Journal:A high-throughput method for the comprehensive analysis of terpenes and terpenoids in medicinal cannabis biomass

Full article title	A high-throughput method for the comprehensive analysis of terpenes and terpenoids in medicinal cannabis biomass
Journal	Metabolites
Author(s)	Krill, Christian; Rochfort, Simone; Spangenberg, German
Author affiliation(s)	AgriBio, the Centre For AgriBioscience
Primary contact	Email: christian dot krill at agriculture dot vic dot gov dot au
Year published	2020
Volume and issue	10(7)
Article #	276
DOI	10.3390/metabo10070276
ISSN	2218-1989
Distribution license	Creative Commons Attribution 4.0 International
Website	https://www.mdpi.com/2218-1989/10/7/276/htm
Download	https://www.mdpi.com/2218-1989/10/7/276/pdf (PDF)

Abstract

Cannabis and its secondary metabolite content have recently seen a surge in research interest. Cannabis terpenes and terpenoids in particular are increasingly the focus of research efforts due to the possibility of their contribution to the overall therapeutic effect of medicinal cannabis. Current methodology to quantify terpenes in cannabis biomass mostly relies on large quantities of biomass, long extraction protocols, and long gas chromatography (GC) gradient times, often exceeding 60 minutes. They are therefore not easily applicable in the high-throughput environment of a cannabis breeding program. The method presented here, however, is based on a simple hexane extract from 40 mg of biomass, with 50 μg/mL dodecane as internal standard, and a gradient of less than 30 minutes. The method can detect 48 individual terpenes and terpenoids and was validated for selectivity, linearity, limit of detection/limit of quantitation (LOD/LOQ), precision, intermediate precision, and accuracy (recovery) for 22 terpenes and terpenoids. The validation parameters are comparable to previously published studies that employ significantly longer runtimes and/or more complex extraction protocols. It is currently being applied to medicinal cannabis precision breeding programs.

Keywords: cannabis volatiles, terpenes, terpenoids, profiling, high-throughput method

Introduction

Cannabis is a genus of annual dioecious plants within the family Cannabaceae with a rich and complex constituency of pharmacologically relevant phytochemicals. Under the Single Convention on Narcotic Drugs (1961), cannabis was deemed to be a plant without medicinal purpose, a conclusion based on very little scientific evidence or clinical trial data. As a result, medicinal use of cannabis was practically prohibited by its status as an illegal narcotic worldwide, and research into cannabis chemistry and biology was largely limited to law enforcement and forensics applications. Changes in the attitudes towards cannabis across various societies and increasing evidence for the benefits of cannabinoids in the treatment of otherwise intractable conditions have precipitated recent changes to legislature in a number of jurisdictions, including Australia. Interest in the medicinal uses of cannabis is increasing, as indicated by a growing number of review articles on the topic.

Cannabis, cannabis extracts, and purified (to varying degrees) major cannabinoids such as Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD) are now being used for the treatment of intractable and debilitating illnesses, such as Dravet’s Syndrome, and investigated for a range of others. When administered as medicinal cannabis extracts, that is, along with the remaining portion of other cannabinoids and non-cannabinoid phytochemicals, less than a quarter of the equivalent dose of purified CBD is required, and markedly lower side effects are commonly observed. This is in-line with a long line of anecdotal evidence that complex plant-based medicines are more effective than their isolated “actives,” which has given rise to the concept of the “entourage effect,” the modulating or synergistic effects of phytochemical compounds such as the minor cannabinoids on those effects related to THC. Despite a current lack of understanding for the underlying mechanisms, the concept is increasingly applied to non-cannabinoid cannabis phytochemicals, such as terpenes and terpenoids.

Terpenes are a class of hydrocarbons synthesized from the five-carbon (C5) “building blocks” dimethylallyl pyrophosphate and isopentenyl pyrophosphate. Depending on the number of C5 units and possible substitutions, they are further classified based on number of units (e.g., C10 monoterpenes [two sub-units], C15 sesquiterpenes [three sub-units]) or functional groups (terpenoids and oxygenated). Mono- and sesquiterpenes are classified as volatile and semi-volatile compounds, respectively, and higher order terpenes (e.g., C20 diterpenes and C30 triterpenes) exist as steroids, waxes, and resins. Cannabis mono- and sesquiterpenes are responsible for the characteristic smell of the plant and its products. They are also present at pharmacologically relevant concentrations, and a growing body of evidence suggests they have pharmacological properties in their own rights. Their common use as food additives and in cosmetics products highlights the inherent safety of these compounds.

As Leghissa and colleagues note, conclusively identifying terpenes in cannabis is challenging due to the large variety of possible candidates and a lack of commercial standards for a large number of them. With the increased interest in medicinal cannabis and the contribution of terpenes to the entourage effect, methods to quantify terpenes in medicinal cannabis biomass and products are becoming more available in the literature. Often, these methods require large quantities of biomass (1–5 g), use large quantities of organic solvents (up to 100 mL/sample), and include separations with gradient runtimes exceeding 60 minutes. These methods are therefore not easily applied in a high-throughput manner. To facilitate terpene quantitation and profiling in order to enable accelerated precision breeding programs in medicinal cannabis, we developed a fast (30-minute separation), microscale (40 mg sample), high-throughput gas chromatography–mass spectrometry (GC-MS) terpene profiling and quantitation method.

Results

Sampling techniques

Three common sampling and sample introduction techniques were evaluated: static headspace, solid-phase microextraction (SPME), and liquid injection of an organic solvent extract. Representative chromatograms for the three tested sampling techniques are shown in Figure 1. All three techniques provide excellent signal strengths for the lower boiling monoterpenes. Sesquiterpenes are underrepresented in the static headspace chromatogram (a), while the SPME chromatogram shows a stronger signal for early eluting sesquiterpenes. Higher boiling sesquiterpenes are only adequately represented in the hexane extract chromatogram.

Fig. 1 Comparison of three cannabis volatile sampling techniques. (a) static headspace; (b) solid-phase microextraction (SPME); (c) liquid extract (Hexane)

Columns

Representative chromatograms for the three columns tested (DB5, DB17, and VF35, in order of increasing polarity) are shown in Figure 2. The DB5 (c) and DB17 (a) columns performed comparably well in terms of resolution and overall run time (≤ 30 minutes), while the VF 35 column required a runtime in excess of 35 minutes for the chosen temperature program.

Fig. 2 Representative chromatograms for the three columns tested: (a) DB17; (b) VF-35; (c) DB-5

Solvent optimisation

Three solvents—hexane, isopropanol, and ethyl acetate—were initially trialed for liquid extraction. All three performed equally well in terms of extraction efficiency, but both isopropanol and, to a lesser extent, ethyl acetate were causing early eluting compounds to show undesirable peak shapes during gas chromatography (GC) analysis. A comparison of peak shapes is shown in Figure 3. Furthermore, hexane extracts appeared as light green, clear extracts, whereas ethyl acetate and isopropanol extracts were opaque and of a much darker color, indicating the presence of a higher proportion of non-target compounds.

Fig. 3 Comparison of peak shapes for early eluting compounds using three different injection solvents: hexane (black), ethyl acetate (green), isopropanol (red). Callout: enlarged detail of peaks at 8.05 min, separated, with vertical line through apex highlighting symmetry; solvents other than hexane cause undesirable peak shapes.

Compound identification and resolution

A total of 48 individual compounds were detected across multiple strains used. Of those, 22 were monoterpenes and -terpenoids, with the remainder consisting of sesquiterpenes and terpenoids. Compound names, base peak (quantifier) ion m/z, retention times and indices, and identification status are summarized in Table 1. Most monoterpene identities predicted by spectral library matching were confirmed using authentic standards, with the exception of fenchol and trans-2-pinanol. The α thujene, β-phellandrene, and thujanol identities were confirmed by matching published retention times indices (RIs). Sesquiterpene preliminary identities were confirmed using authentic standards for eight out of 25 compounds (Table 1), and RI matching for a further three.

Table 1. Compounds detected in this study, their base peak (quantifier) ions (m/z), retention times (RT) and indices (RI), resolution to the closest preceding (Rp) and following peaks (Rf), and identification status (specID: based on library match; RI: based and library match and retention time index; confirmed: matches spectrum and RT of authentic standard; Val: confirmed compound used for method validation). Name m/z RT (Minutes) RI Rp Rf Status α Thujene 93.0 7.866 932 n/a 1.7 RI α-Pinene (+/−) 93.0 8.056 940 1.9 1.6 Val Camphene 93.0 8.505 958 4.8 1.7 Val Sabinene 68.1 9.067 979 1.5 1.2 Val β-Pinene (+/−) 69.2 9.215 985 1.3 1.4 Val Myrcene 93.0 9.380 991 2.6 1.4 Val 3-Carene 79.0 10.007 1014 1.9 1.6 Val α-Terpinene 119.0 10.230 1023 2.0 1.6 Val p-Cymene 93.0 10.437 1031 4.1 1.9 Val Limonene 121.0 10.565 1035 3.0 15.7 Val β-Phellandrene 93.1 10.660 1039 1.2 1.0 specID, RI Eucalyptol 93.0 10.678 1040 1.7 2.1 Val Ocimene isomer 105.1 10.960 1050 1.8 2.2 Val γ-Terpinene 93.1 11.367 1064 2.9 2.3 Val 4-Thujanol 93.1 11.750 1077 3.4 2.9 specID, RI Terpinolene 93.0 12.150 1090 1.7 2.5 Val Fenchone 81.1 12.333 1096 1.8 1.7 confirmed Linalool 81.0 12.496 1101 1.4 2.7 Val Fenchol 81.1 13.201 1128 7.4 2.0 specID trans-2-Pinanol 93.1 13.421 1135 2.0 13.0 specID Isopulegol 93.0 14.022 1156 1.6 2.8 Val Borneol 95.1 14.758 1181 5.4 4.5 confirmed Dodecane 57.1 15.370 1201 6.7 7.2 IS Nerolidol 93.1 16.783 1253 2.6 22.5 Val β-Bergamotene 119.1 21.015 1406 3.0 3.4 specID α-Bergamotene 93.1 21.341 1420 3.6 2.3 specID trans-Caryophyllene 93.1 21.600 1430 2.9 3.4 Val g-Elemene iso1 121.0 21.768 1437 1.4 1.8 specID Bergamotene iso3 119.0 21.848 1440 0.8 3.5 specID, RI α-Guaiene 105.0 21.944 1444 1.7 1.8 specID Farnesene 69.2 22.270 1456 4.1 2.1 confirmed Humulene 69.1 22.524 1466 3.9 32.1 Val epi-β-Selinene 93.1 23.396 1499 1.4 1.5 specID (−)-α-Selinene 105.1 23.567 1506 1.7 1.1 specID, RI Sesquiterpenes, coeluting 93.1 23.574 1506 1.5 1.4 specID δ-Guaiene 161.0 23.662 1510 1.4 2.6 specID, RI β-Guaiene 161.0 24.139 1530 1.2 1.5 specID Farnesol 69.0 24.222 1534 3.2 3.7 confirmed α-Bisabolene 93.1 24.512 1546 1.1 0.7 specID Guaia-3,9-diene 161.1 24.572 1548 1.5 1.6 specID 3,7(11)-Selinadiene 161.1 24.680 1553 1.6 1.1 specID β-cis-Caryophyllene 93.0 24.968 1564 2.5 4.1 Val γ-Elemene iso2 121.1 25.152 1572 1.4 2.8 specID Caryophyllene Oxide 93.0 25.691 1593 1.6 4.8 Val Guaiol 93.0 25.970 1605 2.3 7.0 Val β-Cadinene 189.1 26.666 1636 1.9 1.3 specID γ-Gurjunene 59.1 27.400 1667 1.8 1.5 specID Sesquiterpene 107.0 27.578 1675 1.2 1.4 specID α-Bisabolol 93.0 27.989 1692 1.9 n/a Val

Resolution was calculated based on the following equation (discussed further in the "Materials and methods" section):

$R_{1,2}=2{\frac {T_{2}-T_{1}}{w_{2}+w_{1}}}$

Values presented in Table 1 are based on the respective base peak (qualifier) ion and are given for the closest preceding (R_p) and following peak (R_s).

Linearity

A combined cannabis terpenes (CT) and sabinene standard was injected at ten concentrations from 100 μg/mL to 0.195 ng/mL, based on a 1:2 dilution series of the 100 μg/mL standard. All calibration curves obtained were linear with R2 values greater than or equal to 0.994. Based on individual detection limits, the curves spanned all ten calibration points, except linalool, ocimene, and trans-caryophyllene (100–0.391 μg/mL); caryophyllene oxide, guaiol, and isopulegol (100–1.56 μg/mL); and α-bisabolol and β-cis-caryophyllene (100–12.5 μg/mL). Nerolidol was removed from the analysis due to poor peak shapes and high detection limits.

Detection and quantitation limits

Limit of detection (LOD) and limite of quantitation (LOQ) were determined as LOD = 3.3 × σ/S and LOQ = 10 × σ/S, with σ representing the standard deviation of the compound and S the slope of the compound’s calibration curve. Detection limits range from 0.017 μg/mL to 1.144 μg/mL (Table 2); quantitation limits range from 0.052 to 3.468 μg/mL (β-pinene and α-bisabolol for both LOD/LOQ).

Table 2. Detection (LOD) and quantitation (LOQ) limits in μg/mL for the validation compounds. Compounds are listed in order of their retention times. S: slope of calibration curve Compound S LOD LOQ α-Pinene (+/−) 0.039129 0.022 0.068 Camphene 0.023849 0.032 0.097 Sabinene 0.038856 0.026 0.079 β-Pinene (+/−) 0.047192 0.017 0.052 Myrcene 0.028807 0.047 0.142 3-Carene 0.030862 0.030 0.091 α-Terpinene 0.021459 0.038 0.115 p-Cymene 0.078714 0.021 0.065 Limonene 0.017349 0.066 0.200 Eucalyptol 0.008849 0.115 0.348 Ocimene 0.015782 0.057 0.174 γ-Terpinene 0.032899 0.046 0.139 Terpinolene 0.018898 0.045 0.137 Linalool 0.013307 0.060 0.181 Isopulegol 0.005376 0.045 0.136 trans-Caryophyllene 0.010502 0.092 0.326 Humulene 0.032275 0.021 0.065 β-cis-Caryophyllene 0.005202 0.388 1.177 Caryophyllene Oxide 0.004599 0.252 0.764 Guaiol 0.008546 0.107 0.324 α-Bisabolol 0.010087 1.144 3.468

Accuracy

To determine method accuracy, three replicates of a sample preparation were spiked with 0.5, 0.25, and 0.05 mg/g of CT and sabinene and extracted as described above. Unspiked biomass was extracted to determine endogenous levels. The recovery requirements of the spike as expressed in % adjusted for endogenous levels were 80–120%. Measured values ranged from 61.8% to 121.4%, with the majority of compounds ranging from 80% to 120% recovery across the three concentrations (Table 3).

Table 3. Recovery of analytes in spiked sample at three concentration levels % Spike Recovery Compound High Mid Low α-Pinene (+/−) 106.4 108.2 97.5 Camphene 112.6 121.4 119.0 Sabinene 108.8 115.7 98.7 β-Pinene (+/−) 110.7 113.7 102.2 Myrcene 106.5 104.8 95.5 3-Carene 112.4 118.7 115.7 α-Terpinene 78.8 84.5 81.8 p-Cymene 112.2 115.6 116.5 Limonene 110.6 114.6 107.9 Eucalyptol 116.6 109.9 119.2 Ocimene 101 105.7 98.6 γ-Terpinene 104.6 108.9 101.6 Terpinolene 97.1 100.4 93.0 Linalool 113.6 114.5 105.5 Isopulegol 61.8 66.2 82.6 trans-Caryophyllene 93.5 99.1 89.7 Humulene 92.5 98.3 90.7 β-cis-Caryophyllene 62.0 72.9 87.7 Caryophyllene Oxide 69.0 76.3 84.8 Guaiol 102.6 104.1 93.3 α-Bisabolol 98.7 100.3 93.5

Notes

This presentation is faithful to the original, with only a few minor changes to presentation. Some grammar and punctuation was cleaned up to improve readability. In some cases important information was missing from the references, and that information was added. The original includes an "in prep" citation at citation 8, but the citation appears to be about grapes and tobacco and not related to cannabis; it could not be verified as published and has been omitted for this version.