Beyond the Arabidopsis Genome: Opportunities for Comparative Genomics

Like most higher eukaryotes, flowering plants are believed to contain surprisingly similar numbers of genes. Nevertheless, angiosperm genome sizes vary over a wide range—from 50 Mb to over 120,000 Mb. Comparative mapping has shown that numerous alterations contribute to genomic diversity among plants. Over time, chromosomes are broken, reassembled, partially or wholly duplicated, and even eliminated, ultimately resulting in reproductive isolation and speciation. However, the mechanisms that create such variation, and the evolutionary forces that fix these changes, are not well understood. Comparative analyses of plant genomes promise to clarify the selective pressures driving these changes; such investigations will elucidate alterations at the level of whole genomes, as well as those at the level of specific sequences, including genes, repetitive elements, and other non-coding regions. Although low-resolution genetic maps can identify gross chromosomal alterations, a clear understanding of the mechanisms behind these changes requires multispecies sequence comparisons. Such analyses reveal the composition, organization, and functional components of genomes and provide insight into regional differences in composition between related species. In addition, sequence comparisons elucidate evolutionary history; for example, the stepwise accumulation of nucleotide insertions/deletions (indels) only becomes clear with the analysis of multiple species. Comparative sequence analysis also aids in gene prediction and sequence annotation, and facilitates the identification and definition of regulatory elements, including promoters, enhancers, and transcription factor-binding sites (Kent and Zahler, 2000; Koch et al., 2001b). The recent analysis of the sequence of the Arabidopsis genome highlighted unexpected aspects of its composition, organization, and function (Arabidopsis Genome Initiative [AGI], 2000). The questions raised by these observations can best be approached through comparative genomics. For example, although Arabidopsis is considered a “true” diploid, its genome has undergone major duplication events, followed by extensive rearrangements and chromosome fusion and loss, hypothesized to have shifted the haploid chromosome number from 4 to 8 and then to 5 (AGI, 2000; Vision et al., 2000). Interestingly, however, evidence of duplications was not found in the sequenced portions of the centromere regions. All five centromeres contain tracts of unique DNA, interspersed with similar types of transposable elements and interrupted by large tandem arrays of satellites. Aside from the repetitive sequences, pair-wise comparisons of the centromere regions did not identify blocks of unique sequence indicative of ancient duplication events (AGI, 2000). In addition to large segmental duplications on the chromosome arms, Arabidopsis also contains a prevalence of gene families, many of which are the result of tandem duplications of individual genes, rather than redundancy of entire chromosome segments. Nearly 40% of the predicted genes in the Arabidopsis genome belong to families that contain more than five members (AGI, 2000). Through studies of related species, it will become possible to discern the timing of genome duplications, the types of DNA eliminated, and the mechanisms responsible for rearrangements and deletions. Moreover, analysis of the genomes of related species will clarify how gene families expand, contract, and diversify. Examination of relatives containing different types of genome duplications will also reveal the changes that occur after such events, including mechanisms that are activated after large-scale genomic perturbation. Clearly, comparative genomic approaches would provide enormous benefit toward understanding the origins of the Arabidopsis genome, as well as other plant genomes. Although previous comparisons with genes from yeast, flies, worms, and mammals have provided functional clues for approximately 70% of Arabidopsis genes, the vast evolutionary distances that separate these species restrict such comparisons to coding regions (AGI, 2000). Even the genomes of Arabidopsis and rice (Oryza sativa), which are separated by approximately 200 million years (MY), have substantially diverged (Wolfe et al., 1989), making the incremental stages of evolutionary change difficult to grasp (Goff et al., 2002; Yu et al., 2002). In this Update, we explore the utility of a comparative genomics approach that relies on Arabidopsis and several other species within the Brassicaceae family. We discuss discoveries made from prior comparisons within the family, prospects for additional genomic studies, and their potential for significantly improving our understanding of plant genomes. Finally, we discuss the community resources required to launch an effort with the necessary breadth to address a wide range of sequence-based evolutionary questions, and suggest a set of candidate species for genomic comparisons in the Brassicaceae. Morphology of several species in the Brassicaceae family. Species left to right include: Thlaspi arvense,Arabidopsis suecica, Olimarabidopsis pumila, and Capsella rubella. A through D, Overall adult plant morphology; plants were grown in 3-inch pots. E through H, Variations in mature fruit shape and size. I through L, Close-up of flowers. Size bars represent 1 mm. The Brassicaceae thrive in a variety of habitats; they are concentrated in the northern temperate regions, the Mediterranean, and the mountains of southwest and central Asia, and have migrated extensively to assume a worldwide distribution (Rollins, 1993). Within this family are species that have adapted to diverse ecological settings, providing vast natural variation in developmental, biochemical, and physiological phenotypes. Some Brassicaceae species grow in the high altitudes of the Himalayas (e.g. Crucihimalaya himalaica), others in aquatic environments (e.g. Nasturtium officinale), and others in desert conditions (e.g. Nerisyrenia camporum). Species such as Lesquerella filiformis, with a range restricted to southwestern Missouri, require specific ecological niches, whereas others like Shepard's purse (Capsella bursa-pastoris) are successful colonizers worldwide (Rollins, 1993). Comparative genetic investigations of such species promise to provide insight into the mechanisms that restrict the range of an organism or allow it to conquer diverse environments. Arabidopsis researchers have exploited the variation among Arabidopsis ecotypes to identify genes that contribute to difference in flowering time, plant size, drought tolerance, and variation in glucosinolate profiles for regulation of plant/insect interactions (for review, seeAlonso-Blanco and Koorneef, 2000). With the appropriate genomic tools, it will become possible to elucidate the genetic mechanisms that underlie the ability of Brassicaceae species to adapt to almost any environmental challenge, yielding insight into how specific genes or biochemical pathways evolve. An investment in the resources required for comparative genomics will not only clarify the underlying genetic basis for natural adaptations, but may also enhance the economic value of the Brassicaceae family, which currently ranks fifth in oil production worldwide and comprises over 10% of the U.S. vegetable crop acreage (Al-Shehbaz, 1984; http://govinfo.library.orst.edu). Cross species comparisons of genomes within the Brassicaceae have primarily been limited to plants in the Brassica genus with the goal of identifying genes useful for crop improvement. These studies relied on Arabidopsis as a reference genome and focused on determining the extent of colinearity of molecular markers on genetic or cytological maps. Such investigations have demonstrated, for example, that the Brassica nigra and Arabidopsis genomes, which diverged 16 to 21 MY ago, share many dispersed chromosomal segments estimated at 8 cM in length, or approximately 1.7 Mb (Lagercrantz, 1998). Similar conclusions have been reached comparingBrassica oleracea and Brassica napus (Cavell et al., 1998; Lan et al., 2000). Both physical and genetic mapping studies have also indicated that segments of Arabidopsis chromosomes are often present in triplicate in the diploid Brassica spp. genomes (Brassica rapa, B. oleracea, and B. nigra), suggesting that modern Brassica spp. likely evolved from an ancient hexaploid (Lagercrantz, 1998; Jackson et al., 2000; O'Neill and Bancroft, 2000). This genomic triplication event likely occurred after the divergence of the Brassica spp. and Arabidopsis lineages (Lagercrantz and Lydiate, 1996). On a more local scale, DNA sequence comparisons have been used to assess synteny at the level of individual genes, confirming large colinear stretches of sequence in the Arabidopsis andBrassica spp. genomes. O'Neill and Bancroft (2000) used sequence data from a 222-kb region of Arabidopsis chromosome 4 (duplicated on Arabidopsis chromosome 5) and compared its gene content and order with homoeologous (ancestrally related) segments of theB. oleracea genome. The two species exhibit high levels of gene conservation and colinearity, although in the regions investigated, B. oleracea lacks homologs of some Arabidopsis genes. Similarly, searches for a B. oleracea counterpart of the approximately 15-kb Arabidopsis ABI-Rps2-Ck1 gene segment found disruptions of gene content in regions that otherwise shared high levels of sequence similarity and a conserved gene order (Quiros et al., 2001). Recently, larger scale sequence-based comparisons of other species in the Brassicaceae have been initiated, including wild relatives of Arabidopsis such as C. rubella (Acarkan et al., 2000;Rossberg et al., 2001). C. rubella and Arabidopsis, estimated to have diverged 11 to 14 MY ago, have chromosomal segments that are extremely similar in gene content, orientation, and order. Studies so far indicate that genes are conserved at >90% nucleotide identity, exon sizes and boundaries remain constant, and only minor differences are observed in the length of introns. Further comparative studies between C. rubella and Arabidopsis are required to determine if these levels of genome similarity are representative. Thus, comparative studies performed to date, primarily between theBrassica and Arabidopsis genera, suggest several important themes: (a) conservation of gene sequence, content, and order are common, a situation that facilitates cross species mapping and identification of syntenic regions and gene homologs; (b) theBrassica genus exhibits rapid rates of chromosome evolution, characterized predominantly by duplications, rearrangements, and fusions (Lagercrantz, 1998); (c) in both the Arabidopsis andBrassica genomes, gene density is high, and repetitive DNA content is low, particularly when compared with grass genomes (Kumar and Bennetzen, 1999). Although it is premature to suggest these themes are general attributes of the Brassicaceae, it is clear that questions of gene duplication, chromosome rearrangement, and alterations in ploidy can be appropriately addressed in this family. Augmenting the comparisons between Brassica spp. and Arabidopsis with additional comparisons to closely related species will aid in understanding the forces driving speciation, including reorganization of chromosomes and changes in genome composition and size. Because relatively few genomic analyses of genera within the Brassicaceae have been conducted, it is not clear if the duplications, fusions, and rearrangements that have occurred in Brassicaspp. genomes are a consequence of domestication and selective breeding, or if they are also characteristic of wild relatives. AlthoughBrassica and Arabidopsis species diverged within the last 20 MY, the significant differences between their genomes blurs the incremental processes that gave rise to differences in chromosome structure and number. Furthermore, although the analysis of large orthologous regions of Brassica spp. and Arabidopsis chromosomes can characterize small-scale alterations in chromosome structure, this approach cannot adequately address the mechanistic changes that drive the formation of new plant species. Instead, a stepwise analysis of multiple species, phylogenetically distributed across the family, is required. Expanding genomic comparisons to additional species within the Brassicaceae will provide opportunities to examine the types of genomic changes that contribute to variation in genome size, and how such alterations are manifested phenotypically. Arabidopsis was selected for complete genome sequencing based, in part, on its extremely small genome (140–175 Mb, Bennett and Smith, 1991); however, this small genome may require mechanisms to minimize its genome that are not representative of other plants. Although comprehensive molecular data is often lacking from plants with relatively large genomes, instances where such data are available have demonstrated the predictive value of genome size for some plant traits. Weedy species tend to have smaller genomes, and in some taxa there is a strong correlation between genome size and flowering time (for review, see Bennett et al., 1998,2000). Measurements of genome size (C value) are currently under way for a large number of Brassicaceae species (H.J. Price and S. Johnston, personal communication). This cost-effective procedure, combined with simple molecular assays that tabulate the amount of repetitive satellite, telomere, and rDNA sequences, will clarify whether repeat amplification contributes significantly to Brassicaceae genome evolution. In addition, comparative genetic mapping can be used to assess the degree of genome-wide colinearity in species with different genome sizes, and reveal large-scale rearrangements, duplications, and deletions. Comparative mapping has shown that genome colinearity persists across vast evolutionary distances in grass genomes, and that the differences between species result more from expansion of intergenic regions than chromosomal rearrangements (Gale and 1998). In addition, in and can comparative genetic mapping to assess the distribution of genomic duplications, and how DNA contribute to genome has been useful in comparisons between the Brassica and Arabidopsis species to size and number of large chromosomal et (2000) used a Arabidopsis genomic as a in B. rapa, and demonstrated the large is present multiple in the B. The regions in B. not to be significantly larger than the Arabidopsis that this genomic region has not the divergence of the Brassica and Arabidopsis in ploidy often contribute to in genome more in plants than in In changes in ploidy have been a major in angiosperm genome evolution, the duplication of an entire genome or the of two genomes to a new species is estimated that to of have undergone a event in their and 40% of species in the Brassicaceae family are to be (Al-Shehbaz, 1984; genome duplications, including as well as chromosome fusion and chromosome loss, have to the variation in the chromosome number in the Brassicaceae, with diploid of or chromosomes of Brassicaceae spp. chromosome Koch et and value and 2001). of genome size from the angiosperm value and 2001). from Koch et our and our length is to size, of in the at the for on of the species in the or in for of and and and and Arabidopsis both C. rubella and Shepard's purse the between these taxa is specific to Shepard's purse are in Arabidopsis and Arabidopsis pumila, Arabidopsis of Brassicaceae spp. chromosome Koch et and value and 2001). of genome size from the angiosperm value and 2001). from Koch et our and our length is to size, of in the at the for on of the species in the or in for of and and and and Arabidopsis both C. rubella and Shepard's purse the between these taxa is specific to Shepard's purse are in Arabidopsis and Arabidopsis pumila, Arabidopsis The prevalence of Brassicaceae with of the events that some of these species, opportunities to characterize how genomes change after duplication or fusion The of events within several Brassicaceae genera have been and the of events has been particularly useful for of rapid genomic changes after et genomic changes in Brassica spp. by from the spp. genomic changes in the were by of between and the natural et a similar with suecica, an from a of the Arabidopsis and Arabidopsis genomes. demonstrated and rapid gene at some suggesting a for regulation in species et al., 2000). studies with have been useful for an in which genes are in In suecica, the Arabidopsis genes are However, studies of genes can also and that the of the in can several et al., 1998). Further studies genomes of and ancient will provide insight into how genomes change over evolutionary time, molecular mechanisms that contribute to successful of genomes within and types of are and in the Expanding to additional Arabidopsis relatives will also allow for the of to characterize molecular events to speciation. et (2000) have to with of to or Arabidopsis different mechanisms species has evolved for reproductive Comparative genomic of closely related species are to understanding the portions of plant genomes. sequences, in change a can regions of conserved function exhibit sequence similarity among the characterized in and region centromere organization, and the formation of occur over evolutionary time even between Arabidopsis ecotypes (AGI, 2000). within these regions are often by the amplification of repetitive DNA such changes are particularly in regions of For example, satellite, and arrays and through gene or and centromeres are by Although the alterations in these regions are in Arabidopsis, they are likely more significant in larger genomes that are with of transposable elements and The accumulation of DNA elements contributes significantly to genome expansion in many plant In which for to of the genome, genomic change and through into intergenic regions et al., 1996). In elements are surprisingly in Arabidopsis, only 10% of the genome (AGI, 2000). analysis of other members of the Brassicaceae will reveal whether like Arabidopsis, have significant amplification of pressures likely contribute to variation in providing opportunities to mechanisms that or and amplification of In addition to and repetitive regions, genes with likely a their analysis from comparisons of relatives. In many genes are species specific and exhibit rapid sequence (for review, see and 2002). In Arabidopsis, a of genes that These genes are diverged from their in B. oleracea, which a of five genes. Although the general of the predicted remain the divergence at the DNA sequence level et al., 2001). Such divergence for analysis of more closely related species. Similarly, genes in and rapid changes both within and between Brassicaceae species (for review, et al., 2001). genes are often found in that vary in gene content, even between Arabidopsis ecotypes et al., 1999). only in Arabidopsis or its relatives are for understanding the changes that the unique and genes within members of the Brassicaceae will provide insight into species the mechanisms that contribute to genome requires analysis of multiple species separated by different evolutionary time For example, comparisons of genome between of different can identify conserved genes, but are useful when regions or genes for a comparisons of relatives elucidate sequences, but are in of boundaries to limited variation in the number of selected for comparative genomic studies on available will be with new species at some the value of that will the large of a of Brassicaceae species from Koch et with from the of indicate estimated of divergence at different in of studies the crop Brassica spp. in the indicated by the The within the Brassicaceae family have also been yielding a of species that share within the last approximately MY et al., 2001). of the between Brassicaceae species several separated from Arabidopsis over a range of evolutionary time Species in the genera and Capsella of which were in the genus are estimated to have diverged from Arabidopsis to 14 MY et al., see the divergence time that and S. and S. into a last with Arabidopsis approximately 16 to 20 MY other studies have crop Brassica spp. in this et al., the Brassicaceae, is separated from Arabidopsis, with a estimated to have occurred approximately MY These it possible to candidate species with an distribution for comparative genomic the value of candidate species for genomic on be with of and other physical traits. such as plant size, time, size for and are for In we have data on some of these for a of Brassicaceae species. In some species useful for genetic but of value to researchers in For example, C. requires on the order of to and has a that is not for assays that require multiple on the other it an for comparisons to Arabidopsis genes required for and C. mechanisms that its in the and spp. containing only two of these it difficult for researchers to the large of required for genetic Furthermore, species such as Arabidopsis are extremely to in but many Brassicaceae are or to Nevertheless, comparative genomic studies of these species have provided on the of For example, identification of the chromosomal region that contains the it is similar to an Arabidopsis region containing of genes et al., 2001). Thus, is likely an and Arabidopsis a species through an accumulation of in the genes required for the resources required for large-scale genome are both and is required significantly in resources for new species. the are sequence or genetic physical genomic useful for large or small-scale and and genome sequence sequencing is a and cost-effective approach to the coding content of a genome, although it is toward and of genome maps that on conserved markers are useful for genome however, this approach is relatively in and relies on and cannot identify small-scale duplications, or the degree of nucleotide sequencing of an entire genome can provide useful for comparative genomics. However, large stretches of sequence requires a Because comparative approaches are most with tracts of sequence on the order of of an and approach is to sequence individual from genomic made with chromosome provide several these can to 200 and are repetitive or other that in With genomic of many Brassicaceae species be with as few as this scale, can be on identification of a set of genes, or chromosomal regions. Moreover, when relies on distributed sites that are with they can be used to the genomic of sequence including and than those available for Arabidopsis, Brassicaceae resources have primarily been from in theBrassica The Brassica genome has for B. oleracea and B. napus and This is the to a physical that will be on the Arabidopsis genome and genome from B. oleracea have been in primarily by at Genome and The for Some have genomic investigations in other Brassicaceae species, and genomic resources In comparisons genomic from et al., and C. rubella (Acarkan et al., have been and additional have been from C. and available from will benefit from a comparative genomic that available resources to represent the breadth of the Brassicaceae family. The of the Arabidopsis genome sequence, of large-scale and recent in of it an appropriate time to additional In as a number of researchers are with a on as to species. is in the of the plant community as a whole to resources that are and of resources is for example, can be distributed through such as the and the or the Arabidopsis (Arabidopsis The of resources may require the of although the and of such to be for distribution are also that currently Brassicaceae species the Arabidopsis and the Arabidopsis is useful to however, that some into these have not been by in some in of species or the of Because the in this family are under a central that contains is The for is for many species, including the Brassicaceae. that are under way the and the of Because are extremely useful and can be in a time our is currently additional species for among the candidate species are suecica, C. S. and We that genomic resources from these species will provide a useful for the extent of genome across the entire family. We Johnston, and members of the for