Once initially formed during development, synapses undergo structural and functional modifications in response to genetic programs or environmental cues. This dynamic synapse feature is critical for the refinement of neural circuits during development and underlies learning and memory during animal life. Increasing evidence suggests that a family of cell adhesion molecules (CAMs), including vertebrate neural CAM (NCAM), Aplysia apCAM, and Drosophila Fasciclin II (FasII), plays an important role in structural synaptic plasticity (Dalva et al., 2007; Packard et al., 2003). NCAM knockout mice display impairments in hippocampal long-term potentiation (LTP) and spatial memory (Bukalo et al., 2004; Cremer et al., 1994; Muller et al., 1996). In Aplysia sensorimotor co-cultures, serotonin-induced long-term facilitation (LTF) is accompanied by the formation of new synaptic connections between sensory and motor neurons (Glanzman et al., 1990). This structural change is associated with downregulation of apCAM surface expression in presynaptic sensory neurons (Mayford et al., 1992). Similarly, at the Drosophila neuromuscular junction (NMJ), mutations that increase neuronal activity (e.g., eag Shaker) and cAMP levels (e.g., dunce) cause a decrease in synaptic FasII and a concurrent increase in bouton number (Schuster et al., 1996b). Consistently, hypomorphic fasII mutations that decrease FasII levels also cause synaptic overgrowth (Schuster et al., 1996a). Importantly, synaptic overgrowth in eag Shaker and dunce is suppressed by transgenic expression to maintain synaptic FasII levels (Schuster et al., 1996b), supporting the role of FasII downregulation in activity-dependent synaptic growth. Collectively, these findings suggest that synaptic growth is negatively regulated by synaptic levels of NCAM-type proteins.

Endocytosis and subsequent lysosomal degradation of NCAM-type proteins are apparently the key mechanisms that reduce cell adhesion during synaptic growth. In Aplysia sensory neurons, serotonin downregulates synaptic apCAM by promoting its internalization and rerouting from a recycling to a lysosomal degradation pathway (Bailey et al., 1992; Mayford et al., 1992). Subsequent experiments suggested that the internalization of apCAM requires mitogen-activated protein kinase (MAPK)-dependent phosphorylation in its PEST sequence (Bailey et al., 1997; Michael et al., 1998). MAPK signaling has also been implicated in activity-dependent downregulation of synaptic FasII at the Drosophila NMJ. Increased neural activity in some paralytic mutants causes persistent activation of neuronal MAPK signaling (Hoeffer et al., 2003), which in turn leads to reduced synaptic FasII and increased synaptic growth (Hoeffer et al., 2003; Koh et al., 2002). Taken together, these findings indicate the important role of endocytic downregulation of NCAM-type proteins in synaptic plasticity. However, knowledge regarding the molecular machinery controlling endocytosis and endosomal trafficking of NCAM-type proteins is limited.

In mammalian cells, efficient recycling of endocytosed receptors and CAMs to the plasma membrane is mediated by two members of the molecules interacting with CasL (Mical) protein family, Mical-like 1 (Mical-L1) and Mical-like 2 (Mical-L2), which share an N-terminal calponin homology (CH) domain, a central Lin11, Isl-1, and Mec-3 (LIM) domain, and a C-terminal coiled-coil (CC) domain that can interact with several Rab-family proteins (Fukuda et al., 2008; Yamamura et al., 2008). Mical-L1 has been proposed to facilitate endocytic recycling of transferrin and β1 integrin receptors by recruiting and linking the endocytic regulators Rab8a and EH domain-containing protein 1 (EHD1) on recycling endosomes (Sharma et al., 2009). In addition, Mical-L2 is shown to interact with Rab8 and Rab13 to facilitate endocytic recycling of E-cadherin and occudin, respectively, in epithelial cells (Terai et al., 2006; Yamamura et al., 2008). Although the interaction of Mical-L2 with Rab13 is essential for neurite outgrowth in cultured PC12 cells (Sakai et al., 2010), its role in the nervous system remains poorly understood.

A single Mical-like ortholog is expressed in Drosophila (Giridharan and Caplan, 2014). In this study, we examine the synaptic function of Mical-like at the Drosophila NMJ. We show that Mical-like acts presynaptically to restrain synaptic growth and maintain normal levels of synaptic FasII. In addition, loss of Mical-like impairs the recycling of internalized FasII molecules to the cell surface and enhances their degradative endosomal trafficking. Our results reveal a novel role for Mical-like in the regulation of FasII membrane trafficking and synaptic development.

Drosophila strains

All fly strains were reared on standard Drosophila food at 25°C. w¹¹¹⁸ was used as the wild-type strain. A P-element allele of mical-like (G8603) was obtained from GenExel (Korea). A transgenic line carrying UAS-mical-like^RNAi was obtained from the Bloomington Stock Center (USA). Expression of this transgene was driven by either the neuron-specific C155-GAL4 driver (Lin and Goodman, 1994) or the muscle-specific BG57-GAL4 driver (Budnik et al., 1996).

Molecular biology

To generate pAc-Myc-fasII, Drosophila fasII cDNA was amplified by PCR from genomic DNA prepared from a UAS-FasII-YFP transgene (Kohsaka et al., 2007) and then subcloned into the pTOP Blunt V2 vector (Enzynomics, Korea). A c-Myc epitope sequence was introduced immediately downstream of the signal sequence of fasII via PCR-based mutagenesis. The resulting c-Myc-fasII insert was subcloned into the pAc5.1 vector (Invitrogen, USA) for expression in Drosophila S2R+ cells.

Cell culture and transfection

S2R+ cells were cultured at 25°C in Schneider’s medium (Invitrogen, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS). The cells were transfected in six-well plates using Cellfectin (Invitrogen, USA). In each experiment, 1 × 10⁶ cells were transfected with 2 μg of plasmid DNA or 5 μg of dsRNA in six well plates.

For RNAi-mediated depletion of Mical-like in S2R+ cells, mical-like dsRNA was produced by in vitro transcription of a DNA template containing T7 promoter sequences at both ends, as previously described (Nahm et al., 2010b). The DNA template for in vitro transcription was generated by PCR using primers containing a T7 promoter sequence upstream of the following mical-like-specific sequences: 5′-TTGGAGTATTGGTGCCGAGTT-3′ and 5′-CTCCTCTGGCTTCGGGTGCTT-3′.

Immunohistochemistry and morphological quantification of larval NMJs

Wandering third instar larvae were collected and dissected in ice-cold Ca²⁺-free HL3 saline. Larval fillets were fixed for 20 min in 4% formaldehyde in phosphate-buffered saline (PBS). Antibody staining for fixed samples was performed as previously described (Nahm et al., 2010a). The following primary antibodies were used in this study: mouse anti-FasII (1D4; DSHB, USA) at 1:20 and FITC-conjugated goat anti-HRP (Jackson ImmunoResearch Laboratories, USA) at 1:200. A Cy3-conjugated donkey anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories, USA) was used at a dilution of 1:200. Confocal images were taken with an Olympus FV300 laser-scanning confocal microscope using a Plan Apo 40× 0.90 NA or U Plan Apo 100x 1.35 NA objective and processed using FLOUVIEW imaging analysis software. To compare different genotypes, samples for each experiment were prepared simultaneously and imaged under identical confocal settings.

Bouton number was quantified at NMJ 6/7 in abdominal segment 2 after anti-HRP staining as previously described (Nahm et al., 2010a). For quantification of synaptic FasII levels, larval NMJs were double-stained with anti-HRP and anti-FasII antibodies. The fluorescence intensity of synaptic FasII was measured using ZEN imaging software (Carl Zeiss, Germany) and then normalized to HRP intensity.

Cell surface FasII expression assay

S2R+ cells transfected with the pAc-Myc-fasII construct with or without mical-like dsRNA were incubated for 72 h and then treated with 100 μg/ml cycloheximide for 5 h to inhibit protein synthesis. To visualize surface FasII molecules, the cells were incubated with rabbit anti-Myc antibody (Cell Signaling, USA; 1:100) in culture medium for 1 h at 4°C and then fixed for 20 min with 4% formaldehyde in PBS. After washing three times with PBS, cells were incubated with Cy3-conjugated donkey anti-rabbit secondary antibody (Jackson ImmunoResearch Laboratories, USA; 1:200) for 30 min. To visualize the total pool of FasII, the same cells were then permeabilized with 0.2% Triton X-100 in PBS for 10 min and sequentially incubated with mouse anti-FasII antibody (DSHB, USA; 1:50) for 1 h and FITC-conjugated donkey anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories, USA; 1:200) for 30 min. S2R+ cells showed no detectable levels of endogenous FasII based on anti-FasII immunostaining. Images of immunostained cells were acquired using a LSM 700 laser-scanning confocal microscope (Carl Zeiss, Germany). Images were processed using the ZEN imaging software. To quantify surface FasII expression, the fluorescence intensity of surface Myc-FasII (red) was normalized to the total Myc-FasII (green) fluorescence intensity.

FasII trafficking assay

For the FasII recycling assay, live S2R+ cells transfected with pAc-Myc-fasII were incubated with rabbit anti-Myc antibody (1:100) in culture medium at 4°C for 1 h to label surface Myc-FasII proteins. Cells were incubated in culture medium at 25°C for 30 min to allow internalization and then washed with ice-cold acid buffer (0.5 M NaCl, 0.2 M acetic acid, pH 4.5) for 4 min to remove the antibodies remaining on the cell surface. After washing three times with culture medium, cells were incubated in culture medium at 25°C for 1 h to allow recycling of the internalized FasII-antibody complex to the cell surface. Following fixation, cells were incubated with Cy3-conjugated anti-rabbit secondary for 30 min to visualize the surface FasII pool. Subsequently, cells were permeabilized with 0.2% Triton X-100 in PBS for 10 min and stained with FITC-conjugated anti-rabbit secondary antibody for 30 min to visualize the internalized FasII pool. For quantification of FasII recycling, the fluorescence intensity of recycled Myc-FasII (red) was normalized to internal Myc-FasII (green) fluorescence intensity.

To analyze endosomal sorting of internalized FasII proteins to lysosomes, S2R+ cells transfected with pAC-Myc-fasII and pAc-GFP-rab7 or pAc-spin-GFP were surface-labeled with rabbit anti-Myc antibody as described above, washed, and then incubated in culture medium for 30 min at 25°C to allow internalization. Following fixation and permeabilization, cells were incubated with mouse anti-GFP antibody (Merck Millipore, Germany; 1:100) for 1 h and sequentially with Cy3-conjugated donkey anti-rabbit and FITC-labeled donkey anti-mouse secondary antibodies for 30 min.

FM1-43 dye experiments

FM1-43 dye loading and unloading at the larval NMJ were achieved as previously described (Verstreken et al., 2008). Briefly, wandering third-instar larvae were dissected in Ca²⁺-free HL3 saline and incubated in HL3 saline containing 90 mM KCl, 5 mM CaCl₂, and 4 μM FM1-43 (Molecular Probes, USA) for 1 min. Excessive dye was washed away with Ca²⁺-free HL3 saline before imaging of loaded NMJ synapses. Subsequently, NMJ synapses were subjected to a second stimulation with 90 mM K⁺ saline for 15 or 30 min, washed with Ca²⁺-free HL3 saline, and imaged for the remaining dye. Images of NMJ synapses were acquired using a Plan-NEOFLUAR 40× 0.75 NA water-immersion lens on the Axio Imager D1 epifluorescence microscope (Karl Zeiss, Germany), and the fluorescence intensity was quantified using NIH ImageJ.

Mical-like is required presynaptically to restrain synaptic growth at the Drosophila larval NMJ

In a genetic screen for mutations that affect synaptic morphology at the Drosophila larval NMJ (Nahm et al., 2010a; 2010b), we identified an enhancer P (EP) insertion, G8603, that resides in the first exon of the mical-like gene (Fig. 1A). mical-like^G8603 mutants were homozygous viable and had more extensive NMJs than wild-type (w¹¹¹⁸) controls (Fig. 1B). To quantify the mical-like phenotype, we counted the number of synaptic boutons on muscles 6 and 7 of abdominal segment A2 of third-instar larvae. Because NMJ growth is coupled with an increase in muscle size (Schuster et al., 1996a), we normalized bouton numbers to muscle surface area. Normalized bouton number was increased by 30% in mical-like mutants compared with wild-type larvae (1.74 ± 0.03 versus 1.33 ± 0.04 ×10³ boutons/μm²; P < 0.001; Fig. 1C).

To determine whether Mical-like is required pre- or postsynaptically, we knocked down mical-like in a tissue-specific fashion using a transgenic RNA interference (RNAi) approach. When UAS-mical-like^RNAi was driven by a pan-neuronal C155-GAL4 driver, we observed a 23% increase in the number of boutons (Fig. 1C). In contrast, expression of UAS-mical-like^RNAi with the muscle-specific BG57-GAL4 driver had no significant effect on overall bouton number (Fig. 1C). These data indicate that Mical-like function is required presynaptically for normal synaptic growth at the NMJ.

Mical-like is essential for the normal abundance of FasII at the NMJ

Given the role of Mical-like in the regulation of synaptic growth, we further examined mical-like third instar NMJs with several synaptic markers. The levels and distribution of the presynaptic vesicle marker Cysteine-string protein (Csp), the presynaptic active zone marker Bruchpilot/NC82, and the postsynaptic subsynaptic reticulum (SSR) marker Discs-large (Dlg) were grossly normal in homozygous mical-like^G8603 mutants (data not shown). However, we observed a 40% reduction in the level of the cell adhesion molecule FasII (P < 0.001; Figs. 2A and 2B). A similar decrease of synaptic FasII levels was also induced by presynaptic knockdown of mical-like based on RNAi (Fig. 2B), consistent with its effect on synaptic growth (Fig. 1C). In contrast, a change in FasII levels following knockdown of mical-like in postsynaptic muscles could not be detected (Fig. 2B). These data reveal a presynaptic role of Mical-like in the regulation of FasII levels.

Mical-like is required for the recycling of internalized FasII proteins to the cell surface

We investigated how Mical-like downregulates synaptic FasII. Mammalian Mical-like homologs (Mical-L1 and Mical-L2) have

been implicated in the regulation of endocytic recycling of junctional proteins in non-neuronal cells (Sharma et al., 2009; Terai et al., 2006; Yamamura et al., 2008). In addition, FasII is shown to cycle between the plasma membrane and intracellular endocytic compartments at the Drosophila NMJ (Mathew et al., 2003). Based on these findings, we hypothesized that Mical-like plays a role in synaptic FasII cycling. To test this possibility, we transiently transfected a Myc-FasII construct into Drosophila S2 cells. The surface and total pools of FasII were visualized by immunostaining with anti-Myc antibody before cellular permeabilization and subsequently with anti-FasII antibody following permeabilization, respectively. The ratio of surface to total Myc-FasII levels was decreased by approximately 29% in cells treated with mical-like dsRNA compared with mock-treated control cells (Figs. 3A and 3B). This change in FasII surface expression supports the role of Mical-like in the endocytic downregulation of FasII.

Next, we directly investigated the role of Mical-like in the recycling of internalized FasII to the plasma membrane. Live S2R+ cells expressing Myc-FasII were surface-labeled with anti-Myc antibody at 4°C, a condition that does not permit endocytosis. Cells were then incubated at 25°C for 30 min to allow internalization of the Myc-FasII-antibody complexes. After stripping away non-internalized antibodies using an acid buffer, the cells were further incubated at 25°C for 1 h to allow recycling of the internalized Myc-FasII-antibody complex back to the plasma membrane. Surface and intracellular Myc-FasII protein levels were determined by anti-Myc immunostaining under nonpermeant and permeant conditions, respectively, after the acid wash or additional 1 h incubation. After the acid wash, there was no significant difference in the amount of internalized Myc-FasII-antibody complexes between mical-like-knockdown and mock-treated control cells (Figs. 3C and 3D). In contrast, the level of surface-recycled Myc-FasII-antibody complexes at 1 h after the acid wash was significantly reduced in mical-like-knockdown cells compared with mock-treated controls (Figs. 3C and 3D). These results suggest that Mical-like is required for endocytic recycling but not internalization of FasII proteins.

Finally, we monitored endosomal trafficking of the internalized FasII in comparison with the late endosomal/lysosomal markers Rab7 and Spinster (Spin). To do so, live control and mical-like knockdown S2R+ cells expressing Myc-FasII and GFP-Rab7 or Spin-GFP were prelabeled with anti-Myc antibody at 4°C and then incubated at 25°C for 30 min to resume endocytosis prior to immunostaining for Myc and GFP under permeant conditions. In control cells, the internalized Myc-FasII poorly overlapped with GFP-Rab7 and Spin-GFP (Figs. 3E and 3F). In contrast, we observed extensive overlap between Myc-FasII and the late endosomal/lysosomal markers in mical-like-knockdown cells (Figs. 3E and 3F). These results indicate that the impaired recycling of FasII in the absence of Mical-like is coupled with its preferential sorting to the degradative compartment.

Altogether, these data imply that Mical-like maintains synaptic FasII levels by promoting the reincorporation of internalized FasII to the synaptic membrane.

Mical-like is required for synaptic vesicle exocytosis at the NMJ

The lipophilic dye FM1-43 can reversibly bind to membranes and has been used to monitor synaptic vesicle (SV) endocytosis and exocytosis at the Drosophila larval NMJ (Verstreken et al., 2008). To examine whether Mical-like also plays a role in exocytosis of SVs in addition to FasII, we monitored the loading and unloading kinetics of FM1-43 at mical-like^G8603 synapses. Larval NMJ arbors were first loaded with FM dye by depolarizing NMJ synapses using saline containing 90 mM K⁺, and then washed with Ca²⁺-free saline to stop SV cycling. Subsequently, FM dye was unloaded by applying a second depolarizing stimulus. Under these conditions, dye loading was not significantly different between mical-like^G8603 and wild-type animals (Fig. 4A). Subsequently, wild-type boutons were almost completely unloaded by the second depolarizing stimulus within 30 min (Figs. 4A and 4B). However, at this time point, mical-like^G8603 boutons showed a high retention of FM1-43 (Figs. 4A and 4B). Taken together, these data indicate that mical-like mutants are severely defective in SV exocytosis, but not endocytosis.

Mammalian Mical-like proteins Mical-L1 and Mical-L2 play a role in endocytic recycling of cell surface receptors or junctional proteins in non-neuronal cells (Sharma et al., 2009; Terai et al., 2006; Yamamura et al., 2008). Although these proteins are expressed in neurons (Sakai et al., 2010), their neuronal function remains unclear. Here, in a genetic screen for mutations causing defects in NMJ development, we identify mical-like encoding a single Drosophila ortholog of Mical-like proteins. mical-like mutant NMJs show an increase in bouton number and a decrease in FasII levels. Our targeted RNAi knockdown experiments demonstrate that Mical-like acts presynaptically to control synaptic growth and FasII levels.

Is there a relationship between two synaptic phenotypes induced by loss of Mical-like? Prior studies on the NMJ have demonstrated two main phenotypes of various FasII mutants depending on FasII levels (Schuster et al., 1996a). In FasII null mutants, synaptic boutons form normally during embryogenesis but subsequently retract during larval development. In contrast, hypomorphic fasII mutants, which express FasII above a threshold for supporting synaptic maintenance, show synaptic overgrowth. These findings suggest that FasII-mediated cell adhesion is essential for synaptic maintenance, but also constrains further synaptic expansion when FasII expression is beyond a certain level. Thus, it is tempting to speculate that Mical-like restrains synaptic growth by maintaining synaptic FasII levels.

At the Drosophila NMJ, FasII is shown to constitutively cycle between the plasma membrane and intracellular compartments (Mathew et al., 2003), suggesting that FasII-mediated cell adhesion could be regulated by endocytic membrane trafficking. Despite this emerging outline for synaptic FasII membrane cycling, the identity of the molecular machinery governing this process is poorly defined. At the postsynaptic terminal, the reincorporation of internalized FasII molecules to the cell surface is accomplished by an amphiphysin/soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE)-dependent mechanism (Mathew et al., 2003; Packard et al., 2003). Our current data uncover a novel mechanism of presynaptic FasII recycling, which is mediated by Mical-like. Loss of presynaptic, but not postsynaptic, Mical-like decreases FasII levels at NMJ synapses. We also show that loss of Mical-like selectively impairs the recycling of internalized FasII molecules to the cell surface without affecting endocytosis. Finally, our FM1-43 dye uptake/release experiments also reveal the selective role of Mical-like in SV exocytosis at the NMJ.

In conclusion, our data demonstrate a conserved role for Drosophila Mical-like in regulating endocytic recycling and indicate that this role of Mical-like is important for the proper regulation of FasII levels at synapses. Given the importance of local FasII regulation for structural synaptic plasticity, it will be interesting to determine in the future whether the activity of Mical-like to promote FasII recycling is finely regulated in an activity-dependent manner.