Identification of a novel, dominant dwarfing gene (Ddw4) and its effect on morphological traits of rye

Shortening rye stems to improve lodging resistance is among the major tasks awaiting breeders of this cereal. The most straightforward way to achieve this goal is the implementation of a dominant dwarfing gene into high yielding cultivars. The choice of dominant dwarfing genes in rye is limited to Ddw1 and Ddw3 loci, which are well characterized with respect to map position and tightly linked molecular markers on the long arms of chromosomes 5RL and 1RL, respectively. This paper reports on the identification and preliminary characterization of a novel dominant dwarfing gene, Ddw4, from line S44. This was mapped within the centromeric region of chromosome 3R. The Ddw4 gene is sensitive to exogenous gibberellin. Its introduction into the rye populational cultivar Dańkowskie Amber decreased plant height by c. 54% without any negative effects on spike length and number of kernels per spike. Further genetic studies are needed to determine the perspectives for application of the newly detected dwarfing gene into breeding programs for short-stem rye.


Introduction
In the majority of the studied mapping populations, plant height shows quantitative variation controlled by complex networks of interacting quantitative loci (QTL), as revealed in rice [1], wheat [2], barley [3][4], and rye [5][6][7]. Because of the high number and small and/or unstable effects, QTL are usually ineffective tools for practical breeding. More interesting for practical purposes are rare monogenic mutations causing radical growth reduction that have been identified among genetic stocks. Dominant dwarfing genes-Rht-B1b (Rht1), Rht-D1b (Rht2) (located on short arms of chromosomes 4B and 4D), and Rht8 (2DS) have been introduced into wheat cultivars to reduce plant height [8][9]. A similar strategy for breeding short-straw cultivars of barley is based on the semidwarf gene denso (sdw1) located on chromosome 3HL [10][11]. Semidwarf sd-1 gene from chromosome 1, introduced in the course of "green revolution" into rice cultivars, is a defective GA20 oxidase allele [12]. Recently, sdw1 and sd-1 semidwarf genes were established as being orthologous [13]. Rye cultivars still represent a tall-straw phenotype, which makes them sensitive to lodging. So far, attempts to identify dominant dwarfing genes in this species have resulted in the characterization of the Ddw1 gene on Editor: Aimin Zhang, Institute of Genetics and Developmental Biology Chinese Academy of Sciences, CHINA chromosome 5RL [14][15], the Ddw2 gene on chromosome 7R [16][17] and the Ddw3 gene on chromosome 1RL [18]. Only Ddw1 and Ddw3 genes are more closely characterized and tagged by molecular markers. This paper reports on the identification of a new dominant dwarfing gene, Ddw4, in rye and presents its basic characteristics in view of its possible application in breeding.

Plant material
A dwarf line was selected from within the S44 population of rye (Secale cereale L.) lodged in the National Center for Plant Genetic Resources: Polish Genebank IHAR, accession number 31147. The population was internally differentiated in respect to plant height and contained high, medium and short plants. Plants of c. 80 cm height were self-pollinated in two consecutive years, which led to the development of a dwarf inbred line denoted line S44. A second line, 541, representing tall plants, originated from the Department of Genetics, Plant Breeding and Biotechnology, West Pomeranian University of Technology in Szczecin, Poland [19]. Dwarf line S44 is 2 days earlier than tall line 541 in respect to heading date and flowering time, while both lines are 5-7 days later than cv. Dańkowskie Amber. The F 1 (541 × S44) cross was developed in 2012. The F 2 progeny was obtained through self-pollination of F 1 plants. Each F 2 plant grown in 2013 was self-pollinated giving seeds of F 3 progeny. Recombinant inbred lines (RILs) were further developed using the single seed descent (SSD) method. F 2 plants were phenotyped as short or tall by visual observation and height measurement. F 2 plant phenotype was further confirmed by observation of the F 3 progeny (7-20 plants per line). The hypothesis of the monogenic inheritance of plant height in F 2 and F 3 progenies was verified using the χ 2 test. F 2 plants were also analyzed in respect to: length and thickness of the second internode from the base, spike length, grain number and weight per spike. In order to obtain a short-straw version of the highly yielding cultivar Dańkowskie Amber, and F 1 progeny of the cross Dańkowskie Amber × S44, was developed. A preliminary characterization of these newly developed plant materials was conducted in respect to morphological traits, including thousand-kernel weight (TKW). All field experiments were performed at the Experimental Station of the West Pomeranian University of Technology in Szczecin, Poland. The significance of the differences between genotypes was tested using a t-student test. Correlation coefficient was calculated on average data. Computations were performed using the Statistica 13.1 software (Dell Inc.).

Genetic mapping
The F 2 mapping population for the 541 × S44 cross consisted of 87 lines, including 64 shortstraw and 23 tall plants drawn from the entire population of 519 F 2 plants based on phenotypic performance within F 3 progenies consisting of at least 15 mature plants. DNA was isolated from each of the 89 F 2 and parental lines in early spring using a DNeasy Plant Mini Kit (Sigma), according to the protocol delivered by the manufacturer. Genotyping by sequencing (GBS) was performed using DArTseq technology (Diversity Arrays Technology Pty Ltd, Australia), according to Li et al. [20]. Markers generated "in silico" and in single nucleotide polymorphisms (SNP) versions were taken for constructing linkage groups. DArTseq markers were assigned to rye chromosomes by applying the reference map of the 541 × 2020 cross [21]. Mapping of the dwarfing gene was conducted as the final step of map construction. A computer package, JoinMap 5.0 [22], was used for linkage analysis. Linked markers were grouped at LOD = 18. An original map drawing the relationships to the reference map was generated using MapChart 2.2 software [23].

Test for sensitivity to exogenous GA 3
Seedlings of the two parental lines 541 and S44 were subjected to exogenous GA 3 treatment according to Worland [8]. Kernels distributed on a water paper soaked with distilled water in a Petri dish (φ = 12 cm) were kept for 48 hours at 4˚C to synchronize their germination. Each line was represented by 10 germinating seeds in 3 replications for distilled water (control) and 5 ppm GA 3 treatments. The lengths of seedlings grown for 14 days (2-3 leaves stage) were measured and the ratio between GA 3 treated and water treated plantlets was determined. A non-parametric Kolmogorov-Smirnov test was used to assess the significance of the reaction to GA 3 (Statistica 13.1).

Genetic analysis of plant height in 541 × S44 crosses
The stability of line S44 across three years in respect to the short-straw phenotype is shown in Table 1. The mean height of this line ranged from 71.3 cm in 2015 to 84.8 cm in 2014. The threeyear mean height was 77.2 cm, which constituted 57% of that found for line 541 grown in the same environment. The height of the F 1 generation was 85 cm and this result suggested complete dominance of the dwarfing gene over the respective allele from line 541. The range of plant height (40.0-190.5 cm) found in the F 2 generation consisting of 518 plants (Fig 1) exceeded that of parental lines, which suggests environmental or genetic background effects, including transgression. The plant height distribution in F 2 progeny was bimodal, showing that this generation consists of two subpopulations: a more numerous short-straw one (40-120 cm) with a mean value of 81.6 cm; and a minor one, representing tall plants (130-190 cm) with a mean value of 156 cm. The lowest height attributed to tall plant phenotypes was 130 cm, since from this value no dwarf plants were observed in F 3 progeny. Analysis of F 3 progenies allowed the differentiation of homozygotes from heterozygotes within the dwarfing gene locus. The average height of dwarf homozygotes in the F 2 generation was 72.8 cm, and that for heterozygotes was 87.7 cm, which suggests a high degree of dominance for the dwarfing gene. The segregation ratio of short-straw vs. tall phenotypes in F 2 confirmed by observation of F 3 progenies was consistent with the 3:1 ratio, suggesting the monogenic inheritance of plant height in the 541 × S44 cross ( Table 2).

Mapping of the dwarfing Ddw4 gene in rye
Out of 77 573 DArTseq markers, 17 317 were found to segregate in a Mendelian mode (3:1 segregation ratio) within the F 2 mapping population of the 541 × S44 cross. Linkage groups  (Fig 2, S1 Table). The high-density mapping of Ddw4 was achieved using 595 markers located within a distance of c. 70 cM to the gene locus, and these were collected with the help of a chromosome 3R reference map. The gene is located within a cluster of 18 markers (linkage < 1 cM) which might be used for its tagging and performing marker assisted selection. Since the map position of the dwarfing gene is here reported for the first time, the novel gene was assigned the symbol Ddw4 showing that it is the fourth dominant dwarfing gene found in rye.

Sensitivity of Ddw4 to GA 3 treatment
The S44 line showed a high degree of sensitivity to exogenous GA 3 at the seedling stage (Table 3, Fig 3). The response (177% increase in seedling length) was much higher than that shown by line 541 expected to be highly sensitive to GA 3 . Thus, it may be concluded that the Ddw4 gene belongs to the group of GA 3 sensitive dwarfing genes.

Assessment of Ddw4 effects on other agronomic traits of rye
The conspicuous difference between parental lines S44 and 541 in respect to plant height is associated with the significant difference between the second internode and spike length (Table 1). However, three other traits of agronomic value, i.e. the 2nd internode thickness, kernel number and kernel weight per spike, seem to be unrelated to plant height. Positive significant correlation coefficients between plant height and kernel number per spike (r = 0.32), plant height and kernel weight per spike (r = 0.56) as well as those between kernel number per spike and kernel weight per spike (r = 0.72) found in the F 2 progeny suggest that these two traits might be negatively affected by the Ddw4 dwarfing gene (Table 4). On the other hand, finding that plant height and spike length are not correlated within the F 2 progeny may provide positive perspectives for application of Ddw4 gene in breeding.
Preliminary results showing Ddw4 effects on agronomic values of cv. Dańkowskie Amber are shown in Table 5. The dwarfing gene is apparently dominant in the F 1 generation as it reduces the height of cv. Dańkowskie Amber by c. 54%. This strong positive effect on plant height is not associated with negative effects on thickness of the second internode, spike length, kernel number or kernel weight per spike. Unfortunately, the Ddw4 gene negatively affects the thousand-kernel weight of this cultivar by c. 31%.

Discussion
Dominant dwarfing genes have not previously been reported on chromosome 3R. There is one early paper by De Vries and Sybenga [24] attributing recessive dwarfing gene dw3 to this chromosome [17]. Since the map position of dw3 was not established it is difficult to decide on its relationship to Ddw4. The newly reported Ddw4 gene was precisely localized on the high-density molecular map of chromosome 3R in the centromeric region on the side of 3RL chromosome arm. This position corresponds to major QTL of the D class for plant height, revealed by Masojć et al. [7] using a 541 × Ot1-3 mapping population of rye and a high-density map of DArT markers. The latter study shows that the height of the 541 line is controlled by a complex system of interacting QTL distributed on each of rye chromosomes, which means that the Ddw4 mutation should represent a defective allele in a key locus. Because of the complexity of QTL systems for plant height in rye, some background effects caused by other possible QTL within the 541 × S44 cross might be expected and indeed they can explain a wide range of the variation found in the F 2 generation. A centromeric region on barley chromosome 3HL has been found to contain the recessive dwarfing gene uzu [25]. Due to the high level of synteny between 3H and 3R chromosomes, Ddw4 is possibly orthologous to its counterpart from barley.
The application of Ddw4 in breeding short-straw, lodging resistant rye varieties seems promising in view of its performance in the F 1 cross between cv. Dańkowskie Amber and S44 line. The achieved short-stature of rye plants without shortening of spikes is satisfying. However, the significant reduction in thousand-kernel weight observed in this cross needs further explanation. This negative effect might result from the genetic background inferred by line S44 or by Ddw4 gene itself. Backcrossing with cultivar components and selection for dwarfing alleles should clarify this problem. The chromosome 3RL region containing the Ddw4 locus is also a QTL localization place for thousand-kernel weight, kernel length, kernel thickness and kernel number per spike [5]. Close linkage between these QTL and the dwarfing gene may  result in association of grain parameters with plant height, which means that allele compositions in these loci in a short-straw donor line is crucial for the agronomic performance of developed progenies. The considered chromosome 3R region was shown to contain QTL controlling benzoxazinoid content and leaf rust resistance [26] which in turn raises the question about the response of Ddw4 to brassinosteroids and its association with disease resistance. Ddw4 may also be associated with preharvest sprouting resistance (PHS) and alpha-amylase activity in the grain (AA), as it is tightly linked with QTL for these traits [27][28][29][30] and with GA 3 oxidase [31], a candidate gene for PHS. The strong response from the S44 line to exogenous GA 3 positions the Ddw4 gene in the same group of GA-sensitive dwarfing genes as Ddw1 and Ddw3 [17][18]. Therefore, it can be Novel, dominant dwarfing gene in rye concluded that Ddw4 does not represent GA signaling genes. Neither of these dwarfing genes is an ideal candidate for implementation of a short-stem phenotype into rye cultivars due to the observed negative effects on yield-related traits underlined both by Stojałowski et al. [18] and in this study. So far, attempts to reduce plant height in rye have been made with the use of the Ddw1 gene, especially in Finland [15]. With three dominant dwarfing genes identified and to be tagged by molecular markers in the near future, a wide-ranging search for the most valuable genetic backgrounds eliminating undesirable effects should be initiated.
Supporting information S1