• The effect of nitrogen (N) recycling on tissue δ 15 N was studied using drought-deciduous shrubs ( Encelia farinosa and E. frutescens ) and deciduous trees ( Quercus rubra and Q. alba ) to determine if isotope discrimination occurs with N reabsorption or reallocation.
• There was substantial reabsorption of leaf nitrogen before abscission in all species, yet there was no difference in the δ 15 N of living vs abscised leaves, suggesting no isotope discrimination with reabsorption. However, the contribution of stored N for leaf production resulted in a difference in foliar and absorbed nitrogen δ 15 N, suggesting that isotope discrimination occurs with reallocation.
• The contribution of stored nitrogen to leaf production varied with timing of leaf development: leaves produced first had a δ 15 N that reflected utilization of both absorbed and stored N, yet the contribution of stored N decreased over time.
• The results from this study as well as previous research on intraplant variation in isotopic composition suggest that caution needs to be exercised when interpreting foliar δ 15 N because it is not always a reliable indicator of whole-plant δ 15N.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
In many ecosystems nitrogen (N) is the element that most often limits plant growth and productivity, and anthropogenic activity is changing both the quantity and quality of N available to plants (Vitousek, 1994). In order to predict how plants will respond to these changes it is imperative that we understand the role of different N sources in the overall plant N budget. Variation in N stable isotopes (15N : 14N) is often regarded as a tool to study plant N dynamics because forms of N available to plants often have different isotopic compositions (Yoneyama, 1996). However the use of 15N natural abundance in nitrogen cycling studies is frequently complicated by the fact that soil N pools may mix, confounding attempts to link N contained in plant tissues to specific pools (Robinson, 2001). A further complication is that in addition to soil N pools, internal recycling of N may represent an unaccounted for nitrogen pool accessed by perennial plants.
Plants generally reabsorb half of their total leaf N, yet there is large interspecific variation in the extent of N reabsorption (Chapin & Kedrowski, 1983; Aerts, 1996; Bausenwein et al., 2001a). Variation in N reabsorption efficiency does not appear to be due to differences in nutrient status (Chapin & Moilansen, 1991; Bausenwein et al., 2001b; Grelet et al., 2001) but instead is more likely due to differences in rates of phloem translocation (Chapin & Kedrowski, 1983) and biochemical differences in the rate and extent of protein hydrolysis (Aerts, 1996). Protein hydrolysis, like many biochemical reactions, may involve discrimination against 15N, which could lead to variation in the isotopic composition of different plant tissues (Shearer & Kohl, 1989). Furthermore, based on isotopic mass balance it has been found that foliar N can also be derived from N recycled from fine root turnover prior to leaf flush (Bausenwein et al., 2001b) representing yet another source of N. Thus, foliar isotopic composition may not be indicative of whole-plant composition nor simply reflect reliance on soil N sources, but instead be a product of a multitude of physiological processes.
Isotope discrimination occurs when the lighter isotope (14N) reacts more readily than the heavier isotope (15N), causing the δ15N of the product to be less than that of the substrate. Discrimination could potentially occur during either the biochemical processes involved in N reabsorption from senescing leaves or during the remobilization of that N to new tissues. In either case, if discrimination does occur then foliar δ15N could be affected and could cloud interpretations of community-wide patterns of foliar δ15N. This research addresses the issue of whether discrimination occurs during leaf N reabsorption by examining patterns of δ15N during all phases of vegetative growth in two groups of deciduous plants: drought-deciduous shrubs and winter-deciduous trees. Specifically, we hypothesized that discrimination during N reabsorption and reallocation will be reflected in substantial intraplant variation in δ15N. Discrimination during N reabsorption or reallocation will change the δ15N of the recycled N pool, which becomes a potential internal source of N that could be used along with N absorbed from the soil for new leaf production. If the δ15N of plant tissue reflects that of the N source used during production of the tissues, then any shift in utilization of N sources should be reflected in a change in the isotopic composition of subsequently produced tissues. Furthermore, internal N recycling represents a N source that may not be allocated equally to both root and shoot tissue, especially if shoot growth precedes root growth.
Materials and Methods
The species used in this study represent two different groups of deciduous plants: drought-deciduous shrubs and winter-deciduous trees. The drought-deciduous shrubs were Encelia farinosa Gray and E. frutescens Gray. Quercus rubra L. and Q. alba L. represented winter deciduous trees. Encelia farinosa and E. frutescens are sympatric species that are widespread in both the Mojave and Sonoran Deserts. However, the microhabitat of the two species differs: E. farinosa occurs on desert hillsides and E. frutescens is found almost exclusively in the neighboring washes (Ehleringer & Clark, 1987). Quercus rubra and Q. alba are native to mesic forests in the eastern United States, and frequently co-occur within the central portion of their ranges (Miller & Samuel, 1985).
Plants were germinated from seeds and transplanted into 15-l elongated pots that were large enough to accommodate unrestricted root growth, as was evidenced by the distribution of the root system upon harvesting. The soil medium consisted of sterilized sand that had not been inoculated with mycorrhizae, which allowed for both the amount and isotopic composition of available N to be controlled throughout the experiment. The plants were well watered throughout the growing period and fertilized with modified Hoagland's solution (Smart & Bloom, 1993) according to the experimental treatment (see below).
Sampling and treatment design
The overall experimental design consisted of two phases: examining the extent of N reabsorption during leaf senescence (Phase I), and the extent of N recycling during leaf production (Phase II). The two phases differed in fertilization treatments and the timing of harvests. All plants were fertilized monthly during Phase I with a nutrient solution containing 0.625 mm ammonium and nitrate with an isotopic composition of δ15N of 0‰. During Phase II, plants were randomly assigned to one of three N treatment groups (described below) to determine the contribution of different N sources to leaf production. Water treatments and growth conditions were kept constant between the two phases of the experiment.
Phase I: Discrimination during leaf senescence and abscission
The objective of Phase I was to determine the degree of isotopic discrimination during N reabsorption. This was accomplished by providing adequate amounts of N that had a known isotopic composition, and then examining the N content and isotopic composition for all plant tissues at two different stages of development: before the onset of leaf senescence and following leaf abscission. At each harvest, five plants of each species were randomly selected, removed from the pot, the roots washed with de-ionized water, and the plant divided into root, stem and leaf tissues. The experimental protocol was identical for the second harvest (after leaf abscission), except that leaves were collected as they were abscised from the plants. To induce leaf abscission, watering frequency was decreased for the Encelia species and the Quercus species was subjected to a night-time chilling treatment in addition to decreased watering frequency. At each harvest biomass, δ15N and %N were measured for each tissue. From these data, N content was calculated as mg N per leaf, and there were no observable differences in leaf area within a species that may lead to ambiguities for within species comparisons. This calculation is preferable to expressing it as a percentage of leaf biomass because carbon is also mobilized from leaf tissue during senescence so the calculations would not be standardized. Similarly, expressing N content on a leaf area basis would also not be appropriate in this study because leaf area will also change with leaf abscission due to tissue dehydration.
Phase II: Discrimination during N reallocation
The objective for the second phase of the experiment was to quantify the impact of translocation of N that had been reabsorbed during leaf senescence on foliar δ15N and determine its relative contribution to new leaf production. This was accomplished by randomly assigning plants to one of three treatment groups, which differed in the quantity and isotopic composition of applied fertilizer. The treatment groups were as follows: no N (all leaf N from storage), supplemental N with an isotopic composition identical to that used in the Phase I (to determine the overall effect of both stored and absorbed N on δ15N), and isotopically enriched supplemental N (to quantify the percentage of N from storage). The foundation for all of the fertilizing solutions was quarter-strength modified Hoagland's solution (Smart & Bloom, 1993) that was lacking N. Plants in the first treatment group did not receive any additional N; therefore, all of the N in the newly produced leaves came from storage. The second treatment group received isotopically unaltered supplemental N, which had an isotopic composition similar to natural soil N sources. The enriched fertilization treatment was used to assess the relative contribution of stored and absorbed N sources to leaf production. The concentration of applied N was identical for both the unlabeled and isotopically enriched treatments (0.625 mm ammonium and nitrate). The levels of enrichment varied between the Encelia and Quercus experiments; the isotopically enriched fertilizer used in the Encelia experiment was c. 60‰ and for the Quercus experiment c. 600‰. The pH of all fertilizer solutions was maintained at 6.0 throughout the experiment through the addition of either H2SO4 or KOH.
As with the first phase of the experiment, five plants were randomly chosen from each species and harvested. This harvest occurred after spring growth had ceased, and all tissues were mature. As with previous harvests, the plants were removed from their pots, roots washed in de-ionized water, and separated by tissue. However, the leaves were subdivided by their time of emergence. A cohort of leaves was defined as consisting of four leaves of approximately equal ages, with cohort one being the leaves that were produced first. All tissues were analyzed for biomass, δ15N and %N.
Prior to analysis, all tissue samples were oven-dried at 70°C for 36 h, ground to a fine powder, and transferred to airtight containers. Nitrogen isotope ratios (15N : 14N) and total N concentrations were analyzed by mass spectroscopy. The Encelia samples were analyzed at the Stable Isotope Ratio Facility For Environmental Research at the University of Utah, USA (Finnigan MAT model Delta S) following the method of Evans & Ehleringer (1994). The Quercus samples were analyzed at Stable Isotope/Soil Biology Laboratory of the University of Georgia Institute of Ecology Finnigan Delta C Mass Spectrometer (Thermo Finnigan, San Jose, CA, USA) coupled to a Carlo Erba NA 1500 CHN Combustion Analyzer (CE Elantech, Inc., Lakewood, NJ, USA). To assess the variability due to analysis at the two different locations, replicate Encelia samples were sent to the Georgia facility for analysis. There was no significant difference between the results from the two laboratories (P = 0.08). Isotope values are presented as δ15N (‰).
Data were statistically analyzed using Minitab statistical software (version 13.1; Minitab Inc., State College, Pennsylvania, USA). Patterns of N allocation patterns before and after leaf abscission were assessed using within tissue t-tests on arcsine-transformed data. Deviations of leaf δ15N from the δ15N of the applied fertilizer were assessed using a one-tailed t-test. All other comparisons were analyzed using an analysis of variance, with five replicates for each species. A Tukey's multiple-comparison test was also used to assess the extent of intraplant variation in δ15N among the different plant tissues. In all cases, data from the two genera were analyzed separately, and where there were significant differences between species each species was analyzed independently.
Mobilization and reallocation of foliar N
Regardless of species, N was remobilized prior to leaf abscission as is shown by the change in amounts of foliar N (Fig. 1a,b). The N content of abscised leaves was significantly less than that of living leaves for both species of Encelia (Fig. 1a). The N content of abscised leaves was 62% less than intact, living leaves of E. farinosa. The percent change in N content was greater for E. frutescens, with a difference between living and abscised leaves of 82%. Likewise, the N content of abscised leaves for both Quercus species was much lower than that of living leaves (Fig. 1b). Nitrogen content of Q. rubra abscised leaves was 57% less than that of living leaves; similarly the difference in leaf N content was 55% for Q. alba leaves.
Nitrogen remobilization prior to leaf abscission is further substantiated by comparison of the N allocation patterns before and after leaf abscission. In all four species, the percentage of total N contained in leaf tissues decreased after leaf fall, while the percentage contained in storage tissues (either stem or roots) increased (Table 1). Prior to abscission 56–65% of total N was located within leaves; however, following leaf senescence, abscised leaves only contained 30–42% of the whole-plant N. Correspondingly, the percentage of N contained in stem or root tissues increased after abscission, suggesting that the remobilized N was shifted to storage. Storage allocation patterns differed between the two species of Encelia. Encelia farinosa had similar percentages of total N stored in root and stem tissues; however, E. frutescens had significantly more N stored in roots than stems. Storage allocation patterns did not differ between the two Quercus species; both Q. rubra and Q. alba allocated similar percentages of total N to root and stem tissues. Although there were differences in the observed patterns, all four species exhibited changes in whole-plant N allocation following leaf abscission suggesting remobilization of foliar N during senescence.
Table 1. Nitrogen allocation patterns expressed as a percentage of total N before and after leaf abscission for Encelia farinosa , E. frutescens , Quercus rubra and Q. alba
Before leaf abscission
After leaf abscission
Values in parentheses are SE of the means (n = 5), and asterisks represent significant differences in nitrogen allocation within a tissue before and after leaf abscission (*P < 0.05 and **P < 0.001).
Reabsorption of leaf N did not appear to affect the isotopic composition of leaves for any of the species investigated. There were no significant differences in the δ15N of living vs abscised leaves for either of the Encelia species (Fig. 2a). Similarly, neither Q. rubra nor Q. alba exhibited any differences in the δ15N of living vs abscised leaves (Fig. 2b).
Intraplant variation in δ15N
Both the Encelia and Quercus species exhibited significant, but contrasting, patterns of intraplant variation in δ15N both before and after leaf abscission. Both of the Encelia species showed similar patterns of intraplant variation before and after leaf abscission; however, the magnitude of the difference in E. farinosa tissue δ15N increased after abscission (Fig. 2a). The patterns of intraplant variation observed for E. frutescens tissues before and after abscission were not significantly different. In all cases, the leaves of both Encelia species had significantly higher δ15N values than their respective stem and roots (Fig. 2a). The overall intraplant variation (leaf δ15N – root δ15N) for living tissues (preabscission) for the two Encelia species ranged from 2.1 to 2.8‰. The pattern of intraplant variation was different for the two Quercus species (Fig. 2b). Root tissues of both Quercus species were significantly more enriched in 15N than the leaves. In Q. rubra there was also a significant difference between the δ15N of stems and roots. Although the pattern of intraplant variation observed for the oak species differed from that seen in the Encelia species, the magnitude was similar (1.5‰ for Q. rubra and 2.3‰ for Q. alba).
Effect of N recycling on leaf δ15N
For both species, there was a significant difference between foliar δ15N and that of the absorbed N source. Both Encelia species had leaves that were at least 1‰ greater than the applied N (Table 2), regardless of whether there was a potential contribution from storage. The effect of stored N had little impact on the overall deviation from the applied N due to the fact that the δ15N of Encelia leaves derived solely from storage was not significantly different from δ15N of the nonenriched fertilizer (Table 2). The isotope composition of Quercus leaves adequately reflected the applied N when there was no contribution of stored N; however, when foliar N was derived from both stored and absorbed sources, there was a significant impact on leaf δ15N (Table 2).
Table 2. Effect of nitrogen source on mean foliar δ 15 N (‰) for plants that utilized only absorbed N (from Phase I of study), only recycled (stored) N or a combination of absorbed and internally recycled N (from Phase II of study)
Stored & Absorbed N
Asterisks indicate a significant difference between mean foliar δ15N and that of the applied N fertilizer, and values in parentheses are the SE of the means (n = 5). Foliar δ15N for plants that only had access to stored N were not compared to the applied fertilizer.
The relative contribution of stored N to new leaf production was assessed by examining the δ15N of leaves produced over time by the plants in the enriched N fertilizer treatment group. Initially, the leaves had an isotopic composition that reflected a combination of both stored and absorbed N sources. Based on a linear mixing model, stored N accounts for c. 35% of N in the first cohort of leaves for the two Encelia species (Fig. 3a), and c. 22% for the Quercus species (Fig. 3b). The leaves produced subsequently had an isotopic composition reflecting primarily absorbed N.
Both types of deciduous plants used in this study exhibited substantial reabsorption of leaf N prior to abscission. This was evident from the decrease in leaf N content following senescence, and also by the different patterns of N allocation before and after leaf abscission. The range of N reabsorption observed in this study (55–82%) was comparable with that found for perennial plants in other studies (Chapin & Kedrowski, 1983; Millard & Proe, 1993; Aerts, 1996; Bausenwein et al., 2001a; Bausenwein et al., 2001b). Despite the substantial N reabsorption observed for both the Encelia and Quercus species, there was no significant difference in the δ15N of living vs abscised leaves. This suggests that there was no detectable isotope discrimination associated with the process of N reabsorption.
Despite the lack of any detectable isotope discrimination with foliar N reabsorption, we found that internal N recycling may contribute to intraplant variation in δ15N. For all species studied, we found significant intraplant variation in δ15N. The magnitude of the intraplant variation ranged from 1 to 3‰, which can be considered to be biologically significant (Handley & Raven, 1992). Furthermore, leaves of Encelia plants whose sole N source was internally recycled N did have a significantly different δ15N than leaves of plants that received supplemental N, suggesting discrimination with reallocation of stored N. However, this was not observed for the Quercus species, which showed no differences in isotopic composition of leaves produced from stored vs absorbed N sources. In previous studies on an evergreen tree species (Picea abies), older needles were enriched in 15N relative to younger needles, and this pattern was interpreted to represent discrimination with catabolic breakdown and reallocation of nitrogenous compounds from the older leaves (Gebauer & Schulze, 1991; Gebauer et al., 1994). Results from this study suggest that instead of discrimination occurring with reabsorption, at least for the two drought-deciduous shrubs investigated in this study it appears to occur with reallocation. More importantly, although both types of deciduous plants used in this study showed reliance on recycled N for new leaf production there was no consistent pattern in the effect that recycling had on foliar δ15N. This suggests that it may be inappropriate to generalize about the specific effect of N recycling on foliar δ15N, as it appears to be species specific.
The nature of the discrimination with recycling of N is likely to be due to the incomplete breakdown of N storage proteins into different amino acids that are transported through the xylem and incorporated into new leaf tissue. Current evidence suggests that a few amino acids are predominately involved in N remobilization: asparagine, glutamine, aspartic acid and arginine (Malaguti et al., 2001 and references cited within). All of these with the exception of aspartic acid contain more than two N atoms, and it has been found that at least with arginine the δ15N value of the amino acid varied greatly depending on the extent of N enrichment at the various positions (Medina & Schmidt, 1982). Molecular heterogeneity in 15N enrichment could be the result of different enzymatic reactions involved in forming the molecules and/or different catabolic pathways leading to slight differences in N source pools (Yoneyama et al., 1998). Another process that may lead to differences in isotopic composition of internal N pools is discrimination during loss of volatile amines or ammonia during the protein hydrolysis that accompanies leaf senescence (Shearer & Kohl, 1986). However since we found no difference in δ15N values of living and abscised leaves, this was probably not a major factor for the plants in this study.
Leaves of both groups of plants exhibited a change in foliar δ15N over time, which reflected a shift in the relative contribution of recycled and absorbed N sources. As more leaves were produced, the contribution of recycled N decreased until all leaf N was derived from absorbed sources. The pattern of reliance on recycled N for early leaf production has also been found in herbaceous perennials (Bausenwein et al., 2001a; Bausenwein et al., 2001b), shrubs (Grelet et al., 2001), deciduous (Weinbaum et al., 1984; Millard & Thomson, 1989; Neilsen et al., 1997; Malaguti et al., 2001) and evergreen trees (Proe & Millard, 1995; Stephens et al., 2001). Thus the reliance of perennial plants on recycled N for spring leaf production appears to a widespread phenomenon, which suggests that many plants augment seasonal uptake of N by sequestering a substantial portion of N from ephemeral tissues. Our study, as well as many of the other studies on this subject, used young plants, yet the contribution of recycled N may become more substantial as plants age (Miller & Miller, 1987).
Large and significant intraplant variation in δ15N values was found in this study as well as in previous research under both natural conditions (Gebauer & Schulze, 1991; Gebauer et al., 1994) and controlled experiments (see References in Evans, 2001). The implications of such intraplant variation in δ15N are that foliar δ15N values do not necessarily reflect the δ15N of the whole plant, and variation in foliar δ15N cannot simply be interpreted as utilization of different N sources. An important contribution of the present research is that it clearly shows that even within a controlled experiment there was no consistent pattern of intraplant variation among the deciduous plants studied. The leaves of both Encelia farinosa and E. frutescens were more enriched than stem or root tissues. However, the observed pattern was opposite for the two Quercus species. The roots of both Quercus rubra and Q. alba were more enriched than either leaves or stem tissues.
One explanation for the different patterns of intraplant variation in δ15N of for the Encelia and Quercus species may be that these species exhibit slightly different modes of N assimilation (Evans, 2001). Unlike ammonium, which is assimilated only in roots, nitrate can be assimilated in either root or shoot tissues (Haynes, 1986), and individual species differ in their capacity for shoot and root assimilation (Andrews, 1986). Assuming that nitrate reductase activity discriminates against the heavier isotope, any unassimilated root nitrate would be enriched in 15N compared with assimilated nitrate. Once this enriched pool is transported to the shoot for assimilation, the δ15N of shoot tissue would become more enriched relative to root tissue (Yoneyama & Kaneko, 1989; Evans et al., 1996). Thus, if Encelia species assimilated relatively more nitrate in shoots than roots it would explain the observed pattern of intraplant variation. Similarly, a possible explanation for the pattern observed for the Quercus species is that these species have a low capacity for nitrate reduction in the roots, and most of the nitrate is translocated through the xylem where it is assimilated in stem and leaf tissues. Nitrate reductase activity was not investigated for the plants in this study, but it is an important piece of information that would help interpretation of N dynamics for these plants. However, a lower capacity for nitrate reductase activity in roots than shoots has been shown for another oak species, Quercus robur (Gebauer & Schulze, 1997). The suspected differences in assimilation patterns between Encelia and Quercus species could be due to either the amount of nitrate reductase present in roots vs shoots or to the sensitivity of the enzyme to the presence of ammonium and amino acids produced during ammonium assimilation, which have been shown to influence nitrate reductase activity (Guerrero et al., 1981).
Two other possible explanations for the pattern of preabscission intraplant variation of the two Quercus species are that there may have been substantial efflux of organic N from the root or substantial export of organic N from root to the shoot (Robinson et al., 1998). Of these two explanations, the former is the more plausible. Export of organic N from root to shoots would occur during new leaf production; however, the large intraplant variation δ15N in this study was quantified at the end of the vegetative growth phase and before leaf senescence was induced. As suggested by Robinson et al. (1998), a missing component in our understanding of the physiological processes that lead to intraplant variation in δ15N is a quantification of the isotopic composition of xylem sap. This information would greatly improve our ability to decipher the patterns of tissue δ15N observed in the Quercus species.
Currently, there is interest in using the natural abundance of 15N to infer patterns of plant utilization of soil N sources as it is minimally destructive and integrates all nitrogenous compounds. Unfortunately, our results suggest that direct interpretations of foliar δ15N values may be complicated because of internal plant physiological processes that may vary between species and also temporally within a single species. Furthermore, one assumption to using stable isotopes in this manner is that leaf δ15N reflects that of the whole plant (Shearer & Kohl, 1989), yet this is not always the case (Evans, 2001). Several researchers have found that under controlled conditions with a single N source (either ammonium or nitrate) there was substantial intraplant variation (Mariotti et al., 1982; Yoneyama & Kaneko, 1989; Yoneyama et al., 1991; Evans et al., 1996; Robinson et al., 2000), indicating that foliar δ15N values do not reflect whole-plant values. This study observed similar intraplant variation in deciduous shrub and tree species, and large intraplant variation has also been observed in plants from two native arid-land communities (N. Hardiman, R. D. Evans and K. Kolb, unpublished) and within an evergreen tree (Gebauer & Schulze, 1991). This work provides evidence that the assumption that foliar δ15N accurately reflects whole-plant δ15N would be invalid if a significant fraction of leaf N is derived from N reabsorbed during leaf senescence. The occurrence of seasonal remobilization of leaf N in perennial plants has been well documented, with substantial variability in the extent of remobilization existing between species. Therefore, we propose that it would be prudent to assume that in most cases foliar δ15N does not accurately reflect whole-plant δ15N, nor should differences in foliar δ15N between species be used to interpret utilization of different N sources.
This material is based upon work supported by the National Science Foundation under Grant no. 9723913 from the Ecological and Evolutionary Physiology Program. We thank Nicole Hardiman, Courtney Hamilton, Lynda Sperry, Rhonda Rimer, and Beth Stone-Smith for their assistance with various stages of this research. We also thank Sharon Billings and two anonymous reviewers for their useful comments.