During nervous system development in both vertebrates and invertebrates, neurons undergo a remodeling process that is necessary for their normal function (Luo and O'Leary, 2005; Neukomm and Freeman, 2014; Schuldiner and Yaron, 2015). Moreover, it is suggested that excess or insufficient remodeling may underlie several neurodegenerative diseases (Cardozo et al., 2019; Courchesne et al., 2007). Glial cells play an essential role in both neuronal remodeling and neurodegenerative diseases. A critical crosstalk between glia and neurons is required in both processes; however, little is known about this intercellular cooperation (Boulanger and Dura, 2022; Neniskyte and Gross, 2017). Orion is a secreted chemokine-like protein, which has been described as a necessary component for neuronal remodeling in different paradigms in Drosophila (Boulanger et al., 2021; Ji et al., 2023; Perron et al., 2023). Orion is a ligand that should interact with its receptor to transform the surrounding glial or epidermal cells into phagocytes for neurite debris elimination. Therefore, the identification of the Orion receptor is of paramount importance to understand the phagocytic clearance of degenerating neurons. Importantly, in the case of the phagocytosis of degenerating dendrites of class IV dendritic arborization (C4da) neurons, it was shown that Orion bridges phosphatidylserine (PS), a conserved “eat-me signal,” and the phagocytic receptor Drpr (draper) (Ji et al., 2023). Drpr is involved in many aspects of neuronal remodeling and debris clearance during development or after injury in Drosophila (Boulanger and Dura, 2022; Freeman et al., 2003; Tasdemir-Yilmaz and Freeman, 2014). We aimed to assess whether Drpr can also be a receptor for Orion in the case of mushroom body (MB) neuronal remodeling. MBs are bilateral and symmetrical brain structures required for learning and memory. During metamorphosis, the MB-larval-specific dendrites and axons of early γ-neurons are pruned and replaced by adult-specific processes necessary for memory (Boulanger et al., 2011; Boulanger and Dura, 2015; Redt-Clouet et al., 2012; Trannoy et al., 2011; Yu and Schuldiner, 2014). We show here that the drpr ^indel3 null allele has a MB-pruning phenotype very similar, if not identical, to that of orion ^null alleles. Furthermore, by using an Orion-CD2 transgene in which Orion is permanently tethered to the surface of the MB axons, we show, first, a high recruitment of glial Drpr inside γ-axon bundles and, second, strong genetic interactions between Orion and glial Drpr. These results strongly suggest that Drpr is the glial receptor for Orion in MB neuronal remodeling.

Our laboratory published that Drpr was not, or at least not the sole, Orion receptor principally based on the different penetrance of the MB-pruning phenotype in orion ^null (either orion ¹ and orion ^ΔC ) and drpr ^Δ5 (Boulanger et al., 2021). We showed here that drpr ^Δ5 is not a null allele contrarily to drpr ^indel3 (Figure 1). The MB-pruning phenotype of drpr ^indel3 was very similar, if not identical, to that of orion ¹ with 100% penetrance of the phenotype (n = 36 and n > 20 MBs, respectively), although most of the drpr ^Δ5 MBs appeared wild-type, with only a minority of MBs showing unclear axonal debris and unpruned axons, as observed with UAS-mGFP under the control of 201Y-GAL4 (Figures 1A–E). We confirm these results by labeling the MBs with an anti-Dlp antibody, which, in adult, particularly labels the MB γ-neurons (Figures 1F–J). Drpr belongs to the Nimrod family of the phagocytic receptor together with NimC1 and Eater (Melcarne et al., 2019). NimC1 and eater double null mutant flies are deficient in the phagocytosis of all types of bacteria, suggesting that they are the two main phagocytic receptors for bacteria in Drosophila (Melcarne et al., 2019). Nevertheless, NimC1 and eater double null mutant flies had no effect on MB remodeling (Figure 1K). This may indicate that phagocytosis of neuronal debris after remodeling requires Drpr as a specialized signaling pathway but does not require the main receptors involved in microbe engulfment. Finally, we showed, through Western blotting, that although drpr ^indel3 pupae did not produce any Drpr protein, some Drpr proteins were still present in drpr ^Δ5 pupae, confirming the null allele status of drpr ^indel3 (Figure 1L). These results are in accordance with drpr ^Δ5 being a mutation affecting strongly, but not completely, the transcription of the drpr mRNA (Freeman et al., 2003). Therefore, in rare cases when a Drpr protein is produced, in drpr ^Δ5 flies, it is a normal-sized protein. Contrarily, drpr ^indel3 is a CRISPR mutation affecting the coding sequence, leading to a protein null (Sapar et al., 2018). It is expected that in an exclusive relationship between a ligand and its receptor for a particular process, the null mutant phenotype of each gene has a similar, if not identical, phenotype. This result is a strong indication that the glial phagocytic receptor Drpr functions as a receptor for the neuron-secreted Orion ligand during the MB remodeling process.

Orion is a signal peptide-bearing protein without any transmembrane domain. Therefore, Orion could move in the extracellular space (Boulanger et al., 2021). We wondered if expression of Orion that is permanently tethered to the plasma membrane of otherwise wild-type MB axons could alter Orion function and produce a gain-of-function phenotype. We examined L3 larvae, a stage when the remodeling process has normally not begun, and found that although expression of a wild-type UAS-orion transgene in the MBs showed no glial phenotype, expression of UAS-orion-CD2 dramatically recruited glia inside γ-axon bundles, as monitored by the presence of the Drpr receptor, which serves as a glial marker (Figure 2). We assessed that similar expression levels of Orion-Myc and Orion-CD2-Myc were observed at the tip of the larval dorsal γ-lobes (10604±561 and 11114±842, respectively, n = 25 measures on 5 MBs for each condition, p = 0.84, Mann–Whitney U test). In the UAS-orion-CD2 transgene, Orion is located on the extracellular side of the CD2 transmembrane domain and could potentially interact with the glial receptor Drpr. Why is there glia infiltration, at the L3 stage, when the UAS-orion-CD2 transgene, but not the wild-type UAS-orion transgene, is used? We can propose two non-exclusive hypotheses. One is that Orion-CD2, contrarily to Orion, is not secreted and therefore is immediately available for the potential protein–protein interaction. Another hypothesis could be a similar model for the MBs and for the C4da neuron dendrite pruning (Ji et al., 2023). The CD2 transmembrane domain contained in Orion-CD2 could mimic an extracellularly exposed molecule through which Orion could bridge Drpr. This molecule could be a glycosaminoglycan (GAG) since the GAG binding is necessary for Orion function (Boulanger et al., 2021) or PS as in the case of the C4da neurons. In L3 larvae, these molecules may not be present, or present but not yet functional. The artificial binding of Orion-CD2 to the γ-axon plasma membrane could circumvent the necessity of Orion binding to GAGs or PS, allowing a premature and likely more stable direct interaction with Drpr.

We then asked whether the gain-of-function phenotype caused by the expression of the UAS-orion-CD2 transgene in L3 (i.e., abnormal early attraction of glia on the MB) could produce a morphological phenotype later in development. We examined 6 h after puparium formation (APF) and found that contrarily to the wild-type situation, when the UAS-orion-CD2 transgene was expressed in MBs, the extremities of the lobes were missing. This phenotype is particularly evident in the dorsal lobe and was termed the “headless” phenotype (Figure 3) (Figures 3A–C). We reasoned that if this “headless” phenotype resulted from an interaction between neuronal Orion-CD2 and glial Drpr, it should be rescued in a drpr ^null genetic background. Noteworthy, we observed a complete rescue of the Orion-CD2-induced headless phenotype in drpr ^indel3 pupae at 6 h APF (Figure 3D). In addition, we hypothesized that ectopic expression of Drpr in MB axons, where it is not normally expressed, might also rescue the phenotype. Interestingly, we found a clear rescue of the “headless” phenotype under these conditions (Figure 3E). Notably, the lack of Drpr expression (in drpr ^indel3 individuals) or the ectopic expression of Drpr alone in MBs did not produce an MB phenotype (Figures 3E,F). These results suggest that ectopic expression of Drpr in the MB neurons titrates Orion-CD2, thus preventing it from an early abnormal interaction with endogenous glial Drpr (Figures 3G–I ). The endogenous glial Drpr receptor is then available to interact normally with endogenous secreted Orion.

We finally asked whether Orion-CD2 could titrate endogenous glial Drpr. To address this, we ectopically expressed Orion-CD2 in glia and examined MB pruning at 24 h APF using an anti-Dlp antibody. Indeed, expression of Orion-CD2 in glia blocked the pruning process compared to the wild-type situation (Figures 4A,B). In addition, the pruning phenotype caused by ectopic expression of Orion-CD2 in glia was rescued when Drpr was simultaneously overexpressed in glia (Figure 4C). These phenotypes suggest that ectopic expression of Orion-CD2 in glia titrates endogenous glial Drpr, thereby preventing it from interacting with endogenous neuron-secreted Orion (Figures 4G,H). In addition, when Drpr is overexpressed in glia together with Orion-CD2, it titrates Orion-CD2, allowing the normal interaction between endogenous glial Drpr and endogenous neuron-secreted Orion (Figures 4D-F,I) and leading consequently to normal pruning.

The requirement of the presence of the Drpr protein to allow the Orion-CD2-induced headless phenotype and the titrating results of Drpr on Orion-CD2 and conversely of Orion-CD2 on Drpr associated to the similar, if not identical, mutant phenotype of both null alleles of drpr and orion strongly support the conclusion that the neuron-secreted chemokine-like Orion ligand interacts with the glial phagocytic receptor Drpr during MB remodeling in Drosophila.