Metabolic and growth responses of maize to successive drought and re-watering cycles

Agricultural Water Management 172 (2016) 62–73

Contents lists available at ScienceDirect

Agricultural Water Management journal homepage: www.elsevier.com/locate/agwat

Metabolic and growth responses of maize to successive drought and re-watering cycles Caixia Sun a,∗ , Xiaoxiao Gao a , Xing Chen a , Jianqi Fu a , Yulan Zhang b a b

College of Life and Health Sciences, Northeastern University, Shenyang 110169, China Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang 110016, China

a r t i c l e

i n f o

Article history: Received 19 August 2015 Received in revised form 16 April 2016 Accepted 18 April 2016 Keywords: Cyclic drought Metabolomic profile NMR Recovery

a b s t r a c t In light of an increase in the severity and frequency of episodic drought predicted due to climate change, the mechanisms underlying plant responses to repeated drought and recovery cycles need to be understood. Maize (Zea mays L.) plants of two inbred lines were subjected to two cycles of drought and re-watering and then compared with plants that were watered daily. Changes in the metabolome and growth were monitored at multiple time points during the experiment. The extent of recovery in plant growth after re-watering strongly depended on the plant lines and drought cycles. Both full and partial recovery was observed. The variation in the pattern and extent of the metabolomic profile and in metabolite levels initiated by re-watering during two drought cycles was complex and diverse. Correlation-based network analysis indicated that maize plants required more coordinated and extensive metabolic shifts to cope with drought in the second cycle than in the first cycle. Metabolic pathways in the maize plants returned to their normal status at different rates during recovery. The results provide valuable insight into the growth, biochemical, and metabolic mechanisms used by maize to adapt to cyclic drought. The analysis is also useful in indicating specific traits to be targeted in breeding programs aiming to better adapt the crop to climatic changes. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Soil drought because of inadequate rainfall and fluctuations in soil moisture has become an acute problem that constrains plant growth and productivity, in numerous regions worldwide, particularly in growing regions that already have a tendency to water deficits (Boyer, 1982; Ahuja et al., 2010; Dawson et al., 2011). Plants respond to water deficit through complex mechanisms, which include differential gene expression, changes in biochemical metabolism and morphological adaptation (Chaves et al., 2003; Harb et al., 2010; Pinheiro and Chaves, 2011; Lipiec et al., 2013). To cope with drought, plants can enhance their capacity for aquiring water or conserve water by increasing their root-to-shoot ratios

Abbreviations: BA, biomass allocation; BAS, ratio of shoot dry mass to total dry mass; BAR, ratio of root dry mass to total dry mass; BM, biomass weight; C1, control in the first cycle; C2, control in the second cycle; D1, drought stress in the first cycle; D2, drought stress in the second cycle; D-W, drought and re-watering treatment; LA, total leaf area per plant; PCA, principal component analyses; NMR, nuclear magnetic resonance spectroscopy; W1, re-watering in the first cycle; W2, re-watering in the second cycle. ∗ Corresponding author. E-mail address: [email protected] (C. Sun). http://dx.doi.org/10.1016/j.agwat.2016.04.016 0378-3774/© 2016 Elsevier B.V. All rights reserved.

under water deficits. This route has been considered an adaptive strategy because a larger investment in roots improves water absorption and a decrease of leaf area reduces the area for transpiration (Blum, 1996; Bargali and Tewari, 2004; Xu et al., 2010). In addition, plants could also improve their ability for osmotic adjustment and alter metabolic pathways to tolerate or resist water deficits (Shulaev et al., 2008; Pinheiro and Chaves, 2011). Multiple metabolic processes are widely recognized to being involved in the adaptive response to water stress (Mahajan and Tuteja, 2005; Verslues and Juenger, 2011; Krasensky and Jonak, 2012). Drought disrupts metabolic processes, including CO2 assimilation and metabolic impairment, and ultimately decreases the rate of biomass accumulation (Flexas et al., 2004). Moreover, turgor is maintained by osmotic adjustment under drought conditions and requires the transport and synthesis of functionally important osmolytes and compatible solutes, such as proline, soluble sugars, glutamate, polyols, and glycine-betaine (Foyer et al., 1998; Chaves et al., 2003; Sicher et al., 2012; Sun et al., 2015). Several compounds that are involved in primary metabolism are key precursors of stress-related metabolites, such as flavonoids, phenylpropanoids, and proteins (Dixon and Pavia, 1995). Another function of primary metabolites in stressed plants is the mitigation of oxidative stress because of the formation of reactive oxygen

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

species in plants in response to water stress (Shen et al., 1997; Dixon, 2001; Ramakrishna and Ravishankar, 2011). Using rapidly developing metabolomic techniques, recent studies have identified different types of metabolites that are associated with drought stress responses in plants, and understanding these adjustments in plant metabolism is important in enhancing plant performance under water-stressed conditions (Fiehn, 2002; Shulaev et al., 2008; Obata and Fernie, 2012). In the field context, soil drought and rainfall often occur alternately, implying that repeated drought-recovery cycles are considerably more common than a prolonged drought event. With the changing climate due to the buildup of greenhouse gasses in the atmosphere and diversion of irrigation water from agriculture to other purposes, an increase in the occurrence of episodic drought, in which plants will be repeatedly exposed to drought, may be likely (IPCC, 2007; Xu et al., 2010; Gallé et al., 2011; Sun Y. et al., 2013). Drought stress and recovery following re-watering may follow different adaptive mechanisms where plants are exposed to episodic drought (Gazanchian et al., 2007; Izanloo et al., 2008). The rate and degree of recovery to pre-stress levels varies among plant species and can exert a substantial impact on plant fitness (Loewenstein and Pallardy, 2002). Thus, in addition to drought resistance/tolerance, drought recovery ability is another important component in the plants ability to cope with predicted increases in the severity and frequency of episodic drought (Vankova et al., 2012). Numerous studies consider the effects of episodic drought on plant growth as one of the most fundamental processes in the life cycle of plants. Rapid and complete growth recovery following re-watering is likely the key to preventing significant declines in plant production after episodic drought (Chaves et al., 2009). Several studies showed that plant growth measured as leaf area and biomass was stimulated by re-watering following drought (Reynolds et al., 2004; Siopongco et al., 2006). Xu et al. (2009) found that new plant growth in perennial grass, Leymus chinensis, was enhanced by re-watering. It may overcompensate for the loss of the grass’s net primary production due to the effect of previous drought. In other studies it was found that plant growth characteristics recovered gradually after re-watering (Miyashita et al., 2005; Shi et al., 2014). By contrast, Yahdjian and Sala (2006) found that previous drought limitation to growth existed in the Patagonian steppe. This limitation, which is due to previous drought, may be portrayed as a plant memory behavior based on past drought stress (Xu et al., 2010). Notably, limitations to previous drought and recoveries by subsequent re-watering may influence biochemical metabolism and signal cascades (Fortunati et al., 2008; Xu et al., 2010). Toscano et al. (2014) recently demonstrated that Mediterranean ornamental shrubs could employ various mechanisms, such as the differential partitioning of dry matter between root and shoot parts, as well as the reduction of leaf area, to allow them to tolerate repeated cycles of drought. Furthermore, whether plant growth exhibits complete recovery after re-watering and the extent of recovery may depend on previous drought intensity or duration, plant species, variety, and number of consecutive drying cycles (Flexas et al., 2004; Miyashita et al., 2005; Xu et al., 2009; Shi et al., 2014). Thus, the effect of previous drought and re-watering on plant growth and plant function must be further clarified. Technological innovation over the past decade has made measuring changes in metabolite levels on metabolome-wide scales possible (Urano et al., 2010), enabling an unprecedented overview of the global metabolic changes occurring under drought stress and re-watering. Plant recoveries from drought stress include shifts in specific metabolic pathways and changes in metabolite levels (An et al., 2013). The degree of recovery from drought has been associated with various biochemical mechanisms, including maintenance of membrane stability, osmotic adjustment, phytohormone

63

accumulation, increased carbon partitioning and carbohydrate storage in plant organs, as well as accumulation of sugars and organic solutes (Xu et al., 2009; Gechev et al., 2013; Sicher et al., 2012; Sun C. et al., 2013; Foster et al., 2015). For example, in addition to regulating gene expression and sustaining growth, sugars have been implicated in stabilization of membrane structures and cell osmotic adjustment during drought stress and in providing carbon skeletons for recovery during the rehydration process (Suguiyama et al., 2014). Although physiologically and ecologically significant, the recovery period of the plant in response to drought has received considerably less attention than the response to the developing drought. Drought recovery mechanisms have been already described for shrubs and trees (Marron et al., 2003; Gallé et al., 2007; Echevarria-Zomeno et al., 2009; Gallé et al., 2011; An et al., 2013; Cao et al., 2014; Correia et al., 2014; Toscano et al., 2014), and grasses (Siopongco et al., 2006; Xu et al., 2009; Kang et al., 2011; Meyer et al., 2014; Suguiyama et al., 2014; Zhang et al., 2014; Foster et al., 2015), but evidence of the capacity for recovery from previous drought and the underlying metabolic processes remain largely unknown in crop plants. Especially, in most maize (Zea mays L.) growing regions, irrigation and rainfall patterns dictate episodic drought that can exist at any of the growth stages during the plant life cycle. Maize, which is one of the most important crops worldwide, is used as food, feed, and an energy source and can be planted in diverse ecosystems. Understanding how maize plants respond to episodic drought, as well as the underlying mechanism will indicate specific traits to be targeted in breeding/selection programs and is likely to help in implementing crop management practices during climatic changes. In this study, we attempt to provide an insight into the mechanism of maize plant responses to two successive drought and re-watering cycles in terms of plant growth and metabolome profiles. We also try to explore if various metabolic pathways in maize recover from the effects of previous drought to the same extent and if any differences exist between responses to repeated drought-recovery cycles in maize. 2. Materials and methods 2.1. Plant growth conditions and treatments In this experiment, we used two maize inbred lines: PH4CV (with the Lancaster background; subsequently abbreviated as line L) and PH6WC (with the Reid background; subsequently abbreviated as line R). These two lines that were developed for high yield and extensive adaptability are parents of popularly planted hybrids and are divided into two major germplasm groups, which are available for breeding and research programs in China. The seeds were sown in pots containing 4 kg of soil and then grown side-by-side in a greenhouse at the Experimental Station of Northeastern University, Shenyang, Liaoning (123◦ 4 E, 41◦ 8 N), at 30 ± 2 ◦ C/20 ± 2 ◦ C (day/night), with a 14 h light/10 h dark cycle at a photon flux density of 700 ␮mol m−2 s−1 and with approximately 65 ± 5% relative humidity. The soil in this experiment is a brunisolic soil having pH 5.72, organic matter 2.52 g kg−1 , total N 1.22 g kg−1 , total P (P2 O5 ) 1.12 g kg−1 , total K (K2 O) 24.24 g kg−1 . Fertilizer consisted of 180–75–75 kg ha−1 of N–P2 O5 –K2 O incorporated before planting. Water stress was minimized with timely irrigation during plant culture. After about one month of growth (the fourth leaf stage), the seedlings were randomly assigned to either the control group or subjected to drought and re-watering (D-W) treatments (Fig. 1). For the experiment described in this study, 40 plants from each line were subjected to two consecutive cycles of D-W, whereas another 40 plants from each line received water every day (control). For each treatment cycle, the degree of soil drying was controlled by measuring the pot weight for 7 days, after which the plants were

64

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

Fig. 1. Timeline for drought and re-watering treatments used in the experiment. Maize plants were irrigated daily (control) or subjected to repeated drought and re-watering treatments (the first 7 days drought stress and followed by 7 days rewatering, the second 7 days drought stress and followed by 7 days re-watering).

plants at the levels of growth and metabolome profiles. Relative water content (RWC) and electrolyte leakage (EL) of maize leaves were measured to monitor the level of drought stress, respectively (see Supplementary Table S1 in the online version at DOI: 10.1016/ j.agwat.2016.04.016). Both monitors were almost at similar level as literatures, in which moderate level of drought stress was applied (Carvalho et al., 2011; Bai et al., 2006). On days 7 (D1), 14 (W1), 21 (D2) and 28 (W2) of the water-withholding treatment period, the plants were sampled. Sample plants of the control group were collected at the beginning (T0; only for growth measurements), and on day 7 (C1) and day 21 (C2) after treatments. Growth measurements on one set of plants and metabolite analysis on another set of plants were performed simultaneously. After harvesting, all samples for metabolic analysis were immediately frozen in liquid nitrogen and stored at −80 ◦ C.

2.2. Growth characteristics analysis re-watered to field moisture capacity for another 7 days. Totally, this procedure of two D-W cycles lasted 28 days. A moderateintensity drought stress (60% field moisture capacity) was selected because of its induction of significant stress responses in maize

The total leaf area per plant (LA) was measured with an automatic area meter (LI-3000, Li-Cor Inc., Nebraska, USA) at each sampling point. All harvested plants were divided into shoots and

Fig. 2. Effects of drought and recovery cycles on the leaf area (A), biomass (B), biomass allocation (C) of maize plants. Plants were irrigated daily in the controls (C1 and C2) or subjected to drought and re-watering treatments [Start trial (T0), 7 days drought stress in the first cycle (D1), 7 days re-watering in the first cycle (W1), 7 days drought stress in the second cycle (D2), 7 days re-watering in the second cycle (W2)]. The asterisks depict statistically significant differences between control and stressed plants at each event (T0, D1, W1, D2, and W2) for each line.

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

65

Fig. 3. Principal component analysis based on polar metabolite profiles of two maize inbred lines in response to two drought and re-watering cycles (A, B, D, and E) and metabolic average trajectories for cyclic drought based on principal components (C, F). The coding of the lines and treatments is the same as that in Fig. 2.

roots. The biomass (BM) samples were dried (80 ◦ C for 48 h) to a constant weight. The biomass allocation (BA) was also calculated with the following traits: BAS (ratio of shoot dry mass to total dry mass) and BAR (ratio of root dry mass to total dry mass).

2.3. NMR and metabolites The sample preparation and NMR detection used the method of Sun et al. (2015). Five hundred milligrams of plant sample were ground in liquid nitrogen. Then, 2 mL of a pre-cooled water-methanol mixture (1:1) and 2 mL of chloroform were added to the tube, vortexed for 30 s, and sonicated in an ice bath for 1 min. The sample was then centrifuged at 4 ◦ C for 20 min. This procedure was performed three times; the polar and non-polar fractions were combined and collected separately. For polar samples, methanol was removed under a vacuum; then, the supernatants were frozen at -80 ◦ C and lyophilized at least 24 h. Non-polar samples were dried under reduced pressure in a rotary vacuum evaporator. Afterwards, 800 ␮L of 100% D2 O and 160 ␮L of phosphate-buffered saline (pH 7) containing 10%

D2 O and 0.02 mM sodium 3-trimethylsilyl [2,2,3,3-D4] propionate (TSP) were added to the dried polar fractions; afterward, 1 mL of chloroform-D containing 0.03% tetramethylsilyl (TMS) was added to the dried non-polar fractions. TSP and TMS were used as internal standards separately. All of the contents were transferred to the tubes and then centrifuged at 10,000 rpm for 5 min. For each sample, 0.65 mL of supernatant was transferred to 5 mm NMR sample tubes. Fifteen biological replicates of each treatment were used for NMR analysis. The samples were scanned by high-resolution 1D 1 H NMR spectroscopy (1 H frequency, 600.13 MHz) generating polar and nonpolar metabolic profiles using a Bruker Avance 600 spectrometer (Bruker Biospin, Germany). Sample handling, automation, and acquisition were controlled using TopSpin 2.1 software (Bruker Biospin). A standard 1 H 90◦ pulse sequence was used for both kinds of samples, and residual water resonance was suppressed in the polar samples. Each spectrum was obtained as 32 k data points at a spectral width of 16 ppm and as the sum of 128 transients with a relaxation delay of 2 s. The spectra of polar samples were phase- and baseline-corrected automatically and referenced manually to the

66

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

TSP residual resonance at ␦H 0.00 ppm. For the non-polar samples, the frequency-domain spectra were phase- and baseline-corrected automatically and referenced manually to the TMS residual resonance at ␦H 0.00 ppm. Metabolite resonances were assigned based on publicly available databases (e.g., Biological Magnetic Resonance Data Bank, Spectral Database for Organic Compounds, and Madison Metabolomics Consortium Database) and previous studies (Fan, 1996; Tate et al., 2001). These assignments were further confirmed by extensive 2D NMR data (see Supplementary Table S2 in the online version at DOI: 10.1016/j.agwat.2016.04.016). 2.4. Data and statistical analysis Two separate data and statistical analyses were performed for growth traits and metabolites. For growth traits, analysis of variance (ANOVA) was performed on the effect of D-W treatments on each trait variation using SPSS 13.0 software. Significant differences among the treatment means were analyzed using the Student–Newman–Keuls post hoc tests. The effects were considered significant if P < 0.05. For the metabolome, principal component analyses (PCA) was performed on the polar metabolic profiles using the SIMCA-P+ (version 11.5, Umetrics, Umea, Sweden) software package to assess the effect of D-W treatments on metabolic profiles. Before PCA analysis, spectral intensities were scaled to TSP for the polar extract and to TMS for the non-polar extract. For the polar phase, the spectral intensities were reduced to integrated regions with an equal width (0.04 ppm) consistent with the region of ␦ 9.00 to ␦ -0.04. Because of the residual signals of water and methanol, the spectral regions of ␦ 5.00–4.70 and ␦ 3.38–3.30 were then excluded from analysis. The PCA results were visualized with score plots, in which each point represents the metabolome of one sample. The metabolite levels were calculated relatively from the least overlapping NMR signals of metabolites with known concentrations of TSP and TMS under the assumption of the minimal intersample variations in the spin–lattice relaxation time of the same protons. Differences in the metabolite levels between each treatment and control were considered significant at P < 0.05. Heat map and hierarchical cluster analyses were performed with MeV version 4.2 on semi-quantitative data for polar and non-polar metabolites. In order to partition the metabolites into discrete groups of event course patterns, we used a clustering approach. K-means clustering (MeV version 4.2) on the mean of the replicates (data mean centred and reduced to unit variance and with distance based on correlation) grouped all significant metabolites showing common trends. Metabolite relationships were visualized and studied by using correlation networks (Shannon et al., 2003; Gehlenborg et al., 2010). Correlation analyses between all metabolite pairs were first performed by using Pearson’s correlation on each of the data profiles obtained from the two lines during the two D-W cycles. To construct a network, we determined the threshold values for Pearson’s correlation coefficient that would ensure a q-value of 0.01. All network visualizations were generated, and network properties were computed by Cytoscape 2.8.3 (http://www.cytoscape.org/). 3. Results 3.1. Growth responses to drought and recovery cycles Under well-watered conditions, lines L and R differed in their patterns of biomass accumulation and leaf area expansion (Fig. 2a and b). Line L exhibited higher total leaf area increases and biomass increment rates than line R. The first drought resulted in decreases in LA and BM, and re-watering was followed by recovery of all traits to near control levels, indicating complete reversibility of responses

Fig. 4. Heat map and cluster tree of metabolic profile in two maize lines in response to two drought and re-watering cycles. The metabolic profile for all of the samples represents an average of the replicate measurements of each stress treatment. Images are made with MeV 4.2. The coding of lines and treatments are the same as those in Fig. 2.

to the first drought cycle. However, with the progression of two D-W cycles, the second drought caused a partial reversible decrease in BM in both lines. A stimulated growth was more visible for line R than for line L and was preceded by a return of LA for line R to control values during the second re-watering only. Increased root-to-shoot ratios were observed in plants of lines L and R under drought conditions (Fig. 2c), which were confirmed by decreased leaf area and shoot biomass. 3.2. Metabolic responses to drought and recovery cycles A PCA based on the polar metabolite profile showed different patterns of variations during two successive D-W cycles (Fig. 3). For both lines, the changes can be clearly observed after 7 days of drought. The changes start to be restored and reach highly similar metabolite profiles as the control after 7 days of re-watering during the first cycle, implying that the plants may possess a high metabolic repairing capacity upon re-watering after previous drought stress (Fig. 3a and d). However, metabolic changes in the second cycle appear to be more pronounced than those observed in the first cycle, in which the metabolite profile after re-watering does not reach the normal level seen in well-watered plants (Fig. 3b and e). PCA was also applied to indicate event-point variations. In general, the first two principal components resolved the event series of different responses (Fig. 3c and f). The metabolome, during the first drought event, starts to show evident separation from the control, and is almost fully restored to the control level after the first re-watering event. It then becomes clearly distinct from the three previous profiles (C1, D1, and W1) again during the second drought event and even after the second re-watering event. PCA fails to resolve the control samples between two cycles. Moreover, the control samples from well-watered conditions show a similar metabolic pattern to the re-watered samples in the first cycle. Changes in line R were more pronounced than those observed in line L, particularly during the second cycle. The heat map in Fig. 4 provides a global perspective regarding the changes in polar and non-polar metabolite profiles in response to cyclic D-W treatments. In general, a distinct treatment-related cluster was detected after plants were exposed to successive D-W cycles. Moreover, the metabolic profiles produced for controls and re-watered samples after drought stress in the first cycle were clustered in one group. The metabolome of maize plants exposed to two

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

67

cycles of drought stress showed a similar response and possessed spatially close clusters. However, the metabolic state of maize line R after the second re-watering was similar to the metabolic state after exposure to drought stress. This finding is consistent with the former result based on PCA (Fig. 3). Further analysis by ANOVA investigated the differences in concentrations of metabolites between treatments (Table 1, P < 0.05). Metabolites including sucrose, glucose, fructose, succinate, inositol, choline, proline, isoleucine, leucine, and valine increased significantly in drought affected plants but decreased in recovering plants during the first treatment cycle. By contrast, shikimate, fumarate, tryptophan and tyrosine showed the opposite pattern; decreasing significantly in drought treatment but almost fully restoring to control levels within seven days in the first cycle and then remaining essentially unchanged as the second D-W cycle progressed. Alanine, fatty acids, polyunsaturated fatty acids, and other non-polar metabolites were almost stable under drought stress but changed significantly during re-watering in the first cycle. Several amino acids in line L, such as alanine, ␥-amino-butyrate, glutamate, and glutamine, over-corrected following rehydration at the end of the experiment. The most evident example of this over-correction was ␥-amino-butyrate. More variations in metabolites during the last cycle were found in stressed leaves than those detected during the first cycle. Furthermore, malate, asparagine, adenine, formate, diacylglyceride, and triacylglyceride remained almost unaltered over D-W treatments. Differences between line L and R were also observed in the response ratios of metabolites. During the second recovery phase, more metabolites returned to their non-stress levels in line L than in R, indicating a faster rate of recovery in L. To visualize the changes in metabolites during two successive D-W cycles, we performed K-means clustering and correlation-based network analysis. K-means clustering indicated that most of known metabolites could be categorized into five expression clusters and showed responses in an event-dependent pattern (Fig. 5). Clusters A and B showed major “accumulation–recovery–accumulation–recovery” changes. Cluster C showed major “reduction–recovery–reduction–recovery” changes. Cluster D showed “precedent accumulation” changes. By contrast, cluster E showed “succedent accumulation” changes. Meanwhile, differences between lines L and R were also observed in the clusters. For example, sucrose was assigned into cluster B for line L, whereas the sugar was assigned into cluster D for line R. In line L, 9 metabolites in cluster C were decreased by water stress and then showed a reversal change, whereas 7 metabolites in line R changed in the same manner. Correlation-based network analysis showed that during the second D-W cycle, the number of edges in line L decreased from 181 to 167, network density decreased from 0.19 to 0.17, and average node degree decreased from 8.2 to 7.6 (Fig. 6a and b). In line R, the number of edges decreased from 287 to 259, network density decreased from 0.30 to 0.29, and average node degree decreased from 13.0 to 12.0 during the second D-W cycle (Fig. 6c and d). Overall, the networks of line R were characterized by greater numbers of edges compared with networks of line L.

avoidance mechanism that allowed the minimization of water loss ˜ through stomatal closure (Mahajan and Tuteja, 2005; Banon et al., 2006; Toscano et al., 2014). Plants may also try to maximize water absorption in soil by enhancing the root system under soil drought. Increased dry matter allocation to roots at the expense of both leaves and stems under drought conditions is thought to improve plant water balance and enhance the likelihood of survival (Xu et al., 2010; Sun C. et al., 2013; Toscano et al., 2014). Our reports also confirmed that biomass allocation was remarkably influenced by the water change factor in which the biomass ratio of root and shoot increased under water stress (Fig. 2c). Tolerance to water deficit as well as growth recovery from water deficit by re-watering are two important and inseparable processes for plant survival. The current experiment indicated that LA and BM in both lines were stimulated dramatically after the first and second re-watering compared with their pre-drought samples (Fig. 2a and b). The stimulation of plant growth by re-watering after drought has also been highlighted by numerous experimental investigations (Siopongco et al., 2006; Izanloo et al., 2008; Efeo˘glu et al., 2009; Xu et al., 2010; Shi et al., 2014; Okami et al., 2015). Rapid leaf area growth during the recovery period is essential for maximizing radiation interception, biomass accumulation, and grain yield and is favored by easily available useful nutritional factors in rewetted soil (Heckathorn and Delucia, 1994; Xu and Zhou, 2006). However, when compared with the well-watered control, the BM and LA in line L were remarkably limited, especially after the second re-watering (Fig. 2a and b). The limitation of growth recovery could also exist in soybean and in certain grasses (Yahdjian and Sala, 2006; Xu et al., 2009; Lobato et al., 2008; Husen et al., 2014). The limitation to growth because of previous drought may be portrayed as a plant memory behavior based on past drought stress (Wiegand et al., 2004; Xu et al., 2010). Moreover, in our work, the biomass increment rate and leaf area expansion rate was not limited for line R when watered again after previous drought, suggesting that the pre-stress’ memory may perform the central function in the growth of new parts rather than the final production, which may derive from the re-triggering of stored resources, such as soil nutrition, besides the water reapplication (Xu et al., 2009, 2010). Our results also showed that the growth response to successive drought cycles between the two lines varied (Fig. 2). The two successive D-W cycles led to a greater reduction in total leaf area in line L than in line R, indicating that the growth advantage seen in L under control conditions appeared to be counterbalanced by its high susceptibility to water stress. Compared with the first drought, the second drought cycle exerted a slight effect on plant growth in terms of LA and BM. However, the second re-watering caused greater stimulation of the growth increment rate than that in the first re-watering (Fig. 2). This work highlighted that recovery extent or previous drought limitation because of re-watering may not only depend on the previous drought duration or intensity but also strongly depend on the lines and cycles, during which a partial/full/compensatory recovery may appear. The underlying mechanism responsible for the difference in these growth responses to episodic drought needs to be elucidated further.

4. Discussion

4.2. Metabolic profiling responses

4.1. Growth responses

Metabolite profiling offers the potential to provide not only deeper insight into complex regulatory processes but also to identify chemical signatures for specific phenotypes (Fiehn, 2002). The PCA results demonstrated that metabolomic changes occurred in maize plants during two D-W cycles in a cycle-dependent manner (Fig. 3). The cycle-dependent responses indicated the progressive development of the maize metabolic response to cyclic drought stresses. Plant metabolic response to stress involves a series of characteristic time-dependent biochemical events caused by the

Leaf growth is the most sensitive process among numerous morphological adaptations to water stress that involve the aboveground portions of plants (Chaves et al., 2003). Drought-induced reductions in both cell division and cell enlargement are responsible for the decrease in leaf area (Marron et al., 2003). In our study, two maize lines showed a reduced leaf area under cyclic drought conditions (Fig. 2a). This response is also attributed to an

68

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

Table 1 Response ratios of metabolites in two maize lines exposed to two drought and re-watering cycles. Metabolites

Suc Gluc Fru Pyr Lac Ino SA For HB Succ Fum MA Ile Leu Val Phe Trp Tyr Glu Gln His Pro GAGB Asp Asn Ala Thr Ch Ade Up Adl PP TAG DAG FA FFA UFA PUFA LFA Fal Unp1 Unp2 Ter SS

Line L

Line R

D1/C1

W1/D1

W1/C1

D2/C2

W2/D2

W2/C2

D1/C1

W1/D1

W1/C1

2.32 2.68 2.32 0.67 0.74 2.40 0.99 0.67 1.01 1.97 0.60 1.09 1.49 2.02 1.27 0.91 0.89 0.95 1.23 1.05 0.90 2.10 1.06 1.37 1.35 0.98 0.73 1.76 1.24 1.24 1.06 1.21 1.15 0.99 0.97 1.13 1.18 1.12 1.01 1.07 1.10 1.12 0.99 1.17

0.35 0.31 0.38 1.15 1.13 0.33 1.01 3.49 0.77 0.38 1.53 0.77 0.69 0.45 0.80 0.95 1.08 1.06 0.93 0.97 1.29 0.40 1.25 0.73 0.64 1.36 1.14 0.50 0.55 0.49 1.45 1.30 1.35 1.62 1.49 1.32 1.34 1.44 1.40 1.39 1.33 1.34 1.07 1.70

0.82 0.82 0.88 0.77 0.84 0.78 1.00 2.35 0.78 0.75 0.93 0.84 1.02 0.91 1.02 0.86 0.96 1.00 1.15 1.02 1.16 0.84 1.33 1.00 0.86 1.34 0.83 0.88 0.68 0.60 1.54 1.58 1.55 1.60 1.45 1.49 1.57 1.61 1.42 1.49 1.45 1.50 1.06 1.99

4.44 2.37 1.78 0.96 0.87 2.14 0.90 1.08 2.42 0.36 0.91 0.94 1.15 0.94 1.21 1.07 1.04 1.10 1.11 1.10 1.05 1.76 1.26 1.17 1.20 1.05 0.94 1.42 1.30 0.46 0.84 1.30 0.80 1.16 0.93 0.85 0.80 0.85 0.91 0.84 1.06 0.89 0.70 1.00

0.65 0.68 0.78 1.57 1.04 0.64 1.29 1.01 0.64 2.95 0.88 1.26 0.80 0.89 0.77 1.03 1.17 1.09 1.34 1.15 0.87 0.65 1.19 0.82 1.29 1.32 0.94 0.82 0.90 2.75 0.78 0.67 1.11 0.80 1.11 1.00 1.13 0.89 1.11 1.14 0.79 1.17 1.28 1.15

2.87 1.61 1.40 1.51 0.90 1.37 1.16 1.09 1.56 1.05 0.80 1.18 0.92 0.84 0.93 1.10 1.22 1.20 1.48 1.26 0.91 1.14 1.50 0.96 1.55 1.39 0.88 1.17 1.17 1.26 0.65 0.87 0.88 0.94 1.03 0.86 0.91 0.75 1.01 0.96 0.84 1.04 0.90 1.15

1.24 1.19 1.42 0.59 1.26 1.15 0.65 0.39 0.96 2.37 0.63 1.10 2.20 3.63 2.43 1.02 0.73 0.76 0.88 0.97 0.91 1.13 1.39 1.38 1.20 1.31 1.04 2.08 0.94 1.38 1.22 1.20 1.28 0.95 1.05 1.15 1.33 1.18 1.07 1.12 1.21 1.07 1.04 1.20

0.72 0.88 0.69 1.77 0.91 0.93 1.51 2.70 0.76 0.41 1.78 1.02 0.53 0.31 0.48 1.06 1.50 1.45 1.22 1.21 1.39 0.88 0.93 0.82 0.87 1.25 1.17 0.65 1.26 0.77 1.29 1.03 1.11 1.07 1.05 1.05 1.16 1.03 1.03 1.11 1.36 1.04 1.05 0.89

0.90 1.05 0.98 1.05 1.15 1.07 0.99 1.05 0.73 0.97 1.13 1.12 1.16 1.11 1.17 1.08 1.10 1.10 1.07 1.17 1.25 1.00 1.29 1.13 1.05 1.63 1.21 1.35 1.18 1.06 1.57 1.24 1.41 1.01 1.09 1.22 1.54 1.21 1.10 1.24 1.65 1.12 1.08 1.07

D2/C2 2.31 1.71 1.90 0.77 1.19 1.69 0.86 3.89 1.85 1.08 0.91 0.90 1.78 2.10 2.07 1.21 0.92 1.08 1.06 1.25 1.11 1.65 1.39 0.87 1.06 2.00 0.86 1.48 1.21 0.55 0.79 0.87 0.93 1.23 1.18 0.75 0.98 0.84 1.18 1.01 0.94 1.16 0.92 1.14

W2/D2 1.15 1.35 0.64 1.74 0.70 1.09 1.28 1.12 1.16 1.31 0.92 1.16 0.48 0.41 0.44 0.74 0.99 0.88 1.31 0.91 0.88 0.62 1.02 1.31 1.11 0.66 0.77 0.71 0.73 1.45 0.92 0.89 0.91 0.68 0.68 0.99 0.87 0.91 0.68 0.82 0.84 0.66 0.72 0.69

W2/C2 2.67 2.31 1.23 1.34 0.84 1.85 1.11 4.34 2.15 1.41 0.83 1.04 0.85 0.86 0.91 0.90 0.91 0.95 1.39 1.13 0.97 1.02 1.41 1.15 1.18 1.32 0.66 1.06 0.89 0.80 0.72 0.77 0.84 0.84 0.80 0.74 0.85 0.77 0.80 0.82 0.79 0.76 0.66 0.79

Metabolites with a significant change between treatments within each cycle are indicated in bold (P < 0.05). Mean of response ratios are shown (n = 15). Metabolites variables are peak areas with respect to TSP or TMS. The coding of lines and treatments are the same as those in Fig. 2. Metabolites in legend: Suc, sucrose; Gluc, glucose; Fru, fructose; Pyr, pyruvate; Lac, lactate; Ino, inositol; SA, shikimate; For, formate; HB, ␤-hydroxybutyrate; Succ, succinate; Fum, fumarate; MA, malate; Ile, isoleucine; Leu, leucine; Val, valine; Phe, phenylalanine; Trp, tryptophan; Tyr, tyrosine; Glu, glutamate; Gln, glutamine; His, histidine; Pro, proline; GABA, ␥-amino-butyrate; Asp, aspartate; Asn, asparagine; Ala, alanine; Thr, threonine; Ch, choline; Ade, adenine; Up, unknown polar compound; Ald, aldehyde group; PP, polyphenol region; TAG, triacylglyceride; DAG, diacylglyceride; FA, fatty acid; FFA, free fatty acids; UFA, unsaturated fatty acid; PUFA, polyunsaturated fatty acids; LFA, linoleyl fatty acid; Fal, fatty alcohols; Unp1, unknown non-polar compound 1; Unp2, unknown non-polar compound 2; Ter, terpene; SS, sterols.

dynamic progression of the induced physiological processes (Dutta et al., 2009). The time of recovery appears to be necessary for metabolism regulation and may be associated with the repair of possible damage under drought stress, as suggested by other studies (Oliver et al., 1998; Suguiyama et al., 2014). Our findings are consistent with those of previous studies that showed duration dependence in the metabolic response to abiotic stresses (Dutta et al., 2009; Gavaghan et al., 2011). Interestingly, metabolism after rehydration in line L returns to its basal levels more rapidly than line R, which can be observed in the PCA (Fig. 3c and f). These results were confirmed by heat mapping (Fig. 4).

4.3. Metabolite responses Approximately 82% of the metabolites measured in this study were affected by cyclic water stresses (Table 1). In agreement with earlier results, most of the metabolic changes attributed to water

stress in this study were reversed by re-watering. However, the rate and extent to which individual metabolites recovered to control levels differed. The levels of compounds such as shikimate, succinate, fumarate, isoleucine, leucine, valine, tryptophan, tyrosine, proline, and choline were fully reversed after re-watering, but sterols, aldehyde group, terpenes, fatty acids, free fatty acids, and polyunsaturated fatty acids were only partially restored and remain below control levels, particularly after the second re-watering. In addition, several compounds, such as sucrose, glucose, fructose, and inositol, were over-corrected following re-watering. This finding suggested that different metabolic pathways in maize plants returned to normal status at different rates. The current study showed that the recovery of metabolic processes initiated by re-watering during cyclic drought was considerably more complicated than previously believed. Metabolites with the same metabolic profiling patterns are likely to be functionally related to each other (Kim et al., 2007). For example, clusters A and B

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

69

Fig. 5. Typical K-means clusters of polar and non-polar metabolites of two maize lines in response to two drought and re-watering cycles. Dark line in each line graph indicates the mean of all metabolites in each cluster (data of each metabolite mean centred and reduced to unit variance). Metabolites in each cluster are reported. The coding of lines and treatments are the same as those in Fig. 2. Abbreviations of metabolites were indicated in Table 1.

showed a rapid increase in concentrations of fructose, choline, proline, isoleucine, leucine, and valine during drought stress and a rapid recovery after re-watering, representing a fast response to water availability and excellent dynamics in osmotic adjustment and amino acid metabolism in two maize lines (Fig. 5). For fumarate, shikimate, tyrosine, and tryptophan in cluster C, drought-induced decline invariably recovered after re-watering and changed in an intensive manner during each drought cycle in the two lines. Most remarkably, the sucrose, glucose, and fructose levels of two lines seemed to be strictly co-regulated during drought and recovery, as indicated by the increases in these parameters

during drought treatments and the full recovery after W1 and over-correction after W2 (Table 1). Sugars perform critical functions in plant growth and adaptation to abiotic stress (Gibson, 2005; Hasanuzzaman et al., 2013). The accumulation of sugars in plants under drought stress is a common response (Rizhsky et al., 2004; Mahajan and Tuteja, 2005; Wingler and Roitsch, 2008; Chaves et al., 2009). Under drought stress conditions, plants may catabolize cellular C reserves to avoid short-term C limitations and provide an energy source, thus preserving cellular function. Alternatively, the accumulation of sugars may reflect leaf osmotic adjustment as numerous sugars act as osmoprotectants to help

70

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

Fig. 6. Correlation-based network analysis of two maize lines in response to two drought and re-watering cycles. Nodes correspond to polar (circles) and non-polar (squares) metabolites; node colors consistent with metabolite classes as detailed in the figure legend. Nodes are ordered into modules that correspond to their metabolite classes. Abbreviations of metabolites are indicated in Table 1. (A) and (B) are the networks of line L in the first and second drought and recovery cycles, separately. (C) and (D) are the networks of line R in the first and second drought and recovery cycles, separately.

stabilize macromolecules and maintain osmotic balance in drought stress responses (Chaves et al., 2003; Reddy et al., 2004; Meyer et al., 2014). Sicher et al. (2012) confirmed that the sugars functioned as osmoprotectants and reduced damage to vital growing regions of barley plants during episodes of stress through accumulation in the growing tip and adjacent expansion zone of roots. The decrease in sugars after re-watering probably indicated that carbohydrates were used as an immediate energy source in the energy-dependent metabolic pathways responsible for repairing the damage caused by water deficiency, as suggested by Suguiyama et al. (2014) for Barbacenia purpurea. Foster et al. (2015) also reported that soluble sugars may act as an important carbon source used for regrowth in B. bituminosa during recovery after subsequent re-watering. We can also reasonably speculate that maize growth recovery likely adopts an emergency repair mechanism to protect against damage to molecular functions under recovery (Fig. 5, Table 1). However, despite significant recovery, concentrations of sucrose, glucose, and fructose remain higher than those of the corresponding control at the second recovery phase, implicating that a more precise defense strategy was used by maize plants subjected to D-W cycles. Sugars are likely to protect plant cells against oxidative stress attributed to the activation of specific reactive oxygen species scavenging systems with a consequent reduction in oxidative damage, as confirmed by the interplay of sugars with redox and hormone signals (Rolland et al., 2006; Ramel et al., 2009). Furthermore, sugars are important signaling molecules and may be the key players in the

integration of cellular responses to internal and environmental alterations at a whole plant level (Alla et al., 2012). Therefore, the high accumulation of sugars at the second event of re-watering must be considered an important mechanism for rapid recovery from the cyclic drought. Amino acids are also known recovery-responsive metabolites, indicating that maize growth recovery has a large demand for building block metabolites (Sun C. et al., 2013). In this study, the levels of isoleucine, leucine, valine, and proline that were significantly accumulated during drought, also decreased after re-watering (Table 1) indicating that accumulated nitrogen may be used as a substrate for protein replacement as well as indicating a latent demand for newly acquired organic nitrogen in plants following re-watering (Shelp et al., 1999; Sicher et al., 2012). However, the aromatic amino acids tyrosine and tryptophan decreased during D1 stress and increased to control levels after W1, indicating that nitrogen input to secondary metabolism is prioritized at well-watered conditions. Moreover, the “reduction–recovery” change in these aromatic amino acids was dependent on the “reduction–recovery” change in shikimate during two successive D-W cycles (Fig. 5). In plants, aromatic amino acids act as precursors for the production of several important compounds, such as phytohormones, electron carriers, enzyme cofactors, and antioxidants (Dixon, 2001). Our results emphasized the importance of aromatic amino acid metabolism in plant response to episodic drought. Although the glutamate family amino acids, namely, glutamate, glutamine, ␥-amino-butyrate, and

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

histidine, were not affected significantly by drought, these metabolites increased above control levels after re-watering. Apparently, maize plants could use different nitrogen metabolism strategies to cope with water pulses. Similarly, Gechev et al. (2013) also observed a decrease in amino acids of the aspartate family, with a succedent increase to control levels after rehydration in Haberlea rhodopensis. Glutamine possibly functions as a N status sensor in maize and participates in a signal transduction pathway involving glutamine synthetase (Seebauer et al., 2004). However, this increase in levels of glutamate, glutamine, ␥-amino-butyrate, and histidine in maize exclusively after re-watering may not be an indicator of general stress and cell damage. Instead, this increase may be an adaptive response to pre-drought stress (Ansari et al., 2005; Ashraf and Foolad, 2007; Fait et al., 2008). Osmotic adjustment is considered to be one of the critical processes in plant acclimation to drought stress (Pang et al., 2011; Sun et al., 2015). Our results also suggest the important function of osmotic adjustment in successive D-W cycles, at least in maize leaves. Osmotic adjustment in plants likely assists in maintaining plant metabolism and growth at extremely low water potential by maintaining turgor pressure in the plant cells. Accumulation of osmolytes during osmotic adjustment that is reduced during recovery could be associated with the synthesis of new biomass and could enable regrowth (Morgan, 1984; Warren et al., 2012; Foster et al., 2015). Proline has been widely recognized as a drought-inducible proteinogenic amino acid with an osmoprotective function. Sharp et al. (2004) proposed that proline, together with hexoses was responsible for half of the osmotic adjustment in maize root tips in response to water deficiency. In certain plants, proline synthesis is activated, and its degradation is repressed during water stress, whereas re-watering triggers the opposite regulation (Szabados and Savouré, 2010; An et al., 2013). We also found that maize plant proline accumulates to higher levels during drought and recovers when re-watered (Table 1). The response in maize appears to be a “standard model” of plant proline metabolism during D-W. Except for its known function as an osmotic agent, proline accumulation can also promote plant damage repair ability by upregulating antioxidant activity during recovery from severe water stress (Szabados and Savouré, 2010; Sharma et al., 2011; Zhang et al., 2014). In addition to traditional functions in plant osmotic adjustment and reactive oxygen species scavenging, proline can be rapidly utilized as a source of carbon and nitrogen (Van Heerden and Krüger, 2002; Szabados and Savouré, 2010). On the other hand, the active catabolism of proline promotes respiration pathways and restoration of chloroplast function to protect the photosynthetic system from permanent damage (Sharma et al., 2011; Foster et al., 2015). In addition, proline accumulation can act as a signaling molecule to modulate mitochondrial functions and influence plant development, flowering, and reproduction (An et al., 2013). Inositol is an osmolyte that functions in osmotic adjustment under stress condition (Valluru and Van den Ende, 2011). Drought stress increased inositol to extremely high levels in maize leaves during two successive D-W cycles. Inositol concentration substantially decreased following re-watering but remained higher than the corresponding control values, particularly in line R, thereby reflecting a slower turnover that likely contributed to metabolism for the growth of new shoots (Table 1). Similarly, Vankova et al. (2012) observed the persistence of elevated concentrations of osmoprotectant in tobacco during an early stage of recovery from drought. Inositol is a versatile cellular compound, being a precursor of several other compounds, such as phosphatidylinositol, myoinositol polyphosphate, and several compatible solutes, such as galactinol, pinitol, raffinose-family oligosaccharides, and cell-wall polysaccharides (Valluru and Van den Ende, 2011). Thus, compounds derived from inositol not only function in osmotic adjustment but also in phosphate storage,

71

raffinose synthesis, signaling, and lipid synthesis (Sicher et al., 2012; Zhang et al., 2014). Furthermore, drought-dependent inositol increase in maize leaves was also related to the accumulation of choline (Table 1). Choline is another compound that provides osmotic protection under stress conditions (McNeil et al., 2001). The accumulation of choline during stress treatments suggests that flux toward choline biosynthesis is enhanced. Choline accumulation may also occur when choline is released from membrane phospholipids by phospholipase D (Mattoo et al., 2006). Those osmolytes showing precedent accumulation and succedent reduction were coincident to the time point during two D-W cycles when other metabolites, such as sucrose, glucose, and fructose, started to decline in amount after drought-dependent accumulation. Our results also suggest an important function of osmolytes in plant recovery from cyclic drought in maize. 4.4. Cyclic differences in metabolic responses At a qualitative level, metabolic responses to D-W treatments were similar during the two cycles, indicating that maize plants employ the same strategies to cope with cyclic drought. However, quantitative differences in responses at metabolic level possibly reflect corresponding differences in water-stress tolerance between the first and second cycles (Figs. 3 and 4). Maize plants showed a more dynamic response (specifically in numbers of response metabolites) during the second cycle compared with the first cycle (Table 1). Evidently, maize plants require coordinated and extensive metabolic changes in the wake of their exposure to stress for a second time. Correlation-based network analysis indicated that there are extensive topological network differences between the two cycles (Fig. 6). The correlation-based network of maize plants exposed to D-W treatments revealed a remarkable decrease in the coordinated metabolic activities during the second cycle in contrast to the first cycle, the former of which could withstand the D-W treatment with less metabolic changes. The decreased network density and connectedness shown during the second cycle, which is suggestive of loose regulation imposed on metabolism, agrees with the hypothesis that stress lowers the number of relations and subsequently exerts a negative effect on network stability (Szalay et al., 2007). By contrast, Sanchez et al. (2011) reported that the correlation coefficient between metabolites in Lotus genotypes increased when the plants were exposed to salt stress, which may result in higher network connectivity. Recently, an increase in metabolic network connectivity in response to drought has been observed in grapevine by Hochberg et al. (2013). They proposed that regulatory mechanisms under drought conditions induce a concerted change in metabolism that allows the cell to adapt to the new conditions leading to more dependent metabolic profiles. Episodic drought occurs frequently in many regions of the world, particularly in arid and semi-arid areas. Maize is a well-adapted plant that can be planted in a wide range of climate conditions from tropical to temperate regions. We found different responses to a cyclic drought stress regime between two maize lines in plant growth and metabolic activity, which indicated that maize plants may use different water use strategies to cope with fluctuation in water status. Maize plants with stronger drought resistance and quicker recovery ability from cyclic drought stress will be of great significance to the optimisation of crop cultivation in an increasingly variable environment. 5. Conclusions Maize plants could employ growth, biochemical, and metabolic mechanisms to adapt to cyclic water stress and recovery by re-watering. The recovery extent or growth limitation caused by previous drought strongly depended on the drought cycles and

72

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73

a full or partial recovery of growth may occur. The PCA results clearly demonstrated that metabolomic changes occurred in maize plants during two successive D-W cycles in a cycle-dependent manner. However, the magnitude at which individual metabolites recovered to control levels differed. Levels of compounds such as shikimate, succinate, fumarate, isoleucine, leucine, valine, tryptophan, tyrosine, proline, and choline were fully reversed after re-watering but several non-polar metabolites were only partially restored and remain below control levels after the second re-watering. In addition, several compounds, such as sucrose, glucose, fructose, and inositol, over-corrected following re-watering. This finding suggested that the metabolic pathways with different relative importance, including energy production, osmotic protection, and signaling regulation, in maize plants returned to normal status at different rates during recovery. However, quantitative differences between the two cycles revealed by metabolic analysis suggested the complexity of metabolic processes initiated by water cycle changes. Further research is needed to fully understand the response to cyclic drought at the physiological and molecular levels. Acknowledgments This study was financially supported by National Natural Science Foundation of China (Nos. 31300331 and 41301325), P.R. of China. We also thank Dr. Wendy Harwood for her linguistic modification of this paper. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.agwat.2016.04. 016. References Ahuja, I., de Vos, R.C.H., Bones, A.M., Hall, R.D., 2010. Plant molecular stress responses face climate change. Trends Plant Sci. 15, 664–674. Alla, M.M.N., Khedr, A.H.A., Serag, M.M., Abu-Alnaga, A.Z., Nada, R.M., 2012. Regulation of metabolomics in Atriplex halimus growth under salt and drought stress. Plant Growth Regul. 67, 281–304. An, Y.Y., Zhang, M.X., Liu, G.B., Han, R.L., Liang, Z.S., 2013. Proline accumulation in leaves of Periploca sepium via both biosynthesis up-regulation and transport during recovery from severe drought. PLoS One 8, e69942. Ansari, M.I., Lee, R.H., Chen, S.C.G., 2005. A novel senescence associated gene encoding ␥-aminobutyric acid (GABA): pyruvate transaminase is upregulated during rice leaf senescence. Physiol. Plant 123, 1–8. Ashraf, M., Foolad, M.R., 2007. Roles of glycine betaine and proline in improving plant abiotic stress resistance. Environ. Exp. Bot. 59, 206–216. ˜ Banon, S., Ochoa, J., Franco, J.A., Alarcón, J.J., Sánchez-Blanco, M.J., 2006. Hardening of oleander seedlings by deficit irrigation and low air humidity. Environ. Exp. Bot. 56, 36–43. Bai, L.P., Sui, F.G., Ti-Da, G.E., Sun, Z.H., Yin-Yan, L.U., Zhou, G.S., 2006. Effect of soil drought stress on leaf water status, membrane permeability and enzymatic antioxidant system of maize. Pedosphere 16, 326–332. Bargali, K., Tewari, A., 2004. Growth and water relation parameters in drought-stressed Coriaria nepalensis seedlings. J. Arid Environ. 58, 505–512. Blum, A., 1996. Crop responses to drought and the interpretation of adaptation. Plant Growth Regul. 20, 135–148. Boyer, J.S., 1982. Plant productivity and environment. Science 218, 443–448. Cao, X., Jia, J., Zhang, C., Li, H., Liu, T., Jiang, X., Polle, A., Peng, C., Luo, Z.B., 2014. Anatomical, physiological and transcriptional responses of two contrasting poplar genotypes to drought and re-watering. Physiol. Plant 151, 480–494. Carvalho, R.C.D., Cunha, A., Silva, J.M.D., 2011. Photosynthesis by six portuguese maize cultivars during drought stress and recovery. Acta Physiol. Plant. 33, 359–374. Chaves, M.M., Maroco, J.P., Pereira, J.S., 2003. Understanding plant response to drought-from genes to the whole plant. Funct. Plant Biol. 30, 239–264. Chaves, M.M., Flexas, J., Pinheiro, C., 2009. Photosynthesis under drought and salt stress-regulation mechanisms from the whole plant to cell. Ann. Bot. 103, 551–560. Correia, B., Pintó-Marijuan, M., Neves, L., Brossa, R., Dias, M.C., Costa, A., Castro, B.B., Araújo, C., Santos, C., Chaves, M.M., Pinto, G., 2014. Water stress and recovery in the performance of two Eucalyptus globulus clones: physiological and biochemical profiles. Physiol. Plant 150, 580–592.

Dawson, T.P., Jackson, S.T., House, J.I., Prentice, I.C., Mace, G.M., 2011. Beyond predictions: biodiversity conservation in a changing climate. Science 332, 53–58. Dixon, R.A., 2001. Natural products and plant disease resistance. Nature 411, 843–847. Dutta, B., Kanani, H., Quackenbush, J., Klapa, M.I., 2009. Time-series integrated Omic analyses to elucidate short-term stress-induced responses in plant liquid cultures. Biotechnol. Bioeng. 102, 264–279. Echevarria-Zomeno, S., Ariza, D., Jorge, I., Lenz, C., Del Campo, A., Jorrin, J.V., Navarro, R.M., 2009. Changes in the protein profile of Quercus ilex leaves in response to drought stress and recovery. J. Plant Physiol. 166, 233–245. Efeo˘glu, B., Ekmekc¸i, Y., C¸ic¸ek, N., 2009. Physiological responses of three maize cultivars to drought stress and recovery. S. Afr. J. Bot. 75, 34–42. Fait, A., Fromm, H., Walter, D., Galili, G., Fernie, A.R., 2008. Highway or byway: the metabolic role of the GABA shunt in plants. Trends Plant Sci. 13, 14–19. Fan, T.W.M., 1996. Metabolite profiling by one- and twodimensional NMR analysis of complex mixtures. Prog. Nucl. Magn. Reson. Spectrosc. 28, 161–219. Fiehn, O., 2002. Metabolomics-the link between genotypes and phenotypes. Plant Mol. Biol. 48, 155–171. Flexas, J., Bota, J., Loreta, F., Cornic, G., Sharkey, T.D., 2004. Diffusive and metabolic limitations to photosynthesis under drought and salinity in C3 plants. Plant Biol. 6, 1–11. Fortunati, A., Barta, C., Brilli, F., Centritto, M., Zimmer, I., Schnitzler, J.-P., Loreto, F., 2008. Isoprene emission is not temperature-dependent during and after severe drought stress: a physiological and biochemical analysis. Plant J. 55, 687–697. Foster, K., Lambers, H., Real, D., Ramankutty, P., Cawthray, G.R., Ryan, M.H., 2015. Drought resistance and recovery in mature Bituminaria bituminosa var. albomarginata. Ann. Appl. Biol. 166, 154–169. Foyer, C.H., Valadier, M.-H., Mige, A., Becker, T.W., 1998. Drought induced effects on nitrate reductase activity and mRNA and on the coordination of nitrogen and carbon metabolism in maize leaves. Plant Physiol. 117, 283–292. Gallé, A., Haldimann, P., Feller, U., 2007. Photosynthetic performance and water relations in young pubescent oak (Quercus pubescens) trees during drought stress and recovery. New Phytol. 174, 799–810. Gallé, A., Florez-Sarasa, I., Aououad, H.E., Flexas, J., 2011. The Mediterranean evergreen Quercus ilex and the semi-deciduous Cistus albidus differ in their leaf gas exchange regulation and acclimation to repeated drought and re-watering cycles. J. Exp. Bot. 62, 5207–5216. Gavaghan, C.L., Li, J.V., Hadfield, S.T., Hole, S., Nicholson, J.K., Wilson, I.D., Howe, P.W.A., Stanley, P.D., Holmes, E., 2011. Application of NMR-based metabolomics to the investigation of salt stress in maize (Zea mays). Phytochem. Anal. 22, 214–224. Gazanchian, A., Hajheidari, M., Sima, N.K., Salekdeh, G.H., 2007. Proteome response of Elymus elongatum to severe water stress and recovery. J. Exp. Bot. 58, 291–300. Gechev, T.S., Dinakar, C., Benina, M., Toneva, V., Bartels, D., 2013. Molecular mechanisms of desiccation tolerance in resurrection plants. Cell Mol. Life Sci. 69, 3175–3186. Gehlenborg, N., O’Donoghue, S.I., Baliga, N.S., Goesmann, A., Hibbs, M.A., Kitano, H., Kohlbacher, O., Neuweger, H., Schneider, R., Tenenbaum, D., Gavin, A.C., 2010. Visualization of omics data for systems biology. Nat. Methods 7, S56–S68. Gibson, S.I., 2005. Control of plant development and gene expression by sugar signaling. Curr. Opin. Plant Biol. 8, 93–102. Harb, A., Krishnan, A., Ambavaram, M.M.R., Pereira, A., 2010. Molecular and physiological analysis of drought stress in Arabidopsis reveals early responses leading to acclimation in plant growth. Plant Physiol. 154, 1254–1271. Hasanuzzaman, M., Nahar, K., Alam, M.M., Roychowdhaury, R., Fujita, M., 2013. Physiological biochemical, and molecular mechanisms of heat stress tolerance in plants. Int. J. Mol. Sci. 14, 9643–9684. Heckathorn, S.A., Delucia, E.H., 1994. Drought-induced nitrogen retranslocation in perennial C4 grass of tallgrass prairie. Ecology 75, 1877–1886. Hochberg, U., Degu, A., Toubiana, D., Gendler, T., Nikoloski, Z., Rachmilevitch, S., Fait, A., 2013. Metabolite profiling and network analysis reveal coordinated changes in grapevine water stress response. BMC Plant Biol. 13, 184–200. Husen, A., Iqbal, M., Aref, I.M., 2014. Growth water status, and leaf characteristics of Brassica carinata under drought and rehydration conditions. Braz. J. Bot. 37, 217–227. Intergovernmental Panel on Climate Change (IPCC), 2007. Climate Change 2007: Synthesis Report. Cambridge University Press, Cambridge, UK, http://www. ipcc.ch/pdf/assessmentreport/ar4/syr/ar4 syr.pdf (accessed 17.11.07.). Izanloo, A., Condon, A.G., Langridge, P., Tester, M., Schnurbusch, T., 2008. Different mechanisms of adaptation to cyclic water stress in two South Australian bread wheat cultivars. J. Exp. Bot. 59, 3327–3346. Kang, Y., Han, Y., Torres-Jerez, I., Wang, M., Tang, Y., Monteros, M., Udvardi, M., 2011. System responses to long-term drought and re-watering of two contrasting alfalfa varieties. Plant J. 68, 871–889. Kim, J.K., Bamba, T., Harada, K., Fukusaki, E., Kobayashi, A., 2007. Time-course metabolic profiling in Arabidopsis thaliana cell cultures after salt stress treatment. J. Exp. Bot. 58, 415–424. Krasensky, J., Jonak, C., 2012. Drought, salt, and temperature stress-induced metabolic rearrangements and regulatory networks. J. Exp. Bot. 63, 1593–1608. Lipiec, J., Doussan, C., Nosalewicz, A., Kondracka, K., 2013. Effect of drought and heat stresses on plant growth and yield: a review. Int. Agrophys. 27, 463–477. Lobato, A.K.D., Neto, C.F.D., dos Santos, B.G., da Costa, R.C.L., Cruz, F.J.R., Neves, H.K.B., Lopes, M.J.D., 2008. Physiological and biochemical behaviour in soybean (Glycine max cv. Sambaiba) plants under water deficit. Aust. J. Crop Sci. 2, 25–32.

C. Sun et al. / Agricultural Water Management 172 (2016) 62–73 Loewenstein, N.J., Pallardy, S.G., 2002. Influence of a drying cycle on post-drought xylem sap abscisic acid and stomatal responses in young temperate deciduous angiosperms. New Phytol. 156, 351–361. Mahajan, S., Tuteja, N., 2005. Cold, salinity and drought stresses: an overview. Arch. Biochem. Biophys. 444, 139–158. Marron, N., Dreyer, E., Boudouresque, E., Delay, D., Petit, J.-M., Delmotte, F.M., Brignolas, F., 2003. Impact of successive drought and re-watering cycles on growth and specific leaf area of two Populus × canadensis (Moench) clones,‘Dorskamp’ and ‘Luisa Avanzo’. Tree Physiol. 23, 1225–1235. Mattoo, A.K., Sobolev, A.P., Neelam, A., Goyal, R.K., Handa, A.K., Segre, A.L., 2006. Nuclear magnetic resonance spectroscopy-based metabolite profiling of transgenic tomato fruit engineered to accumulate spermidine and spermine reveals enhanced anabolic and nitrogen-carbon interactions. Plant Physiol. 142, 1759–1770. McNeil, S.D., Nuccio, M.L., Ziemak, M.J., Hanson, A.D., 2001. Enhanced synthesis of choline and glycine betaine in transgenic tobacco plants that overexpress phosphoethanolamine N-methyltransferase. Proc. Natl. Acad. Sci. U. S. A. 98, 10001–10005. Meyer, E., Aspinwall, M.J., Lowry, D.B., Palacio-Mejía, J.D., Logan, T.L., Fay, P.A., Juenger, T.E., 2014. Integrating transcriptional metabolomic, and physiological responses to drought stress and recovery in switchgrass (Panicum virgatum L.). BMC Genom. 15, 527–542. Miyashita, K., Tanakamaru, S., Maitani, T., Kimura, K., 2005. Recovery responses of photosynthesis transpiration, and stomatal conductance in kidney bean following drought stress. Environ. Exp. Bot. 53, 205–214. Morgan, J.M., 1984. Osmoregulation and water stress in higher plants. Ann. Rev. Plant Physiol. 35, 299–319. Obata, T., Fernie, A.R., 2012. The use of metabolomics to dissect plant responses to abiotic stresses. Cell Mol. Life Sci. 69, 3225–3243. Okami, M., Kato, Y., Kobayashi, N., Yamagishi, J., 2015. Morphological traits associated with vegetative growth of rice (Oryza sativa L.) during the recovery phase after early-season drought. Eur. J. Agron. 64, 58–66. Oliver, M.J., Wood, A.J., Mahony, P.O., 1998. To dryness and beyong-preparation for the dried state and rehydrationin vegetative desiccation-tolerant plants. Plant Growth Regul. 24, 193–201. Pang, J., Yang, J., Ward, P., Siddique, K.H.M., Lambers, H., Tibbett, M., Ryan, M., 2011. Contrasting responses to drought stress in herbaceous perennial legumes. Plant Soil 348, 299–314. Pinheiro, C., Chaves, M.M., 2011. Photosynthesis and drought: can we make metabolic connections from available data? J. Exp. Bot. 62, 869–882. Ramakrishna, A., Ravishankar, G.A., 2011. Influence of abiotic stress signals on secondary metabolites in plants. Plant Signal. Behav. 6, 1720–1731. Ramel, F., Sulmon, C., Gouesbet, G., Couee, I., 2009. Natural variation reveals relationships between pre-stress carbohydrate nutritional status and subsequent responses to xenobiotic and oxidative stress in Arabidopsis thaliana. Ann. Bot. 104, 1323–1337. Reddy, A.R., Chaitanya, K.V., Vivekanandan, M., 2004. Drought-induced responses of photosynthesis and antioxidant metabolism in higher plants. J. Plant Physiol. 161, 1189–1202. Reynolds, J.F., Kemp, P.R., Ogle, K., Fernández, R.J., 2004. Modifying the ‘pulse-reserve’ paradigm for deserts of North America: precipitation pulses, soil water, and plant responses. Oecologia 141, 194–210. Rizhsky, L., Liang, H., Shuman, J., Shulaev, V., Davletova, S., Mittler, R., 2004. When defense pathways collide: the response of Arabidopsis to a combination of drought and heat stress. Plant Physiol. 134, 1683–1696. Rolland, F., Baena-Gonzalez, E., Sheen, J., 2006. Sugar sensing and signaling in plants: conserved and novel mechanisms. Annu. Rev. Plant Biol. 57, 675–709. Sanchez, D.H., Pieckenstain, F.L., Szymanski, J., Erban, A., Bromke, M., Hannah, M.A., Kraemer, U., Kopka, J., Udvardi, M.K., 2011. Comparative functional genomics of salt stress in related model and cultivated plants identifies and overcomes limitations to translational genomics. PLoS One 6, e17094. Seebauer, J.R., Moose, S.P., Fabbri, B.J., Crossland, L.D., Below, F.E., 2004. Amino acid metabolism in young maize earshoots: implications for assimilate movement and nitrogen signaling. Plant Physiol. 136, 4326–4334. Shannon, P., Markiel, A., Ozier, O., Baliga, N.S., Wang, J.T., Ramage, D., Amin, N., Schwikowski, B., Ideker, T., 2003. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504. Sharma, S., Villamor, J.G., Verslues, P.E., 2011. Essential role of tissue-specific proline synthesis and catabolism in growth and redox balance at low water potential. Plant Physiol. 157, 292–304. Sharp, R.E., Poroyko, V., Hejlek, L.G., Spollen, W.G., Springer, G.K., Bohnert, H.J., Nguyen, H.T., 2004. Root growth maintenance during water deficits: physiology to functional genomics. J. Exp. Bot. 55, 2343–2351.

73

Shelp, B.J., Bown, A.W., McLean, M.D., 1999. Metabolism and functions of gamma-aminobutyric acid. Trends Plant Sci. 4, 1360–1385. Shen, B., Jensen, R.G., Bohnert, H.J., 1997. Increased resistance to oxidative stress in transgenic plants by targeting mannitol biosynthesis to chloroplasts. Plant Physiol. 113, 1177–1183. Shi, J., Yasuor, H., Yermiyahu, U., Zuo, Q., Ben-Gal, A., 2014. Dynamic responses of wheat to drought and nitrogen stresses during re-watering cycles. Agric. Water Manage. 146, 163–172. Shulaev, V., Cortes, D., Miller, G., Mittler, R., 2008. Metabolomics for plant stress response. Plant Physiol. 132, 199–208. Sicher, R.C., Timlina, D., Bailey, B., 2012. Responses of growth and primary metabolism of water-stressed barley roots to rehydration. J. Plant Physiol. 169, 686–695. Siopongco, J.D.L.C., Yamauchi, A., Salekdeh, H., Bennett, J., Wade, L.J., 2006. Growth and water use response of doubled-haploid rice lines to drought and re-watering during the vegetative stage. Plant Prod. Sci. 9, 141–151. Suguiyama, V.F., Silva, E.A., Meirelles, S.T., Centeno, D.C., Braga, M.R., 2014. Leaf metabolite profile of the Brazilian resurrection plant Barbacenia purpurea Hook (Velloziaceae) shows two time-dependent responses during desiccation and recovering. Front. Plant Sci. 5, 1–13. Sun, Y.Q., Yan, Fei, Liu, F.L., 2013. Drying/rewetting cycles of the soil under alternate partial root-zone drying irrigation reduce carbon and nitrogen retention in the soil-plant systems of potato. Agric. Water Manage. 128, 85–91. Sun, C.B., Fan, X.W., Hu, H.Y., Liang, Y., Huang, Z.B., Pan, J.L., Wang, L., Li, Y.Z., 2013. Pivotal metabolic pathways related to water deficit tolerance and growth recovery of whole maize plant. Plant Omics 6, 377–387. Sun, C.X., Gao, X.X., Fu, J.Q., Zhou, J.H., Wu, X.F., 2015. Metabolic response of maize (Zea mays L.) plants to combined drought and salt stress. Plant Soil 388, 99–117. Szabados, L., Savouré, A., 2010. Proline: a multifunctional amino acid. Trends Plant Sci. 15, 89–97. Szalay, M.S., Kovács, I.A., Korcsmáros, T., Böde, C., Csermely, P., 2007. Stress-induced rearrangements of cellular networks: consequences for protection and drug design. FEBS Lett. 581, 3675–3680. Tate, A.R., Damment, S.J.P., Lindon, J.C., 2001. Investigation of the metabolite variation in control rat urine using 1 H NMR. Anal. Biochem. 291, 17–26. Toscano, S., Scuderi, D., Giuffrida, F., Romano, D., 2014. Responses of Mediterranean ornamental shrubs to drought stress and recovery. Sci. Hortic. 178, 145–153. Urano, K., Kurihara, Y., Seki, M., Shinozaki, K., 2010. Omics’ analyses of regulatory networks in plant abiotic stress responses. Curr. Opin. Plant Biol. 13, 1–7. Valluru, R., Van den Ende, W., 2011. Myo-inositol and beyond—emerging networks under stress. Plant Sci. 181, 387–400. Van Heerden, P.D., Krüger, G.H.J., 2002. Separately and simultaneously induced dark chilling and drought stress effects on photosynthesis, proline accumulation and antioxidant metabolism in soybean. J. Plant Physiol. 159, 1077–1086. Vankova, R., Dobra, J., Storchova, H., 2012. Recovery from drought stress in tobacco: an active process associated with the reversal of senescence in some plant parts and the sacrifice of others. Plant Signal. Behav. 7, 19–21. Verslues, P.E., Juenger, T.E., 2011. Drought metabolites, and Arabidopsis natural variation: a promising combination for understanding adaptation to water-limited environments. Curr. Opin. Plant Biol. 14, 240–245. Warren, C.R., Aranda, I., Cano, F.J., 2012. Metabolomics demonstrates divergent responses of two Eucalyptus species to water stress. Metabolomics 8, 186–200. Wiegand, T., Snyman, H.A., Kellner, K., Paruelo, J.M., 2004. Do grasslands have a memory: modeling phytomass production of a semiarid South African grassland. Ecosystems 7, 243–258. Wingler, A., Roitsch, T., 2008. Metabolic regulation of leaf senescence: interactions of sugar signalling with biotic and abiotic stress responses. Plant Biol. 10, 50–62. Xu, Z.Z., Zhou, G.S., 2006. Nitrogen metabolism and photosynthesis in Leymus chinensis in response to long-term soil drought. J. Plant Growth Regul. 25, 252–266. Xu, Z.Z., Zhou, G.S., Shimizu, H., 2009. Are plant growth and photosynthesis limited by pre-drought following rewatering in grass? J. Exp. Bot. 60, 3737–3749. Xu, Z.Z., Zhou, G.S., Shimizu, H., 2010. Plant responses to drought and rewatering. Plant Signal. Behav. 5 (6), 649–654. Yahdjian, L., Sala, O.E., 2006. Vegetation structure constrains primary production response to water availability in the Patagonian steppe. Ecology 87, 952–962. Zhang, J.Y., de Carvalho, M.H.C., Torres-Jerez, I., Kang, Y., Allen, S.N., Huhman, D.V., Tang, Y., Murray, J., Sumner, L.W., Udvardi, M.K., 2014. Global reprogramming of transcription and metabolism in Medicago truncatula during progressive drought and after rewatering. Plant Cell Environ. 37, 2553–2576.