Gain-of-function screen identifies the role of miR-2 in mitochondrial homeostasis in Drosophila

LI Hao-Miao, ZHAO Mei-Qi, ZHANG Feng-Chao,SHEN Jie, ZHANG Jun-Zheng,*

(1. MOA Key Lab of Pest Monitoring and Green Management, Department of Entomology, College of Plant Protection,China Agricultural University, Beijing 100193, China; 2. College of Grassland Science and Technology,China Agricultural University, Beijing 100193, China)

Abstract: 【Aim】 The homeostasis of mitochondria is maintained by a wide range of molecular processes in eukaryotic cells. Very recently, microRNAs (miRNAs) are emerging as vital players of mitochondrial homeostasis. However, our understanding of how miRNAs regulate mitochondrial homeostasis is still incomplete. Our study aims to examine the roles and mechanisms of miRNAs in mitochondrial homeostasis regulation. 【Methods】 A gain-of-function screen was performed in Drosophila melanogaster during which miRNAs were over-expressed in the larval fat body cells and the mitochondrial integrity was monitored. The targets of miRNAs were predicted by bioinformatics methods and RNAi knock-down experiments were performed to examine their effects on mitochondrial morphology. 【Results】 A total of 106 miRNAs were screened in larval fat body of D. melanogaster, among which the over-expression of 21 miRNAs resulted in developmental defects of fat body. When over-expressed in the larval fat tissue of D. melanogaster, 10 miRNAs led to lethality at the early larval stage, while the other 11 miRNAs led to pupal lethality. Over-expression of miR-2 was found to result in abnormal aggregation of mitochondria in the larval fat body cells of D. melanogaster. Sequence analysis revealed that the Pink1 gene may be a target of miR-2. The expression level of Pink1 gene was shown to be down-regulated by miR-2 over-expression. Genetic interaction experiments demonstrated that over-expression of Pink1 was sufficient to rescue the abnormal aggregation of mitochondria caused by miR-2. 【Conclusion】 Our data support the view that miR-2 likely targets Pink1 to modulate mitochondrial homeostasis.

Key words: Drosophila melanogaster; miRNA; mitochondria; homeostasis; miR-2; Pink1

1 INTRODUCTION

Mitochondria are double-membrane organelles that generate chemical energy in the form of ATP with which cells carry out various vital functions (Mishra and Chan, 2014). During the power generating process, mitochondria interact with multiple metabolism pathways to utilize sugars, fats, and other metabolites as chemical fuels (Spinelli and Haigis, 2018). In addition to energy production, mitochondria perform multifaceted functions such as generation of reactive oxygen species (ROS), regulation of cell signaling and apoptosis (Kasahara and Scorrano, 2014). Therefore, it is not surprising that mitochondrial dysfunction has been implicated in neurodegenerative diseases, metabolic disorders, as well as tumorigenesis (Kasahara and Scorrano, 2014; Mishra and Chan, 2014; Giampazolias and Tait, 2016).

The homeostasis of mitochondria is maintained by balanced actions of mitochondrial biogenesis and degradation (Ploumietal., 2017). These regulatory machineries are particularly important for energetically demanding tissues such as muscles and nerves. Mitochondrial biogenesis is tightly regulated at the transcriptional level, while clearance of damaged mitochondria is majorly carried out by mitophagy, a selective form of autophagy (Palikarasetal., 2018). Recent discoveries have revealed that the mitochondrially targeted kinase Pink1 plays crucial roles in mitophagy (Ploumietal., 2017; Palikarasetal., 2018). In healthy mitochondria, Pink1 is cleaved by several proteases at the inner membrane and subsequently degraded by the ubiquitin-proteasome system (Clarketal., 2006). Mitochondrial damages lead to stabilization of Pink1 on the outer membrane, where it phosphorylates Parkin (Park) to stimulate its ubiquitin ligase activity (Parketal., 2006). Activated Park protein ubiquitinates several outer membrane components and facilitates mitophagy to remove the unhealthy organelle (Yangetal., 2006). Loss of function mutations ofPink1 andParkare strongly associated with familial forms of Parkinson’s disease, illuminating the importance of mitochondrial homeostasis in the health of the nervous system (Mishra and Chan, 2014; Spinelli and Haigis, 2018).

MicroRNAs (miRNAs) are small non-coding RNA molecules comprising of 18-25 nucleotides that can cause degradation or translational inhibition of target genes by binding to mRNAs, usually at the 3′-untranslated region (UTR) (Vendraminetal., 2017). Based on the mode of action, each miRNA is estimated to regulate tens of target genes while each gene can be regulated by multiple miRNAs (Geiger and Dalgaard, 2017). Most miRNAs did not show strong phenotypes when loss-of-function mutants were analyzed, but over-expression of miRNAs normally gives rise to observable effects (Chenetal., 2014). Ample evidence revealed that miRNAs regulate mitochondrial function by modulating expression of mitochondrial proteins encoded by nuclear genes (Tomasettietal., 2014). As of thePink1 gene, more than 20 miRNAs are predicted to bind at the 3′UTR region but only two of them have been experimentally validated (Wangetal., 2015; Kimetal., 2016). Clearly, our understanding of how miRNAs regulate mitochondrial function is incomplete.

Composed by a single layer of large polyploid cells, theDrosophilalarval fat body is recognized as a great model for studying cellular organelles. A transgenic fly line that marks mitochondria with the yellow fluorescent protein (mtYFP) has been successfully used to visualize mitochondriainvivo(LaJeunesseetal., 2004). Taking advantage of this genetic tool, we exploited the potential of studying mitochondrial homeostasis in the fat body of fly larvae. The mtYFP reporter was recombined with aCg-Gal4 driver (Hennigetal., 2006), enabling genetic manipulations in the larval fat body cells by the UAS-Gal4 system. The Hedgehog (Hh) signaling pathway inhibits fat body formation in early developmental stages (Pospisiliketal., 2010), and regulates autophagy as well as lipolysis in mature fat body cells (Jimenez-Sanchezetal., 2012; Zhangetal., 2020). Whether the Hh signaling is involved in mitochondrial homeostasis is unknown. The specificity and sensitivity of this system were examined with transgenes targeting the Hedgehog (Hh) signaling pathway as well as essential mitochondrial components.

Using the fat body mitochondrial morphology as readout, we report a gain-of-function screen inDrosophilamelanogasterthat leads to identification of specific miRNAs as modulators of mitochondrial function. We show that miR-2 is capable of regulating the mitochondrial homeostasis in the larval fat body. We further identifiedPink1 as the potential target regulated by miR-2invivo. These results provide novel insights into the molecular mechanisms by which miRNAs regulates mitochondrial function and laid foundation for further studies.

2 MATERIALS AND METHODS

2.1 Fly stocks

AdultD.melanogasterflies were maintained in standard medium and crosses were performed at 29℃. The RNAi stocks were obtained from the Tsinghua Fly Center (Table 1). Other RNAi stocks used in this study include:HLHm3 (TH01967.N),HLHm5 (TH01968.N),HLHmbeta(TH01969.N),Kr-h1 (TH01976.N),rpr(TH02185.N),CG4911 (TH02589.N),CG11665 (TH02820.N),Cdk4 (TH04358.N),Cht7 (TH04597.N),os(THU0981),awd(THU1022),upd2 (THU1288),E(spl) (THU2198),skl(THU3063),Rab5 (THU3215),Rab21 (THU3217),Sos(THU3521),Myd88 (THU3533),Pink1 (THU3783),Sik3 (THU3935),upd3 (THU4873),HLH106 (THU5848),HLH4 (THU5849),HLHmgamma(THU5851), andcas(THU5863). The transgenic miRNA lines ofD.melanogaster(Bejaranoetal., 2012) were obtained from Bloomington Drosophila Stock Center. A stock carrying both theCg-Gal4 driver (#7011) and thesqh-MitoEYFP(#7194) marker was generated by standard genetic recombination. The miR-2 sponge stock was obtained from Bloomington Drosophila Stock Center (#61367) and UAS-Pink1 was a gift from Dr. WANG Tao. The UAS-Ptc(Johnsonetal., 1995) and UAS-HhM1 (Palmetal., 2013) stocks have been described before.

2.2 Screen design and phenotype scoring

To gain a more thorough view of miRNA function in mitochondrial biology, we screened a collection of 106 transgenic miRNA lines (Bejaranoetal., 2012). Each transgenic miRNA line was crossed with theCg-Gal4;sqh-MitoEYFPstock individually. At least 50 F1progenies were examined for developmental defects. The mitochondrial morphology in the fat body of the 3rd instar larvae ofD.melanogasterfor each cross was examined. Similarly, each transgenic RNAi line was crossed with theCg-Gal4;sqh-MitoEYFPstock individually and examined for mitochondrial morphology.

Table 1 Phenotypes of larval development and mitochondrial morphology of Drosophila melanogaster after knock-down of mitochondrial components

Table 2 Phenotypes of larval development and mitochondrial morphology of Drosophila melanogaster after miRNA over-expression

2.3 Fat body preparation and microscopy

The fat body was dissected from the 3rd instar larvae ofD.melanogasterand mounted in 80% glycerol without fixation, and samples were examined within 4 h after preparation. The fluorescence images were acquired with a Zeiss Axio Imager Z1 microscope equipped with an ApoTome or a Leica SP8 confocal microscope. The figures were assembled in Adobe Photoshop CS5 with minor adjustments (brightness and/or contrast).

2.4 Real-time quantitative RT-PCR

The fat body was dissected from the 3rd instar larvae ofD.melanogasterand the total RNA was extracted by Trizol (Invitrogen, USA). The cDNA was prepared by the First-Strand cDNA Synthesis SuperMix for qPCR (TransGen Biotech, China). Real-time quantitative PCR was conducted using qPCR SYBR Green Master Mix (Yeasen Biotech, China) and the ABI QuantStudio 6 Flex System (Thermo Fisher, USA). Reactions were performed in a 20 μL reaction mixture containing 10 μL Master Mix, 50 nmol/L of each primer, 1 μL cDNA sample and nuclease-free water. The PCR protocol used consisted of a 20 s denaturation at 95℃ followed by 20 s annealing at 59℃ and 20 s elongation at 72℃ for 40 cycles. The housekeeping genebetaTub85Dwas used as the internal control. The primer pairs used for qPCR are as follows: Pink1-F: 5′-TCAATCCCAACCCGTCCAAG-3′; Pink1-R: 5′-CCACTGTAGGATCTCCGGACT-3′; Tub85D-F: 5′-CGGTCAATGCGGTAACCAGAT-3′; Tub85D-R: 5′-ACTATCGCCGTAGTACGTTCC-3′. The primer pairs were designed to amplify a 128 bp fragment ofPink1 (NCBI reference sequence: NM_001031878.2) and a 94 bp fragment ofbetaTub85D(NCBI reference sequence: NM_079566.4).

3 RESULTS

3.1 Fly larval fat body is feasible for mitochondrial homeostasis study

We found that transgenes that either activate or repress the Hh signaling were insufficient to alter the mitochondrial morphology when driven byCg-Gal4. As in wild type cells (Fig. 1: A), EYFP-Mito was distributed as punctate structures throughout the cytoplasm in Hh signaling defective fat body cells (Fig. 1: B, C). It has been shown that over-expressed Hh proteins could be secreted from the fat body and reach the wing imaginal disc through circulation (Palmetal., 2013). These external Hh proteins compete with endogenous molecules and dampen the signaling activity in the developing wing (Palmetal., 2013). In agreement with previous studies, Hh over-expression in the larval fat body cells byCg-Gal4 resulted in wing defects resembling Hh signaling reduction. The wings ofCg-Gal4>HhM1 flies displayed reduced distance between longitudinal vein L3 and L4, as well as disappearance of the anterior cross vein (Fig. 1: D, E), both are classical Hh deficient phenotypes (Johnsonetal., 1995). These experiments suggest that the Hh signaling functions specifically during fat body differentiation but is dispensable for mitochondrial homeostasis in the mature tissue.

The sensitivity of the mtYFP reporter was examined after knock-down of essential mitochondrial components by RNAi (Table 1). Disruption of mitochondrial function by reducing the expression of the electron transport chain complex subunits led to significant changes of the mitochondrial morphology. In general, respiratory chain defective mitochondria formed large aggregates, as shown by knock-down ofCIA30 (Fig. 2: A, B),Scox(Fig. 2: C),ND-18 (Fig. 2: D),ND-49 (Fig. 2: E) andND-5 (Fig. 2: F) by RNAi. Inhibition ofTMEM70 by RNAi led to clustering of mitochondria (Fig. 2: G). It is likely that acute malfunction of mitochondria overwhelms the homeostasis machinery, leaving behind many damaged mitochondria to be cleaned up (Zhouetal., 2019). Attenuation of the mitochondrial activity in larval fat body often resulted in systemic growth retardant, and eventually lethality before reaching the adulthood (Table 1). This observation is consistent with the findings that fat body serves as a requisite reservoir for stored lipid during metamorphosis (Lietal., 2019). Interestingly, we found that knocking down the expression ofParkresulted in mitochondrial aggregation in the larval fat body cells (Fig. 2: H, I). Such phenotype is consistent withParkmutants in adult tissues such as muscle and neurons (Clarketal., 2006; Parketal., 2006; Yangetal., 2006), suggesting that mitochondrial dynamics is controlled by the Park-mediated mitophagy pathway in the larval fat tissue (Zhouetal., 2019). The combination of a fat tissue specific Gal4 and mitochondrial fluorescent reporter thus provides an opportunity for systemically studies of mitochondrial homeostasisinvivo.

Fig. 1 Hh signaling is dispensable for mitochondrial homeostasis in larval fat body cells of Drosophila melanogaster

Fig. 2 Knock-down of essential components disturbs mitochondrial morphology in larval fat body cells of Drosophila melanogaster

3.2 A genetic screen identifies miRNAs involved in mitochondrial homeostasis

Fig. 3 Gain-of-function screen identifies miRNAs involved in mitochondrial regulation in larval fat body cells of Drosophila melanogaster

Overall, 83 miRNAs were insufficient to cause neither developmental defects nor visible changes of the mitochondrial morphology. Early larval stage lethality was resulted from over-expression of 10 miRNAs, and lethality at the pupal stage was induced by the over-expression of another 11 miRNAs (Table 2). The 11 miRNAs that caused pupal lethality were found to inhibit the formation of fat tissue, the remaining fat body cells were scattered and the mitochondrial morphology were aberrant (Fig. 3: B, C). For the 10 miRNAs that resulted in early larval stage lethality, very few fat body cells could be recovered but similar mitochondrial morphology defected were recorded (Fig. 3: D). Only two miRNAs were found to regulate mitochondrial morphology in the fat tissue without disrupting fly development. We found that in miR-992 over-expressing fat tissue, mitochondria became clustered and formed multiple rings inside one cell (Fig. 3: E). In particular, over-expression of miR-2b-1 led to aggregation of mitochondria, a phenotype resembling that of defective mitophagy caused byParkknock-down (Fig. 3: F). Therefore, we performed further genetic studies on the function of miR-2b-1 in mitochondrial homeostasis.

3.3 miR2 likely targets Pink1 to regulate mitochondrial health

We first conducted bioinformatics analyses to unveil the potential targets of miR-2b-1 by TargetScan (http:∥www.targetscan.org/). We next performed a small-scale RNAi screen targeting 25 potential target genes of miR-2b-1. Interestingly,Pink1 was the only target of miR-2b-1 showing mitochondrial phenotype in our RNAi screen. Knock-down ofPink1 in the fat body cells led to similar aggregation defects as over-expressing miR-2b-1 (Fig. 4: A, B). We further examined whetherPink1 and miR-2b-1 genetically interacts with each other to regulate mitochondrial homeostasis. Knock-down ofPink1 together with miR-2b-1 over-expression significantly enhanced the mitochondrial aggregation (Fig. 4: C).Pink1 over-expression alone does not cause marked changes in mitochondrial morphology (Fig. 4: D). However, whenPink1 was co-expressed with miR-2b-1 in the fat tissue, the mitochondrial aggregates disappeared and the mitochondria looked grossly normal (Fig. 4: E). The rescue efficiency ofPink1 is comparable with that of the miR-2 specific sponge (Fig. 4: F) (Loyaetal., 2009). The expression level ofPink1 gene was down-regulated by miR-2b-1 when examined by real-time quantitative RT-PCR (Fig. 4: G). Taken together, these results suggest that miR-2b-1 likely regulates mitochondrial homeostasis through suppressingPink1 expression.

Fig. 4 miR-2b-1 likely regulates Pink1 in larval fat body of Drosophila melanogaster

4 DISCUSSION

During our screen, we found that activation of miR-2b-1 in the larval fat body ofD.melanogasterled to mitochondrial homeostasis defects without disrupting the overall larval development. miR-2b-1 belongs to the conserved miR-2 family which is involved in developmental events such as suppressing apoptosis (Starketal., 2003; Leamanetal., 2005), embryonic patterning (Rödeletal., 2013), insect metamorphosis (Songetal., 2019), oogenesis (Lozanoetal., 2015) and wing morphogenesis (Lingetal., 2015). To our knowledge, the role of miR-2 in mitochondrial homeostasis has not yet been reported. It has been shown that miR-2 mutant flies displayed slightly higher viability and longer mean median lifespan (Chenetal., 2014). Interestingly, over-expression ofPink1 andParkis also reported to extend lifespan in fly by reducing proteotoxicity and altering mitochondrial dynamics (Todd and Staveley, 2012; Ranaetal., 2013). We reckon that miR-2 might function as a general inhibitor ofPink1 in various tissues and developmental stages, therefore removing of miR-2 elevatesPink1 expression to protect fly against aging related degeneration. Further studies are needed to demonstrate whether miR-2 directly regulatesPink1 expression and to which extent is miR-2 involved in mitochondrial homeostasis.

ACKNOWLEDGEMENTSWe thank Dr. WANG Tao, the Bloomington Drosophia Stock Center, the Tsinghua Fly Center for providing fly stocks. This work was supported by grants from National Natural Science Foundation of China (31772526 and 31970478 to ZHANG Jun-Zheng, and 31872295 to SHEN Jie). The authors declare that there is no conflict of interest.