1

3 Synthesis

3.1 N storage and remobilization

N storage and remobilization for spring growth in R. acetosa, A. capillaris and F. rubra

  • internal N sources contribute substantially to growth in spring and summer in all three species studied. At the initial harvest in spring, last year's N contributed to the total N in the new aboveground growth between 82% and 56%, depending on the species (see Fi. 21). significant differences exist between the three species for this contribution at any given time in spring, with R. acetosa consistently having the lowest and A. capillaris the highest contribution of remobilized N to the total N in the new aboveground growth.

Fig. 21: Contribution of remobilized N to the total N in the new aboveground parts of the three study species at the initial spring harvest and the last shared harvest in summer.

  • marked differences exist between the dicot R. acetosa and the grasses with respect to where N was remobilized from:

(1)R. acetosa remobilized mainly from stores belowground, whereas the grasses remobilized N from aboveground.

(2)R. acetosa invested in the construction and the filling of a specialised storage structure, from which most remobilized N originated from, whereas both grasses recycled N out of the overwintering leaves. There was no increase in %N in overwintering leaves over winter, indicating that no accumulation of N in excess of the immediate requirement for growth or maintenance had occurred.

  • Reproductive growth resulted in additional remobilization form fine roots and leaves in the case of R. acetosa, or from vegetative tillers in case of the grasses.
  • R. acetosa did not use a VSP for N storage in its taproot; both a general protein pool and free amino acids (primarily arginine) were degraded in spring. Storage compounds were not determined for the grasses.

N remobilization during different seasons and with different N supply

  • For all harvest intervals (seasons) studied, the internal N recovered in new leaves was N remobilized from old leaves.
  • The contribution of remobilized N to the total N in the new leaves was very similar for plants with both high and low N status. Only at H2 were measurements greatly lower for HN plants than for LN plants, and this was due to the fact that the current uptake of N (in relation to the plant N capital) was higher for HN plants.
  • The remobilization rate from old leaves of the grasses (as mg N g plant N-1 day-1) showed no consistent pattern for treatment or species. The remobilization rate was highest in spring. The proportion of initial old-leaf N, which was remobilized during an interval, was slightly higher for LN than for HN plants.

In both experiments, significant differences were observed for both species and genotypes with respect to the amount remobilized N contributed to the N within new tissues. Such differences in the contribution of remobilized N to the total N in the new aboveground growth could be due to a number of reasons. (1) Plants could differ in the amount of N available from storage. (2) They could have a different time course of N remobilization. (3) Differences could be due to differences in the timing of growth. As shown in Section 2.3 differences in the contribution of remobilized N at a given time between F. rubra and A. capillaris become insignificant when expressed in relation to the biomass produced. Furthermore, in the second experiment, where no differentiation in timing was observed for F. rubra and A. capillaris, no major differences were seen between the species with respect to this contribution. (4) Plants might differ in their allocation of remobilized N (root versus shoot). (5) Plants could differ in the amount of N taken up during the remobilization period.

The processes of N remobilization and uptake are not necessarily independent from each other, as N uptake is regulated at the whole plant level (Imsande & Touraine 1994). To what extent N remobilization from overwintering stores is dependent on the external N supply is poorly known for herbaceous plants. For trees the amount of N remobilized is dependent on the amount of N in store (Millard 1996). This behaviour was also seen for Cichorium intybus (witloof chicory), with the amount of N remobilized similar irrespective of the nitrate concentration in the nutrient solution during sprouting (Limami et al. 1996). On the other hand, the total amount of N remobilized from basal internodes and roots of M. caerulea during spring growth was higher when plants were defoliated as opposed to undefoliated (Thornton & Millard 1993). My own observations (unpublished) on M. caerulea have shown that N remobilization is dependent on the amount of N available during spring growth: plants being deprived of N withdrew twice as much N from storage pools in the basal internodes as plants receiving N. Apparently, N remobilisation from overwintering stores is - in at least some herbaceous perennials – more flexible than in trees.

The interdependence of internal and external N sources also works the other way around: the size of the N store can influence N uptake. For Bistorta bistortoides, N fertilization increased the amount of N in the rhizomes, making plants completely independent from N from current uptake during the second year of growth. Unfertilized plants achieved the same aboveground growth but derived nearly half the N required for it from uptake (Lipson et al. 1996). The lack of whole plant N data for the grasses makes comparisons among species with respect to total plant uptake and remobilization difficult. Estimates suggest that the grasses acquired more N during the first year of growth than R. acetosa, and that N uptake during the second year - in relation to last year's N - was higher for R. acetosa. The very different storage strategies (specialised storage organ versus leaves) between the growth forms can potentially have implications for N uptake. Grass tillers are normally relatively constant in the number of living leaves they have (Hunt & Brougham 1966; Jones et al. 1982), which means that the production of new leaves is balanced with the death of old leaves. Therefore N becomes available from leaf turnover continuously. The availability of these internal N sources can, in form of the proteolysis-products amino acids, possibly prevent uptake of external N when the demand of the plant is low, and could explain the lower contribution of uptake N in the new leaves of the grasses. It is known that amino acids delivered through the phloem can inhibit nitrate uptake (Muller et al. 1994), and, as demonstrated in split-root experiments, the expression of the high affinity transporter for nitrate is regulated via shoot-to-root signalling of N demand (Gansel et al. 2001). The situation for ammonium is less clear, but for the high affinity transporter a control via glutamine (via gene expression) has been reported (Rawat et al. 1999). There is some indication for an interdependence for remobilization and uptake in the second experiment: the contribution of remobilized N to the total N in new leaves was higher in spring than in winter (the number of new leaves produced during the labelling period were similar), comcomitant with an increase in the remobilization rate, while the uptake rate was either constant or even declining. This, however, could have been an experimental artifact.

A specialised storage organ such as the taproot of R. acetosa could act as a sink for external N even at times of low plant demand for construction of new tissues, and could result in a more efficient exploitation of nutrient pulses and patches. These differences might be important for the N balance of environments such as grazed grasslands where N inputs through e.g. urine deposition can severely exceed plant demand, and would therefore be susceptible to loss from the system. In N-poor grasslands, dicot species increase at the expense of the abundance of grasses, and while certainly not all dicots have a specialised storage structure, this could be one reason for their success under these conditions.

The second experiment has shown that, if the grass plants were grown continuously in either LN or HN, and therefore differed markedly in their N status, the differences in the contribution of internal N to the total N in the new leaves due to treatment were small. In fact, this contribution was the least plastic trait. Again, it was remobilization from old, senescing leaves, which provided the unlabelled N in the new leaves, without major changes due to the season. The apparent absence of reserves (in the sense of reserve formation, Chapin et al. (1990)) in grasses means that the extent of N remobilization is dependent on (1) the turnover rate of the leaves and (2) the resorption efficiency for nutrients during senescence. It is believed that in N-limiting situations, plants from nutrient-poor environments are successfully outcompeting plants from nutrient-rich environments due to their longer tissue life spans, and therefore, smaller N losses through litter (Chapin 1980; Grime 1979). The longer life span of leaves is correlated with higher tissue density and higher investment in secondary compounds (Ryser 1996). This could result in smaller resorption efficiency. These two factors combined would suggest that the amount of remobilized N within a specific time period could actually be smaller for plants from nutrient-poor environments than for plants from nutrient-rich habitats, given that uptake and the N capital in the living leaves are the same. Based on their work with four grass species subjected to defoliation to a constant height, Thornton et al. (1996) observed that the contribution of remobilized N in the regrowth was inverse to the order in which species ingress in a L. perenne sward. These authors suggested that late-successional species allocated more N to storage in roots and stubble, and that these reserves were responsible for the higher contribution of internal sources in the new leaves. An alternative explanation for this behaviour is that tall growing early-successional species lost relatively more N through defoliation than the smaller late-successional species, and that therefore species effects were related to different N pool sizes removed from the plant. No comparative data are available for the importance of remobilization in undefoliated grasses. The data from this work suggests that species differences are small for coexisting grasses. It is surprising that so much work has been carried out on remobilization after defoliation, without remobilization during undisturbed growth ever having been quantified. This, however, is really the measurement post-defoliation remobilization should be compared with to see whether the rates of remobilization of internal N are actually altered after defoliation. In the context of defoliation, N remobilization is throughout the literature referred to as a response to defoliation. If N remobilization however is a process contributing substantially to the growth of new leaves also under undefoliated conditions, defoliation responses would be changes in this contribution, changes in the site of where N remobilization occurs from, changes in the rate of remobilization and changes in the enzyme complement involved.

Experiments like the ones described here cannot be used to assess the importance of remobilization under field conditions, where an array of factors such as competition and natural variations of temperatures and nutrients acts on the plants. The presence of grazing animals during winter could result in the loss of an important storage pool for the grasses, but affect R. acetosa to a lesser extent. Belowground herbivory, which can be significant especially in old, long-established pastures (Brown & Gange 1990) could make storage in specialised storage organs risky unless (costly?) chemical and mechanical defenses are in place. Furthermore, competitive interactions among the species for both below- and aboveground resources can alter any behaviour observed in individually grown plants. Here speed of resource capture or capability for the exploitation of nutrient patches might be important. In a broad comparison of species derived from various habitats, Robinson & Van Vuuren (1998) have shown that dicots are better able to exploit local nutrient patches than grasses. Whether this holds true for the species under investigation here is not known, but would be worth exploring.

3.2 15N signatures in genotypes of A. capillaris and F. rubra, and their potential as a measure for N remobilization

Plant 15N signatures depend on both the 15N value of the N source and processes within the plant. The experiments carried out here were aimed at determining the pattern and magnitude of the latter. It was found that plant 15N signatures, measured on plants grown under controlled conditions with a N source of uniform 15N, showed a high degree of variation in response to environmental conditions and genotypic factors. These were:

  • Plants behaved differently with different forms of N given: (1) whole plant 15N was under genotypic control if plants were grown with nitrate, but, if grown with ammonium, neither genotypic nor species-specific differences were seen. (Unfortunately, the A. capillaris genotypes were not identical in the two experiments). (2) Shoots and roots within a plant were not differently enriched when grown with nitrate but were significantly different when grown with ammonium.
  • Plants behaved differently with different amounts of N given: (1) Plants grown with LN (as ammonium) had a lower shoot than root 15N value whereas the opposite was seen for plants grown with HN. The same was true for the comparison of youngest leaf with the remainder of the living leaves. (2) Differences among different plant parts in 15N were greater under LN than HN.
  • 15N signatures of the living shoot material varied in time, despite any changes in the N source.
  • Even mild stresses resulted in significant, genotype-specific changes in 15N.
  • 15N signatures (whole plant, plant parts) were frequently subject to interactions of environment with genotype or species.

These factors will be important when considering the potential of measurements of 15N signatures of plants grown under natural conditions in the field. An underlying assumption of many field measurements is that 15N values measured in the plant are a representation of the signature of N taken up by the plant. Different plant signatures are supposedly indicative of different N sources or different routes of N acquisition (e.g. Hobbie et al. 2000; Michelsen et al. 1996). The experiments presented here have clearly shown that genotypic factors, against a background of identical N source 15N, can result in a variation of 15N which can match the magnitude of variation observed in plants measured in the field, where both genetic and environmental factors act together (Table 18).

Table 18: The range of 15N signatures of A. capillaris and F. rubra genotypes in comparison with the range found in other plant species under field conditions.

Plant type

/ Environment / 15N range in foliage / Reference
Agrostis capillaris (five genotypes) / Controlled experiment with nitrate as N source / 2.2 / Section 2.4
Agrostis capillaris (five genotypes) / Controlled experiment with ammonium as N source (LN) / 1.5 / Section 2.5
Festuca rubra (five genotypes) / Controlled experiment with ammonium as N source (LN) / 0.6 / Section 2.5
Agrostis capillaris / Scottish upland pasture / 1.7 / Marriott et al., unpublished
Savanna grasses (six species) / Savanna in Cote d’Ivoire / 1.7 / Abbadie et al. (1992)
Picea abies / German spruce forest / 1.6 / Gebauer & Schulze (1991)

Pinus sylvestris

/ Swedish pine forest / 4.4 / Högberg & Johannisson (1993)
Rainforest trees (14 species) / Amazonian rainforest / 2.7 / Guehl et al. (1998)
Taiga trees and shrubs (six species) / Alaskan taiga forest (three forest types) / 9.0 / Kielland et al. (1998)
Mulga vegetation (~100 species of different life forms) / Eastern and Western Australian mulga woodlands / 8.0 / Pate et al. (1998)

Environmental influences and their interaction with genotypes will further complicate interpretation of 15N values. However, it is not only the (so far) unpredictable extent to which plant 15N will deviate from its source, which poses a problem for meaningful field measurements, but also the difficulty in determining 15N of the soil N pools. Due to discrimination during microbial transformations of soil N compounds (Mariotti et al. 1981; Yoshida 1988), variations in the 15N values of different forms of N available for uptake by the plant are likely. It is also known that soil 15N can vary seasonally (Neilson et al. 1998; Handley & Scrimgeour 1997). It is now possible to isolate nitrate from the soil solution to determine its 15N value (Johnston et al. 1999), but satisfactory methods for ammonium and the various forms of organic N are lacking. Furthermore, the 15N of N left in the soil is not necessarily identical to the 15N of N taken up by the plant.

Another problem with field measurements will be associated with the sampling strategy. For practicality, whole plant sampling is often not feasible, and leaves are the most convenient samples. However, as shown here, 15N of leaves can be different from roots, and also among leaf fractions of different ages significant differences exist. As N nutrition was able to influence the pattern of intra-plant enrichments, no reliable predictions can be made about the relationships of the 15N of a sampled leaf to the rest of the plant.

The parallel experiment with 15N-enriched N and N at natural abundance has shown that the 15N signature of remobilized N was close to, and at least initially slightly depleted, to the 15N of its source tissue. This was consistent for both N treatments. However, due to the variability described above and the uncertainty of 15N signatures of the soil solution components, the initial aim of using 15N signatures as a measure for N remobilization seems unrealisable. Nevertheless, plant 15N values appear to be very sensitive signals of different routes of N metabolism among genotypes or in response to N treatment, and for that reason deserve further attention. More lab-based experiments are needed for to be able to understand and interprete these signals. One of the drawbacks of my experiments were that plants were not grown under sterile conditions, therefore N transformations in or on top of the sand due to microbial activity cannot be excluded. This could impact on the 15N of the N source, making any statements on whole plant discrimination invaluable. Furthermore, documentation of all forms of N losses and their 15N signatures is required in order to explain whole plant discriminations, and the genotypic and environmental differences between them. Ultimately, a better knowledge of the physiology of the plant is essential, with focus on the genotypic and phenotypic variability of site of N assimilation, N transport compounds and internal N cycling.