Control of Synaptic Levels of Nicotinic Acetylcholine Receptor by the Sequestering Subunit Dα5 and Secreted Scaffold Protein Hig

Abstract

The presentation of nicotinic acetylcholine receptors (nAChRs) on synaptic membranes is crucial for generating cholinergic circuits, some of which are associated with memory function and neurodegenerative disorders. Although the physiology and structure of nAChR, a cation channel comprising five subunits, have been extensively studied, little is known about how the receptor levels in interneuronal synapses are determined and which nAChR subunits participate in the regulatory process in cooperation with synaptic cleft matrices and intracellular proteins. By a genetic screen of Drosophila, we identified mutations in the nAChR subunit Dα5 gene as suppressors that restored the mutant phenotypes of hig, which encodes a secretory matrix protein localized to cholinergic synaptic clefts in the brain. Only the loss of function of Dα5 among the 10 nAChR subunits suppressed hig mutant phenotypes in both male and female flies. Dα5 behaved as a lethal factor when Hig was defective; loss of Dα5 in hig mutants rescued lethality, upregulating Dα6 synaptic levels. By contrast, levels of Dα5, Dα6, and Dα7 subunits were all reduced in hig mutants. These three subunits have distinct properties for interaction with Hig or trafficking, as confirmed by chimeric subunit experiments. Notably, the chimeric Dα5 protein, which has the extracellular sequences that display no positive interaction with Hig, exhibited abnormal distribution and lethality even in the presence of Hig. We propose that the sequestering subunit Dα5 functions by reducing synaptic levels of nAChR through internalization, and this process is blocked by Hig, which tethers Dα5 to the synaptic cleft matrix.

SIGNIFICANCE STATEMENT Because the cholinergic synapse is one of the major synapses that generate various brain functions, numerous studies have sought to reveal the physiology and structure of the nicotinic acetylcholine receptor (nAChR). However, little is known about how synaptic levels of nAChR are controlled and which nAChR subunits participate in the regulatory process in cooperation with synaptic cleft matrices. By a genetic screen of Drosophila, we identified mutations in the nAChR subunit Dα5 gene as suppressors that restored the mutant phenotypes of hig, which encodes a secretory matrix protein localized to cholinergic synaptic clefts. Our data indicate that Dα5 functions in reducing synaptic levels of nAChR, and this process is blocked by Hig, which tethers Dα5 to the synaptic cleft matrix.

Introduction

Nicotinic acetylcholine receptors (nAChRs) mediate diverse neurophysiological effects of acetylcholine in synapses, thus contributing to the activities of neural circuits that regulate cognitive behaviors such as the performance of attention and learning tasks (Ballinger et al., 2016; Barnstedt et al., 2016; Sabec et al., 2018). nAChRs are also implicated in numerous human disorders such as myasthenia gravis, Alzheimer's disease, nicotine addiction, and schizophrenia (Steinlein and Bertrand, 2008; Albuquerque et al., 2009; Higley and Picciotto, 2014; Wallace and Bertrand, 2015), and pathogenic amyloid β peptides bind to or interact with particular nAChRs (Wang et al., 2000; Nagele et al., 2002; Grassi et al., 2003; Snyder et al., 2005; Hu et al., 2008; Dziewczapolski et al., 2009). Therefore, these receptors are promising therapeutic targets.

nAChRs comprise five subunits arranged around the receptor pore to form either a homomeric or heteromeric pentamer. In mammals, 16 nAChR subunits, classified as α or non-α (β, γ, δ, or ε), are incorporated into receptors in a distinct proportion. The combinations of these subunits produce various receptors whose physiological properties differ. The most common nAChR subtypes in the brain are α4β2-containing and α7-containing nAChRs. These have characteristic properties in terms of permeability to calcium (Fucile, 2004) and affinity for acetylcholine (Gotti et al., 2006). In addition, α7-containing receptors are restricted to the perisynaptic dendritic membrane and excluded from the postsynaptic membrane (Williams et al., 1998; Temburni et al., 2000). Furthermore, nAChRs containing α6 interact with a specific set of accessory components (Gu et al., 2019). These data suggest that at least some subunits have a complex function beyond simply contributing to the physiological properties of the receptor. However, how synaptic levels of the nAChR are regulated in the CNS and which subunits are critical for the regulatory mechanism remain unclear.

Drosophila has 10 nAChR subunits, classified as α or β (Dα1−7, Dβ1−3). Although widely observed roles of vertebrate nAChRs in the CNS include a modulation of neurotransmitter release at presynapses (Dani and Bertrand, 2007), accumulated data have shown that a main function of Drosophila nAChRs is more limited to mediating neurotransmission postsynaptically (Fayyazuddin et al., 2006; Barnstedt et al., 2016). In the Drosophila brain, the expression of Dα5 is most strongly correlated with that of Dα6 and Dβ1 (Croset et al., 2018). Dα5 and Dα7 subunits also form functional homomeric and heteromeric channels in heterologous cells (Lansdell et al., 2012). Particular nAChR subunits participate in various neural functions, including light response and sleep (Fayyazuddin et al., 2006; Shi et al., 2014; Wu et al., 2014; Somers et al., 2017).

In a previous study, we showed that synaptic levels of Dα6 and Dα7 are reduced in hig mutants, indicating that Hig promotes the maintenance of nAChR levels (Nakayama et al., 2014). Hig is a complement control protein (CCP) domain-containing and Ig domain-containing secretory protein localized to the synaptic clefts of cholinergic synapses (Hoshino et al., 1993, 1996, 1999; Nakayama et al., 2014). Another synaptic cleft protein, Hasp, which also contains multiple CCP domains, is required for Hig localization at the clefts (Nakayama et al., 2016). Hig and Hasp form distinct molecular compartments within a cleft, indicating that the synaptic cleft is organized with a heterogeneous architecture. When the hig or hasp gene is mutated, most homozygotes die during development, and the surviving adult flies exhibit decreased locomotor activity and longevity.

In the present study, we performed a genetic screen for suppressors of hig and subsequently identified suppressor mutations in the Dα5 gene. Loss of function of Dα5 rescued lethality of hig mutants, concomitant with an increase in synaptic Dα6 levels. Data for chimeric subunits indicated that Dα5, Dα6, and Dα7 differ in terms of interactions with Hig or trafficking properties in postsynapses. We propose that Dα5 controls the synaptic levels of nAChRs by both facilitating the internalization of the receptors into postsynapses and interacting extracellularly with Hig. The data also indicate that one particular nAChR subunit can impair synapses and cause lethality, depending on the circumstances at synaptic clefts.

Materials and Methods

Drosophila strains.

hig^dd37 is a null mutation caused by deficiency in the hig gene (Hoshino et al., 1993). elav^C155-GAL4 and OK107-GAL4 were used to drive GAL4 (galactosidase-4) expression in all neurons and Kenyon cells, respectively, in the GAL4/UAS (upstream activating sequence) binary expression system. The mutant strain nAcRα-30D^DAS1 (hereafter called Dα6^DAS1; Watson et al., 2010), GAL4 driver lines (elav^C155, OK107-GAL4), and UAS lines (UAS-Dα7; stock #64148 and UAS-yfp-Rab5; stock #24616) were obtained from the Bloomington Drosophila Stock Center. Dα7^PΔEY6 was obtained from H.J. Bellen (Howard Hughes Medical Institute, Baylor College of Medicine, Houston, TX; Fayyazuddin et al., 2006). The RNAi lines for AChR subunits used in this study are as follows: Dα1 (stock #28688), Dα2 (stock #27493), Dα3 (stock #27671), Dα4 (stock #31985), Dα5 (stock #25943), Dα6 (stock #25835), Dα7 (stock #27251), Dβ1 (stock #31883), Dβ2 (stock #28038), and Dβ3 (stock #25927; all from Bloomington Drosophila Stock Center).

EMS mutagenesis and genome sequencing.

The econd chromosomes of hig^dd37 flies were isogenized by crossing with w¹¹¹⁸. The isogenized hig^dd37/CyO male flies were starved, exposed to 25 mm ethyl methanesulfonate (EMS) in 1% sucrose for 24 h, and then crossed with Sco/CyO female flies. Over 400 independent lines carrying EMS-treated second chromosomes harboring hig^dd37 were established for screening. To isolate suppressors of hig mutants, the recovery of locomotor activity was examined in flies homozygous for both hig^dd37 and secondary mutations. Following three rounds of outcrossing to wild-type flies to remove irrelevant mutations, the candidate line was analyzed by whole-genome sequencing (HiSeq 1500, Illumina) or Sanger sequencing to identify the mutations responsible for the suppression of hig.

Whole-genome sequencing.

Genomic DNAs from flies homozygous for hig^dd37 and two lines of EMS18 independently outcrossed to w¹¹¹⁸ flies were extracted using a genomic DNA extraction kit (catalog #28-9042-75, GE Healthcare). Thirty flies were used for each extraction. Shearing of the genomic DNA was performed in two stages on a sonicator (model S220, Covaris). Sonication was first performed at peak incidence power (PIP) 105 W, 5% duty cycle, and 200 cycles per burst conditions to obtain DNA fragments of ∼500 bp, and subsequently at PIP 140 W, 10% duty cycle, and 200 cycles per burst conditions to obtain fragments of ∼300 bp. In each step, sonication was conducted at 6°C bath temperature and 80 s duration in microTUBE AFA Fiber Pre-Slit Snap-Cap 6 × 16 mm tubes (Covaris). The resultant DNA fragments were purified using AMPure XP beads (Beckman Coulter). Genomic DNA libraries were constructed using the KAPA Library Preparation Kit (Roche) without PCR amplification. The libraries were quantified with the KAPA Library Quantification Kit (Roche) and sequenced with the Rapid Run mode of HiSeq1500 (Illumina) using the TruSeq Rapid PE Rapid Cluster Kit (Illumina) and the TruSeq Rapid SBS Kit (Illumina) to obtain paired-end reads of 151 nucleotides. Base calling was performed with RTA version 1.17.21.3, and fastq conversion was performed with bcl2fastq version 1.7 (Illumina). Sequencing data were assessed with FastQC 0.10.1 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Removal of adapter and low-quality regions was performed using Trim Galore version 0.3.3 (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) with the parameters “-e 0.1 -q 20.”

The sequencing data have been deposited in the DNA Data Bank of Japan (DDBJ) under the accession number DRA011100.

Mutation detection.

The processed reads were aligned with the genome assembly FlyBase dm5.54 using BWA version 0.7.5a (Li and Durbin, 2009) with the BWA-aln algorithm and default parameters. After removal of PCR duplicates using the MarkDuplicates function of Picard Tools version 1.103 (https://broadinstitute.github.io/picard/), local realignment of the mapped reads and SNP/InDel (single nucleotide polymorphism and insertion/deletion) calling were performed using the RealignerTargetCreator and UnifiedGenotyper functions of GenomeAnalysisToolKit version 2.7–4 (McKenna et al., 2010), respectively. To detect SNPs/InDels specific to the mutant samples, we extracted the mutations that were shared between the two ESM18 lines but not detected in the control. The extracted mutations were processed with SnpEff version 3_2 (Cingolani et al., 2012) to annotate the effects of the mutations on gene functions; those classified as “high” or “moderate” by the SnpEff definition were selected. All candidate SNPs/InDels were visualized and evaluated in the UCSC Integrative Genomics Viewer (IGV) version 2.2.23 (Thorvaldsdóttir et al., 2013).

Measurement of antigravity locomotion.

The wings of male flies were cut off with microscissors, and the wing-clipped flies were subjected to an assay of antigravity locomotion at 4–9 d after eclosion. Flies were tapped down to the bottom of a graduated plastic tube (15 ml high-clarity polypropylene conical tube, 17 × 120 mm style; Falcon) and allowed to freely climb the wall for 4 s at 25°C. The graduation each fly reached was scored three times, and the average value was used for statistical analysis. At least 44 flies were examined for each genotype.

Measurement of longevity.

Up to five flies were reared at 25°C in a vial containing normal fly media and transferred to a new vial every few days until the flies did not exhibit any movement. At least 25 flies were examined for each genotype.

Construction of transgenic lines.

To generate UAS-Dα6, Dα6 was PCR-amplified from GH15518 (DGRC) and cloned into the XhoI/XbaI sites of pJFRC7-20XUAS-IVS-mCD8::GFP (Addgene) to replace mCD8::GFP with Dα6. To generate UAS-Dα5^wt, UAS-Dα5^C437S, and UAS-Dα5^S597F, cDNAs of Dα5 were amplified by RT-PCR from hig^dd37, EMS18, and EMS312, respectively. The resultant cDNAs included a 6 bp insertion at N-terminal domain (ntd) 790 and a 9 bp insertion at ntd 847 relative to the sequence of nAChRα5-RB reported in FlyBase. To generate UAS-Dα5-FLAG, Dα5 was PCR amplified from Dα5 cDNA and cloned into the XhoI/XbaI sites of pJFRC7-20XUAS-IVS-mCD8::GFP (Addgene) to replace mCD8::GFP with Dα5-FLAG. To introduce the FLAG sequence following Dα5, PCR was performed with a synthetic oligonucleotide (5′-CTCGAGAGACT ATGAAAAATGCACAACTGAAACTGAC-3′) as the forward primer, and a synthetic oligonucleotide encoding 3× FLAG (5′-TCTAGATTAGCCGCCCTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCCGAGACAATAATATGTGG-3′) as the reverse primer.

To construct UAS-5Ex6Tm, the extracellular domain of Dα5 (amino acid positions 1–526 of Dα5) and the transmembrane domain of Dα6 (amino acid positions 244–494 of Dα6) were amplified and fused by a PCR-based method. Similarly, to construct UAS-6Ex5Tm, the extracellular domain of Dα6 (amino acid positions 1–243 of Dα6) and the transmembrane domain of Dα5 (amino acid positions 527–808 of Dα5) were amplified and fused. To construct UAS-5Ex7Tm, the extracellular domain of Dα5 and the transmembrane domain of Dα7 (amino acid positions 275–560 of Dα7) were amplified and fused. To construct UAS-7Ex5Tm, the extracellular domain of Dα7 (amino acid positions 1–274 of Dα7) and the transmembrane domain of Dα5 were amplified and fused. The resultant cDNAs were cloned into the XhoI/XbaI sites of pJFRC7-20XUAS-IVS. Transgenic fly lines were generated by Phi-C31-mediated site-specific integration at attP2. The primers used for amplification were as follows: 5 extracellular domain (5Ex) and 6 transmembrane and cytoplasmic domains (6Tm): 5Ex, Forward (Fw), CTCGAGAGACTATGAAAAATGCACAACTGAAACTGAC; Reverse (Rv), ACATGGCACAATTAAGTTGAAGAAATAGTACAGTGTTCG; 6 Tm, Fw, TACTATTTCTTCAACTTAATTGTGCCATGTGTGC; Rv, CCTCTAGATTATTGCACGATTATGTGCGGAGC; 6Ex5Tm: 6Ex, Fw, CTCGAGAGGACATGGACTCCCCGC; Rv, ACAAGGTATGATCAGATTGAAAAAATAATATAATGTACGGC; 5 Tm, Fw, TATTATTTTTTCAATCTGATCATACCTTGTGTACTG; Rv, CCTCTAGACTACGAGACAATAATATGTGGTGC; 5Ex7Tm: 5Ex, Fw, CTCGAGAGACTATGAAAAATGCACAACTGAAACTGAC; Rv, GCACGGCACAATCAGGTTGAAGAAATAGTACAGTGTTCG; 7 Tm, Fw, TACTATTTCTTCAACCTGATTGTGCCGTGCG; Rv, CCTCTAGATTACGGGAAAATGAAATGCGG; 7Ex5Tm: 7Ex, Fw, CTCGAGGAGTTATGAAGAAGCCATCACGCAG; Rv ACAAGGTATGATCAGATTGAAAAAATAGTACAACGTTTTGC; 5 Tm, Fw, TACTATTTCTTCAACCTGATTGTGCCGTGCG; Rv, CCTCTAGACTACGAGACAATAATATGTGGTGC.

Disruption of the Dα5 gene by CRISPR/CAS9.

To express Dα5-targeting gRNAs, we introduced the sequence 5′-GGCGAAGGTGATGTCTATAT-3′ into pBFv-U6.2B (Kondo and Ueda, 2013) and established transgenic flies. The resultant transgenic flies were crossed with nos-Cas9 flies (NIG-FRY). Genomic deletion of Dα5 in established homozygotes was determined by PCR and Sanger sequencing.

Antibody production.

To generate anti-Dα5 antibodies, the synthetic peptide RPMTPGGTLPHNPAFYRTVC, derived from the cytoplasmic loop region between the M3 and M4 transmembrane domains, was used as an antigen to immunize rabbits and mice. To generate anti-Dα7 antibody, the synthetic peptide CVGPAGPVVDGRLHEAIS, derived from the cytoplasmic loop region between M3 and M4, was used as an antigen to immunize rats. The anti-Dα5 and anti-Dα7 antisera were purified by affinity binding to the corresponding antigens.

Immunohistochemistry.

The brains of adult male flies were dissected in PBS, fixed on ice for 1 h with 4% paraformaldehyde in PBS, and stained with the following antibodies: anti-Hig (1:1000), anti-Bruchpilot (1:20; catalog #nc82, Developmental Studies Hybridoma Bank; Wagh et al., 2006), anti-DN-cadherin Ex#7 (1:20; Iwai et al., 1997), or anti-GFP (1:500; Thermo Fisher Scientific). Alexa Fluor 405 (1:200; Thermo Fisher Scientific), or Cy2, Cy3, or Cy5 conjugated antibodies (1:200; Jackson ImmunoResearch) were used as secondaries. For staining with rabbit or mouse anti-Dα5 (1:100; this study), rabbit or guinea pig anti-Dα6 (1:200; Nakayama et al., 2014), and rat anti-Dα7 (1:100; this study), dissected brains were fixed for 10 min on ice with 2% paraformaldehyde in PBS. Samples were observed by sequential scanning on FV1000 (Olympus) or SP2 (Leica) confocal microscopes.

Quantification of confocal images was performed using the Olympus confocal software, and fluorescence intensity was compared with that of control samples stained simultaneously in the same tube. Fluorescence detection for these samples within each experiment was performed with identical settings for laser power, detector sensitivity, scan speed and mode, pinhole size, and image magnification and resolution. A single optical section was used to measure the fluorescence for each sample, and the normalization of fluorescence intensity was performed by calculating the ratio of the intensity to the number of pixels in the region of interest (ROI). For quantification of fluorescence in the calyx of the mushroom body (MB), ROI corresponds to an area containing a group of microglomerulus but not the cell bodies of Kenyon cells. The background levels of fluorescence signals were measured in the visual field outside the brains, normalized as previously described, and then subtracted from the signal levels in the synaptic regions.

Quantitative RT-PCR.

Total RNA was extracted from heads using the RNeasy Mini Kit (QIAGEN). Over 100 heads were extracted for each experiment. cDNAs were synthesized using ReverTra Ace (TOYOBO). Quantitative RT-PCR was performed in AriaMx Real-Time PCR System (Agilent). The DyNAmo Flash SYBR Green qPCR Kit (Thermo Fisher Scientific) was used for quantitative PCR (qPCR) amplification and detection.

The following primers were used for qPCR: GAPH1, Fw, 5′-ATTTCGCTGAACGATAAGTTCGT-3′; GAPDH1, Rv, 5′-CGATGACGCGGTTGGAGTA-3′; Dα6, Fw: 5′-AATTGATCGGCGATTGGAAG-3′; Dα6, Rv, 5′-CCGACGTATCCGTAGCTTAATG-3′; Dα7, Fw, 5′-AGAAGCCATCACGCAGTCG-3′; Dα7, Rv, 5′-GTTAGTCCGAAGCTCAGTTGC-3′.

The primer pairs for Dα6 and Dα7 were positioned on different exons to prevent amplification of genomic DNA.

Preparation of whole-brain extracts.

Brains were dissected from ten adult flies and collected on dry ice. The brain samples were ground with a plastic pestle in 15 µl of homogenization buffer (10 mm HEPES, pH 7.5, 100 mm KCl, 1 mm EDTA, 10% glycerol, 0.1% Triton X-100, 5 mm DTT, 5 mm PMSF, and protein inhibitors; cOmplete, Mini EDTA-free Protease Inhibitor, Roche) and subjected to SDS-PAGE.

Immunoprecipitation and blotting.

Head extracts and subcellular fractions were prepared as previously reported (Schloss et al., 1988), with slight modifications. After whole adult bodies were frozen in liquid nitrogen, the heads were fractionated through a metal mesh and then ground with a mortar and pestle. The sample was homogenized in buffer A (10 mm Tris-HCl, pH 7.5, 280 mm sucrose, protease inhibitors; cOmplete, Mini EDTA-free Protease Inhibitor, Roche) with a glass–glass homogenizer. Following centrifugation at 1000 × g for 10 min at 4°C, the supernatant was further centrifuged at 20,000 × g for 30 min at 4°C. The precipitate was resuspended in buffer A and centrifuged again under the same conditions. The resultant pellet, which was expected to mainly contain cell membrane fragments and large organelles, was lysed in membrane extraction buffer (10 mm Tris-HCl, pH 8.5, 0.2 m NaCl, 1.8% Triton X-100, 0.6% sodium deoxycholate, 10% glycerol) and centrifuged at 48,600 × g for 60 min at 4°C. The supernatant was used as the membrane fraction, which was the source of synaptic membrane proteins used for immunoprecipitation. To obtain a soluble fraction as a source of Hig protein, adult heads were homogenized in HBST (10 mm HEPES, pH 7.4, 150 mm NaCl, 0.5%, w/v, Triton X-100, and protease inhibitors) with a plastic pestle in a 1.5 ml tube and centrifuged at 20,000 × g for 30 min at 4°C. The supernatant was passed through a 0.45 µm membrane filter (Millex) and used as the soluble fraction. To examine the formation of the Hig–Dα6 complex, the membrane fraction was incubated with rabbit anti-Dα6 antibody (1:500) and Protein G Mag Sepharose (GE Healthcare). After washing with HBST, the beads were incubated in HBST using the soluble fraction extracted from the same strain used to prepare the membrane fraction. To examine the formation of the Hig–Dα5-FLAG complex, anti-DDDDK-tag mAb-Magnetic Beads (MBL) were used for immunoprecipitation. For immunoblotting of Hig, Dα6, Dα5-FLAG, N-cadherin, and actin, guinea pig anti-Hig antibody (1:4000), guinea pig anti-Dα6 antibody (1:2000), rabbit anti-FLAG antibody (1:1000; MBL), rat anti-N-cadherin antibody (1:100; Iwai et al., 1997), and mouse anti-β-actin antibody (1:5000; Abcam) were used, respectively.

Experimental designs and statistical analyses.

Statistical differences between two experimental groups were assessed using the unpaired Student's t test or Mann–Whitney test. Differences among three or more experimental groups were evaluated with one-way ANOVA followed by Tukey's or Dunnett's multiple-comparisons test for normally distributed data and the Kruskal–Wallis test followed by Dunn's multiple-comparisons test for data not normally distributed. More descriptions of the statistical tests and data comparisons are provided in the corresponding figure legends. Statistical analyses were performed using Prism 8. Differences with p values < 0.05 were considered statistically significant. In all figures, error bars indicate the mean ± SD, ns represents “not significant,” and *, **, and *** represent p < 0.05, p < 0.01, and p < 0.001, respectively. The p values >0.001 are shown accurately in the respective figure legends.

Results

Loss of function of Dα5 suppresses the phenotypes of hig mutants

To elucidate the molecular mechanism by which Hig contributes to the differentiation of cholinergic synapses, we attempted to identify factors that closely interact with Hig by performing a screen for mutations that suppress the reduced locomotor activity of hig mutants (Fig. 1A). Of >400 hig^dd37 lines treated with EMS, two independent lines, EMS18 and EMS312, clearly exhibited restoration of antigravity locomotion when homozygous for the second chromosome, which carries both hig^dd37 and a putative secondary mutation, sup18 or sup312 (Fig. 1B). This restoration is not simply caused by an additive effect of the secondary mutations, because the flies carrying only the secondary mutation, but no hig defect, generated by chromosome crossover, exhibited a locomotor activity not distinguished from that of the wild type (Fig. 1B), suggesting that the suppression detected here arose from specific interactions of the mutated genes with hig. The two suppressor lines also recovered their longevity, which was shortened in hig^dd37 (Fig. 1C). The mutated genes in these lines functioned as weak dominant alleles but more effectively suppressed hig^dd37 when homozygous (see hig^dd37sup18/hig^dd37 vs hig^dd37sup18, and hig^dd37sup312/hig^dd37 vs hig^dd37sup312; Fig. 1C). Notably, trans-heterozygotes (hig^dd37sup18/hig^dd37sup312) for the second chromosomes of EMS18 and EMS312 exhibited phenotypes indistinguishable from those of each homozygous line (Fig. 1C), suggesting that the two mutations responsible for suppression were in the same gene. We performed whole-genome sequencing of EMS18 and identified a candidate suppressor mutation, sup18, in the gene encoding nAChR subunit Dα5. In the mutant protein, a serine residue was substituted for a cysteine residue at amino acid 437 of the Cys-loop structure (Fig. 2A,B). Sequencing the Dα5 gene of the other line, EMS312, also revealed a missense mutation, sup312, that replaced a serine residue at amino acid 597, in the third transmembrane region, with a phenylalanine residue (Fig. 2A,B). The synaptic levels of these mutant Dα5 proteins were significantly lower than the levels of wild-type protein in the calyces of mushroom bodies (Fig. 2C,D). In addition, the mutant proteins ectopically expressed in the brain tissues using the Gal4 driver OK107 (Fig. 2E) failed to normally accumulate at synapses (Fig. 2F,G). These data suggest that the missense mutations affect the stability or synapse-directed transport of Dα5 proteins.

Figure 1.

Mutations suppressing hig phenotypes were identified in Dα5. A, Strategy for isolating hig suppressor mutations. EMS-treated hig^dd37/CyO male flies were crossed to females carrying a second chromosome balancer. Among the progeny lines, the lines of flies that restored locomotor activity when homozygous for the second chromosome were isolated in the screen. B, Antigravity locomotion of hig-suppressor mutants. Each mutant line of flies is homozygous for the second chromosome carrying hig^dd37, hig^dd37, and suppressor (sup18 or sup312), or only suppressor mutations. N ≥ 44 flies for each genotype. ***p < 0.001, **p = 0.0019. All p values designated as “ns” are >0.99 by Kruskal–Wallis test followed by Dunn's multiple-comparisons test. C, Mean longevity of adult flies carrying hig^dd37, both hig^dd37and one of suppressors, or both hig^dd37and one of the nAChR subunit gene mutations. N ≥ 27 flies for each genotype. Statistical significance is presented relative to hig^dd37. ***p < 0.001, *p = 0.0165. p values designated as ns for Dα6^DAS1hig^dd37 and Da7^PΔEY6hig^dd37 are >0.999 and >0.998, respectively. One-way ANOVA followed by Tukey's multiple-comparisons test: F_(9,642) = 201.5, p < 0.0001. D, Suppression of reduced longevity of hig mutants by RNAi-mediated knockdown of nAChR subunit genes. Among the genes encoding nAChR subunits, only knockdown of Dα5 partly restored longevity, which was shortened in hig^dd37 flies. Each UAS-RNAi was expressed in all neurons by the Gal4 driver elav^C155. The genotype of control flies is elav^C155/Y; hig^dd37. Statistical significance is shown relative to the control. N ≥ 17 flies for each genotype. ***p < 0.001. All p values designated as ns are >0.993. One-way ANOVA followed by Tukey's multiple-comparisons test: F_(11,454) = 197.8, p < 0.0001.

Figure 2.

The suppressor mutations resulted in amino acid substitutions of Dα5. A, Substituted amino acids caused by the suppressor mutations in Dα5 protein. A schematic protein structure of Dα5 is shown with a lipid bilayer. M1–M4 are transmembrane domains. C437S (sup18) and S597F (sup312) were found in EMS18 and EMS312, respectively. B, The positions of the mutations and antigens in Dα5. The positions of the two suppressor mutations (magenta) and CRISPR/Cas9-mediated mutation (green) are shown along with the transmembrane regions M1–M4 and the amino acid sequence chosen as the antigen to raise the Dα5 antibody. The orange box represents the extracellular Cys-loop with two cysteines on either side. C, Distribution of wild-type and mutant Dα5 proteins in the calyx of the mushroom body in the adult brains. The signal for Dα5 disappeared in Dα5^CR5, indicating the specificity of the antibody used in this experiment. N-cadherin (Ncad) was used as a synaptic marker. D, Immunofluorescence intensities of Dα5 in Figure 2C. N ≥ 10 calyces were examined for each genotype. ***p < 0.001, *p = 0.0215 and 0.0353 for Dα5^C437S versus Dα5^S597F, and Dα5^S597F versus Dα5^CR5, respectively. One-way ANOVA followed by Tukey's multiple-comparisons test: F_(3,38) = 318.0, p < 0.0001. E, Expression patterns of the OK107-Gal4 driver. GFP (green) expressed under the control of OK107-Gal4 was visualized by staining with GFP antibody in the left figure. The GFP-positive areas include the calyx (ca), which is a synaptic cluster in the mushroom body. The whole synaptic regions in the brain are labeled with the Brp antibody (magenta). F, Distribution of wild-type and mutant Dα5 proteins overexpressed in the calyx of the mushroom body in adult brains. Dα5 expression is driven by OK107-Gal4. Enlarged views of cell bodies are shown in the insets. Note that Dα5^S597F protein accumulated in cell bodies. Scale bar: inset, 2 µm. G, Immunofluorescence intensities of wild-type and mutant Dα5 proteins in F. Values measured in calyces are presented relative to the wild type. N ≥ 11 calyces were examined for each genotype. ***p < 0.001. The p values designated as “ns” are 0.1093 and 0.6032 for wild-type versus OK107> Dα5^C437S, and wild-type versus OK107> Dα5^S597F, respectively. One-way ANOVA followed by Tukey's multiple-comparisons test: F_(3,46) = 189.2, p < 0.0001.

To further confirm that Dα5 is a suppressor of hig, we generated a knock-out mutation of the gene using the CRISPR/Cas9 technique. The frame-shift mutation Dα5^CR5, caused by a 2 bp deficiency just upstream of the transmembrane regions (Fig. 2B), also suppressed the reduced longevity of hig^dd37 (Fig. 1C). These data demonstrate that the loss of Dα5 function can suppress hig phenotypes. Notably, this finding suggests that runaway activity of Dα5 contributes to lethality in flies lacking Hig.

The Drosophila genome possesses the genes encoding 10 nAChR subunits (seven α and three β). Dα5, Dα6, and Dα7 share high amino acid sequence similarity. Hence, we investigated whether null mutations of Dα6 and Dα7 suppressed hig in the same way as Dα5 mutations. However, neither of those mutations restored longevity (Fig. 1C). To determine whether the loss of function of nAChR subunits other than these three proteins could suppress the mutant phenotypes of hig, we knocked down each of the nAChR subunits (Dα1–7 and Dβ1–3) by RNAi in the hig^dd37 background. Although knockdown of Dα5 restored the longevity of hig^dd37, reduced expression of the other subunits did not suppress the mutant phenotype (Fig. 1D). Collectively, these data indicate that, among all of the nAChR subunits, only the loss of function of Dα5 suppressed the mutant phenotype of hig.

Dα5 and Dα7 interact with hig at synaptic clefts

Identification of Dα5 as a suppressor of hig suggested that the two proteins closely interact with each other or function in the same signaling pathway. Therefore, we investigated how the distribution of Dα5 and Hig proteins were altered in the synaptic regions of hig and Dα5 mutant flies, respectively. In hig^dd37 flies, the level of Dα5 was reduced throughout the entire brain (Fig. 3A). The calyx of the mushroom body in the posterior brain contains a number of microglomeruli (Fig. 3B,C) comprising mainly cholinergic synapses (Yasuyama et al., 2002), each of which associates with Hig in the synaptic cleft (Fig. 3I, schematic illustration of a microglomerulus). In the calyx of hig^dd37 flies, the level of Dα5 was reduced to 50% of the level in wild-type flies; however, it was restored by the expression of Hig-GFP (Fig. 3B,D). Conversely, in Dα5^CR5 mutants, the level of Hig was reduced in most synaptic regions (Fig. 3A) and was reduced to ∼50% of the wild-type level in the calyx (Fig. 3C,E). This change was also rescued by expression of Dα5. Thus, Hig and Dα5 mutually regulate each other to maintain their synaptic localization. This interaction between Hig and Dα5 was confirmed by immunoprecipitation experiments in which a short fragment of Hig (Fig. 3F, Hig-SF; Nakayama et al., 2016) containing at least a part of the 40–266 aa sequence used as an antigen to generate the anti-Hig antibody (Nakayama et al., 2014) was coprecipitated with FLAG-tagged Dα5, indicating that these two proteins form a protein complex (Fig. 3F). Full-length Hig was not detected in the complex, suggesting that Hig protein might need to be processed either before or after complex formation with Dα5 to execute its function in synapses.

Figure 3.

Interaction between Hig and Dα5 mediate mutual localization at synapses. A, Antibody staining for Dα5 (green) and Hig (magenta) in wild-type, hig^dd37, Dα5^CR5, and Dα5^CR5Dα7^PΔEY6 brains. Arrowheads indicate Hig staining associated with the dendrites of giant fiber neurons in which Dα7 is prominently expressed (Nakayama et al., 2016). N-Cadherin (Ncad) is used as a marker for synaptic regions. B, Dα5 staining in mushroom body calyces of wild-type, hig^dd37, and hig-rescued brains. higGFP expression is driven by the Gal4 driver elav^C155. B, C, The calyx contains a number of microglomerulus. C, Hig staining in mushroom body calyces of wild-type, Dα5^CR5, and Dα5-rescued brains. Dα5 expression is driven by the Gal4 driver elav^C155. D, Relative fluorescence intensity of Dα5 staining in mushroom body calyces of wild-type, hig^dd37, and hig-rescued brains. Values are normalized against the fluorescence intensity of N-cadherin staining. ***p < 0.001, **p = 0.0013. N ≥ 11 calyces for each genotype. Kruskal–Wallis test was followed by Dunn's multiple-comparisons test. E, Relative fluorescence intensity of Hig staining in mushroom body calyces of wild-type, Dα5^CR5, and Dα5-rescued brains. Values are normalized against the fluorescence intensity of N-cadherin staining. ***p < 0.001, **p = 0.0026. N = 12 calyces for each genotype. Kruskal–Wallis test followed by Dunn's multiple-comparisons test. F, Coimmunoprecipitation of Hig with Dα5. The soluble fraction (SFr) containing Hig was prepared from the heads of wild-type flies and added to the membrane fraction (MFr) prepared from flies expressing Dα5-FLAG or wild-type as a control, and the mixture was immunoprecipitated with anti-FLAG antibody. Anti-FLAG and anti-Hig antibodies were used to detect Dα5-FLAG and Hig, respectively, in Western blotting. Hig-FL and Hig-SF are the full-length and short forms of Hig, respectively (Nakayama et al., 2016). IP, Immunoprecipitation; In, input. The genotype of the flies expressing Dα5-FLAG is elav^C155/+ or Y; UAS-Dα5-Flag/+. G, Relative fluorescence intensity of Hig staining in mushroom body calyces of wild-type and Dα7^PΔEY6 brains. N = 17 calyces for each genotype. Mann–Whitney test, ^nsp = 0.832. H, Relative fluorescence intensity of Hig staining in mushroom body calyces of Dα5^CR5 and Dα5^CR5Dα7^PΔEY6 brains. N ≥ 12 calyces for each genotype. Unpaired t test: t = 5.44, df = 24, ***p < 0.001. I, Schematic structure of a microglomerulus in the mushroom body calyx. The presynaptic bouton (pb) is formed by the axon terminal of a projection neuron. The claw-like structures surrounding pb are dendritic terminals (d) of Kenyon cells, and nAChRs (green) are present on the membranes of the terminals. Hig (magenta) is localized in the synaptic cleft.

In a previous study, we showed that Dα7 is required for the localization of Hig at the dendrites of giant fiber neurons (Fig. 3A, arrowheads) in the brain, indicating that Dα7 maintains Hig at synaptic clefts (Nakayama et al., 2016). Hence, we investigated how Hig is distributed in the absence of both Dα5 and Dα7. Hig signals were reduced and barely detectable in most synaptic regions of Dα7 ^PΔEY6 Dα5^CR5 double-mutant brains (Fig. 3A), including mushroom body calyces (Fig. 3H), indicating that Dα5 and Dα7 are important subunits for maintaining Hig in synaptic clefts. However, Hig levels were unchanged in the calyces of Dα7^PΔEY6 mutants compared with those in the wild type (Fig. 3G). This result suggests that Hig might preferentially interact with Dα5 with a higher affinity than with Dα7 and that the interaction of Hig with Dα7 becomes experimentally apparent in the absence of Dα5.

Loss of Dα5 in hig mutants increases the synaptic levels of Dα6 and Dα7

Our previous findings indicated that the synaptic levels of Dα6 and Dα7 subunits are reduced in hig mutants (Nakayama et al., 2014). To elucidate the mechanism underlying suppression of hig by loss of Dα5 function, we examined the synaptic levels of Dα6 in the calyx of the double mutant Dα5^CR5hig^dd37. Notably, Dα6 levels, which were reduced in hig^dd37, were restored to or even above the wild-type level in double-mutant brains (Fig. 4A,E), although alteration in Dα7 levels was not significantly detected (Fig. 4F), possibly because of its faint signal in the tissue. Furthermore, both Dα6 and Dα7 were detected at higher levels in Dα5^CR5 single mutants than in the wild type (Fig. 4A,B,E,G). Therefore, Dα5 inhibits accumulation of Dα6 and Dα7 at synapses in the wild-type brain. By contrast, the synaptic levels of Dα5 in the calyx were maintained at the wild-type level in both Dα6^DAS1 (Fig. 4C,H) and Dα7^PΔEY6 (Fig. 4I) flies. Moreover, the Dα5 levels in hig^dd37 were indistinguishable from those in both Dα6^DAS1hig^dd37 (Fig. 4C,H) and Dα7^PΔEY6hig^dd37 (Fig. 4D,J) flies. These data suggest that, among the subunits, Dα5 serves a characteristic function in reducing the levels of nAChR.

Figure 4.

Synaptic levels of Dα6 and Dα7 are increased by the loss of Dα5 in calyces. A–D, Immunostaining in calyces of the indicated fly lines, using antibodies against Dα6 (A), Dα7 (B), and Dα5 (C, D). E–J, Quantification of staining signals for Dα6 (E), Dα7 (F, G), and Dα5 (H–J) in calyces of mutant flies. ***p < 0.001. ^nsp shown below is a p value designated as not significant. Ordinary one-way ANOVA, followed by Tukey's multiple-comparisons test when applicable, for E–H and J. Unpaired t test for I. N ≥ 8 calyces for each genotype. E, **p = 0.0054, ^nsp = 0.722, F_(3,48) = 43.9, p < 0.0001. F, F_(2,31) = 1.38, p = 0.266. G, **p = 0.0080; ^nsp = 0.201, 0.582, and 0.0823 for wild type versus Dα6^DAS1, wild type versus Dα7^PΔEY6, and Dα5^CR5 versus Dα6^DAS1, respectively. F_(3,40) = 12.46, p < 0.0001. H, ^nsp = 0.143 and 0.120 for wild type versus Dα6^DAS1, and hig^dd37 versus hig^dd37Dα6^DAS1, respectively. F_(3,49) = 52.2, p < 0.0001. I, t = 0.835, df = 18, ^nsp = 0.415. J, ^nsp = 0.729. F_(2,36) = 27.98, p < 0.0001. K, L, Increased longevity of hig^dd37 flies on overexpression of Dα6 (K) or Dα7 (L). The genotype of the flies used in this experiment was elav^C155/Y; hig^dd37; UAS-Dα6 or UAS-Dα7/+. ***p < 0.001 Mann–Whitney test. N ≥ 30 flies for each genotype. M, N, Total levels of Dα6 mRNA and Dα7 mRNA in Dα5 mutants relative to the wild type. Quantitative RT-PCR was performed to quantify Dα6 (M) and Dα7 (N) mRNAs extracted from the heads of wild-type and Dα5 flies. mRNA levels are normalized against the corresponding level of GAPDH mRNA. M, ^nsp > 0.999; N, ^nsp = 0.700; Mann–Whitney test.

The restored longevity and locomotion of suppressed mutants could be ascribed to recovery of the synaptic levels of nAChR subunits, including Dα6 and Dα7. To test this possibility, we overexpressed Dα6 or Dα7 in hig^dd37 mutants using a pan-neuronal elav-Gal4 driver and detected partial rescue of the longevity of hig mutants (Fig. 4K,L). These data suggest that one of the causes of hig phenotypes is a reduction in synaptic levels of Dα6 and Dα7 nAChR subunits, and that the suppression of hig by loss of Dα5 function is at least partly a result of the recovery of these subunit levels.

To examine whether the Dα5-mediated regulation of nAChR subunit levels occurs at a transcriptional process, we compared each level of Dα6 and Dα7 mRNAs in both wild type and Dα5^CR5 mutants using qPCR experiments. The data showed no significant alteration in each mRNA level by loss of Dα5 (Fig. 4M,N), suggesting that the increases in Dα6 and Dα7 levels in Dα5 mutants were caused by a post-transcriptional process.

Dα5, Dα6, and Dα7 have distinct properties for Hig interaction or synaptic localization

Although Dα5, Dα6, and Dα7 share a high similarity in amino acid sequences, only the loss of Dα5 suppresses the hig phenotypes. Our data suggest that this specific genetic interaction is based on the characteristic properties of Dα5 protein in terms of the inhibitory effects on synaptic levels of other subunits as well as the interaction with Hig. To further clarify these properties of Dα5 compared with those of Dα6 and Dα7, we overexpressed these subunits to examine whether the levels of Hig and the endogenous subunits are altered in the synaptic regions.

When Dα5 was overexpressed in mushroom body neurons, the synaptic level of Hig was elevated in the microglomeruli of calyx (Fig. 5A,B), consistent with the mutual requirement of Dα5 and Hig for synaptic localization. However, in calyces with overexpressed Dα5, the synaptic level of Dα6 was reduced, whereas no change in Dα7 level was detected (Fig. 5A,D,E). Next, when Dα6 was overexpressed, both Hig and Dα5 levels were significantly reduced (Fig. 5A–C). In this experiment, the reduction in the levels of Dα5 that positively regulates Hig localization could offset the ability of Dα6 to increase the Hig level. We then overexpressed Dα6 in Dα5^CR5 mutants but still detected a decrease in the Hig levels (Fig. 6A,B). Therefore, Dα6 can reduce synaptic Hig levels in a manner independent of Dα5. This negative regulation of Hig by Dα6 might arise from their physical interaction in the synaptic clefts or, more indirectly, from inhibition of the accumulation of Hig-interacting proteins other than Dα5. Finally, overexpression of Dα7 decreased the levels of Dα5 and Dα6, but did not significantly change the level of Hig (Fig. 5A–D). However, when Dα7 was overexpressed in the Dα5^CR5 mutant background to eliminate the effect of Dα5 on Hig levels, the level of Hig was higher than that in Dα5^CR5 mutants (Fig. 6C,D). Collectively, these results indicate that both Dα7 and Dα5 positively regulate Hig to increase its synaptic level in a dosage-dependent manner, but that Dα6, when overexpressed, negatively regulates Hig levels. In addition, the finding that the overexpression of Dα5, Dα6, or Dα7 decreases the levels of other subunits suggests that there might be an upper limit of postsynaptic capacity for recruiting nAChRs into the membrane.

Figure 5.

Overexpression of Dα5, Dα6, and Dα7 affects synaptic levels of Hig and the nAChR subunits. Each subunit was expressed in Kenyon cells by the Gal4 line OK107. A, Immunostaining signals for Hig, Dα5, Dα6, and Dα7 in the calyces of mushroom bodies when Dα5, Dα6, or Dα7 was overexpressed. B–E, Quantification of the signals for Hig (B), Dα5 (C), Dα6 (D), and Dα7 (E) appearing in A. Significant differences are presented relative to the wild type. ***p < 0.001. One-way ANOVA followed by Dunnett's multiple-comparisons test. N ≥ 10 calyces for each genotype. B, ^nsp = 0.207; F_(3,58) = 64.26, p < 0.0001. C, **p = 0.0014; F_(3,58) = 467.3, p < 0.0001. D, *p = 0.0342 and 0.0138 for OK107>Dα5 and OK107>Dα7, respectively; F_(3,38) = 853.5, p < 0.0001. E, ^nsp = 0.778 and 0.778 for OK107>Dα5 and OK107>Dα6, respectively; F_(3,38) = 248.7, p < 0.0001.

Figure 6.

The extracellular domains of Dα5, Dα6, and Dα7 have distinct effects on the synaptic levels of Hig. Each protein was expressed by Gal4-OK107 in Kenyon cells of Dα5^CR5 mutants. A, C, Immunostaining with antibodies for Hig (αHig), Dα5 (αDα5), and N-cadherin (αNcad) in calyces of the mushroom body of the indicated fly strains. The anti-Dα5 antibody recognized the cytoplasmic loop in the 5 Tm sequences of 6Ex5Tm and 7Ex5Tm. See E. Scale bar, 10 µm. B, Quantification of staining signals appearing in A. ***p < 0.001, ^nsp > 0.999. N ≥ 12 calyces were examined. Ordinary one-way ANOVA followed by Tukey's multiple-comparisons test: F_(4,75) = 269.8, p < 0.0001. D, Quantification of staining signals appearing in C. ***p < 0.001, *p = 0.0263, ^nsp = 0.458 and 0.189 for Dα5^CR5, OK107>Dα7 versus Dα5^CR5, OK107 > 5Ex7Tm; and Dα5^CR5, OK107 > 5Ex7Tm versus Dα5^CR5, OK107 > 7Ex5Tm, respectively. N ≥ 12 calyces were examined. One-way ANOVA followed by Holm–Sidak's multiple-comparisons test: F_(4,71) = 29.83, p < 0.0001. E, Schematic structure of chimeric proteins constructed by swapping the extracellular domains of Dα5, Dα6, and Dα7 subunits. Each antibody raised for these subunits recognizes the cytoplasmic loop in the respective Tm sequences.

These data further raise the question of why only Dα5 mutations suppress the phenotypes of hig mutants, although Dα5 and Dα7 share the ability to positively regulate Hig levels. We hypothesized that Dα5 might more effectively cause the intracellular trafficking of nAChRs, leading the receptors to an intracellular degradation pathway or to a reduction of the receptor levels via destabilization on the synaptic membranes by an unknown mechanism. When Dα5 was overexpressed in the mushroom body MB calyx of wild-type flies, Dα5 signals associated with postsynaptic terminals in a microglomerular structure (Fig. 3I, schematic patterns of nAChR and Hig in a microglomerulus) simply increased the level and displayed a continuous doughnut-like pattern (Figs. 7A,8A). However, when Dα5 was overexpressed in hig mutants, the Dα5 signals were discontinuous, failing to delineate the microglomerular structures, and typically remained as punctata in the entire calyces (Figs. 7B,8A). Notably, such Dα5 signals were partly colocalized to the puncta of early endosome marker YFP-Rab5 (Fig. 7B). However, the Rab5 puncta did not appear in hig mutants unless Dα5 was overexpressed (Fig. 7B–D). These results suggest that Dα5 induces intracellular trafficking and internalizes itself into postsynaptic terminals. By contrast, the overexpression of either Dα6 or Dα7 in hig mutants increased their signal levels, displaying doughnut-like patterns, which showed forms with less or no disruption (Fig. 8B,C). Therefore, Dα5 has a distinguishing feature in terms of subcellular localization at synapses, which is pronounced in the absence of Hig. These data suggest that Dα5 can more frequently internalize nAChR, thereby restricting receptor levels on the postsynaptic membranes.

Figure 7.

Overexpressed Dα5 changes its subsynaptic distribution in hig mutants. A, B, Distribution of overexpressed Dα5 and YFP-Rab5 in the mushroom body calyx of wild-type (A) and hig^dd37 (B) flies. Enlarged views of the dashed squares are shown in the insets. The brackets in A and the arrowheads in B in the insets indicate the positions of microglomeruli and punctate signals, respectively. Scale bar: inset, 2 µm. C, D, Distribution of overexpressed YFP-Rab5 in the calyx of wild-type (C) and hig^dd37 (D) flies. Distribution of YFP-Rab5 was undiscernible between hig mutants and the wild type. OK107-Gal4 was used to drive overexpression in A–D. Scale bars: C, D, 10 µm.

Figure 8.

Overexpressed Dα5, but not Dα6 or Dα7, exhibits disrupted localization in the postsynaptic terminals of hig^dd37 mutant calyces. A–C, Distribution of Dα5 (A), Dα6 (B), and Dα7 (C) overexpressed by OK107-Gal4 in the calyces of wild-type (top) and hig^dd37 (bottom) flies. Calyces were immunostained with antibodies against Hig (left) and either of the subunits (middle and right). The images in the small squares in the middle panels are enlarged in the right panels. Brackets and arrowheads indicate microglomerular structures and punctate patterns, respectively.

To further confirm this possible role of Dα5, we conducted biochemical experiments to investigate the amounts of Dα6 in the extracts of mutant flies. We initially compared the total amount of Dα6 in the extracts prepared from the whole adult brains of Dα5 mutants and wild types, and also similarly examined the amount of Dα6 in Dα5Dα7 double mutants relative to that in Dα7 mutants (Fig. 9A,B). No significant difference was observed in each comparison, although the Dα5 mutation introduced into wild-type or Dα7 mutants tended to cause an increase in Dα6 levels, possibly reflecting the elevated levels in the synaptic regions, as observed in the immunohistochemical experiments. We next compared the amount of Dα6 in the membrane fraction prepared from adult heads of wild-type and Dα5 mutant flies. Dα6 was present at a significantly higher level in Dα5 mutants than in the wild-type, even after normalization to the total amount of Dα6 in the whole-brain extract of each fly line (Fig. 9C,D). These data are consistent with the idea that Dα5 influences trafficking of nAChR in postsynapses. Moreover, Dα6 levels, which tended to decrease in the membrane fraction of Dα7 mutants relative to the wild type, increased significantly in Dα5 Dα7 double mutants (Fig. 9C,D), again supporting the role of Dα5 in nAChR trafficking. Collectively, these data suggest that Dα5 controls nAChR levels by facilitating the trafficking event and that Hig blocks the runaway function of Dα5.

Figure 9.

Dα6 levels are increased in the membrane fractions when Dα5 is impaired. A, Levels of Dα6 protein in whole-brain extracts of wild-type flies and flies harboring the indicated mutations. The multiple bands corresponding to Dα6 disappeared in extracts of Dα6 mutants. B, Quantification of Dα6 signals detected in A. Each value is normalized against the corresponding level of β-actin protein. ^nsp = 0.690 and 0.998 for wild type versus Dα5, and Dα5 versus Dα5Dα7, respectively. Repeated-measures one-way ANOVA with the Geisser–Greenhouse correction, followed by Tukey's multiple-comparisons test: F_(2.28,11.4) = 29.5, p < 0.0001. C, Levels of Dα6 in the membrane fractions prepared from the head extracts of wild-type and mutant flies. Dα6 was immunoprecipitated and detected with anti-Dα6 antibody. N-cadherin (Ncad), a synaptic membrane protein, was used as a control to compare the amounts of membrane fractions input for immunoprecipitation. Dα6 was detected only in the immunoprecipitated samples, but not in the membrane fractions before immunoprecipitation. D, Quantification of Dα6 in the membrane fractions in C. The values were normalized against the amount of N-cadherin and the Dα6 levels in the whole-brain extracts of the respective lines in B. *p = 0.0326 and 0.0489 for wild-type versus Dα5, and Dα7 versus Dα5 Dα7, respectively. Repeated measures one-way ANOVA with the Geisser–Greenhouse correction, followed by Tukey's multiple-comparisons test: F_(1.12,3.37) = 95.15, p = 0.0013.

The complex function of Dα5 in the regulation of nAChR levels is mediated by both the N-terminal extracellular and intracellular domains

Because Hig is a synaptic cleft protein, we considered it likely that Hig interacts with the extracellular domains of Dα5 and Dα7. To test this idea, we exchanged the extracellular domain of Dα5 with that of Dα6 to construct chimeric proteins 6Ex5Tm and 5Ex6Tm, which contained the extracellular domain (Ex) and the transmembrane and cytoplasmic domains (Tm) of the respective subunits (Fig. 6E). When 5Ex6Tm was expressed in Dα5^CR5 mutants, the synaptic level of Hig was increased (Fig. 6A,B), as observed when Dα5 was overexpressed (Fig. 5A,B). By contrast, when 6Ex5Tm was expressed in Dα5^CR5 mutants, Hig levels were rather reduced at the synaptic region (Fig. 6A,B). This change was also observed as a result of Dα6 overexpression in wild type or Dα5^CR5 mutants (Figs. 5A,B,6A,B). These data indicate that the extracellular domain, but not the transmembrane and cytoplasmic domains, of Dα5 positively regulates Hig levels and that the extracellular domain of Da6 negatively affects Hig levels. In similar experiments, 5Ex7Tm and 7Ex5Tm overexpressed in Dα5^CR5 mutants increased the synaptic levels of Hig (Fig. 6C,D), as did both Dα5 (Fig. 5A,B) and Dα7 in Dα5^CR5 mutants (Fig. 6C,D). These results are consistent with the idea that the extracellular domains of Dα5 and Dα7 function in maintaining or increasing Hig levels. This function of the Dα7 extracellular domain is supported by the results showing that the expression of 7Ex5Tm and 6Ex5Tm in Dα5^CR5 mutants increased and decreased Hig levels, respectively. Thus, the extracellular domain of Dα7 as well as that of Dα5 positively regulates Hig levels, serving to maintain Hig at synaptic clefts.

To identify the domains of Dα5 required for control of trafficking, we examined the distributions of chimeric proteins overexpressed in MB calyces. When overexpressed in the wild-type background, 5Ex6Tm, 5Ex7Tm, and 7Ex5Tm, all of which exhibited a positive interaction with Hig by the extracellular domains, were distributed normally along the microglomerular structures (Fig. 10A); however, only the signal of 6Ex5Tm, which exhibited negative interaction with Hig, presented a punctate pattern at a reduced overall level (Fig. 10A), as observed for Dα5 overexpressed in hig mutants (Figs. 7B,8A). Moreover, when overexpressed in hig mutants, both 7Ex5Tm and 6Ex5Tm exhibited similar punctate patterns (Fig. 10B), recapitulating the distribution of Dα5 in hig mutants; by contrast, signals of 5Ex6Tm and 5Ex7Tm in hig mutants displayed a doughnut-like pattern associated with microglomerular structures (Fig. 10B), as observed for Dα6 and Dα7 in hig mutants (Fig. 8B,C). These results indicate that the chimeric proteins are localized in discontinuous or punctate patterns when they contain 5Tm and do not positively interact with Hig. Although the artificial construction of chimeric subunits might result in unexpected misfolding of the proteins, which can cause the abnormal trafficking, the altered distributions of 7Ex5Tm in hig mutants relative to the wild type were similar to those of the intact Dα5. In addition, the punctate distribution of 6Ex5Tm in both wild type and hig mutants resembled that of the intact Dα5 in hig mutants. This similarity suggests that the 5Tm in the chimeric proteins present the authentic function of the Tm domain in Dα5 and that the punctate subcellular distribution is not caused by the artificial outcome of the chimeric construction. We therefore conclude that the protein portion spanning the transmembrane and cytoplasmic domain of Dα5 is responsible for control of Dα5 trafficking in the postsynaptic terminals.

Figure 10.

The Tm region of Dα5 contributes to the disrupted localization of nAChR and lethality. A, B, Distribution of chimeric proteins in mushroom body calyx of wild-type (A) and hig mutant (B) flies. The antibodies used for staining are shown in each panel. An enlarged view of each square is indicated on the right side. Brackets and arrowheads indicate microglomerular structures and punctate staining labeled by chimeric proteins, respectively. C, Mean longevity of flies overexpressing Dα5, Dα6, or chimeric proteins under the control of OK107-Gal4. ***p < 0.001, **p = 0.0061, *p = 0.0349, and ^nsp > 0.999, ^nsp = 0.320, and ^nsp = 0.0676 for OK107 versus OK107>Dα5, OK107 versus OK107 > 7Ex5Tm, and OK107>Dα5 versus OK107>Dα6, respectively. N ≥ 36 flies for each genotype. Kruskal–Wallis test followed by Dunn's multiple-comparisons test: p < 0.0001.

Notably, although overexpression of Dα5 or 7Ex5Tm in the wild type, each of which interacts positively with Hig by the extracellular domain, did not affect longevity (Fig. 10C), overexpression of 6Ex5Tm, which lost a positive interaction with Hig, exerted a lethal effect and greatly decreased the longevity of survivors, even in the presence of Hig. However, the overexpression of Dα6 did not reduce longevity (Fig. 10C). Therefore, 5Tm is responsible for the lethality when the interaction of the associated extracellular domains with Hig is compromised. Collectively, the data suggest that the contribution of endogenous Dα5 to the lethality of hig mutants is at least partly attributable to the runaway function of the Tm portion. This portion can normally serve for the regulation of receptor trafficking given that its associated extracellular domain maintains a positive interaction with Hig at synaptic clefts.

Discussion

Because the cholinergic synapse is one of the major synapses that transmit neural information throughout the CNS, many studies have sought to reveal the physiological function and structure of nAChR. However, very little is known about how nAChR levels are regulated on synaptic membranes, and the question remains a challenge for the field. Obstacles to research efforts include the difficulty of cell surface expression of pentamer receptors containing a particular set of subunits in cultured heterologous cells, as well as the challenge of analyzing a receptor whose levels are influenced by extracellular matrix in the synaptic clefts. Our in vivo study circumvented these problems, uncovering the regulatory scheme that determines nAChR levels.

A genetic modifier screen is a potent means for identifying factors that interact with a gene of interest. We successfully identified Dα5 as a suppressor gene of hig. Introduction of Dα5 mutations on a chromosome harboring defect in hig restored locomotor activity, longevity, and synaptic levels of Dα6, all of which were compromised in hig mutants. RNAi experiments performed for all 10 subunits encoded on the Drosophila genome indicated that only knockdown of Dα5 could rescue the hig phenotype. This observation implies that the Dα5 activity contributes to lethality in the absence of Hig, whereas Hig prevents runaway activity of Dα5 by tethering it to the synaptic cleft matrix or stabilizing it on the postsynaptic membrane. Although Hig might intracellularly help transport of Dα5 to postsynaptic membranes, Hig ectopically expressed in glia is secreted and localized to synaptic areas, and can rescue the reduced longevity of hig mutants (Nakayama et al., 2016). This suggests that Hig functions extracellularly at synaptic clefts rather than within the Dα5-expressing cells. Dα5 is a functional subunit that constitutes a pentameric nAChR, as evidenced by the observation that its homopentamer as well as its heteropentamers combined with other subunits acts as a cation channel in Xenopus oocytes (Lansdell et al., 2012). However, Dα5, like Dα6 and Dα7, is apparently dispensable for viability, at least under laboratory conditions, as homozygotes for Dα5-null mutations can be maintained through generations, presumably because of compensation by other subunits. These data suggest that Dα5 coevolved with Hig as a pair of synaptic proteins that cooperate to regulate the synaptic levels of nAChR. Given that both Dα5 and Hig are widely distributed in similar patterns in the entire adult brain, these molecules may together serve as general controllers of nAChR levels and of brain functions produced by circuits comprising cholinergic neurons.

Based on the data presented in this study, we propose a model to explain how nAChR levels are regulated (Fig. 11A–D). Although there is a possibility that Dα5 induces degradation of nAChR on the postsynaptic membranes, the data in the present study suggest that Dα5 is involved in the control of receptor trafficking. In the presence of Hig at synaptic clefts, Dα5 undergoes a basal level of endocytosis that normally regulates the surface level of nAChRs (Fig. 11A). In the absence of Hig, Dα5 induces excess endocytosis of nAChR, thereby decreasing surface receptor levels (Fig. 11B). The endocytosis might only locally affect the Dα5-containing receptors or might occur in a wide area on the postsynaptic membrane to involve surrounding receptors that do not contain Dα5. In addition, we cannot exclude the possibility that Dα5 either prevents the endosomes containing Dα5 from entering a recycling pathway or causes them to take a degradation pathway. The intracellular trafficking prominently caused by Dα5 is likely key to the suppression of hig and to the regulation of nAChR levels. Among the Dα5, Dα6, and Dα7 subunits, both Dα5 and Dα7 positively interact with Hig. Therefore, in the presence of Hig, Dα7, together with Dα5, reinforces the maintenance of nAChR on the postsynaptic membrane (Fig. 11A). Dα6 may be neutral in the regulation of nAChR trafficking. Thus, these three subunits, despite their sequence similarities, have distinct properties in regard to the determination of nAChR level on postsynaptic membranes. Although we here propose a model of Dα5-dependent trafficking of nAChRs, the regulatory system of nAChR localization might be more complex. The reduction in levels of endogenous nAChR subunits caused by the overexpression of other subunits suggests that there might be an upper limit of nAChR amounts on the postsynaptic membranes. This limit might serve to maintain the extent of neurotransmission within a certain range.

Figure 11.

Model of the regulation of nAChR levels by Hig and Dα5. A, In the wild type, Hig tethers nAChRs to the synaptic cleft by directly or indirectly interacting with Dα5 and Dα7 to restrict Dα5-dependent endocytosis when the receptors contain those subunits. Endocytosis of nAChRs occurs at a basal level. nAChR is schematically shown here as a heteropentamer containing Dα5, Dα6, and Dα7. However, a fraction of nAChRs at synapses might exclude some or all of these subunits or be their homopentamers. B, In hig mutants, Dα5-containing receptors are excessively endocytosed, which contributes to lethality. Although the figure illustrates that Dα5-dependent endocytosis only affects the single nAChR containing Dα5, the endocytosis might widely occur on the postsynaptic membrane and involve the surrounding receptors that do not contain Dα5. This wide-range endocytosis might also occur in the wild type. C, D, In Dα5 (C) and hig Dα5 (D) mutants, Dα5-dependent endocytosis does not occur, because the sequestering subunit Dα5 is missing.

nAChR subunits are complex proteins with multiple domains including an extracellular ligand-binding site, the transmembrane sequences, and the cytoplasmic loop that potentially binds to accessory proteins. In previous studies, the cytoplasmic domains of particular subunits were reported to mediate the subcellular localization or surface expression of the receptors. In chick, homopentameric α7-nAChRs, which are largely confined to perisynaptic regions of adult ciliary ganglion neurons in vivo (Jacob and Berg, 1983; Shoop et al., 1999), are targeted to the postsynaptic membrane when the cytoplasmic loop of the α7 is replaced with the homologous region of α3 (Williams et al., 1998). The binding of Bcl-2 to an intracellular motif of α7 upregulates the surface expression of the receptor (Dawe et al., 2019). In addition, the lysosomal-associated membrane protein LAMP5 and chaperone-related protein SULT2B1 enhance surface trafficking of α6-containing nAChRs in cultured cells (Gu et al., 2019). Thus, subsets of nAChR subunits likely have specific functions beyond their contributions to ion-channel physiology. The present study revealed that Dα5 in the Drosophila brain is a notable regulatory subunit with dual functional regions involved in intracellular trafficking as well as extracellular Hig interaction. To understand the molecular mechanism responsible for the trafficking that controls receptor levels and brain functions, future studies should identify the proteins that bind to the intracellular regions of Dα5.

Extracellular interactions between neurotransmitter receptors and secreted proteins have been reported in a few animal species. Ly6/neurotoxin 1 superfamily proteins GPI (glycosylphosphatidylinositol) anchored to organellar membranes and to the outer leaflet of the cell surface regulate the surface levels of nAChRs, as reported for Ly6 in vertebrates (Wu et al., 2021) and Qvr/SSS in Drosophila (Wu et al., 2014). However, these proteins function in reducing the levels of nAChRs, opposing a number of secretory proteins including Hig, which maintains the levels of nAChRs or other neurotransmitter receptors. For instance, in Caenorhabditis elegans, synaptic clustering of l-AChR relies on extracellular scaffold proteins assembled in the synaptic cleft of the neuromuscular junction (Gally et al., 2004; Gendrel et al., 2009; Rapti et al., 2011). In the vertebrate CNS, Lgi1, C1q family proteins, and Nptxs recruit and cluster ionotropic glutamate receptors at postsynaptic sites (O'Brien et al., 1999, 2002; Xu et al., 2003; Matsuda et al., 2010; Uemura et al., 2010; Yuzaki, 2018), and the CCP-containing protein SRPX2 regulates the formation of excitatory synapses (Sia et al., 2013). These results, together with our findings, indicate that extracellular proteins are important for establishing the surface expression of the receptors and presumably reflect expansion in the complexity of synapses along with an increase in their number over the course of evolution. In human and mouse brains, a number of secreted proteins (Human Protein Atlas, Mouse Brain Atlas), including CCP repeat-containing proteins such as Hig and Hasp, might constitute the synaptic cleft matrix and positively interact with specific nAChR subunits to modulate receptor trafficking or prevent their runaway activity. Defects in the interaction can cause synaptic dysfunction or degeneration. Further studies will elucidate the general mechanisms underlying the regulation of nAChR levels and might provide insight into the development of therapies for the disorders associated with the loss of nAChR.