K.A. VandenBosch1, D.R. Cook2
'Department of Plant Biology, 220 BioSci Center, 1445 Gortner Ave, University of Minnesota, St. Paul, MN, 55108, USA
department of Plant Pathology, Room 254 Hutchinson Hall, University of California, Davis, CA 95616, USA
Approximately ten years after the elucidation of Nod factor, the lipochito-oligosaccharide made by rhizobia in response to their host plants, investigators have identified many responses to this signal in plant roots. In addition, many details of cellular function in infected roots and nodules have been discovered, principally by identifying differentially regulated genes or by analyzing function of candidate genes. However, understanding of many aspects of nodulation remains sketchy. How is the Nod factor signal transduced in the host? Do the regulatory mechanisms that govern nodulation overlap with networks controlling other aspects of development? Are components of the nodulation machinery restricted to legumes, or perhaps to a larger clade of nodule-formers that includes actinorhizal plants? Do homologous rhizobia avoid triggering a defense response, or are such means used to optimize nodule number?
Realizing that answering such complex questions would require a facile genetic model organism, investigators set out more than a decade ago to identify legume species that could serve as a model for the legume family. Medicago truncatula, a self-compatible diploid species, has attracted many adherents because of its small genome, rapid life cycle and other attributes. This growing group of investigators has contributed to the development of tools for genetic analysis. Advancement of the system has been chronicled with a series of reviews (Barker et al. 1990; Cook 1999; Cook et al. 1997; Frugoli, Harris 2001; Harrison 2000). Because M. truncatula is congeneric with alfalfa, M. sativa, it is also host to the well-characterized Sinorhizobium meliloti, which is itself the topic of genomic and proteomic analysis. M. truncatula also hosts pathogens that attack alfalfa and related legumes. This enables the use of microbial partners on M. truncatula that have been well characterized on crop species, and indicates that this annual medic will be a tractable model for understanding resistance mechanisms in the legume family.
Motivated by a desire to dissect symbioses and to answer some of the questions posed above, a group of nodulation biologists has been at the core of several parallel efforts to develop genomic tools for Medicago truncatula. In the United States, the NSF Plant Genome Project and the Noble Foundation have sponsored these efforts, while in Europe, INRA has initiated genomic analysis in France, and efforts are continuing in several countries under the auspices of the EU. This overview will emphasize the accomplishments to date of the NSF-sponsored project, and will conclude by outlining the goals of the Medicago community for developing and applying genomic tools to additional problems in the biology of this important plant family.
The large number of host genes likely to be involved in legume/microbe interactions, and the genome complexity of the major crop legumes, necessitate a coordinated effort in a legume species with tractable molecular and genetic attributes. Such a species could be used as a node, or central point, for comparison with more complex related species. This multi-institutional project, entitled "Medicago truncatula as the Nodal Species for Comparative and Functional Legume Genomics," aims to develop M. truncatula to play a pivotal role, as a point of comparison with other members of the Fabaceae. The project is advancing detailed knowledge of the structural and functional genome elements that underlie aspects of plant biology unique to, or best-studied in, legumes. The biological emphases are symbiotic nitrogen fixation and nitrogen metabolism, mycorrhizal associates and phosphate metabolism, and key legume/pathogen interactions.
Research on the project encompasses the following approaches: (1) comparative genomics, which involve comparing the organization of genes between M. truncatula and pea, alfalfa and soybean; (2) functional genomics, emphasizing characterization of expressed genes and initiating large-scale analysis of expression patterns to study gene function; and (3) bioinformatics, which has emphasized development of database resources of the analysis and dissemination of M truncatula genome information collected from the NSF project and other sources. The bioinformatics tools support both of the other goals, and relate the two to each other. The project web site, http://www.medicago.org, also serves as a discussion forum for the Medicago community, a catalog of published works, a directory of M. truncatula researchers, and points to other Medicago research sites. The discussion below highlights several advances in the project.
The close relationship between M. truncatula and M. sativa is reflected in the extensive macro- and microsynteny between the two species. To date, only one locus, the nucleolar-organizing region, occupies a position on different linkage groups (LGs) in the two species. This enabled investigators to use the same nomenclature for LGs and chromosomes in both Medicago species. Even more important from a practical standpoint, the markers from the M. sativa map are being adapted for placement on the M. truncatula map as a means to swiftly populate the latter, and to enable comparative mapping projects in the two species. Gyorgy Kiss discusses one such successful comparative mapping project, and the contribution towards map-based cloning of an important symbiotic gene, in a chapter in this volume. Comparative mapping between M. truncatula and pea (Pisum sativum) indicates strongly conserved gene arrangement on five of the eight LGs in M. truncatula. This is expected to assist cloning efforts of loci identified in pea, where map-based cloning efforts are severely hampered by a large genome and ample repetitive sequences. One such project, directed by René Geurts, is underway in Wageningen Agricultural University (Gualtieri et al. submitted). The extent of microsynteny among these taxa is being further documented with an intensive study LG 5 in Medicago spp., which corresponds to LG 1 in pea.
The microsyntenic relationships among temperate legumes were expected because of their close phylogenetic relationships. Therefore, we are interested to see to what extent M. truncatula genome structure may be a good model for legumes that are less closely related, such as the tropical legume tribe Phaseolidae that contains important crops such as soybean (Glycine max) and beans (Phaseolus spp.). This work is being carried out collaboratively between the Cook lab and Nevin Young at the University of Minnesota. To date, macro- and microsynteny between M. truncatula and soybean has been demonstrated, although quantification of similarity is complicated by duplication within the soybean genome, and limited polymorphism in the soybean mapping populations. Preliminary work shows that synteny between M. truncatula and G. max is somewhat less than between duplicated regions of the G. max genome, but far exceeds synteny between the model dicot Arabidopsis thaliana and either of the two legumes (Yan et al. submitted). We predict that M. truncatula will have experimental value for at least some regions of the soybean genome, following continued development of a pan-legume map. A major, on-going objective that will facilitate this is the development of gene-based markers that will be used to coordinate the maps of M. truncatula and crop legumes.
Integration of the M. truncatula genetic and cytogenetic maps was achieved in collaboration with Ton Bisseling and colleagues in Wageningen. Working with pachytene chromosomes, the Dutch group has mapped at least two bacterial artificial chromosome (BAC) clones from the
M. truncatula genomic library onto each chromosome. Visualization indicates that the chromosome arms are composed of long, typically uninterrupted stretches of euchromatic DNA (Kulikova et al. 2001). Heterochromatin is largely clustered in the pericentromeric regions, and is estimated to comprise up to 80% of the M. truncatula genome. This simple genome arrangement, which predicts organization of the gene-rich regions in as little as 100 Mbp of mostly continuous euchromatin, will greatly aid map-based cloning efforts. In the longer term, this will also facilitate sequencing of the gene-rich portions of the genome.
Towards our goal of identifying legume genes involved in nutrient acquisition and microbial interactions, 17 cDNA libraries were made and subjected to high throughput sequencing. Five of these libraries are from Sinorhizobium-mocu\s.ted or infected tissues, contributing more than 19,000 ESTs from the approximately 55,000 ESTs from the project to date. All ESTs have been deposited in GenBank and used (with other public data) to construct a gene index that represents a minimally redundant set of sequences. The gene index (MtGI), constructed at The Institute for Genome Research, groups M. truncatula ESTs into tentative consensus sequences (TCs) according to sequence overlap (http://www.tigr.Org//tdb/mtgi). The third release of MtGI (April, 2001) predicted -13,000 TCs and -19,000 singletons, totaling -32,000 predicted unique sequences. The fourth edition of MtGI was due out in August, 2001. MtGI also assigns probable function to expressed genes, based on BLAST results.
The gene index can also indicate expression patterns of highly expressed genes, based on tissue of origin of the ESTs in the TCs. Analysis reveals many examples of tissue-specific patterns of expression. For example, about 3000 TCs, or nearly one fourth, are composed of ESTs exclusively from tissues responding to microbial infection or elicitation. Of these, approximately 1750 TCs appear to have a symbiosis-specific pattern of distribution, including about 900 TCs unique to rhizobium-inoculated tissues, and about 300 common between nodules and mycorrhizae. In addition to nodule-specific genes and putative markers of early responses to rhizobium, we have also identified root-specific markers and expressed sequences common to both symbiotic and pathogenic responses. This approach permits the identification of marker genes to be used as developmentally specific controls in planned transcriptional profiling studies.
Our long-term aims are to establish a unigene set of this EST resource and to investigate thoroughly the genome-wide patterns of gene expression by using hybridization of probes to DNA microarrays. As a first step, we have assembled a small-scale array of-1000 cDNA clones (called the "kiloclone set"), that contains positive and negative controls and some developmental markers. The kiloclone set is also rich in clones encoding proteins with putative functions in signal transduction, transcriptional regulation, control of cell division and cell death, pathogen responses, secondary metabolism, and a number of genes of unknown function. Hybridization experiments to date have monitored differences in gene expression during the early steps of the symbiotic process or during pathogenic interactions, compared to uninoculated plants. These early experiments have instilled confidence in the assays by demonstrating both technical and biological repeatability. Novel genes expressed during nodulation and responses to the root rot pathogen Phytophthora medicagenis have been identified to date. Verification of the identity of the cDNA clones and their expression patterns is underway.
5. Concluding Remarks: The Long-term View
On-going work on our project will continue to emphasize structural, comparative and functional genomics approaches. The specific objectives of the structural genomics team are to create a comprehensive physical map for M. truncatula, and to determine the correspondence between 3000 ESTs and BAC addresses in the physical map. The combination of a physical map, BAC end sequencing, and the assignment of the location of several thousand ESTs are expected to delimit the gene-rich portions of the genome, and to lay the foundation for developing a low-pass gene inventory in these regions. The M. truncatula gene-based markers will also continue to be evaluated for their utility as markers for the genetic maps of alfalfa, pea and soybean. This comparative approach will be instrumental in extending the impact of M. truncatula genomics to other legumes. Fruits of the functional genomics activities will include assembly of a set of cDNA clones of minimal redundancy that represents the breadth of transcribed genes identified to date. The objective will be to include at least 24,000 representative clones in this "unigene set", which will be freely available to the public. The second major aim will be to use the unigene set in DNA microarray experiments that will monitor global patterns of gene expression. Of relevance here, we plan to identify genes that are differentially regulated from the earliest perception of Nod factor through nodule senescence, in wild type plants and developmental mutants
Genomic research on Medicago truncatula is having a significant impact on legume research, worldwide. Currently a database of M truncatula researchers lists approximately 197 researchers, from 16 different countries (http://www.medicago.org). A strength of the Medicago community is its open communication and the development of public resources. A series of international workshops on this species has been key to involving a broad community of researchers in the development of resources, such as BAC and EST libraries, genetic markers and maps, and curated populations containing both natural and induced genetic variation. Beyond nodulation, these tools are being applied to many topics, including productivity and forage quality, seed development and nutrition, responses to abiotic and biotic stresses, and natural product biosynthesis. The long-term impact of these efforts will be to integrate genetic and functional information in Medicago truncatula specifically, and legumes generally. This knowledge will enable more efficient cloning and characterization of valuable genes and traits, and become the focus of crop improvement strategies throughout the world. Thus, the genetic and genomic system that started as a means to better understand nitrogen fixation has developed into a major resource for legume biology.
Barker DG et al. (1990) Plant Mol. Biol. 8, 40-49 Cook D (1999) Curr. Opin. Plant Biol. 2, 301-304 Cook D et al. (1997) Plant Cell. 9, 275-281 Frugoli J, Harris J (2001) Plant Cell. 13, 458-463 Harrrison M (2000) Trends Plant Sci. 5, 414-415 Kulikova O et al. (2001) Plant J. 26
We gratefully acknowledge support from the NSF Plant Genome Project, award number DBI-9872664, and the contributions of our co-principal investigators Steve Gantt, Michael Hahn, Maria Harrison, Dongjin Kim, Ernie Retzel, Debby Samac, Chris Town, Carroll Vance, and Nevin Young. A special acknowledgement is due to post-doctoral fellows and other associates on the project, including Maria Fedorova and Jinyuan Liu, who contributed to EST library construction, and Gabriella Endre and Silvia Penuela, who have pioneered the cDNA microarrays. We thank Jennifer Cho and Joe White for their roles in developing and utilizing MtGI. Lastly, we are grateful to collaborators Ton Bisseling, René Geurts and Gyorgy Kiss who have contributed to construction and utilization of comparative genomic tools outlined here.
Was this article helpful?