Marker Haplotype Construction for the Hybrid Necrosis Gene Ne2 and Its Distribution in Old and New Wheat Varieties

Abstract

Hybrid necrosis in wheat is caused by an interaction between two genes, Ne1 and Ne2, that triggers the gradual death of plant tissue. This trait affects wheat breeding as the gene Ne2 is the same as the gene Lr13 for leaf rust resistance. We have built a three-marker haplotype that consists of single nucleotide polymorphism (SNP) marker information already available on genotyping arrays for the determination of the presence and absence of Ne2. In this work, test crosses of eight bread wheat varieties with known and unknown Ne1 carriers showed that six of them possessed Ne2. We analyzed a set of wheat varieties which had partial SNPs and phenotypic data, i.e., hybrid necrosis and leaf rust reactions, using Kompetitive Allele-Specific PCR (KASP) markers previously available for Ne2. The observed haplotypes of the SNP markers RAC875_c1226_652, Ra_c4397_542, and AX-110926324 perfectly matched the KASP marker variants for Ne2 and ne2. A prediction, based on these SNP haplotypes, of the distribution of Ne2 in wheat varieties, predominantly from Germany and released between 1900 and 2024, showed that breeding steadily increased the proportion of Ne2 in the German gene pool.

1. Introduction

Hybrid necrosis is a type of postzygotic genetic incompatibility and a component of the mechanisms to prevent gene flow between different plant populations and species [1]. In wheat, hybrid necrosis occurs when the two complementary dominant genes necrosis 1 (Ne1) and necrosis 2 (Ne2) are merged through crossbreeding [2]. Ne1 and Ne2 are located on chromosome arms 5BL and 2BS, respectively [3]. The degree of necrosis of plant tissues in hybrid plants, which can go as far as the complete death of entire plants, depends on the interaction of certain alleles at these two loci: three alleles (Ne1^w—weak, Ne1^m—moderate, and Ne1^s—strong) were described for the Ne1 locus, whereas five alleles (Ne2^w, Ne2^mw—moderately weak, Ne2^m, Ne2^ms—moderately strong, and Ne2^s) were reported for the Ne2 locus [4]. Moreover, it was reported that Ne1 and Ne2 are incompletely dominant in controlling the timing and severity of necrosis [5]. Here, homozygosity at Ne2 had a much stronger effect than homozygosity at Ne1. Numerous test crosses involving wheats with Ne1^sNe1^sne2ne2 and ne1ne1Ne2^sNe2^s genotypes have shown that Ne alleles are widespread in wheat germplasm. These screenings began as early as 1963 [4] and are still ongoing [6,7].

Hybrid necrosis is a tremendous challenge regarding the gene transfer between related wheat species. Severe hybrid necrosis of pentaploid F1 hybrids from most crosses with hexaploid wheat was reported in a durum wheat program at the Plant Breeding Institute in Cambridge [8]. In a recent study [9], the frequency of hybrid necrosis in 83 triploid F1 hybrids derived from crosses between T. durum cv. ZY1286 (Ne1Ne1ne2ne2) and multiple H. villosa accessions was 69.9%. This study identified a new necrosis gene Ne-V that is independent of Ne2, but probably interacts with the durum wheat Ne1 gene. In addition, these interspecific crossings have essentially contributed to combining complementary genes for hybrid necrosis in gene pools and are responsible for failures in the breeding of elite lines.

Major advancements in understanding hybrid necrosis in wheat were made by the cloning of both genes. Ne1 encodes an α/β hydrolase, which activates an innate immunity response to induce hybrid necrosis when interacting with Ne2 [10]. The gene Ne2 was independently cloned by three research groups in 2021 [11,12,13] and encodes a nucleotide binding leucine-rich repeat (NLR) protein. These studies have also focused on the known relationship between the hybrid necrosis gene Ne2 and the high-temperature leaf rust resistance gene Lr13: genetic, mutational and transgenic analyses have shown that Lr13 and Ne2 are the same gene. Thus, Lr13 not only confers resistance to leaf rust but also contributes to hybrid necrosis when interacting with Ne1. This discovery has significant implications for wheat breeding, as it suggests that selecting for leaf rust resistance may inadvertently introduce a genetic disposition for hybrid necrosis into breeding lines. In addition, there is a functional allele of Ne2 that does not show a pleiotropic effect on leaf rust resistance [11,12].

Convenient molecular markers for the prediction of Ne alleles in wheat germplasm help breeders select suitable crossing partners and minimize the risks of hybrid necrosis in future breeding programs. In the work by Hewitt et al. [11], Kompetitive Allele-Specific PCR (KASP) markers were developed to distinguish between Ne2/Lr13, Ne2/lr13, ne2, and a non-functional gene variant found in the wheat cultivar Arina. In addition, Si et al. [13] developed a marker for generally detecting Ne2 independent of Lr13 as well as cleaved amplified polymorphic sequence (CAPS) markers for differentiating Ne2^w, Ne2^m, and Ne2^s. The aim of this study is to (a) provide the Ne2 genotype of eight current wheat varieties using test crosses and (b) to increase genotyping platform diversity by establishing a marker haplotype for the determination of the presence or absence of Ne2 using markers from whole-genome genotyping arrays (Wheat 20k and 25k Illumina Infinium XT arrays), which are tools that are preferably used by breeders. For haplotype building, we analyzed a validation set of 131 wheat varieties, for which phenotypic and single nucleotide polymorphism (SNP) marker data were partially available, using known simple PCR markers for Lr13 and Ne2. A haplotype consisting of three markers was used for the prediction of genotypes at the Ne2 locus in addition to 440 old and new wheat varieties and the analysis of the occurrence of Ne2 over time.

2. Materials and Methods

2.1. Plant Materials

Test crosses were made between the hexaploid winter wheat (WW) varieties Felix (Ne1^sNe1^sne2ne2) and Asory, Axioma, Campesino, Julius, and SU Selke. We included data from observations on the occurrence of hybrid necrosis when transferring genes from the durum wheat variety Icaro to the bread wheat varieties Absolut (WW), Quatrox (spring wheat), and Viki (WW), and from the wheat line PI 119333 (WW) to the bread wheat variety Rosatch (WW). Except for Felix, the genotypes at the Ne1 and Ne2 loci of the wheats were unknown. For cross-pollination, the two main florets of spikelets were hand-emasculated before anthesis, following the removal of the inner florets. Within 2–5 days, one anther with mature pollen was inserted into the emasculated florets. The phenotypes of the F1 hybrids were evaluated either in the field or in the greenhouse based on the necrosis grades presented in [4]. The designation of Ne2 allele also followed [4]. A diversity panel of 104 entries, for which Ne2 and/or leaf rust gene postulations were available from the literature [4,6,7,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32], was analyzed, with the known functional markers for Lr13 and Ne2, to confirm the presence of the postulated alleles. Information was also obtained from the GRIS database (http://wheatpedigree.net, accessed on 8 April 2025). An additional 27 wheats with no leaf rust and hybrid necrosis data were analyzed with these markers and were used for marker haplotype building for Ne2. For 76 of the 131 lines, whole-genome SNP genotyping data were available. The distribution of Ne2 gene variants was predicted in 440 traditional (predominantly from Bavaria) and modern (predominantly from Germany) wheat varieties.

2.2. Simple PCR Marker Analysis for Lr13 and Ne2

The CAPS marker HBAU-Lr13 used in this study was developed by Yan et al. [12]. The PCR reaction contained 80 ng genomic DNA, 0.35 µM of each primer, 0.5 U GoTaq® Flexi DNA polymerase (Promega GmbH, Walldorf, Germany), 4 µL of 5× PCR buffer, 1.5 mM MgCl₂, and 0.2 mM each of dNTPs in a total volume of 20 µL. The PCR reaction was carried out for 35 cycles at 95 °C for 10 s, 58 °C for 30 s and 72 °C for 50 s with a final step at 72 °C for 15 min. After the 431 bp-long PCR products (5 µL) were incubated for 1 h at 37 °C with 1 U HindIII (New England Biolabs GmbH, Frankfurt am Main, Germany) in a reaction volume of 20 µL, they were loaded on 1.5% agarose gels: cleaved fragments (297+134 bp) indicated the gene variant for leaf rust resistance at the Lr13 locus. Two KASP markers were taken from the study of Hewitt et al. [11]: Lr13haplo1 amplifies gene variants Lr13_ha (resistance and hybrid necrosis variant) and Lr13_hb (only hybrid necrosis variant), whereas Lr13haplo2 distinguishes the Arina variant from either Lr13_ha or Lr13_hb. Genotypes not carrying any of these variants do not amplify (we call them the No call variant). According to [11], the designation of Ne2 gene variants was as follows: A for Lr13_ha, B for Lr13_hb, and Z for Lr13haplo2. KASP primers were synthesized by biomers.net (Ulm, Germany). The 10 µL-KASP assays contained 25-50 ng template DNA and 0.14 µL PACE^TM Genotyping Assay Mix (12 µM allele-specific primer 1–FAM, 12 µM allele-specific primer 2–HEX, 30 µM common reverse primer) in 1× PACE^TM Genotyping Master Mix (3CR Bioscience, Harlow, UK). The DNA template was denatured for 15 min at 94 °C and cycled 10 times for 20 s at 94 °C and 1 min at 65 °C (drop 0.8 °C per cycle). This was followed by 34 cycles of 20 s at 94 °C and 1 min at 57 °C. Samples were cooled to 37 °C for fluorescence reading. PCR reactions for varieties that showed unexpected marker genotypes were repeated one or two times.

2.3. Marker Haplotype Construction for Ne2 Using SNP Markers from Genotyping Arrays

To allow the concurrent determination of genotypes at the Ne2 locus when using whole-genome genotyping data, we assessed three markers within and nearby the Ne2 gene (TraesCS2B01G182800, 157688966–157696282 bp, RefSeq v.1.0) on chromosome 2B [11,13]: RAC875_c1226_652 (157693607 bp), Ra_c4397_542 (157694672 bp), and AX-110926324 (157756308 bp). SNP genotypes of wheat lines were sourced from our projects, using Wheat 20k and 25k Illumina Infinium XT arrays, BRIWECS (Wheat 15k Illumina iSelect array) [33] and NIAB (wheat 90k Illumina iSelect SNP array, see acknowledgments). SNP haplotype classes were compared to KASP marker classes using a Χ² test based on a contingency table.

3. Results

3.1. Phenotypic Screens for Hybrid Necrosis

F1 plants of the cross between Felix and Julius showed strong hybrid necrosis, whereas in crosses of Felix with Asory, Axioma and SU Selke moderately strong necrotic phenotypes were obtained (Figure S1). Normal plant development was observed in F1 hybrids of the cross Felix × Campesino. F1 hybrid necrosis was strong in crosses Absolut × Icaro and Viki × Icaro (Figure S2) and resembled that of the cross Felix × Julius. No symptoms were recorded for Quatrox × Icaro. The F1 between PI 119333 and Rosatch showed weak hybrid necrosis symptoms and produced seeds (Figure S3). The strong necrotic reactions in the F1 of the crosses that involved Julius, Absolut, and Viki were characterized by the fact that the plants had only one shoot, necrotization began at the 1–2 leaf stage and death occurred at the 3–6 leaf stage. F1 plants of the crosses with Asory, Axioma and SU Selke developed a few tillers but died at the flag leaf stage.

3.2. Marker Analyses

Of the 131 wheat varieties investigated, 38 were previously postulated to carry Ne2 (

Table 1. Distribution of phenotypes and marker genotypes in 131 wheat cultivars.

Postulated Genes a	No. of Lines	HBAU-Lr13 b	Lr13haplo c	Observed Ne2 Haplotypes d
Ne2	22	R	A	AAA
	15	S	B	AAA
	1 e	S	Z	ACA/G
ne2	14	S	Z	AAA/AAG f, AA/CG, ACA
	9	S	-	AAG, ACA, AC*, *C*, GAG
	1 g	S	B	NA
Lr13 data only	27	R	A	AAA
	1 h	S	-	GA*
lr13 data only	2	S	A	AAA
	1	S	Z	NA
	4	S	-	AC*, GAG
Lr other data only	2	R	A	AAA
	1	S	Z	NA
	2	S	-	AC*
Ne2/ne2 carrier lines	3	S	-	GAG
No data available	10	R	A	AAA
	3	S	B	AAA
	6	S	Z	AAG, ACA, AC*, GA*
	7	S	-	ACA, GAG, GA*

^a The literature and this work; ^b HBAU-Lr13 = co-dominant marker with variants R (resistant) and S (susceptible); ^c Lr13haplo = A (resistance and hybrid necrosis variant) or B (only hybrid necrosis variant) determined using marker Lr13haplo1, Z (Arina variant) determined using marker Lr13haplo2,—(No call variant) determined using Lr13haplo1 and Lr13haplo2; ^d Ne2 haplotype = Allelic configuration at marker loci RAC875_c1226_652, Ra_c4397_542 and AX-110926324; ^e Niobrara; ^f Schweigers Taca (see discussion); * Not available; ^g Nimbus; ^h Cardos.

, Table S1). Twelve of the thirty-eight varieties carrying Ne2 simultaneously expressed Lr13, which was corroborated by the corresponding genotypes R and Lr13_ha (resistance and hybrid necrosis variant) of markers HBAU-Lr13 (Figure S4) and Lr13haplo1 (Figure S5), respectively (Table S1). Another 10 Ne2 carriers were postulated by PCR analysis to also carry Lr13 (Table S1). Except for the variety Niobrara, which carries the non-functional Arina variant, the remaining 15 Ne2 carriers were found to possess the only hybrid necrosis variant based on their genotype Lr13_hb for marker Lr13haplo1 (Figure S5, Table 1, Table S1). Of the 24 wheat varieties with the postulated ne2ne2 genotype, 14 carried the Arina variant and nine carried the No call variant. The variety Nimbus was an exception as it was found to carry the only hybrid necrosis variant. Forty varieties had no phenotypic data for Ne2 but did for leaf rust resistance (Table 1). The presence of Lr13 and thus Ne2 was confirmed for 27 varieties, whereas the variety Cardos postulated to carry Lr13 showed the susceptible allele for HBAU-Lr13 and the No call variant for markers Lr13haplo1 and Lr13haplo2. Leaf rust susceptible varieties (Table 1) had either the only hybrid necrosis variant (Astron and Bussard), the Arina variant (Alcedo), or the No call variant (Armada, Aron, Inspiration and Zentos) (Table S1). Varieties assumed to carry Lr genes other than Lr13 had the Arina variant (Robigus) or did not amplify (Alidos and Premio). In the varieties Hussar and Torfrida, additionally, Lr13 could be postulated by observing that the genotype R of HBAU-Lr13 and Lr13_ha was present. For Klein Aniversario, Minhardi and Zorba, Ne2 carrier and non-carrier lines of the variety are available. Here, we accessed the Ne2 non-carrier line of the varieties. Of the 26 varieties with no phenotype data at all, ten were found to possess Lr13_ha, three carried Lr13_hb, six carried the Arina variant, and seven did not amplify any of these variants (Table 1). Although the varieties Benno, Disponent, Ibis, Komponist, and Riebesel 47/51 were found to possess Lr13_hb, they did not amplify Lr13_ha/b (Table S1). For the variety Kalyansona, we observed amplification for Lr13haplo2 but not for Lr13haplo1, although it was previously shown to amplify Lr13_hb.

3.3. Ne2 Haplotype Construction

We have chosen three SNP markers, RAC875_c1226_652, Ra_c4397_542 and AX-110926324, from the Ne2 region for haplotype analysis in the validation panel. Of the 131 diverse wheat lines, there were 76 with whole-genome SNP data from different genotyping array sources. In addition to the four haplotypes AAA, AAG, ACA and GAG, we observed three incomplete haplotypes (AC*, *C* and GA*), most probably derivatives of the latter two haplotypes, and two mixed haplotypes (AA/CG and ACA/G) (Table 1, Table S1). Thirty-eight varieties that carried the AAA haplotype, always carried the resistance and hybrid necrosis variant (Lr13_ha) or the only hybrid necrosis variant (Lr13_hb) of the Ne2 gene. In contrast, varieties with the AAG, ACA and GAG haplotypes and incomplete derivatives always possessed the Arina variant or the No call variant. The two observed, mixed haplotypes, AA/CG and ACA/G, were associated with the Arina variant. Two accessions with different Ne2 haplotypes were available for the variety Schweigers Taca, including one that fits the KASP marker results of the accession investigated. Overall, SNP haplotype classes significantly fitted the KASP marker classes (df = 4, p < 0.001, Table S2) and were in accordance with the classification into Ne2 and ne2 carriers.

3.4. Ne2 Prediction in Traditional and Modern Wheat Varieties and Change in Ne2 Distribution over Time

Haplotype analysis of the 440 varieties with RAC875_c1226_652, Ra_c4397_542 and AX-110926324 predicted that 200 varieties are carriers of Ne2, and 239 varieties are carriers of ne2 (Table S3). For the variety Federation, the genotype of the Ne2 locus could not be determined. However, the marker haplotype of the advanced backcross-derived line T. dicoccum/5*/Federation also indicates that Federation is a non-carrier of Ne2. In the prediction set, 84 varieties had phenotypes for the hybrid necrosis trait. Of the 24 varieties with Ne2, 19 were predicted correctly, while there were 57 correct predictions for the 60 varieties with ne2 (Table S3).

The proportion of Ne2 carriers increased across the last century and is presented in Figure 1. Due to the low number of available varieties until 1940, we combined the decades from 1900 to 1920 and 1921 to 1940. In these time periods, the proportion of Ne2 in varieties was 25.8% and 27.6%, respectively, and it decreased to 19.4% in the decade 1941–1950. In the seven decades from 1951 to 1960 and 2011 to 2020, the proportion of varieties with Ne2 increased from 20.0% to 61.3%. The proportion of Ne2 further increased to 69.6% in the period 2021–2024.

4. Discussion

According to [4], the strong Ne1 gene variant in the variety Felix allowed us to infer the Ne2 genotype of the varieties involved in these testcrosses. The suggested genotypes are as follows: ne1ne1Ne2^sNe2^s for Julius, ne1ne1Ne2^msNe2^ms for Asory, Axioma and SU Selke, and ne1ne1ne2ne2 for Campesino. Although the Ne1 genotype of durum wheat variety Icaro is not known, we propose, due to the strong necrotic reactions of the F1 plants, that Absolut and Viki both carry genotype ne1ne1Ne2^sNe2^s, whereas the spring wheat variety Quatrox possesses ne1ne1ne2ne2. The F1 of the cross Rosatch × PI 119333 showed the slightest necrosis symptoms. Based on our genotyping results, we suggest that Rosatch harbors Ne2 and, thus, that PI 119333 harbors Ne1. But we could not determine the allele of Ne2 because the genotype of Ne1 in PI 119333 is not known. The Ne2 genotypes of the remaining seven varieties were correctly predicted with both the KASP markers and the Ne2 marker haplotypes established in this study.

The validation set consisted of 52 lines that had only phenotypes for reaction to hybrid necrosis and/or leaf rust and 24 lines that had only SNP genotypes. An overlap of phenotypic and SNP data was observed for 52 lines. Except for two lines that had neither phenotypic nor SNP data, all lines were linked by analyzing them with known markers for Ne2 and Lr13, increasing the number of lines to 76 with overlapping information for trait and SNP data for Ne2 haplotype building. This procedure not only confirmed the gene postulations from previous testcrosses with strong Ne1 carriers in most cases, but also provided new information. Ten varieties that were confirmed as carriers of Ne2 are now known also to express Lr13, and another twelve varieties that were confirmed to carry Ne2 did not carry the Ne2/Lr13 gene variant. In addition, Lr13 and thus Ne2 were newly postulated in 12 varieties, three other varieties possessed the only hybrid necrosis variant of Ne2, and 22 varieties were found not to be carriers of Ne2. In the validation panel, the varieties Niobrara, Nimbus, and Cardos showed discrepancies between phenotype and HBAU-Lr13/KASP marker alleles. This suggested that previous gene postulations were either incorrect or indicative of within-variety heterogeneity. In addition, we identified Lr13 in the varieties Hussar and Torfrida. Although these varieties have been investigated in previous gene postulation studies, Lr13 has never been detected.

We identified a haplotype for the determination of the presence or absence of Ne2 using three SNP markers that spanned the gene locus and are available on genotyping arrays. In the validation set, the prediction accuracy of the marker haplotypes was just as high as with the analysis by the KASP markers. Based on this consistent comparison result, we undertook haplotype analysis on a collection of 440 old and new varieties, of which 84 varieties again had phenotypes for the presence or absence of Ne2. We achieved 90.5% correct predictions of the genotype at the Ne2 locus in these varieties. The incorrect predictions included two varieties, Schweigers Taca and Anza, for which whole-genome genotyping data of two accessions of each variety were available. One of the two accessions of Schweigers Taca had the marker haplotype AAG that matched the ne2 result of the test cross; the other (AAA) did not. For the variety Anza, the two accessions also had different haplotypes (ACA and GAG), but none matched the Ne2 result of the test cross. In addition, Anza was also reported to carry Ne1 [24], so there must be accessions that do not carry Ne2. This observation directly shows that the presence of within-variety heterogeneity is a reason for misclassifications. Within-variety heterogeneity is common in pure line breeding (21 varieties of the prediction panel were heterogeneous for their Ne2 haplotype, Table S3) and reflects the breeders’ view that the conscious or unconscious inclusion of several closely related pure lines in a variety provides useful flexibility, but sublines can be lost during seed multiplication [34], so that the originally existing heterogeneous gene loci can no longer be addressed. The impact of within-variety heterogeneity also influences the results on leaf rust; for example, the variety Anza was reported to be heterogeneous for Lr13 [27].

We found the highest proportion of Ne2 in the most recent wheat varieties (69.6%), whereas the lowest proportions were present in the traditional wheat varieties (25.8% and 27.6%) and in the varieties derived from the initial phase of strict pure-line breeding (20.0%, decade 1951–1960) [35], which indicates that breeding from 1900 to 2024 has steadily increased the proportion of Ne2 in the released wheat varieties from Germany (Figure 1). The decrease in Ne2 carriers by 19.4% in the decade 1941–1950 marks the transition phase from traditional to modern varieties. A study on the distribution of Ne genes in German wheat varieties, released between 1961 and 1986, already reported a high proportion of Ne2 (63.8%) [22]. A similar dominance of Ne2 carriers (65.4%) was also discovered in Indian spring wheat varieties [23], whereas studies from China indicate lower but variable Ne2 proportions of 8.3% [13] and 33.2% [24]. High (56.0% [36] and 53.5% [37]) and low (14.2% [38]) Ne2 frequencies were observed in the winter and spring wheat varieties, respectively, from Russia. However, an increase in the frequency of Ne2 was also observed in China [24] and in the spring wheats from Russia [38]. These increases confirm that Ne2 is under positive selection because of its association with rust resistance [10]. Therefore, breeders should be aware of hybrid necrosis when integrating crosses with Ne1 carriers into breeding programs.

5. Conclusions

Our three SNP marker-based haplotypes show the same capacity for discriminating Ne2 from ne2 as the KASP markers developed in the study of Hewitt et al. [11] and can therefore be used with confidence. We can also assume that the steady increase of Ne2 in German wheat varieties must be associated with an adaptive advantage such as the protection against the leaf rust fungus through Lr13 that increases resistance in gene combinations. Finally, there is no reason to give up the useful genetic variation of a Ne1 carrier line that is to be crossed into a Ne2 carrier line, as there is always the possibility of achieving this transfer via a bridge cross with a non-carrier line.