Cannabinoids in seed extracts of Cannabis sativa cultivars

H. Mölleken1 and H. Husmann2

1Bergische Universität, Physiologische Chemie der Pflanzen, Gaußstraße. 20, 42119 Wuppertal, Germany

2Max-Planck-Institut für Kohlenforschung, Abt. Chromatographie, Kaiser-Wilhelm-Platz 1, 45470 Mülheim/Ruhr, Germany

        Mölleken, H. and H. Husmann 1997. Cannabinoids in seed extracts of Cannabis sativa cultivars. Journal of the International Hemp Association 4(2): 73, 76-79.  Cannabinoids are of great interest for pharmaceutical and medical purposes.  However, as some of them are also used as recreational drugs, forensic and economic problems could result if they occur in food products made from hemp seeds with cannabinoid concentrations that are high enough to have a psychoactive effect.  We have therefore (a) developed a method to determine cannabinoids in microgram quantities in every hemp product or in any part of the plant within 15-30-minutes analysis time and (b) analyzed various hemp oils and plant extracts of Cannabis cultivars from the EU and other regions to determine their cannabinoid content.  Our investigations have proven that the active substance cannot be found inside the achene, but may be attached to the shell (pericarp) of the fruit.  Thus, the content of THC in hemp oil can come only as a result of the technical process of harvesting the fruits.

Cannabis fruits show great variability both in their morphological and chemical characteristics.

        Many chromatographic methods have been applied to the forensic detection of cannabinoids, especially Δ9-tetrahydrocannabinol (THC), in media such as blood and urine. Rarely, however, have researchers been concerned with the cannabinoids in hemp oil manufactured for nutritional purposes (Alt and Reinhardt 1996, Brenneisen 1984, Callaway et al. 1997, Lanyon et al. 1981, Law et al. 1984, McBurney et al. 1986, Meesters and Eggink 1996, Norman et al. 1971, Novotny et al. 1976, Paul et al. 1987, Nakamura et al. 1990, Pertwee 1997, Struemper et al. 1997, Vicki et al. 1973, Vree 1972, Vree et al. 1973, Whiting and Manders 1982).
     We have thus developed a method to analyze cannabinoids in hemp oil. This new method can be used either directly, for analysis of an oil dissolved in methylene chloride, or indirectly by extraction of the oil or the fruits in methanol, whereby the methanol supernatant is analyzed after sedimentation of the oil phase or of the fruits. An advantage to this method is that after derivatization, each methylated cannabinoid (CB-OME), and each fatty acid (FA) as its fatty acid methyl ester (FAME), can be determined in a single GC analysis. It allows a quick and precise determination of Δ8-THC, Δ9-THC, cannabinol (CBN), cannabidiol (CBD) and most probably, other cannabinoids we have not investigated. We have used this method to look for cannabinoids in various hemp oils, in extracts of the “seeded“ flowers and in parts of these fruits.

Materials and Methods
        Various Cannabis strains including industrial hemp cultivars from the European Union and Eastern Europe were analyzed. Hemp fruits (1 g), either with or without removal of adherent particles through agitation in methanol, were homogenized in methanol, then treated with trimethylsulfonium hydroxide (TMSH) to produce FAMEs from the FA residues of the triglycerides (Mölleken and Theimer 1997 a, b) and methyl ethers of the cannabinoids. Hemp oils and various parts of the plant were directly dissolved in methanol before methylation as described previously.
     The resulting CB-OMEs and FAMEs were analyzed on an HP 5890 gas chromatograph (GC) equipped with either an FID or an HP 5870 Mass Selective Detector (GC-MS). OV-1 columns were used.

Figure 1. Chromatograms with A: indirect detection of Δ9-THC as Δ9-THC-O-Me after derivitisation with TMSH in methanol of the cannabinoid of the hemp oil from C. sativa var. Alice; B: direct detection of Δ9-THC after extraction with methanol of the hemp oil from C. sativa var. Alice.

Figure 2. Mass spectra of the four cannabinoids after methylation with TMSH in methanol.


Methodological Approach
        Figure 1 compares the direct and indirect methods for detecting Δ9-THC in the oil of one hemp variety. It becomes clear that both analyses are successful, though the retention times (RT) are shifted, with the methylated substances being detected earlier. The methylation seems to result in better resolution, however, so that monoterpenes and sesquiterpenes in flower extracts and FAMEs in hemp oils can be determined at the same time.
     As the above analyses characterized only Δ9-THC, we undertook a second experiment in which we looked for the methylation of other cannabinoids such as Δ8-THC, CBD and CBN, as well. We first examined a mixture of these substances by GC-MS (Fig. 2) after methylation (compare Pfleger et al. 1993). It is clear that (a) all cannabinoids are methylated and (b) the molecular ion of CBD shows that both hydroxy groups are methylated. Further details of the GC-conditions may be published after testing the method on urine and blood specimens after ingestion of (a) hemp oil and (b) hashish.
     Thus, we have found a method that allows the determination of cannabinoids alongside any triglyceride (TG) in the sample. It is also possible to determine THC directly, without derivatisation of the hemp oil. However, the high-boiling TGs cannot be eliminated from the column under the chromatographic conditions used for the detection of cannabinoids, and so in the course of time, will accumulate on the column, hindering the precision of subsequent analyses.

Figure 3. Comparison of the hemp oil extraction from the fruits with the female flower of the Hungarian variety Alice.

Figure 4. Analysis of hemp fruit extracts (without cleaning the shell) from Chinese, USSR/Dagestan and Spanish varieties for cannabinoids as CB-OMe, derivitisation with TMSH. FAME = fatty acid methyl ester.

Cannabinoids in hemp oil and fruit extracts
        First we compared the Δ9-THC in a hemp oil extracted from the fruits of the Hungarian variety “Alice“ with that found in its female flower (Fig. 3). Whereas the flower extract shows a peak like that of the Δ9-THC-standard, the “seed“ oil does not contain any Δ9-THC.
     To verify this analysis, we selected fruits from varieties of several origins (i.e., China, Dagestan, Spain) whose dried flowers normally contain 2-4 % Δ9-THC by dry weight (VIR 1977). We analyzed the effect of the fruit extraction method on cannabinoid content (Fig. 4). Whereas extracts from the complete fruits of the Chinese and Russian (Dagestan) varieties do not show any cannabinoid content, the Spanish variety shows large peaks for CBD and CBN. We therefore agitated fruits of this hemp variety with methanol for a few minutes to remove adherent particles (Fig. 5). The analyses of the clean fruits confirm that there are no cannabinoids in the seed. The only opportunity for cannabinoids to enter the oil is in the course of the harvest or other technical processes. These viscous substances are bound in the outer wax layer of the shell (pericarp), and they could enter the oil when it is being pressed (compare with Bósca 1995).
     To support this interpretation, we have compared the analyses for cannabinoids of oil extracts of the fruits (Fig. 6) and various hemp oils (Fig. 7) with a cannabinoid standard. Oils from varieties of the EU do not contain cannabinoids (for example Félina 34), with the exception of one oil from a mixture of several hemp varieties (Fédora/Futura), which shows a small peak representing CBN in picogram quantities/sample. Only one oil, from Hungary, contains all three cannabinoids in concentrations that hint at microgram quantities/sample.

Discussion and Conclusion
        In general, Cannabis contains three major cannabinoids, CBD, CBN, and Δ9-THC (Alt and Reinhardt 1996, Brenneisen 1984, Turner et al. 1980, Vree et al. 1971, 1972) that require an unambiguous detection in hemp oil for forensic reasons. Therefore, to analyze the cannabinoids in hemp oil, we established a method that allows a quick, but very precise determination, whereby these three compounds are well separated. The methylation guarantees a precise determination and at the same time, the triglycerides do not accumulate on the column. If desired, the FAMEs can be determined in the same analysis. Thus, this chromatographic method is sufficiently simple and reproducible to be used for applications of forensic interest. As Nakamura et al. (1990) established a comparable analysis for the Δ9-THC-metabolite (11-nor-Δ9-THC-COOH in urine), we will test this method on urine and blood samples in the future.
     With this method, we are not only able to identify the psychoactive cannabinoids, but we also get many peaks representing such cannabinoids, as cannabichromene or cannabigerol. Furthermore, Δ8-THC may be used as an internal standard in the analysis of hemp oils for cannabinoids, since it is not naturally occurring.
     Up to now, we have analyzed only Δ9-THC and three additional cannabinoids, as we have no other standards for comparison. We are, however, confident that all the cannabinoids can also be detected by this method. Others have already established methods that give good analyses, but these methods have not been tested for hemp oil analyses and require more time for measurement and preparation of the sample (Alt and Reinhardt 1996, Brenneisen 1984, Lanyon et al. 1981, Law et al. 1984, Mc Burney et al. 1986, Doorenbos et al. 1971, Meesters and Eggink 1996, Norman et al. 1971, Novotny et al. 1976, Paul et al. 1987, Nakamura et al. 1990, Pertwee 1997, Struemper 1997, Vicki et al. 1973, Vree et al. 1972, Vree et al. 1973, Whiting and Manders 1982).
     There have been only a few analyses of cannabinoids in hemp oil (Alt and Reinhardt 1996, Brenneisen 1995, Máthé and Bócsa 1995). Máthé and Bócsa analyzed only one variety, ‘Kompolti’, and Alt and Reinhardt made tests of urine and blood from test subjects that ate various hemp oils.
     These analyses prove by detailed investigations that hemp fruits do not contain any cannabinoids. Though Doorenbos et al. (1971) describe a cannabinoid content of plant parts decreasing in the following order: bracts, flowers, leaves, small stems, large stems, roots and seed, we could not find an indication of cannabinoids in the hemp fruit varieties we analyzed. Fetterman et al. (1971) also found traces of CBD, THC and CBN in some varieties after cleaning the fruits with chloroform. Vogelmann (1988) found to the contrary, that cannabinoids seem to appear first in the seedlings, cannabichromene being the first cannabinoid that can be detected.
     The opinion that hemp oil contains cannabinoids (Alt et al. 1996, Struempler et al. 1997) must be qualified. Hemp oils from fruits with cleaned shells should not contain Δ9-THC. Other oils may have low to high concentrations. An industrial technique that separates the shell from the fruits would make it possible to produce a hemp oil without any cannabinoids.

Figure 5. Analysis of particles washed from the shell of the Spanish hemp variety (derivitisation of the cannabinoids to CB-OME with TMSH in methanol).

Figure 6. Analysis of Δ9-THC-O-Me in oil extracts of fruits of some Cannabis varieties (fruits are pressed in methanol and derivitised with TMSH).

Figure 7. Chromatograms of cold pressed hemp oils (100 ml oil/Methanol extract 2:1 + 200 ml TMSH). 7a: hemp oil from ‘Felina 34’; 7b: hemp oil from a mixture of ‘Fedora’/’Futura’; 7c: hemp oil from a Hungarian variety.