Introduction
Wheat is the most important crop in Afghanistan and accounts for 77% of the total cereal production in this country (World Bank Report, 2012). However, wheat production has been unstable due to various abiotic and biotic stresses such as drought, yellow rust and Sunn pest (Eurygaster integriceps). Although previous studies have reported the importance of landraces for developing crops with drought tolerance (Reynolds et al., Reference Reynolds, Fernanda and Trethowan2007), disease resistance (Bonman et al., Reference Bonman, Bockelman, Jin, Hijmans and Gironella2007) and high-quality traits (Black et al., Reference Black, Panozzo, Wright and Lim2000), caution is necessary to avoid deleterious linkage drag when using landraces in breeding programmes. Therefore, searching wheat landraces to identify novel genes for critical traits is important for future breeding programmes, particularly those of rain-fed wheat improvement in Afghanistan and elsewhere.
The scientific expeditions initiated by the late Dr Hitoshi Kihara in 1955, and followed by those of other Japanese researchers in 1965 and 1978, collected Afghan wheat landraces that are now preserved at the Kihara Institute for Biological Research, Japan. Recent advancements in molecular markers with high-throughput systems show great potential for genome-wide characterization of many crops and plants (Tuberosa et al., Reference Tuberosa, Graner and Varshney2011). In particular, diversity array technology (DArT) and single-nucleotide polymorphism (SNP) markers are gaining importance in germplasm studies following reductions in the cost of genotyping. Irrespective of wheat genome size, many diagnostic markers have been developed for vernalization (Vrn), photoperiod response (Ppd), grain colour (R), leaf rust (Lr), yellow rust (Yr), stem rust (Sr) and Fusarium head blight (Fhb) (MAS wheat, http://maswheat.ucdavis.edu/; Himi et al., Reference Himi, Maekawa, Miura and Noda2011; Liu et al., Reference Liu, Pumpherey, Gill, Trick, Zhang, Dolezel, Chalhoub and Anderson2008), which aid in the characterization of any new wheat germplasm. Therefore, the aims of this study were to characterize Afghan wheat landraces using SNP and DArT markers and to identify allelic variation at the Vrn, Ppd, R, Lr, Yr, Sr and Fhb loci to find novel germplasm for Afghan wheat breeding programmes.
Materials and methods
Afghan wheat landraces numbering 446 (hereafter called ‘KAWLR’, Kihara Afghan wheat landraces) were used in this study (Fig. 1). Control samples for each gene were included for diagnostic marker screening. We extracted genomic DNA and used a Genotyping by Sequencing (GBS) 1.0 V array (www.triticarte.com.au) with approximately 50,000 probes. Diversity analysis was carried out using Phylip program. Diagnostic markers for Vrn, Ppd, R, Lr, Yr, Sr and Fhb were used after standardizing PCR conditions (Table 1). The growth habits of landraces were classified as winter, spring and/or facultative based on the allelic combination of all the three Vrn homoeologous loci. In the case of Ppd, lines were categorized as either photoperiod sensitive or insensitive. For rust and Fhb1 resistance, ‘susceptible’ landraces were identified based on reported information and the unamplified one reported as an unknown or null allele.
Afghan wheat landrace collection. Number of lines collected from each province is given in embedded boxes. The whole country is divided into eight agroecological zones (A–H) (FAO Report, 2003) and the provinces under each zone are listed.
Diagnostic molecular markers for vernalization (Vrn), photoperiod response (Ppd), grain colour (R), leaf rust (Lr), yellow rust (Yr), stem rust (Sr) and Fusarium head blight (Fhb) genes
Results and discussion
Molecular characterization is vital for any germplasm study to reveal the genomic constitution or diversity or to accelerate the process of further improvement by complementation with impeccable phenotyping. In particular, recent advancement in array-based genotyping is quickening this characterization process with more accuracy (Kilian and Graner, Reference Kilian and Graner2012). The GBS array identified 39,855 polymorphic markers in this germplasm collection. The diversity analysis of KAWLR revealed a complex diversity pattern with large diversity within the groups (see online supplementary Fig. S1). Among the 20 clusters, five were prevalent. The major clusters were well aligned with collection sites. Nei's genetic distance analysis indicated that the closest groups were Herat and Ghor, followed by Badakhshan and Takhar with genetic distances of 0.027 and 0.035, respectively. Takhar, however, exhibited the largest genetic distances with Kandahar and Badghis, with values of 0.267 and 0.199, respectively. This result differs from those of previous work on this germplasm reporting a much lower diversity (Terasawa et al., Reference Terasawa, Kawahara, Sasakuma and Sasanuma2009).
The analysis of Vrn genes (Vrn-A1, Vrn-B1 and Vrn-D1) in a total of 376 wheat landraces revealed a slightly higher frequency of the winter type (55%) than of the spring type. The classification of landraces based on agroecological zones revealed that landraces from the north-eastern and north-western regions possessed winter-type alleles when compared with those from other zones. Winter-type wheat is generally more tolerant to frost than spring-type wheat (Fujita et al., Reference Fujita, Kawada and Tahir1992) and is thus better adapted to cold areas with average winter temperatures ranging from − 5 to 5°C. In Afghanistan, unlike spring-type wheat, most winter-type wheat is grown under rain-fed conditions. Therefore, the untapped landraces of Afghanistan can be effectively used in winter wheat breeding programmes. Landraces with a spring-type Vrn allele (Vrn-A, Vrn-B or Vrn-D) were identified based on the presence of either one or a combination of Vrn homoeologous alleles on chromosome 5. The Vrn-D1 allele was present in 111 landraces followed by Vrn-B1 (70) and Vrn-A1 (16) (see online supplementary Table S1). Unexpectedly, a combination of spring alleles was not found to classify the lines with the exception of a few having the combination of Vrn-A1, -B1, Vrn-A1, -D1 or Vrn-B1, -D1. The lower frequency of Vrn-A1 alleles and their combinations suggests that even landraces classified as the spring type need medium-to-mild vernalization. In particular, the spring wheat cultivars carrying the Vrn-D1 allele for normal flowering may have a higher grain yield in a Mediterranean environment in Turkey (Andeden et al., Reference Andeden, Yediay, Baloch, Shaaf, Kilian, Nachit and Özkan2011). The majority (97%) of the landraces are photoperiod sensitive and distributed throughout the country without much dependence on agroecological zones. A similar distribution has been reported in Chinese landraces (Zhang et al., Reference Zhang, Liu, Li, Wei, Hu, Li, Lu and Yang2010). The combination of two growth habit traits (i.e. vernalization and photoperiod) is critical for determining the adaptive nature of accessions and their use in breeding programmes. According to Kato and Yamashita (Reference Kato and Yamashita1991), genetic variation in these three traits influencing heading date is highest in areas in Iran and Afghanistan and wheat landraces can adapt to these areas by combining different alleles of such traits.
Grain colour is an end-use quality trait, and white-coloured seeds are preferred in Afghanistan. Based on the classification of lines by combining all the three alleles, 40% of the landraces were found to be of white grain type, whereas the remaining were red and amber grain types. Landraces collected in the south-western region tended to have a white R allele, whereas those collected in the north-western region had a red R allele. When analysing individual loci, the presence of the white allele was prevalent than that of the red allele in all the three homoeologous alleles on chromosome 3 (see online supplementary Table S1). Overall, the white allele (white seed) may have been selected in Afghanistan by the agriculture history.
To determine the potential of KAWLR, diagnostic markers for Lr, Yr, Sr and Fhb were screened. We primarily chose durable rust-resistant genes that were not specific to any races. In the KAWLR set, 17 landraces had positive alleles for the mentioned diseases. The entire set was screened for the Lr34 gene, and a subset of lines was selected based on preliminary field screening (data not shown) and diversity analysis. Four lines were found to be positive for Lr46 and three lines had a Sr-resistant allele. Both Lr34 and Lr46, having a durable resistance mechanism, also showed resistance to multiple diseases (Singh et al., Reference Singh, Huerta-Espino, Bhavani, Herrera-Fiessel, Singh, Singh, Velu, Mason, Jin, Njau and Crossa2011). In Fhb1 screening, only five accessions were found to have resistant alleles and 67 accessions susceptible alleles. Though the gene-specific genotyping revealed that only a few landraces had resistant alleles compared with susceptible ones, it was possible to identify the haplotypes of the known genes. Additionally, our results suggest that new resistance alleles might be present in the accessions evaluated in this study.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1479262114000203
