Genetically modified bacteria in agriculture.

Certain bacteria isolated from soils possess properties that allow them to exert beneficial effects on plants either by enhancing crop nutrition or by reducing damages caused by pathogens or pests. Some of them, such as rhizobia, azospirilla, and agrobacteria, have been traditionally released in fields as seed inoculants and they often lead to increases in the yield of different crops while the application of others, such as pseudomonads, often fails to give the expected results. Bacteria genetically modified to be easily traceable and/or to be improved in their expression of beneficial traits have been constructed and released with plants in a number of experimental field plots. With these releases, it has been possible to monitor the modified inoculant bacteria after their introduction in field ecosystems and to assess their impact on the resident microflora. Local environmental factors appeared as playing a crucial role in the survival and persistence of bacteria once released in fields and in the expression of the beneficial traits whether improved or not. The spread of inoculant bacteria from their point of dissemination was limited. Transient shifts in favour of the released bacteria and in disfavour of some members of the bacterial and fungal populations present in the plant rhizosphere might occur with certain released bacteria. The changes observed were, however, less important than those observed under usual agricultural practices. Gene transfer from resident population to introduced bacteria was detected in one case. The transconjugants were found only transiently in the phytosphere of plants but not in soils. No differences between the survival, spread, persistence in field and ecological impacts of genetically modified bacteria and of the corresponding unmodified parent strain could be detected.

An interdisciplinary research strategy to improve symbiotic nitrogen fixation and yield of common bean (Phaseolus vulgaris) in salinised areas of the Mediterranean basin.

The main findings of a cooperative research group of agronomists, plant breeders, microbiologists, physiologists and molecularists to improve the symbiotic nitrogen fixation (SNF) and N2-dependent yield of common bean under moderate salinity in the Mediterranean basin are summarised. Agronomic surveys in reference production areas show large spatial and temporal variations in plant nodulation and growth, and in efficiency of utilisation of the rhizobial symbiosis. The latter was associated with a large rhizobial diversity, including new bean nodulating species. Macrosymbiont diversity in SNF and adaptation to NaCl was found. However, contrasts between plant genotypes could be altered by specific interactions with some native rhizobia. Therefore, variations in soil rhizobial population, in addition to agronomic practices and environmental constraints, may have contributed to erratic results observed in field inoculations. At the mechanistic level, nodule C and N metabolisms, and abcissic acid content, were related to SNF potential and tolerance to NaCl. Their relation with nodule conductance to O2 diffusion was addressed by in situ hybridisation of candidate carbonic anhydrase and aquaporin genes in nodule cortex. The limits and prospects of the cooperative strategy are discussed.

Genetic diversity of rhizobial symbionts isolated from legume species within the genera Astragalus, Oxytropis, and Onobrychis.

The genetic diversity of 44 rhizobial isolates from Astragalus, Oxytropis, and Onobrychis spp. originating from different geographic locations was evaluated by mapped restriction site polymorphism (MRSP) analysis of 16S rRNA genes and by PCR DNA fingerprinting with repetitive sequences (REP-PCR). A comparison of tree topologies of reference strains constructed with data obtained by MRSP and by 16S rRNA gene sequence analyses showed that the topologies were in good agreement, indicating that the MSRP approach results in reasonable estimates of rhizobial phylogeny. The isolates were distributed into 14 distinct 16S rRNA gene types clustering into three major groups which corresponded with three of the genera within the legume symbionts. Most of the isolates were within the genus Mesorhizobium. Five were identified with different genomic species nodulating Lotus spp. and Cicer arietinum. Three Astragalus isolates were classified as Bradyrhizobium, one being similar to Bradyrhizobium elkanii and another being similar to Bradyrhizobium japonicum. Six of the isolates were related to species within the genus Rhizobium. Two were similar to Rhizobium leguminosarum, and the remainder were identified as Rhizobium gallicum. DNA fingerprinting by REP-PCR revealed a high level of diversity within single 16S ribosomal DNA types. The 44 isolates were distributed into 34 REP groups. Rhizobial classification at the genus and probably also the species levels was independent of geographic origin and host plant affinity.

Rhizobium gallicum sp. nov. and Rhizobium giardinii sp. nov., from Phaseolus vulgaris nodules.

Thirty-one strains of two new genomic species (genomic species 1 and 2) of rhizobia isolated from root nodules of Phaseolus vulgaris and originating from various locations in France were compared with reference strains of rhizobia by performing a numerical analysis of 64 phenotypic features. Each genomic species formed a distinct phenon and was separated from the other rhizobial species. A comparison of the complete 16S rRNA gene sequences of a representative of genomic species 1 (strain R602spT) and a representative of genomic species 2 (strain H152T) with the sequences of other rhizobia and related bacteria revealed that each genomic species formed a lineage independent of the lineages formed by the previously recognized species of rhizobia. Genomic species 1 clustered with the species that include the bean-nodulating rhizobia, Rhizobium leguminosarum, Rhizobium etli, and Rhizobium tropici, and branched with unclassified rhizobial strain OK50, which was isolated from root nodules of Pterocarpus klemmei in Japan. Genomic species 2 was distantly related to all other Rhizobium species and related taxa, and the most closely related organisms were Rhizobium galegae and several Agrobacterium species. On the basis of the results of phenotypic and phylogenetic analyses and genotypic data previously published and reviewed in this paper, two new species of the genus Rhizobium, Rhizobium gallicum and Rhizobium giardinii, are proposed for genomic species 1 and 2, respectively. Each species could be divided in two subgroups on the basis of symbiotic characteristics, as shown by phenotypic (host range and nitrogen fixation effectiveness) and genotypic data. For each species, one subgroup had the same symbiotic characteristics as R. leguminosarum biovar phaseoli and R. etli biovar phaseoli. The other subgroup had a species-specific symbiotic phenotype and genotype. Therefore, we propose that each species should be subdivided into two biovars, as follows: R. gallicum biovar gallicum and R. gallicum biovar phaseoli; and R. giardinii biovar giardinii and R. giardinii biovar phaseoli.

Distribution of Symbiotic Genotypes in Rhizobium leguminosarum biovar viciae Populations Isolated Directly from Soils.

The distribution of symbiotic (Sym) plasmid types across background genotypes was investigated in two field populations of Rhizobium leguminosarum biovar viciae isolated directly from soils. PCR-based methods were used to characterize the background genotypes and the Sym gene types. Identical Sym gene types were associated with a variable range of background genotypes, while the same background genotype could harbor distinct Sym gene types. Random distributions of Sym gene types in the background genotypes were observed in the two soil populations. These results suggest that Sym plasmid transfer is less restricted than previously thought on the basis of the analysis of strains isolated from legume nodules.

Diversity of repC plasmid-replication sequences in Rhizobium leguminosarum.

Homologues of the plasmid replicator gene repC were detected and characterized in a sample of Rhizobium leguminosarum strains. Conserved PCR primers were designed from published sequences of repC; they amplified a fragment of about 750 bp from 39 out of 41 strains tested, and also from several Sinorhizobium strains, including S. meliloti. Restriction endonuclease digestion showed that the PCR product from individual strains, though uniform in size, was often heterogeneous in sequence. PCR products from 24 field isolates of R. leguminosarum from France, Germany and the UK were cloned and partially sequenced from both ends. Phylogenies constructed from the 5' and 3' ends (200 bp each) were largely congruent and demonstrated four clearly defined groups plus several unique strains. Published Agrobacterium repC sequences fall within the phylogeny of R. leguminosarum sequences, though not within any of the four groups. Specific pairs of PCR primers were designed for each of the four groups; 29 out of 41 R. leguminosarum strains gave a PCR product of the expected size with more than one group-specific primer pair. We hypothesize that the sequence groups correspond to incompatibility groups of Rhizobium plasmids.

Typing of rhizobia by PCR DNA fingerprinting and PCR-restriction fragment length polymorphism analysis of chromosomal and symbiotic gene regions: application to Rhizobium leguminosarum and its different biovars.

Characterization of 43 strains of Rhizobium leguminosarum biovars viciae, trifolii, and phaseoli was performed by two methodologies based on PCR amplification, i.e., PCR DNA fingerprinting of interrepeat sequences and restriction fragment length polymorphism (RFLP) analysis of PCR -amplified chromosomal and symbiotic gene regions. Groupings generated by PCR DNA fingerprinting with either extragenic palindromic repetitive primers or two different single random primers were correlated with similar levels of resolution. Although less discriminating, PCR-RFLP analysis of intergenic spacer between genes coding for 16S and 23S rRNA (16S and 23S rDNA) yielded intraspecific polymorphisms. The classification of strains was independent of the biovar status and was in agreement with those obtained by PCR DNA fingerprinting. Intrabiovar variation within symbiotic gene regions was detected by PCR-RFLP analysis of nifDK and nodD gene regions, but the strains were grouped according to the biovar. The rDNA intergenic spacer and nif primers were verified to be universal for rhizobial species by testing of various reference strains, whereas the nod primers designed in this study were biovar or species specific for R. leguminosarum and Rhizobium etli. Classifications of R. leguminosarum strains by the PCR-based methods were correlated with those previously obtained by conventional total DNA restriction profile comparisons and RFLP analysis using chromosomal and symbiotic gene probes. Ranges of discriminating powers were also equivalent between the two approaches. However, the PCR-based methods are much less time-consuming and are therefore more convenient.

Characterization, distribution, and localization of ISRl2, and insertion sequence element isolated from Rhizobium leguminosarum bv. viciae.

An insertion sequence (IS) element, ISR12, from Rhizobium leguminosarum bv. viciae strain MSDJ4184 was isolated by insertional inactivation of the sacRB gene of pSUP104-sac, which allows positive selection. ISRl2 is 932 bp long, is flanked by 17-bp imperfect terminal inverted repeats, and generated a 3-bp target site duplication. ISRl2 was found to be 63 to 77% homologous to insertion elements of the IS5 group of the IS4 superfamily. A probe incorporating a full-length copy of ISRl2 was used to screen genomic DNAs from a collection of strains and from two field populations of R. leguminosarum to detect and estimate the copy numbers of homologous sequences. Among the collection of 63 strains representing the different species and genera of members of the family Rhizobiaceae, homology to ISRl2 was found within strains belonging to Sinorhizobium meliloti and S. fredii; within four of the six recognized Rhizobium species. R. leguminosarum, R. tropici, R. etli, and R. galegae; and within Rhizobium sp. (Phaseolus) genomic species 2. The apparent copy numbers of ISRl2 varied from one to eight. Among 139 isolates of R. leguminosarum from two field populations, homology to ISRl2 was detected in 91% of the isolates from one site and in 17% from the other. Analysis of the 95 isolates that hybridize to ISRl2 revealed a total of 20 distinct hybridization patterns composed of one to three bands. Probing blots of Eckhardt gels showed that sequences with homology to ISRl2 may be found on plasmids or the chromosome. Analysis of their genomic distribution demonstrated relationships and diversity among the R. leguminosarum isolates tested.

Rapid Identification of Rhizobia by Restriction Fragment Length Polymorphism Analysis of PCR-Amplified 16S rRNA Genes.

Forty-eight strains representing the eight recognized Rhizobium species, two new Phaseolus bean Rhizobium genomic species, Bradyrhizobium spp., Agrobacterium spp., and unclassified rhizobia from various host plants were examined by restriction fragment length polymorphism (RFLP) analysis of 16S rRNA genes amplified by polymerase chain reaction (PCR). Twenty-one composite genotypes were obtained from the combined data of the RFLP analysis with nine endonucleases. Species assignments were in full agreement with the established taxonomic classification. Estimation from these data of genetic relationships between and within genera and species correlated well with previously published data based on DNA-rRNA hybridizations and sequence analysis of 16S rRNA genes. This PCR-RFLP method provides a rapid tool for the identification of root nodule isolates and the detection of new taxa.

Genomic heterogeneity among French Rhizobium strains isolated from Phaseolus vulgaris L.

Levels of DNA relatedness between strains isolated from root nodules of Phaseolus vulgaris and reference strains of different Rhizobium species were determined by performing DNA-DNA hybridization experiments (S1 nuclease method). The nine strains examined were members of three genomic groups previously delineated by a restriction fragment length polymorphism analysis among strains isolated from P. vulgaris at different sites in France. In agreement with the results of the restriction fragment length polymorphism analysis, three genomic species were found. We confirmed that one of these species corresponded to Rhizobium leguminosarum since the strain examined was 100% related to the type strain of this species. The other two species were new genomic species which were less than 21% related to reference strains belonging to other Rhizobium species, including Rhizobium etli and Rhizobium tropici, and were 18% related to each other. As determined by an analysis of partial 16S ribosomal DNA sequences, each of the genomic species was found to belong to a lineage independent from the lineages of previously described Rhizobium species. Nevertheless, they were included in the group formed by the fast-growing Rhizobium species. Both genomic species 1 and genomic species 2 contained a majority of strains which were capable of nodulating both P. vulgaris and Leucaena leucocephala, like R. tropici. However, they also contained strains with a nodulation phenotype restricted to P. vulgaris, like R. leguminosarum bv. phaseoli and R. etli bv. phaseoli. Our data are the first evidence that in Europe species other than R. leguminosarum nodulate P. vulgaris.

Quantitative study of nodulation competitiveness in Rhizobium strains.

We compared the nodulation competitiveness of three strains of Rhizobium leguminosarum by counting the number of nodules formed on faba bean plants after the application at sowing time of different concentrations of the strains to soils already containing Rhizobium strains of the same species. A relationship of type y = ax was found to exist between the ratio of the nodules formed by the applied inoculum strain to the nodules formed by the soil strains and the ratio of Rhizobium cells in the inoculum to the cells in the soil. This relationship was also confirmed in another competition experiment in which two R. meliloti strains of identical competitiveness were mixed in various proportions. The relationship can also be applied to the majority of results reported in the literature. Should it prove to be more widely applicable, it could be used to estimate the relative competitiveness of Rhizobium strains and thus predict the performance of an inoculum in a given soil.

N abundance of nodules as an indicator of N metabolism in n(2)-fixing plants.

This paper expands upon previous reports of (15)N elevation in nodules (compared to other tissues) of N(2)-fixing plants. N(2)-Fixing nodules of Glycine max (soybeans), Vigna unguiculata (cowpea), Phaseolus vulgaris (common bean), Phaseolus coccineus (scarlet runner bean), Prosopis glandulosa (mesquite), and Olneya tesota (desert ironwood) were enriched in (15)N. Nodules of Vicia faba (fava beans), Arachis hypogaea (peanut), Trifolium pratense (red clover), Pisum sativum (pea), Lathyrus sativus (grass pea), Medicago sativa (alfalfa), and Lupinus mutabilis (South American lupine) were not; nor were the nodules of nine species of N(2)-fixing nonlegumes. The nitrogen of ineffective nodules of soybeans and cowpeas was not enriched in (15)N. Thus, (15)N elevation in nodules of these plants depends on active N(2)-fixation. Results obtained so far on the generality of (15)N enrichment in N(2)-fixing nodules suggest that only the nodules of plants which actively fix N(2) and which transport allantoin or allantoic acid exhibit (15)N enrichment.

Nitrogen Isotope Fractionation Associated with Nitrate Reductase Activity and Uptake of NO(3) by Pearl Millet.

Nitrogen isotope fractionation by Pearl Millet (Pennisetum americanum L. and P. mollissimum L.) grown on nitrate was associated with nitrate reductase activity. Fractionation was evidenced at the step of nitrate reduction when the substrate-to-enzyme ratio was high (possibly saturating for the active sites of the nitrate reductase enzyme), for instance in young seedlings having a low nitrate reductase activity or in seedlings grown on high nitrate concentration.When the substrate concentration was low (and, hence, the active sites of the enzyme were possibly not saturated), the isotopic discrimination could only be associated with the uptake of nitrate into the cell. In that case, isotopic fractionation was null. It is concluded that the uptake of nitrate does not discriminate among nitrogen isotopes.

[Symbiotic efficiency in spontaneous mutants of Rhizobium legumino-sarum resistant to streptomyocin, spectinomycin, or kanamycin]. / Efficience symbiotique de mutants spontanés de Rhizobium leguminosarum résistants á la streptomycine, spectinomycine ou kanamycine

Symbiotic effectiveness of 45 mutant strains selected from four wild effective strains of Rhizobium leguminosarum for resistance to streptomycin, spectinomycin or kanamycin was determined on Vicia faba. Loss of effectiveness occurred in twenty of these mutants; distribution of ineffective mutants was uniform among the three types of antibiotic resistant mutants but varied with the parent strain from which mutants have been derived.

Nitrite formation from hydroxylamine and oximes by Pseudomonas aeruginosa.

Nitrite was formed from hydroxylamine and several oximes by intact cells and extracts of Pseudomonas aeruginosa. The activity was induced by the presence of oximes in the culture medium. Nitroalkanes were not intermediates in the conversion of acetaldoxime, acetone oxime, or butanone oxime to nitrite, since nitromethane inhibited the formation of nitrite from the nitro compounds but not from the corresponding oximes. The oxime apparently functions as a constant source of hydroxylamine during growth of the bacterium. Hydroxylamine at low concentration was converted stoichiometrically to nitrite by extracts of the bacterium; high concentrations were inhibitory. Nicotinamide adenine dinucleotide phosphate, oxygen, and other unidentified cofactors were necessary for the reaction. Actively nitrifying extracts possessed no hydroxylamine-cytochrome c reductase activity. Hyponitrite, nitrous oxide, and nitric oxide were not metabolized.