Quality-controlled ceramide-based GPI-anchored protein sorting into selective ER exit sites

Article Quality-controlled ceramide-based GPI-anchored protein sorting into selective ER exit sites Graphical abstract Highlights dCeramide moiety of GPI-anchor is required for clustering and ER exit of GPI-APs dGPI-AP with ceramide moiety is monitored by the GPI-glycan remodelase Ted1 dGPI ceramide remodeling is required for sorting of GPI-APs into selective ERESs dLipid-based protein sorting is quality controlled in the ER membranes Authors Sofia Rodriguez-Gallardo, Susana Sabido-Bozo, Atsuko Ikeda, ..., Kazuo Kurokawa, Manuel Mun ˜iz, Kouichi Funato Correspondence [email protected] (K.K.), [email protected]s (M.M.), [email protected] (K.F.) In brief GPI-APs undergo remodeling before exiting the ER. Rodriguez-Gallardo et al. show that GPI-APs remodeled with a very-long acyl chain ceramide moiety are monitored by a GPI-glycan remodelase. This quality-control system is necessary to ensure the lipid-based sorting of GPI- APs into selective ERESs for differential ER export. Rodriguez-Gallardo et al., 2022, Cell Reports 39, 110768 May 3, 2022 ª2022 The Authors. https://doi.org/10.1016/j.celrep.2022.110768 ll Article Quality-controlled ceramide-based GPI-anchored protein sorting into selective ER exit sites Sofia Rodriguez-Gallardo, 1,4 Susana Sabido-Bozo, 1,4 Atsuko Ikeda, 2,4 Misako Araki, 2 Kouta Okazaki, 2 Miyako Nakano, 2 Auxiliadora Aguilera-Romero, 1 Alejandro Cortes-Gomez, 1 Sergio Lopez, 1 Miho Waga, 3 Akihiko Nakano, 3 Kazuo Kurokawa, 3, *Manuel Mun ˜iz, 1, *and Kouichi Funato 2,5, * 1 Department of Cell Biology, Faculty of Biology, University of Seville and Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocı ´o/CSIC/Universidad de Sevilla, 41012 Seville, Spain 2 Graduate School of Integrated Sciences for Life, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8528, Japan 3 Live Cell Super-Resolution Imaging Research Team, RIKEN Center for Advanced Photonics, Wako, Saitama 351-0198, Japan 4 These authors contributed equally 5 Lead contact *Correspondence: [email protected] (K.K.), [email protected] (M.M.), [email protected] (K.F.) https://doi.org/10.1016/j.celrep.2022.110768 SUMMARY Glycosylphosphatidylinositol-anchored proteins (GPI-APs) exit the endoplasmic reticulum (ER) through a specialized export pathway in the yeast Saccharomyces cerevisiae. We have recently shown that a verylong acyl chain (C26) ceramide present in the ER membrane drives clustering and sorting of GPI-APs into selective ER exit sites (ERES). Now, we show that this lipid-based ER sorting also involves the C26 ceramide as a lipid moiety of GPI-APs, which is incorporated into the GPI anchor through a lipid-remodeling process after protein attachment in the ER. Moreover, we also show that a GPI-AP with a C26 ceramide moiety is monitored by the GPI-glycan remodelase Ted1, which, in turn, is required for receptor-mediated export of GPI-APs. Therefore, our study reveals a quality-control system that ensures lipid-based sorting of GPI-APs into selective ERESs for differential ER export, highlighting the physiological need for this specific export pathway. INTRODUCTION Lipidation is an essential post-translational modification by which proteins are covalently modified with specific lipids that regulate protein localization, function, and stability. Glycosylphosphatidylinositol (GPI) anchoring is a special type of lipidation present in all eukaryotes that occurs at the endoplasmic reticulum (ER) and targets GPI-anchored proteins (GPI-APs) to the cell surface, where they play a range of essential physiological roles (Kinoshita et al., 2013;Mun ˜iz and Riezman, 2016;Lopez et al., 2019;Kinoshita, 2020). The core structure of the GPI-an- chor precursor is largely conserved in evolution and consists of a phospholipid moiety (acylphosphatidylinositol) with a glycan backbone [(EtNP)Man3-(EtNP)Man2-(EtNP)Man1-GlcN], where EtNP is a side-branch ethanolamine-phosphate, Man is mannose (the numbers represent the positions of the Man in the anchor), and GlcN is glucosamine. Once the GPI-anchor precursor has been made by a series of sequential reactions at the ER membrane, it is then attached en bloc in the ER lumen by a GPI-transamidase complex to newly synthesized proteins containing a GPI attachment signal sequence at their C terminus. Immediately after GPI attachment and during GPI-AP secretory transport to the cell surface, the lipid moiety of the GPI anchor undergoes structural remodeling, which is important for GPI-AP function and trafficking (Tanaka et al., 2004;Bosson et al., 2006;Fujita et al., 2006a,2006b;Maeda et al., 2007;Fujita and Jigami, 2008;Fujita and Kinoshita, 2012;Kinoshita and Fujita. 2016). In yeast, lipid remodeling entirely occurs at the ER (Figure 1A), with an inositol deacylation by Bst1 followed by fatty-acid remodeling, which involves the removal of an unsaturated fatty acid at the sn2 position by Per1 and its replacement with a very-long chain saturated fatty acid (C26) by Gup1 (Bosson et al., 2006;Fujita et al., 2006a,2006b). In most cases, the C26 diacylglycerol (DAG), formed as part of the anchor, is replaced with a ceramide that also contains a very-long chain saturated fatty acid (C26), by Cwh43 (Conzelmann et al., 1992; Reggiori et al., 1997;Ghugtyal et al., 2007;Umemura et al., 2007). In addition to the GPI lipid remodeling, the glycan portion of the GPI anchor also undergoes another remodeling process prior to ER exit (Fujita et al., 2009;Manzano-Lopez et al., 2015). The initial side-chain EtNP on the second mannose of the glycan portion is removed by the specific phosphodiesterase Ted1 (Manzano-Lopez et al., 2015). The remodeled GPI glycan is then recognized by the transmembrane cargo receptor p24 complex, which, in turn, selectively recruits Lst1, a specific isoform of the major COPII cargo binding subunit Sec24 required to form GPI-AP-enriched COPII vesicles (Manzano-Lopez et al., 2015). Cell Reports 39, 110768, May 3, 2022 ª2022 The Authors. 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ll OPEN ACCESS A B E C D (legend on next page) 2Cell Reports 39, 110768, May 3, 2022 Article ll OPEN ACCESS While in mammals, the Golgi complex is the site where GPI-APs are sorted fromother secretory proteins, inyeast, the ER is the major sorting station. Although how the difference is made remains unknown, clustering of GPI-APs seems to be one of the determinants of thissorting process (Lebreton et al., 2019;Rodriguez-Gal- lardo et al., 2020). In yeast, GPI-APs having a C26 ceramide lipid moiety are clustered into discrete ER zones associated with specific ER exit sites (ERESs) (Rodriguez-Gallardo et al., 2020). This clustering and subsequent sorting process are hindered by decreasing the acyl chain length of free ceramide in the membrane from C26 to C16-C18. These data indicate that the chain length of ceramide in the ER membrane is critical for the sorting of GPI-APs and is consistent with a model by which C26 ceramides and GPI- APs having C26 ceramide can be cotransported to the Golgi in the same specialized COPII vesicles (Kajiwara et al., 2008;Loizides- Mangold et al., 2012;Mun ˜iz and Riezman, 2016;Funato et al., 2020). It is, however, unclear whether GPI lipid remodeling is involved in the sorting of GPI-APs or in how ceramide-based GPI-AP sorting into ERESs is regulated. Here, we investigated the role of a lipid moiety of GPI-APs in its sorting process and demonstrated that ceramide remodeling is required for clustering and subsequent sorting of GPI-APs. In addition, biochemical analyses revealed that a GPI-AP with a C26 ceramide moiety remodeled by Cwh43 is monitored by the GPI-glycan remodelase Ted1 to ensure recognition by the p24 complex and ER export of GPI-APs with C26 ceramide, providing a mechanism by which GPI-AP cargo sorting is quality controlled in the ER membranes. Our data provide insight into the physiological role of the ceramide moiety of GPI-APs on GPI-AP sorting. RESULTS ER clustering of Gas1 is driven by GPI ceramide remodeling Gas1 is a GPI-AP that has a C26 ceramide as a GPI lipid moiety (Yoko-O et al., 2013;Rodriguez-Gallardo et al., 2020). In a previous study, deletion of CWH43 wasshowntocauseaccumulation of the ER form of Gas1 (Yoko-O et al., 2013;Yoko-O et al., 2018). We confirmed and showed that, like in bst1D mutant cells, the ER form of Gas1 was accumulated in cwh43Dcells (Figure 1B), suggesting that ceramide remodeling of GPI anchors is required for the efficient ER-to-Golgi transport of GPI-APs. By contrast, the maturation of carboxypeptidase Y (CPY), a non-GPI-AP, was not affected in the bst1Dor cwh43Dcells (Figure 1B). Consistent with this, we observed that deletion of either BST1 or CWH43 retains Gas1-GFP in the ER, as is evident from the ER-characteristic nuclear ring staining (Figure 1C). These findings led to the idea that ceramide remodeling might contribute to sorting of GPI- APs into ERESs. We have recently shown that very-long acyl chain ceramides present in the ER membrane drive clustering and subsequent sorting of newly synthesized Gas1 tagged with GFP into selective ERESs (Rodriguez-Gallardo et al., 2020). Therefore, we tested whether Gas1-GFP is aggregated into clusters in cwh43Dmutant cells. For this purpose, we used the temperature-sensitive COPII allele sec31-1 as a genetic tool to reversibly retain newly synthesized cargo proteins at the ER. Gas1-GFP concentration and cluster formation at the ER membranes can be observed in the sec31-1 mutant cells under fluorescence microscopy by inhibiting COPII vesicle budding at a restrictive temperature (37C) and releasing the budding block upon shifting down to a low temperature (24C), as recently established (Rodriguez-Gallardo et al., 2020). Gas1-GFP was clustered in sec31-1 mutant cells, but it was mainly unclustered and distributed throughout the ER membrane in sec31-1 cwh43Dlike in sec31-1 bst1D(Figure 1D). In addition, analysis using super-res- olution confocal live imaging microscopy (SCLIM) also confirmed that Gas1-GFP was unclustered in the lipid-remodel- ing mutant cells and showed that Gas1-GFP clusters in the sec31-1 mutant cells were adjacent to ERESs labeled with Sec13-mCherry (Figure 1E) (Rodriguez-Gallardo et al., 2020). Figure 1. Ceramide remodeling is required for clustering of GPI-APs in ER membranes (A) A scheme for the overall GPI-AP biosynthetic pathway. The pathway can be divided into three parts: biosynthesis of the GPI anchor, attachment of the GPI anchor to protein, and remodeling of GPI-APs. After the attachment of the GPI anchor to protein, structural remodeling occurs on the lipid and glycan portions of the GPI anchor. The acyl chain linked to inositol is removed by Bst1. Then, the unsaturated fatty acyl chain within the sn2 position is eliminated by Per1 and replaced with a long saturated (C26:0) fatty acid by Gup1 followed by the replacement of the diacylglycerol (DAG) moiety by ceramides, a reaction that requires Cwh43. Side-chain ethanolamine phosphate (EtNP) residue attached to the second mannose is removed by a phosphodiesterase Ted1, and GPI-APs exit the ER in COPII-coated vesicles through direct interaction with the Emp24 complex as a cargo receptor. (B and C) Lipid remodeling is required for efficient ER-to-Golgi transport of Gas1, a ceramide-based GPI-AP. Deletion of BST1 or CWH43 causes accumulation of the ER form of Gas1. (B) Extracts prepared from wild-type and mutant cells were analyzed by western blot for Gas1 and CPY. Mature, mature forms of Gas1 and CPY; ER, immature ER form of Gas1; p1 and p2, ER and Golgi forms of CPY, respectively. Quantification of the percentage of immature Gas1 was performed. The bars represent the average of three independent biological replicates. Statistics: error bars indicate the SD. Student’s t test. *p < 0.05. (C) Gas1-GFP localization in wild-type and deletion mutant cells was analyzed by conventional fluorescence microscopy. Scale bar, 5 mm. (D and E) Newly synthesized Gas1-GFP forms clusters in the wild-type ER membrane adjacent to specific ERESs but not in the bst1Dor cwh43Dmutant ER membrane. (D) sec31-1,sec31-1 bst1D, and sec31-1 cwh43Dcells expressing galactose-inducible Gas1-GFP (green) and constitutive ERES marker Sec13-mCherry (ERES, magenta) were incubated at 37C for 30 min and then shifted down to 24C to release a secretion block and imaged by conventional fluorescence microscopy after 20 min. White arrowheads: ER Gas1-GFP clusters; open arrowheads: unclustered Gas1-GFP distributed throughout the ER membrane showing the ER- characteristic nuclear ring staining. Scale bar, 5 mm. (E) The same strains as in (D) were incubated at 37C for 30 min, shifted down to 24C to release a secretion block, and imaged by SCLIM after 20 min. Left panels show representative merged two-dimensional (2D) images of Gas1-GFP and Sec13-mCherry at one xyz plane. Scale bar, 1 mm. Right panels show representative merged 3D images of Gas1-GFP and Sec13-mCherry. Scale unit, 0.65 mm. Gas1-GFP was detected in discrete ER zones or clusters adjacent to specific ERESs in wild-type cells. Gas1-GFP was unclustered in lipid-remodeling-deficient mutant cells. Cell Reports 39, 110768, May 3, 2022 3 Article ll OPEN ACCESS Gas1-GFP IP T IP T IP T IP T IP TIP T - + + + + + IP: an-GFP Emp24p Pgk1p Gas1-GFP D C WT gpi7Δcwh43Δgpi7Δcwh43Δ Gas1- GFP A ER 0 10 20 30 40 Gas1p p2 (Golgi) mature CPY p1 (ER) mature gpi7Δ cwh43Δ cwh43Δgpi7ΔWT *** Immature Gas1p (%) B Mid2- VENUS P EtN KN lysine-asparagine ethanolamine phosphate mannose glucosamine inositol (mature) [M+2H]2+ = 860.762±10 ppm 0.00% 0.00% 68.23% 1.26% 89.05% 0.00% 1.65% 0.00% 0.00% [M+2H]2+ = 841.240±10 ppm 0.00% 89.14% 8.58% 1.83% [M+2H]2+ = 779.736±10 ppm 2.56% 2.29% 26.63% 0.00% 7.27% (ER) 1.64% 88.51% 1.85% [M+2H]2+ = 718.232±10 ppm 5.33% 8.57% 3.50% [M+2H]2+ = 799.258±10 ppm 92.11% proposed GPI-glycan structures, monoisotopic masses of each GPI- glycan relave abundances (%) of each GPI-glycan from Gas1-GFP WT gpi7 Δ cwh43 Δ gpi7 Δcwh43 Δ ted1 Δ (mature) (mature) (ER) Figure 2. EtNP trimming is required for ceramide-remodeling-dependent association of Gas1p with p24 complex and ER export of Gas1 (A) Coimmunoprecipitation (coIP) assay between Gas1-GFP and the p24 complex subunit Emp24. gpi7Dbypasses the requirement of Cwh43 for the association of Gas1 with p24 complex. Gas1-GFP is specifically recognized by the p24 complex in a Cwh43-dependent manner. This physical interaction is specific because Pgk1, an unrelated cytosolic protein, is not associated with Gas1-GFP after IP, and the gpi7Dmutation suppresses the defect in an interaction between (legend continued on next page) 4Cell Reports 39, 110768, May 3, 2022 Article ll OPEN ACCESS These results demonstrate that ceramide remodeling of the GPI anchors is required for ER clustering of Gas1. Gas1 with a C26 ceramide moiety is monitored by the GPI-glycan remodelase Ted1 to ensure ER export As Gas1-GFP clustering at the ER requires removal of EtNP from the second mannose residue in GPI anchors by Ted1 and subsequent recognition by the transmembrane cargo receptor p24 complex (Rodriguez-Gallardo et al., 2020), GPI ceramide remodeling by Cwh43 could be indirectly required for Gas1 clustering through p24 binding. Therefore, we addressed whether Gas1 could bind with the p24 complex in cwh43Dby examining the extent of interaction between Gas1-GFP and Emp24, a component of the p24 complex. As a negative control, we used ted1D mutant cells in which the interaction between GPI-APs and Emp24 is disrupted (Manzano-Lopez et al., 2015). As shown in Figure 2A, in ted1Dmutant cells, Emp24 did not coprecipitate with Gas1-GFP, whereas wild-type cells exhibited an association of Emp24 with Gas1-GFP. By using this method, we found that Emp24 could not be efficiently coprecipitated with Gas1-GFP in the cwh43Dcells, suggesting that GPI ceramide remodeling is required for the recognition of Gas1 by the p24 complex. Furthermore, this binding defect observed in cwh43Dcells was rescued by the additional deletion of GPI7 (gpi7Dcwh43D). Since GPI7 encodes an enzyme that adds the EtNP on the second mannose during GPI anchor biosynthesis that is later removed by Ted1, the loss of function of Gpi7 leads to the production of GPI anchors without the EtNP on the second mannose (Benachour et al., 1999). Therefore, the binding rescue by gpi7 mutation suggests that the Emp24-binding defect in cwh43Dcells could be due to the inability of Ted1 to trim the second EtNP. We then addressed this possibility by mass spectrometry analysis of the GPI glycan of Gas1-GFP. Most of the peptides bearing GPI glycan from the mature Gas1-GFP in wild-type and gpi7D mutant cells had no side-chain EtNP attached to mannose (m/z 799.258 corresponding to KN-EtNP-Man(Man-Man)-Man- Man-GlcN-Ino-P), whereas GPI glycan from the immature Gas1-GFP in the cwh43Dor ted1Dcells had two side-chain EtNPs (m/z = 841.240 corresponding to KN-EtNP-Man(Man)- Man(EtNP)-Man(EtNP)-GlcN-Ino-P) (Figure 2B). Therefore, these findings confirm that Ted1 only acts once the GPI anchor of Gas1 has been remodeled to ceramide by Cwh43. If ER export of Gas1 is dependent on the trimming of the second EtNP and subsequent recognition by the p24 complex, the ER-export defect in cwh43Dcells should be rescued by the gpi7Dmutation because the Emp24-binding defect in cwh43D was suppressed by deleting GPI7 (Figure 2A). As seen in Figure 2C, deletion of GPI7 suppressed the accumulation of the ER form of Gas1 caused by the cwh43Dmutation. Gas1-GFP retained in the ER of cwh43Dwas also diminished by the gpi7D mutation (Figure 2D). These data are consistent with the model that ceramide remodeling is required for clustering and ER export of Gas1 through the trimming of EtNP in GPI anchors. Consistently, the role of ceramide remodeling is specific for GPI-APs because transport of non-GPI-anchored cargo, CPY (Figure 2C) or Mid2-VENUS (Figure 2D), was not affected in cwh43Dcells. Taken together, these results suggest that glycan remodelase Ted1 monitors GPI-AP with a ceramide moiety remodeled by Cwh43 to ensure that exported Gas1 has ceramide as a GPI lipid moiety. We have recently shown that the chain length of ceramide in the ER membrane is critical for the clustering and sorting of Gas1 remodeled with C26 ceramide (Rodriguez-Gallardo et al., 2020). This is consistent with a model in which free ceramides and ceramide-based GPI-APs are cotransported in the same vesicles to the Golgi (Kajiwara et al., 2008;Loizides-Mangold et al., 2012;Mun ˜iz and Riezman, 2016;Funato et al., 2020). Since deletion of GPI7 suppressed the ER accumulation of unremodeled GPI-APs, we next analyzed if the free-ceramide requirement for ER exit of GPI-APs can be restored by the gpi7Dmutation. For this purpose, we used myriocin, a specific inhibitor of serine palmitoyltransferase (SPT), that is the enzyme catalyzing in the first step in sphingolipid synthesis (Horvath et al., 1994). We observed that in myriocin-treated wild-type cells, Gas1 and Gas1-GFP but not non- GPI-APs (CPY, Mid2-VENUS) accumulated in the ER, whereas in myriocin-treated gpi7Dcells, the GPI-APs were not localized to the ER (Figures 3A and 3B). Similar results were obtained with temperature-sensitive lcb1-100 mutant background strains (Figure 3C), which are defective in SPT at a non-permissive temperature (37C) (Zanolari et al., 2000). These results indicate that removal of EtNP from the second mannose overrides not only ceramide remodeling but also the free-ceramide requirement for ER exit of GPI-APs. When GPI7 is deleted, the retention of Gas1 in the ER caused by emp24Dwas not rescued (Figures 4A and 4B). Similar results were obtained in arv1Ddefective in GPI-anchor synthesis (Kajiwara et al., 2008)(Figure 4C), bst1D,orgup1D(Figure 4D). Therefore, the ceramide-bypass pathway in the absence of EtNP on the second mannose requires remodeling of the GPI anchor to C26 DAG, in addition to recognition by the p24 complex, because the gup1Dmutant is deficient in the generation of a pG1-type anchor consisting of C26 DAG (Bosson et al., 2006; Fujita et al., 2006a;Umemura et al., 2007;Ghugtyal et al., 2007). Gas1-GFP and the p24 complex caused by the deletion of CWH43. Enriched ER fractions of wild-type and mutant cells expressing Gas1-GFP were processed for native IP. Bound material (IP) was resolved by SDS-PAGE and analyzed by immunoblotting using antibodies against Emp24, Pgk1, and GFP. Total (T) represents a fraction of the solubilized input material. (B) Structural analysis of the GPI glycan of Gas1-GFP. Extracts prepared from cells expressing Gas1-GFP were treated with PI-specific phospholipase C (PIPLC), and then Gas1-GFP was purified, digested with trypsin, and analyzed by liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS). The table shows the percentage of different GPI-glycan structures found in the mature or ER forms of Gas1-GFP purified from wild-type and mutant cells. The relative abundance (%) of each GPI glycan was calculated from the peak intensity of monoisotopic masses of each GPI glycan. (C and D) gpi7Dmutation suppresses the accumulation of the ER form of Gas1 caused by the cwh43Dmutation. (C) Extracts prepared from cells were analyzed by western blot for Gas1 and CPY as in Figure 1B. Quantification of the percentage of immature Gas1 was performed. The bars represent the average of three independent biological replicates. Error bars indicate the SD. Statistics: Student’s t test. ***p < 0.001. (D) Localization of Gas1-GFP and Mid2-VENUS was analyzed by conventional fluorescence microscopy. Scale bar, 5 mm. Cell Reports 39, 110768, May 3, 2022 5 Article ll OPEN ACCESS 0 10 20 30 40 WT gpi7Δ gpi7ΔWT myriocin (μg/mL) 0 10 200 10 20 n.s. *** Gas1p ER p2 (Golgi) mature CPY p1 (ER) mature □ ■myriocin 20μg/mL none Immature Gas1p (%) gpi7ΔWT WT gpi7Δ Gas1p ER mature 0 10 20 30 40 Immature Gas1p (%) temperature (℃)37 37 37 3725 25 *** B A C Gas1- GFP Mid2- VENUS myriocin － ＋ － ＋ Figure 3. Elimination of EtNP attached to second mannose bypasses ceramide-dependent transport of GPI-APs (A–C) gpi7Dmutation suppresses the accumulation of the ER form of Gas1 caused by inhibition of sphingolipid synthesis. (A) Extracts prepared from cells treated with indicated concentration of myriocin were analyzed by western blot for Gas1 and CPY as in Figure 1B. Quantification of the percentage of immature Gas1 was performed. The bars represent the average of three independent biological replicates. Error bars indicate the SD. Statistics: Student’s t test. n.s., not significant, and ***p < 0.001. (B) Localization of Gas1-GFP and Mid2-VENUS in the cells treated with or without myriocin (10 mg/mL, for 1.5 h at 24C) was analyzed by conventional fluorescence microscopy. Scale bar, 5 mm. (C) Cells grown at 25C were shifted to 25Cor37 C for 120 min, and the extracts prepared from the cells were analyzed by western blot for Gas1 as in Figure 1B. Quantification of the percentage of immature Gas1 was performed. The bars represent the average of three independent biological replicates. Error bars indicate the SD. Statistics: Student’s t test. ***p < 0.001. 6Cell Reports 39, 110768, May 3, 2022 Article ll OPEN ACCESS Gas1 sorting into selective ERESs depends on GPI ceramide remodeling We have recently shown that Gas1-GFP and the transmembrane plasma membrane protein Mid2-iRFP are segregated into different ERESs during ER export (Rodriguez-Gallardo et al., 2020). In addition, the data presented in the previous paragraph show that GPI7 mutation, which generates GPI-APs with C26 DAG-based GPI (Benachour et al., 1999;Fujita et al., 2006a; Umemura et al., 2007), bypasses the Ted1 requirement for p24-dependent ER export of Gas1. Therefore, we took advantage of the gpi7Dmutant to address whether GPI ceramide remodeling is required for sorting of Gas1, since C26 DAG-based Gas1 in cwh43Dmutant cells was unclustered (Figures 1D and 1E). Thus, we constructed single (sec31-1), double (sec31-1 gpi7D), and triple (sec31-1 gpi7Dcwh43D) mutants of sec31-1, gpi7D, and cwh43Dexpressing Gas1-GFP and Mid2-iRFP with the COPII outer coat protein Sec13 tagged with mCherry. We then analyzed them by SCLIM and quantified the proportion of ERESs in which only one type of cargo or both are present (Figures 5A and 5B). SCLIM analysis showed that after incubation at 37C and subsequent release at 24C for 20 min, most of the ERESs (about 75%) contained only one type of cargo in sec31-1 mutant cells. The top panels of Figure 5A show typical examples of ERESs with only one cargo in sec31-1 mutant cells. About 15% of ERESs contained both cargos overlapping in the same area (Figures 5A, yellow filled arrowheads), and 10% of ERESs contained both cargos but segregated in clearly distinct zones. These results indicate that Gas1-GFP is sorted into different ERESs from Mid2-iRFP upon ER exit. Most importantly, we found that in either sec31-1 gpi7Dor sec31-1 gpi7Dcwh43D mutant cells, a large percentage of ERESs (40%) include both cargos overlapping in the same area, and the percentage of 0 10 20 30 40 50 60 Gas1p ER mature C Immature Gas1p (%) Gas1p ER mature D 0 10 20 30 40 50 60 Immature Gas1p (%) A 0 10 20 30 40 Gas1p ER p2 (Golgi) mature CPY p1 (ER) mature Immature Gas1p (%) WT gpi7Δemp24Δ gpi7Δ emp24Δ Gas1- GFP Mid2- VENUS B n.s. n.s. n.s. n.s. Figure 4. ER export of GPI-APs in gpi7Drequires GPI anchoring, PI with a C26:0 fatty-acid moiety, and p24 complex (A, C, and D) Extracts prepared from wild-type and mutant cells were analyzed by western blot for Gas1 (A, C, and D) and CPY (A) as in Figure 1B. Quantification of the percentage of immature Gas1 was performed. The bars represent the average of three independent biological replicates. Error bars indicate the SD. Statistics: Student’s t test. n.s., not significant. (B) Localization of Gas1-GFP and Mid2-VENUS was analyzed by conventional fluorescence microscopy. Scale bar, 5 mm. Cell Reports 39, 110768, May 3, 2022 7 Article ll OPEN ACCESS A BC (legend on next page) 8Cell Reports 39, 110768, May 3, 2022 Article ll OPEN ACCESS Culture conditions Strains were grown either in rich YP medium (1% yeast extract, 2% peptone) supplemented with 0.2% adenine and containing either 2% glucose (YPD), 2% raffinose (YPR) or 2% galactose (YPG) as carbon source or in synthetic minimal medium (0.15% yeast nitrogen base, 0.5% ammonium sulfate) supplemented with the appropriate amino acids and bases as nutritional requirements, and containing either 2% glucose (SD) or 2% galactose (SG) as carbon source. Immunoblotting Cell lysates denatured with SDS sample buffer were separated by SDS-PAGE followed by immunoblotting (Manzano-Lopez et al., 2015;Ikeda et al., 2016). Blots were probed with rabbit polyclonal antibodies against Gas1, CPY, Emp24 and GFP, or rat monoclonal antibody against HA (3F10) and detected by chemiluminescence using peroxidase-conjugated affinity-purified goat anti-rabbit IgG or anti-rat IgG antibodies. Fluorescence microscopy Fluorescence observation by conventional fluorescence microscopy or super-resolution confocal live imaging microscopy (SCLIM) was performed as described (Manzano-Lopez et al., 2015;Rodriguez-Gallardo et al., 2020). For conventional fluorescence microscopy of Gas1-GFP, Mid2-Venus and Ted1-Venus, cells expressing Gas1-GFP, Mid2-Venus or Ted1-Venus were grown to log phase overnight in YPD, collected by centrifugation, and washed twice with phosphate-buffered saline. The cells were then incubated at least 15 min on ice before being examined by microscopy as described previously (Rodriguez-Gallardo et al., 2020). For visualization of Gas1-GFP clusters and into ERES, temperature-sensitive sec31-1 mutant cells expressing galactose-inducible cargos (Gas1-GFP and/or Mid2-iRFP) and constitutive ERES marker Sec13-mCherry were grown to log phase overnight in YPR. After induction for 1 h at 24C in YPG, cells were incubated for 30 min at 37C in SG, shifted down to 24C for releasing secretory block, imaged by conventional fluorescence microscopy or SCLIM, and analyzed with Volocity software as described (Rodriguez-Gallardo et al., 2020). Native coimmunoprecipitation Native coimmunoprecipitation experiment was performed as described (Rodriguez-Gallardo et al., 2020). The ER-enriched fraction was solubilized by treatment with 1% digitonin for 1h at 4C and immunoprecipitated with GFP-Trap beads (Chromo Tek). The immunoprecipitated GFP protein complexes were separated by SDS-PAGE and analyzed by immunoblot. MS analysis of GPI-glycan MS analysis of GPI-glycan of Gas1-GFP was carried out exactly as described (Rodriguez-Gallardo et al., 2020). After treatment of Gas1-GFP with PI-PLC, immunoprecipitation with GFP-Trap, separation by SDS-PAGE, staining with Coomassie brilliant blue, and in-gel digestion with trypsin, the peptides carrying GPI-glycan were subjected to HPLC/mass spectrometry. MS and MS/MS data were obtained in positive ion mode over the mass range mass/charge ratio (m/z) 300 to m/z 3,000 by LTQ Orbitrap XL. Lipid labelling with [ 3 H]myo-Inositol In vivo labelling of lipids with [ 3 H]myo-inositol was carried out as described (Ikeda et al., 2021). Radiolabeled lipids were extracted with chloroform-methanol-water (10:10:3, vol/vol/vol), and analyzed by thin-layer chromatography using a solvent system (chloroform-methanol-0.25% KCl (55:45:10, vol/vol/vol)) for complex sphingolipids. Radiolabeled complex sphingolipids were visualized and quantified on an FLA-7000 system. Pulse-chase analysis Pulse-chase analysis of Gas1 with [ 35 S]methionine was performed as described previously (Kajiwara et al., 2008). After pulse-labeled and chased, the samples immunoprecipitated with anti-Gas1p rabbit antiserum and protein A-Sepharose were separated by SDSPAGE and analyzed with a Phosphor Imager. QUANTIFICATION AND STATISTICAL ANALYSIS Statistical analysis Statistical analysis was performed using Student’s t-Test (n.s. not significant; *, p < 0.05; **, p < 0.01 and ***, p < 0.001). The mean ±s. d. for three independent experiments is shown. Cell Reports 39, 110768, May 3, 2022 e2 Article ll OPEN ACCESS