The QTL interval mapping studies for stripe rust in wheat started in mid-1990s (Line et al. 1996). Since then, more than 70 reports have been published, leading to identification of >350 QTLs, so that stripe rust resistance is now treated both as a quantitative trait involving quantitative resistance loci (QRLs) and a qualitative trait controlled by a number of Yr genes, the latter following a gene-for-gene relationship with Avr genes in the pathogen. The relationship of QRLs in the host and the corresponding QTLs for virulence in the pathogen has not been worked out so far, although quantitative nature of virulence in pathogen has been worked out in some cases including wheat pathogens like Zymoseptoria tritici where several large and small effect QTLs were identified for virulence (Stewart et al. 2018)
The 61 MQTLs identified during the present study were derived from 184 QTLs, which indicated roughly three-times reduction in redundancy for the genomic regions controlling stripe rust resistance in wheat genome. Earlier, while conducting meta-QTL analysis in wheat for fusarium head blight, roughly five-fold reduction in redundancy was reported (Venske et al. 2019). The absence of MQTLs on the 3 of the 7 D sub-genome chromosomes (5D, 6D, 7D; Figure 2) agrees with earlier reports on QTL analysis (Gutierrez-Gonzalez et al. 2019; Rimbert et al. 2018; Gardener et al. 2016; Iehisa et al. 2016; Poland et al. 2012). A similar situation was earlier reported for MQTL analysis for leaf rust resistance (Soriano and Royo 2015) and MQTLs for fusarium head blight (Venske et al. 2019).
Sixty-one (61) MQTLs (including four MQTLs derived from the QTLs belonging to durum wheat) is a fairly large number indicating a high degree of redundancy of QTLs which agrees with a large number of Yr genes for stripe rust resistance reported in wheat genome. Such a redundancy of genes/MQTLs is a requirement for providing resistance against large number of ever-evolving races of stripe rust, distributed in different wheat growing regions of the world (Pradhan et al. 2020). Earlier, for resistance against fusarium head blight also, 65 MQTLs were identified. However, the number of MQTLs identified in this study far exceeds the number of MQTLs identified for leaf rust resistance (35) (Soriano and Royo 2015). Perhaps this is due to relatively fewer QTL studies (19) available for meta-analysis in case of leaf rust resistance.
Most of the MQTLs identified in the present study controlled more than one parameters/traits (Table 2; Table S2). This probably indicated either a tight linkage of genes for different traits, or occurrence of pleotropic genes or a bias due to the use of related traits measuring the same resistance component by different means as also reported earlier in case of MQTL analysis for leaf rust resistance in wheat (Soriano and Royo 2015). In the present study, five MQTLs (MQTL3-1B, MQTL10-2A, MQTL19-2B, MQTL20-2B and MQTL24-2D) also showed co-localization of 1 to 9 different known Yr genes including four cloned Yr genes. These genomic regions may be involved in controlling both, qualitative and quantitative resistance and thus may be more important. Earlier, a number of Yr genes (out of the 82 reported Yr genes) have been deployed in commercial wheat cultivars; however only few are still effective. For instance, some of the Yr genes which are still effective in India include Yr5, Yr10, Yr15, Yrsp, Yr47, Yr57 and Yr63 (Prasad 2020; Sharma et al. 2020). Four of these Yr genes (Yr5, Yr10, Yr15, Yrsp) that were found to be co-located with MQTLs (Figure 3) identified during the present study and four other Yr genes (Yr53, Yr61, Yr65,Yr69) are known to be effective worldwide (Zhang et al. 2019; Zhou et al. 2014). The MQTLs showing co-localization with Yr genes may be important targets for introgression into susceptible wheat lines for improvement of stripe rust resistance. While breeding for stripe rust resistance, generally Yr genes are deployed and there is only one report where two QTL (QYr.nafu-2BL and QYr.nafu-3BS) have been utilized for developing stripe rust resistance in wheat cultivars (Hu et al. 2020).
Some of the MQTLs also overlap Yr genes that have already been cloned. For instance, MQTL20-2B was co-localized with two cloned Yr genes (Yr5/Yrsp and Yr7) whereas MQTL1-1A was co-localized with cloned gene Yr10. Co-localization of Yr5/Yrsp and Yr7 in the same MQTL region is perhaps due to the allelic nature of both the genes which were earlier shown to be closely linked (Zhang et al. 2009). All the above three cloned genes encode proteins for NBS-LRR (Liu et al. 2014; Marchal et al. 2018).
Efforts were also made during the present study to identify MQTLs and MQTL hot spots that may prove useful for breeding; we describe these MQTLs as breeders’ MQTL. For selecting these breeders’ MQTLs, we utilized a number of criteria including the following two criteria suggested in an earlier study (Loffler et al. 2009): (i) the low CI and high average PVE of the MQTLs and (ii) the number of QTLs carried by individual MQTL. Additional criteria were also used in the present study for prioritizing and selecting breeders’ MQTLs and MQTL hotspots. For instance, the relationship between MQTLs and the pathotypes occurring in specific wheat growing regions may be an important criterion. While doing this we also have to keep in mind that virulence can also be quantitative in nature as mentioned earlier. MQTLs showing resistance against more than one pathotypes may also be important for achieving broad spectrum resistance. Such important MQTLs showing resistance against multiple pathogen races were also identified in a recent study on MQTL analysis reported for tan spot resistance in wheat (Liu et al. 2020).
MQTLs identified in the present study consisted of original QTLs which showed resistance at either APR, HTAP (high temperature adult plant resistance), SR (seedling resistance) or all stage resistance (ASR). Also, almost all the MQTLs (except MQTL36-4A) showed resistance against more than one pathotypes indicating that these MQTLs exhibit race non-specific resistance and may be containing a number of novel genes which may be involved in providing resistance against broad spectrum of pathotypes. Keeping in view the different criteria listed above, a number of breeders’ MQTLs were identified, which are listed in Table 4.
Some of the Yr genes have been shown to overlap the QRLs/MQTLs and also the CGs that were identified during the present study. The 409 CGs identified during the present study have been shown to encode a variety of proteins; at least some of them are known to be involved in disease resistance (Table 3). The differential expression of CGs observed during the present study agrees with earlier reports (Wang et al 2021; Dobon et al. 2016; Zhang et al. 2014). These genes are largely involved in important processes like protein phosphorylation, photosynthesis, protein ubiquitination, transmembrane transport, oxidation-reduction processes, etc. which are relevant to disease resistance. In an earlier study also, a reduction in photosynthesis was shown to enhance stripe rust resistance due to the interaction of Yr36, encoding for wheat kinase STARTI (WKSI) with Psbo, a member of photosystem II (Wang et al. 2019) without having any adverse effect on yield. Similarly, in another study, a number of genes encoding PR (pathogenesis-related) proteins, involved in a number of defense responses were shown to get induced in response to stripe rust infection in a number of wheat lines carrying different genes for ASR (YrTr1, Yr76, YrSP, YrExp2) and HTAP (Yr5, Yr59, Yr62 and Yr7b) (Farrakh et al. 2016). A number of downstream genes, apparently similar to the CGs identified in the present study and involved in processes mentioned above were also identified in a transcriptome study conducted using a pair of introgression lines, which differed for Yr5 (Dobon et al. 2016).
CGs identified during the present study also deserve to be discussed. In wheat, CGs underlying the MQTLs were also identified for several traits including drought tolerance (Kumar et al. 2020), tan spot resistance (Liu et al. 2020) and fusarium head blight resistance (Venske et al. 2019). However, the strategy used by us was novel and not used in any of these earlier reports. For instance, in most of the earlier reports, the complete physical interval of the MQTL regions was considered for identification of CGs. However, in the present study, we calculated the exact physical position of the MQTL based on the MQTL peak position available from the BioMercator software. The 1 Mb interval on either sides of the MQTL peak was considered for identification of genes, which were used for identification of CGs responsible for stripe rust resistance.
Important differentially expressed CGs identified in the present study are presented in Figure 4. Fifty-nine (59) CGs out of the total 409 CGs were available in breeder’s MQTL indicating that these CGs are more important. Out of the 59 CGs, 32 CGs also showed differential expression and encoded important R genes, S/TPK, SLC transporter, Mitogen-activated protein (MAP) kinase, UDP-glucosyltransferases, S1/P1 nuclease, etc. The role of some of the important CGs (shown in Figure 4) during disease resistance can be summarised as follows: (i) Earlier, STPK-V, a member of Pm1 gene was reported to confer powdery mildew resistance in wheat (Cao et al. 2011). (ii) NBS-LRR domain containing genes are the protein products of the cloned Yr genes like Yr10, Yr5, etc. as mentioned earlier (Liu et al. 2014; Marchal et al. 2018). (iii) TaMAPK4, a type of MAPK gene is reported to act as a positive regulator of stripe rust resistance in wheat (Wang et al. 2018). (iv) The above transporters may also possibly encode Yr genes similar to Yr46 which was shown to encode for hexose transporter (Moore et al 2015). (v) UDP-glucosyltransferases were earlier reported to show differential expression due to stripe rust infection in wheat genotypes indicating their role in Yr39 mediated stripe rust resistance (Coram et al. 2008). Some other important CGs like those encoding for WRKY domains, Ankyrin repeat and F-box domain containing genes were also identified in different MQTLs, although the expression data was not available for these genes. WRKY and Ankyrin repeat domain containing genes were recently found to encode for proteins of cloned YrU gene (Wang et al. 2020). Similarly, F-box domain containing gene was identified as a CGs underlying the YrR39 locus in wheat and it was shown to upregulate due to stripe rust infection (Yin et al. 2018).
In summary, the present study allowed us to identify 6 MQTLs and 4 MQTL hotspots to be used by breeders particularly for high yielding wheat cultivars which are susceptible for stripe rust (Table 4). Four of these 10 genomic regions also showed co-localization with known Yr genes. Some of the important CGs which were identified during the present study may be further validated/edited using approaches like gene editing, overexpression, gene knockout strategies or CG based association mapping (CGAM). Reports are available where some of these strategies have been used for validation of genes for their role in stripe rust resistance. For instance, overexpression of TaWRKY62 provided high temperature seedling plant resistance to stripe rust by activating other genes encoding for PR proteins, salicylic and jasmonic acid responsive genes and ROS associated genes (Wang et al. 2017b). Similarly, in another study overexpression of TaLHY (a type pf MYB TF) in leaf blade and sheath reduced the negative impacts of stripe rust on wheat plant (Zhang et al. 2015). This knowledge may prove to be useful for validating similar CGs identified in this study.
Gene editing or mutation is still unexplored in case of stripe rust resistance except a single study where the function of Yr15 gene (encoding wheat tandem kinase 1 or WKS1) in stripe rust resistance was validated using mutation analysis (Klymiuk et al. 2020). EMS mutations were created in this gene in resistant wheat lines to develop susceptible lines. The resulting susceptible lines showed point mutations in the three amino acids,i.e. Gly54, Ala149 and Ala460 leading to disruption in gene function, thereby validating the role of WSK1 in resistance. Therefore, a similar strategy may be certainly explored for at least three cloned genes (Yr5/Yrsp, Yr7 and Yr10) which are co-located in two important MQTL regions mentioned in Table 4 as well the important CGs shown in Figures 4a and 4b. Similarly, techniques involving CRISPR/Cas9 or base editing may also be employed for the above cloned Yr genes as well as the CGs after the identification of causal SNPs involved in providing stripe rust resistance through CGAM approach. Similar report involving CRISPR/Cas9 for fusarium head blight are available where successful editing (using CRISPR-Cas9) of three genes, TaABCC6, TaNFXL1, and TansLTP9 showed enhanced resistance (Cui et al. 2017).