INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

High salinity causes pleiotropic effects in plant growth such as reduced cell expansion, decreased protein synthesis, and accelerated cell death. Soil salinity is one of the major limitations of crop productivity worldwide. Halophytes are native flora of saline environments that have the characteristic to overcome the ion and osmotic imbalance caused by high NaCl concentrations. Halophytes possess a set of unique salt adaptation mechanisms. Through comparisons with glycophytes (nonhalophytes), the mechanisms of salt tolerance in halophytes have been studied at physiological, biochemical, and molecular levels (for review, see Hasegawa et al., 2000). In general, three adaptation strategies are commonly found in halophytes: compartmentation of toxic ions, accumulation of osmolytes, and conservation of water (Bohnert et al., 1995). The main strategy for glycophytes is to control of ion flux into root xylem and as the result, restrict ion movement to the shoot (for review, see Hasegawa et al., 2000).

Ice plant (Mesembryanthemum crystallinum) has been used as a halophytic model because the mechanisms of salt tolerance can be induced when the plants reach a certain developmental stage. Although ice plant is not suitable for genetic manipulation, comparisons of the physiological, biochemical, and gene expression changes before and after salt stress has provided useful data. In a time course progression of salt-induced changes in ice plant, the increased level of Pro was detected as early as 3 h after salt stress (Michalowski et al., 1989), the expression of osmolyte production genes Inps and Imt started 12 h after stress (Ishitani et al., 1996), the expression of potassium channel (Su et al., 2001) and gene MipA encoding a water channel protein decreased in the first 30 h and recovered 3 d after the salt stress (Yamada et al., 1995), and the induction of transcripts for Crassulacean acid metabolism pathway Ppc1 and Ppdk occurred 2 d after salt stress (Cushman et al., 1990). Based on the northern-typed analyses, the expression of these salt-induced mRNAs did not change significantly within the first 30 h of stress (Vernon et al., 1993). Studying the expression of salt-induced genes has provided a detailed network of the salt adaptation mechanism in ice plant (Bohnert et al., 1995). However, basic questions such as the length of time it takes for Na⁺ ions to reach the photosynthetic organs and the effectiveness at which ice plants make an appropriate response to the increased Na⁺, remained largely unknown. In this article, we examined the early effect of salt stress on the chemical composition and structural details of ice plant leaves using Fourier transform infrared (FT-IR) spectroscopy. This technique is highly suitable for the examination of initial response to stress because the acquisition time of FT-IR spectrometry can be shorter than a second (Griffiths and de Haseth, 1986). Effects of salt stress on Arabidopsis were also examined to compare the different responses to high salinity between halophytes and glycophytes.

The vibration of chemical bond absorbs radiation in the IR region between 4,000 and 400 cm¹. Each functional group in a molecule has characteristic absorption frequencies in the IR spectrum (Griffiths and de Haseth, 1986). The sensitivity of IR spectroscopy has been successfully applied to in vitro and in vivo detection of biological systems. During chemical extraction of plant cell walls, components and possible crosslinks of each fraction were identified by FT-IR microspectroscopy (McCann et al., 1992; Séné et al., 1994). The IR method provides a unique way to study the conformation of proteins (Susi et al., 1967). The C=O, -NH₂, and C-N bonding of the peptide linkage absorbs radiation in the 1,800 to 1,200 cm¹ region. The absorption band of C=O stretching vibrations of the amide group depends on the nature of hydrogen bonding between C=O and N-H moieties and is particularly useful for determining the secondary structure of a polypeptide chain (for review, see Surewicz et al., 1993). In addition to the purified proteins, the secondary structures of proteins in complex biological samples have also been analyzed. This noninvasive method has been applied to detect changes of the overall protein secondary structure during dehydration in maize (Zea mays) embryo (Wolkers et al., 1998) and pollen (Wolkers and Hoekstra, 1995).

Mathematical approaches such as deconvolution and curve fitting are generally applied to extract information from the raw IR spectra to resolve the overlapping band components in the heterogeneous matrix. Deconvolution is a band-narrowing technique that can enhance small features buried in an overlapped band. Comments about this technique can be found in the review article by Surewicz and Mantsch (1988). The curve-fitting technique enables further quantitative analysis of individual bands buried in an overlapping band. A series of Gaussian or Lorenzian curve shapes is used to compose a synthesized spectrum. By subtracting the raw spectrum from the synthesized spectrum, the residual can be used to determine the fitness between synthesized and raw spectrum. In a statistical manner, the number of synthesized spectrum with low residual may be more than one. To reduce the possibility of fitting the curves in unwanted directions, the band positions obtained by deconvolution spectrum are typically used (Surewicz and Mantsch, 1988). Decomposition of the amide I band by curve fitting into its constituents and the assignment of these components to a protein structure has been successfully applied in predicting the structure of membrane proteins (for review, see Arrondo and Govi, 1999).

Given the complex biological matrix, the changes of a specific functional group cannot be assigned to a particular molecule in the cells. However, the changes in chemical composition reflect the overall changes in the metabolic processes, and can be detected more sensitively than the traditional northern or western-blotting methods. In this report, the early salt-induced changes in chemical constituent and protein conformation in ice plant and Arabidopsis were detected by FT-IR spectrometry. Leaves were dried to remove the spectral interference caused by blanket absorption of water over most of the IR region. The changes in amount of cell wall pectin and carbohydrate were calculated according to the band intensities of the original IR spectra. The IR signals were further processed by deconvolution and curve fitting to reveal the rapid changes in protein conformation. The differential responses to salt stress underline the different strategies to high salinity between halophytes and glycophytes.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Changes in Sodium Content and Relative Water Content (RWC) at an Early Stage of Salt Stress

Long-term salt-induced accumulation of Na⁺ in the leaves and xylem sap has been measured in the halophyte ice plant (Demmig and Winter, 1986; Adams et al., 1992), but the short-term change of Na⁺ has not yet been reported. To monitor the changes of Na⁺ and water content in the leaves at an early stage of salt stress, pot-grown 5-week-old ice plants were irrigated with 200 mM NaCl solution. Under continuous illumination, leaves were collected hourly for 24 h. A transient increase of Na⁺ in the leaves was observed at the 1st h of stress along with a marked decrease in RWC (Fig. 1A). After 3 h of salt stress, Na⁺ content in the leaves rapidly decreased to a level lower than the Na⁺ content of prestressed condition. At the same time, the RWC of the leaves returned to the prestressed condition. After 24 h of stress, the Na⁺ level decreased about 50% and the RWC increased to about 2% of the prestressed level. After 2 to 3 d of continuous stress, Na⁺ content returned to a level that was slightly higher than the prestressed level (data not shown). The results indicated that Na⁺ ions were rapidly taken up by roots and reached the leaves within 1 h, and that the Na⁺ efflux system was actively engaged to prevent Na⁺ toxicity of the mesophyll cells.

View larger version (21K): [in this window] [in a new window]   Figure 1.   Effect of salt on the sodium content and RWC in ice plants and Arabidopsis. Under continuous illumination, 5-week-old ice plants (A) and 3-week-old Arabidopsis (B) were irrigated with 200 and 50 mM NaCl, respectively. Leaves collected from at least three different plants were pooled as one sample.

To assess the different strategies of Na⁺ detoxification between halophytes and glycophytes, 3-week-old Arabidopsis plants were treated with 50 mM NaCl solution. This level of salt did not cause an immediate death to the plants, but the acceleration of leaf senescence and branched inflorescence was observed (data not shown). The Na⁺ concentration in the nonstressed Arabidopsis leaves was 50 times lower compared with that of ice plant (1.83 versus 86.5 mg g¹ dry weight). Unlike the fluctuation observed in the ice plant, there was no change of Na⁺ content and RWC in the leaves after the 1st h of stress (Fig. 1B). The concentration of Na⁺ started to increase at 3 h, and a 4-fold increase was observed at the end of the 24 h. As the stress persisted, there was a 14-fold increase of leaf Na⁺ content after 48 h (data not shown). The RWC was negatively correlated to the Na⁺ content by a 2% decrease after 24 h of stress. The results show that even under mild salt stress, the capacity to prevent Na⁺ reaching the photosynthetic tissues and maintain leaf water status is limited in Arabidopsis.

Identification of the Chemical Composition at Different Developmental Stages of Ice Plants

The feasibility of FT-IR spectrometry for obtaining the chemical information from leaf was investigated. Typical IR spectra obtained from three developmental stages (seedling, juvenile, and adult) of the ice plant are shown in Figure 2. Primary leaves taken from seedlings generated large numbers of sharp peaks in the mid-IR region (2,000-1,000 cm¹), indicating that primary leaves have a rich chemical composition. Several absorption regions were identified, and the band assignments are labeled in Figure 2A. Absorption bands located around 3,400 cm¹ correspond to O-H and N-H stretching vibrations that mainly occur from proteins and carbohydrates. The bands around 3,000 cm¹ represent C-H stretching vibrations that are mainly caused by lipid and carbohydrates. Absorption raised from C-H bending modes was located around 1,200 to 1,500 cm¹, but overlap with other absorption bands within this region. Three protein absorption bands located around 1,680 (C=O); 1,550 (N-H); and 1,250 (C-N) cm¹ were assigned as amide I, II, and III bands, respectively. Absorption bands around 1,745 cm¹ correspond to isolated carbonyl group (COOR), indicating ester-containing compounds commonly found in membrane lipid and cell wall pectin. Bands raised by acids (-COO) were located around 1,600 and 1,420 cm¹, but were seriously overlapped by other bands in these regions. Bands around 1,100 cm¹ in the "fingerprint" region indicate several modes such as C-H bending or C-O or C-C stretching. Carbohydrates in the leaves were the major constituents that contributed to these absorption bands.

View larger version (22K): [in this window] [in a new window]   Figure 2.   A, Absorption FT-IR spectra in the 4,000 to 1,000 cm1 region in three stages of ice plants. Characteristic group frequencies are indicated at the top. B, Deconvolved absorption FT-IR spectra in the 2,000 to 1,000 cm1 region in three stages of ice plants. The major chemical constituents in dried leaves that contribute to the formation of bands at specific wave numbers are indicated at the top. C, The assignment of absorption bands to their major chemical components in the leaves. Dried ice plant leaves were taken from seedlings (1-week-old), juvenile (5-week-old), and adult (10-week-old; leaves emerged from side branches) plants. All plants were grown in a low-salt soil. Samples were collected from at least three plants of each stage and at least three spectra were obtained for each sample. Only one representative spectrum of each stage is shown.

Two significant observations are presented in Figure 2A. First, the amides II band of protein around 1,550 cm¹ was prominent in seedlings and decreased in band intensity at the juvenile and adult stages (Fig. 2A, arrowhead). This suggests the primary leaves are protein rich. Second, leaves at the juvenile and adult stages generated similar IR spectra and the peak shapes were extensively broader than the spectrum obtained from the seedling stage, especially in the 1,000 to 2,000 cm¹ region. This result suggests that new molecules are produced during development and/or that the interactions between the existing molecules became stronger.

To distinguish the small difference in patterns caused by overlapping bands, a deconvolution technique was used. The results of deconvolved spectra in the region between 1,000 and 2,000 cm¹ are shown in Figure 2B. Deconvolution enhanced the resolution of small bands that were buried under other bands in the original spectra. For example, the fingerprint region around 1,100 cm¹ clearly consisted of four bands. However, the change in acid was still hard to determine in the deconvolved spectra, as the acid bands at 1,420 and 1,600 cm¹ overlapped with the C-H absorption band (1,400 cm¹) and the amide I band (around 1,680 cm¹), respectively.

Two observations are presented in Figure 2B. First, similar band patterns in ester and the fingerprint region were observed among all three leaf stages, indicating that the carbohydrate components and cell wall pectin structure do not change significantly during development. Second, distinct patterns of protein absorption bands were found in the three stages. For the amide I band around 1,680 cm¹, a side band at a lower wave number was revealed in the seedling and adult stages, but was not found at the juvenile stage. Two amide II bands around 1,550 cm¹ were found in all three stages, but the major band shifted from a higher to lower wave number during development. A side band of amide III band around 1,250 cm¹ was present at the seedling stage, but disappeared as the plants became older. These results show a unique protein conformation at all three stages of development.

After carefully evaluating the IR spectra generated from the dried leaves, three absorption regions were chosen and were used to study the effect of salt: the 1,745 cm¹, 1,680 cm¹, and 1,100 cm¹ (fingerprint) bands. The assignment of absorption bands to their major contributing chemical components is presented in Figure 2C.

IR Monitors the Changes in Carbohydrate and Cell Wall Pectin by Salt Stress

IR measurement followed by the deconvolution technique was able to show the differences of chemical composition during development in the ice plant. Using this method, salt-induced changes in chemical composition were examined. Under continuous illumination, 5-week-old ice plants were salt stressed with 200 mM NaCl. The effects of the salt were analyzed by FT-IR spectrometry (Fig. 3A). To compare, Arabidopsis was stressed with 50 mM NaCl, a nonpermissive salt concentration for its growth. One set of typical IR spectra is shown in Figure 3B. As can be seen in these two figures, salt stress did not cause major alternations in the IR spectra. The area of two relative isolated bands, around 1,745 cm¹ (ester) and 1,100 cm¹ (fingerprint) bands, were integrated and plotted against time of stress (Fig. 3C). There was little change in the 1,745-cm¹ band area throughout the test period in Arabidopsis. In ice plant, an increase in the 1,745-cm¹ band intensity was observed after 1 d of stress. This result suggests that the accumulation of cell wall pectin, as indicated by the increase in the ester band area, continued in ice plants but stopped in Arabidopsis. Although there was an initial decrease in Arabidopsis, the accumulation of carbohydrates continued in both plants after 48 h of salt stress, as indicated by the increase in the 1,100-cm¹ band area (Fig. 3C). The results show that in glycophytes, the synthesis of cell wall is much more sensitive to salt than the synthesis of carbohydrates.

View larger version (27K): [in this window] [in a new window]   Figure 3.   The effect of salt on the FT-IR absorption spectra of 5-week-old ice plants stressed with 200 mM NaCl (A) and 3-week-old Arabidopsis stressed with 50 mM NaCl (B). Samples were collected at different time intervals, indicated at the left end of each spectrum. At least three leaves were collected for each time point and at least three spectra were obtained from each sample. Only one representative spectrum of each treatment is shown. The major chemical constituents in the leaves that contribute to the formation of bands in the particular wavenumbers are indicated at the top. C, The relative band area of 1,745 cm1 (ester) and 1,100 cm1 (fingerprint region) versus length of stress. Straight lines indicate changes occurred in ice plants; dashed lines represent changes in Arabidopsis. The band area at time 0 (before the salt stress) was set to 100%. Bars represent SD of the means.

Deconvolution of the IR Spectrum Detects the Changes in Protein by Salt Stress

The original IR spectra showed salt-induced changes in the bands around 1,680 cm¹ (amide I band) after the first few hours of stress in ice plant (Fig. 3A). However, this region yielded broad and overlapping bands. To enhance the resolution of the IR spectra and to precisely determine the band positions for later quantitative analyses, deconvolution was used. The results are shown in Figure 4A. After deconvolution, band positions corresponding to esters, proteins, and carbohydrates were clearly distinguished. The most prominent change was associated with amide I band (Fig. 4A). The positions of amide I band are sensitive to the secondary structure of the proteins. A side band at 1,640 cm¹ responsible for the increase of the 1,680-cm¹ bandwidth shown in Figure 3A started to emerge after the 1st h of stress. At 24 h after salt stress (Fig. 4A, arrowhead), the shape of the amide I band already strikingly resembled the amide I band of the 10-week-old adult leaves shown in Figure 2B. This result indicates a rapid protein conformation change within 1 d of salt stress in ice plant leaves; this change is consistent with the developmental transition from juvenile to adult stage.

View larger version (24K): [in this window] [in a new window]   Figure 4.   The effect of salt on the deconvolved FT-IR absorption spectra of 5-week-old ice plant stressed with 200 mM NaCl (A) and 3-week-old Arabidopsis stressed with 50 mM NaCl (B). IR spectra shown in Figure 3 were deconvolved. The arrowhead shown in A indicates a newly emerged amide I band at 1,640 cm1. The arrowhead shown in B indicates the position of amide II band at 1,550 cm1. The major chemical constituents in the leaves are indicated at the top. C, A curvefitting result obtained for ice plant leaves at time 0. The dashed line shows the original IR spectrum. The band positions obtained by deconvolution spectrum are indicated at the top. The amide I absorption band is composed of two bands located at 1,680 cm1 (major) and 1,640 cm1 (minor).

The IR spectra of salt-stressed Arabidopsis were also deconvolved in the region of 2,000 to 1,000 cm¹ (Fig. 4B). Distinct patterns of deconvolved spectra were found in Arabidopsis. For example, the relative intensities of four bands in the fingerprint region were different between Arabidopsis and ice plants. The band positions corresponding to esters and proteins were not identical to those of ice plants. The amide I band located around 1,660 cm¹ in Figure 3B is clearly composed of two bands located at 1,668 and 1,630 cm¹. A decrease in the 1,630-cm¹ amide I band was observed between 5 and 24 h of salt stress. The intensity of the amide II band located around 1,550 cm¹ decreased after 48 h of salt stress (Fig. 4B, arrowhead). The results show that the protein synthesis pathway in Arabidopsis is sensitive to mild salt stress.

Specific Changes in Protein Conformation at Early Stages of Salt Stress

To quantitatively analyze the change in intensities of the amide I band, the IR spectra of the ice plants (Fig. 3A) and Arabidopsis (Fig. 3B) were curve fitted. The Gaussian peak shape was used to fit the spectrum collected at each time point, and a typical curve-fitting result is shown in Figure 4C. As can be seen in this figure, the spectrum can be decomposed into several bands with high fitness. The amide I band of ice plant is clearly composed of two bands located at 1,680 and 1,640 cm¹. The major amide I band is located at 1,680 cm¹, whereas the 1,640-cm¹ band is the minor side band observed in the deconvolved spectra. The fitted curve also shows an acid band at 1,600 cm¹ (Fig. 4C) that was unresolved in the deconvolved spectra. For Arabidopsis, the amide I band is also composed of two bands, with the major band located at 1,668 cm¹ and a minor side band located at 1,630 cm¹ (data not shown). The total area of amide I bands assigned by the fitting program (i.e. 1,680 and 1,640 cm¹ for ice plant and 1,668 and 1,630 cm¹ for Arabidopsis), was calculated and plotted against time of salt stress (Fig. 5A).

View larger version (17K): [in this window] [in a new window]   Figure 5.   Salt-induced changes in amide I bands in ice plants (straight lines) and Arabidopsis (dashed lines). The relative band area of total amide I (A), major amide I (B), minor amide I (C), and ratio of major/minor amide I band (D) were calculated based on the results obtained by curve fitting. Spectra presented in Figure 3, A and B were curve fitted, and the band area at time 0 (before salt stressed) was set to 100%. Bars represent SD of the means.

After 24 h of stress, the total leaf protein increased 15% in ice plant and decreased 10% in Arabidopsis. The result indicates that ice plants are able to maintain protein synthesis under salt stress, whereas the protein synthesis machinery in glycophytes is sensitive to salt. Similar salt-induced changes are also observed in the amide II band located at 1,550 cm¹ and amide III band located at 1,250 cm¹ (data not shown). The result confirms the observation seen by the deconvolved spectra; the curve-fitting technique was able to quantify the changes of overlapping bands at a specific wave number.

Further examination of the changes in individual amide I bands revealed some interesting observations. In ice plant and Arabidopsis, changes in the major amide I band (1,680 cm¹ for ice plant and 1,668 cm¹ for Arabidopsis) were similar: a raise in band intensity during the 1st h of stress, with a return to the prestressed level (Fig. 5B). However, the changes in the minor amide I band (1,640 cm¹ for ice plant and 1,630 cm¹ for Arabidopsis) were different between ice plant and Arabidopsis (Fig. 5C). In Arabidopsis, a rapid decrease of the 1,630-cm¹ band intensity at first 5 h was recorded, with a 20% decrease in band intensity after 24 h of stress. In the ice plants, an initial decrease at the 1st h was followed by a large increase in band intensity and a 40% increase in the minor amide I band after 24 h of salt treatment (Fig. 5C). This suggests the changes in the minor band intensity are the major factor that causes salt-induced changes in the protein content.

In proteins, the most important hydrogen bonds are those between peptide bonds. In the results, the positions of amide I bands reflect the degree of hydrogen bonding: the higher the wavenumber, the weaker the H bonding (i.e. the less-ordered protein structure). Therefore, the ratio between two amide I bands is a useful indicator of the overall protein ordering at a particular point of stress. The 1,668/1,630 ratio rapidly increased in Arabidopsis and reached a maximum value 12 h after salt stress, followed by a slow recovery (Fig. 5D). The results suggest that under salt stress, a rapid rise in relative proportion of the less-ordered form of protein in Arabidopsis occurs. In ice plants, the 1,680/1,640 ratio decreased after an initial rise at 1 h. The minimum value of 1,680/1,640 ratio occurred 12 h after the stress (Fig. 5D). The first 12 h poststress seems to be a critical period that determines the degree of salt tolerance in higher plants. The result shows that halophyte ice plant is able to maintain a higher-ordered form of protein in the leaves at the early stages of salt stress.

To examine whether sodium ions are the factor that causes the change in protein conformation, the band ratio (Fig. 5D) was plotted against Na⁺ concentration in the leaves (Fig. 1). As presented in Figure 6, a positive correlation between band ratio and Na⁺ content in ice plants (R² = 0.8045) and a poor correlation in Arabidopsis (R² = 0.0369) is evident. This suggests that halophyte ice plants are able to make rapid and positive protein conformation changes in response to the fluctuation of Na⁺ level in the leaves.

View larger version (17K): [in this window] [in a new window]   Figure 6.   Relationship between leaf sodium content (Fig. 1) and amide I band ratio (Fig. 5D). The straight line indicates a correlation between sodium content and amide I band ratio in ice plants. The dashed line reveals no correlation in Arabidopsis. The square of the regression coefficient (R2) is shown for each plot.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

In this work, we used IR spectroscopy for early detection of chemical and conformational changes in plant leaves under salt stress. Bulk chemical analysis of changes in carbohydrate, protein, and cell wall were performed with two model plants: Arabidopsis and ice plant, which have been used extensively for salinity stress studies (for review, see Hasegawa et al., 2000). The results reveal different strategies for coping with a sudden rise of soil salinity in glycophytes and halophytes.

The halophytic ice plants actively respond to increased Na⁺. Sodium ions are required for the maximum growth of cultured ice plant cells; these cells are able to use Na⁺ for the maintenance of cellular osmotic potential under high salt (Yen et al., 1995, 1997). The rapid fluctuation of leaf Na⁺ and RWC (Fig. 1) indicates that ice plant roots have no apparent prevention barrier against a sudden increase of Na⁺ in the soil. Na⁺ is transported to the leaves via a myo-inositol-dependent sodium uptake system (Nelson et al., 1999). The incoming sodium ions cause a rapid protein conformation change (Fig. 5) that may serve as a signal to initiate a signal transduction cascade for salt adaptation mechanisms in ice plants. An active Na⁺ efflux system was triggered and a rapid decrease of leaf Na⁺ level was observed 3 h after salt stress. High levels of Na⁺ were found in the xylem sap of salt-stressed ice plants (Adams et al., 1992), suggesting that excess Na⁺ might be temporally excluded into the vascular tissues. At the same time, several salt-tolerant mechanisms are engaged in mesophyll cells. These include transient increased expression of leaf-specific K⁺ channel genes at 6 h (Su et al., 2001), increased expression of vacuolar H⁺-ATPase gene increased at 8 h, followed by a decrease (Löw et al., 1996), and increased expression of genes encoding enzymes for compatible solutes synthesis 12 h after stress (Vernon and Bohnert, 1992; Ishitani et al., 1996). Once the mechanisms for ion homeostasis and osmotic adjustment are activated, the synthesis of protein, cell wall pectin, and carbohydrate are able to continue under salt stress in this halophyte. In a study of salt-adapted tobacco (Nicotiana tabacum) suspension cells, an increase of protein, pectin, and poly-GalUA in the cell wall was observed by FT-IR microspectroscopy (McCann et al., 1994).

A passive defense to high salinity was observed in glycophytes. Upon sensing the increased salinity in the soil, Arabidopsis was able to delay sodium ions reaching the photosynthetic tissues for 12 h (Fig. 1). Exclusion of Na⁺ from the roots is the first line of defense when glycophytes are salt stressed. Several Na⁺ compartmentation systems have been found in the root of Arabidopsis, such as SOS1 (Shi et al., 2000) and AtNHX1 (Gaxiola et al., 1999). Overexpression of a single Na⁺/H⁺ antiport greatly enhances salt tolerance (Apse et al., 1999; Zhang and Blumwald, 2001). Nonetheless, under continuous illumination and prolonged stress, toxic Na⁺ ions inevitably reached the leaves. Protein synthesis was most sensitive to the increased concentration of Na⁺ (Fig. 5), possibly through the inhibition of the ribosomal complex. The different sensitivity to salt observed in Arabidopsis is similar to the results obtained for other glycophytes (Boyer, 1970) in that the rate of cell wall expansion is much more sensitive to salt than the carbon assimilation at early stages of osmotic stress.

FT-IR has been widely used for analysis of protein conformation (Surewicz and Mantsch, 1988). The assignment of IR bands to specific secondary structure such as -helices, -sheets, turns, and nonordered structure is a difficult task. Nevertheless, IR detection of changes in protein secondary structure in maize embryo (Wolkers et al., 1998, 1999) has been reported. Based on the amide I absorption band, an increased ratio of -helix/-sheet was found to associate with the increase of desiccation tolerance in developing maize embryo. In this report, we have also found changes in amide I band during plant development and under salinity stress. We did not attempt to assign each individual amide I band to a particular secondary structure in a mixed protein population of plant leaves. Instead, we provided the ratio between individual amide I band as a good indicator of the dynamics of protein conformation. The increased ratio of higher to lower amide I bands represents an increased proportion of unordered structures (less H-bonding) in proteins (Susi, 1969). In Arabidopsis, the ratio increased upon salt stress, suggesting that the protein conformations in the leaves rapidly became less ordered. The ratio reached its maximum value after 5 h, but at the same time, the Na⁺ concentration increased only 2-fold. This indicates that a portion of the leaf protein in Arabidopsis is sensitive, but not responsive, to the elevated level of Na⁺. Subtle changes in certain amino acids in salt-sensitive proteins would increase their stability under salt stress. This may be one of the reasons that the two plant species have very similar genomic and protein compositions but have different capacity to salt tolerance.

A portion of leaf protein in ice plants was salt responsive, as indicated by the positive correlation between Na⁺ content and ratio of higher to lower amide I band (Fig. 6). Furthermore, the changes of amide I band occurred not only rapidly from salt stress, but also gradually during development. Emerge of the 1,640-cm¹ lower amide I band occurred in 10-week-old nonstressed adult plants (Fig. 2B), the stage when full competency to osmotic stress is achieved (Adams et al., 1998). Although the increased portion of more ordered protein structure may not be a prerequisite for salt tolerance in higher plants, a plant's responsiveness to salt stress is an important factor for growth regulation under salt stress (Zhu, 2001). In vitro assay of leaf protein extract using FT-IR microspectroscopy is currently underway to reveal the relationship between salinity and protein conformation.

Dried leaves were used in this study because water is such a strong IR absorber. Its absorption bands can obscure most of the chemical information generated from other bands. Although we cannot eliminate the possibility of chemical changes caused by the process of sample drying (McCann et al., 1992), the generation of typical spectra in dried leaves is one step toward the establishment of the response spectra to various stresses in situ. Based on the information generated by dried leaves, an in situ monitoring of the effects of salt and wounding in intact ice plants was performed (Tsai, 2001). The changes in cell wall pectin and carbohydrate in salt-stressed fresh leaves were similar to the changes reported in this article. A nondestructive analysis using FT-IR and FT-Raman has successfully characterized the changes of lignin structure in transgenic tobacco (Stewart et al., 1997). A database containing such spectra can then be used to predict the responses to a specific stress in a mixed plant population. IR technique is currently restricted in detection of dried leaves and relative simple systems. With the assistance of spectral subtraction technique (Griffiths and de Haseth, 1986) and suitable arrangement of the optical system, it is possible to detect IR signals directly from living plants. The use of optical fiber (Yang and Huang, 2000; Yang and Lin, 2000) can further eliminate the tedious optical arrangement of the detection system. IR technique is a perspective technique for the detection of stress and screening mutants even before stress symptoms appear.

In conclusion, FT-IR is able to detect the chemical changes at early stages of salt stress. Ice plants show a positive growth tendency under salt stress, but the growth of Arabidopsis is obviously sensitive to salt. The results presented in this article underline the changes of chemical property, and contribute to the current knowledge of physiological, biochemical, and genetic components of salt adaptive processes in higher plants.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Growth Conditions and Salt Treatments

Seeds of ice plants (Mesembryanthemum crystallinum) were germinated directly in a soil mixture (peat moss:vermiculite:sand, 3: 1: 1, v/v) and were maintained in a growth chamber under a 16-h light, 30°C/8-h dark, 18°C cycle. After 2 weeks, the seedlings were thinned to one plant per pot. The irradiance was 600 to 700 µmol quanta m²s¹. Plants were watered twice a day and were supplemented with a diluted Peter's fertilizer (3 g L¹) once a week. One-week-old (seedling), 5-week-old (juvenile), and 10-week-old (adult) nonstressed plants were used to represent three developmental stages. Salt treatment was initiated by irrigating 5-week-old juvenile plants with 200 mM NaCl (50 mL pot¹) 2 to 3 h into the light period. Under continuous illumination, the newly expanded third and fourth pair leaves were excised at different time points up to 48 h and were immediately dried in an oven for 2 d at 55°C. Arabidopsis plants were germinated directly in a soil mixture (peat moss:vermiculite:perlite, 8: 1: 1, v/v) and were maintained in a growth chamber under continuous illumination at 22°C. After 1 week, the seedlings were thinned to one plant per pot. The irradiance was 400 to 500 µmol quanta m² s¹. Plants were watered twice a day and were supplemented with a diluted Peter's fertilizer (3 g L¹) once a week. Three-week-old plants that were on set of inflorescence were used in the response kinetic to salt. Salt treatment was initiated by irrigating plants with 50 mM NaCl (50 mL pot¹). Under continuous illumination, the fully expanded leaves (approximately 4-5 cm in length) were excised at different time points and were immediately dried in an oven for 2 d at 55°C. Three to four leaves were taken from different plants and were pooled as one sample point. Dried leaves were first used for FT-IR measurement (see below). After IR spectra were generated, leaves were ground into powder and subjected to determination of Na⁺ content.

Na⁺ Content in the Leaves

HCl (6 M) extracts were prepared from the ground leaf powder (Lambert, 1976). Sodium content was analyzed by an inductively coupled plasma Atomic Emission Spectrometer (JY138 Ultrace; Instruments SA, Edison, NJ).

RWC

RWC of leaf tissue was monitored and calculated in a separate set of samples according to Chu et al. (1990).

IR Spectroscopy

A Fourier transform infrared spectrometer (Magna 550; Nicolet, Madison, WI) was used to obtain the IR spectra. This spectrometer was equipped with a mercury-cadmium-telluride detector with a 1-mm² sensing area. Spectra were collected in 4-cm¹ resolutions and coadded 100 scans. A 45-degree reflection-absorption optical accessory was used to perform the measurements. In this optical system, a gold-coated plate was used as a reflection reference. To remove the spectral interference from water absorption bands, leaves were dried before measuring the reflection-absorption IR spectra. Dried excised leaves were placed on top of the gold-coated plate to obtain the reflection spectra. IR spectra were displayed in an ordinary absorption unit. Two to three spectra were generated from each sample by measuring the different area of the leaf, and only one representative spectrum was shown in the results. All the data points for salt treatment were repeated three times using different sets of leaf samples and highly reproducible spectra were obtained.