MYC cis-Elements in PsMPT Promoter Is Involved in Chilling Response of Paeonia suffruticosa

The MPT transports Pi to synthesize ATP. PsMPT, a chilling-induced gene, was previously reported to promote energy metabolism during bud dormancy release in tree peony. In this study, the regulatory elements of PsMPT promoter involved in chilling response were further analyzed. The PsMPT transcript was detected in different tree peony tissues and was highly expressed in the flower organs, including petal, stigma and stamen. An 1174 bp of the PsMPT promoter was isolated by TAIL-PCR, and the PsMPT promoter::GUS transgenic Arabidopsis was generated and analyzed. GUS staining and qPCR showed that the promoter was active in mainly the flower stigma and stamen. Moreover, it was found that the promoter activity was enhanced by chilling, NaCl, GA, ACC and NAA, but inhibited by ABA, mannitol and PEG. In transgenic plants harboring 421 bp of the PsMPT promoter, the GUS gene expression and the activity were significantly increased by chilling treatment. When the fragment from -421 to -408 containing a MYC cis-element was deleted, the chilling response could not be observed. Further mutation analysis confirmed that the MYC element was one of the key motifs responding to chilling in the PsMPT promoter. The present study provides useful information for further investigation of the regulatory mechanism of PsMPT during the endo-dormancy release.


Introduction
The mitochondrial phosphate transporter (MPT) shuttles inorganic phosphate (Pi) into the mitochondrial matrix, where Pi is utilized for oxidative phosphorylation to synthesize ATP from ADP. MPT encoding genes have been cloned from mammals [1][2][3], yeast [4], and wood frogs [5] with most studies focusing mainly on the structure and catalytic function of the transporters.
Recently, cloning and characterization of MPT were reported in several plants [6][7][8][9][10][11][12]. Plant MPT genes were identified to be involved in abiotic stress responses, and their expression patterns showed tissue preferences. Birch Mpt1 was ozone-inducible and highly expressed in the tissue of dividing cells, such as root tips, shoot apices and developing root nodules [6]. AtMPTs play an important role in response to salt stress in Arabidopsis. Furthermore, with different expression profiles in various tissues and conditions, transcription of AtMPTs has been detected in all tissue except siliques [12]. The sequences and structures of 26 potential PT family genes in rice were analyzed, and six MPTs also showed tissue preferential expression profiles, among which OsPT17 and OsPT19 were differently regulated under hormone treatment conditions. In addition, six putative cis-elements were found in all of the OsPT genes including ARR1AT, CAATBOX1, CACTFTPPCAL, GATABOX, GT1CONSENSUS and GTGANTG10. Specifically, GATABOX and GT1CONSENSUS are light-responsive cis-elements, and CACTFTPPCAL is necessary for carbon metabolism [11]. Current knowledge of MPT regulation and the molecular mechanisms mediating its biological functions in plants is still incomplete.
Tree peony (Paeonia suffruticosa Andrews) is one of the most well-known horticultural and medicinal plants in the world. One of the main production mechanisms in the tree peony industry, especially for the Spring Festival flower market in China, is forcing culture. Dormancy is a major obstacle for the forced culture of tree peony in winter, and sufficient chilling is an efficient way to break dormancy. Therefore, it is important to determine how chilling induces dormancy release in tree peony. PsMPT was previously isolated from the tree peony subtractive cDNA library of burst buds and strongly induced by chilling treatment to promote ATP production during the release of bud dormancy. In addition, ectopic-expression of PsMPT in Arabidopsis showed that PsMPT enhanced ATP synthesis and affected plant growth and development [10]. These results suggested that PsMPT plays an important role in energy production during bud dormancy release in tree peony [13]. However, the expression characteristics of PsMPT and its regulatory mechanisms are unclear.
In this study, we isolated the promoter of PsMPT and constructed PsMPT promoter::GUS engineered Arabidopsis. We investigated: 1) the temporal and spatial characteristics of the PsMPT promoter in Arabidopsis and PsMPT expression in tree peony; 2) how plant hormones and abiotic stresses, including chilling, affects the activity of the PsMPT promoter; 3) which one of cis-elements among the PsMPT promoter is involved in the chilling response.

Plant materials
Four-year-old tree peonies (Paeonia suffruticosa 'Luhehong') were obtained from the Tree Peony Research Center of Heze (Shandong, China). According to the method of Huang et al. [10], plants were treated in cold conditions (0-4°C) for 21 days to break bud dormancy, as the daily mean temperature was under 10°C in Qingdao, Shandong, China. The plants were then transferred to a greenhouse (18-22°C, 8-h-light/16-h-dark cycle) to resume growth. Tissues (root, stem, leaf, calyx, petal, stamen and carpel at the early stage of flowering) were collected and stored at -80°C until use. One hundred μmolÁL -1 ABA and 50 μmolÁL -1 GA 3 were applied to non-chilling buds with double-distilled water as the control, and buds were collected after 0, 1, 6, 12, 24 and 48 h. Three replicates (3 plants/replicate) were performed for all treatments.

Isolation of the PsMPT promoter
Genomic DNA was extracted from tree peony buds using the cetyltrimethylammonium bromide (CTAB) extraction method as previously described [14]. DNA samples were qualified photometrically, then checked on agarose gel, and stored at -20°C for use. Based on the cDNA sequence of PsMPT (Genbank accession No.: EU072922), three gene-specific primers, SP1, SP2 and SP3, in nested positions were designed with primer premier 5.0 and synthesized ( Table 1). The PsMPT promoter was amplified using gene-specific primers and four short arbitrary primers (AP) with the Genome Walking Kit (TaKaRa). The PCR products were ligated into the pMD18-T vector (TaKaRa), and sequenced at BGI-Beijing, China.

Construction of the promoter-reporter plasmids
To construct the binary vector consisting of the β-glucuronidase (GUS) coding sequence driven by the PsMPT promoter, a fragment from -1117 to -1 relative to the translation initiation codon was obtained using the high fidelity DNA polymerase and promoter specific primers (FP6 and RP, Table 1). The serial deletions of the PsMPT promoter (-909, -621, -574, -421, and -282 to -1) were amplified by PCR with the corresponding forward (FP1, FP2, FP3, FP4, FP5) and reverse primers (RP) that contained a Hind III and BamH I site at the 5' end of each primer, respectively. For chilling response element identification, the MYC element -413 relative to the translation initiation site was subsequently deleted from the P2 construct with the FDP primer, and mutated to form the MP construct by recombination PCR with the FM1 and RM1 primers (Table 1). PCR products were retrieved and cloned into a pMD18-T simple vector (TaKaRa) followed by sequencing conformation at BGI Beijing, China. The Hind III/BamH I digested DNA fragments were inserted into the corresponding sites of pBI121, in place of the deleted CaMV35S promoter upstream of the GUS coding sequence. The pBI121 vector with the CaMV35S::GUS was used as the positive control. The resulting constructs were named P1

Arabidopsis transformation
Arabidopsis thaliana wild-type (Col-0) plants were transformed using the Agrobacterium tumefaciens strain EHA105 carrying the above constructs according to the floral dip method [18]. The transformants were screened on MS medium containing 50 mg L -1 kanamycin, and positive plants were identified by PCR amplification using GUS specific primers (GUSFW and GUSRV, Table 1). The corresponding T 1 transgenic seedlings that segregated at a ratio of 3:1 (resistant: sensitive) were selected to propagate T 2 individuals for further analysis. Five to ten transgenic lines were obtained for every construct.

GUS histochemical assay and quantitative analysis of GUS activity
GUS activity was determined in tissue extracts using a standard protocol [19]. GUS fluorescence was measured with a Microplate Spectrofluorometer, the data were obtained by subtracting the background 4-methyiumbelliferyl glucuronide of the PPsMPT::GUS transgenic plants.
The average GUS activity was obtained from at least five independent transformants, and each assay was repeated three times. GUS histochemical staining was performed using identified homozygous transgenic plants by a modified Jefferson's method [19]. In brief, plant tissues were incubated in a 100 mmol L -1 sodium phosphate buffer (pH 7.0) containing 0.1% Triton X-100, 10 mmol L -1 EDTA, 1 mmol L -1 X-gluc and 0.5 mmol L -1 potassium ferricyanide at 37°C overnight. The stained tissues were then washed several times with 70% ethanol to bleach the chlorophyll.

Quantitative real-time PCR (qPCR)
Total RNA was extracted from the inflorescence of Arabidopsis and buds of tree peony as previously described [20], and then treated with DNase I (TaKaRa) according to the manufacturer's instructions. First strand cDNA was synthesized from 2 μg of total RNA using the PrimerScript™ RT reagent Kit (TaKaRa). PCR reactions were performed in a 25 μL system including 12.5 μL 2×SYBR Green Master mix (TaKaRa), 300 nmol L -1 each primer (Table 1) and 2 μL 10-fold diluted cDNA template. PCR reactions were run in a Roche LightCycler 1 480 (Roche, Germany) using the following program: 95°C for 2 min and 45 cycles of 95°C for 5 s, 57°C for 30 s and 72°C for 30 s. The reactions were run in triplicate. The expression was normalized to beta-actin. Quantification of the relative gene expressions was performed using the 2 -ΔΔCt method [21]. Statistical analyses were performed using SPSS 13.0 (SPSS, USA).

Expression characteristics of PsMPT in tree peony
The expression of PsMPT was previously reported during dormancy release in tree peony [13]. In this study, the temporal and spatial expression of PsMPT was further detected at the early stage of flowering in tree peony. The results of qPCR indicated that the transcription of PsMPT was detected in all tree peony tissues; however, the PsMPT transcript was very low in root, stem, leaf and calyx, but high in flower organs, including petal, stamen and stigma (Fig 1A). The PsMPT transcripts in the stamen were expressed 6-fold as compared to that of the root. The results indicated that PsMPT was expressed preferentially in flower organs of tree peony. The response of PsMPT in dormant buds to Gibberellic Acid (GA) and Abscisic Acid (ABA) were analyzed by qPCR. When GA was applied, the PsMPT transcript was quickly promoted and peaked at 24 h, then declined slightly. Conversely, ABA application dramatically decreased the expression of PsMPT (Fig 1B).

Isolation of the PsMPT promoter and the putative cis-acting element
In this study, the 1174 bp upstream genomic DNA sequence of PsMPT was isolated by TAIL-PCR (Fig 2). The adenosine of the translation initiation codon (ATG) of the PsMPT gene was defined as +1 (Fig 2). A motif search was carried out using PLACE and PlantCare to analyze the putative cis-elements. As shown in Fig 2, the putative TATA box was found at position -189/-185 and the CAAT box at position -210/-207. In addition, a number of regulatory motifs potentially related to environmental signals were found, which included auxin-, GAand dehydration-responsive and tissue-specific elements. Among the cis-elements, one putative pyrimidine box and two GA-responsive elements (GAREs) were located at -991/-986, -913/-907 and -780/-771, respectively, which were related to GA response and sugar repression [22,23].  [24][25][26]. Several MYB and MYC motifs were thought to respond to chilling or freezing in Arabidopsis [24][25][26]. There were two putative sulfur-responsive element (SURE) motifs at positions -647/-643 and -387/-382, and one putative TGA element (auxin responsive element) at position -764/-759. SURE contains the auxin response factor binding sequence [27]. Therefore, PsMPT might be regulated by chilling and auxin. Interestingly, eight putative GATA boxes were present in the sequence, which were thought to be involved in tissue-specific expression and light response [28,29]. Moreover, three putative pollen1lelat52 elements, related to the pollen specific expression [2,30], were found at -775/-771, -782/-777, and -1005/-1001, respectively. Temporal and spatial expression of the PsMPT promoter in Arabidopsis To identify the expression patterns of the PsMPT promoter, the promoter::GUS chimeric construct (PPsMPT::GUS) was transformed into Arabidopsis, and histochemical GUS staining was carried out in various organs throughout plant development (Fig 3I). These results showed that GUS activity was not detected in the seedling tissues (Fig 3IA and 3IB) but were observed in the flower organs (Fig 3IC, 3IE-3IG). GUS activity in transgenic plants was more pronounced in the stigma and stamen compared with the sepals (Fig 3II-3IL). GUS staining was not observed in the silique and seeds (Fig 3IH). These results suggested that PsMPT is preferentially expressed in the flower tissues.
qPCR was performed to evaluate the spatial expression of the PsMPT promoter. The results showed that the GUS transcript was only detected in the flower (Fig 3II). No transcript of GUS was detected in the roots, rosette leaves, cauline leaves and silique. GUS transcript levels in bloomed flowers were approximately 10-fold higher compared with flower buds. In bloomed flowers, the most abundant expression was found in stigma, followed by in stamen, with the lowest amount in the petals. In summary, the GUS staining results are in accordance with that of qPCR. The GUS reporter did not appear in the immature flower buds possibly due to the low abundance of the transcripts. The flower-specific expression characteristics of the promoter implied that PsMPT might participate in plant anthesis and gametophyte development.

Responses of the PsMPT promoter to hormones and abiotic stresses
The transgenic plants carrying the PPsMPT::GUS cassette were treated with hormones and abiotic stresses, and the transcription of GUS was evaluated by qPCR, respectively. Overall, GUS activity changed rapidly and fluctuated during the entire period for all of the treatments (Fig  4). Chilling increased the GUS transcript during 3 h to 6 h after treatment, with a peak approximately 2.5-fold higher at 6 h (Fig 4A), which was also verified by GUS staining (Fig 4B). The results of qPCR showed that GA 3 and NAA treatments enhanced GUS expression at 1 h, then decreased, followed by another peak at 24 h. ACC accelerated the transcript of GUS during the entire process. Conversely, ABA inhibited GUS expression throughout the process. In addition, the GUS transcript was continuously decreased by mannitol and PEG until 3 h, followed by a slight increase at 12 h and 24 h; however, the GUS transcript levels were lower than that of control. Notably, NaCl dramatically enhanced GUS activity, and it reached a peak at 3 h with an approximate 25-fold increase (Fig 4C). The results of GUS activity were consistent with GUS expressions of qPCR (Fig 4D).

Identification of chilling response elements
As reported previously, PsMPT was chilling inducible. Furthermore, PsMPT was involved in chilling induced dormancy release in tree peony. To confirm the PsMPT promoter region involved in the chilling response, a number of truncated promoter fragments (P1 (-282), P2 (-421), P3 (-574), P4 (-621), P5 (-909) and P6 (-1117)) were isolated and fused to the GUS reporter gene into pBI121 vector (Fig 5A). Transgenic Arabidopsis plants with each promoter-GUS construct were generated. The transcription levels of GUS were detected by qPCR ( Fig  5B). In the transgenic plants, the highest level of GUS expression was detected in the engineered Arabidopsis with the P6 construct, which contained the full-length PsMPT promoter (−1117/−1). GUS transcript decreased in order from the P6 to P1 construct, and P2 and P3 had similar abundance. These results suggested that the transcription enhancer might exist in the upstream of the PsMPT promoter.
We also detected GUS expression of the successive deletions at 4°C for 6 h compared to 18°C (Fig 5B). After exposure to 4°C, the GUS transcripts were up-regulated in the P2 to P6 constructs. The largest increase in GUS abundance was observed in plants with the P6 construct (6.34-fold). The next-largest increase was observed in the P5 construct (-909/-1, 5.40-fold). Similar enhancement was observed in the P2 and P3 constructs with 2.45 and 2.41-fold, respectively. However, the P1 construct (-282/-1), which only contained basic transcriptional elements, could not induce GUS expression at 4°C. These results indicated that the −421/−282 region of the PsMPT promoter could lead to efficient chilling induction. Bioinformatics analyses showed that a MYC cis-element existed in this region, and several MYCs were Characteristic of PsMPT Promoter and a Key Chilling Response Motif found within the promoter. We speculated thus that the MYC elements in the promoter might be involved in the chilling response.
To verify our hypothesis, another deletion construct (DP) (-404) and one mutated construct (MP) (-574) were constructed (Fig 5A). In the DP construct, the MYC element located at -412 in the promoter was deleted. When exposed to 4°C, no significant difference was observed as compared to that of 18°C treatment, which indicated that there was a loss of function for chilling induction with the MYC element deleted in the PsMPT promoter (Fig 5B). In the MP construct where the MYC element (-412) was mutated from CAATTG to ACCTTG, a complete elimination of the chilling response was observed (Fig 5B).
To confirm the results obtained by qPCR, a quantitative measurement of GUS activity was performed using the series constructs, P1-P6, DP and MP. Consistent with the results obtained by qPCR, the GUS activity was observed to increase from the P2 to P6 after exposed to 4°C, and no difference was observed in the P1, DP and MP. In summary, we concluded that the MYC element involved in chilling treatment responses (Fig 5C).

Discussion
MPT can shuttle inorganic phosphate (Pi) into the mitochondrial matrix, where ATP is synthesized. Thus, MPT plays a key role in cellular ATP regeneration. ATP is essential for almost all biological processes in the cell, and MPTs have been reported to be involved in abiotic responses [6,12], bud dormancy release, growth and development [13]. Although several MPTs have been cloned and functionally annotated, their characteristics and regulatory mechanisms are poorly understood.
We previously cloned PsMPT, a chilling induced gene, in tree peony, which accelerated ATP synthesis and dormancy release [13]. In this study, we isolated the PsMPT promoter and detected its activity using GUS as the reporter in transgenic Arabidopsis. GUS-staining and qPCR of the transgenic plants revealed that the PsMPT promoter was preferentially expressed in the flower, mainly in stamen and stigma. Therefore, the tissue-specific expression may be related to the putative GATA boxes and pollen1lelat52 elements founded in the promoter [30,31]. The results suggested that PsMPT might play an important role during gametophyte development, pollination and fertilization, and Pi might be transported to reproductive organs during the reproductive development stage. Differing from ectopic expression of the PsMPT promoter, PsMPT mRNA was detected in all tissues of tree peony by qPCR. This discrepancy has also been reported in mice [32,33] and is believed to be related to different biological species or incomplete isolation of promoter sequences.
The temporal and spatial expression of MPT has been reported in several plants, and all of the results showed tissue-specific expression. Birch MPT1 was highly expressed in tissues containing dividing cells [6]. There were six MPT members in rice, and a microarray analysis also revealed tissue-specific expression [11]. In Arabidopsis, the transcription of the AtMPT1 was pronounced in the stamens of flowers, and AtMPT2 mRNA was abundant in rosette leaves; whereas, AtMPT3 was strongly expressed in leaves and weakly expressed in the roots and flowers [12]. Overall, the spatial pattern of PsMPT was similar to AtMPT1. In contrast, PsMPT showed only 51% sequence identity with AtMPT1, and 79% sequence similarity with AtMPT3 (S1 Table). The organization of promoters between PsMPT and AtMPTs was also compared, and large differences were observed (S1 Fig). Similarly discrepant MPT sequences and expression patterns between tree peony and Arabidopsis were also found in grape and Arabidopsis [11]. Taylor et al. investigated the relationship between the promoter and coding sequence selective constraint and suggested that they were generally uncorrelated [34], which implied partially independent evolution of promoters and their coding sequences between species. Therefore, we speculated that the discrepancy might be due to species-specific independent evolution of MPTs and their promoters.
The response of the PsMPT promoter to abiotic stress and hormones was analyzed by qPCR. GUS transcript increased during chilling treatment in the transgenic plants driven by the PsMPT promoter. In addition, the PsMPT promoter was also induced by salt, GA 3 , ACC and NAA, while ABA, mannitol and PEG suppressed its activity. In rice, OsPT19 was suppressed by five hormone treatments including ABA, 2, 4-D, GA 3 , KT, and NAA, whereas OsPT17 was induced by the hormone treatments with NAA or GA 3 [11]. These responses might be due to the cis-elements in the promoter, for instance, the putative pyrimidine box and two GARE motifs to GA response, putative TGA and SURE elements to NAA response, and MYB and MYC to abiotic stress response, such as chilling. Interestingly, NaCl treatment could significantly increase the expression of the reporter gene with a maximum of approximately 25-fold. Similar results were reported for Arabidopsis [12].
Among all of the factors influencing promoter activity, chilling and GA 3 treatments are of interest because they effectively accelerate the dormancy release in winter [35][36][37][38]. Huang et al. found that PsMPT was induced by chilling [13]. Interestingly, it was reported that the chillinginduced expression of PsMPT was not maintained after being transferred to a greenhouse (18-22°C) when less than 21 days of chilling were applied. On the other hand, the levels of PsMPT transcripts remained high with a 21 d or longer chilling duration after returning to growth temperature [13]. In this study, ectopic expression analyses provided more evidence that the PsMPT promoter could be induced by chilling. We speculated that the early increase of the PsMPT transcript might be induced by chilling, and GA production might be a downstream effect of chilling, as proposed for dormant seeds [39,40]. Meanwhile, chilling temperature was reported to enhance the accumulation of endogenous GA, and exogenous GA could partially replace chilling to accelerate endo-dormancy release [41]. In this study, we found that exogenous GA could activate the expression of the PsMPT promoter. Therefore, the reactivation of the PsMPT transcripts might be due to the high endogenous GA induced by sufficient chilling when transferred to growth condition. Buds chilled for less than 21 days had relatively low GA levels that could not activate the PsMPT expression required for the recovery of plant growth ability.
Considering the central role of PsMPT in energy metabolism during dormancy release, it is important to elucidate how chilling accelerates PsMPT expression. It is well-known that the transcription of mRNA is mainly regulated through the cooperation of transcript factors and corresponding cis-elements. Several cis-elements have been identified to be involved in chilling or cold responses, such as ABRE (ABA responsive element), DRE/CRT (dehydration-responsive element/C-repeat element, A/GCCGAC), MYB, MYC, and the E-box [24,26,42,43,44,45]. Bioinformatics analysis of the isolated PsMPT promoter showed that four MYB and four MYC elements were present upstream of the promoter, which might be responsible for the chilling response. Based on the location of the MYC and MYB elements, deletion experiments were conducted to identify the candidate chilling response elements in the promoter. GUS expression and activity revealed that the P2 construct containing one MYC element (-412/-408, CAATTG) effectively responds to chilling, and the addition of a MYB element (P3 construct) did not improve the chilling response ability. When MYC was deleted or mutated, the chilling response character abated. Alternatively, increase of MYC elements in the P4, P5 and P6 constructs enhanced the chilling response activity, indicating there was an additive effect of MYC elements in the chilling response. This study demonstrates that the MYC elements in the PsMPT promoter play a crucial role in the chilling responses.
In conclusion, we isolated the PsMPT promoter in tree peony and found that it is a floralpreferential promoter. The promoter of PsMPT responded to chilling, ACC, PEG, NaCl, mannitol, auxin and GA. Deletion and mutation analyses demonstrated that the MYC cis-element functioned in the chilling response. This work provides useful information for further investigation of the regulatory mechanisms of PsMPT promoter during endo-dormancy release.