Variations in Terpene Profiles of Different Strains of Cannabis sativa L.
S. Casano and G. Grassi
Consiglio per la Ricerca e la Sperimentazione in
Centro di Ricerca per le Colture Industriali
V. Martini and M. Michelozzi
Consiglio Nazionale delle
Istituto di Genetica Vegetale
Sesto Fiorentino (Firenze)
Keywords: aroma volatiles, cannabinoids, chemosystematic studies, medical effects,
Secondary compounds of the plant are indispensable to cope with its often
hostile environment and the great chemical diversity and variability of intraspecific
and interspecific secondary metabolism is the result of natural selection. Recognition
of the biological properties of secondary compounds have increased their great utility
for human uses; numerous compounds now are receiving particular attention from
the pharmaceutical industry and are important sources of a wide variety of
commercially useful base products. Medical and other effects of Cannabis sativa L. are
due to concentration and balance of various active secondary metabolites, particularly
the cannabinoids, but including also a wide range of terpenoids and flavonoids. A wide
qualitative and quantitative variability in cannabinoids, terpenoids, and flavonoids
contents in Cannabis species are apparent from reports in the literature. Terpenes are
strongly inherited and little influenced by environmental factors and, therefore, have
been widely used as biochemical marker in chemosystematic studies to characterize
plant species, provenances, clones, and hybrids. This study investigated the variability
of terpene profiles in C. sativa. The terpene composition in inflorescences of samples
collected from progenies of 16 plants derived from different strains was analysed by
GC/FID. The amount of each terpene (in sufficient quantities to be considered in
statistical analysis) was expressed as a percentage of total terpenes. Results showed a
large variation between different strains in the relative contents for several mono-
terpenes (α-pinene, camphene, β-pinene, sabinene, Δ-3-carene, α-phellandrene,
β-myrcene, α-terpinene, limonene, 1.8-cineole, γ-terpinene, cis-β-ocimene, trans-β-
ocimene, α-terpinolene) and one sesquiterpene, β-caryophyllene. This variability in
terpene composition can provide a potential tool for the characterization of Cannabis
biotypes and warrant further research to evaluate the drug’s medical value and, at the
same time, to select less susceptible chemotypes to the attack of herbivores and
The psychotropic effects of Cannabis, primarily due to the main psychotropic
cannabinoid, Δ9-THC (delta9-tetrahydrocannabinol), have been intensely studied as pure
compounds for medicinal activity. The pharmaceutical industry, however, is interested in
the plant as a source of raw material and studying the variability and synergy among the
various secondary metabolites. Other cannabinoids, terpenoids, and flavonoids may
reduce Δ9-THC-induced anxiety, cholinergic deficit, and immunosuppression, while at
the same time increase cerebral blood flow, enhance cortical activity, kill respiratory
pathogens, and provide anti-inflammatory activity (McPartland and Russo, 2001).
Terpenoids possess a broad range of biological properties, including cancer chemo-
preventive effects, skin penetration enhancement, antimicrobial, antifungal, antiviral, anti-
hyperglycemic, anti-inflammatory, and antiparasitic activities (Paduch et al., 2007).
Plants exhibit dynamic biochemical changes when attacked by diseases and
herbivores and in response to abiotic stresses, resulting in the induced production and
Proc. XXVIIIth IHC
IHC Seminar: A New Loo
at Medicinal and Aromatic Plants
Eds.: Á. Máthé et al.
Acta Hort. 925, ISHS 2011
release of aroma volatiles that are beneficial for direct or indirect defense. In Arabidopsis
thaliana (Huang et al., 2010) and in Medicago truncatula (Navia-Ginè et al., 2009) a
significant quantitative variation in the emission of the monoterpene trans-β-ocimene
occurs as a consequence of the attack by herbivorous insects. Two monoterpenes
generally present in the aroma volatiles of Cannabis, limonene and α-pinene, as well as
other monoterpenes, have been shown to powerfully repel herbivorous insects (Nerio et
al., 2010), while sesquiterpenes tend to be related to intake by grazing animals. Potter
(2009) demonstrated that in Cannabis the monoterpene:sesquiterpene ratios in leaves and
inflorescences are very different because of the dominant presence of sessile trichomes on
foliage and of capitate stalked trichomes on floral material, with the most volatile
monoterpenes dominating in inflorescences to repel insects and the most bitter
sesquiterpenes dominating in leaves to act as antiherbivory for grazing animals. Being
that pharmaceutical Cannabis is normally cultivated in facilities not accessible to grazing
animals, the major pest problem remains herbivorous insects, especially the most
common and destructive spider mites, thrips, and whiteflies, thus the analysis of
monoterpenes and the study of their variability may play a strategic role into select plants
less susceptible to the attack of these and other insects.
Terpenes are strongly inherited and little influenced by environmental factors and,
therefore, have been widely used as biochemical marker in chemosystematic studies to
characterize plant species, provenances, clones and hybrids. A wide variability in
terpenoids content in different strains of Cannabis have been reported (Mediavilla and
Steinemann, 1997; Novak et al., 2001; Hillig, 2004; Fischedick et al., 2010). The
variability on secondary metabolism combined with genetic data has recently re-opened
the old debate on its taxonomic treatment. In fact, Hillig (2005) proposed a polytypic
concept which recognizes three species (Cannabis sativa, Cannabis indica and Cannabis
ruderalis) and seven putative taxa, but at present the majority of researchers continue to
agree on the monotypic treatment and identify the species as Cannabis sativa L. The
differentiation of strains in ‘pure sativa’, ‘mostly sativa’, ‘sativa/indica hybrid’, ‘mostly
indica’, ‘pure indica’ and ‘ruderalis hybrid’ is generally adopted by breeders and growers
to distinguish the different biotypes. The current study investigated the variability in
terpene profiles of Cannabis strains and explored the utility of monoterpenes in the
distinction between ‘mostly sativa’ and ‘mostly indica’ biotypes.
MATERIALS AND METHODS
Several strains with Δ9-THC profile were obtained from breeders of private
companies. Assignment of strains exclusively to ‘mostly sativa’ or ‘mostly indica’
biotypes was based on the genetic background declared by breeders of the strains.
Assignment to ‘pure indica’ and ‘pure sativa’ biotypes was not used because of the
uncertain information on these strains. Each strain consisted of a commercial pocket,
generally of ten viable seeds. Preliminary evaluations on the declared genetic background
were performed by growing these strains during the spring-summer term in a greenhouse
at CRA-CIN (Rovigo). At the beginning of the flowering stage staminate plants were
eliminated while pistillate plants were treated with silver thiosulfate solution to artificially
induce the production of staminate inflorescences. Self-pollination of all the pistillate
plants was performed by physically isolating plants from each other by individual white
paper bags. Only 16 pistillate plants derived from 16 different strains were finally
selected. 8 plants (ID: 5, 6, 7, 8, 9, 10, 11 and 12) were derived from ‘mostly sativa’
strains and the other 8 plants (ID: 2, 3, 4, 13, 14, 15, 16 and 17) were derived from
‘mostly indica’ strains.
Progenies of the 16 plants were grown in indoor conditions at CRA-CIN (Rovigo).
In total 99 plants (3 to 7 plants for each strain) were grown under 600 W/m2 high pressure
sodium lamps (Philips Son-T). Photoperiod was kept at 18 hours of light for the first 4
weeks of cultivation and then decreased to 12 hours of light until the harvest. Temperature
and relative humidity of the air were respectively maintained at 25±3°C and 50-70%.
Plants were individually grown in 1.5-L pots in finely ground flakes of coconut fibre
(CANNA B.V.) and they were daily ferti-irrigated, by using an automatic irrigation
system, with a dose of nutrient solution depending on requirement. The nutrient solution
used (EC=1.7) was obtained by mixing equal parts of Coco A and B (CANNA B.V.) with
tap water, and then the pH level was adjusted to 5.5. Ferti-irrigation was interrupted 2
weeks before the harvest and pots were flushed with tap water adjusted to pH=5.5.
The harvest of early strains (‘mostly indica’) occurred after 105 days from sowing
while the harvest of late strains (‘mostly sativa’) was deferred at 133 days. Fresh
inflorescence tissues of plants were sampled during the harvest for analyses of terpenoids.
The sample material (80 mg of fresh inflorescence tissues) was ground in liquid nitrogen,
extracted in 4 ml of n-pentane and then 1 ml of the extract was transferred to GC vials.
The terpene composition was analyzed by GC/FID. In total, 28 compounds were detected,
15 were fully identified while 13 remained unknown (unk). Terpenoids were identified by
matching their retention times with those of pure compounds under the same conditions.
Depending from their retention times, peaks were identified as following:
α-pinene, unk1, unk2, camphene, β-pinene, sabinene, Δ-3-carene, α-phellandrene,
β-myrcene, α-terpinene, limonene, 1.8 cineole, γ-terpinene, cis-β-ocimene, trans-β-
ocimene, α-terpinolene, unk3, unk4, β-caryophyllene, unk5, unk6, unk7, unk8, unk9,
unk10, unk11, unk12 and unk13. Terpenoids identified were mostly monoterpenes with
the exception of one sesquiterpene, β-caryophyllene.
Relative content of each monoterpene was expressed as a percentage of total
monoterpenes, while each sesquiterpene was calculated as a percentage of total mono-
terpenes plus sesquiterpenes. Data were not normally distributed (Kolmogorov-Smirnov
one sample test) and were analysed by the non-parametric Kruskal-Wallis ANOVA
followed by the Mann-Whitney U Test for multiple comparisons. Differences were
accepted when significant at the 5% level. Statistical analyses were performed by using
SYSTAT 12.0 software (Systat Software Inc., USA).
RESULTS AND DISCUSSION
The relative content of terpenoids is strongly inherited while total yield per weight
of tissue is more subjected to environmental factors. Expression of composition on a
+ tissue basis (mg/g) is used for quality control and standardization of Cannabis cultivars,
as well as for chemosystematic studies (Fischedick et al., 2010), but the relative content
(%) of terpenoids is more often used for chemosystematic studies.
The average relative contents of dominant compounds detected in the aroma
volatiles of all the strains were: β-myrcene (46.1±2.6%), α-pinene (14.0±1.5%),
α-terpinolene (10.2±1.8%), limonene (7.3±1.3%), trans-β-ocimene (6.6±0.7%), β-pinene
(6.1±0.4%), α-terpinene (3.6±1.0%), β-caryophyllene (1.2±0.2%), 1.8 cineole
(1.1±0.2%), α-phellandrene (0.7±0.1%) and Δ-3-carene (0.6±0.1%). The average relative
contents of camphene, unk1, cis-β-ocimene, unk5, unk8, unk7, unk13, sabinene,
γ-terpinene, unk3, unk4, unk6, unk10, unk2, unk9, unk11 and unk12 were lower than
0.5%. Results of Kruskal-Wallis ANOVA between different strains (d.f.=15, N=99)
showed significant changes in relative contents of all the compounds: α-pinene (X2=71.6,
P<0.001), unk1 (X2=71.5, P<0.001), unk2 (X2=43.6, P<0.001), camphene (X2=67.2,
P<0.001), β-pinene (X2=53.2, P<0.001), sabinene (X2=72.5, P<0.001), Δ-3-carene
(X2=69.4, P<0.001), α-phellandrene (X2=59.6, P<0.001), β-myrcene (X2=47.7, P<0.001),
α-terpinene (X2=36.3, P<0.01), limonene (X2=77.1, P<0.001), 1.8 cineole (X2=67.5,
P<0.001), γ-terpinene (X2=30.9, P<0.01), cis-β-ocimene (X2=79.5, P<0.001), trans-β-
ocimene (X2=82.1, P<0.001), α-terpinolene (X2=78.7, P<0.001), unk3 (X2=37.6,
P<0.001), unk4 (X2=33.7, P<0.01), β-caryophyllene (X2=55.7, P<0.001), unk5 (X2=65.6,
P<0.001), unk6 (X2=74.4, P<0.001), unk7 (X2=50.1, P<0.001), unk8 (X2=64.7, P<0.001),
unk9 (X2=63.2, P<0.001), unk10 (X2=61.1, P<0.001), unk11 (X2=80.1, P<0.001), unk12
(X2=61.8, P<0.001) and unk13 (X2=52.8, P<0.001).
β-myrcene was detected in high % in all the strains, with strain 17 having the
highest relative content (80.1±7.3%) and strain 8 having the lowest relative content
(16.1±3.4%) (Table 1). β-myrcene was the dominant terpene in almost all the strains, with
the exceptions of strains 6, 7, 8 and 12. α-terpinolene was detected in high % in some
‘mostly sativa’ strains (7, 8, 9, 10 and 12), with strains 7 and 8 having α-terpinolene as the
dominant terpene (respectively 41.8±7.2% and 37.3±3.5%), while it was not detected or it
was detected in traces in ‘mostly indica’ strains and in some ‘mostly sativa’ strains (5, 6
and 11). α-pinene and β-pinene were detected in all the strains and their relative contents
were commonly lower than 10%. α-pinene was detected in higher relative contents (up to
10%) in some strains (3, 6, 8, 11, 12, 14, 15 and 16), with strains 6 and 12 having α-
pinene as the dominant terpene (respectively 46.3±5.7% and 24.2±15.6%).
β-pinene was detected in higher relative contents (up to 10%) in strains 3 (12.6±1.6%)
and 6 (13.2±0.8%). Limonene was detected in low % or traces in some ‘mostly indica’
strains (3, 14, 15 and 16) and in ‘mostly sativa’ strains, while it was detected in much
higher % (up to 10%) in some ‘mostly indica’ strains (2, 4, 13 and 17), with these strains
having limonene as second most abundant terpenoid. Trans-β-ocimene was not detected
or it was detected in low % in one ‘mostly sativa’ strains (6) and in ‘mostly indica’ strains,
while in some ‘mostly sativa’ strains (5, 7, 8, 9, 10, 11 and 12) it was detected in much
higher % (up to 5%), with strains 5 and 11 having trans-β-ocimene as second most
abundant terpenoid (respectively 18.7±1.9% and 16.8±2.2%). α-terpinene was detected in
low % or traces in almost all the strains, with strains 4 having a much higher relative
content (18.0±8.0%). The sesquiterpene β-caryophyllene was detected in all the strains
and its relative content was commonly lower than 2%, with some strains (2, 9, 13 and 17)
having relative contents up to 2%. 1.8 cineole was detected in low % (up to 2%) in some
‘mostly sativa’ strains (7, 8, 9, 10 and 12), while it was detected in lower % or traces in
‘mostly indica’ strains and in some ‘mostly sativa’ strains (5, 6 and 11). Δ-3-carene and α-
phellandrene were detected in low % (up to 1%) in some ‘mostly sativa’ strains (7, 8, 9,
10 and 12), while they were not detected in ‘mostly indica’ strains and in some ‘mostly
sativa’ strains (5, 6 and 11).
Mann-Whitney U test between ‘mostly sativa’ strains and ‘mostly indica’ strains
(d.f.=1, N=99) showed significant changes in relative contents of several compounds
except for α-pinene, unk2, β-pinene, α-terpinene, γ-terpinene, β-caryophyllene, unk7,
unk12 and unk13 (Fig. 1). Relative contents of camphene (X2=22.7, P<0.001), β-myrcene
(X2=23.1, P<0.001), limonene (X2=27.8, P<0.001), unk3 (X2=15.4, P<0.001), unk6
(X2=29.9, P<0.001) and unk11 (X2=42.3, P<0.001) were significantly higher in ‘mostly
indica’ strains than in ‘mostly sativa’ strains (Fig. 1). Plants derived from ‘mostly sativa’
strains showed significantly higher relative proportions of unk1 (X2=33.4, P<0.001),
sabinene (X2=24.9, P<0.001), Δ-3-carene (X2=39.6, P<0.001), α-phellandrene (X2=31.97,
P<0.001), 1.8 cineole (X2=19.2, P<0.001), cis-β-ocimene (X2=48.6, P<0.001), trans-β-
ocimene (X2=52.6, P<0.001), α-terpinolene (X2=13.2, P<0.001), unk4 (X2=15.3,
P<0.001), unk5 (X2=29.6, P<0.001), unk8 (X2=24.3, P<0.001), unk9 (X2=7.5, P<0.01)
and unk10 (X2=9.5, P<0.01) than plants derived from ‘mostly indica’ strains (Fig. 1).
Although Hillig (2004) stated that differences on terpenoids in Cannabis are of
limited use for taxonomic discrimination at the species level, with sesquiterpenes
generally more useful than monoterpenes, we found that several monoterpenes markers
can be powerful tools for discerning between ‘mostly sativa’ and ‘mostly indica’ biotypes
(Table 1 and Fig. 1). Our results are also supported by results recently obtained by
Fischedick et al. (2010) showing that monoterpenes are able to distinguish cultivars with
similar sesquiterpenes and cannabinoids levels.
The main differences between terpene profiles of the evaluated strains belonging
to the two principal biotypes were that ‘mostly indica’ strains were characterized by
dominancy of β-myrcene, present in high relative contents, with limonene or α-pinene as
second most abundant terpenoid, while ‘mostly sativa’ strains were characterized by more
complex terpene profiles, with some strains having α-terpinolene or α-pinene as dominant
terpenoid, and some strains having β-myrcene as dominant terpenoid with α-terpinolene
or trans-β-ocimene as second most abundant terpenoid.
This wide variability in terpene composition can provide a potential tool for the
characterization of Cannabis biotypes, and warrant further researches in order to evaluate
the drug’s medical value and, at the same time, to select less susceptible chemotypes to
the attack of herbivores and diseases. More detailed studies on the variability in
monoterpenes and sesquiterpenes are needed. Breeding for specific terpenoids in plants is
a fascinating research topic; in fact, the various biological activities of these compounds
make the analysis of terpenoids a valuable tool for improving a considerable number of
traits in pharmaceutical and industrial cultivars of Cannabis.
Terpenoids analysis, combined with cannabinoids and flavonoids analyses, are
essential for the metabolic fingerprinting of pharmaceutical cultivars. Pharmaceutical
cultivars of the two principal biotypes may exhibit distinctive medicinal properties due to
significant differences in relative contents of terpenoids, thus the synergy between the
various secondary metabolites must be investigated in deeper details in the future in order
to better elucidate the phytocomplex of Cannabis and to allow selection of chemotypes
with specific medical effects.
Thanks to Phytoplant Research S.L. for financial assistance.
Fischedick, J.T., Hazekamp, A., Erkelens, T., Choi, Y.H. and Verpoorte, R. 2010.
Metabolic fingerprinting of Cannabis sativa L., cannabinoids and terpenoids for
chemotaxonomic and drug standardization purposes. Phytochemistry 71:2058-
Hillig, K.W. 2004. A chemotaxonomic analysis of terpenoid variation in Cannabis.
Biochemical Systematics and Ecology 32:875-891.
Hillig, K.W. 2005. Genetic evidence for speciation in Cannabis. Genetic Resources and
Crop Evolution 52:161-180.
Huang, M., Abel, C., Sohrabi, R., Petri, J., Haupt, I., Cosimano, J., Gershenzon, J. and
Tholl, D. 2010. Variation of herbivore-induced volatile terpenes among Arabidopsis
ecotypes depends on allelic differences and subcellular targeting of two terpene
synthases, TPS02 and TPS031. Plant Physiology 153:1293-1310.
McPartland, J.M. and Russo, E.B. 2001. Cannabis and Cannabis extracts: greater than the
sum of their parts? Journal of Cannabis Therapeutics 1:103-132.
Mediavilla, V. and Steinemann, S. 1997. Essential oil of Cannabis sativa L. strains.
Journal of the International Hemp Association 4:80-82.
Navia-Ginéa, W.G., Yuanb, J.S., Mauromoustakosd, A., Murphye, J.B., Chenb, F. and
Kortha, K.L. 2009. Medicago truncatula (E)-β-ocimene synthase is induced by
insect herbivory with corresponding increases in emission of volatile ocimene.
Plant Physiology and Biochemistry 47:416-425.
Nerio, L.S., Olivero-Verbel, J. and Stashenko, E. 2010. Repellent activity of essential oils:
a review. Bioresource Technology 101:372-378.
Novak, J., Zitterl-Eglseer, K., Deans, S.G. and Franz, C.M. 2001. Essential oils of
different cultivars of Cannabis sativa L. and their antimicrobial activity. Flavour
and Fragrance Journal 16:259-262.
Paduch, R., Kandefer-Szerszeń, M., Trytek, M. and Fiedurek, J. 2007. Terpenes:
substances useful in human healthcare. Archivum Immunologiae et Therapiae
Potter, D. 2009. The propagation, characterisation and optimisation of Cannabis sativa L.
as a phytopharmaceutical. Ph.D. Thesis in Pharmaceutical Sciences. Department of
Pharmaceutical Science Research, King’s College, London.
Tabl e s
Table 1. Terpene profiles of different ‘mostly indica’ and ‘mostly sativa’ strains of Cannabis sativa L.
ND = not detected.
Fig. 1. Comparison of terpene profiles in ‘mostly indica’ (black histograms) and in
‘mostly sativa’ (white histograms) strains of Cannabis sativa L. Break on Y-axis is
0.7-0.8. Numbers on X-axis refer to individual compounds: 1=α-pinene, 2=unk1,
3=unk2, 4=camphene, 5=β-pinene, 6=sabinene, 7=Δ-3-carene, 8=α-phellandrene,
9=β-myrcene, 10=α-terpinene, 11=limonene, 12=1.8 cineole, 13=γ-terpinene,
14=cis-β-ocimene, 15=trans-β-ocimene, 16=α-terpinolene, 17=unk3, 18=unk4,
19=β-caryophyllene, 20=unk5, 21=unk6, 22=unk7, 23=unk8, 24=unk9, 25=unk10,
26=unk11, 27=unk12 and 28=unk13.