Burkholderia phymatum Strains Capable of Nodulating Phaseolus vulgaris Are Present in Moroccan Soils

Abstract

Phylogenetic analysis of 16S rRNA, nodC, and nifH genes of four bacterial strains isolated from root nodules of Phaseolus vulgaris grown in Morocco soils were identified as Burkholderia phymatum. All four strains formed N₂-fixing nodules on P. vulgaris and Mimosa, Acacia, and Prosopis species and reduced acetylene to ethylene when cultured ex planta.

Until 2001 all known bacteria involved in root nodule symbioses with leguminous plants were classified as members of the order Rhizobiales of the Alphaproteobacteria, including Azorhizobium, Bradyrhizobium, Mesorhizobium, Rhizobium, and Sinorhizobium (28, 36, 38). Moulin et al. (21), however, described two Burkholderia nodule-forming strains isolated from Machaerium lunatum in French Guiana and Aspalathus carnosa in South Africa, respectively, this being the first report on the presence of a betaproteobacterium within root nodules of legumes. Later, the strains were formally classified as Burkholderia phymatum STM815^T and Burkholderia tuberum STM678^T, respectively (33). Burkholderia species are the predominant isolates from nodules of mimosoid legumes from Panama (2), Costa Rica (3), Taiwan (4, 6), Brazil (5, 7), Venezuela (5), and Madagascar (24), which indicates a high affinity of Burkholderia for forming effective symbioses with Mimosa. Diazotrophy is well represented in Burkholderia; among the more than 55 species presently classified as Burkholderia, 9 have been shown to fix N₂ ex planta by using either the acetylene reduction activity (ARA) assay or the presence of nifH genes encoding nitrogenase reductase (3, 5, 11, 24) and more recently by ¹⁵N₂ isotopic dilution experiments (17).

Common bean (Phaseolus vulgaris) is an herbaceous leguminous plant which establishes N₂-fixing symbiosis with at least 5 species of the genus Rhizobium. Rhizobium etli is the predominant species in America (29) and is also detected in Europe and Africa (13, 20). Rhizobium leguminosarum bv. phaseoli is commonly found in Europe (13) and has also been reported to be present in Tunisia (20) and Colombia (10). Rhizobium tropici is found in acid soils of South America and is also present in Europe and several African countries (18). Rhizobium giardinii has been detected only in European and Tunisian soils (1, 20), and Rhizobium gallicum has been found nodulating beans in Europe, North Africa, and Mexico (1, 13). In this study we report on the isolation and characterization of B. phymatum from root nodules of P. vulgaris grown in alkaline soils from Morocco. Our results show that strains formed effective nodules on species of Mimosa, Acacia, and Prosopis and fixed atmospheric N₂ under free-living conditions.

Soil was taken from a field near Oulade Mansour (34°47′N, 2°15′W, Oujda province, Morocco) where maize and common bean have traditionally been grown as rotational crops without N fertilization. Soil had a sandy-clay texture and the following characteristics: pH (in water), 8.1; 55.18% sand; 17.17% silt; 27.65% clay; 6.1% carbonates; 7.69% organic carbon; 0.069 total nitrogen. Seeds of P. vulgaris cv. Flamingo were surface sterilized (96% ethanol for 30 s followed by immersion in 15% [vol/vol] H₂O₂ for 8 min), washed several times with sterile water, germinated in the dark, and planted in 1-kg pots containing soil and sterile sand (1:1, vol/vol). Plants were grown for 30 days in controlled environmental chambers under conditions previously described (9). Nodules were collected, pooled together, surface sterilized with 2.5% HgCl₂ for 5 min, and rinsed thoroughly with sterile distilled water. Then, 12 nodules were placed independently on petri dishes and crushed in a drop of sterile water with a sterile glass rod. The resulting suspension was streaked onto petri dishes containing either yeast extract-mannitol (YEM) medium (35) or peptone-mineral salts-yeast extract (PSY) medium (25). After incubation of the plates at 30°C for 7 days, CFU which represented all of the colony types that could be distinguished by microscopic observation of living cells were chosen. After identification, Burkholderia strains were routinely grown in BAc medium (12).

For DNA extraction and PCR amplifications, genomic DNA was isolated from bacterial cells using the RealPure genomic DNA extraction kit (Durviz, Spain) according to the manufacturer's instructions. Repetitive extragenic palindromic (REP) fingerprinting was performed using primers REPIR-I and REP2-I according to the method of de Bruijn (8). PCR amplifications of the 16S rRNA gene fragments were done with the Bphym-F and Bphym-R species-specific primer pair (37).

After isolation from root nodules, 52 strains forming morphologically different colonies were obtained and grouped in two main clusters after REP-PCR fingerprinting (data not shown), a technique that is extensively used to group bacteria at subspecies or strain level (8, 34) and has proven to be a powerful tool for studies of microbial ecology and evolution (14). The nearly complete sequence of the 16S rRNA gene from a representative strain of each REP-PCR group was obtained and compared with those held in GenBank. Forty-six strains in cluster I were members of the family Rhizobiaceae from the Alphaproteobacteria. Another 4 strains, here referred to as GR strains (GR01 to GR04), grouped in cluster II and were classified into the family Burkholderiaceae within the Betaproteobacteria. The remaining two strains have not been clearly classified as yet. The four GR strains have almost identical 16 rRNA gene sequences, and BLAST searches showed that they were phylogenetically close (99% identity) to B. phymatum STM815^T, a strain originally isolated from the papilionoid legume Machaerium lunatum (21, 33). A phylogenetic analysis including 30 Burkholderia reference strains showed that strains from root nodules of P. vulgaris form a tight cluster with B. phymatum STM815^T (Fig. 1).

FIG. 1.

Phylogenetic tree showing the positions of four P. vulgaris-isolated strains, GR01, GR03, GR05, and GR06, within the genus Burkholderia based on 16S rRNA gene sequence comparisons. One thousand bootstrap samplings were performed. The NCBI GenBank accession number for each strain is shown in parentheses. The bar represents one nucleotide substitution per 1,000 nucleotides. The multiple alignments of the sequences were performed with CLUSTAL W software (30). The tree topology was inferred by the neighbor-joining method (27), based on 1,310 DNA sites, and the distance matrix method was performed according to the method of Jukes and Cantor (15) using the program MEGA version 2.1 (16).

The nodC gene was amplified with the primer pairs and conditions previously described (11). Amplification of the nodC gene from each GR strain yielded a DNA fragment of about 0.4 kb (data not shown) whose nucleotide sequences were identical for all four strains and showed 99% identity to those of B. phymatum strains STM815^T and NGR195A (11). A phylogenetic tree inferred from NodC sequences from B. phymatum strains and members of the order Rhizobiales is shown in Fig. 2. Primers IGK (23) and NDR-1 (31) were used for amplification of the nifH genes as indicated earlier (22). PCR amplifications of the nifH gene and further sequencing from each GR strain revealed that they all had almost identical DNA sequences, which were 99% identical to those of B. phymatum STM815^T and NGR195A (11). A phylogenetic tree based on NifH sequences showing the relationships between B. phymatum and other Burkholderia and rhizobial species is shown in Fig. 3.

FIG. 2.

Phylogenetic tree inferred from NodC sequences shows the positions of four P. vulgaris-isolated strains, GR01, GR03, GR05, and GR06, within the genus Burkholderia. The tree topology was inferred by the neighbor-joining method (27) based on 195 sites. The bar represents the number of amino acid substitutions per site. One thousand bootstrap samplings were performed. The NCBI GenBank accession number for each strain is shown in parentheses.

FIG. 3.

Phylogenetic tree inferred from NifH sequences shows the positions of four P. vulgaris-isolated strains, GR01, GR03, GR05, and GR06, within the genus Burkholderia. The tree topology was inferred by the neighbor-joining method (27) based on 195 sites. The bar represents the number of amino acid substitutions per site. One thousand bootstrap samplings were performed. The NCBI GenBank accession number for each strain is shown in parentheses.

For nodulation tests, seeds of Glycine max, Cicer arietinum, Pisum sativum, Lens culinaris, Lotus corniculatus, and Medicago sativa were surface sterilized as described above for common beans. Seeds of Mimosa, Leucaena, Prosopis, and Acacia were surface sterilized with concentrated sulfuric acid for 10 min followed by 3% sodium hypochlorite for 10 min and then washed thoroughly with sterile water. Plant cultivation was carried out as indicated above. Acetylene reduction activity (ARA) by nodulated plants was assayed on detached root systems excised at the cotyledonary node as previously described (19). The GR strains are true symbionts of P. vulgaris as, after nodule isolation, they were able to establish new effective symbiosis with common beans, with values of ARA ranging from 492 to 525 μmol ethylene/plant/h. B. phymatum STM815^T also infected P. vulgaris, but the efficiency of the symbiosis, determined as plant dry weight (1.06 ± 0.18 g/plant/h), was half of that found in plants nodulated by the GR strains. These strains also nodulated Mimosa pigra, Acacia cochliacantha, Acacia bilimeki, Leucaena glauca, and Prosopis laevigata but were unable to form nodules on P. sativum, L. culinaris, L. corniculatus, M. sativa, G. max, and C. arietinum.

Diazotrophy is common among Burkholderia species, as shown recently by N₂ isotopic dilution studies (17). Under free-living conditions, ARA by the GR strains was tested in semisolid JMV medium as indicated earlier (26). At the end of the experiments, the culture purity was routinely checked by plating to verify uniform colony morphology. Like strain STM815^T, strains isolated from P. vulgaris also had nitrogenase activity when grown ex planta. Values of activity, however, were about half of that detected in strain STM815^T (83 ± 15 nmol C₂H₄/h).

Preparation of whole-cell proteins and sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) assays were performed as described previously (12). Protein profiles were compared with type and reference strains of legume-nodulating Burkholderia species. P. vulgaris-isolated strains showed SDS-PAGE protein profiles (evaluated by visual comparison) almost identical to those from the STM815^T type strain of B. phymatum but clearly different from those of other legume-nodulating Burkholderia species (Fig. 4). No differences were found when API 20 NE and API 50 CH strips were used to check for differences in nitrogen and carbon sources between B. phymatum STM815^T and the GR strains (data not shown).

FIG. 4.

Protein electrophoregrams (SDS-PAGE) of P. vulgaris-isolated strains and type strains of known legume-nodulating Burkholderia species. Lane 1, B. tuberum STM678^T; lane 2, B. phymatum STM815^T; lanes 3 to 6, B. phymatum GR01, GR03, GR04, and GR06, respectively; lane 7, B. mimosarum PAS44^T; lane 8, B. nodosa Br3437^T.

Based on 16S rRNA gene sequences and protein profiles, which provide strong evidence for the delineation of bacterial species (32), the Phaseolus-isolated strains could be assigned to the species B. phymatum. Moreover, analysis of the phylogenetic relationships of such sequences and other Burkholderia species showed that they formed a robust clade with B. phymatum STM815^T. In addition, sequencing of the nodC and nifH genes revealed that the phylogenetically closest bacterial species was B. phymatum STM815^T. All these results support the affiliation of Phaseolus-isolated strains as B. phymatum. Since genomic DNAs from the GR strains had the same DNA band pattern after REP-PCR fingerprinting and extremely similar 16S rRNA and nifH gene sequences, as well as identical nodC sequences, the four strains could be derived from a single clone. Our results also suggest that strains of B. phymatum isolated from Mimosa and Phaseolus have acquired their symbiosis genes either from a common ancestor or by lateral transfer between them, the direction of transfer being unknown. Although limited to three isolates, strains NGR114 and NGR195A from Mimosa invisa and Mimosa pudica in Papua New Guinea, respectively (11), and STM815^T from M. lunatum in French Guiana (21, 33), and four strains from P. vulgaris in Morocco, our results raise questions concerning the biogeographical, environmental, and host taxon distribution of B. phymatum nodule symbionts. B. phymatum was originally discovered in nodules from M. lunatum in French Guiana, and most other strains in this lineage have been found associated with host legumes in the genus Mimosa, primarily in the Neotropics. Thus, the current results extend both the known host distribution and geographic range of this group of nodule symbionts. Whether B. phymatum is prevalent on rhizobial species within nodules of P. vulgaris in the geographic site where soil samples were taken cannot be elicited from the present results. Accordingly, it will be important in future work to survey additional sites both within the native geographic range and elsewhere to understand the consistency of the association between Phaseolus and Burkholderia.