Free amino acid, ammonium and nitrate concentrations in soil solutions of a grazed coastal marsh in relation to plant growth

Authors


Correspondence: H. A. L. Henry. Fax: +1 416 978 5878; e-mail: hhenry@botany.utoronto.ca

Abstract

Soluble free amino acids, ammonium and nitrate ions as sources of nitrogen for plant growth were measured in soils of a coastal marsh grazed by snow geese in Manitoba, Canada. Amounts of nitrogen, primarily ammonium ions, increased in the latter half of the growing season and over winter, but fell to low values early in the growing season. Free amino acid concentrations relative to ammonium concentrations were highest during the period of rapid plant growth in early summer, especially in soils in the intertidal zone, where the median ratio of amino acid nitrogen to ammonium nitrogen was 0·36 and amino acid concentrations exceeded those of ammonium ions in 24% of samples. Amino acid profiles, which were dominated by alanine, proline and glutamic acid, were similar to goose faecal profiles. In a continuous flow hydroponic experiment conducted in the field, growth of the salt-marsh grass, Puccinellia phryganodes, on glycine was similar to growth on ammonium ions at an equivalent concentration of nitrogen. When supplies of soil inorganic nitrogen are low, amino acids represent a potentially important source of nitrogen for the re-growth of plants grazed by geese and amino acid uptake may be as high as 57% that of ammonium ions.

Introduction

Although soil inorganic nitrogen (N) is often the primary source of plant N (Glass et al. 1999; Murphy et al. 2000), soluble organic nitrogen in soils may also be utilized by plants (Lawes & Gilbert 1881). In Arctic and alpine soils, for example, low temperatures, a short growing season and often low precipitation slow rates of N mineralization, which results in the accumulation of organic N, including free amino acids which can be readily taken up by plants (Kielland & Chapin 1992; Nadelhoffer et al. 1992; Chapin, Moilanen & Kielland 1993; Kielland 1994; Atkin 1996; Raab, Lipson & Monson 1996, 1999). In addition, in some polar soils, these environmental effects on availability and uptake of nitrogen by plants may be further exacerbated by soil salinity. The detrimental effects of sodium chloride on inorganic N acquisition are both direct and indirect; sodium chloride inhibits nitrification (McClung & Frankenberger 1985; Wilson et al. 1999) and the presence of chloride ions inhibits nitrate uptake by roots (Cram 1973; Deane-Drummond & Glass 1982). Although free amino acids may be an important alternative source of nitrogen for plants growing in saline soils, particularly when inorganic N availability is limited, data on amino acid availability in these soils are lacking, and the ability of halophytes to grow on amino acids as the sole source of nitrogen has not been assessed.

In the present study, concentrations of free amino acids, ammonium and nitrate ions were determined in soil solutions of an Arctic coastal marsh in which lesser snow geese (Chen caerulescens caerulescens L.) heavily graze the dominant grass Puccinellia phryganodes that has a diminutive growth form (Jefferies 1988a, b). The growth of the vegetation is nitrogen-limited (Cargill & Jefferies 1984a) and the addition of goose faecal droppings to swards dominated by Puccinellia results in increased standing crop within the growing season (Bazely & Jefferies 1985; Hik & Jefferies 1990). Much of the soluble faecal nitrogen is leached from droppings within 48 h (Bazely & Jefferies 1985; Ruess, Hik & Jefferies 1989). Faecal inputs can be high (≤ 22 m−2 week−1) during the growing season and droppings are widely dispersed (Jefferies & Rockwell 2002). Results from an experimental study indicated that although re-growth of Puccinellia swards following defoliation by geese was dependent on droppings, estimated amounts of soluble inorganic nitrogen in faeces were inadequate in some grazing treatments to account for the N sequestered in new plant growth (Hik, Sadul & Jefferies 1991). A likely source of additional N in faecal material is the presence of free amino acids. Consequently, we have compared free amino acid profiles in soil and faeces and determined the ability of Puccinellia to grow on an amino acid as the sole source of N.

The following questions were addressed. (1) How do seasonal changes in the concentrations of free amino acids in the soil solution compare with those of ammonium and nitrate ions? (2) Are there differences in the N profiles between intertidal sites that are flooded with sea water in late summer and autumn and supratidal sites which are flooded no more than twice every 3 years? (3) How much temporal and spatial variation in amino acid profiles is present at a site and are the free amino acid profiles of faecal extracts and soil solutions similar? (4) Is the growth response of tillers of Puccinellia phryganodes grown in continuous flow hydroponic solutions, in which either glycine or ammonium ions is the sole source of N, similar?

Materials and methods

Study sites

The study was conducted in a grazed Arctic coastal salt marsh at La Pérouse Bay, Wapusk National Park, located approximately 30 km east of Churchill, Manitoba, Canada (58°43′ N, 94°26′ W), on the south-west coast of Hudson Bay. The marsh can be subdivided into an intertidal zone, that is tidal in late summer and autumn, but not from snow melt until late July, and a supratidal zone that is rarely flooded with sea water. The vegetation of the intertidal marsh is dominated by Puccinellia phryganodes (Trin.) Scribn. and Merr. and Carex subspathacea Wormskj, which form discontinuous swards because of early spring grubbing by lesser snow geese that leads to sward destruction (Jefferies 1988a, b). Net above-ground production of intact swards has been estimated to be between 100 and 150 g m−2 year−1 (Cargill & Jefferies 1984b; Hik & Jefferies 1990). Further inland, the supratidal zone also is dominated in low-lying areas by swards of P. phryganodes and C. subspathacea, but low shrubs, predominately Salix brachycarpa Nutt., and grasses, especially Festuca rubra L and Calamagrostis deschampsioides Trin., colonize frost-heave sites. The nomenclature follows Porsild & Cody (1980).

The soils were classified as Regosolic Static Cryosols (Agriculture Canada Expert Committee on Soil Survey 1987, Wilson & Jefferies 1996). Soils in the intertidal zone have a greyish mineral horizon of marine sediment in which the top 1–2 cm contains organic material. The soils are often highly reduced (Eh − 50 mV) in spring, especially where they are close to drainage channels (Wilson & Jefferies 1996). In the supratidal marsh, a mineral base is covered by 3–4 cm of dark brown-black highly humified organic material. Soil pH averages 7·12 ± 0·04 (n = 24) in both the intertidal and supratidal marshes. The salinity of extracted soil solutions can rise to as high as 40 g Na+ L−1 in late summer in intertidal sites devoid or nearly devoid of vegetation. However, beneath intact swards the salinity of the bulk soil solution in summer is at or less than that of sea water (approximately 12 g Na+ L−1) (Srivastava & Jefferies 1995; Wilson & Jefferies 1996).

Soil and faecal material sampling

Soils were sampled from rooting zones of intact Puccinellia swards at three tidal and three supratidal sites (> 1 km distance between sites) on nine sampling dates starting 29 May 1999 and ending 30 July 2000 (refer to Fig. 1 for sampling dates). On 1 November 1999 and 29 April 2000, only two and five sites were sampled, respectively, as the other sites were snow or ice-bound. Spring melt occurred in late May or early June and the soils froze again in early October. Permafrost was 30–50 cm beneath the surface (Wilson & Jefferies 1996). Within each site, three replicate samples (15 cm × 15 cm × 3 cm deep) were collected approximately 50 m apart. Samples were kept cool (3–5 °C) and extracted within 24 h by squeezing (and by centrifugation on 1 July and 21 July 1999). Results from squeezing were compared with those based on water extraction using a ceramic cup tensiometer (Model 1905; Soil Moisture Equipment Corp., Goleta, CA, USA) and data sets were similar for concentrations of individual amino acids (r2 = 0·95). Raber et al. (1998) have shown that centrifugation is a suitable method for obtaining soil solutions. Fresh goose faeces (n = 6) were collected from both intertidal and supratidal marshes on 30 June and 21 July 1999 and extracted in 60% ethanol (15 g FW in 75 mL) for 6 h. For all analyses, latex gloves were worn and glassware was acid-washed to minimize contamination. Appropriate blanks and standards were used for all analyses to correct for potential contamination and deionized water washings from gloves yielded no amino acid contamination. Extracts were filtered through Whatman GF/A glass filter paper (Whatman, Maidstone, UK) and frozen at −20 °C until analysis.

Figure 1.

Seasonal trends in persulphate oxidation estimates of total soluble nitrogen in soil solution samples collected from (a) the intertidal zone and (b) the supratidal zone of a salt-marsh at La Pérouse Bay, Manitoba, over 14 months. Each bar represents the mean of three sites (three replicates chemically pooled per site) and lines denote standard error.

Soluble nitrogen analyses

Free amino acids in soil solutions, and faecal extracts were separated by reverse-phase high-performance liquid chromatography (HPLC) through two Waters (Milford, MA, USA) PICO-TAG columns (3·9 mm × 300 mm; packing material 3 µm NOVAPAK) at a temperature of 46 °C, following the method of Bidlingmeyer et al. (1987). The samples were precolumn derivatized using a methanol–water–triethyamine–phenylisothiocyanate (7 : 1 : 1 : 1) reagent to form phenylthiocarbamyl derivatives of the amino acids, eluted with a Waters PICO-TAG gradient and fluorescence was measured at 254 nm.

Ammonium and nitrate (as nitrite) concentrations were determined colormetrically following reaction with phenol–sodium nitroprusside (Solorzano 1969) and Marshall’s reagent, respectively, using an auto-analyser (Pulse Instrumentation Ltd, Saskatoon, Saskatchewan, Canada). Nitrate was reduced to nitrite prior to analysis (Keeney & Nelson 1982). The total soluble nitrogen was measured for pooled samples at a site, on each sampling date, using alkaline persulphate oxidation (Cabrera & Beare 1993), followed by analysis for nitrate (as nitrite). By subtracting ammonium, nitrate and total free amino acid concentrations from the total nitrogen concentration, an estimate of the concentration of unidentified soluble nitrogen sources was obtained. The fraction of unidentified soluble nitrogen that comprised proteins and peptides was estimated for a small subset of samples (n = 3) by evaporating and hydrolysing aliquots of samples prior to HPLC analysis.

Continuous flow experiment

A continuous flow hydroponic experiment was assembled under field conditions in an experimental garden in the supratidal marsh at La Pérouse Bay. Plant material was collected from a single patch to minimize genetic variation. Individual tillers (shoots approximately 2·5 cm long and roots approximately 5 cm long) of P. phryganodes, that had only actively growing white roots produced within the season, were placed in 250 mL culture vessels sunk into the ground. The roots were bathed in one-tenth strength N-free Hoagland’s solution containing 50 mm NaCl, to which was added the appropriate nitrogen source to give final concentrations of either 10, 100 or 1000 µm of ammonium or 100 µm of glycine. The solutions were gravity-fed from 45 L polypropylene reservoirs into sets of six culture vessels, drained into 40 L reservoirs sunk into the ground and fed back to the 45 L reservoirs via Rio 180 powerhead pumps (TAAM Inc., Camarillo, CA, USA). The solutions in the reservoirs were well agitated and the flow rate into culture vessels of 350 mL min−1 ensured that the solutions were replaced every 45 s. Reservoirs and culture vessels were connected via Tygon 80-H tubing and the culture vessels were positioned randomly within three blocks. The solutions were changed every second day to avoid bacterial growth and to ensure a nitrogen depletion of less than 5%, and light was excluded from culture vessels, reservoirs and tubing to stop algal growth. The solutions remained clear at all times and examination of roots under a light microscope showed no accumulation of slime on roots and no evidence of the build-up of bacterial plaque on the root surface under high magnification (600×).

Tillers were transplanted on 22 June 2000 soon after melt and plants were harvested on 5 July, 20 July and 4 August (six replicates per nutrient treatment per harvest). Root and shoot materials were washed in deionized water, dried at 60 °C for 4 d and then weighed and ground. The nitrogen content was determined using Kjeldahl digestion followed by analysis for ammonium concentrations as indicated above.

Sub-samples of root material were collected and preserved in 60% ethanol for estimations of root surface area in order to calculate net nitrogen fluxes into roots during the experiment. The root surface area was estimated by measuring the length and diameter of primary and secondary roots and roots hairs under a light microscope. For diameter measurements, the roots were immersed in glycerine and diameter, root hair length and number were quantified for 0·4 mm long segments at 3 mm intervals along the root length (n= 2 tillers per treatment per sampling date).

In a separate study, transpiration rates were determined in order to assess whether mass flow to the root surface could alone account for the estimated net fluxes of N. Water loss from sealed culture vessels with tillers was measured over 24 h and measurements were made four times during the season using six replicates at a range of air temperatures and light conditions. Leaf surface area was estimated in a similar manner as root surface area on the basis of leaf length and incremental diameter measurements.

Data analyses

Mean free amino acid, ammonium and nitrate concentrations were compared between the intertidal and supratidal marshes on each sampling date using a Tukey HSD multiple comparisons test. This test is an exact alpha-level test when sample sizes are equal (Tukey 1953).

For the continuous flow experiment, mean responses and standard errors for all treatments at each harvest date were estimated and plotted. Full factorial anova models were used to test for significant differences between harvest dates and treatments for plant dry weight, N content of plants as a percentage of dry weight (both log-transformed) and the root : shoot ratio. Within harvest dates, treatment differences were assessed using pairwise t-tests. In all cases, blocks were pooled due to a lack of significance of block effects (P > 0·75).

For each treatment, regression lines were fitted to log–linear plots of estimated root surface area and cumulative total N uptake over time. Regression equations were used to estimate the daily flux of N per unit of root surface. Estimates of daily N uptake were used to calculate the rate of transpiration required to supply sufficient N to the root surface only via mass flow assuming steady-state conditions (Tinker 1969). These rates of transpiration were also expressed as the required transpiration rate per unit leaf area, in order to meet the estimated net N influx and compared with measured transpiration rates.

Results

Temporal and spatial trends in soluble N

For both intertidal and supratidal sites, concentrations of soluble N were relatively low at the beginning of the growing season after the snow had melted. They increased later in the growing season (late July in both 1999 and 2000) and over the winter, and fell back to low values at the beginning of the following growing season (Fig. 1). The seasonal trends in concentrations of soluble N were determined largely by the presence of ammonium ions, which, on average, were the dominant form of soluble nitrogen (Fig. 2). Although the total free amino acid concentrations rose over the winter in the intertidal zone and declined in the supratidal zone, seasonal trends in free amino acids (and nitrate concentrations) were less pronounced than those of ammonium ions.

Figure 2.

Seasonal trends in free amino acids, ammonium and nitrate concentrations in soil solution samples collected from (a) the intertidal zone and (b) the supratidal zone of a salt-marsh at La Pérouse Bay, Manitoba, over 14 months. Each bar represents the mean of three sites (three replicates per site) and lines above the bars denote standard error. Common lower case letters denote a lack of a significant different between soluble nitrogen species within a given sampling date (Tukey’s HSD test).

The proportion of total free amino acids relative to inorganic nitrogen was highest during the growing season, when ammonium concentrations were at their lowest (Fig. 2). This proportion was also highest for intertidal sites, where ammonium and total free amino acid concentrations ranged from 53 to 203 µm N and 25–47 µm N, respectively (nitrate concentrations ranged between 8 and 37 µm N) and the median proportion of total amino acids to ammonium was 0·36 throughout the growing season. Due to both high spatial heterogeneity and a lack of dependence between ammonium and total amino acid concentrations, this ratio varied widely between samples (Fig. 3). As a result, amino acids were present at concentrations greater than those of ammonium ions in 24% of soil extracts sampled from intertidal sites during the growing season.

Figure 3.

Distribution of total free amino acid nitrogen as a proportion of ammonium nitrogen in soil solution samples collected during the growing seasons of 1999 and 2000. Samples are grouped and ordered within sites in (a) the intertidal zone and (b) the supratidal zone of a salt-marsh at La Pérouse Bay, Manitoba.

Mean concentrations of unidentified soluble nitrogen compounds were higher in intertidal sites (188 ± 37 µm nitrogen, n= 25) than in supratidal sites (77 ± 19 µm nitrogen, n= 25), where they accounted for 42 and 24% of the total soluble nitrogen pools, respectively. From 52 to 98% of unidentified soluble nitrogen compounds were proteins and peptides. For both intertidal and supratidal sites, concentrations of unidentified soluble nitrogen compounds were highest during the spring melt (30 May−10 June 2000; spring melt soil water was not sampled in 1999, as the melt occurred in April before our arrival).

Soil amino acid profiles were dominated by alanine and to a lesser extent proline and glutamic acid (Fig. 4). Other amino acids present at relatively high concentrations included gamma-aminobutyric acid and valine, that were prevalent in intertidal sites, and leucine, lysine and tyrosine, which were present at all sites. Lysine concentrations were highly correlated with air temperature; they were high throughout the growing season and low in winter in both intertidal and supratidal sites (Fig. 4). In the intertidal marsh, other amino acids, such as valine and glycine, showed the opposite trend and high concentrations were present at low temperatures. However, these increases were not proportional increases, as the total amino acid pool also increased at this time. The free amino acid profile of faecal matter extracts collected in 1999 closely resembled that of soil solution collected over the same period (Fig. 5). Concentrations of individual amino acids expressed as a proportion of total amino acid concentration were statistically indistinguishable between faecal extracts and soil solutions for 16 of the 22 amino acids tested, although concentrations did differ significantly between soil and faeces for glutamic acid, glycine, lysine, phenylalanine, tryptophan and valine (Fig. 5).

Figure 4.

Mean concentrations of amino acids present in bulk soil solutions of (a) the intertidal zone and (b) the supratidal zone of a salt-marsh at La Pérouse Bay, Manitoba. Data are pooled for sampling dates where the mean soil temperature was greater than 2 °C (black bars: 13 June, 1 July, 21 July 1999 and 28 June, 30 July 2000) and less than 2 °C (white bars: 29 May, 1 November 1999 and 29 April, 6 June 2000). Amino acid key for Figs 4 and 5: asp, aspartic acid; glu, glutamic acid; asn, asparagine; ser, serine; gln, glutamine; gly, glycine; his, histidine; arg, arginine; gaba, gamma-aminobutyric acid; thr, threonine; ala, alanine; pro, proline; aaba, alpha-aminobutyric acid; tyr, tyrosine; val, valine; met, methionine; cys, cysteine; ile, isoleucine; leu, leucine; phe, phenylalanine; try, tryptophan; lys, lysine.

Figure 5.

Mean concentrations of amino acids present in faecal extracts (black bars) and bulk soil solution (white bars) expressed as a percentage of the total free amino acid concentration for samples collected on 30 June and 21 July 1999. Error bars denote standard error. Significant differences between faecal extract and soil solution percentages as determined using pairwise t-tests denoted as follows: *P < 0·05, **P < 0·01 and ***P < 0·001. Refer to Fig. 4 for amino acid key.

Continuous flow uptake experiment

At the final harvest, biomass of P. phryganodes tillers grown in a solution of 100 µm glycine was comparable to the biomass of tillers grown at an equivalent ammonium concentration (Fig. 6a). In response to a higher ammonium concentration, tillers increased biomass production and nitrogen uptake, although total nitrogen uptake per plant at the final harvest appeared to saturate between external ammonium concentrations of 100 and 250 µm (Fig. 6b). Total nitrogen uptake per plant was comparable for tillers grown on both 100 µm glycine and ammonium (Fig. 6b). Tissue N as a percentage of dry weight increased until day 28 then levelled off between days 28 and 43 for all ammonium treatments. It continued to increase for the glycine treatment, which reached the same percentage N-values by day 43 as those in tissues of plants grown at the two highest ammonium treatments (Fig. 6c). Root : shoot ratios decreased sharply between days 28 and 43 for all treatments with the exception of the 10 µm ammonium treatment, in which the root : shoot ratio remained high (Fig. 6d). From the results of the continuous flow experiment, we can estimate that soil amino acids contribute up to 57% of the nitrogen taken up by swards of Puccinellia in the intertidal zone during the snow-free season relative to ammonium N. This estimate is obtained by calculating the total uptake of ammonium and glycine N per tiller over 43 d as a function of the external concentrations of N (Fig. 6b). A constant proportion of glycine uptake relative to ammonium uptake was assumed over all concentrations based on the proportion at 100 µm. Values of N uptake were adjusted to reflect uptake expected when the concentrations of these N sources in the bulk soil solution were expressed as mean values for the growing season (ammonium: 103 µm, amino acids: 39 µm). Mean nitrate concentrations in intertidal soils were low (20 µm) and nitrate uptake was not estimated as the growth response to nitrate was not determined experimentally.

Figure 6.

Plots of (a) total dry weight versus day; (b) total nitrogen at final harvest versus N concentration in the growth solution; (c) whole plant percentage N versus day; and (d) root : shoot ratio versus day for individual tillers of Puccinellia phyryganodes grown in hydroponic culture at fixed nitrogen concentrations in the field at La Pérouse Bay, Manitoba, from 22 June to 4 August 2000. Error bars denote standard error (n = 6). Full factorial anova results are as follows: (a) total dry weight versus day (F = 210, P < 0·001), treatment (F = 2·86, P = 0·044), day * treatment (F = 2·00, P = 0·124) (c) %N versus day (F = 33·9, P < 0·001), treatment (F = 8·55, P < 0·001), day * treatment (F = 0·51, P = 0·678) (d) root : shoot versus day (F = 11·85, P = 0·001), treatment (F = 3·44, P = 0·022), day * treatment (F = 3·24, P = 0·028). Lower case letters denote treatment differences within sampling dates as determined using pairwise t-tests. Curve fitted to data for total N uptake versus N concentration in (b) was obtained using OLS regression: total N uptake = 0·347 + 2·87(1 − e−0·0178 concentration).

Estimates of the mean net nitrogen flux across the root surface necessary to sustain mean tissue concentrations ranged between 0·17 and 1·26 µmol cm−2 d−1 for plants in the solutions of 10 and 1000 µm ammonium chloride, respectively (Table 1). Estimated mean rates of transpiration from leaves required to supply sufficient nitrogen to the root surface solely as a result of mass flow were 14 and 0·20 g H2O cm−2 d−1, respectively, for these plants (Table 1). These values greatly exceeded the observed transpiration rates in the field, which ranged from 0·005 to 0·04 g H2O cm−2 d−1, indicating that diffusion as well as mass flow influences the nitrogen supply rate to the root surface.

Table 1.  Range of estimates for daily nitrogen flux across root surface and transpirational flow required to maintain a constant net supply of nitrogen to the root surface via mass flow for tillers of P. phryganodes grown in hydroponic culture
Daily nitrogen flux (µmoles N cm−2 d−1)
10 µm NH4Cl100 µm NH4Cl1000 µm NH4Cl100 µm glycine
0·17–0·580·62–1·050·49–1·260·45–0·87
Required transpirational flow (g H2O cm−2 d−1)
10 µm NH4Cl100 µm NH4Cl1000 µm NH4Cl100 µm glycine
7·6–14·01·6–2·90·20–0·221·9–2·4

Discussion

Temporal and spatial trends in soluble N

Despite seasonal and between-site variation in free amino acid concentrations in soils, the size of the soluble N pool was determined primarily by changes in the concentration of ammonium ions, the dominant form of soluble N. Nitrate concentrations were particularly low as a percentage of total soluble N (Fig. 2). However, during much of the early and mid-growing season when there was a rapid increase in standing crop, ammonium concentrations in soil solutions were also low. Free amino acids contributed most to the soluble nitrogen pool at this time, especially in intertidal sites, where N present in amino acids exceeded NH4+-N in nearly one-quarter of samples collected during the growing season (Fig. 3). Tillers may exploit soil patches rich in amino acids in order to meet the N demand for growth following defoliation by geese. The degree to which this occurs is dependent on both the extent of amino acid rich patches and the ability of plant roots to exploit these microsites (Hodge et al. 2000). Although the relative contribution of free amino acids to the soluble N pool is lower than that found in other Arctic sites (Chapin et al. 1993; Kielland 1995), where free amino acids exceed inorganic N by an order of magnitude, the results are comparable to those reported for alpine sites (Raab et al. 1996, 1999), and they far exceed the relative contribution of free amino acids to the N pool in temperate or subtropical soils (Schobert & Komor 1987, Schobert, Kockenberger & Komor 1988, Turnbull et al. 1996). In Arctic soils where free amino acids exceed soluble ammonium ions by an order of magnitude, enhanced extracellular protease activity at low soil pH may be responsible for the high levels (Kielland 1995).

In both intertidal and supratidal sites, soluble N concentrations rose during late summer and in winter (Fig. 1). Increases in soluble N in winter may be linked to freeze–thaw events (16 per year, on average, in the Churchill area; cf. R. Bello, unpublished results) brought about by the lysis of bacterial and root cells (Skogland, Lomeland & Goksoyr 1988; Hobbie & Chapin 1996), and leaching of soluble nitrogen from senescent leaves (Kielland 1995). This build-up of soluble N is responsible for the pulse of nitrogen present in spring melt water, much of which may be lost as runoff because plant uptake is low due to low root biomass, cold temperatures and frozen conditions immediately below the soil surface (Hobbie & Chapin 1996; Brooks, Williams & Schmidt 1998). In both supratidal and intertidal sites, spring melt was characterized by relatively high concentrations of unidentified soluble N compounds (comparison of Figs 1 & 2), that is consistent with cell lysis and litter breakdown associated with freeze–thaw cycles. Although this assertion may appear to be contradicted by the fact that the high NH4+-N concentrations of frozen samples from the intertidal zone in late April 2000 exceeded the estimated total soluble nitrogen, the estimate may have been biased by high concentrations of dissolved organic carbon, which can compete with soluble nitrogen for oxygen during the persulphate reaction (Cabrera & Beare 1993). Ammonia volatilization may also result in incomplete N recoveries from solutions containing high initial concentrations of ammonium ions when alkaline oxidation is employed (Ross 1992). From 52 to 98% of unidentified soluble N compounds were proteins or peptides. Since small peptides can be taken up by plant roots (Stribley & Read 1980), this pool of nitrogen may represent a further source of organic N for coastal plants, beyond that measured in this study. On the basis of the findings of Lee (1988), it is likely that amines (and urea) are also important components of the remaining unidentified nitrogen compounds. However, aside from samples collected in April 2000, soluble inorganic nitrogen and free amino acids accounted for the majority of the total soluble nitrogen (58–76%).

In contrast to intertidal sites, where amino acids concentrations rose over the winter, amino acid concentrations declined in soils in the supratidal marsh (Fig. 2). This disparity between supratidal and intertidal sites may be related to differences in snow-pack and ice cover, which influence soluble nitrogen concentrations due to their effects on soil temperature and water availability (Brooks et al. 1998). In intertidal sites, a thick layer of ice (30 cm) was present in winter and early spring, under which soils were both frozen and highly anoxic. Freeze–thaw cycles of a few days duration do not lead to the disappearance of this ice layer. In supratidal sites, in contrast, where snow was frequently blown elsewhere exposing the soil surface and highly reduced conditions were not present, mineralization rates may have been higher, especially during freeze–thaw cycles, leading to a decline in free amino acid concentrations. Breakdown and decomposition of litter can occur beneath snow-packs in winter during freeze–thaw cycles (Hobbie & Chapin 1996).

Amino acid profiles

Among the amino acid profiles of algae, bacteria, fungi and yeasts (Block 1956), the mean free amino acid composition in soil shows the greatest similarity to that of bacteria (Sowden, Chen & Schnitzer 1977). In addition, in salt-marsh soils, high concentrations of proline are likely the result of fine root and microbial cell lysis, as proline is used as an osmoregulatory solute in Puccinellia species (Stewart & Lee 1974) and in micro-organisms (Flowers, Troke & Yeo 1977). Lysine, which was present at high concentrations in both marshes at warmer temperatures does not normally occur at high concentrations in plant tissue, suggesting that its origin also may be a product of microbial decomposition.

The free amino acid profiles of both supratidal and intertidal sites during the growing season showed marked differences from those reported for other Arctic and alpine sites (Chapin et al. 1993; Kielland 1995; Raab et al. 1996, 1999). Although different Arctic communities have been shown to exhibit unique amino acid profiles (Kielland 1995), typically, the dominant amino acids reported in Arctic sites have been glycine, serine and glutamic acid, with asparagine, arginine, aspartic acid, alanine and cysteine present at high concentrations at some sites. While glutamic acid concentrations were moderately high in both supratidal and intertidal soils and glycine and serine reached occasional high concentrations in supratidal sites, salt-marsh soils were dominated primarily by alanine, proline, lysine and leucine (Fig. 4). Alanine and gamma-aminobutyric acid may be microbial decarboxylation products of aspartic and glutamic acids, respectively, and the high levels of glycine and serine may have originated from mats of diatoms present on these salt-marsh soils (Lee 1988).

Continuous flow experiment

The comparable biomass at final harvest of tillers grown on equivalent concentrations of ammonium ions and glycine indicates the ability of P. phryganodes to utilize this amino acid as the sole source of nitrogen (Fig. 6a). From this result, we can estimate that amino acid uptake by Puccinellia may be as high as 57% that of ammonium in the intertidal zone (see Results section). Results from short-term uptake experiments using doubly labelled 13C15N-glycine strongly suggest that much of the glycine is taken up intact by roots of Puccinellia and that the presence of ammonium and nitrate ions does not inhibit glycine uptake (Henry & Jefferies, in review). Glycine has a low molecular weight and rates of uptake of amino acids appear to be inversely proportional to their molecular weights (Kielland 1994), hence plants may not grow as well on other amino acids, or on mixtures of amino acids. Nevertheless, in these salt-marsh soils, the other amino acids present at high concentrations include alanine and proline which also have a low molecular weight and consequently may be readily taken up by Puccinellia. In addition, given the range of specific neutral, basic and acidic amino acid transporters expressed in roots (Fischer et al. 1998), growth on a mixture of amino acids may be beneficial to plants because competition between individual amino acids for transporter sites is minimized.

Results from the continuous flow experiment indicate that at typical soil concentrations of 100 µm nitrogen, observed transpiration rates are insufficient to replenish nitrogen at the root surface via mass flow alone assuming steady state conditions (Table 1). Diffusion must contribute to nitrogen movement to the root surface, especially at low external concentrations of soluble nitrogen. Because depletion zones occur in the vicinity of roots, as a result of uptake and microbial demand for nitrogen in the rhizosphere (Nye & Tinker 1977; Jones 1999), bulk soil solution concentrations are generally inaccurate as estimates of ionic concentrations at the root surface (Nye 1969; Tinker 1969). Hence, calculations of the mass-flow contribution based on the bulk soil concentration that help maintain the concentration of nitrogen at the root surface tend to overestimate this contribution.

Under field conditions depletion of N at the root surface is likely to be exacerbated compared with that in water culture, because of enhanced root competition and restricted water flow in soils. Concentrations of NH4+-N in the soil solution fall coincident with the increase in biomass of this forage grass and the onset of intensive goose grazing (Fig. 2; Cargill & Jefferies 1984b; Jefferies 1988a, b; Wilson & Jefferies 1996). Puccinellia has no well developed storage organs: it produces a new set of roots each year, and the small number of leaves that survive the winter are replaced shortly after spring melt, hence the internal N pool is limited and is mainly confined to the shoot (Bazely & Jefferies 1989a, b; Srivastava & Jefferies 1995). Re-growth of shoots following defoliation by geese creates a demand for N at a time when exchangeable and soluble inorganic N in soils is low (Fig. 2; Wilson & Jefferies 1996). Experimental studies indicate that not only are faecal droppings necessary for re-growth of vegetation following grazing, but also that multiple grazings (with attendant droppings) are more likely to meet the demands for N to support continuous growth of the grazed grass crop (Hik & Jefferies 1990; Hik et al. 1991; see also Prins, Ydenberg & Drent 1980). The feedback results in increased net above-ground primary production compared with that of ungrazed or infrequently grazed swards. When soluble nitrogen sources are injected into the swards they are incorporated into plant tissue within 48 h, based on tracer studies with 15N (Kotanen 2002). The similarity between the amino acid profiles in faeces and soils (Fig. 5) indicates that the ubiquitous presence of fresh droppings throughout the season and the rapid loss of soluble nitrogen from faeces probably accounts for the similarity.

These results, together with those published elsewhere, indicate that soluble organic nitrogen appears to play a significant role in the nitrogen economy of this forage grass at a time when grazing is intense and inorganic N supplies, as exchangeable and soluble ammonium ions, are low (Wilson & Jefferies 1996). Facilitation by the consumer ensures re-growth at a critical time in the phenology of grazed swards.

Acknowledgments

We wish to thank Paul Matulonis, Rachel Sturge, Deborah Tam, Donny Walkoski, Michael Warnock and the staff and students at the La Pérouse Bay Field Station for assistance in the field and in the laboratory. This research was supported financially by the Natural Sciences and Engineering Research Council of Canada, the Canadian Northern Studies Trust and the Royal Canadian Geographic Society. We are grateful for logistic support from Parks Canada, the Churchill Northern Studies Centre and Hudson Bay Helicopters.

Ancillary