Expression profiling of Chickpea genes differentially regulated during a resistance response to Ascochyta rabiei

Using microarray technology and a set of chickpea (Cicer arietinum L.) unigenes, grasspea (Lathyrus sativus L.) ESTs and lentil (Lens culinaris Med.) resistance gene analogs, the ascochyta blight (Ascochyta rabiei (Pass.) L.) resistance response was studied in four chickpea genotypes, including resistant, moderately resistant, susceptible and wild relative (Cicer echinospermum L.) genotypes. The experimental system minimized environmental effects and was conducted in reference design, where samples from mock-inoculated controls acted as references against post-inoculation samples. Robust data quality was achieved through the use of three biological replicates (including a dye-swap), the inclusion of negative controls, and strict selection criteria for differentially expressed genes including a fold change cutoff determined by self-self hybridizations, Students t test and multiple testing correction (P<0.05). Microarray observations were also validated by quantitative RT-PCR. The time-course expression patterns of 756 microarray features resulted in differential expression of 97 genes in at least one genotype at one time-point. K-means clustering grouped the genes into clusters of similar observations for each genotype, and comparisons between A. rabiei-resistant and susceptible genotypes revealed potential gene 'signatures' predictive of effective A. rabiei resistance. These genes included several pathogenesis-related proteins, SNAKIN2 antimicrobial peptide, proline-rich protein, disease resistance response protein DRRG49-C, environmental stress-inducible protein, leucine-zipper protein, polymorphic antigen membrane protein, as well as several unknown proteins. The potential involvement of these genes and their pathways of induction are discussed. This study represents the first large-scale gene expression profiling in chickpea, and future work will focus on functional validation of the genes of interest. Keywords: time course disease state analysis Overall design: Total RNA was extracted from pooled stem and leaf samples for each genotype (FLIP94-508C, ICC3996, ILWC245 and Lasseter) at each time-point (including control samples) using the RNeasy® Plant Mini Kit (Qiagen, Valencia, CA). The quantity and quality of the total RNA samples were assessed by OD260/OD280 ratios and gel electrophoresis respectively. Fluorescent-labeled targets were prepared and hybridized to array slides according to [Coram, TE. and Pang, ECK. 2006. Expression profiling of Chickpea genes differentially regulated during a resistance response to Ascochyta rabiei. Published in Plant Biotechnology Journal]. All hybridizations were performed with six technical replicates and three biological replicates, incorporating dye-swapping (i.e. reciprocal labelling of Cy3 and Cy5) to eliminate any dye bias. Overall, 360 images were analyzed from 60 slides, resulting in 18 data points for each time-point of each genotype. Slides were scanned at 532 nm (Cy3 green laser) and 660 nm (Cy5 red laser) at 10 µm resolution using an Affymetrix® 428 array scanner (Santa Clara, CA), and captured with the Affymetrix® Jaguar software (v. 2.0, Santa Clara, CA). Image analysis was performed using Imagene 5 (BioDiscovery, Marina Del Rey, CA) software. Quantified spot data was then compiled and transformed using GeneSight 3 (BioDiscovery, Marina Del Rey, CA). Data transformations consisted of a local background correction (mean intensity of background was subtracted from mean signal intensity for each spot), omitting flagged spots, Lowess normalisation of the entire population, creating a Cy5/Cy3 mean signal ratio, taking a shifted log (base 2), and combination of duplicated spot data. To identify differentially expressed (DE) genes, expression ratio results were filtered to eliminate genes whose 95% confidence interval for mean fold change (FC) did not extend to 1.8-fold up or down, followed by Students t test with False Discovery Rate (FDR) multiple testing correction to retain only genes in which expression changes versus untreated control were significant at P < 0.05.