Abscisic acid modifies the changes in lipids brought about by water stress in the moss Atrichum androgynum


  • Irina A. Guschina,

    1. Institute of Ecology of the Volga River Basin RAS, Togliatti 445003, Russia;
    2. School of Biosciences, Cardiff University, PO Box 911, Cardiff CF10 3US, UK;
    Search for more papers by this author
  • John L. Harwood,

    Corresponding author
    1. School of Biosciences, Cardiff University, PO Box 911, Cardiff CF10 3US, UK;
      Author for correspondence: John L. Harwood Tel: +44 (0)29 20874108 Fax: +44 (0)20 20874116 Email: Harwood@Cardiff.ac.uk
    Search for more papers by this author
  • Mike Smith,

    1. School of Botany and Zoology, University of Natal, Pietermaritzburg, Scottsville 3209, Republic of South Africa
    Search for more papers by this author
  • Richard P. Beckett

    1. School of Botany and Zoology, University of Natal, Pietermaritzburg, Scottsville 3209, Republic of South Africa
    Search for more papers by this author

Author for correspondence: John L. Harwood Tel: +44 (0)29 20874108 Fax: +44 (0)20 20874116 Email: Harwood@Cardiff.ac.uk


  • • Mosses are particularly able to withstand drought stress. Moreover, abscisic acid (ABA), which is intimately involved during stress in higher plants, has also been implicated in bryophytes. Because membrane damage is a common feature of drought stress, we have studied changes in lipid composition during desiccation and rehydration of the moss Atrichum androgynum and the effect of exogenous ABA on these processes.
  • • In order to correlate any membrane changes with drought stress, we analysed different lipid classes by thin-layer chromatography, fatty acid composition by gas–liquid chromatography and lipid peroxidation.
  • • Water stress caused changes in phosphoglyceride composition consistent with an activation of phospholipase D and of phosphatidylinositol metabolism. Recovery of phosphoglyceride composition towards original levels occurred during rehydration and ABA treatment reduced the overall extent of changes. Reduction in thylakoid lipids and chlorophyll coincided with loss of photosynthesis.
  • • The data show that mosses respond to drought stress similarly to higher plants and that ABA may reduce membrane damage by diminishing the lipid changes.


Plants respond to water stress by modifying aspects of their lipid metabolism, and these changes depend on the plant species and the conditions of desiccation (Liljenberg, 1992). Typically, glycerolipids and phosphoglycerides decrease while neutral lipids increase (Meyer et al., 1992; Stevanovic et al., 1992; Dakhma et al., 1995; Olsson et al., 1996; Repellin et al., 1997; Lauriano et al., 2000). Consistent with this, radiolabelling from [1–14C]acetate of membrane lipids was decreased and that into neutrals was increased by water stress (Pham Thi et al., 1985). Fatty acid desaturation was also reduced (Pham Thi et al., 1987).

Not only can lipid biosynthesis be changed by water stress, but degradation of polar lipids may also be altered (Sahsah et al., 1998). Phospholipase D and galactolipases were stimulated significantly, especially in sensitive species. Expression of a gene encoding phospholipase D was increased in drought-sensitive cowpea plants but not in tolerant cultivars (El-Maarouf et al., 1999). Moreover, the activity and expression of soybean lipoxygenase (which can give rise to lipid signalling molecules) was also enhanced by water deficit (Maccarrone et al., 1995). Recently, Matos et al. (2001) have detected a drought-induced acyl hydrolase in cowpea leaves which may be involved in the galactolipid degradation referred to above.

A well-known effect of water stress is lipid oxidation that can lead to disruption or destabilization of membrane structure (Liljenberg, 1992). This oxidation is caused by generation of reactive oxygen species (ROS) during desiccation (Smirnoff, 1993; Jiang, 1999). Although both desiccation-tolerant and sensitive plants form ROS, they are much more damaging to the latter (Seel et al., 1991). The water stress-induced changes in lipid metabolism discussed above may reflect direct effects of ROS or may be a result of the need to repair damaged membrane lipids.

The plant hormone abscisic acid (ABA) is an essential mediator in triggering the plant responses to many environmental stresses (for reviews see Bray, 1997; Leung & Giraudat, 1998). It is involved in desiccation tolerance in higher plants, ferns and bryophytes through regulation of gene expression and synthesis of new proteins that might be responsible for tolerance (Werner et al., 1991; Bopp & Werner, 1993; Reynolds & Bewley, 1993; Hellwege et al., 1994; Leung & Giraudat, 1998; Machuka et al., 1999; Iuchi et al., 2000). In two jute species, Corchorus capsularis and Corchorus olitorius, subjected to water stress, foliar sprays of ABA increased the activity of the ROS scavenging enzymes, superoxide dismutase and catalase, and reduced malondialdehyde formation. This suggests that part of the role of ABA is to activate signal transduction pathways that reduce ROS formation which will, in turn, reduce lipid peroxidation and membrane damage (Chowdhury & Choudhuri, 1989).

In bryophytes, limited information is available on the endogenous levels of ABA and how they change under different environmental conditions (see Bopp & Werner, 1993; Christianson, 2000). Slow drying of Funaria hygrometrica increased its endogenous ABA content sixfold, and was correlated with increased tolerance to rapid drying. Moreover, exogenously added ABA could also increase tolerance to rapid drying (Werner et al., 1991). Bopp and Werner (1993) noted that, in several different bryophytes, water deficit enhanced endogenous ABA, while exogenously applied ABA could directly alter aspects such as stomatal closure in Anthoceros, differentiation in Riccia and drought hardening in Funaria. Beckett et al. (2000) and Mayaba et al. (2001) used chlorophyll fluorescence to show that ABA pretreatment increased nonphotochemical quenching (NPQ) in the moss Atrichum. Increased NPQ can reduce ROS formation around photosystem II following stress (Gilmore, 1997). In addition to activating signal transduction pathways that reduce ROS formation, ABA can also interact with specific membrane phospholipids, as noted both in vivo and in vitro (Leshem et al., 1990; Stillwell et al., 1991; Shripathi et al., 1997).

Direct evidence for ABA involvement in the regulation of lipid metabolism and a role for phospholipase D-mediated hydrolysis of membrane lipids in ABA-signalling pathways have been reported in the last decade (Frank et al., 2000; Ritchie & Gilroy, 2000; for review see Wang, 2000). In mungbean leaves, exogenous application of ABA increased phospholipid levels and decreased free fatty acid and monogalactosyldiacylglycerol (MGDG) contents (Aghofack-Nguemezi et al., 1991). Changes in triacylglycerol metabolism as well as alterations in fatty acids were found also as a result of ABA treatment of developing wheat embryos (Rodriguez-Sotres & Black, 1994) and in microspore-derived embryos of oilseed rape (Finkelstein & Somerville, 1989; Pomeroy et al., 1994; Wilmer et al., 1997; Wilmer et al., 1998). Among a number of plant genes induced by water stress were ABA-inducible cis- and trans-acting factors, a number of genes for protein kinases and enzymes involved in phosphatidylinositol turnover (Shinozaki et al., 1998). In addition, an ABA pretreatment of the moss Atrichum androgynum or of leaf discs of Vigna unguiculata enhanced membrane integrity, as evaluated by less electrolyte efflux, during rehydration following desiccation (Campos & Thi, 1997; Beckett, 1999, 2001).

Many bryophytes are desiccation tolerant and photosynthesize and grow only when water is available (Oliver et al., 2000; Proctor, 2000). Lipids of bryophytes differ from higher plants by the presence of the betaine lipid diacylglyceryltrimethylhomoserine (DGTS), high levels of esterified sterols and very long chain fatty acids and triacylglycerols (TAGs) often with acetylenic fatty acids (Dembitsky, 1993). However, limited information is available on the effect of water stress on the lipid metabolism of these plants (Stewart & Bewley, 1982; Hakala & Sewon, 1992).

Here, we report changes in lipid composition during desiccation and rehydration of the moss A. androgynum and the effect of exogenous ABA on these processes. Atrichum androgynum was chosen because it is an ABA-responsive organism, and specifically because ABA treatment reduces ion leakage during rehydration following desiccation (Beckett, 1999, 2001; Beckett & Hoddinott, 1997). This study was conducted to increase our knowledge of drought responses in bryophytes and to determine if ABA-induced changes in membrane permeability correlated with alterations in lipid biochemistry.

Materials and Methods

Plant material

The moss A. androgynum (Hedw.) P. Beauv. was collected during the moist months (February–April, 2000) from the Doreen Clarke Nature Reserve, Hilton, KwaZula-Natal Province, Republic of South Africa. This nature reserve is a small pocket of Afromontane forest in the mist belt region of KwaZulu Natal. Before the experiments, plants were stored on wet filter paper at 20°C and a photosynthetic photon fluence rate (PPFR), measured across photosynthetically available wavelengths, of 75 µmol photons m−2 s−1 under continuous fluorescent light for 2 d. Apical green 2 cm segments were then cut and material divided into replicates of 1.6 g fresh weight. Experiments were typically repeated three times with independently harvested samples.


Fatty acid standards were from Nu-Chek (Elysian, MN, USA) and silica gel G plates from Merck (Darmstadt, Germany). Abscisic acid was obtained from Sigma (St Louis, MO, USA). Other reagents were of the best available grades and were from Sigma or from BDH (Capital Enterprises, Pinetown, South Africa).

Pre-treatment with ABA and desiccation/rehydration course

Abscisic acid was dissolved in a drop of 1 m NaOH, diluted with deionized water and the pH of the resulting solution adjusted to 5.6 with HCl. Moss samples were gently shaken (c. 120 r.p.m.) in 10 ml of deionized water or 100 µm ABA for 1 h then stored for 3 d in the light as described earlier. Control samples were similarly treated but with water instead of 0.1 mm ABA.

Material (1.6 g) was left exposed to slowly dry under normal laboratory conditions (PPFR of 10 µmol photons m−2 s−1, 20°C) for 30 h, then over silica gel for a further 12 h. The relative water content (RWC) at the end of drought period was about 0.02. The moss was then rehydrated in 10 ml of deionized water for 8 h. The leakage of potassium (K+) in replicate material was determined using the sequential elution technique of Brown and Buck (1979). Pretreating A. androgynum with ABA reduced the proportion of K+ lost during subsequent rehydration from c. 70 ± 18% to 41 ± 6%, which is typical of previous results (Beckett, 1999, 2001).

Lipid analysis

Lipids were extracted using the method of Garbus et al. (1963), following heating of samples in hot isopropanol at 70°C for 30 min to ensure that lipid catabolic enzymes were inactivated.

Neutral lipids were separated by one-dimensional thin-layer chromatography (TLC) on 10 × 10 cm silica gel G plates with double development, first with toluene–hexane–formic acid (140 : 60 : 1, by volume) for the whole plate height followed by hexane–diethyl ether–formic acid (60 : 40 : 1, by volume) to half height. Polar lipids were separated by two-dimensional TLC on 10 × 10 cm silica gel G plates using chloroform–methanol–toluene−28% ammonium hydroxide (65 : 30 : 10 : 6, by volume) in the first dimension and then chloroform–methanol–toluene–acetone–acetic acid–water (70 : 30 : 10 : 5 : 4 : 1, by volume) in the second direction.

Identification of lipid classes was made by reference to authentic standards and confirmed using specific colour reagents (Kates, 1986). Quantification of triacylglycerols, DGTS and glycolipids was achieved by determination of ester bonds (Higgins, 1987) and phospholipids by phosphorus (Vaskovsky et al., 1975).

Fatty acid methyl esters (FAMEs) were prepared by transmethylation with 2.5% H2SO4 in dry methanol and analysed by capillary gas–liquid chromatography (GLC) using a gas chromatograph (Varian-3500; Varian UK Limited, Surrey, UK). A 30-m Carbowax capillary column (Supelco, Poole, Dorset, UK) was run with a temperature programme (initial temperature of 140°C for 10 min, then increasing at 6°C min−1 to 265°C) and with nitrogen as carrier gas. Routine identification was by reference to standards and quantification was made using a 3395 integrator (Hewlett–Packard, Wilmington, DE, USA) and an internal standard of heptadecanoate. Full details of methods used to analyse lipids are presented in Bychek-Guschina (2002).

To measure the level of lipid hydroperoxidation, individual lipids were separated by TLC, then scraped and eluted with chloroform–methanol (1 : 1, by volume). Hydroperoxides were determined following the method of Stine et al. (1953).

Photosynthetic capacity

Net photosynthesis was measured at 25°C and 50% relative humidity (r.h.) using an Analytical Development Corporation (ADC) Mark III portable infra-red gas analyser (IRGA) (Analytical Development Co Ltd., Hoddeston, UK) with a barrel-shaped Parkinson leaf chamber, modified with a water-cooled jacket. The flow rate through the leaf chamber was 120 ml min−1. Photosynthesis was measured at a saturating PPFD of 150 µmol photons m−2 s−1. Equilibrating samples for 10 min was found to give steady-state rates of gas exchange without causing enough water loss to reduce photosynthesis. Chlorophyll extraction and quantification were performed by the method of Barnes et al. (1992).


The composition of individual phospholipids in the moss A. androgynum was similar to that of other moss species (Bychek, 1994). Phosphatidylcholine (PC), phosphatidylglycerol (PG), phosphatidylethanolamine (PE), phosphatidylinositol (PI), phosphatidic acid (PA) and phosphatidylserine (PS) were detected in this species. As expected, PC and PG were the major phosphoglycerides. Water stress and rehydration did not change significantly the total content of phosphoglycerides. Figure 1 shows the changes in the content of individual PLs during desiccation and subsequent rehydration. The two major extra-plastid phosphoglycerides, PC and PE, were unchanged by desiccation but were significantly reduced during the initial period of rehydration (Fig. 1a,b). Phosphatidylcholine then showed recovery to original concentrations after 8 h rehydration. By contrast, the major plastid phosphoglyceride, PG, was reduced significantly during desiccation. It did not recover during rehydration. Two lipids showed increases in their relative amounts by 1 h of rehydration. Phosphatidylinositol was increased about 50% and then returned to pretreatment concentrations by 8 h rehydration (Fig. 1e). Phosphatidic acid showed the biggest changes, being increased nearly fourfold by 1 h of rehydration (Fig. 1d). Concentrations of this lipid then reduced gradually during rehydration. At 8 h rehydration phosphatidic acid were present at double the amount in untreated tissue.

Figure 1.

Effect of desiccation and the following rehydration on the phospholipid content in abscisic acid (ABA)-treated (open columns) and untreated (closed columns) Atrichum androgynum . (a) Phosphatidylethanolamine; (b) phosphatidylcholine; (c) phosphatidylglycerol; (d) phosphatidic acid; (e) phosphatidylinositol. Rehydration (RH) was for 1, 2, or 8 h and the phospholipid content is expressed as µg P g−1 d. wt. Results are means ± SD for three independent experiments. Bars with different letters are statistically different at P < 0.01 by Student's t-test, comparing control with desiccated/rehydrated mosses. The concentration of total phosphoglycerides was 10.5 ± 1.5 µg P g−1 d. wt in both initial samples and did not change significantly during desiccation or rehydration.

Pretreatment of tissue with ABA was able generally to reduce the phospholipid changes seen on desiccation and rehydration. Thus, it prevented the stress-induced changes in the proportions of PC and PI (Fig. 1b,e) and reduced those found in PE and PA (Fig. 1a,d). In addition, ABA-treated moss showed no significant desiccation or rehydration-induced changes in its PG content, although ABA treatment reduced its total amount about 25% compared with untreated moss (Fig. 1c).

Monogalactosyldiacylglycerol and DGDG were the major glycosylglycerides in A. androgynum. Desiccation significantly reduced the concentration of MGDG, but had no significant effect on the level of DGDG (Fig. 2). The MGDG content recovered quickly after desiccation, and reached the control level within 1 h of rehydration (Fig. 2a). Concentrations of DGDG did not change during the early stages of rehydration but after 8 h were reduced (Fig. 2b). While ABA treatment had little effect on the concentrations of these two glycosylglycerides, it did prevent the reduction in DGDG after 8 h rehydration (Fig. 2b).

Figure 2.

Effect of water stress and abscisic acid (ABA) treatment on the glycosylglyceride content of Atrichum androgynum. Monogalactosyldiacylglycerol (MGDG) levels are shown in (a) and digalactosyldiacylglycerol (DGDG) in (b). The results are means ± SD (n = 3). Open columns, ABA treatment; closed columns, control. Bars with different letters are statistically different at P < 0.05 by Student's t-test, comparing control with desiccated/rehydrated mosses.

As mentioned in the Introduction, mosses contain appreciable amounts of the betaine lipid, DGTS, as a membrane constituent (Dembitsky, 1996). Desiccation lowered the amount of DGTS, and concentrations did not recover during 8 h of rehydration (Fig. 3a). In ABA-treated moss, however, the concentration of DGTS was unaffected by desiccation or the initial rehydration, but after rehydration for 2 h the DGTS content of ABA-treated moss fell to a similar concentration as the controls. By contrast, desiccation approximately doubled the concentration of TAGs, and these returned to their initial values during rehydration (Fig. 3b). The ABA-treatment did not alter the changes in TAGs seen for control samples during desiccation and rehydration.

Figure 3.

Effect of desiccation and the following rehydration on diacylglyceryltrimethylhomoserine (DGTS) (a) and triacylglycerol (b) levels in ABA-treated (open columns) and untreated (closed columns) Atrichum androgynum. Results are means ± SD (n = 3). Bars with different letters are statistically different at P < 0.01 by Student's t-test, comparing control with desiccated/rehydrated mosses.

Tables 1 and 2 present the fatty acid composition of total lipids in untreated and ABA-treated mosses. Typical for many bryophyte species, very long chain (> 18C) polyunsaturated fatty acids were identified in A. androgynum , but their total relative proportions did not exceed 10–17%. No statistically significant differences in the total content of fatty acids between control and desiccated moss were found in either ABA-treated or untreated plants. However, the relative proportions of fatty acids changed during rehydration. After 8 h of rehydration, the proportion of palmitic (C16 : 0) acid decreased while the proportions of 16C unsaturated and some long-chain polyunsaturated fatty acids increased. The same general changes in fatty acid composition occurred in the rehydrated moss pretreated with ABA, ( Table 2 ).

Table 1.  Effect of water stress on the fatty acid composition of Atrichum androgynum
Fatty acidControlDry1 h Rehydration2 h Rehydration8 h Rehydration
  1. Fatty acids expressed as mole percentage. They are indicated with the number before the colon showing the number of carbon atoms and the figure afterwards denoting the number of double bonds (position of first double bond indicated, where known). Data are means ± SD for n = 3 for independent samples. Statistical significance was compared with control by Student's t-test. (P < 0.10*, P < 0.05, **); tr, trace (< 0.05%).

< 16C13.4 ± 2.310.6 ± 1.616.8 ± 2.414.2 ± 2.311.2 ± 1.9
16 : 021.4 ± 1.318.7 ± 1.016.6 ± 2.616.9 ± 1.5*15.3 ± 1.6**
16 : 1 0.5 ± tr. 0.5 ± tr. 0.3 ± tr.tr. 1.9 ± 0.5**
16 : 2 0.8 ± tr. 0.4 ± tr.** 2.2 ± 0.3** 1.0 ± 0.1* 1.3 ± 0.1**
16 : 3 2.5 ± 0.4 2.0 ± 0.2 3.0 ± 0.3 1.9 ± 0.2 1.2 ± 0.4
18 : 1 n-9 8.7 ± 2.8 9.3 ± 0.2 9.7 ± 1.7 9.6 ± 0.412.1 ± 0.4
18 : 2 n-623.2 ± 1.725.3 ± 2.222.6 ± 2.324.1 ± 2.122.0 ± 2.5
18 : 3 n-6 2.6 ± 0.3 2.2 ± 0.8 3.2 ± 1.0 2.9 ± 0.2 4.0 ± 0.9
18 : 3 n-314.8 ± 1.917.1 ± 0.712.4 ± 0.614.6 ± 1.412.3 ± 2.6
20 : 5 n-3 2.6 ± 0.3 2.9 ± 0.2 2.6 ± 0.4 2.9 ± 0.3 4.4 ± 1.0*
22 : 2 n-6 0.7 ± 0.4 1.3 ± 0.2tr. 1.3 ± 0.1 2.4 ± 0.9**
22 : 3 n-6 0.5 ± 0.1 0.6 ± 0.1tr. 0.7 ± 0.2 1.6 ± 0.2**
Others 8.3 ± 1.5 9.1 ± 0.810.6 ± 2.3 9.9 ± 1.110.3 ± 1.5
Table 2.  Effect of desiccation and the following rehydration on the fatty acid composition of abscisic acid-treated Atrichum androgynum
Fatty acidControlDry1 h Rehydration2 h Rehydration8 h Rehydration
  1. Fatty acids expressed as mole percentage. They are indicated with the number before the colon showing the number of carbon atoms and the figure afterwards denoting the number of double bonds (position of first double bond indicated, where known). Data are means ± SD for n = 3 for independent samples. Statistical significance was compared with control by Student's t-test. (P < 0.10*, P < 0.05, **); tr, trace (< 0.05%).

< 16C12.4 ± 2.010.2 ± 1.411.2 ± 3.014.2 ± 2.312.1 ± 1.9
16 : 021.2 ± 1.219.5 ± 1.214.4 ± 1.4.**14.9 ± 1.5**17.4 ± 2.1*
16 : 1 0.5 ± tr. 0.3 ± tr.* 0.6 ± tr.tr. 3.1 ± 0.7**
16 : 2 0.8 ± tr.tr. 0.5 ± tr.** 1.7 ± 0.2* 0.4 ± tr.**
16 : 3 1.8 ± 0.1 1.4 ± 0.3 1.5 ± 0.2 1.9 ± 0.2 2.8 ± 0.1**
18 : 1 n-910.4 ± 1.9 9.0 ± 2.012.0 ± 1.2 9.6 ± 0.410.3 ± 1.0
18 : 2 n-623.5 ± 3.326.0 ± 1.425.2 ± 3.424.1 ± 2.119.6 ± 3.3
18 : 3 n-6 2.4 ± 0.1 2.5 ± 0.3 3.1 ± 0.2 2.9 ± 0.2 4.1 ± 0.4**
18 : 3 n-314.2 ± 2.917.0 ± 1.215.5 ± 1.914.6 ± 1.411.3 ± 2.2
C20 : 5 n-3 3.8 ± 0.7 3.6 ± 0.3 3.3 ± 0.6 2.8 ± 0.7 3.8 ± 0.3
C22 : 2 n-6 1.2 ± 0.2 1.1 ± 0.1 0.6 ± 0.3 1.3 ± 0.2 4.4 ± 0.4**
C22 : 3 n-6tr.tr. 1.0 ± 0.1** 0.7 ± 0.2** 3.5 ± 0.1**
Others 7.8 ± 1.2 9.4 ± 0.911.1 ± 1.011.3 ± 0.8* 7.2 ± 0.6

Previous work by others has shown that significant lipid oxidation can occur during drought stress (for reviews see Liljenberg, 1992; Smirnoff, 1993). Accordingly, we examined peroxidation in typical lipid classes. Figure 4 summarizes the changes in the amounts of lipid peroxidation that occurred in the phosphoglycerides and the galactosylglycerides (MGDG and DGDG) in A. androgynum. In the galactolipids of untreated moss, drought increased hydroperoxide production by c. 70% and this level was maintained throughout rehydration (Fig. 4a). The ABA pretreatment delayed the increase in peroxidation of the galactosylglycerides to the beginning of rehydration. Furthermore, the galactolipid peroxide levels were subsequently lowered during rehydration for ABA-treated moss (Fig. 4a). In the phosphoglycerides, drought stress also increased lipid peroxidation, which had returned to original levels by the end of the rehydration period (Fig. 4b). Once more, ABA pretreatment delayed the maximum level of peroxidation until 1 h of rehydration, but again with a reduction to original levels for phosphoglycerides during the remaining rehydration period.

Figure 4.

Changes in hydroperoxide levels in galactosylglycerides (a) and phospholipids (b) caused by water stress in Atrichum androgynum. Open columns, abscisic acid (ABA) treatment; closed columns, control. Data are means of three replicates of a pooled sample (six preparations).

Desiccation a significantly (23%) reduced the total chlorophyll concentration, and this had not recovered after 8 h of rehydration (Fig. 5). Moss treated with exogenous ABA had a lower initial chlorophyll content. This was not altered by desiccation but subsequent rehydration caused increases in chlorophyll concentrations (Fig. 5). Desiccation completely inhibited photosynthesis (Table 3). Photosynthesis only slightly recovered during rehydration, and recovery was not improved by ABA treatment.

Figure 5.

Effect of desiccation/rehydration and abscisic acid (ABA) pretreatment on total chlorophyll content of Atrichum androgynum. Open columns, ABA treatment; closed columns, control. Results are means ± SD of three separate experiments for each graph. Asterisks show statistically different points at P  < 0.01 by Student's t -test, comparing untreated control with desiccated and rehydrated mosses (*) or ABA-treated mosses (#) compared with initial values.

Table 3.  Effect of water stress and abscisic acid (ABA) pretreatment on the photosynthetic capacity of the moss Atrichum androgynum
 Photosynthesis (µmol CO2 mg−1 d. wt)
 InitialDry1 h Rehydration2 h Rehydration8 h Rehydration
  1. nd, None detected. Results are means ± SD (n = 3 separate experiments).

Control3.35 ± 0.05nd0.40 ± 0.350.85 ± 0.200.82 ± 0.10
ABA-treated3.45 ± 0.10nd0.41 ± 0.370.48 ± 0.320.65 ± 0.45


While only a limited number of angiosperms can survive prolonged desiccation, the ability of some mosses to survive water potentials of −150 MPa is remarkable (Proctor, 2000). Although the involvement of lipids in the response of plants to water stress has been suggested for many higher plants, little information is available for bryophytes. In the present work, the effects of desiccation and subsequent rehydration on the lipid composition of the moss A. androgynum were studied as a part of continuing research on drought-tolerance mechanisms in this species (Beckett & Hoddinott, 1997; Beckett, 1999, 2001; Beckett et al., 2000; Mayaba et al., 2001, 2002). Atrichum androgynum is a moderately drought-sensitive species with an inducible (protection-based) tolerance mechanism, whereas extremely tolerant mosses possess a constitutive protection system and a rehydration-induced recovery mechanism (Beckett, 1999; Oliver et al., 1998). Pretreatment of A. androgynum with exogenous ABA has been shown to increase its tolerance to ion leakage during rehydration following desiccation (Beckett, 1999, 2001), implying that ABA might induce structural changes in membrane composition.

Desiccation and rehydration of A. androgynum under our experimental conditions produced a number of effects on endogenous lipids. First, desiccation significantly decreased the concentrations of two key chloroplast lipids (Harwood, 1980), MGDG (Fig. 2a) and PG (Fig. 1e). While MGDG concentrations returned to those of the control during rehydration, concentrations of PG did not. The ABA treatment did not affect these changes in chloroplast lipids.

Conversely, the proportions of extra-plastidic phosphoglycerides changed, mostly during the initial rehydration. The proportions of the major components PE and PC were reduced, while that of PA increased (Fig. 1a,b,d). Further rehydration allowed PC levels to recover, while proportions of PA decreased. The PI concentrations were significantly increased after rehydration for 1 h (Fig. 1c). Unlike the results obtained from chloroplast lipids, ABA treatment of A. androgynum tended to reduce and, in some cases, prevent the stress-induced changes in extra-plastidic membrane phospholipids.

In the highly desiccation-tolerant moss species Tortula ruralis, changes in phosphoglyceride metabolism were also noted which, again, largely recovered during rehydration (Stewart & Bewley, 1982). Moreover, several studies on higher plants produced similar results. Thus, it is well known that activation of phospholipase D is often involved in plant stress responses (for review see Wang, 2000). Indeed, the activation of phospholipase D and/or increases in phosphatidic acid formation during water stress has been reported for a number of plants or algae (English, 1996; Sahsah et al., 1998; El-Maarouf et al., 1999; Frank et al., 2000; Munnik et al., 2000; Meijer et al., 2001;) and our results (Fig. 1d) are consistent with these data. Moreover, the induction of genes for phosphatidylinositol metabolism following water stress in Arabidopsis (Shinozaki et al., 1998) would also fit with the observed increase in PI (Fig. 1e). Such changes could also be connected with the PI–PA cycle that is well established in other eukaryotic tissues (see Gurr et al., 2002).

Changes in the proportions of the phosphoglycerides generally occurred during the first part of rehydration rather than during desiccation (Fig. 1). This may be related to changes in tonicity and, possibly, disruption of membrane structures that require changes in metabolism for repair. Interestingly, these lipid changes also agree with the time-course of drought-induced ion leakage, which occurs during rehydration (Minibayeva & Beckett, 2001). Although the phospholipid changes are similar to those observed in water-stressed higher plants, the timing is different. We suggest that, because bryophytes are well-adapted to, and capable of much more dehydration than higher plants, the desiccation phase can be tolerated better for many aspects of biochemistry and physiology than in higher plants.

Phospholipase D has also been suggested to be involved in the response of barley aleurone cells to ABA (Ritchie & Gilroy, 2000). Our studies showed that pretreatment of A. androgynum with ABA tended to reduce the changes in phosphoglycerides caused by water stress and rehydration (Fig. 1). This would agree with the protective effect of ABA on ion leakage in this species, believed to be caused by damage to the plasma membrane (Beckett, 1999, 2001). Other studies have shown effects of exogenous ABA on lipid biochemistry or membrane structure in a few higher plant species also (Aghofack-Nguemezi et al., 1991; Campos & Thi, 1997; Shripathi et al., 1997).

The reductions in two key chloroplast lipids (MGDG and PG), caused by desiccation, were not altered by ABA treatment (Fig. 2). The degradation of MGDG was probably caused by increased acyl hydrolase activity, which usually favours MGDG as substrate(Galliard, 1980). It is notable that drought induces the expression of a gene coding for this enzyme in cowpea (Matos et al., 2001). There have also been a number of reports that water stress causes a change in galactosylglyceride metabolism in different plants (Pham Thi et al., 1985; El-Hafid et al., 1989; Tarano et al., 1993; Sahsah et al., 1998). In particular, a decline in MGDG causes an increase in the DGDG–MGDG ratio (Navari-Izzo et al., 1995; Quartacci et al., 1997) in excellent agreement with our data (Fig. 2). Because MGDG is the major lipid of photosynthetic thylakoids (Harwood, 1998), it is not surprising that changes in its metabolism have been correlated with alterations in membrane integrity (Tarano et al., 1993; Platt et al., 1994) and photosynthetic capacity following water stress (Navari-Izzo et al., 1995). In the present study, the significant losses in MGDG content (Fig. 2) and a similar reduction in chlorophyll (Fig. 5) correlated with a dramatic loss of photosynthetic capability (Table 3). Partial recovery of photosynthetic capacity after 8 h of rehydration (Table 3) was associated with a return of MGDG to normal levels (Fig. 2). In keeping with the beginning of thylakoid structural repair, ABA treatment had no statistically significant effect on these parameters during the recovery period. In earlier studies with Atrichum spp., Beckett et al. (2000) and Mayaba et al. (2001) found better recovery of photosynthesis following desiccation, which, by contrast to the present study, was significantly improved by ABA pretreatment. However, the desiccation period in their studies was shorter, and physiological stress would have been less than in the present investigation.

Two other lipid classes were also examined. DGTS is a significant component of the lipids of lower plants (Dembitsky, 1996). In Chlamydomonas, it is thought to be partly located in chloroplasts (Mendina-Morgenthaler et al., 1985) so it is interesting that desiccation reduced its concentrations (Fig. 3) as for MGDG (Fig. 2). Pretreatment with ABA reduced the stress-induced decrease in DGTS concentration, as for the extra-chloroplast phosphoglycerides (Fig. 1). The TAG content of A. androgynum increased during water stress (Fig. 3). This was in keeping with a typical rise in this lipid that often occurs during stress (Gurr et al., 2002). In higher plants, other environmental stresses cause large, reciprocal changes in the concentrations of MGDG and TAG (Sakaki et al., 1994), in good agreement with our results (Figs 2 and 3).

Abscisic acid pretreatment hardens A. androgynum to ion leakage during rehydration and we therefore predicted that ABA might reduce the formation of lipid oxidation products. Reactive oxygen species are produced during the desiccation of bryophytes (Seel et al., 1991) and can abstract hydrogen radicals from membranous lipids, thus initiating peroxidation. For example, the perhydroxy radical (HOO) generated by the protonation of O2 − and OH can extract the bis-allylic hydrogen atom from unsaturated fatty acids (LH) forming lipid alkyl radicals (L) that are further oxidized by molecular oxygen to generate lipid peroxy radicals (LOO). The LOO react with LH yielding LOOH and L. Thus, a radical chain reaction is propagated. Lipid peroxides decompose to give aldehydes such as 4-hydroxynonenal (HNE) (Esterbauer et al., 1991) and malondialdehyde (MDA) (Valenzuela, 1991) and other products such as volatile hydrocarbons (e.g. ethane and pentane). These aldehydes are considered as toxic second messengers that disseminate initial free radical events. In this way, lipid peroxides and their degradation products cause extensive damage (Howlett & Avery, 1997). The consequences of changes in lipid and protein structure, and the inactivation of membrane-bound enzymes, are loss of membrane integrity and selective permeability. Therefore, following membrane damage, leakage of electrolytes, such as K+, can occur (for review see Cuny et al., 2002). Only limited work on lipid oxidation has been carried on bryophytes, but Dhindsa and Matowe (1981) and Seel et al. (1992) found that desiccation induces more MDA production in desiccation-sensitive mosses than in tolerant species. Interestingly, our recent work with A. androgynum showed that ABA pretreatment greatly reduced H2O2 formation during rehydration following desiccation (Mayaba et al., 2002). To our surprise, however, in the present study, we found that ABA pretreatment did not reduce hydroperoxide formation (Fig. 4). Possibly, the reduction in ion leakage results from ABA inducing dehydrins (Oliver et al., 1998) that stabilize membrane channel proteins. However, our results suggest that ABA does not increase the tolerance of A. androgynum to electrolyte leakage by reducing lipid peroxidation.

In conclusion, desiccation-induced changes in the lipid composition of A. androgynum were consistent with a role of phospholipase D in the overall process, and of glycerolipids as a sensitive target for membrane damage. Our results show that ABA treatment reduces the size of the stress-induced changes in lipid composition in A. androgynum. Thus, in bryophytes, as in other plant species, ABA alters the changes in lipids and/or lipid metabolism which are induced by water stress.


IAG is grateful to NRF and The Royal Society for fellowships. We thank the University of Natal Research Fund for running expenses.