Inoculation with the native Rhizobium gallicum 8a3 improves osmotic stress tolerance in common bean drought-sensitive cultivar

Abstract

Symbiotic nitrogen fixation potential in common bean is considered to be low in comparison with other grain legumes. However, it may be possible to improve the nitrogen fixation potential of common bean using efficient rhizobia. In order to improve osmotic stress tolerance of a drought-sensitive common bean cultivar (COCOT) consumed in Tunisia, plants were inoculated either by the reference strain Rhizobium tropici CIAT 899 or by inoculation with rhizobia isolated from native soils Rhizobium gallicum 8a3. Fifteen days after sowing, osmotic stress was applied by means of 25 mM mannitol (low stress level) or by 75 mM mannitol (high stress level). Fifteen days after treatment plants were harvested and different physiological and biochemical parameters were analysed. Results showed no significant differences between the studied symbioses under control conditions. However after exposure to osmotic stress our results showed better tolerance of COCOT to osmotic stress when inoculated with the native R. gallicum 8a3. This can be partially explained by better water-use efficiency in both leaves and nodules, better relative water content in nodules and better efficiency in utilization of rhizobial symbiosis as compared with COCOT-CIAT 899 symbiosis. Hence, the present study suggested the better use of native soil isolated strains for the inoculation of common bean in order to improve its performance and nitrogen fixation potential under stressful conditions.

Keywords:

• Common bean
• improvement
• mannitol
• nitrogen fixation
• osmotic stress

Abbreviations

ARA	=	Acetylene reduction activity
DAS	=	Days after sowing
DW	=	Dry weights
EURS	=	Efficiency in utilization of the rhizobial symbiosis
FW	=	Fresh matter weight
LRWC	=	Relative water content of leaves
NAR	=	Net assimilation rate
NDW	=	Nodule dry weights
NF	=	Nitrogen fixation
Nn	=	Nodule number
NRWC	=	Relative water content of nodules
NWUE	=	Water-use efficiency in nodules
PWUE	=	Plant water use efficiency
SLA	=	Specific leaf area
SNF	=	Symbiotic nitrogen fixation
SWUE	=	Water-use efficiency in shoots
TSS	=	Total soluble sugars
TW	=	Turgid fresh matter weight.

Introduction

Common bean (Phaseolus vulgaris L.), a traditional crop originating from Latin America, is the most important food legume for human consumption worldwide, especially in Africa, where its cultivation as a staple food extends into marginal areas. Symbiotic nitrogen fixation (SNF) potential in common bean is considered to be low (Pereira and Bliss 1987, Isoi and Yoshida 1991) in comparison with other grain legumes. However, it may be possible to improve the SNF potential of common bean (Bliss 1993, Hardarson et al. 1993). Yield potential of legumes depends on the rhizobia association and plant genotype which together influence the symbiotic performance (Sadiki and Rabih 2001, Mhadhbi et al. 2004). Therefore, inoculation with efficient rhizobia might improve symbiotic nitrogen fixation (SNF) and productivity of common bean.

In the Mediterranean zone, little or no rainfall occurs during extended periods of the year. Tunisia is mostly located in the semi-arid, arid and Saharan climatic zones where the annual rainfall varies from 300 to less than 100 mm (Le Houérou 1990). Water deficits (commonly known as drought) can be defined as the absence of adequate moisture necessary for plants to grow normally and complete their life cycle (Zhu 2002). The lack of adequate moisture leading to water stress is a common occurrence in rainfed areas, brought about by infrequent rains and poor irrigation (Wang et al. 2005). Common bean appears to be particularly sensitive to this stress (Kirda et al. 1989) with considerable reduction in N₂ fixation (Ladrera et al. 2007, Sassi et al. 2008b) as a consequence of changes in nitrogenase activity and nodule biomass (Gálvez et al. 2005).

Water deficiency and drought directly affect nodule activity and function (Davey and Simpson 1990). Regardless of the physiological mechanism of N₂ fixation inhibition by drought stress, there is evidence that legume species have significant genetic variation in their ability to fix N₂ under drought conditions, e.g. Pimentel et al. (1990). Several studies have explained the effect of water stress on plant physiology and SNF in common bean (Ramos et al. 2003, Gálvez et al. 2005), nevertheless, few studies have been conducted under hydroaeroponic conditions. Hydroaeroponic environment enables the comparison of different symbiotic associations and the selection of the most tolerant symbiosis under stressed conditions (Jebara et al. 2001) which are the main objectives of this study.

In Tunisia, Mhamdi et al. (2002) showed that P. vulgaris is nodulated by a diversity of species including Rhizobium gallicum, R. leguminosarum bvs. Phaseoli and viciae, R. etli, R. giardinii, Sinorhizobium fredii, S. meliloti and S. medicae and Mnasri et al. (2007) showed the efficiency of the R. gallicum for bean cultivation. The present work focused on the enhancement of osmotic stress tolerance of a drought-sensitive cultivar consumed in Tunisia by inoculation with rhizobia isolated from native soils R. gallicum 8a3 and compared with inoculation by the reference strain R. tropici CIAT 899. For that purpose, we analysed several physiological and biochemical traits in order (i) to look for the main traits inducing osmotic stress tolerance amelioration, (ii) to understand the likely mechanisms involved in such improvement and (iii) to determine useful criteria for genetic improvement of drought tolerance.

Methods

Plant growth and conditions for imposing osmotic stress

The biological material was bean (Phaseolus vulgaris L.) seeds of COCOT blanc (provided by M. Trabelsi, ESA Mateur, Tunisia). Seeds were surface sterilized and pre-germinated in agar 0.9% then transferred in 1 dm³ glass bottles wrapped with aluminium foil to maintain darkness in the rooting environment. The nutrient solution contained 0.25 mM KH₂PO₄, 0.7 mM K₂SO₄, 1 mM MgSO₄·7H₂O, 1.65 mM CaCl₂, 22.5 µM Fe for macronutrients, and 6.6 µM Mn, 4 µM Bo, 1.5 µM Cu, 1.5 µM Zn, 0.1 µM for micronutrients. Medium pH was maintained at 7.0 by adding 0.2 g dm^-3 CaCO₃. It was aerated with a flow of 400 cm³ min^–1 of filtered air via a compressor and ‘spaghetti tube’ distribution system. Plants were grown in a temperature-controlled glasshouse with night/day temperatures of c. 20/28 °C, relative humidity 90/75% and a 16 h photoperiod. The irradiance was supplied by mercury vapour lamps (OSRAM HQI-T400W/DH).

Osmotic stress treatments

Osmotic stress was applied by means of 25 mM mannitol (low osmotic stress level) or 75 mM mannitol (high osmotic stress level). The water potentials of nutrient solutions were: -0.5 MPa, −0.7 MPa and -1.2 MPa for control, 25 mM and 75 mM mannitol respectively. Mannitol is an osmotic component used generally to generate osmotic stress when added to nutrient solution. Mannitol was added to 15-day-old plants corresponding to the initial period of nodules formation and the establishment of N₂ fixation. Simultaneously, nitrogen source was provided to plants as 1 mL of (Rhizobium tropici ) CIAT 899 or local (Rhizobium gallicum) 8a3 strain that was previously isolated from the Cap Bon region in Tunisia, characterized at the phenotypic and molecular levels by Mhamdi et al. (1999) and kindly provided and maintained in culture in the Laboratory of Legumes Micro-organisms Interactions (LILM), Centre of Biotechnology Borj Cedria (CBBC). At the beginning of flowering, 30 DAS, plants were harvested for growth parameters determination and compared with non-stressed plants (controls).

Dry weight and leaf area

After harvest, different plant parts were separated. Leaves, roots and nodules were then weighed for fresh weight determination. Leaf areas were determined with a portable Area Metre (Model LI-3000A, LI COR). Dry weights (DW) of different plant parts were determined after drying for 3 days at 70 °C.

Relative water content

The relative water content of leaves (LRWC) and nodules (NRWC) were measured respectively in the second or third youngest fully expanded leaf that was harvested in the morning and on fresh nodules harvested at the end of the treatment period. This parameter was determined using the following equation:

FW is the fresh matter weight determined within 2 h after the harvest and TW stands for the turgid fresh matter weight (Schonfeld et al. 1988). TW was obtained after soaking the leaves in distilled water in test tubes for 4 h at room temperature (c. 20 °C) under low light condition or after 16 h at 4 °C for nodules. Later, leaves and nodules were quickly and carefully blotted dry with tissue paper for determining turgid weight.

The water-use efficiency

Water-use efficiency in shoots (SWUE) and nodules (NWUE) means were calculated as the ratios of the total dry mass produced over the total water used (Boyer 1996).

Proline assay

Free proline was quantified spectrophotometrically using the method of Bates et al. (1973). The protocol is based on the formation of red coloured formazone by proline with ninhydrin in acidic medium, which is soluble in organic solvents like toluene.

Nitrogenase activity and nitrogen fixed

Three plants from each treatment were used for the assessment of nodule nitrogenase activity (EC 1.7.9.92) estimated by the acetylene reduction assay (Hardy et al. 1968). Ten per cent of C₂H₂ was added to the nodulated-root atmosphere and, after incubation, the rate of ethylene evolution was measured using a Hewlett-Packard 4890 gas chromatograph equipped with a Porapak-T column. This assay has been shown to be prone to give inaccurate nitrogen fixation (NF) measurements and, therefore, absolute values may not be reliable (Minchin et al. 1983). However, the results obtained herein are consistent with plant biomass and nitrogen content parameters. Nitrogen content was determined, according to the Kjeldahl method, at the beginning and the end of the osmotic treatment. Nitrogen fixed was then calculated as the N content at harvest minus the N content of the plants at the onset of the treatment (Sassi et al. 2008a).

Total soluble sugars determination

Total soluble sugars were quantified using the anthrone method. The 20 mg DW homogenate in deionized water was incubated in a water bath at 70 °C then centrifuged at 3000 g for 10 min. 100 µL of the supernatant was added to 4 ml of anthrone solution (0.15 g anthrone in 100mL 80% H₂SO₄) and incubated in a boiling water bath. The absorbance of the samples was determined spectrophotometrically at 620 nm using glucose as standard (Aydi et al. 2010).

Extraction and assay of leghaemoglobin

Nodules (100 mg) were homogenized in a mortar and pestle with 3 mL Drabkin's solution. The Drabkin's solution is prepared with 52 mg KCN, 198 mg K₃Fe(CN)₆ and 1 mg NaHCO₃ in 1 L distilled water. The homogenate was centrifuged for 15 min at 500 g, samples of the supernatant were adjusted to 10 mL by Drabkin's solution, then centrifuged at 20?000 g for 30 min. The absorbance of supernatant was measured at 540 nm against the Drabkin's solution (Shiffmann and Löbel 1970).

Efficiency in utilization of the rhizobial symbiosis (EURS)

The EURS was estimated by the slope of the regression model of shoot biomass as a function of nodule biomass. For a linear adjustment-curve, i.e. y = ax + b, b corresponds to the shoot biomass production without nodules (g sDW₀), and a corresponds to the EURS as (g sDW –g sDW₀) g^–1 nDW (Aydi et al. 2004).

Statistical analysis

Statistical analysis was carried out using the Statistica software (version 5, StatSoft, France). The analysis of variance (ANOVA) and the lowest standard deviation (LSD) of the means were used to determine statistical significance (p ≤ 5%) between treatments. Data are presented as mean values of six replicates (three for ARA, total soluble sugar and proline content) and their corresponding standard errors.

Results

Symbioses effects on growth response to osmotic stress

Table I summarized growth parameters measured on plants growing on control nutrient solution and those growing on the same solution with 25 mM mannitol (low osmotic stress) or 75 mM mannitol (high osmotic stress). Under controlled conditions, plant dry weight, leaf area, leaf number and root to shoot ratio (R/S) did not exhibit any changes according to the inoculated rhizobia strain. Similarly, the net assimilation rate (NAR) and the specific leaf area (SLA) showed no significant difference between symbioses. Such data showed the similar effect of the studied rhizobia strain on cv. COCOT grown under control conditions. Conversely, after 30 days of exposure to both osmotic stress levels, a different pattern was observed (Table I). Independently of the associated rhizobia osmotic stress limited significantly all growth parameters of the common bean cultivar. However, it seems that symbiosis implicating R. gallicum 8a3 strains was more tolerant than the R. tropici CIAT one. Plant dry weight decreased at low (25 mM) and high (75 mM) mannitol-induced osmotic stress in both symbioses. However, COCOT-8a3 symbiosis exhibited lower inhibition rate mainly at high stress level as compared with COCOT-CIAT symbiosis, reaching 42% in the former while it exceeded 60% in the latter. This was reflected by keeping better leaf area and leaf number under stressed conditions. Data herein showed increased root to shoot ratio (R/S) (Table I). This parameter showed its highest increase under 75 mM mannitol in COCOT-8a3 symbiosis. Decreased net assimilation rate (NAR) and specific leaf area (SLA) were also reported in Table I, this decrease being higher in COCOT-CIAT symbiosis mainly under severe stress conditions where inhibition rates reached 71% and 64% respectively.

Table I. Effect of osmotic stress on growth parameters: PDW (plant dry weight, g. plant⁻¹); Leaf area (cm² plant⁻¹); Leaf number (Leaf plant⁻¹); R/S (root to shoot ratio); NAR (Net assimilation rate, g DW cm⁻² day⁻¹) and SLA (Specific leaf area, cm² g LDW⁻¹) in a drought-sensitive common bean ‘cv. Coco blanc’ inoculated with CIAT (reference strain) or 8a3 (native strain) and submitted to low (25 mM) and high (75 mM) mannitol-induced osmotic stress during 15 days. Values represent mean±SE (n = 6). Numbers followed by a different letter within a column are significantly different at p ≤ 0.05 according to LSD analysis.

Symbiosis	Coco-CIAT			Coco-8a3
Mannitol (mM)	0	25	75	0	25	75
PDW	2.8±0.2 a	2.1±0.1 b	0.9±0.2 d	3.3± 0.2 a	2.8±0.2 b	1.9±0.1 c
Inhibition rate (%)	–	25	68	–	15	42
Leaf area	622±22 a	506±17 b	310±9 c	655±242 a	623±21 a	487± 12 b
Inhibition rate (%)	–	19	50	–	12	26
Leaf number	10±2 a	7±1 b	4±1 c	10±2 a	8±2 a	6±1 b
Inhibition rate (%)	–	30	60	–	30	50
R/S	0.2±0.1 c	0.3± 0.1 b	0.4±0.1 a	0.2±0.1 c	0.3±0.1 b	0.5±0.1 a
Inhibition rate (%)	–	1.5	2	–	1.5	2.5
NAR	0.14±0.1 a	0.11± 0.1 b	0.04±0.1 c	0.17±0.1 a	0.15±0.1 a	0.08±0.1 b
Inhibition rate (%)	–	21	71	–	12	53
SLA	650±25 a	534±21 b	231±8 d	679±22 a	600±21 a	344±11 c
Inhibition rate (%)	–	18	64	–	12	49

Symbioses effects on nodule performance under osmotic stress

Nodule growth and NF parameters changes between control and stressed conditions are given in Table II. In general, under control conditions, no obvious differences were observed between both studied symbioses. Furthermore, low osmotic stress level (25 mM mannitol) did not discriminate well between symbioses since data did not reveal large differences. Actually, at high stress level (75 mM mannitol), although the inhibition rates were lower than those of plant growth, osmotic stress induced significant inhibition of nodule dry weights (NDW), which in turn strongly inhibited nitrogenase activity assayed by acetylene reduction activity (ARA). The superiority of COCOT-8a3 symbiosis is mirrored by minor NDW decreases with only 40% paralleled with lower ARA inhibition rates not exceeding 51% compared with 58% and 72% respectively in COCOT-CIAT symbiosis. In opposition, marked inhibition was observed in the nodule number (Nn) mainly in COCOT-8a3 symbiosis under severe stressed conditions reaching up to 70%. In addition, the above-mentioned symbiosis nodules kept higher leghaemoglobin content. Nodule to root ratio (N/R) was also determined to verify whether the inhibition of nodule number under osmotic stress was mainly due to lower root surface or lower aptitude of nodule establishment by roots. The data reported in Table II showed decreased N/R with increasing osmotic stress level. This result revealed that osmotic stress inhibited more strongly the establishment of new nodules generation than the root growth itself.

Table II. Effect of osmotic stress on nodule performance: NDW (Nodule dry weight, mg plant⁻¹); nodule number (nodule plant⁻¹); ANW (average nodule weight, mg nod⁻¹); N/R (Nodule to root ratio); ARA (Acetylene reduction activity, µmol C₂H₄ h⁻¹ plant⁻¹); Fixed N (mmol Plant⁻¹); Lb (Leghaemoglobin, mg gFW⁻¹) and EURS (efficiency of utilization of the rhizobial symbiosis) in a drought-sensitive common bean ‘cv. Coco blanc’ inoculated with CIAT (reference strain) or 8a3 (native strain) and submitted to low (25 mM) and high (75 mM) mannitol-induced osmotic stress during 15 days. Values represent mean±SE (n = 6) only for ARA determination (n = 3). Numbers followed by a different letter within a column are significantly different at p ≤ 0.05 according to LSD analysis.

Symbiosis	Coco-CIAT			Coco-8a3
Mannitol (mM)	0	25	75	0	25	75
NDW	200±18 a	165±12 b	98±9 d	210±15 a	188±11 a	131±08 c
Inhibition rate (%)	–	18	51		10	38
Nod. number	225±24 a	109±13 b	85±6 c	218±19 a	101±9 b	66±5 d
Inhibition rate (%)	–	52	62	–	54	70
ANW	0.9±0.01 b	1.5±0.05 a	1.2±0.05 b	1.0±0.03 b	1.9±0.07 a	2.0±0.08a
Inhibition rate (%)	–	–70	–30	–	–93	−106
N/R	0.25±0.1 a	0.22±0.1 a	0.12±0.1 b	0.23±0.1 a	0.20±0.1 a	0.12±0.1 b
Inhibition rate (%)	–	12	52	–	13	48
ARA	19±2 a	15±2 a	5±1 c	19±1 a	17±2 a	8±1 b
Inhibition rate (%)	–	17	72	–	11	58
Fixed N	8±2 a	5±1 b	2±2 c	8±3 a	6±2 a	4±1 b
Inhibition rate (%)	–	29	71	–	14	43
Lb	25±5a	21±5 a	10±2 b	28±4 a	25±4 a	16±3 b
Inhibition rate (%)	–	16	60	–	11	43
EURS	1.4±0.3 a	1.3±0.4 a	0.9±0.1 b	1.6±0.4 a	1.5±0.3 a	1.5±0.1 a
Inhibition rate (%)	–	9	34	–	5	5

Symbioses effects on water relations under osmotic stress

To understand how water relations of both symbioses COCOT-CIAT and COCOT-8a3 were affected by osmotic treatment we monitored relative water content in leaves (LRWC) and nodules (NRWC) in both symbioses. Data from Table III analyses indicated that the control treatments of both symbioses showed similar values. Following exposure to osmotic stress, both parameters decreased, this decrease being higher under the high osmotic stress level. Nevertheless, the effect of osmotic stress was more pronounced on NRWC as compared with LRWC. The symbiosis COCOT-CIAT exhibited the highest decreases reaching 33% and 73% respectively at low and high osmotic stress. In line with RWC data, PWUE as well as SWUE and RWUE were affected by both osmotic stress levels while controls showed similar trends in both symbioses. The one and only difference was observed in NWUE where no significant change was detected with both osmotic stress levels, although reduction reached 63% as compared with controls.

Table III. Effect of osmotic stress on water relations: LRWC (leaf relative water content,%); NRWC (nodule relative water content,%); WUE (water use efficiency, g DW ml⁻¹) in a drought-sensitive common bean ‘cv. Coco blanc’ inoculated with CIAT (reference strain) or 8a3 (native strain) and submitted to low (25 mM) and high (75 mM) mannitol-induced osmotic stress during 15 days. L denotes leaves, N denotes nodules, S denotes shoots and IR denotes inhibition rate (%). Values represent mean±SE (n = 6). Numbers followed by a different letter within a column are significantly different at p≤0.05 according to LSD analysis.

Symbiosis	Coco-CIAT			Coco-8a3
Mannitol (mM)	0	25	75	0	25	75
LRWC	70±13 a	67±11 b	34±5 c	68±14 a	61±10 a	31±4 c
Inhibition rate (%)	–	4	51	–	10	54
NRWC	80±15 a	54±10 b	22±7 d	81±16 a	59±9 b	35±5 c
Inhibition rate (%)	–	33	73	–	27	57
PWUE	0.14±0.01 a	0.06±0.01c	0.04±0.01 d	0.17±0.01 a	0.1± 0.02 b	0.06±0.01 c
Inhibition rate (%)	–	57	71	–	41	65
SWUE	0.34±0.02 a	0.21±0.03 b	0.17±0.01 c	0.27±0.02 a	0.21±0.02 b	0.15±0.01 c
Inhibition rate (%)	–	38	50	–	22	44
NWUE	0.022±0.004 a	0.008±0.002 b	0.002±0.001 c	0.022±0.005 a	0.008±0.001 b	0.008±0.001 b
Inhibition rate (%)	–	64	91	–	64	64

Symbioses effects on osmotic adjustment under osmotic stress

The accumulation of proline and total soluble sugars either in leaves or in nodules of both symbioses is shown in Table IV. Results showed no significant differences between both symbioses under control conditions. When compared with control leaves, the accumulation of total soluble sugars (LTSS) appeared to be 3-fold and 5-fold higher in COCOT-CIAT symbiosis and 2-fold and 4-fold higher in COCOT-8a3 symbiosis respectively at low (25mM mannitol) and high (75 mM mannitol) osmotic stress levels. Similar trends were reported in nodules where increased total soluble sugars content (NTSS) exceeded 1-fold under both osmotic stress levels. On the contrary, in both symbioses, osmotic stress had a significant inhibitory effect on proline accumulation either in leaves or in nodules; this effect was more pronounced at higher osmotic stress level Likewise, data showed no obvious differences between both symbioses in terms of proline accumulation (Table IV).

Table IV. Effect of osmotic stress on osmotic adjustment: LTSS (Leaf total soluble sugar, mmol g^-1 DW); NTSS (Nodule total soluble sugar, mmol g DW ^-1); L. Proline (Leaf proline content, mmol FW g^-1); N. Proline (Nodule proline content, mmol FW g^-1); in a drought-sensitive common bean ‘cv. Coco blanc’ inoculated with CIAT (reference strain) or 8a3 (native strain) and submitted to low (25 mM) and high (75 mM) mannitol induced osmotic stress during 15 days. TRC denoted treated/control ratio. Values represent mean±SE (n = 3). Numbers followed by a different letter within a column are significantly different at p ≤ 0.05 according to LSD analysis.

Symbiosis	Coco-CIAT			Coco-8a3
Mannitol (mM)	0	25	75	0	25	75
LTSS	112±15 c	324±22 b	534±27 a	152±11 c	376±13 b	554±26 a
TCR	–	3	5	–	2	4
NTSS	274±17 c	312±22 b	422±32 a	277±13 c	300±25 b	352±21 b
TCR	–	1.1	1.5	–	1.1	1.3
L Proline	0.5±0.01 a	0.3±0.01 b	0.1±0.01 c	0.5±0.01 a	0.3±0.01 b	0.1±0.01 c
IR (%)	–	40	80	–	40	80
N Proline	1.4±0.1 c	2.8±0.3 b	5.2±0.5 a	1.7±0.1 c	2.7±0.1 b	5.9±0.4 a
TCR	–	2.0	3.7	–	1.6	3.5

Discussion

Rhizobial partner involvement in growth conservation under osmotic stress

Results presented herein revealed the negative effect of osmotic stress on all growth parameters namely PDW, LA, LN, NAR and SLA. This was in accordance with our previous data recently published using four cultivars of common bean submitted to 50 mM mannitol (Sassi et al. 2008a). Nevertheless, in the present work, plants were inoculated with two different strains separately: (Rhizobium tropici) CIAT 899 and the local one (Rhizobium gallicum) 8a3. Under unstressed conditions, both rhizobial partners showed similar behaviour, indicating no significant difference between both studied symbioses: COCOT-CIAT and COCOT-8a3. This result does not support those of Tajini et al. (2008) that showed differences between both symbioses under control conditions in the field. This could be associated to various others conditions influencing growth parameters under field conditions mainly rhizobial competitiveness (Tajini et al. 2008). However, under stressed conditions, our results showed different behaviours between both symbioses at low and mainly at high osmotic stress levels (Table I). Actually, all growth parameters declined in both symbioses, but it seems that reductions were lower in R. gallicum–COCOT symbiosis. This could indicate that this symbiosis was able to maintain higher growth potentialities even under low water availability. This can be partially explained by maintaining higher root to shoot ratio and lower leaf area reduction (26%) even at high osmotic stress level (75 mM mannitol). Indeed, it is possible that under osmotic stress the plants spend more photosynthetic energy on root production in search of water and/or reducing water loss (Kafkafi 1991), which enables common bean to avoid harmful effects of osmotic stress (Sassi et al. 2008a).

Rhizobial partner involvement in maintaining nitrogen fixation under osmotic stress

To estimate symbiotic effectiveness of both symbioses, COCOT-CIAT 899 and COCOT-8a3, nitrogen-fixing capacity and nodules features were monitored through 30 days of exposure to low and high levels of osmotic stress (Table II). It seems that under control conditions, both symbioses behaved similarly showing the typical efficiency of both rhizobia. On the contrary, when submitted to both osmotic stress levels, the decline in all NF related parameters confirms the high contribution of the rhizobial partner to the symbiotic performance under stressed conditions. These results are in agreement with reports that mention importance of bacterial partner contribution in symbiotic effectiveness (Aouani et al. 1997, Mhadhbi et al. 2004, 2008, Tejera et al. 2004). However, the better performance of COCOT-8a3 symbiosis strengthens the importance of examining the interaction between the diversity of native rhizobia with local cultivars (Tajini et al. 2008). This suggests also the involvement of rhizobial strain in nitrogen-fixing capacity, and therefore selection of a suitable rhizobial partner can increase common bean production through improvement of symbiotic nitrogen fixation. As well, our data suggested that the superiority of COCOT-8a3 symbiosis was well established at a high stress level (75 mM mannitol) and that this superiority was not mirrored by higher nodule number under osmotic stress since the data showed significant reduction of this parameter under the high level osmotic stress condition (Table II). Nevertheless, this decline was alleviated by both producing bigger nodules (reaching 2-fold higher than respective controls) and maintaining constant the efficiency in utilization of symbiotic rhizobia (EUSR) even under osmotic stressed conditions. Such behaviour has been widely reported in osmotic stress tolerant symbioses (Saadallah et al. 2001, Aydi et al. 2004, 2008, Sassi et al. 2008a). Indeed, water stress-tolerant N₂ fixation associated with increased individual nodule dry weight was also reported by Serraj and Sinclair (1998) as a consequence of decreased respiration and ureid export resulting in increased carbon concentration in larger nodules as compared with well-watered conditions. Thus large nodules will favour photosynthate and water allocation, maintain favourable nodule relative water content and provide continued supply of water for exporting ureids in nodule xylem (King and Purcell 2001). Actually, we demonstrated by the presented data that even if the studied symbioses behaved similarly under control conditions, they did not have the same performance under osmotic stress. This result was not in accordance with data of Mhadhbi et al. (2009) and Pimratch et al. (2008) who reported that the superiority of a given symbiosis under stressful conditions in terms of high biomass production and nitrogen-fixing capacity was mirrored by its behaviour under non-stressed circumstances.

Rhizobial partner involvement in keeping adequate water status under osmotic stress

Given that the ability of plants to survive severe water deficits depends on their ability to restrict water loss (El Jaafari 2000), the reported work scrutinizes the water status of cv. COCOT as inoculated with either the reference strain CIAT 899 or the local one 8a3 and submitted to increasing levels of osmotic stress induced by mannitol. In accordance with growth and NF parameters, no significant changes were observed between both studied symbioses under control conditions (Table III). Under stressful conditions and mainly under higher osmotic stress level the superiority of the symbiosis COCOT-8a3 was linked essentially to maintaining lower NRWC reductions and constant NWUE at high mannitol concentration in the growing medium (75 mM). This demonstrates that the superiority of COCOT-8a3 symbiosis in terms of water relations is well established at nodule level which could be the origin of the maintenance of better NF capacity reported by this work. Indeed, relationships between maintaining higher NRWC and better tolerance to osmotic stress were previously reported (Sassi et al. 2008b). This could be mainly linked to better ureid export from nodules being easier by adequate nodule water status (Serraj and Sinclair 1998). It should be noted also that the maintaining of lower NRWC reduction in COCOT-8a3 symbiosis was mainly attributed to the accumulation of total soluble sugars notably under a high osmotic stress level (Table 4). Osmotic stress-induced increased soluble sugar in nodules was reported earlier (Fougère et al. 1991). It was generally used for osmotic adjustment (OA). It was also reported that the lowering of the osmotic potential by osmolyte accumulation in response to stress improves the capacity of the cells to maintain physiological processes such as photosynthesis, enzyme activity and cell expansion (Granier et al. 2000, Kiani et al. 2007). However, concerning COCOT-CIAT 899 symbiosis the accumulation of soluble sugar was mirrored by more than a 70% reduction in NRWC which suggests that soluble sugar seems not to have an important role in OA but their accumulation was consistent with the decline in sucrose synthase activity previously reported with this symbiosis (Sassi et al. 2008b). This accumulation presents also a metabolic cost due to synthesis and compartmentation of osmolytes (Bajji et al. 2000), which could impede adequate nodule growth.

In conclusion, this work confirms the relationship between osmotic stress tolerance improvement and inoculation with native soil-isolated R. gallicum 8a3 as compared with inoculation by the reference strain R. tropici CIAT 899. This can be partially explained by better water-use efficiency in both leaves and nodules, better relative water content in nodules and better efficiency in utilization of rhizobial symbiosis. Consequently, the present study recommends the better use of native soil-isolated strains for the inoculation of common bean in order to improve its performance and NF potential under stressful conditions. Nevertheless, further research is needed to explain osmotic stress tolerance in common bean symbiosis via the better understanding of the osmotic stress effect on limiting nodulation through its effects on root-hair colonization and infection by rhizobia.

Acknowledgements

The authors thank Dr Moez Jebara for technical assistance in measurement of acetylene reduction activity (ARA) and Laboratoire d'interaction Legumineuses-microorganismes in Centre de Biotechnologie de Borj Cedria (CBBC) for providing rhizobia. This work was supported by the AQUARHIZ project: ‘Modulation of plant-bacteria interactions to enhance tolerance to water deficit for grain legumes in the Mediterranean dry lands’ FP6 Project INCO-CT-2004-509115, and by the Tunisian Ministry of Higher Education and Scientific Research (LR10CBBC02).