Post-uptake metabolism affects quantification of amino acid uptake

Summary

• •
  The quantitative significance of amino acids to plant nutrition remains controversial. This experiment determined whether post-uptake metabolism and root to shoot export differ between glycine and glutamine, and examined implications for estimation of amino acid uptake.
• •
  Field soil containing a Eucalyptus pauciflora seedling was injected with uniformly ¹³C- and ¹⁵N-labelled glycine or glutamine. I quantified ¹⁵N and ¹³C excess in leaves and roots and intact labelled amino acids in leaves, roots and stem xylem sap. A tunable diode laser quantified fluxes of ¹²CO₂ and ¹³CO₂ from leaves and soil.
• •
  60–360 min after addition of amino acid, intact molecules of U-¹³C,¹⁵N glutamine were < 5% of ¹⁵N excess in roots, whereas U-¹³C,¹⁵N glycine was 30–100% of ¹⁵N excess in roots. Intact molecules of glutamine, but not glycine, were exported from roots to shoots.
• •
  Post-uptake metabolism and transport complicate interpretation of isotope labelling such that root and shoot contents of intact amino acid, ¹³C and ¹⁵N may not reflect rates of uptake. Future experiments should focus on reconciling discrepancies between intact amino acid, ¹³C and ¹⁵N by determining the turnover of amino acids within roots. Alternatively, post-uptake metabolism and transport could be minimized by harvesting plants within minutes of isotope addition.

Introduction

Soil contains a complex mixture of nitrogen (N)-containing compounds, from inorganic N (nitrate and ammonium) through organic N monomers such as amino acids to organic N polymers such as peptides and proteins. Over recent decades we have amassed considerable qualitative evidence showing that plants can take up intact amino acids from soil (Chapin et al., 1993; Schimel & Chapin, 1996; Lipson & Näsholm, 2001; Näsholm & Persson, 2001). In recent times, the major research challenge has shifted from determining if amino acids are taken up to determining the quantitative significance of amino acid uptake (Näsholm et al., 2009). The problem is that the methods that have proved useful in demonstrating the (qualitative) occurrence of amino acid uptake are not necessarily quantitative.

The most common method for estimating uptake is to use amino acids in which carbon (C) and/or N is labelled with stable isotope (e.g. uniformly ¹³C- and ¹⁵N-labelled glycine: U-¹⁵N,¹³C Gly) and quantify isotope label in plant tissues (Näsholm et al., 1998; Näsholm & Persson, 2001; Warren, 2009b). Intact uptake is implied if the slope of the correlation of ¹³C to ¹⁵N excess in the plant tissue is the same as in the intact amino acid (e.g. 2 ¹³C : 1 ¹⁵N in the case of added U-¹⁵N,¹³C Gly). However, the same correlation between ¹³C and ¹⁵N could also arise if a double-labelled amino acid were mineralized in the soil to inorganic compounds (i.e. ¹⁵NO₃⁻, ¹⁵NH₄ ⁺  and ¹³CO₂) that were subsequently taken up independently (Jones et al., 2005a). Moreover, many studies have found that the correlation of ¹³C to ¹⁵N excess is not significant, weak or different from expectations (e.g. ≠ 2 ¹³C : 1 ¹⁵N in the case of added U-¹⁵N,¹³C Gly) (Bardgett et al., 2003; Weigelt et al., 2005; Harrison et al., 2007). Discrepancies between expected and measured ratios of ¹³C to ¹⁵N may arise as a consequence of rapid metabolism and loss of ¹³C, in particular if the isotope label is in a position prone to metabolism (e.g. the carboxy-C of glycine) (Schimel & Chapin, 1996). An additional complication is that the ¹³C isotope is strongly ‘diluted’ by the high C : N ratio of plant tissues and this can make it difficult to detect significant ¹³C in bulk plant tissues and lead to anomalous ratios of ¹³C : ¹⁵N (Näsholm & Persson, 2001).

Amino acids labelled with ¹⁴C have also been used extensively to estimate uptake of amino acids (e.g. Jones et al., 2005a,b). Use of the radio isotope ¹⁴C overcomes the problem of dilution by the high C content of plant tissues and leads to a very large signal. One major drawback with ¹⁴C-labelled amino acids is that ¹⁴C label could enter the plant as intact amino acid (as assumed) or following deamination (e.g. as ¹⁴CO₂ or ¹⁴C-carboxylate). Use of dual ¹⁴C, ¹⁵N-labelled amino acids (e.g. Xu et al., 2008) can provide somewhat stronger evidence for amino acid uptake than ¹⁴C amino acids, and overcomes the ‘dilution’ problem that afflicts use of ¹³C, ¹⁵N-labelled amino acids. However, the problem remains that the presence of both labels within the plant could be attributable to uptake of intact amino acid molecules or independent uptake of ¹⁴C and ¹⁵N.

The strongest evidence for uptake of intact amino acids comes from the presence of intact molecules of labelled amino acids within plant tissues (e.g. as determined by gas chromatography–mass spectrometry (GC-MS); Persson & Näsholm, 2001). However, interpretation is rarely straightforward because the amount of labelled amino acid measured in plant tissues is commonly smaller than the total amounts of ¹⁵N and ¹³C taken up (Warren, 2009a,b). This raises the problem of determining whether some proportion of amino acid N was taken up after deamination or whether the amino acid was taken up intact and subsequently metabolized within the plant (Näsholm et al., 2000).

Post-uptake metabolism and within-plant transport of amino acids may explain anomalous relationships between ¹³C and ¹⁵N and discrepancies between amounts of labelled amino acid and ¹⁵N or ¹³C. In general terms, any metabolic reaction will lead to an apparent ‘loss’ of intact labelled amino acid molecules because of the transfer of ¹⁵N and/or ¹³C to other molecules, whereas effects on ¹³C and ¹⁵N depend on the specific reaction (e.g. deamination vs transamination) and fate of products (retained vs exported). One study reported that glycine taken up via roots was primarily metabolized via transaminase reactions and a fraction of labelled glycine was detected in the xylem sap (Schmidt & Stewart, 1999). However, with the exception of this one paper, we know little about the fate of amino acids following uptake. Several authors have suggested that lower than expected ¹³C:¹⁵N ratios and small amounts of labelled amino acid are because amino acids are rapidly metabolized with consequent loss of ¹³C label as ¹³CO₂ (e.g. Persson & Näsholm, 2001; Bennett & Prescott, 2004; Hawkins et al., 2005). Unfortunately, in most cases the putative loss of ¹³C has not been quantified, or a hydroxide trapping method has been used. Major drawbacks of the hydroxide trapping method are that it cannot be used with photosynthesizing leaves and generally involves placing a plant in a closed system that is not at steady state. These drawbacks of the hydroxide trapping methods make it difficult to relate measures of ¹³CO₂ efflux to real-world conditions (e.g. a photosynthesizing plant in a well-mixed, steady-state atmosphere).

An additional complication that has not received any attention is that there are probably differences between amino acids in rates of metabolism and within-plant transport, and these may confound studies comparing uptakes of different amino acids (e.g. Weigelt et al., 2005; Harrison et al., 2007, 2008). This study contrasts uptakes of glycine and glutamine by seedlings of Eucalyptus pauciflora. Glycine and glutamine were chosen because it was hypothesized that they would have contrasting patterns of metabolism and transport. I test the hypothesis that amino acids that are minor components of the xylem sap (e.g. glycine) are metabolized within roots before transport to shoots, whereas amino acids that are major components of the xylem sap (e.g. glutamine) are exported from root to shoot as intact molecules. Moreover, the differing locations of metabolism (roots for glycine; shoots for glutamine) will affect relationships among amounts of ¹³C, ¹⁵N and intact amino acid molecules. To determine the fate of glycine and glutamine added to the soil, uniformly ¹³C- and ¹⁵N-labelled glycine and glutamine were injected into the soil. The ¹⁵N and ¹³C excess of leaves and roots was determined by isotope ratio mass spectrometry while the presence of intact labelled amino acids in roots, leaves and xylem sap was determined by GC-MS. To test the hypothesis that the loss of ¹³CO₂ from leaves is significant (e.g. Persson & Näsholm, 2001; Bennett & Prescott, 2004; Hawkins et al., 2005), instantaneous fluxes of ¹²CO₂ and ¹³CO₂ in leaves were quantified with a tunable diode laser. Simultaneous measurements of ¹²CO₂ and ¹³CO₂ efflux from the soil surface provided an indication of amino acid metabolism in roots and soil.

Materials and Methods

Plant material and soil

Soil was collected (0–15 cm) from a subalpine Eucalyptus pauciflora Sieber ex Spreng. woodland in the Snowy Mountains of Australia (36°06′S; 148°32′E; 1500 m altitude). The soil is classified as a humic umbrosol (World Reference Base) or chernic tenosol (Isbell, 2002) and has been described previously (Warren, 2009b; Warren & Taranto, 2010). The soil is a well-drained sandy loam without coarse fragments > 2 mm, pH (H₂O) is 4.5, organic C (Walkley and Black method) is 12–17%, and total N is 0.2–0.3%. Seeds of E. pauciflora (seedlot 19626; Australian Tree Seed Centre, ACT, Canberra, Australia) were sown into soil, stratified at 4°C for 4 wk and then germinated in a fully sunlit glasshouse at the University of Sydney (Camperdown, NSW, Australia). Germinants were transferred to individual 210-ml plastic tubes filled with the same soil and grown for 1 yr before being re-potted into 1.89-l glass preserving jars (wide-mouth preserving jar; Ball Corporation, Broomfield, CO, USA) filled with c. 1.5 l of soil packed to field density. Measurements of ¹⁵N uptake and ¹³C evolution were made when the seedlings were 18 months old and weighed c. 12 g each. The mean temperature inside the glasshouse during the period of plant growth was 22.1°C with an absolute maximum of 35.8°C and minimum of 11.7°C. Average daily photosynthetically active radiation inside the glasshouse was 472 μmol m⁻² s⁻¹ (18.8 mol m⁻² d⁻¹) with an absolute maximum of 1701 μmol m⁻² s⁻¹.

Gas exchange system for measuring fluxes of ¹²CO₂ and ¹³CO₂

To determine if added isotope label was subsequently respired as ¹³CO₂, fluxes of ¹²CO₂ and ¹³CO₂ from leaves and soil were quantified. Measurements were made inside the laboratory in a temperature-controlled incubator fitted with two LED arrays and two fluorescent tubes that provided c. 150–300 μmol photosynthetically active radiation (PAR) m⁻² s⁻¹ at plant height for a 10–11-h photoperiod. Air (and soil) temperatures during measurements were 24–27°C. ¹²CO₂ and ¹³CO₂ exchange by soil and leaves was determined with a tunable diode laser (TGA100a; Campbell Scientific, Logan, UT, USA) from continuous measurements of ¹²CO₂ and ¹³CO₂ from the soil headspace and one leaf per plant (Fig. 1). To permit measurements of ¹²CO₂ and ¹³CO₂ exchange from the soil headspace, inlet and outlet quick-connect bulkhead fittings were fitted to the lids of the glass preserving jars, as described previously (Anonymous, 2009). To accommodate the stem of a seedling, a 10-mm-diameter hole was drilled in the centre of the lid, and then an 8–10-mm-wide removable tongue was cut from the centre hole to the periphery. The stem was positioned in the centre of the lid and sealed with plastic reusable adhesive (Blu-Tak; Bostik, Kings Park, NSW, Australia), and the 8–10-mm-wide tongue was replaced and sealed with cloth tape (cloth gaffa tape, 3M; Pymble, NSW, Australia) and plastic reusable adhesive (Blu-Tak). The screw-on sealing ring was placed over the top of the seedling and screwed onto the jar. For measurements of leaf gas exchange I used a custom-built 18-cm² chamber covered with transparent Propafilm connected to an open gas exchange system (Li-6400; Li-Cor, Lincoln, NE, USA) and the tunable diode laser, as described previously (Douthe et al., 2011; Warren et al., 2011). When a leaf was placed in the chamber it was kept in its natural orientation and received illumination from the LED arrays and fluorescent tubes of the refrigerated incubator.

Figure 1. Schematic diagram of the system used to obtain real-time measurements of ¹²CO₂ and ¹³CO₂ exchanges in the headspace of a soil chamber and the leaf of a seedling. The plant and soil chamber were set up inside the refrigerated incubator that controlled air temperature and provided LED and fluorescent lighting for the plant. To simplify the diagram, I have omitted the seedling and plumbing details for the manifold system and tunable diode laser.

Air from outside the laboratory was drawn at 2.5 l min⁻¹ into a 10-l buffer volume and humidified to a dew-point of 15°C (Li-610 dew-point generator; Li-Cor) (Fig. 1). Air from the dew-point generator was split into four streams: soil headspace, leaf gas exchange system, ‘reference gas’ entering the soil chamber, and a vent to avoid over-pressurizing the soil headspace or leaf gas exchange system. Samples of air from the soil headspace, and reference and sample gas from the leaf gas exchange chamber were continuously drawn into the manifold system of the tunable diode laser at 200 ml min⁻¹. The tunable diode laser was programmed to measure¹³CO₂ and ¹²CO₂ of, in turn, the two calibration gases, and the four intakes (air entering and exiting the soil and leaf chambers). Each intake was measured for 45 s, but the first 15 s was ignored to minimize carryover and enable stabilization between intakes. Calibration of the tunable diode laser was a two-step process with an initial linear interpolation between two calibration cylinders, and then a minor recalculation to account for some nonlinearity and permit extrapolation beyond the range of the two calibration cylinders (Douthe et al., 2011). In brief, the initial (‘working’) calibration used two calibration cylinders with absolute concentrations of ¹²CO₂ and ¹³CO₂ of 414.5 and 4.49 μmol mol⁻¹ for cylinder 1 and 286.9 and 3.10 μmol mol⁻¹ for cylinder 2, respectively. Recalculating data required knowledge of the deviation between the carbon isotope ratio (δ¹³C) estimated with the linear interpolation and actual δ¹³C. To determine this deviation, an 8-g cartridge of pure CO₂ (soda charger; iSi, Vienna, Austria) was diluted with the gas mixer of an Li-6400 to create CO₂ mole fractions from 200 to 2000 μmol mol⁻¹ with the same δ¹³C. The Li-6400 was plumbed to the tunable diode laser and replicated measurements were made at a range of CO₂ concentrations. The tunable diode laser was highly linear across the entire range of measured ¹²CO₂ and ¹³CO₂, with the highest concentrations of ¹²CO₂ and ¹³CO₂ differing from actual concentrations by < 0.5%. There was an offset of < 3‰ in calculated δ¹³C at the largest CO₂ concentrations measured in this experiment. The relationship between measured (with linear interpolation) and actual δ¹³C was used to recalculate data. The precision of the measurement of δ¹³C was determined by substituting the source of external air with a cylinder of CO₂ in air (¹²CO₂ = 405.5 μmol mol⁻¹ and ¹³CO₂ = 4.54 μmol mol⁻¹), and then measuring δ¹³C of each of the four intakes (i.e. air entering and exiting empty soil and leaf chambers). The precision of each of the four intakes was better than 0.06‰ (SD; n = 10).

Rates of soil respiration were calculated separately for ¹²CO₂ and ¹³CO₂ based on tunable diode laser measurements of (dry) concentrations of ¹²CO₂ and ¹³CO₂, and flow measured within the manifold system:

• (Eqn 1)

where flow is the molar flow of air through the soil headspace, and CO₂s and CO₂r are concentrations of ¹²CO₂ or ¹³CO₂ measured in the sample and reference air, respectively. Carbon isotope discrimination by leaves (Δ) was determined as described previously (Farquhar et al., 1982; Evans et al., 1986):

• (Eqn 2)

where δ_ο and δ_e are the carbon isotope composition (¹³C/¹²C) of air exiting the chamber (δ_ο) or entering the chamber (δ_e). Carbon isotope composition was expressed against the carbon isotopic composition of the Pee Dee Belemnite formation (PDB), where δ = 1000 [((¹³C_sample/¹²C_sample)/(¹³C_PDB/¹²C_PDB)) – 1]. ξ = uC_e/(sA) where u is the molar flow rate through the chamber, C_e is the concentration of CO₂ entering the chamber, A is the rate of photosynthesis and s the projected leaf area. To determine if ¹³C-labelled amino acids were being metabolized and ¹³CO₂‘lost’ within leaves, it was necessary to fit data to the model of leaf-level carbon isotope discrimination (Farquhar et al., 1982; Evans et al., 1986). The discrimination of carbon isotopes during photosynthesis and respiration is a function of the different diffusivities of ¹³CO₂ and ¹²CO₂ and fractionation by enzymes (Farquhar et al., 1982; Evans et al., 1986):

• (Eqn 3)

where Δ = R_a/R_p– 1 and R_a and R_p are the molar ratios of ¹³CO₂ : ¹²CO₂ in the air and the photosynthetic product, respectively. In this model, discrimination is a function of the concentrations of CO₂ in air (C_a), at the leaf surface (C_s), in the intercellular air spaces (C_i) and in the chloroplast (C_c); and fractionations resulting from diffusion through the boundary layer (a_b, 2.9 ‰), diffusion through stomata (a, 4.4 ‰), and diffusion and dissolution of CO₂ into water (a_i, 1.8 ‰); net fractionation by Rubisco and PEP carboxylase (b, 28 ‰); fractionation resulting from mitochondrial respiration (e); and fractionation resulting from photorespiration (f). k is the carboxylation efficiency computed as (Farquhar et al., 1982) k = (A + R_d)/(C_i – Γ*), where Γ* is the CO₂ photocompensation point (38 μmol mol⁻¹ for a variety of Eucalyptus spp.; C. Warren unpublished). Fractionation caused by the boundary layer was negligible with our well-mixed leaf chamber and thus was omitted from the model. R_d, C_a and C_i were provided by the Li-6400. C_c was estimated based on the known relationship between net photosynthesis (A) and mesophyll conductance (g_m = max {0.05, 0.019 A}) and then by calculating C_c (C_c = C_i–A/g_m)( Warren, 2008; Douthe et al., 2011; Warren et al., 2011). f was assumed to be 11‰ (Lanigan et al., 2009). The model was solved for e for groups of five data points by using the Solver add-in of Microsoft Excel to find the value of e that provided the best fit to measured Δ. The carbon isotope composition of the respiratory substrate was then determined as described previously (Wingate et al., 2007).

¹⁵N uptake and fluxes of ¹²CO₂ and ¹³CO₂ after adding glycine or glutamine to soil

Eight seedlings of similar size were used in this experiment. Three seedlings received glutamine, with one seedling harvested at each of 60, 120 and 360 min. Five seedlings received glycine with one seedling harvested at each of 60, 360 and 5760 min, while two seedlings were harvested at 120 min. Collection of root and xylem sap samples involved termination of the experiment and ¹²CO₂ and ¹³CO₂ measurements, whereas leaf samples could be collected nondestructively and thus were collected at each time-point (until roots and xylem sap were collected). Leaf samples were collected 30, 60, 120 and 360 min after addition of amino acids. It was not possible to obtain replicate measurements of root and xylem sap for each time-point because of the large amount of expensive ¹³C- and ¹⁵N-labelled amino acid required to label the soil.

Plants were installed in the gas exchange system and fluxes of ¹²CO₂ and ¹³CO₂ from leaves and soil were recorded for at least 12 h before addition of 50.0 ml of 250 μg N ml⁻¹ (equivalent to 17.86 mM of N) U-¹³C,¹⁵N glycine or glutamine (99 atom %; Isotec Inc., Miamisburg, OH, USA). This amount of added amino acid N is proportional to what was used in previous experiments with E. pauciflora growing in the same soil (Warren, 2009a,b) after scaling for differences in mass of soil between studies. Seedlings were not watered for 1 d before solutions were injected into the soil. This let the soil dry to 50–100 ml below field capacity, thereby avoiding the risk of waterlogging. Solutions were injected into the soil 90–120 min after lights were turned on (i.e. while plants were actively photosynthesizing and transpiring). A total of 50 ml of solution was injected into the soil in five 10-ml aliquots using a 4-sideport Cass needle (18 gauge × 150 mm long; Victor-G & Company, Kanpur, India). It was not necessary to irrigate plants harvested within 360 min of amino acid addition, whereas the seedling harvested at 5760 min received c. 40 ml of water each day to replace what was lost via evaporation and transpiration (determined gravimetrically by weighing jars). To disentangle evaporation from soil from transpiration, I determined the rate of water loss of soil-filled jars without seedlings.

Collection of leaves, roots and xylem sap

Uptake of ¹⁵N, ¹³C and labelled amino acids was traced into roots, leaves and xylem sap. Leaf samples were collected from the leaves closest to the leaf enclosed in the chamber by punching 4–6 leaf discs (0.56 cm² each) into a 2-ml microfuge tube (Safe-Lock tube 2.0 ml; Eppendorf AG, Hamburg, Germany), immediately freezing in liquid N and subsequently storing at − 80°C. Following termination of ¹²CO₂ and ¹³CO₂ measurements, xylem sap was collected by cutting off a 10–15-cm length of shoot with at least five attached leaves, placing the shoot in a Scholander pressure chamber and exuding xylem sap by applying a slight over-pressure (0.05–0.1 MPa above the balancing pressure). The xylem sap was collected from the cut end of the shoot with a Pasteur pipette, transferred into a 2-ml microfuge tube, frozen in liquid nitrogen and stored at − 80°C. Roots were collected by transferring the entire jar of soil plus roots into a 1-mm sieve and rapidly separating roots from soil. A random subsample of roots (of all diameters) was washed three times with 50 mmol l⁻¹ KCl and three times with ultra-pure water, patted dry, placed into a 2-ml microfuge tube, frozen in liquid nitrogen and stored at − 80°C.

Measurement of ¹⁵N and ¹³C in leaves and roots

Leaves and roots were freeze-dried and weighed to the nearest 0.01 mg. Subsamples were ground to a fine powder with a bead mill (30 s at 25 Hz with a 5-mm stainless steel bead) (TissueLyser; Qiagen, Doncaster, VIC, Australia). A subsample was weighed into tin capsules and analysed for ¹³C and ¹⁵N by isotope-ratio mass spectrometry (IRMS) at the UC Davis Stable Isotope Facility. There was insufficient xylem sap to measure ¹⁵N and ¹³C by IRMS because of the large amount of xylem sap required for GC-MS measurements (see the next section).

Measurement of amino acids in leaves, roots and xylem sap

Leaf and root samples were extracted with methanol:chloroform:water and analysed by GC-MS of tert-butyldimethylsilyl derivatives as described previously (Warren, 2009b), except that acetonitrile replaced dimethylformamide as the derivatisation solvent. Amino acids were identified based on retention indices and mass spectra of authentic standards run under the same conditions. ¹⁴N,¹²C glycine was quantified from m/z 246; U-¹³C,¹⁵N glycine was quantified from m/z 249; ¹⁴N,¹²C glutamine was quantified from m/z 431; and U-¹³C,¹⁵N glutamine was quantified from m/z 438. Natural isotopes of the unlabelled amino acid did not contribute to peak area for the labelled amino acid. Each sample was analysed twice and the duplicate measurements were averaged. The flux of isotope labelled amino acids in the xylem was calculated by multiplying the measured concentration of amino acid in xylem sap (in nmol ¹⁵N ml⁻¹ xylem sap) by the transpirational flux of water (in ml h⁻¹) and then dividing by the dry mass of leaves. The cumulative amount of labelled amino acids delivered to leaves via the xylem was determined at 60, 120 and 360 min after isotope addition.

Results

Amino acid profile of roots, leaves and xylem

Twenty-one amino acids were present at quantifiable concentrations, but profiles were dominated by a handful of abundant amino acids (particularly Ala, Asp, Asn, Glu and Gln). There were large differences in amino acid profiles between xylem and leaves or roots (Fig. 2). Glutamine dominated the amino acid profile of xylem sap (73% of amino acids by mass), but was only 5–6.5% of amino acids in leaves and roots. Glycine was 3.7% of amino acids in leaves, 6.8% of amino acids in roots, and 0.2% of the amino acids in xylem sap.

Figure 2. Amino acid profiles of aqueous extracts of leaves and roots, and xylem sap of Eucalyptus pauciflora. Amino acids were determined by GC-MS (see the Materials and Methods section for details). Data are means of five replicate control plants.

Fluxes of ¹²CO₂ and ¹³CO₂ after adding glycine or glutamine to soil

Rates of soil ¹²CO₂ efflux differed among replicates (data not shown) and were 10–15% higher when lights were on and the plant photosynthesizing than when the lights were off (Fig. 3). ¹³CO₂ efflux was c. 1% of rates of ¹²CO₂ efflux and, in the absence of labelling, had a similar diel pattern. ¹³CO₂ efflux from soil increased dramatically within minutes of injecting U-¹³C,¹⁵N glycine (or glutamine) into the soil (Figs 3, 4), whereas ¹²CO₂ efflux was unaffected (Fig. 3). There was a general trend for rates of ¹³CO₂ efflux to peak c. 360 min after addition of isotope-labelled amino acid (Fig. 4). Most experiments were terminated at or before 360 min, but one soil sample receiving glycine was followed for 4 d (Fig. 3). In the 4 d after addition of U-¹³C,¹⁵N glycine, the cumulative amount of ¹³CO₂ efflux above baseline rates was 12 384 μg (1032 μmol) or c. 60% of the added ¹³C (20 400 μg; 1700 μmol).

Figure 3. Rate of efflux of ¹²CO₂ and ¹³CO₂ from the soil headspace of a 1.89-l glass jar containing c. 1.5 l of humic umbrosol with a seedling of Eucalyptus pauciflora. The 10–15% diel variation in rates of CO₂ efflux was associated with lights on (unshaded region) vs lights off (shaded region). At time zero, 50.0 ml of 250 μg N ml⁻¹ U-¹³C,¹⁵N -glycine (99 atom %) was added. Data are for one seedling.

Figure 4. Rate of efflux of ¹³CO₂ from the soil headspace of 1.89-l glass jars containing c. 1.5 l of humic umbrosol with a seedling of Eucalyptus pauciflora. At time zero, 50.0 ml of 250 μg N ml⁻¹ uniformly labelled ¹³C- and ¹⁵N-labelled glycine or glutamine (99 atom %) was added. Labelled amino acids were added at least 1 h after seedlings were illuminated. Data are for individual plants receiving glycine (five plants labelled: Gly 1,..., Gly 5) or glutamine (three plants labelled: Gln 1,..., Gln 3).

During illumination, rates of net photosynthesis varied among replicates between 4.5 and 6 μmol m⁻² s⁻¹, while in the dark the rate of respiration was 0.8–1.1 μmol m⁻² s⁻¹. Δ varied between c. 25 and 37‰. The carbon isotope composition of respiratory substrate differed by c. 10‰ between light and dark measurements, but in the absence of ancillary measurements it is unclear if this difference was real or an artefact of the model fit. There was no evidence that addition of uniformly ¹³C-labelled glycine or glutamine affected isotope composition of respiration (Table 1, Fig. 5).

Table 1.   Carbon isotope composition of the respiratory substrate of photosynthesizing leaves of Eucalyptus pauciflora before and after addition of 50.0 ml of 250 μg N ml⁻¹ U-¹³C,¹⁵N glycine (99 atom %) to the soil

	Glycine			Glutamine
	n	δ^13C respiratory substrate	Paired t-test, P	n	δ^13C respiratory substrate	Paired t-test, P
For each plant the average carbon isotope ratio (δ^13C) was determined for 30 min before application of amino acid, and 15–60 and 60–120 min after application of amino acid. Application of amino acids involved some manipulation of the aboveground part of the plant and gas exchange system, and for this reason the first 15 min of measurements after amino acid application were ignored. n, the number of replicate plants; P, the result of a paired t-test comparing pre- with post-application δ^13C.
30 min before	4	−31.3 (1.4)		3	−27.7(0.7)
15–60 min after	4	−29.0 (1.6)	0.14	3	−28.5 (1.6)	0.61
60–120 min after	3	−29.3 (1.2)	0.09	3	−28.2 (1.0)	0.62

Figure 5. Overall measured discrimination against ¹³CO₂ by a leaf of Eucalyptus pauciflora (Δ), calculated carbon isotope composition of respiratory substrate, rate of net photosynthesis and rate of transpiration. Data are for one leaf of a seedling of E. pauciflora growing in a 1.89-l glass jar containing c. 1.5 l of humic umbrosol. At time zero, 50.0 ml of 250 μg N ml⁻¹ U-¹³C,¹⁵N glycine (99 atom %) was added. A 10-point moving average was used for all data. Soil efflux data for the same seedling are shown in Fig. 3.

¹³C and ¹⁵N uptake

In roots, the maximum ¹³C labelling and largest ¹³C:¹⁵N ratio were for roots collected 60 min after glycine or glutamine addition. After 60 or 120 min there were no further increases in ¹³C excess despite increases in ¹⁵N excess. In soils receiving glycine there was no significant ¹³C or ¹⁵N excess of leaves until 5760 min (4 d), while in soils receiving glutamine there was ¹³C excess in leaves at all four time-points (30, 60, 120 and 360 min) but significant ¹⁵N excess was not detected in leaves until 360 min (Fig. 6).

Figure 6. The relationship between ¹⁵N excess and ¹³C excess of roots and leaves of Eucalyptus pauciflora growing in 1.89-l glass jars with c. 1.5 l of humic umbrosol. At time zero, 50.0 ml of 250 μg N ml⁻¹ uniformly labelled ¹³C- and ¹⁵N-labelled glycine or glutamine (99 atom %) was injected into the soil. Numbers indicate time elapsed (in minutes) since addition of glycine or glutamine. Error bars are 1 SE; see Table 1 for details of number of replicates.

Uptake of U-¹³C,¹⁵N amino acid

In roots, intact molecules of U-¹³C,¹⁵N glycine accounted for 30–100% of ¹⁵N excess at 60, 120 and 360 min after isotope addition, whereas in leaves U-¹³C,¹⁵N glycine was not detected – consistent with the nonsignificant ¹⁵N excess (Table 2). At 5760 min, intact molecules of U-¹³C,¹⁵N glycine were not detected in roots or leaves despite significant ¹⁵N excess. Intact molecules of U-¹³C,¹⁵N glycine were not detected in xylem sap.

Table 2.   Uptake of glycine and glutamine implied from isotope-ratio mass spectrometry (IRMS) measurements of total amounts of ¹⁵N and ¹³C (IRMS), and GC-MS measurement of intact U-¹³C,¹⁵N glycine and glutamine molecules in xylem sap and aqueous extracts of leaves and roots of Eucalyptus pauciflora

Time (min)	n	Roots			Xylem sap	n	Leaves
		U-^13C,^15N Gly (nmol ^15N g^−1)	^15N excess (nmol g^−1)	^13C excess (nmol g^−1)	U-^13C,^15N Gly (nmol ^15N ml^−1)		U-^13C,^15N Gly (nmol ^15N g^−1)	^15N excess (nmol g^−1)	^13C excess (nmol g^−1)
30						5	nd	−3 (2)	−50 (16)
60	1	219	333	458	nd	5	nd	−3 (2)	−56 (38)
120	2	172 (85)	314 (60)	284 (35)	nd	4	nd	0 (2)	−14 (27)
360	1	1566	1328	1157	nd	2	nd	9 (6)	15 (37)
5760	1	nd	3756	646	nd	1	nd	5151	−25

Time (min)	n	Roots			Xylem sap			n	Leaves
		U-^13C,^15N Gln (nmol ^15N g^−1)	^15N excess (nmol ^15N g^−1)	^13C excess (nmol ^13C g^−1)	U-^13C,^15N Gln (nmol ^15N ml^−1)	U-^13C,^15N Gln flux (nmol ^15N g leaves^−1 h^−1)	Cumulative U-^13C,^15N Gln flux (nmol ^15N g leaves^−1)		U-^13C,^15N Gln (nmol ^15N g^−1)	^15N excess (nmol ^15N g^−1)	^13C excess (nmol ^13C g^−1)
Data are ^15N and ^13C excesses relative to control plants or samples collected before ^15N addition. Data are means; the standard error is given where the number of biological replicates is > 1. n, the number of biological replicates. Data for root and xylem sap are means of two duplicate measurements on each biological replicate. nd, not detected.
30								3	nd	−2 (2)	1020 (279)
60	1	96	3245	4556	185	77	77	3	nd	1 (3)	791 (216)
120	1	63	1916	2831	115	48	125	2	nd	13 (7)	717 (46)
360	1	185	3575	4227	57	23	173	1	nd	557	1580

In roots, intact molecules of U-¹³C,¹⁵N glutamine accounted for < 5% of ¹⁵N excess at 60, 120 and 360 min after isotope addition. Significant amounts of U-¹³C,¹⁵N glutamine were detected in xylem sap and, based on known whole-plant transpiration rate, the cumulative amounts of U-¹³C,¹⁵N glutamine transported to leaves were 77 nmol ¹⁵N g⁻¹ leaves at 60 min and 173 nmol ¹⁵N g⁻¹ leaves at 360 min. Surprisingly, U-¹³C,¹⁵N glutamine was not detected in leaves despite the large amounts of U-¹³C,¹⁵N glutamine transported in the xylem sap.

Discussion

Contrasting patterns of metabolism of glycine and glutamine

In recent years there has been a growing acceptance that plants can take up amino acids, but there is still the unanswered question of quantitative significance (Näsholm et al., 2009). One of the reasons that quantification is difficult is because interpretation of isotope labelling is dependent on knowing the fate of amino acids following uptake. The results of this study support the hypothesis that there are fundamental differences between amino acids in terms of what happens after uptake and this affects patterns of isotope labelling. In the 60–360 min after addition of U-¹³C,¹⁵N glutamine to the soil, intact molecules of U-¹³C,¹⁵N glutamine accounted for < 5% of ¹⁵N excess in roots, whereas U-¹³C,¹⁵N glycine accounted for 30–100% of ¹⁵N excess in roots. The striking difference between glycine and glutamine is possibly because glutamine is metabolized faster than glycine (e.g. 200-fold faster flux of labelled glucose to Gln than Gly in tubers of potato (Solanum tuberosum); Roessner-Tunali et al., 2004).

Post-uptake metabolism makes it difficult to estimate the uptake rate of intact amino acids. The uptake of amino acids is most commonly determined from addition of double-labelled (¹³C and ¹⁵N) amino acids and IRMS analysis of ¹³C and ¹⁵N in plant tissues. Intact uptake is implied if there is significant ¹³C and ¹⁵N labelling and the slope of the correlation of ¹³C to ¹⁵N excess is the same as in the intact amino acid (e.g. 2 ¹³C : 1 ¹⁵N in the case of added U-¹⁵N,¹³C Gly). However, the correlation of ¹³C with ¹⁵N is rather weak evidence for intact uptake (Jones et al., 2005a,b). The presence of intact labelled amino acid molecules within plant tissues provides far stronger evidence for intact uptake, but there are commonly large discrepancies between amounts of intact amino acid and total ¹⁵N (or ¹³C). In most cases authors make the ‘leap of faith’ that lower-than-expected amounts of intact amino acid are caused by post-uptake metabolism leading to transfer of ¹⁵N and/or ¹³C to other molecules (Warren, 2009a,b); however, the same discrepancy could also arise if metabolism occurred pre-uptake(i.e. in soil) such that ¹⁵N and ¹³C were taken up independently. To disentangle these possibilities requires quantification of fluxes and pools of amino acids in roots, and/or that post-uptake metabolism is minimized by very rapid harvests (i.e. within minutes of adding amino acid to soil).

The transport of some amino acids to leaves and subsequent metabolism mean that interpretation of excess ¹⁵N and ¹³C in leaves is even more problematic than for roots (Persson & Näsholm, 2001). For example, glutamine is one of the main transport forms of N in E. pauciflora (and other Eucalyptus spp.; Marsh & Adams, 1995; Pate et al., 1998) and was transported to leaves (at least partially) as an intact molecule. Once it reached the leaves it must have been rapidly metabolized because U-¹³C,¹⁵N glutamine was not detected in leaves despite the large amounts of U-¹³C,¹⁵N glutamine in the xylem sap (Table 2). Interestingly, ¹³C from glutamine was not ‘lost’ from leaves as ¹³CO₂ because ¹³C excess of leaves remained elevated and the tunable diode laser measurements did not detect changes in isotope signature of the respiratory substrate (Table 1). The absence of a significant effect on isotopic composition of leaf respiration is consistent with the suggestion that amino acids are metabolized primarily via transamination reactions (Schmidt & Stewart, 1999; Näsholm et al., 2009). By contrast, glycine-N was probably transported after deamination in roots because U-¹³C,¹⁵N glycine was not detected in xylem sap (Table 2), leaves did not contain excess ¹³C (Fig. 5) and the isotope signature of the leaf’s respiratory substrate was unchanged by addition of U-¹³C,¹⁵N glycine (Table 1). These findings indicate that amino acids may be metabolized within roots before loading into the xylem (e.g. all glycine and some glutamine) and after transport to leaves (glutamine).

In many cases it is sufficient to know the minimum uptake rate or that some amino acid is taken up intact, but to compare uptake of different amino acids requires quantification of absolute or relative rates of uptake. Unfortunately, standard protocols (most commonly ¹³C,¹⁵N labelled amino acid, and roots harvested ≥ 60 min after labelling) are unlikely to yield data that are useful for relative or absolute quantification. The first problem is that amino acids and/or amino acid N and C are already being transported out of roots within 60 min of applying label to soil, and this pattern of export will probably differ among amino acids (e.g. Gly vs Gln). The second problem is that the putative differences in metabolism of amino acids (e.g. Gly vs Gln) lead to a large and variable discrepancy between amounts of intact amino acids vs ¹⁵N (or ¹³C) excess (Table 2), and there is no rational way of knowing whether intact amino acid or ¹⁵N (if either) gives the correct estimate of uptake. Hence, comparisons of uptake of different amino acids will almost certainly be confounded by differences in post-uptake metabolism and transport. At longer time scales (e.g. > 48 h used to contrast uptake of different amino acids by, for example, Weigelt et al., 2003, 2005; Harrison et al., 2007, 2008) isotope labelling will be even more strongly affected by the patterns of transport and metabolism that differ among amino acids.

The power and limitations of combining multiple technologies

One of the aims of this work was to demonstrate the power of combining multiple technologies to trace the fate of added isotope label. Combining IRMS measurements of ¹³C and ¹⁵N with GC-MS measurements of intact amino acids demonstrated rapid metabolism of added amino acids, and large differences between glycine and glutamine in metabolism and transport. The tunable diode laser system enabled real-time measurements of ¹²CO₂ and ¹³CO₂ exchange and demonstrated that loss of ¹³CO₂ from leaves is an insignificant pathway (Table 1). The ability to perform long-term real-time measurements in illuminated or darkened leaves is a significant advantage compared with the more traditional approach of placing the plant or soil in a darkened, enclosed system and trapping evolved CO₂ with hydroxide (Hawkins et al., 2005). Continuous measurements of ¹³CO₂ efflux from the soil headspace holds great promise as a tool for investigating the fate of added ¹³C label (e.g. to investigate plant–microbe competition for added amino acid). To exploit the full power of the system it would be useful to make parallel measurements in soil-free systems so as to disentangle root from soil metabolism.

Double-labelled amino acids such as glutamine are expensive and this meant that it was not feasible to obtain multiple replicates for each time-point. However, limited replication was not a problem because biological and analytical variability (e.g. as indicated by replicate measurements of δ¹³C of respiratory substrate; Table 1; ¹³C, ¹⁵N and intact amino acids in leaves; Table 2) were many times smaller than the differences in key contrasts (e.g. 100-fold difference between glycine and glutamine in ¹³C or ¹⁵N uptake; up to 100-fold difference between ¹⁵N or ¹³C and intact amino acid). Lessons learned in this and other studies suggest that there is little value in using expensive double-labelled amino acids and future experiments may be able to use less expensive isotope labels (e.g. U-¹³C). Use of double-labelled amino acids has been recommended for GC-MS analyses of amino acid uptake so that mass spectral fragments representing natural isotopes of the unlabelled amino acid do not overlap with the labelled amino acid (Persson & Näsholm, 2001). With cheaper singly labelled amino acids it is possible to limit overlap of GC-MS mass spectral fragments by using monisotopic derivatizing reagents (e.g. acylation reactions), or by using an alternative analytical platform that does not require derivatization of samples (e.g. LC-MS). Another incentive for exploring alternative analytical methods is that many methods are more sensitive than GC-MS of t-BDMS (tert-butyldimethylsilyl) derivatives (e.g. GC-MS with negative chemical ionization, and LC-MS), and thus better detection limits permit use of smaller amounts of label and/or shorter labelling periods.

The results of this study are broadly indicative of patterns of amino acid uptake by seedlings, but extrapolation to other species and the field should be done cautiously. The marked differences between glycine and glutamine may not be reflected in all species because of species differences in post-uptake metabolism and transport of amino acids (e.g. depending on major transport forms of N and sites of metabolism). Results may also differ depending on mycorrhizal status. Seedlings were visually confirmed to be ectomycorrhizal and studies on other Eucalyptus spp. have shown that this may affect their ability to grow on amino acids (Turnbullet et al., 1995). Hence, it is plausible that the observed patterns of post-uptake metabolism of amino acids are at least partially attributable to mycorrhizal fungi, and could vary if plants were amycorrhizal or had differing symbionts.

Conclusions

Determining whether or not amino acids are taken up intact requires knowledge of what happens after uptake. Many things can happen to an amino acid taken up by a plant. The fidelity with which amounts of ¹⁵N, ¹³C or intact amino acid reflect actual rates of uptake will depend on the specific metabolic reactions (deamination, transamination, protein synthesis, etc.), their location (roots, xylem or shoots), the circulation of amino acid derived N and C within the plant, and the length of time between isotope addition and harvest. This study has demonstrated that there are large differences between glycine and glutamine in patterns of post-uptake metabolism and within-plant transport. Differences between amino acids in metabolism and transport will confound simple interpretations of isotope labelling such that root and shoot contents of intact amino acid, ¹³C and ¹⁵N may not accurately reflect rates of uptake.

In future experiments it may be possible to avoid confounding estimates of uptake with patterns of post-uptake metabolism and transport by quantifying rates of amino acid turnover within roots (e.g. as has been done for potato tubers; Roessner-Tunali et al., 2004). Determining rates of turnover would help clarify whether a small pool of intact labelled amino acid within roots is a result of plants taking up amino acids as intact molecules that are rapidly metabolized, or plants taking up the mineralization products of amino acids. Unfortunately, determining the turnover of amino acids is not trivial and thus is unlikely to be routine. A pragmatic approach is to limit the problem of metabolism by performing harvests within minutes of adding isotope-labelled amino acids. A drawback of rapid harvest is that the amount of label within tissues would be small; however, this problem can almost certainly be overcome with analytical methods (e.g. GC-MS with NCI ionization, and LC-MS) that are more sensitive than those normally used to quantify uptake.

Acknowledgements

This work was supported by funding from the Australian Research Council (LP0989129, LE0882935) and Forestry Tasmania. Sam Ruggeri manufactured the custom leaf chamber for the Li-6400.