Introduction

The medicinal plant Catharanthus roseus, commonly known as the Madagascar periwinkle, produces the life-saving anticancer medicines vinblastine and vincristine. Vinblastine and vincristine are produced in extremely low concentrations in C. roseus (0.0002–0.0005% by weight), contributing to a shortage of these critical drugs (Noble 1990; Pan et al. 2016b; Rabin 2019; Shuman et al. 2020). Vinblastine and vincristine are produced from the terpenoid indole alkaloid (TIA) pathway, a tightly controlled pathway comprised of over 30 steps (Kulagina et al. 2022). TIA biosynthesis is initiated with the condensation of the terpenoid and indole branches to form strictosidine (Fig. 1). Strictosidine is enzymatically converted to produce over 130 different TIAs (van der Heijden et al. 2004).

Fig. 1
figure 1

Overview of the C. roseus terpenoid indole alkaloid (TIA) biosynthetic pathway. The TIA biosynthetic pathway is illustrated with major metabolites named. Dashed lines indicate multistep conversions while solid lines indicate direct conversions. Select enzymes are noted along arrows in blue (color figure online)

Full size image

While the enzymatic steps of the pathway have been elucidated, an understanding of its regulation is still emerging (Colinas et al. 2021; Liu et al. 2021). When C. roseus is subjected to herbivory, the jasmonate-regulated defense pathway is activated, initiating a cascade of both activating and repressing transcription factors, contributing to the tight regulation of TIA biosynthesis (Goklany et al. 2013; Peebles et al. 2009a; Rizvi et al. 2016). Here, we characterize the Zinc finger Catharanthus Transcription factors (ZCTs), a class of putative transcriptional repressors in C. roseus. The expression of ZCT1, ZCT2, and ZCT3 is elicited by methyl jasmonate (Chebbi et al. 2014; Goklany et al. 2013; Mortensen et al. 2019b; Pauw et al. 2004), yeast extract (Pauw et al. 2004), or auxin (Mortensen et al. 2019b). Structurally, ZCTs contain two characteristic Cys2-His2 zinc fingers and the EAR motif, one of the most common repressor domains involved in plant transcriptional regulation (Chow et al. 2023; Ciftci-Yilmaz and Mittler 2008; Liu et al. 2022; Yang et al. 2018).

Previous studies showed that ZCT1, ZCT2, and ZCT3 may play a repressive role in TIA biosynthesis. Using transient assays, the overexpression of ZCT1, ZCT2, and ZCT3 in cell suspension cultures of C. roseus resulted in the repression of the upstream TDC and STR promoters (Pauw et al. 2004). ZCT1 and ZCT2 were also identified as repressors of the HDS promoter, a gene upstream of the terpenoid branch (Chebbi et al. 2014). Further evidence included negative correlations between ZCT expression and TIA production. For instance, elicitation of the TIA pathway with methyl jasmonate led to an increase in the expression of both activators such as ORCA2 and ORCA3 and repressors such as ZCT1, ZCT2, and ZCT3 (Goklany et al. 2013); at moderate levels of methyl jasmonate, the ratio of ORCA to ZCT levels was high and TIA production was optimized; but at high levels of methyl jasmonate, the ratio of ZCT to ORCA levels was high and TIA production was inhibited (Goklany et al. 2013). Similarly, the overexpression of ORCA3 in C. roseus hairy root cultures led to increased ZCT levels and an unexpected decrease in alkaloid production (Peebles et al. 2009a).

In this paper, we identified three additional Cys2-His2 zinc finger transcription factors (ZCT4, ZCT5, and ZCT6) in the ZCT subgroup that includes the previously identified ZCT1, ZCT2, and ZCT3. Since ZCT1, ZCT2, and ZCT3 are putative repressors of TIA biosynthesis, we characterized the activity of ZCT4, ZCT5, and ZCT6 to evaluate if these six ZCTs could act redundantly or compensate as repressors of the TIA pathway. We also investigated if these six ZCTs differed in their induction by characterizing their tissue-specific expression and responsiveness to methyl jasmonate. Since ZCTs are induced by methyl jasmonate, we investigated the role of the jasmonate-associated CrMYC2a in regulating the expression of ZCTs. Our characterization suggests that ZCTs can be induced by CrMYC2a and that they can act as regulators of TIA metabolism, potentially as an adaptative response to jasmonate elicitation.

Materials and methods

Identification of ZCT4, ZCT5, and ZCT6

Coding sequences of ZCT1, ZCT2, and ZCT3 were aligned with blastn against the C. roseus transcriptome (Góngora-Castillo et al. 2012). Hits with E values < 1E-4 were considered potentially homologous ZCTs. This led to the discovery of three new ZCTs, ZCT4, ZCT5, and ZCT6. The Cra_locus_86538 was not further characterized (Figures S1) since it did not pass the threshold of an E value < 1E−4; in addition, based on an amino acid alignment, it contains only one rather than two Cys2-His2 zinc finger domains as found in the ZCTs.

Cladogram of Cys2-His2 transcription factors in C. roseus

We also searched for additional Cys2-His2 zinc finger transcription factors in the C. roseus genome using the TFDB search tool in the Plant Transcription Factor Database (v3.0) (Jin et al. 2014). Initially, 127 Cys2-His2 zinc fingers were identified. However, correcting for isoforms yielded 65 independent protein sequences identified by the Plant TFDB; a cladogram of all 65 Cys2-His2 transcription factors encoded in the C. roseus transcriptome (Góngora-Castillo et al. 2012) was generated with Qiagen CLC Main Workbench 8 (Figures S1; protein alignment parameters: gap open cost = 10.0; gap extension cost = 1.0; end gap cost = free; alignment mode = very accurate, redo alignments = no; use fixpoints = no). The Neighbor Joining algorithm and Jukes-Cantor distance measure were used with the Bootstrap method (100 replicates) to generate the tree as described in “Identification of Cys2-His2 zinc finger transcription factors in C. roseus” section.

Expression analysis of ZCTs

RNA-Seq reads were downloaded from the NCBI-SRA databases (BioProject: PRJNA252611) and processed to obtain sequencing reads in FASTQ format. Reads were processed to trim the sequencing adapter sequences and read position with low sequencing quality using Trimmomatic version 0.36 (Bolger et al. 2014). Clean reads were aligned to the C. roseus reference genome v2 using STAR version 2.5.2b (Dobin et al. 2013). Aligned sequences were counted using the featureCount version 1.22 package (Liao et al. 2014) in R to get the number of reads aligned to a gene feature. Read counts data were normalized to get gene expression values in FPKM (fragments per kilo base million). FPKM values for TIA pathway genes were extracted from the following CRO numbers: MYC2: CRO_T124533, ORCA2: CRO_T110365, ORCA3: CRO_T110360, BIS1: CRO_T107535, BIS2: CRO_T107539, ZCT1: CRO_T105646_ZCT1, ZCT2: CRO_T114616, ZCT3: CRO_T130011, ZCT4: CRO_T105646_ZCT4, ZCT5: CRO_T124775, ZCT6: CRO_T110996, IO: CRO_T138994, G8O: CRO_T133061, 7DLH: CRO_T106494, IS: CRO_T130026, GES: CRO_T119458, HL2: CRO_T119458, MAT: CRO_T120028, GS2: CRO_T113153, STR: CRO_T125329, GS1: CRO_T113154, 8HGO: CRO_T107879, T19H: CRO_T119486, Redox1: CRO_T129272, 7DLGT: CRO_T131714, Redox2: CRO_T132421, TDC: CRO_T125328, LAMT: CRO_T103723, GO: CRO_T127440, SLS1: CRO_T109448, HYS: CRO_T116107, SGD: CRO_T128799, T3R: CRO_T124298, T16H2: CRO_T110598, 16OMT: CRO_T110596, NMT: CRO_T111273, T3O: CRO_T113994, DAT: CRO_T120021, D4H: CRO_T127167, HL1: CRO_T139139, T16H1: CRO_T110599. Inspection of the original CRO_T105646 gene prediction from the C. roseus v2 genome (Franke et al. 2018) revealed that this locus was a fusion of the ZCT1 and ZCT4 loci. In this paper, the fused locus was manually reannotated to create the CRO_T105646_ZCT1 and CRO_T105646_ZCT4 loci used for gene expression analysis and correlation analysis. In the most recent C. roseus vr3 genome (Li et al. 2023), ZCT1 and ZCT4 were correctly annotated and matched our manually annotated CRO_T105646_ZCT1 and CRO_T105646_ZCT4 loci. Heatmaps were generated in R using the gplots package (Warnes et al. 2009).

Construction of ZCT and MYC2a overexpression plasmids

Once identified, ZCT and MYC2a wild-type coding sequences were amplified from C. roseus complementary DNA (cDNA) in fragments, introducing synonymous A to T mutations to remove MoClo-incompatible restriction enzyme sites (Eco31I and BpiI). Fragments of coding sequences were harnessed into the universal pL-1 plasmid pAGM1311 and carried through the subsequent MoClo levels to a final level 2 destination vector pSB90. Coding sequences and associated oligonucleotides are listed in Table S1 and Table S2. Modular cloning was used to assemble the ZCT coding sequences under control of the strong CaMV 2 × 35S promoter (plasmid part # pICH51288, Engler et al. 2014) and the Mas (A. tumefaciens) terminator (plasmid part # pICH77901, Engler et al. 2014). The pSB90 backbone vector containing the constitutively active VirG (virGN54D mutation) expressed the assembled ZCT overexpression cassette (Mortensen et al. 2019a). The mutated CrMYC2a plasmid (CrMYC2a[D126N]) was generated from the wild-type expression vector via site-directed mutagenesis (Azenta Life Sciences). Constructed vectors were confirmed by restriction enzyme digestion.

Construction of native promoter-reporter constructs

Coding sequences for genes of interest were identified from the C. roseus transcriptome (Góngora-Castillo et al. 2012) and aligned by BLAST against the C. roseus Sunstorm Apricot genome (Kellner et al. 2015). A region of approximately 1 kilobase upstream of the start codon was amplified and cloned upstream of an intron-containing firefly luciferase coding sequence and Ocs (A. tumefaciens) terminator. This firefly luciferase cassette was then assembled into the multigene level 2 destination vector with an intron-containing Renilla luciferase gene expressed with the Nos (A. tumefaciens) promoter and terminator. Constructs were validated by double restriction enzyme digests. Plasmids were transformed into Agrobacterium tumefaciens (GV3101 (pMP90)) for use in transient evaluation. Oligonucleotides used to amplify promoter fragments from the genome and the promoter sequences may be found in Table S3 and Table S4. The construction of the native promoter-reporter plasmids for the vindoline pathway is described in Cole–Osborn et al. (Cole-Osborn et al. 2024).

Transient overexpression assays in C. roseus seedlings

Catharanthus roseus seedlings were transiently transformed with A. tumefaciens containing the effector and promoter driving reporter constructs using vacuum infiltration, as previously described (Mortensen et al. 2019a, 2022). For pooled overexpression studies evaluated by either dual-luciferase or RT-qPCR, Agrobacterium strains containing the individual ZCT (ZCT1, ZCT2, ZCT3, ZCT4, ZCT5, ZCT6) effector plasmid, each at an OD600 of 0.06, were combined. For single effector trans-activation, the Agrobacterium strain containing the ZCT overexpression plasmid was infiltrated at a final OD600 of 0.34. In both single and pooled effector transactivation, the Agrobacterium strain containing the promoter driving reporter plasmid was at an OD600 of 0.06 to achieve an effector to reporter ratio of 6 to 1. For overexpression studies evaluated by RT-qPCR, the reporter construct was omitted. For pZCT1 and pZCT3 trans-activation experiments, the Agrobacterium strain containing the effector (CrMYC2a) and reporter expression plasmids were each infiltrated at a concentration of OD600 of 0.2 each for a final OD600 of 0.4 (1–1 ratio).

After transient overexpression through vacuum infiltration, seedlings recovered for 2 days in the dark and one day in the light prior to harvest. For dual-luciferase analysis, 10 biological replicates of 2 pooled seedlings were collected for each condition per each experiment. For RT-qPCR analysis, 5 biological replicates of 15 pairs of cotyledons were collected for each condition per each experiment.

Relative promoter activity via dual luciferase assays

Relative promoter activity via luciferase activity was analyzed as previously described (Mortensen et al. 2019a) with the Luc-Pair™ Duo-Luciferase HT Assay Kit (Genecopeia). The plant protein extract from two pooled seedlings (20 µL) was mixed with 20 µL of each substrate solution in accordance with the kit protocol. Samples were measured in a 96-well white plate (Corning © 3693). Luminescence was measured in relative light units (RLU) with a plate reader (Bio-Tek Synergy HTX).

RNA extraction and RT-qPCR

Tissue (leaf or seedlings) was harvested into RNAse-free 2 mL conical tubes containing 10 × 3 mm glass beads. Upon harvest, tissue was immediately flash-frozen in liquid nitrogen and stored at − 80 °C until the day of extraction. RNA was extracted using the Direct-zol RNA Miniprep Kit (SKU R2071). RNA integrity was visually confirmed through gel electrophoresis. Single stranded complementary DNA (cDNA) was then synthesized using the LunaScript® RT SuperMix Kit (NEB #E3010). Transcript levels were then monitored via RT-qPCR using SYBR Reverse Transcriptase (RT) (ABClonal). Two technical replicates were included for each sample. Reactions were run on either the Stratagene MX3000P or Bio-Rad CFX.

Design of viral silencing fragments for C. roseus seedlings

The SGN VIGS design tool (vigs.solgenomics.net) (Fernandez-Pozo et al. 2015) was used to design fragments for Viral Induced Gene Silencing (VIGS) of ZCTs in seedlings. Coding sequences from the C. roseus transcriptome (Góngora-Castillo et al. 2012) were used as queries. The database “Catharanthus roseus v 2” was selected for reference. Parameters included: N-mer size: 23, Fragment length: 200–300 bp, mismatches 0–2. Top results were analyzed, and off-targets were entirely mitigated except in the cases of ZCT1 and ZCT2 which are too homologous to accommodate independent VIGS fragments. VIGS fragments were PCR-amplified from cDNA and cloned into the Golden Gate-compatible pTRV2-GG plasmid (Addgene Catalog # 105349) (Gantner et al. 2018). Oligonucleotides used to construct VIGS fragments in pTRV2 plasmids and the VIGS sequences may be found in Table S5 and Table S6.

Viral silencing of ZCTs in C. roseus seedlings

Catharanthus roseus seeds (0.4 g, Vinca Little Bright Eye, NEseed Cat #19140) were surface-sterilized in 70% ethanol for 45 s followed by 10% bleach for 10 min. Seeds were triple rinsed in sterile deionized water and then submerged and agitated in a 3% vol/vol Plant Preservative Mixture (PPM, Caisson Labs) solution for 16 h overnight in the dark (25 °C, 60 RPM). Then, seeds were planted on Gamborg’s B5 medium without sucrose and germinated in the dark for approximately 7 days. After germination, seedlings were transferred to light (16 h photoperiod) for 2 days to allow for photomorphogenesis. Seedlings were then planted in soil and grown under the same 16 h photoperiod until two true leaves emerged (~ 40 days). Once developed, seedlings were infected with Agrobacterium tumefaciens strains harboring pTRV1 and pTRV2 viral silencing vectors.

In preparation for infecting plants, A. tumefaciens (GV3101 (pMP90)) cultures were grown and induced as previously described (Mortensen et al. 2019a). After induction, cultures were normalized to an OD600 of 4 and the strains harboring pTRV1 and pTRV2 plasmids were mixed at a 1:1 ratio. Modified tweezers were dipped into the Agro-containing solution and each plant was pinched two times just below the highest node. Plants were transferred to the dark for 2 days and then placed back into the 16 h photoperiod until photobleaching in the ChLH-silenced condition was apparent, approximately after 21 days (Figure S6). After silencing occurred, the second emerging leaf pair after the infection point was harvested, flash frozen in liquid nitrogen and stored at − 80 °C prior to RNA extraction and evaluation via RT-qPCR. Oligonucleotides used for qPCR are listed in Table S7.

Results

Identification of Cys2-His2 zinc finger transcription factors in C. roseus

We identified the Cys2-His2 zinc fingers encoded in the C. roseus transcriptome (Góngora-Castillo et al. 2012) using the Plant Transcription Factor Database (TFDB) (Tian et al. 2020). Our search yielded 65 transcription factors containing from one to six Cys2-His2 zinc fingers (Figure S1). The number of Cys2-His2 zinc finger-containing transcription factors in C. roseus is fewer than that of Arabidopsis (176) (Englbrecht et al. 2004), rice (189) (Agarwal et al. 2007), soybean (321) (Yuan et al. 2018), or tomato (112) (Ming et al. 2020) and is closer to that of alfalfa (58) (Pu et al. 2021).

Of the 65 transcription factors containing Cys2-His2 zinc fingers identified in C. roseus, we focused on the ZCT subgroup, which contains two Cys2-His2 zinc fingers and a specific arrangement of defining motifs (B-box, L-box, and LxLxL motifs). The ZCT subgroup consists of the previously characterized, putative repressors of the TIA pathway, ZCT1, ZCT2, and ZCT3, and the newly identified and uncharacterized ZCT4, ZCT5, and ZCT6. In the cladogram with just these six ZCTs (Fig. 2), ZCT2 and ZCT4 clustered most closely while ZCT3 and ZCT6 clustered most closely.

Fig. 2
figure 2

C. roseus ZCT gene tree based on amino acid sequence. C. roseus ZCTs were identified using the Plant Transcription Factor Database. The Neighbor Joining algorithm and Jukes–Cantor method were used with the Bootstrap method to generate the cladogram based on amino acid sequence alignments. Bootstrap values are indicated

Full size image

We confirmed via an amino acid alignment that these newly identified ZCT sequences (ZCT4, ZCT5, ZCT6) contained the motifs or domains that are characteristic of the ZCT subgroup (ZCT1, ZCT2, and ZCT3) (Fig. 3). These six ZCTs share high sequence conservation mainly within their two Cys2-His2 zinc fingers but differ in the spacer length and sequence between their two Cys2-His2 zinc fingers. Cys2-His2 zinc fingers interact with target DNA through coordination with a zinc ion (Elrod-Erickson et al. 1996; Pabo et al. 2001) and through the characteristic plant specific QALGGH sequence within the Cys2-His2 zinc finger motif; the QALGGH sequence allows binding into the major groove of DNA (Takatsuji 1999). A single zinc finger provides relatively weak binding, while multiple fingers enhance binding (Iuchi 2001). In addition, the spacer length and sequence between the two Cys2-His2 zinc fingers affect the binding specificity of the ZCTs to their target (Kubo et al. 1998; Takatsuji and Matsumoto 1996). Therefore, the targets regulated by the ZCTs may differ and must be determined experimentally.

Fig. 3
figure 3

Amino acid alignment of C. roseus ZCT proteins shows conservation of characteristic motifs. The B-box and L-box regions are thought to play a role in nuclear localization and protein–protein interaction, respectively. The Cys2-His2 zinc fingers (with the Cys and His residues highlighted in blue) contain the QALGGH motifs responsible for binding to the major groove of the target DNA. The EAR or LxLxL motif is a strong repression domain and is consistently present in most ZCT proteins. tZCT6 refers to the truncated version of the ZCT6 protein and contains one instead of the two Cys2-His2 zinc finger domains found in the other ZCTs (color figure online)

Full size image

While amplifying ZCT6 from cDNA, both a complete and a truncated transcript were identified. The complete transcript is referred to as ZCT6 and the truncated transcript is referred to as tZCT6 (Fig. 3). We identified two QALGGH sequences in most ZCTs, the first of which was present in ZCT1 through ZCT6 but was absent from tZCT6, likely rendering it a weaker binder; therefore, we focused on characterizing the two-fingered ZCTs.

Each of these ZCTs also contains a B-box which is suspected to serve as a nuclear localization signal and a leucine-rich box (L-box) which is presumed to assist in protein interaction (Kubo et al. 1998; Sakamoto et al. 2004). The final conserved element among these ZCTs is the EAR motif, a characteristic LxLxL sequence. The EAR motif is a strong repression domain and is consistently present in most of the ZCTs, with ZCT3 harboring a second LxLxL motif within its L-box. The EAR motif is one of the most common transcriptional repressor domains found in plants. This domain works by recruiting chromatin remodeling factors that prevent expression of genes (Chow et al. 2023).

A UniProt and BLAST search using the C. roseus ZCT protein sequences showed that the six ZCTs are most closely related to the Arabidopsis ZAT and AZF proteins (Table S8). The C. roseus ZCTs and Arabidopsis ZAT and AZF proteins share the highest sequence identity in the Cys2-His2 zinc finger domains but most of these Arabidopsis proteins did not contain all the characteristic motifs found in the C. roseus ZCTs (Fig. 3; Figure S2). Among the best characterized Arabidopsis ZATs are ZAT10 and ZAT12 (Table S8); these ZATs are involved in response to environmental stress, including hypoxia, cold, heat, salinity, drought, light, UV-B, wounding, and others. For example, ZAT10 overexpression facilitates response to osmotic stress (Mittler et al. 2006). Similarly, ZAT12 has been shown to improve tolerance to low temperature and drought stress (Davletova et al. 2005; Vogel et al. 2005). Among the Cys2-His2 zinc fingers, the C. roseus ZCTs (i.e., the previously characterized ZCT1, ZCT2, ZCT3) are unique in that they play a putative role in repressing alkaloid biosynthesis.

In our search for additional ZCTs, we identified ZCT4, ZCT5, and ZCT6 that clustered with the previously characterized putative repressors of the TIA pathway, ZCT1, ZCT2, and ZCT3. The ZCTs share two highly conserved Cys2-His2 zinc fingers and the EAR repression motif (with ZCT6 possessing two motifs) but differ in spacer sequence and length outside of the Cys2-His2 zinc fingers; these differences can affect the promoters they bind and regulate. Thus, we proceeded with evaluating if ZCT4, ZCT5, and ZCT6 could act redundantly or compensate as repressors of TIA biosynthesis; if the ZCTs act redundantly and are similarly induced, then increasing TIA biosynthesis would require the inhibition (i.e., knockout or silencing) of multiple ZCTs rather than one or two ZCTs (i.e., ZCT1, ZCT2), as previously demonstrated in Rizvi et al (2016).

Tissue-specific and jasmonate-responsive expression of ZCTs

To assess the role and redundancy of the expanded members of the ZCT family, we explored their expression profile in various tissue types and in response to jasmonate. We used previously published transcriptome data (Góngora-Castillo et al. 2012) to identify relative transcript levels of ZCTs in various tissue types (Fig. 4). ZCT3 was the most highly expressed ZCT in most tissue types, except for the flower where ZCT5 and ZCT6 were most highly expressed. Flowers expressed most of the six ZCTs with the stem tissue being the next highest. Interestingly, the tissues that expressed the highest ZCT levels (flowers and stems) also contained the lowest alkaloid levels while the tissues that expressed the lowest ZCT levels (immature and mature leaves) contained the highest alkaloid levels (Pan et al. 2016b). This observation is consistent with the putative role of ZCTs as repressors of TIA biosynthesis.

Fig. 4
figure 4

Expression of ZCTs in C. roseus tissue. Publicly available transcriptome data (Góngora-Castillo et al. 2012) was used to generate a heat map of ZCT expression in various tissue types. Expression level is reported as fragments per kilobase transcripts per million (FPKM)

Full size image

Next, we evaluated the jasmonate responsiveness of ZCT4, ZCT5, and ZCT6 in hairy root cultures where increasing ZCT1, ZCT2, and ZCT3 levels were correlated with repressed TIA gene expression (Goklany et al. 2013; Peebles et al. 2009a). A promoter scan (Figure S3) confirmed that ZCT4, ZCT5, and ZCT6 contain cis-regulatory elements that play a role in jasmonate-inducibility (i.e., WUN: wound-responsiveness; TGACG and CGTCA-motifs: methyl jasmonate responsiveness) as well as tolerance to environmental stressors (i.e., MBS: drought-inducibility; HSE: heat stress responsiveness; TC-rich repeats: defense and stress responsiveness). We subsequently determined if the newly identified ZCT4, ZCT5, and ZCT6 were also jasmonate-responsive and could contribute to the repression of TIA biosynthesis associated with high jasmonate dosage in hairy root cultures, as observed with ZCT1, ZCT2, and ZCT3.

We elicited C. roseus hairy roots with two dosages of methyl jasmonate (250 and 1000 µM MJ) and monitored ZCT transcript levels at 30 min and 24 h (Fig. 5). The selected jasmonate dosages of 250 and 1000 µM were previously shown to optimize and inhibit TIA pathway gene expression and TIA levels, respectively (Goklany et al. 2013). At both jasmonate dosages, all ZCTs (excluding ZCT5) exhibited a burst of expression at 30 min, a rapid response suggesting regulation by a pre-existing transcription factor. At the optimal dosage for alkaloid production in hairy roots (Goklany et al. 2013) (250 µM MJ), ZCT1 and ZCT3 were elevated 5–10-fold at 24 h whereas ZCT2, ZCT4, and ZCT6 had returned to near basal levels. At the inhibitory dosage for alkaloid production in hairy roots (1000 µM MJ), the expression of ZCT1 increased ~ 100-fold while the expression of ZCT2, ZCT3, ZCT4, and ZCT6 increased 5–20-fold by 24 h; the high expression level of ZCTs was sustained at 24 h as compared to that at the optimum jasmonate dosage. Unlike other ZCTs, the expression of ZCT5 was inhibited with jasmonate at both dosages. Our results are consistent with previous results that ZCT1, ZCT2, and ZCT3 are induced by jasmonate. The new candidates, ZCT4 and ZCT6, were also responsive to jasmonate and elevated at the high jasmonate dosage associated with repressed alkaloid production, suggesting their potential role in regulating the TIA pathway.

Fig. 5
figure 5

Jasmonate-responsiveness of ZCTs in hairy root cultures. C. roseus hairy root cultures were treated with two dosages of methyl jasmonate (250 or 1000 µM MJ), and tissue was collected at the designated timepoints (0, 0.5, and 24 h). Genes of interest (ZCTs) were monitored by RT-qPCR. Relative transcript level was first normalized to the housekeeping gene SAND and then to the control condition (0 h) using the 2−ΔΔCt method. Each data point represents the average from tissue harvested from three independent flasks (i.e., three biological replicates) with error bars representing the standard deviation

Full size image

Finally, a co-expression analysis (Figure S4) using transcriptomic data acquired across tissues, timepoints, and treatment conditions was performed to identify TIA genes correlated with ZCT expression and therefore potentially regulated by ZCTs. ZCT2, ZCT3, and ZCT4 were strongly correlated with each other and negatively correlated with the expression of the vindoline pathway genes. The correlation between the vindoline pathway and ZCTs was tested experimentally in the following section.

The role of ZCTs in regulating TIA biosynthesis

To elucidate their regulatory role in the TIA pathway, the activity of ZCTs was assessed through three complementary approaches: transient co-expression of ZCTs with promoter-driving reporter constructs, transient overexpression of ZCTs with transcript monitoring, and transient silencing of ZCTs through viral induced gene silencing (VIGS) followed by transcript monitoring. To evaluate gene function, we used several transient expression methods since the development of transgenic C. roseus plants is time-intensive and has low transformation efficiencies (Choi et al. 2004; Kumar et al. 2018; Pan et al. 2012; Sharma et al. 2018; Verma and Mathur 2011; Wang et al. 2012).

ZCTs were transiently co-expressed with TIA promoter-driving reporter constructs using Agrobacterium in C. roseus seedlings. First, we pooled Agrobacterium strains overexpressing ZCT1 through ZCT6 and co-infiltrated C. roseus seedlings with Agrobacterium strains containing the candidate promoter-reporter constructs (Fig. 6). Promoters screened included genes from the terpenoid branch (pG10H, pLAMT), the condensation and upstream steps (pSTR1, pGO), the root-associated pathway (pT19H), and the vindoline branch (pT16H1, pT16H2, p16OMT, pT3O, pT3R, pNMT, pD4H, pDAT). These promoters were selected based on prior evidence: repression of the terpenoid branch (Goklany et al. 2013) and the condensation step (Goklany et al. 2013; Pauw et al. 2004) and from our co-expression analysis (Figure S4) (negative correlation with the vindoline branch). A dual luciferase assay was used to quantify promoter driven reporter activity.

Fig. 6
figure 6

Promoter trans-activation screen of TIA pathway promoters by pooled ZCTs. Candidate transcription factors (ZCT1 through ZCT6) harbored in individual cultures of Agrobacterium tumefaciens were pooled and overexpressed in the presence of the promoter-driving reporter plasmid with a 6 to 1 ratio of effector to promoter::reporter for a total OD600 of 0.4. Each biological replicate or data point consists of two seedlings that were pooled and flash-frozen in liquid nitrogen. Proteins were extracted and quantified by measuring the luminescence of the intron-containing firefly luciferase (FLUC-I) and Renilla luciferase (RLUC-I). Relative promoter activity was calculated by normalizing (FLUC-I/RLUC-I)ZCTs/(FLUC-I/RLUC-I)Average GUS, where GUS served as the control condition and was set to 1. The graph combines one to three replicate experiments with 7–15 biological replicates for each promoter per experiment. The horizontal line in the box plot represents the median, the ends of the box represent the 1st and 3rd quantile, and the error bars represent the range. The data were log-transformed to achieve a normal distribution prior to statistical testing. Significance based on Student’s t test is denoted as (*) p < 0.05, (**) p < 0.01, (***) p < 0.001, and (****) p < 0.0001

Full size image

Significant repression of the promoter-reporter constructs of pLAMT (p < 0.05), pSTR1 (p < 0.0001), pT16H2 (p < 0.01), and pT3O (p < 0.001) was observed in the screens where ZCTs were pooled (Fig. 6). In contrast, enhanced activity of pT16H1 (p < 0.05), pNMT (p < 0.0001), and pD4H (p < 0.0001) also occurred with the pooled ZCTs compared to the GUS overexpression control. To complement the promoter transactivation studies, we also overexpressed ZCTs in seedlings (in the absence of the promoter-driving reporter construct) and evaluated their effect on the native gene profile (Figure S5). No significant repression of the monitored genes was observed across two experimental replicates. However, a 2-fold activation of T16H1 expression (p < 0.01) was apparent in one experiment where ZCT expression levels were particularly high (Figure S5A), consistent with the transactivation of the pT16H1 promoter (Fig. 6).

Similarly, we performed trans-activation experiments with individually expressed ZCTs on a subgroup of candidate promoters across the TIA pathway (Fig. 7). We observed significant repression of pSTR1 by ZCT4 (p < 0.001), ZCT5 (p < 0.0001), and ZCT6 (p < 0.01). We also identified ZCT1 (p < 0.01) and ZCT2 (p < 0.05) to be significant repressors of pT16H2. Consistent with the pooled evaluation strategy, no effect on pG10H expression was observed with the overexpression of individual ZCTs.

Fig. 7
figure 7

Promoter trans-activation screen of TIA pathway promoters by independently overexpressed ZCTs. Candidate transcription factors (either ZCT1, ZCT2, ZCT3, ZCT4, ZCT5, or ZCT6) harbored in individual cultures of Agrobacterium tumefaciens were independently overexpressed in the presence of the promoter-driving reporter plasmid with a 6 to 1 ratio for a total OD600 of 0.4. Each biological replicate or data point consists of two seedlings that were pooled and flash-frozen in liquid nitrogen. Proteins were extracted and quantified by measuring the luminescence of the intron-containing firefly luciferase (FLUC-I) and Renilla luciferase (RLUC-I). Relative promoter activity was calculated by normalizing (FLUC-I/RLUC-I)ZCTs/(FLUC-I/RLUC-I)Average GUS, where GUS served as the control condition and was set to 1. The graph combines three replicate experiments with 10 biological replicates for each condition per experiment. The horizontal line in the box plot represents the median, the ends of the box represent the 1st and 3rd quantile, and the error bars represent the range. Data were log-transformed to achieve a normal distribution prior to statistical testing. Significance based on one-way ANOVA and Dunnett’s test is denoted as (*) p < 0.05, (**) p < 0.01, (***) p < 0.001, and (****) p < 0.0001

Full size image

To evaluate their role in planta, we performed Viral Induced Gene Silencing (VIGS) of the ZCTs in young plants (Figure S6). Because ZCTs repressed specific TIA promoters in our trans-activation experiments, the VIGS of ZCTs would be expected to upregulate these target genes. To account for possible compensation by redundant ZCTs, we designed 200 bp VIGS fragments that would either target individual ZCTs per experiment (i.e., ZCT1/2 or ZCT3), clusters of ZCTs per experiment (jasmonate responsive cluster of ZCT1, ZCT2, and ZCT3), or all six ZCTs simultaneously per experiment (ZCT1 through ZCT6). Transcript levels were monitored via RT-qPCR and compared to that of a control condition that targeted GFP. Effective silencing was confirmed using a ChLH silencing fragment that resulted in photobleaching.

Although there was successful photobleaching of ChLH plants, no silencing of ZCTs was evident. The resiliency of the ZCTs to RNA silencing machinery in transiently transformed plants suggests a potential feedback mechanism. For instance, we previously observed the repression of the pZCT1 promoter with the overexpression of ZCT1 (Mortensen et al. 2019b). A similar resistance to viral silencing of a transcription factor was observed with the C. roseus BIS1 VIGS experiments (Van Moerkercke et al. 2015). In a later publication, BIS was found to amplify its own expression (Schweizer et al. 2018), which would explain the difficulty of silencing BIS.

CrMYC2a can induce the jasmonate-responsiveness of ZCTs

Results from the promoter trans-activation experiments suggested that ZCTs act as regulators of the TIA pathway when transiently over-expressed (Fig. 6) and in addition, that ZCTs are highly induced with jasmonate (Fig. 5). Since ZCTs, like ORCAs (Zhang et al. 2011), are rapidly induced within 30 min of jasmonate addition (Fig. 5), we hypothesized that CrMYC2a, a pre-existing jasmonate-responsive transcription factor, activates ZCT expression. When jasmonate is added, JAZ is ubiquitinated and degraded (Chini et al. 2007; Thines et al. 2007), releasing MYC2a to activate the expression of the ORCAs (Schweizer et al. 2018; Zhang et al. 2011) and BISs in C. roseus (Van Moerkercke et al. 2016). We evaluated the role of CrMYC2a in regulating ZCTs through a promoter trans-activation assay in transiently transformed C. roseus seedlings (Fig. 8).

Fig. 8
figure 8

Trans-activation of ZCT1 and ZCT3 promoters with CrMYC2a. GUS (control), wild-type CrMYC2a (wtMYC2a), and mutated CrMYC2a (MYC2a[D126N]), harbored in individual cultures of Agrobacterium tumefaciens, were overexpressed in the presence of the ZCT1 and ZCT3 promoter-driving reporter plasmids with a 1 to 1 ratio for a total OD600 of 0.4. Each biological replicate or data point consists of two seedlings that were pooled and flash-frozen in liquid nitrogen. Proteins were extracted and quantified by measuring the luminescence of the intron-containing firefly luciferase (FLUC-I) and Renilla luciferase (RLUC-I). Relative promoter activity was calculated by normalizing (FLUC-I/RLUC-I)MYC2a/(FLUC-I/RLUC-I)Average GUS, where GUS served as the control condition and was set to 1. The graph combines two to three replicate experiments with 10 biological replicates for each promoter per experiment. The horizontal line in the box plot represents the median, the ends of the box represent the 1st and 3rd quantile, and the error bars represent the range. Data were log-transformed to achieve a normal distribution prior to statistical testing. Significance based on one-way ANOVA and Dunnett’s test is denoted as (*) p < 0.05, (**) p < 0.01, (***) p < 0.001, and (****) p < 0.0001

Full size image

We evaluated trans-activation of the most jasmonate-inducible ZCT promoters (pZCT1 and pZCT3) by overexpressing CrMYC2a and its de-repressed mutated CrMYC2a[D126N] (Schweizer et al. 2018). In addition to activating the expression of the ORCAs and BISs, CrMYC2a activates the expression of JAZ, leading to the regeneration of JAZ and the subsequent repression of CrMYC2a, thereby limiting the activation of TIA biosynthesis (Chini et al. 2007; Thines et al. 2007). Unlike CrMYC2a, the CrMYC2a[D126N] mutant lacks the negatively charged aspartate (D), rendering it incapable of binding to JAZ; this mutation results in a constitutively active or de-repressed CrMYC2a.

Both pZCT1 and pZCT3 promoters were significantly activated, i.e., 5-fold increases, by overexpression of the de-repressed CrMYC2a[D126N]. pZCT1 was also activated by the wild-type CrMYC2a while the pZCT3 promoter was unaffected. Since ZCT3 expression is higher than ZCT1 in seedlings (Fig. 4), the pZCT1 promoter activity may have been lower than that of the pZCT3 promoter, and therefore its trans-activation was more sensitive and easily observed. The activation of both pZCT1 and pZCT3 promoters by jasmonate-associated CrMYC2a suggests a molecular mechanism by which jasmonate-responsive ZCT expression (i.e., ZCT1, ZCT2, ZCT3, ZCT4, ZCT6) can be turned on with methyl jasmonate.

Discussion

In this study, we investigated the role of an expanded number of ZCT transcription factors as regulators of the TIA pathway in C. roseus. First, we performed a bioinformatic analysis and identified 65 transcription factors containing Cys2-His2 zinc fingers in the C. roseus genome; we focused on the characterization of the ZCT subgroup, which included the previously characterized, putative repressors ZCT1, ZCT2, and ZCT3, and the newly identified ZCT4, ZCT5, and ZCT6. These six ZCTs contained the repressive EAR motif and were responsive to jasmonate. Thus, we investigated the potential targets regulated by these six two-fingered Cys2-His2 ZCTs.

To investigate potential targets of ZCTs, we screened promoters of genes from the terpenoid branch (pG10H, pLAMT), the condensation and upstream steps (pSTR1, pGO), the root-associated pathway (pT19H), and the vindoline branch (pT16H1, pT16H2, p16OMT, pT3O, pT3R, pNMT, pD4H, pDAT; Fig. 1). The transient overexpression of pooled ZCTs in seedlings regulated the promoters of the terpenoid (pLAMT), condensation (pSTR1), and the vindoline branches of the TIA pathway (Fig. 6). The significant repression of pLAMT from the terpenoid branch (by 27%) was consistent with previously published results; for instance, (Chebbi et al. 2014) reported that ZCT1 and ZCT2 acted as repressors of HDS, involved in catalyzing the production of isopentenyl diphosphate upstream of LAMT.

The overexpression of pooled ZCTs also significantly repressed the condensation step, pSTR1 (by 57%). By overexpressing ZCTs individually in seedlings, we showed that ZCT4, ZCT5, and ZCT6 could repress pSTR1 by 41–61% (Fig. 7). Pauw et al. (Pauw et al. 2004) transformed cell suspensions via particle bombardment and showed repression of pTDC and pSTR1 with ZCT1, ZCT2, and ZCT3 overexpression; in contrast, we observed that ZCT1, ZCT2, and ZCT3 did not repress pSTR1 in seedlings transformed via Agro-infiltration. Differences in methodology and in tissue-specific ZCT levels could have impacted the observed repression. For instance, particle bombardment introduces wounding to transform plant tissue, initiating jasmonate biosynthesis responses (Bidney et al. 1992) and potentially inducing ZCT expression. In contrast, Agrobacterium infection activates the salicylic acid pathways that compete with the jasmonic acid pathways (Bidney et al. 1992; León et al. 2001; Reymond et al. 2000). In addition, promoter activity or native transcript levels will differ between seedlings and cell suspensions; thus, a lack of repression may be attributed to the low activity of that promoter where further decrease in promoter activity is undetectable. The lack of repression may also be attributed to the relatively high native ZCT level in that tissue. For instance, native ZCT3 levels were already high in seedlings (Fig. 4), potentially limiting the effect of overexpression as no effect was observed with ZCT3 overexpression in our seedlings; in contrast, the native levels of ZCT4, ZCT5, and ZCT6 were low in seedlings, increasing the likelihood of observing an effect with their overexpression. Therefore, while the specific ZCT responsible for repression differed between studies, the repression of pLAMT and pSTR1 by the ZCT family was consistent with that in the literature.

In our study, the overexpression of ZCTs repressed the vindoline pathway promoters pT3O (by 35%) and pT16H2 (29%) and surprisingly activated pT16H1 (42%), pNMT (97%), and pD4H (122%) (Fig. 6). When individually tested, both ZCT1 and ZCT2 repressed pT16H2 by 38–47% (Fig. 7). Our study is the first to investigate the role of ZCTs in regulating the vindoline pathway. The EAR-motif is a main feature of the ZCTs and is responsible for the recruitment of co-repressors. Therefore, the observed activation of genes such as pT16H1, pNMT, and pD4H is likely through an indirect mechanism. For example, ZCTs might repress a repressor of the vindoline genes, resulting in increased vindoline gene expression.

As a complement to the promoter trans-activation studies, we overexpressed pooled ZCTs in seedlings (in the absence of the promoter-driving reporter construct) and evaluated its effect on gene expression, specifically for LAMT, STR1, T16H1, T16H2, T3O, NMT, D4H, genes whose promoters were regulated by ZCTs in the promoter trans-activation studies. We observed varying levels of ZCT overexpression between experimental replicates, particularly ZCT4 and ZCT5 levels (Figure S5). No repression of monitored genes was observed in these two experimental replicates (Figure S5). Due to the potential stability of these monitored mRNA transcripts, repression of TIA genes may not have been detectable via RT-qPCR. In the experimental replicate where ZCTs were most highly overexpressed, a significant 2-fold activation of T16H1 was observed but NMT and D4H were not (Figure S5A), potentially due to their already high basal transcript levels. As a third approach, we attempted to evaluate the role of ZCTs in planta through VIGS. While ChLH-silencing was successful, ZCT levels remained unaffected (Figure S6), partly due to the already low basal levels of specific ZCTs in immature leaves (Fig. 4).

In our study, the expression of ZCTs (except ZCT5) increased with jasmonate in a dosage-dependent manner (Fig. 5). The rapid 30-min response of ZCTs to methyl jasmonate was evident, like the rapid 30-min response of ORCA3 to methyl jasmonate previously observed (Goklany et al. 2013). Therefore, we investigated a mechanistic explanation for their high induction upon jasmonate elicitation. The activator protein MYC2a is a well characterized player in jasmonate response. This constitutively expressed protein exists in a repressed state, bound by JAZ proteins. Upon elicitation by methyl jasmonate, rapid ubiquitination and degradation of JAZ proteins occurs, freeing MYC2a to activate other genes (Chini et al. 2016). Thus, the rapid response of ZCTs to methyl jasmonate (within 30 min) suggests regulation by MYC2a, rather than mechanisms which require synthesis from the genome level. Previously, we demonstrated that ORCA3 did not transactivate the pZCT1 promoter (Mortensen et al. 2019b). In this paper, we confirmed significant activation of the pZCT1 and pZCT3 promoters by the constitutively active or de-repressed CrMYC2a, suggesting that the jasmonate-responsive expression of the ZCTs (ZCT1, ZCT2, ZCT3, ZCT4, ZCT6) is mediated by CrMYC2a (Fig. 8).

MYC2a is an activator in jasmonate signaling pathways but also serves to regulate other pathways such as salicylic acid, gibberellic acid, and auxin biosynthesis (Kazan and Manners 2013). It is possible that ZCTs are activated by MYC2a for environmental adaptation. This role in environmental adaptation has been observed for zinc finger proteins in several other plant species. In Arabidopsis, AtZFP11 responds to jasmonate and mediates stress-responsive genes (Dinkins et al. 2003). In rice (O. sativa), ZFP179 (Sun et al. 2010), ZFP182 (Huang et al. 2007), and ZFP245 (Huang et al. 2005, 2009) transcription factors support plant metabolism in response to salinity and drought stress. In soybean, SCOF-1 is activated by low temperature (Kim et al. 2001). In tomato, all 112 Cys2-His2 zinc finger proteins have been identified, and are induced to various levels upon cold, heat, salinity, and osmotic stresses (Ming et al. 2020). Therefore, due to their homology with Arabidopsis zinc fingers, the C. roseus ZCTs may play a similar role in abiotic stress response.

Conclusion

For nearly two decades, ZCT1, ZCT2, and ZCT3 have been considered putative repressors of the TIA biosynthetic pathway in C. roseus based on its original studies (Chebbi et al. 2014; Pauw et al. 2004). Here, we identified additional ZCTs, ZCT4, ZCT5, and ZCT6, and characterized the expanded subgroup of ZCTs (ZCT1 through ZCT6) in C. roseus. The structure of ZCTs (two-fingered DNA binding domains and EAR repressor domain) paired with its jasmonate-responsiveness suggests a role in regulating the defense-activated TIA pathway. In this paper, we showed that ZCTs can repress as well as indirectly activate TIA promoters when expressed at high levels and that their jasmonate-responsiveness can be mediated by CrMYC2a.