The seeds of cotton (G. hirsutum L.) Yiluzao-7, a major cotton cultivated in north China, were sown in 34 cm diameter plastic pots containing 5.5 kg of washed sand. Each pot contained five seedlings. The pots were watered daily with Hoagland nutrient solution at 17:00–18:00. All pots were placed outdoors but were kept from rain. The temperature range at day time was 23–27°C and 19–22°C at night time.

Two alkali salts (NaHCO₃ and Na₂CO₃) were selected based on the salt content and pH of the majority of alkali soils in northeast China. For the alkali stress treatment, two alkali salts were mixed in a 9:1 molar ratio (NaHCO₃:Na₂CO₃). The total salt concentration was set at 80 mM (pH, 9.06). In the 80 mM alkali stress treatment solution, a mixture of 72 mM NaHCO₃ and 8 mM Na₂CO₃ resulted in ion concentrations of 88 mM Na⁺, 72 mM HCO₃^-, and 8 mM CO₃^2-.

Firstly, all plants were grown in full Hoagland nutrient solution for 30 days (since sowing). After growing for 30 days, ten pots of seedlings were randomly divided into two sets: one set (five pots) was used as control, which was watered with nutrient solution daily for three times; another set (five pots) was watered with the nutrition solution with the alkali salts daily for three times. The plants were subjected to alkali stress solution by watering from 17:00 to 18:00 for six consecutive days. Five seedlings of a pot were pooled as a biological replicate, therefore, each treatment have five pots as five biological replicates. The five seedlings in each pot were harvested after treatment for 6 days.

Chlorophylls a and b and carotenoid were extracted with acetone; each sample was spectrophotometrically analyzed for five times at 440, 645, and 663 nm five times in accordance with previously described methods. We used the equations proposed by Arnon (1949) for pigment concentration estimations.

Dry old and young leaves were ground, and ∼100 mg of tissue samples were dry-digested in a muffle furnace at 500°C for 6 h, and then 10 mL of HNO₃:H₂O (1:1) was added to extract ions. The contents of Na, K, Ca, Mg, P, Fe, Cu, Zn, and Mn were determined using an ICP-OES spectrometer (iCAP 6000 series, Thermo Fisher Scientific Inc.) according to the operation manual. After water extraction was performed, the quantities of anions (NO₃^-, Cl^-, SO₄^2-, and H₂PO₄^-) were determined through ion exchange chromatography (DX300 ion chromatographic system; AS4A-SC ion-exchange column, CD M-II electrical conductivity detector; DIONEX, Sunnyvale, CA, USA) with a mobile phase comprising 1.7 mM/1.8 mM Na₂CO₃/NaHCO₃. Each measurement was repeated five times.

Approximately 100 mg of each frozen tissue sample was transferred into 2 ml centrifuge tubes, and 60 μl of water containing ribitol as an internal standard was added to each tube. After the mixtures were vortexed with 0.3 ml of methanol and 0.1 ml of chloroform, a 70 Hz grinding mill system (Jinxin Biotech LTD. Shanghai, China) was used to grind the samples for 5 min; the samples were then incubated at 70°C for 10 min. The tubes were centrifuged at 12,000 rpm at 4°C for 10 min. Supernatant (0.35 ml) was decanted into a 2 ml volume screw-top glass tube; the samples were dried in a vacuum concentrator at 30°C for 2 h. Afterward, each sample was dissolved in 80 μl of methoxamine hydrochloride and incubated at 37°C for 2 h. The samples were further derivatized with N,O-bis(trimethylsilyl)-trifluoroacetamide (BSTFA) containing 1% trimethylchlorosilane (TMCS; 100 μl) at 70°C for 1 h (Lisec et al., 2006). GC-TOF/MS analysis was performed using a 1D Agilent 7890 gas chromatograph system coupled with a Pegasus 4D time-of-flight mass spectrometer. The system was equipped with a DB-5MS capillary column coated in 5% diphenyl cross-linked with 95% dimethylpolysiloxane (30 m × 250 μm inner diameter and 0.25 μm film thickness; J&W Scientific, Folsom, CA, USA). An aliquot of the analyte (1 μL) was injected in a splitless mode. Helium was used as carrier gas, the front inlet purge flow was 3 mL min^-1, and the gas flow rate through the column was 1 mL min^-1. The initial temperature was maintained at 90°C for 0.25 min; temperature was increased to 180°C at a rate of 10°C min^-1 and to 240°C at a rate of 5°C min^-1. Temperature was further increased to 285°C at a rate of 20°C min^-1 for 11.5 min. Injection, transfer line, and ion source temperatures were 280, 270, and 220°C, respectively. Energy was set at -70 eV in an electron impact mode. Mass spectrometry data were acquired in a full-scan mode with an m/z range of 20–600 at a rate of 100 spectra per second after a solvent delay of 492 s.

Photosynthetic pigments, metal elements, and inorganic anions were statistically analyzed using SPSS 13. All data were presented as average, along with the standard error (SE), of five biological replicates. Significant difference between old and young leaves at same stress condition was determined by T-test. Metabolites were identified by searching FiehnLib (GC-TOF), a commercial EI-MS library (Kind et al., 2009). The resulting three-dimensional data, namely, peak number, sample name, and normalized peak area, were run in SIMCA-P 13 software package (Umetrics, Umea, Sweden) and subjected to principal component analysis (PCA) and orthogonal projections to latent structure-discriminant analysis (OPLS-DA). Non-commercial databases, including Kyoto Encyclopedia of Genes and Genomes (KEGG)¹, were utilized to search for metabolite pathways. Format data were uploaded in the MetaboAnalyst website ² for further analysis (Paz and Martinez-Ramos, 2003).

Alkali stress-induced growth inhibition was more evident in old leaves than in young leaves (Figure 1). In addition, alkali stress increased the concentrations of chlorophylls a, b, a+b, and carotenoids in young leaves but it reduced their amounts in old leaves (Figure 1). In response to 6-day alkali stress treatment, Na content in old and young leaves significantly increased by 8.67 and 3.39 times, respectively (Table 1). K content in old leaves decreased, whereas no significant change was observed in young leaves (Table 1). The effects of alkali stress on Ca, Mg, Fe, Cu, Zn, and Mn were not significant both in old and young leaves (Table 1). Moreover, under alkali stress, Ca and K became the dominant component in the old and young leaves of total mineral element, respectively (Table 1). Alkali stress increased Cl^- contents in both old and young leaves but it reduced NO₃^- and H₂PO₄^- contents (Figure 2). Moreover, SO₄^2- contents increased in young leaves but decreased in old leaves under alkali stress (Figure 2).

A total of 133 metabolites were identified in the leaves of cotton seedlings, and all samples were distributed within the 95% confidence interval of Hotelling’s T2 ellipse (Figure 3). The scores plot of PCA results showed that ∼75% variability in the four groups of samples can be explained using two principal components (Figure 3A). Additionally, heatmap analysis (Supplementary File S2) showed an obvious separation between samples treated with and without alkali stress treatment in different tissues. OPLS-DA was conducted using one orthogonal and one predictive component calculated from all models derived from two classes of samples. The scores plot of OPLS-DA results clearly showed the separation between the leaves of cotton leaves treated with 80 mM alkali stress for 6 days and with good model quality (Figures 3B–E). A total of 133 metabolites, including organic acids, amino acids, sugars and poly-sugars, and nucleotide derivatives, were identified. Detailed information on these identified metabolites is shown in Supplementary File S1. Most of the metabolites detected in this study were predominantly classified under the general biochemical pathways, such as TCA cycle, glycolysis, GABA pathway, GS/GOGAT cycle, Shikimic acid pathway, and amino acid metabolism (Table 2) based on search results in Plant Metabolic Network³ and KEGG.

Alkali stress strongly affected the metabolites, including organic acids, amino acids, sugars and poly-sugars, and nucleotide derivatives in leaves (Figure 4). Alkali stress had little effect on young leaves but it considerably increased sugar and poly-sugar contents (Figure 4). In addition, alkali stress significantly reduced the amounts of organic acids but it increased amino acid contents in old leaves (Figure 4). In the TCA cycle, the amounts of oxalic acid, citric acid, aconitic acid, α-ketoglutaric acid, and succinic acid significantly decreased, and the degree of reduction was higher in old leaves than in young leaves under alkali stress treatment (Table 2). Mannose was up-accumulated in cotton in addition to the metabolites involved in glycolysis and which were obviously up-accumulated in young leaves (Table 2). Glutamic acid (used in proline synthesis) was depleted dramatically probably because of the synthesis of downstream metabolites; proline increased by 5.02- and 2.62-folds in young and old leaves under alkali stress (Table 2). In glutamine synthetase/synthase (GS/GOGAT), the amounts of transamination-related metabolites clearly changed in young leaves under alkali stress; asparagine contents increased, whereas aspartic acid and glutamine contents decreased (Table 2). GABA pathway was enhanced under alkali stress, resulting in increased GABA and putrescine contents (Table 2). Moreover, the concentration decline of ethanolamine and increase of glycine and serine is indicative that the metabolism of plasma membrane was inhibited (Table 2). Shikimic acid pathway in young leaves was not significantly affected by alkali stress. However, the amount of shikimic acid significantly decreased and that of quinic acid significantly increased in old leaves, demonstrating that shikimic acid pathway was inhibited in old leaves under alkali stress (Table 2).

The young and old leaves of cotton exposed to 80 mM alkali stress for 6 days showed distinct differences in terms of growth; alkali stress exerted a stronger inhibitory effect on old leaves than on young leaves. This result implies that young organs are protected at the expense of the old organs; this phenomenon is possibly an adaptive strategy of plants to alkali stress. Alkali stress considerably increased Na⁺ content of both old and young leaves; however, old leaves accumulated higher Na⁺ concentration than young leaves (Table 1). Moreover, alkali stress did not significantly influence K⁺ content of young leaves but strongly reduced K⁺ content in old leaves (Table 1). Na⁺ enters the plant cells via the K⁺ transporter pathways and through the non-selective cation channels, and Na⁺ exclusion mechanism is dependent on H⁺ gradient across cell membrane (Zhu, 2003; Yang et al., 2012). These results maybe indicated that cotton could compartmentalize Na⁺ into vacuoles of old leaves to prevent the onset of ion toxicity in the entire young organ. Cotton plants possibly utilize a specific regulatory mechanism of Na⁺ transmission in old leaves. The results indicated that alkali stress exerted no significant effect on Ca, Mg, Fe, Cu, Zn, and Mn contents in both leaves, and this result was possibly caused by the short-term treatment with 80 mM alkali stress (Table 1). Accumulation of large amounts of Na⁺ in tissue causes ionic imbalance; to maintain ionic balance and pH homeostasis, plants usually accumulate inorganic anions, such as Cl^-, NO₃^-, and SO₄^2- (Yang et al., 2008, 2009). Our results showed that old and young leaves accumulate Cl^-, indicating that Cl^- accumulation is a general adaptive response to excessive Na⁺ in cotton leaves (Figure 2). However, under alkali stress, SO₄^2- content was increased in young leaves but descreased in old leaves. This phenomenon is possibly a special adaptive response to excessive Na⁺ in young leaves and possibly plays important roles in maintenance of ionic balance and pH homeostasis in young leaves (Figure 2).

Plants growing in saline environments suffer from osmotic stress and are exposed to excessive Na⁺, which induces generation of reactive oxygen species (ROS) and causes protease activation and intracellular hyperammonemia (Parida and Das, 2005; Zhang J.T. et al., 2011). To avoid accumulation of excessive Na⁺ and hyperammonemia-induced cytotoxicity, plant cells normally react through ion transport, compartmentation, synthesis of compatible solutes, and transamination metabolites, which is abundant in plant tissues and involved in active synthesis and metabolism of energy (Zhang J.T. et al., 2011; Yang et al., 2012). However, interference between salinity and metabolites is a very complex network affecting nearly all metabolic and developmental processes in plants.

Metabolic changes caused by alkali stress (80 mM) are dependent on age of tissues (Supplementary File S1). Proline plays as osmoprotectant in plants subjected to drought or salinity stress (Delauney and Verma, 1993; Hare and Cress, 1997). Under alkali stress, young leaves and old leaves increased 2⁵- and 2²-fold (fold was calculated by the formula: log₂^{Alkali/control}) in proline content compared to their controls (Table 2). The same phenomenon in cultivated cotton leaves was reported by Hu et al. (2009) and Chen et al. (2011). Moreover, reduction in glutamic acid contents indicated that glutamic acid was converted into proline by Δ1-pyrroline-5-carboxylate synthase (P5CS), especially in young leaves. Sugars, the amount of which increased significantly in young leaves under alkali stress, are compatible solutes produced in response to salinity stress (Chen and Murata, 2002). The amounts of inositol and mannose in both old and young leaves significantly increased in cotton subjected under alkali stress, and the extent of increase was higher in young leaves than in old leaves. Inositol plays important functions in membrane biosynthesis, and they protect the membrane by acting as free radical scavengers; our result is similar to that for Actinidia (kiwifruit) leaves subjected under salinity conditions as reported by Klages et al. (1999). Organic acid synthesis was significantly inhibited in old leaves. Citric acid, succinic acid, and malic acid, which are effective ROS-scavenging metabolites, were significantly downregulated in old leaves. By contrast, synthesis of amino acids, such as GABA, valine, and serine, was dramatically enhanced in old leaves, and this phenomenon is favorable for osmotic adjustment and membrane stability. The results suggest that proline is a common compatible solute that showed great changes under alkali stress in leaves; in addition, sugars (glucose, inositol, and mannose) were the compatible solutes specific in young leaves, whereas some amino acids (GABA, valine, and serine) were the solutes specific in old leaves. Under alkali stress, TCA cycle was inhibited significantly in old leaves compared with that in young leaves. The results suggest that alkali stress (high pH) caused mass dissipation of energy, and organic acid synthesis was inhibited in old leaves, implying that alkali stress strongly influenced the energy production in old leaves. Under alkali stress, young leaves showed high amounts of photosynthetic pigments and accumulation of increased number of metabolites (PEP, 3PGA, F6P, G6P, and glucose) that are associated with glycolysis. This finding suggests that alkali stress stimulated production of reducing force and enhanced N metabolism, which in turn increased sugar production and enhanced glycolysis in young leaves. In GS/GOGAT cycle, aspartic acid and glutamine contents decreased and asparagine content increased in young leaves under alkali stress, suggesting that these changes in the amounts of transamination-related metabolites are consistent with the shift of metabolic activities toward proline biosynthesis. GABA pathway apparently converts transamination products into compatible osmolytes (Shelp et al., 2006). In the present case, GABA contents increased dramatically in cotton leaves, indicating that GABA functioned as osmoregulator and as cytosolic pH regulator in response to alkali stress; moreover, GABA serves as an intermediate that assists in biosynthesis of osmolytes, such as myo-inositol and proline (Bouche and Fromm, 2004). Ethanolamine and shikimic acid are plant metabolites that participate in synthesis of membrane phospholipids, phosphatidylethanolamines, and lignin, which are used in the synthesis of plant cell membranes and cell wall (McNeil et al., 1999; Wang et al., 2002). Our results suggested that suppression of ethanolamine and shikimic acid restricts cell membrane and wall elongation, leading to growth inhibition of old leaves of cotton under alkali stress.

In summary, alkali stress exerted a considerably stronger limiting effect on old leaves than on young leaves. Under alkali stress, young leaves can maintain low Na and high K contents, as well as maintain a relatively stable pigment accumulation and TCA, resulting in increased accumulation of photosynthetic metabolites. Comparison of the metabolic response of the old and young leaves clearly shows that they display different mechanisms of osmotic regulation. Under stress condition, organs at different developmental stages possibly utilize different mechanisms of metabolic regulation. We suggest that future investigations on alkali stress should include more organs obtained at different development stages.