Analysis of the Rhizobium leguminosarum siderophore-uptake gene fhuA: differential expression in free-living bacteria and nitrogen-fixing bacteroids and distribution of an fhuA pseudogene in different strains

Kay H. Yeoman1, Florence Wisniewski-Dye2, Christopher Timony1, James B. Stevens1, Nicola G. deLuca1, J. Allan Downie2 and Andrew W. B. Johnston1

School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK1
John Innes Centre, Norwich Research Park, Colney Lane, Norwich NR4 7UH, UK2

Author for correspondence: Andrew W. B. Johnston. Tel: +44 1603 592264. Fax: +44 1603 592250. e-mail: a.johnston{at}uea.ac.uk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
A mutation was isolated in the Rhizobium leguminosarum gene fhuA, which appears to specify the outer-membrane receptor for the siderophore vicibactin. The mutant was defective in iron uptake and accumulated the siderophore vicibactin in the extracellular medium. Expression of fhuA was regulated by Fe3+, transcription being higher in iron-depleted cells. Transcription of fhuA was independent of a functional copy of rpoI, a neighbouring gene that specifies a putative ECF {sigma} factor of RNA polymerase and which is involved in siderophore production in Rhizobium. Mutations in fhuA did not detectably affect symbiotic N2 fixation on peas. An fhuA::gus fusion was expressed by bacteria in the meristematic zone of pea nodules but not in mature bacteroids. Some other strains of R. leguminosarum also contain a pseudogene version of fhuA. The sequences of some of these and the ‘real’ fhuA genes were determined.

Keywords: ECF {sigma} factor, fhu genes, iron-mediated regulation, pseudogene, rhizobia, siderophores

Abbreviations: CAS, chrome azurol sulphonate; ECF, extracytoplasmic factor; NTA, nitrilotriacetate; X-Gluc, 5-bromo-4-chloro-3-indolyl ß-D-glucuronide

The GenBank/EMBL/DDBJ accession number for the sequence determined in this work is AJ238208.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Apart from the lactobacilli, bacteria need iron, which they use as a component of many enzymes. In many biologically relevant conditions, iron exists in the extremely insoluble oxidized ferric form. Therefore, many bacteria make siderophores, small organic molecules that they excrete and which bind Fe3+. The Fe–siderophore complexes are internalized by dedicated transport systems (Braun et al., 1998 ; Crosa, 1997 ). Since the enzyme complex nitrogenase comprises iron-containing proteins, diazotrophs have a particularly high demand for iron. In the case of the root-nodule bacteria known as the rhizobia, which fix N2 symbiotically, the requirement must also be satisfied in competition with the host plant; note that the most abundant single plant protein in nodules is the iron-containing leghaemoglobin (see Fett et al., 1998 ).

Rhizobium leguminosarum bv. viciae, the symbiont of peas, lentils, vetches and some beans, makes vicibactin, a cyclic trihydroxamate siderophore with three residues each of N2-acetyl-N5-hydroxy-D-ornithine and D-hydroxybutyrate (Dilworth et al., 1998 ). Different rhizobia make other hydroxamates (Persmark et al., 1993 ), catechols (Roy et al., 1994 ), citrate (Guerinot et al., 1990 ) or anthranilate (Barsomonian et al., 1992 ). In some cases, rhizobial mutants defective in siderophore synthesis fix N2 normally (Reigh & O’Connell, 1993 ; Fabiano et al., 1995 ), but in others, Sid- mutants fail to fix N2 symbiotically (Barsomian et al., 1992 ). Yeoman et al. (1997) found that cyc (ccm) mutants of R. leguminosarum bv. viciae which are defective in cytochrome c maturation were, for reasons that are not clear, also compromised for vicibactin synthesis. Such mutants were also Fix- on peas, due to the defect in electron transport.

Stevens et al. (1999) identified some of the fhu genes of R. leguminosarum. These are homologues of the corresponding genes in (for example) Escherichia coli which are involved in the uptake of hydroxamate siderophores. In E. coli, FhuA is an outer-membrane receptor, FhuD a periplasmic transporter, FhuB an integral cytoplasmic membrane protein and FhuC an ATPase (Braun et al., 1998 ). In R. leguminosarum, fhuCDB are in one operon whose expression is enhanced in cells grown in low concentrations of iron. Mutations in fhuCDB caused cells to make larger haloes on plates containing the ‘universal’ siderophore indicator chrome azurol sulphonate (CAS) (Schwyn & Neilands, 1987 ) and were defective for vicibactin and iron uptake (Stevens et al., 1999 ). These mutants nodulated and fixed N2 normally on peas, indicating that vicibactin is not important in iron nutrition in bacteroids. It is not known if these bacteria make another, bacteroid-specific siderophore. It may also be the case that bacteroids acquire iron in the ferrous form; although bacteroids can take up both Fe2+ and Fe3+ iron, the efficiency is greater with the former (LeVier et al., 1996 ; Moreau et al., 1998 ).

In E. coli, fhuA is in the fhuACDB operon. In R. leguminosarum strain 8401pRL1JI, there is a different arrangement in which there is a version of fhuA oriented divergently from fhuCDB. However, this copy of fhuA appears to be a pseudogene; it has many stop codons and is not detectably expressed (Stevens et al., 1999 ). LeVier & Guerinot (1996) identified a gene, fegA, in Bradyrhizobium japonicum which was a homologue of fhuA and which specified an outer-membrane protein, made in response to iron deprivation.

A functional fhuA gene of R. leguminosarum has now been discovered and is described here. The effects of iron availability and of the regulatory genes rpoI, fur and feuQ on its transcription are described as are its expression in pea root nodules. We also looked for the presence of fhuA and its pseudogene version, {psi}fhuA, in a number of field isolates of R. leguminosarum.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Media and bacterial growth conditions.
Strains and plasmids used in this study are shown in Table 1. Strains of E. coli and R. leguminosarum were grown routinely as described by Beringer (1974) . In high-iron medium, FeCl3 (20 µM) was added; in low-iron medium, there was no added iron, but, instead, 2,2'-dipyridyl (20 µM) was present. Peas were inoculated, grown and assayed for N2 fixation by C2H2 reduction as in Beynon et al. (1980) . Qualitative CAS tests on agar plates were done as described by Yeoman et al. (1997) .


View this table:
[in this window]
[in a new window]
 
Table 1. Strains and plasmids

 
Enzyme assays.
ß-Galactosidase and ß-glucuronidase assays were done as described by Rossen et al. (1985) and by Wilson et al. (1992) , respectively. Nodules were stained for ß-glucuronidase activity in situ with 5-bromo-4-chloro-3-indolyl ß-D-glucuronide (X-Gluc) (Jefferson et al., 1987 ).

In vivo genetic manipulations.
Plasmids were transferred by conjugation into R. leguminosarum using the helper plasmid pRK2013 (Figurski & Helinski, 1979 ). Strain 8401 was mutagenized with Tngus by using it as a recipient in a conjugational cross with E. coli strain MM294, which contains a derivative of the plasmid pRK600 into which the transposon Tngus has been inserted (Sharma & Signer, 1990 ). Since pRK600 is mobilizable into Rhizobium but fails to replicate in that host, it acts as a ‘suicide’ plasmid. Thus, by selecting Kanr transconjugants (specified by Tngus), derivatives of strain 8401 with random insertions of the transposon into the genome were obtained, at frequencies of approximately 10-6. Kanr colonies were picked to minimal (Y) medium containing CAS. Transduction of Tngus from mutant A691 was done, using rhizobiophage RL38, as described by Buchanan-Wollaston (1979) .

In vitro DNA manipulations.
Routine transformations, restriction digestions, ligations, Southern blotting and hybridization were done essentially as described by Downie et al. (1983) . R. leguminosarum genomic DNA was isolated using a Promega genomic preparation kit. Sequencing was done by the dideoxy chain-termination method, in some cases by MWG Ltd, Germany. Data were analysed with the DNA-Star package. Searches of databases used BLAST in the EGCG package. The primers used to amplify the two fhuA genes were: fhuA, 5'-TCCATAGGTTCCGCCCGCATCCGT-3' and 5'-TTTCGACGATGTGATAGGCGACCG-3'; {psi}fhuA, 5'-GGAGCAGATCGGCAAGGTCGGCGTG-3' and 5'-CGCCGATCGCCGTAATATTCTGTGC-3'. The primers used to amplify the fhuA promoter region were: 5'-CGCAGATCTTCGCAGCCATCGAGGGGGC-3' and 5'-CGCGCATGCCGTAATTGATATAGGGCTGGC-3'.

Iron uptake.
Uptake of iron from 55Fe-NTA (prepared with 55FeCl3 and sodium nitrilotriacetate) was measured as described by Yeoman et al. (1997) . Cells were grown in minimal (Y) medium with FeCl3 (20 µM) or in the absence of added iron but with 2,2'-dipyridyl (20 µM). Vicibactin was identified by electrospray mass spectroscopy as described by Yeoman et al., 1999 .


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Isolation of a siderophore-overproducing mutant
Following mutagenesis of R. leguminosarum strain 8401 with Tngus (see Methods), one mutant, termed A691, with a larger halo on CAS plates (the diameter was about twice that of the wild-type) was isolated following the screening of approximately 4000 transconjugants. This phenotype was similar to that of fhuCDB mutants that were defective in vicibactin uptake (Stevens et al., 1999 ). However, we found that A691 was not corrected by pBIO400, a cosmid from a pLAFR1-based gene library of R. leguminosarum DNA containing cloned fhuCDB (Stevens et al., 1999 ). The mutation from A691 was transduced into strain 8401, selecting kanamycin resistance. All the transductants co-inherited the CAS phenotype of A691, demonstrating that Tngus caused this phenotype. One transductant, A775, was chosen for further study.

It was shown that A775 was defective in iron uptake. Cells were grown in low-iron medium and exposed to 55Fe-NTA. No detectable uptake of this substrate was observed (<2% of the wild-type value), over a period of 15 min. It was also confirmed that the extra CAS-staining material was vicibactin, not some novel siderophore. Cells of A775 were grown and the cell-free growth medium assayed for hydroxamate. Strain A775 contained approximately four times more hydroxamate than did the wild-type and it was confirmed by electrospray mass spectroscopy that this hydroxamate corresponded to vicibactin.

Cloning and analysis of the fhuA gene
To locate the Tngus insertion precisely, DNA was isolated from strain A775 and digested with EcoRI. Fragments were ligated to pBluescript and used to transform E. coli, selecting Kanr transformants. By such means, part of Tngus, containing Kanr and gus, together with Rhizobium genomic DNA immediately upstream of the gus reporter, was cloned to form pIJ9116. The DNA adjacent to gus was sequenced; this indicated that the transposon was located in DNA whose deduced protein product had similarity to the C-terminus of FhuA of E. coli and other bacterial hydroxamate receptors (Table 2; and see below). Significantly, this 3' end of an fhuA homologue was similar in sequence to the corresponding region of the pseudogene version, {psi}fhuA, identified by Stevens et al. (1999) in the same strain of R. leguminosarum. However the sequences were not identical (see below). Sequencing also showed that in A775, the gus reporter Tngus was in the same orientation as fhuA.


View this table:
[in this window]
[in a new window]
 
Table 2. Comparison of R. leguminosarum FhuA and {psi}FhuA with other proteins of this family

 
Yeoman et al. (1999) identified a putative ECF regulatory gene, rpoI, in R. leguminosarum bv. viciae. During the course of other studies on rpoI, the region upstream of the rpoI coding sequence had been determined (K. H. Yeoman, unpublished). We noted that this upstream sequence overlapped the 3' region of fhuA, which was sequenced (see above). The rpoI gene had previously been cloned as part of a cosmid, termed pBIO1096 (Yeoman et al., 1999 ). We used this cosmid as a source of DNA to confirm the overlap between the part of fhuA identified here and the previously identified DNA upstream of rpoI (see Fig. 1). pBIO1096 DNA was digested with different restriction enzymes, and following electrophoresis and blotting, was probed with pIJ9116 DNA. A single 3·7 kb PstI fragment hybridized. This fragment was subcloned from pBIO1096 into the wide-host-range vector pRK415 to form pBIO1097, which was then mobilized into strain A775. The transconjugants were restored to wild-type phenotype on CAS plates, showing that this fragment contained a functional fhuA gene. This fragment was sequenced; it was found to contain the whole of fhuA, plus 333 bp of upstream sequence (accession no. AJ238208).



View larger version (8K):
[in this window]
[in a new window]
 
Fig. 1. Diagram of the fhuA region of R. leguminosarum. The approximate locations and directions of transcription of the genes in the rpoI region are indicated. Restriction sites for EcoRI (E) and PstI (P) are shown, as are the dimensions of the cloned DNA in plasmids pBIO1096 and pBIO1097. The ‘flag’ above the restriction map indicates the location of the Tngus insertion in strain A775. The asterisk indicates that the cloned DNA in pIJ9116 also contains a piece of Tngus.

 
The deduced FhuA protein had a molecular mass of 79·5 kDa, and was similar throughout its length to FhuA proteins of other bacteria (Table 2) the greatest similarity, overall, being to FegA of B. japonicum (34·8% identity). The other proteins in the FhuA family are all predicted to be in the outer membrane and most have a typical signal sequence. The predicted fhuA gene product from R. leguminosarum has a potential signal sequence but it is not entirely typical (MARVFLNVSNNVSRIYRDSLFVTTAIVLIGIAASPAASQS). Thus, the first 18 residues are fairly hydrophilic and include three positively charged residues (R) and one acidic residue (D), followed by a hydrophobic stretch and a potential signal peptidase cleavage site (PAA/S). This predicted 40-residue signal sequence is somewhat larger than normal and the presence of the aspartate is atypical. However, the net positive charge of the N-terminal domain followed by the hydrophobic residues may be consistent with this region acting as a signal peptide. TonB-dependent outer-membrane proteins, such as FhuA, share amino acid sequences, termed the TonB boxes I, II and III (see Postle, 1999 ). FhuA of R. leguminosarum has a potential Ton box I (EVPRS) near the N-terminus and putative Boxes II and III are present between amino acid residues 693–699 and 167–190, respectively.

In the region 5' of fhuA there were no obvious regulatory sequences (e.g. a fur box) and the promoter of this gene has not yet been identified. A gap of 1·5 kb separates the 3' end of fhuA and the start of rpoI. Within this DNA there was no ORF larger than 300 bp and there was no homology of this DNA nor any potential peptide products to sequences in databases. However, this is a rather large intergenic space, and it may be that it does contain a short gene of unknown function. In the intergenic region, 57 bp 3' of fhuA is a perfect inverted repeat, (5'-CCGTCGCCCACCAGGCCCGTCGACCTCGACGGCCTGGTGGGCGACGG-3'), which might act as a {rho}-independent transcriptional terminator.

Expression of fhuA: effects of iron, rpoI, feuQ and fur
The fhuCDB operon of R. leguminosarum is expressed at higher levels in cells that are depleted for iron than in those that are replete for the metal (Stevens et al., 1999 ). A derivative of the fhuA::gus mutant strain A775 containing pBIO1097 (to correct the defect in iron uptake of strain A775 itself), was grown in high- and low-iron media (see Methods) and assayed for ß-glucuronidase. As with fhuCDB, expression was higher in the latter than the former medium, the activity being 54±12 and 858±67 Miller units, respectively.

We had previously identified two R. leguminosarum genes, feuQ and rpoI, which are believed to be regulatory and affect iron uptake. Yeoman et al. (1997) showed that a mutation in feuQ (which is likely to be a sensor in another representative of the two-component family of transcriptional regulators) severely affected iron uptake, although siderophore production appeared to be unaltered. Mutations in rpoI, which is downstream of fhuA, nearly abolish vicibactin production and the presence of cloned rpoI enhances siderophore production in R. leguminosarum.

To measure the effects of feuQ and rpoI on fhuA expression, an fhuA::lacZ transcriptional fusion plasmid was made as follows. A 1 kb PCR fragment, containing 470 bp of the N-terminal coding region of fhuA plus 530 bp upstream of fhuA was made, using the primers shown in Methods, and with pBIO1096 as template. This fragment was cloned first into pUC18 and thence into the EcoRI–SphI sites of the wide host-range promoter-probe plasmid pMP220 to form pBIO1111. This plasmid was then mobilized into wild-type strain 8401pRL1JI and the feuQ and rpoI mutant derivatives, J100 and J256, respectively. The transconjugants were grown in high- and low-iron media and were assayed for ß-galactosidase. As with the fhuA::gus fusion, it was found that addition of Fe3+ to the growth medium reduced expression of the fusion. This was true for the wild-type background (211 Miller units in high-iron and 1555 units in low-iron medium) and in the two mutants J100 (216 and 1455 units) and J256 (188 and 1318 units). These results showed that neither rpoI (strain J256) nor feuQ (strain J100) is required for transcription of fhuA, nor do they mediate the iron-dependent control of its expression. It had also been shown previously that neither rpoI nor feuQ had any detectable effect on expression of an fhuB::lacZ fusion in either high- or low-iron media (Stevens et al., 1999 ; Yeoman et al., 1999 ).

deLuca et al. (1998) described a homologue of the ‘global’ regulator fur in R. leguminosarum but could not obtain a knockout mutation in it, suggesting that this gene is essential. Nevertheless, since fur regulates expression of the fhu genes of other bacteria in an iron-dependent way (Crosa, 1997 ), we examined the effects of fur on expression of the fhuA::gus fusion in strain A775 by conjugating into it plasmid pBIO929, which contains fur of R. leguminosarum. This derivative was grown in high- and low-iron media and assayed for ß-glucuronidase; it had no effect on the fhuA::gus expression in either medium. However, in the absence of a knockout mutant, we cannot be sure whether fur has a role in regulating fhuA transcription or not.

fhuA and symbiotic nitrogen fixation
The original fhuA mutant, A775, was derived from R. leguminosarum strain 8401, which lacks a symbiotic plasmid and so fails to nodulate. To examine the effects of the fhuA::gus mutation on nodulation, peas were inoculated with strain J253, a derivative of A775 into which plasmid pRL1JI had been introduced by conjugation. Judged by the numbers and sizes of the nodules, their time of appearance and the levels of C2H2 reduction, J253 appeared to be unaffected in symbiotic N2 fixation. Bacteria isolated from these nodules were found to retain the large-halo phenotype of the input fhuA inoculant and were all Kanr. This Nod+ Fix+ phenotype is similar to what is seen with fhuCDB mutants (Stevens et al., 1999 ) and points to the fact that the fhu system of Fe3+ uptake is unimportant in N2-fixing bacteroids. Consistent with this were the observations obtained with nodules stained with X-Gluc. The only part of the nodule with significant ß-glucuronidase activity was near the meristem, where non-differentiated bacteria, some still in infection threads, are located. In the zone containing mature, N2-fixing bacteroids, no staining was seen (Fig. 2).



View larger version (103K):
[in this window]
[in a new window]
 
Fig. 2. Pea root nodule containing fhuA::gus fusion strain A253 and stained with X-gluc. Note that staining for ß-glucuronidase occurs only at the nodule tip, containing the undifferentiated bacteria. The mature, N2-fixing zone, nearer the root, has no detectable staining.

 
Since these nodules were induced by mutants that were defective in vicibactin uptake, their pattern of expression of fhuA::gus might not reflect the normal situation. Therefore, pBIO1097, which contains cloned fhuA, was mobilized into J253 and transconjugants were used to inoculate peas. The same pattern of staining with X-gluc of the nodules was found as with strain J253 itself (not shown).

Pseudogene {psi}fhuA occurs in different R. leguminosarum strains
Since a mutation in the version of fhuA described here causes a defect in iron uptake and a large-halo phenotype on CAS, it is clear that this gene is functional. In contrast, the {psi}fhuA pseudogene that is adjacent to fhuCDB, but which is unlinked to the functional fhuA gene identified here, is a non-functional pseudogene and is not expressed (Stevens et al., 1999 ).

A comparison of parts of the C-terminal regions of FhuA and {psi}FhuA is shown in Fig. 3. The similarity is no greater than that between R. leguminosarum and members of the FhuA family of proteins in other bacteria (Table 2). We wished to see if genes corresponding to {psi}fhuA were widespread in strains of R. leguminosarum. Probes corresponding to parts of fhuA and {psi}fhuA were made by PCR using as templates pBIO1096 and pBIO400, which respectively contain the two versions of the gene. The locations of the primers used for {psi}fhuA are shown in Fig. 3; those used for fhuA are given in Methods and are located in the 5' half of that gene.



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 3. Comparison of the C-terminal regions of the FhuA and corresponding FhuA gene products of R. leguminosarum strain 8401pRL1JI. The deduced amino acid sequences are shown in single-letter code. Asterisks indicate identical residues. The bold ‘N’ indicates the difference in the sequences of the {Psi}FhuA gene product in strains BC1 and RES-6 compared to that sequence in strain 8401pRL1JI. The locations corresponding to those used for generating the primers P1 and P2, used to amplify {psi}fhuA, are shown.

 
Two PCR products of the expected sizes were generated from fhuA and {psi}fhuA. Each product was then used to probe DNA obtained from 10 field isolates of this species. These strains had been shown from RFLP to be distinct from each other and from 8401pRL1JI (Rigottier-Gois at al., 1998 ). Genomic DNAs were digested with EcoRV and PstI and, after electrophoresis, were probed separately with the two PCR 32P-labelled products. With the fhuA probe, bands hybridizing as intensely as with strain 8401pRL1JI were seen for all 10 strains (Table 3). These exhibited some RFLP, different patterns being obtained which were consistent with the relatedness of the strains (Rigottier-Gois et al., 1998 ) (Table 3). With the {psi}fhuA probe, strains BC1 and RES-6 hybridized strongly and to the same-sized bands as with 8401pRL1JI. Strain BB18 gave a very weak signal, the band being larger than for 8401pRL1JI; seven strains had no detectable signal with the {psi}fhuA probe. The two strains hybridizing to {psi}fhuA had similar fhuA RFLP patterns to each other and to strain 8401pRL1JI (Table 3). Following hybridization of the EcoRV-digested DNAs with the {psi}fhuA probe, the filter was stripped and reprobed with fhuA DNA. The original bands did not reappear, but ‘new’ bands, corresponding to those expected for fhuA, were seen (Table 3). Thus, there was no cross-hybridization between {psi}fhuA and fhuA in the strains used here.


View this table:
[in this window]
[in a new window]
 
Table 3. Sizes (kb) of EcoRV and PstI genomic fragments hybridizing to fhuA and {psi}fhuA probes in field isolates of R. leguminosarum bv. viciae

 
The {psi}fhuA primers were used to amplify the corresponding DNA from the genomes of strains BC1 and RES-6 and the resulting fragments were sequenced. These products were identical to each other and differed by one nucleotide (causing a D to N substitution in the protein) from that of strain 8401pRL1JI (Fig. 3); by chance this difference resulted in the loss of a SalI site in strains RES-6 and BC1. It was confirmed that there was a SalI RFLP at this location. We failed to get a genomic PCR product from strain BB18, which showed weak hybridization to {psi}fhuA.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
We have identified a functional copy of fhuA which is homologous to fhuA genes which, in other bacteria, encode the receptor for hydroxamate siderophores (see Braun et al., 1998 ). In B. japonicum, an FhuA homologue, termed FegA, was shown to be present in enhanced levels in bacteria depleted of iron (LeVier & Guerinot, 1996 ).

Using transcriptional fusions to gus and to lacZ, we found that fhuA of R. leguminosarum was transcribed at elevated levels in cells starved of iron. Stevens et al. (1999) had similarly found that the unlinked fhuCDB operon of R. leguminosarum was Fe3+ regulated. We do not know, however, what regulatory gene mediates this iron-dependent expression. It is apparent from the results obtained here that two regulatory genes, rpoI and feuQ, which affect iron uptake in R. leguminosarum, are not involved. Given the adjacent locations of rpoI and fhuA, together with the facts that mutations in rpoI abolish siderophore synthesis and that overexpression of rpoI causes enhanced production of vicibactin (Yeoman et al., 1999 ), we were surprised at the lack of interaction between fhuA and rpoI. At present, the ‘target’ gene(s) that require the RpoI {sigma} factor for their transcription remains to be identified.

An fhuA::gus fusion was used to show that although fhuA is expressed in undifferentiated bacteria in the nodule, the mature N2-fixing bacteroids do not transcribe this gene at detectable levels. The basis of this switch-off regulation is not known. Several genes that are expressed in free-living rhizobia are quiescent in bacteroids; these include the pss (exo) genes for polysaccharide synthesis (Latchford et al., 1991 ), the amtB gene that is involved in ammonium transport (Tate et al., 1999 ), and some nod genes required for the early steps in the infection process (Schlaman et al., 1991 ; Marie et al., 1992 ). The mechanisms involved in the down-regulation of genes in bacteroids have received little attention and it remains to be seen if there is some ‘global’ control or if individual genes have specific shut-down systems. The finding that fhuA is not expressed in bacteroids is consistent with the lack of symbiotic defects found with various fhu mutants (this study; Stevens et al., 1999 ). To date, the only R. leguminosarum mutants as yet identified that are defective both in siderophore synthesis and in N2 fixation are the cyc (ccm) mutants described by Yeoman et al. (1997) . In these cases, it seems almost certain that it is the respiratory defect rather than that in siderophore synthesis that is responsible for the symbiotic phenotype (Delgado et al., 1995 ). The negative results with defined fhu mutants strongly indicate that vicibactin is not used for iron uptake in R. leguminosarum bacteroids. It may be that there is another, unknown bacteroid-specific siderophore system. Alternatively, there is circumstantial evidence that bacteroids acquire iron in the reduced, ferrous form. Bacteroids of B. japonicum in soybean nodules can import both Fe3+ and Fe2+, but the uptake of the latter is more efficient (Moreau et al., 1998 ). However, in the absence of (feo) mutants that are defective in Fe2+ uptake, it is impossible to know the relative importance of the two forms of iron in bacteroid nutrition.

Stevens et al. (1999) identified a pseudogene version of fhuA, next to the functional fhuCDB genes. In that study, no hybridization to any other DNA was observed when a probe spanning {psi}fhuA was used. It is clear, though, that R. leguminosarum strain 8401pRL1JI does contain a functional gene that is unlinked to {psi}fhuA; however, a comparison of the sequences of fhuA and {psi}fhuA in this strain shows that there is DNA and protein homology in only limited areas. The similarity of the potential products of the fhuA and {psi}fhuA sequences in R. leguminosarum strain 8401pRL1JI is no greater than that found between the FhuA of this strain and those of other bacteria. This suggests that these two genes did not arise via recent gene duplication followed by limited divergence. Rather, it points to fhuA and {psi}fhuA having had quite different origins, one, perhaps, having been acquired by gene transfer.

It is apparent that other different field isolates of R. leguminosarum contain homologues of both the functional and the pseudogene versions of fhuA. However, the two strains that contained a close homologue of {psi}fhuA appeared to be closely related to each other and to strain 8401pRL1JI as judged from their RFLP patterns at other loci (Rigottier-Gois et al., 1998 ; this study). In these two strains, the sequenced regions of the {psi}fhuA homologues were identical to each other and differed in only one base pair from the allele in 8401pRL1JI. We were surprised that a pseudogene version of the gene differed so little in different strains since, by definition, such genes are not subject to the constraints that are required to maintain gene function. It will be of interest to know the precise sequences of the regions around the pseudogene versions of fhuA in different strains of R. leguminosarum. Are they in the same relative positions in the chromosomes? How large are the regions that distinguish these pseudogene regions from those in the strains that do not harbour them (or which, perhaps, contain different versions of fhuA pseudogenes)?

Pseudogenes are relatively rare in prokaryotes. It was noted by Stevens et al. (1999) that in bacteria, the deduced original products of several pseudogenes, including R. leguminosarum {psi}fhuA, are located at the cell surface. FhuA of E. coli is a receptor for several coliphages, colicins and at least one antibiotic (Killman & Braun, 1992 ; Killman et al., 1995 ). It may be that there is particularly strong selection pressure to lose versions of such cell-surface proteins that can act as targets for such antimicrobial agents. The finding here that pseudogenes exist in only a minority of strains of R. leguminosarum suggests that such selection pressure may be intermittent and not uniform in different populations.


   ACKNOWLEDGEMENTS
 
We are grateful to Peter Young for field isolates of R. leguminosarum and to Alan Cavell for sequencing PCR products. The work was funded by the BBSRC of the UK and EU contract (BIO4-CT96-0181) to J.A.D.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Barsomonian, G. D., Urzainqui, A., Lohman, K. & Walker, G. C. (1992). Rhizobium meliloti mutants unable to synthesize anthranilate display a novel symbiotic phenotype. J Bacteriol 174, 4416-4426.[Abstract]

Beringer, J. E. (1974). R factor transfer in Rhizobium leguminosarum. J Gen Microbiol 84, 188-198.[Medline]

Beynon, J. L., Beringer, J. E. & Johnston, A. W. B. (1980). Plasmids and host-range in Rhizobium leguminosarum and Rhizobium phaseoli. J Gen Microbiol 120, 421-429.

Braun, V., Hantke, K. & Koster, W. (1998). Bacterial iron transport: mechanisms, genetics, and regulation. Metal Ions Biol Syst 35, 67-145.

Buchanan-Wollaston, A. V. (1979). Generalised transduction in Rhizobium leguminosarum. J Gen Microbiol 112, 135-142.

Carson, K. C., Glenn, A. R. & Dilworth, M. J. (1994). Specificity of siderophore-mediated transport of iron in rhizobia. Arch Microbiol 161, 333-339.

Crosa, J. H. (1997). Signal transduction and transcriptional and post-transcriptional control of iron-regulated genes in bacteria. Microbiol Mol Biol Rev 61, 319-341.[Abstract]

Delgado, M. J., Yeoman, K. H., Wu, G., Vargas, C., Davies, A. E., Poole, R. K., Johnston, A. W. B. & Downie, J. A. (1995). Characterisation of the cycHJKL genes involved in cytochrome c biogenesis and symbiotic nitrogen fixation in Rhizobium leguminosarum. J Bacteriol 177, 4927-4934.[Abstract]

Dilworth, M. J., Carson, K. C., Giles, R. G. F., Byrne, L. T. & Glenn, A. R. (1998). Rhizobium leguminosarum bv. viciae produces a novel cyclic trihydroxamate siderophore, vicibactin. Microbiology 144, 781-791.

Downie, J. A., Hombrecher, G., Ma, Q. S., Knight, C. D., Wells, B. & Johnston, A. W. B. (1983). Cloned nodulation genes of Rhizobium leguminosarum determine host-range specificity. Mol Gen Genet 190, 359-365.

Fabiano, E., Gill, P. R., Noya, F., Bagnasco, P., Delafuente, L. & Arias, A. (1995). Siderophore-mediated iron transport iron acquisition mutants in Rhizobium meliloti 242 and its effect on the nodulation kinetics of alfalfa nodules. Symbiosis 19, 197-211.

Fett, J. P., LeVier, K. & Guerinot, M. L. (1998). Soil microorganisms and iron uptake by higher plants. Metal Ions Biol Syst 35, 188-214.

Figurski, D. H. & Helinski, D. R. (1979). Replication of an origin-containing derivative of plasmid RK2 dependent on a plasmid function provided in trans. Proc Natl Acad Sci USA 76, 1648-1652.[Abstract]

Guerinot, M. L., Meidl, E. J. & Plessner, O. (1990). Citrate as a siderophore in Bradyrhizobium japonicum. J Bacteriol 172, 3298-3303.[Medline]

Jefferson, R. A., Kavanagh, T. A. & Bevan, M. W. (1987). Gus fusions – ß-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J 6, 3901-3907.[Abstract]

Keen, N. T., Tamaki, S., Kobayashi, D. & Trollinger, D. (1988). Improved broad-host range plasmids for DNA cloning in Gram-negative bacteria. Gene 70, 191-197.[Medline]

Killman, H. & Braun, V. (1992). An aspartate deletion mutation defines a binding-site of the multifunctional FhuA outer-membrane receptor of Escherichia coli K-12. J Bacteriol 174, 3479-3486.[Abstract]

Killman, H., Videnov, G., Jung, G., Schwarz, H. & Braun, V. (1995). Identification of receptor-binding sites by competitive peptide mapping – phage-T1, phage-T5, and phage-{phi}80 and colicin-M bind to the gating loop of FhuA. J Bacteriol 177, 694-698.[Abstract]

Lamb, J. W., Hombrecher, G. & Johnston, A. W. B. (1982). Plasmid-determined nodulation and nitrogen-fixation abilities in Rhizobium phaseoli. Mol Gen Genet 186, 449-452.

Latchford, J. W., Borthakur, D. & Johnston, A. W. B. (1991). The products of the Rhizobium genes psi and pss which affect exopolysaccharide production are associated with the bacterial cell surface. Mol Microbiol 5, 2107-2114.[Medline]

LeVier, K. & Guerinot, M. L. (1996). The Bradyrhizobium japonicum fegA gene encodes an iron-regulated outer-membrane protein with similarity to hydroxamate-type siderophore receptors. J Bacteriol 178, 7265-7275.[Abstract]

LeVier, K., Day, D. A. & Guerinot, M. L. (1996). Iron uptake by symbiosomes from soybean root-nodules. Plant Physiol 111, 893-900.[Abstract/Free Full Text]

deLuca, N. G., Wexler, M., Pereira, M. J., Yeoman, K. H. & Johnston, A. W. B. (1998). Is the fur gene of Rhizobium leguminosarum essential? FEMS Microbiol Lett 168, 288-295.

Marie, C., Barnie, M.-A. & Downie, J. A. (1992). Rhizobium leguminosarum has two glucosamine synthases, GlmS and NodM, required for nodulation and development of nitrogen-fixing nodules.Mol Microbiol 6, 843-851.[Medline]

Messing, J., Crea, R. & Seeburg, P. H. (1983). A system for shotgun DNA sequencing. Nucleic Acids Res 9, 309-314.[Abstract]

Moreau, S., Day, D. A. & Puppo, A. (1998). Ferrous iron is transported across the peribacteroid membrane of soybean nodules. Planta 207, 83-87.

Persmark, M., Pittman, P., Buyer, J. S., Schwyn, B., Gill, P. R. & Neilands, J. B. (1993). Isolation and structure of rhizobactin 1021, a siderophore from alfalfa symbiont Rhizobium meliloti 1021. J Am Chem Soc 115, 3950-3956.

Postle, K. (1999). Active transport by customized ß-barrels. Nature Struct Biol 6, 3-6.[Medline]

Reigh, G. & O’Connell, M. (1993). Siderophore-mediated iron transport correlates with the presence of specific iron-regulated proteins in the outer membrane of Rhizobium meliloti. J Bacteriol 175, 94-102.[Abstract]

Rigottier-Gois, L., Turner, S. L., Young, J. P. W. & Armarger, N. (1998). Distribution of repC plasmid-replication sequences among plasmids and isolates of Rhizobium leguminosarum bv. viciae from field populations. Microbiology 144, 771-778.[Abstract]

Rossen, L., Shearman, C. A., Johnston, A. W. B. & Downie, J. A. (1985). The nodD gene of Rhizobium leguminosarum is autoregulatory and in the presence of plant root exudate induces the nodABC genes.EMBO J 4, 3369-3374.

Roy, N., Bhattacharyya, P. & Chakrabartty, P. K. (1994). Iron acquisition during growth in an iron deficient medium by Rhizobium sp. isolated from Cicer arietinum. Microbiology 140, 2811-2820.

Schlaman, H. R. M., Horvath, B., Vijgenboom, E., Okker, R. J. H. & Lugtenberg, B. J. J. (1991). Suppression of nodulation gene expression in bacteroids of Rhizobium leguminosarum biovar viciae. J Bacteriol 173, 4277-4287.[Medline]

Schwyn, B. & Neilands, J. B. (1987). Universal chemical assay for the detection and determination of siderophores. Anal Biochem 160, 47-56.[Medline]

Sharma, S. B. & Signer, E. R. (1990). Temporal and spatial regulation of the symbiotic genes of Rhizobium meliloti in planta revealed by transposon Tn5-gusA. Genes Dev 4, 344-356.[Abstract]

Spaink, H. P., Okker, R. J. H., Wijffelman, C. A., Pees, E. & Lugtenberg, B. J. J. (1987). Promoters in the nodulation region of the Rhizobium leguminosarum symbiotic plasmid pRL1JI. Plant Mol Biol 9, 27-39.

Stevens, J. B., Carter, R. A., Hussain, H., Carson, K. C., Dilworth, M. J. & Johnston, A. W. B. (1999). The fhu genes of Rhizobium leguminosarum specifying siderophore uptake proteins: fhuCDB are adjacent to a pseudogene version of fhuA. Microbiology 145, 593-601.[Abstract]

Tate, R., Cermola, M., Riccio, A., Iaccarino, M., Merrick, M., Favre, R. & Patriarca, E. J. (1999). Ectopic expression of the Rhizobium etli amtB gene affects the symbiosome differentiation process and nodule development. Mol Plant–Microbe Interact 12, 515-525.

Wilson, K. J., Hughes, S. G. & Jefferson, R. A. (1992). The Escherichia coli gus operon: introduction and expression of the gus operon in E. coli and the occurrence and use of GUS in other bacteria. In Gus Protocols: Using the GUS Gene as a Reporter of Gene Expression, pp. 7-22. Edited by S. R. Gallagher. New York: Academic Press.

Wood, W. B. (1966). Host specificity of DNA produced by Escherichia coli; bacterial mutations affecting the restriction and modification of DNA. Mol Biol 16, 118-133.

Yeoman, K. H., Delgado, M. J., Wexler, M., Downie, J. A. & Johnston, A. W. B. (1997). High affinity iron acquisition in Rhizobium leguminosarum requires the cycHJKL operon and the feuPQ gene products which belong to the family of two-component regulators. Microbiology 143, 127-134.[Abstract]

Yeoman, K. H., May, A. G., DeLuca, N. G., Stuckey, D. B. & Johnston, A. W. B. (1999). A putative ECF {sigma} factor gene, rpoI, regulates siderophore production in Rhizobium leguminosarum. Mol Plant–Microbe Interact 12, 994-999.[Medline]

Received 9 November 1999; revised 16 December 1999; accepted 10 January 2000.