Epigenome Editing for Cannabinoid Yield: Targets, Tools, and Priorities in Cannabis sativa
1Department of Agriculture, Gopalganj Science and Technology University, Gopalganj 8100, Bangladesh; smvahsan@gmail.com (S.M.A.); injamumrassel@gmail.com (M.I.-U.-H.); nayanhowladar@gmail.com (N.C.H.); rakibhasanhlc@gmail.com (M.R.H.)
2Institute of Cannabis Biotechnology, Andong National University, Andong 36729, Republic of Korea
3Department of Plant Medicals, Andong National University, Andong 36729, Republic of Korea
4Institute of Genomics for Crop Abiotic Stress Tolerance, Department of Plant and Soil Science, Texas Tech University, Lubbock, TX 79409, USA; skeya@ttu.edu
5Department of Biotechnology, Yeungnam University, Gyeongsan 38541, Republic of Korea; azizul@ynu.ac.kr
6Department of Biochemistry and Molecular Biology, Dhaka University, Dhaka 1000, Bangladesh; hossainshekhar@du.ac.bd
7Biosciences Division, Oak Ridge National Laboratory, 1 Bethel Valley, Oak Ridge, TN 37831, USA
*Correspondence: hwchoi@anu.ac.kr (H.W.C.); mdmerahm@ttu.edu (M.M.R.)Abstract
Cannabinoid content can vary several-fold across Cannabis sativa L. cultivars that carry functional, closely related CBDA synthase (CBDAS) and THCA synthase (THCAS) alleles, a difference that coding-sequence variation among functional alleles does not fully explain. Structural and copy-number variation at the synthase loci, the functional or pseudogenised state of synthase alleles, and linkage at the B locus account for much of the qualitative drug-versus-hemp chemotype; however, these genetic factors do not fully explain the quantitative, several-fold variation in cannabinoid content among cultivars that carry functional, near-identical synthase alleles. Three independent lines of evidence now indicate that chromatin-level regulation contributes to this variation. H3K4me3 and H3K56ac co-occupy the promoters and gene bodies of THCAS, CBDAS, OLS, and OAC exclusively in glandular trichome tissue while H3K27me3 marks the identical loci in vegetative tissues, indicating that a Polycomb-to-Trithorax chromatin switch is associated with trichome-specific cannabinoid gene expression, potentially independently of transcription factor availability. Progressive DNA hypomethylation accumulates during micropropagation in a cultivar-specific manner, with promoter-region differentially methylated positions reaching up to 22% by twenty subcultures, identifying DNA methylation maintenance as a candidate determinant of epigenetic stability in clonally propagated material, although the functional consequences for cannabinoid pathway gene expression and yield remain to be quantitatively established. In Cannabis indica cell suspension cultures, UV irradiation increases genome-wide DNA methylation (detected by MSAP, which does not resolve locus-specific changes) and elevates CBDAS transcript levels approximately 4-fold relative to non-irradiated controls. This coupling of an environmentally triggered methylation change to cannabinoid pathway gene expression in the Cannabis genus is suggestive, but the genome-wide MSAP signal and the use of C. indica cell cultures mean the locus-specific methylation change at the CBDAS promoter remains to be demonstrated. These findings identify specific, experimentally documented chromatin states as candidate targets for dCas9-effector intervention, including the H3K27me3 repressive state at cannabinoid loci that could be addressed by dCas9-KDM6A and the propagation-induced CG hypomethylation that could be addressed by dCas9-DNMT3A; these editing strategies remain hypotheses that have not yet been tested in Cannabis. We synthesize functional evidence from Catharanthus roseus, Papaver somniferum, and Artemisia annua demonstrating that equivalent chromatin switches control secondary metabolite yield in medicinal plants, evaluate which dCas9-effector architectures are most appropriate for each Cannabis chromatin target, and identify the critical mechanistic gaps that must be closed before epigenome editing can be rationally deployed for cannabinoid yield enhancement in Cannabis.
1. Introduction
Cannabis is among the fastest-growing pharmaceutical crops globally, with global legal Cannabis sales projected to reach USD 58 billion by 2028 [1]. Its therapeutic and commercial value is determined almost entirely by the content and composition of cannabinoids and terpenes accumulated in the secretory glandular trichomes of unfertilized female inflorescences [2,3]. Cannabidiol (CBD) and Δ9-tetrahydrocannabinol (Δ9-THC) are the two phytocannabinoids of greatest clinical relevance, with approved or late-stage therapeutic applications in epilepsy, neuropathic pain, anxiety disorders, and anti-inflammatory therapy [4,5]. Despite this pharmacological significance, cannabinoid content varies by as much as 10-fold across cultivars and more than 3-fold across developmental stages within the same variety [6]. No master transcriptional regulator of the cannabinoid pathway has been identified, no promoter architecture for its rate-limiting genes has been fully resolved, and no predictive model connects genotype to chemotype at the molecular level. A complementary explanation lies in epigenomics. The epigenome, defined as the complete set of DNA and histone modifications that regulate gene expression without altering the underlying DNA sequence, operates in plants through three interacting mechanisms: DNA methylation in the CG, CHG, and CHH contexts, covalent histone modifications such as H3K4me3, H3K27me3, and histone acetylation, and ATP-dependent chromatin remodeling, which together set which genes are accessible for transcription in a given cell type. This review asks whether these mechanisms govern cannabinoid biosynthetic gene expression and whether they can be harnessed to enhance cannabinoid yield. This question is especially pertinent in Cannabis sativa, where cannabinoid and terpene profiles vary several-fold across developmental stages and environments despite a stable genome, and even among cultivars carrying functional, near-identical synthase alleles, a plasticity that sequence-level variation alone does not explain. Whereas earlier Cannabis genomics centered on sequence-level features such as genome assembly, synthase gene structure, and copy-number variation, epigenome-scale studies remain scarce and have not functionally connected the chromatin state to cannabinoid pathway gene expression. This review consolidates that emerging evidence and evaluates how it can be advanced.
Cannabinoids are C21 terpenophenolic compounds, formed by the condensation of a resorcinol-type polyketide moiety with a monoterpene-derived isoprenoid unit, and they constitute a class of specialized metabolites largely restricted to Cannabis [7,8]. In the living plant, they accumulate predominantly as carboxylic (acidic) forms, such as cannabidiolic acid (CBDA) and Δ9-tetrahydrocannabinolic acid (THCA), which decarboxylate non-enzymatically to their neutral, pharmacologically active counterparts on heating, drying, or aging. More than one hundred and twenty distinct phytocannabinoids have been isolated from C. sativa, conventionally grouped into around eleven structural subclasses (the CBG, CBD, THC, CBC, CBND, CBE, CBL, CBN, CBT, cannabielsoin, and miscellaneous types), all of which derive from the single central precursor cannabigerolic acid (CBGA) and its propyl analog cannabigerovarinic acid through variation in side-chain length, cyclisation, and oxidation [7,8]. This shared biosynthetic origin coupled with extensive downstream tailoring accounts for the marked chemical diversity of the cannabinoid pool across tissues, developmental stages, and chemotypes. Beyond their pharmacological value to humans, cannabinoids appear to serve adaptive functions for the plant. They are synthesized and stored in the secretory cavity of capitate-stalked glandular trichomes that are most dense on the bracts of female inflorescences, a deployment consistent with protection of the developing seed. Acidic cannabinoids such as THCA and CBGA show antimicrobial and insecticidal activity in vitro, and a recent quantitative-genetic study reported that foliar cannabinoid concentration is inversely correlated with chewing-herbivore damage in the field and that Trichoplusia ni larvae grow more slowly and survive less well on high-cannabinoid leaves and on cannabinoid-amended diet, supporting a defensive role against herbivory [9]. Cannabinoid accumulation is also induced by ultraviolet-B exposure, suggesting an additional role in mitigating UV and oxidative stress, although the relative contributions of herbivore defense, antimicrobial protection, and abiotic-stress tolerance remain incompletely resolved [10]. Understanding how the plant regulates this defensive chemistry is therefore of both ecological and applied interest and motivates the focus of this review on the control of cannabinoid biosynthetic gene expression.
Cannabinoid biosynthesis converges on a single central precursor, cannabigerolic acid (CBGA), whose formation defines the metabolic commitment point for the entire pathway. CBGA is produced by condensation of olivetolic acid (OA) with geranyl pyrophosphate (GPP) by the prenyltransferase GOT at the surface of the secretory trichome [11]. OA itself is assembled from hexanoyl-CoA through a polyketide pathway requiring tetraketide synthase (TKS) and olivetolic acid cyclase (OAC) for correct aldol cyclisation; without OAC, TKS produces the shunt product olivetol rather than OA, demonstrating that OAC activity is the first rate-limiting checkpoint controlling carbon flux into the pathway [12]. GPP is supplied from the plastidic methylerythritol phosphate (MEP) pathway, and the intersection of two compartmentalized precursor streams at the trichome surface creates a metabolic bottleneck whose regulation is poorly understood [6]. CBGA is subsequently converted to cannabidiolic acid (CBDA) or Δ9-tetrahydrocannabinolic acid (THCA) by the substrate-competing enzymes CBDA synthase (CBDAS) and THCA synthase (THCAS), whose relative expression levels directly determine the CBD:THC chemotype ratio of a cultivar [13]. Both acids undergo spontaneous non-enzymatic decarboxylation to yield their pharmacologically active forms upon heating. The transcriptional output of four genes, OAC, OLS, CBDAS, and THCAS, controls cannabinoid identity and yield at every branch point of the pathway. What regulates the expression of these four genes at the chromatin, transcription factor, and post-translational levels remains the central unresolved question in Cannabis molecular biology.
Cannabinoid content varies by as much as 10-fold across cultivars, and at least part of this variation occurs among cultivars carrying functional, closely related CBDAS and THCAS alleles, a residual difference that coding-sequence variation among functional alleles does not fully explain [14]. Cannabinoid concentration increases more than five-fold between early and late floral stages, whereas transcript levels of core biosynthetic genes do not rise proportionally. This decoupling indicates that cannabinoid accumulation is shaped by regulation beyond transcript abundance, including post-transcriptional, translational, and enzyme-level control [3,6]. These observations require important genetic qualification, because several sequence-level factors beyond simple allelic identity are known to shape chemotype. The synthase genes reside within large, retrotransposon-rich, minimally recombining genomic regions that differ structurally between drug- and hemp-type alleles and contain one active enzyme alongside numerous pseudogenes [15]; copy-number variation at THCAS and CBDAS is common [16]; inactivating single-nucleotide polymorphisms, frameshifts, and pseudogenization determine whether a given synthase allele is functional and thereby set the qualitative drug-, hemp-, or CBG-dominant chemotype [17,18]; and the synthases are tightly linked to each other and to the aromatic prenyltransferase locus associated with total cannabinoid content [14,15]. What these structural, copy-number, and allelic-state determinants do not fully account for is the quantitative, several-fold variation in cannabinoid content among cultivars that share functional, near-identical synthase alleles, and it is this residual variation that the present review attributes in part to chromatin-level regulation. CBDAS and THCAS transcripts are essentially absent from leaves and stems but reach their highest levels in secretory glandular trichomes, implying that chromatin accessibility, not transcription factor availability, gatekeeps pathway expression in a cell-type-specific manner [19]. In Catharanthus roseus, jasmonate-responsive bHLH factors activate the alkaloid pathway promoters [20,21].
In Papaver somniferum, alkaloid biosynthesis is confined to specific cell types and cultivar-specific yield differences correlate with differential CHH methylation at biosynthetic loci [22,23,24], and in Artemisia annua, jasmonate-induced histone acetylation is implicated in activation of the artemisinin pathway [25,26]. In all three medicinal systems, a tissue-specific or environmentally responsive chromatin switch at rate-limiting pathway gene promoters controls metabolite yield independently of coding-sequence variation; Cannabis presents the same phenotypic profile, yet no equivalent chromatin-level analysis has been performed for any cannabinoid pathway gene.
In Cannabis glandular trichomes, the activating marks H3K4me3 and H3K56ac and the repressive mark H3K27me3 show reciprocal, tissue-specific distributions at cannabinoid and terpenoid loci [27]; micropropagation drives cultivar-specific accumulation of differentially methylated positions in promoter and intergenic regions [28]; and UV-B raises cannabinoid accumulation, although whether it acts through locus-specific methylation at pathway promoters is unresolved [10]. These lines of evidence, examined in detail in Section 3, leave open which enzymes write and erase these marks at pathway promoters and which signals recruit them to trichome chromatin.
Targeted epigenome editing with dCas9 fusion proteins has been demonstrated in plants with quantitative effects on pathway gene expression: dCas9-directed DNA demethylation activates ripening genes in tomato (Solanum lycopersicum) [29], dCas9-SunTag de novo methylation heritably silences target loci in Arabidopsis [30], and CRISPRa activates endogenous genes in Nicotiana benthamiana [31]. No equivalent experiment has been conducted in Cannabis: the chromatin states at OAC, OLS, CBDAS, and THCAS have not been functionally interrogated, and no dCas9-methyltransferase, dCas9-HAT, or CRISPRa approach has been applied to the cannabinoid pathway in any tissue [27].
This review therefore has a single aim: to assess whether chromatin-level regulation governs cannabinoid biosynthetic gene expression in glandular trichomes and whether it can be exploited to enhance cannabinoid yield. We map the histone modifications and DNA methylation patterns that demarcate tissue-specific cannabinoid gene expression, draw on functional evidence from model medicinal plants linking chromatin state to metabolite flux, evaluate which dCas9-effector architectures best fit each Cannabis chromatin target (Figure 1), and define the experimental priorities that must be met before epigenome editing can be deployed for cannabinoid crop improvement.
3. Epigenetic Regulation of Cannabinoid Biosynthesis in Cannabis sativa: Current Evidence
Five studies now define the empirical foundation of Cannabis epigenomics, spanning DNA methylation dynamics during micropropagation, the histone modification landscape of glandular trichomes, environmental epigenetic responses, and chromatin dynamics during in vitro organogenesis. Together, they establish that epigenetic mechanisms operate at every level of Cannabis biology relevant to cannabinoid yield. They also expose five specific mechanistic gaps whose resolution is required before epigenome editing can be rationally designed for cannabinoid enhancement (Figure 2; Table 1).
3.1. Progressive DNA Hypomethylation Accumulates During Micropropagation
Genome-wide 3D-GBS and enzymatic methyl sequencing of three Cannabis cultivars, including Critical Purple Kush, Green Crack, and Gelato, subcultured across 20 subcultures over 60 weeks, revealed that differentially methylated positions accumulate with subculture number in a cultivar-specific pattern, with the majority of mutations arising during culture initiation and the first five subcultures [28]. The rate and genomic distribution of epimutation differed significantly among cultivars, promoter-region DMPs reaching 22% in CPK, 9% in GC, and 13% in GEL, demonstrating that CG and CHG methylation maintenance fidelity varies with genotype under identical tissue culture conditions (Figure 2) [28]. Progressive DNA hypomethylation has also been documented in Cannabis shoots during in vitro subculture using MSAP-Seq, confirming that epigenetic drift is a reproducible outcome of in vitro propagation; however, the direct functional consequences for cannabinoid biosynthetic gene expression and yield have not been quantitatively established [28]. Whether the observed hypomethylation reflects impaired DRM2-mediated CHH re-establishment, reduced CMT3–KYP feedback loop activity, or enhanced ROS1-driven active demethylation at pathway loci has not been determined.
3.2. Trichome-Specific Histone Modification Switches Activate Cannabinoid Pathway Gene Loci
ChIP-seq and RNA-seq analysis of glandular trichome, leaf, and stem tissues from Cannabis sativa mapped the active and repressive histone modification landscape at cannabinoid and terpenoid biosynthetic gene loci across three tissue types, providing the first genome-wide histone modification maps for C. sativa glandular trichomes [27]. H3K4me3 and H3K56ac co-occupy the promoters and gene bodies of cannabinoid biosynthetic genes, including CBDAS, PT4/CBGAS, and terpenoid synthases, specifically in glandular trichome tissue but are absent at the same loci in leaf and stem, establishing that trichome-specific transcriptional activation is marked by a permissive chromatin state absent in vegetative tissues (Figure 2) [27]. H3K27me3 and H2A.Z show the reciprocal pattern, co-localizing in the gene bodies of silenced cannabinoid pathway genes in vegetative tissues and depleted in trichomes, indicating that the developmental transition to trichome identity involves a coordinated Polycomb-to-Trithorax switch at pathway gene promoters (Figure 2) [27]. Trichome-specific intergenic regions of high H3K56ac enrichment without corresponding H3K4me3 at transcription start sites identify candidate putative enhancer elements whose role in co-expression of cannabinoid and terpenoid pathway gene clusters has not been functionally tested [27]. Genes encoding starch and sucrose metabolism enzymes, which supply precursor pools for cannabinoid biosynthesis, also show trichome-specific H3K4me3 and H3K56ac enrichment; complementary ATAC-seq profiling further shows that chromatin accessibility at fatty acid biosynthesis and trichome initiation gene promoters, rather than at core cannabinoid synthase promoters, is the primary epigenetic driver of cannabinoid yield differences between cultivars [27,53]. The identities of the transcription factors that read the H3K4me3 marks at cannabinoid pathway gene promoters, and the upstream signals that recruit responsible Trithorax methyltransferases to these loci during trichome differentiation, remain unknown.
3.3. UV Irradiation Induces Genome-Wide DNA Hypermethylation and Cannabinoid Gene Upregulation in Cannabis indica
Exposure of Cannabis indica cell suspension cultures to UV radiation at a 4 h photoperiod increased genome-wide DNA methylation, as detected by methylation-sensitive amplified polymorphism analysis, and elevated CBDAS and THCAS transcript levels approximately 4-fold relative to non-irradiated controls, with progressive upregulation of OAC and OLS also observed (Figure 2). This result, observed in C. indica rather than C. sativa, demonstrates that environment-triggered DNA methylation changes are mechanistically coupled to cannabinoid pathway gene expression in the Cannabis genus, though the species distinction warrants caution in extrapolation. The MSAP method used detects methylation at CCGG sites genome-wide but cannot resolve whether the hypermethylation occurs at cannabinoid gene promoters directly or at flanking repressor loci whose silencing releases pathway gene expression. In Arabidopsis, the DNA demethylase ROS1 preferentially targets transposable element loci proximal to protein-coding genes, preventing methylation spreading into flanking genic regions; whether analogous UV-B-responsive ROS1-mediated demethylation at pathway gene-adjacent TE loci occurs in Cannabis is unknown [37]. Whole-genome bisulfite sequencing at single-base resolution in C. sativa tissues subjected to UV-B, drought, and heat stress is required to determine whether stress-responsive methylation changes at cannabinoid pathway gene promoters constitute a conserved environmental induction mechanism.
3.4. Chromatin Remodeling Governs the Embryogenic Competency Transition During Cannabis Organogenesis
Transcriptomic comparison of non-embryogenic and embryogenic Cannabis callus identified 247 differentially expressed epigenetic regulator genes spanning DNA methyltransferases, histone-modifying enzymes, chromatin remodeling complex subunits, and small RNA pathway components, indicating that the acquisition of embryogenic competency in Cannabis involves wholesale reprogramming of the chromatin regulatory machinery (Figure 2) [57]. HAT and HDAC activities are central to this transition: chromatin remodeling factors, including CHROMATIN REMODELING 35 (CRF35), mediate crosstalk between H3K27me3 and HDAC activity to repress embryogenesis-related transcription factors (LEC1, LEC2), and their upregulation in embryogenic calli correlates with the repression of somatic embryogenesis progression [57]. This mechanism is consistent with findings across multiple plant species, where HDAC inhibition with trichostatin A de-represses H3 acetylation at embryogenic loci and significantly enhances somatic embryogenesis frequency, including in Arabidopsis and Cannabis root explants [46,58]. The practical implication is direct: pharmacological or dCas9-HDAC-based modulation of histone deacetylase activity at embryogenic competency loci may improve the genotype-dependent transformation and regeneration efficiencies that currently constrain Cannabis biotechnology.
3.5. Five Mechanistic Gaps Define the Frontier for Cannabis Epigenome Editing
Five mechanistic gaps define the frontier for Cannabis epigenome editing, and their closure is required before dCas9-effector strategies can be rationally designed for cannabinoid yield enhancement. First, no comprehensive, whole-genome base-resolution methylome exists for any cannabinoid-relevant Cannabis tissue. The available base-resolution methylation data (enzymatic methyl sequencing of micropropagated shoots; [28]) derive from a reduced-representation 3D-GBS protocol that samples only a fraction of the genome and were generated in whole shoot tissue rather than glandular trichomes. Whole-genome bisulfite sequencing or EM-seq covering CG, CHG, and CHH contexts in glandular trichomes, leaf epidermis, and vascular tissue has not been published, and the RdDM pathway’s role at transposable element-proximal cannabinoid pathway genes remains entirely unexplored [32,35]. The existing trichome ChIP-seq dataset provides four histone modification maps but no DNA methylation data at single-base resolution, leaving the relationship between TE-flanking CHH methylation and H3K27me3-to-H3K4me3 switching at CBDAS, THCAS, and PT4/CBGAS loci unresolved [27]. Second, the histone modification landscape across trichome development remains incomplete. The existing dataset covers H3K4me3, H3K56ac, H3K27me3, and H2A.Z in mature trichome tissue only [27]; a developmental series additionally integrating H3K36me3, H3K9me2, H3K27ac, and H3K9ac from trichome progenitor cells through secretory stage maturation would identify the full complement of chromatin-state transitions required to define rational epigenome editing targets. Third, stress-responsive epigenomic data in C. sativa are entirely absent. How drought, heat, UV-B, and light intensity reprogram the methylome and histone landscape at cannabinoid pathway gene loci is unknown. The only available stress-responsive epigenetic data in the Cannabis genus comes from UV-irradiated C. indica cell suspension cultures, which show genome-wide MSAP-detected hypermethylation coincident with approximately 4-fold CBDAS and THCAS upregulation but cannot resolve locus-specific methylation changes [55]. Promoter characterization confirms that CsCBDAS and CsPT1 loci respond to IAA, GA3, SA, ABA, and abiotic stresses including heat, cold, light, salt, and drought, establishing that upstream signaling inputs converge on cannabinoid gene promoters; whether these signals act through chromatin-level mechanisms remains unresolved [59]. Fourth, transgenerational epigenetic inheritance in Cannabis has not been tested. Whether propagation-induced or environmentally induced methylation changes are reset between sexual generations or transmitted to progeny with consequences for cannabinoid profiles is unknown, and cultivar-specific differential methylated positions that accumulate during micropropagation have demonstrated potential phenotypic impact without confirmed heritability [28]. Fifth, the trichome chromatin-reading transcription factors that establish cannabinoid pathway gene locus identity remain unidentified. AP2/ERF, WRKY, bHLH, bZIP, MYB, and NAC family members modulate cannabinoid pathway expression through hormonal and elicitor signaling, but whether any of these interact specifically with H3K4me3-marked nucleosomes at CBDAS and THCAS promoters, or recruit Trithorax methyltransferases to these loci during trichome differentiation, has not been determined [60,61]. Each of these five gaps maps directly to a specific dCas9-effector targeting decision, covering methyltransferase, demethylase, HAT, HDAC, or chromatin remodeler deployment, and their experimental resolution defines the minimum knowledge base for rational Cannabis epigenome editing (Figure 2) [53,61].
4. CRISPR/dCas9-Based Epigenome Editing: Tools, Targets, and Implementation in Cannabis
No epigenome editing experiment has been conducted in Cannabis. The trichome ChIP-seq atlas, the micropropagation methylation data, and the UV-B induction findings reviewed in Section 3 (Figure 2) collectively define at least six specific chromatin targets whose manipulation could alter cannabinoid pathway gene expression. The dCas9-effector toolkit has been validated in tomato, Arabidopsis, and Nicotiana benthamiana at secondary metabolite pathway loci with quantitative outcomes, and its application to Cannabis is technically feasible given existing transformation protocols (Figure 3; Table 2) [62]. What is required is a systematic mapping of which effector architectures are most appropriate for each Cannabis target, followed by a phased experimental program beginning in hairy root cultures and progressing to stable trichome-targeted transformation.
4.1. dCas9-Effector Architecture Determines Editing Specificity and Mark Durability
A catalytically dead Cas9 (dCas9), carrying D10A and H840A substitutions that inactivate both RuvC and HNH nuclease domains, retains full sgRNA-directed genomic targeting but cannot cleave DNA, providing a programmable sequence-specific anchor for epigenetic effector recruitment (Figure 3) [63]. The sgRNA base-pairs with a 20-nucleotide protospacer upstream of a 5′-NGG-3′ PAM, and multiple sgRNAs can be deployed simultaneously to coordinate epigenetic modification across all four cannabinoid pathway genes in a single transformation event [64]. The field of plant dCas9-based chromatin regulation has advanced substantially since these foundational reports, with tools now available for targeted deposition and removal of DNA methylation, H3K4me3, H3K9me2, H3K27me3, and H3K27ac at endogenous plant loci [65,66]. The SunTag recruitment system extends single dCas9 binding events into cooperative effector deposition by recruiting up to 24 copies of a single-chain antibody-effector fusion through a repetitive GCN4 peptide array, achieving histone mark deposition levels unattainable by single dCas9-effector fusions at loci with low chromatin accessibility (Figure 3) [67]. In Arabidopsis, SunTag-dCas9-VP64 activated endogenous targets more than 100-fold compared to dCas9-VP64 alone, and SunTag-dCas9-JMJ13 achieved targeted H3K27me3 removal at the CUC3 locus with developmental consequences, establishing that the SunTag architecture is the most effective amplification platform for both mark deposition and erasure in plants [30,68]. A modular dCas9-based recruitment platform enabling combinatorial epigenome editing has further demonstrated that coordinated deposition of multiple marks at a single locus produces synergistic transcriptional outcomes unachievable by single-effector systems [69]. The reversibility of epigenetically modified states, unlike permanent DNA sequence changes, means that editing outcomes can decay after dCas9-effector expression ceases unless mark maintenance machinery reinforces them, a property that is both a design challenge for durability and an opportunity for temporal control of pathway output [70].
4.2. DNA Methylation Editing at Cannabinoid Pathway Gene Promoters
Targeted CG methylation writing using dCas9 fused to the DNMT3A-DNMT3L catalytic domain silences target gene promoters by recruiting methyl-CpG-binding domain proteins and occluding transcription factor binding sites; in Arabidopsis, SunTag-dCas9-DNMT3A targeting endogenous loci achieved robust on-target CG methylation deposition with minimal detectable off-target methylation by whole-genome bisulfite sequencing [30]. Applied to Cannabis, dCas9-DNMT3A targeting the THCAS promoter in chemotype II cultivars could suppress THC branch competition and redirect CBGA flux toward CBDA production without altering the THCAS coding sequence, preserving the genetic integrity of the cultivar (Figure 3). Targeted CG demethylation using dCas9 fused to the TET1 catalytic domain oxidizes 5-methylcytosine stepwise to 5-hydroxymethylcytosine, 5-formylcytosine, and 5-carboxylcytosine, resolved to unmethylated cytosine through base excision repair; in Arabidopsis, this SunTag-dCas9-TET1cd system achieved heritable reactivation of the silenced FWA locus and targeted demethylation of the CACTA1 transposable element with minimal effects on global methylation patterns (Figure 3) [71]. The progressive CHG and CHH hypomethylation documented during Cannabis micropropagation represents a near-term application for dCas9-DNMT3A: targeted re-methylation of differentially methylated positions at cannabinoid pathway gene-associated loci in advanced-passage cultures could stabilize the epigenotype and potentially recover propagation-induced reductions in cannabinoid yield, a hypothesis testable in existing micropropagation systems by transient protoplast delivery of dCas9-DNMT3A ribonucleoprotein complexes [28].
4.3. Histone Modification Editing at Trichome Pathway Gene Loci
The H3K27me3 marks enriched at CBDAS, THCAS, OLS, and OAC loci in vegetative Cannabis tissues are an experimentally documented chromatin feature and therefore a candidate target for de-repression by dCas9 fused to the KDM6A or KDM6B JmjC demethylase domain, which actively removes H3K27me3 and restores transcriptional competency at Polycomb-silenced loci. In Arabidopsis, dCas9-JMJ13 SunTag-mediated targeting of the CUC3 boundary gene locus achieved H3K27me3 reduction with measurable developmental consequences (Figure 3), establishing that single-locus H3K27me3 erasure is sufficient to override Polycomb repression at a developmentally regulated plant gene [72]. Conversely, dCas9 fused to the p300 HAT catalytic domain deposits H3K27ac at target promoters; in Arabidopsis, dCas9-p300 targeted to the AREB1 promoter increased H3K27ac enrichment and activated drought-responsive gene expression, demonstrating that H3K27 acetylation written by dCas9-HAT is a viable transcriptional activation strategy at endogenous plant loci [73]. In Cannabis, dCas9-p300 targeted to the CBDAS or OAC promoters in trichome progenitor cells could amplify cannabinoid pathway flux by converting H3K27me3-marked bivalent states to H3K27ac-marked active states during the developmental window of trichome differentiation. The CMT3-H3K9me2 self-reinforcing loop at heterochromatic loci means that for Cannabis pathway genes flanked by transposable element-derived H3K9me2 regions, dual targeting of both the DNA methylation and histone modification layers simultaneously may be required for durable state changes; systematic profiling of available dCas9-based epigenome editors confirms that coordinated multi-layer editing produces more stable transcriptional outcomes than single-effector deployment [70].
4.4. Multiplexed CRISPRa and CRISPRi for Coordinated Pathway Modulation
The dCas9-VPR tripartite activation domain, combining VP64, p65, and Rta transactivation domains, has activated endogenous secondary metabolite pathway genes at endogenous loci in Nicotiana benthamiana and tomato; multiplexed CRISPRa activation of three flavonoid pathway genes simultaneously produced specific and predictable metabolite profiles without detectable off-target transcriptional changes by RNA-seq, establishing that CRISPRa is transferable across Solanaceae and potentially across broader crop phylogenies [30,74]. A multiplexed CRISPRa strategy using sgRNAs simultaneously targeting the CBDAS, OAC, and OLS promoters could amplify transcriptional flux through the CBD branch of the cannabinoid pathway across all three rate-limiting steps in a single transformation event, an approach that has no equivalent in conventional overexpression strategies because it avoids the dosage and codon optimization problems of transgene-based pathway engineering (Figure 3). CRISPRi using dCas9 fused to three tandem repeats of the SRDX EAR-motif repressor domain has silenced endogenous Arabidopsis genes with high efficiency and represents the most straightforward approach to suppressing THCAS expression in CBD-producing chemotype III cultivars without introducing a THCAS knockout mutation [64]. Combined deployment of CRISPRa at CBDAS and CRISPRi at THCAS in a single chemotype II cultivar provides a quantitative test of whether the CBD:THC chemotype ratio is primarily determined by the relative transcriptional outputs of these two competing synthases, a mechanistic question that has not been resolved experimentally (Table 2).
4.5. Delivery, Durability, and Validation Challenges Specific to Cannabis
Agrobacterium-mediated stable transformation of Cannabis sativa remains genotype-dependent and highly variable in efficiency across commercially relevant cultivars, representing the primary bottleneck for constitutive dCas9-effector expression [56,75]. The large size of dCas9-effector fusions, typically 160 to 240 kDa for SunTag architectures, imposes constraints on T-DNA cassette design and transformation efficiency relative to standard Cas9 constructs, and trichome-targeted expression requires a trichome-specific promoter whose regulatory elements have not been fully characterized in Cannabis [27,59]. Transgene-free ribonucleoprotein delivery into protoplasts avoids these constraints and is the recommended first-line delivery method for Phase I proof-of-concept experiments, enabling transient epigenome editing validation before stable transformation is attempted. Mark durability across vegetative propagation cycles is the most significant unresolved technical challenge: epigenetic marks introduced by dCas9-effectors must persist across the subculture generations used in commercial Cannabis propagation, yet the endogenous CMT3 and MET1 maintenance machineries may dilute or reset them during each cell division [70]. Dual-effector systems targeting both DNA methylation and histone marks at the same locus, or inducible dCas9-effector expression systems reactivated at defined intervals during propagation, represent the two most tractable strategies for maintaining editing outcomes across extended vegetative cycles. Validation of epigenome editing outcomes requires four measurements at each target locus: targeted bisulfite sequencing or CUT&RUN to confirm mark deposition or erasure, RT-qPCR to verify transcript-level response, targeted metabolite profiling by HPLC or LC-MS/MS to confirm cannabinoid phenotype, and whole-genome methylation and histone profiling to quantify off-target effects (Figure 3).
4.6. A Phased Experimental Program for Cannabis Epigenome Editing
The first priority is functional validation of dCas9-effector constructs in Cannabis using hairy root cultures and transient protoplast transfection, which circumvent stable transformation recalcitrance and allow rapid iteration of sgRNA design and effector selection. Specific first-phase targets are dCas9-TET1 demethylation of CBDAS promoter CG sites verified by targeted bisulfite sequencing and RT-qPCR, CRISPRa targeting of OAC and OLS to test transcriptional amplification of upstream polyketide flux, and CRISPRi targeting of THCAS in chemotype II hairy root cultures to quantify competitive pathway suppression. Once functional constructs are validated in hairy root systems, the second priority is stable transformation of two chemotype backgrounds using trichome-specific promoters derived from THCAS or CsLTP1 regulatory sequences to drive trichome-targeted dCas9-effector expression, with methylation mark durability assessed by EM-seq at targeted loci across ten successive vegetative propagation cycles (Figure 3). The third priority addresses translational deployment: selection of stable epigenotype lines with verified cannabinoid profiles across developmental stages, assessment of epigenetic stability across sexual generations if breeding pipelines are used, and regulatory pathway evaluation in jurisdictions where non-mutagenic epigenome editing may qualify for reduced oversight relative to sequence-modifying technologies. Each phase generates specific quantitative outputs, methylation percentages, transcript fold-changes, cannabinoid yields, off-target profiling data, that define the go/no-go criteria for progression to the next phase.
| Effector (dCas9 Fusion) | Molecular Action | Model-System Precedent | Proposed Cannabis Target | Hypothesized Effect (Untested in Cannabis) |
|---|---|---|---|---|
| dCas9-DNMT3A/3L (SunTag) | Writes CG methylation; silences promoters | On-target CG methylation at Arabidopsis loci [30] | THCAS promoter (chemotype II) | Suppress THC branch; redirect CBGA flux to CBDA |
| dCas9-TET1 (SunTag) | Removes CG methylation (oxidation + BER) | Heritable reactivation of FWA; demethylation of CACTA1 in Arabidopsis [56] | Methylated pathway-gene promoters | Reactivate silenced biosynthetic loci |
| dCas9-p300 | Writes H3K27ac; activates transcription | dCas9-p300 at AREB1 activated drought genes in Arabidopsis [30] | CBDAS or OAC promoters (trichome progenitors) | Amplify cannabinoid pathway flux |
| dCas9-KDM6A/KDM6B (JMJ13) | Removes H3K27me3; relieves Polycomb repression | dCas9-JMJ13 reduced H3K27me3 at CUC3 in Arabidopsis [72] | H3K27me3-marked CBDAS, THCAS, OLS, OAC | De-repress Polycomb-silenced pathway genes |
| CRISPRa (dCas9-VPR/SunTag-VP64) | Recruits transactivators; upregulates genes | Multiplexed activation of flavonoid genes in Nicotiana/tomato [74] | CBDAS, OAC, OLS promoters (multiplexed) | Amplify flux through CBD branch at three steps |
| CRISPRi (dCas9-3xSRDX) | EAR-motif repression; silences genes | High-efficiency silencing in Arabidopsis [64] | THCAS (chemotype III) | Suppress THCAS without a knockout mutation |
5. Conclusions
Three bodies of evidence now anchor Cannabis epigenomics as a field with direct implications for cannabinoid yield engineering. H3K4me3 and H3K56ac co-occupy the promoters and gene bodies of cannabinoid biosynthetic genes exclusively in glandular trichome tissue, while H3K27me3 marks the same loci in leaf and stem, establishing that trichome-specific cannabinoid gene expression is controlled by a Polycomb-to-Trithorax chromatin switch rather than by transcription factor availability alone. Progressive CG and CHG differential methylation accumulate during micropropagation in a cultivar-specific pattern, with promoter-region changes reaching up to 22% by 20 subcultures, identifying DNA methylation maintenance as a candidate determinant of epigenetic stability in clonally propagated Cannabis material; whether this epigenetic drift alters cannabinoid output has not yet been tested against metabolite data. UV irradiation of Cannabis indica cell suspension cultures increases genome-wide DNA methylation and elevates CBDAS and THCAS transcript levels approximately 4-fold (Figure 2), demonstrating that environmentally triggered methylation changes are coupled to cannabinoid pathway gene expression in the Cannabis genus. These three findings define six specific chromatin targets for dCas9-effector intervention: the H3K27me3 repressive state at cannabinoid loci in non-trichome tissues, addressable by dCas9-JMJ13 or dCas9-KDM6A; the H3K27me3-to-H3K4me3 transition in trichome progenitors, addressable by dCas9-p300; the CG hypomethylation accumulating at pathway gene-associated loci during micropropagation, addressable by dCas9-DNMT3A; the competitive THCAS transcriptional activity in CBD-producing chemotypes, suppressible by dCas9-3xSRDX CRISPRi; the upstream OAC and OLS transcriptional bottleneck, amplifiable by dCas9-VPR or dCas9-SunTag CRISPRa; and the stress-responsive methylation-cannabinoid induction axis, exploitable through dCas9-DRM2 targeting at UV-responsive loci (Table 2). Each target is supported by experimental evidence from at least one plant model system; none has been tested in Cannabis. Five knowledge gaps must be addressed before these strategies can be rationally deployed. No comprehensive, whole-genome base-resolution methylome exists for glandular trichomes or other cannabinoid-relevant Cannabis tissues; the existing base-resolution data come from reduced-representation enzymatic methyl sequencing of micropropagated shoots and do not provide whole-genome coverage of the relevant cell types. The trichome histone modification atlas covers only four marks in mature tissue and lacks a developmental series from progenitor to secretory stage. No stress-responsive epigenomic dataset exists for C. sativa at locus resolution. The transcription factors reading H3K4me3 at cannabinoid pathway gene promoters in trichomes remain unidentified. Whether propagation-induced methylation changes persist across sexual generations or are reset in progeny has not been determined. Each gap maps directly to a tractable experimental design: EM-seq in trichome-enriched tissue, developmental CUT&RUN from progenitor to mature secretory trichome, locus-resolved bisulfite sequencing under abiotic stress, chromatin immunoprecipitation with MYB, AP2/ERF, and WRKY antibodies at cannabinoid pathway gene promoters, and methylation profiling across F1 progeny of hypomethylated propagation lines. C. sativa epigenome editing does not require resolution of all five gaps before the first experiment is attempted. Functional validation of dCas9-SunTag-JMJ13 at the CBDAS locus in hairy root cultures, confirmed by CUT&RUN and RT-qPCR, is achievable with existing tools and existing Cannabis transformation protocols. That single experiment would establish whether H3K27me3 erasure at a cannabinoid pathway gene promoter is sufficient to elevate transcript levels in a non-trichome tissue context, providing the first direct causal evidence linking a specific chromatin mark to cannabinoid pathway gene expression in Cannabis. This is the experiment the field needs next.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
Conflicts of Interest
The authors declare no conflicts of interest.
Abbreviations
3D-GBS: three-dimensional genotyping by sequencing; AaHAT: Artemisia annua histone acetyltransferase; ABA: abscisic acid; ADS: amorpha-4,11-diene synthase; AP2/ERF: APETALA2/Ethylene Response Factor; ATAC-seq: assay for transposase-accessible chromatin with sequencing; bHLH: basic helix-loop-helix; BRM: Brahma; bZIP: basic leucine zipper; Cas9: CRISPR-associated protein 9; CBDA: cannabidiolic acid; CBDAS: cannabidiolic acid synthase; CBGA: cannabigerolic acid; CBD: cannabidiol; CG: cytosine–guanine methylation context; CHD: chromodomain helicase DNA-binding; CHG: cytosine–(any base except G)–guanine methylation context; CHH: cytosine–(any base except G)–(any base except G) methylation context; ChIP-seq: chromatin immunoprecipitation sequencing; CLF: Curly Leaf; CMT3: Chromomethylase 3; CRISPR: clustered regularly interspaced short palindromic repeats; CRISPRa: CRISPR activation; CRISPRi: CRISPR interference; CRF35: Chromatin Remodeling Factor 35; CsHDA1/5/7: Cannabis sativa Histone Deacetylase 1/5/7; CsSRT1: Cannabis sativa Sirtuin 1; CYP71AV1: cytochrome P450 71AV1; CYP80B1: cytochrome P450 80B1; dCas9: catalytically dead Cas9; DBR2: double bond reductase 2; DME: Demeter; DMPs: differentially methylated positions; DNMT3A: DNA methyltransferase 3A; DRM2: Domains Rearranged Methyltransferase 2; ELF6: Early Flowering 6; EM-seq: enzymatic methyl sequencing; EZH2: Enhancer of Zeste Homolog 2; GA3: gibberellin A3; GCN5/HAG1: General Control Non-repressible 5/Histone Acetyltransferase of the GNAT family 1; GOT: geranylpyrophosphate:olivetolate geranyltransferase; GPP: geranyl pyrophosphate; H2A.Z: histone H2A variant Z; H3K4me3: histone H3 lysine 4 trimethylation; H3K9ac: histone H3 lysine 9 acetylation; H3K9me2: histone H3 lysine 9 dimethylation; H3K27ac: histone H3 lysine 27 acetylation; H3K27me3: histone H3 lysine 27 trimethylation; H3K36me3: histone H3 lysine 36 trimethylation; H3K56ac: histone H3 lysine 56 acetylation; HAT: histone acetyltransferase; HDA6: Histone Deacetylase 6; HDA19: Histone Deacetylase 19; HDAC: histone deacetylase; HPLC: high-performance liquid chromatography; IAA: indole-3-acetic acid; JAZ: Jasmonate ZIM-domain repressor; KDM6A/UTX: Lysine Demethylase 6A; KYP/SUVH4: Kryptonite/Suppressor of Variegation Homolog 4; LC-MS/MS: liquid chromatography–tandem mass spectrometry; LEC1/LEC2: Leafy Cotyledon 1/2; MEA: Medea; MeJA: methyl jasmonate; MEP: methylerythritol phosphate; MET1: DNA Methyltransferase 1; MIA: monoterpene indole alkaloid; MSAP: methylation-sensitive amplified polymorphism; MYB: Myeloblastosis transcription factor; NAC: NAM/ATAF/CUC domain transcription factor; nt: nucleotide; OA: olivetolic acid; OAC: olivetolic acid cyclase; OLS: olivetol synthase; ORCA: octadecanoid-responsive Catharanthus AP2-domain transcription factor; p300: histone acetyltransferase p300; PAM: protospacer adjacent motif; PKL: Pickle; PRC2: Polycomb Repressive Complex 2; PT4/CBGAS: prenyltransferase 4/cannabigerolic acid synthase; RdDM: RNA-directed DNA methylation; REF6: Relative of Early Flowering 6; RNA-seq: RNA sequencing; ROS1: Repressor of Silencing 1; RT-qPCR: reverse-transcription quantitative polymerase chain reaction; SA: salicylic acid; SAM: Synergistic Activation Mediator; SDG8/ASHH2: SET Domain Group 8/Ash1 Homolog 2; sgRNA: single guide RNA; siRNA: small interfering RNA; SRDX: Super-repressor Domain X; STR: strictosidine synthase; SunTag: SuperNova Tagging system; SWI/SNF: SWItch/Sucrose Non-Fermentable; SWN: Swinger; T-DNA: transfer DNA; TDC: tryptophan decarboxylase; TE: transposable element; TET1: Ten-eleven translocation methylcytosine dioxygenase 1; THCA: Δ9-tetrahydrocannabinolic acid; THCAS: tetrahydrocannabinolic acid synthase; THC: Δ9-tetrahydrocannabinol; TKS: tetraketide synthase; TrxG: Trithorax group complex; TSA: trichostatin A; T6ODM: thebaine 6-O-demethylase; VPR: VP64-p65-Rta; WRKY: WRKY domain transcription factor.
| Epigenetic Mechanism | System | Target Genes/Locus | Key Finding | Reference |
|---|---|---|---|---|
| DNA methylation (CHH) | Papaver somniferum | Alkaloid biosynthetic loci | Cultivar-specific differences in alkaloid yield correlate with differential CHH methylation | [23,24] |
| Histone acetylation | Artemisia annua | Artemisinin pathway | Jasmonate-induced histone acetylation implicated in pathway activation | [25,26] |
| TF/chromatin control | Catharanthus roseus | STR, TDC promoters | Jasmonate-responsive bHLH factors activate alkaloid promoters via JAZ derepression | [20] |
| Histone modification (H3K4me3, H3K56ac vs. H3K27me3) | Cannabis sativa (trichome vs. leaf/stem) | CBDAS, PT4/CBGAS, terpenoid synthases | Activating marks occupy pathway promoters in trichomes but are absent in leaf/stem; reciprocal H3K27me3 in vegetative tissue | [27] |
| DNA methylation (CG, CHG) | Cannabis sativa (micropropagation) | Promoter and intergenic regions | Cultivar-specific accumulation of differentially methylated positions over 20 subcultures; yield effect not yet established | [28] |
| DNA methylation (genome-wide, MSAP) | Cannabis indica (cell suspension, UV) | Genome-wide; locus not resolved | UV increases genome-wide methylation and elevates CBDAS/THCAS transcripts; promoter-specific methylation not demonstrated | [55] |