Department of Biological Sciences, John Tabor Laboratories, University of Essex, Wivenhoe Park, Colchester CO4 3SQ, UK and Environmental Biology and Instrumentation Division, Building 318, Brookhaven National Laboratory, Upton, New York 11973, USA
Lolium perenne, a main component species in managed grassland, is well adapted to defoliation, fertilization, and regrowth cycles; and hence, to changes in the assimilatory carbon source-sink ratio. In the Swiss Free Air CO2 Enrichment experiment the source-sink ratio is (i) increased by elevated partial pressure of CO2 (pCO2), (ii) decreased by enhanced carbon use under high N fertilization, and (iii) gradually increased during regrowth after defoliation. Since sucrose synthesis plays a central role in leaf carbohydrate metabolism in this fructan-accumulating species, we investigated how sucrose-phosphate synthase (SPS) responds to the differing assimilatory carbon fluxes and source-sink ratios in the field. Assimilatory carbon flux, as estimated by leaf gas exchange, strongly depended on pCO2. Surprisingly, the SPS content per leaf area did not increase with pCO2, but increased with N fertilization. During later regrowth, when a dense canopy had formed, the SPS content decreased; in particular, SPS was decreased at high N under elevated pCO2. Further, the higher assimilatory carbon flux through SPS at elevated pCO2 was accompanied by a higher activation state of SPS. The SPS content correlated very strongly with the ratio of free sucrose to free amino acid in leaves, which represents the carbon source-sink ratio. Hence, SPS content in L. perenne appears to be regulated by the current, strongly nitrogen-dependent, source-sink relation.
Plants of managed grassland are subject to severe and quick changes in their carbon and nitrogen source-sink ratio by regular defoliation with subsequent fertilization: (i) defoliation diminishes the leaf area and therefore the assimilatory capacity of the carbon source; (ii) fertilization increases N availability for regrowth, thus increasing the carbon sink. In the course of regrowth, carbon assimilatory capacity is quickly re-established with increasing leaf area, whereas N available for plant growth decreases rapidly, so that further growth is often limited by lack of N.
In such a scenario, the extent of change in the source-sink ratio after defoliation depends on the amount of N fertilizer applied and on the partial pressure of CO2 (pCO2). Increasing N enhances the carbon sink, and increasing pCO2 enhances the carbon source by stimulating photosynthesis ( Bowes 1991; Long & Drake 1992). In perennial ryegrass (Lolium perenne L.), an important component of grassland in mild temperate regions, shoot biomass increases with N fertilization, and a further biomass increase by means of elevated pCO2 depends on sufficient N supply ( Schenk, Jäger & Weigel 1996; Hebeisen et al. 1997 ). Leaf carbohydrate metabolism under these changing source-sink relations, with special emphasis on its interaction with pCO2 and N supply, was the focus of our interest.
In source leaves of L. perenne, only very little carbohydrate is stored as starch ( Fischer et al. 1997 ); most of the photosynthetically fixed carbon appearing there is immediately directed into the synthesis of sucrose, which is used for storage, fructan synthesis and export ( Pollock & Cairns 1991; Isopp et al. 2000 ). Since sucrose-phosphate synthase (SPS) is the key enzyme of sucrose synthesis ( Huber & Huber 1996), it ultimately sustains the main assimilatory carbon flux.
Indeed, the SPS content of leaves has been shown to be regulated mainly to fit the photosynthetic rate. When the source-sink balance of plants is altered, photosynthesis and SPS generally change likewise and in parallel. SPS increases along with photosynthesis in high light ( Klein, Crafts-Brandner & Salvucci 1993), high N ( Makino, Nakano & Mae 1994), after a reduction in the number of leaves ( Suwignyo et al. 1995 ) or, as in most studies, under elevated pCO2 (e.g. Seneweera et al. 1995 ; Wang & Nobel 1996; Hussain, Allen & Bowes 1999). Even in the two reports in which SPS was decreased due to elevated pCO2, the photosynthetic rate was also decreased due to acclimation ( Peet, Huber & Patterson 1986; Socias, Medrano & Sharkey 1993). Yet, in a recent study with 16 species, no correlation was found between SPS content and acclimation of Rubisco protein; but leaf gas exchange rates, however, were not measured ( Moore et al. 1998 ). In L. perenne, the photosynthetic rate remains unchanged when the source-sink ratio differs after defoliation or in different N treatments ( Rogers et al. 1998 ). Hence, the question arises whether SPS changes in accordance with changes in the photosynthetic rate or rather in accordance with source-sink relations.
The source-sink ratio is strongly dependent on the carbon and nitrogen status of the plant. This status is indicated by the sugar and amino acid content ( Koch 1997). In tobacco plants, in which the carbon source-sink ratio was increased by withholding N, the ratio of free hexose to free amino acid correlated with a decline in photosynthetic activity ( Paul & Driscoll 1997).
The Swiss Free Air CO2 Enrichment (FACE) experiment offers a unique opportunity to investigate the response of SPS to different source-sink relations under field conditions. In the FACE experiment, swards of L. perenne are subjected to a regular cutting regime, and are additionally influenced in their source-sink relations by growing at ambient or elevated pCO2, combined with low or high rates of N fertilizer application ( Fischer et al. 1997 ; Hebeisen et al. 1997 ; Rogers et al. 1998 ). An experiment was designed using those swards to test, under field conditions, whether SPS in L. perenne leaves, like Rubisco, is influenced by source-sink relations, as indicated by the sugar to amino acid ratio, or whether SPS is solely dependent on the source activity, as indicated by the photosynthetic rate. Further, as a secondary objective, the main in vitro regulatory properties of L. perenne SPS had to be established.
MATERIALS AND METHODS
Swards of Lolium perenne L. (cv. Bastion) were grown in the FACE facility in Eschikon near Zürich, Switzerland ( Fischer et al. 1997 ). The facility consists of three pairs of control and fumigated plots, which are left at ambient pCO2 (35 Pa) and kept at elevated pCO2 (60 Pa), respectively. The swards were established in autumn 1995, with subplots maintained at a low (14 g N m−2 a−1) and a high (56 g N m−2 a−1) rate of nitrogen supply (as NH4NO3) and all defoliated at 5 cm above ground five times per year. Plants with a reduced source-sink capacity were sampled shortly after a defoliation event at regular intervals during 24 h starting at 0700 h Central European Summer Time on 25 June 1997 (9 d after defoliation, DAD; 7 d after fertilization). Plants with a high carbon source-sink ratio were sampled similarly before the next defoliation when leaves formed a dense canopy (9 July, 23 DAD). From each plot, samples of 10 leaves (for SPS determination) and 20 leaves (for chemical analysis) were taken at each sampling time. Since grass leaves represent a linear developmental gradient with the tip the oldest tissue, the distal 4 cm of the exposed blade of the emerging leaf of a tiller, assessed by the fact that it was the longest leaf in which the ligule had not emerged, was selected. Area and fresh mass of the sampled material were determined on the samples for chemical analysis; then these samples were plunged into liquid N2, freeze dried and finely ground prior to analysis. The leaf samples for SPS determination were immediately frozen in liquid N2 and stored at −80 °C until analysis.
Leaf gas exchange
Leaf gas exchange rates were measured at the same time of leaf sampling using open gas-exchange systems (CIRAS-1, PP Systems, Hitchin, UK), as described by Rogers et al. (1998) . Leaf gas exchange rates were measured in situ under growth pCO2 and ambient temperature, using incident light. In each plot six to seven individual leaves were selected as above and measured at each sampling time. The daily net CO2 uptake for the 24 h period was calculated by integrating the CO2 exchange rates. The mean of the photosynthetic rates at the beginning and end of each interval was multiplied by the duration of the interval; these values were added up to give the daily net CO2 uptake. Photosynthetic rates around noon (from 1100 to 1500 h) were used to estimate the maximum carbon flux in leaves.
SPS activity was measured using a procedure modified from Walker & Huber (1989). Frozen leaf samples were ground in a chilled mortar with four volumes of cold extraction buffer containing 50 mol m−3 3-[N-morpholino]-propanesulfonic acid (Mops)-KOH pH 7·4, 15 mol m−3 MgCl2, 1 mol m−3 EDTA, 5 mol m−3 DTT, 0·1% (v/v) Triton X-100, 0·1% (w/v) bovine serum albumin, 2 mol m−3 benzamidine and 2 mol m−3ɛ-aminocapronic acid. Extracts were passed through a double layer of miracloth, centrifuged for 3 min at 13 000 g, and 200 mm3 of the supernatant were desalted in prespun Poly Prep Chromatography columns (BioRad, Munich, Germany) filled with 2 cm3 Biogel P6 (BioRad) and equilibrated with extraction buffer without Triton. All steps were carried out quickly at temperatures between 0 and 4 °C. The SPS activity was assayed under saturating (Vmax) and limiting (Vlim) substrate conditions. The Vmax assay consisted of 50 mol m−3 Mops-KOH pH 7·4, 15 mol m−3 MgCl2, 1 mol m−3 EDTA, 8 mol m−3 fructose-6-phosphate (Fru-6-P), 32 mol m−3 glucose-6-phosphate (Glc-6-P), 10 mol m−3 uridine-diphospho-glucose, and 45 mm3 extract (equivalent to 9 mg of fresh mass) in a final volume of 100 mm3. The Vlim assay was the same as above, except that it contained 3 mol m−3 Fru-6-P, 12 mol m−3 Glc-6-P and, additionally, 10 mol m−3 inorganic phosphate. Assays were incubated for 10 min at 25 °C and stopped by boiling for 3 min. Product formation in the assays was linear up to 20 min. The supernatant of the centrifuged assay was diluted and analysed for sucrose by high-performance liquid chromatography (HPLC) as described below. Preliminary experiments had shown that sucrose-6-phosphate was only present in trace amounts in the assay (~0·1% of sucrose formed); thus sucrose-6-phosphate could be neglected. Invertase activity under assay conditions was negligible. SPS activity is expressed as the rate of sucrose formation. The activation state of SPS in per cent was calculated as (Vlim/Vmax) × 100.
A procedure modified from Reimholz et al. (1997) was used to quantify SPS protein. Frozen plant tissue was ground in liquid N2. The powder was immediately added to hot SDS gel loading buffer, the slurry heated for 4 min to 90 °C, and subsequently centrifuged. Total proteins equivalent to 20 mm2 leaf area were separated by sodium dodecyl sulphate (SDS)-polyacrylamide gel electrophoresis (PAGE) on 7·5% gels (Mini-Protean II, BioRad) and transferred to PVDF membranes (BioRad) by semi-dry blotting (Multiphor II Novablot Electrophoresis Transfer Kit, Pharmacia, Dübendorf, Switzerland). The SPS was immunostained using polyclonal rabbit antibodies raised against SPS fragments 30 and 90 (1 : 1 mixture in a dilution 1 : 2000) of maize ( Bruneau et al. 1991 ), and visualized using the chemiluminescence assay with horse-radish peroxidase (Boehringer, Mannheim, Germany). The SPS was quantified on scanned autoradiographs using the software package ScionImage (Scion Corp., Frederick, MD, USA). On each gel the same amount of the same maize leaf extract was loaded as a reference to allow comparison between gels.
Starch was determined as in Fischer et al. (1997) . Water-soluble carbohydrate was determined in hot water extracts of lyophilized material: extracts and hydrolysed extracts were deionized by the use of ion-exchange resins (Dowex 50–100 mesh: 50WX8, H+-form; 1 X 8, formate form; Fluka, Buchs, Switzerland). Free glucose, fructose and sucrose in extracts were separated by HPLC on an anion-exchange column (Carbopac PA100, 250 × 4 mm, Dionex, Sunnyvale, CA, USA) using 100 mol m−3 NaOH as the eluent; for sucrose in the SPS assays 150 mol m−3 NaOH was used. The HPLC system and instrument settings were as described in Lüscher, Frehner & Nösberger (1993). Compound water soluble carbohydrates (mainly sucrose and fructan) in extracts were hydrolysed by incubation of an aliquot of the extract with 750 U β-Fructosidase (Boehringer) in acetate-buffer (25 mol m−3, pH 5·0) at 40 °C for 15 h. The reaction was stopped by boiling for 5 min. Total glucose and fructose in the hydrolysed extracts were determined as described above. The fructan content was calculated as total hexose content minus the content of free hexoses and hexoses from sucrose. The TNC was calculated as the sum of the individually determined carbohydrate contents.
Amino acid and protein determination
The content of free amino acid was determined in 80%-ethanol extracts with ninhydrin, following the procedure of Yemm & Cocking (1955). Total protein content was determined in 100 mol m−3 NaOH extracts from lyophilized material using the Bradford dye-binding procedure (Bio-Rad) and bovine serum albumin as standard.
Statistical data analysis
Statistical analyses were carried out as in Fischer et al. (1997) using the General Linear Model procedure of the SAS statistical analysis package (version 6·12, SAS Institute, Cary, NC, USA) to determine the significance of observed differences at the 95% confidence level.
Photosynthesis and carbohydrate content
In young source leaves of L. perenne, photosynthetic rates per leaf area around noon were strongly increased under elevated pCO2 (+ 36%, P = 0·042), whereas they were not influenced by N or time during regrowth ( Table 1). When expressed on a protein basis ( Table 1), photosynthetic rates were highly stimulated by elevated pCO2 (+ 62%, P = 0·039), whereas they decreased at high N (P = 0·026) and increased late in regrowth (P < 0·001). The daily net CO2 uptake of leaves was significantly stimulated by elevated pCO2 (P < 0·001; Table 1) and late in regrowth (P < 0·001). There was no overall effect of N fertilization on daily net CO2 uptake (P = 0·648).
Table 1. Photosynthesis, sucrose-phosphate synthase and total non-structural carbohydrate content in source leaves of Lolium perenne from the Swiss FACE. Plants were grown at ambient and elevated pCO2, combined with two levels of N supply, and sampled early (9 DAD) and late (23 DAD) after partial defoliation. Net photosynthetic CO2 uptake rate (A) was measured around noon and expressed on a leaf area as well as on a protein basis. SPS activity (Vmax) was expressed on a protein basis. Daily net CO2 uptake was calculated from the photosynthetic CO2 uptake rates of young source leaves measured at intervals during 24 h. To allow a direct comparison with CO2 exchange rates, total non-structural carbohydrate content (TNC) at 0700 h is expressed in carbon equivalents
35 Pa CO2
60 Pa CO2
35 Pa CO2
60 Pa CO2
Mean values (± SE, n = 3; for A, n = 9) within a given parameter followed by the same superscript letter are not significantly different (P < 0·05).
A (μmol m−2 s−1)
14·8 ± 0·8a
20·2 ± 2·2b
14·2 ± 1·2a
17·1 ± 1·2a
14·6 ± 1·1a
18·7 ± 1·9b
15·1 ± 1·1a
21·0 ± 1·7b
A (μmol g−1 s−1)
3·59 ± 0·19a
6·88 ± 0·76b
6·06 ± 0·52bd
9·12 ± 0·45e
3·50 ± 0·25a
4·98 ± 0·43cd
4·86 ± 0·42c
8·19 ± 0·67e
SPS Vmax (μmol g−1 s−1)
0·61 ± 0·03a
0·87 ± 0·07bc
0·64 ± 0·04ae
0·85 ± 0·05bc
0·80 ± 0·03bd
0·95 ± 0·04c
0·83 ± 0·04bd
0·75 ± 0·03de
CO2 uptake (mmol C m−2 d−1)
394 ± 10a
665 ± 32bc
518 ± 23d
686 ± 45c
365 ± 20a
573 ± 39bd
547 ± 16d
823 ± 53e
TNC (mmol C m−2)
50 ± 6ab
122 ± 9a
263 ± 73d
632 ± 13e
36 ± 9a
60 ± 7ac
115 ± 32bc
197 ± 50d
The TNC content of leaves was higher at elevated pCO2 (P < 0·001; Table 1), even at 9 DAD, when plants still had a small leaf area and therefore a low carbon source-sink ratio. Free hexoses and sucrose were only slightly enhanced under elevated pCO2 ( Fig. 1). The high N fertilization rate led to lower carbohydrate contents within each CO2-treatment. The TNC content increased significantly by three- to five-fold from 9 to 23 DAD (P < 0·001). The increases in TNC were mainly due to higher contents of the storage compound fructan ( Fig. 1). The lowest TNC content was observed at 9 DAD, high N and ambient pCO2, and the highest at 23 DAD, low N and elevated pCO2.
The protein content of leaves of L. perenne was generally increased by high N (P = 0·006), but was decreased by elevated pCO2 (P = 0·022, Table 2). The decrease due to elevated pCO2 was more pronounced under low N supply. In leaf tissue of regrown plants (23 DAD), the protein content was lower than in leaf tissue of the same physiological age shortly after fertilization (9 DAD, P < 0·001). Similarly, the content of free amino acid was strongly increased at high N (P < 0·001), but it decreased with time (P < 0·001; Table 2). However, growth at elevated pCO2 led only to a marginal decline of free amino acid content (P = 0·527).
Table 2. Protein, free amino acid content, hexose : amino acid ratio and sucrose : amino acid ratio in source leaves of Lolium perenne from the Swiss FACE. Plants were grown at ambient and elevated pCO2, combined with two levels of N supply, and sampled early (9 DAD) and late (23 DAD) after partial defoliation
35 Pa CO2
60 Pa CO2
35 Pa CO2
60 Pa CO2
Mean values (± SE; n = 9 for protein; n = 33 for amino acid and hexose : amino acid ratio; n = 3 for sucrose : amino acid ratio) within a given parameter followed by the same superscript letter are not significantly different (P < 0·05)
Protein (g m−2)
4·26 ± 0·25a
3·20 ± 0·28b
2·47 ± 0·10cd
1·96 ± 0·12d
4·36 ± 0·21a
4·29 ± 0·26a
3·13 ± 0·14b
2·66 ± 0·09bc
Amino acid (mmol m−2)
3·49 ± 0·28a
2·84 ± 0·22b
2·05 ± 0·09d
1·72 ± 0·05d
9·24 ± 0·52c
9·83 ± 0·46c
4·13 ± 0·24a
3·89 ± 0·38a
Hexose : amino acid (molar ratio)
0·97 ± 0·08a
1·24 ± 0·07b
1·16 ± 0·04ab
1·33 ± 0·05b
0·32 ± 0·02c
0·38 ± 0·02c
1·17 ± 0·09ab
1·70 ± 0·13d
Sucrose : amino acid (molar ratio)
1·32 ± 0·19a
3·00 ± 0·63b
6·75 ± 0·75c
6·16 ± 1·06cd
0·70 ± 0·12a
0·99 ± 0·11a
3·92 ± 0·46e
5·20 ± 0·73de
The ratio of free hexose to free amino acid has been suggested to represent the relation of ‘active pools’ of carbon and nitrogen ( Paul & Driscoll 1997). At low N, this ratio was high and was almost constant for all CO2 treatments and harvest dates ( Table 2). At high N it increased markedly from 9 to 23 DAD. At high N and 23 DAD, the hexose : amino acid ratio was strongly increased under elevated pCO2 (+ 45%). At high N similar trends were observed for the ratio of sucrose : amino acid ( Table 2) but at low N and 9 DAD there was a considerable CO2 effect, and the ratios increased during regrowth.
General properties of L. perenne SPS
A new assay method was developed making it possible to accurately measure SPS activity in L. perenne leaf extracts, as measurement of uridine-diposphate ( Copeland 1990) was inappropriate because of the high activity of uridine diphosphatase. Similarly, the high fructan background concentration made a colorimetric measurement of bound fructose ( Copeland 1990) too insensitive. However, separation and quantification of the assay product, sucrose, by HPLC was a reliable method of measuring SPS activity in L. perenne.
The main components in the SPS assay were optimized in order to be able to accurately quantify SPS activity in raw extracts of L. perenne (e.g. Fru-6-P concentration, Fig. 2a). The next step was to examine the short-term regulatory properties of SPS: inorganic phosphate in the assay (10 mol m−3) decreased the apparent activity by 30% at 3 mol m−3 Fru-6-P in extracts of light-adapted leaves ( Fig. 2a). Feeding mannose – as opposed to sorbitol – to leaves led to an almost complete activation of SPS even in the dark, probably by sequestration of cytosolic phosphate ( Fig. 2b). SPS activity in L. perenne leaves appeared to undergo a short-term regulation similar to that of the covalent modification of SPS in spinach ( Huber & Huber 1996), in which SPS at Vmax is constant throughout the day, but there is clearly an activation of SPS at Vlim due to light.
Indeed, under field conditions, the activation state of SPS exhibited considerable diurnal changes ( Fig. 3), whereas SPS at Vmax showed no changes during a 24 h period. Therefore, daily averages of SPS activity at Vmax are presented ( Fig. 4a). In all treatments, the activation state was high during the day ( Fig. 3); In particular, during early regrowth it decreased sharply at the end of the light period, whereas later in regrowth it started to decline in the afternoon. The activation state rose again before dawn, similar to that in Lolium temulentum ( Pollock & Housley 1985) and tomato leaves ( Jones & Ort 1997).
The amount of SPS protein in leaves was estimated with polyclonal antibodies raised against maize SPS ( Bruneau et al. 1991 ; Fig. 2c); to avoid proteolytic artifacts, the sample preparation of Reimholz et al. (1997) was used. Western blot analysis revealed two prominent bands in L. perenne leaves ( Fig. 2c) which were similar to the bands of SPS 1a, 1b and of SPS 2 of potato ( Reimholz et al. 1997 ) and which probably represent isoforms of the SPS enzyme. In this study we were interested in the total amount of SPS, and so we calculated the protein amount from the sum of both bands. In L. perenne leaves, the SPS protein content correlated closely with SPS activity at Vmax (r2 = 0·92; Fig. 5); hence, both Vmax and protein content represent total SPS content.
Response of SPS to treatments
In the present experiment, SPS at Vmax based on leaf area showed a very strong positive response to the N treatment (+ 39%, P = 0·023), but no overall response to elevated pCO2 (P = 0·918), whereas SPS decreased from 9 to 23 DAD (− 37%, P < 0·001; Fig. 4a). This decrease disappeared when SPS at Vmax was expressed on a leaf protein basis (− 3%, P = 0·388; Table 1). SPS per protein was slightly higher under elevated pCO2 (+ 19%, P = 0·191) and at high N (+ 10%, P = 0·359). On a leaf area basis, the same pattern as at Vmax was observed for SPS protein by Western blotting, but with higher variation ( Figs 4b, 5). At 23 DAD and high N, elevated pCO2 produced a significant reduction of SPS at Vmax both per leaf area and per protein. In contrast to Vmax, the activation state of SPS was generally increased at elevated pCO2 (P = 0·014), whereas high N led to a decrease of the activation state (P = 0·029, Fig. 3). For the single treatments, these effects were not significant, except at 23 DAD and at high N, when elevated pCO2 enhanced the activation state significantly during the light period compared with that at ambient pCO2.
Correlations similar to those of Paul & Driscoll (1997) were tested for SPS with several cell constituents of leaves ( Fig. 6a–f). Since we were interested in the relation of SPS with the carbon flux and, hence, the photosynthetic rate, SPS activity and constituents were expressed per leaf area. However, no correlation of SPS was observed – neither with the daily net CO2 uptake (r2 = 0·20, not shown), nor with the photosynthetic rate (r2 = 0·00 on area basis, Fig. 6a; r2 = 0·05 on a protein basis, not shown). SPS was weakly negatively correlated with the morning sucrose contents (r2 = 0·56, Fig. 6c), whereas the correlation disappeared when the daily average sucrose contents were used (r2 = 0·02, not shown). For the other leaf constituents and the ratios, morning and average values gave very similar correlations, thus only those using average values are given: SPS did not correlate with the free hexose content (r2 = 0·08, Fig. 6b), and it correlated positively with the free amino acid (r2 = 0·88, Fig. 6d) and the protein content (r2 = 0·77, not shown), negatively with the free hexose : free amino acid ratio (r2 = 0·75, Fig. 6e) and the sucrose : free amino acid ratio (r2 = 0·85, Fig. 6f).
In this work we present new evidence that in leaves of L. perenne, under field conditions, the SPS protein content responded to the carbon source-sink ratio. The latter appeared to be mediated by compounds close to those in primary carbon and nitrogen metabolism. The activation state of SPS changed in order to adjust the rate of sucrose synthesis to the rate of photosynthesis, even over the long-term.
Carbon-sink limitation is indicated by leaf constituents
In L. perenne leaves, the contents of carbohydrate, protein, and amino acid indicate that the pCO2 and N treatments, and the stages of regrowth, resulted in plants with a wide range of source-sink ratios. The high accumulation of carbohydrate late in regrowth, especially at low N and elevated pCO2, indicates a severe carbon source-sink imbalance ( Stitt 1991; Fischer et al. 1997 ). The accumulation of carbohydrate occurred mainly in the form of fructan, which, in all likelihood, is synthesized in leaves when carbon assimilation exceeds carbon use ( Pollock & Cairns 1991). Lower contents of protein and free amino acid in leaves, both late in regrowth as well as in the low N treatment, indicate a decreased N availability for the plant, thus restricting further growth. Indeed, carbon sink strength, as modulated by N supply, significantly affects the overall growth response of L. perenne to elevated pCO2 ( Schenk et al. 1996 ; Hebeisen et al. 1997 ).
Furthermore, the ratio of free sugars to free amino acid indicates the source-sink ratio or, in other words, the whole-plant carbon and nitrogen status ( Koch 1997; Paul & Driscoll 1997). Our experiment yielded plants exhibiting a wide range in their hexose : amino acid ratio or the sucrose : amino acid ratio – corresponding to a variety of source-sink ratios. The largest change in the hexose : amino acid ratio occurred at high N and elevated pCO2, thus confirming our expectation that plants subjected to that particular treatment would undergo the most severe changes in their source to sink ratios.
Sucrose-phosphate synthase content is strongly influenced by growth conditions
Lolium perenne accumulates sucrose and fructan rather than starch. During fructan synthesis, glucose units are re-cycled to sucrose, thus generating a flux through SPS in addition to that due to carbon assimilation. In our experiment, however, the flux resulting from daily or longer term fructan accumulation, respectively, was nil ( Isopp et al. 2000 ) or small (estimated from Fig. 1), in comparison with the daily carbon uptake ( Table 1). Thus, sucrose synthesis was expected to closely correlate with the photosynthetic rate in L. perenne.
However, the higher photosynthetic rates at elevated pCO2 were not accompanied by a corresponding increase in SPS activity at Vmax when compared on a leaf area basis, and only partially when compared on a protein basis ( Table 1 and Fig. 4a). At high N and 23 DAD, SPS was decreased under elevated pCO2 on both the leaf area and protein basis, whereas photosynthetic rate remained strongly increased. Elevated pCO2 consistently led to a higher activation state of SPS. This is a surprising result in the long term, since (i) higher assimilation rates usually result in an increased SPS content ( Klein et al. 1993 ; Seneweera et al. 1995 ; Hussain et al. 1999 ), and (ii) a higher activation state in association with increased assimilation rates has been described in short-term experiments only ( Battistelli, Adcock & Leegood 1991). The present study is the first to report that, even in the long term, the higher carbon flux at elevated pCO2 was associated mainly with a higher activation state rather than with an increased SPS content.
In contrast to elevated pCO2, high N fertilization had a strong positive effect on SPS content per leaf area, without affecting carbon assimilation. Analogous to the effect produced by elevated pCO2, apparent SPS activity was adjusted to the current assimilatory carbon flux by a corresponding reduction of the activation state. In rice, SPS at Vmax was found to increase together with photosynthesis at higher N ( Makino et al. 1994 ).
The decrease of SPS per leaf area during later regrowth ( Fig. 4a) was due to a general decrease in leaf protein, since it disappeared when SPS was expressed on a protein basis ( Table 1). However, SPS did not exactly follow the protein decrease under elevated pCO2, nor did it follow the protein pattern at low and high N (compare Tables 1 and 2). Therefore, SPS does not seem to be solely related to total leaf protein.
Both SPS and Rubisco of L. perenne responded in parallel to temporal changes of the carbon source-sink ratio in the same experimental system. However, they differed in their specific response pattern to N fertilization (compare Fig. 4a and Rogers et al. 1998 , Fig. 6); this suggests that different regulatory mechanisms control the level of each enzyme in response to the source-sink ratio. Since regulation of SPS content did not parallel the photosynthetic rate, the carbon flux through SPS was adjusted by changing the activation state accordingly (especially in treatment high N and at 23 DAD in Figs 3 and 4a). These results show that a high flexibility of key enzymes (regulatability, as used by Stitt & Sonnewald 1995) such as SPS is necessary to allow sufficient carbon flux even when the enzyme content is decreased relative to the apparent flux. For example, in tubers of transgenic potato, carbon flux into sucrose is not related to the genetically altered SPS content, but it is related to the activation state and metabolite contents ( Krause et al. 1998 ). Our study with SPS of L. perenne from the Swiss FACE is one of the first to report that fine regulation (via activation state) can compensate for large changes in enzyme content in field-grown plants.
Possible regulation of sucrose-phosphate synthase content
The positive correlations of leaf SPS content with the total protein and the free amino acid content emphasize the importance of nitrogen in the regulation of SPS content. Since the SPS content does not appear to simply reflect the changes in protein content nor to be directly linked to the photosynthetic rate, the regulation of SPS content appears to be quite intricate.
Indeed, the correlations of the SPS content with the hexose : amino acid ratio and the sucrose : amino acid ratio suggest an involvement of the C/N balance, probably a result of source-sink relation, in the regulation of SPS content. On the one hand, the relation between SPS and the sugar : amino acid ratio could be indirect, as in spinach where sugars and amino acid contents are strongly related to plant growth parameters ( Buysse, Smolders & Merckx 1993). On the other hand, the relation could be quite direct: In N-deficient tobacco, the sugar : amino acid ratio is related to the photosynthetic activity ( Paul & Driscoll 1997). Further, selected sugars and amino acids can directly affect gene expression ( Koch 1997; Oliveira & Coruzzi 1999). However, a direct effect of sugars alone on the expression of SPS appears unlikely because of the missing correlation. Further, in a short-term experiment, sugar feeding had no influence on the SPS transcript level in sugar beet leaves ( Lee & Daie 1997). High contents of sugar in source leaves do not necessarily lead to a feedback inhibition of gene expression ( Stitt et al. 1995 ), since higher sucrose concentrations might be needed, for example, to sustain increased carbon export rates at elevated pCO2 ( Farrar & Williams 1991; Isopp et al. 2000 ).
In conclusion, the following two hypotheses can explain the correlation of the sugar : amino acid ratio with L. perenne SPS: (1) the sugar : amino acid ratio simply reflects the source-sink relation, which controls SPS via a mechanism that is independent of the sugar and amino acid signals; alternatively (2) sugars and amino acids, while reflecting the source-sink relation in their ratio, directly influence the expression of the SPS gene(s).
The Swiss FACE experiment was supported by: The Swiss Federal Institute of Technology (ETH), Swiss National Science Foundation, The Federal Offices for Agriculture, for Energy, and for Science and Education (COST), and the Brookhaven National Laboratory. We would like to thank Calgene Inc. for kindly providing SPS antibodies, Jon Anderson and Paul Humphries for gas exchange measurements, José P. Almeida, Paola Curioni and Sandra Zangger for their help in sampling, and the team in Eschikon for maintaining the FACE facility.