Rhizopine (L-3-O-methyl-scyllo-inosamine, 3-O-MSI) is an opine-related compound which is synthesized in alfalfa nodules induced by Rhizobium meliloti L5-30. 3-O-MSI is thought to be a selective growth substrate for R. meliloti L5-30, which harbors genes essential for its synthesis (mos) and catabolism (moc). The moc/mos system has been postulated to provide a competitive advantage in the rhizosphere to bacterial strains harboring it. We are studying the possibility of constructing a novel selectable marker cassette, which is based on the essential moc genes and the utilization of rhizopine as sole nitrogen and carbon source. Moreover, we are exploring the feasibility of the "biased rhizosphere" concept. With this system, we want to study whether transgenic plants, synthesizing and secreting rhizopine into their rhizosphere, are able to selectively enrich for (beneficial) soil bacteria, which are harboring the moc-cassette. In this paper, the characterization of the moc locus and the construction of a minimal moc gene cassette is described. This moc cassette was shown to be able to confer the ability to degrade rhizopine to a Moc^- rhizobial host.

Key Words: rhizopine, L-3-O-methyl-scyllo-inosamine, Rhizobium meliloti L5-30, alternative selectable marker, biased rhizosphere

An opine-like compound, rhizopine (L-3-O-methyl-scyllo-inosamine; L-3-O-MSI) has been identified in nitrogen-fixing nodules induced by Rhizobium meliloti strain L5-30 on its host plant Medicago sativa. Rhizopine has been shown to serve as a unique carbon (C) and nitrogen (N) source specifically for R. meliloti strain L5-30 (Tempe et al., 1982, Murphy et al., 1987). The genes conferring the synthesis (mos) and catabolism (moc) of 3-O-MSI have been found to be of bacterial origin, and have been cloned and characterized (Murphy et al., 1987, 1988).

It has been postulated that the ability to synthesize and degrade rhizopine may confer a competitive advantage upon the bacteria harboring the mos/moc system (Murphy and Saint, 1992). The underlying concept is that the rhizobia carrying the mos genes are capable of directing the synthesis of the unusual nutritional mediator rhizopine from plant derived compounds in the nodule. The rhizopine compound, in turn, can only be utilized as N- and C- source by those bacteria carrying the corresponding catabolism (moc) genes, in the nodule and after plant decomposition in the soil. Based on this concept, we would like to explore the possibility of using the mos/moc system to achieve two interrelated long term goals: i. the construction of a novel selectable marker (moc) cassette for the detection and tracking of microorganisms in the environment, and ii. the utilization of the moc cassette and transgenic plants carrying the mos genes to create biased rhizospheres.

With regard to the rationale behind the first goal, a number of beneficial microorganisms have already been identified, including plant growth promoting rhizobacteria (see Davison, 1988) and microorganisms able to degrade environmental pollutants (Ka et al., 1994; Fan et al., 1993). However, only a limited number of scientific studies have addressed the questions of spread and persistence of (genetically modified) beneficial soil microbes in the environment (see Wilson and Lindow, 1993). This may be due, in part, to the lack of suitable genetic markers for employment in the soil environment. For example, antibiotic resistance marker genes, widely used in microbiological laboratories, are not optimal for environmental studies because of the potential adverse effect of spreading these genes. Other marker genes, such as -galactosidase (lacZ, Drahos et al., 1986) or luciferase (lux, see Meighen, 1989) are also difficult to use in large scale environmental microbial enumeration studies, because of the fact that they cannot be used for microbial selection, may not be selective enough or may be more cumbersome/expensive to detect. Therefore, we reasoned that if the ability to catabolize rhizopine conferred by the moc genes was sufficiently unique and rhizopine could be provided as unique substrate (N- and C-source) for selective plating experiments, then the moc system could be employed as a safe and effective selectable marker system for environmental studies.

With regard to the rationale behind the second goal, field studies with bacterial inoculants have shown that, in most cases, newly introduced (genetically modified) strains are rapidly outcompeted by the indigenous microbial populations. Thus, attempts have failed to inoculate soybean fields with highly efficient nitrogen-fixing strains of the genus Bradyrhizobium japonicum in order to replace an indigenous strain with lower nitrogen-fixing capability (see Triplett, 1990; Triplett and Sadowsky, 1992). We would like to create biased rhizospheres in order to study whether it is possible to enrich for certain beneficial bacteria in the rhizosphere of specific (crop) plants. Transgenic plants will be created expressing the rhizopine synthesis (mos) genes and these plants will be examined for their ability to synthesize and excrete rhizopine into the rhizosphere (Mos phenotype). The rhizopine catabolism (moc) genes will be introduced into various beneficial soil bacteria and the resulting strains examined for their ability to degrade rhizopine as sole N- and C- source (Moc phenotype). Ultimately, we will examine whether these Moc⁺ bacteria can obtain a competitive advantage in the rhizosphere of Mos⁺ plants.

Here, we describe how we used our recently obtained knowledge about the structure and organization of the rhizopine catabolism genes (Rossbach et al., 1994) to construct a rhizopine catabolism (moc) cassette. The constructed moc-cassette contains the minimal number of genes required for rhizopine catabolism and is functional in a Moc^- rhizobial genetic background.

Bacterial strains and plasmids used. The bacterial strains and plasmids used in this study are described in Table 1.

Bacterial growth media. The media and antibiotic concentrations used have been described by de Bruijn et al. (1989). The Minimal Medium used for the rhizopine catabolism assay has been described by Rossbach et al. (1994).

Tn5 mutagenesis, conjugation and gene-replacement protocols. Tn5-mutagenesis, di- and tri-parental matings, and homogenotizations were carried out as described by de Bruijn and Rossbach (1994).

DNA sequencing strategy, determination, and analysis. The DNA sequencing strategy, and analysis used have been described by Rossbach et al. (1994).

Construction of the moc-cassette. Plasmids pSR8500 and pSR8504 were linearized and used as template in PCR. For mocC, pSR8504 was incubated with the oligonucleotides 5'CCGGCGGTACCTACTAGATCTCCCCACTTAACGTTTCC3' and 5'GTCCGCATCTAGAAGGAGGAAGTCGC3' as primers, both synthesized at the Macromolecular Facility at Michigan State University. For mocBA, pSR8500 was incubated with oligonucleotides 5'GTATTCCTGCAGTC

GTTGGCCTCTTC3' and 5'GATGTTCTAGAGCTCCGAAC

GAAGTGG3' as primers. The conditions for PCR were 15 cycles with 1 minute at 94^oC, 1 minute 50^oC, 2 minutes 68^oC, followed by 5 minutes at 68^oC. The obtained PCR products were analyzed and purified via agarose gel electrophoresis, digested with the restriction enzymes PstI and XbaI (mocBA) and KpnI and XbaI (mocC) and inserted together in the intermediate vector pK18. Subsequently, pK18 containing mocBAC on a PstI-BglII fragment was restricted with PstI and BglII and the mocBAC fragment was inserted into pTE3, which had been previously digested with PstI and BamHI.

Preparation of nodule extracts, catabolism studies and high voltage paper electrophoresis. Preparation of nodule extracts, catabolism studies and high voltage paper electrophoresis have been described in Rossbach et al. (1994).

Delineation of rhizopine catabolism (moc) genes. In order to construct a selectable marker cassette which is based on the rhizopine catabolism (moc) genes, the essential moc genes needed to be delineated. Previously, a plasmid carrying a large (15 kb) DNA segment with the moc locus (pPM1031: Murphy et al., 1987) had been shown to confer the ability to catabolize rhizopine to a Moc^- R. meliloti strain (strain 1021). In order to delimit the responsible moc genes, we subjected pPM1031 to transposon Tn5 mutagenesis, and mapped the location of ~50 different Tn5 insertions. The exact location of 24 Tn5 insertions is shown in Figure 1. We examined the phenotype of each plasmid-borne single Tn5 insertion in R. meliloti 1021. In addition, most of the Tn5 insertions were reintroduced into the genome of R. meliloti strain L5-30 via marker exchange ("reverse genetics"), thereby creating proper moc::Tn5 insertion mutants in the original strain from which the moc locus was cloned. The Moc phenotype of each Tn5 insertion mutation was examined using the rhizopine catabolism assay (Rossbach et al., 1994).

The results of the rhizopine catabolism assays are summarized in Figure 1. We identified Moc^- Tn5 insertions in two different regions, region I and II, which are separated by three kilobases of DNA. The construction of several deletion clones and subclones of the moc locus confirmed the results of the Tn5 mutagenesis experiments. Plasmids pSR8510, 8511, and 8533, lacking at least part of the regions responsible for rhizopine degradation, were not able to confer the ability to catabolize rhizopine to a Moc^- R. meliloti strain. However, plasmids pPM1031 and pDK8568, containing the complete regions I and II, were able to confer the ability to catabolize rhizopine (Moc⁺; see Fig. 1).

Determination of the DNA sequence of the moc locus. The determination of the DNA sequence of the moc locus has been described in detail by Rossbach et al. (1994). Here we focus on the relevant features of the deduced moc gene products in respect of their usefulness for the construction of a minimal moc-cassette.

The DNA sequence of 12 kilobases, including region I and region II was determined. A CodonUse program was used to predict open reading frames (ORF's), based on the codon usage in R. meliloti. Four ORF's were predicted in the two regions, in which the Tn5 insertions were responsible for generating the Moc^- phenotype, three in region I and one in region II (Fig. 1). These genes appear to be indispensable for rhizopine catabolism and have been named mocA, B, R, and C. The deduced amino acid sequences of mocA, B, C, and R were compared to protein sequences in GenBank, Release No 78.0. By analyzing the DNA sequence of the different moc genes, and the amino acid sequence of the proteins deduced from these sequences, it is possible to make predictions about the possible functions of three out of four moc gene products. These predictions will be sequentially discussed in the order that the moc gene products are likely to act.

The deduced mocB product exhibits a signal sequence at its N-terminal end which is typical for bacterial proteins transported through or inserted into bacterial membranes (see Pugsley, 1993). Moreover, MocB shows significant homology to periplasmic binding proteins, including MglB and RbsB of Escherichia coli. These periplasmic binding proteins are responsible for sensing and binding compounds like (methyl)galactoside (MglB; Hogg et al., 1991), and ribose (RbsB; Groarke et al., 1983), and interact with membrane components to facilitate the transport of these compounds through the bacterial membrane (see Ames, 1986). It is feasible that the transport of the sugar-alcohol 3-O-MSI is also dependent on a periplasmic binding protein, and we conclude that MocB may be involved in facilitating rhizopine uptake and transport through the bacterial membrane.

The deduced mocR product displays a typical helix-turn-helix motif at its N-terminus, which is known to be involved in DNA-binding (Pabo and Sauer, 1984). In addition, MocR shares significant homology with members of the GntR regulatory class of bacterial proteins (see Haydon and Guest, 1991), which have been shown to regulate the expression of several operons, including fatty acid biosynthesis and degradation or plasmid transfer. We assume that MocR is involved in regulating moc gene expression, possibly as a positive regulator.

The deduced mocA product shows the characteristic amino acids of an NAD-binding motif, which is present in many dehydrogenases (see Thompson and Donkersloot, 1992). Moreover, MocA shares significant homology with the inositol dehydrogenase of Bacillus subtilis. B. subtilis can grow on inositol as sole carbon source, and inositol dehydrogenase is the first enzyme involved in its degradation (Fujita et al., 1991). Because rhizopine is an inositol compound, we assume that MocA is directly involved in degrading L-3-O-methyl-scyllo-inosamine, the rhizopine.

No obvious motifs or significant homologies to any other proteins in the database were detected for the deduced mocC product. The experimental data clearly show that it is essential for rhizopine degradation; but its expected function is unknown at this time.

Designing a minimal moc-cassette. Our results enabled us to design a minimal moc-cassette for use as a selectable marker. We decided to reconstruct the mocBA and mocC genes with the Polymerase Chain Reaction (PCR). Because Tn5 insertions in the mocR gene abolished functional rhizopine catabolism, it is most likely that the mocR product plays the role of a positive regulatory protein. Thus, we postulated that mocBAC should be sufficient for rhizopine catabolism, when expressed from a strong vector promoter.

Figure 2 outlines our strategy for the construction of a minimal moc cassette. We designed oligonucleotides, which were homologous to the 5' and 3' end of mocC, as well as oligonucleotides homologous to the 5' end of mocB and the 3' end of mocA coding sequences. In addition, all oligonucleotides carried at their 5' ends convenient restriction sites for use in successive cloning steps. These oligonucleotides were used in PCR experiments and resulted in amplification of DNA products of the expected size. The PCR products were purified and cloned adjacent to each other (in the order mocBAC) behind the strong trp promoter of the vector pTE3 (Egelhoff and Long, 1985), resulting in the recombinant plasmid pSR8612. The pTE3 vector is a broad-host range plasmid and allows us to test whether pSR8612 is able to confer the ability to catabolize rhizopine to a variety of different bacterial hosts.

The minimal moc cassette confers the ability to catabolize rhizopine to a Moc^- rhizobial strain. Plasmid pSR8612, along with control plasmids, was conjugated into R. meliloti 1021, a strain not able to synthesize or catabolize rhizopine (Mos^-/Moc^-). Figure 3 shows the results of a rhizopine catabolism assay, visualized by a high voltage paper electrophoresis (HVPE). R. meliloti containing plasmids pPM1031 (lane4), pDK8568 (lane 8), and pSR8612 (lane 6) was able to catabolize rhizopine (Moc⁺). In contrast, R. meliloti without plasmid (lane 2) or containing the vector pTE3 (lane 5), was not able to degrade rhizopine (Moc^-). Based on these data, we conclude that the mocBAC genes are sufficient to confer the ability to degrade rhizopine to other rhizobial Moc^- host strains.

Synthesis of L-3-O-methyl-scyllo-inosamine. Since the isolation of rhizopine from alfalfa nodules is quite laborious, we have begun to explore the possibility of chemically synthesizing L-3-O-methyl-scyllo-inosamine in collaboration with Dr. Rawle Hollingsworth (Department of Biochemistry, MSU). Thus far, we have completed the synthesis up to the related compound scyllo-inosamine (SI). Previous results have shown that R. meliloti strain L5-30, which can degrade 3-O-MSI, also is able to degrade scyllo-inosamine (Murphy et al., 1987). The completed synthesis of scyllo-inosamine (which is not available commercially) allowed us to test the different moc::Tn5 mutants for their ability to catabolize SI. Interestingly, all the mutants, which could not degrade 3-O-MSI, also were not able to degrade SI. Thus, the same protein products most likely involved in the transport of rhizopine (MocB), activation of the moc genes (MocR), and degradation of rhizopine (MocA and MocC), also are required for the degradation of SI.

Creation of transgenic Mos⁺ plants. We obtained the plant transformation/expression vector pPCV91 from Drs. C. Koncz and N. Strizhov (Max-Planck-Institute, Cologne, FRG), which allows the concurrent expression of up to three different genes under the control of the strong 35S and the pmas1' and 2' promoters. We have inserted restriction fragments containing mosA, B, and C, which are essential for rhizopine catabolism (Murphy et al., 1993), separately downstream of these three promoters. With the resulting plasmid pMOS7, we have transformed Lotus corniculatus and tobacco plants. We have regenerated the prospective transgenic plants. In a preliminary screening we could not detect rhizopine in the transgenic callus of tobacco or hairy roots of Lotus. This result will remain unclear until sufficient plant material can be obtained to analyze the structure of the newly integrated foreign DNA. In the meantime, we are pursuing alternative transformation strategies in our efforts to produce Mos⁺ plants. We have reconstituted the mosA,B, and C genes via PCR and are cloning them into the same vector (pPCV91). This alternative approach has been initiated because the restriction fragments we had previously used carried additional upstream and downstream sequences, which might interfere with mos gene expression in transgenic plants. We will use the new constructs to transform Lotus and tobacco plants, and test the resulting transgenic plants for rhizopine synthesis and excretion. Prospective Mos⁺ plants will then be used to assemble the biased rhizosphere and to study whether it is possible to selectively enrich for beneficial soil bacteria in the rhizosphere of (crop) plants (see Fig. 4).

We have described how the delineation of the rhizopine catabolism genes via Tn5 mutagenesis and the subsequent determination of the DNA sequence enabled us to predict the genes necessary for the construction of a minimal moc-cassette. The genes mocBAC, possibly encoding proteins involved in transport (MocB) and degradation (MocA and MocC) of rhizopine, have been shown to be sufficient for rhizopine catabolism in a Moc^- rhizobial host strain. However, for the construction of a universal rhizopine catabolism cassette, which maybe useful as a selectable marker, the moc genes need to be expressed in a wide range of different (beneficial) soil bacteria. Experiments to transfer the plasmid pSR8612 into several other soil bacteria are underway. The broad host range origin of pTE3 will enable us to tranfer the minimal moc-cassette into a wide range of different bacterial strains, including plant growth promoting Pseudomonas strains. The resulting exconjugants containing pSR8612 will be tested for their ability to utilize rhizopine as sole nitrogen and carbon source. Subsequently, it may be desirable to integrate the rhizopine catabolism genes into the chromosome of the host strain. This would ensure that the moc genes will not be lost easily during subsequent experiments for their use as selectable marker cassette or in the biased rhizosphere.

At this point, we can only speculate about whether the mocBAC-cassette is indeed sufficient to confer the ability to degrade rhizopine to non-rhizobial strains. In order to gain a better insight into the minimum set of genes necessary to degrade opines or opine-like compounds in general, we compared the structure and organization of several regions encoding opine catabolism and transport genes (see Fig. 5). The DNA sequence of the genes involved in catabolism and transport of the opines octopine and nopaline have been published (Habeeb et al., 1991, Schindler et al., 1989; Schrell and Schroeder, 1993; Valdivia et al., 1991; Zanker et al., 1992). The genetic region encoding the catabolism and transport of mannityl-opines has been analyzed by transposon-induced lacZ fusions and transcriptional units have been defined (Hong et al., 1993).

It is interesting to note that, in general, octopine and nopaline catabolism and transport genes share a high degree of similarity in organization and function (see Zanker et al., 1992). Common to the octopine and nopaline catabolic loci is the presence of a regulatory gene, which is divergently transcribed with respect to the structural genes encoding catabolic and transport functions. Similarly, the mocR gene, the presumptive regulatory gene of the rhizopine locus, is transcribed in the opposite direction to the presumptive structural mocBA genes. All, the mocR, occR, and nocR gene products exhibit a helix-turn-helix motif at their N-termini thought to be involved in DNA-binding. However, MocR shares the greatest homology with the GntR class of bacterial regulatory proteins, whereas OccR and NocR share common features with the LysR-type regulatory family. Some members of the LysR-type regulators are well described (Henikoff et al., 1988). Most of the LysR-type regulators act as activators, only a few act as repressors. They often act on genes which are divergently transcribed from their own direction of transcription. They are able to repress their own synthesis; that is, they are negatively autoregulated. The LysR-type regulators are activated by specific ligands. However, in contrast to many other regulatory proteins, they bind to their operators in both the presence and absence of the ligand. Because the ligands do not greatly affect the binding affinity of the regulator to their target promoter, they seem to alter other properties. Recently Wang et al. (1992) reported that in the case of OccR, octopine might cause relaxation of a bend in the DNA of the promoter region and thus permits transcription of the structural genes involved in octopine transport and catabolism. Thus far, the GntR-type of regulatory proteins are not studied as well as members of the LysR-regulators. The function of only a few of them are known, including HutC of Klebsiella aerogenes, a repressor of the histidine utilization operons (Schwacha and Bender, 1990), and FadR of E. coli, which has multiple roles in fatty acid biosynthesis and degradation (DiRusso et al., 1988). At this time we can only speculate about the specific role and mechanism of action of MocR. Interestingly, its central and C-terminal domains share some conserved amino acids with amino aspartate transferases (see Rossbach et al. 1994). It could be that these portions of MocR are involved in ligand (rhizopine?) binding.

It is also interesting to compare the gene product most likely involved in transport of rhizopine (MocB) with those involved in the transport of the opines octopine and nopaline. The gene products involved in octopine transport (OccQ, M, P, T), as well as the gene products involved in nopaline transport (NocP, T, Q, M; Zanker et al., 1992) belong to a family of transport proteins called "traffic ATPases" or "shock-sensitive permeases" (see Ames et al., 1990). These transport systems are responsible for transporting sugar compounds, and basic or branched-chain amino acids. They usually consist of two to four integral membrane units and one periplasmic binding protein (Ames, 1986). One subunit displays an ATP-binding site, and it is assumed that the energy required for transport derives from ATP hydrolysis (see Ames et al., 1990). Normally, all subunits of these traffic ATPases are encoded by genes located next to each other in one operon, as is the case in the octopine and nopaline catabolic operons. MocB shows significant homology to the periplasmic binding protein MglB, responsible for sensing and binding (methyl-) galactoside (see Rossbach et al., 1994). Surprisingly, we could not identify any other open reading frame in the rhizopine catabolic region, which would resemble the integral membrane units of this type of permeases. However, some periplasmic transport systems have been described to be multifunctional; they are able to transport more than one substrate and can interact with more than one type of periplasmic binding protein (Ames et al., 1990). Thus, we speculate that MocB, after having bound rhizopine, is able to interact with a multifunctional shock-sensitive permease, possibly a sugar or sugar-alcohol transport system.

Directly adjacent to the genetic region responsible for octopine transport, genes are located encoding proteins involved in degradation of octopine. The two subunits of the octopine oxidase encoded by ooxBA are adjacent to the ocd gene encoding ornithine cyclodeaminase. These enzymes, together with a chromosomally encoded arginase, are necessary for octopine degradation via arginine, ornithine and proline. Homologous genes can be detected in the locus responsible for nopaline catabolism. However, in the nopaline catabolic region there are additional genes, encoding an arginase (arc) and a 40 kD product which shares similarity with dehydrogenases (Schrell and Schroeder, 1993). Because rhizopine is an inositol compound, but octopine and nopaline are basic amino acid derivatives, we do not expect, and did not find, any further homology between rhizopine and opine degradation gene products.

Our results enabled us to construct a minimal cassette consisting of the rhizopine catabolism genes mocBAC. We are eager to explore its use as a marker cassette and as part of a biased rhizosphere system. In the biased rhizosphere we can test whether a transgenic plant expressing the rhizopine synthesis genes and secreting the rhizopine, can enrich (beneficial) soil microorganisms harboring the moc-cassette in its rhizosphere (see Fig 4). Preliminary data obtained by other investigators, support the feasibility of the biased rhizosphere. It has been demonstrated, that mannityl-opines are accumulated and excreted by transgenic tobacco plants harboring the corresponding synthesis genes (Savka and Farrand, 1992). Also, transgenic Lotus plants producing mannityl opines have recently been reported to promote growth of opine degrading agrobacteria (Guyon et al., 1993). In addition, Hwang and Farrand (1994) have described a plant-growth promoting Pseudomonas strain, which carries an agrobacterial mannityl-opine catabolic region, and was subsequently able to use mannopine and agropine as sole carbon source. These results and our own data let us suggest that the biased rhizosphere is a realistic concept worth exploring for the purpose of modifying microbial populations in the rhizosphere of (crop) plants.

We thank Uwe Rossbach, Marlene Cameron, and Kurt Stepnitz for help in preparing the figures. This work was supported by the US Department of Agriculture (92-39210-8224), the Rackham Foundation, the NSF Center for Microbial Ecology (BIR 9120006), and the US Department of Energy (DE-FG1290ER20021). SR also acknowledges the OECD Project on Biological Resource Management for a short-term fellowship.

Ames GF-L (1986) Bacterial periplasmic transport systems: structure, mechanism, and evolution. Ann Rev Biochem 55:397-425

Ames GF-L, Mimura CS, Shyamala V (1990) Bacterial periplasmic permeases belong to a family of transport proteins operating from Escherichia coli to human: Traffic ATPases. FEMS Microbiol Lett 75:429-446

de Bruijn FJ, Rossbach S (1994) Transposon Mutagenesis. In: Gerhardt P, Murray RGE, Wood WA, Krieg NR (eds) Methods for General and Molecular Bacteriology. American Society for Microbiology, Washington, DC, pp 387-405

de Bruijn FJ, Rossbach S, Schneider M, Ratet P, Messmer S, Szeto WW, Ausubel FM, Schell J (1989) Rhizobium meliloti 1021 has three differentially regulated loci involved in glutamine biosynthesis, none of which is essential for symbiotic nitrogen fixation. J Bacteriol 171: 1673-1682

Davison J (1986) Plant beneficial bacteria. Bio/Technology 6:282-286

DiRusso CC (1988) Nucleotide sequence of the fadR gene, a multifunctional regulator of fatty acid metabolism in Escherichia coli. Nucl Acids Res 16: 7995-8009

Drahos D J, Hemming BC, and McPherson S (1986) Tracking recombinant organisms in the environment: B-galactosidase as a selectable non-antibiotic marker for fluorescent Pseudomonas. Bio/Technology 4:439-444

Egelhoff TT, Long SR (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

Eardly BD, Materon LA, Smith NH, Johnson DA, Rumbaugh MD, Selander RK (1990) Genetic structure of natural populations of the nitrogen-fixing bacterium Rhizobium meliloti. Appl Environ Microbiol 56:187-194

Fan S, Scow KM (1993) Biodegradation of trichlorethylene and toluene by indigenous microbial populations in soil. Appl Environm Microbiol 59:1911-1918

Fujita Y, Shindo K, Miwa Y, Yoshida K (1991) Bacillus subtilis inositol dehydrogenase-encoding gene (idh): sequence and expression in Escherichia coli. Gene 108:121-125

Groarke JM, Mahoney WC, Hope JN, Furlon CE, Robb FT, Zalkin H, Hermodson MA (1983) The amino acid sequnece of D-ribose binding protein from Escherichia coli K12. J Biol Chem: 12952-12956

Guyon P, Petit A, Tempe J, Dessaux Y (1993) Transformed plants producing opines specifically promote growth of opine-degrading agrobacteria. Mol Plant-Mirobe Interact 6: 92-98

Habeeb LF, Wang L, Winans SC (1991) Transcription of the octopine catabolsim operon of the Agrobacterium tumor-inducing plasmid pTiA6 is activated by a LysR-type regulatory protein. Mol Plant-Microb Interact 4:379-385

Haydon DJ, Guest JR (1991) A new family of bacterial regulatory proteins. FEMS Microbiol Lett 79:291-296

Henikoff S, Haughn GW, Calvo JM, Wallace JC (1988) A large family of bacterial activator proteins. Proc Natl Acad Sci USA 85:6602-6606

Hogg RW, Voelker C, Von Carlowitz I (1991) Nucleotide sequence and analysis of the mgl operon of Escherichia coli K12. Mol Gen Genet 229:453-459

Hong SB, Dessaux Y, Chilton WS, Farrand SK (1993) Organization and regulation of the mannopine cyclase-associated opine catabolism genes in Agrobacterium tumefaciens 15955. J Bacteriol 175:401-410

Hwang I, Farrand SK (1994) A novel gene tag for identifiying microorganism released into the environment. Appl Environ Microbiol 60:913-920

Meade HM, Long SR, Ruvkun GB, Brown SE, Ausubel FM (1982) Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J Bacteriol 149: 114-122

Meighen EA (1991) Molecular biology of bacterial bioluminescence. Microbiol Rev 55:123-142

Ka JO, Holben WE, Tiedje JM (1994) Genetic and phenotypic diversity of 2,4-dichlorophenoxyacetic acid (2,4-D)-degrading bacteria isolated from 2,4-D-treated field soils. Appl Environ Microbiol 60:1106-115

Kowalski M (1970) Transducing phages of Rhizobium meliloti. Acta Microbiol Polon Ser A 19: 109-114

Murphy PJ, Heycke N, Banfalvi Z, Tate ME, de Bruijn FJ, Kondorosi A, Tempe J, Schell J (1987) Genes for the catabolism and synthesis of an opine-like compound in Rhizobium meliloti are closely linked and on the Sym plasmid. Proc Natl Acad Sci USA 84: 493-497

Murphy PJ, Heycke N, Trenz SP, Ratet P, de Bruijn FJ, Schell J (1988) Synthesis of an opine-like compound, a rhizopine, in alfalfa nodules is symbiotically regulated. Proc Natl Acad Sci USA 85: 9133-9137

Murphy PJ, Saint CP (1992) Rhizopines in the legume-Rhizobium symbiosis. In: Verma DPS (ed) Molecular Signals in Plant-Microbe Communications. CRC Press Boca Raton, FL, pp 377-390

Murphy PJ, Trenz SP, Grzemski W, de Bruijn FJ, Schell J (1993) The Rhizobium meliloti rhizopine mos locus is a mosaic structure facilitating its symbiotic regulation. J Bacteriol 175: 5193-5204

Pabo CO, Sauer RT (1984) Protein-DNA recognition. Ann Rev Biochem 53: 293-321

Pridmore RD (1987) New and versatile cloning vectors with kanamycin resistance marker. Gene 56:309-312

Pugsley AP (1993) The complete general secretory pathway in Gram-negative bacteria. Microbiol Rev 57: 50-108

Rossbach S, Kulpa DA, Rossbach U, de Bruijn, FJ (1994) Molecular and genetic characterization of the rhizopine catabolism (mocABRC) genes of Rhizobium meliloti L5-30. Mol Gen Genet, in press

Savka MA, Farrand S (1992) Mannityl opine accumulation and exudation by transgenic tobacco. Plant Physiol 98: 784-789

Schindler U, Sans N, Schroeder J (1989) Ornithine cyclodeaminase from octopine Ti plasmid Ach5: Identification, DNA sequence, enzyme properties, and comparison with gene and enzyme from nopaline Ti plasmid C58. J Bacteriol 171:847-854

Schrell A, Schroeder J (1993) Characterization of a nopaline-induced gene for 40 kDa protein in the nopaline catabolic gegion of Ti plasmid pTiC58. Biochim Biophys Acta 1174:303-304

Schwacha A, Bender RA (1990) Nucleotide sequence of the gene encoding the repressor for the histidine utilization genes of Klebsiella aerogenes. J Bacteriol 172: 5477-5481

Tempe J, Petit A, Bannerot H (1982) Presence de substances semblables a des opines dans des nodosites de Luzerne. C R Acad Sc Paris, Serie III, 295: 413-416

Thompson J, Donkersloot JA (1992) N-(carboxyalkyl)amino acids: occurrence, synthesis, and functions. Ann Rev Biochem 61: 517-557

Triplett EW (1990) The molecular genetics of nodulation competitiveness in Rhizobium and Bradyrhizobium. Mol Plant-Microbe Interact 3:199-206

Triplett EW, Sadowsky MJ (1992) Genetics of competition for nodulation of legumes. Annu Rev Microbiol 46:399-428

Valdivia RH, Wang L, Winans SC (1991) Characterization of a putative periplasmic transport sytem for octopine accumulation encoded by Agrobacterium tumefaciens Ti plasmid pTiA6. J Bacteriol 173:6398-6405

Wang L, Helman JD, Winans SC (1992) The A. tumefaciens transcriptional activator OccR causes a bend at a target promoter, which is partially relaxed by a plant tumor metabolite. Cell 69:659-667

Wilson M, Lindow SE (1993) Release of recombinant microorganisms. Annu Rev Microbiol 47:913-944

Zanker H, von Lintig J, Schroeder J (1992) Opine transport genes in the octopine (occ) and nopaline (noc) catabolic regions in Ti plasmids of Agrobacterium tumefaciens. J Bacteriol 174:841-849

Rossbach Table 1

Rossbach fig 1

Rossbach fig 2

Rossbach fig 3

Rossbach fig 4

Rossbach fig 5

1. Corresponding author: Dr. Frans J. de Bruijn, DOE Plant Research Laboratory, Michigan State University,East Lansing, MI 48824, Phone: (517) 353 2229, FAX: (517) 353 9168, E-mail: debruijn@msu.edu