MPI-blastn and NCBI-TaxCollector: improving metagenomic analysis with high performance classification and wide taxonomic attachment.

Metagenomic sequencing technologies are advancing rapidly and the size of output data from high-throughput genetic sequencing has increased substantially over the years. This brings us to a scenario where advanced computational optimizations are requested to perform a metagenomic analysis. In this paper, we describe a new parallel implementation of nucleotide BLAST (MPI-blastn) and a new tool for taxonomic attachment of Basic Local Alignment Search Tool (BLAST) results that supports the NCBI taxonomy (NCBI-TaxCollector). MPI-blastn obtained a high performance when compared to the mpiBLAST and ScalaBLAST. In our best case, MPI-blastn was able to run 408 times faster in 384 cores. Our evaluations demonstrated that NCBI-TaxCollector is able to perform taxonomic attachments 125 times faster and needs 120 times less RAM than the previous TaxCollector. Through our optimizations, a multiple sequence search that currently takes 37 hours can be performed in less than 6 min and a post processing with NCBI taxonomic data attachment, which takes 48 hours, now is able to run in 23 min.

The intestinal microbiome in type 1 diabetes.

Few concepts in recent years have garnered more disease research attention than that of the intestinal (i.e. 'gut') microbiome. This emerging interest has included investigations of the microbiome's role in the pathogenesis of a variety of autoimmune disorders, including type 1 diabetes (T1D). Indeed, a growing number of recent studies of patients with T1D or at varying levels of risk for this disease, as well as in animal models of the disorder, lend increasing support to the notion that alterations in the microbiome precede T1D onset. Herein, we review these investigations, examining the mechanisms by which the microbiome may influence T1D development and explore how multi-disciplinary analysis of the microbiome and the host immune response may provide novel biomarkers and therapeutic options for prevention of T1D.

Determination of the amino acid residues required for the activity of the anti-rhizobial peptide antibiotic trifolitoxin.

AIMS: The first aim was to determine those amino acid residues required for the biological activity of the potent peptide antibiotic, trifolitoxin (TFX). The second aim was to determine the concentrations of TFX1 and TFX2 that cause 50% inhibition of bacterial growth (Ki), the two predominant isomeric forms of TFX made by Rhizobium. METHODS AND RESULTS: Site-directed mutagenesis of tfxA was used to produce strains that made mutant TFX peptides. The mutant tfxA genes were placed on a vector and inserted in Rhizobium leguminosarum b. trifolii Tn54A112, a tfxA mutant of strain T24 that lacks trifolitoxin activity. Our standard bioassay was used to assess the activity of these mutants. TFX1 and TFX2 were purified by reverse phase chromatography. Several concentrations of each peptide were assayed for biological activity to determine Ki. The unmodified TFX peptide (DIGGSRQGCVA) was synthesized and was found to lack any biological activity. Four of the 11 amino acid residues in ribosomally synthesized, post-translationally modified peptide were required for TFX activity. These required amino acids include arginine (R37), glutamine (Q38), glycine (G39) and cysteine (C40). S36T and S36Y mutants showed reduced TFX activity. The numbering system is based on the 42-amino acid TfxA peptide that is post-translationally modified to form the active TFX peptide. The Ki of TFX2 was determined to be 10-fold lower than TFX1. CONCLUSIONS: The post-translational modifications of the TfxA peptide are required for biological activity. TFX2 is far more active than TFX1. SIGNIFICANCE AND IMPACT OF THE STUDY: The sequence of the TfxA peptide appears to have been optimized for maximum activity through the course of evolution. Even conservative changes to any of the amino acid residues required for activity results in a complete loss of activity. The understanding of the action of this peptide is critical for its proposed action as a control agent for crown gall disease.

Seasonal dynamics of phytoplankton and planktonic protozoan communities in a northern temperate humic lake: diversity in a dinoflagellate dominated system.

Species diversity and richness, and seasonal population dynamics of phytoplankton, planktonic protozoa, and bacterioplankton sampled from the epilimnion of Crystal Bog in 2000, were examined in order to test the hypothesis that these groups' diversity and abundance patterns might be linked. Crystal Bog, a humic lake in Vilas County, Wisconsin, is part of the North Temperate Lakes Long-Term Ecological Research Site. Phytoplankton and planktonic protozoa were identified and enumerated in a settling chamber with an inverted microscope. Bacterial cells were enumerated with the use of fluorescence 4', 6'-diamidino-2-phenylindole (DAPI)-staining procedures, and automated ribosomal intergenic spacer analysis (ARISA) was used to assess bacterioplankton diversity. Bacterial cell counts showed little seasonal variation and averaged 2.6 x 10(6) cells/mL over the ice-free season. Phytoplankton and planktonic protozoan numbers varied by up to two orders of magnitude and were most numerous in late spring and summer. Dinoflagellates largely dominated Crystal Bog throughout the ice-free period, specifically Peridiniopsis quadridens in the spring, Peridinium limbatum in summer, and Gymnodinium fuscum and P. quadridens in fall. Brief blooms of Cryptomonas, Dinobryon, and Synura occurred between periods of dinoflagellate domination. The dominant dinoflagellate, Peridinium limbatum, was calculated to have a growth rate of 0.065 day(-1) and a doubling time of 10.7 days. Heterotrophic nanoflagellates (HNFs) were a consistent component of the planktonic protozoa; seasonal patterns were determined for three genera of HNFs (Monosiga, Bicosoeca, and Desmarella moniliformis). Three genera of ciliates (Coleps, Strobilidium, and Strombidium) comprised the greater part of the planktonic protozoa in Crystal Bog. The number of species of planktonic protozoa was too low to calculate a diversity index. Shannon-Weaver diversity indices for phytoplankton and bacterioplankton in the epilimnion followed very similar seasonal patterns in this lake, supporting the hypothesis that in freshwaters, diversity patterns of these groups are linked.

Annual patterns in bacterioplankton community variability in a humic lake.

Bacterioplankton community composition (BCC) was monitored in a shallow humic lake in northern Wisconsin, USA, over 3 years using automated ribosomal intergenic spacer analysis (ARISA). Comparison of ARISA profiles of bacterial communities over time indicated that BCC was highly variable on a seasonal and annual scale. Nonmetric multidimensional scaling (MDS) analysis indicated little similarity in BCC from year to year. Nevertheless, annual patterns in bacterioplankton community diversity were observed. Trends in bacterioplankton community diversity were correlated to annual patterns in community succession observed for phytoplankton and zooplankton populations, consistent with the notion that food web interactions affect bacterioplankton community structure in this humic lake. Bacterioplankton communities experience a dramatic drop in richness and abundance each year in early summer, concurrent with an increase in the abundance of both mixotrophic and heterotrophic flagellates. A second drop in richness, but not abundance, is observed each year in late summer, coinciding with an intense bloom of the nonphagotrophic dinoflagellate Peridinium limbatum. A relationship between bacterial community composition, size, and abundance and the population dynamics of Daphnia was also observed. The noted synchrony between these major population and species shifts suggests that linkages across trophic levels play a role in determining the annual time course of events for the microbial and metazoan components of the plankton.

Temporal patterns in bacterial communities in three temperate lakes of different trophic status.

Despite considerable attention in recent years, the composition and dynamics of lake bacterial communities over annual time scales are poorly understood. This study used automated ribosomal intergenic spacer analysis (ARISA) to explore the patterns of change in lake bacterial communities in three temperate lakes over 2 consecutive years. The study lakes included a humic lake, an oligotrophic lake, and a eutrophic lake, and the epilimnetic bacterial communities were sampled every 2 weeks. The patterns of change in bacterial communities indicated that seasonal forces were important in structuring the behavior of the bacterial communities in each lake. All three lakes had relatively stable community composition in spring and fall, but summer changes were dramatic. Summertime variability was often characterized by recurrent drops in bacterial diversity. Specific ARISA fragments derived from these lakes were not constant among lakes or from year to year, and those fragments that did recur in lakes in different years did not exhibit the same seasonal pattern of recurrence. Nonetheless, seasonal patterns observed in 2000 were fairly successful predictors of the rate of change in bacterial communities and in the degree of autocorrelation of bacterial communities in 2001. Thus, seasonal forces may be important structuring elements of these systems as a whole even if they are uncoupled from the dynamics of the individual system components.

The Diversity of Archaea and Bacteria in Association with the Roots of Zea mays L.

The diversity of bacteria and archaea associating on the surface and interior of maize roots (Zea mays L.) was investigated. A bacterial 16S rDNA primer was designed to amplify bacterial sequences directly from maize roots by PCR to the exclusion of eukaryotic and chloroplast DNA. The mitochondrial sequence from maize was easily separated from the PCR-amplified bacterial sequences by size fractionation. The culturable component of the bacterial community was also assessed, reflecting a community composition different from that of the clone library. The phylogenetic overlap between organisms obtained by cultivation and those identified by direct PCR amplification of 16S rDNA was 48%. Only 4 bacterial divisions were found in the culture collection, which represented 27 phylotypes, whereas 6 divisions were identified in the clonal analysis, comprising 74 phylotypes, including a member of the OP10 candidate division originally described as a novel division level lineage in a Yellowstone hot spring. The predominant group in the culture collection was the actinobacteria and within the clone library, the a-proteobacteria predominated. The population of maize-associated proteobacteria resembled the proteobacterial population of a typical soil community within which resided a subset of specific plant-associated bacteria, such as Rhizobium- and Herbaspirillum-related phylotypes. The representation of phylotypes within other divisions (OP10 and Acidobacterium) suggests that maize roots support a distinct bacterial community. The diversity within the archaeal domain was low. Of the 50 clones screened, 6 unique sequence types were identified, and 5 of these were highly related to each other (sharing 98% sequence identity). The archaeal sequences clustered with good bootstrap support near Marine group I (crenarchaea) and with Marine group II (euryarchaea) uncultured archaea. The results suggest that maize supports a diverse root-associated microbial community composed of species that for the first time have been described as inhabitants of a plant-root environment.

Genomic interspecies microarray hybridization: rapid discovery of three thousand genes in the maize endophyte, Klebsiella pneumoniae 342, by microarray hybridization with Escherichia coli K-12 open reading frames.

In an effort to efficiently discover genes in the diazotrophic endophyte of maize, Klebsiella pneumoniae 342, DNA from strain 342 was hybridized to a microarray containing 96% (n = 4,098) of the annotated open reading frames from Escherichia coli K-12. Using a criterion of 55% identity or greater, 3,000 (70%) of the E. coli K-12 open reading frames were also found to be present in strain 342. Approximately 24% (n = 1,030) of the E. coli K-12 open reading frames are absent in strain 342. For 1.6% (n = 68) of the open reading frames, the signal was too low to make a determination regarding the presence or absence of the gene. Genes with high identity between the two organisms are those involved in energy metabolism, amino acid metabolism, fatty acid metabolism, cofactor synthesis, cell division, DNA replication, transcription, translation, transport, and regulatory proteins. Functions that were less highly conserved included carbon compound metabolism, membrane proteins, structural proteins, putative transport proteins, cell processes such as adaptation and protection, and central intermediary metabolism. Open reading frames of E. coli K-12 with little or no identity in strain 342 included putative regulatory proteins, putative chaperones, surface structure proteins, mobility proteins, putative enzymes, hypothetical proteins, and proteins of unknown function, as well as genes presumed to have been acquired by lateral transfer from sources such as phage, plasmids, or transposons. The results were in agreement with the physiological properties of the two strains. Whole genome comparisons by genomic interspecies microarray hybridization are shown to rapidly identify thousands of genes in a previously uncharacterized bacterial genome provided that the genome of a close relative has been fully sequenced. This approach will become increasingly more useful as more full genome sequences become available.

Dyadobacter fermentans gen. nov., sp. nov., a novel gram-negative bacterium isolated from surface-sterilized Zea mays stems.

A Gram-negative bacterium, designated NS114T, was isolated from duplicate treatments of surface-sterilized Zea mays stems. The plants were grown in synthetic soil under greenhouse conditions and watered with fertilizer containing no nitrogen. Strain NS114T could not be isolated from plants watered with the standard level or 20% (w/v) of the standard level of nitrogen. Cells occurred as pairs in young cultures that attached to form angled arrangements in R2A broth and occasionally formed rounded, horseshoe arrangements in YM broth. Cell variation resulted in flocculent chains of coccoid cells in old cultures. Strain NS114T fermented glucose and sucrose. The G + C content was 48 mol%. Phylogenetic analysis of the 16S rRNA gene showed that the strain was a member of the domain Bacteria and branched from a point equidistant from an aquatic organism, Runella slithyformis and a marine isolate, 'Microscilla furvescens'. Phenotypic and genotypic analyses indicated that strain NS114T could not be assigned to any recognized genus; therefore a new genus and species, Dyadobacter fermentans gen. nov., sp. nov., is proposed, for which NS114T is the type strain.

Immunolocalization of dinitrogenase reductase produced by Klebsiella pneumoniae in association with Zea mays L.

The endophytic lifestyle of Klebsiella pneumoniae is described, including the production of dinitrogenase reductase by bacteria residing in maize root tissue. The green fluorescent protein (GFP) was used to detect the colonization of maize by K. pneumoniae strains 2028 and 342. These strains were found to reside in intercortical layers of the stem and within the region of maturation in the root. The production of dinitrogenase reductase by GFP-tagged bacteria was visualized using immunolocalization. This activity was only apparent when bacteria were supplied with an exogenous carbon source. The results suggest that maize provides a suitable habitat for K. pneumoniae and that this species is capable of producing nitrogenase under the appropriate plant cultivation conditions.

Automated approach for ribosomal intergenic spacer analysis of microbial diversity and its application to freshwater bacterial communities.

An automated method of ribosomal intergenic spacer analysis (ARISA) was developed for the rapid estimation of microbial diversity and community composition in freshwater environments. Following isolation of total community DNA, PCR amplification of the 16S-23S intergenic spacer region in the rRNA operon was performed with a fluorescence-labeled forward primer. ARISA-PCR fragments ranging in size from 400 to 1,200 bp were next discriminated and measured by using an automated electrophoresis system. Database information on the 16S-23S intergenic spacer was also examined, to understand the potential biases in diversity estimates provided by ARISA. In the analysis of three natural freshwater bacterial communities, ARISA was rapid and sensitive and provided highly reproducible community-specific profiles at all levels of replication tested. The ARISA profiles of the freshwater communities were quantitatively compared in terms of both their relative diversity and similarity level. The three communities had distinctly different profiles but were similar in their total number of fragments (range, 34 to 41). In addition, the pattern of major amplification products in representative profiles was not significantly altered when the PCR cycle number was reduced from 30 to 15, but the number of minor products (near the limit of detection) was sensitive to changes in cycling parameters. Overall, the results suggest that ARISA is a rapid and effective community analysis technique that can be used in conjunction with more accurate but labor-intensive methods (e.g., 16S rRNA gene cloning and sequencing) when fine-scale spatial and temporal resolution is needed.

A transposable partitioning locus used to stabilize plasmid-borne hydrogen oxidation and trifolitoxin production genes in a Sinorhizobium strain.

Improved nitrogen-fixing inoculum strains for leguminous crops must be able to effectively compete with indigenous strains for nodulation, enhance legume productivity compared to the productivity obtained with indigenous strains, and maintain stable expression of any added genes in the absence of selection pressure. We constructed a transposable element containing the tfx region for expression of increased nodulation competitiveness and the par locus for plasmid stability. The transposon was inserted into tetA of pHU52, a broad-host-range plasmid conferring the H2 uptake phenotype. The resulting plasmid, pHUTFXPAR, conferred the plasmid stability, trifolitoxin production, and H2 uptake phenotypes in the broad-host-range organism Sinorhizobium sp. strain ANU280. The broad applications of a transposon conferring plasmid stability are discussed.

Molecular microbial diversity in soils from eastern Amazonia: evidence for unusual microorganisms and microbial population shifts associated with deforestation.

Although the Amazon Basin is well known for its diversity of flora and fauna, this report represents the first description of the microbial diversity in Amazonian soils involving a culture-independent approach. Among the 100 sequences of genes coding for small-subunit rRNA obtained by PCR amplification with universal small-subunit rRNA primers, 98 were bacterial and 2 were archaeal. No duplicate sequences were found, and none of the sequences had been previously described. Eighteen percent of the bacterial sequences could not be classified in any known bacterial kingdom. Two sequences may represent a unique branch between the vast majority of bacteria and the deeply branching, predominantly thermophilic bacteria. Five sequences formed a clade that may represent a novel group within the class Proteobacteria. In addition, rRNA intergenic spacer analysis was used to show significant microbial population differences between a mature forest soil and an adjacent pasture soil.

Isolation and characterization of the UDP-glucose 4'-epimerase-encoding gene, galE, from Brucella abortus 2308.

The UDP-glucose 4'-epimerase-encoding gene, galE or exoB, was isolated from Brucella abortus 2308 by complementation of an exoB mutant of Sinorhizobium meliloti. Confirmation of the identity was done by constructing an in-frame deletion of 660 bp with galE of the B. abortus genome by marker exchange. The resulting galE mutant lacked UDP-glucose 4'-epimerase activity. This activity was restored by in trans complementation with the intact gene. The B. abortus gal E mutant is not altered in colony morphology compared to wt 2308. The lack of UDP-glucose 4'-epimerase activity in the mutant and PCR analysis strongly suggest that only one copy of galE exists in B. abortus 2308. The galE sequence of B. abortus 2308 is more similar to galE from other animal-inhabiting bacteria than it is to exoB from the Sinorhizobium legume symbionts. We propose that galE in B. abortus evolved by lateral transfer from other animal-inhabiting bacteria rather than from a common ancestor of Brucella and Sinorhizobium.

Inoculation with Sinorhizobium meliloti RMBPC-2 Increases Alfalfa Yield Compared with Inoculation with a Nonengineered Wild-Type Strain.

Inoculation of alfalfa seeds with any of three recombinant strains of Sinorhizobium meliloti significantly increased overall plant biomass compared with inoculation with the wild-type strains over a 3-year period at three locations. A high proportion of nodules were occupied by the inoculum strains throughout the 3-year period.

A newly discovered gene, tfuA, involved in the production of the ribosomally synthesized peptide antibiotic trifolitoxin.

Trifolitoxin (TFX) is a gene-encoded, posttranslationally modified peptide antibiotic. Previously, we have shown that tfxABCDEFG from Rhizobium leguminosarum bv. trifolii T24 is sufficient to confer TFX production and resistance to nonproducing strains within a distinct taxonomic group of the alpha-proteobacteria (E. W. Triplett, B. T. Breil, and G. A. Splitter, Appl. Environ. Microbiol. 60:4163-4166, 1994). Here we describe strain Tn5-2, a Tn5 mutant of T24 defective in the production of TFX, whose insertion maps outside of the tfx cluster. It is not altered in growth compared with T24, nor does it inactivate TFX in its proximity. The wild-type analog of the mutated region of Tn5-2 was cloned. Sequencing, transcriptional fusion mutagenesis, and subcloning were used to identify tfuA, a gene involved in TFX production. On the basis of computer analysis, the putative TfuA protein has a mass of 72.9 kDa and includes a peroxidase motif but no transmembrane domains. TFX production studies show that extra copies of the tfxABCDEFG fragment increase TFX production in a T24 background while additional copies of tfuA do not. Lysate ribonuclease protection assays suggest that tfuA does not regulate transcription of tfxA. Upstream of tfuA are two open reading frames (ORFs). The putative product of ORF1 shows high similarity to the LysR family of transcriptional regulators. The putative product of ORF2 shows high similarity to the cytosine deaminase (CodA) of Escherichia coli.

Molecular microbial diversity of an agricultural soil in Wisconsin.

A culture-independent survey of the soil microbial diversity in a clover-grass pasture in southern Wisconsin was conducted by sequence analysis of a universal clone library of genes coding for small-subunit rRNA (rDNA). A rapid and efficient method for extraction of DNA from soils which resulted in highly purified DNA with minimal shearing was developed. Universal small-subunit-rRNA primers were used to amplify DNA extracted from the pasture soil. The PCR products were cloned into pGEM-T, and either hypervariable or conserved regions were sequenced. The relationships of 124 sequences to those of cultured organisms of known phylogeny were determined. Of the 124 clones sequenced, 98.4% were from the domain Bacteria. Two of the rDNA sequences were derived from eukaryotic organelles. Two of the 124 sequences were of nuclear origin, one being fungal and the other a plant sequence. No sequences of the domain Archaea were found. Within the domain, Bacteria, three kingdoms were highly represented: the Proteobacteria (16.1%), the Cytophaga-Flexibacter-Bacteroides group (21.8%), and the low G+C-content gram-positive group (21.8%). Some kingdoms, such as the Thermotogales, the green nonsulfur group, Fusobacteria, and the Spirochaetes, were absent. A large number of the sequences (39.4%) were distributed among several clades that are not among the major taxa described by Olsen et al. (G.J. Olsen, C.R. Woese, and R. Overbeek, J. Bacteriol., 176:1-6, 1994). From the alignments of the sequence data, distance matrices were calculated to display the enormous microbial diversity found in this soil in two ways, as phylogenetic trees and as multidimensional-scaling plots.

Expression of tfx and sensitivity to the rhizobial peptide antibiotic trifolitoxin in a taxonomically distinct group of alpha-proteobacteria including the animal pathogen Brucella abortus.

Three phylogenetically distinct groups within the alpha-proteobacteria which differ in trifolitoxin sensitivity are described. Trifolitoxin sensitivity was found in strains of Agrobacterium, Brucella, Mycoplana, Ochrobactrum, Phyllobacterium, Rhodobacter, Rhodopseudomonas, Rhodospirillum, and Rhizobium. Strains of Agrobacterium, Brucella, Phyllobacterium, Rhizobium, and Rhodospirillum were capable of producing trifolitoxin upon conjugal transfer of tfxABCDEFG.

Alfalfa yield response to inoculation with recombinant strains of Rhizobium meliloti with an extra copy of dctABD and/or modified nifA expression.

The construction of rhizobial strains which increase plant biomass under controlled conditions has been previously reported. However, there is no evidence that these newly constructed strains increase legume yield under agricultural conditions. This work tested the hypothesis that carefully manipulating expression of additional copies of nifA and dctABD in strains of Rhizobium meliloti would increase alfalfa yield in the field. The rationale for this hypothesis is based on the positive regulatory role that nifA plays in the expression of the nif regulon and the fact that a supply of dicarboxylic acids from the plant is required as a carbon and energy source for nitrogen fixation by the Rhizobium bacteroids in the nodule. These recombinant strains, as well as the wild-type strains from which they were derived, are ideal tools to examine the effects of modifying or increasing the expression of these genes on alfalfa biomass. The experimental design comprised seven recombinant strains, two wild-type strains, and an uninoculated control. Each treatment was replicated eight times and was conducted at four field sites in Wisconsin. Recombinant strain RMBPC-2, which has an additional copy of both nifA and dctABD, increased alfalfa biomass by 12.9% compared with the yield with the wild-type strain RMBPC and 17.9% over that in the uninoculated control plot at the site where soil nitrogen and organic matter content was lowest. These increases were statistically significant at the 5% confidence interval for each of the three harvests made during the growing season. Strain RMBPC-2 did increase alfalfa biomass at the Hancock site; however, no other significant increases or decreases in alfalfa biomass were observed with the seven other recombinant strains at that site. At three sites where this experiment was conducted, either native rhizobial populations or soil nitrogen concentrations were high. At these sites, none of the recombinant strains affected yield. We conclude that RMBPC -2 can increase alfalfa yields under field conditions of nitrogen limitation, low endogenous rhizobial competitors, and sufficient moisture.

DNA sequence of the common nodulation genes of Bradyrhizobium elkanii and their phylogenetic relationship to those of other nodulating bacteria.

A 6.6-kb BamHI fragment containing the common nodulation genes of Bradyrhizobium elkanii USDA94 was identified by southern hybridization using the common nod genes of B. japonicum as a probe. This fragment was cloned and sequenced. Analysis of the sequence showed open reading frames highly homologous to nolA, nodD2, nodD1, and nodKABC from other bradyrhizobial sources. The sequence showed higher homology to the common nod genes of Bradyrhizobium sp. (Parasponia) than to those from B. japonicum. The open reading frame identified between nodD1 and nodA in the B. elkanii sequence was far more similar to nodK from Bradyrhizobium sp. (Parasponia) than to nodY from B. japonicum. The molecular phylogeny of nodD and nodAB from many sources was analyzed. The genetic distance between the nod genes is far greater than the distance between the 16S rRNA and nifH genes. The differences between the nod genes among the species of Rhizobium is as great as that between Bradyrhizobium and Rhizobium. The host range of the microsymbiont was found to be a better predictor of the similarities of the common nod genes than the 16S rRNA or nifH genes. We propose two groups of nod genes among the rhizobia and bradyrhizobia, based on molecular phylogenetic analysis: those which nodulate legumes of temperate origin in the tribes Vicieae or Trifolieae and those which nodulate legumes of tropical origin in the tribe Phaseoleae.