Elevated atmospheric CO2 concentration and diurnal cycle induce changes in lipid composition in Arabidopsis thaliana

Authors

  • Åsa Ekman,

    1. Department of Crop Science, Swedish University of Agricultural Sciences, PO Box 44, SE-23053 Alnarp, Sweden;
    2. Department of Pure and Applied Biochemistry, Centre for Chemistry and Chemical Engineering, Lund University, PO Box 124, SE-22100 Lund, Sweden
    Search for more papers by this author
  • Leif Bülow,

    1. Department of Pure and Applied Biochemistry, Centre for Chemistry and Chemical Engineering, Lund University, PO Box 124, SE-22100 Lund, Sweden
    Search for more papers by this author
  • Sten Stymne

    1. Department of Crop Science, Swedish University of Agricultural Sciences, PO Box 44, SE-23053 Alnarp, Sweden;
    Search for more papers by this author

Author for correspondence: Åsa Ekman Tel: +46 40 415549 Fax: +46 40 415519 Email: Asa.Ekman@vv.slu.se

Summary

  • • Few studies regarding the effects of elevated atmospheric CO2 concentrations on plant lipid metabolism have been carried out. Here, the effects of elevated CO2 concentration on lipid composition in mature seeds and in leaves during the diurnal cycle of Arabidopsis thaliana were investigated.
  • • Plants were grown in controlled climate chambers at elevated (800 ppm) and ambient CO2 concentrations. Lipids were extracted and characterized using thin layer chromatography (TLC) and gas liquid chromatography.
  • • The fatty acid profile of total leaf lipids showed large diurnal variations. However, the elevated CO2 concentration did not induce any significant differences in the diurnal pattern compared with the ambient concentration. The major chloroplast lipids monogalactosyldiacylglycerol (MGDG) and phosphatidylglycerol (PG) were decreased at elevated CO2 in favour of phosphatidylcholine (PC) and phosphatidylethanolamine (PE). Elevated CO2 produced a 25% lower ratio of 16:1trans to 16:0 in PG compared with the ambient concentration. With good nutrient supply, growth at elevated CO2 did not significantly affect single seed weight, total seed mass, oil yield per seed, or the fatty acid profile of the seeds.
  • • This study has shown that elevated CO2 induced changes in leaf lipid composition in A. thaliana, whereas seed lipids were unaffected.

Introduction

The global atmospheric concentration of CO2 is increasing as a result of human activities during the last two centuries. The pre-industrial concentration of approx. 280 ppm is expected to reach approx. 700 ppm in not more than 100 years from now (IPCC, 2001). Although plants have had to adapt to changes in atmospheric CO2 concentration throughout evolutionary history (Post et al., 1990), it is the pace of the present change that is extraordinary and that raises questions about how this change in photosynthetic conditions will affect plants in the coming century.

Elevated CO2 can affect biological processes at the molecular, physiological, and ecological levels (Ward & Strain, 1999). At current CO2 concentrations, the carbon-fixing enzyme ribulose-biphosphate carboxylase/oxygenase (Rubisco) is not saturated with CO2 in photosynthesis, causing the oxygenase reaction of the enzyme to compete with the carboxylase reaction, which leads to carbon and energy loss for the organism. At least in theory, an elevated CO2 environment would increase carboxylation of Rubisco and result in more efficient photosynthesis. In a survey of 156 species, elevated CO2 indeed produced average increases in biomass production of 41 and 22% in C3 and C4 plants, respectively (Poorter, 1993), although acclimation of photosynthetic activity after long-term exposure has been seen in several species (Nie et al., 1995; Cheng et al., 1998; Ainsworth & Long, 2005). Physiological differences in carbon storage and demand (sink–source status) among plant species are thought to be important factors in their different responses to CO2, which produce different abilities to benefit from the extra carbon available in an elevated CO2 environment (Stitt, 1991; Reekie et al., 1998; Paul & Foyer, 2001; Ainsworth & Long, 2005).

While the effects of elevated CO2 concentrations on starch and sugar metabolism have been extensively studied in different plant species (for reviews, see Heineke et al., 1999; Stitt, 1991; Sharkey et al., 2004), the effects on lipid metabolism have only been dealt with in a few species. Plant lipids have been shown to be affected by environmental factors such as temperature, light, and atmospheric constituents (Harwood, 1994) and studies on wheat (Triticum aestivum) leaves have shown that alterations in lipid composition can be expected for growth at elevated CO2 (Robertson & Leech, 1995; Williams & Harwood, 1997; Williams et al., 1998a,b). As lipids are major components of cell membranes and are involved in energy storage and signalling systems, their metabolism can influence the growth and development of an organism. Therefore, this study of lipid composition in mature seeds and in leaves during the diurnal cycle of the model plant Arabidopsis thaliana is of interest for predicting how future environments in which CO2 concentrations are elevated will change the conditions for plant lipid metabolism in oil-producing plants.

Plants were grown in controlled growth chambers at ambient (approx. 380 ppm) and elevated (800 ppm) CO2 concentrations. For the first time in A. thaliana, diurnal variations in the fatty acid profile of total and separated leaf lipids were determined to elucidate whether elevated CO2 produced any differences. Mature seeds from plants grown under different nutrient supplies and CO2 concentrations were analysed with the purpose of determining whether there were any quantitative or qualitative changes in oil composition.

Materials and Methods

Growth conditions and sampling

Arabidopsis thaliana (L.) Heynh. Columbia (Col-0) seeds (Lehle Seeds, Round Rock, TX, USA) were grown in a medium consisting of peat:vermiculite:perlite (4 : 2 : 1 by volume) in controlled growth chambers (Biotronen, SLU-Alnarp, Sweden). The relative humidity was kept at 70% and the temperature was kept at a constant 22°C during the growth period. Cool-white light was provided by fluorescent tubes (215 W; Sylvania, Mississauga, Canada) with a photon flux of 200 µmol m−2 s−1 photosynthetically active radiation (PAR) (quantum sensor; Li-Cor, Lincoln, NE, USA) under a 16 h light: 8 h dark regime. The plants were watered three times a week and rotated in the chamber to avoid internal chamber effects. The CO2 concentration was measured with an infrared gas analyser (The Analytical Development Co., Hoddesdon, UK) and controlled to 800 ppm in the elevated chamber using locally developed software (Biotronen). Thirteen days after the cotyledons had opened and at defined time intervals during a 48-h period, whole leaf rosettes were cut from the root, their fresh weight (FW) was measured and then they were either immediately frozen in liquid nitrogen and saved at −80°C for lipid extraction or dried at 70°C before dry weight (DW) and starch analysis. At each sampling time three samples, each consisting of three leaf rosettes, were taken for lipid analysis. Ten leaf rosettes were pooled for starch analysis. Sampling during the night was performed in dim light to avoid light-induced fatty acid synthesis.

For seed analysis, plants were grown as already described, with the difference that half of the plants were fertilized twice a week (Rika® macro and micronutrients; Weibulls AB, Svalöv, Sweden) and the other half did not receive any fertilizers. Seeds from 12 separate plants from each treatment were harvested and stored in open tubes at room temperature before seed lipid analysis.

Biomass determination

Plants (whole leaf rosettes) were dried at 70°C until a constant weight was achieved. Single seed weight was determined by weighing 500 seeds.

Lipid analysis

Lipids from frozen leaf samples stored at −80°C were extracted according to Bligh & Dyer (1959), with the difference that 1 mm ethylenediaminetetraacetic acid (EDTA) in 0.15 m acetic acid (HAc) was added to prevent lipase activity. A small portion of the extract was evaporated under nitrogen before lipids were transmethylated to fatty acid methyl esters in 2% H2SO4 in MeOH at 90°C for 45 min. Fatty acid methyl esters, including an internal standard (methyl-heptadecanoate), were extracted into hexane and the fatty acid profile was determined by gas liquid chromatography (GC-17A; Schimadzu, Bergman Labora, Upplands Väsby, Sweden) on a fused silica capillary column (WCOT, CP-Wax 58, 50 m, 0.32 mm internal diameter; Varian, Bromma, Sweden) using a flame ionization detector.

Leaf lipids, with three replicates for each sampling time, were separated on silica thin layer chromatography (TLC) plates (Silica Gel 60; Merck, Darmstadt, Germany) in a polar system with CHCl3:MeOH:HAc:H2O (90 : 15 : 10 : 3 by volume). The different lipid classes were identified relative to standards by staining lightly with iodine, and the gel areas were scraped from plates, dried under nitrogen gas to remove traces of iodine, methylated and analysed using gas chromatography as already described.

Seeds were direct-methylated in 2% H2SO4 in MeOH at 90°C for 90 min and fatty acid profiles were determined as already described.

Starch analysis

Dried plant material (10 pooled leaf rosettes) was ground with a mortar and pestle to a fine powder from which three samples were taken. The starch concentration was determined by an enzymatic method using a total starch determination kit (Megazyme, Wicklow, Ireland) according to the manual, except that the samples were first washed twice with 80% EtOH at 80°C to remove soluble sugars.

Statistical calculations

Treatment effects were analysed by analysis of variance (ANOVA) using the general linear model (minitab 14; Minitab, State College, PA, USA), in which time in the diurnal analyses was treated as a block, and all factors were fixed. Significant mean differences between treatments were calculated using pairwise comparisons with the method of Tukey at levels P ≤ 0.05, P ≤ 0.01, and P ≤ 0.001. If necessary, data were transformed according to Box & Cox (1964) before the ANOVA and Tukeys tests were performed to stabilize variance and obtain the approximately normal distribution of residuals required for valid statistical inference.

Results

Physiological observations

Arabidopsis thaliana plants were grown at ambient and elevated CO2 concentrations in controlled climate chambers. By day 13 after the cotyledons had opened, the plants at both elevated and ambient CO2 concentrations were considered to be approximately at the same development stage before flowering. During the 2 d of sampling, plants were at the development stages designated 1.07–1.10 and 1.08–1.11 (i.e. they carried 7–10 and 8–11 leaves in the rosette, respectively) in the ambient and elevated CO2 chambers, respectively, according to the growth stages of A. thaliana outlined by Boyes et al. (2001). Biomass production at elevated CO2 was 115% higher (P ≤ 0.001) than that at ambient CO2 during the 2-d sampling period (Fig. 1a). Plants at elevated CO2 contained less water in green tissue during the whole diurnal cycle. The percentage DW to FW differed by 0.67 units (P ≤ 0.001; Fig. 1b). Plants at both ambient and elevated CO2 concentrations showed the same diurnal variations in percentage DW, having a lower water content during the day. Flowering time did not differ between the two different CO2 concentrations.

Figure 1.

Biomass production per plant (a), dry weight (DW) as a percentage of fresh weight (b), and starch concentration on a dry weight basis (c) in Arabidopsis thaliana leaves during the diurnal cycle (dashed areas are dark hours) at ambient (open circles) and elevated (closed circles) CO2 concentrations. Results are the mean ± standard deviation for 10 plants. Results for ambient and elevated CO2 treatments are significantly different (P ≤ 0.001).

Starch content of leaves

The starch content in leaves was higher in the daytime compared with the night in both CO2 regimes (Fig. 1c). The mean starch content on a DW basis during the sampling period was 2.3% units higher at elevated CO2 compared with the ambient concentration (P ≤ 0.001). The leaf starch concentration was higher by the end of the day under elevated CO2 than under ambient CO2 (c. 9.5 and 6.0% of DW, respectively). Starch storage was depleted in plant leaves at the ambient CO2 concentration by the end of the night, whereas at elevated CO2 leaves still contained starch concentrations of approx. 1–2% of DW (Fig. 1c).

Total leaf lipid content and lipid class profile

Total fatty acid concentration (based on the weight of acyl groups of total lipids) in green tissue on a DW basis was significantly lower at elevated CO2 than at the ambient CO2 concentration (3.2 and 3.9%, respectively; a difference of 0.7% units; P ≤ 0.001; Table 1). The membrane lipid composition (mol%) was altered by CO2 concentration such that phosphatidylcholine (PC) and phosphatidylethanolamine (PE) were increased by 2.1 and 2.6% units, respectively, and phosphatidylglycerol (PG) and monogalactosyldiacylglycerol (MGDG) were decreased by 0.9 and 3.6% at the elevated CO2 concentration compared with the ambient concentration (P ≤ 0.001; Table 1). The amount of digalactosyldiacylglycerol (DGDG) was not significantly influenced by CO2 concentration.

Table 1.  Relative amounts of the major membrane lipids (expressed as mol% of their sum) and total fatty acid concentration (dry weight percentage) in Arabidopsis thaliana leaves at ambient (Am) and elevated (El) CO2 concentrations
LipidAmElDifference
  1. Results are the diurnal mean ± standard deviation from the analysis of variance (ANOVA). Significant differences compared with ambient CO2 concentration are shown: ***, P ≤ 0.001; ns, not significant.

  2. DGDG, digalactosyldiacylglycerol; DW, dry weight; MGDG, monogalactosyldiacylglycerol; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol.

DGDG (%)18.9 ± 0.918.8 ± 0.9−0.1 ns
MGDG (%)40.2 ± 1.736.6 ± 1.7−3.6***
PC (%)21.4 ± 1.323.5 ± 1.3+2.1***
PE (%)12.0 ± 1.314.6 ± 1.3+2.6***
PG (%) 7.4 ± 0.6 6.5 ± 0.6−0.9***
Total fatty acids (DW%) 3.9 ± 0.2 3.2 ± 0.2−0.7***

Fatty acid profile of total and individual leaf membrane lipids

The fatty acid profile (mol%) of total leaf lipids showed large diurnal variations, with 18:1 increasing twofold (from approx. 2.5 to 5%) during the first 9 h of the light period and decreasing correspondingly during the night (Fig. 2a). At the same time, the sum of 18:2 and 18:3 decreased during the day and increased during the night. The only difference in the diurnal fatty acid profile pattern of total leaf lipids produced by elevated CO2 was a lower content of 18:2 and 18:3 and a higher content of 18:0, especially during the second 24-h sampling period (Fig. 2a). The fatty acids 16:0 and 16:1 showed very small or no diurnal variations in the total leaf lipid fraction (data not shown).

Figure 2.

Fatty acid profile (mol%) in Arabidopsis thaliana leaves of total lipid (a), phosphatidylcholine (PC) (b), phosphatidylglycerol (PG) (c), and phosphatidylethanolamine (PE) (d) during the diurnal cycle (dashed areas are dark hours) at ambient (open symbols) and elevated (closed symbols) CO2 concentrations. Acyl groups are 16:0 (▴), 16:1trans (▪), 18:0 (▾), 18:1 (◆), 18:2 (▸), 18:3 (◂), and the sum of 18:2 and 18:3 (•). Note that in the graph of PG (c) the symbols ▪ and • should be read on the right axis. Results are the mean ± standard deviation for three samples.

Among the leaf lipid classes, PC, PG and PE showed diurnal variations in the fatty acid profile (Fig. 2b–d). MGDG and DGDG showed very small variations irrespective of CO2 concentrations or diurnal rythms (data not shown). The diurnal variation in 18:1 seen in the total lipid fraction was more pronounced in PC, with an increase from approx. 5 to 13% by the end of the day. Variation in 18:1 could also be seen in PE and PG, although not of the same magnitude. The reversed diurnal changes of 18:2 and 18:3 seen in the total lipid fraction could also be observed in PC (Fig. 2b), and to some extent in PE (Fig. 2d). However, the changes in 18:2 and 18:3 in PG did not correspond to the changes in 18:1, and the diurnal variations observed were not repeated diurnally (Fig. 2c).

The fatty acids 16:0, 16:1, and 16:1trans (trans-Δ3-hexadecenoic acid) showed diurnal variations in PC, PG, and PE, but variations did not show exactly the same trends during the first and second 24-h periods (Fig. 2b–d). However, the effect of elevated CO2 on the relative amounts of 16:0 and 16:1trans in the chloroplast lipid PG was striking. The ratio of 16:1trans to 16:0 was reduced by 25% compared with the ambient CO2 concentration (P ≤ 0.001; Fig. 3). These two fatty acids in PG also showed a diurnal behaviour, with higher 16:1trans to 16:0 ratios during the first hours of light at both CO2 concentrations.

Figure 3.

The ratio of 16:1trans to 16:0 in phosphatidylglycerol (PG) (based on mol%) in Arabidopsis thaliana leaves at ambient (open circles) and elevated (closed circles) CO2 concentrations. Results are the mean ± standard deviation for three samples. Results for ambient and elevated CO2 treatments are significantly different (P ≤ 0.001).

Seed harvest and seed lipid analysis

CO2 concentration and nutrient availability were shown to have a synergistic effect on the oil content of seeds (Fig. 4a). When there was a good nutrient supply, CO2 had no significant effect on seed oil content. However, at ambient CO2, nutrient limitations decreased seed oil content by 20%, whereas at elevated CO2 the decrease was only 10% (P ≤ 0.001). Seed weight and seed harvest were not significantly affected by the CO2 concentration; however, nutrient limitations decreased single seed weight by 26%, and total seed harvest by 96% (P ≤ 0.001; Fig. 4b,c).

Figure 4.

The effect of CO2 concentration and nutrient supply on oil content per seed (a), total seed weight harvest per plant (b) and single seed weight (c) in Arabidopsis thaliana plants grown at either ambient (Am) or elevated (El) CO2 concentration and under either good (+) or scarce (–) nutrient supply. Results are the mean ± standard deviation for 10 (a, b) and five (c) plants. Letters indicate significant differences between the means for Tukey's test (P ≤ 0.001).

CO2 concentrations did not affect seed fatty acid composition, but nutrient limitations altered the fatty acid profile significantly (at P ≤ 0.001 for all fatty acids; Fig. 5). Total seed lipids from plants grown under nutrient limitations showed a relative increase of 16:0 (+0.8%), 18:0 (+0.4%), 18:2 (+1.8%), and 20:0 (+0.4%) and a decrease of 18:1 (−0.8%), 18:3 (−1.7%), and 20:1 fatty acids (−1.2%) compared with seed oils from plants grown under a good nutrient supply (P ≤ 0.001).

Figure 5.

The effect of CO2 concentration and nutrient supply on the fatty acid profile of total seed lipids in Arabidopsis thaliana plants grown at either ambient (Am) or elevated (El) CO2 concentration and under either good (+) or scarce (–) nutrient supply. Results are the mean ± standard deviation for 12 plants. Letters indicate significant differences between the means for Tukey's test (P ≤ 0.001).

Discussion

Physiological observations

In studies of plants exposed to elevated CO2 concentrations, it is of utmost importance to consider the possible side effects of increased development rates produced by faster growth. However, by day 13 after the cotyledons had opened, the plants at elevated and ambient CO2 concentrations were considered to be at approximately the same early developmental stage before flowering, defined by the number of leaves (Boyes et al., 2001). Therefore, changes observed in leaves at this age could not be attributed to increased development rates. The enhanced growth effects of plants exposed to elevated CO2 concentration have been extensively studied, and the approximate doubling of leaf biomass production observed in this study of the C3 plant A. thaliana is in agreement with previous observations (Poorter, 1993). Light microscopy indicated that no change in leaf thickness was induced by elevated CO2 (data not shown); this parameter could therefore not explain the large increase in leaf biomass at elevated CO2. In a previous study by Teng et al. (2006), the leaf thickness of A. thaliana was increased by 5% at elevated CO2. However, it should be noted that samples were taken at a later development stage than in our study. Increased carbohydrate content (see next section) together with increased leaf surface area can probably explain the increase in biomass induced by elevated CO2 in our study.

The results of this study indicate that plants contain proportionally less water at elevated CO2 than plants grown at ambient CO2, in accordance with earlier findings (Idso et al., 1988). However, no difference in the diurnal variations of water content could be observed. The higher %DW of FW of leaves at elevated CO2 is mainly a result of the increased carbohydrate content (see next two sections).

Starch content of leaves

It is well known that elevated CO2 induces accumulation of starch in plants, including A. thaliana (Cheng et al., 1998; Li et al., 2006a; Teng et al., 2006), and sink capacity is thought to exert a strong control on the response to elevated CO2 (Rasse & Tocquin, 2006). However, the diurnal variations of leaf starch content in A. thaliana exposed to elevated CO2 have not been reported before. Starch storage was depleted in plant leaves at the ambient CO2 concentration by the end of the night, whereas at elevated CO2 leaves still contained approx. 2% starch on a DW basis. This corresponds to a mobilization rate of starch during the night on a DW basis of, on average, 0.75 and 0.94 µg mg−1 h−1 at ambient and elevated CO2 concentrations, respectively, which is an increase of 25% at elevated CO2. This corroborates the findings of a study on Ricinus communis (Grimmer & Komor, 1999), which suggested that plants grown at ambient CO2 concentrations are operating under sink limitation during the day and under source limitation during the night.

Total leaf lipid content during the diurnal cycle

The total fatty acid concentration in green tissue on a DW basis was decreased from 3.9 at the ambient CO2 concentration to 3.2% at the elevated concentration. Although an increased amount of starch in leaves at the elevated CO2 concentration was observed in this study, this alone cannot explain the decrease in the amount of leaf lipids. In a previous study, the leaf cellulose content of A. thaliana grown at elevated CO2 was found to be increased, probably because of an increase in cell wall thickness (Teng et al., 2006). Taken together, the increases in starch and cellulose, which increase the DW content, may thus explain the decrease in fatty acid concentration.

Elevated CO2 produced differences in the relative amounts of the leaf membrane lipids of A. thaliana. We suggest that elevated CO2 reduces the need for the chloroplast lipids MGDG and PG in favour of the extraplastidic lipids PC and PE in A. thaliana. In previous studies, elevated CO2 produced no significant changes in leaf lipid class composition in young wheat leaves (Williams et al., 1998a), whereas in mature wheat leaves an increase in PC content was observed (Williams et al., 1998b). In the macroalga Ulva rigida, the PE:PG ratio was increased by growth at high CO2 (Gordillo et al., 2001), which is consistent with our findings. Elevated CO2 concentrations have previously been shown to increase the proportion of stroma to grana thylakoid membranes in the chloroplasts in four tree species (Griffin et al., 2001), in Beta vulgaris (Kutik et al., 1995) and recently in A. thaliana (Teng et al., 2006). It should be noted that this proportion is unaffected by growth temperatures and light intensities (Albertsson & Andreasson, 2004). The relative decrease in MGDG at elevated CO2 can be related to the increased stroma to grana thylakoid ratio, given that grana thylakoids are enriched with MGDG compared with stroma thylakoids (Gournaris et al., 1983). This suggestion is not consistent with the lateral homogeneity of lipids in the thylakoid membranes suggested by Duchene & Siegenthaler (2000).

Other changes in chloroplast structure have been shown to be induced by elevated CO2 in previous studies. Elevated CO2 resulted in a decrease in the concentrations of proteins involved in photosynthesis in A. thaliana and in wheat leaves exposed to elevated CO2 (Robertson & Leech, 1995; Cheng et al., 1998). Further, transcript levels for proteins involved in photosynthetic reactions in A. thaliana and wheat were shown to decrease at elevated CO2 (Nie et al., 1995; Cheng et al., 1998; Li et al., 2006a). These findings suggest that plants acclimate to an environment with elevated CO2 by adjusting the amounts of the constituents of the photosynthetic apparatus. Therefore, it is not surprising that the difference in lipid composition we observed at elevated CO2 seems to be targeted to the chloroplast.

Fatty acid profile of leaf membrane lipids during the diurnal cycle

Diurnal variations in 18-carbon fatty acids of total leaf lipids, with 18:1 increasing twofold during the day and decreasing correspondingly during the night, have only been reported once before, in spinach (Spinacia oleracea) leaves (Browse et al., 1981). A light dependence of de novo fatty acid synthesis has been reported previously (Ohlrogge, 1997; Hunter & Ohlrogge, 1998). This diurnal pattern of polyunsaturation in leaf lipids, including the reversed decrease of the sum of 18:2 and 18:3 during day and its increase during the night, has been suggested to be a result of the light dependence of fatty acid synthesis combined with the non-light dependence of further desaturation of 18:1 (Browse et al., 1981). It is pertinent to point out here that sampling time substantially influences the fatty acid profile of green plant tissue, including that of A. thaliana, during the diurnal cycle.

The diurnal changes in 18-carbon fatty acids were more pronounced in PC, which is in agreement with this lipid as a main site for further desaturation of 18:1 (Stymne & Appelqvist, 1978; Stobart et al., 1980). That the same trends could be seen in PE, to some extent, is consistent with previous studies suggesting that this lipid could also be a site for further desaturation of 18:1 (Sánchez & Stumpf, 1984; Griffiths et al., 1988). The fatty acid 18:1 was also increased in PG during the day and decreased during the night, but these changes did not match corresponding changes in 18:2 and 18:3 and were not repeated diurnally.

The only difference induced by elevated CO2 in the diurnal fatty acid profile pattern of the 18-carbon fatty acids was an increase in 18:0 and a reversed decrease in 18:2 and 18:3 in total lipids during the second 24-h sampling period, suggesting faster de novo fatty acid synthesis during the first hours of light at elevated compared with ambient CO2. At the same time, 18:1 was unaffected by CO2 concentrations, indicating that Δ9-desaturation could be a rate-limiting step in green tissue in an elevated CO2 environment. However, the same pattern was not observed during the first 24-h sampling period, for which we cannot presently give any explanation. The 16-carbon fatty acids showed diurnal variations in PC, PG and PE, but trends were not repetitive, suggesting a more complex regulation of these fatty acids compared with the diurnal dependence of 18-carbon unsaturated fatty acids.

The effect of elevated CO2 on the relative amounts of 16:0 and 16:1trans in the chloroplast lipid PG was pronounced. Interestingly, the ratio of 16:1trans to 16:0 in PG from leaves at elevated CO2 was decreased by 25% compared with that at ambient CO2. The 16:1trans fatty acid is specific for the chloroplast lipid PG in thylakoid membranes in eukaryotic photosynthesizing organisms. The involvement of 16:1trans-containing PG in the stabilization of light-harvesting complex II (LHCII) trimerization and the formation of grana stacks has been extensively studied (for reviews, see Siegenthaler, 1991; Trémolières & Siegenthaler, 1998). In a recent study of winter rye (Secale cereale), it was shown that the ratio of 16:1trans to 16:0 in PG was positively correlated with the oligomeric form of LHCII, and the ratio increased with growth irradiance and temperature (Gray et al., 2005). LHCII is concentrated in the grana thylakoids (Andersson & Anderson, 1980; Danielsson et al., 2004) and, as 16:1trans PG is known to stabilize LHCII oligomers, this suggests that the relative decrease in 16:1trans in the total leaf fraction of PG seen in this study is a result of the decrease in the grana to stroma thylakoid ratio induced by elevated atmospheric CO2, as discussed in the previous section. Again, this assumption is not consistent with the lateral homogeneity of 16:1trans acid-containing PG in the thylakoid membranes suggested by Duchene & Siegenthaler (2000) but is in agreement with other studies suggesting a higher level of 16:1trans PG in thylakoid fractions enriched with oligomeric LHCII (Trémolières, 1991).

Seed harvest and lipid analysis of seeds

It is of agricultural interest to know whether a future environment with elevated CO2 will alter the oil yield and lipid composition of oil crops. Arabidopsis thaliana is a plant that produces seeds with a high oil content (approx. 40%) and can be regarded as a plant that has a strong carbon sink during the seed-filling period. Arabidopsis thaliana is also in the same family as the important oil seed rape (Brassica napus) and is often used as a model for this crop, although the applicability of this comparison should be considered with care in the light of the differences in seed characteristics between the two species (Li et al., 2006b). Different effects of elevated CO2 on oil yield in oil crops have been reported. No change in oil content could be observed in peanut (Arachis hypogaea) (Wu et al., 1997) and oil seed rape (Frick et al., 1994). CO2 enrichment stimulated the seed yield of soybean (Glycine max), but the effects on seed oil content were variable, with differences among cultivars (Heagle et al., 1998; Ziska et al., 1998).

Single seed weight and total seed weight harvest per plant were not significantly influenced by CO2 concentration in this study, and no change in time to maturation of seeds could be observed. These results are consistent with those of a study on the effect of elevated CO2 on the growth and reproduction of eight different genotypes of A. thaliana (Ward & Strain, 1997). No differences in seed number, reproductive mass or day to first flower could be observed between the 350 and 700 ppm CO2 treatments. However, a significant increase was observed between the 280 and 350 ppm CO2 treatments (Ward & Strain, 1997). In other studies of the influence of elevated CO2 on plant reproductive output and time to reproduction, various results for different species have been presented, suggesting species-specific responses (Ward & Strain, 1999; Poorter & Navas, 2003).

Scarce nutrient supply drastically decreased the reproductive output of A. thaliana. An approximately 20-fold reduction in total seed weight harvest per plant was observed in this study, and single seed weight was decreased by 20%, irrespective of CO2 concentration. The results indicate that limited nutrient conditions will still stress plants in a future atmosphere with elevated CO2, as at ambient CO2 concentrations.

The oil content of A. thaliana seeds was not altered by elevated CO2 concentration under a good nutrient supply, but nutrient limitations were associated with pronounced decreases in oil content. The decrease in seed oil content caused by nutrient limitations was larger at ambient than at elevated CO2. Nutrient limitation altered the fatty acid profile of seeds significantly, but the influence of CO2 concentrations was very small. The changes induced by a nutrient-limited environment, with small relative increases of 16:0, 18:0, 18:2, and 20:0, and decreases of 18:1, 18:3, and 20:1 fatty acids, indicate that the oil quality is somewhat sensitive to the nutrient status of the plant. As increased light intensities have been shown to positively influence seed characteristics such as seed yield, seed size, and oil content in A. thaliana (Li et al., 2006b), it would be of value to investigate whether a combined increase of light and atmospheric CO2 concentration would influence these parameters in an additive manner.

The increased carbon flow through photosynthesis at elevated CO2 concentrations demonstrated by the higher amounts of starch in the leaves of A. thaliana does not increase the total amount of carbon transport to the sink during seed filling. If A. thaliana is regarded as a model for Brassicaceae oil crops, our results indicate that future increases in atmospheric CO2 concentration will not produce any significant alteration in oil yield in these crops. However, species-specific responses can be expected, and more studies need to be carried out on other oil-producing plants before any general conclusions can be drawn.

Acknowledgements

The work was supported by grants from the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS) and SLU-Alnarp faculty grant for LTH-SLU collaboration. We would also like to thank Per-Åke Albertsson for valuable comments and discussions.

Ancillary