Nitrate is an important nitrogen source for plants, but also a signal molecule that controls various aspects of plant development. In the present study the role of nitrate on seed dormancy in Arabidopsis was investigated. The effects of either mutations affecting the Arabidopsis nitrate reductase genes or of different nitrate regimes of mother plants on the dormancy of the seeds produced were analysed. Altogether, data show that conditions favouring nitrate accumulation in mother plants and in seeds lead to a lower dormancy of seeds with little other morphological or biochemical differences. Analysis of germination during seed development indicated that nitrate does not prevent the onset of dormancy but rather its maintenance. The effect of an exogenous supply of nitrate on seed germination was tested: nitrate in contrast to glutamine or potassium chloride clearly stimulated the germination of dormant seeds. Data show, moreover, that the Arabidopsis dual affinity nitrate transporter NRT1.1 (CHL1) may be involved in conveying the nitrate signal into seeds. Thus, nitrate provided exogenously or by mother plants to the produced seeds, acts as a signal molecule favouring germination in Arabidopsis. This signalling may involve interaction with the abscisic acid or gibberellin pathway.
Seed germination starts with the uptake of water by the dry seed and ends with the elongation of the embryo axis which is seen at a macroscopic level as the protrusion of the radicle from the seed coat. Under certain circumstances, however, intact imbibed seeds are metabolically active but fail to complete germination even though their environment is favourable, a physiological block imposed by seed dormancy. Seed dormancy contributes to the adaptation of plants to their environment by optimizing the germination to the time period of the year, enabling for example the germination of annual plants in spring but not in autumn. Seed dormancy depends on many factors including mother plant growth conditions (McCullough & Shropshire 1970), seed storage conditions and environmental/chemical treatments imposed onto the seeds (Bewley 1997). In Arabidopsis seed coat contributes largely to dormancy: thus, seed coat removal and mutations affecting seed coat can alleviate seed dormancy (Debeaujon & Koornneef 2000; Debeaujon, Leon-Kloosterziel & Koornneef 2000).
Physiological, genetic and biochemical approaches have shown the major role of two hormones in determining dormancy/germination of seeds (Bentsink & Koornneef 2002). Abscisic acid (ABA) is a key hormone promoting dormancy (Finkelstein, Gampala & Rock 2002). Thus mutants deficient in the synthesis of this hormone and some ABA response mutants show altered dormancy (Finkelstein et al. 2002). Genetic analyses demonstrated that it is the ABA determined by the embryo genotype that is important in controlling seed dormancy (Karssen et al. 1983). Germination experiments using an ABA synthesis inhibitor and biochemical analyses show in addition that dormancy is an active process linked with de novo synthesis of ABA during seed imbibition in sunflower, barley and Nicotiana plumbaginifolia (Finkelstein et al. 2002). Gibberellins (GAs) on the other hand promote germination (Yamaguchi & Kamiya 2002). Thus, GA-deficient mutants fail to germinate (Ross, Murfet & Reid 1997). In Arabidopsis and tomato GAs have been proposed to overcome the mechanical constraints imposed by the tissues surrounding the embryo (Debeaujon & Koornneef 2000) and promote the growth potential of the embryo (Karssen & Laçka 1986). The importance of the ABA/GAs balance for germination is supported by the fact that reduced dormancy of some of the ABA mutants is linked with a lowered requirement for GAs for germination (Koornneef, Bentsink & Hilhorst 2002). Much effort has been devoted to understanding the signalling pathways of these two key hormones controlling dormancy/germination in Arabidopsis, more recently by transcriptomic or proteomic approaches for GAs (Gallardo et al. 2002; Ogawa et al. 2003; Yamauchi et al. 2004). The complexity of this physiological process is further underlined by the fact that other hormones such as brassinosteroids and ethylene also appear to affect seed dormancy/germination (Bentsink & Koornneef 2002), and that some of the Arabidopsis rdo mutants – isolated as displaying a reduced dormancy – do not appear to be affected in their sensitivity to ABA nor in ABA content or GA requirement (Leon-Kloosterziel et al. 1996b; Peeters et al. 2002).
In this study, we have chosen to analyse the effect of nitrate on the germination of Arabidopsis seeds. Nitrate is a major nitrogen source for many plant species. It is assimilated via its reduction by nitrate reductase (NR) and other enzymes leading ultimately to the production of amino acids and nitrogen compounds. In addition to its role as nutrient, nitrate was shown to act as a signal molecule, that independently of its assimilation controls numerous aspects of plant development and metabolism (Scheible et al. 1997a, b; Wang et al. 2003). Plants that are deficient in NR activity were instrumental in uncovering the signalling effects of nitrate since they accumulate nitrate and display an enhanced signalling by nitrate (Scheible et al. 1997a, b). Nitrate has for long been known to stimulate germination in a large number of plant species, but initial studies in Arabidopsis did not detect a clear effect of nitrate on germination (Hilhorst & Karssen 1988; Derkx & Karssen 1993). This could in part be due to the conditions under which the studied seeds were obtained (in the greenhouse). For example nitrate levels in Arabidopsis seeds were shown to vary enormously (about a factor 600, Derkx & Karssen 1993) and could account possibly for the variations in responses of different seed batches to exogenous nitrate, as was proposed in Chenopodium album (Saini, Bassi & Spencer 1985). In a closely related plant species, however, Sisymbrium officinale, nitrate was shown to promote germination, possibly by enhancing GA synthesis (Hilhorst & Karssen 1988) in close interaction with light. This effect was independent of nitrate reduction, suggesting possibly a signalling role for nitrate (Hilhorst & Karssen 1989). More recently, nitrate provided exogenously, was shown to promote germination in Arabidopsis Landsberg erecta ecotype by reducing light requirement of seeds (Batak et al. 2002) and in the Cape Verde Island ecotype by affecting ABA levels in imbibed seeds (Ali-Rachedi et al. 2004). We chose to analyse the effect of nitrate on germination of Arabidopsis seeds in a low dormant accession (Col0) by producing seeds under controlled conditions in growth chambers. We have developed an approach based on a controlled nitrate supply (low, standard or high nitrate) to mother plants and on the use of the NR-deficient mutants available in this ecotype of Arabidopsis. We tested the effect of N regime of mother plants and the impact of a deficiency in nitrate assimilation on the dormancy of the produced seeds. Effects of exogenously provided nitrate on the germination of freshly harvested wild-type or NR-deficient seeds were also studied as well as the involvement of nitrate transporters in sensing/transmitting this possible signal for germination.
MATERIALS AND METHODS
Plant material and growth conditions
For growth in controlled chambers seeds of wild type Arabidopsis thaliana (L) Heyhn, ecotype Columbia (Col0) and the G5, G′4-3 (Wilkinson & Crawford 1993), chl1-5 (Tsay et al. 1993), abi1-5 (Leon-Kloosterziel et al. 1996a) mutants that are in the Col0 background; as well as the chl1-10 mutant in a Wassilewskija (Ws) background (Munos et al. 2004) were stratified for 48 h at 4 °C in the dark in a 0.1% agar solution. Then approximately five seeds were sown in a small pot filled with homogeneous non-enriched compost and grown in a chamber (22 °C day/18 °C night and 80% relative humidity) under a light cycle of 16 h of light (100 µE m−2 s−1) and 8 h of dark. Six days after sowing, only one seedling per pot was retained. The pots were watered three times per week by immersion of the base of the pots in a solution containing either 3 or 10 m m nitrate (Loudet et al. 2003). One week after bolting, in some cases, the nitrate regime of plants was changed to 25 or 50 m m nitrate.
Seed germination experiments
For each experiment, all genotypes in the various growth regimes (3, 10 and 50 m m nitrate) were harvested on the same day. Freshly harvested seeds refers to seeds that were harvested and stored for 1 week at 4 °C before sowing. For time-course studies on developing seeds, only the primary shoot was kept and the flowers were tagged at the day of anthesis. Four independent batches of 50–80 immature or mature seeds were sown on 0.5% agarose plates (Litex agarose; FCM A/C, Vallensbaek Strand, Denmark) containing in addition in certain cases potassium nitrate, potassium chloride or glutamine. The plates were incubated in a growth chamber (Sanyo MLR-350; Avon, France) at 25 °C, 16 h light (100 µE m−2 s−1)/8 h dark, 70% relative humidity. Germination was scored as positive when the radicle protruded from the seed. To test the effect of ABA (Lomon Bio Technology Co., Ltd, Chengdu, China) or paclobutrazol (Sopra, Velizy, France), a cold treatment at 4 °C for 3 d was given to mature after-ripened seeds (2 months old) before incubation at 25 °C. After 7 d in the growth chamber, germination on paclobutrazol was scored as positive when both the radicle and the cotyledons emerged from the tegument.
Seed structural analyses
Developing seeds were cleared by a 1–24 h incubation in a chloralhydrate solution (8 : 2 : 1, w : v : v, respectively, of chloralhydrate, water and glycerol) then observed under a microscope as described by Boisson et al. (2001). Microscopic observation was performed with Nomarski differential interference contrast optics using a light microscope (Axioplan 2; Zeiss, Jena, Germany).
Mature seeds were fixed by vacuum infiltration and incubation overnight at 4 °C in a 0.1 m phosphate buffer pH 7.2 solution containing 5% glutaraldehyde (v/v) and 0.1% Triton. Subsequently seeds were dehydrated by successive incubation in solutions containing increasing ethanol concentrations (10 to 95%) and then embedded in Technovit 7100 resin (Heraus Kulzer, Wehrheim, Germany). Four-micrometre seed sections were obtained with a microtome (Leica RM 2055; Leica, Rueil-Malmaison, France), stained with toluidine blue (1% w/v) and observed under a light microscope.
The tetrazolium salts test (Debeaujon et al. 2000) was performed by incubating seeds in a solution of 1% 2,3,5-triphenyltetrazolium chloride (w/v) for 24 h in the dark at 30 °C.
Metabolite content of seeds
Fatty acids, sugar, starch and amino-acid contents of seeds were determined as described by Baud et al. (2002) using 15–40 seeds per extract. For the extraction of soluble compounds and starch, 20 seeds were homogenized in 500 µL 80% (v : v) ethanol at 4 °C and extracted for 60 min. In our case the extract obtained was also used for analysing nitrate content of seeds by high-performance liquid chromatography on a DX-120 analyser (Dionex, Sunnyvale, CA, USA).
GUS staining and activity measurements
Histochemical GUS staining was performed according to the method described by Jefferson (1987), with some modifications. Dry seeds were first sown on agarose plates for 1–48 h and the seed coat was removed before staining embryos. Siliques were opened to increase X-glucuronide penetration. Embryos, plantlets and siliques were vacuum infiltrated for 1 h in a 50 m m potassium phosphate buffer, pH 7.2 (1% triton, 0.5–5 m m ferro/ferricyanide and 2 m m Xglucuronide). Subsequently samples were incubated overnight in the dark at 37 °C. Stained embryos and plantlets were cleared by incubation in a chloralhydrate solution (8 : 2 : 1, w : v : v, respectively of chloralhydrate, water and glycerol) and then observed under a light microscope (Axioplan 2; Zeiss). Siliques were also cleared in a 70% ethanol before observation.
GUS activity measurements were performed according to Leydecker et al. (2000) with 5 µg protein and fluorescence was quantified using the Labsystems Fluoroskan II fluorimeter (Farnborough, UK).
In Arabidopsis, two genes (NIA1 and NIA2) encode the NR apoenzyme, and contribute differentially to total NR activity in the shoot (Wilkinson & Crawford 1993): a deletion mutant (G5) affected in the NIA2 gene displays about 10% NR activity, whereas the G′4-3 double mutant (nia1 nia2) retains only 0.5% wild type (WT) NR activity in the shoot. To study the effect of nitrate feeding of mother plants on the dormancyof the produced seeds, we grew the three Arabidopsis genotypes Col0 (WT), G5 and G′4-3 in controlled growth chambers under long days. The analysis of the behaviour of the G′4-3 mutant helps uncouple the signalling and the nutritional effects of nitrate since this mutant is affected by nitrate nutritional effects due to its severe impairment in nitrate assimilation, but it can still mediate nitrate signalling due to its propensity to accumulate nitrate (Zhang & Forde 1998). Plants were cultivated on non-fertilized peat and watered every other day with a nutrient solution containing either 3 m m nitrate (low nitrate nutrition) or 10 m m nitrate (standard nutrition) under a long day regime.
G′4-3 seeds are less dormant than G5 and WT seeds produced under standard or low nitrate nutrition
The development of the G5 mutants was very similar to that of WT plants whether the plants were grown under low or standard nitrate nutrition (compare Fig. 1a middle to Fig. 1a top). Indeed deficiency in NR activity affects plant growth under standard conditions only when NR activity falls below 10% of that of WT plants (Wilkinson et al. 1993). For both genotypes leaves were more developed on 10 m m nitrate (Fig. 1a right) than on 3 m m (Fig. 1a left), revealing the limitation of plant growth by the low nitrate supply. The G′4-3 plants were smaller than WT or G5 plants under either nitrate regimes (Fig. 1a bottom) as a consequence of their severe impairment in NR activity (Fig. 1b top). They accumulated nitrate (Fig. 1b bottom) in leaves and bolted about 2 weeks later than WT and G5 plants. Therefore, in order to perform germination analysis on seeds produced from the three different genotypes at the same time, G′4-3 seeds were systematically sown 2 weeks before WT and G5 plants.
We tested the capacity of 3-week-old WT, G5 and G′4-3 seeds to germinate. For this, flowers were marked right after pollination and the siliques produced were collected 3 weeks after. Seeds were sown in vitro on agarose and radicle protrusion was scored 7 d after sowing. Figure 2a shows that G′4-3 seeds germinate better than WT and G5 seeds regardless of whether the three seed genotypes were produced under 3 or 10 m m nitrate feeding of mother plants. This lower capacity to germinate of the WT and G5 seeds is not due to a lower seed viability since scoring germination of the seeds after stratification (3 d at 4 °C) led to 90–100% germination of all genotypes (data not shown). Thus 3-week-old WT and G5 seeds are more dormant than G′4-3 seeds produced under the same nitrate regime. In addition, the G′4-3 seeds themselves were less dormant when produced under 10 m m nitrate than under 3 m m feeding (Fig. 2a). This indicated a positive correlation between the nitrate dose given to the mother plant and the germination capacity; a higher dose of nitrate led to lower dormancy of G′4-3 seeds whereas a lower dose of nitrate (3 m m) resulted in the production of more dormant G′4-3 seeds.
The same germination test was performed on bulk-harvested seeds from the three different genotypes grown under low or standard nitrate nutrition: seeds were in this case all harvested when the plant siliques had yellowed and dried and sown directly afterwards on agarose for germination test. Again a higher percentage of germination was observed for the G′4-3 seeds obtained under standard nitrate nutrition than on low nitrate (data not shown); in all cases the percentage of G′4-3 seeds that germinated was greater than that of WT and G5 seeds that were dormant. Thus fresh bulk-harvested seeds behave like seeds collected 3 weeks post-pollination.
Nitrate dose during silique formation affects the dormancy of the G′4-3 seeds produced
A possible explanation for the difference in dormancy observed between the G′4-3 seeds produced on 3 and 10 m m was that the higher nitrate dose given to the mother plant led to lower seed dormancy. Alternatively, as plants were more developed under standard nutrition than under limiting nitrate (see Fig. 1a), the difference in dormancy could be related to developmental effects. We therefore tried to limit these differences in plant development by restricting the differences in N regime to the after-bolting stage of plants. We grew WT and G′4-3 plants on 3 m m nitrate until bolting then transferred half of the plants to 10 m m nitrate nutrition while keeping the remaining half on 3 m m nitrate nutrition. Germination was scored on bulk-harvested seeds from the different sets of plants. Figure 2 summarizes the data obtained by continuous feeding of mother plants with either 3 or 10 m m nitrate. Again the WT seeds were dormant whether mother plants were fed 10 or 3 m m nitrate during silique formation. G′4-3 seeds were also less dormant than WT seeds in all cases, and less dormant when produced on 10 m m than on 3 m m nitrate. These data stress the importance of nitrate nutrition during silique formation in determining seed dormancy and confirm the dose effect of nitrate that was demonstrated in the continuous feeding experiment (Fig. 2a).
Dormancy is linked to a maternal effect
Crosses were performed to analyse whether the dormancy status of the produced seeds was inherited maternally or not. WT plants and G′4-3 plants grown on 10 m m nitrate were crossed using either genotype as mother plant. F1 seeds resulting from the crosses were analysed for germination on agarose. As reported in Table 1, F1 seeds were dormant when the mother plant was WT, and non-dormant when the mother plant was G′4-3. Thus at a genetic level the observed dormancy is determined maternally.
Table 1. Germination analyses of F1 seeds resulting from crosses between wild-type and G′4-3 mutant plants grown under 10 m m nitrate nutrition. n indicates the number of F1 siliques analysed.
Germination percentage of the F1 seeds ± SE
Col0 female × Col0 male
1.6 ± 0.2 (n = 4)
Col0 female × G′4-3 male
2.0 ± 1.2 (n = 11)
G′4-3 female × Col0 male
78 ± 4.3 (n = 12)
G′4-3 female × G′4-3 male
81 ± 6.5 (n = 4)
The G′4-3 seeds and WT seeds produced are very similar morphologically and structurally
The lower dormancy observed for the G′4-3 seeds could result from differences in the development of the mother plants and of the produced seeds since the G′4-3 plants grew more slowly than WT plants. Alternatively structural differences in the seed affecting for example the testa (a maternal tissue) could affect seed dormancy which in Arabidopsis is essentially imposed by the coat (Debeaujon & Koornneef 2000; Debeaujon et al. 2000). Thus, another hypothesis to explain the effect of nitrate feeding of mother plants on seed dormancy would be that nitrate could affect the dormancy of the produced seed by indirectly controlling the seed coat composition.
Analysis of the development of the embryos of WT and G′4-3 plants during silique formation was performed by interference optics or by sectioning mature seeds followed by toluidine blue staining. Embryo development and the structure of mature seeds are very similar for both genotypes grown on 10 m m nitrate (Fig. 3a & b) and on 3 m m nitrate (data not shown). In particular the testa structure is normal in the G′4-3 mutant (Fig. 3b bottom). In addition, staining of the testa with tetrazolium salts was performed to analyse the integrity of the testa of the produced seeds. Tetrazolium salts reveal the permeability of the seed coat; thus Arabidopsis testa mutants display often altered staining with this dye (Debeaujon & Koornneef 2000; Debeaujon et al. 2000). Figure 3c shows that again WT and G′4-3 seeds behave very similarly, indicating that the G′4-3 seeds apparently are not affected structurally, at least as assessed by these methods. Other physical parameters of the seed including seed length, width and weight were measured in WT and G′4-3 seeds produced under 10 or 3 m m nitrate regime. Again, no significant difference was observed between the two genotypes for the tested parameters (data not shown). Thus the difference in dormancy between the two genotypes apparently was not linked to big morphological/structural changes in the produced seeds.
High nitrate feeding of WT plants leads to the production of less dormant seeds
Experiments were performed to test whether WT plants could be induced to produce more or less dormant seeds, depending on nitrate nutrition. As WT seeds produced on 3 and 10 m m nitrate were dormant in both cases, we sought to produce less dormant WT seeds by growing plants on intermediate (25 m m) or high nitrate doses (50 m m) during silique formation, following vegetative development on 10 m m nitrate. Figure 4a shows that in these conditions WT seeds produced under 50 m m nitrate are less dormant than those produced under 10 m m nitrate, and that 25 m m-produced WT seeds display intermediate dormancy. Thus even WT seeds can be induced to be less dormant by increasing the nitrate dose during silique formation. Tetrazolium salt staining of the high nitrate (50 m m)-produced WT seeds revealed as for the G′4-3 seeds no difference in staining (Fig. 3c) or in seed size/weight (data not shown) compared with the WT dormant 10 m m nitrate-produced seeds.
Low dormancy of WT and G′4-3 seeds is correlated with a higher nitrate content of seeds
Since nitrate nutrition of mother plants affected seed dormancy, we analysed the metabolite contents of seeds, to test whether nitrate could be accumulated differentially depending on the mother plant genotype and N nutrition. Nitrate, sugars, starch, amino acids and lipid contents were measured in seeds displaying different degrees of dormancy: dormant seeds (WT and G5 seeds produced under 10 m m nitrate), seeds with intermediate dormancy (G′4-3 seeds produced under 3 m m nitrate) and non-dormant seeds (G′4-3 seeds produced under 10 m m nitrate). No significant differences (P < 0.05) in sucrose and amino acids (Fig. 5a top and middle graphs) as well as starch and fatty acid contents (data not shown) were detected between the different seeds. Nitrate, however, was accumulated at higher levels in G′4-3 seeds than in G5 and WT seeds (Fig. 5a bottom graph). Furthermore seed nitrate content in the G′4-3 genotype reflected mother plant nutrition (Fig. 5a bottom): 10 m m-produced G′4-3 seeds accumulated more nitrate than 3 m m ones.
Nitrate content was also measured in WT seeds produced from plants cultivated on 10, 25 or 50 m m nitrate during silique formation and maturation. WT seeds produced on 50 m m nitrate displayed a higher nitrate content than those obtained on 10 m m whereas 25 m m-produced WT seeds contained intermediate levels of nitrate (Fig. 5b) stressing again a link between nitrate nutrition of mother plant, nitrate accumulation in seeds and lower seed dormancy. The absolute values of seed nitrate contents fluctuated widely depending on the culture experiment. For example in WT seeds obtained from plants cultivated on 10 m m nitrate, they varied from barely detectable levels in one experiment (Fig. 5a bottom) to 0.15 nmol seed−1 in another culture (Fig. 5b), but the G′4-3 seeds reproducibly displayed higher nitrate contents than the G5 and WT seeds from the same experiment (Fig. 5b).
Nitrate provided exogenously in the medium stimulates germination of dormant WT seeds but not of chl1-5 mutant seeds at low concentrations
The data above suggested that nitrate provided to the mother plant alleviated seed dormancy, possibly through its accumulation endogenously in seeds. Experiments were performed to test whether nitrate provided exogenously in the germination medium could affect the germination of the seeds. To this end, freshly harvested WT Col0 seeds (dormant) obtained from plants grown on 10 m m nitrate were tested for their capacity to germinate in the presence or absence of 10 m m nitrate. As shown in Fig. 6a, nitrate but not glutamine (5 m m) nor KCl (10 m m) clearly stimulated germination of dormant WT seeds. A more detailed analysis of the nitrate effect was performed by sowing WT, G5 and G′4-3 seeds obtained from 10 m m nitrate-fed mother plants on agarose with different concentrations of nitrate (ranging from 0 to 10 m m). The data (Fig. 6b) indicate that nitrate stimulates the germination of WT and G5 seeds in a dose-dependent manner: higher percentages of germination were observed for higher external nitrate concentrations. For the G′4-3 seeds a small stimulatory effect of nitrate on germination was observed (compare germination on agarose to that in the presence of 0.1 m m nitrate that shifts from 80 to 100% germination, Fig. 6b). However, due to the low dormancy of the latter seeds, no further increase in the germination percentage was observed in the presence of higher external nitrate concentrations.
Since nitrate-stimulated germination in a dose-dependent manner, we wondered whether mutants affected in nitrate transport would be affected in nitrate stimulation of germination. Nitrate transport at low concentrations (< 1 m m) involves high-affinity nitrate transport systems in roots that include (depending on the physiological conditions under which plants are grown) the dual high/low affinity NRT1.1 (CHL1) nitrate transporter (Liu & Tsay 2003) and the high-affinity NRT2.1 transporter (Filleur et al. 2001). WT Col0 and chl1-5 plants affected in the NRT1.1 gene were grown in controlled chambers with 10 m m nitrate and germination of the produced seeds was assessed in the presence of different concentrations of nitrate. The germination of the chl1-5 seeds, although equally stimulated by high nitrate doses (10 m m) failed to respond to 1 m m nitrate (Fig. 6c), indicating that the deletion of CHL1 in the mutant affected nitrate stimulation of germination at low nitrate concentrations. This phenotype of the chl1-5 seeds was confirmed using another mutant allele (chl1-10, Munos et al. 2004) in a Wassilewskija (Ws) background (Fig. 6d).
Expression of NRT1 during imbibition and seed development was studied using transgenic plants containing an NRT1::GUS fusion (Guo et al. 2001). Figure 7a shows that NRT1 is expressed both during seed imbibition (after 24 h imbibition) and – although to a variable degree – during seed development since GUS activity is detected at these stages by histochemical staining. The presence of GUS activity in siliques was not detected in the initial histochemical staining analyses (Guo et al. 2001). This discrepancy with our data is most likely due to the lack of permeability of the siliques to the GUS substrate since we had to open the siliques prior to infiltration with the GUS substrate to visualize GUS staining in the developing seeds. GUS activity measurements were performed to confirm the histochemical staining. Although the absolute levels of GUS activity were variable among independent plants, the same expression profile was observed (Fig. 7b): GUS activity increased during seed development and decreased at seed maturity, but was still well expressed in dry seeds and after 24 h imbibition.
The same germination experiments were performed with mutant seeds affected in the NRT2.1 gene. No differences in the nitrate dose–response curve of germination was detected between the latter mutant seeds compared with WT, nor was the NRT2.1 gene expressed early during seed imbibition or in developing siliques (data not shown).
Thus, nitrate provided exogenously to the dormant seeds stimulates germination, and this signalling may involve the NRT1.1 but not the NRT2.1 transporter when low concentrations of nitrate are present.
G′4-3 seeds and high nitrate-produced WT seeds require less GAs for germination
Since ABA and GAs play an important role in controlling seed dormancy and germination, experiments were performed to analyse whether nitrate feeding of mother plants affected the germination of the produced seeds by interacting with the ABA or GA pathways.
G′4-3 seeds and seeds developing on high nitrate-fed WT plants experience partial dormancy during their development, but this dormancy is rapidly alleviated at seed maturity
During seed development on the mother plant, dormancy develops in WT seeds after a temporary period of germinability of the isolated seeds and is linked to the presence of embryonic ABA (Karssen et al. 1983). In ABA-deficient mutants, dormancy fails to install and mutant seeds acquire progressively full germination capacity during their development (Karssen et al. 1983). Germination tests were performed on isolated immature seeds during their development (from 8 to 21 d after pollination, DAP) on mother plants fed with 10 m m nitrate. As shown in Fig. 8a, WT and G5 seeds behaved similarly: they germinated at the earlier stages of seed development (8–14 d after pollination) then progressively failed to germinate as dormancy was set up (18–22 DAP). G′4-3 seeds in general experienced also the onset of dormancy as shown by the fall in germination percentage that was observed in independent experiments 18–19 DAP; some G′4-3 seed batches, however, failed to display this onset of dormancy. This explains the high variability of the germination percentage at this stage (Fig. 8a). The absolute germination percentage of G′4-3 developing seeds was always higher than that observed for the G5 and WT immature seeds, and at 22 DAP the mature G′4-3 seeds had lost almost all their dormancy (about 90% germination). Similar to the G′4-3 seeds, WT seeds during their development on high nitrate-fed mother plants (50 m m) displayed a higher tendency to germinate, but dormancy was set up reproducibly in independent experiments at the same time as WT seeds from 10 m m nitrate grown plants and quickly lost afterwards (Fig. 8b). Thus nitrate did not affect the onset of dormancy but rather its maintenance in mature seeds.
Altered resistance of G′4-3 and high nitrate-produced WT seeds to paclobutrazol but not to ABA
Seeds from WT plants grown on 10 or 50 m m or from 10 m m nitrate-grown G′4-3 plants were tested for their capacity to germinate in the presence of different concentrations of ABA or in the presence of the gibberellin synthesis inhibitor paclobutrazol, after 3 d stratification to break dormancy. No difference in germination sensitivity to the inhibitory effects of ABA was detected between the different seeds, showing that nitrate effect on seed dormancy probably did not involve a change in sensitivity to ABA (Fig. 9a). G′4-3 seeds, however, were more resistant to paclobutrazol than the 10 m m nitrate-produced WT seeds (Fig. 9b), indicating a lower requirement for GA for germination for the former seeds. Their resistance to paclobutrazol, however, was lower than that of the aba1-5 mutant at the higher doses of paclobutrazol (Fig. 9b). High nitrate-produced WT seeds displayed also a slight resistance to paclobutrazol, although this resistance was lower than that observed for the G′4-3 seeds (Fig. 9b). The absolute level of resistance to paclobutrazol of the G′4-3 and high nitrate-produced seeds varied from one culture to another, but the former seeds were always more resistant to paclobutrazol than the latter.
The degree of seed dormancy can also be assessed by the degree of dark germination. Thus, a very low degree of dormancy is linked with a high dark germination. For example, freshly harvested seeds from the very low dormant aba2-1 abi3-1 and aba3-1 abi3-1 double-mutant plants display high dark germination compared with the single mutants or WT seeds (Leon-Kloosterziel et al. 1996a). We therefore assessed the dark germination of Col0 seeds obtained under standard or high-nitrate nutrition as well as that of G′4-3 seeds from standard nutrition: all seed lots still required light for germination (data not shown), showing that the low degree of dormancy of WT 50 m m nitrate-produced seeds and G′4-3 10 m m-produced seeds was not sufficient so as to affect their light requirement for germination.
An effect of nitrate on seed dormancy
In this study we have shown that nitrate affects seed dormancy in several ways. First, the nitrate regime of the mother plants had an impact on the dormancy of the seeds produced: high nitrate (50 m m) feeding of WT mother plants resulted in produced seeds that were less dormant than those obtained under normal nitrate nutrition (10 m m); the latter seeds, however, were less dormant than seeds from N-limited mother plants (3 m m nitrate) as shown by the germination analysis in the G′4-3 mutant background. This first effect of nitrate operated during plant bolting and seed development since only the nitrate nutrition of the mother plant after bolting was important in determining the resulting seed dormancy. This is in accordance with the fact that nitrate is taken up by Arabidopsis plants during the whole course of the plant development including during flowering and seed set, and that seed N comes essentially from N taken up and assimilated by the plant rather than from mobilization of N from source organs except when plants are severely limited in N (Schulze et al. 1994). Thus higher nitrate feeding of mother plants was correlated with a lower dormancy of the produced seeds. This is, to our knowledge, the first time mother plant N regime has been shown to affect seed dormancy. Second, a severe impairment in nitrate assimilation (in the G′4-3 mutant) resulted in nitrate accumulation in plants and a lower dormancy of the produced seeds. Thus, although G′4-3 plants displayed symptoms of nutritional limitation (slower growth and development, lower seed set) their produced seeds resemble seeds from WT plants cultivated under high nitrate and were even less dormant than the latter seeds. This suggested that nitrate feeding of mother plants decreased the dormancy of the produced seeds by a signalling effect and not a nutritional effect of nitrate. In favour of this hypothesis, the nitrate content of seeds was higher in the G′4-3 seeds and in the WT seeds obtained under high nitrate nutrition than in WT seeds obtained under standard nitrate feeding. In contrast to nitrate, seed contents of other metabolites such as sugars, starch, lipids, and amino acids did not vary significantly between the different genotypes or N regimes. Our data thus sustain the hypothesis that nitrate accumulated in the seed affects seed dormancy as proposed by others (Saini et al. 1985; Hilhorst 1990). The third way nitrate affected seed dormancy was when it was provided exogenously to WT dormant seeds and alleviated dormancy in a dose-dependent manner. This effect of nitrate was not nutritional since glutamine – another N source – was ineffective in breaking seed dormancy. Although we did not measure whether glutamine was effectively transported into the seeds, Garciarrubio, Legaria & Covarrubias (1997) showed that amino acids relieved the germination block imposed by abscisic acid, suggesting that amino acids could indeed enter into the seeds. Furthermore ammonium, a smaller molecule than glutamine, did not lead to higher germination percentages of seeds than water (A. Alboresi, unpublished experiments) sustaining the hypothesis that nitrate acts as a signal. Whether the effect of nitrate feeding of mother plants on the dormancy of the produced seeds reflects the same phenomenon observed when providing mature seeds with nitrate exogenously remains to be proven. In favour of a common mechanism between the two effects of nitrate, both processes involve a signalling and not a nutritional regulation by nitrate, as discussed above. In addition the NRT1.1 gene possibly involved in the response of dormant seeds to low concentrations of exogenous nitrate is expressed during seed development and early after imbibition. This suggests that this gene could be involved in taking up nitrate both in imbibed mature seeds and during seed development on mother plants and lead thus to alleviated dormancy. In Sisymbrium the existence of a nitrate receptor occurring in two conformations, a high- and a low-affinity state, was hypothesized after a detailed analysis of the germination response of seeds to different exogenous nitrate doses (Hilhorst 1990). The CHL1 dual affinity nitrate transporter could thus be a candidate for this receptor gene and set up thus the molecular basis for the analysis of nitrate stimulation of germination in Arabidopsis. Interestingly, although initially characterized as a root-specific nitrate transporter, NRT1.1 was shown subsequently to have a broader role in plants that could include signalling by low levels of nitrate, pH homeostasis during lateral root elongation (Guo et al. 2001) and regulation of the NRT2.1 gene expression (Munos et al. 2004). It is possible that in the case of germination also, NRT1.1 contributes similarly to signalling by nitrate. No effect of a mutation in the NRT2.1 gene on the response of seed germination to nitrate was detected, as expected from the absence of expression of this gene early during seed imbibition. However, both NRT1.1 and NRT2.1 belong to multigenic families in Arabidopsis (Crawford & Forde 2002). It would be of interest to analyse the expression of other members of these families in siliques and during imbibition so as to assess their potential involvement in nitrate signalling in seeds, in particular at higher exogenous nitrate concentrations. Indeed the normal response to high exogenous concentrations of nitrate of the chl1 seeds could be due to the participation of other nitrate transporters to nitrate uptake into seeds at these concentrations.
How nitrate affects seed dormancy
High nitrate feeding of mother plants or a severe impairment of nitrate assimilation in mother plants led to the production of seeds that accumulated nitrate and displayed lower dormancy. Whether nitrate in seeds alleviates dormancy directly or through metabolic changes in the seeds remains to be assessed more thoroughly by analysing more closely the seed contents of other metabolites than sucrose, starch, amino acids or fatty acids. Other nitrogenous compounds such as nitric oxide (Batak et al. 2002; Bethke et al. 2004) or nitrite (A. Alboresi, unpublished experiments) stimulate germination in Arabidopsis. It would be interesting to analyse whether nitrate acts per se on seed germination or through the production of N-related signals.
Our results support the hypothesis that nitrate accumulation in seeds is correlated with a lower requirement of GAs for germination. As shown in Fig. 9b the G′4-3 seeds that accumulate nitrate were clearly resistant and high nitrate-produced WT seeds slightly resistant to paclobutrazol. This lower resistance to paclobutrazol of the latter seeds may be due to the fact that the dormancy phenotype of the latter seeds is less clear-cut than that of the G′4-3 mutant. Indeed high nitrate-produced WT seeds always experienced the onset of dormancy during their development on mother plants (Fig. 8b) whereas sometimes immature G′4-3 seeds did not become dormant (Fig. 8a). Similarly G′4-3 seeds accumulated more nitrate than high nitrate-produced WT seeds. The use of another GA inhibitor such as tetcyclasis or of a GA-deficient background (Peeters et al. 2002) could help assess more finely the altered GA requirement for germination of the G′4-3 and high nitrate-produced WT seeds.
Nitrate could affect the seed requirement of GAs for germination in several ways. The seed coat structure, as assessed by staining with tetrazolium salts or sections was not obviously affected in nitrate-accumulating seeds, making it unlikely that seed coat alterations as in seed coat mutants (Debeaujon et al. 2000) were causing the reduced GA requirement. Nitrate has been proposed to stimulate germination by acting as an osmoticum and thus enhance water uptake in dormant caryopses of Avena fatua (McIntyre, Cessna & Hsiao 1996; McIntyre 1997). This effect of nitrate could act in parallel to the GA pathway that was postulated also to enhance the embryo growth potential (Karssen & Laçka 1986). Our data show, however, that KCl was far less effective in stimulating germination than KNO3, suggesting that the nitrate effect was not purely osmotic. Furthermore the experiments in oat involved the use of very high external nitrate concentrations (100 m m, McIntyre et al. 1996), so their physiological significance may be questioned.
Alternatively the reduced seed GA requirement for germination observed in nitrate-accumulating seeds could be due to altered GA or ABA synthesis/signalling. Immature high nitrate-produced WT seeds and G′4-3 seeds showed a higher propensity to germinate than WT seeds produced under 10 m m nitrate, although dormancy was set up in most cases. This germination profile of immature seeds resembles strongly that of immature WT seeds sown in the presence of GAs (Karssen & Laçka 1986), suggesting that indeed nitrate could signal germination in Arabidopsis by enhancing GA synthesis or signalling as hypothesized in S. officinale (Hilhorst & Karssen 1988). Alternatively, although nitrate did not seem to alter ABA sensitivity of seeds (Fig. 9a), it could change seed ABA levels. Indeed, in the recently published work on the Arabidopsis dormant Cvi ecotype (Ali-Rachedi et al. 2004), nitrate was shown to accelerate the decrease in ABA levels early during seed imbibition and to prevent the de novo synthesis of ABA that operates subsequently in dormant seeds (Ali-Rachedi et al. 2004). It is noteworthy that other processes linked to nitrate signalling or assimilation are connected with the ABA pathway. For example NR through its NO producing activity was reported to be necessary for ABA-mediated stomatal closure (Desikan et al. 2002), and nitrate control of root architecture (Zhang et al. 1999) was shown similarly to interact with ABA (Signora et al. 2001).
A more thorough understanding of nitrate signalling on seed dormancy and its interaction with ABA or GA pathways will be attained by transcriptomic or proteomic approaches on seeds produced under different nitrate feedings or imbibed with or without nitrate.
Physiological importance of nitrate control of seed dormancy
In conclusion, we have shown that nitrate is another environmental factor that in addition to light and temperature (McCullough & Shropshire 1970; Hayes & Klein 1974; Munir et al. 2001) controls Arabidopsis seed dormancy. The stimulatory effect of nitrate on germination was observed in the Arabidopsis low-dormant ecotype Col0, but also in Landsberg erecta ecotype (Batak et al. 2002), in the more dormant Ws (Fig. 8) and highly dormant Cvi ecotypes (Ali-Rachedi et al. 2004). This effect is mediated by a signalling effect of nitrate in Arabidopsis and probably other plant species such as S. officinale (Hilhorst & Karssen 1989). In their natural environment, plants do not usually have to cope with high concentrations of nitrate such as 50 m m. Yet it is possible that in our drastic conditions we uncovered a factor (nitrate) regulating dormancy that operates under physiological conditions but whose effect is usually masked by the numerous other factors affecting dormancy (Hayes & Klein 1974; Van der Schaar et al. 1997; Alonso-Blanco et al. 2003). Indeed, although nitrate alleviates dormancy, it is by no way in general the most effective way for breaking dormancy: stratification, for example, was shown to be the most effective treatment reducing dormancy in high-dormancy nearly isogenic lines (Alonso-Blanco et al. 2003) or in dormant low nitrate seeds (our data), showing that other factors in addition to nitrate do control seed dormancy. In their natural environment, plants experience changes in N content of soil due for example to nitrate leaching after rain. It is possible that a limiting N environment sensed by mother plant leads to an enhanced dormancy of the seeds to prevent germination under unfavourable conditions. When better environmental conditions arise (for example higher N in soil), dormant seeds would then be stimulated to germinate. This signalling would thus participate in adapting plant development to its environment.
This work was in part funded by the EU fifth framework programme PLUSN (http://www.plusn.org). We thank A. Marion-Poll, I. Debeaujon and B. Sotta for helpful discussions, A. Marion-Poll for critical reading of the manuscript. The help of M. Miquel, S. Boutet, B. Gaudin with, respectively, fatty acids, amino-acid, and seed section analyses is gratefully acknowledged as well as the gift of the transgenic NRT1::GUS seeds by N. Crawford (San Diego, USA) and the chl1-10 and NRT2::GUS seeds by M. Lepetit and B. Touraine, respectively (Montpellier, France). We thank J. Talbotec and F. Gosse for care of plants in the greenhouse.