Drosophila and human Headcase define a new family of ribonucleotide granule proteins required for stress response

Abstract

Cells have means to adapt to environmental stresses such as temperature fluctuations, toxins, or nutrient availability. Stress responses, being dynamic, extend beyond transcriptional control and encompass post-transcriptional mechanisms allowing for rapid changes in protein synthesis. Previous research has established headcase as a fundamental gene for stress responses and survival of the Drosophila adult progenitor cells (APCs). However, the molecular role of Headcase has remained elusive. Here, we identify Headcase as a component of ribonucleoprotein (RNP) granules. We also show that, Headcase is required for proper RNP granule formation and remodeling upon stress and is crucial for translation control. Likewise, the human Headcase homolog (HECA) is identified as a component of RNP granules and has similar roles in translational regulation and stress protection. Thus, Headcase proteins define a new family contributing to specific roles among the RNP heterogeneous network.

Heca proteins, conserved across species, are essential for stress adaptation and regulate RNP granules and protein synthesis.

INTRODUCTION

Throughout development and into adulthood, organisms face numerous environmental stimuli and require robust mechanisms to adapt and maintain homeostasis. Cellular mechanisms that balance these demands allow for adaptation to environmental stresses such as temperature fluctuations, toxins, or nutrient availability. These mechanisms involve stress response gene activation, cellular metabolism modulation, and alterations in protein synthesis. The capacity for a smooth and efficient response to stress is critical for organismal health, as persistent or inappropriate responses can result in disease and developmental anomalies. Stress responses, being dynamic, extend beyond transcriptional control and encompass post-transcriptional mechanisms such as RNA splicing, editing, transport, and degradation. These processes allow for rapid changes in protein synthesis, aligning it with developmental cues and environmental challenges.

Our previous research and that of others has established that the Drosophila gene headcase (previously shortened as hdc and presently renamed as heca by Flybase), is pivotal for managing stress responses and survival (1–4). heca is expressed in the adult progenitor cells (APCs) (5), the diploid cells specified during the embryonic stage that will later give rise to the adult structures. heca not only regulates systemic ecdysone levels (3) but in the APC clusters, it also modulates tissue growth by affecting the TOR pathway (2, 3), predominantly via the dRpS6 branch rather than the 4E-BP branch (3). Without heca, APCs are vulnerable to apoptosis, possibly due to an overactive unfolded protein response. Similarly, the HECA human homolog plays a conserved role in growth control by influencing the TORC1 pathway (2).

Despite the importance of Heca proteins in development and stress response, their molecular roles in cellular context remain underexplored. In particular, no functional domains have been identified in Heca proteins that might account for a molecular function. Here, our study sheds light on these roles, revealing a network of interactions for Heca. First, we show that Drosophila Heca is a component of ribonucleoprotein (RNP) granules and analyze its interaction with other components of well-known RNP complexes such as the P-bodies (PBs) or stress granules (SGs). In particular, we find that, upon stress, Drosophila Heca not only accumulates at SGs but it is also required for their proper formation. We also demonstrate that Drosophila Heca is crucial for translational control. In addition, we show that human HECA, mirroring Drosophila homolog’s function, also modulates translation and stress protection, engaging with molecular partners integral to the RNA binding protein (RBP) network that form phase-separated granules. Given the association of Heca proteins with RNP granules and their necessity for granule formation during stress, we hypothesized that they might also function as RBPs, which we subsequently confirmed. Thus, this work identifies a new family of RNP granule interacting proteins across species that regulate translation and confer stress protection. At the same time, this work underscores the heterogeneity among PBs and SGs structures, both regarding their components and their subsequent roles in cell physiology.

RESULTS

Drosophila Heca interacts with RBPs that form RNP granules, under normal conditions and after stress induction

Analysis of Heca distribution in the APCs by means of a validated Heca antibody (5, 6) (see also Materials and Methods and fig. S1) shows a diffuse cytoplasmic localization (Fig. 1, A and B). Intrigued by the potential of a distinctive subcellular distribution pattern, we used the APCs of Tr2 tracheal region, which is formed of large diploid progenitors, for detailed imaging. Using super-resolution microscopy, we observed that Heca accumulation manifests a punctate pattern within the cytoplasm (Fig. 1, C and D). Given the known association of Heca with cellular stress responses (3), we investigated whether the protein’s distribution or localization pattern is altered in response to stress stimuli. Following the ex vivo application of oxidative stress, using arsenite as an oxidative agent, a notable transformation in Heca distribution was observed (Fig. 1, E to H, S, and T), with the formation of aggregate-like structures reminiscent of RNP granule formation, which was not associated to significant changes in the protein levels (Fig. 1Q). In addition, to evaluate its potential to participate in phase separation, we used the catGRANULE algorithm (7), which predicted that both isoforms of Heca (8) can form granules with scores of 1.27 and 1.35 above SD (for the Heca long and short isoforms, respectively). We did not observe significant changes in Heca aggregate sizes in tissues of heat-shocked or starved L3 larvae (Fig. 1, I to R), suggesting that the Heca response is specific to oxidative stress. This behavior is common among proteins, which exhibit stress-specific localization to RNP granules (9–12).

Fig. 1. Heca accumulation and localization pattern under normal and stress conditions.

(A to P) Cells from L3 wing discs (A to H) and Tr2 tracheal cells (I to P) labeled with anti-dHeca antibody (gray) and DAPI (blue) under normal conditions (A, B, I, and J) after arsenite treatment (C, D, K, and L), heat stress (E, F, M, and N) and starvation (G, H, O, and P). Heca shows a granule-like cytoplasmic organization most prominently after oxidative stress induction. (Q) Quantification of Heca staining intensity under normal conditions (blue) and after arsenite treatment (magenta). (R) Comparison of Heca puncta volumes between normal conditions (blue) and after stress induction (arsenite stress in magenta, heat stress in black, and starvation stress in gray) in L3 Tr2 tracheal cells. Kruskal Wallis H, P < 0.0001 (n = 10 individuals per condition, n_{normal conditions} = 722 and n_stress = 1001). (S and T) Imaris 3D projection of cytoplasmic Heca aggregates in L3 Tr2 tracheal cells under normal conditions (S) and arsenite stress (T). (U) Frequency distribution of nearest-neighbor distances (NNDs) for Heca puncta under normal (blue) and arsenite stress conditions (pink), showing a shift toward closer proximity under stress. Kolmogorov-Smirnov D = 0.1731, P < 0.0001 (n = 10 individuals per condition, n_{normal conditions} = 664 and n_stress = 706).

As a next step, we aimed at determining whether Heca colocalizes with established components of RNP granules under both standard and stress conditions. We first examined colocalization under standard conditions of Heca with PBs, which are prevalent during normal cellular conditions, against components such as Trailer Hitch (Tral) [Like-Sm protein 14 (LSM14) in humans], Decapping protein 1 (Dcp1), and Maternal expression at 31B (Me31B) [DEAD-box helicase 6 (DDX6) in humans] (Fig. 2, A to F) (13–17). Using super-resolution microscopy and performing our analysis through the DiAna tool (18) in Fiji software (19), we observed varied degrees of spatial distances (defined here as the center to center distance amongst the segmented particles from their closest neighbors) between Heca and the PB components. The mean calculated distances between Heca and Dcp1, Tral, and Me31B reflect the proximity between the proteins, while the mean distance for the Heca-Dcp1, which is found close to the diffraction limit is indicative of a closer interaction between this pair (Fig. 2M). The variability in colocalization values between Heca and different granule components could suggest distinct biological interactions or functionalities associated with each component.

Fig. 2. Heca spatial localization in comparison to granule components under normal and stress conditions.

(A to C and G to I) Tr2 cells of control L3 Oregon-R larvae. Projections of super-resolution image acquisitions showing colocalization of Heca with Dcp1 (A and D), Tral (B and E), and Me31B (C and F) under normal conditions (A to C) and stress conditions (G to I). Heca is in magenta, granule components are in green, and DAPI stains nuclei in blue. (D to F and J to L) The plots represent fluorescence intensity profiles along the yellow lines drawn in (A) to (C) and (G) to (I), for each of the fluorophores. (M) Violin plots of center to center distances calculated with the DiAna Fiji plugin between Heca and Dcp1, Tral, and Me31B under normal and stress conditions, indicating increased proximity under stress. Kruskal Wallis H, P < 0.0001, followed by Dunn’s multiple-comparisons tests. Exact P values for each comparison performed are shown on the graphs. Dotted lines in the violin plots represent the median as well as the upper and lower quartile (n = 6 individuals per condition; normal conditions: n_Dcp1 = 412, n_Tral = 179, and n_Me31B = 168; stress conditions: n_Dcp1 = 49, n_Tral = 60, and n_Me31B = 199). Scale bars, 2 μm. (N to U) Colocalization of Heca with the SG markers Caprin (N and R) and Rasputin (O and S) under normal condition (N and O) and arsenite stress (R and S). Heca is in magenta, Rasputin and Caprin in green, and DAPI stains nuclei in blue. Scale bars, 2 μm. (P, Q, T, U) The plots represent fluorescence intensity profiles along the yellow lines in (N), (O), (R), and (S).

Upon exposure to stress, the proximity of Heca with all PB markers increased significantly. In this case, all calculated mean distance values for Heca-Dcp1, Heca-Tral, and Heca-Me31B pairs were found closer to the diffraction limit, suggesting co-aggregation of the protein and the granule structures (Fig. 2, G to L and M). Heca’s presence within SG formed upon arsenite treatment was further substantiated through costaining with the SG marker and core component Rasputin [Ras GTPase-activating protein binding-protein 1 (G3BP1) in humans] (Fig. 2S) and Caprin (Capr) (Fig. 2R), another protein known to accumulate in SG during stress. Typically dispersed in the cytoplasm under normal conditions, these proteins transition to forming large condensates poststress induction together with PB components (20–24). Note that upon stress induction, Heca-positive puncta appear to colocalize with Rasputin, but not all Rasputin-positive aggregates contain Heca, indicating the heterogeneity in SG composition and their possible biological diversity.

To study further the interactions between Heca and RNP granule proteins, we performed in situ proximity ligation assays (PLAs) (see Materials and Methods). These assays targeted Heca with Dcp1 and Rasputin (Rin) under both normal and stress conditions. During normal cellular states, we confirmed the interaction of Heca with Dcp1 by an increased number of PLA spots per area compared to the negative controls (Fig. 3, A, B, and G). For the Heca-Rasputin pair, the interaction detected under normal conditions (Fig. 3, H to M) was more pronounced following stress, as evidenced by a greater number of PLA spots per area relative to both the corresponding controls and the nonstressed conditions (Fig. 3N). Notably, as indicated by the PLA experiments, even under normal conditions, Heca appeared to interact with Rasputin, highlighting the potential functional interplay between these proteins in cellular homeostasis before the initiation of the stress response. Rasputin is known to be a key player in the formation and regulation of SGs, and its role in RNA metabolism under stress is well-documented (25). Under nonstress conditions, Rasputin is distributed throughout the cytoplasm where it is thought to partake in the regulation of mRNA turnover and translation, possibly through its interaction with the RNA decay machinery (21, 26). Upon stress induction, Rasputin quickly relocates to SGs, where it contributes to the sequestration and triage of mRNA, a process that is vital for cell survival during stress. The observed interactions between Heca and Rasputin suggest that Heca may be involved in these processes as well, assisting in the stabilization of certain mRNAs or facilitating their decay when no longer needed.

Fig. 3. PLAs for Heca interactions with Dcp1 and Rasputin in Tr2 cells.

(A to F) Interaction between Heca and Dcp1 in Tr2 cells of control L3 Oregon-R larvae, under normal conditions (A), in heca^−/− used as negative control (B), and after arsenite treatment (C), shown by DAPI stain in gray and PLA spots in green (D to F). (G) Quantification of PLA spots per area showing significant interaction between Heca and Dcp1 under normal conditions and after arsenite treatment compared to the negative control. Brown Forsythe ANOVA, P < 0.0001, followed by Dunnett’s T3 post hoc tests. Exact P values for the comparisons performed are noted on the graphs (n = 4 individuals per condition, two ROIs analyzed per individual. n = 8 ROIs per condition). (H to M) Interaction between Heca and Rasputin in Tr2 cells, under normal conditions (H), in heca ^−/− used as negative control (I), and after arsenite treatment (J). PLA spots are shown in green in (K) to (M). Scale bars, 10 μm. (N) Quantification of PLA spots per area showing interaction between Heca and Rasputin under normal conditions, which is significantly increased after arsenite treatment. Brown Forsythe ANOVA, P < 0.0001, followed by Dunnett’s T3 post hoc tests. Exact P values for the comparisons performed are noted on the graphs (n = 4 individuals per condition, two ROIs analyzed per individual. n = 8 ROIs per condition). Error bars represent SD of the mean.

Heca modifies RNP granule formation, stress responses, and protein synthesis

Upon verifying the interaction between Heca protein and RNP granule components, we explored whether Heca influences the formation of aggregates, shedding light on its functional role in the regulation of stress response elements. We initially attempted to knock down heca expression using RNA interference (RNAi) in tracheal cells with the btlGal4 driver. The aim was to assess changes in the pattern of aggregate formation for PB components. Nonetheless, control RNAi experiments targeting green fluorescent protein (GFP), unexpectedly revealed altered aggregate formation (see fig. S2), suggesting a possible effect of the RNAi induction in the organization of the granules (27).

Consequently, we used a heca mutant background to examine potential effects on granule components in L3 larvae tissues (see Materials and Methods for the exact genotypes). In Tr2 cells of heca mutants, we observed an impaired formation of PBs, by a reduction of the partition coefficient for the Dcp1 protein (defined as the ratio of the intensity of the protein detected in the granule over the average surrounding cytoplasmic intensity), and by an altered subcellular distribution including increased nuclear localization (Fig. 4, B and J). Given that a higher portion of the overall Dcp1 protein accumulated in the nucleus, as indicated by an increased nuclear to cytoplasmic ratio (Fig. 4, B and K), could imply a potential buffering action for its cytoplasmic activity of RNA decapping (28). The effect of Heca on the stability and constitution of PBs was also confirmed by altered partition coefficients for other PB members like Tral and Me31B (Fig. 4, E and L) and Me31B (Fig. 4, H and M).

Fig. 4. Effects of Heca on granule component distribution and protein synthesis.

(A to I) Localization of Dcp1 (A to C), Tral (D to F), and Me31B (G to I) in control (A, D, and G), heca^−/− (B, E, and H), and btl > UAS-heca (C, F, and I) Tr2 cells of L3 larvae. Granule markers are in gray and nuclear DAPI in blue. Scale bars, 5 μm (J to M) Violin plots showing partition coefficients (n = 6 individuals and n = 18 ROIs per genotype) and nuclear to cytoplasmic intensity ratios for Dcp1 (n = 6 individuals and n = 12 cells per genotype) (J and K), partition coefficients for Tral (n = 6 individuals and n = 18 ROIs per genotype) (L), and Me31B (n = 6 individuals and n = 18 ROIs per genotype) (M), in control, heca^−/−, and btl > UAS-heca, indicating changes in granule component localization. (N and O) Snapshots of in vivo time-lapse for RIN::mCherry in gray, 29 min after stress induction with arsenite in control (N) and heca^−/− (O) Tr2 cells. (P) Graph showing time course of SG formation upon arsenite treatment in control and heca^−/− cells, with notable delays in heca^−/− (n = 3 individuals, error bars show SD of the mean in each time point). (Q and R) Images showing O-propargyl-puromycin (OPP) incorporation to quantify new protein synthesis in wing discs with altered Heca levels in hecaRNAi (Q and R) and in UAS-heca, in both cases induced by the CiGal4 (T and U). The Ci domain is labeled by a UAS-GFP construct induced by the same driver. (R) and (U) are heatmaps of OPP incorporation levels in (Q) and (T) (n = 9 individuals, n = 18 ROIs of the anterior, and n = 18 ROIs of the posterior disc compartment per genotype). Scale bars, 20 μm. (S and V) One sample t test was used to test the null hypothesis for the differences in paired values of mean intensities between the posterior versus the anterior compartment. The exact P values are noted on the graph.

To further investigate the impact of Heca under stress, we examined stressed tissues in heca mutants, using the SG markers Rasputin (Rin) and Capr (29). Upon 45 min post arsenite exposure of dissected tissues ex vivo, SGs in heca mutants exhibited considerable variability, ranging from a complete absence to near-normal aggregate formation compared to controls (fig. S3). As this variability of phenotypes observed could point to a possible effect on the formation dynamics rather than the incapacity of the mutant cells to form SG structures, we used a Rin::mCherry knock-in line (15) that enabled us to monitor the time course of SG formation via time-lapse imaging in ex vivo cultured tissues (see Materials and Methods for details). In control tissues, SGs were detectable from approximately 20 min following arsenite addition. In contrast, in heca mutant tissues, SG appearance was delayed poststress induction (movie S1 and Fig. 4, N to P). A similar, albeit less marked, delay in SG formation was also observed following exposure to alternative stressors, such as temperature shifts and nutrient deprivation (fig. S4).

Together, these findings suggest that Heca is integral to the proper formation and function of RNP granules, affecting not only the physical properties and kinetics of formation of these structures but also modulating their composition and potentially their functional activity. To further test this hypothesis, we overexpressed heca in the trachea using a btlGal4 driver and examined the morphology of RNP granule components. In Tr2 tracheal cells overexpressing heca in standard conditions, we also confirmed changes in the subcellular distribution of PB components. The partition coefficient of Dcp1 was found decreased in comparison to the control (Fig. 4, C and J). Tral distribution was profoundly affected, forming large aggregates in the cytoplasm with a significant increase of its partition coefficient (Fig. 4, F and L). In the case of Me31B, we observed a reduction in the partition coefficient, while we observed a notable variation in the calculated values, indicating that Me31B exhibits a unique behavior upon Heca overexpression (Fig. 4, I to M). Our results indicate a specific regulatory function of Heca on each granule component, which could have significant implications for cellular physiology and metabolism. In particular, the deleterious effect or the excess of Heca protein suggests fine-tuned interactions between components of RNP granules; absence of Heca does not allow proper aggregate formation, while an excess of the protein could potentially interfere with other interactions impairing as well proper granule formation.

Given the pivotal role of granules in RNA metabolism and the observed effects of Heca on RNP granule composition and dynamics, we postulated that these perturbations could be also mirrored by alterations in the protein synthesis machinery. To test this hypothesis, we used the O-propargyl-puromycin (OPP) assay to quantify new protein synthesis in tissues with altered heca levels. For these experiments, we used wing discs as a model. Using the CiGal4 driver, which allowed us to directly compare in the same disc the anterior (experimental) and posterior (control) compartment, we either knocked down or overexpressed heca and observed the subsequent effects on OPP incorporation. heca knockdown was associated with an increased OPP signal (Fig. 4, Q to S), while its overexpression did not significantly alter OPP incorporation (Fig. 4, T to V). This suggests that a specific threshold of Heca is essential for maintaining control over bulk protein synthesis. It should be noted that since OPP incorporation is quantitative rather than qualitative method, future work would be required to determine whether specific transcripts or proteins are preferentially affected following modifications in Heca levels.

Functional parallels between Drosophila and human HECA

Building upon the results concerning the relationship between Drosophila Heca and RNP granule components, we investigated the potential functional conservation for its human homolog. Our results on the role of heca in stress protection in Drosophila, led us to hypothesize that human HECA might similarly had a role in stress protection. To test this, we used shRNA to knock down HECA expression in HeLa cells and monitored reactive oxygen species (ROS) production (see fig. S5 for shRNA efficiency in HeLa cells). Using the CellROX reagent, we observed an increase of ROS in cells with lower HECA levels compared to the respective controls (Fig. 5, A to D), supporting HECA’s involvement in cellular stress responses.

Fig. 5. Human HECA loss induces ROS production.

(A to C) Imaging of reactive oxygen species (ROS) production in HeLa cells stained with CellROX under control (A), scramble shRNA (B), and HECA shRNA (C) conditions, showing increased ROS in HECA knockdown cells. (D) Violin plots showing quantification of ROS levels in control, scramble, and HECA shRNA conditions, indicating significantly higher ROS levels in HECA knockdown cells. Kruskal Wallis H, P = 0.0003, followed by Dunn’s multiple-comparisons tests. Exact P values of post hoc tests noted on the graph (n_control = 215, n_scramble = 225, and n_shRNA-HECA = 290 cells).

To ascertain whether like the Drosophila homolog, also HECA is found in proximity with RNP granule components, we examined its spatial relationship with DCP1A and G3BP1. Imaging analysis revealed close interaction between HECA and DCP1A in nonstressed cells (Fig. 6A). Only following stress induction, it was also detected in SGs marked by G3BP1 (Fig. 6B). PLAs also confirmed the HECA interaction with DCP1A under normal conditions (Fig. 6, E, F, and M), while this interaction became more evident after stress induction (Fig. 6, I, J, and M). Positive PLA spots were detected also in stressed cells for the HECA-G3BP1 pair (Fig. 6, K, L, and N), confirming the presence of HECA in SGs.

Fig. 6. Colocalization and Interaction of HECA with DCP1A and G3BP1 in HeLa cells under normal and stress conditions.

(A to D) Colocalization of HECA with DCP1A under normal conditions (A) and G3BP1 upon arsenite treatment (B) in HeLa cells. Scale bars, 2 μm. Inset images show magnified views of colocalization areas. The plots (C and D) show fluorescence intensity profiles along the yellow lines drawn in the insets (A and B). PLA spots showing interaction of HECA with DCP1A (E, F, I, and J) and G3BP1 (G, H, K, and L) under normal conditions (E to H) and after arsenite stress induction (I to L) in control and HECA shRNA cells DAPI stains nuclei in blue and PLA spots labeled in magenta. (M and N) Violin plots showing quantification of PLA spots per cell for HECA with DCP1A (M) and G3BP1 (N) under normal and stress conditions. Kruskal Wallis H, P < 0.0001, followed by Dunn’s multiple-comparisons post hoc tests. Exact P values for each test performed shown on the graph (for HECA-DCP1A PLA, cells analyzed under normal conditions: n = 30 control and n = 24 shRNA and cells analyzed in stress: n = 26 control and n = 28 shRNA; for HECA-G3BP1 PLA, cells analyzed under normal conditions: n = 25 control and n = 23 shRNA and cells analyzed in stress: n = 26 control and n = 21 shRNA).

Mirroring the observations in Drosophila, HECA knockdown in HeLa cells led to higher nuclear to cytoplasmic ratio and a reduced partition coefficient for DCP1A, indicative of altered subcellular localization and aggregation propensity. Control experiments with scramble RNAi did not manifest any prominent effects on DCP1A protein accumulation, underscoring the specificity of HECA’s influence (Fig. 7, A to F).

Fig. 7. Effects of HECA knockdown on DCP1A granule stability, localization, and protein synthesis in HeLa cells.

(A to C) Super-resolution images showing changes in DCP1A granules stability and subcellular localization in control (A), scramble (B), and HECA shRNA (C) cells. (D to F) Violin plots showing particle volume (particles analyzed: n_control = 42, n_scramble = 68, and n_shRNA-HECA = 103) (D), partition coefficient (particles analyzed: n_control = 8, n_scramble = 9, and n_shRNA-HECA = 10) (E), and nuclear to cytoplasmic ratio (cells analyzed: n_control = 9, n_scramble = 9, and n_shRNA-HECA = 10) (F), for DCP1A under control, scramble, and HECA shRNA conditions. Kruskal Wallis H, followed by Dunn’s multiple-comparisons tests performed for all parameters analyzed. Exact P values of post hoc tests noted on the graph. (G to I) Images showing O-propargyl-puromycin (OPP) incorporation to quantify new protein synthesis in HeLa cells with control (G), scramble (H), and HECA shRNA (I) treatments. Heatmaps show OPP incorporation levels. (J) Violin plot showing quantification of OPP incorporation levels in control, scramble, and HECA shRNA cells, indicating increased protein synthesis in HECA knockdown cells. Kruskal Wallis H, P = 0.0003, followed by Dunn’s multiple-comparisons tests. Exact P values of post hoc tests noted on the graph (n_control = 28, n_scramble = 13, and n_shRNA-HECA = 19 cells). Dashed lines in violin plots represent median, upper, and lower quartiles.

Last, we extended our investigation to explore the potential effects of HECA knockdown on bulk protein translation. Following HECA shRNA treatment, we detected an increase in OPP incorporation, suggesting an up-regulation in protein synthesis akin to the effects observed with the Drosophila homolog (Fig. 7, G to J).

In summary, our results indicate that human HECA shares functional similarities with its Drosophila counterpart in stress protection mechanisms. HECA’s involvement in reactive oxygen species (ROS) management, its subcellular localization in relation to RNP components, and its effects on protein translation during stress induction highlight its potential role in cellular homeostasis and stress response pathways.

Heca proteins identified as RBPs

The identification of Heca proteins as components of RNP granules prompted us to question their functional role within these structures. Notably, the ability to bind RNA is a hallmark of many granule-associated proteins, enabling them to scaffold granule formation, regulate RNA stability, and influence translation during stress (30, 31). This raised a key question of whether Heca proteins directly bind RNA. To address this, we used orthogonal organic phase separation (OOPS) (32), a robust method that isolates RNA-protein complexes by leveraging phase separation properties. Applying OOPS in Drosophila S2 and human HeLa cell lines, we observed that typical to the RBP behavior, both proteins were present in the interphase, demonstrating their potential to bind RNA directly. For Drosophila S2 cells, Fragile X Mental Retardation 1 (dFMR1), a well-studied RBP (33, 34), served as a positive control, confirming the expected RNA-binding activity. Syntaxin 1A (Syx1A), a SNARE complex member participating in vesicle fusion (35, 36), served as a negative control showing no such association (Fig. 8A). Similarly, for HeLa cells, DCP1A was used as a positive control and was detected in the interphase, whereas Vinculin (VCL), an actin binding protein (37) used as negative control, did not (Fig. 8B). This discovery fills a significant gap in understanding their molecular role, offering a mechanistic explanation for their influence on RNP granule formation and translational control. The RNA binding activity suggests that Heca proteins might directly interact with specific transcripts, playing an active role in RNA metabolism. This could include stabilizing certain mRNAs, promoting their storage within granules, or facilitating their degradation during stress. Moreover, their ability to bind RNA likely underpins their role in SG dynamics, as RNA binding is essential for sequestering transcripts into SGs and modulating translation during stress. The conservation of RNA binding activity across species underscores its evolutionary importance, further supporting the hypothesis that Heca proteins function as central regulators of RNA dynamics.

Fig. 8. Identification of Heca proteins as RBPs using orthogonal OOPS.

(A and B) Western blots following OOPS of Drosophila S2 and human HeLa cells extracts, revealed that Heca proteins cosegregate with RNA binding proteins (RBPs) in the interphase, indicating their potential RNA binding activity. Positive controls, dFMR1 in S2 cells (A) and DCP1A in HeLa cells (B), were detected in the interphase, confirming RNA binding activity. Negative controls, Syx1A (S2 cells) (A) and VCL (HeLa cells) (B), showed no interphase association, validating the specificity of the assay. The results suggest that Heca proteins are classified amongst the RBPs. The OOPS workflow (C) allows the isolation of RNA-protein complexes through phase separation. RNA-protein complexes segregate into the interphase, distinct from the aqueous (RNA) and organic (protein) phases, enabling their specific identification and analysis. For presentation clarity, the images of Western blots shown here were adjusted using Adobe Photoshop. Uncropped versions of the blots are available in the Supplementary Materials.

DISCUSSION

Heca proteins as regulators of RNP granules and stress responses

In this study, Heca proteins emerge as central regulators of RNP granules, influencing cellular responses to stress and contributing to RNA metabolism. In particular, we have shown Heca interacting with granule components including Dcp1, Tral, Me31B, and Rasputin, each of them playing a distinct role within the cellular framework. Tral is involved in mRNA trafficking and stability and is critical for efficient protein synthesis and secretion, while in the context of PBs, it facilitates the storage and decay of mRNAs (38–40). DCP1 plays a pivotal role in the degradation of mRNA by removing the 5′ cap, a key step in the regulation of mRNA decay. It functions together with DCP2, its catalytic partner, to orchestrate this decapping activity, which is a crucial determinant of mRNA half-life and thereby gene expression (40, 41). Me31B is identified as a DEAD-box protein implicated in a wide array of RNA metabolic processes. It has been found to participate in translational repression, mRNA decay, and the regulation of mRNA stability and localization, playing a pivotal role in post-transcriptional gene regulation. Me31B is also involved in the silencing of translation during Drosophila oogenesis (16, 40, 42). Rasputin in Drosophila is the homolog of G3BP1, which is implicated in the stress response and mRNA metabolism. It functions under both normal and stress conditions to regulate mRNA stability and translation, thus playing a crucial role in the cell’s ability to manage environmental stress (25). Rasputin under normal conditions aids in the stabilization and translation of its target mRNAs, contributing to cellular growth and homeostasis. Under stress conditions, it is essential for the formation and function of SGs, where it sequesters mRNAs to halt their translation and protect them under adverse conditions (20). In this capacity, Rasputin is a positive regulator of RNA metabolism, maintaining mRNA stability and ensuring the appropriate localization of mRNAs for their future translation when the stress subsides.

Mechanistic insights for the role of Heca in RNA metabolism and cellular adaptation to stress

Our results suggest that Heca controls the subcellular distribution of Dcp1, contributes to the recruitment and stabilization of Tral and Me31B in the PB structures, and interacts with Rasputin both under normal and stress conditions, fine-tuning the kinetics of the SG formation. Combining the previous knowledge about each of these proteins, we would expect that Heca has a direct effect on the decapping process of mRNAs contributing to the targeted decay of transcripts, facilitates the trafficking of RNAs from the ER contributing to their subsequent storage and/or decay, and enhances the translation of selected transcripts. Under stress conditions, it allows the timely formation of SGs, which are critical for cells to cope with stress as they temporarily sequester mRNAs and translational machinery components, to prevent the translation of nonessential proteins and to protect mRNAs from degradation. Notably, our identification of Heca as an RBP further supports its direct involvement in these processes. However, RNA-protein binding profiling techniques would be needed to establish Heca target specificity and further elucidate its role in RNA stability and translation in various stress contexts.

Dual role of Heca in development and stress resilience

Beyond its evident connection to stress-related responses, the overall effects of Heca in cellular and organismal physiology may also stem from its roles in translational regulation and the modulation of the TOR pathway (2, 3, 43). In addition, a particularly compelling aspect of heca’s function is its involvement in ecdysone production and secretion (3). Ecdysone acts as a systemic hormonal regulator driving critical developmental transitions, including molting and metamorphosis. However, the activation of ecdysone signaling imposes considerable metabolic demands, necessitating extensive cellular reorganization, such as enhanced protein synthesis, cytoskeletal remodeling, and increased energy metabolism. These processes can create stress-like conditions at the cellular level, necessitating adaptive mechanisms to maintain homeostasis. The dual role of Heca, regulating ecdysone production at the systemic level while autonomously protecting cells from stress, positions it as a central mediator in coordinating developmental and stress-related responses. This functional versatility likely underlies the developmental defects observed in heca mutants under both normal and stressed conditions, as the loss of heca disrupts key pathways required for both stress resilience and normal physiological processes.

Interconnection between translational regulation, stress responses, and development

To date, the characterization of Heca proteins has largely focused on their role in regulating the TORC1 pathway in both Drosophila and humans, with particular attention given to its association with Unkempt, a suppressor of the same pathway (2, 3, 44, 45). Unkempt an RBP that plays a pivotal role in the regulation of key developmental transcripts (46) was recently shown to interact with the CCR4-NOT effector complex for the specific recognition of mRNAs (47). The functions of Unkempt and the exclusive expression of heca in APCs, underscore the critical importance of Heca for organismal development. It is conceivable that Heca may extend its functionality beyond stress response, potentially engaging directly with Unkempt in translational control mechanisms that are essential during normal physiological states. Furthermore, Heca’s identification as a constituent of RNP granules adds a layer of complexity to its influence over the TORC1 signaling axis. TORC1 is known to be modulated by SG components in response to stress stimuli (48, 49), suggesting that Heca’s association with these structures could have profound implications on TORC1 dynamics. This association presents a possible mechanism through which Heca may exert finely tuned control over cell physiology, bridging the gap between stress response, developmental regulation, and growth control. Such intricate regulation by Heca, in particular in the context of RNP granules, illuminates its potential as a key molecular orchestrator, balancing cellular growth with the demands of environmental adaptation and stress resilience.

MATERIALS AND METHODS

Drosophila strains

All Drosophila melanogaster strains were raised at 25°C under standard conditions. The Oregon-R strain was used as control. The RIN::mCherry is described in (15), the btl Gal4 in (50), and the Ci Gal4 in (51). hdcRNAi was obtained from the Vienna Drosophila Resource Center, #104322. The following stocks were obtained from the Bloomington Drosophila Stock Center (BDSC): heca^BG23007 BDSC #12410, UAS-heca BDSC #64056, heca ^Df(3R)ED6332 BDSC #24141, and UAS-GFP-RNAi BDSC #9331. The larvae noted as heca^−/− are heteroallelic combinations of the heca mutant allele heca^BG23007 over heca ^Df(3R)ED6332, a deficiency for the heca gene. To test for SG formation in a heca mutant background, we created a recombinant chromosome carrying the RIN::mCherry insertion with the heca^BG23007 allele. SG formation was tested in heca^BG23007::RIN::mCherry/heca ^Df(3R)ED6332 flies. All experiments involving D. melanogaster were performed in accordance with institutional standards for the use of invertebrate model organisms. Ethical approval was not required for this study, as D. melanogaster is not subject to animal ethics regulations under the European Directive 2010/63/EU.

Drosophila tissue immunostaining

Tracheae or wing discs from L3 larvae were dissected in 1× phosphate-buffered saline (PBS) and fixed in 4% formaldehyde (Invitrogen, catalog no. 15650599) for 20 min at room temperature (RT). The tissues were rinsed in 0.1% Triton X-100 (PBT), blocked in PBT + bovine serum albumin (BSA) 0.5% for 1 hour and incubated at 4°C with primary antibodies diluted in PBT + BSA 0.5% overnight.

The following primary antibodies were used for the immunostainings: mouse anti-dHeca (1:3), Developmental Studies Hybridoma Bank (DSHB; U33) (5); rat anti-Tral (1:100), a gift from E. Wahle (42); rabbit anti-dDcp1 (1:20), a gift from T.-B. Chou (41); mouse anti-Me31B (1:3000), a gift from Florence Besse; rabbit anti-Rasputin (1:700), a gift from E. R. Gavis (21); and rabbit anti-Caprin (1:700), a gift from O. Papoulas (52).

After incubation with primary antibodies, the tissues were washed with PBT (3 × 10 min washes) and incubated with the corresponding secondary antibodies (Alexa-conjugated dyes 488, 555, and 647, Life Technologies, 1:450), for 2 hours at RT, followed by 3 × 10 min washes with PBT, and then rinsed with PBS before mounting in Vectashield with 4′,6-diamidino-2-phenylindole (DAPI; Palex Medical SA, catalog no. h1200).

Cell culture

HeLa (ATCC CCL-2) and human embryonic kidney (HEK) 293T (ATCC CRL-11268) were maintained and cultured at 37°C, 5% CO₂, in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 25 mM HEPES, 10% fetal bovine serum (FBS), streptomycin (100 U/ml), and penicillin (100 U/ml). Cells were routinely grown in plastic tissue culture dishes and harvested with a solution of trypsin-EDTA (Cultek 5525-053-CI) while in logarithmic phase.

Lentiviral infection and SiRNA generation

For HECA silencing, HeLa cells were infected with shRNA-HECA lentiviral particles that were purchased from Sigma MISSION shRNA Library (TRC, clone ID TRCN0000161697). Cells were infected with lentiviral particles for 1 day in polybrene medium (2 μg·ml⁻¹). On day 3, the culture medium was removed and replaced with complete DMEM. After 5 days, stable clones expressing shRNA were selected with puromycin (5 μg·ml⁻¹) (53) and last GFP-positive cells expressing shRNAs were flow sorted. To generate Scramble negative control, HEK293T cells were transfected with scramble shRNA plasmid (PLKO.1 scramble, Addgene, plasmid catalog no. 1864) packaging (psPAX2, Addgene, catalog no. 12260) and envelope (pMD2.G, Addgene, catalog no. 12259) plasmids in a three-plasmid lentivirus packaging system (54). We collected transfected cell supernatants containing lentivirus particles over 36 hours and aliquoted and stored these lentivirus-containing supernatants at −80°C. We used 5 μl of lentiviral supernatant per well in six-well plate to infect HeLa cell at 70% confluency. Cells were infected with lentiviral particles for 1 day in polybrene medium (2 μg·ml⁻¹). On day 3, the culture medium was removed and replaced with complete DMEM. After 5 days, stable clones expressing shRNA were selected with puromycin (5 μg·ml⁻¹).

Real-time quantitative polymerase chain reaction analysis

We used quantitative real-time quantitative polymerase chain reaction (RT-qPCR) to determine whether shRNAs affected HECA mRNA. Total RNAs were extracted with TRIzol Reagent (Life Technologies, 15596018), purified with an NZY total RNA isolation kit (NZYTech, MB13402) and treated with deoxyribonuclease I. cDNAs were prepared from 0.8 μg of RNA using the Revert Aid H Minus First Strand cDNA Synthesis Kit and oligo-dT primers (Life Technologies, K1632). The analysis was carried out on a LightCycler 480 Real-Time PCR System with SYBR Green PCR Master Mix (Thermo Fisher Scientific, K022). RNA levels were normalized to the housekeeping gene. The relative quantification was calculated using the ΔΔC_t method. An Actin control gene was used for normalization. Error bars indicate the SE from biological replicates, each consisting of technical triplicates. The oligonucleotides used were as follows: HECA FORWARD, GCACTATGTACACCTACGACATCC; HECA REVERSE, TCCCACAGTGCTTACAGTTCA; ACTIN FORWARD, CACCATTGGCAATGAGCGGTTC; and ACTIN REVERSE, AGGTCTTTGCGGATGTCCACGT.

Cell fixation and fluorescence immunostaining

HeLa cells were cultured on chambered microscope glass slides (Cell culture Slide, eight-well, Lab Clinics PC30108). Cells were washed three times in PBS and then fixed in 4% paraformaldehyde (PFA) for 15 min at RT and 10 min in cold methanol. After being washed three more times in PBS, cells were permeabilized by applying 0.2% TritonX-100 for 45 min at RT. Permeabilized cells were blocked in PBS + BSA 1%, and the following primary antibodies diluted in PBS + BSA 1%, were added: rabbit anti-HECA (1:200), Thermo Fisher Scientific, PA5-31372; mouse anti-G3BP1 (1:1000), Sigma-Aldrich, SAB1406936; and mouse anti-DCP1A (1:200), Abnova, ABNOH00055802-M06.

After 1-hour incubation at RT, cells were washed three times in PBS and incubated with the corresponding secondary antibodies (Alexa-conjugated dyes 488, 555, Life Technologies, 1:400) for 30 min at RT. Cells were mounted using Vectashield Mounting Media With DAPI (Palex Medical SA, catalog no. h1200).

For stress induction, arsenite treatment was used; Hela cells were cultured in the completed DMEM for 24 hours until 80% confluency, then medium was changed to DMEM with 20 mM arsenate (Sigma-Aldrich, 06277) for 30 min. The cells were washed with PBS three times before fixation.

Proximity ligation assays

The PLA was performed using the Duolink In Situ Red Starter Kit Mouse/Rabbit (Sigma-Aldrich, catalog no. DUO92101) according to the manufacturer’s instructions with some modifications.

Larvae under normal condition or upon exposure to stress (1 mM arsenite in Schneider’s medium for 45 min at RT) were dissected and tissues were fixed in 4% formaldehyde for 20 min and rinsed three times with PBS. After permeabilization, the samples were blocked in PBT for 1 hour at 37°C, left at RT for 1 hour more and then incubated overnight at 4°C with the following antibody combination: mouse anti-dHeca (1:3), DSHB (U33) (5)–rabbit anti-dDcp1 (1:20) (42); and mouse anti-dHeca (1:3), DSHB (U33) (5)–rabbit anti-Rasputin (1:700) (21).

The following day, the samples were incubated with the MINUS and PLUS PLA probes corresponding to the primary antibodies used, followed by 45-min ligation and 100-min amplification, using Texas Red–labeled oligos to generate the signal. Last, the tracheae were dissected in PBS and were mounted in Vectashield DAPI-containing medium. All incubations were performed in a humidity chamber using a volume of 20 to 40 μl per well. For imaging, the Zeiss 880 confocal microscope was used.

For PLA, HeLa cells were grown on chambered microscope glass until 80% confluency or under normal condition or upon exposure to arsenate (20 mM arsenite in complete DMEM for 30 min at 37°C). After three washes with PBS, cells were fixed with formaldehyde 4% for 15 min. Cells were permeabilizated and blocked with PBT for 1 hour at 37°C, left at RT for 1 hour more, and then incubated overnight at 4°C with the following antibody combination: rabbit anti-HECA (1:200), Thermo Fisher Scientific PA5-31372–mouse anti-DCP1 (1:200), Abnova ABNOH00055802-M06; and rabbit anti-HECA (1:200), Thermo Fisher Scientific PA5-31372–mouse anti-G3BP1 (1:1000), Sigma-Aldrich, SAB1406936.

The following day, the samples were incubated with the MINUS and PLUS PLA probes corresponding to the primary antibodies used, followed by 45-min ligation and 100-min amplification, using Texas Red–labeled oligos to generate the signal. Cells were mounted in Vectashield with DAPI and immediately analyzed. Samples were protected from light throughout the entire experimental procedure. Images were acquired with a Zeiss LSM880 confocal microscope.

OPP incorporation assays

The Click-iT Plus OPP Alexa Fluor Protein Synthesis Assay Kit (Invitrogen, Thermo Fisher Scientific, catalog no. C10457) was used as reagent for fluorescent labeling of nascent protein synthesis, according to the manufacturer’s instructions.

Larvae were inverted in Schneider’s Drosophila Medium (Gibco, Thermo Fisher Scientific, 21720024) at RT and transferred in Schneider’s medium, 10% FBS with 20 μM OPP for 30 min. Inverted L3 larvae were washed in PBS 3 × 10 min. Wing discs were dissected and fixed with 4% formaldehyde and permeabilized in PBT for 15 min. After two PBS washes, the samples were incubated in OPP staining solution for 40 min protected from light followed by a rinse in component F. Tissues were mounted in Vectashield with DAPI.

HeLa cells were plated at 80% confluency on chambered microscope glass slides and recovered overnight. Cells were incubated for 30 min in fresh complete DEMEM with Click-iT OPP 20 μM at 37°C, 5% CO₂. Cells were washed in PBS and then fixed in 4% PFA for 15 min. Cells were permeabilized in PBT for 15 min. After two PBS rinses, cells were incubated in OPP staining solution for 30 min while protected from light, followed by a rinse in component F. Cells were mounted in Vectashield with DAPI.

Cell ROX assays

For ROS detection, HeLa cells were grown on chambered microscope glass until 80% confluency and then were incubated for 20 min in fresh DMEM complete medium containing 5 μM CellROX Deep Red Reagent (Thermo Fisher Scientific, catalog no. C10422) at 37°C, 5% CO_2. After one wash with PBS, cells were fixed with formaldehyde 4% for 15 min. Cells were mounted in Vectashield with DAPI. Samples were protected from light throughout the entire experimental procedure. Images were taken with a Zeiss LSM880 confocal microscope immediately after the procedure described here.

OOPS

For OOPS, we followed the protocol described in detail by Queiroz et al. (32) with some adjustments. In brief, Drosophila S2 cells or human HeLa cells were grown until 80% of confluency and ultraviolet irradiated or not, to account for the methods specificity. After three washes in PBT, cell lysis was performed adding first 20 μl of 80 mM dithiothreitol or 100 μl of urea buffer (8 M urea, 1% SDS in PBS) and lastly, 1 ml of TRIzol. Phase separation was obtained by adding 200 μl of chloroform (VWR) and centrifugation at 12,000g for 15 min at 4°C. Aqueous (transparent) and organic (pink) phases were separated, preserving the interphase. Three additional phase separations were performed. The resulting interphase was precipitated with TRIzol:methanol (1:9) and centrifuged at 14,000g for 10 min at RT. Pellets were partially resuspended in water and digested with 2000 U of ribonuclease T1 (RNase T1; Thermo Fisher Scientific, EN0541) and 4 μg of RNase A (Sigma-Aldrich) for 10 min at 4°C. The pellet was further resuspended by adding 1% SDS and 100 mM TEAB (Honeywell, 17902) followed by sonication for six cycles (30-s ON/OFF). Subsequently, the samples were incubated overnight at 37°C in RNase mix (2000 U of RNase T1 and 4 μg of RNase A) that was adjusted to 1 mM MgCl₂. For Western blots of the obtained samples, the following primary antibodies were used: mouse anti-dHeca, DSHB (U33) (1:20); mouse anti-dFMR1, DSHB (5B6) (1:2000); mouse anti-Syx1A, DSHB (8C3) (1:5000); rabbit anti-HECA, Thermo Fisher Scientific PA5-31372 (1:2000); mouse anti-DCP1A, Abnova ABNOH00055802-M06 (1:2000); and mouse anti-Vinculin (Santa Cruz Biotechnology) H10 sc-25336. (1:500). For chemiluminescence detection, we used IRDye 800 CV Donkey anti-mouse and IRDye 680RD Goat anti-rabbit (Licor) (1:10,000).

Ex vivo treatments and stress induction of Drosophila tissues

For ex vivo arsenite or temperature treatments, larvae were inverted in Schneider’s medium with 10% FBS. For oxidative stress induction, the medium was supplemented with 500 μM arsenite. Subsequently the tissues were either washed with PBS and fixed in 4% formaldehyde or used for time lapse imaging to observe the formation of SGs. For nutrient deprivation experiments, the larvae were dissected in PBS and transferred immediately to KRB (Thermo Fisher Scientific, catalog no. J67591.K2). Heca subcellular localization analysis after heat stress or starvation was performed on tissues of L3 larvae that were either heat-shocked for 1 hour at 37°C or starved for 9 hours in empty bottles containing water soaked filter paper.

Image acquisition and image analysis

Images were obtained with the Zeiss 880 confocal microscope, using the Plan-Apochromat 40×/0.95, 63×/1.4, and 100×/1.4 Oil DIC M27 objectives. Time-lapse in vivo imaging was performed with the Fast Airyscan mode using the 40× objective. Tissue mounting for time-lapse imaging was performed as described in (55). Super-resolution images were obtained with the Airyscan mode followed by processing by the Zen Software or the SIM mode of the microscope. The same microscope settings were used for all comparative analyses. Volumetric calculations of aggregates were performed using the Imaris Software (Oxford Instruments). A lower cutoff of 0.1μm³ was used for the volume of the particles selected for quantifications. The nearest-neighbor distance (NND) calculation was performed in two-dimensional (2D) projections of 0.2-μm stacks, where in the selected regions of interest (ROIs), the foci maxima were detected and a mask was created to analyze the NND distances in Fiji. For the distance analysis, we used the DiAna Fiji plugin for classical segmentation after applying a Gaussian filter of 2 and using manual thresholding for each image. The distance analysis was subsequently performed in the segmented 3D objects. The find maxima function of Fiji was used to detect the PLA puncta in projections of the selected acquisitions. The partition coefficients were calculated in average projections of the respective images, dividing the intensity of manually selected puncta to the background intensity of the nearby cytoplasmic region of the same area, within a maximum distance of 20 pixels from the border of the particle. The nuclear-to-cytoplasmic ratios were also calculated measuring the intensities in sum projections in ROIs within the nuclear area defined by the DAPI staining, versus cytoplasmic region of the same cell. Quantifications of SGs were performed by applying background subtraction and subsequently a Gaussian filter to max projections of the time lapse acquisitions per each time point, using the analyze particles function of Fiji to calculate the segmented SGs in ROIs of the same area among the acquisitions. Final figures presented here were produced in the Adobe Photoshop CC software.

Statistical analysis

An unpaired t test with Welch’s correction was used to compare two populations with unequal variances. We used the Kolmogorov-Smirnov test to compare probability distributions. For multiple-comparisons analysis of variance (ANOVA), the Brown-Forsythe test corrected for unequal variances among samples, followed by Dunnett’s test for post hoc pairwise comparisons. The nonparametric Kruskal-Wallis H test was used to compare multiple groups with skewed data distributions, followed by Dunn’s test for multiple comparisons using rank sums. The Kolmogorov Smirnov, D’Agostino Pearson, and Shapiro-Wilk tests were used to assess the normality of data distributions. Statistical analysis, data processing, and graphical representations were performed in GraphPad Prism 10.2.2 software.

Supplementary Materials

The PDF file includes:

Figs. S1 to S5

Legend for movie S1

Uncropped Western Blots

Legend for dataset S1

Other Supplementary Material for this manuscript includes the following:

Dataset S1

Movie S1