Introduction

Wheat (Triticum aestivum L.) is an important staple food crops worldwide. With the increasing population and the improving standard of living, demand for wheat yield and quality has been increasing annually. However, wheat grain yield and quality are seriously threatened by nearly 50 diseases and pests (Singh et al. 2016). Powdery mildew, caused by Blumeria graminis f. sp. tritici (Bgt), is one of the most destructive foliar diseases, which generally decreases wheat yield by 10–15%, and up to 50% in severe cases (Juroszek and von Tiedemann 2012). Traditional strategies for controlling this disease include the use of host resistance, application of fungicides, and integrated disease managements. Compared with other methods, host resistance has been demonstrated as the most effective and environmentally friendly strategy to control or slow epidemics of powdery mildew.

To date, more than 100 formally and/or temporarily designated wheat powdery mildew (Pm) resistance genes/alleles have been identified (Zhu et al. 2022). However, most of them confer race-specific resistance. New virulent Bgt isolates can easily defeat such type of Pm genes. For example, Pm1, Pm2, Pm3, Pm4, Pm5, and Pm8 are losing or have almost lost their resistance to Bgt isolates currently prevalent in many wheat production regions in China (Wang et al. 2005; Xiao et al. 2022; Zeng et al. 2014). Even the most effective gene Pm21, derived from Dasypyrum villosum, is also undergoing strong selection pressure and the risk of losing its resistance (He et al. 2020). Currently, efficient utilization of disease resistance genes and rapid pyramiding of major genes are effective means to develop new disease resistance wheat varieties.

Wheat Pm genes can be divided into two categories: race-specific and non-race-specific resistance genes. The former usually confers qualitative resistance, which follows the Mendel’s Law of Segregation, and accounts for most of the documented Pm genes, especially those of nucleotide-binding site leucine-rich repeat (NBS-LRR) receptor protein-encoding genes (Wu et al. 2022a). The latter confers mostly partial quantitative resistance. Multiple fungal pathogen resistance genes Pm38/Yr18/Lr34/Sr57, Pm39/Lr46/Yr29/Sr58, and Pm46/Yr46/Lr67/Sr55 also provide the quantitative resistance (Krattinger et al. 2009; Lillemo et al. 2008; Moore et al. 2015). Qualitative and quantitative resistance genes together constitute the genetic basis against powdery mildew. The qualitative resistance genes endow plants with resistance in a manner of ‘gene for gene,’ which are generally believed to provide a relatively high level of resistance but the limited species of pathogens (Alemu et al. 2021; Limpert et al. 1987). Quantitatively inherited genes usually provide more durable but general levels of resistance, and they often deliver partial resistance to adult plants, and the genetic analysis is more complicated (Spielmeyer et al. 2005). Besides, according to the stages that the resistance expresses, Pm genes could be categorized as all-stage resistance (ASR) genes and adult-plant resistance (APR) genes. Most of the Pm genes confer ASR, except for Pm38, Pm39, Pm46, and Pm62; Pm6 exhibits a moderate immunity at the seedling stage and high resistance from the fourth leaf stage onward. In addition, two wheat-D. villosum translocation lines harboring Pm55 and Pm67 confer the tissue-differentiated ASR resistance (Fig. 1).

Fig. 1
figure 1

Resistance stages of Pm genes. This figure shows the resistance period of the officially named Pm gene. ASR: all-stage resistance; APR: adult-plant resistance

Full size image

Exploration and location of Pm genes

From the early genetic research in the 1930s to the modern gene mining based on high-quality genomes of common wheat and its relatives in the last decade, mining of wheat Pm genes has entered a new era (Avni et al. 2017; IWGSC 2014, 2018; Li et al. 2022b; Ling et al. 2018; Luo et al. 2017; Waterhouse 1930; Xiao et al. 2022). Pm genes or alleles have been identified from species in genera Triticum, Thinopyrum, Secale, Aegilops, and Dasypyrum. Most Pm genes were derived from common wheat. Besides, wheat relatives also contribute a large proportion of these Pm genes. For instance, Pm40 and Pm43 and other three temporarily named Pm genes are derived from Thinopyrum spp.; five Pm genes (Pm7, Pm8, Pm17, Pm20, and Pm56) are identified from S. cereale; Ae. tauschii is the diploid progenitor of hexaploid wheat providing the D subgenome, and 10 Pm genes originate from four Aegilops species; D. villosum is another excellent resistance resource, and Pm21/PmV, Pm55/Pm5V, Pm62, and Pm67 have been introgressed from this species into common wheat in forms of compensating Robertsonian translocations (Chen et al. 1995, 2006; Zhang et al. 2016, 2018b, 2021b). Modern genomic and genetic techniques have greatly accelerated mining of the Pm genes in the last years. Research progresses of the Pm genes in Triticeae crops are introduced in the following (Fig. 2).

  1. (1)

    Pm genes on the homoeologous group 1 chromosomes. These Pm genes include the Pm3 alleles (Bourras et al. 2019; Zeller et al. 1993), Pm10 (Tosa et al. 1987), Pm24, Pm25 (Shi et al. 1998), Pm28 (Peusha et al. 2000), Pm39 (Lillemo et al. 2008), and MlHLT (Wang et al. 2015) from Triticum Spp., Pm8 (Hurni et al. 2014), Pm17 (Singh et al. 2018) and PmSESY (He et al. 2021a) from Secale spp., Pm32 (Hsam et al. 2003) from Ae. speltoides, and Pm67 (Zhang et al. 2021b) from D. villosum. Pm67, a dominant Pm resistant gene, is derived from a wheat-D. villosum introgression line and physically located in the distal bin (FL 0.70–1.00) of 1VS#5. The introgression line harboring Pm67 is immune to powdery mildew at the seedling stage and sheaths, stems, and spikes at the adult-plant stage and meanwhile a low level of susceptibility without the formation of conidiophores although a spot of colonies in leaves at adult-plant stage. Another wheat-D. villosum translocation line T5V#4S·5DL harboring Pm55 shows a nearly different immune reaction at the adult-plant stage from Pm67 (Zhang et al. 2016). When the Pm55 translocation line was crossed to the Pm67 donor, all the F1 plants showed immunity at the adult-plant stage. This suggested that the combination of the two genes might provide comprehensive powdery mildew resistance in wheat, and these two genes could become the valuable germplasms for adult-plant resistance to powdery mildew. PmSESY, derived from the rare rye wild species S. sylvestre, is identified as a dominant Pm gene and confers a broad-spectrum resistance to different Bgt isolates. He et al. (2021a) mapped PmSESY to the 1.88-cM genetic interval at the end of chromosome 1RL.

  2. (2)

    Pm genes on the homoeologous group 2 chromosomes. The second homoeologous chromosome group has the largest number of Pm genes, including Pm4a (Mcintosh and Bennett 1979), pm50 (Mohler et al. 2013), and Pm49 (Ml5323) (Piarulli et al. 2012) from T. turgidum ssp. dicoccum, Pm4b (Mcintosh and Bennett 1979), Pm33, and PmPS5A (Zhu et al. 2005) from T. carthlicum, Pm4c (Hao et al. 2008), Pm4e (Li et al. 2017), Pm52 (Zhao et al. 2013), Pm63 (Tan et al. 2019), Pm65 (Li et al. 2019), PmHNK54 (Xu et al. 2011), pmX (Fu et al. 2013), pmQ (Li et al. 2020e), PmKN0816 (Wang et al. 2021), PmLS5082 (Wu et al. 2022a), PmCG15-009 (Zhang et al. 2023), and PmYN99102 (Mu et al. 2022) from T. aestivum, Pm4d from einkorn wheat T. monococcum (Schmolke et al. 2012), Pm6 from T. timopheevii (Helmsjϕrgensen and Jensen 1973), pm26 (Rong et al. 2000), pm42 (Hua et al. 2009), Pm64 (Zhang et al. 2019), MlIW170 (Liu et al. 2012), MlWE74 (Zhu et al. 2022), MLZec1 (Mohler et al. 2005), and MlIW39 (Qiu et al. 2021) from wild emmer wheat T. turgidum var. dicoccoides, Pm43 (He et al. 2009), PmL962 (Shen et al. 2015), and PmWE99 (Ma et al. 2016) from Th. intermedium, Pm51 from Th. ponticum (Zhan et al. 2014), Pm57 from Ae. searsii (Liu et al. 2017), Pm58 from Ae. tauschii (Wiersma et al. 2017; Xue et al. 2022), Pm62 from D. villosum (Zhang et al. 2018b), Pm68 from T. durum (He et al. 2021b), and MlHubel from T. aestivum ssp. spelta (syn. T. spelta) (Peng et al. 2014).

Fig. 2
figure 2

The distribution, position, origin, and existence form of the wheat powdery mildew resistance genes. The chromosomes of common wheat are represented by a cylindrical shape with black border, and the chromosomes from other wheat relatives are represented by a cylindrical shape with blue border. The black dot in the middle of cylindrical shape represents the centromere. The orange rectangles on chromosomes represent the genetic distance between Pm genes and the closely flanking marker. The number on the left side of chromosomes represents the physical location, and the number on the right side represents genetic distance. *: Pm genes encoding NBS-LRR protein; △: Pm genes encoding kinase; and ○: Pm genes encoding transporter. The map was performed based on the information from Chinese Spring RefSeq v1.0 chromosomes, The International Wheat Genome Sequencing Consortium (IWGSC)

Full size image

Among these Pm genes, six were mapped on the chromosome arm 2AL, including three formally named genes Pm4, pm50, and Pm65, and the other three temporally named genes PmPS5A, PmHNK54, and pmX. pm50 is located in the distal end of the marker Xgwm294, and the pm50 donor maintains high level of powdery mildew resistance at the adult-plant stage (Mohler et al. 2013); Pm65 is a dominant gene located between the markers Xstars355 and Xstars356 (Li et al. 2019); PmPS5A and pmX may be allelic to Pm4 (Fu et al. 2013; Zhu et al. 2005); PmHNK54 is located on chromosome 2AL-1-0.85 and tightly linked to markers Xbarc5 and Xgwm312 (Xu et al. 2011).

Nine Pm genes were identified on the chromosome arm 2BS, including the formally named genes pm26, pm42, and Pm68 and the temporarily named genes Ml5323, MlIW170, PmL962, PmWE99, MlWE74, and MlIW39. Pm68 was identified from resistant Greek durum wheat TRI 1796 and flanked by markers Xdw04 and Xdw12/Xdw13 with the corresponding physical region of 16.2–24.8 Mb (He et al. 2021b); pm26 and pm42 were recessive genes derived from wild emmer wheat (WEW) and molecular marker analysis indicated that they were not located at the same locus (Hua et al. 2009); MlIW170 is located in 2BS3-0.84–1.00 and pm26 and MlIW170 appear to be allelic (Liu et al. 2012); MlWE74 is located in the genomic region of 799.9 kb of the Zavitan reference sequence (Zhu et al. 2022). Whether comparing the resistant spectrum of pm26, MlIW170, MlWE74, or using the alleles of co-segregated marker WGGBD425 and flanking markers WGGBD412 and WGGBH346, it is difficult to identify their positional relationship, and therefore, it needs further evidence to distinguish them.

The chromosome arm 2BL harbors the largest number of Pm genes, including Pm6, Pm33, Pm51, Pm52, Pm63, Pm64, pmQ, MlZec, PmKN0816, PmLS5082, PmCG15-009, and PmYN99102. Pm6 was located on the wheat-T. timopheevii 2B/2G introgression. In 2016, TaRLK was first considered to be the candidate gene of Pm6; however, the follow-up study indicated that it was not located on the 2G segment in the introgression line IGV1-465, which was excluded (Chen et al. 2016). Due to severe recombination inhibition between chromosomes 2B and 2G, genetically mapping Pm6 was very difficult, and Pm6 was mapped and flanked by markers CIT02g-20 and CIT02g-18 (Wan et al. 2020); Pm52, derived from the wheat cultivar Liangxing 99, was initially localized in the 2BL2-0.36–0.50 region and later narrowed to the interval between the markers Xicssl326 and Xicscl795 (Wu et al. 2019); pmQ is a recessive gene located between markers Xicsq405 and WGGBH913, and its physical interval overlapped those of Pm51 and Pm63, which needs to be distinguished from each other (Li et al. 2020e).

A few Pm genes were identified on chromosome 2D. MlHubel is the first recessive Pm gene identified on the chromosome arm 2DL, and Pm43 is another Pm gene on this chromosome arm with the dominant inheritance and located between markers Xwmc41 and Xbarc11 (He et al. 2009; Peng et al. 2014). Pm58 is located on the chromosome arm 2DS and has been narrowed to a 0.22-cM interval between markers Xsts20220 and Xkasp61553. AET2Gv20068500 is considered to be the candidate gene for Pm58, which encodes a typical NLR protein (Wiersma et al. 2017; Xue et al. 2022).

In addition, Pm57 was derived from Ae. searsii and transferred into common wheat in the form of T2BS·2BL-2SS#1L translocation. Pm57 was then mapped in a 5.13-Mb physical region between markers X67593 (773.72 Mb) and X62492 (778.85 Mb) (Dong et al. 2020; Liu et al. 2017). Pm62 was derived from D. villosum and transferred into the common wheat in the form of T2BS·2 V#5L translocation, and it was a novel adult-plant resistance gene (Zhang et al. 2018b).

In summary, many formally and/or tentatively named Pm genes have been found on the second homoeologous chromosome group, suggesting that it is most likely an enrichment region for the Pm genes. Cloning these genes and exemplary genomic assembly of this homoeologous group could clarify this issue.

  1. (C)

    Pm genes on the homoeologous group 3 chromosomes. The third homoeologous chromosome group contains fewer Pm genes, including Pm44 on chromosome 3A (Alam et al. 2011), Pm13 from the translocated chromosome 3S of Ae. longissima (Ceoloni et al. 1992), and Pm41 on chromosome 3BL (Li et al. 2009); besides, TaRPP13L1-3D on chromosome 3D was reported to respond to powdery mildew (Zhang et al. 2021a). Pm41 is derived from wild emmer accession IW2, encoding a typical NBS-LRR protein, and the other Pm genes on the third homoeologous chromosome group have not been cloned (Li et al. 2020c). Recent studies showed that TaRPP13L1-3D is a truncated CC-NB-ARC gene, and its overexpression enhances the resistance to Bgt isolates and up-regulates the expression of defense-related genes. Interestingly, TaRPP13L1-3D is located in the plasma membrane and nucleus, and migration was detected from the plasma membrane to the nucleus (Zhang et al. 2022). The RanGAP2-WPP complex is the plant-specific nuclear pore complex mediating the connection between nucleoplasm and cytoplasm (Zhou et al. 2012). TaRPP13L1 directly interacts with the WPP domain-containing protein 1 (TaWPP1), while it does not interact with RanGAP2L in the yeast two-hybrid (Y2H) system (Zhang et al. 2022). This implies the variable interaction in NBS-LRR proteins conserved domains. So, TaRPP13L1 should be an essential component for plant basal defense, probably via migration from the plasma membrane to the nucleus, although the molecular function remains to be elucidated.

  2. (D)

    Pm genes on the homoeologous group 4 chromosomes. The fourth homoeologous chromosome group contains several Pm genes, including pm61, MLIW30, and PmPBDH on the chromosome arm 4AL (Geng et al. 2016; Liang et al. 2022; Sun et al. 2018); Pm66 on the translocation chromosome T4SlS·4BL of Ae. longissima (Li et al. 2020a); and Lr67/Yr46/Sr55/Pm46/Ltn3, a multi-resistance locus on the chromosome arm 4DL (Herrera-Foessel et al. 2014; Moore et al. 2015). QPm.osu-4A, QPm.uga-4A, and QPm.tut-4A were reported in the same interval as MLIW30, but they were distinct loci (Chen et al. 2009; Geng et al. 2016; Hao et al. 2015; Jakobson et al. 2012). The recessive gene pm61 from Chinese wheat landrace Xuxusanyuehuang (XXSYH) is also located on 4AL-0.8–1.00 as MLIW30 and QPm.uga-4A, but they should be different from each other (Hao et al. 2015; Jakobson et al. 2012; Sun et al. 2018). Further studies revealed that pm61 is localized at a genetic interval of 0.71 cM, corresponding to the Chinese Spring reference genome of 0.61 Mb, between MlIW30 and QPm.tut-4A (Hu et al. 2019). In addition, a wheat-rye line WR41-1 carrying T4BL·4RL and T7AS·4RS translocations is resistant to powdery mildew, which implies that WR41-1 may contain a new Pm gene (An et al. 2013).

  3. (E)

    Pm genes on the homoeologous group 5 chromosomes. Many Pm genes are located on the fifth homoeologous chromosome group, including pm2026 on the chromosome arm 5AL (Xu et al. 2008); Pm36 (Blanco et al. 2008), Pm53 (Petersen et al. 2015), MIWE4 (Zhang et al. 2015), MI3D232 (Zhang et al. 2010), and PmAS846 (Xue et al. 2012a) on the chromosome arm 5BL; Pm16 (Chen et al. 2005; Reader and Miller 1991) and Pm30 (Liu et al. 2002) on the chromosome arm 5BS; Pm34 and Pm35 on the chromosome arm 5DL (Miranda et al. 2007, 2006); Pm48 (reconfirmed as Pm2) (Wu et al. 2023), Pm2 and its alleles on the chromosome arm 5DS of common wheat and Ae. tauschii (Briggle 1969; Fu et al. 2017; Ma et al. 2015a; Ma et al. 2014; Ma et al. 2015b; Sun et al. 2015a; Sun et al. 2015b; We, 2014; Xu et al. 2015); Pm7 on the translocated chromosome T4BS·4BL-5RL (Heun and Friebe 1990); and Pm55 on the chromosome arm 5VS of D. villosum (Zhang et al. 2016).

Pm2 is one of the most widely used Pm genes in wheat production. Many Pm2 alleles have been reported in wheat cultivars, but they are all the Pm2a allele (Jin et al. 2022; Ma et al. 2015a; Yu et al. 2022). To identify the new Pm2 allele, a large number of Ae. tauschii accessions were screened, and Pm2d–Pm2j were then identified (Manser et al. 2021). pm2026 is a recessive Pm gene and the first tentatively named resistance gene on chromosome 5A from T. monococcum (Xu et al. 2008). WEW is an essential germplasm for wheat powdery mildew resistance, and Pm16, Pm30, Pm36, PmAS846, Ml3D232, and MIWE4 are all derived from WEW. Recently, Pm36 was finely mapped and located in the genetic interval of 1.5 cM. Thirteen protein-coding genes were found in the Pm36 region, and three of them had no homologous components in the WEW Zavitan v2.0 reference genome. Among the remaining 10 protein-coding genes, six encode the BTB/POZ (broad-complex, tram track, and bric-abrac/poxvirus and zinc finger) protein. Recent studies indicate that BTB/POZ plays an essential role in plant growth and development, and the gene encoding BTB/POZ protein may be a potential candidate gene for Pm36 (Nigro et al. 2022; Shalmani et al. 2021). In addition, by comparing the physical interval of Pm36 and Ml3D232, it was found that they were not the same gene, but it was difficult to determine whether Pm36, PmAS846, and MlWE4 were alleles. PmAS846 is located in bin 5BL14 (0.75–0.76) (Xue et al. 2012a). Using weighted gene correlation network analysis (WGCNA), specific genes induced by Bgt invasion were divided into three modules (black module, darkolivegreen4 module, and plum1 module) (Hu et al. 2020). Candidate genes in the PmAS846 region and protein–protein interaction networks were also constructed. This study provided the new idea and methods for the discovery of Pm genes.

  1. (F)

    Pm genes on the homoeologous group 6 chromosomes. Several Pm genes are found on chromosome of this homoeologous group, including PmGZ1 on chromosome 6A (Li et al. 2022a); Pm54 and Pm69 (PmG3M) on 6BL (Hao et al. 2015; Xie et al. 2012b); Pm11 and Pm14 on 6BS (Tosa and Sakai 1990; Tosa et al. 1988); and Pm45 (Ma et al. 2011) and Pm07J126 (Yu et al. 2012) on 6DS. In addition, there are five Pm genes from relative species in forms of chromosome translocations: Pm12 (T6BS-6SS·6SL) (Jia et al. 1996), Pm20 (T6BS·6RL) (Friebe et al. 1994), Pm21 (T6AL·6 V#4S) (Chen et al. 1995), Pm27 (6B/6G) (Järve et al. 2000), and Pm56 (T6AL·6RS) (Hao et al. 2018). Among these genes, Pm11, Pm14, Pm45, Pm54, and PmGZ1 are all derived from T. aestivum. PmGZ1 was found in the wheat cultivar Guizi 1 (GZ1). By genotyping-by-sequencing (GBS) and QTL analysis, PmGZ1 was localized on chromosome 6A, within the confidence interval of chr6a-307802221 and chr6a-309885836, corresponding to the genetic interval of Chinese spring 58.6–60.9 cM (2.3 cM) (Li et al. 2022a). Pm56 was derived from Qinling rye. Chromosomes 6A and 6R dimonomer individuals are developed by infiltrating the rye 6R chromosome into wheat. Further analysis of hybrid line harboring incomplete forms of 6R revealed that Pm gene(s) is present on both arms of chromosome 6R; meanwhile, the presence of a short arm can give complete immunity, but the presence of the long arm can only confer a moderately to highly resistant level (Hao et al. 2018).

  2. (G)

    Pm genes on the homoeologous group 7 chromosomes. Many Pm genes have been reported on the chromosome in this homoeologous group, including Pm1 (and its alleles) (Hsam et al. 1998), pm9 (Fiona 1984), Pm37 (Perugini et al. 2008), Pm59 (Tan et al. 2018), Pm60 (and its alleles) (Ji et al. 2008; Li et al. 2021b; Wu et al. 2022b, 2021a; Zou et al. 2022, 2018), Mlm80 and Mlm2033 (Yao et al. 2007) on the chromosome arm 7AL; Pm5 (and its alleles) on the chromosome arm 7BL (Heun and Fischbeck 1987; Hsam et al. 2001; Huang et al. 2003; Lebsock and Briggle 1974); Pm40 (Luo et al. 2009) and pm47 (Xiao et al. 2013) on the chromosome arm 7BS; PmH on the chromosome 7B (Zhou et al. 2005); Pm29 on the chromosome arm 7DL (Zeller et al. 2002); and Pm15 (Tosa and Sakai 1990), Pm19 (Lutz et al. 1995), and Pm38 (Dyck 1977) on the chromosome arm 7DS.

Pm15, pm9, Pm29, pm47, Pm59, and PmH are all derived from T. aestivum. Among the rest, pm47 is a gene on a new unique locus by marker correlation alignment with Pm40 on 7BS and Pm5e, mlxbd, PmH, and Mlmz on 7BL. The EST marker BE606897 and the microsatellite marker Xgwm46 were located on both sides of the single intransigent gene pm47, and the genetic distances were 1.7 cM and 3.6 cM, respectively (Xiao et al. 2013). Pm59, a dominant gene, was mapped between STS markers Xmag1759 and Xmag1714, with genetic distances of 0.4 cM and 5.7 cM, respectively (Tan et al. 2018).

Pm40 is derived from a wheat-Elytrigia intermedium translocation line. After twice gene mapping, a full-length TraesCS7B01G164000 encoding an NBS-LRR protein was cloned in the Pm40 region by analyzing transcriptional data and identifying gene expression differences between the resistant donor and the susceptible line (Luo et al. 2009; Zhong et al. 2016). Co-isolation analysis by SNP of TraesCS7B01G164000 in L658- and L958-based markers and qRT-PCR analysis both suggested that this gene is a candidate for Pm40 (Yang et al. 2021).

Cloning and functional analysis of the Pm genes

Understanding the molecular nature of the Pm genes is a prerequisite for rational application of them to fight powdery mildew. Although more than 100 Pm genes/alleles have been documented, only a dozen of them have been cloned so far, and most of the cloned Pm genes encoded NBS-LRR proteins, including Pm1a, Pm2, Pm3b, Pm5e, Pm8, Pm12, Pm17, Pm21, Pm41, Pm60, and Pm69. Pm4b and Pm24 encode kinase-MTC and tandem kinase, respectively, and both Pm38 and Pm46 encode transporters conferring a quantitative resistance.

At present, methods for cloning Pm genes mainly include map-based cloning, target sequence enrichment sequencing (TEnSeq) including mutagenesis chromosome flow sorting and short-read sequencing (MutChromSeq) and mutagenesis and resistance gene enrichment sequencing (MutRenSeq), homologous-based cloning, core genome targeting sequencing, re-sequencing, and genome-wide association study (GWAS). Most cloned Pm genes used the map-based cloning strategy, including Pm1a, Pm2b, Pm3, Pm5e, Pm21, Pm24, Pm38, Pm41, and Pm60. However, the map-based cloning relies on a large segregation population and acceptable mapping interval, often needs extensive efforts, and spends many years. Moreover, common wheat is an allohexaploid species, with a large genome (exceeding 16 Gb) and a large number of repetitive sequences, which increases the difficulty of map-based cloning. In recent years, target sequence enrichment and re-sequencing techniques emerged, which accelerated cloning of wheat Pm genes. MutChromSeq was applied to the cloning of Pm2 and Pm4b. MutChromSeq is independent of recombination and based on the purification of individual chromosomes from susceptible mutants, making it convenient, fast, and highly accurate. MutRenSeq is based on the precise capture of NLR-targeted sequences and therefore is only available for NLR resistance genes. Pm1a and Pm21 are identified by MutRenSeq. Homologous cloning is according to the reference genome of related species to design target gene primers for amplifying the target gene, which have resulted in the isolation of Pm8 and Pm17. In addition, more precise genome sequencing technology, such as long-read Oxford nanopore technology (ONT) and core genome-targeted sequencing (CGT-Seq), combined with bulked segregant RNA sequencing (BSR-Seq) or short-read transcriptome sequencing (RNA-Seq), facilitates cloning of the Pm genes, which has been used in isolating Pm69 and MlWE18.

Here, we provide an overview of the Pm genes that have been cloned, summarize their discovery, location, cloning, and functional analysis, and highlight the recent advances in the molecular function analysis of these Pm genes.

NBS-LRR protein

Plant NBS-LRR receptors are typical plant pathogen-resistant proteins that can recognize the pathogen effectors inducing effector-triggered immunity (ETI). NBS-LRR protein consists of a carboxyl-terminal LRR domain, a central NB domain that binds ATP or ADP, and an amino-terminal Toll, interleukin-1 receptor, resistance protein domain (TIR), or coiled-coil (CC) domain (Dodds and Rathjen 2010). However, in monocot plants, including wheat, the N-terminal domain of NBS-LRR proteins appears to be restricted to the CC domain, but not the TIR domain (Araújo et al. 2019). According to the resistant genes that have been identified in cereal crops at present, 85% of wheat race-specific resistance genes encode NLRs (Sánchez-Martín and Keller 2021). As follows, we will introduce the progress of Pm genes that belong to the NLR family and propose the further research direction of NLR protein and its application in resistance breeding.

Pm1

Pm1 is the first officially cataloged Pm gene in wheat, and it has been studied since the 1950s, combined with the leaf rust and stem rust resistance genes Lr20 and Sr15 (Waterhouse 1952; Watson and Luig 1966). More Pm1 alleles, including Pm1a–e, were identified in T. aestivum, T. monoccum, and T. spelta; and Pm18 and Pm22 were also identified as a member of the Pm1 cluster and identical to Pm1c and Pm1e, respectively (Hsam et al. 1998; Singrun et al. 2003). For decades, great efforts have been made to finely map and clone the Pm1 locus (Alam et al. 2011; Worthington et al. 2014; Yao et al. 2007). In 2021, Pm1a was successfully cloned using MutRenSeq, and its avirulence gene AvrPm1 was also isolated simultaneously using the strategy of map-based cloning combined with whole-genome re-sequencing (Hewitt et al. 2021). Stable transformation of Pm1a confers a strong resistance in susceptible cv. Fielder and heterologous co-expression of Pm1a and AvrPm1 proved their receptor–effector relation. Interestingly, through phylogenetic analysis and the terminal chromosome arm 7AL rearrangement, Pm1a is closer to D. villosum-derived Pm21 at the evolutionary level.

After long-term popularization in production, a dozen of hypothetical Pm1 alleles were reported, including Pm1a–1e, MlAG12, Mlm2033, Mlm80, PmTb7A.2, HSM1, MlUM15, PmU, Pm37, Pm59, pm9, and mlRD30; however, whether allelic or distinct should be questioned because of the suppressed recombination in this region (Hewitt et al. 2021; Liang et al. 2016; Perugini et al. 2008; Tan et al. 2018). The cloning and identification of Pm1a in a highly chromosome-rearranged region provide a new strategy for the isolation of heterologous resistance genes. Moreover, we speculate that it could be possible to explore more Pm genes in this region with the ancient resistance cluster.

Pm2

Pm2 was officially named in the 1950s (Briggle 1966; Pugsley and Carter 1953) and then located on the chromosome arm 5DS in the 1970s (McIntosh and Baker 1970). Later, molecular markers linked to Pm2 were continuously developed and utilized, including the RFLP marker Xbcd1871, SSR markers Xcfd81, Xgwm190, Xcfd18, and Xgwm114, and the RAPD marker OPP15(900) (Ma et al. 1994; Qiu et al. 2006; Wang et al. 2004). It is also a multi-allelic locus including several alleles, such as Pm2a–2j. Through MutChromSeq, Pm2a was cloned and encoded a typical NLR protein (Sánchez-Martín et al. 2016). Through further homologous sequence analysis, the Pm2 alleles in the common wheat were all Pm2a haplotype (Chen et al. 2019), but in the Ae. tauschii, several new alleles were identified, including Pm2d–2j (Manser et al. 2021). Among the new Pm2 alleles, Pm2e, Pm2f, Pm2i, and Pm2j have higher similarity with Pm2a, while Pm2d, Pm2g, and Pm2h have considerable variation compared with Pm2a. Recently, the function of Pm2b was further verified by barley stripe mosaic virus-induced gene silencing, and PM2b could self-associate and interact with the transcription factor TaWRKY76-D, which was proven to regulate wheat powdery mildew resistance negatively (Jin et al. 2022).

Among the documented Pm genes, Pm2 is one of the most widely used Pm genes in wheat breeding. Only in our lab, 33 wheat cultivars/breeding lines were identified to carry Pm2 alleles from 641 wheat genotypes. Unfortunately, these alleles were also confirmed to be the Pm2a haplotype, which implies a serious simplification trend of Pm2 alleles in wheat production. We have reasons to believe the new alleles Pm2d–2j could enrich the genetic diversity of this locus and have high breeding value once they were introgressed in wheat background.

Pm3/8/17

Pm3, located on the chromosome arm 1AS, formed the largest allelic series of Pm genes, which include about 17 alleles (Pm3a–g; Pm3k–Pm3t) (Bhullar et al. 2010a; Bourras et al. 2019; Yahiaoui et al. 2006). Pm3 is the first cloned Pm gene, and it also encodes an NBS-LRR protein. According to the sequence conservation of the A genome among the diploid T. monococcum (AmAm genome), the tetraploid T. durum (AABB genome), and the hexaploid wheat, Pm3b was the first cloned allele in locus Pm3 (Yahiaoui et al. 2004). The molecular approaches used for the Pm3b candidate gene confirmation contained bacterial artificial chromosome (BAC) libraries-based physical mapping from diploid and tetraploid wheats, genetic mapping in hexaploid wheat, γ-ray mutagenized population sequencing, southern blot analysis, chromosome walking, and 3'RACE-PCR. Functional validation of the candidate gene was also carried out. Infection with a fungal isolate followed by transient co-bombarding into epidermal cells of pUbi:GUS and the wild-type or mutant Pm3b concluded that the candidate Pm3b confers the AvrPm3b-dependent resistance (Yahiaoui et al. 2004).

The successful identification of Pm3 benefits from the conversion and difference among Triticeae genomes. Although all Pm3 alleles showed sequence similarity and CC/NBS-domain conservatism, if other Pm3 alleles are true or encoded by other resistance gene-like sequences closely linked on chromosome 1A is remained to be determined (Bhullar et al. 2010b). It is reported that the combination of different Pm3 alleles will suppress the Pm3-mediated resistance (Stirnweis et al. 2014). Different Pm3 alleles usually display different polymorphisms in the NBS and/or the LRR domains, thus determining various signaling activities and recognition spectra (Stirnweis et al. 2014). For example, only two polymorphic residues are sufficient to convert susceptible PM3CS to the functional PM3E; similarly, a single residue shared in PM3B and PM3C also plays a decisive role in the powdery mildew resistance (Brunner et al. 2010). A recent study demonstrated the importance of naturally occurring polymorphic residues in the LRR domain to the PM3 function in protein abundance, AVR recognition, and immune response activities (Lindner et al. 2020). In summary, a better understanding of molecular structure and function will support the artificial NLR receptor improvement and molecular breeding in agricultural application like the Pm3 locus.

Pm8 and Pm17 are located on chromosome arm 1RS of rye. It has been confirmed that Pm8 is a homoeologous gene of Pm3, and the homoeologous relationship between these two genes has been conformed (Hurni et al. 2013). According to the genetic analysis, Pm17 also proved to be an allele of Pm8, and the similarity between Pm17 and Pm8 was as high as 87% (Hsam and Zeller 1997; Singh et al. 2018). Furthermore, the dominant inhibitory effect of Pm3CS against the disease resistance of Pm8 was directly confirmed using the single-cell transient expression determination. The transgenic lines (Pm3a, Pm3f, and Pm3b) were separately crossed with the Pm8 transgenic lines, confirming that these three Pm3 alleles also suppressed the resistance of Pm8 to the avirulent Bgt isolates (Hurni et al. 2014).

Based on the homologous relationship with Pm3, Pm8, and Pm17 were cloned. The cloning and functional validation of Pm8 were performed via the Pm3 gene sequence (Hurni et al. 2013). The PCR marker aIAG95 facilitates the identification of Pm17 (Mohler et al. 2001). Functional analyses of southern blot, transient expression technology, and transgene demonstrated that Pm17 encodes an NLR-type R protein (Singh et al. 2018).

Pm5

Pm5 is located on chromosome arm 7BL. Five alleles (Pm5a–5e) were identified (Heun and Fischbeck 1987; Hsam et al. 2001; Lebsock and Briggle 1974). Recently, Pm5e was finely mapped to a 0.04-cM region, corresponding to a 13.5-kb genomic region, within which there are three protein-coding genes. DNA sequencing, EMS mutation, and transgene experiments showed that Pm5e encodes an NLR resistance protein. The allelic association analysis of Pm5e showed that a mutation of Met to Ile in the C-terminal LRR domain conferred its powdery mildew resistance (Xie et al. 2020).

Pm5e was initially genetically identified from the Chinese wheat cultivar Fuzhuang 30 (FZ30), which maintained powdery mildew resistance for over 80 years (Huang et al. 2003). Multiple Pm5 alleles were found in Chinese wheat landraces; however, the resistance Pm5e allele (NLRA) is rare. It is inferred that natural point mutation in the susceptible NLRG allele and disease screening pressure is the main reason for the acquisition of disease resistance in Pm5e. The mutation of Met to Ile in the Pm5e LRR domain represents powdery mildew resistance obtained, and the concomitant change in recognition and interaction with the AVR in Bgt is still nebulous. Uncovering the molecular mechanism of Pm5e-mediated resistance will contribute to enriching the genetic source of powdery mildew resistance for wheat improvement.

In addition to the multi-allelic Pm5 (Pm5a–Pm5e), several hypothetical Pm5 alleles were also reported. The mlxbd was located between markers 7BLSSR49 and WGGC5746, with genetic distances of 0.4 and 0.3 cM, respectively (Jin et al. 2020). The pmDHT was located in the region between Xwmc526/XBE443877 and Xgwmm611/Xwmc511, with a genetic distance of 0.8 and 0.3 cM, respectively (Qie et al. 2019). PmBYYT was located between SNP markers W7BL-8 and W7BL-15, with genetic distances of 3.0 and 2.9 cM, respectively. Gene allelic analysis has shown that pmDHT, mlxbd, PmBYYT, and Pm5e are alleles or tightly linked (Qie et al. 2019; Xu et al. 2018). Recently, PmAL11 was mapped in the Pm5 locus and sequenced to be identical to recessive Pm5e, although it was mediated by a dominant inheritance pattern (Han et al. 2023). The relationship between PmAL11 and Pm5 needs further determination.

Pm12/Pm21

Pm12 was derived from Ae. speltoides and transferred to wheat cv. Wembley, generating highly resistant common wheat Line#31 (Wembley Line#31, WL31). As early as 1991–1992, an RFLP molecular marker for the Pm12 gene was developed in an F2 generation segregating population, and it was clarified that the chromosome carrying Pm12 is a T6BS-6SS·6SL translocation. Meanwhile, there was serious recombination suppression between the alien chromosome segment from Ae. speltoides and the corresponding segment of common wheat (Jia et al. 1996). An EST-SSR marker Xcau127 that co-segregated with Pm12 was found, and this marker can detect both Pm12 and Pm21 in different genomes, which firstly implied the potential homologous relationship between Pm12 and Pm21 (Song et al. 2007, 2009).

Pm21, originating from the chromosome arm 6VS of D. villosum, is a highly effective resistance gene against almost all the tested Bgt isolates (Chen et al. 1995; Zhu et al. 2023). However, it is challenging to isolate Pm21 due to the recombination inhibition between the chromosome arm 6VS and the homoeologous wheat chromosome arms. In 2018, two groups of scientists simultaneously cloned Pm21 and confirmed that it encodes an NBS-LRR protein using different strategies (He et al. 2018; Xing et al. 2018). He et al. (2018) used a map-based cloning strategy based on the segregation population between resistant and susceptible D. villosum accessions. Xing et al. (2018) used a distinct-different MutRenSeq strategy to clone this gene. In addition, three D. villosum-derived Pm21(#4), Pm21(#4)-H, and PmV were also confirmed to be allelic to Pm21 (Li et al. 2005, 2020d; Zhao et al. 2019). Recently, a resistance Pm21 haplotype Pm21(8#) from a new wheat-H. villosa translocation T6DL·6 V#8S was cloned and showed dramatic sequence polymorphism, variable transcriptional patterns, and different resistance performance from other Pm21 alleles. The Pm resistance conferred by Pm21 was challenged by rapid evolution of Bgt population in wheat production; diverse resistance mechanism among Pm21 haplotypes will help increase the utilization of Pm21 cultivars in breeding process (Huang et al. 2023).

Pm12 was then cloned using deletion mutation and physical positioning, and the homologous relationship between Pm12 and Pm21 was finally determined (Zhu et al. 2023). The coding sequence of Pm12 shares 89.9% identity with that of Pm21. It is worth noting that these two orthologous genes differed in their resistance to Bgt isolates. We found that different intramolecular interactions may be the reason for the differences in cell death-inducing activity. Pm12 and Pm21 are ancient genes with abundant evolutionary dynamics, and transposable elements (TEs) insertion in the second intron of Pm12/Pm21 orthologs may be the reason for the loss of function in Chinese Spring. The cloning of Pm12 and Pm21 provides an excellent promotion for their rational and durable breeding use. Further studies on amino acid variations between Pm12/Pm21 and their orthologs may reveal the molecular mechanism of effector recognition and the interaction network in the disease resistance mediated by Pm12/Pm21.

Pm41

Pm41 was identified on chromosome arm 3BL. It was first mapped in the 2.7-cM region and then finely narrowed to a region of 0.6 cM (Li et al. 2009; Song et al. 2014). A NLR disease resistance gene was then identified as the Pm41a candidate by map-based cloning (Li et al. 2020c). Strikingly, haplotype analysis indicated that Pm41b and other Pm41 haplotypes were silent alleles of pm41a, probably due to transposon insertions. This result represents an unexplored hidden variation of Pm resistance in modern wheat breeding (Li et al. 2020c). Recently, a similar phenomenon was also reported. Sixteen wheat accessions containing Pm41 were detected among 332 wheat varieties, whereas all are susceptible to Bgt isolate at the seedling stage, unlike in IW2 (Cheng et al. 2022). It is implied that natural mutants frequently happened in Pm41, and more accurate molecular markers are required to distinguish the resistant and susceptible versions. Pm41 is the first cloned Pm gene from the WEW accession IW2 and confers the all-stage resistance. It is reported that Pm41 is only limitedly distributed among minority populations but harbors a high resistance to Bgt in the hexaploid Fielder and the tetraploid WEW IW2. Consequently, it is expected that Pm41 could be exploited as a valuable powdery mildew resistance resource and plays a vital role in wheat genetic improvement after introgression into cultivated wheat.

Pm60

Pm60 was isolated recently from T. urartu. Molecular markers locked the Pm60 in a 356-kb region of the T. urartu genome; cloning and sequence analysis identified two candidate genes, PmR1 and Pm60, both encoding NBS-LRR proteins. By gene silencing, single-cell transient expression, and stable transformation analysis, Pm60 was functionally validated. The Pm60 alleles Pm60a and Pm60b were then obtained by homologous cloning in T. urartu accessions PI428210 and PI428215, respectively (Zou et al. 2022). Pm60a showed sensitivity to Bgt isolate E18, whereas Pm60 and Pm60b were resistant. Further analysis showed no inhibition of allelic resistance function among these three alleles. This indicated that the amino acid deletion in Pm60a narrowed the resistance spectrum. PI428210 (Pm60a) was found to carry other uncharacterized resistance genes through hybridization analysis (Zou et al. 2022). Also, a non-functional allele of Pm60a named Pm60a' was identified from T. urartu accession PI662227, and two molecular markers, M-Pm60-S1 and M-Pm60-S2, were developed to distinguish Pm60a from Pm60a' (Zhao et al. 2020). Except for Pm60a, Pm60b, and Pm60a', several other alleles were also described from different wheat genotypes and related species, including MlWE18, PmG16, MlIW72, and MlIW172 (Ben-David et al. 2010; Ji et al. 2008; Ouyang et al. 2014; Zhao et al. 2020; Zou et al. 2018). The dominant gene PmU was also located on chromosome arm 7AL, which was later considered to be Pm60 (Qiu et al. 2005; Zhang et al. 2018a).

MlWE18 was mapped in the common wheat line 3D249 to a 0.09-cM genetic interval between markers WGGC4660 and WGGC4657, corresponding to a 334-kb physical region. Using RT-PCR, it was found that resk36_3047604 (NLRWE18) was highly homologous to the protein encoded by Pm60. Sequencing and transgenic technology showed that NLRWE18 was the allele MlWE18 (Wu et al. 2021a). A high-density SNP genetic map of the G18-16 × LDN RIL population (147 strains) was conducted to locate PmG16 in the 1.4-cM genetic interval between markers uhw386 and uhw390, in which the NLR gene TRIDC7AG077150.1 with high sequence similarity with Pm60 was found. Marker M-Pm60-S1 showed separation from PmG16, indicating that TRIDC7AG077150.1 was a candidate gene of PmG16 and named TdPm60 (Ben-David et al. 2010; Li et al. 2021b; Zhao et al. 2020). It is found that MlIW72 and MlIW172 also contain TdPm60 alleles (Li et al. 2021b). Using molecular markers associated with Pm1a, Pm60 MlWE18, Mlm2033, as well as polymorphic markers of Mo75, MlIW172 was mapped to the 0.048-cM interval between markers M405 and WGGC4656 (Hewitt et al. 2021; Yao et al. 2007). There were four CGT-Seq markers in IW172, and NLRIW172 was expressed after inoculating with E09. The subsequent function validation proved NLRIW172 as a new allele of Pm60 (Wu et al. 2022b).

Pm69

Pm69, derived from WEW, was mapped on chromosome arm 6BL, a complex genomic region with structural variations and suppressed recombination. So, it is difficult to clone it by the conventional positional cloning method. Recently, Pm69 was cloned and confirmed to encode an NLR protein using a long-read sequencing strategy Oxford Nanopore Technology (ONT) (Li et al. 2022d). High polymorphism was detected in the Pm69 region among different Triticeae genomes, implying a rapidly evolution event in this NLR cluster. This implied that we could mine more functional Pm69 alleles with different reaction patterns to the Bgt isolates. Pm69 was a rare NLR only found in individual WEW accessions and hence was expected as an undeveloped and valuable genetic resource for resistance breeding.

Kinase

Pm4 and Pm24 are the only two cloned Pm genes that encode kinases. Pm4 is located on the chromosome arm 2AL and has several alleles, including Pm4a, Pm4b (Mle), Pm4c, Pm4d, and Pm4e (Hao et al. 2008; Li et al. 2017; Mcintosh and Bennett 1979; Schmolke et al. 2012; Ullah et al. 2018; Wu et al. 2018; Yao et al. 2022). Pm4b proved to play a vital role in powdery mildew resistance in many wheat cultivars (Vincent et al. 2017). After confirming the precise physical localization of these alleles, AET2Gv21296200 was identified as Pm4b, which encodes the chimeric kinase-MCTP (multiple C2 domains and transmembrane region protein) (Sánchez-Martín et al. 2021). Firstly, EMS-mutant lines were obtained from the Fed-Pm4b wheat lines, and the chromosome 2A carrying Pm4b was flow-sorted and sequenced by the MutChromSeq approach. Interestingly, it is proven that only the existence of two spliceosomes, Pm4b_V1 and Pm4b_V2, could confer powdery mildew resistance by the transgenic and VIGS analyses. Furthermore, subcellular localization and in vitro and in vivo interaction analysis demonstrated that Pm4b_V2 interacts with and recruits Pm4b_V1 from cytosol to endoplasmic reticulum to form an ER-associated complex. This study revealed a new notion of race-specific powdery mildew resistance conducted by receptor-like cytoplasmic kinases. However, the downstream signal molecules and interacting partners are not clear, and the exposition of how the Pm4-ER complex detects and manipulates pathogen effectors will help understand Pm4-mediated resistance at the mechanistic level.

Pm24 and its alleles Pm24b, PmDTM, and MlHLT are localized on the chromosome arm 1DS, and they were confirmed as the identical alleles responsible for the resistance to powdery mildew (Huang et al. 2000; Lu et al. 2020a, 2020b; Wang et al. 2015; Xue et al. 2012b). In 2020, Pm24 from Chinese wheat landrace Hulutou was identified as TraesCS1D02G058900 according to the IWGSC CS RefSeq, and a 6-bp natural deletion of lysine–glycine is critical for the powdery mildew resistance (Lu et al. 2020b). Another team owned the same result from Datoumai (Lu et al. 2020a). TraesCS1D02G058900 encoded a receptor-like serine/threonine protein kinase. Receptor-like cytoplasmic kinases usually act as the preliminary defense in pattern recognition receptor-mediated immunity and have been reported and well described to confer pathogen resistance in Arabidopsis (Bi et al. 2018; Liu et al. 2019).

Transporter

Pm38/Yr18/Lr34/Sr57 from bread wheat acts as one of the few genes with durable disease against multiple fungal diseases for over 50 years. To understand the molecular nature of this resistance gene, great efforts have been made for decades.

At first, Lr34 was localized between the two markers gwm1220 and SWM10 on the chromosome arm 7DS and then locked in a 363-kb physical interval (Bossolini et al. 2006; Spielmeyer et al. 2008). Eight open reading frames were identified in this candidate region. Further molecular marker assays, consistent sequence polymorphism in the alleles, and candidate DNA sequencing analysis in loss of function of the Lr34 mutants limited the candidate to an ABC transporter consisting of 24 exons (Krattinger et al. 2009). Lr34 sequencing comparison between Chinese Spring (+ Lr34) and Renan (a French winter cultivar, -Lr34) indicated only three polymorphisms, among which two were located in exons containing one deletion of a phenylalanine residue and a tyrosine to a histidine in Chinese Spring (Krattinger et al. 2011). However, both amino acid changes were found only in domesticated bread wheat (Lr34res), but not in the wild wheat progenitors and susceptible wheat cultivars (Krattinger et al. 2016). It is presumed that the conservative phenylalanine residue plays a pivotal role in Pm38-resistance. A study on barley (Hordeum vulgare L.) cultivar Golden Promise showed that a deletion of a single phenylalanine codon in the D genome is sufficient to convert the lr34-susceptible allele to a functional Lr34 resistance gene (Chauhan et al. 2015). These data demonstrated that Lr34/Pm38 encodes an ABC-type transporter conferring a single origin and durable fungal diseases resistance to which the domesticated phenylalanine residue absence is critical.

To explore the molecular mechanism of Pm38, the Lr34/Pm38 transgenic wheat was generated, and resistance assays confirmed that a minimum transcriptional threshold of Lr34 was required. This evidence explained the discrepant resistance between adult plants and field-grown seedlings and between hexaploid wheat and durum wheat or barley (Rinaldo et al. 2017). It revealed a transcription mechanism in the Pm38-regulated resistance.

In Lr34-expressing rice, transcriptomics, physiology, genetics, and in vitro and in vivo transport assays clarified that LR34 transports ABA and alters the ABA-regulated gene expression and biological processes (Krattinger et al. 2019). A similar result showed that the redistribution of plant hormone abscisic acid and a series of phenotypes of increasing ABA signaling, including reduced stomatal conductance or leaf tip necrosis, was shown in Lr34-expressing wheat and barley (Braunlich et al. 2021). It is worth noting that Lr34 was found to interact with other rust resistance genes, including Lr13, Lr37, and Lr46, to maintain durable and effective plant resistance (McCallum and Hiebert 2022); however, the interaction mechanism in Pm38-mediated powdery mildew resistance has not been studied.

Pm38, acting as an almost ‘all-around’ resistance gene in cereal crops, plays a pivotal role in wheat breeding. The molecular essence and action mechanism have been investigated preliminarily; nevertheless, the intracellular interaction network conferring the pathogen resistance needs to be further explored.

Like Pm38, Pm46/Yr46/Lr67/Sr55 also confers multiple pathogen resistance in adult plants. In 2015, Lr67 was identified as a hexose transporter. At first, a wheat bacterial artificial chromosome (BAC) clone carrying a closely linked (0.4 cM) marker with Lr67 on chromosome 4DL was isolated in the near-isogenic wheat cultivar Thatcher + Lr67 line, RL6077. In the BAC clone, two conservative genes were predicted, and then, more markers were designed to distinguish the resistant and susceptible lines. After scoring these markers and sequence analysis in susceptible EMS mutants of RL6007, a predicted sugar transporter (STP) was preliminarily identified. Amino acid changes resulting from nucleotide transitions or fragment deletions in STP were found in all EMS mutants. In addition, STP-transgenic wheat and barley conferred the adult-plant resistance to stripe rust or leaf rust. Conclusively, the sugar transporter gene was identified to confer the Lr67 multi-pathogen resistance (Moore et al. 2015). Functional analysis of Lr67 was also carried out. Conservative Gly144 and Val387 residues were mutated to Arg and Leu in Lr67-resistant alleles. Using hexose transport deficient yeast mutant, [14C] glucose uptake assay was conducted. Resistant LR67 protein was incapable of glucose import, whereas the replacement of Arg144 with Gly restored the transport activity, but Leu387 could not. Kinetic assay and inhibitor treatment suggested that LR67sus functions as an H+/hexose symporter. Bimolecular fluorescence complementation (BiFC) assay in LR67 ectopic expressed Nicotiana benthamiana indicated that LR67 can form homo- and heterodimers. According to the dimerization-mediated dominant-negative interference on transporter activity in other plant sugar transporter families, it is inferred that LR67res reduced hexose transport by forming inactive heteromultimeric protein complexes. Together, LR67-mediated pathogen resistance has proven to be closely related to hexose transportation; however, more evidence, especially in wheat, is still needed to illuminate the specific mechanism and regulation network of LR67 during plant defense responses.

Pm38 and Pm46 are both pleiotropic APR genes that confer a broad spectrum of powdery mildew resistance. However, they are usually difficult to be selected by conventional breeding methods because they are always pyramided with other resistance genes or QTL. Therefore, precision breeding utilizing modern molecular biology may be a better choice; and the in-depth study of the molecular mechanism and interaction network of Pm38 and Pm46 will significantly accelerate this process.

Various Pm genes that may be effective against single or multiple races of Bgt isolates have been characterized. Notably, most of the cloned Pm genes encoded NBS-LRR protein that belongs to the typical pathogen resistance factors; and other Pm genes are identified as kinase or transporter. Undoubtedly, these resistance genes possess diverse acting mechanisms and form huge molecular interaction networks. Pathogen perception, effector identification, defense signaling, variable splicing, transcription regulation and other resistance response mechanism will further be explored and discovered to expand our knowledge of wheat powdery mildew resistance at the molecular level.

Powdery mildew avirulent genes and the interaction mechanism with resistance genes

Upon the interaction between plants and pathogens, effectors are secreted by pathogens and then recognized by plant NBS-LRR proteins, therefore activating the ETI immune response. Most Pm resistance genes identified in wheat encode NBS-LRR immune receptors that conform gene-for-gene hypothesis and correspond to avirulent effectors (encoded by Avr gene) in pathogens. Identification of avirulent gene diversity in pathogen populations is a critical factor that may be related to the durable resistance in wheat production. In recent years, many insights about the AvrPm genes are obtained using the map-based cloning, GWAS, RNA sequence, BAC sequence, and other technologies. At present, several AvrPm genes have been described, including AvrPm1a, AvrPm2, AvrPm3a2/f2, AvrPm3b2/c2, AvrPm3d3, AvrPm4, AvrPm8, and AvrPm17 (Hewitt et al. 2021; Müller et al. 2022; Praz et al. 2017).

AvrPm1a is identified as BgtE-5612_THUN12, and the RNase-like folding was found in the AvrPm1a structural modeling. Utilizing transient co-expression assay in the N. benthamiana system, specific recognition between Pm1a and AvrPm1a has confirmed and led to a hypersensitive (HR) response. However, further physiological and biochemical evidence is still needed to analyze the interaction mechanism between AvrPm1a and Pm1a. AvrPm2 encoding as BgtE-5845 belongs to the family of RNase-like effectors, and RNase effectors play an essential role in controlling powdery mildew virulence and pathogen race specificity. AvrPm2 also shares a structural homology with known RNase-like ribonuclease, which is required for haustorium formation (Pedersen et al. 2012; Pennington et al. 2016; Pliego et al. 2013). Therefore, it was hypothesized that AvrPm2 could not only interact with Pm2 to trigger the HR response in hosts and participate in the defense process but also participate in haustorial formation and inhibit the defense response of plant cells. In the study on the specific resistance mechanism of Pm4, it is suggested that AvrPm4 could directly interact with Pm4 during haustorial formation to elicit a defense response. AvrPm4 can bind to any Pm4 variant, and the heterologous complex can result in kinase activation and disease resistance (Sánchez-Martín et al. 2021).

The diversification of AvrPm3 is inseparable from the diversification of Pm3 alleles. It can be predicted that AvrPm3 will continue to evolve to avoid R gene recognition (Sela et al. 2014). At present, the research on AvrPm3 is relatively comprehensive. AvrPm3a2/f2 and its alleles AvrPm3b2/c2 and AvrPm3d3 were described; meanwhile, a specific suppressor gene has been identified in the pathogen. AvrPm3a2/f2 corresponding to Pm3a and Pm3f was cloned, and it could specifically recognize these two alleles (Bourras et al. 2019). Furthermore, it was found that Bcg1vir from the pathogen inhibited the resistance mediated by AvrPm3a2/f2-Pm3a/f. A new model Avr-R-Svr was proposed and confirmed by the addition of the third interaction element ‘suppressor of avirulence (Svr)’ (Bourras et al. 2015, 2019). SvrPm3a1/f1 involved in this model may inhibit the expression of AvrPm3 or Pm3, impede the recognition of AvrPm3, and lead to the reduction of the HR phenotype. Another study implied that Pm3f is controlled by a single locus AvrPm3f, while the AvrPm3c has two independent loci, and AvrPm3c1 is epistatic to AvrPm3c2 (Parlange et al. 2015). The interaction mechanism of AvrPm3-Pm3 is complex. It has been shown that SvrPm3a1/f1 can not only inhibit AvrPm3a2/f2 and reduce the intensity of HR but also have the same effect on AvrPm3b2/c2 (Bourras et al. 2019). The interaction between other Pm3 alleles can also inhibit the interaction of AvrPm3-Pm3 (Bourras et al. 2015; Stirnweis et al. 2014). Svr, an avirulent inhibitor encoded by pathogens, also plays a role in the AVR-R-SVR model. However, many isolates are not suitable for the AVR-R-SVR model, which further indicates the complexity of the molecular mechanism of Pm3 avirulence genes.

Pm8 and Pm17 are the homologous versions of Pm3 in rye. AvrPm17 is encoded by a gene pair of BgTH12-04537 and BgTH12-04538 and functionally validated recently. Co-expression and protein detection showed that Pm17 could recognize AvrPm17_THUN12 and AvrPm17_96224. According to the sequence analysis, effector expansion occurred in wheat powdery mildew after the divergence from the barley mildew lineage, and AvrPm17 is encoded in an expanded gene cluster. It is supposed that effector expansion is coevolution with the host immune system and is important for resistance durability in introgression species. AvrPm17 mutation occurred earlier than the time when Pm17 entered wheat, which indicated that ancient variation in the AvrPm17 limits the effectiveness of rye Pm17 resistance genes in wheat (Müller et al. 2022). AvrPm8 was identified by GWAS and functionally validated in N. benthamiana. Multiple unique gain-of-virulence mutations of AvrPm8 were sequenced from a global collection of 219 Bgt isolates. According to the statistical analysis between the worldwide distribution of these mutants of AvrPm8 and the avirulent variant, Kuna et al (2023) inferred that the high prevalence of Pm8 wheat results in continual gain-of-virulence mutations. On the contrary, AvrPm3 and AvrPm17 gain-of-virulence mutations were relatively more conservative and relied on single amino acid polymorphisms meanwhile preserving at least one functional ORF (Bourras et al. 2019; McNally et al. 2018; Muller et al. 2022). Regional breeding programs impact the immune escape ability from Avr effector mutant; perhaps, only the ancient genetic variation that occurred before introgression to wheat could allow the more durable resistance like in AvrPm17. The ultimate approach to breeding is to continue to track the avirulence gene mutations and then manipulate them by molecular means.

Although many potential avirulence genes are predicted to exist in the powdery mildew fungi, only a few Avrs have been cloned or described. Further studies on their composition and function are needed for the remaining undiscovered Avrs. The identification and cloning of Avr genes can provide new ideas and help understand the molecular mechanism during wheat resistance to powdery mildew. In addition, specific markers for Avr genes allow easy determination of their presence and evolution, which contributes to the practical analysis of pathogen avirulence patterns and rational deployment of resistance genes in specific agricultural events. Furthermore, breeders will benefit from the precise selection of specific Avr proteins and lower risks for pathogen adaptability. Predictably, breeding strategies that integrate with pathogen molecular biology will make great improvement in powdery mildew resistance breeding and assist agricultural development.

Genomic-assisted breeding for powdery mildew resistance

Conventional breeding has a long history and still plays a key role in modern crop improvement. This approach involves the use of natural germplasm collection, hybridization and integration of complementary genetic sources, and extensive screening of excellent traits in the different generations. Using conventional breeding approaches, many elite cultivars have been produced, bringing a significant promotion in wheat production. However, conventional breeding requires multiple generations of screening and is easily affected by complex environmental and climatic conditions; hence, it is time-consuming, labor-intensive, and inefficient.

Molecular breeding provides means and possibilities for accelerating the conventional breeding process. Understanding the performance of parents, especially the main genes, can predict the performance of hybrid combinations, providing the possibility for purposeful variety cultivation. With the help of molecular breeding technology, the breeding process should be more effective and reasonable.

Molecular markers are complementary tools to conventional breeding since they are easy to assay, highly heritable, more accurate, cheaper, faster, and not affected by environments. Marker-assisted selection (MAS) transfers phenotype screening in the breeding process to the molecular detection. In recent years, ‘speed breeding’ combined with the molecular selection tools has been used to transfer and pyramid R genes into elite susceptible cultivars, which greatly shortens breeding cycle. By extending the photopericycle, common wheat and durum wheat could breed six generations per year. Speed breeding accelerates the growth and development of normal plants without affecting the excellent phenotype (Watson et al. 2018). Recently, a new comprehensive speed breeding for long-day crops was also proposed (Song et al. 2022). Furthermore, integration of precise MAS tools, high-throughput genotyping, and speeding breeding may efficiently accelerate the breeding use of Pm genes, increase the wheat genetic variation, and help fight with powdery mildew (Fig. 3).

Fig. 3
figure 3

Comparison of four breeding strategies and utilization of Pm genes in modern breeding approach. Four breeding strategies including conventional breeding, MAS breeding, speed breeding, and gene editing and genetic engineering are displayed above the arrow. The bottom represents modern breeding methods based on powdery mildew- resistant and susceptible genes

Full size image

Development and utilization of breeding markers for Pm genes

Due to the continuous variation of virulent Bgt isolates and the high barrier in introducing resistant cultivars, the current situation of high-quality wheat cultivars with high yield and Pm resistance is worrying. Therefore, the introduction of diversified Pm genes and the enhancement of germplasm innovation are urgent and practical approaches for the high-yield and multi-resistance breeding of wheat (Fig. 4).

Fig. 4
figure 4

Turnaround of wheat powdery mildew resistance research. The main aspects of wheat powdery mildew resistance research include phenotyping, sequencing, genotyping, gene function, and breeding. These aspects complement and depend on each other

Full size image

Molecular markers have been widely used to detect Pm genes. For example, 659 resistant accessions to Bgt isolate E09 were screened from 2978 wheat accessions, and specific molecular markers closely linked to Pm2, Pm4, Pm5, Pm6, Pm8, Pm21, and Pm24 were used to test the 659 resistant accessions, 328 resistance accessions may carry single Pm gene and 191 accessions carry combined Pm genes consequently (Jin et al. 2021). Similarly, six accessions carrying Pm3, 18 accessions carrying Pm8, five accessions carrying Pm16, and three accessions carrying Pm21 were identified in the germplasms from Yunnan Province, China (Wu et al. 2021b). In summary, molecular markers screening reveals resistance information of local wheat cultivars and provides a favorable reference for the rational planting of different resistant cultivars in the production process (Table 1).

Table 1 List of officially named Pm genes in wheat resource bank
Full size table

In addition, MAS technology can help transfer or pyramid different Pm genes for developing broad-spectrum germplasms, which has been the mainstream direction of breeding in recent years. For example, KASP markers that can simultaneously diagnose Pm12, PmV, and Pm21 are developed, which help pyramid these genes to other genotypes (Zhang et al. 2021c). Based on SLAF-Seq, 404 PCR and 14 KASP rye-specific markers on a large scale were developed, of which two KASP markers, SWK5282 and SWK252224, were successfully applied in rye molecular breeding, which will help the detection and transfer of rye chromosomes 1RS and 2RL in the wheat background (Han et al. 2020a). Four new varieties with elite agronomic traits are developed using molecular markers linked to Pm40 (Tang et al. 2018). Using MAS and phenotypic identification, Pm2b was introduced into three elite susceptible cultivars, and Pm2b near-isogenic lines (NLRs) with improved agronomic performance were developed (Xu et al. 2017). The combination of MAS technology and genetic engineering will make the breeding process more efficient and predictable, and there will be important breakthroughs in molecular mark-assisted breeding.

As a simple and low-cost technique, molecular markers also play an essential role in the molecular identification of distant hybridization materials between wheat and its relatives. Wheat-rye T1RS·1BL translation line harbors many disease resistance genes, including Pm8, Yr9, Lr26, and Sr31. Although Pm8 and Lr26 have lost its resistance due to the pathogen evolution, T1RS·1BL is still widely used in wheat breeding because of its yield potential and extensive environmental adaptability. By molecular markers analysis of rye chromosome arm 1RS, R2207 was confirmed as a cytogenetically stable T1RS·1BL translocation line with highly resistant to stripe rust and powdery mildew with superior agronomic traits (Han et al. 2020b).

At present, molecular markers have made large contribution in the wheat powdery mildew breeding process; however, it is still an urgent task to develop more efficient molecular markers for Pm resistance screening. In the breeding process, it is also necessary to select high-quality genes for pyramiding according to the immunization periods and tissues for different genes. In future, the combination of MAS technology with transgenic and gene editing technologies will significantly improve the accuracy and effective utilization of Pm resistance and other key agronomic traits.

Gene editing and intelligent breeding

With the innovation of molecular sequencing and cloning technologies, more and more Pm genes will be revealed. However, there is still a massive gap between gene cloning and the actual production application. The integration and introgression of resistance genes from wild relatives to wheat landraces and local cultivars have attracted widespread attention. Genetic engineering is available only for the cloned Pm genes; introgressive hybridization is an alternative method, but the rearrangement of wheat genetic material, distance incompatibility, and undesirable agricultural traits are always accompanied and resulting in a low efficiency.

Increasing studies demonstrated that structural variation, including alternative splicing, fragment deletion or insertion, and resistance-determined SNP play vital roles in resistance gene-mediated defense reactions (Andersen et al. 2020; Jin et al. 2022; Lu et al. 2020b). CRISPR/Cas9, a powerful gene editing tool, makes it possible to artificially modify the molecular structure in susceptible alleles to a resistant version. CRISPR/Cas9 has been used to edit multiple genes in wheat to alter morphogenesis, create male sterile mutant, analyze the gene function, and improve Pm resistance (Hyde et al. 2023; Li et al. 2020b; Liu et al. 2020; Zhang et al. 2017). A notable breakthrough has been reported that the targeted deletion of susceptible Pm gene Mlo by Crispr generates broad resistance to powdery mildew without yield penalty (Li et al. 2022c). However, the bottleneck imposed by tissue culture and regeneration restricts the promotion of wheat genetic transformation. Recently, a Barley stripe mosaic virus-based sgRNA (BSMV-sg) delivery system was developed and helps achieve culture-free and heritable genome editing in wheat plants (Chen et al. 2022; Li et al. 2021a). Consequently, basic researches on Pm genes structural variation combined with powerful genome editing tools may provide new and more intelligent breeding strategies to promote broad-spectrum powdery mildew resistance.

Perspective

Mining Pm resistance genes is the foundation for exploring host resistance. However, the process is long, complex, and full of uncertainty. For instance, wheat cultivars harboring Pm2 are widely used in wheat production worldwide. We identified 33 wheat cultivars/breeding lines carrying Pm2 alleles. These alleles were all confirmed to be Pm2a haplotype by sequencing although these genotypes showed different reaction patterns to different Bgt isolates. It is supposed that the various resistance phenotypes in these wheat cultivars are influenced by different genetic backgrounds. Another possibility is that Pm2 does not function alone in these wheat cultivars/breeding lines. There may exist other factors or enhancers closely linked with Pm2, which form a genetic unit to function together with Pm2. In addition, the heterozygous or mixed population of isolates may increase diversity of Avr genes and uncertainty in the phenotype identification. Therefore, further and deep evidence are needed to clarify the genetic and/or molecular mechanisms of Pm2-mediated resistance.

The ideal Pm resistance is a broad-spectrum and durable resistance in the whole growth period, but most of the Pm genes only show partial resistance. Pyramiding multiple Pm genes into a same genetic background may be the best option for achieving ideal Pm resistance. However, not all Pm genes are suitable for pyramiding. For example, the combination of Pm3 and Pm8 does not have a good breeding effect. Moreover, two quality resistance genes with high resistance may over allocate photosynthetic energy, leading to a decrease in yield. Therefore, a better choice is to pyramid different types of Pm genes, such as a major gene and an adult resistance gene or a QTL, which may bring better resistance without losing excellent agronomic traits. Furthermore, the combination of different germplasms with tissue-differentiated resistance may provide new ideas for resistance breeding. For example, Pm67 showed seedling and sheaths, stems, and spikes at the adult-plant stage immunity, but a low level of susceptibility on adult-plant leaves. In contrast, Pm55 conferred a near immunity response on leaves after 5-leaf stage but susceptibility on sheaths and spikes. The F1 hybrid plants between Pm55 and Pm67 harboring lines exhibited complete immune across all tissues (Zhang et al. 2021b). Therefore, pyramiding these two genes may conduct a different resistance mechanism and could become the novel and potential valuable germplasms for adult-plant powdery mildew resistance.

Pm genes from wild relatives usually confer a high-level resistance. However, linkage drags are usually difficult to overcome, which impede the effective and rapid utilization of most alien Pm genes in breeding. To make use of these Pm genes in the actual breeding process, the best way is to clone them and directly transfer these genes to wheat background using transgenic technology. Alternatively, some strategies, such as creating smaller translocation fragment, multigenerational backcrossing, mutagenesis treatment, or cutting the linkage drags by gene editing, could also accelerate their application in breeding.