J3 3





Figure 1. Genome structure downstream of the phe-tRNA gene or symbiosis island in strains CJ3, CJ4, R88B, MAFF303099 and R7A. The ;MAFF303099 sequence is from Kaneko et al. (2000). Genes designations are as for MAFF303099 where homologs are present. The R88B gene labeled mus-20 is similar to mus-20 of P. putida (Espinosa-Urgel et al. 2000).

Figure 1. Genome structure downstream of the phe-tRNA gene or symbiosis island in strains CJ3, CJ4, R88B, MAFF303099 and R7A. The ;MAFF303099 sequence is from Kaneko et al. (2000). Genes designations are as for MAFF303099 where homologs are present. The R88B gene labeled mus-20 is similar to mus-20 of P. putida (Espinosa-Urgel et al. 2000).

Sequence downstream of the int gene in strain CJ4 (Figure 1) contained a cluster of genes found elsewhere in the genome of MAFF303099, including a gene mll2848 whose product showed similarity to a family of outer membrane adhesins. This family includes yadA of Yersinia and uspA of Moraxella (Hoiczyk et al. 2000). Hence the int genes in R88B and CJ4 are both associated with genes encoding outer membrane adhesins that have been identified as essential colonization factors, strongly suggesting that the acquired DNA is of adaptive value.

The sequence immediately downstream of the symbiosis island in strain R7A showed identity to a region 35.4 kb downstream of the symbiosis island in MAFF303099 (Figure 1). The identity included the 3' 17-bp of the tRNA gene except that MAFF303099 showed a 1-bp deletion. These results clearly demarcate the island associated with the int gene in MAFF303099. The R7A DNA is not found in the other mesorhizobia analyzed, indicating that it represents another acquired region that R7A has in common with MAFF303099.

5. Concluding Remarks

In summary, the symbiosis island of R7A is a dynamically evolving element shaped by multiple recombination events that adapts a soil saprophyte to a symbiotic interaction with a plant host. In addition, we have found evidence for fifteen different acquired DNA regions adjacent to a phenylalanine tRNA gene in members of a population of soil mesorhizobia. Some of these acquired regions encode traits such as iron acquisition and seed adhesion that are likely to endow the strains carrying them with a competitive advantage. Overall, this work has shown remarkable genetic diversity in the soil mesorhizobial population that is due to the acquisition of chromosomal DNA and has provided insight into the extensive role of horizontal gene transfer in microbial evolution.

6. References

Espinosa-Urgel M et al. (2000) J. Bacteriol. 182, 2363-2369

Guerinot ML (1994) Annu. Rev. Microbiol. 48, 743-772

Hacker J, Kaper JB (2000) Annu. Rev. Microbiol. 54, 641-679

Hochhut B et al. (1997) J. Bacteriol. 179, 2097-2102

Ochman H et al. (2000) Nature 405, 299-304

Perna NT et al. (2001) Nature 409, 529-533

Ravatn R et al. (1998) J. Bacteriol. 180, 5505-5514

Sullivan JT et al. (2000) In Triplett E (ed), Prokaryotic Nitrogen Fixation: A Model System for

Analysis of a Biological Process, pp 693-704, Horizon Scientific Press, Wymondham, UK Sullivan JT, Ronson CW (1998) Proc. Natl. Acad. Sei. USA 95, 5145-5149 Sullivan JT et al. (1996) Appl. Environ. Microbiol. 62, 2818-2825 Sullivan JT et al. (1995) Proc. Natl. Acad. Sei. USA 92, 8985-8989 VipreyVA etal. (1998) Mol. Microbiol. 28, 1381-1389

7. Acknowledgements

This work was supported by grants from the Marsden Fund administered by the Royal Society of New Zealand, the US DOE (DE FG02-91ER200021) and by Otago and Michigan State Universities.


P. van Berkum1, S. Reiner2, B.D. Eardly3

'Soybean Genomics and Improvement Laboratory, USD A, ARS, Beltsville, MD 20705, USA

2Insect Biocontrol Laboratory, USDA, ARS, Beltsville, MD 20705, USA

3Penn State Berks-Lehigh Valley College, Berks Campus, P.O. Box 7009 Reading, PA 19610, USA

1. Introduction

Phylogeny is the analysis of character sets to reconstruct evolutionary paths of extant species. Often the character sets are based on morphological variation among extant and extinct organisms. However, rhizobia and other bacteria have few such informative characters and no fossil record is available. In such cases gene sequences are compared in an evolutionary context. This approach relies on the assumption that evolution throughout the genome progresses at an approximately constant rate in a hierarchical manner largely via the mechanism of mutation and Darwinian selection. Although many genes are potentially useful for the purpose of reconstructing evolutionary history, in bacteria it has become common practice to use the 16S rRNA genes. Use of the 16S rRNA gene further assumes that each genome contains a single copy or multiple identical copies and that evolution of the 16S rRNA gene approximates that of the entire genome. However, there are increasing concerns that these assumptions may not strictly apply. For instance, individual bacterial cells may harbor multiple divergent copies of the 16S rRNA gene (Amann et al. 2000; Carbon et al. 1979; Dreyden, Kaplan 1990; Rainy et al. 1996; van Berkum et al. 1999; Wang et al. 1997). Also, evolution of the 16S rRNA gene may be reticulate (Eardly et al. 1996; Sneath 1993; Wang, Zhang 2000; Yap et al. 1999). Because of these concerns we decided to examine the 16S rRNA gene to provide perspective on these concerns as they relate to rhizobia and closely related taxa. Our approach was to first evaluate how the addition of various taxa would influence the consistency of 16S rRNA gene tree topologies in a phylogenetic analysis of rhizobia and closely related non-rhizobial taxa; then to compare tree topologies reconstructed from the 16S rRNA gene, the 23S rRNA gene and the Internally Transcribed Space (ITS) region between them; and finally to investigate the possibility of gene conversion in the 16S rRNA gene of rhizobia and selected related bacteria.

2. Procedure (Material and Methods)

The sequences examined in this study (16S rRNA, 23S rRNA and ITS region) were generated either using an automated sequencing protocol described previously (van Berkum, Fuhrmann 2000) or in the case of the 16S rRNA genes were in part obtained from public databases. Sequences were aligned using PILEUP of the Wisconsin GCG package and alignments were checked manually by using Genedoc (Nicholas, Nicholas 1997). Neighbor-joining trees were constructed from Jukes-Cantor distances using MEGA (Kumar et al. 1993) or trees were assembled in a stepwise manner with Parsimony analysis using PAUP (Swofford 2001). Parsimony and distance trees also were generated with aligned sequences of the 16S rRNA genes constraining the topologies to resemble the 23S rRNA tree and were drawn with MacClade (Maddison, Maddison 1999). This was done to investigate differences in the number of steps required to construct trees and to compare tree topologies using the Shimodaira-Hasegawa test (Shimodaira, Hasegawa 1999). This was done by generating likelihood scores of tree files and then subtracting the score closest to zero from those of the other trees. The significance in differences among the likelihood scores was determined with a one-tailed bootstrap test using 1000 permutations of the data. Finally, the Geneconv program (Sawyer 1989) was used to test the possibility of a history of recombination among the 16S rRNA genes. With this method the distribution of polymorphic nucleotide positions along a gene is examined to estimate the likelihood that distinct segments have differing phylogenies.

3. Results and Discussion

We investigated consistency in 16S rRNA tree topology because topologies of published trees may vary, depending on differences in alignments, software used, or taxa sampled. To examine the effects of the latter, the 16S rRNA gene sequences for the three non-rhizobial taxa Blastobacter aggregatus, Bl. capsulatus, and Ensifer adhaerans were included in the analysis and effects on tree topology were noted. The resulting tree placed the two Blastobacter species within the genus Agrobacterium, and E. adhaerens was placed in close proximity to the genus Sinorhizobium (data not shown). The species R, galegae and R. huautlense also were moved from a position adjacent to Ag. vitis to one adjoining R. gallicum.

In the next phase of our analysis, we compared the tree topologies for three different loci within the rRNA operon. In order to determine whether the topologies of each were equally parsimonious, the 16S rRNA tree was constrained to the topology of the 23S rRNA tree (Figure 1). However, the constrained tree required more steps for construction than the unconstrained tree, indicating that the constrained tree was less parsimonious than the unconstrained tree. Subsequently we compared topologies of the constrained and the unconstrained 16S rRNA trees to determine whether the same or different phylogenetic information was obtained from analysis of the 16S rRNA

24 trees 909

24 trees 909

Unconstrained 16S rRNA tree

Ag. tumefaciens Ag. rubi Ag. vitis R. galegae R. huautlense Ag. rhizogenes R. tropici R. leguminosarum R. etli R. gallicum S. meliloti S. arboris S.fredii S. saheli S. kostiense S. terangae My. dimorpha Och. anthropi M. ciceri M. loti M. amorphae M. huakuii Phy. myrsinacearum B. japonicum Af. felis Rho. palustris Bl. denitrificans B. elkanii Az. caulinodans Rhod. sphaeroides

15 trees 1016

My. dimorpha Ag. vitis Ag. tumefaciens Ag. rubi Ag. rhizogenes R. tropici R. etli S. meliloti S. arboris S. fredii S. saheli S. kostiense S. terangae R. leguminosarum R. gallicum R. galegae R. huautlense M. amorphae M. loti M. ciceri M. huakuii Phy. myrsinacearum Och. anthropi Bl. denitrificans B. elkanii B. japonicum Af. felis Rho. palustris Az. caulinodans Rhod. sphaeroides

Unconstrained 16S rRNA tree

Constrained 16S rRNA tree

Figure 1. Comparison of the 16S rRNA tree to a 16S rRNA tree constrained to the topology of the 23S rRNA tree generated by Parsimony analysis and drawn in MacClade.

and 23 S rRNA genes. Topologies of the 24 unconstrained trees (Figure 1) were similar as tested by the Shimodaira-Hasegawa test, while topologies of the unconstrained reference tree and the 15 constrained trees (Figure 1) were significantly different (P<0.001). A similar test was done comparing distance trees generated from the ITS region and the 16S rRNA gene (Figure 2). These two trees also had significantly different topologies as determined by the Shimodaira-Hasegawa test (PO.OOl).

USDA 122 USDA 62 USDA 110 USDA 129 —USDA 126

_Bl. denitrificans

I 1_2281 B. liaoningense

USDA 6 B.japonicum USDA 38 Rho. palustris Af. felis USDA 130 USDA 94 USDA31 USDA 46

USDA 76 B. elkanii USDA 3177 (A. americand) Meth. extorquens Meth. rhodinum Meth. organophylum Rho. sphaeroides


USDA 122 USDA 126 USDA 62

_USDA 129

-USDA 110



I_2281 B. liaoningense

-USDA 124

USDA 130 USDA 76 B. elkanii


_USDA 3177 (A. americana)

—Af.felis I—Bl. denitrificans 1—Rho. palustris

Meth. extorquens

Meth. rhodinum Meth. organophylum Rho. sphaeroides

Figure 2. Comparison of trees reconstructed from the 16S rRNA gene and the ITS region.

We concluded from comparing tree topologies that the phylogenetic information obtained by analysis of thel6S rRNA gene is incongruent with that of the 23 S rRNA gene or with that of the ITS region. A comparison of the 23 S rRNA gene and ITS region was not made.

Using tests for recombination we predicted that several different regions within the 16S rRNA gene potentially resulted from gene conversion (data not shown). We focused on a 229 bp segment identified from comparing the 16S rRNA gene sequences of Mesorhizobium huakuii and S. fredii. Topologies of trees reconstructed either from this small region or from the alignment after removing this short segment were significantly different from trees reconstructed from the entire length of the aligned 16S rRNA gene sequences. Therefore, it is possible that 16S rRNA gene sequences of rhizobia and related taxa contain short segments resulting from lateral gene transfer and genetic recombination with divergent alleles.

From our data we concluded that our concerns over the assumptions made when reconstructing bacterial evolutionary history from 16S rRNA gene sequence divergence are relevant to rhizobia and related taxa. Therefore, it may be inappropriate to use the 16S rRNA gene alone for reconstructing bacterial phylogenies and subsequently to use these results as primary evidence for deciding rhizobial nomenclature. Some examples where rhizobial classification decisions were made largely on the basis of analysis of the 16S rRNA gene include separation of Sinorhizobium from Rhizobium (DeLajudie et al. 1994), the proposal of the new genus Allorhizobium (DeLajudie et al. 1998), the suggestion for combining Agrobacterium and Allorhizobium into the genus Rhizobium (Young et al. 2001), and proposing separate genera for Bradyrhizobium japonicum and B. elkanii (Willems et al. 2001). Such changes in nomenclature may not be warranted since the evidence for phylogenetic placement of these genera is at best inconclusive. As an alternative we suggest that a more conservative approach be taken where taxonomic decisions are based on the analysis of a variety of loci, and that comparative analytical methods be used to estimate phylogenetic relationships among the species being considered.

4. References

Amann G el al. (2000) Extremophiles 4, 373-376 Carbon C et al. (1979) EMBO J. 11, 4175-4185 DeLajudie P et al. (1994) Int. J. Syst. Bacteriol. 44, 715-733 DeLajudie P et al. (1998) Int. J. Syst. Bacteriol. 48, 1277-1290 Dreyden SC, Kaplan S (1990) Nucleic Acids Res. 18, 7267-7277 Eardly BD et al. (1996) Plant Soil 186, 69-74

Kumar S et al. (1993) MEGA: Molecular Evolutionary Genetics Analysis, Version 1.01,

The Pennsylvania State University, University Park, PA Maddison WP, Maddison DR (1999) MacClade, Analysis of Phylogeny and Character Evolution,

Version 3.08, Sinaur Associates, Sunderland, MA Nicholas KB, Nicholas HB (1997) Alignment Editor and Shading Utility, Version 2.6.001 Rainey FA et al. (1996) Microbiol. 142,2087-2095 Sawyer SA (1989) Mol. Biol. Evol. 6, 526-538 Shimodaira H, Hasegawa M (1999) Mol. Biol. Evol. 16, 1114-1116 Sneath PHA (1993) Int. J. Syst. Bacteriol. 43, 626-629

Swofford DL (2001) PAUP, Phylogenetic Analysis Using Parsimony, Version 4.0b8,

Sinaur Associates, Sunderland, MA van Berkum P, Fuhrmann JJ (2000) Int. J. Syst. Evol. Microbiol. 50, 2165-2172 van Berkum P et al. (1999) In Martinez E, Hernandez G (ed) Highlights of

Nitrogen Fixation Research, pp. 267-269, Kluwer Academic/Plenum Publisher, New York Wang Y, Zhang Z (2000) Microbiol. 146, 2845-2854 Wang E et al. (1997) J. Bacteriol. 179, 3270-3276 Willems A et al. (2001) Int. J. Syst. Evol. Microbiol. 51,111-117 Yap WH etal. (1999) J. Bacteriol. 181, 5201-5209 Young JM et al. (2001) Int. J. Syst. Evol. Microbiol. 51, 89-103


F.J. de Bruijn ''2'3, F. Ampe3, J. Batut3, H. Berges3, M. Davalos3, M.E. Davey1'2, J. Gouzy3,

D. Kahn3, E. Kiss3, E. Lauber3, C. Liebe3, A. Milcamps1, C. Ronson4, P. Struffi1, J. Sullivan4'

J. Trzebiatowski1, C. Vriezen1

'MSU-DOE Plant Research Laboratory zDept Microbiol. Michigan State U., E. Lansing, MI 48824, USA

3INRA/CNRS Laboratoire de Biologie Moleculaire des Relations Plantes-Microorganismes, BP 27, 31326 Castanet-Tolosan Cedex, France

4Dept Microbiol., U. of Otago, Dunedin, New Zealand

1. Introduction

In nature, bacterial growth is restricted by a wide variety of environmental factors, including the lack of essential nutrients, osmotic stress, oxygen limitation and oxidative stress. In most natural settings at least some essential nutrients are limiting, and stress conditions are prevalent, therefore periods of negligible growth or dormancy are the more typical physiological state of bacteria. Understanding how bacteria are able to monitor, sense, and respond to their environment in nutrient-deprived, stressed physiological states or in eukaryotic host tissues is fundamental to our understanding of microbial biology and ecology, as well as symbiotic plant-microbe interactions.

While studies on the E. coli model system, and selected marine systems, have revealed a considerable number of basic genetic components of microbial stress responses, comparatively little is understood about how soil bacteria respond to nutrient limitation conditions. Soil is generally a harsh, oligotrophic environment. Nutrient limitation and oxygen limitation may represent the most prevailing stress conditions for soil bacteria. A paucity of organic matter is present in most soils, which is often insoluble or in a form inaccessible to bacteria (e.g. humus and lignin). In fact, non-growth or extremely limited growth, may be the rule, rather than the exception (Matin et al. 1989). Rhizosphere soil has been reported to be a somewhat more supportive environment for bacteria than bulk soil, due to the presence of plant root exudates as readily accessible nutrients. However, even in the rhizosphere, bacterial growth and activity are generally limited to short periods during which these exudates are available (Lynch, Whipps 1990).

We have focused our studies on stress mediated gene expression during nutrient deprivation and under other stress conditions in the indigenous soil bacterium, Sinorhizobium meliloti. This bacterium is a capable of establishing a symbiosis with the legume alfalfa {Medicago sativa), during which a new specialized organ is formed, the nitrogen-fixing root nodule. These nodules provide the proper physiological conditions for the bacteria to survive in the absence of competing microflora, and to reduce atmospheric dinitrogen to ammonia, which is then assimilated by the plant. Little is known about the manner in which these bacteria are able to persist in their free-living state, in the soil and rhizosphere or how they respond to specific physiological stress conditions in planta. S. meliloti has become one of the leading model organisms to study microbial persistence, competition and stress responses in soil, as well as plant-microbe interactions leading to indeterminate nodule formation, since its genetic analysis is highly developed, and its entire genomic DNA sequence has been determined and annotated (Galibert et al. 2001).

Moreover, one of its hosts, Medicago truncatula, has become a model organism to study the plant components of symbiotic nitrogen fixation, the formation of indeterminate nodules, as well as comparative legume genetics (Cook et al. 1997). Similar statements can be made for Mesorhizobium loti and its host Lotus japonicus, which have been developed as a model system for determinate nodule formation (Cook et al. 1997). The M. loti system is of particular interest, since the genomic sequence of an M. loti strain MAFF303099 (Kaneko et al. 2000) has been determined and most, if not all of the symbiotic genes of another strain, R7A, as well as several putative persistance- and competition-related loci have been found to be located on a transferable, well defined Symbiosis Island (Sullivan et al. 1995; Ronson et al. this volume).

Therefore, we have been interested in using molecular genetic, genomic and post-genomic approaches to study environmental control of gene expression in S. meliloti, and to employ comparative structural and functional genomics with other rhizobia, such as M. loti (and especially its transferable Symbiosis Island), in order to elucidate pathways involved in rhizosphere microbial persistence and competition, as well as other aspects of symbiotic plant-microbe interactions

We initially focused on a limited number of key issues in the area of S. meliloti gene regulation in response to environmental stress conditions that are relevant to molecular microbial ecology in general, and symbiotic plant-microbe interactions in particular, namely:

1. What is the nature of bacterial genes responding to environmental stress?

2. What is the role of stress-responsive bacterial genes in persistence or competition in bulk soil or the (endo)rhizosphere of plants?

3. Do novel, common global regulatory loci exist in (soil) bacteria that are responsible for persistence and competition in the bulk soil, in the rhizosphere of plants and/or bacterial differentiation (bacteroid formation)?

2. Isolation of TnSluxAB Tagged S. Meliloti Loci

Using a promoterless Tn5luxAB (Wolk et al. 1991), transposon mutagenesis of S. meliloti strain 1021 was carried out and 5000 transcriptional fusions were analyzed for luminescence under N- or C- deprivation or O2 limitation. Using this protocol, 22 S. meliloti strains carrying Tn5luxAB gene fusions induced by N-deprivation, 12 by C-deprivation and 24 by 02-limitation were isolated. Cloning and DNA sequence analysis of the tagged loci revealed genes sharing similarities with previously identified bacterial loci involved in N- or C-metabolism and 02-responses, as well as novel genes. The N- and/or C-deprivation induced loci include exopolysaccharide biosynthesis (1exoF and exoY), nitrite and nitrate assimilation (nasD,E, A), amino acid transport (braF and ilvH), amino acid synthesis (dapA) or degradation (speB, arcC), as well as ribose transport (rbsA) genes. Among the loci carrying 02-limitation induced gene fusions, similarities were discovered with exopolysaccharide (exoO), nitrogen fixation (fixN), cytochrome oxidase (•cyoC), heat shock (htpG), and novel genes.

Our genetic screen for N-, C-deprivation and 02-limitation induced loci identified genes known to be induced under these conditions (e.g. nas D, E, A under N-deprivation; rbsA under C-deprivation; fixN under 02-limitation), suggesting that our approach to isolate multiple environmentally controlled S. meliloti genes had been successful (Lim et al. 1993; Milcamps et al. 1998; Trzebiatowki et al. 2001). We do realize, however, that not all N- or C-deprivation or 02-limitation induced genes would be picked up in a screen of 5000 Tn5luxAB insertions (Milcamps et al. 1998). Therefore our future studies will be focused on the use of micro- and macro-arrays to analyze gene expression patterns in a more global fashion (see below)

3. Characterization of Novel S. Meliloti Genes Induced Under Nutrient Deprivation Conditions

In order to identify genes involved in multiple stress responses and their regulators, we first focused on strains carrying gene fusions induced under N- and C-deprivation and/or 02-limitation. Interestingly, we observed that only a limited number of Tn5luxAB tagged loci (13 out of a collection of 57) were induced by more than one stress. We chose strains N4 and C22 for further studies.

Strain N4 was found to carry a TnSluxAB fusion inducible under N- and C-deprivation conditions, and in the presence of tyrosine. This strain was found to be unable to grow on tyrosine as sole C-source and produced a brown pigment, which was very pronounced in the presence of 0.2% tyrosine. The tagged locus was cloned from the S. meliloti genome by the plasmid rescue procedure described by Wolk et al. (1991), its DNA sequence was determined, and it was identified as the hmgA gene, encoding homogentisate dioxygenase. This enzyme is involved in the degradation of tyrosine. A very high similarity of the deduced protein with the corresponding eukaryotic enzymes of human, mouse and fungal origins was observed (50% identity, 57% similarity) for the human homolog. This was the first report of the presence of a homogentisate dioxygenase gene in bacteria, and a phenotype of the rhizobial mutant strain in culture closely resembling human phenylketonuria, namely the production of dark pigments in the urine of PKU patients. The Tn5-luxAB induced mutation in the hmgA locus does not affect the symbiotic properties (nodulation and nitrogen fixation) of S. meliloti, but does appear to be involved in stationary growth phase survival (Milcamps, de Bruijn 1999).

S. meliloti mutant strain C22 harbors a Tn5luxAB insertion in a gene that is induced by a variety of stresses, including carbon, nitrogen or iron deprivation, oxygen limitation, as well as high salt concentrations (400 mM NaCl). This Tn5luxAB tagged locus was cloned and its DNA sequence was determined. The tagged gene is part of an operon consisting of two ORFs, ndiA and ndiB, for nutrient deprivation induced genes A and B, which are unknown in other bacteria thus far. Comparison of the deduced amino acid sequences of both ndiA and ndiB to the protein databases at NCBI did not reveal significant similarity with any known gene products (Davey et al. 2000).

4. Isolation and Characterization of Regulatory Loci, Controlling the Nutrient Deprivation Responses in Strains N4 and C22

In order to identify the regulatory genes (circuits) responsible for the multiple nutrient and/or oxygen deprivation responses observed with strains N4 and C22, two types of studies were carried out. Since both N4 and C22 fusions are N-deprivation induced, we first examined whether the genes carrying the Tn5luxAB fusions were controlled by the well known ntr (nitrogen regulation) system, first described in enteric bacteria (Merrick, Edwards 1995). Therefore, we tested the expression of the tagged loci of strain N4 and C22 in ntr A and ntrC mutants of strain 1021 (Ronson el al. 1987; Szeto et al. 1987). Neither the expression of the N4 nor the C22 fusion were found to be ntr controlled, indicating that novel regulatory pathways are likely to be involved in their expression.

Our second approach to identify putative regulators, controlling the nutrient deprivation response of the tagged N4 and C22 loci, consisted of a secondary transposon mutagenesis of these strains. The aim of this approach was to inactivate gene(s) encoding trans-acting factor(s) which would alter the observed luminescence pattern of the Tn5luxAB reporter gene fusion. As secondary mutagen, the Tn3 derivative Tnl721 was chosen (Schoffl et al. 1981). With Tn 1721, individual collections of 3600 double mutant derivatives of strains N4 and C22 were constructed

Out of 3600 strain N4 double mutant strains, two strains with a very reduced luminescence under N-deprivation conditions were selected, the Tn 1721 tagged loci were cloned and DNA sequence analysis revealed that both insertions were located in a single ORF. Database searches revealed significant similarity with a group of transcriptional regulators of the ArsR family. This gene was called nitR (nitrogen regulation; Millcamps et al. 2001), and will be one of the regulatory loci to be examined using macro- and micro-arrays (see below).

Out of 3600 strain C22 double mutant strains, one strain was isolated which no longer displayed luminescence under the conditions of N-, C-, Fe-deprivation or 02-limitation. The Tn/ 721 tagged locus was cloned and the Tn/ 721 insertion was found to reside in an ORF encoding a protein with a high degree of similarity to the mitochondrial benzodiazepine receptor of rat, human, mouse and bovine (50% identity); as well as to the outer membrane oxygen sensor protein

TspO of Rhodobacter sphaeroides (42% identity; McEnery et al. 1992; Yeliseev, Kaplan 1995). This type of outer membrane receptor (pkl8 /TspO) has been found in vertebrates and invertebrates, but has not been observed thus far in Saccharomyces cerevisiae, or in the majority of prokaryotes analyzed (Davey et al. 2000). Since a member of the alpha subdivision of purple bacteria is the likely source of the endosymbiont that gave rise to the mammalian mitochondrion (Yang et al. 1985), the finding of a pkl8/TspO ortholog in R. sphaeroides, a member of this family, is intriguing. Its rhizobial equivalent will be the subject of future studies, including macro- and micro-arrays, to identify all the loci controlled by this interesting gene (see below).

5. Macro- and Micro-array Whole Genome Profiling

In addition to using genetic approaches, we have started to exploit the complete determination and annotation of the S. meliloti genome to enter the post-genomic era by using transcriptome analysis and proteomics to study the microbial genes involved in nutrient-deprivation and stress responses, as well as those involved in symbiotic plant-microbe interactions. The interest of large-scale gene expression analysis using DNA macro- and micro-array lies in the ability to simultaneously visualize the expression patterns of all the predicted ORFs of an organism under varying physiological conditions. Moreover, the analysis of expression patterns in particular mutant backgrounds allows the identification of unique or overlapping regulatory networks controlling the responses to particular conditions.

DNA sequence annotation performed by the Toulouse group has revealed that approximately 2500 S. meliloti genes (40% of the 6204 ORFs) cannot be defined functionally or via establishing similarities with gene/protein sequences in the available databases. Therefore the true post-sequencing challenge lies in the identification of the function of these unknown genes. One of the approaches available to achieve this goal is transcriptome analysis, since genes that are coordinately expressed/regulated are likely to be involved in particular functions, especially if they are clustered in the genome.

Towards this purpose, we are currently in the process of generating or using existing protocols to:

1. Use PCR to generate secondary and unique DNA probes corresponding to every ORF of S. meliloti identified;

2. Use the corresponding DNA products, or single oligonucleotides specific to each ORF, to generate macro-arrays;

3. Use the total genomic arrays to analyze S. meliloti gene expression under a variety of environmental stress conditions (including N-, C- and 02-limitation; osmotic and oxidative stress), as well as during infection and in planta;

4. Use well characterized S. meliloti regulatory mutants generated previously in our laboratories or the regulatory loci described above, to study their genetic targets and the general genetic circuitries involved in environmental control of gene expression in this model organism.

We will investigate regulatory networks involved in several environmental challenges met by S. meliloti in the bulk soil, in the rhizosphere of plants and in symbiotic conditions in nitrogen-fixing nodules. An emphasis will be put on responses to oxygen, nitrogen, and carbon limitation and the impact of the corresponding genes and regulatory systems on the symbiotic interaction of S. meliloti with the model legume Medicago truncatula.

6. References de Bruijn FJ et al. (1995) In Tikhonovich IA et al. (ed) Nitrogen Fixation: Fundamentals and Applications, pp. 195-200, Kluwer Academic Publishers, Dordrecht, The Netherlands de Bruijn FJ (1998) In Elmerich C et al. (ed) Biological Nitrogen Fixation for the 21st century, pp. 195-200, Kluwer Academic Publishers, Dordrecht, The Netherlands Cook D et al. (1997) Plant Cell 9, 275-281

Davey ME, de Bruijn FJ (2000) Appl. Environ. Microbiol. 66, 5353-5359 Galibert F et al. (2001) Science 293, 668-672 Kaneko T et al. (2001) DNA Res. 7, 331-338

Lim PO et al. (1993) In Guerrero R, Pedros-Alio C (ed) pp. 97-100, Spanish Society for

Microbiology, Madrid, Spain Lynch JJ, Whipps JM (1990) Plant Soil 129, 1-10 Matin A et al. (1989) Ann. Rev. Microbiol. 43, 293-316 McEnery MW et al. (1992) Proc. Natl. Acad. Sci. USA 89, 3170-3174 Merrick MJ, Edwards RA (1995) Microbiol. Rev. 59, 604-622 Milcamps AP et al. (1998) Microbiol. 144, 3205-3218 Milcamps AP, de Bruijn FJ (1999) Microbiol. 145, 935-947 Milcamps AP et al. (2001) Appl. Environ. Microbiol. 67, 2641-264 Ronson et al. (1987) Cell 4, 579-581 Schoffl R et al. (1981) Mol. Gen. Genet. 181, 87-94 Sullivan et al. (1995) Proc. Natl. Acad. Sci. USA 92, 8985-8989 Szeto et al. (1987) J. Bacteriol. 169, 1423-1432 Trzebiatowski J et al. (2001) Appl. Environ. Microbiol. 67 Wolk CP et al. (1991) Proc. Natl. Acad. Sci. USA 88, 5355-5359 Yang D et al. (1985) Proc. Natl. Acad. Sci. USA 82, 4443-4447 Yeliseev AA, Kaplan S (1995) J. Biol. Chem. 270, 21167-21175


R. Palacios

Nitrogen Fixation Research Center, National University of México, Ap. postal 565-A, Cuernavaca, Morelos, México

Genomic science is changing the focus of biology from the gene to a new paradigm centered in the genome. Genomics is rapidly developing through the integration of very powerful analytical techniques - DNA sequencing and transcript and protein detection from whole genomes - with the computational tools that allow the interpretation of the information obtained. The aim of genomics is to understand the biological significance of the information coded in a genome. This includes the acquisition of the sequence and its evolutionary correlation with that of other genomes; its integrated transcriptional patterns under different conditions; the complete set of proteins present at a given time and the structural characteristics and functional metabolic pathways that such sets of proteins confer to the organism.

Nitrogen fixation research has fully entered the era of genomics. The first landmark was sequencing the symbiotic plasmid of the broad host range Rhizobium strain NGR234 (Freiberg et al. 1997) that was reported on the 11th International Congress on Nitrogen Fixation. By the time of this 13th International Congress, we have the complete sequence of the genomes of different nitrogen-fixing organisms: the Archeae Methanobacterium thermoautotrophicum (Smith et al. 1997); and the symbiotic bacteria Mesorhizobium loti (Kaneko et al. 2000) and Sinorhizobium meliloti (electronic address: toulouse.inra.fr./meliloti.html; S. Long, this volume). In addition, sequences of the genomes of other nitrogen-fixing organisms has been obtained. These include the photo synthetic bacterium Rhodobacter capsulatus (R. Haselkorn, personal communication) and the filamentous cyanobacterium Nostoc punctiforme (T. Thiel, this volume).

A large amount of sequence information now exists in regard to the symbiotic regions of different Rhizobia. In addition to Sinorhizobium meliloti, Mesorhizobium loti and Rhizobium sp. NGR234, the nucleotide sequence of the symbiotic plasmid of Rhizobium etli (G. Dávila, this volume) and that of a region of 410 kb of the chromsome of Bradyrhizobium japonicum that contains most of the symbiotic genes (Góttfert et al. 2001) have been obtained. This information has led to the discovery and characterization of new genes that affect the symbiotic process.

From an integral viewpoint, the sequence information from symbiotic regions of different organisms is now being used for comparative genomic studies. Preliminary conclusions indicate that the overall order of symbiotic genes is not conserved and that the set of symbiotic genes present in a particular genome might be acquired from different evolutionary routes. This argues in favor of reviewing the taxonomy of Rhizobium and related bacteria in the light of the new integral genomic information.

Functional genomics of nitrogen-fixing organisms has also started both at the transcriptomic level (Perret et al. 1999) and the proteomic level (Natera et al. 2000; M. Djordjevic, this volume). Furthermore, comparative and functional genomics approaches are now being used to analyze complex ecological functions such as rhizobial soil persistence and competitivity for nodule formation (F. de Bruijn, this volume) as well as to characterize novel nitrogen-fixing endophytes (E. Triplett, this volume).

The knowledge of the DNA sequence of whole genomes or replicons has suggested new forms of genomic manipulation. In our laboratory we have developed an experimental strategy -natural genomic design - to obtain derivative rhizobial populations containing alternative genomic structures (Flores et al. 2000; P. Mavingui, unpublished).

It is clear that genomic science is introducing new horizons into nitrogen fixation research and that will be a key element to achieve the long term goals in our field.


Freiberg C etal. (1997) Nature 387, 394-401

Smith DR etal. (1997) J. Bacteriol. 179, 7135-7155

GottfertM etal. (2001) J. Bacteriol. 183, 1405-1412

Perret X et al. (1999) Mol. Microbiol. 32, 415-425

Natera S et al. (2000) MPMI 13, 995-1009

Flores M et al. (2000) Proc. Natl. Acad. Sci. USA 97, 9138-9143


This work was supported in part by grants L0013N and 0028 from CONACyT-México; and IN214498 from DGAPA-UNAM, México.

Was this article helpful?

0 0

Post a comment