Comparison of Three Gas Chromatographic Methods – Identification of Terpenes and Terpenoids in Cannabis sativa L.

1. Introduction

Today we know, without any doubt, that the presence of terpenes and terpenoids in the Cannabis sativa L. plant is important, not only from the biogenetical point of view but also from the medical perspective. These bioactive compounds play an important role concerning use of cannabis as a medicament. Unfortunately, up-to-date there is not enough knowledge concerning the importance of these compounds and their ratio with others bioactive compounds, mainly cannabinoids. At present, it is very important to clarify the importance of the compounds quantity, the content compound types and their ratios, and to understand the medicinal power of this plant for the treatment of various diseases.

The biosynthesis of terpenes in cannabis is comprised of two different pathways [1,2] . The first pathway is the plastidial methylerythritol phosphate pathway [3], which starts with the substrates pyruvate and glyceraldehyde-3-phosphate [4,5]. Through several steps geranyl diphosphate is produced, which then interacts with olivetolic acid for cannabigerolic acid origination. Additionally, this pathway is also a precursor for monoterpene origination [6,7]. The second pathway is the cytosolic mevalonate pathway [8] which starts with acetoacetyl-CoA and after several steps forms farnesyl diphosphate followed by sesquiterpenes formation.

Terpenes found in hemp are not unique to this plant, as they are also found in other plants. Many terpenes have medical potential but their bioactivity obviously depends on the cannabis chemotype, due to terpenes and cannabinoids having different quantitative contents and ratios [9,10,11,12,13,14,15]. Of course, both the cannabis chemotype used and the patient’s genetics play a major role in the treatment. In addition, the amount used in one patient may be too high and in another one insufficient. At present, we need to know much more about the biodynamic effect of terpenes in cannabis in order to be able to breed cannabis plants that are suitable for therapeutic use. We must understand that the therapeutic effectiveness of terpenes contained in cannabis is different when we use plant material (by smoking, vaporization or in capsules), or as extracts (in oil, capsules, suppositories, creams and the other preparations). We must also understand that many terpenes are not stable compounds. It is possible that some of these phytochemicals are artifacts that formed in the resin on the plant due to various reasons (as described below). For instance, UV sunlight exposure, harvesting, drying, storage and processing of the plant, and even during the analysis. Smoking and vaporization gives rise to other substances, not fully studied, that can also affect patients (or other users) either positively or negatively. So far, there is a discussion on whether terpenes act as such or have a synergistic or entourage effect [16,17,18,19,20,21].

LaVigne [22] found that the terpenes α-humulene, geraniol, linalool, and β-pinene produced cannabinoid tetrad behaviors in mice [23,24,25,26], suggesting cannabimimetic activity. Further, some mice behaviors could be blocked by cannabinoid or adenosine receptor antagonists, suggesting a mixed mechanism of action.

We must take into consideration that upon aging essential oils can undergo oxidation and polymerization which may result in a loss of pharmacological properties. Heat, light and air can lead to its oxidation, polymerization, isomerization, thermal rearrangement, or dehydrogenation [27]. Inflorescence stored using the novel packaging approach is a significant step towards providing patients with cannabis inflorescence of reproducible and reliable terpene content, an important component of inflorescence efficacy [28].

The aim of this study was to compare three different gas chromatography/mass spectrometry methods - hexane based liquid extraction (Liq), static head-space extraction (HS), and head-space solid phase microextraction (SPME), - in order to determine which method best identified volatile compounds in cannabis samples, mainly monoterpenes/monoterpenoids and sesquiterpenes/sesquiterpenoids. We used four chemotypes to also compare possibilities of content compounds identification in different samples.

2. Experimental

2.1. Methods

Standards: Commercially available standards for α-pinene, camphene, β-pinene, myrcene, Δ³-carene, α-terpinene, p-cymene, limonene, 1,8-cineole, α-ocimene, trans-β-ocimene, γ-terpinene, terpinolene, linalool, isopulegol, geraniol, β-caryophyllene, α-humulene, cis-nerolidol, trans-nerolidol, caryophyllene oxide, guaiol, and α-bisabolol were obtained from Restek (Bellefonte, PA, USA).

Plant material: Dry female flowering tops of four different chemotypes (LOH LL1 - Green fields chemotype, LOH LL2 – Titan chemotype, LOH LL3 – Black Domina chemotype, LOH LL4 – Neptune chemotype) used for medical treatment, cultivated in Israel, were used for analysis. These four varieties were selected randomly to compare different chemotypes.

Sample preparation: The ground plant material was extracted with n-hexane (final concentration was 1 mg/ml) with occasional shaking for half an hour. One microgram of the sample thus prepared was injected for Liq analysis. 25 mg of plant material was used for HS and 0.3 mg for SPME analysis.

Instrument: GC/MS [Agilent 7890B GC, Agilent 5977B MSD, PAL 3 (RSI 85)].

Column: Agilent Technologies, Inc., HP-5MS UI, 30 m x 250 μm, film 0.25 μm.

2.2. Experimental Conditions for HS

Incubation time: 6 min; Incubation temperature: 80ºC.

2.3. Experimental Conditions for SPME:

Incubation time: 10 min, incubation temperature: 60ºC, GC cycle time: 5 min, fiber conditioning station temperature: 250 ºC, pre desorption conditioning time: 2 min, sample extraction time: 10 min, sample desorption time: 1 min

2.4. Experimental Condition for All Three Analyses

The column temperature was initially 35 °C for 5 min, followed by temperature ramping from 35-150 °C at 5 °C/min, then to 250 °C at 15 °C/min (inlet: 250 °C; detector: 280 °C; split ratio 5:1;); gas: helium (flow rate: 1 mL/min).

Analytical method validation - selectivity, specificity, accuracy, precision, linearity, range, limit of detection, limit of quantification, ruggedness, and robustness were performed [29]. They are beyond the scope of this manuscript and will be published in another publication.

2.5. Identification

The content compounds were identified by comparison to standards, retention times, retention indices, mass spectra of particular compounds, and the spectral matching of libraries NIST/EPA/NIH Mass Spectral Library 2017, Wiley Registry of Mass Spectral Data 11th Edition, FFNSC3, ©2015, and Adams Essential Oils Library.

3. Results

As we did not have all the main compounds as standards, it was impossible to quantify all these terpenes/terpenoids exactly. We therefore “quantified” volatile compounds by a normalization method. We chose the biggest peak (by space) as 100% and calculated the relative ratios of the other ones (in Tables as. % norm). The ten main peaks in each method are marked in Tables below in bold.

Sample LOH LL1 (Table 1) revealed β-myrcene as the biggest peak in HS and β-caryophyllene in Liq and SPME methods. There are 7 different compounds above 50%; in norm Liq – 7 ones above 50%, in HS – 1 above 50%, and in SPME – 6 above 50%. Between the ten main compounds above 50% were identified β-myrcene - 3 x, β-caryophyllene – 2 x, γ-elemene – 1 x, α-humulene – 2 x, α-bulnesene 2 x, γ-selinene – 2 x, selina-3,7(11)-diene - 2 x. Altogether within the ten main compounds 14 different terpenes/terpenoids were identified. Altogether 51 compounds were identified by HS (97.53% of total volatiles), 46 compounds by SPME (88.36%) and 37 compounds by liquid (72.73%) GC/MS, what was together 67 different volatile compounds. The same 28 compounds were found in all three analyses. Comparison of all three methods of analysis is presented in Chromatogram 1 - Liq (Chromatogram 1a), HS (Chromatogram 1b), and SPME (Chromatogram 1c).

Chromatogram 1a. Hexane based liquid extraction method (Liq) - sample LOH LL1 chemotype. Counts vs. Acquisition Time (min)

Chromatogram 1b. Static head-space extraction method (HS) - sample LOH LL1 chemotype. Counts vs. Acquisition Time (min)

Chromatogram 1c. Head-space solid phase microextraction method (SPME) - sample LOH LL1 chemotype. Counts vs. Acquisition Time (min)

Chromatogram 1. Comparison of all three methods of analysis - sample LOH LL1 chemotype. Counts vs. Acquisition Time (min).

Chromatogram 1. Comparison of all three methods of analysis - sample LOH LL1 chemotype. Counts vs. Acquisition Time (min).

Results of analysis of the sample LOH LL2 are presented in Table 2. Limonene was the biggest peak in HS and β-caryophyllene in Liq and SPME methods. Three different compounds were above 50%; in norm Liq – 3 above 50%, in HS – 1 above 50%, and in SPME – 2 above 50%. Of the ten main compounds identified above 50% were limonene 1 x, β-caryophyllene – 2 x, and α-humulene – 1 x. Altogether 17 different terpenes/terpenoids were found between the ten main compounds. The same 24 compounds were found in all three analyses. Altogether 74 compounds were identified by HS (97.87% of total volatiles), 57 compounds by SPME (90.80%) and 38 compounds by liquid (80.02%) GC/MS, what was together 94 different volatile compounds. Comparison of all three methods of analysis is presented in Chromatogram 2.

In Table 3 limonene was the biggest peak in HS and SPME and selina-3,7(11)-diene in Liq. Five different compounds were above 50%; in norm Liq – 4 above 50%, in HS – 2 above 50%, and in SPME – 5 above 50%. Between the ten main compounds identified using all three methods was above 50% β-myrcene – 2 x, limonene – 3 x, β-caryophyllene – 2 x, γ-selinene – 2 x, selina-3,7(11)-diene – 2 x. Altogether there were 20 different terpenes/terpenoids between the ten main compounds. The same 21 compounds were found in all three types of analysis. Altogether 55 compounds were identified by HS (96.45%), 49 compounds by SPME (85.43%) and 45 compounds by Liq (86.85%) GC/MS, what was together 84 different volatile compounds. Comparison of all three methods of analysis is presented in Chromatogram 3.

Table 4 pointed out limonene as the biggest peak in HS and β-caryophyllene in Liq and SPME. Four different compounds were above 50%; in norm Liq – 1 above 50%, in HS – 3 above 50%, and in SPME – 2 above 50%. Between the ten main compounds gathered from all three methods were above 50% β-myrcene – 1 x, limonene – 1 x, β-caryophyllene – 3 x, and α-humulene – 1 x. Altogether 17 different terpenes/terpenoids were identified between the ten main compounds were. The same 21 compounds were found in all three types analysis. Altogether 57 compounds were identified by HS (97.17%), 47 compounds by SPME (88.85%), and 34 compounds by Liq (88.10%) GC/MS, what was together 77 different volatile compounds. Comparison of all three methods of analysis is presented in Chromatogram 4.

A comparison of the number of identified substances for the four different chemotypes by the three different methods can be found in the Table 5 and results of quantitative determination of compounds for which we had commercially available standards are in Table 6.

4. Discussion

Gas chromatography analysis of the essential oil from Cannabis sativa was published already in 1957 [30]. In oil from fresh large leaves of female Cannabis sativa the compounds identified by gas chromatography were myrcene, limonene, α-humulene, and β-caryophyllene [31]. By steam distillation of fresh Indian Cannabis sativa L. from Kashmir 21 terpenes/terpenoids were identified in the essential oil [32].

Static headspace gas chromatography has already been used for marijuana and hashish analysis by Hood et al. [33,34]. They identified 16 terpenes and 1 terpenoid in the samples. For simultaneous quantification of 93 terpenoids present in air-dried Cannabis inflorescences and extracts static headspace – GC/MS/MS was used [35]. We also used GC/MS for identification of volatiles in different chemotypes of cannabis (medical and industrial) and published content volatiles, mostly terpenes/terpenoids, and their ratios in cannabis inflorescences and essential oils [36]. Thirteen chemotypes with different main terpenes/terpenoids were presented.

Solid-phase microextraction GC/MS was used to identify cannabidiol, Δ⁸-tetrahydrocannabinol, Δ⁹-tetrahydrocannabinol, and cannabinol in pure water and human saliva [37]. For cannabinoids determination in cannabis samples GC/MS method was developed [38]. Yang et al. [39] identified 13 monoterpenes, 4 monoterpenoids, and 14 sesquiterpenes in cannabis essential oil. The three mentioned gas chromatography techniques (HS, SPME, and liquid injection) were compared [40]. All three were excellent for the lower boiling monoterpenes. In HS sesquiterpenes were underrepresented. SPME gave a stronger signal for early eluting sesquiterpenes. Higher boiling sesquiterpenes were only adequately represented in liquid injection (hexane extract). Myers et al. [41] compared headspace-syringe and liquid injection-syringe techniques to the more modern headspace-solid phase microextraction arrow and direct immersion-SPME arrow. They used 23 terpene/terpenoids standards and determined from the results that the liquid injection-syringe method is the most straightforward and robust method.

We compared three different gas chromatography/mass spectrometry methods - static head-space extraction, head-space solid phase microextraction, and hexane based liquid extraction - to identify volatile compounds in cannabis samples, mainly monoterpenes/monoterpenoids and sesquiterpenes/sesquiterpenoids. We found hexane to be the best solvent for analysis of a liquid samples. Liquid samples give the most complex spectrum of the main mono- and sesquiterpenes/terpenoids as sesquiterpenes/sesquiterpenoids can be seen with higher retention times. Such liquid samples also have the advantage for absolute quantification of terpenes. The static headspace chromatogram gives the best representation of monoterpenes and monoterpenoids but a weaker signal for sesquiterpenes and sesquiterpenoids. Solid-phase microextraction gives significant spectrum of sesquiterpenes and sesquiterpenoids with shorter retention times and weaker signal for monoterpenes and monoterpenoids.

It seems that the results of Liq and SPME are the most similar (but not in all cases), so the analysis of the extract prepared with an organic solvent (hexane) will be the most suitable for the quantitative determination of these substances.

Altogether 26 terpenes/terpenoids were between the ten main present in four different chemotypes. They were divided with Liq containing 16 terpenes/terpenoids (2 terpenes, 1 terpenoid, 9 sesquiterpenes, and 4 sesquiterpenoids), 15 in HS (5 terpens, 2 terpenoids, 8 sesquiterpenes) and 17 in SPME (4 terpenes, 3 terpenoids, 10 sesquiterpenoids). The main terpene in chemotype LOH LL3 from the HS and SPME methods was limonene, but as we can see from Table 3 the other terpenes/terpenoids do not follow the same relative ratio. Chemotype LOH LL4 has a similar situation. The main terpene in Liq and SPME is β-caryophyllene, but as we can see from Table 4 the other terpenes/terpenoids do not follow the same relative ratio. As we can see from Tables 1 to 4, the terpenes found most often from amongst the ten main were β-myrcene, limonene, β-caryophyllene, α-humulene, γ-selinene, and selina-3,7(11)-diene. The most reproducible analysis for quantitative analysis appears to be from Liq samples.

It is now well known that different cannabis chemotypes (also mentioned as variety, strain, chemovar, cultivar, phenotype, or genotype) with the same content of major cannabinoids act differently in the treatment of the same patient. Not every chemotype is suitable for a given patient. Cannabis constituents (whether cannabinoids, terpenes/terpenoids or the other bioactive substances) interact with each other and can thus increase (or decrease) the effectiveness of a given chemotype [16,42]. If these effects are independent, synergistic or entourage, they still need to be thoroughly studied. In addition, each of us is genetically different and therefore one chemotype that is suitable for one patient might not be suitable for another one. It should also be emphasized that a given chemotype of cannabis that is suitable for a particular patient in a given disease may not (but sometimes may) cure a different disease.

5. Conclusions

In conclusion we can say that for the final evaluation of the comparison of the three methods of analysis, extraction with hexane gives balanced results (which is advantageous for quantitative analysis), although the other two methods allowed for the identification of more substances. This means that the same method should be used everywhere for the quantitative evaluation of constituents in cannabis. Only in this way it will be possible to objectively compare the results from different laboratories. The differences for the same sample analyzed in different laboratories will then be within the allowable error range.