N content of phloem and xylem exudates during the transition to flowering in Sinapis alba and Arabidopsis thaliana


Correspondence: L.Corbesier. Fax: +32 (0)4 366 38 31; e-mail: laurent.corbesier@ulg.ac.be


The involvement of nitrogenous substances in the transition to flowering was investigated in Sinapis alba and Arabidopsis thaliana (Columbia). Both species grown in short days (SD) are induced to flower by one long day (LD). In S. alba, the phloem sap (leaf and apical exudates) and the xylem sap (root exudate) were analysed in LD versus SD. In A. thaliana, only the leaf exudate could be analysed but an alternative system for inducing flowering without day-length extension was used: the displaced SD (DSD). Significant results are: (i) in both species, the leaf exudate was enriched in Gln during the inductive LD, at a time compatible with export of the floral stimulus; (ii) in S. alba, the root export of amino acids decreased in LD, whereas the nitrate remained unchanged – thus the extra-Gln found in the leaf exudate should originate from the leaves; (iii) extra-Gln was also found very early in the apical exudate of S. alba in LD, together with more Glu; (iv) in A. thaliana induced by one DSD, the leaf export of Asn increased sharply, instead of Gln in LD. This agrees with Asn prevalence in C-limited plants. The putative role of amino acids in the transition to flowering is discussed.




Since early in the twentieth century, the importance of nutrition in the control of the flowering process has been recognized (reviewed in Bernier, Kinet & Sachs 1981). In 1913, Klebs proposed that promotion of reproductive development requires a high ratio of carbohydrates to nitrogen (the so-called C : N ratio) within plant tissues. This theory was based on the observations that (a) conditions favouring photosynthetic CO2 fixation are generally promotive for flowering and flower development, and (b) a high N supply (fertilizer) may reduce reproductive development in some plants, perhaps because it enhances vegetative growth. Later, Sachs & Hackett (1977) postulated that, irrespective of the environmental factors inducing flowering (daylength, temperature, etc.), the basic process is a modification in source–sink relationships within the plant in such a way that the shoot apex receives a better supply of assimilates (mainly carbohydrates) than under non-inductive conditions. This was known as the ‘Nutrient diversion’ theory. Elaborating on these ideas, Raper et al. (1988) and Rideout, Raper & Miner (1992) hypothesized that floral transition is stimulated by an imbalance in the relative availability of carbohydrates and N in the shoot apex.

In photoperiodic plants, there is strong experimental evidence that leaves produce promoters and inhibitors of flowering when exposed to favourable and unfavourable conditions, respectively (Bernier et al. 1981). These signals are transported from the leaves to the apical meristem in the phloem with carbohydrates. Thus phloem is an obvious material to look at in any attempt to elucidate the nature of the floral signals. It is thus surprising that only few studies tackled the biochemical modifications of the phloem sap during the induction of flowering. Carbohydrates, which are by far the most abundant component, were analysed in several photoperiodic plants that can be synchronously induced to flower by exposure to a single long day (LD) or short day (SD) (Bernier et al. 1998). In all cases, an increase in sucrose export out of the leaves was observed during the transition to flowering. Other physiological signals that are good candidates for controlling the transition to flowering and received some attention are phytohormones such as cytokinins, gibberellins and polyamines (Bernier et al. 1993, 1998; Koornneef et al. 1998; Levy & Dean 1998).

Surprisingly, changes in the amino acid content of the phloem sap at floral transition have not been investigated so far, despite the fact that amino acids are the second prevalent compounds (behind carbohydrates) of this sap (Peoples & Gifford 1990; Lam et al. 1995). Scarce analyses mainly concerned extracts of organs, including reproductive structures. Maeng & Khudairi (1973) showed that several amino acids increase in the SD plant Lemna perpusilla and the LD plant Lemna gibba during floral transition. In addition, exogenous treatments with some amino acids may stimulate the flowering process in various Lemna species (Tanaka & Takimoto 1977; Kandeler 1985). In several other species, namely chicory, cabbage and tobacco, an increase in the content of amino acids, especially Pro, is observed at the time of floral transition, even when this transition occurs in the absence of a cold treatment (Shvedskaya & Kruzhilin 1966; Vallée, Perdizet & Martin 1968; Bouniols et al. 1973). In Populus x euramericana and Salix x smithiana, amino acids increase drastically in the xylem sap during growth of catkins (Sauter 1981; Babin & Bonnemain 1993). Recently, Valle, Boggio & Heldt (1998) observed an increase in the Gln content during early fruit development in tomato, which they suggested might be due to import via phloem sap.

Finally, transgenic plants of Lotus corniculatus overexpressing a cytosolic Gln synthetase gene from soybean are enriched in amino acids and flower earlier than untransformed plants (Vincent et al. 1997). Interestingly, vegetative growth is not stimulated in those plants and flowering is hastened without any increase in shoot and root dry weights.

The aim of the present study was to provide data concerning amino acid changes during the transition to flowering in Sinapis alba and Arabidopsis thaliana. In S. alba induced to flower by a single LD, both the xylem sap (root exudate) and the phloem sap collected from mature leaves (leaf exudate) and at the apical part of the shoot (apical exudate) were analysed. For technical reasons, only the leaf phloem sap could be collected in A. thaliana but, in that species, the effects of both the LD- and the displaced short day (DSD)-inductive systems (Corbesier et al. 1996) on the leaf export of amino acids were investigated. In the DSD system, floral transition occurs without extension of the photoperiod, thus without any increase in photosynthesis.


Growth conditions and photoperiodic treatments

Growth conditions for S. alba and A. thaliana ecotype Columbia were as described elsewhere (Lejeune, Kinet & Bernier 1988; Corbesier et al. 1996). Briefly, plants of S. alba were sown and grown on a mixture of perlite and vermiculite (1 : 1) in 8 cm pots and were watered alternatively with demineralized water and a complete Hoagland solution. Seeds of A. thaliana were first vernalized in the dark at 2 °C on wet filter paper for 6 weeks, then sown and grown on a mixture of leaf mould, clay and sand. Six plants were grown in a tray containing 1 L of substrate and were watered daily with tap water. All plants were grown in controlled cabinets. Light was provided by Very High Output fluorescent tubes (Sylvania, Zaventem, Belgium) at an irradiance of 150 and 48 μmol m−2 s−1 (PAR) for S. alba and A. thaliana, respectively. Temperature was 20 °C and relative humidity was about 80%.

After 65 or 56 d of culture in 8 h SD, respectively, plants of S. alba and A. thaliana were induced to flower by a single photoperiodic treatment, then returned to the SD regime. For both species, the single LD was 22 h long and consisted of the extension of the photoperiod without any change in light quality or quantity. In A. thaliana, the alternative DSD was used, which was an 8 h SD delayed by 10 h within a daily cycle.

Because of the different photoperiodic treatments to which the plants were submitted, time zero had to be clearly defined: ‘The start of the experiment’ as presented in the Results was fixed at the beginning of the light period of the SD and the LD. In DSD, plants were still in the dark at h 0 and the light period started at h 10. Dissection of shoot apices 2 weeks after the experiment showed that the 22 h LD induced 100% flowering in both S. alba and A. thaliana. In A. thaliana, the DSD induced 85% of the plants to flower. Control plants continuously kept in standard SD remained vegetative.

Collection of xylem sap

Root exudate of S. alba was collected under a mild vacuum during 4 h periods as described elsewhere (Lejeune et al. 1988). In summary, exudation was achieved by fitting a 5 mL disposable syringe with silicon rubber tubing on the cut stump of plants decapitated at the cotyledonary node. By pulling back the piston and maintaining it with a small rod, enough vacuum was applied to ensure exudation. The exudates from 10 plants were pooled and stored at −20 °C until analysis.

Collection of leaf phloem sap

Leaf exudates were collected using the EDTA-method (King & Zeevaart 1974) as previously described by Lejeune et al. (1988) for S. alba and Corbesier, Lejeune & Bernier (1998) for A. thaliana. Briefly, the uppermost five leaves below the half-expanded one of S. alba plants were placed together for 4 h in a 250 mL beaker containing 20 mL of 20 mM EDTA (pH 7·5). For A. thaliana, the seven youngest mature leaves were collected per plant and placed in a 500 μL microcentrifuge tube containing 400 μL of 10 mM EDTA (pH 8·5) for 8 h. Ten plants were harvested at each sampling time. The vessels containing the leaves were enclosed in airtight clear chambers containing water to ensure maximum relative humidity and to prevent EDTA uptake by the leaves. During exudation, the leaves were subjected to the same light : dark regime as intact plants. After collection, exudates were stored at −20 °C until analysis.

Collection of apical phloem sap

Exudates of S. alba plants were collected using the method previously described by Lejeune et al. (1993). Briefly, 10 plants per batch were detopped as close to the apex as possible and a 1·5 mL microcentrifuge tube containing 1·25 mL of 20 mM EDTA (pH 7·5) with 1% agarose (electrophoresis grade; Pharmacia, Uppsala, Sweden) was placed immediately on the cut stump. After 4 h, the tubes were removed from the plants and stored at −20 °C until analysis.

Extraction of apical exudates

The exudation media collected from 10 plants were pooled, homogenized and extracted during 1 h in 15 mL methanol, after addition of norvaline (9 × 10−5M, final) as internal standard. Agarose was centrifuged (19 000 g, 30 min, 4 °C) and re-extracted for 30 min with 15 mL methanol. After a second centrifugation, the supernatants were combined and reduced to 5 mL under vacuum at 30 °C to remove methanol. Apical exudates were then treated as root and leaf exudates.

Analysis of amino acids

Amino acids were analysed using a method adapted from Einarsson, Josefsson & Lagerkvist (1983). In leaf and root exudate, norvaline was first added to the samples as an internal standard (5 × 10−4M, final) and the exudates were filtered on Acrodisc LC13PVDF 0·2 μm (Vel, Leuven, Belgium). Using siliconed 1·5 mL microtubes, 100 μL of exudate was made up to 150 μL with 0·5 M borate buffer (pH 7·7). The derivatization was then initiated by the addition of 100 μL 3 × 10−3M 9-fluorenylmethoxycarbonyl chloride (FMOC) dissolved in acetone (Merck-Belgolabo, Overijse, Belgium). After 60 s, 200 μL of 4 × 10−2M 1-adamantamine (in 75% acetone) was added to quench the excess of FMOC, and 60 s later, 20 μL of the reaction mix was sampled and diluted 20-fold in ultrapure water for high-performance liquid chromatography (HPLC) analysis. Derivatization was performed using a Gilson 231XL automatic injector (Analis, Namur, Belgium) equipped with a 20 μL sample loop. A 10 μL aliquot of the diluted reaction mixture was analysed by reverse phase HPLC and fluorescence detection (excitation, 265 nm; emission, 313 nm). The equipment consisted in a pump (LKB2150; LKB, Bromma, Sweden), a Spherisorb C18 analytical column (100 × 2 mm i.d.; Achrom, Zulte-Machelen, Belgium) kept at 45 °C and a fluorescence detector (Merck-Hitachi L-7480; Merck-Belgolabo, Overijse, Belgium). The mobile phase was a mixture of methanol : acetic acid buffer (pH 6·3 adjusted with triethylamine) with an elution gradient from 30 : 70 (v/v) to 80 : 20 and a flux of 0·3 mL min−1 for 50 min. Amino acids were quantified using a calibration curve ranging from 2 to 35 pmol injected. Identification by retention times and quantification based on peak areas were performed using the Borwin integrator software (JMBS Développements, Le Fontanil, France). Recovery of the internal standard ranged from 90 to 95% for the leaf and root exudates and from 60 to 75% for the apical exudate.

The identity and relative abundance of the amino acids detected were confirmed using the same derivatization procedure and fluorescence detection in a different HPLC system composed of a Lichrospher 100 RP-18 analytical column (125 × 5 mm i.d.; Merck-Belgolabo) operating at 45 °C, and a mixture of acetonitrile : acetic acid buffer as the mobile phase. Amino acid separation was achieved using an elution gradient from 25 : 75 (v/v) to 50 : 50 with a flux of 1 mL min−1 which increased to 2 mL min−1 after 11·5 min (results not shown).

Nitrate quantification

Root exudates were purified by passing through an Adsorbex RP18 cartridge (Merck-Belgolabo). The effluent was then injected into an HPLC system to quantify NO3 using a calibration curve of NaNO3 (Merck-Belgolabo) ranging from 0·5 to 2 mg L−1. The equipment consisted of a pump (Merck 6200; Merck-Belgolabo), a Hamilton PRPX-100 column (150 × 4·1 mm i.d.; Vel) kept at room temperature, and an UV detector (Merck L4000; Merck-Belgolabo). A guard column (Hamilton PRPX-100, 25 × 4·1 mm i.d.; Vel, Leuven, Belgium) was introduced between the injector and the analytical column. Samples were injected using an automatic injector (Merck AS4000; Merck-Belgolabo) equipped with a 100 μL sample loop. The isocratic system was operated at 2 mL min−1 using a degassed mixture of 4 mM K hydrogen phtalate (Merck-Belgolabo) in 1 : 39 methanol : water (v/v) as the mobile phase. Determination of the retention times and quantification based on peak areas measured at 277·5 nm were performed using the Chrompack integrator software (Merck-Belgolabo).

Estimation of total N

The amounts of amino acids were converted to equivalent N atoms. N from nitrate was added for the root exudate.

Statistical analysis

Three independent experiments were conducted and the results shown are mean data. Analysis of the significance of the changes occurring in the export of amino acids and nitrate in the leaf, root and apical exudates was performed by the one-way analysis of variance (ANOVA) test using the SigmaStat statistical software (Jandel Scientific, Erkrath, Germany) with P≤ 0·05.


The five most abundant amino acids found in the phloem sap of S. alba and A. thaliana were Gln, Glu, Asn, Asp and Ser. Together, they accounted for about 80% of the total amino acids of the phloem sap. As observed for other species, other amino acids were also detected (Ala, Thr, Gly, Val, Met, results not shown) but in lower amounts (Weiner et al. 1991; Winter, Lohaus & Heldt 1992; Lam et al. 1995). In the xylem sap of S. alba, Gln was the major N-component as generally observed in non-leguminous plants (Peoples & Gifford 1990; Lam et al. 1995): Gln was at least 10-fold more abundant than any other amino acid detected in that sap.

As the amounts of leaf and apical exudates were very small and largely diluted in the exudation media, which had a fixed volume, the results presented are the net production per plant per h of exudation. Similarly, as the volumes of the root exudates collected in SD controls and LD-induced plants were not statistically different (around 7 and 8 mL per plant), results presented below are the net production per plant per h of exudation

Amino acids and nitrate in the xylem sap of S. alba

In SD, Gln, Asn, and Ser exhibited a diurnal fluctuation in the xylem sap, with higher levels in the light than in the dark, whereas Asp and Glu were quite constant (Fig. 1). Interestingly, the same three amino acids were significantly less exported when plants were induced to flower by one LD; their levels returned to normal by the following SD. Gln, which was by far the predominant amino acid of the root exudate (note the different scales in Fig. 1), exhibited a decrease of approximately 40% during the LD in comparison with the SD controls. Asp export remained unchanged during the LD but increased afterwards. Glu remained unchanged throughout the experiment.

Figure 1.

Gln, Glu, Asn, Asp, Ser and nitrate content of the root exudate of S. alba plants during an 8 h SD (●) and a 22 h LD (○). Exudates were collected during 4 h periods starting at various times after start of the experiment. Experimental points correspond to the middle of each period of exudation. * = statistically different from SD controls (P≤ 0·05). Error bars = 2 × standard deviation.

Nitrate was also analysed in the root exudate of S. alba and found to fluctuate only slightly in SD. In LD-induced plants, root export of nitrate remained unchanged during the LD but significantly increased afterwards, during the following SD.

This late increase probably accounts for the higher total N content estimated at that time (Table 1), whereas earlier changes observed in LD versus SD were not statistically significant when considered together.

Table 1.  Total N content in the root, leaf, and apical exudates collected at various time intervals in non-induced (SD) and induced (LD, DSD) plants of S. alba and A. thaliana
 Exudation periodsTotal N (nmol plant–1 h–1)
 (h from start of experiment)SDLDDSD
  • *

    statistically different from SD controls

S. alba
Root exudate 8–12474·7 ± 66·5437·9 ± 69·4
 12–16401·7 ± 62·6337·9 ± 96·1
 16–20323·3 ± 26·1281·6 ± 81·5
 20–24387·9 ± 79·2305·7 ± 53·7
 24–28507·6 ± 112·7498·9 ± 54·3
 28–32656·8 ± 26·4911·6 ± 21·5*
Leaf exudate 8–12288·1 ± 35·2318·2 ± 34·7
 12–16178·1 ± 19·8177·9 ± 11·3
 16–20306·8 ± 45·7451·7 ± 44·8*
 20–24231·6 ± 21·1473·6 ± 39·9*
 24–28225·7 ± 32·2535·7 ± 43·9*
 28–32196·3 ± 29·7457·8 ± 57·1*
Apical exudate 8–12133·3 ± 15·3133·7 ± 28·2
 12–16 96·8 ± 11·6184·0 ± 33·9*
 16–20 54·9 ± 5·0110·6 ± 14·6*
 20–24 69·3 ± 12·9100·1 ± 7·3*
 24–28168·7 ± 25·6140·3 ± 9·0
 28–32171·5 ± 28·7166·2 ± 35·6
A. thaliana
Leaf exudate 8–16 12·5 ± 1·9 14·3 ± 2·014·7 ± 3·0
 16–24 12·6 ± 2·0 17·4 ± 1·8*18·5 ± 2·9*
 24–32 13·2 ± 2·0 22·4 ± 2·7*19·8 ± 3·6*

Amino acids in the leaf phloem sap of S. alba

In SD, the export of Gln and Asn remained more or less constant during a 24 h cycle whereas the Glu, Asp and Ser export was somewhat higher around midnight than at any other experimental time (Fig. 2a). In LD, there was a marked increase in the export of Gln by the leaves: from the exudation period h 16–20, its concentration in the phloem sap was three to four times higher than at the same time in SD. Glu and Asn amounts were not significantly altered during the LD, but were significantly higher during the following SD, although to a much lower extent than Gln. On the contrary, Asp and Ser export decreased during the LD.

Figure 2.

Gln, Glu, Asn, Asp and Ser content of the (a) leaf and (b) apical exudates of S. alba plants during an 8 h SD (●) or a 22 h LD (○). Exudates were collected during 4 h periods starting at various times after start of the experiment. Experimental points correspond to the middle of each period of exudation. * = statistically different from SD controls (P≤ 0·05). Error bars = 2 × standard deviation.

Taken together, these changes gave a significant increase in total N in the leaf exudate of LD plants from h 16–20 (Table 1), with the same kinetics as Gln content.

Amino acids in the apical phloem sap of S. alba

In SD, all five amino acids analysed showed a diurnal fluctuation, with higher levels during the day than during the night, and Gln was by far the most abundant (Fig. 2b). In LD, Gln and Glu arrival at the apex was significantly enhanced during the extension of the photoperiod but Gln was the most abundant at any time. Glu seemed to precede Gln since its amount in the sap increased first. There was no increase in the amounts of Asn, Asp and Ser directed towards the apex during the LD. During the following SD, all amino acids were less abundant in LD-induced plants than in SD vegetative ones but returned to control levels at the end of that SD.

When total N was calculated, the transient enrichment of the apical exudate during the extension period of the LD was statistically significant (Table 1), but not the fluctuations observed afterwards.

Amino acids in the leaf phloem sap of A. thaliana

As the length of the exudation period required in A. thaliana to collect sufficient amount of exudate was 8 h versus 4 h in S. alba, the timing obtained in A. thaliana was less precise (Fig. 3).

Figure 3.

Gln, Glu, Asn, Asp and Ser content of the leaf exudate of A. thaliana plants during an 8 h SD (●), a 22 h LD (○), and an 8 h DSD (bsl00084). Exudates were collected during 8 h periods starting at various times after start of the experiment. Experimental points correspond to the middle of each period of exudation. * = statistically different from SD controls (P≤ 0·05). Error bars = 2 × standard deviation.

In SD controls, leaf export of the five most abundant amino acids remained quite constant during the entire 24 h cycle. In LD-induced plants, there was a significant increase in the export of Gln during the photoperiod extension and the following SD: in comparison with SD controls, its level in the phloem sap rose to 2- to 2·5- fold. Among the other four amino acids, only the amount of Asn showed a significant change in LD: a small decrease in the exudation period h 16–24.

In contrast, in DSD-induced plants, the export of Asn increased markedly, about 3·5- to 4-fold. Similarly to the Gln export in LD-induced plants, the increase in Asn level persisted during the following SD. Export of Glu, Asp and Ser remained unchanged whereas the Gln export was reduced.

In terms of total N, an enrichment was clearly found in both LD and DSD, from h 16–24 (Table 1).


In S. alba, the five most abundant amino acids in the xylem sap (root exudate) and in the phloem sap (leaf and apical exudates) were analysed and their levels were found to change during the inductive LD with: (1) a decrease in the amino acid export out of the roots; (2) an increase in Gln export out of the mature leaves; and (3) an increase in Gln and Glu arrival at the apical bud. In A. thaliana, the composition of the leaf exudate also changed during the transition to flowering, with an increased leaf export of Gln in LD and of Asn in DSD. These observations lead to four main conclusions and raise interesting questions:

First, in LD-induced plants, the Gln content of the leaf exudate increases in both species (Figs 2a & 3). Interestingly, this increase starts at the same time as the floral stimulus is translocated out of the leaves in S. alba or even before in A. thaliana. Indeed, the translocation of the floral stimulus – in fact the movement of its slowest component – out of the mature leaves was previously followed by defoliation experiment, allowing timings to be compared. In S. alba, the extra-Gln in the leaf exudate is detected from h 16–20 whereas defoliation experiments showed that export of the floral stimulus by induced leaves starts around h 16 after beginning of the LD (Bernier 1989). In A. thaliana, the stimulation in Gln export out of the leaves was observed at h 8–16, whereas the slowest component of the floral stimulus was previously shown to be translocated from h 20 (Corbesier et al. 1996). These observations suggest that Gln may play a promoting role in the transition to flowering. Several examples were reviewed in the Introduction where amino acids do stimulate flowering. Results of Vincent et al. (1997) are of special interest since they observed an early flowering phenotype in transgenic plants of L. corniculatus enriched in amino acids. The present study suggests that S. alba and A. thaliana could behave in the same way. In that respect, analysis of the floral behaviour of transgenic or mutant plants of A. thaliana affected in the biosynthesis of Gln would be informative.

Second, in S. alba, the extra Gln found in the leaf exudate during floral transition does not originate from an increase in nitrate reduction in the roots since Gln export in the xylem sap is decreased (Fig. 1) and the total N exported by the roots tends to decrease during the LD (Table 1). Thus, the extra Gln should be either synthesized, from stored precursors, or mobilized in the leaves themselves. However, it is unlikely to originate from a specific increase in leaf protein degradation since Kinet (1975) observed that the total leaf protein content was higher in LD-induced S. alba plants than in SD-controls.

Third, in the apical exudate of S. alba, Gln and Glu increase during the extension period of the LD (Fig. 2b). Despite the fact that this increase in Gln arrival was expected because of its increased export out of the leaves, this change surprisingly occurs earlier in the apical exudate than in the leaf exudate. A similar timing discrepancy was previously observed for sucrose and cytokinin movement by Lejeune et al. (1993, 1994) who measured earlier increases in cytokinin and sucrose export in the apical exudate than in the leaf exudate. These authors suggested that this could be related to the exudation techniques. Indeed apical exudates are collected on detopped plants with all leaves normally exposed to environmental conditions whereas leaf exudate is collected from isolated leaves packed and kept in 100% relative humidity. Thus changes observed in apical exudates may be more reliably timed in comparison with processes occurring in the intact plant, and it is crucial to keep in mind the experimental design when integrating results. In that respect, it must be emphasized that Gln and Glu are directed towards the apex well before any cytological changes and mitotic stimulation can be detected in the meristem of intact plants (Bernier et al. 1993), thus their increased export can hardly be a consequence of an increased sink activity of the apex.

Now, the fact that the apical exudate is poorer (2- to 3-fold) in Gln than the leaf exudate seems normal since amino acids exported by leaves are not only translocated towards the apex but also to other sinks, such as young leaves and roots (Scheible et al. 1997). Concerning Glu changes that are observed in the apical exudate and not in the leaf exudate, Thorpe & Minchin (1996) reviewed other cases where phloem sap is modified along its path, but the mechanisms involved are poorly understood. Interestingly, a gene that encodes an enzyme controlling the Glu-derived Pro biosynthesis has been recently identified as a target of the CONSTANS gene promoting flowering of A. thaliana in response to LD (Samach et al. 2000).

Fourth, when flowering is induced by one DSD in A. thaliana, i.e. without a lengthening of the photosynthetic period, the increased export of amino acids out of the leaves still occurs at the same time as in LD, but now consists of Asn instead of Gln (Fig. 3). This can be explained by the fact that the plants receive a protracted period of darkness before the DSD, and behave like dark-grown or dark-adapted A. thaliana plants (Lam et al. 1995). Indeed, because of the higher N : C ratio in Asn versus Gln, Asn is preferred for nitrogen transport and storage in this species when C supply is limited.

The DSD is regarded as the alternative system to induce flowering in LD plants that helps to discriminate between events related to photoperiod extension per se and floral induction (Périlleux, Bernier & Kinet 1994). Thus the increased export of one major amino acid – Gln or Asn – in both inductive systems strengthens the hypothesis that such compounds may be involved in floral transition. Moreover, neither Gln nor Asn increases during the light period in SD, thus the rise is not just a light effect.

Although our comparative analysis of S. alba and A. thaliana gave consistent results, whether the changes observed are specifically related to flowering cannot be asserted. As a general rule, physiological correlations gain significance by their recurrence in various systems, and this question could be further addressed by comparing more plant species, including SD plants. It is puzzling, however, that a promoting role of amino acids may seem contradicted by the fact that a high nitrate supply is known to inhibit flowering in many species, including S. alba (Bernier 1969) and A. thaliana (L. Corbesier, unpublished results). However, Gln or Asn increase in the phloem sap collected from leaves when other amino acids decrease, suggesting a fine regulation instead of a gross stimulation of amino acid export at floral transition. This fine regulation may be affected by mineral nutrition since nitrate and ammonium are known to control different steps of their own assimilation (Campbell 1999; Stitt 1999; Zhuo et al. 1999). Finally, the inhibition of flowering by high nitrate could be explained by some signalling role of this anion, independent of its nutritional function, as recently shown in the case of the root system morphology (Zhang et al. 1999).


L.C. is grateful to the F.R.I.A. and F.N.R.S. for the award of research fellowships. We wish to thank M.-C. Requier for nitrate analysis. This research was supported by grants from the European Community Contract BI04 CT97–2231; the Interuniversity Poles of Attraction Programme (Belgian State, Prime Minister's Office – Federal Office for Scientific, Technical and Cultural Affairs; P4/15) and the University of Liège (Fonds Spéciaux).