Transcriptomic Signature of Spatial Navigation in Brains of Desert Ants

Abstract

Navigation is crucial for central-place foragers to locate food and return to the nest. Cataglyphis ants are renowned for their advanced navigation abilities, relying on landmark cues and path integration. This study aims to uncover the transcriptomic basis of exceptional spatial learning in the central nervous system of Cataglyphis niger. Ants navigated a maze with a food reward, and we examined expression changes linked to correct decisions in subsequent runs. Correct decisions correlated with expression changes in the optic lobes, but not the central brain, showing a downregulation of genes associated with sucrose response and Creb3l1. The latter gene is homologous to Drosophila crebA, which is essential for long-term memory formation. To understand how ants use distance information during path integration, we analyzed expression shifts associated with the last distance traveled. We uncovered a transcriptomic footprint in the central brain, but not in the optic lobes, with genes enriched for energy consumption and neurological functions, including neuronal projection development, synaptic target inhibition, and recognition processes. This suggests that transcriptional activity in the central brain is necessary for estimating distance traveled, which is crucial for path integration. Our study supports the distinct roles of different brain parts for navigation in Cataglyphis ants.

Keywords: gene expression; learning; maze; path integration; step integration.

FIGURE 1

Worker of Cataglyphis niger in the Tel Baruch sand dunes near Tel Aviv. Credit to Adi Bar.

FIGURE 2

Maze and ant behavior. (A) Maze showing the entrance to the nest (E) from the acclimatization chamber and the chamber with reward (R). (B) The proportion of correct turns (correctness) to reach the reward (n = 23 individuals). (C) Correlation matrix of behavioral variables: Last distance and speed last run are the distance and average speed of an ants' last run; time improvement is the time difference reaching the reward between the last and the first run, proportional to the total time in the runs; distance improvement (L _dist) and correctness improvement (L _prop) refer to the change in the proportion of distance and correct turns over time, respectively; correctness last runs is the mean proportion of correct turns made in the second and third runs; mean time drinking is the average drinking time per run. Only statistically significant (p < 0.05) Pearson's correlations are shown.

FIGURE 3

Central brain expression for the last distance traveled. (A) PCA of the variance transformed transcriptomic reads of central brain expression of 19 Cataglyphis niger workers. In total, 51% of the variance in expression profiles are represented by the first two axis of the PCA. The gray‐blue corresponds to the scaled and centered traveled distance (cm) in the maze in the last run before sacrifice. In black distances superior to 1500 cm are depicted. (B) Heat map of the normalized expression of DEGs in the central brain. Z‐scores are computed on a gene‐by‐gene basis by subtracting the mean and then dividing by the standard deviation after the clustering.

FIGURE 4

Gene expression in the central brain based on the distance traveled in the last run. (A) Log‐fold change in upregulated genes (red) and downregulated genes (blue) in ants walking larger distances. The 10 most significant genes are named. The horizontal dashed line represents the significance threshold of our differential expression analysis p < 0.05. (B) Representative subset of enriched GO terms in the central brain. Revigo visualization and clustering are based on the semantic similarities and hierarchical structure (parent–child terms) of GO terms. One single representative GO term enriched in workers that walked longer distances in the last run is shown for each cluster, with p < 0.05, and represented by more than one gene. The most statistically significant (and nonredundant) 19 upregulated GOterms and, (C) 14 downregulated DEGs GOterms. The bubble size is relative to the number of annotations for the GO Term ID in the underlying EBI GOA database.

FIGURE 5

Optic lobe expression for the correctness test. (A) PCA of the variance transformed transcriptomic reads of the optic lobe's expression of 15 workers. (B) Heat map of the normalized expression of DEGs in the optic lobes. The gray‐blue corresponds to the average of the correct proportion of movements made in the maze in the test(s) run. Z‐scores are computed on a gene‐by‐gene basis by subtracting the mean and then dividing by the standard deviation after the clustering.

FIGURE 6

Optical lobe gene expression in Cataglyphis niger workers. (A) Log‐fold change in upregulated genes (red) and downregulated genes (blue) in ants exhibiting higher correctness in last runs. The 10 most significant genes are named. The horizontal dashed line represents the significance threshold of our differential expression analysis p < 0.05. (B) Revigo semantic clustering of the five most statistically significant (and nonredundant) GO terms with p < 0.05 for downregulated DEGs. The bubble size is relative to the number of annotations for the GO Term ID in the underlying EBI GOA database.

FIGURE A1

Relation between the number of ants that entered the reward chamber before the tracked individual in their first training run (n = 23) and correctness. The correctness is measured as the proportion of correct turns until the reward chamber is reached.