SEARCH

SEARCH BY CITATION

Keywords:

  • acclimatization;
  • ionomic;
  • Lotus;
  • metabolomic;
  • salt stress;
  • transcriptomic

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The model legume Lotus japonicus was subjected to non-lethal long-term salinity and profiled at the ionomic, transcriptomic and metabolomic levels. Two experimental designs with various stress doses were tested: a gradual step acclimatization and an initial acclimatization approach. Ionomic profiling by inductively coupled plasma/atomic emission spectrometry (ICP-AES) revealed salt stress-induced reductions in potassium, phosphorus, sulphur, zinc and molybdenum. Microarray profiling using the Lotus Genechip® allowed the identification of 912 probesets that were differentially expressed under the acclimatization regimes. Gas chromatography/mass spectrometry-based metabolite profiling identified 147 differentially accumulated soluble metabolites, indicating a change in metabolic phenotype upon salt acclimatization. Metabolic changes were characterized by a general increase in the steady-state levels of many amino acids, sugars and polyols, with a concurrent decrease in most organic acids. Transcript and metabolite changes exhibited a stress dose-dependent response within the range of NaCl concentrations used, although threshold and plateau behaviours were also observed. The combined observations suggest a successive and increasingly global requirement for the reprogramming of gene expression and metabolic pathways to maintain ionic and osmotic homeostasis. A simple qualitative model is proposed to explain the systems behaviour of plants during salt acclimatization.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant salt stress has become a major concern worldwide due to the salinization of agricultural land caused by irrigation, so-called secondary salinization. This is of particular relevance as most crops are salt-sensitive and progressing desertification imposes a need for irrigation. Salinity imposes at least two primary stresses on plants: hyperosmotic stress caused by the reduction of water potential and consequently reduced water availability, and hyperionic stress, related to the toxic effects of the accumulated ions. Consequently, salinized plants are subjected to dehydration, metabolic toxicity, nutrient deficiencies, membrane dysfunction and oxidative stress, which lead to tissue damage and early senescence (Tester and Davenport, 2003). However, plants under salt stress do not show symptoms of cellular damage provided that the stress dose is not prolonged or is below the tolerance threshold. As sessile organisms, plants have evolved a number of strategies to acclimatize to various kinds of deleterious conditions and thus to increase competitiveness in various ecological niches. Unlike adaptation, which is a consequence of evolutionary mechanisms acting at the genetic level in populations over many generations, acclimatization is a proximal phenotypic response to changes in the environment (Orcutt and Nilsen, 2000). Over seconds, minutes, hours or days, plants can reprogram their metabolism, physiology and morphology to attain new physiological states, which ensures fitness and survival under abiotic and biotic constraints (Lichtenthaler, 1996). Among others, salinity acclimatization responses include: (i) the maintenance of ion homeostasis, including ion exclusion, compartmentation, redistribution, organ-specific allocation and excretion; (ii) osmotic adjustment and compatible solute accumulation; (iii) water balance and control of transpiration; and (iv) structural and anatomical changes, for example the modification of apoplastic barriers (Hose et al., 2001; Tester and Davenport, 2003). Recent evidence also demonstrates that growth and developmental responses are fundamental to plant survival under prolonged salt stress (Achard et al., 2006).

Despite knowledge that plants respond progressively to hyperosmotic and hyperionic stress over time, the mainstream of current molecular and biochemical research is focused on short-term signalling and early responses triggered by salts, and experiments are typically performed using lethal treatments (Munns, 2005). As a result, detailed knowledge of the molecular basis of salt-stress acclimatization is still largely lacking.

Legumes are second in importance to agriculture after grasses, and cover around 12–15% of the world’s agricultural land and supply 33% of human dietary nitrogen needs (Graham and Vance, 2003). They also play a critical role in natural and agricultural ecosystems, due to their ability to fix nitrogen. Given their importance, further knowledge of legume stress physiology is required to address present and future threats to food security. Here we present ionomic, transcriptomic and metabolomic analyses of the glycophyte model legume Lotus japonicus subjected to long-term regimes of non-lethal levels of salinity. The existence of perennial cultivated Lotus species that are particularly salt-tolerant provides a reason for using this model (Teakle et al., 2006). The results presented below reveal new molecular and metabolic components of the salt-stress response in legumes, and provide systems-level insights into the plastic acclimation process. Finally, a simple model is proposed that incorporates these observations to qualitatively explain the dose dependency, plateau and threshold behaviour of the responses.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Experimental set-up and physiological assessment of salt-acclimatized plants

To address the responses of plants during salt acclimatization, two long-term experimental designs were used. The first regime was based on a gradual acclimatization to salts, whereas the second was an initial acclimatization (ia) approach involving seed germination and growth on a range of defined salt concentrations (Figure 1a). Three independent experiments were performed in a greenhouse, each comprising controls and six treatments. These were labelled according to experimental design and final NaCl concentration as 50, 100 and 150 (gradual acclimatization), and ia25, ia50 and ia75 (initial acclimatization).

image

Figure 1.  Experimental design and physiological assessment of salt-acclimatized plants. (a) Gradual step acclimatization and initial acclimatization (ia) experimental designs (see Experimental procedures). i, imbibition; t, transplanting; s, salinization; d, days. (b) Plant growth and final shoot biomass. (c) Shoot sodium content. For (b) and (c), data are the mean ± SD of three independent experiments. Plants photographs were taken at the end of the experiment after 28 days under greenhouse conditions. FW, fresh weight. (d) Soil conductivity. Data are the mean ± SD of three independent experiments, each containing three independent replicated samples taken from pooled soil from five pots.

Download figure to PowerPoint

As expected, the final shoot biomass of salt-acclimatized plants decreased with increasing NaCl concentration (Figure 1b). Shoot Na+ content correlated linearly with increasing levels of salt added (r2 = 0.9418 and 0.9471 for the gradual and initial regimes, respectively), but the slope of the regression differed depending on the experimental approach (Figure 1c). In addition, soil conductivity was also linearly correlated (r2 = 0.9523) with the amount of NaCl added, indicating that salts did not accumulate differentially in the soil due to experimental design or the varying transpiration rates of plants (Figure 1d). As a consequence, the basic difference between the two approaches was that plants that were gradually acclimatized faced a higher level of osmotic stress in the roots for a given internal Na+ accumulation than those that were initially acclimatized, while the latter were under higher ionic stress than the former at the same final soil salt content.

Nutrient profiling

Inductively coupled plasma/atomic emission spectrometry (ICP-AES) was used to profile changes in shoot micro- and macronutrient contents (Table S1). anova analysis was performed to identify elements with altered levels associated with salt stress, using a false discovery rate lower than 1% (FDR < 0.01, Figure 2). As expected, Na+ and K+ levels were negatively correlated under increasing salinity. Calcium and magnesium were slightly increased under all salt treatments and manganese in some, to not more than 150% of controls. The strongest salt-induced decrease was observed for molybdenum, reaching approximately 30% of the control content. Sulphur and phosphorus decreased compared to the control to a minimum of 60% under the ia75 treatment. In addition, zinc decreased only under the most extreme doses of the initial acclimatization design, reaching around 70% of the control content. Iron and boron levels were not significantly altered by any of the salt treatments, and the contents of cadmium, cobalt, chromium, copper, nickel and selenium were below the detection limits.

image

Figure 2.  Shoot nutrients exhibiting statistically significant changes compared to the control treatment. Bars indicate control (white), gradual step acclimatization treatment (black) and initial acclimatization treatment (grey). Data are the mean ± SD of three independent experiments. Asterisks indicate statistically significant changes compared to the control at FDR < 0.01.

Download figure to PowerPoint

Gene expression analysis

Transcriptomic analysis using the Affymetrix Lotus Genechip® was performed in three independently replicated experiments to identify genes that were differentially expressed during salt acclimatization. Gene expression was analysed after hierarchical clustering/supporting tree (HCL-ST) analysis applied to the complete set of probesets. Treatments did not cluster either according to the experimental design or increasing soil conductivity. Rather, a trend coincident with shoot Na+ accumulation was observed (Figure 3a, compare with Figure 1c,d). The data set was further analysed by a significance-based comparison of salt-treated plants and controls, applying an FDR < 0.01 and a twofold change threshold. An increasing number of differentially regulated genes resulted from increasing salinity under both experimental designs (Figure 3b). A total of 912 probesets that matched both statistical and threshold criteria were classified as salt-stress-responsive (522 up-regulated and 390 down-regulated, Table S2). In agreement with the concept of increasing salt toxicity within shoot tissue, treatments with the highest salt accumulation showed an increased number of differentially regulated genes.

image

Figure 3.  Transcriptomic fingerprinting and microarray results and validation. (a) Hierarchical clustering/supporting tree analysis (HCL-ST) of the transcriptomic profiles. Bootstrap analysis comprised 10 000 iterations. (b) Number of probesets from microarray profiling declared salt-responsive in each treatment, based on the false discovery rate (FDR < 0.01), fold change (> twofold) and intersection of the thresholds. (c) Comparison of microarray and quantitative real-time RT-PCR data from 40 probesets coding for putative L. japonicus transcription factors. Each symbol represents the mean expression level (log2-transformed) of the 150 mm NaCl treatment relative to control treatment from three independent experiments. The open square represents the probeset chr2.CM0249.113 (see Results).

Download figure to PowerPoint

As transcription factors are often expressed at low levels and may be difficult to quantify by microarrays (Czechowski et al., 2004), quantitative real-time RT-PCR was used to validate array data for 40 probesets representing putative L. japonicus transcription factors. Genes classified or not as salt-regulated were tested, and the mean expression ratio between the 150 mm NaCl treatment and control samples was calculated from both chip and quantitative real-time RT-PCR data for the three independent experiments (Figure 3c and Table S3). Remarkably good agreement was found between the two technologies for 39 probesets, with a linear regression slope of 0.9989 and r2 = 0.7976. A single probeset failed the validation test (chr2.CM0249.113); it was strongly down-regulated according to the chip analysis but not changed according to the quantitative real-time RT-PCR data.

The expression pattern of the 50 most up-regulated and 50 most down-regulated probesets is shown in Figure 4(a). Dose-dependent induction or repression of these genes was evident, irrespective of the experimental approach. Remarkably, a subset of genes deviated from the general trend, responding only at high salt concentrations, while others appeared to reach a plateau of high or low expression. Representative expression patterns are shown in Figure 4(b) for genes exemplifying these behaviours. Functional annotation of salt-regulated probesets was obtained by comparison of all translated sequences to the Arabidopsis thaliana genome, and 807 probesets (88%) had a significant hit (E value ≤1e-5). These identifications were used to visualize the functional classification using MapMan software (Usadel et al., 2005). After manual checking, the most common functional groups were: transcription and RNA processing (10%), large enzyme families (miscellaneous, 10%), transport (8%), protein modification and degradation (7%), signalling (7%), stress and defence (6%), cell wall (5%) and hormone metabolism (4%), secondary metabolism (4%) and amino acid metabolism (2%) (Figure 4c). All categories comprised up- and down-regulated probesets, with the exception of the amino acid and cell-wall-related genes. Most of the probesets were up-regulated in the former and down-regulated in the latter, suggesting a potential requirement for salt-stress-specific regulation of these metabolic pathways (Figure S1). Subsequently, we focused on particular functional groups, for which expression of selected target genes was validated with quantitative real-time RT-PCR. Their expression was also tested in whole shoots subjected to salt acclimatization and drought, in seedlings under salt shock, and in various aerial organs under control and salt-acclimatization conditions.

image

Figure 4.  Overview of transcript profiling results. (a) Expression patterns of the 50 most up-regulated and 50 most down-regulated salt-responsive probesets from microarray profiling experiments. Each expression level (log2-transformed) represents the mean of three independent experiments. To aid comprehension, error bars were removed and a horizontal line was added at control expression level. (b) Representative mean expression patterns (log2-transformed) of six probesets among the 10 most up-regulated and 10 most down-regulated genes. (c) Non-redundant functional categories of salt-regulated probesets, according to the MapMan software.

Download figure to PowerPoint

Transcription and RNA processing-related genes.  Many transcription factors were found within this group, which presumably coordinate global transcriptional changes during the salt-acclimatization response. All transcription factor hits were verified at http://daft.cbi.pku.edu.cn/index.php. The most abundant transcription factor families were AP2/ERF (24%) and MYB (20%). Both families include characterized members involved in the control of plant stress regulons, such as the cold/osmotic stress-induced CBF/DREB sub-family, or the MYB transcription factors implicated in ABA-independent and dependent signalling pathways (Nakano et al., 2006; Yanhui et al., 2006).

Within the AP2/ERF A-5 sub-family of A. thaliana transcription factors, no biological function is currently known (Nakano et al., 2006). Therefore, we selected as a target gene the probeset TM0715.25, encoding a putative DREB-like transcription factor similar to those classified in the A-5 sub-family. The transcript levels of this gene were increased by both salt acclimatization and shock, but not by drought, suggesting a role related to ionic but not osmotic stress (Figure 5a). Expression was higher in stems and developing leaves than in mature leaves (Figure 5b), but was induced by salt acclimatization specifically in mature leaves (Figure 5c).

image

Figure 5.  Expression level of selected probesets monitored by quantitative real-time RT-PCR. Bars represent the mean relative expression level (log2-transformed) ± SD for three independent replicates. The dotted lines indicate a twofold change. Probesets were sorted according to the respective expression levels in (a). (a) Gene expression in shoot samples after gradual step acclimatization to 150 mm NaCl (black), in vitro salt shock (grey) and drought (white). (b) Relative gene expression compared to mature leaves, in stems (grey) and developing leaves (white) under control conditions. (c) Transcriptional regulation after salt acclimatization in mature leaves (black), stems (grey) and developing leaves (white), determined after gradual step acclimatization to 150 mm NaCl and compared to control treatment.

Download figure to PowerPoint

As flavonoid metabolism is involved in salt-stress responses in rice (Walia et al., 2005), we selected another gene represented by probeset TM1624.23. This putative transcription factor was 79% similar to the Zea mays C1 factor protein and 60% similar to AtTT2, both of which are involved in the regulation of pro-anthocyanidin metabolism (Nesi et al., 2001). Expression of this gene was induced by salt acclimatization and drought, but not by salt shock (Figure 5a). It was more highly expressed in stems and developing leaves than in mature leaves, but the salt elicitation was restricted to mature and developing leaves (Figure 5b,c). In addition, several other probesets representing genes putatively involved in the metabolism of phenylpropanoid derivatives were transcriptionally regulated, including homologues of AtPAL1, AtTT4, AtTT7 and AtTT19 (Table S2). We therefore predict a role for TM1624.23 in flavonoid metabolism controlling long-term environmental and developmental responses.

Transport-related genes.  Most of the probesets listed in this functional group were putative oligo-peptide, amino acid, sugar or organic acid transporters of the ABC family, suggesting an important role for changes in metabolite allocation under salt acclimatization. As the celery sucrose transporter AgSUT1 has been reported to be down-regulated under salt stress (Noiraud et al., 2000), we focused on two probesets coding for sucrose transporters: the up-regulated putative LjSUT1 (probeset chr5.CM0344.58, 80% similar to PsSUT1) and the down-regulated putative LjSUF1 (probeset chr2.TM0134.37, 85% similar to PvSUF1). The transcript levels of both genes were highly responsive during salt acclimatization and salt shock, but almost non-responsive to drought (Figure 5a). LjSUT1 was preferentially expressed in leaves compared to stems under control conditions, while the opposite was true for LjSUF1 (Figure 5b). However, both were elicited under salt stress in all organs, particularly mature leaves (Figure 5c). Because sucrose transporters are involved in loading of sucrose in the phloem (Gottwald et al., 2000), we hypothesize that a complex change in sucrose partitioning between source and sink tissues may be part of the salt-acclimatization process.

No Na+ transporters or Na+/H+ antiporters from the NHX or HKT families were shown to be up-regulated under salt stress, in contrast to previous results (Maathuis, 2006). The lack of probesets representing orthologous NHX genes in the Lotus Genechip® partially explains this discrepancy (only one is present, similar to AtNHX6). However, at least two putative HKT genes are represented. The apparent inconsistency may also be explained by our particular choice of long-term acclimatization treatments and restriction to non-lethal stress doses. On the other hand, we found a gene encoding a putative Cl channel that was repressed (Ljwgs_016759.2, 86% similar to AtCLC-b, Hechenberger et al., 1996) and identified one putative K+ transporter as down-regulated by salt treatment (chr5.CM0911.54.1, 79% similar to AtKUP1, Kim et al., 1998). In addition, several genes encoding putative NO3 and NH4+ transporters exhibited reduced expression, including a NO3 transporter that is 71% similar to AtNRT1.2 (probeset chr1.CM0206.164), a NH4+ transporter that is 84% similar to AtAMT2 (probesets gi15799271 and Ljwgs_114175.1_s), and a novel NH4+ transporter of the LjAMT1 family with 89% similarity to LeAMT1.3 (probesets Ljwgs_028040.1 and Ljwgs_054494.1) (Table S2).

Signalling-related genes.  This functional group included putative membrane and cytoplasmatic receptor-like kinases, protein kinases and phosphatases, calmodulin-binding proteins and mitogen-activated protein kinases, suggesting a substantial change of the sensitivity and mode of control exerted by signalling pathways. Homologues of characterized genes integrating environmental signals included the probesets chr4.CM0617.35, encoding a protein 63% similar to the phosphatase type 2C AtABI1 (Leung et al., 1997), and Ljwgs_014485.1 and TM0845.12, which are highly similar to the CBL-interacting protein kinases of A. thaliana (Batistic and Kudla, 2004). Among the most up-regulated genes was probeset gi4336433 (Figure 4b), coding for a protein phosphatase type 2C associated with nodule development and highly expressed in nodules and flowers (LjNPP2C1, Kapranov et al., 1999). The expression of LjNPP2C1 was highly induced by salt acclimatization and salt shock, and to a lesser extent by drought (Figure 5a). Interestingly, transcript levels were equally abundant and elicited in mature leaves, stems and developing leaves (Figure 5b,c). These results may indicate an unexpected link between the nodulation process and the physiology of abiotic stress tolerance. In line with this, other probesets coding for nodulins were transcriptionally regulated, including LjNOD16, LjNOD21, LjENOD40-2 and nodulin-like proteins similar to MtN21 and MsENOD8 (Table S2).

Other genes of interest.  The stress and defence-related functional group comprised several putative LEA genes, heat shock- and cold- or dehydration-responsive genes. Because LEA and LEA-like genes are among the most up-regulated during stress, we used two highly regulated probesets as positive controls: Ljwgs_145133.1 (LEA group 3) and chr1.TM0221.11 (LEA group 4). Expression of both genes was induced in all tissues tested under salt acclimatization, salt shock and drought (Figure 5a–c).

Among the most down-regulated genes (Figure 4b), the probeset Ljwgs_007448.1 was of particular interest because of the sequence similarity to PsRMS1 and AtMAX4 (89% and 83%, respectively). These orthologues encode the carotenoid-cleaving deoxygenase that is implicated in the control or generation of a long-range transmissible branching signal (Foo et al., 2005). Expression of Ljwgs_007448.1 was low in developing leaves and equally high in stems and mature leaves (Figure 5b). Salt acclimatization resulted in strong reduction, but a similar trend was observed under salt shock and drought. Moreover, reduction of expression under salt stress was strongest in mature leaves compared to stems and developing leaves (Figure 5a,c). As RAMOSUS/MAX genes interact with auxins and cytokinins for control of morphological patterns, transcriptional down-regulation under stress of this putative LjRMS1 may indicate a novel link between stress-induced signalling and morphogenetic responses in legumes.

Metabolite profiling

Profiling of soluble metabolites was performed for all treatments in the three independent experiments using gas chromatography/mass spectrometry technology (GC/EI-TOF-MS). Probabilistic principal component analysis (PPCA) was applied to the complete set of observed mass fragments without prior knowledge of metabolite identity (Figure 6a). This fingerprinting analysis revealed a clear trend of metabolic re-adjustment upon salt acclimatization, illustrated by the first two principal components covering the major variance of the data set. The change in metabolic phenotype was clearly coincident with the increasing salt dose. Sample-to-sample variation, however, also increased substantially at high salt concentrations.

image

Figure 6.  Overview of metabolic profile results. (a) Metabolic fingerprinting displayed using probabilistic principal component analysis (PPCA). The first two principal components are shown, which comprise the major variation of the data set. Symbols indicate control treatment (open boxes), gradual step acclimatization (black) and initial acclimatization (grey). (b) Chemical categories for the soluble, salt-responsive metabolites. (c) Representative patterns of relative pool size changes from the applied chemical categories. Each symbol represents the mean of three independent experiments. Error bars were omitted for improved visualization. The dotted horizontal lines indicate relative control levels.

Download figure to PowerPoint

A screen of the metabolic profiles for statistically significant changes (FDR < 0.01) in metabolite pool sizes upon salt acclimatization revealed 147 mass spectral tags; 88 accumulating and 59 decreasing (for definition, see Desbrosses et al., 2005). Due to the restricted availability of reference substances, only approximately one third of the metabolites can currently be identified. As a general metabolic trend, we found a strong decrease in most organic acids and a few amino acids, concomitant with increases of multiple amino acids, sugars, polyols and specific organic acids (Figure 6b). Notable exceptions to these changes were the increase in glucuronic and gulonic acids and the decrease in the primary products of nitrogen assimilation, namely glutamine and asparagine (Table S4). Interestingly, some metabolites related to general stress responses in other species did not accumulate in salt-stressed L. japonicus plants. For example, myo-inositol levels did not change, while levels of β-alanine and galactinol were actually slightly decreased (Table S4).

Selected examples of the metabolic changes demonstrated a qualitative similarity to the observed changes in patterns of gene expression (Figure 6c). In most cases, altered metabolites showed strict positive or negative dose dependency, although some changed only at elevated salt concentrations, demonstrating dose-threshold behaviour. Other metabolites appeared to reach an upper or lower pool size, which is particularly evident for most of the organic acids shown in (Figure 6c). This was not caused by analytical limitations resulting from saturation of chromatographic or mass detector capacity.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Nutrient homeostasis under salt acclimatization

Salinity-induced nutritional disorders are typically discussed as deficiencies or changes in the requirement for nutrients. However, the influence of salt stress on plant nutrition is highly variable, and depends on the genotype, tissue, growth conditions and chemical characteristics of the soil (Grattan and Grieve, 1999). Exposure of L. japonicus plants to increasing concentrations of salts led, as expected, to increases in shoot Na+ with concomitant decreases in shoot K+ (Figure 2). Reduction in plant K+ is the most commonly recognized nutritional change under salt stress, and has been implicated in growth and yield reduction in crops (Grattan and Grieve, 1999). However, many glycophytes are able to substitute Na+ for K+ without negative effects on growth (Marschner, 1995). Increases in calcium, magnesium and manganese rule out an NaCl-induced deficiency of these elements in L. japonicus shoots under our experimental conditions (Figure 2). The same applies to iron and boron, which appeared to be under homeostatic control. However, we found decreased concentrations of zinc, phosphorus and sulphur under severe salt stress (Figure 2). Salt-induced deficiency of these elements has been reported previously in other species, and could arise from changes in nutrient availability, uptake, transport or partition within the plant (Grattan and Grieve, 1999). As it is generally accepted that an increased supply of nutrients does not necessarily improve the growth of salt-stressed plants under nutrient-sufficient conditions, it is difficult to assess whether a real deficiency exists in our experiments (Grattan and Grieve, 1999). In addition, molybdenum was the most salt-sensitive of all micronutrients in our ionomic profile, decreasing even under mild treatments (Figure 2). To our knowledge, a connection between salinity and Mb content has not been recognized previously. Currently, we have no evidence that salt-acclimatized L. japonicus plants reach critical molybdenum deficiency levels, and further analyses will be required to assess this possibility. Interestingly, shoot nitrate reductase activity vital for nitrate assimilation has been shown to decrease under salt stress in tomato (Debouba et al., 2007). As molybdenum is a co-factor of nitrate reductase, our data may provide an explanation for these results.

Salinity-induced changes of gene expression

Major changes in the expression of genes involved in amino acid metabolism as well as nitrogen and organic compound transport, including putative sucrose, amino acid and organic acid transporters, indicated profound changes in central metabolism under long-term salt stress. General metabolic changes were also reflected within the miscellaneous gene group, which included large enzyme families including peroxidases, lipases, glucosidases, glycosyl- and glutathione-S-transferases, cytochrome P450s and oxidases, linked to a myriad of cellular processes (Figure S1). Further global control of the acclimatization response is reflected by changes in the transcription and RNA processing, signalling and hormone metabolism categories (Figure 4c).

We identified new molecular candidates that may represent important factors in salt acclimatization in legumes, and further analysed the expression of selected genes within the transcription, transport and signalling functional groups. We also identified a multitude of genes from L. japonicus that are homologous to genes that are stress-regulated in other species, not only in the stress and defence-related group but also in other functional categories. For example, many probesets representing genes from secondary metabolism were transcriptionally regulated by salt stress, including flavonoid, phenylpropanoid and phenol metabolism (Figure S1), which have been previously correlated with biotic and abiotic stress responses (Kliebenstein, 2004; Walia et al., 2005). In addition, we found many transcriptionally regulated genes relating to the cell wall, including expansins, cellulose synthases and glycosyl transferases. In contrast to results reported for rice (Walia et al., 2005), most L. japonicus genes in this functional group were down-regulated by salt stress (Table S2 and Figure S1). This observation may be a consequence of the decreased growth and reduced requirement for cell-wall synthesis under long-term salt acclimatization.

Some transcriptional changes were reflected in the metabolomic data. For instance, proline accumulation in plants is common under salt stress (Figure 6c), and seven up-regulated probesets encode proteins that are putatively involved in proline metabolism: Δ-1-pyrroline 5-carboxylase synthetases (P5CS, probesets Ljwgs_006172.2, Ljwgs_032463.1, Ljwgs_053689.1 and chr1.CM0147.99) and Δ-1-pyrroline-5-carboxylate dehydrogenases (P5CDH, probesets chr4.CM0170.37, Ljwgs_019593.1_s and Ljwgs_052588.1_s). We also found induction of two probesets encoding putative myo-inositol-1-phosphate synthases (Ljwgs_091497.1_s and chr4.CM0307.12), concomitant with changes in osmoprotectants from the inositol family (see below). The activity of this enzyme was found to be a rate-limiting step in the biosynthesis of inositol-containing compounds, and is known to be involved in the salt-stress responses of halophytes (Nelson et al., 1998). The depletion of asparagine may be under transcriptional control, as indicated by down-regulation of asparagine synthase 1 (LjAS1, probesets gi897770, gi897770_s and chr5.CM0071.60_s) and up-regulation of two putative asparaginases (Ljwgs_021574.1 and chr5.CM0096.107). In addition, the slight decrease in galactinol was paralleled by down-regulation of a putative galactinol synthase (chr1.CM0122.56).

Approximately one third of the probesets identified in L. japonicus showed a significant hit (E value ≤1e-5) when matched to those reported under long-term salt stress in A. thaliana, suggesting a certain degree of inter-species similarity in the molecular responses (Table S5) (Sottosanto et al., 2004). Future comparative systems analysis under well-controlled environmental and nutritional conditions will be required to unravel inter-species conservation of salt-stress acclimatization mechanisms.

Salinity-induced changes of the metabolic phenotype

Results from the non-targeted metabolite profiling demonstrated a major and reproducible change of the metabolic phenotype in the course of salt acclimatization, which was most evident for amino acid, sugars and polyols and organic acid metabolism (Figure 6c).

Accumulation of amino acids and other nitrogen-containing compounds is a remarkable biochemical feature of almost all plant stress responses reported so far. This change in nitrogen metabolism has been interpreted as an accumulation of compatible solutes, generation of carbon and nitrogen reserves for future needs, a sink for detoxification of excess nitrogen or for redox potential cycling (Gilbert et al., 1998; Rabie, 1999). This broad response reflects a tightly controlled metabolic shift, and is inconsistent with nitrogen deficiency as a mechanism of salt injury (Grattan and Grieve, 1999). Although it has been extensively shown that nitrogen content may be affected under salinity due to alterations in NO3 uptake, non-nodulated salt-stressed legumes exhibited accumulation of NH4+ and even increased total nitrogen content (Huq and Larher, 1983; Speer et al., 1994). Therefore, the accumulation of nitrogen-containing compounds may represent a response to a decrease in nitrogen demand caused by reduced growth rates under stress. This contention is in line with the hypothesis of reduced assimilation of inorganic nitrogen in salt-acclimatized L. japonicus plants (see below). In addition, the enhanced expression of most of the genes involved in amino acid metabolism (Figure S1) indicates the requirement for de novo synthesis already observed in other plant species (Gilbert et al., 1998).

Notable exceptions to the general amino acid behaviour were the two amides glutamine and asparagine. The observed decrease in both glutamine and glutamate may suggest a reduced capacity of NH4+ assimilation through the glutamine synthetase/glutamate synthase pathway, as they play a pivotal role in this process (Forde and Lea, 2007). Indeed, these enzymatic activities have been reported to decrease under salinity (Debouba et al., 2007). Although we did not find any regulation of the genes involved in this pathway, we demonstrated consistently reduced expression of two probesets encoding putative nitrite reductases (gi9968472 and chr4.CM0227.40_s, Table S2), which are also involved in nitrogen assimilation, and also down-regulation of genes encoding putative NO3 and NH4+ transporters. In contrast, the depletion of asparagine was paralleled by transcriptional changes of genes putatively involved in its metabolism (see above). As a major nitrogen transport compound of L. japonicus (Waterhouse et al., 1996), the transcriptional control of asparagine levels in the shoot may also reflect decreased inorganic nitrogen assimilation.

Along with nitrogen-containing compounds, sugars and polyols also increase under stress and are known to have protective roles as osmoprotectants (Munns, 2005; Orcutt and Nilsen, 2000). Several compounds related to these chemical groups accumulated in salt-acclimatized L. japonicus plants, including the disaccharides maltose and sucrose and the polyols arabitol and erythritol (Figure 6c and Table S4). We also identified the salt-induced methylated inositols pinitol and ononitol, in line with transcriptional changes of the myo-inositol pathway as described above.

The general depletion in organic acids observed in L. japonicus shoots has been demonstrated previously in salt-stressed roots and nodules of the legume alfalfa (Fougere et al., 1991). This phenomenon may reflect an increased energy demand that is met by intensified respiration, carbon allocation to amino acid or sugar pools that are required as compatible solutes, or transport from shoots to roots as carbon source. It is interesting to consider a possible role in compensation for uneven charges entering the plant; organic acids are known to counterbalance unequal uptake of ions as they occur as carboxylic anions under physiological pH (Hinsinger et al., 2003; Marschner, 1995). Moreover, many organic acids accumulate in other species under a variety of stress regimes, such as drought, cold and heat, suggesting that this metabolic change is due to ionic misbalance rather than a decrease in the amount of fixed carbon under stress conditions (Kaplan et al., 2004; Timpa et al., 1986). Notable exceptions to the general depletion of organic acids were glucuronic and gulonic acids (Table S4). As glucuronate is involved in the myo-inositol oxidation pathway that synthesizes nucleotide sugars for cell-wall polysaccharides, the accumulation of glucuronic acid may reflect an inhibition of cell-wall biosynthesis due to decreased growth (Kanter et al., 2005). Alternatively, it might represent an important feature of ascorbic acid metabolism, in line with the parallel increase in gulonic acid. Glucuronic and gulonic acids are intermediates of the uronic pathway that synthesizes ascorbate from myo-inositol (Ishikawa et al., 2006).

Gradual step increase in salts compared to initial acclimatization

The salt-stress physiology of glycophytes is currently interpreted in terms of the biphasic growth model (Munns, 2002, 2005). In essence, this model proposes two growth inhibition phases in response to a gradual increase of salinity. In the first phase, growth inhibition is caused by decreased osmotic potential and thus reduced water availability. After prolonged exposure to salinity, accumulation of ions within plant tissues triggers a second mode of growth inhibition caused by ion toxicity. The biphasic growth model formalizes three essential aspects of salinity: (i) that plant salt-stress physiology ultimately depends on toxicity of ions per se, (ii) that the time of exposure is an important variable, and (iii) that appropriate experimentation requires long-term progressive acclimatization to differentiate between (a) hyperosmotic and hyperionic stresses, and (b) physiological acclimatization responses and cellular damage or senescence. In this work, we compared the experimental approach of the biphasic growth model, i.e. the conventional gradual step increase design, to an alternative experimental design based on initial acclimatization. The rationale behind the initial acclimatization is that, given a particular genotype and environment, a range of non-lethal soil salt concentrations should exist that allow the genotype to geminate and develop. Such an experimental design mimics natural and agricultural topsoil-associated salinity, where not only successful plant growth but also establishment on salt-rich soils is required.

Nutrient, transcriptome and metabolome profiling data were used to test whether both experimental designs equally address salt-stress physiology. Firstly, non-supervised analysis suggested that changes do not cluster according to the experimental approach but to the salt-stress dose (Figures 3a and 6a). Secondly, we found only rare statistical evidence for a qualitative differentiation between the acclimatization regimes. For example, under ia50 and ia75 treatments, zinc content decreased while glucose and fructose were increased, a behaviour that is not observed in gradual acclimatization (Figures 2 and 6c). However, it cannot be ruled out that such a difference is due to the differential stress doses perceived by plants between the two designs, as described previously (see Results). Other changes showed only quantitative trends differing between the experimental approaches, with a few changes arising specifically in ia75 plants, probably due to the highly stressful nature of this treatment. Considering the broad scope of the profiling techniques used herein, it could be argued that no major differential effects were observed between the gradual acclimatization and the initial acclimatization experiments. Our results reveal a general equivalence of both salt regimes, supporting the use of the gradual acclimatization approach and the biphasic growth model interpretations, despite previous concerns raised against them (Neumann, 1997).

The stress acclimatization process requires fine tuning of responses

Given that plants acclimatized to a non-lethal salt-stress dose may have reached a stable physiological state essential for survival, it could be argued that most, if not all, of the molecular and metabolic responses reflect the set of plastic physiological changes that, as a whole, allow the plant to cope with the environmental constraint (Lichtenthaler, 1996). If this is the case, how are the various traits regulated and coordinated within the plant in response to changes in the stress dose? From a molecular perspective, a temperature-dependent adjustment has been described for expression of CBF/DREB transcription factors controlling cold-stress responses in A. thaliana, demonstrating that stress sensing and response mechanisms are not binary on/off systems (Zarka et al., 2003). Such ‘rheostat’ control of responses may be interpreted in a simple deterministic model, where all available traits needed to cope with the stress are elicited in a strictly linear dose-dependent manner (Figure 7a). Based on the results of our work, two refinements of this model are necessary (Figure 7b). First, some responses may reach a systems constraint, such as a concentration plateau (trait A). These constraints may be of a genetic or metabolic nature and reflect limitations in the minimal or maximal elicitation of transcript or metabolite pools. Second, some responses are not required at low salt concentrations and will be activated only when a threshold is reached (trait C), suggesting that sensitive responses at the molecular and metabolic systems level are sufficient to compensate for the induced change under low stress intensities. We propose that most major plastic changes under salt acclimatization may be qualitatively explained by this fine tuning model, which includes linear, plateau and threshold dose-dependent responses (Figure 7b). Whether this model may fundamentally be applied to other abiotic stresses remains to be determined.

image

Figure 7.  Models of dose-dependent salt-stress acclimatization responses. (a) Strictly linear dose-dependent responses. (b) Fine-tuning model, including linear, plateau and threshold dose-dependent responses. A, B and C represent measurable molecular and metabolic plant traits.

Download figure to PowerPoint

In conclusion, we performed a systems investigation of salt acclimatization in L. japonicus, designed to be as non-biased and comprehensive as possible given the current technological limitations. We found a complex pattern comprising ionomic, transcriptomic and metabolomic responses to salt stress, some of which have not been described in plants before. In addition, we showed that the general transcriptional regulation under long-term salinity is mainly dominated by ion accumulation and toxicity rather than osmotic effects, in line with the proposed physiological biphasic growth model (Munns, 2002, 2005). Finally, we demonstrated the need to refine simple dose-dependence models of salt acclimatization. We now venture to predict that molecular and metabolic differences in acclimatization responses between tolerant and sensitive cultivars may be characterized by three possible features within the framework of the fine-tuning model. Increased tolerance may arise from: (i) changes in the slope of dose-dependent responses, (ii) changes in the upper or lower concentration constraints of a response, or (iii) a shift in the stress dose threshold of factors that do not respond at low stress levels.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant material, growth conditions, experimental design and sampling

Seeds of Lotus japonicus var. Gifu were germinated on agar plates containing agarified half-strength BD solution (Broughton and Dilworth, 1971) plus 2 mm KNO3 and 2 mm NH4NO3. Four days after imbibition, seedlings were transplanted to soil (Einheit, type null) in 10 cm pots, and irrigated with the above solution. Two salt-stress treatments were implemented: (i) gradual step acclimation to various final NaCl concentrations, or (ii) initial acclimation growth at various NaCl concentrations (see Figure 1a). The gradual acclimation started 8 days post-imbibition, and the salt concentration was increased in three steps of 4 days from 0 to 50, 100 and 150 mm NaCl. A subset of plants was kept at each salt level. Initial acclimation growth exposed plants from germination onwards to nutrient solution supplemented with 25, 50 and 75 mm NaCl. In both cases, fresh nutrient solution was prepared every 4 days. The total duration of greenhouse culture was 28 days under a 16/8 h day/night regime, 23 ± 2°C, 55–65% relative humidity. Whole shoots, excluding cotyledons, were sampled in situ into liquid nitrogen in the middle of the light period. Three successive independent experiments (experiments 1, 2 and 3) were performed during the spring season. Each experiment consisted of seven sample sets: control (no salts), 50, 100, 150 (gradual step acclimatization), ia25, ia50 and ia75 (initial acclimatization). Each set had seven independent biological replicate pools of four plants, with the exception of ia75 treatment in experiments 2 and 3 where fewer replicates were available (five and four, respectively). The stress doses used in the experimental designs were not lethal within the cultivation period, and at harvest all plants were in the vegetative stage, and roots did not show nodules. Growth was estimated by determination of mean fresh weight of the pooled shoots, and soil electrical conductivity was determined on 1:2 dried soil:water extraction.

An independent experiment was conducted for verification of the expression patterns of selected target genes, comprising control, gradual acclimatization (150 mm NaCl) and drought-treated plants. In the latter, irrigation was stopped at day 15 after imbibition (the soil water content was 3.79 ± 0.08 and 0.96 ± 0.02 g H2O g−1 dry soil for control and drought treatments, respectively). Whole shoots or separate pools of mature leaves, stems and developing leaves were harvested. In addition, in vitro salt-shock experiments were performed using 7-day-old seedlings grown in agarified MS medium without sucrose (Murashigue and Skoog, 1962). Independent biological replicates, containing at least 50 seedlings each, were sampled 24 h after exposure to 250 mm NaCl together with non-treated controls.

Nutrient profiling analysis

For micro- and macronutrient profiling, 100 mg of plant material was digested with 2 ml HNO3 (Merck; http://www.merck.de) at 140°C until complete digestion. Then 100 μl of a 100 g l−1 LiCl solution (Fluka; http://www.sigmaaldrich.com) was added as a modifier, and the final volume adjusted with ultra-pure water to 10 ml. Element concentrations were determined by inductively coupled plasma/atomic emission spectrometry (ICP-AES) using an IRIS Advantage Duo ER/S (Thermo Fisher; http://www.thermofisher.com). The element emission lines used were: B_2089, Ca_3181, Fe_2599, K_7698, Mg_2790, Mn_2605, Mo_2020, Na_5889, P_1782, S_1820 and Zn_2062. Elemental quantification was validated using IC-CTA-VTL2 Virginia tobacco leaves as a certified reference material (Institute of Nuclear Chemistry and Technology; http://www.ichtj.waw.pl).

Gene expression analysis

Sample tissue for all the biological replicates were pooled to obtain representative RNA for each independent experiment. Total RNA was isolated using the hot borate method (Wan and Wilkins, 1994), and quality and quantity were assessed using a Bioanalyzer-2100 with RNA 6000 NanoChips (Agilent Technologies; http://www.agilent.com) and a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies; http://www.nanodrop.com), respectively. For the microarray experiments, RNA was labelled using the One-Cycle Target labelling kit, hybridized to the Genechip® Lotus1a520343 and scanned, according to the manufacturer’s instructions (Affymetrix; http://www.affymetrix.com).

For quantitative real-time RT-PCR analysis, total RNA was DNAseI-treated with TURBO DNA-free (Ambion; http://www.ambion.com) and first-strand cDNA was synthesized using an oligo(dT) primer and SuperScript III transcriptase (Invitrogen, http://www.invitrogen.com/). Real-time PCR was performed using 2 × SYBR Green I PCR Mastermix and an ABI Prism 7900HT sequence detection system (Applied Biosystems, http://www.appliedbiosystems.com/). Primer design, reaction conditions, cycling and dissociation-curve parameters, DNA contamination and 3′ to 5′ ratio checks were performed as described by Czechowski et al. (2005), without spike-in control and using a reaction volume of 10 μl. Amplification efficiency was assessed using the LinRegPCR program (Ramakers et al., 2003). Analysis of expression data was performed as previously described (Czechowski et al., 2004, 2005), using the geometric mean of four housekeeping genes for normalization (Vandesompele et al., 2002). The housekeeping genes were LjUBQ4 (chr5.CM0956.27), LjGPI-anchored protein (chr3.CM0047.42), LjPP2A (chr2.CM0310.22) and LjUBC10 (chr1.TM0487.4), which were selected from the most stably expressed genes in the plants (Czechowski et al., 2005). A list of all primers used is provided in Table S6.

Metabolic profiling analysis

Frozen plant tissue (60 mg) was extracted using methanol/chloroform, and the polar fraction was prepared by liquid partitioning into water and derivatized (Desbrosses et al., 2005). Gas chromatography coupled to electron impact ionization/time-of-flight mass spectrometry (GC/EI-TOF-MS) was performed using an Agilent 6890N24 gas chromatograph with split or splitless injection connected to a Pegasus III time-of-flight mass spectrometer (LECO Instrumente GmbH; http://www.leco.de) (Wagner et al., 2003). Metabolites were quantified after mass spectral deconvolution (ChromaTOF software version 1.00, Pegasus driver 1.61, LECO) of at least three mass fragments. The peak height representing arbitrary mass spectral ion currents was normalized using the sample fresh weight and ribitol content for internal standardization.

Metabolites were identified using NIST05 software (http://www.nist.gov/srd/mslist.htm) and the mass spectral and retention time index (RI) collection of the Golm metabolome database (Kopka et al., 2005; Schauer et al., 2005). Mass spectral matching was manually checked, and accepted with thresholds of match >650 (maximum 1000) and RI < 1.0 %. RIs represent Kovàts indices (Kovàts, 1958) calculated from additions of C12, C15, C19, C22, C32, C36n-alkanes. Table S4 lists not only metabolites identified by standard addition but also mass spectral tags that are as yet unidentified (Desbrosses et al., 2005).

Data analysis and statistics

Hierarchical clustering/supporting trees (HCL-ST) and probabilistic principal component analysis (PPCA) were used as clustering algorithms to analyse data in a non-supervised approach, using TIGR multiple experiment viewer software (TMEV_3.1) and the MetaGeneAlyse webpage (http://metagenealyse.mpimp-golm.mpg.de).

Ionomic and metabolomic data were log10-transformed prior to statistical analysis, which was performed by anova using the following linear model for each of the two experimental designs separately: = β0 + β1 + β2 + ε, where y is the measured intensity, t indicates treatment or control, e indicates the experimental block, and ε is the error. The P values associated with the null hypothesis H0: β1 = 0 for every nutrient or metabolite were extracted using Student's t-test. A summary P value (Pcombined) for both experimental designs was obtained using H0: β1,step = 0 or β1,ia = 0 and Pcombined = 2 × min (Pstep, Pia). The false discovery rate (FDR) correction (Benjamini and Hochberg, 1995) was applied to the Pcombined value, and nutrients or metabolites were declared significant if FDRcombined < 0.01. Both HCL-ST and PPCA were tested for outlier detection, and the biological replicates exp1_ai25_7, exp2_Ctrol_3, exp2_ai50_2, exp2_50_1, exp3_Ctrol_7, exp3_100_7 and exp3_150_7 were omitted from the metabolomic profile (not shown).

Microarray data were analysed using the bioconductor software package for the R programming language (Gentleman et al., 2004). Data quality was assessed using the affy (Gautier et al., 2004) and AffyPLM packages, and expression estimates were obtained using the RMA algorithm (Irizarry et al., 2003). Control and bacterial probesets were removed, and only genes assigned as present (< 0.05) using the MAS5 present/absent algorithm were retained. Statistical testing for differential expression was performed using mixed models with the LIMMA bioconductor package (Smyth, 2004). values describing control versus treatment comparisons were corrected for multiple testing using the FDR (Benjamini and Hochberg, 1995). Raw data are deposited at Array-Express (http://www.ebi.ac.uk/arrayexpress) as E-MEXP-1204.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

This work was conducted within the framework of the European LOTASSA project (INCO-CT-2005-517617). We greatly acknowledge the long-standing support of all directors at the Max Planck Institute for Molecular Plant Physiology (MPIMP), the technical assistance of Ines Fehrle for GC/EI-TOF-MS profiling analyses, and the MPIMP ‘greenteam’, particularly Britta Hausmann who was in charge of the legume greenhouse. We would like to thank Dr Florian Wagner and the German Resource Center for Genome Research (RZPD, Berlin) team for expert microarray hybridization, Dr Björn Usadel for support with the MapMan software application and for sequence matching, and Dr Ina Talke for support with ICP analysis. D.H.S., F.L., J.K. and M.U. would also like to thank Dr Armin Schlereth for his selfless commitment to our research group and valuable day-to-day assistance.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Figure S1. MapMan overview windows for metabolism, large enzyme families (miscellaneous) and transport.

Table S1. Nutrient profile data for all detectable elements.

Table S2. Transcriptomic profile data.

Table S3. Data used to produce (Figure 3c).

Table S4. Metabolic profile data.

Table S5. Sequence matching of L. japonicus salt-responsive probesets against A. thaliana salt-responsive genes, and list of L. japonicus probesets showing a significant hit.

Table S6. Primers used for quantitative real-time RT-PCR.

Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
TPJ_3381_sm_TPJ_3381_Fig_S4a.jpg126KSupporting info item
TPJ_3381_sm_TPJ_3381_Fig_S4b.jpg52KSupporting info item
TPJ_3381_sm_TPJ_3381_Fig_S4c.jpg55KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S1.xls17KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S2.xls265KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S3.xls16KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S5.xls45KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S6a.xls1014KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S6b.xls29KSupporting info item
TPJ_3381_sm_TPJ_3381_Table_S7.xls25KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.