Synaptic Development in Diverse Olfactory Neuron Classes Uses Distinct Temporal and Activity-Related Programs

Antennal lobe synapses are added over two distinct phases of pupal development

In the antennal lobe (AL), ORNs, PNs, and LNs (Fig. 1A) comprise three major neuron types that contribute to the detection and subsequent transfer of olfactory information to higher order brain structures (Vosshall et al., 2000; Jefferis et al., 2001; Tanaka et al., 2009). Different classes of each neuron type project to the roughly 50 glomeruli that make up the AL, which are subdivided based on the type of olfactory information they receive (Fig. 1B), and form synapses with each other to generate functional circuits (Suh et al., 2004; Hallem and Carlson, 2006; Jefferis et al., 2007; Grabe and Sachse, 2018). To begin to understand how synapse number changes over time in the antennal lobe at large, we measured the levels of Bruchpilot, an active zone protein and presynaptic marker (Wagh et al., 2006; Fouquet et al., 2009), during pupal development. During the roughly 4-d metamorphic process of pupal development, larvae break down their cellular architecture, including their current nervous system, and re-form as an adult fly (Thummel, 2001). Neuronal wiring, axon pathfinding, and partner matching are largely complete by 48 h after puparium formation (APF), indicating that synapses are unlikely to form before the component neurons of the AL have reached their final destination (Komiyama and Luo, 2006). We used antibody staining to measure Bruchpilot intensity (Laissue et al., 1999) from 48 to 96 h APF (Fig. 1C–G'). After 96 h APF, pupation is complete and the adult fly ecloses. Quantification of Bruchpilot levels can serve as a measurement of the amount of active zone protein present; in cases of synapse development, the levels of Bruchpilot would necessarily increase concomitantly with synapse addition as sites of neurotransmitter release grow more numerous. Changes in Bruchpilot are consistent with synapse formation and are regulated by genetic and molecular pathways that influence synaptic development (Wagh et al., 2006; Fouquet et al., 2009; Mosca and Luo, 2014; Mosca et al., 2017). We found that average Bruchpilot intensity in the AL increases significantly from 60 to 72 h APF as well as from 84 to 96 h APF (Fig. 1H). Bruchpilot intensity remains largely unchanged, however, between 72 and 84 h APF, suggesting that there are multiple phases of general synapse addition (Fig. 1H). In contrast, the average intensity of N-cadherin, which functions as a general neuropil marker in the fly brain (Hummel and Zipursky, 2004), did not change during the course of pupal development (Fig. 1I). This increase in Bruchpilot levels during development suggests that the total number of synaptic connections within the entire antennal lobe circuit increases during specific time frames.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Synapse addition across the antennal lobe during pupal development occurs during distinct time frames. A, B, Schematic of the Drosophila antennal lobes. Three major neuron classes project to the antennal lobes to relay and process olfactory information: olfactory receptor neurons (green), projection neurons (magenta), and local interneurons (blue; A). Each antennal lobe consists of ∼50 glomeruli (B) with the three glomeruli examined throughout this study (DA1, VA1lm, and DL4) highlighted (orange). C–G', Representative confocal image stacks of pupal antennal lobes stained with antibodies to Bruchpilot (red) and N-cadherin (blue) at 48 h (C), 60 h (D), 72 h (E), 84 h (F), and 96 h (G) after puparium formation. H, I, Quantification of average staining intensity for Brp (H) and N-cadherin (I) from 48 to 96 h APF. Brp intensity increases significantly from 60 to 72 h APF as well as from 84 to 96 h APF. N-cadherin intensity does not change throughout this time frame. For each time point, n ≥ 9 antennal lobes from 8 brains. ***p < 0.001. ns = not significant. Scale bar = 20 μm.

Pheromone-sensing olfactory receptor neurons develop synapses during pupation then remove synapses in late adulthood

Our results suggest that the multiple classes of antennal lobe neurons, when considered together, add synapses during specific time frames of pupal development (Fig. 1). However, our data represents the aggregate of all neuronal classes and does not distinguish between each of the major classes of antennal lobe neurons (ORNs, PNs, and LNs). To understand how the cell-type specific synaptic development of each class of neurons progresses, we expressed Bruchpilot-Short (Brp-Short), an active zone label associated with presynaptic contacts (Mosca and Luo, 2014), in each class of neurons using the UAS-GAL4 binary expression system (Brand and Perrimon, 1993). We performed a high-resolution time course analysis of synaptic development throughout pupation and adulthood using confocal microscopy followed by three-dimensional rendering (see Materials and Methods). Brp-Short is a fluorescently tagged truncated variant of the active zone protein Bruchpilot (Fouquet et al., 2009; Mosca and Luo, 2014) that interacts with full-length Bruchpilot without interfering with its function or adding ectopic synapses (Wagh et al., 2006; Fouquet et al., 2009; Mosca and Luo, 2014; Mosca et al., 2017). Brp-Short signal manifests as synaptically localized puncta, acting as a proxy label for active zones, and accurately reports synapse number from specific cell populations and fold changes in synapse number because of genetic perturbation, as supported by genetic and electron microscopy studies (Fouquet et al., 2009; Kremer et al., 2010; Christiansen et al., 2011; Berger-Müller et al., 2013; Mosca and Luo, 2014; Coates et al., 2017, 2020; Mosca et al., 2017). We also labeled each neuronal class examined with membrane-bound GFP to fluorescently label neurites (Lee and Luo, 1999), enabling the quantification of neurite volume for each neuron class within a given glomerulus (Mosca and Luo, 2014; Mosca et al., 2017). To establish a time course of synaptic development, we quantified Brp-Short puncta and neuronal membrane volume marked by mCD8-GFP every 12 h from 48 to 92 h APF and then every 3–5 d, at regular intervals, from 0 to 24 d after eclosion in ORN classes that innervate distinct glomeruli. We specifically assayed pupal synapse development at 92 h APF to ensure that all pupae were collected before eclosion (at 25°C, adult eclosion typically occurs between 92 and 96 h APF).

ORNs are the primary sensory neurons responsible for conveying odorant information from the environment to the antennal lobes (Wilson, 2013). We first sought to assess the temporal synaptic development of ORNs that project to the VA1lm glomerulus (Fig. 2A) and are genetically accessed by Or47b-GAL4 (Vosshall et al., 2000). This glomerulus receives pheromone odorant information (Vosshall et al., 2000; Fishilevich and Vosshall, 2005), is involved in oviposition as well as courtship decisions, influencing preference for younger mates (Zhuang et al., 2016; Chin et al., 2018; Sethi et al., 2019), and is sexually dimorphic, with male glomeruli being ∼40% larger than female glomeruli (Fishilevich and Vosshall, 2005; Mosca and Luo, 2014). We quantified Brp-Short puncta and neurite volume at 48, 60, 72, 84, and 92 h APF as well as at 0, 5, 10, 15, and 20 d after eclosion (Fig. 2B–M'') in both male and females separately to assay any sexual dimorphisms in synaptic development. In male pupal ORNs, there were no synapses detectable at 48 h APF, but synapses were then steadily added over pupal development with the number of Brp-Short puncta doubling from 60 to 92 h APF (Fig. 2N; Extended Data Table 2-1). Neurite volume was also negligible at 48 h APF, possibly because the neurons must become more mature before mCD8-GFP can stably localize to the membrane, but then increased 3-fold from 60 h APF to 92 h APF (Fig. 2O; Extended Data Table 2-1). This glomerulus-specific time course of synapse formation differs from the time course of gross AL synapse formation (Fig. 1) in that synapses were added steadily throughout pupal development rather than in two separated time periods. After eclosion, there was a 36% increase in the number of Brp-Short puncta from 0 to 5 d of age followed by a 13% decrease in puncta number from 5 to 10 d of age (Fig. 2P; Extended Data Table 2-1). Following 10 d of age, synapse number remained constant throughout the analysis. Adult neurite volume increased by 29% between 0 and 5 d after eclosion and then an additional 21% by 15 d after eclosion (Fig. 2Q; Extended Data Table 2-1). Together, these data suggest that synapses are steadily added from 48 h APF to 5 d after eclosion, demonstrating that individual classes of olfactory neurons may show distinct synapse addition phases than the aggregate antennal lobe. This is followed by a pruning phase where synapses are removed before the completion of synaptic development and establishment of a steady state synapse number. Neurites continually grow and add volume until 15 d after eclosion, after which point, they achieve a steady state volume. We additionally determined the time course of female VA1lm ORN synaptic development (Fig. 3A–K'') to determine whether any of the developmental mechanisms displayed sexual dimorphism or whether they occurred similarly to male ORNs. There were fewer Brp-Short puncta and lower neurite volume at all time points in female VA1lm ORNs compared with male VA1lm ORNs, consistent with established sexual dimorphisms in size and synapse number (Stockinger et al., 2005; Mosca and Luo, 2014). We found that females also had a 2-fold increase in puncta and 3-fold increase in neurite volume from 60 to 92 h APF (Fig. 3L,M; Extended Data Table 3-1). Adult female ORNs exhibited a similar increase in Brp-Short puncta to the male ORNs from 0 to 5 d of age that was followed by a decrease from 5 to 10 d after eclosion (Fig. 3N; Extended Data Table 3-1). Adult neurite volume, however, significantly increased from 0 to 5 d after eclosion and then did not significantly change afterward (Fig. 3O; Extended Data Table 3-1), suggesting that female ORNs achieve their mature adult volume earlier than male ORNs. This suggests similarities in the general addition of synapses throughout development despite sexual dimorphisms in absolute synapse number but a minor difference in the growth rate of neurites during development between males and females. Overall, these time courses demonstrate a general, previously unappreciated, pattern of synaptic development for a sexually-dimorphic, pheromone-specific glomerulus in which synapses continually form until 5 d after eclosion and then decrease before stabilizing to a mature quantity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

A developmental time course of synapse number and neurite volume in male VA1lm ORNs. A, Schematic of the Drosophila antennal lobes showing ORNs (green) of the VA1lm glomerulus (orange). B–F'', Representative confocal image stacks of male pupal VA1lm ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–G'', High-magnification, single optical section of ORNs aged to 92 h APF from inset in F showing synaptic labels. Note that synaptic puncta are not evident until 60 h APF. H–L'', Representative confocal image stacks of male adult VA1lm ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F'' at 0 d (H), 5 d (I), 10 d (J), 15 d (K), and 20 d (L) after eclosion. M–M'', High-magnification, single optical section of 20-d-old ORNs from inset in L showing synaptic labels. N, O, Quantification of Brp-Short-mStraw puncta (N) and membrane GFP volume (O) for pupal male ORNs. Both synaptic puncta and neurites are not measurable until 60 h APF, at which point they increase during the remainder of pupal development. P, Q, Quantification of Brp-Short-mStraw puncta (P) and membrane GFP (Q) volume for adult male ORNs. Synaptic puncta significantly increase from 0 to 5 d old and then decrease between 5 and 10 d after eclosion. Neurite volume increases from 0 to 15 d of age before achieving a steady state. For each time point, n ≥ 8 glomeruli from 4 brains. *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 2-1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Synapse formation for female ORNs of the VA1lm glomerulus. A, Schematic of the Drosophila antennal lobes showing ORNs (green) of the VA1lm glomerulus (orange). B–F'', Representative confocal image stacks of female pupal VA1lm ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. Note that synaptic puncta do not appear until 60 h APF. G–K'', Representative confocal image stacks of female adult VA1lm ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F'' at 0 d (G), 5 d (H), 10 d (I), 15 d (J), and 20 d (K) after eclosion. L, M, Quantification of Brp-Short-mStraw puncta (L) and membrane GFP volume (M) for pupal female ORNs. Both synaptic puncta and neurites are not measurable until 60 h APF, at which point they increase during the remainder of pupal development. N, O, Quantification of Brp-Short-mStraw puncta (N) and membrane GFP (O) volume for adult female ORNs. Synaptic puncta significantly increase from 0 to 5 d old and then decrease between 5 and 10 d after eclosion. Neurite volume increases from 0 to 5 d of age before achieving a steady state. For each time point, n ≥ 10 glomeruli from 5 brains. ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 3-1.

We next sought to determine whether another sexually dimorphic class of ORNs, the Or67d-positive ORNs that project to the DA1 glomerulus (Fig. 4A), develop similarly to VA1lm ORNs or whether there are differences based on glomerulus (Datta et al., 2008). DA1 ORNs detect male pheromones, including 11-cis-Vaccenyl acetate, that are involved in courtship and aggression for both male and female flies (Kurtovic et al., 2007; Datta et al., 2008; Wilson, 2013). Similar to VA1lm, DA1 is a sexually-dimorphic glomerulus with increased glomerular size in males compared with females (Stockinger et al., 2005; Mosca and Luo, 2014). We used a GAL4 driver that labels DA1 ORNs via GAL4 expression under the control of the endogenous Or67d promoter (Stockinger et al., 2005) to examine the time course of synapse formation during adult stages; Or67d-GAL4 is not active during pupal development precluding our analysis of synapse development in these ORNs from 48 to 92 h APF. We quantified Brp-Short puncta number and neurite volume for male DA1 ORNs every 3 d from 0 to 24 d of age (Fig. 4B–J'') and found that Brp-Short puncta increased by 50% between 0 and 6 d of age before achieving and maintaining a steady state synapse number until day 15. Surprisingly, we observed a 25% decline in synapse number between days 15 and 18 (Fig. 4K; Extended Data Table 4-1). Neurite volume doubled from 0 to 6 d old and then did not significantly change for the remainder of adulthood (Fig. 4L; Extended Data Table 4-1), thus maintaining a steady state. We found congruent results with female DA1 ORNs (Fig. 5A–J''). Brp-Short puncta number in female ORNs significantly increased by 20% (Fig. 5K; Extended Data Table 5-1) from 0 to 3 d of age, maintained a steady state from 3 to 12 d, and then trended down by 15% from 12 to 18 d of age. Puncta number then significantly increased by 19% from 18 to 21 d age before stabilizing for the remainder of adulthood. Neurite volume significantly increased by 60% from 0 to 12 d of age and then did not significantly change for the rest of adulthood (Fig. 5L; Extended Data Table 5-1). Thus, the general female time course of DA1 ORNs was nearly identical to the male time course, with slight variability in the exact times of synapse change. Compared with VA1lm ORNs, DA1 ORN synapse development showed similar synapse addition until 5–6 d after eclosion but did not display a synapse reduction phase until 12–18 d old, while VA1lm ORNs displayed synapse pruning from 5–10 d old. These data imply that pheromone-specific ORNs follow similar general patterns of synaptic development but have critical differences in the exact timing of developmental phases.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

A developmental time course of synapse number and neurite volume in adult male DA1 ORNs. A, Schematic of the antennal lobes showing ORNs (green) of the DA1 glomerulus (orange). B–J'', Representative confocal image stacks of male adult DA1 ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 0 d (B), 3 d (C), 6 d (D), 9 d (E), 12 d (F), 15 d (G), 18 d (H), 21 d (I), and 24 d (J) of age. K, L, Quantification of Brp-Short-mStraw puncta (K) and membrane GFP volume (L) for adult male DA1 ORNs. Both Brp-Short puncta number and neurite volume steadily increase from 0 to 6 d of age; after 6 d after eclosion, neurite volume stabilizes while Brp-Short puncta maintain a steady state before decreasing between 15 and 18 d of age. For each time point, n ≥ 14 glomeruli from 7 brains. **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 4-1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Synapse formation for the DA1 glomerulus in female ORNs. A, Schematic of the antennal lobes showing ORNs (green) of the DA1 glomerulus (orange). B–J'', Representative confocal image stacks of female adult DA1 ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 0 (B), 3 (C), 6 d (D), 9 d (E), 12 d (F), 15 d (G), 18 d (H), 21 d (I), and 24 d (J) of age. K, L, Quantification of Brp-Short-mStraw puncta (K) and membrane GFP volume (L) for adult female DA1 ORNs. Both Brp-Short puncta number and neurite volume steadily increase from 0 to 3 d of age, after which neurite volume stabilizes. Brp-Short puncta maintain a steady state before decreasing and subsequently increasing between 15 and 21 d of age. For each time point, n ≥ 11 glomeruli from 6 brains. *p < 0.05, **p < 0.01, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 5-1.

Bruchpilot represents a core component of the active zone and is an accurate reporter of synaptic release sites (Kittel et al., 2006; Wagh et al., 2006; Mosca and Luo, 2014). However, Bruchpilot protein levels can show changes that are independent of changes in synapse number (Sugie et al., 2015; Woźnicka et al., 2015). Therefore, we also sought to examine the time course of DA1 ORN synaptic development using an additional synapse marker that was independent of Brp-Short to ensure that the differences we observed were reflective of synaptic connections and not simply changes in the Bruchpilot protein itself. The presynaptic active zone is comprised of multiple proteins that function together to recruit and release synaptic vesicles during neurotransmission (Südhof, 2012). RIM-BP, the Rab3-interacting molecule (RIM)-binding protein, is an essential active zone protein that couples synaptic vesicle recruitment machinery to vesicle release architecture (Hibino et al., 2002; Kiss et al., 2011; Petzoldt et al., 2020). In Drosophila, RIM-BP is a central component of the active zone (Liu et al., 2011) and labels most, if not all active zones in the fly nervous system. To examine the time course of synaptic development, we expressed a GFP-tagged transgenic RIM-BP (RBP-GFP) using Or67d-GAL4 and quantified RBP-GFP puncta from 0 to 24 d of age for both males (Fig. 6A–I') and females (Fig. 6J–R') in a fashion analogous to our Brp-Short experiments. When co-expressed in the same Or67d-positive ORNs, we observed near (but not complete) overlap between Brp-Short and RBP-GFP puncta (Fig. 6S–U'), suggesting that they were both labeling a nearly identical population of synaptic contacts. In male ORNs, we found that RBP-GFP puncta significantly increased by 66% (Fig. 6V) from 0 to 9 d of age and then maintained a steady-state throughout most of adulthood. RBP-GFP puncta then significantly decreased by 13% from 21 to 24 d of age. For female ORNs, RBP-GFP puncta number significantly increased by 66% (Fig. 6W) from 3 to 6 d of age before maintaining a steady-state for much of adulthood. Similarly to the male ORNs, RBP-GFP puncta significantly decreased by 18% between 12 and 15 d of age before stabilizing for the remainder of observed adulthood. Just as we found with the Brp-Short time courses of DA1 ORNs (Figs. 4 and 5), the RBP-GFP time courses of synapse formation for males and females were nearly identical with only slight variations in the timings of synapse change. The similarity of the time courses obtained using either Brp-Short or RIM-BP imaging validate that our findings can accurately demonstrate the temporal pattern of synaptic development for ORNs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Developmental time courses of synapse formation using RIM-Binding Protein in male and female DA1 ORNs. A–I', Representative confocal image stacks of male adult DA1 ORNs expressing RIM-Binding Protein-tagged GFP (RBP-GFP) and stained with antibodies against GFP (green) and N-cadherin (blue) at 0 d (A), 3 d (B), 6 d (C), 9 d (D), 12 d (E), 15 d (F), 18 d (G), 21 d (H), and 24 d (I) of age. J–R', Representative confocal image stacks of female adult DA1 ORNs expressing RBP-GFP and stained with antibodies as in A–I' at 0 d (J), 3 d (K), 6 d (L), 9 d (M), 12 d (N), 15 d (O), 18 d (P), 21 d (Q), and 24 d (R) of age. S–U', Single optical sections with high-magnification insets (dashed boxes) of female DA1 ORNs at 24 d of age expressing Brp-Short-mStraw and RBP-GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue). V, Quantification of RBP-GFP puncta for male DA1 ORNs. RBP-GFP puncta number increases until between 6 and 9 d age. RBP-GFP puncta maintain a steady-state until decreasing at 24 d of age. W, Quantification of RBP-GFP puncta for female DA1 ORNs. RBP-GFP puncta number increases until 6 d of age. RBP-Puncta maintain a steady-state until puncta number decreases at 15 d of age, stabilizing for the remainder of adulthood. For each time point, n ≥ 8 glomeruli from 6 brains. *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm.

Food-sensing olfactory receptor neurons develop synapses during pupation and early adulthood then stabilize

Drosophila rely on olfaction to find and interact with mates as well as identify suitable food sources. Food sensing glomeruli lack sexual dimorphism seen in size and synapse number of pheromone-sensing glomeruli (Couto et al., 2005; Stockinger et al., 2005; Mosca and Luo, 2014), likely because all flies have a shared, unabating need to identify food sources. Our assays on sexually dimorphic glomeruli thus far followed the synaptic development of pheromone-sensing ORNs but we further sought to understand whether the same developmental features apply in glomeruli that sense food odorants or whether these glomeruli followed a distinct developmental program. We first established a time course for male ORNs that project to the DL4 glomerulus (Fig. 7A), a class of neurons that conveys information about volatile chemical odorants related to food (Endo et al., 2007; Martin et al., 2013; Mosca and Luo, 2014), throughout pupal (48–92 h APF) development (Fig. 7B–F'') and into adulthood from 0 to 24 d old (Fig. 7G–O''). We found that pupal Brp-Short puncta increased 1.5-fold from 48 to 72 h APF and then did not significantly change (Fig. 7P; Extended Data Table 7-1) throughout the rest of pupation. Pupal neurite volume also exhibited a nearly 1.5-fold change from 48 to 72 h APF before reaching a steady-state neurite volume that was maintained until adulthood (Fig. 7Q; Extended Data Table 7-1), suggesting much of the development of pupal DL4 ORNs occurs well in advance of eclosion. Adult male Brp-Short puncta in DL4 ORNs significantly increased by 1.25-fold from 0 to 12 d after eclosion before achieving their mature synapse number and maintaining that number of puncta throughout the remainder of observed adulthood (Fig. 7R; Extended Data Table 7-1), further demonstrating that synapse formation continues into adulthood. Adult neurite volume did not significantly change from 0 to 6 d after eclosion, maintaining a similar level to that measure during pupal stages; intriguingly, however, neurite volume declined from 6 to 15 d by 39% and from 18 to 21 d by 60% (Fig. 7S; Extended Data Table 7-1). We additionally generated a time course of development for female DL4 ORNs (Fig. 8A–O'') to examine any potential sexual dimorphism. Female DL4 ORNs showed a 1.5-fold increase in both Brp-Short puncta (Fig. 8P; Extended Data Table 8-1) and neurite volume (Fig. 8Q; Extended Data Table 8-1) from 48 to 60 h APF with no significant changes for the remainder of pupal development. Adult female DL4 Brp-Short puncta significantly increased between 0 and 6 d after eclosion by 33% before maintaining a steady-state. Brp-Short puncta then decreased by 25% from 18 to 24 d of age. (Fig. 8R; Extended Data Table 8-1). Female DL4 ORN adult neurite volume also did not significantly change from 0 to 6 d of age but significantly declined between 6 and 9 d (Fig. 8S; Extended Data Table 8-1). While neurite volume steadily decreased as the fly aged in both sexes, this is likely an artifactual effect of the AM29-GAL4 driver weakening over time (Endo et al., 2007). We do not observe significant decrease of glomerular size in a general neuropil stain, suggesting volume remains largely constant. Overall, however, the absolute puncta numbers and neurite volumes at each stage were comparable between males and females, implying that synaptic development for food sensing ORNs is not sexually dimorphic and that both sexes use similar developmental programs for food-sensing ORNs. Thus, food-sensing ORNs use a different time course of synaptic development than pheromone-sensing glomeruli like DA1 and VA1lm, one that emphasizes early development and lacks sexual dimorphism. Taken together, this indicates that ORNs of sexually dimorphic glomeruli follow similar rules for synaptic development which are distinct from the rules followed by nonsexually dimorphic ORNs and that within classes, each ORN subtype has subtle, individual variations in development.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

A developmental time course of synapse number and neurite volume in male DL4 ORNs. A, Schematic of the Drosophila antennal lobes showing ORNs (green) of the DL4 glomerulus (orange). B–F'', Representative confocal image stacks of male pupal DL4 ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–O'', Representative confocal image stacks of male adult D4 ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F'' at 0 d (G), 3 d (H), 6 d (I), 9 d (J), 12 d (K), 15 d (L), 18 d (M), 21 d (N), and 24 d (O) of age. P, Q, Quantification of Brp-Short-mStraw puncta (P) and membrane GFP volume (Q) for pupal male DL4 ORNs. Both synaptic puncta and neurite volume increase throughout pupal development. R, S, Quantification of synaptic puncta (R) and neurite volume (S) for adult male DL4 ORNs. Synaptic puncta trends upward from 0 to 3 d of age and then stabilize. Neurite volume steadily decreases over time. For each time point, n ≥ 11 glomeruli from 6 brains. *p < 0.1, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 7-1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

A developmental time course of synapse number and neurite volume in female DL4 ORNs. A, Schematic of the Drosophila antennal lobes showing ORNs (green) of the DL4 glomerulus (orange). B–F'', Representative confocal image stacks of female pupal DL4 ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–O'', Representative confocal image stacks of female adult DL4 ORNs expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F'' at 0 d (G), 3 d (H), 6 d (I), 9 d (J), 12 d (K), 15 d (L), 18 d (M), 21 d (N), and 24 d (O) of age. P, Q Quantification of Brp-Short-mStraw puncta (P) and membrane GFP volume (Q) for pupal female DL4 ORNs. Both synaptic puncta and neurite volume increase from 48 to 60 h APF. They then trend upward for the remainder of pupal development. R, S, Quantification of synaptic puncta (R) and neurite volume (S) for adult female DL4 ORNs. Synapses trend upward from 0 to 6 d old and then remain largely stable. Neurite volume remains steady until 12 d of age and then begins to decline. For each time point, n ≥ 10 glomeruli from 5 brains. *p < 0.1, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 8-1.

Projection neurons continually develop synapses during pupation and early adulthood

Projection neurons (PNs) comprise the second order neurons of the olfactory system, relaying odorant information from ORNs that project onto PN dendrites in the antennal lobe to higher order brain structures such as the mushroom bodies and the lateral horn (Modi et al., 2020). ORN to PN communication is vital for making behavioral decisions, including foraging, oviposition, and courtship, based on the odor information available to the fly (Jefferis et al., 2001). PNs also arrive in the antennal lobes before ORNs and begin prepatterning the various glomeruli in preparation for ORN arrival (Jefferis et al., 2004). Prior studies, however, have only examined wiring and not synapse formation. Wiring and synaptogenesis are separable but equally essential events, and despite the importance of these neurons, we lack a detailed understanding of how they develop synapses over time. We addressed this by constructing a time course of synapse formation for male DA1 PNs (Fig. 9A) using genetic access via the Mz19-GAL4 driver (Jefferis et al., 2004); Mz19-GAL4 drives expression in DA1, VA1d, and DC3 PNs, but we restricted our analysis to DA1 PNs to enable a direct comparison to the ORNs (and LNs, see below) of the same glomerulus (Figs. 4 and 5). This allowed us to assess whether different classes of neurons belonging to the same microcircuit, such as those observed within an olfactory glomerulus, would share developmental programs or use distinct modes of synapse addition. As with ORNs, we quantified Brp-Short puncta number and neurite volume from 48 to 92 h APF and from 0 to 24 d after eclosion (Fig. 9B–O''). Unlike ORNs, however, Brp-Short puncta were clearly visible in PNs at 48 h APF and steadily increased in number over 2.5-fold by 92 h APF (Fig. 9P; Extended Data Table 9-1). Pupal neurite volume also continually increased during pupation, doubling from 48 to 92 h APF (Fig. 9Q; Extended Data Table 9-1). These data show that PN synapses develop throughout the course of pupation. During adulthood, Brp-Short puncta continued to be added after eclosion with puncta number increasing 2-fold from 0 to 6 d of age, followed by no significant changes for the rest of adulthood (Fig. 9R; Extended Data Table 9-1). Neurite volume increased nearly 2-fold from 0 to 3 d after eclosion and then remained largely unchanged throughout the rest of adulthood (Fig. 9S; Extended Data Table 9-1). Therefore, DA1 PNs develop synapses throughout pupation and even early into adulthood before stabilizing at their mature quantity. Unlike ORNs, however, they lack a pruning phase and reach their mature state more quickly.

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.

A developmental time course of synapse number and neurite volume in male DA1 projection neurons. A, Schematic of the antennal lobes showing PNs (magenta) of the DA1 glomerulus (orange). B–F'', Representative confocal image stacks of male pupal DA1 PNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–O'', Representative confocal image stacks of male adult DA1 PNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F at 0 d (G), 3 d (H), 6 d (I), 9 d (J), 12 d (K), 15 d (L), 18 d (M), 21 d (N), and 24 d (O) of age. P, Q, Quantification of Brp-Short-mStraw puncta (P) and membrane GFP volume (Q) for pupal male DA1 PNs. Brp-Short puncta and accompanying neurites are visible at 48 h APF and increase steadily during pupal development. R, S, Quantification of synaptic puncta (R) and neurite volume (S) for adult male DA1 PNs. Brp-Short puncta and neurite volume increase until 6 d after eclosion after which both stabilize for the remainder of adulthood. For each time point, n ≥ 10 glomeruli from 5 brains. **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 9-1.

We additionally generated a time course of development for female DA1 PNs to examine any potential sexual dimorphisms in PN synaptic development (Fig. 10A–O''). In female PNs, both the number of Brp-Short puncta (Fig. 10P; Extended Data Table 10-1) and neurite volume increased 2-fold from 48 to 92 h APF (Fig. 10Q; Extended Data Table 10-1), showing that males and females both develop synapses during the entirety of pupation. Adult female PNs exhibited a nearly 2-fold increase in Brp-Short puncta from 0 to 6 d of age, and then did not change for the remainder of adulthood (Fig. 10R; Extended Data Table 10-1). Female neurite volume increased 1.5-fold from 0 to 3 d after eclosion and then did not significantly change (Fig. 10S; Extended Data Table 10-1; similar to males). These data demonstrate that male and female PNs have nearly identical time courses of synaptic development. Furthermore, their exact time frames of synapse addition and stabilization are distinct from those of ORNs. Therefore, PNs that receive information predominantly from pheromone-sensing ORNs have a unique pattern of synapse formation, demonstrating that different neuronal classes within the same microcircuit have distinct patterns of synaptic development. This may underlie their specific function in the antennal lobe olfactory circuit (see Discussion).

• Download figure
• Open in new tab
• Download powerpoint

Figure 10.

Synapse formation for female PNs of the DA1 glomerulus. A, Schematic of the antennal lobes showing PNs (magenta) of the DA1 glomerulus (orange). B–F'', Representative confocal image stacks of female pupal DA1 PNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–O'', Representative confocal image stacks of female adult DA1 PNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F at 0 d (G), 3 d (H), 6 d (I), 9 d (J), 12 d (K), 15 d (L), 18 d (M), 21 d (N), and 24 d (O) of age. P, Q, Quantification of Brp-Short-mStraw puncta (P) and membrane GFP volume (Q) for female pupal DA1 PNs. Synaptic puncta are added between 48 and 60 h APF as well as between 72 and 84 h APF. Neurite volume steadily increases until ∼84–92 h APF. R, S, Quantification of synaptic puncta (R) and neurite volume (S) for female adult DA1 PNs. Synapses addition occurs from 0 to 6 d of age and then stabilizes. Neurite volume slightly increases from 0 to 3 d and the remains constant. For each time point, n ≥ 12 glomeruli from 6 brains. *p < 0.1, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 10-1.

Local interneurons rapidly undergo synapse formation late in pupation to reach maturity

In the antennal lobe, the most diverse class of cells are local interneurons (Ng et al., 2002; Tanaka et al., 2008, 2012; Chou et al., 2010). Compared with ORNs and PNs, less is known about their precise physiological functions. LNs have been implicated in a number of processing steps in olfactory information including gain control, regulating interglomerular communication, and refining olfactory signal (Tanaka et al., 2009; Chou et al., 2010; Yaksi and Wilson, 2010; Hong and Wilson, 2015). Within the antennal lobe, different types of LNs synapse onto both ORN and PN targets. The multiglomerular LNs comprise the most numerous class and innervate multiple glomeruli throughout the antennal lobe. This innervation occurs after PN and ORN projections have entered the antennal lobes (Chou et al., 2010). Again, however, as wiring and synaptogenesis are separable events, there is little known about how synapse addition and active zone development occurs in these essential cells. Given their prominence among the olfactory LNs, we wanted to understand how their synapses developed. Therefore, we established a time course of synapse addition and neurite growth during development for a subset of multiglomerular LNs accessed genetically by NP3056-GAL4 (Chou et al., 2010) for males (Fig. 11A) as well as females (Fig. 12A). To enable a direct comparison with Or67d-expressing ORNs and Mz19-accessed PNs, we focused our examination specifically to the projections of those multiglomerular LNs within the DA1 glomerulus (Fig. 11B–O''). As NP3056-GAL4 expresses in a heterogeneous population of LNs (Chou et al., 2010; Liou et al., 2018; Suzuki et al., 2020), these data represent the sum of all multiglomerular LN synapses in the DA1 glomerulus. This allowed us to compare synapse formation between each neuron class for the same precise region of the antennal lobe, eliminating cellular geography as a potential variable influencing synaptic development.

• Download figure
• Open in new tab
• Download powerpoint

Figure 11.

A developmental time course of male local interneuron synapse number and neurite volume in the DA1 glomerulus. A, Schematic of the antennal lobes showing LNs (blue) of the DA1 glomerulus (orange). B–F'', Representative confocal image stacks of male pupal DA1 LNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–O'', Representative confocal image stacks of male adult DA1 LNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F'' at 0 d (G), 3 d (H), 6 d (I), 9 d (J), 12 d (K), 15 d (L), 18 d (M), 21 d (N), and 24 d (O) of age. P, Q, Quantification of Brp-Short-mStraw puncta (P) and membrane GFP volume (Q) for pupal male DA1 LNs. Synaptic puncta and neurites begin to significantly form between 60 and 72 h APF and rapidly increase for the remainder of pupal development. R, S, Quantification of synaptic puncta (R) and neurite volume (S) for adult male DA1 LNs. There is no significant difference in the number of synaptic puncta or neurite volume across all adult times examined. For each time point, n ≥ 10 glomeruli from 5 brains. *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 11-1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 12.

Synapse formation in the DA1 glomerulus for female LNs. A, Schematic of the antennal lobes showing LNs (blue) of the DA1 glomerulus (orange). B–F'', Representative confocal image stacks of female pupal DA1 LNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies against mStraw (red), GFP (green), and N-cadherin (blue) at 48 h (B), 60 h (C), 72 h (D), 84 h (E), and 92 h (F) APF. G–O'', Representative confocal image stacks of female adult DA1 LNs (dashed white lines) expressing Brp-Short-mStraw and membrane-tagged GFP and stained with antibodies as in B–F'' at 0 d (G), 3 d (H), 6 d (I), 9 d (J), 12 d (K), 15 d (L), 18 d (M), 21 d (N), and 24 d (O) of age. P, Q, Quantification of Brp-Short-mStraw puncta (P) and membrane GFP volume (Q) for female pupal DA1 LNs. Addition of quantifiable synaptic puncta and neurite volume begins between 48 and 60 h APF and then occurs rapidly until 84–92 h APF. R, S, Quantification of synaptic puncta (R) and neurite volume (S) for female adult DA1 LNs. Synapses remain largely constant throughout adulthood, except for a decrease and subsequent increase from 3 to 9 d of age. Neurite volume is constant throughout adulthood. For each time point, n ≥ 11 glomeruli from 6 brains. **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm. Raw data are provided in Extended Data Table 12-1.

From 48 to 60 h APF, we observed very few LN synapses in DA1 and neurite elaboration within the glomerulus was minimal (Fig. 11B–C''). Between 60 and 92 h APF, however, there was significant addition of active zones and neurite growth (Fig. 11D–F''). In the final 30 h of pupation, active zone number increased nearly 10-fold (Fig. 11P; Extended Data Table 11-1). LN neurite volume within DA1 also increased nearly 2.5-fold during the same time period (Fig. 11Q; Extended Data Table 11-1). We also determined the time course of synapse addition and neurite growth for multiglomerular LNs during the first 24 d of adulthood. At day 0, immediately after eclosion, we observed a higher number of active zones than at 92 h APF, indicating that even during the short time between 92 h APF and eclosion, there was a significant amount of synapse addition. The same was true of neurite outgrowth. Surprisingly, however, we found that both synapse number (Fig. 11R; Extended Data Table 11-1) and neurite outgrowth (Fig. 11S; Extended Data Table 11-1) remained constant from 0- to 24-d-old adult flies. Female DA1 LNs appeared nearly identical to male LNs (Fig. 12B–O''). From 60 to 92 h APF, female Brp-Short puncta increased 7-fold (Fig. 12P; Extended Data Table 12-1). In the same time frame, neurite volume increased 2-fold (Fig. 12Q; Extended Data Table 12-1). Female adult LNs differed from males in that Brp-Short puncta decreased 37% from 3 to 6 d of age and then re-increased 49% at 9 d of age (Fig. 12R; Extended Data Table 12-1). Neurite volume, however, did not change after eclosion (Fig. 12S; Extended Data Table 12-1). Overall, this suggests that LNs complete most synapse addition during pupal development. Males and females both showed the same general pattern of synapse formation, with the expected sex-specific dichotomy in synapse number and neurite volume (as observed previously in ORNs and PNs; Stockinger et al., 2005; Figs. 4, 5, 9, and 10). Female LNs, however, showed more variation throughout the adult time course than males, who retained their steady state of synapse number once achieved. This unique timeline of development for multiglomerular LNs in the DA1 glomerulus highlights an additional developmental program that differs from the programs previously observed in ORNs and PNs.

Decreasing neuronal activity lowers synapse number in antennal lobe neurons

In a glomerular microcircuit, ORNs, PNs, and LNs have interconnected projections and fulfill distinct individual roles to support the larger goal of the antennal lobe in processing olfactory information. Our data suggest that each class of antennal lobe neurons, despite a shared goal and common microcircuit, utilizes different temporal programs to complete synaptic and neurite development. As even closely related neurons of the same microcircuit use different developmental programs, this raised the possibility that each class of neurons may further use different molecular and cellular mechanisms to accomplish synaptic formation. One well-characterized modulator of synapse formation and plasticity is neuronal activity (Cang and Feldheim, 2013; Simi and Studer, 2018). Neurons with increased activity can outcompete other, less active neurons for synaptic partners as well as increase the number of synaptic connections made to other neurons (Miller, 1996). In Drosophila, there is a rich history of studying activity-dependent synaptic development at peripheral neuromuscular synapses (Keshishian et al., 1996); there, changes in activity influence synaptic development in terms of bouton number, ectopic projection, active zone organization, and physiological output (Jarecki and Keshishian, 1995; Schmid et al., 2008; Harris and Littleton, 2015; Akin and Zipursky, 2020; Hazan and Ziv, 2020). The role of activity-dependent synapse organization in the CNS is less well understood, raising questions as to whether all synapses follow similar activity-dependent requirements for development or, as our temporal data might suggest, there is also heterogeneity within molecular and cellular programs that different central neurons use to complete development.

We first sought to determine whether antennal lobe neuron synapse development in ORNs, PNs, or LNs could be regulated by activity-dependent mechanisms. To accomplish this, we either decreased or increased neuronal activity chronically throughout development in ORNs, PNs, or LNs and quantified Brp-Short puncta number and neurite volume in those neurons. Because our developmental analyses showed that ORNs, PNs, and LNs all achieved stable synapse number by 10 d after eclosion (Figs. 2–12), we specifically quantified synapse organization in 10-d-old adult flies. This allowed us to observe how changes in neuronal activity during development influence mature synapse density and to compare our findings between each neuron class. To decrease neuronal activity, we expressed an active tetanus toxin (TeTxLc) in either ORNs, PNs, or LNs that innervate the DA1 glomerulus and compared Brp-Short puncta numbers to neurons that expressed an inactive variant of the toxin as a control (Sweeney et al., 1995). Tetanus toxin lowers neuronal activity by cleaving Synaptobrevin at the synapse, impairing synaptic vesicle fusion and blocking neurotransmission (Sweeney et al., 1995). For male and female ORNs, there was a 29% and 21% decrease, respectively, in Brp-Short puncta number in neurons where active TeTxLc was expressed compared with the inactive control (Figs. 13A–B',G, 14A–B',G). However, neurite volume was unaffected by reduced activity (Figs. 13D–E',G, 14D–E',G). PNs also exhibited a significant decrease (by 60% for males and 36% for females) in Brp-Short puncta number following blocked synaptic transmission (Figs. 13H–I',N, 14H–I',N). Unlike ORNs, however, neurite volume was also significantly decreased in PNs expressing TeTxLc (Figs. 13K–L',N, 14K–L',N). Finally, we observed that TeTxLc expression in LNs caused a reduction in Brp-Short puncta (43% in males, 28% in females; Figs. 13O–P',U, 14O–P',U) but did not influence neurite volume (Figs. 13R–S',U, 14R–S',U). These data suggest that correct synapse addition requires a baseline level of neuronal activity, regardless of antennal lobe neuron class. However, PN neurite outgrowth appears to be activity dependent, while ORN and LN neurite outgrowth seems to be activity independent.

• Download figure
• Open in new tab
• Download powerpoint

Figure 13.

Normal neuronal activity is required for normal synaptic development in male ORNs, PNs, and LNs of the antennal lobe. A–F', Representative confocal image stacks of ORNs of the DA1 glomerulus expressing either Brp-Short-mStraw (A–C) or a membrane-tagged GFP (D–F) and either an inactive (A, D) variant of tetanus toxin light-chain (TeTxLc), an active (B, E) variant of TeTxLc, or two copies of EKO (C, F). Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). G, Quantification of Brp-Short puncta number and neurite volume between male DA1 ORNs expressing either active or inactive TeTxLc or 2xEKO. Expression of either active tetanus toxin or 2xEKO reduces the number of synaptic puncta without altering neurite volume. H–M', Representative image stacks of 10-d-old DA1 PNs (dashed white lines) expressing Brp-Short-GFP (H–J) or membrane-tagged tdTomato (K–M) and either inactive TeTxLc (H, K), active TeTxLc (I, L), or 2xEKO (J, M). Brains were stained with antibodies against tdTomato (green) or GFP (red) and N-cadherin (blue). N, Quantification of Brp-Short puncta number and neurite volume between male DA1 PNs expressing TeTxLc dead, TeTxLc active, or 2xEKO. Decreasing neuronal activity in PNs results in fewer synaptic puncta and reduced neurite volume. O–T', Representative image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (O–Q) or membrane-tagged GFP (R–T) as well as inactive TeTxLc (O, R), active TeTxLc (P, S), or 2xEKO (Q, T). Brains were stained with antibodies as in A–F'. U, Quantification of Brp-Short puncta number and neurite volume between male DA1 LNs expressing TeTxLc dead, TeTxLc active, or 2xEKO. Decreasing neuronal activity in LNs decreased synaptic puncta but did not affect neurite volume. For each experimental group, n ≥ 8 glomeruli from 4 brains. ***p < 0.001, ns = not significant. Scale bar = 5 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 14.

Decreasing neuronal activity in female antennal lobe neurons of the DA1 glomerulus. A–F', Representative confocal image stacks of ORNs of the DA1 glomerulus expressing either Brp-Short-mStraw (A–C) or a membrane-tagged GFP (D–F) and either an inactive variant of TeTxLc (A, D), an active variant of TeTxLc (B, E), or 2xEKO (C, F). Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). G, Quantification of Brp-Short puncta number and neurite volume between female DA1 ORNs expressing either active or inactive TeTxLc or 2xEKO. Expression of active tetanus toxin or 2xEKO reduces the number of synaptic puncta without altering neurite volume. H–M', Representative image stacks of 10-d-old DA1 PNs (dashed white lines) expressing Brp-Short-GFP (H–J) or membrane-tagged tdTomato (K–M) and either inactive TeTxLc (H, K), active TeTxLc (I, L), or 2xEKO (J, M). Brains were stained with antibodies against tdTomato (green) or GFP (red) and N-cadherin (blue). N, Quantification of Brp-Short puncta number and neurite volume in female DA1 PNs expressing TeTxLc dead, TeTxLc active, or 2xEKO. Decreasing neuronal activity in PNs results in fewer synaptic puncta and reduced neurite volume. O–T', Representative image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (O–Q) or membrane-tagged GFP (R–T) as well as inactive TeTxLc (O, R), active TeTxLc (P, S), or 2xEKO (Q, T). Brains were stained with antibodies as in A–F'. U, Quantification of Brp-Short puncta number and neurite volume between female DA1 LNs expressing inactive TeTxLc, TeTxLc active, or 2xEKO. Decreasing synaptic transmission in LNs decreased synaptic puncta but did not affect neurite volume. Decreasing excitability in LNs also decreased synaptic puncta as well as neurite volume. For each experimental group, n ≥ 10 glomeruli from 5 brains. *p < 0.1, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm.

We also sought to examine the consequences of reduced neuronal activity via a nonsynaptic method on synapse number. To do so, we expressed two copies of the transgenic Electrical Knock-out channel (2xEKO; White et al., 2001) in each olfactory neuron class. EKO is a modified Drosophila Shaker K⁺ channel that decreases neuronal excitability by activating at a low voltage (∼–60 mV) for long periods of time, resulting in K⁺ influx and subsequent prevention of excitation (White et al., 2001). Similarly to TeTxLc, when 2xEKO was expressed in male and female ORNs, there was a 40% and 39% decrease in puncta number, respectively (Figs. 13C,C',G, 14C,C',G) without any change in neurite volume (Figs. 13F,F',G, 14F,F',G). In PNs, decreasing neuronal excitability resulted in a reduction of Brp-Short puncta number (20% for males and 31% for females; Figs. 13J,J',N, 14J,J',N) as well as a significant decrease in neurite volume (Figs. 13M,M',N, 14M,M',N). Expression of 2xEKO in LNs also decreased the number of Brp-Short puncta similarly to blocking neurotransmission with TeTxLc (50% decrease in males and 25% decrease in females; Figs. 13Q,Q',U, 14Q,Q',U). However, decreasing excitability in female LNs reduced neurite volume by 25%, but did not affect male neurite volume (Figs. 13T,T',U, 14T,T',U). These results further indicate that synapse addition is dependent on a baseline level of neuron activity across all antennal lobe neuron classes, but that neurite growth is activity dependent in PNs and largely activity independent in ORN and LN classes. This suggests that different activity-related mechanisms exist to control specific aspects of synapse organization in different classes of cells.

Increasing neuronal activity increases ORN but not PN or LN synapse formation

Neuronal activity regulates synapse formation and the strength of connections within neural circuits (Cang and Feldheim, 2013; Vonhoff and Keshishian, 2017; Simi and Studer, 2018; Pan and Monje, 2020). We established that reduced neural activity results in fewer synapses in all classes of neurons observed and sought next to determine whether the reciprocal relationship was true for antennal lobe neurons: does increased electrical activity result in additional synapses? To increase neuronal activity in DA1 ORNs, PNs, and LNs, we expressed the bacterial sodium channel NaChBac, which increases electrical activity by enhancing the inward Na⁺ current into the cells, resulting in additional action potentials and neuronal firing (Ren et al., 2001; Nitabach et al., 2006). We observed that NaChBac expression in male or female DA1 ORNs (Figs. 15A–D', 16A–D') led to a 49% or 38% increase, respectively, in Brp-Short puncta (Figs. 15E, 16E) but did not significantly alter neurite volume (Figs. 15E, 16E). This suggests a direct relationship between cell-autonomous neuronal activity and ORN synapse formation. Intriguingly, we found that this relationship did not exist in PNs: NaChBac expression in DA1 PNs had no effect on either Brp-Short puncta number (Figs. 15F–G',J, 16F–G',J) or neurite volume (Figs. 15H–I',J, 16H–I',J) in males or females. This suggests that PN synapse formation is resistant to increases in electrical activity, but sensitive to reduced neuronal firing, indicating partial activity dependence. Finally, NaChBac expression in LNs exhibited sexually dimorphic changes in synapse number. Increased LN activity in males had no effect on Brp-Short puncta number (Fig. 15K–L',O) or neurite volume (Fig. 15M–O) but significantly reduced Brp-Short puncta (Fig. 16K–L',O) in females by 19% without altering neurite volume (Fig. 16M–O). This indicates a number of aspects about the relationship between LN synapse formation and neural activity: (1) that there is a sexual dimorphism between males and females in the activity dependence of synapse formation; (2) in females, there is likely to be an optimal electrical activity that promotes synapse formation whereby any alteration (increase or decrease) results in impaired synapse formation; and (3) in males, there is likely to be some level of independence from increased activity in regulating synapse formation but sensitivity to reduced activity in regulating synapse formation (similar to PNs but in contrast to ORNs). Collectively, these results demonstrate that each class of neuron is influenced differently by increased neuronal activity and that the rules that govern synapse development are not the same for each class of neuron.

• Download figure
• Open in new tab
• Download powerpoint

Figure 15.

Increased neuronal activity in male ORNs, but not PNs or LNs, results in an increase in synapse number. A–D', Representative confocal image stacks of 10-d-old DA1 ORNs expressing either Brp-Short-mStraw (A, B) or membrane-tagged GFP (C, D) in control flies (A, C) or in flies expressing the NaChBac (B, D) construct to increase neuronal activity. Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). E, Quantification of Brp-Short puncta and neurite volume in control and NaChBac expressing male DA1 ORNs. Increasing neuronal activity increased the number of synaptic puncta without affecting neurite volume. F–I', Representative image stacks of 10-d-old DA1 PNs (dashed white lines) expressing Brp-Short-GFP (F, G) or membrane-tagged tdTomato (H, I) in control (F, H) or NaChBac expressing (G, I) flies. Brains were stained with antibodies against tdTomato (green) or GFP (red) as well as N-cadherin (blue). J, Quantification of Brp-Short puncta and neurite volume for control male DA1 PNs or those expressing NaChBac. Increasing neuronal activity in PNs did not affect synaptic puncta number or neurite volume. K–N', Representative image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (K, L) or membrane-tagged GFP (M, N) in control (K, M) or NaChBac expressing (L, N) flies with antibody staining as in A–D'. O, Quantification of Brp-Short puncta and neurite volume for male DA1 LNs. Increasing excitability did not affect puncta number or neurite volume. For each experimental group, n ≥ 11 glomeruli from 6 brains. ***p < 0.001, ns = not significant. Scale bar = 5 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 16.

Increasing neuronal activity in female antennal lobe neurons of the DA1 glomerulus. A–D', Representative confocal image stacks of 10-d-old DA1 ORNs expressing either Brp-Short-mStraw (A, B) or membrane-tagged GFP (C, D) in control flies (A, C) or in flies expressing a NaChBac (B, D) transgene to increase neuronal activity. Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). E, Quantification of Brp-Short puncta and neurite volume in control and NaChBac expressing female DA1 ORNs. Increasing neuronal activity increased the number of synaptic puncta without affecting neurite volume. F–I', Representative image stacks of 10-d-old DA1 PNs (dashed white lines) expressing Brp-Short-GFP (F, G) or membrane-tagged tdTomato (H, I) in control (F, H) or NaChBac expressing (G, I) flies. Brains were stained with antibodies against tdTomato (green) or GFP (red) as well as N-cadherin (blue). J, Quantification of Brp-Short puncta and neurite volume for control female DA1 PNs or those expressing NaChBac. Increasing neuronal activity in PNs did not affect synaptic puncta number or neurite volume. K–N', Representative image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (K, L) or membrane-tagged GFP (M, N) in control (K, M) or NaChBac expressing (L, N) flies with antibody staining as in A–D'. O, Quantification of Brp-Short puncta and neurite volume for female DA1 LNs. Increasing excitability slightly decreased puncta number without affecting neurite volume. For each experimental group, n ≥ 12 glomeruli from 6 brains. *p < 0.1, ***p < 0.001, ns = not significant. Scale bar = 5 μm.

Altering kinase activity affects synapse number differently in ORNs compared with PNs and LNs

Diverse signal transduction and molecular pathways are responsible for translating changes in neuronal activity to cellular modifications in neurons, including altered neurite volume, synapse number, and gene expression (Chiang et al., 2009; Heinz and Bloodgood, 2020; Faust et al., 2021; Lee and Fields, 2021). Our data suggested that perturbing electrical activity had both shared and differential effects on synapse formation across distinct neuronal class (ORN vs PN vs LN): reduced electrical activity reduced synapse number in all classes, but increased electrical activity affected synapse number in some (ORN) but not all (PN and LN) neuronal classes. We next sought to identify a molecular determinant that could underlie at least some of the changes we observed in activity-related synapse number. Shaggy (sgg), the Drosophila homolog of the mammalian kinase GSK-3β, functions with neuronal activity to mediate neuronal stability in the CNS (Chiang et al., 2009) and at the NMJ to regulate bouton formation and microtubule organization (Franco et al., 2004; Ataman et al., 2008; Miech et al., 2008). However, it is unclear whether kinase activity influences synapse formation in addition to neuronal stability. In ORNs, there is an inverse relationship between neuronal activity and sgg kinase levels, whereby increased neuronal activity leads to decreased kinase activity, promoting axon stability, while decreased neuronal activity leads to increased kinase levels and axon degeneration (Chiang et al., 2009). Our data indicates that ORN, PN, and LN synapse number are all reduced when neuronal activity is decreased (Figs. 13 and 14). Thus, we hypothesized that if Shaggy was the relevant downstream kinase to mediate these changes in synapse number (and given the inverse relationship between neuronal activity and kinase levels), overexpression of a constitutively active sgg (sgg-CA) would result in a similar reduction in synapse number. To test this, we expressed a constitutively active variant of sgg (Sgg-CA; Bourouis, 2002) in all three classes of cells to increase kinase activity along with Brp-Short or mCD8-GFP to quantify synaptic puncta number and neurite volume. Consistent with this hypothesis, Sgg-CA expression in ORNs or LNs that innervate the DA1 glomerulus both similarly resulted in a cell autonomous decrease in Brp-Short puncta number in male (Fig. 17A–B',E,F–G',J) and female (Fig. 18A–B',E,F–G',J) brains. In male and female ORNs, neurite volume was unaffected (Figs. 17C–E, 18C–E), while in LNs, there was a 28% decrease in neurite volume in males while female LN neurite volume was unaffected (Figs. 17H–J, 18H–J). We were unable to assess the effects of Sgg-CA expression on Brp puncta number in PNs as expression via Mz19-GAL4 was lethal. This is likely because of constitutive developmental expression causing lethality, as previously observed (Berdnik et al., 2006). This ORN and LN data, however, supports the hypothesis that electrical activity may signal through the GSK-3β homolog Shaggy to regulate olfactory synapse number across all observed classes of antennal lobe neurons. However, as male LN neurite volume is affected by Sgg activity but not neuronal activity (and vice versa for female LN neurite volume), there are also likely Sgg-dependent, activity-independent and Sgg-independent, activity-dependent phenomena that regulate neuronal growth.

• Download figure
• Open in new tab
• Download powerpoint

Figure 17.

Increased kinase activity decreases male ORN and LN synapse number. A–D', Representative confocal image stacks of 10-d-old ORNs of the DA1 glomerulus expressing Brp-Short-mStraw (A, B) or membrane-tagged GFP (C, D) in control flies (A, C) or in flies expressing a constitutively active (B, D) variant of Shaggy (GSK3β) to increase overall kinase activity. Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). E, Quantification of Brp-Short synaptic puncta and membrane GFP volume in control and Sgg-CA expressing male DA1 ORNs. Increasing kinase activity in ORNs caused a significant decrease in synaptic puncta but did not affect neurite volume. F–I', Representative confocal image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (F, G) or membrane-tagged GFP (H, I) in control flies (F, H) or in flies expressing constitutively active (G, I) Shaggy and stained with antibodies as in A–D'. J, Quantification of Brp-Short puncta and neurite volume for control and Sgg-CA expressing male DA1 LNs. Increased activity decreased both puncta number and neurite volume in male DA1 LNs. For each experimental group, n ≥ 8 glomeruli from 4 brains. *p < 0.1, **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 18.

Increasing kinase activity in female olfactory neurons of the DA1 glomerulus only decreases synapse number. A–D', Representative confocal image stacks of 10-d-old female ORNs of the DA1 glomerulus expressing Brp-Short-mStraw (A, B) or membrane-tagged GFP (C, D) in control flies (A, C) or in flies expressing a constitutively active variant of Shaggy (GSK3β; C, D) to increase overall kinase activity. Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). E, Quantification of Brp-Short synaptic puncta and membrane GFP volume in female DA1 ORNs. Increasing kinase activity in ORNs caused a significant decrease in synaptic puncta but did not affect neurite volume. F–I', Representative image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-GFP (F, G) or membrane-tagged tdTomato (H, I) in control flies (F, H) or in flies expressing constitutively active Shaggy (G, I) and stained with antibodies as in A–D'. J, Quantification of Brp-Short puncta and neurite volume for female DA1 LNs. Increased activity decreased puncta number without affecting neurite volume in female LNs. For each experimental group, n ≥ 4 glomeruli from 2 brains. *p < 0.1, ***p < 0.001, ns = not significant. Scale bar = 5 μm.

We similarly assessed the relationship between neuronal activity and kinase levels by expressing a dominant negative sgg transgene (sgg-DN) to decrease kinase levels (Franco et al., 2004). Because of the inverse relationship between the two activity levels, we hypothesized that decreasing Shaggy would phenocopy increasing neuronal activity. Intriguingly, however, decreasing kinase activity in ORNs decreased Brp-Short puncta in male (Fig. 19A–B',E) and female (Fig. 20A–B',E) brains by 27% and 30%, respectively. However, decreasing sgg levels in PNs and LNs did not affect Brp-Short puncta in males (Fig. 19F–G',J,K–L',O) and females (Fig. 20F–G',J,K–L',O), matching the effects of increased excitability in these neuron classes. Decreased kinase activity increased neurite volume 26% in male PNs but did not affect ORNs or LNs (Fig. 19C–E,H–J,M–O) nor any class of neuron in females (Fig. 20C–E,H–J,M–O). This suggests additional heterogeneity in the molecular cascades controlling neuronal and synapse development as well as sexual dimorphism. Overall, our data shows that development is not a uniform process between neuronal classes, even in neurons belonging to the same microcircuit with the same general processing goal, and that each class of neuron forms synapses using its own unique temporal, activity-dependent, and molecular programs. Thus, considerable temporal and molecular heterogeneity exists in the development of central synaptic organization.

• Download figure
• Open in new tab
• Download powerpoint

Figure 19.

Decreasing kinase activity in male olfactory neurons decreases synapse number in ORNs, but not PNs or LNs. A–D', Representative confocal image stacks of 10 d old male ORNs of the DA1 glomerulus expressing Brp-Short-mStraw (A, B) or membrane-tagged GFP (C, D) in control flies (A, C) or in flies expressing a dominant negative (B, D) variant of Shaggy (GSK3β) to decrease overall kinase activity. Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). E, Quantification of Brp-Short synaptic puncta and membrane GFP volume in male DA1 ORNs of each group from A–D'. Decreasing kinase activity in ORNs caused a significant decrease in synaptic puncta but did not affect neurite volume. F–I', Representative image stacks of 10-d-old male DA1 PNs (dashed white lines) expressing Brp-Short-GFP (F, G) or membrane-tagged tdTomato (H, I) in control flies (F, H) or in flies expressing dominant negative (G, I) Shaggy. Brains were stained with antibodies against tdTomato (green) or GFP (red) and N-cadherin (blue). J, Quantification of Brp-Short puncta and neurite volume in male DA1 PNs from F–I'. Decreasing kinase activity did not influence puncta number in PNs but did increase neurite volume. K–N', Representative confocal image stacks of 10-d-old male DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (K, L) or membrane-tagged GFP (M, N) in control flies (K, M) or in flies expressing dominant negative (L, N) Shaggy and stained with antibodies as in A–D'. O, Quantification of Brp-Short puncta and neurite volume for male DA1 LNs of the groups described in K–N'. Decreased kinase activity did not affect synaptic puncta or neurite volume in LNs. For each experimental group, n ≥ 14 glomeruli from 7 brains. **p < 0.01, ***p < 0.001, ns = not significant. Scale bar = 5 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 20.

Decreasing kinase activity in female olfactory neurons of the DA1 glomerulus. A–D', Representative confocal image stacks of 10-d-old female ORNs of the DA1 glomerulus expressing Brp-Short-mStraw (A, B) or membrane-tagged GFP (C, D) in control flies (A, C) or in flies expressing a dominant negative variant of Shaggy (GSK3β; B, D) to decrease overall kinase activity. Brains were stained with antibodies against mStraw (red) or GFP (green) and N-cadherin (blue). E, Quantification of Brp-Short synaptic puncta and membrane GFP volume in female DA1 ORNs. Decreasing kinase activity in ORNs caused a significant decrease in synaptic puncta but did not affect neurite volume. F–I', Representative image stacks of 10-d-old DA1 PNs (dashed white lines) expressing Brp-Short-GFP (F, G) or membrane-tagged tdTomato (H, I) in control flies (F, H) or in flies expressing dominant negative (G, I) Shaggy. Brains were stained with antibodies against tdTomato (green) or GFP (red) and N-cadherin (blue). J, Quantification of Brp-Short puncta and neurite volume in female DA1 PNs. Decreasing kinase activity did not alter puncta number or neurite volume in PNs. K–N', Representative confocal image stacks of 10-d-old DA1 LNs (dashed white lines) expressing Brp-Short-mStraw (K, L) or membrane-tagged GFP (M, N) in control flies (K, M) or in flies expressing dominant negative Shaggy (L, N) and stained with antibodies as in A–D'. O, Quantification of Brp-Short puncta and neurite volume for female DA1 LNs. Decreased kinase activity did not affect synaptic puncta or neurite volume in LNs. For each experimental group, n ≥ 10 glomeruli from 5 brains. ***p < 0.001, ns = not significant. Scale bar = 5 μm.