Given the relatively short period since it was first suggested as a model species, the genomic resources comprising the Brachypodium Tool Box have been very rapidly established. Most recently, a very high quality draft genome sequence was published (IBI, 2010). The final genome assembly was remarkably complete, with 99.6% of all the sequences incorporated into the final assembly and only 0.4% of the sequence predicted to be missing based on paired end information. Similarly, the initial annotation was of high quality and was facilitated by a large number of ESTs (Vogel et al., 2006). Thus, the genome sequence is now in place to serve as a foundation for a myriad of applications. For example, a combination expression and tiling Affymetrix microarray has been created for Brachypodium (T. Mockler, pers. comm.). Other genomic resources include BAC libraries (Hasterok et al., 2006; Huo et al., 2008), BAC end sequences and a physical map based on contigs derived from a BAC library using the SNaPshot high-information-content-fingerprinting (HICF) method (Huo et al., 2009). However, other challenges, which are considered later in this text, must be overcome for the Brachypodium Tool Box to achieve its full potential.
1. Comparative genomics: Brachypodium as a Pooid reference genome?
The value of Brachypodium was first seen in its ability to act as a bridge species (Jenkins et al., 2005) to aid the cloning of genes in temperate cereals with extremely large genomes. A bridge species is necessary because larger genomes have many DNA repeats that effectively isolate genes or ‘gene islands’ in map-based cloning approaches (Devos, 2010). By contrast, small genome grasses have gene-rich euchromatic regions with one gene per 5–15 kb, on average. Brachypodium and rice have similar proportions of retrotransposon sequences (21.4% and 26%, respectively), which is much less than sorghum (54%) or bread wheat (over 80%) (IBI, 2010). Even before the genome sequence was available, the chromosomal pairing locus Ph1 was accurately mapped to a cluster of kinase genes within a heterochromatic segment of wheat chromosome 5B using markers derived from a much smaller orthologous region in B. sylvaticum (Griffiths et al., 2006). Interestingly, although the rice genome sequence was available, its sequence was too divergent from wheat in this region to provide markers and it could not be used to fine map the Ph1 locus. In other examples, markers from Brachypodium were used to identify the Lr34/Yr18 wheat rust resistance gene (Spielmeyer et al., 2007) and the barley Ppd-H1 photoperiod response gene (Turner et al., 2005).
The availability of the Bd21 genomic sequence, EST collections and resequencing data from other accessions (soon to be available) means that more genomic sequence data and markers are now coming on line to facilitate genomic comparisons. Simple sequence repeat (SSR) microsatellite markers are widely used as anchor markers in genetic mapping, and in marker-assisted breeding. A large number of SSR markers have been derived from B. distachyon (Azhaguvel et al., 2009; Vogel et al., 2009; Garvin et al., 2010) offering the possibility of rapidly identifying genetic loci associated with a trait and translating this to other cereals. Garvin et al. (2010) populated the Brachypodium genome with a total of 139 SSR markers derived from ESTs, BAC end sequences (BES) and conserved orthologous sequences (COS) from a range of grass species. In macrosyntenic comparisons, 13 out of 20 Brachypodium linkage groups have equivalents in the rice genome sequence. Azhaguvel et al. (2009) developed and used 160 EST- and 21 derived genomic microsatellite markers to evaluate genetic diversity among the Brachypodium accessions and to relate the observed genetic races to the feeding preferences of the wheat greenbug (Schizaphis graminum Rondani) and the Russian wheat aphid (RWA) (Diuraphis noxia). Further, phylogenetic analysis suggested that Brachypodium is closer to Aegilops tauschii (the D genome donor of common wheat) than to rice. When EST contigs were compared, Brachypodium exhibited orthology with wheat EST contigs from all 21 of the wheat chromosomes (Kumar et al., 2009).
Some astonishing examples of orthology between Brachypodium and the temperate cereals at a single gene level have been reported. In Triticeae genomes, a tandem duplication occurred in a globulin gene leading to the selection of a high molecular weight glutenin gene – an essential trait in bread wheat. This globulin duplication was present in Brachypodium but not in any tropical grass genome (Gu et al., 2010). The Eps-A (m) 1 locus, which controls morphological changes, such as spiking in Brachypodium and in one wheat genome progenitor (A. tauschii), is very similar to orthologues at this locus, including, crucially, Mot1 and FtsH4, which are tightly linked to the earliness per se phenotype (Faricelli et al., 2010). Nevertheless, some studies have suggested other genomic relationships. Yu et al., (2009) associated the Hessian fly resistance gene H26 (mapped to the wheat 3DL distal region 3DL3-0.81-1) with a Brachypodium supercontig. Encouragingly, 14 of the 15 ESTs were collinear between the distal region of wheat 3DL and Brachypodium supercontig 13. However, of 46 STS (sequence tagged site) primer markers derived from this supercontig, only one could be mapped to wheat chromosome 3D. The apparent lack of conservation was such that the authors advised caution when using the Brachypodium genomic sequence for molecular mapping and gene cloning in wheat.
Reconciliation of these seemingly contradictory data came about with the publication of the Brachypodium genome sequence. On a global level, Brachypodium and rice exhibit a high degree of conservation of gene order such that entire rice chromosomes or chromosome arms can be mapped to their Brachypodium counterparts. By contrast, the alignment of diploid wheat ancestor (A. tauschii) and barley genetic maps to the Brachypodium genome is more fragmentary, although there are still large segments of conserved alignment. This indicates that there were a large number of genomic rearrangements in the lineage containing wheat and barley after the divergence from the Brachypodium lineage. Thus, instead of simply using Brachypodium as a roadmap for the wheat and barley genomes through simple pairwise comparisons, it will be more effective to use multiple comparisons of gene order in sequenced grass genomes. However, because Brachypodium shares a higher degree of nucleotide sequence conservation with the temperate grains, markers created using Brachypodium sequence have a much higher conversion rate.
When colinearity of genes within seven large gene families was examined four exhibited a high degree of conservation in gene order between cereal genomes, but this was not the case for the nucleotide binding site (NBS)-leucine-rich repeat (LRR) and F-box gene families (IBI, 2010). This could indicate rapid diversification owing to strong natural selection driven by pathogen pressure in the case of NBS-LRR and by the regulation of both developmental and stress responsive traits in the case of F-boxes (Meyers et al., 2003; Xu et al., 2009). Thus, the degree of synteny between Brachypodium and wheat in any particular region will vary owing to both the historical rearrangements after the divergence of the lineage and evolutionary history of the gene, which may be accelerated in specific gene families.
Interestingly, when only gene sequence is compared, the overwhelming majority of gene families are highly conserved between rice, sorghum, Brachypodium, wheat and barley; only 265 out of 16 215 gene families were specific to the Pooideae (Brachypodium, wheat and barley). Thus, there are considerable genomic similarities between the Ehrhartoideae (rice), Panicoideae (Sorghum) and Pooideae (B. distachyon, T. aestivum and Hordeum vulgare), undoubtedly reflecting their relatively recent evolutionary divergence. Thus, in addition to serving as a structural model for the large genomes of the temperate grains, the overall similarity at the gene level indicates that, for the majority of traits, Brachypodium can serve as a functional genomic model for all the grasses. As a corollary to this, depending on the trait, Brachypodium might not be the best source of comparative genomic information for the Triticeae and a comparative approach using all available genomes is advised. This may be more important for traits that have come under selection pressure during rapid evolutionary diversification, and are therefore specific to the Triticeae, such as traits conferred by the NBS-LRR and F-box’s genes (IBI, 2010).
2. Transformation and reverse genetic tools within the Brachypodium Tool Box
Efficient transformation is a prerequisite for a modern model organism and we are fortunate that Brachypodium has been very amenable in this regard. Biolistic bombardment-based transformation of a polyploid line was first reported in 2001 (Draper et al., 2001). A more efficient method that worked on a diploid line was later reported in 2005 (Christiansen et al., 2005). However, Agrobacterium is the preferred transformation method, where it is important to have simple, low copy integration events and methods with efficiencies up to 30% have been published (Vain et al., 2008; Vogel & Hill, 2008). Today, the average efficiency in a production setting where the emphasis is on minimizing the labour per transgenic line rather than maximizing the efficiency of each transformation is c. 45% (up-to-date methods are available at http://brachypodium.pw.usda.gov/). With these improvements Brachypodium is arguably one of, if not the, most easily transformed grasses.
The efficiency of Brachypodium transformation makes feasible the creation of collections of sequence indexed T-DNA mutants. An excellent example of the power of a T-DNA population to reveal gene function is the SALK T-DNA tagged Arabidopsis resource (Ecker, 2002). Such a resource is a crucial component of the Brachypodium Tool Box. Two groups have established projects to create Brachypodium T-DNA mutants. Researchers at the John Innes Centre have reported the generation of a collection of 4500 T-DNA lines. Analysis of 741 accessions showed that 660 T-DNA loci could be assigned to a unique location in the Brachypodium genome sequence (Thole et al., 2010). Of these, c. 60% could be associated with ESTs. The T-DNA lines generated by the BrachyTAG programme are available as a community resource and have been distributed internationally since 2008 via the web site (http://www.brachytag.org/). Similarly, researchers at the USDA-ARS Western Regional Research Center are generating thousands of tagged lines. At the time of writing > 10 000 have been created, and over 6000 flanking sequence tags (FSTs), have been assigned to a unique location in the genome. Rather than concentrating only on T-DNA tagging, this effort is also using gene-trapping vectors to identify promoters with useful expression patterns, and activation tagging vectors that contain transcriptional enhancers to ‘activation tag’ nearby genes. Details and ordering instructions can be found at http://brachypodium.pw.usda.gov/TDNA.
In addition to insertional mutants, a method to mutagenize seeds with ethyl methanesulphonate (EMS) has been optimized (http://brachypodium.pw.usda.gov), and irradiation with fast neutrons has also been used to create mutant collections (D. Laudencia-Chingcuanco and M. Byrne, pers. comm.). These EMS mutants are very useful for forward genetic screens because of the large number of mutants per plant. In addition, EMS mutants can also be used to identify mutations in particular genes for reverse genetic studies by a TILLING (Targeted Induced Local Lesions in Genomes) approach (McCallum et al., 2000) which employs a mismatch specific endonuclease to identify mutated PCR amplicons. The authors are aware of the creation of one TILLING population (http://www-ijpb.versailles.inra.fr/en/institut/actualite.htm) which currently consists of c. 6000 individuals.
Although these emerging mutagenic approaches are impressive, it will take some time before sufficient genome coverage is achieved for most Brachypodium genes to have a corresponding mutant. Given this, the development of a targeted gene disruption strategy based on virus-induced gene silencing (VIGS) represents a significant advance as it represents an immediately applicable strategy through which the expression of targeted genes can be disrupted. Barley stripe mosaic virus (BSMV) is a single-stranded tripartite RNA virus where infection of the host can occur following simple rubbing of leaves with naked (nonenveloped or capsidated) genome. Derivatives of this virus, including fragments of a targeted gene, have been used to suppress gene expression in barley and wheat (Holzberg et al., 2002; Scofield et al., 2005). Recently, two studies have shown the efficacy of BSMV-induced VIGS in Brachypodium. Demircan & Akkaya (2010) silenced a phytoene desaturase gene and Pacak et al. (2010) suppressed the expression of genes IPS1, PHR1 and PHO2 known to participate in phosphate (Pi) uptake and reallocation. Following these successes, we expect VIGS to be widely employed in Brachypodium research.
3. Cytogenetic tools for genome analyses
An important outcome of Brachypodium research has been the rapid development of cytogenetic tools. Routine FISH of BAC clones is able to identify and delimit specific chromosomes, chromosome arms, and particular chromosome regions (Hasterok et al., 2006; Jenkins & Hasterok, 2007). This has provided unprecedented insights into the genomic relationships within the Brachypodium genus, and should help to reveal evolutionary relationships between members of the Pooideae and more distant grass species (Wolny et al., 2010). From a more practical viewpoint, the use of ordered, labelled BAC clones representing the definitive Brachypodium nuclear sequence has allowed the construction of supercontigs covering large regions of the B. distachyon chromosomes, which has proved invaluable for the validation of linkage group assembly (Febrer et al., 2010; IBI, 2010). In addition, these resources enabled robust chromosome ‘painting’, which is now at our disposal for investigating with greater resolution the structural relationships between related genomes and the early association of chromosomes during meiosis (Fig. 2).
Figure 2. Pachytene in Brachypodium (2n = 10) with physically mapped alternating clusters of red and green bacterial artificial chromosome (BAC) probes hybridizing to the short arm of chromosome Bd1. Bar, 5 μm (courtesy of Dr Dominika Idziak, Department of Plant Anatomy and Cytology, University of Silesia, Katowice, Poland).
Download figure to PowerPoint
As an example of the application of such tools, the Jenkins and Hasterok teams are actively utilizing these cytogenetic resources to understand and manipulate meiosis and recombination in grasses. A major obstacle to plant breeding programmes is not so much sexual and genomic incompatibility, as the retention over many backcross generations of undesirable wild germplasm. Purging potential new cultivars of this material is costly and time consuming, and largely governs the speed at which new cultivars are released. These negative effects of linkage drag are inversely proportional to the extent of genetic recombination in successive backcross generations.
Recombination is assayed traditionally by direct cytological scoring of chiasmata at metaphase I of meiosis, which has shown that the chromosomes of temperate, large-genome cereals and grasses, such as wheat, barley, rye and ryegrass, each have regularly two or more chiasmata. However, cytological inspection shows that the vast majority of these are confined to distal regions, meaning that interstitial and proximal chromosome segments seldom engage in recombination events. This uneven distribution of crossovers can also be inferred from measurements of recombination frequencies between genetic markers in mapping populations, which are then correlated to physical chromosome positions using deletion and introgression mapping or recombination nodules. Such studies in, for example, wheat (Erayman et al., 2004; Saintenac et al., 2009) have revealed that there is a common gradient of increasing recombination from proximal to distal regions of chromosomes. Because it is now known that the interstitial and proximal regions of cereal and grass chromosomes harbour a sizeable proportion of the genes of the genome (albeit at a lower density), the inevitable conclusion is that a large part of the genome of these species is not regularly involved in recombination events, effectively consigning many genes to recombination backwaters. This limits the potential of genetic variation in these crops, it prolongs linkage drag in introgression programmes, and it confounds the scope of map-based cloning approaches. Indeed, it has been acknowledged that ‘> 30% of wheat genes are in recombination-poor regions and thus are inaccessible to map-based cloning’ (Erayman et al., 2004). Clearly, it would be desirable to crack open these recombination coldspots in order to release new genetic variation which could be exploited in advanced breeding programmes. However, even if sites of recombination were shifted, map-based cloning strategies would still be constrained by the inordinately large genome sizes of these species.
Preliminary studies of meiosis in a diploid and an allotetraploid accession of Brachypodium (Jenkins et al., 2005) have shown that chiasmata are not distally localized as in the other members of the Poaceae mentioned above. Thus, if there is variation in recombination among Brachypodium germplasm there is a real prospect of establishing the genetic basis of chiasma location and frequency. Variation in genomic recombination has been noted in Arabidopsis (Sanchez-Moran et al., 2002), rye (Rees, 1961), Lolium and Festuca (Jones & Rees, 1966; Rees & Dale, 1974), and barley (Gale & Rees, 1970).
4. Looking into the Brachypodium Tool Box: bioinformatic tools
In the multinational effort required to develop Brachypodium as a model, parallel development of appropriate bioinformatic tools allows the collection, curation and interrogation of genomic and post-genomic data. Readers unfamiliar with Brachypodium are recommended to use the bioinformatic tools available through the http://www.brachypodium.org website. From this website, it is possible to access the Brachypodium genome browser Brachybase (http://www.brachybase.org), where the 8X Assembly genome can be viewed, downloaded it in its entirety, and related to EST databases and Affymetrix (BdArray) array probes. A blast tool (http://www.brachybase.org/blast/) allows nucleotide and amino acid sequence comparisons. For excellent cross-species genomic comparisons and as a repository of much information on Brachypodium, the Gramene website (http://www.gramene.org/Brachypodium_distachyon/) is also a good place to start.
Other databases (http://mips.helmholtz-muenchen.de/plant/brachypodium/) allow whole genome, protein sequence/structures and motifs to be investigated and promoter regions to be accessed. http://www.modelcrop.org/ has many of the functions of these other sites but also allows the Brachypodium physical map to be displayed and easily compared with the rice and sorghum genomes. With http://www.phytozome.net it is possible to search for orthologous or homologous genes among all sequenced plant genomes based on name or sequence. Matches may be extracted and alignments compared by progressive alignment algorithms (dynamic programming) while relationships can be displayed using phylogenetic approaches or multivariate principal component analysis.
One result of all of these bioinformatic data and tools has been the development of an Affymetrix array based on the Bd21 genome sequence and EST databases (S. E. Fox et al., unpublished data). The derived oligonucleotides were used to generate the array representing unique single copy sequences, with mean probe spacing of 42 bases with 95% of probe pairs < 126 nucleotides apart enabling studies of gene-specific expression. S. E. Fox et al. (unpublished) are using this array resource to characterize Bd21 expression during development, diurnal and circadian cycling as well as abiotic or biotic stress to represent a ‘Bd21 expression atlas’. It is expected that these data will be made available via a web portal analogous to the Arabidopsis genomic site (e.g. TAIR Microarray Experiments).
5. Diversity within the Brachypodium Tool Box: germplasm collections
If Brachypodium is to act as functional genomic model for the Pooideae and other grasses, germplasm collections must contain accessions with economically relevant traits and encompass sufficient variation for mapping projects to succeed. Thus, an essential part of the Brachypodium Tool Kit must be well-characterized germplasm collections (Filiz et al., 2009). Sequence data from comprehensive germplasm collections will yield markers linked to specific loci and help identify gene function, for example, improved stress tolerance or yields, enabling the isolation of corresponding alleles in grasses and cereals. Brachypodium genes could also be used directly to generate GM cereal derivatives. Beyond cereal/forage grass improvement strategies, a well-curated Brachypodium germplasm collection will also allow macro and micro level evolutionary trends to be modelled and related to past or present selection pressures. It should also be possible to predict losses in genetic diversity by genetics and ecological niche modelling, due to environmental change or human activities.
Until recently, a salient feature of Brachypodium research has been the relative paucity of available germplasm. Much early work focused on only seven inbred lines (Bd1-1, Bd2-3, Bd3-1, Bd18-1, Bd21, Bd21-3 and Bd29) developed from USDA collections (http://www.ars-grin.gov/npgs) and another small collection of ABR accession (ABR1 through ABR7) originating mainly from Spain (Stace and Catalán collection, Leicester, UK). Now, however, there is a large collection of Turkish germplasm available: 195 diploid lines collected from 53 locations in Turkey (Filiz et al., 2009; Vogel et al., 2009). Within this large collection considerable variation in flowering time, seed size, and plant architecture was noted. Sixty-two wild accessions (first generation) were analysed with 43 SSR markers and the vast majority of individuals were homozygous, despite the presence of multiple alleles in the local population. Under laboratory conditions, intimately grown lines failed to outcross. This reflects near-cleistogamy in the diploid lines which, while facilitating genotype conservation, poses a considerable barrier to full exploitation of the Brachypodium Tool Box as this makes crosses and derivation of mapping families quite difficult. However, following on from an early demonstration of successful crossing of two Brachypodium accessions (Routledge et al., 2004), recombinant inbred lines have been derived (Garvin et al., 2008) and simple, very efficient step-by-step crossing protocols are now available (see http://www.ars.usda.gov/SP2UserFiles/person/1931/BrachypodiumCrossing.pdf and http://brachypodium.pw.usda.gov). These relatively simple protocols represent an essential component of the Brachypodium Tool Box and we expect the populations of recombinant inbred lines developed by Garvin et al. (2008) will soon be added to by those from other groups.
We are now greatly expanding the germplasm collection with numerous collections from Northern Spain (Fig. 3a). In this collection, we have concentrated on developing inbred lines from various environments. Thus, we have sampled individual seeds from 46 localities, from high altitudes in the foothills of the Pyrenees, to lowland areas around the flood plain of the Ebro river basin, coastal areas around Catalunya and a Balearic Island population collected from Ibiza (Table 1). Between 10 and 30 individuals were sampled from each site, with each sample separated from the others by at least 1 m to avoid sampling close relatives. This collection is being developed specifically to search for intrapopulation and interpopulation variation and to relate this genetic diversity to potential adaptation to environmental selection pressures. We also are seeking to determine the relationships between sympatric 2n = 10 and 2n = 30 Brachypodium populations, and we have at least one site where the two ploidy-level species are sympatric (Table 1, Bierge, Huesca, Spain).
Figure 3. The development of an ecotypic Brachypodium germplasm collection. (a) Sites of Brachypodium sampling in north-east Spain: ‘Lowland’, light blue circles; ‘Highland’, orange circles; ‘Coastal’, green circles; ‘Island’, dark blue circles. Insets are example images of a ‘lowland’ (Alfranca, Zaragoza, Spain), ‘highland’ (Puerto del Perdón, Navarra, Spain), ‘coastal’ (Port Lligat, Cadaqués, Girona, Spain) and ‘Island’ (Port des Torrents, Ibiza, Spain) region. (b) Preliminary assessment of genotypic variation of this new germplasm collection has been undertaken using 14 bacterial artificial chromosomes end sequences (BES) -derived tetranucleotide- trinucleotide- and dinucleotide-based repeats simple sequence repeat (SSR) markers described at the USDA Brachypodium web site (http://brachypodium.pw.usda.gov/SSR/) (loci DB069I22, DB078E03, DB088E23, DB080E01, DH044F02, DB078M18, DH038F15, DB060E14, DB092H16, DB082M05, DB009L08, DB071F17, DH048N18, DH038H13, DB088J19). Descriptions of the germplasm together with either confirmed or tentative estimations of ploidy are given in Table 1. Neighbour-joining phenogram from 96 shared SSR alleles based on Nei & Tajima (1983) Da distance between 44 Spanish and Turkish–Iraqi Brachypodium wild-type germplasm lines. Bootstrap values on branches are based on 10 000 replications. Symbols: blue circle, Spanish highlands; red circle, Spanish lowlands; dark green square, Turkish–Iraqi (the Bd21 whole sequenced-genome sample is encircled); bright green square, Tekirdag 10 (Turkey); purple triangle, Balearic coastal.
Download figure to PowerPoint
Table 1. Inbred Brachypodium distachyon lines
|Accession name||Site||Latitude||Longitude||Alt||Chromosome number (2n)||Genome size 2C (pg)||Standard deviation|
|Bie2||Bierge, Sierra de Guara, Huesca, Spain||42.17305 N||0.09075 W||730||10||0.681||0.009|
|Bie3|| || || || ||10||0.662||0.008|
|Bie10|| || || || ||30||1.314||0.016|
|Bie13|| || || || ||30||1.316||0.02|
|Bou2||Cala De Bou, San Antonio, Ibiza, Spain||38.96774 N||1.26819 E||3||30||1.319||0.028|
|Bou13|| || || || ||30||1.293||0.018|
|Bou18|| || || || ||30||1.336||0.015|
|Cel4||Santa Cilia, Sierra de Guara, Huesca, Spain||42.23836 N||0.16599 W||791||10||0.693||0.009|
|Foz1||Foz de Lumbier, Navarra, Spain||42.63651 N||1.30484 W||434||10||0.694|| |
|Gal2||Murillo de Gallego, Zaragoza, Spain||42.34552 N||0.74299 W||515||10||0.670||0.016|
|Men5||Menarguens, Lleida, Spain||41.72018 N||0.72452 E||234||30||1.287||0.011|
|Men7|| || || || ||30||1.278||0.009|
|Men14|| || || || ||30||1.275||0.013|
|Men19|| || || || ||30||NM||NM|
|Men30|| || || || ||30||NM||NM|
|Mig3||San Miguel de Foces, Ibieca, Huesca, Spain||42.14799 N||0.19497 W||572||10||0.671||0.008|
|Mon3||Puerto de Pallaruelo, Castejón de Monegros, Zaragoza, Spain||41.65132 N||0.21042 W||515||10||0.687||0.019|
|Mur3||Castillo de Mur, Lleida, Spain||42.09763 N||0.87750 E||487||10||0.682||0.013|
|Rei7||Puente de la Reina, Huesca, Spain||42.56319 N||0.78688 W|| ||10||0.669||0.011|
|Sar2||Laguna de Sariñena, Huesca, Spain||41.78619 N||0.18278 W||292||10||0.668||0.016|
|Tor2||Port des Torrents, San Antonio, Ibiza, Spain||38.96740 N||1.2818 E||6||30||1.291||0.02|
|Tor7|| || || || ||30||1.293||0.011|
|Tor8|| || || || ||30||1.307||0.02|
|Uni2||Escuela Politécnica Superior, Huesca, Spain||42.117408 N||0.445046 W||480||10||0.675||0.009|
|Yas3||Yaso, Sierra de Guara, Huesca, Spain||42.20237 N||0.12236 W||731||10||0.662||0.008|
Preliminary assessment of genotypic variation in a total of 38 wild individual samples, representative of most of the Spanish Brachypodium germplasm localities (Table 1, Fig. 3a), involved comparison with five individual samples representing different Turkish Brachypodium lines (Vogel et al., 2009) and Bd21. The number of alleles surveyed per locus ranged from 2 to 14, with an average of 6.85 across the 44 samples studied. Six out of 14 screened loci showed total or predominant levels of homozygosity across the 44 germplasm lines assessed (with observed heterozygosity values ranging from 0 to 0.159). Relationships among individual samples were evaluated through Nei et al.’s (1983) Da distance-based neighbour-joining (NJ) reconstruction. The unrooted NJ phenogram (Fig. 3b) showed: a large Spanish clade formed by samples from medium to relatively high altitude Pyrenean and Prepyrenean localities (Fig. 3a‘highland’) and their close low altitude Ebro river valley localities (Fig. 3a‘lowland’); a Turkish–Iraqi clade formed by most of the samples surveyed from different altitudinal localities of Turkey, including the sequenced line Bd21 and a line developed from the same location, Bd21-3; and a ‘Miscellaneous’ clade formed by the coastal Balearic isles (Ibiza: Fig. 3a‘island’) samples plus one sample from low altitude Tekirdag (Tek10, Turkey) and four samples from both highland and lowland Iberian localities.
Although only a small number of Turkish–Iraqi germplasm representatives were tested, a greater genetic diversity was observed in wild Brachypodium individuals from the western Mediterranean region compared with those from the eastern. Variation in DNA sequences within other temperate Mediterranean grasses (e.g. Hordeum marinum, Jakob et al. 2007), has similarly suggested greater diversity in the lines from the western part of the region, which may have contained a larger number of glacial refugia during the last ice age from which a larger number of ‘bottlenecked’ populations would have been derived. More exhaustive comparative phylogeographical studies of Brachypodium populations across its native Mediterranean distribution area are currently underway by Catalán’s group to test this hypothesis.