Hormonal control of promoter activities of Cannabis sativa prenyltransferase 1 and 4 and salicylic acid mediated regulation of cannabinoid biosynthesis

In silico analysis of CsPT1 and CsPT4 promoters

As shown in Fig. 1b,c, the 600 bp region of CsPT4 pro we cloned was sufficient to drive strong expression of coding sequences cloned downstream of this promoter region. Analysis using PLACE and PlantCARE, the online query software for cis-acting regulatory element analyses, revealed many putative responsive elements in the cloned promoter fragments of CsPT1 and CsPT4. These elements are conserved sequences which have been demonstrated, in other organisms, to bind to transcription factors related to various environmental cues such as stress, wounding, drought, and hormone signaling²⁰.

The hormone responsive elements for both CsPT1 pro and CsPT4 pro are listed in Table ​Table1,1, and contextually visualized in Supplementary Figs. ³ and ⁴. The in silico analysis showed that in CsPT4 pro, there are three responsive elements for cytokinin, one for SA, one for GA, one for ethylene, and one for ABA. The analysis also revealed that in CsPT1 pro, there are five GA responsive elements, two SA, six auxin, and three ABA responsive elements. MeJA signaling is vital for glandular trichome formation and specialized metabolite biosynthesis^21–26. However, no JA responsive elements were found in either promoter. It is possible that JA may not activate CsPT1 and CsPT4.

CsPT4 pro exhibits a strong response towards various hormones

To understand hormonal regulation of the CsPT genes, Dual Luciferase (LUC) Reporter Assay (DLR) was performed in Arabidopsis protoplasts. The same promoter sequences used for the GUS assay were cloned to create CsPT pro:LUC constructs (Supplementary Fig. ¹). The CsPT4 pro that was cloned contains three putative cytokinin responsive elements; CsPT4 pro was predicted to have no auxin responsive elements. Therefore, NAA was used as a negative control. The activities of LUC driven by the promoters were measured after hormone treatment using luminescence.

At 3 and 5 h post treatment (hpt) of ZR, CsPT4 pro activated the luciferase reporter at significant levels (Fig. 2a). This result is consistent with our aforementioned in silico analysis that cytokinin may regulate CsPT4 expression. Because there are no predicted auxin responsive elements in CsPT4 pro, it is expected that there would be no increased luciferase activity upon auxin treatment. Following application of NAA, there was no increased activity measured from CsPT4 pro, indicating that auxin may not modulate CsPT4 transcription in Arabidopsis protoplasts (Fig. 2b).

In addition, the 600 bp CsPT4 promoter region contains one ethylene, one ABA, and one GA putative responsive element. At 5 hpt with ACC (causing ethylene rise), CsPT4 pro was able to be activated at significant levels (Fig. 3a). In response to ABA and GA₃, CsPT4 pro was activated at 3 and maintained at 5 hpt (Fig. 3b,c). These three hormones had similar levels of CsPT4 pro activation. The results indicate that CsPT4 can be involved in different hormonal signaling pathways; this supports the hypothesis that hormones may coordinate to control glandular trichome development and cannabinoid biosynthesis in cannabis female flowers^27–30.

Auxin and ABA activate CsPT1 pro

The 1500 bp region of CsPT1 pro contains three auxin, three ABA, and five GA putative responsive elements. Following treatment from NAA, CsPT1 pro was activated significantly as early as 1 hpt. The activity was slightly higher at 3 hpt and retained at 5 hpt (Fig. 4a). ABA was able to induce significant activation of CsPT1 pro only at 3 hpt and the activity returned to the basal level 5 hpt (Fig. 4b). Treatment with GA₃ prompted slight increases in relative luminescence, however there was no significant difference compared to the solvent control at each time point (Fig. 4c).

MeJA is a major hormone involved in plant defense responses to stress stimuli^31–33. In a previous study, we showed that MeJA foliar spray increased cannabinoid production¹². Although no MeJA responsive elements were found in CsPT1 pro, the response of CsPT1 pro towards MeJA was still examined. Interestingly, there was slight decrease of luminescence 3 hpt and then slight increase of luminescence 5 hpt (Fig. 4d). However, there was no significant difference observed at either time point. MeJA could regulate other cannabinoid biosynthetic genes or promote glandular trichome morphogenesis, resulting in higher cannabinoid accumulation.

CsPT1 pro and CsPT4 pro are responsive to SA

CsPT1 pro and CsPT4 pro were examined collectively for their response to SA. Both promoters were activated after treatment with SA (Fig. 5). CsPT1 pro was activated sooner at 1 hpt and the activity was retained at similar levels from 1 to 5 hpt (Fig. 5a). CsPT4 pro was not activated until 3 hpt with a slight decrease at 5 hpt (Fig. 5b). Intriguingly, the increased activity of CsPT4 pro was substantially higher than that of CsPT1 pro despite it containing one more SA Responsive Element (SARE), suggesting a greater demand of CsPT4 transcripts after SA elicitation.

To further understand whether the identified putative SARE in CsPT4 pro is truly responsive to SA, site-directed mutagenesis was performed to introduce mutations in the sequence. Four base pairs were mutated to diminish or eliminate responses of CsPT4 pro to SA elicitation (Fig. 5c and Supplementary Fig. ⁵). The results revealed that the mutated CsPT4 pro was no longer responsive to SA elicitation (Fig. 5c), indicating that the identified SARE in CsPT4 pro is a bona fide cis-acting element involved in SA signaling.

Hormone regulation of CsPT promoters

Specialized metabolite biosynthesis is regulated by plant hormones^26,36. Results from this research demonstrated that CsPT4 pro was responsive towards cytokinin, ABA, ethylene, SA, and GA, while CsPT1 pro was only activated following treatment with auxin, ABA, and SA (Figs. ​(Figs.2,2, ​,3,3, ​,4,4, ​,5),5), providing hints on the association between hormone signaling and cannabinoid biosynthesis.

ABA could activate both promoters in our study (Figs. 3b, ​b,4b).4b). ABA plays essential roles in the regulation of floral development and plant responses to abiotic stress³⁷. Our previous study showed that both CsPT1 and CsPT4 were upregulated during female flower maturation and associated with glandular trichome development¹². Previous studies demonstrated that ABA alone did not affect flower sex determination, however, ABA inhibited indole-3-acetic acid (IAA) induced female flower formation and impaired GA induced male flower development^38,39. ABA treatment increased THC concentration in both leaf and female flower tissues³⁰ and has been shown to promote trichome formation⁴⁰. The activation of the promoters of CsPT1 and CsPT4 by ABA suggests that ABA may also play important roles in glandular trichome development and cannabinoid biosynthesis during female flower maturation in cannabis plants.

The promoter of CsPT4 was also activated by cytokinin, ethylene, and GA but not auxin (Figs. 2, ​,3).3). These hormones have been shown to be important for sex determination during floral development in dioecious or gynoecious plants^41–44. Cytokinin production in roots was shown to be essential for female sex determination in hemp plants generated from shoot cuttings of young plants⁴⁵. Ethylene was shown to promote female flower development in industrial hemp^38,46–48. Exogenous GA application induced male flower formation on female cannabis plants, which was inhibited by ABA³⁹. GA signaling associated genes were also shown to be upregulated in induced-male (on female flowers) and male flowers compared to female flowers⁴⁴.

In our study, NAA did not activate CsPT4 pro (Fig. 4a), even though auxin (IAA) was shown to be involved in the induction of female flowers³⁸. Promoters of CsPT1 and CsPT4 were activated by hormones important for both female and male flower induction, indicating that CsPT1 and CsPT4 could be active in both female and male flowers. These findings also imply that cannabinoid biosynthesis involves the interaction of different hormonal pathways that are essential for female flower induction and development. Monitoring of hormone content during cannabis flower development will enable us to further understand how these hormones function in cannabinoid or other specialized metabolite biosynthesis.

The DLR assays for hormone regulation of CsPT1 pro and CsPT4 pro were conducted using Arabidopsis protoplasts. Although many transcription factors are conserved among different plant species, there are also many with distinct functions in certain species. The results we observed in Arabidopsis protoplasts may be inconsistent or even contradictory. It would be interesting and important to further investigate the functions of these hormones in cannabis plants. Our findings still provide insights and clues to the hormonal modulation of glandular trichome development and cannabinoid biosynthesis in cannabis female flowers.

Cloning of promoters and in silico promoter analysis

Promoters of Cannabis sativa CsPT1 and CsPT4 were identified using the reference genome information on NCBI. For CsPT4 pro, we were able to clone 600 bp of the promoter, while we were able to clone 1500 bp of the CsPT1 pro. The promoters were PCR amplified using iProof High Fidelity Polymerase (Bio-Rad, CA, USA). CsPT1 pro and CsPT4 pro were inserted into an entry vector pDONR221, and then cloned into a destination vector, either P2GWL7 (for luciferase assay) or pBGWFS7 (for GUS assay). The orientation of insertion is shown in Supplementary Fig. ¹.

After the final insertion, plasmid maxiprep was conducted, with Macherey Nagel’s NucleoBond Xtra Maxi EF kit (Duren, Germany). CsPT1 pro and CsPT4 pro were analyzed using plant cis-acting regulatory DNA element databases, PLACE and PlantCARE. The predicted hormone responsive elements of CsPT4 pro and CsPT1 pro are shown in Supplementary Figs. ³ and ⁴ respectively.

GUS assay

Cannabis seed sterilization and seedling growth was conducted according to Sorokin et al.⁶⁷. After germination, seedlings were placed on full MS (pH 5.6) plates for 10 more days. Then the seedlings were used for transient expression. Leaves were taken from flowering Auto Tsunami plants, an auto-flowering variety, grown in a growth chamber with LED lights of 16 h light/8 h dark at 26 °C. Agrobacterium mediated transient gene expression in cannabis seedlings and leaves followed the method published by Deguchi et al.⁶⁸. The Agrobacterium strain used was GV3101. GUS staining was performed according to the protocol in Arabidopsis: A Laboratory Manual⁶⁹. GUS-stained seedlings and leaves were recorded using a Cannon digital camera.

Hormone treatment of protoplasts and real-time quantitative PCR (qPCR)

Arabidopsis leaves were collected from wt plants of 3–4 weeks old. Two protocols were followed for protoplast isolation. The tape-sandwich method was followed directly to separate the lower epidermal layer of the leaves, exposing the mesophyll cells⁷⁰. The remaining leaves were used to enzymatically isolate the protoplasts following a protocol published by Yoo et al.⁷¹.

Hormones were administered to Arabidopsis protoplasts at concentrations of 5, 10, and 50 µM, to determine the concentration that would be used in the dual luciferase reporter assay (DLR). These concentrations were shown to cause a response in Arabidopsis⁷². Seven hormones were administered to protoplasts, obtained from Sigma-Aldrich (MO, USA): cytokinin (trans-zeatin riboside, ZR), auxin (1-naphthaleneacetic acid, NAA), gibberellic acid 3 (GA)_, salicylic acid (SA), abscisic acid (ABA), ethylene (1-aminocyclopropane-1-carboxylic acid, an ethylene precursor, ACC), and methyl jasmonate (MeJA).

Representative hormone responsive genes were chosen for qPCR (Supplementary Table ¹) to ensure the hormones were activating the signaling pathways required for DLR assays^72–75. Protoplasts were treated with various concentrations of hormones, then collected after 4 h of shaking at 40 rpm in the dark, for RNA isolation and qPCR analysis.

Total RNA was isolated using TriReagent (Molecular Research Center, OH, USA) following the manufacturer’s protocol. cDNA was synthesized using iScript™ Reverse Transcription Supermix (Bio-Rad). qPCR was performed using iTaq Universal SYBR Green Supermix (Bio-Rad) in a CFX Connect system, with AtGAPDH as the housekeeping gene. qPCR program used: 95 °C for 3 min, followed by 39 cycles of 95 °C for 10 s, and 55 °C for 30 s. Primers are listed in Supplementary Table ². The hormone concentration which caused the highest fold increase in the gene, relative to the control, was chosen for further promoter analysis and DLR (Supplementary Table ³).

Dual luciferase reporter assay

Dual Luciferase Assays were conducted in which CsPT1 pro and CsPT4 pro were cloned upstream of the firefly LUC gene; CaMV 35S promoter driven Renilla LUC was used as control. Arabidopsis protoplasts were transfected with both constructs (10 μg firefly LUC plasmid, 10 ng Renilla LUC plasmid⁷⁶). Relative luminescence is calculated by dividing the recorded firefly luminescence by the Renilla luminescence. An empty P2GWL7 plasmid was also transfected as a negative control, which ensured there was no background noise.

Following overnight transfection, the protoplasts were treated with hormones. CsPT4 pro received 10 µM ZR, 10 µM ACC, 50 µM SA, 50 µM GA, 10 µM NAA and 10 µM ABA. CsPT1 pro received 50 µM GA, 50 µM SA, 10 µM NAA, 10 µM MeJA and 10 µM ABA. Samples were collected at 1, 3, and 5 h post treatment (hpt). ZR, SA, GA, and ABA were dissolved in DMSO, while ACC, NAA, and MeJA were dissolved in DI water. The control samples were administered equal proportions of the solvent.

DLR reporter assay system was used (Promega, WI, USA) with the only alteration being the lysis buffer. Cell culture lysis reagent (Promega) was used instead of passive lysis buffer. A CLARIOstar Plus plate reader (BMG Labtech, Germany) was used. Statistical analysis was determined using a one tailed Student t-test. The experiment was repeated three times.

Site-directed mutagenesis

Site directed mutagenesis was conducted on the SA responsive element (SARE) in the cloned region of the CsPT4 pro. Q5 site directed mutagenesis kit (NEB, MA, USA) was used in which point mutations were made to the sequence of the putative SARE. See Supplementary Table ² for primer design and Supplementary Fig. ⁵ for sequence comparison. Transfection was repeated with the mutated promoter, and luminescence was measured following SA treatment as mentioned above.

RNA isolation from cannabis tissues and qPCR

Approximately 250 mg of tissue was collected from the cannabis plants. Samples from GG included root, leaf, petiole, and flower. Samples from Wife were flowers. Tissues were flash frozen with liquid nitrogen and stored at − 80 °C. Total RNA was isolated from samples using Plant and Fungi RNA Extraction Kit (Macherey Nagel, Duren, Germany). Single-stranded cDNA was synthesized as described above.

Primers for qPCR are listed in Supplementary Table ². CsPP2A was used as the housekeeping gene for Wife samples, with 4 biological replicates for each treatment. CsUbiquitin was used for GG samples, with 3 biological replicates for each organ from the plant. qPCR was conducted as stated previously, with same polymerase and program. Samples were analyzed using the delta delta ct method and an ANOVA was conducted using a paired Student’s t-test.

Root Drench: SA was dissolved in 1 mL of DMSO and further diluted in 1 L of DI water, to obtain a 1 mM concentration. There were 4 biological replicates per treatment group. DMSO (control) or SA was administered to the plants as a soil drench during the fourth week of flowering; treatment methods from other studies were followed^60,61. Flower samples were collected at hour 0, 3, 6, 12, 24, 48, and 72 h post treatment. Approximately 250 mg fresh flower samples were collected and flash frozen for qPCR assay.

MeSA treatment and HPLC analysis

Floral Spray: MeSA (Sigma-Aldrich) was dissolved in 1 mL of DMSO, then diluted in 1 L DI water to create a concentration of 100 µM. DMSO or MeSA was applied by foliar spray until flowers were fully soaked; treatment methods from other studies followed^77,78. The spray was applied during the fourth week of flowering, as a one-time treatment. Flower samples were collected from plants prior to treatment and two weeks post treatment. Cannabinoids were sampled by milling the entirety of each plant’s flowers into a composite mixture. Approximately 500 mg of flower tissue was used for cannabinoid extraction using 20 mL 9:1 methanol:dichloromethane (DCM). Samples were shaken for 1.5 h, then discarded. Samples were then diluted with 100 µL extraction to 900 µL 9:1 methanol:DCM.

Cannabinoid concentration was analyzed on a Shimadzu High Pressure Liquid Chromatograph (HPLC) Instrument (LC-10). The column used was a Raptor Arc C18 (150 × 4.6 mm, 2.7 μm particle size) (Restek, PA, USA). The HPLC method followed Shimadzu’s protocol⁷⁹. Peaks on the chromatogram were identified using cannabinoids standards (Restek, PA).

We thank Frederick Pettit and Shelly Durocher for their assistant in maintaining cannabis and tobacco plants in the greenhouse. We thank Dr. Huanzhong Wang for providing us the p2GWL7 and pBGWFS7 vectors. This work was supported by Fine Fettle LLC. This work was supported in part by Olds Softgels Inc. This work is supported by Foundational Knowledge of Plant Products 2022-67014-37049 from the USDA National Institute of Food and Agriculture (YM and GAB), and Hatch project CONS01027 (GAB).