Rhizodeposition of Nitrogen and Carbon by Mungbean (Vigna radiata L.) and Its Contribution to Intercropped Oats (Avena nuda L.)

Compounds released by mungbean roots potentially represent an enormous source of nitrogen (N) and carbon (C) in mungbean-oat intercropping systems. In this study, an in situ experiment was conducted using a 15N - 13C double stem-feeding method to measure N and C derived from the rhizodeposition (NdfR and CdfR) of mungbean and their transfer to oats in an intercropping system. Mungbean plants were sole cropped (S) or intercropped (I) with oat. The plants were labeled 5 weeks after planting and were harvested at the beginning of pod setting (Ip and Sp) and at maturity (Im and Sm). More than 60% and 50% of the applied 15N and 13C, respectively, were recovered in each treatment, with 15N and 13C being quite uniformly distributed in the different plant parts. NdfR represented 9.8% (Sp), 9.2% (Ip), 20.1% (Sm), and 21.2% (Im) of total mungbean plant N, whereas CdfR represented 13.3% (Sp), 42.0% (Ip), 15.4% (Sm), and 22.6% (Im) of total mungbean plant C. When considering the part of rhizodeposition transferred to associated oat, intercropping mungbean released more NdfR and CdfR than mungbean alone. About 53.4–83.2% of below-ground plant N (BGP-N) and 58.4–85.9% of BGP-C originated from NdfR and CdfR, respectively. The N in oats derived from mungbean increased from 7.6% at the pod setting stage to 9.7% at maturity, whereas the C in oats increased from 16.2% to 22.0%, respectively. Only a small percentage of rhizodeposition from mungbean was transferred to oats in the intercropping systems, with a large percentage remaining in the soil. This result indicates that mungbean rhizodeposition might contribute to higher N and C availability in the soil for subsequent crops.


Introduction
Legume-cereal intercropping is an agronomic technique that is known to potentially improve crop yield by increasing resource use efficiency, and has been widely and consistently used in

Study site
The experiment was conducted at the farm of Baicheng Academy of Agricultural Sciences, Baicheng, Jilin, China (45°37 0 N, 122°48 0 E, 152 m above sea level). The local climate is temperate continental monsoon, with an average annual rainfall of 407.9 mm. During the experimental period in 2011, precipitation was 231.4 mm, and the average temperature was 20.2°C. The soil is a chernozem and was collected from a field that had been fallow for 5 years. The soil chemical properties, rainfall, and temperature at the site between June and September 2011 are shown in Table 1 and Fig. 1, respectively.
Experimental design and 13 C- 15

N double labeling
In the spring of 2011, 16 polyvinyl chloride plastic (PVC) columns (diameter, 20 cm; length, 55 cm) with a plastic pallet (diameter, 25 cm) at the bottom were buried in the field. The surface soil (0-30 cm) from the field was collected, air-dried, and passed through a 3 mm sieve to remove gravel and plant residue. About 19 kg of the dry soil was placed in each column and watered daily to maintain a water holding capacity of between 40 and 80%.
The experiment was established in a RCBD (randomized complete block design) with 4 replicates for each treatment. Ten columns were used for intercropping of mungbean with oats, of which 8 were used for the label treatment and 2 for the control. The other 10 columns were used for sole mungbean cropping, of which 8 were used for the label treatment and 2 for the control. Six oat and 6 mungbean seeds were sown in each intercropped column and thinned to 2 oat and 2 mungbean seedlings after emergence. The intercropping treatment was planted as shown in Fig. 2. Six mungbean seeds were sown in the mono-cropped columns and thinned to 2 plants per column after emergence. Corresponding to each 15 N and 13 C labeling treatment, an unlabeled control treatment was planted and harvested the same as the pot treatment to obtain background values of 15 N and 13 C for the different plant and soil parts (as show in Fig. 2).
The mungbean variety used in this experiment was Bailv 11, which matures in about 100 days, while the oat variety was Baiyan 2, which matures in about 85 days. Both varieties were provided by Baicheng Academy of Agricultural Sciences. Mungbean was planted on June 29 after accelerated germination, whereas the oats were planted 20 days later. This planting technique is the normal practice in the study area.
The isotope feeding method used in this experiment was the 15 N-13 C double-labeling method [29,34,40]. In brief, a needle was used to pass a cotton wick through the stem at approximately 3 cm above the soil surface. The end of the wick was inserted into a 1.5-mL liquid chromatography injection bottle that contained the labeling liquid. Transpiration was prevented by covering the wick with a soft plastic tube, except for the part that was inserted into the stem. The labeling solution was 0.2% (w/v) 15 N enriched (99 atom%) urea and 4% (w/v) 13 C enriched (99 atom%) glucose. The solutions were prepared separately using sterile deionized water under aseptic conditions at twice the target concentration, and were mixed together in the injection bottle when the labeling system was set. All of the materials used in the system were sterilized for 20 min at 121°C. Measurements were performed in 2 stages; namely, mungbean pod setting and maturation. Therefore, a total of 4 measurements were performed when combined with the cropping  methods; namely, S p = mungbean harvested at pod setting; S m = mungbean harvested at maturity; I p = intercropping system harvested at mungbean pod setting; and I m = intercropping system harvested at mungbean maturity. S p , I p , and the respective controls were harvested at the onset of mungbean pod setting (August 24, 2011), whereas S m , I m , and the respective controls were harvested at mungbean maturity (September 19, 2011).
Mungbean was labeled twice at weekly intervals starting on the fifth week after planting (August 7, 2011). For each labeling, 1 mL labeling solution was used. Furthermore, absorption of all the solute by the plant was ensured by adding a further 1 mL of deionized water twice after the solution had been taken up by the plants. The vials were placed for 24 h and 48 h (depending on the speed of uptake) and were removed after the completion of labeling. After labeling, a 1-mm mesh was placed on the ground around the labeled plants to collect any fallen leaves and to prevent the soil from being enriched by plant material contaminants.

Plant and soil sampling
During sample collection, the whole of the above-ground plants (AGP) were directly cut above the soil surface. Mungbean was separated into the leaves, stems, and pods, while the AGP of oats was sampled as a single plant part. The soil and roots were also sampled at the same time. First, the columns were removed from the field, the top of each column was sealed with a plastic bag, and the soil from each column was weighed and placed on a plastic sheet. Subsequently, the soil was passed through a 2-mm sieve, and all visible roots were collected manually. The sieved soil was thoroughly mixed, and a 300-g soil sample was obtained for analysis. A subsample of 100 g was dried at 60°C for at least 72 h to determine the weight, and 13 C and 15 N contents. The other 200 g was washed using a 200-μm sieve to determine the amount of remaining rootlets. After the samples were dried at 60°C for at least 72 h, the weight was determined. The samples were then ground to a fine powder using a centrifuge mill and a ball mill, the total N and C content and 15 N and 13 C enrichment in the plant and soil samples were determined using an isotope ratio mass spectrometer (Vario EL, Elementar, Germany) coupled with a Vario PYRO Cube Elemental analyzer. Plant samples were dried at 105°C for 1.5 h and at 60°C to constant weight, the dry weight was determined.

Calculations and statistical analysis
The percentage of soil N and C derived from rhizodeposition (%NdfR and %CdfR, respectively) was calculated using the following equations [41]: %CdfR ¼ atom% 13 C Soil Labeled À atom% 13 C Soil Control atom% 13 C Root Labeled À atom% 13 The quantities of NdfR and CdfR were calculated as follows [29,34]: CdfR N and C transfer from the labeled mungbean to oats was measured by comparing 15 N and 13 C atoms in the mungbean roots and in the receiver's (oat) biomass. The control was the reference unlabeled crop. The proportion of N and C in the oat AGP and roots was calculated using the following equation [33]: %Ctransfer ¼ atom% 13 C Receiver À atom% 13 C Control atom% 13 C Root Labeled À atom% 13 In this experiment the receiver means oats and labeled plants were mungbean, while the control means the corresponding unlabeled plants. The quantities of N and C that were transferred to the oats were calculated as follows [33]: The average data and errors were calculated by Microsoft Excel (Version 2010; Microsoft Inc., USA), while SigmaPlot (Version 12.5; Systat Software Inc., USA) was used to compile the Fig. 1. The equations were edited by MathType (Version 6.0, Design Science Inc., California, USA). The significant effect was compared using Tukey's HSD (Honestly Significant Difference) test at p < 0.05. The analyses of variance were performed using JMP (Version 9.0, SAS Institute Inc., USA) to determine the treatment effects.

Results
Recovery and distribution of 15 N and 13 C 15 N enrichment was the highest in the stem fraction, followed by the grain, leaves, and roots for mungbean, and was the lowest in the AGP and root fraction of oats. 15 N enrichment was generally below 0.3 atom% 15 N excess. However, atom% 15 N excess of the different plant parts did not show any significant difference between treatments (p < 0.05). The total recovery of applied 15 N was between 62% and 82%, and it was slightly higher at maturity than at pod setting, but there was no significant difference between different treatments (p < 0.05). More than 80% of the recovered isotopes were present in the above-ground parts of mungbean. The amount of 15 N that was recovered increased from pod setting to maturity in the grains, but decreased in the stems, leaves, and below-ground plant (BGP) fractions over the same period. At the pod setting stage, more than 60% of the recovered 15 N was detected in the leaves and stems of mungbean. Less 15 N was recovered in the intercropping system than in the sole cropping system. At maturity, more than half of the recovered 15 N was detected in mungbean grains, which was 3 times higher than that recorded at pod setting. The amount of 15 N recovered from rhizodeposition increased over time, with a maximum of 9.82% and 7.06% 15 N being recovered from S m and I m , respectively, with 0.53% of recovered 15 N being found in the intercropped oats (Table 2). 13 C enrichment was similar to 15 N enrichment, with no significant difference between the different treatments (p < 0.05). 13 C enrichment ranged between 0.001 atom% 13 C excess and 0.25 atom% 13 C excess, and was the highest in the mungbean stem fraction. More than 70% of the recovered isotopes were present in the above-ground parts of mungbean. The amount of 13 C recovered from rhizodeposition in the I p treatment was 20.62% (including the C transferred to oats), which was twice that of the S p treatment. In both the S m and I m treatments, about 20% of the recovered 13 C was detected in the BGP. The recovered 13 C in the intercropped oats was 0.19% and 0.38% of the total recovered 13 C at pod setting and maturity, respectively (Table 3).

Quantity and partitioning of N and C
Mungbean dry matter showed no significant difference in the sole or intercropping systems. However, significantly more dry matter was obtained at the mature stage in mungbean grains and leaves (p < 0.05) ( Table 4). At the pod setting stage, NdfR represented 9.8% and 9.2% of the total N, contributing to 57.2% and 53.4% of BGP-N, in the S p and I p treatments, respectively. When combined with the N transferred to the intercropped oat, the total NdfR in I p was 97.0 mg plant -1 , which was 3.7% higher than that in S p . About 40% of the C in I p was detected in the BGP fraction, which was more than fifth that of the corresponding sole cropping treatment. In addition, CdfR was higher than NdfR, representing 13.3% and 42.0% of total C, contributing to 58.4% and 85.9% of BGP-C, in the S p and I p treatments, respectively. After adding the C that was transferred to intercropped oats, the total C derived from rhizodeposition was 10.15 g plant -1 in I p , which was almost 6 times that of S p , but there was no significant difference (p > 0.05) ( Table 5).
The NdfR was higher at maturity in both the sole and intercropping treatments, representing 20.1% and 21.2% of total N, and contributing to 82.5% and 83.2% of the BGP-N, respectively. When combined with the amount of N transferred to oats, total NdfR was 10.8% higher in I m compared to S m . The C derived from rhizodeposition in the S m and I m treatments represented 15.4% and 22.6% of total recovered C, accounting for 72.8% and 79.2% of BGP-C, respectively (Table 5). Mungbean were labeled five weeks after planting and harvested at the beginning of pod setting and at maturity. Mungbean were divided into grain, stem, leaves and root, while oats were divided into above-ground part and root, and rhizodeposition.
The abbreviation in the table represent different treatments. S p : sole mungbean harvested at pod setting, S m : sole mungbean harvested at maturity, I p : intercropping system harvested at mungbean pod setting, and I m : intercropping system harvested at mungbean maturity. Values are means±standard error (n = 4). Values with different letters within a column indicate significant differences between the treatment S p , I p , S m and I m (Tukey HSD, p<0.05). doi:10.1371/journal.pone.0121132.t002

N and C transfer from mungbean to intercropped oat
In the I m treatment, the N content of the AGP and roots of oats was 131.8 mg plant -1 and 20.4 mg plant -1 , respectively. These values were about 3 times higher than those in the I p treatment.
In the AGP of oats, 0.5 and 5.5 mg plant -1 N was derived from mungbean in treatments I p and I m , which represented 1.5% and 4.6% N content, respectively. In addition, the total N derived from mungbean was significantly higher in treatment I m compared to I p (p < 0.05). The percentage of N derived from mungbean in oats increased from 7.6% at pod setting to 9.7% at maturity (Table 6). Mungbean were labeled five weeks after planting and harvested at the beginning of pod setting and at maturity. Mungbean were divided into grain, stem, leaves and root, while oats were divided into above-ground part and root, and rhizodeposition.
The abbreviation in the table represent different treatments. S p : sole mungbean harvested at pod setting, S m : sole mungbean harvested at maturity, I p : intercropping system harvested at mungbean pod setting, and I m : intercropping system harvested at mungbean maturity. Values are means±standard error (n = 4). Values with different letters within a column indicate significant differences between the treatment S p , I p , S m and I m (Tukey HSD, p<0.05). yyyy doi:10.1371/journal.pone.0121132.t003 The C content in the AGP of oats in the I p and I m treatments was 0.42 and 1.73 g plant -1 , respectively. In comparison, the corresponding values of the roots were 0.16 and 0.51 g plant -1 . In the oat AGP fraction, 0.03 and 0.45 g plant -1 C was transferred from mungbean in the I p and I m treatments, representing 8.2% and 24.5% of oat AGP-C, respectively. In addition, the ratio of transferred C from mungbean to oats in relation to total C in oats was 16.2% at pod setting and 22.0% at maturity (Table 6).

Discussion
Recovery of 15 N and 13 C Less than 70% 15 N was recovered from the S p , I p , and S m treatments, which was lower than that obtained by similar studies conducted under controlled conditions, in which recovery rates of  [29,30,34,40]. However, this level was higher than that reported by previous studies conducted under field conditions, in which the recovery rates were reported to be about 60% [29]. The 15 N that was not recovered in the plant-soil system might have been lost because of experimental errors in the mass balance between the calculated 15 N added through labeling and its recovery from the various plant parts [34,42], or loss as NH 3 in urea solution from the wick system. The latter explanation was probably the most likely in this experiment. Specifically, at the first application, the 15 N urea solution was taken up by the plants in less than 1 day, whereas the speed of uptake declined markedly during the second application, requiring at least 2 days. In addition, the recovery rates were higher at maturity than at pod setting, which might be attributed to the lower amount of 15 N lost in the plant-soil system after pod setting.
Relatively higher amounts of applied 13 C (more than 50%) were recovered in our study compared to that reported by similar field studies [29]. However, lower amounts of 13 C were recovered compared to 15 N, which might be partly attributed to the loss of assimilated 13 C as CO 2 , due to either plant or microbial respiration [5,36]. There are two possible explanations for the lower recovery of 13 C at maturity than at pod setting in our study. First, the continued loss of 13 C as CO 2 , supporting previous studies on pea and oat [29]. Second, the solution remained in the plant-soil system for a long time, resulting in an increase in the loss of the applied isotope by microorganisms. Therefore, short-term labeling has advantages over longterm labeling under field conditions, with the short time between labeling and sampling yielding better isotope recovery rates [13].

Suitability of the 15 N-13 C double-labeling method in the intercropping system
To our knowledge, this is the first study in which the cotton-wick method was used for the 15 N-13 C double labeling of an intercropping system. This technology allowed the higher recovery of 15 N and 13 C and did not damage the root structure, which is a drawback of the split-root method [35]. The results of our study showed relatively higher and uniform isotope enrichment of mungbean plants, which was sufficient to trace the 15 N and 13 C in intercropped oat AGP and root fractions (Table 2 and 3). However, the plants in the intercropping system were not homogeneously labeled, with oat fractions (receiver) having lower enrichment than mungbean fractions. This difference may have been a consequence of higher enrichment closer to the label source, which influenced the distribution of the isotope in plants [14]. This trend has also been recorded in other 15 N labeling experiments that used stem feeding [29,34] and root labeling [19].
Furthermore, estimation of rhizodeposition by using Janzen and Bruisma's equation needs to be performed carefully, particularly when the pulsed labeling method is used, because of the assumptions supporting the calculations [29,43]. According to Mahieu et al. [44], rhizodeposition is overestimated for plants subject to the pulse-labeling technique compared to the continuouslabeling method. Labeling experiments lasting more than 1 year need be conducted, because short-term studies may be strongly affected by the weather and other uncontrollable factors.

Quantity and partitioning of N and C
The NdfR (at both pod setting and maturity) was lower for mungbean than for peas labeled using similar methods [29,31]. However, the percentage of NdfR at maturity was higher for mungbean than for peas, faba beans, and lupin that were labeled using a similar method [34], as well as for mungbean (17%) from a pot experiment performed using the petiole-feeding method [45], peas labeled using the split-root method [19,21,22], and alfalfa and soybean labeled under sterile conditions [46]. These differences indicate that the experimental conditions, labeling methods, and environment factors strongly influence NdfR estimates.
The NdfR was 3 times higher at the mature stage than at the pod stage. These results are consistent with those recorded for the pea (Pisum sariuum L.) by Jensen [19] using the splitroot method, with lower NdfR at flowering than at maturity. However, our results contradicted those of Wichern et al. [29], who used a similar method. In addition, more NdfR was obtained in the intercropping system than in the sole cropping system, particularly at maturity. However, the dry matter of mungbean was not significantly different between the sole and intercropping system (Table 4). These results indicate that the N transferred to oats increases the NdfR of mungbean and might strengthen N-fixation. Furthermore, the ratio of NdfR to BGP-N in mungbean was similar to that reported in a previous study [13].
Overall, lower CdfR to total plant C was obtained for mungbean than for pea at both the pod and mature stages [29], but was higher than that for cereal and grasses reported in other studies [36,47,48]. This result might be attributed to the strong rhizodeposition ability of legumes compared to that of cereals, in addition to differences in the root construction and rhizospheric microorganisms of different plants. However, higher amounts of CdfR were obtained in the present study compared to those obtained by Wichern et al. [29] for peas and oats, despite the same method being used. A higher percentage of CdfR to total plant C was obtained in the mungbean-oat intercropping system than in the sole mungbean cropping system, under similar mungbean dry matter content (Table 4). This is because the transfer of C to intercropped oats enhanced the release of mungbean CdfR in the intercropping system.

N and C transfer from labeled mungbean to oat plants
The labeling method used in the current study has been previously used to evaluate C and N rhizodeposition in field and laboratory studies [14, 29-31, 34, 40, 49, 50]. However, to our knowledge, this study is the first time the cotton-wick method was used to evaluate rhizodeposition and the transfer of C and N in a mungbean-oat intercropping system. The current study showed that N transfer represents more than 7% of total N in associated oats, particularly in the roots, which accounted for more than one-third of N transfer. Furthermore, the amount of N that was transferred to oats increased over time. Less N was transferred from mungbean to oats in this study than that reported in previous studies performed using the direct labeling method [51,52] or other approaches [19]. For example, several experiments of intercropped barley and peas reported that 19% of N in barley was acquired from peas by using the split-root method [20]. However, other studies found little or no evidence of N transfer in legume-cereal intercropping systems [2,49,53,54]. Thus, the mechanism of N transfer from legumes to cereal crops might be different, because of differences in plant photosynthesis intensity, root intermingling, and nodule activity.
In this study, mungbean contributed to 16.2% and 22.0% of total C uptake by oats at the pod setting and mature stages, respectively. The oat plants were at the filling stage when mungbean plants were mature. Thus, the strong nutritional need of oats at this stage led to a strong nutrient absorbing ability, which resulted in a higher proportion of C being present at harvest. Most previous studies have focused on the below-ground C of monocropping systems, and found part of the C was respired as CO 2 incorporated into the soil organic matter or used by microbial [36]. However, these previous studies did not investigate the transfer of belowground C in intercropping systems, even though the amount was comparatively low. Our result show that mungbean contributes to about 20% of the C in intercropped oat, and that this part of C cannot be easily neglect. The increased yield of non-legumes might be partly attributed to the amount of C transferred. Nevertheless, information about the importance of C transfer from legumes to non-legumes remains limited, we cannot easily coming to a conclusion.
Although our study provides new insights into the effects of intercropping mungbean with oats on NdfR and CdfR, as well as the transfer of N and C from mungbean to associated oats, the exact mechanisms that influence the amount and quality of NdfR and CdfR during intercropping remain unknown. Moreover, considerable research is required to determine the dynamics of NdfR and CdfR during the plant growth period. Subsequent research should focus on the influence of intercropping on the release of NdfR and CdfR into different soil pools and their utilization by successive crops, by quantifying the nutritional benefit provided to subsequent crops in rotation. Other parameters that warrant further studies include the influence of genotypic variation on rhizodeposition in important crops, N and C transfer from plants to soil microorganisms, and the role of rhizodeposition on C sequestration and N nutrition.

Conclusions
The mungbean and oat intercropping system can be efficiently labeled and studied using the in situ 15 N-13 C double stem-feeding method under field conditions, although maybe short term duration is one of the problems. This study showed that mungbean NdfR contributes to more than half of BGP-N, whereas CdfR accounts for about 60% of BGP-C. In addition, intercropping and time extension can increase the NdfR and CdfR of mungbean. Furthermore, significant amounts of BGP-N were transferred from mungbean to oats in the intercropping system under field conditions, accounting for about 10% total N of associated oats, whereas C transfer accounted for about 20%. It shows the C transferred from mungbean to oats in the intercropping system should not be overlooked. The results of this study show that only a small proportion of rhizodeposition from mungbean is transferred to oats under the intercropping system, with a large portion remaining in the soil. This finding implies that mungbean rhizodeposition could represent a potentially large source of nutrients in the cropping systems of northeast China.
Supporting Information S1 Table. 15 N and 13 C nature enrichment (‰) in the plant and soil parts. (DOCX)