| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Biology, University of Regina, 3737 Wascana Parkway, Regina, SK S4S 0A2, Canada
2 Department of Biology, University of Calgary, 2500 University Drive, Calgary, AB T2N 1N4, Canada
Correspondence
Christopher K. Yost
chris.yost{at}uregina.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Present address: Department of Biology, York University, 4700 Keele St, Toronto, ON M3J 1P3, Canada.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Galibert et al., 2001). The occurrence of large plasmids is a common feature of these bacteria, and in the fast-growing rhizobia, including the genera Sinorhizobium and Rhizobium, plasmids can make up as much as 40 % of the total genome size (Hynes et al., 1989; Hynes & Finan, 1998; Galibert et al., 2001). Recent genome sequencing projects (Kaneko et al., 2000, 2002; Galibert et al., 2001; Wood et al., 2001; DelVecchio et al., 2002; Streit et al., 2004) have confirmed the variety and complexity of the genomes of the rhizobia and their closest relatives among the alphaproteobacteria, as well as revealing much about their biology. The genomes of Rhizobium leguminosarum and Rhizobium etli have also been completely sequenced and, once published, should allow even more insight into genomic organization in the rhizobia.
We are interested in the role that plasmid-encoded catabolic genes might play in influencing rhizosphere competition. Plasmids in the fast-growing rhizobia carry genes involved in nodulation, nitrogen fixation and a number of other processes, including catabolism of a number of sugars and amino acids, melanin synthesis, LPS synthesis and many others (Hynes et al., 1988; Hynes & McGregor, 1990; Baldani et al., 1992; Brom et al., 1992, 2000; Oresnik et al., 2000; Barnett et al., 2001; Finan et al., 2001). The presence of individual plasmids has been shown to affect the ability of strains to compete for nodulation of the appropriate host plant in several cases (Martínez-Romero & Rosenblueth, 1990; Brom et al., 1992; Moënne-Loccoz & Weaver, 1995; Brom et al., 2000), though the exact determinants that influence this competition have not been established. There is evidence that plasmid-mediated catabolism of proline (Jiménez-Zurdo et al., 1995), trigonelline (Boivin et al., 1991; Goldmann et al., 1991), calystegines (Guntli et al., 1999) and rhizopines (Heinrich et al., 1999) in Sinorhizobium meliloti plays a role in competition, and catabolism of mimosine by Rhizobium strain TAL1145 (Soedarjo & Borthakur, 1998) has been shown to confer a competitive advantage for nodulation of Leucaena spp. We have also established that rhamnose catabolism is important for competition for nodulation of clover by R. leguminosarum bv. trifolii (Oresnik et al., 1998). Genome sequences suggest that plasmids of the fast-growing rhizobia carry large numbers of genes encoding predicted transport systems and catabolic pathways (Barnett et al., 2001; Finan et al., 2001; Streit et al., 2004), so it is likely that many other loci important to the saprophytic competence and competitive ability of rhizobia remain to be discovered.
In this paper, we describe a plasmid-encoded erythritol utilization gene region which appears to be one of the determinants on R. leguminosarum bv. viciae strain VF39SM plasmid pRleVF39f that contribute to this plasmid's importance for competitive nodulation. The genes involved in uptake and catabolism of erythritol are identified for the first time in any of the rhizobia, and evidence is provided that these genes have been subject to lateral transfer between Brucella spp. and R. leguminosarum.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. R. leguminosarum strains were routinely cultured at 30 °C on TY medium (Beringer, 1974) or in HP broth (Hynes et al., 1985), or Vincent's minimal medium (VMM; Vincent, 1970) with carbon sources added at 10 mM final concentration. When succinate was used as a carbon source, R. leguminosarum strains were grown using GMS minimal medium and 5 mM succinate (Poole et al., 1994). Plasmid-cured derivatives of strain VF39SM (Hynes & McGregor, 1990) were tested for utilization of 96 carbon sources simultaneously using Biolog plates as described previously for S. meliloti (Oresnik et al., 2000). The manufacturer's protocol was used to conduct the tests. Escherichia coli strains were cultured on LB medium (Sambrook et al., 1989) at 37 °C. When necessary, Rhizobium strains were cultured in media containing antibiotics at the following concentrations (expressed in µg antibiotic ml–1): erythromycin (Em), 100; gentamicin (Gm), 30; neomycin (Nm), 100; spectinomycin (Sp), 100; streptomycin (Sm), 500; and tetracycline (Tc), 5. E. coli was cultured with the following antibiotic concentrations when required (expressed in µg antibiotic ml–1): ampicillin (Amp), 100; Gm, 15; kanamycin (Km), 50; Sp, 100; and Tc, 10. Antibiotics were obtained from Sigma-Aldrich.
|
) as described by Hynes et al. (1985) and modified by Hynes & McGregor (1990). Probe labelling, Southern blots and detection procedures were performed using the digoxigenin (DIG) labelling and detection system as specified by the manufacturer's instructions (Roche Biochemicals).
PCR cloning of the eryA gene and eryA promoter.
PCR amplifications were performed using a PCR kit from Qiagen and using Qiagen instructions. The locations of the primers designed for amplification of eryA and the promoter region are shown in Fig. 1. The sequences are: AR1, 5'-GCATTGAGCAGCACTTCA; AR2, 5'-GGCGTGAAGCGGTCATA; AR3, 5'-CGAAAAAGGGCAGCAAAT. In a 100 µl PCR reaction, 5 µl pCos42 DNA was used as the template, with 1 µl RNase (10 mg ml–1) and 1 µl of each primer at a stock concentration of 50 µM. PCR amplification of eryA was performed for 37 cycles, with an annealing time of 1 min at 56 °C, an extension time of 2 min at 72 °C, and a melting time of 45 s at 94 °C.
|
and plated on LB containing Amp and 2 % X-Gal. White colonies were screened for the correct insert by digestion with EcoRI, and then sent to the University of Calgary Core DNA sequencing facility for verification and to determine the orientation of the insert.
PCR cloning of the hypothetical gene and eryE gene.
Fragments of the hypothetical gene and eryE gene were cloned using the following primer pairs respectively: HypF, GAATGACATGCGTGAAGC; HypRev, GCTCCTGGTGAATGATGG; EryEF, AATTCGGCCAGTTTACCA; EryERev, GATGCTGAGCCCACTGAC. PCR amplification and cloning of PCR products were performed as stated above.
Mutagenesis of eryA, eryC, eryD, deoR, tpiA2, and rpiB, hyp and eryE.
Insertional mutagenesis was carried out using the vectors pJQ200SK and pJQ200mp18 and the strategy described by Quandt & Hynes (1993). The cloning strategies are described below. The eryA PCR-amplified fragment in pCR2.1-TOPO was cloned into pJQ200SK via double digest using the SstI and XhoI sites found in the multiple cloning sites of pCR2.1-TOPO and pJQ200SK. The eryA gene was disrupted by cloning a Sp-resistance cassette from pCY : : 
Sp (Yost, 1998) into the SalI site of eryA. eryD was mutated by first cloning a BamHI fragment from pCos42 that contained the eryC, eryD and deoR genes into pBluescript II SK+, then a Km-resistance cassette (Fellay et al., 1987) was cloned into the BglII site between the eryC and eryD genes. The resultant fragment was cloned into pJQ200mp18 via a BamHI digest. eryC was mutated by first cloning the BamHI fragment containing a portion of eryA, eryB and a small portion of eryC into pBluescript SK+. This fragment was transferred from pBluescript II SK+ to pJQ200mp18 using a BamHI/SstI double digest. A single cross-over into the VF39SM genome resulted in the disruption of eryC. The deoR gene was disrupted by first cloning the BamHI fragment from pCos42 that contained the eryC, eryD and deoR genes into pJQ200mp18. A GusNm cassette from pCRS530 (Reeve et al., 1999) was then inserted into the XhoI site in the deoR gene. To mutate tpiA2 and rpiB, the 1.5 kbp BamHI/PstI fragment containing both tpiA2 and rpiB (Fig. 1
) was cloned into pJQ200SK. A Km-resistance cassette was liberated from pHP45Km (Fellay et al., 1987) via a SmaI digestion and was cloned into the internal NruI site of tpiA2. To mutate rpiB, a Km-resistance cassette was liberated from pHP45Km (Fellay et al., 1987) via a SmaI digestion and cloned into the internal NsbI site of rpiB. Construction of the hyp gene mutant was performed by cloning a 
1200 bp PstI fragment of pTOPO : : hyp into pJQ200mp18 digested with PstI. This construct was subsequently digested with NotI and a GusNm cassette was ligated into the internal NotI site of the hyp gene. The eryE mutant was created by cloning the eryE PCR fragment from pTOPO : : eryE via a BamHI/PstI digest and ligating into pJQ200sk digested with BamHI/PstI. A Km-resistance cassette was liberated from pHP45Km (Fellay et al., 1987) via a SmaI digestion and cloned into the internal SmaI site of eryE.
Construction of eryABCD, deoRtpiA2rpiB and hyperyEFG promoter fusions.
The eryA promoter fragment contained in pCR2.1-TOPO was cloned into pFus1par via an EcoRI digest. Clones for both the forward and reverse orientation of the eryA promoter fragment relative to the gusA gene were isolated. A deoR : : gusA fusion was created by cloning a EcoRI/BglII fragment containing a portion of the deoR gene and upstream region into the vector pFus1par digested with EcoRI/BglII. A hyp : : gusA fusion was created by cloning the 855 bp EcoRI fragment, containing the first 538 bp of the hyp gene and 317 bp of upstream region from pTOPO : : hyp into pFus1par digested with EcoRI. pTOPO : : eryE was digested with EcoRI and the eryE fragment was ligated to pFus1par digested with EcoRI. This resulted in a gusA fusion with a fragment that contains 191 nt upstream from the start codon of eryE.
Assays for 
-glucuronidase activity.
Assays for gusA activity in both free-living and bacteroid states were performed according to Miller (1972) and Wang et al. (1989), respectively, with modifications described by Yost et al. (2004).
Nodulation competition assays with plasmid-cured derivatives of VF39SM.
For all plant tests, sterilized pea seeds (Pisum sativum cv. Trapper) were prepared and germinated as described previously (Hynes et al., 1988). Bacteria were inoculated onto seedlings in growth pouches at concentrations of approximately 109 bacteria per seedling. In growth pouch experiments, six replicates of each treatment were always done. Plants were grown in a controlled-environment growth chamber as described previously (Hynes et al., 1988). Mixtures of two strains were prepared so that the bacteria were roughly in a 1 : 1 ratio as judged by optical density. The actual numbers for each strain were determined by plate counts on the inoculum. For tests in sterile sand/vermiculite and non-sterile soil, the experiments were done in 8-inch pots; four seedlings were planted per pot and there were three replicates for each treatment. Inoculum was added at a rate of approximately 2x109 bacteria per pot. Soil was collected from a native pasture on the Lethbridge Research Station, which had previously been shown to contain low numbers of indigenous R. leguminosarum (Hynes & O'Connell, 1990). Plants were grown in a greenhouse with natural light supplemented with fluorescent light. Nodules were collected from plants after 4 weeks growth, and were surface-sterilized and processed using microtitre plates and a ‘nodule crusher’ as described by Beattie & Handelsman (1989). In general, 96 nodules were taken from each treatment (16 per plant in growth pouches, 32 per pot in other experiments). Identification of bacteria reisolated from nodules was based on antibiotic resistances and confirmed by examination of plasmid profiles; the production of melanin by VF39SM derivatives was also used to confirm identity, as was the unusual colony morphology of 336 derivatives when grown on VMM. Data from competition experiments were evaluated by application of the binomial probability theory. The null hypothesis was that both competing strains were equally competitive; the probability of obtaining the observed ratio of nodule occupancy given that there was no difference in competitiveness was calculated using a FORTRAN program written specifically for that purpose. If this value was less than 0.05, it was assumed that there was indeed a difference in competitiveness. The validity of these results was verified using a chi-squared test, based on the null hypothesis that expected ratios of bacteria in nodules would be the same as the ratio in the inoculum.
Use of plant exudates.
The ability of various strains to use plant exudate for growth was tested in VMM from which carbon and nitrogen sources were deleted. Plant exudates were prepared from seeds and roots as described by Mulligan & Long (1985), and used to supplement VMM as outlined by Hynes & O'Connell (1990). Medium depletion experiments were carried out by growing cultures of various strains to saturation (until there was no increase in optical density) in test tubes containing 5.0 ml VMM with plant exudates. Cultures were filtered through 0.2 µm nitrocellulose filters and the filtrate was then inoculated with 0.1 ml (approx. 2x107 cells) of an overnight culture of strain VF39SM, which had been washed twice in sterile distilled water. Growth of VF39SM on the filtrates was determined after growth for 72 h by examining cultures and scoring them for growth based on optical density, and by enumerating bacteria (c.f.u.) by plating on TY plates. Experiments were done in duplicate and have been repeated at least three times with similar results.
Nodulation competition assays with strain 19B-3.
Trapper pea (P. sativum) seeds were surface-sterilized by soaking the seeds in 2.6 % sodium hypochlorite for 5 min, followed by a 5 min soak in 70 % ethanol. The seeds were then rinsed three times in sterile distilled water. Seeds were incubated on water agar plates [12.5 g agar (l distilled water)–1] at room temperature in the dark for 3 days. Germinated seedlings were transferred to modified magenta jars that were designed to resemble Leonard jars (Brom et al., 1992; Oresnik et al., 1998) and contained a growth substrate of vermiculite.
R. leguminosarum cultures were grown overnight in TY. The seedlings were co-inoculated with VF39SM and 19B-3 or 19B-3cos42 at 1 : 1 and 1 : 10 ratios. Approximately 106 cells ml–1 were added to the seedlings. The proportion of wild-type to mutant strain was confirmed by performing viable plate counts on the inoculant suspension. The inoculated peas were then grown for 3 weeks, after which the nodules were harvested, and surface-sterilized by washing them in a 1 % solution of sodium hypochlorite for 5 min, followed by a 5 min wash in 70 % ethanol. The nodules were then rinsed twice in sterile distilled water. Surface-sterilized nodules were placed in individual wells of a 96-well microtitre tray containing 40 µl sterile distilled H2O and crushed using inoculating sticks. Seven microlitres of the macerate was spotted in duplicate onto TY plates containing Sm and TY plates containing Sm and Tc, to distinguish which strain had formed the nodule. For each competition experiment 96 nodules were collected for analysis. Three independent competition experiments were conducted.
DNA sequence analysis.
The databases used for analysis of the ery genes were: GenBank for Brucella spp; Rhizobase (http://www.kazusa.or.jp/rhizobase/) for rhizobial strains; and the Wellcome Trust Sanger Institute website (http://www.sanger.ac.uk/cgi-bin/blast/submitblast/r_leguminosarum) for R. leguminosarum 3841. Phylogenetic analysis of CLUSTALW alignments was performed using the online server of the Wageningen University and Research Centre Laboratory of Bioinformatics (http://www.bioinformatics.nl/tools/clustalw.html), with the standard parameters specified by the online server.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Hynes & McGregor, 1990) allowed us to verify that the four largest of the six plasmids in this strain are necessary for competition for nodulation of peas and lentils. Since three of these plasmids, pRleVF39c, pRleVF39d and pRleVF39e, are required for formation of nitrogen-fixing nodules (Hynes & McGregor, 1990), their requirement for competition is not surprising. However, the largest plasmid, pRleVF39f, is not required for nodulation or nitrogen fixation, so we decided to examine its role in competition in more detail. When strain LRS39601 was assayed in competition experiments against several standard test strains, in sterile growth pouches, in sterile vermiculite, or in non-sterile soil, it invariably formed a much smaller percentage of the nodules on both pea and lentil than the parent strain VF39SM. Table 2 shows some typical data.
|
Catabolic genes for many carbon sources are localized to plasmids in the genome of VF39SM
In order to identify specific compounds whose utilization was encoded by plasmids in VF39SM, we tested the cured strains in VMM supplemented with several sugars and amino acids as sole carbon source. Since strain LRS39501 does not grow on VMM (Hynes & McGregor, 1990), we also used Biolog plates to assay for the presence of catabolic processes missing in the cured strains. Based on these experiments we determined that pRleVF39b is required for growth in VMM containing gluconate, glucuronate, or malonate; pRleVF39c is required for growth in VMM containing glycerol or melibiose; pRleVF39d is required for growth in VMM containing adonitol, alanine, hydroxy-L-proline, or trigonelline; pRleVF39e is required for catabolism of histidine, rhamnose, serine and sorbitol; and pRleVF39f is required for growth in VMM containing arginine, citrate, erythritol, ornithine, or proline. Based on these experiments and the recent genome sequence of the closely related R. leguminosarum strain 3841, it is clear that many carbon catabolic or transport genes are spread amongst the plasmids. We were particularly interested in the fact that pRleVF39f apparently carries genes involved in erythritol utilization, since the ability to catabolize erythritol is a characteristic that can be used to differentiate rhizobia from biovar I Agrobacterium strains (New & Kerr, 1971; Jordan, 1984), and the molecular basis for this phenotype has not been investigated. We were able to confirm by plasmid transfer experiments that transfer of pRleVF39f to A. tumefaciens UBAPF2 conferred the ability to grow on erythritol on the transconjugants. Experiments were then initiated to identify the genetic determinants for erythritol catabolism and to assess their importance in competition.
Screening of a VF39SM transposon mutant library for strains deficient in erythritol utilization
A previously constructed transposon mutant library of VF39SM (García-de los Santos et al., 2002) was screened for mutants unable to catabolize erythritol. Approximately 4000 mutants were screened, resulting in the identification of four strains unable to utilize erythritol. Three of these mutants were also affected in the catabolism of other carbon sources and were not analysed further since the mutated gene was likely a component of a general catabolic pathway. The remaining mutant was selected for further analysis. The DNA flanking the transposon insertion was cloned and submitted for DNA sequencing. A BLASTX (Altschul et al., 1997) search using the DNA sequence identified a ribose ABC transporter permease from Brucella melitensis (GenBank accession YP_223164) as a best fit (E value 2x10–73) with 89 % identity and 98 % similarity over 154 residues. A BLASTN search of the R. leguminosarum 3841 genome identified the region in 3841 corresponding to the 19B-3 flanking DNA, and it was interesting that the genes adjacent to the insertion showed strong homology to erythritol catabolism genes from Brucella spp.
A previously constructed cosmid library of VF39SM (Yost et al., 1998) was probed using the DNA flanking the transposon insertion. Three cosmids were identified that hybridized to the probe: pCos42, pCos718 and pCos733. pCos42 restored the ability to grow on erythritol to 19B-3, whereas pCos733 partially restored erythritol catabolism by allowing 19B-3 to grow slowly and produce punctiform colonies on VMM erythritol medium, while pCos718 did not restore erythritol catabolism to 19B-3. pCos42 also restored erythritol utilization to LRS39601, thus indicating that pCos42 contained all the genes necessary for erythritol catabolism that are carried on pRleVF39f.
Genetic characterization of ery loci using the R. leguminosarum 3841 genomic sequence and mutants in R. leguminosarum VF39SM
The R. leguminosarum strains VF39SM and 3841 are quite similar, as both contain six plasmids, of very similar sizes, and encoding similar genes (Hynes et al., 1989; our unpublished work with cured strains of 3841), and we have found most genes we have sequenced in VF39SM to be 98–99 % identical to their homologues in 3841. We thus took advantage of the availability of the genomic sequence of 3841 in order to study the erythritol utilization locus carried on pCos42. It should be noted that the restriction sites predicted by the DNA sequence of this region from 3841 were all exactly conserved in pCos42, as we were able to confirm experimentally. Clones containing eryA, eryB, hyp and eryE amplified from pCos42 DNA were verified by sequencing and proved to be 99 % identical to the 3841 sequence. Additionally, the cloned DNA flanking the Tn insertion of 19B-3 was 99 % identical to the 3841 sequence. The total length of DNA sequence used in these alignments was 3820 nt. The genes adjacent to the Tn5-B22 insertion in 19B-3 were likely to be associated with erythritol metabolism based on their significant homology to Brucella abortus genes implicated in erythritol catabolism (Sangari et al., 2000). Table 3 provides the homology scores from a BLASTP alignment comparing the ORFs predicted from the 3841 genome sequence to the ery genes from B. abortus (Sangari et al., 2000; Halling et al., 2005). Fig. 1 provides a genetic map of the arrangement of the ery and related genes as determined from the R. leguminosarum 3841 genome sequence. We assigned the names eryE, eryF and eryG to the transporter genes transcribed in the opposite direction from eryA. We also noted that the genes immediately downstream of the eryABCD operon were conserved in both the 3841 and B. melitensis 16M genomes (Fig. 1b). To verify the predicted functions from the BLASTP results and to determine how many genes are involved in erythritol catabolism, we undertook a strategy to mutate the ery genes, deoR, tpiA2 and rpiB (mutagenesis strategies are described in Methods).
|
) and phenotypic analysis of ery mutants, it is likely that the ery genes exist in three operons: (1) a transport operon with a gene encoding a hypothetical protein followed by the genes eryE, F, G, encoding a predicted ATP-binding transporter protein, a predicted permease, and a predicted periplasmic-binding protein, respectively; (2) a catabolic gene operon with the genes eryA, B, C, D; and (3) an operon encoding a DeoR-type regulator followed by a predicted triosephosphate isomerase gene (tpiA2), and then a homologue of a ribose-5-phosphate isomerase gene (rpiB). A rho-independent terminator was predicted following the erythritol transporter genes. A terminator could not be readily predicted from the DNA sequence following eryD.
Mutagenesis of the eryA, eryC, hyp, eryE, eryF, deoR, tpiA2 or rpiB genes resulted in an inability of these strains to utilize erythritol as a sole carbon source. In addition, our results strongly suggest that the Brucella spp. genes with homology to eryE and eryF (Table 3) are incorrectly annotated. The Brucella annotation suggests these genes are involved in ribose transport; however, the eryE and eryF mutants were capable of growing on ribose as a sole carbon source. As expected from the sequence homology, mutagenesis of negative regulator eryD did not diminish growth on erythritol. Addition of erythritol to VMM medium containing mannitol inhibited the growth of eryC, deoR, tpiA2 and rpiB mutants but did not impair growth of the eryA, eryD, hyp, eryE or eryF mutants. The toxic effect of erythritol on the eryC, deoR, tpiA2 and rpiB mutants suggests that a build-up of phosphorylated intermediates in the erythritol catabolic pathway is detrimental to the cell. The tpiA2 mutant was capable of growing on glycerol as a sole carbon source. Complementing mutants with pCos42 restored growth patterns to that of the wild-type.
Expression of the eryABCD operon was induced in the presence of erythritol (Table 4). The pFusPEryA fusion was not induced by erythritol when present in the VFeryE background (data not shown). There was a slight reduction in expression when VF39SM pFusPEryA was grown in the presence of both erythritol and glucose, glycerol or mannitol, but a slight increase when succinate was included with erythritol. Mutation of the eryD gene resulted in constitutive expression of the eryA : : gusA fusion on pFusPEryA in all the media conditions (Table 4). Constitutive expression of the eryA : : gusA fusion was also observed in an eryA mutant background, supporting the hypothesis that eryD is contained within an operon with eryA. The eryA fusion was not induced in nodules formed by wild-type VF39SM nor was the eryA fusion expressed at the levels observed in free-living conditions in the nodules formed by VFeryD (Table 4). Expression of the deoR-tpiA2-rpiB operon was induced in the presence of erythritol (Table 5). However, the fusion was not induced when in the VFeryA background (Table 5). In addition, expression of the fusion in the VFdeoR background results in unregulated expression as evidenced by Table 5. Expression of the hyp : : gusA gene fusion did not exhibit an induction by erythritol and in fact levels of expression were similar for all carbon sources tested, with a slight reduction in expression when grown in succinate (Table 6).
|
|
|
indicates, this mutant is impaired in its ability to compete against wild-type for nodulation. Selecting a mutant (19B-3) predicted to be involved in erythritol uptake to conduct the nodulation competition test rather than a catabolic mutant ensured that the competition results were bona fide and not an experimental artefact due to erythritol toxicity that could be experienced by a catabolic mutant. The growth rate of 19B-3 versus wild-type was identical in TY medium and in minimal medium with mannitol (data not shown) and therefore the diminished competition ability was unlikely to be due to a general decrease in growth rate.
|
) alignments using the nucleotide sequences or predicted amino acid sequences from all the erythritol genes group the Brucella and R. leguminosarum genes into the same cluster and on a separate branch from the other members of the Rhizobiaceae for which the erythritol genes have been sequenced. Fig. 3(b) shows the neighbour-joining tree of an eryB alignment. The groupings are contrary to current taxonomic predictions from phylogenetic markers such as 16S rRNA (Fig. 3a). To illustrate further the possibility of horizontal transfer of the ery genes between Rhizobium leguminosarum and Brucella spp., the tpiA genes were examined for similarity. There are two tpiA homologues (tpiA1 and tpiA2) found in the published sequenced genomes of the Rhizobiaceae. tpiA2 is associated with the ery genes (Fig. 1) and is required for erythritol metabolism while tpiA1 is the conventional triosephosphate isomerase involved in glycolysis (I. Oresnik, personal communication). The outcome of a CLUSTALW alignment of the tpiA genes is shown in Fig. 3(c). The tpiA1 alignment is in agreement with the current phylogenetic view of the Rhizobiaceae while the tpiA2 alignment groups R. leguminosarum and Brucella together. These sequence alignments and apparent phylogenies suggest that the ery gene clusters may have been transferred horizontally between Brucella and R. leguminosarum at some point after the divergence of the Sinorhizobium and Rhizobium clades.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), whereas the homologous genes in S. meliloti 1021 and Mesorhizobium loti are chromosomally encoded (Kaneko et al., 2000; Galibert et al., 2001).
The seminal work on erythritol catabolism was published by Sperry & Robertson (1975a, b) and was based on their studies of erythritol metabolism in B. abortus. Genetic characterization of the genes involved in erythritol metabolism followed later (Sangari et al., 2000). These publications provide the current knowledge of erythritol catabolism in bacteria. Although the catabolic pathway is largely well characterized, the genes involved in erythritol transport remained unknown. Our isolation of an erythritol-utilization-defective R. leguminosarum strain (19B-3) with a mutation in a gene predicted to be a permease strongly suggests that this gene functions in erythritol transport. The fact that the permease appears to be in an ABC-transporter operon, and that this operon is adjacent to the eryABCD catabolic operon (Fig. 1), also supports our conclusion. Furthermore, Table 5 provides evidence that the eryE gene is involved in erythritol transport since the pFus1PDeoR fusion is not induced by erythritol when present in a VFeryE background. The slight increase in gusA activity compared to mannitol background levels (Table 5) may be due to the non-specific transport of a low amount of eryrthritol by the same transporter system found on pCos733 that partially restored growth of 19B-3 on VMM erythritol medium. Based on sequence similarity, the ABC-transporter operon is an orthologue to the ABC-transporter operon found adjacent to the eryABCD operon in the genomes of B. abortus, B. melitensis and Brucella suis (Table 3, Fig. 1). It is highly likely that this transporter functions in erythritol transport in Brucella. However, the genes are currently annotated as encoding ribose transporters. R. leguminosarum 19B-3, VFeryE and VFhyp are capable of growing on ribose as a sole carbon source and it is likely that the transporter genes have been mis-annotated in the Brucella genomes. Interestingly, the most likely orthologues to eryEFG in S. meliloti 1021 are found on the megaplasmid pSymB, while the ery catabolic genes are located on the chromosome (Finan et al., 2001; Galibert et al., 2001).
Regulation of eryA expression in R. leguminosarum is similar to that of B. abortus (Sangari et al., 2000). Expression is inducible by growth in erythritol and is negatively regulated by the repressor EryD. A close to eightfold induction of the eryA fusion occurs when VF39SM is grown in erythritol compared to glucose (Table 6). Catabolite repression of polyol metabolism by glucose has been observed in R. leguminosarum bv. trifolii (Ronson & Primrose, 1979). The effect was observed with polyols that were poor carbon sources (such as ribitol) but not with stronger carbon sources (Ronson & Primrose, 1979). The fact that expression of the eryA fusion was not significantly repressed when other carbon sources (glucose, mannitol) were added to the growth media in addition to erythritol may suggest that the polyol erythritol is a good carbon source for R. leguminosarum. Root exudate from P. sativum did not induce expression of the eryA fusion, and we assume this is due to an absence or low concentration of erythritol in the exudate preparation. The eryA : : gusA fusion is constitutively expressed in an eryD mutant, indicating that EryD functions as a transcriptional repressor. The arrangement of the eryABCD genes in R. leguminosarum suggests that the genes are transcribed as an operon. Mutational analysis supports this hypothesis, as mutations in eryA or eryC have a polar effect on eryD function; hence eryA : : gusA fusions are constitutively expressed in eryA or eryC mutant strains. Although a rho-independent terminator was not predicted following eryD, it is likely that the operon ends at eryD because an eryD mutant is still capable of growing on erythritol, whereas a deoR, tpiA2, or rpiB mutant cannot grow on erythritol. If transcription extended past eryD and into these genes an eryD mutation should have a polar effect and result in an erythritol-utilization-defective phenotype (Fig. 1). Erythritol is unlikely to be a carbon source supplied to R. leguminosarum bacteroids during symbiosis. Therefore it is not surprising that expression of the eryABCD operon is down-regulated during nodulation (Table 4). However, the fact that an eryA : : gusA fusion remains down-regulated in the eryD mutant during nodulation is unexpected and suggests that other unidentified regulators function to repress the expression of the eryABCD operon in the bacteroid. We are investigating this further.
Characterizing the region immediately upstream of the eryE gene using a gusA fusion has demonstrated that there is no promoter in this region and that the erythritol transporter genes are expressed from a promoter upstream of the hyp gene (Table 6). The defective erythritol utilization phenotype of the hyp mutant (VFhyp) also supports this conclusion. The hyp gene is conserved in a limited number of genomes presently sequenced. A BLASTP search reveals a periplasmic lipoprotein conserved domain, and in the genomes of Erwinia carotovora, S. meliloti, M. loti, Yersinia pseudotuberculosis and Chromohalobacter salexigens the orthologous hyp gene is always associated with a predicted ABC-transporter operon (data not shown). The fact that the hyp : : gusA fusion was not induced by erythritol and appears to be expressed constitutively is somewhat surprising and further investigation will be required to determine the full significance of this result.
Sperry & Robertson (1975b) demonstrated that erythritol is toxic to an eryC mutant of B. abortus. We observed similar results in R. leguminosarum and our results also support the view that direct involvement of eryA is required for erythritol toxicity since growth of an eryA mutant (VFeryA) is not inhibited by the presence of erythritol in the growth media (Table 5). Sperry & Robertson (1975b) attributed the toxicity to an ATP drain caused by the activity of the erythritol kinase (eryA). However, impairment of growth may not be due to an ATP drain but rather to the toxic effect of a phosphorylated intermediate. A toxic effect caused by a build-up of phosphorylated intermediates in sugar catabolism has been proposed for a number of sugars although the exact mechanism of action is unknown (Englesberg et al., 1962; Reiner, 1977; Richardson et al., 2004).
The proposed catabolic pathway for erythritol catabolism ends with the production of dihydroxyacetone phosphate (Sperry & Robertson, 1975a). Triosephosphate isomerase (tpiA) converts dihydroxyacetone phosphate to glyceraldehyde 3-phosphate, which acts as substrate for the Krebs cycle (Irani & Maitra, 1977). Using the genome sequence of R. leguminosarum 3841, we identified a putative operon adjacent to the eryABCD operon that contained a tpiA homologue (tpiA2) closely flanked by a deoR homologue, and an rpiB homologue (Fig. 1). We were interested to determine the role of these genes in erythritol catabolism. The fact that a mutation in the deoR gene yielded a similar phenotype to mutations in tpiA2 and rpiB provides evidence for a polar effect and that these three genes are part of an operon. The tpiA2 mutant was unable to grow on erythritol but was able to grow on glycerol. Growth on glycerol is not surprising considering the presence of tpiA1 in the genome. Similar to an eryC mutation, growth of the deoR, tpiA2 or rpiB mutant is inhibited in the presence of erythritol. This suggests that accumulation of at least two different phosphorylated intermediates in the erythritol catabolism pathway is toxic to the cell and that toxicity involves a general mechanism rather than the action of a single specific intermediary compound. The effector molecule of the DeoR repressor is unlikely to be erythritol (the inducer of the eryABCD operon) based upon the fact that the deoR : : gusA fusion is not induced by erythritol in a VFeryA mutant background. Therefore it is likely that an intermediate of the catabolic pathway acts as the effector molecule to release repression of the deoR promoter.
A mutation in erythritol metabolism affects strain VF39SM's ability to nodulate competitively (Fig. 2). Catabolism of other sugars and components of root exudates has been previously shown to influence nodule competition in various rhizobia (Jiménez-Zurdo et al., 1995; Soedarjo & Borthakur, 1998; Oresnik et al., 1998; Fry et al., 2001). The mutation found in 19B-3 affects erythritol uptake and therefore the diminished competition ability is unlikely to be due to a toxic effect of erythritol on an erythritol catabolic mutant. This could be expected if an eryB, eryC or tpiA2 mutant were used in the competition experiments. Future investigations will focus on 19B-3's ability to colonize the rhizosphere and will determine at what stage erythritol metabolism is important. R. leguminosarum VF39SM can nodulate a large variety of host plants within the Viceae tribe, including plants from the genera Lens, Vicia and Lathyrus. To place the significance of erythritol metabolism in the broader context of R. leguminosarum–legume interactions, future work will also examine the host range of the competition-related phenotype.
Taxonomic studies, using a variety of phylogenetic markers, confirm that Brucella spp. are phylogenetically distinct from the symbiotic nitrogen-fixing clade of Rhizobium and Sinorhizobium (Gaunt et al., 2001; Eardly et al., 2005). Therefore, the high sequence homology between the erythritol metabolic genes of B. melitensis and R. leguminosarum (Table 3; Fig. 3b, c) strongly suggests that a horizontal transfer has occurred between these organisms. B. melitensis was selected as the type species for the CLUSTALW alignments. However, all three Brucella genomes (B. abortus, B. suis and B. melitensis) gave identical alignments to those found in Fig. 3(b) and 3(c) (data not shown). This is not surprising as the three species are extremely closely related (Halling et al., 2005) and, in fact, there is disagreement between taxonomists as to whether the abortus and suis species should be grouped as biovars of B. melitensis (Verger et al., 1985). It is unclear in which direction the horizontal transfer of the ery genes proceeded, but the absence of eryA genes and the ability to grow on erythritol from one strain of R. leguminosarum and all strains of R. etli that we examined suggests transfer from Brucella to Rhizobium as the most parsimonious explanation. Codon bias analysis did not reveal any clues that would help address this question (data not shown). However, Medrano-Soto et al. (2004) suggest that codon usage bias is a poor predictor for horizontal gene transfer. Unfortunately at this time there are very few complete genome sequences that appear to have genes for erythritol catabolism (based on a BLASTP search of EryC). Based on the alignment searches, ery gene sequences are currently restricted to the members of the alphaproteobacteria: Brucella spp., M. loti, R. leguminosarum, Rhizobium sp. NGR234 and S. meliloti. As more genomes are sequenced and released into the public databases we will gain a better understanding of the distribution of ery genes and hope to place the horizontal transfer of these genes in a broader phylogenetic context.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Baldani, J. I., Weaver, R. W., Hynes, M. F. & Eardly, B. D. (1992). Utilization of carbon substrates, electrophoretic enzyme patterns, and symbiotic performance of plasmid-cured clover rhizobia. Appl Environ Microbiol 58, 2308–2314.
Barnett, M. J., Fisher, R. F., Jones, T. & 23 other authors (2001). Nucleotide sequence and predicted functions of the entire Sinorhizobium meliloti pSymA megaplasmid. Proc Natl Acad Sci U S A 98, 9883–9888.
Beattie, G. A. & Handelsman, J. (1989). A rapid method for the isolation and identification of Rhizobium from root nodules. J Microbiol Methods 9, 29–33.[CrossRef]
Beringer, J. E. (1974). R factor transfer in Rhizobium leguminosarum. J Gen Microbiol 84, 188–198.[Medline]
Boivin, C., Barran, L. R., Malpica, C. A. & Rosenberg, C. (1991). Genetic analysis of a region of the Rhizobium meliloti pSym plasmid specifying catabolism of trigonelline, a secondary metabolite present in legumes. J Bacteriol 173, 2809–2817.
Brewin, N. J., Wood, E. A., Johnston, A. W. B., Dibb, N. J. & Hombrecher, G. (1982). Recombinant nodulation plasmids in Rhizobium leguminosarum. J Gen Microbiol 128, 1817–1827.
Brom, S., García-de los Santos, A., Stepkowsky, T., Flores, M., Davila, G., Romero, D. & Palacios, R. (1992). Different plasmids of Rhizobium leguminosarum bv. phaseoli are required for optimal symbiotic performance. J Bacteriol 174, 5183–5189.
Brom, S., García-de los Santos, A., Cervantes, L., Palacios, R. & Romero, D. (2000). In Rhizobium etli symbiotic plasmid transfer, nodulation competitivity and cellular growth require interaction among different replicons. Plasmid 44, 34–43.[CrossRef][Medline]
Capela, D., Barloy-Hubler, F., Gouzy, J. & 25 other authors (2001). Analysis of the chromosome sequence of the legume symbiont Sinorhizobium meliloti strain 1021. Proc Natl Acad Sci U S A 98, 9877–9882.
Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T. J., Higgins, D. G. & Thompson, J. D. (2003). Multiple sequence alignment with the CLUSTAL series of programs. Nucleic Acids Res 31, 3497–3500.
DelVecchio, V. G., Kapatral, V., Redkar, R. J. & 23 other authors (2002). The genome sequence of the facultative intracellular pathogen Brucella melitensis. Proc Natl Acad Sci U S A 99, 443–448.
Eardly, B. D., Nour, S. M., van Berkum, P. & Selander, R. K. (2005). Rhizobial 16S rRNA and dnaK genes: mosaicism and the uncertain phylogenetic placement of Rhizobium galegae. Appl Environ Microbiol 71, 1328–1335.
Eckhardt, T. (1978). A rapid method for the identification of plasmid desoxyribonucleic acid in bacteria. Plasmid 1, 584–588.[CrossRef][Medline]
Englesberg, E., Anderson, R. L., Weinberg, R. N. L. P. H., Huttenhauer, G. & Boyer, H. (1962). L-Arabinose-sensitive, L-ribulose 5-phosphate 4-epimerase-deficient mutants of Escherichia coli. J Bacteriol 84, 137–146.
Fellay, R., Frey, J. & Krisch, H. (1987). Interposon mutagenesis of soil and water bacteria: a family of DNA fragments designed for in vitro insertional mutagenesis of gram-negative bacteria. Gene 52, 147–154.[CrossRef][Medline]
Finan, T. M., Weidner, S., Wong, K. & 9 other authors (2001). The complete sequence of the 1,683-kb pSymB megaplasmid from the N2-fixing endosymbiont Sinorhizobium meliloti. Proc Natl Acad Sci U S A 98, 9889–9894.
Fry, J., Wood, M. & Poole, P. S. (2001). Investigation of myo-inositol catabolism in Rhizobium leguminosarum bv. viciae and its effect on nodulation competitiveness. Mol Plant Microbe Interact 14, 1016–1025.[Medline]
Galibert, F., Finan, T. M., Long, S. R. & 47 other authors (2001). The composite genome of the legume symbiont Sinorhizobium meliloti. Science 293, 668–672.
García-de los Santos, A., Morales, A., Baldomá, L., Clark, S. R., Brom, S., Yost, C. K., Hernández-Lucas, I., Aguilar, J. & Hynes, M. F. (2002). The glcB locus of Rhizobium leguminosarum VF39 encodes an arabinose-inducible malate synthase. Can J Microbiol 48, 922–932.[CrossRef][Medline]
Gaunt, M. W., Turner, S. L., Rigottier-Gois, L., Lloyd-Macgilp, S. A. & Young, J. P. (2001). Phylogenies of atpD and recA support the small subunit rRNA-based classification of rhizobia. Int J Syst Evol Microbiol 51, 2037–2048.[Abstract]
Goldmann, A., Boivin, C., Fleury, V., Message, B., Lecoeur, L., Maille, M. & Tepfer, D. (1991). Betaine use by rhizosphere bacteria: genes essential for trigonelline, stachydrine, and carnitine catabolism in Rhizobium meliloti are located on pSym in the symbiotic region. Mol Plant Microbe Interact 4, 571–578.[Medline]
Guntli, D., Heeb, M., Moënne-Loccoz, Y. & Defago, G. (1999). Contribution of calystegine catabolic plasmid to competitive colonization of the rhizosphere of calystegine-producing plants by Sinorhizobium meliloti Rm41. Mol Ecol 8, 855–863.[CrossRef]
Halling, S. M., Peterson-Burch, B. D., Bricker, B. J., Zuerner, R. L., Qing, Z., Li, L. L., Kapur, V., Alt, D. P. & Olsen, S. C. (2005). Completion of the genome sequence of Brucella abortus and comparison to the highly similar genomes of Brucella melitensis and Brucella suis. J Bacteriol 187, 2715–2726.
Heinrich, K., Gordon, D. M., Ryder, M. H. & Murphy, P. J. (1999). A rhizopine strain of Sinorhizobium meliloti remains at a competitive nodulation advantage after an extended period in the soil. Soil Biol Biochem 31, 1063–1065.[CrossRef]
Hirsch, P. R. (1979). Plasmid-determined bacteriocin production by Rhizobium leguminosarum. J Gen Microbiol 113, 219–228.
Hynes, M. F. & Finan, T. M. (1998). General genetic knowledge. In The Rhizobiaceae: Molecular Biology of Plant Associated Bacteria, pp. 25–34. Edited by H. P. Spaink, A. Kondorosi & P. J. J. Hooykaas. Dordrecht: Kluwer Academic.
Hynes, M. F. & McGregor, N. F. (1990). Two plasmids other than the nodulation plasmid are necessary for formation of nitrogen-fixing nodules by Rhizobium leguminosarum. Mol Microbiol 4, 567–574.[Medline]
Hynes, M. F. & O'Connell, M. P. (1990). Host plant effect on competition among strains of Rhizobium leguminosarum. Can J Microbiol 36, 864–869.
Hynes, M. F., Simon, R. & Pühler, A. (1985). The development of plasmid-free strains of Agrobacterium tumefaciens by using incompatibility with a Rhizobium meliloti plasmid to eliminate pAtC58. Plasmid 13, 99–105.[CrossRef][Medline]
Hynes, M. F., Brucksch, K. & Priefer, U. (1988). Melanin production encoded by a cryptic plasmid in a Rhizobium leguminosarum strain. Arch Microbiol 150, 326–332.[CrossRef]
Hynes, M. F., Quandt, J., O'Connell, M. P. & Pühler, A. (1989). Direct selection for curing and deletion of Rhizobium plasmids using transposons carrying the Bacillus subtilis sacB gene. Gene 78, 111–120.
