The serotonergic central nervous system of the Drosophila larva: anatomy and behavioral function

Abstract

The Drosophila larva has turned into a particularly simple model system for studying the neuronal basis of innate behaviors and higher brain functions. Neuronal networks involved in olfaction, gustation, vision and learning and memory have been described during the last decade, often up to the single-cell level. Thus, most of these sensory networks are substantially defined, from the sensory level up to third-order neurons. This is especially true for the olfactory system of the larva. Given the wealth of genetic tools in Drosophila it is now possible to address the question how modulatory systems interfere with sensory systems and affect learning and memory. Here we focus on the serotonergic system that was shown to be involved in mammalian and insect sensory perception as well as learning and memory. Larval studies suggested that the serotonergic system is involved in the modulation of olfaction, feeding, vision and heart rate regulation. In a dual anatomical and behavioral approach we describe the basic anatomy of the larval serotonergic system, down to the single-cell level. In parallel, by expressing apoptosis-inducing genes during embryonic and larval development, we ablate most of the serotonergic neurons within the larval central nervous system. When testing these animals for naïve odor, sugar, salt and light perception, no profound phenotype was detectable; even appetitive and aversive learning was normal. Our results provide the first comprehensive description of the neuronal network of the larval serotonergic system. Moreover, they suggest that serotonin per se is not necessary for any of the behaviors tested. However, our data do not exclude that this system may modulate or fine-tune a wide set of behaviors, similar to its reported function in other insect species or in mammals. Based on our observations and the availability of a wide variety of genetic tools, this issue can now be addressed.

Figure 1. Anatomy of the Serotonergic System in the Larval CNS Based on anti-5HT Staining.

5HT positive cells (green) of Canton-S wild type larvae are shown in combination with anti-FasciclinII (FasII)/anti-Cholineacetyltransferase (ChAT) neuropil markers (magenta) (A and D–G). (A) The CNS of the third instar larva comprises 19 different 5HT-positive bisymmetrical clusters of one to three cells each. (B–G) In the brain hemispheres, five serotonergic clusters, SP1, SP2, LP1, SE0 and IP, were detected (in B and C only the anti-5HT channel is shown). (D) 5HT cells innervate the antennal lobe (AL; right arrow) and the suboesophageal ganglion (SOG; left arrowhead). (E) The mushroom body lobes (MB; arrow) and the (E) MB calyx (arrow) show only very week – if any - innervation. (F) By contrast, the larval optic neuropil (LON; arrow) is innervated by serotonergic arborizations. (B) and (C) show a frontal view of the anterior or posterior half of the brain, respectively. In (D–G) lateral is always to the right and medial to the left. Scale bars: A–C: 50 µm; D–G: 25 µm.

Figure 2. Expression Pattern of the Driver Line TRH-GAL4 in the Larval CNS.

Triple staining of TRH-GAL4/UAS-mCD8::GFP third instar larvae in the first column shows cell membrane-bound CD8 labeling (green) combined with 5HT-immunoactivity (red) and anti-FasII/anti-ChAT staining for visualizing the neuropil (blue). The second (CD8), third (5HT) and fourth columns illustrate the three channels separately. The first row (A–A’’’) shows the whole CNS. The other rows represent higher magnifications of the brain in frontal view (B–B’’: posterior; C–C’’: anterior) and the ventral nerve cord (VNC) (D–D’’). A high co-localization of CD8- and 5HT-positive cells is found in the posterior hemisphere clusters SP1, SP2 and LP1 (B–B’’) as well as in the anterior clusters IP and SE0-3 (C–C’’). Nearly all cells of the VNC clusters T1-3 and A1–A8/A9 (D–D’’) show anti-CD8 and anti-5HT double staining. In addition some non-serotonergic CD8-expressing cells were detected (asterisks). Scale bars: 50 µm.

Figure 3. Expression Pattern of the Driver Line TPH-GAL4 in the Larval CNS.

First column: CNS of TPH-GAL4/UAS-mCD8::GFP third instar larvae stained with anti-CD8 (green), anti-5HT (red) and anti-FasII/anti-ChAT (neuropil markers; blue). The second, third and fourth columns represent the three channels separately. The first row (A–A’’’) shows an overview of the CNS. Other rows represent higher magnifications of the brain in frontal view with slightly shifted brain hemispheres (B–B’’: posterior; C–C’’: anterior) and the ventral nerve cord (VNC) (D–D’’). In the posterior (B–B’’) and anterior brain (C–C’’) as well as in the VNC (D–D’’) most cells are both anti-CD8 and anti-5HT positive. Only in some clusters (e.g. SE3 and T1) a few CD8-positive cells do not show 5HT expression. Details are also presented in Table 1. Scale bars: 50 µm.

Figure 4. Post- and Presynaptic Organisation of the Larval Serotonergic System.

By crossing TRH-GAL4 (A,B) and TPH-GAL4 (C,D) with UAS-Dscam17.1::GFP and UAS-n-syb::GFP, postsynaptic and presynaptic regions, respectively, were visualized. The brains of third instar larvae were stained with anti-GFP (green) and with anti-FasII/anti-ChAT (magenta). The expression patterns for the postsynaptic innervation are similar for the two driver lines, the same is true for the presynaptic labeling. Scale bars: 50 µm.

Figure 5. Ablation of the Serotonergic Neurons via UAS-hid,rpr Expression.

UAS-hid,rpr (head involution defective; reaper) was crossed with TRH-GAL4 (A) or TPH-GAL4 (B) and stained with anti-5HT (green) and anti-FasII/anti-ChAT (magenta). Nearly all serotonergic neurons undergo apoptosis. Only a small number of 5HT cells in the VNC, the hemispheres and the SOG were not ablated by hid and reaper expression. A similar expression pattern compared to wild type (Figure 1) was detectable in all control groups, by crossing either the two GAL4-lines (C, D) or the UAS-line (E) with white¹¹¹⁸ control flies. Scale bars: 50 µm.

Figure 6. Morphology of the SP1 Cell.

SP1-1 type 5HT cell as shown in single-cell flp-out clones via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti-ChAT (blue) staining (A). The three channels of the staining are presented individually in panels B–D. Scale bar: 50 µm.

Figure 7. Morphology of the SP2 Cells.

SP2-1 and SP2-2 type 5HT cells shown in single-cell flp-out clones via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti-ChAT (blue) staining (A and E). The three channels are presented individually in panels B–D and G–H. In A and B three cells are labeled by the flp-out technique. Besides the SP2-1 cell (arrow), weak expression was detectable in an additional cell body (arrowhead) and a third cell of the LP cluster (asterisk). The SP2-2 cell was only weakly labeled and therefore likely misses a comprehensive visualization of its entire morphology. Scale bars: 25 µm.

Figure 8. Morphology of the LP1 Cells.

LP1-1 type 5HT cells shown in single cell flp-out clones via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti-ChAT (blue) staining (A and E). The three channels are presented individually in panels B–D and G–H. Two examples for different flp-out clones are shown in A and E. Due to the variation in their morphology it is not possible to clarify, if the two clones label the same cell or two different cells of the LP cluster. Thus, in more restricted manner we categorized both clones as LP1-1. Scale bars 10 µm (in A) and 25 µm (in E).

Figure 9. Morphology of the IP Cells.

IP1-1, IP1-2 and IP1-3 type 5HT cells shown in single-cell or two-cell flp-out clones via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti- ChAT (blue) staining (A, E and I). The three channels are presented individually in panels B–D, G–H and J–L. The IP1-1 cell (B, arrow) is visualized in a double flp-out clone that shows an additional weakly labeled cell body in the right hemisphere (arrowhead). The IP1-2 cell (F) is also visualized in a double flp-out clone together with the SP1-1 cell (see also Figure 6). The arrow marks the cell body of the IP1-2 cell that innervates the ipsi- and contralateral hemispheres by crossing the midline more dorsal (arrow) compared to the SP1-1 cell that crosses the midline next to the pharynx. The expression in the SOG belongs to third cell of a different type that does not innervate the brain hemispheres. Scale bars 25 µm.

Figure 10. Morphology of the SE Cells.

SE type 5HT cells shown in single cell flp-out clones stained via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti- ChAT (blue) (A, E, I). The three channels are presented individually in panels B–D, F–H, J–L. The SE2-1 cell (A) is visualized by a single cell flp-out clone of the larval brain hemispheres; however there is an additional projection from an additional non-5HT descending neuron of the abdominal cluster (arrowhead). Similarly, the SE2-2 cell type bifurcated close to its cell body and sent branches ipsi- and contralaterally. The contralateral branch split again and its extensions covered the contralateral hemineuromere completely from ventral to dorsal. The SE2-3 cell (I-J) is shown as a double flp-out clone that also visualizes the cell body of an additional 5HT cell (arrowhead). The SE2-3 cell is only weakly labeled by anti-5HT (S, arrow). Scale bars 25 µm.

Figure 11. Morphology of the T1 Cells.

5HT cells in the T1 neuromere as shown in single cell flp-out clones via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti- ChAT (blue) staining (A, E, and I). The three channels are presented individually in panels B–D, F–H and J–L. For the T1-3 cell (J) only limited information is presented due to the low quality of the GFP staining of the flp-out clone. Scale bars 25 µm.

Figure 12. Morphology of the A1–A7 Cells.

5HT cells in A1–A7 neuromeres shown in single-cell flp-out clones via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti- ChAT (blue) staining (A, E, I and M). The three channels are presented individually in panels B–D, F–H, J–L and N–P. Similar to Chen and Condron (2008) we were able to characterize to types of 5HT neurons for A1–A7 called type1 and type2. The two types of neurons are representatively depicted for A4-1 (A–D) and A4-2 (E–H). For the 5HT positive neurons innervating the outer neuromeres A1 and A7 there was a trend to restrict their innervation to the anterior (for A1) and posterior (A2) boarders. In B and J there are additional cell bodies labeled on a lower level. Scale bars 25 µm.

Figure 13. Morphology of the A8/A9 Cell.

The single 5HT cell of the A8/A9 neuromere shown in a single cell flp-out clone via anti-CD8 (green), anti-5HT (red) and anti-FasII/anti- ChAT (blue) staining (A). The clone is not presented as a frontal view but rather as a sagittal view. The three channels are presented individually in panels B–D. Scale bar 25 µm.

Figure 14. Serotonergic Neurons of the CNS are not Necessary for Olfactory Chemotaxis towards Amylacetate and Benzaldehyde.

Third instar larvae with almost completely ablated serotonergic neurons were tested for naïve amylacetate (AM) (A, B) and benzaldehyde (BA) (C, D) preferences. TRH-GAL4/UAS-hid,rpr larvae showed preference for AM (p<0.01 compared to zero) (A) and for BA (p<0.01 compared to zero) (C). Compared to the controls UAS-hid,rpr/+ and TRH-GAL4/+, TRH-GAL4/UAS-hid,rpr did not perform significantly different either in AM or in BA preference tests (p>0.05). Similar results were found by testing TPH-GAL4/UAS-hid,rpr larvae. They preferred AM (p<0.01) (B) as well as BA (p<0.001) (D) and showed in both assays no significant difference to any control line (p>0.05). Under each boxplot of the figure for each genotype the sample size is shown; n = 15−20. Asterisks above each boxplot indicate, if the data is significantly different from zero. *<0.05; **<0.01; ***<0.001.

Figure 15. The Role of the Serotonergic System of the CNS for Gustatory Choice Behavior.

Larvae were tested for their gustatory preference to different sugars (A–F) or salt (G–J) at varying concentrations. TRH-GAL4/UAS-hid,rpr larvae showed a strong preference for 0.2M sucrose (p<0.01) (A) and for 0.2M fructose (p<0.001) (C) with no significant difference to any control. Interestingly, TRH-GAL4/UAS-hid-rpr larvae did not prefer 2M fructose, whereas all controls did (p<0.05). In comparison with the control lines, no significant difference was found, except for 0.2M sucrose, where experimental larvae showed a slightly decreased preference compared to UAS-hid,rpr/+ (p<0.05) (B). TPH-GAL4/UAS-hid,rpr animals strongly preferred (p<0.01) 0.2M sucrose (B), 0.2M fructose (D) and 2M fructose (F). (E). Concerning 1.5M and 2.32M sodium chloride, we noticed a strong avoidance for both TRH-GAL4/UAS-hid,rpr ((G) p<0.01, (I) p<0.01) and TPH-GAL4/UAS-hid,rpr ((H) p<0.01, (J) p<0.05) experimental groups. The performance indices of TRH-GAL4/UAS-hid,rpr at 1.5M salt (G) and of TPH-GAL4-UAS-hid,rpr at 2.32M salt (J) were slightly reduced compared to the corresponding GAL4 control lines. Under each boxplot of the figure for each genotype the sample size is shown; n = 13−24. Asterisks above each boxplot indicate, if the data is significantly different from zero. n.s.>0.05; *<0.05; **<0.01; ***<0.001.

Figure 16. The Serotonergic Neurons of the CNS are not Necessary for Phototaxis.

Preference for darkness was tested for TRH-GAL4/UAS-hid,rpr and TPH-GAL4/UAS-hid,rpr larvae as well as for driver and effector line controls. (A) TRH-GAL4/UAS-hid,rpr larvae showed a strong preference for light (p<0.001) and did not show any significant difference to UAS-hid,rpr/+ nor to TRH-GAL4/+. Also, TPH-GAL4/UAS-hid,rpr larvae preferred darkness (p<0.001), whereas TRH-GAL4/UAS-hid,rpr did not show any difference to the controls (A), TPH-GAL4/UAS-hid,rpr had a slightly higher preference compared to UAS-hid,rpr controls (B). Under each boxplot of the figure for each genotype the sample size is shown; n = 15. Asterisks above each boxplot indicate, if the data is significantly different from zero. *<0.05; ***<0.001.

Figure 17. The Serotonergic Neurons of the CNS are not Necessary for Appetitive and Aversive Olfactory Learning.

For testing appetitive olfactory learning, we utilized a two-group, reciprocal training design consisting of two half trials that give rise to a final performance index. Third instar larvae lacking serotonergic neurons preferred an odor that was paired with 2-M fructose (A, B). Using a single odor, non-reciprocal standard assay for aversive odor-shock learning third instar larvae lacking serotonergic neurons avoided the odor paired with pulses of electric shock (C, D). In both learning experiments, TRH-GAL4/UAS-hid,rpr larvae achieved relatively high performance scores (A) (p<0.01) (C) (p<0.001). Similar results were obtained for TPH-GAL4/UAS-hid,rpr larvae, which showed significant sugar learning (p<0.01) and electric shock learning (p<0.01). In none of the learning assays significant differences between experimental and control larvae were found. Under each boxplot of the figure for each genotype the sample size is shown; n = 10−16. Asterisks above each boxplot indicate, if the data is significantly different from zero. **<0.01; ***<0.001.