain length in a suffix that has no mnemonic logic, which makes the literature harder to read than it needs to be. The table maps them. Note that -varin is the only suffix most readers will have met, that -butol and -hexol were coined by the same group in consecutive papers, and that a compound written as THC-C7 in an analytical paper and THCP in a commercial one is the same molecule.

Sources: Citti C 2019 · Linciano P 2020 · Linciano P 2020 · Hanuš LO 2016 · Martin BR 1999

The structure-activity relationship: affinity rises, peaks, and falls

The relationship between side-chain length and cannabinoid receptor affinity is one of the oldest and best-replicated findings in the field, and it has a shape rather than a direction. Affinity is low for very short chains, rises steeply through the middle of the series, reaches an optimum in the region of seven to eight carbons, and then declines as the chain grows past the length of the hydrophobic channel it occupies in the receptor. The Martin and Crocker work of 1999, done on the Δ8-THC scaffold specifically so that the side chain could be varied without confounding the ring chemistry, mapped this in detail and found something more interesting than a potency curve: manipulating the side chain does not merely tune affinity, it changes the character of the interaction, so that one scaffold with different chains yields agonists, partial agonists and antagonists. That is the paper the phrase "delineates agonists, partial agonists, and antagonists" comes from, and it is the primary reference for anyone who wants the SAR rather than the marketing.

• Δ9-THC (C5), CB1 Ki approximately 40 nM. Partial agonist.
• Δ9-THCB (C4), CB1 Ki 15 nM, CB2 Ki 51 nM — described by its isolating group as comparable to Δ9-THC.
• Δ8-THC-C8 (C8), CB1 Ki 8.5 nM in the Martin 1999 series — roughly five times the affinity of Δ9-THC.
• Δ9-THCP (C7), CB1 Ki 1.2 nM, CB2 Ki 6.2 nM — roughly 33 times the CB1 affinity of Δ9-THC measured in the same study.
• Δ9-THCV (C3), CB1 Ki 75.4 nM at mouse brain membranes and 62.8 nM at human CB2, behaving as a competitive antagonist rather than a weak agonist.

Sources: Martin BR 1999 · Crocker PJ 1999 · Showalter VM 1996 · Citti C 2019 · Linciano P 2020 · Thomas A 2005

The modern confirmation, and why efficacy is a second axis in vitro

A 2026 study in Molecular Neurobiology profiled the C3 to C8 homologues of Δ9-THC, Δ8-THC and hexahydrocannabinol across two signalling outputs — inhibitory G-protein activation and beta-arrestin recruitment — and reproduced the classical curve while adding the axis the classical work could not measure. Side-chain elongation was the dominant determinant of signalling efficiency, with potency gains that peaked at the C7 and C8 homologues. The C4 through C8 homologues behaved as high-efficacy agonists in the G-protein pathways while remaining only partial agonists for beta-arrestin recruitment, which is pathway bias: the same molecule is a strong agonist on one output and a weak one on another. The C3 homologues showed minimal agonism and instead produced functional CB1 antagonism, partially inhibiting the G-protein pathways and almost completely inhibiting beta-arrestin recruitment. This is an independent confirmation, in a modern functional assay, of the 2005 finding that the propyl homologue reverses sign.

Sources: Durydivka O 2026 · Thomas A 2005 · Martin BR 1999

The propyl reversal is the most important qualitative fact in the series contested in vitro

Most of the side-chain series is a story about degree. The C3 homologue is a story about kind. Δ9-THCV was shown by the Pertwee group in 2005 to displace the reference radioligand from CB1 and CB2 sites with respectable affinity and then to antagonise agonist-induced signalling rather than produce it — a competitive antagonist at both receptors, with apparent KB values in the tens of nanomolar. The behaviour is dose-dependent and preparation-dependent: at higher concentrations, and in some tissues, THCV shows partial agonist character, and the literature contains both descriptions because both are real under their respective conditions. For a research programme this is the single most useful qualitative handle in the homologous series, because it means chain length is not a potency dial but a functional switch, and it means a chemotype rich in the propyl series is pharmacologically different in kind rather than merely weaker.

Contested — caveat. THCV is reported as a CB1/CB2 antagonist at the concentrations used in the 2005 binding and GTPγS work, and as a partial agonist at higher doses and in some tissue preparations. Both characterisations are in the literature; neither is wrong, and a flat statement that "THCV is an antagonist" omits the dose dependence.

Sources: Thomas A 2005 · Durydivka O 2026

THCP: what the 2019 paper actually reported animal

Citti and colleagues isolated Δ9-THCP from a medicinal cannabis variety, characterised it, and measured its receptor binding and its in vivo activity in the standard mouse tetrad. The binding result is the one that travelled: CB1 Ki of 1.2 nM and CB2 Ki of 6.2 nM, against approximately 40 nM at CB1 for Δ9-THC measured in the same work, giving a ratio near 33-fold. The commonly quoted "30 times more potent" is a rounding of that affinity ratio, and it is worth being precise about what it is a ratio of: binding affinity, not subjective potency, not dose equivalence, and not toxicity. The paper also reported in vivo cannabimimetic activity greater than that of Δ9-THC, which is a stronger claim than binding alone, and it reported natural abundance in the fractions of a hundredth of a percent — low enough that the compound is a research and analytical target in plant material rather than a bulk constituent. THCP has since been identified and quantified in recreational products by forensic laboratories, which is how a trace phytocannabinoid becomes an analytical routine.

Sources: Citti C 2019 · Caprari C 2024 · Showalter VM 1996

The THCJD 19-fold claim: not verified, and the located number is much smaller contested anecdotal

The claim that THCJD is approximately nineteen times as potent as THC circulates throughout the hemp trade and appears in both operator source documents. No primary source for a nineteen-fold figure was located in this pass. Searching for it returns vendor pages, cannabinoid-shop explainers and secondary summaries, all of which reference "studies" without citing one. What the primary literature does contain is a measured affinity for the octyl compound: in the Martin 1999 side-chain series, Δ8-THC-C8 has a CB1 Ki of 8.5 nM, which against Δ9-THC at approximately 40 nM is a factor of roughly five in binding affinity, not nineteen. That is a substantial number and it is consistent with the C7-to-C8 optimum, but it is not the trade figure. Until someone produces the primary source, the honest description is: the octyl homologue binds CB1 several-fold more tightly than Δ9-THC, and the nineteen-fold multiplier is an unsourced trade claim. The same treatment applies to every potency multiplier attached to a novel cannabinoid in commercial copy — ask for the assay, the species, the endpoint and the reference compound, and if those four are missing the number means nothing.

Contested — caveat. Trade claim without located primary support. The "19 times THC" figure for THCJD could not be traced to any peer-reviewed source. The located primary datum for the octyl homologue is a CB1 Ki of 8.5 nM (Martin 1999), roughly five-fold the affinity of Δ9-THC. Do not repeat the 19-fold figure as fact.

Sources: Martin BR 1999 · Van Kush Family Research Institute 2026* · Van Kush Family Research Institute 2026* · Showalter VM 1996

Affinity is not effect, and higher affinity is a toxicological question before it is a commercial one

A binding constant describes how tightly a molecule occupies a receptor. It does not describe what the receptor then does, how much of the molecule reaches the receptor, how long it stays, what its metabolites do, or what the experience is like. Three separate cautions follow, and all three are routinely elided in novel-cannabinoid marketing. First, affinity and efficacy are independent: the Martin and Crocker work exists precisely because one scaffold with different side chains spans agonism, partial agonism and antagonism at similar affinities. Second, pharmacokinetics intervene: a compound with thirty times the affinity delivered in a product whose dose-per-unit is unknown is not thirty times the experience, it is an unquantified exposure. Third, and most important, a more potent full-spectrum agonist at CB1 is a toxicological concern rather than a feature. The entire clinical history of the synthetic cannabimimetics is the demonstration of what happens when CB1 is driven harder and longer than a partial agonist can drive it. The phytocannabinoid homologues are not those compounds and this page is not claiming they are — but the direction of travel, toward higher CB1 occupancy in products whose dose arithmetic is frequently wrong, points at the same hazard.

Sources: Martin BR 1999 · Crocker PJ 1999 · Banister SD 2018 · Tai S 2014 · Durydivka O 2026

The historical precedent: side-chain modification was the first big potency lever ever found contested historical / ethnographic

This is not a modern discovery. The Adams group, working on synthetic Δ6a,10a analogues in the early 1940s because those were the isomers whose stereochemistry could be controlled, found that replacing the natural pentyl chain with a branched dimethylheptyl chain produced very large increases in activity — the historical literature describes gains of several hundred-fold. The dimethylheptyl motif went on to appear in the most potent classical cannabinoids ever made, including the Hebrew University HU series. The modern rediscovery of THCP as the naturally occurring heptyl homologue is therefore a rediscovery: the SAR was mapped on synthetic scaffolds eighty years before the plant was found to be doing a version of it.

Contested — caveat. The "several hundred-fold" figure for the dimethylheptyl side chain is reported in the historical and review literature by potency in whole-animal assays of the period, not by a modern receptor binding constant. It is not directly comparable to the Ki ratios quoted elsewhere on this page.

Sources: Adams R 1940 · Adams R 1941 · Martin BR 1999 · Crocker PJ 1999 · Citti C 2019