¹³C and ¹⁵N allocations of two alpine species from early and late snowmelt locations reflect their different growth strategies

Abstract

Intense efforts are currently devoted to disentangling the relationships between plant carbon (C) allocation patterns and soil nitrogen (N) availability because of their consequences for growth and more generally for C sequestration. In cold ecosystems, only a few studies have addressed whole-plant C and/or N allocation along natural elevational or topographical gradients. ¹²C/¹³C and ¹⁴N/¹⁵N isotope techniques have been used to elucidate C and N partitioning in two alpine graminoids characterized by contrasted nutrient economies: a slow-growing species, Kobresia myosuroides (KM), and a fast-growing species, Carex foetida (CF), located in early and late snowmelt habitats, respectively, within the alpine tundra (French Alps). CF allocated higher labelling-related ¹³C content belowground and produced more root biomass. Furthermore, assimilates transferred to the roots were preferentially used for growth rather than respiration and tended to favour N reduction in this compartment. Accordingly, this species had higher ¹⁵N uptake efficiency than KM and a higher translocation of reduced ¹⁵N to aboveground organs. These results suggest that at the whole-plant level, there is a compromise between N acquisition/reduction and C allocation patterns for optimized growth.

Introduction

Plant carbon assimilation, allocation, and yield are strongly affected by variations in soil fertility and nitrogen balance. Therefore, intense efforts are currently devoted to elucidating the relationships between plant carbon allocation patterns and nitrogen availability in order to improve our understanding of carbon cycles and sequestration pathways in terrestrial ecosystems (Waring, 1993; Farrar and Gunn, 1998; Trumbore, 2006). This is particularly important in the context of global climate change, in which carbon fixation and growth tend to be promoted by increased atmospheric CO₂ but depend on the availability of N (or other nutrients) (Long et al., 2006).

Alpine ecosystems are particularly sensitive to such ecological effects because (i) alpine habitats are strongly N-limited in many instances (Bowman et al., 1993); (ii) snow cover gradients influence the nature of plant communities (Körner, 1999; Choler, 2005); and (iii) the composition of alpine vegetation varies in response to deposition of anthropogenic nitrogen (Bowman et al., 2006). However, the physiology of carbon allocation patterns of alpine plants in relation to N uptake is not well understood.

In fact, ecophysiological studies of alpine plants so far have mostly focused on carbon uptake (Körner, 1982; Körner and Diemer, 1987) or on the ecology of storage (Jaeger and Monson, 1992; Lipson et al., 1996; Monson et al., 2006). Only Körner (1999) presented a whole-plant C allocation of a dominant graminoid in alpine tundra, Carex curvula, after long-term ¹³C labelling. Other authors have examined the constraints of nutrient availability on leaf physiological properties and plant production of various alpine communities (Bowman et al., 1993, 1995; Theodose and Bowman, 1997), but they did not disentangle the C and N fluxes sustaining plant growth.

Furthermore, alpine plants are generally exposed to contrasting microenvironmental conditions that vary along the snow cover gradient (nutrients and water availability, temperature, length of the favourable period for carbon uptake), adding complexity to variations in the C and N balance. In late snowmelt locations, the growing season is shorter and accompanied by relatively high nutrient availability (particularly at the beginning of the growing season) (Bowman, 1992; Baptist and Choler, 2008). Plants growing under such conditions are thus energy limited rather than nutrient limited. In contrast, plants growing in early snowmelt habitats are much less energy limited but soil nutrients are less abundant. Contrasting growth rates (fast- and slow-growing species are observed in late and early snowmelt conditions, respectively) and leaf functional diversity related to nutrient economy (which is higher in fast-growing species, Wright et al., 2004) have been described along snow cover gradients (Kudo, 1996; Choler, 2005). However, to what extent the dynamic allocation of C and N is associated with these contrasting strategies remains unknown.

It is well known that carbon fixation and growth are promoted by higher N content at the leaf or the whole-plant level (Wright et al., 2004) and are accompanied by higher leaf and root respiration rates (Reich et al., 1997; Craine et al., 2002; Tjoelker et al., 2005; Atkinson et al., 2007). However, the extent to which carbon allocation patterns are related to N uptake efficiency and N allocation remains uncertain (Garnier, 1991; Osone et al., 2008). For example, fast-growing species have higher leaf N content and maximum carboxylation rates V_cmax compared with slow-growing species found at similar altitude (Baptist and Choler, 2008). These features are assumed to sustain rapid growth of the aboveground compartment at the expense of belowground organs, thereby supposedly causing lower root/shoot biomass ratios (Chapin, 1980). Nevertheless, there is no direct evidence of such a relationship, and thus that using shoot/root ratios as allocation estimates (as is commonly done: see, for example, Grulke and Bliss, 1985) may be questionable (Craine et al., 2002).

¹²C/¹³C and ¹⁴N/¹⁵N isotope techniques were used to investigate C fixation, N assimilation, and C and N allocation patterns of two alpine species. Carex foetida (CF) and Kobresia myosuroides (KM) were selected as case studies because they represent a fast-growing species inhabiting late snowmelt locations and a slow-growing species inhabiting early snowmelt locations, respectively. To our knowledge, the present work provides the first attempt to compare the coupled dynamics of C and N for alpine plants along a snow cover gradient. While the respiratory properties of the two species appear to be similar, the present labelling experiments show that, in contrast to KM, CF has different leaf/root allocation ratios for C and N. In other words, CF allocates more C to its roots whereas reduced N was more likely to be allocated to the aboveground compartment. It is therefore concluded that, in the present pair of species, the adaptation of fast-growing species for an optimal growth involves increased C and N assimilation efficiencies to feed belowground biomass production.

Materials and methods

Plant material

Twenty-eight plants each of KM and CF (both Cyperaceae) and their associated soil were collected in the vicinity of the Galibier pass [2646 m above sea level (a.s.l.)] and the Agnel pass (2744 m a.s.l.) in the French Alps in October 2005. They were transferred into pots (20×20×30 cm) in a common garden located in the Grand Galibier mountain range in the south-western Alps (45°7′N, 6°5′E, 2100 m a.s.l.). Snowmelt occurred on 15 May 2006. The pulse–chase labelling experiments (¹³C labelling and ¹⁵N labelling were independent) started on 5 July 2006 and ended on 26 July 2006. All the experiments were conducted at the Station Alpine Joseph Fourier, except the isotope measurements, which where carried out at the scientific facilities of the University of Barcelona. The δ¹³C of CO₂ of the air at the Galibier Pass was approximately –11.5±0.1‰ (Nogués et al., 2006).

¹³C labelling procedure

In July 2006, at the peak of standing biomass, 12 out of 15 monoliths (hereafter referred to replicates) were labelled in a ¹³CO₂-enriched atmosphere. The other three were used as controls (for the initial carbon isotope composition before labelling, the corresponding sampling time is denoted as T_init). After labelling, a first set of three was immediately harvested (T₀). For the other sets, the chase times lasted 24 h (1 d, T₁), 82 h (3.5 d, T₃), and 274 h (11.5 d, T₁₁), respectively. The day before pulse labelling, the replicates were arranged in controlled conditions: 12 h photoperiod, and mean air temperature and PPFD (photosynthetic photon flux density) of 18 °C and 550 μmol m⁻² s⁻¹, respectively. Labelling was applied following a dark period of 12 h for all replicates. After labelling and until the end of the chase period, replicates were kept in the same controlled conditions. Every 2 d, the monoliths were watered with 0.5 l of distilled water.

The isotope label was applied over a period of 5 h by enclosing the monoliths, two by two, in a 36.0 l Perspex^® labelling chamber. The atmospheric air was initially CO₂ depleted (decarboxylated) by passing through a soda-lime column and then mixed with ¹³CO₂ fluxes from a gas cylinder (enriched at 5%, Euriso-top, Saint-Aubin, France) in a mixing chamber. The mixing chamber was then connected to the sample air hose of the HCM 1000 Infra Red Gas Analyser (IRGA; Heinz Walz GmbH, Effeltrich, Germany) and the CO₂ concentration was estimated. The CO₂ concentration within the chamber was kept between 380 ppm and 420 ppm throughout the labelling procedure by mass flow controllers located before the mixing chamber (ROD-4, Aera, Fort Collins, CO, USA). The air flux passing through the labelling chamber was controlled by the HCM 1000 at a rate of 1.0 l min⁻¹_. A second pump was added at the output of the labelling chamber to avoid overpressurization. A fan was placed into the chamber to ensure air mixing. Aluminium tubes arranged at the bottom of the chamber and connected to a water bath maintained the chamber temperature at 22.5±0.4 °C during the pulses. Light intensity above the vegetation was kept at 550 μmol m⁻² s⁻¹.

At the end of each chase time (T₀, T₁, T₃, T₁₁) and also for T_init, the three replicates were harvested. The living aboveground vegetation was harvested first by clipping the leaves at the soil surface. Half of the biomass was immediately lyophilized, whereas the rest was enclosed for 5 min in a 1.0 l Perspex^® chamber to measure respiration (see below for procedure) and the isotopic signature of dark-respired ¹³CO₂. To measure the latter, the 1.0 l chamber connected to an IRGA model Li-6200 (LI-COR, Inc., Lincoln, NE, USA) was flushed with CO₂-free air to ensure that only the CO₂ respired in the chamber was measured. CO₂ was allowed to accumulate up to a concentration of ∼1000 ppm, and then air samples were collected using a special 50 ml syringe (SGE, Ringwood, Australia) and a needle (model microlance 3, BD, Plymouth, UK). The gas samples were passed through a magnesium perchlorate column (water vapour trap), then immediately injected into a 10 ml vacutainer (BD vacutainer, Plymouth, UK) as previously described (Noguès et al., 2008). To avoid contamination with the air present in the syringe and needle, both were flushed with N₂ before taking each sample. The vacutainers were also overpressurized with N_2, so the pressure inside the vacutainer was above the ambient pressure.

Non-woody (new) and woody (old) roots were subjected to a similar procedure after being harvested and washed. CO₂ accumulation time was longer for old roots as respiration was much lower than that of new roots and leaves. Leaves and roots were subsequently lyophilized for isotopic analysis of organic matter (OM). OM analyses were done using an elemental analyser with a zero-blank autosampler (EA1108, Series 1, Carbo Erba Strumentazione, Milan, Italy) coupled to an isotope ratio mass spectrometer (Delta C, Finnigan Mat, Bremen, Germany) operating in continuous flow.

The δ¹³C of respired CO₂ was measured using gas chromatography-combustion-isotope ratio mass spectrometry (GC-C-IRMS) as previously described (Nogués et al., 2008). Briefly, water vapour and oxygen were removed from the gas samples, and carbon dioxide, argon, and nitrogen were separated by gas chromatography (6890N, Agilent Technologies, Palo Alto, CA, USA) coupled to a Delta^plus isotope ratio mass spectrometer through a GC-C Combustion III interphase (ThermoFinnigan, Bremen, Germany). The column used was a 30 m×0.32 mm i.d. GS-GASPRO (J&W Scientific Inc., Folsom, CA, USA). The carrier gas was helium at a flow rate of 1.2 ml min⁻¹. Injection port temperature was 220 °C. The oven temperature was kept at 60 °C during the whole run. The injection was done in split mode (injected volume 0.3 ml, split flow 20 ml min⁻¹).

CO₂ gas exchange measurements

In order to evaluate ¹³C fixation during pulse labelling, CO₂ fluxes were measured on each replicate with a Li-6200 IRGA (LI-COR) before and just after labelling. Light intensity, air temperature, and relative air humidity were recorded during all measurements. The net photosynthetic fixation rate (A_net) was calculated as the sum of the net CO₂ fluxes in light (NEP) and in darkness (ER, which takes into account the CO₂ evolution rate by both the belowground and aboveground compartments) as follows:

(1)

In darkness, respiration by the aboveground compartment was small compared with that of the belowground compartment. In other words, the contribution of photosynthetic organs to ER was small (typically <8%), so the overestimation of the net photosynthetic rate A_net was negligible. At the end of the chase period (see section below), leaf and new and old root respiration was estimated every minute for 5–10 min by enclosing them in a dark 1.0 l Perspex^® chamber connected to the Li-6200 IRGA.

¹⁵N labelling procedure

The [¹⁵N]-NO₃, [¹⁵N]-NH₄, and [¹⁵N]-glycine uptake by KM and CF was assessed independently from the ¹³C labelling, i.e. on the remaining nine monoliths for each species (n=3 replicates for each compound and each species). Glycine was chosen as the amino acid as Lipson et al. (1999) indicated that it is the soil amino acid most available to plants. A 100 ml aliquot of a 1 mM solution of [¹⁵N]-NO₃, [¹⁵N]-NH₄, or [¹⁵N]-glycine was added with a 5 ml syringe; the concentration of N added was thus equivalent in the three treatments. The syringe was inserted to a depth of 5 cm in the soil following a 2×2 cm grid layout. Plants were then watered with 500 ml of demineralized water to ensure homogeneous labelling in the soil.

¹⁵N isotope sampling and processing

After a 24 h chase, leaves and new and old roots were harvested, sorted, washed with demineralized water, and lyophilized for isotopic analysis. ¹⁵N/¹⁴N ratios were analysed at the scientific facilities of the University of Barcelona using an elemental analyser (Flash 1112 EA, ThermoFinningan, Germany) operating in continuous flow mode. The analysis for obtaining the natural abundance of ¹⁵N in leaves and new and old roots of KM and CF was performed on the 15 monoliths used for the ¹³C labelling experiment.

Isotope labelling calculations

To estimate ¹³C and ¹⁵N enrichment in each organ of the plants, %Atom (¹³C or ¹⁵N proportion) for ¹³C and ¹⁵N was calculated using the following equation:

(2)

where δ is the isotopic signature of CO₂ respired or of OM. R_standard is the international standard reference (i.e. ¹³C/¹²C, PeeDee Belemnite, and ¹⁵N/¹⁴N, atmospheric air).

%Atom excess was then calculated as the %Atom ¹⁵N or ¹³C differences between labelled and unlabelled organs (control, at T_init):

The labelling-derived ¹³C content per DW (γ¹³C, in μg ¹³C g⁻¹ DW) in each organ of the plant was calculated as follows:

(3)

where %C is the percentage of carbon in the organ. The dynamics of the labelling-derived ¹³C content in leaves were described by a double-exponential equation.

The labelling-derived ¹³C flux associated with root and leaf respiration (γ¹³C_R, in μg ¹³C g⁻¹ h⁻¹) was calculated as follows:

(4)

where mass_organ is the mass of the organ (g) considered, R_organ is the respiration rate (μg C h⁻¹), and %Atom excess is here the ¹³C atom excess in CO₂.

As the plants experienced similar conditions during the chase period and the respiration measurements, the cumulative labelling-derived ¹³C content over time (in mg ¹³C g⁻¹) was estimated by (i) fitting an exponential decay constant to the labelling-derived ¹³C flux over chase time (γ¹³C_R):

(5)

(where t is the time in hours and a and b are constants); and (ii) integrating this exponential over time. For CO₂ respired by new and old roots, the maximum concentration sampling point was used as the ‘time zero’ for the exponential curve fit.

Total labelling-derived ¹³C mass (γ¹³C_M, μg or mg ¹³C) at chase time T was calculated by averaging the organ mass over the 15 replicates (see Table 1). For this purpose, the labelling-derived ¹³C mass for each organ and the loss through leaf and new/old root respiration from T₀ to T were added as follows:

(6)

By doing so, it was possible to see whether the labelling-derived ¹³C mass was balanced over time for each species, i.e. whether the total amount of ¹³C (remaining+respiratory losses) remained constant (Supplementary Table S1 available at JXB online). γ¹³C_M was also compared with the amount of ¹³C fixed by each replicate during the 5 h labelling period based on A_net values integrated over the labelling period.

Table 1.

Morphological and physiological characteristics of C. foetida and K. myosuroides

	K. myosuroides	C. foetida
Morphological characteristics
Mean dry biomass (g)
Aboveground	2.6 (0.1)	1.9 (0.1)
New roots	0.51 (0.07)	0.80 (0.09)
Old roots	21.6 (1.8)	28.00 (2.0)
Belowground productivity (g m−2 d−1)	6.9 (1.8) a	12.5 (0.6) b
Physiological characteristics
Assimilation rate during labelling (ng C g−1 leaf DW s−1)	791.5 (20.3) a	622.4 (13.3) b
Leaf respiration in darkness (ng C g−1 leaf DW s−1)	63.5 (6.7) a	100.3 (7.4) b
New root respiration (ng C g−1 new root DW s−1)	233.6 (21.0) a	206.3 (15.0) a
Old root respiration (ng C g−1 old root DW s−1)	56.4 (4.3) a	68.6 (13.0) a

See text for further statistical details. Values are the mean ±SE (n=15, except for belowground productivity, where n=4, and assimilation rate, where n=12). Different letters indicate significant differences between the two species (P <0.05).

The labelling-derived ¹⁵N content per whole plant dry matter (γ¹⁵N, μg ¹⁵N g^-1) was calculated as follows:

(7)

where %N is the percentage of nitrogen in the organ, mass_organ, is the mass of the organ (g), and mass_plant is the mass of the whole plant (g).

Similar to the ¹³C data analysis, the labelling-derived ¹⁵N mass (μg¹⁵N) was calculated for each organ based on the average mass over the nine replicates of each species (see Table 1).

Belowground productivity

Belowground productivity was estimated using a root ingrowth core method. The ingrowth cores (5×15 cm) were installed on 30 May 2006 by removing one soil core in a set of four replicates for each species and by filling the holes with the same soil sieved through 2 mm mesh. Roots present within the cores were sampled on 15 July 2006, washed free of soil, dried at 60 °C for 48 h, and weighed. Root production was then calculated as the root biomass divided by time between 30 May 2006 and 15 July 2006, and expressed per unit core surface for each replicate.

Statistical procedure

For isotopic data, a non-parametric Kruskall–Wallis test was performed, except for ¹⁵N natural abundance for which a one-way analysis of variance (ANOVA) was used. Similarly, a non-parametric Kruskall–Wallis test was used to compare belowground productivity (n=4). The species effect on flux measurements and root/shoot ratios were analysed using a one-way ANOVA. Finally, an analysis of covariance (ANCOVA) was applied to test for the regression between (i) the labelling-derived quantity of ¹³C (γ¹³C) in new root OM against that of leaf OM and (ii) the γ¹³C of root-respired CO₂ against that of leaf-respired CO₂. Species was considered as a qualitative factor. All analyses were performed with Jump software (SAS Institute Inc., Cary, NC, USA).

Results

Biomass production

Mean biomass and belowground productivity data are indicated in Table 1 (upper part). Under the present experimental conditions, KM had a larger aboveground biomass and a lower belowground biomass than CF and hence a higher shoot/root ratio. In addition, the root productivity per unit ground area was higher in CF as compared with KM (χ²=5.0₁, P=0.02, Table 1).

Photosynthetic rates and total ¹³C assimilation

During ¹³C pulse time, the net assimilation rate of the above compartment (A_net) was significantly larger in KM than in CF (F_1,21=5.1, P=0.03). This does not reflect a general trend: in homogeneous light conditions, CF had higher photosynthetic rates (see Supplementary Fig. S1 at JXB online; Baptist and Choler, 2008). The lower value in CF observed here was caused by the non-saturating light conditions and is in contrast to values of maximal leaf CO₂ assimilation (in saturated light, see Supplementary Fig. S1). Assimilation values were summed over the pulse time to calculate the total labelling-derived ¹³C mass (γ¹³C_M) fixed by photosynthesis, giving 1.33±0.2 mg for KM and 0.93±0.1 mg for CF (Supplementary Table S1 at JXB online).

Respiratory properties

Leaf respiration in darkness was largely and significantly higher in CF (F_1,28=13.2, P=0.001) (Table 1). Thus, in the present experimental conditions, the leaf respiration to net assimilation ratio was 8.0% in KM and 16.1% in CF. New and old root respiration did not differ significantly between the two species.

¹³C fixation and partitioning

The ¹²C/¹³C isotope composition (δ¹³C) of OM and of respired CO₂ after labelling is shown in Fig. 1. Clearly, at T₀, leaves were the most labelled organs, followed by new roots (Fig. 1). Old roots showed a low degree of labelling throughout the experiment (i.e. until the end of the chase time). The kinetics were very different for leaves and roots: the δ¹³C value of leaves continuously declined during the chase time while that of new roots increased within 1.5 d and reached a plateau.

Fig. 1.

δ¹³C (‰) of leaves, and new and old roots of K. myosuroides (a) and C. foetida (b). δ¹³C (‰) of CO₂ respired by the leaves, and new and old roots of K. myosuroides (c) and C. foetida (d) following pulse labelling. On the left of the dashed vertical line is shown the natural abundance of ¹³C in the OM of leaves, and new and old roots, and the CO₂ respired by the leaves, and the new and the old roots. Leaves, open circles; new roots, open squares; old roots, filled squares. All x-axes show the time elapsed since pulse labelling (in days). Values are the mean ±SE (n=3).

The decline in the amount of ¹³C in leaves may be caused by: (i) isotopic dilution (natural ¹²CO₂ fixed during the daytime of the chase period); (ii) a loss of ¹³C by respiration (dark-respired CO₂ was strongly ¹³C enriched, Fig. 1c and d); and (iii) export (¹³C increases in roots within a couple of days). It is noteworthy that fixed ¹³C in CF leaf OM was exported more rapidly compared with KM [the calculated half-time of the exponential decay (t_1/2 of the fastest turnover pool) is nearly 21 h for KM and only 13 h for CF].

A larger isotopic dilution is probably not responsible for such a difference: plants of both species experienced similar conditions, and KM had a slightly higher photosynthetic rate than CF under the conditions of the experiment (see paragraph above, Table 1). In contrast, respiration contributed to these leaf ¹³C kinetics simply because the respiration rate was larger in the case of CF (Table 1) while having a similar ¹³C enrichment (Fig. 1c, d).

The faster decline of leaf ¹³C in CF than KM also came from the export of larger amounts of assimilates to roots: the δ¹³C value of OM in new roots was higher in CF than in KM. Furthermore, in the steady state, the δ¹³C values of new root OM were 0‰ and 40‰ for KM and CF, respectively (Fig. 1a, b). That is, the labelling-derived carbon in new roots accounted for ∼0.6% and 1.8% of the total C in KM and CF, respectively.

The labelling-derived quantity of ¹³C (γ¹³C) in new roots is plotted against that of leaves in Fig. 2. There was a significant correlation between the γ¹³C value of new roots and that of leaves when the species were considered together (F_1,22=46.1, R²=0.71, P <0.0001, Fig. 2a). The slope of the regression was slightly (but not significantly) steeper in CF, again showing more rapid ¹³C kinetics in leaves. In addition, the more rapid transfer of carbon from leaves to roots was reflected by the shift of data points from the right-hand side (T₀) to the left-hand side (the chase measurements, at T₁–T₃, clustered on the left, Fig. 2A, open symbols) in CF, while there were intermediate data points in KM. γ¹³C_R values of leaf-respired CO₂ and root-respired CO₂ were correlated (F_1,23=22.2, R²=0.56, P=0.0001, Fig. 2b) when species were considered together. However, neither the slope of the regression (F_1,23=0.44, P=0.51) nor the mean value (F_1,23=2.57, P=0.12) differed between the two species.

Fig. 2.

(a) Labelling-derived ¹³C content of new roots in relation to labelling-derived ¹³C content of leaves and (b) labelling-derived ¹³C content of CO₂ respired by the new roots in relation to the labelling-derived ¹³C content of CO₂ respired by the leaves for K. myosuroides (filled symbols) and C. foetida (open symbols) at each chase time. Values are the mean ±SE (n=3). See text for further statistical details.

Whole-plant carbon partitioning

Eleven days after pulse labelling, carbon allocation patterns were calculated, taking into account integrated respiratory losses. The results are shown in Fig. 3 (raw data are shown in Supplementary Fig. S2 at JXB online). In KM, 45% of the labelling-derived ¹³C remained in the leaves, and nearly 5% of the labelling-derived ¹³C was allocated to new roots. For CF, 29% of the labelling-derived ¹³C remained in the leaves, while new roots accounted for ∼20%. Respiratory losses associated with leaves were more pronounced in KM than in CF, while those associated with new root respiration were larger in CF. However, in the case of new roots of CF, the respiration/OM ratio of ¹³C allocation was much smaller (Fig. 3, light grey), clearly suggesting that ¹³C was directed to new root growth to a greater extent than to respiration.

Fig. 3.

Percentage allocation of the total labelling-derived ¹³C mass (γ¹³C_M) recovered in the leaves and the new and old roots and lost through leaf, and new and old root respiration 11^d after pulse labelling. Leaves, white; new roots, light grey; old roots, dark grey. Hatched patterns correspond to C lost through respiration.

Nitrogen allocation and partitioning

The natural ¹⁴N/¹⁵N isotope composition (δ¹⁵N) of unlabelled organs and soil is indicated in Table 2. Soil N was always ¹⁵N enriched (P <0.05) by 1.7–7.1‰, so a ¹⁴N/¹⁵N isotope fractionation during N reduction/assimilation is apparent. KM leaves were clearly and significantly ¹⁵N depleted compared with new and old roots, whereas CF leaves had nearly the same δ¹⁵N value as new and old roots and were slightly enriched compared with KM leaves.

Table 2.

¹⁵N natural abundance (δ¹⁵N, ‰) of unlabelled leaves, new roots, old roots, and soil in K. myosuroides and C. foetida monoliths

δ15N (‰)	K. myosuroides	C. foetida
Leaf	–3.40 (0.37) a	0.19 (0.23) a
New roots	0.61 (0.62) b	–0.02 (0.37) a
Old roots	–0.76 (0.14) b	2.73 (1.12) ab
Soil	3.78 (0.12) c	4.49 (0.95) b

Values are the mean ±SE (n=15). Different letters indicate significant differences between the organs and the soil (P <0.05).

The nitrogen allocation after 24 h ¹⁵N labelling with [¹⁵N]-nitrate, [¹⁵N]-ammonium, or [¹⁵N]-glycine is indicated in Fig. 4. Clearly, CF was more ¹⁵N-labelled than KM at the whole-plant level, no matter which labelling molecule was applied. In other words, CF had a higher nitrogen uptake efficiency (NupE) than KM: the recovery of ¹⁵N from the labelling solution was 13.7±2.5% glycine, 19.8±3.1% NO₃^–, 2.9±1.2% NH₄⁺ for CF; and 4.6±0.5% glycine, 10.4±3.8% NO₃^–, and 1.8±0.4% NH₄⁺ for KM. For nitrate, the CF/KM ratio of NupE was then as high as 1.9.

Fig. 4.

Labelling-derived ¹⁵N content in whole-plant dry matter (μg ¹⁵N g^-1 DW) after amendment with [¹⁵N]-NO₃^– (a), [¹⁵N]-NH₄⁺ (b), or [¹⁵N]-glycine (c) in K. myosuroides (grey bar) and C. foetida (white bar). Circle graphs, percentage allocation of the total labelling-derived ¹⁵N mass recovered in the leaves (white), new roots (grey), and old roots (black) (μg ¹⁵N μg^{-1 15}N in whole-plant dry matter). Values are the mean ±SE (n=3). ns, not significant, * indicates significant differences between both species (P <0.05). See text for further statistical details.

In addition, nearly twice as much N was allocated to leaves in CF as in KM when ¹⁵N was supplied as nitrate or ammonium (circle graphs, Fig. 4a, b). Such a difference was less pronounced with [¹⁵N]-glycine (Fig. 4c). However, old roots represent a large ¹⁵N sink in Fig. 4 simply because of their high biomass. In other words, the specific ¹⁵N abundance (μg ¹⁵N g⁻¹ DW) was always low in old roots (data not shown). In the case of nitrate, the specific ¹⁵N abundances of leaves were 72.6±26.8 and 20.3±9.7 μg ¹⁵N g⁻¹ DW in CF and KM, respectively, giving a CF/KM ratio of 3.5. Such a ratio is larger than the NupE ratio (1.9, see above), clearly demonstrating that preferential N allocation to leaves, rather than whole-plant uptake efficiency, was responsible for the larger ¹⁵N mass in CF leaves (Fig. 4a).

Discussion

This study aimed to clarify C and N partitioning patterns at peak standing biomass of the alpine plant C. foetida (CF), a fast-growing species from late snowmelt locations, and K. myosuroides (KM), a slow-growing species from early snowmelt locations. Both natural isotopic abundances and isotopic labelling (¹³C, ¹⁵N) were used, and C and N assimilation patterns were followed.

Carbon fixation and partitioning

CF exhibits high relative growth rates (e.g. belowground productivity, Table 1) and photosynthetic capacity (Supplementary Fig. S1 at JXB online). This agrees with the larger maximal carboxylation rate V_cmax and higher leaf N elemental content in CF (Choler, 2005; Baptist and Choler, 2008), which is an indicator of the maximum velocity of ribulose bisphosphate (RuBP) carboxylation by Rubisco. Whole-plant carbon allocation favoured the root compartment in CF as indicated by the larger ¹³C transfer to new roots (Figs 1–3) as compared with KM. Such a pattern might correspond either to a larger flux of assimilates from aboveground to belowground organs or to a higher ¹³C-specific abundance of assimilates. Since the kinetics of the ¹³C decline in leaf OM are faster in CF (Fig. 1; t_1/2 values of 13 h versus 21 h in KM; Fig. 2a) and the initial leaf ¹³C abundance (carbon source) was very similar in both species, the first hypothesis is favoured.

The mass balance after an 11 d chase period indicated that the larger carbon flow from leaves to roots was directed to new root growth rather than respiration in CF plants (Fig. 3). In fact, the ¹³C abundance in root-respired CO₂ is dissimilar to the amount in total OM: while the ¹³C allocation (labelling-derived ¹³C content) to new root total OM is nearly three times higher in CF (Fig. 1, and Supplementary Fig. S2 at JXB online), the maximum ¹³C content in root-respired CO₂ was only doubled. In other words, the turnover (consumption) of the root respiratory pool was lower in CF.

Nevertheless, a large difference was found between woody (old) and non-woody (new) roots. Old roots were weak ¹³C sinks (very low ¹³C abundance after labelling, Fig. 1) with low respiratory activity as compared with new roots (nearly 70% less, Table 1). This reflects differences in metabolic activities: new roots are responsible for root growth and nutrient absorption while old roots have a conduction and storage role (Comas et al., 2000; Lipp and Andersen, 2003; Volder et al., 2005). However, at the whole-plant level, old roots accounted for substantial ¹³C content because of their very large biomass (Table 1), while the C allocation pattern to old roots was somewhat similar in both species (Figs 1, 3).

Nitrogen uptake and assimilation

Variations in ¹⁵N uptake between both species (Fig. 4) are consistent with findings of previous studies that suggest that fast-growing species display higher specific nitrogen absorption rates than slow-growing species (Garnier, 1991; Poorter et al., 1991). In addition, CF and KM exhibited different N allocation patterns: CF experienced higher ¹⁵N allocation to leaves after a 24 h chase period as compared with KM (Fig. 4). Consequently, one may assume that N translocation toward the aboveground compartment was more efficient in CF than in KM plants.

However, there is a predominance of leaf N reduction over root N reduction in KM as compared with CF (Table 2). In fact, in normal conditions where nitrate is reduced by both leaves and roots, the natural ¹⁴N/¹⁵N isotope composition (δ¹⁵N) of leaves is higher (¹⁵N enriched) than that of roots, regardless of the δ¹⁵N and N content in the soil. This is because nitrate reduction fractionates against ¹⁵N, thereby enriching the remaining nitrate molecules exported to leaves (for a recent review, see Tcherkez and Hodges, 2007). This was typically the case in CF (Table 2). In contrast, KM leaves were not ¹⁵N enriched (and are in fact slightly ¹⁵N depleted) as compared with new roots, demonstrating that N reduction occurred mainly in leaves (Table 2). Accordingly, under non-limiting nutrient availability, shoots generally appear to be the predominant site of NO₃^– reduction because of the higher content of excess reductants produced by photosynthesis (Scheurwater et al., 2002).

The mechanisms underlying such a shoot versus root pattern of N reduction are uncertain. It is believed that it is influenced by the photosynthetic rate (which generates reductants and carbon skeletons) as well as biochemical parameters (maximum organ-specific velocity of nitrate reductase activity), biomass allocation, and environmental conditions (Gojon et al., 1994). Here, is it possible that the high C flux allocated to the belowground compartment in CF promotes significant levels of NO₃^– reduction in the roots (Pate, 1980). Although arctic and alpine slow-growing species such as KM are thought to have both a smaller NupE and a low nitrate reductase activity (Atkin, 1996), it remains plausible that a high rate of NO₃^– reduction would contribute to mitigating nitrate retroflux (back-diffusion to the soil), thereby increasing NupE in KM (Mata et al., 2000; Miller and Cramer, 2004).

Biomass and energetic root/shoot balance

A larger rate of root respiration was expected in CF as compared with KM plants as more respiratory energy was necessary to sustain higher root productivity and nitrogen uptake (almost 2-fold larger). Nevertheless, root respiration did not differ between the two species (Table 1), i.e. the fast-growing species respired at a lower rate than would be expected from their C and N metabolism. Previous studies have reported similar results concerning the absence of apparent relationships between the relative growth rate and the root respiration rate (Van der Werf et al., 1988; Poorter et al., 1991).

Root respiration can be decomposed into three components, i.e. the respiratory cost of root growth and of ion uptake, and maintenance respiration. A classical relationship has been proposed between total root respiration (denoted as R) and such components, as follows (modified version from Van der Werf et al., 1994):

where R is root respiration, R_m is maintenance respiration, NupE is nitrogen uptake efficiency, P_r is root productivity, and C_u and C_g are the specific respiratory costs of ion uptake and growth, respectively.

It has been shown that both nitrate uptake efficiency and root growth are 2-fold higher in CF, while both species exhibit similar root respiration rates (Table 1). Consequently, it may be assumed that the respiratory costs of ion uptake and growth (C_g and C_u coefficients) are much lower in CF as compared with KM unless there is a large difference in the value of maintenance respiration. Such an assumption is not likely. Scheurwater et al. (1998) and Van der Werf et al. (1988) demonstrated that within a plant life form (e.g. Graminoids), maintenance respiration differs only very slightly between fast- and slow-growing species.

Although several studies have suggested that the construction costs of roots and leaves may not differ between fast- and slow-growing species (Van der Werf et al., 1988; Navas et al., 2003; Roumet et al., 2006), it was not possible to obtain experimentally the cost coefficients (C_g and C_u) in the present study, and thus it remains plausible that both coefficients differ between the two species.

Conclusions

CF and KM represent contrasting examples of C and N dynamic allocation during the growth period. Relative to KM, CF exhibits an improved photosynthetic capacity and a lower N assimilation capacity in its leaves, which are compensated for by (i) a preferential carbon allocation to roots favouring root growth (storage) and NO₃^– reduction and (ii) a larger translocation of reduced N to aboveground organs. Unexpectedly, then, CF root respiration was much lower than expected, possibly due to a lower respiratory cost for ion uptake. The large C allocation to the belowground compartment during the growing season might also ensure sufficient root storage for sustaining the extremely rapid greening of snowbed plants just after snowmelt in the following year.

The results obtained here suggest that an allocation-based balance between root N reduction and leaf CO₂ assimilation is involved in growth strategies of alpine species growing under short, energy-limited vegetation periods. This coupling between C and N fluxes was less apparent in the case of KM, which benefited from a longer growing season but was nutrient limited. Nevertheless, it is recognized that other environmental parameters may be involved in C and N allocation patterns (e.g. temperature, CO₂ mole fraction), impacting on growth, respiration, and N metabolism. The importance of temperature-mediated changes in whole-plant C allocation as a result of growing season progress also needs to be established (Atkin et al., 2007). These aspects will be addressed in a future work.

Supplementary data

Supplementary data, comprising Table S1, and Figures S1 and S2, are available at JXB online.