Assessing Field Prunus Genotypes for Drought Responsive Potential by Carbon Isotope Discrimination and Promoter Analysis

In order to improve the effectiveness of breeding practices for Prunus rootstocks, it is essential to obtain new resistance resources, especially with regard to drought. In this study, a collection of field-grown Prunus genotypes, both wild-relative species and cultivated hybrid rootstocks, were subjected to leaf ash and carbon isotope discrimination (Δ13C) analyses, which are strongly correlated to water use efficiency (WUE). Almond and peach wild relative species showed the lowest Δ13C ratios, and therefore, the highest WUE in comparison with hybrid genotypes. In addition, drought-related cis-regulatory elements (CREs) were identified in the promoter regions of the effector gene PpDhn2, and the transcription factor gene DREB2B, two genes involved in drought-response signaling pathways. The phylogenetic analysis of these regions revealed variability in the promoter region sequences of both genes. This finding provides evidence of genetic diversity between the peach- and almond-relative individuals. The results presented here can be used to select Prunus genotypes with the best drought resistance potential for breeding.


Introduction
Drought stress is a significant challenge to agriculture, especially in arid and semi-arid climates [1] such as the Mediterranean region, where water availability is the most important factor for plant survival. In plants, water stress response is a complex combination of different factors at the biochemical, molecular and physiological levels leading to plant adaptation under drought conditions [2,3]. Late embryogenesis abundant (LEA) proteins are involved in this functional adaptation. Their accumulation plays a crucial role in protecting protein structure and binding metals under osmotic and oxidative stresses induced by drought, cold and salinity [4,5]. Dehydrins, which belong to the Group II LEA proteins, are one of the most important proteins that accumullate during water stress [6,7]. The role of dehydrins in abiotic stress tolerance has been demonstrated in different woody species [8][9][10][11][12]. In particular, three dehydrin genes (Ppdhn1, Ppdhn2 and Ppdhn3) have been described in peach confirming their induction by cold and/or drought [13][14][15], and the presence of specific cis-regulatory elements (CREs) in their promoter regions is thought to contribute to their induction by several abiotic stresses [14,15]. Recently, Bielsa et al. [16] confirmed the drought-induction of two genes: a gene encoding a homologous protein to D-29 LEA protein and

Plant Material and Growth Conditions
A total of 48 individuals, listed in Table S1, were used in this study. The genotypes were located at the CITA (Centro de Investigación y Tecnología Aroalimentaria de Aragón) facilities in Zaragoza, Spain (41 • 43 26" N, 0 • 48 31" W) belonging to a rootstock and wild relatives collections, respectively. Conventional orchard practices were used in tree training and weed control. Water requirements were supplied by surface irrigation for the hybrids and their parentals, and drip irrigation for the almond wild-relative species.

Leaf Ash Content Analysis and Carbon Isotope Discrimination Analysis
WUE was estimated from two analyses: (i) leaf ash content and (ii) leaf carbon isotope discrimination (∆ 13 C). Approximately 15 leaves per tree were collected, washed with deionized water, and air dried at 60 • C for 48 hours (h). The tissue was dried further at 70 • C for 72 h, ground to a degree that would allow passage through a 40-mesh screen, and analyzed for 13 C content (University of California, Davis Stable Isotope Facility, Department of Plant Sciences, Davis, CA, USA). Carbon isotope discrimination (∆ 13 C) was calculated according to [49]. The carbon dioxide isotope composition in air was assumed to be −7.8 parts per thousand [50]. The same sample leaf tissue weight (0.5 g) was placed in a preheated porcelain crucible and burnt in a muffle furnace at 550 • C for 24 h to determine ash content using a thermogravimetric analyzer (Leco, Inc., St. Joseph, MO, USA, model TGA701). Correlation analysis was performed to relate leaf ash content with ∆ 13 C using IBM SPSS Statistics v21.0 (SPSS Inc./IBM Corp., Chicago, IL, USA).

DNA Isolation
Leaves were collected and stored at −20 • C. Total DNA was extracted from 50 mg of frozen leaves as described by Doyle and Doyle [51]. In brief, each sample was ground in a mortar with liquid N 2 . The ground material was lysed with 700 µL of CTAB (100 mM Tris-HCl C 4 H 11 NO 3 , 20 mM EDTA, 2% CTAB, 1.4 M NaCl, pH 8, 1% PVP-40, 0.1% NaHSO 3 ) and 0.4 µL of 2-mercaptoethanol and transferred to a 1.5 mL Eppendorf tube. Cellular lysis was further assisted via incubation at, at 65 • C for 25 min. Next, 700 µL of chloroform-isoamyl alcohol (24:1, v/v) was added. Samples were then homogenized, it was centrifuged at 5590× g for 15 min, at room temperature. After centrifugation, 450 µL from the upper phase were transferred to a new 1.5 mL Eppendorf tube and an equal volume (450 µL) of cold isopropanol was added and samples thoroughly mixed. The precipitated nucleic acid was recovered by centrifugation at 10,956× g at room temperature for 5 min, washed in 800 µL of 10 mM ammonium acetate in 76% ethanol for 45 min. After the washing step, the sample was centrifuged again at 10,956× g at room temperature for 5 min. Finally, the supernatant was removed and the pellet dried at room temperature. DNA was re-suspended in 100 µL of TE buffer (10 mM Tris-HCl, 0.1 mM EDTA, pH 8.0) and stored at 4 • C overnight. The following day, the samples were quantified using a NanoDrop ® ND-1000 UV-vis spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA).

PCR Amplification
In order to obtain the approximate 1000 bp upstream sequence of the translation start codon to represent the promoter region, primers were designed based on the nucleotide sequences of the PpDhn2 (ppa011637m.g) and DREB2B (ppa022996m.g) according to the sequences of the assembled and annotated peach genome (P. persica genome v1.0; http://www.rosaceae.org/). Approximately 150 ng of genomic DNA were amplified using a Platinum Taq DNA Polymerase High Fidelity kit according to the manufacturer's instructions (Invitrogen, Life Technologies, Carlsbad, CA, USA) and the PpDhn2-specific primers, forward 5 -TTGAGCAGCAGTATCACAAGC-3 , and reverse: 5 -GGTGGTTCCGGTCGTAGTAG-3 ; and the DREB2B-specific primers, forward 5 -ACGTGGGACAAAACAGGGTA-3 , and reverse: 5 -TACCAAGCCAAAGACGACTG-3 . The PCR conditions used were 1 min at 94 • C, followed by 35 cycles of 30 s at 94 • C, 1 min at 60 • C and 2 min at 68 • C, followed by a final extension of 10 min at 72 • C. After agarose gel electrophoresis, the PCR products were purified using a DNA Clean and concentrator™-5 kit (Zymo Research, Orange, CA, USA) following the manufacturer's recommendations.

Cloning and Sequencing
The gDNA fragments of 1074 bp and 1003 pb obtained for the putative promoter regions of the PpDhn2 and DREB2B genes, respectively, were subsequently cloned into the pCR™2.1-TOPO ® vector (Invitrogen, Life Technologies, Carlsbad, CA, USA) following the manufacturer's instructions. The plasmid DNA of the positive transformants was isolated using GeneJET™ Plasmid Miniprep kit (Thermo Fisher Scientific, Waltham, MA, USA). After digestion with EcoR1 using EcoR1-HF™ RE-Mix ® (New England, BioLabs Inc., Ipswich, MA, USA) for checking the quality and the integrity of the gDNA insert within the vector, positive clones were sent to Beckman Coulter Genomics (Danvers, MA, USA) and Secugen S.L. (Madrid, Spain) for sequencing using the universal M13 forward and reverse primers.
The phylogenetic trees for each promoter region of the PpDhn2 and DREB2B gene were constructed to classify our individual plant lines on the basis of their respective promoter sequences. This analysis was done using MEGA 6.0 [55] with the neighbour-joining (NJ) method [56], and a bootstrap analysis was conducted using 1000 replicates [57]. The evolutionary distances were determined using the Kimura 2-parameter method [58].

Relationship between Leaf Ash Content and ∆ 13 C
Mean ∆ 13 C ratios varied among genotypes and ranged from 17.71‰ to 23.17‰ and mean ash content varied from 5.96 to 17.97% (Table 1). There was a significant (p < 0.05) positive relationship between ∆ 13 C and leaf ash content (Figure 1) P. davidiana individuals had the lowest value for both ∆ 13 C ratio and leaf ash content (Table 1 and Figure 1). The ∆ 13 C values of almond-related wild species were close to the average (20.99‰) with ratios between 19.96‰ to 20.87‰ (Table 1). Genotypes with highest ∆ 13 C ratios were 'Nemared' (23.17‰), 'Monegro' (23.11‰) and 'Mira × Pecher' (22.95‰). ∆ 13 C ratios of the individuals belonging to G × N series, except for the genotype 'GN-8', were above average (Table 1). Genotype 'GF-677' had the highest leaf ash content and the fourth highest ∆ 13 C value ( Table 1). Variability of ∆ 13 C values was low with an overall standard deviation value of 1.31 and coefficient of variation (CV%) of 6.40, while the overall standard deviation of ash content values was 3.10 with a CV% of 34.99 (Table 1). Comparing both ∆ 13 C and ash content values, P. davidiana individuals had the lowest values ( Figure 1), indicating higher WUE than the other genotypes. Conversely, 'GF-677' and 'Mira × Pecher' hybrids had the highest values for ash content and ∆ 13 C, indicating the lowest WUE (Table 1 and Figure 1). The almond wild-relative species had similar low ash content values to P. davidiana except for P. webbii individuals F3 and F17, and low ∆ 13 C values compared to the peach and the peach hybrid values. Overall, these peach relatives had higher ∆ 13 C and ash content values than almond wild-relative species. Among the G × N series, 'GN-8' and 'GN-10' had lower ash and ∆ 13 C values than 'Felinem', 'Garnem', 'Monegro' and 'Nemared' (Table 1).

Figure 1.
Relationship between carbon isotope discrimination [∆ 13 C (‰)] and leaf ash content (%) for Prunus genotypes. Arrows indicate the negative relation between these two parameters and WUE.

Phylogenetic Analysis Based on Promoter Regions of PpDhn2 and DREB2B Genes
The 5 regulatory region of the PpDhn2 gene amplified from the 47 genotypes and species assessed were classified into six clusters ( Figure 2) based on the dendrogram tree obtained by the NJ method. Cluster I contained 28 individuals, including all the hybrids and their parentals, except one individual belonging to P. mira genotype, which was grouped in cluster II, as well as 12 individuals belonging to 6 different wild-relative species (Figure 2). Cluster II included one P. mira individual, P. mira T1, as above-mentioned, and another six plant lines from four different wild-relative species (Figure 2). Clusters III, IV and V were the only clusters containing just one individual from 2 wild-relative species, P. gorki for Cluster III and the P. webbii F3T2 and F3T1 lines for clusters IV and V, respectively. (Figure 2). Finally, cluster VI was formed by nine wild-relative almond species (Figure 2). These results revealed the diversity in the promoter region of the PpDhn2 gene.
The promoter region of the DREB2B TF encoding gene from 48 plant lines was grouped in four clusters ( Figure 3). The largest cluster, cluster I, contained 29 individuals including the 'Garfi' almond and individuals belonging to the almond wild-relative species P. vavilovi, P gorki, P. webbii, P. bucharica, P. orientalis, P. zabulica, and P. kotschii, as well as the 'GF-667' hybrid ( Figure 3). Cluster II contained four wild-relative almond individuals belonging to P. zabulica and P. kotschii species (Figure 3). The smallest group was cluster III and was formed by the single genotype P. mira T2 (Figure 3). All hybrid individuals and most of the parentals were found in cluster IV (Figure 3). This dendrogram showed evolutionary distances close to zero, indicated a high level of conservation in the 5 regulatory region of the DREB2B gene for each individual analyzed.

Drought-Related cis-Regulatory Elements Fround in PpDhn2 and DREB2 Promoters
In order to construct a more detailed understanding of the expression regulation of the PpDhn2 and DREB2B genes in response to drought, we next searched for CREs in the putative promoter regions of each analyzed plant line. Based on the dendrograms resulting from the phylogenetic analysis, the nucleic acid sequences of selected individuals from each cluster were aligned. Individuals of each group of the alignment were selected again depending on the nucleic acid differences found in the alignment analysis. Finally, CREs were found not only responsive to drought stress, but also to other processes and stresses such as light, development, hormone, biotic and abiotic stress responses in both promoter regions For the PpDhn2 gene, we analyzed the promoter regions of P. mira T2 and P. webbii F17T2 from cluster I, P. gorki T4 from cluster III, and P. zabulica F1T2 from cluster VI as representatives of each cluster. For CREs analysis, clusters II, IV and V were represented by P. gorki T4 from cluster III because the promoter regions of all these grouped plant lines harbored the same complement of CREs in their respective promoter regions. Different families of CREs associated with drought stress and ABA signaling response were predicted in both sense and antisense positions. Four CRE classes were found in all genotypes: different ABA-and dehydration-responsive elements; several (basic leucine zipper) bZIP TFs also related to ABA signaling; an element regulated by calcium signals; several myeloblastosis (MYB) motifs, as well as a myelocytomatosis (MYC) and the SEF4 TF (Figure 4a, and Table 2 and Table S2). Among the CREs, EBOXBNNAPA was the most abundant element with a repetition range of 18 to 4 in the promoter region of each genotype, followed by ACGTATERD1 with a range of 8 to 6 repetitions (Table S2). Clear differences between individuals from cluster VI and individuals from the rest of the clusters (I, II, III, IV and V) were identified (Figure 4a and Table S2). Three different CREs families were only represented in the promoter region of genotypes from cluster VI: a heat shock promoter element (HSE); a low-temperature-responsive element (LTRE-1); and three MYB elements (Figure 4a, and Table 2 and Table S2). Six CREs were found in individuals from clusters I, II, III, IV and V, but not in individuals from cluster VI: the ABA-responsive element (ABRE) motif ABREDISTBBNNAPA; a T-box ACGTTBOX; the dehydration-responsive (DRE) element DRE1COREZMRAB17; the MYC elements MYCATERD1 and MYCATRD22, and the MYC recognition site G-box. Furthermore, individuals from clusters I, II, III, IV and V contained a GT3 box in their promoters (Figure 4a, and Table 2 and Table S2). The ethylene-responsive element (ERE), ERELEE4, was only identified in cluster I, but not in clusters II, III, IV and V (Figure 4a, and Table 2 and Table S2). Finally, five CREs were found only in individuals from clusters II, III, IV and V, but not in clusters I and VI: four different dehydration-responsive (DRE) elements CBFHV, DRE, DRE1COREZMRAB17 and DRECRTCOREAT; and a LTRE element LTRECOREATCOR15 (Figure 4a, and Table 2 and Table S2).
The study of the DREB2B TF gene promoter region was conducted in 'Garfi', 'GF-677', P. orientalis T4, P. vavilovi T4, P. bucharica F7T2, P. kotschii T1 and P. bucharica F7T1 from cluster I; P. kotschii T3 from cluster II; P. mira T2 from cluster III; and P. davidiana T3, 'Mira × Pecher', P. persica and 'Garnem' from cluster IV. CREs were located in both the sense and antisense orientation, presenting a more conserved sequence than the PpDhn2 gene promoter region. We identified in all individuals several ABA-, and dehydration-responsive elements also identified in the PpDhn2 promoter region however, we also identified an additional ABRE-element, namely the ABARE-element HEXMOTIFTAH3H4. Other CRE families were also identified in this analysis, including ERELEE4 motif; HSE element; the motif LTRE1HVBLT49; several MYB elements; the calmodulin-binding motif CAMTA3; SR1, and the MYC element EBOXBNNAPA (Figure 4b, and Table 2 and Table S3). The motif most frequently identified was the ACGTATERD1 being identified on 8 occasions. Interestingly, the cis-element MYB2CONSENSUSAT was only identified among cluster I members, but not in individuals belonging to clusters II, III, IV and V. (Figure 4b and Table S3). The bZIP TF DPBFCOREDCDC3 element was found in clusters I, II, III and IV, but not in cluster V (Figure 4b and Table S3). The motif SEF3MOTIFGM was presented in individuals from cluster II, III and IV (Figure 4b, and Table 2 and Table S3). Finally, the SEF4 element was only found in clusters I and V and interestingly, the position of this TF binding site was identified at different positions within the promoter regions of members of each cluster (Figure 4b, and Table 2 and Table S3).

Discussion
Drought tolerance must be one of the primary criteria when selecting a rootstock that we be cultivated in areas where water availability is limited. The drought response in plants is controlled by complex prototypical and physiological components. The relationship between genotype and phenotype is crucial in order to understand this response for enhancing the expression of desired traits related to drought tolerance, such as WUE. Therefore, increasing WUE in rootstocks is important to ensure future economical fruit tree production in less water-friendly environments [61]. Here, leaf ash content and carbon isotopic composition were carried out to estimate the long-term WUE [37,62] in peach and almond wild-relative species, in a number of interspecific Prunus hybrids and their parental genotypes. Furthermore, in the same plant material, a molecular genetics approach was used to assess the promoter region landscapes of two drought-responsive genes involved in key responsive pathways, the effector gene, PpDhn2 and the gene encoding the DRE2B TF was performed. Our data revealed the genotypes with the highest WUE ( Figure 1); as well as documenting the variability between the PpDhn2 and DREB2B genes for each assessed plant line. (Figures 2-4).
Leaf ash content and ∆ 13 C were returned a positive correlation, and further; the ratios were similar to ratios obtained in previous reports in apple [37] and peach [38]. Also, it is known that ash content and ∆ 13 C can be used to evaluate long-term WUE in fruit trees [37]. Based on these criteria, all P. davidiana, P. mira and the almond wild-relative species presented higher WUE than the hybrid genotypes and their parentals. This improved WUE could be due to the natural adaptation of these species to severe conditions, which represent a different growing strategy than the hybrid rootstocks studied. These wild-relative species originate from the arid steppes, deserts, and mountainous areas [63][64][65][66] in which the lack of water is a common factor. In these species stomata closure, a proven adaptation to water restrictive environments, results in decreased ∆ 13 C, and therefore; ∆ 13 C would be a reliable phenotypic measure of long-term drought survival [30,40,67]. Both leaf ash content and ∆ 13 C appear to be suitable phenotypic parameters for assessing drought stress in Prunus. Further, Brendel et al. [61] were able to identify different quantitative trait loci (QTLs) for WUE as estimated by leaf ∆ 13 C in Quercus robur L.
Promoter analysis of the PpDhn2 and DREB2B genes revealed the presence of CREs associated with ABA-and dehydration-response. We found that all individuals shared ABREs in both gene promoter regions, although the number of ABREs varied depending on the genotype and the gene promoter sequence. ABRE is the most abundant CRE in ABA-responsive gene expression, and at least two copies of an ABRE are necessary for ABA-responsive induction of transcription [3]. Different MYB motifs and a MYC element were also distributed throughout the promoter regions of both genes in each assessed genotype. Both MYB and MYC TF binding sites have been associated with drought responsiveness and are fundamental to ABA-and drought-responsive expression [68][69][70]. Further, specific CREs for PpDhn2 gene from each cluster were identified. This finding indicates additional expression regulation opportunities for the PpDhn2 locus compared to the DREB2B gene among the plant lines analyzed, which indicate more diversity of that promoter region compared to DREB2B promoter along the studied genotypes. The nine almond individuals from this cluster, that belong to P. bucharica, P. zabulica and P. vavilovi also had in common a HSE, a LTRE, as well as three MYB motifs that were more abundant elements than in the other individuals. However, we identified in the promoters of individuals from clusters I, II, III, IV and V an ABRE motif, one DRE element, three MYC recognition sites, and a GT3 box, the CRE to which the negative regulator of WUE Trihelix TF binds to, to regulate the expression of the SDD1 locus [71]. Other specific CREs were found in the almond wild-relative species and in a P. mira T1 belonging to clusters II, III, IV and V. Their PpDnh2 promoter regions harboured 3 DREs, one C-repeat binding factor and one LTRE. In previous reports, the promoters of the PpDhn2 gene in peach was studied and founding ABRE and MYC CREs, but not MYB binding sites, nor DRE/CRT or LTRE elements on the sense strand [14,15]. In our work, the DRE motifs and LTRE were located in the negative strand, but not in the sense strand. The influence of this cis-element and its orientation in gene promoter regions remains an area of debate with both dependent and independent orientation motifs [72,73]. Recent research did not find evidence of the influence of motif orientation in regulatory gene expression in a number of cis-elements studied in A. thaliana [74]. Similar to our findings, Bassett et al. [14] and Wisniewski et al. [15] observed that no DRE elements were found in positive sense in the PpDhn2 promoter and suggested that the absence of this cis-element was related to the lack of expression in response to cold. However, other reports have confirmed the presence of one DRE/CRT element in the promoter region of the Y n SK n dehydrin class, which includes PpDhn2 [75]. In spite of this observation, it is known that Y n SK n dehydrins are not expressed in response to cold. García-Bañuelos et al. [76] concluded that MdDhn, which shows great similarity with PpDhn2, was accumulated after a period of acclimation in apple trees. Based on that, Zolotarov and Strömvik [75] affirmed that cold-induced expression of Y n SK n -type dehydrins would not be detected in some cases because of a limited time of exposure to low temperature. So that, the presence of the DRE and LTRE elements found in the anti-sense position in our individuals could have some effect in the expression of PpDhn2 in a possible response to cold.
All species shared essentially the same CREs in their DREB2B promoter region, evidencing a lower variability among the promoter sequences of every studied genotype. Beside the elements described before, we identified in sense orientation a HSE element, which binds to heat shock factors responsible for heat stress tolerance [77]. Moreover, although several reports demonstrated that DREB2B is not induced by low temperatures [1,17,67], a LTRE element, an important motif for the induction of cold regulated genes [78], was located upstream of the transcription start codon. The presence of ABREs motifs in the promoter region of DREB2B denoted the implication of this TF in ABA-dependent signal transduction pathway [79]. In the literature, the relation between dehydrin expression and an increase of WUE in cereals has been demonstrated. Sivamani et al. [80] confirmed an improvement of biomass and WUE in transgenic barley plants expressing HVA1 gene under drought conditions. Furthermore, Melišová et al. [33] suggested that elevated expression of the HvDhn4 gene, which is also a Y n SK n -type dehydrin and similar to PpDhn2, was associated with the high WUE observed in a drought-tolerant variety of barley at 12 h after ABA treatment. Moreover, DREB TFs improved tolerance to abiotic stress in transgenic plants by regulation of genes involved in abiotic stress responses, so DREB TFs could increase WUE under water deficit conditions [81]. Furthermore, tobacco transgenic lines with overexpression of SbDREB2A, homolog to DREB2B, showed higher WUE and also, a higher expression of different dehydrins including ERD10B, ERD10D and LEA5, conferring drought tolerance [82].
The promoter regions of both genes also contained multiple cis-elements related to other plant responses. For example, SORLIP or I-box motifs which are usually upstream elements are regulated by light and the circadian clock; other elements are associated with to development responses, including the O 2 -site involved in zein metabolism regulation and a CAT-box linked to meristem expression. Some motifs are related to hormone responses including an ARR1AT motif (cytokinin response regulator), several CGTCA-motifs involved in methyljasmonate-responsiveness, and the GARE-motif associated with gibberellin-responsiveness, as well as others linked to additional stress. The presence of these CREs could reflect the role of DREB2B and PpDhn2 in other processes in addition to cold and drought [79,[83][84][85].
Based on our data, the presence of DRE elements in PpDhn2 promoter belonging to the genotypes from clusters I, II, III, IV and V, suggest that PpDhn2, in addition to the ABA-dependent pathway, is also induced in an ABA-independent manner by the binding of DREB2B TF to these DRE elements under drought conditions [18][19][20][21]. From our promoter analysis of the PpDhn2 gene, we are unable to find a definitive association between PpDhn2 and different WUE and Ash measurements in a variety of Prunus genotypes.
In conclusion, our phylogenetic classification of the Prunus collection based on the CREs identified in the promoters of both genes showed a clear distinction between peach relatives and almond relatives. Nevertheless, in both PpDhn2 and DREB2B phylogenetic trees, it was demonstrated that P. davidiana (the highest WUE), a peach wild-relative specie [86], was closer to the other parental plant lines and their hybrids (lowest WUE). Our results show that phenotyping data is useful as an early selection criteria and that there are relatively few differences in promoter regions of the genes examined here, which suggests that improving drought survival could be accomplished by introgressing one or more of these genes/promoters into standard Prunus rootstock germplasm. According to our results, almond wild-relative species would be the genotypes with the best drought resistance potential for incorporation into future breeding programs aimed at generating new cultivars with drought tolerance potential.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4395/8/4/42/s1, Table S1: List of the 48 individuals used in this study, Table S2: Cis-regulatory elements of PpDhn2 promoter gene region in each of the cluster-representative individual. Cells in grey color, CREs outside of the first 1000 bp 5 of the translation start site. In red color, CREs in negative strand. In black color, CRES in positive strand, Table S3: Cis-regulatory elements of DREB2B promoter gene region in each of the cluster-representative individual. Cells in grey color, CREs outside of the first 1000 bp 5 of the translation start site. In red color, CREs in negative strand. In black color, CRES in positive strand.