5.1 General Description of the Subtribe

Whereas the economically important cereals, Triticum L., Secale L., Hordeum L., and Aegilops L., have been subjected to intensive taxonomic, cytogenetic, molecular, and evolutionary studies, several other Triticeae genera received less attention. These “orphan” genera are Agropyron Gaertner, Eremopyrum (Ledeb.) Jaub. & Spach, Henrardia C. E. Hubbard, Dasypyrum (Coss. & Dur.) Dur., Heteranthelium Jaub. & Spach., Taeniatherum Nevski, and Crithopsis Jaub. & Spach. Several diploid Elymus species, having the St and E genomes, are also included in this group. They are small genera containing few species, most of which diploids that are characterized by a distinct morphology, and grow in different regions of the tribe distribution area, some in more arid environments. Their genetic relationships to the other well-studied genera of the tribe are vaguely known. Studies of these small genera may provide additional knowhow on the range of genetic diversity in the tribe, on processes that have led to diverge evolutionary developments as well as on the phylogenetic relationships among members of the tribe. As relatives of the crops, species of these orphan genera may contain valuable genes that, may be transferred to the crops and enhance greater tolerance to biotic and abiotic stresses, improve quality and performance. As such, these orphan genera deserve greater attention.

A number of interspecific and intergeneric hybrids involving those genera were produced during the years (e.g., Cauderon 1966; Sakamoto 1967, 1968, 1969, 1972, 1973, 1974, 1979; Sakamoto and Muramatsu 1963; Dewey 1969, 1970, 1984; Frederiksen 1991a, b, 1993; Frederiksen and von Bothmer 1986, 1989, 1995). Successful production of the hybrids suggests fairly good genetic or cytoplasmic compatibility among those species. However, there is very little chromosome pairing in F1 hybrids between them as well as between wheat and these species, indicating limited homology between their genomes. Hence, these genera seem to be highly differentiated from the taxonomic and the genetic viewpoints.

5.2 Elymus Species with St or E Genome

5.2.1 Group Description

The delimitation of the genera Elymus L. and Agropyron Gaertner has been the subject of controversy over the years (Assadi and Runemark 1995), primarily due to the absence of clear-cut generic characters and from the presence of numerous intergeneric hybrids that gave rise to conflicting results as discussed below (Melderis 1978). Their delimitation using different taxonomic treatments was changed several times over the last eight decades. Nevski (1933), Bor (1968), Tzvelev (1976), and Sakamoto (1974) kept these two genera separated, while Gould (1947) and Runemark and Heneen (1968), assuming that the traditional subdivision into Elymus s. l. and Agropyron s. l. (including Pseudoroegneria, Elytrigia, and Thinopyrum) is artificial, united them into a single genus, Elymus.

The species within the Elymus-Agropyron group have traditionally been referred to as Agropyron, if the spikelets are solitary, and as Elymus, if they are arranged in pairs or larger numbers at each rachis node. Yet, this division is not very distinct and Runemark and Heneen (1968) and Melderis (1978) pointed out that the number of spikelets at each node has a limited taxonomic value since several Agropyron species contain pairs of spikelets at several rachis nodes, especially in the lower or in the middle part of the spike, while several species of Elymus contain only one spikelet on each rachis node. Also, with regard to leaf anatomy, no difference was found between Elymus and Elytrigia that was included in Agropyron (Runemark and Heneen 1968). Dewey (1969, 1970) found homology between the genomes of several Elymus and Elytrigia species and Runemark and Heneen (1968) noted similar chromosome morphology in Agropyron elongatum (now Elymus elongatus) and Elymus caninus. In reality, the two genera only represent different levels in the reduction of a paniculate inflorescence (Runemark and Heneen 1968). Cytogenetic studies (e.g., Cauderon 1966; Sakamoto 1973; Dewey 1984, and reference therein) contributed to a better understanding of the genomic relationships among species of Elymus and Agropyron and, as a result, to modification of the delimitation of the species in these two genera. Based on the above information, as well as on the absence of morphological discontinuities between the taxa Pseudoroegneria, Elytrigia, Thinopyrum, and Elymus, Melderis (1978, 1980) included these taxa in the genus Elymus s. l., while retaining Agropyron s. str. as a separate genus for the crested wheatgrasses, that contains only species with a solitary spikelet at each rachis node. The restricted Agropyron genus contains diploid and polyploid species that are based on the P genome (Table 2.4) and are morphologically distinct from other genera in Triticeae. The genus Elymus is treated by Melderis (1978, 1980, 1985a, b) in a broad sense, as comprising the genera Elytrigia Desv., Pseudoroegneria (Nevski) A. Löve, Thinopyrum A. Löve, Lophopyrum A. Löve, and Trichopyrum A. Löve (Table 2.2). Wang (1989) supported Melderis’ classification which viewed separation of Pseudoroegneria from Elymus as unjustified for both evolutionary and morphological reasons, since several Elymus species include St genomes from different Pseudoroegneria diploids, and Pseudoroegneria and Elymus can hardly be distinguished from each other. This taxonomic classification makes the discrimination between these two genera more straightforward and has been accepted by various taxonomists, e.g., Clayton and Renvoize (1986), Assadi and Runemark (1995), Watson and Dallwitz (1992), Watson et al. (1985). Therefore, this book follows the Melderis classification.

A large number of hybrids within and between the Elymus s. l. and Agropyron s. str. genera have spontaneously emerged in nature. Many hybrids are sterile, but a considerable number are more or less fertile, at least upon backcross to one of the parents. Apparently, introgressive hybridization has played an important role in the evolution of these two genera.

Melderis (1978, 1980) transferred the following two sections from Agropyron s. l. to Elymus s. l.: Caespitosae (Rouy) Melderis, comb. nov. (Syn.: Agropyron sect. Caespitosa Rouy; Elytrigia sect. Caespitosae (Rouy) Tzvelev) and Junceae (Prat) Melderis, comb. nov. [Syn.: Agropyron sect. Junceae (Prat) Tzvelev]. The subdivision of these two sections was mainly based on caespitose or rhizomatous habit. The constituent species contain the St, Ee (=E) and the Eb (=J) genomes (Table 5.1), three genomes that occur in species that were included by Nevski’s classification (1933), in the genus Elytrigia, but, later on, Nevski himself (1934a, b) included Elytrigia as a section in Agropyron. Tzvelev (1973) maintained the generic status of Elytrigia, but pointed out that species of this genus are close to species of Elymus.

Table 5.1 Species of Elymus having the St or E genomes

The Elymus L. section Caespitosae is characterized by caespitose plant, lax and erect spikes, tough rachis, solitary or sometimes two spikelets on the node at the lower part of the spike, usually with 6–13 florets, and glumes 5–8 mm long, unkeeled with 5–9 veins. In some species, the rachilla is fragile and disarticulates above the glumes and beneath each floret (floret-type disarticulation). This section contains about 13 species (Table 5.1) comprising a polyploid series (2n = 14, 28, 56, and 70), and most of them are allogamous, and have long anthers.

The Elymus L. section Junceae is characterized by rhizomatous or caespitose plants, lax, erect, and sometimes curved spikes, a fragile rachis, disarticulating at maturity into spikelets with the rachis segment below them (Fig. 2.3; wedge type dispersal unit), solitary spikelets on the rachis, with 2–9 florets and keeled glumes, 5–18 mm long. This section comprises only three species, E. farctus (Viv.) Runemark ex Melderis, occurring in Europe and the Middle East, E. curvifolius (Lange) Melderis, occurring in south and central Spain, and E. distichus (Thumb.) Melderis, native to South Africa. Elymus farctus and E. distichus grow on maritime coasts.

The species of sections Caespitosa and Junceae presumably originate in Europe or west Asia. Genome St is found in several diploid and polyploid species, Ee exists in the diploid taxon E. elongatus (Host) Runemark subsp. elongatus [formerly Agropyron elongatum (Host) Beauv.] and in several auto- and allo- polyploids, whereas genome Eb occurs in the diploid taxon E. farctus (Viv.) Runemark ex Melderis subsp. bessarabicus (Savul. & Rayss) Melderis [formerly Agropyron junceum subsp. bessarabicum Savul. & Rayss; Thinopyrum bessarabicum (Savul. & Rayyss) A. Löve], in the allopolyploids of this species and of E. distichus (Thunb.) Melderis and in the autopolyploid E. curvifolius (Lange) Melderis (Table 5.1).

Löve (1984) used the genome symbols J for the genome of Elymus farctus and E for that of Elymus elongatus. Endo and Gill (1984) questioned the equivalence of J and E and based on differences in C-banding patterns, justified the separation of these two genomes. However, Dewey (1984) and Dvorak et al. (1984b), based on evidence from karyotype and genome analyses, considered the J and E genomes as the same basic genome. Previous studies of chromosome pairing in hybrids carrying these two genomes had already shown that they are closely related (Cauderon and Saigne 1961; Heneen and Runemark 1972; Dvorak 1981a; McGuire 1984) and more recent studies supported Dewey’s consideration (Wang 1985b; Wang and Hsiao 1989). However, Jauhar (1988) reached a different conclusion by studying chromosome pairing in the hybrids analyzed by Wang (1985b) and Wang and Hsiao (1989). Since a recent literature review indicated that most studies regarded J and E genomes as members of the same cluster (see Table 1 in Wang and Lu 2014), it is now generally accepted to regard them as very closely related genomes, supporting the use of a common basic genome symbol, E (Seberg and Frederiksen 2001; Yen et al. 2005; Fan et al. 2007; Liu et al. 2008; Sha et al. 2010; Yan et al. 2011; Wang and Lu 2014). Thus, these genomes were designated Ee for Elymus elongatus and Eb for Elymus farctus, respectively, as proposed by Dvorak (1981a) and McGuire (1984).

While the use of one basic genome symbol for these two species was rejected by some researchers (Jauhar 1988, 1990a, b; Jarvie and Barkworth 1992; Jauhar et al. 2004), several studies using different methodologies, have further confirmed the close relationship between genomes Ee and Eb. The studies included chromosome pairing (de V Pienaar et al. 1988; Forster and Miller 1989; Wang and Hsiao 1989), random amplified polymorphic DNA (RAPD) and sequence-tagged site (STS) markers (Wei and Wang 1995; Li et al. 2007), genomic in situ hybridization (GISH) (Kosina and Heslop-Harrison 1996; Chen et al. 1998a, b, 2003), chloroplast DNA sequences (Mason-Gamer et al. 2002; Liu et al. 2008), sequences of a gene encoding plastid acetyl-CoA carboxylase (Fan et al. 2007, 2009), and nuclear rDNA internal transcribed spacer (ITS) sequences (Hsiao et al. 1995; Liu et al. 2008; Yu et al. 2008). At the second International Triticeae Symposium, the Genome Designation Committee (Wang et al. 1995) adopted a system for the application of nuclear genome symbols in the tribe Triticeae. This system is based mainly on prevailing symbols, but since the number of basic nuclear genomes in the Triticeae exceeds the number of single letters in the Roman alphabet, some basic genomes are designated with an uppercase letter followed by a lowercase letter, e.g., Ee or Eb, for the genome in Elymus elongatus and E. farctus, respectively. An uppercase letter followed by a superscript in small letters are used when modified versions of a basic genome is referred to, e.g., Am for the genome found in Triticum monococcum.

Melderis transferred the Asiatic diploid species of Elymus, namely, libanoticus, reflexiaristatus (subsp. reflexiaristatus and subsp. strigosus), and the diploid cytotypes of stipifolius and tauri from Agropyron (=Elytrigia) to Elymus (1978, 1980). These species carry the St genome (Wang et al. 1995) that also exists in the Elymus polyploid species containing the Ee genome, i.e., elongatus, nodosum, bungeanus, and hispidus (Table 5.1). The St genome is related to the Ee and Eb genomes (Wang 1989). Bieniek et al. (2015) found that nucleotide sequences of the diploid Ee, Eb and St taxa are almost identical, with only one substitution within the matK gene, differentiating genome Eb from the Ee and St genomes. Petersen and Seberg (1997) and Wang and Lu (2014) confirmed this very close relationship between the three genomes. The St genome almost always has a dominant influence on the morphology of the taxa of which it is a component (Assadi and Runemark 1995) and since exists in more primitive Elymus species, it is reasonable to assume that Ee and Eb evolved from St.

5.2.2 Elymus Species with St Genome

5.2.2.1 Species Description

The St-genome species of Elymus were previously recognized as a biological unit and placed as a separate section, Elytrigia, in the traditional Agropyron s. l. (Nevski 1934a). Due to the fact that all these species contain one genome, Love (1980), treated them as a separate genus, Pseudoroegneria. However, due to the absence of morphological discontinuities between Pseudoroegneria and Elymus, Melderis (1978, 1980) included Pseudoroegneria in the genus Elymus s. l.

The Elymus species bearing the St genome include approximately 15 different taxa that consists of about equal numbers of diploids and tetraploids. The type species of this group is E. reflexiaristatus (Nevski) Melderis subsp. strigosus (M. Bieb.) Melderis [formerly Pseudoroegneria strigosa (M. Bieb.) Á. Löve] (Dewey 1984; Lӧve 1984; Watson and Dallwitz 1992; Yan and Sun 2011). Interspecific hybrids between the St diploid species exhibit almost complete chromosome pairing at fist meiotic metaphase, but with high or complete sterility, indicating divergence of the same basic genome in each diploid (Stebbins and Pun 1953; Dewey 1975). Some of the species, e.g., stipifolius, and spicatus, have diploid and tetraploid cytotypes and the tetraploids behave cytologically as autotetraploids or near autoploids (Dewey 1975). The tetraploid taxa of two other species, tauri and panormitanus, are allopolyploids containing the St and P subgenomes.

A large amount of the allopolyploid species of Elymus s.l. share a common St genome with diploid Elymus species in different combinations with H, Y, P, and W subgenomes (Table 2.4). The maximum likelihood tree constructed, using nuclear ribosomal internal transcribed spacer region (nrITS) data, showed that diploid Elymus, Hordeum and Agropyron species served as the St, H and P subgenomes donors, respectively, for the Elymus allopolyploids (Dong et al. 2015). The maximum likelihood tree for the chloroplast genes (matK and the intergenic region of trnH-psbA) suggests that the Elymus diploid donors of the St genome to Elymus allopolyploids served, in most cases, as the maternal donor. Moreover, the chloroplast genes data suggest that diploid St Elymus species from Central Asia and Europe are more ancient than those in North America (Dong et al. 2015). Thus, it was hypothesized that the Elymus s. l. species originated in Central Asia and Europe, and then spread to North America.

The St genome species are perennials, caespitose, and cross-pollinating, with culms between 30 and 90 cm tall, narrow, linear spikes with single, distantly spaced spikelets, 5–8 mm long glumes of equal length in E. reflexiaristatus, or unequal in E. spicatus, tauri and libanoticus, 8–30 mm long glume awns, absent in E. tauri and libanoticus, and long anthers. These species grow in the northern Hemisphere, from southwestern and southeastern Europe, the Middle East, Transcaucasia across Central Asia and Northern China to Western North America (Dewey 1984). They occur on open rocky hillsides, are exceptionally drought and salt tolerant and have excellent quality forage that is palatable to animals (Dewey 1984).

The Elymus libanoticus-related species, that exemplify all St genome diploid species, are described below.

5.2.2.2 Elymus libanoticus (Hack.) Melderis—A Representative Example

5.2.2.2.1 Morphological and Geographical Notes

Elymus libanoticus (Hackel) Melderis [Synonym: Agropyron libanoticum Hackel; Pseudoroegneria libanotica (Hackel) D.R. Dewey; Elytrigia libanotica (Hackel) Holub); Pseudoroegneria tauri ssp. libanotica (Hackel) Á. Löve; Agropyron sosnovskyi Hackel; Elytrigia sosnovskyi (Hackel) Nevski; Elymus sosnovskyi (Hack.) Melderis; Pseudoroegneria sosnovskyi (Hackel) A. Love; Agropyron gracillimum Nevski; Elytrigia gracillima (Nevski) Nevski; Pseudoroegneria gracillima (Nevski) Á. Löve], is perennial, caespitose, with short rhizomes, 45–85 cm high culms, 5–15 cm long linear spikes, with 4–7 spikelets, each 10–15 mm long, one per node, tough rachis, 3–6 florets, unequal, lanceolate, 3–5-veined glumes, the lower ones 6–8 mm long, typically 3/4 or nearly as long as lower floret, and upper ones 7–9 mm long, 8–9 mm long, lanceolate, 3-veined, unawned, lemma, palea shorter than lemma, sparsely ciliate on keels, 4–5 mm long anthers, and caryopsis adherent to palea and lemma. Chromosome number 2n = 2x = 14 (Dewey 1972) (Fig. 5.1a).

Fig. 5.1
figure 1

Mediterranean Elymus species; a A plant and spikes of E. libanoticus (Hack.) Melderis (2n = 2x = 14), (Photographed by the late Prof. Avinoam Danin); b A plant and spike of E. elongatus (Host) Runemark ssp. elongatus (2n = 2x = 14); c A plant and spike of E. farctus (Viv.) Runemark ex Melderis ssp. farctus Runemark ex Melderis (2n = 6x = 42)

Unlike E. libanoticus, E. sosnovskyi (Hack.) Melderis [=Agropyron sosnovskyi Hack.: Elytrigia sosnovskyi (Hack.) Nevski] bears acuminate glumes with 3 veins. Agropyron gracillimum Nevski differs from E. libanoticus by their smaller leaf thickness. These differences, however, fall within the variation of the Iranian E. libanoticus material (Assadi 1996). Moreover, hybrids between E. sosnovskyi or A. gracillimum with E. libanoticus were highly fertile, with regular meiotic metaphase. Therefore, the three names are considered synonymous (Assadi 1996). E. libanoticus is closely related to E. tauri subsp. libanoticus, differing only in several morphological traits (Assadi 1996).

Elymus libanoticus grows in Lebanon, Syria, northern Israel, south and southeastern Anatolia, northern Iraq, Iran, and Caucasus. It thrives on dry mountain slopes and limestone ravines, usually on more xeric habitats, 1000–3050 m a.s.l. It is an Irano-Turanian element.

5.2.2.2.2 Cytology, Cytogenetics and Evolution

Hsiao et al. (1986) analyzed the karyotype of diploid St genome species, including Elymus spicatus, E. Reflexiaristatus subsp. Strigosus, E. libanoticus, and E. stipifolius. All four species possess similar karyotypes and chromosomal lengths. The karyotypes of all species have one pair of small and one pair of large satellites on the short arms of chromosomes 2 and 5, respectively (Hsiao et al. 1986). The karyotypes are symmetric; most chromosomes are metacentric and a few are sub-metacentric (Wang et al. 1985; Hsiao et al. 1986; Deng et al. 2004). The St genome consists of smaller chromosomes than those of the R, P, and Eb genomes. Despite their wide geographical distribution, the karyotype patterns of the St genome species have not been dramatically altered. The karyotype of E. spicatus has been reported previously (Schulz-Schaeffer and Jurasits 1962; Dvorak et al. 1984a, b).

Endo and Gill (1984), using the acetocarmine-Giemsa C-banding technique, studied heterochromatin distribution in somatic chromosomes of diploid Elymus and Agropyron species. With the exception of E. elongatus, which is moderately self-fertile, all other species are cross-pollinating and self-sterile. The cross-pollinating species showed large terminal C-bands and a high level of C-band polymorphism, whereas E. elongatus showed small terminal and interstitial bands and a minimal C-band polymorphism. C-banding patterns show that the Eb genome of diploid E. farctus appears to be distinct from the Ee genome of diploid E. elongatus and may constitute an intermediate link between the Ee and St genomes (Endo and Gill 1984).

E. spicatus, E. libanoticus, and E. stipifolius have similar C-band patterns, although C-bands were less prominent in E. stipifolius than in the others. Thus, the C-banding patterns and morphology of satellite chromosomes supported previous evidence that E. spicatus, E. libanoticus, and E. stipifolius share a common St genome. Variation in the intensity of terminal C-bands was observed in E. stipifolius, which is to be expected in a basic genome of species with worldwide distribution (Dewey 1981).

Wang (1989) produced the tetraploid hybrid (genome StStStH) from crossing the hexaploid Elymus transhyrcanus (genome StStStStHH) with E. libanoticus (genome StSt). This F1 hybrid exhibited at first meiotic metaphase 13.94 univalents, 0.16 rod and 6.78 ring bivalents (6.94 total bivalents) and 0.06 trivalents. The reciprocal hybrid showed an average of 10.22 univalents, 2.34 rod and 5.24 ring bivalents (7.58 total bivalents), 0.74 trivalents and 0.10 quadrivalents. The amount of pairing in the hybrid and particularly that of trivalent cnfiguraton was much less than expected in the case of three fully homologous St genomes. Hence, either the two St sugenomes of the hexaploid had diverged from one another or both had diverged from the St genome of the diploid.

In the F1 tetraploid hybrid (genome StStEeEe) between Elymus libanoticus (genome StSt) and E. hispidus (=Thinopyrum intermedium; genome StStEeEeEeEe) Wang (1989) observed at first meiotic metaphase an average of 6.68 univalents, 4.96 rod and 3.66 ring bivalents (8.62 total bivalents), 1.06 trivalents, 0.20 quadrivalents and 0.03 pentavalents. These data show that, in addition to the autosyndetic pairing in the form of bivalents between the Ee subgenomes, the presence of multivalents indicates some allosyndetic pairing between St and Ee chromosomes, indicating that the two genomes are related.

5.2.2.2.3 Crosses with Other Triticineae Species

Studies of meiotic chromosome pairing in F1 hybrids between diploid Agropyron cristatum (genome PP) and several different diploids species of Elymus with St genome (genome of all hybrids was PSt), showed that the two genomes, P and St, are related (Wang 1985a, 1986, 1987a, b, 1988, 1989, 1990, 1992; Wang et al. 1985). Size differences between Agropyron (large) and St genome Elymus (small) chromosomes facilitated interpretation of chromosome pairing in the F1 hybrids. The average chromosome pairing at first meiotic metaphase of the diploid hybrid A. cristatum x E. libanoticus included 7.71 univalents, 2.77 bivalents, 0.22 trivalents, 0.01 quadrivalents and 0.01 pentavalents (Wang 1986), while that between A. cristatum and E. stipifolius displayed a similar amount and pattern of pairing (Wang 1985a). These pairing data indicate allosyndetic pairing between the homoeologous chromosomes of the two genomes, demonstrating a close relation between the St and the P genomes.

Meiotic chromosome pairing in the F1 hybrid Elymus spicatus (genome StSt) x Secale strictum (genome RR) exhibited an average of 12.97 univalents, 0.49 bivalents and 0.01 trivalent (Wang 1987b). The F1 hybrid Agropyron. mongolicum x S. strictum, which had the PR genome, showed an average of 12.86 univalents, 0.51 bivalents, 0.03 trivalents and 0.004 quadrivalents. The hybrid between Elymus spicatus and A. mongolicum (genome StP) had a mean configuration of 8.05 univalents, 2.86 bivalents, 0.07 trivalents and 0.01 quadrivalents. All hybrids were sterile. The meiotic pairings of these hybrids indicated that chromosome homology between the St and P genomes is higher than between St and R and between P and R. The degree of meiotic pairing in the E. spicatus x A. mongoicum hybrid was similar to those in other diploid hybrids bearing the same genome constitution, i.e., A. cristatum x E. stipifolius and A. cristatum x E. libanoticus (Wang et al. 1985; Wang 1986).

Following hybridization of the diploid Elymus Stipifolius (genome StSt) with tetraploid Elymus elongatus (genome EeEeEeEe), Dvorak (1981a) obtained a triploid hybrid (genome StEeEe), that exhibited 7.8 univalents, 5.9 bivalents and 0.41 trivalents at first meiotic metaphase. This pattern of pairing was attributed primarily to autosyndesis between homologous chromosomes of the Ee genomes. Stebbins and Pun (1953) had speculated that the Ee and St genomes might be variations of the same basic genome, yet the hypothesis was contradicted by Dvorak’s (1981a) data, which showed that the Ee and St genomes are distinctly different.

Wang (1989) crossed Elymus libanoticus (genome StSt) with the tetraploid cytotype of Agropyron cristatum (genome PPPP) and observed an average meiotic pairing profile of 11.30 univalents, 3.40 rod and 1.50 ring bivalents (4.90 total bivalents) in the resulting triploid hybrid. Most pairing in this triploid hybrid was autosyndetic, indicating a difficulty in learning about the relationship between two genomes when one of the genomes exists in two doses.

Interpretation of chromosome pairing in St-Elymus and Agropyron hybrids is aided by size differences between the Agropyron (large) and St-Elymus (small) chromosomes. Chromosomes of autotetraploid Elymus spicatus (genome StStStSt) paired only rarely with chromosomes of diploid Agropyron cristatum (genome PP) in their triploid hybrids (StStP) (Dewey 1964). In the tetraploid hybrids (PPStSt) of A. desertorum (genome PPPP) and tetraploid E. spicattus (genome StStStSt), all chromosome pairing was attributed to autosyndesis between the PP and StSt genomes (Dewey 1967). Wang et al. (1985) crossed Agropyron desertorum (2n = 4x = 28; genome PPPP) with the autotetraploid cytotype of Elymus stipifolius (2n = 4x = 28; genome StStStSt). The tetraploid hybrid averaged 3.09 bivalents, most of which resulted from autosyndetic pairing between the P or the St genomes (Wang et al. 1985). In this tetraploid hybrid, because of the presence of homologous chromosomes, the P genome chromosomes rarely paired with the St genome chromosomes.

Wang (1989) used the mean C-values (the ratio between the number of chiasmata and the number of chromosome arms) to assess the relationships between genomes in diploid hybrids of the perennial Triticeae. He found that a C-value of 0.55 in diploid hybrids can serve as a critical value (in conjunction with other evidence, e.g., karyotype characteristics) to separate intergenomic from intragenomic divergence. Using this rule, he found that the Secale R genome, the Hordeum H genome and the Psathyrostachys N genome are distinct from each other and from other Triticeae genomes (C-values 0.03–0.17), while the St, Ee, Eb, and P genomes show considerable homoeology (C-values 0.24–0.36). Thus, the Eb and St genomes, despite considerable differences in total genome size (Hsiao et al. 1986), show considerable homoeology, with a mean C-value of 0.35 in the diploid hybrid between them (Wang 1989). Similar homoeology was recorded by Liu and Wang (1993) in the triploid hybrids (genomes StStEb and StEeEe).

5.2.3 E. elongatus (Host) Runemark (Based on Ee Genome)

5.2.3.1 Species Description

Elymus elongatus [syn. Triticum elongatum Host; Agropyron elongatum (Host) Beauv.; Agropyron elongatum subsp. scirpeum; Elytrigia elongata (Host) Nevski; Lophopyrum elongatum (Host) Á. Löve; Thinopyrum elongatum (Host) D. R. Dewey] is a perennial, caespitose, 30–100 cm high, with robust, glabrous culms, 10–25 cm long lax and erect spikes, tough rachis, 10–25 mm long solitary spikelets on each rachis node, sometimes two spikelets on one node, with 6–13 (9–25) awnless florets, glumes shorter than spikelet, 6–8 mm, 5–9 veined, without keels, 7–10 mm long, keeled lemma, keeled palea, 4–4.5 mm long anthers, caryopsis with adherent pericarp. The rachilla is fragile and disarticulates above the glumes and beneath each floret (floret-type disarticulation). This type of seed dispersal is characteristic of the Arctic-Temperate group and especially of species of Elymus.

The cytotaxonomy of E. elongatus was studied by several researchers, e.g., Peto (1930), Simonet (1935), Cauderon (1958, 1966), Schulz-Schaeffer and Jurasits (1962), Schulz-Schaeffer et al. (1971), Evans (1962), Runemark and Heneen (1968) Heneen (1972), Heneen and Runemark (1972, 1977), and Luria (1983), who showed that E. elongatus comprises a polyploid complex of diploid, tetraploid, octoploid, and decaploid taxa (Table 5.1). The diploid and decaploid taxa are well documented in the literature and evidence desmonstrates that the autotetraploid taxon also belongs to this group (Heneen and Runemark 1972). Hexaploid chromosome number was also found in material collected from Istria (Heneen and Runemark unpubl.). Schulz-Schaeffer and Jura (1967) reported the existence of hexaploid types in plants collected from Turkey. However, this hexaploid was not recognized as a valid subspecies. In addition to the diploid and tetraploid subspecies, an octoploid subsp. of Elongatus, subsp. turcicus (P. E. McGuire) Melderis, from Turkey, was described (McGuire 1984). Thus, it appears that the Elymus elongatus complex is represented in nature by types that form a complete polyploid series, ranging from diploids to decaploids.

Several authors described variants of this species as separate species or as subordinate taxa. Since the morphological differences between these taxa are not clear, Melderis (1980) recognized only two taxa that merit the subspecies status, namely, subsp. elongatus, a diploid, and subsp. ponticus, a decaploid. Later, Melderis (1985a) recognized an additional subspecies, subsp. turcicus (P. E. McGuire) Melderis, an octoploid taxon. Heneen and Runemark (1972) included an autotetraploid taxon from Cyprus and the Aegean islands in E. elongatus as ssp. flaccidifolius. Breton-Sintes and Cauderon (1978) classified an accession from Sicily of Heneen and Runemark (1972) autotetraploid subspecies as Agropyron elongatum (Host) ssp. scirpeum (C. Presl.) Cifferi et Giacom. The taxon ssp. flaccidifolius was elevated by Melderis (1978) to the specific rank Elymus flaccidifolius (Boiss. & Heldr.) Melderis. However, since there are only minor morphological differences (mainly quantitative) between this species and other autotetraploids of E. elongatus, it is more appropriate to classify it, along with all the other autoetraploids of E. elongatus, as a subspecies of elongatus. Hence, in this book, all the autotetraploid taxa of E. elongatus (=Agropyron elongatum var. flaccidifolium Boiss. & Heldr.; Agropyron flaccidifolium (Boiss. & Heldr.) Candargy; Elymus flaccidifolius (Boiss. & Heldr.) Melderis; Agropyron elongatum Host subsp. scirpeum (C. presl.) Ciferri & Giacom.; Lophopyrum scirpeum (C. Presl) Á. Löve; Thinopyrum scirpeum (Presl) D. R. Dewey; Agropyron scirpeum C. Presl; Elytrigia scirpea (C. Presl) Holub) are referred to as Elymus elongatus (Host) Runemark ssp. flaccidiffolius (Boiss. & Heldr.) Runemark, and were grouped together as one subspecies.

The diploid and the tetraploid subspecies of E. elongatus exhibit wide morphological variation in the number of spikelets per spike, number of florets per spikelet, hairiness, and plant color. The decaploid subspecies, subsp. ponticus, exhibits wider variation than subsp. elongatus and flaccidifolius.

E. elongatus is found in all parts of the Mediterranean basin, in southwestern, southeastern, and eastern Europe, in North Africa, and the Middle East, Caucasus, western Asia, and Arabia. It was introduced or invaded Australasia, South America, and North America. It is a Mediterranean element (chorotype) and grows among Mediterranean plant communities. This species grows in salt marshes and near salty springs and is salt tolerant (Moxley et al. 1978; Dewey 1960; McGuire and Dvorak 1980).

5.2.3.2 Ssp. elongatus (2n = 2x = 14)

5.2.3.2.1 Morphological and Geographical Notes

E. elongatus (Host) Runemark subsp. elongatus [Agropyron elongatum (Host) Beauv.; Lophopyrum elongatum (Host) Á. Löve; Thinopyrum elongatum (Host) D.R. Dewey] is a diploid subspecies with tall stems (50–80 cm high); 10–25 cm long spikes, with 9–26 spikelets per spike, internodes at the base of the spike as long as the spikelets; 10–17 mm long spikelets with 7–8 florets, usually one spikelet at each rachis node, 7–10 mm long glumes with 5–9 veins, where the lower glume is shorter (about 2/3–3/4) than the lower floret, 9–10 mm long lemmas, 4–4.5 mm long anthers and 4 mm long caryopsis (Fig. 5.1b).

Subsp. elongatus grows in the Mediterranean basin. In Israel, the diploid taxon grows in the Coastal Plain from the Shfela (Einot Gibton) and northwards (Acre plain). It grows in salt marshes, near salty springs and on maritime sands, from sea level to 100 m a.s.l., throughout the range of the species. These saltmarsh habitats are characterized by high underground water that forms floods in the winter and salty soil with salty crust in the summer. The subspecies also grows on sandy soils near river mouths, on silt near river’s banks or springs or on clay soil. When growing in wet soils, the amount of annual rainfall is not a limiting factor.

5.2.3.2.2 Cytology, Cytogenetics and Evolution

Matsumura and Sakamoto (1956), Cauderon (1958), Evans (1962), Schulz-Schaeffer and Jurasits (1962), Runemark and Heneen (1968), Heneen (1972) and Luria (1983) described the karyotype of the diploid subspecies. The karyotype is symmetric, consisting of four metacentric pairs and three sub-metacentric pairs. The differences in length and arm ratio among the chromosomes of this subspecies are relatively small (Dvorak and Knott 1974). However, the homologous chromosomes can be visually identified (Evans 1962). Two chromosome pairs have satellites, with one metacentric pair carrying a large satellite and one sub-metacentric pair bearing a small satellite. The constrictions between the satellites and the chromosome arms carrying them are the nucleolar organizing regions (NORs). The NORs contain a set of argyrophilic proteins which are selectively stained by silver. After silver staining, the NORs can be easily identified as black dots that are called Ag-NORs. Thus, in agreement with the number of satellite (SAT)-chromosomes mentioned above, four Ag-NORs are regularly observed in somatic cells of diploid E. elongatus (Lacadena et al. 1984). Giemsa C-banding analysis of the chromosomes of several accessions of this subspecies revealed small terminal and interstitial bands and a minimal C-band polymorphism (Endo and Gill 1984).

Heneen and Runemark (1972) observed karyotype differences between plants of ssp. elongatus collected from different locations, the major difference lying in the appearance of the SAT-chromosomes. Runemark and Heneen (1968), comparing the karyotype of subsp. elongatus with that of diploid E. farctu (genome EbEb), found that the two karyotypes resemble one another, but the chromosomes of E. elongatus are somewhat smaller than those of E. farctus. In addition, differences in morphology of the SAT-chromosomes exist between the two taxa (Heneen 1962); the pair with large satellites in E. elongatus has more median centromeres than the equivalent pair in E. farctus. The constriction in this SAT-chromosome divides the short arm in E. elongatus into two unequal parts, with the part proximal to the centromere being longer than the satellite, which is not the case in E. farctus.

Chromosomal pairing at meiosis in the diploid subspecies is regular (0.08–0.15 univalents and 6.92–6.95 bivalents per meiocyte; 12.04–12.86 chiasmata/cell (Cauderon 1958; Luria 1983). However, in accordance with the karyological observations, meiotic analysis of several inter-varietal hybrids showed the existence of structural heterozygosity among several accessions of this subspecies (Heneen and Runemark 1972). Similarly, in one inter-varietal cross, Luria (1983) observed a quadrivalent, suggesting the existence of a reciprocal translocation between the two accessions. These findings may indicate the occurrence of initial steps of karyotype divergence among and within accessions of the diploid subspecies of E. elongatus.

5.2.3.2.3 Crosses with Other Triticineae Species
5.2.3.2.3.1 Crosses with Diploids Species

Cross of Elymus farctus ssp. bessarabicus (genome EbEb), as female, with E. elongatus ssp. elongatus (genome EeEe) was successful, while the reciprocal cross failed (Wang 1985a). Karyotypes of mitotic chromosomes in the parental species revealed that three of the seven chromosomes in the Eb and Ee genomes were similar in length and arm ratio. Meiosis in the F1 hybrids substantiated this observation, but four chromosomes had undergone some structural rearrangements such as reciprocal translocations (Wang 1985b). Chromosomal pairing at meiotic first metaphase of the F1 hybrid averaged 2.68 univalents, 4.68 bivalents, 0.27 trivalents, 0.27 quadrivalents, and 0.01 pentavalents (Wang 1985b). The F1 hybrids were completely sterile upon self-pollination. From the relatively high pairing, Wang (1985a) concluded that the Eb and Ee genomes are closely related, supported the transfer of Lophopyrum elongatum to the genus Thinopyrum, as was suggested by Dewey (1984), as opposed to keeping them as two separate genera, as suggested by Löve (1984). GISH studies substantiated this conclusion by showing that genomes Ee and Eb are closely similar in their repetitive DNA (Kosina and Heslop-Harrison 1996).

Considering the suppression of pairing by the Ph1 gene that inhibits homoeologous pairing between the chromosomes of ssp. bessarabicus and ssp. elongatus in the tri-generic hybrid with durum wheat (genome ABEbEe), Jauhar (1988, 1990a, b) argued that the genomes Eb and Ee are homoeologues rather than homologues and should be assigned distinct genome symbols (J and E, respectively).

Jauhar et al. (2004) later analyzed chromosomal pairing in meiosis of the tri-generic hybrids between durum wheat, with and without the Ph1 gene, and the amphidiploid E. farctus ssp. bessarabicus-E. elongatus ssp. elongatus. Meiotic chromosome pairing was studied using both conventional staining and fluorescent genomic in situ hybridization (fl-GISH). As expected, the Ph1-intergeneric hybrids (genome ABEbEe) showed low chromosome pairing (23.86% of the total chromosome complement paired), whereas 49.49% of the trigeneric hybrids without Ph1 showed pairing. Fl-GISH analysis provided insight to the study of the specificity of chromosome pairing: wheat with Elymus (AB with Ee and/or Eb), wheat with wheat (A with B), or E. elongatus with E. farctus (Ee with Eb). The analysis revealed that without the Ph1 gene in the tri-generic hybrid, there were 3.97 chiasmata/cell between chromosomes of the Eb and Ee genomes, 2.29 chiasmata/cell between wheat chromosomes, and 2.6 chiasmata/cell between wheat–Elymus chromosomes. Thus, the two E genomes are more closely related to each other than A and B to one another.

Similarly, Forster and Miller (1989) reported that the chromosomes of ssp. bessarabicus and subsp. elongatus rarely paired in the presence of the Ph1 gene, i.e., in the hybrid between the two amphiploids Triticum aestivum-diploid E. farctus x T. aestivum–diploid E. elongatus. However, they concluded that, because of the relative high frequency of pairing between chromosomes of these two species at the diploid level, their genomes warrant a common genome symbol. Yet, since the two genomes do not pair in a wheat genetic background, their differentiation should also be indicated. Therefore, Forster and Miller (1989) proposed that the genome symbol of E. elongatus be E and that of ssp. besarabicus Eb, as suggested by Dvorak (1981a) and McGuire (1984).

Crossess with other Triticeae diploids revealed very little homology. Dvorak (1981b) succeeded in crossing Aegilops tauschii (=Ae. squarrosa) with ssp. elongatus, while crosses between ssp. elongatus and Ae. speltoides or Triticum monococcum ssp. aegilopoides (=T. boeoticum), were not successful. Mean chromosome pairing at first meiotic metaphase of the diploid hybrid yielded 10.7 univalents, 1.5 (0–5) bivalents, 0.027 trivalents per cell, indicating a certain degree of homoeology between the genomes of the two species (Dvorak 1981b). The F1 hybrid plants were sterile, with very low pollen fertility.

5.2.3.2.3.2 Crosses with Tetraploid Species

Cauderon (1958) and Cauderon and Saigne (1961) crossed the allotetraploid Elymus farctus ssp. boreo-atlanticus (=Agropyrum junceum boreo-atlanticum) (genome EbEbEeEe) with Elymus elongatus ssp. elongatus (genome EeEe) and studied chromosome pairing at first meiotic metaphase of the triploid F1 hybrid (genome EbEeEe). Meiotic pairing showed 3.40 univalents, 4.50 bivalents, 2.76 trivalents, and 0.08 quadrivalents. From the relatively high frequency of trivalents at the hybrid meiosis, they concluded that the Eb and Ee genomes are closely related. A similar conclusion was drawn following karyotype analysis (Cauderon 1958).

When diploid E. elongatus was crossed with tetraploid (durum) wheat (genome BBAA) (Jenkins and Mochizuki 1957; Mujeeb-Kazi and Rodriguez 1981), the F1 hybrids showed very little pairing (0.3–2.6 bivalents per cell), suggesting that in the presence of one dose of Phl, the Ee genome chromosomes of elongatus showed little, if any, pairing with those of the subgenomes A and B of durum wheat. However, the level of chromosomal pairing reported by Jenkins and Mochizuki (1957), i.e., 2.6 bivalents/cell in the hybrid durum wheat x diploid E. elongatus was significantly higher than expected on the basis of pairing in haploid durum wheat, i.e., 0.37 bivalents/cell, as reported by Kihara (1936) and Lacadena and Ramos (1968).

In a later study, Mochizuki (1960, 1962) studied chromosomal pairing between individual E. elongatus chromosomes and tetraploid wheat chromosomes in monosomic addition lines, where single elongatus chromosomes were added to the durum complement. No chromosome associations were observed between wheat and elongatus chromosomes in three lines, while a high frequency of trivalent associations was noted in the remaining four lines. From these results, he concluded that four elongatus chromosomes are partially homologous to durum chromosomes. However, Dvorak and Knott (1974) assumed that the trivalents resulted from translocations between the durum and elongatus chromosomes that occurred during the production of the monosomic addition lines and actually, in the presence of two doses of Ph1 of durum wheat, there was no pairing between the elongatus and the durum chromosomes. Following this controversy, Ono et al. (1983) re-examined Mochizuki (1962) durum-elongatus addition lines and found that apart from 5Ee, no elongatus chromosomes paired with wheat chromosomes. Evans (1962) found that the nucleolar of Elymus were suppressed in the amphiploid Triticum durum-diploid Elymus elongatus by the durum NORs.

5.2.3.2.3.3 Crosses with Hexaploid Species

The F1 hybrid between hexaploid wheat (BBAADD) and diploid E. elongatus (EeEe) exhibited very little chromosomal pairing at meiosis (Jenkins 1957). A low level of pairing between Ee genome chromosomes of diploid E. elongatus and those of common wheat was also observed in elongatus addition lines to common wheat (Dvorak and Knott 1974). Study of pairing of single diploid E. elongatus chromosomes with common wheat chromosomes in monosomic addition lines, in the presence of two doses of the homoeologous-pairing suppressor Ph1, showed that elongatus chromosomes do not pair with wheat chromosomes, with the exception of chromosome IV [assigned later to homoeologous group 3, and designated 3Ee by Dvorak (1980)], that very rarely paired with a wheat chromosome (Dvorak and Knott 1974). The researchers thus concluded that elongatus genomes did not play any role in the evolution of the polyploid series of Aegilops and Triticum.

When ten ditelosomic addition lines, comprising of diploid E. elongatus telosomes added to the common wheat complement, were crossed to Aegilops speltoides, that suppresses the activity of the Ph1 gene, all ten elongatus telosomes paired with common wheat chromosomes (Dvorak 1979). But, because this pairing only occurred when Ph1 was not active, Dvorak concluded that none of the ten elongatus-chromosome arms has a homologous partner among the three common wheat subgenomes A, B, and D, and the involved speltoides genome.

Likewise, in crosses between elongatus substitution lines, where the activity of Ph1 was suppressed, elongatus chromosome 6Ee paired, to some extent (4.6%), with wheat chromosomes of homoeologous group 6 (Dvorak 1979). Similarly, pairing between elongatus and wheat chromosomes was also observed by Johnson and Kimber (1967), Dvorak (1979, 1981b), and Sears (1973) in hybrids between elongatus and common wheat, when the Ph1 gene of wheat was suppressed or absent.

If the interpretation of these data, as well as that of Jenkins and Mochizuki (1957) on pairing in the hybrid T. durum x diploid E. elongatus, is correct, then there must be considerable homology between Elymus elongatus and common wheat chromosomes. However, Dvorak and Knott (1974) assumed that this degree of pairing does not result from chromosomal homology but, rather, from the presence of E. elongatus genes that promote homoeologous pairing. Indeed, Dvorak and Knott (1974) found that chromosome IV (designated later 3Ee by Dvorak 1980) and chromosome I (1Ee; Dvorak 1980) increased significantly the pairing of wheat chromosomes, i.e., they carry genes that promote pairing of homoeologous chromosomes. If this is the case, then, the pairing reported by Jenkins and Mochizuki (1957) between elongatus and durum wheat chromosomes, resulted presumably from homoeologous pairing between chromosomes of the A and B subgenomes of durum in addition to that between elongatus and durum.

Dvorak (1987) assumed that genes promoting or suppressing pairing of homoeologous chromosomes are ubiquitous among Triticeae diploid species. To identify such genes in diploid E. elongatus, he crossed common wheat lines with added or substituted E. elongatus chromosomes with Hordeum bulbosum to obtain haploids, and with Triticum urartu to obtain interspecific hybrids. Studies of chromosome pairing at first meiotic metaphase in the resulting haploids and hybrids and in the parental addition and substitution lines revealed genes affecting homologous or homoeologous chromosome pairing. Genes promoting pairing were found on the short and long arms of chromosome 3Ee, on the short arms of 4Ee and 5Ee, and on chromosome 6Ee of E. elongatus. Genes suppressing pairing of homoeologous chromosomes were found on the long arms of chromosomes 4Ee and 7Ee (Dvorak 1987). That may explain why different results were found when using lines that may contain different alleles of these pairing genes.

While eight Ag-NORs were observed in many cells of the amphiploid common wheat–diploid E. elongatus, four on the wheat chromosomes 1B and 6B and on four on two elongatus chromosomes, in some cells the Ag-NORs of elongatus were suppressed by the wheat chromosomes (Lacadena et al. 1984).

5.2.3.3 Ssp. flaccidifolius (Boiss. & Heldr.) Runemark (2n = 4x = 28)

5.2.3.3.1 Morphological and Geographical Notes

Ssp. flaccidifolius [Syn.: Boiss. & Heldr.) Runemaks [Syn.: Agropyron scirpeum C. Presl; Agropyron scirpeum var. flaccidifolium Boiss. & Heldr.; Agropyron elongatum var. flaccidifolium (Boiss. & Heldr.) Boiss. & Heldr.; Agropyron flaccidifolium (Boiss. & Heldr.) Candargy; Agropyron elongatum Host ssp. scirpeum (C. Presl.) Ciferri & Giacom.; Elymus elongatus ssp. flaccidifolius (Boiss. a Heldr.) Runemark; Elytrigia scirpea (C. Presl) Holub; Lophopyrum scirpeum (C. Presl) Á. Löve; Thinopyrum scirpeum (C. Presl) D. R. Dewey} is a perennial caespitose, more or less glaucous grass with erect culms, 70–115 cm high, with 11–23 cm long spikes, with 5–17 spikelets per spike, 10–22 mm long spikelets, with 5–10 florets and glumes with 5–6 veins, 4–6 mm long anther and 5 mm long caryopsis. In several accessions, two spikelets are located at each rachis node in the lower part or the center of the spike.

The diploid and tetraploid subspecies of E. elongatus cannot be morphologically distinguished with certainty, primarily because they differ mainly in quantitative traits (Breton-Sintes and Cauderon 1978; Luria 1983). Luria (1983) found several tetraploid accessions of E. elongatus ssp. flaccidifolius in Israel, in addition to diploid accessions of ssp. elongatus. The tetraploid accessions morphologically resemble the tetraploid subsp flaccidifolius (Breton-Sintes and Cauderon 1978; Luria 1983). The Israeli tetraploid differs from the diploid cytotype of ssp. elongatus only in its somewhat taller plants, shorter flag leaf, larger stomata, larger pollen grains and longer caryopses.

The tetraploid subspecies grows in the Mediterranean basin (Heneen and Runemark 1972; Luria 1983; Gabi and Dogan 2010). In Israel, the tetraploid subspecies grows from Einot Gibton, Shfela, southwards (was found in Nahal-Zin springs in the Negev). The distribution of this taxon is fragmentary and the populations are isolated from one another. In the salty spring of Einot Gibton, the only site in Israel where the diploid and the tetraploid subspecies occur together, the diploid grows near the spring and the tetraploid in the outer ring (Luria 1983). Hence, the diploid subspecies can tolerate higher concentrations of salt than the tetraploid subspecies.

5.2.3.3.2 Cytology, Cytogenetics and Evolution

Based on karyomorphological data, Heneen and Runemark (1972) assumed that the tetraploid subspecies (2n = 4x = 28; genome EeEeEeEe) is an autotetraploid, derived from the diploid subspecies by chromosome doubling, or rather, via inter-varietal hybridizations followed by chromosome doubling. An inter-varietal origin of natural autopolyploids in different groups of plants is a widespread phenomenon, as discussed by Stebbins (1950, 1971).

Heneen and Runemark (1972), Breton-sintes and Cauderon (1978), and Luria (1983) arranged the chromosomes of the tetraploid subspecies in seven groups of four. These groups morphologically correspond to the seven pairs of the diploid subspecies, supporting the likelihood of an autopolyploid origin of the tetraploid (Heneen and Runemark 1972; Breton-Sintes and Cauderon 1978). Yet, detailed karyomorphological studies showed the existence of small differences between pairs within groups of four (Heneen and Runemark 1972; Breton-Sintes and Cauderon 1978; Luria 1983). Consequently, it was proposed that this taxon originated from hybridization between diploid varieties that underwent some chromosomal divergence and therefore, possess two partially diverged genomes, namely, Ee1Ee1Ee2Ee2 (Breton-Sintes and Cauderon 1978).

Chromosome pairing in the F1 triploid hybrid between the tetraploid and the diploid subspecies, is only slightly lower than that expected for a hybrid between autotetraploid and its diploid progenitor (Dvorak 1981b; Charpentier et al. 1986). The assessment of the homology between the two genomes of the tetraploid showed that differentiation had occurred in all chromosome arms that could be tested. From the pairing frequencies of individual telosomes of the diploid subspecies of E. elongatus with chromosomes of the tetraploid subspecies, Dvorak (1981a) concluded that slight differentiation occurred in every chromosome of the two subspecies. Thus, both genomes of the tetraploid subspecies appear to be a slightly modified version of the genome of the diploid subspecies (Dvorak 1981b).

To account for this genomic divergence, Heneen and Runemark (1972) suggested that the autotetraploid originated from hybridization(s) between different diploid lines whose karyotype underwent some structural chromosomal differentiation. On the other hand, Dvorak (1981b), Dvorak and Scheltgen (1973) and Dvorak and McGuire (1981) proposed that this differentiation resulted from changes in nucleotide sequences, rather than chromosomal aberrations such as inversions, translocations and other structural rearrangements. Alternatively, Eilam et al. (2009, 2010) suggested that the tetraploid subspecies underwent some cytological diploidization at the tetraploid level due to elimination of DNA sequences from two chromosomes in each group of four. Indeed, the tetraploid subspecies contained a significantly smaller amount of nuclear DNA (about 10% less) than the expected additive value of the diploid parent (Eilam et al. 2009, 2010). Also, a newly synthesized autotetraploid line of E. elongatus, produced by Charpentier et al. (1986), had significantly less DNA (8.57%) than the expected additive value (Eilam et al. 2009) (Fig. 5.2). The similarity in nuclear DNA content between the synthesized and the natural autotetraploids of E. elongatus indicates that the reduction in DNA content in the natural autotetraploid occurred immediately after its production, with only small changes in genome size over the history of the autotetraploid. Elimination of DNA sequences from two out of the four homologous chromosomes in each set of four, or elimination of sequences from one pair and other sequences from the second pair, augments the differentiation between the constituent subgenomes. Hence, the two subgenomes that became slightly divergent as a consequence of this pattern of elimination, underwent cytological diploidization. This reduction in nuclear DNA may lead to exclusive bivalent pairing between fully homologous chromosomes and consequently, disomic inheritance. The eliminated sequences are likely to include those that participate in homologous recognition and initiation of meiotic pairing.

Fig. 5.2
figure 2

Nuclear DNA content in diploid species and their polyploid derivatives, natural (black) and synthetic (gray): a Diploid and autotetraploid cytotypes of Elymus elongatus (taken from Eilam et al. 2009), and b Diploids, allotetraploid and allo-auto-hexaploid of Elymus farctus (taken from the Angiosperm C-value database (http://www.rbgkew.org.uk.cval/) at Kew Botanic Gardens. *The expected value, for each polyploid derivative, is the sum of the observed values of its parents

The chromosomes in other tetraploid Elymus species, such as E. farctus ssp. boreo-atlanticus (Heneen 1962) and E. rechingeri (Heneen and Runemark 1962), could not be grouped into groups of four. These species most likely have an allopolyploid origin. An autopolyploid origin of the tetraploid subspecies of E. elongatus is also indicated by the occasional formation of quadrivalents at meiosis (Heneen and Runemark 1972). This conclusion was also supported by genome analysis; Dvorak (1981b) and Charpentier et al. (1986) observed extensive pairing at the first meiotic metaphase in the triploid hybrid of these two subspecies and concluded that the three genomes of the triploid (Ee, Ee1, and Ee2) are closely related.

Meiosis is generally regular in subsp. flaccidifolius (Heneen and Runemak 1972; Dvorak 1981b; Luria 1983; Charpentier et al. 1986). For instance, Charpentier et al. (1986) observed 0.12 univalents, 13.9 bivalents 0.025 quadrivalents; 26.55 chaismata/cell at first meiotic metaphase. The majority of the cells had all the chromosomes paired as ring bivalents, indicating a high degree of homology within chromosome pairs. Multivalents, represented mainly by quadrivalents, occurred very rarely. The preferential bivalent pairing and the rarity of multivalent pairing in the autotetraploid subspecies, indicate either that this taxon is an autotetraploid that underwent cytological diploidization (Eilam et al. 2009, 2010) or that the tetraploid subspecies is a segmental allopolyploid (Breton-sintes and Cauderon 1978).

An induced autotetraploid of E. elongatus, produced by colchicine treatment of a diploid plant, was found morphologically indistinguishable from the natural tetraploid (Charpentier et al. 1986). The F1 hybrid between the natural and the induced autotetraploid had almost complete chromosome pairing, with an average of 1.0 univalents, 7.9 bivalents, 2.8 quadrivalents and 23.8 chiasmata per cell, nearly similar to the chromosomal pairing observed in the induced autotetraploid parent (Charpentier et al. 1986). This pairing pattern further supports the autopolyploid nature of the natural tetraploid subspecies. Because of this slight genomic divergence, Dvorak (1981a, b) suggested classifying the tetraploid and the diploid taxa in two separate species. However, since autotetraploids and their diploid progenitors are usually included in the same species (Stebbins 1950), diploid and tetraploid elongatus are two cytotypes and were classified as members of a single biological species (Heneen and Runemark 1972; Breton-Sintes and Cauderon 1978).

5.2.3.3.3 Crosses with Other Triticineae Species

Homology between genomes Eel and Ee2 was also inferred from the meiotic behavior of the F1 hybrid between tetraploid E. elongatus and common wheat. Despite the presence of one dose of Phl, the F1 hybrid (2n = 5x = 35; genome BADEe1Ee2) exhibited at the first meiotic metaphase five to seven bivalents, interpreted as autosyndetic pairing of elongatus Eel and Ee2 chromosomes (El Gawas and Khalil 1973; Dvorak 1981b; Sharma and Gill 1983; Charpentier et al. 1988a). This number of bivalents shows that most, if not all, chromosomes of the two Ee subgenomes were involved in pairing. Pairing of the elongatus Ee1 and Ee2 chromosomes in the presence of one dose of the Ph1 gene indicated that these two genomes still retained their homology, further supporting the autoploid nature of tetraploid E. elongatus.

Indications that genomes Ee1 and Ee2 are closely related were also reported by Han and Li (1993). In two crossing combinations, Triticum timopheevii ssp. timopheevii (2n = 4x = 28; genome GGAA) x tetraploid E. elongatus and T. turgidum ssp. durum (2n = 4x = 28; genome BBAA) x tetraploid E. elongatus, chromosome pairing at first meiotic metaphase included a mean 9.10 univalents, 9.11 bivalents, 0.20 trivalents and 13.78 univalents, 6.87 bivalents, and 0.15 trivalents, respectively (Han and Li 1993). Since pairing between A and B wheat subgenomes is very low in haploids of tetraploid wheat containing the Ph1 gene (Kihara 1936; Lacadena and Ramos 1968), pairing in the hybrid containing the BAEe1Ee2 genomes was likely due to autosyndesis between Ee1 and Ee2 chromosomes of tetraploid E. elongataus. Similar homologous relationships were observd between the two Eb genomes of hexaploid E. farctus (Charpentier 1992) and between the two Eb genomes of tetraploid E. farctus (de V Pienaar et al. 1988).

The hybrid formed between tetraploid E. elongatus and common wheat, with zero dose of Ph1, exhibited a relatively high degree of autosyndetic pairing between elongatus two-subgenome chromosomes and between wheat-subgenome chromosomes and allosyndetic pairing between wheat and elongatus chromosomes (range of chromosomal pairing was 9.6–11.2 bivalents and 1.2–1.9 trivalents per cell; quadrivalents, and some pentavalents were also observed) (Charpentier et al. 1988a). In contrast, a drastic reduction in pairing was observed in hybrids carrying one dose of Ph1. Altogether they showed a means 4.6–7.7 bivalents per cell, multivalents were rare. The number of chiasmata/cell dropped from 19–20 in Ph1-deficient hybrids to 6–10 in hybrids with one dose of this gene (Charpentier et al. 1988a).

Ph1-deficient haploid bread wheat was found to form at first meiotic metaphase 3.2–4.2 bivalents, 0.9–2.0 trivalents and very few quadrivalents (0.02–0.12) and pentavalents per cell (Riley 1960). Assuming a similar level of pairing between the wheat chromosomes in hybrids generated from the tetraploid subspecies of E. elongatus and Ph1-deficient bread wheat, then 6–7 bivalents of the observed 9.6–11.2 should be the results of pairing between Eel and Ee2 elongatus chromosomes. The number of quadrivalents and pentavalents that were observed in these hybrids indicate allosyndetic pairing. The relatively high level of allosyndetic pairing in F1 hybrids between tetraploid E. elongatus and Ph1-deficient bread wheat, and the low level of allosyndetic pairing in the presence of Ph1, indicates that the subgenomes of tetraploid elongatum do not have genes that suppresses or promote homoeologous pairing.

Interestingly, in contrast to the effect of Ph-suppressors or homoeologous pairing promoters on diploid E. elongatus (Dvorak 1987), such an effect was not observed in hybrids between common wheat and tetraploid E. elongatus (Dvorak 1981b).

To determine the chromosomal location of these and other genes that control pairing in diploid E. elongatus, disomic addition lines of chromosomes derived from the diploid subspecies of E. elongatus in the background of Chinese Spring, were crossed with the tetraploid subspecies of E. elongatus, and pairing was then compared to those observed in hybrids between Chinese Spring and tetraploid E. elongatus, whose Ee1 and Ee2 chromosomes were previously defined as homologues (Charpentier et al. 1986, 1988a). The resultant F1 hybrids (2n = 5x = 36), each carrying three doses of a given elongatus chromosome, enabled evaluation of the effect of each elongatus chromosome on pairing of homologues (Eel with Ee2) and homoeologues (A, B, D, and Ee). The study of chromosomal pairing in these hybrids enabled classification of the elongatus chromosomes into those that suppress (6Ee), promote (5Ee, 3Ee, and possibly also 1Ee), or have no effect on pairing (4Ee). The effect of chromosomes 2Ee and 7Ee was not studied. Chromosomes 5Ee and 3Ee differed in their effect on the degree and pattern of chromosome pairing, namely, the effect of 5Ee was stronger than that of 3Ee. Pairing analysis in such addition lines and in substitution lines, in their haploid derivatives and in hybrids between these lines and Triticum urartu, led Dvorak (1987) to allocate genes that promoted homologous or homoeologous pairing to chromosome arms 3EeS, 3EeL, 4EeS, 5Eep and to chromosome 6Ee of diploid E. elongatus. Genes suppressing homoeologous pairing were allocated to chromosome arms 4EeL and 7Eeq. In accord with Charpentier et al. (1988b), chromosomes 3Ee and 5Ee of diploid elongatus promoted homoeologous pairing.

In the presence of an extra dose of chromosome 6Ee of elongatus, the number of bivalents per cell was reduced, indicating suppression of pairing between the Eel and Ee2 chromosomes (Charpentier et al. 1988b), bringing Charpentier et al. to conclude that chromosome 6Ee of diploid E. elongatus carries gene(s) that inhibit(s) pairing or chiasma formation. This is in contrast to the finding of Dvorak (1987), who assigned a pairing-promoting effect to this chromosome. Chromosome 4E had no effect on pairing in hybrids with wheat (Charpentier et al. 1988b). This is in accord with Dvorak (1987), who reported pairing suppression by the long arm of chromosome 4Ee, but assumed the presence of a pairing promoter on the short arm of 4Ee, thus accounting for the lack of pairing effect by the entire 4E chromosome. He also found a suppressive effect of chromosome arm 7Eq, a chromosome arm that was not studied by Charpentier et al. (1988b).

Charpentier et al. (1988b) also studied chromosome pairing at first meiotic metaphase in hybrids between the bread wheat cultivar Chinese Spring and a synthetic autotetraploid line derived from diploid E. elongatus. The hybrids exhibited a high level of homoeologous pairing. Apparently, the genome of the diploid, from which the autotetraploid was synthesized, promoted pairing even in the presence of Phl. A similar effect was reported for gene(s) derived from another diploid accession of E. elongatus (Dvorak 1981b). Promotion of homoeologous pairing by diploid E. elongatus was also observed in an amphiploid between allotetraploid E. farctus (subsp. boreali-atlanticus; genome EbEbEeEe) and diploid E. elongatus (genome EeEe) (Yvonne Cauderon, personal communication). The amphiploid had genome (EbEbEeEeEeEe). While tetraploid and hexaploid E. farctus exhibited mostly bivalents at meiosis, the amphiploid had several multivalents per cell, mostly quadrivalents but also some hexavalents. Evidently, in this amphiploid, genome Ee of dipoid E. elongatus promoted pairing between the homologues Ee genomes and between homoeologues Eb with Ee genoes. Three different accessions of diploid E. elongatus were found to promote homoeologous pairing: the accession used by Jenkins (1957) to produce the initial hybrid from which Dvorak and Knott (1974) derived their disomic addition lines, an accession from south France used by Cauderon in the cross with the tetraploid form of E. farctus and the Israeli accession, from which the induced autotetraploid was derived. Thus, the ability to promote homoeologous pairing may be a common feature of many accessions of diploid E. elongatus. Promotion of homoeologous pairing in the presence of Phl was described in several diploid Triticinae, viz. Aegilops speltoides (Riley 1960; Riley et al. 1961; Dvorak 1972), Amblyopyrum muticum (Riley 1966a, b; Dover and Riley 1972), Ae. longissima (Mello-Sampayo 1971b), Secale cereale (Riley et al. 1973; Lelley 1976; Dvorak 1977) and Dasypyrum villosum (Blanco et al. 1988b) (Table 5.2). However, chromosomal allocation of the promoters was only determined in rye (Lelley 1976) and in E. elongatus (Dvorak 1987; Charpentier et al. 1988b). In rye (Lelley 1976) chromosome 3R, and possibly also 5R, 4R, and 7R, were found to carry genes that promote homoeologous pairing in hybrids with wheat. This finding corresponds to the allocation of pairing promoters in diploid E. elongatus (Dvorak 1987; Charpentier et al. 1988b); the genes of 3Ee and 5Ee are presumably homoeoalleles to those of rye (Lelley 1976), as well as to those of homoeologous groups 3A, 3B and 3DL and 5A, 5D, and 5BS in bread wheat (Sears 1976).

Table 5.2 Promoters and suppressors of chromosomal pairing in Triticineae

5.2.3.4 Ssp. turcicus (P. E. McGuire) Melderis (2n = 8x = 56)

5.2.3.4.1 Morphological and Geographical Notes

Ssp. turcicus [=Elytrigia turcica P. E. Maguire; Elytrigia elongata ssp. turcica (P. E. McGuire) Valdés & H. Scholz; Elytrigia pontica ssp. turcica (P. E. McGuire) Jarvie & Barkworth; Lophopyrum turcicum (P. E. McGuire) McGuire ex Löve; Thinopyrum turcicum (P. E. Maguire) Cabi & Dogan], is a perennial caespitose, more or less glaucous grass with erect culms, 70–115 cm high, with 10–20 cm long spikes, spikelets with 7–9 florets and glumes with 7–9 veins. Anthers are 2.5–3.5 mm long.

Within the polyploid complex of E. elongatus, ssp. turcicus most resembles the decaploid ssp. ponticus. It differs from ponticus by the laxer leaves, less prominent ligules, more rounded apex of glumes, lack of hairs inside the glumes at the apex, smaller anthers, and in chromosome number (McGuire 1983). Although morphologically similar, the octoploid and the decaploid taxa were treated as separate species by McGurie (1983), but Melderis (1978), Dewey (1984), and Moustakas (1989) considered the morphological differences between the two taxa insufficient for separation on the specific level and consequently, classified them as two subspecies.

This subspecies distributes in Thassos Island, Greece (Moustakas 1993), Turkey, Georgia, and northern Iran (Jarvie 1992). It grows on dry calcareous, saline land from sea level to dry and saline mountain habitats, 1800 m above sea level, in low rainfall areas.

5.2.3.4.2 Cytology, Cytogenetics and Evolution

Chromosome counts in accessions of Elymus elongatus from eastern Turkey and northern Iran, showed the presence of octoploid plants with 2n = 8x = 56 (Lorenz and Schulz-Schaeffer 1964; Sculz-Schaeffer et al. 1971; McGuire 1983), implying that the octoploids are not just sporadic individuals, arising in populations with other ploidy levels, but represent established populations (McGuire 1983).

Moustakas (1993) performed computer-aided karyotype analysis and found that the karyotype of the octoploid taxon is asymmetric, namely, it is composed of 10 metacentric chromosome pairs, 15 sub-metacentric chromosome pairs and 3 sub-telocentric chromosome pairs. Only two chromosome pairs have secondary constrictions, implying that a number of NORs are inactive in this octoploid.

All the chromosomes pairs of the octoploid can be matched to the chromosome pairs of the decaploid. Only the sub-telocentric satellited chromosome pairs differ slightly (Moustakas 1993). Yet, the karyotype analysis (Moustakas 1989, 1991, 1993) indicated that the chromosomes of the present-day diploid elongatus and those of the octoploid and decaploid diverged from each other.

Analysis of seed protein polymorphism patterns (Moustakas 1989) revealed that the octoploid originated from a speciation event more recent than that associated with the decaploid from which it presumably evolved. Moustakas (1989) found that the patterns of seed-protein electrophoresis of ssp. turcicus were qualitatively similar to those of ssp. ponticus.

Taking into consideration the results of the karyotype analysis, Moustakas (1993) concluded that ssp. turcicus is a segmental allopolyploid, with genome designation JjJjJjJjJeJeJeJe. (Genome designations Jj and Je represent the same genome but with some structural nuances.) On the other hand, Jarvie (1992) thought that the genome of subsp. turcicus is EEEEJJJJ is an auto-allo polyploid. Yet, if subsp. turcicus derived from the decaploid subsp. ponticus, its genome designation should be EeEeEbEbStStStSt.

5.2.3.5 Ssp. ponticus (Podp) Melderis (2n = 10x = 70)

5.2.3.5.1 Morphological and Geographical Notes

Ssp. ponticus (commonly known as tall wheatgrass and rush wheatgrass) [Syn.: Triticum ponticum Podp; Agropyron elongatum ssp. ponticum (Podp.) Senghas; Agropyron incrustatum Adamovic; Elymus ponticus (Podp.) N. Snow; Elytrigia pontica (Podp.) Holub; Elytrigia elongata ssp. pontica (Podp.) Gamisans; Elytrigia ruthenica (Griseb.) Prokudin; Lophopyrum ponticum (Podp.) Á. Löve; Thinopyrum ponticum (Podp.) Barkworth & D. R. Dewey] is a perennial, caespitose, tall plant with a 50–100 cm high stem. Its leaves are green or glaucous bluish with flat to curling blades that are often covered with short, stiff hairs, its lower sheaths usually ciliate, spikes are 10–35 cm long, lower internodes are usually much longer than the spikelets, rachis is not fragile, spikelets are 17–25 mm long, with 8–18 florets, glumes are thick and hardened, and 9–11 mm long, with 5–7 veins, lemmas are also thick and hardened, and 10–13 mm long, palea with cilia are seen along the entire length of keels and anthers are 4–7 mm long. Under certain conditions, this perennial grass can grow up to 2 m tall, and spikelets up to 3 cm long, each containing up to 12 flowers.

Ssp. ponticus is native to southeastern Europe, Turkey near the Black Sea and southern Russia. It grows well in dry and saline habitats, especially alkaline soils, as well as in disturbed habitats, such as waste ground and roadsides. This subspecies is found generally 360–1740 m above sea level.

5.2.3.5.2 Cytology, Cytogenetics and Evolution

Peto (1930), Simonet (1935), and Vakar (1935) reported that accessions of E. elongatus from Russia are decaploid, with 2n = 10x = 70 chromosome number. Heneen and Runemark (1972) superficially described the karyotype of the decaploid subspecies. The large number of chromosomes rendered it difficult for them to construct the karyotype and to identify all the SAT-chromosomes. Generally, chromosome morphology and SAT-chromosome type seemed similar to those of the diploid and tetraploid subspecies of E. elongatus, indicating that ssp. ponticus is interrelated to these subspecies (Heneen and Runemark 1972), supporting the view that different subspecies of E. elongatus are involved in the origin of the decaploid subspecies.

On the diploid and tetraploid levels, there is a correlation between the number of SAT-chromosomes and degree of ploidy. This correlation is not obvious at the decaploid level, since the number of the barely detectable SAT-chromosomes in the decaploid subspecies was not proportional to the degree of ploidy (Heneen and Runemark 1972).

Up to ten nucleoli were recorded in pre-meiotic cells of the decaploid subspecies (Schulz-Schaeffer and Jura 1967). Brasileiro-Vidal et al. (2003), using silver nitrate staining to determine the number of nucleoli and NORs, revealed 17 AG-NOR sites on mitotic metaphase cells—a number similar to that of 45S rDNA detected via FISH. However, the mean number of nucleoli per interphase nucleus was much lower; in most cells, the number ranged from four to nine, indicating that at interphase, the active Ag-NOR sites tend to coalesce, as suggested by Lacadena et al. (1988).

Li and Zhang (2002) used FISH to study the distribution of the 18S-5.8S-26S rDNA in the decaploid subspecies and in its related diploid taxa, E. elongatus ssp. elongatus (genome EeEe), E. farctus subsp. bessarabicus (genome EbEb) and E. stipifolius (=Pseudoroegneria stipifolia) (genome StSt). The distribution of rDNA genes was similar in all three diploid taxa, i.e., two pairs of loci were observed in each somatic cell at metaphase and interphase. The first pair was located near the terminal end and the second in the interstitial regions of the short arms of a pair of chromosomes. The maximum number of major rRNA loci detected on metaphase spreads of the decaploid subspecies was 20, which corresponded to the additive sum of that of its progenitors. However, in the decaploid, all of the major loci were located on the terminal end of the short arms of the chromosomes. Apparently, the interstitial loci that exist in the possible diploid donors of genomes to the decaploid, changed their position during the formation and evolutionary history of the decaploid. These results suggest that there has been distinct differentiation between ponticus and its diploid relatives during the evolutionary process (Li and Zhang 2002). Positional changes of 18S-5.8S-26S rDNA loci between ssp. ponticus and its candidate genome donors, indicate that it is almost impossible to find a genome in the decaploid that is completely identical to that of its diploid donors (Li and Zhang 2002). The interstitial position is likely an ancestral trait, whereas the terminal position is probably a later-derived trait (Dubcovsky and Dvorak 1995). During polyploidization of ssp. ponticus, all of the interstitial loci have been either deleted and novel loci have been positioned on terminal regions of the chromosomes, or, alternatively, migrated to terminal positions. A similar phenomenon has been observed in other Triticeae species (Gill and Apples 1988; Dubcovsky and Dvorak 1995), although the exact underlying mechanism remains unknown.

Using FISH to determine the number and position of 45S and 5S rDNA sites in another accession of ssp. ponticus, Brasileiro-Vidal et al. (2003) detected both 45S and 5S rDNA sites on the short arms of 17 chromosomes, while on three other chromosomes, only the 5S rDNA site was observed. In ssp. ponticus, the 45S rDNA loci were always distally located in relation to the 5S rDNA loci. The occurrence of these sites in 17 instead of 20 chromosomes, as observed by Li and Zhang (2002), most likely indicates a reduction in the number of 45S rDNA sites in the accession used by Brasileiro-Vidal et al. (2003).

In ssp. elongatus, the 5S rDNA sites were associated with chromosomes 1Ee and, possibly, 5Ee (Scoles et al. 1988; Dvorák et al. 1989). Considering the distribution of these sites in diploid elongatus, the chromosomes carrying the 5S rDNA in the decaploid might also belong to homoeologous groups 1 and 5 (Brasileiro-Vidal et al. 2003).

Meiosis in the decaploid was less ordered. Cauderon (1958) observed 1.04 univalents, 19.9 bivalents, and a number of chain and ring multivalents (0.76 trivalents, 2.71 quadrivalents, 0.81 pentavalents, 0.52 hexavalents, 0.76 heptavalents, 0.19 octovalents, 0.05 ennevalents (=nine valents), and 0.04 decavalents). Similar patterns of chromosomal pairing were also observed by Zhang et al (1993, 1996) and Muramatsu (1990). The high frequency of multivalents in the decaploid subspecies may be the results of the activity of pairing promoters that exist in the ssp. ponticus genome (Zhang et al. 1993, 1995; Cai and Jones 1997).

Genomic relationships between the genomes of ssp. ponticus and its related taxa, have been the subject of several studies, and, due to chromosome pairing complexity, different genome formulae have been proposed for the decaploid. Peto (1936), assuming that the decaploid is an auto-allo-polyploid, tentatively assigned the genome formula AAXXXXYYYY, whereas Matsumura (1949) proposed the genome formula BBXXXXYYYY. Both researchers assumed that the A and B subgenomes of Triticum exist in ssp. ponticus. Muramatsu (1990) and Wang et al. (1991) concluded from the high frequency of multivalents in pollen mother cells (PMCs) of both the decaploid and the polyhaploid of ssp. ponticus, that this taxon contains several closely related or identical genomes, and therefore, is an autodecaploid with the genomic formula J1J1J2J2J3J3J4J4J5J5 and JJJJJJJJJJ, respectively. The J genome is from E. farctus ssp. bessarabicus (currently designated as Eb), is closely related to the Ee genome of ssp. elongatus and possesses modified versions of the same basic genome, namely, E (Forster and Miller 1989; Wang and Hsiao 1989; Wang 1990). Dvorak (1975) and Wang et al. (1991) regarded ssp. ponticus to be an autodecaploid that behaves as an allodecaploid, due to a bivalentization system.

Konarev (1979) assumed that the diploid subspecies of E. elongatus contributed to the karyotype of the decaploid taxon. Moustakas (1993), based on karyotype analysis, concluded that ssp. ponticus is a segmental allopolyploid, with genome formula JjJjJjJjJjJjJeJeJeJe, where genome designations Jj and Je represent the same genome but with structural differences. Dvorak (1975, 1981a) showed pairing of the chromosomes of diploid elongatus with some of the chromosomes of the decaploid, but the pairing was poor in every case, indicating that differentiation of the chromosomes had occurred. Consequently, Dvorak (1975) postulated that the decaploid evolved from an ancestral elongatus-like diploid taxa, by primary and secondary chromosome doubling of inter-ecotypic or inter-specific hybrids. In accordance, the karyotype analysis performed by Moustakas (1989, 1991), indicated that the chromosomes of the present-day diploid elongatus and those of the decaploid diverged from one another.

According to Dvorak (1981b), the decaploid appears to have one group of three closely related genomes and another group of two closely related genomes. Moreover, he suggested that the chromosomes of the diploid elongatus are more closely related to the doublet of the decaploid genome than to the triplet. Thus, Zhang and Dvorak (1990) suggested the genome designation ExExExExEyEyEyEyEyEy for subsp. ponticus. Jarvie (1992), assuming that ssp. ponticus is an auto-allo-polyploid, proposed the genome symbol EEEEEEJJJJ. Zhang et al. (1996), on the basis of GISH studies and genome specific markers, also suggested that this subspecies is an auto-allo-polyploid, but with the genome symbol StStStStEeEeEbEbExEx, where the St genome is homologous to the St genome of Elymus stipifolius. Moreover, GISH revealed that the centromeric region might be the critical area for discrimination between the St and E subgenomes (either Ee or Eb) in ssp. ponticus. Mitotic cells of several accessions of ssp. ponticus, when hybridized with the St probe and blocked by E genomic DNA, had 28 chromosomes strongly hybridized by the St probe at regions near the centromere (Zhang et al. 1996). When Ee or Eb was labeled as the probe and St was used as the blocker, all 70 chromosomes were labeled with FITC (fluorescein isothiocyanate). However, there were about 28 chromosomes lacking hybridization signals at the centromeric regions (Zhang et al. 1996). These consistent results were interpreted by Zhang et al. (1996) to mean that ssp. ponticus has 28 St genome chromosomes and 42 E genome chromosomes. The chromosome pairing data of Wang (1992) and the molecular studies of Hsiao et al. (1995) showed that the St, Ee, and Eb genomes are very closely related. The GISH results of Zhang et al. (1996) also revealed the close relationships between these three genomes and that GISH cannot distinguish between Eb and Ee. Taken together, the centromere and the region nearby may be the critical areas that discriminate the St from the E genomes in ssp. ponticus.

The 70 chromosomes of ssp. ponticus all fluoresced bright yellow when probed either with DNA from the Ee genome of E. elongatus ssp. elongatus or from the Eb genome of E. farctus ssp. bessarabicus (Chen et al. 1998a, b). This demonstrated that a substantial affinity exists between these probes and the subgenomes present in ponticus. Conversely, no obvious hybridization signal was detected in ponticus when probing either with DNA from the Ee genome and blocking with Eb genome DNA, or in the reverse analysis, using Eb genome DNA as probe and Ee genome DNA as blocker. Since this is expected when an effective DNA probe is used to block itself, the results suggest that the Ee and Eb genomes are closely related to one another and to the chromosomes of ssp. ponticus.

Chen et al. (1998a, b) also performed GISH using genomic DNA probes from E. elongatus ssp. elongatus (genome Ee), E. farctus ssp. bessarabicus (genome Eb), and E. strigisus (=Pseudoroegneria strigosa) (genome St), to investigate the genomic constitution of ssp. ponticus. Their findings indicated that the decaploid subspecies had only the two basic genomes Eb and Ebs (=Js). The Ebs genome of ponticus is homologous with E (Ee and Eb) genomes, but is quite distinct at the centromeric regions, which strongly hybridize with the St genomic DNA probe. This may indicate that the Ebs genome is a modified Eb (=J) genome whose chromosomes exchanged St segments via translocations between the two (Chen et al. 1998a, b). Support of this hypothesis also came from lack of centromeric hybridization signals upon hybridization of mitotic chromosomes of the diploid subspecies of elongatus and farctus with St genome DNA in the presence of Eb or Ee genome blocker (Chen et al. 1998a, b). Likewise, mitotic chromosomes of E. strigosus, probed with Eb genome DNA and blocked with St genome DNA, showed no hybridization signal. It appears that the chromosomes of ponticus, which show hybridization affinity with the centromeres of E. strigosus DNA, were not simply derived from any of these three diploid species, but rather, have a more complicated origin (Chen et al. 1998a, b).

Consequently, the group proposed that ssp. ponticus contains only segments of the St genome rather than any intact St genome or chromosomes. Based on the GISH results, namely, that all 70 chromosomes of ssp. ponticus hybridized extensively with Eb or Ee genome DNA probes, even in the presence of St genome blocker, Chen et al. (1998a, b) redesignated the genomic formula of ssp. ponticus as EEEEEEEbsEbsEbsEbs, where E refers to the Ee- or Eb-type chromosomes closely related to the genomes of ssp. elongatus and ssp. bessarabicus, respectively, while Ebs refers to a modified Ee- or Eb-type chromosomes distinguished by the presence of St genome-specific sequences close to the centromere.

The major disagreement between Zhang et al. (1996) and Chen et al. (1998a, b) centered around the explanation of the GISH results of ssp. ponticus probed by St genomic DNA and blocked by E genomic DNA. The St genomic probe hybridized all 70 chromosomes, but more strongly hybridized with 28 chromosomes at their centromeres and nearby regions. In the reverse GISH analysis, these 28 chromosomes were also hybridized by the E genomic probe, except for their centromeric and nearby regions, that were completely blocked by the St genomic DNA (Zhang et al. 1996). Zhang et al. opined that the unexpected signals appearing beyond the probe genome chromosomes were mainly caused by cross-hybridization between St and E genomes, arising from their close relationship in ssp. ponticus. Therefore, Zhang et al. (1996) proposed that the centromeres and nearby regions might be critical in the discrimination of St and E genomes.

Li and Zhang (2002) and Liu et al. (2007) accepted this interpretation, and used StStStStEeEeEbEbExEx as the genome formula for ssp. ponticus. Accordingly, the candidate donors of genomes to ssp. ponticus have been narrowed down to a few species, including the diploid species of the genus Elymus, namely, E. elongatus ssp. elongatus, E. farctus ssp. bessarabicus and E. stipifolius.

Since cytogenetic data indicate that the St, Ee, Eb (=J), and Ebs (=Js) genomes are very closely related (Wang 1992), the latest genomic designations are consistent with the earlier autopolyploid designation for these subspecies (Fedak et al. (2000). The study of Chen et al. (1998a, b) indicates that ssp. ponticus is not a characteristic autodecaploid and its five subgenomes are most likely modified versions of the Eb or Ee genomes. Since 28 chromosomes of ssp. ponticus containing centromeric region of St chromosomes are recombinant chromosomes, they concluded that ponticus can be regarded as a segmental autodecaploid with three sets of the E genome (Ee or Eb) genomes plus two sets of Ebs genome (Chen et al. 1998a, b). In contrast, Zhang et al. (1996) and Li and Zhang (2002) found that the decaploid is an auto-allo-polyploid containing six Ee genomes and four St genomes. The existence of multivalents in meiosis of the decaploid subspecies indicates that intergenomic recombination occurs quite frequently in this subspecies. Therefore, the genomes in the decaploid are recombinant genomes and differ from those of their donors.

5.2.3.5.3 Crosses with Other Triticineae Species

Peto (1936) suggested that the relatively high number of paired chromosomes in F1 hybrids between Triticum turgidum ssp. dicoccon (2n = 4x = 28; genome BBAA) x Elymus elongatus ssp. ponticus (2n = 10x = 70; genome EeEeEbEbEbEbStStStSt) indicates homology between some wheat and E. elongatus chromosomes. Although B-subgenome chromosomes of wheat occasionally paired with other chromosomes at first meiotic metaphase of these F1 hybrids, aceto-carmine Giemsa N-banding analysis of chromosomes in root tip cells and PMCs of the F1 hybrid between Triticum aestivum cv. Fukuhoc and ssp. ponticus, showed that the latter does not contain the B subgenome of wheat (Zhang et al. 1993).

The average pairing in first meiotic metaphase of the F1 hybrid between tetraploid wheat, Triticum turgidum subsp. durum and E. Elongatus ssp. ponticus included 14.93 (9–25) univalents, 11.92 (5–18) bivalents, 2.14 (0–5) trivalents, 0.54 (0–3) quadrivalents, and 0.42 (0–20) pentavalents (Zhang et al. 1993). The average chromosome pairing in the F1 hybrid between hexaploid wheat, Triticum aestivum cv. Fukuhoc and ssp. ponticus displayed 10.87 (5–19) univalents, 16.40 (6–22) bivalents, 2.78 (0–6) trivalents, 0.55 (0–3) quadrivalents, and 0.23 (0–3) pentavalents (Zhang et al. 1993). Likewise, the F1 hybrid T. aestivum cv. Chinese Spring x ssp. ponticus had 11.19 (4–17) univalents, 14.73 (7–20) bivalents, 3.12 (0–6) trivalents, 0.67 (0–2) quadrivalents, and 0.63 (0–2) pentavalents (Zhang et al. 1993). Cai and Jones (1997) used GISH to distinguish autosyndetic from allosyndetic pairing in the hybrid between ssp. ponticus and Triticum aestivum cv. Chinese Spring. Chromosome pairing in this hybrid occurred mainly among wheat chromosomes and among ssp. ponticus chromosomes, whereas allosyndetic pairing between wheat and ponticus chromosomes was very low. These results showed that the relationships among T. aestivum subgenomes and among ssp. ponticus subgenomes are much closer than the relationship between the subgenomes of the two species. The higher frequencies of autosyndetic pairing among the chromosomes of ssp. ponticus than among bread wheat chromosomes in the hybrid, indicated that the relationships between the five subgenomes of ssp. ponticus are closer than those between the three subgenomes of T. aestivum.

Comparing the observations in aestivum x ponticus to those in durum x ponticus, shows that adding the D subgenome leads to pairing of more than eleven chromosomes. This may indicate either that ponticus contains gene(s) that promote homoeologous pairing in the presence of Ph1, located in several different subgenomes of ponticus (Zhang et al. 1993, 1995), or that ponticus has a subgenome(s) related to the D subgenome of bread wheat (Zhang et al. 1993). Cai and Jones (1997) reported relatively high autosyndetic pairing frequencies among bread wheat chromosomes in the hybrid of subsp. ponticus x Triticum aestivum cv. Chinese Spring. The mean autosyndetic pairing frequency between wheat chromosomes in the hybrid was much higher than between those of euhaploids Chinese Spring with Ph1 (Jauhar et al. 1991). Since this hybrid carries the Ph1 gene, Cai and Jones (1997) suggested that ssp. ponticus carries gene(s) that can promote homoeologous chromosome pairing in the presence of Ph1.

Jauhar (1995) analyzed the F1 hybrid between ssp. ponticus and bread wheat, which proved perennial and morphologically resembled the Elymus parent. The hybrid (2n = 8x = 56; genome EeEbEbStStBAD) showed high chromosome pairing (average pairing included 9.24 univalents, 6.23 rod- and 11.68 ring-bivalents, 2.07 trivalents, 1.09 quadrivalents, 0.02 pentavalents, and 0.05 hexavalents). Like Cai and Jones (1997), also Jauhar (1995) suggested that this high pairing was due to the inactivation of Ph1 by ssp. ponticus promoters.

Addition lines of bread wheat bearing chromosomes of diploid E. elongatus, which are homoeologous with wheat chromosomes of groups 6 and 7, were crossed with addition lines of bread wheat carrying chromosomes of the decaploid subspecies of E. elongatus, which are also homoeologous with wheat chromosomes of groups 6 and 7 (Dvorak 1975). The chromosomes of the two elongatus subspecies paired with one another in the presence of the Ph1 gene of common wheat. Since pairing between diploid E. elongatus and decaploid E. elongatus chromosomes in the presence of the Ph1 gene, was generally low, it was suggested that the chromosomes assigned to the same group are not homologous, but rather, closely homoeologous (Dvorak 1975). In contrast to the low pairing between chromosomes of the two elongatus subspecies in the presence of Ph1, Dvorak (1975) assumed that these chromosomes could pair quite regularly in the absence of Ph1.

Johnson and Kimber (1967) produced complex hybrids bearing 29 chromosomes, including one telocentric chromosome (in different hybrids, different telocentric chromosomes represent a different chromosome arm) and twenty complete chromosomes of T. aestivum (2n = 6x = 42), seven complete chromosomes of Aegilops. speltoides (2n = 2x = 14) and one telocentric chromosome derived from E. elongatus ssp. ponticus, corresponding to homoeologous group 6. The presence of the Ae. speltoides genome induced pairing between homoeologous chromosomes at meiosis, even in the presence of Ph1. The elongatus telocentric chromosome paired with wheat chromosomes homoeologous to group 6. There was no evidence that it paired with chromosomes of any other group.

5.2.4 E. farctus (Viv.) Runemark Ex Melderis (Based on Eb Genome)

5.2.4.1 Species Description

E. farctus (Viv.) Runemark ex Melderis [Syn. Triticum farctum Viv.; Triticum junceum L.; Agropyron junceum (L.) Beauv.; Agropyron junceum ssp. mediterraneum Simonet; Elytrigia juncea (L.) Nevski; Agropyron farctum (Viv.) Rothm.; Elymus multinodus Gould; Elymus farctus (Viv.) Runemark; Elytrigia juncea ssp. mediterranea (simonet) Hyl.; Thinopyrum junceum (L.) Á. Löve; Thinopyrum farctum (Viv.) Cabi & Dogan comb. nov.] is rhizomatous, perennial, with rigid and glabrous 30–60 cm high stems, 2–5 mm broad, glaucous-green, leaves, densely pubescent on ribs of upper surface, with no auricles; 15–25 cm long spikes, fragile rachis (wedge type disarticulation), 5–12 veined and keeled 10–18 mm long glumes, 10–18 mm long lemmas and 6–12 mm long anthers, Caryopsis with adherent pericarp, 9.0 mm long.

This species (commonly known as sand couch-grass) comprises a polyploid complex represented by diploid, tetraploid, hexaploid, and possibly also octoploid subspecies, some of which are not well defined morphologically (Simonet and Guinochet 1938; Heneen 1972, 1977; Cauderon 1979). The diploid (2n = 2x = 14), tetraploid (2n = 4x = 28), and hexaploid (2n = 6x = 42) subspecies carry the genomes EbEb, Eb1Eb1Eb2Eb2, and Eb1Eb1Eb2Eb2EeEe (=JJ, J1J1J2J2, and J1J1J2J2EE), respectively (Cauderon 1958). While the tetraploid subspecies arose as an autopolyploid whose two genomes underwent some differentiation, it is assumed that the hexaploid was derived from a cross between the tetraploid subspecies and diploid E. elongatus (genome EeEe), through alloploidy (Cauderon 1958). The two Eb subgenomes (Eb1 and Eb2) of the tetraploid and the hexaploid subspecies are still very closely related, as is evident by the almost regular autosyndetic pairing of their chromosomes in hybrids with other polyploid Elymus species (Ostergen 1940a; Cauderon 1958). The Eb and Ee subgenomes are also related, as demonstrated by hybrids between tetraploid E. farctus and diploid E. elongatus, which show an average of 2.8 trivalents per cell (Cauderon and Saigne 1961). In fact, based on the high frequency of trivalents, Dvorak (1981a, b) and Dewey (1984) suggested that the Eb and Ee subgenomes be regarded as variations of the same genome. Yet, despite the close relationship between these subgenomes and the ability of their corresponding chromosomes to pair with each other, the hexaploid is characterized by almost complete bivalent pairing at the first meiotic metaphase; multivalents are rarely found (Charpentier et al. 1986).

All the subspecies are facultative cross-fertilizing and are capable of self-fertilization (Melderis 1978; Luria 1983). Nearly all subspecies possess long-creeping rhizomes, except for subsp. rechingeri, which grows in tufts on maritime rocks. All subspecies share a common smooth spike-rachis, readily disarticulating at maturity (wedge-type disarticulation), long anthers and flat, often convolute leaves, with densely and minutely hairy prominent ribs.

Simonet (1935) and Simonet and Guinochet (1938) were the first to notice the intraspecific differentiation of E. farctus. On the basis of karyological, morphological and chorological studies, they divided E. farctus (then Agropyron junceum (L.) P. Beauv.) into two subspecies: northwestern European (Atlantic) 2n = 28 (named A. junceum ssp. boreali-atlanticum) and southern European (Mediterranean) 2n = 42 (named A. junceum subsp. mediterraneum). Prokudin (1954) divided Elytrigia juncea (L.) Nevski (=Agropyron junceum) into three separate species: E. juncea (L.) nevski s.str., and E. mediterranea (simonet et Guinochet) Prokudin, both having a south-Europe distribution range, and E. junceiformis A. & D. Löve, occurring in the northern coasts of Europe. However, Melderis (1978, 1980) transferred the Elytrigia genus to Elymus and named the species Elymus farctus, which he then further divided into the following four subspecies: ssp. boreali-atlanticus, occurring in the northern and western coasts of Europe, ssp. farctus, occurring in western Europe and the Mediterranean basin, ssp. bessarabicus and ssp. rechingerii, occurring in the southern part of Europe. Hence, Elymus farctus occurs in Europe and the Middle East and grows in maritime sands, near sea level.

5.2.4.2 Ssp. bessaribicus (Savul. & Rayss) Melderis (2n = 2x = 14)

5.2.4.2.1 Morphological and Geographical Notes

Ssp. bessaribicus (Savul. & Rayss) Melderis [=Agropyron bessarabicum Savul. & Rayss; Elytrigia juncea ssp. bessarabicum (Savul. & Rayss) Tzvelev; Thinopyrum bessarabicum (Savul. & Rayyss) Á. Löve]) is perennial, with shortly creeping or absent rhizomes, 50–80 cm high plants, with rigid, fairly thick, culms, usually not swollen at base, 15–35 cm long, erect or slightly curved spikes, fragile rachis, breaking at maturity above each spikelet (wedge type disarticulation), where rachis internodes are usually longer than lower spikelets; 10–25 mm long 5–9 flowered spikelets, appressed to rachis and laterally compressed, 10–18 mm long 6–12 veined, asymmetrically keeled, and unawned glumes, 10–20 mm long unawned lemma, keeled towards apex, and 10–12 mm long anthers.

Ssp. bessaribicus distributes in coasts of the Black Sea from Bulgaria to Crimea, Sea of Azov, Aegean and N.E. Mediterranean Sea. Corotype of subspecies is Mediterranean. Habitat: Seashores, sandy soil or sandy loam.

5.2.4.2.2 Cytology, Cytogenetics and Evolution

Cauderon (1958), Moustakas and Coucoli (1982), and Moustakas (1993) studied the chromosome number and karyotype of ssp. bessarabicus. All accessions studied had 2n = 2x = 14, with a symmetric karyotype, consisting of four metacentric pairs and three sub-metacentric pairs. Two chromosome pairs are SAT-chromosomes, one metacentric pair carries a large satellite and one sub-metacentric pair has a small satellite. The karyotype of ssp. bessarabicus is similar to the karyotype of diploid E. elongatus (Cauderon 1958; Runemark and Heneen 1968; Moustakas and Coucoli 1982; Wang 1985b; Hsiao et al. 1986). Yet, the chromosomes of E. farctus are larger than those of E. elongatus and differences in morphology of the SAT-chromosomes exist between the two taxa (Heneen 1962); the pair with large satellites in E. elongatus has more median centromeres than the equivalent pair in E. farctus, whose short arm is divided into two equal parts by the secondary constriction. The second pair of ssp. bessarabicus has small satellites, which are somewhat larger than those of the equivalent pair in diploid E. elongatus. Endo and Gill (1984) also found a distinction between the two species in C-banding patterns. From their cytological analyses, Heneen and Runemark (1972), Moustakas and Coucoli (1982), and Moustakas (1993) concluded that the seven pairs of E. bessarabicus show striking similarity in chromosome size and centromere positions with the seven largest pairs of tetraploid E. boreali-atlanticus and the fourteen pairs of subsp. rechingeri.

5.2.4.2.3 Crosses with Oter Triticineae Species

Wang (1985a) crossed E. bessarabicus with E. elongatus, and, from the relatively high chromosomal pairing in their F1 hybrid, concluded that the chromosomes of the two taxa show a high degree of homology, which counterweighed the C-banding differences reported by Endo and Gill (1984). Hsiao et al. (1986) assumed that the C-banding differences of the two taxa are due to structural rearrangements.

McGuire (1984) studied chromosomal pairing at meiosis of F1 triploid hybrids between Elymus curvifolius (2n = 4x = 28; genome Eb1Eb1Eb2Eb2) x ssp. bessarabicus and observed an average of 3.71 univalents, 2.29 rod- and 1.82 ring-bivalents, 2.64 trivalents, and 0.29 quadrivalents. These pairing data indicate that the two subgenomes of E. curvifolius are closely related to the genome of ssp. bessarabicus and that the tetraploid had an autopolyploid origin.

Similarly, McGuire (1984) studied chromosomal pairing at meiosis of the F1 triploid hybrid between Elymus elongatus ssp. flaccidifolius (=Elymus scirpeus; Elytrigia scirpea) (2n = 4x = 28; genome Ee1Ee1Ee2Ee2) and ssp. bessarabicus (genome EbEb). Mean chromosome pairing included 5.14 univalents, 1.28 rod- and 3.86 ring-bivalents, 1.47 trivalents, 0.11 quadrivalents, and 0.1 pentavalents. Also this hybrid provides evidence that the two Ee subgenomes of E. flaccidifolius are homologues and are related to genome Eb of ssp. bessarabicus. This is in accord with the finding of Wang (1985a, b), who demonstrated that genome Eb of ssp. bessarabicus is closely related to the genome Ee of diploid E. elongatus.

Wang (1988) produced diploid hybrids between ssp. bessarabicuus and Elymus spicatus (=Pseudoroegneria spicata) (genome StSt), as well as with Secale strictum (formerly S. montanum) (genome RR). Meiotic chromosome pairing of the F1 hybrid spicatus x bessarabicus averaged 4.34 univalents, 2.77 rod- and 1.42 ring-bivalents, 0.24 trivalents, and 0.14 quadrivalents. On the other hand, chromosome pairing of the F1 hybrid bessarabicus x strictum included 11.05 univalents, 1.22 rod- and 0.04 ring-bivalents, 0.13 trivalents, and 0.01 quadrivalents. These meiotic data suggest that the ST genome of E. spicatus and the Eb genome of ssp. bessarabicus are more closely related to each other than Eb is with the R genome of Secale.

Alonso and Kimber (1980) and Sharma and Gill (1983) produced an F1 hybrid between Triticum aestivum cv. CS and diploid ssp. bessarabicus. The morphology of the hybrid was closer to wheat, but showed intermediate expression of some traits. The hybrid seemed not to be perennial, since it produced few tillers. Since hybrids between bread wheat and tetraploid or hexaploid E. farctus exhibit chromosomal pairing that result from autosyndesis of Elymus chromosomes, the study of the relationships between the Eb genome of ssp. bessarabicus and the subgenomes of bread wheat is incomprehensible. Hence, the relationships between the two species can be studied directly in hybrids with diploid E. farctus. A mean 0.2 bivalents per cell (always rod) was observed at meiosis of the hybrid; most of the cells had 28 univalents (Alonso and Kimber 1980). Somewhat higher chromosomal pairing was observed in such a F1 hybrid by Sharma and Gill (1983), with a mean 0.83 rod bivalents, 0.04 ring bivalents, and 0.01 trivalents per cell. The very low pairing in these hybrids indicates that genome Eb of diploid E. farctus is only distantly related to the A, B, and D subgenomes of bread wheat. Moreover, the Eb genome of diploid E. farctus did not induce homoeologous pairing of wheat chromosomes in the hybrids containing the Ph1 gene.

5.2.4.3 Ssp. rechingeri (Runemark) Melderis (2n = 4x = 28)

Ssp. rechingeri (Runemark) Melderis [Syn.: Agropyron rechingeri Runemark; Elymus rechengeri Runemark in Runemark & Heneen; Elytrigia rechingeri (Runemark) Holub] is perennial, without rhizomes, with culms usually swollen at base, short ligule (c. 0.5 mm), 2–10 cm long spikes, rachis internodes that are usually shorter than the lower spikelets, 5–12 mm, 4–5 veined glumes, and palea ciliate in upper half of keel only.

Ssp. rechingeri is a tetraploid (2n = 4x = 28) cross-fertilizing taxon. The karyotype of ssp. rechingeri shows great similarities to the karyotype of E. furctus ssp. boreali-atlanticus (2n = 4x = 28) (Heneen 1977). Meiosis is generally normal, but some asynapsis and multivalent formation at first meiotic metaphase and separation difficulties at first anaphase have been noted (Heneen 1977). Chromosomal polymorphism, manifested by changes in the three pairs of satellite chromosomes, was observed both between and within populations (Heneen 1977). Some offspring plants also show structural and numerical chromosome deviations. Fertility is lower in crosses between populations, indicating some genetic differences between populations (Heneen 1977).

This subspecies grows in small, isolated populations, mainly on the Aegean islands. It also occurs in west Turkey, coasts of Greece, Crete, and the Mediterranean coast of Egypt. Corotype: E. Mediterranean element.

5.2.4.4 Ssp. boreali-Atlanticus (Simonet & Guinochet) Melderis (2n = 4x = 28)

5.2.4.4.1 Morphological and Geographical Notes

Ssp. boreali-atlanticus (Simonet & Guinochet) Melderis [Syn.: Agropyron junceum subsp. boreali-atlanticum Simonet & Guinochet; A. junceiforme (Á. & D. Löve) Á. & D. Löve; Elytrigia juncea subsp. boreo-atlantica Hyl.; Thinopyrum junceum (L.) Á. Löve] is perennial with long rhizomes, up to 55 cm high, fragile, glabrous culms, erect or slightly curved, 5.0–14.0 cm long, spike with 4–10 nodes and 16–23 mm long internodes, spikelets longer than internodes, with 3–5 flowers, rachis breaking up between each spikelet at maturity (wedge-type disarticulation), 10–16 mm long glumes, 10–17 mm long awned lemma, sometimes with very short awns.

Ssp. boreali-atlanticus occurs in the northern part of the distribution of the species, from Portugal to Finland, however, its stands are concentrated mainly between Portugal and Germany, whereas in the east, their numbers decrease rapidly.

5.2.4.4.2 Cytology, Cytogenetics and Evolution

The chromosome number given for ssp. boreali-atlanticus from different areas of distribution is 2n = 4x = 28 (Peto 1930; Simonet and Guinochet 1938; Östergren 1940a; Moustakas et al. 1986; de V Pienaar et al. 1988). The karyotype is symmetric, exhibiting only two, out of the four possible, secondary constrictions in somatic metaphases (Simonet 1935). The genome of ssp. boreali-atlanticus was formulated as Eb1Eb1Eb2Eb2 (=J1J1J2J2), assuming an autopolyploidy origin that later sustained some differentiation in the two subgenomes (Cauderon 1958). Alternatively, Moustakas et al. (1986) and de V Pienaar et al. (1988), on the basis of genome analysis of the whole polyploid complex of E. farctus, concluded that ssp. boreali-atlanticus is a segmental allopolyploid containing a basic genome E (=J) (more or less modified at different polyploid levels of the entire complex). In line with this suggestion, Liu and Wang (1993) proposed EbEbEeEe as the genomic formula of this subspecies.

Östergren (1940a) reported that meiosis in ssp. boreali-atlanticus is regular; 14 bivalents were observed in almost all cells, whereas quadrivalents were very rare. Few individuals were found to be heterozygous for a paracentric inversion (Östergren 1940a).

5.2.4.4.3 Crosses with Other Triticineae

McGuire (1984) analyzed the F1 hybrid between ssp. boreali-atlanticus (genome EbEbEeEe) x E. curvifolius (genome Eb1Eb1Eb2Eb2) and observed mean chromosomal pairing of 3.00 univalents, 0.93 rod- and 1.57 ring-bivalents, 1.36 trivalents, 1.79 quadrivalents (0pen), 1.14 quadrivalents (close), and 0.79 pentavalents. The high frequency of quadrivalents and pentavalents resulted presumably, from the close relatedness of Eb and Ee. Östergren (1940a) studied chromosomal pairing at meiosis of the F1 hybrid between ssp. boreali-atlanticus and Elymus repens (2n = 6x = 42; genome StStStStHH). Average chromosomal pairing included 11.8 univalents and 11.6 bivalents. From the pairing data, Östergren (1940a) concluded that ssp. boreali-atlanticus is not entirely autopolyploid.

Östergren (1940b) crossed T. turgidum subsp. turgidum (2n = 4x = 28; genome BBAA) with ssp. boreali-atlanticus, and obtained at the first meiotic metaphase of the F1 hybrid an average chromosomal pairing of 18.4 univalents and 4.8 bivalents; most of which were rod shaped. Chromosome pairing in haploid tetraploid wheat was found by Kihara (1936), Lacdena and Ramos (1968) and Jauhar et al. (1999), to be very low and therefore, most of the pairing in the hybrid (genome BAEbEe) was autosyndetic of Eb and Ee chromosomes. This pairing, that took place in the presence of the Ph1 gene of ssp. turgidum, again indicates that the two subgenomes of ssp. boreali-atlanticus, namely, Eb and Ee, are closely related.

5.2.4.5 Ssp. farctus (Viv.) Runemark Ex Melderis (2n = 6x = 42)

5.2.4.5.1 Morphological and Geographical Notes

Ssp. farctus (Viv.) Runemark ex Melderis [syn.: Triticum farctum Viv.; Agropyron farctum Viv.; Agropyron junceum (L.) Beauv.; Agropyron junceum ssp. mediterraneum Simonet & Guinochet; Thinopyrum junceum (L.) Á. Löve; Elytrigia juncea (L.) Nevski] is perennial, usually with long-creeping rhizomes, 50–80 cm high plants, with rigid, thick culms, glaucous, usually rolled leaves, 15–35 cm long, erect spike, more or less fragile rachis, breaking at maturity above each spikelet (wedge type disarticulation), 10–25 mm long, 5–9 flowered, glabrous, laterally compressed, awnless spikelets, appressed to rachis, 10–18 mm long, lanceolate, 6–12 veined, asymmetrically keeled glumes, 10–20 mm long lemma, keeled towards the apex, palea ciliate nearly along the entire length of keels, 10–12 mm long anthers (Fig. 5.1c).

Distribute along the coasts of the Mediterranean Sea. Grows near seashores, on sandy soil or sandy loam. Chorotype: Mediterranean element.

5.2.4.5.2 Cytology, Cytogenetics and Evolution

Simonet (1935) and Simonet and Guinochet (1938) reported that the number of chromosomes in this subspecies is 2n = 6x = 42, and its genome was formulated by Cauderon (1958) as Eb1Eb1Eb2Eb2EeEe (=JIJ1J2J2EE). Charpentier et al. (1986) studied the degree and pattern of chromosomal pairing at the first meiotic metaphase of lines of ssp. farctus collected from several sites along the Israeli Mediterranean coast. Despite the relatedness between its subgenomes, ssp. farctus exhibited almost complete bivalent pairing, i.e., 21 bivalents; multivalents were rarely found.

A similar pattern of bivalent pairing with rare multivalents is characteristic of other autopolyploid or segmental allopolyploid species of Elymus, e.g., E. hispidus (=Agropyron intermedium) (genome EeEeEeEeStSt) (Cauderon 1958, 1966), and of tetraploid E. elongatus (genome Ee1Ee1Ee2Ee2) (Charpentier et al. 1986). The almost strict bivalent pairing in these auto-allopolyploids may be brought about by a gene system that induces bivalentization by restricting pairing to fully homologous chromosomes. A similar gene system, determining a bivalent rather than a multivalent pattern of pairing, was described for the autotetraploid Avena barbata (Ladizinsky 1973) and in an autotetraploid line of Aegilops longissima (Avivi 1976).

5.2.4.5.3 Crosses with Other Triticineae Species

Charpentier et al. (1986) analyzed chromosomal pairing at the first meiotic metaphase of F1 hybrids generated from lines of ssp. farctus and the bread wheat cultivar Chinese Spring. In the presence of the wheat Phl gene, chromosome pairing in the hybrid (genome BADEb1Eb2Ee) included 25.88 univalents, 5.43 rod- and 0.87 ring-bivalents, 1.08 trivalents, and 0.06 quadrivalents. Most of the pairing was autosyndetic of the Eb1 and Eb2 chromosomes of ssp. farctus. Based on this pattern of pairing, Charpentier et al. (1986) concluded that the corresponding chromosomes of the Eb1 and Eb2 subgenomes are distant homologues.

The seed set in the crosses CS x ssp. farctus was rather low (1.8–7.0%), and most seeds were very shriveled, with a poorly developed endosperm. Embryos were well differentiated, and when cultured on Orchid agar, 19–33% of the embryos germinated. The F1 plants exhibited strong heterosis, i.e., they were tall (1.20–1.60 m) and had vigorous tillering. All plants were perennials with a non-brittle rachis.

The small seed set in crosses between bread wheat (as female) and hexaploid E. farctus, shrivelling of most of the seeds, and their poor germination, as well as the fact that the F1 plants were completely male sterile with anthers that failed to dehisce, attest to the existence of significant strong chromosomal and/or genetic barriers between these two taxa.

Eleven disomic addition lines, and nine partial amphiploids have been obtained from hybridization between bread wheat cv. CS and the hexaploid subspecies of E. farctus (Charpentier 1992). The genomic structure of the disomic addition lines consists of 21 pairs of wheat chromosomes, plus an additional pair of ssp. farctus.

McArthur et al. (2012) used FISH to characterize thirteen disomic addition lines of chromosome pairs of ssp. farctus that were added to bread wheat cv. CS. Several lines were those produced by Charpentier (1992). Five disomic addition lines (AJDAj5, 7, 8, 9, and HD3508) were identified to contain a farctus chromosome that corresponded to homoeologous group 1. Addition lines AJDAj2, 3, and 4 contained a farctus chromosome that corresponded to homoeologous group 2, HD3505 to group 4, AJDAj6 and AJDAj11 to group 5, and AJDAj1 probably to group 6. Several ssp.. farctus chromosomes in the addition lines were found to contain genes for resistance to Fusarium head blight, tan spot, Septoria nodorum blotch, and stem rust (Ug99 races).

5.2.5 Phylogenetic Relationships of St, Ee and Eb Genome Elymus Species with Other Triticineae Species

Elymus s. l. is the largest, most morphologically diverse and most widely distributed genus in the Triticeae, consisting of approximately 150 diploid and polyploid species (Table 2.1). Cytogenetic analyses have shown that there are eight basic genomes within Elymus, namely, St, E (either Ee or Eb), L, Xp, H, P, W, and Y. The St, L, and Xp genomes exist in diploid species that grow in the Temperate-Arctic region whereas other diploids with St genome and those with E genome grow in the Mediterranean-Central Asiatic region.

The St genome of several diploid species of Elymus (Table 5.1) is closely related to Ee and Eb genomes of diploid E. elongatus and E. farctus, respectively (Wang 1989, 1992). Bieniek et al. (2015) found that the nucleotide sequences at three chloroplast loci (matK, rbcL, trnH-psbA) are almost identical in the diploid Ee, Eb and St taxa, with only one substitution within the matK gene, differentiating genome Eb from the Ee and St genomes. Petersen and Seberg (1997), based on rpoA sequence data, and Wang and Lu (2014), based on a literature review, corroborated the very close relationship between the Ee, Eb and St genomes. This close relationship was also substantiated by study of the sequences of a gene encoding plastid acetyl-CoA carboxylase (Fan et al. 2007). A study using 5S rDNA further strengthened the reported close relationship between St and Eb (Shang et al. 2007). The St genome is also closely related to the P genome of Agropyron species (Wang 1992). The St and Eb genomes were shown to be more closely related to the R genome of Secale than to the Vv genome of Dasypyrum villosum (Shang et al. 2007). The Ns, H, and R genomes are remotely related to the E-St-P cluster (Wang 1992). Since the St genome exists in primitive diploid species of Elymus and in many allopolyploid species of this genus, it is assumed that genomes Ee and Eb evolved from St.

The St and E genomes are two important basic genomes in the perennial species of the Triticeae. In addition to their existence in diploids, the St and E genomes also exist in almost all polyploid (allopolyploid and autopolyploids) Elymus species, whereas the H, P, W, and Y genomes exist only in allopolyploid species (Dewey 1984; Wang et al. 1995). The H genome originated from one of the species of Hordeum, the P from Agropyron, the W from Australopyrum, and the origin of the Y genome is unknown (Dewey 1984; Yan et al. 2011; Petersen et al. 2011). Liu et al. (2006) and Okito et al. (2009) postulated a common origin of St and Y, whereas Dewey (1984) assumed that the donor of the Y genome is a distinct, yet undiscovered or extinct, diploid Asian species. Sun et al. (2008), Sun and Komatsuda (2010), and Yan et al. (2011)) supported Dewey’s (1984) view, and separated the St and Y genomes into distinct groups. Studies on the chloroplast DNA of Elymus showed that the St genome of diploid Elymus species is the maternal donor of the St genome in all the allopolyploid Elymus species (Mason-Gamer et al. 2002; Liu et al. 2006; Hodge et al. 2010; Yang et al. 2017).

Recent molecular studies (Mason-Gamer et al. 2010; Sun et al. 2008; Yu et al. 2008; Sun and Komatsuda 2010; Yan and sun 2011) showed the presence of several distinct clades within the diploid St genome Elymus species, but analysis of various DNA sequences suggested different combinations of clades. For instance, E. libanoticus and E. tauri have similar sequences and form one clade, while the other diploid St genome species form another clade (Mason-Gamer et al. 2010; Sun et al. 2008; Sun and Komatsuda 2010; Yan and Sun 2011). Yu et al. (2008) found that E. libanoticus, diploid E. tauri and diploid E. spicatus are more closely to one another than they are to E. stipifolius and E. reflexiaristatus. Yan et al. (2011) grouped the Eurasian St genome species E. libanoticus, E. reflexiaristatus and diploid E. tauri into one clade and the North American E. spicatus into a separate clade. Evidently, there are great discrepancies between various phylogenetic studies performed on this group of diploid species.

By analyzing the sequence diversity in the internal transcribed spacer (ITS) region of nuclear ribosomal DNA, Di̇zkirici et al. (2010) found that all the species of Elymus s. l. fall into one clade. Fan et al. (2013), using two single-copy nuclear gene (Acc1 and Pgk1) sequences, found that that the St genome Elymus species are closely related to the Ee genome of E. elongatus. Yang et al. (2017) using nuclear internal-transcribed spacer and the chloroplast trnL-F sequences, analyzed phylogenetic relationships among Elymus and related genera, and obtained four major clades: (1) the St/E clade, comprised all of the St and E genome species of Elymus; (2) the P/W clade, including Agropyron and Australopyrum; (3) the Ns clade which included Psathyrostachys; and (4) the H clade, which consisted of Hordeum species. The results suggested that: (a) diploid St genome species were the maternal donors of St in allopolyploid species in Elymus s.l. and that the trnL-F sequences are highly similar among these species; (b) the trnL-F sequences of Agropyron species and Australopyrum species are similar, and the P genomes are closely related to the W genome; and (c) the trnL-F sequences of species with the H or Ns genomes diverged considerably from that of species with the St, E, P, or W genomes.

Petersen and Seberg (1997) and Escobar et al. (2011) classified the diploid Triticeae into five major clades: (1) Psathyrostachys; (2) Hordeum; (3) Elymus; (4) Agropyron (includes Australopyrum)–Eremopyrum: and (5) Aegilops–Triticum–Secale–Taeniatherum. These major clades were defined recently also on the basis of the nuclear phosphoglycerate kinase (PGK) gene that codes for plastid PGK isozyme (Adderley and Sun 2014). Similarly, Bieniek et al. (2015), studying phylogenetic relationships among the Triticeae diploid species through analysis of three chloroplastic genes, also supported the finding of the above five clades. The Elymus clade is the largest clade, containing the diploid species of the genus Elymus s. l. (Melderis 1980), namely, the St-genome species, E. elongatus, and E. farctus.

Hsiao et al. (1995) and Kellogg et al. (1996) considered T. monococcum to be the sister group to Elymus elongatus. Based on internal transcribed spacers (ITS) of the nuclear rDNA sequences, Hsiao et al. (1995) reported that Ee and Eb jointly clustered with subgenomes A, B, and D. In accord with this finding, Liu et al. (2007), using genomic hybridization (both Southern and in situ hybridization), also showed that the St and Eb genomes are very closely related to the A, B and D subgenomes of common wheat, but are more closely related to the D subgenome than to the A and B subgenomes. These observations provide a possible explanation as to why most of spontaneous translocations and substitutions occurring in the common wheat—E. elongatus ssp. ponticus, usually take place in the D genome, some in the A subgenome and rarely in the B subgenome. In accord with the above findings, both genomes Ee and Eb were found to be closely related to subgenomes A, B, and D of the wheat group (reviewed in Wang and Lu 2014) and thus, the latter genomes may derive from the E genome(s). Interestingly, from their study of two single-copy nuclear gene (Acc1 and Pgk1) sequences, Fan et al. (2013) concluded that the relationship between Elymus farctus subsp. bessaribicus (genome Eb) and Triticum/Aegilops is closer than between Elymus elongatus (genoe Ee) and Triticum/Aegilops.

5.3 Agropyron Gaertner Senso Stricto

5.3.1 Taxonomic Notes

The genus Agropyron Gaertner s.l. (the name Agropyron is derived from the Greek terms ‘agros’ meaning field and ‘puros’ meaning wheat) was one of the largest genera in the Triticeae, including more than 100 species (Gaertner 1770). However, Nevski (1933, 1934a) divided the perennial species of Agropyron to four genera, Agropyron s. str., Elytrigia Desv., Roegneria C. Koch and Anthosachne Steud. Gould (1947) included Agropyron s. str. in Elymus, while Runemark and Heneen (1968) further expanded the generic concept of Elymus to also include Elytrigia, Roegneria, Aneurolepidium, Terella, Hystrix (Asperella) and Sitanion. Melderis (1978, 1980, 1985a) followed this expansion of the generic concept of Elymus but retained Agropyron s. str. as a separate genus for the crested wheatgrasses.

Agropyron s. str. is currently considered to be a small genus, including only 10–15 species (Dewey 1984; Sakamoto 1991), all of which are the crested wheat grasses (Nevski 1933; Hitchcock 1951; Tzvelev 1983). The genus is morphologically well characterized by its distinctly keeled glumes, its short rachis internodes, spikelets divergent from the rachis at an angle of more than 45° and typically pectinate spikes. It should be emphasized that Nevski (1933, 1934a) treated the genus Agropyron on the basis of morphological features only; later cytogenetic analyses, mainly genome analyses, confirmed the validity of Nevski’s treatment. Genome analyses revealed that Agropyron is genomically homogeneous with diploid, tetraploid, and hexaploid taxa, all displaying the P genome (Löve 1984), thus rendering it the only complex with just one genome and genomically distinct from other taxa of the Agropyron-Elymus complex (Melderis 1978; Löve 1984). The P genome also exists in several polyploid species of Elymus, but has a negligible influence on the morphology of these species (Assadi and Runemark 1995). Therefore, it seems reasonable to maintain Agropyron as a separate genus. The generic concept of Nevski was accepted by many taxonomists (e,g., Tzvelev 1973, 1976; Melderis 1978, 1980, 1985a; Clayton and Renvoize 1986; Watson and Dallwitz 1992; Assadi and Runemark 1995).

The species of Agropyron s. str. are perennials, rhizomatous or caespitose and cross-fertilizing (Melderis 1978). They have firm, straight, 20–70 cm long culms, spikes with hairy and tough rachis (rarely fragile), that do not disarticulate at maturity, with short internodes, solitary, sessile, pectinate and strongly laterally compressed, spikelets with 2–12 florets at each node of the rachis. The rachilla disarticulates at maturity above the glumes and beneath the florets (floret-type disarticulation). In this type of disarticulation, lemma and palea fall off, while glumes persist, firmly attached to the rachis. The glumes are boat-shaped, with a prominent keel, 1–2 inconspicuous lateral veins and a wide margin, lemma are membranous, 5-vined and keeled and both glumes and lemmas are awnless, or lemma have a short awn. Their anthers are 3.5–5.0 mm long.

The Agropyron species form an autopolyploid series consisting of diploids, tetraploids and hexaploids (Dewey and Asay 1975; Dewey 1967, 1982, 1983; Melderis 1978; Assadi 1995; Knowles 1955; Jensen et al. 2006). The autopolyploid series is based on the P genome of diploids A. cristatum and A. mongolicum (Knowles 1955; Dewey 1967, 1969, 1984; Löve 1982).

Much confusion prevailed in regard to the number of species in this genus (Cabi 2010). Tzvelev (1976) recognized 10 species in Agropyron, Yilmaz et al. (2014) reported that this genus contains ± 15 species, and Sakamoto (1991) assumed that it contains 19 species, with a distribution of 5 diploids, 13 tetraploids and 1 hexaploid. The taxonomy of the species has been complicated due to extensive interspecific introgressive hybridizations, as well as by the fact that many of the interspecific hybrids are fertile (Knowles 1955; Dewey 1983; Asay and Dewey 1979).

Thirteen Agropyron species, their synonyms, common name, ploidy level, genome formula, and geographic distribution, are presented in Table 5.3. The species grow in a wide range of habitats, including steppe-like habitats, mountains, saltmarshes, and seashores, on sands or stony mountain slopes, but primarily in the grasslands of Eurasia, at altitudes ranging from a few meters to more than 5000 m above sea level (Dewey and Asay 1975; Tzvelev 1983; Yang et al. 2014). Due to their environmental adaptability, tolerance to aridity and infertile soils, resistance to pest and disease damage, and palatability, several Agropyron species have been used extensively as an ecological resource and for feed research (Dewey 1984).

Table 5.3 Species of Agropyron senso stricto, their synonyms, common name, ploidy level, genome formula and geographic distribution

5.3.2 Agropyron cristatum—The Genus Type

The typification of the genus Agropyron was based on A. cristatum (L.) Gaertn. (Tzvelev 1976), which was first described as Triticum cristatum (Linnaeus 1753) and, later classified as Agropyron by Gaertner (1770). This species (Fig. 5.3a) comprises a polyploid complex of diploid, tetraploid and hexaploid cytotypes. It is a very polymorphic species, and was subdivided on the intraspecific level to the following eight subspecies: Ssp. cristatum (contains diploid, tetraploid and hexaploid cytotypes); Ssp. pectinatum (M. Bieb.) Tzvelev (contains diploid and tetraploid cytotypes); Ssp. sabulosum Lavrenko; Ssp. brandzae (Panfu & Solacolu) Melderis; Ssp. ponticum (Nevski) Tzvelev; Ssp. sclerophyllum Novopokr. ex Tzvelev; Ssp. bulbosum (Boiss.) Á. Löve; Ssp. incanum (Nábĕlek) Melderis (hexaploid cytotype). The species A. deweyi Á. Löve, A. incanum (Nábĕlek) Tzvelev, A. imbricatum Roem. & Schult., and A. bulbosum Boiss. are considered synonyms of A. cristatum.

Fig. 5.3
figure 3

Species of the Agropyron-Eremopyrum-Henrardia clade: a A spike of Agropyron cristatum (L.) Gaertn.; b A plant and a spike of Eremopyrum distans (K. Koch) Nevski; c A plant and a spike of diploid Eremopyrum buonapartis (Spreng.) Nevski

A. cristatum is facultative allogamous (Cabi 2010). It is native to Europe and Asia, growing from Portugal in the west to China in the east. It was introduced from Russia and Siberia to North America in the first half of the twentieth century, where it was often used as forage and in erosion control. However, currently it is considered a weed in the USA and Canada. Agropyron cristatum is the most widely distributed species of Agropyron, it exhibits significant inter- and intra-population variation in maturity time, height, texture, rhizome development, fertility, and seed size. It is a xerophytic species, which probably originated from central Asia, and is indigenous to this area, including parts of the former USSR, China, Afghanistan, Turkey, and Iran (Dewey and Asay 1975; Tzvelev 1976; Cabi 2010). It grows in a variety of steppes and steppe-like habitats. In its native range, it is frequently found on carbonate slopes in the forest steppe belt, on dry terraces, and in steppe woodlands (Cabi 2010). It grows from 1500 to about 2200 m above sea level and prefers well-drained, deep, loamy soils. It tolerates frost, drought, and salinity and prefers moderately alkaline conditions (Cabi 2010). It is best adapted to areas with poor precipitation (200–400 mm annual rainfall).

The broad pectinate spiked A. cristatum (L.) Gaertner contains three cytotypes (2n = 14, 28 and 42 chromosomes) (Araratian 1938); Dewey 1982, 1983; Dewey and Asay, 1982; Yang et al. 2014), all three of which occur in Iran (Dewey and Asay (1975). Tetraploids are the most common, exhibit high morphological variation, and are found throughout the entire distribution area. Hexaploid populations occur only in the Azerbaijan province in northwestern Iran. The diploid cytotype is rare in northwestern Iran but is known from Europe and other regions. The polyploid races behave cytologically as autoploids. Heterozygous chromosome interchanges are common in the tetraploids, and aneuploidy is uncommon.

Tetraploid A. cristatum has been used in wheat breeding for many years, has been hybridized and chromosomal addition, substitution and translocation lines were formed (Li and Dong 1991; Jensen and Bickford 1992; Yang et al. 2010).

5.3.3 Cytology, Cytogenetics and Evolution

Agropyron cristatum and A. mongolicum are the only diploid taxa in the genus Agropyron s. str. Both are cross-pollinating species, but differ morphologically: A. cristatum (an Eurasian species) has broad pectinate spikes, whereas A. mongolicum (an East Asian species) has narrow linear ones (Hsiao et al 1986). Schulz-Schaeffer et al. (1963), McCoy and Law (1965), Taylor and McCoy (1973), Endo and Gill (1984) Hsiao et al. (1986) and Yang et al. (2014) described the karyotype of A. cristatum; that of A. mongolicum was reported by Hsiao et al. (1986). The chromosomes of these two diploids are all metacentric or sub-metacentric, relatively large and with a symmetric karyotype (Hsiao et al. 1986; Yang et al. 2014). The karyotypes are very similar and differ slightly in the centromere positions of chromosomes 5 and 7. Despite differences in plant morphology, the two species hybridize readily (Dewey and Hsiao 1984). The F1 hybrids showed reasonably good chromosome pairing at meiosis, with an average of five to six bivalents per cell. They probably differ only in minor structural rearrangements of certain chromosomes (Hsiao et al. 1986).

Yang et al. (2014) studied the karyotype of six Agropyron cristatum populations distributed from Northern Europe (Sweden) to Southwest Asia (Iran). The European (Swedish and Bulgarian) populations were diploids, the two populations from the Russian Federation and one from Iran were tetraploids, while the second Iranian population was hexaploid. Differences in the centromere position in a number of chromosomes and in the relative length of the longest chromosome indicated the existence of karyological variation among these populations (Knowles 1955; Yang et al. 2014). Satellites were not observed in all populations (Yang et al. 2014). The karyotypes of the two Iranian populations were different from those reported by Hsiao et al. (1986) and Hsiao et al. (1989), who showed that the diploid cytotype had two small satellites on the fourth and the sixth chromosomes. The minute satellites appear as small dots visible only at early metaphase. The occurrence of minute satellites on two chromosome pairs of A. cristatum, was also reported by Knowles (1955), McCoy and Law (1968), Watson and Dallwitz (1992) and Endo and Gill (1984). The results reported by Yang et al. (2014) supported the relationship between distribution and ploidy levels (Dewey 1984; Dewey and Asay 1975; Yen and Yang 2006), with diploids distributed in small and scattered areas, tetraploids showing a universal distribution, and hexaploids distributed narrowly in Northeastern Turkey and Northwest Iran.

The karyotype of tetraploid A. cristatum was compared to that of a colchicine-induced autotetraploid of diploid A. cristatum. The idiograms of the two tetraploid taxa were strikingly similar, suggesting that the tetraploid cytotype evolved through autopolyploidy (Taylor and McCoy 1973).

Peto (1930) determined the chromosome number of A. desertorum as 2n = 4x = 28. Knowles (1955) determined chromosome numbers in A. desertorum, A. sibiricum, A. fragile, and A. michnoi, all of which are tetraploids (2n = 4x = 28). Although several works reported no satellite chromosomes in A. desertorum (Sarkar 1956; Schultz-Schaeffer and Jurasits 1962), McCoy and Law (1965) reported the existence of four to six such chromosomes in a number of clones of A. desertotum.

Endo and Gill (1984) used the acetocarmine-Giemsa C-banding technique to study heterochromatin distribution in somatic chromosomes of two diploid Agropyron taxa, A. cristatum, and A. imbricatum (a synonym of A. cristatum). While most cross-pollinating Triticeae species show large terminal C-bands and a high level of C-band polymorphism, A. cristatum exhibited only small to medium terminal bands in most of the chromosomes with low C-band polymorphism. Both A. cristatum and A. imbricatum showed gross similarity in C-banding patterns, although small differences were discernible. This confirms Dewey’s (1983) assumption that A. imbricatum may carry a genome similar to the P genome of A. cristatum (Endo and Gill 1984).

Yousofi and Aryavand (2004) used flow cytometry to determine the ploidy levels of six different populations of A. cristatum in Iran. The mean nuclear 2C DNA content ranged from 26.41 to 27.56 pg for two varieties of A. cristatum ssp. pectinatum (five populations), and 43.47 pg for Agropyron cristatum ssp. incanum (one population). These results were supported by chromosome counting; chromosome number in the tetraploid populations varied from 28 to 31, and in the hexaploid population from 35 to 44. The frequency of aneuploidy was lower (3–4%) in tetraploids and much higher (about 18.9%) in the hexaploid population (Yousofi and Aryavand 2004).

Mean 2C DNA content was 26.26 pg for three tetraploid A. cristatum populations and 27.50 pg for the other two tetraploid populations of this species, a difference of 1.24 pg. Small differences in DNA content at the intraspecific level may be due to the presence or absence of accessory chromosomes (B-chromosome) (Vogel et al. 1999), or due to the aneuploidy observed within these populations. Yousofi and Aryavand (2004) argued that the differences in DNA content between the sub-specific taxa exceeds the probable DNA content of accessory chromosomes, since the average 2C DNA content of an Agropyron chromosome has previously been reported to be about 1 pg (Vogel et al. 1999).

Vogel et al (1999) determined the mean DNA content of three diploid accessions of A. cristatum, but did not analyze the tetraploid and hexaploid cytotypes. The data of Vogel et al. (1999) and those of Yousofi and Aryavand (2004) show that compared to other Triticeae species, the size of the haplome genome of A. cristatum, the P genome, and probably of the other Agropyron species, is intermediate to small.

Dewey and Asay (1982) hybridized three morphologically distinct taxa of diploid A. cristatum. Mean chromosome pairing at meiotic first metaphase of the three F1 hybrid combinations included a range of 1.38–2.25 univalents, 5.05–5.83 bivalents, 0.03–0.52 trivalents, and 0.005–0.18 quadrivalents. The pairing data indicated that the three diploids contain the same basic genome, which differ by structural rearrangements of some chromosomes. The moderately high sterility in the F1 hybrids serves as a genetic barrier but does not preclude gene flow among the diploids. Hence, the diploid taxa were identified as three different subspecies of A. cristatum (Dewey and Asay 1982).

Agropyron mongolicum was hybridized with the diploid cytotype of A. cristatum (Hsiao et al. 1989; Chen et al. 1992b). Chromosome pairing at first meiotic metaphase in the F1 hybrid averaged 1.40 univalents, 5.59 bivalents, 0.35 trivalents, and 0.09 quadrivalents per cell (Hsiao et al. 1989) and 0.22 (0–2) univalents, and 0.79 (0–3) rod- and 6.10 (4–7) ring-bivalents (Chen et al. 1992b). The Fl hybrids were partially fertile. The presence of seven bivalents in many pollen mother cells (PMCs) of the F1 hybrid A. cristatum x A. mongolicum indicated that the two diploid species contain the same basic P genome. However, the occurrence of multivalents revealed that the genomes of these two diploids differ by a reciprocal translocation(s). These two diploids are the likely source of morphological and cytological variation in the tetraploid species of Agropyron (Hsiao et al. 1989). In accord with this view, Mellish et al. (2002), using AFLP markers, concluded that A. desertorum is an allopolyploid of A. cristatum and A. mongolicum.

Assadi (1995) analyzed chromosome numbers and meiotic behavior in A. cristatum ssp. incanum (2n = 4x = 42). The existence of multivalents at the first meiotic metaphase of this subspecies, which averaged 2.73 quadrivalents and 0.64 hexavalents per cell, indicated that this taxon is an autohexaploid (Assadi 1995).

Hybrids between the diploid cytotype of A. cristatum and tetraploid A. desertorum showed a high frequency of trivalents at the first meiotic metaphase (Knowles 1955), showing considerable homology between the cristatum and desertorum genomes, thus indicating an autoploid origin of A. desertorum (Knowles 1955). All crosses between the tetraploid species A. sibiricum, A. fragile, and A. michnoi with A. desertorum produced fertile hybrids (Knowles 1955), implying phylogenetic closeness and likely autoploid origin of these tetraploid species. Myers and Hill (1940) observed an average quadrivalent frequency of 3.8 (3.4–4.5) per PMC of tetraploid A. cristatum, suggesting the autotetraploid derivation of this cytotype.

Artificial crosses between diploid and hexaploid cytotypes of A. cristatum were not successful (Dewey 1969), but those between tetraploid A. desertorum and hexaploid A. cristatum produced viable and highly fertile seeds.

Dewey and Pendse (1968) crossed Agropyron desertorum and an induced-tetraploid derived from diploid A. cristatum. The A. desertorum used had 2n = 4x = 31 (the three extra A. desertorum chromosomes were believed to be B chromosomes), and the induced-tetraploid A. cristatum had 2n = 4x = 28. From a cytological aspect, the parents behaved as autoploids. Chromosome pairing at diakinesis of the 28-chromosome F1 hybrids included an average of 0.02 univalents, 8.54 bivalents, 0.02 trivalents, 2.25 quadrivalents, 0.22 hexavalents, and 0.06 octavalents. Hexavalent and octavalent associations at diakinesis and bridge-fragment formations at first and second anaphase signified structural heterozygosity between the A. cristatum and A. desertorum genomes. However, the F1 hybrids were fertile.

Dewey (1969) crossed doubled-diploid Agropyron cristatum and tetraploid A. desertorum with hexaploid A. cristatum. Meiosis in the parent plants was typical of that in autoploids, and the 35-chromosome F1 hybrids exhibited pentavalent associations, with up to five per cell. Occasional higher multivalent associations and bridge-fragment formations at first anaphase indicated the existence of some structural heterozygosity. The pentaploid hybrids were surprisingly fertile. The high pentavalent pairing indicated close homology between the parental genomes. These results and others led Dewey (1969) to assume that all crested wheatgrasses, whether diploid, tetraploid, or hexaploid, contain one basic genome that has undergone some structural rearrangements including both translocations and inversions. Hence, autopolyploidy has played an important evolutionary role in the evolution of this genus.

Likewise, Knowles (1955) suggested that Agropyron desertorum is an autopolyploid of diploid A. cristatum. Sarkar (1956) concluded that evolution occurred primarily through autopolyploidy, followed by structural and genic changes in the chromosomes. However, since the degree of morphological variation among the tetraploid species implied the contribution of more than one genome, Sarkar (1956) suggested that segmental allopolyploidy must have been involved in the evolution of these species. Likewise, Schulz-Schaeffer et al. (1963) suggested segmental allopolyploidy in evolution of the tetraploid species, while the hexaploid cytotype of A. cristatum originated from autopolyploidy only. They suggested the possibility that A. desertorum is an allopolyploid involving diploid A. cristatum and an unknown diploid. The karyotype analysis performed by McCoy and Law (1965) supported the assumption of a segmental alloploidic nature of the tetraploids.

In line with these works, the chromatographic study of Lorenz and Schulz-Schaeffer (1964) showed that the tetraploid species A. desertorum, A. pectinatum (currently A. cristatum) and A. sibiricum contained more phenolic compounds than the diploid Agropyron species. They concluded that doubling of a single genome would not lead to proliferation of phenolic compounds, and, consequently, the tetraploids may have been derived through allopolyploidy, with the second diploid parent still unidentified.

The phenolic profile of the tetraploid species also brought Taylor and McCoy (1973) to conclude that while tetraploid Agropyron cristatum ssp. pectinatum (formerly A. pectiniforme) is a natural autopolyploid, another tetraploid subspecies of A. cristatum, namely, ssp. imbricatum, as well as A. desertorum, A. fragile, and A. sibiricum are segmental allopolyploids that derived from hybridization of different diploid subspecies of A. cristatum. To further support their conclusion, Taylor and McCoy (1973), produced a colchicine-induced autopolyploid from two different clones of A. cristatum, and confirmed that autopolyploidy does not, in itself, result in the production of phenolic compounds absent in the diploid progenitor.

While segmental allopolyploidy that resulted from intergeneric hybridizations and introgressions, e.g., from tetraploid Elymus species, can not be ruled out, the possibility exists, as suggested by Dewey (1969), that the tetraploids originated through autoploidy of different diploids of A. cristatum and A. mongolicum, that later hybridized with each other.

Yousofi and Aryavand (2004) found a genome size of 26.4 and 27.6 pg 2C DNA in different tetraploid lines of A. cristatum, respectively 7% and 3% less than the additive amount of the diploid (14.2 × 2 = 28.4) (Vogel et al. 1999). These reductions in DNA content in the tetraploids, as well as structural chromosomal rearrangements, may have led to reduced multivalent formation and increased bivalent pairing and consequently, to a disomic mode of inheritance (Eilam et al. 2009). Autotetrapoids contain duplications of most of their gene loci. While the activity of most duplicated genes might be of adaptive value, the activity of some of the duplicated genes may lead to overproduction of proteins and other chemical compounds and consequently, to a disadvantageous or even deleterious effect (Birchler and Veitia 2007). Natural selection will favor changes leading to sub-functionalization or neo-functionalization in these loci. Sub-functionalization may occur when an ancestral gene with two functions becomes duplicated and each of the duplicated genes specializes in one of the ancestral gene functions, while neo-functionalization describes gain of a new, nonancestral function in a duplicated locus. Neo-functionalization of duplicated genes in autotetraploids may generate the formation of new gene products, such as new phenolic compounds that are not present in the diploid progenitor. Actually, new phenotypes often arise with polyploid formation and can contribute to the success of polyploids (Osborn et al. 2003, and reference therein).

Chromosome numbers in the tetraploid ssp. pectinatum of A. cristatum vary from 28 to 33 (Assadi 1995). Different PMCs within the same anther can display chromosome counts ranging from 28 to 32 in one plant and from 32 to (rarely) 33 in another plant. Aneuploid chromosome numbers have been reported in various tetraploid collections of A. cristatum (Myers and Hill 1940; Dewey and Asay 1975), a phenomenon that may have derived from cytologically unstable pentaploid hybrids between tetraploid and hexaploid cytotypes of A. cristatum (Dewey 1974). Hence, the plants that had the somatic chromosome numbers 2n = 32 and 33, were probably a derivative of such unstable pentaploid hybrids (Assadi 1995). The variable chromosome number in different PMCs may be caused by elimination of chromosomes in archesporial division or at an early stage of the meiotic cycle. Alternatively, the extra chromosomes may be B chromosomes, which were frequently observed in PMCs of A. cristatum (Knowles 1955; McCoy and Law 1965; Assadi 1995; Asghari et al. 2007). Baenziger (1962) found no B chromosomes in adventitious root-tips of diploid A. cristatum but reported the presence of these chromosomes in stem meristems, in primary roots, and in PMCs but they were absent in adventitious roots of the tetraploid A. desertorum and diploid A. cristatum. The B chromosomes are usually smaller than the basic (A) chromosomes, not heterochromatic, and show sub-terminal centromeres. At meiosis, there is good pairing between the B chromosomes but not between B and A chromosomes (Baenziger 1962). A B chromosome with a sub-terminal constriction, which is either a centromere or a secondary constriction, was also observed in A. desertorum (McCoy and Law 1965). The group reported a mitotic chromosome count of 2n = 28, whereas the meiotic chromosome count was 2n = 32. Evidently, as noted by Knowles (1955), the mitotic chromosome number may not agree with the meiotic number.

5.3.4 Crosses with Other Triticineae Species

Studies of meiotic chromosomal pairing in F1 hybrids between diploid A. cristatum and several different diploid species of Elymus bearing genome StSt, (genome of all hybrids was PSt) showed that the P and St genomes are related (Wang 1985a, b, 1986, 1992). Size differences between Agropyron (large) and Elymus (small) chromosomes facilitated interpretation of chromosome pairing in these hybrids. Average chromosome pairing at first meiotic metaphase of the diploid hybrid between Agropyron cristatum and Elymus stipifollus included 7.65 univalents, 2.88 rod- and 0.21 ring-bivalents (total 3.09 bivalents), 0.04 trivalents and 0.01 quadrivalents (Wang 1985b). These pairing configurations indicate allosyndetic pairing between the homoeologous chromosomes of the two genomes, showing that the St and the P genomes are related. Later, Wang (1986) produced hybrids between diploid Agropyron cristatum and Elymus libanoticus. Chromosomal pairing at first meiotic metaphase of the F1 hybrid (genome PSt) averaged 7.71 univalents, 2.77 bivalents, 0.22 trivalents, 0.01 quadrivalents, and 0.01 pentavalents per cell. As with the hybrid of A. cristatum x E. stipifolius (Wang 1985b), chromosome pairing in the hybrid between diploid A. cristatum and E. libanoticus was mainly between the P genome (large) and the S-genome (small) chromosomes, i.e., allosyndetic. Meiotic pairing also in this hybrid suggests that the P and the St genomes are related.

Wang (1992) produced hybrids between diploid species of Agropyron (genome P), Elymus (genomes St, Ee and Eb), Psathyrostachys (genome Ns), Hordeum (genome H), and Secale (genome R). Chromosome pairing patterns in these diploid hybrids enabled the estimation of genomic similarity between the various genomes. The results showed that Ee of Elymus elongatus and Eb of E. farctus are the most closely related genomes, followed by the St and P genomes. The N, H, and R genomes are remotely related to the E-St-P cluster. These relationships are also reflected in hybrids of higher ploidy levels, when genes controlling chromosome pairing are kept in check (Wang 1992). Similarly, the average meiotic chromosome pairing in the intergeneric diploid hybrid E. spicatus (genome StSt) x Secale strictum (formerly montanum) (genome RR) included 12.97 univalents, 0.49 bivalents, and 0.01 trivalents (Wang 1987b). The hybrid A. mongolicum x S. strictum, which have the PR genomes, had an average of 12.86 univalents, 0.51 bivalents, 0.03 trivalents, and 0.004 quadrivalents. The hybrid between E. spicatus and A. mongolicum (genome StP) had a mean configuration of 8.05 univalents, 2.86 bivalents, 0.07 trivalents, and 0.01 quadrivalents. All hybrids were sterile. The meiotic pairings of these hybrids indicated that chromosome homology between the St and P genomes is higher than between both St and R or between P and R. The degree of meiotic pairing in the E. spicatus x A. mongoicum hybrid was similar to that observed in other diploid hybrids bearing the same genome constitution, i.e., A. cristatum x E. stipifolius and A. cristatum x E. libanoticus (Wang et al. 1985; Wang 1986). Interestingly, mitotic preparations of root-tip cells of these hybrids suggested that the chromosomes of different genomes were spatially separated (Wang 1987b). As was found in other plant species (Finch et al. 1981; Avivi et al. 1982), the separated genome distribution in the A. cristatum x Elymus species in the majority (50–67%) of root-tip cells, suggested that each genome in these hybrids occupy a different part of the nucleus.

Monoploids and hybrids were obtained from the cross of diploid Elymus elongatus (genome EeEe) and A. mongolicum. The monoploid was a result of gradual and eventually complete elimination of A. mongolicum chromosomes in the hybrid. About 95% of the root-tip cells, and nearly all of the pollen mother cells, had only seven chromosomes (Wang 1987a). The genome in the monoploid cells was identified as Ee, by its characteristic satellited chromosomes (Wang 1987a). This was the first report of chromosome elimination following intergeneric hybridization in the Triticeae that did not involve species of Hordeum or Critesion. The monoploid plant had only a few root-tip cells that contained as many as seven additional chromosomes, whereas none of the PMCs had more than eight chromosomes. These observations indicated that chromosome elimination commenced some time after zygote formation and was nearly complete in the PMCs. Chromosome elimination in the hybrids between tetraploid H. vulgare and H. bulbosum begins at maximum, 3–5 days after pollination and was frequently complete 9 days after pollination (Fukuyama and Hosoya 1983).

Very little autosyndesis between chromosomes within the Ee genome occurred in the monoploid. On the other hand, extensive chromosome pairing was observed at first meiotic metaphase of the F1 hybrid E. elongatus x A. mongolicum (genome EeP), averaging 6.42 univalents, 2.53 rod- and 0.85 ring-bivalents, 0.25 trivalents, and 0.02 quadrivalents. Bridges and fragments were present in many first anaphase cells. The hybrid was sterile and had non-dehiscent anthers. This pairing profile revealed a degree of chromosome homology between Ee and P, indicating a close phylogenetic relationship between these two species. The amount of pairing between the Ee and the P genomes, especially the occurrence of three ring bivalents in some cells, suggests a close relation between the two. Clark et al. (1986), who studied the spacer region of rDNA units, also found a close relationship between Ee and P genomes. However, the number of univalents in this intergeneric hybrid (6.42) exceeds that (2.89) observed in the interspecific hybrids of diploid E. farctus (genome Eb) and diploid E. elongatus (genome Ee) (Wang 1985b). Despite the capacity of the Ee and P genome chromosomes to pair, they have differentiated to a degree that the two genomes have different chromosome lengths and karyotypic patterns (Hsiao et al. 1986). These karyotypic differences, as well as morphological differences between E. elongatus and A. mongolicum, justify their classification into separate genera (Wang 1987a).

Wang (1985b) crossed tetraploid A. desertorum (genome PPPP) x tetraploid E. stipifolius (genome StStStSt) and found that most F1 hybrid (genome PPStSt) pairing was autosyndetic, namely, pairing between chromosomes of the St genome and pairing between those of the P genome. Average pairing configurations included 4.48 univalents, 5.79 rod- and 5.07 ring-bivalents (total 10.86 bivalents), 0.53 trivalents and 0.05 quadrivalents (Wang 1985b). These data imply that the two parental tetraploid species are autoploids, that several chromosomes in each genome underwent some structural rearrangements, and that the P and St chromosomes tended to pair with their homologues (autosyndesis) rather than with their homoeologues (allosyndesis). Hence, chromosomal pairing in such hybrids cannot disclose the degree of relatedness between the P and St genomes.

Assadi and Runemark (1995) crossed E. libanoticus) (genome StSt) with a tetraploid cytotype of A. cristatum, namely, ssp. pectinatum (genome PPPP). Average meiotic configurations in the F1 hybrid (genome StPP) displayed: 11.30 univalents, 3.40 rod- and 1.50 ring-bivalents (4.90 total bivalents). The hybrid was sterile. Most chromosomal pairing, if not all, was autosyndetic between P chromosomes. The preferential pairing between the homologous P chromosomes precluded the assessment of the relationships between genomes P and St.

Dewey (1963a) reported that meiosis in plants of Elymus hispidus (formerly Agropyron trichophorum) (2n = 6x = 42; EeEeEeEeStSt] was basically regular; average chromosomal associations at the first meiotic metaphase showed 0.09 univalents, 20.56 bivalents, 0.05 trivalents, and 0.16 quadrivalents, and was therefore described as an auto-allohexaploid. Dewey (1963a) analyzed chromosomal pairing at first meiotic metaphase of F1 hybrids between Elymus hispidus and A. desertorum (genome PPPP) and between E. hispidus and hexaploid cytotype of A. cristatum (genome PPPPPP). The hexaploid A. cristatum parent averaged 0.18 univalents, 7.44 bivalents, 0.81 trivalents 2.86 quadrivalents, 0.08 pentavalents, and 2.11 hexavalents at diakinesis and consequently, was described as an autohexaploid (Dewey 1963b). Chromosome pairing at the fist meiotic metaphase of the F1 hexaploid hybrid (genome EeEeStPPP) presented 5.08 univalents, 8.94 bivalents, 4.33 trivalents, 1.11 quadrivalents, 0.27 pentavalents, and 0.05 hexavalents per cell. On the basis of chromosome pairing in the parent species and their hybrids, it was concluded that one of the E. hispidu genomes was partially homologous with the P genomes of hexaploid A. cristatum and tetraploid A. desertorum.

Martín et al. (1999) reported the production of the amphiploid Aegilops tauschiiAgropyron cristatum, obtained by crossing an induced autotetraploid of Ae. tauschii (genome DDDD) with tetraploid A. cristatum (genome PPPP). They used multicolor fluorescence in situ hybridization (FISH), using total genomic DNA probes, to distinguish between the chromosomes of Ae. tauschii and those of A. cristatum at meiosis of the amphiploid. Analysis of chromosomal pairing at first meiotic metaphase of the amphiploid (genome DDPP) showed the presence of multivalents (trivalents, quadrivalents and pentavalents), which were of A. cristatum origin. Moreover, pairing between Ae. tauschii chromosomes was higher than between A. cristatum chromosomes. The high frequency of multivalents, presumably due to translocations, plus the reduced pairing of the A. cristatum chromosomes, indicated that the two P genomes of the latter underwent some structural changes. FISH analysis also showed a rare event of pairing between Ae. tauschii and A. cristatum chromosomes. The presence of homologous chromosomes (DD and PP) competed with the homoeologous pairing between D and P and thus, the low level of homoeologous pairing between these two genomes is not indicative of the relationships between these two genomes. In the absence of homologous chromosomes, this pairing could be presumably higher.

Martin et al. (1998) studied chromosome pairing at meiosis between the diploid cytotype of A. cristatum (genome PP), Aegilops tauschii (genome DD) and Hordeum chilense (genome HchHch) in the trigeneric hybrid Ae. tauschii-A. cristatum x H. chilense. Since this trigeneric hybrid (genome DPHch) had a single dose of each genome and thus, lacks homologous chromosomes, analysis of its pairing pattern can reveal the level of affinity between the genomes of these species. Using FISH, the pairing of these hybrids at first meiotic metaphase showed higher pairing between the D and Hch genomes than between each of them and the P genome.

Hybridization of Agropyron species with Triticum species has been difficult. White (1940) was unsuccessful in producing hybrids between Triticum and diploid A. cristatum. Smith (1942a, b) reported one hybrid plant produced from 882 Triticum aestivum florets pollinated with A. cristatum, and Mujeeb-Kazi et al. (1987) reported that no viable embryos were obtained from a cross of T. aestivum with tetraploid Agropyron cristatum. Chen et al. (1989) succeeded to produce two hybrid plants from the pollination of 952 T. aestivum florets with diploid A. cristatum pollen, but in both instances, the hybrid plants died before reaching maturity. Nevertheless, Chen et al. (1989, 1990), Li and Dong 1990, 1991), and Ahmad and Comeau (1991) recently succeeded to obtain hybrid plants between T. aestivum and various tetraploid Agropyron species, namely, A. desertorum, A. michnoi, and A. fragile. Studies of chromosomal pairing at first meiotic metaphase of these F1 pentaploid hybrids showed higher levels of meiotic pairing than expected. But, the genomic relationships of the P genome of Agropyron with the A, B, and D subgenomes of T. aestivum could not be clearly assessed due to the presence of the wheat Ph1 homoeologous-pairing suppressor and two homologous genomes in the tetraploid Agropyron species.

Ahmad and Comeau (1991) analyzed chromosome pairing in ten F1 hybrids between T. aestivum cv. Fukuho and A. fragile. Mean chromosome configurations at first meiotic metaphase of the 10 pentaploid hybrids (genome BADPP) included 17.29 univalents, 6.57 rod- and 1.97 ring-bivalents, 0.18 trivalents, 0.03 quadrivalents, and 0.002 hexavalents per PMC. However, there was a considerable intra-hybrid variation in the mean number of bivalents per PMC, ranging from 5.88 to 11.03. Since one would expect to find up to seven bivalents in the polyhaploid of A. fragile and up to three bivalents (Ahmad and Comeau 1991) in the polyhaploid of T. aestivum cv. ‘Fukuho’, the expected maximum bivalents in these F1 hybrids should be 10. The higher number of bivalents in some of the hybrids, which presumably occurred between wheat chromosomes, was attributed to the a pairing-promoter gene(s) present in A. fragile (Ahmad and Comeau 1991). Allosyndetic pairing between wheat and A. fragile chromosomes are not expected in these hybrids, since the A. fragile chromosomes tend to pair preferentially with their own homologues, rather than with wheat homoeologues. Such a pairing-promoter gene system has been previously reported in tetraploid A. cristatum (Chen et al. 1989), as well as in other Triticeae species, e.g., Amblyopyrum muticum, Aegilops speltoides, Secale cereale, and Dasypyrum villosum (Dover and Riley 1972; Chen and Dvorak 1984; Lelley 1976; Blanco et al. 1988b; Table 5.2). Such a gene system was not found in diploid A. cristatum (Limin and Fowler 1990). Since the pairing promotion system in tetraploid Agropyron species is determined by a polygenic system (Jubault et al. 2006), and different accessions may have different alleles of the system, the use of bulk pollen from many different plants to pollinate wheat (Ahmad and Comeau 1991), might produce hybrids that differ in the allelic composition of the Ph suppressors.

Jauhar (1992) produced and analyzed intergeneric hybrids between T. aestivum and tetraploid Agropyron cristatum. The F1 pentaploid hybrids (genome BADPP) were perennial like the male wheatgrass parent and morphologically intermediate between the two parents. Two types of hybrids were obtained: a low-pairing (LP) hybrid and a high-pairing (HP) hybrid. The LP hybrid, with an apparently functional Ph1 (the suppressor of homoeologous pairing), had a mean display of 25.91 univalents, 3.17 rod- and 1.16 ring-bivalents, and 0.14 trivalents. If A. cristatum were a true autotetraploid, its haploid complement (PP) in the hybrid should form approximately 7 bivalents. The mean 4.33 bivalents (of which about 1.0 bivalent probably involved the A, B and D subgenomes of wheat) suggests a certain degree of divergence between the two P genomes (Jauhar 1992). The degree of divergence between chromosomes that is required for the Ph1 suppressor to operate on is not known. It would appear, however, that the degree of similarity between the two P genomes is inadequate to pass the discrimination limits of Ph1 (Jauhar 1992).

The HP hybrid had 15.73 univalents, 5.89 rod- and 2.98 ring-bivalents, 0.47 trivalents, and 0.03 chain quadrivalents, a pairing profile that likely involved both autosyndesis (pairing within the BAD component and within the PP component of the BADPP hybrid) and allosyndesis (pairing between the parental complements), as indicated by the frequent formation of heteromorphic bivalents and asymmetrical trivalents (Jauhar 1992). This kind of pairing could have occurred only if Ph1 was partially suppressed by genes of tetraploid A. cristatum. The existence of LP and HP plants among the F1 pentaploid hybrids presumably results from segregation of the Ph1 suppressors of A. cristatum.

Limin and Fowler (1990) reported the first successful hybridization of T. aestivum cv. Chinese Spring with diploid A. cristatum. Average chromosomal pairing per PMC at the first meiotic metaphase of the F1 tetraploid hybrid (genome BADP) included 27.69 univalents, 0.15 rod- and 0.00 ring-bivalents, and 0.003 trivalents (0.16 chiasmata per cell) (Limin and Fowler 1990). Haploid Chinese Spring wheat had 0.24–0.27 chiasmata per cell (Miller and Chapman 1976; McGuire and Dvorak 1982). Therefore, the chiasma frequency observed in the study of Limin and Fowler (1990) could be explained on the basis of autosyndesis pairing between the wheat subgenomes. The pairing data of the tetraploid hybrid indicates that pairing-promoting gene(s) does not exist in the diploid accession of A. cristatum used as parent in the described cross.

Triticum aestivum cv. Chinese Spring was crossed with diploid Agropyron species from Inner Mongolia, A. cristatum and A. mongolicum, with or without B chromosomes, generating intergeneric F1 tetraploid hybrids with 2n = 27, 28, 32, and 33 chromosomes (Chen et al. 1992a, b). The extra chromosomes in the hybrids with 2n = 32 and 33 were assumed to be B chromosomes. Average meiotic pairing in the euploid hybrid (2n = 4x = 28; genome BADP), derived from the cross of Chinese Spring x (A. cristatum x A. mongolicum), included 14.38 univalents, 3.56 rod- and 1.36 ring-bivalents, and 1.26 trivalents, and the maximum number of bivalents was seven. This level of pairing is higher than expected and was likely due to homoeologous pairing between wheat chromosomes. Hence, the data indicated that the P genome of diploid Agropyron originated from Inner Mongolia, as those of the tetraploid Agropyron species, possess a genetic system that suppresses the Ph1 genes of wheat. The chromosome pairing observed in hybrids of CS x A. cristatum/A. mongolicum included 4.92 bivalents and 1.26 trivalents on average, and was much higher than that previously reported for the hybrid CS x diploid A. cristatum produced by Limin and Fowler (1990), which had an average of 27.69 univalents, 0.15 bivalents and 0.003 trivalents per cell. This difference between the hybrids of common wheat x diploid A. cristatum could be attributed to the use of different accessions of A. cristatum used in the study of Chen et al. (1992a, b) versus that of Limin and Fowler (1990).

The observed pattern of pairing in the tetraploid hybrid (genome BADP) (Chen et al. 1992a, b) was very similar to that of the pentaploid hybrid CS x tetraploid A. cristatum (genome BADPP), previously produced by Chen et al. (1989). Both diploid and tetraploid A. cristatum, used in these crosses, have genes that suppress the Ph-system of wheat. In the pentaploid hybrids, average chromosome pairing included 8.18 univalents, 11.88 bivalents, 0.97 trivalents, and 0.03 quadrivalents (Chen et al. 1989). However, as the tetraploid A. cristatum used was a true autotetraploid (Dewey 1984; Chen et al. 1992a, b), in most PMCs, seven bivalents representing autosyndesis pairing of the 14 Agropyron chromosomes, was observed. Thus, the meiotic behavior of wheat chromosomes shows 8.18 univalents, 4.88 bivalents, 0.97 trivalents, and 0.03 quadrivalents, corresponding to the level of pairing in a wheat haploid deficient for the Ph gene (7.93 univalents, 5.20 bivalents, 0.53 trivalents, and 0.02 qiadrivalents; Kimber and Riley 1963). Hence, the high level of pairing in the pentaploid may be ascribed to the suppression of the Ph gene of wheat (Chen et al. 1989).

In the study of Chen et al. (1992a, b), the mean chromosome pairing observed in the tetraploid hybrids dislayed 14.38 univalents, 4.92 bivalents, and 1.26 trivalents. If we assume that the seven chromosomes of the P genome of Agropyron did not pair with those of wheat in the hybrid, and consequently, were univalents, the meiotic behavior of wheat chromosomes in the hybrid would therefore be 7.38 univalents, 4.92 bivalents, and 1.26 trivalents. This level of pairing is very similar to that of the wheat chromosome in the pentaploid hybrid. Hence, the P genome of both diploid and tetraploid A. cristatumn has a gene that suppress the Ph gene system of wheat. The fact that the P genome of another accession of diploid A. cristatum did not suppress the Ph effect (Limin and Fowler 1990), indicates the existence of variability within diploid Agropyron concerning the Ph suppressors.

Li and Dong (1990) produced intergeneric hybrids between T. aestivum cv. Chinese Spring and tetraploid Agropyron desertorum. Average meiotic chromosome pairing at the first metaphase of the F1 hybrid (genome BADPP) showed 6.62 univalents, 4.16 rod- and 8.20 ring-bivalents, 0.57 trivalents, 0.35 quadrivalents, 0.06 pentavalents and 0.03 hexavalents. The number of bivalents and multivalents in the F1 hybrid was higher than the expected seven bivalents between the PP genomes of Agropyron and one bivalent between the wheat genomes.

Li and Dong (1991) also produced intergeneric hybrids between T. aestivum cv. Chinese Spring and Agropyron michnoi. The average meiotic chromosome pairing at the first meiotic metaphase of F1 pentaploid hybrid (genome BADPP) included 6.39 univalents, 3.75 rod- and 8.64 ring-bivalents, 0.81 trivalents, 0.30 quadrivalents and 0.04 pentavalents; the bivalent and multivalent formation was much higher than expected.

Chen et al. (1989) assumed that the higher pairing in hybrids between bread wheat and tetraploid Agropyron species resulted from Agropyron genes that suppress the wheat Ph effect and thus lead to wheat homoeologous pairing. Li and Dong (1990) suggested that the duplicated dosage of the P genome induced pairing between the homoeologues. Their conclusion was inspired by the report of Riley et al. (1973) who showed that an extra dose of the rye R genome increased the level of pairing in hybrids between common wheat and tetraploid rye, i.e., the mean number of wheat-chromosome bivalents increased from 0.24 to 2.25. Li and Dong suggested a similar effect of the P genome, when present in an extra dose. However, the lack of dosage effect of the P genome of Agropyron on suppression of the Ph gene of wheat is indicated from the similar level of pairing between the wheat chromosomes in the tetraploid and the pentaploid hybrids (Chen et al. 1992a, b). It appears that wheat-Agropyron allosyndetic associations, if any, are rare, even if the level of chromosome pairing observed in the hybrid is high.

Hybrids between disomic addition lines of A. cristatum chromosomes or chromosome arms to the complement of bread wheat x Aegilops peregrina (=Ae. variabilis) (2n = 4x = 28; genome SvSvUU), can be used for studying the effect of individual Agropyron chromosomes and chromosome arm on homoeologous pairing between wheat and Ae. peregrina chromosomes and to assess the Ph-suppressing effect of different P genome chromosomes. Jubault et al. (2006) used five disomic addition lines (1P, 3P, 4P, 5P and 6P) and five ditelosomic addition lines (2PS, 2PL, 4PS, 5PL and 6PS) of wheat—A. cristatum addition lines, produced by Chen et al. (1994, 1992a, b), in crosses with Ae. peregrina. Chromosome configurations in each hybrid, which had either 2n = 36 or 35 + t, were recorded, and the pairing level for each of them was compared with that of the control hybrid T. aestivum CS–Ae. peregrina. All the genotypes, except those with 2PS and 2PL chromosomes, displayed a significantly higher level of homoeologous pairing than the control. Consequently, all the P chromosomes tested, with the exception of chromosomes arms 2PS and 2PL, seemed to promote homoeologous pairing. The A. cristatum Ph-suppressing system appeared polygenic. However, the pairing-promoting effect of every Agropyron chromosome was weaker than the effect of the absence of Ph1.

In addition, Jubault et al. (2006) assessed the level of pairing between individual A. cristatum chromosomes and those of common wheat, in hybrids lacking Ph1. Allosyndetic pairing between P and BAD chromosomes were very rare even in the absence of Ph1. Only telosome 5PL paired, at a very low frequency, with wheat chromosomes (Jubault et al. (2006). Since the addition lines did not provide any evidence for structural rearrangements between the P and the A, B, D subgenomes, it is assumed that the lack of ability of P chromosomes to pair with wheat chromosomes stems from divergence of the DNA sequences that are involved in homology recognition and initiation of meiotic pairing.

5.3.5 Phylogenetic Relationships of Agropyron with Other Triticineae

The study of Escobar et al. (2011) showed that the clade of Agropyron, Astralopyrum, Eremopyrum and Henrardia was not affected by introgression and/or incomplete lineage sorting. The analyses of 5S DNA sequences using Wagner parsimony and NJ distance methods (Baum and Appels 1992), placed consistently Agropyron (genome P), Pseudoroegneria (currently Elymus; genome St) and Australopyrum (genome W) in one clade. Agropyron evolved, most probably, from Elymus species having the St, Ee or Eb genomes that are moderately related to genome P of Agropyron (Wang 1989). The close phylogenetic relationship between Agropyron and Ermopyrum is supported by the data of Escobar et al. (2011); the latter might have evolved from the former.

The internal transcribed spacer (ITS) region of nuclear ribosomal DNA sequence phylogeny indicated that the endemic Australian grasses Australopyrum pectinatum (genome W) are closely related to species of Agropyron (genome P) (Hsiao et al 1995). Species of the W and P genomes share certain gross morphological similarities and Australopyrum was once treated as a member of Agropyron (Löve 1984). The karyotypes of P and W genome species are also similar, but the chromosomes of the W genome are smaller (Hsiao et al. 1986). The differences in chromosome size could simply be due to a low copy number of the repetitive DNA, because the chromosomes of Australopyrum pectinatum ssp. velutinum contain much less C-banded heterochromatin than do those of Agropyron cristatum (Endo and Gill 1984).

The P genome of the allohexaploid Elymus species most likely derived from a diploid species of Agropyron (Petersen et al. 2011; Fan et al 2013; Dong et al. 2015). Refoufi et al. (2001) found that Elymus pycnanthus (Godr.) Melderis (=Elytrigia pycnantha or Thinopyrum pycnanthum) is a hexaploid containing genomes St, Ee, and P. Using genomic in situ hybridization (GISH) techniques, they also proposed that the P genome of E. pycnanthus is closely related to that of A. cristatum. In accord with the above, Dizkirici et al. (2010), constructing a phylogenetic tree by the maximum parsimony method, based on sequence diversity in the internal transcribed spacer (ITS) region of nuclear ribosomal DNA, revealed that Elymus pycnanthus clustered with species of Agropyron. Molecular diversity statistics also indicated that E. pycnanthus is close to Agropyron species (Dizkirici et al. 2010).

The studies of Fan et al. (2009) and Sha et al. (2010) indicated that the Xm genome of the allopolyploid Leymus species might have originated from an ancestral lineage of Agropyron (genome P) and Eremopyrum triticeum (genome F).

5.4 Eremopyrum (Ledeb.) Jaub. & Spach.

5.4.1 Morphological and Geographical Notes

Eremopyrum (eremia ‘desert’, and pyros ‘wheat’ in Greek) was described by Ledebour (1853) and included in Triticum sect. Eremopyrum. However, since Eremopyrum was morphologically similar to the P genome-bearing Agropyron crested wheatgrasses, Bentham and Hooker (1883) included them in Agropyron Gaertner. But, due to its annual habit, Jaubert and Spach (1851) distinguished this taxon from Agropyron as a separate genus. Since then, a number of different Eremopyrum species have been described. Currently, there are five universally accepted species in the genus (Gabi and Dogan 2010), namely, E. bonaepartis (Spreng.) Nevski, E. confusum Melderis, E. distans (C. Koch) Nevski, E. orientale (L.) Jaub. et Spach, and E. triticeum (Gaertn.) Nevski (Figs. 5.3b and c).

In her review of the taxonomy of the genus Eremopyrum, Frederiksen (1991b) recognized only the following four species: E. triticeum, E. orientale, E. distans, and E. bonaepartis. Due to the absence of clear-cut delimitation between three previously considered species, she included them in E. bonaepartis: E. confusum, characterized by awned glumes and lemmas, E. bonaepartis s. str., with sharp pointed glumes and lemmas, and Triticum sinaicum Steud., characterized by gradually tapering glumes and lemmas on 1–3 lower spikelets, but distinctly awned lemmas on upper spikelets. Moreover, Frederiksen (1991b) determined the chromosome number of these taxa and found that E. confusum and E. bonaepartis s. str. are tetraploids, while Triticum sinaicum is a diploid. In contrast, Gabi and Dogan (2010) indicated clear differences between the three taxa. More specifically, they confirmed that E. confusum Melderis is a valid species and consequently, like Melderis (1985b), recognized five species in Eremopyrum. Since tetraplois E. bonaepartis is an allotetraploid (Sakamoto 1979), and differs morphologically from diploid E. bonaepartis (Gabi and Dogan 2010), they concluded that the two taxa should be treated as separate species.

All the five species are annual, short plants (30–40 cm high), with a short, compact, laterally compressed spike, rachis with very short internodes, solitary spikelets, seated distichously at a wide angle to the rachis, and with spikes that disarticulate at maturity at each rachis node beneath each spikelet (wedge-type disarticulation), but only in E. triticeum is the disarticulation at the base of each floret (floret-type disarticulation). The Eremopyrum species feature spikelets with 2–5 bisexual florets, distal or no sterile florets, and very short anthers (0.4–1.3 mm) indicating facultative self-pollination. Based on differences in disarticulation of spike and spikeletes, Nevski (1936) divided Eremopyrum into two sections, Micropuryum Nevski (includes E. triticeum) and Eremopyrum Nevski (includes the other four species) (Table 5.4).

Table 5.4 Sections and species of Eremopyrum, their disarticulation type, ploidy level, genome formula and geographic distribution

Eremopyrum species grow in steppes and semi-desert regions, from the Balkan, through the East Mediterranean to Asia (Balkan, Turkey, Syria, Jordan, Israel, Sinai Peninsula, Caucasia, Turkmenistan, Iraq, Saudi Arabia, Iran, Afghanistan, Pakistan, and China) (Bor 1968, 1970; Davis et al. 1988). In their native ranges, they serve as valuable fodder on ephemeral spring pastures. E. triticeum, E. bonaepartis, and E. orientale have been found in North America.

The distribution of E. triticeum extends from southeastern Europe and Turkey in the west, to China in the east. The species appears to be widespread in the northern region of the genus distribution area. E. distans is an Asiatic species, widely distributed from the East Mediterranean and Eastern Turkey to Afghanistan in the east. E. orientale is a widespread species that distributes from Morocco and Algeria in the west, to China in the east. E. bonaepartis is the most common of the species, and widely distributed, growing from Morocco in the west through the Middle East, the Sinai Peninsula, the Arabian Peninsula, Iran and Central Asia, to Afghanistan, Pakistan, and the Chinese province Xinjiang in the east. It is a variable species that has been divided into several subspecies (Table 5.4).

5.4.2 Cytology, Cytogenetics and Evolution

Studies of somatic chromosome number showed that Eremopyrum includes both diploid and tetraploid taxa; E. triticeum and E. distans, are diploids (2n = 2x = 14), E. ponaepartis contains diploid and tetraploid cytotypes, and E. confusum and E. orientale are tetraploids (2n = 4x = 28) (Avdulov 1931; Sakamoto and Muramatsu 1965; Frederiksen 1991b). Interspecific Eremopyrum hybrids were produced and studies of their chromosome pairing at meiosis showed: (1) very little pairing between the diploid species, indicating that their genomes are remarkably diverged from one another; (2) that the tetraploid species are allotetraploids, each containing two different genomes; tetraploid E. bonaepartis contains one genome of diploid E. bonaepartis and a second genome of E. distans, while tetraploid E. orientale contains one genome of E. triticeum and a second genome of E. distans (Sakamoto 1972). Since tetraploid E. bonaepartis is allotetraploid, the two ploidy types of E. bonaepartis should be separated into two different species. Based on these findings, Sakamoto (1979) classified the genomes of the Eremopyrum species as follows: diploid E. bonaepartis (genome AA), E. distans (genome BB), E. triticeum (genome CC), tetraploid E. bonaepartis (genome AABB) and tetraploid E. orientale (genome BBCC). Since genomic symbols A, B, and C, were previously given to Triticum and Aegilops species, Dewey (1984) changed the genome symbol of Eremopyrum species given by Sakamoto (1979). Later, Wang et al. (1995) suggested the presence of two different genomes in Eremopyrum, F and X. Following the genome analysis performed by Sakamoto (1979), the genome symbol of the Eremopyrum species are as follows: E. triticeum FF, diploid E. bonaepartis XbXb, tetraploid E. bonaepartis XbXbXdXd, E. distans XdXd, E. orientale XdXdFF (Table 5.4). The genome of E. confusum was not determined, but it is probably similar to that of tetraploid E. bonaepartis, i.e., XbXbXdXd.

The diploid species E. triticeum and E. distans, are more easily distinguishable than the other species. The allotetraploid species often grow in mixed populations with their diploid progenitors and exhibit wider morphological variation (Sakamoto 1979). The three allotetraploid species are of recent origin, and most probably arose in the dry steppe zones of northwestern Iran, the assumed distribution center of this genus (Sakamoto 1979).

Intergeneric hybridizations showed that there are strong sterility barriers between Eremopyrum species and those of other Triticeae genera (Sakamoto 1967, 1968, 1972, 1974; Frederiksen 1991b, 1993, 1994; Frederiksen and von Bothmer 1995). Sakamoto (1974) succeeded in producing Heteranthelium piliferum x diploid Eremopyrum bonaepartis and E. bonaepartis x tetraploid Hordeum depressum hybrids. The diploid H. piliferum x E. bonaepartis hybrid exhibited abnormal growth and very little chromosomal pairing at first meiotic metaphase (average of 13.93 univalents, and 0.04 bivalents per cell). Growth of the triploid hybrid E. bonaepartis x H. depressum was highly vigorous and chromosomal pairing (averaged 9.97 univalents, 5.50 bivalents, 0.01 trivalents, and 0.00 quadrivalents) resulted mainly from autosyndesis of chromosomes derived from the autotetraploid Hordeum parent. From the very little pairing between the chromosomes of E. bonaepartis and the other two species, Sakamoto (1974) concluded that there is no homology among the genomes of the three species.

5.4.3 Phylogenetic Relationships with Other Triticineae Species

Phylogenetic studies, based on morphology (Seberg and Frederiksen 2001), chloroplast DNA (Mason-Gamer et al. 2002; Hodge et al. 2010), chloroplast, mitochondrial, and nuclear DNA sequences (Seberg and Petersen 2007; Escobar et al. 2011) and nuclear genes (Hsiao et al. 1995; Mason-Gamer et al. 2010), included Eremopyrum species in the same clade with species of Agropyron s. str.

5.5 Henrardia C. E. Hubbard

5.5.1 Morphological and Geographical Notes

Henrardia is a small genus containing two species, H. persica (Boiss.) C. E. Hubbard and H. pubescens (Bertol.) C. E. Hubbard (Hubbard 1946). The two species, having a characteristic morphology, differ from other genera of the tribe. Both species are annuals short plants with relatively long, cylindrical spike (5–15 cm long), anthers are small (1.5–2.2 mm in H. persica and 0.7 mm in H. pubescens) indicating a mating system of facultative self-pollination. Rachis harbors a solitary spikelet at each node, fragile, disarticulating at maturity just below the nodes (barrel type), so that each spikelet falls with the rachis segment beside it. Caryopsis is free but tightly enclosed by the glumes (Fig. 5.3c).

Both species are distributed in Turkey (Anatolia) and from there have dispersed eastwards through Armenia and Transcaucasia to Central Asia and southwards to Iraq, Iran, Afghanistan and Baluchistan (Hubbard 1946). H. persica is not common than H. pubescens (Bor 1968). The latter species may be found also in Syria (Bowden 1966).

5.5.2 Cytology, Cytogenetics and Evolution

The two species are diploids (2n = 2x = 14); Sakamoto and Muramatsu (1965) and Sakamoto (1972) reported 2n = 14 in H. persica and Bowden (1966) observed 2n = 14 in H. pubescens. The karyotype of both species is extremely asymmetric, consisting of large chromosomes, of which four pairs have sub-telocentric and three pairs have telocentric chromosomes (Asghari-Zakaria et al. 2002). One of the chromosomes has a small satellite located at the end of its long arm (Asghari-Zakaria et al. 2002). Henrardia species have a most asymmetric karyotype; all other species have metacentric or sub-metacentric chromosomes, except Aegilops caudata, Ae. umbellulata, and Ae. uniaristata that have several sub-telocentric chromosomes (Chennaveeraiah 1960) and Eremopyrum triticeum (all seven pairs are sub-telocentric), E. bonaepartis (five pairs are sub-telocentric) and E. distans (two pairs are sub-telocentric) (Frederiksen 1991b). In all other Triticeae species the NOR region and the satellite are located on the short arm whereas in the Henradia species (and in Eremopyrum distans) they are located on the long arm (Asghari-Zakaria et al. 2002). Study of the C-banded karyotype of H. persica showed that each chromosome has a unique, easily recognizable C-banding pattern (Asghari-Zakaria et al. 2002). The karyotype of the Henrardia species is unique, differing from those of all other Triticeae.

In an attempt to study the genetic relationships between Henrardia and other genera of the tribe, Sakamoto (1972) crossed H. persica, as either female or male parent, with a number of species from different Triticeae genera. Hybrids were obtained only in the cross of tetraploid Eremopyrum orientale x H. persica. These hybrids were intermediate in spikes morphology but their spikelets were of Eremopyrum type. Disarticulation of ripe spikelets of E. orientale is of the wedge-type and that of H. persica is of the barrel-type. The F1 showed the wedge-type disarticulation of the Eremopyrum parent. Chromosomal pairing at first meiotic metaphase was very low (13–21 univalents and 0–4 bivalents per cell), indicating reduced homology between the genomes of these two species. Another hybrid of H. persica x diploid Eremopyrum distans was obtained by Frederiksen (1993) but the plant was very weak, did not develop normal roots and died within a short time.

5.5.3 Phylogenetic Relationships with Other Triticineae Species

Because of its very peculiar morphology, Henrardia was earlier included in genera outside the Triticeae. Yet, in his taxonomical revision of these taxa, Hubbard (1946) noticed that Henrardia shares several diagnostic traits with the Triticeae, e.g., ovary and caryopsis hairy at the apex, lodicule hairy, lemma three or more nerved and seed longitudinally grooved with simple starch grains of the Triticeae-type in the endosperm (Tateoka 1962; Seberg et al. 1991). Consequently, since these characters have been regarded to be of diagnostic value in distinguishing the tribe Triticeae from other Poaceae tribes, Hubbard (1946) transferred this taxon to the Triticeae as a new genus, Henrardia C. E. Hubbard.

The deviating morphology of Henrardia led first to its classification in a separate sub-tribe, Henrardiinae Pilger, within Triticeae (Tzvelev 1976; Löve 1984). Later on, Clayton and Renvoize (1986) considered Henrardia as an offshoot of Aegilops and included it in the sub-tribe Triticineae. This is in accordance with Kellogg (1989) and Frederiksen and Seberg (1992) who, based on a cladistics analysis, concluded that Henrardia and the diploid species of Aegilops form a clade. As a consequence Kellogg (1989) suggested inclusion of Henrardia into Aegilops s. lat. As Henrardia possess a unique morphology and exceptional karyotype as well as low crossability with Aegilops species (Sakamoto 1972), it seems at present most convenient to consider it a part of the Triticineae, but to maintain it as a separate genus.

Seberg and Frederiksen (2001) performed a cladistic analysis, primarily based on morphology, of the monogenomic diploid genera of the Triticeae, and found that the large Aegilops clade consists of taxa traditionally included in Aegilops (van Slageren 1994), the closely related Amblyopyrum muticum, and Henrardia persica. But, using β-amylase gene sequences, Mason-Gamer (2005) carried out a phylogenetic study of the monogenomic Triticeae and found that Henrardia persica is close to Eremopyrum bonaepartis and both are in the same clade with Psathyrostachys. Eeremopyrum species grouped with Elymus (=Agropyron) and Henrardia on the chloroplast DNA tree (Mason-Gamer et al. 2002). A dual placement of Eremopyrum and Henrardia with Elymus (=Agropyron) has also been supported by other data sets (reviewed in Mason-Gamer 2005). Also Hodge et al. (2010), using chloroplast gene encoding ribosomal protein S16, found that Eremopyrum bonaepartis and Henrardia persica formed a well supported clade. Phylogenies based on chloroplast DNA contradict the phylogeny based on morphological data of Petersen and Seberg (1997) and Seberg and Frederiksen (2001). The agreement of the rps16 data (Hodge et al. 2010) and the results of Mason-Gamer et al. (2002) support the placement of Henrardia in the HenrardiaEremopyrum clade. A similar conclusion was reached by Hsiao et al. (1995) who analyzed nuclear DNA sequences. Escobar et al. (2010) determined the mating system of Triticeae species and combined the data with those obtained from molecular analysis of 27 protein-coding loci. They found that Henrardia persica is very close to Eremopyrum bonaepartis and form a clade (clade III) with Eremopyrum triticeum and Agropyron mongolicum. Using most comprehensive molecular data set of one chloroplastic and 26 nuclear genes, Escobar et al. (2011) found two well-supported clades, the first is formed by Australopyrum (clade IIIA), Henrardia and Eremopyrum bonaepartis (clade IIIB), and Agropyron mongolicum and E. triticeum (clade IIIC).

Seberg and Frederiksen (2001), based on morphology, placed Elymus farctus subsp. bessaribicum (=Thinopyrum bessarabicum) and Henrardia persica at the bases of their respective tree. However, the accumulated trees indicate that either Psathyrostachys or Hordeum is basal to the rest of the tribe and Henrardia is a more advanced type (Mason-Gamer 2005; Escobar et al. 2011).

5.6 Dasypyrum (Coss. & Durieu) T. Durand

5.6.1 Morphological and Geographical Notes

The taxonomy of and relationships of the Dasypyrum species have been the subject of controversy. Originally, they were placed in Secale L. and later in various other Triticeae genera until Schur (1866) recognized that this taxon was morphologically distinct from Secale, Triticum and other genera in the tribe. Schur (1866) placed it in a new genus, Haynaldia Schur, named in acknowledgement of Cardinal Haynald (1816–1891) and his interest in science and botany (Bor 1970). To avoid confusion with other Haynaldia genera, the genus name was later changed to Dasypyrum Cosson et Durieu and its generic rank was validated by Durand in 1888 (de Pace et al. 2011). According to Löve (1984), the genus name derives from the Greek words dasy (bushy, hairy) and pyros (wheat), and was selected to reflect the distinctive hairy keels of the glumes (de Pace et al. 2011). At first, three species were included in Dasypyrum (Candargy 1901): D. villosum, D. hordeaceum, and D. sinaicum, the latter being an annual species (Humphries 1978) occurring in eastern Mediterranean environments (Durand 1888) and was recognized as a species by Candargy as well (1901). However, in her taxonomical revision of the genus Dasypyrum, Frederiksen (1991a) noted that Dasypyrum sinaicum (Steudel) Candargy is based on Triticum sinaicum Steudel, whose lectotype belonged to Eremopyrum bonaepartis (Sprengel) Nevski, thereby rendering it inappropriately assigned to Dasypyrum. Thus, she recognized only two species in the genus: the annual diploid D. villosum (L.) Candargy [=Haynaldia villosa (L.) Schur] and the perennial tetraploid D. breviaristatum (=D. hordeaceum), and demonstrated that the inclusion of D. hordeaceum was based on a later homonym, and for that reason, changed the name to D. breviaristatum. D. breviaristatum is commonly known as a perennial tetraploid, 2n = 4x = 28. However, Sarkar (1957) isolated both a tetraploid and a diploid cytotype from a 1954 collection of D. breviaristatm assembled by G. L. Stebbins in Morocco. Later on, Ohta et al. (2002) also reported on the existence of a diploid cytotype of D. breviaristatum among populations of the tetraploid cytotype in the Atlas Mountains. Morphologically, the diploids were similar to but smaller than the tetraploids in plant height, spike length, and spikelet number (Ohta et al. 2002). A distinct difference between the two cytotypes was only found in the number of trichomes on the leaf surfaces (Ohta et al. 2002). The two cytotypes are perennial, the diploid being slower in growth than the tetraploid and also with smoother leaves (Sarkar 1957).

The genus Dasypyrum bears distinctive two-keeled glumes with tufts of bristles along the keels, rendering them easily distinguishable from other genera in the Triticeae. Its plants are annuals or perennials, with 20–100 cm high culms, and terminal spikes that are 4–12 cm long, including the awns, with 1 spikelet per node. Their rachis disarticulates above each spikelet (wedge-type disarticulation). Spikelets are more than three times the length of the rachis internodes, and laterally compressed, with 2–4 florets. The lower two florets are usually fertile, the terminal florets are sterile, glumes are awned with two hairy keels, lemmas are awned and anthers are 4–7 mm long (Fig. 5.4a).

Fig. 5.4
figure 4

Species of the Dasypyrum-Secale-Heteranthelium clade; a A spike of D. villosum (L.) Candargy; b A plant and spike of S. cereale L. ssp. cereale; c Giemsa C-banding of S. Cereale L. ssp. cereale chromosomes. The seven chromosomes are from left to right, 1R to 7R. (From Gill and Kimber 1974); d A plant and a spike of Heteranthelium piliferum Hochst

Dasypyrum (Cosson and Durieu) T. Durand (=Haynaldia Schur) comprises two allogamous (predominantly out-crossing) species: the annual D. villosum (L.) Candargy and the perennial D. breviaristatum (Lindb. f.) Frederiksen. In the recent literature, Haynaldia villosa (L.) Schur is most commonly known as D. villosum, although the former name is still occasionally used (Gradzielewska 2006a, b; de Pace and Qualset 1995).

Dasypyrum villosum (L.) P. Candargy–mosquito grass–[syn.: Agropyron villosum (L.) Link, Haynaldia villosa (L.) Schur, Secale villosum L., Pseudosecale villosum (L.) Degan; Triticum villosum (L.) Link] is an annual species with 20–100 cm long culms. Blades are light green, spikes are 4–12 cm long, and glumes have tufts of hair on the two keels, with a tuft of stiff hairs below the awns, which are straight and 15–60 mm long; anthers are 4–7 mm long. In contrast, Dasypyrum breviaristatum (H. Lindb.) Frederiksen–[syn. D. hordeaceum (Cosson & Durieu) Candargy; Haynaldia hordeacea (Coss. & Durieu) Hack] is perennial, has short rhizomes, and features dark green leaves, glume keels with hairs that are not in tufts, and awns that are ~ 15 mm long.

Morphologically, the most conspicuous evolutionary divergence between diploid and tetraploid D. breviaristatum and D. villosum is apparent in the vegetative propagation device, with presence of rhizomes in diploid and tetraploid D. breviaristatum and absence of rhizomes in D. villosum (de Pace et al. 2011). These differences clearly reflect the major trends of adaptive radiation between D. villosum and the two cytotypes of D. breviaristatum, that occurred during colonization of high altitude habitats and further adaptation and differentiation of the tetraploid cytotype of D. breviaristatum to the environmentally disturbed habitats in forests and pastures at high altitude (de Pace et al. 2011). The F1 hybrids between D. villosum and D. breviaristatum produce rhizomes, indicating that the perenniality trait is dominant (Ohta and Morishita 2001; Blanco and Simeone 1995).

All D. villosum and D. breviaristatum plants show dimorphism for kernel color, with a yellow and dark red kernel within every spikelet (Onnis 1967). Yellow kernels are more frequent on the second floret, are heavier than dark-red kernels, and germinate faster (de Pace et al. 1994). Similar dimorphism in kernel color exists in wild emmer, Triticum turgidum subsp. dicoccoides, where the light-colored kernels, develop in the second floret, germinate in the first year after seed dispersal, and the dark–colored kernels develop on the lower floret, in the second year. The inheritance of the kernel color does not show any Mendelian segregation differences are reported between the kernels color classes: the dark-red seeds have longer seed dormancy than the yellow ones and maintain longer germination ability (after eight years of storage) than that of the yellow seeds (Stefani et al. 1998).

The distribution area of the genus Dasypyrum is in the southwestern part of the distribution region of the sub-tribe Triticineae, and it is known more as the region of D. villosum than as the geographical range of D. breviaristatum (Sarkar 1957). The core distributional center of D. villosum is in the Mediterranean Basin of southern Europe, e.g., Italy (including Sicily and Sardinia), Slovenia, Croatia, Bosnia-Erzegovinia, Serbia, Albania, Macedonia, Greece, (including Crete). It also sporadically grows in Spain (the Baleares Islands), southern France, southern Switzerland, Austria, Hungary, Romania, Bulgaria, Moldova, Ukraine (Krym), Turkey, Caucasus, Armenia, Azerbaijan, Georgia, Southeastern Russia, and western Turkmenistan (Maire 1952: Frederiksen 1991a; de Pace et al. 2011). In its core distributional centers in southern Europe, D. villosum is a vigorous plant that grows at low altitudes and is absent in habitats above 1350 m (de Pace et al. 2011). It is common in open herbaceous plant formations, often in dense stands, and also occupies disturbed habitats (de Pace et al. 2011). Genetic studies revealed lower inter-population and higher intra-population genetic diversity, as expected of an out-crosser species (de Pace et al. 2011). However, there is evidence of a positive relationship between spatial distance and genetic distance (de Pace et al. 2011). Therefore, to capture more genetic variation of D. villosum, samples should be collected from distant populations.

The distribution of Dasypyrum breviaristatum is rather restricted, and is mainly outside the range of the annual D. villosum. It grows in two isolated mountainous regions, each located over 1000 m above sea level, i.e., the Atlas Mountains of Morocco and Algeria, and Mt. Taygetos, in the Peloponnisos, Greece (Frederiksen 1991a; Ohta and Morishita 2001). Both the diploid and the autotetraploid grow in mixed populations, but the majority of the plants are tetraploids (Ohta et al. 2002). Recent investigations show that the diploid cytotype of D. breviaristatum is found only in Morocco and Algeria, while the tetraploid cytotype grows in Greece as well (Sarkar 1957; Ohta and Morishita 2001; Ohta et al. 2002).

Contrary to D. villosum, D. breviaristatum is common in the pastures and forests of the mountains of Algeria and Morroco, at an altitude ranging from 1000 to 2200 m above sea level. In Mt. Taygetos, Greece, it was found at an altitude of 1080 m (Frederiksen 1991a). The habitats of diploid D. breviaristatum in Morocco are disturbed oak forests and calcareous bedrock (Ohta et al. 2002). The distribution of the diploid is more restricted than that of the tetraploid. These ecological aspects of local, narrow distribution of the diploid cytotype and more expansive geographic distribution of the tetraploid cytotype, match the trends observed for other diploid–tetraploid taxa in the Triticeae (e.g., Zohary and Feldman 1962).

5.6.2 Cytology, Cytogenetics and Evolution

5.6.2.1 Karyotype and Genome Size

The karyotype of the Dasypyrum species is symmetric. D. villosum contains five pairs of metacentric (of which two are satellited (SAT)-chromosomes) and two pairs of sub-metacentric chromosomes (Linde-Laursen and Frederiksen 1991). de Pace et al. (2011) reported on a similar karyotype, but with only one SAT-chromosome, in D. villosum. The diploid cytotype of D. breviaristatum contains six pairs of metacentric chromosomes, one being satellited, and one pair of sub-metacentric chromosomes (Ohta et al. 2002), while the tetraploid cytotype has 13 pairs of metacentric chromosomes (of which three pairs are SAT-chromosomes—two with large and one with small satellites), and one pair of sub-metacentric chromosomes (Linde-Laursen and Frederiksen 1991). Another line of this species contains only two pairs of SAT-chromosomes with large satellites (Linde-Laursen and Frederiksen 1991), indicating some degree of polymorphism regarding the number of SAT-chromosomes in tetraploid D. breviaristatum. In addition, nucleolar dominance of the breviaristatum-NOR region on the villosum NOR and wide C-band karyotype differences between the genomes of the two species were observed by Linde-Laursen and Frederiksen (1991).

The use of different banding techniques, including staining with fluorochromes, C-banding, and Ag-NOR (Gill 1981; Linde-Laursen and Frederiksen 1991; Blanco et al. 1996), and chromosomal localization of a species-specific 380 bp long satellite DNA sequence (de Pace et al. 1992), allowed for reliable identification of each chromosome pair of D. villosum, in different genomic backgrounds after interspecific or intergeneric hybridization. In contrast, chromosomal identification was not possible in D. breviaristatum because of the overall similarity of banding patterns and chromosome morphology (Linde-Laursen and Frederiksen 1991).

The 1C genome size of D. villosum is 5.065 pg (Obermayer and Greilhuber 2005; Eilam et al. 2007), although large intraspecific variation, either between or within populations, has been detected (Greilhuber 2005). Genome size of tetraploid D. breviaristatum is twice that of D. villosum (Blanco et al. 1996).

5.6.2.2 Cytogentic Relationship Within and Between the Two Cytotypes of D. breviaristatum

To elucidate the cytogenetic relationship between the diploid and the tetraploid cytotypes of Dasypyrum breviaristatum, the two cytotypes were reciprocally crossed with one another and chromosome pairing at first metaphase of meiosis and fertility were studied in the F1 hybrids (Ohta and Morishita 2001). The researchers used two diploid and nine tetraploid plants of D. breviaristatum for within and between cytotype crossings. The diploids were collected from one population in Morocco, while two of the tetraploids were collected from the same population as the diploid plants, five from other regions of the Atlas Mountains and two from Greece (Ohta and Morishita 2001). F1 hybrid between the two diploid plants exhibited almost complete chromosomal pairing at MI (0.14 univalents, 6.92 bivalents, and 0.003 quadrivalents per cell; 12.70 chiasmata/cell.) and pollen fertility of the F1 hybrid was high (82.3%). The F1 hybrids between the different ecotypes of the tetraploid cytotype from Morocco, displayed chromosomal configurations typical of autotetraploid (0.14–0.75 univalents, 3.90–5.81 bivalents, 0.06–0.78 trivalents, 2.54–4.45 quadrivalents, and 0.02–0.26 hexa- or octo-valents) configurations. The range of chiasmata/cell was 23.75–25.44 and pollen fertility was high (69.2–77.9%). The F1 hybrid between the tetraploid plants from Morocco and Greece showed somewhat more univalents, bivalents, and trivalents per cell (1.01–1.61 univalents, 6.21–6.30 bivalents, 0.46–0.87 trivalents), fewer quadrivalents and higher configurations per cell (2.77–3.28 quadrivalents, 0.01–0.02 higher configurations), and somewhat fewer chiasmata/cell (22.00–23.22). Nevertheless, pollen fertility was as high as that of the F1 generated tetraploid cytotypes from different ecotypes in Morocco. The pattern of chromosomal pairing at meiosis in hybrids between tetraploid Moroccan ecotypes from different site of Morocco, indicated that this hybrid was an autotetraploid and, therefore, the tetraploid cytotype is an autotetraploid.

The presence of higher multivalents (hexa- and octo-valents) in several crosses between tetraploid D. breviaristatum ecotypes indicated the occurrence of reciprocal translocations between these ecotypes (Ohta and Morishita 2001). Ohta and Morishita further revealed the presence of aneuploid plants in natural populations of the tetraploid cytotype. Such chromosomal variation is also maintained in the natural populations of tetraploid D. breviaristatum by vegetative propagation. These researchers also observed the following mean pairing configurations in D. breuiaristatum (4x) x D. breuiaristatum (2x) F1 hybrid: 3.38 univalents, 3.20 bivalents, 3.74 trivalents, and 0.005 quadrivalents per cell. The mean arm pairing frequency and relative affinity were 0.915 and 0.641, respectively, indicating homology of the diploid genome to the two genomes of the tetraploid cytotype. The seven trivalents, observed in many meiocytes of the F1 hybrid between the diploid and the tetraploid cytotypes of D. breviaristatum, supported this conclusion.

5.6.2.3 Origin of the Tetraploid Cytotype of D. breviaristatum

The origin and genomic composition of the tetraploid cytotype of D. breviaristatum is under debate (de Pace et al. 2011). Several authors suggested an autoploid origin of the tetraploid cytotype (Sarkar 1957; Sakamoto 1986; von Bothmer and Claesson 1990; Galasso et al. 1997; Ohta and Morishita 2001; Ohta et al. 2002). Sarkar (1957) studied chromosomal pairing at meiosis of the tetraploid cytotype and found 11.7 (2–14) bivalents and 1.1 (0–6) quadrivalents per cell. The occurrence of up to six quadrivalents in a cell may indicate an autoploid derivation of the tetraploid D. breviaristatum, making the diploid cytotype is the most likely ancestral type (Sarkar 1957).

The pattern of chromosomal pairing at meiosis observed in F1 hybrids between the tetraploid and the diploid cytotypes (Ohta and Morishita 2001), and particularly the existence of seven trivalents in several meiocytes of the F1 triploid hybrid, supported the autotetraploidy hypothesis of Sarkar (1957) regarding the speciation event that led to formation of the tetraploid D. breviaristatum. A similar conclusion, based on the similarity in karyotype and in plant morphology of the two cytotypes of D. breviaristatum, was reached by Ohta et al. (2002), who also suggested that the diploid cytotype is the most probable ancestral form of the tetraploid cytotype. Indirect evidence for this hypothesis came from experiments of Nakajima (1960), where the diploid cytotype of D. breviaristatum produced several unreduced gametes. Also, the fact that diploid D. breviaristatum is perennial, as is the tetraploid cytotype, designates the diploid cytotype as the most likely progenitor in which the genome duplication event occurred. This hypothesis is also supported by the studies of Sakamoto (1986), and von Bothmer and Claesson (1990).

On the other hand, using DNA fragment analyses and isozymes, as well as FISH of structural genes sequences, Blanco et al. (1996) concluded that tetraploid D. breviaristatum originates from D. villosum. Other authors proposed an allopolyploid origin for tetraploid D. breviaristatum, most likely with D. villosum as one of the parents and diploid D. breviaristatum as the other parent (Frederiksen 1991a; Linde-Laursen and Frederiksen 1991). However, a direct derivation from D. villosum was ruled out following molecular cytogenetic analyses (Galasso et al. 1997), RAPD analyses of genomic DNAs (Yang et al. 2006), and studies of the meiosis in reciprocal crosses of the two species (Sakamoto 1986; Ohta and Morishita 2001). Following failure to detect D. villosum-specific DNA sequences in tetraploid D. breviaristatum fluorescent in situ hybridization (FISH), Uslu et al. (1999) concluded that D. villosum is not related to tetraploid D. breviaristatum. Therefore, the hypothesis that the tetraploid cytotype of D. breviaristatum was derived from the diploid cytotype of this species by autopolyploidy, is the only one, to date, that benefits from concurrent support (Ohta et al. 2002; de Pace et al. 2011).

5.6.2.4 Cytogenetic Relationships Between D. breviaristatum and D. villosum

To clarify the genomic relationships between the two species of Dasypyrum, Ohta and Morishita (2001) reciprocally crossed the two cytotypes of D. breuiaristatum with D. villosum and examined chromosome pairing at the first metaphase of meiosis and fertility in the F1 hybrids. Some seed setting (36.1%) was obtained in D. villosum x diploid D. breviaristatum; 30% of the shriveled caryopses germinated. The reciprocal cross did not produce any seeds. The mean pairing configurations and mean arm pairing frequency per ell in the diploid D. villosum x D. breuiaristatum (2x) hybrids were 11.12 univalents and 1.44 bivalents, and 0.107 mean arm pairing, and the hybrids were almost completely sterile. Based on these results, Ohta and Morishita (2001) concluded that the genome of diploid D. breuiaristatum is only distantly related to that of D. villosum. They, therefore, proposed, in accordance with Wang et al. (1995), to use of the symbol Vb for the haploid genome of the diploid cytotype of D. breuiaristatum and Vv for the haploid genome of D. villosum. Furthermore, since they concluded that tetraploid D. breviaristatum is an autotetraploid, with diploid D. breviaristatum as the immediate ancestor, they proposed the genome symbol VbVb for the haploid genome of tetraploid D. breviaristatum.

A similar pattern of seed set and chromosome homoeology was observed after hybridization of D. villosum (as female) with tetraploid D. breviaristatum. About 50% and 12% F1 hybrids seed setting was reported by Ohta and Morishita (2001) and Blanco et al. (1996), respectively, and over 80% of them germinated. The reciprocal combination did not produce any seed. Chromosome pairing at meiosis of the F1 hybrid between D. villosum and tetraploid D. breviaristatum from Morocco contained 8.75–9.49 univalents, 5.51–5.68 bivalents, and 0.05–0.25 trivalents per cell, as well as 8.70–8.76 chiasmata/cell (Ohta and Morishita 2001). In contrast, the F1 hybrid between D. villosum and tetraploid D. breviaristatum from Greece showed somewhat higher chromosomal pairing (7.28 univalents, 6.64 bivalents, 0.14 trivalents, and 0.004 quadrivalents per cell; 11.58 chiasmata/cell). The higher chromosomal pairing in the hybrid between D. villosum and the Greece ecotype of tetraploid D. breviaristatum may indicate exchange of some chromatin between these two taxa (Ohta and Morishita 2001). The very low pairing between the F1 hybrid chromosomes of D. villosum and diploid D. brevisaristatum clearly indicated no homology between their genomes, which is supported by the almost complete sterility of the F1 hybrid (Ohta and Morishita (2002). This conclusion coincides with previous results reached following karyotype analysis (Ohta et al. 2002).

Sakamoto (1986) also achieved interspecific hybridization between D. villosum, used as female, and tetraploid D. breviaristatum, as male. At meiosis, the triploid F1 hybrid displayed chromosome configurations with an average of 6.5 bivalents and 7.9 univalents per cell. As the tetraploid parents contained many quadrivalents, the bivalents observed in the interspecific hybrid were ascribed to autosyndesis of the tetraploid D. breviaristatum chromosomes (Sarkar 1957; Ohta and Morishita 2001).

Because in the pollem mother cells (PMCs) of F1 of D. villosum x tetraploid D. breviaristatum, the observed number of trivalents per cell was 0.14, while 3.74 were observed in the cross of the diploid cytotype and the tetraploid cytotype of D. breviaristatum, the contribution of D. villosum genome to the tetraploid was ruled out. In accord with this conclusion, Galasso et al. (1997) observed seven bivalents of D. breviaristatum and seven univalents of D. villosum after simultaneously labeling F1 hybrid D. breviaristatum (4x) and D. villosum chromosomes in first meiotic metaphase with fluorescein (FITC)-labeled D. breviaristatum DNA. Yang et al. (2005, 2006) confirmed the divergence of the Vv and Vb genomes using GISH and RAPD markers. When the entire Vb genomic DNA was labeled to hybridize a somatic mitotic metaphase of a partial amphiploid with 42 chromosomes (genome BBAAVbVb), generated from the selfing population of the amphiploid bread wheat (cv. Chinese Spring)-D. breviaristatum, 14 chromosomes were strongly and uniformly labeled to Vb (Yang et al. 2005). In contrast, when D. villosum was used as a probe, many arms of the 14 Vb chromosomes displayed large regions in their distal half with less intense labeling. These differentiated GISH patterns not only reflected the large genomic divergence between the Vb and Vv genomes, as described by Galasso et al. (1997), but also helped identify the chromosome pairs of D. breviaristatum in wheat background labeled with Vv-DNA.

5.6.2.5 Cytogenetic Relationships Between Dasypyrum Species and Species of Other Triticineae Genera

Several researchers crossed D. villosum with various species from other Triticeae genera, in pursuit of identification of cytogenetic and evolutionary relationships between these species. A complete list of hybrids between Dasypyrum species and Triticeae species and their level of pairing at first meiotic metaphase, is provided in Tables 4.7 and 4.8 of the publication of de Pace et al. (2011). Little chromosome homology was observed between Dasypyrum species and any other species of the Triticeae. Similarly, hybridizations performed by Lucas and Jahier (1988) between D. villosum and other diploid Triticeae species, indicated a low average number of chromosome pairings between homoeologous arms.

Crosses between diploid wheat (wild and domesticated T. monococcum and T. urartu) with D. villosum produced F1 hybrids that displayed very low chromosomal pairing at first meiotic metaphase (Sando 1935; Kihara 1937; Sears 1941; Lucas and Jahier 1988; von Bothmer and Claesson 1990). Similarly, studies of chromosome pairing in F1 hybrids between tetraploid wheat, T. turgidum ssp. durum, and D. villosum, revealed very low pairing at meiosis, demonstrating that the Vv genome is not homologous to the A and B subgenomes of tetraploid wheat (Kihara and Nishiyama 1937; Nakajima 1966; von Blanco et al. 1983a, b, 1988b). Using the C-banding technique on PMCs of the F1 T. turgidum ssp. durum cv. Capelli x D. villosum hybrid, Blanco et al. (1988b) estimated that of the 194 observed bivalents, 82% involved A-B chromosome associations, 11.3% A-Vv, 1% B-Vv, 4.1% A-A, 1.6% B-B, and 0% VvVv, and deduced that A and B subgenomes are related, while the Vv genome is more distant, but closer to the A than to the B genome.

Yu et al. (1998, 2001) analyzed chromosome pairing in F1 hybrids between the common wheat cultivar Chinese Spring and D. villosum. On average, they observed 1.61 bivalents per cell. Chen and Liu (1982) reported a cytogenetic study of T. aestivum x D. villosum F1 plants, in which they were able to identify the villosum chromosomes in the T. aestivum background and found very little pairing at first meiotic metaphase, i.e., 22.5–27.8 univalents, 0.11–0.37 bivalents, and 0.10 trivalents per cell. The F1 hybrid had reduced vigor and tillering ability, as compared to Chinese Spring (Chen and Liu 1986).

Nakajima (1953) studied meiosis in the F1 hybrid of T. timopheevii x D. villosum, and noted, on average, one bivalent per cell. He later reported (Nakajima 1960) meiotic chromosome pairing in F1 hybrids between Triticum turgidum and tetraploid D. breviaristatum, and concluded that there is no homology between the genomes of D. breviaristatum and the two subgenomes of T. turgidum. The mean bivalent frequency of 8.06 per cell in the tetraploid F1 hybrid of this cross, and 8.9 in the pentaploid F1 T. aestivum x tetraploid D. breviaristatum hybrid, resulted from autosyndesis of the breviaristatum chromosomes. The combination T. turgidum subsp. durum x tetraploid D. breviaristatum produced an average of 7.86 bivalents and 11.5 chiasmata/cell (Blanco and Simeone 1995). Seven out of the 7.86 bivalents resulted from autosyndesis of the breviaristatum chromosomes.

The T. aestivum x tetraploid D. breviaristatum F1 hybrid seed set was 2.9% (von Bothmer and Claesson 1990). Two hybrid plants displayed a high level of pairing, with up to 12 bivalents and an average of 7.8 bivalents. The pairing of seven bivalents was attributed to autosyndesis of breviaristatum chromosomes.

Oehler (1933, 1935), Sando (1935), Kihara and Lilienfeld (1936), von Berg (1937), Sears (1941), Lucas and Jahier (1988), and Deng et al. (2004) crossed diploid and tetraploid species of Aegilops with D. villosum. The F1 hybrids displayed very low chromosomal pairing and were completely sterile. An intergeneric triploid hybrid between Aegilops tauschii and the tetraploid cytotype of D. breviaristatum was produced and at meiosis, displayed 8.80 univalents and 6.15 bivalents per cell (Sakamoto 1986).

Sando (1935) produced hybrids between Secale fragile and D. villosum. In general, the F1 plants resembled the Secale parent. Kostoff and Arutiunova (1937) analyzed pairing in a trigeneric hybrid (T. turgidum subsp. dicoccon x D. villosum) x S. cereale (genome BAVvR) and found that Vv chromosomes were not homologous with the Secale chromosomes (genome R) and both Vv and R chromosomes were not homologous to the wheat A or B chromosomes. Similarly, Nakajima (1951) observed very low pairing in MI of the F1 hybrid between D. villosum and Secale cereale and concluded that no homology existed between the Vv and R genomes. Extremely low pairing ability was also observed between Vv and R chromosomes in PMCs of hybrid plants generated by crossing two amphiploids: Ae. UniaristataD. villosum (2n = 28; genome NNSvSv) and Ae. uniaristataS. cereale (2n = 28; genome NNRR) (Jahier et al. 1988). However, Jahier et al. (1988) did not reject the working hypothesis that Vv and R chromosomes share homologous sequences. Rather, they attributed the Vv and R asynapsis to factors such as asynchronous meiotic rhythm between R and Vv genomes. Similar causes can explain the lack of pairing between Vv and Vb reported above.

5.6.3 Phylogeny and Time of Origin

The shared peculiar spike morphology, unilocus molecular and biochemical markers (Blanco et al. 1996), and signals on the chromosomes of tetraploid D. breviaristatum, but not on most other Triticeae species, hybridization with the pHv62 D. villosum-species-specific repeated sequence (Uslu et al. 1999), seen between the two species of Dasypyrum, are indicative of common ancestry. The genomic distance between D. villosum and D. breviaristatum, as determined by 301 RAPD loci, was smaller than their distance from Secale species (Yang et al. 2006). Therefore, the formation of the dasypyrum species and their biological and taxonomical status may be explained by a cascade of events, which began in the earlier stages of Triticineae separation from the Hordeineae (13–15 MYA), and continued through the reproductive isolation of the lower-altitude D. villosum ecotypes from the high–altitude diploid D. breviaristatum prototype, followed by the autopolyploidization event of the tetraploid cytotype from the diploid one, and incipient reproductive isolation between the two cytotpes. Such divergence has not occurred for other syntenic and gene-rich DNA segments of genomes Vv and Vb, as suggested by the strong similarity between D. villosum and D. breviaristatum genomes in restriction fragment patterns of genomic DNA, the phenotypes for some isozyme systems, and the location of gliadin genes (Blanco et al. 1996).

The phyletic relationships within the Dasypyrum genus and among Dasypyrum species and other Triticineae species, has been assessed at the levels of morphology, protein, chromosome, chloroplast and nuclear fragments and nucleotide sequences. Morphology-based phylogenetic analyses showed that Dasypyrum branched in a sister group of Secale within the same clade (Baum 1978a, b, 1983; Kellogg 1989; Frederiksen and Seberg 1992; Seberg and Frederiksen 2001). Kellogg (1989) placed Dasypyrum near Agropyron and Triticum monococcum, and Baum (1978a, b, 1983) considered Secale cereale and D. villosum as evolutionarily more contiguous to Triticum and Aegilops than to the rest of the Triticineae.

Phylogenetic studies based on molecular data suggested that Secale is the closest relative of the Triticum-Aegilops genera (Kellogg et al. 1996; Huang et al. 2002a; Mason-Gamer et al. 2002). When Hordeum and Dasypyrum were analyzed with Secale, Triticum and Aegilops, they were positioned at the base of the tree topology, as out-groups (Yamane and Kawahara 2005; Kawahara et al. 2008), implying a much earlier divergence between D. villosum and the common ancestor of Triticum-Aegilops.

The analysis from two cpDNA data sets, one based on restriction site variation (Mason-Gamer and Kellogg 1996b) and the other on sequences encoding the rpoA-subunit of the RNA-polymerase (Petersen and Seberg 1997), placed Elymus species possessing the E genome, e.g., E. elongatus and E. farctus, and Dasypyrum together on cpDNA cladograms. Sequencing of the nuclear starch synthase gene also revealed a close affinity between E. farctus and Dasypyrum (Mason-Gamer and Kellogg 2000). The cpDNA tree contained a well-supported clade, including Dasypyrum and diploid Elymus with St genome. The Dasypyrum-Elymus with St genome monophyly was also observed in cladograms obtained from similar RFLP profiles of 14 cloned fragments covering the entire cpDNA of T. aestivum (Kellogg 1992), morphological data (Kellogg 1989), and 5S RNA (Appels and Baum 1991).

However, the nuclear DNA data (Hsiao et al. 1995; Kellogg and Appels 1995) are incongruent with the cpDNA data, in that they suggest different affinities of diploid Elymus with St genome and Dasypyrum within Triticineae. The Heteranthelium element of the transposon Stowaway is present in Dasypyrum, but absent in other Triticeae species (Petersen and Seberg 2000). Heteranthelium piliferum and D. villosum stand at one extreme of the phylogenetic relationships, determined by variation in the PCR sequences of 6-SFT (sucrose:fructan 6-fructosyltransferase), whereas diploid Secale, Triticum and Aegilops species are at the other extreme (Wei et al. 2000). Molecular phylogeny of the RPB2 (the second-largest subunit of RNA polymerase II) gene sequence revealed that the Vv genome of D. villosum is sister to the St genome of Elymus and that both diverged from the H-genome of barley (Sun et al. 2008). These findings fall in line with the phylogenetic relationships of mono-genomic species of Triticeae inferred from nuclear rDNA (internal transcribed spacer) sequences, where Heteranthelium and Dasypyrum demonstrated close relation to diploid Elymus with St genome.

DNA/DNA hybridization of the genomes of rye and D. villosum with labeled nuclear DNA from wheat and rye, revealed greater homology between the Vv- and R- than between Vv and A-, B-, and D-subgenomes of wheat (Lucas and Jahier 1988). Fish analysis involving hybridization of the genomes of different Triticeae species with species-specific molecular probes prepared from tandem repeated DNA sequences of D. villosum (pHv62) and S. cereale (pSc119.2), demonstrated a greater homology between the R- and Vv-genomes than between R- or Vv-genomes and between those of Triticum and Aegilops (Uslu et al. 1999).

Escobar et al. (2011; Fig. 2.2) obtained the most comprehensive molecular dataset to date in Triticeae, including one chloroplast and 26 nuclear genes. They found that Dasypyrum, Heteranthelium and genera of clade V, grouping Secale, Taeniatherum, Triticum and Aegilops, evolved in a reticulated manner. Their evidence supported the following clades (Escobar et al. 2011): The first includes Australopyrum (clade IIIA), Henrardia and Eremopyrum bonaepartis (clade IIIB), and Agropyrum and E. triticeum (clade IIIC), while the second consists of Dasypyrum and Heteranthelium (clade IV), on the one hand, and Secale, Taeniatherum, Triticum and Aegilops (clade V), on the other hand. St genome Elymus does not group with Hordeum but is sister to Dasypyrum. Consequently, Heteranthelium branches at the base of clade V and these two newly inferred clades (Elymus-Dasypyrum and Heteranthelium-clade V) are closely related to each other.

Lucas and Jahier (1988) concluded that the differentiation of D. villosum (and Secale cereale) from the genera Aegilops and Triticum occurred earlier than speciation in the latter two genera. They assumed that the closest Aegilos-Triticum species to D. villosum is wild diploid wheat, T. monococcum ssp. aegilopides, and not diploid species of Aegilops.

5.6.4 Use of Dasypyrum in Wheat Improvement

5.6.4.1 Production of Cytogenetic Lines Facilitating Identification of Useful Genes, Their Allocation to Chromosomes and Construction of Genetic Maps

D. Villosum has been crossed to Secale, Aegilops, Agropyron, and Triticum species, but permanent introgression of its chromosomal segments occurs only in wheat, following controlled backcrosses (de Pace et al. 2011). At the beginning of the twentieth century, Nazareno Strampelli was the first to show the feasibility of crossing wheat species with D. villosum and of transferring genes from the wild species to tetraploid and hexaploid wheat, by crossing, backcrossing and then implementing proper selection rules (Strampelli 1932). The T. aestivum cv. Rieti x D. villosum cross produced an F1 hybrid, from which (after a putative process of backcrossing to cv. Rieti) the “gigas” winter bread wheat cultivar Cantore was derived (Strampelli 1932). Strampelli succeeded to release several interesting cultivars of bread wheat, such as the so-called Triticum giganteum (which produced large spikes and kernels as large as a coffee seed) or lines with sweet kernels.

Following the pioneering work of Strampelli, several researchers used D. villosum in intergeneric hybridizations to study the morphology, chromosomal pairing at meiosis, and fertility of the hybrids. The studies were of interest to the evolutionist, biosystematist, and cytogeneticist, but they rarely went further in the backcrossing and selection programs for releasing cultivars. However, several cytogenetic lines have facilitated meaningful assessment and use of the potential of D. villosum for wheat improvement.

Production of wheat (tetraploid and hexaploid)—D. villosum amphiploids was the first methodological approach used to assess the potential of Dasypyrum species to contribute genes that may improve wheat. Tschermak-Seysenegg (1934) produced the first hexaploid amphiploid (2n = 6x = 42; genome BBAAVvVv) from the cross of T. turgidum with D. villosum. McFadden and Sears (1947) produced the hexaploid amphiploid T. turgidum ssp. dicoccoidesD. villosum and Jan et al. (1986) produced the octoploid amphiploid T. aestivum cv. CS–D. villosum with 2n = 8x = 56 (genome BBAADDVvVv). Several other amphiploid BBAADDVvVv were produced by Mini et al. (1988), but all octoploid amphiploids had agronomically poor plant type.

In contrast to the octoploid amphiploids, hexaploid amphiploids (T. turgidum ssp. durumD. villosum) proved more promising. Such amphiploids were produced by several researchers (see review in de Pace et al. 2011), and exhibited a plant habitus and spike morphology resemblant of wheat, good seed quality, but brittle rachis. As a primary amphiploid, the overall performance, with the exception of brittle rachis, of the hexaploid BBAAVvVv was equal to or better than, most hexaploid primary triticale. Thus, the BBAAVvVv amphiploids deserve consideration as new crop plants, much as triticale did in its early stages of development (de Pace et al. 2011).

Dasypyrum addition and substitution lines in wheat are suitable material for studying the role of single Dasypyrum chromosomes in determining various traits in the wheat background. Sears (1953) and Hyde (1953) produced the first set of six out of seven possible monosomic addition lines, using D. villosum chromosomes to complement the standard laboratory bread wheat cultivar Chinese Spring host genome. Later, Lukaszewski (1988) completed the set of seven monosomic addition lines. Using C-banding, plant morphology, and molecular markers, Lukaszewski (1988) assessed the homoeology of the added Vv chromosomes to those of wheat, and assigned each chromosome to the wheat homoeologous groups. Similarly, Blanco et al. (1987) produced a set of six monosomic addition lines by donating D. villosum chromosomes to the durum wheat cultivar Creso. Each added Vv chromosome had a specific effect on plant morphology and fertility. Liu et al. (1995) described six D. villosum substitution lines in bread wheat and reported the homoeology assignment of the Vv chromosomes to the six different wheat homoeologous groups.

Various novel disease-resistance genes have been identified on specific Vb chromosomes of the perennial tetraploid D. breviaristatum. Addition and substitution lines were isolated in the progeny of wheat—D. breviaristatum amphiploids crossed with cultivated wheat, including different addition lines carrying genes for stripe rust (Yang et al. 2008), as well as stem rust and powdery mildew (Liu et al. 2011a, b) resistance. Marker data indicated that the Vb chromosomes in the latter two addition lines were rearranged with respect to wheat homoeologous groups. On the other hand, various molecular markers confirmed a group 2 homoeology for the Vb chromosome substituted into a Chinese bread wheat in place of chromosome 2D, able to confer stripe rust resistance at the adult plant stage (Li et al. 2014). Interestingly, FISH, C-banding, and PCR-based molecular marker analyses indicated that the 2Vb of D. breviaristatum was completely different from 2Vv of D. villosum, in line with the current view about the origin of 4 × D. breviaristatum.

5.6.4.2 Production of Translocation Lines of D. villosum Chromosomal Segments in Wheat Chromosomes

5.6.4.2.1 Production of Translocations via Induction of Homoeologous Pairing

Halloran (1966) crossed D. villosum with the bread wheat cultivar Chinese Spring monosomic for chromosome 5B, and obtained two types of hybrids, with (2n = 28) and without (2n = 27) chromosome 5B. Very low pairing (0.25 bivalent/cell) was observed in meiosis of the 28-chromosome hybrid containing 5B, leading to the conclusion that D. villosum does not possesses gene(s) that remove the inhibition to homoeologous pairing due to the Ph1 gene in chromosome 5B (Halloran 1966). The 27-chromosome hybrid (deficient for chromosome 5B) showed much higher chromosome pairing (9.6 univalents, 3.8 rod and 1.06 ring bivalents, 0.86 trivalents and 0.7 quadrivalents) than the 28-chromosome hybrid (Halloran 1966). Yu et al. (1998, 2001) also noted enhanced homoeologous pairing between wheat and D. villosum chromosomes in the absence of Ph1. These observations indicated that it is possible to induce transfer of chromosomal segments from D. villosum to bread wheat by promoting homoeologous pairing in the absence of Ph1 or in the presence of its mutant ph1ph1.

In contrast to Halloran (1966), von Bothmer and Claesson (1990) suggested that D. villosum genotyopes might influence the pairing frequency in the F1 hybrids between D. villosum and Triticum-Aegilops species. Likewise, Blanco et al. (1983a, b) concluded that D. villosum contains genes that promote homoeologous pairing in the presence of Ph1, explaining the similar proportions of homoeologous pairings observed in F1 hybrids between T. turgidum and D. villosum, in the presence of Ph1 or its mutant allele ph1. To explain the discrepancy between Halloran’s results and theirs, Blanco et al. (1988b) presumed that ecotypes of D. villosum may vary in their ability to promote homoeologous pairing in the presence of Ph1. Genetic variation for the promotion of homoeologous pairing has been demonstrated in various accessions of several Triticeae species, e.g., Amblyopyrum muticum (Dover and Riley 1972), Ae. speltoides (Dvorak 1972), Ae. longissima (Mello-Sampayo 1971a), Elymus elongatus (Mochizuki 1962), and Secale cereale (Dvorak 1977) (Table 5.2).

Yu et al. (1998, 2001) also analyzed chromosome pairing in F1 hybrids of the bread wheat cultivar Chinese Spring (with Ph1) and its ph1b mutant (a defciency for Ph1) with D. villosum. On average, 1.61 chromosomes per cell paired in the hybrid with Ph1, but 14.43 in the hybrid with ph1b. GISH revealed three types of homoeologous associations between wheat (W) and D. villosum (D) chromosomes (W-D, D-W-W and D-W-D) in PMCs of the CSph1 x D. villosum hybrid, and only one type (W-W) in the CSPh1 x D. villosum hybrid. Translocations of chromosome segments or entire arms, were detected by GISH in the BC1 plants from the backcross of CSph1 x D. villosum to CSph1b.

5.6.4.2.2 Production of Translocations via Irradiation

Irradiating mature female or male gametes of plans having addition or substitution of whole chromosomes or chromosome arm of D. villosum in either bread or durum wheat background, with 60Co-gamma-ray, is a new and highly efficient means of eliciting small segment structural changes in chromosomes, especially interstitial translocations of D. villosum segments in wheat chromosomes (Chen et al. 2008; Cao et al. 2009; Bie et al. 2007).

5.6.4.3 Allocation of Useful Genes to Chromosomes

The various genetic lines mentioned above allowed allocation of useful genes of Vv to chromosomes of Vv, namely, genes for morphological (e.g., Sears 1982a, b; Mariani et al. 2003; Chen et al 2008), biochemical (e.g., Resta et al. 1987; Shewry et al. 1987, 1991; Montebove et al. 1987; Blanco et al. 1991; de Pace et al. 1988, 1992, 2011), molecular (e.g., Gil and Appels 1988; de pace et al. 1992; Liu et al. 1995; Galasso et al. 1997), disease resistance (e.g., Pasquini et al. 1978; Panayotov and Todorov 1979; Chen et al. 1997;Yildirim et al 1998, 2000; Oliver et al. 2005; Huang et al. 2007; Bizzarri et al. 2009), and abiotic stresses tolerance (e.g., Scarascia-Mugnozza et al. 1982; Schlegel et al. 1998) genes. An alternative approach in which genes are assigned to specific chromosomes of D. villosum using nullisomic amphiploids, has been proposed (Zhong and Qualset 1990).

5.6.4.4 Contribution of Dasypyrum Genes to Wheat Improvement

In an evaluation of forage crop potential, the amphiploid T. turgidum ssp. DurumD. villosum showed high biomass quality (N yield) and quantity (biomass) (de Pace et al. 1990). The role of D. villosum chromosome segments introgressed in hexaploid wheat in pre-breeding and primary population mapping for complex genetic trait analysis, was evidenced by Mariani et al. (2003). Vaccino et al. (2007) found that wheat lines with chromosome 1BL containing chromatin introgressed from D. villosum, were early-heading, good grain yielders, improved bread making quality and were environmentally stable over the years. Analysis of several hexaploid lines derived from the backcross of the F1 hybrid T. turgidum ssp. durum x D. villosum to bread wheat, exhibited good agronomic performance in field trials, in term of yield, kernel weight, and bread-making quality when compared to the best standards (Vaccino et al. 2009). Successful transfer of the D. villosum gene for powdery mildew resistance was described by several researchers, e.g., Blanco et al. 1988a; Shi et al. (1996) Chen et al. (2008), Liu et al. (1996), and Qi et al. (1995).

5.7 Heteranthelium Hochst

Heteranthelium is a monotypic genus containing the species H. piliferum (Banks et Sol.) Hochst. This species is annual, consisting of short plants with a very peculiar spike morphology that is different from that of other Triticeae genera (Fig. 5.4d). Its spike consists of two kinds of interspersing spikelets, fertile and sterile, both strikingly different in appearance. In the fertile spikelets the lower florets are bisexual and the upper ones are barren and scale-like, whereas the sterile spikelets are composed of barren scale-like florets only. The glumes and the lower parts of the awns are hairy. The spikelets are solitary at each rachis node. Rachis only partially fragile and consequently, ripe spikes of this species do not disarticulate between individual spikelets, but break up into a number of sections, each section contains at least three fused rachis-segments diminishing in length from below upwards, the lowest spikelet is fertile, the next above is smaller, fertile or sterile with a reduced lemma, and succeeding spikelet(s) reduced to a bunch of awned glumes. Thus, H. piliferum has an exceptional dispersal unit consisting of three spikelets of which 1–2 are sterile. The small size of the anthers (1 mm) indicates mating system of facultative self-pollination; bagged spikes were fertile almost as non-bagged spikes (Luria 1983). There is some variation in plants size, leaves size, spike color (from green to red), number of seeds in the fertile spikelets and in seed size.

Heteranthelium piliferum is native to the Eastern Mediterranean region and central Asia (Israel, Jordan, Syria, Lebanon, Turkey, Transcaucasia, Iraq, Iran, Pakistan, Afghanistan, Turkmenistan, Tajikistan and Kyrgyzstan). It is an Irano-Turanian element growing on rocky and dry slopes of foothills in semi-arid steppes as well as at the edges of the Mediterranean region. Its distribution is limited by drought (desert) and cold (higher mountainous regions). In the sub-Mediterranean regions it grows in 550–750 m above sea level in annual rainfall of 250–300 mm while in mountainous area (e.g., Hermon Mt.) it grows at elevation of 1400–1700 m above sea level, with annual rainfall of 1000–1300 mm, in plant community consisting of sub-Mediterranean and many Irano-Turanian plants.

H. piliferum is a diploid (2n = 2x = 14) (Sakamoto and Muramatsu 1965; Bowden 1966), with symmetric karyotype with median or sub-median centromere, one chromosome pair is satellite with a small satellite (Chennaveeraiah and Sarkar 1959; Ferederiksen 1993; Bowden 1966; Sakamoto 1974). Each chromosome had a distinct C-banding pattern and this technique provided adequate information to identify all of H. piliferum chromosomes (Asghari-Zakaria 2007). Homologous chromosomes were identified based on position of the centromere and similarities of C-banding patterns (Asghari-Zakaria 2007). Its genome symbol is Q (Wang et al. 1995).

In an attempt to elucidate the genetic relationships to other genera of the tribe, Sakamoto (1974) crossed H. piliferum, as the female parent, with a number of Triticeae species. Yet, only the inter-generic crosses of H. piliferum with 2 x Eremopyrum bonaepartis and autotetraploid hordeum depressum were successful. The F1 hybrid H. piliferum x E. bonaepartis showed poor growth and the morphology of the spikes was similar to that of Eremopyrum, while the spikelets were intermediate. Very little chromosome pairing was observed at first meiotic metaphase of the hybrid (0.04 bivalents per cell), indicating that the genome of H. piliferum is only distantly related that of E. bonaepartis (Sakamoto 1974). Growth of the F1 hybrid H. piliferum x tetraploid Hordeum depressum was vigorous and the spike morphology was intermediate between the parents, i.e., a solitary spikelet exists at each rachis node like the Heteranthelium parent but no sterile spikelets characteristic of Heteranthelium were found. The Heteranthelium characteristic of one spikelet per rachis node dominated the multiple spikelet character of Hordeum. Similar dominance of the solitary spikelet trait over multiple spikelets at each rachis node was also found in bread wheat x Hordeum vulgare hybrids (Muramatsu 2009). The triploid F1 hybrid H. piliferum x tetraploid Hordeum depressum exhibited 5.5 bivalents and 0.01 trivalents per cell, resulting from autosyndesis of the Hordeum parent chromosomes and indicating lack of homology between H. piliferum and H. depressum.

Considering the unique morphology of the spike and the dispersal unit, distribution in the east and southeast periphery of the distribution area of the sub-tribe Triticineae, inter-generic cross-ability and cytogenetic relationships of H. piliferum, Sakamoto (1974) concluded that the monotypic genus Heteranthelium is a distinctive entity, representing a specialized group that occupies an isolated position in in the tribe Triticeae. This taxon has evolved as an annual during the process of adaptation to rather dry habitats of the Mediterranean climatic regions.

In different classifications of Triticeae Heteranthelium is supposed to be related to either Triticum/Aegilops complex (Nevski 1934a; Tzvelev 1976) or Hordeum (Love 1984; Clayton and Renvoize 1986; Kellogg 1989). Clayton and Renvoize (1986) regarded it as an advanced offshoot of Critopsis. Yet, Heteranthelium has one spikelet per node like the sub-tribe Triticineae and awnlike glumes like Hordeum (Frederiksen 1993). Thus, the phylogenetic relationships of Heretanthelium are still ambiguous. Studies of Hodge et al. (2010) on the chloroplast gene encoding ribosomal protein S16, showed that H. piliferum is in the same clade with Triticum monococcum, Secale cereale, and all the Aegilops species. Their results are consistent with the finding of Mason-Gamer et al. (2002), based on combined cpDNA sequences of tRNA genes, spacer sequences, rpoA genes, and restriction sites. Phylogenetic relationships based on mating systems showed that H. piliferum is in a clade with Dasypyrum Villosum (Escobar et al. 2010). This is in spite of the fact that H. piliferum is facultative self-pollinated whereas D. Villosum is an out-crosser. Escobar et al. (2011) studied one chloroplastic and 26 nuclear genes and found that H. piliferum comprises a clade with Dasypyrum villosum (clade IV) that branches at the base of clade V (Triticum, Aegilops, Secale and Taeniatherum). Cllades IV and V are closely related to each other (Escobar et al. 2011).

5.8 Taeniatherum Nevski

5.8.1 Morphological and Geographical Notes

Because of its diverse morphological features, Taeniatherum Nevski has previously been included in several different genera. Linnaeus (1753) considered it as a single species belonging to the genus Elymus (E. caput-medusae L.), while Schreber (1772) concluded that this taxon consists of two Elymus species (E. caput-medusae and E. crinitum Schreb.), and Link (1827) defined three species (E. cput-medusae, E. crinitum and E. platatherus Link). Other authors thought that Taeniatherum belongs either to the genus Cuviera Koeler (Simonka 1897) [now Hordelymus (Jessen) Harz] or Hordeum L. (Cosson and Durieu 1855; Ascherson and Graebner 1902). However, Nevski (1934a) pointed out that Taeniatherum differs from the above genera in several principal features, e.g., it differs from Elymus in its one-flowered spikelets with connate, subulate glumes, and annual life cycle. Taeniatherum differs from Hordelymus in its sessile spikelets with connate glumes, flattened lemma awns and annual life cycle, and differs from Hordeum in its tough rachis, rigid spike with a terminal spikelet and sessile spikelets in pairs (Frederiksen 1986). Consequently, Nevski (1934a) considered it a separate genus, Taeniatherum (Fig. 5.5a).

Fig. 5.5
figure 5

Species of Taeniatherum-Crithopsis clade; a A spike and a plant of T. caput-medusae (L.); b A plant and spike of C. delileana (Schult.) Rozhev

The genus Taeniatherum is characterized by a broad morphological variation, making the classification of its specific rank difficult. Consequently, the taxonomic literature contains different classifications of the specific rank in this genus. In his monograph on Taeniatherum, Nevski (1934a) defined three species, T. caput-medusae, T. crinitum, and T. asperum, while Bor (1968) recognized only two: T. crinitum and T. asperum, and noted that they only differ very little from one another in several morphological traits. The distinguishing characters between the taxa that Nevski (1934a) and Bor (1968) defined, were glume length and spreading of glumes after ripening, length of the lemma, and width of the lemma awn base. The culm length to spike length ratio was sometimes used to distinguish between the taxa Nevski 1934a, 1936; Maire and Weiller 1955; Humphris 1978). However, all these traits show gradual transition between the taxa and merely indicate quantitative, rather than qualitative differences, rendering them impractical to use for identification purposes. In his taxonomical treatment of the genus, Humphris (1978) combined the above-mentioned species of Nevski (1934a) and Bor (1968) to one, T. caput-medusae, which contains several morphological variants. He argued that the morphological characteristics that distinguished between those defined by Nevski and Bor were unstable and variable, and did not justify a split of the species.

While revising the taxonomy of Taeniatherum, Frederiksen (1986) thought that the morphological characters that led Nevski to split the species into three, were indistinct, with many intermediates occurring between them, rendering it rather difficult to define these taxa morphologically. Since, morphological features, geographical extension and crossing experiments (Sakamoto 1969; Frederiksen and von Bothmer 1986) indicate some taxonomic differentiation, Frederiksen (1986) treated the three Nevski species as subspecies of T. caput-medusae (L.) Nevski, namely, subsp. caput-medusae, subsp. crinitum (Schreb.) Melderis, and subsp. asperum (Simk.) Melderis. She stated that the absence of discontinuities in the morphology supports a sub-specific rank.

Taeniatherum caput-medusae (L.) Nevski (common name medusa-head) is a small genus of annual, short plants (20–61 cm tall) with dense spikes, bearing one terminal spikelet and otherwise paired sessile spikelets, and a tough rachis, that does not disarticulate at maturity. It has two-flowered spikelets, with a lower hermaphrodite floret, and an upper floret reduced to a scale-like rudiment. At maturity, glumes are divergent from the rachis, forming an acute or obtuse angle. The glumes and lemma are awned. Glumes 3–5 cm long (including awns). Lemma’s awns are flattened (0.4–1.5 mm wide) at the base, 6.0–13.5 cm long (Incl. lemma), and twisted when dry. The palea is a little longer than the body of the lemma. Anthers are small (about 1 mm in length), and the mature caryopsis is firmly adhered to the lemma and palea, which are released from the glumes, that remain on the rachis. The very small anthers that shed most of the pollen inside the floret indicate a facultative autogamous species (Frederiksen and von Bothmer 1986) (Fig. 5.5a).

The species distributes from West Mediterranean to Central Asia, i.e., Portugal, Spain, Southern France, Morocco, Algeria, Italy (incl. Sardinia and Sicily), Croatia, Serbia, Bosnia-Herzegovina, Macedonia, Albany, Turkey, Greece (incl. Crete), Romania, Bulgaria, Southern Hungary, Southwestern Russia, Crimea Caucasia, Cyprus, Tunisia, Libya, Egypt (incl. Sinai), Israel, Jordan, Syria, Lebanon, Iraq, Iran, West Pakistan, Afghanistan, Turkmenistan, Uzbekistan, Tajikistan, and Kirgizstan. Outside this area, it has been introduced as a weed in the northern and northwestern parts of Europe, in North and South America and Australia. T. caput-medusae inhabits low mountains and plateau areas, growing on altitudes from 600 to 800 m in Yatir, Israel, to 1850 m in south Sinai, Egypt, in a range of annual precipitation from 100 mm in Sinai to 1000–1300 mm in Mt. Hermon. All sites it grows in are arid, somewhat extreme habitats, in sub-Mediterranean small-shrub formations and Irano-Turanian steppes, and is usually found on Nubian sandstone, basalt, and calcareous soil, as well as on stony or gravelly soils.

T. caput-medusae ssp. caput-medusae [(Syn.: Elymus caput-medusae L.; Hordeum caput-medusae (L.) Coss. & Dur.; Cuviera caput-medusae (L.) Simk.; Hordeum caput-medusae subsp. bobartii Asch. & Graebn.; Elymus caput-medusae var. typicus Halácsy; Elymus caput-medusae subsp. bobartii (Asch. & Graebn.) Maire; Hordelymus caput-medusae (L.) Pign.; Taeniatherum caput-medusae var. caput-medusae Humphries] bear glumes that are 3.5–8.0 mm long and horizontal. Paleae are 5.0–8.5 mm long. The awn base is 0.5–0.8 mm wide. This subspecies is restricted to the western part of the species distribution area, i.e., Portugal, Spain and southernmost France in Europe, and Morocco and Algeria in Africa.

T. caput-medusae ssp. crinitum (Schreb.) Melderis (1984) [syn.: Elymus crinitus Schreb.; Hordeum crinitum (Schreb.) Desf.; Elymus caput-medusae var. crinitus (schreb.) Ball.; Elymus caput-medusae subsp. crinitus (Schreb.) Nyman; Hordeum caput-medusae subsp. crinitum (Schreb.) Asch. & Graebn.; Taeniatherum crinitum (Schreb.) Nevski; Hordelymus caput-medusae subsp. crinitus (Schreb.) Pign; Taeniatherum caput-medusae var. crinitum (Schreb.) Humphries] bear glumes that are 1.5–4.0 mm long, erect or curved. Glume awns are 0.4–1.0 mm wide at base. Lemmas are 8.5–14.5 mm long. Lemma awns are 0.6–1.0 mm wide at base. This subspecies is found from Italy and eastwards into Asia. Within a large part of the distribution area, morphological transitions are found between this subspecies and subsp. asperum.

T. caput-medusae ssp. asperum (Simk.) Melderis [syn.: Cuviera caput-medusae var. aspera Simk; Hordeum caput-medusae subsp. asperum (Simk.) Degen in Asch. & Graebn.; Cuviera aspera (Simk.) Simk.; Eltmus caput-medusae var. asper (Simk.) Halácsy; Elymus asper (Simk.) Brand in Hallier & Brand; Hordeum caput-medusae var. asperum (Simk.) Fom. & Wor. Ex Fedtsch.; Taeniatherum asperum (Simk.) Nevski; Elymus caput-medusae subsp. critinus var. asper (Simk.) Maire; Hordelymus asper (Simk.) Beldie in Savulescu; Hordelymus caput-medusae subsp. asper (Simk.) Pign.] bears glumes that are 1.5–4.0 mm long, erect or curved. Lemmas are 5.0–14.5 mm long. The awn base is 0.4–1.0 mm wide. This subspecies has a very broad distribution, as it is found in most parts of the distribution area of the species. Morphological intermediary specimens between this subspecies and the others are found.

5.8.2 Cytology, Cytogenetics and Evolution

Chromosome number was determined by Sakamoto and Muramatsu (1965), Sakamoto 1969), Bowden (1966), Coucoli and Symeonides (1980), Luria (1983), Frederiksen (1986) and Linde-Laursen and Frederiksen (1989), who found that all taxa are diploids (2n = 2x = 14). The genome symbol of Taeniatherum caput-medusae is Ta (Wang et al. 1995). The three subspecies of T. caput-medusae have the same karyotype, with no interspecific variation (Frederiksen 1986). The karyotype is symmetric, with six metacentric pairs (one of which is a SAT-chromosome with a small satellite) and one sub-metacentric. All seven pairs of chromosomes were of nearly equal size, so that it was difficult to distinguish between them with certainty.

In an attempt to characterize the karyotype of each subspecies, Linde-Laursen and Frederiksen (1989) studied the C-banding patterns of somatic metaphases in plants from several populations of all three taxa. The C-banding patterns of all three subspecies were rather similar and characterized by a majority of small or very small bands. The number of bands per chromosome varied from 2 to 12 and had no preferential disposition. The banding pattern polymorphism was narrow but was sufficient for identifying the homologues of each of the seven chromosome pairs. No larger polymorphism was found within the sub-species, indicating that this trait is of no diagnostic value in distinguishing between the subspecies, and supported the view of a close relationship between them (Frederiksen 1986).

Frederiksen and von Bothmer (1986) hybridized the three subspecies of Taeniatherum and studied chromosomal pairing at meiotic first metaphase and fertility in the F1 intra-specific hybrids. Crosses within a subspecies were as difficult to perform as crosses between subspecies. The intra-specific hybrid plants were vigorous but completely sterile (Frederiksen and von Bothmer 1986). Meiotic pairing in the parental plants was high, predominantly in the form of ring bivalents and 13.10–14.27 chismata/cell. Chromosomal pairing in the F1 intra-specific hybrids also showed a very high degree of chromosomal pairing (average of 11.0 chiasmata /cell; no multivalents were observed), indicating that all three sub-species have the same basic genome. The almost complete lack of multivalents in the hybrids also shows that the basic genome has not been subjected to any larger structural rearrangements. However, this finding is in contradiction with the observations reported by Sakamoto (1969), who observed in several pollen mother cells of a sterile F1 hybrid of ssp. asperum x ssp. crinitum, a trivalent and in several others, a quadrivalent, in addition to PMCs with seven bivalents or six bivalents and two univalents (average chromosome pairing per cell was 1.9 univalents, 5.8 bivalents, 0.1 trivalents, and 0.2 quadrivalent). From these observations, Sakamoto (1969) concluded that the two taxa had very similar but structurally differentiated genomes. However, it is most likely that one of the accessions used by Sakamoto (1969) contained a small reciprocal translocation that did not exist in the accessions used by Frederiksen and von Bothmer (1986). Consequently, Frederiksen and von Bothmer (1986) concluded from the pairing data, that the three subspecies share a high degree of genome homology, and thus, supported Frederiksen’s (1986) treatment of these taxa as subspecies. This conclusion fails to account for the complete sterility of the intra-specific F1 hybrids, indicating that certain genetic barriers exist between the subspecies. Frederiksen and von Bothmer (1986) assumed that these genetic barriers might have resulted from the processes of adaptation to the different conditions of the large variety of ecological habitats occupied by this species, that together with the effect on adaptive gene complexes, also selected genes involved in the intra-specific barriers.

The crossability of T. caput-medusae with other Triticeae species has been low, making it difficult to produce intergeneric hybrids, and consequently, the cytogenetic relationships between this species and other Triticeae species are poorly known. For instance, Sakamoto (1973) tried to cross T. caput-medusae with Crithopsis delileana and did not succeed to obtain F1 seeds. However, several intergeneric crosses between Taeniatherum caput-medusae and other Triticeae species were successfully performed. Schooler (1966) crossed T. caput-medusae with Aegilops cylindrica and obtained a highly sterile triploid F1 hybrid. At meiosis, univalents were the most common configuration, while rod bivalents and especially ring bivalents were only rarely observed (Schooler 1966). Thus, he concluded that the three genomes, Ta of T. caput-medusae and CD of Ae. cylindrica, are only distantly related.

Sakamoto (1991) analyzed intergeneric hybrids between T. caput-medusae ssp. crinitum and Eremopyrum orientale (4x) and Agropyron tsukushiense. Chromosome pairing in several plants of the F1 triploid hybrid E. orientale (4x) x T. caput-meusae ssp. crinitum (2x) was low (average of 19.3–20.6 univalents and 0.2–0.8 bivalents per cell) and the hybrid was completely sterile. Chromosome pairing in the F1 tetraploid hybrid between Agropyron tsukushiense (6x) x T. caput-medusae ssp. crinitum (2x) was also low (average of 26.1 univalents and 0.9 bivalents per cell), and no seed set was obtained. No genomic homology was found between these three genera (Sakamoto 1991).

Frederiksen and von Bothmer (1989) crossed the 3 subspecies of Taeniatherum caput-medusae with 30 different species representing 11 genera of the Triticeae. A seed set were observed in 15 combinations only. However, most of the seeds lacked an embryo or the embryo was unable to form a vigorous seedling, and consequently, only six resulted in adult plants. In the combination of Taeniatherum caput-medusae with Hordeum bulbosum (2x), a haploid of T. caput-medusae was obtained as a result of selective elimination of the H. bulbosum chromosomes. All other hybrids included the following five combinations: T. caput-medusae ssp. caput-medusae x Psathyrostachys fragilis, T. caput-medusae ssp. crinitum x Psathyrostachys juncea, ssp. crinitum x Dasypyrum villosum, ssp. asperum x Hordeum brevisubualatum, and ssp. crinitum x Eremopyron. orientale (4x). All F1 hybrids were morphologically intermediate between the parents. Perenniality dominated over annuality in the combinations with the Hordeinae species, and annuality in hybrids with Dasypyrum and Eremopyrum. All hybrids were highly sterile due to very low chromosome pairing at meiosis. The pairing data supported the doctrine that Taeniatherum is a distinct genus within the Triticeae, but slightly related to the genomes of Psathyroatachys, Dasypyrum, Eremopyrum or Hordeum.

Löve (1984) assumed that the tetraploid genus Hordelymus possesses one genome from Hordeum and another from Taeniatherum. As no hybrids between these two genera survived (Frederiksen and von Bothmer 1989), homology between the genomes of these two genera could not be proven nor rejected. Yet, von Bothmer et al. (1994) ruled-out the presence of the H genome in Hordelymus and demonstrated the presence of the Ns genome instead. Likewise, Bieniek et al. (2015) showed that Psathyrostachys juncea (genome NsNs) contributed its genome as a maternal parent to the allotetraploid Hordelymus eyuropaeus and not Taeniatherum caput-medusae, and suggested a genomic formula of NsNsXrXr for Hordelymus.

Frederiksen (1994) crossed Taeniatherum caput-medusae with Triticunm aestivum and obtained an F1 hybrid. Morphologically, the hybrid looked like T. aestivum, with broad leaves and one spikelet per node, with three florets. The hybrid was completely sterile. Analysis of meiotic pairing at first meiotic metaphase showed a high number of univalents (20.02), an unexpected high number of bivalents (3.52) (rod 3.44 and ring 0.08), and trivalents (0.08), resulting in 1–10 chiasmata/cell. The relatively high pairing was ascribed to pairing between the homoeologous chromosomes of the A, B, and D subgenomes of wheat. It is concluded that the genome of Taeniatherum eases the restriction on homoeologous pairing in T. aestivum, imposed by the Ph1 gene.

As Taeniatherum seems to be a distant relative of Triticum, it is reasonable to suppose that the observed pairing in the hybrid only occurred between wheat chromosomes. The mean number of chiasmata in euhaploid T. aestivum containing Ph1 was reported to be 1.06, while it was 8.48 in euhaploid mutants lacking Ph1 (Jauhar et al. 1991). The number of chiasmata per cell in the hybrid studied by Frederiksen (1994) was 4.26 and thus, fell between the range referred to above. Thus, the genotype of Taeniatherum seems to have some repressive effect on the Ph1 gene, resulting in increased pairing of homoeologous wheat chromosomes.

5.8.3 Phylogeny

In most morphological trees, Taeniatherum is linked to the Hordeum group, mainly because it shares the characteristic of multiple spikelets per node (Baum 1983; Baum et al. 1987; Kellogg 1989; Frederiksen and Seberg 1992). In contrast, phylogenetic studies based on molecular analyses have placed Taeniatherum close to Secale, Triticum and Aegilops (e.g., Mason-Gamer et al. 2002, based on chloroplast DNA, and Hsiao et al. 1995, Seberg and Petersen 2007, and Escobar et al. 2011, based on nuclear DNA sequences).

Hsiao et al. (1995) used the sequences of the internal transcribed spacer (ITS) region of nuclear ribosomal DNA and sequences of tRNA to estimate phylogenetic relationships among 30 diploid Triticeae species representing 19 genomes. They found that most of the annuals of Mediterranean origin, i.e., species of Triticum, Aegilops, Crithopsis, Taeniatherum, Eremopyrum, Henrardia, Secale, and two perennials, Elymus farctus and Elymus elongatus, constitute a monophyletic group. In a more restricted species sampling, the two perennial parsimony tree species Elymus farctus and E. elongatus, formed a sister group with Triticum monococcum, Aegilops speltoides, and Ae. tauschii. Crithopsis, Taeniatherum, Eremopyrum, and Henrardia were close to Secale. Based on this finding, Hsiao et al. (1995) suggested that the Triticeae apparently independently evolved similar sorts of characters in two parallel lineages, one in the Mediterranean and one in the Arctic-temperate region. Several conspicuous morphological characters appear in both groups, for example, a similar number of spikelets per node, a similar number of florets per spikelet, linear spikes, and keeled glumes. Species in the Mediterranean group that formed a monophyletic lineage in the ITS tree, were grouped with various perennials of Arctic-temperate origin on the morphology trees (Kellogg 1989; Frederiksen and Seberg 1992; and following Seberg Frederiksen, unpublished data). This apparent agreement between the molecular and bio-geographic data, and conflict of both of these lines of evidence with the morphological trees, suggest that large-scale morphological parallelism occurred in the evolutionary history of the Triticeae. The results of Hsiao et al. (1995) supported the suggestion of Sakamoto (1973) to classify Triticeae as two major groups, a Mediterranean group and an Arctic-temperate group (Table 2.3), with the Mediterranean lineage evolving from the Arctic-temperate species.

Mason-Gamer et al. (2002) analyzed new and previously published chloroplast (cp) DNA data from Elymus and from most of the mono-genomic genera of the Triticeae, and presented additional cp DNA data from Elymus and from mono-genomic genera and constructed the phylogeny for the mono-genomic genera. Their analysis was in agreement with previous cpDNA studies with regard to the close relationship between Secale, Taeniatherum, and TriticumAegilops. Further, their analysis provided moderate support for some additional still unresolved relationships or for those that were very weakly supported in the previous cpDNA studies. These included (i) the sister relationship between Taeniatherum and Triticum–Aegilops; (ii) the placement of Heteranthelium with a Secale + Taeniatherum + TriticumAegilops clade; and (iii) the placement of the Elymus farctus and E. elongatus + Dasypyrum clade with the Secale + Taeniatherum + AegilopsTriticum + Heteranthelium clade. A close relationship between Taeniatherum and TriticumAegilops is completely at odds with the DMC1 tree (Petersen and Seberg 2000) and the morphology-based cladogram (Seberg and Frederiksen 2001). The ITS data left the relationship unresolved (Hsiao et al. 1995), and Taeniatherum was not included in the 5S long spacer data set (Kellogg and Appels 1995). But, Escobar et al. (2011) found that the clade containing Taeniatherum and TriticumAegilops is also seen on the 5S short spacer data tree, but only if Elymus farctus and E. elongatus are included in the clade. Some conflict appears in the position of Secale in the starch synthase tree, in which a close relationship among Triticum, Aegilops, and Heteranthelium was detected, while Secale was placed in a SecaleTaeniatherumElymusDasypyrum group (Mason-Gamer and Kellogg 2000).

The above phylogenetic studies suggest that Taeniatherum is closer to the species of the sub-tribe Triticineae than to those of the Hordeineae. Hence, use of the number of spikelets on each node, as a diagnostic marker for placing genera in one of the two sub-tribes (Tzvelev 1976; Clayton and Renvoize 1986), is not sufficient for such a grouping.

5.9 Crithopsis Jaub. & Spach.

5.9.1 Morphological and Geographical Notes

The genus Crithopsis was previously included in one of the Elymus, Hordeum and Eremopyrum genera, however, Jaubert and Spach (1851) recognized it as a separate genus. Crithopsis is monotypic and is represented by the species C. delileana (Schult.) Rozhev.

Crithopsis delileana [syn.: Elymus delileana Schult.; Elymus geniculatus Del.; Elymus aegyptiacus Spreng.; Elymus rhachitrichus Hochst. ex Jaub. & Spach; Elymus subulatus Forssk.; Hordeum delileanum (Schult.) Hack.; Hordeum geniculatum (Delile) Thell.; Eremopyrum cretense (Coustur. & Gand.) Nevski; Crithopsis rhachitricha Jaub. & Spach; Crithopsis brachytricha Walp] is annual, culms 10–30 cm high, spike dense and bristly, 2–5 cm long (excl. awns), with paired sessile spikelets at each node of rachis. The rachis is densely hairy and fragile and disarticulates above each node (wedge-type disarticulation). Spikelets contain two florets, where the lower is hermaphrodite and the upper rudimentary. Glumes are equal in size and linear, tapering to an awn. Lemma taper to form a long flat awn, equal in length to the glume awn. Anthers are about 0.8 mm long. The caryopsis is firmly adherent to lemma and palea. The dispersal unit is the rachis internode, which carries two spikelets with a single grain in each (Fig. 5.5b).

The species is characterized by a wide morphological variation, mainly in the color of the leaf sheath, hairiness of the leaves and glumes and lemmas, size of rachis internodes, glumes and lemmas. Luria (1983) found no difference in the seed sets of bagged and unbagged spikes, indicating that the pollination system is autogamous. This finding was corroborated by Frederiksen (1993) who, on account of the short anthers, assumed this species to be predominantly autogamous.

The distribution of the species is from North Africa, through the southeastern Mediterranean basin to Central Asia [Morocco, Algeria, Tunisia, Egypt, Libya, Greece (incl. Grete), Turkey, Cyprus, Syria, Lebanon, Jordan, Israel, Iraq, Iran, Afghanistan, and west Pakistan (Baluchistan)]. The species grows in a wide range of habitats, on a variety of climatic conditions and on different soils [terra rossa, rendzina (dark and light), basalt, grumusol, sandy soil, loess, and desert lithosol]. It is found at 100–950 m above sea level, from xeric climates with 100 mm annual rainfall to more mesic conditions, with 800 mm annual rainfall. Crithopsis delileana grows in dry steppe grassland and batha (Mediterranean small shrub formations), as well as in semi-disturbed habitats, in plant communities consisting of Irano-Turanian and sub-Mediterranean plants.

5.9.2 Cytology, Cytogenetics and Evolution

All analyzed accessions of Crithopsis delileana are diploids, 2n = 2x = 14 (Sakamoto and Muramatsu 1965; Bowden 1966; Sakamoto 1991, 1973; Luria 1983; Frederiksen 1993; Linde-Laursen et al. 1999). Löve (1984) designated its haploid genome as K. The karyotype of all analyzed accessions was similar, i.e., symmetric, consisting of five pairs of metacentric chromosomes, one pair of sub-metacentric chromosomes, and one pair of metacentric SAT-chromosomes, with rather small satellites (Frederiksen 1993; Linde-Laursen et al. 1999). Only the SAT-chromosome pair and the sub-metacentric pair could be identified reliably by morphology, whereas the morphological differences between the other five pairs of metacentric chromosomes were insufficient for safe identification. Linde-Laursen et al. (1999) made use of Giemsa C-banding, Giemsa N-banding and silver nitrate staining to discriminate between the chromosome pairs of C. delileana. The Giemsa C-banding patterns included a few small to very small, mainly centromeric or telomeric bands (Linde-Laursen et al. 1999). The banding patterns of the different populations were polymorphic, but within populations, the variation in banding patterns was sufficient for identifying the homologous chromosomes of each of the seven pairs. N banding produced no or few weakly developed bands in the chromosomes at the same positions as C bands. Silver nitrate staining identified two nucleolar organizer regions (NORs), at the nucleolar constriction of the pair SAT-chromosomes. The number of NORs was confirmed by observing a maximum of two nucleoli at interphase (Linde-Laursen et al. 1999).

Intergeneric hybridization between Crithopsis and other members of the Triticeae are very difficult. Attempts to produce intergeneric hybrids between C. delieana and Aegilops peregrina (=Ae. variabilis), Eremopyrum bonaepartis, E. triticeum Taeniatherum caput-medusae ssp. caput-medusae and Agropyron tsukushiensis have always failed (Sakamoto (1973). Likewise, Frederiksen and von Bothmer (1989) had no success in crossing C. delileana with Taeniatherum caput-medusae. However, a single weak hybrid was produced in the cross of C. delileana and Eremopyrum distans (Frederiksen 1993).

Crithopsis delileana and Taeniatherum caput-medusae are considered close taxonomically to each other (Clayton and Renvoize 1986) and morphologically similar. Both feature spikes with two sessile spikelets at each rachis node and a single hermaphroditic floret at each spikelet. In contrast to Crithopsis, which has a fragile rachis, Taeniatherum bears a tough rachis and a disarticulating rachilla. The karyotype of C. delileana is morphologically similar to that of T. caput-medusae, as shown by Linde-Laursen and Frederiksen (1989). However, the C-banding patterns of the two species exhibit differences in the distributions of the C-bands, with C. delileana having more telomeric and fewer intercalary bands than T. caput-medusae. Yet, altered distribution of C-bands is a weak diagnostic characteristic, as activity of transposable elements, that can affect the distribution and quantity of the C-banding, may be different in closely related species and even within a species.

5.9.3 Phylogeny

The genera Crithopsis and Taeniatherum are traditionally considered related to Hordeum and Psathyrostchys and therefore, are included in the Hordeineae (Tzvelev 1976; Clayton and Renvoize 1986). All four genera have more than one spikelet per rachis node and only one hermaphroditic floret per spikelet. Morphologically-based phylogenetic analyses supported inclusion of the four genera within the same, albeit not fully resolved, clade (Kellogg 1989; Frederiksen and Seberg 1992). Following her cladistics analysis, Kellogg (1989) even proposed to include Crithopsis and Taeniatherum in Hordeum in a single clade. Similar analysis performed by Frederiksen and Seberg (1992) placed Crithopsis, Taeniatherum, Hordeum and Psathyrostachys in a single clade.

Yet, the cytological studies of Linde-Laursen et al. (1999) indicate a closer relationship between the genera Crithopsis and Taeniatherum than between Hordeum or Psathyrostachys. In contrast to the comparatively minor differences in chromosome morphology distinguishing between the karyotypes of C. delileana and T. caput-medusae, the karyotypes of members of the genera Pasthyrostachys and Hordeum show several distinct characteristics (Linde-Laursen et al. 1999). Species of Psathyrostachys have significantly larger chromosomes, including SAT-chromosomes with generally minute satellites, as compared to the species of the other three genera, while species of Hordeum have chromosomes that produce bands after N-banding (Linde-Laursen 1981; Morris and Gill 1987; Xu and Kasha 1992). This banding pattern, unobserved in the other genera, indicates a significant qualitative difference in the composition of the constitutive heterochromatin (Gill 1987). The above differences support the taxonomic approach relating to them as individual genera.

There is also a discrepancy between the phylogenetic trees derived from morphological analysis versus those generated following molecular analysis of either chloroplast, mitochondrial or nuclear DNA sequences (Seberg and Frederiksen 2001; Seberg and Petersen 2007; Mason-Gamer 2005; Petersen et al. 2006; Escobar et al. 2011). In the morphological trees, Kellogg (1989), Frederiksen and Seberg (1992), Seberg and Frederiksen (2001), and Seberg and Petersen (2007) included Crithopsis and Taeniatherum in the Hordeum group, mainly because they share the characteristic of multiple spikekets per rachis node. Seberg and Petersen (2007) showed an incongruence between morphological evaluations and nucleotide sequence analysis of two plastid genes (rbcL, rpoA), one mitochondrial gene (coxII), and two single-copy nuclear genes (DMC1, EF-G). In addition, they found that data derived from chloroplast genes were somewhat incongruent with those from the mitochondrial and nuclear sequences. Similar limited agreement was found between gene trees derived from nuclear genes (the two 5S rDNA arrays and ITS) and those derived from the chloroplast (cpDNA RFLPs and rpoA; Kellogg et al. 1996; Petersen and Seberg 1997), both of which deviated considerably from the morphologically based phylogenies.

Based on the chloroplast data, Seberg and Petersen (2007) placed Crithopsis close to Secale, both of which were close to Aegilops/Triticum and Taeniatherum. Similar results showing that Crithopsis is close to Secale were obtained by Petersen and Seberg (1997), who analyzed the plastid genome spanning the entire rpoA gene, and by Petersen et al. (2006), who analyzed the plastid gene ndhF. They found that Crithopsis and Secale were sister species to a clade that includes Aegilops, Triticum, and Taeniatherum. Yet, based on the mitochondrial and the nuclear data, Seberg and Petersen (2007) found that Crithopsis was linked to the Aegilops/Triticum group. This finding is in agreement with that of Mason-Gamer and Kellogg (1996a), whose nuclear data supported classification of the clade of Ae. tauschii, Triticum monococcum, Taeniatherum caput-medusae, Elymus farctus ssp. bessarabicus, Crithopsis delileana and Elymus elongatus. Similar results were also obtained by Petersen et al. (2006), who studied the nuclear gene DMC1, and found that Crithopsis was close to Aegilops/Triticum and Elymus elongatus. Likewise, Hsiao et al. (1995) analysed the internal transcribed spacer (ITS) region of nuclear ribosomal DNA and found that Crithopsis is close to Secale and included both in the Aegilops/Triticum clade.

Mason-Gamer (2005) constructed a phylogenetic tree based on analysis of the β-amylase genes of the Triticeae. This tree consists on an evidence-supported clade included Aegilops, Triticum, Crithopsis, and Taeniatherum. Placement of Crithopsis (and Taeniatherum) within or very near Aegilops was supported by most of the molecular data sets that included C. delileana (see Table 5 of Mason-Gamer 2005), including the analyses of cpDNA restriction sites (Mason-Gamer et al., 2002), the rpoA gene (Petersen and Seberg 1997), the integrated highly repetitive genes (Kellogg et al. 1996), and the DMC1 gene (Petersen and Seberg 2002). In full agreement with the above, Seberg and Petersen (2007) presented a highly resolved, strongly supported, consensus phylogenetic tree, based on all their data, which included morphological assessments, and analyses of two chloroplast genes, one mitochondrial gene, and two nuclear genes. As a rule, the tree positions Crithopsis and Taeniatherum in the same clade, which includes several other taxa as well, mostly Aegilops, but invariably excludes Psathyrostachys and Hordeum.

Close homology between DNA sequences of Crithopsis delileana and Taeniatherum caput-mesdusae with Elymus elongatus was found by Arterburn et al. (2011). Diploid E. elongatus and C. delileana were most closely related. E. elongatus was equally close to E. farctus ssp. bessarabicus and T. caput-medusae, but ssp. bessarabicus was shared less similarity with T. caput-medusae and C. delileana than with E. elongatus.

The traditional taxonomic subdivision of the genera of the Triticeae into two sub-tribes, the Hordeineae and the Triticineae (e.g., Tzvelev 1976; Clayton and Renvoize 1986), is not supported by the phylogenetic schemes derived from molecular analyses. Sakamoto (1973, 1991) classified the Triticeae into two groups based on their geographical distribution: the Arctic-Temperate group and the Mediterranean-Central Asiatic group (Table 2.3). In their phylogenetic analysis of ITSs of nuclear rDNA, Hsiao et al. (1995) recovered the Mediterranean group. The Arctic-Temperate group includes most of the Hordeineae (characterized by three spikelets per rachis node), while the Mediterranean-Central Asiatic group includes all of the Triticineae (characterized by solitary spikelets at each rachis node), as well as species of Hordeum, Crithopsis and Taeniatherum (characterized by two to three spikelets per rachis node). The latter group is younger and assumed to have developed from the Arctic-Temperate group (Runemark and Heneen 1968; Sakamoto 1973). DNA sequence-based phylogenetic analyses, indicate that the two annual genera, Crithopsis and Taeniatherum, that are classified in the Hordeineae on the basis of morphological features, are linked to the genera of the sub-tribe Triticineae. Their inclusion in the Mediterranean-Central Asiatic group is not merely due to their geographical distribution, but also to close phylogenetic relationships between the two genera, Crithopsis and Taeniatherum. Their relationships with genera of the Triticinae may indicate their evolvement from the same ancestral group and suggests that evolutionary development of diagnostic morphological characteristics, e.g., number of spikelets on each rachis node and number of fertile florets in each spikelet, occurs at varying rates. The dispersal unit of many species belonging to the Mediterranean-Central Asiatic contains two seeds, either one in each of the two spikelets or two in a single spikelet, indicating different routes for achieving analogous adaptive traits to brace the long, dry summer of the Mediterranean and Central Asiatic regions.

5.10 Concluding Comments

5.10.1 Evolution of the Sub-tribe Triticineae

The cold climate that prevailed during the Oligocene geological era (33.7–23.8 MYA; Table 2.5) channeled the evolution of the early Triticeae toward the development of perennial and allogamous genera adapted to mesophyllic habitats that are characteristic to the temperate-arctic zones. During the Miocene era (23.8–5.3 MYA; Table 2.5), the climate warmed up and the Tethys Sea, that covered a large area of the Mediterranean basin and southwest Asia, disappeared and the lands in these regions rose. The climate in the east Mediterranean and central Asia became seasonal, namely, with cold and humid winters and hot and dry summers. These geological and climatic changes, characterized by a relatively short growth period in the winter and long inactivity in the summer, have led to the development of diversified ecosystems, expansion of grasslands and opening of new ecological niches (Table 2.5). Such ecological changes were suitable for the development of the annual, autogamous Triticineae genera. It is therefore assumed that the radiation of species in the Triticeae tribe might have been triggered by the middle to late Miocene climate (Fan et al. 2013).

Indeed, several molecular studies indicated that the radiation of genera and species in the sub-tribe Triticineae, the Mediterranean and central Asiatic lineage, occurred during the later part of the Miocene era. Sequencing of chloroplast genes indicated that the divergence of the sub-tribe Triticineae from the subtribe Hordeineae took place either 15 (Marcussen et al. 2014), 10.6 (Gornicki et al. 2014), or 8–9 (Middleton et al. 2014) MYA, whereas sequencing of nuclear genes showed that this divergence occurred 11.6 (Chalupska et al. 2008), 11.0 (Huang et al. 2002a, b) or 10.1 (Dvorak and Akhunov 2005) MYA (Table 2.6). The molecular data and biogeography of the sub-tribe suggest that the Mediterranean lineage derived from the Arctic-temperate lineage and that the two lineages have since evolved in parallel (Hsiao et al. 1995).

5.10.2 Appearance of Advanced Traits in Different Genera

Ancestral trait is a trait that exists in a group of taxa that are all descendent from a common ancestor in which the trait first developed. Primitive trait represents the original condition of the trait in the common ancestor, whereas an advanced trait or a derived trait signifies an important change from the original condition.

Major evolutionary changes that facilitated rapid adaptation of the sub-tribe Triticineae to the newly opened Mediterranean and central Asiatic habitats, included the transition from perennial to annual growth habit, from allogamy to autogamy, from tall to short plants, from a few to many tillers, from long to short spikes, from several to a single spikelet on each rachis nod, and from un-awned to awned spikelets. In the Mediterranean and Central Asiatic climate, annual growth habit is an adaptive trait, since annual plants can pass the long, dry and hot summer as seeds. The evolution of species with an annual growth habit occurred independently several times in the Triticineae, either between genera (Agropyron and Eremopyrum), or within genera (in Dasypyrum and Secale) (Table 5.5). Perennial growth habit is a dominant trait controlled by a small number of genes (Charpentier et al. 1986; Lammer et al. 2004), where mutations in these genes lead to the development of annual plants. Allogamy is considered a primitive character since most perennial species in the tribe Triticeae are allogamous. Stebbins (1957) pointed out that in many plant groups, autogamy derived from allogamy and therefore, autogamy might be considered a more advanced trait. Indeed, autogamy developed independently several times from the cross-pollinating species in the Triticineae (Escobar et al. 2010; Table 5.5). While annual growth habit requires reestablishment of the populations each year, autogamy is well fit for mass reproduction of plants with similar well-adapted genotypes, that can occupy successfully the same habitats, and facilitates the rapid colonization of newly opened habitats (Mac Key 2005). Since autogamy in the Triticineae species is not an obligatory trait, occasional out-crossing may result in the production of sufficient genetic flexibility that effectively used by the rapid generation shift (Allard and Kannenberg 1968; Allard et al. 1968; Allard 1975). The annual plants pass the dry and hot summers as dormant seeds that are buried and stored in the soil until autumn rains begin and growth starts again. The change from dormancy to germination is a fundamental process superimposed by a series of genetic trigger mechanisms, which selectively respond to environmental factors like water, temperature, light, and oxygen (Mac Key 1987, 1989). Selection for ability to compete against other species in dense stands may be combined with lower competition within the species itself (Mac Key 2005). Short plants with a large number of tillers can better tolerate drought and hot winds. Awned spikelets are better protected from drought and herbivores and more efficient in seed dispersal and self-sowing. Not all these changes occurred in all the genera and presumably did not happen simultaneously (Table 5.5). The Triticineae genera in Table 5.5 are classified in five clades, each containing genus or genera with ancestral and advanced traits. Genera (or species) having more advanced traits than ancestral ones, e.g., in clade 1, Eremopyrum and Henrardia have more advanced traits than Agropyron, in clade 2, Heteranthelium has more advanced traits than Dasypyrum and Secale, in clade 3, Aegilops is more advanced than Amblyopyrum and Triticum, and in clade 4, Crithopsis is more advanced than Taeniatherum. The genus Aegilops contains species with ancestral traits and species with advanced traits displaying evolutionary changes within the genus. Interestingly, genera or species exhibiting more advanced traits, namely, species of Eremopyrum, Taeniatherum, Crithopsis, Heteranthelium and Henrardia, occupy more xeric habitats in the peripheries of the Triticineae distribution area.

Table 5.5 Possession of ancestral (bold) and advanced traits in Triticineae genera

5.10.3 Chromosomal Pairing Level in Intergeneric F1 Hybrids

Despite the fact that the various Triticineae genera are relatively young and still maintain a great deal of genetic relatedness, their chromosomes are homoeologous rather than homologous, due to genetic and structural changes that have occurred during their divergence. Consequently, F1 intergeneric hybrids display minimal meiotic pairing between the homoeologous chromosomes of the parental genera. For instance, pairing between Secale species and other Triticineae is affected by chromosomal rearrangements, such as translocations and inversions, that drove the evolution of Secale genomes. Also, the accumulation of large amounts of subtelomeric heterochromatin can change the relative position of the telomeric regions at the beginning of meiosis, hindering pairing initiation of Secale and other Triticineae chromosomes (Devos et al. 1995; Lukaszewski et al. 2012; Megyeri et al. 2013). Furthermore, genetic and epigenetic changes, such as mutations, inactivation or elimination of DNA sequences that are involved in homology recognition and initiation of pairing at the beginning of meiosis, may cause pairing failure in hybrids between various Triticineae genera (Feldman et al. 1997; Ozkan et al. 2001).

Sybenga (1966) postulated that zygomeres control meiotic chromosomal pairing. The zygomeres may be made of homologous-specific DNA sequences that are clustered in “pairing initiation sites”. Clustering of telomeres in the bouquet at the leptotene–early zygotene meiotic stages may facilitate the approaching of terminal regions of homologous chromosomes (Naranjo and Corredor 2008; Naranjo and Benaveste 2015). These terminal regions may contain homologous-specific DNA sequences that become active at these early meiotic stages and play a role in homology recognition and initiation of pairing. Pairing between chromosomes in interspecific and intergeneric hybrids only occurs when the chromosomes of the two parents share the same homologous-specific DNA sequences. It seems likely that the sequences initiating chromosome pairing between homologous chromosomes of one Triticineae genus may be different from those of another genus, thus hindering the alignment of homoeologous chromosomes at the beginning of meiosis and the initiation of chromosome pairing at early zygotene. Moreover, meiotic pairing in interspecific or intergeneric F1 hybrids may either be promoted or suppressed by various pairing genes and B chromosomes that exist in these taxa (Table 5.2). Consequently, the reduced meiotic chromosomal pairing in intergeneric Triticineae hybrids, brought about by homoeologous pairing suppressors, cannot be used as an indication of the degree of evolutionary differentiation between the genomes of these genera.

5.10.4 Phylogenetic Scheme of the Triticineae Genera

5.10.4.1 Elymus Species with St, Ee and Eb Genomes

Escobar et al. (2011) classified the diploid Triticeae in five major clades: (1) Psathyrostachys; (2) Hordeum; (3) Elymus; (4) Aegilops–Triticum–Secale–Taeniatherum; and (5) EremopyrumAgropyron. The Elymus clade is the largest clade, consisting, among others, of representatives of the St, Ee and Eb genomes. These genomes are three important basic genomes from which, presumably, the subtribe Triticineae evolved. The St genome of several diploid species of Elymus (Table 5.1) is moderately related to the Ee genome of diploid E. elongatus and to the Eb genome of diploid E. farctus (Wang 1989). Petersen and Seberg (1997), based on rpoA sequence data, and Wang and Lu (2014), based on a literature review of chromosomal pairing data, confirmed the very close relationship among the St, Ee, and Eb genomes. The close relationship between St, Eb and Ee genomes was also substantiated by analysis of the sequences of a gene encoding plastid acetyl-CoA carboxylase (Fan et al. 2007), and between St and Eb by the study using 5S rDNA (Shang et al. 2007). Bieniek et al. (2015) found that the nucleotide sequences at three chloroplast loci (matK, rbcL, trnH-psbA) are almost identical in the diploid Ee, Eb and St taxa, with only one substitution within the matK gene, differentiating genome Eb from the Ee and St genomes. Since the St genome exists in more primitive diploid species of Elymus, it is assumed that genomes Ee and Eb evolved from St. Other Triticineae genera can be classified in the following four, semi-independent clades (Table 5.5).

5.10.4.2 The Clade Agropyron–Eremopyrum–Henrardia

The genus Agropyron most probably evolved from Elymus species bearing the St genome that is moderately related to genome P of Agropyron (Wang 1989). This assumption is supported by the analysis of 5S DNA sequences that consistently placed Elymus species with an St genome and Agropyron (genome P) in one clade (Baum and Appels 1992). Additional phylogenetic studies, based on morphology (Seberg and Frederiksen 2001), chloroplast DNA (Mason-Gamer et al 2002; Hodge et al. 2010), chloroplast, mitochondrial, and nuclear DNA sequences (Seberg and Petersen 2007) and nuclear genes (Hsiao et al. 1995; Mason-Gamer et al. 2010), included Eremopyrum species in the same clade as species of Agropyron. Likewise, the close phylogenetic relationship between Agropyron and Ermopyrum is also evident from the data of Escobar et al. (2011), who placed these two genera in the same clade. It is assumed therefore, that Eremopyrum, whose species are self-pollinating annuals that exhibit many advanced traits (Table 5.5), may have evolved from diploids of the genus Agropyron.

Eeremopyrum species were grouped with Agropyron and Henrardia on the chloroplast DNA tree (Mason-Gamer et al. 2002). Similarly, using β-amylase gene sequences, Mason-Gamer (2005) found that Henrardia persica is close to Eremopyrum bonaepartis. Dual placement of Eremopyrum and Henrardia with Agropyron has also been supported by other data sets (reviewed in Mason-Gamer 2005). Hsiao et al. (1995) reached a similar conclusion upon analysis of nuclear DNA sequences. Upon analysis of the chloroplast gene encoding ribosomal protein rps16, Hodge et al. (2010) also placed Eremopyrum bonaepartis and Henrardia persica in a single clade. Upon combination of Triticeae species mating system observations and data obtained from molecular analysis of 27 protein-coding loci, Escobar et al. (2010) found that Henrardia persica is very close to Eremopyrum bonaepartis and form a clade with Eremopyrum triticeum and Agropyron mongolicum. Likewise, Escobar et al. (2011) found two well-supported sub-clades, the first formed by Henrardia and Eremopyrum bonaepartis, and the second by Agropyron mongolicum and E. triticeum. The study of Escobar et al. (2011) showed that the genera Agropyron, Eremopyrum and Henrardia were not affected by introgression and/or incomplete lineage sorting.

5.10.4.3 The Clade DasypyrumSecaleHeteranthelium

Two cpDNA data sets, one based on restriction site variation (Mason-Gamer and Kellogg 1996b) and the other on sequences encoding the rpoA subunit of the RNA polymerase (Petersen and Seberg 1997), placed Elymus species possessing the E genome, e.g., E. elongatus and E. Farctus, together with Dasypyrum on cpDNA cladograms. Sequencing of the nuclear starch synthase gene also revealed a close relation between Elymus St species and Dasypyrum (Mason-Gamer and Kellogg 2000). The cpDNA tree contained a well supported clade that included Dasypyrum, with Elymus (St and Ee species). The Dasypyrum-Elymus (St species) monophyly was also observed in cladograms obtained from RFLP similarity patterns obtained from 14 cloned fragments covering the entire cpDNA of T. aestivum (Kellogg 1992), morphological data (Kellogg 1989) and 5S RNA (Appels and Baum 1991).

The nuclear DNA data (Hsiao et al. 1995; Kellogg and Appels 1995) are incongruent with the cpDNA data, as they suggest different affinities of St species of Elymus and Dasypyrum. However, molecular phylogeny of the RPB2 (the second largest subunit of RNA polymerase II) gene sequence reveals that the Dasypyrum villosum genome is sister to the St genome of Elymus and that both diverged from the H genome of barley (Sun et al. 2008). This finding is in line with the phylogenetic relationships of monogenomic species of Triticeae inferred from nuclear rDNA (internal transcribed spacer) sequences, which established a close relation of Dasypyrum and Heteranthelium to the St genome species of Elymus.

The phyletic relationships among Dasypyrum and other Triticineae genera have been assessed at the level of morphology, protein and chloroplast and nuclear DNA sequences. Morphology-based phylogenetic analyses showed that Dasypyrum branched from a sister group of Secale within the same clade (Baum 1978a, b, 1983; Kellogg 1989; Frederiksen and Seberg 1992; Seberg and Frederiksen 2001). Dasypyrum villosum is morphologically similar to Triticum, in general (Baum 1978a, b), and to Triticum monococcum (Seberg and Frederiksen 2001), in particular. Kellogg (1989) placed Dasypyrum near Agropyron and Triticum monococcum, and Baum (1978a, b, 1983) considered Secale cereale and D. villosum as evolutionarily more contiguous to Triticum and Aegilops than to the rest of the Triticineae. Yang et al. (2006), studing genome relationship based on species-specific PCR markers, concluded that the formation of the Dasypyrum species started at the earlier stages of the separation of the sub-tribe Triticineae from the sub-tribe Hordeineae (13–15 MYA).

Phylogenetic studies based on molecular data suggested that Secale is the closest relative of the Triticum-Aegilops genera (Kellogg et al. 1996; Huang et al. 2002a; Mason-Gamer et al. 2002). When Hordeum and Dasypyrum were assessed with Secale, Triticum and Aegilops, they stood at the base of the tree topology as out-groups (Yamane and Kawahara 2005; Kawahara et al. 2008), implying a much earlier divergence between D. villosum and the common ancestor of Triticum-Aegilops.

DNA/DNA hybridization experiments, in which the genomes of Secale cereale and D. villosum were hybridized with labeled nuclear DNA from Triticum aestivum and S. cereale, revealed greater homology between Dasypyrum and Secale than between Dasypyrum and Triticum (Lucas and Jahier 1988). FISH analysis, in which the genomes of different species of the Triticeae were hybridized with species-specific molecular probes prepared from tandem repeated DNA sequences of D. villosum (pHv62) and S. cereale (pSc119.2), exhibited greater homology than the homology observed between Secale or Dasypyrum villosum genomes and those of Triticum and Aegilops (Uslu et al. 1999).

Escobar et al. (2011) included Dasypyrum and Heteranthelium in the same clade, and Secale, Taeniatherum, Triticum and Aegilops in another clade. St genome species of Elymus are sister to Dasypyrum. Lucas and Jahier (1988) concluded that the differentiation of D. villosum and Secale cereale from the other Triticineae genera occurred before speciation in Aegilops and Triticum. They assumed that T. monococcum subsp. aegilopides is the closest species to D. villosum from the Aegilos-Triticum group, and not diploid species of Aegilops. Shang et al. (2007) concluded that the St and Eb genomes of Elymus were also more closely related to the genome of Secale cereale than to the genome of Dasypyrum villosum.

The Heteranthelium transposon Stowaway is present in Dasypyrum but absent in other Triticeae species (Petersen and Seberg 2000). Variations in the PCR sequences of 6-SFT (sucrose-fructan 6-fructosyltransferase) placed Heteranthelium piliferum and D. villosum at one extreme of the phylogenetic relationships, whereas diploid species of Secale, Triticum and Aegilops were at the other extreme (Wei et al. 2000).

Taken together, the reported phylogenetic relationships of Heretanthelium remain ambiguous. Assessments of the chloroplast gene encoding ribosomal protein rps 16, performed by Hodge et al. (2010), placed H. piliferum in the same clade with Triticum monococcum, Secale cereale, and all the Aegilops species. Their results are consistent with the conclusions of Mason-Gamer et al. (2002), which were based on combined cpDNA sequences of tRNA genes, spacer sequences, rpoA genes and restriction sites. Phylogenetic relationships determined by mating systems showed that Heteranthelium piliferum is in a clade with Dasypyrum villosum (Escobar et al. 2010), despite the fact that H. piliferum is a facultative self-pollinated species, whereas D. villosum is an out-crosser. Similarly, Escobar et al. (2011) found that Heteranthelium piliferum comprises a clade with Dasypyrum villosum that branches at the base of the clade of Triticum, Aegilops, Secale and Taeniatherum. Thus, these two clades are closely related to each other.

Escobar et al. (2011) concluded that Dasypyrum, Heteranthelium, Secale, Taeniatherum, Triticum and Aegilops, evolved in a reticulated manner. The ambiguous reports concerning the phylogenetic relationships of Dasypyrum, Secale, Heteranthelium, Triticum and Aegilops may result from introgression between these genera and/or incomplete lineage sorting (Escobar et al. 2011).

5.10.4.4 The Clade AmblyopyrumTriticumAegilops

Most morphological and molecular trees included Amblyopyrum in the Aegilops clade (Seberg and Petersen 2007); several molecular trees also included several species of Elymus in this clade (Mason-Gamer et al., 1998). In phylogenetic analyses based on nuclear DNA sequences, Amblyopyrum is a monophyletic taxon, although, the relationships of Amblyopyrum within the Aegilops clade remain unresolved (Frederiksen and Seberg 1992; Frederiksen 1993). Nevertheless, its position as an intermediate between Elymus elongatus and Sitopsis species of Aegilops was suggested by Eig (1929b). Numerical analysis (Baum 1977, 1978a, b; Schultze-Motel and Meyer 1981) indicated the close relationship of Amblyopyrum with Aegilops and Triticum, but also confirmed the morphological differences (Baum 1977).

Hsiao et al. (1995) and Kellogg et al. (1996) considered diploid wheat, Triticum monococcum, to be a sister group to Elymus elongatus. Upon studying internal transcribed spacers (ITS) of the nuclear rDNA sequences, Hsiao et al. (1995) reported that Ee and Eb jointly clustered with subgenomes A, B, and D of Triticum aestivum. In accord with this finding, Liu et al. (2007), using genomic hybridization (both Southern and in situ hybridization), demonstrated that the St and Eb genomes are very closely related to the A, B and D subgenomes of bread wheat, with the closest relation being to the D subgenome. Their findings provide a possible explanation as to why most of the spontaneous bread wheat—E. elongatus ssp. ponticus translocations and substitutions occur in the D subgenome, while only some occur in the A subgenome and rarely any in the B subgenome. Taken together, the diploid donors of subgenomes A, B, and D of the wheat group may have derived from the E genome of Elymus (reviewed in Wang and Lu 2014).

Studies of nuclear DNA sequences (genes or repetitious DNA) and chloroplast DNA sequences of the wheat group (the genera Amblyopyrum, Aegilops and Triticum), have shown significant inconsistencies, possibly due to both ancient and recent inter-specific and inter-generic hybridizations and introgressions (Kawahara 2009) that usually involve distal chromosomal region much more than proximal ones. Incongruence between chloroplast and nuclear genomic data has often been reported (Sasanuma et al. 2004; Kawahara 2009; Li et al. 2014). A monophyletic origin of Aegilops and Triticum was inferred from some of analyses (e.g., Hsiao et al. 1995; Kellogg and Appels 1995; Kellogg et al. 1996; Huang et al. 2002a, b), whereas a polyphyletic origin was deduced from others (Petersen and Seberg 1997, 2000; Seberg and Frederiksen 2001; Sallares and Brown 2004; Mason-Gamer 2005; Petersen et al. 2006). It is probable that intergeneric hybridizations and introgressions from other genera of the Triticineae e.g., Elymus, Secale and others, blurred the monophyletic origin of the wheat group.

Hammer (1980) assumed that allogamous, self-incompatible plants, resembling species of Amblyopyrum muticum and Aegilops speltoides, were ancestral types to the autogamous genera of the wheat group. He reached this assumption by combining anther length and amount of pollen produced, with reductions in anther size and/or amount of pollen, which suggested increasing autogamy, for which he found strong positive correlations. Long anthers that produce substantial amounts of pollen are an indication of allogamy or of a transitional state to autogamy. Based on these findings, Hammer (1980) produced an evolutionary model delineating the sequence of origin of the various species. Hammer’s (1980) phylogenetic model assumes that Amblyopyrum separated from the ancestral Aegilops lineage at an early stage, of which Ae. speltoides is thought to be the most primitive representative of Aegilops. Phylogenetic speculation that assumes a change from allogamy (as in A. muticum) towards facultative autogamy (as in species of Aegilops, Eremopyrum, Heteranthelium, Henrardia and Triticum), coinciding with divergent morphological development, suggests separation from the ancestral Aegilops lineage at an early stage (Hammer 1980).

Dvorak and Zhang (1992a, b), analyzing repeated DNA sequences, concluded that A. muticum is close to Ae. caudata, Ae. comosa, Ae. uniaristata and Ae. umbellulata. Wang et al. (2000) compared the internal transcribed spacer (ITS) region of the ribosomal DNA in the diploids of the wheat group (including A. muticum) and observed wide divergences of this sequence between species. The highest divergence was between Ae. speltoides and A. muticum. In situ hybridization with repeated DNA markers and C-banding patterns, suggest that A. muticum occupies an isolated position, which is relatively closer to the Sitopsis species (Ae. speltoides, Ae, bicornis, Ae. sharonensis, Ae. longissima and Ae. searsii) than to other species of Aegilops (Badaeva et al. 1996a, b). Sallares and Brown (2004), who analyzed the transcribed spacers of the 18S ribosomal RNA genes, reached a similar conclusion, namely, that A. muticum has a basal position and that it is close to Ae. speltoides.

5.10.4.5 The Clade TaeniatherumCrithopsis

Taxonomists assigned the two genera Taeniatherum and Crithopsis to the subtribe Hordeineae. In most morphological trees, Taeniatherum and Crithopsis were linked to the Hordeum group, mainly because they all present multiple spikelets per rachis node (Baum 1983; Baum et al. 1987; Kellogg 1989; Frederiksen and Seberg 1992). Yet, phylogenetic studies based on molecular analyses placed Taeniatherum and Crithopsis closer to Secale, Triticum and Aegilops than to Hordeum [e.g., Mason-Gamer et al. (2002), based on chloroplast DNA, and Hsiao et al. (1995) and Escobar et al. 2011), based on nuclear DNA sequences]. Hsiao et al. (1995) used the sequences of the ITS region of nuclear ribosomal DNA and sequences of tRNA to estimate phylogenetic relationships among 30 diploid Triticeae species representing 19 genomes. They found that most of the annuals of the Mediterranean origin, i.e., species of Triticum, Aegilops, Crithopsis, Taeniatherum, Eremopyrum, Henrardia, and the perennials, Elymus farctus, E. elongatus, and Secale strictum, comprise a monophyletic group. In the parsimony tree from a more restricted species sampling, the two perennial species Elymus farctus and E. elongatus formed a sister group to Triticum monococcum, Aegilops speltoides, and Ae. tauschii, whereas Crithopsis, Taeniatherum, Eremopyrum, and Henrardia, were close to Secale. Based on this finding, Hsiao et al. (1995) supported Sakamoto (1973), who suggested that the Triticeae should be classified as two major groups, a Mediterranean group and an Arctic-temperate group, where the Mediterranean lineage evolved from Arctic-temperate species.

Mason-Gamer et al. (2002) analyzed new and previously published chloroplast DNA data from Elymus and from most of the mono-genomic genera of the Triticeae, and presented additional cpDNA data to construct the phylogeny for the mono-genomic genera. They concluded that their analysis was in agreement with previous cpDNA studies with regard to the close relationship between Secale, Taeniatherum, and TriticumAegilops. Further, their analysis provided moderate support for some relationships that were unresolved or very weakly supported in earlier cpDNA studies. These included (i) the sister relationship between Taeniatherum and Triticum–Aegilops; (ii) the placement of Heteranthelium in a Secale + Taeniatherum + TriticumAegilops clade; and (iii) the placement of the St- and Ee-genome Elymus species + Dasypyrum clade in the Secale + Taeniatherum + AegilopsTriticum + Heteranthelium clade. Yet, a close relationship between Taeniatherum and TriticumAegilops is completely at odds with the DMC1 tree (Petersen and Seberg 2000) and the morphology-based cladogram (Seberg and Frederiksen 2001). The relationship was unresolved by the ITS data (Hsiao et al. 1995), and Taeniatherum was not included in the 5S long spacer data set (Kellogg and Appels 1995). But, Escobar et al. (2011) reported that the clade containing Taeniatherum and TriticumAegilops is also seen on the 5S short spacer data tree, but only if Elymus farctus and E. elongatus are included in the clade. Some conflict exists with regard to the position of Secale in the starch synthase tree, in which a close relationship among Triticum, Aegilops, and Heterranthelium was detected, while Secale was put in a SecaleTaeniatherumElymus (Ee genome species)–Dasypyrum group (Mason-Gamer and Kellogg 2000).

The above phylogenetic studies suggest that Taeniatherum is closer to the species of the subtribe Triticineae than to those of the Hordeinae. Hence, the number of spikelets on each node (Tzvelev 1976; Clayton and Renvoize 1986), is not an adequate feature for placing such genus in one of the two subtribes.

5.10.5 Evolutionary Changes in the Polyploid Triticineae

Several Triticineae genera contain polyploid species or subspecies that developed in the Mediterranean-Central Asiatic region (Table 2.8). These genera are Elymus, Agropyron, Eremopyrum, Dasypyrum, Triticum and Aegilops. The perennial polyploid taxa of Agropyron and Dasypyrum are typical autopolyploids exhibiting multivalents at meiosis due to pairing between the homologous chromosomes of the multiple genomes in the tetraploid and hexaploid cytotypes of these species (Dewey 1961; Ohta and Morishita 2001). In these taxa, although it causes some sterility, multivalent pairing increases plant heterozygosity, that, together with the perennial growth habit and allogamy, emphasizes the evolutionary principle of genetic flexibility. On the other hand, autopolyploids of the subspecies of Elymus elongatus and E. farctus, are diploidized autopolyploids exhibiting almost exclusive bivalent formation at meiosis due to strict intra-subgenomic pairing between fully homologous chromosomes (Cauderon 1958, 1966; Charpentier et al. 1986). This type of autopolyploidy, combined with facultative self-pollination, leads to full fertility, homozygosity and permanent heterozygosity between subgenomes, emphasizing the evolutionary principle of immediate fitness.

Eilam et al. (2009) determined genome size in several diploid and autotetraploid cytotypes of several Triticeae species. DNA content in the typical autotetraploids did not undergo downsizing, namely, it did not deviate significantly from the expected sum of the two diploid genomes. In contrast, the DNA content of the diploidized autotetraploids underwent downsizing, which was presumably required for the bivalent pattern of pairing due to restriction of pairing to intra-subgenomic homologs. Studies of Eilam et al. (2009) show that the DNA content of the diploidized autotetraploids, E. elongatus and E. farctus underwent downsizing, whereas that of the typical autopolyploids of Agropyron showed the expected DNA content (Table 2.4).

All the annual polyploid species of the genera Eremopyrum, Triticum and Aegilops are allopolyploids, which facilitates adaptation to the Mediterranean and Central Asiatic climate. Leitch and Bennett (2004) showed that there is genome downsizing in polyploids as compared to their diploid parents. Eilam et al. (2008) determined DNA content in allopolyploid species of Aegilops and Triticum and found that most contain less DNA than expected. Feldman et al. (1997), Ozkan et al. (2001) Shaked et al. (2001), Kashkush et al. (2002) and Salina et al. (2004) showed that DNA sequences are eliminated in newly formed allopolyploids at the same times as or very soon after chromosome doubling. It was assumed that elimination of these DNA sequences may ameliorate the harmonious co-existence of the two or more homoeologous genomes that reside in one allopolyploid nucleus. Furthermore, it is assumed that the homologous-specific DNA sequences play a role in homologous recognition at the beginning of meiosis and initiation of pairing, thus, their elimination from one of the subgenome, leads to diploid-like meiotic behavior in the allopolyploids, i.e., exclusive intra-subgenomic pairing of homologous chromosomes. This pattern of pairing ensures full fertility and prevents intergenomic gene exchanges, thereby sustaining plausible inter-subgenomic genetic interactions that may lead to permanent positive heterosis between subgenomes.

Allopolyploid patterns of gene expression might be intermediate (between that of the two parents), dominant (similar to one of the parents) or overdominant (greater than that of the parents). Overdominance can produce novel traits not found in the parents. It can be caused by novel cis–trans interactions between regulatory elements of the different subgenomes that harbor the same nucleus, as shown in yeast (Tirosh et al. 2009). The genetic system of the allopolyploids may contribute to the build-up of genetic variability and creation of populations with archipelagoes of genotypes via interspecific introgression (Zohary and Feldman 1962), thereby increasing their adaptability, fitness, competitiveness and capacity to colonize rapidly newly-opened ecological niches.