Introduction

Human MPV17 is a putative channel-forming protein of the inner mitochondrial membrane and it is involved in mitochondrial deoxynucleotide homeostasis1,2,3. MPV17 mRNA is widely expressed in tissues, including the liver, kidney, muscle, lung, brain, and heart (https://www.proteinatlas.org/ENSG00000115204-MPV17, Human Protein Atlas proteinatlas.org). Autosomal recessive mutations in the MPV17 gene cause MPV17-related hepatocerebral mitochondrial DNA depletion syndrome (MDDS), including Navajo neurohepatopathy, which typically has hepatic and gastrointestinal manifestations, metabolic derangements, and neurological involvement, including developmental delay, hypotonia, microcephaly, and motor and sensory peripheral neuropathy; most patients die by childhood4. Patients with Navajo neurohepatopathy show more severe symptoms and often die in infancy or early childhood due to progressive liver dysfunction associated with depletion of mitochondrial DNA (mtDNA)5. Autosomal recessive mutations in the MPV17 gene have also been identified in patients with Charcot-Marie-Tooth (CMT) disease, the most common inherited motor and sensory peripheral neuropathy6,7,8.

To elucidate the mechanism of the pathogenesis of MPV17-related hepatocerebral MDDS, model organisms such as MPV17 knockout (KO) mice and KO zebrafish have been established and analyzed. MPV17 KO mice show significantly reduced longevity, lower body weight, and hair graying, with defects in the kidney, skin, and peripheral nervous system9,10,11,12. MPV17 deficiency in mice leads to a reduced deoxynucleotide pool resulting in mtDNA depletion in the liver, but not in the kidney or brain2. In zebrafish, diverse phenotypes have been reported in two mpv17 mutants (roy orbison (roy) and transparent (tra)), and Mpv17 KO zebrafish generated by the CRISPR/Cas9 system13,14,15,16,17. While roy and tra mutants show only transparent phenotypes with a loss or strong reduction of iridophores and melanophores, Mpv17 KO zebrafish show reduced longevity and defects in muscle, liver, and energy supply during development in addition to a strong reduction of iridophores and melanophores13,14,15,16,17. However, in these studies, the neurological symptoms have not been fully analyzed, albeit human patients with MPV17-related hepatocerebral MDDS commonly show them. Drosophila models have been successfully used to study various neurological disorders18,19,20,21,22. In the present study, we establish a Drosophila model to analyze the neurological pathogenesis of MPV17-related hepatocerebral MDDS.

Results

CG11077 is a Drosophila ortholog of human MPV17

In the Drosophila genome, there are ten MPV17/PMP22 family genes, including nine uncharacterized genes named with CG numbers. To identify the Drosophila ortholog of human MPV17, the amino acid sequence similarity of MPV17/PMP22 family proteins among model organisms and the human MPV17 family was analyzed (Fig. 1A). Because the phylogenetic tree clearly shows that CG11077 is the best candidate for a Drosophila ortholog of human MPV17, we named CG11077 Drosophila Mpv17 (dMpv17). The alignment of Mpv17 proteins among model organisms indicates that these proteins, including the MPV17/PMP22 domain marked by underlining, are well conserved (Fig. 1B). The mutations identified in patients with MPV17-related hepatocerebral MDDS (R50Q)4 and CMT (P98L, R41Q)6,7,8, which both show peripheral neuropathy, are identical or similar between Drosophila and humans (Fig. 1B). Because these mutations have been identified as recessive mutations in patients and predicted to be hypomorphic mutations6,8, we knocked down dMpv17 specifically in the nervous system during Drosophila development to elucidate the effect of dMpv17 depletion on nerve cells.

Figure 1
figure 1

CG11077 is an ortholog of human MPV17. (A) A phylogenetic tree was constructed with amino acid sequences of MPV17/PMP22 family proteins, human (MPL, FKSG24, PXMP22, and MPV17), S. cerevisiae SYM1, D. melanogaster (CG7970, CG14777, CG144778, CG32262, CG32263, CG1662, CG12355, and CG11077), C. elegans Mpv17-like, and Mpv17s (mouse, rat, zebrafish, and X. laevis). The sequence similarity of MPV17/PMP22 family proteins among model organisms and the human MPV17 family was analyzed. (B) Multiple alignment of amino acid sequences of Mpv17 among model organisms and human, including CG11077, was carried out with the Clustal Omega multiple sequence alignment program. Amino acids conserved among all organisms are marked by asterisks and those conserved between D. melanogaster and others are highlighted in gray. The black and white arrowheads indicate the amino acids mutated in MPV17-related hepatocerebral MDDS (R50Q) and CMT (R41Q and P98L) patients, respectively. Underlining indicates the amino acid sequences conserved between PMP22 and MPV17 (Mpv17/PMP22: IPR007248).

Full size image

Phenotypes of neuron-specific dMpv17 knockdown flies

To investigate the effect of dMpv17 depletion on nerve cells, we knocked down dMpv17 in the nervous system by using a pan-neuronal GAL4 driver, embryonic lethal abnormal vision (elav)-GAL4. The knockdown efficiency of dMpv17 mRNA in the larval central nervous system (CNS) of two independent knockdown lines (elav > dMpv17-IR42671 and elav > dMpv17-IR56889) was determined by RT-qPCR. dMpv17 mRNA levels were reduced to 46% and 54%, respectively, in elav > dMpv17-IR42671 and elav > dMpv17-IR56889 compared to the control elav > GFP-IR (Fig. 2A). We thus confirmed the effective knockdown of dMpv17 in two independent RNAi lines. With these lines, we examined whether dMpv17 knockdown flies showed neurological defects, such as peripheral neuropathy, seizures, or cognitive disability, that correspond to neurological symptoms of human MPV17-related MDDS.

Figure 2
figure 2

Neuron-specific dMpv17 knockdown larvae show neurological-defect phenotypes. (A) Total RNA was extracted from the CNS of third instar larvae carrying elav > GFP-IR (UAS-GFP-IR/+; elav-GAL4/+;+), elav > dMpv17-IR42671 (elav-GAL4/UAS-dMpv17-IR42671;+), and elav > dMpv17-IR56889 (elav-GAL4/UAS-dMpv17-IR56889;+). The mRNA levels of dMpv17 and G6PD in each strain were detected by qRT-PCR. G6PD was used as an internal control. ***P < 0.01, statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 3. (B, C) The learning ability of larvae carrying elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 was tested in the odor-taste learning assay. AM preference (B) and larval learning index (C) were quantified. *P < 0.05, ***P < 0.001, using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 3. (D, E) Larval locomotor ability was tested in the crawling assay with elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889. (D) The average crawling speed was quantified. Statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 25–31. (E) Typical trajectories of larvae with each genotype.

Full size image

Because MPV17-related MDDS patients often show cognitive impairment4, we investigated the learning ability of dMpv17 knockdown larvae by testing whether the larvae could learn the relationship between reward and odor. Whereas n-amyl acetate (AM), a compound with a fruity odor, was still attractive to dMpv17 knockdown larvae, as it is to control larvae (Fig. 2B), the larvae of both knockdown lines showed a significantly diminished ability to learn the relationship between AM and the reward, sucrose (Fig. 2C). These results indicate that dMpv17 knockdown larvae had a significant learning defect.

Because larvae with impairment of the nervous system often exhibit defects in locomotor behavior23, we next evaluated larval locomotor ability by carrying out a crawling assay with wandering-stage larvae. Both knockdown lines showed a significant decrease in average crawling speed (Fig. 2D) and abnormally shortened trajectories (Fig. 2E). In addition to larval locomotor ability, we also examined the locomotor ability of adult flies by performing a climbing assay with elav-GAL4. However, no apparent defects in climbing ability were observed (Fig. S1). We next performed a bang-sensitivity assay with adult flies to investigate whether dMpv17 knockdown induced seizures, but dMpv17 knockdown flies showed no seizures after exposure to mechanical stress by vortexing (Fig. S2). We also examined the longevity of the neuron-specific dMpv17 knockdown flies. Although the neuron-specific dMpv17 knockdown male flies survived longer than elav > GFP-IR flies (Fig. S3), the conclusion has to be waited for further analysis. For some reason, the longevity of the control elav > GFP-IR male flies is much shorter than other control such as Act5C >  + (data not shown).

dMpv17 knockdown impairs mitochondrial function in the CNS

Mammalian Mpv17 encodes an inner mitochondrial membrane protein involved in dNTP uptake1,2,24. In addition, in both MPV17 KO mice and human MPV17-related MDDS patients, the mtDNA copy number is significantly reduced2,10,25. Therefore, we measured the relative mtDNA copy number in the CNS of neuron-specific dMpv17 knockdown larvae as indicated by the ratio of mitochondrial-encoded 16s rRNA to nuclear-encoded RpL32. There were statistically significant decreases in relative mtDNA copy number to 22% and 27%, respectively, in elav > dMpv17-IR42671 and elav > dMpv17-IR56889 compared to the control elav > GFP-IR (Fig. 3A).

Figure 3
figure 3

A dMpv17 knockdown induced mitochondrial dysfunction in the larval CNS. (A) The relative mtDNA levels in the larval CNS extract of elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 larvae were quantified. The relative levels of 16s rRNA (for mtDNA) and RpL32 (for nuclear DNA) were measured by qRT-PCR. **P < 0.01, statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 3. (B) The ATP levels in the larval CNS extracts of elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 larvae were measured with Cell titer-Glo assay. *P < 0.05, statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 5. (C) The lactate level in the larval CNS extract of elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 larvae were measured with the Lactate-Glo assay. **P < 0.01, *P < 0.05, statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 3. (D) Pyruvate levels in the CNS extracts of elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 larvae were measured with a pyruvate assay kit. **P < 0.01, *P < 0.05, statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs elav > GFP-IR. n = 3. (E) Third instar larval CNSs were stained with CM-H2DCFCA for the detection of ROS. Representative ROS images of the CNS of elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 are shown. Scale bar, 100 μm. (F) The fluorescence intensity in the CNS area enclosed by the white line was quantified. **P < 0.01, statistical analysis was performed using Student’s t-test vs elav > GFP-IR. n = 5.

Full size image

Because a decrease in the copy number of mtDNA leads to functional impairment of mitochondria26,27, we measured the amount of ATP in the CNS to evaluate the level of mitochondrial activity. Relative ATP levels were reduced to 68% and 65%, respectively, in elav > dMpv17-IR42671 and elav > dMpv17-IR56889 compared to the control elav > GFP-IR (Fig. 3B). The reduction in ATP levels in neuron-specific dMpv17 knockdown flies indicates that dMpv17 knockdown resulted in perturbation of the mitochondria. In addition to the defect in the mitochondria, high levels of reactive oxygen species (ROS) have been detected in the glomeruli and embryonic fibroblasts from Mpv17−/− mice28,29. We therefore measured ROS in the CNS of dMpv17 knockdown larvae and found that ROS levels in both knockdown lines were significantly increased (Fig. 3E,F).

Furthermore, in MPV17-related MDDS patients, lactic acidosis is often found in addition to the reduction in relative mtDNA copy number30. We therefore examined lactate levels in the CNS in both neuron-specific dMpv17 knockdown fly lines. The relative lactate levels in the CNS increased by 2.3- and 2.8-fold, respectively, in elav > dMpv17-IR42671 and elav > dMpv17-IR56889 flies compared to the control elav > GFP-IR (Fig. 3C). These data are very similar to symptoms of MPV17-related MDDS patients. We also investigated the effect of dMpv17 knockdown on glycolytic activity by measuring pyruvate levels. However, pyruvate levels were not significantly changed (Fig. 3D). Our findings for lactate and pyruvate levels in the CNS of dMpv17 knockdown larvae are consistent with findings in HEK293T cells expressing human MPV17 carrying the mutation R50W, which was identified in an MPV17-related MDDS patient25. In addition, the level of ROS production in R50W-expressing cells is greater than that in wild-type Mpv17-expressing cells25, a result that is broadly consistent with our results for dMpv17 knockdown larvae.

Reduction of dMpv17 induces abnormal morphology of the NMJ

The defective locomotor ability exhibited by dMpv17 knockdown larvae may be due to a functional abnormality of neurons. We therefore examined the morphology of neural synaptic terminals at the neuromuscular junctions (NMJs) in third instar larvae. The presynaptic terminals were visualized with the neuronal membrane marker anti-horseradish peroxidase (HRP) IgG and the postsynaptic domains with the postsynaptic marker anti-Discs large 1 (Dlg1) IgG (Fig. 4A). The total length of the synaptic branches, the number of boutons, and the average size of the boutons in the NMJs on the fourth muscle of neuron-specific dMpv17 knockdown larvae (elav > dMpv17-IR56889) were measured and compared with those of control larvae (elav > GFP-IR). While the total length of the synaptic branches and the number of boutons were significantly decreased in dMpv17 knockdown larvae compared with control (Fig. 4B), the average size of the boutons was not changed (data not shown). At the synaptic terminal there is an active zone where synaptic vesicles fuse to release neurotransmitters into the synaptic cleft and transmit signals to the next synapse or muscle. Therefore, we performed immunostaining with anti-Bruchpilot (Brp) IgG, a marker of the active zone, and measured the density and size of the active zone in the NMJs (Fig. 4C). The area of the active zone of the synapse was reduced in the NMJs of neuron-specific dMpv17 knockdown larvae (Fig. 4D). These data suggest that the reduction of dMpv17 leads to the locomotor impairment following the morphological defects in the larval NMJs.

Figure 4
figure 4

A dMpv17 knockdown influences larval NMJ morphology. (A) Synapses in the NMJ on the fourth muscle of third-instar larvae were inspected in elav > GFP-IR, elav > dMpv17-IR42671, and elav > dMpv17-IR56889 larvae. The presynaptic terminals were visualized with anti-HRP IgG and the post-synaptic domains with anti-Dlg1 IgG. Scale bar, 10 μm. (B) Synapse branch length, bouton numbers were quantified. **P < 0.01, *P < 0.05, statistical analysis was performed using Student’s t-test vs elav > GFP-IR. n = 15. (C) The active zone in the NMJ of the fourth muscle was visualized with anti-HRP and anti-Brp IgGs. Scale bar, 10 μm. (D) The area of Brp signals merged with HRP was quantified. **P < 0.01, statistical analysis was performed using Student’s t-test vs elav > GFP-IR. n = 13.

Full size image

Discussion

There are ten MPV17/PMP22 family genes in the Drosophila genome based on the predicted amino acid sequences. However, only one gene, Pasang Lhamu, has been reported to date31, and none of the others has yet been characterized. In this study, we demonstrated that CG11077, an MPV17/PMP22 family gene, is a Drosophila ortholog of mammalian Mpv17 and named it dMpv17. Human Mpv17 is known as a causal gene of MPV17-related hepatocerebral MDDS and CMT4,6,8, and models of these diseases have been developed in several organisms such as the mouse and zebrafish10,17. However, in spite of the fact that most patients show neurological symptoms, neurological defects have not been investigated in these model organisms.

Drosophila is a useful model to study the pathogenic mechanisms of various neurological diseases and disorders20,21,22,32 because the Drosophila nervous system is relatively simple, yet it shows good conservation in its basic structure and function with the mammalian nervous system33,34. In Drosophila, the larval CNS contains 15,000 cells, including 1000 glial cells, and the adult CNS contains 167,000 cells, including 16,000 glial cells35,36,37, while the human brain consists of 100 billion neurons and glial cells38. In addition, Drosophila larval and adult behavior is less complex than that of mammals but is still complex enough to analyze the effect of changes in gene function. We therefore specifically reduced the expression of dMpv17 in the Drosophila nervous system. Neuron-specific dMpv17 knockdown larvae showed neurological defects, including locomotor and cognitive impairments. Furthermore, in the larval CNS, the mtDNA copy number and ATP levels were significantly decreased and lactate and ROS levels were increased. These results show good correlation with the symptoms of human MPV17-related MDDS and the results obtained in cultured cells derived from patients. In Mpv17 KO mice, the relative mtDNA copy number is decreased in the liver and muscle but not in the kidney or brain2,10. In contrast, neuron-specific dMpv17 knockdown leads to a reduction in the copy number of mtDNA in the Drosophila larval CNS, suggesting that the fly model is more suitable than the mouse model to study the neurological symptoms of human MPV17-related MDDS and CMT.

dMpv17 knockdown induced both locomotor and mitochondrial defects in the larvae like the knockdown larvae of Solute carrier family 25 member 46a (Slc25A46a) or Mitochondrial trifunctional protein β subunit (Mtpβ/dHADHB), which are both functional in the mitochondria39,40. However, during the adult stage, while the Slc25A46a or Mtpβ/dHADHB knockdown flies showed locomotor defect, we did not detect any neurological defects, such as those responsible for seizures after severe mechanical stimulus (Fig. S2) or impairment of climbing ability (Fig. S1), even though dMpv17 is expressed in the adult CNS based on the RNA-seq data in the FlyBase, predicting that dMpv17 functions in the adult brain. In addition, surprisingly, the climbing ability of elav > GFP-IR was not deceased compared to dMpv17 knockdown flies throughout adult stage despite the longevity of elav > GFP-IR is significantly shorter (Figs. S1, S3). The observed difference between larvae and adults may be due to inefficient knockdown of dMpv17 in the adult brain. Because the larval motor neurons in the CNS and PNS are reorganized during metamorphosis41,42,43,44,45, if dMpv17 does not function in adult flies, it might not lead to behavioral impairment. In addition, the other genes which are shown in Fig. 1A might compensate for the dMpv17 function in adult flies. Alternatively, it might be related to the difference in metabolism between larvae and adult flies. Larvae tend to use more glycolysis because in their short time they need to grow quickly and thus need to use carbons for growth instead of fuel. In contrast, adults primarily use aerobic respiration for their ATP46,47,48. However, further analysis is necessary to clarify this point.

In conclusion, we first showed that CG11077 is an ortholog of mammalian Mpv17 and that neuron-specific CG11077 knockdown induced both locomotor and cognitive defects in Drosophila larvae, depending on the mitochondrial dysfunction involved. dMpv17 knockdown flies showed phenotypes that are very similar to the symptoms of MPV17-related hepatocerebral MDDS and CMT caused by Mpv17 mutation, and at least somewhat similar to the phenotypes of other Mpv17−/− model organisms, including mouse and zebrafish. Furthermore, it is also interesting whether the oenocyte or fat body, corresponding mammalian liver-specific dMpv17 knockdown flies mimic the symptoms of MPV17-related hepatocerebral MDDS and CMT caused by Mpv17 mutation. In addition to the similarities of the knockdown phenotypes to the human diseases, Drosophila provides various genetic tools to investigate genetically interacting factors and therapeutic candidates relatively easily, making the neuron-specific dMpv17 knockdown fly a potentially useful model to study the diseases.

Materials and methods

Fly stocks

Flies were raised on standard food containing 10% glucose, 5% corn flour, 4% dry yeast, 3% rice bran, and 0.65% agar at 25 °C49. White mutant (108479: w1118) and yellow and vermillion double mutant (101249: y1 v1) fly strains were sourced from the Kyoto Drosophila Stock Center. UAS-GFP-IR (9331: w1118; P{w+mC = UAS-GFP.dsRNA.R}143), elav-GAL4 (8760: w*; P{w+mC = GAL4-elav.L}CG16779[3]), UAS-dMpv17-IR56889 (56889: y1 sc* v1 sev21; P{y[+ t7.7] v[+ t1.8] = TRiP.HMC04110}attP40) targeting the region corresponding to nucleotide residues 598–618 and 688–708, were sourced from the Bloomington Drosophila Stock Center. UAS-dMpv17-IR42671/CyO (42,671: w1118; P{GD3155}v42671/CyO) targeting the region corresponding to nucleotide residues 252–577 was obtained from the Vienna Drosophila Resource Center. In order to minimize genetic background effects, the fly lines used in this study were backcrossed several times with the white or yellow vermillion mutant strain.

Comparison of amino acid sequences of MPV17/PMP22 family proteins

The amino acid sequences of MPV17/PMP family proteins were obtained from UniProt (http://www.uniprot.org). Multiple sequence alignments of protein sequences and the construction of a phylogenetic tree were carried out with the Clustal Omega program50.

Quantitative RT-PCR analysis

RNA was purified from the CNS of male third instar larvae using the RNeasy Lipid Tissue Mini Kit (74804, Qiagen). RNA (100 ng) was used as a template for cDNA synthesis using the PrimeScript RT reagent kit (RR037; Takara). Real-time PCR was performed using SYBR Premix Ex Taq II (RR820; Takara) with 10 ng of cDNA as a template. Specific primers were designed for dMpv17 (fw 5′-CTTCAATAAGCTGAGCATCG-3′/rev 5′-CGACAATTCTCTGGTTGC-3′) and glucose-6-phosphate dehydrogenase (G6PD) (fw 5′-AAGCGCCGCAACTCTTTG-3′/rev 5′-AGGGCGGTGTGATCTTCC-3′). PCR was performed using a CFX96 Touch Real-Time PCR Detection System (Bio-Rad), and data were analyzed using CFX Manager Software. The values were read from the calibration curve calculated by CFX Manager Software.

ROS detection

ROS levels in the larval CNS were measured by the general oxidative stress indicator CM-H2DCFDA (C6827, ThermoFisher Scientific). The dissected third instar larval CNSs were shaken in 10 µM CM-H2DCFDA in PBS in the dark for 10 min. Then the CNSs were washed with PBS for 5 min and mounted with Vectashield Mounting Medium (Vector Laboratories). Fluorescent patterns were visualized by a confocal laser scanning microscope (Olympus Fluoview FV10i). The fluorescence intensity in the CNS was analyzed by ImageJ software (NIH, USA).

ATP measurement

ATP levels in the larval CNS were measured by CellTiter-Glo (G7570; Promega) as previously described51 with some modifications. Ten instar larval CNSs were homogenized in 100 µL of ATP assay buffer (ab83355; Abcam), and the homogenate was centrifuged at 12,000×g at 4 °C for 10 min. The supernatant was then transferred to a new tube, 10 µL of cold trichloroacetic acid (ab204708; Abcam) was added to the supernatant, and the sample was kept on ice for 15 min. After centrifugation at 12,000×g for 5 min, the supernatant was transferred to a new tube, and 7.5 µL of neutralization solution (ab204708; Abcam) was added to the supernatant. The sample was incubated on ice for 5 min. In each well of a microtiter plate, 10 µL of sample, 40 µL of ATP assay buffer, and 50 μL of CellTiter-Glo were mixed. Luminescence was read on a Lumat LB 9507 luminometer (Berthold Technologies).

Lactate measurement

The Lactate-Glo Assay (J5021; Promega) was used to measure lactate levels52. After homogenizing ten larval CNSs in 100 µL of PBS, samples were centrifuged at 12,000×g at 4 °C for 10 min, and the supernatant was transferred to a new tube. In each well of a microtiter plate, 1 µL of sample, 49 µL of PBS, and 50 µL of lactate detection reagent were mixed and the luminescence was read on a Lumat LB 9507 luminometer.

Pyruvate measurement

A Pyruvate Assay Kit (MAK071; Sigma) was used to measure pyruvate levels53. After homogenizing ten larval CNSs in 100 µL of pyruvate assay buffer, samples were centrifuged at 12,000×g at 4 °C for 10 min. The supernatant was then transferred to a new microtiter plate. In each well, 50 µL of sample and 50 µL of pyruvate detection reagent were mixed and the fluorescence was read on Fluoroskan Ascent plate reader (ThermoFisher).

Climbing assay

Twenty male flies were transferred to a glass vial (9820TST-F25-150, Iwaki) without CO2 anesthesia and allowed to habituate for 5 min. The flies were knocked to the bottom by banging the vial twice, then the flies were allowed to climb the wall for 7 s. The height to which each fly climbed was scored as follows: 0 (< 2.0 cm), 1 (2.0–3.9 cm), 2 (4.0–5.9 cm), 3 (6.0–7.9 cm), 4 (8.0–9.9 cm), and 5 (> 10.0 cm). The test was carried out 5 times with a 1-min interval between tests. The climbing index was calculated after arranging the scores of each test in descending order and the averages of the five tests were calculated.

Bang-sensitivity assay

The bang-sensitivity assay was performed as previously described54 with modifications. Seven-day-old flies were transferred to a test vial without CO2 anesthesia and allowed to acclimatize for 5 min. The flies were then vortexed at the highest speed for 20 s and were immediately video recorded. The test was repeated after 10 min. The number of flies showing seizure-like responses, characterized by complete immobilization on the back and/or uncontrollable movements for 3 s or more, were counted.

Crawling assay

Larval locomotion was assessed by crawling assay23. Male third instar larvae climbing a wall were used. Assays were performed on 2% agarose gel plates and larval movements were recorded with a video camera for 1 min at a density of 3 larvae per plate. Larval movement distance, length and average speed were analyzed using Image J's wrMTrck plugin.

Visualization of the neuromuscular junction

To visualize the morphology of the neuromuscular junction (NMJ), male third instar larvae were dissected in HL3 saline solution and fixed in 4% paraformaldehyde in PBS at 25 °C for 30 min. After washing 3 times with PBS containing 0.3% Triton X-100 (PBST) for 10 min, the samples were blocked with PBS containing 0.15% Triton X-100 and 10% normal goat serum for 30 min and incubated with mouse anti-Discs large 1 IgG (4F3; 1:200; DSHB) or mouse anti-Bruchpilot IgG (nc82; 1:200; DSHB) at 4 °C for 16 h. Samples were then washed 3 times with PBST for 10 min and treated for 2 h with anti-mouse IgG labeled with Alexa 594 (A-11032; 1:200; Molecular Probes) and FITC-conjugated goat anti-HRP IgG (1:200; Jackson ImmunoResearch). After washing 3 times with PBST for 10 min, the samples were embedded in Vectashield Mounting Medium and inspected with a confocal laser scanning microscope (Olympus Fluoview FV10i). At the neuromuscular junction, branch length and the numbers of branches and boutons were measured using MetaMorph Microscopy Automation and analyzed using Image Analysis Software (Molecular Devices). Synapses and active zones were observed on the fourth muscle in segments A2 to A6.

Quantification of mitochondrial DNA copy number

Total DNA was purified from 20 third instar larval CNSs using Mitochondrial DNA Isolation Kit (K280-50; Biovision), omitting the step of separating the genomic and mitochondrial DNA. mtDNA copy number was assessed by qPCR using primers for 16S rRNA (for mtDNA) and RpL32 (for nuclear DNA)44. Real-time PCR was performed using SYBR Premix Ex Taq II (RR820; Takara) with 10 ng of total DNA as a template. Specific primers were designed for 16S rRNA (fw 5′-TCGTCCAACCATTCATTCCA-3′/rev 5′-TGGCCGCAGTATTTTGACTG-3′) and RpL32 (fw 5′-AGGCCCAAGATCGTGAAGAA-3′/rev 5′-TGTGCACCAGGAACTTCTTGAA-3′). PCR was performed using a CFX96 Touch Real-Time PCR Detection System, and data were analyzed using CFX Manager Software. The values were read from the calibration curve calculated by CFX Manager Software.

Odor-taste learning assay

PCR tubes containing 10 µL of n-amyl acetate (AM) diluted 50-fold with liquid paraffin were placed at either end of a 10-mm plate filled with 2% agarose gel containing 2 M sucrose (SUC). Third instar larvae in the food were quickly washed in PBS, placed in the center of the plate, and exposed to AM for 5 min to allow them to learn the relationship between AM and SUC. Next, after washing with PBS, the larvae were transferred onto gel without SUC in which PCR tubes containing 10 µL of 1-octanol (OCT) were placed (25506; Nacalai Tesque), and left for 5 min. The training was carried out 3 times and the larvae were transferred to a SUC-free test plate divided into two compartments by a 1-cm-wide neutral zone drawn in the center. After 3 min, the number of larvae that moved to the OCT or AM side was recorded. The period from conditioning to testing was defined as AM/OCT+. The "+" here represents reward. The learning index (LI) was calculated from the AM and OCT preferences as previously described55.

Lifespan analysis

Adult male flies were raised on the standard food under 12 h light–dark cycles at 28 °C. Flies were transferred to new food every 3 or 4 days and the number of dead flies was counted.

Data analysis

GraphPad Prism 9.4.1 was used to statistically analyze all results. In the qPCR analysis, ATP, lactate, and ROS measurements, and the crawling, climbing, and learning assays, p-values were also calculated by one-way ANOVA with Dunnett’s multiple comparison test. For immunostaining of NMJs (synapse and active zone), p-values were calculated using the unpaired two-tailed Student’s t-test. All data are shown as means ± SEM.