Down- and up-regulation of methylation-related genes
In the present study, hierarchical clustering based on the Euclidean distance showed that 16 methylation-related genes were down-regulated in a susceptible line, 2-2890 (Fig. 1). Corresponding genes were up-regulated in a resistant line 2-2618 upon F. graminearum inoculation.
Genomic DNA methylation levels have been shown to change in plants to adapt to the stresses including pathogen infections (Viggiano and de Pinto 2017). By this epigenetic mechanism, stable alteration in gene expression without changes in the underlying DNA sequence occurs that results in normal plant development (Verhoeven et al. 2010; Zhang et al. 2010; Ganguly et al. 2017; Hewezi 2018).
In the present study since a susceptible line, 2-2890 could not adapt to the F. graminearum stress 16 methylation-related genes were seen to be down-regulated. This indicates that genomic DNA methylation did not occur in a susceptible line and hence surrendered to the pathogen infection resulting in a rapid progression of the FHB disease. However, in a resistant line, 2-2618 genomic DNA methylation did occur since the methylation-related genes were up-regulated checking the spread of the FHB disease.
Gene ontology (GO) analysis associated these methylation-related genes with L-methionine salvage from methylthioadenosine and S-adenosylmethionine and steroid biosynthesis (p-value 0.001) (Fig. 2).
In the present study GO analysis of methylation-related genes suggested that methylation of the plant DNA might activate the steroid biosynthesis pathway in a resistant line. This plausibly helped protect the wheat plants from infection upon F. graminearum inoculation. In plants, various steroid products such as steroid glycoalkaloids and sesquiterpenoid phytoalexins, sterols, brassinosteroids and cytokinins, farnesyl and geranyl groups, dolichols and ubiquinone have been reported to require for defence-related functions including membrane biogenesis, control of growth and development, protein prenylation, protein glycosylation and respiration (Stermer et al. 1994; Chappell 1995).
Co-expression analysis of methylation-related genes
Co-expression analysis of methylation-related genes in FHB S 2-2890 NIL with methionine S-methyl-transferase gene (MSM; TraesCS1A02G013800) resulted in 3-hydroxy-3-methyl-glutaryl coenzyme A reductase (HMGCR; TraesCS5A02G269300) (Fig. 3).
The HMGCR catalyzes the first essential step of a mevalonate (MVA) pathway for isoprenoid biosynthesis (Caelles et al. 1989). This pathway provides precursors for a wide variety of isoprenoid products that are required for very diverse functions, including defence against pathogen attacks. In the present study, a negative correlation (-0.82) between methylation-related genes and a gene encoding MSM in FHB S 2-2890 NIL was found (Fig. 3). This indicates that although the expression levels of methylation-related genes might have increased initially upon pathogen infection the MSM levels could have been decreased. This is possible because of a certain counter-defence mechanism of pathogen against MSM, leading to the suppression of MSM encoding gene during the further infection process. As a result, a lower level of HMGCR could have been produced in a susceptible line and hence, less isoprenoid biosynthesis. Varying levels of HMGCR isoforms (from higher levels of HMGR1S to 10 times lower levels of HMGR1L and HMGR2) in inflorescences, seedlings and roots of Arabidopsis thaliana in response to a variety of stress conditions have been reported previously (Enjuto et al. 1994; Enjuto et al. 1995; Lumbreras et al. 1995). Plant HMGCR has also been reported to be modulated by myriad endogenous signals and external stimuli including pathogen attack (Stermer et al. 1994; Rodríguez-Concepción et al. 2013).
Further, the HMGCR was found to be negatively correlated (-1.00) to genes encoding pathogenesis-related and detoxification proteins and xylanase inhibitors in the present study (Fig. 4). This means that as the level of HMGCR decreases the levels of pathogenesis-related and detoxification proteins and xylanase inhibitors increases in the pathogen-infected plants. This suggests that a decrease in HMGCR level is needed for a blockage of certain biosynthetic routes to redirect photoassimilates and energy for the urgent synthesis of adaptive compounds. However, probably due to the less efficient recognition or signal response pathways the susceptible plants may not take advantage of this defence mechanism. The changes in HMGR levels may occur at different rates, depending on the plant system and the type and severity of the stress condition as suggested previously (Leivar et al. 2011). These results suggest that the synthesis of HMGCR is independent of the synthesis of adaptive compounds such as pathogenesis-related and detoxification proteins and xylanase inhibitors. We speculate that the two different pathways for synthesis of HMGCR and adaptive compounds respectively might get activated sequentially upon pathogen infection. However, plants use HMGCR modifications to expand their adaptations against stresses for maximizing energy consumption efficiency and resistance responses when needed.
GO analysis of the negatively correlated methylation-related genes
To study the regulation of DNA methylation in plants in response to pathogen infection the gene set enrichment analysis was carried out and the results are displayed on the Venn diagram (Fig. 5). GO analysis associated the genes, encoding pathogenesis-related and detoxification proteins as well as xylanase inhibitors, with methionine S-methyl transferase activity (p-value 0.001). The transferase activity and methionine S-methyl transferase activity shared the most co-expressing genes (3401 + 4 + 1 = 3406), indicating similar functions between these two gene groups.
Results of the GO analysis suggest that the methyltransferases regulate the expression of defence-related genes including PR proteins, detoxification proteins and xylanase inhibitors. Recently, DNA methylations were shown to have a significant role in plant resistance responses by controlling the expression level of the defence genes (Geng et al. 2019). The expression of PR2 genes was activated in methylation-defective mutants of a diploid wheat species obtained through silencing of a methylation-related gene DRM2. DNA hypomethylation caused due to infection of an obligate biotrophic fungus Blumeria graminis f. sp. Tritici was thought to upregulate some defence-related genes in that study. Similarly, methylation-defective mutants of Arabidopsis were found to be more susceptible than the wildtype to Botrytis cinerea and Plectosphaerella cucumerina (López et al. 2011). Previously several genes encoding methyltransferases have been reported to confer resistance in plants against various pathogens (He and Dixon 2000; Seo et al. 2001; Xu et al. 2006; Berr et al. 2010; Yang et al. 2017; Salvador-Guirao et al. 2018).
2 – Gene plot analysis of negatively correlated genes
Expression levels of 3-hydroxy-3-methyl-glutaryl coenzyme A reductase (HMGCR) were higher (Log2 levels from 3.25 to 4.00) in pathogen inoculated compared to methionine S-methyl-transferase (MSM) (Log2 levels from 1.25 to 3.25) in mock-inoculated FHB S line, 2-2890.
HMGCR was also reported in cotton and potato following inoculation of Verticillium dahliae and Phytophthora infestans, respectively in previous studies (Joost et al. 1995; Yoshioka et al. 1996). Expression of HMGCR in pathogen inoculated susceptible lines in the present study indicates its specialized role in plants under stress conditions including pathogen infection. Whereas, the expression of MSM in the mock-inoculated susceptible line indicates its generalized role in the plant’s normal DNA methylation processes. Continuous expression of HMGCR has also been reported in a susceptible cotton cultivar by Joost et al. 1995. As HMGCR is required for the biosynthesis of isoprenoids (Caelles et al. 1989) its increased levels may lead to a burst of isoprenoids, halting the initial spread of a pathogen. However, due to the phytotoxic effects of excessive isoprenoids, the susceptible plants may succumb to further infection processes as their recognition or signal response pathways are weak.
Effects of miRNAs on methylation-related genes
Forty-three methylation-related genes were down-regulated by a miRNA miR9678 (Fig. 7a). These genes were enriched in the GO categories of responses to biotic stimulus and glucan endo-1,4-beta-glucanase among others (Fig. 7b).
Previously miRNAs have been found to target the methyltransferases and therefore reported as the negative regulators of DNA methylation (Das et al. 2010). The authors affirmed that a miR153 targets the methyltransferase DNMT1 and downregulates the DNA methylation of DNMTt targeted genes. From our results, we speculate that the miR9678:methyltransferase-RNA duplexes may lead to the down-regulation of methyltransferase genes. This is evident from a previous study conducted by Khraiwesh et al. (2010) on miR166-target interaction in moss (Physcomitrella patens). The authors revealed the involvement of miRNA:target-RNA duplexes through interactions of miR166 with its target genes, PpC3HDZIP1 and PpHB10, respectively in a ΔPpDCL1b mutant. Although this is not direct evidence for miR9678 this miRNA has been shown to modulate abscisic acid/gibberellin (ABA/GA) signalling in wheat by targeting a long noncoding RNA (WSGAR) (Guo et al. 2018). Overexpression of this miRNA in transgenic wheat lines reduced the GA content and affected the expression of WSGAR genes. However, there are possibilities that miR9678 might target genes other than WSGAR and the addition of genome sequences to available wheat genomic resources may support this hypothesis. To this end, the downregulation of methylation-related genes in the present study might result from the suppression of certain miRNA9678 target genes including the methyltransferases. Guo et al. (2018) also proposed different genes, such as α-amylase and genes required for the production of specific GAs (i.e., TaCPS, TaKO, and TaKS), as the targets of miR9678.
Guo et al. (2018) proved the role of the ABA-miR9678-GA mechanism in wheat seed germination; however, the role of this mechanism in plant defence can be ascertained from the present study. Some plant pathogenic fungi including F. graminearum have been reported to produce ABA via an MVA biosynthetic pathway (Siewers et al. 2004; Siewers et al. 2006; Qi et al. 2016). ABA produced by F. graminearum seems to suppress the wheat immune responses as reported earlier (Lievens et al. 2017). Upon F. graminearum infection ABA signalling regulators in wheat like TaVP1/TaABF/TaABI5 may bind to the promoter of miR9678 and activate its expression. In a previous study miR9678 has been shown to upregulate upon ABA treatment (Guo et al. 2018). Once the expression of miR9678 increases upon F. graminearum infection the miR9678:methyltransferase-RNA duplexes may lead to downregulation of methyltransferase genes. Due to the downregulation of methyltransferase genes, methylation of the genes responsive to biotic stimulus including glucan endo-1,4-beta-glucanase may not occur. This may lead to the susceptibility of a host plant to the invading pathogen. Previously ABA has been shown to increase the susceptibility of Arabidopsis to F. oxysporum (Anderson et al. 2004; Trusov et al. 2009). Similarly, exogenous applications of ABA has been shown to increase the susceptibility of rice and barley to Magnaporthe grisea and M. oryzae, respectively (Koga et al. 2004; Ulferts et al. 2015).
Like plants GA, a phytohormone can also be produced by many plant pathogens including species of Fusarium (Yabuta et al. 1950; MacMillan 2001). However, exogenous application of this phytohormone on plants has been shown to induce defence-related genes in plants in response to pathogens (Casacuberta et al. 1992). Subsequent studies have also reported that GA enhances the resistance of plants to diseases caused by ascomycetes fungi (Eshel et al. 2002; Tanaka et al. 2006). Although the functioning of GA has not been studied during plant-pathogen interactions we speculate that the resistant plants produce higher levels of GA than ABA upon pathogen infection. These higher levels of GA may misregulate the expression of pathogen genes that are involved in early to later-stage infection events. Repression of nitrogen metabolic genes of F. graminearum upon GA application has been reported to reduce bioenergetic resources and redox regulation during early-stage infection (Buhrow et al. 2016). The authors also suggested that GA may regulate gene expression in F. graminearum during later-stage infection events.
In the present study, we provide theoretical evidence about the ABA-miR9678-GA mechanism that affect defence responses in plants against invading pathogens (Fig. 8a, b). Once F. graminearum infects both susceptible and resistant wheat plants it produces higher levels of ABA to facilitate its invasion. Following this, ABA signalling regulators in susceptible wheat plants binds to a promoter of miR9678 and increases its expression; however, in resistant wheat plants levels of GA increases. Once the expression of miR9678 increases the miRNA downregulates its target methyltransferase genes in susceptible plants. Whereas, in resistant plants, the increased levels of GA misregulate the expression of pathogen genes. In susceptible plants, methylation of genes responsive to biotic stimulus including glucan endo-1,4-beta-glucanase may not occur due to downregulated methyltransferase genes. This may lead to the progression of pathogen infection in susceptible plants. In the case of resistant plants, misregulation of the pathogen’s metabolic genes required for the establishment and progression of infection events may result in enhanced resistance.