Introduction
In bread wheat (Triticum aestivum L.), at least 21 major genes dispersed over a number of chromosomes determine the overall height of plant. The Green Revolution semi-dwarfing genes Rht-B1b and Rht-D1b have widely been used for cultivar development (Knopf et al., Reference Knopf, Becker, Ebmeyer and Korzun2008). Both of these semi-dwarfing alleles are insensitive to exogenous gibberellic acid (GA3; McIntosh et al., Reference McIntosh, Hart, Gale, Li and Xin1995) and were sourced from the Japanese cultivar ‘Norin-10’, from which they were transferred to the Green Revolution wheat bred at CIMMYT, and thereafter to an estimated 70% of commercial cultivars grown across the whole range of wheat production containing at least one of them (Evans, Reference Evans1998). While the dwarfing effect of the Rht-1 alleles cannot be reversed by the provision of exogenous GA3, a number of other dwarfing genes are sensitive to GA application. Different GA3-sensitive genes such as Rht4 have been mapped on chromosomes 2BL, Rht5 on 3BS, Rht8 on 2DS, Rht9 and Rht12 on 5AL, and Rht13 on 7BS (Ellis et al., Reference Ellis, Rebetzke, Azanza, Richards and Spielmeyer2005), while the GA-insensitive Rht-B1 and Rht-D1 genes reside on the short arms of, respectively, chromosomes 4B and 4D (Gale and Marshall, Reference Gale and Marshall1975; McVittie et al., Reference McVittie, Gale, Marshall and Westcott1978; Börner et al., Reference Börner, Röder and Korzun1997). Following the isolation of the latter genes by Peng et al. (Reference Peng, Carol, Richards, King, Cowling, Murphy and Harberd1997, Reference Peng, Richards, Hartley, Murphy, Devos, Flintham, Beales, Fish, Worland, Pelica, Sudhakar, Christou, Snape, Gale and Harberd1999), it became clear that both semi-dwarfing alleles derive from a single base pair mutation that abolishes the plant's ability to respond to GA, and this difference soon led to the elaboration of diagnostic polymerase chain reaction (PCR)-based assays for their presence (Ellis et al., Reference Ellis, Spielmeyer, Gale, Rebetzke and Richards2002).
Both Rht-B1b and Rht-D1b reduce plant height and increase yield under favourable climatic conditions (Worland et al., Reference Worland, Korzun, Röder and Ganal1998; Petsova et al., Reference Pestsova, Korzun and Börner2008), but they are not beneficial in the Mediterranean environment, where high temperatures and drought are commonly encountered in the period between anthesis and grain filling (Worland and Sayers, Reference Worland and Sayers1995). As a result, alternative dwarfing genes, such as the GA-sensitive Rht8, have become widespread among cultivars grown in the Mediterranean basin and Southern Europe. This gene, along with Ppd-D1, another important gene associated with adaptation, is located on the short arm of chromosome 2D. When favourable alleles are present at both Rht8 and Ppd-D1, as in the Japanese cultivar ‘Akakomugi’, plant height is reduced by 10 cm, flowering is accelerated by 8 d and spikelet fertility is markedly improved (Worland et al., Reference Worland, Korzun, Röder and Ganal1998; Zhang et al., Reference Zhang, Yang, Zhou, He and Xia2006). These genes were introduced to the European wheat gene pool by the Italian breeder Nazareno Strampelli in the early 1940s, who used them to avoid heat stress during grain filling. Since then, Rht8 has been exploited throughout the Mediterranean basin (Worland et al., Reference Worland, Korzun, Röder and Ganal1998) and has more recently been spread to Australia, where it is favoured in some wheat-growing areas as it is not associated with the poor seedling emergence in impacted soils that do affect carriers of the Rht-1 (Rht-B1b and Rht-D1b) semi-dwarfing alleles (Rebetzke and Richards, Reference Rebetzke and Richards2000; Bonnett et al., Reference Bonnett, Ellis, Rebetzke, Condon, Spielmeyer and Richard2001; Ellis et al., Reference Ellis, Rebetzke, Azanza, Richards and Spielmeyer2005, Reference Ellis, Bonnet and Rebetzke2007). Mapping experiments have established that Rht8 is closely linked to the microsatellite locus Xwgm261 (Korzun et al., Reference Korzun, Röder, Ganal, Worland and Law1998), specifically, the 192 bp allele of Xgwm261 is associated with the dwarfing allele of Rht8. However, Ellis et al. (Reference Ellis, Bonnet and Rebetzke2007) reported that 192 bp allele at Xgwm261 is not always associated with Rht8 dwarfing gene in wheat. They further explained that wide spread use of Norin-10-derived germplasm from the CIMMYT green revolution germplasm introduced a second haplotype into international germplasm, in which Xgwm261 has no association with Rht8.
Turkey is a major producer of wheat, with some 9Mha sown annually leading to an annual production of 20 Mt (Altintas et al., Reference Altintas, Toklu, Kafkas, Kilian, Brandolini and Ozkan2008). Modern breeding in Turkey started in 1925 with the goal to select well-adapted lines from local landraces for wheat improvement. In 1967, the national wheat release and training project was established, and international organizations contributed by introducing several cultivars from foreign countries (Braun et al., Reference Braun, Zincirci, Altay, Atli, Avci, Eser, Kambertay, Payne, Bonjean and Agnus2001). Large parts of Turkey are located within the Fertile Crescent, the region where wheat originated and where wheat was brought in cultivation and domestication (Kilian et al., Reference Kilian, Özkan, Pozzi, Salamini, Feuillet and Muehlbauer2009). It is essential to have information about allelic variations at agricultural important loci in Turkish wheat collections at hand. Some characterization of Turkish germplasm has been undertaken, in particular with respect to the presence of the important 1B/1R and 1A/1R wheat–rye translocations (Yediay et al., Reference Yediay, Baloch, Kilian and Ozkan2010) and the allelic constitution at the major vernalization and photoperiod requirement genes (Andeden et al., Reference Andeden, Yediay, Baloch, Nachit, Shaaf, Kilian and Ozkan2011). In this study, we report the allelic constitution at the major semi-dwarfing genes of the Turkish bread wheat core collection.
Material and methods
Plant material, DNA extraction and PCR analysis
The bread wheat core collection consists of 65 accessions, comprising 56 cultivars released for cultivation in Turkey over the past 70 years and nine landraces (Supplementary Table S1, available online only at http://journals.cambridge.org). As controls, four isogenic lines in ‘Norin-10’ (Rht1+Rht2; Rht1+rht2; rht1+Rht2 and rht1+rht2) were included. From each accession, eight plants were grown under greenhouse conditions, from which young leaf was collected from 10-d-old seedlings. DNA was extracted from snap-frozen leaf material from four seedlings per accession individually, following the Doyle and Doyle (Reference Doyle and Doyle1987) Cetyl trimethylammonium bromide method, as modified by Ozkan et al. (Reference Ozkan, Brandolini, Pozzi, Effgen, Wunder and Salamini2005). We applied six PCR primer pairs diagnostic for the Rht-B1 and Rht-D1 alleles (Ellis et al., Reference Ellis, Spielmeyer, Gale, Rebetzke and Richards2002; Zhang et al., Reference Zhang, Yang, Zhou, He and Xia2006) to obtain allele calling at these two loci. Experimental details relating to these assays are given in Table 1. As the Rht-B1 targeted primer combinations BF-WRI and BF-MRI (Ellis et al., Reference Ellis, Spielmeyer, Gale, Rebetzke and Richards2002) were not fully diagnostic when tested on the control plants, the Zhang et al. (Reference Zhang, Yang, Zhou, He and Xia2006) procedure was preferred, in which the relevant primer combinations were NH.BF2-MR1.2 and NH.BF2-MR1. Gradient PCR experiments were used to determine an optimum annealing temperature of 66°C. For Rht-D1, the Ellis et al. (Reference Ellis, Spielmeyer, Gale, Rebetzke and Richards2002) primers DF-WR2 and DF2-MR2 were effective. All amplicons were electrophoretically separated through 2% agarose gels and visualized by ethidium bromide staining. All experiments were repeated three times. For Rht8 typing, amplicon length at the Xwgm261 locus was obtained following the methods described by Korzun et al. (Reference Korzun, Röder, Ganal, Worland and Law1998). All PCR were performed according to Cömertpay et al. (Reference Cömertpay, Baloch, Kilian, Ülger and Özkan2011). The amplicons were separated by capillary electrophoresis on an ABI 3130xl Genetic Analyser (Applied Biosystems, Foster City, USA) platform, and the output was handled by GeneMapper software v3.7 (Applied Biosystems).
Table 1 Specific DNA molecular markers for three dwarfing gene considered in this study
Results
The allelic state at Rht-B1 and Rht-D1 for each of the accessions is given in Supplementary Table S1 (available online only at http://journals.cambridge.org). At Rht-B1, 41 of the 65 accessions carried a 372 bp NH.BF2-WR1.2 fragment, diagnostic for the wild-type (non-dwarfing) Rht-B1a allele, while the remaining 24 accessions were presumed to carry Rht-B1b because no product was amplified. The presence of Rht-B1b in these accessions was confirmed by amplification profiles obtained using the primer pair NH.BF2-MR1.2, since a 380 bp fragment was present in all of them but was absent from each of the 41 lines typed as carrying Rht-B1a. With respect to the allelic status at Rht-D1, the only accession from which the 264 bp fragment diagnostic for Rht-D1b was amplified was cultivar ‘Pandas’. The other 64 accessions produced a 220 bp fragment when their DNA was amplified with the DF-MR2 primer pair diagnostic for Rht-D1a. Amplification profiles for Rht-B1a and Rht-B1b for some bread wheat genotypes are shown in the Fig. 1. Finally, for Rht8, at least seven distinct Xgwm261 amplicon lengths were represented in the core collection (Fig. 2). At microsatellite Xgwm261 locus, 17 bread wheat cultivars and landraces (26%) carried 192 bp allele; 35% (23 entries) carried 165 bp allele, whereas only 12% (8 entries) carried 174 bp allele. Of the 65 accessions, 48 carried either the 165 bp, the 174 bp or the 192 bp alleles, while six carried the 202 bp allele, five carried the 211 bp allele, three carried the 196 bp allele and one carried the 169 bp allele. The 192 bp allele diagnostic for Rht8 was present in 18 accessions (Supplementary Table S1, available online only at http://journals.cambridge.org).
Fig. 1 Amplification profiles derived from primer combination NH-BF.2-WR1.2 diagnostic for Rht-B1a and primer combination NH-BF.2-MR2 diagnostic for Rht-B1b. 1, Atay-85; 2, Melez13; 3, Yektay; 4, Gerek 79; 5, Doğankent-1; 6, Kate A-1; 7, Kutluk-94; 8, İkizce-96; 9, Palandöken 97; 10, Süzen 97; 11, Demir 2000; 12, Bayraktar 2000; 13, Pandas; 14, Bağcı-2002; 15, Nenehatun; 16, Pehlivan; 17, Türkmen; 18, Yıldız 98; 19, Cumhuriyet 75; 20, Sertak 52; 21, Kırmızı başak; 22, Adana99. MW, molecular weight.
Fig. 2 Allelic variation at Xgwm261. The arrows show the calculated size of each DNA fragment. Saraybosna (192 bp), Pandas (165 bp), Kırik (211 bp) and Porsuk 2800 (174 bp).
Discussion
Dwarfing genes have been credited with an important contribution to yield improvement, both because they permit a more efficient utilization of assimilate and reduce the extent of lodging-induced yield loss. In warm humid environments, taller plants tend to produce less leaf area, and therefore are less effective as photosynthesizers and assimilators (Ahmad and Sorrells, Reference Ahmad and Sorrells2002). The various dwarfing genes vary with respect to their effect on height, grain yield and other aspects of agronomic performance (Worland et al., Reference Worland, Korzun, Röder and Ganal1998; Ahmad and Sorrells, Reference Ahmad and Sorrells2002; Ganeva et al., Reference Ganeva, Korzun, Landjeva, Tsenov and Atanasova2005; Zhang et al., Reference Zhang, Yang, Zhou, He and Xia2006; Ellis et al., Reference Ellis, Bonnet and Rebetzke2007; Knopf et al., Reference Knopf, Becker, Ebmeyer and Korzun2008; Pestsova et al., Reference Pestsova, Korzun and Börner2008; Guedira et al., Reference Guedira, Brown-Guedira, Van Sanford, Sneller, Souza and Marshall2010).
The presence of dwarfing genes can be monitored directly from the behaviour of the seedling (Pestsova et al., Reference Pestsova, Korzun and Börner2008; Tang et al., Reference Tang, Jiang, He and Hu2009), but these techniques can be rather time and labour intensive, and sometimes are also influenced by the testing environment. The use of exogenously supplied GA is very effective for discriminating between semi-dwarf and tall types but is unable to differentiate between the Rht-B1b and Rht-D1b alleles. The acquisition of the DNA sequence of these two key genes now permits their unambiguous monitoring by marker technology (Korzun et al., Reference Korzun, Röder, Ganal, Worland and Law1998; Ellis et al., Reference Ellis, Spielmeyer, Gale, Rebetzke and Richards2002; Zhang et al., Reference Zhang, Yang, Zhou, He and Xia2006), while the presence of Rht8 can be fairly reliably predicted by the allelic state of a tightly linked microsatellite locus. However, Ellis et al. (Reference Ellis, Bonnet and Rebetzke2007) report that the 192 bp allele at the Xgwm261 locus is not always associated with the Rht8 dwarfing gene in wheat.
In the Turkish bread wheat core collection, 37% of the accessions (24 entries) carried Rht-B1b, and the rest carried the wild-type Rht-B1a allele (Supplementary Table S1, available online only at http://journals.cambridge.org). Correspondingly, the Rht-D1b allele was carried by only one Turkish cultivar, which was released within the last 10 years. Rht-B1b is rather uncommon in German wheat germplasm (Knopf et al., Reference Knopf, Becker, Ebmeyer and Korzun2008), while it dominates Chinese wheat (Yang and Liu, Reference Yang and Liu2006; Zhang et al., Reference Zhang, Yang, Zhou, He and Xia2006). Rht-D1b appears to be rather more frequent than Rht-B1b among semi-dwarf US soft winter wheat (Guedira et al., Reference Guedira, Brown-Guedira, Van Sanford, Sneller, Souza and Marshall2010). Nearly, all Turkish cultivars released after 1967 were bred directly and/or indirectly from materials provided by the CIMMYT (Altintas et al., Reference Altintas, Toklu, Kafkas, Kilian, Brandolini and Ozkan2008; Yediay et al., Reference Yediay, Baloch, Kilian and Ozkan2010), but although the CIMMYT programme exploited both Rht-B1b and Rht-D1b, the percentage of Rht-B1 is remarkably small and Rht-D1b has been found in only one line. Low frequency of Rht-B1b and the nearly absence of Rht-D1b in Turkish wheat might be also correlated with higher temperature during ear emergence and drought conditions especially in the Mediterranean and South-east Anatolian regions. Worland and Law (Reference Worland and Law1985) reported that the distribution of GA-insensitive dwarfing genes is restricted to areas where heat and drought stress condition prevails during grain filling.
The Rht-D1b dwarf allele was found only in the cultivar ‘Pandas’ having the wild tall allele Rht-B1a (Supplementary Table S1, available online only at http://journals.cambridge.org). This cultivar originated from Italy, carried the 165 bp allele for Xgwm261 locus, besides having Bezostaya as one of the parents in its pedigree. ‘Pandas’ is a semi-dwarf wheat cultivar with approximately 67–75 cm plant height, good yield potential, wide adaptation and is widely grown in the Mediterranean part of Turkey (Yücel et al., Reference Yücel, Baloch and Özkan2009).
It has been suggested that the presence of the GA-insensitive dwarfing genes is associated with a reduction in coleoptiles length and hence a poorer rate of seedling emergence from impacted soils (Allan, Reference Allan1989). In addition, they appear to exert a negative effect on spikelet fertility and grain yield in environments suffering from frequent episodes of heat stress (Worland and Law, Reference Worland and Law1985). Ellis et al. (Reference Ellis, Rebetzke, Azanza, Richards and Spielmeyer2005) have suggested that both Rht-1 dwarfing alleles also negatively affect the growth of young plants, but that this effect is not mirrored by Rht8. Therefore, Rht8 might be more suitable in reducing final plant height, without compromising early plant growth (Ellis et al., Reference Ellis, Rebetzke, Azanza, Richards and Spielmeyer2005). In this study, the Xgwm261 locus is quite polymorphic, and seven alleles (165, 169, 174, 192, 196, 202 and 211 bp) were represented in the Turkish core collection. About 76% of the accessions carried either the 165 bp, the 174 bp or the 192 bp alleles, and the other four alleles were each rather rare.
The 211 bp allele was present in ‘Melez-13’, ‘4-22’ and ‘Kırik’. Cultivar Melez-13 came from a cross of Italian cultivar Mentana (Strampelli cultivar carrier of 165 bp allele) with unknown local landrace, and later on cultivar ‘4-22’ was selected from Melez-13 population. This finding suggested that 211 bp allele was already present in old Turkish local cultivars. Worland et al. (Reference Worland, Sayers and Korzun2001) reported the presence of this same allele in a number of observed Turkish cultivars. Ganeva et al. (Reference Ganeva, Korzun, Landjeva, Tsenov and Atanasova2005) also suggested that the 211 bp present in the Bulgarian cultivar ‘Ivanov’ may have originated from Turkish germplasm.
About 26% of the studied Turkish wheat cultivars and landraces carried the 192 bp allele at microsatellite Xgwm261 locus. However, the frequency of the 192 bp allele in Turkish wheat is much lower than that in Italian (60%) and Yugoslavian (86%) wheat (Guedira et al., Reference Guedira, Brown-Guedira, Van Sanford, Sneller, Souza and Marshall2010). Correlated selection for this allele may have been driven by its linkage with photoperiod-insensitive allele at Ppd-D1, For example, the cultivar ‘Çukurova86’ and Doğankent1, Pehlivan, Golia having 192 bp allele at Xgwm261 locus, also carried photoperiod-insensitive allele Ppd-D1a (Andeden et al., Reference Andeden, Yediay, Baloch, Nachit, Shaaf, Kilian and Ozkan2011). However, we also observed that 192 bp allele is not always associated with Ppd-D1a. Some wheat cultivars having Ppd-D1a allele carried alleles other than 192 bp (Andeden et al., Reference Andeden, Yediay, Baloch, Nachit, Shaaf, Kilian and Ozkan2011).
It is very interesting to note that Xgwm261-192 allele was also present in Turkish wheat landraces (Koca Buğday, Kırmızı Başak and Arpatan), explaining the one of the probable source of 192 bp allele into Turkish wheat cultivars. In this case, one could expect that Xgwm261-192 allele in Turkish wheat is not associated with height-reducing gene. It could be assumed that 192 bp allele might appear around 1965–1970. At that time Norin-10-derived CIMMYT material was heavily used in Turkish wheat programmes, which shows the second possibility of height neutral Xgwm261-192 allele. Both of these possibilities have been described by Ellis et al. (Reference Ellis, Bonnet and Rebetzke2007); they stated that N Borlaug's semi-dwarf Mexican varieties have different haplotype in which Xgwm261-192 allele is not associated with height-reducing gene. They proposed a hypothesis that Xgwm261-192 arose prior to the evolution of Rht8 and, in such case, it would be expected to persist in landraces as observed in Turkish landraces (Supplementary Table S1, available online only at http://journals.cambridge.org). Third possible pathway of 192 bp allele in the Turkish wheat-breeding programme is the use of Russian wheat cultivars ‘Bezostaya and Kavkas’. Bezostaya, which is universally recognized photoperiod insensitive, semi-dwarf wheat cultivar carrying Xgwm261-192, contained Strampelli variety ‘Ardito’ in its pedigree (Worland et al., Reference Worland, Korzun, Röder and Ganal1998; Borojevic and Borojevic, Reference Borojevic and Borojevic2005). Bezostaya and Kavkas were among the key varieties used in Turkish wheat-breeding programme, thus providing an independent source of Akakomugi 192 bp allele in Turkish wheat gene pool. In such case, cultivars carrying Akakomugi Xgwm261-192 allele should be associated with height-reducing genes. Turkish cultivars carrying photoperiod-insensitive allele (Ppd-D1a) should have Akakomugi Xgwm261-192 haplotypes associated with reduced plant height. At the moment, it is difficult to distinguish the above-mentioned Xgwm261-192 haplotypes in Turkish wheat without having phenotypic data for plant height, although it can be depicted from their pedigree. However, detailed field trial is also needed to elucidate the effect of different alleles on the plant height of Turkish wheat under different production environment.
The most frequent Xgwm261 allele was the 165 bp amplicon in the Turkish wheat varieties, which was also common in a sample of CIMMYT-based Australian, Chinese, Greek, Portuguese, Spanish and Turkish cultivars (Zheleva et al., Reference Zheleva, Todorovska, Jacquemin, Atanassov, Christov, Panayotov and Tsenov2006). Worland et al. (Reference Worland, Korzun, Röder and Ganal1998) reported that Strampelli cultivar Mentana carries Xgwm261 165 allele. This cultivar used as parental line in some crossing experiment in Turkish wheat programme as well. Mentana was also one of the important parents in the Mexican and CIMMYT wheat research and breeding programme and may be the source of 165 bp allele in the CIMMYT cultivars. It is also important to note that 165 bp allele started to be appeared in Turkish cultivars in 1970.
Xgwm261 174 allele was present only in eight genotypes. Interestingly, the 174 bp allele mostly exists in cultivars released until 1979 before introducing the CIMMYT material to Turkey. The two cultivars Sertak-52 and Sivas-111/33 carried the 174 bp allele, and these were directly selected from local landraces and subsequently registered as cultivars. Andeden et al. (Reference Andeden, Yediay, Baloch, Nachit, Shaaf, Kilian and Ozkan2011) mentioned that these cultivars carry the photoperiod-sensitive allele (Ppd-D1b). Both of these cultivars were grown in the Central Anatolian region and Northern Turkey, where growing period of the wheat crop is longer due to cold temperature compared with the other parts of Turkey. Worland et al. (Reference Worland, Korzun, Röder and Ganal1998) reported that the 174 bp allele was common in Northern European wheat, where most of the wheat cultivars are photoperiod sensitive due to long life cycle of wheat-growing period. Ahmad and Sorrells (Reference Ahmad and Sorrells2002) also found that the 174 bp allele to be present in Great Britain, German and French wheat cultivar.
All the cultivars in the core set released prior to 1979 were Rht-B1a+ Rht-D1a. This indicates the reality that selection for reduced plant height was not undertaken in Turkey until the CIMMYT germplasm had been introduced. The Turkish ‘National Wheat Release and Training Project’ was established in 1967 and was initiated with genetic input from CIMMYT and other international organizations. Since that time, the proportion of cultivars that were semi-dwarf (either Rht-B1b or Rht8) has increased rapidly. In the Mediterranean and South-east Anatolian regions of Turkey, where high temperatures are common place from anthesis onwards and especially during grain filling, since the GA-insensitive Rht-1 dwarfing genes are not well adapted to these conditions, Rht8 would be a valuable resource for Turkish wheat breeding.
In this study, we have provided information concerning the allelic state at the three major semi-dwarfing genes across a core collection of Turkish bread wheat. Together with matching data concerned with vernalization and photoperiod requirement genes (Andeden et al., Reference Andeden, Yediay, Baloch, Nachit, Shaaf, Kilian and Ozkan2011) and the 1B/1R and 1A/1R wheat–rye translocations (Yediay et al., Reference Yediay, Baloch, Kilian and Ozkan2010), this information should allow for a more rational use of this collection for the purpose of wheat improvement. We expect that this will facilitate advances in cultivar adaptation, raising grain productivity through the application of marker-assisted selection. To understand the history and origin of different alleles of dwarfing genes in Turkish wheat germplasm, further detailed study is needed over large collection of germplasm containing all old Turkish cultivars and local landraces.
Acknowledgements
We warmly acknowledge TÜBİTAK (The Scientific and Technological Research Council of Turkey, TOVAG-1070207) and University of Çukurova, Scientific Research Projects Unit (ZF2007YL28), for their financial support. We thank Dr T.R. Endo (Genetic Resource Bank, Kihara Institute for Biological Research, Japan) for the kind provision of dwarfing gene isogenic lines.
