Hepatic reference gene selection in adult and juvenile female Atlantic salmon at normal and elevated temperatures

Abstract

Background: The use of quantitative real-time polymerase chain reaction (qPCR) has become widespread due to its specificity, sensitivity and apparent ease of use. However, experimental error can be introduced at many stages during sample processing and analysis, and for this reason qPCR data are often normalised to an internal reference gene. The present study used three freely available algorithms (GeNorm, NormFinder and BestKeeper) to assess the stability of hepatically expressed candidate reference genes (Hprt1, Tbp, Ef1α and β-tubulin) in two experiments. In the first, female Atlantic salmon (Salmo salar) broodstock of different ages were reared at either 14 or 22 °C for an entire reproductive season, therefore a reference gene that does not respond to thermal challenge or reproductive condition was sought. In the second, estrogen treated juvenile salmon were maintained at the same temperatures for 14 days and a reference gene that does not respond to temperature or estrogen was required. Additionally, we performed independent statistic analysis to validate the outputs obtained from the program based analysis.

Results: Based on the independent statistical analysis performed the stability of the genes tested was Tbp > Ef1α > Hprt1 > β-tubulin for the temperature/reproductive development experiment and Ef1α > Hprt1 > Tbp for the estrogen administration experiment (β-tubulin was not analysed). Results from the algorithms tested were quite ambiguous for both experiments; however all programs consistently identified the least stable candidate gene. BestKeeper provided rankings that were consistent with the independent analysis for both experiments. When an inappropriate candidate reference gene was used to normalise the expression of a hepatically expressed target gene, the ability to detect treatment-dependent changes in target gene expression was lost for multiple groups in both experiments.

Conclusions: We have highlighted the need to independently validate the results of reference gene selection programs. In addition, we have provided a reference point for those wishing to study the effects of thermal challenge and/or hormonal treatment on gene stability in Atlantic salmon and other teleost species.

Figure 1

Abundance of hepatic mRNA transcripts for Tbp, Hprt1 and Ef1α and β-tubulin in female maiden Atlantic salmon reared during the reproductive season. The average quantification cycle (C_q) of all groups analysed, and the standard error is shown at the top left of the graph. Bars show the average C_q (+SEM) for each group of fish (n = 7). Different superscripts denote statistical significance between groups at a given sample point at p ≤ 0.05.

Figure 2

Vitellogenin gene expression normalised to either Tbp or β-tubulin in female maiden Atlantic salmon reared at either 14 or 22°C and repeat salmon reared during the reproductive season. The mean (+ SEM) gene expression levels for each group of fish (n = 7) are displayed. The asterisk is placed above sample points where different target gene expression results were obtained after normalization.

Figure 3

Abundance of hepatic mRNA transcripts for Tbp, Hprt1 and Ef1α in juvenile Atlantic salmon given a blank silastic implant. The average quantification cycle (C_q) of all groups analysed, and the standard error is shown at the top left of the graph. Bars show the average C_q (+ SEM) for each group of fish (n = 7). Different superscripts denote statistical significance between groups at a given sample point at p ≤ 0.05.