Background

The global human population is predicted to continually expand and reach 10 billion by 2050, significantly increasing the need for the safe and reliable food production. The improvement of crops using advanced technologies provides an effective strategy to meet food production demands in the future [1]. Wheat represents one of the most important staple crops worldwide, and the identification of quantitative trait loci (QTLs) important for agronomical traits, such as plant height (PH) and spikelet compactness (SC), offers critical information to ensure food security [2].

PH is tightly associated with lodging resistance and thus influences grain yield. To date, a total of 25 reduced height genes (Rht1–Rht25) have been documented in wheat [3], and these dwarf genes have been classified into gibberellic acid (GA)-sensitive or -insensitive based on response to GA treatment. The GA-insensitive Rht-B1b (Rht1) and Rht-D1b (Rht2) genes on chromosomes 4B and 4D, which encode truncated DELLA proteins, significantly increase the harvest index, and their use in breeding resulted in the well-known ‘Green Revolution’ [4, 5]. their allelic variations such as Rht-B1c (Rht3) [6], Rht-B1e (Rht11) [7], Rht-B1p (Rht17) [8], and Rht-D1c (Rht10) [9] have been identified. Moreover, several GA-sensitive dwarf genes including Rht8 [10, 11], Rht12 [12], Rht13 [13], Rht18 [14], and Rht24 [15] have been cloned or intensively studied in wheat. Rht12 was located on chromosome 5A and mutations in GA2oxA13 gene produced tall overgrowth phenotype in the Rht12 background [12]. A missense mutation of NB-LRR gene in Rht13 caused height reduction [13]. The dominant Rht18 gene was identified by isolating and sequencing chromosome 6A of overgrowth mutants, and the dwarf phenotype of this mutant was found to be caused by the increased expression of GA2oxA9 resulting in a reduction of active GA content [14]. Map-based cloning suggested that GA2oxA9 was the causal gene of Rht24, which affected GA homeostasis and led to plant height reduction [15].

Since the Rht8 gene does not influence coleoptile length, it well complements Rht-B1b and Rht-D1b weakness, and it has been widely used in wheat breeding for several decades [16, 17]. Rht8 was mapped on the short arm of chromosome 2D, and the SSR marker Xgwm261 was regarded as a perfect diagnostic marker for Rht8 previously [17, 18]. Using two wheat mutants Rht8-2 and Rht8-3 for construction of segregation populations, the Rht8 gene was cloned recently, and it was found to encode an RNase H-like protein that affects bioactive GA content and changes plant height [10]. Similar results were obtained simultaneously in a map-based cloning study using a wheat variety containing the Rht8 gene [11], and both studies also showed that the dwarf allele of Rht8 was positively selected during wheat breeding [10, 11].

Previous analyses of near isogenic lines (NILs) [19] and transgenic plants [11] have suggested that Rht8 not only reduces PH but also significantly decreases spike length (SL) and thus increases SC. The modification of SL or SC plays an important role in the improvement of yield potential in wheat [20, 21], and identification of associated QTLs is critical for wheat improvement. In hexaploid wheat, spike morphology is regulated by three major genes, namely Q, C (Compactum), and S (Sphaerococcum) located on chromosomes 5A, 2D, and 3D, respectively [22,23,24]. These genes exert pleiotropic effects on SC and SL, PH, and grain shape. To date, a large number of QTLs associated with spike morphology have been identified on nearly all wheat chromosomes [25]. In addition, the VRN, Ppd, and Eps genes have also been found to be involved in SL and development and to affect SC in wheat [26,27,28].

In the present study, we dissected the genetic control of PH and SC by performing QTL mapping using a recombinant inbred line (RIL) population derived from a cross between the wheat variety ND5181 and the mutant line Rht8-2 and identified natural variation in the homoeologous gene Rht8 that contributes to PH change for qPH2B.1. The linkage markers and genes influencing PH and SC identified here can be applied to molecular breeding and the promotion of wheat production.

Materials and methods

Plant materials and phenotypic evaluation

We established an F7 RIL population comprising 139 lines derived from a cross between wheat cultivar NongDa5181 (ND5181) and the mutant Rht8-2 using the single seed descent method. The RIL population and the parents were planted in the Zhongpuchang and Changping experimental fields at the Institute of Crop Sciences, Chinese Academy of Agricultural Sciences (39°97′N, 116°34′E; Beijing, China) during the 2021–2022 crop season. Each genotype was planted in one row (15 seeds per row), which was 1.5 m long with a row spacing of 0.3 m. PH and SC (calculated by dividing spikelet number per spike by the SL) were measured at maturity with eight replicates for each line [29].

Genotyping and QTL mapping

Genomic DNA of the RIL population and parent lines was extracted and assessed as previously described [29]. The DNA samples were hybridized to the GenoBaits Wheat 40 K Panel containing 202,971 markers. Genotyping was performed at the MOLBREEDING (Shijiazhuang) Biotech Co., Ltd. (http://www.molbreeding.com). A total of 15,258 homozygous SNPs were selected from ND5181 and Rht8-2 for follow-up analyses. The BIN and MAP functions of IciMapping 4.1 were used to remove redundant markers and construct the genetic map, respectively. The genetic linkage map contained 1847 Bin markers across 21 chromosomes. The threshold of the logarithm of odds (LOD) score was set to 2.5, and the Kosambi map function was used to calculate the map distance from recombination frequencies. Composite interval mapping (ICIM) on IciMapping 4.1 was selected to identify QTLs for PH and SC. The mean values of the phenotypic traits in each line were used for QTL analysis. QTLs detected in two environments were regarded as stable QTLs.

KASP marker development and QTL validation

We developed the KASP assay based on the SNPs identified within or around ND5181 and Rht8-2 from the 40 K SNP array genotyping data. The online primer design pipeline PolyMarker (http://polymarker.tgac.ac.uk/) was used to design specific primers. The KASP assays were performed as previously described [29]. End-point fluorescence data were screened using the microplate reader FLUOstar Omega SNP (BMG LABTECH, Germany) and analyzed by the Klustering Caller Software. The KASP markers were tested on the two parents, and then the developed polymorphic KASP markers were used for the identification of genotypes in the F7 RIL population. QTL analysis was conducted using IciMapping 4.1.

Cloning and sequencing of Rht8-B1

Gene-specific primers were designed to amplify the full-length sequence of Rht8-B1. The genomic DNA of ND5181 and Rht8-2 was extracted and used as a template in PCR. The components of each 20 µL reaction were as follows: 10 µL buffer, 4 µL dNTPs, 1 µL genomic DNA, 0.8 µL forward primer, 0.8 µL reverse primer, 0.4 µL KOD FX, and 3 µL ddH2O. The reaction conditions were 94 ℃ for 2 min, followed by 32 cycles of 98 ℃ for 10 s, annealing at 65 ℃ for 20 s, and 68 ℃ for 2 min, with a final extension of 68 ℃ for 5 min. The PCR products were sequenced at the Shanghai Sangon Biotech Co., Ltd. (https://www.sangon.com/).

Generating Rht8-B1 and Rht8-D1 mutants by gene editing

For CRISPR/Cas9-based gene editing, single guide RNA (sgRNA) target sequences were designed and plant transformations were performed as previously described [10]. The sgRNA sequence targeting Rht8 genes was 5’-GCCGCCGGAGAGCAGCTGCC-3’. We sequenced the Rht8 genes in T2 plants and validated the mutations produced by CRISPR/Cas9-based gene editing. Single mutants of Rht8-B1 and Rht8-D1 were successfully selected, and the heights of edited and wild-type (WT) plants recorded and compared.

Quantitative RT-PCR

Quantitative RT-PCR was performed according to our previous report [30]. Briefly, total RNA was isolated from the first internode below the spike using TRNzol-A+ Reagent (Tiangen Biotech), and then purified using an RNA purification kit (Tiangen). The first-strand cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad), and the SsoFast EvaGreen Supermix Kit (Bio-Rad) was used for quantitative RT-PCR. This experiment was conducted on a CFX 96 Real-Time System (Bio-Rad) following the manufacturer’s instructions. The actin gene was used as an internal control. The primers used for quantitative RT-PCR are listed in Supplementary Table 5.

Distribution analysis of Rht8-B1b in wheat accessions

A total of 305 worldwide accessions with genotypic information obtained from the Wheat Union Database (http://wheat.cau.edu.cn/WheatUnion/) were used for analysis of allelic variation in Rht8-B1 [31,32,33,34]. The frequency of Rht8-B1b in wheat accessions from different geographical regions was calculated according to the number of accessions carrying Rht8-B1b allele.

Accession numbers

Sequence data of Rht8-B1a and Rht8-B1b have been deposited in the GenBank data library under the accession numbers OQ512875 and OQ512876, respectively.

Statistical analysis

Statistical analyses, namely Student’s t-tests and correlation analyses, were performed using SPSS v21.0 software (IBM, USA).

Results

Phenotypic variation of two parent lines and the RIL population

We have previously identified a semi-dwarf wheat mutant line Rht8-2 with high yield potential [10]. To explore QTLs associated with important agronomic traits, we constructed a RIL population including 139 lines derived from a cross between the wheat variety ND5181 and the mutant line Rht8-2. The two parent lines showed significant differences in PH and SC in two environments (Zhongpuchang and Changping) (Fig. 1A). Specifically, compared with ND5181, the PH of Rht8-2 was 11 cm shorter, and SC was larger by 0.45 (Fig. 1; Table S1). In the RIL population, PH and SC displayed obvious transgressive segregation. PH and SC showed normal distributions in the two environments, suggesting they are controlled by multiple genes (Fig. 2). Pairwise correlation analysis between PH and SC showed correlation coefficients of -0.45 and -0.41 with statistical significance in the Zhongpuchang and Changping field experiments, indicating that a negative correlation between PH and SC in the RIL population.

Fig. 1
figure 1

Phenotype comparisons between ND5181 and Rht8-2. (A) Phenotype of ND5181 and Rht8-2. (B) Plant height of ND5181 and Rht8-2. (C) Spikelet compactness of ND5181 and Rht8-2

Full size image
Fig. 2
figure 2

The frequency distributions of plant height, and spikelet compactness in the RIL population. A-B. Frequency distribution of plant height in the Zhongpuchang (A) and Changping (B) environments. C-D. Frequency distribution of spikelet compactness in in the Zhongpuchang (C) and Changping (D) environments. Phenotypic values of the two parental lines are marked by vertical arrows

Full size image

QTL mapping analysis

A total of 23 QTLs associated with PH and SC were detected in the two environments on chromosomes 1A, 1B, 2B, 2D, 3A, 3B, 4B, 5A, 5B, 6B, 6D, 7A, 7B, and 7D. Among these QTLs, qPH2B.1, qPH2D, qPH4B, qSC1B, qSC2B.1, qSC2D.1, and qSC7D, were detected in both environments (Fig. 3; Table 1).

Fig. 3
figure 3

Chromosomal locations of the identified stable QTLs associated with PH and SC. The stars and triangles represent PH and SC, respectively. The red and green colors indicate data from the Zhongpuchang and Changping environments, respectively

Full size image
Table 1 QTLs associated with plant height (PH) and spikelet compactness (SC) in the Zhongpuchang and Changping environments identified with IciMapping 4.1
Full size table

Three stable QTLs associated with PH were identified on chromosomes 2B, 2D, and 4B in both environments. The major QTL qPH4B showed the highest LOD scores (30.1 in Zhongpuchang and 40.2 in Changping) and explained 29.7% and 41.8% of the phenotypic variation, respectively. qPH2D had LOD scores 22.7 (Zhongpuchang) and 23.6 (Changping) and accounted for 19.2% and 17.8% of the phenotypic variation, respectively, and qPH2B.1 had LOD scores of 11.5 (Zhongpuchang) and 4.0 (Changping) and explained 7.9% and 2.1% of the phenotypic variation, respectively. The qPH4B allele from ND5181 decreased PH, while the alleles of qPH2D and qPH2B.1 from Rht8-2 reduced PH (Table 1).

We identified four stable QTLs associated with SC on chromosomes 1B, 2B, 2D, and 7D in the two environments. qSC2B.1 had LOD scores 24.6 and 12.7 and explained 25.0% and 15.0% of the phenotypic variation in Zhongpuchang and Changping, respectively; qSC2D.1 had LOD scores of 13.6 and 19.4 and explained 13.0% and 27.6% of the phenotypic variation, respectively; qSC1B had LOD scores of 5.0 and 6.2 and explained 3.8% and 7.0% of the phenotypic variation, respectively; and qSC7D had LOD scores of 2.8 and 5.4 and explained 1.9% and 5.6% of the phenotypic variation, respectively. Among these QTLs in ND5181, the qSC1B and qSC7D alleles increased SC and the qSC2B.1 and qSC2D.1 alleles decreased SC (Table 1).

Rht8-B1 is the candidate gene of qPH2B.1

To validate the mapped region of qPH2B.1, we successfully developed nine KASP makers around the qPH2B.1 region based on the results of 660 K SNP array analysis between the ND5181 and Rht8-2 varieties. These markers validated this region and delimited it to a physical interval of 3.5 Mb between markers 2B-4 and 2B-5. This QTL showed a LOD score of 3.4 and explained 11.8% of the observed phenotypic variation (Fig. 4). Given that this region included Rht8-B1 (TraesCS2B02G073600), a homoeologous gene of Rht8, we sequenced the region and uncovered two genetic variants (39,418,567–39,418,568, GC in ND5181 and TT in Rht8-2) in the coding region of the gene, which resulted in an amino acid change from G (ND5181) to V (Rht8-2) at the 175th position. We then analyzed the effects of Rht8-B1 on PH and SC in the RIL population and found that TT allele reduced PH by 6.2% and 3.6%, shortened SL by 6.6% and 4.7%, and increased SC by 5.8% and 6.3% in the Zhongpuchang and Changping field experiments, respectively (Fig. 4).

Fig. 4
figure 4

Validation of qPH2B.1 and the predicted effects of Rht8-B1b in the RIL population. (A). Validation of qPH2B.1. B-D. The effects of Rht8-B1b on plant height (B), spike length (C), and spikelet compactness (D)

Full size image

Rht8-B1 exhibited a lower impact on PH reduction than Rht8-D1

We compared the effects of Rht8-B1 and the previously reported Rht8-D1 gene on PH by knocking out both genes in a Fielder background and evaluating the PH of the T2 lines. The results showed that the PH of edited plants was significantly lower than that of control plants. Specifically, edited Rht8-B1 and Rht8-D1 plants showed a PH reduction of 5.6% and 17.5%, respectively, compared with the WT control, suggesting that Rht8-B1 had a smaller effect on PH than Rht8-D1 (Fig. 5; Figure S1).

Fig. 5
figure 5

The effects of Rht8-B1 and Rht8-D1 on plant height of CRISPR-Cas9-edited plants. (A) Phenotype of Fielder and edited lines (Rht8-B1 and Rht8-D1). (B) Plant height of Fielder and edited lines (Rht8-B1 and Rht8-D1)

Full size image

Comparison of the expression of Rht8-B1 and Rht8-D1

To examine the difference in expression patterns between Rht8-B1 and Rht8-D1, we analyzed their transcript levels using the Hexaploid Wheat Expression Dataset [35] and found that although both genes were highly expressed in the stem at the jointing stage, the expression of Rht8-D1 was significantly higher than that of Rht8-B1 (Fig. 6A). Our previous study suggested that a frameshift mutation in Rht8-D1 caused the dwarfism phenotype in Rht8-2 and that its expression was significantly lower in the mutant compared with that in WT [10]. To investigate the effects of mutation of Rht8-D1 on the B subgenome of Rht8, we analyzed the expression of Rht8-B1 in the first internode below the spike of the mutant Rht8-2 and WT. The results showed that the expression of Rht8-B1 was remarkably higher in Rht8-2 (Fig. 6B), indicating that the mutation of Rht8-D1 affected the transcript level of Rht8-B1.

Fig. 6
figure 6

A. Expression data for Rht8-B1 and Rht8-D1 from the Hexaploid Wheat Expression Database. B. Analysis of Rht8-B1 expression in the first internode below the spike in WT and Rht8-2

Full size image

Distribution of Rht8-B1 in wheat varieties worldwide

We used a total of 305 worldwide accessions from the Wheat Union Database (http://wheat.cau.edu.cn/WheatUnion/), namely 193 accessions from China and 112 accessions from other countries, for the analysis of Rht8-B1 allelic variation. We found that 68 Chinese accessions (35.2%) contained the Rht8-B1b TT allele, compared with only 6 accessions (5.4%) in other countries (Fig. 7). Of the Chinese accessions, 20.3% of the 118 modern varieties and 58.7% of the 75 landrace accessions had Rht8-B1b allele, (Table S2). These results suggest that Rht8-B1b allele was not widely used historically for wheat breeding.

Fig. 7
figure 7

Analysis of the distribution of Rht8-B1b in wheat accessions worldwide. GC and TT represent Rht8-B1a and Rht8-B1b, respectively

Full size image

Discussion

Semi-dwarf wheat usually possesses high lodging resistance with high yield stability [4]. The identification of novel genes regulating PH is pivotal for improving lodging resistance through wheat breeding. In this study, we identified Rht8-B1 as a novel regulator of PH by performing genetic mapping and gene editing analyses. By combining mapping using the Wheat40K array and molecular markers, qPH2B.1 was mapped within a physical interval of 36.5–40.0 Mb between markers 2B-4 and 2B-5 on chromosome 2B of the Chinese Spring reference genome v1.0 (Fig. 4A). The Rht8-B1 gene was located to position ~ 39.4 Mb on chromosome 2B, making it a candidate gene for qPH2B.1. Genotyping of Rht8-B1 in the RIL population showed that genetic variants in this gene were associated with PH (Fig. 4B). Further gene editing of Rht8-B1 validated its function in the regulation of PH (Fig. 5). These results suggest Rht8-B1 is the causal gene of qPH2B.1.

We recently identified Rht8 on chromosome 2D (Rht8-D1) by performing map-based cloning using two dwarf mutants, Rht8-2 and Rht8-3; this gene encodes a Ribonuclease H-like protein that modifies PH by regulating the bioactive GA content [10]. Here, we used Rht8-2 for the construction of a RIL population and found a major QTL – qPH2D – on chromosome 2D with a high LOD score, which probably represents Rht8-D1. In contrast to the Rht8-D1 (qPH2D) locus, qPH2B.1 had a minor effect on PH (Figs. 3 and 4 A), which is consistent with gene editing results (Fig. 5) and reports by Chai et al. [11]. The larger effect of Rht8-D1 on PH relative to Rht8-B1 is also in accordance with the higher expression of Rht8-D1 in the stem (Fig. 6A). Previous studies extensively investigated the function of the Rht8 orthologous gene TAC4/sg2 in rice [36, 37]. A stop-gain mutation in TAC4 was found to lead to a greater tiller angle and a dwarf phenotype that may result from changes in IAA content and distribution. In addition, an 8-bp deletion in sg2 was found to result in a smaller grain phenotype due to repressed cell expansion in spikelet hulls and a semi-dwarf phenotype. These observations indicate that orthologous genes to Rht8 affect multiple agronomic traits and play different roles, highlighting the need for further exploration of the effects of Rht8-B1 and Rht8-D1 on PH in the future.

The semidwarf Rht8-B1b allele has high potential for utilization in wheat breeding. By analyzing the genetic sequence of Rht8-B1 in the mutant Rht8-2 and WT variety Jing411, we found that the Rht8-B1b allele was derived from the WT, indicating that it is a natural variant. Importantly, the distribution of the Rht8-B1b allele in global wheat varieties suggests it has not been widely utilized historically in wheat breeding outside of China (Fig. 7). Similar to Rht8-D1b, the Rht8-B1b allele had no effects on thousand grain weight. Several studies have suggested that dwarf or semidwarf genes are associated with a decrease in thousand grain weight and ultimately affect wheat yield [38,39,40]. The combination of Rht8-B1b and other dwarf genes in wheat breeding can thus be an alternative approach for developing lodging-resistant wheat.

We found consistency between the stable QTLs associated with PH and SC detected in this study and previous reports. For example, Rht-B1b was located at ~ 30.8 Mb on chromosome 4B within the chromosomal region corresponding to qPH4B [5]; qSC2D.1 was identified between molecular markers A61578 and A61731 in the interval 20.8–30.3 Mb, and is closely linked to Rht8-D1, which was reported to significantly reduce SL [11]; qSC7D was located between molecular markers A202015 and A202077 in the interval 584.5–588.2 Mb, and is closely linked to WAPO1-7D, which regulates spikelet number per spike [41]. We also identified several putatively novel QTLs for PH and SC, including qPH7B.1 on chromosome 7BL, which had a LOD score 10.3 and explained 7.0% of the phenotypic variation in PH, and qSC1B located on the long arm of chromosome 1B between molecular markers A18513 and A19081, which is associated with SC.

Conclusion

We identified seven stable QTLs for PH and SC in two environments using a RIL population. Using mapping and gene editing analyses we found that Rht8-B1 is the causal gene of qPH2B.1 and affects PH variation. Rht8-B1 had less of an effect on PH than Rht8-D1 and has not been widely utilized in wheat breeding. This implies that combining Rht8-B1b with other favorable Rht genes has great potential for breeding lodging-resistant wheat varieties. Our study revealed novel markers and QTLs, providing important information for marker-assisted selection in wheat.