Symbiotic Hydrogenase Activity in Bradyrhizobium sp. (Vigna) Increases Nitrogen Content in Vigna unguiculata Plants

Bacteria belonging to the group collectively known as rhizobia form nitrogen-fixing nodules in symbiosis with leguminous plants. In these symbioses, a large amount of hydrogen is released from nodules as a by-product of the nitrogenase reaction. Certain rhizobial strains are able to oxidize this hydrogen through the expression of uptake hydrogenases. Hydrogen recycling has been shown to reduce energy losses associated with diazotrophy (19). Consequently, incorporation of the hydrogen oxidation capability into rhizobial strains has been proposed as a way to improve symbiotic nitrogen fixation (11). However, enhancement of legume host productivity associated to the activity of the hydrogenase enzyme has been reported only for certain symbiotic associations, such as Bradyrhizobium japonicum/soybean (1, 8).

To date, a detailed molecular characterization of hydrogenase gene clusters has been reported for Rhizobium leguminosarum bv. viciae, B. japonicum, and Azorhizobium caulinodans (3, 11, 16). Hydrogenase gene clusters share a common core of 18 genes, named hup, hyp, and hox genes, that encode functions involved in enzyme biosynthesis and regulation. In certain rhizobia, hydrogenase activity is induced in microaerobic free-living cells, in addition to symbiotic conditions (17, 21). In previous work, we analyzed the presence of hup, hyp, and hox genes in several Bradyrhizobium sp. (Lupinus) and Bradyrhizobium sp. (Vigna) strains (2). Significant differences in the presence of hup regulatory genes were observed by Southern hybridization assays. In addition, phylogenetic analysis of partial hupS and hupL sequences revealed that Bradyrhizobium sp. (Vigna) hup sequences cluster apart from those of Bradyrhizobium sp. (Lupinus). In the light of these differences within the Bradyrhizobium genus, we have characterized the symbiotic hydrogenase activity of the previously studied Bradyrhizobium sp. (Lupinus) and Bradyrhizobium sp. (Vigna) strains and the relevance of their hydrogenase systems for plant productivity.

The hydrogen oxidation capabilities of several Bradyrhizobium sp. (Lupinus) and Bradyrhizobium sp. (Vigna) strains were analyzed in symbiosis with Lupinus albus cv. Unicrop and Vigna unguiculata cv. Blackeye, respectively. Plants inoculated with bacterial cultures were grown under bacteriologically controlled conditions as previously described (9), by using a nitrogen-free plant nutrient solution supplemented with 20 μM NiCl₂ in order to optimize hydrogenase activity (6). Hydrogen metabolism was characterized by measuring hydrogen evolution in intact nodules (5) and hydrogenase activity in bacteroid suspensions (15). In this analysis, we found that the Bradyrhizobium sp. (Lupinus) strains displayed a wide range of hydrogen uptake activity levels (Table 1). Strains 624 and Z89 showed the lowest and highest levels of hydrogenase activity, respectively, whereas strains UPM860, 466, and IM43B exhibited intermediate values. None of them was able to fully recycle the hydrogen produced by nitrogenase, since nodules evolved hydrogen in all cases (Table 1). Since a significant plant host effect on hydrogenase activity has been previously shown (12), assays were also carried out using Ornithopus compressus as a plant host. However, similar values of bacteroid hydrogenase activity were obtained in this legume host (data not shown). For Bradyrhizobium sp. (Vigna) strains, different levels of symbiotic hydrogenase activity were also observed, but in this case, a large proportion of strains recycled all the hydrogen generated in the nitrogen fixation process (Table 1). Hydrogenase activities from Bradyrhizobium sp. (Lupinus) and Bradyrhizobium sp. (Vigna) strains were similar to those previously obtained without nickel supplementation (12, 14). Due to the higher levels of hydrogenase activity exhibited by Bradyrhizobium sp. (Lupinus) Z89 and Bradyrhizobium sp. (Vigna) M5, we selected these two strains for further functional characterization of the hydrogen uptake system.

To generate the Bradyrhizobium sp. (Lupinus) Z89.1 and Bradyrhizobium sp. (Vigna) M5.1 hupSL mutant strains, 1.5-kb DNA fragments containing both the hupS 3′ end and the hupL 5′ end from the Z89 and M5 wild-type strains were obtained by PCR amplification by using the degenerate primers hupSL1 and hupSL2 and PCR conditions previously described (2). Each PCR product was cloned in the PCR2.1-TOPO vector (Invitrogen BV, Groningen, The Netherlands), excised with KpnI-XbaI, and introduced into the pSS10 vector (3). The resulting plasmids pSCB2 and pSCB3 were introduced by conjugation into strains Z89 and M5, respectively, and transconjugants were selected in Rhizobium minimal medium (13) plates supplemented with spectinomycin. Genomic DNA of transconjugant strains was analyzed by Southern hybridization experiments (18) using a Rhizobium leguminosarum hupS gene probe in order to confirm the single-crossover event leading to the plasmid integration in the corresponding hupSL region (data not shown).

The effect of hupSL gene disruption in Bradyrhizobium sp. (Lupinus) Z89.1 and Bradyrhizobium sp. (Vigna) M5.1 mutants was assessed by measuring hydrogenase activity in symbiosis with L. albus and V. unguiculata plants, respectively (Table 2). Bacteroids from both mutant strains exhibited basal levels of hydrogenase activity compared with those of the corresponding wild-type strains. In agreement with this, nodules from L. albus and V. unguiculata induced by these mutants evolved large amounts of hydrogen. These results confirmed inactivation of the hydrogenase system and the absence of other functional hup systems in these strains.

Disruption of hupSL genes induced a differential effect on the levels of nitrogenase activity of the Bradyrhizobium sp. (Vigna) M5.1 and Bradyrhizobium sp. (Lupinus) Z89.1 hup mutants (Table 2). Acetylene reduction activity in Bradyrhizobium sp. (Vigna) M5.1 was significantly reduced compared with that of the wild-type strain, whereas no difference between values from Bradyrhizobium sp. (Lupinus) Z89 and mutant strain Z89.1 was observed. These results indicate that hydrogenase activity might be necessary for optimal levels of nitrogenase activity in Bradyrhizobium sp. (Vigna) bacteroids.

To test whether the observed reduction of nitrogenase activity in Bradyrhizobium sp. (Vigna) M5.1 bacteroids could affect legume productivity, we analyzed the effect of hydrogenase inactivation on nitrogen accumulation and dry weight of the corresponding plant host. V. unguiculata plants inoculated with the Bradyrhizobium sp. (Vigna) M5.1 mutant showed levels of nitrogen accumulation lower than those inoculated with the wild-type strain (Table 3). This result was consistently observed in five independent experiments carried out with 30- and 40-day-old plants grown under controlled chamber conditions. The reduction in nitrogen content was also correlated with a reduction of plant dry weight in one experiment at 30 days (Table 3). Similar experiments were carried out in the Bradyrhizobium sp. (Lupinus)-L. albus symbiosis. In this case, no significant difference in nitrogen content or dry weight was observed between 40-day-old L. albus plants inoculated with the Bradyrhizobium sp. (Lupinus) Z89.1 hup mutant or the wild-type strain Z89 (Table 4).

Our results show that inactivation of hydrogenase genes impairs bacterial nitrogenase activity in Bradyrhizobium sp. (Vigna) and has an effect on total nitrogen content of the symbiotic plant partner, V. unguiculata. Several mechanisms by which hydrogenase activity might increase the efficiency of the nitrogen fixation process have been proposed (7). Among them, prevention of H₂ inhibition of the nitrogenase reaction, protection of nitrogenase from the O₂ damage, and provision of an additional source of energy appear as critical factors that might influence nitrogenase activity. At this point of the investigation, it is difficult to determine whether one of these factors or a combination of them is responsible for the beneficial effect of hydrogenase activity on the nitrogen fixation process in this specific symbiosis. Such analysis requires further knowledge of the system, and we are now pursuing the molecular characterization of the Bradyrhizobium sp. (Vigna) hup cluster. These data will hopefully shed some light on the role of hydrogen recycling in this symbiosis, and they also might provide some clues to explain why a very high proportion of Bradyrhizobium sp. (Vigna) isolates from Vigna unguiculata nodules are hydrogenase positive (4, 14).

The use of mutant strains carrying antibiotic resistance genes is somewhat controversial (10), and it can be argued that the resistance markers might unbalance bacterial fitness and affect the nitrogen fixation process. In this work, however, the Bradyrhizobium sp. (Lupinus) Z89.1 hup mutant carries the same resistance gene as Bradyrhizobium sp. (Vigna) M5.1, and no significant differences either in acetylene reduction assays or in the nitrogen content of Lupinus plants were associated with the hup mutant in the same experimental conditions. Similarly, no effect on plant productivity was detected in the A. caulinodans-Sesbania rostrata symbiosis using a hup mutant strain carrying this antibiotic resistance cassette (3). Future work will be focused on the construction and testing of a hup deletion mutant in plant productivity experiments to confirm these data and definitely establish the contribution of the hydrogenase system to the productivity of the Bradyrhizobium sp. (Vigna)-V. unguiculata association.

• Copyright © 2005 American Society for Microbiology