**4. Discussion**

**Library miRNA FI Putative target genes**

22 Plant Engineering

miR169 −3.8 RAPB protein—*Oryza* 

miR395 −15.5 ATP sulphurylase

miR909 5 Inhibin, beta B subunit;

miR1432 4.5 Para-hydroxybenzoate-

miR916 3.3 Zein protein-body ER

miR1877 −3 Putative protein binding

miR2092 2.6 Unknown miR2863 3.5 Unknown

miR5169 2.2 Unknown

miR479 3 Unknown

miR2592 3 Unknown

FI = fold induction compared to control libraries. Inf = infected, L = leaf, R = root, sys = systemic, zma = maize miRNAs.

**Figure 6.** Northern blot analysis of miR395 expression. The signs + indicates *C. graminicola* infection, - control tissue. H= herbivore (*Spodoptera frugiperda*, non-fungus control). The tRNA and 5S rRNA are shown as a control for equal loading

Inf L sys L miR397 2.2 Laccase; multicopper

Inf R sys L miR395 −2.7 ATP sulphurylase

**Table 2.** Maize miRNAs differently regulated upon *C. graminicola* infection.

and were stained with ethidium bromide.

*sativa*; allene oxide synthase—*Zea mays*

(LOC541653), mRNA

vinculin; heavy metal transport/detoxification

polyprenyltransferase (LOC100282174)

protein

oxidase;

protein

membrane protein

(LOC541653), mRNA

It has been documented that plant sRNAs can act as regulators of gene expression during plant‐defence responses as reviewed in Ref. [36]. However, the mechanisms of sRNA-medi‐ ated immunity remain largely elusive, especially for host‐fungi interactions. In rice cultivars that are susceptible to *Magnaporthe oryzae*, enhanced resistance could be achieved by overex‐ pressing miR160 and miR398 [37]. MicroRNA160 targets auxin-responsive factor 16 (*ARF16*), and miR398 regulates superoxide dismutase 2 (*SOD2*), both known defence-related genes. In wheat, *B. graminis* infection was demonstrated to lead to massive adaptations of the miRNA expression profile, where miRNAs only induced in either resistant or susceptible cultivars where identified [19]. Various sRNA expression studies upon fungal infections point towards a role as fine-tuners in the concert of setting up efficient and targeted antifungal defences, rather than having direct defensive impacts [38]. In this regard, it is not surprising that sRNAs were identified as regulators of basal immunity and *R*-gene-mediated resistance. In cotton, bioin‐ formatic approaches revealed over 300 NBS-LRR genes potentially controlled by the miR482 family [39], which cleave NBS-LRR transcripts, resulting in the generation of secondary siR‐ NAs that even enhance the silencing of multiple NBS-LRR genes. *V. dahliae* infection leads to a down‐regulation of miR482, hence to a de‐repression of *R*‐genes.

Notably, plant RNAi pathway components were shown in specific cases to be important for mounting proper antifungal-defence responses. RDR6-deficient plants were found to be more susceptible to *Verticillium* spp. but not to *Botrytis cinerea, Alternaria brassicicola* or *Plectosphaerella cucumerina* [40]. Similarly, *ago4* mutants were discovered to be more susceptible to *B. cinerea* and *P. cucumerina* [41, 42], possibly due to the over-induction of the SA-defence pathway which leads to diminished JA‐defence responses that are important in controlling necrotro‐ phic pathogens.

The present study widens the understanding of the putative role of sRNAs in fine-tuning plant-hormonal pathways during fungal infection. First of all, Arabidopsis sRNA pathway components were demonstrated to be required for antifungal responses against *C. higginsia‐ num*. HYL1- and HEN1-deficient plants were more susceptible than the wild type. The higher susceptibility was accompanied by a de-regulated hormonal response. The sRNA mutants *hen1‐1* and *hyl1‐2* exhibited higher SA-, JA- and ABA-induction levels. The hormonal imbal‐ ance might explain the altered susceptibility to *C. higginsianum*, as enhanced SA is known to be important during biotrophic infections, whereas high JA levels are typical for defence against necrotrophs [43]. On the other hand, mutation of *RDR6* did not affect the susceptibil‐ ity against *C. higginsianum* suggesting that the tasiRNA (trans‐acting siRNA) pathway is not involved in antifungal responses. Recent studies demonstrated that RDR6-deficient plants were more resistant to the hemibiotrophic pathogen *Pseudomonas syringae DC3000*, presumably by a constitutive activation of the SA-dependent-defence pathway. Hence, it was speculated that RDR6 acts as a negative regulator of PTI and basal defence in Arabidopsis [44]. Notably, the hormonal imbalance discovered in the sRNA mutants could only partially explain the altered susceptibility, as higher JA levels were found in *hyl1‐2* mutants, which would lead to an enhanced resistance during the necrotrophic stage of *C. higginsianum*. This suggests that sRNAs act as putative-defence coordinators beyond hormonal pathways. Consequently, the metabolomic analysis uncovered additional layers of sRNA-regulated antifungal responses, namely the proper induction of defence-related secondary metabolites. Especially *hen1‐1* and *hyl1‐2* mutants were found to exert a massively altered defence metabolome, and to a lesser extent also the analysed *rdr6‐15* mutants. sRNAs are known to be directly involved in the regulation of secondary metabolites; overexpression of miR393 for instance was shown to increase levels of glucosinolates and decreases camalexin [45], which indicates that miR393 is involved in the re-direction of the metabolic flows. Similarly, a possible link between miR163 and the biosynthesis of secondary metabolites was described [46]. Loss or overexpression of miR163 alters the transcription of target genes and the profiles of secondary metabolites after induction by the fungal elicitor alamethicin. On the other hand, *rdr6‐15* mutants showed a wild-type-like metabolomic profile in both infected and control-treated conditions, despite the levels of some compounds being slightly different in the mutant after infection. The minor differences in the metabolomic profile between *rdr6‐15* and wild-type plants might be explained by the complex redundancies between the members of these protein families [3, 47]. Notably, some sRNA mutants such as *hen1‐1* exhibit developmental defects, thus the genetic and metabolomic phenotype observed in response to fungi might be significantly affected by developmental pathways. However, this issue underlies all genetic studies using knock-out mutants with severe phenotypes. Using rigid statistical criteria for compound clustering, it is possible to partially differentiate developmental from antifungal responses, as shown in the PCA analysis of infected and control *hen1‐1* mutants.

than having direct defensive impacts [38]. In this regard, it is not surprising that sRNAs were identified as regulators of basal immunity and *R*-gene-mediated resistance. In cotton, bioin‐ formatic approaches revealed over 300 NBS-LRR genes potentially controlled by the miR482 family [39], which cleave NBS-LRR transcripts, resulting in the generation of secondary siR‐ NAs that even enhance the silencing of multiple NBS-LRR genes. *V. dahliae* infection leads to a

Notably, plant RNAi pathway components were shown in specific cases to be important for mounting proper antifungal-defence responses. RDR6-deficient plants were found to be more susceptible to *Verticillium* spp. but not to *Botrytis cinerea, Alternaria brassicicola* or *Plectosphaerella cucumerina* [40]. Similarly, *ago4* mutants were discovered to be more susceptible to *B. cinerea* and *P. cucumerina* [41, 42], possibly due to the over-induction of the SA-defence pathway which leads to diminished JA‐defence responses that are important in controlling necrotro‐

The present study widens the understanding of the putative role of sRNAs in fine-tuning plant-hormonal pathways during fungal infection. First of all, Arabidopsis sRNA pathway components were demonstrated to be required for antifungal responses against *C. higginsia‐ num*. HYL1- and HEN1-deficient plants were more susceptible than the wild type. The higher susceptibility was accompanied by a de-regulated hormonal response. The sRNA mutants *hen1‐1* and *hyl1‐2* exhibited higher SA-, JA- and ABA-induction levels. The hormonal imbal‐ ance might explain the altered susceptibility to *C. higginsianum*, as enhanced SA is known to be important during biotrophic infections, whereas high JA levels are typical for defence against necrotrophs [43]. On the other hand, mutation of *RDR6* did not affect the susceptibil‐ ity against *C. higginsianum* suggesting that the tasiRNA (trans‐acting siRNA) pathway is not involved in antifungal responses. Recent studies demonstrated that RDR6-deficient plants were more resistant to the hemibiotrophic pathogen *Pseudomonas syringae DC3000*, presumably by a constitutive activation of the SA-dependent-defence pathway. Hence, it was speculated that RDR6 acts as a negative regulator of PTI and basal defence in Arabidopsis [44]. Notably, the hormonal imbalance discovered in the sRNA mutants could only partially explain the altered susceptibility, as higher JA levels were found in *hyl1‐2* mutants, which would lead to an enhanced resistance during the necrotrophic stage of *C. higginsianum*. This suggests that sRNAs act as putative-defence coordinators beyond hormonal pathways. Consequently, the metabolomic analysis uncovered additional layers of sRNA-regulated antifungal responses, namely the proper induction of defence-related secondary metabolites. Especially *hen1‐1* and *hyl1‐2* mutants were found to exert a massively altered defence metabolome, and to a lesser extent also the analysed *rdr6‐15* mutants. sRNAs are known to be directly involved in the regulation of secondary metabolites; overexpression of miR393 for instance was shown to increase levels of glucosinolates and decreases camalexin [45], which indicates that miR393 is involved in the re-direction of the metabolic flows. Similarly, a possible link between miR163 and the biosynthesis of secondary metabolites was described [46]. Loss or overexpression of miR163 alters the transcription of target genes and the profiles of secondary metabolites after induction by the fungal elicitor alamethicin. On the other hand, *rdr6‐15* mutants showed a wild-type-like metabolomic profile in both infected and control-treated conditions, despite the levels of some compounds being slightly different in the mutant after infection. The

down‐regulation of miR482, hence to a de‐repression of *R*‐genes.

phic pathogens.

24 Plant Engineering

To extend the view on antifungal responses possibly linked to sRNA pathways, the miRNA transcriptome of the agricultural important model crop *Z. mays* infected with *C. graminicola* was analysed. During this interaction, maize was found to set-up an organ-specific miRNA profile. In the locally and systemically induced fungal-specific miRNAs, only a few were found to target defence genes. In particular, zma‐miR1432, which targets a para‐hydroxy‐ benzoate-polyprenyl transferase, was found to be up-regulated during both biotrophic and necrotrophic infection stages in maize leaves and roots. The miRNA target is essential in terpenoid-quinone synthesis. Hence, it could be speculated that its down-regulation could divert the flow of secondary metabolites from terpenoids towards flavonoid biosynthesis. This would be coherent with the fact that terpenoids play only minor roles during *C. gramini‐ cola* infection in maize [48]. The second identified miRNA linked to defence pathways was zma‐miR169. This miRNA is down‐regulated in response to fungal root infections, and it putatively targets a gene encoding an allene oxide synthase (*AOS*). AOS is a key enzyme in JA synthesis, thus it can be speculated that zma-miR169 acts as a suppressor of JA signalling under non‐stressed conditions, whereas the down‐regulation of zma‐miR169 during fungal infection could promote JA synthesis. The enhanced JA levels in some Arabidopsis miRNA mutants support this hypothesis. Another yet intriguing altered miRNA was zma‐miR395, which was down‐regulated in *C. graminicola*‐infected maize roots and systemic leaves upon root infection. The down-regulation of zma-miR395 was accompanied by the up-regulation of two of its putative targets in roots, one of them encoding an ATP sulphurylase. APS plays an important role in sulphate assimilation and glutathione synthesis; inhibiting glutathione synthesis in Arabidopsis was shown to trigger the suppression of miR395 [49], thus mimick‐ ing fungal infection. It can be speculated that the down-regulation of zma-miR395 positively regulates sulphate‐mediated defence and/or the glutathione pathway. Intriguingly, miR168 that targets AGO1 was induced upon leaf infection, consistent with recent work demonstrat‐ ing a similar fold induction of miR168 in Arabidopsis treated with elicitors of *F. oxysporum* [50]. Thereby, a majority of elicitor-responsive miRNAs were shown to be associated with development and miRNA homeostasis [50], corroborating the observation that sRNA path‐ ways likely do not regulate direct-defence pathways.

Although some sRNA pathways components were shown here to be required for battling *C. higginsianum*, they seem to act as fine-tuner of defence schemes rather than to directly regu‐ late defence genes and defensive compounds. A common picture found for sRNA mutants exhibiting higher susceptibility to *C. higginsianum* was rather a hormonal and metabolomic imbalance. Moreover, the absence of altered miRNAs targeting direct-defence genes in maize suggested an indirect defensive role of sRNAs against *Colletotrichum* spp.
