Brain Transcriptomic Response to Social Eavesdropping in Zebrafish (Danio rerio)

Abstract

Public information is widely available at low cost to animals living in social groups. For instance, bystanders may eavesdrop on signaling interactions between conspecifics and use it to adapt their subsequent behavior towards the observed individuals. This social eavesdropping ability is expected to require specialized mechanisms such as social attention, which selects social information available for learning. To begin exploring the genetic basis of social eavesdropping, we used a previously established attention paradigm in the lab to study the brain gene expression profile of male zebrafish (Danio rerio) in relation to the attention they paid towards conspecifics involved or not involved in agonistic interactions. Microarray gene chips were used to characterize their brain transcriptomes based on differential expression of single genes and gene sets. These analyses were complemented by promoter region-based techniques. Using data from both approaches, we further drafted protein interaction networks. Our results suggest that attentiveness towards conspecifics, whether interacting or not, activates pathways linked to neuronal plasticity and memory formation. The network analyses suggested that fos and jun are key players on this response, and that npas4a, nr4a1 and egr4 may also play an important role. Furthermore, specifically observing fighting interactions further triggered pathways associated to a change in the alertness status (dnajb5) and to other genes related to memory formation (btg2, npas4b), which suggests that the acquisition of eavesdropped information about social relationships activates specific processes on top of those already activated just by observing conspecifics.

Fig 1. Behavioral paradigm and selected behavioral groups for transcriptomic analysis.

(A) Schematics of experimental treatments bystanders to interacting conspecifics (magenta), bystanders to non-interacting conspecifics (lime); and isolated fish (blue). ROI in light grey. (B) Schematic of the subject fish’s mean resultant directional vector composed by the vector’s length R, mean angle α (0° opposite and 180° directed towards the stimulus) and R projected onto 180° (Rproj). (C) Clustering analysis of all focal fish from the experimental treatments using time spent in ROI and Rproj. Cluster A–strongly attentive profile; cluster B–weakly attentive profile. (D) Linear histograms and 2D heatmaps of time spent in each position of the arena (left), and polar directional histograms (right) of one individual per behavioral group (for complete set of samples, see S1 Fig). Heatmaps are scaled from maximum relative value (red) to minimum relative value (dark blue). Linear and polar histograms represented in arbitrary scale. (E) Scatter plots of selected fish from the four behavioral groups [bystanders to interacting conspecifics (BIC, magenta), bystanders attentive to non-interacting conspecifics (BANIC, green), bystanders inattentive to non-interacting conspecifcs (BINIC, lime), and isolated fish (ISOL, blue) for time in ROI (left), Rproj (center) and mean resultant directional vectors (right).

Fig 2. Changes in gene expression in the brain of bystander zebrafish.

(A) Venn diagram showing DE genes between behavioral groups (BIC, bystanders to interacting conspecifics; BANIC, bystanders attentive to non-interacting conspecifics; and BINIC, bystanders inattentive to non-interacting conspecifics), and the reference group (ISOL, isolated fish) (up-regulated = ▴; down-regulated = ▾). (B) Hierarchical clustering of the individuals from each selected behavior group (columns) and of DE genes (lines). Heatmap represents normalized gene expression levels (blue = low expression, yellow = high-expression).

Fig 3. Transcription factor motifs enriched in differentially expressed genes for behavioral groups.

(A) Single motifs enriched in at least one of the behavioral groups (BIC, bystanders to interacting conspecifics; BANIC, bystanders attentive to non-interacting conspecifics; or BINIC, bystanders inattentive to non-interacting conspecifics). (B) Pairs of motifs involving GATA2 enriched in BIC and/or BANIC. (C) Pairs of motifs involving TAL1::GATA1 enriched in BIC and/or BANIC. Associations found in each behavior group can be strongest with genes up-regulated (orange) or down-regulated (purple). Grey cells indicate no significance of associations to any group of differentially expressed genes. Significance was calculated using uncorrected (P) and corrected (FDR) P-values.

Fig 4. Transcription networks of the different behavioral groups.

Networks consisting of DE genes and enriched transcription factors TF for the behavioral groups: (A) bystanders to interacting conspecifics (BIC); (B) bystanders attentive to non-interacting conspecifics (BANIC); and (C) bystanders inattentive to non-interacting conspecifics (BINIC). The thickness of the edges correspond to the confidence score of the genes’ association, yellow nodes indicate up-regulated DE genes, blue nodes indicate down-regulated DE genes, orange nodes indicates TF motifs mainly associated with up-regulated DE genes, and purple indicates TF motifs mainly associated with down-regulated DE genes.