A compressed Yiwu sheng pu’er cake on white wrapping paper.

Tea and Terps · Episode 8

Biosynthesis and Yiwu

Nik and Travis follow how a cannabis plant builds its cannabinoids, terpenes and cannflavins from a few shared carbon building blocks: one pathway that works like an assembly line, one that improvises, and the prenyl groups the plant attaches to almost everything. In the cup is a 2018 Yiwu sheng pu'er cake, and the tea brings its own chemistry: catechins, galloyl groups and the many teas one leaf can become.

  • Nik and Travis
  • Yiwu sheng pu'er
  • 1:48:39

Yiwu Zheng Shan 2018 cake, photographed by Nik in an earlier tea session.

The cup

Yiwu Zheng Shan, a pressed sheng from 2018

Nik's tea for this episode is a 2018 Yiwu Zheng Shan from Tea Hong: a sheng pu'er from the Yiwu tea mountains in Mengla, Yunnan, pressed into a cake and left to age. Sheng is the raw style of pu'er. The leaf is withered, pan-fired to slow its enzymes, rolled and sun-dried, then steamed and pressed. Nik breaks pieces off by hand: this cake is pressed loosely enough to come apart, unlike the tightest cakes.

He rinsed the leaf and brewed 10 grams gong-fu in his 150 mL zini teapot, steeping it again and again through the whole conversation. Even the dry leaf smelled of dates, plums, apricots and raisins over tropical hardwood. The liquor deepened from gold toward red as the steeps went on, and an hour and forty minutes in he was tasting honey, ginseng root, mint and camphor, with the leaf still giving flavor.

Tea
Yiwu Zheng Shan / 易武正山
Origin
Yiwu, Mengla, Yunnan, China, 雲南・勐臘・易武
Style
Compressed sheng pu’er
Vintage
2018
Brewing
Gong-fu style, 10 grams of leaf, 150 mL zi-ni Yixing teapot by Hu Xiao Xiang (胡小香), Black-and-white cup
Source
Tea Hong

What Nik noticed

  • Dates, plums, apricots and raisins on the dry leaf
  • Tropical hardwood and a clean forest floor
  • Rum-soaked raisins with a minty aftertaste
  • Honey, ginseng root, mint and camphor in the late steeps

Special notes

Pressed sheng, prepared in Nik’s zini teapot.

Amber sheng pu’er liquor in a small cup on a red floral coaster.
Yiwu Zheng Shan 2018, brewed and photographed by Nik in an earlier tea session
Opened olive and copper sheng tea leaves and thick stems inside a clay pot.
Yiwu leaf after brewing, photographed by Nik in the earlier session
1 of 2
How a sheng pu'er cake is made
  1. PickedLeaves and buds from the Yiwu tea mountains
  2. WitheredThe leaves soften while they are still alive
  3. Pan firedHeat slows the leaf's enzymes
  4. Rolled and sun driedThis makes maocha, loose raw tea
  5. Steamed and pressedSoftened by steam and pressed into a cake
  6. AgedStored for years while the tea slowly changes

Chapter by chapter

Inside the episode

Press any time to play the episode from that moment. While it plays, the chapter you are in lights up.

  1. Opening
  2. Play from 0:17 An episode without a rehearsalNik and Travis did not prepare together this time, so the conversation pulls threads from earlier episodes together.
  3. Play from 5:16 A compressed Yiwu cakeNik shows the 2018 cake: leaf that was steamed, pressed and left to age, loose enough to break apart by hand.
  4. Play from 6:07 Pressing tea before pressing hashCakes can be pressed under tons of force, and on the tightest ones caffeine can crystallize on the leaf surface.
  5. Play from 7:49 Why pressed tea gets betterPressing began for transport. Unlike coffee, a pu'er cake improves for years, and this one is eight years old.
  6. The cup
  7. Play from 9:40 Yunnan, villages and priceFamous villages, tiny growing areas and marketing that pushes the best leaf past 20 dollars a gram.
  8. Play from 12:35 Why Yiwu tastes sweetYiwu is a whole sourcing region near the Laos border, known for sweeter, mellower pu'er than the punchy young cakes.
  9. Play from 14:18 Fruit and forest on the dry leafDates, plums, apricots, raisins and tropical hardwood, before any water touches the leaf.
  10. Play from 15:04 Wild trees and pigment chemistryUnpruned, unwatered tea trees, and a look back at how a carotenoid pigment breaks down into the violet scent of beta-ionone.
  11. Two assembly lines
  12. Play from 17:38 More than THCNik's central idea: breeders could select for a whole family of cannabinoids, starting with THCV.
  13. Play from 19:55 Two parent pathwaysThe polyketide side builds like an assembly line; the terpene side improvises, which is why no two terpene profiles match.
  14. Play from 23:58 Counting the carbons in THCSix carbons, plus three malonyl units that each leave one carbon behind as CO2, plus a ten-carbon GPP: 22 carbons, then 21 after decarboxylation.
  15. Play from 25:33 A prenyl group on everythingThe same five- and ten-carbon units go onto polyketides, flavonoids and some sulfur aroma compounds.
  16. Play from 28:43 A family of chain lengthsShorter and longer starters give THCV, THCB, THCH and THCP alongside THC.
  17. Play from 31:01 Isoprene units, stackedMonoterpenes have ten carbons, sesquiterpenes fifteen, carotenoids forty, and natural rubber keeps going.
  18. Play from 33:29 Rings have limits, terpenes do notA ring can only close on a short chain, while tens of thousands of terpenes come from a much smaller set of enzymes.
  19. Travis's lab results
  20. Play from 36:01 Steady THCV, swinging THCAcross five or six tests of one plant, THCV held near 5 percent while THC swung between about 9 and 12.
  21. Play from 42:00 Where terpenes startPyruvate and glyceraldehyde 3-phosphate, a product of photosynthesis, feed the plastid pathway behind every terpene.
  22. Play from 42:46 CO2 in the curing jarDecarboxylation, respiration and leftover metabolism may all add to the CO2 a sealed jar collects.
  23. Play from 45:51 THCV is decided at the startDivarinolic acid is not olivetolic acid with two carbons removed: the choice happens before the ring forms.
  24. Play from 48:32 One plant sharing out its resourcesWhy cannabinoids and terpenes are less in competition than one plant allocating, and a breeding target of high CBDA and THCVA.
  25. Play from 51:23 Minor cannabinoids in autoflowers?Travis wonders whether autoflowers carry minor cannabinoids that photoperiod lines lost.
  26. Play from 53:27 Cannflavins and sulfur compoundsThe same prenyl move shows up in the flavonoids of the leaf and in the plant's sulfur aromas.
  27. Play from 55:52 Top of the plant, bottom of the plantNo two profiles are the same, even top and bottom of one plant, and Travis plans a staggered-harvest test.
  28. Old trees and harvest windows
  29. Play from 59:19 A tree over 100 years oldSingle-tree tea, deep roots that reach a mineral layer, and why good oolong should taste like a soaked rock.
  30. Play from 1:02:36 Harvest windowsHops can have a harvest window of only a few days. Cannabis changes over its window too.
  31. Play from 1:04:08 GPP at the hubThe ten-carbon GPP goes into every cannabinoid and every monoterpene, so a plant's pull on it shapes the whole profile.
  32. Play from 1:09:21 What the next generation keptIn one small cross, Travis saw about half the ocimene and three quarters of the THCV carried into the offspring.
  33. Play from 1:11:59 One terpene, many jobsNik on terpenes as defenses: limonene against soft-bodied insects and fungi, and nerolidol alongside thrips.
  34. Tea chemistry
  35. Play from 1:17:09 Catechins and the galloyl groupCannabis adds prenyl groups; tea's specialty is the galloyl group, which gives catechins like EGCG.
  36. Play from 1:20:15 One leaf, every teaPolyphenol oxidase browns catechins like a cut apple, and makers steer it into white, green, yellow, black and pressed teas.
  37. Play from 1:23:55 Are tea aromas as fragile?Travis asks; Nik explains charcoal roasting, Maillard chemistry and terpenes bound in the leaf.
  38. Play from 1:28:49 Trichome towers and leaf cellsCannabis builds aroma in trichome heads far from the leaf; tea makes its catechins inside the leaf itself.
  39. Play from 1:30:23 Boiling water for white teaWhy white tea can take boiling water, and why Nik brews one Japanese black tea cooler to keep its yuzu note.
  40. Play from 1:31:13 Does the setting choose the molecule?Does the plant decide what a synthase makes, or does the molecule's surroundings? Travis suggests both.
  41. Play from 1:34:13 Do not toss the low-THC plantA plant that thrives in a strong room but tests low on THC is making something else. Follow it for a few generations.
  42. Play from 1:36:16 Beyond the flowerCannflavins in the leaves and compounds in the roots: parts of the plant few growers ever test.
  43. Closing
  44. Play from 1:39:57 Honey in the late steepsAn hour and forty minutes in, the leaf is still giving flavor: honey, ginseng root, mint and camphor.
  45. Play from 1:41:45 Nine hundred years of processingTea makers' processing shaped which plants they kept, down to the leaf thickness charcoal roasting needs.
  46. Play from 1:44:25 A message for breedersWork with what the plant can already express, rather than modifying it or spraying it.

Two pathways

One pathway works like an assembly line, the other improvises

Nik describes two parent pathways behind a cannabinoid. The polyketide side is neat and predictable. It starts from hexanoyl-CoA, a six-carbon starter, and an enzyme adds one small unit at a time, so each step has one expected product, like a square peg in a square hole.

The terpene side behaves differently. One starting molecule can feed several terpene synthases, and one synthase can turn the same starting molecule into several different products. That flexibility is real enzymology, and it is why Nik thinks no two terpene profiles will ever match exactly, even for the same cultivar in the same room under the same light. Travis pictured it as a factory with a steady conveyor belt and robot arms that improvise.

Two ways to build
Step The polyketide side The terpene side
Starts fromHexanoyl-CoA, six carbons GPP, ten carbons
Each enzymeAdds one unit, one expected product Can make several products from one start
Leads toOlivetolic acid, the ring in a cannabinoid Monoterpenes in shifting ratios
Nik's pictureAn assembly line The wild west

Many terpene synthases are known to make more than one product from a single substrate.

Carbon accounting

Counting the carbons in THC

Nik walked through THC one carbon at a time. The polyketide starts as hexanoyl-CoA, six carbons. Three malonyl-CoA units are added, three carbons each, and each one gives up a carbon as CO2 as it joins, so the chain grows by two carbons per step. Six plus six is twelve: olivetolic acid, the ring at the center of a cannabinoid, folded into shape by an enzyme called olivetolic acid cyclase.

Next comes the ten-carbon geranyl group from GPP, the same building block behind the terpenes. A second enzyme, a prenyltransferase, attaches it to olivetolic acid to make CBGA, 22 carbons, and another enzyme turns CBGA into THCA without losing any. The last carbon goes during drying, curing or heat: THCA loses CO2 and becomes THC, C21H30O2. Nik's point is that every step of the count is accounted for, and that the CO2 released along the way is part of the plant's own bookkeeping.

From six carbons to twenty-one
  1. Hexanoyl-CoAThe starter, six carbons
  2. Plus three malonyl-CoAEach adds two carbons and releases one CO2
  3. Olivetolic acidTwelve carbons, folded into a ring
  4. Plus GPPA ten-carbon geranyl group attached: CBGA, 22 carbons
  5. THCAThe plant's enzyme reshapes CBGA, still 22
  6. THCDecarboxylation releases CO2, 21 carbons left

The cyclase folds the ring; the prenyltransferase, a different enzyme, attaches the geranyl group (Gagne and colleagues, 2012; Luo and colleagues, 2019).

The varins

THCV is decided at the very first step

THCV is not THC with two carbons shaved off. Nik's point is that the difference is settled before any ring forms. When the plant starts the polyketide from a four-carbon starter instead of the six-carbon one, the chain becomes divarinolic acid rather than olivetolic acid. From there the same steps follow, CBGVA and then THCVA, and the finished THCV has 19 carbons to THC's 21.

Starters of other lengths give a whole family, from THCV up to THCP, most of them found only in small amounts. Travis has seen how steady this branch can be: across five or six lab tests of one plant, THCV held near 5 percent while THC swung between about 9 and 12 percent. Nik's question for breeders is why some plants favor the shorter starter, and whether that can be selected for.

One family, different starters
Cannabinoid Side chain Carbons
THCVThree carbons 19
THCBFour carbons 20
THCFive carbons 21
THCHSix carbons 22
THCPSeven carbons 23

The side chain comes from the starter that began the polyketide. THCB, THCH and THCP have been reported from cannabis in small amounts.

Real structures

Six molecules from the episode

Four steps on the road to THC and THCV, one cannabis flavone and the tea catechin from the cup, rendered from real PubChem geometry rather than drawn. Pick one to read why it matters, then turn it in three dimensions.

Olivetolic acid

The ring at the center of a cannabinoid: a six-carbon starter plus three malonyl units, folded by olivetolic acid cyclase.

Structure data: PubChem open chemistry database

Molecular structure

Drag to rotate. Scroll or use +/- to zoom.

C12H16O4

Olivetolic acid, twelve carbons.

Two specialties

Cannabis adds prenyl groups, tea adds galloyl groups

Attaching a five- or ten-carbon prenyl group is how cannabis makes a cannabinoid, and Nik pointed out that the plant does the same thing elsewhere. Cannflavins A and B are flavones carrying a ten- or five-carbon group, found in the leaves as well as the flowers. Some of the sulfur compounds behind skunky cannabis aromas also carry a five-carbon prenyl skeleton, although how the plant makes them is still unknown.

Tea has its own specialty. Both plants build flavonoids the same way, from phenylalanine and three malonyl-CoA units, and tea turns them into catechins. Then it attaches a galloyl group, seven carbons from gallic acid, to some of them. That gives the eight major catechins: four without the galloyl group and four with it, the best known being EGCG.

Same toolkit, different finish
Plant Cannabis Tea
Signature decorationA prenyl group, 5 or 10 carbons A galloyl group, 7 carbons
Added toOlivetolic acid, flavones Catechins
ExamplesCBGA, cannflavins A and B ECG, EGCG

Cannflavin A and B prenylation was shown in enzyme assays by Rea and colleagues (2019).

The maker decides

One leaf, every kind of tea

Every true tea starts as the same plant. What changes is how the maker handles the catechins in the leaf. An enzyme, polyphenol oxidase, browns them the way a cut apple browns, linking them into theaflavins and thearubigins that color and flavor the cup. Heat stops it, and breaking the leaf's cells speeds it up.

White tea is withered without rolling, so its cells stay whole, and it can still oxidize deeply over a long wither. Green tea is heated early to stop the enzymes. Yellow tea is covered while warm and moist so the leaves sweat, which builds sweetness and takes skilled work. Black tea is rolled to break the cells and fully oxidized. And sheng pu'er like Nik's is pressed into a cake and changes slowly for years. Even brewing follows the leaf: white tea can take boiling water, while Nik keeps one Japanese black tea at 185°F to hold on to its yuzu note.

Same leaf, different handling
  1. WhiteWithered, cells left whole
  2. GreenHeated early to stop the enzymes
  3. YellowCovered while warm and moist, so the leaves sweat
  4. BlackRolled to break the cells, fully oxidized
  5. Sheng pu'erPressed into a cake, changing for years

For breeders

Don't throw away the low-THC plant

The thread Nik kept returning to is breeding. Most selection chases THC and terpenes, but he sees more room on the polyketide side, where the chemistry is orderly: a plant that favors divarinolic acid could anchor a cultivar with high CBDA and THCVA and little THCA. He is clear that this is an idea to explore, not a method he has worked out.

His practical advice follows from it. If a plant thrives in a strong room but tests low for THC, its energy has not gone to waste: it is making something else. Look at the rest of the lab report before culling it, and follow it for a few generations. Tea makers did something similar over centuries, keeping only plants whose leaves suited their processing. Travis's numbers show why it is worth testing: one of his plants passed about half its ocimene and three quarters of its THCV to its offspring in one small cross.

One plant's lab results
  • THCabout 9 to 12%
  • THCVabout 5%
  • Total terpenesabout 3 to 8%

Travis's results for one indoor plant across five or six tests, as he recalled them on air. One plant, not a study.

Moments worth rewinding

Short cuts into the conversation

Moments from the clip archive. Each one plays the full episode from where it starts.

  • Why pu'er comes pressed into a cake

    The leaf is processed, softened with steam and pressed. Nik compares the pressure to a rosin press and explains why a sheng cake keeps changing for years while coffee is best fresh.

  • Same room, same light, different terpenes

    The polyketide side of a cannabinoid works like an assembly line. On the terpene side one starting molecule feeds several enzymes, which is why Nik expects no two terpene profiles to match exactly.

  • Where THC's 21 carbons come from

    Nik counts out loud: a six-carbon starter, three units that add two carbons each, a ten-carbon piece from the terpene side, and one last CO2 lost to heat.

  • The piece the plant attaches to everything

    Cannabis takes building blocks from its terpene pathway and attaches them to other molecules. The same move makes the first cannabinoid acid and the cannflavins, which are found mainly in the leaves.

  • So few enzymes, so many terpenes

    Five-carbon units stack head to tail, and a fairly small set of terpene synthases turns them into tens of thousands of products. Nik calls that flexibility "a pretty cool engine to have."

  • THCV is not THC with two carbons shaved off

    The difference is set at the start, with a building block two carbons shorter, before the ten-carbon piece is attached and the ring is closed. A finished THC molecule cannot be trimmed into THCV.

  • Breeding for the minor cannabinoids

    How much of each cannabinoid a plant makes is heritable. Nik floats an idea for breeders: find the plants that lean toward the varin precursor and breed with them. It is his idea, not a tested protocol.

  • Tea adds galloyls, cannabis adds prenyls

    Tea and cannabis start some signature molecules with the same chain-building chemistry. Tea finishes its catechins with a galloyl group, which gives EGCG, while cannabis attaches prenyl groups.

  • Tests low in THC? Read the rest of the test

    Which acid a plant makes from CBGA depends mostly on its genes. A healthy plant that tests low in THC is probably making something else, Nik says, so check the full test before you throw it out.