Allelic Variations in Vernalization (Vrn) Genes in Triticum spp.

Abstract

Rapid climate changes, with higher warming rates during winter and spring seasons, dramatically affect the vernalization requirements, one of the most critical processes for the induction of wheat reproductive growth, with severe consequences on flowering time, grain filling, and grain yield. Specifically, the Vrn genes play a major role in the transition from vegetative to reproductive growth in wheat. Recent advances in wheat genomics have significantly improved the understanding of the molecular mechanisms of Vrn genes (Vrn-1, Vrn-2, Vrn-3, and Vrn-4), unveiling a diverse array of natural allelic variations. In this review, we have examined the current knowledge of Vrn genes from a functional and structural point of view, considering the studies conducted on Vrn alleles at different ploidy levels (diploid, tetraploid, and hexaploid). The molecular characterization of Vrn-1 alleles has been a focal point, revealing a diverse array of allelic forms with implications for flowering time. We have highlighted the structural complexity of the different allelic forms and the problems linked to the different nomenclature of some Vrn alleles. Addressing these issues will be crucial for harmonizing research efforts and enhancing our understanding of Vrn gene function and evolution. The increasing availability of genome and transcriptome sequences, along with the improvements in bioinformatics and computational biology, offers a versatile range of possibilities for enriching genomic regions surrounding the target sites of Vrn genes, paving the way for innovative approaches to manipulate flowering time and improve wheat productivity.

1. Introduction

Wheat species (Triticum spp.) are classified based on ploidy level into diploids (2n = 2x = 14), tetraploids (2n = 4x = 28), and hexaploids (2n = 6x = 42), including wild and domesticated species as well as landraces, and old and modern elite cultivars [1,2].

The species of the Triticum genus derive from spontaneous hybridization and domestication events with closely related goat-grass species (Aegilops spp.). In an allopolyploid wheat, the genomes of the diploid progenitors become homoeologous subgenomes, because multiple paralogues (copies of a gene) are found at each locus. Many genes in the subgenomes are expected to be similar in sequence and regulation, but others might be divergent. Polyploidy appears to have occurred spontaneously in the Triticeae tribe in different periods of history, making the Triticum genus an exceptional lineage for investigating plant allopolyploidy, evolutionary dynamics, agricultural domestication processes, genetic diversity, and the mechanisms governing adaptation to diverse local environments or stress conditions [1]. Several studies demonstrated the important role played by polyploidy in the diversification of plant species, gene evolution, and the domestication processes in crops [3,4,5,6]. In wheat, domestication led to the appearance of soft glumes, nonfragile rachis, and free-wheat characteristics, increasing the diversity of wheat grains to end uses, and improving grain yield and adaptability to diverse environmental conditions [7,8]. Additional traits such as grain yield, shape and size of the seed, plant height, and flowering time were modified during domestication and by the subsequent breeding process [9]. Specifically, the spread of domesticated wheat originating from the Fertile Crescent has necessitated adaptation to new environments, facilitated by favorable alleles at critical genetic loci [10]. These traits provided allohexaploid bread wheat a competitive advantage in different environmental contexts. Bread wheat (T. aestivum) is generally more salt-tolerant than its tetraploid progenitor emmer (T. Turgidum subsp.) [11] and it has been associated with adaptation to cold stress [12,13]. However, among all these traits, the change in heading date and/or flowering time was one of the key traits allowing wheat to spread to new regions. The control of flowering time is crucial for the adaptation of the crop to different climatic conditions and has a significant impact on grain yield in Triticeae [14,15]. The regulation of flowering time is primarily determined by three groups of loci, of which two interact with environmental conditions [16]. Vernalization genes (Vrn) control the requirement for a cold period to transition from the vegetative to the reproductive phase, whereas photoperiod genes (Ppd) determine the response to day length (Distelfeld et al. [16]). The third group of loci is known as ‘narrow-sense earliness’ or ‘earliness per se’ (Eps), reflecting its independent role from vernalization and photoperiod [17]. Vrn genes play crucial roles in controlling growth habits and they represent the most important adaptative strategy to postpone heading after the winter stage, preventing frost damage [18,19,20]. The identification of Vrn genes has contributed to the classification of wheat cultivars into winter growth habit, in which exposure to cold temperatures is required to induce flowering, spring growth habit, which does not require exposure to cold temperatures to induce flowering, and intermediate growth habit [16,19,21,22,23]. Recently, a dozen functional and non-functional Vrn alleles have been identified in different wheat species, revealing their different contribution to wheat phenology. Identifying variation in these genes is a key point in controlling wheat phenology and satisfying the growing conditions prevalent in each agro-environment.

However, to better exploit the new information and define new phenological models of wheat, it is necessary to share with the scientific community the classification of the known allelic variants of the Vrn genes, highlighting ambiguous cases of nomenclature, which should be resolved.

Considering the above reasons, in this review, we first describe the main features and functions of Vrn genes and alleles, named according to Boden et al. [24], and then we summarize recent comprehensive studies of Vrn alleles at different levels of wheat ploidy that could be of great support in developing resilient wheat cultivars better adapted to higher or lower latitudes and different climatic regions. Furthermore, the review emphasizes the need for innovative genomic tools to effectively discriminate allelic variants, a critical obstacle in advancing the phenotypic characterization of individual alleles.

2. Characteristics and Functions of Vrn Genes

To date, four main Vrn genes have been described in wheat (Vrn-1, Vrn-2, Vrn-3, and Vrn-4) [19,25,26,27].

The Vrn-1 gene encodes a MADS-box transcription factor (TF) that is homologous to APETALA1 (AP1) in Arabidopsis and plays a crucial role in the vernalization response by regulating the transition of the vegetative shoot apical meristem to the reproductive phase [19,28,29,30,31]. The Vrn-1 gene was also named WAP1 by Trevaskis et al. [31], who reported that Vrn-1 was strongly expressed in spring wheat, moderately expressed in semi-spring wheat, and not expressed in winter wheat. Bread wheat harbors three copies of Vrn-1 (Vrn-A1, Vrn-B1, and Vrn-D1) located in the middle of the long arms of homoeologous chromosome 5 [19,32,33,34,35], in the same regions where the two copies of durum wheat (Vrn-A1 and Vrn-B1) have been mapped [36,37].

The Vrn-2 gene plays a crucial role as a flowering repressor, accelerating wheat flowering [19,25]. The Vrn-A2 locus includes two tandemly duplicated CCT domain genes (CONSTANS, CO-like, and TOC1), namely ZCCT1 and ZCCT2. Both are repressed during vernalization under long days [25], enhancing the activation of Vrn-1 and promoting flowering [38]. In diploid and tetraploid wheat, the Vrn-2 gene consists of two homoeologous copies, namely Vrn-A2 and Vrn-B2 [25], mapping on chromosomes 5AL and 4BL, respectively. Vrn-A2 probably translocated from 4AL to 5AL in hexaploid wheat [39]. Indeed, Vrn-B2 was mapped on 4BL but not 5BL, although highly similar sequences were also reported on chromosomes 2BS, 4BS, and 5DL, suggesting that duplication, deletion, and translocation events have impacted the Vrn-B2 locus across different hexaploid wheat varieties [39]. Vrn-D2 was detected in a genome contig sequence from 4DL but not from 5DL, suggesting that no translocations occurred between 4DL and 5DL in hexaploid wheat [39]. One or more homoeologous copies might also exist for Vrn-D2 on chromosome 4D, although further analysis is needed to clarify its genomic position in bread wheat [16,40].

Vrn-3 produces a mobile protein that functions as a flowering activator [26] and, as homologous to FLOWERING LOCUS T (FT) of Arabidopsis [26,41], is induced by long days, further accelerating reproductive apex development [26,38,42]. In hexaploid wheat, the three copies of Vrn-3 (Vrn-A3, Vrn-B3, and Vrn-D3) were mapped on chromosomes 7A, 7B, and 7D, respectively [26,43,44,45,46]. Yan et al. [26] first suggested that Vrn-3 was linked to a gene similar to Arabidopsis FT; the latter was then proposed by Tamaki et al. [47] as the candidate for Florigen, a locus encoding the mobile signal that induces floral initiation at the shoot apex [47]. FT encodes a poly ethanol amine binding protein (PEBP) [48,49], a class of proteins implicated in cellular signaling [50,51]. In Arabidopsis, FT expression increases in leaves when plants are exposed to long days [48,49], allowing its protein to be transported to the shoot apex to promote flowering [52]. Also, in wheat and barley, the expression of FT is induced by long days and promotes flowering [53].

Vrn-4 encodes a MADS-box TF very similar to Vrn-1 but, unlike Vrn-1, Vrn-2, and Vrn-3, its function has not yet been well clarified [54,55]. Vrn-D4 probably operates upstream of the positive regulatory feedback loop connecting Vrn-1, Vrn-2, and Vrn-3 [54,56]. It is located on the centromeric region of chromosome 5D in hexaploidy wheat [44] and it originated by the insertion of a segment on chromosome5AL, carrying the Vrn-1 gene (~290-kb), into the proximal region of chromosome 5D [27]. The insertion of the 5AL region includes a copy of Vrn-A1 with distinctive mutations in its coding and regulatory regions [27]. This insertion is almost fixed in the ancient wheat from South Asia [27], especially in India and nearby regions [57,58]. The gene was found for the first time in a wheat variety of Australian origin (i.e., Gabo) [59], which probably inherited it from an Indian variety (i.e., Muzaffarnagar) [60]. Vrn-4 was mainly found at a higher frequency among wheat accessions of the ancient subspecies T. aestivum subsp. sphaerococcum from South Asia [58], a subspecies with an increased differentiation in the centromeric region of chromosome 5D, indicating that Vrn-4 likely played a role in local adaptation through positive selection [27]. The centromeric region of homoeologous group 5 includes the TaVIL1 gene [55], which encodes a homolog of the Arabidopsis VIL1 (Vernalization Insensitive 3 VIN3-like 1). This gene, together with VIN3, is crucial for the epigenetic memory of vernalization and plays a role in the photoperiodic regulation of flowering time in Arabidopsis [61]. Taking into account its similarity with the Arabidopsis VIL1 gene, its centromeric location, and its increased expression during vernalization, the wheat TaVIL-D1 gene was previously considered as a potential candidate gene for Vrn-D4 [55].

3. Allelic Variations within Vrn Genes at Different Ploidy Levels

Figure 1 shows the number of alleles investigated in this review, with about 100 alleles examined, including 69 from Vrn-1, 6 from Vrn-2 and 23 from Vrn-3. The highest number of alleles was identified in the promoter region (about 60), followed by the intron region (about 20) (Figure 1; Tables S1–S3).

Natural variation in vernalization requirements in temperate cereals is strongly associated with mutations at Vrn-1, Vrn-2, Vrn-3, and Vrn-4 genes [16,19,25,26,27]. Spring and facultative growth habits harbour one or more dominant alleles at the Vrn-1, Vrn-3, or Vrn-4 loci, while winter wheat harbour dominant allele(s) at the Vrn-2 locus but recessive alleles at the other three loci [54]. Molecular analysis of Vrn loci allowed the scientific community to establish that multiple dominant alleles of these loci emerged through mutations occurring in two key regulatory regions: the promoter and the first intron. Specifically, Vrn-A1 predominately carries mutations in the promoter region, whereas Vrn-B1 and Vrn-D1 predominately harbor deletions within the first intron [56,62,63,64]. In addition, copy number variations (CNV) at one of the Vrn-1 loci were also reported in bread wheat by Diaz et al. [65] and Strejčková et al. [63].

3.1. Allelic Variation of Vrn-1 at the Promoter Level

Genetic variations at the promoter level may significantly impact Vrn-A1 expression and regulation [66] (Figure 1, Figure 2 and Figure 3; Table S1). Thus, understanding their diversity in genotypes with different ploidy levels can provide a valuable resource to further investigate the genetic basis of FT regulation in wheat. The promoter of Vrn-1 is considered a repertoire of regulatory elements, of which CArG-box, VRN-box, and ACGT-motif are the most studied [66]. VRN-box is characterized by a 16 bp region (“TTAAAAACCCCTCCCC”) and is considered the most influential on the “winter-spring” growth habit [56], whereas CArG-box (a common binding site for MADS-box) is not critical, since genotypes with a fully deleted CArG-box region show a preserved vernalization machinery [62,67]. Distinct novel genetic variations have been revealed to be situated within the regulatory region of Vrn-A1 (Figure 2).

Vrn-A1a stands out as one of the most significant and potent spring alleles [36]. It has a duplicated promoter region carrying characteristic foldback elements. The two fragments differed from the recessive vrn-A1 allele by the insertion of a 222 bp foldback element in the larger fragment and a 131 bp foldback element in the smaller one [36]; it was reported that Vrn-A1a was predominant in spring varieties released in the United States and Argentina between 1970 and 2004 and hypothesized that the increase in Vrn-A1a frequency in this germplasm was related to the introduction of the semi-dwarf germplasm from CIMMYT during the 1970s. Indeed, the allele Vrn-A1a was not present in a collection of durum wheat landraces analyzed by Royo et al. [68], which were typically characterized by tall plants, long coleoptiles, and early vigor. Later, Muterko et al. [62] described three different variants of Vrn-A1a designated as Vrn-A1a.1, Vrn-A1a.2, and Vrn-A1a.3. Vrn-A1a.1 and Vrn-A1a.3 corresponded to the known Vrn-A1a allele described by Yan et al. [36] in hexaploid and tetraploid wheat, whereas Vrn-A1a.2 was novel and compared to Vrn-A1a was characterized by two deletions (16 bp and 4 bp) within the MITE element [36,37,62] (Figure 1, Figure 2 and Figure 3; Table S1).

Yan et al. [36] also described the allele called Vrn-A1b in tetraploid accession (T. durum), which is characterized by a 20 bp deletion in the 5-UTR and two mutations in the identical host direct duplications (HDD) region. Within the Vrn-A1b allele, six variants (Vrn-A1b.1, Vrn-A1b.2, vrn-A1b.3, vrn-A1b.4, Vrn-A1b.5, and Vrn-A1b.6) were described by Muterko et al. [62]. These alleles are characterized by variants within the A-tract and C-rich regions of the VRN-box [62,69]. Mutations within the VRN-box differentiate variants of Vrn-A1b and Vrn-A1i from the recessive vrn-A1 [62]. The VRN-box of all Vrn-A1b alleles along with Vrn-A1i harbor polymorphisms in the A-tract, although Vrn-A1b.5 and Vrn-A1b.6 can be distinguished from Vrn-A1i and vrn-A1b.4 by an SNP within the C-rich segment. Moreover, Konopatskaia et al. [64] described a novel allele, named Vrn-A1b.7, characterized by deletions of 20 bp located 137 bp upstream of the start codon and mutations within the VRN-box (Figure 1, Figure 2 and Figure 3; Table S1).

Tranquilli and Dubcovsky [70] also identified variants within the VRN-box. vrn-A^m1 and Vrn-A^m2 were found in diploid T. monococcum and were reported as dominant for spring and winter growth habits. Sequence analysis revealed SNPs in the A-tract of the VRN-box in T. turgidum and T. durum and validated the identification of the vrn-A^m1 allele for the accession of T. monococcum [69]. Subsequently, Muterko et al. [69] demonstrated the existence of a 10 bp deletion in diploid wheat (T. monococcum), as well as some natural variants, Vrn-A^m1a, vrn-A^m1b [67], and Vrn-A^m1g [20], that exhibited deletions or a complete absence of the CArG-box, as in the specific case of vrn-A^m1b. Natural variants within the other regulatory regions (i.e., CarG box and/or G box) were also identified in tetraploid and hexaploid wheat [20,62,71]. Two alleles, named Vrn-A1d and Vrn-A1e, harbored 32 bp and 54 bp deletions within the CarG box, respectively [36], whereas Vrn-A1f exhibited a substantial 50 bp deletion within the −62 and −112 bp region; it also displayed a smaller 8 bp deletion within the G box [71] as well as a polymorphism within the A-tract (A replaced by G) [69]. This allele was first described by Golovnina et al. [71] in a collection of wild diploids (T. boeoticum and T. urartu) and tetraploid (T. araraticum and T. timopheevii) wheat. In addition to Vrn-A1f, Golovnina et al. [71] described two other variants called Vrn-A1g and Vrn-A1h as having 34 bp and 20 bp deletions near the CArG- box, respectively, in addition to the minor deletion of 8 bp in the G box. Among them, the dominant Vrn-A1g allele was reported as extremely rare in both T. monococcum and T. boeoticum [71]. Ivaničová et al. [72] designed a Vrn-A1f-like allele from T. militinae (Zhuk. and Migush.) (2n = 4x = 28, AtGG genome), a wild wheat that originated from a hybridization event separate from emmer wheat and belongs to the T. timopheevii (Zhuk.) group. Comparison between Vrn-A1f-like and Vrn-A1a revealed major mutations in the promoter region [the nonexistence of the Spring fold element (SFE) insertion and two deletions (8 base pairs and 50 base pairs) positioned downstream of the CArG box] but also within the first intron [72]. In spring T. dicoccum, a dominant allele known as Vrn-A1k, characterized by a 42 bp insertion at −108 bp, was reported by Muterko and Salina [73], whereas Vrn-A1j was described in T. compactum as carrying a deletion of 54 bp between −140 and −87 in the promoter [74] (Figure 1; Table S1).

Figure 3. Schematic representation of natural variations identified within Vrn-A1. The figure depicts different polymorphism types identified in the literature not based on ploidy level. Different alleles are represented by different shapes: the star denotes polymorphism, rectangle deletion, and triangle insertions at the promoter, intron and/or exon regions. For each allele, the deposited GenBank code are specified on the right, while the corresponding references are available in Table S1. Haplotypes indicate the presence of different polymorphisms at exon-4 (Ex4C/Ex4T), intron-4 (Ex4C.s, Ex4C.m, Ex4C.f, and Ex4C.sph), and exon-7 (Ex7C/Ex7T) [74].

Figure 3. Schematic representation of natural variations identified within Vrn-A1. The figure depicts different polymorphism types identified in the literature not based on ploidy level. Different alleles are represented by different shapes: the star denotes polymorphism, rectangle deletion, and triangle insertions at the promoter, intron and/or exon regions. For each allele, the deposited GenBank code are specified on the right, while the corresponding references are available in Table S1. Haplotypes indicate the presence of different polymorphisms at exon-4 (Ex4C/Ex4T), intron-4 (Ex4C.s, Ex4C.m, Ex4C.f, and Ex4C.sph), and exon-7 (Ex7C/Ex7T) [74].

Miroshnichenko et al. [75] reported the absence of overlapping between Vrn-A1f, Vrn-A1k, and Vrn-A1j and the majority of known VRN box-related mutations. Recently, Zhang et al. [76] discovered two additional alleles (vrn-A1n and Vrn-A1o) that showed a linked duplication in the promoter region. The common copy was identical to the recessive allele vrn-A1, whereas the additional copy of vrn-A1n contained a 54 bp deletion in the promoter that did not influence flowering, and Vrn-A1o harbored an 11 bp deletion that conferred spring growth [76] (Figure 1; Table S1).

Regarding the Vrn-B1 gene, Chu and colleagues [22] revealed a novel retrotransposon located at ~100 bp upstream of the start codon at the promoter level of the Vrn-B1 allele. This retrotransposon is responsible for conferring the spring growth habit in durum wheat [22]. Later, Muterko et al. [69] recognized novel allelic variants in the promoter regions of Vrn-B1 by investigating 178 tetraploid and hexaploid wheat accessions. The new variants were named Vrn-B1.f, Vrn-B1.s, and Vrn-B1.m for Vrn-B1 and were extensively distributed in both hexaploid and tetraploid wheat. Furthermore, the novel allele Vrn-B1.m, along with Vrn-B1.m.1 and Vrn-B1.m.2, was detected by Muterko and collaborators [62] in T. dicoccum, and characterized by a polymorphism and deletion in the promoter region (Figure 1; Table S1).

As for Vrn-D1, compared with the recessive vrn-D1, the allele named Vrn-D1c harbors a 174 bp insertion in bread wheat [77,78] (Figure 1; Table S1).

Finally, novel allele Vrn-G1 and Vrn-G1.a were detected by Shcherban and collaborators [79] in T. timopheevii, characterized by an insertion of 215 bp at −99 bp in the promoter region (Figure 1; Table S1).

3.2. Allelic Variation of Vrn-1 at Gene Body Level

Regarding the allelic variation at the gene body level, Vrn-A1c [37] and Vrn-A1L [80] alleles were discovered in tetraploid wheat, which were characterized by 5.5 kb and 7.2 kb deletions in the first intron, respectively [80] (Figure 1, Figure 2 and Figure 3; Table S1). Compared to the recessive vrn-A1 allele, Vrn-A1c in hexaploid wheat had eight unique SNPs and five unique 1 bp indels in the first intron [37]. Additionally, an allele called Vrn-A1ins was identified, which possesses a 0.5 kb insertion within intron 1 of the diploid T. monococcum [80]. Furthermore, the vrn-A1u allele was observed, and was characterized by a 1.4 kb deletion within intron 1 of T. urartu and polyploid species with an A-genome [80]. Sehgal et al. [81] and Steinfort et al. [82] described the Vrn-A1f and VRN-A1AUS28709 Ai2 alleles in T. aestivum, respectively, harboring a deletion in intron 1. Furthermore, T. araraticum and T. timopheevii as the tetraploid species of the Timopheevi group are characterized by Vrn-A1f-del (2.7 kb deletion at intron 1 in T. araraticum), Vrn-A1f-ins (0.4 kb insertion at intron 1 in T. timopheevii), and Vrn-A1f-del/ins (0.4 kb insertion and 2.7 kb deletion at intron 1 in T. timopheevii), plus the deletions and the polymorphism in the promoter as described for allele Vrn-A1f [79], while T. militinae possesses an MITE transposon (0.4 kb insertion) and a 2.7 kb deletion in intron 1, and also exhibits a host duplication of nine base pairs in the first intron, and two synonymous SNPs in exon 7 and exon 8 [72]. Intriguingly, a polymorphism in the coding sequence of the recessive allele has been exclusively identified for Vrn-A1 [83,84]. Based on the presence of “C → T” transition within exon 4 at position 20 bp of Vrn-A1, two different haplotypes were initially distinguished (Ex4C, wild type and Ex4T, mutant type). Similarly, the same transition (“C → T”) which led to the substitution of alanine for valine (Ala180/Val180) within exon 7 was observed [83]. Muterko and Salina [74] reported then a survey of exon 4 haplotypes in 12 tetraploid and hexaploid wheat species. The authors found that the Ex4T haplotype was present only in the hexaploid wheat vrn-A1 allele, and exclusively in combination with the Ex4C haplotype in accessions of hexaploid wheat carrying Vrn-A1 multi-copies. In addition, to denote the Vrn-A1 exon 4 haplotype, Muterko and Salina used the previously available nomenclature [73], further expanding it. Using the abovementioned nomenclature, mutations within intron-4 were used to distinguish four haplotypes (Ex4C.s, Ex4C.m, Ex4C.f, and Ex4C.sph) [73]. The first three were named based on their migration velocity (s: slow, m: middle, f: fast), whereas Ex4C.sph was detected only in T. sphaerococcum. Furthermore, Muterko and Salina [74] identified two polymorphisms in exon 4 and exon 7 on the Vrn-A1j (exon 7) and Vrn-A1k (both exon 4 and 7) alleles.

The dominant alleles of the Vrn-B1 and Vrn-D1 loci exhibit variations from the recessive alleles, mainly characterized by insertions or deletions within the first intron [37,56,85]. The allele Vrn-B1a, identified in 2005 by Fu and colleagues [37], was characterized by a 6850 bp deletion in intron 1, whereas a similar allele called Vrn-B1b (the same 6850 bp deletion of Vrn-B1a plus a 36 bp indel) was described by Santra et al. [86]. Vrn-B1c, discovered by Chu et al. [22] and later by Milec et al. [87], differs from the others by an 817 bp deletion and 432 bp duplication in intron 1. Zhang et al. [88] reported a novel dominant allele, Vrn-B1d, in the Chinese spring Hongchunmai. The allele contained several genetic divergences within intron 1 compared to vrn-B1, including a large 6850 bp deletion (670–7519 bp), one small 187 bp deletion (7851–8037 bp), an SNP (T/C at 7845 bp), and one 4 bp mutation (TTTT to ACAA, 7847–7850 bp). In 2021, Strejčková and colleagues [63] found a novel allele called Vrn-B1f, which was characterized by an 836 bp insertion within intron 1 in bread wheat.

Recently, several mutations in both encoding and non-encoding regions of Vrn-A1 and Vrn-B1 were identified in 95 out of 263 wild emmer (T. dicoccoides) wheat genotypes with diverse growth habits and flowering times collected in the Fertile Crescent [89]. In 2023, Strejčková et al. [89] identified 15 and 7 SNPs in exons of Vrn-A1 and Vrn-B1, respectively, as well as one insertion in exons of Vrn-B1. The Vrn-D1a and Vrn-D1b dominant alleles of spring bread wheat have the same deletion (4235 bp) but they differ from each other by an SNP (C/A at translation site of CArG-box in Vrn-D1b) [37,90].

3.3. Copy Number Variations of Vrn-1

Copy number variation (CNV) can also greatly impact Vrn-1 gene function [65], thus influencing wheat adaptation and flowering time [65,91,92]. In bread wheat, CNV in recessive and dominant Vrn-1 alleles has been reported [65,92,93]. A different number of copies of Vrn-A1 led to different vernalization requirements among winter wheat cultivars [65,91]. The heading date of winter wheat was affected by allelic variation associated with CNV at the Vrn-A1 locus [94]. The earlier flowering after a short vernalization period relates to a low copy number at Vrn-A1 [65]. In other words, the CNV of the Vrn-A1 gene strongly impacts vernalization requirements and late flowering [65]. Zhu et al. [91] recommended that choosing wheat varieties with three copies of the recessive vrn-A1 gene would be a viable method to increase the frost tolerance ability of wheat because of the association between increased Vrn-A1 copy number and greater frost tolerance.

More than 90% of winter varieties of T. aestivum carry two to three copies of the Vrn-A1 gene [92]. Muterko and Salina [93] represented the copy number of Vrn-A1 with the alternative exon 4 haplotype in spring and winter accessions of tetraploid and hexaploid wheat. Another study reported the duplication of Vrn-A1b.3 in T. dicoccum and the Vrn-A1b.3 and Vrn-A1b.2 in hexaploid T. spelta [95]. Muterko [95] described that duplicated Vrn-A1b.2 was related to the awnless spikes in T. spelta, whereas Würschum et al. [92] found that the geographical patterns of Vrn-A1 copy number variations were compatible with their roles in promoting wheat’s worldwide adaptability.

CNV at the Vrn-B1 locus was also reported by Muterko and Salina [93] in T. compactum (Host) and T. spelta (L.), although Strejčková et al. [63] reported that Vrn-B1 and Vrn-D1 exist in a single copy. By contrast, the authors found that recessive Vrn-A1 has one to four copies, whereas the dominant Vrn-A1 has one or two copies [63].

3.4. Allelic Variation of Vrn-1 at Different Ploidy Levels

On the AA genome, three recessive alleles (vrn-A^m1, vrn-A1u, and vrn-A^m1b) have been identified in diploid species [19,20,71,80] (Figure 4).

The vrn-A^m1 allele was found in all diploid species, and to date, it represents the only variant reported in Triticum sinskajae A. Filat. et Kurk. [19,20,71,80]. By contrast, vrn-A1u, identified in T. urartu Thum. ex Gandil by Golovnina et al. [71], is identical to the recessive vrn-A1 reported in polyploid wheat and differs from vrn-A^m1 for a deletion in the promoter region [71,80]. The vrn-A^m1b allele instead was only detected in accessions of T. monococcum L. [19,67]. Dominant alleles were also identified in diploid wheat (e.g., T. monococcum) [71]. For example, two dominant alleles (Vrn-A^m1f and Vrn-A^m1a Vrn-A1h) were found in T. boeoticum Boiss. and T. monococcum [71,80], whereas, so far, no dominant alleles have been identified in T. urartu [71,80].

In tetraploid species, the recessive allele vrn-A1 was inherited from diploids, presumably from T. urartu, since no differences were observed at the promoter level (Figure 3) [64,80], and to date, three recessive alleles [vrn-A1(vrn-A1u), vrn-A1b.3, vrn-A1b.4] have been described in both Timopheevii A. Filat. et Dorof. and Dicoccoides Flaksb. sections [62,69,80]. As suggested by Konopatskaia et al. [64], dominant alleles such as Vrn-A1a.3, Vrn-A1e, Vrn-A1i and Vrn-A1b might originate through deletion (Vrn-A1b and Vrn-A1e), insertion (Vrn-A1a.3), or substitution (Vrn-A1i) events from the recessive vrn-A1. Interestingly, dominant alleles of Vrn-A1b except Vrn-A1b.7 and Vrn-A1e were distributed only in the dicoccoides section (AABB), suggesting that they evolved from vrn-A1 after the section separation [34,64]. By contrast, Vrn-A1b.7 was found in both the Emmer lines (AABB) and the Timopheevii lines (AAGG), suggesting that they originated from a common tetraploid ancestor [64]. Shcherban and Salina [85] reported that the presence of new dominant Vrn-1 alleles was not related to the origin in diploids, since the allele set found in T. dicoccoides differs from Timopheevii, indicating an independent origin of dominant alleles within these two allopolyploids [85]. In T. timopheevii Zhuk. and T. araraticum Jakubz. have only one dominant allele (Vrn-A1f), which originated from the recessive vrn-A^m1 of T. monococcum, T. urartu, T. boeoticum, and was described at the Vrn-A1 locus [32], whereas ten dominant alleles were identified in different tetraploid wheat species of section Dicoccoides Flaksb. [Vrn-A1a(Vrn-A1a.3), Vrn-A1b(Vrn-A1b.1), Vrn-A1b.2, Vrn-A1b.5, Vrn-A1b.6, Vrn-A1e, Vrn-A1f, Vrn-A1i, and Vrn-A1d] [34,69,71,80]. Vrn-A1a.3 was restricted to T. dicoccum and T. dicoccoides, whereas the dominant Vrn-A1d allele has been found in both Timopheevii A. Filat. et Dorof. and Dicoccoides Flaksb. sections and it probably arises from Vrn-A1b variants due to an extended deletion. Konopatskaia et al. [64] alternatively reported that the two deletions within vrn-A1 could originate from the Vrn-A1d locus [64]. Vrn-A1d probably originated at the tetraploid level, and it was not inherited in hexaploid wheat, as suggested by Konopatskaia et al. [64], even though most of the known dominant Vrn-1 alleles in common hexaploid wheat originated at the tetraploid stage [Vrn-A1a.1, Vrn-A1a.2, Vrn-A1b(Vrn-A1b.1), Vrn-A1b.2, Vrn-A1b.6, Vrn-A1c, and Vrn-A1f] [22,37,77,96].

Figure 4. Chord diagram illustrating the allelic variation of Vrn-1. Colors represent the allelic variations, changing based on their association with different wheat as well as ploidy groups: diploid (2x), tetraploid (4x), and hexaploid (6x). The plot was drawn using the chordDiagram() implemented in the R library circlize (version 0.4.15) [97].

Figure 4. Chord diagram illustrating the allelic variation of Vrn-1. Colors represent the allelic variations, changing based on their association with different wheat as well as ploidy groups: diploid (2x), tetraploid (4x), and hexaploid (6x). The plot was drawn using the chordDiagram() implemented in the R library circlize (version 0.4.15) [97].

In hexaploid wheat, before the identification of vrn-A1b.3 in T. vavilovii (Thum.) Jakubz. and T. spelta L. by Muterko et al. [62,69], vrn-A1 was the only recessive allele identified [36,37,62].

In tetraploid wheat, four dominant alleles at the Vrn-B1 locus were described [62,69], each characterized by mutations within the promoter region (such as insertion of repeated elements or short deletions) [22,69,71,98].

Vrn-B1a is the only dominant allele identified in the dicoccoides section and durum accessions [37,69,71], whereas Vrn-B1c probably originated from Vrn-B1a due to an additional deletion of 0.8 kb and a duplication of 0.4 kb [62]. Also, the Vrn-B1b allele appears to have originated from Vrn-B1a, since along with a deletion in the first intron, it also harbors a 36 bp deletion plus an additional SNP [86]. This allele was described in common wheat originating from North America and was associated with the spring growth habit [96]. The Vrn-B1dic promoter differs from vrn-B1 for 29 nucleotide substitutions, one deletion, and one SNP insertion in the region spanning −220 to −155 bp upstream of the start codon, and it was found only in a genotype belonging to T. dicoccoides [64].

Shcherban et al. [85] identified one accession of T. turanicum Jakubz. (AABB) with the Vrn-B1a allele that does not correspond to the dominant Vrn-B1a for an insertion in the promoter [71]. Interestingly, the insertion was homologous to that identified in the Vrn-A1a allele, although the position was different (−100 from the start codon).

The dominant Vrn-D1a allele was found in the near-isogenic line TDE and it abounded in spring wheat adapted to tropical and subtropical regions [99,100]. Vrn-D1b arises from Vrn-D1a due to SNP in the CArG-box region [90]. The Vrn-D1c allele was found in three out of 205 Chinese wheat cultivars [77]. In the same year, Muterko et al. [101] found the Vrn-D1s allele, which is associated with spring form. Shcherban et al. [85] reported that the distribution of spring forms along with different alleles at Vrn-1 is largely due to artificial selection based on different climatic conditions. For example, dominant haplotypes at the Vrn-A1 and Vrn-B1 loci were observed in cultivars from northern and central Europe and from Russia [85], whereas the monogenic dominant haplotypes contained at either Vrn-B1 or Vrn-D1 were mostly widespread in cultivars for southern Europe [85,102]. Therefore, the monogenic Vrn-B1/Vrn-D1 haplotypes were suggested to gain a breeding advantage for subtropical southern regions, providing a longer vegetative period [85]. By contrast, digenic dominant haplotypes might be useful in regions with a temperate climate [103].

Unfortunately, previous studies focused only on the promoter or gene body levels, whereas most of the other studies covered accessions of a single species or a specific ploidy level. Vrn-A1f is a good example. Golovnina et al. [71] described the Vrn-A1f allele at the promoter level in a collection of four diploid wheat species (T. urartu, T. boeoticum, T. monococcum, and T. sinskajae), seven genotypes belonging to Aegilops speltoides and Ae. squarrosa (syn. Ae. tauschii), and 17 accessions belonging to the sections of Dicoccoides, Triticum, Timopheevii. Later, Sehgal et al. [81] described a 6Kb deletion in intron 1, naming the allele as Vrn-A1f in wheat germplasm composed of landrace accessions, synthetic hexaploids developed at CIMMYT by crossing durum wheat (T. turgidum subsp. durum) or emmer wheat (T. turgidum subsp. dicoccum) with diverse Aegilops tauschii accessions, and elite lines. However, the promoter region was not covered and investigated, thus leaving gaps at both species and gene levels. Similarly, the Vrn-B1c allele reported by Milec et al. [104] and Chu et al. [22] was renamed as Vrn-B1d in the Catalog of Gene Symbols for Wheat [105], but it is unrelated to the Vrn-B1d allele reported by Zhang et al. [88]. In addition, in most cases, the focus was carried out without considering the CNV at the Vrn-A1 gene. For example, roughly 90% of winter hexaploid wheat varieties carry two to three copies of this gene [92], with different haplotypes at exon 4 and exon 7 [65,106,107].

3.5. Allelic Variation of Vrn-2, Vrn3, and Vrn4 Genes

The identification of natural variations in Vrn-2 genes may prove difficult due to the limited characterization of the Vrn-2 gene in hexaploid wheat. Indeed, few natural variations in the promoter and/or in the first intron of Vrn-2 genes (Vrn-A2, Vrn-B2, Vrn-D2, and Vrn-S2) were identified and characterized (Figure 1; Table S2). They were originally observed in diploid wheat (T. monococcum) [25]. Furthermore, a previous development of a tetraploid wheat line lacking functional copies of Vrn-2 has been documented [104]. In addition, various hexaploid wheat cultivars may have undergone multiple events of duplication, deletion, and translocation involving Vrn-B2. Consequently, the task of identifying specific variations becomes challenging [39]. Unlike Vrn-1, Vrn-3, and Vrn-4 genes that are dominant for spring growth habit, Vrn-2 genes are dominant for winter growth habit [25]. Vrn-B2 is generally functional, whereas Vrn-A2 is non-functional in tetraploid wheat [40,108]. Tan and Yan [39] isolated Vrn-2 from hexaploid winter wheat cultivars Jagger and 2174, reporting no differences at Vrn-A2 or Vrn-D2, while two copies of Vrn-B2 were found in 2174, indicating that Jagger carried a null allele. The first copy (Vrn-B2a.1) was 2327 bp long and had a 2087 bp insertion between the start and stop codon plus a 144 bp insertion before the start codon, and a 96 bp insertion after the stop codon, whereas Vrn-B2a.2 had an extra ‘CAC’ motif at positions 136–138 from the start codon and five SNPs compared with Vrn-B2a.1 [39]. The cloned Vrn-D2 was 2364 bp in length, where 239 bp corresponded to an insertion before the start codon and 96 bp to an insertion after the stop codon [39]. Distelfeld et al. [108] reported Vrn-S2 in Ae. speltoides and Vrn-D2 in Ae. tauschii, concluding that the winter growth habit of most of the Ae. speltoides and Ae. tauschii accessions was probably due to functional Vrn-2. The ZCCT1 and ZCCT2 proteins from both species showed no mutations in the conserved amino acids of the CCT domains [108].

Several natural variations were also detected and characterized in the promoter and/or in the first intron of Vrn-3 (Vrn-A3, Vrn-B3, and Vrn-D3) (Figure 1; Table S3). Recently, Nishimura et al. [109] reported in wild emmer wheat six Vrn-A3 alleles with the 7- and 25 bp insertions in the promoter region, namely, Vrn-A3a-h2, Vrn-A3a-h3, Vrn-A3a-h4, Vrn-A3a-h5, Vrn-A3a-h6, and Vrn-A3c-h2. Similar insertions (i.e., Vrn-A3a-h2 and Vrn-A3c-h1) were also found in cultivated tetraploid and hexaploid wheat [109]. Yan et al. [26] identified the vrn-A^m3 allele in T. monococcum, which is characterized by a polymorphism in the promoter region. The Vrn-B3 locus in tetraploid and hexaploid wheat is defined by five dominant alleles, all linked to modifications in the promoter. Yan et al. [26] identified the Vrn-B3a allele characterized by the insertion of 5300 bp in the promoter region. Later, Chen et al. [84] showed two novel alleles: Vrn-B3b, with an insertion of 890 bp in the promoter, and Vrn-B3c, characterized by two deletions (20 bp and a 4 bp) in the promoter of Vrn-B3a. Berezhnaya et al. [110] discovered two novel allelic variants of the Vrn-B3 gene in common wheat varieties from Russia. These alleles were designated the Vrn-B3d and Vrn-B3e alleles and had 1615 bp and 160 bp insertions in the promoter, respectively [110]. Among the alleles described for Vrn-3, Muterko et al. [62] reported a high frequency of Vrn-B3a in T. durum varieties from Ukraine, Russia, and Kazakhstan. Finally, Bonnin et al. [42] demonstrated the presence of polymorphic sites within four haplotypes in the A genome (TAFTAh1, TAFTAh2, TaFTAh3, and TAFTAh4), whereas two were identified in the D genome (TAFTDh1 and TAFTDh2), and only one line (BT21) showed a polymorphism in the B genome (TaFTBBT21) of Vrn-3. All five affected sites (three SNPs and two deletions) were found within the first intron [43]. Additionally, a single polymorphism for genome D was observed, consisting of an INDEL of one G in the third exon [43].

Vrn-4 is an early flowering allele and is comparatively less comprehended in comparison to the preceding three vernalization genes. The Australian cultivar Gabo was the first to identify Vrn-4 [59,111], and subsequently, it was backcrossed into Triple Dirk to produce an isogenic line called TDF [111]. This locus was assigned to chromosome 5D [112] and is now recognized as Vrn-D4 [113]. Although only the D genome has been identified thus far as having the natural variation for flowering time in the centromeric region of homologous group 5 chromosomes, the arm position of Vrn-D4 in wheat is yet unclear [44]. The Vrn-D4 locus might play a crucial role in the variation in flowering time in hexaploid wheat germplasm, and it seems to have undergone independent evolution from the vernalization pathway in dicot species [25].

4. Conclusions and Future Perspectives

Phenology is a key target for adapting wheat varieties to ongoing climate changes and Vrn genes play a crucial role in accelerating reproductive development after prolonged exposure to low temperatures during the winter [85,114,115,116,117]. Matching the development phases and growth of the wheat crop to the thermal regime of the growth environment allows us to maximize the exploitation of natural resources by limiting climatic stress.

This review highlights the effort of the scientific community to understand the molecular mechanisms underlying wheat vernalization in the last two decades. Molecular advances have not only unveiled a rich reservoir of allelic variability within Vrn-1 genes, impacting critical developmental stages, but they have also brought to light the complexity of the allelic landscape. However, the quest for a comprehensive understanding remains a captivating puzzle.

Recent advances in sequencing technologies can now help to untangle vernalization, covering all the gene space of Vrn-1 in different species at one time. Precise genotyping of Vrn-1 genes, by assigning weights to each Vrn allele, might not only improve the accuracy of genomic prediction models but also streamline the varietal selection process [118,119,120,121], identifying the best Vrn allelic combination. This information could also be used to feed phenological predictive models to rapidly identify genetic material with the optimum phenology within different target regions and/or untested advanced breeding lines early in a breeding program. Additionally, the possibility of inducing site-specific mutations in the genetic sequences of these genes offers new perspectives for developing resilient wheat varieties.