Circumventing glycosylation in heterologous cannabinoid production

ABSTRACT

Throughout the ages, the Cannabis plant has been used either as a source of fibre, nutrition or for medicinal and recreational use. For the latter aspects, the phytocannabinoids are responsible. A class of metabolites, originally found in C. sativa, that interact with the endocannabinoid system, which is associated with the reward system in the brain. As part of their pharmacological activity, they have been found to modulate pain and hunger, have anti inflammatory and anti tumour properties and show psychotropic effects. While a plethora of different phytocannabinoids is produced by the plant, the most abundant are Δ9 tetrahydrocannabinol (THC), cannabidiol (CBD) and their common precursor cannabigerolic acid (CBGA). While most of the pharmacological spectrum is highly desirable, the psychoactive properties of THC are considered a major limitation for clinical use. This causes the high pharmacological potential of this species to still be largely unharnessed, as non psychoactive phytocannabinoids only have a low yield in the plant and are difficult to separate from the psychoactive THC. As chemical production of synthetic cannabinoids is costly, biological production in heterologous host systems could provide a suitable avenue to produce specific, desired cannabinoids. This would not only enable industrial scale production of rare or novel cannabinoids as pharmaceuticals but could also decrease regulatory hurdles associated with Cannabis cultivation. Previous research aiming for the heterologous production of cannabinoids can broadly be classified according to the used host organism and all encounter different challenges. First is bacterial production, with E. coli being the first host with whom heterologous production of THCA and CBGA has been successful. Unfortunately, the combination of a low yield due to a low precursor supply of GPP and observed antibacterial activity of cannabinoids makes this host unsuitable for large scale industrial production. Second are yeasts and with them, heterologous production could also be successfully shown. While the yield is much better compared to bacterial production, a high degree of metabolic engineering was necessary to achieve mediocre amounts of cannabinoids. Especially the supply of GPP had to be massively improved. Additionally, cytotoxic activity has been observed again, limiting the potential yield of the desired compounds. Plants on the other hand offer the potential of compartmentalization and can provide higher quantities of GPP as a precursor. While the production of OA and the downstream conversion of CBGA to specific cannabinoids like THC and CBD has been shown, so far, no production of CBGA from OA and GPP has been successfully performed in plants. Instead, the produced OA is glucosylated and not prenylated. This study aims to close this last bottleneck in the reaction cascade to achieve heterologous cannabinoid production in planta. Two different core methods are utilized to achieve the desired CBGA formation, one being stable chloroplast transformation in N. tabacum cv. petit Havana and the other being Agrobacterium mediated transient gene expression in N. benthamiana. Advantages of chloroplast transformation include the absence of small molecule glycosylation, precursor availability in the form of hexanoic acid and GPP and high transgenic protein accumulation. The main disadvantage of this method lies in the time needed to generate stable transformed plants which can be used for testing. Transient expression on the other hand generates a much smaller timeframe for experimental procedures, allowing more parameters to be tested. As part of this study tested combinations include cytosolic or chloroplast enzyme localization, an increase of the intracellular GPP supply, metabolic channelling using the SpyTag/SpyCatcher system, and identification and addition of a β glucosidase from C. sativa. Neither using stable chloroplast transformation, nor any variant using transient gene expression was sufficient to achieve CBGA production in planta. For chloroplasts, no integration event of the expression cassette could be observed, prohibiting the production of the desired compounds or any intermediates. For transient transformation, neither enhancement of the intracellular GPP pool, nor metabolic channelling was enough to achieve CBGA formation. In all samples, the reaction seemed to halt due to the glucosylation of OA to form OA Glc. Literature suggested that this loss of available intermediate due to glucosylation does not occur in the original host and is a side effect of heterologous expression in Tobacco spp. Nevertheless, we found that OA Glc is also present in C. sativa in a much higher quantity than free OA. We then identified five different potential β-glucosidases from transcriptomic data which could be responsible for the cleavage of OA-Glc in C. sativa and thus reintroduce OA to the pathway. One of the five candidates, namely the vicianin hydrolase like enzyme (CsVH), showed a positive impact on the intracellular supply of free OA and was able to raise the concentration to a similar level compared to the tested C. sativa cultivars. Nevertheless, prenylation of OA to form CBGA could still not be observed. Literature suggests no intermediates or other routes of production than direct prenylation of OA with GPP to form CBGA in C. sativa. Especially since pathway reconstruction in bacteria and yeast has been successful under this assumption, the biosynthesis route was believed to be fully elucidated. Nevertheless, as part of this study, both OA-Glc and free OA were found in the original host with a clear imbalance towards the glycosylated form. Additionally, the identification of a glucosidase from C. sativa with a positive impact on the intracellular concentration of free OA in Tobacco was possible, thus OA Glc cleavage activity of this enzyme is assumed. Further testing of said enzyme is necessary to verify the observed in planta activity and its importance for phytocannabinoid biosynthesis. The former can be done using in vitro enzyme characterization and the latter by gene silencing and subsequent measuring for a change in phytocannabinoid content. Even though production of the desired compound CBGA was not successful, by identifying CsVH an important step towards heterologous cannabinoid production in planta was made. The presence of OA Glc in C. sativa suggests that the phytocannabinoid production pathway must be updated to account for this intermediate. Additionally, even though further testing is still necessary, CsVH must be added as potentially responsible for the cleavage reaction and reintroduction of OA into the pathway.