The autophagic and non-autophagic functions of the S. cerevisiae PROPPIN Hsv2

ABSTRACT

Autophagosome formation depends on PtdIns3P, its presence is deciphered by PROPPINs, a family of β-propellers, which in yeast consists of Atg18, Atg21 and Hsv2 and in mammals of WIPI1, WIPI2, WDR45B/WIPI3 and WDR45/WIPI4. While Atg18 is required for scaffolding the Atg2-Atg9 complex, which mediates non-vesicular membrane transport to the phagophore, Atg21 organizes the Atg8 lipidation machinery. Atg18 further acts as part of a retromer complex in vacuole fragmentation. So far, the function of Hsv2 remained elusive. Here we show that Hsv2 is required for autophagy of large cargos such as the fatty acid synthase complex (FAS) and ribosomes. We further found that Hsv2 interacts with the key retromer component Vps35 and mediates vacuole fission cooperatively with Atg18. Interestingly, the residues for interaction of Hsv2 with Atg2 and Vps35 are distinct from those of Atg18. Hsv2 is known to affect the biogenesis of the spore wall, which prompted us to include diploid cells in our analyses. We found that Hsv2 interacts with the SNARE Pep12, and that Pep12 mislocalized to the vacuole in diploid but not haploid hsv2∆ cells. This suggests a role of Hsv2 in protein sorting in diploid cells. The loop 6C/D of PROPPINs partially inserts into membranes causing their bending. We found that the membrane bending activity of Hsv2 is required for vacuole fragmentation and sorting in diploids but not for its autophagic function. Mutations in WDR45/WIPI4, the presumed mammalian homolog of Hsv2 cause the neurodegenerative disease BPAN, our study thus also helps to understand its underlying principles.

Abbreviations: CSC: cargo specific complex, FAS: fatty acid synthase, PAS: phagophore assembly site; PROPPIN: beta-propeller that binds phosphoinositides.

Introduction

Macroautophagy/autophagy delivers superfluous or damaged intracellular material including parts of organelles to lysosomes (vacuoles in yeast) for degradation and reuse. It is highly conserved from S. cerevisiae to humans and its mechanism has gained a lot of interest due to its relevance for numerous diseases such as cancer and neurodegenerative diseases [1,2]. Here we focus on autophagy in S. cerevisiae. It is upregulated when nutrients are scarce by modulating the activity of the Atg1 kinase [3]. This leads to the assembly of double-membraned phagophores, which are then expanded and closed to form autophagosomes. Eventually, the autophagosome fuses with the vacuole to release its content [4]. Expansion of the phagophore requires non-vesicular influx of membrane lipids from the ER and probably also the vacuole mediated by Atg2 and their distribution by the scramblase Atg9 [5–7]. Another crucial component is ubiquitin-like Atg8, which is covalently coupled to phosphatidylethanolamine (PE). This coupling depends on ubiquitin-like conjugation systems and the E3-like component Atg5 coupled to Atg12 in complex with Atg16 [8]. Atg8–PE is involved in phagophore expansion and is thought to be part of a coat-like structure during autophagosome biogenesis [9,10] and in selective autophagy it is involved in binding cargo receptors to the inside of the phagophore. The recruitment and arrangement of these components to phagophores largely depends on PtdIns3P, which is generated by an autophagy specific PtdIns 3-kinase complex [5]. The presence of PtdIns3P is decoded by a family of proteins, which possess two PtdIns3P-binding sites at the circumference of their seven bladed β-propeller fold [11–14]. These proteins have been termed PROPPINs for β-propellers that bind phosphoinositides [15], they are generally thought to act as scaffolds for protein complexes. In S. cerevisiae the PROPPIN family consists of Atg18, Atg21 and Hsv2. Atg18 is a core autophagy protein, which has been shown to organize the Atg2-Atg9 complex at the ER and vacuolar contact sites of phagophores [16–18]. Atg21 on the other hand scaffolds the Atg8 lipidation machinery at the vacuolar contact site of the phagophore by binding Atg8 and the Atg12–Atg5-Atg16 complex [19]. PROPPINs also bind PtdIns(3,5)P₂ and thus can fulfill additional functions beside their role in autophagy [12]. Atg18 for example regulates through Vac14 the PtdIns3P 5-kinase Fab1 and it interacts with the myosin 5-specific adaptor Vac17 [20–22].

Recently, Atg18 has been shown to be part of a novel type of retromer complex, which affects vacuole fragmentation upon hyperosmotic stress, but not autophagy [23,24]. In S. cerevisiae the retromer complex consists of an arch-shaped cargo specific complex (CSC) formed by Pep8/Vps26, Vps29 and Vps35, which assembles with the SNX-BAR sorting nexins Vps5 and Vps17 [25–27]. The sorting nexins bind via their PX domain to PtdIns3P and their BAR domain confer membrane bending. The Vps5-Vps17-retromer complex is involved in retrograde trafficking from the endosome to the Golgi. Atg18 binds the CSC competitively to Vps5 and Vps17 and replaces them in retromer complexes [23,24]. As the SNX-BAR sorting nexins, Atg18 strengthens membrane association by binding PtdIns3P and PtdIns(3,5)P₂ and can induce membrane bending by partial insertion of its amphipathic loop 6C/D into membranes. Interestingly, this membrane bending activity of Atg18 is only needed for retromer-dependent vacuole fragmentation, but not for autophagy [28].

Similar to the yeast PROPPINs autophagic and non-autophagic functions have been attributed to their four human homologs WIPI1, WIPI2, WDR45B/WIPI3 and WDR45/WIPI4 [14]. For WIPI1 a retromer but not autophagy-dependent role in transferrin receptor cycling from the endosome to the plasma membrane was observed. Thus, a role of WIPI1 in PtdIns(3,5)P₂-dependent fission of endosomal transport carriers was proposed [24,29].

WIPI1 can enhance the autophagic role of WIPI2 [30], which organizes the LC3 lipidation machinery including ATG16L1 [31,32]. In this function WIPI2 resembles yeast Atg21. WDR45 (WD repeat domain 45): WDR45B/WIPI3 and WDR45/WIPI4 were both linked to phagophore expansion, they interact with the non-vesicular membrane-lipid transfer protein ATG2A [6,7,33–37]. WDR45B/WIPI3 and WDR45/WIPI4 in this regard correspond to Atg18. The high sequence similarity of WDR45/WIPI4 with Hsv2 gave a hint for an interaction of Hsv2 with Atg2. In a two-hybrid assay indeed such an interaction was detected, but no molecular function was assigned [35]. Mutations in WDR45/WIPI4 cause BPAN (beta-propeller protein associated neurodegeneration) a severe neurodegenerative disease [14], which has been linked to a disturbed interaction of WDR45/WIPI4 with ATG2A [35]. However, detailed studies suggest that not the loss of the autophagic function causes BPAN, but mislocalization of ATG2A [38,39]. In single knock down experiments only WIPI2 was essential for autophagy, while only a double deletion of WDR45B/WIPI3 and WDR45/WIPI4 led to significant autophagic defects [39].

Here we uncover a role of Hsv2 in the autophagy of very large cargos such as fatty acid synthase and ribosomes. We further detected in co-immunoprecipitations interaction of Hsv2 with the retromer component Vps35 and describe a function of Hsv2 in hyperosmotic vacuole fragmentation in cooperation with Atg18. Interestingly, we found that Hsv2 uses distinct residues for binding Atg2 and Vps35. Furthermore, we demonstrate an involvement of Hsv2 in Pep12 sorting in diploid cells. Mutational analyses of the membrane bending Hsv2 loop 6C/D show its requirement for vacuole fragmentation and protein sorting in diploids, but its dispensability for autophagy.

Results

Hsv2 cooperates with Atg18-retromer in hyperosmotic vacuolar fragmentation

Recently, an Atg18-retromer complex involved in vacuolar fragmentation upon hyperosmotic stress was identified [23,24]. We analyzed if Hsv2 can also form retromer complexes. Indeed, chromosomally expressed Vps35-6xHa co-immunoprecipitated with Hsv2-GFP, but not with GFP alone (Figure 1(A)). While atg18∆ cells exhibited a strong vacuolar fragmentation defect in the presence of 0.4 M NaCl, hsv2∆ cells showed wild-type like vacuolar fragmentation [40]. We therefore speculated that the role of Hsv2 in retromer complexes might be masked by other PROPPINs. Since vacuole fragmentation is severely reduced in atg18∆ cells additive effects of the other PROPPINs are hard to measure. We thus decided to dissect putative additive effects in vps5∆ cells. vps5∆ cells have highly fragmented vacuoles even in the absence of Atg18 and thus allow analyses of vacuolar morphology with higher sensitivity. Furthermore, Vps5 competes with the formation of Atg18-retromer. Accordingly, lack of Vps5 significantly increased the levels of Atg18-retromer, while overexpression of Atg18 led to decreased interaction of Vps5 with Vps35 [23]. We counted the number of FM 4–64-stained vacuoles one h (Figure 1(D,E) and S1) and 15 min (Figure 1(F)) after incubation with 0.4 M NaCl. We further analyzed the vacuole fragmentation without salt (Figure S2). Quantification and statistical evaluation showed that under these conditions deletion of either ATG18 or HSV2 in the vps5∆ background had only slight effects on vacuole fragmentation, while concomitant deletion of both significantly reduced it. On the other hand, lack of Atg21 alone or in combination with the other PROPPINs did not largely affect vacuole fragmentation of vps5∆ cells. This indicates that in the absence of Vps5, Atg18 and Hsv2 cooperatively mediate vacuole fission. In line with such an overlapping function, the steady state expression level of Hsv2 in Western blots increased ~50% in the absence of Atg18 (Figure 1(A,B)). Interestingly, this increase did not lead to increased interaction of Hsv2-GFP with Vps35-6xHA in atg18∆ cells, only in vps5∆ atg18∆ cells increased interaction was detectable (Figure 1(C)). This fits with the idea that lack of Vps5 causes increased interaction of Atg18 with Vps35 and additional lack of Atg18 increases the interaction of Hsv2 with Vps35.

Figure 1.

Hsv2 interacts with Vps35 and affects vacuole fragmentation in cooperation with Atg18. (A) Coimmunoprecipitation of the indicated strains in the WCG background chromosomally expressing Vps35-6xHA and Hsv2-GFP from plasmid with its endogenous promoter with anti-GFP micro beads. As negative control, GFP alone was expressed. The strains were grown in selective media to the early stationary phase. (B) Expression levels of Hsv2 increase in cells lacking Atg18. Five independent Western blots (A) were analyzed. Values are normalized to WT-levels. Error bars are SEM. (C) Double deletion of VPS5 and ATG18 increases binding of Hsv2 to Vps35. Five independent experiments (A) were quantified using Fiji and statistically evaluated with GraphPad Prism 9. Error bars are SEM. (D) BY4742 deletion strains were grown to OD₆₀₀ 0.5–1 in YPD medium. After incubation with fm 4–64 for 30 min cells were resuspended in YPD with 0.4 M NaCl and incubated for 1 h before microscopy. From left to right fm 4–64 fluorescence, Nomarski and a merge of the single plane images are shown. (E,F) Double deletion of Hsv2 and Atg18 reduces vacuole fragmentation. Quantification of six experiments after 1 h incubation in salt with 200 cells each. (F) Vacuole fragmentation as in (e) after only 15 min. salt incubation. Statistical relevance with an unpaired two-tailed t-test. p-values: ns, not significant p > 0.05. * for p < 0.05. ** for p < 0.01. *** for p < 0.001. **** for p < 0.0001. Error bars indicate SEM.

Membrane fission activity of Hsv2 depends on its amphipathic loop 6C/D

The Atg18 loop 6C/D folds into an amphipathic helix, which is thought to partly penetrate the membrane. This induces membrane bending required for the fission activity of Atg18-retromer [28] and enhances membrane binding. In line, Hsv2 amino acids 318 to 335 of loop 6C/D are predicted to form an amphipathic helix (Figure 2(A)). We thus generated a Hsv2^I325D,L321K mutant in which the hydrophobic region of the amphipathic helix is disrupted by inserting charged amino acids (Figure 2(A)). As expected, Hsv2^I325D,L321K is released to the cytosol under non-starvation conditions (Figure 2(B)), while upon starvation for nitrogen additional punctae appear, which as outlined below correspond to the PAS. We expressed this mutant in vps5∆ atg18∆ hsv2∆ cells and after one h in 0.4 M NaCl counted the number of FM 4–64-stained vacuoles (Figure 2(C)). Quantification showed that in contrast to wild-type Hsv2 the mutated Hsv2^I325D,L321K is unable to restore increased vacuole fragmentation (Figure 2(D)). This demonstrates that similar to Atg18 the membrane fission activity of Hsv2 depends on its amphipathic loop 6C/D. Co-immunoprecipitations showed that binding of Hsv2^I325D,L321K to Vps35 is reduced to about half of the wild-type Hsv2 (Figure 2(E,F)). This suggests that Hsv2-retromer formation is favored by membrane association.

Figure 2.

The amphipathic helix of the Hsv2 loop 6C/D is required for vacuole fragmentation. (A) Helical wheel projection of Hsv2 loop 6C/D, the hydrophobic side of the amphipathic helix is marked by a blue arc. The mutations Hsv2^L321K and Hsv2^I325D are marked with red arrows. (B) Fluorescence microscopy confirms release of Hsv2^L321K,I325D-GFP to the cytosol under non-starvation conditions (0 h), after 2 h incubation in SD-N punctae appear. Vacuolar staining with fm 4–64 is enclosed. (C) vps5∆ atg18∆ hsv2∆ cells in the BY4742 background expressing Hsv2 and Hsv2^L321K,I325D from plasmid with the endogenous promoter were grown in selective medium to OD₆₀₀ of 0.5–1. After incubation with fm 4–64 for 30 min cells were resuspended in YPD with 0.4 M NaCl and incubated for 1 h. (D) Five experiments were conducted with 200 cells counted each. Statistical relevance was determined by an unpaired two-tailed t-test. Asterisks indicate p-values * for p < 0.05. Error bars indicate SEM. (E,F) Co-immunoprecipitation with anti-GFP beads shows binding of Hsv2^L321K,I325D to Vps35-Ha, with is about half of the wild-type. Statistical relevance was determined by an unpaired two-tailed t-test. Asterisks indicate p-values * for p < 0.05. Error bars are SEM.

Hsv2 localizes to both edges of the phagophore

Based on the cooperative action of Hsv2 and Atg18 in vacuole fragmentation, we speculated that Hsv2 might also act during autophagy. This is supported by work on the putative mammalian Hsv2 homologs WDR45B/WIPI3 and WDR45/WIPI4, which indicated an interaction with ATG2A [33,35]. In extension of this work an interaction of Hsv2 with Atg2 was detected with the 2-hybrid system, but the functional relevance of this interaction was not uncovered [35]. We first analyzed whether Hsv2 localizes to the PAS, the phagophore assembly site. Indeed, Hsv2-GFP colocalized with the PAS marker mCherry-Atg8 (Figure 3(A,B)). Interestingly, normal PAS recruitment of Hsv2-GFP required both Atg18 and Atg2. In line with the PtdIns3P-binding ability of Hsv2, Atg14 as a component of the PtdIns 3-kinase complex at the PAS and the key PAS components Atg1 and Atg9 are needed (Figure 3(A,B)). Overexpression of the selective autophagy cargo prApe1 leads to the formation of a giant cargo complex, which allows direct evaluation of phagophores in fluorescence microscopy [41]. We found that Hsv2-GFP is typically located at both edges of the mCherry-Atg8 positive phagophore (Figure 3(C-F)). The membrane-bending defective mutant Hsv2^I325D,L321K-GFP showed a comparable localization at both edges, but with a somewhat increased localization at the non-vacuolar edge (Figure 3(D-F)).

Figure 3.

Hsv2 localizes to both edges of the phagophore. (A) Strains in the WCG background expressing Hsv2-GFP and mCherry-Atg8 from their endogenous promoters were grown to log phase in selective medium and starved for 30 min in SD-N. (B) Quantification of five independent experiments in which 50–100 cells were counted each. Statistical relevance was determined by an unpaired two-tailed t-test. Asterisks indicate p-values ** for p < 0.01. Error bars indicate SEM. (C) Localization of Hsv2-GFP on mCherry-Atg8 positive phagophores, which form around giant prApe1 structures. prApe1 was expressed from plasmid with the CUP1-promoter. Expression was induced with 100 µM CuSO₄ to generate the giant prApe1 structures. Hsv2-GFP localized to both edges of the enlarged phagophore. (D) as in (C) the membrane bending defective Hsv2^L321K,I325D-GFP was monitored. (E) 3D representations were generated using Huygens Professional software from fluorescence microscopic images. Size bars 0.5 µm. (F) Vacuolar edges were identified with the help of the Nomarski images. Statistical relevance was determined by an unpaired two-tailed t-test from three experiments. * for p < 0.05. ** for p < 0.01. Error bars are SEM.

Hsv2 affects autophagy of very large cargos

We speculated that the interaction of Hsv2 with Atg2 might help to allow membrane influx for phagophore growth. Membrane delivery to phagophores is most critical when very large cargos are engulfed. We thus tested the giant prApe1 complex generated by overexpression of prApe1 as a putative cargo. Indeed, hsv2∆ cells starved for nitrogen in SD-N medium showed a reduced maturation rate compared to wild-type cells (Figure 4(A,B)). Endogenously expressed prApe1 does not form the giant prApe1 complex, well in line maturation of non-overexpressed prApe1 was not affected in hsv2∆ cells (Figure S3). Since overexpressed prApe1 is a well-known, but artificial large cargo, we extended our analyses to natural cargoes. We used Rpl9b-GFP, a component of the large ribosomal subunit as a known bulky autophagic cargo [42,43]. Again, hsv2∆ cells showed a reduced degradation rate compared to wild-type cells (Figure 4(C,D)). Fatty acid synthase is a large 2.6 MDa enzyme complex known to be preferentially degraded by autophagy upon nitrogen-starvation [44]. We followed degradation of the fatty acid subunit Fas2-GFP in SD-N and found it to be severely affected in the absence of Hsv2 (Figure 4(E,F)). Together, this shows the important role of Hsv2 for autophagy of bulky cargoes.

Figure 4.

Hsv2 affects autophagy of bulky cargoes. (A) Maturation of giant prApe1 complexes generated by overexpression from a CUP1-promoter with 100 µM CuSO₄. Strains were grown in selective media lacking histidine to the log phase and then starved of nitrogen in SD-N. Samples were taken as indicated and separated by western blots. (B) Quantification of six independent experiments was done using Fiji and represented as bar graphs using GraphPad Prism 9. Values were normalized to 2 h of the wt. The error bars are SEM. (C) Degradation of Rpl9b-GFP, a component of the large ribosomal subunit. Cells grown in selective media to the log growth phase were starved in SD-N and samples were taken as indicated and processed in Western blots. (D) Quantification of these blots was done as in (B). The graph depicts a ratio of free GFP to the total GFP signal. The wt ratio after 24 h was set to 100%. The error bars represent SEM. (E) Degradation of Fas2-GFP, a subunit of the fatty acid synthase. Cells grown in selective media to the log phase were starved in SD-N and samples were taken as indicated and processed in Western blots. (F) Quantification of these blots was done as in (B). The graph depicts a ratio of free GFP to the total GFP signal. The wt ratio after 24 h was set to 100%. The error bars represent SEM.

The autophagic function of Hsv2 depends on interaction with Atg2, but not on its membrane bending ability

To substantiate our model that Hsv2 recruits more Atg2 or strengthens its binding to the phagophore to enhance membrane influx, we next analyzed Hsv2-mutants unable to bind Atg2. Studies on WDR45B/WIPI3 and WDR45/WIPI4, putative mammalian orthologs already identified residues required for interaction with mammalian ATG2A [33,35], among them mutations of patients suffering from BPAN. Furthermore, with the yeast 2-hybrid system an interaction defect of Atg2 with Hsv2^N78G and Hsv2^D100G was reported [45]. We verified this in co-immunoprecipitations, as expected Hsv2-GFP clearly coprecipitated with Atg2-3xHA, while Hsv2^N78K-GFP failed to interact with Atg2-3xHA. Hsv2^N78K was expressed at levels similar to Hsv2 (Figure 5(A,B)), in contrast both Hsv2^D100G and Hsv2^N78K,D100G were only present at very low levels (Figure 5(A)). Accordingly, the precipitated amounts of Atg2-3xHA were also low for these mutants (Figure 5(B,C)), which most likely does not reflect impaired interaction but rather the low expression level. We conclude that the interaction of Hsv2 with Atg2 mainly depends on Hsv2-N78. Using maturation of the giant prApe1-complex generated by prApe1-overexpression, we evaluated the autophagic activities of these Hsv2-mutants. Hsv2^N78K was unable to fully complement the maturation defect seen in hsv2∆ cells, confirming that the interaction of Hsv2 with Atg2 is required (Figure 5(C,D)). The Hsv2^I325D,L321K mutant, defective in membrane bending via its loop 6C/D, complemented the maturation defects of hsv2∆ cells, indicating that the Hsv2 autophagic function does not require its membrane bending function. This is in agreement with findings showing that the autophagic function of Atg18 also does not depend on its membrane bending ability [28].

Figure 5.

Autophagy of bulky cargoes by Hsv2 depends on its interaction with Atg2. (A) Co-immunoprecipitation of cells expressing Atg2-3xHA and the indicated Hsv2-GFP mutants with anti-GFP beads. As a negative control, GFP alone was included. The strains were grown in selective media to the early stationary phase. Hsv2^N78K was unable to bind Atg2. Hsv2-GFP variants containing the D100G mutation were only expressed at very low levels. (B) Quantification of four co-immunoprecipitations of (a) was done using Fiji and represented as a bar graph using GraphPad Prism 9. The graph depicts a ratio of Atg2-3xHA to the Hsv2-GFP signal and is normalized to the wt. The error bars are the standard error of the mean. (C) Maturation of the giant prApe1-complex generated by expression from a CUP1-promoter in the presence of 100 µM CuSO₄. Cells were grown in selective media lacking histidine and uracil and harvested in the log phase and starved in SD-N. Samples were taken as indicated and separated by western blots. (D) Quantification of at least three blots was done using Fiji and represented as bar graphs using GraphPad Prism 9. While the Hsv2^N78K mutant unable to bind Atg2 was inactive, the Hsv2^L321K,I325D mutant with defects in membrane binding was active. The error bars are the standard error of the mean.

Interaction of Hsv2 with Atg2 is dispensable for retromer-binding and hyperosmotic vacuole fragmentation

Having established the role of N78 of Hsv2 for Atg2 binding, we analyzed its importance for hyperosmotic vacuole fragmentation and retromer binding. As shown in Figure 6(A), Hsv2^N78K complemented the vacuole fragmentation defect of hsv2∆ cells. We again included Hsv2^D100G and Hsv2^N78K,D100G mutants, but expect that their low expression levels are the main reason for the observed lack of fission activity.

Figure 6.

Vacuole fragmentation does not require interaction of Hsv2 with Atg2, but its amphipathic helix in loop 6C/D. Hsv2 and Atg18 use different residues for binding to the CSC of retromer. (A) vps5∆ atg18∆ hsv2∆ cells of the BY4742 background transformed with plasmids expressing the indicated Hsv2 variants from the endogenous promoter were grown in selective medium to log phase (OD₆₀₀ of 0.5–1). After incubation with fm 4–64 for 30 min, they were incubated in YPD with 0.4 M NaCl for 1 h. (B) Quantification of five independent experiments with 200 cells each was done. Statistical relevance was determined by an unpaired two-tailed t-test. Asterisks indicate p-values: * for p < 0.05. Error bars are the standard error of the mean. (C) Co-immunoprecipitation with anti-GFP beads using the strain WCG hsv2∆ chromosomally expressing VPS35-6xHA and the indicated Hsv2 variants from a plasmid with the endogenous promoter. As negative control, GFP alone was used. The strains were grown in selective media to the early stationary phase. (D) Quantification of the blots from (C) shows normal binding of both Hsv2^N78K and Hsv2^T77E to Vps35. The graph depicts a ratio of Vps35-6xHA to the Hsv2-GFP signal and is normalized to Hsv2-GFP. The error bars are the standard error of the mean. Hsv2^D100G containing variants are only expressed in low amounts, which makes interpretation difficult. (E) Atg18 binds the retromer CSC via T56, sequence alignment identifies Hsv2-T77 as the corresponding residue.

Hsv2-N78 within loop 2A/B corresponds to Atg18-S57 (Figure 6(B)). Interestingly, an Atg18^T56E mutant is unable to interact with retromer [24]. This raises the question if Hsv2-T77, which corresponds to Atg18-T56 (Figure 6(B)) also plays a role in retromer binding. We found that Hsv2^T77E can restore the vacuole fragmentation defect of vps5∆ atg18∆ hsv2∆ cells similar to wild-type Hsv2 (Figure 6(A)), and in co-immunoprecipitations normally interacts with Vps35 (Figure 6(C,D)). This shows that Hsv2 and Atg18 use different residues for retromer binding. Consistent with the intact function in vacuole fragmentation, co-immunoprecipitations also confirmed intact interaction of Hsv2^N78K with Vps35-Ha (Figure 6(C,D)). Again, the interpretation of the interactions of Hsv2^D100G and Hsv2^N78K,D100G is hampered by their low expression level. Together our data shows that N78 of Hsv2 is only required for interaction with Atg2, but not with Vps35 of retromer.

Hsv2 affects the formation of the spore wall

Hsv2 has been identified in a screen of homozygous diploid deletion mutants for defects in spore wall organization [46]. During sporulation, which occurs when diploid cells are starved for nitrogen in the presence of a non-fermentable carbon source, four haploid nuclei surrounded by a double-membraned prospore membrane are formed [47]. Then the spore wall is built up between the inner and outer prospore membranes in a consecutive manner. The spore wall consists from the in- to the outside of a mannan, a β-1,3-glucan, a chitosan and a N-N-bisformyl-dityrosine containing layer [47]. Formation of the dityrosine layer depends on the correct assembly of the underlying layers. The dityrosine is formed in the spore cytoplasm and exported into the prospore membrane. Upon incorporation into the spore wall about half of the L,L-dityrosine is converted into D,L-dityrosine. hsv2∆ diploids showed a significantly lower level of D,L-dityrosine in total hydrolyzates and an increased level of dityrosine in the soluble fraction pointing to defects in proper assembly of the outermost dityrosine layer [46]. Core autophagy mutants are generally unable to sporulate, but hsv2∆ diploids sporulated wild-type like (Figure 7(A)). Defects in the dityrosine layer can be detected by Calcofluor white staining [48]. Calcofluor white fluorescently stains chitosan, but in wild-type spore walls the intact dityrosine layer prevents its access to the underlying chitosan layer. Indeed, spores derived from hsv2∆ cells showed significantly more calcofluor white staining compared to wild-type cells (Figure 7(B,C)). This phenotype was complemented by Hsv2, but not the membrane bending-defective Hsv2^I325D,L321K mutant (Figure 7(C)). This mutant only affected the Hsv2-retromer, but not its autophagic function. We thus speculated that Hsv2-dependent protein sorting might be required for normal formation of the dityrosine layer of the spore wall. This prompted us to include diploid cells in our attempt to uncover a putative role of Hsv2-retromer in protein sorting.

Figure 7.

Hsv2 and its amphipathic helix within loop 6C/D are required for spore wall biogenesis. (A) BY4743 wt and hsv2∆ diploid cells were transformed with either an empty plasmid, Hsv2-GFP or Hsv2^L321K,I325D-GFP. Cells were grown to stationary phase (OD600 of 4) and incubated in enriched sporulation medium for 3 days. The percentage of diades, triades and tetrades was counted in four experiments. Absence of Hsv2 did not affect sporulation. (B) Samples from (A) were stained with calcofluor white 30 min before microscopic visualization. (C) Percentage of calcofluor-stained spore walls. Both absence of Hsv2 or the Hsv2^L321K,I325D mutant cause defects in spore wall assembly indicated by enhanced calcofluor staining. Four independent experiments were conducted with 100–200 cells each. Statistical relevance was determined by an unpaired two-tailed t-test. Asterisks indicate p-values, error bars indicate SEM.

Hsv2 interacts with Pep12 and affects its sorting in diploid, but not haploid cells

Cargo proteins typically interact with their sorting machinery. We thus searched for putative interaction partners of Hsv2. From previous studies we had a collection of constructs for the split-ubiquitin system. This is similar to the yeast 2-hybrid system, but better suited to detect interactions between membrane associated proteins. In an unbiased approach, we screened these constructs and detected a possible interaction of Hsv2 with Pep12 (Figure S4). Co-immunoprecipitation of Pep12-HA with GFP-Hsv2, but not with GFP alone confirmed this interaction (Figure 8(A,B)). The SNARE Pep12 localizes to endosomes and to some extent to the vacuole membrane [49]. In non-starved and starved haploid wild-type and haploid hsv2∆ cells GFP-Pep12 was mostly absent from the vacuole lumen. However, almost 80% of all diploid hsv2∆ cells showed mislocalization of GFP-Pep12 to the vacuole lumen (Figure 8(C,D)). Expression of wild-type, but not of the membrane-bending defective mutant Hsv2^I325D,L321K complemented this mislocalization phenotype. Together, our experiments point to a role of Hsv2 in protein sorting in diploid cells, which depends on its ability to bend membranes with its loop 6C/D.

Figure 8.

The SNARE Pep12 is mislocalized to the vacuole lumen in diploid hsv2∆ cells, but not in hsv2∆ haploids. Pep12 sorting depends on the amphipathic helix in the loop 6C/D of Hsv2. (A) Co-immunoprecipitation of GFP-Hsv2 with Pep12-HA using anti-GFP beads. (B) Quantification of (A) Free GFP was included as negative control. (B) Haploid wt and hsv2∆ cells (left panel) or BY4743 wt and diploid hsv2∆ (right panel) cells were transformed with GFP-Pep12. Cells were grown in selective medium to log phase (OD₆₀₀ of 0.5–1) and imaged (0 h SD-N) before they were transferred to SD-N medium for 2 h. (C) BY4743 wt and diploid hsv2∆ cells were transformed with GFP-Pep12 and either an empty plasmid, Hsv2-3xHA or Hsv2^L321K,I325K-3xHA. After 2 h in SD-N, the number of cells with GFP signal in their vacuoles was counted. GFP-Pep12 missorting into the vacuole in hsv2∆ diploid cells was complemented by Hsv2, but not by Hsv2^L321K,I325D, indicating the importance of the amphipathic helix in loop 6C/D of Hsv2 for its sorting function. Five independent experiments were conducted with 100–200 cells each. Statistical relevance was determined by an unpaired two-tailed t-test. Asterisks indicate p-values: ns, not significant p > 0.05. * for p < 0.05. ** for p < 0.01. *** for p < 0.001. **** for p < 0.0001. Error bars indicate SEM.

Discussion

Almost 25 years ago, Atg18 was identified as a core autophagy protein in S. cerevisiae [50,51], while it became clear that its homolog Atg21 also plays a role in autophagy, the functional relevance of the third family member Hsv2 so far remained elusive. Atg18 scaffolds the Atg2-Atg9 complex at the phagophore edges to allow non-vesicular membrane influx for phagophore expansion [5–7]. We now found the requirement of Hsv2 in the autophagic removal of bulky natural cargoes such as the large ribosomal subunit Rpl9b and the fatty acid synthase subunit Fas2 (Figure 4(A-F)). As shown with the Hsv2^N78K mutant, this role depends on the interaction of Hsv2 with Atg2 (Figure 5(A-D)). What might be the molecular function of Hsv2 during autophagy of bulky cargos? Atg24/ Snx4 has been shown to localize to the phagophore rim in ring-like structures dependent on its BAR and PX domains [43,52]. Here, it widens the phagophore rim to allow incorporation of large cargo. We do not expect such a function for Hsv2 for two reasons. First, similar to the localization of Atg18, we detected Hsv2 not in ring-like structures at the phagophore rim, but found it restricted at both phagophore edges (Figure 3(C-F)). Secondly, the Atg18-Atg2 complex was reported to promote constriction of the phagophore rim [43]. Together, we favor the idea that similar to Atg18, Hsv2 scaffolds a complex of Atg2 with Atg9 and thus enhances membrane transfer to allow engulfment of very large cargos. But why do we need Hsv2 in addition to Atg18, if they share rather similar functions? A strong hint comes from the finding that Atg18 and Hsv2 use distinct residues for their interaction with Atg2. Hsv2-N78 within loop 2A/B is crucial for binding Atg2. Hsv2-N78 corresponds to Atg18-S57, but the interaction sites of Atg18 with Atg2 have been mapped to Atg18-P72R73 and Atg18-T90FPTS94 [53,54]. Due to the use of two distinct interaction residues Atg18 and Hsv2 can either individually recruit more Atg2 to the phagophore or could interact simultaneously with Atg2. Indeed, for the recruitment of the Atg18-Atg2-Atg9 complex to phagophores PtdIns3P binding by Atg18 is not sufficient. Additionally, an amphipathic helix of Atg2 and interaction of Atg2 and Atg9 is required [16,17]. Cooperative binding of Atg18 and Hsv2 to Atg2 thus could enhance and stabilize complex recruitment and its membrane transfer activity. Cooperative action of Atg18 and Hsv2 is further supported by the requirement of Atg18 for efficient recruitment of Hsv2 to phagophores and the increased expression level of Hsv2 in the absence of Atg18 (Figures 1(B) and 3(A,B)).

A role of Hsv2 in engulfment of large cargos is further supported by its involvement in micronucleophagy (piecemeal microautophagy of the nucleus) [40], where non-essential parts of the nucleus are engulfed into vacuolar invaginations. Indeed, the intermediate structures of micronucleophagy have a quite large diameter of 580–770 nm [55,56]. Since absence of Hsv2 has only mild effects on micronucleophagy, we did not follow this here. Hsv2 has further been implicated in an Atg5- and Atg7-independent type of alternative autophagy [57,58]. This Golgi-membrane-associated degradation pathway (GOMED) can be induced by amphotericin B, which disturbs the transport from the Golgi to the plasma membrane.

Recently, we and others detected Atg18 to form a novel retromer complex together with the CSC of retromer, which consists of Pep8/Vps26, Vps29 and Vps35 [23,24]. Using co-immunoprecipitations, we also found that Hsv2 interacts with the core retromer component Vps35 (Figure 1(A,C)). The role of Atg18-retromer in vacuole fission during hyperosmotic stress is thought to mainly depend on PtdIns(3,5)P₂. Accordingly, cells lacking the PtdIns3P 5-kinase Fab1 almost completely lack fragmented vacuoles [12,29]. In contrast, atg18∆ cells still showed some residual vacuole fragmentation suggesting the existence of an additional fission machinery. Deletion of HSV2 alone did not affect vacuole fragmentation [40], we thus speculated that its putative function in vacuole fission might be masked by Atg18. In atg18∆ cells only few cells show vacuole fragmentation making it experimentally difficult to identify another fission machinery. We thus choose a more sensitive alternative approach. Atg18 and Vps5 bind competitively to the retromer CSC [23,24], to shift this equilibrium toward the formation of Atg18-retromer, we used a vps5∆ background. Deletion of ATG18 in vps5∆ cells only had minor effects on the strong vacuolar fragmentation phenotype (Figure 1(D-F)). Additional deletion of HSV2, but not of ATG21 significantly reduced vacuole fragmentation. This indicates a cooperative role of Atg18 and Hsv2 in vacuole fission. Since the single deletion of ATG18, but not of HSV2 affects vacuole fission, we expect preferential binding of Atg18 to the retromer CSC. In line with such a fission activity, both Atg18 and Hsv2 bind PtdIns(3,5)P₂ with higher affinity than PtdIns3P, while Atg21 shows less affinity to PtdIns(3,5)P₂ [59]. Again, this cross-talk between Atg18 and Hsv2 is strengthened by a ~ 50% increase in the Hsv2-level in atg18∆ cells and the two-fold increase in binding of Hsv2 to Vps35 in atg18∆ vps5∆ cells (Figure 1(B,C)).

Once more, a major question is why Hsv2 is needed in addition to Atg18 in the formation of retromer complexes? As for the situation during autophagy, we expect the major reason is the use of distinct residues for interaction of Atg18 and Hsv2 with the CSC of retromer. Atg18^T56E within the loop 2A/B is unable to interact with the retromer CSC, but its autophagic function is almost normal [24]. Atg18-T56 corresponds to the conserved Hsv2-T77 (Figure 6(B)). However, Hsv2^T77E binding to Vps35 was unaffected and it was functional during hyperosmotic vacuole fragmentation (Figures 5(A,B) and 6(A,B)). As discussed above, Hsv2-N78 corresponding to Atg18-S57 within the same loop 2A/B mediates interaction with Atg2. In sum, Atg18 binds with its loop 2A/B to the retromer CSC, while the same loop in Hsv2 mediates binding to Atg2. We propose that the use of different binding interfaces of Atg18 and Hsv2 to retromer CSC modulates the geometry of the resulting PROPPIN-retromer assemblies on membranes. For example, Atg18 oligomerizes and forms helical tubes reminiscent of the COPII outer coat [60]. Furthermore, different binding modes of Atg18 and Hsv2 result in distinct exposure of additional interaction sites allowing interaction with additional proteins. This would fit with mammalian retromer complexes which interact with multiple partners such as the WASH complex [25]. This could also include interactions with cargo receptors, we thus expect that Hsv2-retromer has sorting functions distinct from Atg18-retromer.

Identification of cargos for a novel retromer complex is not easy, we thus first focused on the known effect of Hsv2 on the assembly of the outermost dityrosine layer of the spore wall [46]. In contrast to diploids lacking the core autophagy protein Atg18, diploid hsv2∆ cells form normal rates of ascospores (Figure 7(A)), but their increased calcofluor staining indicates defects in the formation of the dityrosine layer (Figure 7(B,C)). The spore wall layers are formed consecutively and disturbed assembly of an inner layer also affects the proper formation of outer layers [47]. We thus speculated that Hsv2 might affect protein sorting in diploid cells. To find hints for putative cargoes we tested with the split-ubiquitin system a collection of constructs, which were already available in the lab and found a putative interaction of Hsv2 with the SNARE Pep12. With co-immunoprecipitations we could confirm the interaction between Hsv2 and Pep12 (Figure 8(A,B)). Indeed, Pep12 mislocalization dependent on the presence of Hsv2 was observed in diploid, but not in haploid cells (Figure 8(C,D)). Together, these experiments suggest a role of Hsv2 in protein transport especially in diploid cells.

A key feature of PROPPINs is their ability to induce membrane tubulation and fission dependent on an amphipathic helix formed in their loop 6C/D, which is thought to partially penetrate into the membrane [24,28,29]. We here generated a Hsv2^I325D,L321K mutant (Figure 2(A)), where the formation of this amphipathic helix is disturbed. Interestingly, similar to Atg18 the membrane bending activity of Hsv2 was not required for its autophagic function (Figure 5C,D). However, for vacuole fragmentation (Figure 2(C,D)), normal formation of the spore wall (Figure 7(C)) and Pep12 sorting in diploids (Figure 8), the amphipathic loop 6C/D helix of Hsv2 was required. This is well understandable, since vacuole fragmentation and the sorting functions should involve membrane fission events. For the autophagic function of Atg18 and Hsv2 on the other hand, the presence of a carboxyterminal amphipathic helix of Atg2 might be sufficient to allow localization of the Atg18-Atg2-Atg9 complex to the highly curved edges of the phagophore [16,17].

How incorporation of Hsv2 in retromer and/or Atg2 complexes is regulated is unclear at this point. However, it is tempting to speculate that post-translation modifications might be involved. In fact, phosphorylation of Atg18 has been observed [61].

Atg18 and Hsv2 both interact with Atg2, which is reminiscent to WDR45B/WIPI3 and WDR45/WIPI4, which thus should be considered as counterparts. Mutations in WDR45/WIPI4 cause the neurodegenerative disease beta-propeller protein-associated neurodegeneration (BPAN), formerly called static encephalopathy of childhood with neurodegeneration in adulthood/SENDA, which is accompanied by iron accumulation. Many WDR45/WIPI4 mutations affect its binding to ATG2A for example the WDR45/WIPI4^N61K mutation, whose corresponding Hsv2^N78K mutation we used here [14,33,35]. However, most likely not defects in the autophagy function of WDR45/WIPI4 are responsible for this disease. Instead mislocalization of ATG2A, which induces ferroptosis independent of autophagy has been identified as a major problem [38,39]. More mechanistic insights into the role of WDR45/WIPI4 are necessary to fully understand the development of this disease. Hopefully, studies on the yeast Hsv2 as a model can contribute to this. Indeed, our study uncovered that Hsv2^D100G corresponding to the disease-associated WDR45/WIPI4^D84G mutant might not per se be unable to interact with Atg2, but is only present in very low amounts. Further analyses of putative retromer associated roles of WDR45/WIPI4 might help to gain a deeper insight into the development of BPAN.

Materials and methods

Strains, plasmids and oligonucleotides

Yeast strains listed in Table 1 are derived from the WT strain WCG4a MATα his2–11,15 leu2–3, 112 ura3 [62]. Deletion strains and chromosomal integrations were generated using the method described by [63]. PCR of genomic DNA verified the correct integration or deletion of the target gene.

Table 1.

A list of Saccharomyces cerevisiae strains used in this study.

Strain	Genotype	Reference
WCG4a (WT)	MATα his2–11,15 leu2–3, 112 ura3	[55]
hsv2∆	hsv2:kanMX4	[34]
hsv2∆ Pep12-6xHA	hsv2∆::kanMX4 Pep12-6xHA:hphNT1	This work
hsv2∆ Vps35-6xHA	hsv2∆::kanMX4 Vps35-6xHA:natNT2	This work
atg18∆ hsv2∆ Vps35-6xHA	atg18∆::kanMX4 hsv2∆::hphNT1 Vps35-6xHA:natNT2	This work
vps5∆ hsv2∆ Vps35-6xHA	vps5∆::hisMX6 hsv2∆::kanMX6 Vps35-6xHA:natNT2	This work
atg18∆ vps5∆ hsv2∆ Vps35-6xHA	atg18∆::kanMX6 vps5∆::hisMX6 hsv2∆::hphNT1 Vps35-6xHA:natNT2	This work
Atg2-3xHA hsv2∆	Atg2-3xHA:hphNT1 hsv2∆::natNT2	This work
BY4741 (WT)	met15∆	Euroscarf, Y00000 WT Mat a
vps5∆	vps5∆::kanMX6
atg18∆ vps5∆	atg18∆::kanMX6 vps5∆::natNT2	This work
atg18∆ atg21∆ vps5∆	atg18∆::kanMX6 atg21∆::hphNT1 vps5∆::natNT2	This work
atg18∆ atg21∆ hsv2∆ vps5∆	atg18∆::kanMX6 atg21∆::hphNT1 hsv2∆::hisMX6 vps5∆::natNT2	This work
atg18∆ hsv2∆ vps5∆	atg18∆::kanMX6 hsv2∆::hisMX6 vps5∆::natNT2	This work
atg21∆ hsv2∆ vps5∆	atg21∆::hphNT1 hsv2∆::hisMX6 vps5∆::natNT2	This work
atg21∆ vps5∆	atg21∆::hisMX6 vps5∆::natNT2	This work
hsv2∆ vps5∆	hsv2∆::hisMX6 vps5∆::natNT2	This work
BY4743 (WT)		WT diploid Euroscarf Y20000
hsv2∆	hsv2∆::kanMX6	Euroscarf
Fas2-yeGFP	FAS2-yeGFP:hphNT1	This work
Fas2-yeGFP hsv2∆	FAS2-yeGFP:hphNT1 hsv2∆::kanMX6	This work
Rpl9b-yeGFP	RPL9b-yeGFP:hphNT1	This work
Rpl9b-yeGFP hsv2∆	RPL9b-yeGFP:hphNT1 hsv2∆::kanMX6	This work

Strains requiring auxotrophic selection were grown in a complete minimal medium (CM) containing 0.67% [w:v] yeast nitrogen base without amino acids (Becton Dickinson, 291,920), 2% [w:v] glucose (Roth, 6780.2) set to pH 5.6 and supplemented with the appropriate amino acids.

To induce autophagy, cells were harvested, washed and incubated in SD-N medium containing 0.17% [w:v] yeast nitrogen base without amino acids and ammonium sulfate (Becton Dickinson, 233,520) and 2% [w:v] glucose (Roth, 6780.2).

Plasmids used are listed in Table 2.

Table 2.

A list of plasmids used in this study.

Strain	Genotype	Reference
Hsv2-GFP	pRS426-Hsv2P-Hsv2-GFP-CYC1T	This work
mCherry-Atg8	pRS315-ATG8P-mCherry-Atg8-ATG8T	[15]
Ape1	pRS423-CUP1P-Ape1-CYC1T	[15]
Hsv2 Hsv2-I325D,L321K	pUG36-HSV2P-Hsv2 Hsv2I325D,L321K -CYC1T	This work
Hsv2N78K	pUG36-HSV2P-Hsv2N78K-CYC1T	This work
Hsv2D100G	pUG36-HSV2P-Hsv2D100G-CYC1T	This work
Hsv2N78K D100G	pUG36-HSV2P-Hsv2N78K,D100G-CYC1T	This work
Hsv2N78K-GFP	pRS426-HSV2P-Hsv2N78K-GFP-CYC1T	This work
Hsv2D100G-GFP	pRS426-HSV2P-Hsv2D100G-GFP-CYC1T	This work
Hsv2N78K,D100G-GFP	pRS426-HSV2P-Hsv2N78K,D100G-GFP-CYC1T	This work
Hsv2T77E-GFP	pRS426-HSV2P-Hsv2T77E-GFP-CYC1T	This work
pUG36	CEN/ARS AmpR ori lacZ’ URA3	[56]
Hsv2-3xHA	pRS315- HSV2P-Hsv2-3xHA-HSV2T	This work
Hsv2I325D,L321K −3xHA	pRS315- HSV2P-Hsv2I325D,L321K −3xHA-HSV2T	This work
Hsv2-GFP	pUG35-HSV2P-Hsv2-GFP-CYC1T	This work
Hsv2-I325D,L321K -GFP	pUG35-HSV2P-Hsv2I325D,L321K -GFP-CYC1T	This work
pYM25	AmpR ori yeGFP TEFP-HygR-CYC1T	[57]

The Oligonucleotides used in this study are shown in Table 3. Mutations were introduced into the endogenous plasmid using the QuikChange Lightning Multi Site-Directed Mutagenesis Kit (Agilent Technologies, 210,515-5). This was achieved following the manufacturer’s protocol. Chromosomal knockouts and C-terminal tagging were generated using the method described by [63].

Table 3.

A list of DNA oligonucleotides used in this study.

Name	Sequence	Application
Hsv2N78K Forward	CCAGGATGCTTTATCGTACGAAgTACATCGCACTAGTCGG	N78K mutation in Hsv2
Hsv2N78K Reverse	CCGACTAGTGCGATGTAcTTCGTACGATAAAGCATCCTGG	N78K mutation in Hsv2
Hsv2D100G Forward	GCTCTAAATAAACTGATCATCTGGGgTGATCTTTTGCAAAAGG	D100G mutation in Hsv2
Hsv2D100G Reverse	CCTTTTGCAAAAGATCAcCCCAGATGATCAGTTTATTTAGAGC	D100G mutation in Hsv2
Hsv2T77E Forward	TCCGACTAGTGCGATGTAATTctcACGATAAAGCAT	T77E mutation in Hsv2
Hsv2T77E Reverse	ATGCTTTATCGTgagAATTACATCGCACTAGTCGGA	T77E mutation in Hsv2
Hsv2 Hsv2I325D,L321K Forward	GTTCGAGCCATCCCaagAAAAATTATgacCCGAAGGGCCTATGG	Loop mutation in Hsv2
Hsv2 Hsv2I325D,L321K Reverse	CCATAGGCCCTTCGGgtcATAATTTTTcttGGGATGGCTCGAAC	Loop mutation in Hsv2
Hsv2 S1	GGCAGCGATTATTAGAGGACAACTATAAGCATACATAACTAGCAGATGcgtacgctgcaggtcgac	Deletion of HSV2
Hsv2 S2	CGTAAATGCACACTTTCTCTATACATATATATATATTTATATTCATGTTAatcgatgaattcgagctcg	Deletion of HSV2
Atg18 S1	GTAATAGTGTTCCAGTTAACTCTGTATCCTTTTCTTCTTCGGCCTGACAATGcgtacgctgcaggtcgac	Deletion of ATG18
Atg18 S2	GTGTATGCGTTGTGACGTACGGAAGGCAGCGCGAGACACTTCCGTGATCAatcgatgaattcgagctcg	Deletion of ATG18
Atg21 S1	CAAAAGACAATTCCACTCCTTTGGATTTGAAATAGACAGATAGAAAAGGATATGcgtacgctgcaggtcgac	Deletion of ATG21
Atg21 S2	CGTGAATACGTACAATATCTATTAAGATTATGAAAACTGCACATATGCATTAatcgatgaattcgagctcg	Deletion of ATG21
Vps5 S1	GCAGCAGGGATTTTATAAACTTTCATACATCCTGCAATAACAAGCCATGcgtacgctgcaggtcgac	Deletion of VPS5
Vps5 S2	CATAAATCCTGAGGAACGTGACACATAAAGTTATTGTATACAGATCATCTAatcgatgaattcgagctcg	Deletion of VPS5
Vps35 S2	GTGTAGTTTTTTTTTATCTTGGGCATGTACGAAGAGCAAGTACGTTATTTAACTAatcgatgaattcgagctcg	C-terminal tagging of Vps35
Vps35 S3	CATTGAAAGTCAAAGAGAAGTTGACGATCGTTTCAAAGTCATATATGTAcgtacgctgcaggtcgac	C-terminal tagging of Vps35
Rpl9b S2	CATATTTAAGCGCCTCTAAGCGCAAAGACACTTTTTTGTTCCATTTAatcgatgaattcgagctcg	C-terminal tagging of Rpl9b
Rpl9b S3	GATGGTATCTACGTTTCCCACAAGGGTTTCATTGTCGAAGACATGcgtacgctgcaggtcgac	C-terminal tagging of Rpl9b
Fas2 S2	CGACACGTTACATATTAAAAGAGGGACTACGTAGTGCTCTCTCTAatcgatgaattcgagctcg	C-terminal tagging of Fas2
Fas2 S3	CACGATGACCTCCAAGCTGTCGCGGTCGCCGTTTCTACTAAGAAAcgtacgctgcaggtcgac	C-terminal tagging of Fas2
PEP12_S2	TATATTATGTATATGATATTTGACGACGTGTGTTGGTTTGGTTTAatcgatgaattcgagctcg	C-terminal tagging of Pep12
PEP12_S3	GTGCTTCTCGTAATGCTTCTTTTTATTTTTCTCATTATGAAATTGcgtacgctgcaggtcgac	C-terminal tagging of Pep12

Antibodies

The anti-Ape1 antibody was used as described by [64].

The other primary antibodies used were anti-GFP (from mouse IgG1κ; Roche, 11,814,460,001) and the HA-probe antibody (F-7, mouse monoclonal IgG2a; Santa Cruz Biotechnology, sc-7392). The secondary antibodies used were horseradish peroxidase-conjugated goat anti-rabbit IgG (Thermo Fisher Scientific, G-21234) and horseradish peroxidase-conjugated goat anti-mouse IgG (Dianova, 115–035-166).

Co-immunoprecipitation

Strains were grown in a complete minimal medium until the early stationary phase, OD₆₀₀ 2.5–3. 200 OD₆₀₀ units were harvested (500 ×g, 5 min, 4°C), and pelleted cells were washed with 50 mM Tris HCl, pH 7.5 before centrifuging again. The pellet was resuspended in lysis buffer (50 mM Tris HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA [Roth, 8043.1], 0.5% [w:v] Tween 20 [Sigma-Aldrich, P7949], 1 mM PMSF [Roth, 6367.1], 1x cOmplete^TM [EDTA-free Protease Inhibitor Cocktail; Roche, 05056489001], 0.1% [w:v] aprotinin [Merck, 616,370], 0.1% [w:v] pepstatin [Merck, 516,481], 0.1% [w:v] leupeptin [Merck, 108,975], 0.1% [w:v] chymostatin [Merck, 230,790]) before being subjected to glass bead lysis at 4°C for 30 min. To remove cell debris, samples were centrifugated at 10,000 ×g for 10 min at 4°C and the supernatant, with a final volume of approximately 800 µl, was transferred to a fresh microcentrifuge tube and a 25-µl load sample was collected. Fifty µl of µMACs anti-GFP micro beads (Miltenyi Biotec, 130–091-125) were added to each sample and incubated on ice for 30 min. Equilibration of the columns (Miltenyi Biotec, 130–042-701) was done with the lysis buffer before samples were loaded onto the columns. The columns were washed three times with lysis buffer equal to the sample volume and once with 50 mM Tris HCl, pH 7.5. Proteins were eluted from the column in two steps, first, 20 µl 2x Laemmli buffer heated to 95°C and incubated for 5 min and second 50 µl 2x Laemmli buffer heated to 95°C. Eluted proteins and the load sample were boiled at 95°C before loading onto an SDS gel, 10 µl of the load sample (≈0.75% of the lysate) and 10 µl of the bound sample were loaded.

Quantitative degradation analysis of Rpl9b-GFP and Fas2-GFP

Cells expressing Rpl9b-GFP or Fas2-GFPwere grown YPD medium to the log growth phase, harvested, washed with SD-N medium and then incubated at 30°C in SD-N medium. Two OD₆₀₀ units of the cell culture were harvested after 0, 6 and 24 h of nitrogen starvation. The protein extracts were lysed using an alkaline lysis buffer (0.28 M NaOH, 1.125% [v:v] β-mercaptoethanol) followed by TCA precipitation. The samples were loaded onto a 10% SDS gel for separation and analyzed by immunoblotting with antibodies against GFP. For visualization and quantification of the PVDF membranes Amersham ImageQuant 800 and Fiji software were used.

Overexpression of Ape1

Cells expressing CUP1 promoter-driven Ape1 were grown in CM medium without histidine and 100 µM CuSO₄ to the log growth phase, harvested, washed with SD-N medium and then incubated at 30°C in SD-N medium. Two OD₆₀₀ units of the cell culture were harvested after 0, 1 and 2 h of nitrogen starvation. The protein extracts were lysed using an alkaline lysis buffer (0.28 M NaOH, 1.125% [v:v] β-mercaptoethanol) followed by TCA precipitation. The samples were loaded onto a 15% SDS gel for separation and analyzed by immunoblotting with antibodies against Ape1. For visualization and quantification of the PVDF membranes Amersham ImageQuant 800 and Fiji software were used.

Delta vision

Fluorescence microscopy was performed using the DeltaVision® microscope (Olympus IX71, Applied Precision) equipped with the UPlanSApo x100, 1.4 numerical aperture, oil immersion objective and a CoolSNAP_HQ2^™ couple-charged device (CCD) camera. Imaging occurred with a 100x objective and a 2 × 2 or 1 × 1 binning. At least 20 focal planes along the z-axis with a distance of <0.2 µm were captured. The resulting images were deconvolved using softWoRX^™ (Applied Precision) and further processed with Fiji or Huygens Professional.

Vacuolar fragmentation

Cells were grown in YPD medium to the log growth phase. Five OD_{600 units} of the cell culture were harvested and resuspended in 1 ml of YPD with 5 μl FM 4–64 (Invitrogen, T3166). The cells were incubated at 30°C for 30 min and subsequently split into two subcultures. One was resuspended in 500 μl YPD, the other in 500 μl YPD +400 mM NaCl. The NaCl culture was incubated at 30°C for an additional 1 h. Cells were imaged using the DeltaVision® microscope (Olympus IX71, Applied Precision). The resulting images were deconvolved using softWoRX^™ (Applied Precision) and further processed with Fiji. Vacuolar fragmentation was determined by counting the vacuoles in each cell.

Cells with plasmids were grown in CM medium without uracil. After the incubation with FM 4–64 all cells were regenerated in YPD medium for 1 h before they were split into subcultures with and without NaCl.

Sporulation

Cells were grown in CM medium without certain amino acids overnight at 30°C. Cells were harvested, washed with enriched sporulation medium [46] and subsequently incubated in enriched sporulation medium at 30°C for 3 days. Before imaging the cells were treated with calcofluor white for 30 min. Cells were imaged using the DeltaVision® microscope (Olympus IX71, Applied Precision). The resulting images were deconvolved using softWoRX^™ (Applied Precision) and further processed with Fiji.

Multi-Site-Directed mutagenesis Hsv2^L321K,I325K

Hsv2 was mutated using the QuikChange Multi Site-Directed (Agilent Technologies, 210,518) using Mutation fwd: GTTCGAGCCATCCCAAGAAAAATTATGACCCGAAGGGCCTATGG and Mutation rev: CCATAGGCCCTTCGGGTCATAATTTTTCTTGGGATGGCTCGAAC.

GraphPad prism 9 and statistics

All graphs and statistics were done using GraphPad Prism version 9. All error bars are standard error of the mean (SEM). Asterisks indicate p-values: ns, not significant p > 0.05. * for p < 0.05. ** for p < 0.01. *** for p < 0.001. **** for p < 0.0001.

Funding Statement

This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Projectnumber 519482673; (Th752/8-1).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2026.2671338