Nitrate inhibition of N2 fixation and its effect on micronutrient accumulation in shoots of soybean (Glycine max L. Merr.), Bambara groundnut (Vigna subterranea L. Vedc) and Kersting’s groundnut (Macrotyloma geocarpum Harms.)

Although nitrate is known to inhibit nodulation and N2 fixation in symbiotic legumes, little is known about its effect on the uptake and accumulation of trace elements such as Fe, Zn, Mn and Cu. The aim of this study was to evaluate the effect of 5 mM NO3 − supply, either with or without rhizobial inoculation, on nodulation, nodule functioning and micronutrient levels in the shoots of soybean (Glycine max L.Merr.), Bambara groundnut (Vigna subterranea L. Vedc) and Kersting’s groundnut (Macrotyloma geocarpum Harm). The results showed reduction in plant growth, nodule formation and nodule dry matter by the supply of 5 mM NO3 − to inoculated seedlings of all three species. Nitrate inhibition respectively caused 1.2, 1.4, and 1.5-fold decrease in nodule number per plant in Bambara groundnut, soybean and Kersting’s bean, which resulted in 2.3, 3.3 and 4.5-fold reduction in nodule dry weight of the test species (in that order). The application of 5 mM NO3 − to soybean plants also resulted in 2.5, 4.0 and 5.4-fold decrease in shoot accumulation of Fe, Zn and Mn, respectively, when compared to the purely symbiotic control plants. Furthermore, we observed 1.3, 1.8 and 1.3-fold decreases in the concentration of Zn, Mn and Cu in shoots of inoculated Bambara groundnut with NO3 − supply, levels lower than those found in soybean. With Kersting’s groundnut, shoot concentration of Fe, Zn and Cu were higher with the application of 5 mM NO3 − to inoculated plants when compared to the purely symbiotic treatment, which was opposite to soybean. But pure NO3 −feeding of this species respectively resulted in 2.0, 1.4 and 1.3-fold decreases in Fe, Zn and Cu relative to inoculated NO3 −-fed plants. Clearly, NO3 − supply to landraces/genotypes of the three legume species did not only inhibit nodule formation and functioning, it also reduced shoot micronutrient levels in soybean and Bambara groundnut, but not Kersting’s bean.


Introduction
Mineral nutrition is important for growth of plants, especially nodulated legumes due to the high demand for nutrients by N 2 -fixing bacteroids in root nodules (Israel 1987;Udvardi and Poole 2013). Recent studies on cowpea have shown a strong relationship between N 2 fixation and mineral accumulation (Belane et al. 2014). However, several factors can inhibit nodulation and N 2 fixation in legumes, which in turn affects mineral accumulation in symbiotic legumes. The enzyme nitrogenase, which reduces N 2 to NH 3, is O 2 -labile and therefore denatures when the O 2 concentration is high (Dakora and Atkins 1989). Mineral N (nitrate and ammonium) is also a potent inhibitor of nodule formation and N 2 fixation (Streeter and Wong 1988;Ayisi et al. 2000). Nitrate, in particular, can combine with leghaemoglobin to form nitrosylleghaemoglobin, which reduces O 2 supply to N 2 -fixing bacteroids (Appleby 1984). Waterlogging of nodulated legumes can also result in reduced N 2 fixation due to low O 2 (Minchin and Summerfield 1976), though some legumes (e.g. Psorolea, Presented at the 17th African Association of Biological Nitrogen Fixation (AABNF) Biennial Conference, Gaborone, Botswana, 17-21 October 2016 Sesbania, Neptunia and Aeschynomene species) are reported to grow, nodulate and fix N 2 under low O 2 conditions as typically found in wetlands and flooded soils (Kanu and Dakora 2017;James et al. 2001).
However, if indeed mineral accumulation in nodulated legumes is symbiotically-linked (Belane et al. 2014), then inhibition of N 2 fixation in root nodules by any of the factors mentioned should alter the uptake and accumulation of minerals in legume organs. Furthermore, imposing an inhibition on nodule functioning ought to reduce N 2 fixation and hence mineral accumulation. Although in this study we measured both micronutrients and macronutrients, the focus is on trace elements because about 263 million African people are suffering from micronutrient deficiency. The aim of this study was to use nitrate as an inhibitor of N 2 fixation to test whether applying this solute to nodulated legumes can reduce nodulation and N 2 fixation, and thus decrease mineral uptake and accumulation in plant organs.

Description of study sites
The study was conducted in a naturally-lit glasshouse located at the Tshwane University of Technology, Pretoria, South Africa during the summer months of October to December 2015. Day and night temperatures were not controlled but average maximum and minimum temperatures were 15 and 28°C respectively.

Source of seeds
The soybean genotypes used in this study included PAN1614 from PANNAR Seed Company in South Africa, as well as TGx1830-20E and TGx2001-25DM from the International Institute for Tropical Agriculture, Mozambique. The Kersting's groundnut/Kersting's bean landraces (Belane, Boli, Heng Milk Mottled, Puffeun, Heng Red Mottled) were sourced from Ghana while the Bambara landraces (Red, Brianbeck, Cream, Black and Mottled) were obtained from South Africa and Ghana.

Planting
Surface-sterilized seeds of landraces/genotypes of the three legume species (soybean, Kersting's and Bambara groundnuts) were planted in autoclaved sand contained in 3-L capacity pots. The experimental treatments included i) rhizobial inoculation, ii) rhizobial inoculation +5 mM KNO 3 supply, and iii) 5 mM KNO 3 -feeding. Bambara and Kersting's groundnuts were inoculated using peat-based commercial inoculant of Bradyrhizobium strain CB756 which is a known commercial inoculant for Bambara and Kersting's groundnuts (Stimuplant, Pretoria). In contrast, soybean was inoculated with Bradyrhizobium japonicum strain WB74, its compatible microsymbiont used for commercial soybean production in South Africa. Each treatment was replicated 5 times. Three surface-sterilized seeds were planted per pot and later thinned to one plant per pot at 7 days after germination. The inoculated plants were watered with full-strength N-free nutrient solution (Broughton and Dilworth 1971). Nitrate-fed plants were watered with 5 mM KNO 3 solution. Watering was done twice a week but the frequency increased as the demand for water also increased with plant growth.

Plant harvest and processing
Plant harvest was based on the time of flowering of each species. Soybean and Kersting's groundnut plants were harvested at 44 days after emergence (DAE), while the Bambara groundnut plants were harvested at 63 DAE. Each plant sample was separated into shoots and roots and kept in separate labelled paper bags. In the laboratory, the shoots were ovendried (60°C) for 48 h, weighed and ground into fine powder (0.85 mm) for analysis of minerals.

Nodulation studies
At harvest, nodulated roots were taken to the laboratory and carefully washed to remove sand. Nodules were then plucked, counted, and oven-dried (60°C) for 72 h for assessing dry matter per plant.

Determination of micro-and macronutrients in plant shoots
The determination of micronutrients in plant shoots was carried out at the Institute for Plant Sciences, Department of Agriculture, Western Cape, South Africa, following the procedure outlined in the Agri-laboratory Association of Southern Africa handbook. Each shoot sample (2.0 g) was ashed overnight in a crucible at 60°C and 5 mL of 1:1 HCl added and left to stand in an oven overnight at 60°C. The solution was filtered, and distilled water added to make it up to 40 mL. The samples were then analyzed for both major and trace elements using inductively coupled plasma-mass spectrometry (ICP-MS).

Correlation analysis
Ground plant shoots were analyzed for both micro-and macronutrients, and correlation analysis performed between shoot mineral concentrations and symbiotic parameters (shoot biomass, nodule number and nodule dry weights) in order to establish physiological relationships (if any) between symbiosis and mineral accumulation.

Statistical analysis
All the collected data were tested for normal distribution before being subjected to analysis of variance to compare treatment means using STATISTICA program (version 10) and GENSTAT (11th edition). A 2-way ANOVA was used to analyze the data set, and where there were significant differences, the Duncan's multiple range test was used to separate the means (p ≤ 0.05).

Effect of genotype/landrace on plant growth and nodulation
Of the three soybean genotypes studied, PAN 1614 and TGx2001-25DM produced significantly more nodule numbers than TGx1830-20E (Table 1). As a result, those two genotypes also produced much greater nodule dry matter (Table 1). Shoot biomass followed the same trend.
There were significant variations in nodulation and plant growth among landraces of Kersting's groundnut (Table 2). Landrace Belane produced the highest number of nodules per plant (34 in number), while Heng Red Mottled produced the lowest, only 17 nodules per plant. The number of nodules were similar for the other landraces. However, the highest nodule dry weight was produced by landrace Puffeun, and the least by Heng Red Mottled landrace. The Belane landrace produced the least shoot biomass, with the other landraces producing similar shoot biomass (Table 2).
Bambara groundnut landraces also differed in plant growth and root nodulation ( Table 3). The Mottled landrace showed greater nodulation (30 nodules plant −1 ), followed by Brianbeck (28 nodules plant −1 ) and then the Red landrace (24 nodules plant −1 ). The Black and Cream landraces produced the least number of nodules (Table 3). However, nodule dry matter was similar for all landraces, except Cream which produced significantly lower nodule dry weight (Table 3). Shoot biomass was markedly higher in the Mottled landrace, followed by Brianbeck, with the Cream and Red landraces producing the lowest shoot biomass (Table 3).

Effect of nitrate on root nodulation
Supplying 5 mM NO 3 − to inoculated soybean plants depressed nodulation (nodule number and dry weight) relative to inoculated purely symbiotic plants. In fact, it caused 1.4fold decrease in nodule number, and 3.3-fold decrease in nodule dry weight relative to purely symbiotic plants ( Table 1). The 5 mM NO 3 − -fed plants (without inoculation) had no root nodules (Table 1).
With Kersting's groundnut, supplying NO 3 − to inoculated plants also markedly depressed nodulation. The provision of 5 mM NO 3 − reduced nodule number per plant by 1.5-fold and nodule dry weight by 4.5-fold relative to purely symbiotic plants ( Table 2). The 5 mM NO 3 − -fed plants (without inoculation) produced no root nodules (Table 2).

Effect of genotype/landrace on micronutrient concentration in shoots
Of the three soybean genotypes tested, TGx1830-20E showed markedly greater levels of Fe, Mn and Cu, while TGx2001-25DM recorded the lowest, except for Zn (Table 1). Kersting's  (Table 3).
With Kersting's bean, the purely symbiotic plants showed lower levels of Fe, Zn and Cu relative to the other treatments. As a result, the inoculated NO 3 − -fed plants revealed 1.9, 1.4 and 1.3-fold increase in Fe, Zn and Cu relative to purely NO 3 − -fed plants, and 1.3, 1.4 and 1.5-fold increase in Fe, Zn and Cu when compared to nodulated purely symbiotic plants (Table 2).
Inoculating soybean plants with Bradyrhizobium japonicum strain WB74 markedly increased the concentration of Fe, Zn and Mn in shoots when compared to purely NO 3 −fed plants or the inoculated + NO 3 − treatment (Table 1). In fact, the inoculated + NO 3 − plants also generally accumulated more Fe, Zn and Mn than the purely NO 3 − -fed plants (Table 1). However pure NO 3 − -feeding accumulated more Cu in soybean shoots than the other two treatments (Table 1). In Kersting's groundnut landraces, the inoculated + NO 3 − treatment increased the concentration of the micronutrients Fe, Zn and Cu over the sole inoculated and purely NO 3 − fed plants (Table 2).
With Bambara groundnut, bacterial inoculation increased the shoot concentration of Zn, Mn and Cu but decreased that of Fe (Table 3). Pure NO 3 − -feeding and inoculated + NO 3 − Fig. 1 Genotype x treatment interaction on a Nodule number b Nodule dry matter and c Shoot dry matter of soybean genotypes grown in the glasshouse plants accumulated more Fe in Bambara groundnut shoots relative to Bradyrhizobium inoculation alone (Table 3).

Genotype x inoculation interaction on nodulation and plant growth
The genotype x inoculation interaction was significant for nodule number, nodule dry weight and shoot biomass of soybean, Kersting and Bambara groundnuts (Tables 1, 2 and 3). For soybean, Bradyrhizobium inoculation alone significantly increased the nodule numbers and nodule dry weight in all three test genotypes, followed by the inoculated + NO 3 − treatment ( Fig. 1a and b). There were no nodules on soybean plants receiving only 5 mM NO 3 − . For PAN 1614, Bradyrhizobium inoculation on its own significantly promoted plant growth and increased shoot biomass, followed by pure NO 3 − feeding and least was inoculated + NO 3 − (Fig. 1c). Plant growth (measured as shoot biomass) was increased equally by bacterial inoculation and NO 3 − feeding of soybean genotype TGx1830-20E while the inoculated + NO 3 − promoted the least plant growth in this genotype (Fig. 1c). With genotype TGx2001-25DM, shoot dry matter was not affected by Bradyrhizobium application or NO 3 − feeding (Fig. 1c).
As found for soybean, the genotype x inoculation analysis showed increased nodule number (except in the Puffeun landrace) and dry mass in all the Kersting's landraces when solely inoculated ( Fig. 2a and b). Plant growth measured as shoot biomass was also markedly increased in the Boli, Heng Milk Mottled, Puffeun and Heng Red Mottled landraces when solely inoculated with Bradyrhizobium over the inoculated + NO 3 − and NO 3 − feeding (Fig. 2c). However, in the Belane landrace, the highest shoot biomass was recorded in the pure NO 3 − fed plants. Relative to the pure NO 3 − fed plants, the inoculated + NO 3 − treatment produced higher shoot biomass in the Boli landrace while in Heng milk mottled, Puffeun and Heng red mottled landrace, similar shoot biomass was produced with both treatments (Fig. 2c).
The genotype x inoculation analysis also showed that Bradyrhizobium application to Bambara groundnut significantly increased nodule number (except Red) and nodule dry weight per plant in all five landraces when compared to inoculated + NO 3 − treatment ( Fig. 3a and b). There were no nodules on pure NO 3 − fed pants. However, plant growth was differentially regulated by the treatments imposed. As shown in Fig. 3c, bacterial inoculation on its own promoted plant growth in only the Black, Brianbeck and Cream landraces when compared to pure NO 3 − feeding and inoculated + . The latter treatment also promoted greater plant growth in only the Mottled and Red landraces (Fig. 3c). Of the five landraces, only Black showed similar and equal shoot growth promotion with NO 3 − feeding and Bradyrhizobium inoculation alone (Fig. 3c).

Genotype x inoculation interaction on micronutrient distribution in plant shoots
The genotype x inoculation interaction was significant for Fe, Zn, Mn and Cu concentrations in shoots of the test genotypes and landraces (Tables 1, 2, and 3). Although shoot Fe concentrations were unaltered by the treatments imposed on soybean genotype TGx2001-25DM, bacterial inoculation on its own markedly increased shoot Fe levels in PAN 1614 and TGx1830-20E when compared to pure NO 3 − feeding or inoculated + NO 3 − (Fig. 4a). The former and latter treatments accumulated similar levels of Fe in the shoots of TGx1830-20E. Shoot Zn concentrations were also altered by the imposed treatments, with Bradyrhizobium application markedly increasing shoot Zn levels over pure NO 3 − feeding and inoculated + NO 3 − in PAN 1614 and TGx2001-25DM (Fig. 4b), but the inoculated + NO 3 − treatment increased Zn accumulation over bacterial inoculation and NO 3 − feeding in TGx1830-20E and over pure NO 3 − feeding in TGx2001-25DM (Fig. 4b). Shoot Mn concentration rose significantly in all three soybean genotypes with bacterial inoculation when compared to pure NO 3 − feeding or inoculated + NO 3 − (Fig. 4c). Pure NO 3 − feeding significantly increased shoot Cu levels in TGx1830-20E. Similar concentrations of Cu in PAN 1614 were accumulated in pure NO 3 − feeding and inoculated + NO 3 − plants but the latter treatment produced greater levels of Cu in TGx2001-25DM. The lowest concentration of Cu in all genotypes was recorded in purely inoculated plants (Fig. 4d). The Kersting's groundnut landraces markedly differed in the concentration of Fe, Zn, Mn and Cu in shoots when planted with the imposed treatments. The inoculated + NO 3 − treatment increased the levels of Fe in the Belane, Puffeun and Heng Red Mottled landraces relative to the other treatments (Fig. 5a). Sole Bradyrhizobium inoculation increased the concentration of Fe in the Boli landrace and together with the pure NO 3 − feeding treatment, increased the level of Fe in the Heng Milk Mottled landrace (Fig. 5a). Zn concentrations increased significantly with the inoculated + NO 3 − treatment in the Belane and Boli landraces, while the Zn concentration of Heng Milk Mottled and Heng Red Mottled landraces were increased with pure NO 3 − feeding (Fig. 5b). Pure NO 3 feeding increased Mn concentration in all five landraces (Fig. 5c). Inoculation alone however increased the Zn concentration in Puffeun landrace. Levels of Cu in shoots of the test landraces were increased by the inoculated + NO 3 − treatment in the Belane and Boli landraces (Fig. 5d). However, pure NO 3 − feeding increased Cu concentration in shoots of Heng Milk Mottled, Puffeun and Heng Red Mottled landraces (Fig. 5d). Bambara groundnut landraces differed in their individual responses to the three experimental treatments. As shown in shoot Fe in the Black, Mottled and Red landraces of Bambara landraces relative to pure NO 3 − feeding and bacterial inoculation alone. However pure NO 3 − feeding also significantly caused greater accumulation of Fe in shoots of Black and Brianbeck landraces. Bradyrhizobium inoculation alone markedly increased Zn concentration in Brianbeck and Cream landraces, and together with the inoculated + NO 3 − treatment also significantly increased shoot Zn levels in the Red landrace (Fig. 6b). It was in only Mottled landrace that pure NO 3 − feeding induced the highest shoot accumulation of Zn (Fig. 6b). Shoot Mn concentration showed a strong response to Bradyrhizobium inoculation as it elicited a marked accumulation of this trace element in the Black, Brianbeck, Cream and Red landraces when compared to the other two treatments (Fig. 6c). Bacterial inoculation also significantly raised the concentration of Cu in the Mottled, Cream and Red landraces, while inoculated + NO 3 − treatment induced much greater Cu level in Black and Brianbeck landraces (Fig. 6d).

Discussion
N 2 fixation is the source of N for meeting the N demand of nodulated legumes. This process can differ between and among legume species and genotypes. In this study, we found differences in N 2 fixation between three soybean genotypes as well as among five Kersting's bean and Bambara groundnut landraces treated to 5 mM NO 3 − and/or rhizobial inoculation (Tables 1, 2 and 3). Here, the three test soybean genotypes responded differently to Bradyrhizobium inoculation, NO 3 − feeding and the combined application of NO 3 − plus bacterial inoculation. Soybean genotypes PAN 1614 and TGx2001-25DM exhibited better plant growth (measured as shoot biomass) with inoculation as a result of higher nodulation and N 2 fixation (Table 1). Increased nodulation and plant growth were also observed in the Kersting's bean and Bambara groundnut landraces with inoculation (Tables 2 and 3). However, soybean, Kersting's bean and Bambara groundnut symbioses revealed some sensitivity to NO 3 − nutrition. As shown in Tables 1, 2 (Dakora et al. 1992;Dakora 1998).
We found some subtle effects of NO 3 − on the symbioses of the three legume species. As shown in Fig. 1, nodule mass was generally more sensitive to NO 3 − than nodule number for the three soybean genotypes. Similarly, nodule mass of Kersting's bean landraces Boli, Heng Milk Mottled and Puffeun was also more sensitive to NO 3 − inhibition than nodule numbers ( Fig. 2). Although the results of Bambara groundnut were similar to those of Kerting;s bean, the effect of NO 3 − on the symbiosis was less pronounced, a clear indication of tolerance of the species to added NO 3 − when compared to soybean (Dakora et al. 1992;Dakora 1998). Ecologically speaking, legumes adapted to growth in nutrient-poor soils tend to underperform with mineral fertilization [e.g. Cyclopia longifolia (Vogel L.)] and usually prefer rhizobial symbiosis as N source and therefore fix more N 2 when nodulated by effective rhizobial strains (Muofhe and Dakora 1999;Maseko and Dakora 2015). Such legumes tend to perform poorly when grown with N and other mineral fertilizers (Ayisi et al. 2000;Mndzebele and Dakora 2017). In contrast, legumes that are adapted to moderately N-rich soils seem capable of depending jointly on symbiosis and mineral N for their nutrition.
In this study, ICP-MS analysis of plant shoots revealed marked differences in micronutrient levels in response to the imposed N sources. Compared to inoculated + NO 3 − treatment, Bradyrhizobium inoculation alone significantly increased Fe concentration in PAN 1614 and TGx1830-20E, Zn concentration in PAN 1614 and TGx2001-25DM, as well as Mn in PAN 1614, TGx1830-20E and TGx2001-25DM (Fig. 4). Coincidentally, Bradyrhizobium application also markedly increased symbiotic performance (i.e. shoot biomass, nodule number and nodule dry matter per plant) in the same soybean genotypes (Fig. 1). With Kersting's groundnut, increased plant growth as a result of Bradyrhizobium inoculation (Fig. 2) did not translate into increased levels of mineral nutrients in shoots of the same plants. However, there was an increase in the concentration of Fe, Zn and Cu in the landrace Belane when inoculated + 5 mM NO 3 − . The same treatment also increased the concentration of Zn and Cu in Boli landrace, while pure NO 3 − -feeding increased Fe, Zn and Cu in Heng Milk Mottled landrace, only Zn and Cu in the Heng Red Mottled landrace, and increased Mn concentration in all five landraces (Fig. 5). The micronutrient data for Bambara groundnut showed a similar pattern and relationship with symbiotic performance as Kersting's bean. Here, Bradyrhizobium inoculation markedly increased the concentration of Zn in Brianbeck, Cream and Red landraces, Mn in Black, Brianbeck, Cream and Red landraces, and Cu in Mottled, Cream and Red landraces when compared to inoculated + NO 3 − treatment (Fig. 6). In contrast, Fe concentration was significantly greater in the inoculated + NO 3 − treatment of the Black, Mottled and Red Bambara groundnut landraces (Fig. 6a). Again, these significant increases in shoot micronutrient accumulation coincided with greater symbiotic performance of the genotypes/landraces that were treated to Bradyrhizobium or inoculated + NO 3 − . It has been shown that accumulation of micronutrients such as Fe, Zn, Mn and Cu in nodulated cowpea is linked to symbiotic efficiency, and is greater where there is higher N 2 -fixing efficiency in root nodules (Belane et al. 2014). The micronutrient data obtained in this study can therefore be interpreted to mean that NO 3 − inhibition of N 2 fixation in soybean and Bambara groundnut altered trace element accumulation in plant shoots. This could be attributed to the suppressive effect of NO 3 − on nodulation and N 2 fixation (Escuredo et al. 1996;Streeter and Wong 1988). It was therefore not surprising that where Bradyrhizobium inoculation increased symbiotic performance, micronutrient levels were also increased in shoots, and where symbiotic performance was low in inoculated + NO 3 − treatment, trace element concentrations were correspondingly low. However, this trend was not observed in Kersting's groundnut landraces, as greater mineral concentrations were recorded in the inoculated + NO 3 − plants, followed by 5 mM NO 3 − fed plants, despite having much higher plant growth when seedlings were solely dependent on bacterial inoculation. Decreasing N 2 fixation by means of NO 3 − application to nodulated plants decreased micronutrient accumulation in some of the test legume genotypes/landraces. This therefore suggests that factors that inhibit N 2 fixation in nodulated legumes (e.g. pO 2 , NO 3 − , waterlogging, drought, etc.) can potentially reduce trace element concentrations in plant organs.
Although it can be argued that NO 3 − -fed and purely symbiotic nodulated legumes may have different physiologies, in this study symbiotic parameters (e.g. shoot biomass, nodule number and nodule dry weights) were significantly correlated with shoot concentration of both macronutrients (P, K, Ca, Mg and S) and micronutrients (Fe, Zn, Mn and Cu). For example, we found significant correlations between shoot biomass of soybean genotype PAN1614 and P (r = 0.86**), S (r = 0.89**), Fe (r = 0.78*), Zn (r = 0.82**) and Mn (r = 0.83), just as its nodule dry matter was also significantly correlated with P (r = 0.70*), K (r = 0.90**), S (r = 0.72***), Fe (r = 0.95***), Zn (0.95***) and Mn (r = 0.92**). We therefore interpret these results to mean that NO 3 − inhibition of |nodulation and N 2 fixation was the main cause of the decreased micronutrient accumulation. In Africa, the soils are low in micronutrients, as a result about 263 million people are suffering from micronutrient deficiency. Therefore, any agronomic practice that further reduces the levels of trace elements in plant shoots is likely to heighten micronutrient deficiency among rural Africans. Although only the grain of the three legumes used in this study is eaten; in the case of cowpea both the leaves and grain are consumed as food. So, reductions in trace element concentration of cowpea shoots would no doubt promote micronutrient deficiency in rural children. It would also be interesting to determine whether these findings apply to nonlegumes. Taken together, this study has demonstrated significant NO 3 − inhibition of nodule formation and N 2 fixation in soybean and Bambara groundnut, but to a lesser extent in Kersting's groundnut. The data have also shown preferences for N sources (NO 3 − , inoculated + NO 3 − and purely symbiotic) by the genotypes and landraces of the three legume species tested. Furthermore, micronutrient concentrations in plant shoots varied with genotype/ landrace and increased with N 2 -fixing efficiency in soybean and Bambara groundnut. Decreasing N 2 fixation with NO 3 − application reduced trace element accumulation in shoots of soybean and Bambara groundnut, but not Kersting's groundnut.