Two Pairs of Neurons in the Central Brain Control Drosophila Innate Light Preference

See allHide authors and affiliations

Science  22 Oct 2010:
Vol. 330, Issue 6003, pp. 499-502
DOI: 10.1126/science.1195993


Appropriate preferences for light or dark conditions can be crucial for an animal’s survival. Innate light preferences are not static in some animals, including the fruit fly Drosophila melanogaster, which prefers darkness in the feeding larval stage but prefers light in adulthood. To elucidate the neural circuit underlying light preference, we examined the neurons involved in larval phototactic behavior by regulating neuronal functions. Modulating activity of two pairs of isomorphic neurons in the central brain switched the larval light preference between photophobic and photophilic. These neurons were found to be immediately downstream of pdf-expressing lateral neurons, which are innervated by larval photoreceptors. Our results revealed a neural mechanism that could enable the adjustment of animals’ response strategies to environmental stimuli according to biological needs.

Preference between light and darkness plays an important role in animal life (14). The fruit fly Drosophila melanogaster avoids light at the first to mid-third instar larval stage, but this photophobic behavior is thereafter reduced, before pupation (57). In addition to the circadian photoreceptor cryptochrome (CRY), the larval visual system includes two bilateral groups of 12 photoreceptors (8, 9): the Bolwig’s organs (BO), which send out Bolwig’s nerves (BNs) to innervate the pace-making neurons, the pigment-dispersing factor (Pdf)–expressing lateral neurons (pdf neurons) in larval central brain (6, 10). Blocking either BO or pdf neurons causes larval blindness, as measured by phototactic assay (6, 11, 12).

To investigate downstream neurons underlying larval phototactic behavior, we screened a batch of up to 800 Gal4 lines [obtained from Drosophila Genetic Resource Center (DGRC), Kyoto] in a simple light-dark choice assay (6) using the Gal4/UAS system to drive ectopic expression of the tetanus toxin light chain (TeTxLC; UAS-TNTG), a neuron-specific toxin that prevents presynaptic release of synaptic vesicles (13). Whereas most Gal4 lines manifested photophobia and several Gal4 lines exhibited loss of light preference with ectopic TeTxLC expression at early to mid-third instar larval stage, one Gal4 line, NP394-Gal4, demonstrated a preference for light. This line showed positive larval phototaxis when TeTxLC expression was driven by NP394-Gal4 [Fig. 1A, performance index (PI) = −0.35 ± 0.07, P < 0.001, n = 16; fig. S1]. Furthermore, temporary TeTxLC expression in NP394-Gal4–labeled neurons was able to confer positive phototaxis at various larval stages (fig. S2). The positive larval phototaxis was reproduced by ectopic expression of a mutated form of the open rectifier potassium channel, dORKΔC (Fig. 1A, PI = −0.25 ± 0.09, P < 0.05, n = 16), the overexpression of which hyperpolarizes neurons and subsequently inactivates neuronal function (14).

Fig. 1

Positive larval phototaxis was induced by inhibition of neurons labeled by three Gal4 lines. (A) Expression of TeTxLC (UAS-TNTG) and dORKΔC with the Gal4/UAS system in the three Gal4 lines led to positive larval phototaxis, whereas the Gal4 and UAS lines alone resulted in negative phototaxis. (B) Expression of shits with the Gal4/UAS system in the three Gal4 lines led to positive larval phototaxis at restrictive temperature. Early to mid-third instar larvae were used. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001; n = 16 for all groups. (C to H) Expression patterns of NP394-Gal4 (C and F), NP423-Gal4 (D and G), and NP867-Gal4 (E and H) in third instar larval CNS. (I and J) Double Gal4 expression patterns of NP394-Gal4 + NP423-Gal4 (I) and NP394-Gal4+NP867-Gal4 (J) in third instar larval CNS. (K) Single-neuron morphology of the NP394-neurons in NP423-Gal4. Background is FasII-labeled neural tracts (magenta). mCD8-GFP (green) was used to label the neuron. Arrows indicate the soma of the NP394-neurons in (C) to (J) and two arborization regions in (K). Scale bars, 20 μm.

To find more phototaxis-positive Gal4 lines sharing common labeling with the NP394-Gal4, we rescreened all the Gal4 lines by overexpressing dORKΔC. This method was chosen because the overexpression of TeTxLC led to lethality or defects in locomotion in a large number of Gal4 lines, meaning that behavioral assays could not be conducted. Two lines, NP423-Gal4 and NP867-Gal4, manifested positive larval phototaxis when the labeled neurons were inhibited by dORKΔC (Fig. 1A, PI = −0.22 ± 0.06 for NP423-Gal4 > dORKΔC and PI = −0.34 ± 0.07 for NP867-Gal4 > dORKΔC, n = 16 for both lines). For further confirmation at the behavioral level, we applied the temperature-sensitive form of Dynamin (shits) that instantly inhibits cell endocytosis at the restrictive temperature (15). In all three Gal4 lines, ectopic expression of shits at the restrictive temperature resulted in positive larval phototaxis (Fig. 1B, PI = −0.22 ± 0.05 for NP394-Gal4 > UAS-shits, PI = −0.22 ± 0.05 for NP423-Gal4 > UAS-shits, and PI = −0.29 ± 0.05 for NP867-Gal4 > UAS-shits, n = 16 for all lines). By contrast, hyperactivation of NP394-Gal4–labeled neurons by expressing the sodium channel NaChBac (16) significantly enhanced light avoidance in late third instar larvae, which generally exhibit reduced light avoidance compared with younger larvae (figs. S3 and S4). Thus, we concluded that regulation of activity in neurons labeled by these Gal4 lines could switch larval phototaxis between negative and positive, suggesting that these neurons mediate the larval preference between light and darkness.

In parallel with behavioral screening, we used a membrane-tethered green fluorescent protein (mCD8-GFP) to visualize the expression patterns of the Gal4 lines at the larval stage. Outside the central nervous system (CNS), common labeling was found only in the salivary gland in all three Gal4 lines (table S1). In the larval CNS, the NP394-Gal4 expression pattern was most restricted of the three Gal4 lines (Fig. 1, C to H, and fig. S5). NP394-Gal4 labeling was most marked in two pairs of mirror-symmetrically arranged neurons in the supraesophageal ganglion from as early as the first instar and throughout larval development (fig. S6). For convenience, we refer to these cells as NP394-neurons. In the other two Gal4 lines, labeling of neurons with morphology and location similar to those of the NP394-neurons was also observed (Fig. 1, C to H). To confirm that the same NP394-neurons were labeled in all three Gal4 lines, we conducted combinatorial Gal4 labeling. In flies carrying combinations of two Gal4 lines such as NP394-Gal4 and NP423-Gal4, or NP394-Gal4 and NP867-Gal4, the number of NP394-neuron–like neurons visualized was the same as that in the single Gal4 line of NP394-Gal4 (Fig. 1, I and J). Such a lack of increment in the neuron number suggests that the NP394-neurons were labeled by all the Gal4 lines. Thus, the NP394-neurons appear to be the only neurons labeled in common in the CNS in all three Gal4 lines.

We propose that the NP394-neurons are crucial for switching the larval light preference between darkness and light. To verify this, we introduced elav-Gal80, which can specifically repress Gal4 activity in neurons, into NP394-Gal4 flies to prevent Gal4 expression in NP394-neurons (17). In parallel with the repression of Gal4 labeling in NP394-neurons by elav-Gal80 (Fig. 2, A and B), the positive larval phototaxis in the TeTxLC-expressing NP394-Gal4 line became negative, possibly because TeTxLC expression driven by NP394-Gal4 was excluded from the NP394-neurons (Fig. 2C, PI = −0.31 ± 0.07 for NP394-Gal4 > UAS-TNTG, and PI = 0.39 ± 0.09 for NP394-Gal4 + elav-Gal80 > UAS-TNTG, n = 16 for both lines). Thus, inhibition of the activity of the NP394-neurons was able to reverse the larval light preference.

Fig. 2

Masking NP394-neurons from inhibition restored negative larval phototaxis. (A and B) Introduction of elav-Gal80 repressed NP394-Gal4 labeling (green) in NP394-neurons but left labeling in other regions intact. Anti-FasII labeling (magenta) was used as background. Scale bars, 20 μm. (C) Larval phototaxis changed from positive to negative when elav-Gal80 was introduced into NP394-Gal4 larvae expressing TeTxLC. Data are presented as the mean ± SEM. ***P < 0.001; n = 16 for all groups.

We then investigated the morphology of the NP394-neurons. The contralateral projections of NP394-neurons from both sides were largely overlapped, making single-neuron morphology difficult to identify. Hence, the FLP-out technique was applied to investigate the morphology of single NP394-neurons (Fig. 1K and fig. S7). Only one type of single-neuron morphology was observed. The cell body was located in the center of each brain hemisphere. The neuron projected posteriorly for a short distance in most cases, then medially, crossing the midline along the same path as that in the corresponding contralateral NP394-neurons and turning posteriorly in the end. Two major arborization regions were formed: one close to the terminal region of lateral neurons labeled by an antibody against Pdf (18), and the other closer to the midline (figs. S7 and S8). The part of the tract that crossed the midline appeared to contain little arborization.

Because the NP394-neurons appeared to be involved in larval light preference, we speculated that they might somehow receive visual information from the larval visual sensory system. We compared the NP394-Gal4–labeled NP394-neurons with Pdf-specific antibody–labeled pdf neurons. The dendrites of NP394-neurons were close to the axonal termini of pdf neurons, although the interface area appeared to be limited (Fig. 3 and fig. S8). Consequently, it seems that the NP394-neurons are postsynaptic to pdf neurons. Thus, the polarity of NP394-neurons was examined. We expressed both the presynaptic-specific markers, such as Syt-GFP (19), and postsynaptic marker Dscam-GFP (20), in NP394-neurons using NP423-Gal4 to determine the pre- and postsynaptic regions of the neurons (fig. S9). The arborization regions were not labeled by the presynaptic-specific marker Syt-GFP, but were strongly labeled by the postsynaptic marker Dscam-GFP. Thus, the arborization region of contact with the pdf neurons was most likely postsynaptic.

Fig. 3

Synaptic contact between NP394-neurons and pdf neurons. (A and B) Patterns of NP394-neurons (green) and pdf neurons (magenta) in larval brain. (A) Patterns in z-axis projection. (B) Patterns in a single scanning layer of 0.3 μm. Inset in (B) shows the overlap between NP394-neurons and pdf neurons in an x-z view. Arrows indicate overlapping regions. (C to G) GRASP signals visualized between pdf neurons and NP394-neurons labeled by NP394-Gal4 (C to E), NP423-Gal4 (F), and NP867-Gal4 (G). GRASP signal in the framed region in (C) is shown at higher magnification in (D) and (E). Insets in (F) and (G) show only the GRASP signals. Arrows indicate the GRASP signal. Scale bars, 20 μm.

To verify the synaptic interaction between the NP394-neurons and the pdf neurons, we used the recently developed GRASP (green fluorescent protein reconstitution across synaptic partners) technique (21, 22). In this technique, the two components of GFP, GFP1-10 and GFP11, are differentially expressed in two neighboring cells, and form fluorescent reconstituted GFP across cellular gaps if the cells are close enough to each other. We expressed GFP1-10 in NP394-neurons and GFP11 in pdf neurons (23) and observed signals of reconstituted GFP in regions where pdf neurons may synapse with NP394-neurons (Fig. 3, C to G, and fig. S10). To further confirm the contact between NP394-neurons and pdf neurons, we examined the morphology of NP394-neurons in the absence of pdf neurons to test whether the presence of pdf neurons is required for the normal morphological development of NP394-neurons. We did not find any substantial morphological change of NP394-neurons in a pdf-DTI line in which pdf neurons in brain hemispheres were missing (fig. S11). Nevertheless, calcium imaging analysis revealed that the NP394-neuron response to light simulation, as indicated by GCAMP3.0 fluorescence (24), was stronger and faster in the absence of pdf neurons than in their presence (Fig. 4, peak ΔF/F: 0.72 ± 0.10 for control and 1.11 ± 0.12 for pdf-DTI; time latency for 20% ΔF/F: 30.19 ± 1.44 s for control and 8.03 ± 0.73 s for pdf-DTI; movies S1 to S3). Thus, our results suggest that functional contact does exist between the NP394-neurons and pdf neurons. Because pdf neurons receive visual information necessary for phototactic behavior, we propose that the NP394-neurons may receive input signals from pdf neurons for the reversion of light preference.

Fig. 4

Ablation of pdf neurons allows stronger and faster response to light stimulation in NP394-neurons. (A) Change in GCAMP3.0 fluorescence (ΔF/F) upon light stimulation. Scale bar, 10 μm. (B) Representative response processes of NP394-neurons in control and pdf-DTI larvae upon light stimulation. (C) Average peak response in control and pdf-DTI larvae. (D) Average time for fluorescence change to achieve 20% (20% ΔF/F) after light stimulation in control and pdf-DTI larvae. Data are presented as the mean ± SEM. *P < 0.05, ***P < 0.001; n = 6 for all groups.

As larvae with inhibited NP394-neurons manifested positive phototaxis whereas those with hyperactivated NP394-neurons showed promoted light avoidance, we propose that NP394-neuron activity is positively correlated with larval light-avoidance ability. Two possible scenarios could be operating in this situation. First, the activity of NP394-neurons itself controls the larval phototaxis by an unknown mechanism. Second, the NP394-neurons activate the pathway that mediates avoidance of light whereas other unidentified neurons activate the pathway that underlies avoidance of darkness, as was shown for the mechanisms underlying odor-taxis in adult Drosophila (25).

Supporting Online Material

Materials and Methods

Figs. S1 to S13

Table S1


Movies S1 to S3

References and Notes

  1. We thank K. Rao for antibody to Pdf; J. Blau, K. Scott, M. Rosbach, T. Lee, C. Feng, Y. Zhong, W. Yi, Y. Jan, and R. Jiao for fly stocks and constructs; and C. Wang and X. Hao for facility setup. We thank the DGRC (Kyoto) and the Bloomington fly stock center for the fly stocks, and Developmental Studies Hybridoma Bank for antibodies. We also thank H. Gong for technical assistance, and J. Fleming and R. Wolf for valuable comments. This work was supported by the National Natural Sciences Foundation of China [grant 30770682 (Z.G.) and grants 30621004 and 30625022 (L.L.)], the ‘973 Program’ [grants 2005CB522804 and 2009CB918702 (L.L.)], and the Knowledge Innovation Program of the Chinese Academy of Sciences [grant KSCX2-YW-R-247 (L.L.)].
View Abstract

Stay Connected to Science

Navigate This Article