**3. Results**

#### **3.1. Arabidopsis sRNA mutants show different levels of susceptibility to** *C. higginsianum*

To test if a functional silencing machinery is required for a proper antifungal-defence response, *A. thaliana* wild types Ler-0 and Col-0 showing lower and intermediate resis‐ tance, respectively, and sRNA pathway mutants were subjected to fungal infection assays to monitor the susceptibility to *C. higginsianum*. To cover important components of sRNA pathways, the loss of function mutants for the genes encoding *HYL1, HEN1*and *RDR6* was analysed. The sRNA pathway mutants were infected with *C. higginsianum*-GFP, and the disease progression was compared to the relative wild‐type ecotype, for *hen1‐1* namely *Landsberg erecta* (Ler-0), for all other mutants Columbia (Col-0). Fungal growth was moni‐ tored at 24, 48, 72 and 96 h postinfection (hpi) (**Figure 3**). These time points were chosen to cover all known infection stages of *C. higginsianum* during hemibiotrophic growth on leaves [25, 29]. The infection assays showed an altered susceptibility of mutants (**Figure 3**). For *hen1‐1,* a significant higher susceptibility was only detected in late infection stages (96 hpi). Comparison of *hyl1‐2* with Col-0 yielded statistically significant differences of fungal growth at all time points (**Figure 3(b)**). The RNA mutant was found to be more susceptible to *C. higginsianum* compared to the wild type. By contrast, *rdr6‐15* was infected by *C. higgin‐ sianum* as efficiently as the wild type (**Figure 3(c)**). Altogether, a defective sRNA machinery seems to render plants more susceptible to fungal attack. However, mutations in *RDR6‐15* did not alter the susceptibility against the *C. higginsianum*.

**2.7. Deep sequencing and Northern blotting of maize sRNAs**

**2.8. Identification and quantification of conserved miRNAs**

blotting techniques as described [33].

14 Plant Engineering

**2.9. Target prediction of maize miRNAs**

**2.10. Statistical analysis**

sigmaplot.com).

**3. Results**

genomic project. Default settings were applied.

For sRNA library preparation, six biological replicates were pooled and total RNA was iso‐ lated using Trizol (Invitrogen, www.invitrogen.com); 10 µg of total RNA was further pro‐ cessed using an Illumina-Solexa deep-sequencing approach at FASTERIS (http://www. fasteris.com). The expression of selected miRNAs was further analysed using sRNA Northern

To identify conserved maize miRNAs, sequences of 4677 mature plant miRNAs were down‐ loaded from miRBase (release 18.0, November 2011). Identical miRNA sequences identified in different species or duplicated loci in a genome were collapsed, resulting in a non-redun‐ dant list consisting of 2228 unique miRNAs. Sequences belonging to the same miRNA family were further analysed by multiple alignment using ClustalW (www.clustal.org) and classified in subgroups to distinguish bona fide mature miRNAs from misannotated miRNA\* forms or sequences generated from different regions of the same precursor. This non-redundant library was then applied to screen the small RNA libraries. All the small RNA reads in the range of 20–24 nt in size, and which are represented and represented by at least two reads in a library were aligned to the 1772 unique miRNAs derived from miRBase. For the screening, a maximum of three mismatches was allowed and up to 2 nt overhanging nucleotides at the 5' and/or 3' end. Alignments were performed using SeqMap [34]. The output was filtered and reformatted with

Putative targets of maize miRNAs were identified using the psRNATarget web server (http:// bioinfo3.noble.org/miRU2/) against *Z. mays* DFCI Gene index (version 19) and *Z. mays* PlantGDB

Variances of quantified levels of metabolites and fungal growth for multiple groups were analysed by a one-way analysis of variance (ANOVA); a *P*‐value of <0.05 was considered significant. The Mann–Whitney *U*-test was used to compare significant differences between two sample groups. All statistical analysis was performed using Sigma Plot 11.0 (http://www.

**3.1. Arabidopsis sRNA mutants show different levels of susceptibility to** *C. higginsianum*

To test if a functional silencing machinery is required for a proper antifungal-defence response, *A. thaliana* wild types Ler-0 and Col-0 showing lower and intermediate resis‐ tance, respectively, and sRNA pathway mutants were subjected to fungal infection assays

custom PERL scripts, classifying the identified miRNAs according to miRBase.

#### **3.2. Arabidopsis sRNA mutants show an altered hormonal balance after** *C. higginsianum* **infection**

Hormone signalling is a key process that regulates stress responses. To evaluate the implica‐ tion of sRNA pathways in hormone‐mediated plant defence against *C. higginsianum*, levels of salicylic acid, jasmonic acid and abscisic acid were quantified by HPLC-MS/MS. All selected mutants and wild‐type accessions were analysed 4 days post *C. higginsianum* infection and hormone levels of both infected and mock were measured. In response to *C. higginsianum* attack, SA and JA were induced to different levels in all genotypes (**Figure 4**). Notably, SA and JA inductions were more pronounced in the mutants *hen1‐1* and *hyl1‐2* compared to their respective wild-type (Ler and Col-0) infected plants. For instance, in infected *hen1‐1* plants, JA levels rose up to 589 ng/100 mg fresh weight, whereas in infected Ler plants, JA only reached 234 ng/100 mg fresh weight. However, *rdr6‐15* did not appear to have significant differences of SA and JA levels compared to wild‐type‐infected plants (**Figure 4(a)(b)**). On the other hand, ABA levels were found to be induced during fungal infection in *hyl1‐2* and *hen1‐1* contrary to *rdr6‐15* that show no significant changes in ABA quantity upon fungal infection (**Figure 4(c)**). These results suggest that the sRNA mutant *rdr6‐15* is likely not implicated in the regulation of hormone levels during antifungal responses, whereas a functional HEN1 and HYL1 pro‐ tein seems to be required to mount a full SA, JA and ABA response to fungal attack.
