The ability to resist and recover from external stresses is important for all living organisms that face stresses such as heat, reactive oxygen species, and irradiation during development and in the adult stage. Damaged cells need to be removed by apoptosis and replaced with newly formed cells by compensatory proliferation. The wing imaginal disc of Drosophila melanogaster is the primordium of the adult wing, and shows a very low level of cell death during normal larval development. In contrast, it shows extensive cell death by environmental stresses, and yet can develop into a normal wing even after 40% to 60% cell death (Jaklevic and Su, 2004; Karpen and Schubiger, 1981).

The centrosome consists of a pair of centrioles and pericentriolar materials (PCMs). DNA damage and mitotic defects cause the overduplication of centrosomes and the formation of multipolar spindles, leading to mitotic failure and cell death (Vakifahmetoglu et al., 2008). Defects in PCMs interrupt spindle assembly and activate the spindle assembly checkpoint (Torres et al., 2011). Fly Anastral spindle 3 (Ana3) is a PCM responsible for the cohesion of centrioles, prevention of premature centriolar segregation, and formation of basal bodies (Stevens et al., 2009). Ana3 and its mammalian homolog Rotatin (RTTN) contain multiple Armadillo repeats known to interact with Wnt signaling components and potentiate the Wnt pathway (Song et al., 2003). Wnt has critical roles in growth, development, adult homeostasis, and regeneration (Clevers and Nusse, 2012; Logan and Nusse, 2004; Raslan and Yoon, 2020). Ana3 and RTTN are also important for the formation of cilia and basal bodies (Kheradmand Kia et al., 2012; Stevens et al., 2009). Loss of RTTN causes polymicrogyria (PMG), situs inversus, isomerism, and heterotaxia in humans (Vandervore et al., 2019).

From a previous genetic screen, we found 28 mutants as responsible for the suppression of lethality caused by the overexpression of Sol narae (Sona) (Kim and Cho, 2020). In the present work, we identified one of suppressors as ana3^m19. Sona is a member of a disintegrin and metalloprotease with thrombospondin motif (ADAMTS) family (Kim et al., 2016). Most ADAMTSs are secreted proteases that cleave components in the extracellular matrix, and their malfunctions result in multiple diseases including cancer (Kelwick et al., 2015). Sona is positively involved in Wingless (Wg) signaling, and secreted by both the exosomal secretion pathway and Golgi transport (Kim et al., 2016; Won et al., 2019). Sona cleaves the linker region of extracellular Wg and generates a new functional form of Wg that is specialized in cell proliferation (Won et al., 2019).

Sona is important for cell survival, with the level of Sona correlated with the extent of cell survival (Tsogtbaatar et al., 2019). Cells expressing a high level of sona are cell autonomously resistant to γ-ray irradiation, while Sona secreted from these cells induces Cyclin D (Cyc D) in the neighboring cells for cell survival and proliferation in a non-cell autonomous manner. Interestingly, Wg-CTD but not full-length Wg induces Cyc D, which demonstrates that Sona is involved in intercellular communication to support the normal development of damaged tissues by regulating Wg signaling. Consistent with this, sona suppressors such as wntless, arrow, pou domain motif 3, and archipelago are related to Wg signaling (Han et al., 2020; Kim and Cho, 2020; Kim et al., 2016; Nam and Cho, 2020; Won et al., 2019).

We report here that Ana3 is also important for cell survival. Furthermore, overexpression of Ana3 increased the survival rate of irradiated flies, and the amount of Ana3 correlated with the extent of organism survival under irradiation. The level of Ana3 in wing discs was significantly increased by 1 h after irradiation, indicating that Ana3 may be one of the proteins that respond to irradiation at the front line. Ana3 expressed in S2 cells increased the level of both intracellular and secreted Sona by negatively regulating the lysosomal degradation pathway, which is consistent with the finding of ana3^m19 as a sona suppressor. Our data demonstrate a new role of Ana3 in the stabilization of Sona.

Fly lines

GMR-Grim, UAS-p35, and Gal4 drivers including ptc-Gal4, nub-Gal4, GMR-Gal4, en-Gal4, and act-Gal4 were obtained from Bloomington Drosophila Stock Center (USA). UAS-ana3 RNAi line was obtained from VDRC (V101280). Ubq-ana3-GFP, UAS-ana3-GFP, and ana3^SH0558 flies were kindly provided by Jordan W. Raff. UAS-sona lines were generated in our laboratory (Kim et al., 2016). Flies were incubated at 25°C unless otherwise indicated.

Adult wing mounting and size measurement

Wings of adult flies less than 3 days old were dissected and mounted in Gary’s Magic Mountant (Mixture of Canada Balsam and methyl salicylate, 4:1). For wing size, the entire wing except for the hinge region was measured as pixel numbers with ImageJ.

Cell culture, transfection, and preparation of various fractions

S2 cell lines obtained from DGRC were cultured in M3 media (S8398; Sigma-Aldrich, USA) containing 10% insect medium Supplementaryement (I7267; Sigma-Aldrich) at 25°C. Transfection was performed with cellfectin (Invitrogen, USA) according to the manufacturer’s instruction. pUAST-sona-HA (Kim et al., 2016) and pUAST-ana3-GFP (Stevens et al., 2009) were used for transfection.

To prepare the P100 fraction, conditioned media (CM) were sequentially centrifuged at 300g, 2,000g, 10,000g to remove dead cells and cell debris (Gross et al., 2012). The cleared CM was centrifuged at 100,000g to obtain two fractions. The supernatant fraction (SN_Δ) containing soluble proteins was concentrated with an Amicon ultra-4 centrifugal filter unit (Millipore, USA) for further analysis. The pellet P100 fraction containing extracellular vesicles including exosomes was washed with phosphate-buffered saline (PBS) and then pelleted again by one additional centrifugation at 100,000g to remove contaminants.

Immunocytochemistry

Sona-Pro and Sona-C antibodies were used to detect Sona protein as previously described (Kim et al., 2016). The Ana3 antibody was generously provided by Jordan W. Raff. For intracellular staining, fly larvae were dissected and fixed with PLP solution (2% paraformaldehyde, 0.1 M lysine, 0.25% sodium M-periodate) as described (McLean and Nakane, 1974). Fixed wing discs were blocked in block buffer (50 mM Tris pH 6.8, 150 mM NaCl, 0.5% NP-40, 5 mg/ml bovine serum albumin [BSA]) for 2 to 6 h at 4°C and washed in wash buffer (50 mM Tris pH 6.8, 150 mM NaCl, 0.5% NP-40, 1 mg/ml BSA). Antibodies were diluted in wash buffer and incubated overnight at 4°C or 2 h at room temperature. After washing several times, samples were treated with DAPI and mounted with a Vectashield mounting medium. Sample images were captured with an LSM laser (Zeiss, Germany) scanning confocal microscope and processed by Adobe Photoshop. Sona-Pro (rabbit, 1:200), Sona-C (mouse, 1:100), Ana3 (Stevens et al., 2009) (rabbit, 1:100), and cleaved Dcp-1 (Cell signaling Asp216, rabbit, 1:100) were used.

For extracellular staining, larvae were dissected in ice cold M3 medium and incubated in M3 medium containing primary antibody for 2 h at 4°C as described in (Strigini and Cohen, 2000). Then, wing discs were washed with M3 media and PBS twice each and fixed in 4% paraformaldehyde in PBS for 50 min. Samples were then processed the same way as described above, except that the buffers did not contain any detergent. The antibodies were used at a 10-time higher concentration for extracellular staining than for intracellular staining.

Western analysis

Western analysis was performed as described (Kim et al., 2016). Protein bands were visualized with West Pico Plus Chemiluminescent Substrate (Thermo Scientific, USA) for an ECL system (Kim et al., 2016). Sona-Pro (rabbit, 1:5,000), HA (Roche 3F10, rat, 1:2,000), Ana3 (Stevens et al., 2009) (rabbit, 1:1,000), Syntaxin 1A (DSHB 8C3, mouse, 1:1,000), GAPDH (MA5-15738, mouse, 1:2,000; Invitrogen), and α-tubulin (T9026, mouse, 1:5,000; Sigma-Aldrich) were used. All samples except Ana3 were loaded in 10% SDS-PAGE gel. Samples for the Ana3 blot were loaded in 7% SDS-PAGE gel.

Irradiation condition

Flies were cultured at 25°C and irradiated during the second or third instar larval stage at the indicated γ-ray doses with a Gammacell 3000. They were cultured and dissected at the designated time or at the late 3rd instar larval stage for immunocytochemical analysis.

Statistical analysis

Statistical analysis was performed based on Student’s t-test to compare controls with experimental groups by using oneway analysis of variance (ANOVA). Data are pre-sented as mean ± SD. The t-tests were used to determine the statistical significance of the difference between two groups. A P value < 0.05 was considered statistically significant.

The m19 sona suppressor has a lethal mutation in the ana3 gene

We previously reported an ethyl methanesulfonate (EMS)-based genetic screen for the identification of suppressors in their heterozygous forms that survive against lethality induced by an overexpression of Sona (Kim and Cho, 2020; Kim et al., 2016). One of the suppressors, m19, was chosen among the 28 established sona suppressors for this study in order to understand the role of the m19 gene in Sona function.

The m19 was embryonic lethal, and its lethal site was mapped between cinnabar (cn) and curved (c) on the right arm of the second chromosome by meiotic mapping (Fig. 1A). Deficiency mapping located the lethal site of m19 in the region between 48F1 and 49A1 that is deleted in Df(2R)BSC425 and Df(2R)Exel6061 (Fig. 1B). In this region, 11 genes were present, so we carried out complementation tests by crossing m19 and available mutants in these genes. We identified ana3^SH0558, a loss of function mutant of ana3 that does not complement the lethality of m19 (Stevens et al., 2009). Although the homozygous ana3^SH0558 was pupal lethal, the transheterozygous m19/ana3^SH0558 was embryonic lethal because m19 is more severe than ana3^SH0558 in the loss of the ana3 phenotype.

Whole-genome sequencing revealed that m19 has a single base change from C to T near the end of the ana3 coding region (Fig. 1C). This mutation changed a glutamine residue to a stop codon that truncates 252 amino acid residues at the carboxyl terminal region from the total 1,977 amino acid residues of wild-type Ana3 (Fig. 1D). This truncated region was the most conserved region in Ana3 homologs in other model organisms, showing 20% to 23% identity and 34% homology and containing Armadillo repeats (Song et al., 2003). Ana3 had at least eight different regions containing multiple Armadillo repeats (Fig. 1D).

Knockdown of ana3 results in small and deformed wings and hinges

To test whether the m19 suppressor is indeed ana3, we examined the loss and gain of Ana3 effects on Sona. Before we initiated this work, we confirmed that Ubq-ana3-GFP, UAS-ana3-GFP, and UAS-ana3 RNAi lines were effective in changing the level of Ana3 protein (Supplementary Fig. S1). Overexpression of Ana3 in actin-Gal4>UAS-ana3-GFP (act>ana3-GFP) and Ubq-ana3-GFP flies induced no morphological changes (Figs. 2A-2C). In contrast, knockdown of ana3 in tubulin>ana3 RNAi (tub>ana3i) flies induced embryonic lethality. Therefore, the gain of Ana3 did not affect fly morphology, but the loss of Ana3 caused developmental defects.

Knockdown of Ana3 induced morphological changes in wings and hinges. Wings of cubitus interruptus (ci)>ana3i, patched (ptc)>ana3i, and engrailed (en)>ana3i flies at 25°C were smaller than control wings by 29%, 35%, and 14%, respectively, and those of nub>ana3i flies remained folded (Figs. 2D-2J, Supplementary Figs. S2A and S2B). The size of the wings was decreased throughout the entire blade regardless of the Gal4-expressing regions but more so along the proximodistal axis than the anteroposterior axis (Figs. 2D-2I). Knockdown of ana3 with nub-Gal4 and ptc-Gal4 at 29°C induced more severe phenotypes than at 25°C (Supplementary Fig. S2). In addition, female ptc>ana3i flies were pupal lethal at 29°C with only 3% escapers while male ptc>ana3i flies with small wings were not lethal, showing that knockdown of ana3 induced more severe phenotypes in females than in males (Fig. 2J, Supplementary Figs. S2C-S2F).

Ana3 was also important for the growth in the region between the hinge and the anterior cross vein. Compared to control hinges, hinges of ci>ana3i, ptc>ana3i, and en>ana3i flies were significantly smaller due to the reduction in the size of the anterior costal cells, alula, and 1st basal cells (Fig. 2K). In sum, Ana3 is important for the development and growth of wings and hinges (Figs. 2K-2N).

ana3 and sona show a mutually positive genetic interaction

To examine the genetic interaction between sona and ana3, we then examined whether the knockdown of ana3 also rescues the lethality induced by sona overexpression, because ana3^m19 as a heterozygous form successfully rescued the lethal phenotype of sona overexpression in the original genetic screen (Fig. 3A). All ptc>sona flies were embryonic lethal and 79% of ptc>ana3i flies developed to adults at 25°C. In contrast, 85% of ptc>sona; ana3i flies survived to adults. Similarly, all en>sona flies were embryonic lethal and 88% of en>ana3i flies developed into adults, while 90% of en>sona; ana3i flies developed into adults (Fig. 3A). Thus, knockdown of ana3 substantially rescued the Sona-induced lethal phenotype.

Coexpression of sona reciprocally rescued the small wing phenotype caused by knockdown of ana3. The small wing phenotype of ptc>ana3i and en>ana3i flies was rescued up to 99% and 94% by the coexpression of sona, respectively (Figs. 3B-3H). The small hinge phenotype of en>ana3i flies was also partially rescued in 20% of en>ana3i; sona adult flies (n = 20, Fig. 3G). Based on this mutually positive genetic interaction between ana3 and sona, as well as the failed complementation between m19 and ana3^SH0558, we concluded that the suppression of the Sona-induced lethality in m19 is due to a lethal mutation in the ana3 gene, which we name ana3^m19.

Loss of ana3 results in apoptosis

Sona as well as some centrosomal proteins such as CEP-55 and RassF7 is important for cell survival (Gulsen et al., 2016; Kalimutho et al., 2018; Tsogtbaatar et al., 2019). Based on the close genetic relationship between ana3 and sona, we tested whether the small wing phenotype by knockdown of ana3 is also due to cell death. Indeed, dying cells with activated Death caspase 1 (Dcp-1) were detected in en>ana3i wing discs (Figs. 4A and 4B). When ana3i and p35 were coexpressed by en-Gal4, the Dcp-1 signal disappeared, establishing that knockdown of ana3 induces apoptosis (Figs. 4C and 4D).

To examine whether the inhibition of cell death by the caspase inhibitor p35 rescues the adult wing phenotype induced by the knockdown of ana3, we compared wings of en>ana3i; p35 flies to those of en>ana3i and en>p35 flies at 29°C (Figs. 4E-4H). In fact, the wing size of en>GFP, ana3i; p35 flies was 93% of the control wings, whereas those of en>GFP, ana3i and en>GFP, p35 were only 77% and 67% of the control wings, respectively (Fig. 4I). Knockdown of ana3 rescued defects in both the shape and the size of the wings and hinges caused by p35 overexpression (Figs. 4E-4H). Thus, apoptosis is responsible for the phenotypes induced by the knockdown of ana3.

We then asked whether overexpression of ana3 can suppress the cell death induced by the overexpression of Grim, one of apoptotic proteins (Chen et al., 1996). GMR>Grim/+ flies had small and rough eyes, but GMR>Grim; ana3-GFP flies had normal eyes just slightly smaller than control GMR/+ eyes (Supplementary Figs. S3A-S3C). This result raised an interesting possibility that Ana3 may actually promote cell survival in cell-death-causing conditions.

Overexpression of Ana3 increases cell survival

We previously reported that the gain of Sona is capable of promoting organism survival under irradiation (Tsogtbaatar et al., 2019). To test whether the gain of Ana3 is also able to promote organism survival, we irradiated Ubq-ana3-GFP, CS, and ana3^SH0558/CyO-GFP larvae at 1,000, 1,250, and 1,500 rad, and counted the number of adult survivors to obtain survival rates (Fig. 5A). The survival rate of ana3^SH0558/CyO-GFP flies was only 47%, while those of Ubq-ana3-GFP and CS flies were about 90% in the unirradiated condition (Fig. 5A, the darker set of bars). To compare the survival rates of these flies in unirradiated vs. irradiated conditions, the survival rates in the unirradiated condition were converted to 100%. At 1,000 rad, the survival rates of both Ubq-ana3-GFP and CS flies were 97% but that of ana3^SH0558 heterozygotes was only 51%, showing that the half amount of Ana3 in ana3^SH0558 heterozygotes is insufficient to resist radiation damage. At 1,250 rad, the survival rate of Ubq-ana3-GFP was 80%, while those of CS and ana3^SH0558 heterozygotes were 46% and 36%, respectively, showing that the high level of Ana3 in Ubq-ana3-GFP flies is responsible for survival against radiation damage. At 1,500 rad, over 90% of all flies died regardless of genotype.

Consistent with the role of Ana3 in survival against radiation, CS wing discs had a high level of Dcp-1 signals while Ubq-ana3-GFP wing discs had no Dcp-1 signals 24 h after irradiation at 1,500 rad (Figs. 5B and 5C). Furthermore, Ubq-ana3-GFP flies developed one day earlier than CS flies in both unirradiated and irradiated conditions (Supplementary Fig. S3D). In case of 4,000 rad, Ubq-ana3-GFP flies entered the pupal stage two days earlier than the CS flies. In sum, Ana3 can increase survival rates under irradiated conditions and accelerate fly development.

To confirm whether the amount of Ana3 correlates with the survival rate, we carried out the following experiment including act>ana3-GFP larvae that express a larger amount of Ana3 than Ubq-ana3-GFP (Supplementary Figs. S4A-S4D). CS, Ubq-ana3-GFP, and act>ana3-GFP larvae were irradiated with γ-rays and their survival rates were measured (Supplementary Fig. S4E). At 1,500 rad, the survival rates of CS and Ubq-ana3-GFP were below 10% but that of act>ana3-GFP was 67%. At 2,000 rad, the survival rates of CS and Ubq-ana3-GFP flies were less than 5% but that of act>ana3-GFP flies was 16%. These results demonstrate that the amount of Ana3 correlates with the survival rate of irradiated flies.

If the amount of Ana3 correlates with the survival rate against irradiation, it is possible that irradiation triggers an increase in the amount of Ana3. Indeed, the level of Ana3 in CS wing discs was increased 4 times in 1 h and then reduced to the normal level 6 h after irradiation (Figs. 5D and 5E). This shows that the amount of Ana3 is dynamically regulated by irradiation, which may be prerequisite for cell survival against irradiation.

Both intra- and extracellular levels of Sona are decreased in ana3 mutants

Since overexpression of Sona also increases the survival rate of flies after irradiation (Tsogtbaatar et al., 2019), and ana3 has a positive genetic interaction with sona, we reasoned that the level of Ana3 may correlate with that of Sona in vivo. We therefore examined the amount of Sona in CS and ana3^SH0558 wing discs visualized with two anti-Sona antibodies, Sona-Pro antibody that recognizes the pro domain and Sona-C antibody that recognizes the carboxyl region (Kim et al., 2016). Intracellular levels of Sona in ana3^SH0558 wing discs were lower than those in control discs by 29% and 40% with Sona-Pro and Sona-C antibodies, respectively, whereas the levels of Dlg were similar (Figs. 6A-6I, Supplementary Fig. S5). Furthermore, the extracellular levels of Sona in ana3^SH0558 wing discs visualized by Sona-Pro were only one fourth of control wing discs in the peripodial epithelium (Figs. 6C and 6D). The level of extracellular Sona in the basal region of the disc proper was also substantially reduced in ana3^SH0558 wing discs (Figs. 6E and 6F). When the level of Ana3 is reduced in the anterior region of ci<GFP, ana3i discs, the extracellular level of Sona was also reduced (Figs. 6G, 6G’, 6H, and 6H’). Therefore, the level of Ana3 positively correlates with that of both intra- and extracellular Sona, suggesting that Ana3 regulates the level of Sona in vivo.

Ana3 is secreted via exosomes

To biochemically verify the role of Ana3 in regulating the in vivo level of Sona, we utilized the S2 cell culture system. S2 cells transfected with actin-Gal4 and different combinations of UAS-ana3-GFP and UAS-sona-HA plasmids were cultured for 3 days, and then the CM after removal of dead cells and cell debris was centrifuged at 100,000g for 3 h to obtain two fractions: the P100 pellet fraction that contains exosomes and the supernatant SN_Δ fraction that contains soluble extracellular proteins secreted by Golgi transport (Gross et al., 2012). The 225 kDa Ana3 protein endogenously expressed in S2 cells and the 252 kDa Ana3-GFP protein were detected in cell extract (CX) regardless of Sona (Fig. 7A). Both Ana3 forms were unexpectedly detected in the P100 fraction, indicating that Ana3 is secreted to extracellular space. Endogenous Ana3 secreted from untransfected S2 cells demonstrated that the presence of Ana3 in the P100 fraction is not due to the overexpression of Ana3 (second row in Fig. 7A). Most extracellular Ana3 was present in the P100 fraction, although some faint bands were detected in the SN_Δ fraction. Therefore, we concluded that extracellular Ana3 is secreted mostly via exosomes.

Ana3 increases the level of exosomal Sona by protein stabilization

To find out whether Ana3 regulates the level of Sona in S2 cells, S2 cells were transfected with increasing amounts of UAS-ana3 cDNA with a constant amount of UAS-sona-HA and act-Gal4 cDNAs. We previously reported that 75 kDa full-length Sona is processed intracellularly, while 37 kDa active Sona and 20 to 25 kDa pro domain fragments are secreted to the extracellular space by both the exosomal secretion pathway and conventional Golgi transport (Won et al., 2019) (Fig. 7B). The 75 kDa full-length Sona-HA is detected by both Sona-Pro and anti-HA antibodies while 20 to 25 kDa pro domain fragments and 37 kDa active Sona-HA are detected by Sona-Pro and anti-HA antibody, respectively (Won et al., 2019) (Fig. 7B). The amounts of these Sona products in both CX and precleared CM were increased as the amount of Ana3 was increased. Since the transcription of transfected sona cDNA is carried out by Gal4 protein expressed by the actin promoter, the increase in the Sona level is due to the stabilization of Sona by Ana3 in S2 cells. Especially, the amounts of 20 to 25 kDa fragment and active Sona-HA were dramatically increased. This indicates that Ana3 specifically stabilizes active Sona in P100, SN_Δ, or both fractions.

We have previously shown that Arrow (Arr) secreted via exosomes can stabilize Sona when it is added to the culture media of Sona-expressing S2 cells (Han et al., 2020). Since both Ana3 and Arr stabilize exosomal Sona and are secreted via exosomes, we reasoned that exosomal Ana3 may also stabilize Sona. To test this possibility, we purified exosomes from the culture media of S2 cells transfected with ana3 cDNA and then added the obtained exosome fraction to the culture media of S2 cells transfected with sona cDNA. Unlike exosomal Arr, however, exosomal Ana3 did not noticeably increase Sona stability (Supplementary Fig. S6).

To identify which fraction of extracellular Sona among P100 and SN_Δ fractions is increased by Ana3, the CM was further fractionated to P100 and SN_Δ fractions. Similar to the data in Fig. 7B, the level of full-length Sona was not noticeably changed when 0.5 to 1 μg of UAS-ana3-GFP and 1 μg of UAS-sona-HA (0.5 to 1:1 ratio) were co-transfected (arrow in the left panel of Fig. 7C), while the amount of active Sona slightly increased in CX (arrowhead in the left panel of Fig. 7C). Interestingly, the amount of active Sona-HA in the P100 fraction was greatly increased while that in the SN_Δ fraction showed little change (right panel of Fig. 7C). Therefore, Ana3 specifically increased the level of exosomal Sona when Sona is coexpressed with Ana3.

We have shown in this report that the ana3^m19 mutant is a suppressor of Sona-induced lethality (Fig. 1). Fly ana3 has a positive genetic interaction with sona that encodes a metalloprotease involved in Wnt signaling, establishing a potential link between Ana3/RTTN and Wnt signaling. Ana3/RTTN is a peripheral member of the centrosome complex whose malfunction leads to embryonic lethality in both ana3 mutant flies and RTTN knockout mice (Faisst et al., 2002; Stevens et al., 2009). Both Ana3 and Sona are involved in cell survival and resistance to irradiation. Consistent with their positive genetic interaction and functional similarity, we found that Ana3 stabilizes Sona and increases the level of Sona in both wing discs and S2 cells. The truncated region in Ana3^m19 protein is the most conserved region in Ana3/RTTN homologs, suggesting that this region plays a key role in stabilizing Sona. Some PMG mutations have also been identified in the Armadillo repeats in this carboxyl region of RTTN protein (Stouffs et al., 2018).

Both lethality and the small wing phenotype induced by Sona overexpression were completely rescued by knockdown of Ana3, suggesting that one of the main functions of Ana3 is to stabilize Sona (Fig. 3). It is worth noting that a degradation of Sona occurs in the lysosome but not in the proteasome complex, as well as that another sona suppressor Arr also stabilizes Sona (Han et al., 2020). Since our original genetic screen was aimed at identifying suppressors that reduce Sona activity, it makes sense that both ana3 and arr mutants are identified as sona suppressors. Interestingly, Ana3 dramatically increased the level of exosomal Sona but not soluble Sona (Fig. 7). This suggests that Ana3 stabilizes Sona in the exosomal secretion pathway that is interconnected with the lysosomal degradation pathway and the endosomal pathway but not in Golgi transport (Repnik et al., 2013).

The loss of ana3 induced cell death, which is a common phenotype of centrosome components (Fig. 4). Interestingly, overexpression of Ana3 enhanced the survival rate of irradiated flies, with wing discs showing the increased level of Ana3 1 h after irradiating the larvae, indicating that signals initiated by irradiation increase the level of Ana3 to prevent cell death. Since Ana3 stabilizes Sona, and knockdown of ana3 completely rescues the lethality caused by overexpressed Sona, the ability of Ana3 in promoting cell survival may stem from stabilized Sona. We have previously shown that Sona-expressing cells are resistant to irradiation in a cell autonomous manner, and Sona secreted from these cells enables neighboring cells to survive and proliferate in a non-cell autonomous manner (Tsogtbaatar et al., 2019). Thus, it is possible that the increased level of Ana3 by irradiation contributes to increasing the level of Sona, which in turn functions to promote cell survival in both cell-autonomous and non-cell autonomous manners.

Extracellular Sona cleaves Wg and generates Wg-CTD that increases the level of Cyc D for initiating cell cycles (Won et al., 2019). Cyc D1 in mammalian cells promotes cell proliferation in response to mitogens, but overexpression of Cyc D1 leads to centrosome amplification, deregulation of the mitotic spindle, and chromosome abnormalities (Nelsen et al., 2005). Cyc D1 is oncogenic in many human cancer cells because it contributes to malignant transformation, with centrosome amplification by ras oncogene depending on Cyc D1 (Zeng et al., 2011). The link between fly Cyc D, Sona, and Wg-CTD, as well as the association of many components in Wnt signaling such as Disheveled, Armadillo/β-catenin, Axin, and Arrow/LRP6 with centrosomes, suggests that Sona may participate in the regulation of centrosomal duplication for the initiation of cell cycles (Kaplan et al., 2004).

We thank our lab members for critically reading this manuscript and giving valuable suggestions. We especially thank Dr. Jordan W. Raff for the generous offering of all fly lines, antibodies, and constructs. We also thank Bloomington Drosophila Stock Center, Drosophila Genetic Resource Center, and Developmental Studies Hybridoma Bank for fly strains and antibodies. This research was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Education, 2017R1A2B4009254 and 2019R1H1A2039726.

D.G.C., S.S.L., and K.O.C. designed the experiments and analyzed the data. D.G.C. and S.S.L. conducted the experiments. D.G.C. and K.O.C. wrote the paper.

The authors have no potential conflicts of interest to disclose.