Building a biofactory: Constructing glandular trichomes in Cannabis sativa.

Flowers of Cannabis sativa L. are densely covered with glandular trichomes containing cannabis resin that is used for medicinal and recreational purposes. The highly productive glandular trichomes have been described as 'biofactories.' In this review, we use this analogy to highlight recent advances in cannabis cell biology, metabolomics, and transcriptomics. The biofactory is built by epidermal outgrowths that differentiate into peltate-like glandular trichome heads, consisting of a disc of interconnected secretory cells with unique cellular structures. Cannabinoid and terpenoid products are warehoused in the extracellular storage cavity. Finally, multicellular stalks raise the glandular heads above the epidermis, giving cannabis flower their frosty appearance.

Genomic analyses of Symbiomonas scintillans show no evidence for endosymbiotic bacteria but does reveal the presence of giant viruses.

Symbiomonas scintillans Guillou et Chrétiennot-Dinet, 1999 is a tiny (1.4 µm) heterotrophic microbial eukaryote. The genus was named based on the presence of endosymbiotic bacteria in its endoplasmic reticulum, however, like most such endosymbionts neither the identity nor functional association with its host were known. We generated both amplification-free shotgun metagenomics and whole genome amplification sequencing data from S. scintillans strains RCC257 and RCC24, but were unable to detect any sequences from known lineages of endosymbiotic bacteria. The absence of endobacteria was further verified with FISH analyses. Instead, numerous contigs in assemblies from both RCC24 and RCC257 were closely related to prasinoviruses infecting the green algae Ostreococcus lucimarinus, Bathycoccus prasinos, and Micromonas pusilla (OlV, BpV, and MpV, respectively). Using the BpV genome as a reference, we assembled a near-complete 190 kbp draft genome encoding all hallmark prasinovirus genes, as well as two additional incomplete assemblies of closely related but distinct viruses from RCC257, and three similar draft viral genomes from RCC24, which we collectively call SsVs. A multi-gene tree showed the three SsV genome types branched within highly supported clades with each of BpV2, OlVs, and MpVs, respectively. Interestingly, transmission electron microscopy also revealed a 190 nm virus-like particle similar the morphology and size of the endosymbiont originally reported in S. scintillans. Overall, we conclude that S. scintillans currently does not harbour an endosymbiotic bacterium, but is associated with giant viruses.

Coral-infecting parasites in cold marine ecosystems.

Coral reefs are a biodiversity hotspot,1,2 and the association between coral and intracellular dinoflagellates is a model for endosymbiosis.3,4 Recently, corals and related anthozoans have also been found to harbor another kind of endosymbiont, apicomplexans called corallicolids.5 Apicomplexans are a diverse lineage of obligate intracellular parasites6 that include human pathogens such as the malaria parasite, Plasmodium.7 Global environmental sequencing shows corallicolids are tightly associated with tropical and subtropical reef environments,5,8,9 where they infect diverse corals across a range of depths in many reef systems, and correlate with host mortality during bleaching events.10 All of this points to corallicolids being ecologically significant to coral reefs, but it is also possible they are even more widely distributed because most environmental sampling is biased against parasites that maintain a tight association with their hosts throughout their life cycle. We tested the global distribution of corallicolids using a more direct approach, by specifically targeting potential anthozoan host animals from cold/temperate marine waters outside the coral reef context. We found that corallicolids are in fact common in such hosts, in some cases at high frequency, and that they infect the same tissue as parasites from topical coral reefs. Parasite phylogeny suggests corallicolids move between hosts and habitats relatively frequently, but that biogeography is more conserved. Overall, these results greatly expand the range of corallicolids beyond coral reefs, suggesting they are globally distributed parasites of marine anthozoans, which also illustrates significant blind spots that result from strategies commonly used to sample microbial biodiversity.

Parallel functional reduction in the mitochondria of apicomplexan parasites.

Extreme functional reduction of mitochondria has taken place in parallel in many distantly related lineages of eukaryotes, leading to a number of recurring metabolic states with variously lost electron transport chain (ETC) complexes, loss of the tricarboxylic acid (TCA) cycle, and/or loss of the mitochondrial genome. The resulting mitochondria-related organelles (MROs) are generally structurally reduced and in the most extreme cases barely recognizable features of the cell with no role in energy metabolism whatsoever (e.g., mitosomes, which generally only make iron-sulfur clusters). Recently, a wide diversity of MROs were discovered to be hiding in plain sight: in gregarine apicomplexans. This diverse group of invertebrate parasites has been known and observed for centuries, but until recent applications of culture-free genomics, their mitochondria were unremarkable. The genomics, however, showed that mitochondrial function has reduced in parallel in multiple gregarine lineages to several different endpoints, including the most reduced mitosomes. Here we review this remarkable case of parallel evolution of MROs, and some of the interesting questions this work raises.

Overcoming the challenges of preserving lipid-rich Cannabis sativa L. glandular trichomes for transmission electron microscopy.

Cannabis glandular trichomes produce and store an abundance of lipidic specialised metabolites (e.g. cannabinoids and terpenes) that are consumed by humans for medicinal and recreational purposes. Due to a lack of genetic resources and inherent autofluorescence of cannabis glandular trichomes, our knowledge of cannabinoid trafficking and secretion is limited to transmission electron microscopy (TEM). Advances in cryofixation methods has resulted in ultrastructural observations closer to the 'natural state' of the living cell, and recent reports of cryofixed cannabis trichome ultrastructure challenge the long-standing model of cannabinoid trafficking proposed by ultrastructural reports using chemically fixed samples. Here, we compare the ultrastructural morphology of cannabis glandular trichomes preserved using conventional chemical fixation and ultrarapid cryofixation. We show that chemical fixation results in amorphous metabolite inclusions surrounding the organelles of glandular trichomes that were not present in cryofixed samples. Vacuolar morphology in cryofixed samples exhibited homogenous electron density, while chemically fixed samples contained a flocculent electron dense periphery and electron lucent lumen. In contrast to the apparent advantages of cryopreservation, fine details of cell wall fibre orientation could be observed in chemically fixed glandular trichomes that were not seen in cryofixed samples. Our data suggest that chemical fixation results in intracellular artefacts that impact the interpretation of lipid production and trafficking, while enabling greater detail of extracellular polysaccharide organisation.

A polarized supercell produces specialized metabolites in cannabis trichomes.

For centuries, humans have cultivated cannabis for the pharmacological properties that result from consuming its specialized metabolites, primarily cannabinoids and terpenoids. Today, cannabis is a multi-billion-dollar industry whose existence rests on the biological activity of tiny cell clusters, called glandular trichomes, found mainly on flowers. Cannabinoids are toxic to cannabis cells,1 and how the trichome cells can produce and secrete massive quantities of lipophilic metabolites is not known.1 To address this gap in knowledge, we investigated cannabis glandular trichomes using ultra-rapid cryofixation, quantitative electron microscopy, and immuno-gold labeling of cannabinoid pathway enzymes. We demonstrate that the metabolically active cells in cannabis form a "supercell," with extensive cytoplasmic bridges across the cell walls and a polar distribution of organelles adjacent to the apical surface where metabolites are secreted. The predicted metabolic role of the non-photosynthetic plastids is supported by unusual membrane arrays in the plastids and the localization of the start of the cannabinoid/terpene pathway in the stroma of the plastids. Abundant membrane contact sites connected plastid paracrystalline cores with the plastid envelope, plastid with endoplasmic reticulum (ER), and ER with plasma membrane. The final step of cannabinoid biosynthesis, catalyzed by tetrahydrocannabinolic acid synthase (THCAS), was localized in the cell-surface wall facing the extracellular storage cavity. We propose a new model of how the cannabis cells can support abundant metabolite production, with emphasis on the key role of membrane contact sites and extracellular THCA biosynthesis. This new model can inform synthetic biology approaches for cannabinoid production in yeast or cell cultures.

Cannabis Glandular Trichome Cell Walls Undergo Remodeling to Store Specialized Metabolites.

The valuable cannabinoid and terpenoid metabolites of Cannabis sativa L. are produced by floral glandular trichomes. The trichomes consist of secretory disk cells, which produce the abundant lipidic metabolites, and an extracellular storage cavity. The mechanisms of apoplastic cavity formation to accumulate and store metabolites in cannabis glandular trichomes remain wholly unexplored. Here, we identify key wall components and how they change during cannabis trichome development. While glycome and monosaccharide analyses revealed that glandular trichomes have loosely bound xyloglucans and pectic polysaccharides, quantitative immunolabeling with wall-directed antibodies revealed precise spatiotemporal distributions of cell wall epitopes. An epidermal-like identity of early trichome walls matured into specialized wall domains over development. Cavity biogenesis was marked by separation of the subcuticular wall from the underlying surface wall in a homogalacturonan and α-1,5 arabinan epitope-rich zone and was associated with a reduction in fucosylated xyloglucan epitopes. As the cavity filled, a matrix with arabinogalactan and α-1,5 arabinan epitopes enclosed the metabolite droplets. At maturity, the disk cells' apical wall facing the storage cavity accumulated rhamnogalacturonan-I epitopes near the plasma membrane. Together, these data indicate that cannabis glandular trichomes undergo spatiotemporal remodeling at specific wall subdomains to facilitate storage cavity formation and metabolite storage.

Cannabis glandular trichomes alter morphology and metabolite content during flower maturation.

The cannabis leaf is iconic, but it is the flowers of cannabis that are consumed for the psychoactive and medicinal effects of their specialized metabolites. Cannabinoid metabolites, together with terpenes, are produced in glandular trichomes. Superficially, stalked and sessile trichomes in cannabis only differ in size and whether they have a stalk. The objectives of this study were: to define each trichome type using patterns of autofluorescence and secretory cell numbers, to test the hypothesis that stalked trichomes develop from sessile-like precursors, and to test whether metabolic specialization occurs in cannabis glandular trichomes. A two-photon microscopy technique using glandular trichome intrinsic autofluorescence was developed which demonstrated that stalked glandular trichomes possessed blue autofluorescence correlated with high cannabinoid levels. These stalked trichomes had 12-16 secretory disc cells and strongly monoterpene-dominant terpene profiles. In contrast, sessile trichomes on mature flowers and vegetative leaves possessed red-shifted autofluorescence, eight secretory disc cells and less monoterpene-dominant terpene profiles. Moreover, intrinsic autofluorescence patterns and disc cell numbers supported a developmental model where stalked trichomes develop from apparently sessile trichomes. Transcriptomes of isolated floral trichomes revealed strong expression of cannabinoid and terpene biosynthetic genes, as well as uncharacterized genes highly co-expressed with CBDA synthase. Identification and characterization of two previously unknown and highly expressed monoterpene synthases highlighted the metabolic specialization of stalked trichomes for monoterpene production. These unique properties and highly expressed genes of cannabis trichomes determine the medicinal, psychoactive and sensory properties of cannabis products.

The ARM Domain of ARMADILLO-REPEAT KINESIN 1 is Not Required for Microtubule Catastrophe But Can Negatively Regulate NIMA-RELATED KINASE 6 in Arabidopsis thaliana.

Microtubules are dynamic filaments, the assembly and disassembly of which are under precise control of various associated proteins, including motor proteins and regulatory enzymes. In Arabidopsis thaliana, two such proteins are the ARMADILLO-REPEAT KINESIN 1 (ARK1), which promotes microtubule disassembly, and the NIMA-RELATED KINASE 6 (NEK6), which has a role in organizing microtubule arrays. Previous yeast two-hybrid and in vitro pull-down assays determined that NEK6 can interact with ARK1 through the latter protein's Armadillo-repeat (ARM) cargo domain. To explore the function of the ARM domain, we generated fluorescent reporter fusion proteins to ARK1 lacking the ARM domain (ARK1ΔARM-GFP) and to the ARM domain alone (ARM-GFP). Both of these constructs strongly associated with the growing plus ends of microtubules, but only ARK1ΔARM-GFP was capable of inducing microtubule catastrophe and rescuing the ark1-1 root hair phenotype. These results indicate that neither the ARM domain nor NEK6's putative interaction with it is required for ARK1 to induce microtubule catastrophe. In further exploration of the ARK1-NEK6 relationship, we demonstrated that, despite evidence that NEK6 can phosphorylate ARK1 in vitro, the in vivo distribution and function of ARK1 were not affected by the loss of NEK6, and vice versa. Moreover, NEK6 and ARK1 were found to have overlapping but non-identical distribution on microtubules, and hormone treatments known to affect NEK6 activity did not stimulate interaction. These findings suggest that ARK1 and NEK6 function independently in microtubule dynamics and cell morphogenesis. Despite the results of this functional analysis, we found that overexpression of the ARM domain led to complete loss of NEK6 transcription, suggesting that the ARM domain might have a regulatory role in NEK6 expression.