The diploid einkorn wheat Triticum monococcum L. subsp. monococcum (2n = 2x = 14, A^mA^m) (hereafter Tm) was among the first crops domesticated in the Fertile Crescent starting from its wild progenitor Triticum monococcum L. subsp. boeoticum (Boiss.) Á. Löve et D. Löve (2n = 2x = 14, A^bA^b) (hereafter Tb). The archeological record of einkorn domestication both in terms of date and place remains elusive maybe because it is a minor crop that usually is associated with emmer. However, using molecular markers, the present-day distribution of a progenitor of domestic einkorn has been located in south-eastern Turkey – between the volcanic mountain ranges of Kartal–Karadaǧ and Karacadaǧ (Heun et al., 1997; Kilian et al., 2007).

The geographical range and the long-time periods shown from the archeological evidence points to a protracted and geographically diffuse processes of domestication for most if not for all Near Eastern crops, rather than a single process (Lev-Yadun et al., 2000; Fuller et al., 2011, 2012; Willcox, 2013), however, some would disagree (Heun et al., 2012). This applies also to einkorn wheat, for which the ‘dispersed-specific model’ has been proposed (Kilian et al., 2007).

On Near Eastern archeological sites, emmer and barley were the staple cereals and those where einkorn dominate are the exception. One example is Tell Qaramel (Willcox and Herveux, 2013). Two-grained einkorn was identified at the late Pleistocene sites of Mureybet and Abu Hureyra. But there was a doubt with regard to the identification because of the confusion with wild Secale species. The term Triticum/Secale was adopted, however, we now know that this taxon is dominated by wild rye (Willcox, 2007, 2014). Einkorn on Near Eastern sites is represented predominantly by two-grained forms (Özkan et al., 2011) while on European sites the single-grained type dominates (Kreuz and Boenke, 2002). During the last millennia, hulled wheats were largely replaced by free-threshing tetraploid (durum wheat) and hexaploid wheats (bread wheat), which deliver higher yields. Today, einkorn is cultivated on a very small scale, i.e., in the Balkans, France, Italy, Morocco, Spain, and Turkey (Schiemann, 1956, 1957; Nesbitt and Samuel, 1996; Perrino et al., 1996; Zaharieva and Monneveux, 2014). This diploid wheat has been re-discovered as a source of genetic variation for wheat breeding, and it is increasingly used by the organic food industry in Europe due to its nutritional value and biodiversity that can be exploited for the improvement of the health-related traits of modern wheat (Arzani and Ashraf, 2017).

Multi-locus re-sequencing analysis of a comprehensive collection of 321 wild and 92 domesticated einkorn lines revealed, for instance, that wild einkorn consists of three distinct Tb races (i.e., alpha, beta, gamma) (Kilian et al., 2007). One of those races, the wild race beta, is genetically much more similar to domesticated einkorn, hence it is the presumed race that was the progenitor of all known present-day domestic einkorns. Today race beta occurs only in the Kartal–Karadaǧ and Karacadaǧ mountains. The diffusion of domesticated einkorn out of the Fertile Crescent was reconstructed based on archaeobotanical remains and genotyping data summarized by (Zaharieva and Monneveux, 2014; Brandolini et al., 2016).

For simplicity, we follow the classification system of J. MacKey (MacKey, 1988), although wild and domesticated einkorn are not separated by any significant crossing barrier (Harlan and Zohary, 1966). The rachis type is one of the most conspicuous features distinguishing Tb from Tm: The Tb forms have a fragile rachis, which promotes seed dispersal in the wild, while the latter’s rachis is non-fragile and breaks when threshed (Figure 1), a prominent component of the domestication syndrome (Hillman and Davies, 1990). The fracture point of rachis in einkorn wheat lies above the node to produce a wedge-shape rachis node (Sakuma et al., 2011). The wedge type is found in diploid and tetraploid wheats harboring the A genome (A, AB, AG) (Li and Gill, 2006) as well as for hexaploid (ABD) (Watanabe et al., 2002). Wild einkorn wheat (Tb) produces wedge-shape spikelets rather similar to those of wild barley (Hordeum vulgare ssp. spontaneum (C. Koch) Thell. In barley, the brittle rachis trait is conditioned by the allelic status at two physically linked genes Non-brittle rachis 1 (btr1) and Non-brittle rachis 2 (btr2): a loss-of-function in either gene results in the formation of a non-brittle rachis. Loss-of-function mutations in both genes are found in the domesticated barley genepool, implying at least three independent origins of the stiff rachis phenotype in barley (Pourkheirandish et al., 2015; Civan and Brown, 2017). The objective of this study was to identify the causative polymorphism responsible for the non-brittle rachis trait of domesticated einkorn and to find closest present-day representatives of the wild ancestors of cultivated einkorn that could indicate the possible location of einkorn domestication.

In barley, Btr1 and Btr2 act as complementary genes, with loss-of-function mutations in either of them generating the non-brittle rachis phenotype (Pourkheirandish et al., 2015). In einkorn wheat, a critical A119T substitution at Btr1 was found to discriminate between wild and domesticated einkorn wheat. A119T lies outside either of the two predicted transmembrane helices formed by BTR1, but it nevertheless appears to represent a component of the active site of the einkorn wheat BTR1 protein. The replacement of a threonine by an alanine residue has been suggested to facilitate receptor internalization in a rat opiod receptor (Wolf et al., 1999), and has also been shown to affect the activity of a histone protein (Jacob et al., 2014). Whether the A119T substitution in some way impairs the receptor-ligand structure, as suggested by Pourkheirandish et al. (2015), remains to be established. These findings suggest that the A119T at Btr1 is the critical polymorphism, which causes the non-brittle (a stiffer) rachis phenotype of domesticated einkorn. In this study, we did not find any einkorn wheat carrying functional allele of Btr1 in combination with loss-of-functional allele of Btr2. Whether Btr1 and Btr2 are functionally linked and act complementary genes as for barley has still to be investigated.

Today the most extensive natural stands of wild einkorn wheat are found in the northern part of the Fertile Crescent (Harlan and Zohary, 1966). However, we do not know the extent of natural stands at the end of the Pleistocene and the beginning of the Holocene. Archeological finds provide evidence for wild einkorn cultivation in northern Syria and Eastern Turkey between 11500 and 10500 cal BP. Signs of morphological domestication only appear after this long period of pre-domestication cultivation (Tanno and Willcox, 2012). Multi-locus genotyping (incl. re-sequencing) studies have concluded that the most likely site of the progenitor of einkorn wheat lies in south-eastern Turkey (Heun et al., 1997; Kilian et al., 2007). In the previous study by Kilian et al. (2007), the closest wild to domesticated einkorn were collected within the Kartal–Karadaǧ and Karacadaǧ mountain ranges on basaltic soils (Kilian et al., 2007).

Here, a haplotype analysis of Btr1 and Btr2 sequences showed that two wild einkorn accessions, which resembled the immediate progenitor – at Btr1 and Btr2 – of all domesticated einkorn wheats included in this study, both were collected between Kahramanmaraş and Gaziantep. This location lies immediately east of the Kartal–Karadaǧ mountain range and is the region where the progenitor of einkorn survives until today (Kilian et al., 2007). Furthermore, it has to be seen, if the A119T mutation led to the foundation of all known ‘eco-geographical’ or ‘geographical provenance’ groups/ lineages of domesticated einkorn (Flaksberger, 1935; Dorofeev et al., 1979; Szabó and Hammer, 1996; Zaharieva and Monneveux, 2014; Brandolini et al., 2016). Re-sequencing of Btr1 and Btr2 in a comprehensive collection of wild and domesticated accessions from diverse regions will answer if the non-brittle einkorn is selected only once (A119T) or multiple times.

One of the few early Neolithic sites where einkorn is the dominant crop, Tell Qaramel (Willcox and Herveux, 2013), lies less than 100 km to the south of where the two Tb accessions were collected (Figure 7). This adds weight to the hypothesis that it was in this general region that einkorn was first taken into cultivation. Willcox and Herveux (2013) suggested on the basis of an association with arable weeds that morphologically wild einkorn was cultivated at Tell Qaramel by approximately 11500 cal BP. Could this be the earliest example of cultivated wild einkorn, which later diffused as a minor component of emmer fields over a wider area of northern Levant and into Cyprus by 10600 cal BP? More archaeobotanical evidences are needed to answer this question but we can speculate that at some point the btr1 allele appeared, fixed and dispersed to agrarian societies far away from the Karadaǧ mountain range. Einkorn is diploid and mutation in Btr1, rather than accumulated mutations in multiple loci or multiple genomes, was sufficient to produce non-brittle rachis. Therefore, producing non-brittle rachis seems to be achieved in rather a shorter time than domestic emmer wheat which required fixation of two non-functional mutations in the homoeologous Btr1 genes in the A and B sub-genomes (Avni et al., 2017). The possible unintentional fixation of non-brittle rachis in diploid wheat or barley may led to an intentional selection of two mutations in tetraploid wheat genome A and B. It is true that QTLs other than the btr1 were identified in wheat and barley but with a milder phenotypic effect (Komatsuda et al., 2004; Jiang et al., 2014).

Archaeobotanical finds indicate that it probably took over one millennium between the start of cultivation and the establishment of populations dominated by semi-brittle types (Tanno and Willcox, 2012). These finds provide evidence for multiple “false starts” and “dead ends” in the long history of plant and animal domestication, i.e., early domesticated lineages went extinct or diffused out of their original domestication area – and became slowly adapted to other environments (Fuller et al., 2011, 2012). In other words, we do not know how many domesticates were lost and we do not know to what extend the modern crop genepool (incl. the crop wild relatives) represents the situation in the past. However, Kilian et al. (2007) presented evidence that einkorn, in contrast to more intensely bred crops, possibly underwent little reduction of diversity during domestication.

The wild ancestors of domesticated einkorn btr1 (A119T) have been found in the same geographical area as the immediate wild ancestors of the btr2 (Pourkheirandish et al., 2015) and btr1b (Civan and Brown, 2017) lineages of domesticated barley. More precisely, the closest wild relatives to these non-brittle rachis phenotypes were all found northwest and southwest of Gaziantep and one of which (wild barley accession FT730) was collected in the Kartal–Karadaǧ mountains. This coincidence supports the ‘Cradle of Agriculture’ theory of Lev-Yadun et al. (2000) and in this case similar results should be found for the other founder crops of the Near East (Emmer wheat, pea, chickpea, lentil, and bitter vetch).

The genomic organization of Btr1/Btr2 and Btr1-like/Btr2-like gene pairs on chromosome 3A of hexaploid bread wheat cultivar CS was exactly the same as that on barley chromosome 3H (Pourkheirandish et al., 2015). This demonstrates that the configuration must have been present in the common ancestor of wheat and barley, which diverged some eight million years ago (Middleton et al., 2014). The implied duplication event, which formed the Btr1/Btr2 and the Btr1-like/Btr2-like pairs must therefore have occurred prior to the separation of the wheat and barley lineages. The sequence of the einkorn Btr genes is highly homologous to that of their barley orthologs. Their predicted products are of the same size and its BTR1 protein forms the same two transmembrane helices seen in both barley and emmer wheat BTR1 (Pourkheirandish et al., 2015; Avni et al., 2017). As for BTR2, neither the ‘CAR’ nor the ‘PIP’ motifs found in barley were reproduced in the wheat protein of different ploidy levels. The wild forms of both einkorn and emmer wheat, just as does that of barley produce spikelets with a wedge type rachis internode, promoting spike disarticulation above each rachis node (Zohary et al., 2013). The implication is that the products of the wild type allele at each of the Btr orthologs all act similarly to form a disarticulation layer above the rachis node to produce a wedge-shaped rachis internode. This form of brittle rachis is produced by species belonging to about half of the Triticeae genera (specifically, the genera Hordeum, Triticum, Secale, Psathyrostachys, Heteranthelium, Crithopsis, Taeniatherum, Thinopyrum, Dasypyrum, Eremopyrum, and Sitanion, along with a few Aegilops species; Sakuma et al., 2011). The extrapolation that allelic variation at the Btr1 and/or Btr2 orthologs is universally responsible for the brittle rachis in these taxa will need to be verified by further research.

The non-brittle einkorn wheat originated by a single nucleotide change at Btr1 that caused an amino acid substitution. How the single amino acid substitution (A119T) can change the specificity of BTR1 protein is an open question. This mutation was selected as a result of cultivation by early farmers. Today the nearest wild relatives of non-brittle einkorn have been found growing only in the northern Levant suggesting the latter may have originated there too along with non-brittle barley. The work described here provides a foundation for future studies on spike disarticulation of wheat and other members of Triticeae tribe.

TKa selected and provided the Tb and Tm materials. MP and SS assessed rachis type and performed QTL analysis. MP constructed the physical contig, which was sequenced by HK, TM, and annotated by MP. FD and MP carried out haplotype analysis. AD provided the tetraploid sequence of btr1 and btr2 orthologs. GW provided the archaeobotanical data. MP and TKo designed the experiments. MP, AD, BK, GW, and TKo wrote the manuscript. All authors have reviewed and commented on the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.