Metabolomic and proteomic changes in the xylem sap of maize under drought


D. P. Schachtman. Fax: 314-587-1521; e-mail:


Plants produce compounds in roots that are transported to shoots via the xylem sap. Some of these compounds are vital for signalling and adaptation to environmental stress such as drought. In this study, we screened the xylem sap using mass spectrometry to quantify the changes in new and previously identified sap constituents under extended drought. We detected and quantified the changes in the concentration of 31 compounds present in the xylem sap under progressively increasing drought stress. We found changes in the hormones abscisic acid (ABA) and cytokinin, and the presence of high concentrations of the aromatic cytokinin 6-benzylaminopurine (BAP). Several phenylpropanoid compounds (coumaric, caffeic and ferulic acids) were found in xylem sap. The concentrations of some of these phenylpropanoid compounds changed under drought. In parallel, an analysis of the xylem sap proteome was conducted. We found a higher abundance of cationic peroxidases, which with the increase in phenylpropanoids may lead to a reduction in lignin biosynthesis in the xylem vessels and could induce cell wall stiffening. The application of new methodologies provides insights into the range of compounds in sap and how alterations in composition may lead to changes in development and signalling during adaptation to drought.


The xylem in plants is the main conduit for water and minerals from roots to shoots. It is composed of cell wall material with specific properties that facilitate the movement of water through the plant. Water is driven from roots to shoots either through root pressure or a force that is set up by the evaporation of water from stomates on the leaf surface. The xylem sap in plants is generally thought to be comprised mainly of water and minerals. However, recent studies have revealed a wide array of compounds that are found in sap include amino acids, organic acids (Gollan, Schurr & Schulze 1992; Senden et al. 1992; Schurr & Schulze 1995), plant hormones and their metabolites (Bano et al. 1994; Sauter, Dietz & Hartung 2002), polyamines (Friedman, Levin & Altman 1986) and proteins (Biles & Abeles 1991; Kehr, Buhtz & Giavalisco 2005; Alvarez et al. 2006).

The characterization of changes in sap composition is important to plant growth and development because under abiotic stress (i.e. water deficit), it is well-known that signals travel via the xylem from roots to shoots that reduce plant growth and transpiration (Wilkinson & Davies 2002; Davies, Kudoyarova & Hartung 2005). The identity of drought-induced root signals is still somewhat controversial (Munns & King 1988; Holbrook et al. 2002) and may comprise a number of different compounds. However, the role of abscisic acid (ABA) as a root-sourced, long-distance signal that causes stomatal closure under conditions of water stress is a dominant factor (Davies et al. 2005) and has been shown to be active both on the whole-leaf level and on individual guard cells (Schroeder, Kwak & Allen 2001). Other plant hormones such as cytokinins also act as root-to-shoot signals (Shashidhar, Prasad & Sudharshan 1996). Studies show that the application to the leaf epidermis of trans-zeatin (Z), trans-zeatin riboside (ZR), 6-benzyladenine (BA) and other cytokinins reverse ABA-induced stomatal closure (Pospisilova 2003). It has been proposed that drought-induced stomatal closure is the result of interactions among plant hormones, chiefly the ABA/ZR ratio (Dodd 2003). Interactions between hormones highlight the importance of the quantification of multiple small molecules in studies on how the xylem sap contributes to the modulation of whole plant processes.

Previous studies that have quantified multiple compounds in xylem sap have been mainly restricted to the changes in cations, anions, amino acids, organic acids and ABA in the xylem sap under water-stressed (WS) conditions of sunflower (Gollan et al. 1992; Schurr, Gollan & Schulze 1992), Ricinus communis and maize (Bahrun et al. 2002; Goodger et al. 2005). From these studies, only malate (mal) and perhaps some inorganic ions have been implicated in root-to-shoot communication (Gollan et al. 1992; Patonnier, Peltier & Marigo 1999; Goodger et al. 2005). A more comprehensive screening of the small molecule content in xylem sap and the quantification of changes in multiple small molecules in response to soil drying has not been reported.

Several recent proteomic studies have shown that multiple proteins are found in the xylem sap from different plant species (Biles & Abeles 1991; Buhtz et al. 2004; Kehr et al. 2005; Alvarez et al. 2006). Changes in the quantity of individual xylem sap proteins have been documented in response to pathogen infection. Xylem sap proteins change in tomato in response to infection by pathogenic fungi (Rep et al. 2002, 2003). In rice, a cationic peroxidase specifically accumulates in xylem vessels after an incompatible interaction with the vascular bacterial pathogen Xanthomonas oryzae (Young et al. 1995). The multiple defence proteins found in maize xylem sap under conditions where pathogens are not present were found to have antifungal activity (Alvarez et al. 2006). Little is known about how the protein composition of the xylem sap changes in response to abiotic stresses such as soil drying, and whether such changes may contribute to root-to-shoot communication or may provide clues to changes in xylem structure and function under stress.

In addition to proteins in the xylem sap, small peptides may be involved in root-to-shoot signalling. Small peptides provide an important signalling mechanism in both plants and animals (Ryan et al. 2002; Matsubayashi & Sakagami 2006). Several peptides have been identified in plants (Ryan et al. 2002) and have been shown to mediate signalling events during plant–pathogen interactions (Pearce & Ryan 2003), cell division (Matsubayashi & Sakagami 2006) and anther–stigma interactions. Plant natriuretic peptides extracted from ivy have been shown to modulate stomatal aperture via the up-regulation of guanylate cyclase (Pharmawati et al. 2001). It is not known whether peptides are present in the xylem sap, but bioactive peptides are known to move systemically in the phloem of plants (Ryan & Pearce 1998) and could be another way in which roots communicate to leaves under soil water deficit.

In most studies on xylem sap composition, only a small number of compounds have been characterized simultaneously. New methods for small molecule and protein characterization now make it possible to take a more comprehensive approach to studying how compounds change in xylem sap in response to environmental perturbations. To document the complex nature of the xylem sap, we screened for the presence of 61 small molecules. Following the screening, we asked the question of how the small molecules and the proteins in xylem sap change under drought stress in order to gain more insight into the overall response to drought and the possible signals being generated by roots that are transported to shoots.


Plant material

The experimental procedure used to grow maize plants for xylem sap collection has been previously described (Goodger et al. 2005). We used Zea mays cv. FR697 in these experiments. Plants were watered daily with Miracid Professional 21-7-7 Acid Special (The Scotts Company, Marysville, WA, USA) supplemented with 5 mm KCl. Supplemental iron was initially supplied to the soil by addition of 20 mL 35 mm Sprint 330 Iron chelate (Becker Underwood, Inc., Ames, IA, USA) 7 d after sowing (DAS). Plants were then sprayed with 82 µm ferric citrate (Sigma-Aldrich, Inc., St Louis, MO, USA) 9, 12, 13, 14 and 15 DAS. Plants were grown in a controlled-environment chamber with a day/night temperature of 26/18 °C, RH of 45%, with a photoperiod of 16 h and a light intensity of 815 µmol m−2 s−1. Water was withheld from half of the pots from 15 DAS. Sap was extracted from both well-watered (WW) and WS plants 7, 10 and 12 d after water was withheld (22, 25, 27 DAS, respectively).


l-Serine, l-proline, l-valine, l-threonine, l-cysteine, l-pyroglutamate, l-leucine, l-isoleucine, l-asparagine, l-ornithine, l-glutamate, l-methionine, l-histidine, l-phenylalanine, l-arginine, l-citrulline, l-tyrosine, l-tryptophan, l-succinate, l-ascorbate, 1-naphtylacetic acid, kinetin, trans-cinnamic acid, caffeic acid, p-coumaric acid, adenine, spermine and spermidine were purchased from Sigma; Z, ZR, 6-benzylaminopurine (BAP) and n-benzyl-9-(2-tetrahydropyranyl)-adenine were from RPI Corp. (Mt. Prospect, IL, USA); l-citric acid, l-aspartate and l-lysine were from Fisher Chemical (Fairlawn, NJ, USA). Brassinolide, ABA glycosyl ester and o-glucosylzeatin were purchased from OlChemlm Ltd. (Olomouc, Czech Republic); l-malic acid was from Spectrum (Gardena, CA, USA); indole-3-acetic acid was from Invitrogen (Carlsbad, CA, USA); and ferulic acid was from Fluka (Buchs, Switzerland). ABA, 3-indolebutyric acid, gibberellic acid and l-glutamine were from Phytotechnologies Laboratories (Shawnee Mission, KS, USA). All the flavonoids were purchased from Indofine Chemical Co. (Hillsborough, NJ, USA). Phaseic acid (PA) not commercially available was prepared from seeds of Robinia pseudacacia (Hirai, Fukui & Koshimizu 1978), and kindly provided by Dr Hirai (Kyoto University, Japan). All chemicals used were of analytical or reagent grade.

Plant water status

Leaf conductance (gs) was measured on the abaxial surface of leaf 4 (avoiding major veins) 6 h into the photoperiod using an AP4 steady-state porometer (Delta-T Devices, Cambridge, UK). At the times of sap harvest, the xylem pressure potential of leaf 4 from each plant was determined using a Scholander pressure chamber (Soil Moisture Equipment Corp., Santa Barbara, CA, USA).

Xylem sap extraction

Plants were de-topped near the top of the exposed mesocotyl, and the pots containing the roots and mesocotyl were placed in custom-built pressure vessels (CHPT Manufacturing, Georgetown, DE, USA). The mesocotyl was rinsed, the initial sap discarded and the silicon tubing was fitted over the mesocotyl. The rate of exudation from the mesocotyl was increased by application of pressure to the root using compressed air (to approximately 3.0–10.0 bars), and sap was collected rapidly for 15–20 min. To ensure that the composition of xylem sap reflected what was flowing in the intact plants, pressure was applied to induce sap flow at a rate that approximated the opposite of the average measured xylem pressure potential at the time of collection. The pressurization of roots to a positive value equal to the negative of leaf water or xylem pressure potential has been shown to be a good guide for approximating sap flow rates. For further discussion and data on how flow rates change sap concentrations in maize, see Goodger et al. (2005).

All sap samples were frozen immediately after sampling and stored at −80 °C until analysed. Three biological replicates of xylem sap from each treatment (WW and WS plants) at each sampling date (7, 10 and 12 after water withholding) were collected. The sap for the proteomic studies and the metabolite profiling were collected during the course of different experiments. The plant growth and rate of soil drying were very uniform between the different experiments because the same soils and experimental procedures were used. At day 7, sap was collected from three to six plants per replicate from WW and WS; at day 10, sap was collected from four to six plants per replicate from WW and WS plants and at day 12 sap was collected from six to eight plants for WW and WS plants. Three biological replicates were collected. Sap was collected and frozen immediately upon exudation. After collection, the sap was thawed on ice, and proteinase inhibitor [Complete, ethylenediaminetetraacetic acid (EDTA)-free Protease Inhibitor Cocktail; Roche Diagnostics, Penzberg, Germany] was added to the xylem sap in order to analyse the protein and amino acid contents of the xylem sap.

Protein separation and analysis

The protein purification from xylem sap has been described previously (Alvarez et al. 2006). Protein concentrations in purified samples were determined using a RediPlate EZQ Protein Quantitation Kit (Molecular Probes, Eugene, OR, USA). The proteins were separated by two-dimensional (2-D) electrophoresis (2-DE) and stained using SYPRO Ruby (Molecular Probes) (Zhu et al. 2006). Protein spots were visualized using the TYPHOON 9410 system (Amersham Biosciences, Piscataway, NJ, USA). Gels containing three biological replicate samples from each treatment (WW and WS plants) and each sampling date (7, 10 and 12 after water withholding) were analysed with Phoretix 2D Evolution Software (Nonlinear Dynamics, Durham, NC, USA) (Zhu et al. 2006). The protein identification procedure using liquid chromatography–mass spectrometry/mass spectrometry (LC–MS/MS) has been previously described (Alvarez et al. 2006). The peptide tandem mass spectra were searched against the MAIZE EST database ( using MASCOT search engine ( The MAIZESEQ EST database contains maize EST and full-length cDNA sequences from Ceres, DuPont/Pioneer and Monsanto. Unambiguous identification was judged by the number of peptide sequence tags, sequence coverage, Mascot score and the quality of tandem MS spectra.

Stock and working standard solutions

Stock standard solutions of the 64 compounds analysed in the xylem sap were prepared by dissolving the appropriate amount of the compounds in solvent A (0.1% formic acid in water). These solutions were stored in the dark at −20 °C. Just prior to injection, a series of aqueous working standard solutions were prepared by serial dilution of the stock standard solution in solvent A to reach a range of concentrations from 1 fmol µL−1 to 400 pmol µL−1 according to the separation and ionization capacity of the compound to quantify the compound in the xylem sap.

Small molecule preparation and analysis

After collection, the xylem sap was lyophilized. Just prior to the LC–MS/MS analysis, two batches of samples were prepared from the three biological replicates from each treatment (WW and WS plants) and each sampling date (7, 10 and 12 d after water withholding). One batch of xylem sap, with ∼4 mg dry weight completely dissolved in 500 µL of solvent A was used to quantify compounds of low concentrations in the xylem sap. Another batch of sap, with 1 mg of dry sap completely dissolved in 15 mL of solvent A, was used to quantify the more highly concentrated compounds.

The small molecule identification and quantification in the xylem sap were performed using LC–MS/MS. The high-performance liquid chromatography (HPLC) (Prominence; Schimadzu, Columbia, MD, USA) used an autosampler (Leap CTL PAL; Leap technologies, Carrboro, NC, USA) to automatically inject the samples. A C18-RP (Gemini 5µ; 150 × 2.0 mm, Phenomenex, Torrance, CA, USA) analytical column was used, with a C18 guard cartridge to prolong its lifetime. The mobile phase consisted of two solvents: 0.1% formic acid in water for solvent A and 0.1% formic acid in ACN for solvent B. Gradients were set up according to the nature of the compounds analysed. To better resolve the hydrophilic compounds (methods 1 and 2), we used a shallow gradient starting at 0–6% B for 5 min and then 6–20% B for 10 min followed by a re-equilibration of the column for 5 min at a flow rate of 0.25 mL min−1 of solvent A. For the hydrophobic compounds (methods 3 and 4), a gradient from 5 to 80% B for 25 min followed by 5 min of re-equilibration at a flow rate of 0.5 mL min−1 was used. Chromatography was performed at room temperature. Because of the higher concentration of hydrophilic than hydrophobic compounds in the sap, a 20 µL sample was injected for analysis of the hydrophilic compounds (methods 1 and 2) and 50 µL for the hydrophobic compounds (methods 3 and 4).

The HPLC was coupled to a hybrid triple quadrupole-ion trap (4000 Q-TRAP, Applied Biosystems, Foster City, CA, USA) mass spectrometer equipped with a TurboIonSpray (TIS) interface operated in the positive ion mode. The electrospray ionization (ESI) parameters for the hydrophilic and the hydrophobic compounds ionization were: curtain gas (CUR) at 25 and 30 psi, the ion source (IS) at 4500 and 5500 V, nebulizer gas (GS1) at 25 and 50 psi, TIS gas (GS2) at 20 and 55 psi and the TIS probe temperature at 100 and 500 °C, respectively. To optimize analyte sensitivity, individual working solutions having the characteristic MRM transition with the best declustering potential (DP) and collision energy (CE) were chosen for each compound (see database Each transition was performed with a 30 or 40 ms dwell time to get a scan time around 1.7 s for all the transitions analysed using a method including an enhanced mass spectrometry (EMS) and an enhanced product ion (EPI) scans before switching to MRM mode.

Statistical analysis

To compare the compound concentrations between the control and treatment across the range of water stress and to determine significance of the variation in protein abundance between WW and WS plants at 7, 10 and 12 d of water stress, t-tests were performed.

Rapid alkalinization bioassay

We measured the change in pH of external medium in which cell suspensions were growing induced by fractions of xylem sap in order to identify potential signalling compounds in those fractions. The peptide interaction with the cell membranes as measured by changes in the external pH is a method previously used to identify signals involved in the biotic stress responses (Pearce et al. 2001). Black Mexican Sweet (BMS) corn suspension cells were maintained in a Murashige and Skoog basal media (containing 10 mg L−1 thiamine, 100 mg L−1 myo-inositol, 1.16 g L−1l-proline, 1 g L−1 casein hydrolysate, 3% sucrose, 1 mg L−1 2,4-D) and adjusted to pH of 5.8 with KOH. Cultures were grown in the dark at 28 °C with shaking at 160 rpm. Suspension cultures were maintained by transferring 5 mL of cells to a 250 mL Erlynmeyer flask containing 50 mL of fresh media every 7 d. Suspension cells were used in a bioassay 5 or 6 d after transfer.

For these bioassays, 1 mL aliquots of cells was transferred to the wells of a 24-well sterile cell culture plate. The plate was placed on an Eppendorf (Westbury, NY, USA) Thermomixer R, and the cells were allowed to equilibrate for 1 h at 28 °C, shaking at 450 rpm. After 1 h, the protein and peptide fractions were added to the cells, and the pH was measured after 5, 15, 30, 45 and 60 min, using a MI-410 Combination pH Electrode (Microelectrodes, Inc., Bedford, NH, USA). Rapid alkalinization factor (RALF), a 5 kDa polypeptide which induces rapid alkalinization of culture medium of tobacco suspension cells was used as a positive control (Pearce et al. 2001). The lyophilized peptide fraction was resuspended in 0.1% trifluoroacetic acid (TFA), and 0.25 nmol (1 µL) was added to 1 mL of BMS cells in the bioassay. TFA (10 µL of 0.1%) was added to 1 mL of suspension cells as a negative control.

Sap lyophilized to approximately half the original volume was thawed and filtered through 0.2 µm cellulose acetate filters, and concentrated on Amicon 15 mL Ultra Centrifugal Filter Devices with a 5 kDa MWCO (Millipore, Bedford, MA, USA) in order to separate the protein and peptide fractions, which were applied to the cell suspensions in separate bioassays. The amount of protein in 1 µL was 1.2 µg for well watered, and 2.3 µg for water stressed as determined using BCA protein quantification (Pierce Biotechnology, Rockford, IL, USA). The protein fraction used for these analyses was collected from plants that were at an early stage of drought because at this stage, the root signals are dominant over possible hydraulic signals (7 d after withholding water). The small peptide fraction was freeze-dried and resuspended in 0.1% TFA, and 10 µL was added to 1 mL of cells. One microlitre of the concentrated protein fraction, from both WW and WS plants, was used in the bioassay. Pronase (Calbiochem, EMD Biosciences, Inc., San Diego, CA, USA) solution was prepared as a 10 µg mL−1 solution in a 0.1 m phosphate buffer, pH 7.4, and heated at 40 °C for 5 min before use. Equal volumes of pronase solution and samples were mixed together, and the samples were incubated at 40 °C for 25 min. In order not to exceed a volume of 10 µL sample added per 1 mL of cells in the bioassays, the peptide fraction was further dried down and concentrated prior to treatment with pronase, so the final volume added to the cells was 6 µL (3 µL of peptide fraction plus 3 µL pronase). RALF treated with pronase or with the phosphate buffer alone, served as the negative control and positive control, respectively. Two microlitres were added to 1 mL of cells for each control.


Plant water status

WS plants had significantly lower leaf conductance 7–8 d after watering was stopped (Fig. 1a), which indicated that water stress was being sensed by plants (Goodger et al. 2005). Figure 1a shows the leaf conductance and xylem pressure potentials from the experiment from which sap was collected for the metabolite profiling experiment. For the proteomic experiments, the only difference in response was that the stomatal conductance in the drought-stressed plants was significantly different starting at day 7. The xylem pressure potential measured at day 7 of drought-stressed plants was not significantly different from that of control in both experiments (Fig. 1b), confirming that days 7 and 8 after withholding water were early stages of water stress before hydraulic signals may play a role.

Figure 1.

Leaf conductance (a) and xylem pressure potential (b) for well-watered (WW) and water-stressed (WS) plants. Water was withheld in the WS treatment from day 0 onwards and xylem sap was extracted from both WW and WS plants at days 7, 10 and 12. Values are the mean ± standard error (SE) of 15 plants for each treatment and each day. Significant differences between treatments are indicated as **P < 0.01 and ***P < 0.001.

Small molecule database and MRM validation

There is no publicly available library for LC–ESI–MS/MS fragmentation of plant compounds; therefore, the first step in the identification of the compounds in xylem sap was to set up a library of fragmentation patterns for compounds that were or might be found in the xylem sap. Standards for 86 compounds, including amino acids; organic acids; plant hormones such as ABA, cytokinins and auxins; and flavonoids such as chalcones, flavones and phenylpropanoid were analysed. Ionization and fragmentation parameters DP and CE were optimized for each standard in order to operate in MRM mode. This feature provides the maximum sensitivity and selectivity for detection and quantification of targeted compounds. The small molecule library is available at This library provides the details about the MS/MS fragmentation and ionization parameters, and MRM transitions for small molecule identification and quantification (see Supplementary Fig. S1 for a database preview). A list of the 86 compounds is provided as supplementary data to this report (Supplementary Table S1).

From these standard compounds, we narrowed down the analysis of the xylem sap to 64 compounds. The identification and the quantification of the small molecules in the xylem sap were then performed by using specific MRM transitions to identify selected compounds. Sixty one compounds were quantified, because for certain amino acids leu and ile, glu and gln, asp and asn, the elution times and molecular weights were too close to differentiate between them with an accuracy of only ±0.7 Da (Supplementary Table S2). In addition for these amino acids, the MS/MS fragmentation patterns were similar preventing the use of different daughter ions to monitor them by MRM mode. Four different methods were used to resolve the hydrophilic compounds (methods 1 and 2) from the hydrophobic compounds (methods 3 and 4) (Supplementary Table S2). To maintain high MRM sensitivity, the dwell time for each transition was set at 30 or 40 ms. The scan times were less than 1.7 s. This provided to high resolution peaks for analysis (approximately 20 measurements per peak).

Standard curves for each compound were constructed in different concentration ranges to test the detection limit and to ensure that the signal intensities were not saturating. For the lowest concentration range, 20 µL sample injections were used in methods 1 and 2, and for 50 µL sample injections in methods 3 and 4 are shown in Supplementary Table S2 with R2 for a linear fit for the concentration range used. These standards curves were used to quantify the small molecules in the xylem sap.

Identification and quantification in the xylem sap

We detected only 31 of the 61 compounds analysed in the xylem sap of maize (Table 1). We detected all the amino acids, organic acids and phenylpropanoids analysed, but none of the chalcones, flavones or flavonols, and only naringenin (nar) among the flavonones. Among the detected plant hormones, we only found ABA and its catabolite PA, Z, its conjugate ZR) and the aromatic cytokinin BAP despite efforts to increase the sensitivity of detection.

Table 1.  Concentration of the identified compounds of maize xylem sap in well-watered (WW) and water-stressed (WS) plants after 7, 10 and 12 d of water stress
 Day 7 WWDay 7 WSDay 10 WWDay 10 WSDay 12 WWDay 12 WS
  1. The percentage of each compound calculated from the total compounds in mm is also indicated. The compounds were grouped in major compounds with a mm concentration range and expressed in percent (%) of total compounds, and the minor compounds with a nm concentration range and per mil (‰) which provides an estimate of the relative proportions. Significant differences between WW and WS are indicated as *P < 0.05 and **P < 0.01. Abbreviations used in the first column are in Supplementary Table S1. ND, not detected.

Major (mm)
 glu/gln3.560 ± 0.18576.79%4.419 ± 0.578*51.76%4.264 ± 0.70552.35%7.198 ± 0.94246.52%3.280 ± 0.14254.24%0.4677 ± 1.15328.86%
 ser0.319 ± 0.0136.88%0.506 ± 0.042**5.92%0.480 ± 0.1075.89%0.845 ± 0.128*5.46%0.401 ± 0.0156.63%1.353 ± 0.364*8.35%
 thr0.174 ± 0.0083.76%0.335 ± 0.044*3.92%0.238 ± 0.0612.92%0.438 ± 0.0742.83%0.228 ± 0.0063.78%0.631 ± 0.172*3.89%
 val0.169 ± 0.0273.65%0.163 ± 0.0571.91%0.175 ± 0.0432.15%0.419 ± 0.044*2.71%0.198 ± 0.0933.28%0.364 ± 0.1392.24%
 mal0.064 ± 0.0271.38%2.197 ± 0.422**25.73%2.256 ± 1.09227.69%5.228 ± 1.28733.78%1.345 ± 0.08022.25%7.138 ± 1.391**44.04%
 his0.060 ± 0.0171.30%0.133 ± 0.021*1.55%0.092 ± 0.0211.13%0.156 ± 0.0261.01%0.072 ± 0.0031.20%0.197 ± 0.051*1.21%
 lys0.052 ± 0.0091.13%0.148 ± 0.037*1.73%0.069 ± 0.0210.85%0.141 ± 0.0490.91%0.083 ± 0.0181.38%0.198 ± 0.0561.22%
 arg0.043 ± 0.0130.93%0.122 ± 0.013**1.43%0.072 ± 0.0120.89%0.125 ± 0.0660.81%0.042 ± 0.0070.70%0.177 ± 0.053*1.09%
 leu/ile0.040 ± 0.0040.86%0.082 ± 0.0250.96%0.079 ± 0.0260.97%0.164 ± 0.0411.06%0.058 ± 0.0030.96%0.389 ± 0.110*2.40%
 cit0.035 ± 0.0040.75%0.176 ± 0.048*2.06%0.104 ± 0.0481.27%0.178 ± 0.0441.15%0.081 ± 0.0151.34%0.089 ± 0.0200.55%
 asp/asn0.023 ± 0.0020.50%0.033 ± 0.003*0.38%0.037 ± 0.0090.45%0.067 ± 0.0120.44%0.040 ± 0.0010.66%0.102 ± 0.027*0.63%
 met0.019 ± 0.0010.41%0.030 ± 0.004*0.34%0.025 ± 0.0070.31%0.036 ± 0.0060.23%0.022 ± 0.0020.37%0.033 ± 0.0100.20%
 phe0.017 ± 0.0020.36%0.022 ± 0.0060.26%0.021 ± 0.0060.26%0.033 ± 0.0060.21%0.021 ± 0.0010.34%0.040 ± 0.0130.25%
 tyr0.015 ± 0.0010.33%0.019 ± 0.0030.23%0.027 ± 0.0070.33%0.057 ± 0.0140.37%0.022 ± 0.0020.36%0.136 ± 0.049*0.84%
 orn0.014 ± 0.0030.30%0.008 ± 0.0010.10%0.008 ± 0.0010.10%0.014 ± 0.002*0.09%0.010 ± 0.0020.16%0.015 ± 0.0050.09%
 pglu0.012 ± 0.0050.27%0.016 ± 0.0080.18%0.015 ± 0.0080.19%0.027 ± 0.0120.17%0.012 ± 0.0040.19%0.014 ± 0.0060.09%
 pro0.009 ± 0.0010.20%0.021 ± 0.0060.25%0.022 ± 0.0090.27%0.035 ± 0.0090.22%0.017 ± 0.0020.28%0.153 ± 0.065*0.94%
 suc0.004 ± 0.0020.09%0.098 ± 0.014**1.15%0.151 ± 0.0821.85%0.287 ± 0.0761.85%0.106 ± 0.0171.75%0.446 ± 0.135*2.75%
 trp0.003 ± 0.001*0.07%0.008 ± 0.0020.10%0.009 ± 0.0030.11%0.023 ± 0.0070.15%0.005 ± 0.0010.09%0.055 ± 0.013*0.34%
Minor (nm)
 sper423.8 ± 203.49.14‰133.8 ± 55.41.57‰47.8 ± 19.60.59‰41.5 ± 12.80.27‰19.4 ± 4.00.32‰17.9 ± 3.30.11‰
 citrul407.3 ± 64.48.79‰358.7 ± 28.14.19‰461.6 ± 89.85.67‰669.1 ± 50.44.32‰424.5 ± 88.47.02‰561.2 ± 174.53.46‰
 cys404.3 ± 54.88.72‰918.1 ± 108.1**10.75‰144.0 ± 22.71.77‰157.5 ± 37.31.02‰ND70.2 ± 35.10.43‰
 fer220.1 ± 29.14.75‰474.4 ± 154.25.56‰613.7 ± 247.47.53‰560.4 ± 2.73.62‰546.9 ± 51.29.04‰310.6 ± 37.1*1.92‰
 cou30.7 ± 6.10.66‰64.4 ± 18.80.75‰82.9 ± 13.41.02‰131.0 ± 7.8*0.85‰59.3 ± 2.40.98‰170.1 ± 45.9*1.05‰
 caf14.7 ± 6.60.32‰47.5 ± 21.90.56‰58.2 ± 24.40.71‰223.0 ± 28.2**1.44‰32.6 ± 8.70.54‰291.7 ± 106.0*1.80‰
 ZR3.2 ± 0.10.07‰2.6 ± 1.00.03‰4.0 ± 1.10.05‰4.3 ± 0.70.03‰3.9 ± 0.40.06‰1.3 ± 0.4**>0.01‰
 BAP2.0 ± 0.20.04‰4.9 ± 1.70.06‰7.2 ± 1.50.09‰12.3 ± 0.2*0.08‰3.0 ± 0.10.05‰14.3 ± 2.9**0.09‰
 ABA0.75 ± 0.210.02‰20.6 ± 8.3*0.24‰60.2 ± 17.10.74‰109.9 ± 7.3*0.71‰0.94 ± 0.350.01‰217.7 ± 65.3*1.34‰
 nar0.56 ± 0.510.01‰0.13 ± 0.04>0.01‰0.29 ± 0.23>0.01‰0.20 ± 0.14>0.01‰0.13 ± 0.05>0.01‰0.60 ± 0.27>0.01‰
 Z0.52 ± 0.010.01‰0.21 ± 0.07**>0.01‰0.86 ± 0.200.01‰0.82 ± 0.30>0.01‰0.60 ± 0.11>0.01‰0.24 ± 0.06*>0.01‰
 PAND0.04 ± 0.02*>0.01‰0.05 ± 0.02>0.01‰0.12 ± 0.01*>0.01‰ND0.15 ± 0.03**>0.01‰
Total (mm)4.636 ± 0.244 8.538 ± 1.181* 8.145 ± 2.097 15.474 ± 2.737* 6.046 ± 0.065 16.207 ± 3.779* 

The imposition and progression of drought brought about some expected as well as previously unreported changes in the metabolite concentrations in xylem sap. The total amount of the compounds that we quantified (excluding mineral ions which are a significant component of the xylem sap) increased over time and was higher in WS plants than in WW plants (Table 1). Approximately 50% of the total mass of the dried xylem sap was accounted for by the compounds listed in Table 1. The remainder of the xylem sap is composed of inorganic ions (Schurr & Schulze 1995; Goodger et al. 2005) and other molecules including proteins and peptides (Kehr et al. 2005; Alvarez et al. 2006).

Among the compounds that we quantified in this study, the amino acids glu/gln, serine (ser) and threonine (thr) were the most abundant (Table 1). In both drought-stressed plants and at days 10 and 12 in WW plants, malate was also present in abundance compared to other compounds. It is interesting to note that ferulic acid (fer) was found to be relatively abundant in the xylem sap in comparison to caffeic (caf) and coumaric acids (cou) (Table 1).

More significant changes in xylem sap metabolite content were measured 12 d after watering was stopped; however, many compounds were found in increased concentrations in WS plants even as early as 7 d after watering ceased. Some trends (e.g. ABA and thr) were consistent over the period of the drought but at day 10, differences were not significant (Table 1 and Fig. 2). A conjugated form of cytokinin ZR was found in much greater abundance than Z and decreased in the xylem sap of drought-stressed plants with the decrease being significant only at day 12 (Table 1 and Fig. 2). We also found BAP, which is another cytokinin, in relatively high abundance in the xylem sap. BAP was significantly higher in sap from drought-stressed plants under conditions of severe drought at days 10 and 12 (Fig. 2). Of the organic acids, malate was present in abundance and was significantly higher in drought-stressed plants at days 7 and 12 (Fig. 2). Amino acids mainly increased under drought stress at days 7 and 12 (Table 1 and Fig. 2).

Figure 2.

Fold change (log2) of selected metabolites found in xylem sap and subject to change after 7, 10 and 12 d of water stress. Abbreviations are found in Supplementary Table S1. Changes less than zero indicate a decrease, and greater than zero indicate an increase. Arrows show a qualitative increase in a particular metabolite which was not detected under control conditions, but present under water stress.

Differential protein expression analysis in response to water stress

Image analysis identified 39 proteins that were differentially expressed in the xylem sap of WW and WS plants (see examples in Fig. 3). Spots 273, 636 and 341 decreased in abundance in WS plants after 12 d, while spots 705, 758 and 759 increased in abundance (Fig. 3). Proteins were classified according to their biological function in Table 2. We identified 32 of the 39 protein spots that were found to be more or less abundant under water stress conditions. Six spots were found to contain two different proteins (spots 190, 211, 229, 266, 284 and 720). The image analysis data and the protein identification are available in the proteomic database ProticDB: This database is a modified version of PROTICdb v-1.2 (Ferry-Dumazet et al. 2005).

Figure 3.

Sections of two-dimensional (2-D) gels from well-watered (WW) and water-stressed (WS) plants showing the protein abundance variation after 7, 10 and 12 d for spots 273, 341, 636, 705, 758 and 759. For spot identification, see Table 2.

Table 2.  Protein name and fold change of the proteins identified as being significantly regulated under water stress classified by their biological function (CWM, cell wall metabolism; PDM, plant defence mechanism; PCD, programmed cell death)
Log2 Fold change WS/WWSpot numberProtein namepI/Mr exppI/Mr theoM score/ pep#Plant species/ accession #
  1. The theoretical and experimental pI and Mr, the Mascot score and the number of matched peptides, the plant species and the gi protein accession number are indicated. The log2 transformation of the fold change WS/WW after 7, 10 and 12 d is shown by the bar figure at the left. Positive fold change indicates the increase of protein abundance under water stress. The asterisks indicate significant changes in protein abundance in WS plants as compared to WW after 7, 10 and 12 d (*P < 0.05; **P < 0.01; ***P < 0.001).

  2. WS, water-stressed; WW, well-watered.

Day 7 Day 10 Day 12CWM
inline image720Aldose-1-epimerase-like protein7.7/399.4/39111/3Oryza sativa 2739168
inline image238Aldose-1-epimerase-like protein7.6/429.4/39222/4Nicotiana tabacum 2739168
inline image714α-Galactosidase preproprotein, putative7.1/388.1/46111/2O. sativa 78708844
inline image229Putative early nodulin 8 precursor7.1/426.9/4275/2O. sativa 50938787
inline image284Putative lipase7.1/328.2/36113/3O. sativa 55296706
inline image266Putative lipase6.1/348.2/36176/4O. sativa 55296706
inline image115Putative pectin methylesterase6.4/746.1/58214/5O. sativa 34907404
inline image190Putative polygalacturonase6.8/455.9/51160/3O. sativa 55775064
inline image754Putative polygalacturonase6.1/455.9/5178/2O. sativa 55775064
inline image303Xyloglucan endotransglycosylase homolog6.0/296.2/31278/8Zea mays 563235
inline image705Xyloglucan endotransglycosylase homolog6.3/296.2/31282/9Z. mays 563235
inline image767Xyloglucan endotransglycosylase homolog6.6/306.2/31192/7Z. mays 563235
Day 7 Day 10 Day 12CWM/PDM
inline image636β-glucanase:ISOTYPE=II6.2/298.9/35170/3Hordeum vulgare 228411
inline image295Putative class IV chitinase4.7/304.8/28180/4O. sativa 50910275
inline image643Putative class IV chitinase4.8/314.8/28170/4O. sativa 50910275
inline image720Peroxidase 52 precursor, putative, expressed7.7/396.9/34112/3O. sativa 77548362
inline image266Peroxidase 52 precursor, putative, expressed6.1/346.9/34162/4O. sativa 77548362
inline image273Peroxidase6.3/327.6/32254/5Cenchrus ciliaris 520570
inline image279Peroxidase6.5/347.6/32314/7C. ciliaris 520570
inline image284Peroxidase7.1/327.6/32244/5C. ciliaris 520570
inline image305Putative β-1,3-glucanase9.2/295.1/59178/5O. sativa 50938049
inline image211Putative peroxidase7.4/415.8/36212/5O. sativa 52076880
inline image339Putative peroxidase precursor4.7/255.1/5999/2O. sativa 50939979
inline image423Thaumatin-like pathogenesis-related protein7.8/96.3/17149/4Avena sativa 662351
inline image604Thaumatin-like pathogenesis-related protein 3 precursor7.2/155.8/1799/3A. sativa 1710785
inline image418Thaumatin-like pathogenesis-related protein 3 precursor7.2/165.8/1787/2A. sativa 1710785
Day 7 Day 10 Day 12PDM
inline image605At3g136507.0/156.7/1899/2O. sativa 62734201
inline image759Putative cupin family protein6.1/245.9/36157/4O. sativa 50919209
inline image337Putative germin A6.3/276.9/2465/2O. sativa 50941859
inline image341Putative germin A5.3/245.9/24191/5O. sativa 50941859
inline image211Putative polygalacturonase inhibitor7.4/415.9/36102/2O. sativa 50931079
inline image229Putative polygalacturonase inhibitor7.1/425.9/36119/2O. sativa 50931079
inline image755Putative polygalacturonase inhibitor5.1/265.9/36104/3O. sativa 50931079
inline image367Zeamatin-like protein9.4/217.3/ 24115/3Zea diploperennis 75993883
Day 7 Day 10 Day 12PCD
inline image629Endonuclease6.4/345.4/34177/3Zinnia elegans 3242447
inline image630Endonuclease6.5/355.4/34155/3Z. elegans 3242447
inline image190Nucleotide pyrophosphatase homolog6.8/455.1/52150/4O. sativa 34903578
Day 7 Day 10 Day 12Unclassified protein
inline image180OSJNBa0070C17.165.9/476.3/51390/8O. sativa 50928681
inline image365Secretory protein4.7/229.3/24167/3Triticum aestivum 5669008

The identified proteins whose abundance changed fell into two major biological categories (Table 2): cell wall metabolism such as peroxidase, xyloglucan endotransglycosylase, polygalacturonase inhibitor and pectin methylesterase, and plant defence mechanisms such as thaumatin-like pathogenesis-related protein, zeatin-like protein, cupin family protein, putative germin A, class IV chitinase and β-1,3-glucanase. No unique proteins were identified in sap from WS plants. Most of the significant quantitative differences in protein abundance were observed after 10 and 12 d of withholding water (Table 2).

All isoforms of the aldose-1-epimerase-like protein and xyloglucan endotransglycosylase homolog classified in the cell wall metabolism category increased in abundance after 10 and 12 d of water stress (Table 2). These enzymes are involved in cell wall loosening for cell elongation (Cosgrove 1999). Several peroxidases, classified in the cell wall metabolism category, but also shown to participate in plant defence mechanisms, were up-regulated under water stress. Most peroxidases identified at a more basic pI (spot numbers 720, 279 and 211; Table 2) increased after 10 and 12 d of water stress, while those identified at a more acidic pI (spot numbers 273 and 339; Table 2) decreased in abundance, mainly after 12 d. There were two exceptions: spot 266 (pI = 6.1) increased after 10, but not 12 d, and spot 284 decreased after 10 and 12 d, but had an approximately neutral pI (7.1). The abundance of other cell wall metabolism proteins, glucanases, chitinases and thaumatins, also known as pathogenesis-related proteins, also varied with water stress. Glucanases and chitinases decreased under water stress, while all isoforms of thaumatin significantly increased in abundance after 12 d of water stress, as did the putative germin A classified in the plant defence mechanism category. Putative polygalacturonases significantly increased in abundance after 10 d, while putative polygalacturonase inhibitors significantly increased after both 10 and 12 d of water stress.

pH Bioassay

To test for the presence of potential bioactive peptides in the xylem sap, we compared two different fractions of the sap using a suspension cell bioassay. The protein fraction which contained greater than 5 KDa proteins from either WW plants or WS plants did not induce alkalinization of maize suspension cells (Fig. 4a). However, the peptide fraction (less than 5 KDa) from both WW and WS plants induced alkalinization (Fig. 4b). In order to test whether small peptides may be responsible for the alkalinization, the sap was treated with pronase. After pronase treatment, the peptide fraction maintained its ability to induce alkalinization (Fig. 4c). These experiments also showed that the compounds inducing pH alkalinization were not water stress specific and may not be proteinaceous. Because peptides may be resistant to proteases, we tried to isolate the activity using reversed-phase C18 HPLC (data not shown). After separation of the sap into multiple fractions, we were not able to measure a pH alkalinization with any of the fractions collected (data not shown).

Figure 4.

pH Change of external medium in which maize suspension cells are bathed in response to the addition of different xylem sap fractions from well-watered (WW) and water-stressed (WS) plants after 7 d of water stress. (a) The protein fraction >5 kDa, (b) the peptide fraction <5 KDa and (c) the peptide fraction after treatment with pronase. The rapid alkalinization factor (RALF) peptide was used as a positive control to establish the degree of alkalinization in maize suspension cell cultures and the buffer alone as a negative control except in (c) where the negative control used is RALF treated with pronase. P, protein; SM, peptide fraction.


This study characterized the changes in 61 metabolites and the proteome of xylem sap during the course of a progressive drought stress. The aim was to gain a more comprehensive understanding of how the sap composition changes under water stress to better understand long-distance root signalling induced at early and later stages of drought. We quantified xylem sap amino acids; organic acids; plant hormones; and their metabolites, polyamines, phenylpropanoids and flavonoids.

ABA and cytokinin levels under water stress – regulation of stomatal aperture

ABA and its catabolite PA increased in abundance under water stress. ABA is known to increase in xylem sap under water stress and is one of the dominant long-distance root-to-shoot signals when soils become dry (Wilkinson & Davies 2002; Davies et al. 2005). The levels of ABA in the xylem depend on synthesis and also other factors involved in ABA metabolism. Increased levels of ABA in sap may occur because of: (1) a higher rate of ABA biosynthesis in roots leading to increased translocation via xylem sap (Peuke, Jeschke & Hartung 1994); (2) increased recirculation of ABA from leaves to roots via the phloem as shown in salt stress and phosphate deficiency (Jeschke et al. 1997); (3) ABA may be released from its conjugate ABA-glycosylesters (ABA-GE) which are present in the xylem sap of sunflower and barley (Bano et al. 1994; Dietz et al. 2000); or (4) reduced catabolism of ABA to PA (Hansen & Dörffling 1999).

The data on ABA, ABA-GE and PA concentrations in xylem sap and those of other studies provide some insights into the possible forms that ABA is transported in drought-stressed maize and the amount of ABA catabolized. The low export of ABA from leaves under water stress conditions (Jia & Zhang 1997) would argue against the recirculation of ABA from leaves suggesting that the ABA coming from the roots via the xylem is synthesized in the roots. Although ABA is probably synthesized in roots, important carotenoid precursors may come from the leaves (Ren et al. 2007). We were not able to detect ABA-GE in maize xylem sap from WW or WS plants, so in our studies on maize this form of ABA does not seem to be a dominant transport form. However, the ABA in the xylem sap may in part come from the hydrolysis of ABA-GE under drought conditions (Sauter & Hartung 2000). An enzyme has been identified in the cortical apoplast of maize roots that cleaves ABA-GE releasing ABA, which may enter the xylem vessels from the apoplast. An enzyme that releases ABA from the conjugated form is a β-d-glucosidase and was recently shown using genetic tools to be involved in increasing free ABA levels under dehydration stress in Arabidopsis (Lee et al. 2006). Although the hydrolysis of ABA-GE to ABA in the root apoplast before release in the xylem sap may explain why we could not detect ABA-GE in the xylem sap, other studies have identified ABA-GE as being present in the xylem sap of maize and many other plant species (Sauter, Davies & Hartung 2001). In one study on drought-stressed maize roots, several β-d-glucosidases in the cell wall increased in abundance under drought (Zhu et al. 2007) that may be involved in hydrolysis of ABA-GE. These β-d-glucosidases were not seen in this proteomic analysis because we analysed proteins in the sap and not in root cell walls.

The low levels of PA in xylem sap suggest that the ABA present is not being catabolized on its way to the leaves (Zeevaart & Boyer 1984). In the xylem sap of sunflower, ABA-GE was found in greater concentrations than PA (Hansen & Dörffling 1999). However, in our study, we detected PA, but not ABA-GE, despite a 10-fold lower detection level for ABA-GE. The PA content increased slightly with water stress. Although PA has been shown to induce stomatal closure (Sharkey & Raschke 1980), it was found in trace concentrations, 500–1300 times lower than ABA under drought stress in maize, which suggests that ABA catabolism is not dominant under these conditions. Furthermore, PA content tracked closely with ABA content which was similar to the study by Ren et al. (2007) who found that the rate of ABA degradation in maize leaves was proportional to the concentration of ABA under both WS and non-stressed conditions.

Cytokinins like Z may also regulate stomatal aperture by keeping stomata open (Dodd 2003). Under water stress, we found that Z concentrations decreased and the conjugated form ZR was in higher concentrations in the xylem sap of WS plants, in agreement with previous works (Bano et al. 1994; Shashidhar et al. 1996). We also report for the first time the presence of BAP in maize which more recently has been recognized as a natural cytokinin in plants (Strnad 1997; Veach et al. 2003). Here, BAP was found in maize xylem sap in relatively high concentrations. In our study, Z and ZR in the xylem sap decreased under water stress, while in the aromatic base cytokinin, BAP increased under water stress. Previous studies showed that the application of BAP in maize at higher levels to what we found in WS plants (days 10 and 12) did not affect stomatal aperture (Pospisilova & Batkova 2004).

Increased levels of BAP in the xylem sap may be important under drought by delaying leaf senescence (McDavid, Sagar & Marshall 1973). BAP at high concentrations is also known to induce proline accumulation (Thomas, McElwain & Bohnert 1992), which acts in osmotic adjustment under water stress. We observed increased proline accumulation after 12 d of water stress, and other studies have also shown that proline accumulation in roots of maize increases under water stress (Ober & Sharp 1994). Our results suggest that BAP may play a role in response to drought, but more studies are needed to clarify the role of BAP in maize under drought stress.

Phenylpropanoid metabolism under water stress: a possible role in lignification and cell wall stiffening

The BAP in maize xylem sap may have other functions in water stress. For example, exogenously applied BAP was shown to induce anthocyanin synthesis in Arabidopsis (Deikman & Hammer 1995), which may in part be because of the post-transcriptional modification of enzymes in the phenylpropanoid biosynthetic pathway such as chalcone synthase and dihydroflavonol reductase. Because anthocyanins in the xylem sap would be unstable, we quantified intermediates of the phenylpropanoid pathway. One intermediate, p-coumaric acid, increased in WS plants. Besides the possible fate of p-coumaric acid in the anthocyanin pathway, p-coumaric acid is also an intermediate in lignin biosynthesis, as are caffeic acid and ferulic acid which were also identified in maize xylem sap. While p-coumaric acid and caffeic acid concentrations both increased in WS plants, ferulic acid concentration decreased over the 12 d course of the water stress to become significantly lower at day 12. Multiple enzymes such as 4-coumarate-CoA ligase (4CL) (Hu et al. 1999), caffeate O-methyltransferase (COMT) (Vincent et al. 2005) and coniferaldehyde 5-hydroxylase (Li et al. 2003a) are involved in regulating the relative abundance of monolignols such as coumaryl, coniferyl and sinapyl alcohols which may be polymerized into lignin. We found that the substrates of the 4CL enzymes, caffeic acid and p-coumaric, accumulated in the xylem sap under water stress. The accumulation of these monolignols in the xylem sap suggests that water stress decreases lignin biosynthesis, which may be related to the coordination of growth along the length of maize leaves (Vincent et al. 2005).

In addition, some peroxidases are also involved in processes related to cell wall metabolism including the cross-linking of monolignols (Passardi, Penel & Dunand 2004). Several isoforms of peroxidases involved in the last steps of cell wall lignification occurring in the tracheid elements are differentially expressed during water stress (Sato et al. 1995). Peroxidases may also function in suberization, cross-linking of cell wall structural proteins, defence against pathogen attack, oxidative stress protection, cell elongation and reactive oxygen species (ROS) production (Hiraga et al. 2001). Furthermore, isoforms of peroxidases may have different functions (Christensen et al. 1998; Bolwell et al. 2002; Li et al. 2003b). Cell wall anionic peroxidases catalyse the reaction between hydrogen peroxide (H2O2) and cinnamyl alcohols to form phenoxy radicals necessary to polymerize and form lignin, while cell wall cationic peroxidases oxidize NADH produced in muro to form H2O2. We found two groups of peroxidases with different pIs and different patterns of variation under water stress including spots 273 and 339 which represent anionic peroxidases that decreased in abundance, while spots 720, 279 and 211 represent cationic peroxidases that increased in abundance, after 10 d. We speculate that the reduced accumulation of the anionic peroxidases in xylem sap may be linked to a water-stress-induced reduction in lignification. The speculation that lignin production is reduced is also suggested by the metabolite analysis which showed an increase in the abundance of free monolignols in xylem sap.

Increases in the abundance of peroxidases might also lead to the increased cross-linking of the other components of cell wall in response to drought as suggested previously (Passardi et al. 2004). Cell wall cross-linking may be increased by water stress as indicated by the drop in ferulic acid content after 10–12 d of water stress. It is known that growth in roots and leaves slows down under drought because of reduced cell elongation. Under drought conditions, the peroxidase-catalysed formation of diferulic bonds between polysaccharide-bound lignin (Veitch 2004) may be enhanced leading to rigidification of the cell wall. The drought-induced accumulation of other enzymes involved in cell wall reinforcement, such as XET as demonstrated in this study further suggests the modification of the tracheid elements (Cosgrove 1999; Bourquin et al. 2002). Cell wall modification under drought is also suggested by the accumulation of the aldose-1-epimerase, which is also known as a mutarotase-like protein that catalyses the conversion of the β-anomer to the α-anomer of galactose (Holden, Rayment & Thoden 2003) and may function in cell wall rearrangement and stiffening.

Early signals induced by water stress versus adaptation to long-term water stress

We found that several compounds increased or decreased in WS plants after 7 d of drought. Among the compounds whose abundance increased early in the drought, ABA is already known as an early long-distance signal of water stress (Davies et al. 2005) inducing stomatal closure. Malate also increased early, making it a candidate for study of possible signalling effects. Many amino acids in the sap increased transiently in WS plants at 7 d, but little is known about why amino acids increase in sap very early in water stress. A number of amino acids also accumulated only under severe water stress. Among these, pro accumulates in WS roots and is induced by ABA accumulation (Ober & Sharp 1994), but the functional significance of the accumulation of other amino acids is not understood.

Most changes in protein abundance came after 10 d of drought. The proteins identified, which have functions in cell wall metabolism and defence mechanisms, are more likely involved in adaptation by cell wall stiffening than as water stress signals. Cell wall stiffening may increase the strength of xylem vessels to better withstand the increased tension occurring during water stress or to restrict water loss from internal tissues.

The role of hormones in the intercellular communication in plants is accepted and understood. In contrast, the role of small peptides in plants is just beginning to be elucidated. Tomato systemin was the first identified plant signalling peptide (Ryan & Pearce 1998), and so far only a few peptides with biological activity have been discovered (Matsubayashi & Sakagami 2006). In our search for peptides as signalling molecules in the root-to-shoot communication in response to water stress, we observed that the peptide fraction from the xylem sap was able to induce an alkalinization of the external medium of corn suspension cells. This alkalinization was induced by both WW and WS xylem sap, but a proteinaceous component could not be purified from sap that induced alkalinization. Tests with an active non-specific protease suggested that the activity was non-proteinaceous. Future technical advances will aid in the identification of the factor responsible for the rapid alkalinization in the peptide fraction of xylem sap.

In this study, we detected 31 compounds in the xylem sap of maize using an LC–MS/MS, and a database was created for use by other groups for the identification and quantitation of these plant-specific small molecules in MRM mode. We detected differences in both the small molecule and protein content of maize xylem sap under droughted versus WW conditions which may contribute to root-to-shoot communication at early stages of drought and during adaptation to more severe drought. Proteomic and metabolomic results suggest a working hypothesis that water stress reduces lignin biosynthesis of the root xylem vessels, and likely increases cell wall stiffening.


This research was funded by a Plant Genome Program Grant No. (NSF-DBI-0211842) to D.P.S. We thank DuPont, Monsanto, Ceres and the National Corn Growers Association for their contribution of the MaizeSeq database to the public domain. Thanks to Drs Oliver Yu (Donald Danforth Plant Science Center, St Louis, MO, USA), Michael Neff (Washington University, St Louis, MO, USA) and Nobuhiro Hirai for providing standards. Thanks to Bert Berla for technical assistance. We thank Greg Pearce and Dr Clarence Ryan (Washington State University, Pullman, WA, USA) for guidance and for providing the RALF peptide in our experiments. Critical comments on the manuscript were provided by Joseph Jez, Oliver Yu, Leslie Hicks and Sixue Chen.