Selection of Reference Genes for Normalizing Quantitative Real-Time PCR Gene Expression Data with Multiple Variables in Coffea spp.

Selection and validation of appropriate reference genes should be the first step to consider in experiments based on quantitative real-time polymerase chain reaction (qRT-PCR). In this study, ten candidate genes were investigated for their stability as suitable reference genes in qRT-PCR data normalization using a diverse set of 12 Coffea cDNAs from plants from three different species/genotypes exposed to single or multiple abiotic stresses (drought and chilling, alone or in combination). Primer amplification efficiencies were calculated for all of the selected genes and varied according to each individual genotype. The expression of each gene was measured by qRT-PCR to evaluate its stability. A multiple analytical approach was followed, based on consensus merged data from four different complementary statistics, namely geNorm, BestKeeper, NormFinder, and coefficient of variation, which produced comparable but not identical results. According to this approach, the most suitable sets of reference genes for data normalization in the five experimental datasets are (1) total assay: GAPHD, Cycl, and UBQ10; (2) genotype: GAPDH, UBQ10, Ap47, and EF-1A; (3) cold stress: UBQ10, GAPDH, ACT, and EF-1A; (4) drought stress: GAPDH, ACT, EF1A, and Apt; and (5) multiple stress: UBQ10, GAPDH, ACT, and elf-4A. Normalization of gene expression using these selected genes was validated by examination of the expression of the photosynthetic-related ApoA2 gene in samples from non-stressed and stressed plants. Our results are useful to assist studies on Coffea physiology with the aim of breeding for increased tolerance to abiotic stress conditions.