Sensory neuron-derived eph regulates glomerular arbors and modulatory function of a central serotonergic neuron

Abstract

Olfactory sensory neurons connect to the antennal lobe of the fly to create the primary units for processing odor cues, the glomeruli. Unique amongst antennal-lobe neurons is an identified wide-field serotonergic neuron, the contralaterally-projecting, serotonin-immunoreactive deutocerebral neuron (CSDn). The CSDn spreads its termini all over the contralateral antennal lobe, suggesting a diffuse neuromodulatory role. A closer examination, however, reveals a restricted pattern of the CSDn arborization in some glomeruli. We show that sensory neuron-derived Eph interacts with Ephrin in the CSDn, to regulate these arborizations. Behavioural analysis of animals with altered Eph-ephrin signaling and with consequent arborization defects suggests that neuromodulation requires local glomerular-specific patterning of the CSDn termini. Our results show the importance of developmental regulation of terminal arborization of even the diffuse modulatory neurons to allow them to route sensory-inputs according to the behavioural contexts.

Figure 1. Glomerular-specific innervation pattern of the CSDn in the AL is regulated by Ephrin.

(A-A″, E) Innervation pattern of the axonal terminals of the CSDn (green) in glomeruli VA1l/m, VA1d and DA1 (anti-Brp in red) in control adults is shown (n>6). Asterisks indicate glomeruli with fewer innervations and arrowhead indicates glomerulus with more innervations from the CSDn. (B-B″, F) In Ephrin^KG09118 hypomorphs, increased terminal innervations can be seen to VA1l/m (n = 5, p<0.001), DA1 (n = 5, p<0.001) and DL3 (n = 9, p = 0.018) while innervations in VA1d (n = 5, p = 0.865) and V (n = 4, p = 0.149) are comparable to controls. (D-D″, G) Targeted expression of Ephrin in the CSDn in Ephrin^KG09118 hypomorphs restores distribution of axonal terminals in VA1l/m (n = 6, p = 0.99), glomerulus DA1 (n = 6, p = 0.606) and glomerulus DL3 (n = 6, p = 0.992). (C-C″, G) Targeted expression of Ephrin in the CSDn does not change overall distribution pattern of axonal tips in VA1l/m (n = 8, p = 0.241), DA1 (n = 8, p = 0.092233) and DL3 (n = 8, p = 0.910) when compared to controls, however a small decrease in overall branch tip number is observed. (E–G) Quantification of total axonal branch tip number in glomeruli V, VA1l/m, VA1d, DA1 and DL3 is plotted in histograms. A one-way repeated measure ANOVA test was performed to assess significant difference between the genotypes (F = 28.544, P<0.001). All pairwise multiple comparisions were performed using Fisher LSD method.. *, p<0.05; **, p<0.01; ***, p<0.0001; n.s. (not significant), p>0.05. (H–L) Ephrin shows broad expression pattern and it is expressed throughout the developing AL (n>5). APF = After puparium formation. All the images hereafter are oriented as indicated in A′ unless otherwise mentioned. D, dorsal; M, medial. Scale bar = 20 µm. See also Table S1.

Figure 2. Sensory neurons differentially express Eph, which is capable of initiating repulsive interaction with Ephrin.

(A–D) Eph is strongly enriched in three anteriorly positioned glomeruli: VA1l/m, DA1 and DL3 glomeruli in the developing AL starting from 50 hAPF (n>5). White dots encircle the developing AL. Arrow indicates antennal commissure. (F-F′) Targeted expression of EphRNAi in sensory neurons (Pebbled-Gal4/+; UAS EphRNAi/+) results in strong reduction of Eph expression (red) in the antennal lobe compared to (E-E′) controls (UAS EphRNAi/+). AL is counterstained with phalloidin (green). (G–H) Eph expression is reduced in the AL of animals lacking majority of the OSNs from trichoid and basiconic sensilla. (G) Eph (red) is prominently expressed in select few glomeruli in the AL at 70 hAPF of control animals. (H) amos¹/Df(2L)M36F-S6 animals show drastic reduction in the Eph expression in the AL. (I–J) Targeted expression of Ephrin in PNs prevents their entry in high Eph-expressing glomerulus VA1l/m (arrowhead). (I) In control animals (Gal4-GH146,mCD8::GFP/+; Or47b::rCD2/+), PN arbors (green) innervate glomerulus VA1l/m (red). (J) Very few PN arbors innervate VA1l/m glomerulus (red) when Ephrin is overexpressed in PNs (Gal4-GH146,mCD8::GFP/UAS Ephrin; Or47b::rCD2/+). Scale bar = 20 µm.

Figure 3. Olfactory sensory neuron-derived Eph controls glomerular-specific arborization pattern of the CSDn.

(A) The CSDn is labeled in control (RN2flp, tub>STOP>LexA::VP16, lexAOpCD2GFP) animals (B) which shows distinct glomerular-specific arborization pattern. (C) RNAi-mediated knockdown of Eph in sensory neurons (RN2flp, tub>STOP>LexA::VP16, lexAOpCD2GFP; Pebbled-Gal4>UAS EphRNAi) leads to increased CSDn arborization in glomerulus DA1 and VA1l/m. (D) Histogram shows quantification of the axonal branch tip number of the CSDn in different glomeruli (n = 3). (E–H) Loss or transformation of antenna leads to uniform arborization of the CSDn in the AL. (E and G) In control animals, terminal arbor of the CSDn shows glomeruli-specific differences in innervation pattern with some glomeruli receiving fewer inputs (asterisks in E and G). Trans-allelic combination of wg^1-16 and wg^LacZ leads to loss of antenna and (F) Axonal terminals of the CSDn from animals lacking antenna (wg^1-16/wg^LacZ; RN2flp, tub>CD2>Gal4, UASmCD8GFP/+) uniformly innervate the AL. (H) In animals where antenna is transformed into leg (RN2flp, tub>CD2>Gal4, UASmCD8GFP/Antp), axonal terminals of the CSDn innervate the AL homogeneously. Scale bar = 20 µm.

Figure 4. Eph function is not required for appropriate targeting of OSNs and uniglomerular PNs.

(A–B) OSN terminals innervating glomerulus VA1l/m appear comparable to (A) controls in (B) Eph null animals. (C–F) α-sNPF (green) labels specific sets of OSN terminals including (C) DA1 and (E) DL3 in the adult antennal lobe of control animals. α-sNPF immunoreactivity appears comparable to controls in the (D) DA1 and (F) DL3 of Eph null mutants. α-Brp (red) labels the neuropil. (G–H) Targeted expression of Eph in the olfactory sensory neurons does not change their overall pattern and OSNs appear comparable to (G) controls. (I–J) Uniglomerular projection neurons appear normal in Eph null animals. (I) Innervation pattern of projection neurons innervating glomeruli VA1d and DA1 in Mz19mCD8::GFP animals is unchanged in (J) in Eph null mutants (Mz19mCD8::GFP; Eph^X652).

Figure 5. Perturbation in Levels of Eph signaling leads to defective glomeruli-specific positioning of the terminal of the CSDn.

(A-A″, H) Innervation pattern of the axonal terminals of the CSDn (green) in glomeruli VA1l/m, VA1d and DA1 (anti-Brp in red) in control adults is shown (n≥6). (B-B″, H) In Eph null animals, axonal terminals of the CSDn show overall reduction in their AL innervation. This defect is pronounced in glomeruli which normally receive more innervations from the CSDn (VA1d (n = 4, p<0.001), VA1l/m (n = 4, p = 0.127), DA1 (n = 4, p = 0.025), DL3 (n = 4, p = 0.745) and V (n = 4, p<0.001). (C-C″, I) Targeted expression of Eph in the CSDn results in exquisite reversal of the terminal arborization pattern in these glomeruli compared to controls; terminals preferentially target VA1l/m (n = 5, p = 0.002), DA1 (n = 5, p<0.001), DL3 (n = 5, p = 0.003) and avoid glomerulus VA1d (n = 5, p<0.001). (H–I) Quantification of total axonal branch tip number is plotted in a histogram. Asterisks indicate glomeruli with fewer innervations and arrowhead indicates glomerulus with more innervations from the CSDn. A one-way repeated measure ANOVA test was performed to assess significant difference between the genotypes (F = 27.341, P<0.001). All pairwise multiple comparisions were performed using Fisher LSD method. *, p<0.05; **, p<0.01; ***, p<0.0001; n.s. (not significant), p>0.05. Scale bar = 20 µm. (D–G) Glomeruli-specific innervation of axonal terminals is achieved by directed growth of axonal terminals of the CSDn. Terminal arbors of the CSDn in (D–E) control (RN2flp, tub>CD2>Gal4, UASmCD8GFP/+) and (F–G) Eph mutant animals (RN2flp, tub>CD2>Gal4, UASmCD8GFP/+; Eph^X652). Developmental profile of the axonal terminals of control CSDn at (D) 50 hAPF and (E) 70 hAPF is shown. (D) At 50 hAPF, very few axonal terminals of the CSDn can be seen extending to region of the AL where VA1l/m, VA1d, DA1 and DL3 are located. (E) Adult-like pattern of glomeruli-specific innervation of axonal terminals is apparent at 70 hAPF where high innervation of VA1d and low innervation of VA1l/m and DA1 by the CSDn terminals is seen. (F) At 50 hAPF, axonal terminals of the CSDn in Eph null mutants can be seen near the region of AL where the above-mentioned four glomeruli are located but (G) fail to innervate these glomeruli even at 70 hAPF. Asterisks indicate glomeruli with fewer innervations and arrowhead indicates glomerulus with more innervations from the CSDn. Scale bar = 20 µm.

Figure 6. CSDn modulates odour-guided behaviour.

The pair of CSDn is specifically labeled by R60FO2Gal4. Targeted expression of GFP using R60FO2Gal4 (R60F02Gal4/+; UAS mCD8GFP/+) shows a pair of neurons with anatomy characteristic of the CSDn and (Ai–Aii) innervations to the antennal lobe (Brp in red; GFP in green). (Aiii–vi) The neurons labeled by R60F02 co-express 5-HT (red) indicating it is indeed the CSDn (green; Brp in blue). Asterisks indicate cell body of the CSDn. (B–F) The CSDn modulates olfactory response of adult Drosophila towards CO₂. (B) Suppression of evoked synaptic transmission by targeted expression of tetanus toxin light chain (TNTG) in the CSDn (R60F02Gal4/+; UAS TNTG/+, n = 10, p = 0.017) leads to an increase in CO₂ avoidance index compared to control animals (R60F02Gal4/+; UAS TNTVIF/+, n = 12). (C) Similar increase in CO₂ sensitivity is observed upon suppression of CSDn excitability by targeted K_ir2.1 expression (R60F02Gal4/+; UAS K_ir2.1/+, n = 11, p<0.01 compared to controls) in the CSDn. (D) CSDn function is required in the adults for modulating olfactory behaviour. Adult-specific expression of K_ir2.1 in the CSDn is achieved by rearing animals (R60F02Gal4/+; UAS K_ir2.1/+; Tub-Gal80^ts/+) at 18°C throughout development (white bars in D) and then shifting to 29°C after eclosion (black bars in D). Adult-specific suppression of CSDn excitability results in increased CO₂ avoidance (n = 17; p = 0.006). (E) In a reporter line for serotonin receptor 5-HT_1BDro (5-HT_1BDro-Gal4/+; UAS-2xEGFP/+), a group of local interneurons are labeled (red arrows) along with mushroom bodies (yellow arrowheads). (F) RNAi-mediated knock down of 5-HT_1BDro in the 5-HT_1BDro expression domain (5-HT_1BDro-Gal4/+; UAS-5-HT_1BDroRNAi/+, n = 11) results in increased CO₂ sensitivity (p<0.05 compared to all control genotypes, n>7). 5-HT_1BDro expression outside the mushroom bodies, likely in the AL, may be necessary for CO₂ sensitivity as blocking 5-HT_1BDroRNAi expression in mushroom body neurons (MB-Gal80/+; 5-HT_1BDro-Gal4/+; UAS-5-HT_1BDroRNAi/+, n = 14) does not ameliorate increased CO₂ sensitivity (p = 0.12 compared to 5-HT_1BDro-Gal4/+; UAS-5-HT_1BDroRNAi/+, n = 11) and animals exhibit increased CO₂ avoidance (p<0.01 compared to all control genotypes, n>7). Significance was assessed by Mann-Whitney test. *, p<0.05; **, p<0.01; ***, p<0.0001; n.s. (not significant), p>0.05.

Figure 7. Glomerular-specific innervation pattern of the CSDn is relevant for odour-specific modulation of odour-guided behavior.

(A) Targeted expression of tetanus toxin light chain in the CSDn (R60F02Gal4/+; UAS TNTG/+, n = 14, p = 0.028) leads to an increase in relative male courtship index towards cVA-treated virgin females compared to controls (R60F02Gal4/+; UAS TNTVIF/+, n = 16) which implies decreased sensitivity towards cVA in test animals. Now to test the effect of increasing arbors of the CSDn in cVA sensitive glomeruli we tested courtship index of Ephrin^KG09118. (B) Ephrin^KG09118 males are more sensitive to cVA as exhibited by highly reduced courtship towards cVA-treated females (n = 25, p<0.001) compared to control males (n = 17). (C) Targeted Ephrin expression in the CSDn in an Ephrin mutant background (RN2flp, tub>CD2>Gal4, UASmCD8GFP/UAS Ephrin; Ephrin^KG09118 in which both the CSD neurons were labeled; n = 26) results in partial rescue of the cVA sensitivity compared to the Ephrin mutant males (p = 0.004). . On the other hand, (D) CO₂ sensitivity of Ephrin^KG09118 (n = 12; p = 0.98) is comparable to controls (n = 19). (E) Eph^X652 animals show increased avoidance (n = 8; p = 0.003) towards CO₂ compared to controls (n = 10). As shown earlier, in Eph^X652 the CSDn innervations to V glomerulus are reduced while in Ephrin hypomorphs, these are comparable to controls. The courtship Index towards cVA treated females are normalized to the respective males Courtship Index towards Acetone (Mock) treated Virgin females. Significance was assessed by Mann-Whitney test. *, p<0.05; **, p<0.01; ***, p<0.0001; n.s. (not significant), p>0.05.