Genomic analysis of Zhou8425B, a key founder parent, reveals its genetic contributions to elite agronomic traits in wheat breeding

Abstract

High-quality genome information is essential for efficiently deciphering and improving crop traits. Here, we report a highly contiguous and accurate hexaploid genome assembly for the key wheat breeding parent Zhou8425B, an elite 1BL/1RS translocation line with durable adult plant resistance (APR) against yellow rust (YR) disease. By integrating HiFi and Hi-C sequencing reads, we have generated a 14.75-Gb genome assembly for Zhou8425B with a contig N50 of 70.94 and a scaffold N50 of 735.11 Mb. Comparisons with previously sequenced common wheat cultivars shed light on structural changes in the 1RS chromosome arm, which has been extensively used in wheat improvement. Interestingly, Zhou8425B 1RS carries more genes encoding AP2/ERF-ERF or B3 transcription factors than its counterparts in four previously sequenced wheat and rye genotypes. The Zhou8425B genome assembly aided in the fine mapping of a new APR locus (YrZH3BS) that confers resistance to YR disease and promotes grain yield under field conditions. Notably, pyramiding YrZH3BS with two previously characterized APR loci (YrZH22 and YrZH84) can further reduce YR severity and enhance grain yield, with the triple combination (YrZH3B + YrZH22 + YrZH84) having the greatest effect. Finally, the founder genotype effects of Zhou8425B were explored using publicly available genome resequencing data, which reveals the presence of important Zhou8425B genomic blocks in its derivative cultivars. Our data demonstrate the value of the Zhou8425B genome assembly for further study of the structural and functional characteristics of 1RS, the genetic basis of durable YR resistance, and founder genotype effects in wheat breeding. Our resources will facilitate the development of elite wheat cultivars through genomics-assisted breeding.

This study reports a highly contiguous and accurate genome assembly for Zhou8425B, a key wheat founder parent with many hundreds of derivative cultivars, and provides insights into the genetic contributions of Zhou8425B to elite agronomic traits, e.g., yield-related traits and durable yellow rust resistance, in wheat breeding. These results will aid genomics-assisted wheat improvement in the future.

Introduction

Wheat, including mainly common wheat (Triticum aestivum, AABBDD, 2n = 6x = 42) and its cultivated relatives, is a staple food for over 35% of the world population (Erenstein et al., 2022). Its efficient improvement is essential for maintaining the food and nutritional security of an ever-increasing global population. However, efficient improvement of wheat traits is a challenging task, as they are controlled by polygenes and affected by environmental factors. Fortunately, advances in plant genomics and genome editing have greatly increased the pace at which crop traits are dissected and enhanced (Gao, 2021; Varshney et al., 2021; Sun et al., 2022). Since 2014, genome sequences have been reported for an increasing number of wheat accessions (International Wheat Genome Sequencing Consortium IWGSC, 2014; IWGSC, 2018; Walkowiak et al., 2020). More recently, there has been growing international interest in sequencing elite wheat cultivars to accelerate the identification of genes that control important agronomic traits (Sato et al., 2021; Shimizu et al., 2021; Akpinar et al., 2022; Athiyannan et al., 2022; Aury et al., 2022; Kale et al., 2022; Shi et al., 2022; Jia et al., 2023). However, among the ∼20 common wheat genomes sequenced to date, only two (Aikang 58 and Kenong 9204) carry the 1RS translocation extensively used in worldwide wheat improvement through introgression breeding (Rabanus-Wallace et al., 2021; Shi et al., 2022; Jia et al., 2023). Lack of sufficient genomic information for 1BL/1RS translocation lines prohibits efficient molecular and functional analysis of the 1RS genes that contribute to wheat growth and control of agronomic traits.

Zhou8425B is a semidwarf 1BL/1RS translocation line that exhibits superior traits including durable adult plant resistance (APR) to yellow rust (YR) and leaf rust (LR) diseases (Jia et al., 2018). Both diseases cause severe economic losses in global wheat production (Bhavani et al., 2022). Using Zhou8425B and derivative varieties, four major APR loci against YR (YrZH22 and YrZH84) or LR (LrZH22 and LrZH84) have been mapped on chromosomes 1B (LrZH84), 2B (LrZH22), 4B (YrZH22), and 7B (YrZH84) (Li et al., 2006; Zhao et al., 2008; Wang et al., 2016, 2017; Zhang et al., 2017). LrZH22 (Lr13) was recently isolated by map-based cloning and found to encode a nucleotide-binding leucine-rich repeat protein (Hewitt et al., 2021; Yan et al., 2021). However, the molecular identities of LrZH84, YrZH22, and YrZH84 remain obscure at present. Zhou8425B also shows elite yield-related traits (such as higher grain number per spike and larger grains), resistance to lodging, and tolerance to abiotic stresses (e.g., drought and late-season heat stress), and more than 70 quantitative trait loci (QTLs) that control such traits have been detected using Zhou8425B-derived genetic populations (Gao et al., 2015; Huang et al., 2020; Li et al., 2021a; Dong et al., 2022; Wei et al., 2022). However, none of the detected QTLs have been molecularly cloned. Lack of genome sequence information may have hindered progress in characterizing the valuable genes carried by Zhou8425B.

Owing to its outstanding agronomic traits, Zhou8425B has been used as a founder parent in wheat breeding in China since its release in 1988 (He et al., 2011, 2018; Xiao et al., 2011; Tang et al., 2015; Li et al., 2018, 2023; Zhang et al., 2021). However, the founder genotype effects of Zhou8425B have not been studied at the whole-genome level, which is not conducive for systematically dissecting the contributions of Zhou8425B to modern wheat breeding. To aid characterization of the valuable genes in Zhou8425B and facilitate genomics-assisted wheat breeding, we developed a highly contiguous genome assembly for Zhou8425B using PacBio HiFi sequencing and high-throughput chromosome conformation capture sequencing (Hi-C) technologies. The quality of the Zhou8425B genome assembly, as judged by contig and scaffold N50 values, average nucleotide accuracy, BUSCO completeness score, and long terminal repeat score (LAI) (Ou et al., 2018), was better than that of recently sequenced hexaploid wheat cultivars (Sato et al., 2021; Shimizu et al., 2021; Akpinar et al., 2022; Athiyannan et al., 2022; Aury et al., 2022; Kale et al., 2022; Shi et al., 2022; Jia et al., 2023).

Comparison of the Zhou8425B assembly with those of Aikang 58, Kenong 9204, and two diploid rye lines (Lo7 and Weining) for which genome sequence information is available (Li et al., 2021b; Rabanus-Wallace et al., 2021) shed light on structural variations in the 1RS chromosome arm. Interestingly, the 1RS arm of Zhou8425B harbored more genes predicted to encode AP2/ERF-ERF or B3 transcription factors (TFs) than did its counterparts in Aikang 58, Kenong 9204, Lo7, and Weining, and we identified 1RS genes likely to be useful for further deciphering the contributions of 1RS to wheat growth and trait improvement. The Zhou8425B genome assembly was used to accelerate the mapping of a new APR locus (tentatively named YrZH3BS) against YR to a 1–2 Mb interval on the 3BS chromosome arm. The genetic effects of YrZH3BS and its combinations with YrZH22 and YrZH84 on reducing YR disease severity and enhancing grain yield under field conditions were validated by analyzing 212 Zhou8425B derivative varieties, thus providing insight into the genetic basis of durable YR resistance and highlighting the benefits of pyramiding multiple APR genes for simultaneous enhancement of wheat YR resistance and grain-yield performance. Finally, the founder genotype effects of Zhou8425B were explored using the wheat genome resequencing data reported by Hao et al. (2020), which included seven Zhou8425B derivative cultivars. Together, our data demonstrate the value of the Zhou8425B genome assembly and related resources for further study of the structural and functional characteristics of 1RS, the genetic basis of durable YR resistance, and founder genotype effects in wheat breeding.

Results

Pedigree, agronomic traits, and breeding application of Zhou8425B

Zhou8425B was developed in China from 1978 to 1988 (Xiao et al., 2011; Tang et al., 2015). Its pedigree included the hexaploid triticale line Guangmai 74 (AABBRR) and several common wheat varieties that originated in European countries and China; its breeding process involved multiple rounds of crossing and backcrossing and a single γ-irradiation treatment of dry seeds (supplemental Figure 1A). Zhou8425B was considerably shorter than Guangmai 74 and had longer spikes and much larger grains (supplemental Figure 1B–1D).

To better understand the origin of important trait genes in Zhou8425B, we performed molecular marker analysis and revealed the differential presence of four known dwarfing genes (Rht1, Rht2, Rht8, and Rht24), the key photoperiod-insensitive gene Ppd-D1a, five APR loci (including the newly identified YrZH3BS locus against YR disease), and seven important grain weight-associated genes (TaCwi-A1, TaSus1-7A, TaSus1-7B, TaSus2-2A, TaGS-D1, TaGS5-A1, and TaGW2-6B) (Khalid et al., 2019; Zhang et al., 2021; Chegdali et al., 2024) in the lines used during the breeding of Zhou8425B (supplemental Table 1). Most of the examined genes were effectively pyramided in Zhou78A, the immediate maternal parent of Zhou8425B, whereas the paternal parent Annong 7959 provided Rht24 and elite alleles of TaCwi-A1 and TaGS5-A1 (supplemental Table 1).

Zhou8425B and its early-generation descendants (e.g., Zhoumai 16, Zhoumai 22, and Aikang 58) have been extensively used as wheat breeding parents in China, and 839 derivative cultivars have passed provincial and/or national certification by 2023 (supplemental Table 2). Many derivative varieties, such as Zhoumai 16, Zhoumai 22, Aikang 58, Zhengmai 7698, Bainong 207, and Bainong 4199, have gained wide application in the major winter wheat production zones of China (He et al., 2011; 2018; Guan et al., 2019). We selected five representative Zhou8425B derivative cultivars (Zhoumai 11, Zhoumai 13, Zhoumai 16, Zhoumai 22, and Aikang 58) (supplemental Figure 2) and used themto investigate the inheritance of the 17 important trait genes analyzed above (supplemental Table 1). Eight of the 17 genes were detected in all five descendants, including three dwarfing genes (Rht1, Rht2, and Rht24), Ppd-D1a, LrZH84, and three grain weight-associated genes (TaSus1-7A, TaSus1-7B, and TaGS5-A1) (supplemental Table 1), indicating that these genes may contribute to the elite traits of Zhou8425B and its derivatives.

Characteristics of the Zhou8425B genome assembly

Before constructing the Zhou8425B genome assembly, we checked its chromosome number and 1RS translocation using fluorescent in situ hybridization assays. Forty-two somatic chromosomes were consistently observed in root tip cells, and one pair of 1BL/1RS translocation chromosomes was identified (Figure 1A). We then constructed a 14.75-Gb genome assembly with contig and scaffold N50 values of 70.94 and 735.11 Mb, respectively (Table 1) by combining HiFi (48× genome coverage) and Hi-C (124× genome depth) sequencing reads (supplemental Data S1). Approximately 98.44% (14.52 Gb) of the assembled sequences were anchored onto 21 chromosomes on the basis of chromatin interactions revealed by Hi-C analysis (Table 1; supplemental Figure 3); the remaining 0.23 Gb of unanchored sequences were mainly rDNA repeats.

Figure 1.

Genome assembly of Zhou8425B.

(A) Characterization of Zhou8425B somatic chromosomes by fluorescent in situ hybridization assays using the oligonucleotide probes pSc119.2 (green) and pTa535 (red). A total of 42 chromosomes, including two 1BL/1RS translocation chromosomes (arrows), were consistently identified.

(B) Schematic representation of the 21 assembled chromosomes of Zhou8425B. The contigs scaffolded onto each chromosome are shown on the left. Centromeres are marked by blue rectangles. Telomeres, represented by green arcs, were assembled for one or both ends of 18 chromosomes.

(C) Circos plot displaying some important features of the Zhou8425B genome assembly. From outside to inside are the characteristics of 21 chromosomes (a), density of high confidence protein-coding genes (b), Gypsy LTR transposable elements (TEs) (c), Copia LTR TEs (d), DNA TEs (e), and syntenic relationships among the chromosomes of the three subgenomes (f). The gene and TE densities shown in (b) to (e) were obtained by scanning each chromosome sequence in 3-Mb windows.

Table 1.

Main features of the Zhou8425B genome assembly and annotation

Assembly
Assembled genome size	14.75 Gb
Sequence assigned to chromosomes	14.52 Gb
Number of scaffolds	1109
N50 scaffold length	735.11 Mb
N90 scaffold length	532.03 Mb
Number of contigs	1729
N50 contig length	70.49 Mb
N90 contig length	9.87 Mb
Longest contig	323.62 Mb
Longest scaffold	866.06 Mb
GC content	46.06%
Annotation
Number of high-confidence genes	106 940
Mean length of high-confidence genes	3569 bp
Total length of high-confidence genes	381.70 Mb
Number of genes supported by transcript evidence	96 728
Retrotransposons	9.60 Gb
DNA transposons	2.77 Gb
Other repeats	544.37 Mb
Total repeat sequences	12.91 Gb
Total repeat sequences percentage	87.53%
Number of telomeres assembled	27

The assembled A, B, and D subgenomes were 5.01, 5.43, and 4.08 Gb in length, respectively (supplemental Data S1). This result is consistent with the finding that more contigs (40–85) were needed to span individual subgenome-B chromosomes, compared with only 11–42 for subgenome-A chromosomes, and 5–15 for subgenome-D chromosomes (Figure 1B and supplemental Data S1). The centromeric regions, highly enriched in Cereba and Quinta retrotransposons (supplemental Figure 4), were deduced for all 21 chromosomes (supplemental Data S1). We assembled 27 telomeres, and 10 chromosomes (1D, 2D, 3B, 3D, 4A, 4B, 5A, 5B, 6A, and 7B) had assembled telomeres at both ends (Figure 1B). The GC content of the assembly was 46.06% (Table 1).

To aid in gene annotation, 12 Zhou8425B plant samples collected at the seedling, heading, and post anthesis stages were used for RNA sequencing, resulting in 161.35 Gb of transcriptome data (supplemental Data S1). By combining these data and the gene sets reported for previously sequenced wheat genotypes, we annotated 106 940 high-confidence (HC) protein-coding genes for Zhou8425B, with 96 728 (90.45% of the HC genes) supported by transcript evidence (Table 1). Moreover, about 99% of the 106 940 HC genes were functionally annotated by BLAST analysis in gene function- annotation databases, including NR, SwissProt, GO, KEGG, and eggNOG (supplemental Data S1). The HC genes were mainly distributed toward the terminal regions of the 21 chromosomes (Figure 1C). Five types of non-coding RNA genes (rRNA, tRNA, miRNA, snRNA, and snoRNA) were also identified in the Zhou8425B genome assembly and numbered 26 237, 51 883, 39 521, 1147, and 3973, respectively (supplemental Data S1).

Approximately 12.91 Gb of the Zhou8425B assembly consisted of repetitive elements, accounting for 87.53% of the assembled genome (Table 1; supplemental Data S1). The content of retrotransposons (9.60 Gb) was higher than that of DNA transposons (2.77 Gb). The three major families of transposons were Gypsy, Copia, and CACTA, which occupied 43.69%, 16.01%, and 15.05% of the Zhou8425B genome, respectively (supplemental Data S1). Gypsy transposons were enriched toward the centromeric regions, whereas Copia and CACTA elements accumulated more in the terminal regions of chromosomes (Figure 1C).

Assembly quality of the Zhou8425B genome sequence

The high quality of the Zhou8425B genome assembly was supported by multiple lines of evidence. First, 99.96% of the 43,318,145 unique HiFi reads could be mapped to the assembly (supplemental Data S1). Second, mapping of 368.15 Gb of Zhou8425B Illumina reads (∼25× of the assembled genome) yielded a mean per base quality value (QV) of 47 and an average nucleotide accuracy greater than 99.99% (supplemental Data S1). Third, a BUSCO completeness score of 99.5% was obtained for the assembly (supplemental Data S1). Finally, the LAI value, which indicates the contiguity of intergenic and repetitive regions of the genome assembly (Ou et al., 2018), was greater than 15 for the A, B, and D subgenomes of the Zhou8425B assembly (supplemental Figure 5), higher than the values reported for recently released hexaploid wheat genomes (Shi et al., 2022; Jia et al., 2023).

To verify the high continuity of the Z8325B genome assembly, we analyzed gluten gene loci, which have an important influence on wheat end-use qualities (Wang et al., 2020). These complex loci specify two types of high-molecular-weight glutenin subunits (HMW-GSs), three types of low-molecular-weight glutenin subunits (LMW-GSs), and four types of gliadins, respectively (Wang et al., 2020). Owing to the presence of numerous paralogs, the loci that encode LMW-GSs and/or gliadins are usually difficult to assemble (Huo et al., 2018a, 2018b). As anticipated, two paralogous genes specifying x- and y-types of HMW-GSs were detected in each of the three homoeologous Glu-A1, Glu-B1, and Glu-D1 loci, with Glu-A1y being a pseudogene (supplemental Figure 6; supplemental Data S2). Consequently, five HMW-GSs were expressed in Zhou8425B grains (supplemental Figure 7). Ten gluten genes, including 4 pseudogenes, were found in the composite locus Gli-A1/Glu-A3, with the 6 active members coding for 1 δ-gliadin, 2 γ-gliadin, and 3 LMW-GS proteins (supplemental Figure 6; supplemental Data S2). Zhou8425B lacked the Gli-B1/Glu-B3 locus because of the 1RS translocation that replaced the 1BS chromosome arm. We therefore examined the Sec-1 and Sec-4 loci carried by 1RS (Li et al., 2021b). Twenty-three gluten genes, including 13 pseudogenes, were present in the Sec-1 locus, with the 10 active members encoding 40K γ-secalins; 18 gluten genes were detected in the Sec-4 locus, with 15 active members encoding ω-secalins (supplemental Figure 6; supplemental Data S2). Seventeen gluten genes, including 4 pseudogenes, were observed in the Gli-D1/Glu-D3 locus, with the 13 active members encoding 1 δ-gliadin, 3 γ-gliadin, 3 ω-gliadin, and 6 LMW-GS proteins (supplemental Figure 6; supplemental Data S2). The homoeologous Gli-A2, Gli-B2, and Gli-D2 loci, residing on group 6 chromosomes, contained 15, 14, and 11 active genes encoding α-gliadins, respectively; these loci also harbored various numbers of inactive α-gliadin genes (3–19) (supplemental Figure 6; supplemental Data S2). Corresponding to the annotation of 9 active LMW-GS genes in Zhou8425B, multiple LMW-GS proteins accumulated in its grains (supplemental Figure 7). Although the 10 gluten-gene loci varied greatly in size (0.06–20.23 Mb, supplemental Figure 6), they were mostly covered by a single contig in the Zhou8425B genome assembly, with the exception of Glu-B1, which encompassed 2 adjacent contigs (supplemental Data S2). These data, together with the evaluation results above, suggest that the Zhou8425B genome assembly is highly contiguous and accurate in both genic and non-genic spaces.

Gene expression features of Zhou8425B based on transcriptome analysis

Principal-component analysis of gene expression data showed separation of the 12 transcriptome samples and clustering of samples with similar gene expression profiles (Figure 2A). On the basis of FPKM (fragments per kilobase of exon model per million mapped fragments) values, the proportions of genes that were expressed at low (FPKM ≤ 1), intermediate (1 < FPKM ≤ 10), or high (10 < FPKM ≤ 100) levels were 61.60%, 28.18%, and 9.53%, respectively; only 0.68% of the genes were strongly expressed (FPKM > 100) (Figure 2B). Construction of gene co-expression networks with 64 692 HC genes expressed in at least 6 of the 12 transcriptome samples revealed 22 co-expression modules, with 1 to 4 clusters of co-expressed genes significantly associated with individual samples (Figure 2C). More detailed analysis of the MEblue module revealed one cluster of co-expressed genes that were significantly associated with developing grains at 15 days post anthesis (DPA); within this cluster, a group of genes was found to be co-expressed with the putative hub gene TraesZ8425B1RS01G036200 (Figure 2D), which resides on the 1RS translocation and may encode a putative B3 TF. Genes co-expressed with TraesZ8425B1RS01G036200 included not only those located on 1RS but also those on wheat chromosomes, many of which were annotated to encode gluten proteins (e.g., HMW-GS), enzymes involved in starch biosynthesis (starch synthase) and metabolism (α-amylase inhibitor), and proteins that function in lipid utilization (lipid transfer proteins) (Figure 2D; supplemental Data S3).

Figure 2.

Transcriptome analysis of 12 samples collected from Zhou8425B plants.

(A) Clustering of 12 samples by principal-component analysis of gene expression. The 12 samples included 10-day seedling leaves (SL), 10-day seedling roots (SR), heading-stage leaves (HL), heading-stage roots (HR), heading-stage stems (HS), heading-stage stem nodes (HSN), heading-stage ineffective tillers (HIT), heading-stage ineffective tiller buds (HITB), immature spikes (1 cm) (IS1), immature spikes (2–3 cm) (IS2), and developing grains at 5 (5DG) or 15 (15DG) days post anthesis.

(B) Differences in the expression levels of Zhou8425B genes revealed by transcriptome analysis of the 12 samples.

(C) Construction of co-expression modules with the transcriptome data from 12 Zhou8425 plant samples. A total of 22 modules (listed on the right) were obtained for the 12 samples (bottom panel). Each significant cell had two values, correlation coefficient (top) and p value (below).

(D) A subset of the genes co-expressed with TraesZ8425B1RS01G036200, which were extracted from the MEblue module associated with developing grains (15DG). The seven illustrated co-expressed genes were predicted to encode high-molecular-weight glutenin subunit (TraesZ8425B1BL01G290100), low-molecular-weight glutenin subunit (TraesZ8425B1D01G008500), α-amylase inhibitor (TraesZ8425B2B01G023600), lipid transfer proteins (TraesZ8425B1A01G278800 and TraesZ8425B1D01G287000), starch synthase (TraesZ8425B1RS01G089600), and γ-secalin (TraesZ8425B1RS01G013900).

Comparative analysis of the 1RS arms in different genetic backgrounds

The 1RS translocation has contributed substantially to global wheat improvement (Rabanus-Wallace et al., 2021). Although the resistance genes (Pm8, Yr9, Lr26, and Sr31) carried by the 1RS originated from Petkus rye have lost their functions, largely because of changes in pathogen races, several other sources of 1RS have been reported to confer resistance to multiple pathogens after being transferred to wheat (Crespo-Herrera et al., 2017; Szakács et al., 2020; Li et al., 2022; Luo et al., 2022). Furthermore, there is increasing evidence that 1RS also carries genes beneficial for wheat tolerance to abiotic stresses such as drought (Gabay et al., 2021; 2023). Genome assemblies for two common wheat cultivars (Aikang 58 and Kenong 9204) carrying the 1RS translocation and two diploid rye lines (Lo7 and Weining) were recently reported (Li et al., 2021b; Rabanus-Wallace et al., 2021; Shi et al., 2022; Jia et al., 2023). We were therefore able to compare the 1RS arm assembled for Zhou8425B with its counterparts in Aikang 58, Kenong 9204, Lo7, and Weining.

Pairwise comparisons revealed that the assembled Zhou8425B 1RS (317 Mb) was longer than the 1RS in Aikang 58 (282 Mb), Kenong 9204 (275 Mb), Lo7 (267 Mb), and Weining (310 Mb); although the 1RS arms we compared were largely collinear, complex chromosome structural changes were apparent, including inversions, duplications, and translocations (Figure 3A). A closer inspection revealed that the syntenic regions between the Zhou8425B 1RS and those of the four other genotypes were 275.47 Mb (Aikang 58), 274.77 Mb (Kenong 9204), 174.40 Mb (Lo7), and 122.47 Mb (Weining) in length, whereas the non-syntenic regions for the four comparisons, calculated without considering large structural changes, were 32.71, 24.80, 83.15, and 88.53 Mb in length, respectively (supplemental Table 3).

Figure 3.

Comparison of the 1RS arms of five different genotypes.

(A) Synteny and structural variations between the 1RS arm of Zhou8425B and its counterparts in two common wheat cultivars (Aikang 58 and Kenong 9204) and two diploid rye lines (Lo7 and Weining). The five 1RS arms differed in assembled length, with a non-collinear segment (∼20.5 Mb) present at the telomere end of Zhou8425B 1RS (arrow).

(B) Interdispersed arrangement of three types of tandemly repeated satellite arrays in a subtelomeric region of 1RS from Zhou8425B; the basic repeat units are Tr380, Tr118, and Tr571.

(C) Genes and gene families annotated for the 1RS arms of the five compared genotypes.

(D) Conserved and unique gene families computed for the 1RS arms of the five genotypes.

(E) Expansion of genes encoding B3 or AP2/ERF-ERF TFs (arrows) in the 1RS arm of Zhou8425B relative to those of Aikang 58, Kenong 9204, Weining, and Lo7.

(F) Expression heatmaps of the genes encoding B3 or AP2/ERF-ERF TFs carried by the 1RS arm of Zhou8425B created from transcriptome sequencing data of 12 Zhou8425B samples.

Likely large structural changes between the Zhou8425B 1RS and its counterparts in Aikang 58, Kenong 9204, Lo7, and Weining were computed using SyRI software, which detects genomic rearrangements and regional sequence differences by comparing whole-genome assemblies (Goel et al., 2019). Relative to the Zhou8425B 1RS, the AK58 1RS exhibited only one inversion, whereas the Kenong 9204 1RS exhibited four inversions and one translocation (supplemental Table 4). The 1RS sequences in Lo7 and Weining differed more extensively from that in Zhou8425B, with six inversions in the Lo7 1RS and 29 structural changes, including 24 inversions, 4 translocations, and 1 duplication, in the Weining 1RS (supplemental Table 4). The fragments involved in the structural changes varied from 0.51 to 17.52 Mb in length, although most (35/41, 85.36%) were less than 4 Mb (supplemental Table 4).

A non-collinear segment of approximately 20.5 Mb located at the telomere end of the Zhou8425B 1RS was consistently found in the comparisons (Figure 3A). This segment was primarily composed of three types of satellite array with basic repeating units of 380, 571, and 118 bp, which we refer to as Tr380, Tr571, and Tr118 hereafter (Figure 3B). Bioinformatic analysis revealed that Tr380, Tr571, and Tr118 corresponded to the previously reported pSc200, pSc250, and pSc119.2 repeat sequences (supplemental Figure 8), which were found to be enriched in subtelomeric regions of rye (Vershinin et al., 1995). Among the interdispersed Tr380, Tr571, and Tr118 tandem arrays, one of the two Tr380 arrays was located immediately proximal to the telomere of 1RS (Figure 3B), which is consistent with a past cytological study showing that pSc220 repeats were consistently detected in more distal chromosomal ends of rye (Vershinin et al., 1995). Examination of the Zhou8425B genome assembly showed that Tr380 and Tr571 arrays were specifically located toward the 1RS telomeric region with relatively high densities (>80%), whereas Tr118 repeats were present in different locations on multiple chromosomes (supplemental Figure 9A). Nevertheless, the density of Tr118 repeats was highest (>60%) toward the telomeric region of 1RS (supplemental Figure 9A). The three types of tandem repeat were also detected in the genome assemblies of Aikang 58, Kenong 9204, Lo7, and Weining, but with very low densities (<10%) (supplemental Figure 9B–9E). This may be due to insufficient assembly of complex satellite arrays in these four genotypes, as none were sequenced using HiFi long-read sequencing technology.

The numbers of HC genes and gene families differed among the 1RS arms from different genotypes (Figure 3C). A total of 440 gene families were conserved among the 5 1RS arms, with 9–46 specific gene families detected for individual 1RS arms (Figure 3D). Interestingly, the genes predicted to encode B3 or AP2/ERF-ERF TFs, which frequently regulate plant tolerance to abiotic stresses (Swaminathan et al., 2008; Feng et al., 2020), were expanded in Zhou8425B relative to Aikang 58, Kenong 9204, Lo7, and Weining (Figure 3E). The B3 gene TraesZ8425B1RS01G036200 was more highly expressed in developing grains at 15 DPA (Figure 3F), consistent with the earlier finding that this gene had numerous co-expressed partners in developing grains at 15 DPA (Figure 2D; supplemental Data S3). The AP2/ERF-ERF gene TraesZ8425B1RS099000 was expressed in multiple plant samples, especially those of immature spikes, heading-stage stems, or heading-stage immature tillers (Figure 3F).

Mapping of YrZH3BS aided by the Zhou8425B genome assembly

APR genes are highly desirable for the development of crop cultivars with durable disease resistance and superior yield performance because they exert less selection pressure on pathogen races and facilitate more efficient coordination between plant growth and defense (Hu and Yang, 2019; DeMell et al., 2023). In general, durable APR involves the action of multiple genes with partial disease-suppression effects, which makes the dissection and breeding of durable APR very challenging (Dinglasan et al., 2022). Among the 86 chromosomal loci officially recognized to control wheat YR resistance to date, most confer major-effect all-stage resistance that often collapses because of changes in pathogen races; only 26 specify APR that is partially effective but tends to be more broad-spectrum and longer lasting (McIntosh et al., 2020; Klymiuk et al., 2022; Feng et al., 2023; Zhu et al., 2023). Consequently, mining and deployment of APR genes represent a valuable strategy for controlling the damage caused by YR disease in global wheat production (Huerta-Espino et al., 2020; Kale et al., 2022; Wang et al., 2023a; 2023b; Lin et al., 2023).

By combining a genome-wide association study (GWAS) with bi-parental mapping aided by the Zhou8425B genome sequence, we identified a new APR locus (i.e., YrZH3BS) that significantly reduced YR disease severity and promoted grain yield under field conditions. The initial GWAS experiment was performed with 245 wheat cultivars grown in 5 different field environments. Flag leaf disease severity (FLDS), calculated as the percentage of leaf area covered by Pst pustules and scored on a 0 - 6 scale (supplemental Figure 10), was recorded for each line. The resulting FLDS values were used for GWAS analysis in conjunction with 612 673 SNP markers with known physical locations (supplemental Figure 11). We detected 214 non-redundant SNPs significantly associated with FLDS, which were mainly distributed in several GWAS peaks on 3BS, 3DS, 7BL, or 7DL (supplemental Data S4). We then focused on the significantly associated SNPs located in a 6.95–21.40 Mb interval of 3BS (Figure 4A). Haplotype analysis showed that the 14 associated SNPs in this block formed 3 major haplotypes, with Hap C having a significantly lower mean FLDS value (19.19; p < 0.0001) than Hap A (50.25) and Hap B (55.40) (Figure 4B). These data pointed to the presence of a new APR locus on 3BS (YrZH3BS). To verify YrZH3BS, we carried out further bi-parental mapping with 199 F_2:3 lines developed using Xinhuamai 818 (Hap C) and Yumai 1 (Hap A) as parents. Under field conditions, Xinhuamai 818, derived from Zhou8425B (Figure 4C), was highly resistant (mean FLDS = 4.50%) to YR, whereas Yumai 1 was strongly susceptible (mean FLDS = 84.17%) (Figure 4D). On the basis of FLDS values, the 199 F_2:3 lines could be divided into three groups, 42 homozygously resistant, 101 heterozygously resistant, and 56 homozygously susceptible (supplemental Table 5), which fitted a 1:2:1 segregation ratio (χ² = 2.02 < χ²_{0.05, 2} = 5.99).

Figure 4.

Discovery and fine mapping of a new APR locus (YrZH3BS) against wheat yellow rust (YR) disease.

(A) Manhattan plot showing a GWAS peak on 3BS that was significantly associated with flag leaf disease severity (FLDS).

(B) Differences in FLDS values (means ± SE) among the three major haplotypes formed by 14 SNPs located in the 3BS GWAS peak. Statistical analysis was performed using Duncan’s multiple comparison test, with significant differences (p < 0.05) indicated by different letters after the FLDS values.

(C) The YrZH3BS locus in Xinhuamai 818 was inherited from Zhou8425B.

(D) Comparison of YR resistance and FLDS between Xinhuamai 818 and Yumai 1.

Xinhuamai 818, but not Yumai 1, was highly resistant to YR in the field. Consistent with this observation, the mean FLDS value of Xinhuamai 818 was substantially lower than that of Yumai 1.

(E) Fine mapping of YrZH3BS to a 1.39-cM interval on 3BS using 10 polymorphic DNA markers developed with the aid of the Zhou8425B genome sequence. The 1.39-cM interval corresponds to a physical distance of 1.73 Mb in Zhou8425B 3BS and 0.91 Mb in CS 3BS.

(F) The percentages of FLDS variance explained by YrZH3BS in two field environments (2020PX and 2023PX).

(G) Schematic representation of the different locations of Yr58, Yr30, and YrZH3BS on the 3BS chromosome arm.

To further map YrZH3BS, we made use of the genome sequence information for Zhou8425B and Chinese Spring (CS) to accelerate the development of polymorphic DNA markers. Zhou8425B was especially valuable for this purpose, because it not only possessed durable APR to YR but also served as a parent for the breeding of Zhoumai 13, which was used to develop Xinhuamai 818 (Figure 4C). The 3BS sequences in Zhou8425B and CS were analyzed, and the resulting indels were used to develop DNA markers. After mapping in Xinhuamai 818, Yumai 1, and derivative F_2:3 lines, 10 polymorphic markers (M174, M199, M207, M210, M287, M328, M311, M344, M392, and M409, supplemental Table 6) were genetically mapped to the region associated with YrZH3BS, with M328 and M311 being the closest flanking markers (Figure 4E). The genetic distance covered by the two markers was 1.39 cM, corresponding to a physical distance of 1.73 Mb on Zhou8425B 3BS (from 18.60 to 20.33 Mb) and 0.91 Mb on CS 3BS (from 13.25 to 14.16 Mb) (Figure 4E). YrZH3BS explained 18.85%–51.60% of the variance in FLDS in two different field environments (Figure 4F).

To date, two APR genes (Yr30 and Yr58) against YR have been found on 3BS. Yr30 has recently been fine mapped to a 0.52-cM interval (Wang et al., 2024) corresponding to the 6.06–6.68 Mb region of 3BS on the basis of the CS genome sequence (Figure 4G). Yr58 was mapped to an interval flanked by the microsatellite markers GWM389 and BARC75 (Chhetri et al., 2016), which spanned the 0.81–3.40 Mb region of 3BS (Figure 4G). The physical location of YrZH3BS (13.25–14.16 Mb on 3BS, Figure 4F) clearly differed from those of Yr30 and Yr58 (Figure 4G), indicating that it is a new gene conferring APR against YR. Interestingly, the interval containing YrZH3BS on Zhou8425B 3BS carried a 240-kb insertion relative to its syntenic region on CS 3BS; this inserted fragment harbored four HC genes, two of which encoded receptor-like protein kinases (RLKs) (supplemental Figure 12), which have frequently been found to control plant resistance against different pathogens (He and Wu, 2016; Jamieson et al., 2018).

Individual and combined effects of three APR loci on YR disease severity and grain yield traits

The identification of YrZH3BS raised the question of whether positive interactions among YrZH3BS, YrZH22, and YrZH84 might reduce YR disease severity and promote grain yield traits. To investigate this question, we collected 212 common wheat cultivars developed using Zhou8425B or its early descendants (Zhoumai 13, Zhoumai 16, Zhoumai 22, and Aikang 58) as parents and grew them in two field environments. FLDS and grain yield-related traits, including spike number per plant (SNPP), grain number per spike (GNPS), thousand grain weight (TGW), and grain yield per m² (GY), were scored for each line in each environment (supplemental Data S5). The 212 cultivars were also genotyped for the presence of YrZH3BS, YrZH22, and/or YrZH84 using their respective molecular markers (supplemental Table 7), enabling the identification of cultivars carrying zero, one, two, or all three genes (supplemental Data S5).

Joint analysis of genotypic and phenotypic data revealed that the presence of YrZH3BS (n = 84), YrZH22 (n = 140), or YrZH84 (n = 72) significantly reduced FLDS (by 31.37%–55.10%) and enhanced GNPS (6.20%–8.30%), TGW (12.00%–17.31%) and GY (14.88%–22.77%) (supplemental Table 8). The three genes formed three double combinations, with YrZH3BS + YrZH22, YrZH3BS + YrZH84, and YrZH22 + YrZH84 detected in 66, 35, and 43 cultivars, respectively. These three combinations reduced FLDS by 52.08%–71.11% and enhanced GNPS (7.59%–8.49%), TGW (18.23%–24.75%), and GY (21.76%–29.62%) (Figure 5; supplemental Table 9). The triple combination of the three genes (YrZH3BS + YrZH22 + YrZH84) was detected in Zhou8425B and 25 derivative cultivars (supplemental Data S5); it strongly reduced FLDS (by 84.09%) and markedly increased GNPS (8.29%), TGW (29.76%), and GY (31.98%) (Figure 5; supplemental Table 9). Hence, the triple gene combination was most effective in reducing FLDS and increasing GNPS, TGW, and GY, followed by the double gene combinations and then the single genes. Finally, none of the individual APR genes or their combinations affected SNPP, although they consistently enhanced GNPS, TGW, and GY (Figure 5; supplemental Tables 8 and 9).

Figure 5.

Effects of double and triple combinations of three APR loci on reducing FLDS and enhancing wheat yield-related traits.

Reductions in FLDS (A) and increases in GNPS (B), TGW (C), and GY (D) were conferred by different double and triple combinations of three APR loci (YrZH3BS, YrZH22, and YrZH84). However, these combinations did not affect spike number per plant (supplemental Table 9). N indicates the number of cultivars with or without the double or triple combination of YrZH3BS, YrZH22, and YrZH84. The values compared were means ± SE. Statistical analysis was performed using Student’s t-test (∗∗∗p < 0.001). FLDS, flag leaf disease severity; GNPS, grain number per spike; TGW, thousand grain weight; GY, grain yield. FLDS was calculated as the percentage of flag leaf area with yellow rust pustules.

Founder genotype effects of Zhou8425B

As a key wheat breeding parent, Zhou8425B is very likely rich in elite trait genes that support its use in the development of numerous derivative cultivars (supplemental Table 2). We therefore investigated the founder genotype effects of Zhou8425B at the whole-genome level using genome resequencing data from Chinese landmark wheat cultivars published by Hao et al. (2020). Seven of these varieties were Zhou8425B derivatives according to their pedigrees, with 1 (Zhoumai 16), 4 (Aikang 58, Zhoumai 22, Fengdecun 5, and Xumai 35) and 2 (Bainong 201 and Bainong 4199) being generations of 1, 2, and 3 descendants, respectively (supplemental Figure 2). The years in which the 7 Zhou8425B derivatives were released spanned from the 2000s to the 2010s, whereas those for the 109 non-Zhou8425B derivatives spanned from the 1930s to the 2010s, with the 25 landraces probably being developed before the 1930s (Hao et al., 2020) (supplemental Data S6). By aligning the resequencing reads of the 141 wheat cultivars to the Zhou8425B assembly, we obtained 88 641 627 high-quality single-nucleotide polymorphism (SNP) markers on 21 chromosomes, most of which were located toward the terminal regions of each chromosome (supplemental Figure 13). Aided by these SNPs, we computed Zhou8425B-related genomic blocks (ZRGBs) in the 141 cultivars. The number of ZRGBs in the Zhou8425B derivatives varied from 34 to 50 (supplemental Data S6), with their combined size accounting for 30.63%–58.57% of the 14.75-Gb Zhou8425B genome assembly (Figure 6A). ZRGBs were also detected in the remaining 134 non-Zhou8425B derivatives, nine of which, released during the 1970s–2010s, carried more and a larger combined size of such blocks (supplemental Data S6). This finding is understandable, as Zhou8425B and the lines used to breed it are themselves the hybridization products of pre-existing wheat germplasms. Nevertheless, the number and combined size of ZRGBs were substantially lower in the 25 landraces (supplemental Data S6).

Figure 6.

Analysis of Zhou8425B-related genomic blocks.

(A) Presence of Zhou8425B-related genomic blocks (ZRGBs) in seven Zhou8425B derivative cultivars (Zhoumai 16, Zhoumai 22, Xumai 35, Aikang 58, Bainong 201, Bainong 4199, and Fengdecun 5). For each cultivar, the total size (Gb) of the ZRGBs and their proportion (%) in the Zhou8425B genome assembly are provided in brackets.

(B) Chromosomal distribution of the 13 ZRGBs shared by 7 Zhou8425B derivative cultivars.

(C and D) Analysis of purifying selection experienced by the 2 ZRGBs (Chr5B ZRGB1 and Chr5B ZRGB2) located on 5B in the 7 Zhou8425B derivative cultivars, using 25 Chinese wheat landraces as controls. In (C), the divergence between the two cultivar groups along different regions of chromosome 5B is shown. The level of nucleotide sequence diversity in Chr5B ZRGB1 and Chr5B ZRGB2 was severely reduced in Zhou8425B derivatives. In (D), chromosomal scans of nucleotide diversity (π) (top panel) and population fixation index (F_ST) (bottom panel) verified the purifying selection encountered by Chr5B ZRGB1 and Chr5B ZRGB2 in Zhou8425B derivatives.

More detailed examination revealed that 13 ZRGBs were common across the seven Zhou8425B derivatives, including the 1RS translocation and one or two fragments on chromosomes 1A, 1D, 2B, 2D, 4A, 5B, 6D, 7A, and 7D (Figure 6B; supplemental Data S7). These shared fragments carried many previously characterized wheat trait genes, such as TaPYL1, Pm8, TaPIF5, TaGRF4, TaWOX2, TaTFL1, VRN1, TaGW2, TaSus1, and TaIPA1 (supplemental Data S7). The finding of TaSus1-7A on a shared ZRGB that spanned from 80 to 89 Mb on Chr7AS (Figure 6B; supplemental Data S7) agreed well with the positive detection of this gene in Zhou8425B and five derivative cultivars in the earlier molecular marker analysis (supplemental Table 1).

Notably, most of the 13 shared ZRGBs had experienced significant purifying selection in the Zhou8425B derivatives relative to their syntenic regions in the 25 landraces (supplemental Data S8). To illustrate this point in more detail, we compared haplotype differentiation for the 2 ZRGBs on the 5B chromosome, i.e., Chr5B ZRGB1 and Chr5B ZRGB2, between 7 Zhou8425B derivatives and 25 Chinese landraces. Chr5B ZRGB1 (105–304 Mb, 199 Mb), which was fixed in the 7 Zhou8425B derivatives, was present in 4 of the 25 landraces, with alternative haplotypes in the remaining 21 lines (Figure 6C). In this region, the nucleotide diversity (π) value was significantly smaller for the Zhou8425B derivatives (7.31 × 10⁻⁴) than for the 25 landraces (1.74 × 10⁻³); the level of population differentiation as measured by the fixation index (F_ST) was also highly significant (Figure 6D; supplemental Data S8). Chr5B ZRGB2 (501–554 Mb, 53 Mb) was also fixed in the 7 Zhou8425B derivatives, but its syntenic region was replaced by alternative haplotypes in all 25 landraces examined in this work (Figure 6C). For this region, the π value of Zhou8425B derivatives (6.92 × 10⁻⁴) was much smaller than that of the 25 landraces (2.32 × 10⁻³); the estimated F_ST value was again highly significant (Figure 6D; supplemental Data S8).

Discussion

We constructed a highly contiguous and accurate genome assembly for Zhou8425B, a key founder parent for wheat breeding with many hundreds of derivative commercial cultivars. Consistent with its improved agronomic traits (supplemental Figure 1), Zhou8425B carries multiple elite QTL alleles and genes (supplemental Table 1). Four dwarfing genes (Rht1, 2, 8, and 24) may play a major role in controlling the short stature and lodging resistance of Zhou8425B, and the combination of Ppd-D1a and Rht8 probably confers early maturity and tolerance to late-season heat stress, according to a previous study (Zhang et al., 2019). The presence of TaCwi-A1, TaSus1-7A, TaSus1-7B, TaSus2-2A, TaGS-D1, TaGS5-A1, and TaGW2-6B, which have been shown to be positively associated with wheat grain weight (Khalid et al., 2019; Zhang et al., 2021; Chegdali et al., 2024), may enable Zhou8425B to develop larger grains. The presence of five APR loci (YrZH22, YrZH84, YrZH3BS, LrZH22/Lr13, and LrZH84) likely makes an important contribution to the durable resistance of Zhou8425B to YR or LR disease (see also below). The finding that 8 of the 17 genes were present in all 5 elite Zhou8425B derivative cultivars (Zhoumai 11, Zhoumai 13, Zhoumai 16, Zhoumai 22, and Aikang 58, supplemental Table 1; supplemental Figure 2) lends support to the above propositions.

Judging from the contig and scaffold N50 values (70.94 and 735.11 Mb, respectively, Table 1), BUSCO completeness score (99.5%), average nucleotide accuracy (>99.99%, supplemental Data S1), and LAI values (>15 for all three subgenomes, supplemental Figure 5), the genome assembly of Zhou8425B is of higher quality than those of other common wheat cultivars reported since 2021. This is mainly due to the integrated use of powerful HiFi long reads and effective Hi-C sequencing data, a strategy previously shown to be efficient for the development of highly contiguous polyploid genome assemblies (Wang et al., 2023a; 2023b). Consequently, the Zhou8425B assembly could resolve highly complex genomic regions such as the large and complex gluten gene loci (supplemental Figure 6; supplemental Data S2). Transcriptome sequencing analysis indicated that more than 90% of the HC protein-coding genes annotated for Zhou8425B were expressed, and their expression data are useful for the construction of co-expression networks and inference of potential hub genes (Figure 2). Together, the high-quality chromosomal sequence and relatively rich gene expression data of the Zhou8425B assembly support its use in wheat genomics and breeding studies.

By comparing the 1RS arms present in three hexaploid wheat cultivars and two diploid rye lines, we made several observations on the structural and molecular variation in 1RS in different genetic backgrounds. First, the 1RS arms varied substantially with respect to the size of their syntenic regions and the number of chromosome structural changes (i.e., inversions, translocations, and duplications). The 1RS translocations in the three hexaploid wheat cultivars were more syntenic with each other and appeared to have fewer large structural changes among them compared with the 1RS arms in diploid rye lines (supplemental Tables 3 and 4). This is consistent with a previous observation on the narrow origin of 1RS translocations used in Chinese wheat breeding (Zhou et al., 2004), thus underscoring the necessity of developing novel 1RS translocations using more divergent diploid rye donors (such as Weining, Ren et al., 2017). Second, the most common form of structural rearrangement in 1RS was probably segmental inversions (supplemental Table 4). This is in line with the finding that segmental inversions are the most frequent type of chromosome structural rearrangement in common wheat (Aury et al., 2022). Recent studies in plants and animals suggest that chromosomal inversions may facilitate population and species differentiation and adaptations to different ecological environments by suppressing meiotic recombination and maintaining multiple beneficial alleles (Harringmeyer and Hoekstra, 2022; Huang et al., 2022b). In this context, the inversion polymorphisms we observed might contribute to the function of 1RS in wheat agronomic control (after translocation into wheat) or to the environmental adaptation of diploid rye lines (e.g., Lo7 and Weining). Lastly, the 1RS arms from different sources differed in HC gene content and the number of specific gene families. Our comparative analysis suggested that the Zhou8425B 1RS contains more genes encoding AP2/ERF-ERF or B3 TFs than do the 1RS arms of Aikang 58, Kenong 9204, Lo7, and Weining (Figure 3E). More detailed microsynteny analysis verified the differential presence of these AP2/ERF-ERF and B3 TF genes among the five 1RS arms (supplemental Figure 14), indicating that the lack or reduced copy number of these genes in Aikang 58, Kenong 9204, Lo7, and Weining is unlikely to be due to poor assembly of local genomic regions. Because the B3 TF gene TraesZ8425B1RS01G036200 was strongly expressed in developing grains at 15 DPA (Figure 3F), it may provide a new target for understanding the transcriptional regulation of seed-related processes and traits. The AP2/ERF-ERF gene TraesZ8425B1RS099000 showed higher expression in immature spikes, heading-stage stems, and heading-stage immature tillers (Figure 3F). It may potentially be involved in coordinating the growth of wheat vegetative and reproductive organs at critical developmental stage(s). Therefore, further investigations of these two genes may shed new light on the contributions of 1RS to wheat improvement.

We assembled a subtelomeric region for the Zhou8425B 1RS composed mainly of three types of satellite array (Figure 3A and 3B; supplemental Figure 8). This complex region is poorly assembled in the genome sequences of Aikang 58, Kenong 9204, Lo7, and Weining (supplemental Figure 9). The three basic repeat units in this region (Tr380, Tr571, and Tr118) corresponded to the pSc200, pSc250, and pSc119.2 repeat sequences (supplemental Figure 8), which were reported nearly 30 years ago and have since been used extensively in cytogenetic studies of 1BL/1RS translocation lines as well as of the genomic and chromosomal characteristics of wheat and related species (Vershinin et al., 1995; Evtushenko et al., 2016; Guo et al., 2019; Luo et al., 2022; Zhang et al., 2022). Our result clarifies the interdispersed organization of Tr380, Tr571, and Tr118 satellite arrays, which may enable more efficient use of pSc200, pSc250, and pSc119.2 probes in future cytogenetic investigations.

Despite the observations discussed above, the analysis of 1RS structural and gene variations performed in this work is still preliminary. This is mainly because there are very few genome assemblies for common wheat varieties carrying the 1RS translocation and for diploid rye lines at present. Moreover, the genome assemblies of Aikang 58, Kenong 9204, Lo7, and Weining rye used in the current analysis were constructed using earlier sequencing technologies and therefore tend to have problems in accurate assembly of complex genomic sequences such as those in telomeres and subtelomeric regions. In the long term, as telomere-to-telomere genome assemblies become available for more 1RS-carrying common wheat varieties and diploid rye lines, more systematic comparisons of syntenic regions and orthologous genes among 1RS arms in different genetic backgrounds will be possible, which will yield a comprehensive understanding of 1RS structural changes and gene functional variations to enhance the utility of 1RS in common wheat improvement.

This work identified a new APR locus (YrZH3BS) against YR disease, mapped it to a small interval (1–2 Mb) on chromosome 3BS, and identified two candidate RLK genes for YrZH3BS with the aid of the Zhou8425B genome sequence (Figure 4; supplemental Figure 12). Pedigree and DNA marker analyses suggested that the YrZH3BS locus in the YR-resistant cultivar Xinhuamai 818 was derived from Zhou8425B (Figure 4C). Hence, Zhou8425B now has three genetically mapped APR loci (YrZH22, YrZH84, and YrZH3BS) against YR. Analyses of YR disease severity and yield-related traits of 212 wheat cultivars clearly demonstrated that the presence of one or more of the YrZH22, YrZH84, and YrZH3BS loci is associated with significantly reduced susceptibility to YR infection and enhanced yield performance, with the triple combination (YrZH22 + YrZH84 + YrZH3BS) having the most pronounced effect (Figure 5; supplemental Tables 8 and 9). Remarkably, YrZH22, YrZH84, and YrZH3BS consistently enhanced GNPS, TGW, and GY but not SNPP (Figure 5; supplemental Tables 8 and 9). We speculate that these genes may act after the wheat tillering stage to reduce YR severity, thus promoting grain number and seed weight. Further molecular cloning and functional studies of YrZH22, YrZH84, and YrZH3BS will help to clarify this question. Consistent with our work, Wang et al. (2023a, 2023b) showed that stacking of three APR genes (Yr18, Yr28, and Yr36) conferred stronger resistance to YR than single genes or their double combinations, although the effects of different combinations of Yr18, Yr28, and Yr36 on yield-related traits were not fully assessed.

Our findings are in line with the notion that durable APR is controlled by polygenes and is beneficial for both plant defense and yield performance (Hu and Yang, 2019; DeMell et al., 2023). More importantly, they offer a useful insight into the genetic basis and efficient utilization of durable APR against YR: pyramiding more APR genes is advantageous for achieving stronger disease resistance and larger increases in grain number and seed weight. Hence, Zhou8425B, along with its 25 derivative cultivars carrying YrZH22 + YrZH84 + YrZH3BS (supplemental Data S5), may be chosen as donors for breeding new wheat lines with durable APR resistance to YR disease and high yield potential. Encouragingly, we noted that YrZH22, YrZH84, and YrZH3BS could be efficiently transmitted from Zhou8425B to Zhoumai 22 and the varieties derived from it in past breeding work (supplemental Figure 15). Pyramiding of YrZH22, YrZH84, and YrZH3BS can now be efficiently achieved using DNA markers developed in previous research and our current work (supplemental Table 7).

The use of founder genotypes to accelerate cultivar development is a common practice in the genetic improvement of crops. Understanding the genetic and genomic basis of founder genotype effects is beneficial for developing better founder parents and elite cultivars through genomics-assisted trait design and breeding (Guo et al., 2018; Hao et al., 2020; Huang et al., 2022a). In wheat, substantial progress has been made in this area of research, especially with respect to identifying the genomic blocks and genes involved in the superior agronomic performance of founder parents (Hao et al., 2020; Li et al., 2023). Here, we identified 13 ZRGBs shared by 7 Zhou8425B derivative cultivars (Figure 6B; supplemental Data S7). The importance of these genomic blocks in agronomic trait control is demonstrated by the finding that most of them underwent significant purifying selection in Zhou8425B derivatives (Figures 6C and 6D; supplemental Data S8).

Interestingly, we noted the presence of one or more previously reported important wheat genes on each ZRGB (supplemental Data S7). For example, two TaPYL1 genes encoding abscisic acid receptors were present on ZRGB1 and ZRGB3, respectively; these genes have been found to regulate drought tolerance and grain yield in wheat (Mega et al., 2019; Mao et al., 2022). The two TaCKX genes located on ZRGB3 and ZRGB9 may play important roles in controlling wheat yield-related traits (Zhang et al., 2012; Szala et al., 2020). TaDA1 on ZRGB6, TaGW2 on ZRGB9, and TaSus1 on ZRGB11 have been shown to affect wheat grain weight and number (Su et al., 2011; Hou et al., 2014; Zhang et al., 2018; Liu et al., 2020; Shen et al., 2024). On the basis of their functions in rice, TaIPA1 on ZRGB12 and TaMoc1 on ZRGB13 may participate in the regulation of wheat plant growth and architecture (Gao et al., 2019). VRN1 and CBF-B12 on ZRGB8 are related to coordinated control of the vernalization response and cold tolerance in winter wheat (Kobayashi et al., 2005; Xu and Chong, 2018). Lastly, TaPPO on ZRGB5 is involved in controlling wheat flour color (Zhai et al., 2016), and TaMFT1 on ZRGB12 regulates wheat tolerance to preharvest sprouting (Liu et al., 2013). In addition, six of the ZRGB-located genes (TaGW2, TaMFT1, TaMoc1, TaPPO, TaSus1, and VRN1) were also found to be important in shaping the elite agronomic traits of the wheat founder parents studied by Li et al. (2023). Hence, the ZRGBs and the genes they carry provide insight into the founder genotype effects of Zhou8425B and may stimulate more systematic research on dissecting the genomic and molecular basis of these effects in the future, thus revealing the contributions of Zhou8425B to wheat breeding.

In summary, our analysis data suggest that the Zhou8425B assembly represents a valuable addition to international wheat genome resources. It enriches the genomic information on 1BL/1RS translocation lines in common wheat and may find wide application in the functional analysis of elite agronomic trait genes and their utilization in genomics-assisted crop improvement.

Methods

Genome sequencing and assembly

High-quality DNA was extracted from fresh leaves collected from Zhou8425B at the tillering stage and used to prepare 15-kb PacBio HiFi libraries using the SMRTbell Express Template Prep Kit 2.0 (Pacific Biosciences, CA) following the manufacturer’s instructions. After quality control, SMRTbell libraries were obtained and sequenced on the PacBio Sequel II/IIe platform (Pacific Biosciences) to obtain the polymerase reads. The PacBio SMRT-Analysis package (https://www.pacb.com) was used to filter the raw reads. High-quality HiFi reads generated by SMRTLink (version 13.0) (https://www.pacb.com/support/software-downloads) with the parameters --min-passes = 3 --min-rq = 0.99 (supplemental Figure 16) were used for genome assembly.

The draft assembly was constructed using Hifiasm (version 0.19.5) with default parameters. The following steps were executed to eliminate low-quality or contaminant sequences: (1) purge_dup (version 1.2.5) (https://github.com/dfguan/purge_dups) was used to eliminate redundancy in the HiFi reads with the following parameters -2 -T changed.cutoffs -l XXX -m XXX -u XXX, (2) sequences were aligned to the NCBI nucleotide sequence database (NT) (ftp://ftp.ncbi.nlm.nih.gov/blast/db/) to filter out contaminant sequences, (3) contig sequence depth and GC content were calculated to filter out contigs with a coverage depth below 5 or a GC content above 50%. The resulting assembly was assessed in three ways. First, HiFi reads were aligned to the assembly using minimap2 (version 2.26) (https://github.com/lh3/minimap2) with the parameters -I6G -ax map-pb -t 60, achieving a mapping rate of 99.96%. Second, on the basis of the above alignment result, samtools (version 1.17) and bedtools (version 2.2.29) were used to calculate depth and GC content in 10-kb windows. The GC-Depth was plotted using R (https://www.r-project.org/). Third, the assembly was evaluated using BUSCO (version 5.1.3) (https://busco.ezlab.org/) with the parameters -c 60 -m geno -offline.

Hi-C-assisted scaffolding

Hi-C libraries were constructed following established protocols (Rao et al., 2014; Xie et al., 2015) and sequenced on the NovaSeq 6000 platform. To obtain high-quality Hi-C reads, the raw Hi-C sequencing data were filtered in three steps: (1) removal of reads with 3 unidentified nucleotides, (2) elimination of reads that aligned to the adaptor, (3) exclusion of reads in which ≥20% bases had phred quality ≤5. The filtered Hi-C reads were mapped to the HiFi genome assembly using Bowtie2 (version 2.3.5). 3D-DNA software (version 180419) was used to obtain uniquely mapped and valid paired-end reads, which were then used to produce chromosome level scaffolds with Juicebox (version 1.11.08). HiCExplorer (version 3.7.2) was used to generate the contact map.

Genome annotation

Repetitive elements were predicted using a hybrid approach that combined de novo and homology-based methods, with different types of repeats identified and annotated as described in a previous study (Li et al., 2021b). A comprehensive approach involving ab initio prediction, protein-based homology searches, and annotation with RNA sequencing data was used to annotate gene structures (Li et al., 2021b). For functional annotation, protein-coding genes were searched against three integrated protein databases, NR, SwissProt, and eggNOG. InterPro-annotated protein domains and Gene Ontology terms were obtained from the corresponding eggNOG-mapper annotation entry for each gene. The tRNAscan-SE software (version 4.09) was used to predict tRNAs, and other types of ncRNA were annotated with the Pfam database using BLAST (version 2.12.0) with default parameters.

RNA sequencing

Total RNA was extracted from 12 Zhou8425B plant samples (supplemental Data S1) for RNA sequencing as reported previously (Li et al., 2021b). Clean reads were aligned to the Zhou8425B genome assembly using HISAT2 (version 2.5.3a), and transcript assembly was facilitated by StringTie (version 1.3.3.b). Gene expression levels, represented by FPKM values, were calculated using the R package Ballgown. Weighted gene co-expression network analysis was performed using the R package WGCNA (https://cran.r-project.org/web/packages/WGCNA/index.html), with Cytoscape (version 3.10.1) used to visualize the co-expression networks.

Detection of structural variations in 1RS

To investigate potential large structural variations among the 1RS arms of Zhou8425B, Aikang 58, Kenong 9204, Lo7, and Weining, we aligned the relevant chromosome sequences using Nummer with the parameters -g 1000 -c 90 -l 40 -t 20 with the Zhou8425B 1RS sequence as the reference. The resulting alignments were filtered using delta-filter (-r -q -l 1000) to retain one-to-one alignments with a minimum length of 1000 bp. The resulting structural variations were analyzed using SyRI (version 1.6) (Goel et al., 2019). A preliminary verification of these structural variations was performed using dot-plot analysis and alignment of HiFi long reads (Ahsan et al., 2023), which confirmed the three inversions between the 1RS arms of Zhou8425B and Kenong 9204 (supplemental Table 4; supplemental Figure 17); however, positive verification of the remaining computed structural changes by both approaches was difficult.

Investigation of TF genes and gene families

To explore the TF genes in the 1RS and 1BS chromosome arms, we extracted the protein sequences of 1RS/1BS and used the iTAK pipeline (http://bioinfo.bti.cornell.edu/tool/itak) to annotate the TFs. We then calculated the expanded TF families in Zhou8425B and examined their expression levels using transcriptome sequencing data generated from the 12 Zhou8425B samples. OrthoFinder (version 2.3.8) was used to perform gene family analysis for the 1RS arms in Zhou8425B, Aikang 58, Kenong 9204, Lo7, and Weining. The shared and unique gene families were computed using Jvenn (https://jvenn.toulouse.inra.fr/app/example.html).

Identification and fine mapping of YrZH3BS

GWAS analysis was performed using the FLDS data from 245 wheat cultivars grown in 5 different field environments with natural and artificially induced YR infections (Zhang et al., 2017). The 5 environments included two in Xindu city and 3 in Pixian city of Sichuan province, China. The 245 cultivars were genotyped using the 660K wheat SNP array (Sun et al., 2020). Bi-parental mapping of YrZH3BS made use of F_2:3 lines (30 plants per line) prepared with Xinhuamai 818 and Yumai 1 as parents. In general, plants of homozygous and resistant F_2:3 lines showed lower FLDS values (<45%), and those of homozygous and susceptible F_2:3 lines showed higher FLDS values (>55%). The heterozygous and resistant F_2:3 lines displayed various FLDS values (20%–80%), as these lines segregated for YrZH3BS, with some possessing YrZH3BS and others lacking this resistance locus. Polymorphic DNA markers were developed for the YrZH3BS region using indels revealed by aligning the 3BS sequences of Zhou8425B and CS.

Analysis of the effects of YrZH3BS, YrZH22, and YrZH84

Two hundred and twelve Zhou8425B derivative varieties were grown in 2 field environments with natural and artificially induced YR infections (Zhang et al., 2017); their FLDS data were collected as described above (supplemental Data S5). Four yield-related traits (SNPP, GNPS, TGW, and GY) were also scored for these lines (supplemental Data S5). The presence of YrZH3BS, YrZH22, and YrZH84 in the 212 cultivars was determined using DNA markers (supplemental Table 6). The resulting genotyping information was used to examine the individual or combined effects of YrZH3BS, YrZH22, and YrZH84 on reducing FLDS and promoting yield-related traits.

Investigation of 17 agronomic trait genes

The six wheat lines used for breeding Zhou8425B were examined for the presence of four dwarfing genes, one photoperiod insensitivity gene, five APR loci, and seven grain weight-associated genes by DNA marker analysis as described previously (Zhang et al., 2021; Mu et al., 2022). The PCR primers used are listed in supplemental Table 7.

Analysis of ZRGBs in Zhou8425B derivatives

This analysis used the genome resequencing data from Chinese landmark wheat cultivars published by Hao et al. (2020). In brief, the resequencing data for 141 cultivars, listed in supplemental Data S6, were downloaded from the Genome Sequence Archive of the BIG Data Center under accession number CRA001870. After read trimming and removal of adapters and low-quality sequences, high-quality paired-end reads were mapped to the Zhou8425B assembly using BWA (version 0.7.17-r1188) with default settings. Variants were computed using the HaplotypeCaller module of GATK (version 4.1). We used VCFtools to retain only biallelic SNPs with quality scores >50, a minimum allele frequency >0.05, missing data <80%, and a read depth between 5 and 50, resulting in 88 641 627 high-quality SNPs (supplemental Figure 13).

ZRGBs were identified using SNPs as described previously (Lai et al., 2010). In brief, we calculated the genomic local distance of each accession from Zhou8425B using a 1-Mb sliding window with the formula: D = $\frac{\sum \text{homozygous}\text{altered}\text{variants}}{\text{window}\text{size}}$. Windows with D values less than 1 × 10⁻⁴ (<100 SNPs/Mb) were regarded as putative ZRGBs. Adjacent putative windows were merged if their distance was less than 2 Mb. The π and F_ST values were calculated in 3-Mb sliding windows with 1-Mb steps using VCFtools.

Data and code availability

The genome assembly and transcriptome datasets of Zhou8425B generated in this work have been submitted to the NCBI database with the accession numbers PRJNA1057534 and PRJNA1059052, respectively. The genome assembly and gene annotations of Zhou8425B are also available from Figshare (https://doi.org/10.6084/m9.figshare.27220875).

Funding

This work was supported by the Ministry of Science and Technology (China) of China (2021YFF1000200), the National Science Foundation of China (32372132), the Natural Science Foundation of Henan Province (232300421033), the Science and Technology Funds of Zhengzhou City, and China Postdoctoral Funds (GZC20230727). We thank Professors Fei Lu, Zhiyong Liu, Yijing Zhang, and Jizeng Jia for constructive advice and discussions.

Acknowledgments

No conflict of interest declared.

Author contributions

G.Y., F.C., G.L., K.Z., and D.W. configured and designed the project. G.L., Y.Y., M.C., and Z.W. sequenced and analyzed the genome. K.Z., C.L., X.L., B.Z., X.S., and Z.H. performed molecular marker analysis of elite genes. S.C., Y.L., C.D., and G.Y. analyzed the effects of three APR genes on YR disease severity and grain yield. T.Z., J.Z., and G.Y. developed Zhou8425B and evaluated its agronomic traits. X.Z., J.T., T.Z., J.Z., Y.P., D.G., G.Y., J.L., and M.Z. curated the seeds of Zhou8425B derivative varieties. Y.R., X.S., and F.C. performed the GWAS experiment and fine mapping of YrZH3BS. K.Z., G.L., H.Z., and C.S. investigated gluten gene loci and the accumulation of gluten proteins in the seeds. D.G., T.Z., G.L., G.K., Y.P., Y.Y., X.Z., and D.W. obtained funds for the project. D.W., G.L., and Y.Y. designed the figures and tables and wrote the manuscript. All authors read and approved the manuscript.

Published: December 16, 2024