Pre-harvest sprouting resistance and haplotype variation of ThVp-1 gene in the collection of wheat-wheatgrass hybrids

The germplasm collection of 87 wheat-wheatgrass hybrids developed in Tsitisin Main Botanical Garden (Russia, Moscow) was evaluated for resistance to pre-harvest sprouting (PHS) by spike sprouting (SS) and germination index (GI) assays as well as for spike and grain features. The PHS resistance variation and haplotype polymorphism of the wheatgrass ThVp-1 and wheat TaVp-1B genes orthologues of Vp-1 was revealed in the studied collection. Four haplotypes of ThVp-1 were revealed: ThVp-1a (41% of the entries), ThVp-1b (13%), ThVp-1c (29%), and ThVp-1d (15%). The association between the allelic state of ThVp-1 and PHS resistance in the wheat-wheatgrass hybrids was shown: haplotype ThVp-1d of the wheatgrass Vp-1 gene is significantly associated with reduced PHS in the wheat-wheatgrass hybrids (mean SS 0.33, mean GI 0.64). The resistant entries may be perspective as a source of PHS resistance in the development of commercial cultivars of perennial wheat.

The wheat-wheatgrass hybrids were developed in many countries, including Russia, USA, Germany, Canada and China. N. V. Tsitsin was the first who crossed wheat and wheatgrass (Thinopyrum sp.) in order to obtain perennial wheat in the 1920s and whose cultivars of partial The Vp-1 genes of wheat, TaVp-1, are localized on the long arms of the homeologous group 3 chromosomes. Six alleles of TaVp-1B were identified and designated as TaVp-1Ba, TaVp-1Bb, TaVp-1Bc, TaVp-1Bd, TaVp-1Be, and TaVp-1Bf [28,[55][56][57]. The alleles TaVp-1Bb and TaVp-1Bc detected by an STS molecular marker are associated with higher seed dormancy and PHS resistance [28]. For the TaVp-1A gene, also some allelic variations associated with PHS resistance were detected [58]. In general, allelic variants of Vp-1 affect PHS resistance in whitegrained wheat. However, the effect of the allelic state of the Vp-1 gene on PHS resistance was shown in red-grained triticale depending on the intensity of the grain color [59].
At the present time, two molecular markers based on the sequences of the orthologous Vp-1 genes of Thinopyrum intermedium, Th. ponticum, Th. bessarabicum (Savul. & Rayss) and Pseudoroegneria spicata (Pursh) have been developed. These markers allow for the identification of various alleles of the Vp-1 gene orthologue of wheatgrass (ThVp-1) in a wheat background [60,61]. The primers for the markers were designed on the conservative regions while the sequence between them is highly polymorphic and can be distinguished directly by PCR using the STS marker Vivip [61] or with preliminary restriction endonuclease digestion using the CAPS marker Vp1BB4_HaeIII [60]. The effect of ThVp-1 on seed dormancy has never been explored.
Although PHS resistance is important for both the development of commercial cultivars of wheat-wheatgrass hybrids and for wheat breeding using wheat-wheatgrass germplasm, this trait has not been studied yet in wheat-wheatgrass hybrids. The aim of this work is to study the germplasm collection of the wheat-wheatgrass hybrids for PHS resistance using different assays, and to assess the effect of grain color, spike parameters and Vp-1 of wheat and wheatgrass on PHS resistance.  (Table 1). Bread wheat cv. Nota and Th. intermedium accession PI 401200 (Germplasm Research International Network) were used as controls for the TaVp-1 and ThVp-1 genes, respectively, in PCR experiments.

Assay of germination index (GI) and spike sprouting (SS)
The entries were grown in the fields of the Department of Distant Hybridization (Tsitsin Main Botanical Garden, Russian Academy of Sciences) at the Snegiri settlement, Istrinsky district of Moscow region (55˚51'32"N 37˚1'54"E). Germination index was determined according to Walker-Simmons method [62]. The spikes were harvested at full ripeness and threshed manually. 25 seeds in four replicates were placed crease down on moist filter paper in Petri dishes, which were then placed in lit climatic chamber (12 hours day, 12 hours night) at +20˚C. The number of germinated seeds was counted daily and removed after counting during 7 days. The seeds were taken as germinated at radical emergence. Water was added to Petri dishes as needed using a sprayer. The remaining seeds were left to germinate for one month to determine their viability. Seeds that did not sprout within a month were excluded from further calculations. The germination index (GI) was calculated according to the formula: where TG were total grains, n 1 , n 2 ,. . . n 7 were the numbers of seeds germinated on the first, second, and subsequent days until the seventh day. Spike sprouting assay (SS) of the wheat-wheatgrass entries was conducted in moist chamber on the shelves wrapped in plastic film. The spikes were harvested at full ripeness, steeped in distilled water for an hour, five spikes per each entry were bound into sheaves and placed on the shelves in an upright position. The sheaves were placed into a chamber and sprayed three times a day for 5 minutes each time. Visual counting of clearly sprouted grains was performed at the third day of sprouting provocation. Then, the spikes were dried and threshed to determine the total number of grains in them. The SS value was calculated as a ratio of sprouted grains to the total number of grains.
In addition, the spike morphology was estimated-presence of awns, glume color, grain color, and threshability.

DNA extraction and molecular markers
The DNA of the wheat-wheatgrass hybrids was extracted from seedlings using CTAB protocol [63].
The PCR analysis was performed using markers Vp1BB4_HaeIII and Vivip for the wheatgrass ThVp-1 genes and Vp-1B3 for the wheat TaVp-1B gene.
The The PCR fragments were separated on 2% agarose gel with TBE buffer at 6 V/cm field strength with GeneRuler 100 bp DNA Ladder (Thermo Fisher Scientific) as a size marker, stained with ethidium bromide, and visualized using UV light.

Statistical analysis
The grouping of the data was carried out using Sturges' rule [64,65]. Analysis of variance (ANOVA) was performed in the 'Statistica 6.0' program. The differences in GI and SS among genotypes with different PCR profiles were tested using Fisher's least significant difference (LSD) test as in Chang

Results
PHS resistance is determined by two groups of traits: spike morphology (presence of awns, glume adherence, spike inclination, etc.) and grain characteristics (color, germination inhibitors in seed coat, hormonal status, etc.) [15,16,62,[67][68][69][70][71]. The effect of the first group can be reflected by evaluating spike sprouting (SS) in intact spikes in moist chamber, and the second group by germination index (GI) of the threshed grains in Petri dishes.

Spike sprouting assay (SS)
For a set of 74 entries spike sprouting (SS) was estimated (Fig 1, Table 1). The proportion of germinated seeds was determined visually on the third day of spikes wetting.

Germination index (GI)
For a set 87 entries the germination index (GI) was evaluated. Using the Sturges' rule, the set was divided into 8 clusters with the corresponding GI ranges ( Table 3).
The entries of different clusters are characterized by different dynamics of germination ( Fig  2). The immediate germination of the majority of grains on the 1st day is represented by 1761 (cluster 8). ZP26 (cluster 7) and 4082 (cluster 6) showed the peak on the 2nd day, 1777 (cluster 5) and 1876 (cluster 4) on the 3rd day, 1735 (cluster 3) on the 5th day. Extended dynamics of germination is intrinsic for 1765-k, the most PHS resistant entry. Therefore, the lower GI value, the more delayed and faint peak of germination can be observed.
Entry 1451 showed both minimum SS value (0.01) and a high GI value (0.87). As 1451 is extremely hard to thresh and has rigid glumes adherent to the caryopsis, its high PHS resistance in the spike is determined primarily by the spike architecture, and the seeds themselves have a low dormancy level. The spike architecture, apparently, also explains PHS resistance in 1748 and 1874 (SS 0.07 and 0.09), which possess low seed dormancy (GI 0.72 and 0.86). 1765-k has low SS (0.02) and low GI (0.15), hence its PHS resistance can be explained equally by the seed and spike characteristics. 1735, 1774 and 150-k showed high resistance in the SS and GI assays, which allows us to consider them as generally resistant to PHS. 4044, 1416-o, 1792, 1803 showed high SS and GI, and thus can be considered as susceptible to PHS. Therefore, the spike features of these entries are unable to provide PHS resistance and to compensate for the low seed dormancy. The entries with low GI and high SS were not revealed (Table 4).

Distribution of wheatgrass ThVp-1 haplotypes
Markers Vp1BB4_HaeIII and Vivip allow the identification of various alleles of the wheatgrass ThVp-1 gene in a wheat background [60,61]. The DNA fragments of Vp1BB4_HaeIII produced by PCR and subsequent restriction endonuclease digestion were analyzed in all entries. Vp1BB4_HaeIII enables the identification of additional fragments different from wheat after restriction of the PCR product. Such additional bands were interpreted as fragments amplified from ThVp-1 [60]. In our study, the analysis of 87 entries using Vp1BB4_HaeIII revealed two sizes of additional fragments of ThVp-1, namely, approximately 600 bp and approximately 550 bp (Fig 3,  Table 5). Consequently, Vp1BB4_HaeIII can distinguish three possible ThVp-1 states: presence of the~600 bp fragment, presence of the~550 bp fragment and absence of both fragments. No entry with simultaneous presence of both the~550 and~600 bp fragments was found.
The Vivip marker was designed based on sequences of another polymorphic regions of the orthologous Vp-1 genes of Th. intermedium, Th. ponticum, Th. bessarabicum, and P. spicata. The results of the identification of the ThVp-1 polymorphism using Vivip are independent from Vp1BB4_HaeIII [61]. In our study, the PCR analysis of 87 entries using Vivip allowed the identification of two additional amplified DNA fragments of wheatgrass origin, namely, approximately 350 bp, and approximately 370 bp (Fig 4). In the studied collection of the wheat-wheatgrass hybrids three possible states of Vivip were revealed: absence of both fragments, presence of the~350 bp fragment, and presence of the~370 bp fragment (Fig 4,  Table 5).  The analysis of the results of two molecular markers have revealed that the 370 bp fragment detected by Vivip occurs in the same wheat-wheatgrass entries as the 550 bp fragment, which is detected by Vp1BB4_HaeIII. Thus, fragments~550 bp of Vp1BB4_HaeIII and~370 bp of Vivip detect the same allele of the wheatgrass ThVp-1 gene.
1795-slo and 4056 showed heterogeneity for the allelic state of ThVp-1, so, they were excluded from the further statistical analysis.
The analysis of 87 entries using both markers Vp1BB4_HaeIII and Vivip showed four possible haplotypes of the wheatgrass ThVp-1 gene that are present in the studied wheat-wheatgrass hybrids (Table 5). These four variants are characterized by the molecular markers, as follows: 1) Absence of the specific wheatgrass DNA fragments of both markers (designated ThVp-1a, 36 entries); 2) Presence of the~600 bp fragment detected by Vp1BB4_HaeIII and absence of specific fragments of Vivip (desisgnated ThVp-1b, 11 entries); 3) Presence of the~550 bp fragment of Vp1BB4_HaeIII and presence of the~370 bp fragment Vivip (desisgnated ThVp-1c, 25 entries); 4) Absence of specific fragments for Vp1BB4_HaeIII and presence of the~350 bp fragment revealed by Vivip (desisgnated ThVp-1d, 13 entries).
Thus, we have identified a polymorphism of the wheatgrass ThVp-1 gene in the wheatwheatgrass hybrid collection and they were grouped into 4 classes (haplotypes) depending on  the results of the combined use of Vp1BB4_HaeIII and Vivip. The distribution of haplotypes in the collection is as follows: the predominant haplotype is ThVp-1a (41%), second place is allele ThVp-1c (29%), followed by haplotypes ThVp-1d (15%) and ThVp-1b (13%). The association between ThVp-1 haplotypes and PHS resistance based on GI in the entire germplasm collection was evaluated statistically using the analysis of variance (ANOVA) ( Table 6).
The ANOVA indicated significant GI differences between the groups of wheat-wheatgrass hybrids with various ThVp-1 haplotypes. Statistically significant differences were observed between haplotypes 'a' and 'd' and between haplotypes 'c' and 'd'. Thus, entries with the ThVp-1d haplotype had significantly lower average GI than ThVp-1a and ThVp-1c genotypes confirming the association of ThVp-1 and seed dormancy.
The ANOVA conducted after arcsin p x transformation of the SS values and Fisher's F-test revealed no significant differences between the groups of wheat-wheatgrass hybrids with different ThVp-1 haplotypes. However, LSD test showed that the group of wheat-wheatgrass hybrids with ThVp-1d haplotype gave lower value of SS than groups with haplotypes 'b' and 'c' ( Table 7).

Distribution of wheat TaVp-1 haplotypes
As wheat-wheatgrass hybrids combine wheat and wheatgrass genomes, our collection was also evaluated for the allelic state of the wheat TaVp-1 gene using the STS marker developed by Yang et al. (2007) [28] ( Table 1). The structure of the studied germplasm collection was as follows: 51% of the entries carried allele TaVp-1Ba, 31% carried allele TaVp-1Bc, 15% showed heterogeneity, 3% were heterozygotes (the latter two groups were not involved in the statistical analysis).
The GI differences between groups of wheat-wheatgrass hybrids with different allelic state of wheat TaVp-1B were not significant with similar mean GI values (Table 8).
No significant differences of SS between groups of wheat-wheatgrass hybrids with different TaVp-1B alleles were revealed (Table 9).  In addition, no statistically significant association was revealed between PHS resistance measured by SS and GI and the presence of awns, glumes color, and grain color (data not shown).

Discussion
In our work for the first time, the PHS resistance polymorphism in wheat-wheatgrass hybrids was revealed by GI and SS evaluation. In our study, the wheat-wheatgrass hybrids significantly differ from each other in PHS resistance. Similarly, wheat-wheatgrass hybrids are also highly polymorphic in other traits such as diseases resistance, grain yield, post-harvest regrowth ability, forage biomass [4,5,[72][73][74]. Among the resistant entries with low GI we identified two types of germination dynamics, namely, an extended germination and delayed peak. In general, the majority of the entries were susceptible to PHS (92% are susceptible to PHS, classes 5-8, Table 3). It may be a serious problem for the introduction of wheat-wheatgrass hybrids as a cereal crop in regions where climate favors PHS. Currently, the main attempts to develop perennial wheat-wheatgrass hybrids as a crop are being made in Australia [4,5,72], where PHS is not a relevant problem and probably that is why it receives little attention.
PHS resistance in grasses is controlled by multiple factors. QTLs for PHS resistance have been found on all chromosomes of wheat in different studies [23, 29, 34-36, [75][76][77][78][79][80][81][82][83][84][85][86][87][88][89][90]. Many studies of the PHS resistance in wheat is focused on TaVp-1, the polymorphism of its haplotypes and its effect on PHS [28, 57-59, 66, 91-94]. In the majority of them the influence of different alleles of TaVp-1B and TaVp-1A was shown primarily in white-grained wheat, though in some studies the effect of TaVp-1 was found in red-grained wheat and triticale as well [59,93]. Given the fact that all wheat-wheatgrass entries possess red or blue grain, which is usually associated with resistance, the absence of resistance in the majority of entries is apparently a result of the negative impact of the wheatgrass genome. For example, a negative effect of chromosome 6Ag i of intermediate wheatgrass on PHS resistance was revealed in the studies of the recombinant inbred substitution lines of wheat [21]. This should be considered in breeding and development of cultivation technology of commercial wheat-wheatgrass cultivars. In addition, one should note, that the proportion of entries demonstrating resistance to sprouting in intact spikes is higher than the proportion of entries demonstrating seed dormancy in the evaluation of GI. Apparently, this is due to the spike architecture of wheat-wheatgrass hybrids. Spikes of wheat-wheatgrass hybrids are generally more dense and hard to thresh in comparison to bread wheat. Four entries (237, 1375, 1654, 1451) differ from others by rough glumes, which adhere very closely to the grain. This feature significantly hardens the process of threshing, and, on the other hand, contributes to the delay of water uptake [40,95]. Apparently, the variability of PHS resistance of wheat-wheatgrass hybrids in our case is mainly due to diversity in the wheatgrass genetic component. We found only two alleles of wheat TaVp-1, namely TaVp-1Bc and TaVp-1Ba, the most common for different wheat cultivars [28, 57-59, 66, 91-94]. We did not reveal any significant differences in PHS resistance between wheat-wheatgrass entries with wheat gene alleles TaVp-1Bc and TaVp-1Ba. As relatively few wheat cultivars were used for the development of the studied wheat-wheatgrass entries, we can hardly expect a wide genetic variability of wheat genes controlling PHS in a wheat-wheatgrass hybrid genome [6,96]. On the other hand, the probability of high polymorphism of wheatgrass genes affecting PHS in different wheat-wheatgrass hybrids is strong, because most of the partial amphiploids carry a synthetic mixed extra genome which has a different combination of chromosomes each time a partial amphiploid is produced [1][2][3][4][5].
Wheat-wheatgrass hybrids carry individual composition of wheatgrass chromosomes, inherited from Th. intermedium and/or Th. ponticum and derived from different subgenomes [1][2][3][4][5]. We may suggest that the recombination between different homeological chromosomes of wheatgrass in the genome of a wheat-wheatgrass hybrid is a rare event. Despite the PCRbased markers Vp1BB4_HaeIII and Vivip applied in our study were developed on the basis of the Vivparous-1 sequence, we may have estimated not only the effect of ThVp-1 but also the effect of the linkage group 3 of Thinopyrum in the present study. . PHS resistance can also be regulated not only by polymorphism in DNA sequence but also by epigenetic factors through DNA and histone methylation [46, 47] that have not been studied in wide hybrids yet and can have its own specificity. As PHS resistance is a complex trait regulated by multiple factors molecular makers Vp1BB4_HaeIII and Vivip cannot be used as the only diagnostic tool in marker-assisted selection of wheat-wheatgrass hybrids or perennial wheat. However, they can be efficiently applied in breeding process after additional exploration and validation in complex with other molecular and phenotypic estimation methods to select PHS resistant wheat-wheatgrass entries and perennial wheat.
We have shown that the studied entries are diverse in wheatgrass component and that this diversity is associated with PHS resistance. The differences in PHS resistance between wheatwheatgrass hybrids can be explained by genetic variations and/or epigenetic modifications. Genetic variations may represent variations in different levels: different combinations of Table 9. Association between TaVp-1B alleles (of wheat origin) and SS values in the wheat-wheatgrass hybrid collection. wheatgrass chromosomes (the genome level), differences in the chromosomes of the homeologous group 3, where genes Vp-1 and R are localized (the chromosome level) or structural and functional differences in the Vp-1 genes (the gene level). Along with that, we have shown that the greatest resistance to PHS is innate to the entries of wheat-wheatgrass hybrids bearing haplotype ThVp-1d of the wheatgrass Vp-1 gene (mean SS is 0.33, and mean GI is 0.64).

Conclusion
We, for the first time, demonstrated the variability of wheat-wheatgrass hybrids in resistance to PHS. Most entries showed susceptibility that may pose a serious problem in wheat-wheatgrass hybrid either as a source of genetic variability in wheat breeding or as a crop in regions with conditions favorable to PHS. Interestingly, although all estimated entries have red grain associated with high seed dormancy, only few showed relatively low GI. Using PCR-based markers we demonstrated the polymorphism of the haplotypes of the wheatgrass ThVp-1 gene in the wheat-wheatgrass hybrid collection and have revealed the significant effect of the ThVp-1 gene on the PHS resistance. The further study may help to shed the light on the role of ThVp-1 gene and its interaction with TaVp-1 in regulation of seed dormancy.