Intronic Sequence Regulates Sugar-Dependent Expression of Arabidopsis thaliana Production of Anthocyanin Pigment-1/MYB75

Sucrose-specific regulation of gene expression is recognized as an important signaling response, distinct from glucose, which serves to modulate plant growth, metabolism, and physiology. The Arabidopsis MYB transcription factor Production of Anthocyanin Pigment-1 (PAP1) plays a key role in anthocyanin biosynthesis and expression of PAP1 is known to be regulated by sucrose. Sucrose treatment of Arabidopsis seedlings led to a 20-fold induction of PAP1 transcript, which represented a 6-fold increase over levels in glucose-treated seedlings. The PAP1 promoter was not sufficient for conferring a sucrose response to a reporter gene and did not correctly report expression of PAP1 in plants. Although we identified 3 putative sucrose response elements in the PAP1 gene, none were found to be necessary for this response. Using deletion analysis, we identified a 90 bp sequence within intron 1 of PAP1 that is necessary for the sucrose response. This sequence was sufficient for conferring a sucrose response to a minimal promoter: luciferase reporter when present in multiple copies upstream of the promoter. This work lays the foundation for dissecting the sucrose signaling pathway of PAP1 and contributes to understanding the interplay between sucrose signaling, anthocyanin biosynthesis, and stress responses.


Introduction
Regulation of genes by sugars such as glucose and sucrose has been well documented in plants and plays an important role in plant growth, development, and physiology [1][2][3][4][5][6]. Although it can be difficult to attribute signaling responses to a specific sugar molecule as opposed to one of its metabolites or catabolites, there is strong evidence for distinct signaling pathways for both glucose and sucrose [7][8][9]. Compared with glucose, sucrose-specific signaling pathways are not well characterized, and a sucrose-specific sensor functioning similar to that of hexokinase in glucose signaling has not been identified although several have been proposed [3,7].
Genes regulated specifically by sucrose are generally defined as those that do not respond in a similar fashion to glucose or fructose. The best known examples of sucrose regulated genes are the Beta vulgaris sucrose transporter BvSUT1 from sugar beet [5,9], patatin from potato [10,11], and the Production of Anthocyanin Pigmentation-1 PAP1/MYB75 transcription factor that regulates anthocyanin biosynthesis in Arabidopsis thaliana [12,13]. For these genes, there is minimal regulation in the presence of glucose or fructose. Other genes such as chalcone synthase [14], rolC [15], and beta amylase [16,17] respond to glucose and other sugars yet show a stronger response to sucrose, while for others comparable responses have been observed for both sucrose and glucose [18][19][20]. It should also be noted that in many cases differences in gene responses to sugars have been reported. For example, glucose and fructose-induced expression of potato proteinase inhibitor II was reported to be equal to or half the response observed with sucrose [21,22]. And higher expression of PAP1 in glucose versus sucrosetreated seedlings has also been reported [23]. This is most likely due to differences in how experiments were performed and/or underlying metabolic changes in sugar content and identity.
For some of these genes sugar response elements or sequences important for sugar regulation have been identified and nuclear binding proteins or activities characterized [4]. Genetic approaches have led to identification of genes that play important roles in regulating sugar expression of specific genes [24,25]. However, overall the molecular pathways involved in sucrose-signaling are not well understood.
We are using Arabidopsis PAP1, a known sucrose-regulated gene, as a model to dissect sucrose signaling. Teng et al first identified PAP1 as important for sucrose-induced anthocyanin accumulation in Arabidopsis and showed that a PAP1 knock-out mutant lacked this response [12]. Solfanelli et al further showed that expression of PAP1 was specific to sucrose and not glucose or fructose [13], although a more recent paper reported higher PAP1 expression in glucose-treated seedlings compared to sucrose-treated seedlings [23]. Sucrose-induced expression of PAP1 has been shown to be modulated by hormones [23,[26][27][28], mutations in the AtSUC1 sucrose transporter [29], and alterations in calcium signaling [30]. In many cases it is not clear if these factors directly impact PAP1 expression or are secondary due to changes in endogenous sugar content. Although there have been several reports on the expression of PAP1 in response to sugars, sequences important for the sucrose-responsiveness of PAP1 have not been identified.
To begin to dissect the sucrose-regulated pathway of PAP1 we have localized sequences important for this response to an intron and further show that this intronic sequence can confer sucrose-responsiveness to a reporter gene. We also discuss the implications of PAP1 regulation by sucrose and how this might contribute to regulation of anthocyanin biosynthesis in plants.

Sucrose specificity of anthocyanin biosynthesis
To confirm the sucrose-specific induction of anthocyanin biosynthesis, Arabidopsis Col-0 seeds were germinated in one-half-strength liquid Murashige and Skoog (MS) media and seedlings were treated with water or various sugars after 4 days of growth in continuous light. Seedlings were harvested 48 hours later and anthocyanins quantified (Fig 1A). Sucrose treated seedlings accumulated the most anthocyanin, although glucose treated seedlings also accumulated anthocyanin relative to the water treated control. These results are similar to those reported by Teng et al [12]. Expression of PAP1 was examined by RT-qPCR 24 hrs after addition of water or various sugars ( Fig 1B). PAP1 transcript abundance paralleled anthocyanin accumulation; PAP1 levels were 20-and 3.6-fold greater after sucrose and glucose treatment, respectively.

Identification of sequences important in conferring sucrose responsiveness of PAP1
To define sequences important in governing the sucrose-induced expression of PAP1, we generated PAP1 promoter: reporter fusions with β-galactosidase and luciferase (PAP1pro:GUS or LUC). The promoter sequence was defined as 2006 bp upstream of the PAP1 start codon. A known sucrose response element (SRE), SURE-2, was identified in the promoter of PAP1 using the Plant cis-acting regulatory DNA element (PLACE) database [31]. SURE-1 (TTTTCTATT) and SURE-2 (AATACTAAT) elements were initially identified as important in conferring sucrose responsiveness of patatin, a lipolytic acyl hydrolase from potato [32]. SURE-1 and SURE-2 share sequence similarity with one another (5 out of 9 identical nucleotides) and with SP8 elements [32,33]. The similarity between the SURE-2 elements in the PAP1 and patatin promoters extended beyond the 9 bp core element to 17 bp with only one discrepancy (patatin sequence: TATATAATACTAATAAA; PAP1 sequence has a T in place of the underlined A).
Constructs were transformed into Col-0 and transgenic plants were generated. For these and all subsequent experiments, 5-10 single copy homozygous lines were generated per construct and data is shown for 2-4 representative lines per construct. When plants were tested for sucrose-induced GUS or LUC activity as described above, no activity was observed (Fig 2A and  2B). RT-PCR of sucrose-treated PAP1pro:GUS plants showed that even though PAP1 transcript was present, the GUS transcript was not expressed in response to sucrose (Fig 2C). Several genes, including sucrose transporters, have been shown to require additional sequences contained within exons and introns for proper gene expression [34]. To test whether sequences in the exons and introns are important for sucrose-induced expression of PAP1, we fused the PAP1 promoter plus all exons and introns to GUS or LUC (PAP1wg:GUS or LUC) ( Fig 3B). Transgenic plants harboring these constructs were generated and tested for GUS and LUC activity. Both reporter constructs showed sucrose-dependent reporter gene activity (Fig  2A and 2B). Furthermore, the PAP1wg:GUS lines had GUS staining at the base of the cotyledon and the upper part of the hypocotyl, reminiscent of anthocyanin pigmentation in seedlings (Figs 1C and 2A). RT-PCR of these plants showed that the GUS transcript was expressed in response to sucrose, similar to the expression of the endogenous PAP1 transcript (Fig 2C).

PAP1 gene expression during plant development
To determine whether PAP1pro:GUS plants correctly reported PAP1 expression at other developmental stages we compared the pattern of GUS expression of PAP1pro:GUS and PAP1wg: GUS lines. Seeds were germinated on MS media plus 2% sucrose and plants were harvested at In flowering plants (28-42 days old) there was some variation in the pattern and intensity of GUS staining in different experiments that may be due to differences in plant development or environmental conditions. In all plants examined, PAP1pro:GUS lines showed GUS staining in petals, ovaries, and filaments of mature flowers and in some experiments staining was also observed in stigma and anthers (Fig 4C and 4K). Interestingly, in PAP1pro:GUS lines flower buds had little to no staining and immature flowers only showed a discrete band of staining just under the stigma, suggesting that the observed GUS expression pattern may be under developmental control (Fig 4C, 4I and 4J). Siliques were usually stained, but seeds were not ( Fig 4D). In some experiments GUS staining could also be observed in cauline leaves (weak staining along the major vein), secondary inflorescences, and pedicels close to the junction with flowers and siliques. No staining was observed in rosette leaves (Fig 4A), stem, primary inflorescences, and sepals ( Fig 4K).
In comparison, GUS staining of PAP1wg:GUS lines was more uniformly distributed and present in most organs including rosette and cauline leaves (most notably along veins), flowers, and pedicels (Fig 4E-4G). In contrast to the staining observed in PAP1pro:GUS lines, PAP1wg: GUS lines showed GUS staining in sepals, but not petals ( Fig 4L). The lack of GUS staining in petals is consistent with the lack of anthocyanins in Arabidopsis petals [35]. GUS staining was also observed in ovaries and filaments in flowers, but not stigma or anthers ( Fig 4L). In some experiments staining was observed in stem and primary and secondary inflorescences, and siliques ( Fig 4H). No GUS staining was observed in seeds.
GUS staining and GUS transcript observed in PAP1wg:GUS lines correlated well with native PAP1 transcript as detected by RT-PCR, which showed PAP1 transcript in rosette and cauline leaves, inflorescence, flower buds, flowers and siliques ( Fig 4M). PAP1pro:GUS lines only showed GUS transcript in flowers, but not in any other tissues. These results suggest that the PAP1 promoter is not correctly reporting PAP1 expression in Arabidopsis. Our expression data for PAP1wg:GUS lines correlates well with expression data in the AtGenExpress atlas of Arabidopsis development [36]. In this dataset expression values of PAP1 are fairly low (with the exception of senescing leaves), but expression is detected in rosette and cauline leaves, 7 day old seedlings, and stage 15 flowers (including petals, stamen, and carpels). Taken together, the whole gene construct is not only necessary for the response to sucrose as it contains sequences necessary for controlling PAP1 expression as a function of tissue and development.
The expression pattern of a similar PAP1pro:GUS construct has been reported [37]. GUS expression was reported in hypocotyls and cotyledons of 2-3 day old seedlings and young emerging leaf tissue/primordia of 5-7 day old seedlings. As discussed above, we did not observe any GUS staining in seedlings of our PAP1pro:GUS lines when grown under our conditions or those reported by Gonzalez et al.

Importance of intronic sequences in conferring sucrose responsiveness of PAP1
To further define sequences important in controlling sucrose-dependent expression of PAP1, we generated additional reporter gene constructs lacking one or both introns of PAP1 (PAP1wgΔInt1:GUS or LUC, PAP1wgΔInt2:GUS or LUC, PAP1wgΔInts:GUS or LUC) (Fig 3C).
Transgenic plants were generated and tested for reporter gene activity. Constructs lacking intron 1 or both intron 1 and 2 (PAP1wgΔInt1:GUS or LUC and PAP1wgΔInts:GUS or LUC) did not have reporter gene activity in response to sucrose (Fig 5A and 5B). This was confirmed by RT-PCR, which showed no GUS transcript expressed in response to sucrose (Fig 5C). However, the construct lacking intron 2 (PAP1wgΔInt2:GUS) retained sucrose-induced reporter gene activity and GUS expression in response to sucrose. The pattern of GUS staining observed for PAP1wgΔInt2:GUS in seedlings and plants was identical to that observed for the construct containing the entire PAP1 gene (PAP1wg:GUS) (Figs 2A & 5A). However, the pattern of GUS staining in PAP1wgΔInt1:GUS and PAP1wgΔInts:GUS plants was much more limited. In these plants staining was observed in the major vein of rosette leaves with more intense staining near the base of the rosette and weaker staining towards the tip of the leaves (S4 Fig). Staining was also observed at the base of cauline leaves at the attachment point to the inflorescence. No other staining was observed in these lines.

Role of SURE-1 in the sucrose-response
Having identified intron 1 as important for the sucrose response of PAP1, we next used PLACE to scan intron 1 for known cis-elements and identified the SURE-1 sucrose response element (TTTTCTATT) (Fig 3A). Interestingly, we also identified another SURE-2 element (AATACTAAT) within intron 2, which was shown to be dispensable for the sucrose response.
To determine whether the SURE-1 element in intron 1 was necessary for the sucrose-responsiveness of PAP1, we generated 2 constructs in which the SURE-1 element was directly mutagenized. For intron1m1, the SURE-1 element (TTTTCTATT) was mutated to TTTGAGATT according to Grierson et al, who showed that this mutation abolished binding of a protein from nuclear extracts of potato tubers in electrophoresis mobility shift assays [32]. In the second construct, intron1m2, 4 nucleotides of the SURE-1 element were mutagenized to convert the SURE-1 element (TTTTCTATT) to the SURE-2 element (AATACTAAT). The SURE-2 element is present in the promoter and intron 2 of PAP1 and does not appear to be essential for the sucrose response as described above. The mutations were incorporated into PAP1wg:GUS constructs and neither mutation was observed to alter splicing. Both mutants retained sucroseinduced GUS activity and GUS expression in response to sucrose (Fig 6). This suggests that the SURE-1 element in intron 1 is not important in conferring sucrose responsiveness to PAP1.
Deletion analysis defines a 90 bp region in intron 1 as important for the sucrose-response of PAP1 In order to identify sequences in intron 1 that are important for the sucrose-response of PAP1 we deleted regions of intron 1 within the context of the PAP1wg:GUS construct. Intron 1 is 540 bp and initially we tested two constructs with non-overlapping 205 bp deletions (intron1m3 and intron1m4, Fig 6). These deletions did not include the first 50 or last 80 nucleotides of the intron so as not to disrupt splicing, which was verified by RT-PCR for all constructs. Nucleotides 51-255 of intron 1 (including the SURE-1 element) were deleted in intron1m3 and nucleotides 256-460 were deleted in intron1m4 (Fig 6). The intron1m3 construct retained sucroseinduced GUS activity and expression, while no GUS expression or activity was observed in sucrose-treated seedlings of intron1m4 (Fig 6). These results further verify that the SURE-1 element is not important for the sucrose-response of PAP1 because removal of the SURE-1 element (in intron1m3) did not affect the sucrose-response.
Smaller (40-60 bp) deletions were made within the 205 bp region deleted in intron1m4 and these constructs were tested for sucrose responsiveness (intron1m5-8, Fig 6). In addition, two potential cis-elements within this 205 bp region were also mutagenized (intron1m9-11, Fig 6). These potential cis-elements included a 10 nt palindromic G-box sequence with similarity to an abscisic acid response element (G/ABRE consensus: (T/C)ACGTG(T/G)C, PAP1 sequence: AACACGTGTT) [38] and a sequence resembling Coupling Element 3 (ACGCGTGTCCTC) that has been found to be associated with ABREs (PAP1 sequence: ACGCTTGTCTTC) [39,40] (Fig  6). Both of these elements were mutated individually (intron1m9 and 10) as well as together (intron1m11). All three of these constructs retained a sucrose-response, indicating these putative elements are not important for sucrose-dependent expression (Fig 6). Deletion constructs intron1m5 and 6 also retained sucrose-responsiveness, but sucroseinduced GUS activity was absent in constructs intron1m7 and intron1m8. RT-PCR and sequencing confirmed that splicing was not affected in these mutants. This identified a 90 bp region of intron 1, that when deleted (in constructs intron1m7&8), abolished the sucrose-response.

A 90 bp region is sufficient for conferring a sucrose response
To determine if this 90 bp region in the intron of PAP1 is sufficient to confer a sucrose response, we cloned 1-3 copies of this region upstream of the 35S minimal promoter and LUC reporter gene (5'-1X, 5'-2X, 5'-3X; Fig 7). We also cloned the same fragments downstream of LUC (3'-1X, 3'-2X, 3'-3X; Fig 7) to test whether the spatial context of the element was important. Relative to the empty yy449 vector, only the 5'-3X construct had a significantly higher sucrose response although there was a trend of increased sucrose response as the number of 90 bp elements increased. Constructs with the 90 bp element 3' of LUC did not show a significant sucrose response. One to 3 copies of the 90 bp SRE was inserted upstream (5') of the 35S minimal promoter or downstream of the LUC coding sequence (3') in the vector yy449. B. Sucrose response of transgenic seedlings harboring the empty vector (yy449) or 1-3 copies of the 90 bp SRE in either the 5' or 3' insertion site. Error bars represent standard deviation of 2-4 independent single-copy homozygous transgenic lines. Significance was calculated using ANOVA to examine the influence of construct on the sucrose to water ratio. After determining there was a significant influence of construct, a post-hoc Tukey's HSD was performed using a confidence level of 0.05.

Intronic sequences are important in regulating the sucrose responsiveness of PAP1
Given the fundamental roles that sugars play in plant metabolism, understanding how sugars also serve as signaling molecules is vital towards gaining a comprehensive understanding of plant growth and development. Sugars such as glucose and sucrose are known to regulate the expression of large numbers of genes as evidenced by reporter gene studies [11, 14-16, 18, 19, 21, 22, 41-44] and large scale gene expression analyses [4,[45][46][47][48][49][50]. We used reporter gene assays in combination with mutant approaches to identify sequences important for regulating the sucrose response of PAP1. Two notable findings have come from this work. First, the sequence shown to be necessary and sufficient for the sucrose response of PAP1 was located in an intron. Although there are many genes that are known to be regulated by intronic sequences, including the floral Arabidopsis MADS box gene AGAMOUS [51,52], and several sucrose transporters including AtSUC1, AtSUC9, and LeSUT1 [34,53], there are no reports of sugar response elements in an intron. Furthermore, bioinformatic approaches to identify cis elements typically limit searches to upstream regions [54,55]. Second, despite the presence of 3 known SUREs in the promoter and first and second introns of PAP1, none were found to be important. The promoter of PAP1, which contains a SURE-2 element with extended identity to the SURE-2 element of patatin was not sufficient for conferring sucrose responsiveness. And removal of the SURE-2 in intron 2 (by deletion of the entire intron) did not abolish the sucrose response of PAP1. In addition, mutation of SURE-1 did not abolish the sucrose response. Although it is possible that our two mutations of SURE-1 failed to alter the element sufficiently for complete loss of function, complete removal of SURE-1 (in construct intron1m3) does not affect the sucrose response thus demonstrating this element is dispensable for sucrose regulation. Although SURE-1 has been implicated in regulating sucrose responsiveness in numerous genes [19,32,43,44,[56][57][58][59], in at least one other case SURE elements were not important [60]. Furthermore, the presence of SURE elements in promoter sequences of Arabidopsis genes was not or only weakly correlated with the regulation of these genes by sucrose [54].
When placed outside the context of an intron upstream of the 35S minimal promoter, the 90 bp sequence was able to function as a SRE, although it only had a robust sucrose response when present in 3 tandem copies. The sucrose response of constructs with one copy of the 90 bp region was not significantly different from the empty vector unlike what was observed in PAP1wg lines, which contain one copy of the 90 bp region in the context of intron 1. This indicates that other sequences may be required for a robust sucrose-response and/or that the spatial context of the 90 bp sequence is important. Evidence for the latter is suggested by the fact that the 90 bp sequence did not function when placed downstream of LUC (in 3' constructs). In these constructs the element was located 1700 bp downstream of the start codon, while in the context of the PAP1 gene the element is 491 bp downstream of the start codon. The difference in distance of the element relative to the transcriptional start may be one reason that the element did not function effectively.

Contribution of sucrose and other regulatory elements in the expression of PAP1
The correlation between sucrose responsiveness and pattern of GUS staining in our transgenic plants suggests that sequences conferring sucrose responsiveness are also responsible for the observed expression pattern of PAP1 in plants. Removal of intron 1 (in constructs PAP1wg-ΔInt1:GUS and PAP1wgΔInts:GUS) severely limits GUS expression quantitatively and qualitatively as compared to PAP1wg:GUS lines. Furthermore, PAP1wgΔInt2:GUS and both SURE-1 mutants (intron1m1 and intron1m2) retained sucrose responsiveness and both had GUS staining patterns nearly identical to PAP1wg:GUS lines. The ability of sucrose to govern the expression of PAP1 may be related to the protective role of anthocyanins during stress responses. For instance, stress conditions such as high light and cold are known to increase sucrose content and accumulation of anthocyanins in these conditions may serve to protect cells from damage by virtue of their ability to absorb light and serve as antioxidants and osmoprotectants [63][64][65][66].
Differences in the expression pattern between PAP1pro:GUS and PAP1wg:GUS lines further underscores the complexity of PAP1 regulation. PAP1wg:GUS plants showed GUS staining that was reminiscent of anthocyanin pigmentation in seedlings and plants (hypocotyls in young seedlings and leaves with noticeable staining along major veins) and in several other tissues as well (cauline leaves, inflorescences, immature and mature flowers, siliques). RT-PCR confirmed that PAP1 transcript was present in all of these organs and tissues in wildtype Col-0 plants. The pattern of GUS staining in PAP1pro:GUS was remarkably different and nearly opposite that of the PAP1wg:GUS plants. We first observed staining in the roots of 20 day old PAP1pro:GUS plants, but this was at a different location than the root staining observed in PAP1wg:GUS plants. No rosette staining was observed in PAP1pro:GUS lines while strong rosette expression was observed in PAP1wg:GUS lines. Both lines had GUS expression in mature flowers but PAP1pro:GUS lines had strong petal and no sepal staining while PAP1wg:GUS lines had strong sepal and no petal staining. Because the PAP1 promoter sequence is present in the PAP1wg:GUS construct, the discrepancy in staining patterns may be explained by the presence of sequences within the introns and/or exons of PAP1 that repress petal staining. Repressor elements are most likely within the exons of PAP1 because petal staining is not observed in the PAP1wgΔInts:GUS line, which contains the PAP1 promoter plus exons, but no introns. Furthermore, because the PAP1wgΔInts:GUS line had very limited GUS staining that was nothing like the staining observed in PAP1pro:GUS lines, PAP1 exons must repress all promoter expression elements and contain sequences conferring the weak expression observed in rosette and cauline leaves.

Conclusions
A 90 bp intronic sequence was identified as important for regulating the sucrose response of PAP1 and was sufficient for conferring a sucrose response to a reporter gene when present in multiple copies. Intronic sequence was also shown to be important for the proper pattern of PAP1 gene expression. Pigmentation of flowers, fruits, and vegetables remains a target for crop improvement [67] and understanding the many layers of PAP1 regulation may lead to new approaches for modifying anthocyanin content of these plants.

Sucrose treatment
Surface-sterilized seeds were put into a 250 ml flask containing 50 ml of sterilized ½-strength MS basal salts pH 5.7 (Phytotechnology Laboratories, Shawnee Mission, KS). Seeds were kept at 4°C for 2 days then transferred to a continuous light or dark chamber with shaking at~100 rpm. Four days after transfer to the chamber, sucrose (or an equivalent volume of water) was added to 90 mM from a 1.8 M stock. Seedlings were harvested at the indicated timepoints following addition of either sucrose or water.

RT-qPCR
Total RNA was extracted using the RNeasy Plant mini kit (Qiagen, Valencia, CA) and treated with TURBO DNA-free kit (Ambion, Austin, TX). One μg of RNA was reverse transcribed using iScript Reverse Transcription Supermix (BioRad, Hercules, CA) according to the manufacturer's protocol. Twenty μl reactions were prepared using 1 ul of cDNA, 0.5 to 0.8 uM primer, and SYBR Premix Ex Taq II (Perfect Real Time) reagent (Takara Bio Inc, Otsu, Shiga Japan). All reactions were performed in triplicate. PAP1 and ubiquitin primer sequences are from Solfanelli et al [13]. Amplification was carried out on a Roche 480 Light Cycler (Roche, Madison, WI) with the following cycling parameters: One cycle of 95°C for 5 min followed by 45 cycles of amplification (95°C for 10 s: 60°C for 10 s; 72°C for 10 s). Cq values were calculated using the Light Cycler 480 SW 1.5 software and the Absolute quantification/2 nd derivative method. Primer efficiencies were calculated from the slope of Cq values plotted as a concentration of cDNA (ranging from 0.01-1 μg) and were 102 and104% for PAP1 and ubiquitin, respectively. Cq values and primer efficiencies were input into the Relative Expression Software tool (REST) 2009 version [68] and fold changes were determined relative to ubiquitin.

Anthocyanin quantification
Anthocyanin quantification was modified from Solfanelli et al [13]. Anthocyanins were extracted in 1 ml 1% HCl in MeOH for 16-24 hours at 4°C. 0.9 ml of the extract was combined with 0.9 ml water and 1 ml chloroform. Extracts were vortexed and clarified by low speed centrifugation. 0.9 ml of the aqueous phase was removed and Abs 530 and 657 were determined. Anthocyanin content is reported as Abs 530-657/mg fresh weight.

PAP1 promoter, whole gene, and delta intron-reporter constructs
The 2 kb 5'-upstream region of the PAP1 gene (At1g56650) was amplified from A. thaliana genomic DNA (Col-0) using the Expand High Fidelity PCR system (Roche Applied Science, Indianapolis, IN) and the following primers, which contain restriction enzyme sites (underlined): PAP1HindIIIF: 5'-AACACTACAAAAAAGCTTAACTGCATTTAG-3' and PAP1proS-maIR: 5'-TGGACGAACCCTCCCGGGAACAAAGATAG-3' or PAP1HindIIIF and PAP1HindIIIR: 5'-TGGACGAACCCTAAGCTTAACAAAGATAG-3'. A 3379 bp fragment containing the 2 kb 5'-upstream region and all exons and introns of PAP1 gene was amplified using PAP1HindIIIF and PAP1wgSmaIR: 5'-CAAATGTTCGAACCCGGGATCAAATTTCAC-3' or PAP1wgHindIIIR: 5'-CAAATGTTCGAAAAAGCTTTCAAATTTCAC-3'. The whole gene constructs were designed to be translationally fused to either the GUS or LUC ORF and the insertion of a HindIII site for cloning the whole gene construct into pLPTV-BAR resulted in a change of the last amino acid of PAP1 from Asp to Glu.
PCR products were cloned into the pDrive vector using the Qiagen PCR Cloning Kit (Qiagen, Valencia, CA) and sequenced. pDrive containing the 2 kb PAP1 promoter amplified with PAP1HindIIIF and PAP1proSmaIR was first digested with Sma I and the resulting 3676 bp fragment (containing the 2 kb PAP1 promoter plus vector sequence) was isolated. This fragment was partially digested with Hind III and a 2005 bp fragment was isolated and cloned into the Sma I and Hind III sites of pBI101.3. pDrive containing the 2 kb PAP1 promoter amplified with PAP1HindIIIF and PAP1proHindIIIR was partially digested with Hind III and the resulting 2005 kb fragment was isolated and cloned into the Hind III site of the pLPTV-BAR vector. pDrive containing the whole gene PAP1 sequence amplified with PAP1HindIIIF and PAP1wgSmaIR was first digested with Sma I and the resulting 5543 bp fragment was isolated. This fragment was partially digested with Hind III and a 3379 bp fragment was isolated and cloned into the Sma I and Hind III sites of pBI101.3. pDrive containing the whole gene PAP1 promoter sequence amplified with PAP1HindIIIF and PAP1wgHindIIIR was partially digested with Hind III and the resulting 3379 kb fragment was isolated and cloned into the Hind III site of the pLPTV-BAR vector. Sequence identity and orientation were confirmed by sequencing.
To generate PAP1wg constructs lacking one or both introns, a full length PAP1 cDNA was amplified from Arabidopsis seedlings. To create PAP1wgΔIntron2, a 531 bp Age I-Spe I fragment (containing Intron 2) was removed and replaced with a 442 bp Age I-Spe I fragment from PAPI cDNA. To create PAP1wgΔIntron1, a synthetic DNA fragment was constructed from the Sca I site in the 5'-upstream region of the PAP1 gene to the Age I site in Exon 2 and omitted Exon 1 (GENEART, Regensburg, Germany). This 430 bp Sca I-Age I fragment was used to replace the corresponding 969 bp fragment from PAP1wg and PAP1wgΔIntron2 to create PAP1wgΔIntron1 and PAP1wgΔIntrons1&2, respectively.
For the intron1m5-8 deletions, synthetic DNA fragments were constructed in the context of the Sca I-Age I fragment that encompasses intron 1, and incorporating the specific deletions shown in Fig 5 (GENEART, Regensburg, Germany). These Sca I-Age I fragments were used to replace the corresponding 969 bp fragment from PAP1wg.

Minimal promoter constructs
For the 5' constructs, a synthetic DNA fragment was constructed with the 90 bp region of interest flanked by HindIII and BamHI (5') and BglII (3') restriction sites (GeneArt 1 Gene Synthesis by Life Technologies): AAGCTTAAAGGGATCCtaaatgaattcgtgggaaaattttgtatg aacacgtgtttctgtgtt ggaacagttctttatttttattggtgtgcatagattcttcct gAGATCT. This fragment was digested with HindIII and BglII and ligated into HindIII and BamHI-digested yy449 vector (Genbank accession AB638628.1). This process was repeated once or twice more to obtain plasmids with two (192 bp) or 3 (288 bp) copies of the 90 bp region. For the 3' constructs, 1-3 copies of the 90 bp fragment were synthesized with flanking XbaI sites. Multiple copy elements were separated with the spacer AGATCC. These fragments were cloned into the XbaI site of yy449.

Generation of transgenic plants
Plasmids were transformed into Agrobacterium tumefaciens strain GV3101 by electroporation. A. thaliana ecotype Col-0 was transformed by the floral dip method as described [72]. Seeds were surface-sterilized by treating with 95% ethanol for 10 min followed by 20% bleach supplemented with 0.1% Tween-20 for 5 min, and rinsed several times in sterile water. Seeds were suspended in 0.1% agargel and plated on MS plates supplemented with 50 μg ml -1 kanamycin (pBI101.3) or sowed directly on soil and sprayed with Liberty herbicide (pLPTV). Resistant plants were checked by PCR for the presence of the transgene.

Luciferase assay
Seedlings ground to a fine powder and luciferase activity was measured using the Luciferase Assay System (Promega, Madison, WI) according to the manufacturer's protocol. Protein content in cell extracts was measured by Bradford assay using BSA as standard.