TcJAMYC: a bHLH transcription factor that activates paclitaxel biosynthetic pathway genes in yew.

: Paclitaxel (Taxol ® ) is a highly modified diterpene anti-cancer agent produced by the gymnosperm Taxus. Taxus suspension cell cultures have the potential to provide a sustainable source of paclitaxel, but the paclitaxel biosynthetic pathway is not fully characterized, making metabolic engineering efforts difficult. Methyl jasmonate (MJ) is used to elicit paclitaxel production in suspension cultures. Here we show that the promoters of five genes encoding enzymes of the paclitaxel biosynthetic pathway are activated by MJ elicitation. Thus, elicitation of paclitaxel production by MJ is regulated at least in part at the level of transcription. Separately. MJ-elicited Taxus cultures were plated onto B5 agar plates and bombarded with the promoter:GUS fusions alone (+MJ, -MYC) or in combination with the CaMV35S:TcJAMYC effector plasmid (+MJ, +MYC). Results are presented as a GUS/LUC ratio. N=4 in all samples, error bars are SE.


Abstract:
Paclitaxel (Taxol ®) is a highly modified diterpene anti-cancer agent produced by the gymnosperm Taxus. Taxus suspension cell cultures have the potential to provide a sustainable source of paclitaxel, but the paclitaxel biosynthetic pathway is not fully characterized, making metabolic engineering efforts difficult. Methyl jasmonate (MJ) is used to elicit paclitaxel production in suspension cultures. Here we show that the promoters of five genes encoding enzymes of the paclitaxel biosynthetic pathway are activated by MJ elicitation.
Thus, elicitation of paclitaxel production by MJ is regulated at least in part at the level of transcription. A transcription factor that positively activates the promoters of paclitaxel biosynthetic genes has been cloned. This transcription factor, TcJAMYC, possesses a high degree of similarity to AtMYC2 and JAMYC2, which are known to regulate the expression of jasmonate inducible genes in other systems. TcJAMYC binds to E-boxes found in the promoters of the paclitaxel pathway genes, and these promoters are transiently activated by TcJAMYC overexpression. Addition of MJ attenuates the effect of TcJAMYC on the pathway promoters, suggesting that TcJAMYC could be engineered into Taxus cells to avoid the negative regulation of MJ-induced genes that follows initial MJinduced positive regulation.
This strategy could prolong the expression of paclitaxel pathway genes, and increase paclitaxel production.

Introduction:
Paclitaxel (Taxol ®; Figure 1) is a plantderived diterpenoid with anti-mitotic activity that is produced by the gymnosperm genus Taxus. Paclitaxel is approved by the US Food and Drug Administration for the treatment of ovarian, breast and non-small cell lung cancer (www.fda.gov). Also, paclitaxel is being incorporated into arterial stents to inhibit scar tissue formation after implant (1), thus the demand for this important compound is expected to increase. Paclitaxel is found in only small quantities in plants, and total synthesis comprises several steps, and is therefore lowyielding. The main current sources of paclitaxel are semi-synthesis from abundant advanced precursors derived from the needles of yew plants and from Taxus suspension cell cultures (2)(3)(4). The biosynthetic pathway leading to paclitaxel has been partially elucidated (5) (Figure 1), but improved understanding of the four or five undefined pathway steps and the overall regulation of paclitaxel synthesis are needed in order to enable bioengineering approaches. This will allow improved overall production of paclitaxel, or other taxanes and potentially may allow production of novel bioactive taxanes in plant cells.
Generally, Taxus suspension cultures are elicited to produce paclitaxel using 100 µM methyl jasmonate (MJ) (6). The mRNA levels of the known paclitaxel biosynthetic pathway genes in Taxus suspension cultures increase within six hours after elicitation. Transcript levels peak at 12 h to 18 h and then taper off to basal levels within 30 h. Accumulation of 10-deacetyl baccatin III and baccatin III parallels transcript profiles, but with a 12-18 h lag. However, accumulation of WITHDRAWN AT AUTHORS REQUEST paclitaxel occurs much later (three to six days after elicitation), long after steady state transcript levels of the known pathway genes have returned to normal (7). The mechanism(s) underlying the poor correlation between steady state mRNA levels and accumulation of the desired metabolite paclitaxel are not understood. Still, the coordinated expression of the known paclitaxel pathway genes suggests that these genes could be under the control of a single regulatory regime.
Here, the mechanism of MJ elicitation of gene expression in Taxus suspension cell cultures was investigated and a regulator of the paclitaxel metabolic pathway was identified. This regulator has been named Taxus cuspidata JAMYC (TcJAMYC). There is similarity in sequence and function between TcJAMYC and the wellcharacterized AtMYC2, indicating a conserved response to MJ despite at least 150 million years of divergence between the angiosperm and gymnosperm lineages (21). The results presented here suggest that TcJAMYC is a key candidate gene for engineering increased paclitaxel accumulation in Taxus cell cultures.

Materials and Methods:
Cell Culture: Taxus cuspidata suspension cultures were used for all experiments. Suspensions were subcultured every two weeks into Gamborg`s B5 (Sigma, St. Louis, MO) basal salts (3.2 g/L) with 20 g/L sucrose, supplemented with 2.7 µmol/L naphthalene acetic acid (NAA) and 0.01 µmol/L benzyladenine (BA). Ascorbic acid (156 mg/L), citric acid (156 mg/L) and glutamine (906 mg/L) were filter-sterilized (Millipore Millex 0.2 µm syringe filters) and added post-autoclaving. Cultures were maintained in 125 mL Erlenmeyer flasks capped with Bellco (Vineland, NJ) foam closures at 24°C and shaking at 125 rpm in the dark. Cells-transfers were accomplished through the addition of 10 mL of 14 day-old suspension cultures into 40 mL of fresh medium. The approximate packed volume of cells transferred was at least 2 mL to maintain optimum culture density. Transient Transformation and MJ Elicitation: Cells were either mock-elicited by adding 50% EtOH or elicited with MJ (dissolved in 50% EtOH) at a 100 µM final concentration. After three h, both batches of cells (~0.5 g) were transferred onto B5-agar plates, spread out in a circle in the center of the plate with a diameter of about three cm, and gently pressed into the agar to immobilize them. Within two hours, the cells were transformed by bombardment using a PDS-1000 (Bio-RAD, Hercules, CA). Preparation of gold microcarriers was performed as described (22). Four replicates of each experiment were performed. Forty-eight h between bombardment and assay was provided to allow for optimal expression of the reporter genes. GUS and LUC Assays: Cells were ground in lysis buffer (1 ml; 100 mM KHPO 4 pH 7.8 with 0.2% Triton X-100) using ~50 µl of glass beads and a plastic pestle-bit driven by a hand-held drill for one min. The cell lysate was incubated on ice for five min, and the debris was removed by two rounds of centrifugation for 10 min at 16,000 x g. Luciferase activity was measured using a Turner Designs (Sunnyvale, CA) 20/20 luminometer set at 70 % sensitivity using 10 µl of the cell lysate. The luciferase activity was detected using the Applied Biosystems Dual Light System (Foster City, CA) per the manufacturer's instructions. GUS assays were performed using a Turner Designs 450 fluorometer with an NB360 excitation filter and a SC430 emission filter. 4-Methyl umbelliferone (Sigma, St. Louis, MO) was used as a calibration standard. GUS assays contained 75 µl protein extract, 50 µl MeOH, and 125 µl 2X MUG buffer, as described (23). Cloning of Paclitaxel Biosynthetic Gene Promoters:

WITHDRAWN AT AUTHORS REQUEST
The upstream flanking regions of the biosynthetic pathway genes were cloned using inverse-PCR (25) using the cDNA sequences of T5αH (AY628434), DBAT (AF193765), PAM (AY582743), BAPT (AY082804) and DBTNBT (AF466397). PCR products were cloned into pDESTG221 to form translational fusions to GUS with 5 to 30 N-terminal amino acid residues of the pathway biosynthetic gene prior to the fusion point with GUS. Vector Construction: A vector, pDESTG221, was constructed by modifying pPZP221 (24) to contain a Gateway recombination cassette B (Invitrogen, Carlsbad, CA), GUS reporter gene (23) and NOS terminator. The Taxus promoter DNA fragments were reamplified from I-PCR fragments using primers that contained attL sites 5' to the gene specific nucleotides. PCR products were placed into pDESTG221 using Gateway LR Clonase II (Invitrogen, Carlsbad, CA), creating translational fusions to the GUS reporter gene. The TcJAMYC effector gene was amplified from Taxus cuspidata suspension cell culture first-strand cDNA and ligated in between the CaMV35S promoter and a NOS 3' sequence. The 35S:GUS construct was taken from pBI121:GUS and placed into the multiple cloning site of pPZP221 using EcoRI and BamHI restriction sites. The 35S: LUC construct was ligated into pTZ19u. Degenerate Primer Amplification of TcJAMYC: The conserved bHLH domain in the JAMYC proteins from Solanum tuberosum (AJ630505) (17) and Arabidopsis (At1g32640) (26) was used to design degenerate primers with the Consensus-Degenerate Hybrid Oligonucleotide Primers (CODE-HOP) program (27).
Five primers were chosen, and two (Forward primer oDEGmyc.1 GAGAAGAACCTCTGAATCATGTTGARGCN GARMG; Reverse primer oDEGmyc.3 CAGCTTACATTTCAGTTCATTA ATATAAGAAATNGCRTCNCC) produced a PCR product that was then used to screen a cDNA library by hybridization. cDNA Library Construction and Screening: Total RNA was extracted from two grams of Taxus cuspidata cell culture line P991, by guanidium isothiocyanate and cesium chloride gradient ultracentrifugation at 104,000 x g for 18 h, followed by phenol-chloroform extraction.
Poly-A RNA (5 µg) was obtained from 1 mg total RNA using Poly-A Purist Mag-Kit (Ambion, Austin, TX). cDNA construction and cloning was performed using the ZAP Express cDNA Synthesis Kit (Agilent Technologies, Cedar Creek, TX). Plaques (1x10 6 ) from the primary library were screened using the Taxus bHLH fragment obtained by PCR. A 2.5 kb cDNA clone was isolated and sequenced, but the 5'-end was truncated. 5'-Rapid Amplification of cDNA Ends (RACE) was performed using RLM-RACE kit (Ambion, Austin, TX). The full-length cDNA was cloned by PCR using the 5'-sequence obtained by RACE, and this product was cloned and sequenced. Phylogenetic Analysis: Each of the bHLH proteins from Arabidopsis and the TcJAMYC were used to create an unrooted phylogram. Only the DNA binding domain was parsimony informative, so this limited sequence was used to construct the unrooted phylogram. Amino acid sequences were aligned using ClustalX with default settings. Maximum parsimony analysis was performed using the Phylogenetics Analysis Using Parsimony (PAUP 4.0) program. Bootstrap values were calculated using PAUP with 1000 repetitions. TcJAMYC Protein Purification: The TcJAMYC cDNA was recombined into pDEST17 (Invitrogen), an E. coli expression vector containing an N-terminal 6X His tag for affinity purification on a Nickel agarose column (Qiagen, Valencia, CA). The Rosetta 2(DE3) pLysS (EMD biosciences, Gibbstown, NJ) strain of E. coli containing this construct was grown to late log phase (OD 600 = 0.8), induced with 1 mM IPTG, and then incubated with shaking for four more hours at 37°C. Cells were pelleted by centrifugation at 4400 x g, resuspended in 50 mM Tris-Cl pH 6.8, 20 mM β-mercaptoethanol, 2% SDS, 10% glycerol and 10 mM imidazole, and 50 µl DNase1 (10 mg/ml), then incubated on ice for 30 minutes.
Debris was removed by centrifugation at 17,000 x g for 20 min, and the TcJAMYC protein was bound to Ni-NTA resin (Qiagen, Valencia, CA). The resin was washed with buffer containing 250 mM NaCl, 50 mM Tris-Cl pH 6.8, 20 mM imidazole, and eluted with buffer containing 250 mM NaCl, 50 mM Tris-Cl pH 6.8, 300 mM imidazole. The eluted protein was further purified using a Centricon YM-3 centrifugal filter device (Millipore, Danvers, MA) and brought to a final protein concentration of 25 ng/µl in 50% glycerol. Electrophoretic Mobility Shift Assays (EMSA): Oligonucleotide probes (see Table 1) contained a six nucleotide E-box at the center of a 22 bp sequence. A four-nucleotide 5'-overhang was included in each double stranded probe to allow for 32 P-labeling.
Results: Paclitaxel Pathway Gene Promoter Activation: Prior work in the Taxus system demonstrated a positive correlation between the steady state mRNA abundance of the paclitaxel biosynthetic pathway genes (hereafter referred to as "pathway genes") and MJ elicitation (7). This suggested that increased transcription of pathway genes was occurring in response to MJ elicitation. To investigate this possibility, promoters of five of the known pathway genes: T5αH (AY628434), DBAT (AF193765), PAM (AY582743), BAPT (AY082804) and DBTNBT (AF466397) were cloned (Figure 2A).
The DNA fragments encoding these promoters were placed upstream of the GUS (GUS) reporter gene to form translational fusion constructs containing the pathway promoter, the first few amino acids encoded by the pathway gene (T5αH: five amino acids; DBAT: six; PAM: ten; BAPT: seven; and DBTNBT, thirty), and the reporter gene.
These expression constructs were used to transform Taxus cuspidata P991 suspension culture cells by particle bombardment; the cells were either mock-elicited or elicited with 100 µM MJ.
To control for variable transformation efficiency, 35S:LUC was co-bombarded in all experiments ( Figure 2B). Reporter gene activity is presented as a GUS/LUC ratio to control for transformation efficiency between samples. Measurements of reporter activity were performed 48 h after bombardment. Expression of the 35S:GUS control construct ( Figure 2B) was not affected by addition of MJ ( Figure 3). For each pathway gene promoter construct, GUS enzyme activity was at least 2-fold higher when MJ had been added to the cultures with the exception of DBTNBT ( Figure 3). Thus, we concluded that the promoters of these pathway genes are activated by MJ. This result, in combination with increased amount of steady state mRNA for the endogenous genes after MJ elicitation (7), implies that MJ treatment causes increased transcription of pathway genes. Promoter Analysis: Since the most of the pathway gene promoters were activated upon MJ elicitation, and the pathway genes all respond in a uniform time course (7), we reasoned that there could be an MJresponsive trans-acting factor that influences these promoters. Considering this, an in silico analysis of the pathway promoters was performed using PLACE software (http://www.dna.affrc.go.jp/PLACE/) (28). This analysis revealed that there are 36 E-boxes (CANNTG) found in the 6544 nt of the collective pathway gene promoter sequence that we cloned (Figure 2A). A disproportionately large number of potential E-box sites are found in the promoters of the pathway genes, since only 25 E-boxes would be expected to occur at random in a sequence of this length. E-boxes have been found in defense gene promoters in other plants (29). bHLH proteins typically bind to these E-box nucleotide motifs (30). For example, the Arabidopsis bHLH MYC2 protein preferentially binds to the E-box sequence CACGTG. Furthermore, E-boxes are

WITHDRAWN AT AUTHORS REQUEST
commonly found on the promoters of genes that respond to MJ (19). The Solanum JAMYC2 binds a related, "G-box" sequence (AACGTG) on the proteinase inhibitor promoter (17). Cloning TcJAMYC: The large number of E-box binding sites in pathway promoters suggested that there could be a jasmonate responsive MYC regulator that binds to E-boxes and activates the pathway genes. The well conserved bHLH sequences from Arabidopsis MYC2 (At1g32640) and Solanum JAMYC10 (AJ630506) were used to design degenerate primers with the CODE-HOP program (27). These primers were used to amplify a 172 nt fragment was amplified from Taxus cDNA derived from suspension cultures that had been elicited with MJ for six hours ( Figure 4A). The amplified sequence displayed 94% identity at the amino acid level to the Solanum JAMYC bHLH region. Only one gene was cloned using this set of degenerate primers. Amplification from genomic DNA also produced the same gene fragment (data not shown). The Taxus fragment was used to obtain a full-length cDNA. The protein encoded by the full-length cDNA possesses 43% overall identity with the Solanum JAMYC, with 59% positives ( Figure 4A).
To clarify the relationship of the Taxus gene to the bHLH genes from Arabidopsis, an unrooted phylogram was generated from the Taxus DNA-binding domain (DBD) and all 154 Arabidopsis bHLH DBD sequences using maximum parsimony analysis ( Figure 4B). Families that were identified previously (31) are indicated. The Taxus bHLH falls into the same clade as AtMYC2, confirming the similarity of the two genes. Based on the similarity to the other well-characterized JAMYC bHLH transcription factors, we refer to the Taxus gene as TcJAMYC, for Taxus cuspidata jasmonate MYC.
Accumulation of mRNA from the previously characterized JA-MYC genes from tomato and Arabidopsis is positively regulated by MJ (17,19). To determine whether TcJAMYC responds to MJ application, semi-quantitative RT-PCR was performed to compare relative transcript abundance at one hour after MJ elicitation ( Figure  4C). This result indicates that there is an increase in TcJAMYC transcripts after MJ elicitation, again emphasizing the resemblance between TcJAMYC and the Arabidopsis JAMYC, AtMYC2.

Activation of Pathway Promoters by TcJAMYC:
To investigate whether TcJAMYC influences transcription of the paclitaxel pathway genes, co-bombardment experiments were performed using a full-length TcJAMYC cDNA under the control of the CaMV35S promoter in combination with the pathway gene promoter:GUS reporter constructs (see Figure  2A). Mock-elicited cells were bombarded with each pathway gene promoter:GUS construct either alone or in combination with 35S:TcJAMYC. TcJAMYC activated the T5αH, PAM, BAPT, and DBTBNT promoters by at least 2-fold ( Figure 5). Only the DBAT promoter was unaffected by cobombardment with TcJAMYC, although activation of the DBTNBT promoter was weak (< 2-fold).
This demonstrates that TcJAMYC positively regulates the promoters of many of the pathway genes that encode enzymes at various points along the paclitaxel biosynthetic pathway.
To determine the effect of MJ elicitation on the promoter-TcJAMYC interaction, co-bombardment experiments were performed using cells that had been elicited with 100 µM MJ six hours prior to bombardment. The elicited cells were bombarded with the promoter:GUS construct alone or in combination with the 35S:TcJAMYC. The T5αH promoter was highly activated when cobombarded with TcJAMYC into elicited cells ( Figure 5).
The DBAT, PAM, and BAPT promoters, however, were repressed following MJ elicitation despite co-bombardment with the 35S:TcJAMYC effector ( Figure 5). The terminal pathway gene, DBTNBT, showed a slight increase in activity with MJ elicitation when co-bombarded with the 35S:TcJAMYC effector ( Figure 5). Thus, the elicitation analysis revealed possible additional regulation of either the promoters by MJ, the TcJAMYC factor itself, or both.

TcJAMYC Binds to Pathway Promoters in vitro:
To determine whether the TcJAMYC protein physically interacts with the E-box elements found in the pathway promoters, electrophoretic mobility shift assays (EMSA) were performed using the TcJAMYC protein. EMSAs are used to determine whether a protein can bind to a specific DNA probe sequence in vitro. This binding is visualized as a retarded migration rate through a native polyacrylamide gel. 6X-HIS

WITHDRAWN AT AUTHORS REQUEST
tagged TcJAMYC was expressed in E. coli and purified on Ni-NTA agarose columns. The expected molecular weight for the tagged TcJAMYC protein is 73kD ( Figure 6A). To determine whether TcJAMYC binds to E-box elements, the most abundant E-box (CATGTG) in the pathway promoters was used initially as a probe sequence (Table 1). The TcJAMYC protein bound the CATGTG sequence ( Figure 6B). To determine if the binding was specific to TcJAMYC, the GUS protein, expressed in and purified from E. coli ( Figure 6A), was used in the EMSA. This assay demonstrated that the binding of the CATGTG sequence is specific to the TcJAMYC protein, as GUS did not bind to the DNA probe ( Figure 6B). Finally, a mutated DNA probe that does not contain an E-box (Table 1) was used in the EMSA ( Figure 6C). The TcJAMYC protein did not bind to this DNA sequence, demonstrating that the TcJAMYC protein specifically binds to the E-box element CATGTG.
To test the specificity of binding further, a competition assay was performed using the CATGTG probe against increasing amounts of non-specific DNA competitor (the mutated DNA described above). The binding efficiency to the CATGTG probe did not diminish ( Figure 6D), again demonstrating that the TcJAMYC protein specifically binds to the CATGTG sequence. Eight additional E-box sequences were found in the pathway promoters (Figure 2A). To determine whether TcJAMYC also binds these E-boxes, oligonucleotides containing these sequences were used in competition assays against the CATGTG probe. As shown in Figure 6E, the only E-box element that competed well against the CATGTG probe was the CACGTG element. This demonstrates that the TcJAMYC protein preferentially binds more strongly to two specific E-box sequences found in the pathway gene promoters: CACGTG and the CATGTG.
To test which of these two sequences, CATGTG or CACGTG, is most efficiently bound by TcJAMYC, competition assays between these two DNA elements were performed.
The CATGTG probe competed against itself, as previously shown in Figure 6E, and this was used as reference for binding specificity ( Figure 6F). A competition assay was performed using the CATGTG probe against increasing amounts of CACGTG competitor ( Figure 6G). The CACGTG fragment competed for binding of the CATGTG probe more effectively than CATGTG selfcompetitor (compare Figure 6F and 6G). This demonstrates that TcJAMYC preferentially binds to the CACGTG as compared to the CATGTG sequence. To further characterize this binding selectivity, a competition assay was performed with the CACGTG probe against the unlabeled CATGTG competitor ( Figure 6H). Increasing amounts of CATGTG competitor did not efficiently compete for binding, confirming the preference for binding of CACGTG. Thus, the TcJAMYC protein specifically binds to CACGTG and secondarily the CATGTG E-box elements.
Discussion: Transcriptional Activation of Pathway Promoters: MJ has been shown to induce terpene production in conifers (32), glucosinolates in Arabidopsis (33,34), alkaloids in Papaver (35) and Catharanthus (36), as well as proteinase inhibitors in many plant species including important agricultural crops (37)(38)(39). MJ elicitation has also been implicated in the activation of the defense related leucine aminopeptidase promoter in Nicotiana, the strictosidine synthase promoter in Catharanthus, and a sesquiterpene synthase promoter in Nicotiana are activated with MJ elicitation (17,36,40). In this investigation, we have shown that five paclitaxel biosynthetic pathway gene promoters are likewise activated by MJ elicitation. This result explains, at least in part, the previously observed increase in mRNA abundance for these genes that follows MJ elicitation (7), and indicates that the transcriptional response to MJ is conserved across land plant phylogeny from gymnosperms to angiosperms.
Our finding that the promoter of the gene encoding the penultimate step in paclitaxel biosynthesis (BAPT, cf. Figure 5) is strongly activated by MJ is somewhat surprising. The endogenous BAPT transcript was undetectable using RNA gel blot analysis, and up-regulation of mRNA expression from this gene was only detectable using a more sensitive RT-PCR assay (7). This difference in expression between the BAPT reporter construct and the endogenous BAPT gene implies a separate level of regulation of the endogenous gene, perhaps at the posttranscriptional level. Endogenous mRNA from WITHDRAWN AT AUTHORS REQUEST the terminal gene of the paclitaxel biosynthetic pathway, DBTNBT, does not accumulate to higher levels after MJ elicitation (7). Likewise, compared to the other pathway promoters, the DBTNBT promoter was only weakly affected by MJ (Figure 3). This suggests that manipulating the expression of the DBTNBT gene using a stronger promoter could allow increased enzyme production and increased flux through this point in the pathway.
The 35S promoter, which in this assay drove GUS expression as a control, exhibited a low level of expression compared to the pathway gene promoters, and was not affected by MJ. This may indicate that the CaMV35S promoter is not a strong promoter in Taxus and that other promoters should be investigated for the genetic engineering of Taxus cultures. A Jasmonate-Responsive bHLH in Taxus: By leveraging information obtained in model systems (Arabidopsis and Solanum) we have potentially obtained a key regulator of important natural product production. This transcription factor, TcJAMYC activates transcription of the genes in the paclitaxel biosynthetic pathway in the gymnosperm genus Taxus. Many of the enzymatic conversions in the pathway are unknown, so obtaining global regulators of the entire pathway may allow for manipulating the pathway in situ without the need to identify missing steps.
The degree of relatedness between TcJAMYC and AtMYC2, compared to all other bHLH proteins in Arabidopsis, suggests a conserved function across a wide span of evolutionary time ( Figure 4B).
There is a correlation between the presence of certain binding sites and the activation of specific promoters. TcJAMYC preferentially binds the sequences CACGTG and CATGTG, and at least three copies of these sequences are found in the highly activated promoters (Figure 2A and Figure 5). Among the promoters tested in our studies, only the DBAT promoter contains fewer than three copies of these sequences, and the DBAT promoter is the only one that is not upregulated when co-bombarded with TcJAMYC into Taxus cells. Since DBAT mRNA expression and DBAT promoter activity do increase following MJ application, there are likely additional regulatory factors that control MJinduced gene expression in Taxus.
Further work in this system is required for a complete description of the factors involved in MJ-elicited gene expression. A limitation of the Taxus system is that stable transformation is not yet possible for most of the commonly used, paclitaxel-producing cell lines, such as the P991 cell line used in this study. Stable transformation of Taxus suspension cultures with the constructs described here would further define whether TcJAMYC is able to cause increased taxane accumulation and could be used to demonstrate whether there is an in vivo interaction between TcJAMYC and the promoters of pathway genes. Negative Regulation of Pathway Promoters Following MJ Elicitation: In the studies presented here, we uncovered evidence for a negative regulation of the promoters of the paclitaxel biosynthetic genes.
In previous studies, we demonstrated that steady state mRNA levels for pathway genes increase by six hours after MJ elicitation, remain high for about 24-30 hours, and return to original low levels by about 48 hours (7). This rapid up-then down-regulation of the endogenous genes indicates positive regulation for the first 24 hours following elicitation, followed by negative regulation after the 24 h time point. In the studies presented here, the level of reporter gene expression was always measured 48 hours after bombardment, which is ~54 hours after MJ elicitation. Thus, when MJ elicitation was used, measurements were taken at a time when endogenous negative regulatory mechanisms may be operative. The measurements were made at this time because following transformation by bombardment, transcription and translation of both the effector construct (TcJAMYC) and the reporter constructs will take an additional period of time, approximately 12-24 hours. Furthermore, reporter activity was not consistent at time points prior to 48 hours post-bombardment, preventing use of earlier time points for measurements.
The promoters of three late pathway genes (DBAT, PAM, and BAPT) were unable to respond to the presence of the TcJAMYC effector when cells had been previously elicited with MJ ( Figure  5). Interestingly, the promoter of T5αH, an early pathway gene, did not demonstrate this negative regulation. Instead, this promoter was highly activated when elicited by both MJ and activated

WITHDRAWN AT AUTHORS REQUEST
TcJAMYC. We hypothesize that, in the bombardment assays reported here, negative regulators are acting to antagonize TcJAMYC regulation of the reporter constructs by 48 hours, thus preventing TcJAMYC's stimulatory effect on pathway promoter expression. It is both interesting and important that TcJAMYC alone (without MJ) appears not to induce this negative regulatory response ( Figure 5). Because of this effect, it may be possible to bypass negative regulation from MJ elicitation by engineering TcJAMYC into Taxus. This strategy may prolong pathway gene expression by allowing positive regulation to occur without the usual MJ-induced negative feedback loop that would usually follow. This could increase metabolic flux through the paclitaxel pathway.
(D) Competition assay with the TcJAMYC protein and the CATGTG probe against increasing amounts of mutated cold competitor at 0X, 1X, 2X, 4X, 8X, and 16X excess.
(E) Competition assays using TcJAMYC protein and the CATGTG radio labeled probe against all other E-box elements found in the pathway promoters. The cold competitor is listed to the right of the panels. +0: no competitor, 20X Self: the CATGTG cold competitor at 20X excess, 20X Comp: the cold competitor at 20X excess.
(G) A competition assay between radio labeled CATGTG against the CACGTG cold competitor.
(H) A competition assay between radio labeled CACGTG against the CATGTG cold competitor.