Carrot top pesto through the looking glass

Isomers are molecules that have the same chemical constituents in different physical arrangements. Some terpenoid isomers have very different aromas and are important food seasonings. A batch of carrot top pesto led to an exploration of intriguing terpenoid isomers in the mint, carrot, and lemon families.

“Oh, c’mon. Try it,” my husband admonished me with a smile. “If anyone would be excited about doing something with them, I should think it would be you.”

The “them” in question were carrot tops, the prolific pile of lacy greens still attached to the carrots we bought at the farmer’s market. I have known for years that carrot tops are edible and have occasionally investigated recipes for them, but that was the extent of my efforts to turn them into food. My excuse is that I harbored niggling doubts that carrot tops would taste good. Edible does not, after all, imply delicious. My husband had thrown down the gauntlet, though, by challenging my integrity as a vegetable enthusiast. I took a long look at the beautiful foliage on the counter.

“Fine,” I responded, sounding, I am sure, resigned. “I’ll make a pesto with them.”

Carrot tops, it turns out, make a superb pesto. I have the passion of a convert about it, and not just because my carrot tops will forevermore meet a fate suitable to their bountiful vitality. The pesto I made combined botanical ingredients from two plant families whose flavors highlight the fascinating chemistry of structural and stereo isomers.

Isomers and the human senses

Isomers are molecules that have the same chemical constituents in different physical arrangements. My carrot top pesto included ingredients with somewhat similarly-flavored structural isomers, carvacrol (oregano flavor; I used oregano) and thymol (thyme flavor; I used thyme and ajwain), that differ only in the position of a phenolic hydroxyl (Xu et al. 2008). Even more fascinating than simple structural isomers, though, are the optical or stereo isomers of chiral molecules.

Chiral molecules are non-superimposable mirror images of one another. Chirality is often described as “handedness” of molecules, one member of the molecule pair being the “right hand” molecule, and the other the “left hand” molecule.

Chiral amino acids and hands, non-superimposable mirror images of one another (image from Wikipedia)

Many biologically important compounds are chiral. The proteins in all life forms on earth, for example, are almost entirely composed of left-handed amino acids, and there is not an excellent explanation for this. Some sugars come in chiral groups (d-mannose, for example, is indigestible and used to treat urinary tract infections). The structural isomers in my carrot top pesto are terpenoids, and as we’ll see below, there are chiral terpenoids, too.

In his book A Fragrant Introduction to Terpenoid Chemistry (Sell 2003), Charles Sell opens his discussion of chirality by quoting Lewis Carroll’s Alice, as she contemplates a glass of milk in the world she discovered on the other side of the looking glass:

“Perhaps looking glass milk isn’t good to drink” (Carroll 1871).

Alice may have been as much a biochemist as a reluctant hero. Whether or not some things are good to eat or drink really does depend on which side of mirror it’s on, and our noses are excellent stereochemists (Bentley 2006).

Alice entering the Looking Glass, by Sir John Tenniel

A structural isomer (also called an enantiomer) may or may not behave chemically identically to its chiral counterpart. Further, enzymes and mammalian physiological receptors (smell, taste, temperature, pain) may respond differently to different enantiomers. The first job of both an enzyme and an olfactory receptor is to bind with a chemical. To continue the hand analogy, imagine that the enzyme or a particular region in an olfactory receptor includes a binding site shaped like a left-hand glove. Only left-hand-shaped enantiomers will fit well in it. Mammalian receptors can bind multiple enantiomers of at least some chiral molecules and may send a distinct neurological message for each isomer. If a mammalian receptor responds to enantiomers of a chiral molecule differently, we perceive (smell, taste, feel) those enantiomers differently. In the case of olfactory (smell) receptors, each enantiomer may have its own aroma, and therefore flavor.

I think the terpenoid carvone is the most dramatic example of distinct enantiomer aromas. One enantiomer, referred to as R-()-carvone, smells like spearmint (Mentha spicata; family Lamiaceae). Its counterpart, S-(+)-carvone, smells like caraway (Carum carvi; Apiaceae). Each respective carvone enantiomer comprises the bulk of the essential oils of spearmint and caraway. I don’t think anyone would describe the aromas of those popular herbs as even remotely similar. Spearmint, for example, makes an excellent jam with peaches, but I don’t think it would taste good on caraway-laced rye toast.

S-(+)-carvone (caraway flavor) is also a major constituent of the flavor of dill, a close relative of caraway. R-()-carvone (spearmint flavor) is a component of the flavor of some other mint species, but it is not a contributor to the flavor of peppermint, even though peppermint is a hybrid between spearmint and mountain mint. The star player in peppermint oil is the terpene menthol. There are eight stereoisomers of menthol, but only one has the well-known, fresh peppermint odor (Mosandl 1988). Another menthol isomer smells kind of musty or moldy, which is how some wild mints smell to me.

Isomer pairings from the mint and carrot families…and beyond

ajwain seeds and fresh oregano

The carvone isomers exemplify what to me is a fascinating terpenoid-derived flavor similarity or complementarity between foods from the mint (Lamiaceae; Lamiales) and umbellifer or carrot (Apiaceae; Apiales) plant families, one which I exploited in my carrot top pesto. I used structural isomers carvacrol from oregano (mint family) and thymol from ajwain (carrot family; explained below) and thyme (mint family). The terpenoid essential oils contributing to the flavor of the carrot tops themselves are also significantly present in some mint-family herbs, including myrcene, present in thyme; limonene; ocimene, methylisoeugenol and methyleugenol (clove-like) in basil, and sabinene, the double-bond isomer of which is thujone, a major flavor constituent of the mint-family herb savory (Kainulainen et al. 2002).

The mint family and carrot family aren’t particularly phylogenetically close (see our crash course on plant evolution). Both families are in the large flowering plant clade (group of species derived from the same common ancestor) called the asterids, but the asterids contains two major sub-clades, the lamiids and the campanulids. The mint family is in the former, and the carrot family is in the latter (Chase et al. 2016). Any terpenoid flavor chemistry similarities between the two families, then, can’t be persuasively attributed to particularly recent common ancestry. We have already demonstrated, though, that terpenoids are very common across all plants, and their evolution appears to be quite labile, leading to numerous instances of terpenoid convergent evolution across widely disparate parts of the plant evolutionary tree, such as the multiple independent occurrences of lemon-scented plants, including in the Apiaceae and Lamiaceae, as I’ll explain below.

As it happens, one of the chemical precursors in the biosynthesis of carvacrol and thymol is p-cymene, which is the dominant flavor compound in cumin (Apiaceae) and a minor flavor contributor to thyme. Another precursor is gamma-terpenine, which has a strong lemon scent and is most abundant in lemon-scented citrus fruits (Russo et al. 1998) but contributes to flavor in both families as well. The extension of terpenoid similarity and affinity to citrus foods doesn’t stop there.  The major citrus terpenoid limonene is part of the biosynthesis pathway for many other derivative terpenoids, including carvone (Gershenzon et al. 1989, Bouwmeester et al. 1998).


Lemon thyme on the left, French thyme on the right

Limonene is the primary compound in orange essential oil and is often found in great abundance in plants with “lemony” aromas. Limonene itself is chiral, with two enantiomers: R-(+)-limonene, the distinctive orange peel smell, and S-(-)-limonene, which smells more medicinal and is a major aroma note in eucalyptus. The chiral “handedness” of a limonene enantiomer is passed down to the carvone derived from it. Some of the limonene in some citruses and lemon-scented plants gets oxidized into carvone. S-(+)-carvone (the caraway-smelling one) is a component of the essential oil of mandarin orange (Citrus reticulata; Rutaceae, Sapindales). The essential oil of at least one of the many charismatic lemongrass (genus Cymbopogon; Poaceae, Poales) species, gingergrass or palmarosa (Cymbopogon martinii), contains both carvone enantiomers. “Lemony” varieties of mint-family herbs including thyme, mint, and basil must contain decent concentrations of the citrus-scented terpenoids.


parsley and carrot leaves show typical Apiaceae laciness

It’s no wonder that savory recipes combining mint-family and carrot-family ingredients also often include citrus zest and/or juice. Think about tabbouleh, the bulgar wheat salad containing copious amounts of spearmint, parsley, and lemon. One of our favorite zucchini preparations is to saute it with basil, parsley, and lemon zest. And, of course, lemon is often added to a pesto of fresh herbs or greens.

Carrot Top Pesto

That first batch of carrot top pesto was one of those spontaneous pantry-raiding endeavors involving scrounging around the kitchen for suitable ingredients. I wanted a strong herbal flavor to complement the mild carrot tops. Thyme is a classic seasoning for carrots, and we had it and its Lamiaceae compatriot, oregano, growing in the garden. It was with those two mint family herbs in mind that my gaze landed on a small jar of ajwain “seeds” (technically dry, hard fruits called schizocarps inextricably bound to the seed contained within).

thumb_2015-09-05 15.59.00_1024

Adding carrot tops to the onions

Carrots and ajwain (Trachyspermum ammi) are both umbellifers, members of the Apiaceae. Ajwain tastes strongly, hotly, of thyme or oregano, depending on who is describing the flavor. It is native to India, the Middle East, and North Africa, and it is common in the cuisines of that region. It is much less well known outside of that region, despite having a pleasing shape, similar to the “seeds” (fruits) of other familiar umbellifer spices (coriander, cumin, caraway, fennel, celery seed, aniseseed, dill) and tasting almost exactly like some of the most popular Mediterranean mint family herbs (thyme, oregano, savory). As soon as I spotted the jar, languishing on the shelf, the nerdy isomeric, dual-plant-family recipe assembled itself in my head: carrot tops, ajwain, and parsley from the Apiaceae; oregano and thyme from the Lamiaceae; all thrown in at the last minute on top of cooked onion and then thrown into the food processor with salt and pepper, olive oil, lemon, and toasted pumpkin seeds.


Green tops from 6-12 carrots, tough midribs removed

1 bunch parsley, tough midribs removed

1 t each minced fresh oregano and thyme, or 1/2 t dried

1 t ajwain seeds

1 onion, diced

1 c toasted pumpkin seeds

salt and pepper to taste

zest and juice of ½ lemon

¼ c olive oil

water if necessary

Sautée the onion over medium heat in oil until translucent and soft, about 10 minutes in a vessel that can also accommodate the herbs and carrot tops. When the onion is cooked, add the carrot tops, parsley, and herbs, and stir quickly just until the greens wilt. Put all ingredients in a food processor and puree until smooth, adding water if necessary to thin, adjusting seasoning as necessary.


Bentley, R. 2006. The nose as a stereochemist. Enantiomers and odor. Chemical Reviews 106:4099–4112.

Bouwmeester, H. J., J. Gershenzon, M. C. J. M. Konings, and R. Croteau. 1998. Biosynthesis of the monoterpenes limonene and carvone in the fruit of caraway – I. Demonstration of enzyme activities and their changes with development. Plant Physiology 117:901–912.

Carroll, L. 1871. Alice through the looking glass.

Chase, M. W., M. J. M. Christenhusz, M. F. Fay, J. W. Byng, W. S. Judd, D. E. Soltis, D. J. Mabberley, A. N. Sennikov, P. S. Soltis, P. F. Stevens, B. Briggs, S. Brockington, A. Chautems, J. C. Clark, J. Conran, E. Haston, M. Möller, M. Moore, R. Olmstead, M. Perret, L. Skog, J. Smith, D. Tank, M. Vorontsova, and A. Weber. 2016. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG IV. Botanical Journal of the Linnean Society 181:1–20.

Gershenzon, J., M. Maffei, and R. Croteau. 1989. Biochemical and Histochemical Localization of Monoterpene Biosynthesis in the Glandular Trichomes of Spearmint (Mentha spicata). Plant physiology 89:1351–1357.

Kainulainen, P., A. Nissinen, A. Piirainen, K. Tiilikkala, and J. K. Holopainen. 2002. Essential oil composition in leaves of carrot varieties and preference of specialist and generalist sucking insect herbivores. Agricultural and Forest Entomology 4:211–216.

Mosandl, A. 1988. Chirality in flavor chemistry-recent developments in synthesis and analysis. Food Reviews International 4:1–43.

Russo, M., G. C. Galletti, P. Bocchini, and A. Carnacini. 1998. Essential Oil Chemical Composition of Wild Populations of Italian Oregano Spice ( Origanum v ulgare ssp. h irtum (Link) Ietswaart): A Preliminary Evaluation of Their Use in Chemotaxonomy by Cluster Analysis. 1. Inflorescences. Journal of Agricultural and Food Chemistry 46:3741–3746.

Sell, C. 2003. A Fragrant Introduction to Terpenoid Chemistry. Royal Society of Chemistry.

Xu, J., F. Zhou, B.-P. Ji, R.-S. Pei, and N. Xu. 2008. The antibacterial mechanism of carvacrol and thymol against Escherichia coli. Letters in Applied Microbiology 47:174–179.

10 thoughts on “Carrot top pesto through the looking glass

  1. Mike DeNoyer

    Jeanne, Great article.  Love your style. Just one question. How did you use the pesto?  On bread?  with meat? Oh. And good on Erik for suggesting the idea. Love, Dad.


  2. Andi Clevely

    Fascinating piece, especially the key reasons why flavours combine or repel. Until now we have used carrot tops with other veg rejects as ingredients in Meg’s lovely (and unrepeatable – never the same ingredients to hand!) and unique soups, but pesto a promising alternative.
    Would love to see you at work in your kitchen – test-tubes next to mouli? 🙂
    Best wishes, keep up the absorbing reports.
    Andi & Meg Clevely


  3. Pingback: #Celery | The Botanist in the Kitchen

  4. Pingback: Angelica: Holiday fruitcake from a sometimes toxic family | The Botanist in the Kitchen

  5. Cynthia

    Wow. Those carrot tops didn’t know what was coming. I work in a biochemistry lab. I just make carrot top mint pesto yesterday. It was so much better than I thought it would be. I used peanuts because I didn’t have pine nuts on hand. Its was really good! I will definitly try your recipe next time. I’m so glad your husband made you stand up as a vegetable epicurean and do something delicious! The science was a welcome part of the story! I’ve also never seen a recipe that has so many citations. Do you use a citation manager luke EndNote or Refworks for all your recipes? I’d love to see more!


    1. Jeanne L. D. Osnas Post author

      Your pesto sounds interesting, like a Thai mint-peanut version. Thanks for sharing, and I hope I didn’t butcher the chemistry here too badly! I’ve been using Mendeley for the past few years for citations. I imported all my old Endnote libraries into it. I’m a fan.


  6. Pingback: Sage, rosemary, and chia: three gifts from the wisest genus (Salvia) | The Botanist in the Kitchen

  7. Pingback: The Botanist in the Root Cellar | The Botanist in the Kitchen

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.