Background

Transition of plant from the vegetative stage to the generative one represents one of the most critical stages in the life cycle of a plant, since on its proper course depends: whether the plant gives the offspring. An active study of genetic mechanisms of transition to the stage of flowering on model plant species has made it possible to identify a complex network of genes that control this transition, depending on internal and external factors, such as temperature, length of daylight, hormone concentration, et cetera. The key genes of this system were identified and functionally characterized in the basic model object of plants, Arabidopsis thaliana (Reviewed in [1]). These genes can be divided into three types: 1) genes of regulatory cascades; 2) genes- integrators; 3) genes of the flowering meristems identity. Genes of the first type are conductors that transmit signals from separate external or internal factors to genes- integrators. The latter combine various regulatory cascades and, under a certain combination of factors, directly activate genes that induce the formation of flowering meristems.

In cereal plants, the duration of the vegetative period is determined by the interconnected systems of genes of vernalization response (VRN1, VRN2), response to photoperiod (PHYC, PHYB, PPD1, CO1, CO2), which trigger the genes of flowering meristems: VRN1 and LFY2 through the gene- integrator FT1 (Reviewed in [2, 3]). FT1 is a homolog of Arabidopsis FT, which encodes a mobile protein critical for the transmission of day-length signals from leaves to the shoot apical meristem [4]. A special role in this system is assigned to the VRN1 gene, which combines the functions of a response to the vernalization signal and induction of the formation of flowering meristems from the apical cells of sprout. This gene encodes a MADS- domain containing transcription factor homologous to the product of the AP1 gene of Arabidopsis [5]. Prolonged exposure to low temperatures (vernalization) induces the transcription of VRN1 in both leaves and at the shoot apex [6, 7]. In the apex, VRN1 initiates the transition of the shoot apical meristem to the reproductive stage whereas, in the leaves, it upregulates FT1 directly [8], or through the VRN2 gene [9]. Initially, deletion of the VRN1 gene was thought to be the cause of irreversible loss of reproductive capacity in a radiation mutant in diploid wheat (Triticum monococcum L.), designated maintained vegetative phase (mvp) [6, 10]. However, a recent studies establish that, besides VRN1, some other genes, like PHYC, VRN-Am1, may be essential for wheat flowering raising the possibility that their deletion is responsible for the nonflowering phenotype of the mvp mutant [3, 11]. It should be noted that in allopolyploid cereals, including common hexaploid wheat T. aestivum L. (2n = 42; BBAADD), the genetic mode of the flowering regulation is greatly complicated due to the presence of homoeologous copies of each gene.

In common wheat the homoeologous loci VRN-A1, VRN-B1 and VRN-D1 were mapped on chromosomes of group 5 [12,13,14]. All three homoeologs are functional and their combination provides a predominant part of biodiversity of winter and spring forms of this species [15]. Winter wheat contains recessive alleles at each of these loci, while spring wheat, in at least one of them, has a dominant allele. The dominant allele VRN-A1 causes complete insensitivity to vernalization and is epistatic with respect to the dominant VRN-B1 and VRN-D1 alleles, which cause little sensitivity to vernalization [16]. A wide range of the heading time in spring wheat is provided by various combinations of dominant alleles of the homoeologous VRN1 loci that also influence the height of plants and yield [17,18,19]. The earliest in the time of heading are genotypes with three dominant alleles, however, they tend to be low-yielding. Genotypes with two dominant alleles (VRN-A1/VRN-B1 or VRN-D1) ripen somewhat later and have higher yields. The presence of only one of the weak alleles VRN-B1 or VRN-D1 provides the latest ripening of spring varieties [16, 20]. The purpose of this mini-review is to analyze the available molecular genetical data on the formation of the VRN1 allelic system at the following stages of wheat evolution: 1) in diploid donor species of A, B, G, and D genomes; 2) in tetraploid species (BBAA and GGAA genomes); 3) in hexaploid wheat (BBAADD).

The main regulatory regions of VRN1 loci affecting their expression

The molecular analysis of various VRN1 homoeoalleles was first carried out, mainly, on polyploid wheat species as an object. It was shown that the dominant alleles of these loci are, mainly, associated with mutations in two basic regulatory regions: the promoter and the first (1st) intron [21, 22]. It is assumed that mutations in these regions disrupt interaction with regulatory factors leading to independent of ambient temperature, constitutive expression of VRN1 in spring forms [2]. The dominant VRN-A1 alleles associated with insertions (VRN-A1a) or deletions (VRN-A1b, VRN-A1d, VRN-A1e) in the promoter have been identified (Table 1). Initially, mutations in the CArG-like motif located 170 bp upstream from the starting ATG- codon were considered as the main determinants of the spring type [5]. However, later this motif was found to be invariant in a number of spring genotypes [23, 24]. Moreover, the complete deletion of the CArG-box retained sensitivity to vernalization. It was assumed that the CArG box can participate in response to photoperiod by binding a repressor protein expressed on a short day [25].

Table 1 Allelic variability in the promoter region and 1st intron of VRN1 genes of wheat
Full size table

The so-called VRN box (“TTAAAAACCCCTCCCC”), directly 5′-adjacent to the CArG box [24], seems to have the greatest impact on the sensitivity to vernalization. Namely this motif is affected by insertions and deletions in the dominant alleles VRN-A1a, VRN-A1d, VRN-A1e, strictly associated with the spring type of growth [21, 26,27,28]. Recently, it has been established that even point mutations in the VRN box (the alleles Vrn-A1b.3, Vrn-A1b.4 and Vrn-A1i) are sufficient for transformation into the spring type and the combination of these mutations can modulate the response to vernalization [29]. Upstream of the VRN box are binding sites for bZIP transcription factors (ACGT sites) [30, 31]. These sites can bind the TaFDL2 protein, which forms a complex with the product of the VRN3 gene-integrator. It is this interaction that is supposed to provide the inducing effect of the VRN3 gene on the transcription of VRN1 [32].

Besides the VRN1 promoter region, the region of the 1st intron also began to be considered as an important determinant of the spring habit after the discovery in wheat and barley of a number of dominant alleles containing only deletions in the 1st intron (the critical region is within 2.8 kb from the beginning of this intron) [22, 33, 34]. For example, the dominant alleles VRN-A1c, VRN-B1a and VRN-D1a, widely distributed in common wheat, contain large deletions in intron 1, whereas their promoter regions do not differ from those in the corresponding recessive alleles of these loci (Table 1). It was found that the repressed state of the VRN1 gene observed in winter forms prior to vernalization is associated with a high level of methylation of histone H3 in the flanking region of the 1st intron and the degree of methylation is significantly reduced by deletions overlapping this critical region [35, 36]. An epigenetic model of the VRN1 gene regulation was proposed, in which the 5′-flanking region of the 1st intron is attributed to the binding of chromatin modification factors affecting the transcription level of VRN1 [37, 38]. According to the post-transcriptional model of regulation, this region contains a binding site for the Glycine-rich RNA-binding protein 2 (TaGRP2), which inhibits VRN1 expression by binding with the VRN1 pre-mRNA [39, 40].

An assessment of the relative level of transcription of various homoeoalleles VRN1 was carried out using three near-isogenic lines Triple Dirk of T. aestivum bearing the dominant alleles VRN-A1, VRN-B1 and VRN-D1 [41]. An established higher level of transcription of VRN-A1, compared to two other homoeoalleles, leads to an earlier heading of the corresponding line. This may be due to the existence of several paralogs of this gene [42, 43] which may provide its overexpression, explaining the epistatic effect of VRN-A1 in the digenic dominant VRN1 genotypes (see above). It should be noted that differences in the level of transcription and phenotypic effect between dominant alleles of one homoeologous locus can be comparable with those between alleles of different loci. This was shown by Shcherban et al. [38] using near-isogenic lines containing VRN-B1a and VRN-B1c alleles. The revealed level of transcription of VRN-B1c was 10 times higher, probably, due to large structural changes (deletion and duplication) in the 1st intron compared with the first allele.

The VRN1 gene- associated prerequisites of spring growth habit in diploid precursors of wheat

Most of the wild Triticeae species both diploid and polyploid have a winter growth habit, suggesting that the recessive VRN1 allele is the ancestral, original form [44]. Spring growth habit determined by dominant VRN1 alleles could result from selection of mutations appearing during the adaptation of plants to environments that differ from the original. Some authors suggest that the spring type might have evolved from a previous winter plant prototype as an adaptation to warmer conditions, the feature that might be characteristic of both diploid and polyploid species [45, 46]. In support of this idea, Kato et al. [47] found that the distribution of spring forms of wild species T. dicoccoides Thell (2n = 28; BBAA) is sporadic and restricted to areas where vernalization is not required. Differentiation into spring and winter types was revealed in all Triticum and Aegilops diploid (2n = 14) species with the frequency of spring forms varying from 2% (T. urartu) up to 100% (T. sinskajae) [44].

In diploid species related to the A-genome (T. boeoticum Boiss., T. monococcum L., T. urartu Tum. ex Gandil.), the dominant VRN1 alleles were most common among accessions of einkorn wheat T. monococcum, one of the earliest cultivated forms of wheat [48]. Shcherban et al. [27] revealed Vrn-A1ins allele (Table 1) in 35% of the studied accessions of T. monococcum and a total frequency of all dominant alleles was close to the percentage of spring forms (47%), established earlier for this species [44]. In the wild progenitor of T. monococcum - T. boeoticum, a few spring forms were associated with the VRN-A1h allele, although it was also found in some winter accessions (an optional type) [27]. The dominant VRN-A1g allele was quite rare in both T. monococcum and T. boeoticum [26]. The structure of regulatory regions of VRN1 was highly conserved in the predominantly winter species, T. urartu, a donor of A-genome to polyploids. Interestingly, the latter species, in comparison with the other two diploids, has different mutations in both the promoter (nucleotide substitutions) and 1st intron (deletion of 1.4 kb) of VRN1 providing specific identification of T. urartu and the corresponding A-genome in polyploids. However, all these mutations displayed no influence on vernalization requirements in all studied accessions, therefore the common VRN1 allele of T. urartu is recessive (vrn-A1u; Table 1).

In Ae. speltoides Tausch, the presumable donor of B and G-genomes to polyploids, the structure of the promoter of VRN-1 was more polymorphic compared to the Triticum A-genome species, however, most of this variation was upstream from the conserved region encompassing about 0.3 kb from the start codon and containing the putative regulatory sites [49]. No large indels were revealed in the VRN-1 1st intron of Ae. speltoides. All the studied accessions of this species were of winter type (allele vrn-B1sp; Table 1). Untill now, the dominant VRN1 alleles associated with spring growth have not been found in Ae. speltoides.

The only dominant allele VRN-D(t)1 was found in 9 out of 211 accessions of Ae. tauschii Coss., the proposed donor of D-genome to common wheat [50]. This allele contains deletion of a 5.4-kb sequence in the critical region of the VRN1 1st intron (see above) and was strongly associated with a lack of vernalization requirement (Table 1).

Thus, the dominant VRN1 alleles within diploid species, presumable donors of A, B(G) and D-genomes, have a rare occurrence as compared with polyploid species (see below) that is consistent with the predominant winter type of these species. The only exception is T. monococcum, in which the spread of the dominant VRN1 alleles and corresponding spring forms could occur due to the selection process carried out by human. This process resulted to a wide distribution of einkorn wheat from the Fertile Crescent to the more northern regions including the Caucasus, the Balkans, and central Europe [51].

Polymorphism of VRN1 loci in tetraploid wheat species of different origin

An important issue is the origin of the VRN1 alleles associated with spring growth habit in the first tetraploid wheat species of two independent evolutionary lines: the Emmer line (BBAA) and the Timopheevii line (GGAA). The first line begins with the wild tetraploid T. dicoccoides, from which all the main cultural allopolyploids, including hard T. durum Desf. (BBAA) and common hexaploid wheat originated. The second line includes the wild species T. araraticum Jakubz. and derived from it, domesticated spring form- T. timopheevii (Zhuk) Zhuk.

Comparison of the VRN1 alleles set of in the first tetraploid wheat species and those in the diploids shows that in the first tetraploids, the emergence of spring forms was associated with the appearance of new dominant VRN1 alleles that are not related in origin to the dominant alleles of diploid predecessors (Table 1). The set of alleles of T. dicoccoides differs from that of Timopheevii tetraploids indicating an independent origin of the spring type of growth within these two branches of allopolyploids. In spring wheat, T. timopheevii, this trait is apparently associated with the insertion of a miniature transposon (MITE) in the 1st intron of the VRN-A1 locus (alleles Vrn-A1f-del/ins and Vrn-A1f-ins; Table 1) [49, 52]. In T. dicoccoides, spring forms were associated with changes in the VRN-A1 promoter: deletions of variable size (Vrn-A1b, Vrn-A1d and Vrn-A1e alleles), nucleotide substitution (Vrn-A1i) and a large deletion in the 1st intron (Vrn-A1L) [26,27,28,29]. Besides, after the divergence of both lines of tetraploids some changes in the VRN-A1 promoter appeared to be not associated with spring growth habit including a 8 bp insertion in all Emmer wheats and a 50 bp deletion in Timopheevii wheats [49].

In contrast to VRN-A1, the VRN-B1 locus in 83 accessions of T. dicoccoides displayed no significant indels, affecting the main regulatory sites and associated with the spring type [27]. The VRN-B1dic allele with a number of minor mutations within the promoter (SNPs, deletions up to 7 bp) was detected in one spring accession of T. dicoccoides, however, its influence on vernalization response has not been established definitely (Table 1).

In one accession of the Turanian wheat T. turanicum Jakubz. (BBAA) allele VRN-B1a (does not correspond to the dominant Vrn-B1a; Table 1) with an insertion in the VRN-B1 promoter was found [26]. Interestingly, this insertion is homologous to the insertion in the VRN-A1a allele, however, it has other location- at position −100 from the start codon. The insertion of a retrotransposon of 5.4 kb in the VRN-B1 promoter was detected in tetraploid Persian wheat T. carthlicum Nevski, and, unlike the previous case, its association with the spring forms was confirmed [53].

Besides T. dicoccoides, the Vrn-A1L allele containing large deletion in the 1st intron was identified in other Emmer tetraploid species of a later origin: T. carthlicum, T. polonicum L., T. durum [29, 54]. In T. durum the dominant VRN-B1a and VRN-B1с alleles appeared, having deletions of 6.8 and 7.6 kb, respectively, in the 1st intron (Table 1). The latter allele seems to have originated from VRN-B1a due to an additional deletion of 0.8 kb and a duplication of 0.4 kb [38, 55]. The both VRN-B1 alleles are higly distributed among spring varieties of common wheat (see below).

The only identified mutation affecting VRN-G1 locus in Timopheevii wheats is an insertion of foldback element in the promoter region at position −99 from the ATG-codon (allele VRN-G1a) [26]. However, in fact this allele is recessive since it was later found in winter accessions of T. araraticum [49].

Thus, in the first allotetraploid wheat species of both Emmer and Timopheevii lines the most variability affected the locus VRN-A1, while the other homoeolog- VRN-B1 (G1) remained unchanged in spring forms (T. dicoccoides), or mutation in it did not result to the spring phenotype (Timopheevii, allele VRN-G1a). On a later stages of evolution of Emmer allotetraploids, especially, in the cultivated T. durum, different VRN-B1 alleles appeared containing mutations in the promoter or 1st intron regions.

Variability of VRN1 loci in hexaploid wheat

Many known VRN1 alleles, common in hexaploid wheat, originated at the tetraploid stage. This refers to such alleles as VRN-A1a, VRN-A1b, VRN-A1c, VRN-B1a, VRN-B1c [21, 22, 26, 29, 55,56,57,58]. The most ubiquitous among them are the VRN-A1a allele with a foldback insertion in the VRN-box and the VRN-B1a allele. Simultaneously, a number of a new alleles including VRN-B1 (VRN-B1b) and, especially, VRN-D1 (VRN-D1a, VRN-D1b, VRN-D1c, VRN-D1s) loci appeared in hexaploid wheat (Table 1).

The VRN-B1b allele appears to have originated from the VRN-B1a allele, since along with large deletion in the 1st intron, characteristic of the latter, it has an additional SNP and a 36 bp deletion [57]. This allele has been detected in common wheat originating from North America and was associated with the spring growth habit [58].

The dominant VRN-D1a allele was first isolated from spring near-isogenic line TDE (Table 1). As shown, VRN-D1a is the predominant allele in spring wheat genotypes adapted to tropical and subtropical regions [59,60,61,62]. The VRN-D1b allele has been originated from the previous allele due to SNP in the CArG-box at the promoter region [63]. Since the F2 population plants with VRN-D1b headed 32 days later than those with VRN-D1a, the authors suggest that a single nucleotide mutation at promoter region may modify the basal activity level of an allele of VRN1 that is already active (due to the loss of segment in intron 1). The VRN-D1c allele with 174-bp insertion in the promoter region was discovered in three out of 205 Chinese wheat cultivars [64]. As suggested, the insertion in Vrn-D1c may contribute to the increase in gene expression and cause early heading and flowering due to transcriptional cis-elements. In the same year Muterko et al. [65] found the VRN-D1s allele resulted from DNA transposon insertion of 844 bp in the 1st intron and associated with spring form (Table 1).

It can be assumed that the wide distribution of spring forms and associated with them combinations of VRN1 homoeoalleles in common wheat is largely due to artificial selection directed to adaptation of wheat to different climatic conditions. As one of the illustrations, the analysis of distribution of different VRN1 haplotypes among varieties of spring common wheat from different eco-geographical regions of Europe and Russia can be presented [34, 59]. For the group of cultivars of the northern and central Europe and the most of territory of Russia, the digenic dominant haplotypes at the VRN-A1 and VRN-B1 loci are characteristic, while for the southern European group of cultivars, the monogenic dominant haplotypes containing either the dominant VRN-B1 or VRN-D1 are mostly widespread. The latter group of cultivars had a later date of heading compared to the cultivars of the first group. Therefore, the monogenic VRN-B1/VRN-D1 haplotypes could have a breeding advantage in the subtropical southern regions, providing a longer vegetative period that is most effective for this zone in terms of yield. The early ripening, digenic dominant haplotypes, in turn, have an advantage in regions with a temperate climate, where there is a risk of early fall frosts [59].

Conclusions

In the diploid predecessors of polyploid wheats, the dominant VRN1 alleles associated with spring forms have a limited distribution, mainly, in a specific climatic regions where such forms may receive a natural selective advantage over winter forms. The origin of spring forms in polyploids, starting from the first wild-growing tetraploid species (BBAA/GGAA) is due to emergence and spread of a new dominant VRN-1 alleles, not related by origin to alleles of the diploid species, presumed donors of A-, B (G) and D- subgenomes. In cultivated wheat polyploids, especially in common wheat, the process of artificial selection significantly enhance the spread of different VRN1 alleles at each of the homoeologous loci and their combinations that determine the optimal heading time for a district geographical regions.