Selection of reference genes for expression studies with fish myogenic cell cultures

Abstract

Background: Relatively few studies have used cell culture systems to investigate gene expression and the regulation of myogenesis in fish. To produce robust data from quantitative real-time PCR mRNA levels need to be normalised using internal reference genes which have stable expression across all experimental samples. We have investigated the expression of eight candidate genes to identify suitable reference genes for use in primary myogenic cell cultures from Atlantic salmon (Salmo salar L.). The software analysis packages geNorm, Normfinder and Best keeper were used to rank genes according to their stability across 42 samples during the course of myogenic differentiation.

Results: Initial results showed several of the candidate genes exhibited stable expression throughout myogenic culture while Sdha was identified as the least stable gene. Further analysis with geNorm, Normfinder and Bestkeeper identified Ef1alpha, Hprt1, Ppia and RNApolII as stably expressed. Comparison of data normalised with the geometric average obtained from combinations of any three of these genes showed no significant differences, indicating that any combination of these genes is valid.

Conclusion: The geometric average of any three of Hprt1, Ef1alpha, Ppia and RNApolII is suitable for normalisation of gene expression data in primary myogenic cultures from Atlantic salmon.

Figure 1

Growth and differentiation of myogenic cells extracted from Salmo salar fast myotomal muscle. Growth is shown at 2 d (a-d), 5 d (e-h), 8 d (i-l), 11 d (m-p) and 14 d (q-t) after cell extraction. Myogenic cells were identified by positive desmin staining (a,e,i,m,q). Actin counterstained with phalloidin (b,f,j,n,r) and nuclei stained with sytox green (c,g,k,o,s) also confirmed the presence of multinucleated myotubes shown in the overlay (d,h,l,p,t). Scale bars represent 50 μm.

Figure 2

Expression values for all genes at all time points from Atlantic salmon myogenic cell culture. The raw quantification cycle (Cq) values (n = 42) are represented by box and whisker diagram (box represents quartiles). The mean value is indicated by the dashed line and the 5^th and 95^th percentiles are indicated by the dots above and below each plot.

Figure 3

Individual expression profiles for each candidate reference gene at each day of culture. Values shown are raw Cq values represented as mean ± SD (n = 6).

Figure 4

Stability indices calculated with geNorm (A) and Normfinder (B). Stability indices are shown for all time points (1), developing myotubes at days 2–11 (2) and in established myotubes at days 11–20 (3). Stability of gene expression is inversely proportional to the stability index, so least stable genes are to the left and the most stable to the right for each graph.

Figure 5

Normalisation of desmin mRNA expression to various combinations of the geometric average for two genes. Data shown are all calculated relative to day 2, so that day 2 values are equal to 1 arbitrary unit (A.U.). Six normalisation factors were derived by calculating the geometric averages of the following gene combinations: A: Hprt1, Ppia (closed circle); B: RNApolII, Hprt1 (open triangle); C: EF1α, Hprt1 (closed square); D: EF1α, Ppia (open diamond); E: RNApolII, EF1α(closed triangle); F: RNApolII, Ppia (open circle). Values shown are the mean normalised value ± S.E. (n = 6).

Figure 6

Normalisation of desmin mRNA expression to various combinations of the geometric average of three genes. Data shown are all calculated relative to day 2, so that day 2 values are equal to 1 arbitrary unit (A.U.). Four normalisation factors were derived by calculating the geometric averages of the following gene combinations: A: EF1α, RNApolII, Hprt1 (closed circle); B: EF1α, Ppia, Hprt1 (open circle); C: EF1α, RNApolII, Ppia (closed triangle); D: Ppia, Hprt1, RNApolII (open triangle). Values shown are the mean normalised value ± S.E. (n = 6).