Introduction

Nitrogen (N) deposition has been an important part of global environmental change in recent decades1. Since the industrial revolution, industrial processes, agricultural development, and greater human activity have accelerated the emission of N oxides (NOx) and ammonia (NH3)2,3, thus increasing N deposition4. Nitrogen deposition in the global ecosystem was 34 Tg year−1 in the 1860s and 100 Tg year−1 in 1995 and it is projected to increase to 200 Tg year−1 by 20501,5. Nitrogen enrichment has affected global ecosystems drastically6,7. Nitrogen is an essential macronutrient for plants in soil8 and also a key limitation to plant growth in most terrestrial forests9. Contrarily, excess N may lead to better growth of invasive or alien species thereby chang the biodiversity and structure of communities in particular ecosystems6,10,11,12.

Nitrogen derived from biological N fixation provides another N resource to plants13,14,15, through decomposition of the root fraction and root exudation by legumes16. Nitrogen fixation, however, is quite costly as compared with absorbing N from soil17, and can be inhibited by soil N18 and N deposition19. Legumes that can fix atmospheric N can alleviate N limitation in soil20. This can be seen in agricultural ecosystems, in which annual N fixation input from oilseed legumes is estimated to be 18.5 Tg21. Nitrogen-fixing plants can even increase the productivity of nearby plants when planted in combination with certain other species22. Having the ability to fix N, many legumes are less sensitive to N supply than other species23,24, and their photosynthetic rates are seldom affected by leaf N25. Therefore, the effects of N deposition on legumes are increasingly attracting the attention of ecologists26.

Many studies have primarily focused on herbaceous N-fixing legumes21,27,28,29, although studies on N-fixing and non-N-fixing woody legumes are also important elements in forestry and revegetation. Studies examining the differences between woody legumes under various levels of N help to clarify the effect of N supply on the N fixation process. To test the effect of N supply on N-fixing and non-N-fixing woody legume species, we chose Robinia pseudoacacia as the N-fixing species, and one native species, Sophora japonica, and one alien species, Senna surattensis, as the non-N-fixing species; both S. japonica and S. surattensis do not form nodules30,31,32 and cannot fix N33.

To test the effect of increased N supply on non-N-fixing and N-fixing legumes, five different levels of N supply were applied to seedlings of the three tree species in a greenhouse. We hypothesised that N will promote the growth of the three legume tree species, although to a lesser extent in the N-fixing species than in the non-N-fixing species.

Results

Plant growth

No significant effects of N were observed on any of the growth parameters of R. pseudoacacia (Fig. 1). The plant height (p = 0.000), relative growth rate in height (RGRH) (p = 0.000), crown area (p = 0.003), basal diameter (p = 0.010), total biomass (p = 0.000), and number of compound leaves (p = 0.004) of S. japonica significantly increased under an elevated nitrogen supply (Fig. 1).Increased N supply significantly increased the crown area (p = 0.008), basal diameter (p = 0.000), total biomass (p = 0.023), and compound leaf number (p = 0.000) of S. surattensis (Fig. 1).

Figure 1
figure 1

Effects of different levels of nitrogen (N) supply on plant growth in Sophora japonica, Robinia pseudoacacia, and Senna surattensis. (A) Height (n = 5–7); (B) Relative growth rate in height (RGRH) (n = 5–7); (C) Crown area (n = 5–7); (D) Basal diameter (n = 5–7); (E) Total biomass (n = 5–7); (F) Number of compound leaves (n = 5–7). Different lower-case letters for each species denote significant differences at various levels of N supply, and different upper-case letters denote significant differences of species under average level of N supply (p ≤ 0.05).

Full size image

Sophora japonica had the lowest height (p = 0.000), RGRH (p = 0.000), crown area (p = 0.000) and basal diameter (p = 0.001), while R. pseudoacacia had the greatest height, RGRH, crown area, basal diameter and total biomass (p = 0.000), but the least number of compound leaves (p = 0.000) among the three species under the average N supply (Fig. 1).

Biomass allocation

Nitrogen treatment did not significantly affect the root-to-shoot ratio and main root-to-lateral root ratio of any of the three species (Fig. 2C,D). Sophora japonica had higher root-to-shoot ratio than the other two species (p = 0.000) (Fig. 2A). The main root-to-lateral root ratio was the highest in S. japonica and the lowest in S. surattensis (p = 0.000) (Fig. 2B).

Figure 2
figure 2

Difference in biomass allocation (A, B) and the effect of nitrogen on biomass allocation (C, D) in Sophora japonica, Robinia pseudoacacia, and Senna surattensis. (A) Root-to-shoot ratio (n = 27–35); (B) Main root-to-lateral root ratio (n = 27–35); (C) Root-to-shoot ratio (n = 5–7); (D) Main root-to-lateral root ratio (n = 5–7). Different letters for each species denote significant differences (p ≤ 0.05).

Full size image

Leaf traits

Chlorophyll concentration in S. japonica significantly increased with increasing N supply (p = 0.000), while leaf phosphorus (P) concentration significantly decreased under elevated N levels (p = 0.005) (Fig. 3). Chlorophyll concentration in R. pseudoacacia also significantly increased with the addition of more N (p = 0.016), while non-photochemical quenching (NPQ) decreased (p = 0.020) (Fig. 3). The leaf P concentration in R. pseudoacacia was also slightly affected by N. None of the leaf traits of S. surattensis were significantly affected by N (Fig. 3).

Figure 3
figure 3

Effects of different levels of nitrogen deposition on leaf traits in Sophora japonica, Robinia pseudoacacia, and Senna surattensis. (A) Specific leaf area (n = 5–7); (B) Chlorophyll concentration (n = 5–7); (C) Leaf nitrogen (N) concentration (n = 5–7); (D) Leaf phosphorus (P) concentration (n = 5–7); (E) Leaf nitrogen to phosphorus (N:P) ratio (n = 5–7); (F) Maximum quantum yield (Ymax) (n = 5–7); (G) Effective quantum yield (Y) (n = 3); (H) Photochemical quenching (qP) (n = 3); (I) Non-photochemical quenching (NPQ) (n = 3). Different letters for each species denote significant differences (p ≤ 0.05).

Full size image

Among the three species, R. pseudoacacia had the highest specific leaf area (p = 0.000), while S. surattensis had the lowest specific leaf area (Fig. 4A). Robinia pseudoacacia had higher chlorophyll concentration than the other two species (p = 0.000) (Fig. 4B). Senna surattensis had the highest leaf N (p = 0.000) and relatively high P concentration, while S. japonica and R. pseudoacacia had the lowest N concentrations (Fig. 4C,D). Leaf N:P ratios of all three species were similar (Fig. 4E).

Figure 4
figure 4

Differences in leaf traits of Sophora japonica, Robinia pseudoacacia, and Senna surattensis. (A) Specific leaf area (n = 27–35); (B) Chlorophyll concentration (n = 27–35); (C) Leaf nitrogen (N) concentration (n = 27–35); (D) Leaf phosphorus (P) concentration (n = 27–35); (E) Leaf nitrogen to phosphorus (N:P) ratio (n = 27–35). Different letters for each species denote significant differences (p ≤ 0.05).

Full size image

Discussion

Increased N supply promoted the growth of S. japonica and S. surattensis, but not the growth of R. pseudoacacia. During the entire experimental period no root nodules were found on plants of S. japonica and S. surattensis while R. pseudoacacia had root nodules under all N levels. Given that legumes cannot fix nitrogen without root nodules33, our data confirm that R. pseudoacacia was the only of the three species that fixed N. Previous studies have reported that growth of non-N-fixing plants is mostly promoted by elevated N supply34,35,36. However, in species that can fix N (e.g., soybean or lupin), growth is often less or even not affected by N fertilization37,38; this was confirmed in our study. Previous studies have shown that N fertiliser can inhibit N fixation18,39,40, which explains a similar growth of R. pseudoacacia at both high and low N supplies. The growth of R. pseudoacacia under low soil N levels was not limited because the plant was able to fix the required N from the atmosphere.

Nitrogen did not affect biomass allocation, leaf N concentration, or leaf N:P ratio of the three species. Biomass allocation is affected by N in many plant species34,41,42,43,44, but it is independent of N in many N-fixing legumes24. In our study, N did not affect biomass allocation in the three species, suggesting that this trait may be unrelated to N fixation and is specific to legumes. However, this remains to be tested in other N-fixing and non-N-fixing species from this plant group. Increased chlorophyll concentrations in S. japonica and R. pseudoacacia at higher N levels suggests higher photosynthetic rate45,46. Leaf N concentrations were not affected by soil N in our experiment, it has been reported that leaf N is unrelated to photosynthesis in most N-fixing legumes25, and this may also be the same in non-N-fixing legumes. Leaf P concentration in S. japonica decreased under higher N levels and was the lowest of the three species, suggesting that the ability of S. japonica for absorbing P from soil was weaker than that in the other two species. This was also evident from the main root-to-lateral root ratio, which was the lowest in S. japonica owing to its lower nutrient exploiting efficiency when compared with the other two species47.

The similar growth of R. pseudoacacia at different N levels is related to photosynthesis and carbon cost. It has long been reported that under low N conditions plants tend to have a higher photochemical efficiency of the photosystem II (PSII) and photochemical quenching (qP), but lower NPQ48,49,50,51,52. In the present study, higher N levels decreased the NPQ, but N supply had no effect on qP and PSII photochemical efficiency. This absence of correlation between N levels and PSII photochemical efficiency and higher chlorophyll concentrations suggested that the photosynthetic rate in R. pseudoacacia was similar at high and low N levels. This similar photosynthetic rate at different N levels may be attributed to the presence of reactive oxygen species. Namely, qP and NPQ may contribute to the decreased formation of singlet oxygen, a type of reactive oxygen species53. Therefore, similar qP and a decreased NPQ under high N levels may indicate increased production of singlet oxygen, which can damage PSII54. At low N supply, R. pseudoacacia may have higher N fixation rate thus expending more energy and carbon17, while at higher N levels, the plants may have similar or higher photosynthetic rates and may also produce more singlet oxygen that can damage PSII. Thus, despite having higher chlorophyll concentration, the photosynthetic rate in R. pseudoacacia at higher N levels, owing to PSII damage, may be similar or even lower than that in plants at low N levels55.

Different source activities were observed in the three species. Nitrogen has increased sink capacity56, which in turn, promotes source strength57,58. Sophora japonica at high N levels tended to increase the strength of the source (photosynthetic rate) and the number of sources (number of compound leaves) to get a higher relative growth rate and biomass accumulation. Robinia pseudoacacia under high N levels seemed to increase the strength of the source although incurring PSII damage. Senna surattensis at higher N levels increased only the number of sources to enhance their growth.

The difference among species and the varied response to N may be related to different ecological strategies of plant species. Robinia pseudoacacia, which was dominant under low N conditions in our experiment, has the ability to fix N59. Its higher biomass use efficiency in leaves, when compared with other species, was reflected in its large specific leaf area, and its predicted high photosynthetic rate (high chlorophyll concentration)45,46, as well as its ability to fix N, make it a more invasive species that can adapt to different soil types60. Invasion by R. pseudoacacia may lead to an increase of soil N thus may cause biodiversity loss61. The relatively low leaf P concentration in R. pseudoacacia may be due to the strong sink of P in its nodules62,63. Senna surattensis had the lowest main root to lateral root ratio, which shows that its roots have a higher exploitation efficiency47. The N and P concentration was relatively high in the leaves of S. surattensis, which also demonstrated its high absorption rate of nutrients from the soil64. This was confirmed as S. surattensis is mainly distributed in South Asia65, where lower soil P concentrations are observed66. Unlike R. pseudoacacia, which is an early successional species with high foliar efficiency, S. surattensis and S. japonica are more likely to be late successional species67. Sophora japonica in our study was never dominant under any N conditions; it is also observed to be non-dominant in China, especially the northern regions68. Its large main root-to-lateral root ratio revealed its low exploitation efficiency of nutrients from the soil. This may explain its distribution in northern China, where there are higher nutrient levels in the soil66. Studies at the molecular level are still needed to quantify the difference of N-fixing and non-N-fixing legumes.

Our study showed that the growth of the N-fixing species Robinia pseudoacacia was not affected by different levels of N supply. In contrast, the growth of the non-N-fixing legume species was promoted by increased N supply. Similar growth in both the non-N-fixing species and R. pseudoacacia was observed under high N supply. Matured leaves, which are the source part of plants, increased in number in non-N-fixing species under increased N supply. The leaf number in R. pseudoacacia as well as total biomass, were similar under all N levels. Biomass allocation and leaf N, that were not affected by N supply, may have been legume-specific properties, rather than being specific to N-fixing species. The traits’ difference between species may reveal a different adaptation ability, which might affect their distribution in the future, characterised by increased N supply. The non-N-fixing species grew better under high N supply conditions than under low N supply, which provides evidence for planted forestry and garden greening when choosing woody species and managing fertilisation. Further studies are required on the effect of N on matured woody legumes.

Materials and Methods

Study site

Our experiments were conducted at the Fanggan Research Station of Shandong University (36°26′ N, 117°27′E) in Laiwu, Shandong Province, China. The study site is characterized by warm temperate monsoon climate, with a mean annual temperature of 13 ± 1 °C and a mean annual precipitation of 700 ± 100 mm. Most of the rainfall is concentrated in summer and early autumn69. The entire experiment was carried out in a greenhouse during the growing season.

Seeds of R. pseudoacacia, S. japonica, and S. surattensis were germinated in late April. Subsequently 35 healthy, similar-sized seedlings of each species were transplanted into pots filled with 6 kg of loam and 2 kg of sand on May 5 (one plant per pot for each species). The substrates, which were thoroughly mixed, had the following chemical properties: 50.20 mg kg−1 available N, 31.14 mg kg−1 available P, and pH 6.51. The plants were watered every 2 days. After being transplanted into pots, the plants were treated with different amounts of N on June 15th.

Experimental design

Each species was treated with five N levels: 0, 1.5, 2.9, 5.9 and 11.4 mg N per plant day−1 (N1, N2, N3, N4, and N5, respectively). Ammonium nitrate (NH4NO3) was used as the source of N. Nitrogen application began on June 15th and the NH4NO3 solution was added to the pots five times at 15-day intervals. No inoculation was performed, and root nodules were formed with natural rhizobia present in the unsterilized soil. All the plants were harvested on September 10th, i.e. the second week after the last N application. There were seven replicates of each treatment for each species. All pots were randomly arranged in the greenhouse. All plants were watered every 2 days during the experiment.

Measurements

Seedling height and crown area (calculated as crown area = 0.5ab, where a and b are the vertical diagonals of the crown area) were measured at the beginning and the end of N treatments. The relative growth rate in height (RGRH) was calculated with the formula: RGRH = (lnH2−lnH1)/t, where H2 and H1 are the seedling height at the end and the beginning of the N treatment, respectively, and t is the duration of the experiment. The basal diameter was measured before harvest. Flower number (if present) and leaf number were counted at harvest time. Before harvest, the fourth mature leaf from the top was sampled to determine the specific leaf area of plants using the formula: specific leaf area = leaf area/leaf dried biomass, where leaf area was calculated by WinFOLIA Pro 2009a software (Regent Instruments, Inc., Quebec, Canada). Before harvest, the fifth fully expanded mature leaf from the top was collected to measure chlorophyll concentration, using a spectrophotometric method70. Chlorophyll fluorescence parameters were measured using a pulse amplitude modulation chlorophyll fluorometer (Mini-PAM; Walz GmbH, Effeltrich, Germany). The parameters were measured as follows: maximum quantum yield of PSII in dark-adapted leaves, Ymax = (Fm − Fo)/Fm; effective quantum yield of PSII, Y = (Fm′ − F)/Fm′; photochemical quenching, qP = (Fm′ − F)/(Fm′ − Fo′); and non-photochemical quenching, NPQ = (Fm − Fm′)/Fm′, where Fm and Fm′ are maximum fluorescence and peak value of fluorescence; Fo is minimum fluorescence after dark adaptation and Fo′ = Fo/((Fm − Fo)/Fm + Fo/Fm′). During harvest, compound leaf numbers of each plant were recorded, plants were cut at ground level, and the underground parts were carefully washed with tap water. All plants were separated into five parts: main root, lateral root, stem, leaf blade, and leaf petiole. The parts of each plant were packed in separate envelopes and oven dried for 48 h at 80 °C. Once the material was completely dried, all parts were weighed. Plant biomass allocation was calculated, based on dried biomass, as follows: total biomass = main root biomass + lateral root biomass + leaf blade biomass + petiole biomass + stem biomass; root to shoot ratio = (main root biomass + lateral root biomass)/(leaf blade biomass + petiole biomass + stem biomass); main root to lateral root ratio = main root biomass/lateral root biomass. Leaf N and P concentrations were measured using the Kjeldahl method71 and colorimetric determination72, respectively.

Statistical analyses

The data were analysed using one-way analysis of variance (ANOVA) and Tukey’s tests at p ≤ 0.05. Data were checked for normality and homogeneity of variance before performing ANOVA. Log transformation was applied when data did not meet the criteria for normality and homogeneity. Transformed data that still did not meet the criteria for normality and homogeneity of variance were analysed by non-parametric tests. All the data were analysed using SPSS 19.0 software (SPSS Inc., Chicago, IL, USA). Figures were drawn using Origin 8.0 software (OriginLab Co., Northampton, MA, USA).