Simulated pollutant nitrogen deposition increases P demand and enhances root-surface phosphatase activities of three plant functional types in a calcareous grassland
Summary
- •
The effects of 7 yr enhanced nitrogen (N) deposition (3.5–14 g N m−2 yr−1) in combination with phosphorus (P) additions, on growth, shoot N and P content, and root-surface phosphomonoesterase (PME) activities were determined along with mycorrhizal infection rates in seedlings of a sedge (Carex flacca), grass (Koeleria macrantha) and forb (Leontodon hispidus).
- •
Seedlings were grown for 14–28 d in mesh-walled cores inserted into turfs from treated field plots enabling complete root recovery.
- •
After 14 d, root-surface PME activity was typically more than doubled by 3.5 and 14 g N m−2 yr−1, and by 28 d the N treatments consistently gave dose-dependent effects. PME activity was reduced by P additions in the sedge and grass by 55 and 65%, respectively, and correlated with soil and shoot N and P concentrations, again most strongly in the sedge and grass. Mycorrhizal infection was least in the sedge (1%) and greater in the grass (49%) and forb (76%).
- •
Long-term N enrichment of calcareous grassland stimulates root-surface PME in representatives of the three major higher-plant functional types. PME response to P additions was greatest in least mycorrhizal-dependant species with roots more adapted for direct P uptake.
Introduction
The latter part of the 20th century saw rapid increases in the rates of reactive atmospheric nitrogen (N) deposition over much of Europe, and although deposition rates in the UK are now declining it remains at historically high levels (NEGTAP, 2001). Increases in N deposition have been implicated in declining floral diversity of species-rich seminatural grasslands, many of which have undergone considerable species change in response to experimental applications of simulated pollutant N (Van Dam et al., 1986; Willems et al., 1993; Bobbink et al., 1998; Carroll et al., 2003). These findings have heightened concern for the long-term sustainability of some of our most species-rich grasslands, especially as a high proportion of these have already been lost due to agricultural ‘improvement’ by nitrogen and phosphorus fertilizer additions (Rich & Woodruff, 1996).
The mechanisms by which reactive N deposition causes loss of some species but not others from grasslands, are incompletely known. There is evidence that in some cases, major floristic changes have occurred mainly due to a ‘fertilizer’ effect of N pollution, such as the expansion of Brachypodium pinnatum at the expense of forbs in Dutch calcareous grasslands. Combined with efficient uptake and retention of P by B. pinnatum, this grass showed enhanced growth and therefore increased shading and competitive exclusion of other species (Bobbink, 1991; Willems et al., 1993). Less clear, however, are the mechanisms of diversity loss in situations where N deposition has not led to increased productivity of the vegetation. For example, in calcareous grasslands on thin Rendzina soils, plants show little or no productivity gains in response to reactive N deposition since these ecosystems are primarily P limited, even before experiencing N pollution, and are therefore readily ‘nitrogen saturated’ (Jeffrey & Pigott, 1973; Willems et al., 1993; Carroll et al., 2003). Furthermore, there is now emerging evidence that in many N saturated ecosystems, high inputs of N have increased P limitation (Bobbink & Lamers, 2002).
In the case of calcareous grassland, simulated pollutant N has been found to decrease rather than increase productivity and to exacerbate P limitation, possibly through impaired cycling of P (Johnson et al., 1999). This may have important consequences for species composition, since it has been suggested that the balance between N and P supply may be critical for the maintenance of ecosystem diversity (Roem & Berendse, 2000). A fuller understanding of the impact of pollutant N on species-rich calcareous grasslands clearly requires the effects on the processes of P acquisition by the main functional groups of plants in these ecosystems to be established. Calcareous grasslands on Rendzina soils have exceptionally low concentrations of plant available P (Jeffrey & Pigott, 1973), and the main fund of the element is present as organic P compounds in the shallow organic-matter-rich surface horizons. The effect of pollutant N on the processes of organic P mineralization and the uptake of P by plants in calcareous grasslands is therefore a central issue for conservation of these increasingly rare, floristically rich communities.
Plants are dependent upon mineralized phosphate as their source of P (Duff et al., 1994) and this is acquired directly from soil solution by plant roots or through their mycorrhizal associates. In addition, inorganic P can be mineralized from organic P by root-surface, mycorrhizal, or soil and microbial phosphatases (Duff et al., 1994; Tarafdar & Marschner, 1994). Root-surface phosphatases may originate from root exudates, mycorrhizas or bacteria (Marschner, 1995). These enzymes include phosphomonoesterases, phosphodiesterases and phosphotriesterases (Florkin & Stotz, 1964). The expression of root-surface phosphomonoesterase (PME) activity is known to be closely regulated by, and a sensitive indicator of, the extent of P limitation to plant growth (Goldstein et al., 1988a, 1988b). In the presence of adequate supplies of orthophosphate, activity of the enzyme is repressed, whereas under P limitation it is stimulated. It has been inferred that root-surface PME plays an important role in P acquisition by plants (Duff et al., 1994; Marschner, 1995) and plant P uptake from specific phosphomonoesters such as inositol hexaphosphate is sometimes strongly correlated with activity of the enzyme (Kroehler & Linkins, 1988; Helal, 1990).
There is now increasing evidence that one of the major impacts of pollutant N on calcareous grasslands is through effects on the P nutrition of the plants. In a previous study of the effect of simulated pollutant N on root PME activity in a calcareous grassland, Johnson et al. (1999) found highly significant increases in the enzyme on roots of Plantago lanceolata seedlings only 7–13 d after transplantation into plots that had received 14 g N m−2 yr−1 for periods of between 18 months and 5 yr. The addition of orthophosphate fertilizer alone, and in combination with the N additions to plots treated for 18 months, reduced phosphatase activity of the seedlings and removed the stimulatory effect of N additions on the enzyme. The increases in PME activity showed strong positive correlations with soil extractable ammonium and shoot nitrogen concentration and a negative correlation with soil extractable phosphate (Johnson et al., 1999).
Crucial questions arising from this study are whether the responses to simulated N deposition seen in P. lanceolata are typical of other plants; if some plant functional types are more sensitive than others to P stress induced by N enrichment; and whether the effects increase over longer periods of N inputs.
Nitrogen treatments to the plots established by Johnson et al. (1999) have been maintained over a period of 8 yr and during this time there has been a gradual shift in plant community composition with an increase in grass and sedge cover at the expense of forbs. These three functional groups of plants differ in their adaptations to acquire P and this is likely to affect their different responses to N enrichment. The main forbs in calcareous grasslands are typically rather course rooted and heavily colonized by arbuscular mycorrhizal fungi, which play a major role in their P acquisition, whereas the grasses produce very fine roots with long root hairs and are less dependant upon mycorrhizas for P uptake (Read et al., 1976; Harley & Harley, 1987). The sedges typically lack mycorrhiza but produce specialist dauciform roots adapted for P uptake (Grime, 2001).
In this study we set out to test the hypothesis that representatives of these three plant functional types would show increases in root-surface PME activity in response to longer-term (7-yr) enhanced N deposition, and that activity of the enzyme is stimulated by higher soil and plant N concentrations, and decreased by addition of P to soil. We also hypothesized that since sedges – and to some extent grasses – have little or no dependence upon mycorrhizas for the uptake of P, and instead are reliant upon their root adaptations (dauciform roots/prolific root hairs), representatives of these plants will have higher PME activity, with greater responsiveness to pollutant N, than a typical forb.
The hypotheses were tested by transplanting bioassay seedlings of a forb (Leontodon hispidus), a grass (Koeleria macrantha) and a sedge (Carex flacca) into turf from a calcareous grassland that has received long-term N enrichment with simulated pollutant N deposition. The species were selected as they are common at the field site and are representative of the functional types’ contrasting root architecture and mycorrhizal dependence. The bioassay seedlings growth was then monitored over 14 and 28 d along with root-surface PME activity and shoot and soil nutrient status.
Materials and Methods
Description of field plots
The calcareous grassland is sited on Wardlow Hay Cop (300 m a.s.l.), a small conical hill in the Derbyshire Dales, UK (NGR SK 1773). The vegetation is classified as a CG2d, Festuca-Avenula (Helictotricon) grassland (NVC Rodwell, 1992) and is located on a shallow (5–10 cm) Rendzina soil overlaying limestone. Further details of the grassland flora can be found in Morecroft et al. (1994). In 1995, 3 m × 3 m plots were established to which nitrogen (N) and phosphorus (P) treatments were applied following a randomized block design (3 N × 2 P × 3 blocks). N was applied as NH4NO3 at 0 (distilled water only), 3.5 and 14 g N m−2 yr−1 with and without P (as NaH2PO4·H2O) at 3.5 g P m−2 yr−1. Ambient N deposition at the site is approximately 2.5 g N m−2 yr−1. Applications of the N and P treatments were made quarterly in the first year and at monthly intervals thereafter (2 l water per plot per application) using back-pack sprayers (Bastion 15, Application Techniques Ltd, Hassocks, Sussex, UK). Application of P was stopped after the first year because dramatic increases in productivity of the vegetation were observed.
Assessment of mycorrhizal colonization of the bioassay species
Three established plants each of C. flacca, K. macrantha and L. hispidus were removed from untreated calcareous grassland turf in late June and their roots cleaned and then cleared by soaking in 10% KOH at room temperature for 1 h. Roots were stained in trypan blue in lactoglycerol after acidification in 10% HCl and the percentage root length colonized by arbuscular mycorrhiza recorded by line intersection method (Giovannetti & Mosse, 1980).
Germination and growth of bioassay seedlings in mesh cores
In the eighth year of the field treatments (April 2002), five soil monoliths with intact turf were removed from each field plot and were cut to fit 15 × 15 cm square plant pots of 10 cm depth. These monolith turfs were housed outside at the University's experimental gardens in Sheffield (estimated ambient deposition c. 2 g N m−2 yr−1). At the field site, each treatment was replicated by 3 plots, so five turfs from each plot allowed for 15 replicate turfs of each treatment. All turfs were considered to be separate replicates, including those originating from the same field plot. This was justified on the grounds that the large size of the field plots (3 m × 3 m) meant that heterogeneity of turfs would be similar within and between plots. The semi-random arrangement of turfs at the experimental gardens ensured that turfs from the same plot were kept separate from each other. Turfs continued to receive the same monthly N treatments as the field plots from which they originated. Monthly additions of P (3.5 g P m−2 yr−1) were recommenced on the turfs that had come from plots treated with P in 1995.
In July, six 2 cm diameter mesh-walled cores, into each of which a bioassay seedlings was later planted, were inserted into each turf after being filled with air-dried and sieved (2 mm) soil taken from the field plots. The six mesh-cores in each turf allowed for the planting of two seedlings each of three species. By conducting the bioassays in mesh-walled cores it is possible to recover the entire root systems of the seedlings without the complication of intermingling and contaminating roots of the surrounding plants. Since root-surface PME activity can vary with the age of the roots we standardized plant age by using seedlings and measured the enzyme activities of entire root systems. The design of the cores is based on that of Johnson et al. (2001). In brief, cores are 10 cm lengths of plastic tubing (ABS water pipe) with a ‘window’, of c. 50% of the core surface area, cut in each side. The windows and bottom opening of the core are covered in nylon mesh (Plastok Associates Ltd, Birkenhead, UK) with a 35-µm pore size.
The cores were left in the turfs for 12 weeks to allow microbial populations and mycorrhizal networks to re-establish from the surrounding turf through the mesh while excluding roots of neighbouring plants (Johnson et al., 2001). The monthly N and P treatments were continued during this period.
In September, seedlings of C. flacca, K. macrantha and L. hispidus were germinated from locally collected seed and after 10 d were planted, one seedling per mesh-core, and left to grow for 14 and 28 d. The six mesh-cores per tuft meant that each turf could contain 2 seedlings of each of the three bioassay seedlings, the two seedlings allowing for the 2 harvests. Death of a few seedlings meant that number of replicates was reduced from 15 at planting to between 11 and 15 for each harvest for each species/treatment combination. The N and P treatments were not applied during the bioassay period to avoid foliar absorption of N and P treatments, which would confound interpretation of the treatment effects on root nutrient absorption.
Entire seedlings were harvested at 14 and 28 d following planting, soil was washed from the roots with distilled water and shoots were excised. Core soil surrounding the seedlings was also removed for N and P analyses.
Seedling shoot biomass and nitrogen and phosphorus concentration
The shoots of the bioassay seedlings were oven dried (48 h at 80°C), placed in a desiccator and weighed. The N and P concentrations of the dried shoots were determined as the ammonium and phosphate content of Kjeldahl digested samples. Ammonium and phosphate concentrations were measured by automated flow injection analyser (FIAflow2, Burkard Scientific, Uxbridge, UK).
Root-surface phosphomonoesterase activity
Root-surface PME activity was determined by measuring the release of p-nitrophenol (p-NP) from the artificial substrate p-nitrophenyl phosphate (p-NPP; Sigma substrate 104), as described by Johnson et al. (1999). The whole, washed seedling root systems were added to glass vials in a shaking water bath at 37°C. The vials contained 10 ml of 4 mmp-NPP in buffer (0.1 m citric acid/0.1 m NaOH) at the pH of the calcareous grassland soil (pH 6.5). Most previous studies of root-surface phosphatases have been conducted at 37°C, so this was chosen to allow comparisons with previous studies (e.g. Johnson et al., 1999). At 30, 60, 90 and 150 min, 200 µl reaction mixtures were removed and added to 3 ml terminating solution (0.1 m Tris adjusted to pH 12 with 0.1 m NaOH). Absorbance at 410 nm (A410) was then determined using a spectrophotometer (Cecil CE1020, Cambridge, UK). The multiple time series was used to check that the reaction rates were linear, allowing for the decreasing volume of reaction mixture at each time point. Release of p-NP was calculated from A410 of 20–400 µmol p-NP standards in 3 ml terminating solution +0.2 ml buffer. PME activity was expressed as nmol p-NP release g−1 f. wt root s−1. Root f. wt was determined by blotting the roots to touch-dry on tissue at the end of the enzyme assays.
Nitrogen and phosphorus concentrations in soil
Ammonium and nitrate were extracted from mesh core soil samples by shaking 2 g d. wt equivalent of soil in 20 ml of 2 m KCl for 30 min. Samples were filtered through Whatman No. 42 ashless filter paper and the ammonium and nitrate content determined using the flow injection analyser. Phosphate was extracted from mesh core soil samples (2 g d. wt equivalent) by shaking for 30 mins in 40 ml 0.5 m NaHCO3 (adjusted to pH 8.5 with 0.5 m NaOH) with 0.1 g activated charcoal (Olsen et al., 1954). Samples were filtered through Whatman No. 42 ashless filter paper and phosphate concentration determined colorimetrically (Murphy & Riley, 1962).
Statistical analyses
Overall N and P treatment effects on PME activity and seedling N and P concentrations were determined using two-way anova in Minitab 13.31 (Minitab Inc., PA, USA) followed by Tukey's HSD test to determine differences between each treatment level. To normalize the data, rates of release of pNP and shoot N and P concentrations and N : P ratios were log10 + 1 transformed while soil nitrate, ammonium and phosphate concentrations were square-root transformed. Correlations between PME activity and the concentrations of N and P in soil and plant shoots, and between PME activity and shoot biomass were determined by regression analysis performed on the transformed data.
Results
Mycorrhizal status of the bioassay plant species
Established plants of the three species had highly significantly different amounts of mycorrhizal colonization (P < 0.001, arcsine transformed data). C. flacca was virtually uninfected with 1% of the root length colonized (standard error = 0.9%). By contrast K. macrantha had 49% of its root length colonized (se = 5%), and L. hispidus had 75% root length colonized (se = 8%).
Effects of long-term enhanced N deposition and P additions on shoot biomass
The effects of the treatments on growth of the plants were generally modest, and in most cases not significant (Table 1). The Tukey HSD found significant effects of the N and P treatments on shoot biomass of K. macrantha and L. hispidus only and there were no significant interaction effects of the treatments (Table 1). In K. macrantha at day 28 the overall effect of the 3.5 g N m−2 yr−1 treatment (with and without additional P) significantly increased shoot biomass (P < 0.05) by nearly one third, and the addition of P (with or without N enrichment) also caused a significant increase (P < 0.05), but by only one sixth (Table 1). In the case of L. hispidus, while the overall effect of the 3.5 g N m−2 yr−1 treatment was to significantly increased shoot biomass at day 14, the overall effect of P treatment was to decrease it at the same time (P < 0.05). Neither of these effects on L. hispidus were significant at day 28, at which time only the 14 g N m−2 yr−1 treatment had an overall significant effect (P < 0.05), increasing the shoot biomass by over one third compared to the ambient N controls (Table 1).
| Carex flacca | Koeleria macrantha | Leotodon hispidus | |||||
|---|---|---|---|---|---|---|---|
| Day 14 | Day 28 | Day 14 | Day 28 | Day 14 | Day 28 | ||
| (a) | 0 N | 0.85a (0.06) | 1.12a (0.06) | 0.24a (0.02) | 0.37a (0.02) | 0.84a (0.05) | 1.05a (0.05) |
| 3.5 N | 0.91a (0.06) | 1.34a (0.08) | 0.29a (0.02) | 0.46b (0.02) | 1.08b (0.05) | 1.26ab (0.08) | |
| 14 N | 0.76a (0.05) | 1.14a (0.09) | 0.28a (0.02) | 0.43ab (0.03) | 0.97ab (0.06) | 1.43b (0.09) | |
| (b) | – P | 0.79y (0.04) | 1.26y (0.06) | 0.26y (0.01) | 0.39y (0.02) | 1.03y (0.05) | 1.29y (1.28) |
| + P | 0.89y (0.05) | 1.13y (0.07) | 0.27y (0.01) | 0.45z (0.02) | 0.88z (0.05) | 1.11y (1.21) | |
| (c) | N | 1.92 | 3.20* | 2.53 | 5.00** | 4.38* | 5.94** |
| P | 3.07 | 2.63 | 0.18 | 6.22* | 5.33* | 0.34 | |
| N*P | 2.04 | 3.21* | 0.49 | 0.16 | 0.24 | 1.46 | |
- Values are overall treatment means, i.e. (a) N treatment means at all levels of P addition; (b) P treatment means at all levels of N addition. (se in parentheses. Within each harvest and each species, means of N treatments (a) or P treatments (b) sharing the same letter are not significantly different (Tukey HSD, P > 0.05). (c) F-values of main N and P treatment effects and N*P interactions calculated from the anova; significance of F-values shown, *P < 0.05; **P < 0.01.
Effects of long-term enhanced N deposition and P additions on shoot nutrient concentrations
anova revealed significant main effects of the N and P treatments on shoot nutrient status of each of the three species, but no significant interaction between the N and P treatments (Table 2). The 3.5 g N m−2 yr−1 treatment had no significant effects on the shoot N concentrations in any of the species. However, the 14 g N m−2 yr−1 treatment had a significant (P < 0.05 in each case) overall effect on shoot N concentrations in all three species at 14 d, with increases of 43% in C. flacca, 26% in K. macrantha and 55% in L. hispidus. However, by day 28 the overall effect of the 14 g N m−2 yr−1 treatment on shoot N concentrations was significant only in the case of C. flacca where it was increased by 35% compared to the ambient N treatments.
| Shoot N | Shoot P | N : P ratio | |||||
|---|---|---|---|---|---|---|---|
| Treatment | Day 14 | Day 28 | Day 14 | Day 28 | Day 14 | Day 28 | |
| (a) | C. flacca | ||||||
| 0 N | 32.2a (2.3) | 40.6a (2.4) | 2.7a (0.3) | 2.1a (0.3) | 19.7a (4.7) | 49.8a (13.3) | |
| 3.5 N | 36.2a (1.6) | 44.0a (3.2) | 3.0a (0.4) | 2.0a (0.3) | 17.5a (2.4) | 58.6a (25.2) | |
| 14 N | 46.0b (2.3) | 54.9b (2.9) | 4.0a (1.1) | 2.9a (0.5) | 26.6a (6.1) | 38.4a (11.0) | |
| (b) | – P | 36.9y (1.7) | 49.4y (2.3) | 2.4y (0.3) | 1.8y (0.3) | 27.2y (4.9) | 71.7y (18.6) |
| + P | 39.2y (2.0) | 42.3z (2.6) | 3.9z (0.7) | 2.9z (0.3) | 15.5z (1.8) | 25.2z (5.5) | |
| (c) | N | 10.47*** | 7.32** | 0.20 | 2.03 | 1.86 | 0.26 |
| P | 0.18 | 5.57* | 7.26** | 12.05** | 7.97** | 14.28*** | |
| N*P | 0.17 | 2.06 | 2.24 | 0.78 | 2.35 | 0.08 | |
| K. macrantha | |||||||
| (a) | 0 N | 47.7a (3.9) | 70.3a (10.6) | 4.3a (0.7) | 4.1a (0.8) | 28.9a (14.8) | 44.4a (19.9) |
| 3.5 N | 59.3ab (5.1) | 57.6a (4.6) | 2.9a (0.5) | 2.3a (0.4) | 34.1a (7.0) | 58.9a (17.7) | |
| 14 N | 60.2b (3.0) | 66.8a (5.3) | 3.4a (0.6) | 3.2a (0.6) | 63.5a (24.5) | 108.0a (55.4) | |
| (b) | – P | 53.3y (3.4) | 62.7y (5.4) | 2.1y (0.3) | 2.1y (0.5) | 61.9y (16.5) | 108.9y (42.2) |
| + P | 57.8y (3.8) | 67.4y (7.3) | 4.5z (0.5) | 4.0z (0.5) | 29.5z (12.5) | 40.9z (12.4) | |
| (c) | N | 3.67* | 0.72 | 1.30 | 1.80 | 2.17 | 1.67 |
| P | 0.55 | 0.32 | 14.71*** | 8.02** | 8.77* | 5.03* | |
| N*P | 2.01 | 0.46 | 0.14 | 2.09 | 0.38 | 1.76 | |
| L. hispidus | |||||||
| (a) | 0 N | 36.7a (1.8) | 51.4a (3.9) | 4.7a (0.5) | 3.6a (0.5) | 9.9a (1.1) | 29.8a (7.5) |
| 3.5 N | 42.4a (2.1) | 56.7a (3.8) | 5.1a (0.6) | 3.4a (0.4) | 10.3a (0.9) | 23.8a (4.2) | |
| 14 N | 57.0b (3.3) | 62.3a (3.6) | 4.2a (0.3) | 3.0a (0.3) | 16.0b (1.9) | 27.8a (4.2) | |
| (b) | – P | 46.4y (2.6) | 59.5y (3.3) | 4.1y (0.3) | 2.6y (0.2) | 13.5y (1.3) | 36.2y (5.6) |
| + P | 43.3y (2.2) | 54.0y (2.9) | 5.2z (0.4) | 4.0z (0.4) | 10.3z (0.9) | 18.5z (2.2) | |
| (c) | N | 21.77*** | 2.50 | 0.76 | 0.18 | 7.85** | 0.81 |
| P | 0.17 | 1.04 | 4.62* | 11.11** | 6.70* | 11.29** | |
| N*P | 0.04 | 1.09 | 0.03 | 0.38 | 1.12 | 0.57 | |
- Table description and statistics as for Table 1.
Phosphorus addition had no significant effects on the shoot N concentrations in K. macrantha and L. hispidus, at any time, but significantly (P < 0.05) reduced the overall N concentration in C. flacca at day 28 (Table 2).
The nitrogen treatments at all levels and both harvests had no significant effects on the shoot P concentrations in the three species (Table 2). However, the addition of P significantly increased shoot P concentrations in all three species at all harvests (P < 0.05). The rise in shoot P concentrations caused by P additions were large, ranging from increases, at day 14 and 28, respectively, of 63 and 61% in C. flacca, 114 and 90% in K. macrantha and 27 and 54% in L. hispidus.
The N : P ratios in the shoots of C. flacca and K. macrantha were not significantly affected by the N treatments, but the 14 g N m−2 yr−1 treatment gave a significant overall increase the ratio in L. hispidus at 14 d only (Table 2). The addition of P strongly and significantly (P < 0.05) decreased the N : P ratios in the shoots of all three species at both harvests and the N : P ratios tended to increase from 14 to 28 d. Despite this, in each case the P treatment effects became more pronounced with time, the decrease in N : P ratio at day 14 and 28, respectively, were 43 and 65% in C. flacca, 52 and 62% in K. macrantha and 24 and 49% in L. hispidus.
Effects of long-term enhanced N deposition and P additions on root-surface PME activity
There were significant overall effects (with and without additional P) of the nitrogen treatments on root-surface phosphatase activity in all three species, and significant effects of P additions (with and without N treatment) on enzyme activity in C. flacca and K. macrantha, but not L. hispidus (Fig. 1a–c). However, the anova found that there were no significant interaction effects of the N and P treatments so only the overall treatment effects are presented in Fig. 1.
Root-surface phosphomonoesterase (PME) activity (pNP release s−1 g−1 root f. wt) after growth of bioassay seedlings of (a) Carex flacca (b) Koeleria macrantha and (c) Leotodon hispidus for 14 and 28 d in turfs that have received ambient (0 N) or enhanced N deposition at rates of 3.5 (3.5 N) and 14 (14 N) g N m−2 yr−1 for 7 yr, alone or in combination with phosphorus supplementation (+ P). The figures present the main treatment effects (i.e. N treatment effects at all levels of P addition; P treatment effects at all levels of N addition). Within each harvest and each species, bars sharing the same letter are not significantly different (Tukey HSD, P > 0.05).
The overall effects of N treatment showed significant increases in PME activity in the 3.5 g N m−2 yr−1 treatment after only 14 d exposure of C. flacca and K. macrantha to the treated soils, but the increase was not significant in L. hispidus (Fig. 1a–c). In both C. flacca and K. macrantha the 14 g N m−2 yr−1 treatment gave no additional increase in PME activity at the 14 d harvest. However, the effects of the N treatments became more clearly dose-dependant at day 28 since at that time in every species the enzyme activity in the 3.5 g N m−2 yr−1 treatment lay between the ambient and 14 g N m−2 yr−1 treatments (Fig. 1a–c). The magnitude of the overall responses of PME activity to the N treatments were very large. In C. flacca the 14 g N m−2 yr−1 treatments increased activity of the enzyme by 120% and 239% at 14 and 28 d (Fig. 1a). In K. macrantha the increase was 102% and 67% at the two harvests (Fig. 1b), and in L. hispidus it was increased by 186% and 120%, respectively (Fig. 1c).
The addition of P caused a very large and significant overall reduction in PME activity in both C. flacca (Fig. 1a) and K. macrantha (Fig. 1b) at both 14 and 28 d harvests. By contrast, the effect of P addition had no significant overall effect on PME activity in L. hispidus (Fig. 1c).
Whilst the patterns of response of PME activity to the N treatments were remarkably similar for the three species, their rates of enzyme activity g−1 f. wt of root varied by an order of magnitude, ranging at the two harvests under control conditions (0 N) from 9 to 10 nmol pNP g−1 s−1 for L. hispidus to 20–29 nmol g−1 s−1 for C. flacca to 106–134 nmol g−1 s−1 in K. macrantha, respectively.
Correlations between root-surface PME activity and shoot and soil N and P concentrations
Correlations between root-surface PME activity and shoot and soil N and P concentrations are shown in Table 3. All significant correlations (P < 0.05) are presented in 2-5.
| Shoot N | Shoot P | Soil NO3− | Soil NH4+ | Soil PO4− | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| r 2 | P | r 2 | P | r 2 | P | r 2 | P | r 2 | P | |
| C. flacca day 14 | 0.101 | 0.008 | < 0.001 | 0.506 | 0.126 | 0.002 | 0.003 | 0.279 | < 0.001 | 0.918 |
| C. flacca day 28 | 0.134 | 0.002 | −0.081 | 0.018 | 0.327 | < 0.001 | < 0.001 | 0.438 | < 0.001 | 0.557 |
| K. macrantha day 14 | < 0.001 | 0.609 | −0.024 | 0.156 | < 0.001 | 0.508 | < 0.001 | 0.860 | −0.016 | 0.143 |
| K. macrantha day 28 | 0.002 | 0.294 | −0.242 | < 0.001 | 0.026 | 0.100 | < 0.001 | 0.759 | −0.079 | 0.012 |
| L. hispidus day 14 | 0.058 | 0.033 | 0.021 | 0.132 | < 0.001 | 0.607 | < 0.001 | 0.344 | 0.001 | 0.313 |
| L. hispidus day 28 | 0.034 | 0.072 | < 0.001 | 0.873 | < 0.001 | 0.342 | < 0.001 | 0.415 | < 0.001 | 0.338 |
- P= level of significance; statistically significant values shown in bold (P < 0.05). Regressions performed on transformed data (see Materials and Methods).
Relationship between root-surface phosphatase activity and shoot N concentration in (a) Carex flacca at day 14 (r2 = 0.10; P = 0.008; y= 0.823x + 0.118) (b) C. flacca at day 28 (r2 = 0.13; P= 0.002; y= 1.18x − 0.561) and (c) Leotodon hispidus at day 14 (r2 = 0.06; P= 0.033; y= 0.72x − 0.083).
Relationship between root-surface phosphatase activity and soil extractable NO3− in (a) Carex flacca at day 14 (r2 = 0.13; P= 0.002; y= 3.96x + 1.21) (b) C. flacca at day 28 (r2 = 0.33; P < 0.001; y= 12.4x + 0.976).
Relationship between root-surface phosphatase activity and shoot P concentration in (a) Carex flacca at day 28 (r2 = −0.08; P= 0.018; y=−0.691x − 1.67) and (b) Koeleria macrantha at day 28 (r2 = −0.24; P < 0.001; y=−0.781x + 2.27).
Relationship between root-surface phosphatase activity and soil extractable P in Koeleria macrantha at day 28 (r2 = −0.08; P= 0.012; y=−9.34x + 2.49).
Significant positive correlations were found between root-surface PME activity and shoot N concentrations (Fig. 2a–c) in C. flacca at days 14 (r2 = 0.10, P < 0.01) and 28 (r2 = 0.13, P < 0.01), and in L. hispidus at day 14 (r2 = 0.06, P < 0.05).
In C. flacca only, significant positive correlations were found between root-surface PME activity and soil nitrate concentrations at both harvests (r2 = 0.13, P < 0.01; r2 = 0.33, P < 0.001, respectively), and the slope of the fitted regression line was greater at 28 d (Fig. 3a,b).
No significant correlations were found between root-surface PME activity and soil extractable ammonium or shoot biomass for any of the species at either harvest.
Significant negative correlations were found between root-surface PME activity and shoot P concentrations (Fig. 4a,b) in both C. flacca at day 28 (r2 = −0.08, P < 0.05), and K. macrantha at day 28 (r2 = −0.24, P < 0.001).
A negative correlation between root-surface PME activity and the concentration of soil extractable P (Fig. 5) was found only in K. macrantha at day 28 (r2 = −0.08, P < 0.05).
Discussion
The results confirm that root-surface PME activity, in representatives of three major higher-plant functional types of calcareous grassland, are highly sensitive and rapidly responsive to the effects of long-term enhanced N deposition. In all three species PME activity was significantly increased, more than doubling in each case, following only 14 d exposure to N treated soil. Similarly, with the exception of L. hispidus, the additions of P dramatically inhibited PME activity even at the first harvest.
The consistent and large effects on PME activity contrasted with the variable, and typically non-significant, effects of both the long-term N treatments and P additions on shoot biomass. The shoot N and P concentrations also showed very modest responses to the treatments compared to those of PME, even though in several species the enzyme activity correlated with these concentrations. The lack of major effects on shoot biomass, and the modest effects on shoot nutrient concentrations may have been in part due to the bioassay plants being slow-growing. This could be partially caused (and further confounded) by competition for, and uptake of, nutrients by the surrounding sward vegetation which could ‘dampen’ the treatment experienced by the bioassay seedling.
Our results extend the observations, made earlier in the treatment history of the same plots, of increased PME activity in roots of Plantago lanceolata after 13 d growth in the N enriched soil (Johnson et al., 1999). The present study importantly demonstrates that the root-surface PME activity of the grass, sedge and forb are all stimulated in a dose-dependant manner by the N treatments, although activity of the enzyme differs by an order of magnitude between the species. Nonetheless, the PME activities lay well within the reported range of values for plant roots (e.g. 0.08–420 nmol g−1 s−1; Lee, 1988; Helal, 1990; Johnson et al., 1999). Since phosphatase is located on the root surface and the enzyme activities are expressed per unit root weight, differences in root surface area to volume ratios of the different species – a characteristic of the functional types to which they belong – most simply explain the different activity rates. For example, C. flacca and particularly K. macrantha have very fine roots and long root hair lengths, characteristic of grasses and sedges, and consequently will have much higher area to volume ratios than the typically course-rooted forb, L. hispidus. The low rates of PME activity in L. hispidus therefore do not reduce the chance of observing treatment effects or correlations with soil and shoot nutrient data, since these low rates are predominantly a result of expressing rates per root f. wt and not of low absorbance values for (and poor resolution of) PME activities during the enzyme assay.
We hypothesized that the three functional types of plants would differ in the magnitude of their responses to P stress induced by simulated pollutant N deposition. This is not clear in the direct responses to the N treatments where all three species showed consistently large increases in PME activity. However, the results clearly show that the three species differ considerably in the magnitude of response to direct changes in alleviation of P limitation (i.e. response to the + P treatment) with responses being greatest in the sedge and grass, with no response apparent in the forb. The greater PME activity and sensitivity to P supply in the grass and sedge than the forb interestingly correspond with the observation from the N treated plots of a loss of cover of forbs and relative increase in cover of sedges and grasses (P. Horswill, G. Phoenix and J. Hodgson, unpublished data). Further work is clearly required to establish whether the floristic changes are due to a direct causal relationship between the root-surface PME activity of species and their tolerance to P limitation in N polluted grassland ecosystems.
PME responses are also likely to be related to the mycorrhizal status of the plants. We did not determine the mycorrhizal status of the bioassay seedlings, but the large differences in mycorrhizal colonization of established plants of the same species growing in the untreated soil will almost certainly have occurred in the seedlings by the 28 d harvest. The established plants of L. hispidus had high rates of mycorrhizal colonization and their broad tap root with relatively little branching indicates that this species will be heavily dependant upon mycorrhizas for P uptake, and its roots poorly adapted for direct uptake of P from the soil (Read et al., 1976; Hetrick, 1991). The PME activity in this species was least sensitive to changes in shoot and soil nutrient status as judged by the limited number and weak strength of correlations between these factors, together with its low enzyme activity per unit biomass. Most notable was the lack of significant reduction in PME activity in this species in the P treated soil. By contrast, PME activities of K. macrantha, and especially the often non-mycorrhizal C. flacca, were very strongly affected by P treatments and more sensitive to changes in shoot and soil nutrient status. K. macrantha, while mycorrhizal, has a finely branching root system with prolific root hairs and has a much lower rate of mycorrhizal colonization than L. hispidus and is almost certainly less reliant upon mycorrhiza for P uptake (Hetrick, 1991). Therefore, at least in the species used in our study, it would seem that the more mycorrhizal dependant the plant species is, and the less adapted the roots are for direct P uptake, the less sensitive root-surface PME activity is to direct changes in P limitation.
Although we confirm that root-surface PME activity is highly sensitive to soil N and P treatments we do not know what fraction of root-surface PME activity originates from root exudates or root associated mycorrhizas and bacteria. Furthermore, the functional significance of the enzyme is still unclear. It is not known whether the higher activity in the N enriched soil has any benefit for P nutrition of the plants, although the negative correlations between shoot P concentration and enzyme activity in C. flacca and K. macrantha indicates tight regulation of activity in these species. However, even if PME does play an important role in plant P nutrition, it is clear that the + 14 N treatment results in high N : P ratios in some seedlings (particularly in K. macrantha, day 28) so it would seem that stimulation of this enzyme alone is not enough to redress the nutrient imbalance of plants in severely N polluted systems.
The inhibitory effects of P additions on PME activity are well established and known to arise both as a consequence of repression of enzyme production together with competitive inhibition of the active site by orthophosphate (Kroehler & Linkins, 1988; Duff et al., 1994; Hunter & McManus, 1999). However, the positive relationships between PME and soil or shoot N status seen in C. flacca and L. hispidus, in the present study, and in P. lanceolata in the earlier study by Johnson et al. (1999) suggest that root-surface PME activity may not be exclusively controlled by shoot and soil P status, though it should be noted that these correlations cannot be used to prove a controlling role for N. In our study, the correlations observed were relatively weaker (lower r2 values) than those observed by Johnson et al. (1999); the weaker correlations resulting from a large variation in enzyme activity compared to the more modest changes in shoot and soil nutrient status that the treatments caused. Goldstein et al. (1988a, 1988b) suggested that stimulation of phosphatase production before P limitation of growth is detectable is a ‘phosphate starvation inducible’ stress response. This pre-empts phosphate starvation by upregulation of phosphatase gene expression to increase the efficiency of uptake of P from the environment and the recapture of P leakage from tissues. In our study, stimulation of PME activity by the N treatments often occurred in the absence of convincing evidence of P limitation to growth, since addition of P to the soil significantly increased the shoot weight of seedlings only in the case of K. macrantha at day 28. Therefore, most of the stimulatory effects of N on PME activity may be phosphate starvation inducible type responses, and our results suggest that the expression of genes in the proposed phosphate starvation rescue system could be sensitive to shoot and soil N as well as P.
Johnson et al. (1999) observed strong correlations between soil extractable ammonium and root-surface PME activity in Plantago lanceolata but no correlation with soil nitrate, and the same results were obtained from Agrostis capillaris seedlings introduced into a parallel series of N-treated plots on an acidic grassland. It was suggested that ammonium was more important than nitrate for stimulating root-phosphatase activity but the ammonium concentrations were much higher and varied more widely than nitrate showing 180-fold and 2-fold variation, respectively. In the present study, variation in extractable ammonium and nitrate concentrations were much more comparable (day 14: ammonium 0–25 µg g−1, nitrate 0–22 µg g−1; day 28: ammonium 0–64 µg g−1, nitrate 0–22 µg g−1), and whilst some positive relationships were found between PME activity and soil nitrate, we found no significant relationships with soil ammonium.
Attempts to elucidate the mechanisms behind ecosystem change resulting from pollutant N have justifiably concentrated on direct effects of N treatments (e.g. shoot N concentrations, nitrate reductase activity, frost hardiness, soil N mineralization/nitrification; Caporn et al., 1994; Carroll et al., 1999; Pearce & Van der Wal, 2002). However, if ecosystem responses to N pollution are to be fully understood, particularly in systems which may be or are becoming increasingly P limited, the indirect effects of pollutant N on plant P nutrition certainly warrants further examination.
Acknowledgements
We acknowledge funding from the UK Natural Environment Research Council (grant GST/02/2683 to JAL, JPG, JRL and DJR) as part of the Global Nitrogen Enrichment (GANE) thematic programme and by DEFRA. We are grateful to Irene Johnson, Paul Horswill and David Johnson for their assistance and Mr Ben Le Bas at English Nature for allowing access to the grassland field site in the Derbyshire Dales NNR.
