Glutamate is an inhibitory neurotransmitter in the Drosophila olfactory system

Abstract

Glutamatergic neurons are abundant in the Drosophila central nervous system, but their physiological effects are largely unknown. In this study, we investigated the effects of glutamate in the Drosophila antennal lobe, the first relay in the olfactory system and a model circuit for understanding olfactory processing. In the antennal lobe, one-third of local neurons are glutamatergic. Using in vivo whole-cell patch clamp recordings, we found that many glutamatergic local neurons are broadly tuned to odors. Iontophoresed glutamate hyperpolarizes all major cell types in the antennal lobe, and this effect is blocked by picrotoxin or by transgenic RNAi-mediated knockdown of the GluClα gene, which encodes a glutamate-gated chloride channel. Moreover, antennal lobe neurons are inhibited by selective activation of glutamatergic local neurons using a nonnative genetically encoded cation channel. Finally, transgenic knockdown of GluClα in principal neurons disinhibits the odor responses of these neurons. Thus, glutamate acts as an inhibitory neurotransmitter in the antennal lobe, broadly similar to the role of GABA in this circuit. However, because glutamate release is concentrated between glomeruli, whereas GABA release is concentrated within glomeruli, these neurotransmitters may act on different spatial and temporal scales. Thus, the existence of two parallel inhibitory transmitter systems may increase the range and flexibility of synaptic inhibition.

Keywords: VGlut; glomerulus; interneuron; olfaction; volume transmission.

Fig. 1.

Glutamatergic LNs in the antennal lobe. (A) Schematic of the antennal lobe circuit. Excitatory neurons are in white, and LNs are in gray. Dashed lines encircle glomeruli. Some cell types and connections are omitted for clarity. (B) Confocal immunofluorescence image of the Drosophila brain. Neuropil is labeled with nc82 antibody (white), and cells that express Gal4 under the control of OK371-Gal4 are labeled with CD8:GFP (green). The somata of Glu-LNs are clustered ventral to the antennal lobes (arrows). Image is a z-projection of coronal optical slices through a 27-µm depth. (C) GFP+ neurons are immunopositive for VGluT (see also ref. 8). Image is a 1-µm confocal slice through one of the clusters of Glu-LN somata shown in B. Note that some VGluT⁺ somata are not GFP⁺ (arrowhead). (D) Coronal optical section through one antennal lobe, with glomerular compartments indicated by a presynaptic marker (nsyb:GFP) expressed specifically in ORNs. One glomerulus (VM4, dashed lines) is outlined as a landmark. (E) Same as D but with staining for the vesicular GABA transporter (VGAT).

Fig. 2.

Morphology and physiology of Glu-LNs. (A) Morphology of a Glu-LN, shown as a z-projection of a traced biocytin fill. This neuron innervated many olfactory glomeruli, and this pattern was seen in 11 of 29 filled cells. Note innervation of both antennal lobes (black circles), which is typical of Glu-LNs. (B) A whole-cell current clamp recording from a Glu-LN of this morphological type. The spikes fired by this cell (arrow) are small. (C) Mean responses of all of the Glu-LNs of this type (±SEM across experiments), quantified as the change in membrane potential averaged over the stimulus period. Odors are 1, butyric acid; 2, pentyl acetate (10⁻²); 3, pentyl acetate (10⁻⁶); 4, water; 5, methyl benzoate; 6, 1-butanol; 7, ethyl acetate. (D) This neuron innervated mainly a single ventral olfactory glomerulus on both sides of the brain. A similar pattern was seen in 8 of 29 fills. (E) A recording from this type of neuron. Spikes (arrow) are small. (F) Mean responses for all of the Glu-LNs of this type. (G) This neuron innervated the putative hygrosensitive/thermosensitive glomeruli just posterior to the antennal lobe. A similar pattern was seen in 10 of 29 fills. (H) A recording from this type of neuron. Note prominent spikes (arrow) and large excitatory postsynaptic potentials (arrowhead). (I) Mean responses for all of the Glu-LNs of this type.

Fig. 3.

GluClα mediates a glutamate-gated chloride conductance in PNs and GABAergic LNs. (A) Whole-cell current-clamp recording from the soma of an antennal lobe PN (Left) and a GABA-LN (Right). A pulse of glutamate in the antennal lobe neuropil (arrow, 10–20 ms) hyperpolarizes both cells. Picrotoxin (100 µM) either abolishes or attenuates the response, depending on the recording. (B) Time course of the effect of picrotoxin on glutamate responses in PNs. Each line is a different PN recording. (C) Effect of picrotoxin on responses to glutamate. Each symbol is a different recording, with means in blue. Overall, the effects of picrotoxin were similar in PNs (n = 12) and GABA-LNs (n = 7). (D) Responses to glutamate before and after applying 100 µM picrotoxin in a wild-type PN (Left) and a PN expressing GluClα RNAi (Right). Arrow indicates iontophoretic pulses. The residual deflection is a stimulus artifact. (E) Hyperpolarizing responses to iontophoresis in both genotypes, before picrotoxin (black) and after picrotoxin (blue). The response to glutamate is significantly smaller in RNAi flies versus wild type (P < 0.05, Student’s t test, n = 6 wild type and 9 RNAi). The percent inhibition by picrotoxin is also significantly smaller (P < 0.0001, Student’s t test).

Fig. 4.

PNs are inhibited by stimulation of either Glu-LNs or GABA-LNs. (A) A recording from a P2X2-expressing Glu-LN, showing that ATP ejection (arrow) depolarized the cell and elicited a train of small spikes. (B) A recording from a PN showing that, when Glu-LNs were stimulated with ATP (arrow), spontaneous spiking paused, and the membrane was slightly hyperpolarized. (C) Mean membrane potential of PNs in response to Glu-LN stimulation, averaged across experiments, ±SEM (n = 13). (D) Mean firing rate of PNs in response to Glu-LN stimulation, averaged across experiments, ±SEM (n = 9; some cells were excluded because they did not spike during the analysis window). (E) Mean membrane potential change of PNs in response to Glu-LN stimulation, averaged across experiments, ±SEM. Picrotoxin significantly reduced the response to Glu-LN stimulation (P < 0.05, paired t test, n = 4). (F–J) Same as above, but this time stimulating GABA-LNs rather than Glu-LNs (n = 13 for H and J, and n = 9 for I). Picrotoxin (5 µM) and CGP54626 (50 µM) significantly reduced the membrane potential change in response to GABA-LN stimulation (P = 0.01, paired t test, n = 8).

Fig. 5.

GABA-LNs are inhibited by stimulation of either Glu-LNs or GABA-LNs. (A) A GABA-LN recording. When Glu-LNs were stimulated with ATP (arrow), spontaneous spiking paused and the membrane potential hyperpolarized. (B) Mean membrane potential of GABA-LNs in response to Glu-LN stimulation, averaged across experiments, ±SEM (n = 6). (C) Mean firing rate of GABA-LNs in response to Glu-LN stimulation, averaged across experiments, ±SEM (n = 6). (D) Mean membrane potential change of GABA-LNs in response to Glu-LN stimulation, averaged across experiments, ±SEM. Picrotoxin significantly reduced the response to Glu-LN stimulation (P < 0.001, paired t test, n = 4). (E–H) Same as above, but this time stimulating GABA-LNs rather than Glu-LNs. Picrotoxin (100 µM) and CGP54626 (50 µM) significantly reduced the response to GABA-LN stimulation (P = 0.001, paired t test, n = 5). Although GABA-LNs lack GABA_B conductances (22), CGP54626 was needed to produce complete block; GABA-LN stimulation may inhibit tonically active ORNs and PNs via both GABA_A and GABA_B receptors, thereby reducing tonic excitation to GABA-LNs.

Fig. 6.

Paired recordings reveal the connectivity of individual LNs. (A) An example of an inhibitory connection from a GABA-LN onto a PN. Depolarizing current was injected into the GABA-LN through the patch electrode (pre, single trial). CGP54626 (50 µM) blocked the response in the PN (post, mean of 50–60 trials). (B) An example of an excitatory connection from a PN onto a GABA-LN. Mecamylamine (50 µM) blocked the response. Scale bars apply to all graphs. (C) In a typical paired recording, there was no effect of stimulating a Glu-LN on a PN. (D) Similarly, in the same pair, there was no effect of stimulating the PN on the Glu-LN. The total number of pairs tested is not identical to C because a few recordings were lost before both directions of connectivity could be tested. (E) An example of an inhibitory connection from a Glu-LN onto a GABA-LN. Note that the presynaptic spikes are very small, which is typical of many Glu-LNs. Picrotoxin (100 µM) blocked the response. (F) An example of an inhibitory connection from a GABA-LN onto a Glu-LN. Picrotoxin (5 µM) blocked the response.

Fig. 7.

Odor responses are disinhibited by knockdown of GluClα in PNs. (A) Odor responses of PNs in four different glomeruli. The membrane potential is low-pass filtered to remove spikes. Each trace represents a different recording, with 10 PNs total. In half of these experiments (blue traces), we used transgenic RNAi to knock down GluClα expression specifically in PNs. Black traces are wild type. Responses are averaged across 5–6 trials. Odor stimuli are pentyl acetate (10⁻² dilution; VM2, VM5, VA1v) and methyl salicylate (10⁻² dilution; DL1). (B) Peristimulus time histograms showing spiking responses of the same PNs. (C) Mean odor-evoked changes in membrane potential (averaged over the 2-s stimulus period) in all cells. Each symbol represents a different recording (n = 5 control, n = 5 RNAi). Responses in wild-type (black) and RNAi flies (blue) are significantly different (P < 0.001, two-way ANOVA). The values for the two wild-type VM5 recordings are so similar that their symbols lie on top of one another. (D) Mean odor-evoked firing rates for the same cells. Responses in wild-type and RNAi flies are significantly different (P < 0.005, two-way ANOVA). (E) Schematic showing interactions between PNs, and LNs in the two genotypes. Some connections are omitted for clarity.