Insights into food preference in hybrid F1 of Siniperca chuatsi (♀) × Siniperca scherzeri (♂) mandarin fish through transcriptome analysis

Abstract

Background: As economically relevant traits, feeding behavior and food preference domestication determine production cost and profitability. Although there are intensive research efforts on feeding behavior and food intake, little is known about food preference. Mandarin fish accept only live prey fish and refuse dead prey fish or artificial diets. Very little is currently known about the genes regulating this unique food preference.

Results: Using transcriptome sequencing and digital gene expression profiling, we identified 1,986 and 4,526 differentially expressed genes in feeders and nonfeeders of dead prey fish, respectively. Up-regulation of Crbp, Rgr and Rdh8, and down-regulation of Gc expression, consistent with greater visual ability in feeders, could promote positive phototaxis. Altered expressions of period, casein kinase and Rev-erbα might reset circadian phase. Down-regulation of orexigenic and up-regulation of anorexigenic genes in feeders were associated with lower appetite. The mRNA levels of Creb, c-fos, C/EBP, zif268, Bdnf and Syt were dramatically decreased in feeders, which might result in significant deficiency in memory retention of its natural food preference (live prey fish). There were roughly 100 times more potential SNPs in feeders than in nonfeeders.

Conclusions: In summary, differential expression in the genes identified shed new light on why mandarin fish only feed on live prey fish, with pathways regulating retinal photosensitivity, circadian rhythm, appetite control, learning and memory involved. We also found dramatic difference in SNP abundance in feeders vs nonfeeders. These differences together might account for the different food preferences. Elucidating the genes regulating the unique food preference (live prey fish) in mandarin fish could lead to a better understanding of mechanisms controlling food preference in animals, including mammals.

Figure 1

Functional annotation of mandarin fish transcripts based on GO categorization. The left y-axis indicates the percentage of a specific category of genes in that main category. The right y-axis indicates the number of genes in a category.

Figure 2

Scatter plot showing the correlation between the expression levels of feeders and nonfeeders by transcriptome sequencing. SC_X and SC_W indicate feeders and nonfeeders, respectively. The x-axis contains Log10 of Reads Per Kb per Million reads (RPKM) of feeders and the y-axis indicates Log10 of RPKM of nonfeeders. Limitations were based on FDR ≤ 0.001, and the absolute value of Log2 (SC_X / SC_W) ≥ 1.

Figure 3

Differential expression analysis of unigenes by DGE. WL and WB indicate liver and brain in nonfeeders, respectively, while XL and XB indicate liver and brain in feeders. The superscripts of each column represent the number of differentially expressed genes between groups.

Figure 4

Differentially expressed genes between feeders and nonfeeders from transcriptome and DGE analysis. The most important pathways related to live prey food preference included retinal photosensitivity (A), circadian rhythm (B), appetite control (C), learning and memory (D). The colors of ellipses were shaded according to significance level (bright red: the mRNA expression levels of feeders were significantly higher than those in nonfeeders (FDR ≤ 0.001, the absolute value of log2[Ratio] ≥ 1); pink: the mRNA expression levels of feeders were slightly higher than those in nonfeeders (FDR ≤ 0.5, the absolute value of log2[Ratio] ≥ 0.5); bright green: the mRNA expression levels of feeders were significantly lower than those in nonfeeders (FDR ≤ 0.001, the absolute value of log2[Ratio] ≥ 1); pale green: the mRNA expression levels of feeders were slightly lower than those in nonfeeders (FDR ≤ 0.5, the absolute value of log2[Ratio] ≥ 0.5).

Figure 5

Validation of differentially expressed genes with Real-time RT-PCR. A. The relative mRNA abundance of retinal photosensitivity genes (Rgr, Rdh8, Crbp and Gc) in feeders and nonfeeders was determined by Real-time RT-PCR. Compared with nonfeeders, the mRNA levels of Rgr and Crbp were significantly increased, whereas Gc was decreased significantly. B. The relative mRNA abundance of clock genes (Per1, Per2, Clock and Bmal1) in feeders and nonfeeders was determined by Real-time RT-PCR. Compared with nonfeeders, the mRNA levels of Per1 and Per2 were decreased significantly and slightly, respectively, whereas Clock and Bmal1 were slightly increased. C. The relative mRNA abundance of appetite control genes (Pomc, Pyy, leptin, Npy and ghrelin) in feeders and nonfeeders was determined by Real-time RT-PCR. Compared with nonfeeders, the mRNA levels of Pomc, Pyy and leptin were significantly increased, whereas Npy and ghrelin were decreased significantly and slightly, respectively. D. The relative mRNA abundance of learning and memory genes (Creb, Ncam, c-fos and Bdnf) in feeders and nonfeeders was determined by Real-time RT-PCR. Compared with nonfeeders, the mRNA levels of Creb and Ncam were slightly decreased, c-fos and Bdnf were decreased significantly. Data are presented as mean ± standard error (n = 6). * indicates significant differences between groups based on one-way analysis of variance (ANOVA) followed by the post hoc test (P < 0.05).