IPK Home

Gene expression profiling in epidermis of barley attacked by powdery mildew


Gene expression profiling in epidermis of barley attacked by powdery mildew

Broad host resistance mediated by the mlo gene.


Uwe Zierold1, Uwe Scholz1 and Patrick Schweizer1,2


1 Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK)
   Corrensstraße 3
   06466 Gatersleben

2 corresponding author

Aim of the study:

This array study is aimed at
  • the identification of pathogen-response genes in barley leaf epidermis.
  • the identification of candidate genes for mediators and effectors of mlo-mediated resistance in barley.
  • the localization of transcript accumulation in the epidermis versus inner leaf tissues.

Biological system:

Plants were grown in pots of compost soil (from IPK nursery) in a greenhouse with automatic shading and supplementary light (sodium-halogen lamps) resulting in a light period of 16 h. Temperature ranged from 18°C (night) to 21°C (day). Plant culture in the greenhouse resulted in lower basal levels of PR protein expression, compared to culturing the plants in growth chambers. Seven-day-old seedlings were used for all inoculation experiments. Plants of two near isogenic lines of barley, cv. Ingrid and Ingrid BC mlo5, were challenge inoculated with Blumeria graminis f.sp. hordei (average spore density = 17 conidia mm-2) and the adaxial epidermis was stripped at 6 h, 12 h and 24 h post inoculation. Epidermis from non-inoculated control plants grown in parallel was stripped at the same time points.

The HO cDNA array from inoculated barley epidermis:

mRNA from barley leaf epidermis at 6 h and 24 h post inoculation with Blumeria graminis f.sp. hordei or Blumeria graminis f.sp. tritici was mixed in a 1:1:1:1 ratio and used for the establishment of a cDNA library (library ID = "HO"). Approximately 5000 5' EST sequences gave rise to a set of 3595 unigenes. Of those unigenes, approximately 3100 were successfully amplified by PCR and spotted onto nylon membranes.

Expression profiling:

cDNA probes were produced by second-strand 33P labelling, as described by Sreenivasulu, N., Altschmied, L., Panitz, R., Hahnel, U., Michalek, W., Weschke, W. and Wobus, U. (2002) Identification of genes specifically expressed in maternal and filial tissues of barley caryopses: A cDNA array analysis. Molecular Genetics and Genomics. 266: 758-767.
Spot detection was done by using the ArrayVision software (Imaging Research Inc., St. Catharines, Canada) and signal intensities were normalized after local background subtraction by applying a statistical procedure, as described by Sreenivasulu et al (2002, see above).

Array experiments carried out in this study:

Question Barley genotype Tissue Treatm. Time pointsa Biol. repl. Techn. repl.
Regulation Ingrid Epid. Contr. 6h, 12h, 24h 2 1
Regulation Ingrid Epid. Bgh-inocul. 6h, 12h, 24h 2 1
Regulation Ingrid BC mlo5 Epid. Contr. 6h, 12h, 24h 2 1
Regulation Ingrid BC mlo5 Epid. Bgh-inocul. 6h, 12h, 24h 2 1
Localization Ingrid & Ingrid BC Mlo5b Epid. Bgh-inocul. 6h & 24hb 1 3
Localization Ingrid & Ingrid BC Mlo5b Leaf Bgh-inocul. 6h & 24hb 1 3
aHours post inoculation.
bbSplit-sample experiment for localization of transcripts in peeled epidermis or in the remainder of peeled leaves: RNA of Ingrid and Ingrid BC mlo5, harvested 6 and 24 h post inoculation, was pooled in an 1:1:1:1 ratio per tissue.

The data from the localization experiment were imported into the table shown here (column "E/M" = epidermis to mesophyll factor).

Remarks to the data table:

  1. Int/Bckg = ratio of background-subtracted signal to local background.
  2. Norm. Int. = Signal intensity after local background subtraction and array normalization.
  3. Gene centered = median values of signal intensities per gene were transformed to zero after log transformation.
  4. Inocul./Control = Ratio of signal intensities of inoculated samples versus corresponding control samples.
  5. UP and DOWN criterium fulfilled: Firstly, a unigene had to be at least twofold induced or repressed in both biological replicates at one or more time points. Secondly, the mean difference of signal intensities between inoculated and corresponding control samples and the 95% confidence interval was calculated per host genotype (n=6). Genes with a positive mean signal difference and positive lower as well as upper limit of the 95% confidence interval were identified as significantly up-regulated. Genes with a negative mean value and negative lower as well as upper limit of the 95% confidence interval were identified as significantly down-regulated.
  6.    formula
    This is a measure for differences in transcript abundance between the near-isogenic barley lines. DI > 0 indicates genes with a higher transcript abundance in Ingrid BC mlo5, compared to cv Ingrid. The oposite applies to DI < 0.
  7. E/M = Epidermis versus Mesophyll Factor for localization of the corresponding transcript in inoculated barley leaves (pooled mRNA from 6 h and 24 h post inoculation and from both near-isogenic lines).
  8. Differentially regulated candidate genes were classified by functional categories, based on Blast X results or, where no Blast X result was available, by Blast N (not shown) results.
  9. A total of 351 differentially regulated candidate genes is shown. In the publication by Zierold et al (submitted) only 315 candidate genes are discussed. The additional 36 candidate genes shown here correspond to genes represented by redundant cDNA clones on the array, due to a re-arraying problem. However, only the final set of 315 candidate genes was functionally categorized, thus allowing easy identification in the table.

Data are based on two biological replicates, i.e. two independent inoculation experiments on different weeks. Only genes with a signal/background ratio of >2.5 on at least 2 (out of the 24) membranes were analysed and are displayed here.