AGPase, ADP glucose pyrophosphorylase
GS, glutamine synthetase
GOGAT, glutamate : oxoglutarate amino transferase
NADP-ICDH, NADP-dependent isocitrate dehydrogenase
NR, nitrate reductase
OPPP, oxidative pentose phosphate pathway
PEPCase, phosphoenolpyruvate carboxylase
Rubisco, ribulose-1,5-bisphosphate carboxylase/oxygenase
SPS, sucrose phosphate-synthase
This review first summarizes the numerous studies that have described the interaction between the nitrogen supply and the response of photosynthesis, metabolism and growth to elevated [CO2]. The initial stimulation of photosynthesis in elevated [CO2] is often followed by a decline of photosynthesis, that is typically accompanied by a decrease of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco), an accumulation of carbohydrate especially starch, and a decrease of the nitrogen concentration in the plant. These changes are particularly marked when the nitrogen supply is low, whereas when the nitrogen supply is adequate there is no acclimation of photosynthesis, no major decrease in the internal concentration of nitrogen or the levels of nitrogen metabolites, and growth is stimulated markedly. Second, emerging evidence is discussed that signals derived from nitrate and nitrogen metabolites such as glutamine act to regulate the expression of genes involved in nitrate and ammonium uptake and assimilation, organic acid synthesis and starch accumulation, to modulate the sugar-mediated repression of the expression of genes involved in photosynthesis, and to modulate whole plant events including shoot–root allocation, root architecture and flowering. Third, increased rates of growth in elevated [CO2] will require higher rates of inorganic nitrogen uptake and assimilation. Recent evidence is discussed that an increased supply of sugars can increase the rates of nitrate and ammonium uptake and assimilation, the synthesis of organic acid acceptors, and the synthesis of amino acids. Fourth, interpretation of experiments in elevated [CO2] requires that the nitrogen status of the plants is monitored. The suitability of different criteria to assess the plant nitrogen status is critically discussed. Finally the review returns to experiments with elevated [CO2] and discusses the following topics: is, and if so how, are nitrate and ammonium uptake and metabolism stimulated in elevated [CO2], and does the result depend on the nitrogen supply? Is acclimation of photosynthesis the result of sugar-mediated repression of gene expression, end-product feedback of photosynthesis, nitrogen-induced senescence, or ontogenetic drift? Is the accumulation of starch a passive response to increased carbohydrate formation, or is it triggered by changes in the nutrient status? How do changes in sugar production and inorganic nitrogen assimilation interact in different conditions and at different stages of the life history to determine the response of whole plant growth and allocation to elevated [CO2]?
Elevated [CO2] leads to increased rates of carboxylation and decreased rates of oxygenation of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) in C3 plants, resulting in a higher net rate of photosynthesis and increased synthesis of carbohydrate (Stitt 1991; Bowes 1996; Drake, Gonzàlez-Meler & Long 1997). Viewed in a broader context, however, carbon dioxide is just one of many inorganic substrates that are required by plants, and the long-term response of photosynthesis and growth to elevated [CO2] will depend, for two basic reasons, on the availability of mineral nutrients and the way in which they are utilized by the plant.
Firstly, the higher rates of growth in elevated [CO2] will lead to an increased demand for mineral nutrients. This could be met by using nutrients more efficiently and/or by increasing the rate at which minerals are absorbed and assimilated by the plant. To understand the response of plant metabolism and growth to elevated [CO2] we will therefore need to consider how mineral nutrients are acquired and used in the plant. It can be expected that the response will vary, depending on the species involved, the level of nutrition the plant is receiving, and the mineral in question.
Secondly, the acceleration of growth and increase of biomass in elevated [CO2] may change the nutrient status in the plant. This is a very real possibility, unless the nutrient supply is super-optimal or is increased in parallel with the rate of plant growth. Although changes in the nutritional status will not invalidate field studies whose aim is to describe the response of natural populations to elevated [CO2], it will be important to recognize if and when the nutrient state is being modified in order to interpret the results and relate them to studies at other locations and on other species. Changes in the nutritional status will have a serious consequence for studies in controlled conditions which are analysing the physiological and molecular mechanisms whereby elevated [CO2] affects plant metabolism and growth, because failure to recognize a change in the nutritional status could lead to misinterpretation of the results.
The following review will concentrate on the interaction between elevated [CO2] and nitrogen nutrition, because this interaction has been more intensively investigated than the interaction with other nutrients, and because the background knowledge about nitrogen uptake, assimilation and utilization is more extensive than for other nutrients. We (i) summarize the results from numerous descriptive studies that have investigated the interaction between elevated [CO2] and nitrogen nutrition. We then discuss the possible physiological and molecular background to this interaction, in particular (ii) how metabolism, allocation and growth are regulated in response to changes in the nitrogen status of the plant and (iii) whether mechanisms are known that could stimulate the uptake and assimilation of inorganic nitrogen or increase the nitrogen use efficiency in response to the conditions pertaining in elevated [CO2], especially an increased rate of carbohydrate formation. We then consider (iv) what is the best way to monitor the nitrogen status of a plant before (v) returning to the studies of plant metabolism and growth in elevated [CO2] and considering how they can be interpreted in terms of our background knowledge about the interactions between carbon and nitrogen metabolism and signalling processes, and to what extent the interpretation may be complicated or falsified by changes in the nitrogen status.
2 THE PHENOMENOLOGY OF THE NITROGEN-ELEVATED CARBON DIOXIDE INTERACTION
2.1 Acclimation of photosynthesis is more marked in nitrogen-limited plants
Although the initial stimulation of net photosynthesis is sometimes retained during long-term exposure to elevated [CO2] (Idso & Kimball 1992; Gunderson, Norby & Wullschleger 1993; Teskey 1995), it is often partially reversed or abolished (Stitt 1991; Oechel et al. 1993; Sage 1994; Nie et al. 1995a; Jacob, Greitner & Drake 1995; Tissue, Thomas & Strain 1997; Drake et al. 1997) in a process termed the ‘acclimation’ of photosynthesis to elevated [CO2]. Acclimation is usually accompanied by alterations in the gas exchange characteristics that are indicative of a decreased carboxylation capacity (Stitt 1991; Sage 1994; but see also Section 5.2.1) and decrease of Rubisco activity or Rubisco protein in many species (Drake et al. 1997), including tobacco (Sicher & Kremer 1994; Sicher, Kremer & Rodermel 1994), tomato (Van Oosten, Wilkins & Beresford 1994), Scirpus olneyi (Jacob et al. 1995), wheat (Nie et al. 1995a; Rogers et al. 1996; Sicher & Bunce 1997), barley (Sicher & Bunce 1997), rice (Vu et al. 1997), Arabidopsis thaliana (Cheng, Moore & Seeman 1998), Panicum laxum (von Ghannoum et al. 1997), loblolly pine (Tissue, Thomas & Strain 1993; Tissue et al. (1997), cherry (Atkinson et al. (1997), oak (George, Gerant & Dizengremel 1996) and cork oak (Faria et al. 1996). Prolonged exposure to elevated [CO2] has also been shown to lead to a decrease of the levels of transcripts for proteins involved in photosynthesis. In tomato, the levels of the transcripts for the nuclear-encoded small subunit of Rubisco (RbcS) and chlorophyll binding protein (Cab) and Rubisco activase (Rca) decrease in elevated [CO2], whereas the levels of several plastid-encoded transcripts including RbcL, PsaA, PsaB and PsbA (encoding the large subunit of Rubisco and core proteins in photosystems I and II) do not decrease (Van Oosten et al. 1994; Van Oosten & Besford 1995). In wheat, elevated [CO2] led to a decrease of the transcripts for RbcS, RbcL and three other Calvin cycle enzymes (Nie et al. 1995b). Recently, Cheng et al. (1998) reported that long-term exposure to elevated [CO2] leads to a decrease of the level of the RbcS and the RbcL transcripts in Arabidopsis. Moore et al. (1998) reported that the level of the RbcS transcript is decreased in elevated [CO2] in parsley, bugle, cotton, sunflower, wheat, bean and radish. It has been proposed that the acclimation of photosynthesis is due to a sugar-mediated repression of Rubisco and other proteins that are required for photosynthesis (Stitt 1991; Jang & Sheen 1994; Van Oosten & Besford 1996).
Acclimation of photosynthesis to elevated [CO2] is usually more marked in nitrogen-limited plants than in well-fertilized plants (Wong 1979; Oberbauer et al. 1986; Arp 1991; Radoglou, Alpho & Jarvis 1992; Tissue et al. 1993; Pettersson, MacDonald & Standenberg 1993; Sage 1994; Pettersson & MacDonald 1994; El-Kohen & Mousseau 1994; Miglietta, Giuntole & Bindi 1996; Bowler & Press 1996). Analysis of the gas exchange characteristics indicates that the shift to a decreased carboxylation capacity in elevated [CO2] is more marked in nutrient-limited than in well-fertilized plants (Arp 1991; Sage 1994), and that elevated [CO2] leads to a larger decrease of Rubisco in nitrogen-limited plants than in well-fertilized plants (Wong 1979; Tissue et al. 1993; Delgrado et al. 1994; Webber, Nie & Long 1994; Riviere-Rolland, Contard & Betsche 1996; Rogers et al. 1996; Ziska et al. 1996). Further, Riviere-Rolland et al. (1996) showed that elevated [CO2] led to a decrease of Rubisco activity and the RbcS transcript level in pea plants growing in limiting nitrogen, but not when the plants were supplied with a high supply of nitrogen, or were grown with limiting phosphate. The latter treatment indicates that the decrease of Rubisco in the nitrogen-limited plants was not just a non-specific effect, due to slow rates of growth.
These results could be interpreted as evidence that sugar-repression of Rubisco expression operates more effectively in nitrogen-limited plants, which would explain the larger decline of Rubisco and the accentuated acclimation of photosynthesis to elevated [CO2] in nitrogen-limited plants. However, there is an alternative explanation: the decrease of Rubisco may reflect a general decrease of leaf protein, due to reallocation of nitrogen to younger leaves or earlier leaf senescence in nitrogen-limited plants (Coleman, McConnaughay & Bazzaz 1993; Nie et al. 1995a; Miller et al. 1997). These processes may be accelerated in elevated [CO2], because the plants are larger and therefore experience a more acute nitrogen limitation (see Section 3.1.1. and Section 5 for further discussion).
2.2 Nitrogen-limited plants contain large amounts of non-structural carbohydrates even in ambient carbon dioxide and accumulate more in elevated carbon dioxide
Elevated [CO2] typically leads to an increased content of non-structural carbohydrates (Stitt 1991) but the extent of the change varies, depending on the species and conditions (Körner et al. 1995; Poorter et al. 1997; Tissue et al. 1997; Körner et al. 1997). In general, the increase of starch or other carbohydrate storage polymers is larger than the increase of sugars (Morin, Andre & Betsche 1992; Den-Hertog et al. 1996; Poorter et al. 1997; Baxter, Ashenden & Farrar 1997; Fonseca, Bowsher & Stulen 1997; Tissue et al. 1997; Moore, Palmquist & Seeman 1997), but the extent of the changes of sucrose, reducing sugars and starch vary considerably between species (see Bowler & Press 1996; Tuba et al. 1996; Schappi & Körner 1997; Moore et al. 1998).
Nitrogen limitation also leads to an increase of non-structural carbohydrates, which is mainly due to starch (Waring et al. 1985; Hofstra, Lanting & De Visser 1985; Rufty, Huber & Volk 1988; Fichtner & Schulze 1992; Fichtner et al. 1993; Paul & Driscoll 1997; Scheible et al. 1997b). Sugars remain relatively low and sometimes even decrease in nitrogen-limited plants (Rufty et al. 1988; Fichtner et al. 1993; Reddy et al. 1996; Scheible et al. 1997c).
When nitrogen-limited plants are exposed to elevated [CO2] there is a further accumulation of non-structural carbohydrates (Wong 1979, 1990; Bowler & Press 1996; Baxter et al. 1997; Ferrario-Mery et al. 1997; Fischer et al. 1997; Geiger et al. 1999). The nitrogen supply may also influence the relative extent of starch and sugar accumulation in elevated [CO2]. Elevated [CO2] led to a large increase of sugars whereas starch did not markedly increase when Nicotiana plumbaginifolia was growing in hydroponic culture in high nitrate, whereas it led to a large increase of starch and a smaller increase of sugars when the plants were growing in pots on a lower nitrogen supply (Ferrario-Mery et al. 1997). A very marked increase of starch in elevated [CO2] has also been reported in nitrogen-limited sugar beet (Kutik et al. 1995) and Pinus palustris (Pritchard et al. 1997). In tobacco, elevated [CO2] led to an increase of starch and also of sucrose and reducing sugars in nitrogen-sufficient plants, whereas it led to a further and marked increase of starch but not of sucrose and reducing sugars in nitrogen-limited plants (Geiger et al. 1999). On the other hand, Bowler & Press (1996) found an increase of sugars in elevated [CO2] when Agrostis capillaris and Nardus stricta were nitrogen-limited, but not when they received a high nitrogen supply. This may be because these grasses accumulate fructans rather than starch.
2.3 The stimulation of plant growth is decreased and eventually abolished as the nitrogen supply is decreased
Whereas elevated [CO2] typically leads to a marked increase of biomass in well-fertilized plants (Bowes 1996), this response is modified when the nitrogen fertilization is suboptimal. Biomass was not significantly increased by elevated [CO2] in nitrogen-limited soybean (Sionit 1983), tobacco (Geiger et al. 1998b), rice (Ziska et al. 1996), Agrostis capillaris and Nardus strictus (Bowler & Press 1996), Calluna vulgaris, Erica tetralix, Molinia caerula, Rumex obtusifolius, Arrhenatherum elatius (Arp 1998), loblolly pine (Thomas, Lewis & Strain 1994; Johnson, Ball & Walker 1995; Tissue et al. 1997), sweet chestnut (El-Kohen & Mousseau 1994) and several other woody species (Eamus & Jarvis 1989). Elevated [CO2] led to an increase of biomass in nitrogen-limited wheat (Hocking & Meyer 1991a; Mitchell et al. 1993; McKee & Woodward 1994; Rogers et al. 1996; Fangmeier et al. 1997), cotton (Wong 1979; Rogers et al. 1993), Xanthium occidentale (Hocking & Meyer 1985), Vaccinium myrtillus (Arp 1991), Eucalyptus grandis (Conroy, Milham & Barlow 1992), oak seedlings (Norby et al. 1986), loblolly pine seedlings (Gebauer, Reynolds & Strain 1996) and birch (Pettersson et al. 1993) but the relative and absolute stimulation was smaller than in well-fertilized plants. Occasionally, elevated [CO2] produces a large stimulation of growth in nitrogen-limited conditions, including Calluna vulgaris (Whitehead, Caporn & Press 1997) and wheat (Wong & Osmond 1991).
Although the variability may be partly due to species differences, it is probably also a result of differences in the degree of nitrogen deficiency and the way in which the plants were fertilized (see Pettersson & MacDonald 1994). When Eucalyptus spp. (Wong, Kriedermann & Farquhar 1992), birch (Pettersson & MacDonald 1994; Johnson et al. 1995; Silvola & Ahlholm 1995), ponderosa pine (Johnson et al. 1995), loblolly pine (Gebauer et al. 1996), Agrostis capillaris, Nardus stricta (Bowler & Press 1996), rice (Ziska et al. 1996), cotton (Rogers et al. 1993), wheat (Hocking & Meyer 1991b; Rogers et al. 1996) and tobacco (Geiger et al. 1999) were grown at three or more nitrogen supplies, elevated [CO2] led to a large increase of biomass at the highest nitrogen supply, a small increase of biomass at moderately limiting nitrogen supply, whereas it had no positive effect on biomass or even led to a slight decrease of biomass at the lowest nitrogen supply. Different results may be found in ecosystems (Lloyd & Farquhar 1996), but this may reflect long-term interactions between the plants and the soil which modifies the nitrogen supply that the plant is receiving.
These results imply that the nitrogen use efficiency (see Section 2.4) and/or nitrogen uptake and assimilation (see Sections 2.5–2.7) can be increased in elevated [CO2]. However, at a very low nitrogen supply, this is no longer possible. Arp (1998) have proposed that when the nitrogen supply is very low, the tissue nitrogen content is at a minimum and cannot be reduced by further redistribution of nitrogen. This appears plausible, although proof will require a better understanding of what a ‘minimum’ nitrogen content is, and why it cannot be decreased further.
2.4 Nitrogen use efficiency is increased in elevated [CO2]
Innumerable studies have reported that the nitrogen use efficiency (the rate of growth per unit of nitrogen in the plant) increases in elevated [CO2] (see, e.g. Wong 1979; Hocking & Meyer 1991a, b; Pettersson et al. 1993; Rogers et al. 1993; McKee & Woodward 1994; Jacob et al. 1995). This increase of the nitrogen use efficiency is, in part, due to the lower nitrate content in elevated [CO2] (Purvis, Peters & Hageman 1974; Yelle, Gosselin & Trudel 1987; Hocking & Meyer 1985, 1991b). Problems in interpreting the physiological significance of an increased nitrogen use efficiency that are due to a decrease of the nitrate content will be discussed in Section 5.1. However, the organic nitrogen use efficiency (which excludes nitrate and ammonium from the denominator) usually also increases in elevated [CO2] (Sage, Sharkey & Seeman 1990; Pettersson et al. 1993; Baxter et al. 1994).
In elevated [CO2], a given rate of photosynthesis can be achieved with lower activities of Rubisco, other Calvin cycle enzymes, and light reaction components because photorespiration is decreased (Stitt 1991; Quick et al. 1993; Woodrow 1994). This, in principle, could permit decreased investment of nitrogen in the nitrogen-intensive processes of photosynthesis and photorespiration in leaves (see Section 2.1), releasing nitrogen for investment in other processes and tissues. However, it is not yet clear if the increased organic nitrogen use efficiency is solely due to Rubisco, which is a major part of the total leaf protein, operating more efficiently thanks to the higher [CO2] concentration, or if there has been an acclimation resulting in re-allocation of nitrogen to other proteins in the plant.
In transgenic plants expressing an antisense RbcS construct, decreased expression of Rubisco indeed leads to reallocation of organic nitrogen to other proteins in the leaf, including thylakoid proteins resulting in a higher organic nitrogen use efficiency (Makino et al. 1997) and increased rates of photosynthesis in low light (Lauerer et al. 1993; Stitt & Schulze 1994). In elevated [CO2], although numerous studies have reported a decrease of Rubisco (see Section 2.1), few studies have extended the analyses to determine whether there is a selective increase of other proteins. The decrease of Rubisco in elevated [CO2] is accompanied by an increase in the amounts of other leaf proteins in pea (Riviere-Rolland et al. 1996) and wheat (Rogers et al. 1996). The nitrogen released from Rubisco and other proteins of the photosynthesis apparatus could alternatively be allocated to promote the growth of non-photosynthetic organs. However, even though elevated [CO2] stimulates root growth it also leads to an increase of the whole plant leaf area (Wong 1979), and there is no consistent evidence for a selective increase of root growth relative to shoot growth in elevated [CO2], or a selective change in the whole plant leaf area (Pettersson et al. 1993; King, Thomas & Strain 1996; Gebauer et al. 1996; Tissue et al. 1997; Baxter et al. 1997; Drake et al. 1997).
2.5 Elevated carbon dioxide leads to a decrease of the nitrogen content in nitrogen-limited plants, but not necessarily in well-fertilized plants
The overall nitrogen concentration usually decreases when plants are grown in elevated [CO2] (Wong 1979; Curtis, Drake & Whigham 1989; Garbutt, Williams & Bazzaz 1990; Coleman et al. 1991; Coleman & Bazzaz 1992; Hocking & Meyer 1991a, 1991b; Gries, Kimball & Idso 1993; Pettersson et al. 1993; Pettersson & MacDonald 1994; Luo, Field & Mooney 1994; Körner & Miglietta 1994; Newberry et al. 1995; Poorter et al. 1997; Ferrario-Mery et al. 1997; Tissue et al. 1997; Cotrufo, Ineson & Scott 1998). The decrease of the nitrogen content and carbon/nitrogen ratio in elevated [CO2] is partly an arithmetic consequence of the increase of non-structural carbohydrates. Plants in elevated [CO2] accumulate more starch (see above) which can account for up to 40–50% of the leaf weight, and will automatically lead to an apparent decrease in the concentrations of other leaf components. However, the nitrogen concentration usually decreases in elevated [CO2], even after correcting for this effect (see Tissue et al. 1997; Poorter et al. 1997 and other references above).
There have been very few systematic investigations of the changes in the individual nitrogen-containing compounds. Elevated [CO2] often leads to a marked decrease of nitrate (Purvis et al. 1974; Yelle et al. 1987; Hocking & Meyer 1985, 1991b; Ferrario-Mery et al. 1997; Geiger et al. 1998b). There have been hardly any studies of the changes in ammonium and amino acid levels. Elevated [CO2] leads to a decrease of ammonium and of the overall amino acid levels in N. plumbaginifolia and tobacco (Ferrario-Mery et al. 1997; Geiger et al. 1999). Protein typically decreases by 10–20% in elevated [CO2] (e.g. Jacob et al. 1995; Nie et al. 1995a; Sicher, Kremer & Bunce 1995; Ferrario-Mery et al. 1997; Geiger et al. 1998b). In a study of 27 species, Poorter et al. (1997) found that the decrease of leaf protein in elevated [CO2] (on average 18%) was slightly smaller than the decrease of total nitrogen (on average, 21%) and of nitrate (on average, 22%).
The overall decrease of the nitrogen concentration in elevated [CO2] has been interpreted in various ways: it might reflect a higher nitrogen use efficiency due to reallocation of protein (see Section 2.4), or ontogenetic drift leading to accelerated senescence as a result of the faster growth in elevated [CO2], inadequate nitrogen fertilization, or inadequate rates of nitrogen uptake and assimilation in elevated [CO2] (see Section 2.6).
With respect to the latter suggestion, several studies have reported that elevated [CO2] has little or no effect on protein content in young leaves but leads to an earlier decrease in older leaves (Besford, Ludwig & Withers 1990; Nie et al. 1995a; Van Oosten et al. 1994; Miller et al. 1997). Coleman et al. (1993) showed for Abutilon threophrasti that elevated [CO2] leads to a decrease of the nitrogen concentration when plants of the same age are compared, but not when plants of the same size are compared, and argued that the carbon-dioxide-induced reduction in plant nitrogen concentration is not due to changes in the plant nitrogen use efficiency but is more likely to be a size-dependent phenomenon resulting from accelerated plant growth. On the other hand, Roumet et al. (1996) still found a decrease in the nitrogen concentration in elevated [CO2] in 11 different species when they were compared at a common dry weight after growth in ambient and elevated [CO2]. Interpretation of age-dependent changes in the nitrogen concentration requires detailed information about the nutrient supply and the nutrient status of the plant, in order to distinguish between true ontogenetic changes that are linked to internal developmental programmes, and plant-size effects that are related to the exhaustion of the nutrient supply.
Significantly, the nitrogen supply has a marked impact on the effect of elevated [CO2] on the nitrogen concentration in the plant. When wheat and cotton (Rogers et al. 1993, 1996), Avena fatua and Plantago erecta (Chu, Field & Mooney 1996), Nardus stricta and Agrostis capillaris (Bowler & Press 1996), Poa alpina (Baxter et al. 1997) and tobacco (Geiger et al. 1999) were grown in elevated [CO2], there was a large decrease of the leaf nitrogen concentration when the plants were grown at a low nitrogen supply, whereas there was only a small and even non-significant decrease when the plants were grown at a high nitrogen supply. Arp (1991) and Poorter et al. (1997) noted that the decrease in total leaf nitrogen in elevated [CO2] was smaller when plants were grown in hydroponics than in pots, which might again be related to the improved nitrogen nutrition in hydroponic culture. Elevated [CO2] actually led to a small statistically non-significant increase of the nitrogen/carbon ratio and the nitrate and amino acid content when N. plumbaginifolia was grown on high nitrate in hydroponic culture (Ferrario-Mery et al. 1997). Elevated [CO2] also led to a marked and statistically significant increase of the nitrogen concentration and the nitrate, amino acid and protein content of young nitrogen-sufficient tobacco seedlings growing in sand-culture on a high and frequently renewed nitrate supply, but not in older plants (Geiger et al. 1998b). Further, elevated [CO2] stimulated the current growth rate in the young tobacco seedlings, but not in the older plants (Geiger et al. 1998b). These results strongly indicate that the widely reported decrease of the nitrogen concentration in elevated [CO2] may be linked to the accessibility of nitrogen rather than a direct or necessary consequence of the elevated [CO2] (see Section 5.1 for further discussion).
2.6 Elevated carbon dioxide has no consistent effect on nitrate and ammonium uptake
The decreased nitrogen concentration frequently observed in elevated [CO2] (see Section 2.5) indicates that nitrate uptake and assimilation often fail to keep pace with photosynthesis and growth in elevated [CO2]. The effect of elevated [CO2] on whole plant net nitrogen uptake is revealed more directly by comparing the change of the nitrogen concentration with the change of plant biomass. In some studies, the nitrogen content per plant still increased in elevated [CO2] even though tissue nitrogen concentration or the nitrogen/carbon ratio decreased (Wong 1979; Hocking & Meyer 1991b; Chu, Coleman & Mooney 1992; Tissue et al. 1997). In other studies, especially when the nitrogen supply was very low, the total amount of nitrogen per plant was unaltered (Norby et al. 1986; O’Niell, Lumore & Norby 1987; Garbutt et al. 1990; Hocking & Meyer 1991a; Van Ginkel, Gorissen & Van Veen 1997) or reduced (Coleman & Bazzaz 1992; Conroy et al. 1992) in elevated [CO2].
Taken together, these results indicate that elevated [CO2] can lead to increased nitrogen uptake, provided there is an adequate supply of nitrogen outside the plant. As discussed in the following paragraphs, elevated [CO2] could modify inorganic nitrogen uptake by altering the access to or the accessibility of nutrients in the soil, and/or by altering the rates of uptake per unit of root surface.
Plants growing in elevated [CO2] possess larger root systems, at least in absolute terms (Pettersson et al. 1993; Jacob et al. 1995; Jackson & Reynolds 1996; see also Drake et al. 1997), which should allow them to exploit a larger soil volume and might be expected to promote nitrate and ammonium uptake. In addition to an overall increase in root biomass (see above), elevated [CO2] has been reported to increase lateral and secondary root formation in birch (Pettersson et al. 1993), loblolly pine (Larigauderie, Reynolds & Strain 1994), and ponderosa pine (Tingey et al. 1996). Elevated [CO2] altered the distribution of root growth in wheat, promoting root growth in the upper soil layers (Van Vuuren et al. 1997). However, the beneficial effect of an increased root volume might be counteracted by the effects of elevated [CO2] on water flow. Increased [CO2] usually leads to partial closure of stomata, resulting in lower transpirative water flow. This may be advantageous under water-limiting conditions because it leads to a higher water use efficiency, but it will decrease the mass flow of water in the soil to the roots and might therefore decrease the root surface concentrations of soil-mobile minerals, including nitrate (Conroy & Hocking 1993; Van Vuuren et al. 1997). This is unlikely to lead to nitrate becoming limiting in well-fertilized soils, but could decrease access to nitrate when the soil nitrate solution is very dilute. Increased exudation of carbon and trapping of nitrogen in the soil microflora might also modify soil nitrogen availability, especially when overall concentration is low (Schenk et al. 1995).
The reported effect of elevated [CO2] on the rate of nitrate or ammonium uptake per unit root weight is also rather variable. Elevated [CO2] increased the rate of nitrate uptake per unit root weight in loblolly pines (Bassirirad et al. 1996) and Prosopis glandulosa (Bassirirad et al. 1997), did not alter the rate of nitrate uptake in Nardus agrostis (Bassirirad et al. 1997), and decreased nitrate uptake in a mixed field community (Jackson & Reynolds 1996). Ammonium uptake was not altered in elevated [CO2] in loblolly pine (Bassirirad et al. 1996), whereas it was increased in a mixed field community (Jackson & Reynolds 1996). The response of nitrate and ammonium uptake to elevated [CO2] may depend on the nitrogen concentration supplied. Whereas elevated [CO2] led to increased rates of nitrate uptake in loblolly pine growing at high nitrate, it led to decreased uptake in plants growing on low nitrate (Larigauderie et al. 1994). Interpretation of these results on the effect of elevated [CO2] on nitrate and ammonium uptake is complicated by the dynamics of nitrate in the soil, and by the multiple transport systems for nitrate, and the complex regulation of nitrate uptake (see Section 3.1).
2.7 There is no consistent evidence for a stimulation of nitrate and ammonium assimilation in elevated carbon dioxide
Nitrate is reduced in the cytosol via nitrate reductase (NR) to nitrite, which enters the plastid and is reduced to ammonium via nitrite reductase. Glutamine synthetase (GS) then incorporates ammonium into glutamate to form glutamine, and ferredoxin-dependent glutamine synthase (Fd-GOGAT) in a reductive transamination transfers the amino group from glutamine to 2-oxoglutarate to form two molecules of glutamate. Aminotransferases then transfer the amino group from glutamate to aspartate and alanine. Glutamate, aspartate and occasionally glutamine act as the amino donors for the synthesis of other amino acids, nucleotides and other nitrogen-containing compounds in the cell. The major regulatory site in nitrate assimilation is thought to be at NR (see Section 3.2). Nitrate assimilation occurs mainly in the leaves in herbaceous plants, but a larger proportion is assimilated in the roots in many woody species (Andrews 1986; Andrews et al. 1992). De novo ammonium assimilation occurs mainly in the roots (Marschner 1995).
Until recently, there was no consistent evidence for an increase of NR activity in elevated [CO2]. Although elevated [CO2] led to a small increase of NR activity in mustard (Maevskaya et al. 1990) and Vigna radiata (Sharma & Sen Gupta 1990), it produced a two-fold decrease of NR activity in wheat (Hocking & Meyer 1991a, 1991b) and maize (Purvis et al. 1974), a 25% decrease in N. plumbaginifolia grown in vermiculite (Ferrario-Mery et al. 1997), and a 15% decrease in N. plumbaginifolia grown in hydroponic culture (Ferrario-Mery et al. 1997). Elevated [CO2] also led to a 30% decrease of nitrite reductase activity in lettuce (Besford & Hand 1989), which is expressed co-ordinately with NR (Hoff, Truong & Caboche 1994; Vincentz et al. 1993).
It is difficult to understand these results, for two reasons. Firstly, a higher rate of nitrate assimilation will presumably be required in elevated [CO2] to support the higher rate of plant growth. As already discussed, even though nitrogen decreases on a weight basis in elevated [CO2] this is partly due to a decrease of nitrate (Purvis et al. 1974; Yelle et al. 1987; Hocking & Meyer 1985, 1991b), and the organic nitrogen content per plant often increases (see, e.g. Wong 1979; Hocking & Meyer 1991a, 1991b). Secondly, based on our current knowledge about the regulation of nitrogen metabolism (see Section 4.2), we would expect that elevated [CO2] should stimulate rather than inhibit nitrate assimilation.
Recently, Fonseca et al. (1997) showed that transfer of Plantago major to elevated [CO2] leads to a short-term increase of the Nia transcript level in leaves and, especially, in the roots during the first 4 d in elevated [CO2], that is accompanied by a transient 50% increase of NR activity in the shoot and a doubling of NR activity of the roots. Geiger et al. (1998b) showed that long-term exposure of tobacco to elevated [CO2] leads to a sustained increase of NR activity due to subtle changes in the diurnal regulation of NR activity in the leaves and the roots (see Section 5.2.2 for further discussion).
Few studies of elevated [CO2] have differentiated between the use of nitrate and ammonium as nitrogen sources. In principle, elevated [CO2] might favour ammonium assimilation in leaves because photorespiration and the associated recycling of ammonium is decreased (see Sections 3.2.2 and 5.2.3). The levels of the transcripts for the plastid GS and for Fd-GOGAT did not decrease in high [CO2] in tobacco leaves (Migge et al. 1997), plastid GS protein was only slightly decreased, and Fd-GOGAT protein was unaffected. Elevated [CO2] also had no effect on overall GS activity in the leaves of N. plumbaginifolia (Ferrario-Mery et al. 1997). The GOGAT pathway capacity that is released by the decreased rate of photorespiration is therefore available for net ammonium assimilation, or for assimilation of ammonium formed during nitrate assimilation in the leaves. Elevated [CO2] might additionally promote ammonium assimilation in the roots, because more carbohydrate is available to supply energy and carbon skeletons. Elevated [CO2] leads to a three- to four-fold increase of overall GS activity in the roots of tobacco growing on limiting ammonium nitrate in elevated [CO2], especially in nitrogen-limited plants (Geiger et al. 1999).
Summarizing, there is a strong interaction between the nitrogen supply and the response of photosynthesis, metabolism and growth to elevated [CO2]. On the one hand, the acclimation of photosynthesis is accentuated and the stimulation of growth in elevated [CO2] is strongly curtailed in nitrogen-limited plants. On the other hand, elevated [CO2] often leads to a decrease of the internal nitrogen levels. The changes of nitrogen metabolism in elevated [CO2] have, however, not been studied in detail and they have not been related to our knowledge of the regulation of nitrogen metabolism and its interaction with carbon metabolism and growth processes. This makes it difficult to distinguish between possible explanations in which changes in the nitrogen status are seen simply as a change in a resource that is needed for growth, or as a factor that modulates sugar-mediated changes in gene expression, metabolism and growth, or in a more general way, modifies allocation, ontogeny or senescence. The following two sections discuss the emerging evidence from experiments in model systems that nitrate and other nitrogen-containing compounds act as an important source of signals to regulate plant metabolism and growth, and consider how carbon metabolism acts to modulate nitrogen uptake, assimilation and utilization.
3 MANY ASPECTS OF PLANT METABOLISM AND GROWTH ARE REGULATED BY SIGNALS THAT DERIVE FROM NITRATE OR NITROGEN METABOLISM
Application of nitrate fertilizer leads to increased rates of nitrate uptake (Jackson, Flesher & Hageman 1973; Clarkson & Lüttge 1991), increased levels of nitrate in the plant, increased activities of NR (Shaner & Boyer 1976; Galangau et al. 1988), nitrite reductase (Wray 1993), GS and Fd-GOGAT (Sakakibara et al. 1991), increased rates of nitrate assimilation resulting in higher levels of ammonium, amino acids, protein and other nitrogen-containing constituents in the plant (Marschner 1995; Scheible et al. 1997a, 1997b) increased levels of organic acids (Martinoia & Rentsch 1994) and decreased levels of starch (Hofstra et al. 1985; Fichtner & Schulze 1992; Scheible et al. 1997a). The rate of growth is increased, and there are marked changes in allocation resulting in preferential growth of the shoot compared with the roots (Brouwer 1962; Bloom, Chapin & Mooney 1985; Lambers et al. 1990), thicker and less branched roots (Grime et al. 1991; Fichtner & Schulze 1992; Marschner 1995) and delayed flowering (Bernier et al. 1993). Localized application of nitrate to part of the root system stimulates lateral root growth at that site (Drew & Saker 1975; Granato & Raper 1989; Lainé, Ourry & Boucard 1995; Sattelmacher & Thoms 1995). Nitrate also stimulates seed germination (Hilhorst & Karssen 1989).
It has long been suspected that nitrate is not only a resource but also acts, directly or indirectly, to trigger signals that modulate gene expression, metabolism and development (Redinbaugh & Campbell 1991; Crawford 1995; Hoff et al. 1994; Daniel-Vedele & Caboche 1996; Daniel-Vedele, Filleur & Caboche 1998; Stitt & Scheible, 1998). By analogy with micro-organisms, we can expect that nitrogen metabolites such as ammonium, glutamine and asparagine will also act in an analogous manner. If this is the case, the changes in the levels of these compounds in elevated [CO2] (see Section 2.5) can be expected to have profound effects on metabolism and growth. The following section will discuss recent advances in understanding the processes involved in nitrate and ammonium uptake and assimilation, and the significance and possible mechanisms of nitrogen-signalling processes in plants. Genotypes with low activity of NR have provided an invaluable experimental tool to dissect nitrogen signalling, because they allow the supply and the internal level of nitrate to be varied independently of the rate of nitrate assimilation, the levels of amino acids and protein in the plant, and the overall rate of plant growth (Scheible et al. 1997a, 1997b).
3.1. Nitrate uptake
3.1.1 Physiological studies
Based on studies of the multiphasic saturation kinetics, it has been proposed that net nitrate uptake can be mediated by several different systems including a high affinity nitrate-inducible transporter and various low affinity nitrate-inducible transporters (Aslam, Travis & Huffacker 1992; King et al. 1993; Aslam, Travis & Rains 1996), and reversed by a nitrate efflux system (Aslam et al. 1996; Volk 1997).
Nitrate uptake is strongly stimulated after pre-incubation with nitrate. Studies of the influence of pre-incubation on nitrate uptake kinetics indicate that the high affinity transporter and one low affinity nitrate-inducible transport system are stimulated by nitrate, and that another low affinity transporter is constitutive (Aslam et al. 1992; King et al. 1993; Aslam et al. 1996). Pre-incubation with nitrate also increases the rate of nitrate efflux (Aslam et al. 1996; Volk 1997).
Nitrate uptake is subject to feedback regulation, and is therefore inhibited by prolonged incubation with nitrate, or by addition of ammonium. It has been suggested that this decrease may be due to feedback effects that are mediated by nitrate itself, or by ammonium or certain amino acids (see Siddiqi et al. 1990; Clarkson & Lüttge 1991; Muller & Touraine 1993; King et al. 1993; Rufty et al. 1993; Imsande & Touraine 1994). Evidence for a feedback regulation of nitrate uptake by down-stream products formed during nitrate assimilation has recently been provided by Gojon et al. (1998), who showed that tobacco and N. plumbaginifolia transformants with increased expression of NR and consequently increased levels of glutamine have a lower nitrate uptake rate than wild-type plants.
On the other hand, large amounts of nitrate accumulate in NR-deficient mutants and transformants of barley (Warner & Huffaker 1989, N. plumbaginifolia (Ferrario et al. 1995), tobacco (Scheible et al. 1997a) and Arabidopsis (Meyer zu Hoerste 1998), indicating that accumulation of nitrate does not strongly inhibit its own uptake, at least in these species. Lauerer (1996) found up to 50% higher rates of nitrate uptake per unit root weight in tobacco transformants with very low NR activity. Gojon et al. (1998) found unaltered or only slightly lower nitrate uptake rates in N. plumbaginifolia and tobacco genotypes with a small decrease in NR activity, again indicating that high internal nitrate does not lead to a strong feedback inhibition of nitrate uptake, although, as discussed by Gojon et al. (1998), the lower levels of nitrate assimilation in these genotypes would be expected to lead to a stimulation of nitrate uptake and this could be masking a weak feedback regulation of nitrate uptake by nitrate.
There is considerable species-dependent variation with respect to the amount of nitrate that accumulates inside the plant. Large amounts accumulate in tobacco and Arabidopsis (Scheible et al. 1997a; Schulze et al. 1994), smaller amounts in pea (Lexa & Cheeseman 1997) and Plantago major (Fonseca et al. 1997), and very low amounts in rice (Makino et al. 1997). This indicates that the feedback mechanisms regulating nitrate uptake may vary, depending on the species.
3.1.2 Molecular studies
These physiological studies of nitrate uptake have recently been extended by the molecular cloning of the transporters involved in nitrate uptake. These advances mean that the regulation of nitrate uptake can now be subjected to a molecular analysis (see Daniel-Vedele et al. 1998, and Forde & Clarkson 1999 for more details).
A gene encoding a low-affinity, passive nitrate uptake system (Chl-1, now renamed Nrt1 : 1At) was cloned from a T-tagged chlorate-insensitive Arabidopsis mutant (Tsay et al. 1993). Chl-1 is preferentially expressed in roots, and is induced by nitrate. The precise role of this transporter is unclear, however, because the Chl-1 mutant shows pleiotropic changes in cellular pH and in the levels of other inorganic ions. Since the Chl-1 mutant shows decreased rates of nitrate uptake in the presence of ammonium, but not when nitrate is given as the sole nitrogen source (Huang et al. 1996), it has been proposed that there may be further ammonium-repressed low affinity transporters that are able to substitute for the CHL-1 protein in the absence of ammonium. Recently, Lauter et al. (1996) cloned two CHL-1 homologs in tomato, and showed that one of them was expressed constitutively in root cells and root hairs, whereas the other was induced by nitrate and expressed specifically in root hairs.
More recently, sequence homology to the Aspergillus nitrate transporter and Chlamydomonas high-affinity nitrate transporters (Quesada, Galvan & Fernandez 1994) has allowed cloning of genes that encode the high-affinity nitrate transporters in barley (Trueman, Richardson & Forde 1996) and N. plumbaginifolia (Quesada et al. 1997). Genomic analyses indicate the presence of a small multigene family in barley (Trueman et al. 1996). The expression of the high affinity Nrt2 : 1 transporter has been investigated in N. plumbaginifolia and tobacco (Krapp et al. 1998). Expression is restricted to the roots. Nitrate addition leads to a rapid but transient increase of the transcript level in wild-type roots. The contribution of feedback from nitrate itself and from products of nitrate assimilation was dissected using tobacco transformants with very low NR activity. In these plants, nitrate addition led to a strong and sustained increase of the level of the Nrt2 : 1 transcript, whereas ammonium addition led to a strong decrease. This implies that Nrt2 : 1 is repressed by ammonium or compounds produced from ammonium, but not by nitrate accumulation. There were also marked diurnal changes of the Nrt2 : 1 transcript level in the roots of wild-type tobacco plants, with a maximum in the first part of the light period and a decrease later in the photoperiod (A. Krapp. W.-R. Scheible, M. Caboche, unpublished results). These diurnal changes of the Nrt2 : 1 transcript level were modified in mutants with lower expression of NR, and were positively correlated with the diurnal changes in NR protein and activity and negatively correlated with ammonium, glutamine and malate accumulation in the leaves. It will be important to identify the precise signals that link Nrt2 : 1 expression to nitrogen metabolism in the shoot, to investigate how rapidly these changes in transcript translate into changes in transporter protein, and also to investigate whether there are faster changes in root uptake or xylem loading due to post-translational regulation or regulation by low molecular weight effectors.
These physiological and molecular studies demonstrate that there is a close interaction between nitrate uptake, and the metabolism of nitrate to further compounds in the plant. The finding that nitrate uptake and Nrt2 : 1 expression are insensitive to nitrate accumulation but very sensitive to feedback by down-stream signals generated as a result of nitrate assimilation explains why nitrate can be accumulated to relatively large levels in tobacco in excess of that required for current growth. This will, of course, also depend on an effective regulation of nitrate assimilation, to prevent accumulation of down-stream metabolites which inhibit nitrate uptake.
3.2 Nitrate assimilation
Shaner & Boyer (1976) observed that NR activity increases dramatically when plants are re-fertilized with nitrate. Subsequent research has shown that nitrate assimilation is subject to transcriptional, post-transcriptional and post-translational regulation at the step catalysed by NR (Hoff et al. 1994; Crawford 1995; Nussaume et al. 1995; Daniel-Vedele et al. 1998).
3.2.1 Regulation of Nia transcription
The Nia transcript and NR protein are very low in nitrate-limited plants or in the presence of ammonium, and high in plants grown on high nitrate (Hoff et al. 1994; Scheible et al. 1997c). These transcripts also show a marked diurnal rhythm. In nitrate-replete tobacco and tomato, the Nia transcript is present at a high level at the end of the night, stays high or rises slightly during the first 1–2 h in the light, declines markedly during the remainder of the photoperiod, and recovers gradually during the dark period (Galangau et al. 1988; Scheible et al. 1997c; Geiger et al. 1998b). NR protein and activity increase by two- to three-fold in the first 4 h of the light period, decrease during the remainder of the light period and the first part of the dark period, and start to recover during the last part of the night (Galangau et al. 1988; Scheible et al. 1997c; Geiger et al. 1998b). The changes of NR protein occur with a slight time delay compared with the changes of the Nia transcript level, as expected if the changes in the Nia transcript level contribute to the diurnal changes in NR protein.
Nia transcription is induced by nitrate (Pouteau et al. 1989; Cheng et al. 1992; Lin et al. 1994) and repressed by glutamine (Hoff et al. 1994). These nitrogen metabolites probably play a major role in the regulation of Nia expression. The decrease of the Nia transcript during the photoperiod is correlated with a decrease of nitrate (Scheible et al. 1997c) and an accumulation of glutamine (Scheible et al. 1997c), and the gradual recovery during the night is correlated with a decrease of glutamine and a gradual increase of nitrate. Sugar-mediated regulation could also contribute to the light-induced changes of the Nia transcript (see Section 4.2). The transcript for nitrite reductase is regulated in a similar manner, except that nitrite reductase activity is about 10-fold higher than NR activity (Hoff et al. 1994). This co-ordinate regulation presumably ensures that nitrite, which is toxic, is converted through to ammonium.
3.2.2 Post-translational regulation of NR
Faster changes due to reversible post-translational regulation of NR are superimposed on these changes of NR protein. When leaves are darkened, NR is inactivated via a two-step process that involves phosphorylation of ser543 and the subsequent magnesium-dependent binding of an inhibitory 14–3-3 protein to NR (Bachmann et al. 1996; Moorhead et al. 1996).
There is substantial evidence that NR activation is regulated in response to sugar-mediated signals (see Section 4.2.2). Until recently, however, there was no evidence that signals from nitrogen metabolism affect the post-translational regulation of NR. Scheible et al. (1997c) showed that dark inactivation of NR is partially or completely reversed in nitrate-limited wild-type plants and in mutants and transformants with decreased expression of NR. This abolition of dark inactivation was correlated with a decreased level of glutamine and ammonium which are the products of nitrate assimilation, and NR activation could be decreased by feeding glutamine to detached leaves (Scheible et al. 1997c; Morcuende et al. 1998). Nitrate does not affect the activation state of NR (Ferrario, Valadier & Foyer 1996).
3.2.3 Post-transcriptional regulation
There is evidence for further, uncharacterized, mechanisms that modulate either the translation of the Nia transcript and/or rate of the degradation of NR protein. For example, NR protein still decreases in the second half of the photoperiod in transformants that express Nia constitutively under the control of the 35S promotor (Vincentz & Caboche 1991; Ferrario et al. 1995). The decline of NR protein is absent in the second part of the photoperiod in mutants that have a lower number of functional Nia genes, even though the Nia transcript level declines in the same way as in wild-type plants (Scheible et al. 1997c). Similar changes are seen in elevated [CO2] (see Section 5.3). There is also evidence that nitrite reductase is subject to post-transcriptional regulation (Crete, Caboche & Meyer 1997).
Protein phosphorylation also may act as a trigger for protein degradation. When a modified form of NR with a truncated N-terminus that was not susceptible to post-translational dark inactivation was overexpressed, the resulting protein did not decline in the second part of the photoperiod (Nussaume et al. 1995). There is also a correlation between the phosphorylation state or the activation state of NR and the rate at which NR protein decreases (Kaiser & Huber 1997; Scheible et al. 1997c; Geiger et al. 1998b).
3.3 Ammonium uptake and assimilation
3.3.1 Ammonium uptake
Ammonium uptake exhibits multiphasic saturation kinetics with low and high affinity components (Ullrich et al. 1984; Wang et al. 1993). A high affinity ammonium transporter has been cloned in Arabidopsis (Ninneman, Jauniaux & Frommer 1994) and tomato (Lauter et al. 1996). Expression was organ-specific, and transcript for the tomato ammonium transporter LeAmt1 was found in roots, but not in leaves or stems. Within the roots, the LeAmt1 transcript was found in root hairs but also in other root tissues. The transcript level was not markedly decreased by nitrate or increased by ammonium (Lauter et al. 1996).
3.3.2. Ammonium assimilation
Ammonium can be assimilated via the plastid GS2 or a series of cytosolic GS1 isoforms (Lam et al. 1996). Cytosolic GS is present as isoforms that show a cell specific distribution (Lam et al. 1996; Dubois et al. 1996). One isoform of the cytosolic GS is phloem specific (Lam et al. 1996).
Any consideration of ammonium assimilation in leaves is complicated by the interaction between de novo nitrate and ammonium assimilation, and the recycling of the ammonium which is released in leaves during photorespiration. Up to 90% of the ammonium assimilated in the leaves of a C3 plant in ambient [CO2] represents to refixation of photorespired ammonium, rather than de novo assimilation of nitrate or ammonium. Arabidopsis mutants with null activity of the plastid GS were able to grow when photorespiration was suppressed in 2–5% carbon dioxide, but died in ambient [CO2] (Somerville 1986; Lea & Forde 1994; Lam et al. 1996). The cytosolic GS activities are obviously too low, or have an inappropriate location, to cope with the fluxes of ammonium during photorespiration. Incidentally, this does not mean that the plastid GS does not contribute to de novo ammonium assimilation in wild-type plants, indeed it is difficult to see how the ammonium formed by nitrite reductase in the plastid can be separated from the plastid GS.
There is substantial evidence that nitrogen signalling modulates the expression of the GOGAT pathway. Nitrate leads to an increased level of the transcripts for Gln1 (encoding the cytosolic glutamine synthetase, GS1), Gln2 (encoding plastid glutamine synthetase; GS2) and Glu (ferredoxin-dependent glutamate synthase; Fd-GOGAT) in maize roots (Redinbaugh & Campbell 1993), tobacco roots and tobacco leaves (Scheible et al. 1997b), and of the Glu transcript in illuminated barley leaves (Pajuelo et al. 1997). The changes of the transcript for the cytosolic GS were much more marked than for plastid GS, and included the appearance of a second Gln1 transcript (Scheible et al. 1997b). The increase of the GS transcript levels after adding nitrate was accompanied by an increase of overall GS activity in the shoots and roots (Scheible et al. 1997b). Migge et al. (1997) reported that nitrate addition resulted in the appearance of a second form of the plastidic GS in tomato leaves, which was absent when tomato plants were grown on ammonium as the sole nitrogen source. In maize roots, nitrate led to an increase of one GS isoform, and of Fd-GOGAT (Sivasankar & Oaks 1996). In anaerobic rice seedlings, nitrate led to an increase of Fd-GOGAT protein, but not of the cytosolic or plastidic GS (Mattana et al. 1996).
Several lines of evidence show that the GOGAT pathway is probably induced in response to nitrate, rather than the assimilation of nitrate. The transcripts for Gln2, Glu and to a lesser extent, Gln1 also increase when nitrate is supplied to NR-deficient genotypes or genotypes with very low expression of NR (Vaucheret et al. 1990; Kronenberger et al. 1993; Scheible et al. 1997b), and this was accompanied by a marked increase in overall GS activity in the roots though not in the shoots. Addition of tungsten, which blocks the synthesis of the molybdenum-pterin cofactor and inhibits NR activity, had no marked effect on the levels of transcript and protein for plastid GS and Fd-GOGAT (Migge et al. 1997). Further, glutamine and asparagine also had no effect on GS and Fd-GOGAT activity in maize roots even though they led to a marked decrease of NR activity (Sivasankar & Oaks 1996).
3.4 Provision of reducing equivalents for nitrate assimilation
When nitrate is assimilated in leaves in the light, the reducing equivalents are delivered by photosynthetic electron transport. There is ample evidence that addition of nitrate leads to an increase rate of linear electron transport (Foyer et al. 1994a, 1994b). In the alga Selenastrum minutum, this leads to a severe oxidation of the photosynthetic electron chain, and reverses the thioredoxin-mediated activation of the Calvin cycle enzymes resulting in a marked inhibition of photosynthesis (Huppe & Turpin 1994; Turpin, Weger & Huppe 1997). In addition to this direct effect, nitrate may also exert further regulatory action on electron transport. Lauerer (1996) observed that accumulation of nitrate in transformants with low NR activity was accompanied by an increase of the chlorophyll b/chlorophyll a ratio, indicating that nitrate-mediated signals may increase the amount or the antenna size of photosystem II which would alter poising in favour of the production of reducing equivalents.
When nitrate assimilation occurs in the dark in leaves, or in roots, the reducing equivalents are provided by the oxidative pentose phosphate pathway (OPPP), and the reduced ferredoxin required by nitrite reductase is synthesized from NADPH via ferredoxin-NADP-oxidoreductase. Nitrate addition leads to rapid induction of ferredoxin (Matsumara et al. 1997) and ferredoxin-NADP-oxidoreductase (Ritchie et al. 1994) in maize roots. In the green algae Selenastrum minutum addition of nitrate, but not ammonium, leads to a rapid activation of glucose-6-phosphate dehydrogenase and a drop in glucose-6-phosphate and an increase in 6-phosphogluconate levels indicating that the OPPP has been stimulated (Huppe, Vanlenberghe & Turpin 1992; Huppe & Turpin 1994; Turpin et al. 1997). The importance of cross-talk between nitrate metabolism and the redox state of the cell is underlined by the finding that NR expression (Mattana et al. 1996) and activation (Botrel & Kaiser 1997) increases in response to anaerobiosis.
3.5 Regulation of organic acid metabolism by signals from nitrate and nitrogen metabolism
Assimilation of nitrate is accompanied by increased synthesis of organic acids. 2-Oxoglutarate is required as an acceptor in the GOGAT pathway, other organic acids are required during the synthesis of other amino acids, and malate is accumulated as a counter-anion to prevent alkalinization during nitrate assimilation (Martinoia & Rentsch 1994; Heldt 1996). The synthesis of organic acids is regulated at several levels in response to signals from nitrate, ammonium and their metabolism.
Following addition of nitrate or ammonium to the alga Selenastrum minutum, there are rapid changes in the concentrations of glutamine, glutamate, aspartate, 2-oxoglutarate and malate that activate pyruvate kinase and PEP carboxylase (PEPCase) and stimulate the flux of carbon into organic acids (Huppe & Turpin 1994; Turpin et al. 1997). There is evidence for an analogous allosteric regulation of these enzymes by organic acids and amino acids in higher plants (Turpin et al. 1997).
Protein phosphorylation (see MacKintosh 1998; for a recent review) allows a coordinate activation of PEPCase (Foyer et al. 1994a; Duff & Chollet 1995; Li, Zhang & Chollet 1996) and inhibition of sucrose phosphate synthase (SPS) (Champigny et al. 1992; Manh et al. 1993; Huber & Huber 1996) after adding nitrate or ammonium. Activation of PEPCase is regulated via de novo synthesis of PEPCase-kinase (Hartwell et al. 1996), resulting in decreased sensitivity to inhibition by malate (Chollet, Vidal & O’Leary 1996). The malate sensitivity of PEPCase decreased in the presence of nitrate in low-NR tobacco (Scheible et al. 1997b), indicating that nitrate-mediated signals contribute to the post-translational regulation of PEPCase.
There is also evidence that signals from nitrogen metabolism allow coordinate transcriptional regulation of a group of key genes for organic acid synthesis. Nitrate addition leads to an increase of the transcripts for PEPCase, the cytosolic pyruvate kinase, citrate synthase and the NADP-dependent isocitrate dehydrogenase (NADP-ICDH) (Scheible et al. 1997b). These enzymes convert phosphoenolpyruvate to 2-oxoglutarate which is the primary carbon acceptor for inorganic nitrogen, and to malate which acts as a counteranion during nitrate assimilation. Crucially, these transcripts also change after adding nitrate to transformants which have very low NR activity, showing that the increase is triggered by nitrate itself rather than other compounds that are formed during nitrate assimilation (Scheible et al. 1997b). The transcript levels change within 2–4 h, showing that nitrate acts rapidly and that high internal concentrations are not required (Scheible et al. 1997b; and A. Krapp unpublished). Significantly, nitrate addition does not lead to a significant increase of the transcript for fumarase, an enzyme in the section of the tricarboxylic acid cycle that is not required for net synthesis of malate or 2-oxoglutarate (A. Krapp, unpublished results). The changes in transcript levels were accompanied by an increase of the activity of PEPCase (Foyer et al. 1994b; Scheible et al. 1997b), pyruvate kinase, citrate synthase and NADP-ICDH (A. Krapp, W.-R. Scheible and M. Stitt, unpublished), and an accumulation of 2-oxoglutarate and other organic acids (Scheible et al. 1997b).
Addition of nitrate or ammonium also leads to a rapid and marked increase of the Ppc transcript and of PEPCase activity in maize (Sugiharto et al. 1992b; Sugiharto & Sugiyama 1992; Suzuki et al. 1994). Comparison of the kinetics of various metabolite pools with the changes of the Ppc transcript indicated in this case that glutamine may play a key role in the induction of PEPcase. Addition of nitrate or ammonium also led to an increase of PEPCase activity in barley, which was prevented when phosphinothricin was added to inhibit GS activity (Diaz, Lacuestra & Munoz-Rueda 1996), again indicating that glutamine plays a key role in the induction of PEPCase in this species.
3.6 Modulation of nitrate storage
The nitrate content of plant tissues varies enormously. Nitrate levels are very low in nitrate-limited plants, and increase when the nitrate supply is increased or the rate of plant growth is decreased, for example in low light or short days (Matt et al. 1998), or in transformants with low rates of photosynthesis (Stitt & Schulze 1994). Most of the nitrate is located in the vacuole (Miller & Smith 1996). It has been suggested that this nitrate is not readily accessible to metabolism. However, an interruption of the nitrate supply leads within hours to remobilization of the nitrate from epidermal cells in the leaf and root (Dietz, Hollenbach & Hellwege 1994; van de Leij et al. 1998). There are also large (> 20-fold) changes in the nitrate pool in source leaves during the day (see, e.g. Scheible et al. 1997c).
Microelectrode measurements have shown that the cytosolic nitrate level is maintained remarkably constant, irrespective of the level in the vacuole and of whether nitrate is being accumulated or released (Miller & Smith 1996; van de Leij et al. 1998). These results imply that the cytosolic nitrate concentration is tightly regulated, and that the movement of nitrate into and out of the vacuole is not related to changes in the cytosolic nitrate level (see discussion in van de Leij et al. 1998).
The transport proteins involved in nitrate transport across the tonoplast, in the entry of nitrate into the xylem, and in the entry and release of nitrate from cells in the stem and leaves have not yet been identified or cloned. Recently, we have observed that nitrate is remobilized very rapidly from the roots and stems of low-NR tobacco transformants, and that this remobilization can be markedly slowed by supplying glutamine (A. Krapp, and M. Stitt, unpublished results). These results indicate that down-stream signals generated during nitrate assimilation regulate the remobilization of stored nitrate.
3.7 Regulation of the level of the AgpS transcript and of starch accumulation by nutrients
Nitrate limitation typically leads to a large increase of the starch content. It has been tacitly assumed that starch accumulation is a passive response to the decreased rates of growth, and the resulting accumulation of sugars. High levels of sugars can lead to increased starch synthesis via feedback regulation of the sucrose synthesis pathway, resulting in an increase of the 3PGA:phosphate ratio and allosteric activation of AGPase, the key regulatory enzyme in the pathway of starch synthesis (Neuhaus & Stitt 1990; Preiss 1991; Stitt 1991, 1996; Huber & Huber 1996). Sugars also exert transcriptional control on AgpS (Müller-Röber et al. 1990; Krapp & Stitt 1995; Koch 1996; Nielsen et al. 1998), the gene encoding the regulatory subunit of AGPase. However, recent research shows that nitrate, and other nutrients including phosphate, are able to modulate starch accumulation in a more direct manner.
The AgpS transcript level decreases within 2–4 h when nitrate is added to nitrogen-deficient tobacco, and increases again after removal or use of the nitrate (Scheible et al. 1997b). Crucially, the same changes are seen after adding nitrate to tobacco transformants with low NR activity, showing that the signal is related to nitrate itself rather than the metabolization of nitrate (Scheible et al. 1997b). Further, the rate of starch remobilization after adding nitrate is similar in wild-type plants and in the low-NR transformants, even though the rate of growth is not significantly altered in the latter (Scheible et al. 1997b). The observation that the changes in starch metabolism can be effectively uncoupled from the changes in growth rates and carbohydrate consumption demonstrates the importance of nitrate-related signalling for the regulation of starch metabolism. Incidentally, this repression of AgpS is not due to a general inhibition of carbohydrate synthesis by nitrate. The Sps transcript and SPS activity are unaltered or even increased after adding nitrate to wild-type plants or low-NR transformants (Scheible et al. 1997b).
The AgpS transcript level also decreases when phosphate is supplied to detached leaves (Nielsen et al. 1998). Phosphate acts antagonistically to exogenous sugars, but the effect of phosphate is not due to changes in the endogenous sugar levels as these do not change after feeding phosphate (Nielsen et al. 1998). It is not yet known if phosphate itself, or specific phosphorylated metabolites, trigger the increase in the AgpS transcript.
Transcriptional regulation of starch metabolism in response to signals that are derived from nutrients like nitrate and phosphate will allow a feedforward regulation of carbon allocation, and permit coordinate changes in carbohydrate allocation and nutrient assimilation and utilization, without this requiring large intervening changes in the sugar pools. In some cases, the direct effects of nutrients on carbon allocation may be so powerful that they replace or even override the regulation by sugars. This would explain why nitrate- and phosphate-deficient plants sometimes contain larger amounts of starch and lower levels of sugars in their leaves than nutrient-sufficient plants (Brooks 1986; Rao & Terry 1989; Fichtner et al. 1993; Scheible et al. 1997a; Geiger et al. 1999).
3.8 The sugar-mediated repression of genes involved in photosynthesis is modulated by nitrogen status
3.8.1 The nitrogen supply affects the levels of enzymes involved in photosynthesis
When the nitrogen supply to a plant is decreased there is a general decrease of protein and chlorophyll, but the relative amounts of Rubisco, thylakoid proteins and other leaf proteins are not markedly affected (Evans 1996). In some cases, there is a slight decrease of Rubisco relative to total protein (Terashima & Evans 1988; Fichtner et al. 1993; Makino et al. 1992, Makino, Nakana & Mae 1994) which explains why Rubisco exerts more control over the ambient rate of photosynthesis in nitrogen-deficient than in nitrogen-sufficient plants (Quick et al. 1993). This is in striking contrast to changes in the growth irradiance, which lead to marked shifts in the relative amounts of chlorophyll and various proteins (Evans 1996).
Although it is hardly surprising that the overall protein concentration and, more specifically, the amount of Rubisco and other enzymes of photosynthesis increase when nitrogen fertilizer is supplied, the molecular mechanisms underlying this increase have not been investigated. It is unlikely that the nitrogen-driven increase of overall protein is a purely passive response (see Section 3.10). Intriguingly, low-NR transformants growing on high nitrate contained slightly higher Rubisco activity than nitrate-limited wild-type tobacco, or the same transformants growing on limiting nitrate (Lauerer 1996). Further, whereas the RbcS transcript level decreases dramatically in nitrate-deficient wild-type tobacco, it remains significantly higher in low-NR transformants when they are grown on high nitrate (W.-R. Scheible, A. Krapp, M. Stitt, unpublished results).
3.8.2 Sugar-mediated regulation of RbcS expression is modulated by the nutrient status
Removal of nitrogen leads to a cessation of growth and a remobilization of protein from the source leaves. Recently, Paul & Driscoll (1997) showed that the decline of leaf protein and Rubisco after removing nitrogen can be prevented if the light intensity is simultaneously decreased, to prevent carbohydrates from accumulating. This result indicates that there is a close interaction between nitrogen- and sugar-signalling events regulating leaf protein levels.
The interactions between nitrogen status and sugar-mediated repression of RbcS have also been investigated by growing tobacco plantlets in nutrient agar on different ammonium nitrate levels in the presence and absence of sucrose (Paul & Stitt 1993; Nielsen et al. 1998). Nitrogen-limited plants contained lower amounts of total protein and lower Rubisco activity, but the RbcS transcript level was not decreased relative to total RNA. When sucrose was supplied to nutrient-sufficient plants, growth was increased, and protein, Rubisco activity and the RbcS transcript level remained high (see also Kovtun & Daie 1995). When sucrose was supplied to nitrogen-limited plants, there was a marked decrease of the RbcS transcript level, Rubisco, protein and chlorophyll.
This increased susceptibility to sugar-repression in nitrogen-limited plants is not just a consequence of slow growth, because sucrose did not lead to a decrease of protein, Rubisco activity or the RbcS transcript level in phosphate-limited plants (Nielsen et al. 1998). Indeed, in older plants, phosphate limitation typically leads to an increase of the activity of Rubisco and other enzymes that are required during photosynthesis (Brooks et al. 1986; Rao & Terry 1989; Fredeen et al. 1990). Low temperature also leads to an increase of the leaf sugar content and an increase of the activities of the Calvin cycle enzymes (Holaday et al. 1992; Hurry et al. 1994, 1995), and Strand et al. (1997) recently showed that low temperature leads to an increase of the levels of the transcripts for RbcS and other Calvin cycle enzymes in Arabidopsis leaves, even though sugars simultaneously accumulate to high levels.
Taken together, these results indicate that sugars only repress the genes for Calvin cycle enzymes in specific circumstances. There is a striking analogy with the effect of elevated [CO2] on photosynthesis, Rubisco activity and the RbcS transcript level in nitrogen-and phosphate-deficient plants (see Riviere-Rolland et al. 1996; discussed in Section 2.1). The resulting specificity may be important, because it allows the acclimation response to be adjusted, depending on the nutrient stress encountered. Thus, decreased expression of the proteins for photosynthesis in older leaves during nitrogen stress will allow reallocation of nitrogen to support growth of roots, young leaves or reproductive organs, whereas increased activity of the Calvin cycle enzymes during phosphate stress will contribute to the maintenance of photosynthetic rate, and increased activity of the Calvin cycle enzymes as a result of cold acclimation will also contribute to the maintenance of photosynthetic rate, protect against photoinhibition, and allow the accumulation of sugars that act as a cryoprotectant (see Strand et al. 1997 and Nielsen et al. 1998 for further discussion).
3.9 Regulation of whole plant allocation, root architecture and phenology by nitrogen signalling
Nitrogen fertilization not only leads to increased growth, but also results in alterations in allocation and phenology (see above). Recent experiments using transformants with low NR activity have revealed that nitrate-mediated signalling triggers at least some of these changes in allocation and development.
Plants growing with a low nitrogen supply typically have a large root biomass relative to the shoot biomass (Brouwer 1962; Van de Werf & Nagel 1996), and the roots are often much more finely branched than in well-fertilized plants (Grime et al. 1991; Fichtner & Schulze 1992). In low-NR tobacco transformants, however, the accumulation of nitrate was accompanied by a strong inhibition of root growth and stimulation of shoot growth, resulting in shoot : root ratios that were higher than in nitrogen-replete wild-type plants (Scheible et al. 1997a) even though the low-NR transformants were severely nitrogen-limited with respect to organic nitrogen. Split-root experiments indicated that the inhibition of root growth was triggered by the accumulation of nitrate in the shoot, not the root. The inhibition of root growth was correlated with decreased allocation of carbon to root growth (Scheible et al. 1997a) and with a selective decrease in the number of lateral roots (Stitt & Scheible 1998; Stitt and Feil 1999).
Similar results have also been obtained in Arabidopsis (Zhang & Forde 1998; Meyer zu Hoerste 1998). Interestingly, Lexa & Cheeseman (1997) did not find any difference in the shoot-root ratio in a pea NR-deficient mutant. This indicates that there may be species-dependent differences in the extent to which nitrate and other signals regulate shoot-root allocation. Significantly, pea (Lexa & Cheeseman 1997) also differed from tobacco (Scheible et al. 1997a) and Arabidopsis (Meyer zu Hoerste 1998) in having relatively low levels of nitrate in the leaves, in not showing such a marked increase of leaf nitrate in response to a rising nitrate supply in the medium, and in having lower nitrate in the shoot than the roots. These differences might, speculatively, be linked to the ability of pea to form root nodules, in which case nitrate would not be the major nitrogen source.
The action of nitrate on root growth is complex, because localized application of nitrate leads to a local stimulation of lateral root proliferation (Drew & Saker 1975; Granato & Raper 1989). Localized application of nitrate still leads to a proliferation of lateral roots in tobacco (Scheible et al. 1997a) and Arabidopsis (Zhang & Forde 1998) genotypes with very low NR activity, and the root proliferation is not accompanied by an increase of the local concentrations of amino acids or protein (Scheible et al. 1997a), showing that it again involves nitrate-mediated signalling, rather than a simple nutrient-mediated increase in growth. It will be important in the future to investigate how the internal nitrate pool and the external distribution in the soil interact, to regulate the overall rate of root growth, the location of root growth, and root architecture.
Further changes induced by nitrogen fertilization include a delay in flowering, and a stimulation of germination (see above). The delay in flowering of Arabidopsis can probably be attributed to the accumulation of nitrate, rather than high levels of amino acids or protein (Meyer zu Hoerste 1998). The induction of seed germination also occurs independently of the assimilation of nitrate (Hilhorst & Karssen 1989).
3.10 Molecular analysis of nitrogen signalling
Summarizing the last sections, nitrate triggers a coordinated change in the expression and activity of a programme of enzymes involved in nitrate and ammonium assimilation, organic acid synthesis, and starch synthesis, resulting in a redirection of metabolism away from starch synthesis and towards nitrogen assimilation and amino acid production, and also triggers changes in whole plant allocation and in development that affect root growth and architecture, flowering and seed germination. It is probable that other compounds in nitrogen metabolism interact with nitrate, for example glutamine antagonizes the nitrate-mediated induction of the high affinity nitrate transporter, NR and nitrite reductase, and is likely to reinforce the positive effects of nitrate on events lying further down-stream in metabolism, allocation and growth. The molecular dissection of nitrate- and other nitrogen-signalling pathways in plants is still in its infancy. The following section briefly outlines the cardinal features of nitrogen-signalling pathways in other model organisms, and then reviews current research on nitrogen signalling in higher plants.
3.10.1 Nitrogen sensing in bacteria
In Enterobacteria, nitrate acts to regulate two genetically distinct different sets of NR and nitrite reductase genes that have different functions (Stewart 1994). One pathway uses nitrate as an electron acceptor in anaerobic conditions and is induced by nitrate at low oxygen tension (Unden et al. 1995). Nitrate is sensed by two overlapping two-component regulatory systems (Kaiser & Sawers 1995; Darwin & Stewart 1995). The sensor components (NARX, NARQ) interact with nitrate and nitrite and modify the activity of the response elements (NARL, NARP), which are DNA-binding proteins. NARX and NARQ both contain a conserved histidine residue that is needed for the phosphorylation and activation of the response regulators NARL and NARP (Cavicchioli et al. 1995). NARX and NARQ are thought to sense nitrate in the periplasmic space (Cavicchioli et al. 1996; Chiang, Cavicchioli & Gunsallis 1997).
The corresponding genes for a genetically and biochemically distinct NR and nitrite reductase that function in the pathway for nitrate assimilation are encoded on the Ntr regulon in Escherichia coli. The feedback regulation of nitrate assimilation by down-stream metabolites has been intensively investigated. Transcription of the Ntr regulon is inhibited when reduced nitrogen sources are available, via a cascade which includes the bi-functional UTase/UR (urididyl transferase/urididyl removing enzyme), the PII protein, the bifunctional kinase/phosphatase NRII, and the transcription factor NRI (Kamberov et al. 1994). The regulation of this cascade by metabolites may involve measurement of the glutamine concentration by the UTase/UR protein, and of 2-oxoglutarate by the PII protein (Kamberov, Atkinson & Ninfa 1995).
In Klebsiella pneumoniae a pathway-specific factor NASR has been identified that mediates nitrate-induction of the nitrate assimilation pathway (Goldman, Lin & Stewart 1994; Lin & Stewart 1996). The leader sequence of the operon contains a transcription terminator, and NASR may act by overriding this termination signal (Lin & Stewart 1996). Expression of the nitrate assimilation pathway is also subject to general nitrogen control mediated by the NTRC protein, which leads to repression when ammonium is supplied (Goldman et al. 1994; Lin & Stewart 1996).
3.10.2 Nitrogen sensing in fungi
A similar picture, in which of nitrate uptake and nitrate assimilation are subject to a pathway-specific induction by nitrate, and are repressed by ammonium and products of ammonium as part of a general response to these metabolites, is also found in fungi. The best studied systems are Aspergilus nidulans and Neurospora crassa (Crawford & Arst 1993; Marzluf 1993a, 1993b). Nitrate induction of Nit3 (encoding NR) is mediated by the pathway-specific regulatory protein NIT4 in Neurospora, and the homolog NIRA in Aspergillus. Ammonium repression is mediated by the globally acting regulatory protein NIT2 in Neurospora and the homolog AREA in Aspergillus, whereby NIT2 and AREA act positively to increase the expression the enzymes for nitrate assimilation and a large number of unlinked structural genes that specify nitrogen-catabolic enzymes in conditions of nitrogen limitation, specifically low glutamine. Operation of NIT2 is regulated by a negative-acting regulator NMR (Xiao, Fu & Marzluf 1995).
3.10.3 Nitrogen sensing in eukaryotic algae
A NIT2 homolog has been cloned from Chlamydomonas, where it is required for expression of three nitrate and nitrite transporter proteins, NR, and nitrite reductase, which form a gene cluster reminiscent of that found in fungi (Galvan, Quesada & Fernandez 1996). Nit2 is expressed in the presence of nitrate and is repressed by ammonium (Quesada et al. 1993; Schnell & Lefebvre 1993). Two loci, Nrg1 and Nrg2 have been identified that are required for the ammonium-repression of the nitrate transporters, NR and nitrite reductase activity, although they may act in a different manner to the NIT4/NIRA system in fungi (Prieto et al. 1996). There is evidence that NR itself regulates the expression of the nitrate transport proteins (Navarro et al. 1996).
NR is also subject to post-transcriptional and post-translational regulation in Chlamydomonas, although the mechanisms differ form those in higher plants. As in tobacco species, constitutive overexpression of Nia leads to an increase of the Nia transcript but not necessarily of NR protein. In Chlamydomonas, nitrate appears essential for the increase of active NR protein (Navarro et al. 1996). Activation of NR may be related to a redox interconversion process (Franco, Cardenas & Fernandez 1987). As in higher plants, there is evidence that inactivation of NR acts as a signal for its subsequent degradation (Franco et al. 1987).
3.10.4 Nitrogen sensing in plants
Nitrogen sensing in higher plants is likely to be even more complex than in micro-organisms. Plants are multicellular organisms, and nitrogen affects many facets of metabolism and physiology at the cellular and the whole plant level. There are two further intriguing differences between plants and micro-organisms. First, due to the prevalence of nitrate in the soil and the problems of pH regulation that arise during ammonium assimilation, most species that have been investigated in detail appear to utilize nitrate as their major and preferred nitrogen source. Second, the large changes in glutamine that can occur in leaves as a consequence of photorespiratory ammonium release may complicate the use of glutamine as a general signal for the nitrogen status of the plant, at least in leaf mesophyll cells.
The analysis of nitrogen sensing in higher plants has been seriously hampered because the types of screen (e.g. insensitivity to chlorate, Nia expression mutants) that yielded large numbers of regulatory mutants in other organisms did not yield regulatory mutants or have proved refractory in higher plants. This may be due to the complexity of the regulation networks and/or the presence of redundant regulatory proteins. The identification of a large number of other genes whose expression is modulated by nitrate or by other nitrogen metabolites (see above, Sections 3.1–3.5) and the identification of nitrate-modulated processes that provide visual phenotypes (Section 3.7) could open new possibilities to isolate mutants in nitrogen-signalling processes in the near future. In addition, several other approaches have started to yield information about nitrogen-signalling processes in plants.
One involves the use of sequence homologies or functional complementation screens to isolate putative homologs to genes that are involved in nitrogen signalling in other organisms. A NIT2 homolog (Daniel-Vedele & Caboche 1993) and an Arabidopsis PII homolog that lacks the uridenylation site (G. Coruzzi, personal communication) have been identified in plants. A functional complementation screen using a yeast mutant deficient in Gln3 (a yeast homolog of AreA and Nit2) has allowed the isolation of two genes designated Rga1 and Rga2 (Truong et al. 1997), which are members of a multigene family that includes SCARECROW which is involved in pattern formation in roots (Di Laurenzio et al. 1996), and have subsequently turned out to be identical with Gai (Peng et al. 1997) and Rga (Silverstone, Ciampaglio & Sun 1998) which are implicated in giberellin signal transduction.
In a second approach, Zhang & Forde (1998) have used differential screening to identify genes that are rapidly induced after addition of nitrate, and isolated a novel nitrate-regulated gene in Arabidopsis (Anr1) which encodes a member of the MADS-box family of transcription factors. Antisense inhibition of Anr1 expression prevents proliferation of lateral roots in response to local application of nitrate, so identifying ANR1 as a important component of the signalling pathway by which nitrate modulates lateral root growth.
In a third approach, Sakakibara et al. (1998) have provided evidence that a cytokinin-inducible protein in maize leaves (pZm CIP1) is involved in the signal transduction pathways of nitrogen signalling. Earlier results from this group had already shown that the C4-cycle PEPCase can be induced by nitrogen sources in intact plants but not in detached leaves where, instead, cytokinins are effective in increasing PEPCase expression (Sugiharto, Burnell & Sugiyama 1992a; Suzuki et al. 1994). The level of the pZ mCip1 transcript and pZmCIP1 protein also increase transiently after supplying nitrate or ammonium to the roots of maize plants, but not after supplying nitrate or ammonium to detached leaves (Sakakibara et al. 1998). Further, isopentyladenosine accumulates in the root within 2 h of adding nitrate (Sakakibara et al. 1998). Taken together, these results suggest a model for a nitrogen-sensing system which involves the synthesis of cytokinins in the roots, followed by translocation of the cytokinins to the shoot and induction of pZ mCip1 in the leaf, although the precise role of pZm CIP1 still has to be established. Intriguingly, the sequence of pZ mCip1 shows homologies to the response regulator element of bacterial two component signalling systems. A small family of response regulator proteins (ARR3-ARR7) that show a similar cytokinin and nitrate-dependent expression has recently been identified in Arabidopsis (Taniguchi et al. 1998).
These results are fascinating because cytokinins are already implicated in other nitrogen-mediated events. There is indirect evidence implicating cytokinins in the increase of the level of the Nia transcript level and NR activity after adding nitrate to barley plants (Samuelson, Campbell & Larsson 1995), and zeatin riboside has been implicated in the regulation of NR in Chenopodium cell suspension cells (Peters, Füchbauer & Beck 1995). The formation of lateral roots is known to be induced by auxin and inhibited by cytokinins (Webster & Radin 1972). Nitrogen fertilization leads to increased levels of cytokinins in the roots and root sap (Kuiper et al. 1989; Samuelson, Eliason & Larsson 1992; Wagner & Beck 1993; Sakakibara et al. 1998), and the effects of nitrate fertilization on shoot-root allocation can be mimicked by adding cytokinin (Fetene & Beck 1993). Further evidence for an interaction between nitrogen signalling and cytokinins is provided by the phenotype of a low nitrogen-sensitive mutant isolated by Faure, Jullien & Caboche (1994). This mutant maps to the Zea3 complementation group that confers resistance to zeatin, and exhibits a light-dependent sensitivity of cotyledon growth to a low nitrogen/carbon ratio.
It will also be important to identify which nitrogen pools or processes in nitrogen metabolism are sensed. As outlined in Section 3.8, there is evidence that shoot nitrate rather than root nitrate may be involved in the regulation of shoot-root allocation in tobacco. The finding that the cytosolic nitrate level is fairly constant (see Section 3.6) implies that sensing systems for nitrate may be located in the apoplast or at the plasmalemma, and in the vacuole or at the tonoplast. The observation that the Nia transcript level responds to re-irrigation with nitrate, even when the leaves still contain a large pool of nitrate (Geiger et al. 1998b; D. Klein and M. Stitt, unpublished) provides independent evidence that the inflow rather than the internal pool of nitrate is being sensed.
With respect to glutamine signalling, an intriguing problem arises in leaves, because ammonium and glutamine are produced during photorespiration as well as during de novo nitrate and ammonium assimilation. The major pools of ammonium and glutamine appear to be shared by both processes (Scheible et al. 1997c; Geiger et al. 1998b), which raises interesting problems with respect to the extent to which signals derived from these pools can give specific information about the nitrogen status. The cytosolic GS1 occurs as cell specific isoforms, including one expressed specifically in the phloem complex (Dubois et al. 1996; Lam et al. 1996). Since GS1 expression responds more specifically to changes in sugars and nitrate (see Sections 3.3 and 4.3), and there may be important cell-specific events that are not seen in overall measurements of metabolites.
Summarizing, although the molecular analysis of nitrogen signalling in higher plants is still in its infancy, the information available paints a tantalizing picture in which signals derived from nitrate and nitrogen metabolism interact with other fundamental processes in the regulation of plant gene expression, plant hormones and plant development. These interactions underline their potential importance for the control of plant metabolism and development.
4 NITROGEN METABOLISM IS REGULATED BY SIGNALS THAT ARE DERIVED FROM CARBON METABOLISM
The potential contribution of sugar-mediated regulation of gene expression for understanding the acclimation of photosynthesis was pointed out in Stitt (1991). Since then sugar-mediated signalling has been implicated in the regulation of a host of further processes (Koch 1996; Graham 1996; Smeekens & Rook 1997), including nitrogen metabolism (Hoff et al. 1994; Krapp & Stitt 1995; Lam et al. 1996). The following section discusses how nitrate uptake, assimilation and utilization can be regulated in response to signals derived from carbon metabolism. The term ‘sugar-mediated’ regulation is used, without implying a particular mechanism whereby the sugars are leading to the observed changes in transcript levels or metabolism that result after feeding sugars (see Section 4.5 for a discussion of the mechanisms of sugar sensing).
There are two overriding reasons why nitrogen metabolism needs to be regulated in response to changes in carbon metabolism. Firstly, the utilization of sucrose in growing sinks will normally depend upon simultaneous provision of suitable amounts of amino acids. Secondly, nitrate assimilation depends on and in some conditions even competes with carbon metabolism. Nitrate assimilation diverts large amounts of carbon away from carbohydrates and towards the synthesis of organic acids (see Section 3.5). The assimilation of nitrate and the synthesis of amino acids requires reducing equivalents and ATP (Section 3.4), and therefore competes with photosynthetic carbon fixation and carbohydrate synthesis in leaves in limiting light, and requires catabolism of carbohydrates to provide ATP and reducing equivalents in leaves in the dark and in non-photosynthetic tissues. Over 20% of the total root respiration may be related to nitrate uptake and assimilation in barley roots (Bloom, Sukrapanna & Warner 1992). Provided carbohydrates are not limiting, the carbon required for nitrate assimilation can be obtained by reducing the rate of starch accumulation or by re-mobilizing starch or sugar reserves that were accumulated at an earlier time point (Waring et al. 1985; Fichtner & Schulze 1992; Scheible et al. 1997a). However, when carbohydrates are in short supply, continued nitrate assimilation could lead to an metabolic imbalance, resulting in accumulation of ammonium, a shortfall of organic acids to counteract acidification, or a lack of carbon for export or respiration. Conversely, when carbohydrates are available in large amounts, it will be important to increase the rate of nitrogen assimilation.
4.1. Nitrate uptake
There have been relatively few studies of the effects of sugar supply on nitrate uptake, and the effects of changes in carbohydrate levels on the expression of the various nitrate transport proteins has not yet been investigated. Physiological evidence that the sugar supply influences nitrate uptake has been provided by studies of the diurnal changes of nitrate uptake in soybean plants (Delhon et al. 1995). Nitrate uptake is faster in the light period than the dark period, and evidence for a role of sugars was provided by three different manipulations. This light-dependent stimulation was prevented by removal of carbon dioxide from the atmosphere on the previous day or girdling of the stem to prevent sugar export to the roots, and was restored by providing sucrose to the roots. Further, the rate of nitrate uptake during the night was decreased even further by decreasing the light intensity during the preceding light period to reduce starch accumulation and, thence, the supply of sucrose during the ensuing dark period. Recently, it has been observed that addition of sugars to the rooting medium leads to a two-fold higher levels of the Nrt2 : 1 transcript in N. plumbaginifolia seedlings (A. Krapp, unpublished).
Nevertheless, plants growing in low light intensities or in short days (Matt et al. 1998) and transformants with low rates of photosynthesis due to antisense inhibition of the expression of Calvin cycle enzymes (Stitt & Schulze 1994; V. Haake and M. Stitt, unpublished) typically contain extremely high levels of nitrate. This indicates that nitrate uptake is inhibited less than nitrate assimilation when carbohydrates are low. The energetic demands of nitrate uptake are much lower than the energetic and carbon costs for nitrate assimilation, and a large pool of nitrate would provide a reserve of nitrogen that can be immediately exploited if conditions change and photosynthesis and growth can be increased.
4.2 Nitrate assimilation
Whereas nitrate can be readily stored and can even perform an important function as an osmoticum, nitrite is toxic, and ammonium may also be toxic at high concentrations. Further, since the protonated forms are volatile, when nitrite and ammonium accumulate to high concentrations they may exit the plant, which would represent a considerable loss of energy and nitrogen. For these reasons, NR represents a key site where the flow of nitrogen into organic compounds is regulated in response to the supply of carbohydrate.
4.2.1 Transcriptional regulation
Sugars are implicated in the transcriptional regulation of Nia in leaves. The Nia transcript level increases in detached leaves after feeding sugars, especially when the leaves have been predarkened to deplete the endogenous sugars (Cheng et al. 1992; Vincentz et al. 1993), and in leaves on intact plants when the petiole is cooled to inhibit phloem export and increase the endogenous sugar levels (Krapp & Stitt 1995). Sugars are also implicated in the regulation of Nia expression in roots. During prolonged darkening, the level of the Nia transcript decreases, and this is reversed when sugars are supplied to the roots (Botrel & Kaiser 1997; Sivasankar, Rothstein & Oaks 1997).
It is less clear, however, whether and when sugar-mediated signals make a major contribution to the regulation of NR assimilation. Correlations between changes of sugars and the Nia transcript level or NR activation in leaves during light–dark transitions and diurnal rhythms cannot be unambiguously interpreted, because many other potential regulators including nitrate and glutamine are changing simultaneously. The evidence for an effect of sugars on Nia transcript and NR activity is mainly based on feeding experiments in which rather high levels of sugars were supplied exogenously (Vincentz et al. 1993; Sivasankar et al. 1997; Botrel & Kaiser 1997). The levels of sugars that accumulate after petiole cooling are much higher than are normally found in leaves (Krapp & Stitt 1995) and, although NR activity decreased dramatically in antisense RbcS transformants, this decrease was only found in lines with a large reduction of Rubisco activity, photosynthesis and sugar levels (Stitt & Schulze 1994). When tobacco plants are grown in short day-conditions or exposed to an extended night, Nia transcript levels and NR activity do not decrease until sugars are below about 3–4 μmol/g FW (Matt et al. 1999). Recently, we investigated the effect of transferring low-NR tobacco transformants to permanent darkness (D. Klein, A. Krapp. M. Stitt, unpublished results). The Nia transcript level did not decrease for 2 to 3 d, by which time the sugar content had fallen to very low levels. These studies all indicate that sugar-mediated regulation only plays a major role in the regulation of Nia transcription when the sugar levels in leaves are very low. The observation that sucrose and glutamine act antagonistically on the level of the Nia transcript level (Morcuende et al. 1998), however, suggests that higher levels of sugars may be required in conditions where the plants are less depleted in amino acids.
4.2.2 Post-translational regulation of NR
Sugars are also implicated in the post-translational regulation (see Section 3.2.2) of NR. Dark-inactivation is reversed if sugars are supplied to leaves in the dark (Kaiser & Brendle-Behnisch 1991; Kaiser & Huber 1994a), and light-activation does not occur if carbon fixation is prevented by water stress or low [CO2] concentrations (Kaiser & Förster 1989; Kaiser & Brendle-Behnisch 1991). Experiments with extended night treatments indicate that post-translational regulation of NR is sensitive to relatively small changes of sugar levels (Matt et al. 1998; D. Klein, A. Krapp and M. Stitt, unpublished results). As discussed in Section 3.2.3, there is evidence that the rate of NR degradation may also be initiated by NR phosphorylation, in which case sugar-mediated regulation may contribute to the regulation of NR protein turnover.
4.3 Ammonium uptake and assimilation
The cytosolic glutamine synthetase (GS1) transcript level increases in the light, and this increase can be mimicked by adding sugars in the dark (Lam et al. 1996). It is not known if this leads to changes in GS1 protein, nor whether post-translational mechanisms or allosteric effectors regulate GS activity. Recently, Morcuende et al. (1998) showed in feeding experiments with detached leaves that sugars lead within 2 to 4 h to an increased flux from nitrate into glutamine and other amino acids, even though the levels of glutamate and ammonium decrease, providing strong evidence that sugars stimulate GS activity in vivo.
4.4 Provision of carbon skeletons for ammonium assimilation, and amino acid synthesis
The subsequent synthesis of amino acids depends on carbon skeletons produced by photosynthesis and sugar catabolism. Relatively little is known about the regulation of organic acid synthesis and amino acid synthesis by sugars. However, some recent experiments indicate that the stimulation of de novo nitrogen assimilation is accompanied by a coordinated regulation of carbon flow into the pools of organic acids and from there into the amino acid biosynthesis pathways.
4.4.1 Synthesis of organic acids
Recent experiments indicate that the sugar-mediated stimulation of nitrate and ammonium assimilation described in Sections 4.2 and 4.3 is accompanied by a coordinated stimulation of organic acid synthesis. Morcuende et al. (1998) showed that sucrose stimulates the synthesis of organic acids in detached tobacco leaves. Irrespective of whether the sucrose was supplied on its own, or with nitrate, or with glutamine, it led to a marked increase of 2-oxoglutarate, a small decrease of isocitrate, and a large decrease of the glycolytic precursors glycerate-3-phosphate (3PGA) and phosphoenolpyruvate (PEP). These results show that organic acid synthesis is being stimulated via activation of specific steps, including those catalysed by PEPCase, pyruvate kinase and NADP-isocitrate dehydrogenase, and not just as a passive response to an increase supply of substrates. These changes were seen within 2 to 4 h, indicating that the sugars are leading to rapid changes in the activity of these enzymes. Leport et al. (1996) reported that darkening or exposure to CO2 free-air led to an inhibition of in vivo PEPCase activity as monitored by 14C fixation, directly showing that changes in photosynthetic carbohydrate production lead to rapid changes in PEPcase activity.
The increased rates of dark fixation found by Leport et al. (1996) were not accompanied by a change in the in vitro activity or a marked change in the malate sensitivity of PEPcase, indicating that the activation may involve allosteric regulation rather than protein phosphorylation. Although PEPCase can be regulated by protein phosphorylation (see Section 3.5), the contribution during sugar-mediated activation has not yet been established.
Sugar feeding also leads to an increase of the levels of the transcripts for several important enzymes involved in the synthesis of organic acids (Koch 1996), including the cytosolic pyruvate kinase (Krapp et al. 1993) and the NADP-dependent isocitrate dehydrogenase (NADP-ICDH) (Fieuw et al. 1995). However, it is not yet known whether the sugar-mediated increase of the transcript levels leads to an increase of the encoded proteins. Rapid diurnal changes in the transcripts for PEPCase, cytosolic pyruvate kinase and NADP-ICDH and two- to three-fold changes in the activities of the encoded enzymes have recently been observed in tobacco leaves (A.Krapp, W.-R Scheible and M. Stitt, unpublished results).
4.4.2. Synthesis of the minor amino acids
The minor amino acids (the aromatic amino acids, branched and unbranched chain aliphatic amino acids, lysine, histidine, arginine, cysteine, methionine, proline) are synthesized from a variety of carbon precursors, using glutamate, aspartate and, occasionally, glutamine as the amino donor. The regulation of amino acid synthesis by feedback regulation from the individual amino acid end products has been extensively researched (Lea & Forde 1994). Amino acid synthesis in mammals and fungi is also regulated by the general nitrogen control system, in which GNC4 acts as a transcription factor to promote transcription of more than 30 amino acid biosynthesis genes in response to starvation for a single amino acid (Dever et al. 1992; Pain 1994). There is evidence that Opaque-2 is a functional plant homolog of GNC4 (Mauri et al. 1993) although its function in processes other than seed storage protein synthesis has not been investigated in detail.
Relatively little is known about the regulation of amino acid biosynthesis by the availability of the carbon precursors. Carbon starvation leads to a general decrease of amino acids in a sugar cane suspension culture (Veith & Komor 1993). This could be due to an inhibition of amino acid synthesis, or a stimulation of amino acid catabolism. Evidence that prolonged carbon starvation leads to amino acid catabolism is provided by the finding that asparagine synthetase is induced when leaves are darkened (Lam et al. 1996) and during starvation of detached maize root tips (Brouquisse et al. 1992; Chevalier et al. 1996), and is repressed again if sugars are added. Matt et al. (1998) showed that a transient decrease of sugars after a single dark night leads to a rapid and reversible decrease of all the minor amino acids and an increase of glutamate and aspartate, and that re-illumination leads to an increase of the minor amino acids and a simultaneous decrease of glutamate and aspartate. Further, Morcuende et al. (1998) showed that when sugars are supplied to detached source leaves the rates of nitrate assimilation and the levels of all the minor amino acids increase, but the levels of glutamate and aspartate decrease. These results indicate that sugars, directly or indirectly, stimulate flux into the amino acid biosynthesis. It will be important in the future to investigate whether sugars stimulate the expression and activity of enzymes in amino acid biosynthesis pathways.
Summarizing the previous sections, there is mounting evidence that sugar-mediated signalling modulates nitrogen metabolism, acting at several sites including nitrate assimilation itself, ammonium assimilation, the synthesis of organic acids, and possibly the flow of nitrogen and carbon into the various amino acid biosynthesis pathways. These sugar-mediated signals will complement the nitrogen-mediated signals discussed in Section 3. In agreement, it is necessary to provide nitrate and sucrose to achieve maximum rates of nitrate assimilation and amino acid synthesis in detached leaves (Morcuende et al. 1998).
4.5 Mechanisms of sugar sensing
The mechanisms involved in sugar sensing in plants have been recently reviewed (Sheen 1994; Thomas et al. 1995; Koch 1996; Graham 1996; Jang & Sheen 1997; Smeekens & Rook 1997; Smeekens, 1998). There is evidence that hexokinase may play an important role as a glucose sensor in plants (Jang & Sheen 1994; Jang et al. 1997), as has also been suggested in yeast. It is, however, likely that the mechanisms involved in sugar sensing in plants will be considerably more complex than in yeast. Plants are multicellular organisms, and contain different sorts of cells with strikingly different requirements for and dependence on sugars.
The evidence for the role of hexokinase as a sugar-sensor is partly based on experiments with glucose analogues like mannose and 2-deoxyglucose to show that phosphorylation is a requirement for the signal but further metabolization is not required (Sheen 1994; Jang & Sheen 1994; Martin et al. 1997; Koch 1996), The use of glucose analogues is, however, compromised because there is considerable metabolization of the phosphorylated derivative in plant cells (Klein & Stitt 1998). The other stronger line of evidence involves the analysis of the effect of glucose on gene expression transformants with decreased expression of hexokinase and, especially, transformants with overexpression of ectotopic hexokinases (Jang et al. 1997). However, the way in which hexokinase functions as a sugar sensor has not yet been elucidated at the molecular level, even in yeast (Smeekens, 1998).
Some evidence is apparently inconsistent with the operation of hexokinase as a glucose sensor in plants. Tobacco transformants with overexpression of yeast invertase showed strikingly different phenotypes, depending upon whether the invertase was overexpressed in the apoplast, the vacuole or the cytosol (Sonnewald et al. 1991). Whereas the apoplast- and vacuole-overexpressors showed a marked bleaching, a loss of chlorophyll (Sonnewald et al. 1991), decreased protein and Rubisco levels (M. Brauer and M. Stitt, unpublished results), decreased levels of the Cab transcript, and induction of wound-induced proteins (Herbers et al. 1996), transformants with overexpression of invertase in the cytosol did not show changes in expression of cab and wound-induced proteins (Herbers et al. 1996) nor a decrease of Rubisco activity (M. Brauer and M. Stitt, unpublished results). Herbers et al. (1996) proposed that hexose sensing occurs in the endomembrane lumen. An alternative explanation (Smeekens, 1998) would be that sensing requires movement of hexoses back into the cytoplasm from either the apoplast of the vacuole, which would still be compatible with a role for hexokinase but would require the additional postulate that it only acts as a sensor in association with specific transport processes.
The picture is further complicated by recent evidence from yeast that sugar transporter homologs can function as hexose sensors to regulate a subclass of the sugar-mediated responses (Özcan et al. 1996). There is already evidence that the transport of hexoses rather than their metabolism acts as the signal to induce some groups of sugar-responsive genes in plants. Thus, transportable but non-metabolizable glucose analogs like 6-deoxyglucose were able to induce invertase and sucrose synthase in Chenopodium cells (Godt, Riegel & Roitsch 1995; Roitsch, Büttner & Godt 1995) and reporter genes linked to the sugar-responsive potato patatin-1 promotor in Arabidopsis (Martin et al. 1997).
Other recent evidence indicates that plants can also sense sucrose. Sucrose was more effective than reducing sugars in inducing patatin expression in detached potato leaves (Wenzler et al. 1989; Jefferson, Goldsbrough & Bevan 1990), a rolC-promotor GUS reporter gene construct was activated more effectively by sucrose feeding than by glucose or fructose in the phloem of tobacco seedlings (Yokoyama et al. 1994), and sucrose was more effective than glucose in decreasing sucrose symporter activity and transcript levels in detached leaves (Chiou & Bush 1998). Sucrose was also more effective than glucose in activating nitrate assimilation and the synthesis of organic acids in tobacco leaves (Morcuende et al. 1998). Recently, it has been shown that sucrose represses synthesis of a protein termed ATB2, probably acting at the level of translation, and that glucose or fructose are less effective (Rook et al. 1998). ATB2 is a vascular-tissue specific BZIP transcription factor, and might therefore play a role in mediating sucrose-induced changes in gene expression in sink tissues. It should be noted however, that a rapid cycle between sucrose and hexose sugars in many plant tissues (see Dancer, Hatzfeld & Stitt 1990; Goldschmidt & Huber 1992; Krapp et al. 1993; Geigenberger & Stitt 1993) will complicate the interpretation of experiments where different sugars are supplied. It is also important to include a control in which similar amounts of glucose or fructose, i.e. double the molarity of the sucrose concentration added, are included and this has not always been done in studies to date that report a selective effect of sucrose on gene expression.
There is mounting evidence that sucrose and reducing sugars exert a differing influence on sink metabolism and development. Invertase activities are high in stolons but following tuberization, the activity of invertase decreases and sucrose synthase increases, with the result that growing tubers contain relatively low levels of hexose sugars and high levels of sucrose (Oparka et al. 1992; Ross et al. 1994; Appeldoorn et al. 1997). Potato tuber-specific overexpression of yeast invertase leads to a decrease of the starch content and a stimulation of respiration (Sonnewald et al. 1997), and these changes are accentuated when glucokinase is introduced into the invertase-overexpressing tubers, even though the concentrations of all of the glycolytic precursors are massively increased (Trethewey et al. 1998). Further, addition of glucose promotes respiration in potato discs, whereas sucrose promotes starch accumulation even though the levels of the glycolytic metabolites decrease (Geiger Stitt & Geigenberger 1998a). In an elegant analysis of gene expression and carbohydrate profiles in developing Vicia seeds, Wobus and coworkers have shown that high invertase activity and high levels of hexoses are found in zones characterized by cell division, whereas starch accumulation is found in zones characterized by high sucrose synthase and high sucrose (Heim et al. 1993; Weber et al. 1995a, 1995b, 1997a, Weber, Borisjuk & Wobus 1997b; Borisjuk et al. 1998).
It is therefore important to avoid premature generalizations about the mechanisms whereby sugars regulate gene expression and metabolism in plants. Sugar feeding or endogenous accumulation of sugars lead to far-reaching changes in many different aspects of plant metabolism and development, including processes related to rapid catabolism, cell division, wounding, photosynthesis, nitrate assimilation and storage processes. Some of these processes are linked or share common features, but some are mutually antagonistic, and different sugars or sensing mechanisms may be involved. This means that caution is needed in assuming that the molecular details of the sensing mechanism that are revealed by analysis of the sugar-mediated regulation of one set of genes in one tissue and physiological or developmental state are relevant for other tissues, states, or groups of genes. Caution is also needed in assuming that effects produced by feeding sugars are caused by the sugars themselves, rather than the metabolization of the sugars, or even more indirect effects due to changes in other processes that result as an indirect consequence of sugar-feeding.
An important recent development has been the isolation of mutants with altered sugar-sensing properties (Smeekens & Rook 1997). The plants becoming available include rsr mutants altered in the susceptibility of the patatin-1 promotor to sugar-induction (Martin et al. 1997), sun mutants showing decreased susceptibility to sugar repression of a plastocyanin promotor-LUC reporter gene construct in tobacco seedlings (Dijkwel et al. 1997), and mutants altered in the sugar-induction of (α– amylase (Mita et al. 1997). These mutants will provide an invaluable resource to analyse the mechanisms of sugar sensing, to define subclasses of responses, and to investigate if the nitrogen supply or elevated [CO2] affect any, some, or all of these responses. In addition, they are already providing strong evidence for cross-talk between sugar sensing and other signal transduction pathways. SUN6 and SUN7 modulate two specific subclasses of PHYA responses in a sugar dependent manner (Dijkwel et al. 1997).
5 THE SIGNIFICANCE OF CARBON:NITROGEN INTERACTIONS FOR UNDERSTANDING AND ANALYSING THE RESPONSE OF METABOLISM AND GROWTH TO ELEVATED CARBON DIOXIDE
As mentioned in Section 2, although numerous studies have described the interaction between elevated [CO2] and nitrogen supply, few of these studies included a detailed analysis of nitrogen metabolism. A detailed analysis of nitrogen metabolism is required to understand the mechanisms that govern the response to elevated [CO2], for two reasons. Firstly, increased rates of plant growth in elevated [CO2] may lead to plants becoming nitrogen limited or exacerbate an existing nitrogen limitation, and some of the changes in metabolism and physiology in elevated [CO2] might be indirect effects, due to the plants becoming more nitrogen limited. It will therefore be important to analyse nitrogen metabolism in parallel with changes in photosynthesis, carbon metabolism and growth to identify conditions in which complicating effects due to changes in the nitrogen status are absent or minimized, and to distinguish between direct effects of elevated [CO2] and an increased supply of photosynthate, and indirect effects due to nitrogen limitation. Secondly, a detailed analysis of nitrogen metabolism is needed to identify changes in specific pools in nitrogen metabolism which might be of significance for the regulation of metabolism, development and growth in elevated [CO2]. This could include factors which increase nitrogen uptake and assimilation, or the allocation of nitrogen in the cell and the plant.
5.1 Evaluation of the nitrogen status in elevated carbon dioxide
Obviously, correct interpretation of experiments with elevated [CO2] will depend on an adequate set of criteria to assess whether the plant nitrogen status has been significantly changed. This has traditionally been done by measuring the total nitrogen content, and in some cases by resolving total nitrogen into nitrate and total organic nitrogen, or measuring total protein (see Section 2.4). Such measurements may be useful when they can be interpreted against a large background of empirical information for a particular species and type of location, but have several serious limitations if this background information is lacking.
The interpretation of changes of total leaf nitrogen is complicated by the large amounts of nitrate that can accumulate in well-fertilized plants. Due to the dramatic diurnal changes in the leaf nitrate content (Gebauer, Melzer & Rehder 1984; Scheible et al. 1997c), the nitrate content measured at an arbitrary time of the day will not give reliable information about the nitrogen status of the plant. Further, the nitrate present in a plant leaf at a given time in its life history represents a surplus, that has not been reduced and used to produce organic nitrogen. To interpret the functional significance of a change in the nitrate pool, harvests at several time points during the growth of the plant will be required to reveal whether or not the nitrate is subsequently utilized. If the nitrate is not later remobilized, a decrease of nitrate in elevated [CO2] would have little or no functional significance for plant growth, although it may affect the nutritional quality of the litter. The amount of nitrate accumulated can anyway vary greatly between species (se Section 3.1.1).
Changes of total organic nitrogen or protein are also difficult to assess on their own because they are not large (see Poorter et al. 1997) and because Rubisco itself represents 20 to 50% of leaf protein (Evans 1989; Fichtner et al. 1993; Makino et al. 1997). Based on changes of total protein, it is not possible to distinguish between (i) a decrease of leaf protein due to selective decrease of Rubisco and possibly other proteins involved in photosynthesis, and (ii) a general decrease of leaf protein. These two potential explanations, however, will have very different functional consequences (see also the discussion in Baxter et al. 1997).
To avoid the complication caused by the special case of Rubisco, Baxter et al. (1997) suggested that the root protein content is a better marker for the plant nitrogen status. They also proposed further useful parameters which can be monitored to detect changes in the plant nitrogen status including an alteration in the shoot : root ratio and premature senescence of older leaves. A further useful criteria is to use a range of nitrogen supplies including concentrations that are shown to be super-optimal for plant growth. Another very powerful approach would be to compare the changes of total leaf protein with the changes of Rubisco and a group of other selected photosynthesis proteins and of a group of housekeeping proteins.
It is also advisable to carry out a detailed analysis of the levels of the various nitrogen metabolites in the plant. There have been remarkably few investigations of the effect of elevated [CO2] on the levels (Ferrario-Mery et al. 1997; Geiger et al. 1998b) and the diurnal changes (Geiger et al. 1998b) of nitrate, ammonium and the individual amino acids. As discussed in Section 3.4, nitrogen-sufficient plants show highly dynamic changes in the pools of nitrogen metabolites. During the light period nitrate decreases, and considerable amounts of ammonium and of selected amino acids accumulate, especially glutamine, glycine and serine. These changes are then reversed during the night. (Ferrario-Mery et al. 1997; Geiger et al. 1998b). Many of the other amino acids, including the minor amino acids, do not show such marked changes (Geiger et al. 1998b). Further, the effect of the nitrogen supply on the levels of nitrate, ammonium and selected amino acids such as glutamine and asparagine or on the glutamine : glutamate or asparagine : aspartate ratios are far faster and, in relative terms, far larger than the changes of total nitrogen, total protein, Rubisco, chlorophyll or total amino acids (Geiger et al. 1998b; Geiger et al. 1999). They therefore provide better information about short- and long-term changes in the plant nitrogen status, and will also provide an early warning that a nitrogen limitation is developing, long before significant changes in source leaf protein or the overall nitrogen concentration are detectable. Several of these parameters including nitrate, glutamine and asparagine may also act as internal signals for the current nitrogen status (see Section 3). It may also be useful to complement measurements of leaf pools with measurements of the pools in the roots, where the diurnal changes are less marked (Scheible et al. 1997c; Geiger et al. 1998b).
Studies that have included a detailed analysis of the changes in nitrogen metabolism in elevated [CO2] provide the following picture. (i) In plants receiving a super-optimal nitrogen supply, elevated [CO2] either leads to an increase of nitrate (Ferrario-Mery et al. 1997; Geiger et al. 1998b) or, even when it leads to a decrease (Geiger et al. 1999), the nitrate level is still high. The overall level of amino acids, the levels of most of the individual amino acids (Ferrario-Mery et al. 1997; Geiger et al. 1998b, 1999), and the protein content (Baxter et al. 1997; Geiger et al. 1998b) in the leaves and the roots do not decrease significantly, and may even show a small increase. Although there is a decrease of ammonium and of selected amino acids including glycine, serine and glutamine (Ferrario-Mery et al. 1997; Geiger et al. 1998b) this related to changes in photorespiration (see below) rather than to changes in the nitrogen status of the plant. (ii) When elevated [CO2] is supplied to nitrogen-limited plants, there is a relatively small decrease of leaf and root protein (Ferrario-Mery et al. 1997; Baxter et al. 1997; Geiger et al. 1999), a decrease of nitrate (Ferrario-Mery et al. 1997; Geiger et al. 1999), and a decrease of leaf amino acids which includes a particularly marked decrease of glutamine, glycine and serine (Ferrario-Mery et al. 1997; Geiger et al. 1999). There is a decrease of total root nitrogen (Baxter et al. 1997), which includes a decrease of the amino acids and protein (Geiger et al. 1999). (iii) The difference between plants growing in ambient and elevated [CO2] is especially marked when the nitrogen supply is just sufficient for nitrogen-sufficient growth in ambient [CO2], because elevated [CO2] leads to the plants becoming nitrogen limited. This is accompanied by large changes of nitrate, ammonium and glutamine, and a marked decrease of protein and a wide range of enzyme activities in the leaf (Geiger et al. 1999). It will be difficult to interpret the effect of elevated [CO2] on metabolic or physiological parameters in intermediate nitrogen regimes, because many of the observed changes may just be indirect effects caused by the change in the nitrogen status. By careful choice of the fertilization regime (see below for further discussion), however, it may be possible to minimize these effects.
5.2 The regulation of nitrogen metabolism in elevated carbon dioxide
As discussed in the last section, exposure of suboptimally fertilized plants to elevated [CO2] results in the appearance or accentuation of a nitrogen limitation. This could be due to a combination of factors, including exhaustion of the available nitrogen in the rooting medium, decreased availability of the nitrogen in the soil (see also Section 2.6), and dilution due to rapid growth and a more rapid progression through the life cycle (see Baxter et al. 1997). On the other hand, when plants which are receiving a super-optimal supply of inorganic nitrogen are exposed to elevated [CO2], they are able to increase the nitrogen uptake and assimilation to keep pace with or even overtake the increased rate of carbon fixation and growth. The following section will relate recent investigations of nitrogen metabolism in elevated [CO2] to the emerging information about the role of sugars and nitrogen metabolites in the regulation of nitrogen metabolism (Sections 3 and 4).
5.2.1 Uptake of nitrate and ammonium
The reported effects of elevated [CO2] on nitrate and ammonium uptake are extremely variable (Section 2.6). The overall rate of nitrate uptake can be increased in elevated [CO2], resulting in an increased concentration of total nitrogen and the major components of the nitrogen household plants growing in high nitrate in hydroponics and in young seedlings (see Section 5.1). The failure to see a stimulation of nitrate uptake in other systems could be due to a variety of factors, including depletion of soil nitrate which is needed to induce nitrate transporters (see Section 3.2). The molecular cloning of the high and low affinity nitrate transporters and the ammonium transporters will enable the effect of elevated [CO2] on the expression of the various transporters to be studied, and compared with the changes of nitrate, ammonium and nitrogen metabolites. It will also allow transformants to be produced with altered expression of the transporters, to test whether nitrate or ammonium uptake is a limiting step for growth in elevated [CO2].
5.2.2 Nitrate assimilation
The reported effects of elevated [CO2] on NR activity are also very variable (Section 2.7). Nonetheless, elevated [CO2] clearly leads to increased NR activity in Plantago (Fonseca et al. 1997) and tobacco (Geiger et al. 1998b). The reasons for this increase can be quite complex. In tobacco, elevated [CO2] did not markedly increase NR activity in the first part of the light period when NR activity is at its diurnal peak, but it decreased or prevented the decline of NR activity in the second part of the photoperiod, and partially reversed the dark-inactivation of NR (Geiger et al. 1998b). It also led to a change in the diurnal regulation of NR in roots (Geiger et al. 1998b).
Although there was a slight and transient increase of the Nia transcript level after transferring Plantago major to elevated [CO2] (Fonseca et al. 1997), prolonged growth in elevated [CO2] led to a slight decrease of the Nia transcript level in N. plumbaginifolia (Ferrario-Mery et al. 1997) and tobacco (Geiger et al. 1998b) leaves. This decrease was seen in nitrogen-sufficient plants (Ferrario-Mery et al. 1997; Geiger et al. 1998b) even in seedlings, where the internal pools of nitrate and the current growth rate were increased in elevated [CO2] (Geiger et al. 1998b). Elevated [CO2] also fails to prevent the rapid decay of the Nia transcript during the photoperiod (Scheible et al. 1997c; Geiger et al. 1998b). The absence of a positive effect of elevated [CO2] on the Nia transcript level is consistent with the evidence (see Section 4.2.1) that plants growing in ambient [CO2] contain enough sugars for them not to be a major factor in regulating the level of the Nia transcript level, except when the light regime is unfavourable for photosynthesis.
Two lines of evidence indicate that elevated [CO2] acts, instead, via post-transcriptional and post-translational mechanisms to increase NR activity. First, elevated [CO2] reduced the decline of NR activity and protein during the second part of the photoperiod (see Section 3.2.1), and partially reversed the dark inactivation (see Section 3.2.2) of NR (Scheible et al. 1997c; Geiger et al. 1998b). Second, when N. plumbaginifolia transformants expressing Nia under the control of the constitutive 35S promotor were compared with wild-type plants in hydroponic culture on high nitrate, NR activity was only 30% higher in the transformants in ambient [CO2] whereas it was 100% higher in elevated [CO2] (Ferrario-Mery et al. 1997). Interestingly, this stimulatory effect of elevated [CO2] on NR expression in the 35S transformants was reversed when the genotypes were compared in pots with a lower nitrogen supply. This indicates that uncharacterized post-transcriptional mechanisms also contribute to the decrease of NR activity in elevated [CO2] when nitrogen becomes limiting.
The increased activation and increased stability of NR in elevated [CO2] might be a result (see Section 4.2.2) of the higher levels of sugars in elevated [CO2] (see Section 2.2). However, an intriguing similarity between the diurnal changes of NR protein and NR activation in elevated [CO2] (see above) and the changes in tobacco mutants with a decreased number of functional Nia gene copies (Scheible et al. 1997c) suggests an alternative explanation. In both cases, the decay of NR protein during the photoperiod was decreased, dark-inactivation was partially reversed and, crucially, these changes correlated with a slower accumulation of glutamine during the light period. These results indicate that the altered diurnal regulation of NR in elevated [CO2] may be partly due to a weakening of the feedback regulation exerted by glutamine or related metabolites that are formed during nitrate assimilation. The lower level of glutamine in elevated [CO2] may be partly due to increased export of glutamine or increased conversion of glutamine to other amino acids as a result of the higher sugar levels (see Section 4.4). It may also (see Section 5.2.3) be partly due to the decreased rate of photorespiration in elevated [CO2].
Elevated [CO2] also leads to increased NR activity in the roots (Fonseca et al. 1997; Geiger et al. 1998b). The increased supply of sugars may play an important part in the regulation of root NR expression and activity. Fonseca et al. (1997) observed an increase of the Nia transcript level and a larger and more sustained increase of root NR activity than shoot NR activity after transferring Plantago major plants to elevated [CO2], and suggested that this was due to the increased availability of sugars in the root. Geiger et al. (1998b) grew tobacco plants for several weeks in elevated [CO2], and found that their roots contained higher NR activity than in ambient [CO2] as a result of a change in the diurnal rhythm of root NR activity. A transient increase of sugars at the start of the photoperiod was followed by an increase of NR activity, which in turn was followed by a decrease of nitrate and an increase of nitrite and amino acids, and then by a decrease of sugars to lower levels than in ambient [CO2], and a decrease of NR activity (Geiger et al. 1998b). The decrease of sugars might be due to their consumption during nitrate assimilation.
Two studies have shown that the extent and timing of the increase in NR activity in elevated [CO2] correlates with the time period during which the current growth rate is stimulated. After transfer of Plantago major to elevated [CO2], growth was only stimulated during the first days, and this growth stimulation was not accompanied by a decrease of reduced nitrogen (Fonseca et al. 1997). When tobacco was grown continuously in elevated [CO2], the relative growth rate was only stimulated between 12 and 20 d compared to that in ambient [CO2], and at this time the levels of amino acids and protein were higher in elevated than in ambient [CO2] (Geiger et al. 1998b). In both cases, it was proposed that the up-regulation of nitrate assimilation plays an important role in allowing higher rates of growth in elevated [CO2].
Several factors could explain why many other studies have found unaltered or even decreased NR activity in elevated [CO2] (see Section 2.7). Firstly, in tobacco the increase of NR activity was due to a change in the diurnal regulation (see above), and studies investigating activity at a single time point might not detect these subtle changes. Secondly, elevated [CO2] could lead to a preferential assimilation of ammonium, resulting in formation of glutamine and (see Section 3.3) repression and/or post-translational inactivation and degradation of NR. There have been very few comparisons of the response to elevated [CO2] in the presence of nitrate and ammonium nitrate, and many studies have used a mixture of nitrate and ammonium fertilization. Geiger et al. (1999) recently observed that elevated [CO2] always led to a 30 to 50% decrease of NR activity in leaves of plants growing on super-optimal ammonium nitrate, whereas it leads to an increase of NR activity in plants growing on super-optimal nitrate. Thirdly, the low NR activity in some studies in elevated [CO2] might be due to the plants having become nitrate-deficient. NR activity decreases when plants are nitrate-limited, because nitrate is an important inducer of Nia expression (see Section 3.3).
5.2.3 Ammonium assimilation
As discussed in Section 3.2.2, any discussion of ammonium assimilation is complicated by the massive recycling of ammonium during photorespiration. This is especially so in the case of elevated [CO2], when there will be a change in the rate of photorespiration. An increase of the external [CO2] from 0·035 to 0·07 Pa will approximately halve the rate of photorespiration (Stitt 1991), releasing substantial GS capacity for de novo assimilation of ammonium, especially since (see Section 2.7) overall GS activity does not decline in the leaves in elevated [CO2].
The decrease of photorespiration explains the substantial decreases in the pools of ammonium, glycine and serine in elevated [CO2] (Ferrario-Mery et al. 1997; Geiger et al. 1998b). This decrease is found, irrespective of whether the plants are receiving nitrate or ammonium nitrate (Geiger et al. 1999), and irrespective of whether they are nitrogen limited or nitrogen sufficient (Ferrario-Mery et al. 1997; Geiger et al. 1999). The decreased gross flux through the GOGAT pathway could be one of the reasons why elevated [CO2] led to a slower accumulation of glutamine during the photoperiod, even in well-fertilized plants (Geiger et al. 1998b). A decrease of glutamine could have considerable consequences for the regulation of metabolism and growth in elevated [CO2] because, as discussed in Section 3.10, glutamine is one of the compounds that is sensed to provide information about the nitrogen status. The possible consequences for the regulation of NR in elevated [CO2] was discussed in Section 5.2.2. Changes in the concentrations of ammonium and glutamine in the leaves might also lead to altered translocation of amino acids to the roots and affect the expression of transporters (see Section 3.1), and could have more far-reaching effects if glutamine turns out to play a central role in nitrogen signalling as it does in other organisms (see Section 3.10).
5.2.4 Amino acid synthesis
Elevated [CO2] only led to a small increase in the levels of the minor amino acids during the light period, but it led to a two-fold increase of almost all the minor amino acids in tobacco leaves at the end of the night, even though glutamate and aspartate were not significantly increased (Geiger et al. 1998b). This might be explained by the increased activity of NR activity during the night in elevated [CO2] (Section 5.2.1) and by the stimulatory effect of sugars on the synthesis of amino acids (Section 4.4). Further studies are needed to investigate the consequences for the rate of protein synthesis, and the synthesis of secondary metabolites that derive from various amino acids.
5.2.5 Nitrogen allocation
Changes in nitrogen allocation occur at two very different levels in plants: at a cellular level due to changes in the distribution of nitrogen between the various proteins and other nitrogen-containing compounds, and at a whole plant level due to changes in the allocation of nitrogen to different organs. Although the changes in nitrogen allocation in elevated [CO2] have been frequently described (see Sections 2.1, 2.4, also Section 5.3.1) and clearly depend on the nitrogen status of the cell and the plant (see Sections 2.1, 2.3, 3, 5.3.1), the molecular basis for these changes is not understood. At the cellular level, this will require a better understanding how the expression of individual genes and programmes of genes is regulated by sugar- and nitrogen-mediated signals (see Sections 3 and 4). At the whole plant level, it will depend upon identifying the processes that initiate and control the rate of growth of different sinks, and regulate their entry into senescence. The identification of a large family of amino acid transporters (Fischer et al. 1998; Rentsch, Boorer & Frommer 1998) that show differences in their expression between different tissues and cells and during development (Hirner et al. 1998) could provide a valuable tool to probe the mechanistic basis for changes in nitrogen allocation. The growing awareness of the interactions between nitrogen and sugar sensing and hormones (see Sections 3.10 and 4.5) opens a complementary route to investigate how allocation is regulated at the whole plant level (see below for further discussion).
5.3 Acclimation of photosynthesis
Several proximal explanations have been advanced for the acclimation of photosynthesis to elevated [CO2], including an inhibition due to damage caused by large starch grains, a direct limitation of photosynthesis due to an inadequate rate of end-product synthesis, and an inhibition due to a decline of Rubisco and other proteins involved in photosynthesis (reviewed in Stitt 1991). As discussed in Section 2.1, a decrease of Rubisco has been observed in very many studies of acclimation. This decrease could, however, have different causes, including sugar-mediated signalling, an interaction between sugar- and nitrogen-mediated signalling, or an indirect effect due to nitrogen limitation leading to earlier senescence. These will have very different functional implications for the response of plants to elevated [CO2], and the following section will discuss whether it is possible to distinguish between these various explanations for the data published to date.
5.3.1 A more complete data base will be required to evaluate the reasons for the decrease in Rubisco expression in elevated carbon dioxide
The decline of Rubisco activity or Rubisco protein in elevated [CO2] is accompanied by a decrease of the transcripts that encode RbcS and other transcripts encoding proteins involved in photosynthesis (see Section 2.1). Similar changes occur when sugars accumulate in leaves as a result of an inhibition of phloem transport in transgenic plants (Stitt 1991; Krapp et al. 1993; Riesmeier, Willmitzer & Frommer 1994; Lerchl et al. 1995; Geigenberger et al. 1996), petiole cooling (Krapp & Stitt 1995), or sugar feeding (Krapp, Quick & Stitt 1991; Paul & Stitt 1993; van Oosten & Besford 1995; Nielsen et al. 1998). This similarity prompted the hypothesis that the acclimation of photosynthesis in elevated [CO2] might be due to a sugar-mediated repression of RbcS and other genes that are required for photosynthesis (Stitt 1991).
Based on present models for sugar sensing in plants (but see Section 4.5 for a word of caution), it has been proposed that increased levels of hexose sugars lead, via hexokinase-related signalling, to repression of RbcS and other genes encoding proteins involved in photosynthesis, resulting in a decrease of Rubisco and the other encoded proteins and an inhibition of photosynthesis (Jang & Sheen 1994; Van Oosten et al. 1995). In this model, the rate of recycling of hexoses would be more important than the absolute amounts of sugars and, particularly, the amount of sucrose in the leaf. This would explain why the changes of the RbcS transcript level and Rubisco protein and activity do not correlate well with the absolute levels of sugars, either after adding exogenous sugars (see, e.g. Krapp et al. 1993) or during growth in elevated [CO2] (see Moore et al. 1999). Goldschmidt & Huber (1992) showed that there is a correlation between acid invertase activity and the susceptibility of photosynthesis to a ‘sink’-inhibition when sugars accumulate after girdling. Recently, Moore et al. (1998) showed in a study of 16 different species that elevated [CO2] leads to a decrease of Rubisco protein in 11 species where acid invertase activity exceeded a threshold value, but only in one of the five species with low invertase activity. This is an attractive model that would involve a specific sugar-sensing step in the acclimation to elevated [CO2]. However, there are still some serious deficiencies in the experimental support for this model, and it is unlikely to be the only explanation for photosynthetic acclimation.
Firstly, most of the experiments that reported a decrease of Rubisco during photosynthetic acclimation were probably carried out with plants where elevated [CO2] led to the appearance or the exacerbation of a nitrogen limitation (see Sections 2.3 and 5.1). In the absence of direct evidence that the plants did not become nitrogen-limited, it is difficult to distinguish whether the changes in Rubisco activity or protein and the changes of transcript in a particular experiment are triggered by sugar signalling, or by nitrogen limitation, or by an earlier onset of senescence as a consequence of nitrogen limitation, or other factors including self-shading or ontogenetic drift. For this reason, most of the experimental evidence for a decrease of the RbcS transcript and Rubisco activity in elevated [CO2] is irrelevant to the discussion of the above model. It is particularly worrying that elevated [CO2] did not lead to a decreased level of transcripts and the encoded proteins or enzymes in several studies where the nitrogen status was deliberately manipulated or monitored, and the plants were nitrogen sufficient (Riviere-Rolland et al. 1996; Habash et al. 1995; Atkinson et al. 1997; Ludewig et al. 1998; Geiger et al. 1999).
In a study of wheat growing in elevated [CO2] in a FACE system, Nie et al. (1995a) concluded that the decrease of Rubisco activity in wheat leaves in elevated [CO2] might have two phases: an early and small decrease which was independent of major changes in chlorophyll and other proteins including LHCII and the thylakoid ATP synthase, and a later and large decrease which was accompanied by a more general loss of protein and occurred in ambient [CO2] but was accentuated in elevated [CO2], and might reflect general leaf senescence which is speeded up in elevated [CO2]. In a companion study, Nie et al. (1995b) noted that elevated [CO2] resulted in a relatively large increase of sugars and a relatively small decrease of the levels of the transcripts for RbcS, RbcL and three other Calvin cycle enzymes in young plants, and that in older plants the difference became smaller with respect to sugar levels but larger with respect to the transcript levels, again indicating that further factors may be involved in the decrease of these transcripts in elevated [CO2] in the older plants. In Scirpus olneyi (Jacob et al. 1995) and another study on wheat (Sicher & Bunce 1997), the decrease of Rubisco and of total leaf protein were of the same order, indicating that the decrease was due to a non-specific decrease of leaf protein. In tobacco (Geiger et al. 1999), elevated [CO2] had no effect on the levels of the transcripts for RbcS, Cab, Rca (Rubisco activase) and NADP-glyceraldehyd-3-phosphate dehydrogenase or on the activity of Rubisco and six other Calvin cycle enzymes or on the overall leaf protein content in nitrogen-replete plants grown on 10 mM ammonium nitrate or 20 mM nitrate, whereas it led to a decrease of the levels of these four transcripts, a 20 to 30% decrease of Rubisco and six other Calvin cycle enzymes, and a similar decrease of total leaf protein when the plants were growing at a nitrogen supply (3 mM ammonium nitrate or 6 mM nitrate) that was just adequate to avoid nitrogen limitation in ambient [CO2] but became limiting in elevated [CO2].
Secondly, several studies have shown that although elevated [CO2] has a marked effect on Rubisco protein or activity in older leaves, it does not lead to a decrease of Rubisco in young leaves (Van Oosten et al. 1995; Nie et al. 1995a; Pearson & Brooks 1995; Miller et al. 1997). Miller et al. (1997) have suggested that elevated [CO2] leads to a shift in the timing of the normal photosynthetic stages of leaf ontogeny, resulting in an earlier onset of the photosynthetic decline.
Leaf senescence is a complex process that is regulated by ontogenetic factors, and by many environmental factors including the local light regime and the nitrogen supply (Gan & Amasino 1997). There is probably a complex interaction between the environmental factors and the endogenous developmental programme. This will make it difficult to distinguish, on the basis of whole plant parameters, between changes caused by a faster progression through the life cycle per se, and changes induced by changes in the nitrogen status or other external factors. This problem is underlined by the complexity of the sugar- and nitrogen-signalling systems (see Sections 3 and 4) and the emerging evidence that they may interact with hormonal- and phytochrome-signalling (see Sections 3.10 and 4.5). It will be important to examine the leaf age-dependence of changes in photosynthesis and Rubisco in parallel with an analysis of the nutrient status and of genes that are specifically induced during leaf senescence to show whether the development of the plant has been speeded up in elevated [CO2], or whether the earlier photosynthetic decline is triggered by nitrogen deficiency.
Thirdly, even if photosynthetic acclimation is not always just a consequence of earlier senescence or accelerated ontogeny, the glucose-sensing model still does not explain why the decrease of Rubisco is more pronounced in nitrogen-limited plants (see Section 2.1 and above). This discrepancy cannot be explained on the basis of species variability. In tobacco transformants with inhibited phloem function and in petiole girdled tobacco leaves, accumulation of sugars resulted in visual bleaching, a decrease of the RbcS transcript level, and a decrease of Rubisco, even though the plants were not nitrogen-limited (see above for references). In contrast, elevated [CO2] does not lead to a decrease of the RbcS transcript or Rubisco activity in nitrogen-sufficient tobacco, even though tobacco has high acid invertase activity (Moore et al. 1998) and there is a substantial accumulation of sucrose and hexoses in these conditions (Geiger et al. 1999). Elevated [CO2] does not lead to a decrease of Rubisco activity in well-fertilized wheat (Habash et al. 1995) and oak (Atkinson et al. 1997), even though the former also has high invertase activity (Moore et al. 1999) and both species show a decrease of Rubisco when less nitrogen is supplied (see, e.g. Nie et al. 1995b; George et al. 1996). It is therefore necessary to postulate either that signals derived from nitrogen metabolism modulate sugar sensing, or that there is cross-talk between the sugar- and nitrogen-signalling pathways.
Fourthly, Moore et al. (1998) found in their species comparison that the extent of the decrease of Rubisco protein was not correlated with the extent of the decrease of the RbcS transcript. Species with modest reductions in Rubisco (10–20%) often showed a marked reduction of the RbcS transcript level, whereas species with a larger reduction in Rubisco (> 30%) only showed a small decrease of the RbcS transcript level. This indicates that other factors may be leading to the decrease of Rubisco in these species that showed marked acclimation of photosynthesis.
These discrepancies might, of course, be due to difficulties in obtaining material at a comparable stage in the acclimation process from a wide range of species. Interpretation of transcript levels measured at a single time point may also be problematic, because transcript levels can vary diurnally and elevated [CO2] may modify these diurnal rhythms. Recently, Cheng et al. (1998) showed that elevated [CO2] indeed modifies the diurnal changes of the RbcS transcript level in Arabidopsis, having little effect on the amount of transcript during the day but leading to lower transcript levels during the night. In tobacco growing in low nitrogen, elevated [CO2] led to a decrease of the RbcS transcript level that was first apparent as a decrease at the end of the light period (Geiger et al. 1999). It should be noted, however, that these complications will also weaken arguments that are based on finding an apparent correlation between transcript levels that have been measured at a single time point and changes in Rubisco levels and acclimation.
5.3.2 End-product synthesis may sometimes colimit photosynthesis in plants growing in elevated carbon dioxide
Stitt (1991) concluded from a literature review of studies published up to that time that the accumulation of end products was not responsible for the acclimation of photosynthesis. However, as discussed already, many earlier studies are compromised by an interaction with nutrient deficiency. The role of a direct end-product limitation deserves re-examination in experiments where a possible interaction with nitrogen supply is explicitly addressed.
One proposed mechanism invokes an accumulation of starch grains leading to disruption of the plastids. In a recent study on Pinus palustris, Pritchard et al. (1997) showed that starch accumulation and plastid disruption is especially marked when elevated [CO2] is combined with low nitrogen. However, when potato transformants with antisense inhibition of AGPase expression where grown in elevated [CO2] at a high nitrogen supply they showed a more marked acclimation of photosynthesis than wild-type plants, showing that at least in this case, starch grains were not directly causing the inhibition of photosynthesis (Ludewig et al. 1998).
An alternative mechanism for a direct end-product inhibition of photosynthesis is based on experiments on isolated chloroplasts, and experiments on leaves subjected to various treatments to sequester phosphate (Walker & Sivak 1986). In this model, low rates of end-product synthesis are proposed to lead to an acute depletion of phosphate, a decrease of ATP, an inhibition of 3PGA reduction, an increase of 3PGA and deactivation of Rubisco (see Stitt 1991 for references). In a literature survey, Stitt (1991) concluded that there was no convincing evidence for a direct end-product limitation of photosynthesis when sugars accumulate in leaves during prolonged exposure to elevated [CO2], or as a result of other treatments that alter the sink–source balance. Numerous investigations have found that sugar accumulation in leaves, irrespective of whether it is the result of sugar-feeding or an inhibition of phloem transport or growth at elevated [CO2], is accompanied by an increase of ATP, a decrease of 3PGA and increased activation of Rubisco (see also Krapp & Stitt 1995; Morcuende et al. 1996, Morcuende, Pérez & Martinez-Carrasco 1997; Morcuende et al. 1998 for more recent publications). This is the opposite of the predicted response if the rate of phosphate recycling during sucrose and starch synthesis is colimiting for photosynthesis (see above). Further, as also pointed out by Sage (1994), the changes in the leaf gas exchange characteristics in plants showing acclimation following long-term exposure to elevated [CO2] are indicative of a decreased carboxylation capacity rather than an inadequate rate of end-product synthesis.
For two reasons, however, this question needs to be re-evaluated. First, the frequently observed changes in carboxylation capacity are probably at least partly due to changes in the nitrogen status (see Section 5.3.1). Second, some recent results indicate that when sugars accumulate in intact unstressed leaves, end-product limitation of photosynthesis might generate a different metabolite profile to that predicted on the basis of studies of isolated chloroplasts or the effects of phosphate-sequestering compounds on leaf metabolism. Morcuende et al. (1998) showed that sugars strongly stimulate nitrogen metabolism and organic acid synthesis (see Section 4.3), and the resulting stimulation of the respiratory pathways which will lead to a decrease of the levels of 3PGA and other glycolytic intermediates, and increased formation of ATP. There is a pressing need for a re-examination of the role of end-product limitation in plants growing in elevated [CO2] that are demonstrably not subject to changes in their nitrogen status. The growth temperature will need to be carefully chosen and might affect the outcome of the experiments, because the susceptibility of photosynthesis to inhibition by Pi recycling increases at lower temperature (Stitt 1991, 1996).
Two subsequent studies of gas exchange characteristics have shown that end-product limitation indeed makes a contribution to the control of photosynthesis in elevated [CO2] in some conditions and species. The oxygen-sensitivity of photosynthesis (an indicator for a limitation of photosynthesis by the rate of Pi recycling in sucrose and starch synthesis, see Stitt 1991 for further discussion) was decreased when bean (Socias, Medrano & Sharkey 1993) and tobacco (Micallef et al. 1995) were grown at elevated [CO2]. Acclimation of photosynthesis in elevated [CO2] in bean was associated with a decrease of SPS activity and decreased carbamylation of Rubisco, whereas Rubisco protein was not reduced (Socias et al. 1993).
Studies of three different transgenic plants with altered expression of enzymes required for sucrose or starch synthesis have also provided evidence that inadequate rates of end-product synthesis can contribute to acclimation. First, tomato transformants with constitutive overexpression of maize SPS and an increased capacity for sucrose synthesis did not show a loss of the oxygen-sensitivity of photosynthesis when they were grown in elevated [CO2], and did not show an acclimation of photosynthesis to elevated [CO2] (Micallef et al. 1995). One transformant line showed increased biomass in elevated [CO2], compared to wild-type plants (Leport et al. 1996). Galtier et al. (1995) also found that overexpression of maize SPS led to a larger stimulation of photosynthesis during a short-term exposure to elevated [CO2]. Second, potato transformants with decreased expression of AGPase and, hence, a decreased capacity for starch synthesis showed a large and faster accclimation to elevated [CO2] than wild-type plants (Ludewig et al. 1998). Significantly, acclimation was not accompanied by a decrease of the RcbS or Cab transcript levels, or Rubisco activity. Third, antisense inhibition of plastid aldolase expression led to a stronger inhibition of photosynthesis in plants growing in elevated [CO2] than in plants growing in ambient [CO2]. This inhibition was accompanied by a dramatic decrease of starch, an accumulation of sugars, and a high 3PGA:triose phosphate ratio, and the inhibition was not accompanied by a preferential decrease of the transcripts for RbcS and other Calvin cycle enzymes, nor was there a preferential decrease of the corresponding enzyme activities (V. Haake, M. Geiger, R. Zrenner, P. Liu-Walsch, C. Engels and M. Stitt, unpublished).
Summarizing, although there are several potential explanations for the acclimation of photosynthesis, few of the published experiments provide a large enough data-base to be able to evaluate their contribution. There is evidence in isolated cases indicating that selective regulation of Rubisco or low rates of end-product synthesis contribute to the acclimation of photosynthesis. Probably, by default, many of the cases where acclimation has been observed have included a contribution from earlier senescence because the plants were not adequately fertilized. In future experiments it will be important to pay careful attention to the design of the experiments in order control the nitrogen supply at a suitable level, to routinely include analyses not only of carbohydrates but also of nitrogen metabolism (Geiger et al. 1998b, 1999), and to monitor the expression and protein levels of a much wider range of proteins involved in photosynthesis (Nie et al. 1995a, 1995b; Geiger et al. 1999) and housekeeping proteins, in order to allow a clearer interpretation of the significance of changes in single protein species. It will also be important to address the possible contribution of diurnal rhythms in transcripts in leaves (Cheng et al. 1998; Geiger et al. 1998b), to analyse the response at different stages through the life-history of a single leaf (Miller et al. 1997) and of a plant (Nie et al. 1995a, 1995b), and to combine the analyses of events in primary metabolism with an analysis of events related to leaf ageing and senescence. Transgenic plants with altered expression of key steps underlying the different explanations for acclimation including enzymic steps in photosynthesis (see above) and processes leading to senescence (Gan & Amasino 1997), and mutants altered in sugar or nitrogen signalling could provide an invaluable resource to dissect these complex interactions. An unambiguous resolution of this problem will ultimately depend on a better molecular understanding of sugar signalling, nitrogen signalling and the regulation of leaf ageing and senescence.
5.4 Photosynthate partitioning
Elevated [CO2] leads to increased levels of carbohydrates in the leaf, with the major form being starch (Section 2.2). It has been tacitly assumed that the accumulation of starch is a response to the accumulation of sugars. Possible mechanisms could include a feed-back inhibition of sucrose synthesis or recycling of sucrose via invertase (see above) leading to an accumulation of phosphorylated metabolites and stimulation of AGPase by an increased 3PGA:phosphate ratio (for more details see Stitt 1991), and increased expression of AgpS as a result of sugar-mediated signalling (Müller-Röber et al. 1990; Krapp & Stitt 1995; Nielsen et al. 1998).
Several studies, however, have found a marked increase of starch in elevated [CO2] even though sugar levels do not change significantly (Morin et al. 1992; Van Der Kooij & De Kok 1996; Ludewig et al. 1998; Geiger et al. 1999) As discussed in Section 3.7, AgpS expression is regulated by signals from nitrate and phosphate metabolism. The decrease of nitrate (see Sections 2.3 and 5.1) in elevated [CO2] could be primarily responsible for the dramatic increase of the AgpS transcript in elevated [CO2] (Geiger et al. 1999), and contribute to the increase in starch accumulation. This might also explain why the increase of starch in elevated [CO2] is especially marked in nitrogen-limited plants (see Section 2.2) and why sugars are present in higher levels in well-fertilized plants than in nitrogen-limited plants in ambient (Fichtner et al. 1993) and elevated (Ferrario-Mery et al. 1997; Geiger et al. 1999) [CO2].
Any stimulation of starch accumulation and reduction of sugar formation due to signals originating from nitrate or other nutrients will affect the ability of the plant to initiate changes in allocation that require sugar signalling. This could result in some extremely complex interactions when elevated [CO2] is provided to plants at a nutrient supply where it results in large changes in the key pools of nutrients and the metabolites formed from them. Careful analyses of nitrogen metabolism in combination with changes in allocation will be required to unravel these complex interactions, and it will probably be useful to employ genotypes with altered expression of selected genes involved in nitrogen metabolism, and to use sophisticated methods to monitor and control the supply of nutrients to the plant, including the Ingestad method (Ingestad 1997) in which the nutrient supply is increased in parallel with the absolute rate of plant growth (see Pettersson et al. 1993; Pettersson & MacDonald 1994 for applications to plants growing in elevated [CO2]).
5.5 Whole plant growth
The effect of elevated [CO2] on growth is strongly modulated by the nitrogen supply (Section 2.3). In this final section, we summarize some of the changes in photosynthesis, carbon allocation and nitrogen allocation that underlie this dramatic modulation of the response to elevated [CO2].
5.5.1 Growth rate
Elevated [CO2] leads to a sustained stimulation of photosynthesis, elevated levels of sugars, and supports increased growth in well-fertilized plants, but not in nitrogen-deficient plants (Section 2). Two interacting factors explain why elevated [CO2] leads to a large increase of biomass in well-fertilized tobacco, but has little or no effect on the biomass of nitrogen-limited tobacco. First, as shown in studies of tobacco transformants with decreased expression of Rubisco, alterations in the rate of photosynthesis have a large effect on the rate of growth of well-fertilized plants but have little or no effect on the rate of growth of nitrogen-limited plants (Quick et al. 1993; Fichtner et al. 1993; Stitt & Schulze 1994). Second (Sections 2.1 and 5.3.1) elevated [CO2] produces a sustained stimulation of photosynthesis and of sugar formation in well-fertilized plants, but not in nitrogen-limited plants.
When nitrogen-replete plants are grown in elevated [CO2], the total and individual amino acids and protein remain high or even increase slightly, and the activities of Rubisco and other Calvin cycle remain high or increase slightly (Section 5.3.1). This allows an increased rate of photosynthesis to be maintained, compared to plants growing in ambient [CO2]. The AgpS transcript remains relatively low in elevated [CO2], and starch only increases slightly compared to plants growing in ambient [CO2] (Section 5.4). The increased rate of net photosynthesis and the relatively low rates of starch accumulation result in higher levels of sugars in the leaves, and make more photosynthate available for export to support growth. Although it is possible that the sugars also act as signals to promote growth (Section 4, Koch 1996; see also discussion in Micallef et al. 1995), they do not act as an effective signal to inhibit the expression of Rubisco or other proteins involved in photosynthesis in these conditions (Section 5.3.1).
When plants with an intermediate or a limiting nitrogen supply are grown in elevated [CO2], NR activity decreases (Section 5.2.2), nitrate, amino acids and total protein decrease (Section 5.1), the activities of Rubisco and other Calvin cycle enzymes decrease (Section 5.3.1), and the rate of photosynthesis in elevated [CO2] is only slightly increased or even decreases compared to plants growing in ambient [CO2] (Section 2.1). Expression of AgpS is increased (Section 5.3.2), and larger amounts of carbon accumulate as starch in elevated [CO2] (Section 2.2). The additional starch that accumulates in elevated [CO2] in nitrogen-limited plants will represent a substantial portion of the net photosynthesis in freshly expanded leaves (Fichtner et al. 1993). As a result, the amount of photosynthate that is available for export to support growth will not be significantly increased, and may even decrease. This, together with possible effects of decreased transpiration on the availability of nitrate in the soil (Section 2.6) may explain why elevated [CO2] sometimes even leads to a slight decrease of biomass in severely nitrogen-limited plants (Section 2.3).
Ultimately the rate of growth will depend upon the number of active meristems, and rates of cell division and cell elongation, and the duration of the division and elongation phases at each individual meristem. Elevated [CO2] results in a shortening of the cell cycle in the shoot and root meristem (Kinsman et al. 1997). It will be important to determine to what extent this is solely due to an increased supply of sugars to the meristem, and to what extent sugar- or nutrient-mediated signalling modulates the activity at the meristems. In this context, it might be noted that the modulation of root meristem growth by internal and external nitrate does not involve changes in the levels of amino acids in the roots (see Section 3.9). It is more likely that the carbon and nutritional status of the plant is sensed and transduced via hormones or other signalling molecules.
5.5.2 Elevated carbon dioxide has an age-dependent effect on the rate of plant growth
Any consideration of the effect of elevated [CO2] on the physiology and growth of plants must, however, also address time-dependent changes in the growth rate of plants. In general, growth is rapid in young plants and then decreases as the plant ages (see, e.g. Masle, Hudson & Badger 1993; Geiger et al. 1998b). The effect of elevated [CO2] on growth also depends on plant age. Studies with alfalfa (Baysdorfer & Bassham 1985), five different wild annual species (Garbutt et al. 1990), Poa alpina (Baxter et al. 1994), Arabidopsis (Van de Kooij & de Kok 1996) and tobacco (Geiger et al. 1998b) have shown that long-term exposure to elevated [CO2] leads to a stimulation of the current relative growth rate in young plants, whereas the relative growth rate in older plants is unaffected. Several studies of trees have also reported that the increase in biomass is largely due to increased rates of growth in the first year of exposure to elevated [CO2], and that growth is stimulated less or is not increased at all in the subsequent years (Cuelemans, Jiang & Shao 1994; Lee & Jarvis 1995; Tissue et al. 1997).
Young seedlings are typically ‘source’-limited, and contain far lower levels of sugars than older plants (Geiger et al. 1998b). This provides one simple explanation why elevated [CO2] leads to a larger stimulation of metabolism and growth in seedlings than in older plants. The fall off of the growth rate in elevated [CO2] in older plants may often be partly due to an earlier exhaustion of the nutrient supply. This problem can, of course, be counteracted by supplying higher amounts of nutrient, for example by growing the plants in hydroponic culture with controlled nitrogen levels, or by increasing the nutrient supply in a parallel with the plant size (see above). However, the gradual fall-off in the growth rate of plants is probably also partly due to endogenous ontogenetic factors. When tobacco plants were grown in hydroponic culture at an excess nitrogen supply, the growth stimulation by elevated [CO2] was observed for a longer time than in sand or soil culture, but the stimulation nevertheless still disappeared in older plants (M. Geiger, P. Walsch-Liu, C. Engels and M. Stitt, unpublished). Since elevated [CO2] allows plants, especially those with determinate life histories, to move through their life history at a faster rate it will lead to an earlier onset of an ontogenetic decline in growth rates, to which several factors may be contributing. A a plant becomes larger, an increasing part of the biomass is committed to stems, older roots and other support structures, and a decreasing part to leaves and young roots which are responsible for the acquisition and assimilation of carbon dioxide and nutrients. Self-shading of leaves, or physical restriction of root growth (Arp 1991), can lead to a further decrease in growth rates. Especially in determinate plants, the number of growth points at active meristems may not keep pace with the production of assimilatory structures.
Seedlings grow rapidly and exponentially, and over a period of days a relatively small 20% increase in the relative growth rate can result in a doubling of the plant biomass (Geiger et al. 1998b). The marked increase of biomass of older plants in elevated [CO2] therefore reflects a small increase in the rate of growth at an earlier period of their life history. This age-dependent effect of elevated [CO2] on the rate of growth has enormous implications for the design and interpretation of experiments that investigate the effect of elevated [CO2] on metabolic and physiological processes (see also Körner 1995 for a discussion of their impact on up-scaling to ecosystem responses). Firstly, even though plants show striking differences in their biomass, the underlying changes in the rate of growth and hence in molecular and physiological processes may be quite small, and difficult to detect. Secondly, in order to interpret the significance of changes in physiological and molecular parameters in elevated [CO2], it is essential to relate them to the current growth rate of the plant. Many comparisons of photosynthesis, allocation and nitrogen metabolism in older plants in ambient and elevated [CO2] may have been studying two sets of plants which have a similar current growth rate but have exhausted the available nutrient supply to a varying extent and/or are at a different stage in their life history. This, for example, may be another reason for the apparent anomaly that plants in elevated [CO2] have lower NR activity and lower levels of nitrogen, amino acids and protein (see Section 2.4) but are apparently growing faster than plants in ambient [CO2].
There are important endogenous changes in regulation during development, that could be modified by elevated [CO2], leading to further complex interactions. For example, shoot morphogenesis is typically characterized by three developmental phases, juvenile, adult and reproductive (Poethig 1990). Tsai et al. (1997) used antisense RbcS tobacco transformants as a experimental system to decrease the rate of photosynthesis and the levels of sugars, and observed that this resulted in the production of a larger number of leaves in the juvenile phase. This result implies that signals related to the levels of sugars may modulate the transition from the juvenile to the adult phase (Tsai et al. 1997). It will be important to investigate whether this shift in ontogeny occurs earlier in elevated [CO2], and discover whether it affects the subsequent ability to maintain increased rates of photosynthesis and growth in elevated [CO2]. It will also be important to investigate whether this developmental shift is solely due to changes in the levels of carbohydrates, or if changes in other nutrients including nitrogen metabolites are also involved. For these reasons, it may be misleading to apply an oversimplified schemata when considering the contribution of ontogenetic drift to the response of tissue composition of growth to elevated [CO2].
5.5.3 Changes in the allocation of biomass in elevated carbon dioxide may sometimes be mediated by changes in nutrient status
Elevated [CO2] leads to changes in biomass allocation and phenology which can include an increase of the root:shoot ratio, tillering, increased branching, and earlier flowering (Bazzaz & Fajer 1992; Tissue et al. 1997).
Some of these changes may be mediated via changes in the nutrient status. For example, signals from nitrate regulate root:shoot allocation in tobacco and Arabidopsis (Section 3.9). Elevated [CO2] typically leads to a decrease of nitrate (Section 2.4) especially if the nitrogen supply is not super-optimal (Section 5.1), and this or other nitrogen-mediated signals could be responsible for the change in shoot : root allocation often reported in elevated [CO2]. When N. plumbaginifolia was grown hydroponically with a high nitrate supply, elevated [CO2] did not affect the root : shoot ratio (Ferrario-Mery et al. 1997). Elevated [CO2] led to a marked increase of the root : shoot ratio in nitrogen-limited Poa alpina, whereas the ratio decreased in well-fertilized plants (Baxter et al. 1997). Elevated [CO2] also leads to an increase of the root : shoot ratio in nitrogen-limited tobacco, but not in well-fertilized tobacco (Geiger et al. 1999). When the rate of photosynthesis was altered by inhibiting expression of Rubisco, root : shoot allocation was only altered compared to wild-type plants when the plants were grown in high nitrate (Fichtner et al. 1993), when the lower rate of photosynthesis in the transformants led to a marked increase of nitrate in the plants (Stitt & Schulze 1994). The increase of tillering (Baxter et al. 1997), branching (Tissue et al. 1997), tuber initiation (see Heineke and Sonnewald, 1999) and earlier flowering (Bazzaz & Fajer 1992) in elevated [CO2] may also not be exclusively due to the increased carbohydrate levels, but involve an interaction with nitrogen.
In conclusion, it is essential to investigate the changes in nitrogen metabolism in parallel with investigations of photosynthesis, carbon allocation, and growth in elevated [CO2]. Unless the effect of elevated [CO2] on the plant nutrient status is known, there is a serious risk that the changes in other parameters will be misinterpreted. A detailed analysis of nitrogen metabolism is needed in order to allow the contributions of sugar-mediated signalling, nitrogen-induced senescence, and ontogenetic drift to be disentangled, and it is likely that some of the responses observed under elevated [CO2] will be due to more specific changes initiated as a result of signals emanating from the changes in nitrate levels and in nitrogen metabolism in the plant.