Multiple Genetic Controls on Rhizobium meliloti syrA, a Regulator of Exopolysaccharide Abundance

Corresponding author: Sharon R. Long, Howard Hughes Medical Institute, Stanford University, Stanford, CA 94305, fa.srl{at}forsythe.stanford.edu (E-mail).

Communicating editor: J. CHORY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Exopolysaccharides (EPS) are produced by a wide assortment of bacteria including plant pathogens and rhizobial symbionts. Rhizobium meliloti mutants defective in EPS production fail to invade alfalfa nodules. Production of EPS in R. meliloti is likely controlled at several levels. We have characterized a new gene of this regulatory circuit. syrA was identified by its ability to confer mucoid colony morphology and by its ability to suppress the colonial phenotype of an exoD mutant. Here we show that syrA encodes a 9-kD hydrophobic protein that has sequence similarity to two other EPS regulatory proteins: ExoX of Rhizobium NGR234 and R. meliloti, and Psi of R. leguminosarum bv. phaseoli. The syrA transcription start site lies 522 nucleotides upstream of a non-canonical TTG start codon. The syrA promoter region is similar to the promoter region of the nodulation regulatory protein, nodD3. We found that in free-living bacteria, syrA expression is activated by the regulatory locus, syrM, but not by nodD3. In planta, syrM is not required for expression of syrA. Instead, expression of the nitrogen fixation (nifHDKE) genes upstream of syrA plays a role. Specific and distinct sets of genetic controls may operate at different times during nodule invasion.

THE development of a nitrogen-fixing alfalfa nodule by the symbiotic soil bacterium, Rhizobium meliloti (also termed Sinorhizobium meliloti; DE LAJUDIE et al. 1994 ), is a complex process involving sequential changes both in the bacterium and the host plant. Compounds produced by the plant induce expression of rhizobial nod genes (MULLIGAN and LONG 1985 ; PETERS et al. 1986 ). The protein products encoded by these genes synthesize lipo-oligosaccharide signal molecules that cause the earliest events in nodule inititiation: root hair membrane depolarization, root hair deformation, and root cortical cell divisions (DENARIE et al. 1996 ; LONG 1996 ). Later in nodule development, the nod genes do not appear to be expressed (SCHLAMAN et al. 1991 ; SHARMA and SIGNER 1990 ).

The interaction of a bacterial cell with a root hair results in the formation of an inwardly growing tunnel of plant cell wall. The bacteria proliferate in this infection thread, penetrating through the root hair into the root cortex (reviewed by BREWIN 1991 ; HIRSCH 1992 ; KIJNE 1992 ). After release from the infection thread, bacteria differentiate morphologically into bacteroids. These developmental changes involve the eventual cessation of cell division and DNA replication and the induction of genes necessary for nitrogen fixation. The nif genes encode the subunits of nitrogenase and other genes necessary for nitrogenase function. The fix genes encode functions unique to symbiotic nitrogen fixation.

In other free-living, nitrogen-fixing bacterial species, regulation of the nif and fix genes occurs via the ntrA and ntrC gene products in response to nitrogen levels (reviewed in DE BRUIJN et al. 1990 ; MERRICK 1992 ). However, induction of the nif and fix genes occurs differently in R. meliloti during symbiosis (DE BRUIJN et al. 1990 ; MERRICK 1992 ). NtrC is not required; instead, nif/fix gene expression is regulated in response to oxygen levels via the two-component regulators, FixL and FixJ (DAVID et al. 1988 ). FixL senses oxygen tension and phosphorylates FixJ, which activates nifA and fixK (DAVID et al. 1988 ; BATUT et al. 1989 ). NifA and NtrA activate expression from nif and fix promoters (GUSSIN et al. 1986 ; SZETO et al. 1984).

nod gene expression requires a positive transcriptional activator, NodD. NodD proteins are members of the LysR activator family and bind to nod boxes, conserved regions upstream of the nod operons. Most Rhizobium species possess multiple nodD genes. The R. meliloti genome encodes three different NodD proteins (GYORGYPAL et al. 1988 ; HONMA and AUSUBEL 1987 ). NodD1 and NodD2 activate nod promoters in concert with inducers synthesized by the host plant (MAXWELL et al. 1989 ; MULLIGAN and LONG 1985 ; PETERS et al. 1986 ; PHILLIPS et al. 1992 ). NodD3 acts independent of plant inducer compounds and requires another LysR family member, SyrM, for nod gene expression (MULLIGAN and LONG 1989 ; SWANSON et al. 1993 ). The nod genes appear to be expressed early in nodule initiation, whereas syrM appears to be produced at intermediate stages of development (SHARMA and SIGNER 1990 ; SWANSON et al. 1993 ).

The regulatory circuit controlling nod gene expression is probably most important during the early stages of symbiosis, whereas the nif/fix circuit is required later in nodule development. During intermediate stages of development, other genes are required for successful invasion of a developing nodule. For example, exo genes, which encode biosynthetic enzymes necessary for the production of acidic exopolysaccharide (also called EPS I or succinoglycan) are necessary for R. meliloti to invade alfalfa nodules (reviewed by LEIGH and WALKER 1994 ). EPS I consists of repeating octasaccharide subunits assembled on a lipid carrier, then polymerized and secreted (AMAN et al. 1981 ; REUBER and WALKER 1993 ).

R. meliloti strain SU47 can produce a second exopolysaccharide, EPS II or galactoglucan, which can substitute for EPS I in nodule invasion of alfalfa, but not in invasion of other plant hosts (BECKER et al. 1997 ; GLAZEBROOK and WALKER 1989 ; KELLER et al. 1995 ; ZHAN et al. 1989 ). EPS II is only made in Exo^- mutant strains when phosphate is limiting, or in strains overexpressing the EPS II biosynthetic (exp) genes (GLAZEBROOK and WALKER 1989 ; ZHAN et al. 1989 ; ZHAN et al. 1991 ).

Many aspects of exo and exp gene regulation remain mysterious. Nitrogen or phosphate limitation stimulates EPS I production, but otherwise the physiological signals that regulate expression of these genes are not known. Exo structural genes are expressed during nodule invasion and their expression decreases during bacteroid maturation (REUBER et al. 1991 ). The regulatory genes that control expression of the exo structural genes include exoR and exoS which negatively affect the amount of EPS I, probably by repressing genes in the exo synthesis pathway (DOHERTY et al. 1988 ; REED et al. 1991B ; REUBER et al. 1991 ). ExoX, a small, probably membrane-associated protein, exerts negative effects on EPS I production, perhaps by a posttranslational mechanism involving ExoY (REED et al. 1991A ; ZHAN and LEIGH 1990 ). expR represses genes in the EPS II biosynthetic pathway (GLAZEBROOK and WALKER 1989 ). The exp genes are affected positively by mucS and negatively by mucR and expR (ASTETE and LEIGH 1996 ; GLAZEBROOK and WALKER 1989 ; ZHAN et al. 1989 ). mucR may also control EPS I synthesis (ZHAN et al. 1989 ).

Another locus that affects the amount of EPS I is syrA. syrA carried on a low-copy plasmid, pRmJT5, that contained about 20 kb of R. meliloti DNA, conferred a mucoid colony phenotype to Rm1021 (MULLIGAN and LONG 1989 ). An insertion in either of two loci, syrM or syrA, abolished this mucoid phenotype (MULLIGAN and LONG 1989 ). Because SyrM activates transcription of nodD3, it was hypothesized to activate transcription of syrA as well (MULLIGAN and LONG 1989 ; SWANSON et al. 1993 ).

LEIGH et al. 1985 independently isolated a plasmid carrying syrA based on its ability to suppress the calcofluor-dim phenotype of an exoD mutant. This plasmid failed to compensate for the invasion defects of the exoD mutant strain; instead, exoD mutants could invade nodules if the plant growth media was buffered to a slightly acidic pH (REED and WALKER 1991B ). These and other observations led to the hypothesis that exoD encodes a membrane protein necessary for the bacteria to survive the alkaline conditions encountered during nodule invasion and that the effects of exoD mutations on EPS production are probably indirect (REED and WALKER 1991B ).

We report here that syrM is necessary and sufficient to activate expression of syrA, which in turn confers the mucoid-colony phenotype: nodD3 participates only indirectly as an activator of syrM expression. We determined that the requirement for syrM in activating syrA could be obviated by cloning DNA containing the syrA locus downstream of a strong exogenous promoter. We used mutational and DNA sequence analysis to locate the syrA open reading frame. syrA is a small hydrophobic protein possessing sequence similarity to other proteins involved in EPS biosynthesis.

We also found that although syrM activates expression of syrA-gusA gene fusions in free-living cells, syrM is not required for the expression of syrA in bacterioids. Instead, syrA expression is dependent on expression of an upstream nifHDK operon.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plasmids and strains:
All plasmids and strains used in this study are listed in Table 1. R. meliloti strains were grown in TY (BERINGER 1974 ) or on LB agar plates. Escherichia coli strains were grown in LB.

Bacterial genetic techniques:
Broad-host range plasmids were transferred into R. meliloti strains via triparental conjugation using pRK2013 (DITTA et al. 1980 ) or pRK600 (FINAN et al. 1986 ) as helper plasmids. Replacement of the Nm^r of Tn5 with Sp^r/Gm^r of Tn5-233 was performed as previously described (DE VOS et al. 1986 ). N3 phage transduction was used to construct strains with multiple insertions (MARTIN and LONG 1984 ).

Plasmid constructions:
The 1.8-kb BamHI-PvuII from pM144 was inserted into BamHI-SmaI digested pUC119 to create pMB54. This insert was removed as an EcoRI-HindIII fragment and cloned into pLAFR3 to make pMB57. A 3.1-kb BamHI-Bgl II partial digestion product was cloned into the BamHI site of pTE3 to create pMB51 and pMB52 and into pLAFR3 to create pMB56. The insert in pMB51 is oriented such that nodD3 is expressed from the lac promoter, whereas in pMB52 it is oriented such that syrA is expressed from the lac promoter. On pMB56, nodD3 is expressed from the trp promoter. pMB55 contains the 2.6-kb SphI insert from pMB2 (BARNETT and LONG 1990 ) in the SphI site of pUC1318 (KAY and MCPHERSON 1987 ). The 2.6-kb BamHI fragment from pMB55 was cloned into the BamHI site of pMB57 to create pMB64 and into the BamHI site of pMB56 to create pMB61. A 1.1–kb StyI deletion derivative of pMB54 was cloned as an EcoRI-HindIII fragment, into pLAFR3 that had been digested with EcoRI and HindIII to make pMB131. A HindIII-linkered derivative of the 2.6–kb BamHI insert was cloned into the HindIII site of pMB131 to make pMB135.

Construction of pTE3 expression clones:
The 1.8–kb insert of pMB54 was removed with digestion by EcoRI and HindIII and blunted with Klenow enzyme. BamHI linkers were added and the fragment was ligated into the BamHI site of pTE3 in both orientations to form pMB85 and pMB86. Inserts from deletion derivatives of pMB54 were cloned into pTE3 in a similar fashion. The 0.55–kb ClaI-Bgl II fragment from pMB70 was blunted and BamHI linkers were added before ligation into the BamHI site of pALTER to form pMB308. Mutant derivatives of pMB308 were cloned as BamHI fragments into the BamHI site of pTE3.

Assays for mucoid phenotype:
Strains were streaked to single colonies on M9 sucrose plates (MEADE and SIGNER 1977 ) with appropriate antibiotic selection. At 5 days' growth, colonies were scored for mucoid morphology.

Site-directed mutagenesis:
Site-directed mutagenesis was performed using an Altered Sites mutagenesis kit (Promega, Madison, WI) Double-stranded miniprep DNA was made from ampicillin-resistant colonies and screened by DNA sequencing for presence of the introduced mutation. The following oligonucleotides were used: Nucleotide 948 CA, 5'-GCTGGCTTGATTGCTGCTC; 838 TA, 5'-CGTCAT TGACCTTCAGA; 882 CT, 5'-CTGTGCGCTAGTTCTC TCG; 948 CTTC, 5'-CGACGCTGGCTTGTGTGCTGCTCT TCC; 837 GA, 5'-GGAGAACGTCATTATCCTTCAG; 835 TA, 5'-GGAGAACGTCAATGTCCTTCAG

DNA sequence analysis:
The 1.1–kb BamHI-BglII fragment from pM144 was sequenced on both strands using Sequenase (United States Biochemical, Cleveland). This sequence has been assigned Genbank accession U90221. That the Bgl II site of this fragment represents the same Bgl II site of the 2-kb Bgl II fragment containing nodD3 (RUSHING et al. 1991 ) was confirmed by synthesizing a specific primer and sequencing across the Bgl II site of pM144. DNA sequence was analyzed using the University of Wisconsin GCG software (DEVEREUX et al. 1984 ). A codon usage table of 55 R. meliloti SU47 genes and the program Codonpreference of the GCG software were used to generate codon usage profiles. R. meliloti DNA containing Tn3 insertions was subcloned as Bgl II fragments into pUC119 and the insertion points were determined using a primer specific to the Tn3 end to sequence across the junction. The insertion point of Tn5 29b was determined in a similar fashion using a primer specific to the Tn5 end. The locations of Tn5 No. 1005 and No. 213 were reported by RUSHING et al. 1991.

Primer extension analysis of syrA transcript:
RNA was isolated from a strain that has syrM, nodD3 and syrA present on a low copy number plasmid, 1021 pRmJT5 (SWANSON et al. 1987 ). RNA purification and primer extension mapping was done according to BARNETT et al. 1996. Two different oligonucleotides were used to map the syrA start site. One was complementary to a region 108 nucleotides downstream of the syrA transcription start site: 5'-CCACGATCCGCAGAAAT CTTGAGCTCGGGTAAGCGGCG. The other was located 547 nucleotides downstream of the syrA transcription start site in the syrA open reading frame: 5'-GTAGATTGCGAGAGA ACTGGCGCACAGGAGCAGCCATAG.

Construction of syrA::gusA fusion strains:
Mutagenesis of plasmid pM136 was performed as previously described by STACHEL et al. 1985 and SWANSON et al. 1993 . Mutagenized plasmids were screened for insertions in syrA by conjugating into Rm1021 and looking for colonies that were less mucoid. syrA gene fusions were marker exchanged into the R. meliloti genome using pPH1JI (HIRSCH and BERINGER 1984 ).

ß-glucuronidase assays:
TY grown cultures were assayed for ß-glucuronidase activity at mid-log phase as previously described (SWANSON et al. 1993 ). For in planta assays, nodules were sectioned and stained as described by SWANSON et al. 1993 . Nodules were counted and observed every few days from 10 days post inoculation (dpi) to 45 dpi.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

syrM, nodD3 and syrA: effects on mucoid phenotype:
Our previous data suggested that both syrM and syrA are required for the mucoid phenotype, with SyrM most likely acting as a positive activator of syrA. Overproduction of acidic exopolysaccharide (EPS I) is presumed to be responsible for the mucoid phenotype since genes in the EPS I synthesis pathway are required (MULLIGAN and LONG 1989 ).

We defined the requirements for the mucoid phenotype by constructing plasmids containing various combinations of syrA, syrM and nodD3 (Figure 1; Table 2). Any plasmid conferring a mucoid colony morphology that was qualitatively similar to that conferred by pJT5 was scored as positive. pS701 (pJT5, syrM::Tn5), pS303 (pJT5, nodD3::Tn5) and pM111 (pJT5, syrA::Tn5) were sufficient to confer a mucoid phenotype as reported previously (MULLIGAN and LONG 1989 ). pMB61, which contains only syrM, nodD3 and syrA, also conferred a mucoid phenotype. nodD3 is not required for the mucoid phenotype: strains containing plasmids with only syrM and syrA were still mucoid (pMB64). syrA alone on a plasmid (pMB57) was not sufficient to confer the mucoid phenotype. Likewise, syrM alone on pM113 or pS73 did not confer a mucoid phenotype.

View larger version (22K): [in this window] [in a new window]   Figure 1. —Physical map of the syrA region and clones used for genetic analyses. (A) Restriction map of nod, syr and nif region of pSymA. Extent of each plasmid insert is represented by a line above the restriction map. pJT5 and pM113 contain additional sequence upstream of nodH not shown on map (indicated by dashes). Locations of relevant transposon insertions are marked with arrowed triangles. (B) Restriction map of the 1.8–kb PvuII-BamHI fragment containing syrA. Potential ORFs are marked with thin arrows. syrA-gusA insertions are represented by arrowed triangles; horizontal arrows within each triangle indicate the direction of gusA transcription. (C) Deletion derivatives of the 1.8–kb PvuII-BamHI fragment shown in (B). The extent of each plasmid insert relative to the map in (B) is shown along with the corresponding mucoid phenotype. Each clone was tested in both orientations with respect to the exogenous trp promoter as described in MATERIALS AND METHODS. The shaded region denotes the smallest region able to confer the mucoid phenotype as defined by these deletion derivatives. Abbreviations: S, SphI; Bg, Bgl II; Pv, PvuII; B, BamHI; P, Pst I; St, StyI, H, HpaI; C, ClaI; Ss, SstI; Bs, BstE II; A, ApaI.

We tested a possible role for nodD3 in determining the mucoid phenotype using pMB51, pMB52 and pMB56, which each contain nodD3 and syrA on a 3.1–kb Bgl II-BamHI insert expressed from different exogenous promoters (see MATERIALS AND METHODS). All of these plasmids confer a mucoid phenotype to 1021 (Figure 1; Table 2, lines 12–14), but only in strains containing a normal genomic copy of syrM (data not shown). Therefore, we conclude the mucoid phenotype conferred by these plasmids is an indirect result of nodD3 increasing expression of syrM. This is supported by the observation that neither, pM136, which contains nodD3 and syrA but no exogenous promoter, nor pE65, which contains only nodD3 strongly expressed from a vector promoter, was able to confer a mucoid phenotype (Table 2, lines 10 and 11).

We determined whether the syrM-syrA mucoid phenotype was dependent on the presence of the syrA upstream region. A strain containing a plasmid-born syrA but with a deletion from the StyI site to BamHI site was not mucoid (Table 2, line 16). Cloning syrM into this plasmid failed to restore the mucoid phenotype (Table 2, line 17). But, when syrM is cloned into a plasmid containing the intact syrA upstream region a mucoid phenotype is seen (Table 2, lines 6 and 15), indicating the importance of a cis-acting region upstream of syrA.

The requirement for plasmid-born copies of syrM could be circumvented by overexpressing syrA from the trp promoter. We took advantage of this syrM-independent expression to define the syrA locus further. Deletion derivatives of the 1.8–kb BamHI-PvuII DNA were cloned in both orientations into the trp promoter expression vector pTE3 (EGELHOFF and LONG 1985 ; Figure 1). These plasmids were conjugated into R. meliloti and scored for their mucoid phenotype.

An example of such a screen for pMB89 is shown in Figure 2. pMB89 contains the ClaI-PvuII fragment of Figure 1B oriented such that the transcription from the trp promoter proceeds in the direction from the ClaI site to the PvuII site. R. meliloti colonies containing this plasmid are extremely mucoid. pMB90, which contains the insert in the opposite orientation, confers no mucoid phenotype. The combined results presented in Figure 1C indicate that the region contained within a 280 bp StyI-Bgl II fragment, shown by shading, is sufficient to confer mucoid colony morphology when expressed from the trp promoter of pTE3.

View larger version (140K): [in this window] [in a new window]   Figure 2. —Plate assay of the syrA mucoid phenotype. The photograph shows the mucoid phenotype of R. meliloti strain Rm1021 containing either pMB89 (left) or pMB90 (right). pMB89 contains the ClaI-PvuII fragment (Figure 1, B and C) oriented such that syrA is overexpressed via the exogenous trp promoter. pMB90 contains the same insert in the opposite orientation.

Nucleotide sequence of syrA:
The 1.1–kb BamHI-Bgl II fragment shown in Figure 1B is adjacent to the nodD3 sequence reported earlier by RUSHING et al. 1991 . Results from the genetic analyses described above, and from gene fusion data, indicated that syrA was probably expressed from right to left as shown in Figure 1B.

Three potential open reading frames (ORFs) inferred from the sequence are shown by thin arrows in Figure 1B. The ORF closest to the BamHI site encodes a polypeptide of 10.5 kD and has an R. meliloti-like codon usage pattern. Tn5 #29b, which abolishes the mucoid phenotype, is located just upstream, and Tn3 #4–11 is located in the N-terminal region of this ORF (Figure 3). However, a region containing this ORF is not sufficient to confer a mucoid phenotype (Figure 1C, pMB81), nor does its removal abolish the mucoid phenotype (Figure 1C, pMB137 and pMB132). In vitro transcription/translation experiments failed to detect a protein corresponding to this ORF. A larger ORF encoding a potential protein of 18.5 kD spans a region outside of the 280–bp StyI-Bgl II segment. Since only the N-terminal half of this protein would be contained within the 280–bp segment, this putative protein is not a likely candidate for SyrA.

View larger version (37K): [in this window] [in a new window]   Figure 3. —Sequence of 1.1–kb BamHI-Bgl II fragment containing the syrA locus (Genbank accession U90221). Restriction sites are those shown in Figure 1A. The amino acid sequences of the ORF encoding a putative 10.5 kD protein and of SyrA are shown below the nucleotide sequence. Transposon insertion points determined by DNA sequencing are shown with triangles. A potential ribosome binding sequence for syrA is underlined. A region upstream of the 10.5 kD ORF similar to a region upstream of nodD3 is marked with # symbols. Each codon containing a point mutation is shown above the nucleotide sequence line. The corresponding amino acid change in SyrA is shown below the amino acid sequence line. Mutations represented by circled codons abolished the mucoid phenotype. Boxed codons represent mutations that had no effect on the mucoid phenotype. All point mutations were created as stated in MATERIALS AND METHODS and were confirmed by DNA sequencing.

Contained entirely within the 280–bp StyI-Bgl II fragment is a smaller ORF that encodes a putative protein of 9 kD. This ORF corresponds to the region whose overexpression is responsible for the syrA phenotype. This 81-amino acid ORF begins with a TTG (leucine) start codon. Upstream lies a good match with a consensus ribosome binding sequence (underlined in Figure 3). We obtained a Tn3::gusA fusion, #1–24, that is located in the carboxy terminus of the protein product of this ORF.

Site-directed mutagenesis of syrA:
Because this putative ORF begins with a nonconventional start-codon and lacks a strong R. meliloti codon usage pattern, we sought to confirm that this ORF encodes SyrA. We used site-directed mutagenesis to introduce point mutations into the syrA region as described in MATERIALS AND METHODS. These mutated DNAs were cloned in vector pTE3 and assayed for mucoid phenotype as described earlier. The location of each introduced change is shown in Figure 3. Mutations that abolish the mucoid phenotype are circled; those that do not are boxed.

One such mutation, a C to T change at nucleotide 882, introduces a stop codon in frame with the ORF encoding the 18.5-kD protein, but is silent with respect to the 9-kD ORF. A syrA plasmid containing this mutation still confers the mucoid phenotype. This confirms that the 18.5 kD ORF is not required for the mucoid phenotype.

Another mutation (948 C to A) introduces a stop codon in the 9-kD ORF and abolishes the mucoid phenotype. To confirm that this phenotypic change results from a truncation of the SyrA protein and not some other reason, such as an alteration of a binding site or a mutation of a nontranslated RNA, we made another mutated derivative which has a CT to TC change at this same position (948-949), but which is silent at the amino acid level. The mucoid phenotype was observed in this mutant.

Although no candidate ORFs were found in the third reading frame, we nonetheless tested it by the introduction of a stop codon (837 G to A). This mutation introduces a conservative serine to threonine change in the 9-kD ORF and a nonconservative valine to glutamate change in the 18.5-kD ORF. The plasmid carrying this mutation still confers a mucoid phenotype to R. meliloti.

The above results strongly support the hypothesis that the 9-kD ORF encodes SyrA. We created two additional mutations to confirm that this ORF begins at the TTG codon. First, we changed the TTG leucine codon of the putative syrA ORF to a TTA leucine codon (Figure 3). TTA codons specify leucine, but have never been shown to be translational start codons. The mucoid phenotype was lost with a plasmid (pMB372) carrying this mutation. We also made a mutation in which SyrA starts with an ATG instead of a TTG codon. pMB374 confers a mucoid phenotype to R. meliloti similar to that of a plasmid carrying wild-type syrA. These results confirm that SyrA initiates translation from a TTG codon.

SyrA has features similar to ExoX proteins:
A TFASTA search of the Genbank database showed SyrA is 34% identical to ExoX from the broad host range Rhizobium sp. NGR234 (GRAY et al. 1990 ), 26% identical to ExoX from R. meliloti (REED et al. 1991A ), and 33% identical to Psi from R. leguminosarum biovar phaseoli (BORTHAKUR and JOHNSTON 1987 ) (Figure 4). All of these proteins have in common a hydrophobic N-terminal region and a hydrophilic C-terminus.

View larger version (34K): [in this window] [in a new window]   Figure 4. —Amino acid sequence alignment of SyrA protein. The amino acid of SyrA is aligned with the amino acid sequences of Rhizobium meliloti ExoX (REED et al. 1991), Rhizobium NGR234 ExoX (GRAY et al. 1990 ), and R. leguminosarum bv. phaseoli Psi (BORTHAKUR and JOHNSTON 1987 ). Amino acids of the ExoX and Psi proteins identical or conserved in SyrA are shown shaded. Conserved groupings used in the alignment are as follows: Leu, Val, Ile; His, Lys, Arg.

Determination of transcription start site:
We mapped the transcription start site as described in MATERIALS AND METHODS. Our data reveal a single start site 522 nucleotides upstream of the TTG translation start site (Figure 3 and Figure 5). The location of this start site may mean that the 10.5-kD ORF upstream of syrA is cotranscribed with syrA. Alternatively, syrA may share a feature with syrM and nodD3: there is no evidence for the existence of ORFs upstream of syrM and nodD3; yet, these genes possess long leader sequences (BARNETT et al. 1996 ).

View larger version (48K): [in this window] [in a new window]   Figure 5. —syrA transcription initiation site. Autoradiogram of primer extension products (P) and sequencing reactions (A, C, G and T, respectively). The extension product shown here was obtained using a primer complementary to a region 108 nucleotides downstream of the syrA start site (MATERIALS AND METHODS). pM149 (J. MULLIGAN, unpublished results) was used as a template for double-stranded sequencing using the same primer as for the primer extension.

Immediately upstream of the start site is a 68 bp region that is 67.6% identical to a region upstream of the nodD3 start site (marked with # symbols in Figure 3). Because expression of both nodD3 and syrA is affected by syrM, this sequence may be required for interaction of SyrM with these promoters. Additional experiments are necessary to determine if this sequence represents a SyrM binding site.

syrM, nodD3, and syrA: effects on syrA-gusA reporter gene fusions:
We used the syrA-gusA gene fusions described above to assay effects of activators carried in trans (Table 2). Both of the syrA-gusA fusions have basal levels of GUS activity that are unaffected by mutation of syrM (BARNETT 1994 ). Plasmid pJT5 greatly enhanced expression of these syrA fusions, and plasmid copies of syrM, but not nodD3 or syrA, were required for pJT5 to confer these high levels of GUS activity (Table 2, lines 3–5).

Smaller plasmids containing syrM also activated syrA expression (Table 2, plasmids 6–8). nodD3 increased activity of the syrA fusions, but only when expressed in certain contexts. nodD3, expressed from its own promoter, failed to activate syrA expression (Table 2, line 9). In pE65, nodD3 is expressed from the trp promoter and in pMB56 nodD3 is expressed from the lac promoter (Table 2, lines 10 and 14). syrA-gusA gene fusion strains containing these plasmids had high GUS activity. As was the case with the mucoid phenotype, nodD3 appears to enhance syrA expression indirectly via syrM. pE65 and pMB56 do not increase activity of the syrA reporter fusions when assayed in strains containing a syrM insertion (BARNETT 1994 ).

It was previously reported that a plasmid expressing syrA can suppress the calcofluor-dim phenotype of an exoD mutant (REED and WALKER 1991A ). We found that a syrA plasmid does not affect expression of exoD::lacZYA gene fusions. However, exoD may have subtle effects on syrA expression. We found that a plasmid containing exoD causes about a sevenfold reduction in expression of the 4–11 syrA fusion, but not the 1–24 fusion (Table 2, line 18).

syrA does not appear to affect its own expression as strains containing syrA alone on a plasmid showed basal levels of expression (Table 2, pMB57). Additional evidence for lack of syrA autoregulation is the observation that a syrA::Tn5 insertion on plasmid pJT5 did not adversely effect syrA expression (Table 2, line 5).

Symbiotic expression of syrA-gusA fusions:
syrA mutants form normal nitrogen-fixing nodules on alfalfa. We tested the 1–24 and 4–11 syrA-gusA fusions for in planta ß-glucuronidase (GUS) activity. Alfalfa plants were inoculated with strains containing these fusions. Nodules were harvested at various time points, sectioned, stained for GUS activity and observed by microscopy. For the strains containing a syrA fusion in an otherwise wild type background, we observed intense staining in bacteroid-containing cells, shown in Figure 6, A and B. To our surprise, we still observed high levels of GUS staining when a mutation was introduced in syrM (Figure 6, C and D). This is contrary to the situation in free-living cells where syrM is required for activation of syrA (Table 2).

View larger version (64K): [in this window] [in a new window]   Figure 6. —Symbiotic expression of syrA-gusA gene fusions. (A–J) Photomicrographs of hand-sectioned nodules stained for ß-glucuronidase activity (see MATERIALS AND METHODS). (A) JAS303 (Tn3-gusA 1–24) 22 days post inoculation (dpi), (B) JAS304 (Tn3-gusA 4–11) 23 dpi, (C) JAS142 (Tn3-gusA 1–24, syrM::Tn5) 12 dpi, (D) JAS143 (Tn3-gusA 4–11, syrM::Tn5) 25 dpi, (E) JAS144 (Tn3-gusA 1–24, ntrA::Tn5) 38 dpi, (F) JAS145 (Tn3-gusA 4–11, ntrA::Tn5) 38 dpi, (G) MB310 (Tn3-gusA 1–24, fixH::Tn5–233) 22 dpi, (H) MB309 (Tn3-gusA 4–11, fixH::Tn5–233) 22 dpi, (I) MB304 (Tn3-gusA 1–24, nif D::Tn5–233) 20 dpi, (J) MB303 (Tn3-gusA 4–11, nif D::Tn5–233) 20 dpi.

We wondered if perhaps an additional mechanism for activation of syrA occurs in the symbiotic state. NtrA is necessary for the activation of diverse functions including nitrogen fixation, dicarboxylic acid metabolism, and growth on nitrate (RONSON et al. 1987 ). An ntrA mutation had no effect on syrA activity when assayed in culture (data not shown). However, when nodules inhabited by syrA-gus strains containing a mutation in ntrA were assayed, no visible staining was detected (Figure 6E) with one exception: in less than 10% of the cases we observed faint clusters of stained cells in young nodules inhabited by 4–11 fusion strains. One such nodule is shown in Figure 6F. FixL is required for expression of nifA which in turn acts with NtrA to activate symbiotic promoters. A fixL insertion had a similar phenotype as the ntrA insertion except that we never observed even faint or patchy staining (data not shown). This supports the conclusion that a component of the NtrA-NifA circuit is necessary for symbiotic expression of syrA.

Several models can account for the lack of syrA expression in an ntrA mutant background. Although NtrA^- bacteria are capable of invasion, they form Fix^- nodules and senesce prematurely. It is possible that syrA is only active in mature, nitrogen-fixing bacteroids and that NtrA is not directly required for syrA activation. On the other hand, NtrA may be more directly involved either by acting on the syrA promoter itself or by acting on an intermediate, which then acts on syrA. Another possibility is that syrA is transcribed from a different, NtrA-dependent promoter in bacteroids.

In order to test this hypothesis, we transduced a fixH insertion into syrA::gusA strains. fixH is located on pSyma, about 200 kb from syrA. The fixGHIS operon is proposed to encode a redox-coupled cation pump (KAHN et al. 1989 ). fixH mutants form ineffective nodules: the bacteroids are not able to fix dinitrogen and senesce prematurely (VASSE et al. 1990 ). In nodules formed by these syrA-gusA, fixH mutant strains, we observed a GUS⁺ phenotype (Figure 6, G and H). Appearance of this phenotype was developmentally delayed by several days compared to the fixH⁺ nodules and was qualitatively less intense. These differences may reflect a decreased ability to invade, or the premature senescence of the fixH mutated bacteria.

The above results indicate that syrA expression does occur in the absence of nitrogen fixation. As a first step in dissecting the pathway of syrA activation, we tested the effect of an insertion in the nif D locus (RUVKUN et al. 1982 ). Nif D is part of the NtrA-activated nifHDKE operon and encodes the MoFe protein subunit of nitrogenase. NifD^- strains are Fix^- (RUVKUN and AUSUBEL 1981 ), but nodules induced by these strains more closely resemble wild-type than those induced by NtrA mutants (HIRSCH et al. 1983 ). We chose a nif D mutation because of the proximity of the nif HDKE operon to syrA (Figure 1A). The entire nucleotide sequence of this operon has not been reported, but genetic and physical analyses indicate that nif HDKE is transcribed in the same direction as syrA and may end 1–2 kb upstream of syrA. Staining of nodules inhabited by the NifD mutant strains was similar to that of NtrA mutant strains (Figure 6, I and J). No staining was observed except for the occasional faint staining of young nodules containing nif D::Tn5-233, syrA::gusA 4–11 strains. Similar to the ntrA insertion, the nif D insertion had no effect on expression of syrA in free-living bacteria (data not shown).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We identified the syrA locus by the phenotype it confers when present on a plasmid: bacterial strains with multiple copies of syrA form mucoid colonies on agar plates (MULLIGAN and LONG 1989 ). Mutations in exoA abolish the ability of SyrA to confer a mucoid phenotype (MULLIGAN and LONG 1989 ). This provides indirect evidence that the mucoid colonies seen with syrA overexpression are overproducing EPS I. A single Tn5 insertion located downstream from nodD3 abolishes the syrA phenotype (MULLIGAN and LONG 1989 ).

Here we report the identification and characterization of the syrA ORF, and analyze expression using syrA-gusA gene fusions. Our results confirm that SyrM activates syrA expression resulting in a mucoid colony phenotype. In addition, we showed that nodD3 is not required for the activation of syrA by SyrM. Plasmids containing only syrM, activated expression of the syrA-gusA fusions 30- to 100-fold above background levels, but failed to confer a mucoid phenotype when tested in a wild-type background. Only those strains containing multiple copies of syrA were mucoid. This may indicate that an excess of SyrA is needed before EPS I abundance increases.

The need for SyrM in conferring the syrA-mediated mucoid phenotype can be overcome if syrA is expressed from an exogenous promoter. We tested various segments of DNA for their ability to confer a mucoid colony morphology in R. meliloti, and found that a 280–bp piece of DNA was sufficient. Using site-directed mutagenesis, we confirmed that an ORF, which begins with a TTG start codon, expresses a 9-Kd protein responsible for the mucoid colony morphology. It is estimated that ~9% of all known prokaryotic ORFs begin with start codons other than ATG, and that ~10% of these begin with TTG (GUALERZI and PON 1990 ). It has been postulated that rare initiation codons are targets for regulatory mechanisms directed at select genes (GUALERZI and PON 1990 ). Whether the TTG initiation codon of the syrA gene serves as such a target remains to be proven.

We mapped the transcription start site of syrA using RNA prepared from cells overexpressing syrM, nodD3 and syrA. Our data show a single start site located 522 nucleotides upstream of the syrA TTG codon. All insertions that abolish the syrA phenotype, as well as the 10.5 kD ORF, are located downstream of the start site. This biochemical data is consistent with our genetic data: both gusA insertions in the syrA region are activated by SyrM. Upstream of the start is a region with similarity to the nodD3 promoter region (BARNETT et al. 1996 ). This conservation suggests these cis sequences are important for control by syrM.

SyrA is similar to other proteins involved in exopolysaccharide production: ExoX of R. meliloti (REED et al. 1991A ) and Rhizobium sp. NGR234 (GRAY et al. 1990 ), and Psi from R. leguminosarum bv. phaseoli (BORTHAKUR and JOHNSTON 1987 ). ExoX and Psi are hypothesized to be membrane proteins (BORTHAKUR and JOHNSTON 1987 ; GRAY et al. 1990 ; LATCHFORD et al. 1991 ). The predicted hydropathy of SyrA is greater than either ExoX or Psi. The SyrA protein contains only 7 charged residues, all but one of which are confined to the last 17 residues.

ExoX and Psi inhibit EPS synthesis, an effect opposite that of SyrA (BORTHAKUR et al. 1985 ; GRAY et al. 1990 ; REED et al. 1991A ). The effects of ExoX and Psi on EPS I synthesis are thought to occur posttranslationally (BORTHAKUR et al. 1988 ; GRAY and ROLFE 1992 ; LATCHFORD et al. 1991 ). SyrA does not significantly alter expression of two genes in the EPS I synthesis pathway, exoF and exoP (BARNETT 1994 ). Therefore, it seems likely that SyrA effects on EPS synthesis are posttranslational as well.

REED and WALKER 1991A showed that the calcofluor-dim phenotype of an exoD mutant is suppressed when syrM and syrA are present on a plasmid. The mechanism by which this suppression occurs is unknown. ExoD mutants produce EPS identical in structure to EPS produced by wild-type strains, but in lower amounts (REED and WALKER 1991A ). Even though exoD is probably not directly involved in EPS synthesis, exoD appears to be necessary for the syrA-mediated mucoid phenotype: an exoD mutant containing a plasmid overexpressing syrA was calcofluor-bright, but not mucoid (data not shown). exoD only slightly affects syrA expression, as tested using syrA-gusA fusion strains, and overexpression of syrA has no effect on expression of an exoD::lacZ fusion (data not shown). Therefore, if there is an interaction between exoD and syrA, it may occur posttranslationally.

We found that syrM is not required for symbiotic expression of syrA. Instead, ntrA and nif D are necessary. The observations that a nif D mutation abolishes syrA expression, whereas a fixH mutation does not, support the hypothesis that something specific in the NtrA circuit is required for syrA expression rather than nitrogen fixation per se. Our observation that an insertion in nif D reduces expression of syrA, weakens the case for direct activation of syrA by NtrA. Moreover, sequence analysis of an 830 bp region upstream of syrA failed to reveal an ntrA-ntrC/ntrA-nifA consensus sequence or a nifA upstream activating sequence (GUSSIN et al. 1986 ). It is possible that the insertion in nif D has polar effects on syrA transcription. The exact distance between the end of the nifHDKE operon and the translational start of syrA is not known, but is approximately 2.5-kb and other genes may map to this region. Preliminary sequence analysis has identified an ORF with similarity to a ferredoxin-like protein from Rhodobacter capsulatus located about 1 kb from the syrA translational start (data not shown). In R. capsulatus this ferredoxin is transcribed with the nif ENX operon (MORENO-VIVIAN et al. 1989 ). The proximity of the nif operon to syrA is the basis for another model which explains the nif D-dependence of syrA expression. In this model, syrA can be expressed via two promoters, a syrM-dependent promoter located upstream of syrA and the ntrA-nifA-dependent promoter upstream of the nif HDKE operon. In the free-living state, expression occurs when syrM is overexpressed. In the symbiotic state expression occurs via transcriptional read-through from the nif promoter. Because NtrA is necessary for activation of the nif HDKE promoter (RONSON et al. 1987 ), mutations in both nif D and ntrA would abolish transcription of the nif operon and therefore of syrA. A similar situation has been observed for the expression of NifA in R. meliloti: nifA can be expressed from its own promoter as well as from the promoter of the upstream fixABC operon (KIM et al. 1986 ). Additional experiments are required to determine if syrA is expressed from a nif promoter and, if so, to determine the functional consequence of coupling expression of a gene involved in EPS production to expression of those involved in nitrogen fixation.

	ACKNOWLEDGMENTS

We thank H. CHRISPEELS and M. WILLITS for subcloning Tn5 29b and J. MULLIGAN for plasmids pM144, pM149 and pM150. We are grateful to F. M. AUSUBEL, J. DÉNARIÉ and G. C. WALKER for strains and plasmids. We thank L. ZUMSTEIN for pD3-25 and D. GAGE for the fixH::Tn5 strain. We thank R. FISHER and D. GAGE for their comments on this manuscript. M.J.B. was supported by a National Institute of Health (NIH) Training Grant in Cell and Molecular Biology to Stanford University. S.R.L. is an investigator of the Howard Hughes Medical Institute. Additional support for this work was funded by NIH grant GM-30692 awarded to S.R.L.

Manuscript received June 25, 1997; Accepted for publication October 6, 1997.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AMAN, P., M. MCNEIL, L. FRONZEN, A. G. DARVILL, and P. ALBERSHEIM, 1981  Structural elucidation using HPLC-MS and GLC-MS of the acidic polysaccharide secreted by Rhizobium meliloti strain 1021. Carbohydr. Res. 95:263-282.

ASTETE, S. G. and J. A. LEIGH, 1996  mucS, a gene involved in activation of galactoglucan (EPS II) synthesis gene expression in Rhizobium meliloti. Mol. Plant-Microbe Interact. 9:395-400[Medline].

BARNETT, M. J. 1994 SyrM and regulatory circuits in Rhizobium meliloti Ph.D. Dissertation. Stanford University.

BARNETT, M. J. and S. R. LONG, 1990  DNA sequence and translational product of a new nodulation-regulatory locus: syrM has sequence similarity to NodD proteins. J. Bacteriol. 172:3695-3700[Abstract/Free Full Text].

BARNETT, M. J., B. G. RUSHING, R. F. FISHER, and S. R. LONG, 1996  Transcription start sites for syrM and nodD3 flank an insertion sequence relic in Rhizobium meliloti. J. Bacteriol. 178:1782-1787[Abstract/Free Full Text].

BATUT, J., M.-L. DAVERAN-MINGOT, M. DAVID, J. JACOBS, and A. M. GARNERONE et al., 1989  fixK, a gene homologous with fnr and crp from Escherichia coli, regulates nitrogen fixation genes both positively and negatively in Rhizobium meliloti. EMBO J. 8:1279-1286[Medline].

BECKER, A., S. RÜBERG, H. KÜSTER, A. A. ROXLAU, and M. KELLER et al., 1997  The 32 kilobase exp gene cluster of Rhizobium meliloti directing the biosynthesis of galactoglucan: genetic organization and properties of the encoded gene products. J. Bacteriol. 179:1875-1384.

BERINGER, J. E., 1974  R Factor transfer in Rhizobium leguminosarum. J. Gen. Microbiol. 84:188-198[Medline].

BORTHAKUR, D. and A. W. B. JOHNSTON, 1987  Sequence of psi: a gene on the symbiotic plasmid of Rhizobium phaseoli which inhibits exopolysaccharide synthesis and nodulation, and demonstration that its transcription is inhibited by psr, another gene on the symbiotic plasmid. Mol. Gen. Genet. 207:149-154[Medline].

BORTHAKUR, D., R. F. BARKER, J. W. LATCHFORD, L. ROSSEN, and A. W. B. JOHNSTON, 1988  Analysis of pss genes of Rhizobium leguminosarum required for exopolysaccharide synthesis and nodulation of peas: their primary structure and their interaction with psi and other nodulation genes. Mol. Gen. Genet. 213:155-162[Medline].

BORTHAKUR, D., J. A. DOWNIE, A. W. B. JOHNSTON, and J. W. LAMB, 1985  Psi a plasmid-linked Rhizobium phaseoli gene that inhibits exopolysaccharide production and which is required for symbiotic nitrogen fixation. Mol. Gen. Genet. 200:278-282.

BREWIN, N. J., 1991  Development of the legume root nodule. Annu. Rev. Cell Biol. 7:191-226.

DAVID, M., M.-L. DAVERAN, J. BATUT, A. DEDIEU, and O. DOMERGUE et al., 1988  Cascade regulation of nif gene expression in Rhizobium meliloti. Cell 54:671-683[Medline].

DE BRUIJN, F. J., U. HILGERT, J. STIGTER, M. SCHNEIDER, H. MEYER et al., 1990 Regulation of nitrogen fixation and assimilation genes in the free-living versus symbiotic state, pp. 33–44 in Nitrogen Fixation: Achievements and Objectives, edited by P. GRESSHOFF, G. STACEY and NEWTON. Chapman and Hall, New York.

DE LAJUDIE, P., A. WILLEMS, B. POT, D. DEWETTINCK, and G. MAESTROJUAN et al., 1994  Polyphasic taxonomy of rhizobia: emendation of the genus Sinorhizobium and description of Sinorhizobium comb. nov., Sinorhizobium saheli sp. nov. and Sinorhizobium teranga sp. nov. Int. J. Syst. Bacteriol. 44:715-733.

DÉNARIÉ, J., F. DEBELLÉ, and J. C. PROMÉ, 1996  Rhizobium lipo-chitooligosaccharide nodulation factors: signaling molecules mediating recognition and morphogenesis. Annu. Rev. Biochem. 65:503-535[Medline].

DE VOS, G. F., G. C. WALKER, and E. R. SIGNER, 1986  Genetic manipulations in Rhizobium meliloti utilizing two new transposon Tn5 derivatives. Mol. & Gen. Genet. 204:485-491[Medline].

DEVEREUX, J., P. HAEBERLI, and O. SMITHIES, 1984  A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395[Medline].

DITTA, G., S. STANFIELD, D. CORBIN, and D. R. HELINSKI, 1980  Broad-host-range DNA cloning system for Gram-negative bacteria: construction of a gene bank of Rhizobium meliloti. Proc. Natl. Acad. Sci. USA 77:7347-7351[Abstract/Free Full Text].

DOHERTY, D., J. A. LEIGH, J. GLAZEBROOK, and G. C. WALKER, 1988  Rhizobium meliloti mutants that overproduce the R. meliloti acidic calcofluor-binding exopolysaccharide. J. Bacteriol. 170:4249-4256[Abstract/Free Full Text].

EGELHOFF, T. T. and S. R. LONG, 1985  Rhizobium meliloti nodulation genes: identification of nodDABC gene products, purification of nodA protein, and expression of nodA in Rhizobium meliloti. J. Bacteriol. 164:591-599[Abstract/Free Full Text].

FIGURSKI, D. H. and D. R. HELINSKI, 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/Free Full Text].

FINAN, T. M., B. KUNKEL, G. F. DE VOS, and E. R. SIGNER, 1986  Sec-ond symbiotic megaplasmid in Rhizobium meliloti carrying exopolysaccharide and thiamine synthesis genes. J. Bacteriol. 167:66-72[Abstract/Free Full Text].

FISHER, R. F., T. T. EGELHOFF, J. T. MULLIGAN, and S. R. LONG, 1988  Specific binding of proteins from Rhizobium meliloti cell-free extracts containing NodD to DNA sequences upstream of inducible nodulation genes. Genes Dev. 2:282-293[Abstract].

GLAZEBROOK, J. and G. C. WALKER, 1989  A novel exopolysaccharide can function in place of the calcofluor-binding exopolysaccharide in nodulation of alfalfa by Rhizobium meliloti. Cell 56:661-672[Medline].

GRAY, J. X. and B. G. ROLFE, 1992  Regulation study of the exopolysaccharide synthesis, exoX and exoY in Rhizobium sp. strain NGR234. Arch. Microbiol 157:521-528.

GRAY, J. X., M. A. DJORDJEVIC, and B. G. ROLFE, 1990  Two genes that regulate exopolysaccharide production in Rhizobium sp. strain NGR234: DNA sequences and resultant phenotypes. J. Bacteriol. 172:193-203[Abstract/Free Full Text].

GUALERZI, C. O. and C. L. PON, 1990  Initiation of mRNA translation in prokaryotes. Biochemistry 29:5881-5889[Medline].

GUSSIN, G. N., C. W. RONSON, and F. M. AUSUBEL, 1986  Regulation of nitrogen fixation genes. Annu. Rev. Genetics 20:567-591[Medline].

GYORGYPAL, Z., N. IYER, and A. KONDOROSI, 1988  Three regulatory nodD alleles of diverged flavonoid–specificity are involved in host-dependent nodulation by Rhizobium meliloti. Mol. Gen. Genet. 212:85-92.

HANAHAN, D., 1985 Techniques for transformation of E. coli, pp. 109–114 in DNA Cloning, A Practical Approach, Vol. I, edited by D. M. GLOVER. IRL Press, Oxford.

HIRSCH, A. M., 1992  Developmental biology of legume nodulation. New Phytol. 122:211-237.

HIRSCH, P. R. and J. E. BERINGER, 1984  A physical map of pPH1JI and pJB4JI. Plasmid 12:139-141[Medline].

HIRSCH, A. M., M. BANG, and F. M. AUSUBEL, 1983  Ultrastructural analysis of ineffective alfalfa nodules formed by nif::Tn5 mutants of Rhizobium meliloti. J. Bacteriol. 155:367-380[Abstract/Free Full Text].

HONMA, M. and F. M. AUSUBEL, 1987  Rhizobium meliloti has three functional copies of the nodD symbiotic regulatory gene. Proc. Natl. Acad. Sci. USA 84:8558-8562[Abstract/Free Full Text].

KAHN, D., M. DAVID, O. DOMERGUE, M. L. DAVERAN, and J. GHAI et al., 1989  Rhizobium meliloti fixGHI sequence predicts involvement of a specific cation pump in symbiotic nitrogen fixation. J. Bacteriol. 171:929-939[Abstract/Free Full Text].

KAY, R. and J. MCPHERSON, 1987  Hybrid pUC vectors for addition of new restriction enzyme sites to the ends of DNA fragments. Nucleic Acids Res. 15:2778[Free Full Text].

KELLER, M., A. ROXLAU, W. M. WENG, M. SCHMIDT, and J. QUANDT et al., 1995  Molecular analysis of the Rhizobium meliloti mucR gene regulating the biosynthesis of the exopolysaccharides succinoglycan and galactoglucan. Mol. Plant-Microbe Interactions 8:267-277.

KIJNE, J. W., 1992 The Rhizobium infection process, pp. 349–398 in edited by Biological Nitrogen Fixation. edited by G. STACEY, R. H. BURRIS, and J. J. EVANS. Chapman & Hall, NY

KIM, C. H., D. R. HELINSKI, and G. DITTA, 1986  Overlapping transcription of the nifA regulatory gene in Rhizobium meliloti. Gene 50:141-148[Medline].

LATCHFORD, J. W., D. BORTHAKUR, and A. W. B. JOHNSTON, 1991  The products of Rhizobium genes, psi and pss, which affect exopolysaccharide production, are associated with the bacterial cell surface. Mol. Microbiol. 5:2107-2114[Medline].

LEIGH, J. A. and G. C. WALKER, 1994  Exopolysaccharides of Rhizobium: synthesis, regulation and symbiotic function. Trends Genet. 10:63-67[Medline].

LEIGH, J. A., E. R. SIGNER, and G. C. WALKER, 1985  Exopolysaccha-ride-deficient mutants of Rhizobium meliloti that form ineffective nodules. Proc. Natl. Acad. Sci. USA 82:6231-6235[Abstract/Free Full Text].

LONG, S. R., 1996  Rhizobium symbiosis: nod factors in perspective. Plant Cell 8:1885-1898[Free Full Text].

MARTIN, M. O. and S. R. LONG, 1984  Generalized transduction in Rhizobium meliloti. J. Bacteriol. 159:125-129[Abstract/Free Full Text].

MAXWELL, C. A., U. A. HARTWIG, C. M. JOSEPH, and D. A. PHILLIPS, 1989  A chalcone and two related flavonoids released from alfalfa roots induce nod genes of Rhizobium meliloti. Plant Physiol. 91:842-847[Abstract/Free Full Text].

MEADE, H. M. and E. R. SIGNER, 1977  Genetic mapping of Rhizobium meliloti. Proc. Natl. Acad. Sci. USA 74:2076-2078[Abstract/Free Full Text].

MEADE, H. M., S. R. LONG, G. B. RUVKUN, S. E. BROWN, and F. M. AUSUBEL, 1982  Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J. Bacteriol. 149:114-122[Abstract/Free Full Text].

MERRICK, M. J. 1992 Regulation of nitrogen fixation genes in free-living and symbiotic bacteria, pp. 835–876 in Biological Nitrogen Fixation, edited by G. STACEY, G. H. BURRIS and H. J. EVANS. Chapman and Hall, New York.

MORENO-VIVIAN, C., S. HENNECKE, A. PÜHLER, and W. KLIPP, 1989  Open reading frame 5 (ORF5), encoding a ferredoxinlike protein, and nifQ are cotranscribed with nif E, nif N, nif X, and ORF4 in Rhodobacter capsulatus. J. Bacteriol. 171:2591-2598[Abstract/Free Full Text].

MULLIGAN, J. T. and S. R. LONG, 1985  Induction of Rhizobium meliloti nodC expression by plant exudate requires nodD. Proc. Natl. Acad. Sci. USA 82:6609-6613[Abstract/Free Full Text].

MULLIGAN, J. T. and S. R. LONG, 1989  A family of activator genes regulates expression of Rhizobium meliloti nodulation genes. Genetics 122:7-18[Abstract].

PETERS, N. K., J. W. FROST, and S. R. LONG, 1986  A plant flavone, luteolin, induces expression of Rhizobium meliloti nodulation genes. Science 233:917-1008.

PHILLIPS, D. A., C. M. JOSEPH, and C. A. MAXWELL, 1992  Trigonel-line and stachydrine released from alfalfa seeds activate NodD2 protein in Rhizobium meliloti. Plant Physiol. 99:1526-1531[Abstract/Free Full Text].

REED, J. W. and G. C. WALKER, 1991a  Acidic conditions permit effective nodulation by invasion-deficient Rhizobium meliloti exoD mutants. Genes Dev. 5:2274-2287[Abstract].

REED, J. W. and G. C. WALKER, 1991b  The exoD gene of Rhizobium meliloti encodes a novel function needed for alfalfa nodule invasion. J. Bacteriol. 173:664-677[Abstract/Free Full Text].