Pex3 promotes formation of peroxisome-peroxisome and peroxisome-lipid droplet contact sites

Pex3 promotes formation of peroxisome-peroxisome and peroxisome-lipid droplet contact sites Lucía Amado1, Louis Percifull2,3, Rico Franzkoch4,5,6,Vico Flatemersch1, Eleni Joana Brüggemann7, Olympia Ekaterini Psathaki4,5, Maya Schuldiner8, Maria Bohnert2,3, Margret H. Bülow7& Ayelén González Montoro1,5 Peroxisomes are ubiquitous organelles that mediate central metabolic functions, such as fatty acid βoxidation, as well as diverse tissueand organism-specic processes. Membrane contact sites, regions of close apposition with other organelles for direct communication, are central to several aspects of their life cycle. Pex3 is a conserved multifunctional peroxisomal transmembrane protein that is involved in the insertion of peroxisomal membrane proteins, in pexophagy, and in the formation of membrane contact sites. Here, we show that high Pex3 levels in Saccharomyces cerevisiae induce the formation of peroxisome clusters surrounded by lipid droplets, mediated by peroxisome-peroxisome and peroxisome-lipid droplet contact sites.This clustering occurs independently of Pex3 partners in other processes Pex19, Inp1, and Atg36. The cytosolic domain of Pex3 binds peroxisomes, suggesting a direct role in homotypic contact site formation. Lipid droplet-peroxisome contact sites require the lipid droplet-localized triacylglycerol lipaseTgl4, which is enriched at this interface along with other lipases. Pex3 overexpression in Drosophila melanogaster similarly alters peroxisome and lipid droplet morphology and promotes contact site formation.Together, our results oer novel molecular insights into homotypic peroxisome contact sites and peroxisome-lipid droplet contact sites across species. Peroxisomes are found in most eukaryotic cells and are the place of important metabolic reactions, including the oxidation of fatty acids, the synthesis of ether lipids, and the detoxication of hydrogen peroxide and glyoxylate. While some peroxisomal functions are tissueor organism-specic, others are highly conserved, e.g. the oxidation of fatty acids, a process that is virtually ubiquitous1. Impaired peroxisome biogenesis as well as defects in peroxisomal metabolic pathways result in severe human diseases collectively known as peroxisomal disorders1. Membrane contact sites are structurally dened regions of close organelle apposition without membrane fusion2,3. A central function of contact sites is the exchange of material among the compartments, including the transport of luminal material and of membrane lipids. Furthermore, the physical attachment of the organelle membranes can aect organelle fusion, ssion, and positioning4. Membrane contact sites exist between virtually all pairs of organelles5–7 and coordinate multi-organelle processes4,8. Contact sites between peroxisomes and other organelles play important roles in dierent aspects of the peroxisome life cycle9. For example, in the yeast Saccharomyces cerevisiae, peroxisome contact sites with the endoplasmic reticulum aect peroxisome proliferation10,11 and the formation of peroxisome contact sites with the cell periphery determines the distribution of these organelles among mother and daughter cells12. Pex3 is a peroxisomal membrane protein with multiple functions and interactors. It is involved in the targeting of peroxisomal membrane proteins, by acting as a docking factor for Pex19, a cytosolic receptor for 1Cellular Communication Laboratory, Department of Biology/Chemistry, Osnabrück University, Barbarastrasse 13, 49076 Osnabrück, Germany. 2Institute of Cell Dynamics and Imaging, University of Münster, Von-EsmarchStr. 56, 48149 Münster, Germany. 3Cells in Motion Interfaculty Centre (CiM), University of Münster, Münster, Germany. 4iBiOs-Integrated Bioimaging Facility, University of Osnabrück, Osnabrück, Germany. 5Center of Cellular Nanoanalytic Osnabrück (CellNanOs), Barbarastrasse 11, 49076 Osnabrück, Germany. 6Division of Microbiology Department of Biology/Chemistry, Osnabrück University, Barbarastrasse 11, 49076 Osnabrück, Germany. 7Group Membrane Contact Sites, CURE3D Research Lab, Clinic for Cardiovascular Surgery, University Hospital Düsseldorf, Moorenstraße 5, 40225 Düsseldorf, Germany. 8Department of Molecular Genetics, The Weizmann Institute of Science, 7610001 Rehovot, Israel. email: [email protected] OPEN Scientic Reports | (2025) 15:24480 1 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports peroxisomal membrane protein precursors. Another function of Pex3 is the recruitment of the pexophagy receptor Atg36. Both of these functions are conserved in the human Pex3 protein13–19. Pex3 has additional roles in contact site formation. It acts as a membrane anchor for Inp1, a tethering protein involved in the retention of peroxisomes in the mother cell during cell division, through its interaction with the cell cortex12,20. In the yeast Hansenula polymorpha extensive vacuole-peroxisome contact sites are formed during growth phases that require peroxisomal growth. Pex3 is enriched at these contact sites, and overexpression of the protein results in contact site expansion during growth in glucose, suggesting that Pex3 might be a tether21. is prompted us to address the eects of Pex3 overexpression in Saccharomyces cerevisiae. We nd that cells that overexpress Pex3 contain peroxisome clusters surrounded by lipid droplets, which are formed by peroxisome-peroxisome and peroxisome-lipid droplet contact sites. We further show that these contact sites are independent of all the known Pex3 interactors. Instead, ecient formation of peroxisome-lipid droplet contact sites requires the lipid droplet-localized TAG-lipase Tgl4 but not other lipid droplet lipases, even though several of them are enriched in the interface. Interestingly, similar eects of Pex3 overexpression on peroxisome morphology and extended contact with the lipid droplets were also observed in Drosophila melanogaster, showing that these aspects of Pex3 function are conserved to metazoa. Altogether, our ndings expand our understanding of peroxisomal contact sites, and of the multi-functional protein Pex3. Results Overexpression of Pex3 causes a change in the morphology of peroxisomes and lipid droplets In Hansenula polymorpha, Pex3 was observed to be enriched in membrane contact sites between peroxisomes and the vacuole, and overexpression of this protein results in an extension of these contact sites21. us, we decided to test the phenotype of Pex3 overexpression in Saccharomyces cerevisiae. Overexpression from the strong constitutive TEF1 promoter resulted in a morphological change in peroxisomes marked by mCherry directed to the peroxisomal lumen by a PTS1 signal consisting of a serine-lysine-leucine sequence (mCh-SKL). While control cells show on average 5.5 mCh-SKL positive structures, cells overexpressing Pex3 contain mainly one structure (Fig.1A and B). is structure was found in close proximity to the vacuole in 55% of cells. We conrmed these observations by using other peroxisomal markers, Pex3 itself and Pex14, obtaining similar results (Supplemental Fig.1A and B). We analyzed the distribution of this peroxisomal structure among mother and daughter cells and found that it can be found in both compartments, with some preference for the bud (Supplemental Fig.1C). To characterize this structure we sought to describe its molecular microenvironment by proximity biotinylation22, by tagging Pex3 with TurboID23 at its C-terminus, which faces the cytosol. Cells expressing Pex3-TurboID under the TEF1 promoter were incubated with biotin for 3h and the biotinylated proteins were isolated by anity chromatography using a streptavidin matrix. e bait protein Pex3 was among the most highly enriched proteins, as was its known interactor Pex19 (Fig.1C). Gene Ontology (GO) Term enrichment analysis showed the GO term “Peroxisome” as the most represented annotation among our enriched proteins, as expected. Interestingly, the GO term “Lipid droplet” was also signicantly enriched (P value = 0.032, Fig.1C). is prompted us to address the subcellular localization of lipid droplets when Pex3 is overexpressed by uorescence microscopy. is analysis revealed that under these conditions, lipid droplets are strongly recruited to the peroxisomal structure (Fig.1D). We observed and quantied the presence of LDs in close proximity of peroxisomal structures in 90 cells from three independent experiments. Upon Pex3 overexpression, 95% of the peroxisomal structures were in close proximity to or completely surrounded by lipid droplets, whereas the remaining 5% of peroxisomal structures did not present any lipid droplet in their vicinity (Fig.1D). e size of lipid droplets is also slightly increased under these conditions (Supplemental Fig.1D). We conclude that overexpression of Pex3 causes a change in the morphology of peroxisomes and lipid droplets, resulting in the observation of a single peroxisomal structure per cell, which is in close proximity to lipid droplets. Pex3 overexpression induces a cluster of peroxisomes surrounded by lipid droplets, which includes peroxisome-peroxisome and peroxisome-lipid droplet contact sites To understand the characteristics of the structure formed by lipid droplets and peroxisomes upon Pex3 overexpression, we sought to enlarge the structure to gain spatial resolution. is was achieved by deleting PEX11, which results in enlarged peroxisomes24,25. Additionally, we cultured the cells in the presence of oleate, which resulted in enlarged lipid droplets. ese conditions allowed sucient spatial resolution to distinguish the peroxisomal matrix from the membrane, as peroxisomal membrane proteins (Pex13 and Pex3) were observed to surround BFP directed to the peroxisomal lumen by a C-terminal PTS1 signal (BFP-SKL) (Fig.2A and B, Supplemental Fig.1E–G). Similar to what we observed before, Pex3 overexpression caused clustering of the peroxisomal signal. e increased resolution provided by these conditions allowed us to observe that these clusters contain multiple individual maxima of the BFP-SKL signal (Supplemental Fig.1E–G). Figure2A and B show examples of these clusters, as well as line proles across them, illustrating that the peroxisomal membrane proteins show peaks between the individual maxima of the BFP-SKL signal (Fig.2A and B). is indicates that the peroxisomal structure formed upon Pex3 overexpression corresponds to a cluster of multiple peroxisomes. e lipid droplets marked by Erg6-2xmKate2 were observed to surround the cluster of peroxisomes. Given the close and specic proximity observed between these organelles upon overexpression of Pex3, we reasoned that such an organelle cluster would likely be formed by peroxisome-peroxisome and peroxisomelipid droplet contact sites. To test this, we performed on-section CLEM tomography of cells overexpressing Pex3-mKate2, using the mKate2 signal to locate the structures. is approach revealed clusters of peroxisomes surrounded by lipid droplets that included peroxisome-peroxisome and peroxisome-lipid droplet contact sites (Fig.2C–D). Figure2C shows a on-section correlative uorescence microscopy and TEM image from a cell Scientic Reports | (2025) 15:24480 2 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ overexpressing Pex3-mKate2, used to identify the relevant region for TEM tomography. Figure2D shows imaging of this region by TEM tomography, overlayed with a 3D model reconstructing the observed organelles. Figure2E and F and Supplemental videos 1 and 2 show example regions of the tomogram, containing peroxisome-lipid droplet and peroxisome-peroxisome contact sites, respectively. We asked if the formation of peroxisome-peroxisome contact sites and peroxisome-lipid droplet contact sites occurred independently, or whether the clustering of the two types of organelles was interlinked. To test this, Fig. 1. Overexpression of Pex3 causes the formation of a single peroxisomal structure surrounded by lipid droplets. (A–B) Overexpression of Pex3 produces the collapse of all peroxisomal signal into one structure. Panel A shows representative pictures of a strain expressing mCherry-SKL construct to visualize the lumen of the peroxisomes, either with Pex3 at endogenous levels (Control) or overexpressed (TEF1pr-PEX3) and the vacuolar lumen stained with CMAC. Cell outlines are shown as white dashed lines. Scale bar: 2μm. Panel B shows the quantication of the amount of peroxisomal structures per cell. ree independent experiments were performed and 30 cells were analyzed for each experiment and condition. Small diamonds correspond to individual cells, bigger diamonds correspond to the average of independent experiments. e dierent strains were compared using an unpaired two-tailed Student’s t-test. *** P < 0.001. (C) Turbo ID of overexpressed Pex3-TurboID enriches peroxisomal and lipid droplets proteins. Volcano plot showing relative protein intensity in a pull-down of biotinylated proteins between a strain overexpressing Pex3 tagged c-terminally with the TurboID protein (TEF1pr-PEX3-TID) and a wild type control strain (wt). Peroxisomal proteins are marked in magenta and lipid droplets proteins in green. GO term enrichment analysis of the group of proteins signicantly enriched in the Pex3-TID pull-down showed an enrichment of the GO Terms “peroxisomes” and “lipid droplets”. (D) e formed peroxisomal structure is surrounded by lipid droplets. Representative pictures of a strain expressing mCherry-SKL construct to visualize the lumen of the peroxisomes, either with Pex3 at endogenous levels (Control) or overexpressed (TEF1pr-PEX3), the vacuolar lumen stained with CMAC and lipid droplets stained with Bodipy. Scale bar: 2μm. e zoomed in region shows a peroxisomal structure surrounded by lipid droplets with a scale bar of 1μm. Cell outlines are shown as white dashed lines. Scientic Reports | (2025) 15:24480 3 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ we overexpressed Pex3 in a strain that is devoid of lipid droplets because it lacks the synthases for triglycerides and sterol esters, namely Dga1, Lro1, Are1 and Are2 (from here on termed ΔLDs)26. Overexpression of Pex3 caused the accumulation of peroxisomal signal into a single structure irrespective of the absence of lipid droplets, showing that lipid droplets are not required for the formation of the peroxisome-peroxisome contact sites (Fig.2G and H). The cytosolic domain of Pex3 interacts with peroxisomes Pex3 is anchored to the peroxisomal membrane by a single transmembrane domain at its N-terminus, which is sucient to cause targeting to the peroxisomes, and contains a globular cytosolic C-terminal domain27–29 (Fig.3A). To address a possible direct involvement of Pex3 in contact site formation, we expressed the cytosolic domain (CD – amino acids 40–441) fused to GFP at its N-terminus and lacking the transmembrane region (Fig.3A). We observed that this construct localized at peroxisomes, marked by Pex14-HaloTag (Fig.3B). e construct did not decorate lipid droplets marked by Erg6-2xmKate2, and was only observed enriched in these structures when peroxisomal signal was also present (Fig.3B, see zoomed in organelles). Consistently, calculation of the Mander´s coecients M1 and M2 for GFP-Pex3(CD) with Pex14-HaloTag showed high values, while the coecients of overlap with the Erg6-2xmKate2 signal were much lower and comparable to the ones observed between Pex14-HaloTag and Erg6-2xmKate2 (Fig.3C). Scientic Reports | (2025) 15:24480 4 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ To test if the interaction of the cytosolic domain of Pex3 with peroxisomes is strong enough to enable organelle tethering, we articially directed this domain to mitochondria. is was achieved by tagging it with the uorescent protein mKate2 and an N-terminal Alfa tag30, In addition, the mitochondrial outer membrane receptor Tom70 was tagged with a nanobody that recognizes the Alfa tag30, so it should direct the Pex3(CD) to mitochondria (Fig.3D). Indeed, Alfa-mKate2-Pex3(CD) decorates the mitochondrial network under these conditions, as conrmed by co-localization with the mitochondrial marker Cit1-HaloTag (Fig.3E). We quantied the minimum distance of each individual peroxisome to the mitochondrial network, for cells either expressing or not expressing the Alfa-mKate2-Pex3(CD) construct. e expression of the construct caused a strong shi to smaller distances (Fig.3F). Figure3E shows a representative image of the experiment, and additional examples are shown in Supplemental Fig.2. is data indicates that the interaction of Pex3(CD) with peroxisomes is strong enough to induce organelle tethering. Based on these results we propose that the homotypic peroxisomal contact sites are formed by Pex3 being anchored to the peroxisomal membrane via its transmembrane domain, and additionally interacting with other peroxisomes through its cytosolic domain. In contrast, the peroxisome-lipid droplet contact site is likely not directly mediated by Pex3, and only indirectly induced by its overexpression. Known interactors of Pex3 are not involved in forming the peroxisome-lipid droplet cluster Pex3 is a multifunctional protein involved in dierent processes related to the life-cycle of peroxisomes, including the targeting of peroxisomal membrane proteins, pexophagy, and targeting of peroxisomes to the cortex, which strongly inuences peroxisome inheritance12–17,19,29. ese functions are mediated by the direct interaction of Pex3 with dierent binding partners (Fig.4A). Targeting of peroxisomal membrane proteins to the peroxisomal membrane involves its interaction with the cytosolic receptor Pex1914,17 while its role in autophagy is mediated by its interaction with the pexophagy receptor Atg3615. Finally, the tethering of peroxisomes to the cell cortex is mediated by Pex3 interacting with Inp112,20 (Fig.4A). Next, we tested if any of the known interactors of Pex3 is involved in the formation of the peroxisome-lipid droplet cluster. To do this, we overexpressed Pex3 in strains lacking either ATG36 or INP1. e representative microscopy images and quantications shown in Fig.4B–G show that the phenotype caused by Pex3 overexpression in these backgrounds does not dier from the control cells, indicating that neither Inp1 nor Atg36 are required for the formation of this structure. e fact that Atg36 is not required for the formation of this structure suggests that it is not an intermediary in the process of pexophagy. However, we decided to test if pexophagy is induced by Pex3 overexpression, since this has been reported in mammalian cell lines19 In yeast, GFP is resistant to vacuolar degradation, and thus the generation of a free GFP band has been used as a readout for vacuolar degradation of dierent proteins and organelles, including peroxisomes31–33. us, we compared the appearance of a free GFP band in cells overexpressing Pex3 with cells grown under known pexophagy-inducing conditions, namely growth in oleate medium followed by nitrogen starvation33. Figure4H and Supplemental Fig.3A, B and C show that in cells expressing Pex14-GFP, a free GFP band appears under pexophagy-inducing conditions (P.I.C), indicating vacuolar degradation of peroxisomes. In contrast, the lack of a free GFP band in the strain overexpressing Pex3 when grown to logarithmic phase in glucose (L.P.G.) shows that overexpression of this protein alone does not induce pexophagy. Interestingly, pexophagy can still be induced in this strain, suggesting that it is also not blocked by the formation of the peroxisomal clusters. To test the requirement of PEX19, we could not delete this gene, as this results in the absence of functional peroxisomes34. We thus used an alternative strategy, by addressing whether overexpression of a mutant version of Pex3 that does not interact with Pex19 still promotes the formation of the peroxisome-lipid droplet cluster. Fig. 2. Overexpression of Pex3 induces an accumulation of peroxisomes surrounded by LDs that contains PexPex and Pex-LD contact sites. (A–B) Enlarged peroxisomes and LDs reveal that the structures contain several maxima for peroxisome lumen signal, with peroxisomal membrane between them. Representative pictures of a strain with overexpressed Pex3 (TEF1pr-PEX3), expressing the BFP-SKL construct to visualize the lumen of the peroxisomes, Erg6-2xmKate2 marking the lipid droplet monolayer, and either Pex3 (A) or Pex14 (B) tagged with mNeonGreen as markers of the peroxisomal membrane. To produce enlarged peroxisomes and lipid droplets, the cells contain a deletion of PEX11, and were grown with oleate as the sole carbon source for 20hs. Cell outlines are shown as white dashed lines. Scale bars: 1μm. Each graph shows the signal of BFPSKL and the corresponding peroxisomal membrane protein over a line across the structure, depicted in the merged image. (C–F) On-section CLEM tomography conrms that the structure involves Pex-Pex and Pex-LD contact sites. Panel C shows a representative image of on-section CLEM done on a strain with overexpressed Pex3 (TEF1pr-PEX3) and tagged with 2xmKate2. Scale bar: 600nm. Panel D shows a tomography image of the same section overlayed with the 3D model recreated from the images, showing peroxisomes in magenta and lipid droplets in yellow. Scale bar: 50nm. Panels E and F show zoomed in regions of the tomogram as examples of the peroxisome-lipid droplet (E) and peroxisome-peroxisome (F) contact sites. Scale bars: 20nm. (G–H) LDs are not necessary for the formation of the cluster of peroxisomes. Panel G shows representative pictures of strains with overexpressed Pex3 (TEF1pr-PEX3) in control cells and in strains that cannot produce lipid droplets (ΔLDs), expressing Pex14-2xmKate2 to visualize the peroxisomes, lipid droplets were stained with Bodipy and the vacuolar lumen was stained with CMAC. Cell outlines are shown as white dashed lines. Scale bar: 2μm. Panel H shows the quantication of the amount of peroxisomal structures per cell. ree independent experiments were performed and 30 cells were analyzed for each experiment and condition. Small diamonds correspond to individual cells, bigger diamonds correspond to the average of independent experiments. e dierent strains were compared using an unpaired two-tailed Student’s t-test. n.s., not signicant. ◂ Scientic Reports | (2025) 15:24480 5 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ e crystal structure of human Pex3 in complex with a fragment of human Pex19 has previously been solved. It was found that the interaction with Pex19 is mediated by a region centered around HsPex3-Trp104, which also includes Leu10716,39. Alignment of the AlphaFold-generated structure prediction of ScPex3 with the structure of HsPex3 indicated that this region is highly conserved and that the equivalent residues in ScPex3 are Trp128 and Leu131 (Fig.4I). We thus introduced the mutations Trp128Lys and Leu131Lys in ScPex3 (Pex3Mut) and tested the ability of this mutant to interact with Pex19, support peroxisome biogenesis, and form the peroxisomelipid droplet structures upon overexpression. Unlike wtPex3, Pex3Mut was not co-puried with msGFP2-Pex19, indicating that the mutations disrupt this interaction (Fig.4J and Supplemental Fig.3 D, E and F). e Pex19Pex3 interaction has been reported to be required to import PTS1 containing peroxisomal matrix proteins14. Consistently, we observed that expression of Pex3Mut in a strain lacking endogenous Pex3 does not rescue the import of BFP-SKL into peroxisomes, conrming that this mutation disrupts the interaction (Fig.4K). Figure4L– N contain representative microscopy images and the corresponding quantication showing that overexpression of this mutant in a background containing endogenous levels of wtPex3 to have functional peroxisomes, causes the same morphological phenotype as the overexpression of wtPex3. us, the interaction with Pex19 is not required for Pex3 to induce clustering of peroxisomes and lipid droplets. The triacylglycerol lipaseTgl4 is involved in the formation of the peroxisome-lipid droplet contact sites We sought to identify additional molecular players involved in the formation of contact sites upon Pex3 overexpression. Since many contact site tether proteins are involved in the formation of more than one contact site35–41, we tested the involvement of known tether proteins of lipid droplets or peroxisomes with other organelles. Scientic Reports | (2025) 15:24480 6 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ Deletion of the peroxisomal-mitochondrial tethers Pex34 or Fzo11 did not aect formation of Pex3-dependent peroxisome and lipid droplet clusters (Supplemental Fig.4A), indicating that they are not required for formation of the contact sites. We also tested the involvement of the the splicing-generated pair of proteins Ldo16-Ldo45, which tether lipid droplets to the vacuole42,43. Deletion of these genes did not aect the clustering of peroxisomes or lipid droplets, nor the proximity of the clustered structure to the vacuole (Supplemental Fig.4B). Finally, it was recently shown that the human protein M1 Spastin tethers lipid droplets to peroxisomes44. We deleted the yeast homolog Sap1, but observed no alteration of the Pex3-dependent phenotype (Supplemental Fig.4C). Peroxisomes and lipid droplets are also linked by the role of the endoplasmic reticulum during their biogenesis. Both lipid droplets and pre-peroxisomal vesicles bud from a subdomain of the endoplasmic reticulum marked by the reticulon-like protein Pex3045. It was reported that the same domain can bind simultaneously to a lipid droplet and a peroxisome, causing a close association between them46. Deletion of PEX30 did not aect the formation of the peroxisomal and lipid droplet clusters (Supplemental Fig.4D), indicating that this protein is not directly involved in forming them. We next performed a genome-wide microscopy-based screen, to identify factors aecting the formation of this cluster. We crossed a strain carrying TEF1pr-Pex3 and the peroxisomal marker Pex14-mKate2 with a genome-wide collection of deletion mutants of non-essential genes47 and hypomorphic DAmP allele mutants of essential genes48 using an automated mating and sporulation procedure49,50. e resulting mutant collection contains TEF1pr-Pex3, Pex14-2xmKate2 and each gene deleted or depleted (Fig.5A). We analyzed this collection by automated microscopy aer labeling cells with Bodipy and CMAC, to stain lipid droplets and vacuoles respectively. e resulting images were analyzed manually searching for strains in which the phenotype was disrupted. We found a single hit in which the phenotype was signicantly disrupted, which carried the deletions of the gene encoding for the lipase Tgl451. is strain was manually re-constructed to conrm that the disruption of the phenotype did not depend on the genetic background used for the screen and re-analyzed by microscopy (Fig.5B). In cells that overexpressed Pex3, 95% of the peroxisomal structures were proximal to lipid droplets, while this number dropped to 51% in cells that in addition lacked Tgl4 (Fig.5C). Re-insertion of the TGL4 ORF with its endogenous promoter in a plasmid fully recovered the interaction between peroxisomes and lipid droplets, indicating that the eect is specic to the lack of the gene, and not a secondary eect of the genomic modication (Supplemental Fig.5A and B). We tested the specicity of the phenotype by deleting other lipid droplet-localized TAG lipases. Neither deletion of the genes encoding for the lipases homologous to Tgl4, namely Tgl3 and Tgl552,53 (Fig.5B and C) nor other ones, Tgl1, Ldh1 or Yeh154–56 (Supplemental Fig.5C and D) caused a disruption of the phenotype. e homologous lipases also did not cause a further disruption when combined with the deletion of TGL4 (Supplemental Fig.5E and F). is suggests that it is the physical presence of the protein Tgl4 that aects the formation of the contact site and not its activity as a lipase. To test this hypothesis, we assessed the eect of mutation S315G, which disrupts its active site57, and observed no reduction in the association of lipid droplets to the peroxisomal cluster (Fig.5D and E). In addition, we analyzed the localization of Tgl3, 4, and 5 on the lipid droplet surface upon induction of these contact sites. Again, we used the strain lacking PEX11 and grew the cells in the presence of oleate to increase the size of peroxisomes and lipid droplets and gain spatial resolution. We observed that all three lipases Fig. 3. e cytosolic domain of Pex3 binds peroxisomes and is able to tether them to another organelle. (A–C) e cytosolic domain of Pex3 binds to peroxisomes. Panel A shows a diagram of the GFP-Pex3(CD) construct. is construct contains the cytosolic domain of Pex3 (aa40-441) fused to GFP tag at the N-terminus instead of the transmembrane domain as in the full length Pex3. Panel B shows representative pictures of the colocalization experiment of GFP-Pex3(CD) with peroxisomes (Pex14-HaloTag) and lipid droplets (Erg62xmKate2). Cell outlines are shown as white dashed lines. Scale bar: 2μm and 0.8μm. Panel C shows the co-localization analysis of the experiment in B using Mander´s coecients M1 and M2 for the overlap of: Pex14-HaloTag and GFP-Pex3(CD) (cyan diamonds), Erg6-2xmKate2 and GFP-Pex3(CD) (green diamonds) or Pex14-HaloTag and Erg6-2xmKate2 (black diamonds). ree independent experiments were performed and 30 cells were analyzed for each experiment. Each small diamond represents a single cell, and the bigger ones represent the average of each of three independent experiments. (D) Diagram of the strategy used to recruit the cytosolic domain of Pex3 to mitochondria articially. e outer mitochondrial membrane protein Tom70 was tagged in the C-terminus with a Nanobody that recognizes the AlfaTag (Tom70-NB-Alfa). e cytosolic domain of Pex3 was tagged with an AlfaTag and an mKate2 uorescent protein in the N-terminus (AlfaTagmKate2-Pex3(CD) construct). is causes the recruitment of AlfaTag-mKate2-Pex3(CD) to mitochondria. (E–F) Targeting the cytosolic domain of Pex3 to mitochondria tethers peroxisomes to this organelle. Panel E shows representative images of the localization of peroxisomes (GFP-SKL) and mitochondria (Cit1-HaloTag) in the presence or absence of AlfaTag-Pex3(CD), with Tom70 fused to Alfa Nanobody in the background. A maximum intensity projection of the Z-stacks is shown for each image. Cell outlines are shown as white dashed lines. Scale bar: 2μm. Panel F shows the measurements of distances between peroxisomes (GFPSKL) and mitochondria (Cit1-HaloTag) in the presence or absence of the AlfaTag-Pex3(CD) construct as described. ree independent experiments were performed and 30 cells were analyzed for each experiment. Each small diamond represents a single peroxisome, and the bigger ones represent the average of each of three independent experiments. e distribution of distances to mitochondria between the two strains was compared using a Kolmogorov-Smirno test, **** P < 0.0001. Comparison of the means of each experiment with a two-tailed unpaired Student´s t-test results in a P value < 0.05. ◂ Scientic Reports | (2025) 15:24480 7 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ were enriched in the region of the lipid droplet that was in contact with the peroxisomes in some cells, as can be appreciated in the example images and line proles (Fig.5F and G). ese enrichments, however, were not equally frequent for all lipases: Tgl4 and Tgl5 were enriched more frequently than Tgl3. is can be observed by the resulting peaks formed by averaging many cells in the line proles (Fig.5G). To explore the requirements for the formation of Tgl4 foci at lipid droplet-peroxisome interfaces, we performed a microscopy-based screen (Supplemental Fig.6A). A PEX3 overexpression allele, genes for lipid droplet and peroxisome visualization (Erg6-mCh and BFP-SKL), and Tgl4-GFP were introduced into the genome-wide deletion and DAmP libraries47,48 by an automated mating approach49,50. Cells were cultured in the presence of oleate to expand lipid droplets and analyzed by automated microscopy. We identied a total of 86 mutants in which the accumulation of Tgl4-GFP foci at lipid droplet-peroxisome interfaces was fully or partially blocked (Supplemental Table 5, example microscopy images in Supplemental Fig.6B). We analyzed the common functions among the genes identied by the screen (Supplemental Fig.6C). e biggest group of genes was related to energy metabolism. is is also evidenced by the enrichment of the GO Terms “mitochondrion organization”, “mitochondrial respiratory chain complex assembly” and “mitochondrion” in our hit list with adjusted p-values of 0.0009, 0.001, and 0.008, respectively. Other groups of hits corresponded to hypoxia signaling, autophagy, and lipid homeostasis. Taken together, this suggests that the presence of Tgl4 at this organelle interface is regulated by the metabolic state of the cell. Additionally, we identied eight genes from the membrane contact site database, representing a threefold enrichment to the expected amount given the fraction of the genome annotated, likely reecting the tight interrelations within the cellular contact site network. Scientic Reports | (2025) 15:24480 8 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ The phenotype of overexpression of Pex3 is conserved in metazoans Since Pex3 is a conserved protein, and many of its functions, like the incorporation of peroxisomal membrane proteins and the role in pexophagy are conserved in metazoans, we decided to test if the phenotype of induction of contact sites is also conserved. us, we overexpressed Pex3-HA in Drosophila melanogaster larvae in the midgut using the GAL/UAS system58, and we observed by uorescence microscopy cells from the midgut, comparing them to cells expressing only GFP-SKL, to observe peroxisomes when Pex3 is not overexpressed (Fig.6A and B). In the larval midgut, we observed that overexpression of Pex3 caused a reduction of the number of peroxisomes, as well as an increase in their size, whereas the number of lipid droplets was unaected (Fig.6B– D). is phenotype was exacerbated during development, with a drastic increase in peroxisomal and lipid droplet size in the adult midgut (Fig.6E–H). We expressed Pex3-HA with another peroxisomal marker, YFPSKL. Using Airyscan confocal microscopy, we found that Pex3-HA clearly labeled the surface of the enlarged peroxisome, while YFP was imported into the peroxisomal lumen by the peroxisomal targeting sequence, as expected (Fig.6I). Furthermore, these enlarged peroxisomes were closely associated with lipid droplets as can be observed by the shape deformation of the organelles when they are next to each other. us, the phenotype closely resembles the one observed in yeast. Discussion In this work, we have described that Pex3 overexpression induces the formation of peroxisome-peroxisome and peroxisome-lipid droplet contact sites. If we make the approximation that Pex3 expressed under the control of the TEF1 promoter would have the levels of the Tef1 protein, and based on the integrated high throughput data in the protein abundance database PaxDB 5.059, the overexpression system that we used would generate approximately a 200-fold increase in protein levels. We do not think of this as mimicking a physiological scenario, but rather as a tool for discovery, which has in the past proven useful to exacerbate one of the functions of a multifunctional protein over the others. For example, for the proteins Vps39 and Cvm1, overexpression allowed identication of their role as tethers of the vacuole-mitochondria contact site60,61. is role could later be conrmed at endogenous levels of the proteins either by addressing the deletion or separation-of-function mutants61,62. Our microscopy-based screens illustrate how useful this tool can be, since it already allowed the Fig. 4. Formation of the structure is independent of known interactors of Pex3. (A) Diagram of Pex3 known interactors and their functions. (B, E) Representative images of strains overexpressing Pex3 (TEF1pr-PEX3) in control cells and strains lacking Inp1 (B) or Atg36 (E). All strains express Pex14 fused to 2xmKate2 to visualize the peroxisomes, lipid droplets were stained with Bodipy and the vacuolar lumen was stained with CMAC. Cell outlines are shown as white dashed lines. Scale bars: 2 μm. (C, F) Quantication of the number of peroxisomal structures per cell in the microscopy experiments described before. Small diamonds correspond to individual cells, bigger diamonds correspond to the average of independent experiments. ree independent experiments were performed and 30 cells were analyzed for each experiment and condition. e dierent strains were compared using an unpaired two-tailed Student’s t-test. n.s., not signicant. (D, G) Quantication of the fraction of peroxisomal structures with accumulations of lipid droplets next to them in the microscopy experiments described above. ree independent experiments were performed and 30 cells were analyzed for each experiment and condition. e dierent strains were compared using an unpaired two-tailed Student’s t-test. n.s., not signicant. (H) Pex3 overexpression does not induce pexophagy. Whole cell lysates of strains expressing Pex14-GFP with or without Pex3 overexpression were analyzed by Western blot. e cells were either grown in media containing glucose to logarithmic phase (L.P.G) or grown in media containing oleate and shied to nitrogen starvation medium to induce pexophagy (P.I.C). e presence of a free GFP band is indicative of vacuolar degradation of peroxisomes. e whole Western blot membrane as well as a loading control is shown in Supplemental Fig.3 A. (I) e amino acids involved in hsPex3 interaction with hsPex19 are conserved in yeast. To the le, the structure obtained for Homo sapiens Pex3 (cyan) interacting with a peptide of hsPex19 (dark blue)17. Amino acids W104 and L107 of hsPex3 are involved in the interaction with hsPex19. To the right, the structure predicted by AlphaFold for Saccharomyces cerevisiae Pex3 (light blue) shows that it contains a structurally conserved W and L in the same positions (W128 and L131 in scPex3). (J) Pex3(W128K, L131K) cannot interact with Pex19. Anity purication of msGFP2-Pex19 co-puries Pex3-mKate2-AlfaTag but not Pex3(W128K, L131K)-mKate2-AlfaTag. e complete Western blot membranes are shown in Supplemental Fig.3 B and C. (K) Pex3(W128K, L131K) does not support BFP-SKL import into peroxisomes. In a strain that expresses BFP-SKL, endogenous Pex3 was deleted and either Pex3wt or Pex3(W128K, L128K) were re-introduced in a plasmid. Quantication of the number of BFP-SKL puncta per cell is shown to the right. (L–N) Overexpression of Pex3(W128K, L131K), which cannot interact with Pex19, still causes aggregation of peroxisomes and recruitment of lipid droplets. Panel K shows representative images of strains overexpressing Pex3 from a plasmid (TEF1pr-PEX3-mKate2) containing either Pex3 wt or Pex3 mutant. Lipid droplets were stained with Bodipy and the vacuolar lumen was stained with CMAC. e strain contains Pex3 wt in the background, expressed from its genomic locus, in order to have normal peroxisomes. Cell outlines are shown as white dashed lines. Scale bars: 2μm. Panel L shows the quantication of the number of peroxisomal structures per cell in the microscopy experiments described before. Small diamonds correspond to individual cells, bigger diamonds correspond to the average of independent experiments. Panel M shows the quantication of the proportion of peroxisomal structures with accumulations of lipid droplets next to them in the microscopy experiments described above. ree independent experiments were performed and 30 cells were analyzed for each experiment and condition. e dierent strains were compared using an unpaired twotailed Student’s t-test. n.s., not signicant. ◂ Scientic Reports | (2025) 15:24480 9 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ Detection of pexophagy via whole cell lysate and Western blot All cells were grown overnight in SDC media as precultures. e logarithmic phase glucose (L.P.G.) samples, were diluted in SDC media and grown for 20hs without exceeding an OD600 = 0.3 and harvested. e pexophagyinducing conditions samples (P.I.C) were diluted in synthetic media containing oleate as the carbon source (0.2% Oleate, 0.1% Tween-80) and grown for 20hs, then shied to nitrogen starvation medium (without nitrogen source or amino acids) for 22hs and harvested. Whole cell lysates were generated by mechanical disruption using glass beads in lysis buer (3M urea, 1.875mM EDTA-KOH p.H. 8.0, 1.2% (w/v) SDS, 37.5mM Tris–HCl p.H. 6.8, 1.5% (v/v) glycerol. Bromophenol blue was added to the samples to a nal concentration of 0.005% and 2-mercaptoethanol toanal concentration of 0.5%. Equivalent amounts of each sample were subjected to SDSPAGE, transferred to a nitrocellulose membrane and processed for Western blot as described in the previous section. e secondary antibody was coupled to horse-radish peroxidase, and the signal was detected using Sigma Lumi-Light Plus Western Blotting Substrate in an Azure 600 imaging System. e whole uncropped western blot membranes is shown in Supplemental Fig.3A, and additional repetitions in Supplemental Fig.3B and C. In the two other repetitions, the pre-cultures were included in the experiment (grown in glucose for the L.P.G condition and in oleate for the P.I.C. condition). Fly husbandry Flies were reared on standard cornmeal food (130g yarn agar, 248g Baker’s yeast, 1223g Cornmeal and 1.5l sugar beet syrup in 20l distilled water) and kept in a 25°C incubator with light–dark-cycle. Fly lines used in this study were mex-Gal4 (kindly provided by the lab of Irene Miguel-Aliaga), UAS-Pex3-GFP, UAS-Pex3HA (kindly provided by the lab of Reinhard Bauer) and UAS-GFP-SKL (Bloomington Drosophila stock center #28882). For studies in the larval gut, the following genotypes were used: w; mex-Gal4; UAS-GFP-SKL and w; mex-Gal4; UAS-GFP-SKL/UAS-Pex3-HA. For studies in the adult gut, the following genotypes were used: w; mex-Gal4; UAS-GFP-SKL, w; mex-Gal4; UAS-GFP-SKL/UAS-Pex3-HA and w; mex-Gal4; UAS-YFP-PTS1/ UAS-Pex3-HA. Animals were reared on standard diet and analyzed as 3rd instar larvae or 5day old adults (male and female), respectively. Imaging of Drosophila guts Antibodies used in this study were α-GFP (Santa Cruz Biotechnology) and α-HA (Invitrogen), α-TOMM20 (Sigma-Aldrich). For immunohistochemistry, guts from 3rd instar larvae or adult ies were dissected in PBS and xed for 1h in 0.5% PBS-Tween20 and 4% formaldehyde. Tissue was washed with 0.5% PBS-Tween20 and blocked with donkey serum before incubation with the primary antibody (overnight at 4°C). e tissue was washed in 0.1% PBS-Tween20 before incubation with BODIPY 581/591 (ermo Scientic) and the secondary antibody at room temperature in the dark for 1h. Secondary antibodies coupled to Alexa or Cyanine dyes were from Molecular probes. e tissue was washed and incubated for 5min with DAPI (4',6-diamidino-2phenylindole). For imaging, we used a Zeiss LSM 710 with a 25 × water lens (Plan-Neouar, Zeiss), 40 × water lens (C-Apochromat, Zeiss), and 63 × water lens (Plan-Apochromat, Zeiss) and a Zeiss LSM 880 with Airyscan detector. We used ImageJ to quantify peroxisome and lipid droplet number and area from at least 3 individual cells from dierent experiments. Library generation and high-throughput microscopy All yeast manipulations were performed in high-density format (384–1,536 strains per plate) using a RoToR bench-top colony array instrument (Singer Instruments). In order to nd key proteins involved in the formation of the peroxisome-peroxisome or peroxisome-lipid droplet contact sites (Fig.5A), strain AGMY1303, with an overexpression of Pex3 (TEF1pr-Pex3) and a peroxisomal marker (Pex14-2xmKate2) was crossed with a genome-wide library of deletion47 and hypomorphic allele48 strains, by the synthetic genetic array method49,50. For analysis of Tgl4 localization (Supplemental Fig.6), strain yMB1326 (TEF2pr-Pex3 Erg6-mCh BFP-SKL Tgl4-GFP) was crossed with the same mutant collections. Cells were mated on rich medium plates, diploids were selected and sporulation was induced by incubating the cells for ve to eight days on nitrogen starvation media plates. Haploid cells were selected on 50mg/L Canavanine and 50mg/L ialysine. Finally, haploid cells containing the combination of all desired manipulations were selected. A subset of strains was veried by microscopy and conrmed by PCR. For the automated imaging of the obtained libraries, cells were rst transferred from agar plates into 384 well plates for growth in liquid medium. For the screen for genes involved in formation of lipid droplet-peroxisome contacts (Fig.5A), cultures were grown overnight at 30°C in SDC. A JANUS liquid handler (PerkinElmer) connected to the incubator was used to dilute the strains to an OD600 of ∼0.2, and plates were incubated at 30°C for 4h. Cells were washed and fresh SDC media was added containing 20µM 7-amino-4-chloromethylcoumarin (CMAC) dye and 1µg/ml of Bodipy dye, followed by half an hour incubation and another wash step. For the Tgl4-GFP screen (Supplemental Fig.6), liquid handling was performed using a MicroPro 300 liquid handler. Following the same overnight culture as the previous screen, cells were grown in the presence of 0.2% oleate in synthetic medium for 24h for lipid droplet enlargement. Strains were then transferred by a liquid handler into glass-bottom 384-well microscope plates (Matrical Bioscience) coated with concanavalin A (Sigma-Aldrich) and incubated for 20min to allow adhesion of cells to the bottom of the plates. Aerwards, wells were washed twice with SDC medium (SC medium for oleate treated cells) to remove non-adherent cells leaving a cell monolayer. Plates were then transferred to an Olympus automated inverted uorescence microscope system. In the screen for lipid droplet-peroxisome contact site proteins (Fig.5A), cells were imaged in SDC at 18–20°C using a 60 × air lens (NA 0.9) and with an ORCAER charge-coupled device camera (Hamamatsu), using ScanR soware. In the Tgl4-GFP screen (Supplemental Fig.6), cells were imaged in SC medium at 18–20°C using a 60 × air lens (NA 0.9) and with an ORCA-ash4.0 Scientic Reports | (2025) 15:24480 16 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ camera (Hamamatsu), using ScanR soware. Aer acquisition, images were manually reviewed using ImageJ (National Institutes of Health). Data availability e proteomics dataset generated in this study is available via ProteomeXchange with identier PXD063514. Other datasets generated during the current study are available from the corresponding author on reasonable request. Received: 13 December 2024; Accepted: 18 June 2025 References 1. Wanders, R. J. A., Baes, M., Ribeiro, D., Ferdinandusse, S. & Waterham, H. R. e physiological functions of human peroxisomes. Physiol. Rev. 103, 957–1024 (2023). 2. Scorrano, L. et al. Coming together to dene membrane contact sites. Nat. Commun. 10, 1–11 (2019). 3. Eisenberg-Bord, M., Shai, N., Schuldiner, M. & Bohnert, M. A tether is a tether is a tether: Tethering at membrane contact sites. Dev. Cell 39, 395–409 (2016). 4. Prinz, W. A., Toulmay, A. & Balla, T. e functional universe of membrane contact sites. Nat. Rev. Mol. Cell Biol. 21, 7–24 (2020). 5. Valm, A. M. et al. Applying systems-level spectral imaging and analysis to reveal the organelle interactome. Nature 546, 162–167 (2017). 6. Shai, N. et al. Systematic mapping of contact sites reveals tethers and a function for the peroxisome-mitochondria contact. Nat. Commun. 9 (2018). 7. Kakimoto, Y. et al. Visualizing multiple inter-organelle contact sites using the organelle-Targeted split-GFP system. Sci. Rep. 8 (2018). 8. Voeltz, G. K., Sawyer, E. M., Hajnóczky, G. & Prinz, W. A. Making the connection: How membrane contact sites have changed our view of organelle biology. Cell 187, 257–270 (2024). 9. Shai, N., Schuldiner, M. & Zalckvar, E. No peroxisome is an island—Peroxisome contact sites. Biochim. Biophys. Acta Mol. Cell Res. 1863, 1061–1069 (2016). 10. David, C. et al. A combined approach of quantitative interaction proteomics and live-cell imaging reveals a regulatory role for endoplasmic reticulum (ER) Reticulon Homology Proteins in peroxisome biogenesis. Mol. Cell. Proteomics 12, 2408–2425 (2013). 11. Yan, M., Rachubinski, D. A., Joshi, S., Rachubinski, R. A. & Subramani, S. Dysferlin domain-containing proteins, Pex30p and Pex31p, localized to two compartments, control the number and size of oleate-induced peroxisomes in pichia pastoris. Mol. Biol. Cell 19, 885–898 (2008). 12. Knoblach, B. et al. An ER-peroxisome tether exerts peroxisome population control in yeast. EMBO J. 32, 2439–2453 (2013). 13. Burnett, S. F., Farré, J. C., Nazarko, T. Y. & Subramani, S. Peroxisomal Pex3 activates selective autophagy of peroxisomes via interaction with the pexophagy receptor Atg30. J. Biol. Chem. 290, 8623–8631 (2015). 14. Fang, Y., Morrell, J. C., Jones, J. M. & Gould, S. J. PEX3 functions as a PEX19 docking factor in the import of class I peroxisomal membrane proteins. J. Cell Biol. 164, 863–875 (2004). 15. Motley, A. M., Nuttall, J. M. & Hettema, E. H. Pex3-anchored Atg36 tags peroxisomes for degradation in Saccharomyces cerevisiae. EMBO J. 31, 2852–2868 (2012). 16. Sato, Y. et al. Characterization of the interaction between recombinant human peroxin Pex3p and Pex19p: Identication of TRP104 in Pex3p as a critical residue for the interaction. J. Biol. Chem. 283, 6136–6144 (2008). 17. Sato, Y. et al. Structural basis for docking of peroxisomal membrane protein carrier Pex19p onto its receptor Pex3p. EMBO J. 29, 4083–4093 (2010). 18. Schmidt, F. et al. e role of conserved PEX3 regions in PEX19-binding and peroxisome biogenesis. Traffic 13, 1244–1260 (2012). 19. Yamashita, S. I., Abe, K., Tatemichi, Y. & Fujiki, Y. e membrane peroxin PEX3 induces peroxisome-ubiquitination-linked pexophagy. Autophagy 10, 1549–1564 (2014). 20. Hulmes, G. E. et al. e Pex3-Inp1 complex tethers yeast peroxisomes to the plasma membrane. J. Cell Biol. 219 (2020). 21. Wu, H. et al. Peroxisome development in yeast is associated with the formation of Pex3-dependent peroxisome-vacuole contact sites. Biochim. Biophys. Acta Mol. Cell Res. 1866, 349–359 (2019). 22. Fernández-Suárez, M., Scott Chen, T. & Ting, A. Y. Protein-protein interaction detection in vitro and in cells by proximity biotinylation. NIH Public Access 23, 1–7 (2008). 23. Branon, T. C. et al. Ecient proximity labeling in living cells and organisms with TurboID. Nat. Biotechnol. 176, 139–148 (2018). 24. Erdmann, R. & Blobel, G. Giant peroxisomes in oleic acid-induced Saccharomyces cerevisiae lacking the peroxisomal membrane protein Pmp27p. J. Cell Biol. 128, 509–523 (1995). 25. Yifrach, E. et al. Systematic multi-level analysis of an organelle proteome reveals new peroxisomal functions. Mol. Syst. Biol. 18, 1–21 (2022). 26. Sandager, L. et al. Storage lipid synthesis is non-essential in yeast. J. Biol. Chem. 277, 6478–6482 (2002). 27. Höhfeld, J., Veenhuis, M. & Kunau, W. H. PAS3, a Saccharomyces cerevisiae gene encoding a peroxisomal integral membrane protein essential for peroxisome biogenesis. J. Cell Biol. 114, 1167–1178 (1991). 28. Soukupovaa, M., Sprenger, C., Gorgas, K., Kunau, W.-H. & Dodtl, G. Identication and Characterization of the Human Peroxin PEX3. Eur. J. Cell Biol. 78. http://www.urbanscher.de/journals/ejcb (1999). 29. Schmidt, F. et al. Insights into peroxisome function from the structure of PEX3 in complex with a soluble fragment of PEX19. J. Biol. Chem. 285, 25410–25417 (2010). 30. Götzke, H. et al. e ALFA-tag is a highly versatile tool for nanobody-based bioscience applications. Nat. Commun. 10, 1–12 (2019). 31. Torggler, R., Papinski, D. & Kra, C. Assays to monitor autophagy in Saccharomyces cerevisiae. Cells h t t p s : / / d o i . o r g / 1 0 .3 3 9 0 / c e l l s 6 0 3 0 0 2 3 (2017). 32. Kolitsida, P. et al. e pyruvate dehydrogenase complex regulates mitophagic tracking and protein phosphorylation. Life Sci. Alliance 9 (2023). 33. Yu, J. et al. Autophagy-related protein UvAtg7 contributes to mycelial growth, virulence, asexual reproduction and cell stress response in rice false smut fungus Ustilaginoidea virens. Fungal Genetics Biol. 159 (2022). 34. Hettema, E. H., Girzalsky, W., Van Den Berg, M., Erdmann, R. & Distel, B. Saccharomyces cerevisiae Pex3p and Pex19p are required for proper localization and stability of peroxisomal membrane proteins. EMBO J. 19, 223–233 (2000). 35. Levine, T. P. & Munro, S. Dual targeting of Osh1p, a yeast homologue of oxysterol-binding protein, to both the Golgi and the nucleus-vacuole junction. Mol. Biol. Cell 12, 1633–1644 (2001). 36. Loewen, C. J. R., Roy, A. & Levine, T. P. A conserved ER targeting motif in three families of lipid binding proteins and in Opi1p binds VAP. EMBO J. 22, 2025–2035 (2003). Scientic Reports | (2025) 15:24480 17 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ 37. Kvam, E. & Goldfarb, D. S. Nvj1p is the outer-nuclear-membrane receptor for oxysterol-binding protein homolog Osh1p in Saccharomyces cerevisiae. J. Cell Sci. 117, 4959–4968 (2004). 38. Elbaz-Alon, Y. et al. Lam6 regulates the extent of contacts between organelles. Cell Rep. 12, 7–14 (2015). 39. Murley, A. et al. Ltc1 is an ER-localized sterol transporter and a component of ER-mitochondria and ER-vacuole contacts. J. Cell Biol. 209, 539–548 (2015). 40. Toulmay, A. & Prinz, W. A. A conserved membrane-binding domain targets proteins to organelle contact sites. J. Cell Sci. 125, 49–58 (2012). 41. Liu, L. K., Choudhary, V., Toulmay, A. & Prinz, W. A. An inducible ER-Golgi tether facilitates ceramide transport to alleviate lipotoxicity. J. Cell Biol. 216, 131–147 (2017). 42. Diep, D. T. V. et al. A metabolically controlled contact site between vacuoles and lipid droplets in yeast. Dev. Cell 59, 740-758.e10 (2024). 43. Álvarez-Guerra, I. et al. LDO proteins and Vac8 form a vacuole-lipid droplet contact site to enable starvation-induced lipophagy in yeast. Dev. Cell 59, 759-775.e5 (2024). 44. Chang, C. L. et al. Spastin tethers lipid droplets to peroxisomes and directs fatty acid tracking through ESCRT-III. J. Cell Biol. 218, 2583–2599 (2019). 45. Joshi, A. S. et al. A family of membrane-shaping proteins at ER subdomains regulates pre-peroxisomal vesicle biogenesis. J. Cell Biol. 215, 515–529 (2016). 46. Joshi, A. S. et al. Lipid droplet and peroxisome biogenesis occur at the same ER subdomains. Nat. Commun. 9, 1–12 (2018). 47. Giaever, G. et al. Functional Profiling of the Saccharomyces Cerevisiae Genome. http://www.kegg.com (2002). 48. Breslow, D. K. et al. A Comprehensive Strategy Enabling High-Resolution Functional Analysis of the Yeast Genome. 49. Cohen, Y. & Schuldiner, M. Advanced methods for high-throughput microscopy screening of genetically modied yeast libraries. In Methods in Molecular Biology Vol. 781 127–159 (Humana Press Inc., 2011). 50. Hin, A., Tong, Y. & Boone, C. Synthetic genetic array analysis in Saccharomyces Cerevisiae. From: Methods in Molecular Biology vol. 313 (2006). 51. Rajakumari, S., Rajasekharan, R. & Daum, G. Triacylglycerol lipolysis is linked to sphingolipid and phospholipid metabolism of the yeast Saccharomyces cerevisiae. Biochim. Biophys. Acta Mol. Cell Biol. Lipids 1801, 1314–1322 (2010). 52. Athenstaedt, K. & Daum, G. Tgl4p and Tgl5p, two triacylglycerol lipases of the yeast Saccharomyces cerevisiae are localized to lipid particles. J. Biol. Chem. 280, 37301–37309 (2005). 53. Athenstaedt, K. & Daum, G. YMR313c/TGL3 encodes a novel triacylglycerol lipase located in lipid particles of Saccharomyces cerevisiae. J. Biol. Chem. 278, 23317–23323 (2003). 54. Jandrositz, A. et al. e lipid droplet enzyme Tgl1p hydrolyzes both steryl esters and triglycerides in the yeast, Saccharomyces cerevisiae. Biochim. Biophys. Acta Mol. Cell Biol. Lipids 1735, 50–58 (2005). 55. oms, S., Debelyy, M. O., Connerth, M., Daum, G. & Erdmann, R. e putative Saccharomyces cerevisiae hydrolase Ldh1p is localized to lipid droplets. Eukaryot. Cell 10, 770–775 (2011). 56. Köel, R., Tiwari, R., Falquet, L. & Schneiter, R. e Saccharomyces cerevisiae YLL012/YEH1, YLR020/YEH2, and TGL1 Genes Encode a Novel Family of Membrane-Anchored Lipases at Are Required for Steryl Ester Hydrolysis. Mol. Cell Biol. 25, 1655– 1668 (2005). 57. Kurat, C. F. et al. Obese yeast: Triglyceride lipolysis is functionally conserved from mammals to yeast. J. Biol. Chem. 281, 491–500 (2006). 58. Brand, H. & Perrimon, N. Targeted Gene Expression as a Means of Altering Cell Fates and Generating Phenotypes (1993). 59. Huang, Q., Szklarczyk, D., Wang, M., Simonovic, M. & Mering, C. von. PaxDb 5.0: Curated protein quantication data suggests adaptive proteome changes in yeasts. Mol. Cell. Proteom. 22 (2023). 60. Hönscher, C. et al. Cellular metabolism regulates contact sites between vacuoles and mitochondria. Dev. Cell 30, 86–94 (2014). 61. Bisinski, D. D. et al. Cvm1 is a component of multiple vacuolar contact sites required for sphingolipid homeostasis. J. Cell Biol. 221 (2022). 62. González Montoro, A. et al. Vps39 interacts with Tom40 to establish one of two functionally distinct vacuole-mitochondria contact sites. Dev. Cell 45, 621-636.e7 (2018). 63. Gurevich, V. V. & Gurevich, E. V. Analyzing the roles of multi-functional proteins in cells: e case of arrestins and GRKs. Crit. Rev. Biochem. Mol. Biol. 50, 440–452. https://doi.org/10.3109/10409238.2015.1067185 (2015). 64. Lubin, J. W., Rao, T., Mandel, E. K., Wuttke, D. S. & Lundblad, V. Dissecting protein function: An ecient protocol for identifying separation-of-function mutations that encode structurally stable proteins. Genetics 193, 715–725 (2013). 65. Schrader, M., King, S. J., Stroh, T. A. & Schroer, T. A. Real time imaging reveals a peroxisomal reticulum in living cells. 3671, 3663–3671 (2000). 66. Bonekamp, N. A., Sampaio, P., de Abreu, F. V., Lüers, G. H. & Schrader, M. Transient complex interactions of mammalian peroxisomes without exchange of matrix or membrane marker proteins. Traffic 13, 960–978 (2012). 67. Gorgas, K. & Zaar, K. Anatomy and embryology peroxisomes in sebaceous glands III. Morphological similarities of peroxisomes with smooth endoplasmic reticulum and Golgi stacks in the circumanal gland of the dog* **. Anat. Embryol. 169 (1984). 68. Mast, F. D., Rachubinski, R. A. & Aitchison, J. D. Peroxisome prognostications: Exploring the birth, life, and death of an organelle. J. Cell Biol. 219, 1–13 (2020). 69. Pan, X. et al. Nucleus-vacuole junctions in Saccharomyces cerevisiae are formed through the direct interaction of Vac8p with Nvj1p. Mol. Biol. Cell 11, 2445–2457 (2000). 70. Filadi, R. et al. TOM70 sustains cell bioenergetics by promoting IP3R3-mediated ER to mitochondria Ca2+ transfer. Curr. Biol. 28, 369-382.e6 (2018). 71. Hollenstein, D. M. et al. Vac8 spatially connes autophagosome formation at the vacuole in S. Cerevisiae. J. Cell Sci. 132 (2019). 72. Veit, M., Laage, R., Dietrich, L., Wang, L. & Ungermann, C. Vac8p release from the SNARE complex and its palmitoylation are coupled and essential for vacuole fusion. EMBO J. 20, 3145–3155 (2001). 73. Wang, Y. X., Catlett, N. L. & Weisman, L. S. Vac8p, a vacuolar protein with armadillo repeats, functions in both vacuole inheritance and protein targeting from the cytoplasm to vacuole. J. Cell Biol. 140, 1063–1074 (1998). 74. Schmidt, O., Pfanner, N. & Meisinger, C. Mitochondrial protein import: From proteomics to functional mechanisms. Nat. Rev. Mol. Cell Biol. 11, 655–667. https://doi.org/10.1038/nrm2959 (2010). 75. González Montoro, A. et al. Subunit exchange among endolysosomal tethering complexes is linked to contact site formation at the vacuole. Mol. Biol. Cell 32, br14 (2021). 76. Binns, D. et al. An intimate collaboration between peroxisomes and lipid bodies. J. Cell Biol. 173, 719–731 (2006). 77. Janke, C. et al. A versatile toolbox for PCR-based tagging of yeast genes: New uorescent proteins, more markers and promoter substitution cassettes. Yeast 21, 947–962 (2004). 78. Generoso, W. C., Gottardi, M., Oreb, M. & Boles, E. Simplied CRISPR-Cas genome editing for Saccharomyces cerevisiae. J. Microbiol. Methods 127, 203–205 (2016). 79. Schäfer, A., Kerssen, D., Veenhuis, M., Kunau, W.-H. & Schliebs, W. Functional Similarity between the Peroxisomal PTS2 Receptor Binding Protein Pex18p and the N-Terminal Half of the PTS1 Receptor Pex5p. Mol. Cell Biol. 24, 8895–8906 (2004). 80. Grimm, J. B. et al. A general method to improve uorophores using deuterated auxochromes. JACS Au 1, 690–696 (2021). 81. Day, K. J., Casler, J. C. & Glick, B. S. Budding yeast has a minimal endomembrane system. Dev. Cell 44, 56-72.e4 (2018). 82. Bolte, S. & Cordelieres, F. P. A guided tour into subcellular colocalization analysis in light microscopy. J. Microsc. 224 (2006). Scientic Reports | (2025) 15:24480 18 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/ 83. Ollion, J., Cochennec, J., Loll, F., Escudé, C. & Boudier, T. TANGO: A generic tool for high-throughput 3D image analysis for studying nuclear organization. Bioinformatics 29, 1840–1841 (2013). 84. Olsen, J. V. et al. Higher-energy C-trap dissociation for peptide modication analysis. Nat. Methods 4, 709–712 (2007). 85. Cox, J. & Mann, M. MaxQuant enables high peptide identication rates, individualized p.p.b.-range mass accuracies and proteomewide protein quantication. Nat. Biotechnol. 26, 1367–1372 (2008). 86. Cox, J. et al. Andromeda: A peptide search engine integrated into the MaxQuant environment. J. Proteome Res. 10, 1794–1805 (2011). 87. Tyanova, S., Temu, T. & Cox, J. e MaxQuant computational platform for mass spectrometry-based shotgun proteomics. Nat. Protoc. 11, 2301–2319 (2016). 88. Ader, N. R. & Kukulski, W. triCLEM: Combining high-precision, room temperature CLEM with cryo-uorescence microscopy to identify very rare events. Methods Cell Biol. 140, 303–320 (2017). 89. Kremer, J. R., Mastronarde, D. N. & Mcintosh, J. R. Computer Visualization of ree-Dimensional Image Data Using IMOD. (1996). Acknowledgements We thank Christian Wingen and Fatmire Bujupi for generating the UAS-Pex3-GFP and UAS-Pex3-HA y lines. We thank Hadar Meyer and Yeynit Asraf for their assistance with the high-throughput screens. Author contributions L.A. performed most yeast experiments, analyzed the data, and prepared gures. L.P. Performed and analyzed the Tgl4 localization microscopy based screen, and some additional yeast experimetns. R.F. Performed and processed the on-section CLEM TEM Tomography. V. F. helped with yeast experiments. E.J.B. helped with y experiments. O.E.P. supervised electron microscopy experiments and data processing. M.S. and M.Bohnert. supervised microscopy-based genome-wide screen experiments. M.Bülow performed and supervised y experiments. A.G.M supervised the whole project, prepared gures and wrote the initial dra of the manuscript. All authors revised and edited the manuscript. Funding Open Access funding enabled and organized by Projekt DEAL. is project was funded through a Deutsche Forschungsgemeinscha (DFG) individual research grant to Ayelén González Montoro (GO3313/1-1)with additional support from SFB944 and SFB1557. Work performed in the Bohnert lab was supported by the DFG, projects SFB1557 P3 (project ID 467522186), SFB1348 A13 (project ID 386797833) and FOR5815 P6 (project ID 538651361). Work in the Bülow lab was supported by DFG grants 417982926 and 535112684. Work on peroxisomes in the Schuldiner lab is supported by an Israel Science Foundation grant ISF 914/22. e robotic system of the Schuldiner lab was purchased through the kind support of the Blythe Brenden-Mann Foundation. MS is an Incumbent of the Dr. Gilbert Omenn and Martha Darling Professorial Chair in Molecular Genetics Declarations Competing interests e authors declare no competing interests. Additional information Supplementary Information e online version contains supplementary material available at h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 5 9 8 - 0 2 5 - 0 7 9 3 4 - 2 . Correspondence and requests for materials should be addressed to A.G.M. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional aliations. Open Access is article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. e images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © e Author(s) 2025 Scientic Reports | (2025) 15:24480 19 | https://doi.org/10.1038/s41598-025-07934-2 www.nature.com/scientificreports/