Previous Article | Next Article ![]()
Applied and Environmental Microbiology, November 2005, p. 7461-7471, Vol. 71, No. 11
0099-2240/05/$08.00+0 doi:10.1128/AEM.71.11.7461-7471.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Laboratory of Microbiology, Department of Seafood Science, National Kaohsiung Marine University, Kaohsiung City 811, Taiwan,1 EMBRAPA-Agrobiologia, Km 47, Seropédica, 23851-970 Rio de Janeiro, Brazil,2 Tropical Soil Biology and Fertility Institute of Centro Internacional de Agricultura Tropical (TSBF-CIAT), A.A. 6713, Cali, Colombia,3 CHIPs, MSI/WTB Complex, School of Life Sciences, University of Dundee, Dundee DD1 5EH, United Kingdom,4 Division of Environmental and Applied Biology, School of Life Sciences, University of Dundee, Dundee DD1 4HN, United Kingdom,5 Department of Biology 3, University of York, P.O. Box 373, York YO10 5YW, United Kingdom6
Received 10 January 2005/ Accepted 22 June 2005
|
|
|---|
|
|
|---|
-Proteobacteria (including the genera Allorhizobium, Azorhizobium, Bradyrhizobium, Mesorhizobium, Rhizobium, and Sinorhizobium) (30, 34), recently there has been an increasing number of reports of members of the ß-Proteobacteria being isolated from nodules. So far, these include Burkholderia tuberum strain STM678 and Burkholderia phymatum strain STM815 (originally isolated from Aspalathus carnosa in South Africa and Machaerium lunatum in French Guiana, respectively [26, 39]), Ralstonia taiwanensis strains (isolated from Mimosa pudica in Taiwan and India and Mimosa diplotricha in Taiwan [8, 41] and now renamed Cupriavidus taiwanensis [38]), and several Burkholderia strains isolated from Mimosa casta, Mimosa pigra (synonym, Mimosa pellita), M. pudica, and another mimosoid legume, Abarema macradenia, in Panama (3). Although symbiotic genes (nifH and nodA) have been identified in the Burkholderia strains STM678 and STM815, so far there is very little physiological and structural evidence of their symbiotic nature, and they have been shown to form only ineffective nodules on the promiscuous legume Macroptilium atropurpureum (26). More convincingly, not only have some of the Panamanian Burkholderia strains been shown to possess symbiosis-related genes (nodB and nifD), but initial nodulation studies have suggested that they may indeed be N-fixing symbionts within nodules on M. pigra (3). However, these last observations await confirmation by microscopy. The evidence of effective nodulation by C. taiwanensis is much stronger than that presented so far for Burkholderia, as Chen et al. (9) have demonstrated using a green fluorescent protein (GFP)-tagged strain and detailed light and electron microscopic studies that this bacterium can readily form N-fixing nodules on Mimosa spp.
Chen et al. (10) have shown that the genus Mimosa has a particular affinity for nodulation by ß-rhizobia. For example, in their study of 190 isolates from symbiotic nodules on M. pudica and M. diplotricha in Taiwan, the vast majority of the isolates were identified as ß-rhizobia, and these consisted mostly of C. taiwanensis (>93%), with the remainder being made up of a small number of "conventional"
-rhizobia (Rhizobium and Sinorhizobium) as well as two strains of Burkholderia caribensis, a previously described species that was not known to nodulate legumes (39).
On the basis of their data from Taiwanese Mimosa isolates, Chen et al. (10) even suggested that C. taiwanensis may actually be the "specific symbiont" of M. pudica and M. diplotricha. However, this may only hold true in Taiwan, as the legume genus Mimosa probably originated in tropical America (2). Mimosa pudica was introduced to Taiwan by Europeans in 1645 as an ornamental, which then escaped and colonized all parts of the island, and M. diplotricha was first reported in Taiwan in 1965 (43). Given that M. pudica and M. diplotricha are nonnative invasive species in Taiwan, there are two likely mechanisms to explain their apparent affinity for C. taiwanensis and other ß-rhizobial symbionts: (i) the bacteria were brought with them from tropical America or (ii) they have coopted local Taiwanese bacteria. The fact that C. taiwanensis has also been isolated from M. pudica nodules in India (41) supports the former hypothesis, and therefore it is likely that ß-rhizobia are also present in Mimosa nodules in other parts of the tropics, particularly in tropical America, where the genus originated. In the present study, we show, by comparing the sequences of their 16S rRNA genes with those of genes from reference strains, that 20 strains isolated from nodules of various Mimosa spp. in South America are all members of the genus Burkholderia. Moreover, evidence is also presented that five of these Burkholderia strains possess symbiotic genes (nifH, nodA, and nod box) and that all five can form functional, symbiotic nodules on Mimosa spp. In particular, we focus on genetically modified transconjugant variants of two of the strains, Br3461 and MAP3-5, both marked with the gfp marker gene, and hence demonstrate definitively that these two Burkholderia strains are functional symbionts of Mimosa.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. - and ß-rhizobial strains examined by ARDRA
|
Phylogenetic analyses.
Sequences were imported into BioEdit 4.8.4 (18), where amino acid sequences were deduced from the nifH and nodA sequences. ClustalW 1.4 (36), used from within BioEdit, was used to align the 16S rRNA gene, NifH, and NodA sequences and to construct neighbor-joining phylogenies with 1,000 bootstrap replicates, using Kimura distance corrections and discarding positions with gaps in any sequence. Trees were displayed using TREEVIEW (28). Sequence identities were calculated using BioEdit.
Plant tests and microscopy.
Plant cultivation and nodulation tests were carried out as described previously (9), using the tube method of Gibson (16). Seeds of Mimosa acutistipula, M. diplotricha, M. pigra, and M. pudica were surface sterilized with concentrated sulfuric acid for 10 min followed by 3% sodium hypochlorite for 10 min and then washed with sterile water. Seeds were germinated on nutrient agar plates at 28°C in darkness to make sure that there was no contamination. The tubes contained a modified Jensen's N-free plant nutrient medium (33) and were incubated at 35°C under an irradiance of 400 µmol m2 s1 with a photoperiod of 16 h. Seven days after germination, the seedlings were inoculated with 100 µl (approximately 105 cells) of a washed suspension of 1 of the 10 bacterial strains. Plants were harvested 60 days after inoculation, and nitrogenase activity was checked using the "closed" acetylene reduction assay according to the method of James and Crawford (21).
A separate tube experiment was performed with GFP-tagged strains of Br3461 and MAP3-5 that had been constructed according to the method of Chen et al. (9), using Escherichia coli S17-1
pir with pUTmini-Tn5gfp. Mimosa pudica was inoculated with either Br3461-gfp or MAP3-5-gfp, Mimosa bimucronata was inoculated with Br3461-gfp, and M. pigra was inoculated with MAP3-5-gfp. Plants were harvested 35 days after inoculation and tested for nitrogenase activity as described above. For the detection of GFP, fresh nodules were taken and sectioned (50 to 100 µm) using a Vibratome 1500 (Agar Aids, Stansted, United Kingdom). The sections were then mounted on slides and examined under either a Nikon Eclipse epifluorescence microscope or a Zeiss 510META confocal laser scanning microscope (CLSM) according to the method of Chen et al. (9). Some nodules were also fixed and embedded in resin, and sections of these were labeled with immunogold according to the method of James et al. (23), using antibodies raised against the nitrogenase NifH protein (1:100).
Nucleotide sequence accession numbers.
The GenBank accession numbers for 16S rRNA gene sequences obtained as part of this study are as follows: AY773185 (BR3405), AY773186 (BR3407), AY773187 (BR3429), AY773188 (BR3432), AY773189 (BR3437), AY773190 (BR3446), AY773191 (BR3454), AY773192 (BR3461), AY773193 (BR3462), AY773194 (BR3464), AY773195 (BR3466), AY773196 (BR3467), AY773197 (BR3469), AY773198 (BR3470), and AY533859 (MAP3-5). Accession numbers for nodA sequences are as follows: AY533869 (MAP3-5), AY533870 (Br3454), AY533871 (Br3461), AY533872 (Br3407), and AY533873 (Br3469). Accession numbers for nifH sequences are as follows: AY533864 (MAP3-5), AY533865 (Br3454), AY533866 (Br3461), AY533867 (Br3407), and AY533868 (Br3469). Accession numbers for intergenic sequences between nodD and nodB are as follows: AY533874 (MAP3-5), AY533875 (Br3454), AY533876 (Br3461), AY533877 (Br3407), and AY533878 (Br3469).
|
|
|---|
![]() View larger version (23K): [in a new window] |
FIG. 1. Dendrogram derived from the unweighted-pair group method average linkages of Dice similarity coefficients (SD) between the combined ARDRA patterns of all strains studied. The coefficient is expressed as the percentage of similarity for convenience.
|
![]() View larger version (44K): [in a new window] |
FIG. 2. Neighbor-joining tree showing phylogenetic positions of South American Mimosa-nodulating strains and Burkholderia species within the ß-proteobacteria based on 16S rRNA gene sequence comparisons. Rhizobium leguminosarum USDA 2671 was used as an outgroup. Legume symbionts are shown in bold. Bootstrap values are indicated on branches. Only bootstrap values of >50% are shown. Scale bar, 1% sequence divergence (one substitution per 100 nucleotides). Representative sequences in the dendrogram were obtained from GenBank (accession numbers are given in parentheses).
|
- and ß-rhizobia (Fig. 3). The NodA protein (116 amino acids) sequence similarities between the five South American Mimosa-nodulating strains showed that they were 89.4 to 100% similar. The nodA sequence similarities to genes encoding other ß-rhizobial NodA proteins ranged from 86.5 to 98.0% for B. phymatum, 84.6 to 87.5% for B. caribensis TJ 182, 83.6 to 86.5% for C. taiwanensis, and 60 to 75.0% for other rhizobia. As shown previously (9), the NodA sequence of B. tuberum was quite distant from that of the other ß-rhizobia (Fig. 3).
![]() View larger version (53K): [in a new window] |
FIG. 3. NodA phylogenetic tree of - and ß-rhizobia. ß-Proteobacterial strains are shown in bold. The tree was reconstructed by using a neighbor-joining approach based on a 116-amino-acid sequence alignment. Values along branches indicate bootstrap percentages of >50%, based on 1,000 replicates. nodA sequences for representative sequences are available from GenBank (accession numbers are given in parentheses).
|
|
View larger version (30K): [in a new window] |
FIG. 4. Comparison of nod box sequences in promoter regions of nodulation genes from five South American Burkholderia strains and C. taiwanensis strain LMG19424. Footnotes: a, Cupriavidus taivanensis LMG19424; b, the consensus nod box sequence (30) is located in the promoter regions of nodulation genes of various rhizobia, including Rhizobium, Sinorhizobium, Mesorhizobium, Bradyrhizobium, and Azorhizobium spp.
|
![]() View larger version (35K): [in a new window] |
FIG. 5. NifH phylogenetic tree. The tree was reconstructed by neighbor joining based on a 218-amino-acid alignment. Values along branches indicate bootstrap percentages of >50%, based on 1,000 replicates. The tree was rooted using sequences from Frankia alni, Vibrio diazotrophicus, Klebsiella pneumoniae, and Azotobacter vinelandii. Rhizobia are shown in bold, and the -, ß-, or -proteobacterial classification is indicated in parentheses. Clusters 1 and 2 contain -rhizobia only, while cluster 3 includes both symbiotic and nonsymbiotic diazotrophic ß-proteobacteria. nifH sequences for published bacteria are available from GenBank (accession numbers are given in parentheses).
|
|
View this table: [in a new window] |
TABLE 2. Nodulation of Mimosa spp. by South American Burkholderia strains
|
![]() View larger version (44K): [in a new window] |
FIG. 6. Light microscopy of fresh nodules of Mimosa spp. inoculated with gfp-tagged or wild-type (WT) Burkholderia strains Br3461 (A to D) and MAP3-5 (E to H). A Mimosa pudica nodule infected with Br3461-gfp was viewed with transmitted light (A) or epifluorescence (B). Panels C and D show a Mimosa bimucronata nodule infected with Br3461-gfp viewed under epifluorescence (C) and an infected cell from the same nodule containing fluorescent bacteroids viewed using CLSM (D). (E and F) Mimosa pudica nodules infected with MAP3-5-gfp (E) or MAP3-5 WT (F) viewed under epifluorescence. Note that the nodule containing WT MAP3-5 does not fluoresce (F). (G) Mimosa pigra nodule infected with MAP3-5-gfp viewed with epifluorescence. Note that the infected zone fluoresces intensely green (*), but also that the meristematic region has red fluorescence, possibly due to the presence of polyphenolic compounds. (H) CLSM of fluorescent bacteroids within an infected cell from an M. pigra nodule infected with MAP3-5-gfp. Bars, 100 µm (A, B, C, E, and F), 5 µm (D), 50 µm (G), and 10 µm (H).
|
![]() View larger version (177K): [in a new window] |
FIG. 7. Transverse (A) and longitudinal (B) sections of fixed and embedded Mimosa bimucronata (A) and M. pigra (B) nodules formed after inoculation with Burkholderia strains Br3461 and MAP3-5, respectively. Both nodules are indeterminate (see panel B), with a meristem (zone I), an invasion zone (zone II), and a N2-fixing zone (zone III and * in panel A). Note the heavily lignified cortex in panel A (arrow). (C and D) Transmission electron micrographs of bacteroids from nodules of M. pudica plus Br3461 (C) and M. pigra plus MAP3-5 (D). Both panels C and D were immunogold labeled with an antibody against the NifH protein of nitrogenase. Bars, 100 µm (A and B) and 300 nm (C and D).
|
|
|
|---|
The phylogenetic relationships of NodA sequences suggest that ß-rhizobia have acquired their symbiosis genes from the "conventional"
-rhizobia and that there has been more than one independent transfer between the
- and ß-proteobacteria because the ß sequences fall into two separate clades (10), with C. taiwanensis, B. caribensis TJ182, and B. phymatum in one clade and B. tuberum in the other. The nodulation genes (nodA and nod box) of all the South American strains were very similar to each other and to those of C. taiwanensis, B. caribensis TJ182, and B. phymatum, but they were quite distant from those of B. tuberum, suggesting that the strains had most likely acquired these genes from the same transfer event as the C. taiwanensis-B. caribensis TJ182-B. phymatum clade. However, the considerable divergence of NodA sequences within these ß-rhizobia indicates that this transfer event was a long time ago, implying that ß-rhizobia have been established for many millions of years. The NodA phylogeny provides strong evidence of a transfer of nodulation genes between the
- and ß-proteobacteria, but the widely held assumption that the direction of transfer was from the
to the ß group has not yet been rigorously proven. Further studies are under way to establish whether long-range horizontal gene transfer has occurred repeatedly and whether the symbiosis genes in each ß-rhizobium can be traced to a single donor source in the
-rhizobia.
Within the
-rhizobia, the nif genes seem to have a different evolutionary history from that of the nod genes, as their phylogenies are not congruent (19). Indeed, there is some evidence that the nif genes of Bradyrhizobium japonicum may have originated in the ß-Proteobacteria (46). It is therefore interesting to consider that the nif genes of the ß-rhizobia might not, in fact, have the same
-proteobacterial origin as the nod genes. This should be evident from phylogenetic analysis, and the indication for strains examined so far is that NifH is indeed the "local" ß-proteobacterial form (10). In the present study, the NifH sequences of all five strains examined were very close to each other and to those of C. taiwanensis, B. caribensis, and B. phymatum but were not so close to that of B. tuberum. Therefore, as with the NodA sequences, this suggests that the former group of bacteria obtained their nif genes from a different source than that of B. tuberum.
Origin of strains.
All 14 Brazilian strains were originally isolated over 20 years ago by de Faria and coworkers (unpublished) and have long been known to have the ability to effectively nodulate their original hosts (all Brazilian natives) as well as other Mimosa spp. Indeed, they have all been extensively tested for high N fixation in an evaluation of tropical woody legumes (including Mimosa spp.) for use in the reclamation of degraded and deforested areas in Brazil (15). With specific regard to the four representative Brazilian strains, Br3454 is of particular interest, as details of its ability to infect and nodulate M. scabrella were reported by de Faria et al. (12). This study showed that "Rhizobium" strain Br3454 had an unusual ability to infect the roots of M. scabrella by direct epidermal infection, and so far M. scabrella is the only legume reported to be infected in this manner. The hosts of the other three strains, M. caesalpiniaefolia (Br3407), M. bimucronata (Br3461), and M. camporum (Br3469), are all common in forested areas of Brazil (2, 13, 32) and have long been known to be nodulated (1, 5, 7, 20, 29, 34). Mimosa caesalpiniaefolia, a fast-growing tree, has been well studied owing to its ability to grow and nodulate in acidic soils (35) and for its potential for use in the reclamation of degraded areas (15, 17), while the smaller, invasive shrubs M. bimucronata and M. camporum are being evaluated for their potential to assist in the recovery of drastically disturbed lands (29) and to withstand prolonged shading (20), respectively.
Mimosa pigra (syn. M. pellita) is likely to have originated in South America (2), where it has long been known to undergo nodulation, particularly in wetland regions (5, 24). However, in contrast to the hosts of the four Brazilian strains, M. pigra is highly invasive and is now becoming a major pest in some countries, such as in wetland regions in tropical Australia (6), Thailand (45), and southern Taiwan (44). Several Burkholderia strains have recently been isolated from M. pigra nodules in Taiwan (9a) and Panama (3), but strain MAP3-5 and the other MAP strains were isolated from nodules of M. pigra growing alongside the Mapire river, a tributary of the Orinoco in Venezuela, by Barrios (unpublished). Strain MAP3-5 was shown to effectively nodulate M. pigra in a preliminary study by James et al. (22). This last study also suggested that M. pigra, at least when grown under flooded conditions, was infected by MAP3-5 via enlarged epidermal/aerenchyma cells and not via root hairs. This is a particularly interesting observation given the very close relationship between MAP3-5 and Burkholderia strain Br3454, which infects M. scabrella via direct epidermal penetration (12), and hence further infection/microscopy studies are being undertaken with the gfp transconjugant strain of MAP3-5 to determine more exactly how it infects different Mimosa spp.
Geographical distribution of Mimosa-nodulating ß-rhizobia.
Although a few species are native to Asia (e.g., Mimosa himalayana) and Africa (particularly Madagascar), most of the 480 or so species of Mimosa are native to Central and South America (2), with the Cerrado region of Central Brazil being the major center of diversification for the genus (2, 32). It has long been known that Mimosa plants may be nodulated by a variety of rhizobia (1, 37), but few have been typified, and prior to 2000, all had been ascribed to known
-rhizobial genera (3, 25, 27, 42, 47). Since the initial report of Moulin et al. (26) on Burkholderia strains STM815 and STM678, many other strains of "ß-rhizobia" have been isolated, but the majority of them have come from Mimosa spp. in Asia (8, 9a, 10, 41), and relatively few have been found in the Americas (3). Even though they came from geographically very distant parts of South America, i.e., the Brazilian Atlantic Forest (most of the "Br" strains) and the flooded forests of the Orinoco basin (MAP3-1 to MAP3-6), and from 10 different Mimosa spp., all 20 strains examined in the present study were ß-rhizobia. Obviously, considerably more South American Mimosa nodule isolates need to be examined before any firm conclusions can be made, but our study not only gives confirmation that they have a particular preference for Mimosa spp. (10 different species) but also suggests that they may be the "dominant" rhizobial type involved with the genus Mimosa across large parts of South America. Indeed, this is supported by the recent report of 27 Burkholderia strains isolated from mimosoid legumes in Central America (3). Furthermore, it is interesting that none of the 20 strains were in the genus Cupriavidus, which appears strange considering that C. taiwanensis is so dominant in Taiwan (10) and possibly India (41). Could it be that C. taiwanensis is an Asian bacterium that has acquired its symbiosis genes from Burkholderia strains resident within Mimosa nodules that were brought from tropical America and the Caribbean by European colonists from 1645 onwards (43)? Clearly, further studies are urgently needed to compare South American ß-rhizobia with those from other continents.
W.-M. Chen was supported by grants from the National Science Council, Taipei, Taiwan, Republic of China (NSC 93-2320-B-022-001 and 003), and A. Prescott, G. N. Elliott, J. I. Sprent, J. P. W. Young, and E. K. James were supported by the Natural Environment Research Council (NERC).
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»