Development of diversity array technology (DArT) markers for assessment of population structure and diversity in Aegilops tauschii.

Aegilops tauschii Coss. is the D-genome donor to hexaploid bread wheat (Triticum aestivum) and is the most promising wild species as a genetic resource for wheat breeding. To study the population structure and diversity of 81 Ae. tauschii accessions collected from various regions of its geographical distribution, the genomic representation of these lines were used to develop a diversity array technology (DArT) marker array. This Ae. tauschii array and a previously developed DArT wheat array were used to scan the genomes of the 81 accessions. Out of 7500 markers (5500 wheat and 2000 Ae. tauschii), 4449 were polymorphic (3776 wheat and 673 Ae. tauschii). Phylogenetic and population structure studies revealed that the accessions could be divided into three groups. The two Ae. tauschii subspecies could also be separately clustered, suggesting that the current taxonomy might be valid. DArT markers are effective to detect very small polymorphisms. The information obtained about Ae. tauschii in the current study could be useful for wheat breeding. In addition, the new DArT array from this Ae. tauschii population is expected to be an effective tool for hexaploid wheat studies.

Introduction

Aegilops tauschii Coss. is a wild, diploid (2n = 2x = 14, DD), self-pollinated species that is considered to be the D-genome donor to hexaploid wheat, also known as bread wheat or common wheat (Triticum aestivum L.; 2n = 6x = 42, AABBDD) (Kihara 1944). Aegilops tauschii is adapted to a variety of environments such as desert margins, steppe regions, stony hills, wastelands, roadsides, sandy shores and even humid temperate forests (van Slageren 1994). It is also found in the edges of wheat fields in eastern Turkey, Iraq, Iran, Pakistan, India (Kashmir), China (the Himalaya), Afghanistan, most of central Asia, Transcaucasia and the Caucasus region (Feldman 2001). Because it carries one of the genomes of bread wheat, Ae. tauschii is regarded as the most suitable species for wheat improvement among the wild species in the tribe Triticeae. The diversity of the D genome of this species is much larger than that of bread wheat and includes many useful genes for resistance to biotic and abiotic stresses and for seed storage proteins (Assefa and Fehrmann 2000, Colmer , Cox , Lubbers , Naghavi and Mardi 2010, Pestsova , Reif , Sohail ). Thus, Ae. tauschii needs to be studied in detail in order to use it most effectively as a germplasm source for the improvement of bread wheat.

Based on morphology, taxonomists have divided Ae. tauschii into two subspecies: ssp. tauschii and ssp. strangulata (Eig) Tzvel. (Eig 1929, Hammer 1980). Ssp. strangulata has quadrate spikelets whereas ssp. tauschii has elongated spikelets (Kimber and Feldman 1987, Matsuoka ). Ssp. tauschii is further divided into three morphological varieties: anathera, meyeri and typica, whereas ssp. strangulata is monotypic. Dvorak pointed out that some Ae. tauschii accessions have intermediate forms and suggested that these might have a hybrid origin. The geographical distribution is also different between the two subspecies. Most ssp. strangulata is limited to the Caucasus and southeastern Caspian costal region, while ssp. tauschii is distributed throughout central and western Asia (Dudnikov and Kawahara 2006, Eig 1929). Ssp. strangulata is considered to be more closely related to bread wheat compared to ssp. tauschii (Dvorak , Nishikawa , Pestsova ).

The diversity of Ae. tauschii has been studied using molecular tools such as chloroplast DNA variation (Matsuoka , 2009, Takumi ), AFLP (Mizuno ), SSR (Naghavi and Mardi 2010), isozymes (Dudnikov and Kawahara 2006) and random amplified polymorphic DNA (RAPD) markers (Okuno ). Here, we used diversity array technology (DArT) markers to study the population structure and diversity of Ae. tauschii. DArT is a sequence-independent system, based on microarray hybridization that can be used to carry out a whole-genome scan. The method is based on the use of “genomic representations”, which are DNA samples produced by using a specific combination of restriction enzymes and PCR primers. The output result is (0, 1); that is, it indicates the presence or absence of each DNA fragment contained in a genomic representation within the genome of the material being tested. DArT markers are biallelic dominant markers that provide a cost-effective, time-saving method of genome-wide genotyping (Jaccoud , Kilian ). These markers have been successfully used for genotyping, diversity studies, and genetic mapping in many crop species (Jing ).

The molecular information provided by the present DArT analyses will be useful for breeding programs.

Materials and Methods

Plant materials and DNA isolation

Eighty-one accessions of Ae. tauschii collected from various regions of its geographical distribution were used in this study. Thirteen of the accessions were ssp. strangulata, 62 were ssp. tauschii, and six were of unknown subspecies (Table 1). The AE accessions were collected by the Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK), Germany; the AT accessions by the Faculty of Agriculture, Okayama University, Japan; the CGN accessions by the Instituut voor Plantenveredeling, Landbouwhogeschool, Wageningen, the Netherlands; the IG accessions by the International Center for Agricultural Research in the Dry Areas (ICARDA), Syria; the KU accessions by the Germ-Plasm Institute, Faculty of Agriculture, Kyoto University, Japan; and the PI accessions by the United States Department of Agriculture (USDA). Fifty-five of these Ae. tauschii accessions have been used for producing synthetic wheat. These Ae. tauschii were kindly provided by Prof. Y. Matsuoka, Fukui Prefectural University, Japan. The accession 28H51 has been used by ICARDA for producing synthetic wheat derivatives. Accession Aus18913 has been used for mapping of chromosome arm 1DS of Ae. tauschii (Spielmeyer ) and to construct an Ae. tauschii BAC library (Moullet ). Total genomic DNA was extracted from 2- to 3-week-old leaves of the 81 accessions by using the CTAB method (Murray and Thompson 1980).

Table 1

Ae. tauschii accessions used in this study

Country (number of accessions)	Accessiona
Afghanistan (n = 6)	KU2039, PI476874, KU2063, KU2035, KU2022, KU2636A
Armenia (n = 7)	IG127015, KU2816, KU2810, KU2824, CGN10734, KU2809 A, IG126991
Azerbaijan (n = 2)	IG47203, KU2806
China (n = 6)	AT55, AT76, AT80, AT47, PI499262, PI508262
Dagestan (n = 2)	KU201, IG120866
Georgia (n = 5)	AE454, KU2826, KU2828, KU2829A, IG48042
India (Jammu and Kashmir) (n = 1)	IG48042
Iran (n = 37)	KU20-10M, KU2069, KU2074S, KU2075S, KU2076S, KU2078S, KU2079S, KU20-8, KU2080S, KU2088S, KU20-9S, KU2090S, KU2091S, KU2092S, KU2093S, KU2096, KU2097, KU2098, KU2100M, KU2103, KU2104, KU2105, KU2106, KU2109M, KU2111, KU2124, KU2126, KU2144, KU2155, KU2156, KU2158M, KU2159, KU2083, KU2101, KU2102, KU2108M, IG49095
Kazakhstan (n = 1)	AE1090
Kyrgyzstan (n = 1)	IG131606
Pakistan (n = 4)	IG46663, CGN10768, CGN10770, CGN10767
Syria (n = 2)	IG46623, 28H51
Tajikistan (n = 1)	IG48554
Turkey (n = 3)	KU2132, KU2136, PI486277
Turkmenistan (n = 2)	IG126387, IG126489
Unknown (n = 1)	Aus18913M

Accessions without superscripts are ssp. strangulata; accessions with superscripts are ssp. tauschii. A, ssp. tauschii var. anathera; M, ssp. tauschii var. meyeri; S, ssp. tauschii var. strangulata.

Preparation of DArT markers and genotyping

A total of 81 genotypes DNA were sent to Diversity Arrays Technology Pty. Ltd., Yarralumla, Australia, for array development and genotyping as described by Wenzl and Akbari . The Ae.tauschii DArT markers are referred to as (aePt-). Each of the original 81 accessions was genotyped with two arrays: Wheat DArT Array Version 3.0 and the Ae. tauschii array developed in this study. The genomic DNA of each accession was labeled and then hybridized to the DArT arrays. The presence or absence of each marker was determined on the basis of the signals from the labeling and image analysis. The DArT marker data were provided to in term of 1, 0 (present, absent) fashion, as described by Akbari .

Analysis of DArT data

To determine the population structure of Ae. tauschii, we applied the Bayesian method by using the model-based program Structure version 2.3.3 (Falush , Pritchard ) with (0, 1) data matrices. We used a burn-in length (number of cycle runs by the simulation before collecting data) of 104 cycles to minimize the effect of starting configuration, and a simulation run length (after the burn-in) of 106 cycles, and applied the admixture model option in the Structure program. We chose cluster values (K) ranging from 2 to 9; four independent runs for each value gave consistent results.

Genetic similarities between accessions were measured by DICE similarity coefficient based on the proportion of shared alleles. The phylogenetic tree was constructed by clustering accessions based on similarity matrix using the unweighted pair group method (UPGMA) with arithmetic average algorithm in the SAHN module. Bootstrap analysis was performed using 1,000 permutations in Winboot (Yap and Nelson 1996). Bootstrap values over 50 are considered significant and indicated on the phylogenetic tree.

To calculate the genetic similarities and genetic distances between the pairs of accessions, the (0, 1) data matrixes obtained for 4449 polymorphic DArT markers (details presented in the Results section) and 81 accessions were analyzed with the following formulae:

Where, S represents the similarity between the ith and jth accessions, N represents the number of common bands present in both the N and N accessions (i.e., the number of markers where 1's are present for both the N and N accessions), N is number of bands in the ith accession (number of markers with 1's in accession N), and N is number of bands in the jth accession (number of markers with 1's in accession N).

By using the program Excel (Microsoft Corporation), we calculated genetic similarity for all 3240 possible pairs [81 × (81–1)/2] of the 81 accessions using the (0, 1) data matrix consisting of 4449 rows (4449 DArT markers) and 81 columns (81 accessions). An 81 × 81 genetic similarities matrix was created in which the values of 1 on the main diagonal (representing the similarity of each accession to itself) were not considered. The values in the genetic similarities matrix were used to calculate average similarity within ssp. strangulata and within the three varieties of ssp. tauschii. The average similarities between ssp. strangulata and each of the varieties of ssp. tauschii were also calculated.

Results

Polymorphism of DArT markers in Ae. tauschii accessions

A total of 7500 DArT markers were tested. Of these, 5500 markers from wheat DArT Array Version 3.0 (wheat markers) were developed from hexaploid bread wheat and the other 2000 (Ae. tauschii markers) were developed from the 81 Ae. tauschii accessions by the DArT company (Table 2). Of these markers, 3776 of the wheat markers (68.6%) and 673 of the Ae. tauschii markers (33.7%) showed polymorphism among the 81 accessions of Ae. tauschii. The polymorphism information content (PIC) of the DArT markers ranged from 0.024 to 0.500 per marker, with an average of 0.259 (Table 2). Thirty-two markers had a very low PIC (0.024), which means that these markers showed little polymorphism. About 96% of the markers had a call rate of more than 90% and 48% of the markers had a call rate of 100%; 95% of the wheat markers had a call rate of 90% or more while almost all of the Ae. tauschii markers had a call rate of more than 87%. To verify the reproducibility of the genotyping, two of the accessions were analyzed in duplicate (i.e., two wells per accession). The results of both pairs were identical, except for a few missing data points.

Table 2

Quality parameters of the two different types of DArT markers used to analyze 81 Ae. tauschii accessions

	Marker type
Parameter	Wheat Array 3.0	Ae. tauschii	Overall
Total number of DArT markers	5500	2000	7500
Number of polymorphic markers	3776	673	4449
Polymorphism (%)a	68.6	33.7	59.3
Polymorphism information contentb	0.246	0.329	0.259
P (%)c	81.2	82.9	81.5
Reproducibility (%)d	99.9	99.8	99.9
Call rate (%)e	97.8	97.6	97.8

Number of polymorphic markers/total number of DArT markers tested.

Measure of polymorphism to describe the usefulness of a marker.

Reflects how well the two phases of the marker are separated.

On the basis of scoring for replicated samples.

The number of genotypes present and not missing for certain marker.

Phylogenetic tree based on DArT marker genotypes

Phylogenetic trees were constructed using the 3776 polymorphic wheat markers, the 673 polymorphic Ae. tauschii markers and the combined total of 4449 markers showing polymorphism. The structures of these three trees were very similar, therefore, the tree made by using all 4449 markers is shown in Fig. 1.

Fig. 1

Phylogenetic tree of the 81 Ae. tauschii accessions constructed using 4449 DArT markers.

The Ae. tauschii population could be clearly divided into three groups designated A, B and C. Group A contained accessions from China, central Asia (Kazakhstan, Kirghizstan, Tajikistan and Turkmenistan), Afghanistan, Pakistan, India and western Asia (Georgia, Armenia, Syria) and only three accessions (KU2144, IG49095 and KU2109) were from Iran. Group B had three accessions, two from Georgia and one from Afghanistan. Accessions in group C were from Iran, with the exception of accession KU2828, IG46623, IG120866 and IG127015, which were from western Asia (Georgia, Syria, Dagestan and Armenia, respectively).

Sixty-six of the accessions in this study belonged to ssp. tauschii, of which 58 were var. typica, 6 var. meyeri and 2 var. anathera (Fig. 1). Twelve accessions were classified as ssp. strangulata and the remaining 3 accessions were of unknown variety. All of the accessions of ssp. strangulata clustered in one clade in Group C. On the other hand, the three varieties of ssp. tauschii did not cluster into a particular clade. Accessions classified as var. anathera were present only in Group A, while those classified as var. meyeri appeared in Group C, with only one (KU2109) in Group A. Using this tree, we could deduce that the 3 accessions without species information belong to ssp. tauschii. Accession IG127015 from Armenia was classified into Group C but was separate from the other accessions in this group. Accessions KU2039 (Afghanistan), AE454 (Georgia) and KU2829A (Georgia) were classified into Group B (Fig. 1).

Structural analysis of the Ae. tauschii population

The structure of the Ae. tauschii population was further studied to assess the degree of relatedness among the accessions and to group genetically similar accessions (Fig. 2). For this purpose, we used the model-based program Structure version 2.3.3 (Falush , Pritchard ). Four independent runs yielded consistent results. Values of K (number of clusters) ranging from 2 to 9 were tested. The values log-likelihood for the observed data from K = 2 to 9 are: −77679.6, −72394.1, −538662.9, −1840057.9, −111610.2, −4742273, −958412 and −1890751, respectively. K = 3 was selected as having the highest log-likelihood for the observed data, indicating that the current Ae. tauschii population can be clearly divided into three groups (Fig. 2). One of the three groups contained the accessions KU2829A, AE454 and KU2039, which are quite distant from the others. Interestingly, these lines carry useful traits for drought tolerance (Sohail , discussed later in detail). The third group corresponds to Group C in the phylogenetic tree, which includes all of the accessions of ssp. strangulata and about half (29) of the accessions of ssp. tauschii.

Fig. 2

Analysis of the population structure of 81 Ae. tauschii accessions constructed using 4449 polymorphic DArT markers and with the number of clusters (K) set to 3.

Genetic relatedness and dissimilarity among subspecies and varieties

We also analyzed the relatedness and dissimilarity between the two subspecies of Ae. tauschii and among the three varieties in ssp. tauschii. The average similarity among those in ssp. strangulata was 0.87, while that in ssp. tauschii was 0.74. Though the average similarity between accessions of var. typica and accessions of ssp. strangulata was low, the range was wide (0.61 to 0.99), indicating that some of ssp. tauschii var. typica had great similarity with some accessions of ssp. strangulata. Another interesting observation that var. typica had very high maximum similarity (0.99) with var. meyeri and ssp. strangulata. Some of the Ae. tauschii accessions had very high genetic similarities (0.99): AT47, AT76 and AT80; KU2096 and KU2098; KU2156 and KU2155; KU2088 and KU2090; KU2108 and KU2103; Aus18913 and KU2102; CGN10767 and CGN10768; KU2076 and KU2078; KU2816 and KU2824 and KU2069 and KU2828.

Discussion

Polymorphism of DArT markers from wheat and Ae. tauschii

The DArT method for performing whole-genome scans is proving to be useful for the studies of many plant species. It is currently being used for diversity studies, population structure analysis, mapping, marker-assisted selection for multiple phenotypic traits, etc. (Howard ). We tested 5500 previously developed wheat markers (Wheat Array 3.0) and 2000 Ae. tauschii markers that were produced in this study to examine the diversity of Ae. tauschii. Out of the 7500 total markers tested, 4449 showed polymorphism among the 81 accessions in this study (Table 2). The wheat markers showed a higher percentage of polymorphism (68.6%) than the Ae. tauschii markers (33.7%). This might be because the wheat markers had previously been screened and selected as good markers for the D genome of common wheat. However, Pestsova also had a similar result in a study using SSR markers that had not been prescreened and preselected: the SSR markers developed based on T. aestivum sequences revealed a higher level of polymorphism in Ae. tauschii accessions than the markers developed from Ae. tauschii itself.

Previous studies using DArT markers in cultivated crops showed less polymorphism than we found in the current study; for example, the polymorphism rate in hop (Humulus lupulus L.) was 11.9% (Howard ); sugarcane (Saccharum officinarum), 7.0% (Heller-Uszynska ); wheat (T. aestivum), 9.4% (Akbari ); cassava (Manihot esculenta), 14.6% (Xia ) and barley (Hordeum vulgare), 10.4% (Wenzl ). Badea reported the polymorphism rate of triticale for DArT markers of hexaploid wheat, triticale, and rye as 8.6%, 23.4% and 23.8%, respectively. Compared to these studies, we found higher polymorphism rate for Ae. tauschii. This might be because Ae. tauschii is a wild species and because the collection used here is representative of the wide diversity of Ae. tauschii. DArT markers of related species can be used for crops and vice versa; for example, Jing reported that T. monococcum DArT markers can be effectively used for hexaploid and tetraploid wheat as well as for diploid Triticum species.

Classification of Ae. tauschii accessions on the basis of country of origin

The phylogenetic tree of Ae. tauschii made in this study contained three largely differentiated groups, A, B and C. Most accessions in Group A were from regions other than Iran. The region encompassed by Group A starts at the Afghanistan–Turkmenistan border, extends through southern Uzbekistan, Tajikistan and Kirghizstan and reaches all the way to southern Kazakhstan. Some patches of this group are distributed along the Pakistan–Afghanistan border, in India (Jammu and Kashmir), and in western and central China. Group B only had three accessions (AE454 and KU2829A from Georgia and KU2039 from Afghanistan) these accessions also formed an independent group in the structure analysis. Two of these lines (AE454 and KU2829A) were reported to be in an unclear genealogical position by Matsuoka , 2009) and Mizuno . Most accessions in Group C were from Iran, mainly around the lower western and southern sides of the Caspian Sea and along the Iran–Turkmenistan border. Mizuno has also reported a similar grouping on the basis of AFLP results. It is noteworthy that the accessions in Group C were collected from areas with a Mediterranean climate whereas the accessions in Group A were collected from areas with an arid or semiarid steppe climate.

Most of the accessions originating from the same country were clustered together in the phylogenetic tree (Fig. 1). Among the six accessions from China, two (PI499262 and PI508262) were clustered together with those from Afghanistan and Kirghizstan and four (AT55, AT80, AT76 and AT47) were clustered together with those from Afghanistan and Turkmenistan. The first two were collected from Xingjian, which borders with Afghanistan and Kirghizstan, and the other four were from Shaanxi province. The accessions from Xinjiang must have been found in the natural distribution area of this species. However, the accessions from Shaanxi might be weedy types that were carried from Turkmenistan by human activity at some point in history. Likewise, some accessions showed high genetic similarity despite being collected from distant sites. For example, KU2069 from Iran and KU2828 from Georgia were very similar (0.99). This suggests the occurrence of migration, as Ae. tauschii has a weedy growth habit and can occur in a mixture with wheat.

The origin of Aus18913, a key accession for genome sequencing, was not known when we began our study. However, our data clearly indicated that it originated from Iran because of its close similarity (0.99) to KU2102, which was collected at 52 km northwest of Ramsar, Iran, on the southwestern coast of the Caspian Sea (http://www.shigen.nig.ac.jp/wheat/komugi/strains/nbrpDetailAction.do?strainId=KU-2102).

Classification of subspecies

Accessions classified as ssp. strangulata clearly clustered together in the phylogenetic tree (Fig. 1), despite the difficulty of classification between ssp. strangulata and ssp. tauschii based on phenotype (Dudnikov and Kawahara 2006, Dvorak , Pestsova ), the DArT markers revealed a clear cluster of ssp. strangulata. This indicates that both subspecies are genetically well diverged and that this taxonomy is probably valid. Diversity analysis using thousands of DArT markers is more powerful than analysis with other markers because of the high number and high degree of polymorphism detected by these markers. Two ssp. meyeri accessions, Aus18913 and KU2108, clustered with accessions KU2102 and KU2103 of var. tauschii respectively, and had a similarity of 0.99 with each other. The reason for this might be a large number of traits with intermediate morphology (Dudnikov and Kawahara 2006, Pestsova , van Slageren 1994) caused by hybridization between the two subspecies (Dvorak , Hammer 1980, Hammer and Knupffer 1979).

In ssp. tauschii, the varieties anathera, meyeri and typica could not be well differentiated based on the DArT markers. Mizuno suggested the reason for this might be that only a small number of genetic loci control the morphological traits used to discriminate between them. Though accessions of var. anathera were included only in Group A, accessions of var. meyeri and var. typica appeared in both groups. On the basis of RFLP markers, Lubbers reported that ssp. strangulata was more similar to var. meyeri than to var. typica. Here, no accessions of var. meyeri were closely grouped in the phylogenetic tree (Fig. 1).

Use of Ae. tauschii diversity information for wheat breeding

Because Ae. tauschii is the D-genome donor to hexaploid bread wheat, it is regarded as the most promising wild species as a genetic resource for wheat breeding (Feldman 2001, Helbaek 1959, Kihara 1944, Mujeeb-Kazi ). To introduce useful genes from this wild species into common wheat, Ae. tauschii (DD) is crossed with durum wheat (AABB) to produce a hexaploid amphiploid (AABBDD) that is called synthetic wheat (has the same genomes as bread wheat). When using this process, we should utilize the large diversity of Ae. tauschii as efficiently as possible. Sohail measured the morphological and physiological traits related to drought tolerance in many Ae. tauschii accessions and their synthetic hexaploid produced by crosses between Ae. tauschii accessions and a durum wheat, Triticum durum cv. Langdon. The results showed great diversity in drought response at both the diploid and the hexaploid levels. They found that synthetic wheats made by accessions from Georgia and central Asia (corresponding to Group B in this study) showed higher performance than others and that these synthetic wheats and may be useful for wheat breeding. The molecular information provided by the present DArT analyses will elucidate the genetic basis of the morphological and physiological characteristics at both ploidy levels. DArT makers have shown some lines to have very high genetic similarity (0.99); this information is very important for breeders to select suitable and more diverse material for breeding programs, to avoid duplication and save time and labor. This information is also important for gene banks to avoid preserving the same genotypes collected by different researchers and tagged with different accession numbers. For example, some of the accessions having high genetic similarity mentioned in the result section might be duplicates.

In conclusion, DArT markers are capable of detecting even very small polymorphisms, are cost-effective and are efficient for whole-genome scans and population structure. The present study will be useful for wheat breeding and will provide useful information with which to choose a range of accessions that represent a high degree of genetic diversity. Additionally, the new array developed here, which represents a large and diverse collection of Ae. tauschii accessions, could be an effective tool for hexaploid wheat studies.