Drosophila melanogaster stocks were obtained from the Bloomington (BL) or the Vienna (VDRC) Drosophila stock centers. The following Gal4 drivers for targeted expression were used: Mef2Gal4, 24BGal4, Dlmo-Gal4 (MS1096 BL:8860), 69B-Gal4 (BL:1774), en-Gal4 (BL:30557;30564). The following RNAi lines were used to knock down gene expression: UAS-prto-RNAi (VDRC: v30420), UAS-prto-RNAi#2 (BL:53700), UAS-GFP-RNAi (BL:9330), UAS-stv-RNAi (BL:42564), UAS-NUAK-RNAi (BL:31885), UAS-cher-RNAi (BL:26307), UAS-Hsc70-4 RNAi (BL:54810), UAS-atg8a RNAi (BL:28989). The UAS-Dcr stocks (BL:24650; BL:24651) were used to increase the RNAi knock-down efficiency. The FUCCI stock to label cell cycle progression was used (BL:55100). The G203-GFP; Sls-GFP, UAS-Cher-90-GFP stocks were kindly gifted by M. Landgraf and J. Ylänne respectively. In addition, stv¹ and NUAK null alleles and UAS-stv-V5 stocks were kindly provided by E. Geisbrecht. The UAS-prto-FL and UAS-GFP-prto and UAS-Flag-prto transgenic flies were newly generated (see below).

Fly crosses were grown in standard cornmeal medium at 25°C but L1 larvae offspring were also shifted to other temperatures such as 18°C and 29°C as stated for controlling the UAS-construct expression level. For climbing assays crosses were kept at 25°C and flies were raised in rich molasses medium to stimulate the number of individuals reaching adulthood.

The following primary antibodies were used: mouse anti-GFP (1:500; Roche), rabbit anti-GFP (1:400; Molecular Probes), mouse anti-FLAG (1:500; Sigma), mouse anti-Kettin/Sls (1:10; Developmental Studies Hybridoma Bank [DSHB]), mouse anti-Laminin (1:20; DSHB), rabbit anti-active Caspase3 (1:1000; BioLabs), mouse anti-Phosphohistone H3 (1:300; Cell Signalling). Secondary Donkey anti-mouse Alexa 488 and anti-rabbit-Alexa 555 conjugated antibodies were used (1:400; Molecular Probes) and Donkey anti-mouse Rhodamine and anti-rabbit-FITC antibodies were used (1:200; Jackson ImmunoResearch). Rhodamine and Cy5-congugated phalloidin were used to label actin filaments (1:50; Molecular Probes). DAPI (Sigma) was utilized for revealing nuclei.

Late 3^rd instar wandering larvae were pinned and then filleted in a 3% agar plate using surgical scissors and then fixed with 4% Paraformaldehyde in PBS for 20 min. Body wall muscles were rinsed 3 times with PBT (PBS with 0.5% Tween) and then blocked with PBT with 2% BSA (Sigma) and 10% horse serum (Vector Laboratories) for 1 h at room temperature. Primary antibodies were incubated overnight at 4°C in blocking buffer. Samples were then washed 3 times with PBT then incubated with secondary antibodies in blocking buffer for 2 h at room temperature or 4°C overnight. The secondary antibodies were then removed, and samples washed 3 times with PBT and 3 times with PBS. Samples were then incubated with fluorescent conjugated Phalloidin and DAPI for 45 min at room temperature to reveal the actin cytoskeleton and myonuclei. Dissection and staining of adult indirect thoracic flight muscles were carried out as described in Magny et al. (2013). Immunocytochemistry of imaginal discs was performed following the protocols described in Galindo et al. (2007).

Embryos staged at 0–22 h AEL were collected and fixed by shaking in 4% PFA: Heptane (1:1) fixative for 20 min. Fixed embryos were washed in PBT (1x PBS and 0.1% Tween) five times, and then incubated in 2.5 µL of proteinase K diluted in 1 mL of PBT with rotation for 2 min. Embryos were then washed twice with 2 mg/ml glycine in PBT followed by several PBT washes and incubation with rotation for 10 min in hybridization solution (HS) diluted 1:1 in PBT. Samples were then pre-hybridized for 1–2 h in 1 mL HS at 55°C followed by hybridisation with the selected probes at 50 ng of labelled probe in 200 µL of HS at 55°C overnight without rotation. After the overnight incubation and post hybridisation washes, embryos were then blocked with 0.1% BSA in PBT for 30 min followed by incubation in a dilution 1:2000 of mouse anti-Dig antibody (Roche) in PBT-0.1% BSA with shaking for 3–4 h. Detection of transcripts was achieved by using a substrate solution (9 µL NBT and 7 µL BCIP in 2 mL of Genius 3 buffer (100 mM Tris-HCl (pH 9.5), 10 mM NaCl, 50 mM MgCl₂) in the dark at room RT. For imaginal disc and larval muscles, protocol used was described in Galindo et al. (2007), and Magny et al. (2013).

S2 cells were cultured in 10% FBS (v/v) Schneider’s media at 25°C to a confluent stage. Transfections and staining protocols were carried out as described in Pueyo et al. (2016b). We used primary rabbit and mouse anti-Flag antibody (1:500; Sigma) and secondary Donkey anti-mouse FITC (1:1000; Jackson ImmunoResearch), anti-mouse Rhodamine (1:400;Jackson ImmunoResearch) and anti-rabbit Cy5 (1:400; Jackson ImmunoResearch) for detection of tagged-smORF peptides. For detection of mitochondria S2 cells were incubated in 500 nM Mitotracker Red CMXRos (Life Technologies) for 45 min and then fixed for immunohistochemistry. For Lysosome colocalization experiments, cells were incubated in Lysotracker-DND99 (1:1000; Molecular Probes) for 15 min, then mounted and observed in vivo. Images of 5–10 cells per transfection culture were captured on the LSM510 Axioskop 2 or on the Leica TCS SP8 confocal microscopes. For colocalization analysis Z-stack images were taken and analysed with the ImageJ plugin “Mander’s Coefficients” which was used to calculate Pearson’s correlation coefficient of tag to tag signal in 2 different channels.

Extraction of RNA from S2 cells was achieved using manufacturer’s TRIZOL RNA protocol, and subsequently a cDNA library was generated using random primers and Moloney Murine Leukemia Virus Reverse Transcriptase Kit (Invitrogen). Amplification of the full length Prto cDNA was achieved by PCR (Qiagen), using the forward primer 5′ GAG​AGA​GAA​TTC​CTC​TAA​TCA​GCT​GTT​TGG​TTT​GT 3′ and the reverse primer 5′ CAC​ACA​CTC​GAG​ATT​TGG​ATG​TTT​GTT​TTA​CTT​AGC 3′ containing EcoRI and XhoI restriction sites respectively. Restriction sites were used for subsequent cloning of the Prto PCR product in the pUASt vector. Similarly, Prto ORF fragments for N-terminal and C-terminal tagging were amplified as above using the following primers: Ct-forward 5′CAC​CCT​CTA​ATC​AGC​TGT​TTG​GTT​TGT, and Ct-reverse 5′ GTC​ATC​CTC​GTC​GTT​GCG​CAC; Nt-forward 5′ CAC​CTC​CGC​ATC​CGC​TGC​CCG​A and Nt-reverse 5′ ACA​TTT​GGA​TGT​TTG​TTT​TAC​TTA​GCT​T. PCR products were cloned in the pENTR™ Directional TOPO® Cloning vector (Invitrogen). Integration of Prto ORF fragments to final destination vectors (T. Murphy, unpublished results; obtained from the Drosophila Genomics Research Center) was achieved by LR recombination following the Gateway system (Invitrogen). UAS-prto constructs were verified by sequencing and injected to generate transgenic flies (BestGene).

Quantification of muscle defect in L2 and L3 whole larvae was achieved by isolating larvae expressing GFP trapped protein labelling the Z-disc in control or prto KD. Larvae were drowned in 70% ethanol for 1 h and mounted in glycerol for visualisation. Images were captured using the fluorescent Zeiss microscope and Zen 2.6 software. Muscle phenotypes were categorised as truncated or thin muscles.

Pupal case phenotypes were analysed by calculating the axial ratio which is the length of the pupae divided by the width as described in Brooks et al. (2020).

Adult male and female wings were dissected, mounted separately, and the wing size was measured by calculating the area (total wing or per wing compartment) using the “outline” tool of the Zen 2.6 software. The number of wing cells were quantified by counting number of trichomes in predetermined square in different wing areas and calculate the average of trichomes and then extrapolate trichome number to the whole wing area as described in Towler et al. (2020).

Rapid iterative negative geotaxis (RING) method, which is based on the fly’s innate escape response so they climb up the walls of the tube after being tapped to its bottom, was used to assess locomotor defects in adult flies (Gargano et al., 2005). Young flies (1–2 days as adults) were separated by sex and genotype and put in groups of 10 in different tubes marked at 4 and 8 cm height. Negative geotaxis and recording of flies climbing (3 repeats) was then performed accounting for what percentage were able to climb and/or reach the marks at 5, 10, 15 and 30 s (n = 30 per genotype and sex).

At each trial of locomotion behaviour, third-instar larvae were washed to remove traces of food and transferred to a 240 × 240 mm² arena coated with a 2 mm thick layer of 0.8% agar, where they crawled freely for 5 min. After acclimatization, movies were recorded for 10 min at 25°C illumination with Infrared light following the Frustrated Total Internal Reflection (FTIR)-based imaging method (Risse et al., 2017). Movies were recorded with a Basler acA2040-180 km CMOS camera at 2048 × 2048 px² resolution and 2 frames per second (Sims et al., 2019), using Pylon and StreamPix software, mounted with a 16 mm KOWA IJM3sHC. SW VIS-NIR Lens and 825 nm high performance long pass filter (Schneider, IF-093). The momentum x:y coordinates for each larva was obtained with FIM track (Risse et al., 2017) and used to calculate the average velocity.

For survival assays standard vials containing cornmeal, sucrose and yeast medium were used. 10 recently emerged flies (males or females) were placed in each vial and kept in an incubator at 25°C and checked the percentage of flies alive every 2 days for 14 days. Survival assays were performed in triplicate per genotype and sex.

The mitotic index was calculated by the ratio of cells going through mitosis (H3P positive cells) in the posterior compartment (en-Gal4;UAS-dsRed positive cells) and the H3P positive cells in the anterior compartment (ds-Red negative cells) in control and experimental conditions (Towler et al., 2020).

RNA was extracted from wandering L3 larval carcases aged with a 3 h egg lay using a standard phenol-chloroform protocol. Larvae were homogenised in QIAzol lysis reagent (Qiagen), mixed with 0.2 volume chloroform and spun at 12,000 g for 15 min. The aqueous phase was precipitated in 1 volume 100% isopropanol and RNA was pelleted at 15,000 g for 30 min. The RNA pellet was washed twice in 80% ethanol before drying and resuspension in H₂O. DNase treatment was performed using the Turbo DNA-free kit (Invitrogen) following manufacturer’s instructions. RNA concentration and purity was assessed on a Nanodrop One.

1 µg of RNA for each sample was reverse transcribed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). A no RT control was performed in parallel. qRT-PCR was performed using the following primers: prto Fw: 5′-GCT​GTT​TGG​TTT​GTT​CGT​TGT​G-3′, prto Rev: 5′-ATG​TCC​GGA​ATC​TCG​TTC​CA-3′, rpl32 Fw: 5′-TAA​GCT​GTC​GCA​CAA​ATG​GC-3′, rpl32 Rev: 5′-TCG​ACA​ATC​TCC​TTG​CGC​TT-3′ which were optimised to give 99.01% and 92.6% efficiency for prto and rpl32 respectively. qRT-PCR was performed using the PowerTrack SYBR Green Master Mix (Applied Biosystems) in technical triplicate including a no template control. Cycling conditions were as follows: 95°C 5s followed by 58°C 30s cycled 40 times. Melt curve analysis was performed following manufacturer’s instructions. Data analysis was performed using the ∆∆ct method with rpl32 as a reference gene.

Clustal W (DNASTAR) and MAFFT (http://mafft.cbrc.jp/aligment/server) (Katoh et al., 2019) were used to align Prto peptide sequences (L-INS-I MAFFT alignment was used with scoring matrix BLOSUM62; gap opening penalty 1.53). For alignment confidence scores and extraction of well-aligned residues the Guidance 2 package was used (http://guidance.tau.ac.il) (Sela et al., 2015). The Phylo.io tools within MAFFT package was used for constructing Neighbor-Joining tree using WAG model, Alpha ∞, Bootstrap resampling = 100 (Robinson et al., 2016).

All statistical analyses were performed in GraphPad Prism 5 (GraphPad Software. Inc., La Jolla, CA). T-tests were used to compare the means of single test groups to single control groups. If multiple comparisons were required, a one-way ANOVA was performed with a post-test to compare the means of each pair of samples.

To investigate the cellular functions of translated smORFs, we focused on the shCDS class, as they are generally highly translated, accrue high conservation scores, and present peptide structures which are generally good indicators of functionality. The CG9034 gene was chosen for detailed characterisation from the shCDS’s pool because its function is unknown, and therefore we named it purriato (prto), which is an invented word with no meaning. The prto gene is located in the X chromosome and has 2 annotated transcripts differing in their 5′UTRs which encode a unique peptide (Gramates et al., 2022) (Supplementary Figure S1A). Prto peptide appears to be conserved in other insects, is highly expressed in specific tissues and produced abnormal phenotypes in RNAi genome-wide screens (Schmidt et al., 2012; Gramates et al., 2022).

The prto gene displays a dynamic temporal expression profile, with its mRNA expression levels varying from very high to moderately high during embryogenesis and throughout larval, pupal and adult stages (Brown et al., 2014). In addition, prto is significantly expressed in specific organs and tissues, such as in the adult carcass, imaginal discs, head or testes (Brown et al., 2014).

To better characterise the developmental and spatial profile of prto mRNA we performed in situ hybridisations in various organs and different stages. During early embryogenesis, prto mRNAs are maternally provided and distributed ubiquitously (stage 3; Supplementary Figure S1C). Slightly later in development, prto transcripts are restricted to the cephalic furrow (arrowhead) and the mesoderm when it begins to form (stage 6; Supplementary Figure S1D). Importantly, prto mesodermal expression is maintained through mesoderm development and differentiation in embryonic somatic and visceral muscles (Figures 1A, B; Supplementary Figures S1D–F) and larval muscles (Figure 1C). In addition, we have found that prto is highly expressed in the pouch of the wing imaginal discs (Figure 1D), which is supported by the modENCODE transcriptome data (Brown et al., 2014) and wing disc single cell transcriptomics (Li et al., 2022).

We also investigated prto expression profile at the translation level. Poly-Ribo-Seq data showed high levels of productive prto translation in S2 cells (Supplementary Figure S1B) (Aspden et al., 2014) and during embryonic stages (Patraquim et al., 2020). These translation profiles are in accordance with Prto peptide levels identified by mass spectrometry in embryogenesis (Casas-Vila et al., 2017). In addition, the prto translation has also been reported in larval muscles (Chen and Dickman, 2017). Thus, these temporal and spatial profiles at the transcriptional and translational levels provide the developmental framework to explore the putative roles of Prto peptides at the molecular and cellular levels.

The Drosophila prto gene encodes a 77 amino-acid peptide with a predicted secondary structure containing an alpha-helix transmembrane domain according to JNet and Phobius algorithms (Käll et al., 2004; Drozdetskiy et al., 2015) (Supplementary Figure S2A; Figure 1F). In addition, the confidence of the 3D model for Prto is very high supporting that a single transmembrane Prto structure is likely (Figure 1H) (Jumper et al., 2021).

Next, we assessed the evolutionary conservation of Prto by performing PSI-BLAST searches (Altschul et al., 1997; Johnson et al., 2008) and pairwise sequence comparisons to identify Prto homologues using MAFFT and Clustal W (Thompson et al., 1994; Sela et al., 2015; Katoh et al., 2019). Our analysis revealed that Prto peptides are conserved among Protostomia. Prto homologues are found in other arthropods, such as insects (Figure 1E; Supplementary Figure S2C). In addition, we identified Prto homologue peptides in other Ecdysozoans such as Nematoda, and Lophotrochozoans such as Annelida and Mollusca. Our searches did not identify Prto in Deuterostomia (Figure 1E; Supplementary Figures S2B, C). Importantly, the predicted Prto peptide secondary structure obtained using the Phobius program (Madeira et al., 2022) seems to be conserved throughout the Prto peptide family (Figures 1F, G). In addition, striking conservation in stretches of amino-acids at the carboxyl-terminus (Ct) were also observed outside the transmembrane domain (Supplementary Figures S2B, C). Thus, this level of amino-acid and structure conservation strongly suggests that transmembrane Prto peptides could have a role in membrane-related cellular processes.

To corroborate the predicted transmembrane structure of Prto peptides and further explore their cellular localisation we generated different N-terminus and C-terminus tagged Prto constructs and expressed them in Drosophila S2 cells. These differentially tagged Prto peptides showed strong co-localization to intracellular membrane compartments with a Pearson’s correlation coefficient (R) of 0.86 for Prto-Venus vs. FLAG-Prto and 0.87 for GFP-Prto vs. Prto-FLAG (Supplementary Figures S2D–D″, E). Next, to identify the membrane compartments to which Prto peptides localises, we repeated the staining using known cellular markers. This revealed that GFP-Prto peptides are highly localised in endosomes displaying a Pearson’s coefficient of 0.87 with the Hemotin peptide (Supplementary Figures S2F–F″, I) (Pueyo et al., 2016b). Reduced co-localisation of GFP-Prto peptides was found with mitochondria (Mitotracker; R = 5.5; Supplementary Figures S2G–G″, I) and lysosomes (Lysotracker; R = 5.7; Supplementary Figures S2H–H″, I). Thus, the deep conservation of Prto transmembrane peptides and their co-localization with endosomes, lysosomes and mitochondria suggest that function in these intracellular organelles.

Several genome-wide RNAi screens have reported that CG9034 knockdown (KD) produce generic defective phenotypes in specific fly organs (Schmidt et al., 2012). For instance, CG9034 KD using the Mef2 (muscle) and Pannier (thorax) Gal4 drivers led to loss of motility (Schnorrer et al., 2010), and shorter bristles respectively (Mummery-Widmer et al., 2009), but further characterisation have not been attempted.

With previous data suggesting a role of prto in muscles, we set out to perform a detailed characterisation of prto function. Specific knockdown in muscles was achieved through driving prto-RNAi (GD4003) with the Mef2Gal4 (Mef2>) driver at 29°C, which resulted in some lethality in late pupal stages, but escapers were able to produce adult offspring. Female and male prto KD adult flies were short lived (Figure 2A). Over time, surviving flies developed a held up wing posture (not shown) and display climbing defects in comparison with controls in negative geotaxis (Gargano et al., 2005) with males showing more severe phenotypes (Figure 2B). These results supported previous published observations (Schnorrer et al., 2010) and thus validating our functional approach. To understand the prto KD progressive deterioration in wing posture the thoracic indirect flight muscles (IFM) were phenotypically analysed. prto-RNAi KD myofibrils were slightly disorganised and displayed faded actin filaments in sarcomeres in comparison to controls (Supplementary Figures 3A, B). Although there were no major defects in the number of these myofibrils. Therefore, our results suggest that prto could be important in the regulation of physiological muscle processes.

To further characterise the prto function in muscles we crossed the prto-RNAi lines with lines expressing Dicer2 gene (Dcr) to stimulate RNAi silencing and enhance the knockdown (Dietzl et al., 2007). Expression of prto-RNAi using Mef2>UAS-Dcr driver at 29°C decreased prto mRNA levels 4 fold (Supplementary Figure S3C; Supplementary Table S1) and resulted in animals that reached larval stages but were developmentally delayed, smaller in size and died prematurely at early pupae stages. These mutants displayed a curved pupal case morphology with shorter spiracles, and failure of abdominal air bubble displacement (Figures 2C, D). This phenotype is characteristic in mutants of sarcomeric proteins, such as large Titin protein (Brooks et al., 2020), or Z-disk Trim 32 thin protein (LaBeau-DiMenna et al., 2012; Domsch et al., 2013) which are unable to properly contract muscles during the larval to pupal transition.

To determine the role of prto in muscle function we measured the pupal case axial ratios (length/width), which is a good indicator of muscle contraction during larval to pupal transition. The mean axial ratio value in Mef2>UAS-Dcr;GFP-RNAi controls and parental isogenic controls were <3 in accordance to previously published data (Brooks et al., 2020) (Figure 2E). In contrast, prto knockdown (Mef2>UAS-Dcr;prto-RNAi) resulted in a significantly higher mean axial ratio with a value >3.5 (Figure 2E). Similar phenotypes were produced by driving prto depletion with a different muscle driver (24B>) or by using with another prto-RNAi construct (#2) (Figure 2E). The muscle over-expression of either full-length wild-type prto (UAS-prto-FL) or Nt-tagged GFP-prto (UAS-GFP-prto) constructs together with UAS-Dcr;prto-RNAi partially rescued the pupal axial ratio towards wild-type (Figure 2E) and the lethality pupal phase (Supplementary Figure S3E) even considering that these constructs are still targeted by RNAi. In contrast, co-expression of UAS-GFP in muscles to control for the Gal4 dilution failed to rescue the UAS-Dcr;prto-RNAi axial ratio (Figure 2E). qRT-PCR experiments show that prto mRNA expression levels were restored to wild-type levels in Mef2>UAS-prto-FL;UAS-Dcr larvae carcasses (Supplementary Figure S3D; Supplementary Table S1) and larval somatic muscles seemed normal (Supplementary Figure S3H). Finally, when over-expression of UAS-prto-FL (increase of expression 16 fold change, Supplementary Figure S3D) or UAS-GFP-prto was driven with Mef2-Gal4 driver in a wild-type background the larvae displayed normal larval somatic muscles (Supplementary Figures S3I, J–J′), the pupae showed wild-type axial ratios (Figure 2E) and viable offspring with normal muscle function was produced (not shown). Altogether these results indicate that the observed phenotype is specific to the reduction in prto function.

Next, we assessed the functional prto requirement in muscle contraction by monitoring L3 larval locomotion in an agar arena for 10 min (Berni, 2015). The exploratory behaviour of either parental Mef2>UAS-Dcr or prto-RNAi control larvae showed that they can fully move across the agar plate (Figure 2F) at an average velocity of 2 cm/min (Figure 2H) whereas Mef2>UAS-Dcr;prto-RNAi larvae velocity was greatly reduced producing a limited exploration of the arena (Figures 2G, H).

Taken together these findings demonstrate a strong correlation between prto levels and muscle function. Strong reduction of prto expression provokes severe L3 larvae locomotion defects and muscle contraction impairments at pupariation (axial ratios) whilst a milder reduction affects adult motor response and motility. Thus, the prto gene has an essential role in the regulation of muscle function and required further investigation.

To investigate the role of prto in larval musculature, we utilised a GFP-trapped Z-disc protein (G203-GFP, Crisp et al. (2008)) to monitor sarcomeric structures in control and Mef2>UAS-Dcr,prto-RNAi knockdown genetic backgrounds. Whole-mount larvae preparations of late L3 (100–120h AEL) showed a stereotypical array of 30 somatic muscles per hemisegment in controls (Figure 1B inset; Figures 3A–C (Bate, 1990)) whereas more than 50% of somatic muscles exhibited drastic morphological defects in prto KDs (Figures 3C–E). These defects were primarily classified as thinned muscles (38%) and detached muscles (14%) (Figure 3C). Interestingly, the somatic musculature of L2 prto KD larvae (65-70hr AEL) seemed normal and similar to the controls indicating that prto KD defects arise when the somatic muscles are rapidly growing and undergoing enormous mechanical stress (Figure 3C). Thus, Prto function seems not to be necessary for the formation of somatic musculature but instead is required for the maintenance of muscle integrity.

Next, we assessed the integrity of sarcomeric structures by monitoring the organisation of thin filaments (F-actin), and Z-disks (G203-GFP) in L3 larval fillets. Control larval somatic muscles showed regular repeated patterns of thin filament and Z discs along the myofibrils (Figures 3F–F‴; Supplementary Figures S3F–F‴). However, in UAS-Dcr;prto-RNAi KD muscles, myofibrils appear to be over-stretched and sometimes detached (Supplementary Figures S3G–G‴). In addition, the sarcomeric architecture is often disrupted by clumps of actin filaments (arrow; Figures 3H–H‴) and Z-disk proteins (arrowheads; Figures 3G–G‴). The prto KD muscles contained similar numbers of nuclei in comparison to controls (Supplementary Figures S4A–D) but they were disorganized and misshaped in comparison with control (Figures 3F″–FF‴, G″–G‴; Supplementary Figure S4). These prto KD muscle phenotypes are reminiscent to that of mutations of genes involved in sarcomerogenesis, such as cofilin (Balakrishnan et al., 2020), or filamin (cherio (cher) in Drosophila) (González-Morales et al., 2017; Brooks et al., 2020), which are necessary for the extensive larval muscle growth and the maintenance of muscle integrity respectively.

To further explore the role of prto in sarcomeric homeostasis, we expressed a GFP-tagged prto construct in larval muscles, as it is partially able to fulfil the prto function in vivo (Figure 2E). Prto peptides showed a punctate and granulated pattern running along sarcomeres (arrowheads; Figures 4A–A″) and in perinuclear regions (arrow; Figures 4A–A″). Specifically, the Prto striated pattern was aligned to the Z-disks as it colocalise with Sallimus (Sls) proteins, which are large proteins that bind to myosin thick filaments within the Z-disks. Sls contribute to sarcomere extension properties of the muscle and are essential to maintain sarcomeric homeostasis and therefore may implicate Prto in muscle maintenance (Poovathumkadavil and Jagla, 2020) (arrowheads; Figures 4B–B″).

Next, we utilize different markers to precisely map Prto-positive cellular compartments. Prto peptides showed some degree of co-localisation with the endosomal FYVE-Cherry and lysosomal mRFP-spinster markers with the former enriched in the perinuclear regions (arrowheads; Figures 4C–C″) and the later in the sarcolemma (arrow; Supplementary Figures S6A–A″). This subcellular localisation is consistent with our earlier observations in embryonic S2 cells (Supplementary Figure S2F), suggesting a possible general prto role in these organelles. Importantly, these Prto-positive organelles are essential for maintaining muscle homeostasis and integrity as they are involved in cellular proteostasis of damaged sarcomeric proteins and tissue remodelling (Johnson et al., 2015; Fujita et al., 2017).

To understand the mechanisms by which Prto membrane peptides regulate the maintenance of sarcomeric structures we explored important cellular processes such as endosome and lysosome distribution. Firstly, in control muscles, the endosomes (FYVE-Cherry) and lysosomes (LAMP1-GFP) are distributed across the sarcolemma and enriched in the perinuclear region, overlapping abundantly in endo-lysosomal compartments (arrows; Figures 4D–D″). In UAS-Dcr;prto-RNAi KD muscles, these compartments are enlarged and accumulated around the aberrant nuclei, even when muscles are not showing major structural defects (arrows; Figures 4E–E″). Thus, it seems that disrupting prto function negatively affects endosome and lysosome functions and therefore compromises sarcomeric integrity leading to loss of muscle function.

The Chaperone-assisted selective autophagy (CASA) pathway is a mechanism involved in the detection of misfolded sarcomeric proteins caused by contractile forces and their subsequent repair or their degradation via autophagocytosis (Arndt et al., 2010). Members of this pathway (Figure 5A) include: 1) the BAG3 (Starvin (Stv) in Drosophila)/Hsc70-4 Chaperone complexes that recognize defective proteins, such as Filamin which is named Cheerio (Cher) in Drosophila; 2) the NUAK kinases that phosphorylate these misfolded substrates for subsequent ubiquitination; 3) The p62/SQSTM1 complex that later binds ubiquitin chains of targeted proteins and triggers autophagosome formation and subsequent lysosomal degradation. Importantly, the functions of the members of the CASA pathway are conserved in metazoans (Arndt et al., 2010), and specific mutations in human homologues have been identified in Charcot-Marie Tooth type 2L, myopathy, cardiomyopathy, neuropathy and their over-expression is associated with poor prognosis in various cancers and also chemoresistance (Leber et al., 2016; Matsushima-Nishiwaki et al., 2017; Shen et al., 2018; Verdonschot et al., 2020; Cristofani et al., 2021; Kirk et al., 2021; Yerabandi et al., 2022). Notably, mutations in members of this pathway in Drosophila provoke muscle defects reminiscent of human myopathies. These include progressive accumulation of aberrant sarcomeric proteins, autophagosomal and lysosomal dysfunction leading to a progressive loss of muscle sarcomeric structures and contractile function (Arndt et al., 2010; González-Morales et al., 2017; Ader et al., 2022).

Since the prto-RNAi KD muscle phenotype (Figures 2, 3) shows a striking resemblance to those observed in CASA mutants we investigated possible genetic interactions between prto and members of CASA pathway. Our phenotypic characterisation showed that similar prto-like locomotive defects in adult flies and early death were observed in Mef2>NUAK-RNAi and in a lesser extent in Mef2>cher-RNAi KD males (Supplementary Figures S5A, B). Second, analysis of the Indirect Flight Muscle structure showed that NUAK-RNAi KD myofibrils were disorganized, and F-actin filaments were slightly faded mimicking the prto-RNAi phenotype (Supplementary Figures S5C–E). Third, strong reduction of gene functions of CASA members in muscles either by co-expressing Dicer or increasing temperature led to prto-like pupal abnormalities such as reduced/lack of anterior spiracles, air bubble retention and longer pupal case axial ratios (Figures 5B, C; Supplementary Table S2). These phenotypes correlated with the progressive deterioration of muscle sarcomeric structure (Figure 5E; Supplementary Figures S6D–D″) and accumulation of endo-lysosomal aggregates in the sarcolemma as observed in prto-RNAi;UAS-Dcr KD muscles (Supplementary Figures S6B-B″, C–C′; compare Figures 4E–E″). Based on these phenotypical similarities we wondered whether the increase of autophagic particles in muscles induced by CASA member mutations (Brooks et al., 2020) was also observed in prto-RNAi;UAS-Dcr KD muscles. Expression of a double fluorescent tagged Atg8a (autophagosome marker) in the muscle showed that prto KD muscles exhibit a remarkable increase in the number Atg8a positive puncta in comparison to controls (Figures 5F–F‴, G–G‴). Thus, these phenocopies at the cellular, physiological and morphological levels strongly suggest that prto function is closely associated to CASA-mediated protein turnover.

To further explore functional relationship between prto and the CASA pathway in sarcomeric homeostasis, we first tested the interaction with NUAK by monitoring the GFP-Prto peptide distribution in NUAK-RNAi KD muscles. Large Prto-positive aggregates were observed in actin-depleted areas in the sarcolemma in NUAK-RNAi KD mutants as previously reported with other Z-enriched sarcomeric proteins, such as Filamin and dCRYAB (Figures 5D–D′; Supplementary Figures S6D′–D″) (Brooks et al., 2020). We subsequently sought to assess the epistatic relationships among NUAK and prto. Decreasing NUAK function with Mef2>NUAK-RNAi;UAS-Dcr resulted in pupae that displayed the highest axial ratios and stronger L3 larval muscle abnormalities of the CASA members (Figures 5B, E–III). In addition, the over-expression of full length UAS-prto in NUAK-RNAi KD background, was unable to recover this phenotype and even enhanced the pupal case phenotype significantly (Figure 5B). A phenotype that also correlates with the formation of large Prto-positive aggregates as revealed by Prto-GFP in NUAK KD muscles (Figures 5D–D″). This suggest that NUAK and prto do not perform equivalent roles in muscles. Finally, over-expression of NUAK (UAS-NUAK) in muscles in a wild-type background results in pupal lethality, with pupae displaying slightly shorter axial ratios compare to controls (Figure 5B). Notably, either reduction or over-expression of prto gene function enhanced the UAS-NUAK phenotype (Figure 5B), with larvae that die in second or third instars before puparation. The over-expression of prto-GFP was less penetrant and did not change the NUAK gain of function phenotype. Thus, these results suggest that prto function could be acting downstream or in parallel to NUAK to modulate its function.

To address this hypothesis, we sought for possible genetic interactions between prto and the downstream members of the CASA pathway. Stv/BAG3 is able to bind NUAK, Cher, and the chaperone Hsc70-4 to form a complex that allows the ubiquitination of misfolded Cher proteins triggering their removal by autophagy (Arndt et al., 2010; Brooks et al., 2020) (Figure 5A). By using a specific RNAi construct and temperature conditions we sought a sensitive genetic background to underpin possible interactions. Decreasing stv (stv-RNAi) or prto (prto-RNAi) in muscles with Mef2>UAS-Dcr at 25°C produce offspring that reached adulthood and their pupal axial ratio were only slightly higher than controls (Figure 5B). Simultaneous knockdown of stv and prto resulted in pupal lethality and displayed a synergistic effect when we assessed the pupal axial ratio compared to each single gene knockdown including the strongest UAS-Dcr; prto-RNAi KD (29°C) pupae (Figure 5B), suggesting that both genes cooperate. This enhanced phenotype correlated with an increase in sarcomeric disruptions observed in the L3 muscles (Figures 5E–VIII compare with Figures 5E–II). Furthermore, removal of a single stv gene copy (stv ¹ /+) in a prto-RNAi KD background was also sufficient to further increase the prto KD pupal ratio (Figure 5B). Finally, the over-expression of stv (UAS-stv-V5) in muscles in a wild-type background produced short larvae (Tubby-like) with muscles shorter in length (Figures 5E–IX). The Mef2>UAS-stv offspring die at puparation and display reduced axial ratios compared to wild-type (Figure 5B). Importantly, Mef2>UAS-stv over-expression in prto-RNAi KD background was larval lethal at L2-L3 stage (Figure 5B). Altogether these results indicate that there is a strong genetic interaction between ptro and stv in maintaining sarcomeric homeostasis.

We next observed a genetic interaction between prto and the chaperone hsc70-4. The pupal axial ratio and sarcomeric larval muscle abnormalities of prto-RNAi and hsc70-4-RNAi double knockdowns were higher than each single gene knockdown (Figure 5B, compare E-VI with E-II and E-V), further supporting the role of Prto as a regulator of the chaperone complex required for protein assembly and turnover in muscles. Importantly, over-expression of UAS-hsc70-4 was able to partially rescue the UAS-Dcr;prto-RNAi axial ratio phenotype at 29°C, suggesting that more efficient reassembling of unfolded sarcomeric proteins overcomes prto KD muscle abnormalities (Figures 5B, E–VII).

Finally, atg8a function acts at the base of the CASA pathway permitting the anchoring of the p62/SQSTM1, ubiquitinated cargo and chaperone complex and formation of autophagosome (Brooks et al., 2020). Reduction of atg8a function increases the pupal axial ratios and muscle sarcomeric abnormalities (Figures 5B, C, E–IV). Importantly, over-expression of UAS-prto-FL cannot rescue the UAS-Dcr;atg8a-RNAi knockdown phenotype (Figure 5B) suggesting that atg8a function is the limiting factor and that prto function could be acting on upstream members of CASA pathway. Altogether, the genetic interactions place Prto between NUAK and Atg8a functions and in close cooperation with the Stv/Hsc70-4 chaperone complex.

To ascertain how Prto peptides interact with CASA-protein turn over machinery, we monitored cellular localisation of GFP-Prto with Stv-V5 in muscles. Co-localisations of both factors were observed in punctae near the Z-disk region where active remodelling and turnover of Cher takes place (Arndt et al., 2010; Brooks et al., 2020) (arrowheads; Figures 5H–H′; Supplementary Figures S6E–E″). Supporting this view, FLAG-Prto peptides colocalised with the GFP-tagged short Cher isoform (Cher90) in the Z-disks (arrowheads; Supplementary Figures S6F–F″). Interestingly, a search of functional motifs (in diverse members of Prto peptide family) using ELM (Kumar et al., 2020) found an Atg8/LC3 binding domain in the Ct region of Prto peptides supporting their association to autophagosomes (Supplementary Figure S2C). Therefore, our genetic and cellular phenotypic characterisations indicate that Prto transmembrane peptides are involved in maintaining sarcomeric homeostasis, possibly by bridging the Stv/Hsc70-4 chaperone complex turnover machinery with membrane organelles such autophagosomes, and endo-lysosomes for degradation of misfolded Cher proteins. Further work would be required to identify if Prto directly interacts with any members of the CASA pathway.

Refolding and degradation of damaged cargo proteins by the CASA pathway is important to maintain the cellular homeostasis in a diversity of developmental contexts, including muscles, neurons and kidneys (Höhfeld et al., 2021). Therefore, we wondered whether the regulation of CASA pathway by Prto peptides is a tissue-specific function or a broader one extending to other organs.

We explored the role of prto and the CASA pathway in wing development since prto is highly expressed in wing imaginal discs and there are no insights into CASA-mediated functions in this proliferative epithelial organ. First, we expressed a GFP-tagged prto construct with the Dlmo-Gal4 driver in the wing discs (Figures 6A, B–B′) and found that GFP-Prto peptides were localised in the plasma membrane (arrows) and in intracellular punctae in epithelial cells (arrowheads). Some of these Prto-positive intracellular compartments were labelled with the FYVE endosomal marker (arrows; Figures 6C–C‴), suggesting that Prto subcellular distribution across different cell types seems to be conserved.

To investigate the role of prto in wing development we knocked down prto expression in different regions of the wing imaginal disc by expressing the prto-RNAi construct with different Gal4 drivers (69B, Dlmo, engrailed (en)) at different temperature conditions. A specific and significant reduction of the size of the wing was observed in the induced prto-RNAi region (Figures 6D, E; Supplementary Figure S7). For example, prto-RNAi knockdown in the posterior compartment causes shortening of the posterior region of the wing without affecting the anterior part (Figures 6D, E). By counting the number of wing cells (trichomes) in the anterior and the posterior regions of the adult wings we determined a significant ∼30% reduction in number of cells in en>UAS-Drc;prto-RNAi wings in comparison to en>UAS-Dcr controls (Supplementary Figures S8A, B). Similarly, reduction of prto function with Dlmo-Gal4, a driver strongly expressed in the dorsal compartment, produces a curled wing phenotype (not shown) due to reduction of the dorsal epithelium (Supplementary Figure S7).

To ascertain which cellular processes lead to a reduction of wing epithelial cells in prto-RNAi KDs we first investigated the effect of prto function on apoptosis and proliferation in the developing wing imaginal discs. Reduction of prto using en> driver produced posterior compartments of an abnormal shape (misfolded epithelium) with over-stretched epithelial cells showing larger gaps among neighbouring cells. Despite this phenotype, no major increase on apoptosis was observed in prto-RNAi KD wing discs (Supplementary Figures S8C, D–D‴, F–F‴). By contrast, a significant decrease in mitosis was observed by comparing the number of phosphorylated Histone 3 (H3P) positive cells in the anterior (control) and posterior (KD) wing compartments (Figures 6F–F′, G–G′). The Mitotic index between Anterior/Posterior compartment was ∼30% higher in prto-RNAi KD than in controls (Figure 6H), suggesting a reduction in the number of mitotic cells following prto depletion which correlates with the reduction in cell number observed in prto-KD adult wing compartment (Supplementary Figures S8A, B).

To further explore the role of prto in growth we monitored the cell cycle stages of wing epithelial cells using the FUCCI system. The FUCCI system compares expression of two different cell cycle stage protein domains tagged to different fluorescent proteins simultaneously (E2F1-GFP and CycB-RFP) allowing the identification of cell cycle phase by fluorescence (Zielke et al., 2014). Most of wing epithelial cells divide in a non-synchronic fashion apart from those in the wing margin (Barrio and Milán, 2020). In control wing imaginal discs (en>UAS-Dcr), FUCCI expression in the posterior compartment revealed a mosaic of cells expressing specific fluorescent cell-cycle markers showing cells at different stages of the cell cycle (Figures 7A–A‴). In contrast, the UAS-Dcr;prto-RNAi KD wing discs showed a qualitative enrichment of cells co-expressing both cell cycle markers (yellow) indicating that cells were stalled in G2 phase (Figures 7B–B‴). We next performed relative quantifications of numbers of cells expressing each cell cycle markers by the automatic image processing DeadEasy Mito-Glia method (Forero et al., 2010). A similar and steady increase in the population of cells expressing each of the cell cycle markers was observed in en>UAS-Dcr,prto-RNAi in comparison to controls (Supplementary Figure S8E). Since prto-RNAi KD wing regions show lower number of cells going through mitosis, this relatively equal increase in both of these cell cycle populations in prto-RNAi KDs suggests that they are arrested at G2/M transition. Altogether, these findings show that prto function is important for proper cell proliferation in the wing imaginal epithelium.

Next, we monitored the distribution of endosomes and lysosomes (Prto-positive intracellular compartments) in the wing disc epithelium by using FYVE and LAMP-1 markers respectively. Both the FYVE and LAMP-1 markers are distributed in cytoplasmic punta in en>UAS-Dcr control wing imaginal discs (Figures 7C–C‴). However, in en>UAS-Dcr; prto-RNAi KD wing discs, larger aggregates of FYVE and LAMP-1 were observed throughout the cell (Figures 7D–D‴). Interestingly, an increase of Atg8a-positive autophagosomes was also observed in prto-RNAi KD wing discs in comparison to controls (Figures 7E–E′, F–F′). Thus, these results suggest that prto depletion disrupts endo-lysosomal homeostasis and provokes an accumulation of autophagosomes leading to a decrease of cell proliferation in wing disc epithelium.

Since prto knockdowns alter endo-lysosome and autophagosome pathways in both somatic muscles and wing imaginal discs, we investigated whether Prto-mediated regulation of the CASA pathway is required for proper wing growth. Knocking down either NUAK or dCRYAB by RNAi in the posterior compartment using the en-Gal4 driver provoked a reduction in the posterior wing area in comparison to the control anterior wing area (Figures 8A–C, G). In addition, the stv-RNAi knockdown using en-Gal4 driver resulted in extensive male pupal lethality, although some adult escapers eclosed displaying wings with the posterior compartment reduced (Figures 8D, G). Thus, these wing size reductions in KD of CASA members phenocopy those of prto-RNAi KD flies.

Finally, we tested whether genetic interactions between prto and CASA members exist in the wing. First, en>stv-RNAi knocking down produced females that reached adulthood displaying reduced wings (Figures 8E, H). Reduction of prto function in stv-RNAi KD females enhanced all wing phenotypes resulting from the independent knockdown of each genes (Figures 8F, H). Similarly, removing a single functional stv gene copy (stv ¹ /+) in a UAS-Dcr;prto-RNAi background further reduced the posterior compartment area with minimal anterior compartment defects (Supplementary Figures S8G–J). Therefore, the observed genetic interactions between prto and genes in the CASA pathway resemble those found in muscles and suggest a putative role of Prto peptides and the CASA pathway in controlling protein turnover to maintain correct growth of the wing imaginal disc.