This study investigated the global expression profile of two apple cultivars, ‘GD’ and ‘M9’, during infection by the canker pathogen N. ditissima. By employing a RNA-seq approach we were able to elucidate the potential mechanisms involved in partial resistance to European canker in cultivated apple. Furthermore, we utilised individuals from a full-sibling family of apple to study the genetic mechanisms underlying QDR to this wood pathogen.
Global trends in response to infection with N. ditissima
The transcriptome analysis revealed that a large number of genes (> 5,000) were differentially regulated in each of the cultivars ‘GD’ and ‘M9’ at approximately a month after infection with N. ditissima. Apple stem tissue was sampled at two distances from the canker lesion to compare expression profiles in cells adjacent to the symptomatic tissue as well as cells located at a further distance to the diseased tissue.
Overall, a similar number of genes were DE in the two apple cultivars at P1. However, there was a substantial proportion of unique DEGs in each cultivar at P1; 34% and 31% for ‘GD’ and ‘M9’, respectively. This indicates that a subset of different genes were activated in each cultivar due to infection, indicating different underlying resistance mechanisms. The use of two cultivars with a larger difference in resistance to European canker may have facilitated the identification of expression patterns specific to partially resistant cultivars.
Whilst there were fewer DEGs at P2 the difference between cultivars was proportionally larger. This showed that a shift in expression appeared at a further distance from the point of infection in ‘GD’ compared to ‘M9’, with approximately 2,000 versus 900 DEGs in the respective cultivar at P2. The wider differential response in ‘GD’ to N. ditissima infection may be indicative of a more rapid and/or more systemic response to the pathogen in this cultivar compared to ‘M9’.
The enrichment analyses for genes annotated with different functional classes show that a range of molecular processes are affected as a result of N. ditissima infection. The protein family analysis indicated an enrichment of genes involved in pathogen and chitin recognition, hormone signalling, response and transport of toxins and xenobiotics, secondary metabolism as well as sugar and carbon metabolism. The results suggest that the host response to N. ditissima is mediated through a combination of extracellular and intracellular immune receptors (31, 32). Pathogen recognition then activates hormone signalling and altered metabolism of sugars, carbon and secondary metabolites. The observed changes in genes associated with catabolism and transport of toxins and xenobiotics (e.g multidrug and toxic compound extrusion (MATE) family genes, GSTs) could either have a role in attenuating host induced oxidative stress and endogenous metabolites or the detoxification of toxins produced by the pathogen (33, 34). The percentage of genes that could be annotated with KEGG or GO terms was low (28 and 47%, respectively) which could potentially skew the enrichment analysis. However, the PFAM enrichment analysis described a similar picture and was based on a larger percentage of annotated genes (79%).
The responses observed here are similar to what has been observed during infection of Valsa mali, another fungal pathogen causing cankers in apple, including pathways related to phenylpropanoid biosynthesis, starch and sucrose metabolism, plant-hormone signal transduction and plant-pathogen interaction (11).
N. ditissima infection alters expression of genes involved in the phenylpropanoid pathway and lignification
The phenylpropanoid pathway is responsible for the biosynthesis of a wide array of secondary metabolites derived from the deamination of phenylalanine to cinnamic acid by phenylalanine ammonia-lyase (PAL) (35). Cinnamic acid is then further converted in order to produce the plant cell wall components lignin and suberin as well as coumarins, flavonoids and stilbenes. DEGs associated with the phenylpropanoid pathway were identified in both cultivars and sampling positions in this experiment, with 61% of those identified showing an increased abundance after infection. A third of the DEGs annotated to the phenylpropanoid pathway were peroxidases. Moreover, 60 DEGs were predicted to be laccases. Although peroxidases and laccases perform various functions in plants, both are involved in the polymerization of monolignols to form lignin (36). A cell-wall degrading enzyme (CWDE) from Botrytis cinerea was shown to alter expression of peroxidases and genes in the phenylpropanoid pathway as well as increase lignin content in tomato (37). Furthermore, laccase and peroxidase activity have been associated with altered lignification and resistance to plant pathogens for some time (38–41). A further support for the importance of the phenylpropanoid pathway in partial resistance to N. ditissima is the identification of two putative 4CL genes (MD16G1112900 and MD16G1113000) within the QTL interval on chr 16. Both genes were significantly more highly expressed in apple progeny with the QTL16-R QDR allele. 4CL is a key enzyme in the beginning of the phenylpropanoid pathway, in which it catalyses the conversion of hydroxycinnamates into corresponding CoA esters for biosynthesis of flavonoids and lignin (42). The activity of 4CLs, and the subsequent accumulation of lignins have previously been linked to plant pathogen resistance in multiple crops (43–45). Furthermore, two putative CYP genes were identified within the QTL regions on chr 2 and 16. CYPs belong to a large enzymatic gene family with important functions in the synthesis of secondary metabolites (46). Our results suggest that one of the responses in apple trees to N. ditissima infection is a shift in expression of phenylpropanoid pathway genes and altered lignin accumulation through peroxidase and laccase activity. However, further studies would be required to evaluate the relative importance of lignin biosynthesis and phenylpropanoids in QDR to European canker.
Pathogen interaction
The infection with N. ditissima altered the expression of a multitude of genes involved in pathogen recognition, including PRs, NLRs, RLKs, and genes with LysM-domains. Our results indicate that N. ditissima is recognised by the apple host by a combination of basal immunity and more specialised NLRs. Nevertheless, it is not clear whether the NLRs have a role in QDR or have been hijacked by the pathogen to function as susceptibility genes (9).
We identified several candidate genes with a role in pathogen interaction within QTL that have been associated with partial resistance to N. ditissima in scion apple germplasm (Karlstrom et al., 2022). Clusters of putative WAKs and WAKLs were identified in the QTL intervals on chr 10 and 2, respectively. WAKs are usually characterised by the three following domains; a serine/threonine kinase, an epidermal growth factor (EGF) and a galacturonan-binding (GUB) domain. Compared with WAK, WAKL usually lacks the extracellular EGF domain (10). WAK/WAKL-genes are a sub-family of RLKs that function in plant growth and stress-response and that in many cases acts as a positive regulator in plant immune response (10). Nevertheless there are examples of host-pathogen systems where WAKs have been shown to negatively regulate host resistance (10, 47, 48). Furthermore, a genome-wide study of WAKs in apple showed these to be both positively and negatively regulated as a response to infection with V. mali, Alternaria alternata and Pythium ultimum (49). There were three WAKs among the candidate genes within the QTL interval on chr 10. All of these had a significantly lower expression in trees with the QDR allele at QTL10. Furthermore, the gene was significantly down-regulated in ‘M9’ upon infection. A cluster of 14 putative WAKLs were identified as candidates underlying the QTL on chr 2. Five of these WAKLs (MD02G1249500, MD02G1273500, MD02G1273700 and MD02G1254300) are particularly interesting as candidate genes as they were validated in the transcriptome data from ‘GD’/’M9 and more highly expressed in apple trees with the QTL2-R allele.
WAK/WAKL receptors exhibit a tendency to interact with a diverse array of pathogens (10), which suggests that they could confer resistance to other pathogens in addition to N. ditissima. Backing this hypothesis, researchers have identified a QTL on the distal end of chromosome 2 that correlates with intermediate resistance to multiple isolates of apple scab (Venturia inaequalis, (50, 51). Furthermore, a putative disease resistance cluster, believed to contribute to resistance against fire blight (Erwinia amylovora), scab, and powdery mildew (Podosphaera leucotricha), is proposed to reside on chromosome 10 (52). These genetic regions lie approximately 1–2 Mbp and 16–28 Mbp away from the WAKL and WAK gene clusters identified in this study on chromosomes 2 and 10, respectively (51–53). Consequently, it can be inferred that the WAK/WAKL genes on chromosomes 2 and 10 are implicated in interactions with multiple pathogens. There is a growing body of evidence indicating the significant role of WAKs/WAKLs in plant-pathogen interactions. Similarly to the situation with N. ditissima, several described WAKs are involved in resistance against hemibiotrophic or necrotrophic Ascomycete fungi of the Dothideomycetes class (10). Their defensive functions span from the detection of effectors or other molecules indicative of pathogen invasion to the initiation of callus deposition and lignin biosynthesis (10)
Several putative NLRs were DE in QTL regions on chr 2, 8, 15 and 16 when comparing individuals with or without a QDR allele at each QTL. However, the association of these genes to N. ditissima infection could not be confirmed in the transcriptome data from ‘GD’ or ‘M9’.
Putative genes underlying quantitative resistance to N. ditissima
Apple trees rely on quantitative resistance to combat infection with N. ditissima. We dissected six QTL associated with QDR to European canker, in order to understand the mechanisms that underpin tolerance to this wood pathogen. In addition to the above mentioned roles of candidate genes in the phenylpropanoid pathway and pathogen interaction, genes with several other functions were identified. A putative UGT was DE within the QTL interval on chr 6. The UGT gene family was also significantly enriched after canker infection in both ‘GD’ and ‘M9’, and > 38% of the genes annotated to this protein family were DE at the sampling positions closest to the pathogen lesion. UGT is a very large superfamily of enzymes in plants, which catalyse glycosidation. These enzymes have been linked to QDR in multiple species through the glycosylation of endogenous phytohormones, defensive compounds and other secondary metabolites but also by reducing the toxicity of pathogen derived xenobiotics (54). The UGT on chr 6, MD06G1103300, is particularly interesting as a candidate gene as it was strongly up-regulated in infected trees with the QDR allele on chr 6 (Fig. 6).
Transcription factors were identified as candidate genes within the QTL regions on chr 8 (heat shock transcription factor), 10, 15 and 16. This group of proteins are known to be important in the transcriptional reprogramming that occurs in response to pathogen infection (55).
Among the candidate genes within the QTL interval on chr 8 was a putative HIPP encoding gene, which had a lower expression in QTL-R trees. HIPP genes have been described as targets of multiple necrotrophic pathogens (56, 57) and as susceptibility genes in nematode-plant interactions (58, 59). The most well-described HIPP in plant disease is probably HIPP05 from Oryza sativa, also known as Pi21, which functions as a susceptibility factor in interactions with the necrotrophic pathogen Magnaporthe oryzae. Loss-of-function mutations of this gene results in field resistance to the pathogen (60)– while overexpression in the non-host Arabidopsis has been shown to result in increased pathogenicity of M. oryzae on this species (61).
This study used a transcriptome approach to identify candidate genes associated with multiple resistance QTL to European canker in apple. However, only a limited subset of the genes that were DE between QTL-R and QTL-S plants could be validated in expression data from ‘GD’ or ‘M9’, despite the presence of QDR alleles for all QTL in ‘GD’ (Table 1). There could be several reasons for this; 1) differences in timing of sampling of infected tissue between progeny and validation. The differences in number of DEGs between P1 and P2, as well as the differences between ‘GD’ and ‘M9’ in their response at these two positions, shows that infection stage has a large influence on gene expression in the host. The progeny and validation were sampled at the same disease stage and at the same distance from the active lesion. However, the response to different disease progression stages will vary between genotypes and even small variations may influence gene expression. 2) The validation was specific to genes which were differentially expressed upon infection by N. ditissima and would therefore miss constitutively expressed genes. 3) Differences between QTL-R and QTL-S are spurious and due to allelic variation but not related to the response to N. ditissima. This is however an unlikely explanation for genes which are only DE in infected trees.
In addition, the “true” genes underlying resistance QTL may not have been detected in this study. The DE-analysis compared the effect of single alleles on gene expression, ignoring the effects of background QTL. This could potentially hinder the identification of candidate genes if the QTL has a small effect on overall disease progression. The parents of the segregating progeny harboured different haplotype alleles associated with QDR for QTL 8. However, the QDR genes underlying the two alleles were assumed to be the same. This assumption may be incorrect and there may indeed be different genetic variation underlying the resistance for each allele. Furthermore, the low representation of genotypes with no QDR allele for QTL6 and QTL8 may have resulted in a limited power to detect differences in transcript abundance for these QTL.
Despite above limitations we have identified a number of candidate genes associated with resistance loci to N. ditissima. Upon functional characterization, these can pave the way to developing highly canker resistant apple varieties.