INTRODUCTION
Ahipa or ajipa (Pachyrhizus ahipa), a South American legume species, is a potential source of raw materials with a carbohydrate-rich tuberous root and protein- and oil-rich seeds (Leidi et al., Reference Leidi, Sarmiento and Rodríguez-Navarro2003). Some countries in South America (Argentina, Bolivia, Peru) traditionally cultivate it for food as a monocrop or intercropped with maize (Ørting et al., Reference Ørting, Güneberg and Sørensen1996; Sørensen, Reference Sørensen1996). In West Africa, ahipa has been introduced jointly with other Pachyrhizus species to test their potential as root crops (Zanklan et al., Reference Zanklan, Ahouangonou, Becker, Pawelzik and Grüneberg2007). Agronomic studies performed in southwest Europe showed ahipa may be competitive in relation to traditional sources of raw materials, mostly under sustainable agriculture systems (Leidi et al., Reference Leidi, Rodríguez-Navarro, Fernández, Sarmiento, Semedo, Marques, Matos, Ørting, Sørensen and Matos2004). The species establishes symbiosis with rhizobia, has a low requirement for pesticides and leaves N-rich harvest residues, which may be used for fodder or organic-N soil amendment (Sørensen, Reference Sørensen1996).
Ahipa can get adequate amounts of N through symbiotic N2 fixation under controlled conditions (Kjær, Reference Kjær1992). However, poor nodulation rate was reported for ahipa grown under natural conditions (Grum and Sørensen, Reference Grum, Sørensen, Sørensen, Estrella, Hamann and Ríos Ruíz1998), in field trials in southwest Europe using a commercial inoculant (Leidi, Reference Leidi2001) or in subtropical northeast Argentina (Fassola et al., Reference Fassola, Pachas, Rohatsch, Uset and Wiss2007). In Mexico, ahipa was able to fix 58–80 kg N ha−1 with an indigenous soil population of rhizobia maintained by frequent cultivation of yambean (P. erosus) (Castellanos et al., Reference Castellanos, Zapata, Badillo, Peña-Cabriales, Jensen and Heredia-García1997). To our knowledge, there has been no selection of rhizobia strains for increasing N2 fixation in ahipa. Seed inoculation with effective strains is a basic strategy to improve plant N nutrition and increase biomass production in low input farming systems (Hardarson and Atkins, Reference Hardarson and Atkins2003). N2-fixing legume crops decrease the negative environmental impact of farming by reducing the requirement for N fertilizers and the production costs and may even enhance soil fertility (Jensen and Hauggaard-Nielsen, Reference Jensen and Hauggaard-Nielsen2003). Improving N2 fixation potential in legumes depends on soil fertility, the plant host and suitable nodulating bacteria (Hardarson and Atkins, Reference Hardarson and Atkins2003). Production of rhizobial inoculants is a rather long process which begins by selecting and testing strains from collections under controlled conditions and ends in field trials to determine nodulation and N2 fixation rates (Stephen and Rask, Reference Stephens and Rask2000).
Field measurement of N2 fixation is a realistic approach to the contribution made by legumes to the N balance of a system allowing the identification of constraints on legume growth and N2-fixing capacity (Peoples and Herridge, Reference Peoples and Herridge1990). Several methods have been proposed to assess the contribution of symbiotic N2 fixation to plant N economy. Among 15N-isotopic techniques, the natural abundance method is a precise and accurate method for estimating the contribution of symbiotic N2 fixation (Peoples and Herridge, Reference Peoples and Herridge1990; Unkovich and Pate, Reference Unkovich and Pate2000). The ureide method is an alternative technique (Herridge et al., Reference Herridge, Bergersen and Peoples1990), which may be used in ahipa as a portion of the N2 fixed is transported in the xylem as ureide-N (Leidi et al., Reference Leidi, Sarmiento, Mazuelos, Rodriguez-Navarro, Sarmiento, Leidi and Troncoso1997), although it provides only a short-term measure of symbiotic performance.
The aims of the present work were therefore: i) to select efficient rhizobia for ahipa cultivars under controlled conditions, ii) to determine the crop yield response to rhizobia seed inoculation and the symbiotic N2 fixation under field conditions; and iii) to assess the effect of cropping ahipa on soil N balance.
MATERIALS AND METHODS
Screening of Rhizobium strains
Screening of the symbiotic performance of Rhizobium strains of different geographical origin (Table 1), provided by M. Grum (IPGRI-CIAT, Cali, Colombia), was performed using P. ahipa accession AC521. Surface-disinfected and pre-germinated seedlings were transferred to sterilized Leonard jars containing N-free nutrient solution and inoculated with 1 ml of 3 day-old bacterial cultures (108−109 cells ml−1). Two uninoculated controls were run: a) −N-control, irrigated with N-free nutrient solution; and b) +N-control, irrigated with nutrient solution containing 50 mM NH4NO3 to provide a total amount of 7.5 mM N per pot.
Rhizobial strains, origin and plant host species and screening of their symbiotic performance on ahipa AC521. Shoot mass, leaf N concentration, nodule mass and number of nodules per plant after 10 weeks of growth under greenhouse conditions.
In brackets, standard error of the mean. Levels of significance for F: *, p < 0.05; **, p < 0.01; ***, p < 0.001. Degrees of freedom: d.f.t, treatments; d.f.e, error.
Studies on ahipa landrace-rhizobia strain interaction
Two factorial experiments were performed under controlled conditions to determine ahipa genotype-rhizobia strain interaction. In the first, two ahipa accessions (AC102, AC521) and three rhizobia strains (Spec1, PAC48, PAC55) were used. Plants were grown until the fruiting stage. A second experiment was carried out until flowering with more ahipa landraces (AC102, AC230, AC521, AC526) and rhizobia strains (Spec1, PAC48, PAC51, PAC55). Seeds of the different ahipa landraces were inoculated with peat-based inoculants of the rhizobia strains, sown in vermiculite- containing vessels, irrigated with nutrient solution (modified Hewitt, N-free, 10× Zn concentration) and grown in an environmental chamber (day/night temperature: 28/22 °C; 14 hr light/10hr dark periods; irradiance 150 μmol m−2 s−1; relative humidity 40/60%). Three replicates per bacterial treatment were set up for each ahipa accession and sampling time. At vegetative and flowering stages (34 and 57 days after sowing, DAS), plants were harvested to determine plant growth (fresh and dry weight), number and mass of root nodules, and plant N content. Xylem exudates were collected from root stumps during 1 hr in Eppendorf vials maintained on ice and frozen until analysis. The concentration of nitrate, ureides and amino acids was determined as previously described by Leidi et al. (Reference Leidi, Sarmiento, Mazuelos, Rodriguez-Navarro, Sarmiento, Leidi and Troncoso1997).
Field trial
The field experiment was implemented from April to November 2001 at the Experimental Station Las Torres (Seville, southwest Spain) to estimate N2-fixation potential of two ahipa accessions (AC102 and AC521) after inoculation with selected rhizobia strains (PAC48, PAC51, and PAC55). The commercial strain Spec 1 was not included in the field trial because it had provided low nodulation in a preliminary trial (Leidi, Reference Leidi2001). Non-inoculated controls were included. Planting was made in 7 m rows 0.50 m apart with a seed density of 10 seeds per metre on 17 April 2001. Watering was done by furrow irrigation approximately every 2 weeks. Soil samples taken before sowing at 0–30 cm depth showed the following chemical properties: pH, 8.13; CaCO3, 27.4%; organic matter, 0.22%; NO3-N, 0.06 ppm; P, 9.6 ppm; K, 196 ppm. At harvest, plant shoots, roots, pods and seeds were weighed, oven dried and milled for further chemical analyses. N concentration was determined after Kjeldahl digestion of dried samples and ammonium measured by a colorimetric reaction (phenol-hypochlorite) in a Technicon Analyzer.
The study of N2 fixation under field conditions was carried out only on the landrace AC521. Xylem sap was collected by applying pressure (Scholander pump) to lateral branches harvested at flowering-fruiting stage. The analysis of N-solutes was performed as indicated above to determine the relative ureides-N in the xylem sap. The 15N/14N analysis was performed at the Stable Isotope Facility (Environmental Biology Group, RSBS, Australian National University) to determine N2 fixation by natural 15N abundance (Peoples and Herridge, Reference Peoples and Herridge1990). Seeds from the field trial were used and reference values were obtained from nodulated plants grown in N-free sand and non-nodulated plants grown in potted soil. Isotopic fractionation among plant parts was partially assessed performing 15N/14N analysis in seeds and tuberous roots.
Experimental design
A completely randomized designed with three replicates was used for the strain selection. The analyses of plant genotype and Rhizobium strain interactions were performed using a factorial design. The field trial was carried out in a randomized complete block design. A statistical software package (Statistix version 7.0) was used for data analysis.
RESULTS
Only nine strains obtained from the rhizobia collection formed nodules and were tested for plant growth and symbiotic parameters (Table 1). Important variations in symbiotic parameters were recorded among rhizobial treatments, with significant differences in leaf N concentration, nodule mass and number of nodules (Table 1). Three strains (PAC48, PAC51 and PAC55) were selected for further studies because they did not show significant differences in shoot growth over the N-fed control (+N treatment) and had the highest leaf N concentration among the inoculated treatments.
The first factorial experiment for the study of the ahipa-rhizobia interaction was set up with three rhizobia strains and two ahipa accessions and showed significant effects of strain on plant growth, shoot N content and ureide concentration in the xylem sap (Table 2). The xylem ureide concentration was significantly correlated with shoot dry weight and the N content in leaves and stems (r = 0.49, p < 0.05; r = 0.48, p < 0.05; r = 0.63, p < 0.01, respectively).
Effect of the inoculation with different rhizobia strains (Spec1, PAC48, PAC55) on the shoot and root growth, N content and symbiotic parameters (number of nodules, nodule mass, nodule size) in ahipa landraces (AC102, AC521) at flowering-fruiting (79 DAS) grown under controlled conditions.
Standard error of the mean are shown in parentheses.
Levels of significance for F: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
When four ahipa landraces and four Rhizobium strains were included, significant differences between ahipa landraces were observed for plant growth at the vegetative stage (34 DAS) with no significant effect of the rhizobial strains (data not shown). However, a significant effect of strains and the interaction landrace × strain was observed for the size of nodules and the plant N content at flowering stage (57 DAS) (Table 3). The interaction of rhizobial strains with the ahipa landrace significantly affected plant shoot growth. Plant shoot growth was significantly correlated with nodule mass (r = 0.59, p < 0.001, n = 32) and the relative abundance of ureides in the xylem sap (r = 0.63, p < 0.001, n = 32). Xylem composition changed significantly with time, showing an increasing concentration of ureides and a decreasing concentration of amino acids, which might be related to nodule maturation and peaking rates of N2 fixation at flowering (data not shown).
Shoot growth, N content and symbiotic parameters (number of nodules, nodule mass, and nodule size) at flowering (57 DAS) in ahipa landraces (AC102, AC230, AC521, AC526) inoculated with different rhizobia strains (Spec1, PAC48, PAC51, PAC55).
Standard error of the mean are shown in parentheses.
Levels of significance for F: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
In the field experiment, good nodulation rates were observed on root-crowns of the inoculated ahipa landraces (AC102 and AC521) and nodules were not found on the uninoculated control. At fruiting stage, leaves from non-inoculated plants showed symptoms of N-deficiency chlorosis indicating that soil N was limiting at times of high plant N requirements (seed filling).
Inoculation with some rhizobial strains produced a significant increase in root, pod and seed yield over the non-inoculated control in some landrace and strain combinations (Table 4). The strain PAC55 provided a significant effect on root growth in AC102 but none in AC521 while PAC48 improved mainly seed yield in both ahipa accessions. The increase in crop yield provided by effective nodulation in comparison with the non-inoculated control was a consequence of greater root weight and fruit load per plant (data not shown). Apart from the effects on crop yield, rhizobia inoculation provided a significantly greater seed protein concentration than the non-inoculated controls (Table 4) although seed oil concentration was lower in the inoculated treatments.
Effect of inoculation with different rhizobia strains on root and seed yield (in kg dry matter ha−1), root sugar concentration (Brix degrees) and seed protein and oil concentration (%) of ahipa landraces AC102 and AC521.
Standard error of the mean are shown in parentheses.
Levels of significance for F: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
The concentration of N in different plant parts (shoot, root, seeds) was significantly increased by seed inoculation (Table 5). The estimation of N2 fixation by means of the natural 15N abundance method showed that at least 60% of the seed nitrogen was obtained through N2 fixation (Table 5). Seeds from uninoculated plants showed a significant P value (% of plant N derived from N2 fixation) which might suggest either contamination through the irrigation water or root N transfer from inoculated plots. The estimate of N2 fixation based in relative ureide content in the xylem sap was related to that provided by the natural 15N abundance method. A preliminary assessment of plant 15N/14N fractionation (δ15N) across strains showed significant differences among roots (1.08 ± 0.13) and seeds (2.03 ± 0.20).
Nitrogen concentration (%) in shoots, roots, seeds and estimates of N2 fixation by natural 15N abundance (P, proportion of plant N derived from nitrogen fixation, %) or relative ureide content of xylem sap (XRU, %) of ahipa landrace AC521 inoculated with different rhizobia strains harvested at 119 DAS.
Standard error of the mean are shown in parentheses.
Levels of significance for F: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
DISCUSSION
Selection of effective rhizobial strains for ahipa inoculation led to significant increases in tuber and seed yield and seed protein content under field conditions. Roots, seeds and shoots from N2-fixing plants showed significantly greater N concentration than non-inoculated controls. As result of symbiotic N2 fixation, seed protein content in the AC521 plants was enhanced up to 45% over the non-inoculated control plants. Interaction effects among rhizobial strains and ahipa landraces were found for shoot growth, N content, number of nodules and nodule size in experiments run under controlled conditions. It may be necessary to produce inoculants for specific plant genotype-rhizobial strain associations for particular soils or management conditions.
The present results show the importance of ahipa seed inoculation with specific strains when the species is introduced in new areas. In local soils, ahipa did not form nodules and the search for nodulating strains was performed in a rhizobial collection. An essential step for improving the legume capacity for N2 fixation is the research leading to the selection of elite strains of rhizobia on the basis of both specificity and high effectiveness traits (Hardarson and Atkins, Reference Hardarson and Atkins2003; O´Hara et al., Reference O´Hara, Yates, Howieson and Herridge2002). In controlled conditions, a commercial strain provided ahipa with sufficient N for growth and development (Kjaer, Reference Kjær1992). However, the same strain had not performed well under field conditions (Leidi, Reference Leidi2001) and showed lower N2 fixation rates than other strains under controlled conditions (Tables 2 and 3). Yield improvement after rhizobial strain selection has been shown for most legume species (O´Hara et al., Reference O´Hara, Yates, Howieson and Herridge2002; van Kessel and Hartley, Reference Van Kessel and Hartley2000). Our results have shown that it is also the case for ahipa, in which an effective symbiosis improves tuberous root and seed yield, and seed protein and root N concentration (Table 4).
Using the natural 15N abundance method, the inoculation of ahipa with selected strains provided as much as 64 % of seed N (Table 5). Roots were more enriched than seeds in N derived from N2 fixation and presented lower δ15N. Differences in isotope fractionation among organs have also been found in other legumes, but the reasons remain unclear (Yoneyama et al., Reference Yoneyama, Kounosuke, Yoshida, Matsumoto, Kambayashi and Yazaki1986). Our observation on fractionation is just preliminary as it was performed across strains. For other legumes, it has been reported that fractionation may be affected by rhizobia strain or plant genotype (Kyei-Boahen et al., Reference Kyei-Boahen, Slinkard and Walley2002; Steele et al., Reference Steele, Bonish, Daniel and O'Hara1983). The proportion of plant N derived from N2 fixation (P values) obtained in this study agree with those reported for soybean (Peoples and Herridge, Reference Peoples and Herridge1990) and were close to the 55–69% reported for ahipa using the 15N isotopic dilution method (Castellanos et al., Reference Castellanos, Zapata, Badillo, Peña-Cabriales, Jensen and Heredia-García1997).
The ureide content in the xylem sap might be overestimating N2 fixation as the method provides only a short-term measure of symbiotic dependence and the sampling was performed at flowering/fruiting, at times when the highest rate of N2 fixation is expected (Peoples and Herridge, Reference Peoples and Herridge1990). In fact, the average 80% for xylem sap N derived from fixation (XRU) in the inoculated treatments (Table 5) was higher than the N2 fixation estimate obtained by the natural 15N abundance method (P). However, the XRU was a reliable predictor for N2-fixing and provided good enough values considering the low cost and simplicity of analyses (Herridge et al., Reference Herridge, Bergersen and Peoples1990). A significant correlation was observed between XRU and P (r = 0.99, p < 0.01, n = 4) as reported by other authors using different legume species showing the close agreement between values of N2 fixation estimated by the ureide content in xylem saps and 15N techniques (Hansen et al., Reference Hansen, Rerkasem, Lordkaew and Martin1993; Herridge et al., Reference Herridge, Bergersen and Peoples1990).
The total amount of N2 fixed by inoculated ahipa in harvestable organs, calculated from root and seed dry matter yields (Table 4) and the N concentration and P values presented in Table 5, reached 156–260 kg N ha−1. These amounts agree reasonably well with the roughly 100–260 kg N ha−1 calculated from the N-difference between inoculated and non-inoculated treatments. The amount of N2 fixed by ahipa might be considered quite high according to published data on other annual legumes (LaRue and Paterson, Reference LaRue and Paterson1981; Peoples and Herridge, Reference Peoples and Herridge1990; Unkovich and Pate, Reference Unkovich and Pate2000). These values agree with the amount of N2 fixed by yambean (P. erosus) but are much higher than those reported for ahipa by Castellanos et al. (Reference Castellanos, Zapata, Badillo, Peña-Cabriales, Jensen and Heredia-García1997) who compared the two species. A reasonable explanation for the disagreement might be in the lower efficiency of soil rhizobia infecting ahipa roots, evolved after years of yambean cultivation. Such native strains would not be as effective for ahipa as those selected and tested in our experimental conditions. Similarly, the symbiotic behavior of a commercial strain (Spec 1) was overcome by other strains in providing greater shoot growth or leaf N concentration to ahipa landraces (Tables 2 and 3).
It should be emphasized that the total amount of N2 fixed might be much higher if considering what is left in the field. The amount of N remaining in soils after legume cropping (shoots, secondary roots, nodules) is an important contribution to soil N economy (Jensen and Hauggaard-Nielsen, Reference Jensen and Hauggaard-Nielsen2003). Below-ground N, that is N in roots and nodules, may represent up to 50% of the total plant N (Khan et al., Reference Khan, Peoples, Chalk and Herridge2002). A significant amount of biomass (rather rich in N and composed of shoots, shed leaves, empty pods) remains in the field. A gross assessment of soil N balance after cropping ahipa was calculated to ascertain its impact on N availability after harvest (Table 6). The N balance presented in Table 6 indicates that an effectively inoculated crop may leave up to 50 kg N ha−1 in the soil as crop residues without considering below-ground N. The resulting positive N balance is similar to values reported for other legume seeds, like groundnut and soybean but greater than those of common bean or pigeon pea (Peoples and Herridge, Reference Peoples and Herridge1990).
Gross estimation of the total N exported in seeds and roots at harvest of uninoculated (control) and inoculated ahipa landrace AC521, N recovery (total N remaining in dried shoots on the soil) and calculation of fixed N2 using the P values reported in Table 5. The amount of N remaining in the soil in secondary roots or pod walls after threshing was not considered. The gross N balance was then calculated as follows: N balance = (N recovery + N2 fixed seed + N2 fixed root) − (seed removal + root removal).
In world regions with weathered soils where ahipa is being introduced as an additional source for carbohydrates and proteins for food and/or fodder (Fassola et al., Reference Fassola, Pachas, Rohatsch, Uset and Wiss2007; Zanklan et al., Reference Zanklan, Ahouangonou, Becker, Pawelzik and Grüneberg2007), seed inoculation is a low cost technique that might ensure tuberous-root production while improving soil fertility. In European agriculture, ahipa might find a niche for raw material production in environmentally sound farming systems or in marginal lands for low input agriculture. The productivity of available ahipa accessions is quite competitive in comparison with traditional crops (sugar beet, potato) and important yield increases might be expected following modern breeding (Leidi et al., Reference Leidi, Rodríguez-Navarro, Fernández, Sarmiento, Semedo, Marques, Matos, Ørting, Sørensen and Matos2004).
Seed inoculation ensures availability (in number and quality) of specific rhizobia in the rhizosphere, which permits the establishment of nodules formed by the selected strain (Hardarson and Atkins, Reference Hardarson and Atkins2003). It is essential for early nodulation when no soil rhizobia are available or to displace low-efficient natural populations (van Kessel and Hartley, Reference Van Kessel and Hartley2000). Inoculation of ahipa with highly effective N2-fixing strains may certainly increase crop yield and provide a positive soil N balance after cropping.
CONCLUSIONS
It has been shown that inoculation of ahipa with selected rhizobia strains may meet crop N demand for growth and increases seed and root yield. Furthermore, the gross estimate of soil N balance indicates cropping inoculated ahipa may even improve soil N fertility. A simple method, like the ureide content in the xylem sap, may be a helpful complementary tool to assist testing new inoculants for ahipa in field conditions.
Acknowledgements
The collection of rhizobia provided by M. Grum (IPGRI-CIAT) is greatly appreciated. Plant analysis by A. de Castro and field work by J. Cobo and J. M. Vidueira is acknowledged. The research was supported by EU Grant FAIR6-CT98-4297.
