The Suppression of WRKY44 by GIGANTEA-miR172 Pathway Is Involved in Drought Response of Arabidopsis thaliana

Water availability is an important environmental factor that controls flowering time. Many plants accelerate flowering under drought conditions, a phenomenon called drought escape. Four pathways are involved in controlling flowering time, but which ones participate in drought escape is not yet known. In this study, plants with loss-of-function mutations of GIGANTEA (GI) and CONSTANS (CO) exhibited abnormal drought-escape phenotypes. The peak mRNA levels of GI and FKF1 (Flavin-binding Kelch domain F box protein 1) and the mRNA levels of CO and FT (Flowering locus T) changed under drought stress. The microRNA factor miRNA172E was up-regulated by drought stress, and its up-regulation was dependent on GI, while other miRNA172s were not. Water-loss analyses indicated that gi mutants were more sensitive while miRNA172 over-expressing (miRNA172-OX) plants were less so to drought stress than wild-type plants. Digital gene expression and real-time PCR analyses showed that WRKY44 was down-regulated by GI and miRNA172. The WRKY44 protein could interact with TOE1 (a target of miRNA172) in a yeast two-hybrid system. We proposed that GI–miRNA172–WRKY44 may regulate drought escape and drought tolerance by affecting sugar signaling in Arabidopsis.


Introduction
Unlike most animals, plants are sessile organisms. They cannot move to escape the biotic and abiotic stresses that threaten them throughout their life cycles. To adapt to unfavorable and sometimes unexpected conditions, plants have evolved many flexible survival strategies, one of which is the control of flowering time [1,2]. Flowering time is finely tuned because it is of critical importance to successful reproduction and maximal seed set [3].
Flowering time is regulated by multiple environmental and endogenous factors [4]. In general, these factors can be grouped into four genetic pathways: the photoperiod, phytohormone, vernalization, and autonomous pathways [5,6]. These pathways ultimately crosstalk at common targets, such as Flowering Locus T (FT) and Leafy, to promote the transition from vegetative to reproductive phase [4,7]. Also, several microRNAs (miRNAs) participate in these pathways to maintain homeostasis and accurate flowering time, i.e., miRNA159 in the phytohormone pathway [8], miRNA156 in the autonomous pathway [9,10], and miRNA172 in the photoperiod pathway. Notably, miRNA156 inhibits the transcription of miRNA172b via SPL9 and, redundantly, SPL10 [11].
In this study, we considered the regulation of flowering time under drought stress. As the greenhouse effect causes global climate warming, drought is becoming a major agronomic threat to crop yields [18,19]. Under excessively dry conditions, plants must balance drought resistance and escape (via reproduction) to maximize the probability of genetic survival. Thus, water availability affects flowering time in many angiosperms [14]. Many terrestrial plants flower earlier when water is deficient, a phenomenon well studied in wheat, Brassica, and Arabidopsis [20,21], but which pathway is involved in drought escape is not yet clear.
Drought has been reported to alter physiological sugar levels. An increase in soluble sugar with a decrease in leaf osmotic potential was observed during drought [22]. The reduced osmotic potential may prevent moisture loss. Also, sucrose promotes flowering in many plant species [23][24][25]. Feeding sucrose in photosynthetic amounts reversed the floret abortion induced by drought stress [26].
Accelerated flowering under drought will reduce crop yields. Therefore our study focused on the genetic mechanism of accelerated flowering under drought stress in Arabidopsis. We observed and characterized drought escape and defense in different genotypes of Arabidopsis. Our results confirmed that photoperiod factor GIGANTEA (GI) was involved in both drought escape and drought resistance. Also, GI-miR172 may function in sugar signaling by down-regulating WRKY44. This study provides a foundation for researching reduced crop yields under long-term drought.

Plant Materials
The Arabidopsis thaliana ecotypes Col-0 and Ler-0 were used as wild types (WT). Mutants gi, co, gai were in the Ler-0 background, and flc-3 was in the Col-0 background. The miRNA172s-OX lines were in the Col-0 background.

Growth Conditions and Drought Treatment
WT and mutant plants were grown in a climate-controlled culture room at 23-25°C with a relative humidity of 40-60% under long day (LD) conditions (16 h light/8 h dark). The plants were grown on a medium containing 9:3:1 vermiculite: sphagnum peatmoss: perlite. The medium was saturated with tap water containing diluted (1000-fold) Hyponex during the first watering. Thereafter, the plants were irrigated with tap water. For a control (CK), plants were thoroughly watered every 4 d without water-logging the soil. For the drought (DR) treatment, the plants were not watered until samples were collected.

Flowering Time Estimation
WT plants (Col-0 and Ler-0) and four loss-of-function mutants were used to estimate the flowering time under drought stress. Mutants of the photoperiod pathway factor GIGANTEA (gi, CS181), the phytohormone pathway factor Gibberellic Acid Insensitive (gai, CS63), and the autonomous pathway factor Flowering Locus C (flc-3, SALK_140021) were purchased from the Arabidopsis Biological Resource Center (http://abrc.osu.edu/). Mutants of the photoperiod pathway factor CO (CONSTANS) was kindly provided by Hongquan Yang (Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences).
Drought treatment began about 10 d before normal (CK) flowering. Specifically, at an age of 10 d for WT, 23 d for gi and co, 10 d for flc, and 14 d for gai. The flowering time was counted in days. Three biological replicates were performed. WT (Col-0) was also grown under short day (SD: 8/16 h; 23-25°C, 40-60% humidity) conditions and DR treatment beginning at 35 d of age.

Expression Analyses of Photoperoid Pathway Genes
Real-time PCR analysis of rhythmic expression of photoperiod pathway-related genes. Rhythmic expression of four photoperiod pathway genes, GI, FKF1 (Flavin-binding Kelch domain F box protein 1), CO (CONSTANS), and FT, was detected by real-time reverse transcription PCR (qRT-PCR). Expression of ACTIN11 was the control for all the qRT-PCR. Primer sequences are listed in Table 1. WT (Ler-0) plants were grown under LD. CK and DR treatments were performed as described above. Drought treatment began at 10  For semi-quantitative RT-PCR of primary miRNA pri-miRNA172, leaf samples were harvested 4 h after dawn at 2 d intervals for 8 d. There were three biological replicates. RNA preparation and cDNA analysis were carried out as described above using the primers listed in Table 2. The cycle conditions were: 3 min at 94°C; 28× (40 s at 94°C; 40 s at 55°C; 40 s at 72°C); and 5 min at 72°C.
To assay mature miRNAs, treatments were as for semiquantitative analysis of pri-miRNA172, and samples were collected after 8 d of DR treatment. Both Ler-0 and the gi mutant were examined. Each treatment was replicated twice, with three samples per replicate. RNA was prepared as described above. cDNA was synthesized using gene specific primers: U6 5'ATT TGG ACC ATT TCT CGA TTT GT3'; miRNA172A/B 5'ATT TGG ACC ATT TCT CGA TTT GT 3'; miRNA172C/D 5'GTC GTA TCC AGT GCG TGT CGT GGA GTC GGC AAT TGC ACT GGA TAC GAC CTG CAG 3'; and miRNA172E 5'GTC GTA TCC AGT GCG TGT CGT GGA GTC GGC AAT TGC ACT GGA TAC GAC ATG CAG3'. The SYBR method (2X SuperArray PCR master mix (Cat. No. PA-112, SABiosciences, Valencia, CA, USA) was used in an ABI PRISM 7900 system (Applied Biosystems). U6 was the internal control. Primers for the following PCRs are shown in Table 3. Reaction conditions were as follows: 10 min at 95°C and 40× (15 s at 95°C; 60 s at 60°C); the annealing temperature for miRNA172 was 60°C. For the melting curve, the reaction conditions were as follows: 2 min at 95°C; 20 s at 60°C; 10 s at 99°C, decreasing to 60°C.

Measurement of Transpiration Rate and Water Loss
Arabidopsis lines (Col-0 ecotype) over-expressing miRNA172 were used in this study. The primers for gene amplification and enzymes for cloning are listed in Table 4. The miRNA172 fragments were cloned into pCAMBIA1301 expression vector and transgenesis was carried out by the floral-dip method mediated by Agrobacterium [27]. Seeds of transgenic lines over-expressing miRNA172 (miRNA172-OX) were selected on MS agar medium with 20 mg/L hygromycin. E1-2 (miRNA172e over-expressing), D6-3 (miRNA172d overexpressing), and A1-10 (miRNA172a over-expressing) were transgenic homozygote lines.
Ler-0, gi, A1-10 (miRNA172a-OX), D6-3 (miRNA172d-OX), and E1-2 (miRNA172e -OX) were either CK or DR treated at 10 d of age. Samples were collected after 10 d. Leaves of similar developmental stage (3 rd -5 th true rosette leaves) were collected and placed abaxial-side up on open Petri dishes. Transpiration rate and water loss were measured according to Kang et al. [28]. Briefly, the leaves were weighed at hourly intervals. The transpiration rate was represented by the change in weight over time for CK-treated plants, i.e., weight/(fresh weight), while water loss was represented by the lost weight for the DR-treated plants, i.e., (fresh weight -weight)/(fresh weight). Three biological replicates were performed.

Digital Gene Expression Analysis of gi under Drought
WT (Ler-0) plants and gi mutants were CK or DR treated as described above for the qRT-PCR analyses. Samples were collected from two independent treatments. Then digital gene expression (DGE) analysis was performed with all four combinations of genotype and treatment. In detail, we extracted 6 μg of total RNA, purified mRNA via Oligo(dT) magnetic bead adsorption, then used Oligo(dT) to guide  Table 3. Primers for microRNA assay of mature miRNA172s. reverse transcription to synthesize double-stranded cDNA. NlaIII was used to cut the CATG sites in the cDNA, then cDNA fragments with 3' ends were purified with magnetic-bead precipitation, and Illumina adapter 1 (Illumina, San Diego, CA, USA) was added to their 5' ends. The junction of Illumina adapter 1 and the CATG site is the recognition site of MmeI, which cuts 17 bp downstream of the CATG site, producing tags with adapter 1. After removing 3' fragments via magnetic-bead precipitation, Illumina adapter 2 was introduced at the 3' ends of tags, producing tags with different adapters at their ends to form a tag library. After 15 cycles of linear PCR amplification, 85 base strips were purified by 6% TBE polyacrylamide gel electrophoresis. These strips were then digested, and the single-chain molecules were fixed onto the Solexa Sequencing Chip (flowcell). Each molecule grew into a single-molecule cluster sequencing template through in situ amplification. Then, labeled nucleotides were added and sequencing by synthesis was performed. Each tunnel generated millions of raw reads 35 bp length. The raw data were normalized by the number of tags per million. The main reagents and supplies were Illumina Gene Expression Sample Prep Kit and Solexa Sequencing Chip (flowcell), and the main instruments were Illumina Cluster Station and Illumina Genome Analyzer System. We focused on genes that showed log 2 ≥ 1 (the relative expression levels between WT and gi under DR) with false discovery rate values ≤ 0.001.

Phylogenetic Analysis
We performed phylogenetic analysis on the sequences listed in Table 6. Sequences were aligned using ClustalX 1.8 [29] and a phylogenetic tree was constructed with MEGA5 [30] using neighbor joining method [31].

Yeast Two-hybrid System
The yeast host strain Y2H Gold (Clontech, Mountain View, CA, USA) was transformed with pGBKT7-TOE1 as the bait. The Y187 strain was transformed with the plasmid pGADT7 with a full-length open reading frame of WRKY20, WRKY44, or WRKY74; an empty pGBKT7 was the control. Transformants with BD (Binding Domain) and AD (activation domain) were mated on 2× YPDA medium at 30°C [32]. Mated colonies were picked and mixed with 5 mL 0.9% NaCl, then spotted on SD/-Leu/-Trp/-His/-Ade/X-α-gal/AbA agar media. The plates were cultured at 30°C and photographed after 2-3 d.

ABA Treatment
The two-weeks old seedlings were treated with ABA (50μM) or ddH2O. Samples were collected every 12hr, from 0 to 48hr. FT was analyzed by relative-quantitative RT-PCR. PCR conditions were as following: 3 min at 94°C; 28× (40 s at 94°C; 30 s at 55°C; 15s at 72°C); 5 min at 72°C. The primer for FT and ACTIN11 were the same as that for the real time PCR. Table 4. Primers of miRNA172 amplification for transgenic plants.

Statistical Analyses
Because we compared two treatments (CK and DR) with small sizes and equal variances, t-tests were used for all statistical tests of differential gene expression.

The Photoperiod Pathway Mediated by GIGANTEA Might Be Involved in Early Flowering under Drought Stress
We examined the drought escape of Arabidopsis carrying mutations in genes of different flowering pathways. Flowering time (mean ± SE of three replicates, with three samples per replicate) was calculated as days after germination. Plants were either watered normally (CK) or deprived of water (DR) beginning about 10 d before normal (CK) flowering. Relative humidity throughout the experiment ranged from 40-60%.
In this experiment, WT Col-0 and Ler-0 plants under LD conditions flowered significantly earlier (P<0.05) under DR than under CK ( Figure 1A,D). The gai and flc mutants also flowered significantly earlier (P<0.05) under DR, but flowering of the gi and co mutants was not induced by drought ( Figure 1B,D). The gi and co plants withered after 10 d of DR. Because the onset and duration of drought treatment (lasting 10 d and beginning 10 d before normal flowering) of gi and co was the same as that of WT and other plants, these results indicated that the photoperiod pathway might be involved in early flowering under drought. WT plants did not flower earlier under DR and SD conditions ( Figure 1C), indicating that day length is important for early flowering under drought.

Expressions of Photoperiod Pathway Genes Changed under Drought Conditions
Genes in the photoperiod pathway are transcribed rhythmically. Several important genes, including GI, CO, and FT, are the main factors in this pathway. A recent study indicated that FKF1 cooperated with GI to activate CO [33]. Therefore, changes in the mRNA levels of GI, CO, FT, and FKF1 were detected by qRT-PCR ( Figure 2). The peak levels of GI and FKF1 mRNAs were up-regulated under drought stress, while the expressions of CO and FT were reduced.
GI expression peaked 4 h before dusk under both CK and DR (Figure 2A). The maximum level under DR was significantly higher than that of CK. After 5 d of recovery with watering, both mRNA rhythm and levels were similar in DR plants to CK plants. FKF1 was also up-regulated under DR ( Figure 2B). Unlike GI, FKF1 expression was developmentally controlled, because it was up-regulated with developmental age under both CK and DR. After recovery for 5 d, the level of FKF1 under DR was substantially higher than that under CK, which may indicate that DR accelerated aging.
For CO under CK, there was one peak within a 24-h cycle ( Figure 2C). The rhythm was not changed under DR during the first 2 d of sample collection (days 11 and 12 of DR), but on the 13 th day of DR treatment, the expression at the peak (late) time was reduced. After recovery, the circadian expression of CO was recovered. FT was apparently down-regulated with the Table 6. WRKY sequences used in phylogenetic analysis.

Gene symbol Accession number
Arabidopsis thaliana  intensified DR condition ( Figure 2D). There were two expression peaks for FT during one 24-h cycle. These two peaks were reduced and ultimately disappeared under DR. When water was restored, the transcription of FT was recovered.

GI Promoted the Level of miRNA172E under Drought Conditions
Because the expression of CO was not promoted under DR as was that of GI, we focused on miRNA172, a factor downstream of GI. The pri-miRNA level of miRNA172E was reduced under DR ( Figure 3A), while its mature miRNA level increased ( Figure 3B). These data suggested that the processing efficiency of miRNA172E was enhanced under drought stress.
In the gi mutant, up-regulation of mature miRNA172E ( Figure  3D) and down-regulation of pri-miRNA172E ( Figure 3C) under DR were not detected, indicating that the enhanced processing efficiency of miRNA172E was dependent on GI.

GI Inhibited the Expression of WRKY44 under Drought Conditions
In addition to the abnormal drought escape of the gi mutant, we observed that gi was more sensitive to drought stress than WT ( Figure 4A). Although their transpiration rates were similar( Figure 4B), the gi plants lost more water than WT in the early stages of dehydration( Figure 4C). The difference was most stark after 0.5 h, while water loss in gi and WT was similar at later stages (after 1 h). However, miR172s-OX plants lost much less water than both WT and gi. The levels of primary and mature miRNA172s in the transgenic plants, as well as their phenotypes, are shown in Figure S1 and flowering times of the transgenic plants were calculated in Table S1.
Given the higher drought tolerance of miRNA172 overexpression plants, we can conclude that GI-miRNA172 may be involved in drought tolerance of Arabidopsis ( Figure 4A,C).
DGE analysis was carried out to probe the differentiallytranscribed genes in gi mutants under DR to gain insight into the relationship between drought defense and escape ( Figure  5A). The resulting Venn diagram ( Figure 5B) identified crosstalk and differential gene expression between WT and gi under CK and DR. Under DR, 1,218 genes were up-regulated in WT but not in gi, while 407 were down-regulated in WT but not in gi. At the same time, 785 genes were specifically up-regulated The Suppression of WRKY44 by GIGANTEA-miR172 PLOS ONE | www.plosone.org and 798 were specifically down-regulated in gi under DR. These data implied that some factors were differentially regulated by GI under drought stress.
According to DGE analysis, several WRKY (WRKY DNA-BINDING PROTEIN) family members exhibited significantly differential expression between gi and WT ( Figure 5C). Under CK conditions, 20 WRKY genes were up-regulated more than two-fold in gi compared with WT, while 16 were downregulated. However, after DR, only four WRKY genes had fold increases of two or more in gi, while 23 were expressed less ( Figure 5C).
The expressions of nine WRKY family members, representing different subfamilies [34,35], were further examined by qRT-PCR ( Figure 6). WRKY44, WRKY 20,WRKY 40, and WRKY 51 were maintained at much higher levels in gi mutants than in WT plants under DR. This finding indicated that GI suppressed the expression of these genes under DR.
WRKY44 was unique in that it was also greatly up-regulated (≈ 100-fold) in WT under DR compared to CK, although not as much as in the gi mutant (≈ 200-fold). In other words, WRKY44 was constitutively suppressed by GI, while WRKY20, WRKY40, and WRKY51 were suppressed only under DR. In contrast, WRKY54 and WRKY72 were positively activated by GI under CK, while WRKY74 was positively regulated under both CK and DR. Regulation of WRKY19 and WRKY65 seemed to be independent of GI.

Interaction between WRKY Protein and TOE1
Because a previous report indicated that subgroup 1 was involved in sugar signaling [36], we detected the interaction of members of this subgroup (i.e., WRKY44 and WRKY20) and a member of another subgroup (WRKY 74) with TOE1 (Target of EARLY ACTIVATION TAGGED 1), a target of miRNA172 and suppressor of flowering. In a yeast two-hybrid system, WRKY44, which was suppressed by GI, was able to interact with TOE1, while WRKY20 and WRKY74 were not (Figure 8).
We examined the expressions of WRKY44 and its comembers in subgroup 1, including WRKY20 and WRKY51, in co mutant and miRNA172-OX line (Figure 9). The level of WRKY44 was significantly reduced in miRNA172-OX plants (P = 0.015 under CK; P = 0.027 under DR) (Figure 9). The levels of WRKY51 and WRKY20 were unchanged in miRNA172-OX. The down-regulation of WRKY44 in miRNA172-OX plants was consistent with its up-regulation in gi mutants, indicating that GI and miRNA172 were in the same pathway suppressing WRKY44. But in co, the level of WRKYs was similar to WT under both CK and DR conditions (Figure 9).

Discussion
Water deficit affects flowering time in many angiosperms [14]. Many plants accelerate flowering under drought conditions, a phenomenon well studied in wheat, Brassica and Arabidopsis [20,21]. Given that earlier flowering under drought will reduce crop yields, we examined the genetic mechanism of this acceleration in Arabidopsis. Our results indicated that the photoperiod factor GI might be involved in drought-induced early flowering in Arabidopsis. Loss-of-function mutants of GI and CO could not flower under drought stress. Drought led to increased peak levels of GI and FKF1. No previous paper has reported a correlation between drought stress and circadian rhythm. But TOC1 (Time of CAB expression 1 ), an important gene in circadian control, has been reported to be a molecular switch connecting the circadian clock with plant drought responses via mutual regulation with ABAR (ABA-Related gene) [37]. GI and TOC1 are both circadian regulators, with GI activating TOC1 and TOC1 repressing GI [38,39]. The changed expression of GI and its related genes under drought might have resulted from the interaction between GI and TOC1.
Drought reduced the peak levels of FT, which was unexpected. The down-regulation of FT as drought conditions worsened may be related with the increased concentration of endogenous abscisic acid (ABA). We performed an ABA treatment and analyzed the expression of FT as described above and found that ABA (50 μM) inhibited the level of FT ( Figure S2). This finding was consistent with our rhythmical expression data. For example, the reduction of FT was not so apparent at the10 th d of DR treatment (i.e., the first day of sample collection). But beginning on the 11 th day, FT levels declined day by day (Figure 2). Nevertheless, by the time samples were collected, the DR-treated plants had flowered, so the reduction of FT did not affect the flowering time of Arabidopsis. Thus, the suppression of FT by DR may result from increased ABA levels in the plant.
We further investigated miRNA172, an important non-coding RNA in the photoperiod pathway that is controlled by GI [40]. Although both level and function of miRNA172 are reported to be enhanced during drought in maize, Arabidopsis, and potato (Solanum tuberosum) [41][42][43][44][45], differential expression of its family members and other regulating mechanisms have not been studied. In this study, genetic and molecular analyses indicated that miRNA172E exhibited the greatest response to drought, with enhanced processing efficiency because of the decreased precursor levels and more mature miRNA172E under DR in WT plants. The changes in precursor and mature miRNA172E levels in the gi mutant indicated that enhanced miRNA172E processing under drought was dependent on GI.
The photoperiod pathway senses light via plant aerial parts, especially leaves [41,46,47]. However, water availability is assessed by roots [14]. Water availability signals may be transmitted from roots to leaves. Another group of photoperiod genes, cryptochromes (CRY), have been indicated to be related to drought tolerance [48]. In addition, CRY2 positively regulates GI in the photoperiod pathway [49]. The involvement of cryptochromes may explain why GI-miRNA172 was implicated in drought response.
According to our observations, the gi mutant was more sensitive and miRNA172-OX was less sensitive to drought than WT plants, indicating that GI-miRNA172 affects drought defenses other than escape. GI is known to protect plants from several abiotic stresses, including cold [50] and oxidative stress [50][51][52]. The involvement of GI in drought defense may be related to oxidative stress resulting from dehydration.
Interestingly, some WRKY genes, which belong to a defense-related gene family, were characterized as downstream factors of the GI-miRNA172 pathway. The most significant was WRKY44, which was significantly suppressed by GI and miRNA172.
Among the targeted genes of miRNA172, TOE1 is the most influential because a single mutant of toe1 exhibited early flowering [40,53]. We performed yeast two-hybrid screening to detect the interaction of WRKYs and TOE1. WRKY44 could interact with TOE1, a target of miRNA172. This further confirmed the regulation of WRKY44 by GI-miRNA172.
The WRKY superfamily can be divided into seven subgroups according to the number of WRKY domains and features of their zinc-finger-like motifs [54]. WRKY44 belonged to subgroup 1 according to our phylogenetic analysis. One recent study indicated that this group of barley (Hordeum vulgare) WRKYs were involved in sugar signaling [36]. GI-miRNA172 may be involved in sugar signaling by inhibiting WRKY44. Consistent with this hypothesis, excess starch accumulation has been observed in leaves of the gi mutant [55], and research indicated that sugar deficiency was responsible for the sensitivity of the gi mutant to freezing [56]. GI-miRNA172 may affect sugar concentration by inhibiting WRKY44.
According to the qRT-PCR, expression of WRKYs was not changed in the co mutant. Because co exhibited a similar drought phenotype to that of gi, and CO and miRNA172 are two independent factors downstream of GI [40], CO may affect drought escape by regulating other factors.
In conclusion, plants prepare to survive increasing drought stress via two strategies. One is to adjust the osmotic potential to defend against impending dehydration. The other is to bloom early to ensure the perpetuation of their genes. Sugar is an ideal signal that can link both strategies, because of its role in both osmotic adjustment [57,58] and the transition from vegetative to reproductive development [59]. This study indicated that GI-miRNA172 and WRKY may be factors connecting these two pathways. Figure 10 summarizes a working model of this hypothesis. In the LD and DR condition, increasing peak expression of GI promoted the processing of  . Transcriptional level of WRKY20, WRKY44, and WRKY51 in co and miRNA172-over-expressing plants (miRNA172-OX) under standard (CK, white rectangles) and drought (DR, black rectangles) conditions. Controls for the co mutant and miRNA172-OX was Col-0, the wild type in their respective ecotype backgrounds. Results are averages of three biological repeats. * Significantly different (P<0.05) expression between miRNA172-OX and WT under both CK and DR conditions. E1-2 line was used as miRNA172-OX. DR treatment began from10 day age and maintained for 10 days. miRNA172. MiRNA172 could suppress the levels of WRKY44 and TOE1, which encode interactive proteins. Because WRKY44 is involved in sugar metabolism and signaling, GI-miRNA172 might function in drought escape and defense by affecting sugar signaling. In future studies, the mechanism of interaction between WRKY44 and TOE1 should be examined to investigate the function of TOE1, an important photoperiod factor, in sugar signaling.
Another important drought-induced factor, ABA, may be involved in both drought responses. ABA can promote drought tolerance in plants [60]. LD conditions promoted ABA levels, indicating that ABA was regulated by the photoperiod pathway [61], but high ABA level will delay flowering [62]. Considering that LD favors flowering, this inconsistency suggests that the concentration of ABA may be a signal for both drought tolerance and drought escape. This is similar to sugar, in that sugar at photosynthetic amounts will promote the floral transition [26], but will inhibit it at excess concentrations [63].