Differences in ureide and amino acid content of water stressed soybean inoculated with Bradyrhizobium japonicum and B. elkanii

The objective of this work was to study the response to water stress of a drought sensitive soybean cultivar inoculated with Bradyrhizobium japonicum (strain CB1809, Semia 586) and B. elkanii (strain 29W, Semia 5019). CB1809 nodulated plants produced a significantly higher root fraction (19%) than 29W (14.6%). Plants inoculated with CB1809 produced less nodules and accumulated more nitrogen than those inoculated with 29W. In general, low amounts of ureides in nodules were found in watered plants inoculated with either CB1809 or 29W strains, but those levels were five-fold increased in stressed plants inoculated with CB1809. Nodules formed by strain CB1809 had aspartate and glutamate as major amino acids, while those formed by 29W had glutamate, asparagine and alanine. In nodules of plants inoculated with CB1809 aspartate showed the highest accumulation (5 μmol g-1); in stressed plants this amino acid reached a value of 26 μmol g-1, and asparagine was not detected. Nodules formed by the strain 29W accumulated 1 μmol g-1 of aspartate, whether plants were stressed or not. Asparagine was the major amino acid found in nodules from watered plants (6 μmol g-1) and the amount of this amino acid was six-fold increased when plants were water stressed.


Introduction
The use of inoculants for increasing soybean yield in Brazil is well established.High yields (up to 4000 kg ha -1 ) can be obtained and, normally, mineral N is not used.Soybean plants can accumulate up to 40% of protein in grains, and almost all grain N is from N 2 fixation, representing a N-fertilizer economic benefit that saves over US$ 2.5 billion per year (Alves et al., 2003).There is a great variability in nodule efficiency among strains of Bradyrhizobium japonicum and B. elkanii; CB1809 is one of the most efficient of them (Dobereiner et al., 1970;Neves et al., 1985;Ramos & Ribeiro Júnior, 1992).Neves et al. (1985) distinguished two groups of soybean rhizobial strains for N 2 fixation: those with higher efficiency (CB1809, DF 383 and 965) and those with lower efficiency (29W, DF 395 and SM1b).Plants inoculated with the more efficient strains had higher ureide content of xylem sap, and higher relative efficiency (less H 2 evolved per N 2 fixed) than the others.The best strains increased yield by up to 30%.CB1809 strain normally produces few nodules, has high nodule efficiency and high N accumulation averaged over cultivars, while 29W produces a large number of nodules, has low nodule efficiency and low N accumulation in the plant (Ramos & Ribeiro Junior, 1992).The superiority of CB1809 has also been confirmed in field experiments (Oliveira et al., 1991).
There are several works showing that soybean cultivars can vary in their capacity of N 2 fixation (Bohrer & Hungria, 1998) and tolerance to water stress (Salinas et al.,1996;Serraj & Sinclair, 1998;Sinclair et al., 2000).A large variation in nodulation sensitivity to water deficit, among soybean cultivars, can be related to nodule formation and growth (Serraj & Sinclair, 1998), and soybean genotypes can be selected for N 2 fixation and tolerance to water deficits (Sinclair et al., 2000).
It is not known whether soybean plants from the same cultivar, inoculated with contrasting strains can alter its growth and physiological behaviour under water stress.
The aim of this work was to study the effect of water stress on a drought sensitive soybean cultivar inoculated with the strains 29W and CB1809.

Material and Methods
Seeds of soybean (Glycine max), cultivar Bragg, were surface sterilized in 80% ethanol for 30 seconds, in 5% sodium hypochlorite for 2 minutes and then washed ten times in sterilized distilled water.Five days after germination, seedlings were inoculated with either 1 cm 3 culture of Bradyrhizobium japonicum (CB1809) or B. elkanii (29W).The strains were obtained from Embrapa Agrobiologia, Seropédica, Brazil.
Each strain was grown in Yeast Extract/Mannitol broth (YEM) (Vincent, 1970) for eight days at 28 o C, on a shaker at 140 rpm, and then 1 cm 3 containing 2.1x10 9 viable cells was used to inoculate each of the seedlings.
Seeds were planted in Leonard jars (Vincent, 1970), containing a mixture of washed sand and vermiculite (1:1 v/v).Jars were previously sterilized by autoclaving at 121 o C for 1 hour, and then 200 cm 3 of nutrient solution (Summerfield et al., 1977) was added.The nutrient solution was prepared in a disinfected container using deionized water.Two seeds per jar were planted and inoculated, and after 5 days they were thinned to one.Sterilized sand was added to the top of the jars in order to avoid cross contamination.Jars with plants without inoculation were grown among inoculated ones, as a control.
Plants were grown in a glasshouse with supplementary light from Phillips SON T lamps giving a quantum irradiance of 500 to 1,000 µmol m -2 s -1 on the top of the plants.The photoperiod was 14 hours, day/night temperatures were 25ºC-30 o C and 15ºC-18 o C, respectively; the relative humidity was maintained between 40%-50%.
The experimental design was a 2x2 factorial, with two strains (CB1809 or 29W) and two levels of water supply (water and stress).Jars were arranged in ten randomized blocks.Stress treatment started 25 days after planting, by withholding water from half of the jars.Thirty days after planting, plants were harvested.Half of them were used for chemical analysis and the other half for dry weights and N content in plant parts.Statistical analyses were made, using ANOVA test.
Leaves were excised and covered immediately with a clean plastic film to avoid dehydration until the water potential was measured in the youngest fully expanded leaf by using a pressure bomb (Scholander et al., 1965).Leaf area was measured using a Delta T meter.
Plant samples were dried in a fan-assisted oven, at 65 o C for 36 hours.They were then ground in a ball mill to a fine powder and the percentage of N was determined in 600-800 mg aliquots, using a Carlo Erba Elemental Analyzer (Model 1106), with atropine as standard.
A leaf just below the youngest fully expanded one was excised: part of the leaf was weighed and immediately transferred to a mortar containing 5 cm 3 of 400 mol m -3 potassium phosphate buffer at pH 7.0.After grinding, samples were filtered through cotton cloth and immediately placed on ice.They were later transferred to a freezer at -20 o C. The same procedure was carried out for nodules.Allantoin and allantoic acid were determined colorimetrically following the method of Vogels and Drift (1970), using pure standards from BDH Ltd.
Parts of the trifoliate leaves were used to quantify amino acids.Samples were weighed and immediately frozen in liquid N 2 , then they were kept in a freezer at -20 o C. Approximately 0.5 g of leaves or nodules were ground with a pestle and mortar in 7.5 cm 3 (3 cm 3 for nodules) of 80% ethanol acidified with 0.25 mol dm -3 HCl.Then, 750 mm 3 of internal standard (2.5 mmol dm -3 nor-leucine) was added, and samples were dried in a rotary evaporator for approximately 20 minutes, at 45 o C. Samples were resuspended in 1 cm 3 acidified ethanol and kept overnight in a freezer.Nodule samples were centrifuged twice in an eppendorf centrifuge and kept in a freezer (-20 o C) for 24 hours before analysis.
A standard solution (5 mm 3 ) of 18 amino acids (Sigma), each one with a concentration of 2.5 mmol dm -3 was used.Asparagine, glutamine and nor-leucine (as internal standard) were added to the amino acid solutions, giving a final concentration of 2.5 mmol dm -3 .Samples were derivatized with phenylisothiocyanate (PITC) to give phenylthiocarbamyl amino acids (Bidlingmeyer et al., 1984).All derivatized samples were kept in freezer (-20 o C) for one or two weeks before running.Analysis was conducted on a Waters HPLC system, using a reverse phase, Nova-Pak C 18 column (3.9x150 mm) at 45 o C, with a gradient of acetate buffer (aqueous phase) and acetonitrile/methanol (organic phase).Amino acids were detected by UV absorbance (254 nm) and quantified by reference to the internal standard, norleucine.Statistical analyses were made, using ANOVA test.

Results and Discussion
Rhizobia species had no significant effect on leaf water potential, leaf area or shoot/root ratio (Table 1).Water stress significantly decreased leaf water potential, leaf area and shoot/root ratio.There were no significant interactions between strain and stress for those variables.
Nodules from stressed plants had significantly lower total and individual fresh weight (Table 2).There was a highly significant strain and stress interaction for total nodule fresh weight per plant.
Plants inoculated with CB1809 accumulated more N than those with 29W, as found by Neves et al. (1985) and Ramos & Ribeiro Júnior (1992).
Plants inoculated with 29W produced greater nodule dry weight than those inoculated with CB1809, especially when plants were stressed, representing 10% of total plant weight (Table 3).Plants under water stress, inoculated with CB1809, produced more roots than those inoculated with 29W.
Total plant dry weight was unaffected by strain (Table 3).Also, no difference was found in nodule dry weight between watered and stressed plants inoculated with CB1809, but 29W had lower nodule dry weight in stressed plants than watered ones (Table 3).
Table 1.The effect of rhizobial strain and water stress on leaf water potential (LWP), leaf area (LA) and ratio shoot/root (S/R) of soybean (1) .Low amounts of ureide content were found in nodules from watered plants (Table 4).However, during stress treatment plants inoculated with CB1809 had increased ureide content 3,3 fold, while stressed nodules from 29W had similar amounts in watered plants.Gonzales et al. (1995) also observed an increase in ureides from nodules of stressed soybean (cv.Clark) inoculated with strain RCR3407.
Ureide content in leaves (Table 4) was higher in watered plants inoculated with CB1809, and ureide content was lower in watered plants inoculated with 29W and higher in stressed plants.These results suggest that Bradyrhizobium strains had an opposite response when plants were stressed.Rao & Venkateswarlu (1987) observed an increase of ureide contents in nodules and shoots of two crop legumes (mung bean and moth bean) under water stress, mainly at high levels of stress (between -1.3 and -1.6 MPa).It appears that grain legumes transporting fixed N as amides, such as chiken pea, faba bean and lupine, are less sensitive to water stress than those transporting ureides, such as soybean.
Amino acids content in nodules and leaves are shown in Tables 5 and 6.In general, nodules formed by CB1809 and 29W showed that different amino acids were produced, depending on strain inoculated (Table 5).Greenwood & Bathurst (1978) did not find differences in major amino acids extracted from nodules of plants inoculated with nine different B. japonicum strains.
Total amount of amino acids in nodules was increased about threefold by stress (Table 5).Nodules formed by strain CB1809 had aspartate and glutamate as the major amino acids, while those formed by 29W had glutamate, asparagine and alanine.
In watered plants inoculated with CB1809, aspartate was the amino acid which showed highest accumulation (5 µmol g -1 FW); in stressed plants this amino acid reached values of 26 µmol g -1 .In watered or stressed nodules asparagine could not be detected.Based on a small peak in that region, asparagine levels were estimated as <0.8 µmol g -1 .
It was possible to estimate the aspartate/asparagine ratio in nodules from plants inoculated with CB1809: in watered plants this ratio was 6:1, and in stressed ones ratio was 153:1.In plants inoculated with 29W asparagine was the major amino acid in nodules from watered plants (6 µmol g -1 ), and the amount of this amino acid was increased six fold when plants were stressed.The aspartate/asparagine ratio was 0.15 in watered nodules, and 0.032 in those stressed and inoculated with 29W.
These results suggest that these strains have different metabolism when associated with soybean plants; it is not clear why symbiotic association between strain CB1809 and the plant produces high amounts of aspartate instead of asparagine, and why the symbiotic association between strain 29W and the plant produces exactly the opposite (high amounts of asparagine in stressed plants and, in general, low amounts of aspartate).
Asparagine is synthesized by asparagine synthetase and it has been observed that asparagine levels in plant tissues normally increase under stresses such as salt, drought and mineral deficiencies (Ireland, 1990).Miflin & Lea (1977) suggest that asparagine can act as a temporary store of reduced N, particularly in conditions of carbon limitation, and the plant should be able to synthesize this amino acid in large amounts to avoid NH 4 + toxicity.They also suggested that if sufficient fixed carbon is available for amino acid and protein synthesis, asparagine formation should be inhibited.Since the enzyme is present in the plant, bacteria could affect plant ability to synthesize these compounds.
There are several pathways for plant aspartate production.It can be produced by the enzyme aspartase Table 4. Ureide content (µmol g FW -1 ) of watered (W) and water stressed (S) soybean inoculated with Bradyrhizobium japonicum (CB1809) and B. elkanii (29W) (1) . (1)Means followed by the same letter in each row do not differ statistically by Duncan's test at 5% of probability.
Table 5.Amino acid content (µmol g -1 FW) in nodules of watered (W) and water stressed (S) soybean inoculated with Bradyrhizobium japonicum (CB1809) and B. elkanii (29W) (1) .(ammonia assimilating enzyme) or by aspartate dehydrogenase (Ireland, 1990); these enzymes have been detected in few plants, but it appears that they do not have a significant contribution to ammonia assimilation (Ireland, 1990).Levels of proline increased in stressed nodules from both strains (Table 5).Kohl et al. (1991) observed high amounts of proline in soybean nodules under water stress.They also observed that an increase in P5CR (pyrroline-5-carboxylate reductase -an enzyme related to proline metabolism) activity in nodules was not correlated with drought.This enzyme had an increase of 55% in stressed 25-day plants, but not in 55-day-old plants.They argued that an increase in P5CR activity in nodules, in response to drought stress, is not necessary for proline accumulation.
Amino acid content in leaves (Table 6) showed, in general, higher values in stressed plants.The concentration of the following amino acids increased in leaves of stressed plants inoculated with CB1809: glutamate, asparagine, threonine, proline, isoleucine, leucine and phenylalanine.In stressed plants inoculated with strain 29W, levels of asparagine, glycine, glutamine, proline, valine, isoleucine, leucine and phenylalanine increased.
Proline was the amino acid with the highest accumulation in stressed leaves of plants, when they reached -1.8 MPa, representing 21% and 26% of the total amino acids detected in plants inoculated with CB1809 and 29W, respectively.Levels of this amino acid can reach up to 58% of the total amino acids detected in stressed plants (Ramos, 1996).Proline is known to accumulate in plants under water and salt stress (Fukutoku & Yamada, 1981); apparently, proline accumulation is a symptom of severe internal water deficit and not an adaptative metabolic response to stress (Hanson & Hitz, 1983).
An increase of isoleucine, valine and phenylalanine in soybean leaves was also found by Fukutoku & Yamada (1981) and Ramos (1996).It appears that these amino acids also have a role when plants are kept under water stress.
Amino acid metabolism of plants inoculated with these two strains is apparently different, and this influences amino acid composition of the nodules.Asparagine is synthesised in seeds by asparagine synthetase at germination, when protein reserves are hydrolyzed, or in roots during nitrate assimilation and N 2 fixation (Sieciechowicz et al., 1988).An alternative explanation is that these strains, in some way, influence the amino acid metabolism of plant cells.
The diverse efficiency of the symbiotic association produced by these two strains may be related to the different amino acid metabolism of their root nodules, particularly in relation to aspartate and asparagine accumulations.

Conclusions
1. Plants inoculated with CB1809 produce a higher concentration of ureides in nodules, particularly in stressed plants, than plants inoculated with 29W.
2. Plants inoculated with CB1809 produce more roots than those inoculated with 29W.
3. Strain CB1809 promotes higher nitrogen accumulation in both stressed and control plants than 29W.

Table 2 .
The effect of rhizobial strains and water stress on nodule fresh weight (NFW), individual nodule weight (INW), nodule fraction (NF), and total nitrogen in plant of soybean(1).
ns Not significant.

Table 3 .
Dry (1) Means followed by the same letter in each row do not differ statistically by Duncan's test at 5% of probability.
Means followed by the same letter in each row do not differ statistically by Duncan's test at 5% of probability.(2)Notdetected.(3)Estimatedvalues.