Polo Kinase Phosphorylates Miro to Control ER-Mitochondria Contact Sites and Mitochondrial Ca²⁺ Homeostasis in Neural Stem Cell Development

SUMMARY

Mitochondria play central roles in buffering intracellular Ca²⁺ transients. While basal mitochondrial Ca²⁺ (Ca²⁺_mito) is needed to maintain organellar physiology, Ca²⁺_mito overload can lead to cell death. How Ca²⁺_mito homeostasis is regulated is not well understood. Here we show that Miro, a known component of the mitochondrial transport machinery, regulates Drosophila neural stem cell (NSC) development through Ca²⁺_mito homeostasis control independent of its role in mitochondrial transport. Miro interacts with Ca²⁺ transporters at the ER-mitochondria contact site (ERMCS). Its inactivation causes Ca²⁺_mito depletion and metabolic impairment, whereas its overexpression results in Ca²⁺_mito overload, mitochondrial morphology change, and apoptotic response. Both conditions impaired NSC lineage progression. Ca²⁺_mito homeostasis is influenced by Polo-mediated phosphorylation of a conserved residue in Miro, which positively regulates Miro localization to, and the integrity of, ERMCS. Our results elucidate a regulatory mechanism underlying Ca²⁺_mito homeostasis and how its dysregulation may impact NSC metabolism/development and contribute to disease.

Graphical Abstract

INTRODUCTION

Mitochondria are important for many aspects of cellular function, from ATP production and intermediary metabolism to apoptosis. Mitochondrial dysfunction has been broadly linked to the pathogenesis of a multitude of neurological disorders (Chan, 2006; Schon and Przedborski, 2011; Tait and Green, 2012; Wallace, 2005), emphasizing the particular importance of this organelle to neuronal function and maintenance. Mitochondrial function and Ca²⁺ signaling have been intimately linked. On one hand, Ca²⁺_mito uptake helps buffer cytosolic Ca²⁺ signals and oscillations such as those generated by neuronal activation, thus amplifying and sustaining signals arising from neuronal activity while protecting neurons from the potentially damaging Ca²⁺ spikes (MacAskill et al., 2010; Sheng and Cai, 2012; Vos et al., 2010). On the other hand, Ca²⁺_mito can positively regulate the activities of TCA cycle enzymes and electron transport chain components, stimulating ATP production (Cardenas et al., 2010; Jouaville et al., 1999; McCormack et al., 1990). Ca²⁺_mito uptake primarily occurs at ERMCS where local Ca²⁺ can reach high levels (Rizzuto et al., 1993; Rizzuto et al., 1998). Proteins localized to ERMCS include those that mediate Ca²⁺ transfer, such as inositol 1,4,5-trisphosphate receptor (IP3R) at the ER, voltage-dependent anion-selective channel (VDAC) at the mitochondria outer membrane (MOM), mitochondrial Ca²⁺ uniporter (MCU) at the mitochondrial inner membrane (MIM) (Giorgi et al., 2015; Hajnoczky et al., 2002; Rowland and Voeltz, 2012), and mitofusin-2 (de Brito and Scorrano, 2008). In addition to transferring Ca²⁺, ERMCS also transfer lipids (Hayashi et al., 2009). Despite the importance of ERMCS to fundamental cellular functions and its potential link to disease conditions (Schon and Area-Gomez, 2013; Rizzuto et al., 2012), its roles during normal development remain mostly unexplored, and the signaling mechanisms regulating its integrity and function are largely unknown.

Mitochondrial rho GTPase (Miro) is a MOM protein containing two GTPase domains and two Ca²⁺-sensing EF hand domains (Fransson et al., 2003). Genetic studies in Drosophila (Guo et al., 2005) and mouse (Nguyen et al., 2014) revealed a key function of Miro in regulating the axonal transport of mitochondria, during which Miro forms a multi-protein complex with Milton and Kinesin heavy chain (KHC) to link mitochondria with motor molecules and the microtubule cytoskeleton (Glater et al., 2006). Through binding to the EF hands of Miro and altering protein-protein interactions within the transport machinery, cytosolic Ca²⁺ dynamically regulates the motility of mitochondria, offering one mechanism to match mitochondrial distribution with neuronal activity (Macaskill et al., 2009; Saotome et al., 2008; Wang and Schwarz, 2009). The PINK1/Parkin pathway, which is critically involved in maintaining mitochondrial function and mutations in which are associated with Parkinson's disease (PD), can regulate Miro stability and thus mitochondrial transport (Liu et al., 2012; Wang et al., 2011). Interestingly, despite the importance of the Miro/Milton/Khc complex in regulating mitochondrial transport in metazoans, Miro is the only component conserved in all eukaryotes, suggesting that it may perform some ancestral function that is independent of the transport complex. Indeed, functions of Miro not directly related to transport are emerging in yeast (Lee and Lu, 2014), where yeast Miro (Gem1) is localized to ERMCS (Kornmann et al., 2011) and regulates ER-associated mitochondrial division (Murley et al., 2013). The prominent roles of Miro in mitochondrial regulation and the importance of mitochondria to neuronal function and survival in humans promoted us to examine the physiological role of Miro in the nervous system of a multicellular organism.

RESULTS

Miro Regulates Drosophila NSC Maintenance in a Manner Independent of Mitochondrial Transport

To investigate the physiological function of Miro in a developmental setting, we used Drosophila central brain NSCs called NBs. We examined the type I and type II NBs in the larval brains of animals carrying the dMiro^B682 mutant allele, presumably a protein null (Figure S1A, S1B) with disrupted mitochondrial distribution in neuronal and muscle tissues and late-larvae or early-pupae lethality (Guo et al., 2005). The brain size and number of Deadpan (Dpn)-positive NBs were dramatically reduced in dMiro^B682 homozygous mutants (Figure 1A, 1B). Loss of NB was also observed when dMiro was specifically knocked down in the NB lineages by RNAi (Figure 1C, 1D). To examine the effect of dMiro mutation on NB lineage architecture, we performed clonal analysis with MARCM (Lee and Luo, 2001). Type I and type II NB lineages of control animals each invariably contained one Dpn-positive primary NB, whereas the primary NBs were frequently lost in dMiro^B682 mutant NB clones (Figure 1E). In cases where the primary NBs remained in the mutant clones, their sizes were noticeably smaller than control NBs (Figure 1E, 1F). These results indicate a critical and cell-autonomous role of Miro in the development and/or maintenance of larval brain NBs.

Figure 1. Regulation of NB Behavior by Miro.

(A, B) Reduction of NB number and brain size in dMiro^B682 null mutant. Larval brains of dMiro^B682 heterozygous and homozygous animals were immunostained for pan-NB marker Dpn, type I NB marker Ase, and actin. The central brain area is outlined with white dashed line.

(C, D) Reduction of NB number by NB-specific knockdown of dMiro. Control (1407>Dcr2) and dMiro RNAi (1407>Dcr2,dMiro-RI) larval brains were immunostained for Dpn.

(E) MARCM analysis of NB lineages in the dMiro^B682 and milton⁹² backgrounds. Brains were stained for Dpn and Pros, a differentiation marker. NB clones are marked with GFP. Upper: Type II NBs, Lower: Type I NBs. Primary NBs are indicated with arrows. Note that the primary NB in the dMiro^B682 type II NB clone is losing Dpn expression.

(F) Quantification of the size of NBs from E.

(G, H) Effects on NB number by pan-NB (1407-Gal4) or type II NB-specific (Pnt-Gal4) dMiro OE. Brains were stained for Dpn. The central brain area is outlined with yellow dashed line.

(I, J) Quantification of total NB (I) or type II NB (J) number from G and H, respectively.

(K-M) Effects of dMiro OE on type II NB size (L) and IP number (M). The type II NB lineages are outlined with white dashed line. Primary NBs are marked with arrows.

Error bar: SEM; *, p<0.05 versus control in Student's t-tests. n = 5. Scale bars, 100 μm (A, C, G); 20 μm (K).

See also Figure S1.

Type II NBs undergo self-renewing asymmetric divisions to generate immature intermediate progenitors (IPs), which progress into transit-amplifying mature IPs that also undergo self-renewing asymmetric divisions to generate ganglion mother cells (GMCs) and differentiated neuronal or glial progenies (Bello et al., 2006; Boone and Doe, 2008; Bowman et al., 2008). IPs sequentially express the transcription factors Dichaete (D), Grainy head (Grh), and Eyeless (Ey) during their transition from young to old IPs (Bayraktar and Doe, 2013). Thus, early-born (old) IPs express Ey, whereas late-born (young) IPs express D. In dMiro^B682 mutant clones, young IPs and old IPs were both significantly reduced in number compared to control WT clones (Figure S1C, S1D). Loss-of-function (LOF) of dMiro thus reduces the number of transit-amplifying IPs, presumably as a result of stem cell loss during lineage progression.

Next, we examined Miro gain-of-function (GOF) effect using 1407-Gal4- or Pnt-Gal4-directed pan-NB or type II NB-specific Miro overexpression (OE), respectively. Miro OE led to a reduction of the number of total NBs in 1407-Gal4>dMiro-OE (Figure 1G, 1I) and occasional loss of type II NBs in Pnt-Gal4>dMiro-OE conditions (Figure 1H, 1J). A reduction of NB size was also seen in dMiro-OE animals (Figure 1K, 1L). In the Pnt-Gal4>dMiro-OE condition, the number of IPs, specifically the D⁺ young IPs in type II NB lineages, was reduced (Figure 1M). These results indicate that NB maintenance and lineage progression are sensitive to Miro dosage.

Miro is a key component of a conserved protein complex that transports mitochondria along microtubule tracks (Brickley et al., 2005; Fransson et al., 2006; Guo et al., 2005; Hirokawa, 1998; Pilling et al., 2006; Stowers et al., 2002). To test whether the function of Miro in mitochondrial trafficking is important for NB maintenance, we examined mitochondrial distribution in dMiro mutant clones by staining for ATPsyn5α. Mitochondrial distribution or morphology was not obviously affected by dMiro mutation in NB lineages (Figure S1E). To further assess the role of mitochondrial transport in NB development, we analyzed the LOF and GOF of Milton, another key component of the transport complex (Glater et al., 2006). Neither the number nor size of primary NBs was affected by NB-specific knockdown of Milton (Figure S1F), in milton⁹² mutant MARCM clones (Figure 1E, 1F), or when Milton was overexpressed (Figure 1I, 1J). The Milton RNAi line used here was known to be effective in altering neuronal mitochondrial transport or function (Liu et al., 2012; Fang et al., 2012; Iijima-Ando et al., 2009). Intriguingly, Miro-OE led to mitochondrial distribution defects in NBs, with mitochondria accumulating asymmetrically to one side of the perinuclear area, and this defect was partially rescued by Milton-RNAi (Figure S1G). However, Miro-OE effect on NB lineage development was not affected by Milton-RNAi (data not shown), suggesting that a transport-independent function of Miro is important for NB maintenance and lineage development.

Miro Regulates Ca²⁺_mito Homeostasis

Miro contains GTPase domains and Ca²⁺-binding EF hand motifs, and its function in transporting neuronal mitochondria is regulated by activity-induced Ca²⁺ influx (Macaskill et al., 2009; Saotome et al., 2008; Wang and Schwarz, 2009). Yeast Gem1 and mammalian Miro1 (Rhot1) are localized to ERMCS (Kornmann et al., 2011), a structure that directs Ca²⁺ transfer from ER to mitochondria (Hayashi et al., 2009). To test whether Miro regulates Ca²⁺_mito homeostasis in NSC lineage, we first measured Ca²⁺_mito levels under dMiro LOF and GOF, using the Ca²⁺-sensitive fluorescent dye Rhod-2 AM. Consistent with a previous study (Guo et al., 2005), Rhod-2 AM preferentially monitored Ca²⁺_mito in Drosophila neurons (Figure S2A). We loaded primary cultured neurons with Rhod-2 AM. Blocking ER Ca²⁺ uptake with thapsigagin (TG) increased cytosolic Ca²⁺ and induced a spike of Ca²⁺_mito as reported (Perocchi et al., 2010), which quickly dissipated in control cells. Notably, in Miro-OE neurons the TG-induced Rhod-2 AM spike was stronger and persisted longer (Figure 2A), whereas TG induced a much weaker spike in dMiro^B682 mutant or Miro-RNAi neurons than controls (Figure 2B, 2C). We also measured cytosolic Ca²⁺ response using Fluo-3AM. Overexpression of wild type but not a mutant form of Miro led to a decrease in the magnitude of cytosolic Ca²⁺ spikes induced by TG (Figure S2B).

Figure 2. Regulation of Ca²⁺_mito by Miro.

(A-C) Effects of pan-neuronal elav-Gal4 driven dMiro OE (A) or dMiro RNAi (B), or dMiro^B682 null mutation (C), on TG-stimulated Rhod-2 AM fluorescence in primary cultured fly neurons. The traces show mean response of cells present in the microscope field and are representative of more than 3 experiments.

(D-F) Effects of OE or RNAi of Miro (D), Marf (E), or Porin (F) on mitochondrial Ca²⁺ in type II NB lineages as monitored using mito-AEQ. The traces show mean response of 10 dissected larval brains present in the microtiter plate and are representative of more than 3 tests.

(G, H) Imaging of basal Ca²⁺_mito using mito-GCaMP in larval brains without (control) or with Miro or Porin OE. White circle outline larval brain, yellow lines separate central brain (left) from the rest of brain. Images are representative of more than 5 samples. H, quantification of mito-GCaMP intensity in NBs from G.

Error bar: SEM; *, p<0.05 versus control in Student's t-tests. Scale bars, 100 μm (G).

See also Figure S2.

To detect Ca²⁺_mito levels in vivo in NBs, we expressed genetically-encoded, mitochondria-targeted Ca²⁺_mito reporters mito-Apoeaquorin (mtAEQ) (Terhzaz et al., 2006) or mito-GCaMP3 (Lutas et al., 2012) in type II NBs. With the mtAEQ reporter, dissected brains from transgenic (Tg) animals were treated with histamine, an IP3-generating agonist, to induce Ca²⁺ release from the ER and increase Ca²⁺_mito uptake (Perocchi et al., 2010), and luminescence signals were immediately measured with an automated luminometer. Under this condition, Ca²⁺_mito was significantly elevated when known components of ERMCS such as Marf and Porin/VDAC were overexpressed, and Ca²⁺_mito was attenuated when they were inhibited (Figure 2E, 2F), validating the utility of the reporters in monitoring the effect of ERMCS on Ca²⁺_mito homeostasis. We found that Ca²⁺_mito was significantly increased by Miro-OE, but decreased by Miro-RNAi (Figure 2D). In contrast, Milton LOF or GOF had no obvious effect on Ca²⁺_mito uptake under similar conditions (Figure S2C). With the mito-GCaMP3 reporter, we observed that basal fluorescence signal was significantly increased by Miro-OE or Porin-OE (Figure 2G, 2H). The expression level of mito-GCaMP protein was not altered by Miro-OE (Figure S2D). These results, and the observation that cytosolic Ca²⁺ level as measured with a cytosolic GCaMP reporter (Figure S2E-F), and mitochondrial membrane potential (Figure S2G, S2H) - the driving force of Ca²⁺_mito entry, were not altered by Miro-OE, support that Miro specifically promotes Ca²⁺_mito uptake. Miro therefore regulates Ca²⁺ transfer from ER to mitochondria, and this occurs in a manner seemingly independent of mitochondrial transport.

Miro-Mediated Ca²⁺_mito Homeostasis Regulates Mitochondrial Activity

Ca²⁺_mito uptake critically regulates many cellular functions, ranging from ATP production to cell death (Rizzuto et al., 2012). Sufficient Ca²⁺ uptake into the mitochondrial matrix is needed for ATP production by enhancing the activities of key metabolic enzymes. For example, the key TCA cycle enzyme pyruvate dehydrogenase (PDH) is inhibited by phosphorylation by PDH kinase (PDK). Ca²⁺_mito activates PDH phosphatase to remove an inhibitory phosphate from PDHE1 subunit (Cardenas et al., 2010). Thus, Ca²⁺_mito negatively influences p-PDHE1 level. Normalized p-S293-PDHE1 level was markedly increased in dMiro mutant but decreased in Miro-OE brain extracts (Figure 3A, 3B, S3A, S3B). Consistently, the intensity of p-PDHE1 immunosignal was increased in dMiro mutant but decreased in Miro-OE NB clones, compared to control NBs located outside of the clones in the same animals (Figure 3C). ATP production in larval brain was decreased in dMiro^B682 mutant, which was rescued by NB-specific Miro-OE (Figure 3G). Measurement of oxygen consumption rate (OCR) confirmed alteration of mitochondrial metabolism under Miro LOF and GOF conditions (Figure 3H). In response to ATP depletion, AMP-activated protein kinase (AMPK) is activated by phosphorylation as a compensatory response to restore energy homeostasis (Hardie, 2007). AMPK was hyperphosphorylated in dMiro mutant (Figure 3A, 3B). These results suggest that reduced Ca²⁺_mito uptake impairs mitochondrial metabolism and ATP production in Miro LOF condition.

Figure 3. Miro-Mediated Ca²⁺_mito Homeostasis Regulates Mitochondrial Activity.

(A, B) Western blot (WB) analysis showing increased p-PDHE1 and p-AMPK in dMiro mutant. PDHE1, Porin, and actin serve as controls. Normalized protein levels are shown as the ratio of dMiro/Porin, pS293-PDHE1/PDHE1 and pT713-AMPK/AMPK from A.

(C-F) Immunostaining of p-PDHE1 (C), mito-SOX (D, E), or activated caspase 3 (F) in dMiro mutant MARCM clones (yellow-outlined in upper panels) or dMiro-OE Flip-out clones (yellow-outlined in lower panels). NBs within clones are marked with white circles and control NBs outside of the clones are marked with red circles in C.

(G) Effect of dMiro-WT OE on ATP level in dMiro mutant larval brain.

(H) Real-time changes in OCR and Glycolytic rate (ECAR) were measured in control, dMiro-WT-OE (Mhc>dMiro-WT) and dMiro mutant animals using the Seahorse Bioscience XF Analyzer.

Error bar: SEM; *, p<0.05 versus control in Student's t-tests. Scale bars, 50 μm (C), 20 μm (D, E).

See also Figure S3.

While mitochondrial metabolism requires sufficient levels of Ca²⁺_mito, prolonged Ca²⁺_mito uptake can lead to opening of the permeability transition pore and release of pro-apoptotic components such as Cytochrome (Cyt) c, a critical event in mitochondria-mediated cell death (Rizzuto et al., 2012). To test whether the detrimental effect of Miro-OE on NB maintenance is related to Ca²⁺_mito overload-induced cytotoxicity, we analyzed mitochondrial redox status using MitoSOX Red, a mitochondrial superoxide indicator. MitoSOX signal was higher in Miro-OE than control clones. No obvious change in MitoSOX signal was observed in dMiro mutant clones (Figure 3D, 3E). Miro-OE but not dMiro mutant clones also showed increased staining of activated caspase-3 (Figure 3F), and release of Cyt c into the cytosol (Figure S3C, S3D), which are markers of apoptotic processes.

Under normal physiological conditions, the bulk of cellular Ca²⁺ resides within the ER lumen. ER Ca²⁺ depletion can cause ER stress and activate an unfolded protein response (Malhotra and Kaufman, 2011). We tested whether, by enhancing Ca²⁺_mito uptake, Miro-OE might lower ER luminal Ca²⁺ level and cause ER stress. The ER stress marker p-eIF2α was significantly elevated in Miro-OE but not dMiro mutant NB clones (Figure S3E), suggesting that elevated Ca²⁺_mito uptake may cause Ca²⁺_mito overload, Ca²⁺ depletion from the ER, and ER stress, all of which may contribute to apoptotic responses in Miro-OE condition. To test this hypothesis, we first examined the effect of ER stress inducers on NB development. Acute treatment of 3^rd instar larvae with 5mM DTT + 1μM TG resulted in loss of IPs (Figure S3G), indicating that ER stress can recapitulate Miro-OE effect in NB development. Importantly, two independent RNAi lines against the ER stress kinase PERK both effectively rescued dMiro-OE induced NB lineage defects (Figure S3F). Next, we estimated ER Ca²⁺ content by measuring ionomycin-releasable Ca²⁺ levels (Tu et al., 2006). Ionomycin is an ionophore that causes more complete emptying of ER Ca²⁺ stores than TG. In the absence of extracellular Ca²⁺, ionomycin elevates intracellular (cytosolic) Ca²⁺ by emptying intracellular stores, and the resulting transient Ca²⁺ peak gives a reasonable approximation of steady state ER Ca²⁺ content. With this method, we found that HeLa cells transfected with hMiro1 have less ER Ca²⁺ content than control cell transfected with empty vector (Figure S3H, I). Together, these results indicate that Miro-regulated Ca²⁺ homeostasis within the ER and mitochondria are both important for NB lineage development.

Miro Interacts with Ca²⁺ Transporters at the ERMCS to Regulate Ca²⁺_mito Homeostasis and NB Maintenance

Transfer of Ca²⁺ from the ER to mitochondria is a major function of the ERMCS carried out by the Ca²⁺ transporters IP3R, VDAC, and MCU localized at the ER, MOM, and MIM respectively. We found that similar to Miro case, knocking down fly IP3R (Venkatesh and Hasan, 1997), Porin (VDAC) (Messina et al., 1996), or MCU resulted in reduced NB number (Figures S4A-S4D), supporting the importance of ER-mitochondria Ca²⁺ transfer to NB maintenance. In addition, overexpressing these Ca²⁺ transporters also modestly reduced NB number (Figure S4E, S4F). We further tested the genetic interaction between Miro and the Ca²⁺ transporters. Although the heterozygous conditions for dMiro or each of the Ca²⁺ transporter genes alone had no obvious effect on NB development, double heterozygous combinations between dMiro and the Ca²⁺ transporters resulted in reduced NB number (Figure 4A, 4B), supporting that Miro functionally interacts with Ca²⁺ transporters at the ERMCS to control Ca²⁺_mito homeostasis, mitochondrial function, and NB maintenance. Consistently, partial LOF of IP3R or Porin reduced Ca²⁺_mito to basal levels in Miro-OE condition and blocked the Miro-OE effect on NB maintenance (Figure S4G, S4H).

Figure 4. Functional Interaction between Miro and Ca²⁺ Transporter at the ERMCS.

(A, B) Transheterozygotes between dMiro and IP3R (Itp-r83A), Porin, or dMCU mutants resulted in reduction of NB number. Larval brains were immunostained for Dpn, Pros, and F-actin (cell cortex). Central brain area is outlined with dashed line.

(C, D) Effects of CaCl₂ feeding in rescuing brain ATP level (C) and brain size (D) of dMiro mutant. Larval brains were immunostained for Dpn and F-actin. Brain lobes are outlined.

(E, F) Quantification of the effects of 2-APB and BAPTA feeding in rescuing the reduced IP number (E) and NB size (F) caused by dMiro OE.

(G, H) Genetic interactions showing rescue of NB number in dMiro mutant by the OE of Porin, dMCU, yNDI1, or PDK-RNAi, but not the H99 heterozygous background. Larval brains were immunostained for Dpn, Pros, and F-actin. H, quantification of number of NBs shown in G.

Error bar: SEM; *, p<0.05 in Student's t-tests. Scale bars, 100 μm.

See also Figure S4.

We next took a pharmacological approach to test the role of Ca²⁺_mito homeostasis in mediating Miro function in NSCs. Previous studies showed that feeding flies with CaCl₂ could increase intracellular Ca²⁺ levels, especially at the larval stage (Dube et al., 2000), and that exogenous Ca²⁺ could overcome genetic deficits in intracellular Ca²⁺ signaling (Howard, 1984). We found that Ca²⁺_mito was increased upon CaCl₂ feeding (Figure S4I, S4J), and that feeding dMiro^B682 mutant larvae with CaCl₂ partially restored ATP production (Figure 4C) and rescued the brain size reduction and NB loss (Figure 4D). Conversely, Ca²⁺_mito overload caused by Miro-OE was attenuated (Figure S4I, S4J), and the reduction of NB size and loss of IPs in Pnt>Miro-OE condition was rescued by reducing Ca²⁺_mito level via Ca²⁺ chleation with bis (2-aminophenoxy) ethane tetra-acetic acid (BAPTA), or IP3R inhibition with 2-aminoethoxydiphenyl borate (2-APB) (Figure 4E, 4F).

We also tested the interaction between Miro and the ER-mitochondria Ca²⁺ transporters using available GOF alleles. NB-specific OE of Porin or dMCU significantly rescued the NB-loss phenotype seen in dMiro^B682 mutant (Figure 4G, 4H). Consistent with mitochondrial metabolic impairment due to reduced Ca²⁺_mito uptake being responsible for the NB phenotype in dMiro^B682 mutant, overexpression of yeast NDI1 (yNDI1), which encodes a single-peptide alternative complex-I and is functional in boosting fly mitochondrial complex-I activity (Bahadorani et al., 2010; Sanz et al., 2010), partially restored NB number in dMiro^B682 mutant. Enhancing PDH activity by PDK RNAi had similar effect (Figure 4G, 4H). This contrasts with the lack of effect of overexpressing the apoptosis inhibitor p35 (data not shown), or the compound removal of three cell death executers in flies (Reaper, Grim, and Hid) using the H99 deficiency (Foley and Cooley, 1998) (Figures 4G, 4H). These results support the notion that the Miro-LOF effect on NB maintenance is due to Ca²⁺_mito depletion and metabolic impairment as apposed to Ca²⁺_mito overload and apoptotic activation in Miro-GOF condition.

Regulation of Miro Function by Polo-Mediated Phosphorylation at Serine 66

Despite the importance of ER-mitochondria Ca²⁺ signaling to cellular physiology, little is known how this process is regulated. Implication of Miro as a key player in this process offers the opportunity to investigate the regulatory mechanisms. By aligning the sequence of Miro homologs from fly to humans, we identified a conserved, putative phosphorylation site for Polo-like kinases (PLKs) in the N-terminal GTPase domain of Miro (Figure 5A, S5A). Of the two GTPase domains of Miro, the N-terminal one is particularly important for ERMCS localization (Kornmann et al., 2011), mitochondrial morphology and transport, and cell survival (Babic et al., 2015). Polo is also of particular interest as previous studies identified it as a regulator of NB self-renewal in Drosophila (Wang et al., 2007). Our immunohistochemical and biochemical analyses indicated that a fraction of Polo protein is localized to mitochondria (Figure S5B-E). To test for a molecular link between Miro and Polo in NB maintenance, we first asked whether Miro and Polo physically interact in vivo. In co-IP experiments using elav>Polo-GFP fly head extracts, we detected Polo-GFP in complex with endogenous dMiro (Figure 5B). Co-expression with Polo resulted in increased dMiro phosphorylation at Ser but not Thr residue(s) (Figure 5C, 5D). To test whether Polo directly phosphorylates Miro, we performed in vitro kinase assays using PLK1 as kinase and GST-dMiro fusion proteins as substrates. PLK1 strongly phosphorylated Miro-FL (aa 1-634) and Miro-N (aa 1–213), but not Miro-M (aa 210-393) or Miro-C (aa 390-634) (Figure 5E). None of the GST-Miro fusion proteins was phosphorylated by a control kinase PKC-zeta (Figure S5F, G). We next searched for the target site of Polo in Miro-N by focusing on conserved Ser residues, including Ser66 that matches the consensus site for PLKs (Figure 5A, S5A). We note that the corresponding sites in mammalian Miro1 and Miro2 may be better targets for PLK2 or 3, which prefers an acidic residue at +1 position (Rozeboom and Pak, 2012). A S66A mutation led to a dramatic reduction of Miro-N phosphorylation by PLK1 compared to mutating other conserved Ser residues (Figure 5F). S66 in dMiro N-terminal GTPase domain thus represents a major phosphorylation site for PLK1 in vitro.

Figure 5. Regulation of Miro Function by Polo-Mediated Phosphorylation at Serine 66.

(A) Comparison of candidate Polo/PLK phosphorylation motifs at the N-terminus of Miro. The consensus D/E-X-S-X-X-D/E motif is underlined, and S66 indicated by asterisk.

(B) Polo-GFP is present in dMiro IP prepared from fly brain extracts.

(C, D) Increased dMiro phosphorylation at Ser but not Thr residue(s) after Polo co-expression in fly brain. D, quantification of normalized phospho-Ser intensity from C.

(E, F) In vitro kinase assays showing phosphorylation of dMiro-N by PLK1 (E), and the strong effect of S66A mutation in blocking PLK1 effect (F). Bar graph shows quantification of normalized phospho-dMiro signal (top). GST-dMiro proteins were visualized by autoradiography (middle panels) or Coomassie blue staining (bottom panels).

(G, H) Rescue of brain size (G) and NB number (H) in dMiro mutant by dMiro-WT but not -S66A. Larval brains were immunostained for Dpn, Pros, and F-actin.

(I) Rescue of reduced brain ATP level in dMiro mutant by dMiro-S66E, but not -S66A.

(J-L) Rescue of type II NB number (J, K) or IP number (J, L) in Polo-CA OE animals by dMiro-S66A or dMiro-RNAi. Larval brains were immunostained for Dpn and GFP. Dashed lines mark type II NB lineages.

(M, N) Rescue of type II NB number in Polo OE animals by RNAi of ERMCS components. Larval brains were immunostained for Dpn, GFP and NICD (cell membrane). N, quantification of number of NBs in M.

Error bar: SEM; *, p<0.05 in Student's t-tests. Scale bars, 100 μm (G, upper panel of J, M ), 20 μm (lower panel of J).

See also Figure S5.

To study the physiological effects of Polo-mediated phosphorylation of Miro in vivo, we generated Tgs expressing wild type dMiro (dMiro-WT) or dMiro-S66A. To circumvent overexpression artifacts, Tgs that express dMiro-WT and dMiro-S66A at comparable, but less than endogenous dMiro expression levels (Figure 6F, S3A), were chosen for further analyses. The NB-loss phenotype seen in dMiro mutant was significantly rescued by 1407-Gal4-driven expression of dMiro-WT but not dMiro-S66A (Figure 5G, 5H). This effect is likely related to the function of Miro in modulating Ca²⁺_mito homeostasis, as dMiro-S66A exhibited compromised ability to induce Ca²⁺_mito uptake compared to dMiro-WT (Figure S5H, S2B). Similarly, reduced ATP production in dMiro mutant larval brain was rescued by 1407-Gal4 driven dMiro-WT, but not -S66A (Figure 5I). Unlike in dMiro-WT case, NB clones overexpressing dMiro-S66A did not show increased mitochondrial ROS (Figure S5J). Phosphorylation of S66 in dMiro thus positively influences its function in regulating mitochondrial function and NB maintenance under basal conditions.

Figure 6. Phosphorylation of Miro Regulates the Integrity of ERMCS.

(A, B) Immunostaining showing localization of p-S66-Miro to ERMCS of larval brain NBs. Circles mark the NBs of interest. Larval brains were stained for the indicated antibodies.

(C) Effects of Polo on p-S66-Miro immunosignal in Flip-out NB clones with Polo overexpressed or knocked down by RNAi. Clones are marked with GFP.

(D, E) Quantification of dMiro and p-dMiro colocalization with ER and mitochondria shown in A and B (D) or p-dMiro colocalization with ER shown in C (E), using Mander's coefficient.

(F, H) Co-IP assays showing the effects of S66A and S66E mutations on Miro interaction with Porin and Marf. H, quantification of proteins pulled down by co-IP in F.

(G, I) Co-IP assays showing the effects of dMiro mutation or OE of Miro variants on IP3R-VDAC interaction. I, quantification of proteins pulled down by co-IP in G.

Error bar: SEM; *, p<0.05 in Student's t-tests. Scale bars, 20 μm.

See also Figure S6.

Consistent with Polo playing a role in regulating Ca²⁺_mito homeostasis, Ca²⁺_mito level was increased by Polo-OE but moderately decreased by Polo-RNAi (Figure S5I). Moreover, Polo-OE caused significant loss of type II NBs and IPs, which was rescued by dMiro RNAi or dMiro-S66A co-expression (Figure 5J-L), or by RNAi of genes mediating ER-mitochondria Ca²⁺ transfer (Figure 5M, 5N). Importantly, the NB and IP loss caused by Polo-OE was also partially rescued by pharmacological interventions that reduce Ca²⁺_mito (Figure S5K, L). This effect was specific, as NB and IP losses caused by Numb-OE (Ouyang et al., 2011) was not rescued by similar treatments (Figure S5M). The Polo-Miro axis thus plays a physiological role in regulating NB maintenance via Ca²⁺_mito homeostasis.

Phosphorylation of Miro by Polo Regulates the Integrity of ERMCS

We next sought to understand the mechanism by which Polo phosphorylation of Miro affects Ca²⁺_mito homeostasis. We first tested whether phosphorylation affects the localization of Miro. For this purpose, we generated an antibody against dMiro phosphorylated at S66 (p-S66-Miro), the specificity of which was demonstrated by the absence of signal in dMiro mutant (Figure S6A-C), its sensitivity to phosphatase treatment (Figure S6D), and its response to altered Polo activity (see later). To specifically examine the localization of Miro or p-S66-Miro with respect to mitochondria and ER, we used red fluorescent protein (RFP) tagged with the ER-retention sequence KDEL, or ER resident protein PDI tagged with GFP as ER markers, and ATPsyn5α as mitochondrial marker. Compared to the rather uniform mitochondrial localization of Miro protein in general, p-S66-Miro exhibited a granular localization pattern that colocalized with ATPsyn5α and overlapped extensively with ER markers in the perinuclear area (Figure 6A, B, D). The abundance of p-S66-Miro in the perinuclear area was reduced by Polo RNAi but increased by Polo-WT or constitutively active Polo (Polo-CA) OE (Figure 6C, E). Phosphorylation of Miro by Polo at S66 thus promotes its localization to the ERMCS.

To test whether Miro interacts with proteins localized to ERMCS and whether this interaction is affected by its phosphorylation status, we performed co-IP assays using fly brain tissues expressing dMiro-WT, dMiro-S66A, or phospho-mimetic dMiro-S66E, which functioned similarly as dMiro-WT in NBs when overexpressed (Figure 6F, 5I). Tg expression was driven by the pan-neuronal elav-Gal4 driver. Miro-WT exhibited physical association with VDAC/Porin and Mitofusin/Marf, two key components of the ERMCS (de Brito and Scorrano, 2008; Rowland and Voeltz, 2012). Importantly, the Miro-Porin and Miro-Marf interactions were significantly enhanced by S66E, whereas the Miro-Marf interaction was attenuated by S66A mutation (Figure 6F, H), suggesting that S66 phosphorylation positively influences the interaction of Miro with proteins at the ERMCS.

We next examined the effect of Miro phosphorylation on ERMCS integrity. To this end, we analyzed the IP3R-VDAC interaction in brain tissues of dMiro mutant or Tg flies expressing Miro variants. Although the expression level of IP3R or VDAC was unaltered in all genotypes tested, their physical association was reduced in dMiro mutant or Miro-S66A Tg tissues, but increased in Miro-WT or Miro-S66E Tg tissues (Figure 6G, I). Since S66 is located within the first GTPase domain of Miro, we examined the effect of S66 phosphorylation status on GTPase activity. In vitro GTPase assays showed time-dependent GTP hydrolysis by the N-terminal GTPase domain of Miro-WT and Miro-S66D or Miro-S66E, with phosphomimetic Miro exhibiting greater GTPase activity than Miro-WT (Figure S6E). Moreover, prior phosphorylation by Polo significantly enhanced the GTPase activity of Miro-WT, but not Miro-S66A (Figure S6F). We conclude that Polo-mediated phosphorylation positively regulates the GTPase activity of Miro and the recruitment of Miro to ERMCS, where p-S66-Miro enhances the integrity of this ER-mitochondria tethering complex.

Miro Plays Conserved Roles in Regulating Ca²⁺_mito Homeostasis in Mammalian Cells

To test the conservation of the mechanisms of Miro function and regulation, we turned to the human cervical cancer cell line (HeLa) and mouse NSCs. Addition of IP3R agonist histamine to HeLa cells enhanced the recruitment of hMiro1 to the perinuclear ER area, and this effect was ameliorated by the co-treatment with BI2536, a specific chemical inhibitor of PLK1 (Figure 7A, B, S7B). Monitoring Ca²⁺_mito with Rhod-2 AM showed that Ca²⁺_mito was significantly increased in cells transfected with dMiro-WT or -S66E, but less so with dMiro-S66A (Figure 7C, 7D). Previous studies showed that hMiro1 (Fransson et al., 2003) or dMiro (Liu et al., 2012) caused aberrant mitochondrial morphology when overexpressed. We found that hMiro1-induced mitochondrial aggregation in HeLa cells was attenuated by treatment with 2-APB or Ru360, cell-permeant pharmacological inhibitor of IP3R or MCU, respectively (Tang et al., 2005; Garcia-Rivas Gde et al., 2006), or by chelating Ca²⁺ with BAPTA-AM (Figure 7E), suggesting that Miro-regulated Ca²⁺_mito homeostasis underlies its effect on mitochondrial morphology. Moreover, Myc-tagged dMiro-WT, -S66E, or -S66D exhibited more pronounced perinuclear and aggregated localization than Miro-S66A (Figure S7C). These results suggest that as in fly tissues, phosphorylation of Miro at S66 site promotes Miro localization to ERMCS in HeLa cells. We also examined the effect of Polo-Miro signaling on the integrity of the ER-mitochondria tethering complexes. The interaction between VDAC and IP3R was weakened by treatment with BI2536 (Figure 7F, 7G). Immunostaining with ER and mitochondrial markers reveal that ER-mitochondria interaction was increased in cell transfected with hMiro1-WT, but not with hMiro1-SA (Figure S7E, S7F), supporting a critical role of PLK signaling in ER-mitochondria interaction and Ca²⁺_mito uptake. Finally, we examined the effect of Miro on the proliferation of mouse NSCs. Transfection of NSCs with hMiro1-WT attenuated NSC proliferation, whereas hMiro1-S59A (equivalent of dMiro-S66A) was less effective in this aspect (Figure 7H). Mitochondrial metabolism as monitored by OCR was increased by hMiro1-WT but decreased by hMiro1-SA in mouse NSCs (Figure S7G). Supporting the importance of ER-mitochondria Ca²⁺ transfer in NSC regulation, pharmacological inhibition of Ca²⁺_mito uptake by Ru360, and to a lesser extent by 2-APB, attenuated NSC proliferation (Figure S7D).

Figure 7. Regulation of Ca²⁺_mito by Miro in Mammalian Cells.

(A, B) Immunostaining showing that the PLK inhibitor BI2536 blocked histamine-induced hMiro1 recruitment to perinuclear ER area in HeLa cells. B, quantification of Miro-Calnexin colocalization using Mander's coefficient.

(C, D) Effects of dMiro variants on Rhod-2 AM fluorescence in HeLa cells. D, quantification of Rhod-2 signals. RU360 treatment serves as control.

(E) Effects of pharmacological intervention of ER-mitochondrial Ca²⁺ transfer (RU360, 2-APB) or Ca²⁺ availability (BAPTA) on hMiro1-induced mitochondrial morphology change visualized with mitochondria-localized hMiro. Bar graph shows quantification of mitochondrial aggregation.

(F, G) Co-IP assay showing effect of PLK inhibition on IP3R-VDAC interaction in HeLa cells. G, quantification of signal intensity change shown in F.

(H) Effects of hMiro1-WT and hMiro1-S59A transfection on mammalian NSC (ReNcell CX) proliferation monitored over a 72h time course.

Error bar: SEM; *, p<0.05 in Student's t-tests. Scale bars, 100 μm.

See also Figure S7.

DISCUSSION

Mitochondrial and Ca²⁺ signaling pathways are intimately connected in many aspects of cellular physiology, from metabolism and ATP production to cell death. Previous studies have emphasized the key role of Miro in regulating mitochondrial transport by sensing local elevation of cytosolic Ca²⁺ concentration elicited by neuronal activity. This mechanism links mitochondrial distribution to the subcellular need for Ca²⁺ buffering. In this study we demonstrate a conserved role of Miro in regulating Ca²⁺_mito homeostasis in an apparently transport-independent manner. Miro localizes to ERMCS and interacts with its protein components. We further show that this function of Miro is regulated by Polo-mediated phosphorylation at a conserved site in the N-terminal GTPase domain. The identification of new regulators of Ca²⁺_mito homeostasis offers the opportunity to elucidate how Ca²⁺_mito uptake is controlled in different cell types of multicellular organisms, and how defects in this process may impact development and disease.

Our results identify the Polo/Miro signaling axis as an important regulator of Ca²⁺_mito uptake. ER is a major source of Ca²⁺_mito, which is transported to mitochondrial matrix through Ca²⁺ transporters located at the ERMCS. Despite detailed biophysical studies of Ca²⁺_mito uptake, little is known how this process is regulated to meet the developmental and physiological needs of distinct cell types in a multicellular organism. We show that Miro is localized to ERMCS, where it interacts with components of the ERMCS to modulate the integrity of this quasi-synaptic structure and the inter-organelle Ca²⁺ transfer. Our genetic and biochemical studies further identify Polo kinase as a key regulator of Miro-controlled Ca²⁺_mito homeostasis. Polo-induced phosphorylation of Miro promotes the localization of Miro to ERMCS, enhancing the interaction between Miro and components of the ER-mitochondria tethering complex and the integrity of the tethering complex. Our identification of S66 in the N-terminal GTPase domain as a key regulatory site of Miro is consistent with the known importance of this GTPase domain to Miro function (Kornmann et al., 2011; Babic et al., 2015). Our results also uncover a new mechanism of action for Polo in NSC self-renewal and differentiation. Dysregulation of PLKs has been implicated in cancer (Craig et al., 2014) and neurodegenerative diseases such as PD (Mbefo et al., 2010). Given that Miro OE-induced cell death involves Ca²⁺_mito overload, mitochondrial dysfunction, oxidative stress, ER stress, and caspase 3 activation, key features broadly implicated in neurodegenerative diseases, it would be interesting to test in various disease conditions whether disease-causing genes or signaling pathways converge on Miro, and whether pharmacological inhibition of PLK-Miro signaling may offer broad therapeutic benefits.

Our analyses of the effect of Miro LOF and GOF in the fly NB lineages reveal a critical role of Ca²⁺_mito homeostasis in cell fate determination. Miro LOF and GOF both impairs NB maintenance and lineage development, but through different mechanisms. Miro LOF causes Ca²⁺_mito depletion and reduced ATP production, impairing cell growth, a key requirement of NB maintenance (Song and Lu, 2011), and causing premature differentiation and loss of NB stemness through a non-apoptotic mechanism. In contrast, Miro GOF causes Ca²⁺_mito overload, oxidative stress, and activation of the apoptotic cascade, leading to failed maintenance of NB and IPs. These results reveal hypersensitivity of NSC maintenance and lineage progression to Ca²⁺_mito levels. A role for Ca²⁺ signaling in progenitor maintenance was previously shown in the fly hematopoietic system (Shim et al., 2013). However, distinct from the importance of Ca²⁺_mito to NB maintenance reported here, the maintenance of hematopoietic progenitors involves cytosolic Ca²⁺, which is mediated by CaMKII signaling and is sensitive to SERCA activity (Shim et al., 2013), whereas NSC maintenance by Miro-mediated Ca²⁺_mito homeostasis is insensitive to SERCA manipulation (data not shown). Together with previous studies implicating mitochondrial regulators or structural proteins in stem cell differentiation (Kasahara et al., 2013; Mitra et al., 2012; Lee et al., 2013; Teixeira et al., 2015), our result broadens the impact of mitochondria on developmental processes. Given the emerging connection between mitochondrial metabolism and epigenetic modifications (Wallace, 2010; Castegna et al., 2015), we speculate that there exist a mechanistic link between Miro-controlled Ca²⁺_mito homeostasis and epigenetic regulation in the nucleus. Future studies will establish, at the molecular level, how developmental pathways dictating stem cell self-renewal vs. differentiation decision are connected with signaling pathways regulating mitochondrial function.

EXPERIMENTAL PROCEDURES

Fly Genetics

Fly culture and crosses were performed according to standard procedures and raised at indicated temperatures. To generate UAS-dMiro-S66A and UAS-dMiro-S66E Tg flies, wild-type dMiro cDNA was modified using the QuikChange Multi kit (Stratagene) to introduce the S66A and S66E mutations. The insert of UAS-dMiro-WT-Myc was replaced with dMiro-S66A and dMiro-S66E cDNAs to generate UAS-dMiro-S66A-Myc and UAS-dMiro-S66E-Myc, respectively. The resulting pUAST-Miro constructs were injected into w^– embryos to obtain Tg lines according to established procedures (Bestgene, Inc.).

MARCM and Flip-Out Clonal Analysis

To generate NB MARCM clones and overexpression clones, larvae at 24h after-larvae-hatching (ALH) were heat-shocked at 37°C for 90 min and further aged for 72 hrs at 25°C before dissection. MARCM analyses were performed essentially as described (Song and Lu, 2011). For making overexpression clones, w, hsFLP; Actin 5c>CD2>Gal4, UAS-GFP-NLS was crossed with the indicated UAS lines and 24h ALH larvae were heat-shocked at 37°C for 90 min and further aged for 72 hrs at 25°C before dissection.

Pharmacological Treatment

2-APB (Sigma), Ru360 (Calbiochem), EDTA/EGTA (Sigma), CaCl₂ (Sigma), BAPTA (Life Technologies), Histamine (Sigma), Thapsigargin (Calbiochem), or BI2536 (Selleckchem) dissolved in 0.5% DMSO was mixed in instant Drosophila media or cell culture at the indicated final concentrations. 0.5% DMSO alone was used as vehicle control. For treatment of flies, embryos were collected on drug-containing food or control food for 6 hrs at 25°C and allowed to develop further to 120 hr ALH before larval brain dissection and immunostaining.

Immunohistochemistry

To generate the p-S66 Miro antibody, a synthetic peptide with the sequence SIVDFpSAVEQS was used to elicit immune response in rabbits. Purification of the phospho-specific antibody was performed by AbFrontier (Seoul, Korea). Purified p-S66 antibody was diluted at 1:200 for immunostaining. For larval brain immunostaining, larvae were dissected in Schneider's medium (Invitrogen) and fixed with 4% formaldehyde in PEM buffer (100 mM PIPES at pH 6.9, 1 mM EGTA, 1 mM MgCl₂) for 23 min at room temperature. For ER-mitochondria interaction analysis, confocal image stacks were automatically thresholded using ImageJ. Colocalization between organelles was quantified using the Manders’ algorithm (Manders et al., 1993).

Co-Immunoprecipitation and Western Blot Analyses

Approximately 25 adult fly thoraces were collected and homogenized in HBS buffer (210 mM Mannitol, 70 mM Sucrose, 5 mM HEPES pH 7.4, 1 mM EGTA) containing protease inhibitor (Sigma). The lysate was centrifuged at 1,300g for 10 min, and the supernatant was collected and centrifuged again at 13,000g for 15 min. For co-IP, the cytosolic and the mitochondrial fractions were incubated with anti-GFP or anti-Myc conjugated protein G-agarose beads (Upstate) at 4 °C for 2 hours. Beads were washed 3x with PBS for 5 min each. Proteins were eluted and analyzed by WB as described (Wu et al., 2013).

In Vitro Kinase and GTPase Assays

The conditions of the in vitro kinase assays were essentially as described (Lee et al., 2010). GST-Miro fusion protein production was done as described (Liu et al., 2012). Briefly, cDNAs encoding fragments of dMiro were cloned into a glutathione S-transferase (GST)-encoding vector, pGEX-6P-1 (GE Healthcare), and GST fusion production was carried out in E. coli strain BL21(DE3) cells according to Manufacturer's instructions. PLK1 and PKC-zeta were purchased from ProQinase. The GTPase assay for dMiro was performed according to a published procedure (Koshiba et al., 2011).

ATP Measurement

The level of ATP in Drosophila larval brain was measured essentially as previously described (Wu et al., 2013), using a luciferase-based bioluminescence assay kit (ATP Bioluminescence Assay Kit HS II, Roche Applied Science).

Calcium Imaging

For calcium imaging in primary neural cells, ~50 third instar Tg larval brains expressing Mito-GFP, Mito-GFP+UAS-dMiro-WT, Mito-GFP+UAS-dMiro-S66A, or Mito-GFP+dMiro RNAi driven by elav-GAL4 were collected. Primary neural cells were prepared according to published procedures (Egger et al., 2013). To monitor mitochondrial or cytosolic Ca²⁺ levels, primary neural cells were loaded with 1 μM Rhod-2 AM or Fluo-3 AM (Molecular Probes), respectively, in physiological salt solution [pH7.4] (150 mM NaCl, 4 mM KCl, 1 mM MgCl₂, 5.6 mM glucose, 5 mM HEPES) for 30 min at 37 °C. Primary neural cells were treated with 1 μM Thapsigargin (Calbiochem) for intracellular Ca²⁺ perturbation. Calcium imaging was carried out using an inverted confocal microscope (LSM 510 META and LIVE 5, Carl Zeiss) with a 40× objective.

For measuring mitochondrial Ca²⁺ in mammalian cells, HeLa cells were transfected with control pCDNA3.0 vector or Myc-hMiro1, Myc-hMiro2, Myc-dMiro-WT, Myc-dMiro-S66E, or Myc-dMiro-S66A plasmids cloned in pCDNA3.0. After transfection, HeLa cells were loaded with Rhod-2 AM (10 μM in DMEM media, Molecular Probe) for 30 min at room temperature. Cells were then washed with DMEM for 30 min. Single cell fluorescence was excited at 545 nm and images of the emitted fluorescence obtained on a Leica TCS SP5 AOBS confocal microscope were processed with LAS AF (Leica).

Statistical Analysis

Statistical significance of all data were evaluated by two-tailed paired or unpaired Student's t-tests. For studies employing multiple testing, we used a one-way ANOVA followed by Bonferroni post hoc test. All data were presented as mean±s.e.m.

Highlights.

• Mitochondrial Rho GTPase Miro regulates NSC maintenance and lineage maturation

• Miro controls mitochondrial Ca²⁺ homeostasis independent of its role in transport

• Miro interacts with Ca²⁺ transporters at the ER-mitochondria contact site (ERMCS)

• Miro activity is positively regulated by Polo kinase through phosphorylation