An Ouchterlony double diffusion study on the interaction between legume lectins and rhizobial cell surface antigens

Acidic exopolysaccharides and O-antigen containing lipopolysaccharides were isolated from Rhizobium japonicum, R. leguminosarum, R. lupini, R. meliloti, R. phaseoli, cowpea Rhizobium sp. and a non-nodulating soil bacterium. Lectins from seeds of soybean (Glycine max), garden pea (Pisum sativum), lentil (Lens culinaris), alfalfa (Medicago sativa), field bean (Phaseolus vulgaris), jackbean (Canavalia ensiformis) and from wheat germ were tested for their capacity to precipitate rhizobial exopolysaccharides and lipopolysaccharides in the Ouchterlony double diffusion test. Soybean lectin precipitated exclusively with the exopolysaccharide of R. japonicum, whereas the lectins from pea and lentil precipitated exopolysaccharides from all the fast growing strains of Rhizobium. Host range specific interactions between lipopolysaccharides and lectins were observed in the pea/lentil-R. leguminosarum and in the alfalfa-R. meliloti systems. Concanavalin A precipitated the exopolysaccharides of all fast growing strains of Rhizobium, the exopolysaccharide of the cowpea strain and several lipopolysaccharides of different Rhizobium species and thus did not show any correlation between polysaccharide binding and symbiotic specificity. Non-leguminous wheat germ agglutinin did not precipitate any of the rhizobial polysaccharides tested and the lipopolysaccharide of the soil bacterium did not precipitate with any of the lectins examined.Abbreviations Con A Concanavalin A - CPC cetylpyridinium chioride - EPS exopolysaccharide - FITC fluorescein isothiocyanate - KDO 2-keto-3-deoxyoctonic acid - LPS lipopolysaccharide - PBS phosphate-buffered saline - PS polysaccharide     

Comparative Characteristics of Carbohydrate Binding by Lectins from Broad Bean,Pea, Common Vetch,and Lentil Seeds

Lectins from the seeds of broad bean (Vicia faba L.), pea (Pisum sativum L.), common vetch (V. sativa L.), and lentil (Lens culinaris Medik.) were isolated and purified by affinity chromatography. The hemagglutinating activity of lectins was most effectively inhibited by methyl--D-mannopyranoside, trehalose, and D-mannose. Other carbohydrate haptens, such as methyl--D-glucopyranoside, maltose, and alginic and D-glucuronic acids were less effective. Two lectins obtained from different lentil cultivars, unlike other lectins, had a relatively high affinity for melecitose, N-acetyl-D-glucosamine, L-sorbose, and sucrose. Furthermore, these lectins interacted with soluble starch. All the lectins examined had similar, but not identical, carbohydrate-binding properties. Because of their similar D-mannose/D-glucose specificity, these lectins interacted with lipopolysaccharides and exopolysaccharides of Rhizobium leguminosarum bv. viciae, root nodule bacteria that infect broad-bean, pea, common-vetch, and lentil plants with the formation of nitrogen-fixing symbiosis. However, owing to individual distinctions of carbohydrate-binding properties, these lectins showed a higher affinity for the polysaccharides of those microsymbionts within the R. leguminosarum bv. viciae species that were better specialized towards one or the other host plant from the cross inoculation group of legumes.     

Role of Lectins in the Specific Recognition of Rhizobium by Lotononis bainesii

Fluorescein isothiocyanate (FITC)-labeled lectin purified from the root of Lotononis bainesii Baker was bound by cells of five out of seven L. bainesii-nodulating strains of Rhizobium under culture conditions. With the exception of a strain of Rhizobium leguminosarum, strains of noninfective rhizobia failed to bind the root lectin under these conditions. The two nonlectin binding L. bainesii-specific strains did not bind root lectin on the L. bainesii rhizoplane although this was observed with three other L. bainesii-nodulating strains. A single Rhizobium japonicum strain bound root lectin on the L. bainesii rhizoplane. There was no evidence of an interaction between the L. bainesii seed lectin and the Rhizobium strains tested.

Root lectin-specific FITC-labeled antibodies were bound to the tips of developing root hairs and lateral growth points of more mature root hairs of L. bainesii seedlings. The damaged edges of severed root hairs always bound FITC-labeled root lectin antibody. Seed lectin-specific FITC-labeled antibodies were not bound to the roots of L. bainesii. The preemergent root hair region of L. bainesii was most susceptible to infection by rhizobia but nodules also emerged in the developing and mature root hair regions. Lectin exposed at growth points on L. bainesii root hairs may provide a favorable site for host plant recognition of infective strains of Rhizobium.

     

Root Exudates of Various Host Plants of Rhizobium leguminosarum Contain Different Sets of Inducers of Rhizobium Nodulation Genes

Rhizobium promoters involved in the formation of root nodules on leguminous plants are activated by flavonoids in plant root exudate. A series of Rhizobium strains which all contain the inducible Rhizobium leguminosarum nodA promoter fused to the Escherichia coli lacZ gene, and which differ only in the source of the regulatory nodD gene, were recently used to show that the regulatory nodD gene determines which flavonoids are able to activate the nodA promoter (HP Spaink, CA Wijffelman, E Pees, RJH Okker, BJJ Lugtenberg 1987 Nature 328: 337-340). Since these strains therefore are able to discriminate between various flavonoids, they were used to determine whether or not plants that are nodulated by R. leguminosarum produce different inducers. After chromatographic separation of root exudate constituents from Vicia sativa L. subsp. nigra (L.), V. hirsuta (L.) S.F. Gray, Pisum sativum L. cv Rondo, and Trifolium subterraneum L., the fractions were tested with a set of strains containing a nodD gene of R. leguminosarum, R. trifolii, or Rhizobium meliloti, respectively. It appeared that the source of nodD determined whether, and to what extent, the R. leguminosarum nodA promoter was induced. Lack of induction could not be attributed to the presence of inhibitors. Most of the inducers were able to activate the nodA promoter in the presence of one particular nodD gene only. The inducers that were active in the presence of the R. leguminosarum nodD gene were different in each root exudate.     

Attachment of Rhizobium to Legume Roots as the Basis for Specific Interactions

Experiments were conducted to elucidate the basis of the observation that different strains of Rhizobium infect particular legumes. Rhizobia specific for a variety of legumes were grown with ¹³PO^2?₄ and exposed to pea roots (Pisum sativum L.), R. leguminosarum 128C53, which nodulates pea, did not attach to the roots in greater numbers than those strains of rhizobia incapable of infecting pea roots. A complex of R. leguminosarum 128C53 conjugated to a fluorochrome-labeled antibody exhibited a striking attachment to the tips of pea root hairs, where infection normally occurs, but this fluorescent complex also bound to the root hairs of Canavalia en siformis DC., Lupinus polyphyllus Lindl., Trifolium pratense L., and Medicago sativa L., which are not infected by this bacterium. A reproducible, quantitative technique developed for studying interactions between fluorochrome-labeled lectins and rhizobia revealed no relationship between lectin-Rhizobium interactions and the capacity to infect a plant. The data are interpreted as suggesting that simple attachment of Rhizobium to a legume root is not the basis of host-symbiont specificity in this system.     

Increased Bean (Phaseolus vulgaris L.) Nodulation Competitiveness of Genetically Modified Rhizobium Strains

Rhizobium leguminosarum bv. phaseoli strain collections harbor heterogeneous groups of bacteria in which two main types of strains may be distinguished, differing both in the symbiotic plasmid and in the chromosome. We have analyzed under laboratory conditions the competitive abilities of the different types of Rhizobium strains capable of nodulating Phaseolus vulgaris L. bean. R. leguminosarum bv. phaseoli type I strains (characterized by nif gene reiterations and a narrow host range) are more competitive than type II strains (that have a broad host range), and both types are more competitive than the promiscuous rhizobia isolated from other tropical legumes able to nodulate beans. Type I strains become even more competitive by the transfer of a non-Sym, 225-kilobase plasmid from type II strain CFN299. This plasmid has been previously shown to enhance the nodulation and nitrogen fixation capabilities of Agrobacterium tumefaciens transconjugants carrying the Sym plasmid of strain CFN299. Other type I R. leguminosarum bv. phaseoli transconjugants carrying two symbiotic plasmids (type I and type II) have been constructed. These strains have a diminished competitive ability. The increase of competitiveness obtained in some transconjugants seems to be a transient property.     

Genomic insight into the taxonomy of Rhizobium genospecies that nodulate Phaseolus vulgaris

Due to the wide cultivation of bean (Phaseolus vulgaris L.), rhizobia associated with this plant have been isolated from many different geographical regions. In order to investigate the species diversity of bean rhizobia, comparative genome sequence analysis was performed in the present study for 69 Rhizobium strains mainly isolated from root nodules of bean and clover (Trifolium spp.). Based on genome average nucleotide identity, digital DNA:DNA hybridization, and phylogenetic analysis of 1,458 single-copy core genes, these strains were classified into 28 clusters, consistent with their species definition based on multilocus sequence analysis (MLSA) of atpD, glnII, and recA. The bean rhizobia were found in 16 defined species and nine putative novel species; in addition, 35 strains previously described as Rhizobium etli, Rhizobium phaseoli, Rhizobium vallis, Rhizobium gallicum, Rhizobium leguminosarum and Rhizobium spp. should be renamed. The phylogenetic patterns of symbiotic genes nodC and nifH were highly host-specific and inconsistent with the genomic phylogeny. Multiple symbiovars (sv.) within the Rhizobium species were found as a common feature: sv. phaseoli, sv. trifolii and sv. viciae in Rhizobium anhuiense; sv. phaseoli and sv. mimosae in Rhizobium sophoriradicis/R. etli/Rhizobium sp. III; sv. phaseoli and sv. trifolii in Rhizobium hidalgonense/Rhizobium acidisoli; sv. phaseoli and sv. viciae in R. leguminosarum/Rhizobium sp. IX; sv. trifolii and sv. viciae in Rhizobium laguerreae. Thus, genomic comparison revealed great species diversity in bean rhizobia, corrected the species definition of some previously misnamed strains, and demonstrated the MLSA a valuable and simple method for defining Rhizobium species.     

Inoculation with indigenous <Emphasis Type="Italic">rhizobium</Emphasis> strains increases yields of common bean (<Emphasis Type="Italic">Phaseolus vulgaris</Emphasis> L.) in northern Spain,although its efficiency is affected by the tillage system

Common bean (Phaseolus vulgaris L.) crops hold the potential to obtain higher yields by enhancing their biological nitrogen fixation (BNF) with Rhizobium. However in contrast to other legumes, common bean has shown a lack of positive response to inoculation with Rhizobium in many cases. This has led to a limited use of rhizobial inoculants in this crop, especially in Europe. The adaptation of bacterial strains to the rhizosphere is a key factor in the success of any inoculant, especially in a promiscuous legume such as common bean. This research aimed at increasing common bean yields via inoculation with effective indigenous Rhizobium leguminosarum strains. Three highly effective strains (LCS0306, LBM1123 and ZBM1008) which were selected according to their effectiveness at BNF in hydroponic conditions were separately inoculated onto common bean in a field experiment. The experiment was carried out under three environments and three tillage systems: conventional-tillage (CONVT), no-tillage (NT) and a cover-crop (CC). The grain yield observed with seed inoculation was significantly higher than the yield obtained with uninoculated seed under CONVT and CC. However, under NT inoculation had no effect. Furthermore, under CONVT and CC, inoculation with R. leguminosarum LCS0306 produced even higher yields than those obtained in nitrogen-fertilised or control plots. This is the first attempt to explain the inoculation performance of common bean under different tillage systems in Europe. A gene–based hypothesis has been used to explain the effectiveness of indigenous common bean rhizobia as nitrogen fixers in this crop.     

The lectin binding sites on the plasma membrane components of human lymphoblastoid cell lines.

The plasma membrane components of five human B-cell lines and three human T-cell lines were separated by dodecyl sulfate polyacrylamide gel electrophoresis, incubated with the radioactive labeled lectins from lentil, castor bean, wheat germ, Phaseolus bean, peanut, gorse and the Roman snail and the molecular weights of the binding sites determined. The lentil, castor bean and wheat germ lectin bound to multiple components from molecular weights (Mr) 20 000 to 200 000 within the plasma membranes, whereas peanut lectin bound preferentially to glycoproteins of Mr 150 000 and 83 000 in B-cells, and 150 000 and 130 000 in T-cells. The gorse lectin bound to a 220 000 component in B-cells which was not labeled in T-cells.     

Rhizobium sophorae is the dominant rhizobial symbiont of Vicia faba L. In North China

Faba bean (Vicia faba L.) is a major introduced grain-legume crop cultivated in China. In this study, rhizobia that nodulated faba bean grown in soils from three sites in North China (Hebei Province) were isolated and characterized. Firstly, isolates were categorized into genotypes by ribosomal IGS PCR-RFLP analysis, then representatives of the different IGS genotypes were further identified by phylogenetic analyses of 16S rRNA, housekeeping (atpD, recA) and nodulation (nodC) gene sequences. Rhizobial distribution based on the IGS genotype was related to the different soil physicochemical features by redundancy analysis. IGS typing and phylogenetic analyses of 16S rRNA and concatenated housekeeping gene sequences affiliated the 103 rhizobial strains isolated into four Rhizobium species/genospecies. A total of 69 strains of 3 IGS types were assigned to R. sophorae, 20 isolates of 5 IGS types to R. changzhiense and 9 isolates of 3 IGS types to R. indicum. The representative strain of the five remaining isolates (1 IGS type) was clearly separated from all Rhizobium type strains and was most closely related to defined genospecies according to the recently described R. leguminosarum species complex. Rhizobium sophorae strains (67% of total isolates) were common in all sites and shared an identical nodC sequence typical of faba bean symbionts belonging to symbiovar viciae. In this first study of rhizobia nodulating faba bean in Hebei Province, China, R. sophorae was found to be the dominant symbiont in contrast to other countries.     

Effect of Seed Treatment with Rhizobium leguminosarum on Pythium Damping-off, Seedling Height, Root Nodulation, Root Biomass, Shoot Biomass, and Seed Yield of Pea and Lentil

Field experiments were conducted in 2004 and 2005 to determine the effects of seed treatment with Rhizobium leguminosarum bv. viceae on damping‐off, seedling height, root nodule mass, root biomass, shoot biomass and seed yield of pea and lentil in a field naturally infested with Pythium spp. Compared with the untreated controls, treatment of pea seeds with R. leguminosarum bv. viceae strains R12, R20 or R21 significantly (P < 0.05) reduced incidence of damping‐off, promoted seedling growth and increased root nodule mass, root biomass and shoot biomass. Seed treatments with R12 or R21 also resulted in a significant (P < 0.05) increase in seed yield of pea. The strain R21 was most effective among the four strains of R. leguminosarum bv. viceae tested in peas. Although, the level of disease control by strain R21 was similar to seed treatment with the fungicide Thiram^TM, R21 was more effective in enhancing root nodule production and promoting plant growth. For lentil, treatment of seeds with R. leguminosarum bv. viceae strains R12 or R21 significantly (P < 0.05) reduced incidence of damping‐off compared with the untreated control. All of the four strains of R. leguminosarum bv. viceae tested increased lentil seedling height, root nodule mass and shoot biomass, and all except R20 increased root biomass. Seed yield was higher for the treatments of R12 and R21. The strain R12 was most effective among the four strains of R. leguminosarum bv. viceae tested in lentil. Although, strain R12 was as effective as Thiram^TM for control of damping‐off of lentil, it was more effective than Thiram^TM for the production of root nodules and promotion of plant growth. The study concludes that seed treatment with R. leguminosarum bv. viceae is effective in control of Pythium damping‐off of pea and lentil and that the efficacy of control is strain specific, strain R21 for control of the disease on pea and strain R12 for control of the disease on lentil.     

Rhizobium pisi sv. trifolii K3.22 harboring nod genes of the Rhizobium leguminosarum sv. trifolii cluster

The taxonomic status of the Rhizobium sp. K3.22 clover nodule isolate was studied by multilocus sequence analysis (MLSA) of 16S rRNA and six housekeeping chromosomal genes, as well as by a subsequent phylogenic analysis. The results revealed full congruence with the Rhizobium pisi DSM 30132^T core genes, thus supporting the same taxonomic position for both strains. However, the K3.22 plasmid symbiosis nod genes demonstrated high sequence similarity to Rhizobium leguminosarum sv. trifolii, whereas the R. pisi DSM 30132^Tnod genes were most similar to R. leguminosarum sv. viciae. The strains differed in the host range nodulation specificity, since strain K3.22 effectively nodulated red and white clover but not vetch, in contrast to R. pisi DSM 30132^T, which effectively nodulated vetch but was not able to nodulate clover. Both strains had the ability to form nodules on pea and bean but they differed in bean cultivar specificity. The R. pisi K3.22 and DSM 30132^T strains might provide evidence for the transfer of R. leguminosarum sv. trifolii and sv. viciae symbiotic plasmids occurring in natural soil populations.     

Phylogenetic multilocus sequence analysis of native rhizobia nodulating faba bean (Vicia faba L.) in Egypt

The taxonomic diversity of forty-two Rhizobium strains, isolated from nodules of faba bean grown in Egypt, was studied using 16S rRNA sequencing, multilocus sequence analyses (MLSA) of three chromosomal housekeeping loci and one nodulation gene (nodA). Based on the 16S rRNA gene sequences, most of the strains were related to Rhizobium leguminosarum, Rhizobium etli, and Rhizobium radiobacter (syn. Agrobacterium tumefaciens). A maximum likelihood (ML) tree built from the concatenated sequences of housekeeping proteins encoded by glnA, gyrB and recA, revealed the existence of three distinct genospecies (I, II and III) affiliated to the defined species within the genus Rhizobium/Agrobacterium. Seventeen strains in genospecies I could be classified as R. leguminosarum sv. viciae. Whereas, a single strain of genospecies II was linked to R. etli. Interestingly, twenty-four strains of genospecies III were identified as A. tumefaciens. Strains of R. etli and A. tumefaciens have been shown to harbor the nodA gene and formed effective symbioses with faba bean plants in Leonard jar assemblies. In the nodA tree, strains belonging to the putative genospecies were closely related to each other and were clustered tightly to R. leguminosarum sv. viciae, supporting the hypothesis that symbiotic and core genome of the species have different evolutionary histories and indicative of horizontal gene transfer among these rhizobia.     

A DNA element recognised by the molybdenum-responsive transcription factor ModE is conserved in Proteobacteria, green sulphur bacteria and Archaea

Background

In general, chemotaxis in Rhizobium has not been well characterized. Methyl accepting chemotaxis proteins are sensory proteins important in chemotaxis of numerous bacteria, but their involvement in Rhizobium chemotaxis is unclear and merits further investigation.

Results

A putative methyl accepting chemotaxis protein gene (mcpG) of Rhizobium leguminosarum VF39SM was isolated and characterized. The gene was found to reside on the nodulation plasmid, pRleVF39d. The predicted mcpG ORF displayed motifs common to known methyl-accepting chemotaxis proteins, such as two transmembrane domains and high homology to the conserved methylation and signaling domains of well-characterized MCPs. Phenotypic analysis of mcpG mutants using swarm plates did not identify ligands for this putative receptor. Additionally, gene knockouts of mcpG did not affect a mutant strain's ability to compete for nodulation with the wild type. Notably, mcpG was found to be plasmid-encoded in all strains of R. leguminosarum and R. etli examined, though it was found on the nodulation plasmid only in a minority of strains.

Conclusions

Based on sequence homology R. leguminosarum mcpG gene codes for a methyl accepting chemotaxis protein. The gene is plasmid localized in numerous Rhizobium spp. Although localized to the sym plasmid of VF39SM mcpG does not appear to participate in early nodulation events. A ligand for McpG remains to be found. Apparent McpG orthologs appear in a diverse range of proteobacteria. Identification and characterization of mcpG adds to the family of mcp genes already identified in this organism.     

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 = axⁿ 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.     

Interaction of pea (Pisum sativum L.) lectins with rhizobial strains

Lectins from two varieties (PG-3 and LFP-48) of pea have been purified by affinity chromatography on Sephadex G-50. The specific activity increased by 23 and 25 folds, respectively. These lectins from both the varieties were found to be specific for mannose. The purified fluorescein isothiocyanate (FITC) – labelled lectins showed binding reaction with homologous as well as heterologous strains of Rhizobium spp. The results revealed that pea lectins are not highly specific to their respective rhizobia. Moreover, these lectins showed a greater stimulatory effect on homologous Rhizobium leguminosarum strains.     

Synthesis of 1-N-glycyl beta-oligosaccharide derivatives. Reactivity of Lens culinaris lectin with a fluorescent labeled streptavidin pseudoglycoprotein and immobilized neoglycolipid.

The lectin from Lens culinaris (lentil) has a binding specificity for glycopeptides bearing 6-O-linked fucose on the reducing terminus on complex-type N-linked oligosaccharides. Lentil lectin therefore provides an excellent example of a carbohydrate binding protein in which high-affinity interactions are dependent on the integrity of the oligosaccharide core structure. We report here the synthesis of the 1-N-glycyl beta-derivative of Gal beta 4GlcNAc beta 2Man alpha 6(Gal beta 4GlcNAc beta 2Man alpha 3)Man beta 4GlcNAc beta 4(Fuc alpha 6)-GlcNAc (Gal-2F) and its subsequent biotinylation and palmitoylation. The biotin derivative when bound to a streptavidin-fluorescein isothiocyanate (FITC) conjugate was able to bind to both concanavalin A (ConA) and lentil lectin affinity columns. In contrast, synthesis of the biotin derivative of the glycamine derivative of Gal-2F and subsequent binding to streptavidin-FITC afforded reactivity to a ConA affinity column but not to a lentil lectin affinity column. Lentil lectin also bound to plastic microtiter plates containing the adsorbed palmitoyl-1-N-glycyl beta-derivative. No binding occurred when the homologous glycamine neoglycolipid was used. These results suggest the 1-N-glycyl beta-derivative of oligosaccharides may have general utility as an intermediate in the synthesis of novel glycoconjugate probes.     

Congo Red Absorption by Rhizobium leguminosarum

Congo red absorption is generally considered a contraindication of Rhizobium. However, R. leguminosarum takes up the dye on yeast extract-mannitol agar. The uptake of congo red varies among strains of R. leguminosarum, as shown elsewhere with strains of R. trifolii and R. meliloti. Congo red absorption does not distinguish rhizobia from other bacteria, but may be useful as a strain marker.     

Identification and Assessment of Symbiotic Effectiveness of Phage-Typed <Emphasis Type="Italic">Rhizobium leguminosarum</Emphasis> Strains on Lentil (<Emphasis Type="Italic">Lens culinaris</Emphasis> Medik) Cultivars

Symbiotic effectiveness of 19 indigenous and two exotic (USDA 2426 and USDA 2431) strains of lentil Rhizobium belonging to different phage-sensitive and phage-resistant groups was compared under axenic condition. Four strains (USDA 2431, BHULR 104, BHULR 113, and BHULR 115) sensitive to different phages were found significantly superior over others in terms of nodule number, acetylene reduction activity, and total dry weight per plant. Inoculation response of these strains was then evaluated on six lentil cultivars under field condition. A significant symbiotic interaction between rhizobial strains and lentil cultivars was observed. Grain yield enhancement was noticed by the compatible interaction of lentil cultivars HUL-57, L-4147, K-75, and PL-4/DPL-15/DPL-62 with rhizobial strains USDA 2431, BHULR 104, BHULR 113, and BHULR 115, respectively. The authentication of rhizobial strains was accomplished through 16S rDNA sequence analysis. All rhizobial strains had close matching with R. leguminosarum bv. viciae strains. The results have shown that phages can trustfully help selecting out the symbiotically efficient most rhizobial strains for advantageous use with lentil cultivars, in order to strengthen the BNF-based future lentil breeding programs.     

Host-Symbiont Interactions: III. Purification and Partial Characterization of Rhizobium Lipopolysaccharides

The lipopolysaccharides of three strains each of Rhizobium leguminosarum, R. phaseoli, and trifolii have been purified and partially characterized. The last step in the purification procedure is gel filtration column chromatography using Sepharose 4B with an elution buffer consisting of ethylenediaminetetraacetic acid and triethylamine. Each of the lipopolysaccharides reported in this paper elutes as a symmetrical peak in the partially included volume of this Sepharose 4B column. The ratio of 2-keto-3-deoxyoctonate acid (a sugar which is characteristic of lipopolysaccharides) to hexose is constant throughout the carbohydrate-containing peaks as they elute from the Sepharose 4B. The compositions and immunodominant structures of the purified lipopolysaccharides vary as much among strains of a single Rhizobium species as among the different species of Rhizobium. There is no obvious correlation between the nodulation group to which a Rhizobium belongs and the chemical composition or immunochemistry of the Rhizobium's lipopolysaccharide. There is extensive crosslysis by phage of strains of R. trifolii, R. phaseoli, and R. leguminosarum. This suggests that the receptors for these cross-lysing phage reside either in nonlipopolysaccharide structures or in common structures within the lipopolysaccharide which are not detected by compositional or immunochemical analysis.