Small RNA discovery in the interaction between barley and the powdery mildew pathogen

Abstract

Background: Plants encounter pathogenic and non-pathogenic microorganisms on a nearly constant basis. Small RNAs such as siRNAs and miRNAs/milRNAs influence pathogen virulence and host defense responses. We exploited the biotrophic interaction between the powdery mildew fungus, Blumeria graminis f. sp. hordei (Bgh), and its diploid host plant, barley (Hordeum vulgare) to explore fungal and plant sRNAs expressed during Bgh infection of barley leaf epidermal cells.

Results: RNA was isolated from four fast-neutron immune-signaling mutants and their progenitor over a time course representing key stages of Bgh infection, including appressorium formation, penetration of epidermal cells, and development of haustorial feeding structures. The Cereal Introduction (CI) 16151 progenitor carries the resistance allele Mla6, while Bgh isolate 5874 harbors the AVR_a6 avirulence effector, resulting in an incompatible interaction. Parallel Analysis of RNA Ends (PARE) was used to verify sRNAs with likely transcript targets in both barley and Bgh. Bgh sRNAs are predicted to regulate effectors, metabolic genes, and translation-related genes. Barley sRNAs are predicted to influence the accumulation of transcripts that encode auxin response factors, NAC transcription factors, homeodomain transcription factors, and several splicing factors. We also identified phasing small interfering RNAs (phasiRNAs) in barley that overlap transcripts that encode receptor-like kinases (RLKs) and nucleotide-binding, leucine-rich domain proteins (NLRs).

Conclusions: These data suggest that Bgh sRNAs regulate gene expression in metabolism, translation-related, and pathogen effectors. PARE-validated targets of predicted Bgh milRNAs include both EKA (effectors homologous to AVR_k1 and AVR_a10) and CSEP (candidate secreted effector protein) families. We also identified barley phasiRNAs and miRNAs in response to Bgh infection. These include phasiRNA loci that overlap with a significant proportion of receptor-like kinases, suggesting an additional sRNA control mechanism may be active in barley leaves as opposed to predominant R-gene phasiRNA overlap in many eudicots. In addition, we identified conserved miRNAs, novel miRNA candidates, and barley genome mapped sRNAs that have PARE validated transcript targets in barley. The miRNA target transcripts are enriched in transcription factors, signaling-related proteins, and photosynthesis-related proteins. Together these results suggest both barley and Bgh control metabolism and infection-related responses via the specific accumulation and targeting of genes via sRNAs.

Keywords: Barley; Blumeria; CSEPs; EKA family; Pathogen effectors; Small RNA-Seq; Transposable elements.

Fig. 1

Small RNA sequencing and PARE sequencing analysis pipelines. (A) Small RNA-Seq Illumina reads were trimmed, filtered, and run through the two plant miRNA identification programs miRDeep-P and ShortStack to identify miRNA/milRNA candidates and DE reads. (B) Sequencing reads from the PARE libraries were trimmed and filtered and analyzed with the sPARTA version 1.21 [31] and CleaveLand (version 4.4) [32]. Additional input data was provided from the barley or Blumeria transcriptome and miRNA/milRNA candidates plus DE reads developed from the sRNA sequencing pipeline. Input or output files are highlighted with blue boxes, programs or processes are highlighted with green ovals, and PARE program inputs are highlighted with red arrows

Fig. 2

Bgh_Cluster_643 structure and encoded PARE-validated milRNAs. a Linear representation of Bgh_Cluster_643 with milRNA encoding regions for 643–1 to 643–7 highlighted. b RNAfold predicted Bgh_Cluster_643 structure with sRNA mapping density scale from blue (no coverage) to purple (> = 10⁴ mapping reads) outputted from the ShortStack [34]. c Details of Bgh_Cluster_643 predicted milRNAs including name, location on Bgh_Cluster_643, predicted transcript target annotation, and number of mismatches/gaps in transcript alignment. Note that in Additional file 4: Table 2, Column “A”; lines 195–206 show the original designations from the ShortStack program, while simplified names used here are shown in parentheses. d Alignments of predicted milRNA to their transcript targets / cleavage sites with adjusted p values (detailed in Additional file 4: Table 2). Cleavage sites are represented by red arrows

Fig. 3

Bgh genome supercontig HF944340 encodes both a predicted natural antisense siRNA (natsiRNA) transcript as well as a member of the EKA effector gene family. The Bgh_Cluster_643 natsiRNA transcript is processed into several milRNAs candidates including Bgh_Cluster_643–6. The EKA transcript (BGHDH14_bgh06737) is encoded antiparallel to the hairpin and is transcribed and targeted for transcript cleavage by Bgh_Cluster_643–2

Fig. 4

Transcript and sRNA sequencing reads mapped to Bgh genome positions near BGHDH14_bgh06737 and BGHDH14_bgh00862. The gene transcript models are highlighted with the blue lines, while the transcript and sRNA reads for each gene are highlighted with the red boxes. a Transcript based RNA-Seq reads mapped to the Bgh genome. b sRNA based RNA-Seq reads mapped to the Bgh genome

Fig. 5

Genotype membership distribution for genotype-specific phasiRNA loci. CI 16151 is designated by purple, mla6 by pink, rar3 by orange, bln1 by green, and mla6 + bln1 by yellow

Fig. 6

PhasiRNA locus phasing score and mapping position relative to barley gene HORVU3Hr1G105020, a NLR gene with homology to wheat CNL9. a Phasing score diagram on chromosome 3 from 667589499 to 667589696. b Gene model section of HORVU3Hr1G105020 overlapped by phasiRNA loci. c sRNA data from panel mapped to barley genome. Maximum sRNA mapping depth of 33 reads at peak highlighted with a * d PARE library data from panel mapped to barley genome. The phasiRNA seed region is highlighted with the red boxes. Maximum PARE read mapping depth of 49 reads at peak highlighted with #