PmCBFs synthetically affect PmDAM6 by alternative promoter binding and protein complexes towards the dormancy of bud for Prunus mume

The survival in freezing temperature for woody plants is exclusively dependent on the perception of coldness and induction of dormancy. CBF/DREB1 transcriptional factors join cold-response conduits and the DAM genes, especially PmDAM6, are well-known regulators of dormancy. Despite the immense importance, little is documented on the association between CBF proteins and the complexity of the promoter region in PmDAM6 with the function of bud dormancy in P. mume. Therefore, this study was based on the cloning of PmDAM6 and six PmCBFs to evaluate their integral roles in the process of bud development. The consistency of expressions in either vegetative or reproductive buds provided a negative control from PmCBFs to PmDAM6 during the onset of dormancy. Besides, PmCBF5 could form heteromeric complexes with PmDAM1 and PmDAM6. PmCBF1, PmCBF3, and PmDAM4 recognized the promoter of PmDAM6 by the alternative binding sites. Therefore, the interactions of these genes formulated the base of an obvious model to respond to the coldness and engendered dormancy release. Findings of this study will further help the unveil the genetic control of bud dormancy and its augmentation in P. mume and may offer an explanation for the vernalization.

roles in specifying the dormancy transitions during the growth curves [8][9][10][11] . In P. persica, the expression profiles of DAM genes are associated with obvious seasonal changes of temperature 12,13 . As for P. persica and P.avium, the expressions of DAM genes are related to the quantification of the cold response and flowering date manipulation under varying environmental circumstances [13][14][15] . PpDAM5 and PpDAM6, showing the expressions closely related to cooling capacity and flowering date 14,16 , can inhibit peach bud growth at low temperature 13 . The expressions of PpDAM5 and PpDAM6 were up-regulated during internal dormancy, but down-regulated during dormancy release 3,17 .
Low temperature significantly affects C-Repeat Binding Factor (CBF), a cold response signal factor isolated by Stockinger 18 . In Arabidopsis, there are three CBF/DREB1 genes guiding the signal pathway to low temperature response 19 . CBFs can function to enhance the cold resistance of plants by inducing the downstream genes 20,21 . Benedict 22 transferred the AtCBF1 gene into poplar, comparing it to the wild-type poplar, and found that the cold resistance of the transgenetic poplar was highly enhanced. CBF genes in flowering peach can increase the cold tolerance and can cause dormancy in short-day conditions 23 . When a peach CBF gene was overexpressed in apple, both non-acclimated and acclimated freezing tolerance was observed and displayed these phenotypes 24 .
For many woody plants, DAM genes are involved in plant dormancy induction, and CBF genes are related to cold-response pathway. On account of the participation of these genes in supervising the plant growth and development in cold climate, it is necessary to investigate their functions in plant growth and dormancy control. Therefore, studying the interactive influence of both the gene types as determinants of bud activity in P. mume could be a nascent concept inviting plentiful directions for future studies. In this research, we cloned PmDAM6 and six PmCBFs, ascertained their expression profiles in flower bud, flower, leaf bud, and leaf, and tested their roles in the simulated bud dormancy. In addition, multiple interaction experiments were carried out to investigate the protein-protein and protein-DNA associations among PmCBFs and PmDAM6 in controlling bud growth and dormancy. This will invite future studies on molecular perceptive of bud control and cold response conduit.

Results
PmCBFs bind to the promoter of PmDAM6. Based on the research of P. mume genome, there exist more binding sites (C-repeat/DRE element) in the promoter of PmDAM6 than those of other PmDAMs. In this research, we obtain the upstream fragment of PmDAM6. Indeed, there are three potential binding sites in the 1 kb region of PmDAM6's promoter, and four in the 2 kb region. To uncover the role of PmCBFs and PmDAM6 in induction and release of flower bud dormancy, we conducted yeast one hybrid (Y1H) experiments to explore the regulation between six PmCBFs and the promoter of PmDAM6.
As shown in Fig. 1A, the promoter of PmDAM6 (Supplementary Data S2) was separated into four potential binding sites based on the position of CCGAC: M1, −1971; M2, −706; M3, −648; M4, −294. Then, baits (length less than 100 bp) were designed as shown in Fig. 1B. A total of six bait vectors were constructed, M2 and M3 were designed in one bait, which got close to each other, the other two, M1 and M4, were independently performed as baits, then the original sequences of these three baits were duplicated respectively to form other three baits (Supplementary Data S3). Five of these six baits were tested negative to perform the following yeast one hybrid. The corresponding working concentrations of Aureobasidin A (AbA) were as follow: pAbAi-1-1, 400 ng/ml; pAbAi-1-3, 200 ng/ml; pAbAi-2-3, 1000 ng/ml; pAbAi-3-1, 400 ng/ml; pAbAi-3-3, 200 ng/ml. And the pAbAi-2-1 was not considered as a bait because of its normally growing under1000 ng/ml AbA.
In vivo protein-protein interactions among PmCBFs and PmDAM6. Based on yeast two hybrid (Y2H) and bimolecular fluorescence complementation (BiFC) assay, a further exploration about PmCBFs and PmDAMs was made. According to Yeast two-hybrid experiments, PmCBF5 showed a strong interaction with PmDAM1 and a general interaction with PmDAM6 ( Fig. 2A). Moreover, protein-protein interactions among PmDAM6 and PmCBFs were established by BiFC with a yellow fluorescent protein. Fluorescence released by these yellow fluorescent proteins was positioned at nucleus, suggesting the strong inclination among PmCBF5 and these two PmDAMs. PmCBF5 could form heteromeric complexes with PmDAM1 and PmDAM6 (Fig. 2B). As shown in Supplementary Fig. S3, there were no interactions in PmDAM1-YFP N / YFP C , YFP C / PmCBF5-YFP N , PmCBF5-YFP C / YFP N , and YFP N / PmDAM6-YFP C . In order to exclude the false positives, a member from the same protein family can be chosen to execute a negative control 25 . Therefore, the interactions of PmDAM2-YFP N / PmCBF5-YFP C and PmDAM1-YFP N / PmCBF6-YFP C were used as the negative controls of PmDAM1-YFP N / PmCBF5-YFP C ; the interactions of PmCBF6-YFP N / PmDAM6-YFP C and PmCBF5-YFP N / PmDAM4-YFP C were used as the negative controls of PmCBF5-YFP N / PmDAM6 -YFP C . Fluorescent was not detected the in these negative tests.
Crossover of the expressions between PmCBFs and PmDAM6 in flower bud dormancy. In Beijing, P. mume 'SanlunYudie' begins to form buds from June to July, undergoes flower bud differentiation from July to November, steps into dormancy from September to October, retains dormancy from November to January, and initiates dormancy breaking from January to February (Fig. 3A). From July to October, PmDAM6 showed up-regulated in the chilling environment with a temperature under 20 °C and down-regulated in the coldness deepened weather. However, the expression tendencies of six PmCBFs during different periods of flower bud development were basically similar with high positive correlations (Fig. 3B) and were induced vastly under low temperature before freezing. The transcript levels of PmCBFs were low during July-October and January-February, but were high during November-December. The high expression of CBF genes seemed to inhibit the expressions of DAM genes. Together with all expression data, we can conclude that PmDAM6 and PmCBFs function in two temperature regions, respectively. This means PmDAM6 actively respond in the chilling temperature (below 20 o C) and PmCBFs dominate in the freezing zone (below 0 °C), and the regulations in this biological process between these genes are continuous with a suppressive tendency.  structures, the blooming process was divided into four stages. On the whole, the expression trends of PmDAM6 also declined from F1 to F4. In addition, PmDAM6 were expressed in all the floral structures, and the expression of PmCBFs kept low levels and got lower in the flower blooming process (Fig. 4A). The expression tendencies of the PmCBFs in the flower blooming stages were basically the same, the transcript levels were relatively high in F1 and F2, but low in F3 and F4. According to the expression patterns of PmCBFs in different flower structures, PmCBFs were divided into two groups. PmCBF1-4 showed high expression specificity in sepals. The other type genes, including PmCBF5-6 were expressed in all of the floral structures without significant differences.
Expression patterns of PmDAM6 and PmCBFs in leaf buds and leaves. The leaf buds experienced a different development process compared to flower bud. They formed in the same period with flower bud, but only enlarged after the flower blooming after dormancy in the cold winter. DAM genes were specifically expressed in leaf buds at different developmental stages. PmDAM1-3 exhibited high expression levels in September, and expression levels of PmDAM4-6 were the highest in October (Fig. 5A). Furthermore, the expression trends of PmDAMs in leaf bud were largely consistent at different development stages and showed significant expressions from September to October which then gradually declined from October to March in the next year. However, the expression levels of PmCBFs were the highest in November and December. Above all, the expression patterns of DAM and CBF genes between leaf bud and flower bud were largely consistent at different development stages. The expressions of PmCBFs manifested two types of similar patterns with high levels. The first group (PmCBF1-3) showed peak expressions in August, up-regulated from April to August, and down-regulated from September to November (Fig. 5B). The second group (PmCBF4-6) showed peak expressions in June and July, then declined till November. But all the PmCBFs own high levels from June to August.

Simulated bud dormancy and dormancy release. In order to study the functions of PmCBFs and
PmDAMs in the bud dormancy, we detected the expressions of all six PmCBFs and six PmDAMs in the simulated condition of bud dormancy and dormancy release.
One year old branches with flower buds were sampled on October 1, 2015. The growing branches were treated with 6 °C ± 2 °C to induce dormancy. The expressions of PmCBFs were shown in Fig. 3C, and were of two types. One category of PmCBFs, including PmCBF2-5, showed rising expressions, while the other category, contain-ingPmCBF1 and PmCBF6, showed higher expressions in 10-40 days as compared to 0 day, suggesting the low temperature significantly affect the expressions of PmCBFs. However, the expression of PmDAM6 increased in 0-20 days but dropped in the following days. On first day of low temperature exposure, PmDAM6 presented a high expression. During the process of dormancy, PmCBFs accumulated in low temperature, while the expression of PmDAM6 declined.
Warm temperature of spring was simulated at 12 °C ± 2 °C to release the bud dormancy with one year old branches on December 10, 2015. After warm temperature treatment, flower buds turned to expand within 10days, and after 20 days, 95% flower buds showed white petals. As shown in Fig. 4B, the expressions of PmCBFs and PmDAM6 reduced overall. The expressions of PmCBF1, PmCBF3, PmCBF4 and PmCBF5 keep stable, but declined remarkablely during 30-40 days. In contrast, PmCBF2, PmCBF6 and PmDAM6 rapidly declined in 10 days. All results suggested that in the release of dormancy, PmCBFs and PmDAM6 showed a down regulation.

Discussion
PmCBFs suppress overexpression of PmDAM6 in flower bud dormancy. Acquirement of protective strategies for reproductive organs in plants possesses vital importance for the continuation of life, however, the uncertain environment changes, like low-temperature stress, remains the major hurdle to the success of plants In peach, PpDAMs has an association with seasonal dormancy transitions in flower buds 26 . SVP genes, orthologs of DAM, in Arabidopsis show regulatory effect during floral transitions and play a significant role in specifying the floral meristems 33 . In the flower bud of Prunus mume, PmDAM6 exhibited a special expression leading to the dormancy under decreasing temperature (Fig. 3A). Moreover, the appearance of high CBF expression levels enhanced the tolerance of coldness in freezing climates and inhibited the expression of PmDAM6. This may lead to a different phenomenon that flowers buds are easier to sense the changes in environment and devise the mechanism of self-protection 34,35 . Nevertheless, this happens only if plants are shocked by cold. When plants were released from cold weathers, PmDAM6 and PmCBFs just relaxed with all decreasing trends and relatively lower levels (Fig. 4). This must be an integrated result from regulation beyond CBFs and DAMs.
In P. persica, DAMs have shown in vegetative and non-vegetative plant parts development including bud 12 . Here, the protection mechanism prefers to appear in young tissues facing coldness. Annually sampled leaf buds gave important clues about potent roles of DAM6 and CBF genes in P. mume wherein PmDAM6 showed high expressions in September to October and the CBFs appeared from November to December (Fig. 5A). Besides, MdCBF1, 2 and 4 were discerned in the leaf tissues of apple subjected to cold conditions, and VrCBF exhibited a short expression (1/2 h-2 days) both in young and old leaves, while VrCBF1, 2 and 3 expressed only in younger tissues 36 . Some reports have indicated a reverse relation between DAM and CBF during growth curves, and dormancy control, wherein FT can be down-regulated by the high expression of DAM 30 . This may provide an explanation for the growth and dormancy of young tissues.
In the mature leaf aging process, PmDAM6 also showed a negative correlation with PmCBFs especially for PmCBF4-6, though this happened under a warm environment. However, it is confirmed that CBFs function in heat and drought as well as cold response in soybean 37 . The ectopically expressing PpCBF1 induced aging in leaf and also delayed bud opening in the spring 24 .
Under artificial control environment, PpDAM5 and PpDAM6 exhibited an upregulation under sustained low temperature and showed a downregulation by warm temperature in peach 38 . The growth of peach buds are negatively correlated with the expression of PpDAM5 and PpDAM6 13,39 . In addition, the sustained low temperature could satisfy the needs of a chilling requirement in buds to promote the release of dormancy with PpDAM5 and PpDAM6 down regulated 16 . In P. mume, high expression of PmDAM6 hastened the flower buds into dormancy. With the accumulation of coldness in winter, the expressions of PmCBFs increased, and that of PmDAM6 decreased, indicating that high amount of PmCBFs seemed to depress the expression of PmDAM6. When Interactions among PmDAMs and PmCBFs. As reported in recent researches, CBF could bind to the promoter of DAM genes based on a fixed motif. For the model plant, Arabidopsis, CBFs recognize DNA sequences with CRT/DRE sites, which cover a core motif of CCGAC. Compared with other herbaceous plant, like oilseed rape, wheat, or rye, the cold response pathway with a member of CBF remains conservative between species suggesting its vital place for the whole pathway. In pear, the interaction between PpCBF and the promoter of PpDAMs give a convincing proof. In Pyrus pyrifolia, the CRT/DRE binding ability between PpCBF and the promoters of PpDAM1and PpDAM3 was verified 6 . In the genome of P. mume, the conserved binding motif exists in the promoters of PmDAM1, PmDAM4, PmDAM5,and PmDAM6. Particularly, it makes the function of PmDAMs more complex that the PmDAM6 promoter contain more CRT/DRE core sites (Fig. 1). For pear, PpCBF does not bind to the bait, when CGGC was transformed into AAAA, suggesting the regulation of DAM depends on the recognition of this core sequence. In our study, Y1H assays exhibited PmCBF1, PmCBF3 and PmCBF4 indeed interacted with the promoters of PmDAM6 by different sites. But PmCBF3 was not only located in the nucleus but also the chloroplast. Besides, PmCBF5 could form heteromeric complexes with PmDAM1 as well as PmDAM6, suggesting that PmDAMs and PmCBFs might function in the bud dormancy by dimerizations (Fig. 6). Therefore, scrutinizing the expression patterns coupled with dimerizations trends can be a reasonable platform to grasp detailed understanding of protein-protein interactions among PmDAMs and PmCBFs during bud transitions. Though the PmCBFs may have matching functions 20 , the alternative interaction abilities still provided a possible regulation from chaos. Especially for the PmDAM6, the expression did not get the bottom, when the PmCBFs exaggerated. Conversely, the expression of PmDAM6 kept a relative stabilization, which drove us to hypothesize that the interaction from protein to protein and protein to DNA conducted a feedback loop in the deep dormancy condition (Fig. 6). However, there were still mysteries because of the multiple CRT/DRE core sites that whether the number of binding sites superpose or subduct the function of downstream gene, all the hypotheses need further researches.
In our study, we observed the crossover of the expression between PmCBFs and PmDAM6, and provided the evidences that besides multiple recognitions to the promoter of PmDAM6, PmCBFs also could form protein complex with PmDAM6, which act as a center role in dormancy induction. This multiplex regulation might reflect the evolvement of plant towards more rapidly changing environment. Moreover, the expression patterns in vegetative bud and simulated dormancy test of bud convince the negative control in the dormancy formation. We believe this novel research will help direct future discussion on the interactive role of DAM and CBF genes as essential regulators of bud dormancy.  (Supplementary Data S1) were amplified through PCR using the primers in Supplementary Table S1. PikoReal real-time PCR system (Thermo Fisher Scientific, Germany) was used to investigate the expressions of PmDAMs and PmCBFs in different organs. The primers of RT-qPCR experiments were shown in Supplementary Table S2 and the experiments were carried out using previous method 40 . PmPP2A (protein phosphatase 2A) was considered as reference gene 41 . Three biological replicates were performed to calculate the standard deviation. The correlations between gene expressions were done by spearman method, and significant was analysed with kruskal-wallis test in R.

Materials and Methods
Gene cloning and Yeast 2 Hybrid assays. Full length cDNA sequences of PmDAMs and PmCBFs were amplified through PCR using specific primers (Supplementary Table S3). These sequences were cloned into pGBKT7 bait vector and pGADT7 prey vector at the EcoRI and BamHI sites, respectively, using an In-Fusion HD Cloning Kit System (Clonetech, USA). The Y2H assays were performed by former methods 42 . Each interaction analysis was performed three times.
BiFC Assays. Specific primers were used for BiFC assessment (Supplementary Table S4). Full length cDNA sequences of PmDAMs and PmCBFs were cloned, pair-wise, into pCambia1300-YFP-C and pCambia1300-YFP-N to get BiFC constructs. Coexpression analysis was carried out on the leaves of Nicotiana benthamiana according to the procedure stated by Schutze et al. 43 . Chimeric fluorescence, emitted by fusion proteins, was examined under Leica TCS SP8 Confocal Laser Scanning Platform with YFPs being motivated at 514 nm.
Promoter cloning and Yeast 1 Hybrid assays. The genomic DNA of P. mume was isolated from flower bud using DNAsecure Plant Kit (DP320-02, TianGen, China). 2 kb up-stream promoter sequence of PmDAM6was extracted by PCR using specific primers (Supplementary Table S5) and the plasmid of pMDTM18-T-proDAM6 was obtained by former method 42 . In 2 kb promoter of PmDAM6, there are four CCGAC sequences called CBF binding site. According these four CBF binding sites, six baits were designed. Three of them were obtained from the original genome, then each of them was duplicated to form the follow three baits. All six fragments were imported to the pAbAi-bait vectors. The special primers related to cloning these fragments were shown in Supplementary Table S6. The plasmids of pGADT7-PmCBFs were obtained in Y2H assays. These plasmids were transformed into the Y1H Gold strains containing pAbAi-bait, respectively, and these tests were screened on SD/-Leu/AbA plates. All transformations and screenings were observed in triplicate. These Y1H assays were executed by a Matchmaker Gold Yeast One-Hybrid System kit (Clontech, America) following its user manual and correlation steps.