Development of Microbiome Biobanks - Challenges and Opportunities

The microbiome research field is rapidly evolving, but the required biobanking infrastructure is currently fragmented and not prepared for the biobanking of microbiomes. The rapid advancement of technologies requires an urgent assessment of how biobanks can underpin research by preserving microbiome samples and their functional potential.

Open resource metagenomics: a model for sharing metagenomic libraries

Both sequence-based and activity-based exploitation of environmental DNA have provided unprecedented access to the genomic content of cultivated and uncultivated microorganisms. Although researchers deposit microbial strains in culture collections and DNA sequences in databases, activity-based metagenomic studies typically only publish sequences from the hits retrieved from specific screens. Physical metagenomic libraries, conceptually similar to entire sequence datasets, are usually not straightforward to obtain by interested parties subsequent to publication. In order to facilitate unrestricted distribution of metagenomic libraries, we propose the adoption of open resource metagenomics, in line with the trend towards open access publishing, and similar to culture- and mutant-strain collections that have been the backbone of traditional microbiology and microbial genetics. The concept of open resource metagenomics includes preparation of physical DNA libraries, preferably in versatile vectors that facilitate screening in a diversity of host organisms, and pooling of clones so that single aliquots containing complete libraries can be easily distributed upon request. Database deposition of associated metadata and sequence data for each library provides researchers with information to select the most appropriate libraries for further research projects. As a starting point, we have established the Canadian MetaMicroBiome Library (CM(2)BL [1]). The CM(2)BL is a publicly accessible collection of cosmid libraries containing environmental DNA from soils collected from across Canada, spanning multiple biomes. The libraries were constructed such that the cloned DNA can be easily transferred to Gateway® compliant vectors, facilitating functional screening in virtually any surrogate microbial host for which there are available plasmid vectors. The libraries, which we are placing in the public domain, will be distributed upon request without restriction to members of both the academic research community and industry. This article invites the scientific community to adopt this philosophy of open resource metagenomics to extend the utility of functional metagenomics beyond initial publication, circumventing the need to start from scratch with each new research project.

Comparison of the symbiotic and competition phenotypes of Sinorhizobium meliloti PHB synthesis and degradation pathway mutants

The competitive abilities of Sinorhizobium meliloti mutant strains containing lesions in the PHB synthesis (phbC) and degradation (bdhA) pathways were compared. While the bdhA mutant showed no noticeable symbiotic defects on alfalfa host plants when inoculated alone, in mixed inoculation experiments it was found to be less competitive than the wild type for nodule occupancy. Long-term survival of the bdhA mutant on a carbon-limiting medium was not affected. However, when subjected to competition with the wild-type strain in periodic subculturing through alternating carbon-limiting and carbon-excess conditions, the bdhA mutant performed poorly. A more severe defect in competition for growth and nodule occupancy was observed with a mutant unable to synthesize PHB (phbC). These results indicate that the ability to efficiently deposit cellular PHB stores is a key factor influencing competitive survival under conditions of fluctuating nutrient carbon availability, whereas the ability to use these stores is less important.

A novel bacteriocin, thuricin 17, produced by plant growth promoting rhizobacteria strain Bacillus thuringiensis NEB17: isolation and classification

AIMS

The aim of this study was to identify and characterize a compound produced by the plant growth promoting bacterium, Bacillus thuringiensis non-Bradyrhizobium Endophytic Bacterium 17.

METHODS AND RESULTS

The bacterial peptide was analysed and purified via HPLC. Using the disk diffusion assay this peptide inhibited the growth of 16/19 B. thuringiensis strains, 4/4 Bacillus cereus strains, among others, as well as a Gram-negative strain Escherichia coli MM294 (pBS42). Both bactericidal and bacteristatic effects were observed on B. cereus ATCC 14579 and bactericidal effects were observed on B. thuringiensis ssp. thuringiensis Bt1267. The molecular weight of the peptide was estimated via SDS-PAGE and confirmed with Matrix Assisted Laser Desorption Ionization Quadrapole Time of Flight mass spectrometry; its weight is 3162 Da. The peptide is biologically active after exposure to 100 degrees C for 15 min, and within the pH range 1.00-9.25. Its activity disappeared when treated with proteinase K and protease, but not with alpha-amylase or catalase.

CONCLUSIONS

We conclude that this is the first report of a bacteriocin produced by a plant growth promoting rhizobacteria (B. thuringiensis) species and have named the bacteriocin thuricin 17.

SIGNIFICANCE AND IMPACT OF THE STUDY

Our work has characterized a bacteriocin produced by a plant growth promoting bacterium. This strain is previously reported to increase soya bean nodulation.

Bradyrhizobium japonicum mutants with enhanced sensitivity to genistein resulting in altered nod gene regulation

Bradyrhizobium japonicum mutants with altered nod gene induction characteristics were isolated by screening mutants for genistein-independent nod gene expression. Plasmid pZB32, carrying a nodY::lacZ transcriptional gene fusion, was introduced into B. japonicum cells that had been subjected to UV mutagenesis. Ten independent transformants producing a blue color on plates containing 5bromo-4chloro-3indolyl-beta-D-galactopyranoside but lacking genistein, indicative of constitutive expression of the nodY::lacZ reporter gene, were isolated. Beta-galactosidase activity assays revealed that while all of the 10 strains were sensitive to low concentrations of genistein, none exhibited truly constitutive nodY::lacZ expression in liquid culture. Soybean plants inoculated with three of the mutants were chlorotic and stunted, with shoot dry weights close to those of the uninoculated plants, indicating the absence of nitrogen fixation. Differences in the kinetics of nodY::lacZ expression and lipochitin oligosaccharide Nod signal production suggested that the strains carried different mutations. Some of these strains may be useful in mitigating the low root zone temperature-associated delay in soybean nodulation at the northern extent of soybean cultivation.

The chvH locus of Agrobacterium encodes a homologue of an elongation factor involved in protein synthesis

The virulence of Agrobacterium tumefaciens depends on both chromosome- and Ti plasmid-encoded gene products. In this study, we characterize a chromosomal locus, chvH, previously identified by TnphoA mutagenesis and shown to be required for tumor formation. Through DNA sequencing and comparison of the sequence with identified sequences in the database, we show that this locus encodes a protein similar in sequence to elongation factor P, a protein thought to be involved in peptide bond synthesis in Escherichia coli. The analysis of vir-lacZ and vir-phoA translational fusions as well as Western immunoblotting revealed that the expression of Vir proteins such as VirE2 was significantly reduced in the chvH mutant compared with the wild-type strain. The E. coli efp gene complemented detergent sensitivity, virulence, and expression of VirE2 in the chvH mutant, suggesting that chvH and efp are functionally homologous. As expected, ChvH exerts its activity at the posttranscriptional level. Southern analysis suggests that the gene encoding this elongation factor is present as a single copy in A. tumefaciens. We constructed a chvH deletion mutant in which a 445-bp fragment within its coding sequence was deleted and replaced with an omega fragment. On complex medium, this mutant grew more slowly than the wild-type strain, indicating that elongation factor P is important but not essential for the growth of Agrobacterium.

An ACC deaminase minus mutant of Enterobacter cloacae UW4 no longer promotes root elongation

The ACC deaminase gene (acdS) from Enterobacter cloacae UW4 was replaced by homologous recombination with the acdS gene with a tetracycline resistance gene inserted within the coding region. Upon characterization of this AcdS minus mutant, it was determined that both ACC deaminase activity and the ability to promote the elongation of canola roots under gnotobiotic conditions were greatly diminished. This result is consistent with a previously postulated model that suggests that a major mechanism utilized by plant growth-promoting bacteria involves the lowering of plant ethylene levels, and hence ethylene inhibition of root elongation, by bacterial ACC deaminase.

Requirement for the enzymes acetoacetyl coenzyme A synthetase and poly-3-hydroxybutyrate (PHB) synthase for growth of Sinorhizobium meliloti on PHB cycle intermediates

We have identified two Sinorhizobium meliloti chromosomal loci affecting the poly-3-hydroxybutyrate degradation pathway. One locus was identified as the gene acsA, encoding acetoacetyl coenzyme A (acetoacetyl-CoA) synthetase. Analysis of the acsA nucleotide sequence revealed that this gene encodes a putative protein with a molecular weight of 72,000 that shows similarity to acetyl-CoA synthetase in other organisms. Acetyl-CoA synthetase activity was not affected in cell extracts of glucose-grown acsA::Tn5 mutants; instead, acetoacetyl-CoA synthetase activity was drastically reduced. These findings suggest that acetoacetyl-CoA synthetase, rather than CoA transferase, activates acetoacetate to acetoacetyl-CoA in the S. meliloti poly-3-hydroxybutyrate cycle. The second locus was identified as phbC, encoding poly-3-hydroxybutyrate synthase, and was found to be required for synthesis of poly-3-hydroxybutyrate deposits.

Sinorhizobium meliloti strain 1021 bioS and bdhA gene transcriptions are both affected by biotin available in defined medium

Sinorhizobium meliloti 1021 responds to external biotin signals from alfalfa plants through the bioS regulatory locus. Immunogold labeling and electron microscopy revealed that the BioS protein is located within the S. meliloti cytoplasm. Under biotin-limiting conditions the S. meliloti cell lumen was filled with polyhydroxybutyrate (PHB) granules suggesting that either PHB synthesis or degradation are influenced by biotin. To test this hypothesis a 3-hydroxybutyrate-dehydrogenase-lacZ (bdhA-lacZ) fusion was mobilized into S. meliloti. beta-galactosidase tests revealed an overall 3.6-5.2-fold higher bdhA transcription in the presence of added biotin. Comparison of the bdhA and the bioS promoter regions identified several common motifs.

Poly-3-hydroxybutyrate degradation in Rhizobium (Sinorhizobium) meliloti: isolation and characterization of a gene encoding 3-hydroxybutyrate dehydrogenase

We have cloned and sequenced the 3-hydroxybutyrate dehydrogenase-encoding gene (bdhA) from Rhizobium (Sinorhizobium) meliloti. The gene has an open reading frame of 777 bp that encodes a polypeptide of 258 amino acid residues (molecular weight 27,177, pI 6.07). The R. meliloti Bdh protein exhibits features common to members of the short-chain alcohol dehydrogenase superfamily. bdhA is the first gene transcribed in an operon that also includes xdhA, encoding xanthine oxidase/dehydrogenase. Transcriptional start site analysis by primer extension identified two transcription starts. S1, a minor start site, was located 46 to 47 nucleotides upstream of the predicted ATG start codon, while S2, the major start site, was mapped 148 nucleotides from the start codon. Analysis of the sequence immediately upstream of either S1 or S2 failed to reveal the presence of any known consensus promoter sequences. Although a sigma54 consensus sequence was identified in the region between S1 and S2, a corresponding transcript was not detected, and a rpoN mutant of R. meliloti was able to utilize 3-hydroxybutyrate as a sole carbon source. The R. meliloti bdhA gene is able to confer upon Escherichia coli the ability to utilize 3-hydroxybutyrate as a sole carbon source. An R. meliloti bdhA mutant accumulates poly-3-hydroxybutyrate to the same extent as the wild type and shows no symbiotic defects. Studies with a strain carrying a lacZ transcriptional fusion to bdhA demonstrated that gene expression is growth phase associated.

Methylmalonyl-CoA mutase encoding gene of Sinorhizobium meliloti

A cluster of genes on megaplasmid pRmeSU47b, bhbA-D, is required for growth on the polyhydroxyalkanoate degradation pathway intermediates 3-hydroxybutyrate and acetoacetate as sole carbon source. DNA sequence analysis of the bhbA gene indicated that it encoded a protein of 712 amino acids (aa) (78kDa) which appeared to be a homodimeric methylmalonyl-CoA mutase enzyme (EC 5.4.99.2). Cell-free extract of a bhbA::Tn5 mutant was devoid of methylmalonyl-CoA mutase activity, thus confirming the identity of the bhbA-encoded enzyme. The reason for the requirement of methylmalonyl-CoA mutase activity for operation of the polyhydroxyalkanoate degradation pathway is not immediately apparent. Situated immediately upstream of bhbA, in the same orientation, is a gene which is predicted to encode a protein that exhibits remarkable sequence similarity to the alpha subunit of propionyl-CoA carboxylase (EC 6.4.1.3). A mutation in this gene did not affect ability to grow on 3-hydroxybutyrate as sole carbon source. Downstream of, and oriented towards bhbA, was identified a member of the GNTR class of transcriptional regulator-encoding genes. It is not yet known whether this regulatory protein is directly involved in modulation of bhbA expression.

Cloning and sequencing of a gene encoding cellobiose dehydrogenase from Trametes versicolor

Cellobiose dehydrogenase (CDH) is an enzyme produced under lignocellulose-degrading conditions by Trametes versicolor strain 52J (Tv) and several other wood-degrading fungi, including Phanerochaete chrysosporium (Pc). In order to understand better the nature and properties of this enzyme, we isolated a genomic clone of Tv cdh using heterologous probes derived from the sequence of Pc cdh. DNA sequence analysis revealed that Tv cdh consists of 3091 bp of coding sequence interrupted by 14 introns. Southern blotting showed that the gene was present in a single copy in the strain of Tv analyzed. Tv cdh was shown by Northern blot analysis to be expressed as a single transcript under cellulolytic conditions. RT-PCR of poly(A)+ RNA isolated under cellulolytic conditions was used to generate a near full-length cDNA copy of the cdh mRNA. The deduced protein encoded by Tv cdh consists of 768 amino acids (aa), including a predicted 19 aa signal peptide. The protein had 73% identity to the corresponding protein from Pc, which is the only other CDH-encoding gene that has been cloned. Based upon its deduced primary structure and alignment to similar sequences, Tv CDH shares a general structural organization with Pc CDH and other hemoflavoenzymes. Amino acid residues H-109 and M-61 in the N-terminal heme domain are hypothesized to function in heme binding; the C-terminal flavin domain contained a consensus sequence for flavin binding between residues 217-222. Although the protein is known to bind to cellulose, no obvious homology to bacterial or fungal cellulose binding domains was observed. However, a strong homology was observed to a region of Pc CDH that is hypothesized to be involved in cellulose binding.

Megaplasmid and chromosomal loci for the PHB degradation pathway in Rhizobium (Sinorhizobium) meliloti

Chromosomal and megaplasmid loci that affect the poly-3-hydroxybutyrate (PHB) degradation pathway in Rhizobium meliloti were identified. A clone that restores the ability of certain R. meliloti mutants with defined deletions in megaplasmid pRmeSU47b to use 3-hydroxybutyrate or acetoacetate as the sole carbon source was isolated from a cosmid library of R. meliloti genomic DNA. Tn5 insertion mutagenesis, followed by merodiploid complementation analysis, demonstrated that the locus consists of at least four transcriptional units, bhbA-D. We also identified loci involved in 3-hydroxybutyrate and/or acetoacetate utilization by screening for mutants that had lost the ability to use 3-hydroxybutyrate as the sole carbon source while retaining the ability to use acetate (thus ensuring an intact glyoxylate cycle and gluconeogenic pathway). These mutants fell into four classes, as determined by replicon mobilization experiments and genetic linkage in phage transduction; one class corresponded to the bhb locus on pRmeSU47b, two classes mapped to different regions on the chromosome and the fourth, bdhA, represented by a single mutant, mapped to another pRmeSU47b locus, near bacA. The bdhA mutant is deficient in 3-hydroxybutrate dehydrogenase activity.

Discrete regions of the sensor protein virA determine the strain-specific ability of Agrobacterium to agroinfect maize

The ability of Agrobacterium strains to infect transformation-recalcitrant maize plants has been shown to be determined mainly by the virA locus, implicating vir gene induction as the major factor influencing maize infection. In this report, we further explore the roles of vir induction-associated bacterial factors in maize infection using the technique of agroinfection. The Ti plasmid and virA source are shown to be important in determining the ability of a strain to infect maize, and the monosaccharide binding protein ChvE is absolutely required for maize agroinfection. The linker domain of VirAC58 from an agroinfection-competent strain, C58, is sufficient to convert VirAA6 of a nonagroinfecting strain, A348,to agroinfection competence. The periplasmic domain of VirAC58 is also able to confer a moderate level of agroinfection competence to VirAA6. In addition, the VirAA6 protein from A348 is agroinfection competent when removed from its cognate Ti plasmid background and placed in a pTiC58 background. The presence of a pTiA6-encoded, VirAA6-specific inhibitor is hypothesized and examined.

Construction of Agrobacterium strains by electroporation of genomic DNA and its utility in analysis of chromosomal virulence mutations

We have extended the technique of electroporation as a genetic tool for manipulating the Agrobacterium tumefaciens chromosome. We used this technique to introduce chromosomal DNA into recipient A. tumefaciens strains by electroporation and constructed isogenic chvE mutants that share the same chromosomal background but differ in their types of pTi (octopine or nopaline). Both nopaline and octopine pTi-carrying chvE mutants were deficient in vir regulon induction and exhibited similar reductions in host range.

Second site mutations specifically suppress the Fix- phenotype of Rhizobium meliloti ndvF mutations on alfalfa: identification of a conditional ndvF-dependent mucoid colony phenotype

Rhizobium meliloti mutants carrying ndvF insertion or deletion mutations induce nodules on alfalfa which contain very few infected cells and fail to fix N2 (Fix-). We have characterized five independent second site mutations (designated sfx) which completely suppress the Fix- phenotype of ndvF mutants on Medicago sativa but not on another R. meliloti host Melilotus alba. Genetic mapping and phenotypic analysis revealed that the suppressor mutations sfx-1, sfx-4 and sfx-5 mapped to a single locus which was distinct from another locus defined by the sfx-2 and sfx-3 mutations. Tn5-mob-mediated conjugal mapping experiments showed that the sfx-1 locus was located clockwise from trp-33 on the R. meliloti chromosome and a detailed cotransduction map of this region was generated. To clone the sfx-1 locus, we prepared a cosmid library from total DNA obtained from an sfx-1, ndvF deletion strain. From this library, a cosmid pTH56, which converted Fix- ndvF mutants to Fix+, was isolated. Southern blot analysis provided direct physical evidence that the insert DNA in plasmid pTH56 was contiguous with the sfx-1 region. On low osmolarity glutamate-yeast extract-mannitol-salts medium (GYM) agar medium, ndvF insertion and deletion mutants were found to have a mucoid colony phenotype, as opposed to the dry colony phenotype of the wild-type strain. This phenotype was shown to be dependent on the exoB and expE genes required for synthesis of exopolysaccharide II in R. meliloti but not to be dependent on genes required exclusively for the synthesis of the succinoglycan or exopolysaccharide I. Transduction of either sfx-1 or sfx-2 or transfer of the cosmid pTH56 into the ndvF mutants restored them to a wild-type dry colony phenotype. The mucoid phenotype is not responsible for the Fix- phenotype of ndvF mutants as the Fix-, ndvF exp double mutants can be complemented to Fix+ by introducing plasmids which carry only the wild-type ndvF genes.

A chromosomally encoded two-component sensory transduction system is required for virulence of Agrobacterium tumefaciens

TnphoA mutagenesis of Agrobacterium tumefaciens identified new extracytoplasmic protein-encoding virulence loci. Mutations in these loci conferred increased sensitivity to detergents and several antibiotics. Clones carrying these loci were isolated from an A. tumefaciens cosmid library by complementation of the detergent sensitivities of the mutants. The locus on one complementing clone was delineated by Tn5 and TnphoA mutagenesis. DNA sequence analysis of the delineated region revealed that this locus is made up of two transcriptional units, chvG and chvI, which were predicted, on the basis of amino acid sequence homology, to encode the members of a two-component sensory transduction system. The membrane-spanning sensor, a histidine protein kinase, was designated ChvG, and the response regulator, presumably a transcriptional activator, was designated ChvI. Surprisingly, ChvG was also predicted to contain a Walker type A consensus nucleotide binding site, which is unusual for sensor histidine protein kinases. Site-specific insertion mutations in either chvG or chvI abolished tumor formation ability, as well as the ability to grow on complex media. Neither the genes which are regulated nor the inducing signal is known yet for this system.

The ntrA gene of Agrobacterium tumefaciens: identification, cloning, and phenotype of a site-directed mutant

A 3.6-kb EcoRI fragment containing the ntrA gene of Agrobacterium tumefaciens was cloned by using the homologous ntrA gene of Rhizobium meliloti as a probe. Construction of an ntrA mutant of A. tumefaciens by site-directed insertional mutagenesis demonstrated the requirement of the ntrA gene for nitrate utilization and C4-dicarboxylate transport but not for vir gene expression or tumorigenesis.

ndvF, a novel locus located on megaplasmid pRmeSU47b (pEXO) of Rhizobium meliloti, is required for normal nodule development

Rhizobium meliloti strains carrying either of two overlapping deletions (delta 5408 and delta F114) of the megaplasmid pRmeSU47b form nodules on alfalfa which fail to fix N2 (Fix-). Strains carrying these deletions also fail to fluoresce on media containing calcofluor, indicating a defect in synthesis of the acidic exopolysaccharide (Exo-) of R. meliloti. We have isolated cosmid clones (pTH21 and pTH22) which complement the Fix- but not the Exo- phenotype of the strains carrying the delta 5408 and delta F114 deletions. In addition, cosmid clones which complement the Exo- phenotype fail to complement the Fix- phenotype of these deletions; thus, the Exo- phenotype is not related to the Fix- phenotype. A 5-kb region within a 7.3-kb BamHI restriction fragment was found to be required for complementation of the Fix- phenotype of the delta 5408 and delta F114 deletion strains. Tn5 insertions in the 5-kb region generated a Fix- phenotype when recombined into the wild-type genome. We have designated this locus ndvF, for nodule development. TnphoA mutagenesis of this region generated active alkaline-phosphatase gene fusions, indicating that ndvF encodes extracytoplasmic protein(s). Induction of nodules by the ndvF mutants was delayed by 2 to 3 days compared with induction by the wild-type strain. Light microscopy of nodules elicited by strains carrying the large 150-kb delta F114 deletion, a 12-kb deletion removing ndvF, or an individual ndvF::Tn5 insertion mutation demonstrated that many nodules contained few infected cortical cells, indicating that nodule development was blocked early in the infection process, before the release of bacteria from the infection threads.

Analysis of a 1600-kilobase Rhizobium meliloti megaplasmid using defined deletions generated in vivo

A series of 120-600 kilobase deletions with defined endpoints were made in the 1600-kilobase Rhizobium meliloti megaplasmid pRmeSU47b, by homologous recombination between the IS50 elements of transposon insertions. Utilizing IS 50-mediated homologous recombination we also made defined reductions in deletion size and combined adjacent deletions. Deletion structure was confirmed by phage transduction and Southern hybridization analysis. Collectively these deletions span 1400 kilobases of pRmeSU47b, indicating that the majority of the plasmid is not essential for cell viability. This was further confirmed by the construction of a strain SU47 derivative which carries only 450 kilobases of the pRmeSU47b megaplasmid. Examination of the deletion mutants for phenotype revealed novel loci required for dulcitol, melibiose, raffinose, beta-hydroxybutyrate, acetoacetate, protocatechuate and quinate utilization. Previously unidentified loci required for effective root nodule development and exopolysaccharide synthesis were also found. Various deletion mutants were deficient in dicarboxylate transport, lactose utilization, and thiamine and exopolysaccharide biosynthesis, as predicted from earlier studies of this megaplasmid.

Genetic map of Rhizobium meliloti megaplasmid pRmeSU47b

A circular linkage map of the Rhizobium meliloti megaplasmid pRmeSU47b was constructed. The map consists of transposon insertions carrying alternating antibiotic resistance markers linked by phi M12 transduction. Data from conjugation experiments utilizing donor strains carrying Tn5-oriT insertions in the megaplasmid supported the proposed genetic map. In addition, the positions of previously identified Fix, exopolysaccharide synthetic, thiamine synthetic, and C4-dicarboxylate transport loci on the megaplasmid map were determined. By converting cotransduction frequencies to physical distance, we calculated the replicon to be 1,600 kilobases in size, which compares favorably with previous physical estimates.

Analysis of C4-dicarboxylate transport genes in Rhizobium meliloti

A 5.1 kbp DNA fragment was isolated which complemented C4-dicarboxylate transport mutants (dct) of Rhizobium meliloti. Characterization of this fragment by subcloning, transposon mutagenesis, and complementation analysis revealed three loci, designated dctA, dctB, and dctD. TnphoA-generated alkaline phosphatase fusions to dctA suggested that this gene encodes the structural transport protein and allowed the determination of its direction of transcription. Analysis of the fusions in various mutant backgrounds demonstrated that dctB, dctD, and ntrA products are required for dctA expression. The dctA fusion was constitutively expressed in a dctA mutant background, but was not expressed in dctA dctB or dctA dctD double mutants. This suggests that the constitutive expression in a dctA mutant background is mediated through dctB and dctD. Three independent second-site Dct+ revertant mutations in ntrA mutant strains mapped to the dct locus. Succinate transport in these revertant strains was constitutive, whereas in the wild type, succinate transport was inducible. These results are consistent with the direct requirement of the ntrA gene product for dctA expression. Alfalfa plants inoculated with the dctB and dctD mutants showed reduced nitrogen-fixing activity. Nodules induced by dctA mutants failed to fix nitrogen. These symbiotic phenotypes are consistent with previous suggestions that dctA expression in bacteroids can occur independently of dctB and dctD.