Selection of reference genes for real-time expression studies in Streptococcus agalactiae

Streptococcus agalactiae, group B streptococci (GBS) is the leading cause of severe bacterial infections in newborns. GBS expression studies allowed the identification and characterization of virulence factors and a better understanding of the host-pathogen-environment interactions. The measurement of transcript levels by quantitative real-time PCR (qRT-PCR) is a widely used technique in GBS; however, a systematic evaluation and validation of reference gene stability for normalization purposes in GBS expression studies is currently lacking. Therefore, we analyzed the stability of 10 candidate reference genes (16SrRNA, glcK, glnA, groEL, gyrA, recA, rpoB, rpsL, sdhA and tkt) in three GBS prototype strains (O90R, NEM316 and 2603V/R) grown at different temperature conditions (37°C and 40°C). Our approach was based on the calibration of transcript levels from each gene against the number of bacteria from the same sample (ratio messenger RNA/genomic DNA). As a complementary analysis, reference gene stability was also investigated through the bioinformatic applications, geNorm and NormFinder. Considering the whole GBS development cycle, only a minority of genes were stable under both growth conditions, but this number increased when restricting the analysis to the logarithmic time-points. The range of stable genes was higher at 37°C, where recA and sdhA were stable simultaneously for the three strains, and six out of 10 genes were stable for at least two strains. At 40°C, recA showed up again as one of the best options, suggesting its potential use as reference gene in future qRT-PCR studies. The results generated with geNorm and NormFinder were consistent with those obtained experimentally and evidenced minor variations either among strains or temperature conditions. In conclusion, the fluctuation of expression of reference genes observed among different GBS strains and growth conditions highlights the importance of carefully validating, for each experimental scenario, the use of reference genes for qRT-PCR normalization purposes. Nevertheless, recA seems to be a good candidate for such optimizations.