Disease resistance in plants including leaf rust resistance in wheat can be either all stage resistance (ASR) or adult plant resistance (APR). ASR provides race-specific resistance during seedling stage to the adult stage, whereas APR provides resistance only at the adult stage. In wheat, APR genes for leaf rust have been classified by some in two groups, one group including 7 genes (Lr12, Lr13, Lr22a/b, Lr35, Lr37, Lr48 and Lr49) that are race-specific and hypersensitive and the other group having 8 genes (Lr34, Lr46, Lr67, Lr68, Lr74, Lr75, Lr77 and Lr78) that are race-nonspecific and non-hypersensitive,41,42. Other workers recognized only the 8 non-race-specific non-hypersensitive genes as APR genes43,44.
The race non-specific APR genes are also described as ‘slow rusting’ genes or ‘partial resistance’ genes, which typically reduce the disease development in adult plants. It has also been shown that while ASR is qualitative in nature following Flor’s gene-for-gene relationship, APR is quantitative in nature, where an individual APR gene or QTL perhaps functions along with a number of minor genes. As many as ~80 QTLs were known for APR for leaf rust as in the year 201442; 170 more such genes must have been reported since then (manuscript under preparation). After recognizing the pathogen, like ASR, APR also initiates a defense response involving signal transduction and spatial/temporal differential expression of a large number of defense genes43. This aspect has not been studied in detail, although limited information is available3,8,9,10,13,14.
The gene Lr48 involved in the present study is a race-specific hypersensitive APR gene33. In our earlier study, a candidate gene for Lr48 was also inferred, which was shown to encode a receptor like wall associated kinase (WAK) and was shown to be an orthologue of rice gene OsWAK813. In order to understand the Lr48-mediated molecular events involved in downstream signal transduction and defense responses, transcriptome studies have also been conducted8,9,13. We believe that the expression of each APR gene as well as downstream signal transduction and defense responses might be partly regulated through epigenetic modifications44,45,46. In particular, the role of DNA methylation in expression of APR genes for bacterial blight in rice was earlier examined, where a correlation was reported between DNA methylation (examined using MSAP) and repression of gene activity examined using Northern hybridization44.
The present study was focused on an analysis of the effect of the presence of the pathogen on genome-wide DNA methylation in Lr48-mediated APR for leaf rust in wheat genotype CSP44. This allowed us to identify changes in differential methylome in different genic and intergenic regions. We assume that majority of the alteration in DNA methylation occurs of downstream genes, some of which encode proteins for LRR, photosystem I and II, Clp protease, peroxidases, glutaredoxin, kinases, transferases, etc. While analysing the data, we noticed that the methylation level (hypo- as well hypermethylation) during normal growth and development (S0 vs R0) is rather high (Fig. 5, Table 2), which means that a fairly large number of genes undergo hypomethylation (Total 491, unique 130) and another set of relatively smaller number of genes undergo hypermethylation (total 206 DMGs, unique 99) during transition from P-AP to AP stage in the absence of pathogen. This methylation level is affected rather drastically due to the presence of the pathogen. A large number of differentially methylated regions (DMRs) were also available in each of the remaining three treatment pairs (S0 vs S96, R0 vs R96, and S96 vs R96), which suggest that the number of hyper- and hypomethylated DMRs is drastically reduced in the presence of the pathogen during susceptible P-AP stage but not during resistant AP stage (Table 2; Fig.5). Similar results were also observed earlier in case of rice under drought stress, where the proportion of DMRs induced due to drought stress were remarkably low in sensitive line relative to a tolerant line28. However, the situation of DMGs differs from the situation of DMRs. It is apparent that very few genes are differentially methylated during P-AP stage (S0 vs S96) relative to AP stage (R0 vs R96) (Fig. 5). At the resistant AP stage (R0 vs R96), the methylation status of DMGs is just the opposite of the condition observed during normal growth and development (S0 vs R0). At the resistant AP stage, the frequencies of hypomethylated DMGs is greatly reduced while that of hypermethylated DMGs is greatly increased. This appears to be a significant observation suggesting that fewer DMGs are activated relative to a large number of DMGs, which are silenced. Some of the important genes which are activated due to hypomethylation include F-box, wall associated kinase (WAK), glutaredoxin, thioredoxin, armadillo type fold, NB-ARC, LRR, etc. Similarly, the genes which are silenced due to hypermethylation encode proteins for photosystem I and II, ATP synthase, NADH-quinone oxidoreductase, cytochrome C oxidase, etc.
The genes that are activated due to hypomethylation may be perhaps involved in positive regulation of Lr48 mediated APR. Some examples of positive regulation reported in the literature are as follows: (i) in an earlier study in wheat, five F-box genes were found to be upregulated during incompatible interaction in wheat-leaf rust pathosystem suggesting the role of F-box genes during leaf rust resistance50. (ii) A gene encoding TaWAK6 in wheat was recently shown to provide APR during leaf rust infection41. (iii) Peroxidase activity was shown to be involved in providing protection against oxidative burst during leaf rust infection in Aegilops tauschii51. (iv) In rice, the expression of OsGRX20, a type of glutaredoxin, was shown to be induced during bacterial blight (BB) infection in a resistant cultivar52. (v) High expression of genes encoding NB-ARC proteins was observed in incompatible interaction during Lr28 mediated leaf rust resistance53.
The genes that are silenced or downregulated due to hypermethylation may be perhaps involved in negative regulation of Lr48 mediated APR so that silencing of these genes will facilitate APR. There are very few examples of genes that are involved in negative regulation of disease resistance. Therefore, this seems to be a new information, which may need validation in future. For instance, the genes involved in photosynthesis like cyt b6/f of photosystem II are earlier reported to induce resistance through negative regulation. These genes exhibited reduced expression in response to P. syringeae infection in soybean leading to resistance involving hypersensitive reaction. This resistance was induced due to damage of PSII system, which leads to oxidative burst leading to expression of other defense related genes54. In fact, the repression or downregulation of genes involved in photosynthesis during biotic stress like leaf rust in the present study was earlier shown to be an adaptive response to biotic stresses55. Our results are also largely in agreement with the results of an earlier study in Arabidopsis, where DMGs associated with a large number of DMRs were shown to express in response to the bacterial pathogen, Pseudomonas syringae pv. tomato DC3000 (Pst)18.
In the present study, 229 unique DMGs (130 hypo- and 99 hyper-) were observed in the treatment pair S0 vs R0 indicating the role of these genes mostly in transition from P-AP to AP stage during normal growth and development (Fig. 4). The methylation status of majority of these genes changes during susceptible P-AP stage (S0 vs S96); these genes encode proteins for peptidases, transferases, zinc finger proteins, glutaredoxins and thioredoxins, SWEET sugar transporters, expansins, heat shock proteins, FAD binding domains, etc. The role of some of these genes encoding proteins like expansins, FAD binding domains, zinc finger proteins in growth and development has been widely recognized56,57,58.
There are only a few genes (20 hypo- and 29 hypermethylated) which are still differentially methylated during pathogen invasion. These genes include those encoding proteins ferredoxins, aldehyde oxidase, ribosomal proteins, methionine-S-methyl transferase, protein kinases, LRR, etc. The role of these genes during susceptibility needs further examination.
The above results and interpretation receive support from GO analysis. For instance, during susceptibility (S0 vs S96), hypo- and hypermethylated DMGs were found to be different involving completely different functions. These included multi-organ process, reproduction, signalling and biological regulation for hypermethylated DMGs and cellular and metabolic processes for hypomethylated DMGs. On the other hand, the hypermethylated genes during the resistant AP stage were found to be largely involved in structural molecule activity, metabolic and cellular processes, whereas the hypomethylated DMGs during resistance were involved in reproduction, signalling, response to stimulus, etc.
RNA-seq data was also available for approximately 38% (340/897 genes) of all the DMGs identified in the present study13. For 312/340 DMGs (Supplementary Tables S19–22), the expression pattern matched with the level of methylation (assuming that hypermethylation was associated with repression of gene expression and hypomethylation was associated with enhanced gene expression). The results of methylation induced changes in some representative genes were also validated through the results of qRT-PCR analysis, where the expression of 11 out of the 15 genes examined showed inverse relationship with DNA methylation. In an earlier study in rice also, under drought stress, level of methylation in more than half of the DMGs showed negative relationship with gene expression28. Further, in the present study, 76 of the above 312 DMGs, exclusive methylation of the promoter region suppressed gene expression. It is known that DNA methylation in promoter directly represses transcription in one or more of the following ways: (i) inhibition of the binding of transcription activators, (ii) promotion of the binding of transcription repressors, (iii) indirect repression of transcription by promoting repressive histone modifications such as H3K9me2, and (iv) inhibiting permissive histone modifications such as histone acetylation59,60.
Another interesting observation of the present study is an abundance of DMRs in intergenic regions, relative to DMRs in genic regions, which included exons/introns, promoters and other regulatory regions (Table 2). This also received support from two recent studies where >70% of the DMRs were found in the intergenic regions. These included one of our own study involving MeDIP analysis for Lr28-mediated ASR16 and a study involving WGBS during lead (Pb) stress in maize31. The intergenic regions are known to carry many enhancers and other regulatory regions like non coding RNA genes and TEs. Therefore, it is possible that intergenic DNA methylation involves methylation of enhancers and other regulatory sequences encoding non-coding RNAs for trans-regulation as observed in case of genes controlling precursor B-cell acute lymphoblastic leukaemia in mammals61. Similar results were available in our earlier study, where we examined the role of histone modifications in leaf rust resistance17. It is also known that the effect of methylation in gene body and the associated 3’ UTR and 5’ UTR regions differs from the effect of methylation in the promoter sequences. DNA methylation in the gene body region is known to promote transcription involving alternate splicing62, while that in the promoter regions inhibits transcription48,63. The results of the present study are in agreement with this general observation.
During the present study, TEs were also examined in the DMRs, although specific methylation of TEs could not be examined. The number of TEs (located in the gene body, promoter and intergenic regions) was higher in the DMRs at AP stage (R0 vs R96) than the P-AP stage (S0 vs S96) (Table 3; Supplementary Tables S2-S5). This suggested increasing role of possible methylation of TEs in silencing of genes with passage of time from P-AP susceptible stage to AP resistant stage. These results are in agreement with the results of an earlier study64. Among different types of TEs, the LTR retrotransposons like GYPSY, COPIA and Em-spm were abundant among DMRs relative to other types of transposons, particularly in S0 vs R0 and R0 vs R96. Occurrence of TEs and their activity has also been shown to be related with methylation patterns in genomes of Arabidopsis and cassava65,66; GYPSY and COPIA were more heavily methylated than other TEs in cassava67.
The results of DNA methylation at the P-AP and AP stages involving APR gene Lr48 in wheat genotype CSP44 were also compared with methylation status in susceptible and resistant NILs for ASR gene Lr28 in wheat cultivar HD232916, 17. The results revealed a similar pattern of methylation change, i.e. hypermethylation mediated gene silencing in resistant reaction indicating some common epigenetic mechanism involving methylation during ASR and APR. Further observations also revealed two common genes encoding NADPH oxidoreductase and cyt C assembly which were hypomethylated in the present study and also showed differential binding sites for H3K27me3 (a repression mark) in an earlier ChIP-seq analysis17. Similarly, while comparing the differentially methylated regions obtained during ASR due to Lr28 in our earlier study and APR due to Lr48 in the present study, 17 hypermethylated genes and 12 hypomethylated genes were found to be common. These genes encoded proteins for photosystem I, cyt C, cyt b559, different ribosomal proteins, etc. suggesting the role of regulated expression of these genes in disease resistance.