Medicago truncatula on the Move!

Plant biology is moving rapidly. With the recent completion of the sequence of the Arabidopsis genome (The Arabidopsis Genome Initiative, 2000), researchers have turned their attention to the genomes of other crop plants (Adam, 2000). Many agronomically important crop plants are legumes, such as soybean, pea, and alfalfa. However, the size and complexity of these genomes makes them unwieldy and has slowed progress on the genetic characterization of these crops. Medicago truncatula recently emerged as a model plant for legume genetics and genomics (Cook, 1999). The small diploid genome, autogamous genetics, and ease of transformation make this close relative of alfalfa (Medicago sativa) a good model system. The recent development of genetic and genomic tools for M. truncatula research has propelled M. truncatula into the forefront of legume research as an ideal legume model. Advances in M. truncatula genomics were the subject of a recent workshop preceding the 9th Annual Plant and Animal Genome Meeting in San Diego, California, on January 11 and 12.

The international scope of this workshop underscored the interest in M. trun-catula as a model system for studying plant–microbe interactions, in particular rhizobial and mycorrhizal symbioses. More recently, interest has turned to M. truncatula as a system for examining the very rich production of secondary metabolites by legumes and legume-specific disease resistance (Cook et al., 2000; Harrison, 2000). This workshop, which started out three years ago as a small band of National Science Foundation (NSF) genome grant collaborators sponsored by the NSF Plant Genome Project (award number 9,872,664), has grown into an international effort spanning national boundaries and legume species. This year's participants represented not only the original NSF genome grant collaborators but members of the Institut National de la Recherche Agronomique–Centre National de la Recherche Scientifique (INRA-CNRS), German mycorrhizal and European Union (EU) M. truncatula genome projects (funded by the Deutsche Forschungsgemeinshaft and the European Union 5th Framework Program, respectively), the Noble Foundation M. truncatula project, and legume scientists working with other species such as pea, soybean, and alfalfa. What emerged from the sharing of data, questions, ideas, and future plans was a coordinated effort using worldwide resources to develop M. truncatula as a model plant as easy to use as Arabidopsis with broad applicability to other legume systems.

EST SEQUENCING–THE 100,000 MARK

The most publicly visible advance in the development of M. truncatula as a model legume system is the large number of expressed sequence tags (ESTs) deposited in the public domain. The INRA-CNRS-Genoscope, Noble Foundation, and NSF projects all include activities that construct cDNA libraries and sequence ESTs, with the result that M. truncatula now contributes approximately as many ESTs to EST databases as soybean and Arabidopsis and likely will soon pass Arabidopsis. Data from three EST projects were presented at the meeting. Pascal Gamas (INRA-CNRS, Toulouse, France) presented data from the INRA-CNRS-Genoscope EST project, Gregory May (Noble Foundation, Ardmore, OK) discussed the Noble Foundation's EST project, and the NSF project was represented by Chris Town (The Institute for Genomic Research, Rockville, MD). Town also presented data on an EST database, the Medicago truncatula Gene Index, that integrates all publicly available M. truncatula ESTs. Together, this consortium of researchers has sequenced ESTs from 24 cDNA libraries of diverse origins. Not only do the libraries encompass most major plant tissues, but they also incorporate developmental stages and treatments with microbial pathogens, bacterial and fungal symbionts, insect pests, and abiotic stressors. Combining and analyzing the data from several efforts has proven extremely powerful. For example, each EST can be traced to the library from which it was sequenced. By cross-referencing this information with the compilation of homologous ESTs into tentative consensus sequences, the Medicago truncatula Gene Index can provide a first indication of expression patterns for the corresponding gene. At the time of the last Gene Index release in December 2000, almost 89,000 M. truncatula ESTs were available in public databases, but by the January date of the meeting the number had grown to more than 105,000 and continues to climb. The 88,906 ESTs used to generate the latest release fall into 10,160 tentative consensus sequences comprising 68,844 ESTs and 19,985 singletons, for a total of 30,145 unique sequences (for comparison, Arabidopsis has 34,491 unique sequences in the Gene Index).

DEVELOPMENT OF A UNIGENE SET AND LARGE SCALE EXPRESSION PROFILING

Although the generation of ESTs has value in and of itself, one focus of all the genome projects represented was to produce these sequences for the development of microarrays and markers. Kate VandenBosch (University of Minnesota, Minneapolis/St. Paul), Helge Kuester (Universitaet Bielefeld, Germany), and Gregory May presented progress and future directions for the development of a DNA-chip-microarray resource. VandenBosch presented exploratory array data gathered during the development phase of the chip resource under the NSF Genome Project and discussed progress toward the development of a “universal unigene set.” The unigene set will be a defined set of nonredundant cDNA clones that will be made available to the community for use in microarray and other experiments. This will help standardize expression profiling experiments and will allow access to the technology by more researchers. The unigene set will evolve over time, beginning with 6000 genes and growing, depending on future funding, to an expected 20,000 to 30,000 clones covering the breadth of the transcriptome. These genes will be chosen from all of the genes available in the database and sequenced from both ends to confirm identity and obtain complete transcript information where possible. Rod Wing has agreed to add the future EST resource to the M. truncatula bacterial artificial chromosome (BAC) libraries that the Clemson University Genomics Institute currently houses and distributes, making the unigene set another publicly available resource. Kuester discussed the EU project's similar goals to use both microarrays and macroarrays, emphasizing the latter as another way to bring the technology to the average user. In addition to transcriptome analysis using microarrays and SAGE (Serial Analysis of Gene Expression), the Noble Foundation has expanded M. truncatula expression profiling to include proteomics and metabolic profiling. It has undertaken a large-scale project in M. truncatula, using two-dimensional gels and matrix-assisted laser desorption ionization time-of-flight mass spectrometry to identify proteins and electro-spray ionization HPLC mass spectrometry to identify metabolites. Linking these data to the microarray and genome sequencing data should prove extremely fruitful.

AN EXPONENTIAL LEAP IN THE NUMBER OF T-DNA–TAGGED LINES

Crucial to the development of M. truncatula as a model system is the availability of reverse genetic tools. Pascal Ratet (CNRS, Gif sur Yvette, France) and Maria Harrison (Noble Foundation) each discussed the T-DNA–tagged lines being generated by their groups, and Maria Fedorova presented the progress of parallel efforts by a collaboration among Steve Gantt, Debby Samac, and Carroll Vance (University of Minnesota). Ratet's group used a transformation and regeneration protocol (Trinh et al., 1998) and a gene fusion vector to generate 350 T-DNA–tagged lines. Tagged loci have been characterized, and another several thousand lines are planned in the frame of the European M. truncatula genome project. Harrison's group developed a vacuum infiltration protocol for M. truncatula (Trieu et al., 2000) and combined this with an activation tagging vector (Weigel et al., 2000) to generate 180 lines in a pilot experiment. These lines are currently in the T2 or T3 generation and demonstrate both dominant and recessive phenotypes. Scaling up from the initial experiment, more than 10,000 plants have been infiltrated to date, which should result in between 9000 and 51,000 lines, based on the efficiencies reported by Trieu et al. (2000). Transformation is ongoing at 3000 plants per week. The Minnesota group has conducted a small scale (4000 seedlings) mutagenesis experiment using vacuum infiltration of seedlings and various vectors and plans to perform the same analysis with the promoter trap vectors before embarking on large scale production of tagged lines. When these projects are combined, the number of T-DNA–tagged lines in M. truncatula will increase from a few hundred a year ago to a conservative estimate of 125,000 lines within another year.

A COMPREHENSIVE M. TRUNCATULA GENETIC MAP

Six years ago, data on the first M. truncatula mutant, TE-7, was published (Benaben et al., 1995). Since that time, there has been an explosion in the number of mutants isolated and described (Penmetsa and Cook, 1997, 2000; Catoira et al., 2000; Nakata and McConn, 2000; Wais et al., 2000). At the time that TE-7 was identified, there were few genetic markers and there was no map, genetic or physical. A highlight of this meeting was the presentation of genetic maps by Thierry Huguet (INRA-CNRS, Toulouse, France) and Dongjin Kim (University of California, Davis). The two maps were based on physical differences between different pairs of M. truncatula ecotypes, using genotype Jemalong A17 as the recurrent parent. Huguet has identified the location of 240 amplified fragment length polymorphism (AFLP) markers and is currently adding restriction fragment length polymorphisms (RFLPs) and microsatellites to this map. In addition, members of his laboratory, in collaboration with J.M. Prospéri (INRA, Montpellier, France), has generated 195 recombinant inbred lines, 75% of which are at the F7 level. These lines will be available publicly. In a parallel effort, Dongjin Kim reported that the Cook laboratory has mapped 230 codominant polymerase chain reaction markers, 150 of which are linked to clones from a growing BAC genomic library resource (now estimated at 32-fold genome coverage). More than 100 of these markers link ESTs to the map, thus connecting the functional and structural genomics efforts of this group. An ongoing collaboration between the laboratories of Nevin Young (University of Minnesota) and Thierry Huguet to link these two genetic maps is resulting in the identification of mi-crosatellite markers that are poly-morphic among all three mapping ecotypes. Integration of a cytogenetic map being constructed by members of Ton Bisseling's laboratory (Wageningen University, the Netherlands) with the genetic map produced by Kim and Cook has helped resolve some ambiguities, allowing 35 BACs to be placed on both maps. These genetic resources will be complemented by a project at the University of California, Davis, where researchers are initiating the development of a complete physical map for M. truncatula.

The strength of these extensively linked genetic and physical maps is that after identifying a linked marker, a researcher can go directly to a BAC clone containing genomic DNA from the region of interest. In many cases, these BAC clones have already been linked into a contig, thus eliminating several steps in the walk from phenotype to map position and ultimately to gene identification.

THE STRUCTURE OF THE ARABIDOPSIS GENOME IS AN INSUFFICIENT MODEL FOR LEGUMES

The recent completion of the Arabidopsis genome sequence has provided an enormous boost for the plant field as a whole. However, reports of very low levels of microsynteny between Arabidopsis and M. truncatula suggest that the structure of the Arabidopsis genome may not be a good indicator of how legume genomes will look. Using different approaches, Nevin Young reported a global estimate of only 8% microsynteny between Arabidopsis and M. truncatula genomes, whereas researchers at the University of California, Davis, estimate ∼10% frequency of microsynteny based on survey sequencing of more than 20 BAC clones in M. truncatula. Nevertheless, some genome regions are conserved between these two models, and Jongmin Baek from the Davis group reported detailed sequence analysis of a segmental duplication in M. truncatula that exhibits conservation with the Arabidopsis genome.

In addition, an analysis of disease resistance genes by Steve Cannon (University of Minnesota) and Nevin Young in collaboration with Doug Cook has identified resistance genes that have radiated in legumes but that are poorly represented or absent in Arabidopsis. This work suggests that there are some very ancient differences in disease resistance genes and that Arabidopsis may be a poor model for studying disease resistance in the crop legumes.

M. TRUNCATULA AS A NODAL SPECIES FOR COMPARATIVE LEGUME GENOMICS

Influenced by the fact that Arabidopsis may not be a good model for the structure of legume genomes, a large focus of this meeting was on possible synteny between M. truncatula and other legumes, with the goal of ultimately developing a composite legume genomic map. Attention centered on three agronomically important legumes: alfalfa, pea, and soybean. Gyorgy Kiss (Hungarian Academy of Sciences, Szeged, Hungary) and Dongjin Kim have placed more than 60 markers on a comparative map linking M. truncatula with its close relative alfalfa, demonstrating a high level of macrosynteny between these two species.

A survey of microsynteny between soybean and M. truncatula starting with ∼20 Mbp of the soybean genome (50 BAC contigs) found very high levels of microsynteny between the two genomes. Nevin Young reported that 54% of the soybean BACs exhibited some microsynteny with M. truncatula, of which most (>70%) exhibited extensive conservation. These observations suggest that information derived about the M. truncatula genome has enormous implications for soybean.

Evidence for such high levels of microsynteny between M. truncatula and other legumes has prompted the development of cross-genome markers that would link the genetic maps of several important legume species. Dongjin Kim reported on progress in developing cross-genome markers for M. truncatula, pea, mung bean, soybean, and Lotus japonicus. The idea behind such a composite map linking these genomes is that researchers studying a gene in one legume species could easily move to another species to map and clone it and then return to the original species for analysis. At 450 Mbp, the M. truncatula genome is only a few times larger than that of Arabidopsis, but it is 10 times smaller than that of pea. Unlike alfalfa and soybean, M. truncatula is a diploid and thus amenable to genetic analysis. The ease of genetic analysis in M. truncatula and its more manageable genome size makes it the species of choice for mapping and cloning. In concept, M. truncatula would be the nodal species that would link the genetic maps of the larger legume genomes together. In addition, these cross-genome markers would link these maps to that of L. japonicus, another legume species that has been the focus of a large molecular genetic effort.

IDENTIFICATION OF GENE-RICH REGIONS

One of the striking features of legume biology is the variation in genome size among closely related species. The pea genome is almost 10 times the size of the M. truncatula genome, even though their haploid chromosome number differs by only one. A comparison of the pea, alfalfa, and M. truncatula genomes by Noel Ellis (John Innes Centre, Norwich, UK), Gyorgy Kiss, and members of the Cook laboratory demonstrates that the difference in size between the pea and Medicago genomes is not due to genome duplication. Rather, the large size of the pea genome seems to be due to the recent accumulation of retrotransposons and noncoding regions of DNA (Noel Ellis).

These observations raised the issue of gene-rich regions. Identification of such regions could increase the productivity of a cloning or massive sequencing effort. In the process of creating a cytogenetic map, members of the Bisseling laboratory have determined that the genome of M. truncatula has a remarkably efficient organization. One estimate suggests that ∼80% of the genome is organized into pericentromeric heterochromatin, whereas the chromosome arms are composed primarily of euchromatic DNA. Thus, the gene-rich region of the genome may be condensed into as little as 100 Mbp of DNA.

Sequencing the M. truncatula genome may not be far off. Jean Denarié (INRA, Toulouse, France) reported on recent international workshops held in Europe and the United States to identify priorities in the use of M. truncatula to facilitate the breeding of crop legumes (Cook and Denarié, 2000) and the possibility of a massive M. truncatula sequencing effort of international proportions. A follow-up meeting has been planned for the spring to discuss strategic plans.

BIOINFORMATICS RESOURCES FOR M. TRUNCATULA

To keep pace with the rapid increase in M. truncatula genomics, several bioinformatics groups have established searchable databases that assemble M. truncatula EST and/or BAC sequences, report homologies, and are freely accessible through the World Wide Web. The National Center for Genome Resources (Santa Fe, NM) in collaboration with the Noble Foundation has established the Medicago genome initiative site at http://www.ncgr.org/research/mgi/ (Bell et al., 2001). Callum Bell (National Center for Genome Resources) pointed out that, as in other informatics fields, Medicago informatics researchers are grappling with how to reduce redundant effort, break down barriers that separate data types, and cross species boundaries. Bringing the informatics leaders from several projects together at this meeting was seen as a step toward solving these problems. The EU and INRA data are available at http://sequence.toulouse.inra.fr/Mtruncatula.html, with a searchable index under development. The NSF project has fostered the development of two databases: The Institute for Genomic Research Medicago truncatula Gene Index, which is accessible at http://www.tigr.org/tdb/mtgi/, and a species database site (MtDB) at http://chrysie.tamu.edu/medicago/. Ernie Retzel (University of Minnesota) presented a new master link to Medicago-related sites under development at http://www.medicago.org/. This site also will contain comparative information for the soybean project housed at the University of Minnesota as well as protein family assignments. These sites are being updated regularly and are proving to be an invaluable tool for M. truncatula genetics and genomics research.

UPCOMING EVENTS

Perhaps the most important outcome of the meeting was the heightened interactions between the research groups, large and small, working on M. truncatula. The international nature of the effort should be evident this summer at the 4th Workshop on M. truncatula, a satellite meeting of the International Molecular Plant–Microbe Interactions conference (http://www.plantpath.wisc.edu/mpmi/satellite.html). In Europe, a European Molecular Biology Organization practical course in November 2001, at Gif sur Yvette, France will introduce the system to new researchers (http://www.isv.cnrs-gif.fr).

CONCLUSIONS

This conference highlighted the recent explosion in the development of genetic and genomic tools for M. truncatula. The reports of the sequencing of more than 100,000 ESTs, extensive genetic maps linked to BAC genomic clones, progress in activation tagging, and the initiation of large-scale T-DNA mutagenesis demonstrate that the genetic and genomic tools are finally in place to use M. truncatula as a model legume. This strategic focus on a tractable legume and the development of a composite legume genetic map will enable researchers to move easily between M. truncatula's simple genome and those of other legumes. With the rapid advances described above and the possibility of initiating a genome sequencing effort, M. truncatula will not be just on the move, it will be flying!