Initial water deficit effects on Lupinus albus photosynthetic performance, carbon metabolism, and hormonal balance: metabolic reorganization prior to early stress responses

Abstract

The early (2–4 d) effects of slowly imposed soil water deficit on Lupinus albus photosynthetic performance, carbon metabolism, and hormonal balance in different organs (leaf blade, stem stele, stem cortex, and root) were evaluated on 23-d-old plants (growth chamber assay). Our work shows that several metabolic adjustments occurred prior to alteration of the plant water status, implying that water deficit is perceived before the change in plant water status. The slow, progressive decline in soil water content started to be visible 3 d after withholding water (3 DAW). The earliest plant changes were associated with organ-specific metabolic responses (particularly in the leaves) and with leaf conductance and only later with plant water status and photosynthetic rate (4 DAW) or photosynthetic capacity (according to the Farquhar model; 6 DAW). Principal component analysis (PCA) of the physiological parameters, the carbohydrate and the hormone levels and their relative values, as well as leaf water-soluble metabolites full scan data (LC-MS/MS), showed separation of the different sampling dates. At 6 DAW classically described stress responses are observed, with plant water status, ABA level, and root hormonal balance contributing to the separation of these samples. Discrimination of earlier stress stages (3 and 4 DAW) is only achieved when the relative levels of indole-3-acetic acid (IAA), cytokinins (Cks), and carbon metabolism (glucose, sucrose, raffinose, and starch levels) are taken into account. Our working hypothesis is that, in addition to single responses (e.g. ABA increase), the combined alterations in hormone and carbohydrate levels play an important role in the stress response mechanism. Response to more advanced stress appears to be associated with a combination of cumulative changes, occurring in several plant organs. The carbohydrate and hormonal balance in the leaf (IAA to bioactive-Cks; soluble sugars to IAA and starch to IAA; relative abundances of the different soluble sugars) flag the initial responses to the slight decrease in soil water availability (10–15% decrease). Further alterations in sucrose to ABA and in raffinose to ABA relative values (in all organs) indicate that soil water availability continues to decrease. Such alterations when associated with changes in the root hormone balance indicate that the stress response is initiated. It is concluded that metabolic balance (e.g. IAA/bioactive Cks, carbohydrates/IAA, sucrose/ABA, raffinose/ABA, ABA/IAA) is relevant in triggering adjustment mechanisms.

Carbon metabolism, hormone balance, LC-MS, metabolic reorganization, organ relative water content, plant performance, plant water status, soil water deficit

Introduction

The effects of water deficit on photosynthesis, plant growth, and source–sink relationships have been studied widely in the last decade (Chaves et al., 2002; Neumann, 2008), with the emphasis on short-term, severely imposed water deficit (WD). Much less information is available on the early stages (2–4 d) of a slowly imposed soil WD and the underlying metabolic events. Moreover, an integrated picture involving several plant organs in relation to water status and photosynthetic performance is usually not available. Specifically, it is not known how the earliest effects on metabolism are related to the decline in water status (both soil water content and plant water status), but some reports show that soil water deficit affects plant metabolism even in the absence of changes in the plant water status (Pustovoitova et al., 2003; Pinheiro et al., 2005; Antonio et al., 2008). The metabolic status provides a link between environmental signals (soil water content), growth regulation, and plant performance. Optimal utilization of the available carbohydrate for growth and development has been suggested (Hanson and Smeekens, 2009; Robertson et al., 2009), with WD leading to sugar accumulation in an organ- and stress-intensity dependent manner (Pinheiro et al., 2001; Pinheiro et al., 2004; Antonio et al., 2008; Wingler and Roitsch, 2008). The cycle of starch synthesis and breakdown provides another way to modulate stress responses (Usadel et al., 2008; Sulpice et al., 2009), with the circadian clock proposed to be involved in such modulation during the early stress response, in order to maximize carbon uptake and growth (Robertson et al., 2009). Moreover, responses to the environment and the regulation of the growth patterns involve extensive interactions between carbohydrate status, reserve utilization, and the plant hormone network, such as abscisic acid (ABA), indole-3-acetic acid (IAA), and cytokinins (Cks) (Gibson, 2004; Gonzali et al., 2006; Sakakibara, 2006; Wingler and Roitsch, 2008). It is still unknown how such regulation occurs at the whole plant level, when several organs that react differently to the same stress are considered (Pinheiro et al., 2001; Pinheiro et al., 2004). It can be hypothesized that there must be a mechanism to discriminate between an environmental fluctuation and a steady, progressive (although minor) soil water deficit, which implies plant adaptation to different degrees of stress and the need to make adjustments on a daily basis (Boyer, 2010). Since water deficit can lead to starvation of photosynthetic products (Boyer, 2010; Pinheiro and Chaves, 2011), the hormonal and carbohydrate balance at the whole plant level are good candidates for the water deficit decision-making mechanism(s).

An experiment has been designed that aims to establish a timeline of events at the organ level, detecting the initial metabolic alterations in a range of different plant organs and relating them to changes in photosynthetic performance, soil water content, and plant water status. Our working hypothesis is that ABA levels are increased only under de facto stress conditions, and that initial stress responses are triggered by alterations in hormone balance and carbohydrate availability. Such responses can range from: (i) metabolic reorganisation (that are defined here as a response to a changing environment, without affecting the plant water status; (ii) initial stress responses when plant water status are affected; and (iii) established water deficit when the photosynthetic capacity is affected and ABA levels are dramatically increased in all organs. With this multilevel and comprehensive approach, the aim was to determine the onset of stress (i.e. when it begins affecting plant metabolism) and to detect critical events for the activation of the stress response mechanism(s).

Materials and methods

Plant material

Lupinus albus L. plants (cv. Rio Maior) were cultivated on a sterilized soil, peat, and sand mixture (1:1:1, by vol.) in controlled-environment growth chambers: photon flux density 290–320 μmol m⁻² s⁻¹ photosynthetically active radiation (PAR), photoperiod (12 h), temperature (19/25 °C, night/day), and relative humidity (65–70%) as previously described by Pinheiro et al. (2005). Under these conditions, and without fertilization, lupin plants developed rhizobia-containing nodules. Twenty-three days after sowing, WD was slowly induced by withholding watering. Plants were collected 3, 4, 5, 6, 9, and 13 d after withholding water (DAW), and 1 d and 2 d after re-watering (RW). Control plants were watered throughout the whole period. Sample collection took place 3–4 h after the beginning of the photoperiod, and the stem was always separated into vascular (stele) and cortical (cortex) tissue (Pinheiro et al., 2004; Antonio et al., 2008). The leaf blade, fleshy root, stem stele, and stem cortex were immediately frozen in liquid nitrogen, lyophilized, and stored at –80 °C until extraction.

Soil water content and plant water status

Soil water content (%) was measured with a ThetaProbe soil moisture sensor (ML2x ThetaProbe coupled to a ThetaMeter type HH2 from Delta-T Devices Ltd, Cambridge, UK). Leaf water potential was measured with a Scholander pressure chamber (PMS instrument Co, Corvallis, Oregon, USA) at predawn (Ψleaf pd). The relative water content (RWC) for leaf blade, stem cortex, stem stele, and root was determined as previously described by Rodrigues et al. (1995).

Gas-exchange measurements

Net CO₂ assimilation rate (A, μmol m⁻² s⁻¹) and stomatal conductance (gs, mmol m⁻² s⁻¹) were measured 2–3 h after the beginning of the photoperiod, in six plants per treatment, using a portable open gas exchange photosynthesis system LI-6400 (Li-Cor Inc., Lincoln, NE, USA). The most recently expanded leaves at the onset of the stress period were marked and used throughout the entire assay. A and gs values were used to calculate the instantaneous intrinsic water use efficiency (WUE, A/gs).

For the sampling dates 0, 2, 4, 6, and 13 DAW and 1 RW, curves of CO₂ assimilation versus intercellular CO₂ concentration (A/Ci) were also measured in three plants per treatment using the open gas exchange system LI-6400 with an integrated fluorescence chamber head (LI-6400-40) following the protocol of Long and Bernacchi (2003). Cuvette conditions were maintained at a photosynthetic photon flux density (PPFD) of 250 or 1000 μmol m⁻² s⁻¹, relative humidity of 60%, and a leaf temperature of 25 °C. Measurements were carried out by changing ambient CO₂ concentration (Ca) in the cuvette, controlled with a CO₂ mixer, in the following order: 400, 300, 200 100, 50, 400, 400, 600, 800, 1000, and 1200 μmol CO₂ mol⁻¹ air. Non-linear regression techniques, based on the Farquhar model (Farquhar et al., 1980) and later modifications (Sharkey, 1985; Harley and Sharkey, 1991), were used to estimate the maximum ratio of Rubisco carboxylation (Vcmax), the maximum electron transport capacity at saturating light (Jmax) and the velocity for triose phosphate utilization (VTPU) for each A/Ci curve. Our analysis was developed in Excel following the specifications of Long and Bernacchi (2003).

Plant hormone extraction and purification

IAA and Cks were extracted and purified as previously described by Dobrev and Kaminek (2002). Briefly, leaf and root samples were ground in liquid nitrogen and extracted overnight with methanol:water:formic acid (15:4:1, vol., pH 2.5, –20 °C). For the analysis of endogenous Cks, 50 pmol of 12 deuterium-labelled Ck standards were added while, for IAA, a tritiated internal standard was used (Dobrev and Kaminek, 2002). The extracts were purified using Sep-Pak Plus Si-C₁₈ columns and Oasis MCX mixed mode columns (Waters, MA, USA) and evaporated. Dried extracts of leaf blades and fleshy roots were re-dissolved in 10% acetonitrile in water and filtered. Cks were quantified by HPLC-MS, using a TSQ Quantum Ultra AM triple-quadrupole mass spectrometer (Thermo Electron, San Jose, USA). The mass spectrometer was operated in the positive single reaction monitoring (SRM) MS/MS mode with monitoring of between two and four transitions for each compound (Dobrev et al., 2002). The quantification was made using calibration curves with [²H]-labelled Cks as internal standards (6–10 concentration points). Detection limits of different Cks varied from 0.05– 0.1 pmol sample⁻¹. Results represent averages of analyses of three independent samples and of two HPLC-MS/MS injections for each sample.

Levels of IAA were determined using two-dimensional HPLC with fluorescence detection (Perkin Elmer), as previously described by Dobrev et al. (2005).

ABA was extracted and quantified as described by Jiang et al. (2004). Freeze-dried tissue samples were homogenized and extracted in 80% aqueous methanol. After partial purification on a Sep-Pak C18 cartridge (Waters, MA, USA), methanol was removed and the aqueous residue was partitioned against ethyl acetate at pH 3.0. The ethyl acetate of the combined organic fractions was removed under reduced pressure. The newly obtained residue was taken up in TBS-buffer (TRIS buffered saline; 150 mmol l⁻¹ NaCl, 1 mmol l⁻¹ MgCl₂, and 50 mmol l⁻¹ TRIS at pH 7.8) and subjected to an immunological ABA assay (ELISA) as described by Peuke et al. (1994). The accuracy of the ELISA has been verified, and recoveries of ABA during the purification procedures were checked using radioactive ABA and found to be higher than 95%.

Extraction of water-soluble carbohydrates

Water-soluble carbohydrates were extracted from the different L. albus organs following the addition of chloroform/methanol as described previously (Antonio et al., 2008). Briefly, lyophilized plant material was finely ground in liquid nitrogen and extracted with 250 μl ice-cold chloroform:methanol (3:7, v/v), vortex-mixed and incubated for 2 h at –20 °C. After incubation, samples were twice-extracted with ice-cold water and, after centrifugation, the upper phases were collected and pooled. The combined supernatants containing the water-soluble carbohydrates were evaporated to dryness (Savant SpeedVac system, Thermo Electron Corporation, Runcorn, UK). Samples were reconstituted in 100 μl water and centrifuged at 6800 g at 20 °C for 30 min followed by liquid chromatography ion trap mass spectrometry (LC-MS) analysis.

Quantification of water-soluble carbohydrates

LC-MS analyses were performed on a Thermo Finnigan Surveyor HPLC system coupled to an ion trap mass spectrometer (LCQ DECA XP Plus, Thermo Electron, San Jose, CA, USA), equipped with a Thermo Finnigan orthogonal electrospray interface (Antonio et al., 2008). Neutral carbohydrates (glucose, sucrose, raffinose, fructose, and trehalose), sugar alcohols (mannitol, sorbitol, galactinol, and maltitol), and the polyol myo-inositol were detected in the negative ion mode. Mass spectra were acquired over the scan range m/z 50–1000, and data were processed using Xcalibur 1.3 software (Thermo Finnigan, San Jose, CA, USA). Precursor ions were selected with an isolation width of 2 m/z units and activated for 30 ms. Chromatographic separation was carried out using a PGC Hypercarb™ column (5 μm, 100×4.6 mm; Thermo Electron, Runcorn, Cheshire, UK) at a flow rate of 600 μl min⁻¹. The sample injection volume was 20 μl and the PGC column was used at ambient temperature (25 °C). The binary mobile phase was composed of (A) water modified with 0.1% (v/v) of formic acid (FA) and (B) acetonitrile modified with 0.1% FA. The gradient elution was as follows: 0–5 min, 96% A+4% B to 92% A+8% B; 5–7 min, 92% A+8% B to 75% A+25% B, and maintained for 3 min, followed by column re-equilibration: 10–11 min, 75% A+25% B to 50% A+50% B, and maintained for 8 min; 19–20 min, 50% A+50% B to 96% A+4% B and maintained for 10 min.

Starch extraction and quantification

The pellet resulting from the chloroform:methanol extraction of the different L. albus organs was washed twice with water. Ten volumes of water were added to the pellet that was boiled for 3 min and incubated at 130 °C for 1 h. After cooling, samples were incubated for 2 h at pH 4.8 and 60 °C with amyloglucosidase (Roche Applied Science, Amadora, Portugal), starch being quantified in the supernatant as previously described by Pinheiro et al. (2001).

Statistical analysis

Two distinct approaches were used to assess for significant alterations induced by soil water deficit. Univariate analysis was used to provide significance levels for the individual parameters (SWC, Ψpd, RWC, gs, A, WUE, Vcmax, Jmax, TPU, ABA, IAA, Cks, starch, glucose, sucrose, raffinose). The non-parametric Mann–Whitney U test was used in STATISTICA^® version 5.0 (StatSoft Inc., Oklahoma, USA). Data for each DAW that differ significantly with respect to day zero were assigned * (P <0.05), ** (P <0.01) or *** (P <0.001).

Unsupervised multivariate analysis was used to explore the relationships between the physiological and metabolic alterations in the early WD stages. In addition, the relative values (log ratios; see Supplementary Tables S15 and Supplementary Data at JXB online) were used in order to evaluate the relevance of metabolic co-occurrences. Principal Components Analysis (PCA) was carried out using software written in-house in the Matlab software environment (Matlab, 2002). The full LC-MS datasets were also analysed to provide a detailed and untargeted analysis of the water-soluble metabolites in the extracts of the different plant organs (root, stem stele, stem cortex and blade). As the samples differed in dry weight, the raw data were normalized to the total ion count to account for any differences in concentration. The adaptive binning method of Davis et al. (2007) was used to obtain bins for m/z values by identifying the minima in a smoothed reference spectrum. The data were averaged over time and the median value over all observations at each point in the time-averaged spectra taken as the reference spectrum, which was then smoothed using a two level non-decimated wavelet transform. The bin ends identified as the minima in this reference spectrum were then applied to the original spectra and a time series for each bin obtained. A similar procedure was then used to bin the time series for each m/z bin separately. The resulting binned values were then used as variables in PCA. To prevent large variables dominating the analysis, the variables were scaled to unit variance.

Results

Soil water deficit imposition and plant water status

Daily monitoring of the soil water content (SWC; see Supplementary Table S1 at JXB online) made it possible to detect a significant difference in relation to day 0 at 3 DAW (from 22.3% down to 19.7%, i.e. a 12% decrease), the wilting point (6.4%) being achieved at 13 DAW (a 75% decrease in the SWC) as previously described by Pinheiro et al. (2005). The plant water status was evaluated through the determination of the leaf water potential at predawn and the relative water content of the different plant organs (see Supplementary Table S2 at JXB online). At 4 DAW (a 30% reduction in the SWC), the leaf Ψpd decreased from –0.36 to –0.48 MPa (∼25% decrease) but only in the stem cortex was the RWC affected (from 73% to 67%, i.e. a 10% decrease). Further alterations in the RWC were recorded at 5 DAW (root), 6 DAW (blade), and 9 DAW (stem stele). On re-watering, the plant water status returned to that prior to withholding water (see Supplementary Table S2 at JXB online).

Leaf gas-exchange measurements

With a small decrease in the SWC (∼12%, i.e. 3 DAW), leaf conductance was affected (from 369 to 189 mmol m⁻² s⁻¹; Fig. 1A) while the photosynthetic rate (Fig. 1B), the plant water status (see Supplementary Table S2 at JXB online), the transpiration rate, and the vapour pressure deficit (see Supplementary Table S3 at JXB online) were not. As a result of stomatal closure, WUE was increased at 3 DAW (Fig. 1C) reaching a maximum at 6 DAW and steadily decreasing until re-watering took place. Considering the soil WD effects on net photosynthesis (Fig. 1B), at 4 DAW an early WD was imposed (15% reduction), 5 DAW and 6 DAW represented a mild WD (25% and 35% reduction) and 9 DAW and 13 DAW represented a severe and late WD (90–95% decrease). On re-watering, the CO₂ assimilation rate recovers to ∼70% of the control (Fig. 1B) and leaf conductance was quantified as ∼30% of the control (Fig. 1A). In order to determine and compare the limitations of photosynthesis under water deficit, the Farquhar parameters were considered (Fig. 2). Soil water deficit effects on the carboxylation rate (Vcmax) and the PSII electron transport rate (Jmax) were detected 4 DAW (PAR 1000, maximum potential) or 6 DAW (PAR 250 that reflect the growing conditions). After 6 DAW, when the plant water status was severely affected by the soil WD, the photosynthetic rate was also limited to the triose-phosphate utilisation (TPU) (9 or 13 DAW, for PAR 1000 or 250, respectively). The gradual decrease in soil water availability not only leads to larger stomatal limitation, but also decreases maximum rubisco carboxylation activity and electron transport, and therefore RuBP regeneration (Flexas et al., 1999). On re-watering, the estimated rates were calculated as 80–110% (PAR 250) or 40–60% (PAR 1000) of the controls.

Fig. 1.

Leaf conductance (A), net photosynthesis (B), and water use efficiency (C) of Lupinus albus leaves during the period of water-deficit imposition and on re-watering (shaded area). Mean values and standard errors (n=6) are presented. The significance levels were calculated using the Mann–Whitney U test, data differing significantly from day zero being labelled with * (P <0.05), ** (P <0.01). i) DAW (days after withholding water); ii) WW (well watered); iii) WD (water deficit). (This figure is available in colour at JXB online.)

Fig. 2.

Estimated Rubisco carboxylation (Vcmax) (A), maximum electron transport capacity at saturating light (Jmax) (B), and the velocity for triose phosphate utilization (TPU) (C) at PAR 250 (solid line) and PAR 1000 (dotted line) of Lupinus albus leaves during the period of water-deficit imposition and on re-watering (shaded area). Data are the means ±standard error of three measurements. The significance levels were calculated using the Mann–Whitney U test, data differing significantly from day zero (P <0.05) being labelled with * (PAR 250) or ^# (PAR 1000). i) DAW (days after withholding water); ii) WW (well watered); iii) WD (water deficit). (This figure is available in colour at JXB online.)

Phytohormone levels

The amount of free ABA was quantified in the different plant organs throughout the whole experiment (see Supplementary Table S4 at JXB online). A dramatic increase in ABA levels was observed in the four organs, and the maximum ABA content was determined 6 DAW (stem) and 9 DAW (leaf blade and root; see Supplementary Fig. S1 at JXB online). Although differences can be found 3 DAW (a minor ABA increase in blade and root and a decrease in the stem stele), the difference is only statistically significant in the case of stem stele at this early stage (from 283 to 149 pmol g⁻¹ DW). Significant ABA increases were later detected in the blade and root (4 DAW) and in the stem cortex (5 DAW). By 5 DAW, ABA levels were also significantly increased in the stem stele in relation to day zero. The initial ABA decrease in the stem stele coincided with small reductions in the soil water content and leaf conductance (10–15%), while more general ABA alterations were associated with a 30% decrease in SWC, a 40% decrease in leaf conductance and Ψpd, and a 25% decrease in CO₂ assimilation rate. The sharp ABA increase (6 DAW) was detected simultaneously with a reduction in the blade and root RWC and with biochemical limitations in CO₂ assimilation (Vcmax and Jmax reduction). A relationship between ABA biosynthesis and reduced RWC (leaf in addition to root) has already been described (Wilkinson and Davies, 2010).

The effects of the soil WD imposition on IAA and Ck content were evaluated in the leaf blade and root during the early WD stages, i.e. until the sharp ABA increase 6 DAW in all organs (see Supplementary Tables S5–Supplementary Data at JXB online). Cks were grouped according to their structure and function: bioactive Cks, cis-derivatives, O-glucosides (storage forms), N-glucosides (deactivation forms), and Ck phosphates (biosynthesis intermediates). Figure 3 (presented as the log ratio between WD and WW) shows that the earliest changes were detected for IAA and Cks (3 DAW) but not for ABA (4 DAW). It was found that the response of the different Ck classes was similar in both blade and root, with a reduction for bioactive-Ck and an increase in the other forms (Fig. 3). However, in these two organs, opposing effects on IAA were observed with a strong decrease in the blade and increase in the root (Fig. 3; see Supplementary Table S5 at JXB online). Taken together, the data show that soil water deficit reduced Ck physiological activity (decrease in the bioactive forms and increase in the deactivated forms, O and N-glucosylated). The pattern of soil water content reflected in the hormone level changes in the different plant organs is similar to that described in tomato for the early stages of exposure to salinity (Albacete et al., 2008). When considering the hormonal balance (calculated as the log ratios for ABA/IAA, ABA/bioactive Cks, and IAA/bioactive Cks; see Supplementary Fig. S2 at JXB online) it becomes clear that the hormone balance (ABA/IAA) is dramatically reversed in the leaf blade as early as 3 DAW, while a clear change is not detected in the root until 6 DAW.

Fig. 3.

Log ratio of ABA, IAA, and several Ck levels in Lupinus albus leaf blade (A) and root (B) during the period of the soil-water-deficit imposition. Values are the log ratio (water deficit/control), day 0 being considered the control. Mean values and standard errors (3≤n≤6) were considered. The significance levels were calculated using the Mann–Whitney U test, data differing significantly from day zero being labelled with * (P <0.05) or ** (P <0.01). (This figure is available in colour at JXB online.)

Carbohydrate analysis (targeted study)

The effects of slowly imposed WD on carbon metabolism (levels of starch, neutral sugars, and sugar alcohols) were investigated in the different L. albus organs (see Supplementary Tables S11–Supplementary Data at JXB online). It was possible to detect and quantify starch, glucose, sucrose, and raffinose but not fructose, trehalose, myo-inositol, mannitol, sorbitol, galactinol or maltitol.

Carbon metabolism was readily responsive to the earliest and smallest changes in the SWC (∼12% decrease; Figs 4, 5), showing distinct patterns in the four organs, Soil water deficit led to transient alterations in starch (Fig. 4), with the leaf blade and the root showing opposite trends: initial accumulation in the blade (3 and 4 DAW) and remobilization on mild and severe stress. Transient effects were also observed for glucose, sucrose, and raffinose (Fig. 5) and, with the sole exception of blade sucrose, these components accumulated at 6 DAW.

Fig. 4.

Starch log ratio in Lupinus albus leaf blade and root during the period of the soil-water-deficit imposition and recovery (RW). Values are the log ratio (water deficit/control), day 0 being considered the control. Mean values and standard errors (3≤n≤6) were considered. The significance levels were calculated using the Mann–Whitney U test, data differing significantly from day zero being labelled with * (P <0.05). (This figure is available in colour at JXB online.)

Fig. 5.

Log ratio of glucose, sucrose, and raffinose levels in Lupinus albus leaf blade (A), fleshy root (B), stem stele (C), and stem cortex (D) during the period of the soil-water-deficit imposition. Values are the log ratio (water deficit/control), day 0 being considered the control. Mean values and standard errors (3≤n≤6) were considered. The significance levels were calculated using the Mann–Whitney U test, data differing significantly from day zero being labelled with * (P <0.05). (This figure is available in colour at JXB online.)

Multivariate data analysis

In order to gain insight into the events occurring at the whole plant level as a consequence of progressive soil water deficit, principal component analysis (PCA) of the data was performed, using as variables: (i) all the physiological parameters and the quantities of hormones, starch, glucose, sucrose, and raffinose from the different organs (Fig. 6A); (ii) the log ratios for the different carbohydrates and hormones, from all organs (Fig. 6B; see Supplementary Fig. S3A at JXB online), from the root (Fig. 6C), the stem stele (Fig. 6D), the leaf blade (Fig. 6E), the leaf and the root (see Supplementary Fig. S3B at JXB online), and the stem cortex (see Supplementary Fig. S3C at JXB online); and (iii) the full LC-MS/MS datasets of the leaf water-soluble metabolite fractions (Fig. 6F). PCA plots showed clear separation of the 6 DAW sampling date from the other dates along PC1, accounting for most of the variance in the data (Fig. 6A–F). The plots show a progression of interrelated events defining three major groups: control (day 0), onset of WD (3 and 4 DAW), and established WD (6 DAW). Loadings show that the leaf Ψpd and ABA (in all organs but particularly in the root) are responsible for the differences in the physiological and metabolic changes in the early WD stages (Fig. 6A) along the first principal component (PC1), while IAA and Cks contribute most to the second principal component (PC2). For the carbohydrate and hormone ratios (Fig. 6B), the important variables for PC1 are the relative values of sucrose to ABA and raffinose to ABA in all organs and the hormone balance in the roots. When considering the organs individually (Fig. 6C–E), the variables that relate hormone balance to ABA (leaf), soluble sugars to ABA (stele, leaf), and soluble sugars to starch (root, stele) contribute most to PC1, accounting for 36–50% of the total variation. The relevance of soluble sugars in the leaf is reinforced when considering the LC-MS/MS full scans (Fig. 6F); several compounds contribute to the separation [with m/z values ∼103, 149, 387 (sucrose), 475, 549 (raffinose), and 728]. The identification of the unknown species is currently under investigation in order to characterize these components in detail.

Fig. 6.

Principal components analysis with variables: (A) the physiological and metabolic alterations on the early WD stages (SWC, Ψpd, RWC, gs, A, WUE, ABA, IAA, Cks, starch, glucose, sucrose, and raffinose); (B) the carbohydrate and hormone log ratios for the leaf blade, root, stem stele, and stem cortex; (C) the carbohydrate and hormone log ratios for the root; (D) the carbohydrate and hormone log ratios for the stem stele; (E) the carbohydrate and hormone log ratios for the leaf blade; (F) the full scan LC-MS data for the aqueous extracts of the different lupin organs. (Dark triangles) Control at day 0; (open circles) 3 d after withholding water; (open triangles) 4 d after withholding water; (open squares) 6 d after withholding water. (This figure is available in colour at JXB online.)

PCA analysis also shows that the various sampling dates can be separated using carbohydrate and hormone log ratios from all organs (Fig. 6B). The separation is due mostly to values from the leaf (Fig. 6E) and, to a lesser extent, from the root and stem stele (Fig. 6C, D). Figure 6B shows that differentiation between 3 DAW and 4 DAW can be achieved along PC2 where the separation is due to the relative values of: carbohydrates to IAA (leaf), IAA to bioactive Cks (leaf and root), carbohydrate balance (leaf, root, stele), and carbohydrate to ABA (root, stele). Clear separation of the different organs is seen when using log ratios as variables (see Supplementary Fig. S3 at JXB online). The relative levels of soluble sugars together with the variables that relate soluble sugars to starch are responsible for the clear separation of leaf from the other organs (see Supplementary Fig. S3A at JXB online); the leaf and the root are discriminated by the relative levels of soluble sugar to starch and carbohydrate to IAA (see Supplementary Fig. S3B at JXB online), and the stem components are separated by the level of soluble sugars in relation to starch and starch to ABA (see Supplementary Fig. S3C at JXB online).

Discussion

In order to detect and differentiate initial soil water deficit effects in the lupin plant (a reduction in the net CO₂ assimilation rate lower than 20%) metabolic analysis have been combined with the standard procedures for plant performance evaluation, i.e. plant water status and photosynthetic rate (Verslues et al., 2006; Pinheiro and Chaves, 2011). In particular, in a new approach, the full scan LC-MS/MS data recorded for the targeted sugar quantitation have been exploited and those data have been analysed further in an unbiased approach using PCA, in order to take advantage of any additional information hidden in those datasets. Small decreases in soil water content (∼12%) observed at 3 DAW are shown (i) not to affect the plant water status; (ii) to reduce stomatal aperture while not affecting photosynthesis; and (iii) to alter the phytophormone and carbohydrate balance in an organ specific manner. In contrast to IAA and Cks, at 3 DAW, ABA only slightly changes in the leaf blade and root. Strikingly, ABA content decreases in the vascular tissue. To our knowledge, just two other publications report a decrease in ABA content as a result of water deficit (Zholkevich and Pustovoitova, 1993; Pustovoitova et al., 2003). This effect of water deficit on ABA levels may result from the stress imposition rate and mode (slowly imposed soil water deficit). This ABA decrease that was observed in the stem vascular tissue may be related to ABA remobilization through the guard cells via the xylem, which would lead to ABA-dependent stomatal closure (Seki et al., 2007; Wilkinson and Davies, 2010). However, it is well established that stomatal aperture does not rely exclusively on the absolute ABA level (Hartung and Witt, 1968; Munns and King, 1988; Trejo and Davies, 1991; Zhang et al., 1997), but is also controlled by other hormones like auxins, ethylene, and Cks (Hartung and Witt, 1968; Tanaka et al., 2006; Wilkinson and Davies, 2010). So, the regulation of stomatal closure can be achieved through signal repression via the hormone balance between IAA, Cks, and ABA. This implies a tightly regulated process that comprises incoming signals and in which local synthesis and action (on the different plant organs) as well as long-distance signals are involved (Wolf et al., 1990; Sakakibara, 2006; Neumann, 2008; Robert and Friml, 2009). The effect of the initial soil water deficit on the hormone balance does not seem to be explained merely through long-distance transport effects, since the transpiration rate is not significantly affected until 6 DAW. Local effects on hormone synthesis/metabolism can be postulated, based on the observed increase in levels of the Ck deactivated forms (O- and N-glucosides), which accumulate under stress (Havlova et al., 2008).

Altered hormone balance will impact on the carbon status, dependent on stomatal regulation, CO₂ assimilation rate, and carbohydrate use and storage. Interactions between carbohydrate metabolism and ABA, auxin, Ck, and ethylene signalling networks are known and can act as connecting nodes for multiple pathways (Leon and Sheen, 2003; Gibson, 2004; Robertson et al., 2009; Cutler et al., 2010; Stitt et al., 2010), co-ordinating nutritional and environmental inputs. Our results show that metabolic alterations occur prior to changes in the plant water status (that is defined here as metabolic reorganization), implying that a mechanism for the distinction between short-term fluctuations and steadier alterations must be active. Metabolic reorganization also implies that the existing metabolic machinery responds and determines the adjustments of the metabolite status to the changing environments (Stitt et al., 2010), namely the use and storage of carbohydrates (Sakakibara, 2006; Albacete et al., 2008; Smeekens et al., 2010). The transient metabolic alterations found in L.albus organs, for example, starch accumulation in the leaf blade, can be related to short-term responses to a small decrease in the soil water content. Transient responses can act as a switching node, the achievement of several thresholds being necessary to activate the mechanisms that lead to a larger and long-standing soil water deficit. The nature of the sensor(s) and signalling pathways has been the subject of intense research (Pinheiro and Chaves, 2011), the most direct candidates for signal metabolites being sugars, such as sucrose, glucose, and trehalose (Gibson, 2004; Smith and Stitt, 2007; Bolouri-Moghaddam et al., 2010; Stitt et al., 2010). Our analysis demonstrates that the carbohydrate balance and relation to hormone levels are relevant to the earliest responses to soil water deficit, that is, in the leaf. The relative values of the different carbohydrates analysed (glucose, sucrose, raffinose, and starch), their ratio to IAA, and the IAA and bioactive Ck balance can discriminate 3 DAW from the 4 DAW sampling dates, providing the initial responses to a slight decrease in soil water availability (10–15% decrease). Further alterations in the sucrose to ABA and in raffinose to ABA relative values (in all organs) indicate that soil water availability continues to decrease. It is concluded that the balances for sucrose, raffinose, ABA, and IAA seem to be metabolically relevant in triggering the adjustment of stress response mechanisms. The role of raffinose does not seem to be compatible with osmoprotection. Although readily responsive to small variations in the soil water content, raffinose is detected in very small amounts (100-fold lower than sucrose and 1000-fold lower than glucose). Further, raffinose accumulation in lupin is higher on initial stress and early recovery than under severe stress (Antonio et al., 2008), which also seems to exclude osmolyte functions. Raffinose accumulation can be related to reactive oxygen species (ROS) signalling and the antioxidant response (Nishizawa-Yokoi et al., 2008; Bolouri-Moghaddam et al., 2010), which occurs independently of the ABA signalling pathways (Urano et al., 2009). Such a signalling role and the interaction between hormone, sugar, and ROS pathways in the responses to environmental conditions require further investigation (Wingler and Roitsch, 2008; Pinheiro and Chaves, 2011). Further studies are also necessary in order to determine the eventual physiological relevance of the detected co-occurrences, to understand how progressive water deficit modulates carbon sensing and signalling (Muller et al., 2011) and to understand the impact of metabolic reorganization under changing environments on plant performance.

Abbreviations

Abbreviations
 
• ABA

  absicic acid

 
• Cks

  cytokinins

 
• DAW

  days after withholding water

 
• DHZ

  dihydrozeatin

 
• DHZ9G

  dihydrozeatin 9-glucoside

 
• DHZ9R

  dihydrozeatin 9-riboside

 
• DHZOG

  dihydrozeatin-O-glucoside

 
• IAA

  indole-3-acetic acid

 
• iP

  N⁶-(Δ²-isopentenyl)adenine

 
• iP7G

  N⁶-(Δ²-isopentenyl)adenine 7-glucoside

 
• iP9G

  N⁶-(Δ²-isopentenyl)adenine 9-glucoside

 
• iP9R

  N⁶-(Δ²-isopentenyl)adeninosine

 
• MS

  mass spectrometry

 
• MVA

  multivariate analysis

 
• PAR

  photosynthetically active radiation

 
• PCA

  principal components analysis

 
• RW

  rewatering

 
• WD

  water deficit

 
• WW

  well watered

 
• Z

  trans-zeatin

 
• Z7G

  trans-zeatin-7-glucoside

 
• Z9G

  trans-zeatin-9-glucoside

 
• Z9R

  trans- zeatin 9-riboside

 
• Z9ROG

  trans-zeatin 9-riboside O-glucoside

 
• ZOG

  trans-zeatin O-β-glucoside

The authors wish to acknowledge Margarida Oliveira (ITQB) for helpful discussions. CP acknowledges the fellowship SFRH/BPD/14535/2003 from Fundação para a Ciência e Tecnologia (Portugal); CA acknowledges the CHEMCELL Marie Curie Early Stage Research Training Fellowship of the European Community's Sixth Framework Programme (MEST-CT-2004-504345); MF Ortuño acknowledges a Postdoctoral research fellowship from Ministerio de Educación y Ciencia (Spain). This work was partially supported by the Treaty of Windsor (Anglo-Portuguese Joint Research Programme, B-18/08).

Author notes

†

Present address: Max-Planck-Institut für Molekulare Pflanzenphysiologie, Wissenschaftspark Golm, Am Mühlenberg 1, D-14476 Potsdam, Golm, Germany.