The common bean genotypes used here were the wild accessions Phaseolus filiformis P1535309, P. acutifolius G40142 and P. vulgaris P1325687, and the commercial cultivars P vulgaris cv. Pinto Villa and P. vulgaris cv. Flor de Mayo. The seeds were donated by Jorge Acosta Gallegos from INIFAP-Celaya.

Glomus intraradices spores were donated by Javier Villegas from IIQB-UMSNH and maintained at −70°C in a 50% glycerol solution. The Rhizobium strains used (Fcin1, Fcin2, Fcin3, Fcin4, Fcin5, Fcin6, and Fcin7) were isolated in a previous experiment [9]. Stock cultures were maintained at −70°C in PY medium (in g L⁻¹: yeast extract 3, CaCl₂ • 2H₂O 1, casein peptone 5), containing 50% (v/v) glycerol.

Seeds were surface disinfected with 2.5 M NaOCl for 5 min and rinsed with sterile distilled water before use. Three inoculation treatments were imposed as follows: a) Rhizobium inoculation was performed by soaking the seeds for 1 h in a mixture of native rhizobial strains prepared as described by Altamirano-Hernandez et al. [9]; b) Glomus inoculation was performed during sowing by adding to each seed 1 mL of a Glomus spore suspension (30 spores mL^-1); c) co-inoculation was performed by first inoculating Rhizobium as described previously, then 1 mL of a Glomus spore suspension was added to each seed during sowing. Control plants were not inoculated, instead they were watered with a Jensen solution (in g L^-1: 1 Ca(H₂PO₄)₂.H₂O, 0.2 K₂HPO₄, 0.2 MgSO₄.7H₂O, 0.2 NaCl, 0.1 FeCl₂.4H₂O, and 1 mL of mineral solution [2.86 boric acid, 0.22 ZnSO₄.7H₂O, 1,81 MnCl₂.7H₂O, 0.09 K₂MoO₄]) containing 0.05% KNO₃.

Plants from different inoculation treatments (previously described) were watered every three days, and once a week a Jensen solution was used. Drought stress was imposed on plants at the flowering stage (R7) by suspending watering for 16 days. At the end of this period, plant total biomass and leaf area were recorded. From the Rhizobium and co-inoculation treatments, nodules were harvested by hand and nodule number and dry weight were measured.

Root samples were collected from co-inoculated and Glomus-inoculated plants to measure mycorrhizal colonization.

Nodule trehalose was extracted by grinding the nodules in boiling 80% (v/v) ethanol and evaporated to dryness before resuspension in 1 mL of 80% (v/v) ethanol. Aliquots of 100 μL from each sample were derivatized to obtain the aldonitrile peracetylated forms of the sugars, 25 µg of perseitol were added as internal standard, and then analyzed by gas chromatography–mass spectrometry (HP6850-HP5973) on an HP-5 MS column (Agilent Technologies, Palo Alto, CA) as described by Macias-Rodriguez et al. [21].

In general terms, it was observed that co-inoculation of Glomus and Rhizobium in bean plants had a negative effect on nodulation parameters (Figs. 1 and 2). In this sense, the wild accession of P. vulgaris P1325687 showed the highest number of nodules under both watering regimes and inoculation treatments. In the well-watered treatment (Fig. 1a), P. vulgaris P1325687 presented on average 500 nodules per plant when inoculated only with the rhizobial mixture, but when co-inoculated with Glomus, the number of nodules in this bean genotype significantly decreased (p<0.05) to an average of 300 per plant. The commercial P. vulgaris genotypes (Flor de Mayo and Pinto Saltillo) inoculated only with Rhizobium presented 65 nodules per plant on average, but when co-inoculated with Glomus, the number of nodules decreased to 45. P. filiformis and P. acutifolius showed a low number of nodules, with an average of 12 nodules per plant; under co-inoculation they showed an average of 7 nodules per plant.

Fig. (1). Number of nodules per plant under A) well-watered treatment and B) drought stress treatment. White bars represent nodules from plants inoculated only with Rhizobium; black bars represent nodules from plants co-inoculated with Rhizobium and Glomus.

Fig. (2). Nodule dry weight under A) well-watered treatment and B) drought stress treatment. White bars represent nodules from plants inoculated only with Rhizobium; black bars represent nodules from plants co-inoculated with Rhizobium and Glomus.

Fig. (3). Percentage of Glomus colonization under: A) well-watered treatment and B) drought stress treatment. White bars represent nodules from plants inoculated only with Glomus; black bars represent nodules from plants co-inoculated with Rhizobium and Glomus.

Fig. (4). Correlation indexes between mycorrhiza colonization and nodule parameters; A) well-watered treatment and B) drought stress treatment. Dashed lines represent confidence interval at p<0.05.

Fig. (5). Percentage of biomass reduction caused by drought in comparison to well-watered plants among the different inoculation treatments. N: Plants inoculated with rhizobia. N+M: Plants co-inoculated with Glomus and rhizobia. M: Plants inoculated with Glomus. C: Non-inoculated control plants.

When drought was imposed on these bean plants, the number of nodules decreased in the two inoculation treatments when compared with their well-watered counterparts (Fig. 1b). P. vulgaris P1325687 had 350 nodules per plant when inoculated with Rhizobium, and only 100 when co-inoculated with Glomus. Under watering conditions, there was a significant difference among the two commercial bean cultivars. P. vulgaris cv. Flor the Mayo presented 45 nodules when inoculated with Rhizobium and 35 when the mycorrhiza was co-inoculated. Interestingly, P. vulgaris cv. Pinto Saltillo had 10 nodules when inoculated only with Rhizobium and 40 when the mycorrhiza was co-inoculated. This was the only case recorded in this work in which the number of nodules increased with Glomus inoculation. In contrast, P. filiformis and P. acutifolius had 10 nodules with rhizobial inoculation and 4 with co-inoculation under this drought treatment.

Variability in the dry weight of nodules among the different bean genotypes was lower as compared to that recorded for the total number of nodules (Figs. 1 and 2). Under well-watered conditions (Fig. 2a), when plants were inoculated with the rhizobial mixture, P. acutifolius showed the largest nodules, with an average dry weight of 0.55 mg nodule^-1. P. vulgaris P1325687 nodules weighed 0.43 mg on average. The last three genotypes had on average nodules of 0.21 mg. Under the co-inoculation treatment, the weight of nodules significantly (p<0.05) decreased in P. vulgaris P1325687 (0.35 mg nodule^-1), P. acutifolius (0.23 mg nodule^-1) and P. vulgaris cv. Pinto Saltillo (0.15 mg nodule^-1).

Nodule weight decreased under drought condition (Fig.2b). As in well-watered plants, P. acutifolius had the largest nodules when inoculated only with Rhizobium, with an average dry weight of 0.38 mg nodule^-1, followed by P. vulgaris P1325687 with 0.30 mg nodule^-1. The smallest nodules were reordered in P. vulgaris cv. Pinto Saltillo with an average of 0.10 mg nodule^-1. When these bean genotypes were co-inoculated with Glomus, nodule weight decreased significantly (p<0.05) in all genotypes, with a maximum value of 0.19 mg nodule^-1 in P. vulgaris P1325687, and a minimum of 0.1 mg nodule^-1 in P. acutifolius. In contrast, P. vulgaris cv. Pinto Saltillo nodules showed significant (p<0.05) increases, with 0.16 mg nodule^-1.

The effect of mycorrhizal co-inoculation on nodule trehalose content is shown in Table 1. Mycorrhizal co-inoculation caused a decrease in trehalose content of nodules. Under well-watered condition, in plants inoculated only with Rhizobium, the average trehalose content in the nodules was 36 mg g^-1, the highest value corresponded to P. filiformis (50 mg g^-1), and the lowest to P. vulgaris cv Flor de Mayo (20 mg g^-1). In the co-inoculation treatment, the average nodule trehalose content of these five genotypes significantly (p<0.05) decreased to 16 mg g^-1, the maximum value of 25 mg g^-1 corresponding to P. filiformis and the minimum of 10 mg g^-1 to P. acutifolius and P. vulgaris cv. Flor de Mayo.

Under drought stress, trehalose content of nodules significantly (p<0.05) increased in both inoculation treatments in comparison to their well-watered counterparts. The average nodule trehalose content for the five genotypes with Rhizobium inoculation in this water regime was of 67 mg g^-1. The highest trehalose content was recorded in nodules of P. vulgaris P1325687 (90 mg g^-1), and the lowest in P. vulgaris cv. Flor de Mayo (45 mg g^-1). In the co-inoculation treatment, the trehalose content of nodules significantly (p<0.05) decreases in all genotypes, with an average value of 31 mg g^-1. The maximum trehalose value corresponded to P. filiformis (40 mg g^-1) and the minimum to P. vulgaris cv. Flor de Mayo (18 mg g^-1).

To better understand the negative effect of Glomus inoculation on the nodule parameters described previously, mycorrhizal colonization was recorded and the results are shown in Fig. (3). Interestingly, mycorrhizal colonization was higher when the plants were only inoculated with Glomus in all genotypes evaluated, and decreased when the co-inoculation Glomus-Rhizobium was performed. Under well-watered conditions, mycorrhizal colonization among the five evaluated bean genotypes was quite uniform, with an average colonization of 22%. The maximum value was recorded for P. vulgaris cv. Pinto Saltillo (27%) and the minimum for P. vulgaris cv. Flor de Mayo (15%). When co-inoculated with Rhizobium, mycorrhizal colonization decreased to an average of 15%. The highest value was 18% for P. filiformis and P. vulgaris P1325687, and the lowest value was 10% for P. vulgaris cv. Flor de Mayo.

Under drought stress it was observed that, as opposed to nodulation, mycorrhizal colonization increased, with an average value of 37%. P. filiformis showed the highest colonization with 50%, while P. vulgaris cv. Flor de Mayo showed the lowest with 10%. In the co-inoculation treatment, Glomus colonization significantly (p<0.05) decreased, as occurred under well-watered conditions, with an average of 21%, the maximum value being for P. filiformis (35%) and the minimum for P. vulgaris cv. Flor de Mayo (5%).

When a correlation analysis was performed to evaluate the effect of mycorrhiza on nodule parameters, it was interesting to observe that under well watered conditions, comparing data from the five bean genotypes, the correlation indices showed a positive relationship between Glomus colonization and nodulation (although it wasn’t statistically significant). This would indicate that bean genotypes that have the highest mycorrhizal colonization showed the highest number of nodules, nodule dry weight and nodule trehalose content (Fig. 4a), regardless of the fact that individually, in each genotype mycorrhizae decreased the values of these parameters. The correlation index values for Glomus colonization were r=0.38 for nodule number, r=0.63 for nodule dry weight, and r= 0.48 for trehalose content. However, under drought conditions, some changes were observed in these correlations indices (Fig. 4b). The number of nodules had a negative correlation with mycorrhization (r=0.15). Even though the correlation index observed with nodule dry weight decreased, a positive correlation was still observed (r=0.26). Interestingly, trehalose correlation index increased to a significant value of 0.98.

Plant leaf area with the different inoculation treatments under both water regimes is shown in Table 2. Under well-watered conditions, the response of the five bean genotypes to the different inoculation treatments was very heterogeneous. Nevertheless, control plants that where chemically fertilized showed the greatest leaf area (149 cm²), followed by the plants co-inoculated with Rhizobium and Glomus (138 cm²). Plants inoculated with Rhizobium had an average leaf area of 118 cm², and the minimum recorded value corresponded to plants inoculated with Glomus (103 cm²). Also, considering only the inoculated treatments, it was observed that in P. filiformis and P. acutifolius plants, co-inoculation resulted in a negative effect by reducing the leaf area when compared to the maximum area obtained under inoculation, which corresponded to Rhizobium inoculation. On the other hand, co-inoculation had a positive effect on P. vulgaris P1325687 and P. vulgaris cv. Flor de Mayo genotypes, which showed the greatest leaf area under this treatment in comparison to single inoculations.

Under drought stress, bean plants inoculated with Glomus presented the greatest leaf area (96 cm²), followed by plants inoculated with Rhizobium (90 cm²). Leaf area in fertilized control plants decreased to 85 cm², and the smallest leaf area was recorded in co-inoculated plants (84 cm²). Again, considering only inoculation treatments, it was observed that in P. acutifolius and P. vulgaris P1325687, the maximum leaf area was obtained under Glomus inoculation, and co-inoculation negatively affected this parameter. On the other hand, P. acutifolius and P. vulgaris cv. Flor de Mayo showed greater leaf area with Rhizobium inoculation, being negatively affected by co-inoculation. P. vulgaris cv. Pinto Saltillo represented the only case in which a positive effect of co-inoculation was observed, with maximum leaf area.

The effect of inoculation on total biomass is shown in Table 3. In this case, the Rhizobium inoculation produced the highest biomass of all five genotypes tested, both under well-watered and drought stressed plants. The total biomass for Rhizobium inoculated plants under well-watered conditions was 2 mg plant^-1, followed by plants inoculated with Glomus, with a biomass of 1.5 mg plant^-1, and of the inoculated plants, those that were co-inoculated showed the lower biomass (1.4 mg plant^-1). Plants under drought stress inoculated with Rhizobium showed a biomass of 1.4 mg plant^-1; plants inoculated with Glomus showed a biomass of 1 mg plant^-1, and those co-inoculated showed 0.7 mg plant^-1.

Considering the plant biomass obtained under well-watered conditions as 100% of growth for the different inoculation treatments, inhibition caused by drought was recorded and the results are shown in Fig. (5). Although there was no statistical significance (p<0.05) among inoculation treatments, the bean plants less affected by drought were the chemically fertilized ones with an average of 30% biomass reduction when compared to their well-water counterpart.

Among the inoculated treatments, plants inoculated with Rhizobium were the least affected by drought (33% biomass reduction), followed by plants inoculated with Glomus (37% biomass reduction), and the plants most sensitive to drought were the ones co-inoculated, with an average biomass reduction of 44%.