Adult expression of the cell adhesion protein Fasciclin 3 is required for the maintenance of adult olfactory interneurons Free

The proper functioning of the nervous system is determined by the establishment and maintenance of intricate connections between neurons. During nervous system development, billions of neuronal axons navigate to their targets in response to axon guidance cues in the extracellular environment (Tessier-Lavigne and Goodman, 1996). This process involves neuronal outgrowth, navigation and pathfinding that relies on a highly dynamic, fan-shaped and actin-rich structure located at the tip of axons and dendrites called a growth cone (Cajal, 1890). For over 100 years, neuroscientists have researched the nature of the guidance cues underlying these processes and developed an understanding of how these cues signal spatial information to growth cones to direct neural circuit formation (Kolodkin and Tessier-Lavigne, 2011). Many families of guidance cues have been characterized including Semaphorins, Netrins, Slits, Ephrins, morphogens, cell adhesion molecules (CAMs) and cytoskeletal-associated proteins (Yaron and Zheng, 2007).

Numerous studies have supported the idea that CAMs can guide axons by promoting adhesion, stimulating neuronal outgrowth, and functioning as signaling molecules in heterophilic or homophilic combinations (Moreland and Poulain, 2022; Pollerberg et al., 2013; Rutishauser, 1993). The immunoglobulin-like cell adhesion molecules (Ig-CAMs) that belong to the immunoglobulin superfamily and cadherins are the two major classes of CAMs. A role for CAMs in regulating axonal fasciculation and motor neuron guidance was first documented for the Ig-CAM Fasciclin 2 (Fas2) (Harrelson and Goodman, 1988; Lin and Goodman, 1994; Lin et al., 1994). Fas2 is homologous to the neural cell adhesion molecule (N-CAM) in vertebrates and is required for selective axon fasciculation in Drosophila. Ig-CAMs are also known to interact with signal transduction pathways during axonal guidance. For example, Fasciclin 1 (Fas1) genetically interacts with Abelson tyrosine kinase (Abl) with Fas1/Abl double mutants exhibiting commissural axon pathfinding defects (Elkins et al., 1990). Fasciclin 3 (Fas3), another Ig-CAM in Drosophila, is expressed by RP3 motor neurons and their synaptic targets (muscles 6 and 7) during embryonic development (Patel et al., 1987; Snow et al., 1989). Target recognition between the motor neuron growth cones and muscles is mediated through Fas3-dependent homophilic interactions. In addition, RP3 motor neurons form synapses on muscles that ectopically express Fas3, confirming that Fas3 acts as a synaptic recognition cue for these growth cones (Chiba et al., 1995). Additionally, when the Fas3-negative aCC and SNa motor neuron growth cones ectopically express Fas3, they innervate the Fas3-expressing muscles as alternative targets (Kose et al., 1997). However, motor neurons can correctly innervate the target muscles 6 and 7 in Fas3-null mutants, suggesting redundancy in targeting mechanisms. These studies provide evidence for Fas3-dependent homophilic synaptic target recognition in the precise formation of neural circuits.

Once neural circuits are assembled during development, it is presumed that a distinct set of molecular programs is likely required to maintain their connectivity throughout the lifetime of the organism. Nonetheless, many mature neurons continue to express axon guidance proteins long after they have reached their targets and made functional connections. Indeed, the modENCODE high-throughput RNA-seq data confirms that axon guidance cues, including Semaphorins, Plexins and CAMs (including Fas3) continue to be expressed in the adult Drosophila nervous system (Graveley et al., 2011). We have recently shown that members of the Semaphorin family of guidance cues have crucial functions in the mature nervous system and are essential for maintaining neuronal survival, adult motility and longevity (Vaikakkara Chithran et al., 2023). Although this previous work has shown that the expression of Fas3 in adults is also crucial for survival, the role of Fas3 in the maintenance of adult neural circuits has not been examined. In this study, we utilize the well-characterized Drosophila olfactory circuit to explore the importance of stable expression of Fas3 for neural circuit maintenance in adults.

In Drosophila, the olfactory lobe (OL) is composed of 56 neuropil regions, known as glomeruli (Laissue et al., 1999; Tanaka et al., 2012), each of which receives input from a specific population of olfactory sensory neurons (OSNs) that typically express a single type of odorant receptor (Couto et al., 2005; Fishilevich and Vosshall, 2005). Within each glomerulus, OSNs establish synapses with projection neurons (PNs), which serve as output neurons of the OL and can be compared to the mitral/tufted cells in mammals (Imai et al., 2010; Lledo et al., 2005). OSNs also make connections with local interneurons (LNs) that have arborizations confined within the OL, similar to interneurons found in the olfactory bulb of mammals, such as granule and periglomerular cells (Lledo et al., 2005; Stocker et al., 1990; Tanaka et al., 2012). Furthermore, PNs establish feedback connections with LNs (Liu and Wilson, 2013; Sudhakaran et al., 2012; Tanaka et al., 2009). LNs play multiple roles in shaping the output of the OL. One important function of LNs is their involvement in olfactory habituation, a phenomenon characterized by a reduced behavioral response to repeated or continuous exposure to an odorant (Twick et al., 2014). This reduced response is attributed to increased inhibitory input from LNs to odor-selective PNs (Das et al., 2011; Larkin et al., 2010; Sadanandappa et al., 2013; Sudhakaran et al., 2012). Interestingly, a genetic screen identified mutations in several CAMs, including Fas3, which disrupted olfactory habituation (Eddison et al., 2012), suggesting an important role in the proper development of the olfactory circuit. However, the role of Fas3 in the normal functioning of olfactory neurons in adults has not been examined. In this study, we present evidence demonstrating the essential role of continuous Fas3 expression in maintaining the olfactory circuitry in adult flies. Specifically, we reveal that Fas3 promotes cell survival in LNs, suggesting a crucial function for CAMs in the maintenance of adult neurons.

Using a Fas3::GFP protein trap allele, we examined the expression and localization of Fas3 in the adult Drosophila central nervous system (CNS). Fas3 is prominently localized in the OLs (yellow box in Fig. 1A), with lower levels observed in the optic lobes (white box in Fig. 1A), suboesophageal ganglion (yellow arrow in Fig. 1A), and a subset of neuronal populations in the ventral nerve cord (white arrows in Fig. 1A). This expression and localization were also confirmed by Fas3 immunoreactivity (Fig. 1B) and Fas3-Gal4-driven expression of nuclear GFP (Fig. 1C). Within the olfactory circuit, it is likely that Fas3 is expressed in a subset of neurons whose cell bodies lie adjacent the OL (as observed by the UAS-nuclear GFP expression driven by Fas3-Gal4 in Fig. 1C) and project their dendrites into the OL (as seen by the strong expression of Fas3-GFP and anti-Fas3 labelling in the glomeruli of the OL in Fig. 1A,B).

To examine whether Fas3 expression in adults is essential for survival, we used the TARGET system (Fas3-Gal4 with tub-Gal80^ts) to knockdown the expression of Fas3 only in adult Drosophila in all Fas3-expressing cells (Fig. 2A) and monitored the survival of adult flies daily until day 22 post eclosion. We picked male flies for all the experiments. In the survival assay, we employed two controls; an isogenic host strain as a genetic background control for the GD RNAi library and a UAS-dsRNA-GFP control, both of which yielded similar results. On day 17, there was a significant reduction in survival in the Fas3-knockdown flies, with only 87% of the knockdown flies surviving compared to 97% of the control flies. This impact on survival increased with age; by day 22, the survival rate of Fas3 knockdown flies dropped to 47%, whereas 97% of the control flies survived (Fig. 2B), confirming that continued expression of Fas3 in the adult is essential for longevity.

Owing to high Fas3 levels in the adult OLs (as shown in the yellow box in Fig. 1A) and the well-characterized circuitry of the adult Drosophila olfactory system (Fig. 3A), we examined the impact of Fas3 knockdown on the maintenance of this circuit. Among the three neuronal populations of the olfactory circuit (OSNs, PNs and LNs), the OSNs have their cell bodies located in the periphery and the PNs send their axonal projections into higher brain regions. However, we observed Fas3-positive cell bodies right outside the OL (Fig. 1C), and after extensive examination we could not identify any projections directed to any higher brain regions (Fig. 1A,B) suggesting that the Fas3-expressing neurons only send local projections and are likely to be LNs.

To confirm Fas3 expression in LNs, we examined colabeling of Fas3-nlacZ reporter (Fig. 3B) with UAS-nGFP driven by one of three LN-Gal4 lines, GH298-Gal4 (Fig. 3C–C‴), NP3056-Gal4 (Fig. 3D–D‴) or 189Y-Gal4 (Fig. 3E–E‴) (Chou et al., 2010; Liou et al., 2018). Almost all GH298-Gal4 neurons were colabeled by Fas3-lacZ (Fig. 3C‴). Also, 85% of NP3056-Gal4 neurons (Fig. 3D‴) and 50% of 189Y-Gal4 neurons (Fig. 3E‴) were colabeled with Fas3-lacZ. This confirmed that Fas3 is expressed in subsets of olfactory LNs.

In order to examine the importance of Fas3 expression in neural circuit maintenance, we decided to examine the impact of Fas3 knockdown specifically on the olfactory circuit. As we observed a significant reduction in the survival of adult flies from day 17 post eclosion (Fig. 2B), we focused our analysis on circuit structure before this time point. Specifically, we examined LNs and their projections at day 14, 2 weeks following initiation of Fas3 knockdown. Using Fas3-Gal4, we drove UAS-dsRNA-Fas3 with UAS-CD8::GFP or, as a control, we drove UAS-CD8::GFP alone. In the controls, we observed the normal distribution of CD8::GFP neuronal cell bodies located immediately adjacent to the OLs (indicated by white arrows in Fig. 4A). In contrast, in Fas3 knockdowns, we observed a striking, significant loss of these neurons (indicated by white arrows in Fig. 4B). This neuronal loss phenotype was confirmed by measuring GFP intensity in control and Fas3-knockdown flies (Fig. 4C) and was observed with multiple dsRNA-Fas3 lines (Fig. S1). Fas3 antibody staining confirmed that Fas3 was severely reduced by day 14 (compare Fig. 4A′,B′). Next, we observed an increase in anti-Caspase 3 immunolabeling as early as day 7, especially in the region surrounding the OLs (indicated by white arrows in Fig. 4A″,B″), which was confirmed by measurements of Caspase3 immunolabeling (Fig. 4D). This suggested that loss of olfactory circuit neurons observed after Fas3 knockdown is due to apoptotic cell death. It is interesting to note that the region of increased Caspase3 appears to extend beyond Fas3 knockdown, suggesting possible apoptosis in other cells immediately surrounding the Fas3-expressing neurons. Further experiments will be required to examine the possible mechanisms underlying this cell death.

To further confirm that Fas3 knockdown-mediated cell death occurs in subsets of LNs, we used LN-Gal4 drivers to restrict Fas3 knockdown to small groups of olfactory interneurons. We conducted this experiment using multiple UAS-dsRNA-Fas3 lines. Owing to the differences in the genetic backgrounds of the UAS-dsRNA-Fas3 lines, we used a panel of relevant controls from each respective RNAi library. These are detailed in Tables S2, S3. The strongest NP3056-Gal4-mediated knockdown of Fas3 resulted in a ∼50% loss of NP3056-Gal4 expressing neurons by day 14 (marked by white arrows in Fig. 5A–G); from 27.15±0.6 neurons in controls to 13.35±0.8 in Fas3 knockdowns (mean±s.e.m.; P≤0.0001) (Fig. 5H,I). A similar cell autonomous cell death phenotype was also observed when Fas3 was knocked down using the other LN-GAL4 drivers [GH298-Gal4 (Fig. S2) and 189Y-Gal4 (Fig. S3)]. Taken together, our data confirm that Fas3 knockdown results in cell autonomous death of LNs.

To better understand the time course of neuronal loss after Fas3 knockdown, we used the pan-neuronal driver, elav-Gal4, to knockdown Fas3. The knockdown was again restricted to the adults using the TARGET system. Here, we examined neuronal numbers at days 3, 5, 7, 10 and 14 after the initiation of Fas3 knockdown (Fig. 6A1–A5,B1–B5). Pan-neuronal knockdown of Fas3 recapitulated the neuronal death we observed when knocking down Fas3 in subsets of olfactory neurons. After pan-neuronal Fas3 knockdown, we started to observe the loss of Fas3 immunoreactivity by day 7, with significant death of olfactory circuit neurons evident by day 10 (Fig. 6A3,A4,B3,B4), which continued through day 14 (Fig. 6A5,B5). These results corroborate our observations using the Fas3-Gal4 and LN-Gal4 drivers and add a time course to the observed cell death.

Given that we observed significantly increased cleaved Caspase3 immunostaining in Fas3 knockdowns (Fig. 4A″,B″) which is an indicator of apoptosis, we tested whether neuronal death can be rescued by blocking the apoptotic pathway. Using elav-Gal4, we knocked down Fas3 in the presence of overexpressed p35 (by driving the baculovirus protein p35 using UAS-p35), an anti-apoptotic protein. We also used UAS-nGFP to count neuronal numbers on day 14. We chose the area of LN neurons around each OL as the region of interest (ROI) for quantification (indicated by the yellow box in Fig. 6C). Notably, co-expressing UAS-p35 with UAS-dsRNA-Fas3 completely rescued the neuronal death phenotype (Fig. 6C,D). The number of neurons (represented by nuclear GFP) in Fas3 knockdowns was reduced by ∼50% compared to that in the control and p35 rescue (Fig. 6D). These results confirm that Fas3 knockdown-mediated neuronal death can be rescued by blocking the apoptotic pathway.

Previous studies have provided insight into the emergence and integration of neurons into the Drosophila olfactory circuit; however, very little is understood regarding the maintenance of these neurons in the adult. This study begins to reveal the importance of the stable expression of Fas3, an axon guidance cell adhesion protein, in the maintenance of the adult olfactory circuit. We have demonstrated that adult-specific knockdown of Fas3 leads to the death of local olfactory interneurons, which can be rescued by expressing the anti-apoptotic protein p35. Although the dynamic incorporation and function of local interneurons into the rodent olfactory circuit have been examined before (Adam and Mizrahi, 2010; Belluzzi et al., 2003), this is the first study to reveal a crucial role for a guidance protein in the maintenance of Drosophila interneuron survival. Given that we used the TARGET system to restrict the knockdown to the adult, the observed phenotypes are not a result of developmental defects but are due to defects in the maintenance of the adult circuit. To address the potential issue of off-target effects while using RNAi (i.e. knockdown of the expression of multiple genes by the same dsRNA or shRNA), we employed several approaches. First, we picked UAS-dsRNA or -shRNA lines targeting Fas3 that had been used in previous studies where no off-target effects were reported (Hu et al., 2013). In addition, the observed phenotypes were confirmed using five different Gal4 drivers (Fas3-Gal4, elav-Gal4, NP3056-Gal4, GH298-Gal4 and 189Y-Gal4) and multiple different UAS-dsRNA/-shRNA lines and controls (Tables S2,S3). This includes UAS-dsRNA lines that are integrated into specific attP sites, ensuring that experimental and control UAS transgenes are controlled for position effect (Heigwer et al., 2018).

Although there is evidence for intraglomerular plasticity in the adult Drosophila olfactory circuit (Berdnik et al., 2006), this study is the first to examine maintenance of the olfactory circuit at the level of a neuronal survival. Cellular plasticity in the adult olfactory system of several species is attributed to their sensory afferents and to subsets of interneurons that are continuously replaced throughout adulthood (Bayramli et al., 2017; Durante et al., 2020; Fernández-Hernández et al., 2020preprint; Graziadei and Okano, 1979; Monti Graziadei and Graziadei, 1979; Weiler and Farbman, 1997). Sustained neurogenesis and apoptosis of sensory neurons has been detected in the antennae of adult Drosophila, supporting an ongoing turnover in the olfactory system of adult flies (Fernández-Hernández et al., 2020 preprint). In vertebrates, olfactory sensory neuron replacement not only compensates for wear-out processes in the periphery but also contributes to additional glomerular innervations in response to the experience of the organism in adulthood (Jones et al., 2008). In the olfactory epithelium, where mature neurons die and are replaced throughout adult life, tight control over the mechanism of neuronal death is required to avoid tumorigenesis or premature depletion of olfactory sensory neurons. Both in vivo and in vitro studies have shown that olfactory sensory neurons die through an intrinsically programed process, involving caspase activity and mediated by Jun N-terminal kinase (JNK) signaling (Gangadhar et al., 2008).

In adult rodents, the local circuitry of olfactory interneurons is highly dynamic both at the population level and within individual cells. At the population level, a fraction of rodent interneurons undergo continuous replacement, whereas at the single-cell level, their dendritic morphology continues to change even after they reach maturity (Adam and Mizrahi, 2010; Belluzzi et al., 2003). The specific subpopulations of interneurons that either remain stable or undergo changes, as well as the extent of these changes, are still not fully understood. However, multiple studies provide compelling evidence for adult neurogenesis of olfactory interneurons (Kornack and Rakic, 2001; Ming and Song, 2011; Rosselli-Austin and Altman, 1979). Their survival has been shown to be dependent on sensory activity, with odor exposure or olfactory learning promoting neuronal survival (Petreanu and Alvarez-Buylla, 2002; Rochefort et al., 2002). In contrast, there is no evidence for adult neurogenesis of local olfactory interneurons in Drosophila. This rules out the possibility that the loss of Fas3 might be inhibiting the proliferation and differentiation of neural stem cells into local interneurons, thereby resulting in a reduced neuronal number. Moreover, the increased cleaved Caspase3 labeling in the Fas3-knockdown brains and the p35 rescue of neuronal death imply that a caspase-dependent cell death pathway is activated in the LNs upon Fas3 knockdown. Although this could be a newly discovered role for Fas3, there is previous evidence associating cell death with the loss of expression of other cell adhesion molecules, such as integrins (Stupack, 2005), and axon guidance genes, such as Semaphorins and Plexins (Vaikakkara Chithran et al., 2023).

Drosophila local interneurons (LNs) have diverse neurotransmitter profiles, glomerular innervation patterns and odor response properties. Neuronal subpopulations labeled by NP3056-, GH298- and 189-Gal4 drivers show a patchy glomerular innervation pattern in the OL. Each of the LN-Gal4 drivers label a subset of both larval and adult LNs with at least 189-Gal4 and NP3056-Gal4 showing no overlap of expression (Chou et al., 2010). No two neurons have identical innervation patterns, suggesting that these innervation patterns might be established through cell–cell interactions among LNs (Chou et al., 2010). We speculate that the expression of Fas3, a homophilic cell adhesion molecule, is required for maintaining the LN–LN interactions in adults, which in turn is required for maintaining their patchy innervation. A role for CAMs in maintaining axonal morphologies has been demonstrated previously. For example, it has been shown that JNK signaling is required cell autonomously for axon pruning in mushroom body neurons by negatively regulating the plasma membrane localization of Fas2, another cell adhesion molecule of the Ig superfamily (Bornstein et al., 2015). Overexpression of Fas2 or other CAMs such as Fas1, Fas3 or Neuroglian (Nrg) was found to be sufficient to inhibit pruning and high levels of CAMs were observed in unpruned axons. Thus, it is possible that Fas3 knockdown disrupts LN–LN interactions, triggering changes in their axonal and/or dendritic morphologies and glomerular innervation, eventually causing neuronal degeneration and death. The importance of maintaining cell–cell interactions is also evident from the aptly called ‘community effect’ observed among precursor cells during development (Saka et al., 2011). Cells in the developing embryo are in constant communication with their neighbors and this interaction is necessary for them to maintain tissue-specific gene expression and differentiate in a coordinated manner. Loss of signaling from neighbors can cause cellular dedifferentiation and death. Similarly, in the adult brain, the loss of cell–cell communication following Fas3 knockdown might result in neuronal degeneration and death. In addition, Fas3 is known to genetically interact with other cell adhesion genes including armadillo (Greaves et al., 1999), shotgun (Greaves et al., 1999), Toll (Rose and Chiba, 1999) and Nrg (Lu et al., 2014). Ongoing studies will examine whether loss of Fas3 downregulates the expression of other CAMs eventually contributing to cell death.

To maintain survival over the course of their lifetimes, neurons have the ability to adapt certain biological processes to prevent cell death. Cell death pathways are often activated in developing neurons to fine-tune the number of neurons required for the precise formation of neural circuits (Davies, 2003; Kole et al., 2013; Oppenheim, 1991). However, mature neurons promote survival by employing multiple mechanisms to prevent cell death. This shift from permitting cell death to ensuring cellular survival is imperative in mature neurons, as in most cases they must maintain the neuronal circuitry for an entire lifetime of the organism (Benn and Woolf, 2004; Kole et al., 2013). We speculate that maintaining cell–cell interactions by continued expression of guidance cues and cell adhesion proteins, such as Fas3, is a survival mechanism in adult neurons and its loss releases the apoptotic breaks in mature neurons, resulting in cell death. However, the mechanisms downstream of Fas3 loss that result in death require further detailed analysis. Previous studies have discovered several intracellular signaling cascades that are activated upon the homophilic binding of CAMs (Kiryushko et al., 2004). For example, Leukocyte-antigen-related-like (Lar), a protein that resembles CAMs of the Ig superfamily, controls motor axon guidance in Drosophila through protein tyrosine phosphatase activity (Krueger et al., 1996 ). The intracellular region of Fas3 contains two serine residues that are potential sites of phosphorylation mediated by Protein Kinase C and a tyrosine residue that might facilitate the attachment of adapter proteins responsible for internalization into coated pits (Snow et al., 1989). Phosphorylation of CAMs such as integrins, N-CAM and L1 has been extensively studied and recognized as a fundamental event in regulating interactions with cytoplasmic proteins to induce alterations in cell adhesion and signaling (Gahmberg and Grönholm, 2022; Matthias and Horstkorte, 2006; Wong et al., 1996). Furthermore, previous research has suggested a role for coated pits in mediating interactions between neuronal growth cones and neighboring cells (Bastiani and Goodman, 1984). Thus, loss of Fas3 might result in the impairment of cell signaling, endocytosis and protein trafficking, leading to the disruption of cell adhesion and possibly cell death. Further research on intracellular signaling cascades following Fas3 homophilic binding is required to understand how Fas3 maintains cell survival in LNs.

The main goal of this study was to examine whether Fas3 has an essential function in the maintenance of adult neural circuits. Surprisingly, we also found that Fas3 is necessary for the survival of adult Drosophila, which is similar to our previous finding that pan-neuronal knockdown of certain axon guidance genes in adults results in reduced longevity (Vaikakkara Chithran et al., 2023). However, we do not suggest that the loss of LNs after Fas3 knockdown is responsible for the decreased longevity in adult flies. Indeed, the observed reduction in the survival rate after Fas3 knockdown could result from the dysfunction of any number of neuronal populations or non-neuronal tissues that express Fas3 in adult Drosophila. Although we have focused on the role of Fas3 in the adult olfactory circuit in this study, it is important to note that Fas3 is also expressed in the optic lobes, suboesophageal ganglion, ventral nerve cord and non-neuronal tissues. We suggest that Fas3 function in these other neural circuits is more likely to be the cause of the observed reduction in longevity after pan-neuronal knockdown. Examining these additional cellular populations will be crucial to determining where Fas3 is required for adult survival.

Reducing axon guidance gene expression by RNAi in the adult nervous system was carried out using the TARGET system. Gal4 driver line virgin females (see Table S1) were crossed with UAS-dsRNA or -shRNA males (see Table S2) or the appropriate negative control line males (see Table S3). The crosses were set up at 18°C and the progeny (F1) was switched to 29°C immediately after eclosion to repress Gal80^ts and activate the Gal4-UAS system. All analyses were performed on 14-day-old adult male flies unless otherwise stated.

Drosophila melanogaster stocks used in this study were maintained on standard cornmeal food (1 L of fly food medium contained 1 L water; 10 g agar; 30 g each of sucrose, glucose, cornmeal and dry yeast; 0.5 g each of CaCl₂.2H₂O and MgSO₄.7H₂O; 6 ml propionic acid; and 1 g methylparaben dissolved in 10 ml absolute ethanol), at 18, 25 or 29°C in environment rooms set at 70% humidity. The stocks used in this study were received from Bloomington Drosophila Stock Center (BDSC), Vienna Drosophila Resource Center (VDRC) and Kyoto Drosophila Stock Center (DGRC) and are listed in Tables S1–S4.

F1 progeny males were collected immediately after eclosion and switched to 29°C. They were maintained in vials of 10 and their survival was recorded every day for the next 20 days. Flies were flipped into fresh vials every week. Multiple UAS-dsRNA or -shRNA lines and controls were tested and for each line, and three vials containing 10 flies were assayed. All results were analyzed using two-way ANOVA and Tukey's multiple comparison tests using GraphPad Prism.

Adult male Drosophila brains and ventral nerve cords were dissected in ice-cold 1× phosphate-buffered saline (PBS), over a period of 30 min. The tissue was fixed, on poly-lysine coated slides, in 4% PFA at room temperature (RT) for 30 min followed by three times for 5 min washes with PBS containing 0.1% Triton X-100 (PBT). The samples were blocked in PBT containing 5% donkey serum (PBTN) for 2 h at RT, and then incubated in primary antibodies overnight at 4°C. The next day, samples were washed three times for 30 min each time in PBT. They were then blocked in PBTN for 1 h at RT and incubated in secondary antibody for 3 h at RT in the dark. For each experiment, controls and knockdowns were dissected and labeled in parallel. The samples were washed three times for 30 min each time in PBT. They were mounted in Vectashield mounting medium (Vector Laboratories Inc.), coverslipped and sealed with nail polish. Primary antibodies used in this study were as follows: mouse-anti-Fasciclin III 7G10 (DSHB ID: 7G10; 1:400), mouse-anti-β-gal (DSHB ID: 40-1a; 1:50), rabbit-anti-cleaved caspase 3 (Asp175) (Cell Signaling Technology, cat. no. 9661; 1:300). Secondary antibodies used in this study were as follows: donkey-anti-mouse-IgG conjugated to Cy3 (Jackson ImmunoResearch; 1:250), donkey anti-rabbit-IgG conjugated to Cy3 (Jackson ImmunoResearch; 1:250). All images were acquired using the Zeiss LSM 880 Fast Airyscan inverted confocal microscope in a Z-stack. Images were processed using the ‘Airyscan processing’ feature using the Zen-black software. The images were acquired in multiple tiles and stitched post Airyscan processing. The images included in this study are maximum intensity projections of representative images and all images were acquired with the same parameters.

To measure the intensity of GFP (driven by Fas3-Gal4) and caspase3 immunostaining, the Z-stack images were opened in Fiji/ImageJ and processed using Image→Stacks→Z Project→SUM slices. The area shown in representative images (OLs+surrounding regions) was selected as the region of interest (ROI) for measuring the mean grey value. The results were analyzed using unpaired two-tailed t-test with Welch's correction on GraphPad Prism.

In elav-Gal4 driven experiments, the area around the OL was selected as the region of interest and neurons (represented by nlsGFP) were quantified using a MATLAB program kindly provided by Katie Goodwin (University of British Columbia; available from us upon request). In LN-Gal4 driven experiments, the olfactory circuits neurons were quantified manually in a 3D model (using Zen-blue image processing software). The results were analyzed using one-way ANOVA and Tukey's multiple comparison tests using GraphPad Prism.

The authors extend sincere thanks to Katie Goodwin for kindly sharing the MATLAB program for image analysis. The antibodies used in this study were obtained from Developmental Studies Hybridoma Bank (DSHB), Abcam, Cell Signaling Technology and Jackson ImmunoResearch Laboratories Inc. Drosophila melanogaster stocks were received from Bloomington Drosophila Stock Center (BDSC), Vienna Drosophila Resource Center (VDRC) and Kyoto Drosophila Stock Center (DGRC).

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.261759.reviewer-comments.pdf